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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d4329c61-d3e5-400f-9298-80d18c374a84
|
raft-rationale-adaptor-for-few-shot-abusive
|
2211.17046
| null |
https://arxiv.org/abs/2211.17046v1
|
https://arxiv.org/pdf/2211.17046v1.pdf
|
RAFT: Rationale adaptor for few-shot abusive language detection
|
Abusive language is a concerning problem in online social media. Past research on detecting abusive language covers different platforms, languages, demographies, etc. However, models trained using these datasets do not perform well in cross-domain evaluation settings. To overcome this, a common strategy is to use a few samples from the target domain to train models to get better performance in that domain (cross-domain few-shot training). However, this might cause the models to overfit the artefacts of those samples. A compelling solution could be to guide the models toward rationales, i.e., spans of text that justify the text's label. This method has been found to improve model performance in the in-domain setting across various NLP tasks. In this paper, we propose RAFT (Rationale Adaptor for Few-shoT classification) for abusive language detection. We first build a multitask learning setup to jointly learn rationales, targets, and labels, and find a significant improvement of 6% macro F1 on the rationale detection task over training solely rationale classifiers. We introduce two rationale-integrated BERT-based architectures (the RAFT models) and evaluate our systems over five different abusive language datasets, finding that in the few-shot classification setting, RAFT-based models outperform baseline models by about 7% in macro F1 scores and perform competitively to models finetuned on other source domains. Furthermore, RAFT-based models outperform LIME/SHAP-based approaches in terms of plausibility and are close in performance in terms of faithfulness.
|
['Animesh Mukherjee', 'Binny Mathew', 'Kushal Kedia', 'Divyanshu Sheth', 'Punyajoy Saha']
|
2022-11-30
| null | null | null | null |
['cross-domain-few-shot', 'abusive-language']
|
['computer-vision', 'natural-language-processing']
|
[ 7.72881880e-02 -1.57269686e-01 -5.58040440e-01 -2.10021734e-01
-1.17373192e+00 -4.60989684e-01 7.63587713e-01 3.29188645e-01
-3.82226944e-01 6.26711369e-01 3.84336710e-01 -1.51816845e-01
2.25613520e-01 -4.73134518e-01 -3.40052783e-01 -1.97889075e-01
4.07662749e-01 5.16255856e-01 4.08440053e-01 -6.71970725e-01
4.28057164e-01 4.99420799e-02 -9.25274670e-01 5.99743962e-01
6.58748388e-01 5.51873267e-01 -4.49406236e-01 5.37904978e-01
-4.62682068e-01 1.08909559e+00 -9.71379876e-01 -9.05688047e-01
-9.62814316e-02 -3.87075931e-01 -7.53037751e-01 -1.15835415e-02
6.94487095e-01 -4.34424043e-01 -6.69800937e-02 9.96457636e-01
5.98088622e-01 -2.94579193e-02 9.57252026e-01 -1.10607910e+00
-7.41930544e-01 9.79088247e-01 -9.83177125e-01 6.16140068e-01
4.17363465e-01 1.43324018e-01 1.31775403e+00 -9.78699863e-01
7.29080558e-01 1.57162654e+00 8.41471016e-01 9.30553854e-01
-1.29386055e+00 -9.53906298e-01 1.61415756e-01 9.34534241e-03
-1.16661584e+00 -6.63402259e-01 8.17471087e-01 -6.51168168e-01
1.00151479e+00 1.66563645e-01 1.43878952e-01 2.02879977e+00
-7.19516128e-02 8.70169818e-01 8.52348447e-01 -3.61731976e-01
1.41842246e-01 3.57150465e-01 4.97373700e-01 6.27281785e-01
4.33066279e-01 -5.26063323e-01 -7.92437494e-01 -6.65991604e-01
-2.70844046e-02 -9.92139429e-02 1.00211829e-01 3.20060328e-02
-5.57465851e-01 1.57932293e+00 1.70814365e-01 6.08412087e-01
-5.32484688e-02 4.14586700e-02 7.65725195e-01 2.23461166e-01
8.96423697e-01 6.56357169e-01 -1.15256213e-01 -4.15353626e-02
-1.16121793e+00 4.38170612e-01 7.80883849e-01 5.68219960e-01
2.54547209e-01 2.33302265e-02 -6.13046944e-01 1.46263826e+00
2.01899752e-01 2.62982637e-01 8.71980190e-01 -4.86375242e-01
7.15998769e-01 5.10644436e-01 6.69574738e-02 -9.41723347e-01
-4.21934485e-01 -3.42218429e-01 -2.46066213e-01 1.07553583e-02
5.29909909e-01 -1.28756955e-01 -8.12588751e-01 1.75546372e+00
9.72474888e-02 -9.97055396e-02 -1.14845119e-01 8.15227926e-01
9.03262734e-01 4.53299284e-01 3.59178811e-01 -2.39305779e-01
1.39939082e+00 -7.86257327e-01 -6.36764407e-01 -7.92853951e-01
1.12419689e+00 -7.88734019e-01 1.58978403e+00 3.85151029e-01
-7.23361790e-01 -2.17681173e-02 -1.19216084e+00 -1.18921660e-01
-3.95749688e-01 -3.44104558e-01 3.40660840e-01 1.00410056e+00
-4.81736034e-01 4.22017008e-01 -1.34900674e-01 -6.36794686e-01
7.35943615e-01 -2.88313687e-01 1.60440442e-03 -8.35293531e-02
-1.44318473e+00 1.09010267e+00 2.25404516e-01 -7.93118238e-01
-9.82927859e-01 -9.05033827e-01 -7.47730613e-01 -7.39298984e-02
4.54009831e-01 -2.29110867e-01 1.43393707e+00 -8.70126247e-01
-1.11612701e+00 1.47798324e+00 1.05887599e-01 -6.67038023e-01
8.84205699e-01 -4.78983104e-01 -4.55781460e-01 -1.65963069e-01
5.49065351e-01 3.76649022e-01 1.21008396e+00 -9.00350213e-01
-2.10542068e-01 -1.27648577e-01 -2.48168800e-02 -1.35496840e-01
-8.74691844e-01 5.95349312e-01 -3.27889100e-02 -6.69370949e-01
-4.80550051e-01 -7.34492362e-01 2.52254248e-01 -1.79891184e-01
-6.10890567e-01 -2.81786710e-01 8.41954648e-01 -7.43519604e-01
1.49423695e+00 -2.17710161e+00 -1.96021244e-01 -1.48645565e-01
3.79496425e-01 5.22677243e-01 1.22497678e-01 2.02007979e-01
-1.56128975e-02 5.53273141e-01 -3.21642697e-01 -6.74961925e-01
5.34581877e-02 7.93477669e-02 -5.89645207e-01 5.18417001e-01
3.24243963e-01 6.07926369e-01 -8.60304654e-01 -6.62108481e-01
-1.27253875e-01 1.12638831e-01 -6.30340755e-01 9.34728906e-02
-1.60935163e-01 -7.57971853e-02 -2.78026819e-01 5.44069529e-01
3.81404012e-01 -2.12825567e-01 8.44372585e-02 4.44902152e-01
2.81632274e-01 5.66805780e-01 -6.74346030e-01 1.35225463e+00
-4.95306164e-01 7.42769897e-01 -8.95638764e-02 -6.32495821e-01
8.83100569e-01 3.21647584e-01 1.48081928e-01 -6.56507075e-01
2.64558583e-01 4.09241170e-01 8.52739140e-02 -5.29348075e-01
3.50443661e-01 -6.25601828e-01 -5.03074765e-01 7.32543647e-01
-1.33325398e-01 -1.09477721e-01 2.22912952e-01 5.33649802e-01
1.32893157e+00 -3.16282243e-01 8.60237598e-01 -9.88140404e-02
5.23619235e-01 1.94822684e-01 3.30301911e-01 8.86119545e-01
-5.54345369e-01 5.14652014e-01 8.23448002e-01 -3.31687719e-01
-8.34709108e-01 -8.19471121e-01 -1.80951983e-01 1.70511818e+00
-2.14159146e-01 -4.31588739e-01 -5.71099460e-01 -1.17489505e+00
1.21347181e-01 1.32562268e+00 -7.62878120e-01 -3.00354362e-01
-5.63929319e-01 -9.80370283e-01 1.06285846e+00 2.16764137e-01
3.19163144e-01 -8.90696943e-01 -6.01884067e-01 2.38196164e-01
-3.41402203e-01 -1.10751462e+00 -4.75496203e-01 1.12322450e-01
-4.57366377e-01 -1.02576280e+00 -4.83720064e-01 -6.45619854e-02
-1.06490992e-01 2.87359953e-01 1.16420472e+00 3.49271953e-01
-1.69753864e-01 -1.15434572e-01 -3.40095550e-01 -4.10337806e-01
-8.36072206e-01 1.41641989e-01 8.02184194e-02 -4.19331603e-02
7.74837255e-01 -3.77534866e-01 -2.87913457e-02 -5.54290274e-03
-7.37031639e-01 -3.43982637e-01 1.42227754e-01 8.45995188e-01
-1.19297788e-01 -5.95535874e-01 7.00974226e-01 -1.28787756e+00
1.19855762e+00 -8.87631178e-01 3.20315659e-02 -1.22250197e-02
-5.47997594e-01 -1.61029205e-01 4.33671355e-01 -6.75284088e-01
-9.25724864e-01 -5.05845785e-01 -2.68938839e-01 -4.91947979e-01
-1.21224679e-01 1.36682913e-01 1.28100410e-01 4.21864361e-01
1.30721128e+00 -3.22971344e-01 -2.36906171e-01 -5.64770997e-01
1.55191511e-01 8.38086545e-01 -3.91990654e-02 -3.85558307e-01
6.10953391e-01 3.68667603e-01 -6.43367231e-01 -8.53176057e-01
-1.46389341e+00 -7.38821149e-01 -3.01936507e-01 -2.46887095e-02
7.43887186e-01 -9.18064654e-01 -1.12385966e-01 1.95123449e-01
-1.32593668e+00 -3.07240546e-01 8.66029188e-02 1.71623635e-03
-2.64243186e-01 1.60276413e-01 -7.63056636e-01 -9.49806213e-01
-5.23158073e-01 -7.94137180e-01 1.01555967e+00 -1.60228938e-01
-1.02831316e+00 -1.16767597e+00 2.60313600e-01 5.38305998e-01
3.06349397e-01 2.32139274e-01 1.07388926e+00 -1.43115580e+00
4.83953953e-01 -1.79411873e-01 -3.58100802e-01 1.06795564e-01
1.48250610e-02 6.82981405e-03 -1.33326101e+00 -8.30129609e-02
6.96848482e-02 -8.72799456e-01 1.17997825e+00 9.39666852e-02
6.72847569e-01 -3.90156031e-01 -4.59962666e-01 -1.63770523e-02
1.17125022e+00 -2.52911747e-01 3.03970933e-01 4.82703656e-01
4.89533722e-01 6.28365815e-01 5.19015133e-01 5.41045249e-01
3.78336012e-02 8.95488679e-01 3.11804861e-01 2.65499592e-01
-3.45802754e-01 -3.16021085e-01 7.72580147e-01 1.59060016e-01
6.04138970e-01 -2.88179457e-01 -1.16888165e+00 6.46043837e-01
-1.70492601e+00 -1.26627696e+00 -2.26653084e-01 1.70950389e+00
9.30019259e-01 6.61061764e-01 6.03707612e-01 2.17444509e-01
8.48911583e-01 3.75947595e-01 -4.71900642e-01 -9.21868622e-01
-1.07606061e-01 -1.33203968e-01 1.72982171e-01 6.00149751e-01
-1.17100823e+00 1.12304699e+00 6.29060125e+00 9.88412261e-01
-1.19296956e+00 9.36947584e-01 6.44772291e-01 -5.58208585e-01
-2.63907433e-01 -3.55402172e-01 -1.09564519e+00 5.48435628e-01
1.00773776e+00 -3.43615413e-02 2.79617552e-02 1.11938846e+00
4.53517400e-03 1.80319056e-01 -9.23377514e-01 8.02078128e-01
5.81391215e-01 -1.24794257e+00 3.01368609e-02 7.38197491e-02
6.80441380e-01 1.42654389e-01 6.50902763e-02 8.85325074e-01
6.33909285e-01 -9.81002450e-01 1.04648590e+00 -2.43695185e-01
6.37052119e-01 -3.30615193e-01 5.60861409e-01 6.23922646e-01
-2.91173607e-01 -3.79297286e-01 -2.82138914e-01 -1.97368652e-01
3.78212295e-02 5.55888534e-01 -1.28887248e+00 -1.18342400e-01
7.10073352e-01 7.46588945e-01 -7.87946761e-01 5.05585611e-01
-2.89328188e-01 8.83728623e-01 -6.24285564e-02 -2.74696231e-01
4.03287381e-01 4.39366400e-01 6.94477439e-01 1.57050252e+00
6.39554188e-02 -2.70075798e-01 1.15286574e-01 9.26430583e-01
-4.29302901e-01 3.39181483e-01 -8.18404794e-01 -2.18533963e-01
2.46252880e-01 1.20004463e+00 -5.21553516e-01 -5.03858507e-01
-4.58665818e-01 8.72955024e-01 7.33229220e-01 -9.97064784e-02
-1.12764418e+00 -1.65940542e-02 7.43537605e-01 3.84174705e-01
-4.07272018e-02 2.83799976e-01 -5.68329036e-01 -1.19748688e+00
-1.99497923e-01 -9.49027181e-01 9.00915802e-01 -4.31829929e-01
-1.73639357e+00 4.41783428e-01 -1.39899421e-02 -1.02020574e+00
-4.50096250e-01 -5.45045614e-01 -7.80338824e-01 5.59908628e-01
-1.38640237e+00 -1.15752137e+00 -3.34644802e-02 4.46862400e-01
9.52727199e-01 -2.21126452e-01 7.10345805e-01 2.70726144e-01
-7.93468952e-01 6.49287641e-01 -3.14217776e-01 3.52046043e-01
1.28419161e+00 -1.07789123e+00 2.49744281e-01 9.77757096e-01
1.77179351e-01 6.70275092e-01 1.07755637e+00 -8.60604823e-01
-5.81397533e-01 -9.19421852e-01 9.65189159e-01 -9.46786702e-01
1.20348823e+00 -4.74804997e-01 -1.03334177e+00 6.88923717e-01
1.46544293e-01 -1.49355054e-01 1.03751421e+00 5.32628119e-01
-1.24090159e+00 2.95253754e-01 -1.31368208e+00 5.68963349e-01
1.16172993e+00 -4.15185064e-01 -1.05378723e+00 6.20400667e-01
5.07046163e-01 -9.76849496e-02 -2.39763185e-01 -1.73822492e-01
4.57778543e-01 -1.10257185e+00 8.21605206e-01 -1.14504838e+00
9.10907924e-01 3.87340993e-01 -3.72444659e-01 -1.31268942e+00
-1.60795942e-01 -5.29092371e-01 -1.48836106e-01 1.46521616e+00
5.78594387e-01 -5.05114794e-01 3.83782476e-01 4.25245136e-01
4.64590527e-02 -3.16254735e-01 -1.12757075e+00 -8.26044798e-01
5.16321123e-01 -6.27082050e-01 2.66939223e-01 1.27779007e+00
2.75422126e-01 9.36861634e-01 -5.38057566e-01 -2.74965286e-01
5.83708644e-01 5.05172908e-02 6.95066869e-01 -1.42176735e+00
-2.12923989e-01 -8.11173022e-01 9.61874239e-03 -4.14217025e-01
5.23337662e-01 -9.39917266e-01 -4.09989506e-02 -1.22567081e+00
6.22444689e-01 -2.81819910e-01 -3.64758521e-02 7.04005778e-01
-3.27271819e-01 6.40053570e-01 2.67628282e-01 3.02220583e-01
-5.61333179e-01 1.35525957e-01 5.75684607e-01 -4.20603365e-01
-3.20051968e-01 -2.63491482e-01 -1.00709355e+00 9.56682742e-01
8.05763423e-01 -9.38793361e-01 -1.10382587e-01 -2.52005786e-01
2.97047555e-01 -2.53081471e-01 2.29458451e-01 -7.49596655e-01
-2.90038496e-01 -3.24965864e-01 1.63254887e-01 -1.76494539e-01
5.53883672e-01 -2.90161073e-01 -4.87667054e-01 6.76590025e-01
-6.08284771e-01 -3.38786066e-01 8.71335492e-02 5.95866084e-01
1.83980703e-01 -6.02106392e-01 1.16450036e+00 -1.64052874e-01
-4.72887963e-01 -7.45789185e-02 -5.39563656e-01 6.32021666e-01
8.94196630e-01 -1.18136264e-01 -7.12817729e-01 -3.53069037e-01
-4.55374241e-01 -2.49890015e-01 4.11272109e-01 7.27265120e-01
3.22708905e-01 -1.02863777e+00 -9.25141811e-01 -1.39533594e-01
6.66316092e-01 -8.37964952e-01 -5.71450070e-02 9.22221959e-01
-1.74462378e-01 4.09841269e-01 -1.69330344e-01 -4.16450858e-01
-1.30227566e+00 4.75974381e-01 2.82581210e-01 -5.88944256e-01
-4.82921988e-01 9.44639564e-01 1.22155003e-01 -2.96276242e-01
-1.03326105e-01 2.11280897e-01 -3.56223881e-01 5.43145180e-01
9.17891562e-01 2.67904013e-01 6.61545396e-02 -7.06402302e-01
-5.90426564e-01 8.05698857e-02 -5.16274929e-01 -2.34204844e-01
1.17143750e+00 1.30000740e-01 2.49150902e-01 6.97043836e-01
1.11555028e+00 5.92805564e-01 -7.65796363e-01 -3.73822480e-01
4.70521957e-01 -5.30602276e-01 -3.54645923e-02 -8.22028220e-01
-5.42413950e-01 1.05139768e+00 9.47378650e-02 3.87854308e-01
3.86465222e-01 2.11308956e-01 6.86394095e-01 5.77882724e-03
3.45641039e-02 -1.12565958e+00 7.78220892e-01 6.32341564e-01
9.76038575e-01 -1.45102704e+00 -7.31660193e-03 -3.35133523e-01
-1.07286322e+00 1.02146399e+00 7.81490386e-01 -4.41318415e-02
2.52209723e-01 2.46931184e-02 1.44863665e-01 -3.80047053e-01
-7.79604316e-01 -2.54077464e-01 8.92839134e-02 3.75238776e-01
6.36159241e-01 3.54929753e-02 -4.44730252e-01 8.37262154e-01
-1.24051042e-01 -4.44899827e-01 6.93300128e-01 5.10460913e-01
-6.48927629e-01 -7.45977283e-01 -5.25169194e-01 5.57340920e-01
-8.43013167e-01 -2.34028205e-01 -1.02537370e+00 7.80069232e-01
-3.37799378e-02 1.17512107e+00 -8.42245296e-02 -3.43715906e-01
1.89040542e-01 6.14220858e-01 6.70481771e-02 -1.14151037e+00
-9.40798402e-01 -5.98546751e-02 6.47342026e-01 -3.39314640e-01
-1.35485202e-01 -7.75971711e-01 -1.05590093e+00 -4.73896593e-01
-1.79361388e-01 -5.33922240e-02 3.47962826e-01 1.15363574e+00
4.00611907e-02 2.35629007e-01 1.82772726e-01 -2.70927727e-01
-8.89850616e-01 -1.38316274e+00 -3.08983922e-01 6.46648347e-01
3.08426082e-01 -8.83421123e-01 -6.21665239e-01 -3.83093119e-01]
|
[8.836777687072754, 10.443801879882812]
|
c3c1cdbb-9ef5-4593-bda2-630bb5d9cfc0
|
conformal-prediction-set-for-time-series
|
2206.07851
| null |
https://arxiv.org/abs/2206.07851v1
|
https://arxiv.org/pdf/2206.07851v1.pdf
|
Conformal prediction set for time-series
|
When building either prediction intervals for regression (with real-valued response) or prediction sets for classification (with categorical responses), uncertainty quantification is essential to studying complex machine learning methods. In this paper, we develop Ensemble Regularized Adaptive Prediction Set (ERAPS) to construct prediction sets for time-series (with categorical responses), based on the prior work of [Xu and Xie, 2021]. In particular, we allow unknown dependencies to exist within features and responses that arrive in sequence. Method-wise, ERAPS is a distribution-free and ensemble-based framework that is applicable for arbitrary classifiers. Theoretically, we bound the coverage gap without assuming data exchangeability and show asymptotic set convergence. Empirically, we demonstrate valid marginal and conditional coverage by ERAPS, which also tends to yield smaller prediction sets than competing methods.
|
['Yao Xie', 'Chen Xu']
|
2022-06-15
| null | null | null | null |
['prediction-intervals']
|
['miscellaneous']
|
[ 4.51288968e-01 -7.07379282e-02 -1.89142495e-01 -8.15907300e-01
-8.54535401e-01 -6.70131087e-01 4.11183804e-01 3.62446196e-02
4.47928486e-03 1.26747870e+00 -2.15835452e-01 -4.90129203e-01
-6.51990712e-01 -8.89565825e-01 -5.77944517e-01 -7.40323305e-01
-2.23172441e-01 2.07018405e-01 -1.13955610e-01 1.04279615e-01
1.90809757e-01 3.97445619e-01 -1.81238282e+00 1.72128528e-01
9.52746809e-01 1.25187290e+00 -3.29304844e-01 5.15497923e-01
4.74222243e-01 6.85137510e-01 -5.10129988e-01 -4.38326448e-01
1.33229509e-01 -3.91071260e-01 -4.33031470e-01 -3.57187867e-01
-2.62988210e-01 -2.50969399e-02 1.52154461e-01 6.20705605e-01
1.75684035e-01 1.35625303e-01 1.24667585e+00 -1.66427159e+00
-5.72102666e-01 9.69805419e-01 -4.99874860e-01 -1.76772475e-01
1.03921629e-01 -2.13186935e-01 9.31112707e-01 -7.44596124e-01
3.79105993e-02 1.00365961e+00 1.05706918e+00 5.25975287e-01
-1.34408343e+00 -7.32797682e-01 4.76744212e-02 5.13643287e-02
-1.35275853e+00 -1.93439409e-01 3.97336394e-01 -6.68810010e-01
6.89758837e-01 6.28979266e-01 2.05002308e-01 1.38870287e+00
5.80797553e-01 4.68823791e-01 1.29370487e+00 -5.55897892e-01
5.39868891e-01 1.94351703e-01 5.55349648e-01 -1.66144028e-01
2.90073097e-01 5.69934428e-01 -1.89649463e-01 -4.17737067e-01
6.39280200e-01 1.49815410e-01 -1.01185419e-01 2.04030145e-02
-8.89394701e-01 9.68580306e-01 -1.80668622e-01 -3.27983946e-02
-4.90500778e-01 9.58408788e-02 3.35050195e-01 5.72833776e-01
8.14574540e-01 3.06794345e-01 -9.06558454e-01 -5.45494668e-02
-7.90715396e-01 2.74243176e-01 8.02660465e-01 1.02920401e+00
3.85469764e-01 5.20042703e-02 -5.04280686e-01 8.11252773e-01
1.13206901e-01 4.67094839e-01 2.20613033e-01 -9.83287394e-01
1.86929375e-01 3.06812346e-01 4.39978868e-01 -7.40273535e-01
-5.40216446e-01 -4.23598051e-01 -1.13897765e+00 -1.23919867e-01
5.48873961e-01 -4.12670285e-01 -5.79489172e-01 1.94445252e+00
2.00439066e-01 1.84176341e-01 1.38916448e-01 3.93262863e-01
2.80683696e-01 4.66016978e-01 1.16203882e-01 -9.02465999e-01
9.32071805e-01 -2.76614130e-01 -5.51240683e-01 2.98895091e-01
5.80204308e-01 -1.76538542e-01 8.58645678e-01 6.84814990e-01
-5.46096623e-01 -4.82288748e-01 -8.80053520e-01 4.61240768e-01
-8.67997482e-02 -1.05579019e-01 4.38500404e-01 9.37228799e-01
-7.43357599e-01 7.71589279e-01 -7.36288548e-01 7.57827833e-02
3.24068904e-01 1.73971504e-01 -1.88753396e-01 3.02507095e-02
-1.38681030e+00 8.47740591e-01 2.89646029e-01 4.05928157e-02
-4.91738409e-01 -9.18872058e-01 -6.26090586e-01 -7.81842880e-03
2.04787731e-01 -4.61352974e-01 1.09760654e+00 -9.54958141e-01
-1.37539768e+00 1.44305229e-01 -1.62961885e-01 -7.40933359e-01
6.96097195e-01 -1.17612265e-01 -5.38417518e-01 -5.02798676e-01
-2.20109373e-01 2.05097541e-01 9.35444176e-01 -1.06011200e+00
-6.12425029e-01 -4.23182815e-01 -2.94272542e-01 -2.89238662e-01
-1.85357928e-01 1.83900613e-02 3.02581936e-01 -6.23359561e-01
-8.17046762e-02 -8.61015916e-01 -4.60638314e-01 -4.62584913e-01
-3.05470020e-01 -5.97712278e-01 2.32361212e-01 -6.56769097e-01
1.54326153e+00 -2.13680506e+00 -1.03957288e-01 6.34029984e-01
-9.58772749e-02 -5.51740944e-01 -3.43704363e-03 6.44818664e-01
-2.49076888e-01 2.42243648e-01 -5.65292895e-01 -1.66481331e-01
1.10976703e-01 3.12065065e-01 -6.82406485e-01 4.96087998e-01
2.39473611e-01 5.67349672e-01 -5.44555306e-01 -2.58437872e-01
8.57737660e-02 2.41683453e-01 -3.62121969e-01 -6.15634359e-02
-2.66035289e-01 5.96520603e-01 -2.98720300e-01 6.89385772e-01
6.47229433e-01 -3.08260471e-01 2.46871144e-01 1.31544963e-01
-1.32762343e-01 -1.04756564e-01 -1.14790022e+00 8.47739875e-01
-5.19593358e-01 3.22735488e-01 -4.62885916e-01 -1.36987782e+00
1.13642943e+00 2.68648982e-01 5.55564821e-01 -2.44385898e-02
1.37773842e-01 1.62362054e-01 -7.36655295e-02 -4.93363710e-04
1.28391251e-01 -1.34811431e-01 -4.57026094e-01 3.05380851e-01
-5.42713031e-02 -8.78314227e-02 -5.35212345e-02 -4.64248031e-01
1.06585836e+00 1.99040905e-01 7.36879706e-01 -4.19292897e-01
3.99210483e-01 -1.84435159e-01 7.57143974e-01 1.09292221e+00
9.10710320e-02 5.43946683e-01 6.24814868e-01 -1.10173471e-01
-1.09741318e+00 -1.01732206e+00 -9.54952836e-01 9.94668722e-01
-2.99689531e-01 -1.16247624e-01 -3.52115214e-01 -5.66354632e-01
2.28925601e-01 1.29549086e+00 -8.33277464e-01 5.76083064e-02
-3.97177935e-02 -8.34375799e-01 3.28162819e-01 8.83632541e-01
-3.88412885e-02 -6.73083901e-01 -5.20284116e-01 3.77955288e-01
3.08897533e-02 -6.31508768e-01 -1.15266442e-01 4.90301430e-01
-8.34895253e-01 -9.56444383e-01 -4.20467347e-01 -2.11359169e-02
3.22231352e-01 -4.92926955e-01 1.09510863e+00 -4.90712196e-01
1.35241076e-01 5.96277595e-01 -4.81539965e-01 -8.62685800e-01
-3.75592977e-01 -2.49122024e-01 2.25866482e-01 -1.33355358e-03
3.25008273e-01 -7.34946549e-01 -4.68074590e-01 6.64156377e-01
-6.68374002e-01 -2.41499349e-01 2.92922229e-01 1.17343915e+00
7.67343462e-01 1.58841327e-01 1.18327045e+00 -8.04768622e-01
5.81198633e-01 -8.71352077e-01 -8.16245139e-01 6.27502084e-01
-1.02013004e+00 1.57330297e-02 7.52762794e-01 -6.41879261e-01
-8.75885904e-01 -6.66336119e-02 9.38523039e-02 -4.60239023e-01
-9.68035087e-02 6.61420524e-01 6.01666830e-02 3.08349550e-01
7.69287288e-01 1.25045255e-01 -3.55009809e-02 -1.69832498e-01
2.17567995e-01 9.37559545e-01 2.93949723e-01 -8.66051376e-01
3.34738553e-01 -1.57157499e-02 3.33681613e-01 -4.01224136e-01
-6.98671103e-01 -6.90955222e-02 -5.45490444e-01 -2.88245171e-01
1.93493754e-01 -7.55842924e-01 -8.55427086e-01 1.10405467e-01
-7.99395800e-01 -2.91496158e-01 -4.14300799e-01 6.41576946e-01
-8.00019145e-01 -2.35412065e-02 -3.21823448e-01 -1.52724254e+00
-2.35246390e-01 -6.43388569e-01 8.49850774e-01 -1.66705064e-02
-5.16310573e-01 -1.04660010e+00 9.42388996e-02 -2.36115366e-01
2.71783441e-01 7.06558347e-01 7.76180506e-01 -1.08429682e+00
-2.78686136e-02 -2.96940774e-01 6.81602657e-02 6.24223292e-01
1.63517222e-01 3.00861955e-01 -1.00499868e+00 -7.53305927e-02
-8.44799541e-03 -1.28668115e-01 7.50040770e-01 8.03609133e-01
1.67068028e+00 -5.23356259e-01 -3.61223459e-01 1.91425189e-01
1.28180218e+00 4.95028973e-01 5.10523200e-01 -1.25409085e-02
-4.77385037e-02 1.09489465e+00 7.89661169e-01 9.87875581e-01
1.03641473e-01 2.51308471e-01 7.78674409e-02 4.99865592e-01
7.66733706e-01 -1.71755105e-01 3.29120934e-01 4.46632326e-01
-9.61928144e-02 -3.94615859e-01 -7.49381065e-01 4.14719999e-01
-1.99777520e+00 -1.04696012e+00 -3.28301013e-01 2.61756778e+00
9.76690948e-01 -1.23311959e-01 2.76005775e-01 2.88260818e-01
7.05166757e-01 -5.68373561e-01 -9.11234796e-01 -2.90107042e-01
-2.29835287e-01 3.71818691e-01 5.29768229e-01 2.67693132e-01
-9.60254252e-01 9.39840302e-02 7.16316319e+00 9.36774254e-01
-6.97986782e-01 7.20445588e-02 9.25316632e-01 -8.13835207e-03
-4.61788714e-01 2.55525615e-02 -8.39254618e-01 6.35595381e-01
1.38182163e+00 -3.91258240e-01 1.78599745e-01 8.64228308e-01
-5.02685606e-02 5.12778908e-02 -1.52910972e+00 7.72935510e-01
-4.85425621e-01 -8.70548248e-01 -4.95043904e-01 3.29608694e-02
8.79135251e-01 -3.65624338e-01 2.58072674e-01 5.57006180e-01
7.57452309e-01 -1.25758231e+00 5.12702644e-01 9.54801083e-01
9.87820566e-01 -9.38069403e-01 9.49116230e-01 4.98726487e-01
-8.06833506e-01 -3.91478449e-01 -2.46539131e-01 -1.51841551e-01
-1.12670489e-01 1.06780529e+00 -5.33102095e-01 9.48032856e-01
5.76936305e-01 8.11374843e-01 -1.08937033e-01 7.35660851e-01
1.54054955e-01 1.11337411e+00 -7.00511754e-01 -2.75551468e-01
-3.78299177e-01 -2.62886137e-01 1.96709931e-01 8.93985748e-01
8.56661022e-01 3.18304449e-01 -9.98697877e-02 7.37568974e-01
3.35626811e-01 -1.75758731e-02 -5.28788567e-01 2.30900794e-01
9.31532979e-01 9.47021663e-01 -3.36092979e-01 -3.90554629e-02
-3.82739812e-01 3.54746640e-01 -1.14389889e-01 3.94954115e-01
-1.04850757e+00 -1.61509901e-01 4.20084506e-01 -1.50763303e-01
2.33334497e-01 2.02892080e-01 -7.79671848e-01 -8.23431611e-01
-1.03147395e-01 -6.89844310e-01 6.28626049e-01 -5.13096392e-01
-2.12202072e+00 3.81245285e-01 4.39603060e-01 -1.58969688e+00
-7.55801320e-01 -6.34195626e-01 -3.84279907e-01 9.56542611e-01
-8.14138830e-01 -9.83018160e-01 5.47078885e-02 5.63746750e-01
2.77296424e-01 1.64803654e-01 8.47937584e-01 -2.00559452e-01
-4.94933069e-01 8.50498974e-01 5.58286548e-01 -3.02027762e-01
6.62699699e-01 -1.16492546e+00 -1.84140354e-01 4.45143789e-01
-2.72268385e-01 6.19298220e-01 9.08745706e-01 -5.89453757e-01
-9.72806394e-01 -1.22268236e+00 3.76546353e-01 -8.96739781e-01
6.83669150e-01 -1.08856931e-01 -9.60578799e-01 8.94837320e-01
-2.02765092e-01 -1.05151117e-01 1.04068804e+00 4.83615547e-01
-2.94993699e-01 -3.08738470e-01 -1.28256750e+00 3.38379741e-01
8.56651902e-01 -1.46364063e-01 -4.18547273e-01 2.15317413e-01
6.75980091e-01 -1.73135743e-01 -1.53500032e+00 9.85008538e-01
9.45965886e-01 -8.77514064e-01 6.70693755e-01 -6.57401383e-01
3.98278326e-01 -4.62268628e-02 -5.85718155e-01 -1.32901967e+00
-2.85497189e-01 -4.02087450e-01 -1.59831837e-01 1.39991319e+00
6.50537610e-01 -1.17552865e+00 9.08369794e-02 9.48292971e-01
-5.39685450e-02 -7.67845631e-01 -1.07898974e+00 -1.14838409e+00
5.64271688e-01 -9.11797464e-01 8.49412501e-01 8.30144823e-01
1.90862924e-01 2.97107920e-02 -5.48064113e-01 1.20182246e-01
6.62302375e-01 1.24086805e-01 5.03350556e-01 -1.65194118e+00
-3.56792390e-01 -3.73667777e-01 -1.33731723e-01 -5.41554987e-01
3.03577542e-01 -4.16993827e-01 1.48078740e-01 -8.65368664e-01
6.19049743e-02 -8.13024998e-01 -7.11705983e-01 5.38295090e-01
-1.41107798e-01 -2.08609626e-01 -2.70212650e-01 2.30129674e-01
-1.68643191e-01 5.08531928e-01 7.38022208e-01 2.45811209e-01
-2.07865849e-01 6.02891207e-01 -6.44413590e-01 3.98414254e-01
9.06128466e-01 -4.36849356e-01 -6.75694704e-01 5.86029112e-01
3.03143799e-01 4.65343356e-01 3.41238856e-01 -6.28575325e-01
-1.79894060e-01 -5.29609442e-01 6.07725978e-01 -8.61775279e-01
1.34216128e-02 -7.88256049e-01 7.03273177e-01 1.96185261e-01
-7.01462865e-01 -1.50082231e-01 2.17520073e-01 8.85642231e-01
-5.34028821e-02 -2.32939214e-01 4.18077469e-01 3.56496900e-01
-1.91047192e-01 1.93059430e-01 -1.79766983e-01 -2.19520494e-01
1.16576421e+00 -2.04696204e-03 -1.83559567e-01 -4.52310890e-01
-7.61290967e-01 4.38002348e-01 5.27826883e-02 1.45072162e-01
4.85017389e-01 -1.24070418e+00 -1.07571673e+00 3.48195480e-03
1.50988594e-01 -1.31482959e-01 4.90668923e-01 9.72706497e-01
3.01921517e-01 4.37339365e-01 4.11572941e-02 -7.74269760e-01
-9.52146947e-01 5.95222354e-01 1.19300887e-01 -2.98258811e-01
-1.02166638e-01 6.69959486e-01 1.47349373e-01 -5.80974638e-01
-2.03811731e-02 -3.53561252e-01 -3.74702990e-01 -1.41748013e-02
5.72224796e-01 6.25021279e-01 -8.72119740e-02 2.33302042e-01
-3.30438703e-01 1.78637460e-01 1.50103584e-01 -2.12245688e-01
1.27173221e+00 -8.02197531e-02 -6.55700341e-02 1.10006475e+00
7.94743359e-01 -2.71419376e-01 -1.37774527e+00 -4.94457223e-02
1.60111725e-01 -1.80140391e-01 -3.44658524e-01 -9.25540388e-01
-3.80695462e-01 6.35638535e-01 6.29565179e-01 5.34616888e-01
1.39905429e+00 -2.32373774e-01 -7.29016885e-02 5.96150793e-02
5.47543824e-01 -8.59476626e-01 -4.86288667e-01 3.18294972e-01
1.16870773e+00 -1.26344121e+00 -1.02826215e-01 -2.25859985e-01
-6.90358520e-01 1.14035690e+00 4.21925664e-01 6.12436123e-02
8.91094983e-01 3.52948248e-01 -6.36394501e-01 6.29097939e-01
-1.32707274e+00 1.80741683e-01 4.95338023e-01 8.57926965e-01
4.54329401e-01 3.55807334e-01 -6.74113333e-01 1.42237318e+00
-2.49992833e-01 2.18190446e-01 3.42781186e-01 5.92758477e-01
-2.18027487e-01 -9.71586943e-01 -4.22796786e-01 9.37037468e-01
-3.79517227e-01 -1.20876938e-01 8.82505067e-03 8.16615641e-01
-1.21753782e-01 1.24428368e+00 2.97348768e-01 -5.28965414e-01
1.76609114e-01 3.94176960e-01 4.50883329e-01 -2.01709405e-01
-2.52989441e-01 2.59099100e-02 1.95487514e-01 -2.42215648e-01
-3.06040734e-01 -9.94217932e-01 -8.53432655e-01 -3.52026492e-01
-5.35346329e-01 1.76114008e-01 3.93840492e-01 9.55489755e-01
4.43997651e-01 4.24730301e-01 1.12684715e+00 -2.12692216e-01
-1.22023737e+00 -1.30545485e+00 -7.23645627e-01 -1.42716780e-01
1.94672108e-01 -7.92438805e-01 -7.70688891e-01 -3.06472294e-02]
|
[7.408056735992432, 3.974581003189087]
|
f36d515b-d419-478c-a788-557ab615f9de
|
dynamic-local-geometry-capture-in-3d
| null | null |
https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9565556
|
https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9565556
|
Dynamic Local Geometry Capture in 3D PointCloud Classification
|
With the advent of PointNet, the popularity of deep neural networks has increased in point cloud analysis. PointNet's successor, PointNet++, partitions the input point cloud and recursively applies PointNet to capture local geometry. PointNet++ model uses ball querying for local geometry capture in its set abstraction layers. Several models based on single-scale grouping of PointNet++ continue to use ball querying with a fixed-radius ball. Due to its uniform scale in all directions, a ball lacks orientation and is ineffective in capturing complex local neighborhoods. Few recent models replace a fixed-sized ball with a fixed-sized ellipsoid or a fixed-sized cuboid to capture local neighborhoods. However, these methods are not still fully effective in capturing varying geometry proportions from different local neighborhoods on the object surface. We propose a novel technique of dynamically oriented and scaled ellipsoid based on unique local information to capture the local geometry better. We also propose ReducedPointNet++, a single set abstraction-based single scale grouping model. Our model, along with dynamically oriented and scaled ellipsoid querying, achieves 92.1% classification accuracy on the ModelNet40 dataset. We achieve state-of-the-art 3D classification results on all six variants of the real-world ScanObjectNN dataset with an accuracy of 82.0% on the most challenging variant.
|
['Chandra Kambhamettu', 'Shivanand Venkanna Sheshappanavar']
|
2021-10-19
| null | null | null |
ieee-4th-international-conference-on
|
['3d-classification']
|
['computer-vision']
|
[-6.85196519e-01 -2.53022552e-01 5.04972041e-02 -2.08925322e-01
-4.45187449e-01 -5.33838034e-01 4.28802937e-01 9.05242190e-02
-1.39813647e-01 1.13730289e-01 -3.69535059e-01 -4.58037406e-01
-2.02062845e-01 -1.39340305e+00 -8.91650259e-01 -2.13961303e-01
-2.82275856e-01 9.77656007e-01 6.52047217e-01 -4.14148241e-01
1.70952618e-01 1.26831615e+00 -1.64764500e+00 1.26612961e-01
6.98408723e-01 1.34077263e+00 3.29745337e-02 3.43043596e-01
-3.08624059e-01 -1.41437754e-01 -5.92057049e-01 1.63465604e-01
7.50738859e-01 4.65533882e-01 -4.94408280e-01 -3.60126853e-01
9.61582720e-01 -5.52631199e-01 -2.24089682e-01 6.73308909e-01
5.94586968e-01 1.06276311e-01 4.20920849e-01 -1.32450223e+00
-6.03492200e-01 2.76592106e-01 -6.98502421e-01 1.39642388e-01
1.03471607e-01 -8.94594640e-02 9.19509053e-01 -1.36890161e+00
4.84621465e-01 1.29881179e+00 1.06795371e+00 3.11996728e-01
-1.16786337e+00 -1.08344436e+00 -2.42167152e-02 -7.55139366e-02
-1.89594483e+00 1.27864376e-01 9.80022609e-01 -4.03469875e-02
1.38052022e+00 3.06458175e-01 1.01040912e+00 3.92295063e-01
2.46393651e-01 3.37145150e-01 4.57716107e-01 -6.64091390e-03
3.49647999e-01 -4.86598998e-01 2.91156080e-02 5.66690981e-01
7.72933289e-02 1.13728002e-01 -1.40791178e-01 -5.24420857e-01
1.47654080e+00 4.63421524e-01 -1.09884925e-02 -9.33254957e-01
-1.29918563e+00 9.36076343e-01 1.14621317e+00 -6.97697252e-02
-3.04013908e-01 5.04983783e-01 1.40093282e-01 1.78364262e-01
7.23394096e-01 4.02198493e-01 -4.57926959e-01 1.48218106e-02
-8.83299649e-01 6.61923349e-01 6.45243168e-01 1.29179847e+00
1.02109051e+00 2.95026060e-02 2.99772233e-01 8.61830950e-01
3.84182841e-01 7.81795621e-01 -2.89629269e-02 -9.96897221e-01
4.94958818e-01 1.15394771e+00 -1.08101197e-01 -1.25595975e+00
-7.62244761e-01 -3.67565244e-01 -8.98686886e-01 6.94289863e-01
3.24955210e-02 3.73472244e-01 -1.09924662e+00 1.13372803e+00
7.18146086e-01 3.19591671e-01 -4.28027630e-01 8.65338504e-01
1.12545609e+00 5.45577049e-01 -3.36145520e-01 7.22635865e-01
1.18393493e+00 -6.61924779e-01 2.39986345e-01 3.28867048e-01
5.21828592e-01 -4.54992443e-01 1.04472339e+00 4.31723028e-01
-1.22079420e+00 -6.75829589e-01 -1.23971915e+00 -4.50600505e-01
-4.08029288e-01 -2.03468829e-01 5.87674201e-01 4.27585989e-01
-1.44765651e+00 5.95589697e-01 -1.00689185e+00 -1.70395002e-01
6.18088961e-01 7.02992499e-01 -2.46870786e-01 4.61774021e-02
-7.67331898e-01 4.92539495e-01 1.21433865e-02 -1.85038447e-01
-5.24874210e-01 -1.27592003e+00 -8.97872746e-01 2.85233140e-01
1.40055135e-01 -8.63828480e-01 1.13674510e+00 2.09538694e-02
-1.18433833e+00 6.32137537e-01 -1.69604775e-02 -5.21319687e-01
4.30282444e-01 -6.64410964e-02 -4.18762751e-02 2.89261907e-01
2.66405195e-01 1.00012457e+00 6.25607967e-01 -1.18013585e+00
-5.04140615e-01 -7.32259154e-01 2.93111622e-01 2.13649347e-01
4.20822157e-03 -2.03703746e-01 -4.97725278e-01 -4.89341706e-01
7.42823243e-01 -9.29333568e-01 -2.30189964e-01 5.71014285e-01
-1.10592917e-01 -7.68676639e-01 1.25266564e+00 2.14000911e-01
9.89542544e-01 -2.14746094e+00 -3.23684305e-01 4.74275023e-01
8.39822233e-01 1.00981466e-01 2.62795622e-03 1.66168243e-01
-1.45686105e-01 4.62525517e-01 -5.62017485e-02 -6.66000962e-01
-7.74706453e-02 9.58095044e-02 -3.33200127e-01 5.41624308e-01
8.98068249e-02 7.85342693e-01 -7.74574399e-01 -3.87430966e-01
7.19628334e-01 6.61611617e-01 -9.79818106e-01 -3.91140252e-01
-1.31378785e-01 -2.26698518e-01 -4.27180737e-01 9.69732583e-01
1.35154223e+00 -3.13398540e-01 -6.48877680e-01 -2.71534652e-01
-1.64458051e-01 1.44883111e-01 -1.33161795e+00 1.90048802e+00
-4.62265313e-01 3.39103222e-01 1.09696209e-01 -4.49483752e-01
1.37865663e+00 1.56941731e-02 8.88291299e-01 -4.49945390e-01
-5.83315752e-02 2.91150957e-01 -1.38593748e-01 3.13335955e-01
6.04135215e-01 -1.83334260e-03 8.92696250e-03 2.55892992e-01
-3.98234099e-01 -7.44637787e-01 -2.66033500e-01 1.88797772e-01
1.08352232e+00 9.70225930e-02 1.95054516e-01 -5.12951016e-01
2.64702469e-01 1.72542185e-01 1.85322687e-01 6.86398745e-01
-1.38045505e-01 1.34510148e+00 -9.72414669e-03 -9.91955459e-01
-1.19279110e+00 -1.35405791e+00 -6.38581753e-01 8.21418822e-01
5.58814466e-01 -6.04377329e-01 -2.53617167e-01 -3.31684113e-01
5.27079225e-01 3.71069044e-01 -5.40779889e-01 -1.41812330e-02
-9.38341200e-01 -3.94823879e-01 5.45410395e-01 9.49158132e-01
6.06785238e-01 -9.09408987e-01 -7.34665096e-01 2.36413822e-01
3.20628494e-01 -8.53181183e-01 -3.15640330e-01 -8.85555707e-03
-1.18847811e+00 -8.94518912e-01 -4.64819580e-01 -5.98625302e-01
5.08773029e-01 5.66892803e-01 1.44423759e+00 3.78448159e-01
-3.01834382e-02 1.36838883e-01 -2.44819537e-01 -7.38805711e-01
6.16206899e-02 4.23250586e-01 9.96532068e-02 -4.68537152e-01
5.19597709e-01 -8.19195390e-01 -9.44989145e-01 6.82140946e-01
-7.70396054e-01 -1.55502319e-01 2.20412135e-01 3.14878255e-01
9.33672309e-01 -2.24955127e-01 1.86276048e-01 -2.43990511e-01
3.21784824e-01 -6.71003103e-01 -4.33137715e-01 -3.38308156e-01
-3.17163378e-01 -3.43325913e-01 6.71105564e-01 -3.79241616e-01
-2.46620402e-01 -1.69386789e-01 -2.40612134e-01 -1.17348266e+00
-1.79341629e-01 1.01667523e-01 1.43966407e-01 -3.98287266e-01
8.23706806e-01 -8.11093953e-03 -5.57916388e-02 -6.35059834e-01
2.17098475e-01 5.57585895e-01 3.60978335e-01 -7.05762386e-01
6.78180575e-01 1.15951657e+00 1.89381406e-01 -8.03843856e-01
-2.40775764e-01 -6.80722833e-01 -7.59737313e-01 8.81000310e-02
6.54044330e-01 -8.26257408e-01 -8.58596742e-01 5.20898402e-01
-1.37198603e+00 -1.80973962e-01 -6.21749043e-01 1.02274783e-01
-4.87687558e-01 1.09161183e-01 -3.29721481e-01 -4.03798640e-01
-6.44227922e-01 -1.17407000e+00 1.77090132e+00 -3.85379046e-02
-2.74417643e-02 -7.13860750e-01 4.83999401e-02 -1.79179132e-01
4.59309518e-01 4.39644068e-01 6.26268029e-01 -5.26484609e-01
-8.82863700e-01 -2.42838547e-01 -4.92110610e-01 6.02938831e-02
1.34599671e-01 2.90856212e-02 -4.51884001e-01 -3.47380489e-01
-3.57082188e-02 1.07210211e-01 5.92510343e-01 3.89680177e-01
1.64219630e+00 2.54709244e-01 -6.09479785e-01 1.21251619e+00
1.50653839e+00 8.27326849e-02 4.81524110e-01 3.78387392e-01
8.98705721e-01 -1.50217324e-01 2.20253810e-01 3.41972947e-01
5.98594427e-01 6.55094981e-01 9.47634697e-01 -2.68470138e-01
-1.77069455e-01 -7.03871250e-02 -2.13769957e-01 3.83511633e-01
-3.50004107e-01 -1.00467309e-01 -1.35965621e+00 5.32882810e-01
-1.47923315e+00 -6.07292771e-01 -2.67109424e-01 2.06241107e+00
3.43803436e-01 2.67591178e-01 1.65231347e-01 5.99457771e-02
5.52187741e-01 9.50833112e-02 -8.29519510e-01 -2.19088897e-01
2.65681148e-02 4.88334358e-01 6.78173184e-01 3.07119757e-01
-9.82293785e-01 8.90981138e-01 6.09678793e+00 8.02193582e-01
-1.39858222e+00 -1.22688599e-01 2.78017521e-01 -4.29876685e-01
-1.80989623e-01 -2.05740988e-01 -1.00354493e+00 4.62674290e-01
5.76451421e-01 5.66535108e-02 1.71628729e-01 1.40388989e+00
5.22229411e-02 1.42476559e-01 -1.08294177e+00 1.16173899e+00
8.75316709e-02 -1.67077208e+00 4.01865274e-01 4.07539666e-01
5.51757276e-01 8.22259843e-01 -8.02235380e-02 3.18475962e-01
2.80029178e-01 -1.11303675e+00 9.16261554e-01 2.61804521e-01
1.19315720e+00 -9.49648440e-01 4.35483068e-01 3.96444410e-01
-1.63627887e+00 1.26885995e-01 -7.66942024e-01 -1.36476830e-01
4.02900241e-02 2.97340304e-01 -1.08117104e+00 4.43778962e-01
1.29399586e+00 7.19867706e-01 -5.44720113e-01 1.10667837e+00
5.32682776e-01 2.23559961e-01 -1.29095674e+00 -3.16547044e-02
4.36728448e-01 -2.27322832e-01 8.74316156e-01 7.60657191e-01
3.31459075e-01 1.08637676e-01 3.71227741e-01 1.15297008e+00
1.60781071e-02 -1.20747045e-01 -7.00038493e-01 8.21505606e-01
8.93870950e-01 9.68982935e-01 -8.76340985e-01 -3.44574183e-01
-2.86803693e-01 2.93611407e-01 4.19619143e-01 -4.97140642e-03
-7.80246317e-01 -5.28628170e-01 9.80034828e-01 6.76504612e-01
5.26389599e-01 -7.13363349e-01 -7.16480851e-01 -7.68799245e-01
1.08533278e-01 -4.19730783e-01 1.72796935e-01 -9.96698081e-01
-1.15543997e+00 7.83582389e-01 4.63644505e-01 -1.48748338e+00
4.11350746e-03 -6.43859208e-01 -6.93009615e-01 7.76600242e-01
-1.19127202e+00 -1.37659335e+00 -7.09786355e-01 8.87658834e-01
5.50850093e-01 1.37544200e-01 5.15956283e-01 2.60228753e-01
8.03270787e-02 4.43335116e-01 6.15973584e-02 2.08188117e-01
2.66742826e-01 -1.42371941e+00 9.47258055e-01 1.69538006e-01
-2.52118930e-02 8.88704360e-01 2.20047578e-01 -6.72143877e-01
-1.21527374e+00 -1.08167410e+00 2.94389606e-01 -8.49490941e-01
3.58307213e-01 -6.36543632e-01 -1.07266998e+00 6.65324509e-01
-5.46010852e-01 4.58816409e-01 1.20501347e-01 -1.66152865e-02
-2.96481460e-01 -3.62103313e-01 -1.50548911e+00 4.80225772e-01
1.49337339e+00 -2.84446120e-01 -6.05749726e-01 1.92513213e-01
1.09001434e+00 -9.12115395e-01 -1.06231368e+00 8.19516897e-01
4.60167766e-01 -1.08843982e+00 1.52218246e+00 -2.98059821e-01
1.33253947e-01 -4.09109652e-01 -2.22826153e-01 -1.14427626e+00
-4.44402903e-01 -2.75273442e-01 -2.32975185e-01 5.13600290e-01
-3.62907089e-02 -8.89609575e-01 1.25649083e+00 3.68135214e-01
-5.72490156e-01 -1.23926628e+00 -1.31716168e+00 -7.58709252e-01
4.12765443e-01 -7.92645931e-01 1.40907955e+00 6.51712239e-01
-5.17992258e-01 -5.58590405e-02 3.90854895e-01 2.48351350e-01
4.91770118e-01 3.21443588e-01 1.20532179e+00 -1.66249847e+00
5.01135051e-01 -6.13029778e-01 -7.66333997e-01 -1.39549983e+00
-3.27983856e-01 -8.54069352e-01 -4.28640038e-01 -1.51359403e+00
-4.39059585e-01 -1.10585308e+00 -8.66156369e-02 3.68282646e-01
4.39896584e-01 6.91476524e-01 1.75162807e-01 4.37757194e-01
-3.38995308e-01 6.13257289e-01 1.28410578e+00 7.40289167e-02
-4.42084253e-01 -4.19427566e-02 -3.50009441e-01 8.52846682e-01
7.18641698e-01 -3.48781973e-01 -3.33714753e-01 -7.89128780e-01
3.85028452e-01 -2.62777776e-01 6.10455990e-01 -1.42527938e+00
3.78722191e-01 -3.90117429e-02 5.28941453e-01 -1.48970687e+00
7.55021334e-01 -9.92585421e-01 3.77202146e-02 1.14692621e-01
3.72936577e-01 6.22842669e-01 3.40426594e-01 4.28046763e-01
1.02235742e-01 3.20998043e-01 8.01335692e-01 -3.17614883e-01
-4.74418998e-01 9.24525559e-01 3.20208013e-01 -2.87198156e-01
1.04804647e+00 -8.53718400e-01 -2.03162566e-01 -2.58875657e-02
-6.91371322e-01 4.02919739e-01 8.90210390e-01 5.81422806e-01
9.32202578e-01 -1.55606461e+00 -7.94744849e-01 5.71223140e-01
-1.01004960e-02 1.22915995e+00 7.44026825e-02 3.11330348e-01
-1.19380105e+00 3.39648277e-01 -7.99652189e-02 -1.39588380e+00
-9.78557885e-01 2.61786282e-01 7.27296770e-01 1.48469716e-01
-1.00215638e+00 8.55300367e-01 2.82061875e-01 -7.86732852e-01
-4.70879488e-02 -1.04856622e+00 1.93319350e-01 -4.69072312e-01
2.86408305e-01 6.63646877e-01 5.04394650e-01 -6.01589024e-01
-5.09016693e-01 9.53502536e-01 -1.69906486e-02 2.28141844e-01
1.32873940e+00 9.51549336e-02 -1.77405924e-01 3.34969997e-01
1.17576575e+00 -4.75949608e-02 -1.04367232e+00 -1.37284636e-01
-5.80785751e-01 -5.30913472e-01 1.77994415e-01 -3.03675979e-01
-1.05370140e+00 7.96608031e-01 4.94472325e-01 3.73301715e-01
6.66629016e-01 2.88990378e-01 8.96983325e-01 4.05872196e-01
8.26315701e-01 -5.90298295e-01 -1.15456238e-01 7.53179133e-01
1.24965942e+00 -1.10703719e+00 1.55200869e-01 -4.51399803e-01
1.96842492e-01 1.13728571e+00 1.03734636e+00 -7.76961148e-01
1.14736652e+00 3.38443667e-01 -1.68799803e-01 -6.79295421e-01
-3.93147171e-01 2.76038975e-01 4.25295919e-01 5.57958901e-01
-4.60500605e-02 6.62219226e-02 3.36759657e-01 2.47777179e-01
-8.93639147e-01 -1.98731914e-01 1.87143937e-01 7.77454853e-01
-4.73714143e-01 -4.71685499e-01 -8.07095349e-01 5.84763288e-01
4.82483469e-02 -1.76202152e-02 -1.36894867e-01 1.14752507e+00
3.38726133e-01 3.93114746e-01 1.07512057e+00 -3.43530983e-01
5.52169681e-01 -3.75552744e-01 1.73047468e-01 -8.49420249e-01
-4.45180237e-01 -1.58383757e-01 -6.04201674e-01 -9.67583239e-01
-1.16932299e-02 -5.13701856e-01 -1.61004007e+00 -6.46936476e-01
-3.04983348e-01 -1.30791832e-02 9.10057485e-01 4.15859222e-01
8.60591888e-01 1.52840212e-01 4.88618612e-01 -1.52800870e+00
-6.35807335e-01 -9.60533977e-01 -4.68198866e-01 1.77785262e-01
5.16856790e-01 -8.07248116e-01 -3.74123037e-01 -6.93650126e-01]
|
[7.900228500366211, -3.509323835372925]
|
90fe825a-3e2f-4da9-bbfa-43174c6a2652
|
combining-residual-networks-with-lstms-for
|
1703.04105
| null |
http://arxiv.org/abs/1703.04105v4
|
http://arxiv.org/pdf/1703.04105v4.pdf
|
Combining Residual Networks with LSTMs for Lipreading
|
We propose an end-to-end deep learning architecture for word-level visual
speech recognition. The system is a combination of spatiotemporal
convolutional, residual and bidirectional Long Short-Term Memory networks. We
train and evaluate it on the Lipreading In-The-Wild benchmark, a challenging
database of 500-size target-words consisting of 1.28sec video excerpts from BBC
TV broadcasts. The proposed network attains word accuracy equal to 83.0,
yielding 6.8 absolute improvement over the current state-of-the-art, without
using information about word boundaries during training or testing.
|
['Georgios Tzimiropoulos', 'Themos Stafylakis']
|
2017-03-12
| null | null | null | null |
['lipreading']
|
['computer-vision']
|
[ 1.70573846e-01 -7.94794932e-02 -6.67023957e-01 -2.25879833e-01
-1.26378167e+00 -3.17454904e-01 4.91535097e-01 -4.85559136e-01
-6.44722223e-01 3.66936535e-01 5.20154059e-01 -8.08382452e-01
6.74326122e-01 9.97510403e-02 -9.80788291e-01 -6.78505301e-01
-6.36471957e-02 -1.98148116e-01 1.31878391e-01 1.95322067e-01
9.74475667e-02 2.12499872e-01 -1.48653162e+00 7.05627680e-01
2.57992148e-01 1.56782007e+00 1.37531251e-01 1.18932045e+00
1.18450120e-01 8.37520778e-01 -6.43762469e-01 -4.50254619e-01
-1.12873860e-01 -1.12863809e-01 -6.47018492e-01 -1.41948849e-01
7.48432040e-01 -4.49294686e-01 -8.49339962e-01 8.83195221e-01
1.26626146e+00 -2.06997126e-01 4.29845065e-01 -9.90798235e-01
-1.12013268e+00 2.07915381e-01 -2.60942698e-01 4.80206490e-01
3.27291906e-01 4.30792212e-01 7.80889690e-01 -1.38072026e+00
4.14065331e-01 1.18514323e+00 5.20824671e-01 9.95931923e-01
-7.97654033e-01 -7.04794705e-01 -3.20499938e-04 5.85356057e-01
-1.45647359e+00 -1.49744391e+00 5.13526380e-01 -1.56487525e-01
1.57317996e+00 -1.12331696e-01 3.56548667e-01 1.93125403e+00
4.66084629e-02 1.07520390e+00 7.05348492e-01 -4.45025533e-01
-4.29963060e-02 -1.13640152e-01 3.14435624e-02 6.40830517e-01
-3.43732953e-01 4.92216349e-01 -1.30386281e+00 1.75263628e-01
2.33485058e-01 -6.84210122e-01 -5.52861273e-01 -2.51965784e-02
-1.06131113e+00 3.93503755e-01 2.29942411e-01 1.28684789e-01
-2.53980726e-01 4.81851399e-01 6.44561648e-01 2.15974957e-01
6.71728671e-01 -5.77330112e-01 -4.31356013e-01 -6.04062200e-01
-1.05800557e+00 -3.24983895e-01 4.27878797e-01 8.61379862e-01
-1.17518090e-01 4.67056811e-01 -3.75089705e-01 1.15237880e+00
6.62230074e-01 7.82595038e-01 8.14400733e-01 -5.40492117e-01
7.50447452e-01 -3.77729207e-01 -1.13612667e-01 -4.94637221e-01
-9.59948674e-02 -3.65500689e-01 -5.16450524e-01 5.74890785e-02
2.01237962e-01 -2.54134297e-01 -1.46767592e+00 1.79965878e+00
-1.96595877e-01 5.72432876e-01 4.17090207e-01 8.84974062e-01
1.47381818e+00 9.45364296e-01 1.47365719e-01 -2.38074049e-01
1.26215160e+00 -1.29439557e+00 -1.00019753e+00 -3.30881923e-01
1.69341296e-01 -7.39881158e-01 9.97367263e-01 2.39613682e-01
-1.33113885e+00 -6.79190993e-01 -9.85513628e-01 -3.25483501e-01
-5.20611823e-01 2.96537459e-01 -6.19270355e-02 7.77914822e-01
-1.62506402e+00 1.71322033e-01 -5.37819147e-01 -2.65894860e-01
4.64174628e-01 1.36596158e-01 -3.56625021e-01 1.91206470e-01
-1.30413091e+00 6.12296760e-01 -3.13181370e-01 1.08149923e-01
-1.31441379e+00 -8.03802192e-01 -1.07029688e+00 3.16275703e-03
1.17843136e-01 -2.58377939e-01 1.49827206e+00 -9.20449674e-01
-1.93391836e+00 1.17460763e+00 -6.96255982e-01 -6.86467350e-01
3.81726861e-01 -6.18904382e-02 -7.53180742e-01 3.00373644e-01
-2.27929488e-01 1.12773764e+00 1.26942277e+00 -8.46492767e-01
-3.31880271e-01 -1.88604087e-01 -3.65891367e-01 1.16593748e-01
-2.81462848e-01 4.79836434e-01 -8.48339558e-01 -8.67931604e-01
-3.52164924e-01 -8.27727675e-01 5.83887875e-01 1.38906389e-01
-3.72492820e-01 -4.34558302e-01 1.06903911e+00 -1.08680999e+00
7.88807213e-01 -2.29982448e+00 -1.18145049e-01 -5.39436936e-01
-1.04169972e-01 6.53228760e-01 -4.92989838e-01 -3.81540433e-02
-1.77944437e-01 1.88051507e-01 2.17211306e-01 -8.88167679e-01
1.17540926e-01 -3.96473229e-01 -3.83758694e-01 6.39602602e-01
-1.60210207e-02 1.27166235e+00 -3.77141923e-01 -3.92419070e-01
2.77263135e-01 8.89911175e-01 3.11012603e-02 2.51880497e-01
-8.91125947e-02 -1.87945552e-02 3.83262902e-01 9.82047975e-01
6.80951834e-01 1.73829906e-02 -2.39614978e-01 -8.31425339e-02
1.56482726e-01 5.98620772e-01 -2.75424719e-01 2.18619537e+00
-6.25626445e-01 1.61204159e+00 2.64841378e-01 -7.48137891e-01
5.94346583e-01 7.61419356e-01 1.24102131e-01 -1.06241000e+00
2.44010299e-01 2.53594577e-01 -3.38511765e-01 -7.32822120e-01
1.02192082e-01 1.33373335e-01 1.51109949e-01 5.19211590e-02
3.49728674e-01 3.17716599e-01 -4.11418676e-01 -1.34397909e-01
8.45301330e-01 -2.12703854e-01 -2.09309161e-01 -5.49166985e-02
5.76387346e-01 -8.85193706e-01 2.42577001e-01 3.80464911e-01
-7.50087619e-01 7.53304422e-01 4.59777504e-01 -1.22638613e-01
-7.84732699e-01 -1.29267132e+00 -1.64911956e-01 1.28493869e+00
-1.02505781e-01 -3.36585134e-01 -9.27442968e-01 -5.96354604e-01
-2.17731759e-01 5.49053907e-01 -4.84499842e-01 -2.16609061e-01
-2.47295335e-01 4.31049280e-02 1.21812332e+00 6.10850394e-01
6.59097314e-01 -1.31040120e+00 -4.70010251e-01 5.37233166e-02
-4.54730183e-01 -1.63705516e+00 -1.10004199e+00 -1.57019779e-01
-2.58404762e-01 -6.88647091e-01 -1.35260630e+00 -1.20341456e+00
-1.35289803e-01 1.47938088e-01 1.03566504e+00 -3.56777877e-01
-2.28545040e-01 3.85856211e-01 -1.03466526e-01 -3.60328346e-01
-3.51501167e-01 -3.03986333e-02 1.58217251e-01 1.98437750e-01
4.58900273e-01 -3.27266991e-01 -7.42522120e-01 3.38653594e-01
-3.59751791e-01 -2.35805772e-02 5.29670000e-01 9.81038988e-01
3.89119446e-01 -6.09485269e-01 8.54709506e-01 5.21599710e-01
7.13458300e-01 -1.93211451e-01 -5.99376798e-01 4.27771688e-01
-3.57613415e-01 -1.40088335e-01 1.13795586e-01 -6.48006618e-01
-7.81025529e-01 -1.55417904e-01 -6.03063047e-01 -9.05717194e-01
-2.66525924e-01 3.59806083e-02 -3.32201511e-01 -1.95406884e-01
9.25347209e-02 6.43059611e-01 1.66823089e-01 -4.17960405e-01
3.86290401e-01 1.50383663e+00 9.34439898e-01 -6.66582286e-02
-2.15517685e-01 2.13623106e-01 -5.67258060e-01 -1.26054728e+00
-2.43120119e-01 -5.39920628e-01 -1.54759586e-01 -3.73133570e-01
1.17891943e+00 -1.19320607e+00 -1.06449866e+00 1.13907301e+00
-1.50019538e+00 -8.87034237e-01 1.61364943e-01 4.13095385e-01
-7.33783484e-01 1.30911142e-01 -6.92950368e-01 -7.41273940e-01
-5.69368243e-01 -1.39001703e+00 1.44156826e+00 9.83498618e-02
2.46569812e-01 -7.56096601e-01 -7.26789087e-02 5.18584907e-01
6.61497414e-01 -4.08943892e-01 3.89513761e-01 -4.40011978e-01
-3.50744605e-01 6.02274425e-02 -2.91555911e-01 7.11754680e-01
-5.46320267e-02 -1.61768034e-01 -1.55711401e+00 -3.83844465e-01
-2.42378458e-01 -8.49982917e-01 1.29204166e+00 6.58655584e-01
1.38784921e+00 -4.73039329e-01 -2.10628122e-01 8.59310806e-01
1.11036801e+00 3.04988861e-01 8.16123605e-01 -1.99582592e-01
4.69377339e-01 4.29560781e-01 1.59604967e-01 6.35834411e-02
4.05203849e-01 9.94263411e-01 3.34108472e-01 -2.29492664e-01
-9.15911496e-01 -3.49654794e-01 8.34796906e-01 5.50499260e-01
6.13523424e-01 -6.15903497e-01 -9.91689265e-01 9.78351355e-01
-1.43633711e+00 -9.87796426e-01 5.02553940e-01 1.98990381e+00
8.09866726e-01 1.78136259e-01 1.65575802e-01 6.82351552e-03
8.27054739e-01 7.47545242e-01 -7.41068482e-01 -7.07119465e-01
-1.74711302e-01 3.00863713e-01 4.40740794e-01 7.68014848e-01
-1.08151388e+00 1.24562538e+00 6.50644636e+00 1.04920685e+00
-1.72784078e+00 4.43644613e-01 9.34481561e-01 -4.92016524e-01
1.28582790e-01 -7.74901450e-01 -7.65219510e-01 4.06768918e-01
1.67045617e+00 3.15407455e-01 3.73402804e-01 6.51297390e-01
3.44002396e-01 1.52287051e-01 -8.85401011e-01 1.47953641e+00
6.38547719e-01 -1.58747828e+00 -3.36474955e-01 3.24852057e-02
4.57184404e-01 6.39670074e-01 6.50407493e-01 2.81923592e-01
-4.73338276e-01 -1.43639743e+00 1.11491370e+00 2.29323685e-01
1.72604072e+00 -5.41685283e-01 4.44359273e-01 7.78594613e-02
-1.28555572e+00 2.35392321e-02 -3.08809690e-02 4.54236299e-01
1.24187067e-01 -6.25773612e-03 -7.47451067e-01 -1.33498624e-01
8.14855754e-01 6.19797111e-01 -3.12276423e-01 7.54032850e-01
-9.66895148e-02 7.80548990e-01 -1.18065447e-01 -2.02216297e-01
4.22654897e-01 7.39792585e-01 5.31713963e-01 1.59364939e+00
1.23387195e-01 -1.77929416e-01 -4.59324807e-01 5.14767230e-01
-5.12974083e-01 -9.99550223e-02 -9.06747282e-01 -1.19849540e-01
3.35047185e-01 6.73152447e-01 6.67848289e-02 -2.71082133e-01
-5.51696897e-01 1.17338717e+00 1.48103595e-01 8.53099465e-01
-1.15307283e+00 -6.07828796e-01 1.07088375e+00 -2.55071104e-01
7.06630111e-01 -2.16158196e-01 6.79744333e-02 -8.65346253e-01
3.21227014e-01 -1.06367016e+00 1.20180016e-02 -1.03916705e+00
-9.67529893e-01 7.14708507e-01 -4.46080238e-01 -7.36950934e-01
-3.95838976e-01 -9.08570170e-01 -5.54598927e-01 8.87826681e-01
-1.86958849e+00 -1.17878723e+00 -1.77458510e-01 9.16679859e-01
1.18259752e+00 -5.17176092e-01 8.12904358e-01 2.92915583e-01
-4.82438982e-01 1.48139322e+00 1.99208468e-01 3.80961627e-01
7.99180806e-01 -6.84922576e-01 9.29994404e-01 8.15523446e-01
3.31504256e-01 1.58873454e-01 3.73637110e-01 -2.94201225e-01
-1.32721150e+00 -1.16508627e+00 1.16312563e+00 -8.99534766e-03
5.50012767e-01 -9.42611933e-01 -5.94774365e-01 3.95093560e-01
7.95668125e-01 6.62375629e-01 4.85648543e-01 -3.46823037e-01
-6.71615303e-01 -3.86960924e-01 -9.74118114e-01 5.11168659e-01
1.16600752e+00 -1.20123732e+00 -3.68916243e-01 2.35887513e-01
1.07387376e+00 -3.83597404e-01 -4.29168463e-01 5.02085865e-01
8.86766851e-01 -7.78639257e-01 1.20225823e+00 -7.35829115e-01
-3.05338763e-02 2.20245913e-01 -3.74549419e-01 -1.18101501e+00
3.85137349e-01 -1.07798362e+00 -3.47607344e-01 1.08128405e+00
7.83256710e-01 -5.48862875e-01 7.96976626e-01 1.34837748e-02
-2.37179637e-01 -9.80092108e-01 -1.57838941e+00 -1.06325459e+00
2.27093741e-01 -8.37715387e-01 2.86804944e-01 2.51442075e-01
1.17047943e-01 2.53384233e-01 -4.72326458e-01 4.12652902e-02
5.14930964e-01 -3.76323700e-01 3.33524197e-01 -3.59857440e-01
6.66306242e-02 -5.20810068e-01 -4.84526485e-01 -1.62157166e+00
9.23309982e-01 -5.85386515e-01 3.21873307e-01 -1.41715264e+00
-1.77036360e-01 1.35057107e-01 -3.66248995e-01 4.32837039e-01
2.39259109e-01 3.94641012e-01 1.33867040e-01 -2.44656384e-01
-5.08337617e-01 7.70588458e-01 7.38717258e-01 -6.27160728e-01
8.02210495e-02 -7.09306374e-02 -1.58372760e-01 4.29460943e-01
9.55775380e-01 -2.16952294e-01 -4.09798622e-01 -5.61515331e-01
-5.73838651e-01 3.55477512e-01 6.11029923e-01 -9.27820563e-01
4.54464018e-01 1.57825619e-01 1.99578740e-02 -8.03673208e-01
9.58612800e-01 -2.90890038e-01 -5.38363039e-01 2.51486003e-01
-7.63964236e-01 -7.44238496e-02 6.27405286e-01 5.41530848e-01
-3.51581514e-01 3.74245197e-01 8.91519487e-01 4.46581185e-01
-7.83402205e-01 3.69704038e-01 -5.07634640e-01 1.67672306e-01
8.06751668e-01 -2.58159302e-02 -7.35600114e-01 -6.64783835e-01
-6.87398911e-01 -2.72332698e-01 2.29840856e-02 8.41699958e-01
1.13898230e+00 -1.41154850e+00 -8.54314506e-01 5.36907792e-01
1.79269761e-01 -8.35076511e-01 1.86080858e-01 7.80254781e-01
-1.33481637e-01 1.09237504e+00 5.59634455e-02 -8.51783156e-01
-1.68853760e+00 5.46373785e-01 7.43508577e-01 3.46562654e-01
-5.51106215e-01 1.35614932e+00 -5.53774042e-03 5.61835021e-02
1.19566178e+00 -3.81241113e-01 -1.93039533e-02 -9.82583091e-02
5.84808826e-01 1.52095884e-01 2.56971270e-01 -8.51610363e-01
-7.90446043e-01 5.48798084e-01 2.31800169e-01 -7.14349806e-01
7.88424313e-01 -2.86573410e-01 5.27875900e-01 5.07293165e-01
1.81812358e+00 -8.29540789e-02 -1.23305666e+00 -8.29312727e-02
-3.42268288e-01 -2.71641195e-01 4.23244447e-01 -9.68234479e-01
-1.18646753e+00 1.49652815e+00 1.24460256e+00 -1.73008248e-01
1.03530228e+00 1.63644642e-01 1.08866620e+00 2.81230778e-01
1.70419235e-02 -9.78151441e-01 3.24856699e-01 5.35601556e-01
1.20175433e+00 -1.54830337e+00 -8.06679308e-01 2.37870276e-01
-5.75397789e-01 9.69465077e-01 2.85350531e-01 3.31117988e-01
7.44865358e-01 4.22695488e-01 4.52571183e-01 3.28675926e-01
-1.17491949e+00 -2.39382625e-01 5.92336237e-01 7.76876211e-01
2.62650758e-01 9.75510944e-03 2.61688292e-01 2.69792646e-01
1.24184377e-01 1.12236217e-02 -6.76488280e-02 3.35909754e-01
-1.86004743e-01 -5.18651068e-01 -7.09698498e-02 -1.30186200e-01
-7.17098951e-01 -5.68300903e-01 -4.71365094e-01 4.27731603e-01
-2.57293105e-01 1.44666541e+00 2.00883985e-01 -4.34567958e-01
1.01276442e-01 3.24190766e-01 3.02531391e-01 9.08041745e-03
-1.82180718e-01 1.48206130e-01 4.32534575e-01 -9.08871770e-01
-1.05451234e-01 -5.76348662e-01 -9.43542242e-01 -2.10106447e-01
-1.86190873e-01 -2.51021951e-01 1.25228631e+00 5.86267650e-01
7.96628773e-01 3.98138493e-01 4.80565608e-01 -8.72014940e-01
-5.48271179e-01 -1.20803666e+00 -1.30090803e-01 -6.80190995e-02
1.18184543e+00 -4.94380206e-01 -5.78064442e-01 9.65454951e-02]
|
[14.337918281555176, 5.020552158355713]
|
28a65515-6c79-4147-b5d8-aa9d54d6d82e
|
emoberta-speaker-aware-emotion-recognition-in
|
2108.12009
| null |
https://arxiv.org/abs/2108.12009v1
|
https://arxiv.org/pdf/2108.12009v1.pdf
|
EmoBERTa: Speaker-Aware Emotion Recognition in Conversation with RoBERTa
|
We present EmoBERTa: Speaker-Aware Emotion Recognition in Conversation with RoBERTa, a simple yet expressive scheme of solving the ERC (emotion recognition in conversation) task. By simply prepending speaker names to utterances and inserting separation tokens between the utterances in a dialogue, EmoBERTa can learn intra- and inter- speaker states and context to predict the emotion of a current speaker, in an end-to-end manner. Our experiments show that we reach a new state of the art on the two popular ERC datasets using a basic and straight-forward approach. We've open sourced our code and models at https://github.com/tae898/erc.
|
['Piek Vossen', 'Taewoon Kim']
|
2021-08-26
| null | null | null | null |
['emotion-recognition-in-conversation']
|
['natural-language-processing']
|
[-3.87013882e-01 5.47067225e-02 2.70419955e-01 -9.99777019e-01
-8.62182856e-01 -6.50263965e-01 5.11973500e-01 -2.13085413e-01
-1.53463289e-01 3.32255512e-01 6.37276828e-01 -1.39828861e-01
4.20630455e-01 1.10407956e-01 -1.13041513e-01 -2.58196622e-01
-3.56495023e-01 3.89528841e-01 -5.26759028e-01 -5.46467066e-01
2.00024601e-02 2.61889279e-01 -1.48921871e+00 8.94882500e-01
2.66033560e-01 1.05921280e+00 -5.58593869e-01 1.28983307e+00
-3.82050842e-01 1.25026429e+00 -6.66687548e-01 -7.08536804e-01
-2.74436593e-01 -5.87124228e-01 -1.50172162e+00 1.44186392e-01
-7.13772774e-02 8.35298151e-02 -2.00539812e-01 5.89090884e-01
6.72760963e-01 4.20726806e-01 4.21291769e-01 -1.38946700e+00
-1.97365120e-01 7.57545233e-01 -1.99494049e-01 7.02257454e-02
8.39369237e-01 3.78730930e-02 8.36023271e-01 -1.04308867e+00
4.86240059e-01 1.52015817e+00 6.59957230e-01 1.21408057e+00
-9.37807381e-01 -5.95972955e-01 3.73710275e-01 3.13650936e-01
-1.11986911e+00 -1.36411452e+00 9.52604949e-01 -8.13669115e-02
1.34754193e+00 8.37658882e-01 5.41449726e-01 1.52938354e+00
-3.23741823e-01 1.16777349e+00 1.20466864e+00 -3.70710194e-01
3.98029953e-01 3.53164613e-01 4.47452843e-01 4.83139187e-01
-1.28845704e+00 -1.77797407e-01 -1.09719181e+00 -4.70133275e-01
-2.54519939e-01 -4.00005937e-01 -2.19590753e-01 2.57255882e-01
-7.20906198e-01 7.62526333e-01 -1.05015375e-01 2.73572892e-01
-4.47253197e-01 9.14408192e-02 9.50330019e-01 6.51496768e-01
7.72229433e-01 1.52144179e-01 -6.40766025e-01 -1.06241202e+00
-4.96868432e-01 7.12576229e-03 1.61623168e+00 7.64321029e-01
3.23442876e-01 2.44365982e-03 -5.67752123e-02 1.17190576e+00
3.73333305e-01 1.43965602e-01 5.19858003e-01 -1.13926494e+00
-2.15158947e-02 2.48583585e-01 1.85269043e-01 -6.29942715e-01
-4.90201950e-01 1.99276641e-01 -3.99766028e-01 -2.76786923e-01
1.35052830e-01 -8.43770325e-01 -5.09788632e-01 1.83596027e+00
5.55891871e-01 4.20136660e-01 5.32304525e-01 6.28391981e-01
1.17264891e+00 8.74512851e-01 4.13798213e-01 -2.31116593e-01
1.63218868e+00 -1.14561951e+00 -1.05522895e+00 -5.14497876e-01
7.94131279e-01 -7.89826989e-01 8.95124137e-01 3.09947163e-01
-1.13744271e+00 5.81961647e-02 -4.02082175e-01 -1.45212457e-01
-6.21225715e-01 2.90327668e-02 9.03014302e-01 7.26294339e-01
-1.20278108e+00 1.81648269e-01 -7.32367814e-01 -3.20665210e-01
1.07491752e-02 1.55082434e-01 -3.53150725e-01 4.47159082e-01
-1.36839616e+00 9.43735898e-01 -2.65707403e-01 4.65704441e-01
-6.97990358e-01 -4.57413763e-01 -9.80315864e-01 4.76218686e-02
2.51972347e-01 -1.08448751e-01 2.06470418e+00 -1.13489735e+00
-2.47277188e+00 1.10166621e+00 -8.53260934e-01 -2.47334138e-01
2.85523176e-01 -2.32276261e-01 -8.72417748e-01 -9.16558132e-02
-4.25027490e-01 5.64582348e-01 4.93297547e-01 -1.04466891e+00
-4.44955349e-01 -3.25447649e-01 -1.81189701e-01 3.52590799e-01
-1.83987617e-02 8.70331407e-01 -4.83637631e-01 4.76717576e-03
-2.60115743e-01 -9.14836287e-01 -8.46970528e-02 -4.56999898e-01
-4.55716789e-01 -7.63272047e-01 6.81412160e-01 -7.46703386e-01
1.02674448e+00 -2.41068625e+00 -1.95143963e-04 8.22761729e-02
1.26573816e-02 7.54122734e-02 -3.88303548e-01 6.46043658e-01
-3.75284553e-01 1.52810872e-01 -2.45408025e-02 -1.02791584e+00
4.66157258e-01 -4.48927619e-02 -3.34724158e-01 3.16255093e-01
5.44891693e-02 8.91959965e-01 -7.81938195e-01 -2.99925596e-01
1.74714550e-01 7.54306316e-01 -4.27169234e-01 6.35264933e-01
-8.75090528e-03 3.09233576e-01 -2.44026184e-01 4.29378390e-01
3.06596071e-01 1.93676099e-01 3.28547597e-01 2.95897782e-01
-1.17547005e-01 8.67032826e-01 -1.10393310e+00 1.57078278e+00
-8.95167887e-01 8.05469573e-01 8.10858071e-01 -7.00066626e-01
8.46370816e-01 8.99728835e-01 2.63718575e-01 -2.90334195e-01
4.76316988e-01 9.97194126e-02 -2.96214551e-01 -6.47220910e-01
5.28968692e-01 -3.45361829e-01 -6.36220276e-01 5.78395128e-01
1.59041733e-01 -2.60421425e-01 -4.05382141e-02 3.42523396e-01
8.75750184e-01 -2.79592633e-01 3.20480853e-01 2.40855012e-02
5.96530795e-01 -4.62183893e-01 4.84239310e-01 4.84026581e-01
-7.28440166e-01 -2.11948249e-02 6.47547781e-01 -1.66412801e-01
-2.12600574e-01 -5.18501043e-01 1.26267627e-01 1.76008022e+00
-4.65149075e-01 -6.46881819e-01 -8.58233094e-01 -6.90173984e-01
-2.46180758e-01 1.03185785e+00 -7.13794351e-01 -1.35186180e-01
-2.21682787e-01 -6.84580058e-02 7.73377538e-01 3.56685758e-01
3.74436647e-01 -1.37831640e+00 -2.89882421e-01 1.92236915e-01
-6.62471712e-01 -1.19454587e+00 -7.29876578e-01 4.00419772e-01
-2.11388558e-01 -5.86262524e-01 -2.18469918e-01 -7.45466113e-01
1.68659110e-02 -1.54469058e-01 1.35987270e+00 -1.43231943e-01
-1.21961750e-01 7.92863369e-01 -4.27969217e-01 -4.89569843e-01
-6.70895994e-01 -2.58644838e-02 -2.36562956e-02 1.75147876e-01
8.39636624e-01 -4.42008615e-01 -4.34333473e-01 1.96562320e-01
-1.42050833e-01 6.76208874e-03 -2.16944620e-01 4.05898929e-01
-1.56207591e-01 -5.14484346e-01 7.83562362e-01 -8.72035325e-01
9.67589676e-01 -5.85946321e-01 -7.05463141e-02 2.61856198e-01
6.85913637e-02 -2.89086491e-01 3.52568746e-01 -2.22592875e-01
-1.33530831e+00 1.70500085e-01 -7.42361784e-01 -2.12812543e-01
-7.11834490e-01 4.50407773e-01 -4.33558226e-02 4.04866844e-01
3.99582028e-01 2.02736735e-01 4.04501939e-03 -3.59002978e-01
7.09915698e-01 1.39064741e+00 5.76861262e-01 -6.90177441e-01
-1.01045385e-01 6.55801073e-02 -8.91315758e-01 -9.13758993e-01
-7.77652860e-01 -8.79172504e-01 -2.41921678e-01 -5.61437070e-01
7.48701990e-01 -1.06230891e+00 -1.61441183e+00 5.90541065e-01
-1.26094842e+00 -7.10153282e-01 -1.23210840e-01 1.90115333e-01
-5.48167884e-01 1.52893543e-01 -1.03038728e+00 -1.58124435e+00
-6.52948320e-01 -7.76043415e-01 1.03323817e+00 3.66354227e-01
-8.31876218e-01 -1.00413418e+00 1.97260156e-01 4.08000797e-01
5.91508865e-01 -2.14129195e-01 2.50672162e-01 -1.15047097e+00
4.77022588e-01 -1.99032366e-01 2.79766202e-01 2.64845967e-01
-7.39193261e-02 4.98479456e-02 -1.51222599e+00 1.48719400e-01
2.46377751e-01 -7.93873489e-01 4.37827796e-01 -1.08436428e-01
1.12935066e+00 -4.84838337e-01 -8.90058428e-02 1.51475862e-01
6.30205870e-01 3.64426613e-01 4.13337976e-01 -2.24845827e-01
5.05245179e-02 1.03233600e+00 3.60175461e-01 7.14657784e-01
8.28126013e-01 5.38965821e-01 -8.61623734e-02 4.07732576e-02
3.50931466e-01 3.07108909e-02 8.23173046e-01 9.54612374e-01
2.42950499e-01 -3.31629992e-01 -7.97131360e-01 6.64946556e-01
-1.79526329e+00 -1.17119622e+00 -4.28003892e-02 1.55790234e+00
1.28758895e+00 -4.20015782e-01 1.95639312e-01 -2.81614572e-01
6.11089766e-01 3.69990438e-01 -4.93308187e-01 -1.35137808e+00
1.35786265e-01 2.30385333e-01 -3.74453068e-01 9.32100058e-01
-1.06713450e+00 1.19830787e+00 6.71324682e+00 3.64985913e-01
-1.31634831e+00 1.72428131e-01 9.30929065e-01 -4.01748985e-01
-8.58942047e-02 -3.25222313e-01 -5.47883153e-01 2.59240776e-01
1.77575850e+00 -2.10232124e-01 8.07428360e-01 1.09734821e+00
3.00991386e-01 8.06771666e-02 -1.28873169e+00 1.15996659e+00
2.37841740e-01 -9.79630053e-01 -7.66640723e-01 -4.03221846e-01
2.71410018e-01 2.62946934e-01 -1.68560773e-01 9.56656635e-01
6.04325056e-01 -9.15987611e-01 6.50409997e-01 2.11475983e-01
4.28199917e-01 -7.88555920e-01 6.35968626e-01 1.42851204e-01
-9.78152990e-01 8.69933292e-02 2.22289756e-01 -1.12082437e-01
3.27439755e-01 1.83083698e-01 -8.52851272e-01 9.39919949e-02
8.72449338e-01 4.57890272e-01 -6.64558262e-02 3.03194404e-01
-2.67912149e-01 9.34743106e-01 -3.69329572e-01 -2.28793204e-01
1.46735340e-01 3.44442278e-02 5.45935988e-01 1.87395060e+00
-1.18735008e-01 6.15650356e-01 3.54412533e-02 5.42195261e-01
-4.06992257e-01 3.62820864e-01 -3.12386125e-01 -1.67038277e-01
5.58218598e-01 1.60960734e+00 -2.42605746e-01 -5.82469463e-01
-2.46288672e-01 1.36604774e+00 4.77145582e-01 5.02559125e-01
-8.95575345e-01 -4.45932627e-01 1.05473983e+00 -6.59093559e-01
1.20885201e-01 2.31897071e-01 4.68071289e-02 -1.16810763e+00
-1.76623300e-01 -1.30516648e+00 5.17507732e-01 -7.42612422e-01
-1.42009556e+00 8.89534771e-01 -4.64412183e-01 -4.60096955e-01
-6.49409652e-01 -5.57977319e-01 -1.20584929e+00 7.79178083e-01
-1.15193522e+00 -8.36750448e-01 -8.07563886e-02 7.77561247e-01
7.35062897e-01 1.39963090e-01 1.42108512e+00 3.47857624e-02
-9.09475267e-01 7.17153907e-01 -1.82774276e-01 3.39176685e-01
8.90476644e-01 -1.34617031e+00 4.29287761e-01 3.46617430e-01
-5.94310351e-02 7.50774920e-01 9.21459675e-01 -1.46239325e-01
-1.47719717e+00 -5.19559324e-01 1.35347092e+00 -4.07873958e-01
7.56525576e-01 -8.63086283e-01 -8.59019399e-01 1.08286822e+00
8.36661160e-01 -2.92122722e-01 1.30993831e+00 9.99118030e-01
-5.31152368e-01 2.04726279e-01 -1.15487087e+00 6.85570359e-01
7.11332202e-01 -1.04064929e+00 -6.02714062e-01 2.47076735e-01
5.48793495e-01 -5.13722241e-01 -8.99245560e-01 -3.91480364e-02
5.51328719e-01 -1.02376306e+00 5.55555284e-01 -8.26066494e-01
2.85203457e-01 3.23176205e-01 -1.07218996e-01 -1.49448490e+00
1.88521177e-01 -1.31056583e+00 -2.32403323e-01 1.54482007e+00
6.25909448e-01 -8.18907976e-01 5.23309767e-01 1.40754390e+00
-1.71893939e-01 -8.00796390e-01 -1.08808136e+00 -2.55042553e-01
-3.29277362e-03 -8.06301117e-01 6.76591456e-01 1.17007911e+00
1.03777742e+00 6.80001736e-01 -4.38923717e-01 -5.72559610e-02
-9.84435380e-02 5.32863326e-02 7.76048601e-01 -8.66952837e-01
-2.11951032e-01 -5.51390469e-01 5.15065193e-02 -9.90290761e-01
7.87213266e-01 -7.42474675e-01 5.02447724e-01 -1.08770645e+00
2.12727245e-02 -8.39251056e-02 -2.08573580e-01 7.48113811e-01
-1.14508949e-01 -5.42807095e-02 1.23946197e-01 -3.66620779e-01
-9.82528031e-01 7.16349185e-01 4.71013725e-01 1.44833447e-02
-4.62995440e-01 7.84664974e-02 -8.84418786e-01 7.91735113e-01
9.19049501e-01 -2.78227180e-01 -2.36972771e-03 3.24466974e-02
-5.32170609e-02 4.53626871e-01 1.65158063e-01 -4.23392981e-01
2.75274038e-01 2.34130733e-02 1.73787773e-02 -4.56703871e-01
9.32541847e-01 -5.23119390e-01 -1.86168849e-01 2.79729553e-02
-7.77882755e-01 -2.07394123e-01 5.95925570e-01 1.95170090e-01
-2.80942470e-01 -1.41419411e-01 5.78293383e-01 8.16765353e-02
-7.96757817e-01 -1.06795870e-01 -9.18617427e-01 9.56410915e-02
7.08628714e-01 4.09205049e-01 -2.10756794e-01 -1.02570844e+00
-1.11031497e+00 5.14615715e-01 -1.66471049e-01 6.60461724e-01
3.61222863e-01 -8.67090166e-01 -8.18886399e-01 -4.22500670e-02
2.53853083e-01 -5.28304935e-01 6.89577699e-01 7.87463784e-01
1.76876485e-01 3.44639152e-01 2.44117558e-01 -2.73404717e-01
-1.81383896e+00 1.52336270e-01 7.84515381e-01 -1.15222007e-01
-2.58233190e-01 1.19585717e+00 -1.66521475e-01 -1.04955316e+00
5.10390997e-01 -8.33461434e-03 -1.53973147e-01 1.96653426e-01
7.43700206e-01 1.46966651e-01 -4.53151055e-02 -7.99116850e-01
-7.05516338e-01 -1.87714621e-01 -2.78683037e-01 -4.99263853e-01
1.36120629e+00 -4.45584953e-01 -2.31045693e-01 8.61956656e-01
1.39187157e+00 2.89437175e-01 -7.54427791e-01 -1.47236586e-01
-1.45062739e-02 2.72600371e-02 1.91915706e-02 -1.23904753e+00
-8.42634559e-01 6.85837030e-01 5.27707934e-01 2.61689007e-01
9.84075487e-01 3.39273125e-01 7.72218347e-01 6.70577228e-01
-4.67070229e-02 -1.35347068e+00 -8.66025835e-02 8.00578177e-01
1.13224864e+00 -1.31629813e+00 -5.98733723e-01 -3.75634253e-01
-1.37441349e+00 9.15456712e-01 5.18203616e-01 4.27045405e-01
9.24743414e-01 4.03527558e-01 8.39577854e-01 -4.78839219e-01
-1.37395787e+00 -5.81362769e-02 -7.24958032e-02 1.92683890e-01
1.05504215e+00 4.18262273e-01 1.09168380e-01 9.08666372e-01
-4.41403300e-01 -1.78214788e-01 3.88067752e-01 8.61953020e-01
-7.74429739e-02 -1.02029359e+00 -6.67259246e-02 -1.36249334e-01
-6.11616910e-01 -1.53472140e-01 -1.02003932e+00 3.85518402e-01
-5.30321538e-01 1.64315796e+00 6.56693205e-02 -3.72703493e-01
5.35148978e-01 1.02023530e+00 -1.50516897e-01 -4.44257379e-01
-1.12493205e+00 1.19556032e-01 8.99653137e-01 -7.48851299e-01
-3.81732762e-01 -8.71178806e-01 -1.51109624e+00 -5.02067745e-01
-1.71790704e-01 6.84629321e-01 8.86195958e-01 8.32396328e-01
9.06046271e-01 3.05770397e-01 1.15797353e+00 -7.88980186e-01
-4.08003122e-01 -1.30867946e+00 -4.51155782e-01 4.62690055e-01
2.94585615e-01 -8.53969082e-02 -6.96863294e-01 -1.09438993e-01]
|
[13.018585205078125, 6.206577301025391]
|
e8b7a90a-4d27-4912-ab80-e75b3d2736e9
|
fast-fourier-color-constancy-and-grayness
|
1908.02076
| null |
https://arxiv.org/abs/1908.02076v2
|
https://arxiv.org/pdf/1908.02076v2.pdf
|
Fast Fourier Color Constancy and Grayness Index for ISPA Illumination Estimation Challenge
|
We briefly introduce two submissions to the Illumination Estimation Challenge, in the Int'l Workshop on Color Vision, affiliated to the 11th Int'l Symposium on Image and Signal Processing and Analysis. The Fourier-transform-based submission is ranked 3rd, and the statistical Gray-pixel-based one ranked 6th.
|
['Yanlin Qian', 'Ke Chen', 'Huanglin Yu']
|
2019-08-06
| null | null | null | null |
['color-constancy']
|
['computer-vision']
|
[ 3.22756290e-01 -6.49091482e-01 1.38437614e-01 -3.82943511e-01
-8.97479773e-01 -4.35341179e-01 3.78630906e-01 -8.93299356e-02
-5.20738661e-01 3.91570568e-01 -2.35524416e-01 3.87888253e-02
3.48572075e-01 -4.63186204e-02 -3.43884021e-01 -7.37374067e-01
-3.13867569e-01 -6.12474918e-01 1.24241719e-02 2.34740600e-01
5.15933156e-01 4.44976598e-01 -1.46544838e+00 3.64719898e-01
7.67082572e-01 1.13108778e+00 -2.71661133e-01 1.16364801e+00
-1.11447915e-01 2.97928959e-01 -3.22841018e-01 -1.44261599e-01
5.36224008e-01 -8.67949069e-01 -4.49071050e-01 1.04758166e-01
9.22946692e-01 -1.52669728e-01 -2.29356974e-01 1.39020002e+00
4.93183702e-01 1.96438149e-01 6.52752161e-01 -1.48917425e+00
-5.00324368e-01 3.55364643e-02 -1.14724362e+00 6.69755518e-01
4.76851434e-01 1.16192937e-01 6.29217565e-01 -1.40928996e+00
6.03906333e-01 1.37883162e+00 7.70103216e-01 1.49792418e-01
-1.15888143e+00 -2.88024813e-01 3.86376083e-01 6.49502575e-01
-1.32985222e+00 -5.84728718e-01 8.59019876e-01 -3.40913445e-01
6.54509902e-01 3.47512990e-01 8.08220267e-01 4.72760379e-01
3.76743138e-01 1.23734581e+00 1.98462605e+00 -8.65978360e-01
1.38301268e-01 -1.35007054e-01 1.14136010e-01 7.00422704e-01
6.94864094e-02 2.64076173e-01 -6.66047394e-01 -2.26593390e-01
8.25016260e-01 -8.44411135e-01 -2.74564654e-01 -1.91655591e-01
-1.36674869e+00 4.80492026e-01 2.22573042e-01 2.59768665e-01
-3.84100646e-01 5.30924797e-01 1.04888342e-01 3.01675618e-01
6.43177629e-01 1.88094571e-01 -2.13595465e-01 -1.43026993e-01
-9.67760265e-01 -8.17941576e-02 4.36998785e-01 6.02767885e-01
4.84148741e-01 2.66739547e-01 -2.68357366e-01 9.70199347e-01
5.22558928e-01 4.57276314e-01 1.65023301e-02 -1.21217906e+00
-1.80681095e-01 -1.11928821e-01 2.58836865e-01 -6.61454201e-01
-3.16429317e-01 -1.61405474e-01 -5.75426042e-01 6.29681587e-01
5.38104892e-01 -3.30409765e-01 -9.99207675e-01 1.23780274e+00
4.93156128e-02 3.60666007e-01 -4.03063297e-01 9.99135911e-01
1.00154448e+00 5.04273295e-01 -1.01745777e-01 -6.04590416e-01
1.30633128e+00 -8.48488033e-01 -7.12369859e-01 1.07862681e-01
-4.08097744e-01 -1.55747819e+00 4.89465773e-01 1.03834784e+00
-1.39311564e+00 -6.56137824e-01 -1.30167162e+00 1.77749731e-02
-3.34889084e-01 3.87872607e-01 5.81512570e-01 8.60169291e-01
-1.40672350e+00 2.08754987e-01 -5.10383964e-01 -5.93638539e-01
2.67662883e-01 -1.07433856e-01 1.46208093e-01 -3.00741494e-01
-7.55239189e-01 1.00657701e+00 -6.17015362e-02 5.97014800e-02
-5.23711860e-01 -6.10694826e-01 -5.73102295e-01 -6.11830354e-01
-1.80111080e-01 -2.62933165e-01 1.10558867e+00 -8.81497920e-01
-1.61299646e+00 1.31550324e+00 -3.82593393e-01 -2.14688674e-01
5.38843036e-01 -1.33865839e-02 -9.41344380e-01 1.43393889e-01
-1.14422441e-01 7.22672522e-01 1.08801496e+00 -1.50694668e+00
-7.83515811e-01 -3.81912172e-01 -1.24311537e-01 2.05328003e-01
1.76541671e-01 3.51006597e-01 -8.65912080e-01 -6.55613899e-01
4.29058373e-01 -6.15345955e-01 -2.20254257e-01 3.29112589e-01
-4.08176124e-01 -1.21760614e-01 9.11972225e-01 -4.50631887e-01
6.84526145e-01 -2.31333447e+00 -4.30343390e-01 2.45504260e-01
-1.24561816e-01 5.77174313e-02 -3.63956064e-01 2.07716703e-01
-2.93391079e-01 -3.01087588e-01 3.78220566e-02 -3.79132926e-01
8.12182575e-02 -1.21776365e-01 2.29032952e-02 9.72237766e-01
-5.73550798e-02 5.92022002e-01 -9.40893948e-01 -4.57687348e-01
5.08504033e-01 7.64795721e-01 1.75421625e-01 -3.13170880e-01
4.75916505e-01 3.14483583e-01 6.60803169e-02 7.61107802e-01
1.03937733e+00 1.87452108e-01 -2.11997434e-01 -8.30463648e-01
-8.05683494e-01 -7.39168003e-02 -1.27489662e+00 1.57513726e+00
-2.73748100e-01 1.22041762e+00 5.13269484e-01 -5.55578649e-01
6.56180978e-01 2.65815407e-01 7.90719628e-01 -9.44462299e-01
-1.12072594e-01 3.33165824e-01 -1.89821959e-01 -1.94682837e-01
4.37922567e-01 1.79400578e-01 3.43462646e-01 3.77620131e-01
-1.30346149e-01 -7.28499770e-01 6.30270422e-01 1.34844616e-01
7.28972018e-01 6.90915108e-01 2.07207501e-01 -5.62070310e-01
7.93454170e-01 -3.63836020e-01 5.42448819e-01 7.17340708e-01
-8.36623192e-01 1.03540242e+00 2.53358781e-01 -4.69660670e-01
-6.23387635e-01 -1.27496147e+00 -2.99984276e-01 1.18782711e+00
2.04210356e-01 -2.75245070e-01 -1.02234507e+00 -3.59534413e-01
-1.21663064e-01 2.86221117e-01 -4.50653493e-01 2.70003229e-01
-3.22456241e-01 -9.15600419e-01 3.59179437e-01 2.17393249e-01
6.94027781e-01 -7.11439550e-01 -7.91373253e-01 -7.69962445e-02
-3.15825492e-01 -1.11158133e+00 -9.03528333e-01 3.68164361e-01
-5.42237043e-01 -1.20309913e+00 -8.94393325e-01 -8.18402290e-01
8.58975351e-01 2.83890337e-01 1.08599281e+00 -1.82697639e-01
-1.20079637e+00 1.05010748e+00 -1.85813874e-01 -6.58406794e-01
4.13508117e-01 -8.18816304e-01 -1.71275288e-01 2.91948617e-01
2.62446046e-01 -1.96043700e-01 -1.17825949e+00 2.53268361e-01
-3.56670260e-01 -1.48797214e-01 5.48098922e-01 6.09610319e-01
9.02907312e-01 -3.10016684e-02 2.15045691e-01 -6.71647847e-01
4.90391254e-01 4.81234223e-01 -9.78087783e-01 2.50553310e-01
-4.00136054e-01 -3.52492690e-01 4.32798594e-01 1.41203880e-01
-1.13806474e+00 3.66544366e-01 1.44252032e-02 1.09180331e-01
-1.06586248e-01 3.67580727e-02 2.23392606e-01 -8.49228382e-01
5.99086642e-01 3.39861393e-01 -5.03116667e-01 -3.11186135e-01
5.43262780e-01 9.87671837e-02 1.04672289e+00 -4.92937148e-01
8.24296951e-01 8.92856956e-01 1.41971141e-01 -1.06791139e+00
-5.51317692e-01 -8.22256267e-01 -6.26910329e-01 -4.88952398e-01
1.08206749e+00 -6.99037611e-01 -5.95095098e-01 9.82622564e-01
-1.21868074e+00 -1.87500715e-01 -2.92562217e-01 7.40177155e-01
-5.92128456e-01 4.45378572e-01 -4.82123286e-01 -8.31235111e-01
-1.47970393e-01 -1.05822039e+00 8.72306466e-01 3.56546193e-01
3.95846814e-02 -1.06520152e+00 3.90097260e-01 2.50884220e-02
5.69090426e-01 3.54616314e-01 6.18944824e-01 2.48603702e-01
-3.45816582e-01 -1.68600008e-01 -6.14519835e-01 5.87828875e-01
1.37870729e-01 4.72691983e-01 -1.38588595e+00 -2.25843579e-01
-2.61604249e-01 -9.97537822e-02 1.04851115e+00 1.15245557e+00
1.10897648e+00 6.86177969e-01 -5.90693876e-02 6.39385343e-01
1.63499391e+00 3.29602748e-01 7.07829118e-01 1.14524297e-01
2.49360368e-01 3.40853870e-01 4.44103956e-01 3.85446399e-01
2.91260302e-01 6.27607524e-01 2.28546441e-01 -6.47136807e-01
-8.08888078e-01 2.87567288e-01 2.51330823e-01 6.94871604e-01
-3.74153316e-01 2.13057976e-02 -5.24913192e-01 2.90296018e-01
-1.36479497e+00 -6.84985876e-01 -6.57191813e-01 2.33026600e+00
4.73473877e-01 -1.64450899e-01 2.98333317e-01 2.04897955e-01
6.94745243e-01 2.94408292e-01 -3.12147707e-01 -6.05262876e-01
-3.28032374e-01 5.98427653e-01 6.53098524e-01 6.18215144e-01
-1.51896489e+00 3.75294894e-01 8.62767410e+00 4.58915383e-01
-1.02500474e+00 6.65258616e-02 6.07733011e-01 9.54429433e-02
-1.67393818e-01 -3.63846757e-02 -4.57780778e-01 3.03215086e-01
5.11735082e-01 -2.53917634e-01 4.57272768e-01 1.33553103e-01
4.46456730e-01 -9.42916155e-01 -9.12723780e-01 1.31947196e+00
4.41538244e-01 -7.37648904e-01 -5.90322316e-01 -1.97836477e-03
9.04786646e-01 2.16990933e-01 3.57938260e-01 -2.71601826e-01
8.90615731e-02 -1.08732831e+00 5.98202586e-01 5.19517541e-01
6.93428040e-01 -7.26398349e-01 2.55189806e-01 -5.14117599e-01
-1.56045723e+00 3.57434541e-01 -2.32279256e-01 2.66258150e-01
9.90815014e-02 8.75342369e-01 4.74854968e-02 3.55203688e-01
9.63811517e-01 7.57697284e-01 -5.65656960e-01 1.75003052e+00
-2.68424183e-01 4.37126338e-01 -1.95326939e-01 1.64371610e-01
-1.25647923e-02 -3.09947163e-01 6.78656399e-01 1.69593322e+00
-1.58524606e-02 4.75494683e-01 3.69037926e-01 4.72375989e-01
-1.12656370e-01 -7.22822249e-02 -1.32509004e-02 2.75502592e-01
-2.26016302e-04 1.75305843e+00 -9.38644886e-01 -2.59578526e-01
-7.11408079e-01 1.23183274e+00 -6.02678001e-01 7.65457332e-01
-6.84393823e-01 -8.86327207e-01 2.94237882e-01 -4.24945325e-01
1.86235860e-01 -3.03746760e-01 -6.04935646e-01 -8.79567027e-01
-2.38446862e-01 -5.60145974e-01 3.57774734e-01 -1.14505231e+00
-1.34085071e+00 3.69650394e-01 8.92282650e-03 -1.04703283e+00
2.19426990e-01 -1.00105119e+00 -9.29133236e-01 1.09823906e+00
-1.82839406e+00 -9.26915050e-01 -4.41424638e-01 8.05024981e-01
3.10299337e-01 1.05110370e-01 8.21471453e-01 3.83155465e-01
-2.68283933e-01 6.11527443e-01 1.72726080e-01 -9.84380320e-02
1.02409506e+00 -1.63284349e+00 4.86088723e-01 1.28419304e+00
2.31542125e-01 2.05451533e-01 7.80848563e-01 1.23897843e-01
-1.84429002e+00 -6.16040230e-01 7.06832767e-01 -2.63621360e-02
3.97776514e-01 -1.08400509e-01 -4.83828723e-01 3.20041090e-01
9.43241835e-01 3.82194698e-01 5.35613060e-01 -2.29648292e-01
-3.81710351e-01 -2.13616610e-01 -1.25014973e+00 2.56026059e-01
5.04469097e-01 -4.77568477e-01 -2.31511950e-01 3.85631204e-01
-2.19045401e-01 -5.35460830e-01 -8.34714949e-01 1.02224030e-01
8.66518915e-01 -1.17911458e+00 1.10646021e+00 3.07379737e-02
3.85140739e-02 -6.53608859e-01 -2.97193050e-01 -1.37849534e+00
-3.24281365e-01 -8.16793978e-01 4.53825027e-01 7.57577717e-01
4.54905033e-02 -6.23064995e-01 5.51447034e-01 1.91614464e-01
-1.81870148e-01 -2.97357053e-01 -9.60209608e-01 -7.95216560e-01
-1.41578123e-01 -4.82304960e-01 -2.51540452e-01 4.28231508e-01
-2.12089628e-01 -3.72691870e-01 -2.62762785e-01 -2.61292219e-01
1.35965872e+00 2.09324107e-01 5.56745827e-01 -1.00941575e+00
-9.74406824e-02 -8.45318794e-01 -6.83157444e-01 -8.52866292e-01
-5.77412963e-01 -2.71833301e-01 5.16662419e-01 -1.84371674e+00
1.34316206e-01 2.19983518e-01 -9.03139770e-01 2.04566672e-01
-1.67433977e-01 1.07918298e+00 2.71462172e-01 -3.27333391e-01
-7.49652624e-01 -2.21577615e-01 1.05715144e+00 -2.84465551e-01
1.11334272e-01 -1.66418076e-01 -4.97212261e-01 8.14784288e-01
3.75018120e-01 2.42323682e-01 -2.48692259e-01 -3.75493169e-01
-1.09146908e-01 -4.50189829e-01 4.41159934e-01 -8.22391629e-01
3.50210905e-01 -2.61990100e-01 8.94249558e-01 -8.23517144e-01
5.68105102e-01 -3.90469611e-01 -2.33733490e-01 5.03321350e-01
-3.27333286e-02 -9.13363248e-02 3.48314494e-01 -5.70606291e-02
-5.31694055e-01 2.38072604e-01 1.56069922e+00 -2.46655345e-01
-1.15650570e+00 4.24776554e-01 -7.22151458e-01 1.87820464e-01
9.41446722e-01 -5.31414390e-01 -5.60313344e-01 -3.39192748e-01
-8.50881040e-01 -6.89142272e-02 1.62573814e-01 2.32937455e-01
9.54962313e-01 -1.31954610e+00 -8.60955834e-01 3.41231018e-01
2.13706959e-02 -9.60356116e-01 1.80754706e-01 1.49279797e+00
-6.29100859e-01 2.23828375e-01 -4.03212070e-01 -5.88529587e-01
-1.27365804e+00 -1.79243430e-01 3.73897523e-01 2.42947057e-01
-4.17067856e-01 1.26661718e+00 7.98381194e-02 2.18529552e-01
1.37595475e-01 -1.61611840e-01 7.19734505e-02 -1.78219274e-01
7.47416258e-01 9.47700739e-01 1.19792342e-01 -5.05107522e-01
-7.81471133e-01 1.10002434e+00 1.39484346e-01 -6.04542792e-01
1.11670411e+00 -2.52477884e-01 -5.74548304e-01 6.43459320e-01
1.43960977e+00 8.98597464e-02 -1.17332697e+00 -1.19754761e-04
-6.66860044e-02 -8.12051594e-01 5.86049139e-01 -1.13347507e+00
-1.14582193e+00 9.28870440e-01 1.10523367e+00 2.25834355e-01
1.73275650e+00 -5.72929740e-01 7.07179010e-01 -2.88142674e-02
3.34821135e-01 -1.66230452e+00 4.34435420e-02 4.36201155e-01
8.65723789e-01 -1.15291107e+00 5.33360302e-01 -5.11253059e-01
-2.38604903e-01 1.48768806e+00 2.91699201e-01 -1.69008926e-01
8.76629531e-01 3.71518731e-01 5.37609458e-01 2.06442460e-01
-5.32992005e-01 -3.70890707e-01 8.84192169e-01 1.15473199e+00
9.08780277e-01 2.82070576e-03 -1.56153515e-01 -2.61603802e-01
1.96294531e-01 -1.07715234e-01 3.50154340e-01 6.08437300e-01
-6.46454632e-01 -9.72083092e-01 -6.08049214e-01 2.21627519e-01
-5.42434454e-01 -2.18064591e-01 -5.20015597e-01 4.90018606e-01
2.66908109e-01 1.19575870e+00 -2.64790840e-02 1.40729204e-01
4.39498276e-01 -1.54299170e-01 1.19781983e+00 -1.56044915e-01
-2.44558334e-01 4.44636375e-01 1.81461107e-02 -1.15561473e+00
-8.48915637e-01 -7.29319096e-01 -9.60215092e-01 -1.52449384e-01
1.32617205e-01 -4.76950854e-02 1.18617761e+00 3.87326330e-01
-1.82965353e-01 4.15351480e-01 8.96955669e-01 -9.29422319e-01
7.31515735e-02 -4.71222878e-01 -9.45159495e-01 5.89111708e-02
3.98220956e-01 -4.94719982e-01 -3.96492064e-01 5.22761583e-01]
|
[10.506363868713379, -2.564770460128784]
|
1b7c7f30-afcf-4d66-8fef-a75abd0a9caa
|
approaching-sign-language-gloss-translation
| null | null |
https://aclanthology.org/2021.mtsummit-at4ssl.7
|
https://aclanthology.org/2021.mtsummit-at4ssl.7.pdf
|
Approaching Sign Language Gloss Translation as a Low-Resource Machine Translation Task
|
A cascaded Sign Language Translation system first maps sign videos to gloss annotations and then translates glosses into a spoken languages. This work focuses on the second-stage gloss translation component, which is challenging due to the scarcity of publicly available parallel data. We approach gloss translation as a low-resource machine translation task and investigate two popular methods for improving translation quality: hyperparameter search and backtranslation. We discuss the potentials and pitfalls of these methods based on experiments on the RWTH-PHOENIX-Weather 2014T dataset.
|
['Kevin Duh', 'Xuan Zhang']
| null | null | null | null |
mtsummit-2021-8
|
['sign-language-translation']
|
['computer-vision']
|
[ 4.81080174e-01 -2.70548612e-01 -5.00595212e-01 -4.65930015e-01
-1.31691003e+00 -7.90392816e-01 7.80828416e-01 -8.77474487e-01
-6.87084198e-01 7.10684597e-01 7.31587470e-01 -3.16381454e-01
2.85653830e-01 -1.36388347e-01 -5.73892534e-01 -5.02946198e-01
4.06998008e-01 9.15688276e-01 4.04029340e-02 -2.34573528e-01
-1.10942358e-02 -2.89022699e-02 -1.28103554e+00 4.76584613e-01
9.59804654e-01 5.69570422e-01 -2.09342554e-01 7.83684433e-01
-5.59459627e-02 3.35640877e-01 -6.89872444e-01 -9.29734945e-01
6.11059964e-01 -8.37662995e-01 -7.66131520e-01 -1.26411155e-01
9.24756944e-01 -7.76706994e-01 -1.38724387e-01 8.26019168e-01
9.20758665e-01 -2.16509357e-01 5.96920848e-01 -1.27452850e+00
-7.84087002e-01 5.00436366e-01 -4.70949672e-02 -1.27486825e-01
3.84225160e-01 4.23601359e-01 1.17237961e+00 -8.74390900e-01
1.18695891e+00 1.24938452e+00 6.58529639e-01 7.63161540e-01
-7.58513153e-01 -7.30157554e-01 -9.01992768e-02 1.51331738e-01
-1.05984545e+00 -4.97280389e-01 1.98150352e-01 -1.32330358e-01
1.09837723e+00 6.38980940e-02 9.47694302e-01 1.62302601e+00
4.70841676e-02 9.95307982e-01 1.50386047e+00 -5.21912873e-01
-7.25644901e-02 -3.20554405e-01 -3.51543158e-01 8.48847508e-01
1.93325594e-01 2.03736797e-01 -1.26803529e+00 -2.27329478e-01
5.73954105e-01 -5.91753840e-01 -3.58748555e-01 8.58227164e-02
-1.39704728e+00 5.97195864e-01 5.24282120e-02 -9.76716131e-02
-2.82418996e-01 3.46882224e-01 3.67439419e-01 8.92676055e-01
4.45751160e-01 2.53017634e-01 -5.42597532e-01 -5.96023023e-01
-9.73218918e-01 2.35747114e-01 8.66774380e-01 1.15118158e+00
4.16725874e-01 -4.30333391e-02 -1.60282314e-01 7.04421461e-01
4.81147498e-01 1.18008137e+00 4.80771303e-01 -7.37448752e-01
8.63445818e-01 1.44249603e-01 4.03052345e-02 -1.21906772e-01
-7.64612630e-02 1.12921007e-01 -1.01625443e-01 1.07155129e-01
5.09353876e-01 -3.43645871e-01 -1.44530821e+00 1.53742385e+00
3.38323981e-01 -2.87144668e-02 1.52312845e-01 1.26738143e+00
7.75089145e-01 2.24403188e-01 3.68234254e-02 -1.09532051e-01
1.21533740e+00 -1.42914903e+00 -8.39955509e-01 -1.67232335e-01
4.96179253e-01 -1.29860055e+00 1.20751369e+00 2.94746935e-01
-9.73088980e-01 7.75021836e-02 -6.98712111e-01 -3.30107927e-01
-2.26596624e-01 5.06928742e-01 5.95453024e-01 3.89057934e-01
-1.17481840e+00 3.90069515e-01 -1.14849627e+00 -8.56553733e-01
1.40104234e-01 3.50402832e-01 -1.99540824e-01 -3.12278897e-01
-1.02560604e+00 1.22487581e+00 -6.77600950e-02 3.13794583e-01
-4.77235973e-01 -3.95271450e-01 -4.35652167e-01 -6.31446540e-01
7.57313818e-02 -1.11130285e+00 1.53566289e+00 -1.07038832e+00
-2.45418453e+00 9.72062290e-01 -3.26364666e-01 -8.12113658e-02
1.01560986e+00 -5.10413945e-01 -2.62849391e-01 1.68847278e-01
-1.83194995e-01 6.31949961e-01 9.78633046e-01 -7.92638838e-01
-6.86786950e-01 -1.77631512e-01 -3.37220222e-01 6.58687055e-01
1.08478703e-01 4.35245872e-01 -7.81409800e-01 -6.41136646e-01
3.42858694e-02 -1.57496405e+00 9.41898748e-02 5.21730259e-02
-4.31532152e-02 -1.76093914e-02 4.67410028e-01 -1.17379081e+00
9.02325451e-01 -1.59850132e+00 6.32748187e-01 1.92379937e-01
-2.89154708e-01 3.69642258e-01 -4.86404568e-01 5.80457807e-01
6.40611410e-01 -5.62063977e-02 -2.18713403e-01 -6.55452192e-01
1.55208036e-01 5.71568668e-01 -5.43851793e-01 2.61100978e-01
1.25821859e-01 1.37111688e+00 -9.27835703e-01 -4.92772281e-01
-8.23106244e-02 6.74731314e-01 -4.61084217e-01 1.10504217e-01
-3.16615641e-01 5.52806556e-01 -1.96760908e-01 1.15435588e+00
1.21918648e-01 2.18533307e-01 2.50826746e-01 -3.38894308e-01
-2.32737258e-01 7.61278987e-01 -3.26058239e-01 2.02260995e+00
-3.25523198e-01 8.12196195e-01 -2.07430780e-01 -1.87703580e-01
6.16161466e-01 4.13265377e-01 3.08452189e-01 -6.25773489e-01
3.03020149e-01 8.02393734e-01 -2.83775300e-01 -7.60009766e-01
3.79512221e-01 -2.11525247e-01 6.54650256e-02 7.87502527e-01
2.63071388e-01 -6.26078665e-01 2.09275663e-01 -2.64069676e-01
1.01448512e+00 8.42376411e-01 1.18949763e-01 1.71413347e-01
-1.96479350e-01 3.28069687e-01 3.95527393e-01 4.46922362e-01
-3.79478961e-01 7.27658927e-01 -1.63222868e-02 -3.14093709e-01
-1.01195884e+00 -1.05984271e+00 4.26272511e-01 1.26557136e+00
-2.52508402e-01 -5.12646854e-01 -5.68674803e-01 -7.39317000e-01
-1.40836447e-01 3.62348050e-01 -4.39875066e-01 -3.51949558e-02
-9.08182025e-01 -6.68564081e-01 1.09604418e+00 4.05814171e-01
4.92741823e-01 -1.04829836e+00 -6.90058827e-01 -3.04670781e-02
-8.61432374e-01 -1.31194794e+00 -8.39637876e-01 -3.11840065e-02
-8.79610717e-01 -8.03369164e-01 -1.06182623e+00 -9.50688481e-01
3.90150011e-01 -1.08794488e-01 1.11143351e+00 -2.71535311e-02
-8.88680816e-02 4.26819652e-01 -5.74594080e-01 -5.05837202e-01
-5.52321970e-01 2.20614538e-01 2.09860727e-01 -1.73620030e-01
5.67961633e-01 -2.92624682e-01 -4.49886322e-01 3.66023034e-01
-6.20433927e-01 1.39108181e-01 1.03619921e+00 8.40385437e-01
6.98143005e-01 -1.04399776e+00 2.47940980e-02 -3.47347468e-01
8.20364833e-01 2.29280934e-01 -6.34160519e-01 7.41306007e-01
-7.94724405e-01 3.55940551e-01 6.05115388e-03 -6.86565161e-01
-9.68334317e-01 1.14383310e-01 7.08311051e-02 -3.66857797e-01
3.58104557e-01 5.25196552e-01 2.19217554e-01 -2.68388480e-01
4.76327568e-01 3.58639359e-01 4.51999996e-03 -3.29119146e-01
5.47320843e-01 7.70448029e-01 4.43656862e-01 -5.66416204e-01
9.81879175e-01 3.67849350e-01 -2.87930757e-01 -7.48306572e-01
-6.44075453e-01 -1.87949061e-01 -7.06609964e-01 -3.53235751e-01
9.00936007e-01 -9.14403856e-01 -3.95915866e-01 5.28251648e-01
-1.06040049e+00 -8.75900269e-01 -2.52645612e-01 7.71603167e-01
-8.06729794e-01 3.44054908e-01 -7.72131801e-01 -3.15604001e-01
-8.41613889e-01 -1.23991215e+00 1.57260525e+00 -1.43450141e-01
-4.17064041e-01 -5.64098299e-01 6.12690091e-01 7.02173293e-01
5.85766852e-01 -1.76030576e-01 6.40854239e-01 -5.10283485e-02
-6.22523189e-01 8.11541229e-02 -2.68966779e-02 2.67241269e-01
1.66393727e-01 2.27511689e-01 -6.55915797e-01 -2.57674217e-01
-6.00613773e-01 -7.92074680e-01 9.60562110e-01 1.67374581e-01
-1.18853346e-01 -4.68344420e-01 1.67323992e-01 7.96914995e-01
1.07433367e+00 -3.51177484e-01 4.68810409e-01 3.02114218e-01
4.36989099e-01 3.24608684e-01 7.29186118e-01 -7.20284209e-02
7.02082336e-01 8.86097968e-01 -1.40185282e-01 1.29698226e-02
-8.09981167e-01 -5.36256135e-01 8.01264465e-01 1.39660418e+00
-7.06881642e-01 -2.08234474e-01 -8.02865505e-01 4.12720352e-01
-1.95658219e+00 -6.83304548e-01 2.57073671e-01 1.93963170e+00
1.13017035e+00 -3.37441415e-01 1.92738995e-01 -3.51075143e-01
1.60468370e-01 -1.31576478e-01 -3.87272626e-01 -4.08532381e-01
-3.34176749e-01 3.96725684e-01 6.80777669e-01 6.06410146e-01
-7.28811741e-01 1.72765017e+00 7.51313400e+00 3.29725623e-01
-1.26472306e+00 1.75611138e-01 -4.01956201e-01 -4.38085496e-01
-1.06726252e-01 8.26984122e-02 -7.00700283e-01 8.68955925e-02
8.63032579e-01 1.83452144e-01 9.83065665e-01 3.60831469e-01
1.16152890e-01 5.34345619e-02 -1.03986788e+00 9.63660777e-01
4.21124429e-01 -9.35894251e-01 3.36841464e-01 2.83588916e-02
8.98700476e-01 7.96324492e-01 1.35854378e-01 3.42979997e-01
7.13304818e-01 -6.23078525e-01 8.27428877e-01 4.65918511e-01
1.18612063e+00 -2.37869903e-01 5.40129781e-01 -3.28115731e-01
-1.05303025e+00 3.00982744e-01 3.77344638e-02 2.61640042e-01
5.18851042e-01 -1.14815943e-01 -1.00530255e+00 3.04927111e-01
6.56564295e-01 5.25738895e-01 -3.53888333e-01 9.62675095e-01
-9.96801615e-01 1.00357902e+00 -7.80873001e-01 -1.07858807e-01
2.56704450e-01 -4.34023827e-01 4.05679971e-01 1.26372600e+00
5.17117977e-01 3.88405062e-02 -9.94053576e-03 3.77152681e-01
-2.15343654e-01 2.58404911e-01 -3.86634886e-01 -3.20446253e-01
1.40817866e-01 6.90754771e-01 -4.01125103e-01 -4.23362225e-01
-3.14075112e-01 1.60728884e+00 1.74534813e-01 8.62971485e-01
-8.01153958e-01 6.44099414e-02 9.19893205e-01 -4.38500196e-01
2.57940233e-01 -4.43265885e-01 -3.13458927e-02 -1.72305357e+00
1.89713672e-01 -1.47289789e+00 3.15203100e-01 -8.26248467e-01
-1.01889288e+00 4.64746416e-01 -1.07933789e-01 -1.37457645e+00
-5.03259122e-01 -5.60133636e-01 -1.79952934e-01 6.53709531e-01
-1.77832711e+00 -1.82996535e+00 -2.94245958e-01 6.73097372e-01
5.18342197e-01 -1.11828290e-01 9.39090192e-01 3.83336425e-01
-1.84790432e-01 9.70726728e-01 -1.70418173e-02 3.76959257e-02
1.37801361e+00 -8.37199867e-01 6.72355294e-01 8.02276492e-01
4.78538364e-01 4.43571419e-01 7.17395365e-01 -8.07131469e-01
-1.47187150e+00 -9.23649967e-01 1.50678003e+00 -6.21897817e-01
8.18243265e-01 -2.32027441e-01 -2.85755992e-01 1.02648747e+00
3.55355710e-01 -4.33606744e-01 6.04376733e-01 -3.94647196e-02
-5.42182922e-01 2.02427328e-01 -8.22349250e-01 1.12650061e+00
1.48459566e+00 -7.88973987e-01 -6.50601685e-01 4.97958899e-01
4.18788403e-01 -6.13125265e-01 -8.67532432e-01 3.54681373e-01
1.28968740e+00 -2.73534536e-01 7.14792192e-01 -7.79338717e-01
2.96226412e-01 -1.38376877e-01 -2.62510300e-01 -1.57315004e+00
2.89487660e-01 -1.27676392e+00 3.58761773e-02 8.12730372e-01
5.78990102e-01 -5.61756492e-01 6.36783063e-01 4.10608411e-01
-6.76851124e-02 -4.15122360e-01 -1.07302439e+00 -7.39526212e-01
8.41558818e-03 -2.18405351e-01 3.61807227e-01 7.43165195e-01
-1.61911950e-01 4.24125403e-01 -7.57758975e-01 -6.34170324e-02
6.17578566e-01 3.28322768e-01 1.08870387e+00 -6.47576392e-01
-3.69429141e-01 -4.89302814e-01 -2.29251578e-01 -1.12638307e+00
8.20811912e-02 -8.23904455e-01 4.11959320e-01 -1.65328610e+00
-7.12622330e-02 8.61322694e-03 2.14255646e-01 8.46853495e-01
-1.29959837e-01 5.71891129e-01 2.18793109e-01 5.97867131e-01
-3.26189935e-01 7.11529315e-01 1.32025826e+00 -2.01598734e-01
-2.72291452e-01 -5.39407916e-02 5.52299945e-03 5.46285212e-01
6.98706090e-01 -4.34339315e-01 -1.15787171e-01 -1.22773564e+00
1.96922228e-01 -2.59157777e-01 4.64611826e-03 -5.27632296e-01
1.90781564e-01 -2.12658331e-01 -2.70455003e-01 -5.12784660e-01
5.04489422e-01 -5.90386152e-01 -1.01484194e-01 4.56548840e-01
-3.89936447e-01 4.16668087e-01 1.80882066e-02 3.44095796e-01
-1.33921087e-01 3.21098894e-01 4.91405845e-01 -3.85916443e-03
-4.79296267e-01 7.19833374e-02 -5.56821942e-01 5.17189465e-02
3.30120713e-01 -2.55747866e-02 -4.06475782e-01 -5.31787336e-01
-4.90812093e-01 1.72298968e-01 4.78910953e-01 6.86891496e-01
5.99723279e-01 -1.40128982e+00 -9.96304095e-01 1.31301865e-01
2.67328531e-01 -5.66779137e-01 -5.08121789e-01 1.10074401e+00
-7.38717735e-01 4.15476352e-01 -4.44984078e-01 -4.79961812e-01
-1.64795327e+00 -2.53407031e-01 3.30861539e-01 -2.82315195e-01
-6.09250724e-01 8.58752787e-01 -6.53379202e-01 -7.48558700e-01
3.09702188e-01 -7.54294634e-01 3.95420581e-01 -1.49543300e-01
2.05995142e-01 8.21154267e-02 1.44975394e-01 -7.41813719e-01
-2.89202869e-01 9.22586441e-01 1.75231799e-01 -8.72806668e-01
1.18777180e+00 -9.21702385e-02 -2.24809740e-02 2.24466190e-01
8.91121149e-01 -1.23041980e-01 -9.99769270e-01 -4.66281712e-01
1.02456301e-01 -3.57995421e-01 -1.26521021e-01 -1.23992813e+00
-7.96953678e-01 5.38850844e-01 6.27577901e-01 -1.07486856e+00
1.10493755e+00 -2.35780641e-01 1.22327530e+00 9.18111324e-01
5.88935316e-01 -1.42745101e+00 1.01691976e-01 6.95621610e-01
1.02710748e+00 -1.32305658e+00 7.30116889e-02 -5.22355698e-02
-7.83165872e-01 1.12249768e+00 2.19687819e-01 6.56692162e-02
2.88388908e-01 2.09979832e-01 9.32709992e-01 -7.68195093e-02
-7.19078898e-01 -5.03364623e-01 5.10299206e-01 5.08734047e-01
3.08601946e-01 3.93359475e-02 -7.88757324e-01 -6.35351092e-02
-5.79517186e-01 4.30549949e-01 1.28508553e-01 1.20104706e+00
-4.00882727e-03 -1.37238908e+00 -2.02979684e-01 7.90801570e-02
-1.94539919e-01 -3.83163661e-01 -1.14021504e+00 7.93552279e-01
-8.02230984e-02 7.55453348e-01 -5.12894511e-01 -5.08378327e-01
4.60192919e-01 5.96031904e-01 1.06787157e+00 -2.63721973e-01
-7.05200195e-01 1.95283428e-01 4.64391947e-01 -7.12781847e-01
-7.75015354e-01 -9.16717887e-01 -1.00672781e+00 -6.89371601e-02
-2.08002582e-01 -3.61944795e-01 1.02821004e+00 1.25611591e+00
3.26108515e-01 -7.93544278e-02 -5.27663976e-02 -7.91936934e-01
-7.76912749e-01 -1.05184698e+00 1.57046780e-01 4.55728114e-01
2.13962808e-01 -3.99731755e-01 -2.47285098e-01 3.24092299e-01]
|
[9.214421272277832, -6.5424113273620605]
|
d7078d42-268d-451e-832a-9371b0c6b06a
|
how-to-choose-good-samples-for-text-data
|
2302.00894
| null |
https://arxiv.org/abs/2302.00894v1
|
https://arxiv.org/pdf/2302.00894v1.pdf
|
How to choose "Good" Samples for Text Data Augmentation
|
Deep learning-based text classification models need abundant labeled data to obtain competitive performance. Unfortunately, annotating large-size corpus is time-consuming and laborious. To tackle this, multiple researches try to use data augmentation to expand the corpus size. However, data augmentation may potentially produce some noisy augmented samples. There are currently no works exploring sample selection for augmented samples in nature language processing field. In this paper, we propose a novel self-training selection framework with two selectors to select the high-quality samples from data augmentation. Specifically, we firstly use an entropy-based strategy and the model prediction to select augmented samples. Considering some samples with high quality at the above step may be wrongly filtered, we propose to recall them from two perspectives of word overlap and semantic similarity. Experimental results show the effectiveness and simplicity of our framework.
|
['Shengyi Jiang', 'Ziyu Yang', 'Yingwen Fu', 'Nankai Lin', 'Xiaotian Lin']
|
2023-02-02
| null | null | null | null |
['semantic-textual-similarity']
|
['natural-language-processing']
|
[ 3.28680366e-01 6.32866612e-03 -2.45258778e-01 -5.95073402e-01
-5.93066096e-01 -3.76884155e-02 4.40455228e-01 3.62875581e-01
-8.25490713e-01 8.36094856e-01 2.51220107e-01 -6.90932199e-02
-7.61399046e-02 -8.93748999e-01 -1.57677561e-01 -6.76437676e-01
3.73107314e-01 4.48667884e-01 -1.68811366e-01 -1.90300643e-01
4.22658920e-01 3.05553358e-02 -1.55117345e+00 4.55823988e-01
1.43145502e+00 8.08630526e-01 2.95325637e-01 1.46914527e-01
-8.35782945e-01 3.30225706e-01 -6.53872132e-01 -4.91208375e-01
1.62164927e-01 -5.31318903e-01 -7.92654634e-01 2.21812531e-01
-2.23188270e-02 -2.32603446e-01 -4.52724136e-02 1.13832843e+00
6.40035748e-01 3.72699827e-01 4.64721054e-01 -1.06262410e+00
-5.06538510e-01 1.22487307e+00 -4.68132019e-01 1.23466037e-01
4.13572975e-02 -7.09578246e-02 9.75768864e-01 -1.19683909e+00
3.55028540e-01 1.16517305e+00 5.27514100e-01 6.63340151e-01
-8.18348825e-01 -7.56769001e-01 4.23488766e-01 2.03988776e-01
-1.30562520e+00 -4.06325877e-01 1.05477405e+00 -9.21345353e-02
5.42453110e-01 3.38846415e-01 6.26256824e-01 1.04845476e+00
-3.36816758e-01 1.16252995e+00 1.10443389e+00 -6.93332136e-01
1.78687260e-01 5.49815178e-01 4.86671209e-01 3.02205205e-01
4.45457220e-01 -3.63377690e-01 -5.01901031e-01 -1.60023570e-01
1.87330827e-01 4.53990363e-02 -2.00231150e-01 2.90105212e-02
-1.10358620e+00 8.99942815e-01 6.27640411e-02 6.39576197e-01
-3.31439048e-01 -4.19513762e-01 6.87933981e-01 2.44497061e-01
8.37333798e-01 5.73631465e-01 -8.11295748e-01 -6.77004009e-02
-9.40278172e-01 1.81090876e-01 4.78444606e-01 8.47833574e-01
7.20174193e-01 -2.09739469e-02 -3.54210705e-01 1.10138357e+00
2.71921903e-01 6.71778992e-02 1.07642305e+00 -1.68004572e-01
7.02586114e-01 1.01459813e+00 -8.46305788e-02 -9.95539546e-01
-4.99067217e-01 -6.57060266e-01 -1.00982785e+00 -2.90260375e-01
2.39108995e-01 -3.10011357e-01 -7.33238637e-01 1.41456902e+00
4.04754400e-01 -6.48786351e-02 1.09656252e-01 9.18428481e-01
9.07302380e-01 6.89656198e-01 3.48537773e-01 -6.00128055e-01
1.40321589e+00 -9.26532626e-01 -1.06733835e+00 -2.30713561e-01
1.09289789e+00 -7.91993976e-01 1.47822773e+00 5.18099129e-01
-7.00186014e-01 -5.18079162e-01 -9.86072004e-01 3.48687991e-02
-4.45439458e-01 6.33679032e-01 6.87175155e-01 7.13210404e-01
-3.71018380e-01 5.78567624e-01 -4.74696428e-01 -1.45801410e-01
5.41908503e-01 2.98351526e-01 -1.49444178e-01 -3.12286578e-02
-1.47104871e+00 5.60863316e-01 1.04538274e+00 9.84243825e-02
-1.67477071e-01 -3.95496339e-01 -5.67362309e-01 7.60365725e-02
6.01183832e-01 -3.20649564e-01 1.10914886e+00 -1.12722206e+00
-1.33690536e+00 4.24647897e-01 -8.31893608e-02 -4.04320180e-01
4.66018766e-01 -3.23034346e-01 -5.21729887e-01 -2.68190205e-01
-1.62298828e-01 3.61279607e-01 4.81246591e-01 -1.31798482e+00
-7.44112611e-01 -3.40675324e-01 -3.89857441e-01 5.74572146e-01
-1.27912879e+00 -1.09546095e-01 -2.36954600e-01 -9.22865391e-01
3.87356043e-01 -4.93284225e-01 -5.65089941e-01 -3.75040740e-01
-5.22225022e-01 -5.48320770e-01 5.82714379e-01 -5.97096086e-01
1.59387457e+00 -1.78559220e+00 -3.41979980e-01 2.48741895e-01
1.45227686e-01 6.44457281e-01 -1.43945366e-01 8.00492018e-02
1.91176757e-01 5.22396684e-01 -2.68546820e-01 -4.00006086e-01
-1.02295779e-01 5.40051749e-03 -1.73192427e-01 2.01789383e-03
2.96042025e-01 5.99762976e-01 -9.68234360e-01 -9.03159618e-01
1.15114719e-01 2.36546621e-01 -4.91805702e-01 2.59892315e-01
-2.38186404e-01 2.44105577e-01 -8.66643965e-01 5.37961543e-01
7.62496471e-01 1.36782825e-01 1.51786968e-01 -2.84775913e-01
9.98981521e-02 4.88455594e-01 -1.35676873e+00 1.47043145e+00
-5.12224078e-01 2.06598207e-01 -1.43069476e-01 -1.18385649e+00
1.24145651e+00 1.79037631e-01 4.55355942e-01 -4.36999261e-01
4.84463662e-01 4.13306415e-01 4.75233793e-02 -7.70023346e-01
1.04440081e+00 -2.92462468e-01 2.58680254e-01 4.46100622e-01
1.43836206e-02 1.85402006e-01 2.88021594e-01 -2.96582133e-02
6.85407281e-01 6.57711700e-02 2.42396608e-01 -5.84692024e-02
6.81522250e-01 6.64254427e-02 6.40713871e-01 5.21678925e-01
-1.49485081e-01 3.69234890e-01 2.14671969e-01 -3.55794638e-01
-1.11521852e+00 -1.83142677e-01 -6.30001649e-02 1.23456800e+00
-6.48307204e-02 -4.30473000e-01 -6.05082989e-01 -1.02725601e+00
-3.37606132e-01 8.81320953e-01 -3.24274898e-01 -3.95004094e-01
-6.03079855e-01 -1.21577859e+00 4.47615772e-01 3.62126231e-01
4.25372869e-01 -1.24728179e+00 -1.31008357e-01 3.70488077e-01
-3.78005654e-01 -7.80197442e-01 -3.48003805e-01 3.09312552e-01
-1.04774439e+00 -8.61006618e-01 -5.82340956e-01 -8.48371148e-01
8.00376058e-01 2.42755055e-01 9.25065815e-01 4.46306080e-01
1.76631927e-01 -5.65468252e-01 -1.02538788e+00 -5.72963834e-01
-5.23320198e-01 5.75846374e-01 2.17786893e-01 -2.30105221e-02
8.75292897e-01 -2.44185492e-01 -2.58727968e-01 3.81046049e-02
-9.71122265e-01 1.10189743e-01 7.70636499e-01 1.07517874e+00
4.80015963e-01 2.92124629e-01 1.11028636e+00 -9.92695928e-01
1.16401446e+00 -4.83882666e-01 -1.80390060e-01 2.66675591e-01
-8.53308499e-01 9.57635716e-02 1.03678155e+00 -7.56111562e-01
-1.23312628e+00 1.37869015e-01 -3.16798359e-01 2.47830208e-02
-2.28319064e-01 7.13079512e-01 -4.87802178e-01 3.91485095e-01
6.25245810e-01 4.35036033e-01 -9.22077596e-02 -7.75081277e-01
3.13478380e-01 1.00556147e+00 -9.81846452e-02 -4.84830588e-01
5.57719529e-01 3.08331940e-02 -4.98984754e-01 -4.68313634e-01
-9.71335232e-01 -3.91202182e-01 -6.55178368e-01 -2.11275481e-02
3.74743909e-01 -6.54564679e-01 -4.47957724e-01 2.00335220e-01
-1.04969168e+00 1.99423030e-01 -3.11526388e-01 7.98339188e-01
4.73980382e-02 5.58643699e-01 -2.74298221e-01 -1.12342048e+00
-7.75108278e-01 -9.47472095e-01 8.38236690e-01 2.98873305e-01
-4.12806123e-02 -7.08685875e-01 -2.18118951e-01 2.59453863e-01
3.06020170e-01 -2.06115410e-01 7.34764755e-01 -1.43118954e+00
1.42097883e-02 -4.03096497e-01 1.00839399e-02 4.02171254e-01
1.87371284e-01 -1.08584106e-01 -9.60438728e-01 -4.55165729e-02
2.47802705e-01 -3.62563938e-01 9.39405918e-01 1.80616707e-01
1.56775141e+00 -4.94562626e-01 -3.22821498e-01 2.36053035e-01
1.09181368e+00 3.20386857e-01 3.85913640e-01 4.76415962e-01
7.86783516e-01 8.17082882e-01 1.08640289e+00 6.83058500e-01
7.15976134e-02 4.20193017e-01 7.08931386e-02 1.49018720e-01
2.22509831e-01 -1.87138692e-01 8.06977451e-02 1.21307933e+00
2.22801983e-01 -5.30294716e-01 -8.60339582e-01 4.45544183e-01
-1.67514598e+00 -7.99670160e-01 -3.08339328e-01 2.06025267e+00
1.28130054e+00 4.64962333e-01 -5.05548753e-02 7.14235365e-01
8.26363444e-01 -1.06580026e-01 -4.15588021e-01 -1.32248983e-01
-2.55703151e-01 1.13379188e-01 1.96086228e-01 1.80080816e-01
-1.14546072e+00 9.92112100e-01 5.09694386e+00 1.31065524e+00
-8.77866924e-01 1.02029368e-01 9.71611083e-01 -1.16320536e-01
-6.52916193e-01 -1.28031194e-01 -1.05326939e+00 5.56079865e-01
6.34619534e-01 -2.15587825e-01 1.30760148e-02 9.27214146e-01
2.11051986e-01 1.51793763e-01 -6.42560184e-01 8.39816093e-01
1.61717564e-01 -9.51106608e-01 3.46458107e-01 -1.52377307e-01
4.72258627e-01 -4.06761408e-01 -2.83902913e-01 6.48497283e-01
4.03239131e-02 -8.17026377e-01 3.83925766e-01 3.92400533e-01
3.54380459e-01 -1.01589334e+00 1.08861840e+00 7.12113678e-01
-9.20783937e-01 -6.95868805e-02 -6.63046300e-01 -3.85365896e-02
5.05731329e-02 1.05928612e+00 -8.86380196e-01 6.17382586e-01
2.66982645e-01 4.40955073e-01 -6.08548701e-01 1.00392342e+00
1.00694530e-01 7.24028528e-01 -3.08095306e-01 -6.26066864e-01
1.25128344e-01 -3.62196892e-01 4.10539895e-01 1.05113244e+00
4.22669232e-01 1.84338585e-01 4.43396568e-01 5.90345204e-01
-2.43419960e-01 9.02835488e-01 -4.62634027e-01 -1.05530567e-01
6.78592861e-01 1.38950765e+00 -7.05934405e-01 -6.77763999e-01
-1.48154378e-01 7.72410452e-01 2.01700971e-01 -4.43941355e-02
-5.53855717e-01 -5.59872329e-01 1.46673068e-01 -1.42462999e-01
-2.04217881e-01 -5.21300845e-02 -9.06939626e-01 -1.32068777e+00
1.04111508e-01 -9.15419161e-01 3.23968977e-01 -3.50924522e-01
-1.46981001e+00 9.07155216e-01 -2.55605757e-01 -1.46031094e+00
1.69902116e-01 -3.95880491e-01 -2.90966690e-01 7.21203864e-01
-1.24317110e+00 -8.97138059e-01 -3.62001181e-01 1.68575779e-01
9.61630404e-01 -4.66270775e-01 7.03096509e-01 5.17776072e-01
-6.86071455e-01 7.83739805e-01 1.89834729e-01 1.05969444e-01
6.20056808e-01 -1.09634936e+00 2.64843971e-01 7.97099173e-01
1.31738544e-01 6.01239145e-01 5.76955140e-01 -7.79957056e-01
-1.02243781e+00 -1.10316801e+00 9.47861910e-01 -5.67664914e-02
2.69118756e-01 -2.46649697e-01 -1.27516973e+00 2.38360077e-01
1.74991086e-01 -3.19622636e-01 8.64880741e-01 2.01302394e-01
7.27717727e-02 -1.12754859e-01 -1.09542966e+00 7.36748159e-01
8.56381774e-01 1.92376375e-02 -6.84064269e-01 4.20753390e-01
8.06376100e-01 -6.81894720e-02 -6.57346845e-01 5.30690789e-01
3.03764611e-01 -4.69521642e-01 5.37097514e-01 -6.38399601e-01
2.69223183e-01 -1.14846870e-01 1.06252141e-01 -1.50138783e+00
-7.51168607e-03 -2.01654062e-01 6.25426939e-04 1.73483670e+00
3.86202604e-01 -4.63884711e-01 9.61387515e-01 5.82785010e-01
-2.42568925e-01 -9.00572479e-01 -6.38868332e-01 -6.16897106e-01
4.73257676e-02 -4.99038965e-01 8.86383712e-01 1.28915238e+00
2.56422758e-01 5.35272300e-01 -4.99328971e-01 -3.08021545e-01
1.94263145e-01 -1.61614299e-01 5.73948681e-01 -1.45579696e+00
2.21894845e-01 -5.67721605e-01 2.18350247e-01 -9.28899467e-01
1.16935529e-01 -8.72172594e-01 1.98237330e-01 -1.44530261e+00
1.83472589e-01 -7.96927750e-01 -4.02637899e-01 5.66254318e-01
-5.97906351e-01 -5.75626194e-02 -3.88249755e-02 2.39964947e-01
-5.03232718e-01 1.05503488e+00 1.24152315e+00 -4.79421057e-02
-5.87038398e-01 7.01289624e-02 -7.92851269e-01 6.97441041e-01
1.25115895e+00 -5.82781196e-01 -4.61449176e-01 -3.04492533e-01
2.90828645e-01 -4.21378642e-01 -3.34047616e-01 -8.21559548e-01
-2.88148653e-02 -2.74861842e-01 3.43214601e-01 -6.65816426e-01
1.09998807e-01 -8.98378432e-01 -4.07865226e-01 4.22066629e-01
-7.59291589e-01 -1.70568138e-01 4.47761377e-05 3.63626689e-01
-2.08222628e-01 -8.13675106e-01 5.98989606e-01 -2.00341493e-01
-4.18291301e-01 2.83225209e-01 -3.58420283e-01 -1.36512211e-02
6.52839065e-01 -8.29429328e-02 -5.00110984e-02 2.94347703e-02
-4.16786730e-01 3.04982394e-01 4.86877784e-02 5.93833268e-01
5.37147760e-01 -1.46110976e+00 -7.32975602e-01 2.54184395e-01
1.61077738e-01 1.73080444e-01 2.22277060e-01 6.32923722e-01
-1.92595392e-01 1.66276202e-01 1.53091982e-01 -1.21017762e-01
-1.23480749e+00 6.25642538e-01 4.03086804e-02 -4.88801599e-01
-3.42010736e-01 8.34467888e-01 -4.53228414e-01 -7.23142922e-01
2.59799361e-01 -3.09781194e-01 -9.01700735e-01 4.66688305e-01
6.52904749e-01 2.68486649e-01 2.58794963e-01 -4.47235286e-01
-1.07528009e-01 7.35668093e-02 -2.55817175e-01 7.95848295e-02
1.15422928e+00 -1.58760130e-01 -1.25014588e-01 3.58536124e-01
1.12424600e+00 -9.18642953e-02 -6.88106358e-01 -5.14639437e-01
3.82280916e-01 -5.15467525e-01 7.89282396e-02 -5.59901297e-01
-1.06074619e+00 8.59793246e-01 5.89924693e-01 4.47847724e-01
1.06770754e+00 -5.25498629e-01 9.58019972e-01 6.86228216e-01
-1.84038375e-02 -1.53357804e+00 -3.82388681e-02 6.26277864e-01
5.74277222e-01 -1.46146250e+00 6.34932965e-02 -4.10791516e-01
-7.10190475e-01 1.16701066e+00 1.06178367e+00 3.26900691e-01
6.56316936e-01 -1.18617527e-02 -8.71136039e-02 1.11250632e-01
-6.89242005e-01 -1.69333220e-01 4.01582628e-01 2.40279049e-01
8.55776191e-01 -5.45321852e-02 -1.28867793e+00 1.06356001e+00
-2.63807952e-01 -1.12351865e-01 4.64274675e-01 8.50171268e-01
-7.85876036e-01 -1.35664046e+00 -3.61031920e-01 1.02668822e+00
-5.18713534e-01 -3.64229351e-01 -2.93161958e-01 3.98699462e-01
3.15586507e-01 1.06553078e+00 -1.38972506e-01 -6.03289247e-01
1.53828546e-01 4.58516002e-01 -1.09200902e-01 -6.35110795e-01
-6.85544014e-01 1.43681616e-01 2.03125969e-01 1.11926131e-01
-5.38286984e-01 -4.90090281e-01 -1.21916556e+00 4.14343029e-02
-8.84702981e-01 4.62630510e-01 7.23008454e-01 1.08300185e+00
1.30267233e-01 5.81470966e-01 7.23164439e-01 -5.14626861e-01
-7.61488318e-01 -1.60254443e+00 -4.82653916e-01 5.05481958e-01
-1.01782650e-01 -4.56841826e-01 -2.76150644e-01 -1.58317074e-01]
|
[10.543789863586426, 7.52340030670166]
|
bbcca804-b77c-432e-a5e9-8a716e73b155
|
recipenlg-a-cooking-recipes-dataset-for-semi
| null | null |
https://aclanthology.org/2020.inlg-1.4
|
https://aclanthology.org/2020.inlg-1.4.pdf
|
RecipeNLG: A Cooking Recipes Dataset for Semi-Structured Text Generation
|
Semi-structured text generation is a non-trivial problem. Although last years have brought lots of improvements in natural language generation, thanks to the development of neural models trained on large scale datasets, these approaches still struggle with producing structured, context- and commonsense-aware texts. Moreover, it is not clear how to evaluate the quality of generated texts. To address these problems, we introduce RecipeNLG - a novel dataset of cooking recipes. We discuss the data collection process and the relation between the semi-structured texts and cooking recipes. We use the dataset to approach the problem of generating recipes. Finally, we make use of multiple metrics to evaluate the generated recipes.
|
['Agnieszka Ławrynowicz', 'Dawid Wiśniewski', 'Wojciech Taisner', 'Martyna Maciejewska', 'Michał Gilski', 'Michał Bień']
|
2020-12-15
| null | null | null | null |
['recipe-generation']
|
['miscellaneous']
|
[ 3.95403802e-01 2.38172710e-01 1.89685315e-01 -2.84723520e-01
-7.25675762e-01 -7.46077597e-01 7.13625371e-01 2.40592301e-01
-1.24457903e-01 1.03790498e+00 7.55116165e-01 1.78441405e-01
2.18154520e-01 -1.20749950e+00 -7.01733351e-01 -3.05259466e-01
4.50937241e-01 5.63098490e-01 -1.73156768e-01 -6.41903996e-01
3.58859777e-01 -3.04061711e-01 -1.59376848e+00 5.90623617e-01
1.39867377e+00 3.90910000e-01 4.78404641e-01 7.11837351e-01
-6.17766261e-01 1.10591877e+00 -8.32340837e-01 -6.53084338e-01
1.35540888e-02 -1.07256198e+00 -9.19928789e-01 -6.58140751e-03
2.12292269e-01 -2.56268889e-01 5.89108169e-02 1.05099881e+00
3.69878858e-01 2.50184238e-01 7.25088000e-01 -1.36614227e+00
-1.13719213e+00 1.63214314e+00 3.74119431e-01 -3.36628735e-01
6.75844610e-01 1.80046767e-01 1.06729865e+00 -6.71153784e-01
9.03525352e-01 1.10687959e+00 6.24269009e-01 9.01492417e-01
-1.20871127e+00 -3.35864514e-01 -3.29785198e-01 -1.41482845e-01
-8.81083846e-01 -4.03449327e-01 8.46619964e-01 -2.86974788e-01
9.49638307e-01 7.23026320e-02 7.22019017e-01 1.30758083e+00
-2.71537036e-01 1.10064077e+00 1.18500102e+00 -6.97987020e-01
2.18829960e-01 4.69142757e-02 3.37340832e-02 4.74858552e-01
3.66550416e-01 1.87034577e-01 -4.50714529e-01 3.25521380e-02
5.99140763e-01 -5.00980854e-01 -2.01818570e-01 -5.85598424e-02
-1.71380818e+00 1.11665022e+00 5.33887148e-02 5.43660223e-01
-5.30512989e-01 1.27814174e-01 4.94087547e-01 1.20983593e-01
2.68135577e-01 1.17595708e+00 -3.59070867e-01 -2.36526191e-01
-1.03502595e+00 8.69138777e-01 1.08204603e+00 1.23221266e+00
4.04283851e-01 2.57890671e-01 -4.55313087e-01 7.60963619e-01
8.60806257e-02 6.61209047e-01 6.07821107e-01 -9.43466842e-01
6.73707664e-01 5.92983723e-01 3.46354395e-01 -1.09117401e+00
-2.76143640e-01 3.74425918e-01 -9.42957938e-01 -1.89976886e-01
5.78413606e-01 -3.75250608e-01 -5.65031946e-01 1.90271914e+00
1.19893685e-01 -4.01945263e-01 3.15985471e-01 6.71266139e-01
1.04857981e+00 9.01366889e-01 2.98685700e-01 -9.96073559e-02
1.04615688e+00 -1.15653062e+00 -1.05813169e+00 -5.59160672e-02
4.92101997e-01 -7.89342701e-01 1.30268013e+00 3.79323602e-01
-1.26503050e+00 -6.75214231e-01 -1.00115180e+00 -1.88960060e-01
-6.93383515e-01 1.60488263e-01 6.71016574e-01 5.82041323e-01
-9.61312830e-01 9.13901269e-01 -1.88271165e-01 -3.15011829e-01
1.32303581e-01 -1.83570668e-01 -1.67644233e-01 -1.22780561e-01
-1.57993984e+00 9.71393585e-01 1.07906258e+00 -1.48957253e-01
-7.20037639e-01 -5.25433123e-01 -1.08485472e+00 -2.10586607e-01
3.38695109e-01 -8.51656914e-01 1.35252869e+00 -1.03864598e+00
-1.55596161e+00 6.85382903e-01 2.05418780e-01 -3.67699116e-01
3.08453500e-01 -1.19930074e-01 -4.76977497e-01 -7.17968047e-02
3.36045533e-01 8.91309857e-01 4.45093453e-01 -1.09866953e+00
-7.34733641e-01 1.07860625e-01 1.48276612e-01 2.37216160e-01
-1.39182940e-01 1.08077889e-02 2.13581532e-01 -9.34101343e-01
-4.05227065e-01 -7.31442988e-01 -4.27486032e-01 -5.06680369e-01
-6.81545556e-01 -4.94611353e-01 1.03642948e-01 -7.70254552e-01
1.20343840e+00 -1.83310783e+00 2.31369078e-01 -3.00256312e-01
-7.69873112e-02 2.29730338e-01 -4.62642640e-01 6.89100623e-01
1.45508334e-01 4.25690085e-01 -3.97485554e-01 -1.75292864e-01
4.90448117e-01 1.19803391e-01 -4.48806554e-01 -4.51472938e-01
4.72517163e-01 1.05676663e+00 -1.52535594e+00 -6.20426536e-01
2.57587910e-01 1.92814186e-01 -4.35917139e-01 4.91185904e-01
-9.28663194e-01 1.98719159e-01 -5.82220912e-01 2.49791011e-01
2.72951365e-01 2.56476123e-02 1.39926836e-01 -6.84933178e-03
-1.61293909e-01 7.36165762e-01 -1.07813692e+00 2.20765996e+00
-4.52924758e-01 4.15275216e-01 -4.79355395e-01 -7.54946172e-01
9.75155830e-01 5.57136118e-01 8.72652903e-02 -5.36869824e-01
2.98755970e-02 2.13263392e-01 -3.38888109e-01 -6.92109048e-01
1.22967327e+00 -4.73991990e-01 -5.27714789e-01 9.21644568e-01
6.39017150e-02 -7.88791716e-01 1.20405507e+00 3.09002936e-01
8.99226725e-01 6.65589452e-01 4.65732276e-01 -2.14685455e-01
3.03212434e-01 7.94478953e-01 1.99615344e-01 5.28966129e-01
9.13854688e-03 6.62703812e-01 2.57435173e-01 -4.41744745e-01
-1.30245054e+00 -1.02665925e+00 4.52943295e-01 1.02614212e+00
-2.07289726e-01 -6.36030793e-01 -1.11146057e+00 -7.39615142e-01
-2.81306267e-01 1.50270271e+00 -5.42652667e-01 -4.46132980e-02
-6.41436338e-01 -6.83726907e-01 8.48838270e-01 3.51670951e-01
4.36487079e-01 -1.89609671e+00 -6.88517153e-01 5.66446841e-01
-9.82972503e-01 -8.74066114e-01 -5.33268332e-01 -8.67915526e-02
-8.25586021e-01 -1.00610006e+00 -5.01911283e-01 -9.28665698e-01
4.83565301e-01 9.60017368e-02 2.02111602e+00 1.07845463e-01
-1.65398642e-01 1.25135168e-01 -8.56530964e-01 -4.93951648e-01
-1.48366225e+00 8.73812437e-02 -3.77125263e-01 -6.53897285e-01
4.30442303e-01 -2.56162673e-01 -9.88898650e-02 -1.72399402e-01
-1.26309013e+00 6.61453664e-01 2.98579365e-01 6.80585325e-01
3.67649764e-01 2.32730955e-01 7.82594442e-01 -8.43012094e-01
1.19086766e+00 -4.24074173e-01 -2.44300410e-01 5.50551593e-01
-3.61126453e-01 4.21421736e-01 1.08474445e+00 -3.47596377e-01
-1.20768106e+00 1.16770625e-01 -1.51220411e-01 3.72172594e-01
-5.47467470e-01 5.06891727e-01 -1.16732799e-01 9.35413957e-01
8.77204776e-01 3.37463111e-01 -3.22839797e-01 -2.26186097e-01
1.07078004e+00 4.51501459e-01 5.02103388e-01 -9.55567837e-01
6.63312078e-01 -5.58835417e-02 -4.21141028e-01 -6.92898333e-01
-1.05680430e+00 1.56928450e-01 -5.45117259e-01 -2.60733932e-01
1.12520111e+00 -6.53725326e-01 -2.81704939e-03 2.36138165e-01
-1.42710626e+00 -6.78605020e-01 -7.03108251e-01 8.25665519e-02
-8.33396971e-01 2.89465457e-01 -6.19533479e-01 -7.07108259e-01
-6.89746678e-01 -7.06660867e-01 7.98533201e-01 1.59520596e-01
-7.55186558e-01 -9.49264646e-01 4.58421856e-01 3.70164156e-01
5.07815063e-01 7.10112751e-01 1.17508650e+00 -4.75840181e-01
-4.07457292e-01 -1.01730257e-01 2.28113718e-02 3.16068202e-01
2.69463420e-01 6.92307502e-02 -5.92836499e-01 4.50374782e-01
-1.57763019e-01 -9.46188092e-01 4.33300078e-01 2.32111812e-01
6.87991798e-01 -4.26303059e-01 1.38360336e-01 5.16548231e-02
1.27123082e+00 7.84813017e-02 7.58763254e-01 1.27873704e-01
4.87213343e-01 8.89762521e-01 6.34086430e-01 3.47177804e-01
6.75231218e-01 2.96918094e-01 7.33886510e-02 1.57959655e-01
-4.79652941e-01 -9.70551550e-01 4.77402836e-01 1.10547638e+00
-6.15377761e-02 -5.11662126e-01 -5.74945331e-01 8.06379259e-01
-1.57316494e+00 -1.40268850e+00 -4.71106559e-01 1.76072061e+00
1.39312482e+00 -2.55202293e-01 1.19789675e-01 2.71418154e-01
5.83775818e-01 9.40068737e-02 -1.56769395e-01 -4.85024095e-01
-4.07040536e-01 2.35199630e-01 -1.42099977e-01 1.89605057e-01
-1.05340862e+00 1.14384520e+00 6.93153000e+00 7.21114099e-01
-5.87530255e-01 -8.51123482e-02 5.01280546e-01 1.45352840e-01
-6.07516468e-01 -1.49387792e-01 -5.52773237e-01 4.50627148e-01
9.65575516e-01 -3.92883271e-01 9.40862417e-01 8.06002736e-01
2.84157127e-01 -1.03347272e-01 -1.39266288e+00 6.09219909e-01
2.88633168e-01 -1.29690349e+00 3.05507094e-01 -4.75293815e-01
1.13795149e+00 -2.53266752e-01 -3.84990245e-01 6.58910751e-01
1.06102788e+00 -1.16399872e+00 1.13501000e+00 3.82681042e-01
5.86525142e-01 -6.42212033e-01 3.91266793e-01 5.34850776e-01
-1.00859272e+00 2.71503121e-01 -5.17799318e-01 -2.54657269e-01
4.04090077e-01 8.54127288e-01 -6.54398263e-01 5.24389982e-01
3.32015395e-01 5.52323699e-01 -4.97030526e-01 5.43657124e-01
-7.73902714e-01 4.88015324e-01 -2.65969825e-03 -5.62450171e-01
1.70422807e-01 -3.33253264e-01 2.87858695e-01 1.26635194e+00
5.17370820e-01 1.55283794e-01 2.16267943e-01 1.56103671e+00
-1.98757425e-01 3.83575857e-01 -8.52412403e-01 -7.66657412e-01
1.31215155e-01 1.38422477e+00 -6.83223009e-01 -5.87264657e-01
-1.18805327e-01 1.04567075e+00 4.24510002e-01 3.65388505e-02
-7.11291850e-01 -4.95115876e-01 2.72723079e-01 -1.70822337e-01
-5.14144562e-02 -1.53577819e-01 -3.77386868e-01 -1.18819165e+00
-1.55234143e-01 -1.25175118e+00 1.57119781e-01 -1.06224489e+00
-1.54068923e+00 6.56456828e-01 -5.85363656e-02 -8.78147900e-01
-6.28710747e-01 -2.74222225e-01 -5.78012824e-01 6.40904069e-01
-1.24088979e+00 -9.28557277e-01 -2.09816560e-01 2.17896298e-01
7.95575380e-01 1.61407664e-02 1.31346250e+00 -1.45983265e-03
-1.69955641e-01 1.13779530e-02 -1.15162030e-01 3.67998421e-01
6.57354295e-01 -1.55929327e+00 7.76185930e-01 8.29460144e-01
2.97977239e-01 5.84645689e-01 9.60488081e-01 -9.16190147e-01
-1.20456338e+00 -1.31359315e+00 1.47728539e+00 -6.65901542e-01
5.97221971e-01 -4.11925167e-01 -4.98642951e-01 2.64254272e-01
7.17327952e-01 -6.53007686e-01 7.85536170e-01 -1.88500956e-01
-2.17273548e-01 4.74471867e-01 -1.26591420e+00 8.49012077e-01
1.16657698e+00 -1.79668829e-01 -1.13044369e+00 4.55450326e-01
9.64542747e-01 -1.71119750e-01 -7.57835448e-01 -2.67919269e-03
1.78771555e-01 -9.16736305e-01 7.17369914e-01 -6.92549109e-01
1.37901580e+00 -3.50403666e-01 -1.56177193e-01 -1.92693603e+00
-2.34188497e-01 -5.67042887e-01 2.30879039e-01 1.44858134e+00
6.50170028e-01 1.33796051e-01 5.25052488e-01 6.58097744e-01
-1.67992026e-01 -8.95764604e-02 -2.30792046e-01 -5.96489668e-01
3.91664147e-01 -3.50426823e-01 9.31773901e-01 1.04722655e+00
4.47113454e-01 6.63285136e-01 -4.78672177e-01 -5.97355902e-01
5.89827180e-01 4.59752709e-01 7.80967534e-01 -1.01270080e+00
-2.33480707e-01 -5.79332769e-01 4.66246784e-01 -6.61100090e-01
9.16635692e-02 -1.18317854e+00 7.58516908e-01 -2.11762929e+00
2.58381188e-01 -2.76102811e-01 2.35016972e-01 4.23232794e-01
-5.15372276e-01 9.57153216e-02 4.05672312e-01 -3.21215034e-01
-7.02809274e-01 5.44689178e-01 1.38547063e+00 -1.53586745e-01
-2.34673008e-01 -4.80337948e-01 -7.98057735e-01 4.28377330e-01
1.15970862e+00 -5.38326621e-01 -4.75353539e-01 -4.79034305e-01
6.76792741e-01 -3.15737240e-02 1.21775590e-01 -9.17011321e-01
-2.46406451e-01 -4.71900016e-01 1.49020135e-01 -6.12754822e-01
-1.76579684e-01 -3.24923247e-01 2.19241485e-01 3.70553702e-01
-8.25237751e-01 1.01117656e-01 4.49616499e-02 1.05561890e-01
-1.40513003e-01 -4.99403358e-01 5.82348049e-01 -7.15129673e-01
-4.87548500e-01 -1.45755902e-01 -5.88436007e-01 6.51609600e-01
7.82445431e-01 2.30745748e-01 -3.61739576e-01 -2.63866246e-01
-2.78390557e-01 6.09222092e-02 5.54354310e-01 6.54717863e-01
4.80022937e-01 -1.60966718e+00 -1.07477260e+00 -1.07853115e-01
2.12302968e-01 2.40360480e-02 1.52254090e-01 1.54600248e-01
-6.94973171e-01 3.44532043e-01 -2.37666294e-01 8.23073238e-02
-7.23923922e-01 7.39844322e-01 2.42972635e-02 -6.45103335e-01
-4.31183368e-01 4.69821572e-01 -2.31451675e-01 -9.14581537e-01
-1.40717164e-01 -5.73683619e-01 -2.73436397e-01 -6.34595007e-02
7.86731422e-01 3.18436384e-01 -3.69254649e-01 -4.05096710e-01
3.43072861e-01 -4.08887230e-02 4.55698133e-01 -2.22699419e-01
1.34446025e+00 1.04033835e-01 -3.29550475e-01 4.87920851e-01
5.82706630e-01 -2.26926804e-01 -9.18369949e-01 5.59272915e-02
3.67722631e-01 -2.94593051e-02 -3.31527114e-01 -1.25805438e+00
-7.04274118e-01 7.75345623e-01 -9.48511884e-02 4.48737711e-01
7.83818483e-01 -3.31726849e-01 1.15499997e+00 5.29577434e-01
4.85637516e-01 -1.45715213e+00 1.40569076e-01 7.75350749e-01
1.04850399e+00 -1.20570362e+00 -3.34174901e-01 -1.49628609e-01
-9.19915259e-01 1.14907241e+00 4.70011443e-01 -6.41800910e-02
6.76248297e-02 2.51874864e-01 1.39576480e-01 -7.98560120e-03
-6.42825127e-01 -1.54119015e-01 1.46944806e-01 8.64247561e-01
9.09390330e-01 4.09800082e-01 -3.81557643e-01 7.87485659e-01
-7.54904568e-01 4.38519865e-01 7.46127725e-01 7.39762068e-01
-5.55679977e-01 -1.36826730e+00 -4.43638057e-01 3.63305956e-01
-2.30391607e-01 -3.78002346e-01 -6.47097945e-01 5.65761507e-01
2.94793159e-01 1.53304291e+00 -3.37343991e-01 -1.02648251e-01
3.12121004e-01 4.11703289e-01 6.17182434e-01 -8.15357685e-01
-8.22256267e-01 -3.73515844e-01 4.99056160e-01 -2.86704242e-01
-6.31482780e-01 -6.56156778e-01 -1.37328565e+00 -5.26972353e-01
-1.86207697e-01 3.57343733e-01 6.68727160e-01 8.84836733e-01
2.33502332e-02 5.07229686e-01 3.05860132e-01 -7.63984621e-01
-6.26168609e-01 -1.14423645e+00 -5.25344193e-01 8.66355717e-01
-3.47357184e-01 -5.62000424e-02 -1.86639115e-01 5.55660248e-01]
|
[11.502715110778809, 4.65281867980957]
|
b6b4c504-2c6d-4561-a413-149426ad8b39
|
data-augmentation-for-low-resource-quechua
|
2207.06872
| null |
https://arxiv.org/abs/2207.06872v1
|
https://arxiv.org/pdf/2207.06872v1.pdf
|
Data Augmentation for Low-Resource Quechua ASR Improvement
|
Automatic Speech Recognition (ASR) is a key element in new services that helps users to interact with an automated system. Deep learning methods have made it possible to deploy systems with word error rates below 5% for ASR of English. However, the use of these methods is only available for languages with hundreds or thousands of hours of audio and their corresponding transcriptions. For the so-called low-resource languages to speed up the availability of resources that can improve the performance of their ASR systems, methods of creating new resources on the basis of existing ones are being investigated. In this paper we describe our data augmentation approach to improve the results of ASR models for low-resource and agglutinative languages. We carry out experiments developing an ASR for Quechua using the wav2letter++ model. We reduced WER by 8.73% through our approach to the base model. The resulting ASR model obtained 22.75% WER and was trained with 99 hours of original resources and 99 hours of synthetic data obtained with a combination of text augmentation and synthetic speech generati
|
['Jordi Luque', 'Mireia Farrús', 'Guillermo Cámbara', 'Nuria Bel', 'Rodolfo Zevallos']
|
2022-07-14
| null | null | null | null |
['text-augmentation']
|
['natural-language-processing']
|
[ 1.55393213e-01 3.10390174e-01 3.56714219e-01 -4.51775014e-01
-1.07792866e+00 -2.80898958e-01 5.21644235e-01 -7.25784376e-02
-7.31261253e-01 7.67922997e-01 3.91152024e-01 -6.90622211e-01
4.50150728e-01 -5.39843261e-01 -4.52438533e-01 -2.61174947e-01
2.62533575e-01 8.72075558e-01 1.84919015e-01 -8.39796722e-01
-1.16386041e-01 6.41511142e-01 -1.48549056e+00 5.81146717e-01
8.11600268e-01 6.27265036e-01 3.65230709e-01 1.07113397e+00
-4.14517552e-01 6.77321434e-01 -1.27535939e+00 -1.43640369e-01
2.04650864e-01 -3.50397497e-01 -8.06432247e-01 1.27842486e-01
5.43757565e-02 -1.54379353e-01 -6.08240306e-01 6.75668597e-01
6.99534774e-01 2.78750569e-01 1.98473185e-01 -7.51545906e-01
-6.54567778e-01 1.05529809e+00 -1.01566955e-01 4.08923745e-01
5.16937971e-01 -3.78955975e-02 5.96777678e-01 -1.02965629e+00
3.02136302e-01 1.24474788e+00 2.87658483e-01 8.76565993e-01
-9.80278075e-01 -5.69154799e-01 -2.02009246e-01 -1.27970288e-02
-1.60410905e+00 -1.15877175e+00 5.43776989e-01 2.06644982e-01
1.74912417e+00 4.48952079e-01 4.38153565e-01 1.08156431e+00
-5.85061491e-01 7.48566508e-01 1.00887752e+00 -8.63398194e-01
5.10359332e-02 3.28787684e-01 -7.51500111e-03 4.51400369e-01
9.86215845e-03 -2.34402955e-01 -7.94380903e-01 1.12914637e-01
4.49105978e-01 -5.02541065e-01 -7.71636516e-02 7.06474483e-01
-1.05589688e+00 8.03996682e-01 4.30044010e-02 6.34554386e-01
-2.98429936e-01 -2.47430250e-01 3.47447842e-01 5.40416121e-01
6.48735404e-01 6.59373343e-01 -6.97854877e-01 -5.50804377e-01
-9.00203347e-01 -1.08931409e-02 8.86361063e-01 1.11800754e+00
4.72129047e-01 8.54236543e-01 1.99614745e-02 1.39925146e+00
1.47451267e-01 8.56123924e-01 9.36045349e-01 -6.16807163e-01
6.76222920e-01 5.68396807e-01 1.98890292e-03 -1.82613373e-01
-1.70490623e-01 -3.00637931e-01 -5.13574719e-01 -2.30630115e-01
2.56081432e-01 -2.09068850e-01 -1.37716317e+00 1.28846633e+00
-8.99951160e-02 -2.05346674e-01 4.52728212e-01 4.66044277e-01
8.55645239e-01 1.27787054e+00 -1.65516078e-01 -2.15835989e-01
1.13468206e+00 -9.47562337e-01 -1.03883910e+00 -4.02511597e-01
9.26644087e-01 -9.82879937e-01 1.49236333e+00 3.78519654e-01
-1.15595293e+00 -5.49817026e-01 -9.83455479e-01 1.75119247e-02
-6.10554218e-01 2.19048694e-01 1.81930304e-01 1.13076460e+00
-1.37730348e+00 1.33989796e-01 -7.79062569e-01 -4.09340948e-01
5.54140173e-02 4.15140122e-01 -3.81867468e-01 -1.05465822e-01
-1.36513662e+00 1.09444344e+00 4.52132553e-01 1.79017901e-01
-6.83742464e-01 -3.41749489e-01 -9.33446169e-01 -6.71091005e-02
2.80041635e-01 1.46116346e-01 1.48184228e+00 -1.05561376e+00
-1.91926181e+00 6.70392931e-01 -2.03017190e-01 -6.77902699e-01
2.03614146e-01 -3.11499715e-01 -9.07154500e-01 -2.59962350e-01
-4.14691836e-01 4.27885860e-01 5.71107447e-01 -9.04112220e-01
-6.02009714e-01 -2.66653270e-01 -3.22841644e-01 2.28953362e-01
-4.29436773e-01 5.55462062e-01 -2.97380954e-01 -6.73129976e-01
-6.34124950e-02 -8.87462258e-01 -5.29915728e-02 -9.69830096e-01
-1.94355428e-01 -1.52367681e-01 8.26686800e-01 -1.37744045e+00
1.33872187e+00 -2.03999734e+00 -2.63582885e-01 2.17621937e-01
-3.94334793e-01 1.00778937e+00 -3.13400894e-01 4.12145048e-01
1.26915034e-02 3.38979036e-01 -1.42861798e-01 -4.21509355e-01
-2.54139632e-01 4.26142752e-01 -3.39559585e-01 -3.75544317e-02
3.94112945e-01 5.66515505e-01 -5.63939869e-01 -6.62357807e-02
3.20448756e-01 6.93472803e-01 -3.83454233e-01 5.95644176e-01
-7.60711450e-03 2.35131159e-01 3.17411572e-02 5.95693707e-01
1.84665874e-01 4.24719155e-01 -1.65280178e-02 2.94159144e-01
-2.91617632e-01 9.57593203e-01 -1.04730678e+00 1.39924479e+00
-8.90841246e-01 7.45702982e-01 -4.96112369e-03 -7.96339154e-01
1.35695052e+00 6.48138523e-01 1.06457166e-01 -7.57754862e-01
1.01059005e-01 6.21739209e-01 4.16917741e-01 -2.56214440e-01
8.94544721e-01 -3.33397393e-03 6.69535398e-02 2.44101912e-01
2.20044136e-01 -3.20522726e-01 1.83727443e-01 9.22461748e-02
1.23775089e+00 -2.07492545e-01 2.83548743e-01 -1.12721277e-02
6.61682248e-01 -7.33368695e-02 1.69028997e-01 7.39176929e-01
-7.82195851e-02 6.44227624e-01 -8.42329636e-02 -3.42087626e-01
-1.49473381e+00 -6.65994942e-01 -1.72071867e-02 1.10235035e+00
-8.05522859e-01 -2.58971304e-01 -9.17110026e-01 -2.82436013e-01
-5.06367087e-01 1.23336279e+00 -8.19442347e-02 -3.67900692e-02
-9.86935973e-01 -6.68842673e-01 1.03265989e+00 4.85324413e-01
4.68977571e-01 -1.50453603e+00 7.58627951e-02 4.48819607e-01
-1.69182271e-01 -1.40849805e+00 -2.90118814e-01 2.57311791e-01
-6.06900871e-01 -2.98555285e-01 -8.17784965e-01 -7.46554494e-01
3.70812654e-01 4.73930277e-02 9.94117320e-01 1.30239233e-01
-3.07919718e-02 1.12325810e-01 -8.73309731e-01 -6.52256131e-01
-1.23865521e+00 5.17337084e-01 2.40198240e-01 -1.50363490e-01
4.76819694e-01 -2.00760543e-01 1.86153203e-01 1.94564596e-01
-8.97016048e-01 -2.84811798e-02 5.41350603e-01 7.03970671e-01
2.54288435e-01 -3.54554832e-01 1.05404508e+00 -9.01870430e-01
6.50339246e-01 -1.51408449e-01 -5.71282446e-01 2.39053100e-01
-3.95847321e-01 1.61626309e-01 6.49355054e-01 -4.81246024e-01
-1.21382475e+00 8.16511437e-02 -9.69223022e-01 -7.40494877e-02
-3.03473204e-01 4.08286989e-01 -4.47321653e-01 1.90109983e-01
7.49806762e-01 2.85962462e-01 -1.00723483e-01 -5.52867353e-01
3.43038797e-01 1.61613429e+00 2.63425469e-01 -3.89123857e-01
5.11907399e-01 -2.06046566e-01 -7.89967000e-01 -1.66825390e+00
-6.41947806e-01 -5.48773587e-01 -3.92838180e-01 -1.06561467e-01
5.27210593e-01 -8.59362602e-01 -9.52783078e-02 6.12675190e-01
-1.28129303e+00 -5.49275696e-01 -4.51870322e-01 4.63582695e-01
-3.48156661e-01 9.37396139e-02 -6.27349138e-01 -1.12165463e+00
-6.02271259e-01 -1.16248643e+00 9.07717407e-01 8.50991532e-03
-2.72856861e-01 -6.63301408e-01 7.21982941e-02 5.46297431e-01
7.77127147e-01 -6.72371328e-01 6.30136728e-01 -1.33906972e+00
-4.90058847e-02 -4.22532529e-01 1.33501783e-01 8.86439085e-01
1.93634257e-01 -3.54436450e-02 -1.29775608e+00 -1.95164129e-01
-7.84274042e-02 -2.47948557e-01 3.64943981e-01 -1.97359286e-02
1.10319352e+00 -3.96854609e-01 1.97938845e-01 2.82146156e-01
8.74635041e-01 6.45916283e-01 1.02353668e+00 2.63645023e-01
5.98399639e-01 5.33001184e-01 4.05896693e-01 1.45659402e-01
7.26027740e-03 8.34438741e-01 -8.23697522e-02 -2.01718546e-02
-3.63442034e-01 -1.63316131e-01 6.78742290e-01 1.59745812e+00
-7.03498125e-02 -5.74781895e-01 -1.20546222e+00 6.92281544e-01
-1.29220784e+00 -6.57487094e-01 -5.06777987e-02 2.36398292e+00
1.06540120e+00 1.75015867e-01 8.51742774e-02 3.90940309e-01
7.77073085e-01 -3.60763855e-02 -7.40917847e-02 -8.73881936e-01
-3.57665807e-01 5.68188250e-01 4.95361477e-01 8.19486558e-01
-7.04854190e-01 1.40521359e+00 6.50610924e+00 7.88029253e-01
-9.44265962e-01 1.49252996e-01 5.65383196e-01 1.05927791e-02
-1.93344966e-01 -3.78951788e-01 -1.14114082e+00 2.52531260e-01
1.93002582e+00 7.54034966e-02 8.44592035e-01 8.19002628e-01
3.12787086e-01 2.20643088e-01 -7.42101789e-01 9.47726607e-01
2.92861074e-01 -1.09851396e+00 2.34550193e-01 -1.63380936e-01
5.32288551e-01 3.71329010e-01 -1.78069413e-01 6.59819007e-01
4.14234787e-01 -1.08764446e+00 5.61751783e-01 -2.42425199e-03
9.70210671e-01 -9.10484314e-01 1.10503757e+00 3.24286640e-01
-8.24216902e-01 1.75611079e-01 -3.58304381e-01 3.73865478e-02
1.43241256e-01 2.27375314e-01 -1.62502003e+00 1.75153449e-01
4.70611840e-01 -3.43509987e-02 -6.51949704e-01 6.92671061e-01
-1.11320324e-01 1.10272503e+00 -5.04902840e-01 -2.96690285e-01
6.17376231e-02 -1.42989848e-02 5.63424766e-01 1.53073430e+00
5.87575972e-01 2.26904005e-02 -9.00112689e-02 2.07786590e-01
-2.60238826e-01 4.68851537e-01 -6.64330781e-01 -3.65422130e-01
7.39104450e-01 9.15381253e-01 -4.71623093e-01 -5.59123576e-01
-4.71288323e-01 8.65833461e-01 3.79621923e-01 1.82354316e-01
-4.83719766e-01 -6.02827787e-01 4.88817304e-01 3.32730800e-01
-1.49029016e-01 -5.30296028e-01 -1.47438645e-01 -1.04409420e+00
-1.29542768e-01 -1.49409008e+00 1.23411492e-01 -8.64240706e-01
-9.60599422e-01 1.25409520e+00 -2.29792058e-01 -7.32221842e-01
-5.15336156e-01 -7.06838667e-01 -1.23318493e-01 1.22607589e+00
-1.26358032e+00 -1.07139444e+00 -2.05345377e-02 4.88939255e-01
1.13135993e+00 -8.78215194e-01 1.23563111e+00 4.88589972e-01
-4.46874380e-01 6.69709921e-01 8.30963999e-02 3.14932883e-01
6.12035275e-01 -1.18130004e+00 9.43020642e-01 1.11121857e+00
5.13427317e-01 4.61318284e-01 6.91325307e-01 -5.71101665e-01
-1.18521035e+00 -9.86567140e-01 1.09837365e+00 -4.21985000e-01
8.05728495e-01 -6.03232920e-01 -1.18072653e+00 7.07225323e-01
3.13917786e-01 -1.80971012e-01 7.15927005e-01 8.55982974e-02
-1.16874121e-01 -1.77953765e-01 -9.98266578e-01 6.63846314e-01
7.88965523e-01 -6.17199719e-01 -6.98621988e-01 1.68556541e-01
1.12588859e+00 -4.04558569e-01 -5.63342690e-01 9.49747264e-02
9.20879096e-02 -3.65694821e-01 5.48461378e-01 -6.99419677e-01
-1.16617203e-01 -2.23149389e-01 -4.51191932e-01 -1.63084137e+00
2.12859005e-01 -8.41978014e-01 4.55354787e-02 1.36510456e+00
9.29386199e-01 -7.01673329e-01 4.99990642e-01 4.98512030e-01
-5.81133187e-01 -8.07649791e-02 -9.12622750e-01 -8.33601117e-01
-1.99309692e-01 -7.25950480e-01 5.91041863e-01 7.18517661e-01
-1.24178901e-01 5.04872262e-01 -3.55070472e-01 1.07083842e-01
-2.82946061e-02 -7.95225441e-01 7.39116848e-01 -8.24986696e-01
-2.09945008e-01 -1.98315959e-02 -2.84196645e-01 -6.17213070e-01
2.11843133e-01 -8.06153893e-01 3.18195820e-01 -1.36051917e+00
-3.59330267e-01 -5.92361152e-01 -1.53880581e-01 7.30065227e-01
-1.25326673e-02 2.15294227e-01 2.13266775e-01 -1.60093814e-01
-1.08303905e-01 4.68043596e-01 6.24165833e-01 -1.37981633e-03
-5.86790025e-01 8.23612437e-02 -4.88916457e-01 5.47702074e-01
1.15061831e+00 -5.02995491e-01 -2.62890577e-01 -6.05138659e-01
2.45598238e-02 -1.51418466e-02 -4.51509923e-01 -9.93466020e-01
-4.74736132e-02 1.38706297e-01 8.25677663e-02 -4.56309974e-01
5.36150753e-01 -6.08317852e-01 -6.54362887e-02 2.18497589e-01
-4.78368670e-01 3.28013688e-01 4.15001750e-01 -8.68856087e-02
-4.20973539e-01 -3.98831964e-01 8.86161804e-01 1.01166768e-02
-5.70778787e-01 -1.20314099e-01 -8.60829413e-01 6.02298789e-02
4.42050576e-01 2.49156673e-02 -2.35407680e-01 -6.58009171e-01
-6.35953784e-01 -2.67777890e-01 1.22964770e-01 6.72973633e-01
7.83452809e-01 -1.04381454e+00 -1.01537728e+00 5.10262966e-01
1.73447713e-01 -1.61897451e-01 -8.96647945e-02 2.91719049e-01
-8.44117045e-01 5.31430960e-01 -8.78916308e-02 -2.37218663e-02
-1.40280592e+00 2.74865627e-01 3.42767894e-01 -1.13541603e-01
-4.04509395e-01 6.57597482e-01 -4.13701922e-01 -6.74197018e-01
3.15844178e-01 -1.35913551e-01 -4.08586115e-01 -3.40038717e-01
7.74672508e-01 4.16451722e-01 7.22485483e-01 -8.65596712e-01
-1.87599793e-01 -3.60120595e-01 -2.12518767e-01 -6.98848903e-01
1.42173445e+00 2.50521451e-02 5.13136983e-02 6.02777600e-01
9.44299579e-01 3.98210913e-01 -5.18459439e-01 -1.42925188e-01
8.18703249e-02 -1.86510578e-01 1.85297027e-01 -8.99869323e-01
-6.73327088e-01 9.73093092e-01 6.19054615e-01 3.87474746e-01
9.02994871e-01 -1.16574660e-01 7.94714630e-01 8.07421207e-01
2.55396873e-01 -1.38564897e+00 -1.69374049e-01 1.06247294e+00
9.40604866e-01 -1.23052955e+00 -4.98912156e-01 -2.20092744e-01
-6.83417439e-01 1.02468979e+00 4.44220930e-01 9.62709114e-02
4.11904842e-01 5.12732089e-01 3.70828539e-01 3.00090939e-01
-6.74905956e-01 -3.21808308e-01 1.46854103e-01 7.62817800e-01
8.21488798e-01 2.89528698e-01 -2.52390057e-01 4.74893570e-01
-5.61516643e-01 -4.01062727e-01 7.92060137e-01 7.70336866e-01
-6.85322821e-01 -1.31700921e+00 -4.71049219e-01 4.05388445e-01
-6.95051551e-01 -4.24907178e-01 -4.93339956e-01 7.87791610e-01
-4.53714967e-01 1.28447568e+00 1.39517739e-01 -4.30628479e-01
5.18144011e-01 4.39819396e-01 1.16243482e-01 -1.02425396e+00
-5.70703030e-01 2.53358066e-01 7.83352017e-01 -2.05091819e-01
-4.21919599e-02 -5.80310285e-01 -1.19606197e+00 -1.44733310e-01
-2.87364602e-01 2.87387490e-01 1.11026120e+00 9.46985066e-01
9.65088084e-02 6.06988251e-01 6.11205935e-01 -5.26858687e-01
-4.63503540e-01 -1.54341817e+00 -4.98332977e-01 8.89926553e-02
3.06179579e-02 -6.44461289e-02 -2.78330147e-01 9.72481444e-02]
|
[14.392953872680664, 6.855511665344238]
|
da912806-d326-4df4-a81d-9a6bca453699
|
improving-diffusion-models-for-scene-text
|
2304.05568
| null |
https://arxiv.org/abs/2304.05568v1
|
https://arxiv.org/pdf/2304.05568v1.pdf
|
Improving Diffusion Models for Scene Text Editing with Dual Encoders
|
Scene text editing is a challenging task that involves modifying or inserting specified texts in an image while maintaining its natural and realistic appearance. Most previous approaches to this task rely on style-transfer models that crop out text regions and feed them into image transfer models, such as GANs. However, these methods are limited in their ability to change text style and are unable to insert texts into images. Recent advances in diffusion models have shown promise in overcoming these limitations with text-conditional image editing. However, our empirical analysis reveals that state-of-the-art diffusion models struggle with rendering correct text and controlling text style. To address these problems, we propose DIFFSTE to improve pre-trained diffusion models with a dual encoder design, which includes a character encoder for better text legibility and an instruction encoder for better style control. An instruction tuning framework is introduced to train our model to learn the mapping from the text instruction to the corresponding image with either the specified style or the style of the surrounding texts in the background. Such a training method further brings our method the zero-shot generalization ability to the following three scenarios: generating text with unseen font variation, e.g., italic and bold, mixing different fonts to construct a new font, and using more relaxed forms of natural language as the instructions to guide the generation task. We evaluate our approach on five datasets and demonstrate its superior performance in terms of text correctness, image naturalness, and style controllability. Our code is publicly available. https://github.com/UCSB-NLP-Chang/DiffSTE
|
['Shiyu Chang', 'Brian Price', 'Zhifei Zhang', 'Bairu Hou', 'Zhaowen Wang', 'Guanhua Zhang', 'Jiabao Ji']
|
2023-04-12
| null | null | null | null |
['scene-text-editing']
|
['computer-vision']
|
[ 7.26583481e-01 3.15831900e-02 2.65203696e-02 -3.87742281e-01
-2.97263861e-01 -6.84433043e-01 8.44356537e-01 -4.88834381e-01
-2.77524710e-01 6.29495800e-01 1.86005250e-01 -3.13713729e-01
5.92915714e-01 -8.06777894e-01 -9.45124090e-01 -6.57831132e-01
7.72250712e-01 3.27150106e-01 3.63348454e-01 -3.17069024e-01
3.08708370e-01 2.79009551e-01 -1.24008858e+00 3.65153223e-01
1.17395413e+00 5.51569521e-01 5.44559300e-01 8.74823570e-01
-3.91541153e-01 8.54446352e-01 -7.84550369e-01 -4.88350272e-01
3.98363501e-01 -8.93428266e-01 -4.50989723e-01 3.71602833e-01
4.20202523e-01 -6.23278975e-01 -4.39671129e-01 1.03587174e+00
5.11897683e-01 7.97215998e-02 6.65473223e-01 -9.89235282e-01
-1.29993057e+00 3.63546103e-01 -7.80781031e-01 -1.64828792e-01
3.22446346e-01 4.84460026e-01 5.70517838e-01 -6.89226866e-01
7.78199553e-01 1.21416128e+00 2.74312943e-01 1.05851769e+00
-1.43361294e+00 -6.50370538e-01 2.19713449e-01 -1.59234151e-01
-1.01373732e+00 -5.37407398e-01 7.47017145e-01 -4.46294665e-01
6.69663489e-01 3.76473159e-01 5.38088858e-01 1.53599989e+00
3.57057810e-01 8.48242640e-01 1.28654373e+00 -6.47895873e-01
2.50987977e-01 3.26397181e-01 -6.33741498e-01 6.86795235e-01
-9.81071591e-02 3.37662026e-02 -4.83773381e-01 2.40100086e-01
1.10637569e+00 -8.20790753e-02 -5.46657383e-01 -3.89570832e-01
-1.27382505e+00 7.97999203e-01 2.74997652e-01 1.07792638e-01
-1.42881259e-01 1.86630771e-01 2.71584779e-01 3.14264268e-01
5.74889183e-01 5.20948172e-01 -3.06528002e-01 -5.43776527e-02
-9.48087156e-01 2.18050089e-02 6.18352532e-01 1.22717345e+00
6.48618102e-01 2.58809835e-01 -5.26398122e-01 9.48745608e-01
-6.92607835e-02 6.09230042e-01 5.38662136e-01 -8.24587047e-01
5.51213920e-01 2.84377009e-01 7.23348483e-02 -6.84736431e-01
1.87684819e-01 -4.64740209e-02 -8.02454591e-01 6.03184104e-01
2.55957872e-01 -3.54339898e-01 -1.28601325e+00 1.76499033e+00
2.05744877e-01 -5.36031686e-02 -1.07211150e-01 7.21148729e-01
4.16073322e-01 8.97846758e-01 -1.15652092e-01 6.50098398e-02
1.08208907e+00 -1.28224957e+00 -8.09457362e-01 -3.28789920e-01
4.16872591e-01 -9.82207656e-01 1.61108184e+00 2.41096526e-01
-1.17167115e+00 -6.15801930e-01 -9.90706086e-01 -2.06692412e-01
-2.73501188e-01 3.61955315e-01 2.69922018e-01 6.33156538e-01
-1.10961497e+00 3.98994565e-01 -8.28853250e-01 -4.23134267e-01
4.53282923e-01 1.05292551e-01 -8.21127072e-02 -8.99581909e-02
-9.59626853e-01 7.11259902e-01 1.36755943e-01 -2.08477601e-01
-7.98340976e-01 -7.54353881e-01 -8.20735753e-01 3.17584397e-03
2.87213027e-01 -9.07826185e-01 1.05423272e+00 -1.64203203e+00
-2.09640026e+00 7.63102710e-01 -7.76823312e-02 -1.42240152e-01
9.36435521e-01 -1.37467802e-01 -1.97996423e-01 -6.03950359e-02
5.84590845e-02 1.14299524e+00 1.32405865e+00 -1.34959626e+00
-4.53759819e-01 1.13682086e-02 7.82844350e-02 2.89910585e-01
-4.52451706e-01 -9.29738805e-02 -7.37282515e-01 -1.17763925e+00
-4.55928862e-01 -1.07035112e+00 -1.05814606e-01 4.96252984e-01
-4.80887711e-01 2.64386863e-01 1.07297075e+00 -6.78244114e-01
1.06888604e+00 -2.19608164e+00 3.59095067e-01 -1.02427013e-01
9.08555165e-02 2.82024771e-01 -4.23021346e-01 4.25702244e-01
1.40329212e-01 1.00469284e-01 -4.00430948e-01 -5.08001149e-01
-5.64733893e-02 3.33548397e-01 -2.64901549e-01 3.99846807e-02
3.43910277e-01 1.00675797e+00 -7.53870487e-01 -2.38864362e-01
3.19804966e-01 6.04844093e-01 -7.08650649e-01 3.59188557e-01
-5.46006322e-01 8.07353199e-01 -3.31702650e-01 2.20556453e-01
5.61929703e-01 -1.96159318e-01 4.57156412e-02 -6.24671280e-02
-2.76777870e-03 -2.53037047e-02 -9.08615947e-01 1.78806865e+00
-6.14064634e-01 7.93580890e-01 6.06078506e-02 -4.88540232e-01
8.47749233e-01 -1.39293941e-02 7.28992298e-02 -7.71421552e-01
1.26236320e-01 -1.55424541e-02 -1.00286059e-01 -4.09438670e-01
5.68260670e-01 8.88418779e-03 1.67803034e-01 5.69271445e-01
-1.47285149e-01 -6.08941793e-01 2.51701266e-01 3.01137388e-01
8.28312576e-01 4.09722626e-01 -9.47243422e-02 -7.85636306e-02
2.96581894e-01 -2.80997634e-01 2.59311706e-01 7.05360889e-01
1.23769626e-01 8.68523836e-01 4.87076044e-01 -1.86212718e-01
-1.36960971e+00 -8.53315234e-01 2.18929633e-01 1.05636740e+00
-7.77571872e-02 -2.30595410e-01 -1.11098933e+00 -7.77497411e-01
-2.73606539e-01 1.10121179e+00 -8.03782046e-01 -2.29703724e-01
-4.89125997e-01 -3.61783415e-01 3.79603118e-01 5.04477203e-01
7.65744805e-01 -1.09029579e+00 -4.90809679e-01 -3.02621275e-02
-1.94141537e-01 -1.01969171e+00 -1.16861975e+00 4.31585088e-02
-5.14134943e-01 -6.94117367e-01 -1.11804235e+00 -8.61011326e-01
1.12226534e+00 1.32961303e-01 9.46737707e-01 -4.18812446e-02
-1.77511394e-01 3.30878019e-01 -3.63594353e-01 -3.46636206e-01
-8.79729450e-01 4.90413755e-02 -4.09548372e-01 1.43103182e-01
-1.76720396e-01 -4.07912821e-01 -7.75383472e-01 3.45144629e-01
-1.41141260e+00 6.62713468e-01 5.07495165e-01 1.04824328e+00
3.95662487e-01 -1.75725222e-01 2.39427418e-01 -1.18987072e+00
7.56065905e-01 -7.22285733e-02 -5.74799240e-01 3.05206358e-01
-5.31723619e-01 1.81424156e-01 8.31807673e-01 -7.66162813e-01
-1.33095717e+00 1.23706184e-01 -1.53539971e-01 -5.26582062e-01
-1.46322548e-01 -3.95043269e-02 -1.55422032e-01 3.59092536e-03
6.51393712e-01 5.13010025e-01 -1.31413070e-02 -1.08243331e-01
6.02087557e-01 5.56043208e-01 3.94779712e-01 -6.86150610e-01
8.39531183e-01 5.18846393e-01 -3.93035978e-01 -7.14696229e-01
-6.82013869e-01 3.09589058e-01 -6.44851089e-01 -7.91097134e-02
1.01282513e+00 -7.10396349e-01 -1.45266866e-02 7.67160118e-01
-1.03258169e+00 -1.06001043e+00 -4.76959080e-01 -3.85340350e-03
-6.94144666e-01 4.34906155e-01 -6.77534461e-01 -3.69094342e-01
-2.80354708e-01 -1.33523571e+00 1.23393250e+00 1.85431808e-01
-1.89191848e-01 -1.04478526e+00 -1.29510000e-01 2.62093931e-01
7.67533362e-01 1.98759496e-01 1.09723055e+00 -8.48662481e-02
-4.91293699e-01 7.74658769e-02 -1.49251014e-01 5.05513728e-01
5.35567164e-01 1.30537987e-01 -7.57322073e-01 -3.89846981e-01
-1.35393202e-01 -1.93482623e-01 7.71135569e-01 2.22857043e-01
1.25970221e+00 -2.96010613e-01 -1.20480664e-01 8.09497893e-01
1.21479750e+00 3.88565034e-01 8.23437154e-01 2.71332949e-01
8.66354704e-01 3.62254769e-01 2.83637375e-01 3.47233385e-01
1.96529686e-01 5.98280013e-01 1.03789248e-01 -3.58711123e-01
-5.04320800e-01 -5.64775527e-01 4.52320993e-01 5.65006077e-01
2.20294699e-01 -6.85568333e-01 -5.78302145e-01 3.22046489e-01
-1.55639327e+00 -7.72114217e-01 1.55034691e-01 2.05311871e+00
1.20620060e+00 2.18639031e-01 -1.89041018e-01 -2.61371195e-01
7.80330598e-01 2.49460265e-01 -8.45063686e-01 -4.79091614e-01
-1.01762168e-01 8.35826397e-02 3.28023523e-01 4.43356991e-01
-8.62025559e-01 1.35266650e+00 5.66338015e+00 8.22121799e-01
-1.44653642e+00 -5.59616368e-03 9.67604995e-01 -1.48069859e-01
-5.67565203e-01 -1.24685392e-01 -6.07037723e-01 7.00447559e-01
4.95829642e-01 8.63668174e-02 8.43028605e-01 4.04168904e-01
3.91807973e-01 -3.53722349e-02 -1.03513229e+00 8.05194438e-01
3.16243619e-01 -1.27434111e+00 4.64559019e-01 -1.53904557e-01
1.19007194e+00 -2.55613714e-01 3.49054307e-01 2.36635730e-01
4.59244102e-01 -8.65444124e-01 8.01771104e-01 5.33510804e-01
1.30273783e+00 -4.14690524e-01 1.61556393e-01 1.99738503e-01
-7.09535003e-01 1.71698064e-01 -2.91127264e-01 1.75870478e-01
1.74271777e-01 3.68028998e-01 -5.52691996e-01 1.49426997e-01
4.54812497e-01 5.97744882e-01 -6.72875524e-01 3.82775664e-01
-4.96499419e-01 5.03189206e-01 -7.36231729e-02 7.57390931e-02
1.61957934e-01 -3.19047064e-01 3.61418366e-01 1.11513317e+00
4.39962804e-01 -2.46222727e-02 6.50442466e-02 1.25182879e+00
-2.80770034e-01 6.11215271e-02 -6.46148980e-01 -8.07129368e-02
1.98782578e-01 1.07192087e+00 -7.30852246e-01 -4.81033474e-01
-6.07138872e-01 1.79808104e+00 2.66856045e-01 6.38368487e-01
-1.06376517e+00 -4.93163258e-01 4.85553682e-01 1.61833972e-01
5.18640757e-01 -2.33328506e-01 -3.58299732e-01 -1.31379533e+00
4.33857925e-02 -1.10115397e+00 -1.92434385e-01 -1.20246851e+00
-1.14354444e+00 6.73118234e-01 -2.28529885e-01 -9.41994369e-01
-9.05059129e-02 -5.88037372e-01 -8.60517263e-01 9.44631398e-01
-1.37248671e+00 -1.28969049e+00 -4.59569901e-01 6.53444350e-01
1.09141028e+00 -1.74948290e-01 5.30286789e-01 1.30958244e-01
-5.55002570e-01 7.59987175e-01 2.33139679e-01 1.03830144e-01
1.12652230e+00 -1.31284785e+00 7.34005392e-01 8.09013486e-01
-1.49660528e-01 3.61221194e-01 7.00190663e-01 -8.10875654e-01
-1.35551214e+00 -1.25937617e+00 3.87863249e-01 -4.43214595e-01
5.55971742e-01 -7.10716248e-01 -8.69083762e-01 7.83064723e-01
6.33754313e-01 -1.95275590e-01 3.43392849e-01 -5.38873851e-01
-2.27148518e-01 1.61097571e-01 -1.00081074e+00 1.11909711e+00
1.03931081e+00 -3.33887637e-01 -1.44051194e-01 3.41583908e-01
7.03121781e-01 -6.57087564e-01 -4.91338819e-01 -7.08837761e-03
3.99058551e-01 -9.50867355e-01 5.52062750e-01 -3.13884497e-01
8.35496366e-01 -2.31383950e-01 5.27032204e-02 -1.68470907e+00
-2.53711462e-01 -7.72416234e-01 2.22328946e-01 1.33524740e+00
3.93066913e-01 -6.18378699e-01 7.20633984e-01 6.03093266e-01
-6.31482229e-02 -6.52638674e-01 -3.79330128e-01 -5.57846844e-01
2.95631677e-01 3.08977887e-02 5.95494270e-01 9.55701768e-01
-4.51183856e-01 4.75929379e-01 -6.19213045e-01 -1.43606454e-01
2.07071945e-01 -7.81584978e-02 9.74303603e-01 -6.39080465e-01
-4.93758261e-01 -5.25944233e-01 1.75128095e-02 -1.40252852e+00
1.85976446e-01 -6.36905193e-01 1.06809750e-01 -1.54142928e+00
2.21138552e-01 -4.02928352e-01 2.99444646e-01 4.32391196e-01
-3.98574978e-01 1.48224667e-01 3.71099442e-01 1.33098245e-01
-2.44398028e-01 8.00730288e-01 1.89787662e+00 -3.77390832e-01
-2.18911126e-01 -1.31137788e-01 -6.72933638e-01 4.70997214e-01
9.70203519e-01 -2.49177992e-01 -6.68868124e-01 -9.41799283e-01
8.86612907e-02 1.10359499e-02 1.40208751e-01 -7.58353233e-01
5.55886179e-02 -1.99033827e-01 6.51887298e-01 -1.10851772e-01
1.75916985e-01 -7.59655237e-01 4.15223315e-02 3.83845806e-01
-5.68422675e-01 2.06576765e-01 3.01982105e-01 5.65051973e-01
2.23117298e-03 -1.70779720e-01 9.38633740e-01 -9.18109566e-02
-5.03013611e-01 2.85841435e-01 -4.60818768e-01 5.61458245e-02
1.00568235e+00 -1.90647349e-01 -4.17635828e-01 -6.73817635e-01
-5.18204451e-01 1.02812521e-01 1.00930130e+00 6.75515056e-01
5.94309688e-01 -1.20515347e+00 -7.67622113e-01 5.30077875e-01
3.55751738e-02 -8.74933526e-02 2.40026072e-01 5.46623945e-01
-6.36102974e-01 3.12743485e-02 -2.75000513e-01 -5.22083044e-01
-1.04232514e+00 6.81930423e-01 4.18440908e-01 -1.58628449e-01
-6.36560082e-01 7.05488324e-01 7.73890615e-01 -3.86172771e-01
5.80912903e-02 -2.99323350e-01 2.16016918e-01 -4.33260053e-01
4.86699373e-01 -2.78537758e-02 -1.93303764e-01 -2.37077087e-01
3.28317523e-01 5.90521514e-01 -2.61369884e-01 -1.92565158e-01
9.57032502e-01 -3.64320695e-01 5.45223504e-02 3.03091794e-01
9.63041425e-01 2.23524958e-01 -1.86805606e+00 -1.93334907e-01
-5.02913713e-01 -6.15623593e-01 -3.30677405e-02 -1.06954396e+00
-1.22615445e+00 9.07884896e-01 6.05594397e-01 -9.29826275e-02
1.25827610e+00 -2.25710616e-01 8.99155676e-01 7.62300864e-02
1.05081089e-01 -1.05347860e+00 4.24044400e-01 5.21500051e-01
1.05601120e+00 -1.26587522e+00 -3.18448991e-01 -3.51481348e-01
-1.00190878e+00 1.12537134e+00 8.60798419e-01 2.67759897e-02
2.66983122e-01 4.28111702e-01 2.87491709e-01 5.99083938e-02
-6.94499195e-01 1.36762545e-01 1.73404902e-01 6.48984849e-01
5.18412530e-01 -8.14763457e-02 3.15798400e-03 -2.89863572e-02
-1.90111607e-01 -4.23776917e-02 5.93039036e-01 8.50216329e-01
-1.47833422e-01 -1.13425505e+00 -1.84747934e-01 4.31450099e-01
-2.50028938e-01 -2.82563359e-01 -5.11426330e-01 5.86034417e-01
-4.01864052e-02 7.25447774e-01 1.00185335e-01 -1.77794501e-01
3.16530228e-01 5.13181165e-02 5.81787705e-01 -6.80108786e-01
-4.82181638e-01 1.81879580e-01 -2.31385708e-01 -3.05141509e-01
-1.39214933e-01 -5.55583596e-01 -9.88433182e-01 -4.15751904e-01
-1.97081432e-01 -2.05474377e-01 5.35945892e-01 6.77675545e-01
4.33997184e-01 8.50826263e-01 6.38382792e-01 -8.23823452e-01
-4.52102900e-01 -8.66071761e-01 -3.73545676e-01 5.42551041e-01
1.98397622e-01 -3.15172464e-01 -2.11072803e-01 4.45941687e-01]
|
[11.46506118774414, -0.2872617244720459]
|
f14f1005-cf2f-48e5-a86e-16f7079a3800
|
gamma-and-vega-hedging-using-deep
|
2205.05614
| null |
https://arxiv.org/abs/2205.05614v4
|
https://arxiv.org/pdf/2205.05614v4.pdf
|
Gamma and Vega Hedging Using Deep Distributional Reinforcement Learning
|
We show how D4PG can be used in conjunction with quantile regression to develop a hedging strategy for a trader responsible for derivatives that arrive stochastically and depend on a single underlying asset. We assume that the trader makes the portfolio delta neutral at the end of each day by taking a position in the underlying asset. We focus on how trades in the options can be used to manage gamma and vega. The option trades are subject to transaction costs. We consider three different objective functions. We reach conclusions on how the optimal hedging strategy depends on the trader's objective function, the level of transaction costs, and the maturity of the options used for hedging. We also investigate the robustness of the hedging strategy to the process assumed for the underlying asset.
|
['Jun Yuan', 'Zeyu Wang', 'Zissis Poulos', 'John Hull', 'Soroush Farghadani', 'Jacky Chen', 'Jay Cao']
|
2022-05-10
| null | null | null | null |
['distributional-reinforcement-learning']
|
['methodology']
|
[-6.33683980e-01 3.80935147e-02 -2.92584822e-02 -3.47542129e-02
-4.63028997e-01 -9.21470702e-01 2.39638269e-01 -1.37178704e-01
-3.22286636e-01 8.87046576e-01 -1.15433529e-01 -6.23613358e-01
-1.47787005e-01 -1.18362939e+00 -2.34677926e-01 -5.21705508e-01
-6.12310506e-02 6.79483771e-01 2.13989213e-01 -2.24232748e-01
5.09466290e-01 8.02718163e-01 -7.04826713e-01 -1.70094430e-01
5.91601908e-01 1.37375653e+00 -3.82110953e-01 6.62310123e-01
-3.43473405e-01 6.25442386e-01 -8.13244820e-01 -7.33763516e-01
7.04365015e-01 -3.51171166e-01 -3.57190073e-01 -1.12498002e-02
-6.48792624e-01 -6.63354337e-01 8.45082626e-02 8.89584541e-01
1.69199705e-01 3.02146971e-01 1.07052422e+00 -1.16550899e+00
-1.19936891e-01 6.07564509e-01 -2.75162458e-01 2.07095146e-01
-3.81782234e-01 -1.00206248e-01 9.54417884e-01 -4.63033050e-01
-4.49437909e-02 8.85730207e-01 5.03665268e-01 3.88286918e-01
-1.05748403e+00 -2.50176728e-01 -2.25611776e-01 -5.55771172e-01
-6.79345906e-01 -1.18083425e-01 3.76160622e-01 -4.64876235e-01
6.95930898e-01 3.49340767e-01 5.73091209e-01 4.03539032e-01
9.71737921e-01 2.41978526e-01 7.73806036e-01 -2.95014977e-01
8.90344024e-01 7.20896050e-02 3.81635712e-03 -3.05952966e-01
6.89234197e-01 2.26118058e-01 1.93607152e-01 -6.04125857e-01
1.14749944e+00 -1.37263298e-01 -5.08463569e-02 -1.35689065e-01
-3.89866382e-01 1.29583991e+00 -2.40173265e-02 9.78776813e-02
-7.48125076e-01 1.59482032e-01 2.47452244e-01 6.61828399e-01
5.79807460e-01 4.46451694e-01 -6.06137812e-01 -1.42659470e-01
-8.30853164e-01 3.24132860e-01 1.46547365e+00 5.86727083e-01
2.72439923e-02 3.05502892e-01 -2.62513369e-01 3.61368746e-01
1.60635948e-01 4.23821986e-01 1.31292924e-01 -1.40451765e+00
5.72846591e-01 -1.29469365e-01 7.69224465e-01 -5.40081002e-02
3.41665111e-02 -2.17378527e-01 -1.11999728e-01 8.48872542e-01
5.97924650e-01 -6.22378647e-01 -3.68972719e-01 1.34263158e+00
-2.17030078e-01 -3.68805826e-01 1.57095850e-01 1.86393112e-01
-6.74002111e-01 6.45730138e-01 -2.33048975e-01 -6.32405818e-01
7.52172589e-01 -5.70337713e-01 -5.67591786e-01 5.65216005e-01
3.91125828e-01 -6.12539351e-01 3.66415381e-01 5.96755326e-01
-1.63160729e+00 1.54835925e-01 -8.43545020e-01 4.52706575e-01
-3.18658203e-02 -3.01119477e-01 2.92567611e-02 9.34633732e-01
-9.77097631e-01 1.17991173e+00 -8.40863526e-01 4.75793511e-01
-8.42707977e-02 1.87725931e-01 4.79691595e-01 8.51420879e-01
-1.05955243e+00 1.15466404e+00 1.94545522e-01 7.43128434e-02
-5.28044581e-01 -6.41669154e-01 -1.17232464e-01 4.00213540e-01
2.64796197e-01 -6.34565651e-01 1.51803505e+00 -8.62142861e-01
-1.69700670e+00 9.81332064e-02 7.11002290e-01 -9.43070531e-01
1.13192284e+00 -1.00220457e-01 -1.78182423e-01 1.93635345e-01
1.52376732e-02 -3.74695688e-01 3.95415545e-01 -4.45914298e-01
-5.81439018e-01 -3.12786371e-01 -9.51744765e-02 2.08931640e-01
2.73697764e-01 5.25108613e-02 5.19703507e-01 -9.25387025e-01
-1.64579958e-01 -8.61786485e-01 -1.01491399e-01 -1.70061558e-01
-2.18228251e-01 2.58074045e-01 8.65491554e-02 -7.50753164e-01
9.27407146e-01 -1.70327008e+00 -2.85328686e-01 3.89797777e-01
-6.24845326e-01 -3.56423199e-01 5.91157436e-01 8.23688626e-01
-6.29594699e-02 4.57344472e-01 -3.38960558e-01 -2.62472476e-03
3.63045901e-01 9.40086544e-02 -1.01474404e+00 4.10179436e-01
-2.31795490e-01 5.93027532e-01 -1.45831645e-01 2.26124123e-01
-1.51677191e-01 9.03055072e-02 -1.24541730e-01 2.99605310e-01
-5.96092641e-01 -6.16827421e-02 -5.00904202e-01 2.95409918e-01
6.72088325e-01 3.19709063e-01 9.22719538e-02 5.83670437e-01
-8.22811484e-01 3.41024250e-01 -1.11332750e+00 5.62869310e-01
-4.79615211e-01 -1.91831924e-02 2.69744564e-02 -5.41305423e-01
8.50395739e-01 3.00852686e-01 3.00834656e-01 -1.95707247e-01
1.50477663e-01 8.63088727e-01 -1.17117934e-01 1.34478852e-01
4.48408604e-01 -1.06714916e+00 -1.22123227e-01 9.54937994e-01
-4.66799378e-01 -1.26594990e-01 -3.62285078e-02 -2.56016672e-01
6.93662822e-01 -1.34310871e-01 2.58602589e-01 -7.19076931e-01
-5.57589494e-02 -2.17320904e-01 5.02526164e-01 1.74732208e-01
2.51892775e-01 3.52921247e-01 1.27252948e+00 -1.22932509e-01
-1.05322134e+00 -1.06715310e+00 -3.27039301e-01 6.58874810e-01
-3.08536470e-01 7.58152246e-01 -5.63391328e-01 -3.34522247e-01
6.33420348e-01 1.42086291e+00 -6.61363304e-01 1.26665950e-01
-2.64740855e-01 -9.27190065e-01 3.20589125e-01 5.47255337e-01
4.65239078e-01 -9.40492749e-01 -1.04727006e+00 5.54447651e-01
5.19718945e-01 -1.46607429e-01 -6.27175748e-01 4.42342788e-01
-1.14850867e+00 -7.80472875e-01 -1.22675490e+00 2.45256945e-01
2.66606659e-01 -2.47631580e-01 8.79130065e-01 -3.49258363e-01
3.67843807e-01 4.93434489e-01 -2.56597716e-02 -5.14087796e-01
-4.94072258e-01 -4.44847047e-02 -3.79276246e-01 1.67090416e-01
-4.50973511e-01 -3.50916475e-01 -6.43608809e-01 2.85060823e-01
-9.55785573e-01 -6.90589190e-01 -1.34211585e-01 5.17135620e-01
3.65623355e-01 4.93363351e-01 7.79008985e-01 -7.31971383e-01
7.41612494e-01 -6.01588368e-01 -1.29340637e+00 5.00783205e-01
-9.08888042e-01 4.43385720e-01 2.23120824e-01 -1.83159605e-01
-1.47859073e+00 -6.95953190e-01 1.62147701e-01 -2.20951587e-01
5.52150488e-01 4.94623423e-01 -4.24982995e-01 -6.16709562e-03
-1.95546106e-01 -1.86925426e-01 1.96363226e-01 -8.86054158e-01
1.08255304e-01 1.67056680e-01 3.80421847e-01 -6.21587694e-01
3.49036366e-01 3.11604023e-01 3.41554314e-01 -2.82700211e-01
-3.03134590e-01 1.02837086e-01 -3.15829307e-01 1.83417365e-01
6.33150697e-01 -5.21785259e-01 -8.81074905e-01 8.39235902e-01
-9.27470446e-01 -6.51809335e-01 -7.16416836e-01 4.22387451e-01
-1.04496253e+00 -5.52187748e-02 -8.23002279e-01 -1.39704001e+00
-2.76021600e-01 -4.45149839e-01 1.71888068e-01 2.70157427e-01
-3.47796921e-03 -1.52344537e+00 5.10931730e-01 -1.80588707e-01
5.69055855e-01 6.15602195e-01 1.02506256e+00 -9.78168964e-01
-6.72827601e-01 -1.55167520e-01 3.10475528e-01 4.55439866e-01
1.67661637e-01 3.02058578e-01 -3.21087390e-01 -7.30222091e-02
9.89716709e-01 1.60923168e-01 6.74539447e-01 6.32330656e-01
3.89060348e-01 -7.25392461e-01 2.50878692e-01 5.08890629e-01
1.45758712e+00 6.76966250e-01 5.81976950e-01 8.88332427e-01
-1.00322448e-01 9.89562392e-01 7.93274760e-01 6.95591211e-01
-5.37378788e-02 3.54338795e-01 2.70729274e-01 6.53884888e-01
9.46440876e-01 -7.72797465e-02 4.78030026e-01 1.64502010e-01
1.56631798e-01 -2.92861521e-01 -6.07797027e-01 4.72225070e-01
-1.56404185e+00 -1.08460093e+00 6.24390971e-03 2.65804005e+00
8.48108351e-01 4.43149507e-01 7.69277751e-01 -4.88937944e-02
6.87492311e-01 -2.00107977e-01 -8.59132230e-01 -7.93048322e-01
-1.45347595e-01 1.14275411e-01 1.12833667e+00 8.69355202e-01
-5.27876437e-01 4.65148874e-02 6.66646385e+00 3.85060698e-01
-1.13487256e+00 -2.39298195e-01 1.00735307e+00 -2.32097745e-01
-7.53325284e-01 3.26722890e-01 -8.59369516e-01 9.97561097e-01
1.21965003e+00 -8.33262682e-01 2.47784644e-01 7.19094038e-01
2.12531030e-01 -1.86488390e-01 -8.64562392e-01 -8.01008642e-02
-7.77420223e-01 -1.15143764e+00 -2.09892064e-01 6.22778654e-01
6.60053194e-01 -4.57437396e-01 4.89316195e-01 -5.47605678e-02
4.42260563e-01 -8.77533734e-01 9.52642858e-01 9.70393717e-01
2.82142073e-01 -1.41217697e+00 7.89457619e-01 4.15093958e-01
-9.17816520e-01 -8.34028330e-03 -3.00529778e-01 -4.61621881e-02
5.74128509e-01 5.70271194e-01 -4.17001605e-01 5.33689439e-01
2.11866926e-02 -2.88898677e-01 7.19446391e-02 1.09047449e+00
1.38231134e-02 5.07871330e-01 -4.20825511e-01 2.14390695e-01
2.98102796e-01 -9.02944148e-01 3.02209198e-01 5.14021456e-01
8.53629470e-01 3.61542761e-01 -3.16491961e-01 1.23978353e+00
2.15679213e-01 -7.22769946e-02 -2.22698718e-01 -1.12193845e-01
4.84926432e-01 5.41532993e-01 -9.30232704e-01 -1.20390072e-01
-2.43029371e-01 7.14178085e-01 -5.44832647e-01 4.87591624e-01
-5.45495808e-01 -4.13935930e-01 3.68102968e-01 4.19700921e-01
6.29905343e-01 -1.03769012e-01 -8.20155978e-01 -8.67110252e-01
2.08757415e-01 -4.35507149e-01 6.63304627e-01 -3.48918617e-01
-1.33079159e+00 2.12805256e-01 2.87148774e-01 -8.16472173e-01
-8.48565817e-01 -4.56570506e-01 -1.23788905e+00 1.44743609e+00
-1.34448409e+00 -2.42284238e-01 8.03951442e-01 2.37016499e-01
2.69322038e-01 -2.16620460e-01 3.65986139e-01 -6.31175518e-01
-3.20795894e-01 9.66544226e-02 8.80962610e-01 -7.89613202e-02
-1.05859907e-02 -1.55995548e+00 8.07870269e-01 5.37265778e-01
-2.37524599e-01 4.74425882e-01 7.25160241e-01 -1.24276209e+00
-8.21497977e-01 -7.20083296e-01 6.08502150e-01 -1.66059703e-01
1.03544426e+00 2.72093624e-01 -1.03065598e+00 8.68373632e-01
1.84154794e-01 -4.71921325e-01 5.85769832e-01 -5.19039333e-01
1.53323174e-01 -4.76029404e-02 -1.41854095e+00 1.49759680e-01
-7.46661350e-02 -3.44086319e-01 -6.63704813e-01 -2.61055559e-01
4.47551608e-01 -1.04999095e-01 -8.44894111e-01 -2.19733194e-02
5.52834570e-01 -1.01143599e+00 5.52976966e-01 -4.45489258e-01
-3.05564255e-01 3.43588173e-01 -1.12598285e-01 -1.15486741e+00
8.30780789e-02 -1.18640733e+00 1.98181104e-02 1.26075828e+00
2.71837860e-01 -1.22972178e+00 4.61926252e-01 1.47232366e+00
5.25912404e-01 -5.39469004e-01 -1.30271518e+00 -1.28022456e+00
9.79388535e-01 1.37222663e-01 7.31002152e-01 3.03450257e-01
-2.37304300e-01 -2.29188114e-01 -2.65823513e-01 -7.20740408e-02
1.02266061e+00 4.68283981e-01 -7.80769959e-02 -1.22665477e+00
-5.70930600e-01 -5.65244138e-01 5.45618415e-01 -4.63274986e-01
-8.44772812e-03 -2.96530008e-01 -2.69668162e-01 -7.97023296e-01
-1.47326604e-01 -3.27275455e-01 -4.25728232e-01 -3.04519609e-02
2.55600065e-01 -5.69979250e-01 4.30127025e-01 4.35770422e-01
4.02179986e-01 6.08151495e-01 6.58974588e-01 4.26932752e-01
-3.90013188e-01 9.81288195e-01 -5.76753616e-01 7.45791256e-01
1.02124441e+00 -3.92389596e-01 -3.77053827e-01 2.01303065e-01
3.55283141e-01 9.32324767e-01 1.78661615e-01 -2.62225032e-01
9.42312405e-02 -6.58071399e-01 -1.22238614e-01 -8.06603312e-01
-4.49782610e-02 -4.74852949e-01 8.22747767e-01 6.66084886e-01
-5.51797688e-01 3.20948750e-01 -2.24457737e-02 3.70013565e-01
2.03049071e-02 -1.04551649e+00 9.02409971e-01 -1.96297035e-01
3.06475937e-01 1.90099031e-01 -4.39688534e-01 2.34238684e-01
1.08465803e+00 9.44711566e-02 -7.36195222e-02 -5.39142311e-01
-9.31265891e-01 3.42233568e-01 9.09821391e-01 -3.66141587e-01
1.59719661e-01 -1.26580083e+00 -4.79986817e-01 -1.96639240e-01
-9.69889343e-01 -4.29450929e-01 -8.30855444e-02 3.61190259e-01
-7.39174008e-01 3.67550284e-01 -4.62812841e-01 4.55328405e-01
-4.77965593e-01 4.62271631e-01 8.25237155e-01 -3.91595542e-01
-3.34321588e-01 3.43258262e-01 2.49246314e-01 4.94138479e-01
-3.27021368e-02 -3.64907473e-01 1.86011493e-01 6.52857959e-01
5.13856411e-01 1.05993581e+00 -2.59070784e-01 -4.71501797e-02
-2.69188546e-02 3.74130279e-01 1.72243655e-01 -1.06415081e+00
1.35844147e+00 -1.60789803e-01 -1.80502310e-01 9.94837642e-01
8.05563092e-01 1.50103971e-01 -1.70150542e+00 5.24147570e-01
4.37553346e-01 -3.02887052e-01 -2.75400847e-01 -5.88148892e-01
-1.03532767e+00 8.13384712e-01 -2.30793983e-01 8.04183185e-01
6.94799960e-01 -5.82160532e-01 7.97842741e-01 -6.97918683e-02
5.16587853e-01 -1.19806755e+00 -2.20591992e-01 2.09849283e-01
1.04698563e+00 -3.90424937e-01 -4.02646000e-03 -1.95397288e-02
-7.78367519e-01 1.48589146e+00 -3.54977071e-01 -5.78962147e-01
8.55014563e-01 4.21629429e-01 -5.67965172e-02 2.48607486e-01
-8.11610937e-01 3.11005533e-01 -3.37684639e-02 4.23603430e-02
4.54929173e-02 3.51272285e-01 -5.73917150e-01 7.91197896e-01
-1.93895400e-01 8.91129747e-02 1.03730762e+00 9.59181130e-01
-8.29719961e-01 -1.17390275e+00 -4.50621694e-01 1.13002993e-01
-1.03553295e+00 -7.11305626e-03 -4.20423687e-01 7.98680127e-01
-3.95438612e-01 2.20033228e-01 2.30788827e-01 6.85101986e-01
4.06503350e-01 4.21413481e-01 1.05266452e-01 -4.57297355e-01
-5.88346362e-01 4.07417923e-01 8.74617994e-02 -6.69174641e-02
2.59496808e-01 -9.67261255e-01 -1.29453874e+00 -6.59086287e-01
-2.27002800e-01 5.09266615e-01 2.53570080e-01 6.32960498e-01
-1.27168655e-01 -8.13614205e-02 1.11523414e+00 -2.51078427e-01
-1.90429974e+00 -4.15664107e-01 -1.50769639e+00 -4.04395431e-01
2.26749703e-01 -3.85744721e-01 -9.28824365e-01 -3.24401408e-01]
|
[4.887043476104736, 3.935810089111328]
|
3d6ea531-c75d-4875-b359-5859f0560b05
|
graphs-constraints-and-search-for-the
|
2210.09880
| null |
https://arxiv.org/abs/2210.09880v2
|
https://arxiv.org/pdf/2210.09880v2.pdf
|
Graphs, Constraints, and Search for the Abstraction and Reasoning Corpus
|
The Abstraction and Reasoning Corpus (ARC) aims at benchmarking the performance of general artificial intelligence algorithms. The ARC's focus on broad generalization and few-shot learning has made it difficult to solve using pure machine learning. A more promising approach has been to perform program synthesis within an appropriately designed Domain Specific Language (DSL). However, these too have seen limited success. We propose Abstract Reasoning with Graph Abstractions (ARGA), a new object-centric framework that first represents images using graphs and then performs a search for a correct program in a DSL that is based on the abstracted graph space. The complexity of this combinatorial search is tamed through the use of constraint acquisition, state hashing, and Tabu search. An extensive set of experiments demonstrates the promise of ARGA in tackling some of the complicated object-centric tasks of the ARC rather efficiently, producing programs that are correct and easy to understand.
|
['Scott Sanner', 'Elias B. Khalil', 'Yudong Xu']
|
2022-10-18
| null | null | null | null |
['program-synthesis']
|
['computer-code']
|
[ 3.29969317e-01 5.90464115e-01 -4.18046057e-01 -1.16528518e-01
-5.42724013e-01 -4.05324161e-01 7.78162479e-01 4.10861135e-01
1.36860430e-01 4.27938461e-01 -6.02872521e-02 -7.63836443e-01
-1.58651829e-01 -9.55033422e-01 -6.70551956e-01 -2.18793720e-01
-1.23531818e-01 7.34189749e-01 4.35497910e-01 -1.81898803e-01
2.47744113e-01 4.62101161e-01 -1.92480731e+00 2.18638256e-01
7.81923115e-01 6.09154046e-01 -1.81249864e-02 7.83270597e-01
-5.35480618e-01 1.07171631e+00 -5.96924007e-01 -4.78889018e-01
1.06040426e-01 -6.86852455e-01 -1.10582352e+00 1.81143075e-01
3.00453305e-01 1.19927965e-01 -1.16328582e-01 1.22152412e+00
-2.36274555e-01 1.18234776e-01 2.84637332e-01 -1.69323146e+00
-3.84271562e-01 6.86284661e-01 -8.47935379e-02 -1.52900785e-01
6.86434686e-01 4.43045229e-01 1.38323939e+00 -3.92249644e-01
8.87242198e-01 1.44739091e+00 4.85146344e-01 7.06674993e-01
-1.67183888e+00 -2.18040422e-01 -1.48284808e-02 2.16563076e-01
-1.25822282e+00 -1.36140570e-01 5.89561701e-01 -4.68206227e-01
1.49341607e+00 4.56915617e-01 8.60663533e-01 7.11967170e-01
1.99111536e-01 9.02105808e-01 8.63885403e-01 -8.82300735e-01
8.33750904e-01 1.36732340e-01 3.41417551e-01 1.23416269e+00
4.53641266e-01 2.92963963e-02 -3.28565210e-01 -4.17159379e-01
3.29927623e-01 -1.95399314e-01 -7.59203658e-02 -1.10974169e+00
-1.18087971e+00 1.02088368e+00 2.86978066e-01 1.66758955e-01
-7.31542008e-03 4.15634453e-01 7.12101340e-01 3.97704542e-01
-1.08505145e-01 1.02390170e+00 -2.65654266e-01 -5.81903271e-02
-9.00183201e-01 5.15000045e-01 1.30632520e+00 1.22200739e+00
7.86777735e-01 1.63559496e-01 -6.07896149e-02 1.75937757e-01
1.03056006e-01 -5.57408929e-02 2.02725410e-01 -1.20894861e+00
9.44413692e-02 1.03104961e+00 -1.30476311e-01 -6.70107841e-01
-1.75851732e-01 1.04045644e-01 -1.39876857e-01 8.10459077e-01
3.09080780e-01 5.38425334e-02 -9.58470285e-01 1.45939624e+00
1.86552614e-01 -1.16905414e-01 3.17956179e-01 6.43377125e-01
4.56330568e-01 7.48044908e-01 1.95595160e-01 -2.93318421e-01
1.20956922e+00 -1.05172038e+00 -3.76633108e-01 -3.83736759e-01
9.74871933e-01 -1.75256990e-02 1.22912717e+00 4.64871496e-01
-1.28867495e+00 -2.79184878e-01 -1.41364336e+00 1.09476708e-02
-5.89646876e-01 -4.74792153e-01 1.28092456e+00 8.58790278e-01
-1.03283429e+00 5.17636001e-01 -7.49235928e-01 -4.00251895e-01
5.09128153e-01 3.26291144e-01 -9.47292149e-02 -4.08118457e-01
-7.08791494e-01 1.00374150e+00 8.83382201e-01 -4.33325976e-01
-7.91986465e-01 -7.15464056e-01 -1.34414005e+00 3.40311885e-01
9.50062633e-01 -7.95336902e-01 1.36045980e+00 -8.52274537e-01
-1.22355723e+00 8.71631980e-01 1.18465036e-01 -7.83418238e-01
1.56253070e-01 3.80559325e-01 -2.10445866e-01 9.42652300e-02
-5.92352003e-02 5.31733394e-01 5.93139350e-01 -1.04666257e+00
-5.19986391e-01 -2.66050071e-01 5.76178789e-01 -1.16575815e-01
3.87298353e-02 1.25136122e-01 -3.12979966e-01 -2.47012526e-01
-1.84000909e-01 -9.98938084e-01 -3.89892995e-01 -6.74128607e-02
-2.37017855e-01 -4.82144445e-01 6.55825794e-01 -2.83624768e-01
1.36003172e+00 -2.08158994e+00 6.38425708e-01 2.44937018e-01
1.78292051e-01 3.03215265e-01 -1.60385165e-02 7.30284333e-01
-1.10899158e-01 2.70055741e-01 -4.15932000e-01 2.27395773e-01
4.76111799e-01 4.99526650e-01 -3.15727919e-01 1.15597993e-01
2.97275782e-01 1.07094753e+00 -1.10843861e+00 -7.23607659e-01
2.69855797e-01 -2.62660742e-01 -9.19623733e-01 3.52961533e-02
-1.23430133e+00 -3.23195845e-01 -4.13904816e-01 7.91993737e-01
6.91762641e-02 -3.27723116e-01 4.80133057e-01 1.35273814e-01
-4.66293916e-02 6.55589392e-03 -1.20451999e+00 1.94188535e+00
-3.48940730e-01 5.93804181e-01 -1.49712354e-01 -1.00944221e+00
7.55898714e-01 7.47625306e-02 1.61935985e-01 -5.03321111e-01
-1.54912174e-01 1.50143221e-01 -4.22922745e-02 -5.91432929e-01
3.41029257e-01 -1.44816294e-01 -5.93015552e-01 4.22437876e-01
7.01319352e-02 -8.90347719e-01 6.19911492e-01 4.29287195e-01
1.39600945e+00 3.40424597e-01 6.03531480e-01 -3.88646811e-01
4.88125652e-01 9.31860268e-01 5.16629279e-01 9.15856183e-01
6.55768663e-02 3.24540317e-01 8.85358334e-01 -8.03915024e-01
-1.12437189e+00 -7.89639413e-01 2.65345603e-01 7.81765103e-01
-1.92455687e-02 -8.88236940e-01 -8.48013639e-01 -6.96713030e-01
-3.31359421e-04 1.29769933e+00 -3.34104478e-01 -3.26538563e-01
-4.87621129e-01 -2.71642834e-01 4.46225256e-01 3.07218105e-01
4.30673122e-01 -1.17679417e+00 -1.15359116e+00 1.90860197e-01
3.03788453e-01 -8.50244999e-01 7.07259029e-03 2.63297439e-01
-6.66644216e-01 -1.19311404e+00 -1.39591143e-01 -7.89484859e-01
6.54824376e-01 -1.51608601e-01 1.41762304e+00 5.65649688e-01
-8.32816064e-01 7.30809689e-01 -3.36599767e-01 -4.33804095e-01
-7.75672197e-01 -4.09081131e-02 -4.81455952e-01 -5.54333270e-01
4.25124228e-01 -3.53791684e-01 1.64096177e-01 -2.57607281e-01
-9.26744282e-01 2.21213594e-01 3.08030725e-01 9.52123046e-01
2.92161226e-01 6.05093658e-01 1.40792467e-02 -1.11059022e+00
6.86643064e-01 -1.97947368e-01 -1.08264613e+00 5.72224259e-01
-7.86488891e-01 4.10535306e-01 5.79639792e-01 -1.72877863e-01
-9.04751122e-01 3.81526351e-01 4.58849937e-01 -3.96587521e-01
-2.32445464e-01 7.49121785e-01 -1.18922301e-01 -9.68859270e-02
9.08054054e-01 1.01647891e-01 9.63559747e-03 4.04949002e-02
3.91065001e-01 1.33422837e-01 6.05631292e-01 -9.50895011e-01
8.92541409e-01 8.38987809e-03 4.74211514e-01 -7.74455786e-01
-6.30724907e-01 -1.53905109e-01 -3.39475900e-01 2.93555483e-02
8.91599119e-01 -5.48551023e-01 -7.54314065e-01 -6.23178743e-02
-9.97823477e-01 -5.91946304e-01 -7.05728769e-01 -5.49119571e-03
-9.60846603e-01 2.20577672e-01 -1.91696540e-01 -9.60598111e-01
-5.12786545e-02 -1.25800896e+00 8.64582837e-01 7.49922842e-02
-7.04877794e-01 -6.93958700e-01 1.18720599e-01 1.91324756e-01
9.87306163e-02 5.08241236e-01 1.73017895e+00 -6.87847018e-01
-1.00046229e+00 -3.23489904e-01 -3.09298262e-02 -9.46092978e-03
-3.09530616e-01 2.68729150e-01 -5.33999562e-01 -1.77384987e-01
-2.27936774e-01 -6.23892009e-01 2.86298692e-01 -2.51524919e-03
1.06883609e+00 -3.32593977e-01 -3.10187727e-01 3.22277158e-01
1.76076043e+00 2.85034031e-01 6.81874335e-01 3.68147731e-01
3.40574890e-01 5.45526206e-01 6.29284680e-01 2.72298545e-01
1.53841034e-01 5.46811461e-01 3.83749992e-01 3.03802580e-01
-2.47498095e-01 -2.16697380e-01 4.40382510e-02 2.00995252e-01
1.29027605e-01 8.22744593e-02 -1.47886753e+00 7.49591708e-01
-2.12872815e+00 -1.08501625e+00 1.21206008e-01 1.84249508e+00
6.87185705e-01 4.09930766e-01 1.78394750e-01 2.62561768e-01
3.46059918e-01 3.14124487e-02 -3.80198419e-01 -9.85233486e-01
3.30965251e-01 4.65260297e-01 2.23396733e-01 5.11755526e-01
-6.78602874e-01 9.81936514e-01 6.62083006e+00 5.56071460e-01
-5.64348221e-01 -4.48629588e-01 1.66932911e-01 1.29728168e-01
-4.11689281e-01 5.26549995e-01 -4.70407039e-01 1.14911020e-01
1.21491313e+00 -7.80857682e-01 1.04289711e+00 1.31641018e+00
-4.16475832e-01 -4.02298540e-01 -1.48676348e+00 6.34021938e-01
3.54744419e-02 -1.79160488e+00 1.86172634e-01 -2.09832996e-01
6.80053532e-01 -4.13767934e-01 -3.18943679e-01 8.40820312e-01
5.86832166e-01 -1.08513021e+00 7.51055121e-01 3.59546363e-01
4.60832030e-01 -9.20424819e-01 3.28314722e-01 4.29896384e-01
-1.08389664e+00 -3.61842483e-01 -1.51896730e-01 -3.01009774e-01
-3.55226606e-01 -1.67552512e-02 -1.00296783e+00 5.34803748e-01
4.92849022e-01 2.73129910e-01 -5.96093714e-01 1.06522262e+00
-2.87475288e-01 1.14505261e-01 -5.09721078e-02 -3.47461462e-01
3.87956411e-01 -4.68996838e-02 5.87126195e-01 1.17269278e+00
-5.32453209e-02 3.06051046e-01 4.76859033e-01 1.25966096e+00
1.31234139e-01 -2.91171163e-01 -9.26429272e-01 -5.41527748e-01
3.09513181e-01 9.23866451e-01 -8.41289341e-01 -6.13367021e-01
-6.47990763e-01 6.15218937e-01 2.39429623e-01 2.60155171e-01
-6.89945817e-01 -6.96841300e-01 4.07482982e-01 -1.22317798e-01
4.57661837e-01 -2.67111689e-01 -3.92789632e-01 -1.09847796e+00
-1.33524746e-01 -1.27599633e+00 6.90546513e-01 -1.02504289e+00
-5.97315311e-01 2.92506665e-01 4.94636416e-01 -6.18701160e-01
-6.59034014e-01 -5.62987685e-01 -6.24630868e-01 6.02384508e-01
-1.02658820e+00 -9.35374975e-01 -2.70668000e-01 4.88022387e-01
8.21666777e-01 -1.53614163e-01 1.14367402e+00 -3.02349031e-01
-5.31721950e-01 2.19879001e-01 -5.56045532e-01 -2.74524778e-01
-1.45719171e-01 -1.45810056e+00 6.50018513e-01 8.71805787e-01
9.70037729e-02 5.89077532e-01 1.14702356e+00 -6.49143755e-01
-2.21398497e+00 -8.07403088e-01 7.11745083e-01 -3.63619447e-01
8.74777079e-01 -3.00798208e-01 -8.85594726e-01 9.43373919e-01
9.85723659e-02 -4.57057916e-02 -5.41179581e-03 8.06152262e-03
-6.25409842e-01 1.10544868e-01 -1.21145463e+00 8.48248184e-01
1.02836275e+00 -6.76066875e-01 -9.26366448e-01 4.95567650e-01
8.36541414e-01 -3.73676121e-01 -7.70579219e-01 1.70516297e-01
6.46557808e-02 -8.90720785e-01 8.84044647e-01 -9.33685958e-01
4.94843930e-01 -4.08482283e-01 -2.29616508e-01 -1.06791377e+00
-3.06277394e-01 -8.50797415e-01 -3.27674001e-01 8.99313450e-01
2.84031630e-01 -4.22493100e-01 1.01291192e+00 1.16204739e+00
-1.97661132e-01 -5.82454741e-01 -6.05689287e-01 -1.02933347e+00
-2.67596602e-01 -2.80325711e-01 7.33294129e-01 8.90695572e-01
7.14683831e-01 4.59368795e-01 2.05215856e-01 6.70094937e-02
7.90311754e-01 6.65401995e-01 8.53093624e-01 -1.33065581e+00
-4.18865591e-01 -4.82631326e-01 -5.53355515e-01 -5.50066903e-02
4.39520121e-01 -1.10563052e+00 1.52191117e-01 -1.56234825e+00
1.14259385e-01 -1.34122610e-01 2.24983945e-01 6.61429286e-01
3.09379101e-01 -4.92259234e-01 2.33653396e-01 -1.18088074e-01
-7.05540538e-01 1.03328340e-01 8.61400247e-01 -4.33598965e-01
-1.25795871e-01 -2.83833712e-01 -3.79074514e-01 6.89508557e-01
6.70074642e-01 -3.20532292e-01 -6.12843931e-01 -8.59299079e-02
2.57858127e-01 4.35804576e-01 4.30273324e-01 -1.22480428e+00
5.19929945e-01 -5.49193442e-01 -2.93933153e-01 -2.10743159e-01
1.63779557e-01 -9.51572418e-01 4.93298441e-01 7.59984612e-01
-4.45194304e-01 9.19270664e-02 4.11489516e-01 4.60337102e-01
-1.05460040e-01 -6.70251548e-01 7.99443722e-01 -6.19458854e-01
-1.39229250e+00 -8.90216045e-03 -3.33219796e-01 1.40329629e-01
1.34048462e+00 -3.40646058e-01 -1.40138373e-01 2.03157403e-02
-6.27468348e-01 1.41752213e-01 8.50015581e-01 1.99481204e-01
4.29131866e-01 -1.00128925e+00 -7.98647404e-02 2.20639914e-01
4.59461272e-01 -1.20313287e-01 -3.10621500e-01 2.47723088e-01
-7.10883737e-01 5.06315947e-01 -4.03330833e-01 -3.67425859e-01
-1.31499684e+00 1.22843003e+00 2.76776165e-01 -2.06668183e-01
-9.59524989e-01 4.99298006e-01 -2.12303445e-01 -2.22943172e-01
4.50275809e-01 -3.25714558e-01 3.02423269e-01 -3.44732851e-01
4.70881462e-01 3.19399029e-01 -1.91202071e-02 3.45062129e-02
-4.44684923e-01 9.62144136e-02 5.83295785e-02 -3.91348116e-02
1.46626878e+00 4.40959513e-01 -3.53189647e-01 3.05192590e-01
8.35324109e-01 -4.78332162e-01 -8.77749681e-01 -1.53389871e-01
6.19411230e-01 -4.62752759e-01 1.05276361e-01 -6.68500423e-01
-4.57726926e-01 6.02352142e-01 -2.56368052e-02 5.82445681e-01
9.35224473e-01 -6.36852905e-02 4.00695056e-01 8.52011561e-01
7.72915125e-01 -1.10609460e+00 1.20025530e-01 4.46152627e-01
6.24961138e-01 -9.36067283e-01 2.26639897e-01 -3.94520044e-01
-4.56517279e-01 1.33683765e+00 6.91285491e-01 -7.67075792e-02
-1.13109529e-01 3.79834801e-01 -6.53982818e-01 -5.76831698e-01
-9.13665593e-01 -1.49986312e-01 -1.08167738e-01 6.08872890e-01
-1.31604999e-01 -6.52701929e-02 7.23236948e-02 1.72741786e-01
-3.32442597e-02 2.18905449e-01 8.06739867e-01 1.49174654e+00
-6.33325815e-01 -1.02365923e+00 -2.18253434e-01 1.89293295e-01
-9.70896333e-02 4.92814630e-02 -3.56033772e-01 1.14057994e+00
-2.48381957e-01 6.68459892e-01 -2.35746205e-01 -1.20536864e-01
3.55157346e-01 4.15527135e-01 7.92846620e-01 -1.07008064e+00
-4.82773155e-01 -5.42714715e-01 5.54676175e-01 -7.80903697e-01
-9.18316543e-02 -4.69973981e-01 -1.35842359e+00 -3.23474258e-01
1.78270657e-02 2.53189176e-01 6.19269729e-01 7.48673737e-01
7.68873990e-02 4.96855706e-01 1.46911114e-01 -5.52418768e-01
-6.40508890e-01 9.77531895e-02 -4.37661737e-01 2.46990532e-01
2.37904549e-01 -3.31578940e-01 -1.99395940e-01 3.00660841e-02]
|
[8.551314353942871, 7.229255676269531]
|
339c93ac-407c-4972-9fb8-d907545889ca
|
unest-local-spatial-representation-learning
|
2209.14378
| null |
https://arxiv.org/abs/2209.14378v1
|
https://arxiv.org/pdf/2209.14378v1.pdf
|
UNesT: Local Spatial Representation Learning with Hierarchical Transformer for Efficient Medical Segmentation
|
Transformer-based models, capable of learning better global dependencies, have recently demonstrated exceptional representation learning capabilities in computer vision and medical image analysis. Transformer reformats the image into separate patches and realize global communication via the self-attention mechanism. However, positional information between patches is hard to preserve in such 1D sequences, and loss of it can lead to sub-optimal performance when dealing with large amounts of heterogeneous tissues of various sizes in 3D medical image segmentation. Additionally, current methods are not robust and efficient for heavy-duty medical segmentation tasks such as predicting a large number of tissue classes or modeling globally inter-connected tissues structures. Inspired by the nested hierarchical structures in vision transformer, we proposed a novel 3D medical image segmentation method (UNesT), employing a simplified and faster-converging transformer encoder design that achieves local communication among spatially adjacent patch sequences by aggregating them hierarchically. We extensively validate our method on multiple challenging datasets, consisting anatomies of 133 structures in brain, 14 organs in abdomen, 4 hierarchical components in kidney, and inter-connected kidney tumors). We show that UNesT consistently achieves state-of-the-art performance and evaluate its generalizability and data efficiency. Particularly, the model achieves whole brain segmentation task complete ROI with 133 tissue classes in single network, outperforms prior state-of-the-art method SLANT27 ensembled with 27 network tiles, our model performance increases the mean DSC score of the publicly available Colin and CANDI dataset from 0.7264 to 0.7444 and from 0.6968 to 0.7025, respectively.
|
['Yucheng Tang', 'Bennett A. Landman', 'Yuankai Huo', 'Zizhao Zhang', 'Richard G. Abramson', 'Thomas A. Lasko', 'Zhoubing Xu', 'Shunxing Bao', 'Thomas Li', 'Ho Hin Lee', 'Riqiang Gao', 'Leon Y. Cai', 'Yinchi Zhou', 'Qi Yang', 'Xin Yu']
|
2022-09-28
| null | null | null | null |
['brain-segmentation']
|
['medical']
|
[ 4.59713578e-01 4.91266638e-01 -2.30083853e-01 -2.22188130e-01
-8.76623988e-01 -3.97056311e-01 2.67494500e-01 5.52971996e-02
-1.78303316e-01 6.08046293e-01 2.18147993e-01 -1.80324748e-01
-2.32523695e-01 -6.75914764e-01 -9.16303396e-01 -8.95117342e-01
-4.14380461e-01 7.23609686e-01 3.76691014e-01 8.83374512e-02
-3.11718941e-01 6.17944479e-01 -8.54669034e-01 4.29510951e-01
8.57367992e-01 1.13721573e+00 4.93503451e-01 5.92660964e-01
-4.17683087e-03 9.73140538e-01 -2.49933541e-01 -1.57850906e-01
2.43974268e-01 -3.93413067e-01 -1.09481823e+00 1.24683477e-01
5.58106422e-01 -1.32875562e-01 -4.34363395e-01 9.63448524e-01
6.12646222e-01 -3.54958892e-01 6.41424060e-01 -6.18361056e-01
-6.05009139e-01 8.85396600e-01 -7.66956925e-01 3.52726698e-01
-3.41387540e-01 4.98241000e-02 8.18175673e-01 -4.19311911e-01
7.15079904e-01 8.69598508e-01 1.02090561e+00 6.12991154e-01
-1.43733549e+00 -7.71021307e-01 5.33557907e-02 -5.98102808e-02
-1.17508471e+00 -1.00892015e-01 4.97456759e-01 -4.60942060e-01
9.63837624e-01 1.69101223e-01 9.20551836e-01 9.93557096e-01
6.88054442e-01 9.56370533e-01 1.14976037e+00 1.88827425e-01
-1.08098559e-01 -4.28725392e-01 2.70048350e-01 1.10013449e+00
3.04661989e-02 -1.43868312e-01 -3.27863634e-01 1.56732742e-02
1.25959980e+00 7.95231983e-02 -3.25687468e-01 -2.18445227e-01
-1.63160288e+00 6.88425481e-01 1.13161266e+00 5.34559011e-01
-4.95061576e-01 2.73206443e-01 4.01354969e-01 8.28792751e-02
4.96215016e-01 2.02075779e-01 -4.71381128e-01 4.35592890e-01
-9.25265193e-01 -2.02744767e-01 5.42844951e-01 8.43539774e-01
4.37190264e-01 -2.23973751e-01 -4.18228209e-01 7.93616831e-01
1.48849994e-01 2.88327426e-01 5.63592732e-01 -8.04159105e-01
1.56282470e-01 6.64521515e-01 -7.42089272e-01 -8.14578235e-01
-9.11305845e-01 -1.08370793e+00 -1.71795690e+00 -1.81351200e-01
2.77220428e-01 8.08343142e-02 -1.38086009e+00 1.76753592e+00
3.59418541e-01 3.90231490e-01 -1.16523884e-01 9.04464364e-01
1.20575678e+00 3.82077694e-01 6.86200783e-02 -1.45323738e-01
1.43051219e+00 -1.12953484e+00 -3.55870873e-01 -2.22045228e-01
6.43614709e-01 -4.72982675e-01 4.48876888e-01 1.23597734e-01
-1.14535797e+00 -4.01871532e-01 -8.05316091e-01 -1.96484670e-01
6.28794730e-02 -1.31472319e-01 7.77680635e-01 2.02276736e-01
-1.33703935e+00 5.45673430e-01 -1.17946053e+00 -2.53681511e-01
1.02594519e+00 6.48577452e-01 -5.01865447e-01 -7.62877092e-02
-8.49985659e-01 7.44994402e-01 1.21753648e-01 1.33859158e-01
-1.31173182e+00 -1.25476170e+00 -7.46311128e-01 -1.74246300e-02
-3.55455764e-02 -1.17007506e+00 1.00722206e+00 -7.82051086e-01
-1.28296769e+00 1.19471157e+00 -7.53591955e-03 -7.73851454e-01
4.79831874e-01 1.71139643e-01 9.54513550e-02 4.87686872e-01
2.67062753e-01 1.14374793e+00 5.46701431e-01 -1.06284094e+00
-2.61123538e-01 -6.97795868e-01 -2.17888549e-01 3.18276495e-01
4.07233555e-03 -4.02886510e-01 -6.19483948e-01 -7.40255952e-01
3.78190756e-01 -9.97010589e-01 -6.24457359e-01 2.63478905e-01
-6.24777615e-01 1.63104072e-01 6.21225238e-01 -7.81181097e-01
6.74252331e-01 -1.99880862e+00 3.76564205e-01 1.42897323e-01
7.37827063e-01 -1.21131256e-01 -1.37422875e-01 -3.78694475e-01
-1.57887444e-01 2.19691694e-02 -6.40742958e-01 -2.59593129e-01
-3.87615442e-01 4.56490219e-01 3.05159599e-01 6.57722175e-01
-4.94720489e-02 1.36221945e+00 -7.06420183e-01 -8.44432771e-01
1.64357945e-01 6.23615503e-01 -6.59936070e-01 8.29942301e-02
2.18108654e-01 8.22522342e-01 -5.83233058e-01 8.56017292e-01
6.06517494e-01 -9.11055207e-01 1.56754062e-01 -5.52816391e-01
3.32132727e-01 2.88156290e-02 -3.20419788e-01 2.19331837e+00
-3.86774570e-01 4.15032327e-01 3.80243599e-01 -1.42071414e+00
6.55506432e-01 3.82610947e-01 1.17709720e+00 -8.00702512e-01
1.06190361e-01 2.04633246e-03 8.48981440e-02 -3.62278402e-01
-2.24273443e-01 -3.16908211e-01 -8.30324963e-02 8.79718140e-02
3.75557721e-01 -1.97738692e-01 3.92241217e-02 1.51986524e-01
1.41806555e+00 -2.42654294e-01 1.26925394e-01 -6.05033040e-01
3.23297560e-01 -7.44922971e-03 4.72501963e-01 8.16272855e-01
-3.32103610e-01 8.49241912e-01 5.48689961e-01 -3.98705810e-01
-8.44066024e-01 -1.22619987e+00 -3.57622892e-01 7.37792373e-01
2.31135979e-01 -1.14575796e-01 -8.25948179e-01 -8.69671881e-01
-2.11824514e-02 1.41484767e-01 -1.01981080e+00 -1.03830509e-01
-7.64341056e-01 -1.05906367e+00 7.82629490e-01 5.08661926e-01
7.47258067e-01 -9.12178695e-01 -6.72331333e-01 3.82088691e-01
-4.23201740e-01 -1.24171567e+00 -4.47859287e-01 5.53729832e-01
-1.19149089e+00 -1.01223958e+00 -1.02639556e+00 -1.15055072e+00
7.37623572e-01 9.57538188e-02 1.31467927e+00 7.29468241e-02
-7.86678970e-01 2.23824188e-01 -5.21043316e-02 3.74631807e-02
-3.04828048e-01 3.08964998e-01 -5.41284978e-01 -2.85830349e-01
-2.81283557e-01 -5.88903010e-01 -7.82808781e-01 3.71179134e-01
-7.26880729e-01 4.86361176e-01 1.00038707e+00 1.28635275e+00
9.17632580e-01 -1.94680527e-01 3.45573008e-01 -9.87603903e-01
1.14760108e-01 -6.13968313e-01 -2.32539311e-01 3.22577655e-01
-3.11937124e-01 -5.35658114e-02 3.87031943e-01 -1.44465700e-01
-8.68214428e-01 2.90194988e-01 -2.22361416e-01 -4.68009502e-01
-1.18679009e-01 4.24840957e-01 4.09878790e-01 -2.95439333e-01
4.53960896e-01 4.94514912e-01 3.86986583e-01 -1.88501060e-01
2.16890275e-01 7.87819326e-02 7.23731279e-01 -5.46012700e-01
3.48382324e-01 5.71619689e-01 2.72604436e-01 -6.52854323e-01
-8.97203863e-01 -3.32862437e-01 -7.20539153e-01 -1.30022854e-01
1.25569797e+00 -9.80965495e-01 -6.98950529e-01 5.69682002e-01
-9.83622909e-01 -5.60776532e-01 -2.63637394e-01 3.32553208e-01
-6.52470112e-01 2.78146476e-01 -1.17533422e+00 7.92618003e-03
-7.09070206e-01 -1.45214808e+00 1.34036279e+00 6.58521950e-02
-6.35940135e-02 -1.01189518e+00 -1.95419341e-01 5.70970833e-01
5.09803474e-01 5.75382292e-01 1.12826443e+00 -2.85961926e-01
-7.47930706e-01 2.21118063e-01 -3.33953977e-01 2.03361481e-01
1.67326078e-01 -5.18069804e-01 -7.04463184e-01 -3.04779053e-01
-1.54885948e-01 -2.85189539e-01 1.15654278e+00 9.78581429e-01
1.61683977e+00 -1.18057072e-01 -6.46532774e-01 9.87836778e-01
1.35064220e+00 -2.36467849e-02 4.32728231e-01 -1.04715936e-01
8.52545917e-01 4.47345823e-01 -1.23371847e-01 7.34615847e-02
4.10519660e-01 3.19102496e-01 6.36521280e-01 -5.75331330e-01
-5.95129609e-01 1.09904028e-01 -6.88668434e-03 1.01928401e+00
-1.67289659e-01 -1.91084594e-01 -9.00391757e-01 6.49013817e-01
-1.73979020e+00 -6.82957709e-01 4.20786999e-02 1.65975332e+00
8.91362548e-01 -4.06392217e-02 -2.75251001e-01 -3.82646650e-01
5.83346486e-01 8.59145746e-02 -7.64072716e-01 2.59195536e-01
-1.60595432e-01 3.54907870e-01 6.88376904e-01 2.48229027e-01
-1.19556093e+00 8.96909535e-01 6.38933182e+00 8.10106993e-01
-1.13418746e+00 4.88765776e-01 1.22115123e+00 -3.60001177e-02
-9.72812325e-02 -5.31496882e-01 -4.33480531e-01 2.83360690e-01
6.60393953e-01 1.64713696e-01 2.48906806e-01 3.96594703e-01
-2.21291661e-01 1.54563621e-01 -1.04472256e+00 1.06495380e+00
-5.74760139e-02 -1.71657825e+00 1.11636937e-01 1.30062133e-01
7.74957657e-01 6.56925023e-01 9.66851190e-02 7.78532848e-02
4.16755736e-01 -1.48027456e+00 4.52274591e-01 3.40185881e-01
8.16941321e-01 -4.67109561e-01 7.40253747e-01 2.93130904e-01
-1.17383552e+00 4.09892350e-02 -2.96104372e-01 4.40908283e-01
1.31975383e-01 8.34025562e-01 -8.17053497e-01 6.52640283e-01
1.04970407e+00 1.05773962e+00 -3.42448354e-01 9.55524564e-01
1.04584575e-01 5.55544853e-01 -4.74880606e-01 2.26508453e-01
4.85065401e-01 6.72988743e-02 4.51831102e-01 1.23898339e+00
2.33553827e-01 2.65289664e-01 2.02806756e-01 8.26971591e-01
-2.79627919e-01 -9.09742415e-02 -4.10445988e-01 2.92071521e-01
-4.52461652e-02 1.45215619e+00 -9.85801458e-01 -3.98104548e-01
-2.65668690e-01 8.77592266e-01 4.17435229e-01 2.79801309e-01
-1.02822483e+00 2.67530888e-01 5.18197179e-01 2.29641143e-03
3.36636275e-01 8.89369007e-03 -5.19112766e-01 -9.95383263e-01
-3.53005439e-01 -7.17332900e-01 5.91662884e-01 -4.98754293e-01
-1.47693920e+00 7.41441309e-01 -3.17860961e-01 -9.98799086e-01
2.90861666e-01 -4.79357719e-01 -1.86618403e-01 4.01655436e-01
-1.46540964e+00 -1.39408064e+00 -4.33657348e-01 8.17692935e-01
5.12466192e-01 4.91244271e-02 7.43179739e-01 4.50594127e-01
-3.79626542e-01 5.13097107e-01 -1.70893833e-01 2.15766594e-01
4.83313739e-01 -1.14105082e+00 1.01568289e-02 4.15794730e-01
-1.04057796e-01 3.19829494e-01 8.81569535e-02 -4.95006770e-01
-1.30234253e+00 -1.33560264e+00 5.03979862e-01 -1.42257154e-01
5.41492939e-01 -2.54397422e-01 -8.32158625e-01 8.71887505e-01
2.86667734e-01 4.84500170e-01 4.84133989e-01 -1.73265025e-01
-1.70887157e-01 -1.55603975e-01 -1.27966475e+00 2.85775155e-01
1.34538352e+00 -2.57756114e-01 -3.44454288e-01 6.08588755e-01
8.87793422e-01 -8.63655448e-01 -1.45371711e+00 7.16190994e-01
4.52028543e-01 -9.36062515e-01 1.19033182e+00 -2.11369410e-01
5.81069231e-01 -1.17397010e-01 -7.98591450e-02 -1.19163656e+00
-7.59990871e-01 -3.97202522e-01 1.52875796e-01 5.83007932e-01
3.64833117e-01 -6.26453638e-01 8.98367345e-01 4.22659293e-02
-7.78532386e-01 -1.09819782e+00 -1.36160135e+00 -3.62272263e-01
3.89200032e-01 -1.06683984e-01 2.91903019e-01 9.81388569e-01
-1.81736439e-01 3.39006990e-01 -1.37924775e-01 1.66083932e-01
9.27624583e-01 1.83058813e-01 2.58233815e-01 -1.07194805e+00
-3.23185474e-01 -7.58123577e-01 -4.62165713e-01 -1.43656635e+00
1.97733492e-01 -1.50351655e+00 3.00651020e-03 -1.66867685e+00
6.37098312e-01 -6.02156043e-01 -5.49406528e-01 7.86313295e-01
1.78583920e-01 5.08657753e-01 1.08631123e-02 2.06379920e-01
-5.13578534e-01 4.11882073e-01 1.98649812e+00 -6.56500101e-01
2.90336788e-01 -2.34320909e-01 -6.97685182e-01 5.75677276e-01
5.33294559e-01 -3.82069528e-01 -3.64545494e-01 -6.25023127e-01
-4.10836011e-01 3.23321253e-01 6.89820945e-01 -1.00016701e+00
4.40639794e-01 3.65952313e-01 7.80722082e-01 -5.76122880e-01
1.79848298e-01 -8.41350794e-01 3.54259193e-01 9.00138378e-01
-3.58802557e-01 -1.73747167e-01 1.49872065e-01 3.71215999e-01
-2.98281699e-01 3.30534995e-01 1.04785287e+00 -5.41819036e-01
-5.14028668e-01 8.53447735e-01 -2.44147986e-01 6.17635734e-02
1.07713842e+00 -1.20737948e-01 -3.64484072e-01 1.30631134e-01
-1.00729680e+00 3.61659676e-01 9.43640806e-03 1.90744191e-01
4.97072309e-01 -1.14248800e+00 -8.47213328e-01 2.21284717e-01
-2.21730500e-01 5.44460595e-01 6.40976489e-01 1.42436731e+00
-7.65898943e-01 3.92288268e-01 -2.87178218e-01 -1.37776577e+00
-1.09315443e+00 1.91041946e-01 7.95584321e-01 -7.17544019e-01
-1.10214543e+00 1.15181732e+00 9.07658577e-01 -6.37615561e-01
1.10977478e-01 -8.32374275e-01 -4.79495972e-02 -2.22270742e-01
6.78642541e-02 -6.05306476e-02 2.13816613e-01 -5.71925342e-01
-5.14451623e-01 9.05351222e-01 -1.95292935e-01 5.25121689e-01
1.51238334e+00 8.85842717e-04 -4.39280868e-01 -7.52375498e-02
1.42355847e+00 -6.72431231e-01 -1.30169034e+00 -3.13902527e-01
-4.61552620e-01 1.15895540e-01 2.95255512e-01 -8.94602358e-01
-1.74880970e+00 7.93715954e-01 7.36617029e-01 -3.42461243e-02
1.30635309e+00 4.40694273e-01 1.13171446e+00 4.58816737e-02
4.77351010e-01 -5.43070197e-01 9.37550217e-02 5.13659775e-01
7.67764032e-01 -1.07879138e+00 -6.14771508e-02 -5.76143384e-01
-5.02273381e-01 9.35051858e-01 5.01888454e-01 -1.32591322e-01
8.56899202e-01 6.09347701e-01 -2.82959826e-02 -5.50374508e-01
-8.06374907e-01 5.59335873e-02 3.68123204e-01 6.04145169e-01
4.85931814e-01 3.04555595e-01 1.26420274e-01 4.79795128e-01
-1.98440284e-01 -1.51967391e-01 1.35689884e-01 5.76683879e-01
-1.90510243e-01 -5.46518028e-01 4.09007445e-02 6.47095203e-01
-6.69242382e-01 -1.51706144e-01 -1.88374054e-02 7.37909734e-01
1.59388453e-01 3.17374527e-01 1.99592561e-01 -7.91084468e-02
5.86695075e-02 -4.25610423e-01 8.10511410e-01 -3.37984473e-01
-8.60725939e-01 3.55450034e-01 -1.37146071e-01 -7.20716059e-01
-5.31564415e-01 -5.53243577e-01 -1.52657700e+00 -4.63634320e-02
7.02442019e-04 -2.45948434e-01 3.76391113e-01 8.39780629e-01
4.40959990e-01 1.06531203e+00 3.64796966e-01 -7.99251199e-01
-2.33095095e-01 -6.92882001e-01 -5.56489527e-01 3.00832808e-01
4.44046557e-01 -4.70259488e-01 -6.17333734e-03 2.64916644e-02]
|
[14.584197998046875, -2.5007331371307373]
|
d0d1945d-8178-470c-8390-b3162b6a8383
|
enhancing-self-disclosure-in-neural-dialog
|
2109.05090
| null |
https://arxiv.org/abs/2109.05090v2
|
https://arxiv.org/pdf/2109.05090v2.pdf
|
Enhancing Self-Disclosure In Neural Dialog Models By Candidate Re-ranking
|
Neural language modelling has progressed the state-of-the-art in different downstream Natural Language Processing (NLP) tasks. One such area is of open-domain dialog modelling, neural dialog models based on GPT-2 such as DialoGPT have shown promising performance in single-turn conversation. However, such (neural) dialog models have been criticized for generating responses which although may have relevance to the previous human response, tend to quickly dissipate human interest and descend into trivial conversation. One reason for such performance is the lack of explicit conversation strategy being employed in human-machine conversation. Humans employ a range of conversation strategies while engaging in a conversation, one such key social strategies is Self-disclosure(SD). A phenomenon of revealing information about one-self to others. Social penetration theory (SPT) proposes that communication between two people moves from shallow to deeper levels as the relationship progresses primarily through self-disclosure. Disclosure helps in creating rapport among the participants engaged in a conversation. In this paper, Self-disclosure enhancement architecture (SDEA) is introduced utilizing Self-disclosure Topic Model (SDTM) during inference stage of a neural dialog model to re-rank response candidates to enhance self-disclosure in single-turn responses from from the model.
|
['Vincent Wade', 'Benjamin Cowan', 'Mayank Soni']
|
2021-09-10
| null | null | null | null |
['open-domain-dialog']
|
['natural-language-processing']
|
[ 5.55375293e-02 1.18001032e+00 -7.57415816e-02 -7.00417280e-01
-3.43358070e-01 -2.67720729e-01 1.13635659e+00 1.55338094e-01
-1.43087670e-01 1.13622034e+00 1.08090889e+00 -6.63000643e-02
3.14567722e-02 -7.75397301e-01 1.95886195e-01 -3.78746033e-01
2.84427971e-01 8.40321600e-01 -4.64927219e-02 -7.75488257e-01
2.08022803e-01 3.97719443e-02 -1.04956794e+00 6.29010141e-01
6.55868530e-01 4.82387841e-01 1.41415656e-01 7.11397886e-01
-9.02999341e-01 1.09368551e+00 -9.23018277e-01 -7.08006799e-01
-1.22136883e-01 -8.03609788e-01 -1.25894797e+00 -8.93017054e-02
-2.69802183e-01 -5.27856708e-01 -1.89270228e-01 7.57258773e-01
4.07035679e-01 5.79353213e-01 6.10865593e-01 -1.25264275e+00
-7.31722713e-01 1.13506675e+00 -1.18799746e-01 -6.23279922e-02
5.36649704e-01 1.19793907e-01 7.31288195e-01 -6.79108202e-01
6.25903428e-01 1.91085935e+00 5.83532512e-01 1.10424101e+00
-1.17356098e+00 -6.54626846e-01 -1.75392225e-01 -1.81280896e-01
-6.23980403e-01 -3.13395590e-01 8.83716285e-01 -4.29315627e-01
1.31665206e+00 1.84477463e-01 5.42814016e-01 1.54907012e+00
4.10778642e-01 5.75085700e-01 1.23848093e+00 -1.10392675e-01
1.33319691e-01 9.46528792e-01 6.49521768e-01 1.94793299e-01
-4.05027807e-01 1.08029217e-01 -9.61140394e-01 -3.34086388e-01
4.54851300e-01 -1.36324003e-01 9.93744731e-02 3.87148917e-01
-7.83056378e-01 1.47188389e+00 5.26324391e-01 6.18853927e-01
-7.24453509e-01 -5.36597252e-01 5.40186644e-01 6.95210636e-01
8.93490434e-01 6.57550514e-01 9.58847161e-03 -4.41928953e-01
-5.87008297e-01 4.34344411e-01 1.42487240e+00 6.38337612e-01
7.09254384e-01 -1.17402323e-01 -5.44912875e-01 1.21337843e+00
2.87017465e-01 -2.58799165e-01 8.02831888e-01 -9.53266382e-01
2.28307590e-01 9.00669396e-01 -9.87802446e-02 -1.13468206e+00
-3.44540060e-01 -1.70514770e-02 -9.42898154e-01 -3.56547795e-02
2.97731072e-01 -7.20554709e-01 -6.02403805e-02 1.77485013e+00
2.74105847e-01 -5.77469051e-01 7.34892607e-01 7.95941651e-01
1.37864220e+00 1.11984336e+00 3.27766389e-01 -2.12117553e-01
1.40682590e+00 -9.35398519e-01 -9.38552082e-01 -4.30279136e-01
4.48988229e-01 -4.59360570e-01 7.88890421e-01 1.96856305e-01
-1.13000560e+00 -6.43267572e-01 -6.57124162e-01 -3.09029490e-01
-4.23688293e-01 -5.85282207e-01 5.06241977e-01 5.61909735e-01
-1.20546234e+00 5.12151241e-01 -8.36148262e-02 -7.92868137e-01
5.79917245e-02 4.41770375e-01 -3.51988018e-01 5.82500041e-01
-1.66456628e+00 1.25577581e+00 1.75619334e-01 -8.89211819e-02
-3.71004760e-01 -7.06549525e-01 -7.41560876e-01 3.46501738e-01
3.33288103e-01 -6.92800820e-01 1.48764324e+00 -8.72938395e-01
-2.17372990e+00 9.57586467e-01 -1.15168549e-01 -9.20027554e-01
4.21172410e-01 -2.72100627e-01 -1.22886799e-01 -2.48226803e-02
-7.25426823e-02 1.39324498e+00 5.90493202e-01 -1.30401981e+00
-2.80691355e-01 -2.66559541e-01 4.83961515e-02 5.46934485e-01
-2.60472029e-01 3.50899607e-01 4.97012079e-01 -2.55326748e-01
-2.21117511e-01 -6.66653574e-01 -2.16100737e-01 -3.60949337e-01
-7.06338048e-01 -7.66942859e-01 9.50230420e-01 -5.74195564e-01
9.28562999e-01 -1.75335348e+00 -1.17946543e-01 -2.80068606e-01
6.12282515e-01 3.26049626e-01 3.18590999e-01 1.10427737e+00
2.06714764e-01 8.90241414e-02 -3.72448340e-02 -7.85228193e-01
2.00296193e-01 2.02446803e-01 -7.11120129e-01 -1.60812795e-01
2.03171447e-01 7.63599098e-01 -6.15766466e-01 -4.29661483e-01
2.26525307e-01 3.08021277e-01 -3.90334845e-01 7.62949646e-01
-4.02692795e-01 8.37373495e-01 -3.98669183e-01 -2.19105124e-01
5.24630427e-01 -1.27911553e-01 1.69084929e-02 4.46100801e-01
-3.53421032e-01 7.13731527e-01 -4.21493858e-01 1.28183663e+00
-2.60203362e-01 8.31677496e-01 3.68523419e-01 -5.44284344e-01
1.52581823e+00 6.31448209e-01 1.30092770e-01 -2.44090691e-01
3.96450281e-01 -3.87134641e-01 4.91908520e-01 -4.45702165e-01
8.99411082e-01 -5.78348458e-01 -2.16206238e-01 1.04509687e+00
6.89025745e-02 -1.56978413e-01 -3.78059894e-01 5.18768430e-01
7.83033073e-01 -4.54600483e-01 7.03090310e-01 -3.36031884e-01
6.25974059e-01 8.37052315e-02 1.56185970e-01 7.51546562e-01
-4.36461657e-01 1.06984593e-01 6.62505746e-01 -2.40823299e-01
-7.98206985e-01 -7.84970522e-01 1.36860609e-01 1.25920558e+00
-9.68952775e-02 -6.16196208e-02 -9.77262020e-01 -2.19713345e-01
-1.65177435e-01 1.09060872e+00 -4.39634979e-01 -2.24406272e-01
-3.85899931e-01 -1.46706536e-01 7.28096783e-01 1.48332983e-01
9.36855733e-01 -1.70521867e+00 -6.96181357e-01 5.79513252e-01
-1.68963656e-01 -8.81384611e-01 -2.52329260e-01 1.71182454e-02
-6.94046080e-01 -2.37419695e-01 -7.92573452e-01 -6.90174699e-01
1.87711284e-01 3.52740884e-02 7.89409935e-01 -1.52569801e-01
1.95304781e-01 2.56110221e-01 -1.34100035e-01 -5.24399638e-01
-1.23214877e+00 2.24387813e-02 8.46077688e-03 -1.05045764e-02
7.10113704e-01 -6.02019608e-01 -4.36712295e-01 2.66728491e-01
-4.22561109e-01 3.51896942e-01 2.50491977e-01 8.17889750e-01
-6.28773749e-01 -4.92891133e-01 1.04927039e+00 -1.05328774e+00
1.88861513e+00 -8.82417202e-01 2.90022731e-01 -1.21419102e-01
-4.35288757e-01 -7.86799863e-02 4.36406702e-01 -5.83312988e-01
-1.91861665e+00 -5.57957292e-01 -1.74968630e-01 8.98170173e-02
-5.29145241e-01 4.65201467e-01 1.89676374e-01 3.27745855e-01
6.74150050e-01 2.26829171e-01 6.41200006e-01 -2.35459000e-01
4.39767510e-01 1.09814286e+00 4.32366759e-01 -3.83636415e-01
1.63346410e-01 8.20399597e-02 -6.21631205e-01 -1.18112838e+00
-5.42798221e-01 -4.73413736e-01 -3.25474292e-01 -4.67087924e-01
9.81816292e-01 -5.49632490e-01 -1.30358243e+00 4.26163971e-01
-1.59061682e+00 -3.27943921e-01 -1.41710386e-01 2.78312117e-01
-2.31121823e-01 2.24431932e-01 -8.72258306e-01 -1.47108197e+00
-6.62567675e-01 -8.03819716e-01 5.86395979e-01 6.39810860e-01
-1.27166510e+00 -1.21037805e+00 2.52392203e-01 5.41283190e-01
8.21467936e-01 -9.59276557e-02 1.06534672e+00 -1.45293021e+00
-1.63320288e-01 1.07875057e-02 -6.07668348e-02 1.61691949e-01
4.49296609e-02 -6.71127975e-01 -1.17650378e+00 1.69042259e-01
6.71957016e-01 -6.62441611e-01 3.43817592e-01 1.96105748e-01
2.08794221e-01 -8.52787733e-01 -2.98535556e-01 -3.07946175e-01
5.55402577e-01 3.68742049e-01 3.88690978e-01 -1.34517536e-01
4.53046598e-02 1.67476034e+00 3.65542144e-01 4.73471373e-01
6.48370445e-01 4.17962641e-01 2.04270869e-03 -3.44470255e-02
1.19534120e-01 -4.12848800e-01 5.54104507e-01 5.54276288e-01
5.27223349e-01 -4.62566614e-01 -6.44916415e-01 2.97640800e-01
-1.82217574e+00 -9.88019049e-01 -1.48907796e-01 1.55548954e+00
1.06789744e+00 1.79233983e-01 1.88196912e-01 -2.51615852e-01
8.54070663e-01 2.83781439e-01 -5.79039812e-01 -1.09205067e+00
-1.17726792e-02 -8.36509690e-02 -9.15030986e-02 1.01021469e+00
-2.45078951e-01 1.19350171e+00 5.59511328e+00 3.07680607e-01
-1.00733089e+00 2.00717345e-01 8.19554687e-01 2.26969957e-01
-1.88707024e-01 -1.35313766e-02 -9.13153410e-01 2.96963274e-01
1.08059251e+00 -6.45200193e-01 1.26691282e-01 7.15021610e-01
3.28935027e-01 -3.55110466e-01 -1.27813411e+00 6.00649953e-01
3.03838905e-02 -1.19621193e+00 1.24396108e-01 1.00555629e-01
2.44472310e-01 -4.76756483e-01 3.16924937e-02 7.21324980e-01
6.02799296e-01 -1.08280361e+00 1.00811526e-01 4.38535571e-01
1.22304618e-01 -3.47393274e-01 7.52779841e-01 9.26065683e-01
-5.02219379e-01 -1.32671550e-01 -2.91730523e-01 -5.11237264e-01
7.19078481e-01 -4.88508418e-02 -1.66837418e+00 5.09739630e-02
3.56305659e-01 7.28889331e-02 1.44877419e-01 2.68786848e-01
-5.48001602e-02 4.55910832e-01 -2.87921745e-02 -6.38831675e-01
5.05215824e-01 -3.59828681e-01 9.77275193e-01 1.09122813e+00
4.57617342e-02 6.37616217e-01 3.01025510e-02 1.44165576e+00
6.29454777e-02 3.22288163e-02 -8.54298711e-01 -2.55891353e-01
5.95569134e-01 1.07548237e+00 -3.36073995e-01 -3.19413185e-01
-1.10771224e-01 7.29535401e-01 2.37166718e-01 7.88474828e-02
-3.30033779e-01 8.48434940e-02 6.07217014e-01 1.79030001e-01
-4.24735367e-01 1.00099854e-01 -3.79197091e-01 -2.91699290e-01
-4.40821528e-01 -8.74415874e-01 2.45088950e-01 -8.17821920e-01
-1.59516478e+00 9.30140615e-01 2.83411205e-01 -4.75054473e-01
-9.05093074e-01 -1.65767357e-01 -1.12978327e+00 1.20227385e+00
-7.60361254e-01 -1.09600651e+00 -1.11282282e-01 4.66852397e-01
1.16670430e+00 -4.52004701e-01 1.05270016e+00 -3.37827682e-01
-1.07627161e-01 3.53632182e-01 -7.23418653e-01 -6.93689138e-02
7.88739860e-01 -1.02763605e+00 4.67279166e-01 2.11428478e-01
-5.15666902e-01 1.05015087e+00 9.20540035e-01 -9.11728203e-01
-7.93086410e-01 -5.10211706e-01 1.33915401e+00 -2.24900961e-01
5.65329552e-01 -5.64407289e-01 -1.29748154e+00 4.73183692e-01
9.54172552e-01 -1.15636015e+00 9.67521369e-01 1.59892127e-01
2.24700555e-01 4.14719582e-01 -1.41153419e+00 8.51676047e-01
5.83528638e-01 -5.75278938e-01 -1.15029895e+00 -6.69535324e-02
1.02217686e+00 -2.76521027e-01 -7.10193276e-01 -1.49388149e-01
2.15758204e-01 -1.29497313e+00 7.07879663e-01 -6.09091818e-01
5.49524069e-01 6.13534033e-01 3.12510282e-01 -1.32143545e+00
8.96249488e-02 -1.44150782e+00 6.48423061e-02 1.54901671e+00
3.24550807e-01 -9.45485175e-01 8.60630751e-01 1.40524483e+00
-1.79084957e-01 -5.01317501e-01 -8.10250044e-01 -1.47830173e-01
3.77146214e-01 1.51548013e-02 4.64060932e-01 9.87049997e-01
8.71277273e-01 1.14019680e+00 -6.43296540e-01 -4.02114511e-01
1.99271888e-01 -3.34097773e-01 1.01572061e+00 -1.63835299e+00
-6.37513101e-02 -5.32398999e-01 2.62125462e-01 -1.41952431e+00
3.86744469e-01 -5.91092825e-01 2.02241391e-01 -1.40767670e+00
1.56599537e-01 -4.39808145e-02 6.62659049e-01 1.36853665e-01
6.05187938e-02 -6.71591461e-01 2.28120327e-01 2.63764948e-01
-3.51247378e-02 8.20091009e-01 1.22147465e+00 1.99055687e-01
-7.26965129e-01 2.47434929e-01 -1.15230644e+00 5.45078337e-01
8.32336068e-01 -5.58106542e-01 -5.71851730e-01 2.48286441e-01
-2.15674445e-01 9.27250981e-01 2.39544198e-01 -5.85246503e-01
8.15232515e-01 1.93275392e-01 -1.82954609e-01 -6.60714328e-01
9.81666446e-01 -4.12492931e-01 1.28570143e-02 6.36785150e-01
-1.19429052e+00 -1.06058065e-02 1.79535314e-01 5.06350517e-01
-2.09509969e-01 -3.20151418e-01 7.64052331e-01 -3.65769029e-01
-3.23678255e-01 -1.73569322e-01 -1.02674818e+00 -2.42768675e-01
8.68182242e-01 -5.57577789e-01 -4.75753069e-01 -1.27512848e+00
-8.31414938e-01 4.29437965e-01 -1.00046761e-01 7.01299191e-01
5.62281847e-01 -8.11802030e-01 -7.57878125e-01 -4.17915359e-02
-2.43186444e-01 -1.25135139e-01 4.42425966e-01 4.76010025e-01
7.21219406e-02 8.17730367e-01 -3.60137522e-01 -4.11478430e-01
-1.46786106e+00 3.86156105e-02 2.27474943e-01 -4.39719409e-01
-8.28746736e-01 1.04743302e+00 5.08139312e-01 -5.38862050e-01
5.48607647e-01 -8.32118168e-02 -6.38712347e-01 7.82189071e-02
5.18318653e-01 3.55339170e-01 -5.02984762e-01 -6.72584236e-01
1.40014365e-01 -3.26305658e-01 -4.39792871e-01 -6.37063920e-01
9.64632034e-01 -3.19562316e-01 -1.70536175e-01 5.84254503e-01
9.95283246e-01 -4.25844491e-01 -1.08787847e+00 -5.62096715e-01
4.14319523e-02 -1.06627278e-01 -3.24599266e-01 -9.30213213e-01
-2.03988194e-01 9.71879900e-01 6.52128011e-02 7.66248405e-01
3.42607230e-01 1.26250207e-01 9.76403832e-01 7.53685832e-01
-3.43241170e-02 -1.07409644e+00 4.00595129e-01 7.91225493e-01
1.28628802e+00 -1.47428286e+00 -4.51861978e-01 -3.75973105e-01
-1.42091882e+00 1.10229564e+00 9.86115813e-01 3.73735689e-02
6.14763796e-01 2.13212729e-01 2.19403282e-01 -4.98327613e-01
-1.27920532e+00 2.84550756e-01 -6.49755001e-02 6.99243724e-01
6.48496270e-01 -1.50111049e-01 -2.44080454e-01 9.91187930e-01
-8.41460228e-01 -3.30569863e-01 5.95510185e-01 5.70437372e-01
-5.72372437e-01 -9.21691537e-01 -2.54013062e-01 2.09730700e-01
-2.10001871e-01 -1.18321680e-01 -1.41669393e+00 6.69083893e-01
-4.93625522e-01 1.51050758e+00 1.77611440e-01 -3.00166577e-01
-4.24153470e-02 3.27271312e-01 -2.59919375e-01 -9.09623206e-01
-1.56409156e+00 -2.95254439e-01 5.13131797e-01 -1.39035508e-01
-3.40550750e-01 -4.74504590e-01 -1.24193668e+00 -5.94427586e-01
-1.16079219e-01 4.88593310e-01 4.24832284e-01 1.08970368e+00
3.47775280e-01 1.52041852e-01 3.96299005e-01 -5.51846504e-01
-8.03986073e-01 -1.59108734e+00 -2.68236309e-01 1.57050490e-01
2.14071229e-01 -4.03126210e-01 -1.46821484e-01 -3.09371948e-01]
|
[12.835681915283203, 7.986604690551758]
|
029a0d97-db7d-432a-8889-a360833459b5
|
using-neural-machine-translation-methods-for
| null | null |
https://aclanthology.org/2022.acl-srw.21
|
https://aclanthology.org/2022.acl-srw.21.pdf
|
Using Neural Machine Translation Methods for Sign Language Translation
|
We examine methods and techniques, proven to be helpful for the text-to-text translation of spoken languages in the context of gloss-to-text translation systems, where the glosses are the written representation of the signs. We present one of the first works that include experiments on both parallel corpora of the German Sign Language (PHOENIX14T and the Public DGS Corpus). We experiment with two NMT architectures with optimization of their hyperparameters, several tokenization methods and two data augmentation techniques (back-translation and paraphrasing). Through our investigation we achieve a substantial improvement of 5.0 and 2.2 BLEU scores for the models trained on the two corpora respectively. Our RNN models outperform our Transformer models, and the segmentation method we achieve best results with is BPE, whereas back-translation and paraphrasing lead to minor but not significant improvements.
|
['Sebastian Möller', 'Eleftherios Avramidis', 'Galina Angelova']
| null | null | null | null |
acl-2022-5
|
['sign-language-translation']
|
['computer-vision']
|
[ 4.24758077e-01 2.58645833e-01 -1.93344057e-02 -5.32871187e-01
-1.19962680e+00 -6.88776374e-01 1.08479321e+00 -4.01020169e-01
-6.81614876e-01 8.10509801e-01 6.92140639e-01 -4.30395722e-01
3.53911370e-01 -1.78794667e-01 -5.09734273e-01 -7.04827785e-01
3.44173461e-01 1.20610642e+00 2.01259568e-01 -4.34259534e-01
1.65843338e-01 4.70068604e-02 -1.03869140e+00 5.24561584e-01
8.40256274e-01 5.16439736e-01 -3.37303467e-02 9.19748783e-01
-4.10232216e-01 5.40128469e-01 -6.86730444e-01 -4.41292197e-01
3.92419010e-01 -5.24911582e-01 -1.07740009e+00 -1.73805207e-01
8.47875655e-01 -1.11250505e-01 -1.43685088e-01 7.22267210e-01
6.40693724e-01 1.31713878e-02 8.52154732e-01 -1.05411661e+00
-6.65695250e-01 7.57566154e-01 -3.78196836e-01 -9.20843631e-02
2.39535064e-01 -6.05159700e-02 9.05586660e-01 -6.85341060e-01
7.31442571e-01 1.28351665e+00 4.93000805e-01 1.01547456e+00
-8.74599278e-01 -4.81156319e-01 -2.71863937e-01 -3.13077681e-02
-9.95349348e-01 -7.06520736e-01 2.27305964e-01 -7.55827427e-02
1.40889239e+00 1.30610526e-01 6.14839613e-01 1.11975050e+00
4.49535996e-02 1.03400993e+00 1.68612063e+00 -9.13077950e-01
-5.05946465e-02 1.88897029e-01 3.29573750e-02 6.73156977e-01
-2.45678365e-01 -1.81011543e-01 -7.52120614e-01 1.05582237e-01
6.34103656e-01 -5.67208290e-01 -6.73293322e-02 2.88696766e-01
-1.35152125e+00 5.68648458e-01 -2.06376821e-01 5.92323482e-01
-3.29230845e-01 2.74078369e-01 6.56601667e-01 6.38281882e-01
3.87648761e-01 1.63274974e-01 -7.35206008e-01 -4.94887829e-01
-1.01523745e+00 2.76076645e-01 1.05116868e+00 1.28962576e+00
1.44669309e-01 2.43439317e-01 -3.29947114e-01 1.08801317e+00
4.08427685e-01 9.30793226e-01 8.11961114e-01 -6.68158114e-01
9.06636238e-01 1.70099914e-01 1.25704795e-01 2.62211636e-02
-1.48745477e-01 2.71708835e-02 -2.77090400e-01 -1.54155597e-01
7.25202620e-01 -5.09518027e-01 -1.78044844e+00 1.64873338e+00
-8.54873564e-03 -3.00123423e-01 4.90185618e-01 7.22728312e-01
6.89770222e-01 7.74029195e-01 1.94997191e-01 -6.88738897e-02
1.41065180e+00 -1.35610402e+00 -8.88044119e-01 -3.57788682e-01
8.36866915e-01 -1.47995305e+00 1.15263534e+00 4.52646732e-01
-1.28225756e+00 -1.62276581e-01 -7.36761391e-01 -1.80828810e-01
-4.77838218e-01 3.80370975e-01 2.28817150e-01 5.89776099e-01
-1.25506139e+00 4.24398631e-01 -8.71609151e-01 -9.56426322e-01
-7.97756016e-02 6.29112005e-01 -3.47423285e-01 5.56990430e-02
-7.80279696e-01 1.37680054e+00 6.33346736e-02 5.49412742e-02
-6.48995876e-01 -1.43557623e-01 -4.39902365e-01 -5.03603816e-01
-4.92111295e-02 -5.11461079e-01 1.63061595e+00 -1.07180917e+00
-2.19791746e+00 1.27733850e+00 -3.15840304e-01 -6.02598011e-01
5.67457259e-01 -3.82540762e-01 -3.14546287e-01 -9.81032029e-02
-3.12081184e-02 9.12288666e-01 6.02159381e-01 -7.70126879e-01
-7.52888143e-01 -2.91136712e-01 -4.78752911e-01 4.49936688e-01
4.96277772e-02 6.35866106e-01 -2.88699299e-01 -5.39289117e-01
5.10027632e-02 -1.15952063e+00 3.58238928e-02 -5.02279520e-01
-3.19349080e-01 -2.45562747e-01 6.32104814e-01 -1.25669527e+00
7.69493103e-01 -1.68669772e+00 3.49081129e-01 -3.07149589e-02
-4.61220294e-01 5.37964046e-01 -2.93479264e-01 7.36216307e-01
3.20381671e-01 3.42955031e-02 -3.12953532e-01 -6.14086688e-01
1.99670941e-01 8.76915336e-01 -3.53661478e-01 1.45040408e-01
-6.19592443e-02 1.03502369e+00 -5.00418425e-01 -6.12383366e-01
1.35537997e-01 3.78079474e-01 -1.08491853e-01 1.43478155e-01
-3.44566524e-01 3.33195925e-01 -2.47957021e-01 5.96325219e-01
2.93390214e-01 4.96021777e-01 1.64986625e-01 -1.97370742e-02
-2.69959331e-01 7.99190938e-01 -4.98504549e-01 1.93222082e+00
-5.84374964e-01 8.52922499e-01 -4.13662614e-03 -6.31252050e-01
9.34264600e-01 8.58023345e-01 5.72471023e-02 -7.16710985e-01
4.23671454e-01 7.06618130e-01 1.01802006e-01 -7.61323214e-01
4.82922792e-01 -5.19642711e-01 6.66861311e-02 5.61324120e-01
2.16812849e-01 -3.21714461e-01 3.39659870e-01 -6.48636892e-02
7.40761638e-01 5.27240396e-01 -8.47189948e-02 -3.34102303e-01
3.42436314e-01 3.18636626e-01 2.52921462e-01 4.15456474e-01
-3.97554114e-02 7.33038485e-01 1.65913224e-01 -1.07530117e-01
-1.40139115e+00 -8.03633273e-01 2.96988428e-01 1.11088097e+00
-6.04016185e-01 -6.16716444e-02 -1.04641855e+00 -6.44245327e-01
-5.19858778e-01 9.97409761e-01 -3.55646610e-01 3.80264372e-01
-1.07430398e+00 -7.65008986e-01 1.21374083e+00 7.57169008e-01
5.82067132e-01 -1.20459104e+00 -2.91508108e-01 2.90419906e-01
-3.73145640e-01 -1.30611145e+00 -7.25392580e-01 1.22502580e-01
-1.02385199e+00 -6.02104187e-01 -1.12161183e+00 -1.30655539e+00
4.23543632e-01 -3.51010203e-01 8.67227852e-01 -1.67452574e-01
1.14210740e-01 1.36182934e-01 -5.37857234e-01 -5.54291725e-01
-8.41475785e-01 2.23546132e-01 -1.01041488e-01 -3.99108171e-01
6.30176246e-01 -5.86269796e-01 -1.02133498e-01 1.71281174e-01
-6.15442812e-01 2.64852494e-01 8.52729440e-01 8.37532282e-01
2.42059544e-01 -1.05036187e+00 2.04772100e-01 -7.68538475e-01
7.51262665e-01 2.40424141e-01 -3.47965717e-01 4.84147221e-01
-7.24356234e-01 4.79657859e-01 3.55120331e-01 -3.64008576e-01
-1.14102042e+00 -1.83435865e-02 -3.16796631e-01 8.27028453e-02
-3.89239937e-01 1.95783257e-01 1.68156430e-01 -6.22894242e-02
4.97109234e-01 5.09402931e-01 -1.39956372e-02 -7.07181573e-01
5.44170976e-01 1.10859847e+00 5.11599422e-01 -6.93783462e-01
5.44095755e-01 2.28449017e-01 -2.40820691e-01 -7.87948012e-01
-4.28532839e-01 -4.96177077e-01 -8.73648584e-01 1.18730441e-02
9.45667863e-01 -5.73832691e-01 -1.15503505e-01 6.23483181e-01
-1.35310757e+00 -5.16550899e-01 -2.74148911e-01 6.72378719e-01
-7.91966259e-01 3.54349822e-01 -9.36285377e-01 -7.04412818e-01
-8.84426057e-01 -9.05299067e-01 1.13558793e+00 1.09139442e-01
-3.69589120e-01 -8.68811548e-01 4.31720734e-01 6.92141414e-01
5.04515350e-01 -4.51195389e-02 9.85595167e-01 -9.34097946e-01
-2.10322782e-01 -7.18826279e-02 7.98912160e-03 5.37682712e-01
9.10763517e-02 6.96104392e-02 -9.75025415e-01 -8.62594992e-02
-3.81630450e-01 -4.85743463e-01 5.92931390e-01 2.47685984e-01
-1.41396046e-01 -5.15497863e-01 -3.16294283e-02 2.25013494e-01
1.25469565e+00 3.99231374e-01 7.05190301e-01 2.35898182e-01
5.26863515e-01 6.47550464e-01 4.99806851e-01 -1.85925141e-01
3.11975658e-01 6.94344699e-01 -2.67287225e-01 -3.84737700e-02
-6.32814050e-01 -4.07984912e-01 6.92384005e-01 1.52879000e+00
-3.61890823e-01 -3.93377095e-01 -9.33853567e-01 9.24389839e-01
-1.79456258e+00 -5.67546248e-01 -3.46966326e-01 2.07264447e+00
1.04790628e+00 -1.06864445e-01 1.75130069e-01 -3.74223590e-02
3.77846092e-01 -1.07138660e-02 1.55107811e-01 -9.63701725e-01
-1.22533523e-01 5.21846712e-01 6.68271363e-01 8.67314458e-01
-7.09045649e-01 1.64604259e+00 7.13807297e+00 6.35695755e-01
-1.16792870e+00 3.12090397e-01 6.77747978e-03 -3.75639759e-02
-9.42877904e-02 7.01341331e-02 -8.75786006e-01 1.16687499e-01
1.29730594e+00 2.78720170e-01 6.31890178e-01 3.77473027e-01
3.26855302e-01 1.26590237e-01 -1.09585023e+00 5.01538575e-01
3.35547745e-01 -9.02496517e-01 2.61008233e-01 -2.02921610e-02
8.27403963e-01 5.19418716e-01 -2.21493706e-01 2.65002280e-01
4.66025531e-01 -8.32017064e-01 8.90655339e-01 2.28832036e-01
7.48626709e-01 -4.06014591e-01 9.17320430e-01 2.15828121e-02
-8.53277743e-01 4.47452247e-01 -1.61584213e-01 7.74754733e-02
4.10578579e-01 -2.39597768e-01 -1.27145493e+00 2.82012820e-01
2.70485640e-01 2.21080571e-01 -2.18917936e-01 8.83678854e-01
-6.67488456e-01 1.08211243e+00 -6.36477828e-01 -4.39226955e-01
6.79914236e-01 -3.14429104e-01 6.37188196e-01 1.49382901e+00
2.46325940e-01 -1.38164535e-01 -2.15487778e-01 3.98128241e-01
7.54569173e-02 5.22486269e-01 -4.12229925e-01 -1.68092042e-01
5.47605529e-02 8.85739326e-01 -4.65467155e-01 -7.08474040e-01
-3.61120194e-01 1.34246254e+00 -1.42391846e-01 5.69083154e-01
-5.39038479e-01 -4.54833388e-01 3.29137385e-01 -2.95207445e-02
1.83415115e-01 -3.12129617e-01 -1.29570529e-01 -9.88606513e-01
1.61376208e-01 -1.21509671e+00 6.82689026e-02 -7.87168741e-01
-1.05163145e+00 8.18775952e-01 1.36503622e-01 -9.30945158e-01
-7.28935778e-01 -8.14844429e-01 -5.31558633e-01 9.64565694e-01
-1.40651655e+00 -1.77129364e+00 1.91813678e-01 4.30962205e-01
7.91595638e-01 -3.39498639e-01 9.82720613e-01 5.46045005e-01
-1.89887747e-01 6.01002276e-01 3.04557025e-01 3.54997694e-01
9.22958016e-01 -1.27585638e+00 7.18615234e-01 8.06437075e-01
4.53171670e-01 5.86431384e-01 8.74415338e-01 -5.59296250e-01
-1.08242095e+00 -6.64098382e-01 1.56514084e+00 -3.87712270e-01
6.39917850e-01 -2.95267195e-01 -5.03531098e-01 1.04130983e+00
7.93773890e-01 -8.36786211e-01 3.71729702e-01 -1.14435539e-01
-1.70983315e-01 1.25350446e-01 -1.09371579e+00 6.59899712e-01
8.11557710e-01 -4.48087573e-01 -1.05605185e+00 5.47986090e-01
4.15502071e-01 -3.91249418e-01 -7.36265898e-01 3.81296389e-02
8.51226985e-01 -3.16033185e-01 3.11333776e-01 -8.71573865e-01
2.02379063e-01 -1.91084400e-01 -2.47881398e-01 -1.18886244e+00
2.83772051e-01 -9.25661206e-01 3.85713279e-01 1.36234999e+00
8.78520131e-01 -5.84905803e-01 8.96017373e-01 6.02885187e-01
-4.00983185e-01 -4.05984461e-01 -1.13547850e+00 -6.99902296e-01
3.84054005e-01 -1.34064883e-01 1.63240582e-01 6.12820089e-01
1.04417272e-01 6.59258246e-01 -6.35789752e-01 -2.86472619e-01
4.72692281e-01 -5.31135574e-02 7.18712449e-01 -7.67552614e-01
-2.84930706e-01 -4.02439237e-01 -1.26806512e-01 -1.35179937e+00
-1.56825781e-01 -7.06822813e-01 4.17753696e-01 -1.78165150e+00
-6.31268322e-02 7.16610486e-03 3.04411829e-01 8.36813509e-01
3.48383486e-01 1.98265672e-01 1.04414769e-01 1.40330508e-01
-1.08806686e-02 3.74413818e-01 1.15971828e+00 -1.49371758e-01
-2.54586250e-01 3.94972004e-02 -4.09265533e-02 6.90144122e-01
8.85991395e-01 -6.10100567e-01 -7.06413835e-02 -9.97133017e-01
-2.14207917e-01 -1.22011162e-01 -8.60766619e-02 -6.21688306e-01
1.76214218e-01 1.61759295e-02 -2.48859942e-01 -4.96061623e-01
3.68089557e-01 -5.75734019e-01 -2.40740567e-01 4.97667789e-01
-4.02116477e-01 2.87236482e-01 3.62885535e-01 7.43727386e-02
-2.54874319e-01 -3.79270881e-01 8.23423147e-01 -3.40406209e-01
-3.66314143e-01 -1.31837577e-01 -9.58409131e-01 1.15454324e-01
3.58374119e-01 -2.30333596e-01 -2.38272518e-01 -4.06539828e-01
-7.48036206e-01 1.06229991e-01 2.03445822e-01 4.02614325e-01
3.75439972e-01 -1.11298645e+00 -9.22696948e-01 1.62405670e-01
-2.52032846e-01 -3.31442237e-01 -4.74781722e-01 1.01913965e+00
-9.40191805e-01 8.41260612e-01 -4.14302647e-01 -2.66643316e-01
-1.72323501e+00 -2.18329877e-01 4.67419356e-01 -3.62242222e-01
-3.92448395e-01 8.09009910e-01 -4.62075889e-01 -7.08371878e-01
4.27704573e-01 -8.45036328e-01 -2.56997207e-03 -3.04106504e-01
1.63370520e-01 5.17995775e-01 8.86802748e-02 -7.49873340e-01
-2.45692655e-01 9.34618831e-01 -2.86829412e-01 -1.05754733e+00
1.27024579e+00 4.49763052e-03 -9.60501432e-02 4.24243420e-01
8.86907279e-01 -3.32783982e-02 -5.73280334e-01 -2.57895440e-01
1.58139944e-01 7.75040314e-02 -1.68887332e-01 -1.28379071e+00
-8.43662500e-01 1.11815560e+00 6.11154318e-01 -3.33195299e-01
9.62896228e-01 -2.62686461e-01 1.19718874e+00 7.10789800e-01
2.79411614e-01 -1.42719996e+00 -3.35825473e-01 7.65365183e-01
8.80572021e-01 -8.55019212e-01 -4.12534684e-01 -8.29749033e-02
-8.18780601e-01 1.02031493e+00 3.93546224e-01 -3.80351208e-02
1.12375438e-01 5.04186809e-01 6.71472311e-01 -3.96070927e-02
-7.47067750e-01 -2.18696848e-01 3.30202818e-01 5.97934544e-01
7.04007983e-01 7.85872489e-02 -9.04763937e-01 8.73920843e-02
-3.47634792e-01 2.86093708e-02 3.16238880e-01 9.54324186e-01
-4.77329344e-01 -1.58417296e+00 -3.63118201e-01 1.41409174e-01
-6.16602063e-01 -5.07505774e-01 -1.04036200e+00 1.06604600e+00
5.15803229e-03 8.25665772e-01 -2.33668864e-01 -3.26824248e-01
5.15159428e-01 7.42184699e-01 8.16688597e-01 -5.78958988e-01
-1.00589848e+00 4.49082792e-01 6.73531175e-01 -9.10771936e-02
-5.60058415e-01 -7.49048352e-01 -1.49879611e+00 -1.63101643e-01
-2.90250152e-01 3.42995375e-01 9.43686843e-01 1.31856024e+00
-1.15272343e-01 6.19863197e-02 -1.09064147e-01 -4.41133976e-01
-5.35482943e-01 -1.73455298e+00 -3.98900658e-01 2.12133318e-01
-3.16926697e-03 3.17837708e-02 -1.05275132e-01 4.67794150e-01]
|
[9.244807243347168, -6.571361064910889]
|
efdcc9b1-0d4e-45b2-ab8b-d08498e692fe
|
clamp-contrastive-language-music-pre-training
|
2304.11029
| null |
https://arxiv.org/abs/2304.11029v3
|
https://arxiv.org/pdf/2304.11029v3.pdf
|
CLaMP: Contrastive Language-Music Pre-training for Cross-Modal Symbolic Music Information Retrieval
|
We introduce CLaMP: Contrastive Language-Music Pre-training, which learns cross-modal representations between natural language and symbolic music using a music encoder and a text encoder trained jointly with a contrastive loss. To pre-train CLaMP, we collected a large dataset of 1.4 million music-text pairs. It employed text dropout as a data augmentation technique and bar patching to efficiently represent music data which reduces sequence length to less than 10%. In addition, we developed a masked music model pre-training objective to enhance the music encoder's comprehension of musical context and structure. CLaMP integrates textual information to enable semantic search and zero-shot classification for symbolic music, surpassing the capabilities of previous models. To support the evaluation of semantic search and music classification, we publicly release WikiMusicText (WikiMT), a dataset of 1010 lead sheets in ABC notation, each accompanied by a title, artist, genre, and description. In comparison to state-of-the-art models that require fine-tuning, zero-shot CLaMP demonstrated comparable or superior performance on score-oriented datasets. Our models and code are available at https://github.com/microsoft/muzic/tree/main/clamp.
|
['Maosong Sun', 'Xu Tan', 'Dingyao Yu', 'Shangda Wu']
|
2023-04-21
| null | null | null | null |
['music-classification', 'music-information-retrieval']
|
['music', 'music']
|
[ 3.99236411e-01 -2.12051958e-01 -2.33299434e-01 -1.98487461e-01
-1.21635377e+00 -8.89848411e-01 4.48842287e-01 4.79789414e-02
-5.11790633e-01 2.96432972e-01 3.78963023e-01 2.13910472e-02
-1.39800131e-01 -6.02622926e-01 -9.38336968e-01 -1.79603815e-01
2.03899503e-01 5.79839468e-01 -1.34744108e-01 -1.05164438e-01
4.31495994e-01 -3.37302059e-01 -1.73243284e+00 8.80771041e-01
6.31213307e-01 1.01192522e+00 3.71025950e-01 7.34332800e-01
-2.66446739e-01 9.07305717e-01 -5.57315588e-01 -7.06716120e-01
1.25924870e-01 -5.32348752e-01 -9.21070039e-01 -5.10274947e-01
6.48136973e-01 -1.82798132e-01 -3.91521245e-01 7.79547811e-01
7.96418190e-01 2.52265126e-01 4.92364258e-01 -1.01076221e+00
-8.99332047e-01 1.38180840e+00 -2.58166641e-01 -5.99295013e-02
3.50710809e-01 2.39207223e-01 1.59239936e+00 -8.66141498e-01
7.11324394e-01 8.83196115e-01 1.00682271e+00 6.35679483e-01
-1.44946849e+00 -1.00909805e+00 -2.38456875e-01 3.45155418e-01
-1.47256160e+00 -6.11733794e-01 8.69081140e-01 -4.79560524e-01
1.17408800e+00 3.00286323e-01 7.39725053e-01 1.59941983e+00
-4.57277030e-01 8.80638838e-01 5.09526193e-01 -3.67160827e-01
5.66074960e-02 -2.78598696e-01 -1.95443407e-01 5.75773537e-01
-3.61030340e-01 1.20236836e-01 -1.09923816e+00 -1.89798385e-01
7.68296778e-01 -2.78972179e-01 -5.30442744e-02 -2.68405020e-01
-1.47309673e+00 5.73703885e-01 3.52741748e-01 3.55774462e-01
-1.08070455e-01 5.09298205e-01 9.08404350e-01 3.13429058e-01
5.56503832e-02 9.29144740e-01 -3.60789865e-01 -7.11538017e-01
-1.23057282e+00 3.14411789e-01 6.02836132e-01 1.00829315e+00
3.14150989e-01 2.35722765e-01 -3.44840556e-01 1.23255730e+00
-2.13087201e-01 2.58158267e-01 5.59161901e-01 -1.26506960e+00
5.55222213e-01 4.70657080e-01 -3.14517230e-01 -5.05289137e-01
-3.51533704e-02 -8.02135885e-01 -5.57770193e-01 1.17473781e-01
2.92592794e-01 2.99767286e-01 -6.64765358e-01 2.01106763e+00
-4.78667855e-01 3.94290566e-01 -9.21212435e-02 9.21645105e-01
9.93750811e-01 3.68026316e-01 -1.75990798e-02 2.96134502e-01
1.31608021e+00 -1.03195810e+00 -4.85257030e-01 -1.21121602e-02
7.04109848e-01 -1.00651908e+00 1.93471062e+00 6.51938081e-01
-1.40168166e+00 -6.85301661e-01 -1.21053612e+00 -5.30840218e-01
-3.23168546e-01 2.04759210e-01 4.71777052e-01 1.36381909e-01
-6.27406478e-01 1.05525911e+00 -6.45380497e-01 -4.14394215e-02
6.42602921e-01 1.81546614e-01 -8.74982551e-02 1.39674112e-01
-1.07279074e+00 4.33523923e-01 5.52778363e-01 -4.40394014e-01
-9.76658583e-01 -1.22553158e+00 -7.33261287e-01 1.99102208e-01
3.82471263e-01 -7.52441823e-01 1.65959620e+00 -9.96274173e-01
-1.48397553e+00 1.11977649e+00 2.04751417e-01 -4.39714372e-01
2.28495747e-01 -3.45065862e-01 -3.56919974e-01 3.07424702e-02
1.76194966e-01 7.82730103e-01 6.25606298e-01 -6.90680981e-01
-1.66789055e-01 -5.88462539e-02 -5.98546192e-02 2.46493831e-01
-2.37375289e-01 5.51337749e-02 -6.28011167e-01 -1.31070220e+00
-3.71761739e-01 -7.19969571e-01 2.72612333e-01 -9.40835699e-02
-4.89175946e-01 3.57417092e-02 3.08399707e-01 -6.15680635e-01
1.47500229e+00 -2.50427246e+00 3.53155106e-01 -1.53353021e-01
-1.63565546e-01 -3.66297029e-02 -5.86851358e-01 6.01808906e-01
-1.66296482e-01 1.17225468e-01 -5.18436790e-01 -5.91742158e-01
4.06567395e-01 2.18120273e-02 -4.48510855e-01 -1.07917031e-02
1.95509978e-02 1.12445521e+00 -8.60225320e-01 -3.53980601e-01
3.46815139e-02 4.80184734e-01 -1.02499807e+00 -2.24885028e-02
-6.64622128e-01 3.22613835e-01 1.47675812e-01 7.20831394e-01
9.39744338e-03 -3.55649889e-01 -7.34326104e-03 -1.70597494e-01
-1.38551921e-01 1.02398610e+00 -1.05941617e+00 2.87508249e+00
-5.01386464e-01 6.07949495e-01 -2.66885906e-01 -6.52683139e-01
7.50641704e-01 4.06346649e-01 5.24621487e-01 -8.63480151e-01
1.07150495e-01 3.79927635e-01 -1.93676814e-01 -2.29570255e-01
5.89224339e-01 -1.43555820e-01 -4.12980556e-01 5.58650255e-01
4.17773485e-01 -1.73598304e-01 3.44670147e-01 1.74169108e-01
1.16730785e+00 4.92289901e-01 -6.72649443e-02 1.00711197e-01
9.57773253e-02 9.33428779e-02 5.00025213e-01 7.01654673e-01
3.97206098e-01 8.37571323e-01 3.64951640e-01 -1.90662280e-01
-1.23475111e+00 -1.18282068e+00 -7.35721439e-02 1.70342982e+00
-2.71882892e-01 -1.13684022e+00 -5.33393443e-01 -2.71900266e-01
8.79477412e-02 9.50218797e-01 -5.10566473e-01 -2.85555989e-01
-5.71244895e-01 -1.61635175e-01 1.14800060e+00 5.22718668e-01
2.56008774e-01 -1.42319000e+00 -4.70939666e-01 2.81490356e-01
-3.67181569e-01 -6.34007692e-01 -8.64733517e-01 5.29802144e-01
-6.01205170e-01 -1.13637030e+00 -6.29541814e-01 -8.82839084e-01
-8.53283927e-02 -3.69259417e-01 1.59080315e+00 -9.04418826e-02
-5.89611888e-01 9.25952718e-02 -3.56527895e-01 -3.52688491e-01
-2.67649144e-01 6.37446463e-01 -1.61122441e-01 -5.55309713e-01
2.15075478e-01 -8.69672298e-01 -5.46310723e-01 2.39227116e-02
-8.01665246e-01 2.75501132e-01 4.75309432e-01 8.67394090e-01
7.13751972e-01 -4.01591003e-01 4.72478122e-01 -6.02429688e-01
5.53096235e-01 -2.20817924e-01 -4.17463124e-01 2.03137081e-02
-3.91673446e-01 1.18776493e-01 4.15948808e-01 -6.31857753e-01
-5.37157476e-01 -4.68126871e-02 -2.90191263e-01 -7.86580086e-01
5.56526035e-02 5.73507249e-01 7.72960261e-02 4.08395261e-01
7.87782431e-01 2.01410666e-01 -8.15582797e-02 -9.34463739e-01
6.09709144e-01 6.52103782e-01 1.17238784e+00 -7.30922699e-01
4.53142107e-01 5.61552681e-02 -3.34976792e-01 -3.92622679e-01
-1.03428078e+00 -3.78155738e-01 -4.85266775e-01 1.60069123e-01
7.34888673e-01 -1.00057125e+00 -8.68662655e-01 1.66463614e-01
-8.74293923e-01 -7.27680087e-01 -7.73107588e-01 4.37178671e-01
-1.06778347e+00 2.15849318e-02 -8.52512240e-01 -4.65332270e-01
-5.90483129e-01 -7.22024977e-01 1.13589764e+00 -3.08347732e-01
-8.05787325e-01 -5.93664050e-01 3.00097644e-01 3.89543772e-01
2.78011471e-01 -3.12592983e-02 1.06026840e+00 -5.45837760e-01
-4.39671278e-01 1.41338751e-01 6.72748163e-02 3.03325832e-01
-2.28595465e-01 -1.94826096e-01 -1.08962762e+00 -9.12698060e-02
-4.86071110e-01 -8.25273693e-01 1.09002340e+00 1.80466101e-01
1.49842477e+00 -3.32474828e-01 9.22474414e-02 1.17767203e+00
1.23340607e+00 -6.67150738e-03 5.49034357e-01 6.03563547e-01
5.82986951e-01 1.16183758e-01 2.83449501e-01 6.12772644e-01
5.13883457e-02 9.96915221e-01 3.48547935e-01 2.52692461e-01
-4.45084304e-01 -7.94709504e-01 3.68297994e-01 9.87432122e-01
6.10907674e-02 5.49912117e-02 -8.84801209e-01 4.76461381e-01
-1.74054575e+00 -1.31043816e+00 3.63542765e-01 2.12475491e+00
1.31274915e+00 1.50440231e-01 2.98561841e-01 3.99648964e-01
2.77061075e-01 1.84952021e-01 -6.80334926e-01 -2.72494078e-01
-2.93936282e-01 6.19248450e-01 2.10153922e-01 3.04467231e-01
-1.01359141e+00 1.15579605e+00 6.20117998e+00 1.01901340e+00
-9.71170306e-01 2.36147225e-01 -3.75621505e-02 -9.08027053e-01
-4.04260367e-01 -1.01598449e-01 -4.04418647e-01 5.85773170e-01
9.38301921e-01 -1.22365288e-01 9.11647379e-01 7.70057321e-01
-8.11080337e-02 4.52284157e-01 -1.18379200e+00 1.28133738e+00
5.40613756e-02 -1.76290333e+00 1.62732840e-01 -1.20508760e-01
5.07318020e-01 3.09725195e-01 2.39639461e-01 7.28424847e-01
1.82203591e-01 -1.15480554e+00 1.12646699e+00 6.65728152e-01
1.29368031e+00 -7.49158204e-01 3.00810188e-01 3.63687947e-02
-1.21750724e+00 -1.42658725e-01 -8.00445080e-02 -2.36987755e-01
6.23048544e-02 4.77856137e-02 -4.47096884e-01 4.06957477e-01
6.85593605e-01 1.07284045e+00 -4.81467247e-01 9.26242113e-01
-1.46988153e-01 6.60921872e-01 -1.96313843e-01 1.90703958e-01
1.10174984e-01 4.09213640e-02 6.18585944e-01 1.23837698e+00
3.50163907e-01 -2.20920578e-01 3.41764875e-02 1.29077399e+00
-3.93128633e-01 1.99691117e-01 -3.77103984e-01 -5.61886787e-01
6.71665072e-01 7.69640982e-01 -2.06188500e-01 -4.33145612e-01
-3.03763568e-01 1.22208548e+00 3.97634387e-01 2.62021780e-01
-8.70045602e-01 -6.98270023e-01 8.19117665e-01 2.11852282e-01
4.03026938e-01 -8.19712579e-02 -6.54184759e-01 -1.17037880e+00
-5.22350483e-02 -9.99034464e-01 5.30293643e-01 -1.06027782e+00
-1.19414234e+00 3.71712953e-01 -3.15263659e-01 -1.25847089e+00
-3.92281741e-01 -3.72120827e-01 -5.37991583e-01 7.64758587e-01
-9.58740115e-01 -1.25701666e+00 -4.84720916e-02 5.26148081e-01
7.00941503e-01 -5.84602058e-01 1.20020854e+00 5.89062393e-01
-2.69735515e-01 9.02099371e-01 -1.05605461e-01 4.24337685e-01
9.06101644e-01 -1.17027533e+00 5.66697776e-01 1.89339519e-01
7.51894593e-01 7.24565566e-01 4.00474817e-01 -5.21748602e-01
-1.23694777e+00 -8.72644067e-01 7.84148335e-01 -5.27407646e-01
8.20288539e-01 -6.48884237e-01 -8.85664642e-01 6.49901509e-01
2.65572935e-01 -4.13475484e-01 1.10311425e+00 5.18026173e-01
-8.17195535e-01 5.91923632e-02 -3.83864045e-01 6.78318322e-01
1.43137550e+00 -9.92412388e-01 -9.37469125e-01 -2.18541548e-02
8.29845607e-01 -3.55392337e-01 -8.45812738e-01 3.77695531e-01
8.37763786e-01 -6.93618774e-01 9.66532648e-01 -6.17973804e-01
6.55999839e-01 -1.66995808e-01 -3.96632403e-01 -1.03241694e+00
-4.41408277e-01 -7.31617212e-01 -1.26575261e-01 1.20826828e+00
5.43776453e-01 1.98393866e-01 7.82453418e-01 -2.27261677e-01
-5.60917079e-01 -5.80359459e-01 -7.88002908e-01 -1.02260268e+00
1.40345216e-01 -8.89243305e-01 5.03916800e-01 1.07158852e+00
3.00257832e-01 5.06207764e-01 -3.55019063e-01 -4.82688159e-01
5.35331905e-01 2.32549861e-01 6.77912354e-01 -1.14895713e+00
-7.62058854e-01 -9.99047637e-01 -1.90885812e-01 -7.01837301e-01
2.86518663e-01 -1.66144371e+00 -1.95196390e-01 -1.42658353e+00
2.96970814e-01 -2.14181647e-01 -6.48110151e-01 8.96346629e-01
2.60089487e-01 6.56842351e-01 3.15993965e-01 3.11986715e-01
-6.43076122e-01 7.26440907e-01 8.69084358e-01 -3.37778360e-01
-2.83512801e-01 -2.41632938e-01 -6.22151673e-01 5.65483212e-01
9.79915023e-01 -5.51326752e-01 -3.33046049e-01 -7.19163716e-01
6.26704574e-01 -2.43222430e-01 6.12807214e-01 -1.22649121e+00
2.17731267e-01 2.70931840e-01 1.83457673e-01 -7.18698084e-01
7.22437918e-01 -2.99122781e-01 2.31398642e-01 2.42288515e-01
-9.45751965e-01 -1.77773595e-01 4.52637404e-01 1.63129136e-01
-3.18042547e-01 -2.38855377e-01 5.55002630e-01 -1.72782183e-01
-5.59325814e-01 1.08665772e-01 -9.53644663e-02 4.40378457e-01
3.56186748e-01 -6.46307617e-02 -2.57014543e-01 -2.10162535e-01
-8.59645665e-01 -9.73998308e-02 6.06924891e-01 6.45682991e-01
4.63239700e-01 -1.64079809e+00 -6.66145742e-01 2.60407656e-01
5.06339788e-01 -3.23511750e-01 1.82485640e-01 6.46703899e-01
-3.07417870e-01 2.72495747e-01 -2.49272063e-01 -3.22183043e-01
-1.18759072e+00 5.58426917e-01 1.50378391e-01 9.24112797e-02
-8.86937916e-01 9.07663524e-01 -1.31895334e-01 -6.02114558e-01
7.38733888e-01 -4.25236940e-01 3.28512698e-01 -6.62138462e-02
5.10862708e-01 1.81016549e-01 -1.62521392e-01 -1.73539937e-01
-1.67374671e-01 5.59896946e-01 2.22592264e-01 -6.11349642e-01
1.37978232e+00 3.01474810e-01 6.96490407e-02 9.91952360e-01
1.16934216e+00 8.59021097e-02 -1.04467511e+00 -2.47717142e-01
3.12262625e-01 -1.81582510e-01 -5.65474108e-02 -1.23452532e+00
-6.99900806e-01 9.03537452e-01 2.88666725e-01 -3.45027536e-01
9.59384143e-01 2.41854623e-01 8.57456326e-01 4.32093114e-01
1.23628024e-02 -1.18054342e+00 3.93992931e-01 8.26305687e-01
1.18763328e+00 -8.69256675e-01 -4.23241705e-01 1.10809922e-01
-6.60726488e-01 8.89227629e-01 3.90002191e-01 -9.45008248e-02
3.82486403e-01 5.23707747e-01 -9.62596536e-02 -1.07436962e-01
-1.03060186e+00 -3.31130445e-01 5.22251070e-01 2.08270699e-01
8.02485228e-01 -6.20037280e-02 1.23778194e-01 1.29799712e+00
-9.70278263e-01 3.33965063e-01 -5.20904176e-02 8.16925883e-01
-2.73922503e-01 -1.16321373e+00 2.68744770e-03 2.92384654e-01
-5.72713673e-01 -5.97350717e-01 -5.67075014e-01 3.82859498e-01
2.45929956e-01 4.88646626e-01 3.31394792e-01 -5.74881792e-01
4.91431147e-01 4.98992771e-01 4.86341327e-01 -7.60097146e-01
-8.31513166e-01 2.03921303e-01 2.77496368e-01 -6.54355288e-01
4.86956500e-02 -5.47173142e-01 -1.29024971e+00 -1.63856819e-01
1.85620561e-01 4.81467061e-02 6.98348522e-01 4.07351911e-01
5.01320958e-01 8.16374302e-01 2.12442726e-02 -7.84584105e-01
-4.07900035e-01 -1.14862788e+00 -5.49161971e-01 5.11193514e-01
2.53623333e-02 -5.36245644e-01 -2.04562232e-01 1.67940497e-01]
|
[15.841741561889648, 5.318155765533447]
|
650e1ba6-9d39-4268-9503-5a266b326ce0
|
conda-a-contextual-dual-annotated-dataset-for
|
2106.06213
| null |
https://arxiv.org/abs/2106.06213v2
|
https://arxiv.org/pdf/2106.06213v2.pdf
|
CONDA: a CONtextual Dual-Annotated dataset for in-game toxicity understanding and detection
|
Traditional toxicity detection models have focused on the single utterance level without deeper understanding of context. We introduce CONDA, a new dataset for in-game toxic language detection enabling joint intent classification and slot filling analysis, which is the core task of Natural Language Understanding (NLU). The dataset consists of 45K utterances from 12K conversations from the chat logs of 1.9K completed Dota 2 matches. We propose a robust dual semantic-level toxicity framework, which handles utterance and token-level patterns, and rich contextual chatting history. Accompanying the dataset is a thorough in-game toxicity analysis, which provides comprehensive understanding of context at utterance, token, and dual levels. Inspired by NLU, we also apply its metrics to the toxicity detection tasks for assessing toxicity and game-specific aspects. We evaluate strong NLU models on CONDA, providing fine-grained results for different intent classes and slot classes. Furthermore, we examine the coverage of toxicity nature in our dataset by comparing it with other toxicity datasets.
|
['Soyeon Caren Han', 'Josiah Poon', 'Siqu Long', 'Xinghong Guo', 'Kunze Wang', 'Tongshu Zhang', 'Jean Lee', 'Guanghao Huang', 'Henry Weld']
|
2021-06-11
| null |
https://aclanthology.org/2021.findings-acl.213
|
https://aclanthology.org/2021.findings-acl.213.pdf
|
findings-acl-2021-8
|
['dota-2']
|
['playing-games']
|
[ 6.21724725e-02 -4.09041584e-01 -2.14075133e-01 -8.79996866e-02
-1.34201813e+00 -1.08820558e+00 5.95930994e-01 4.07820344e-01
-3.26829523e-01 7.69725740e-01 8.36163461e-01 -2.70512164e-01
-2.96862006e-01 -7.26137102e-01 -2.91059881e-01 -2.97353208e-01
-1.90420657e-01 3.46116513e-01 3.60489249e-01 -4.76549745e-01
1.37315810e-01 -1.51715681e-01 -1.40945864e+00 9.99076426e-01
4.68457788e-01 8.32813561e-01 7.08063170e-02 9.00793135e-01
-2.42709801e-01 1.76075613e+00 -1.08780730e+00 -4.54841137e-01
-1.67692125e-01 -5.78777492e-01 -1.34044385e+00 -1.97474435e-01
-1.59100190e-01 -3.16857100e-01 -4.90764081e-01 6.08746767e-01
6.85534596e-01 2.60064155e-01 6.38066292e-01 -1.34738708e+00
-1.87379420e-01 8.76091063e-01 5.34993783e-02 2.44158074e-01
1.21952438e+00 3.35655570e-01 1.37102664e+00 -2.64498442e-01
6.83869004e-01 1.22620285e+00 6.79135203e-01 8.88878405e-01
-7.13001668e-01 -5.22459388e-01 1.74031153e-01 4.18008059e-01
-1.19243145e+00 -1.82498470e-01 3.03429246e-01 -7.17635870e-01
1.44533658e+00 5.81518829e-01 2.36278549e-01 1.53234315e+00
-6.86441511e-02 1.01290894e+00 9.94045198e-01 1.76228359e-02
5.51369369e-01 -3.65952551e-01 2.26357624e-01 5.47851920e-01
-7.35595107e-01 2.08902452e-02 -9.84716952e-01 -5.95794261e-01
-1.17842928e-01 -1.09665975e-01 -5.66908382e-02 6.79386497e-01
-5.19955099e-01 1.03048396e+00 -2.10677594e-01 -3.22518311e-02
-1.29468128e-01 2.12881550e-01 1.16543388e+00 3.98196071e-01
7.30325222e-01 5.93182206e-01 -7.21997797e-01 -1.13289070e+00
-2.24751502e-01 4.86031443e-01 9.86133099e-01 1.00624168e+00
4.17821974e-01 -4.54468518e-01 -8.49029958e-01 1.20978999e+00
1.13769405e-01 -1.66992933e-01 4.72956508e-01 -7.63306201e-01
3.82033288e-01 7.67177820e-01 -7.92819336e-02 -3.78781825e-01
-7.82362163e-01 5.18401526e-02 6.72733188e-02 -1.73135117e-01
3.83890748e-01 -7.89945722e-02 -5.16261280e-01 2.06519413e+00
2.25747496e-01 1.95082948e-01 2.30675787e-02 5.02485633e-01
1.29786301e+00 4.35129017e-01 7.40778625e-01 -1.39702961e-01
1.73098278e+00 -6.73414409e-01 -7.40870059e-01 3.69622149e-02
1.46281207e+00 -5.24298549e-01 1.48436606e+00 5.41078866e-01
-5.00567079e-01 1.53762266e-01 -9.00906980e-01 -1.89120367e-01
-6.78273737e-01 -4.18728709e-01 8.20624888e-01 9.15447414e-01
-7.61441708e-01 4.35311407e-01 -4.93101776e-03 -4.59864467e-01
2.52974212e-01 1.90862224e-01 -1.58920363e-01 -1.41020685e-01
-1.85735202e+00 7.78858483e-01 2.46222749e-01 -6.84752405e-01
-1.40657628e+00 -1.11218214e+00 -1.08403170e+00 -1.59985781e-01
6.73214972e-01 -2.50470042e-01 1.75163531e+00 3.81579876e-01
-1.64204895e+00 6.53483391e-01 2.40894750e-01 -1.95517048e-01
3.79752249e-01 -9.97123793e-02 -3.52381617e-02 -1.88453808e-01
3.91213030e-01 3.21412146e-01 -5.85900247e-02 -4.21523988e-01
-5.53630233e-01 -2.63976246e-01 6.98498666e-01 4.45570946e-01
-3.08133662e-01 6.84823513e-01 -4.30888563e-01 -4.04220998e-01
-6.52664721e-01 -3.27218503e-01 -2.20316589e-01 -6.26354098e-01
-5.62265217e-01 -8.15659285e-01 4.95790273e-01 -5.19999504e-01
1.70268214e+00 -1.98226047e+00 -1.50250003e-01 -4.20343995e-01
4.93837953e-01 -5.00984550e-01 -3.60474378e-01 8.55969250e-01
-2.03858167e-01 1.69381842e-01 8.05215016e-02 -6.50719404e-01
5.25128424e-01 7.82381222e-02 -2.27869466e-01 1.65379554e-01
-5.26006147e-02 9.02821898e-01 -1.19446325e+00 -2.29811475e-01
1.70238331e-01 -2.20984817e-01 -8.69879842e-01 4.00651813e-01
-8.91375780e-01 1.14370227e-01 -5.78420341e-01 9.31328535e-01
2.91669101e-01 1.24193333e-01 4.84909154e-02 3.04860413e-01
-7.61239007e-02 9.30327356e-01 -4.26467448e-01 1.89765704e+00
-4.68745828e-01 2.78840810e-01 -3.11625332e-01 -4.42171246e-01
4.86085951e-01 5.01537919e-01 6.69012606e-01 -1.03492367e+00
3.52323830e-01 -7.07124099e-02 -2.52553195e-01 -7.20111668e-01
5.61934292e-01 -2.50767469e-01 -1.15547228e+00 6.79400265e-01
1.16562314e-01 -3.84473726e-02 3.79370719e-01 5.17058372e-01
2.13810372e+00 -2.46670514e-01 3.83007526e-01 -1.69458643e-01
2.17293411e-01 1.46686777e-01 3.57528985e-01 9.88692701e-01
-5.74122071e-01 4.32732552e-01 9.75359797e-01 -2.37626761e-01
-4.03505683e-01 -6.93967938e-01 -3.56152914e-02 1.87735939e+00
-1.89714983e-01 -1.28206313e+00 -1.06672680e+00 -8.22341561e-01
-3.03379625e-01 3.47773969e-01 -8.02955508e-01 -2.10783139e-01
2.12062865e-01 -1.00942564e+00 1.35479808e+00 3.85944188e-01
3.07598203e-01 -1.10368979e+00 -5.02295382e-02 5.15337348e-01
-6.48460150e-01 -1.30905223e+00 -5.75890839e-01 5.76656759e-01
-3.85847315e-02 -1.36318386e+00 -2.80878805e-02 -2.57805735e-01
-5.26714146e-01 -1.01936392e-01 1.22437370e+00 1.32396951e-01
-5.98127961e-01 3.37063938e-01 -7.50490367e-01 -1.99373379e-01
-5.99344850e-01 -1.47949457e-01 2.13349462e-01 -4.94571567e-01
7.24374175e-01 -5.98639131e-01 -2.45765168e-02 6.02876306e-01
-8.98672342e-01 -3.91773105e-01 -3.82043831e-02 5.88133037e-01
1.52688503e-01 6.59756288e-02 5.84192276e-01 -9.97778833e-01
1.19217527e+00 -9.61899459e-01 -9.47595537e-02 1.23762682e-01
1.89624116e-01 -3.60094070e-01 6.02856219e-01 -3.03297579e-01
-7.95105040e-01 -7.38819018e-02 -7.09950805e-01 9.68130082e-02
-5.77866316e-01 7.12228537e-01 -6.47866607e-01 1.73777252e-01
1.01749074e+00 2.82735884e-01 -4.66003358e-01 -4.65914756e-01
5.54911017e-01 1.07880247e+00 4.02728021e-01 -1.08594441e+00
8.76487717e-02 -3.24026532e-02 -5.74557424e-01 -9.10045803e-01
-8.76982629e-01 -1.00550163e+00 -2.09919706e-01 -4.40328300e-01
8.85571301e-01 -1.00694430e+00 -1.44205904e+00 7.59109080e-01
-1.25472832e+00 -8.42985570e-01 -7.52969757e-02 -1.91260111e-02
-8.66232157e-01 5.97286761e-01 -1.31797981e+00 -8.83713186e-01
-5.78843355e-02 -1.25664663e+00 1.17771518e+00 -2.42093042e-01
-9.01192248e-01 -8.44194949e-01 4.95965421e-01 7.47300208e-01
-1.67873085e-01 9.76378769e-02 1.15514123e+00 -8.96374762e-01
-1.23836361e-01 -8.25711861e-02 6.64027110e-02 -4.53075647e-01
2.94238836e-01 -4.24571633e-01 -1.34044576e+00 3.06996524e-01
3.38099934e-02 -1.07836485e+00 6.02290332e-01 1.46959931e-01
1.13838339e+00 -3.22537005e-01 5.03852442e-02 2.56715771e-02
8.73725235e-01 5.29549599e-01 7.47904956e-01 3.33968759e-01
4.48185176e-01 7.57545292e-01 6.56164825e-01 9.80877817e-01
4.35728490e-01 8.86948884e-01 4.80276316e-01 3.51821333e-01
8.98909941e-02 -5.29106617e-01 7.58939505e-01 6.58393204e-01
3.95658255e-01 -6.10846043e-01 -7.13143468e-01 5.50048709e-01
-1.90235984e+00 -9.65407491e-01 -1.02193266e-01 1.49681759e+00
1.20926583e+00 1.28731225e-02 6.42561734e-01 2.97438979e-01
3.29770952e-01 2.09440023e-01 -4.45415676e-01 -7.02439487e-01
-2.53152531e-02 6.37030452e-02 1.56189933e-01 8.60067546e-01
-1.14171231e+00 1.13256419e+00 7.09404230e+00 1.97544837e+00
-2.50529736e-01 3.97629797e-01 6.67713821e-01 -1.99593142e-01
-4.46914703e-01 -3.15433621e-01 -4.06397223e-01 6.36164010e-01
1.00949979e+00 -8.22623447e-02 4.40762550e-01 8.58217955e-01
3.35282236e-01 -3.13404649e-01 -1.32253897e+00 7.15517521e-01
-1.87919989e-01 -1.47409368e+00 -9.65554360e-03 1.62672505e-01
2.15833113e-01 9.54343006e-03 -7.58638605e-02 7.68483818e-01
8.69609714e-01 -1.30627847e+00 6.97435796e-01 -1.05743341e-01
9.71574783e-01 -6.82194829e-01 4.62299973e-01 5.44566095e-01
-1.32870686e+00 -2.93080688e-01 -2.51900166e-01 -7.80009449e-01
-9.51970369e-02 2.57675618e-01 -7.99225867e-01 5.18838048e-01
8.34259987e-01 8.50887239e-01 -2.55966246e-01 7.24587917e-01
-1.07125014e-01 5.64040244e-01 -2.97556251e-01 -5.88337421e-01
4.68036413e-01 1.40137196e-01 2.89917469e-01 1.36359763e+00
-2.80792713e-01 6.08969867e-01 3.36994797e-01 7.10259140e-01
-2.92650670e-01 3.02222580e-01 -7.44345069e-01 -3.28115761e-01
5.55747926e-01 9.38253701e-01 -5.64173043e-01 -2.34638125e-01
-3.02566379e-01 9.42411900e-01 3.87698203e-01 -2.53397822e-02
-9.68243122e-01 -1.74460709e-01 1.47863662e+00 -2.56019711e-01
-5.34666896e-01 2.43173733e-01 -3.24983776e-01 -6.82848573e-01
-2.07563713e-01 -8.75436902e-01 9.54795122e-01 -5.45425057e-01
-1.53467035e+00 5.10849833e-01 -6.34948239e-02 -1.25584745e+00
-8.11358020e-02 -7.98502326e-01 -1.02940512e+00 6.18540049e-01
-8.41897488e-01 -7.68599331e-01 1.32230157e-02 6.34090245e-01
1.06996059e+00 1.89500898e-02 1.13060331e+00 6.47692263e-01
-6.74278915e-01 8.06043983e-01 -2.61268795e-01 7.87347779e-02
5.74254990e-01 -1.29117262e+00 3.77489418e-01 3.26967180e-01
-2.38199115e-01 4.15267229e-01 6.10862255e-01 -8.00423384e-01
-1.03832030e+00 -1.08626342e+00 8.64928782e-01 -1.18898082e+00
1.05521512e+00 -9.47259963e-01 -4.07948196e-01 2.11094111e-01
-2.98692100e-02 -8.00517738e-01 1.44195020e+00 4.48394179e-01
-4.84922558e-01 7.46371090e-01 -1.11623573e+00 7.54008949e-01
1.47308397e+00 -1.25697696e+00 -2.67460734e-01 7.58323848e-01
1.28630590e+00 -4.92965698e-01 -8.72861445e-01 7.31362961e-03
2.36637846e-01 -7.10361540e-01 7.90515780e-01 -1.28405833e+00
7.88105607e-01 -1.22598976e-01 -4.48505372e-01 -1.09696734e+00
-3.24146330e-01 -7.51622021e-01 1.75646901e-01 1.08285964e+00
6.15223050e-01 9.36676934e-02 1.06011415e+00 8.02853465e-01
-4.17475700e-01 -7.74456561e-01 -1.08506632e+00 -8.17128003e-01
1.52902022e-01 -1.41352022e+00 6.57822013e-01 9.15717423e-01
1.25946057e+00 6.13580704e-01 -6.80005431e-01 -1.60080627e-01
3.18310767e-01 -3.04806918e-01 4.42186952e-01 -8.31162214e-01
-5.14888406e-01 -4.89372522e-01 -5.74489713e-01 -1.06905389e+00
3.01997662e-01 -8.41017127e-01 2.93737322e-01 -1.11970413e+00
6.86945617e-01 -4.91293222e-02 -3.68619375e-02 8.54783595e-01
-1.94601402e-01 3.40678900e-01 -1.48918092e-01 -1.00008093e-01
-1.16260386e+00 7.85498917e-01 6.42308950e-01 -5.20012021e-01
-1.22296639e-01 -1.69552252e-01 -9.96691465e-01 5.81999421e-01
7.06931829e-01 -5.07763326e-01 -7.07404017e-01 -5.67839183e-02
2.62495011e-01 2.44271174e-01 -1.49197876e-01 -8.19797575e-01
2.71735191e-01 -4.10221606e-01 -3.70489120e-01 -3.65716487e-01
5.21186769e-01 -7.14773536e-02 -1.77517548e-01 2.56486982e-01
-5.79946876e-01 -3.67421210e-01 5.18854737e-01 7.17938960e-01
-6.97439536e-02 -1.08777329e-01 2.84487814e-01 -2.56587863e-01
-9.83282328e-01 5.47427595e-01 -1.22466874e+00 4.51631039e-01
8.26403677e-01 -1.74086109e-01 -5.75993955e-01 -7.41038382e-01
-1.00019455e+00 5.12348592e-01 1.31744593e-01 5.61265349e-01
4.94352907e-01 -1.30522346e+00 -4.52881873e-01 -4.57980126e-01
9.00130808e-01 -5.29476285e-01 8.33515882e-01 4.71373767e-01
-2.89893121e-01 3.59979182e-01 1.44126505e-01 -1.97139874e-01
-1.30899024e+00 3.40245068e-01 4.51393604e-01 -5.36827266e-01
-6.69790804e-02 1.15270579e+00 6.90448642e-01 -9.12375569e-01
4.96527553e-01 -3.11284125e-01 -4.64243561e-01 1.09688371e-01
9.53280926e-01 2.25242138e-01 4.14958775e-01 -3.14624757e-01
-6.22608364e-01 1.46746878e-02 -1.28678963e-01 -9.51559693e-02
8.95517170e-01 1.03144594e-01 -1.27521813e-01 1.85420081e-01
1.27838373e+00 -9.04825628e-02 -9.97242451e-01 -1.70820415e-01
3.64907473e-01 -1.61784992e-01 -4.49569345e-01 -1.29140282e+00
-3.13801885e-01 7.93988049e-01 -7.06433803e-02 1.15587935e-01
7.54064023e-01 3.13532263e-01 7.18657136e-01 4.95426446e-01
5.11545777e-01 -1.04898536e+00 4.90374804e-01 1.14009440e+00
6.34182215e-01 -7.89823532e-01 -5.33618748e-01 -7.09219337e-01
-6.43134892e-01 5.57869494e-01 8.40112627e-01 5.91324270e-01
3.67686450e-01 6.30611956e-01 2.58186549e-01 -6.43094122e-01
-1.23557186e+00 -4.87301052e-01 -4.63153601e-01 7.47931719e-01
4.13017899e-01 2.29248106e-01 -2.12880954e-01 1.56158864e+00
-3.99266660e-01 -3.10803622e-01 8.56146395e-01 9.10989165e-01
-4.44739431e-01 -1.23984778e+00 5.26769981e-02 3.02355796e-01
-5.99649668e-01 -4.53921795e-01 -1.12734497e+00 2.94512779e-01
-2.22674794e-02 1.52505541e+00 -4.91671413e-01 -1.20461118e+00
4.04255897e-01 1.66189909e-01 1.29871190e-01 -1.20580459e+00
-9.37505066e-01 -1.62657276e-01 7.47479022e-01 -9.81628537e-01
1.44982412e-01 -3.34843636e-01 -1.14853215e+00 -6.74723983e-01
-8.52700397e-02 5.67974508e-01 4.49829102e-02 1.27197409e+00
2.65959110e-02 6.27250135e-01 5.28653502e-01 -2.20480829e-01
-1.47961959e-01 -1.02731729e+00 -8.44403982e-01 3.94844264e-01
-1.06454402e-01 -6.79727972e-01 -2.28265047e-01 -3.95230412e-01]
|
[9.047736167907715, 10.433893203735352]
|
f2c51885-9913-4642-9888-ab5811934bce
|
personalized-state-anxiety-detection-an
|
2304.09928
| null |
https://arxiv.org/abs/2304.09928v1
|
https://arxiv.org/pdf/2304.09928v1.pdf
|
Personalized State Anxiety Detection: An Empirical Study with Linguistic Biomarkers and A Machine Learning Pipeline
|
Individuals high in social anxiety symptoms often exhibit elevated state anxiety in social situations. Research has shown it is possible to detect state anxiety by leveraging digital biomarkers and machine learning techniques. However, most existing work trains models on an entire group of participants, failing to capture individual differences in their psychological and behavioral responses to social contexts. To address this concern, in Study 1, we collected linguistic data from N=35 high socially anxious participants in a variety of social contexts, finding that digital linguistic biomarkers significantly differ between evaluative vs. non-evaluative social contexts and between individuals having different trait psychological symptoms, suggesting the likely importance of personalized approaches to detect state anxiety. In Study 2, we used the same data and results from Study 1 to model a multilayer personalized machine learning pipeline to detect state anxiety that considers contextual and individual differences. This personalized model outperformed the baseline F1-score by 28.0%. Results suggest that state anxiety can be more accurately detected with personalized machine learning approaches, and that linguistic biomarkers hold promise for identifying periods of state anxiety in an unobtrusive way.
|
['Laura E. Barnes', 'Mehdi Boukhechba', 'Bethany A. Teachman', 'Congyu Wu', 'Mark Rucker', 'Emma R. Toner', 'Maria A. Larrazabal', 'Mingyue Tang', 'Zhiyuan Wang']
|
2023-04-19
| null | null | null | null |
['anxiety-detection']
|
['medical']
|
[ 4.17813092e-01 3.39528054e-01 -2.88029253e-01 -8.84780407e-01
-1.10341477e+00 -5.76499701e-01 1.77969739e-01 7.99921870e-01
-1.93331018e-01 2.01322943e-01 4.85427082e-01 1.53929442e-01
-2.00754076e-01 -5.80064595e-01 -5.82256652e-02 2.77158529e-01
-4.24532086e-01 1.18427522e-01 -2.49336764e-01 -2.09700376e-01
3.13178003e-01 2.57485718e-01 -1.30448151e+00 9.09223974e-01
8.17397773e-01 8.62545669e-01 -2.93241531e-01 5.00366449e-01
4.02749069e-02 3.62336516e-01 -5.33279657e-01 -4.98410076e-01
-1.02297932e-01 -3.61119866e-01 -6.19533062e-01 -3.39612458e-03
1.09583461e+00 -5.47791421e-01 2.92260081e-01 1.15003037e+00
5.22442698e-01 -5.72659932e-02 1.49126723e-01 -9.19488072e-01
-8.19048762e-01 1.10521758e+00 -3.61075103e-01 3.28812540e-01
1.11744106e+00 1.35219514e-01 1.05558109e+00 -1.37704507e-01
4.34311688e-01 1.56345868e+00 1.08155346e+00 7.59240806e-01
-1.65661407e+00 -8.97097170e-01 3.16040039e-01 -1.88565418e-01
-7.52443790e-01 -7.84614027e-01 7.74252951e-01 -6.20646298e-01
1.00102186e+00 2.10054785e-01 1.20983028e+00 1.48074281e+00
2.20669240e-01 3.75354022e-01 1.52467048e+00 2.26869747e-01
3.80352706e-01 4.48535174e-01 6.58311009e-01 4.94792372e-01
2.76003212e-01 -5.20512819e-01 -5.16238511e-01 -8.19925249e-01
-1.68730184e-01 1.00986600e-01 4.37700152e-01 3.33582848e-01
-6.04641795e-01 6.88703299e-01 2.56091714e-01 8.76135290e-01
-3.78675073e-01 -2.05153778e-01 8.12666178e-01 4.47552621e-01
9.55881834e-01 7.10868120e-01 -4.07252759e-01 -7.73624897e-01
-5.46871364e-01 1.63037583e-01 4.59971398e-01 -1.23691082e-01
5.48733294e-01 -3.84021457e-03 -1.39157265e-01 1.14839232e+00
2.02309623e-01 2.36865252e-01 9.85696197e-01 -8.44576895e-01
2.84710407e-01 1.07800055e+00 -1.47713095e-01 -1.78760087e+00
-9.77226973e-01 -6.61100820e-02 -2.37986334e-02 -5.47787726e-01
-8.41130018e-02 -5.46639800e-01 -3.56156021e-01 2.07046556e+00
2.36615196e-01 5.08377790e-01 -2.16988340e-01 3.18256617e-01
4.80974078e-01 5.15206717e-02 4.95619684e-01 -2.44109258e-01
1.15148497e+00 1.55967131e-01 -6.13637865e-01 -1.20784557e+00
1.36421359e+00 -2.22697839e-01 1.43123007e+00 4.39967245e-01
-1.17502213e+00 -3.21883172e-01 -9.10028756e-01 2.61795111e-02
-3.19005996e-01 -2.56517947e-01 1.12663066e+00 1.53428054e+00
-1.40272880e+00 6.05893195e-01 -8.39727402e-01 -9.29383457e-01
4.23589826e-01 3.43153328e-01 -2.31783107e-01 1.03779674e-01
-1.13486516e+00 5.46147406e-01 7.90935606e-02 -1.08297147e-01
-3.99719864e-01 -6.68229818e-01 -8.84117424e-01 -6.66416436e-02
-8.30700174e-02 -1.82022065e-01 1.02606797e+00 -1.28360283e+00
-1.47327888e+00 1.28790450e+00 -1.33866156e-02 -4.01672959e-01
-1.28291190e-01 -3.52305770e-01 -7.69166231e-01 4.46400404e-01
3.87426674e-01 4.82601076e-01 5.87929010e-01 -2.52451539e-01
-1.16273664e-01 -1.16998291e+00 1.47676378e-01 9.65956002e-02
-8.56040061e-01 4.33198214e-01 5.90251088e-01 1.17415577e-01
2.96153158e-01 -9.79329228e-01 -2.94090390e-01 -2.37927511e-01
-2.67306298e-01 -1.95476726e-01 6.65673852e-01 -5.53201139e-01
1.14496493e+00 -2.34576511e+00 -4.36137229e-01 2.39422798e-01
5.63980341e-01 2.93797463e-01 -1.91463694e-01 1.62414923e-01
-5.91346435e-02 7.18067348e-01 1.43565774e-01 -4.66837376e-01
1.74335748e-01 -3.33676338e-01 -6.08974742e-03 4.71665621e-01
3.37962329e-01 8.13295841e-01 -1.12486732e+00 -2.39750609e-01
-8.53647143e-02 3.34026247e-01 -1.04736328e+00 -2.17197299e-01
7.51925167e-03 2.34832298e-02 -5.23272812e-01 8.01571071e-01
5.25712430e-01 -7.60365948e-02 4.83219594e-01 4.43026811e-01
2.56848067e-01 7.79139876e-01 -6.24921560e-01 1.38629091e+00
-1.70629382e-01 6.90552533e-01 6.35184765e-01 -6.94880068e-01
9.01246905e-01 1.43586755e-01 6.62049413e-01 -4.92002547e-01
3.57011765e-01 -1.41007537e-02 2.91715086e-01 -7.85492718e-01
4.01279956e-01 -4.18317705e-01 -4.00707006e-01 4.50947315e-01
-4.52201098e-01 -2.31696174e-01 -3.09033036e-01 4.27120358e-01
1.67824233e+00 -6.70502543e-01 -5.26761338e-02 -8.58610123e-02
5.46013638e-02 -5.29021144e-01 6.52887583e-01 7.24085271e-01
-1.05059230e+00 -1.64314300e-01 8.90863419e-01 1.25262871e-01
-1.37587562e-01 -8.04441392e-01 -2.03116521e-01 1.22303092e+00
-5.04462838e-01 -8.44235003e-01 -8.50451350e-01 -3.49069268e-01
6.67985380e-02 8.14180374e-01 -7.30684817e-01 -9.48316455e-01
2.56982982e-01 -8.79879117e-01 6.77906036e-01 3.06967273e-02
2.93923914e-01 -7.40312696e-01 -1.40134230e-01 3.49680871e-01
-1.30619884e-01 -1.26344407e+00 2.01325983e-01 -5.68934023e-01
-9.18206215e-01 -6.13596737e-01 2.76835680e-01 -2.24205807e-01
1.27389222e-01 2.35481754e-01 8.70345831e-01 1.31211476e-02
-5.20275533e-01 1.15834177e+00 -2.80677378e-01 -3.74319166e-01
-2.80509293e-01 2.65327878e-02 4.56190288e-01 3.97343963e-01
1.06645334e+00 -8.06176305e-01 -4.69520986e-01 -2.37691715e-01
-7.56910086e-01 -6.91599131e-01 2.34556437e-01 -2.24958137e-01
8.14070366e-03 -3.13959837e-01 1.16438222e+00 -8.73615324e-01
1.21651340e+00 -1.06507778e+00 2.66016703e-02 -3.00882638e-01
-5.25710225e-01 -8.73073637e-01 1.14570834e-01 -7.25753725e-01
-5.74301779e-01 -1.09997459e-01 -1.72311723e-01 3.91838819e-01
-5.97589374e-01 5.31806648e-01 1.80492878e-01 -8.82190932e-03
8.19488943e-01 -6.15463614e-01 2.89428998e-02 -1.66029617e-01
-4.01175767e-02 9.72157955e-01 -4.10874933e-01 -4.09763396e-01
6.52044266e-02 3.06640148e-01 -2.33365446e-01 -1.31481016e+00
-1.19318211e+00 -1.35598734e-01 -3.14014256e-01 -2.56922096e-01
9.44333971e-01 -8.50347459e-01 -1.05068326e+00 2.94149637e-01
-3.45324069e-01 -5.55056930e-01 1.54586270e-01 5.42289019e-01
-3.04328680e-01 2.50110388e-01 -9.73578393e-01 -9.85150456e-01
-3.31815571e-01 -9.15889919e-01 1.23245716e+00 -2.26961389e-01
-9.75867569e-01 -1.02318740e+00 2.69412965e-01 6.48764431e-01
2.63375491e-01 4.82676715e-01 7.87620366e-01 -1.23252892e+00
2.96197414e-01 -9.76163149e-01 -1.32621229e-01 1.58988163e-01
4.77858603e-01 3.47778127e-02 -1.12196732e+00 -1.26887172e-01
1.29293531e-01 -8.33404541e-01 3.80699217e-01 1.84209749e-01
7.97432303e-01 -2.45175332e-01 -3.90303046e-01 -4.16604489e-01
7.99030960e-01 -2.88684666e-02 4.95099664e-01 -1.01721123e-01
4.67726678e-01 6.83474243e-01 1.41257346e-01 3.57137382e-01
3.83555800e-01 4.29304749e-01 4.65472117e-02 7.21469820e-01
5.86694181e-01 -1.85811222e-01 1.08678639e+00 2.81941950e-01
1.09375322e+00 2.50943542e-01 -9.09603477e-01 4.51713741e-01
-1.21693051e+00 -1.10532570e+00 -7.59534612e-02 2.03717470e+00
7.42583394e-01 3.25301141e-01 3.70804250e-01 -1.96174905e-02
4.81683940e-01 2.65253305e-01 -7.01401412e-01 -9.69765782e-01
1.72117040e-01 -5.40283881e-02 -3.96638811e-01 6.01876318e-01
-7.27731109e-01 8.83513153e-01 6.98578787e+00 9.72175784e-03
-1.43479514e+00 1.00861169e-01 1.01071835e+00 -7.70251811e-01
-2.17600152e-01 -2.44910538e-01 -7.00184643e-01 4.27111566e-01
1.92076099e+00 -1.02763265e-01 5.26948154e-01 7.60022938e-01
4.06900376e-01 -8.13286975e-02 -1.08091724e+00 6.12044394e-01
3.90563399e-01 -4.13979530e-01 -6.58298731e-01 3.31600159e-02
3.00003529e-01 5.85277118e-02 5.03449261e-01 5.79814851e-01
3.52021046e-02 -6.18521214e-01 6.15014471e-02 6.39609635e-01
3.19739908e-01 -3.50403935e-01 2.00816274e-01 7.86622614e-02
-4.72083420e-01 -5.83095670e-01 -1.71191111e-01 -7.78631628e-01
-1.09512694e-01 8.73011947e-01 -1.39914274e+00 -5.41001022e-01
6.70810521e-01 9.27268803e-01 -1.09053314e+00 5.18085301e-01
4.29857284e-01 9.84795690e-01 -6.04524434e-01 -2.82815099e-01
2.55564630e-01 -1.27997711e-01 3.00831079e-01 9.50303674e-01
2.95623720e-01 1.11039340e-01 2.02647507e-01 6.93156362e-01
3.84747609e-02 4.76519167e-01 -1.15511978e+00 -1.04063451e+00
2.35688820e-01 1.45883739e+00 -7.63697684e-01 -4.33764189e-01
-2.83203870e-01 8.37811768e-01 5.12123227e-01 -1.93062648e-01
-2.65121281e-01 -1.61788002e-01 7.93138266e-01 5.74978650e-01
-7.20079839e-01 -4.22181701e-03 -3.13141733e-01 -1.07407880e+00
-1.42467782e-01 -8.38766992e-01 4.98109907e-01 -9.67209637e-01
-1.47651815e+00 2.48155445e-01 -1.27196252e-01 -4.24022019e-01
-4.07721817e-01 -3.47384453e-01 -5.30544102e-01 2.71175265e-01
-4.24964041e-01 -8.17177355e-01 -1.68740898e-02 1.61426470e-01
8.93833041e-02 4.41649914e-01 8.28941584e-01 2.97213644e-01
-9.32510495e-01 5.15599728e-01 -3.57312351e-01 -3.23235720e-01
1.18626118e+00 -9.34745014e-01 -5.67901209e-02 4.62869376e-01
-4.54324603e-01 9.46792305e-01 5.03344953e-01 -1.30345678e+00
-1.30615258e+00 -9.70263660e-01 9.02388692e-01 -8.18321407e-01
1.14447737e+00 -6.83716118e-01 -9.28851604e-01 1.22060418e+00
-7.10058659e-02 -6.52707100e-01 1.24813366e+00 7.84846008e-01
-2.56485015e-01 -1.45207360e-01 -1.92877603e+00 1.02434373e+00
1.24867690e+00 -9.82687593e-01 -3.05938989e-01 8.10824037e-01
9.31484401e-01 2.02757865e-01 -1.00921392e+00 1.83495149e-01
4.94983584e-01 -1.13690031e+00 7.84075677e-01 -6.61289036e-01
2.22255766e-01 8.68702948e-01 -2.39133790e-01 -1.25692987e+00
-3.48994970e-01 -5.72163522e-01 4.09886569e-01 1.63689935e+00
3.75581443e-01 -9.33602035e-01 6.14092946e-01 1.73891246e+00
-1.39367774e-01 -3.21840882e-01 -8.00785303e-01 -7.58166835e-02
4.76493360e-03 -8.68194699e-01 4.40781474e-01 1.28342450e+00
8.92859042e-01 3.20472956e-01 2.27001801e-01 2.79056162e-01
2.83911265e-02 -4.92136419e-01 3.56111974e-01 -1.53384316e+00
3.22694033e-02 -5.12410939e-01 -9.19272065e-01 4.96883206e-02
7.50237167e-01 -1.03116524e+00 -5.85909843e-01 -1.15020978e+00
1.29193678e-01 -1.62862837e-01 -2.52628982e-01 6.21625483e-01
-2.60801643e-01 2.93736368e-01 -1.72495231e-01 -8.02485645e-01
-8.28342915e-01 2.56469488e-01 3.66336346e-01 4.44704331e-02
-7.07329750e-01 -2.01316938e-01 -1.55583978e+00 9.44477320e-01
1.23085725e+00 -1.27642587e-01 -5.90434253e-01 1.40672307e-02
7.30573177e-01 -1.49910063e-01 4.03571934e-01 -1.17352712e+00
-4.29767221e-01 -1.60132229e-01 2.42904991e-01 -1.80184748e-02
9.83456671e-01 -5.52833259e-01 -5.44448614e-01 2.17269510e-01
-7.46662736e-01 1.56343415e-01 6.70894206e-01 3.40501964e-01
7.00284243e-01 -2.96502769e-01 5.65877140e-01 -1.63304403e-01
-3.76571231e-02 1.68321412e-02 -9.13093865e-01 -2.14201072e-03
7.09590256e-01 -3.56771536e-02 -1.24416411e-01 -5.92933595e-01
-1.10294569e+00 -2.25070164e-01 3.67881179e-01 4.23454881e-01
5.11536360e-01 -9.35276330e-01 -2.65077591e-01 1.81838691e-01
7.02903718e-02 -1.07212126e+00 8.45973313e-01 1.08763242e+00
2.65055448e-01 7.36419633e-02 2.40531005e-02 -3.45839858e-01
-1.22479272e+00 4.59151983e-01 3.19966584e-01 2.45530292e-01
-2.51000673e-01 7.63763130e-01 -3.54270965e-01 -3.65135044e-01
-1.17025487e-01 -5.21481156e-01 -1.34000435e-01 6.48070037e-01
1.05937994e+00 3.46959144e-01 -7.44774938e-02 -6.01528227e-01
-1.02009624e-01 5.43432645e-02 -3.46072644e-01 -1.96216211e-01
1.34784245e+00 -2.99220502e-01 -5.14573514e-01 9.73070383e-01
1.36080217e+00 3.67977470e-01 -2.96630323e-01 -1.26898140e-01
1.67912975e-01 -4.82447654e-01 7.95956329e-02 -6.91956699e-01
-4.51484352e-01 7.00690567e-01 1.10518122e+00 7.51187384e-01
8.49845290e-01 3.75248939e-02 5.90040386e-01 4.33559179e-01
1.31829992e-01 -1.07086575e+00 2.72067487e-01 2.01517701e-01
6.42454505e-01 -1.04910111e+00 -1.49266124e-01 3.08416169e-02
-6.48821831e-01 2.91133702e-01 7.97430456e-01 -3.36854696e-01
7.42579520e-01 -1.09419554e-01 -1.04153097e-01 -5.88027596e-01
-8.19920957e-01 1.90259367e-01 -3.31721157e-02 6.64551497e-01
8.21665108e-01 4.07628655e-01 -4.01408583e-01 1.10061574e+00
-5.92696369e-01 -2.03756824e-01 6.16891563e-01 9.08331394e-01
-7.00981379e-01 -8.43537688e-01 -2.83164024e-01 1.18414557e+00
-9.03546214e-01 -3.69913764e-02 -1.56976008e+00 7.79215321e-02
-8.79332498e-02 1.50214040e+00 1.94714487e-01 -6.90220654e-01
1.41770199e-01 8.47984612e-01 6.45809546e-02 -1.07968330e+00
-9.31263804e-01 -1.89368606e-01 5.34978211e-01 -1.01354539e+00
-8.55621770e-02 -1.15992963e+00 -1.01431596e+00 -2.74564862e-01
1.81410000e-01 -1.48512676e-01 6.07964873e-01 8.92885089e-01
1.11440361e+00 1.03105135e-01 3.99024516e-01 -6.52086675e-01
-1.85617492e-01 -1.26350749e+00 -7.63949871e-01 3.07773918e-01
4.28392172e-01 -3.63291711e-01 -4.20279384e-01 -7.13997483e-01]
|
[8.786940574645996, 10.35095500946045]
|
ca0b153b-e078-420c-b21a-538ebc8b63a7
|
timemae-self-supervised-representations-of
|
2303.00320
| null |
https://arxiv.org/abs/2303.00320v3
|
https://arxiv.org/pdf/2303.00320v3.pdf
|
TimeMAE: Self-Supervised Representations of Time Series with Decoupled Masked Autoencoders
|
Enhancing the expressive capacity of deep learning-based time series models with self-supervised pre-training has become ever-increasingly prevalent in time series classification. Even though numerous efforts have been devoted to developing self-supervised models for time series data, we argue that the current methods are not sufficient to learn optimal time series representations due to solely unidirectional encoding over sparse point-wise input units. In this work, we propose TimeMAE, a novel self-supervised paradigm for learning transferrable time series representations based on transformer networks. The distinct characteristics of the TimeMAE lie in processing each time series into a sequence of non-overlapping sub-series via window-slicing partitioning, followed by random masking strategies over the semantic units of localized sub-series. Such a simple yet effective setting can help us achieve the goal of killing three birds with one stone, i.e., (1) learning enriched contextual representations of time series with a bidirectional encoding scheme; (2) increasing the information density of basic semantic units; (3) efficiently encoding representations of time series using transformer networks. Nevertheless, it is a non-trivial to perform reconstructing task over such a novel formulated modeling paradigm. To solve the discrepancy issue incurred by newly injected masked embeddings, we design a decoupled autoencoder architecture, which learns the representations of visible (unmasked) positions and masked ones with two different encoder modules, respectively. Furthermore, we construct two types of informative targets to accomplish the corresponding pretext tasks. One is to create a tokenizer module that assigns a codeword to each masked region, allowing the masked codeword classification (MCC) task to be completed effectively...
|
['Enhong Chen', 'Rujiao Zhang', 'Hao Zhang', 'Zhiding Liu', 'Qi Liu', 'Mingyue Cheng']
|
2023-03-01
| null | null | null | null |
['time-series-classification']
|
['time-series']
|
[ 2.97427118e-01 -6.82900473e-02 -4.21724059e-02 -2.98205495e-01
-4.80735630e-01 -6.07475579e-01 6.35689497e-01 4.56363186e-02
-2.71411240e-01 3.50717753e-01 3.82633865e-01 -3.60076934e-01
-1.54573232e-01 -9.91396785e-01 -7.21437275e-01 -9.27996814e-01
-6.02185011e-01 6.85696080e-02 -2.76618898e-02 -3.51193845e-01
-3.73565499e-03 4.76652414e-01 -1.97562838e+00 4.35800523e-01
8.34270358e-01 1.06363881e+00 1.38764158e-01 4.59265798e-01
-2.41244242e-01 9.48687315e-01 -6.10921919e-01 -6.45492375e-02
2.52686441e-01 -5.79816878e-01 -5.26867509e-01 4.16291674e-04
-2.53867149e-01 -2.18526214e-01 -5.37641764e-01 7.81203508e-01
1.84136435e-01 2.10252702e-01 5.56254923e-01 -1.25891781e+00
-7.79349804e-01 8.58003914e-01 -3.62133831e-01 3.09568584e-01
7.12953624e-04 -1.02245033e-01 1.01394928e+00 -5.10964096e-01
3.47437978e-01 8.08432758e-01 6.93541467e-01 4.43336070e-01
-1.23776555e+00 -6.84144139e-01 -4.50753607e-03 3.76875222e-01
-1.21119320e+00 -3.12075853e-01 1.24563873e+00 -3.80996346e-01
9.49205041e-01 4.70854253e-01 8.18229973e-01 1.33643830e+00
1.16453640e-01 5.21968424e-01 1.12716579e+00 -3.51363361e-01
4.06211913e-01 9.11083817e-02 -4.04124819e-02 5.56118786e-01
-3.68092567e-01 2.95070499e-01 -4.86189783e-01 -8.12582523e-02
7.29454160e-01 6.64764285e-01 -1.60460874e-01 -3.76569957e-01
-1.40921783e+00 8.65988255e-01 6.00252688e-01 9.30205941e-01
-3.86415303e-01 -2.59632375e-02 6.31193340e-01 6.78775012e-01
6.80132270e-01 4.59604323e-01 -4.89647955e-01 9.00555775e-03
-9.65926409e-01 -6.95683807e-02 4.28076148e-01 6.13907874e-01
7.99664140e-01 4.11480755e-01 -1.68366339e-02 6.07456684e-01
-2.11214945e-01 1.76756904e-01 1.10501611e+00 -4.36160475e-01
2.18935713e-01 5.78706205e-01 -2.36327782e-01 -1.02952242e+00
-3.24304104e-01 -6.49518847e-01 -1.06042945e+00 -3.08361202e-02
1.79244906e-01 -5.56416214e-02 -8.98744166e-01 2.07104778e+00
1.41382858e-01 4.48016584e-01 2.63580084e-01 7.32783973e-01
4.27434832e-01 1.11625612e+00 1.01700397e-02 -4.34988409e-01
1.28506446e+00 -7.46519327e-01 -7.32909679e-01 1.35854818e-02
7.05730736e-01 -4.93879020e-01 9.41244006e-01 -3.22781913e-02
-7.18798399e-01 -6.51502848e-01 -1.28151035e+00 -3.68308532e-03
-8.61094952e-01 1.39485687e-01 7.84343719e-01 2.22405761e-01
-8.26358616e-01 8.78802836e-01 -9.14566875e-01 -2.39494056e-01
2.44039938e-01 2.13542551e-01 -6.71590686e-01 3.16129535e-01
-1.34714603e+00 7.41179824e-01 4.19798017e-01 1.50809288e-02
-1.04211557e+00 -8.60742152e-01 -1.02338886e+00 2.55103409e-01
-7.00982288e-02 -2.51384407e-01 8.06723773e-01 -1.39880621e+00
-1.34937549e+00 7.71571517e-01 -5.15021347e-02 -7.11456478e-01
1.53148085e-01 2.27276143e-02 -6.81069195e-01 1.17720619e-01
2.57762849e-01 4.42311347e-01 1.17387784e+00 -9.42028284e-01
-3.63109946e-01 -4.88640785e-01 3.81239690e-02 1.16849439e-02
-8.97807479e-01 -1.10468897e-03 2.11296991e-01 -1.11876833e+00
1.54199868e-01 -5.50930262e-01 -2.00156406e-01 -1.41129404e-01
-1.42839357e-01 -1.65679589e-01 1.00626230e+00 -6.49376333e-01
1.33885789e+00 -2.51814747e+00 2.38941714e-01 6.29194500e-03
2.36610666e-01 7.46852607e-02 -2.52568513e-01 7.53660619e-01
-6.08112454e-01 -2.57061332e-01 -4.29636300e-01 -3.37024242e-01
-6.40909299e-02 2.65172452e-01 -1.09453106e+00 4.41681832e-01
4.39509958e-01 7.21390426e-01 -9.61526215e-01 -1.85467616e-01
3.07885408e-01 5.57898700e-01 -5.10590792e-01 4.42284495e-01
-8.81587993e-03 4.16949898e-01 -2.42125779e-01 3.70926052e-01
4.39558148e-01 -1.16341896e-01 6.98917434e-02 -3.29259783e-01
-4.20747131e-01 3.35826218e-01 -6.86264932e-01 1.81764889e+00
-5.03385067e-01 4.73610699e-01 -2.58811086e-01 -1.65556741e+00
9.44602549e-01 4.92916167e-01 8.78027380e-01 -7.25750029e-01
1.38337955e-01 1.53268322e-01 -1.30597457e-01 -5.35049319e-01
2.94476211e-01 -5.83868623e-01 -1.93612307e-01 5.78726947e-01
3.57271969e-01 1.96821839e-01 -1.14624821e-01 -1.85560942e-01
1.26778412e+00 1.73622087e-01 1.92996159e-01 -1.29047692e-01
3.24008673e-01 5.20688528e-03 4.42628682e-01 1.98077768e-01
-7.93314204e-02 5.18206239e-01 2.70217985e-01 -7.37660408e-01
-1.18885505e+00 -1.04241431e+00 -1.86274934e-03 1.33112419e+00
-1.90019876e-01 -4.67000395e-01 -4.95110065e-01 -4.39624578e-01
-1.53248489e-01 5.70315540e-01 -9.52411592e-01 -5.32903373e-01
-6.48200333e-01 -6.34957373e-01 6.67840362e-01 6.72091067e-01
3.17557544e-01 -1.11941004e+00 -8.48304331e-01 3.42475444e-01
-1.15463383e-01 -7.81930506e-01 -2.84093708e-01 9.33256567e-01
-9.68964279e-01 -7.50262260e-01 -6.97303355e-01 -9.29301918e-01
6.83400929e-01 4.56660002e-01 6.92200005e-01 -2.14338019e-01
-2.55676448e-01 6.89732581e-02 -5.73122621e-01 -1.32192865e-01
-1.12551548e-01 -1.19609185e-01 1.57969758e-01 2.80780256e-01
3.51685822e-01 -1.18537199e+00 -4.90531594e-01 1.89285517e-01
-1.32773125e+00 1.08171649e-01 4.44886893e-01 9.06151891e-01
4.56904113e-01 3.50643367e-01 5.63839018e-01 -5.05542159e-01
4.55193311e-01 -6.80468202e-01 -2.85089791e-01 6.06381483e-02
-1.87427849e-01 1.52873158e-01 1.21564567e+00 -8.40550959e-01
-7.47279823e-01 -8.07873085e-02 -1.84005469e-01 -8.25207293e-01
-1.28671508e-02 5.43086946e-01 1.04405485e-01 2.24545732e-01
6.19879603e-01 7.74968326e-01 1.80336088e-01 -3.70467365e-01
5.48071146e-01 5.85340500e-01 5.58697999e-01 -5.66596985e-01
1.05820453e+00 7.35143483e-01 -3.48529488e-01 -6.48859799e-01
-6.94911838e-01 -2.32169345e-01 -5.17834544e-01 2.16133781e-02
9.18314397e-01 -8.77492547e-01 -4.13829744e-01 2.43529543e-01
-8.81904483e-01 -2.20760778e-01 -6.66081250e-01 4.34770137e-01
-6.49863064e-01 2.15084463e-01 -5.38976550e-01 -6.48550332e-01
-1.28978804e-01 -7.49327898e-01 9.01749492e-01 -8.28042030e-02
-2.22429335e-01 -9.67158437e-01 2.22065359e-01 -1.02920666e-01
5.86790860e-01 3.63113642e-01 1.15773070e+00 -8.19163084e-01
-2.44846985e-01 -2.36718461e-01 6.82449266e-02 4.66808379e-01
2.86026090e-01 -2.51872122e-01 -1.24100423e+00 -3.66479307e-01
5.90808511e-01 -3.59730303e-01 7.93098330e-01 1.39056250e-01
1.54473054e+00 -4.67763513e-01 -3.03578883e-01 8.12771916e-01
1.26027954e+00 3.94833237e-01 5.02357543e-01 2.37353623e-01
5.44862092e-01 6.51699662e-01 2.13288039e-01 4.91093129e-01
3.01232368e-01 4.09163058e-01 2.90983796e-01 -8.71588513e-02
1.80690259e-01 -6.74771786e-01 5.75704038e-01 1.26289189e+00
-1.00485615e-01 8.80696550e-02 -6.80993021e-01 7.80010998e-01
-1.77317381e+00 -1.27470708e+00 5.65962195e-01 2.10215354e+00
7.77027249e-01 -4.50875759e-02 8.95624235e-03 5.69227278e-01
6.46090567e-01 5.16617000e-01 -4.78729010e-01 -1.70194089e-01
-1.17255978e-01 4.75028217e-01 1.26388386e-01 -1.34158835e-01
-1.27743387e+00 6.05612218e-01 5.93984985e+00 8.08992505e-01
-1.47825444e+00 2.66211182e-01 4.04963136e-01 8.99465159e-02
-4.93723333e-01 2.51834579e-02 -8.54693055e-02 5.64650178e-01
1.12965190e+00 -2.92898536e-01 5.47632158e-01 7.76001096e-01
2.83147525e-02 6.37067258e-01 -1.28103113e+00 9.65568483e-01
2.46462785e-02 -1.31374860e+00 2.05619499e-01 -1.75445825e-01
4.65383112e-01 -2.75989383e-01 1.61461681e-01 6.46984100e-01
1.93725023e-02 -1.04111147e+00 8.49340379e-01 2.49117747e-01
8.20748329e-01 -5.06780922e-01 3.55125308e-01 4.34968323e-01
-1.45199335e+00 -4.68934655e-01 -4.21564072e-01 -3.58724594e-01
-1.87358081e-01 5.32228529e-01 -3.21045548e-01 6.79640651e-01
6.51530743e-01 9.57922935e-01 -1.42544389e-01 5.96062601e-01
5.13812080e-02 6.23267710e-01 -1.77002326e-01 2.36140952e-01
3.02949876e-01 -1.15117215e-01 3.95992905e-01 8.54983866e-01
5.66778839e-01 1.12805702e-01 -5.45680895e-02 9.51294065e-01
8.34861770e-02 1.43779479e-02 -9.95272398e-01 -3.08973640e-01
5.35160005e-01 1.11867249e+00 -7.08414435e-01 -2.39686906e-01
-4.36803758e-01 1.03114891e+00 4.33552533e-01 4.38552678e-01
-1.01725984e+00 -5.42991459e-01 7.25116551e-01 8.65151882e-02
4.66454864e-01 -3.08085471e-01 -4.01172452e-02 -1.46952760e+00
1.62369728e-01 -7.83124387e-01 5.09741187e-01 -6.53363228e-01
-1.42626595e+00 9.15984750e-01 -5.75532392e-02 -1.81763136e+00
-2.75648385e-01 -3.81775647e-01 -6.57025337e-01 5.89647293e-01
-1.42169142e+00 -1.26219058e+00 -7.60881007e-02 8.30183685e-01
4.67649370e-01 -2.08661720e-01 1.18273532e+00 4.64820921e-01
-3.81689370e-01 4.51752156e-01 1.15032457e-01 2.52007216e-01
4.97587204e-01 -1.08027148e+00 1.19515218e-01 8.88493001e-01
2.44869843e-01 8.65973175e-01 5.22278547e-01 -1.68383479e-01
-1.57609773e+00 -1.34182405e+00 7.66085207e-01 -1.31262437e-01
1.02434647e+00 -6.32583320e-01 -1.04316449e+00 8.18418503e-01
3.83405089e-02 1.24733567e-01 8.69115114e-01 -1.11671155e-02
-7.03878760e-01 -4.69507635e-01 -8.63069832e-01 5.79661131e-01
1.06774139e+00 -1.09393525e+00 -1.03307927e+00 2.59037703e-01
9.38207686e-01 -7.48209953e-02 -8.50661576e-01 4.21014249e-01
3.47501516e-01 -9.04313982e-01 1.01097190e+00 -7.19781995e-01
6.81903660e-01 -3.83469641e-01 -2.14837760e-01 -1.36970925e+00
-3.47899228e-01 -6.91651106e-01 -1.10545188e-01 1.21426189e+00
3.53958785e-05 -8.59121561e-01 4.97756958e-01 -2.99118496e-02
-3.00035268e-01 -6.06678426e-01 -1.10573542e+00 -7.94815063e-01
9.75299720e-03 -4.71367210e-01 8.25554848e-01 1.36201525e+00
3.84394377e-01 1.93315789e-01 -4.67284381e-01 7.51527697e-02
3.92583430e-01 5.22944033e-01 4.01293635e-01 -1.04874372e+00
-1.95760220e-01 -4.22557354e-01 -4.71512288e-01 -9.00040329e-01
3.87368679e-01 -1.03864467e+00 -5.46190999e-02 -9.10415769e-01
-4.06180583e-02 -4.45150584e-01 -6.55079305e-01 6.32199287e-01
1.66753501e-01 7.98885897e-02 -1.90321341e-01 3.81360888e-01
-1.37659878e-01 8.77398670e-01 9.85421836e-01 -1.89037159e-01
3.45432125e-02 -8.85199234e-02 -6.05801165e-01 4.18630898e-01
6.73637807e-01 -3.46966535e-01 -6.97916567e-01 -5.36640942e-01
2.85280100e-03 3.41299832e-01 5.05265534e-01 -1.08134544e+00
2.58173078e-01 8.31515566e-02 3.23973387e-01 -3.92274320e-01
2.88113862e-01 -1.10040975e+00 3.02563041e-01 4.31094140e-01
-5.04520714e-01 2.18403503e-01 1.97632477e-01 5.50821066e-01
-6.55448854e-01 8.13315287e-02 5.11889815e-01 -1.62403181e-01
-7.64369011e-01 2.92776495e-01 -3.96247000e-01 -3.33357543e-01
1.06139970e+00 -2.25486293e-01 -2.14139745e-01 -1.82342842e-01
-7.17415512e-01 -1.78060025e-01 2.32915461e-01 6.14871144e-01
5.60680091e-01 -1.65305388e+00 -3.19864273e-01 7.00526595e-01
3.38067025e-01 -3.05522472e-01 3.95668387e-01 5.71043193e-01
-2.20253080e-01 3.26125145e-01 -6.67761445e-01 -4.65854198e-01
-7.46307552e-01 9.33918595e-01 2.04718396e-01 -2.23591283e-01
-8.46088171e-01 5.44249237e-01 3.11577767e-01 -4.66551632e-01
2.44971707e-01 -5.51801324e-01 -2.85651624e-01 1.80611581e-01
5.61383188e-01 9.84620396e-03 5.72650768e-02 -5.23616016e-01
-2.40409091e-01 5.01908362e-01 2.25754157e-02 5.96383736e-02
1.73739111e+00 1.67860582e-01 -2.90029198e-01 7.25008845e-01
1.49479902e+00 -1.64363593e-01 -9.72842395e-01 -2.72735327e-01
-1.87774733e-01 -1.35577604e-01 -1.00194737e-01 -2.72936553e-01
-9.61484790e-01 1.04664373e+00 5.26804686e-01 6.14731193e-01
1.52292931e+00 -1.56753659e-01 8.56424093e-01 7.67729059e-02
4.26670462e-01 -7.03577161e-01 1.53290227e-01 3.36679786e-01
7.39991248e-01 -8.72354925e-01 -4.71544564e-01 -5.80674186e-02
-3.67937595e-01 1.17499113e+00 2.93475121e-01 -1.86355859e-01
7.08835363e-01 3.25251997e-01 4.09565680e-02 -2.58093804e-01
-8.54981720e-01 -2.95058191e-01 2.14471415e-01 5.19550443e-01
3.78769606e-01 1.85211062e-01 -2.85346899e-02 8.20938945e-01
-3.56871337e-01 -1.09361418e-01 1.07282378e-01 9.17446673e-01
-2.23427847e-01 -9.13612604e-01 -2.65129596e-01 2.70623535e-01
-2.46188775e-01 2.26683766e-02 2.34941430e-02 5.45802414e-01
2.15960160e-01 6.78511262e-01 2.71355808e-01 -7.47420371e-01
1.41396359e-01 2.48740762e-01 1.12676881e-01 -5.70217729e-01
-6.46127105e-01 -2.50616018e-02 -3.48643512e-01 -4.28314567e-01
-5.06536722e-01 -3.41392875e-01 -1.09710753e+00 -1.38474077e-01
3.18534635e-02 3.36073756e-01 3.68586272e-01 8.75488758e-01
3.36604834e-01 6.83970511e-01 1.04577518e+00 -9.15288687e-01
-6.53219938e-01 -8.56060207e-01 -5.66399097e-01 6.92840397e-01
5.83064258e-01 -6.67862594e-01 -5.90217829e-01 2.09127784e-01]
|
[7.246419906616211, 3.0421650409698486]
|
29a0b056-b162-4bec-b18b-631302b8f0df
|
wide-range-mri-artifact-removal-with
|
2210.07976
| null |
https://arxiv.org/abs/2210.07976v2
|
https://arxiv.org/pdf/2210.07976v2.pdf
|
Wide Range MRI Artifact Removal with Transformers
|
Artifacts on magnetic resonance scans are a serious challenge for both radiologists and computer-aided diagnosis systems. Most commonly, artifacts are caused by motion of the patients, but can also arise from device-specific abnormalities such as noise patterns. Irrespective of the source, artifacts can not only render a scan useless, but can potentially induce misdiagnoses if left unnoticed. For instance, an artifact may masquerade as a tumor or other abnormality. Retrospective artifact correction (RAC) is concerned with removing artifacts after the scan has already been taken. In this work, we propose a method capable of retrospectively removing eight common artifacts found in native-resolution MR imagery. Knowledge of the presence or location of a specific artifact is not assumed and the system is, by design, capable of undoing interactions of multiple artifacts. Our method is realized through the design of a novel volumetric transformer-based neural network that generalizes a \emph{window-centered} approach popularized by the Swin transformer. Unlike Swin, our method is (i) natively volumetric, (ii) geared towards dense prediction tasks instead of classification, and (iii), uses a novel and more global mechanism to enable information exchange between windows. Our experiments show that our reconstructions are considerably better than those attained by ResNet, V-Net, MobileNet-v2, DenseNet, CycleGAN and BicycleGAN. Moreover, we show that the reconstructed images from our model improves the accuracy of FSL BET, a standard skull-stripping method typically applied in diagnostic workflows.
|
['Kevin Smith', 'Lennart Alexander Van der Goten']
|
2022-10-14
| null | null | null | null |
['skull-stripping']
|
['medical']
|
[ 3.96283329e-01 3.26892018e-01 1.47043973e-01 -6.32833391e-02
-5.88335216e-01 -1.25548005e-01 2.91673809e-01 1.19331680e-01
-2.75572419e-01 9.12646234e-01 2.13002607e-01 -4.47553694e-01
-3.59376699e-01 -6.39706135e-01 -8.30841184e-01 -6.99540615e-01
-1.27867147e-01 5.57699978e-01 5.20827115e-01 8.63488689e-02
-9.89519432e-02 5.99960983e-01 -1.14905512e+00 2.15459853e-01
6.98788404e-01 1.09361911e+00 4.05745387e-01 1.63305119e-01
5.41516617e-02 1.13811707e+00 -5.37961185e-01 4.09468152e-02
6.76305816e-02 -4.75052118e-01 -8.39658320e-01 -6.16572285e-03
1.54820442e-01 -3.57440412e-01 -2.02851921e-01 8.29026401e-01
4.92277980e-01 -1.11552574e-01 4.62176532e-01 -8.17217469e-01
-6.85483217e-02 8.01631033e-01 -4.97091800e-01 3.16490829e-01
2.16678809e-02 2.40291074e-01 1.84688717e-01 -8.19095969e-01
9.09999251e-01 5.03442883e-01 1.09365737e+00 4.60579485e-01
-1.18493044e+00 -5.37460744e-01 -1.49648279e-01 2.02963099e-01
-1.23736107e+00 -3.23683470e-01 6.91408515e-01 -5.82714915e-01
7.20594168e-01 5.64562738e-01 8.98510158e-01 1.18550658e+00
5.94534695e-01 2.76126921e-01 1.03864181e+00 -2.73058027e-01
3.34560901e-01 1.28088847e-01 -2.68368274e-02 5.94123304e-01
3.53135914e-01 3.68908085e-02 -3.38133931e-01 -9.31405053e-02
9.81296897e-01 1.22150362e-01 -7.10940838e-01 -5.36695123e-01
-1.28455496e+00 5.84741950e-01 4.88327771e-01 6.60278738e-01
-7.47602582e-01 2.21416980e-01 3.99397522e-01 -1.53414086e-02
3.96610767e-01 4.50818032e-01 -2.33968377e-01 5.79438210e-02
-1.32886159e+00 -8.41998775e-03 5.59864163e-01 5.17455578e-01
2.75557995e-01 1.84369579e-01 -2.23571807e-01 6.02259934e-01
4.33144346e-02 2.34059006e-01 9.55201566e-01 -4.29579377e-01
1.16360318e-02 4.42963630e-01 -5.58205321e-03 -8.58149230e-01
-8.39687347e-01 -7.98317134e-01 -1.06868196e+00 2.42863327e-01
2.05365166e-01 3.87052298e-02 -1.11706579e+00 1.44409978e+00
2.85186201e-01 3.76089871e-01 -4.38090354e-01 1.01273501e+00
8.89179289e-01 -1.04589108e-02 5.39397895e-02 -4.33152944e-01
1.21378756e+00 -8.11271012e-01 -8.19481611e-01 -1.56467453e-01
5.22910833e-01 -6.06689930e-01 6.51868343e-01 7.12513208e-01
-1.22798538e+00 -2.12160930e-01 -1.10048020e+00 1.81170329e-01
-1.56781659e-01 -1.99237764e-01 5.07427931e-01 6.89304471e-01
-1.01919532e+00 9.81704056e-01 -1.05185723e+00 -1.27592022e-04
3.52469385e-01 4.15260792e-01 -3.00900161e-01 1.73656359e-01
-1.01014018e+00 1.33094573e+00 1.87984243e-01 3.02539587e-01
-9.21544850e-01 -1.19994390e+00 -6.17039680e-01 -8.20084438e-02
5.76657116e-01 -6.21592045e-01 1.17585540e+00 -1.01352119e+00
-1.27734435e+00 6.38642013e-01 -1.34228943e-02 -7.57216215e-01
1.13893664e+00 -6.80679083e-02 -4.80464160e-01 1.52952343e-01
4.42761108e-02 1.55560866e-01 9.43352759e-01 -1.15935886e+00
-1.51432557e-02 -2.76629776e-01 -3.40232909e-01 -1.65615320e-01
8.43545943e-02 -1.95623264e-01 -1.64211765e-01 -7.45900333e-01
3.74704093e-01 -7.04048455e-01 -4.01981175e-01 1.69686437e-01
-4.43864673e-01 5.13543546e-01 7.18373954e-01 -1.01685297e+00
1.16264856e+00 -1.83194590e+00 -5.25503755e-02 5.86325228e-01
5.17246604e-01 3.57570708e-01 3.67861331e-01 9.97069553e-02
-5.00548363e-01 9.29864962e-03 -5.68780124e-01 -9.67090502e-02
-6.15963221e-01 9.78249162e-02 -2.28339955e-02 6.80804968e-01
-2.38926753e-01 7.05900550e-01 -7.20570445e-01 -4.02623028e-01
5.43999970e-01 6.75873339e-01 -3.02856922e-01 9.86025855e-03
1.07905157e-01 8.71870041e-01 -2.76856393e-01 2.90789485e-01
6.99739814e-01 -4.62964147e-01 4.22805727e-01 -4.37350541e-01
-1.50179848e-01 2.33342037e-01 -1.17782640e+00 1.79189551e+00
-7.38604367e-01 3.39441001e-01 3.98472808e-02 -8.84540796e-01
5.47466278e-01 6.76817834e-01 6.96477354e-01 -7.74301231e-01
3.19619387e-01 3.91381264e-01 -1.43642193e-02 -6.49846077e-01
2.85344362e-01 -2.54010350e-01 4.90240604e-01 4.26210165e-01
-1.02725327e-01 6.74399212e-02 -2.30561227e-01 -6.80494606e-02
1.30421126e+00 6.81628138e-02 4.09628153e-01 -2.85510272e-01
3.82252008e-01 -1.12856880e-01 4.19579744e-01 8.30634832e-01
-9.71727446e-02 8.80050302e-01 2.42682189e-01 -5.12023270e-01
-9.71546292e-01 -1.12540078e+00 -4.32934165e-01 2.22591043e-01
8.83356258e-02 -6.84721246e-02 -6.78556025e-01 -6.42331183e-01
-2.49344915e-01 6.53668106e-01 -7.30988145e-01 -1.51021019e-01
-7.29288399e-01 -7.04661012e-01 4.05047148e-01 4.91845846e-01
3.08989316e-01 -1.07536829e+00 -1.05813372e+00 6.62322760e-01
-4.25700724e-01 -9.49873388e-01 -3.00233394e-01 4.54087287e-01
-1.26878262e+00 -1.33322704e+00 -5.99461198e-01 -2.49964654e-01
6.41461134e-01 9.02487412e-02 1.12265861e+00 3.92129481e-01
-5.00050306e-01 1.16736755e-01 -2.18226075e-01 -2.78631628e-01
-3.94345939e-01 -1.81799963e-01 -8.81332830e-02 3.00090052e-02
-2.21336260e-01 -8.66915166e-01 -7.83100665e-01 2.45035902e-01
-1.00054884e+00 2.02734724e-01 5.07762551e-01 8.21377575e-01
6.24744654e-01 -1.29715502e-01 4.40735340e-01 -1.17223740e+00
4.07271206e-01 -5.45070589e-01 -3.18469673e-01 3.02829221e-02
-6.24709666e-01 -8.99730064e-03 5.64660311e-01 -3.60750198e-01
-9.05961514e-01 9.89701748e-02 -2.60221034e-01 -7.19926953e-01
-1.08023256e-01 4.20002788e-01 1.90624863e-01 -3.09484869e-01
7.75511861e-01 2.41931483e-01 1.70731600e-02 -4.92729664e-01
-9.71891731e-02 3.26920927e-01 5.37374735e-01 -1.83532447e-01
5.49923658e-01 7.27497041e-01 9.21464860e-02 -8.21557701e-01
-3.67907941e-01 -1.45369247e-01 -5.76914847e-01 -5.55955470e-01
7.12604344e-01 -6.03733778e-01 -4.03940082e-01 3.13927829e-01
-9.91926610e-01 -1.61699578e-01 -4.97401536e-01 4.64807093e-01
-3.14938009e-01 3.07843089e-01 -4.99293625e-01 -5.30989647e-01
-4.85407054e-01 -1.40372348e+00 5.36922753e-01 2.57841703e-02
-3.13649386e-01 -9.08442318e-01 -1.68862611e-01 2.18958229e-01
8.74935329e-01 4.37099576e-01 8.77204597e-01 -6.46653175e-01
-6.10700607e-01 -8.48776624e-02 -6.44274428e-02 2.52397716e-01
1.72637716e-01 -3.62976938e-01 -1.09272766e+00 -8.74103308e-02
3.20123732e-01 8.18307996e-02 5.57405710e-01 6.95620239e-01
1.45795536e+00 -1.21592596e-01 -4.03028101e-01 6.99471533e-01
1.44914126e+00 3.40918779e-01 8.44963014e-01 4.17315006e-01
5.45039415e-01 3.10916722e-01 2.54304968e-02 3.24323088e-01
5.89306988e-02 6.59094095e-01 6.93711042e-01 -4.01232868e-01
-3.74500930e-01 -3.01377382e-03 -2.80474812e-01 6.49082363e-01
-5.07386208e-01 2.76605099e-01 -9.39554691e-01 4.72874314e-01
-1.65903771e+00 -7.46180058e-01 -5.03547907e-01 2.26793146e+00
7.55300045e-01 1.02649100e-01 -1.93547547e-01 1.46613926e-01
5.95875919e-01 -4.24983427e-02 -4.31478471e-01 -1.11255072e-01
1.57754064e-01 5.86752236e-01 7.38008142e-01 3.29489589e-01
-8.14587116e-01 3.85153085e-01 5.95638561e+00 6.04748785e-01
-1.54809785e+00 6.31557941e-01 4.46416557e-01 -1.31248474e-01
-2.45682731e-01 -3.53840292e-01 -6.57167360e-02 5.12690067e-01
7.61419177e-01 1.66398808e-01 2.51131117e-01 7.84908116e-01
3.85408431e-01 -3.50633115e-01 -1.08744502e+00 7.26597428e-01
1.04105987e-01 -1.41388023e+00 -3.30371559e-01 -1.15636513e-01
2.98913240e-01 5.23747951e-02 -1.70459688e-01 -2.07309760e-02
-5.83868660e-02 -1.08711267e+00 8.93314540e-01 5.75690627e-01
7.95751274e-01 -4.28716630e-01 7.36948788e-01 1.94752574e-01
-7.83888578e-01 2.17981115e-01 -4.26131487e-02 2.04841509e-01
2.63100803e-01 9.17025447e-01 -9.28564787e-01 7.35287428e-01
6.81176424e-01 3.48161429e-01 -4.11242098e-01 1.29849613e+00
-2.73025990e-01 5.89660823e-01 -2.70725876e-01 4.52000231e-01
-1.59975905e-02 1.45491913e-01 7.42339134e-01 1.07668507e+00
4.06078726e-01 -1.17413579e-02 -1.94361761e-01 8.73408675e-01
1.30991280e-01 8.43535140e-02 -5.77709556e-01 6.87442303e-01
1.22935653e-01 1.27874351e+00 -9.23163950e-01 -3.05562079e-01
-2.09440485e-01 8.36497545e-01 -5.00600897e-02 1.40306756e-01
-9.51876223e-01 8.97854939e-02 8.49837065e-02 5.37930310e-01
3.20977598e-01 2.22901419e-01 -4.63000983e-01 -1.14811254e+00
2.12099388e-01 -5.84553301e-01 -5.59038250e-03 -9.45655107e-01
-9.60046232e-01 8.85475814e-01 -1.42209738e-01 -1.34979701e+00
-1.79316327e-01 -3.64325881e-01 -6.01598799e-01 6.81053340e-01
-1.44464457e+00 -9.05146122e-01 -5.04810572e-01 6.66605771e-01
3.03957850e-01 2.25541338e-01 7.83765435e-01 5.92299402e-01
-1.89576462e-01 2.10292518e-01 -1.10770509e-01 -9.96467397e-02
5.52718878e-01 -9.80495632e-01 -2.19895169e-01 7.97324657e-01
-1.65693313e-01 4.17430609e-01 8.36431146e-01 -7.69218683e-01
-1.05733204e+00 -1.10117710e+00 6.15856171e-01 -1.24279059e-01
5.96685946e-01 -1.10822663e-01 -1.09297824e+00 7.71286905e-01
4.94683832e-02 2.60105073e-01 3.38734895e-01 -1.93444490e-01
-4.09095399e-02 -1.20040648e-01 -1.38581896e+00 2.49907374e-01
8.99144173e-01 -2.05250233e-01 -6.03847146e-01 4.12362486e-01
4.80814695e-01 -7.77403593e-01 -8.35165024e-01 4.30064529e-01
5.31135857e-01 -1.27951121e+00 1.01680672e+00 -3.45240295e-01
5.56143165e-01 2.56944690e-02 2.67790020e-01 -1.39469993e+00
-2.22482458e-01 -2.27145895e-01 -3.84405395e-03 5.71447611e-01
3.43748868e-01 -7.40310311e-01 6.82030737e-01 5.40518641e-01
-4.18176293e-01 -6.74483597e-01 -1.34776366e+00 -6.55681252e-01
-1.31046847e-01 -7.38326252e-01 6.69484079e-01 1.13700235e+00
-2.88510740e-01 -2.56171674e-01 -4.65389550e-01 2.34282479e-01
6.91542447e-01 -2.75671273e-01 2.02728882e-01 -1.32275152e+00
-3.78569961e-01 -2.72432774e-01 -2.41967708e-01 -3.35515261e-01
-2.96636283e-01 -8.41589391e-01 -9.15976521e-03 -1.45214689e+00
5.16278520e-02 -6.88338578e-01 -3.66227359e-01 5.20389557e-01
2.95201898e-01 4.69719589e-01 6.69176728e-02 1.92455709e-01
-4.39081043e-02 1.34939343e-01 1.45714450e+00 -5.60211763e-03
-1.77345380e-01 2.95569971e-02 -3.52830291e-01 1.06253994e+00
5.17183959e-01 -5.95523357e-01 -2.10578337e-01 -3.40331495e-01
2.63103873e-01 6.22859478e-01 7.02659249e-01 -1.24520993e+00
2.29204416e-01 3.01990390e-01 5.07059097e-01 -4.77458030e-01
2.06408933e-01 -1.24812365e+00 6.90279901e-01 8.31156254e-01
8.62123892e-02 -9.78327692e-02 1.50371715e-01 2.14460909e-01
-1.34274084e-02 -4.01273966e-01 8.88870478e-01 -3.82188499e-01
-2.41042316e-01 8.66016001e-02 -5.78106940e-01 -1.30344614e-01
9.04382467e-01 -2.55784988e-01 -1.42721012e-01 -1.53546780e-01
-1.17793548e+00 -1.57019213e-01 1.16320953e-01 1.42043561e-01
6.52562261e-01 -9.31371152e-01 -3.54839325e-01 2.04851627e-01
-1.96513027e-01 7.58327171e-02 6.34564281e-01 1.48911309e+00
-7.24650204e-01 3.68571758e-01 -7.00763613e-02 -6.33281112e-01
-1.12218297e+00 3.49388182e-01 7.80175865e-01 -6.16016746e-01
-1.12900603e+00 6.63483858e-01 1.15794756e-01 -2.04289615e-01
1.12529524e-01 -5.03630817e-01 -2.50832904e-02 -4.34667803e-02
4.96369243e-01 2.97812462e-01 8.34792852e-01 -3.40997547e-01
-4.03542310e-01 2.88396418e-01 -1.99124321e-01 1.01487964e-01
1.60302615e+00 7.48530403e-02 -1.55127600e-01 2.37896934e-01
6.33742988e-01 -1.12101629e-01 -7.91935623e-01 -1.93310678e-01
-2.64182329e-01 -2.55873442e-01 3.54547471e-01 -1.16965175e+00
-1.43103981e+00 6.65247440e-01 8.14273119e-01 8.16028342e-02
1.27334547e+00 -1.81939080e-01 6.47902787e-01 -1.96118161e-01
7.23417640e-01 -7.64122427e-01 -1.84038356e-01 3.37506011e-02
9.30252135e-01 -8.26939762e-01 4.30461019e-02 -3.52326691e-01
-5.34143150e-01 1.14674938e+00 3.98544848e-01 -7.96948969e-02
7.22403705e-01 5.57937622e-01 -1.16803711e-02 -3.42708737e-01
-3.52944613e-01 2.07117394e-01 6.89507574e-02 4.82000321e-01
3.48188221e-01 4.50771824e-02 -4.07735765e-01 6.27805293e-01
-1.33081824e-01 5.31995535e-01 7.82761574e-01 1.00714719e+00
-2.23657623e-01 -7.96790659e-01 -5.85678339e-01 6.16061568e-01
-5.45322776e-01 -1.86185181e-01 2.24415734e-01 9.65776622e-01
4.48731869e-01 4.21927273e-01 -5.93206398e-02 -4.84849662e-02
3.87427866e-01 -9.36784409e-03 6.28710568e-01 -4.05075818e-01
-1.04299378e+00 1.29934058e-01 -4.04080153e-02 -7.55240142e-01
-3.71900022e-01 -5.90967238e-01 -1.14519155e+00 -7.21060187e-02
-4.07918155e-01 -5.39207757e-02 9.08215523e-01 1.04496896e+00
5.94600625e-02 1.23497486e+00 2.16682047e-01 -7.52605200e-01
-2.94748247e-01 -1.01316142e+00 -6.22726977e-01 2.12228939e-01
3.77108991e-01 -8.19487572e-01 -2.38853827e-01 -3.00222896e-02]
|
[13.571914672851562, -2.519087076187134]
|
65d4a0b4-4c37-455f-b21b-1e0409761be2
|
confronting-ambiguity-in-6d-object-pose
|
2305.15873
| null |
https://arxiv.org/abs/2305.15873v1
|
https://arxiv.org/pdf/2305.15873v1.pdf
|
Confronting Ambiguity in 6D Object Pose Estimation via Score-Based Diffusion on SE(3)
|
Addressing accuracy limitations and pose ambiguity in 6D object pose estimation from single RGB images presents a significant challenge, particularly due to object symmetries or occlusions. In response, we introduce a novel score-based diffusion method applied to the $SE(3)$ group, marking the first application of diffusion models to $SE(3)$ within the image domain, specifically tailored for pose estimation tasks. Extensive evaluations demonstrate the method's efficacy in handling pose ambiguity, mitigating perspective-induced ambiguity, and showcasing the robustness of our surrogate Stein score formulation on $SE(3)$. This formulation not only improves the convergence of Langevin dynamics but also enhances computational efficiency. Thus, we pioneer a promising strategy for 6D object pose estimation.
|
['Chun-Yi Lee', 'Hsuan-Kung Yang', 'Hao-Wei Chen', 'Tsu-Ching Hsiao']
|
2023-05-25
| null | null | null | null |
['6d-pose-estimation']
|
['computer-vision']
|
[ 5.05261011e-02 -7.21141621e-02 2.53399014e-01 -2.34984696e-01
-1.11060154e+00 -7.07646370e-01 2.91673213e-01 -2.25996330e-01
-5.22312462e-01 3.45488369e-01 -7.08187670e-02 3.20654631e-01
-5.01740277e-01 -2.84332752e-01 -5.72318792e-01 -7.35580981e-01
-1.67077124e-01 6.35644138e-01 -1.29842060e-02 1.19000554e-01
4.47355688e-01 9.28051889e-01 -1.11337066e+00 -2.92915583e-01
5.69454491e-01 1.33126211e+00 1.50969952e-01 5.68843961e-01
2.41369441e-01 2.87582010e-01 -6.39024198e-01 -4.02775228e-01
5.73159993e-01 -2.37019882e-02 -5.79325199e-01 1.79174706e-01
8.50769460e-01 -5.54318964e-01 -2.50523031e-01 1.04156971e+00
6.73918903e-01 2.44745046e-01 8.15519154e-01 -8.42689872e-01
-4.37961817e-01 1.84396598e-02 -6.00304365e-01 1.37950450e-01
3.84311408e-01 2.88708657e-01 8.16297948e-01 -1.20255876e+00
9.87927318e-01 1.18812323e+00 8.14063668e-01 6.57851636e-01
-1.21006942e+00 -2.98501670e-01 3.25598538e-01 -8.77351761e-02
-1.57410371e+00 -3.75843912e-01 1.05259633e+00 -5.02492607e-01
1.02539933e+00 2.38531098e-01 7.99963832e-01 1.16562688e+00
1.24928690e-01 7.98351288e-01 1.24950361e+00 -1.03983708e-01
3.08843374e-01 -2.22234860e-01 1.46702975e-02 6.50494516e-01
2.30457783e-01 5.40686846e-02 -1.04285502e+00 -2.59142756e-01
9.47566509e-01 -3.37849319e-01 -5.05330898e-02 -1.05099750e+00
-1.35491168e+00 5.84561646e-01 4.30253983e-01 -3.21160018e-01
-2.97382772e-01 5.80868185e-01 1.91733707e-02 -1.83231428e-01
6.62333250e-01 3.90677810e-01 -3.08060497e-01 -3.61531496e-01
-7.26836503e-01 5.78721702e-01 3.58453423e-01 1.19449639e+00
3.32900047e-01 9.66422334e-02 2.58788932e-02 5.17699122e-01
5.99082232e-01 7.72988141e-01 -2.24451631e-01 -1.40640271e+00
4.74846333e-01 3.76518697e-01 5.02106190e-01 -8.54495406e-01
-6.94509745e-01 -7.85401464e-01 -6.16579771e-01 4.95757520e-01
5.17447054e-01 7.42861927e-02 -8.04807782e-01 1.83526373e+00
7.27837980e-01 -2.30649456e-01 -1.76209301e-01 1.08430803e+00
5.92495501e-01 1.87890291e-01 -1.64808810e-01 -3.93874139e-01
1.11885345e+00 -5.62635481e-01 -3.54034573e-01 -5.06297201e-02
2.74109274e-01 -1.00103772e+00 6.74280286e-01 5.14080107e-01
-1.40415490e+00 -2.44112745e-01 -8.70354950e-01 -1.63688704e-01
7.68274516e-02 2.03117177e-01 6.89237535e-01 6.55904710e-01
-1.07738745e+00 6.18952930e-01 -9.68085766e-01 -1.85446590e-01
4.60997343e-01 8.07327271e-01 -1.04328789e-01 3.56671400e-03
-4.75039303e-01 1.24588013e+00 -1.88515544e-01 4.38593835e-01
-7.79020250e-01 -9.08498824e-01 -5.86067140e-01 -5.54416239e-01
4.69379306e-01 -9.38444674e-01 1.05330706e+00 -1.64216429e-01
-1.64950442e+00 1.05192757e+00 -2.13077500e-01 -2.76743174e-01
1.07290113e+00 -3.39711517e-01 2.01706588e-01 2.42338523e-01
-7.72638842e-02 8.07647824e-01 1.32750881e+00 -1.28122580e+00
2.15920564e-02 -8.45680177e-01 -9.79239345e-02 5.10710895e-01
-6.90683955e-03 -1.46874130e-01 -2.93784589e-01 -7.13288605e-01
7.58075416e-01 -1.16740549e+00 -3.54944438e-01 4.57345396e-01
-4.10985341e-03 -1.82093725e-01 7.22168446e-01 -4.57490116e-01
7.32455850e-01 -2.11952758e+00 4.51827645e-01 3.15357327e-01
3.69922549e-01 -2.26778183e-02 -8.29717517e-02 7.12171942e-02
3.11577588e-01 -1.90687060e-01 -4.15525824e-01 -5.09663224e-01
9.86168385e-02 -3.16990577e-02 -1.14287704e-01 8.23848665e-01
4.93771136e-01 1.06600571e+00 -6.85189426e-01 -2.14125261e-01
2.10763171e-01 5.40355384e-01 -7.39694834e-01 -1.23340219e-01
-2.21368775e-01 7.36878037e-01 -5.24856269e-01 8.63917232e-01
9.40985560e-01 -1.32082939e-01 -1.02272935e-01 -6.29491091e-01
-2.15111658e-01 -3.96116897e-02 -1.55401945e+00 2.02104735e+00
-2.45057344e-01 2.77995527e-01 3.60655755e-01 -4.75482911e-01
8.93232167e-01 -1.93765983e-01 1.04860914e+00 -4.52125490e-01
9.48219523e-02 2.88574547e-01 -2.77896911e-01 -3.01061124e-01
4.37817097e-01 -3.06005478e-01 -8.62922445e-02 3.78993392e-01
8.55147168e-02 -7.73741603e-01 -2.30849802e-01 -6.01508701e-03
8.87689412e-01 4.64894027e-01 5.11001155e-04 -4.16031569e-01
2.06298426e-01 -5.71551844e-02 1.80116385e-01 6.67144597e-01
-5.99914432e-01 8.83138001e-01 8.48502889e-02 -2.06390515e-01
-1.04742527e+00 -1.45745575e+00 -4.41292703e-01 3.46232086e-01
4.28336948e-01 -1.91242874e-01 -5.78082442e-01 -7.52817273e-01
4.30604517e-01 3.63537371e-01 -6.07540250e-01 -1.81337342e-01
-7.31330931e-01 -9.35640633e-01 2.79806703e-01 5.89913666e-01
4.26486731e-01 -3.45547795e-01 -6.93225145e-01 1.84169501e-01
-1.39274433e-01 -1.06545734e+00 -5.88911891e-01 -6.25374764e-02
-1.12210047e+00 -9.80716884e-01 -1.03066456e+00 -2.31287315e-01
6.33054376e-01 2.83325851e-01 8.61239791e-01 -5.02044678e-01
-3.82639021e-01 1.08997393e+00 5.63947149e-02 -3.02844733e-01
1.16997212e-01 -1.02206700e-01 4.25660163e-01 -9.90870595e-02
4.18193219e-03 -2.26112813e-01 -9.89994526e-01 5.70489109e-01
-5.24890900e-01 -5.23030758e-01 3.34374309e-01 4.42725688e-01
6.44346058e-01 -3.44522327e-01 1.13584645e-01 -7.85738043e-03
4.37147439e-01 1.51876718e-01 -8.77295196e-01 -7.22449869e-02
-7.14211762e-01 8.26233327e-02 -2.09190667e-01 -6.24334574e-01
-1.20277655e+00 3.43282640e-01 -1.08386770e-01 -5.36608458e-01
2.12069437e-01 2.30391651e-01 1.44120216e-01 -8.01267147e-01
5.80619454e-01 -2.14314386e-01 2.50964284e-01 -6.48630559e-01
3.13961148e-01 1.40313447e-01 2.68458188e-01 -7.28067756e-01
8.04594159e-01 1.03125095e+00 5.42828918e-01 -8.68671477e-01
-6.08145535e-01 -4.07406747e-01 -7.97580898e-01 -4.21898931e-01
7.81514466e-01 -8.96185875e-01 -1.15338969e+00 6.29435062e-01
-1.18680584e+00 -3.57304066e-02 -5.71994245e-01 7.23428011e-01
-8.50060165e-01 5.58071375e-01 -3.66667867e-01 -9.90338266e-01
-3.45286690e-02 -1.27631867e+00 1.31734717e+00 -1.16983250e-01
-3.72185677e-01 -6.85877085e-01 -8.67256448e-02 6.28928661e-01
3.36061835e-01 2.67561108e-01 5.66876531e-01 5.36333174e-02
-1.06254363e+00 -1.51092231e-01 -2.04000682e-01 2.77140945e-01
-4.08544421e-01 -1.20726630e-01 -7.75229990e-01 -3.84017885e-01
4.23663259e-01 -2.98309356e-01 6.31353915e-01 5.63068569e-01
6.83095574e-01 1.51143342e-01 8.69500544e-03 6.25693202e-01
1.19054031e+00 -9.15790647e-02 2.39081815e-01 1.97723985e-01
6.25733197e-01 6.19222760e-01 7.05258429e-01 6.04758978e-01
2.32113957e-01 1.04266334e+00 6.58706486e-01 2.45972246e-01
-4.76276517e-01 -1.30649671e-01 3.94009173e-01 8.40772331e-01
-2.72919148e-01 9.13295895e-02 -8.56672704e-01 2.35910028e-01
-1.65605259e+00 -3.73991787e-01 -3.56891543e-01 2.24689960e+00
3.60380709e-01 -4.62483754e-03 6.22019880e-02 -3.84186387e-01
2.84825414e-01 1.83328927e-01 -9.45133388e-01 6.17211163e-02
-1.67937413e-01 3.70221317e-01 6.28344357e-01 6.41836703e-01
-7.87999392e-01 9.22890663e-01 7.63830471e+00 8.48990023e-01
-8.04172397e-01 2.37056509e-01 1.50152653e-01 -6.03785098e-01
-2.24164531e-01 -1.56923980e-01 -1.10092461e+00 1.02226518e-01
3.32869172e-01 2.21634403e-01 2.71557510e-01 6.85910046e-01
1.39536709e-01 -3.18364263e-01 -9.67389464e-01 1.28246248e+00
3.42554688e-01 -1.24029446e+00 -6.74536675e-02 1.60722777e-01
8.28323543e-01 1.42763555e-01 5.01233995e-01 -2.91912735e-01
-1.17405750e-01 -6.63616598e-01 1.03735828e+00 4.32009816e-01
7.38596439e-01 -5.30182779e-01 2.74490058e-01 1.40540957e-01
-9.14707959e-01 1.16443209e-01 -2.62769639e-01 8.44160914e-02
4.70088840e-01 6.62468433e-01 -6.86629176e-01 3.04941714e-01
7.31172442e-01 5.60748041e-01 -4.06175822e-01 8.47312748e-01
-5.12405857e-02 -1.05185755e-01 -7.24779606e-01 -8.72497633e-03
8.08398575e-02 -4.03255731e-01 1.09303319e+00 7.46425271e-01
6.10786915e-01 3.24128896e-01 -2.62615561e-01 1.10527003e+00
6.77423477e-02 -3.24234307e-01 -4.34735924e-01 1.07218787e-01
-7.71845430e-02 8.78905356e-01 -8.25602531e-01 4.80317064e-02
9.09385756e-02 1.03382158e+00 1.79353997e-01 4.44753617e-01
-7.69371510e-01 2.40856096e-01 8.47633302e-01 8.33331943e-02
4.16642427e-01 -1.11836433e+00 -5.49867034e-01 -1.35718715e+00
4.40130681e-01 -5.41641116e-01 7.29115307e-02 -9.90726948e-01
-1.17401600e+00 5.80752753e-02 2.77047038e-01 -1.15705347e+00
2.40733009e-02 -9.85552847e-01 2.40556210e-01 6.41902983e-01
-1.21675217e+00 -1.10140312e+00 -1.07060142e-01 4.68216002e-01
4.80760455e-01 3.75698954e-01 5.40475488e-01 1.32816330e-01
-9.79394317e-02 2.97347367e-01 1.09101638e-01 -5.71942806e-01
5.12044787e-01 -9.62976992e-01 2.46833250e-01 5.66544712e-01
3.72001603e-02 7.94434905e-01 7.23604739e-01 -5.48737228e-01
-1.94633770e+00 -5.33362627e-01 5.86531997e-01 -1.11599278e+00
5.46364665e-01 -5.96577108e-01 -4.19399887e-01 4.11171377e-01
-4.42119420e-01 6.58041164e-02 2.85770029e-01 -5.81314601e-02
-3.25164169e-01 -2.41961107e-01 -1.37833345e+00 5.42677820e-01
1.64783847e+00 -6.57469392e-01 -3.37368399e-01 3.18480611e-01
3.58328372e-01 -1.03614223e+00 -1.05783844e+00 7.86387205e-01
7.31645644e-01 -8.11360180e-01 1.46640742e+00 -2.17018723e-01
7.40665570e-03 -2.41324052e-01 -3.26914638e-01 -9.74578083e-01
-1.41592324e-01 -8.94791961e-01 -3.59796375e-01 8.97107899e-01
5.26630878e-02 -4.70973283e-01 1.06383324e+00 8.72505546e-01
1.38134398e-02 -5.99757135e-01 -1.59445870e+00 -8.74366701e-01
2.32950728e-02 -7.18171239e-01 1.58414498e-01 5.05860567e-01
-6.25509202e-01 -2.05195919e-01 -2.15972170e-01 3.47719669e-01
1.12142408e+00 5.66125512e-02 6.48383617e-01 -9.20319915e-01
-3.19432884e-01 -4.68673497e-01 -4.42295551e-01 -1.55261171e+00
-1.60078511e-01 -6.00442410e-01 -7.69117996e-02 -1.08610010e+00
7.19755888e-02 -5.85353315e-01 -1.25529110e-01 -2.28113741e-01
1.58410385e-01 6.50115907e-01 4.10944223e-01 2.78210431e-01
-5.29221654e-01 6.73426926e-01 1.46641350e+00 -6.47062659e-02
4.70651761e-02 1.90529644e-01 -3.74695271e-01 9.02144074e-01
2.69717038e-01 -4.03897256e-01 -1.52820021e-01 -6.40593290e-01
6.33514345e-01 -1.35810927e-01 7.83587277e-01 -8.31927598e-01
5.28418459e-02 -1.07943542e-01 3.16236645e-01 -7.27896988e-01
7.31007457e-01 -9.52680826e-01 2.64690042e-01 4.74104047e-01
-9.83473361e-02 -1.17716081e-02 1.07760563e-01 7.52649724e-01
9.95418578e-02 -2.21186563e-01 7.69466460e-01 -2.10120529e-01
-6.55429602e-01 5.27254820e-01 -2.69995555e-02 3.28903012e-02
9.79330003e-01 -4.06993777e-01 3.09897289e-02 -2.94033617e-01
-1.03323066e+00 -8.29159394e-02 5.83364367e-01 2.72029161e-01
7.30805516e-01 -1.23324108e+00 -3.59014660e-01 3.06730419e-01
1.88100457e-01 3.07778239e-01 2.91338652e-01 1.18845582e+00
-5.46663582e-01 2.40306079e-01 1.61260173e-01 -1.10044527e+00
-1.18696475e+00 1.59484997e-01 3.01249266e-01 -7.37778023e-02
-4.17077988e-01 1.37765777e+00 -2.45861802e-02 -3.75682831e-01
3.53388280e-01 -5.53207994e-01 5.39085567e-01 7.75160119e-02
-1.52853969e-02 7.89209545e-01 2.06232086e-01 -6.18063748e-01
-5.91891885e-01 1.26007426e+00 1.13605723e-01 -4.75473434e-01
1.26420295e+00 -4.13096815e-01 6.03927672e-02 2.04722032e-01
1.20631540e+00 6.50527775e-02 -1.87976468e+00 8.07885528e-02
-2.14940667e-01 -5.49689293e-01 -1.48002030e-02 -7.57883310e-01
-6.68726325e-01 8.93098533e-01 9.19107497e-01 -3.59631151e-01
6.16301477e-01 2.31273919e-01 5.75292170e-01 2.29752108e-01
6.87993884e-01 -1.06247282e+00 4.97517616e-01 4.96727198e-01
1.11964166e+00 -1.17992699e+00 2.57535696e-01 -5.97353280e-01
-3.99228603e-01 9.55536187e-01 3.13634217e-01 -2.79497743e-01
5.87678373e-01 1.43743247e-01 7.24336058e-02 -4.23542410e-01
-5.14063314e-02 -3.07769030e-02 4.58826154e-01 6.89881146e-01
9.37276483e-02 -2.08161816e-01 -3.29259366e-01 8.12368616e-02
4.23772782e-02 -3.10458153e-01 1.10392375e-02 1.05964088e+00
-1.42892301e-01 -1.11493540e+00 -4.66939896e-01 5.24130128e-02
-1.18010052e-01 2.05522552e-01 -3.96833867e-01 7.75069594e-01
1.64412335e-01 5.18692434e-01 -1.41017839e-01 3.98599803e-02
5.50108612e-01 -4.12825719e-02 1.33252597e+00 -2.90104300e-01
-4.79617655e-01 3.34898651e-01 -1.73780650e-01 -8.30963016e-01
-6.62221074e-01 -9.08616126e-01 -1.00252616e+00 -7.96509385e-02
-3.62242848e-01 -3.20518941e-01 1.03118658e+00 8.46007526e-01
6.40407264e-01 8.57501402e-02 2.57266819e-01 -1.10426784e+00
-1.13363421e+00 -5.03545403e-01 -5.22037745e-01 4.08764273e-01
4.30156618e-01 -1.01858556e+00 -3.98684144e-01 -2.92885572e-01]
|
[7.2125654220581055, -2.3331751823425293]
|
475b2018-4718-4f71-8a19-9ca24db08c77
|
self-supervised-machine-learning-model-for
|
2203.13875
| null |
https://arxiv.org/abs/2203.13875v2
|
https://arxiv.org/pdf/2203.13875v2.pdf
|
Semi-supervised machine learning model for analysis of nanowire morphologies from transmission electron microscopy images
|
In the field of materials science, microscopy is the first and often only accessible method for structural characterization. There is a growing interest in the development of machine learning methods that can automate the analysis and interpretation of microscopy images. Typically training of machine learning models requires large numbers of images with associated structural labels, however, manual labeling of images requires domain knowledge and is prone to human error and subjectivity. To overcome these limitations, we present a semi-supervised transfer learning approach that uses a small number of labeled microscopy images for training and performs as effectively as methods trained on significantly larger image datasets. Specifically, we train an image encoder with unlabeled images using self-supervised learning methods and use that encoder for transfer learning of different downstream image tasks (classification and segmentation) with a minimal number of labeled images for training. We test the transfer learning ability of two self-supervised learning methods: SimCLR and Barlow-Twins on transmission electron microscopy (TEM) images. We demonstrate in detail how this machine learning workflow applied to TEM images of protein nanowires enables automated classification of nanowire morphologies (e.g., single nanowires, nanowire bundles, phase separated) as well as segmentation tasks that can serve as groundwork for quantification of nanowire domain sizes and shape analysis. We also extend the application of the machine learning workflow to classification of nanoparticle morphologies and identification of different type of viruses from TEM images.
|
['Arthi Jayaraman', 'Todd Emrick', 'Brian Montz', 'Shizhao Lu']
|
2022-03-25
| null | null | null | null |
['morphology-classification']
|
['computer-vision']
|
[ 8.18100393e-01 1.26036450e-01 1.34438887e-01 -6.20617926e-01
-5.91266215e-01 -8.24310124e-01 3.04800242e-01 4.45945770e-01
-7.43144870e-01 7.10849047e-01 -5.59304774e-01 -6.52757704e-01
2.82728702e-01 -8.16425025e-01 -8.97527456e-01 -9.10220325e-01
2.62388825e-01 9.64547932e-01 3.55420887e-01 2.32426792e-01
2.63703585e-01 5.46259761e-01 -1.27474117e+00 5.65536201e-01
4.99839962e-01 1.14194298e+00 7.19246507e-01 7.57033288e-01
-2.66398966e-01 7.27983654e-01 -2.11067989e-01 -1.11678183e-01
7.68246129e-02 -1.65330693e-01 -1.10856795e+00 4.18953449e-01
7.60984197e-02 -1.74258992e-01 4.75866735e-01 9.27019954e-01
2.81427264e-01 -4.91364390e-01 1.05463827e+00 -7.64341235e-01
-8.14131141e-01 5.46844184e-01 -2.74498045e-01 2.39663929e-01
-1.12452552e-01 1.32375047e-01 6.98923111e-01 -7.45742857e-01
7.85748720e-01 7.63241768e-01 6.57597303e-01 5.62797785e-01
-1.69877875e+00 -5.43218136e-01 -3.86000037e-01 1.67757645e-01
-9.26618934e-01 -3.22412878e-01 5.33560336e-01 -8.56631696e-01
1.02700305e+00 -1.34793982e-01 4.73175496e-01 8.96905839e-01
1.33222997e-01 3.46352756e-01 1.40782130e+00 -5.38237929e-01
4.26632851e-01 4.25172687e-01 1.75259426e-01 8.77934277e-01
1.34133622e-01 -1.60640225e-01 9.06449854e-02 1.51372135e-01
8.35879862e-01 1.15062952e-01 3.46477538e-01 -3.04424375e-01
-1.16431868e+00 8.57453167e-01 1.12618923e-01 4.48966801e-01
-1.04561381e-01 -1.24396481e-01 5.40812790e-01 2.56899565e-01
5.20748019e-01 5.25807440e-01 -5.89068472e-01 9.28378776e-02
-7.75111914e-01 -3.30481917e-01 6.67819798e-01 7.75846541e-01
1.26281238e+00 -1.83768451e-01 4.07362670e-01 7.24830031e-01
1.10775948e-01 5.92801094e-01 4.45233285e-01 -8.76448452e-01
-3.99988815e-02 8.00178826e-01 -3.19944248e-02 -3.77151906e-01
-4.88944292e-01 3.90633523e-01 -6.80716097e-01 5.25954485e-01
5.79712570e-01 -7.91099593e-02 -9.10821557e-01 1.27604878e+00
8.63134116e-02 -5.78126431e-01 6.36523739e-02 5.72780848e-01
7.67970145e-01 4.86510247e-01 -1.34821728e-01 -2.01055035e-01
1.17833877e+00 -8.22937548e-01 -4.48353231e-01 -6.46990165e-03
7.08130002e-01 -6.85542822e-01 8.60305548e-01 2.64255911e-01
-8.89916182e-01 -4.07506883e-01 -8.70533466e-01 4.38418165e-02
-8.21303546e-01 4.54054743e-01 4.72867668e-01 6.19685352e-01
-8.17622900e-01 8.29406500e-01 -1.25602555e+00 -8.06772709e-01
8.56535852e-01 7.44932353e-01 -6.41777217e-01 1.53157875e-01
-2.38394216e-01 9.00840640e-01 3.25272322e-01 -3.58787030e-01
-8.33659410e-01 -5.51258087e-01 -7.62641847e-01 -2.99734265e-01
-1.73976272e-01 -2.10980400e-01 1.11202919e+00 -9.82462406e-01
-1.53721869e+00 1.43424487e+00 -1.18218169e-01 -4.28485751e-01
1.01581097e-01 3.51638407e-01 6.34642020e-02 3.89448106e-01
2.43387297e-01 7.36617625e-01 8.77373815e-01 -1.37622476e+00
-5.11990428e-01 -2.92142183e-01 -3.29102963e-01 -4.71436918e-01
-5.50864875e-01 9.86784920e-02 3.69876951e-01 4.16589482e-03
-1.06623925e-01 -7.73808897e-01 -8.33508670e-02 -5.39403893e-02
-3.48317981e-01 -1.99837282e-01 1.21468174e+00 -5.20138025e-01
4.89280492e-01 -1.76685476e+00 5.09037338e-02 2.10381374e-01
3.86071891e-01 2.05811620e-01 -1.72962565e-02 4.08911109e-01
6.38558483e-03 1.36706293e-01 -3.86406392e-01 -2.79316247e-01
-2.23633274e-01 3.01576972e-01 1.38482511e-01 5.31314492e-01
3.62364560e-01 9.75486040e-01 -7.85287797e-01 -5.40798366e-01
5.27057767e-01 1.63036093e-01 -2.35041827e-01 4.79243606e-01
-1.40145212e-01 1.06712663e+00 -1.63709551e-01 6.63193464e-01
4.06162769e-01 -9.89861488e-01 1.87475964e-01 -5.40360689e-01
-5.33854306e-01 2.34898757e-02 -4.74682987e-01 1.34222448e+00
-3.44102651e-01 5.73484540e-01 2.03523025e-01 -1.44488943e+00
9.49649096e-01 1.07580647e-01 6.68014824e-01 -4.86687869e-01
5.45525610e-01 3.13752443e-01 1.32717729e-01 -9.80867684e-01
-4.29893434e-01 -3.93556833e-01 3.97132218e-01 1.01607668e+00
3.38856369e-01 -2.58195460e-01 3.47811192e-01 -3.26286763e-01
1.21030021e+00 3.44025232e-02 1.14811622e-01 -3.69221389e-01
4.53032196e-01 3.30706686e-01 3.46675701e-02 3.85345578e-01
7.20354244e-02 5.29476702e-01 1.06163993e-01 -8.66479874e-01
-1.89765990e+00 -9.74641442e-01 -3.73075038e-01 1.13175833e+00
3.51018831e-02 -6.76457211e-02 -9.27435994e-01 -4.26268220e-01
-2.02881172e-01 -6.01394437e-02 -7.15927362e-01 1.62830040e-01
-3.95251274e-01 -9.54171598e-01 3.81330907e-01 6.83196783e-01
5.58120050e-02 -1.36460710e+00 -9.22972560e-01 -1.00010249e-03
1.84599414e-01 -1.46701539e+00 2.01391548e-01 8.58123958e-01
-7.38329649e-01 -1.30737484e+00 -4.39132869e-01 -1.20378673e+00
1.19434500e+00 -2.81880982e-02 8.18652153e-01 6.51096180e-02
-7.15086341e-01 3.68217409e-01 -4.59400594e-01 -5.27214110e-01
-8.00763488e-01 2.33778179e-01 -9.91568435e-03 -4.78982851e-02
3.12579542e-01 -7.03720868e-01 -6.08028054e-01 2.81301767e-01
-8.68260264e-01 2.58511811e-01 7.04878688e-01 9.51719224e-01
7.60913253e-01 3.70985568e-02 4.90981847e-01 -1.13897002e+00
4.07517195e-01 -2.27479905e-01 -5.85509241e-01 1.84757784e-01
-6.36450768e-01 4.01249006e-02 1.02339566e+00 -3.49136710e-01
-7.15653479e-01 3.36002290e-01 -2.33368605e-01 -3.66521329e-01
-5.92873693e-01 2.79654831e-01 1.65247202e-01 -3.24051678e-01
5.74004710e-01 1.11785874e-01 6.22028708e-01 -2.17143774e-01
4.98159043e-02 8.97371650e-01 3.71436566e-01 -5.35691738e-01
6.69245481e-01 9.72948253e-01 5.64868003e-02 -1.05055773e+00
-6.39718354e-01 -5.14106572e-01 -1.06836712e+00 -4.22432311e-02
1.47559774e+00 -3.93052727e-01 -6.94207013e-01 4.26667899e-01
-1.02960837e+00 -8.11938286e-01 -1.98341206e-01 2.20314458e-01
-6.43004477e-01 3.60900044e-01 -7.62092352e-01 -5.66547036e-01
-5.13279498e-01 -1.34083939e+00 1.15097368e+00 -4.36995588e-02
-2.85360307e-01 -1.43282223e+00 -4.99318466e-02 2.97400445e-01
3.81356508e-01 1.32266387e-01 1.43867040e+00 -7.89998114e-01
-2.07487434e-01 -1.70093760e-01 -3.47112805e-01 5.55344045e-01
6.32440388e-01 1.35695040e-01 -9.85566735e-01 -3.01202118e-01
-4.36151400e-02 -8.98047686e-01 7.07038522e-01 3.51328403e-01
1.19830346e+00 -8.89159143e-02 -5.29981494e-01 4.80641156e-01
1.38896942e+00 3.17002892e-01 4.71597642e-01 3.78531188e-01
7.80593872e-01 8.95928323e-01 1.10194549e-01 1.96110189e-01
8.90231058e-02 2.74142563e-01 3.21681440e-01 -5.22086620e-01
2.15626657e-01 2.36169994e-01 3.08733396e-02 7.58901060e-01
-3.18170130e-01 8.44350606e-02 -8.63699555e-01 2.91157275e-01
-1.61146963e+00 -8.31634164e-01 -3.48688383e-03 1.97899115e+00
8.78122389e-01 -3.98479030e-02 3.72323662e-01 3.20533961e-01
8.29616547e-01 -6.28213882e-01 -6.38465226e-01 -3.17574203e-01
-6.62021935e-02 5.83573997e-01 5.29806018e-01 2.21452385e-01
-1.22115743e+00 9.21427131e-01 6.92600346e+00 4.02425200e-01
-1.40374732e+00 9.80689377e-02 7.36143827e-01 4.54381585e-01
1.39991911e-02 -1.08810335e-01 -5.24188042e-01 4.36763078e-01
8.12562943e-01 4.14443105e-01 6.72882915e-01 6.17878854e-01
2.01994807e-01 -1.03685528e-01 -1.48067403e+00 8.62478912e-01
-2.15269238e-01 -1.70112681e+00 -1.54094115e-01 1.16369478e-01
6.14824831e-01 2.36575693e-01 -1.38875067e-01 -3.05678487e-01
3.39764386e-01 -1.22821236e+00 4.53865826e-01 1.03786036e-01
1.14583766e+00 -3.26529294e-01 7.36144006e-01 3.08336526e-01
-9.79469776e-01 -1.37543395e-01 -5.48196256e-01 -1.28505811e-01
8.05999860e-02 3.81660521e-01 -1.11559439e+00 -4.06632312e-02
9.65820193e-01 8.21395457e-01 -5.04443288e-01 2.84414113e-01
1.74679920e-01 6.39983714e-01 -1.90418333e-01 -5.07187545e-02
-3.52154002e-02 -5.72578311e-01 -1.06621057e-01 1.36281025e+00
7.06142709e-02 -1.81185782e-01 2.29945138e-01 1.11145234e+00
9.90518406e-02 -1.10029742e-01 -6.62992179e-01 -5.19022107e-01
2.65003115e-01 1.75953901e+00 -1.42733788e+00 -2.64879346e-01
-5.79230011e-01 5.80112576e-01 6.28765285e-01 5.00633754e-02
-4.52200800e-01 -5.04052863e-02 8.65361169e-02 4.39552873e-01
7.11011469e-01 -2.09798127e-01 -5.76413393e-01 -7.78471351e-01
-4.46237028e-01 -5.53387880e-01 -6.01117536e-02 -6.38982534e-01
-1.71750844e+00 4.63207752e-01 -3.43675166e-01 -9.16205525e-01
-2.24411320e-02 -1.32936406e+00 -7.86809266e-01 3.68418187e-01
-1.36202466e+00 -1.33368373e+00 -3.46109092e-01 2.89393425e-01
2.11017311e-01 -2.95634627e-01 1.11885655e+00 1.18382841e-01
-4.25137848e-01 9.75366384e-02 2.83076018e-01 3.76946986e-01
5.66766322e-01 -1.40322304e+00 3.68935019e-02 2.30689257e-01
-2.06103072e-01 3.93613309e-01 4.53884602e-01 -4.34168249e-01
-1.33487284e+00 -1.33548176e+00 4.57333118e-01 -4.57639575e-01
8.74629259e-01 -5.14498770e-01 -6.65305078e-01 7.16844380e-01
3.07841539e-01 6.99819252e-02 1.20435798e+00 -3.37549597e-01
-1.28703818e-01 7.34043568e-02 -1.24848974e+00 1.31484494e-01
5.55585027e-01 -5.85788906e-01 -2.99358428e-01 6.39356315e-01
2.05858186e-01 6.71500191e-02 -1.15439785e+00 3.84779513e-01
6.02276027e-01 -9.22406971e-01 6.68967187e-01 -5.65070689e-01
7.78808951e-01 -2.53214896e-01 -7.70750791e-02 -9.07217145e-01
-2.69797862e-01 -2.33267039e-01 4.59063768e-01 1.11851203e+00
6.09951019e-01 -4.32260096e-01 9.57119524e-01 3.44163090e-01
-1.03925072e-01 -7.46190548e-01 -5.83793104e-01 -6.84584022e-01
2.94969499e-01 1.45649947e-02 1.04998134e-01 7.79128492e-01
2.97405750e-01 6.15383387e-01 2.14938596e-01 -1.28644586e-01
6.08467937e-01 3.86861056e-01 5.62854707e-01 -1.30843532e+00
-4.98266481e-02 -1.92329660e-01 -5.33888996e-01 -5.51365256e-01
3.90727133e-01 -1.21035194e+00 1.25128850e-01 -1.57998085e+00
5.76578259e-01 -5.02574503e-01 7.90813640e-02 6.60576522e-01
4.05613989e-01 6.68449163e-01 -3.58041674e-01 5.12349367e-01
-7.39010572e-01 -6.34458195e-03 1.22209632e+00 -2.75757074e-01
2.13256553e-01 -1.96082965e-01 -2.33889759e-01 6.99829876e-01
8.92787576e-01 -4.10786897e-01 -1.51298478e-01 -4.87523675e-01
1.63068324e-01 -4.56929713e-01 4.56022322e-01 -9.04045880e-01
1.62996382e-01 7.06068054e-02 6.06768191e-01 -3.14333141e-01
-6.69498462e-03 -9.64853406e-01 -1.45700976e-01 4.80924875e-01
-3.07013482e-01 -4.61491290e-03 -1.21684454e-01 2.16760382e-01
1.91301599e-01 -4.86387283e-01 1.03756118e+00 -5.59281349e-01
-3.29257280e-01 1.89968362e-01 -1.04461253e+00 -3.29127729e-01
1.08927321e+00 -4.33125764e-01 -4.48870867e-01 9.39052030e-02
-8.19556892e-01 -7.99980536e-02 9.43676174e-01 -3.21537644e-01
5.98426223e-01 -9.41789806e-01 -3.24104458e-01 2.62076676e-01
1.63379937e-01 2.11367503e-01 -2.25711495e-01 9.43737864e-01
-6.35909736e-01 2.38815472e-01 -6.21006310e-01 -1.07892358e+00
-1.24561429e+00 6.41896963e-01 1.63943648e-01 -2.18243644e-01
-3.62077355e-01 5.60667574e-01 1.82505742e-01 -7.14254916e-01
-2.82212913e-01 -3.39017421e-01 -3.00093055e-01 -3.95359695e-01
2.77085215e-01 3.12648743e-01 2.99642384e-01 -5.28605223e-01
-7.83123225e-02 6.97318971e-01 -1.65786326e-01 1.35979414e-01
1.63796031e+00 -1.42024741e-01 -5.74309289e-01 5.98899364e-01
1.41191685e+00 -4.15983647e-01 -1.27217877e+00 1.09761700e-01
1.10017076e-01 2.95186043e-01 -3.27812910e-01 -4.77087557e-01
-8.01620543e-01 1.11811066e+00 5.48318684e-01 4.83537525e-01
8.35880220e-01 4.45112914e-01 7.01985121e-01 5.50923645e-01
3.32501501e-01 -1.11451793e+00 2.99351752e-01 4.68570620e-01
1.67613938e-01 -1.60624325e+00 -1.82940349e-01 -5.76455653e-01
-2.55421579e-01 1.49928212e+00 4.90495801e-01 -1.65708140e-01
6.19082987e-01 1.05989301e+00 3.15057516e-01 -4.90077168e-01
-6.69655561e-01 -2.25542381e-01 -1.49570599e-01 1.00287175e+00
6.54241145e-01 -1.30648360e-01 9.04090628e-02 4.84467089e-01
-2.05541357e-01 -2.61146482e-02 3.53659302e-01 1.19929516e+00
-6.93957031e-01 -1.36473274e+00 -9.24787968e-02 8.91290665e-01
-5.33201158e-01 1.40292123e-02 -5.98470330e-01 3.36290985e-01
3.64717871e-01 9.20965850e-01 3.29927802e-01 -5.53307950e-01
-1.73018143e-01 2.84624435e-02 9.29771781e-01 -9.96723652e-01
-3.62693846e-01 1.05650730e-01 -1.84548959e-01 5.88817261e-02
-9.09893692e-01 -5.66355944e-01 -1.55330849e+00 6.17165267e-02
-4.03728873e-01 1.79300100e-01 7.03616977e-01 1.40399683e+00
6.02643937e-02 3.73382956e-01 6.72882855e-01 -1.19069588e+00
-1.45981312e-01 -1.01605237e+00 -4.99182343e-01 7.26336062e-01
1.74931422e-01 -4.63800669e-01 -2.74472803e-01 9.72399950e-01]
|
[14.247958183288574, -2.9848368167877197]
|
2446638f-9ee1-4899-af06-7ea1208b512e
|
a-deep-learning-based-pipeline-for-efficient
|
1910.10549
| null |
https://arxiv.org/abs/1910.10549v3
|
https://arxiv.org/pdf/1910.10549v3.pdf
|
A Deep Learning based Pipeline for Efficient Oral Cancer Screening on Whole Slide Images
|
Oral cancer incidence is rapidly increasing worldwide. The most important determinant factor in cancer survival is early diagnosis. To facilitate large scale screening, we propose a fully automated pipeline for oral cancer detection on whole slide cytology images. The pipeline consists of fully convolutional regression-based nucleus detection, followed by per-cell focus selection, and CNN based classification. Our novel focus selection step provides fast per-cell focus decisions at human-level accuracy. We demonstrate that the pipeline provides efficient cancer classification of whole slide cytology images, improving over previous results both in terms of accuracy and feasibility. The complete source code is available at https://github.com/MIDA-group/OralScreen.
|
['Jan-Michaél Hirsch', 'Nataša Sladoje', 'Jiahao Lu', 'Christina Runow Stark', 'Joakim Lindblad', 'Eva Darai Ramqvist']
|
2019-10-23
| null | null | null | null |
['oral-cancer-classification']
|
['medical']
|
[ 2.26004943e-01 -2.08520025e-01 -5.86281955e-01 9.82135311e-02
-1.56242275e+00 -4.65429574e-01 2.09289446e-01 7.94349194e-01
-7.09205329e-01 6.39440417e-01 -2.96783098e-03 -6.45586669e-01
3.88700634e-01 -8.34725380e-01 -2.95941561e-01 -1.14417124e+00
3.24024975e-01 5.80831707e-01 2.52837360e-01 2.15445474e-01
2.17119455e-01 6.65391088e-01 -1.23263240e+00 4.18611050e-01
6.88967586e-01 7.49413133e-01 1.98647052e-01 1.41271091e+00
2.84012612e-02 4.35727596e-01 -3.79883528e-01 -1.02718160e-01
-4.95045215e-01 -2.16222093e-01 -6.89907491e-01 -1.05247609e-01
2.68051088e-01 -4.68023628e-01 -2.12751707e-04 9.26754534e-01
7.53578007e-01 -6.60004318e-01 8.11922789e-01 -7.13563859e-01
-7.58152306e-02 -3.31067629e-02 -4.12306756e-01 3.76837045e-01
5.92385307e-02 1.78420633e-01 8.10771227e-01 -7.85570860e-01
7.89112449e-01 8.00598085e-01 8.39308500e-01 7.82319784e-01
-1.09200072e+00 -4.93419617e-01 -5.74184120e-01 1.88468248e-02
-1.40975058e+00 -4.37347651e-01 -4.24386635e-02 -3.01000148e-01
9.40991104e-01 3.85243326e-01 9.86787319e-01 5.87899268e-01
6.48324847e-01 9.69980597e-01 8.93065572e-01 -5.44846892e-01
4.62935477e-01 4.85287346e-02 2.58927554e-01 9.48529780e-01
6.10836983e-01 -3.01524431e-01 -7.05429763e-02 -3.85494113e-01
6.33998394e-01 4.20648128e-01 -1.91621691e-01 3.81424755e-01
-7.89190531e-01 8.93086851e-01 4.57424283e-01 4.38345373e-01
-1.12838618e-01 1.93328187e-01 5.34544349e-01 -1.01410568e-01
2.88610280e-01 -1.09549686e-02 -3.53137821e-01 1.39174119e-01
-7.29049027e-01 -1.02963246e-01 6.31911576e-01 2.68511921e-01
3.15027922e-01 -7.17612922e-01 -1.16956711e-01 7.99290180e-01
3.70101392e-01 6.80010498e-01 7.89065063e-01 -6.99426770e-01
-7.57340372e-01 9.94206190e-01 -1.11278459e-01 -3.19745570e-01
-8.49473834e-01 -4.45588261e-01 -7.85403490e-01 6.69064075e-02
6.44571364e-01 1.55264631e-01 -1.35223258e+00 1.05693674e+00
5.99124312e-01 -2.10009426e-01 -1.75178573e-01 4.84532744e-01
9.00998533e-01 2.12412551e-01 2.10671067e-01 -6.55720532e-02
2.02352405e+00 -7.13046968e-01 -7.47272551e-01 1.22437716e-01
1.31112981e+00 -7.40611136e-01 7.92878747e-01 3.17937165e-01
-6.85662270e-01 1.74426973e-01 -7.59166956e-01 -5.14031351e-01
-6.43174767e-01 6.11580968e-01 6.96820021e-01 7.90581465e-01
-1.33395088e+00 -2.68415779e-01 -1.39394486e+00 -9.27277029e-01
9.29755807e-01 7.51997352e-01 -5.41848063e-01 -2.72945106e-01
-4.05823022e-01 3.54858041e-01 1.33162543e-01 -5.16357459e-02
-5.85061491e-01 -6.96740210e-01 -7.32434928e-01 -2.24793524e-01
-1.21492893e-01 -7.14395046e-01 1.87281895e+00 -3.68934691e-01
-1.54316342e+00 1.17838490e+00 -6.60291612e-01 -1.07754469e-01
3.46341133e-01 1.92022607e-01 6.80125132e-02 4.12718624e-01
4.84486111e-03 7.66678393e-01 9.50566318e-04 -7.61748612e-01
-1.22233617e+00 -5.53024232e-01 -6.33110046e-01 -1.69506431e-01
-2.57463753e-01 -1.47670418e-01 -9.55977857e-01 -1.17219225e-01
-1.23398103e-01 -1.06789041e+00 -4.25663084e-01 5.75344384e-01
-4.68985409e-01 -3.39359105e-01 6.55443370e-01 -7.21795321e-01
8.29595089e-01 -2.07762051e+00 -5.30407369e-01 2.89627910e-01
3.65710467e-01 4.04867232e-01 1.27574485e-02 1.45703852e-01
3.40804607e-01 3.81949395e-01 7.53373653e-02 -5.19470096e-01
-4.24098819e-01 -3.22488934e-01 4.53304172e-01 9.04666305e-01
3.09671134e-01 1.06004250e+00 -9.59848523e-01 -8.45866621e-01
2.28533998e-01 9.78760183e-01 -4.22888100e-01 -1.59249585e-02
-2.33138382e-01 1.79441735e-01 -3.89706612e-01 1.38534760e+00
6.70954406e-01 -8.67450655e-01 3.84747624e-01 -2.03923866e-01
4.20942791e-02 1.32639527e-01 -6.03193343e-01 1.15373254e+00
-2.28782356e-01 6.94176078e-01 4.31349605e-01 -1.12416372e-01
2.63233036e-01 3.69169086e-01 3.50139201e-01 -5.43731451e-01
5.17330468e-01 5.88594258e-01 -1.66553125e-01 -3.29017848e-01
2.09001854e-01 -2.06650630e-01 3.58309299e-01 1.69263631e-01
-1.73555002e-01 -1.10926986e-01 3.79883915e-01 2.34699994e-01
1.34360051e+00 -5.37114084e-01 7.63410151e-01 -2.44480520e-01
5.38314700e-01 4.69382733e-01 4.20392036e-01 3.80318254e-01
-5.54019213e-01 7.04990208e-01 6.72190070e-01 -1.18160278e-01
-5.60577273e-01 -6.34252250e-01 -4.90041673e-01 8.12052429e-01
-1.85370132e-01 -4.73286547e-02 -4.93787259e-01 -7.03658998e-01
2.16244712e-01 3.88409719e-02 -1.16949975e+00 3.48663658e-01
-5.35566926e-01 -1.12789011e+00 6.24887884e-01 3.60957861e-01
7.72789726e-03 -6.79701030e-01 -2.90867865e-01 9.86691192e-02
-8.19089189e-02 -7.25733340e-01 -3.47077519e-01 5.38151026e-01
-7.48963237e-01 -1.53120410e+00 -1.00178218e+00 -1.21343410e+00
1.05611193e+00 3.66765678e-01 6.48216546e-01 7.29453862e-01
-1.16362345e+00 -5.49927242e-02 -5.10362945e-02 -7.99461365e-01
-5.90798438e-01 2.93020397e-01 -5.41701913e-01 -3.65266919e-01
7.62967348e-01 2.19544709e-01 -1.18521738e+00 -2.67536100e-02
-7.99379170e-01 -1.91396788e-01 1.00173473e+00 9.16680932e-01
1.22195411e+00 -2.77015150e-01 3.31713080e-01 -1.05144823e+00
1.06024042e-01 -4.23718959e-01 -5.62322974e-01 1.16830617e-01
-2.41423905e-01 -3.82749081e-01 2.04163700e-01 -1.97254419e-01
-7.41887152e-01 3.44285548e-01 -6.67198718e-01 5.07822812e-01
-3.07659239e-01 3.61633509e-01 2.65222222e-01 -1.65499389e-01
5.58966517e-01 6.12931289e-02 4.90318120e-01 -1.56354100e-01
-3.46557319e-01 9.32968795e-01 2.65023261e-01 2.79767752e-01
9.70112234e-02 1.04309070e+00 3.58090214e-02 -1.05253923e+00
-5.75414419e-01 -1.06645560e+00 -2.68122554e-01 -6.30435795e-02
8.83477092e-01 -1.12403190e+00 -1.13467622e+00 9.14178073e-01
-6.68797910e-01 -7.85905719e-01 1.97643340e-01 4.72911596e-01
6.58026785e-02 5.67740165e-02 -1.43176675e+00 -5.21798015e-01
-7.71591544e-01 -1.11920524e+00 1.62394857e+00 7.52597034e-01
-2.63327837e-01 -1.16553009e+00 4.83786702e-01 2.92828590e-01
4.61217523e-01 2.56003290e-01 5.99278271e-01 -6.43398046e-01
-3.21978062e-01 -7.37607777e-01 -4.20213014e-01 -3.18846196e-01
4.53910768e-01 6.74809635e-01 -9.83799100e-01 -4.99073267e-01
-6.16270959e-01 -4.04810488e-01 1.24368000e+00 8.18286657e-01
9.46396887e-01 7.10188672e-02 -1.25384939e+00 6.50583506e-01
1.73834956e+00 6.46818876e-02 6.20022953e-01 3.97178054e-01
3.17040980e-01 2.95790642e-01 4.62109566e-01 1.59262821e-01
3.57559532e-01 1.16569892e-01 3.14543188e-01 -4.23505694e-01
-4.85925466e-01 1.55599907e-01 -1.57477349e-01 3.79136443e-01
3.45467955e-01 -5.59565723e-01 -1.10103559e+00 9.32709813e-01
-1.21941519e+00 -3.98019850e-01 -4.81614918e-02 1.67712975e+00
9.02946353e-01 -9.64962766e-02 2.09096894e-01 2.00832754e-01
7.26284683e-01 -6.43567324e-01 -7.02978969e-01 1.06830187e-01
-7.03861043e-02 2.14282200e-01 5.46158612e-01 8.16709995e-01
-9.76704121e-01 6.50623024e-01 6.74124765e+00 8.52310359e-01
-1.52476621e+00 -5.66269853e-04 1.31124401e+00 -1.66429892e-01
1.24683544e-01 -6.86703920e-01 -1.27890539e+00 1.59998715e-01
7.49324441e-01 2.83869117e-01 -6.73048854e-01 6.17831826e-01
2.33294759e-02 -4.80157763e-01 -7.74263203e-01 5.22396147e-01
-1.68703109e-01 -1.63516963e+00 1.01419212e-02 7.96868861e-01
4.49695766e-01 2.46586010e-01 -1.24486303e-02 4.22287881e-02
3.54870081e-01 -8.86728287e-01 -1.89708933e-01 2.20736310e-01
1.14243627e+00 -5.07404268e-01 1.26325488e+00 -8.08620900e-02
-1.02860403e+00 6.88781887e-02 -3.54509093e-02 5.01890898e-01
-1.66279972e-01 8.76261294e-01 -1.50758708e+00 -1.34088829e-01
5.35202444e-01 3.44664663e-01 -6.43495202e-01 1.29350078e+00
2.26186067e-01 7.64654994e-01 -5.89985371e-01 -4.54757333e-01
-1.06122553e-01 4.72559214e-01 -7.26864338e-02 1.65603769e+00
4.14440185e-01 2.45309219e-01 -1.44834891e-01 -3.84837799e-02
-5.86430058e-02 4.10499632e-01 1.05349816e-01 -3.94720025e-02
5.21024466e-01 1.61215341e+00 -1.49189138e+00 -1.73982635e-01
-2.51573086e-01 7.30578482e-01 2.84623951e-01 -7.03785045e-04
-2.64532715e-01 -4.25573677e-01 4.76657063e-01 1.77420110e-01
3.65983933e-01 3.49886775e-01 -2.64950573e-01 -6.77760065e-01
-6.62977934e-01 -7.00592697e-01 8.43187869e-01 -2.11040676e-01
-9.33661163e-01 2.50124365e-01 -8.15371633e-01 -7.58348525e-01
3.14232670e-02 -1.14846206e+00 -7.78134942e-01 2.82403648e-01
-1.76307786e+00 -1.37650967e+00 -6.44945443e-01 3.66077721e-01
3.88104826e-01 3.00430477e-01 1.17553329e+00 9.83549729e-02
-9.60119903e-01 9.49015558e-01 2.65884876e-01 2.73245752e-01
7.55462229e-01 -1.36315811e+00 -1.06711768e-01 2.51047432e-01
-7.25208759e-01 5.16418993e-01 3.98096293e-01 -7.14146435e-01
-1.40815318e+00 -1.06488681e+00 8.83466303e-01 -2.19925329e-01
5.19269228e-01 -2.93408841e-01 -6.90619946e-01 4.57800627e-01
1.43185619e-03 1.06617905e-01 1.42142296e+00 -1.23604864e-01
-8.74166265e-02 -1.28166173e-02 -1.33743358e+00 6.60853624e-01
3.03932950e-02 -2.43436366e-01 4.93522465e-01 7.98355162e-01
3.92282903e-01 -5.67947924e-01 -9.56055224e-01 2.25420639e-01
8.36115003e-01 -6.48041129e-01 5.79783380e-01 2.32022151e-01
1.53808504e-01 -2.54105926e-01 8.55057836e-02 -6.12489343e-01
-2.93044388e-01 -2.06259236e-01 2.10529845e-02 7.15427637e-01
6.54771984e-01 -8.43933642e-01 1.39182687e+00 7.12903440e-02
-6.25650138e-02 -1.17868590e+00 -9.58171129e-01 -1.06421113e-01
2.11364776e-01 1.63600028e-01 2.27318943e-01 3.67788136e-01
1.48329079e-01 -9.61284041e-02 6.19656980e-01 1.89004362e-01
5.29128671e-01 -2.76305795e-01 6.07200623e-01 -1.04359591e+00
-5.44633046e-02 -8.45525086e-01 -5.16977310e-01 -4.08689111e-01
-4.17392582e-01 -5.87915361e-01 -7.58420303e-03 -1.78709030e+00
6.46541774e-01 -8.69846120e-02 -3.80185783e-01 7.48580992e-01
-5.41036725e-01 9.93253171e-01 -4.82334167e-01 1.87350795e-01
-5.06122410e-01 -1.54659599e-01 1.20920026e+00 -1.92085817e-01
-2.63298899e-01 -7.69307539e-02 -7.53796160e-01 4.12455887e-01
1.32091963e+00 -3.11930209e-01 3.58705193e-01 7.94456303e-02
2.89066136e-02 -1.97953776e-01 2.54462421e-01 -9.92885232e-01
4.60441798e-01 -1.26757994e-01 6.43154740e-01 -7.69275367e-01
3.23001564e-01 -2.60187119e-01 -1.12610571e-01 1.40438628e+00
-1.72541574e-01 -5.91282368e-01 5.03933668e-01 3.96641254e-01
-1.98335215e-01 1.33013830e-01 1.01245654e+00 -3.81077267e-02
-1.94000334e-01 2.49420598e-01 -8.52922261e-01 -6.72115147e-01
9.08884823e-01 -3.64278585e-01 -9.05775845e-01 -1.05959363e-01
-4.52585995e-01 3.25057298e-01 7.03599095e-01 -4.11262661e-01
3.43223542e-01 -8.40417564e-01 -9.54719603e-01 9.05425400e-02
5.62358379e-01 1.40451342e-01 4.73400623e-01 1.12251818e+00
-1.27414358e+00 6.02615237e-01 2.17201620e-01 -8.27680051e-01
-1.72390795e+00 5.35183847e-02 7.50820160e-01 -6.26116991e-01
-3.24832320e-01 1.26513219e+00 1.84075698e-01 -4.49350059e-01
2.37942368e-01 -5.38262486e-01 -2.78367400e-01 -9.94467586e-02
1.05541313e+00 2.93655992e-01 2.80664504e-01 -2.33378977e-01
-5.07422507e-01 4.59223539e-01 -7.73174047e-01 1.29642263e-01
1.08149910e+00 6.90452605e-02 -3.21145684e-01 1.72042191e-01
1.39990735e+00 1.81209564e-01 -9.03726280e-01 7.77675882e-02
-2.94310987e-01 -4.01194468e-02 2.96225965e-01 -9.11927223e-01
-8.98390770e-01 4.81075704e-01 1.01679921e+00 -1.45531163e-01
1.03395915e+00 2.56995082e-01 8.19501698e-01 4.21230823e-01
-1.12474620e-01 -8.05262744e-01 1.89448725e-02 1.56831935e-01
3.45513850e-01 -1.61493742e+00 1.80807933e-01 -7.81611025e-01
1.39105707e-01 1.09415579e+00 4.83040363e-01 -6.62377477e-02
8.88283432e-01 7.51935422e-01 6.65403068e-01 -4.12041247e-01
-9.21764135e-01 -1.58805266e-01 -1.03366703e-01 5.64277351e-01
9.28730369e-01 2.13646635e-01 -2.96770781e-01 2.16954634e-01
6.06367476e-02 2.37180203e-01 4.96661276e-01 1.22109568e+00
-6.48449242e-01 -1.02841341e+00 -4.25313950e-01 7.57790089e-01
-8.88516247e-01 9.92687270e-02 -4.25032914e-01 9.14173067e-01
-2.50918478e-01 8.25762689e-01 3.11830789e-01 9.05930772e-02
-1.89606830e-01 -2.28220001e-01 1.42150015e-01 -5.59870958e-01
-5.08942902e-01 5.42430401e-01 -1.95972212e-02 -3.67606401e-01
-1.29850388e-01 -7.22637773e-01 -1.54694927e+00 -3.00547123e-01
-5.25273979e-01 -3.04769482e-02 8.68255258e-01 5.65329790e-01
3.35158199e-01 5.34701705e-01 1.96338594e-01 -6.69355094e-01
9.54944566e-02 -1.11988521e+00 -5.78435242e-01 -1.82252571e-01
7.14845061e-01 -8.08701217e-02 -6.03168964e-01 1.17915228e-01]
|
[15.126811027526855, -3.1064934730529785]
|
183d6377-4b0f-444d-8fc3-bddc504c13ff
|
temporally-consistent-horizon-lines
|
1907.10014
| null |
https://arxiv.org/abs/1907.10014v2
|
https://arxiv.org/pdf/1907.10014v2.pdf
|
Temporally Consistent Horizon Lines
|
The horizon line is an important geometric feature for many image processing and scene understanding tasks in computer vision. For instance, in navigation of autonomous vehicles or driver assistance, it can be used to improve 3D reconstruction as well as for semantic interpretation of dynamic environments. While both algorithms and datasets exist for single images, the problem of horizon line estimation from video sequences has not gained attention. In this paper, we show how convolutional neural networks are able to utilise the temporal consistency imposed by video sequences in order to increase the accuracy and reduce the variance of horizon line estimates. A novel CNN architecture with an improved residual convolutional LSTM is presented for temporally consistent horizon line estimation. We propose an adaptive loss function that ensures stable training as well as accurate results. Furthermore, we introduce an extension of the KITTI dataset which contains precise horizon line labels for 43699 images across 72 video sequences. A comprehensive evaluation shows that the proposed approach consistently achieves superior performance compared with existing methods.
|
['Bodo Rosenhahn', 'Florian Kluger', 'Michael Ying Yang', 'Hanno Ackermann']
|
2019-07-23
| null | null | null | null |
['horizon-line-estimation']
|
['computer-vision']
|
[ 2.96790600e-01 -2.75145829e-01 -1.54958859e-01 -6.98047519e-01
-4.49024469e-01 -2.58392125e-01 6.05237126e-01 -1.71509594e-01
-7.43695855e-01 5.19106328e-01 -2.98476726e-01 -3.85865510e-01
-2.78410107e-01 -6.78070009e-01 -1.02200019e+00 -5.33423603e-01
-2.40032062e-01 3.31873633e-02 5.97952127e-01 -1.02220662e-01
3.46771568e-01 7.97486424e-01 -1.71091187e+00 1.87011704e-01
4.07062531e-01 1.51475608e+00 3.48417759e-01 5.47138572e-01
3.38153206e-02 9.39480364e-01 -2.74558991e-01 -1.07114203e-01
5.67001164e-01 -1.01438977e-01 -4.96812612e-01 2.65695840e-01
8.53998065e-01 -6.72106028e-01 -5.62534392e-01 9.21651065e-01
3.25477868e-01 4.61595058e-01 3.59405905e-01 -1.06021321e+00
4.01217304e-03 1.46866158e-01 -5.98316789e-01 2.89210051e-01
-1.34543806e-01 -1.27095670e-01 4.45935130e-01 -6.67315125e-01
7.05362201e-01 9.26678061e-01 8.11611533e-01 1.77078083e-01
-4.89237964e-01 -4.67171341e-01 7.19876513e-02 9.42896664e-01
-1.13762045e+00 -4.52077717e-01 7.92191207e-01 -4.12155747e-01
9.68761027e-01 -2.27271486e-02 6.10709071e-01 6.71097577e-01
5.88072479e-01 8.49454880e-01 7.37793624e-01 -4.07604069e-01
1.28481373e-01 -1.34079278e-01 5.49163632e-02 7.56760597e-01
-7.17066973e-02 3.30962747e-01 -4.20701593e-01 5.50338209e-01
8.20313573e-01 -4.36902605e-02 -2.72495717e-01 -6.31342590e-01
-1.10831368e+00 7.21354067e-01 6.36612654e-01 1.21656783e-01
-3.44145030e-01 4.88338977e-01 4.95762587e-01 1.88277349e-01
3.82907718e-01 1.16503827e-01 -3.64195257e-01 -9.33054537e-02
-8.92156065e-01 3.48127544e-01 3.11705172e-01 9.71297681e-01
7.43021548e-01 4.50187445e-01 1.35440221e-02 8.27921212e-01
7.49536604e-02 5.47373295e-01 3.62570524e-01 -1.37837076e+00
3.37100476e-01 1.75719917e-01 2.12914556e-01 -1.11387432e+00
-7.11170793e-01 -5.09673059e-01 -6.85208201e-01 6.11213684e-01
3.35987359e-01 1.69203296e-01 -1.06141686e+00 1.39991403e+00
2.34717101e-01 4.31289345e-01 -4.66091149e-02 8.62751782e-01
7.91047692e-01 6.03585005e-01 -3.70886773e-01 -1.65731326e-01
1.07073271e+00 -1.22052002e+00 -7.83857048e-01 -3.69773149e-01
6.14223003e-01 -7.49031603e-01 3.50755721e-01 5.40007353e-01
-9.08630848e-01 -6.78591728e-01 -1.23542356e+00 -1.98191434e-01
-3.59106719e-01 1.80693701e-01 4.08407509e-01 3.33323002e-01
-1.08321607e+00 5.54426193e-01 -6.51511252e-01 -1.92874283e-01
2.61559963e-01 1.80623963e-01 -4.32184041e-01 -1.61319003e-01
-1.18006885e+00 1.18290210e+00 6.49869382e-01 5.74024856e-01
-5.58248878e-01 -4.61991012e-01 -1.25583756e+00 -1.79114178e-01
4.63346332e-01 -5.97595751e-01 1.40968311e+00 -9.39441144e-01
-1.48090446e+00 6.28297567e-01 -1.61644250e-01 -1.31449068e+00
8.48252058e-01 -4.64379430e-01 -2.50965774e-01 3.13886434e-01
-1.40013844e-01 9.67214286e-01 1.07651353e+00 -1.05731666e+00
-9.86807585e-01 -1.62428811e-01 1.32297069e-01 2.54812717e-01
1.82426319e-01 -2.44946063e-01 -7.11328745e-01 -3.97921264e-01
1.35084704e-01 -9.32786763e-01 -3.39888781e-01 1.84975967e-01
1.13546883e-03 4.64282930e-02 1.18683517e+00 -8.16495895e-01
7.53509820e-01 -1.89943385e+00 -2.65610874e-01 1.68113932e-01
4.81323600e-02 3.42804074e-01 5.76363094e-02 -3.20996419e-02
1.51397631e-01 -3.86296779e-01 -3.46721053e-01 -4.50106442e-01
-3.45574409e-01 3.56023997e-01 -2.66961634e-01 6.70047879e-01
1.27472296e-01 8.77831578e-01 -6.63915277e-01 -2.90064901e-01
1.03568661e+00 5.44092953e-01 -2.14190140e-01 -9.99783352e-02
-1.73568726e-01 3.69050682e-01 -9.98854190e-02 2.73420841e-01
7.35739112e-01 8.61861035e-02 -3.06237459e-01 -2.90246695e-01
-4.45626557e-01 -1.44850135e-01 -8.67030501e-01 1.69302356e+00
-5.83370626e-01 1.38319635e+00 -2.44115144e-01 -1.00724590e+00
1.01203406e+00 2.50136387e-02 5.67696095e-01 -1.30063546e+00
2.99963653e-01 3.41000438e-01 -1.06978908e-01 -4.70218033e-01
9.44332302e-01 2.24401355e-01 2.92311043e-01 -2.29859322e-01
-3.32539678e-01 -2.96958178e-01 2.42522836e-01 -3.11894864e-01
6.28120959e-01 2.91179478e-01 2.44302094e-01 -6.45323321e-02
8.07134867e-01 4.31169420e-02 6.47064328e-01 6.32832825e-01
-3.63521129e-01 7.53017604e-01 5.07020450e-04 -8.70822251e-01
-1.36334038e+00 -6.36930346e-01 -2.73073286e-01 2.54172921e-01
5.69253743e-01 1.70563594e-01 -5.01782894e-01 -3.14040422e-01
-1.68437451e-01 7.28631854e-01 -6.59107745e-01 -5.21689728e-02
-8.25458467e-01 -1.21837825e-01 3.09863508e-01 7.98079789e-01
1.10637045e+00 -8.69479537e-01 -1.11847138e+00 2.99223900e-01
-1.67797133e-01 -1.65910923e+00 -2.29069352e-01 -5.11469021e-02
-6.90642118e-01 -1.03098083e+00 -6.93694890e-01 -6.43877327e-01
2.09577248e-01 5.98640740e-01 8.42644155e-01 -1.33406952e-01
-1.81612149e-01 2.47821987e-01 -3.46046597e-01 -4.00075078e-01
-1.32897392e-01 -2.03549057e-01 -8.27279836e-02 3.49500850e-02
5.71710728e-02 -1.33854657e-01 -6.45927429e-01 3.73697340e-01
-7.66189039e-01 3.31533700e-01 3.05388123e-01 8.12764585e-01
5.18284440e-01 9.80874002e-02 3.39606851e-01 -5.46109080e-01
-3.94658942e-04 -1.60311937e-01 -1.06475961e+00 -6.93397895e-02
-6.19729996e-01 -1.38279840e-01 4.74240333e-01 8.97668675e-02
-1.10843539e+00 1.09676495e-01 -4.39432204e-01 -8.17780197e-01
-8.27829018e-02 3.69450122e-01 2.86164463e-01 -4.22894120e-01
3.98208112e-01 1.50029033e-01 5.45209125e-02 -1.45209223e-01
3.54494005e-01 2.07139477e-01 8.69722486e-01 4.42273133e-02
4.95229781e-01 8.14418733e-01 3.02782029e-01 -1.24184573e+00
-7.90093660e-01 -7.59837985e-01 -7.21517682e-01 -6.15517557e-01
9.40483809e-01 -8.58732581e-01 -5.47130525e-01 7.38774240e-01
-1.31226432e+00 -5.08579075e-01 -8.03456157e-02 6.24120951e-01
-9.89278734e-01 5.19010425e-01 -4.18946803e-01 -7.81642973e-01
-1.90137684e-01 -1.22092545e+00 1.04198265e+00 2.40132913e-01
2.27427781e-01 -1.12843037e+00 -2.29546741e-01 3.00987870e-01
3.11984867e-01 4.72365379e-01 4.45148975e-01 -1.00920677e-01
-8.12248409e-01 -2.69878715e-01 -3.75380576e-01 4.82407600e-01
-1.51304498e-01 -7.04235509e-02 -9.59155500e-01 -1.44021273e-01
-5.67480661e-02 -1.16618732e-02 1.18587387e+00 8.44885767e-01
1.13893604e+00 -3.49158496e-02 -8.69874954e-02 8.91766787e-01
1.36005211e+00 6.72781229e-01 8.68192852e-01 1.05372310e+00
7.55356610e-01 7.44447291e-01 8.87116194e-01 3.22493970e-01
2.04329938e-01 8.59755933e-01 6.54509008e-01 -8.72881785e-02
-1.32470757e-01 2.19164819e-01 1.42354950e-01 3.15446556e-01
-4.90782969e-02 -3.01385492e-01 -8.01630855e-01 5.54811537e-01
-2.08190536e+00 -1.07129288e+00 -3.16019088e-01 2.15061116e+00
5.59876265e-04 2.60145366e-01 -2.65348852e-01 3.24311823e-01
4.94229555e-01 3.62958103e-01 -5.18626809e-01 -5.13429642e-01
-2.36047521e-01 -2.56385833e-01 1.08746183e+00 6.41184688e-01
-1.22367942e+00 9.11525905e-01 6.17579508e+00 7.86186516e-01
-1.38768578e+00 -3.31941210e-02 5.36002874e-01 2.09236011e-01
5.84813282e-02 -2.63252497e-01 -7.05744803e-01 8.04248676e-02
7.59409666e-01 -7.06693083e-02 2.19930373e-02 8.37181807e-01
5.00089943e-01 -2.79991359e-01 -8.04764330e-01 1.30542147e+00
2.46134341e-01 -1.59436607e+00 -2.12930232e-01 -1.02334015e-01
9.24389720e-01 2.60402948e-01 1.79920197e-01 -1.66188538e-01
-2.72966951e-01 -8.97382438e-01 1.10208929e+00 6.46714330e-01
4.62170035e-01 -9.63572621e-01 1.07527268e+00 3.10305059e-01
-1.24739122e+00 -5.99316768e-02 -2.41154030e-01 -8.75300169e-02
5.08012354e-01 4.33124006e-01 -1.10186481e+00 7.89416134e-01
8.01579714e-01 9.87631679e-01 -5.29909968e-01 1.46286106e+00
-7.99096972e-02 1.03542998e-01 -3.49725544e-01 3.17498475e-01
7.68077016e-01 -2.85844982e-01 5.03388524e-01 9.86967564e-01
4.30507690e-01 -2.85722256e-01 7.18418974e-03 3.87173533e-01
3.08901638e-01 4.30561602e-03 -8.99145246e-01 2.82551557e-01
7.82328472e-02 9.42269325e-01 -8.74718785e-01 -2.97966599e-01
-4.56263244e-01 7.96987772e-01 1.00267142e-01 4.76372212e-01
-8.34002376e-01 -1.76728547e-01 5.12150466e-01 9.36244875e-02
4.37524974e-01 -8.08983088e-01 -2.60115415e-01 -7.51550913e-01
5.78693524e-02 -3.41012925e-01 1.08473293e-01 -9.05525148e-01
-5.19782364e-01 7.67905235e-01 1.23781398e-01 -1.36036479e+00
-7.83873975e-01 -8.04137111e-01 -3.83090645e-01 4.64146882e-01
-1.93204772e+00 -1.07892716e+00 -6.10850871e-01 4.39917177e-01
1.02863884e+00 -9.85424221e-02 1.28328875e-01 3.75795275e-01
-2.86363453e-01 1.60200596e-01 4.74316746e-01 -6.55156150e-02
4.83867377e-01 -9.49978411e-01 5.49573302e-01 1.21146178e+00
-2.87343208e-02 8.25212821e-02 1.00082576e+00 -3.84201139e-01
-9.33290780e-01 -1.33317411e+00 5.74896455e-01 -1.27795368e-01
3.42451751e-01 1.25883147e-01 -8.58314574e-01 6.92917883e-01
1.36573374e-01 4.05197777e-03 -1.03238441e-01 -5.96323252e-01
-8.84901434e-02 -3.35049331e-01 -8.33071768e-01 6.01188123e-01
8.96718740e-01 -2.63423026e-01 -2.15013176e-01 9.95169505e-02
7.08492935e-01 -8.01164985e-01 -4.96627718e-01 8.56339753e-01
6.44376814e-01 -1.37472224e+00 1.04523921e+00 1.34689072e-02
3.03862065e-01 -4.85094041e-01 -9.71399322e-02 -9.09814835e-01
9.06671360e-02 -1.87254086e-01 -1.54779360e-01 5.61784625e-01
1.16809666e-01 -6.37751222e-01 8.87344658e-01 3.63012463e-01
-6.30625248e-01 -7.78179526e-01 -1.02828896e+00 -8.94761562e-01
-2.47904360e-01 -1.00925386e+00 2.50643253e-01 5.03164649e-01
-8.18052173e-01 -1.64929837e-01 -7.44026482e-01 1.67500302e-01
8.43171895e-01 3.62738618e-03 8.14779401e-01 -7.47021317e-01
9.30335820e-02 -5.54177523e-01 -8.42256606e-01 -1.28895020e+00
2.94191450e-01 -3.88113797e-01 3.90900671e-01 -1.72416151e+00
-3.92777920e-01 -4.07969832e-01 -1.35143235e-01 5.53736947e-02
3.26048136e-01 4.68571097e-01 9.36221108e-02 9.17901099e-02
-4.85652655e-01 6.51454687e-01 1.05302024e+00 -2.14643523e-01
-8.72899294e-02 8.30142125e-02 3.38853121e-01 9.47527885e-01
8.47462237e-01 6.45187572e-02 -6.39522374e-01 -6.09188616e-01
5.39518520e-02 1.41285241e-01 6.06673896e-01 -1.30236709e+00
4.10681844e-01 3.58915329e-02 4.79701370e-01 -1.13711047e+00
5.96017838e-01 -9.10180330e-01 1.39650539e-01 4.47432995e-01
-1.87222004e-01 1.57163724e-01 3.93085837e-01 7.30762780e-01
-5.26841044e-01 -4.20435727e-01 8.84493589e-01 5.76349124e-02
-1.66179466e+00 5.38012087e-01 -3.68977100e-01 -4.37257797e-01
1.20684588e+00 -6.42437577e-01 -2.16759652e-01 -6.25077248e-01
-2.46093065e-01 4.28031594e-01 4.67468351e-01 8.18558812e-01
1.06837618e+00 -1.23381913e+00 -4.05837059e-01 2.57596254e-01
1.73287496e-01 1.69755608e-01 5.49285114e-01 8.50778341e-01
-1.10656953e+00 7.72281289e-01 -3.79107177e-01 -9.54608083e-01
-1.22305310e+00 4.28784937e-01 6.38228118e-01 9.64945927e-02
-9.92962062e-01 6.10365748e-01 2.93150604e-01 2.20502578e-02
2.82873124e-01 -5.08679509e-01 -4.37717587e-01 -2.23339573e-01
6.43820524e-01 3.11171830e-01 3.66787821e-01 -8.58412027e-01
-2.05326706e-01 8.24103415e-01 -1.21702060e-01 -1.55707538e-01
1.04251945e+00 -5.17142057e-01 2.62598634e-01 5.78403831e-01
1.36655748e+00 -4.98402476e-01 -1.57706630e+00 -2.44811848e-01
2.17556998e-01 -6.53928399e-01 2.46702075e-01 -2.13736460e-01
-1.05475557e+00 9.45204020e-01 7.88567305e-01 -9.79053155e-02
9.99105096e-01 -5.08051336e-01 1.02237356e+00 4.84197289e-01
4.87444133e-01 -1.07218254e+00 -1.01334192e-01 8.66972148e-01
7.55893528e-01 -1.52613914e+00 -1.90513909e-01 -3.33587170e-01
-4.93109405e-01 1.44918883e+00 4.43571031e-01 -3.09891719e-02
4.04798210e-01 4.45038751e-02 2.56815434e-01 2.97642909e-02
-5.90648770e-01 -3.14943105e-01 4.83273298e-01 5.76152980e-01
8.07550699e-02 -2.92887270e-01 -9.88242775e-02 -2.67499298e-01
-1.93912983e-01 1.06359189e-02 5.98257303e-01 8.93328786e-01
-6.17889345e-01 -5.99713027e-01 -2.34867543e-01 2.99826384e-01
-1.64006814e-01 1.24426812e-01 2.50207126e-01 9.26850080e-01
1.18837163e-01 7.46061981e-01 3.30315977e-01 -3.36234510e-01
1.40561312e-01 -2.39335075e-01 5.09941339e-01 -2.43436918e-02
1.34356543e-01 -1.52637109e-01 3.64319295e-01 -8.29106152e-01
-7.18688309e-01 -4.46602494e-01 -1.25579166e+00 -2.21281067e-01
-2.09355876e-01 -1.38598204e-01 9.96951461e-01 1.23482692e+00
-1.24815293e-02 6.82914972e-01 5.73758245e-01 -1.01115525e+00
-6.73596635e-02 -5.48570633e-01 -5.74943721e-01 9.51377079e-02
6.30546510e-01 -8.93702030e-01 -2.03747034e-01 2.21480772e-01]
|
[8.357603073120117, -1.7098510265350342]
|
692bd1ab-ebce-4118-b10e-210071c6cdfd
|
depth-from-semi-calibrated-stereo-and-defocus
| null | null |
http://openaccess.thecvf.com/content_cvpr_2016/html/Wang_Depth_From_Semi-Calibrated_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Wang_Depth_From_Semi-Calibrated_CVPR_2016_paper.pdf
|
Depth From Semi-Calibrated Stereo and Defocus
|
In this work, we propose a multi-camera system where we combine a main high-quality camera with two low-res auxiliary cameras. The auxiliary cameras are well calibrated and act as a passive depth sensor by generating disparity maps. The main camera has an interchangeable lens and can produce good quality images at high resolution. Our goal is, given the low-res depth map from the auxiliary cameras, generate a depth map from the viewpoint of the main camera. The advantage of our system, compared to other systems such as light-field cameras or RGBD sensors, is the ability to generate a high-resolution color image with a complete depth map, without sacrificing resolution and with minimal auxiliary hardware. Since the main camera has an interchangeable lens, it cannot be calibrated beforehand, and directly applying stereo matching on it and either of the auxiliary cameras often leads to unsatisfactory results. Utilizing both the calibrated cameras at once, we propose a novel approach to better estimate the disparity map of the main camera. Then by combining the defocus cue of the main camera, the disparity map can be further improved. We demonstrate the performance of our algorithm on various scenes.
|
['Ting-Chun Wang', 'Ravi Ramamoorthi', 'Manohar Srikanth']
|
2016-06-01
| null | null | null |
cvpr-2016-6
|
['stereo-matching']
|
['computer-vision']
|
[ 4.44943756e-01 -2.29102075e-01 2.80494690e-01 -2.48801317e-02
-7.10553348e-01 -5.98885417e-01 1.70277163e-01 -4.15715665e-01
-5.49076557e-01 6.21030927e-01 -1.57509476e-01 -6.74157962e-02
4.33629423e-01 -9.64793682e-01 -6.53259397e-01 -8.39423835e-01
9.82080758e-01 1.97450787e-01 7.46287167e-01 7.81641155e-03
4.33079571e-01 4.09863710e-01 -1.72148800e+00 -2.95573566e-02
7.02085912e-01 9.69168961e-01 8.06635618e-01 8.15629303e-01
1.90222546e-01 5.98441482e-01 -3.60361218e-01 -3.26776236e-01
6.30946517e-01 -4.47759122e-01 -3.03488642e-01 3.93856287e-01
6.78285360e-01 -9.77972448e-01 -3.88723522e-01 1.21544397e+00
3.49410772e-01 -2.58504927e-01 4.45003528e-03 -1.00201678e+00
-1.14915539e-02 -1.74331725e-01 -9.58959103e-01 -1.63473681e-01
8.42369914e-01 6.08465485e-02 6.95180237e-01 -6.02538407e-01
6.80293441e-01 1.03752244e+00 2.45541692e-01 4.55336332e-01
-1.11740410e+00 -6.50429606e-01 -7.88038131e-03 2.66380142e-02
-1.34798229e+00 -5.73336065e-01 9.01458859e-01 -2.41426647e-01
2.59665549e-01 2.08076630e-02 7.86149740e-01 7.01623499e-01
2.53019631e-01 1.71735808e-01 1.16589534e+00 -4.96341407e-01
1.87488735e-01 2.73843437e-01 -2.00214729e-01 5.89344800e-01
4.48908567e-01 2.73851901e-01 -5.35798490e-01 1.16376892e-01
1.32396066e+00 4.52554315e-01 -8.41725647e-01 -4.89205092e-01
-1.20833755e+00 3.57018888e-01 3.10793728e-01 2.80677199e-01
-3.44758511e-01 2.80856304e-02 -4.02762115e-01 -5.31463549e-02
4.17130291e-02 9.41868946e-02 -2.74004713e-02 -1.97445601e-02
-7.74399459e-01 -2.24872887e-01 4.33146149e-01 1.15380073e+00
1.11621928e+00 -4.32294726e-01 5.17498076e-01 4.53873128e-01
3.40882421e-01 8.33773255e-01 2.38380328e-01 -1.26517594e+00
6.62110865e-01 6.37771726e-01 2.89904058e-01 -9.16316390e-01
-1.48571096e-02 1.21050112e-01 -7.48743176e-01 8.77081573e-01
5.36266267e-01 -6.66989982e-02 -6.63343549e-01 1.30029666e+00
3.50679129e-01 2.19951898e-01 2.00388446e-01 1.01912701e+00
4.30438966e-01 5.51847458e-01 -8.39390934e-01 -3.13532323e-01
1.22106421e+00 -7.64504731e-01 -6.72306955e-01 -5.32947719e-01
9.92056448e-03 -1.04105771e+00 8.59312892e-01 6.98966026e-01
-1.27496493e+00 -5.19234478e-01 -1.43093026e+00 -2.69795597e-01
1.70624852e-01 1.42626226e-01 2.62098640e-01 3.76868904e-01
-1.13319230e+00 1.41524449e-01 -9.09286082e-01 -8.70680511e-02
-2.01542467e-01 2.48703808e-01 -5.47271073e-01 -6.26932263e-01
-7.37806737e-01 8.93083870e-01 3.64733905e-01 -5.41248880e-02
-5.48767328e-01 -3.13973188e-01 -8.17749441e-01 -1.72900230e-01
3.85228246e-01 -8.72987986e-01 1.02125978e+00 -8.68208408e-01
-1.80608141e+00 8.40880454e-01 -4.08775538e-01 2.32962713e-01
6.04132891e-01 -1.26817673e-01 -6.32021204e-02 6.24273598e-01
1.18511550e-01 4.34489429e-01 7.04567850e-01 -1.40362692e+00
-1.12104428e+00 -4.42987502e-01 3.32835525e-01 4.26250070e-01
-1.78364441e-02 -3.62080634e-01 -1.01437879e+00 1.47849545e-01
6.59314990e-01 -8.30638349e-01 -1.85792387e-01 3.19465399e-01
-1.85140789e-01 7.08083510e-01 9.06745434e-01 -2.30958596e-01
8.80504489e-01 -2.27176094e+00 2.20438331e-01 -1.02298707e-01
3.55659068e-01 1.04817197e-01 7.65519738e-02 1.10660583e-01
1.88823178e-01 -5.49236357e-01 -7.06035569e-02 -4.49902564e-01
-8.43945801e-01 1.79623559e-01 -2.37131249e-02 5.20906866e-01
-3.52734417e-01 3.87202084e-01 -9.53560472e-01 -4.88906324e-01
7.11460352e-01 6.87946618e-01 -4.74210411e-01 4.34527934e-01
1.79194853e-01 6.52525485e-01 -7.11827502e-02 5.58486938e-01
1.05379617e+00 -2.10046992e-01 1.34176493e-01 -3.52458864e-01
-4.54293549e-01 -4.02072817e-02 -1.62717116e+00 1.70095730e+00
-5.62516570e-01 6.87659144e-01 3.89825314e-01 -3.02178949e-01
9.86910343e-01 3.04725915e-01 2.72434175e-01 -8.22003365e-01
-2.55884193e-02 4.22568649e-01 -3.90582472e-01 -1.97348267e-01
4.65040058e-01 -2.95567870e-01 2.58216619e-01 4.20365036e-01
-2.58297682e-01 -5.40555120e-01 1.26597866e-01 5.74829755e-03
9.12463307e-01 -5.73907164e-04 2.97073036e-01 2.44768068e-01
7.86242366e-01 -3.14146191e-01 5.61397791e-01 2.23514721e-01
1.39364600e-01 1.08075905e+00 1.91404462e-01 -2.82656729e-01
-1.16645312e+00 -1.14710677e+00 3.25455368e-02 3.10177207e-01
8.38317811e-01 -2.10602343e-01 -4.62381512e-01 -2.42765453e-02
-1.81833684e-01 8.58659372e-02 -2.21134603e-01 1.72102705e-01
-6.23188317e-01 -2.08893806e-01 -5.84524758e-02 4.04791802e-01
9.56512332e-01 -1.20115213e-01 -9.61654663e-01 6.23850226e-02
-3.88740212e-01 -1.34331632e+00 -5.82025051e-01 -4.15845327e-02
-1.05032718e+00 -1.44629192e+00 -6.83158398e-01 -7.06079543e-01
9.58884656e-01 1.05744708e+00 8.78136218e-01 1.84875056e-02
1.31635159e-01 2.33274147e-01 -6.85734525e-02 8.22344050e-02
-2.39476234e-01 -6.10079646e-01 -8.18534717e-02 1.98959574e-01
4.39660661e-02 -6.53909445e-01 -8.53059173e-01 6.61050379e-01
-9.95431781e-01 5.73921800e-01 5.58936238e-01 5.41035235e-01
7.30178952e-01 3.12922150e-02 -3.30431581e-01 -5.93587518e-01
-1.22710891e-01 1.98103175e-01 -1.33273280e+00 -7.34868795e-02
-4.33678776e-01 -9.91030633e-02 6.56545162e-01 -1.34028867e-01
-1.12584567e+00 3.89660031e-01 1.87063683e-02 -6.17378831e-01
-1.29859326e-02 -2.42162034e-01 -4.79333907e-01 -2.34558657e-01
3.64785880e-01 1.81566119e-01 1.12440340e-01 -4.62584913e-01
1.43615857e-01 7.74519682e-01 8.94211769e-01 -1.11342296e-01
8.56275797e-01 9.54885423e-01 2.10504815e-01 -8.16741526e-01
-5.83818495e-01 -6.41222715e-01 -9.14672434e-01 -2.86945254e-01
8.15750837e-01 -1.12818682e+00 -8.24933648e-01 7.93698013e-01
-1.34217429e+00 1.02078766e-01 9.08715501e-02 7.38178074e-01
-4.63209361e-01 4.43690777e-01 -4.78670299e-01 -5.32671571e-01
2.29322072e-02 -1.43727577e+00 1.37103057e+00 4.70454246e-01
3.29273254e-01 -9.70984578e-01 6.49475530e-02 4.50772017e-01
1.85560957e-01 1.31058723e-01 4.32982087e-01 4.59052563e-01
-1.30997121e+00 -1.33751765e-01 -2.61190593e-01 3.40936840e-01
3.82172942e-01 4.91341986e-02 -1.12767541e+00 -2.84261078e-01
4.58429188e-01 6.02124780e-02 5.81986904e-01 3.77258867e-01
5.98793805e-01 1.79354995e-01 -3.63157243e-01 1.11593509e+00
2.05518746e+00 5.48075199e-01 8.37035894e-01 4.87041533e-01
7.74461448e-01 3.94064277e-01 6.36456788e-01 2.81186163e-01
5.40788352e-01 8.23627710e-01 5.49071491e-01 -3.38704526e-01
-1.20368958e-01 -1.55543298e-01 3.79109889e-01 6.79447591e-01
-1.97690085e-01 -1.90872163e-01 -7.26430953e-01 1.11191496e-01
-1.48139596e+00 -7.99483538e-01 -2.75537521e-01 2.56374216e+00
6.49813116e-01 -3.48374359e-02 -2.13064864e-01 2.38349035e-01
9.70601678e-01 6.41796440e-02 -5.30491889e-01 1.11040033e-01
-1.29718661e-01 -1.53296381e-01 5.82014799e-01 9.51918244e-01
-5.96921742e-01 4.64571416e-01 6.20650625e+00 2.05153033e-01
-1.28999197e+00 -1.31672263e-01 1.84921652e-01 -3.67090762e-01
-4.33822572e-01 1.36621639e-01 -1.04032028e+00 6.75074756e-01
3.40746880e-01 -1.40000373e-01 3.90255988e-01 6.24922335e-01
8.73574615e-02 -8.52666378e-01 -1.45088577e+00 1.55476284e+00
2.88669109e-01 -9.65892553e-01 -1.36464328e-01 3.70982647e-01
6.53394282e-01 -2.92234749e-01 -5.98963946e-02 -5.60742259e-01
-6.28589913e-02 -4.06220913e-01 4.55670357e-01 4.15935546e-01
1.02626300e+00 -6.71047509e-01 7.12104857e-01 4.64807093e-01
-1.17282510e+00 1.39530823e-01 -5.23229301e-01 -1.88117817e-01
3.88638616e-01 7.47180283e-01 -3.74354213e-01 5.02793252e-01
5.26812971e-01 6.30437970e-01 -2.92101055e-01 9.08466816e-01
-4.67034429e-01 -3.70269030e-01 -3.77387375e-01 5.33116698e-01
-1.00912079e-01 -7.58413970e-01 3.68829250e-01 5.40752590e-01
4.91599917e-01 4.90602314e-01 4.31989618e-02 7.48642862e-01
7.87650868e-02 -3.65773320e-01 -7.04894602e-01 5.87727487e-01
4.52846467e-01 1.21853387e+00 -4.85696316e-01 -3.69173437e-01
-8.65661860e-01 1.39328158e+00 -4.16356400e-02 2.78258294e-01
-6.10649049e-01 -5.07309139e-01 5.90864599e-01 1.90979794e-01
1.48475483e-01 -2.58870006e-01 -1.25333220e-01 -1.52592456e+00
2.99173236e-01 -5.82319915e-01 3.15506719e-02 -1.22851062e+00
-6.10607505e-01 6.39780700e-01 -2.21747249e-01 -1.65961576e+00
-3.31658900e-01 -6.42410398e-01 -3.68715674e-01 1.09845281e+00
-1.88234818e+00 -9.39142466e-01 -8.72258902e-01 1.02469718e+00
2.78929681e-01 2.20616788e-01 5.92106462e-01 4.07037228e-01
-3.14911991e-01 8.63504112e-02 2.88931519e-01 4.10184897e-02
7.92080104e-01 -1.12696815e+00 -1.32177979e-01 1.17027879e+00
-2.19327912e-01 3.62637192e-01 3.34274143e-01 -3.96760851e-01
-1.49922287e+00 -5.65413117e-01 5.89042962e-01 -3.34604681e-01
1.43483177e-01 -3.68227959e-01 -6.54660046e-01 6.31751895e-01
1.58016726e-01 -5.59905060e-02 3.73760849e-01 -4.29499507e-01
-6.21780604e-02 -5.92495620e-01 -9.97570157e-01 3.37059110e-01
6.92534089e-01 -7.58237123e-01 -4.90717709e-01 -5.77281751e-02
4.86102730e-01 -7.88395047e-01 -5.06981790e-01 -2.39212234e-02
7.20024824e-01 -1.59664524e+00 9.48465049e-01 5.94798744e-01
6.75278604e-01 -6.12673640e-01 -2.49909043e-01 -1.24679625e+00
-5.42756915e-02 -3.58497083e-01 2.93133706e-01 1.04192781e+00
4.76205945e-02 -9.60048556e-01 8.55322659e-01 5.99153757e-01
5.64458966e-02 -2.87205815e-01 -8.42131317e-01 -5.26615322e-01
-5.16408801e-01 -1.06912494e-01 4.68879253e-01 5.18257916e-01
-6.00973777e-02 4.05371904e-01 -4.44361448e-01 5.31855583e-01
8.88467312e-01 5.01054227e-01 9.73141909e-01 -1.08920312e+00
-4.60891604e-01 2.95742769e-02 -4.66308862e-01 -1.60973680e+00
-4.02871549e-01 -2.18619242e-01 3.92982550e-02 -1.43947732e+00
2.99036473e-01 -2.78887033e-01 2.34467417e-01 -6.74050450e-02
-2.13565126e-01 2.81099439e-01 2.21341923e-01 3.50378573e-01
-2.95821488e-01 7.06876740e-02 1.51167727e+00 2.49744475e-01
-2.97256202e-01 2.77316589e-02 -5.71731806e-01 9.81873035e-01
3.31120878e-01 -1.78952828e-01 -4.48564649e-01 -8.83275628e-01
2.02247575e-01 6.94286644e-01 1.75482392e-01 -1.13652456e+00
6.53236330e-01 -1.09751023e-01 5.75427949e-01 -6.16638601e-01
7.02155530e-01 -1.20305622e+00 6.18231773e-01 3.89237553e-01
3.03758353e-01 1.02422163e-01 -1.01381965e-01 3.54862690e-01
-4.30304646e-01 -2.45108768e-01 1.19982123e+00 -3.31015766e-01
-6.11908555e-01 2.19436795e-01 -1.30041778e-01 -3.63323838e-01
1.19996548e+00 -6.67488754e-01 -4.82364774e-01 -4.89295155e-01
-2.34132722e-01 2.66163722e-02 1.32181287e+00 6.35141656e-02
1.05002224e+00 -1.28430891e+00 -2.79753178e-01 8.14544916e-01
4.81260084e-02 3.99556935e-01 1.39176264e-01 5.92823207e-01
-1.03749406e+00 3.89023066e-01 -3.42179000e-01 -9.47477698e-01
-1.45395124e+00 5.96566737e-01 5.59393048e-01 5.32166176e-02
-6.44086242e-01 5.02286792e-01 4.69528675e-01 1.27761528e-01
5.78487180e-02 -3.03791136e-01 8.23735148e-02 -1.54499412e-01
9.34084356e-01 4.98786479e-01 1.40248388e-02 -5.63933253e-01
-2.26106495e-01 1.31604338e+00 1.28734663e-01 -4.98649329e-01
1.18720293e+00 -6.40848517e-01 -1.17001541e-01 3.76147538e-01
1.27946246e+00 3.74509037e-01 -1.36044490e+00 -1.44435957e-01
-8.21917951e-01 -1.37236142e+00 2.88848966e-01 -2.43151605e-01
-1.31392908e+00 1.11058438e+00 5.13322592e-01 -1.50990203e-01
1.61408722e+00 -2.49061853e-01 7.57885516e-01 1.51970044e-01
9.30133998e-01 -8.97458732e-01 1.41894832e-01 2.21159324e-01
3.75149727e-01 -1.35193992e+00 3.18843164e-02 -4.48725462e-01
-4.32302296e-01 1.40713561e+00 7.85539687e-01 1.35943532e-01
1.96573704e-01 4.74772602e-01 3.15291226e-01 -5.01805246e-02
-5.76833725e-01 -2.24794954e-01 -1.41792476e-01 5.26019573e-01
1.01970702e-01 -3.18631828e-01 1.72400326e-01 -3.47975343e-01
9.39021334e-02 2.35299859e-02 1.11645508e+00 7.95846403e-01
-6.77687705e-01 -1.24575269e+00 -7.67406642e-01 -4.66335304e-02
-1.87595993e-01 4.13535908e-02 -5.89925162e-02 5.49579442e-01
9.60009843e-02 1.08531570e+00 2.79137284e-01 -2.67739117e-01
5.59912801e-01 -4.30456907e-01 6.10018075e-01 -5.39029002e-01
1.53233171e-01 3.22212368e-01 -3.95768493e-01 -9.46480572e-01
-5.82101524e-01 -4.59105521e-01 -8.85280132e-01 -5.31870842e-01
-4.22239840e-01 -5.31885289e-02 6.19936705e-01 4.80348378e-01
1.05788916e-01 2.07663719e-02 1.02375579e+00 -1.01305485e+00
8.64045396e-02 -4.99350160e-01 -8.70163858e-01 2.70641536e-01
7.76942015e-01 -3.78518492e-01 -7.08092809e-01 1.09999292e-01]
|
[9.2568359375, -2.5644049644470215]
|
fce42baa-6ac1-466e-81a8-65c4c5390a05
|
centroid-based-text-summarization-through
| null | null |
https://aclanthology.org/W17-1003
|
https://aclanthology.org/W17-1003.pdf
|
Centroid-based Text Summarization through Compositionality of Word Embeddings
|
The textual similarity is a crucial aspect for many extractive text summarization methods. A bag-of-words representation does not allow to grasp the semantic relationships between concepts when comparing strongly related sentences with no words in common. To overcome this issue, in this paper we propose a centroid-based method for text summarization that exploits the compositional capabilities of word embeddings. The evaluations on multi-document and multilingual datasets prove the effectiveness of the continuous vector representation of words compared to the bag-of-words model. Despite its simplicity, our method achieves good performance even in comparison to more complex deep learning models. Our method is unsupervised and it can be adopted in other summarization tasks.
|
['Pierpaolo Basile', 'Giovanni Semeraro', 'Gaetano Rossiello']
|
2017-04-01
| null | null | null |
ws-2017-4
|
['extractive-document-summarization']
|
['natural-language-processing']
|
[ 9.17896181e-02 1.90180540e-01 -2.02921540e-01 -2.88046032e-01
-5.88054180e-01 -2.44127855e-01 1.02603590e+00 1.00589037e+00
-7.78172016e-01 7.60524511e-01 1.03382325e+00 -6.47311807e-02
-3.08851331e-01 -6.99419975e-01 -2.77248204e-01 -6.45498812e-01
1.58897445e-01 4.48706746e-01 2.59394765e-01 -5.60682237e-01
6.47349358e-01 3.61484885e-01 -1.51266241e+00 1.51588261e-01
1.09562147e+00 3.49679410e-01 2.87229508e-01 4.83892977e-01
-6.60919964e-01 6.39764786e-01 -9.23734546e-01 -2.96802104e-01
-2.02736810e-01 -4.85626668e-01 -8.30753326e-01 -2.06310585e-01
7.62213469e-01 -1.74793959e-01 -2.57734001e-01 1.02973986e+00
6.61911726e-01 4.58810627e-01 1.04590952e+00 -7.42922664e-01
-7.50185430e-01 8.38350296e-01 -5.30868471e-01 4.14836556e-01
4.30412799e-01 -5.58479965e-01 1.47729897e+00 -7.06229746e-01
6.17126048e-01 1.15057600e+00 6.86622739e-01 3.35806578e-01
-9.71351504e-01 -1.24955729e-01 2.68136829e-01 2.39610583e-01
-9.77681577e-01 -3.45740885e-01 8.74751210e-01 -2.01540738e-01
1.26970792e+00 2.91697502e-01 6.59973621e-01 9.60319281e-01
3.44792068e-01 8.54855835e-01 6.47140861e-01 -4.75411683e-01
2.36317560e-01 4.08055261e-02 5.86202621e-01 4.96359617e-01
7.32226491e-01 -7.20315576e-01 -5.99731803e-01 -2.39519760e-01
3.20136733e-02 1.23519816e-01 -4.06059206e-01 -3.89336526e-01
-1.19082594e+00 1.13504052e+00 3.66203427e-01 9.82549787e-01
-5.79410732e-01 2.15004012e-01 9.02570903e-01 8.98862481e-02
8.53934765e-01 8.40804935e-01 -1.06755815e-01 -9.07085687e-02
-1.49258626e+00 4.10039216e-01 5.60622573e-01 6.70326829e-01
5.10828495e-01 1.93009615e-01 -5.41114986e-01 9.18074131e-01
-3.45139503e-02 2.88955629e-01 1.05873382e+00 -4.41073865e-01
3.79241168e-01 7.15233207e-01 -2.53175735e-01 -1.35498619e+00
-5.48413455e-01 -6.09040678e-01 -1.08875966e+00 -3.56427610e-01
-1.51137233e-01 3.57282069e-03 -5.12056947e-01 1.44506860e+00
-6.66805208e-02 -1.78210720e-01 3.10995817e-01 6.17925644e-01
1.22782183e+00 7.91797936e-01 -1.87729165e-01 -4.00715947e-01
1.30321097e+00 -1.01687062e+00 -1.10009873e+00 -1.39292836e-01
6.72321796e-01 -6.79433763e-01 8.44460070e-01 4.17599455e-02
-9.83601570e-01 -3.69709671e-01 -1.22521996e+00 -2.36149445e-01
-7.34189212e-01 5.12882397e-02 6.43900573e-01 5.29814482e-01
-1.21802115e+00 7.91557431e-01 -5.04367352e-01 -8.98616254e-01
4.15220201e-01 1.74219325e-01 -4.74308074e-01 2.87674457e-01
-1.07065380e+00 1.15024590e+00 1.01280129e+00 -2.94874430e-01
-6.09412827e-02 -4.60405916e-01 -9.46428597e-01 4.62216705e-01
6.32253289e-02 -8.64266574e-01 1.06515968e+00 -6.78836465e-01
-1.41050112e+00 6.05527103e-01 -2.72309989e-01 -7.46694565e-01
3.15211028e-01 -4.18377578e-01 -1.17853388e-01 5.95966756e-01
1.48547262e-01 5.65875709e-01 8.40541363e-01 -9.98483181e-01
-3.94127488e-01 -2.74194807e-01 -9.63635072e-02 3.38781923e-01
-8.36980820e-01 -3.53756920e-02 1.39370739e-01 -9.96554494e-01
3.13933119e-02 -4.00316000e-01 -2.94195116e-02 -6.07602894e-01
-3.69215876e-01 -6.11241817e-01 6.60400152e-01 -6.88044906e-01
1.40689242e+00 -1.73485422e+00 4.64677185e-01 -2.23845422e-01
3.09027195e-01 5.72706640e-01 -1.83916181e-01 1.08386052e+00
4.07894365e-02 1.04435049e-01 -3.44640106e-01 -5.02831161e-01
7.01110959e-02 5.27578872e-04 -4.73006546e-01 3.93721133e-01
1.60421506e-01 8.59999001e-01 -1.05365694e+00 -7.11433411e-01
1.79775760e-01 3.68652254e-01 -4.34078872e-01 -1.90859452e-01
-3.39122713e-02 -1.26741588e-01 -5.30834794e-01 8.62590969e-02
5.57934761e-01 2.07394689e-01 2.15282798e-01 -3.43814552e-01
-8.50509405e-02 3.48587871e-01 -6.78137839e-01 1.98998189e+00
-3.71021330e-01 8.77071142e-01 -4.58170295e-01 -1.31399786e+00
9.41520333e-01 3.40826392e-01 3.34344238e-01 -5.74348211e-01
3.10214520e-01 1.73072457e-01 -1.16365805e-01 -4.76365298e-01
1.40252554e+00 -2.15638950e-01 -2.48565480e-01 5.65781653e-01
4.28811878e-01 -4.00272846e-01 5.51577151e-01 6.73616171e-01
7.80822694e-01 -1.09666921e-01 8.60465646e-01 -6.26427710e-01
6.90424383e-01 -1.26961514e-01 1.23717390e-01 7.66890347e-01
1.22321390e-01 5.83753586e-01 5.09701908e-01 -1.35370702e-01
-1.04230356e+00 -8.97659183e-01 -6.24039695e-02 1.02687371e+00
3.51218134e-02 -8.82593453e-01 -7.04597414e-01 -6.96702182e-01
8.06487799e-02 1.06626630e+00 -5.43888152e-01 -3.40043098e-01
-5.36180019e-01 -6.86261892e-01 5.95163941e-01 5.37605524e-01
3.00313145e-01 -8.31748545e-01 -6.03503704e-01 2.34271228e-01
-2.49587342e-01 -9.67095852e-01 -4.06475693e-01 -8.73011798e-02
-1.10659838e+00 -7.50296414e-01 -8.66693795e-01 -7.75287449e-01
3.41529787e-01 3.82834285e-01 9.94245887e-01 -1.79988980e-01
-7.27079064e-02 5.22829831e-01 -7.57422328e-01 -5.28854549e-01
-5.61682761e-01 4.06301022e-01 1.21553369e-01 -1.58876091e-01
4.88185883e-01 -5.93688190e-01 -4.81894106e-01 -4.62822020e-01
-1.14324617e+00 -2.38092512e-01 5.12886226e-01 8.17123413e-01
1.83226038e-02 -1.38658047e-01 1.01317608e+00 -7.24068165e-01
1.44215631e+00 -2.46021435e-01 1.17791034e-01 2.72529989e-01
-4.52925086e-01 3.58400762e-01 7.08789647e-01 -2.70045221e-01
-7.89232194e-01 -5.46032012e-01 -1.14177398e-01 2.24419564e-01
9.14139003e-02 6.79120123e-01 1.54431134e-01 4.72387135e-01
5.06775558e-01 5.49193919e-01 7.04887584e-02 -6.13337338e-01
7.85274923e-01 8.45532835e-01 2.63642251e-01 -3.39283109e-01
5.05290687e-01 4.84661877e-01 -9.34692994e-02 -1.44850481e+00
-6.86021924e-01 -8.34032893e-01 -7.69449651e-01 -6.08533714e-03
9.37607825e-01 -7.28143096e-01 -3.36636305e-01 1.88287437e-01
-1.50125229e+00 5.49069822e-01 -5.51337719e-01 5.92729449e-01
-3.96452934e-01 1.03264868e+00 -3.93696994e-01 -5.42503536e-01
-7.99559236e-01 -6.87027514e-01 1.07244825e+00 2.61148721e-01
-5.72545648e-01 -1.30824971e+00 2.52921462e-01 2.72711545e-01
6.33935511e-01 4.16100994e-02 9.35022414e-01 -1.34629202e+00
2.45539948e-01 -3.92195374e-01 -1.05665952e-01 6.03157282e-01
3.08092505e-01 1.85704920e-02 -7.27356374e-01 -2.90531814e-01
7.39183426e-02 -1.29218236e-01 1.43353879e+00 3.97586316e-01
7.61151314e-01 -3.74194324e-01 -1.83149844e-01 -8.22086558e-02
1.32415235e+00 -3.45258147e-01 6.07527375e-01 3.33590955e-01
6.97881818e-01 6.20071113e-01 3.69066209e-01 4.62376207e-01
2.38416642e-01 4.50040132e-01 1.95971042e-01 9.76907164e-02
-7.89017379e-02 -2.42658943e-01 2.15171024e-01 1.40385211e+00
6.67084977e-02 -3.90815467e-01 -7.82406390e-01 7.52336502e-01
-1.92076838e+00 -1.22998357e+00 -2.62509614e-01 1.91123056e+00
7.15355635e-01 3.91323268e-02 9.38011184e-02 2.29617417e-01
5.95720172e-01 7.94300139e-01 -5.20182997e-02 -9.62212622e-01
-3.71400356e-01 2.38996312e-01 3.54285538e-01 4.38347012e-01
-9.90955532e-01 9.82948840e-01 6.20097017e+00 1.16000140e+00
-8.56485784e-01 2.06828684e-01 -7.15197101e-02 -1.64358635e-02
-4.99069870e-01 -2.14207277e-01 -5.29303193e-01 2.73201913e-01
7.08065867e-01 -5.39236665e-01 -3.04764748e-01 4.64247435e-01
2.56985813e-01 -2.95242131e-01 -9.89776373e-01 8.55454504e-01
9.65030849e-01 -1.48418427e+00 6.41734123e-01 -2.69192457e-01
7.06333756e-01 -1.44611582e-01 -2.11332142e-01 2.60376185e-01
-1.38475507e-01 -8.94941211e-01 4.34553653e-01 5.45655668e-01
2.13097513e-01 -8.65516543e-01 1.04885066e+00 2.88853914e-01
-7.23191082e-01 1.77054584e-01 -7.23277032e-01 -1.69170052e-01
1.15326196e-01 7.27944195e-01 -8.51844609e-01 1.13958454e+00
2.17860505e-01 8.41277659e-01 -7.57547438e-01 8.39161277e-01
-1.49116278e-01 3.53585958e-01 4.34428006e-02 -5.62588215e-01
5.01087368e-01 -3.69581044e-01 8.49742830e-01 1.66294420e+00
3.05476636e-01 -3.97055805e-01 -1.25111952e-01 6.23003602e-01
-1.27846852e-01 7.66347051e-01 -7.38969624e-01 -4.16903228e-01
2.34090641e-01 1.19080067e+00 -8.37446153e-01 -6.35193527e-01
-2.16092452e-01 1.12184966e+00 2.93417454e-01 1.71658337e-01
-4.53419626e-01 -7.21916616e-01 3.76775593e-01 -9.42488164e-02
4.95122939e-01 -4.04851466e-01 -4.75019701e-02 -1.17791283e+00
9.92593244e-02 -6.67780459e-01 6.33248836e-02 -5.60310066e-01
-1.26311266e+00 5.59653342e-01 3.41908902e-01 -1.05541396e+00
-2.56445080e-01 -3.71346086e-01 -8.94290566e-01 4.77616012e-01
-1.51493394e+00 -1.09674823e+00 2.49320865e-02 1.79183096e-01
8.25487733e-01 -5.33185542e-01 9.88944888e-01 -4.37386110e-02
-4.03959632e-01 2.36803025e-01 6.51673377e-01 -1.03254385e-01
7.96167016e-01 -1.47933209e+00 3.18011850e-01 7.45323241e-01
2.33078688e-01 7.64641345e-01 1.06979895e+00 -4.90205467e-01
-1.08554542e+00 -8.22282076e-01 1.33348787e+00 -2.00892746e-01
7.45584011e-01 -2.50835598e-01 -9.67963636e-01 3.58396292e-01
1.01895618e+00 -7.53280878e-01 9.71507072e-01 2.13687807e-01
-3.80625069e-01 -5.48648126e-02 -8.22094321e-01 7.58139670e-01
6.91996694e-01 -3.70989859e-01 -1.63418961e+00 4.05874789e-01
1.00639212e+00 2.94543833e-01 -6.04042172e-01 2.08675504e-01
2.71100134e-01 -1.10172975e+00 7.77302563e-01 -6.32505536e-01
6.60199404e-01 3.99114238e-03 -1.17714673e-01 -1.77432168e+00
-1.42462388e-01 -3.86424512e-01 -2.20224410e-02 1.52726340e+00
6.43756315e-02 -7.35620201e-01 3.61203492e-01 -2.42379278e-01
-1.36579484e-01 -4.87916380e-01 -9.30648804e-01 -8.80440056e-01
5.44892490e-01 4.75247838e-02 4.05563951e-01 9.39255178e-01
4.70124513e-01 7.71494687e-01 -1.55386627e-01 -4.76354837e-01
4.27627772e-01 1.29119426e-01 5.96071124e-01 -1.34281802e+00
1.45272896e-01 -9.77093875e-01 -7.70550013e-01 -7.91890562e-01
7.04304039e-01 -1.24871445e+00 -2.56557047e-01 -2.13522577e+00
3.12271595e-01 3.17296565e-01 -1.64817587e-01 7.33938292e-02
-2.45380118e-01 -8.29805583e-02 1.99394092e-01 4.30224948e-02
-7.95239687e-01 1.09557903e+00 8.13836336e-01 -4.90153313e-01
-2.87928469e-02 -3.80060017e-01 -7.50626087e-01 6.35992646e-01
1.01793969e+00 -4.34450328e-01 -4.20119911e-01 -3.68728131e-01
2.05226958e-01 -4.36582178e-01 3.65655795e-02 -9.33543026e-01
3.32377732e-01 3.31526488e-01 -8.38260632e-03 -8.83755803e-01
2.43971944e-01 -5.75604618e-01 -5.19079804e-01 4.96924073e-01
-4.98076290e-01 2.18569189e-01 1.71261579e-01 6.94594681e-01
-5.88267028e-01 -7.60563612e-01 3.80478472e-01 -7.43202269e-02
-4.50944722e-01 -2.87859827e-01 -6.06417656e-01 1.80432722e-01
7.71981418e-01 -3.00793737e-01 -3.07688385e-01 -4.08677429e-01
-2.22722188e-01 1.42163560e-01 4.28547561e-01 6.27567053e-01
7.75099039e-01 -1.16134286e+00 -1.11069667e+00 -1.59258172e-01
3.24239820e-01 -3.53115052e-01 2.67100692e-01 8.35320532e-01
-7.60210097e-01 7.52620935e-01 -2.11309448e-01 -4.64612216e-01
-1.39650035e+00 6.65577888e-01 7.69851729e-02 -2.64461130e-01
-8.84083748e-01 3.72312874e-01 -6.30285814e-02 -2.51951128e-01
1.47905767e-01 -3.19625556e-01 -7.76746869e-01 8.15341651e-01
5.72338700e-01 4.99097854e-01 1.82096541e-01 -7.19202042e-01
-3.05915773e-01 5.73893905e-01 -3.36764395e-01 -8.01510364e-02
1.54837918e+00 -9.60203111e-02 -5.86864412e-01 7.93011367e-01
1.37779772e+00 2.51103222e-01 -3.32483113e-01 -2.12823361e-01
2.77231812e-01 -1.48108557e-01 9.29198489e-02 -1.60448164e-01
-5.55090725e-01 9.97127175e-01 1.74550444e-01 3.91384751e-01
8.03912640e-01 -1.08990766e-01 7.44577050e-01 6.77251279e-01
-1.87508196e-01 -1.40762067e+00 1.69039577e-01 5.55036724e-01
9.94665861e-01 -1.05976343e+00 4.81006026e-01 -2.73248553e-03
-7.24437892e-01 1.39903545e+00 -4.25910130e-02 -3.24289411e-01
2.88646758e-01 -2.51990318e-01 -2.95004785e-01 -3.09892416e-01
-5.18761933e-01 -3.55819315e-01 3.40050966e-01 4.84017789e-01
6.67582333e-01 -7.67742246e-02 -1.20091939e+00 3.48561883e-01
-2.98193425e-01 -4.59882021e-01 8.77301216e-01 7.76332498e-01
-7.65467346e-01 -1.09536207e+00 -1.27211675e-01 4.88881201e-01
-5.04382849e-01 -2.57141739e-01 -6.66259944e-01 6.76598608e-01
-2.95644790e-01 1.18107092e+00 9.90784019e-02 -1.57369319e-02
2.30919793e-01 2.79255688e-01 4.88689244e-01 -7.63059378e-01
-8.49880040e-01 -1.60933882e-01 1.89545915e-01 -1.71065569e-01
-8.40154469e-01 -6.49570167e-01 -1.16813135e+00 -2.47785702e-01
-3.33619654e-01 4.39779043e-01 9.02903020e-01 9.47174430e-01
3.32569331e-01 7.69473433e-01 4.02597696e-01 -8.82536948e-01
-7.49190152e-01 -1.32490158e+00 -6.25567257e-01 6.46428347e-01
2.93904066e-01 -3.37651014e-01 -3.85453612e-01 -3.10825527e-01]
|
[12.39920425415039, 9.479446411132812]
|
fe990d25-188b-4a21-8806-dffe9ed52f22
|
global-autoregressive-models-for-data
|
1909.07063
| null |
https://arxiv.org/abs/1909.07063v2
|
https://arxiv.org/pdf/1909.07063v2.pdf
|
Global Autoregressive Models for Data-Efficient Sequence Learning
|
Standard autoregressive seq2seq models are easily trained by max-likelihood, but tend to show poor results under small-data conditions. We introduce a class of seq2seq models, GAMs (Global Autoregressive Models), which combine an autoregressive component with a log-linear component, allowing the use of global \textit{a priori} features to compensate for lack of data. We train these models in two steps. In the first step, we obtain an \emph{unnormalized} GAM that maximizes the likelihood of the data, but is improper for fast inference or evaluation. In the second step, we use this GAM to train (by distillation) a second autoregressive model that approximates the \emph{normalized} distribution associated with the GAM, and can be used for fast inference and evaluation. Our experiments focus on language modelling under synthetic conditions and show a strong perplexity reduction of using the second autoregressive model over the standard one.
|
['Jean-Marc Andreoli', 'Marc Dymetman', 'Tetiana Parshakova']
|
2019-09-16
|
global-autoregressive-models-for-data-1
|
https://aclanthology.org/K19-1084
|
https://aclanthology.org/K19-1084.pdf
|
conll-2019-11
|
['small-data']
|
['computer-vision']
|
[ 7.68787637e-02 3.04758549e-01 2.11588308e-01 -5.71986556e-01
-9.94184196e-01 -5.44480324e-01 7.99090147e-01 -2.33714655e-01
-5.96091747e-01 8.97490919e-01 5.19917309e-01 -4.66299951e-01
1.94048166e-01 -8.26285362e-01 -7.00172901e-01 -8.39190662e-01
1.02455206e-02 7.79553294e-01 9.45444405e-02 -2.42558613e-01
-2.11362094e-01 1.72413528e-01 -1.19531620e+00 1.01922937e-01
8.63682508e-01 5.87694705e-01 3.76370877e-01 1.00977564e+00
-4.34662223e-01 1.06665802e+00 -8.15384865e-01 -4.74475950e-01
2.32638698e-02 -7.00386226e-01 -5.52241206e-01 -4.04097587e-01
-2.61422135e-02 -4.75630850e-01 -1.26026928e-01 1.01599228e+00
6.89707875e-01 3.20816100e-01 7.07787514e-01 -7.70988643e-01
-2.63541311e-01 8.65102291e-01 -1.67669058e-01 -4.93358113e-02
2.60822624e-01 1.54573828e-01 9.98523772e-01 -7.70184517e-01
4.71554190e-01 1.62510490e+00 5.81436574e-01 5.41712761e-01
-1.30118752e+00 -5.89209437e-01 4.20702957e-02 -3.04239064e-01
-1.22864652e+00 -4.73219961e-01 4.88971472e-01 -3.40407461e-01
9.72860932e-01 2.13558778e-01 3.19065422e-01 1.31747890e+00
8.07666183e-02 8.25830758e-01 1.08081520e+00 -3.99283588e-01
3.37163955e-01 1.02296069e-01 1.12942241e-01 4.49970990e-01
-4.92654413e-01 1.01995030e-02 -1.84497878e-01 -2.16987222e-01
7.73437262e-01 -2.92595625e-01 -9.92000476e-03 1.79323956e-01
-9.43199337e-01 1.03950810e+00 6.69152886e-02 2.72296816e-01
-7.10833192e-01 3.71502936e-01 3.75574231e-01 1.82102442e-01
6.09485388e-01 5.12758672e-01 -4.18937773e-01 -6.85380578e-01
-1.04415369e+00 4.07655597e-01 9.41695213e-01 8.81569088e-01
7.03244269e-01 3.54603827e-01 -3.89649808e-01 1.21246254e+00
3.62520963e-01 5.63756585e-01 6.75063491e-01 -9.44818079e-01
3.62791926e-01 4.91793677e-02 6.97880760e-02 -3.77032638e-01
-3.54821891e-01 -3.27425510e-01 -8.72928858e-01 -1.18237726e-01
5.79932213e-01 -6.34618282e-01 -9.82268810e-01 1.87376118e+00
-3.24402303e-02 2.28041083e-01 2.50792474e-01 5.81876576e-01
6.53935671e-01 1.08910465e+00 4.02065128e-01 -2.71927774e-01
1.04864359e+00 -9.09572840e-01 -6.73226178e-01 -2.69468009e-01
9.22288537e-01 -5.97973287e-01 1.09518898e+00 4.88280356e-01
-1.19011545e+00 -4.07712191e-01 -5.74185371e-01 3.05222739e-02
-2.99707472e-01 1.01979487e-01 4.31464165e-01 7.41108358e-01
-1.18086636e+00 5.90258956e-01 -9.27187562e-01 -8.17522332e-02
-1.95722625e-01 2.85793394e-01 -1.45425871e-01 7.43533373e-02
-1.48060250e+00 7.83453465e-01 4.11474138e-01 1.84021845e-01
-7.30157554e-01 -5.99030793e-01 -9.66622949e-01 1.98621169e-01
8.64581391e-02 -5.67052960e-01 1.32580459e+00 -1.18177307e+00
-2.07539153e+00 4.84907955e-01 -3.22857141e-01 -5.41197002e-01
6.29163384e-01 -3.83029699e-01 -1.28205940e-01 -2.69676298e-01
-3.51969957e-01 6.67761981e-01 6.17415786e-01 -1.00436544e+00
-8.36943388e-02 -1.24915607e-01 1.54086158e-01 1.12890610e-02
-1.36290163e-01 3.06466579e-01 -5.55462658e-01 -7.37945735e-01
-3.59952003e-01 -8.87045205e-01 -3.83518606e-01 -6.10369325e-01
-4.99634087e-01 -2.60689914e-01 1.99049488e-01 -9.13585663e-01
1.40297437e+00 -1.93631732e+00 1.37661919e-01 1.75843298e-01
-2.93185323e-01 4.23310429e-01 -4.19263750e-01 4.44577307e-01
-1.80008173e-01 3.21800381e-01 -2.81653076e-01 -6.83616459e-01
-2.72291247e-02 3.82170707e-01 -3.48202795e-01 -1.29602820e-01
3.43760461e-01 8.79567504e-01 -1.00818157e+00 -3.59285831e-01
1.65099084e-01 6.05836689e-01 -6.88726902e-01 6.05932057e-01
-3.81882966e-01 4.39180017e-01 -2.64161497e-01 5.48744947e-03
6.72163367e-01 1.08878732e-01 2.76273608e-01 1.91129938e-01
-2.09984064e-01 5.27066052e-01 -1.12370765e+00 1.19732726e+00
-6.99670434e-01 4.99810457e-01 -1.90237001e-01 -9.27169204e-01
1.16494334e+00 2.54355609e-01 2.73693830e-01 -4.30564463e-01
1.52612105e-01 6.27417788e-02 -4.58516646e-03 -3.53520691e-01
6.06057048e-01 -3.05258781e-01 -7.35531896e-02 2.97126025e-01
2.98485905e-01 -3.75522643e-01 1.94305927e-01 -2.70686019e-02
8.22865844e-01 4.29354072e-01 2.94368595e-01 -1.57365277e-01
5.11224687e-01 -4.75953192e-01 5.12298048e-01 1.06783056e+00
3.47674906e-01 7.70905137e-01 9.79013324e-01 -1.78044766e-01
-1.21207678e+00 -9.91943836e-01 2.63558835e-01 1.35494494e+00
-6.43937707e-01 -4.00745094e-01 -9.07562256e-01 -4.19977933e-01
-4.54519480e-01 1.31475675e+00 -4.00500745e-01 6.36873767e-02
-7.46409893e-01 -1.02276707e+00 7.51303077e-01 6.23681664e-01
1.67350560e-01 -1.11802721e+00 -1.81502134e-01 3.52715969e-01
-1.31096169e-01 -9.23468947e-01 -2.77232200e-01 4.08959210e-01
-6.71839654e-01 -4.11324143e-01 -9.36708748e-01 -3.70282620e-01
3.59191895e-01 -4.89217877e-01 1.14386296e+00 -2.57815301e-01
3.93312424e-01 7.35157430e-02 -3.42288822e-01 -4.58097965e-01
-8.84172142e-01 1.03765666e-01 -1.66730776e-01 -1.21633843e-01
1.49298951e-01 -5.24703741e-01 -1.52757898e-01 5.20745292e-02
-9.68261003e-01 2.18101330e-02 4.61193442e-01 9.88727510e-01
4.33802009e-01 -4.24434185e-01 5.58824182e-01 -1.13234198e+00
7.43717730e-01 -4.68392730e-01 -7.41764247e-01 3.03150360e-02
-2.54237324e-01 3.61327350e-01 9.80473995e-01 -4.85301942e-01
-1.22074437e+00 -6.52890429e-02 -9.27143633e-01 -4.85751808e-01
-4.07270975e-02 8.77671838e-01 -1.37829646e-01 5.81836283e-01
4.78750825e-01 3.35616201e-01 2.13492289e-03 -6.02811158e-01
4.24283743e-01 6.71004057e-01 3.55720550e-01 -4.98661280e-01
3.85919154e-01 -1.48723930e-01 -1.47652343e-01 -1.13968778e+00
-6.68399036e-01 -2.12842733e-01 -4.92502958e-01 2.29491070e-02
9.51977968e-01 -8.77458930e-01 -6.49474621e-01 6.04845107e-01
-1.28892052e+00 -7.46411681e-01 -3.73474985e-01 7.71694481e-01
-7.89077461e-01 2.65276700e-01 -5.39514244e-01 -1.28683436e+00
-2.65636742e-01 -9.35511827e-01 1.13571346e+00 1.15832329e-01
-3.59316617e-01 -1.26459813e+00 3.41039032e-01 -9.53624099e-02
4.61889386e-01 -1.23318493e-01 9.37127888e-01 -1.09833276e+00
-1.75026834e-01 -3.25549722e-01 1.59214307e-02 8.38787496e-01
-3.10300052e-01 2.89320976e-01 -1.17576790e+00 4.06011939e-02
-2.01943040e-01 -1.93101376e-01 8.25580835e-01 5.19154072e-01
1.24774694e+00 -3.49335849e-01 5.30374795e-02 5.49597919e-01
1.11307502e+00 1.10408239e-01 1.06421554e+00 -1.23657808e-01
4.25033689e-01 3.93696100e-01 4.12433386e-01 5.50313830e-01
3.13383192e-01 6.39855266e-01 2.23096699e-01 4.16858643e-02
2.38374650e-01 -5.49772561e-01 7.56727576e-01 1.09292889e+00
-2.08618298e-01 -7.48315632e-01 -8.97474945e-01 4.01278913e-01
-1.73008585e+00 -8.99837434e-01 -1.36457577e-01 2.38862133e+00
9.17248726e-01 2.33698711e-01 2.03909293e-01 -2.98965961e-01
3.64206582e-01 2.62662321e-01 -1.25334561e-01 -7.05649555e-01
-2.44320020e-01 3.48811477e-01 3.66496891e-01 8.13777506e-01
-9.17177618e-01 1.12814248e+00 7.17939091e+00 8.85782838e-01
-1.13170838e+00 -1.18855163e-01 6.12042665e-01 3.21345516e-02
-4.96648282e-01 -3.61945927e-02 -9.16682422e-01 6.16490722e-01
1.47423911e+00 -1.91053137e-01 3.71371925e-01 8.36948156e-01
5.62866628e-01 -1.01516239e-01 -9.85016704e-01 6.31376565e-01
-1.50847077e-01 -8.05470347e-01 -1.79371275e-02 -3.16770524e-02
5.45595348e-01 1.25198215e-01 -1.11820571e-01 8.56495321e-01
8.88623595e-01 -1.07684898e+00 4.96094137e-01 8.65795672e-01
5.10116100e-01 -7.02140272e-01 1.09051597e+00 7.54750550e-01
-7.89456606e-01 8.78380463e-02 -4.59326386e-01 1.42862485e-03
4.10487860e-01 5.50848603e-01 -1.01928246e+00 4.00920242e-01
2.65595883e-01 3.26163292e-01 -8.55399072e-02 8.31433952e-01
-4.04576987e-01 1.04454792e+00 -4.85292643e-01 -2.95618176e-01
3.87977988e-01 -4.72542107e-01 6.47608995e-01 1.55865443e+00
5.35134435e-01 -1.07010052e-01 6.19569197e-02 9.18800533e-01
-4.77115959e-02 2.87278324e-01 -4.17213649e-01 -2.00916052e-01
3.06193143e-01 1.02077448e+00 -1.76616728e-01 -4.73453432e-01
-2.11545765e-01 9.71566319e-01 3.31513613e-01 5.00844359e-01
-8.53615761e-01 -2.59758890e-01 4.68573272e-01 -8.04237127e-02
2.67563611e-01 -3.69080454e-01 3.37576377e-03 -9.42754447e-01
-3.24208677e-01 -7.98272610e-01 3.11283499e-01 -1.00813854e+00
-1.16114843e+00 7.25460887e-01 2.27960736e-01 -8.47643614e-01
-1.19829440e+00 -5.06526113e-01 -6.52691722e-01 1.35939884e+00
-1.46394861e+00 -1.10112453e+00 5.06446930e-03 1.76518261e-01
5.45009553e-01 5.61250821e-02 1.14981544e+00 9.10973027e-02
-4.22906220e-01 6.19246900e-01 4.06031132e-01 1.84518889e-01
6.30786121e-01 -1.50611961e+00 5.04672050e-01 4.88032639e-01
-2.51848072e-01 7.16291904e-01 1.01574028e+00 -5.60455143e-01
-1.02905166e+00 -1.17803133e+00 1.17792547e+00 -3.13614339e-01
6.58367395e-01 -5.29857814e-01 -1.04792213e+00 8.76029432e-01
3.45418006e-02 -3.26920092e-01 5.63940287e-01 1.67904451e-01
-1.97699800e-01 1.60347790e-01 -7.56803215e-01 6.58952653e-01
5.03959894e-01 -4.04225796e-01 -4.10270959e-01 2.70594120e-01
8.46335590e-01 -2.93728977e-01 -1.04963779e+00 1.29831910e-01
5.64849913e-01 -7.77030945e-01 6.96188629e-01 -8.87439907e-01
5.07507443e-01 8.14710632e-02 -6.75133914e-02 -1.66167045e+00
-1.85044393e-01 -1.01952958e+00 2.18121588e-01 1.38570237e+00
6.91742420e-01 -5.79621613e-01 6.27331972e-01 5.49214244e-01
-3.67061570e-02 -5.53444445e-01 -8.12810004e-01 -9.26676929e-01
5.47796845e-01 -7.69647539e-01 5.42550445e-01 5.91423452e-01
-1.02984324e-01 3.10635597e-01 -7.04139233e-01 -2.04107404e-01
9.78653207e-02 -4.49184954e-01 9.69729841e-01 -9.20287251e-01
-5.76691747e-01 -3.58674973e-01 -2.31656775e-01 -1.61045158e+00
3.57887417e-01 -5.67595601e-01 5.04490972e-01 -1.24314749e+00
8.05590376e-02 -3.51054043e-01 -7.92315900e-02 3.89729857e-01
-4.05791461e-01 -1.57607704e-01 1.59682423e-01 -2.23368123e-01
-3.32773417e-01 7.84785986e-01 9.17319179e-01 2.28308812e-01
-4.24343228e-01 2.65382379e-01 -3.39134306e-01 7.33904064e-01
5.73202491e-01 -3.90059620e-01 -2.31549591e-01 -3.47798258e-01
2.95481324e-01 2.43883952e-01 7.30946511e-02 -5.95132768e-01
5.62141724e-02 -6.93752989e-02 1.68323647e-02 -5.18084526e-01
6.29404783e-01 -5.09187460e-01 1.52985245e-01 4.07524370e-02
-5.18112659e-01 1.46483779e-01 1.43288657e-01 3.19612622e-01
-3.13213259e-01 -6.31010175e-01 7.80237317e-01 -1.97420478e-01
-2.02104807e-01 -6.14276938e-02 -7.00989842e-01 8.08029696e-02
4.93390411e-01 5.73521741e-02 5.26428893e-02 -9.50264573e-01
-9.03017700e-01 2.39026114e-01 1.67926595e-01 1.94963083e-01
2.21550509e-01 -1.07858825e+00 -1.01917613e+00 2.14718819e-01
-2.69494861e-01 4.23139855e-02 2.05363214e-01 8.82371426e-01
-3.61692309e-01 3.75534296e-01 1.66672409e-01 -4.97068763e-01
-1.01920295e+00 4.00108248e-01 4.94213730e-01 -7.67326176e-01
-9.88003463e-02 7.55882025e-01 3.25515330e-01 -8.31088841e-01
9.51220244e-02 -2.07184836e-01 -3.39588076e-01 1.15450183e-02
5.71308315e-01 3.15980941e-01 -4.95083444e-02 -5.14363229e-01
-4.45535630e-02 1.60631731e-01 7.35335797e-02 -3.55521888e-01
1.43932259e+00 1.04239862e-02 -1.22307902e-02 7.58625865e-01
1.19748187e+00 8.46945345e-02 -1.24321127e+00 -8.11176077e-02
-3.24979015e-02 -1.19585842e-01 -9.48860571e-02 -5.98397315e-01
-6.30936682e-01 1.16720879e+00 2.05559522e-01 3.30336124e-01
8.88035119e-01 -2.84932941e-01 2.98325181e-01 2.42373645e-01
-1.49245456e-01 -9.79884326e-01 -2.50948280e-01 1.19481981e+00
8.11135352e-01 -8.68648767e-01 -4.38092798e-01 -2.90104538e-01
-9.05088902e-01 1.27277744e+00 2.57833719e-01 -8.14926550e-02
6.37384653e-01 6.64255440e-01 1.39308795e-01 1.99205160e-01
-9.20048594e-01 -1.62418753e-01 1.02571763e-01 6.69456065e-01
7.88824677e-01 7.03997016e-02 -1.88697711e-01 8.09737206e-01
-5.69667041e-01 1.20241068e-01 4.52593744e-01 3.70503336e-01
-3.61840963e-01 -1.10515904e+00 -1.51516974e-01 3.17447811e-01
-5.46218693e-01 -4.12133813e-01 -2.91517973e-01 6.31565988e-01
-3.25980753e-01 8.52848589e-01 3.65935266e-01 -2.20345184e-01
3.34536105e-01 5.04173994e-01 1.14950255e-01 -4.51310366e-01
-5.66558599e-01 3.81499052e-01 4.15578961e-01 -3.37009668e-01
5.51737621e-02 -6.21987224e-01 -1.03322589e+00 -1.92148760e-01
-2.73258060e-01 1.57069921e-01 8.78816664e-01 1.03425002e+00
1.21097863e-01 6.03947103e-01 5.01880705e-01 -6.37183845e-01
-9.52337325e-01 -1.35044551e+00 -4.68024105e-01 2.14932740e-01
1.10308826e-01 -1.31521657e-01 -2.94626653e-01 1.19710229e-01]
|
[11.917976379394531, 9.128934860229492]
|
0b01496a-777b-4b5a-8b97-6a5b8230470d
|
a-deep-optimization-approach-for-image
|
1904.07516
| null |
http://arxiv.org/abs/1904.07516v1
|
http://arxiv.org/pdf/1904.07516v1.pdf
|
A Deep Optimization Approach for Image Deconvolution
|
In blind image deconvolution, priors are often leveraged to constrain the
solution space, so as to alleviate the under-determinacy. Priors which are
trained separately from the task of deconvolution tend to be instable, or
ineffective. We propose the Golf Optimizer, a novel but simple form of network
that learns deep priors from data with better propagation behavior. Like
playing golf, our method first estimates an aggressive propagation towards
optimum using one network, and recurrently applies a residual CNN to learn the
gradient of prior for delicate correction on restoration. Experiments show that
our network achieves competitive performance on GoPro dataset, and our model is
extremely lightweight compared with the state-of-art works.
|
['Zhijian Luo', 'Siyu Chen', 'Yuntao Qian']
|
2019-04-16
| null | null | null | null |
['image-deconvolution']
|
['computer-vision']
|
[-1.81917965e-01 2.22637519e-01 5.93493059e-02 -2.04814836e-01
-5.56984186e-01 -2.91808277e-01 2.72602916e-01 -8.17294419e-01
-3.68319213e-01 6.85795605e-01 8.72103095e-01 -3.22390914e-01
3.13286334e-02 -2.56229013e-01 -9.50804114e-01 -7.72529125e-01
1.41135976e-01 1.74187034e-01 3.54668908e-02 -4.19298820e-02
5.82671165e-01 1.30420387e-01 -8.34047794e-01 -1.44883124e-02
8.71286869e-01 8.22683394e-01 5.30099034e-01 4.97095644e-01
1.63881138e-01 1.34116781e+00 -3.69709879e-01 -5.19098878e-01
5.56513786e-01 -4.03706789e-01 -7.66207576e-01 1.09176174e-01
4.14764643e-01 -8.80339622e-01 -8.27438235e-01 1.43521667e+00
6.04715824e-01 1.29457995e-01 6.09581947e-01 -6.82901442e-01
-1.09499025e+00 7.59075463e-01 -6.92783356e-01 3.46125633e-01
7.79826147e-03 4.90678519e-01 9.12379265e-01 -1.10592842e+00
2.00372845e-01 1.26347482e+00 9.53355312e-01 6.40747130e-01
-1.04012668e+00 -5.66827476e-01 3.89977247e-01 1.45927474e-01
-1.08315277e+00 -8.50497246e-01 6.16629958e-01 -3.45883012e-01
7.21515656e-01 -1.89834177e-01 4.76199001e-01 1.05121613e+00
1.07994445e-01 7.90069461e-01 8.59420598e-01 -6.03018068e-02
8.82473513e-02 -4.19292718e-01 -1.06447995e-01 6.38356805e-01
7.63487443e-02 1.95726857e-01 -6.82850182e-01 -3.83524713e-03
1.17287445e+00 5.59892282e-02 -1.00121951e+00 2.64542609e-01
-9.51851845e-01 7.85328567e-01 7.17346907e-01 -2.92426407e-01
-5.72713554e-01 6.90917253e-01 -6.30919356e-03 2.99665868e-01
5.83718896e-01 6.56943321e-01 -3.62457335e-01 -6.55582780e-03
-1.07393491e+00 2.84208238e-01 7.11171091e-01 7.27349222e-01
8.42476130e-01 1.28325135e-01 -4.39021826e-01 9.07363355e-01
5.76648474e-01 1.28648400e-01 5.99849284e-01 -1.54025161e+00
2.91550159e-01 -8.07872880e-03 3.83670807e-01 -7.15319991e-01
-2.67816663e-01 -7.37847567e-01 -7.91020632e-01 3.59237045e-01
5.61900556e-01 -2.38991544e-01 -1.20278871e+00 1.64654350e+00
-8.74081776e-02 4.87801254e-01 -1.68197528e-01 1.51637113e+00
6.18324161e-01 4.33536828e-01 -3.86126935e-01 9.67561305e-02
1.07077014e+00 -1.55661762e+00 -6.75346792e-01 -8.28373671e-01
2.48759799e-02 -8.75833035e-01 1.14006233e+00 6.09293997e-01
-1.36413252e+00 -6.78846613e-02 -1.12218499e+00 -4.35439020e-01
2.73203015e-01 1.89916372e-01 7.86805749e-01 2.73947209e-01
-1.29695451e+00 1.12354302e+00 -7.86394715e-01 1.70295283e-01
8.07573378e-01 2.25253463e-01 -4.35267314e-02 -2.02819109e-01
-8.71006250e-01 9.91342604e-01 1.55082390e-01 5.40906727e-01
-1.39150953e+00 -9.31880951e-01 -5.66658258e-01 1.33442402e-01
2.37505287e-01 -9.57442701e-01 1.42702413e+00 -8.64320397e-01
-2.00176048e+00 6.22061312e-01 -1.71782404e-01 -6.59132242e-01
7.34477043e-01 -5.43765366e-01 1.38889521e-01 1.30733386e-01
2.42020879e-02 5.35675943e-01 1.47480047e+00 -1.21616602e+00
-1.89203098e-01 -1.33596703e-01 4.19728130e-01 1.93313271e-01
-3.65029514e-01 -7.13712275e-02 -7.21619844e-01 -8.00617516e-01
4.66980100e-01 -6.58091307e-01 -4.56523329e-01 2.06947997e-01
-5.73809564e-01 1.11888722e-01 3.57145876e-01 -1.09846377e+00
1.06824291e+00 -2.13333416e+00 2.26207241e-01 -1.30085230e-01
6.24575377e-01 -4.91277827e-03 -8.67750868e-02 2.49986108e-02
-8.85230079e-02 6.44209608e-02 -5.43013990e-01 -7.78499961e-01
8.32600147e-03 3.27451319e-01 -6.87702775e-01 8.70252252e-01
2.96921078e-02 6.98205233e-01 -9.44254398e-01 -7.90940691e-03
-1.75241992e-01 5.39935768e-01 -9.65337038e-01 4.88500446e-01
-1.88457415e-01 5.98317623e-01 -1.34105921e-01 5.46803832e-01
8.46104145e-01 -6.10558450e-01 -9.14908946e-02 -2.90598959e-01
-3.40858009e-03 5.96496105e-01 -1.11329365e+00 1.91842020e+00
-3.87131393e-01 7.35438347e-01 3.91391695e-01 -9.03043747e-01
4.10344571e-01 3.05048913e-01 -1.28124403e-02 -3.87756020e-01
2.08351538e-01 1.26200020e-01 -2.63606697e-01 -5.63933611e-01
3.78924370e-01 -8.86327215e-03 6.81074798e-01 4.70576912e-01
2.01530263e-01 -1.76188439e-01 -1.22276237e-02 2.03665942e-01
1.08138227e+00 3.38084221e-01 -1.82932436e-01 -2.71723121e-01
2.26078838e-01 -3.23620230e-01 8.21791053e-01 9.07216489e-01
-2.27778539e-01 1.18274665e+00 5.44175923e-01 -3.45760196e-01
-8.16019356e-01 -1.01831496e+00 2.17160061e-02 9.85556483e-01
3.63316327e-01 -2.66052186e-01 -7.52083361e-01 -4.73640978e-01
-1.25374898e-01 3.60490739e-01 -4.47968066e-01 -1.05490327e-01
-6.46906555e-01 -1.13255262e+00 3.61641049e-01 5.34082115e-01
7.12919414e-01 -7.29930222e-01 -4.19863090e-02 2.92587489e-01
-3.09881181e-01 -1.01718080e+00 -1.00005233e+00 2.13533968e-01
-1.08953261e+00 -9.65311110e-01 -1.15839756e+00 -7.28012085e-01
9.09658074e-01 4.13083464e-01 1.24112380e+00 3.51520747e-01
1.30508527e-01 -9.29012224e-02 4.76461975e-03 -1.64286852e-01
9.19399317e-03 -2.71107495e-01 -1.17538072e-01 -1.61148589e-02
-1.70438632e-01 -1.07717419e+00 -1.20356262e+00 1.98766887e-01
-6.25443697e-01 -4.48361188e-02 5.01236022e-01 1.13160026e+00
4.00132030e-01 -1.05169959e-01 3.34577560e-01 -8.26788306e-01
7.91517198e-01 -4.55489993e-01 -6.85163558e-01 -1.02311976e-01
-7.59794593e-01 3.33482862e-01 6.69519484e-01 -4.11010802e-01
-1.20573366e+00 -6.76352978e-02 -1.11918924e-02 -7.54744947e-01
4.26237792e-01 5.30998111e-01 1.61607102e-01 -4.40293342e-01
8.09803903e-01 1.05854012e-01 -4.12539728e-02 -8.62302065e-01
5.13961732e-01 3.54075104e-01 1.12301397e+00 -5.58992684e-01
8.30499709e-01 6.99903607e-01 -3.85457337e-01 -2.68974572e-01
-1.29766166e+00 -2.84827679e-01 -1.64394379e-01 -1.01551205e-01
7.43044317e-01 -1.21165025e+00 -7.72986412e-01 6.85939431e-01
-1.49833417e+00 -6.17108405e-01 -4.60207686e-02 5.47808051e-01
-3.63094777e-01 4.29989576e-01 -1.02966046e+00 -5.20112872e-01
-4.70281869e-01 -1.15369225e+00 6.86042905e-01 5.15859246e-01
1.79155126e-01 -8.84848237e-01 -3.18200737e-02 4.69540298e-01
6.36675596e-01 -4.13049310e-01 2.67228037e-01 3.50990780e-02
-8.83200943e-01 1.23844050e-01 -6.59094870e-01 5.30596852e-01
-1.54490203e-01 -2.90103257e-01 -1.31781030e+00 -3.34391177e-01
3.10503513e-01 -2.85543919e-01 1.44907296e+00 6.65534973e-01
1.46978366e+00 -6.58957183e-01 5.09110503e-02 1.30556774e+00
1.30422246e+00 -4.03889984e-01 8.96184206e-01 3.63063425e-01
6.79393470e-01 1.02398843e-01 -1.51618406e-01 5.26486456e-01
5.55360496e-01 1.93676054e-02 9.45186734e-01 -8.20377748e-03
-5.69186091e-01 -2.45125607e-01 4.30693984e-01 7.97200978e-01
-4.45717722e-01 -1.98312566e-01 -5.44498265e-01 5.62086701e-01
-1.88125563e+00 -7.12855875e-01 2.20255516e-02 1.81846607e+00
1.12582898e+00 2.07762361e-01 -2.84061670e-01 -3.09786111e-01
6.89233422e-01 3.45670849e-01 -6.28911078e-01 2.42771909e-01
-1.50167570e-01 4.56487294e-04 9.56273317e-01 7.96014488e-01
-1.05485642e+00 9.63724613e-01 7.60283375e+00 4.68009144e-01
-1.04368246e+00 3.62749547e-01 4.25546110e-01 -2.01951653e-01
-3.42561096e-01 3.41341496e-01 -3.91171694e-01 6.63631082e-01
3.84844065e-01 2.60355204e-01 1.32011855e+00 6.52356923e-01
2.61630744e-01 -3.78096998e-02 -9.44663227e-01 1.05674934e+00
-2.47989427e-02 -1.54506135e+00 -3.12340617e-01 -9.37112197e-02
8.30375254e-01 5.22042513e-01 1.71862245e-01 -3.49165723e-02
8.03240657e-01 -1.16458237e+00 9.85321999e-01 8.75701070e-01
3.81515056e-01 -3.28871816e-01 6.60305858e-01 2.46425167e-01
-5.65888405e-01 -1.70946524e-01 -6.53485954e-01 -4.43086773e-01
3.77958149e-01 9.32200432e-01 -5.31670153e-01 -1.00060692e-02
8.99387181e-01 9.67335701e-01 -2.09529683e-01 1.41588950e+00
-9.60684180e-01 8.63329649e-01 -1.16591208e-01 5.08731067e-01
2.16331616e-01 -2.84126818e-01 6.84707880e-01 1.00565994e+00
2.61320114e-01 9.91429910e-02 -2.16079652e-02 1.23295701e+00
-5.87837756e-01 -4.72858638e-01 4.57471795e-03 1.11498907e-02
3.39448214e-01 9.99430954e-01 -2.91925907e-01 -1.81385949e-01
-4.69413966e-01 1.30118513e+00 5.78912258e-01 8.26723814e-01
-8.20354164e-01 -1.32041380e-01 7.60523975e-01 -1.12923950e-01
4.03787374e-01 -2.63325036e-01 -7.51650155e-01 -1.48703706e+00
1.92256011e-02 -6.16565049e-01 2.00917214e-01 -1.14815283e+00
-1.60032630e+00 3.47696334e-01 -6.23680532e-01 -1.00658977e+00
1.82489291e-01 -7.26257563e-01 -9.87829149e-01 1.16312540e+00
-2.14878225e+00 -7.42361784e-01 -3.58280718e-01 6.17039263e-01
3.53637427e-01 -6.75539076e-02 3.79814804e-01 4.29783642e-01
-8.76943707e-01 5.33649564e-01 1.01281621e-01 1.22144751e-01
9.63514686e-01 -1.25691664e+00 2.91067183e-01 1.27197313e+00
-1.89443812e-01 8.10429573e-01 8.47640216e-01 -4.38743591e-01
-1.42753315e+00 -8.41710329e-01 4.96491909e-01 -2.80857235e-01
7.37560928e-01 1.15283981e-01 -9.25228417e-01 8.60992849e-01
4.02840793e-01 2.34795213e-01 1.05664864e-01 2.89883707e-02
-5.72450519e-01 -7.09844157e-02 -9.08840418e-01 6.57984495e-01
1.13542402e+00 -6.62794709e-01 -6.21041119e-01 6.13060534e-01
8.16433728e-01 -8.61929297e-01 -4.27139014e-01 -1.19868949e-01
2.22402662e-01 -1.04068494e+00 1.34664464e+00 -4.23481464e-01
8.17107856e-01 -3.58022034e-01 2.56825574e-02 -1.50196517e+00
-6.83218479e-01 -1.22812915e+00 -4.69208539e-01 8.43994260e-01
3.65686506e-01 -7.86672115e-01 6.46161914e-01 5.99631429e-01
-6.78397298e-01 -4.70322639e-01 -7.38069415e-01 -5.81884682e-01
-1.81766916e-02 -3.30117047e-01 6.26765132e-01 7.75832891e-01
-3.62506479e-01 7.17056096e-02 -8.75806928e-01 6.26325488e-01
9.46954370e-01 -3.54782604e-02 5.20935833e-01 -9.51074123e-01
-5.93878090e-01 -7.23862588e-01 9.64905769e-02 -1.74963331e+00
-8.06935038e-03 -9.20477867e-01 4.38607484e-01 -1.56696630e+00
2.13804573e-01 -1.35119289e-01 -3.19542825e-01 5.85790575e-01
-3.78981143e-01 5.32550156e-01 2.70701349e-02 5.99327743e-01
-3.23822737e-01 6.20209694e-01 1.68512249e+00 -3.04748803e-01
-6.56606853e-02 -2.13147983e-01 -1.15615714e+00 1.12475002e+00
4.42486048e-01 -5.34885705e-01 -2.40165472e-01 -1.16692138e+00
4.73884583e-01 -9.59921405e-02 6.31490350e-01 -6.00715220e-01
6.74671829e-01 -1.31094366e-01 3.96235049e-01 -3.83106291e-01
3.18266958e-01 -3.60339969e-01 -2.96656996e-01 1.33131802e-01
-1.01425208e-01 -4.59839642e-01 -3.06705505e-01 5.27839899e-01
-1.27174675e-01 -4.53735948e-01 8.82414043e-01 -4.38866705e-01
-5.98165452e-01 6.16141081e-01 -6.78701326e-02 3.69461328e-01
2.90603633e-03 7.54003823e-02 -5.04814506e-01 -5.91504037e-01
-6.64192200e-01 3.92893583e-01 3.47654790e-01 -7.32405335e-02
6.41705155e-01 -1.11236835e+00 -7.70452678e-01 -9.84514579e-02
-4.91424412e-01 2.70472974e-01 1.56741276e-01 1.04234016e+00
-7.36167312e-01 -2.53915131e-01 1.18444033e-01 -3.75528127e-01
-2.98679471e-01 3.48069131e-01 7.37169445e-01 -3.37783732e-02
-1.12348807e+00 1.41955435e+00 3.29635918e-01 -2.66077310e-01
5.44603169e-01 -1.42999724e-01 -6.89847842e-02 -1.83508590e-01
8.01685691e-01 3.83312821e-01 6.85427990e-03 5.26767410e-02
-1.51332989e-01 3.11102122e-01 -7.87241235e-02 -7.92941749e-02
1.62004018e+00 -4.16572064e-01 -4.75391924e-01 -2.05008373e-01
9.59298491e-01 -2.16992516e-02 -2.03312373e+00 -3.16283494e-01
-3.26974899e-01 -6.78251922e-01 6.20323777e-01 -8.41983795e-01
-1.55786133e+00 8.28792989e-01 1.96465895e-01 -1.75962761e-01
1.23699272e+00 -1.37986660e-01 9.68731701e-01 3.87832522e-01
7.28811545e-04 -1.07541502e+00 2.30943903e-01 7.75533378e-01
1.32353783e+00 -1.32846701e+00 1.58478245e-01 -1.79177046e-01
-3.21674526e-01 1.24346864e+00 5.40362895e-01 -5.11387527e-01
8.07565212e-01 4.04832959e-01 2.02041626e-01 -2.26949170e-01
-3.31404746e-01 4.62268246e-03 1.59603566e-01 4.01312858e-01
2.29209498e-01 -2.99655467e-01 -1.44974232e-01 9.56658363e-01
-3.03725213e-01 1.57193735e-01 6.45976067e-01 5.52501380e-01
-4.45384979e-01 -6.25056565e-01 -5.39670527e-01 3.24444562e-01
-5.70273101e-01 -5.91157436e-01 5.48986010e-02 1.00829579e-01
-8.67669210e-02 8.38861406e-01 -2.18866587e-01 -7.38073811e-02
3.70076392e-03 -2.82245249e-01 4.35738742e-01 -2.79742569e-01
-4.29372519e-01 2.52668768e-01 -2.90264953e-02 -8.02414238e-01
-3.04769397e-01 -3.44996750e-01 -1.21678245e+00 -4.35114801e-01
-1.98775977e-01 -2.25566953e-01 4.00102466e-01 1.08440137e+00
1.99016556e-01 6.90547526e-01 6.57342970e-01 -1.19696987e+00
-7.17782497e-01 -1.01904798e+00 -4.98583317e-01 1.05519883e-01
7.71320999e-01 -5.67771196e-01 -8.48464668e-01 2.04527617e-01]
|
[11.637998580932617, -2.6745567321777344]
|
244a9c6a-7eca-4475-a0b7-0cd7b3b6ec79
|
large-scale-visual-speech-recognition
|
1807.05162
| null |
http://arxiv.org/abs/1807.05162v3
|
http://arxiv.org/pdf/1807.05162v3.pdf
|
Large-Scale Visual Speech Recognition
|
This work presents a scalable solution to open-vocabulary visual speech
recognition. To achieve this, we constructed the largest existing visual speech
recognition dataset, consisting of pairs of text and video clips of faces
speaking (3,886 hours of video). In tandem, we designed and trained an
integrated lipreading system, consisting of a video processing pipeline that
maps raw video to stable videos of lips and sequences of phonemes, a scalable
deep neural network that maps the lip videos to sequences of phoneme
distributions, and a production-level speech decoder that outputs sequences of
words. The proposed system achieves a word error rate (WER) of 40.9% as
measured on a held-out set. In comparison, professional lipreaders achieve
either 86.4% or 92.9% WER on the same dataset when having access to additional
types of contextual information. Our approach significantly improves on other
lipreading approaches, including variants of LipNet and of Watch, Attend, and
Spell (WAS), which are only capable of 89.8% and 76.8% WER respectively.
|
['Andrew Senior', 'Lorrayne Bennett', 'Hank Liao', 'Utsav Prabhu', 'Cían Hughes', 'Nando de Freitas', 'Ben Coppin', 'Matthew W. Hoffman', 'Kanishka Rao', 'Hasim Sak', 'Ben Laurie', 'Yannis Assael', 'Thomas Paine', 'Marie Mulville', 'Brendan Shillingford']
|
2018-07-13
|
large-scale-visual-speech-recognition-1
|
https://openreview.net/forum?id=HJxpDiC5tX
|
https://openreview.net/pdf?id=HJxpDiC5tX
|
iclr-2019-5
|
['lipreading']
|
['computer-vision']
|
[ 1.37309432e-01 5.83285242e-02 -2.06696004e-01 -2.78495997e-01
-1.21182001e+00 -3.97329003e-01 6.66250288e-01 -3.57829690e-01
-4.94029760e-01 4.98719335e-01 4.80161756e-01 -3.83081526e-01
7.68262863e-01 -6.01697676e-02 -7.81013668e-01 -5.97408354e-01
3.59567314e-01 8.74199811e-03 1.82545766e-01 3.12193125e-01
5.33499837e-01 2.74765074e-01 -2.14678574e+00 4.33942825e-01
3.71298581e-01 1.29734039e+00 4.77536827e-01 1.28315759e+00
-2.05810621e-01 7.47361541e-01 -4.65713531e-01 -3.32951933e-01
-2.18309332e-02 -1.29013881e-01 -5.44976771e-01 4.92549352e-02
8.63222599e-01 -7.64904082e-01 -6.63004160e-01 1.02979147e+00
9.57911611e-01 7.23998398e-02 6.27063572e-01 -1.41803145e+00
-8.51145566e-01 2.39447683e-01 -2.49445289e-01 -5.72328158e-02
6.26638830e-01 6.18116498e-01 7.18644679e-01 -1.40729237e+00
4.59612012e-01 1.30701280e+00 4.36921835e-01 9.71870601e-01
-8.09628427e-01 -6.39205694e-01 -3.60476136e-01 3.90418470e-01
-1.51864326e+00 -1.89892590e+00 3.52757990e-01 -3.28540146e-01
1.47025967e+00 -3.37739848e-02 3.56094897e-01 1.41040957e+00
-1.92385092e-01 8.37627888e-01 7.41024911e-01 -6.13400102e-01
9.93598551e-02 3.03461015e-01 -1.53365374e-01 4.53816205e-01
-1.52932331e-01 -1.03048809e-01 -1.07933760e+00 3.31855386e-01
4.40932423e-01 -4.55319762e-01 -5.30403733e-01 -1.82513922e-01
-1.22991014e+00 4.28372175e-01 -2.29741484e-01 -2.06422716e-01
-1.57589689e-01 1.09460279e-01 6.17534637e-01 1.33885428e-01
2.84739971e-01 -4.92974758e-01 -1.71620786e-01 -3.09589833e-01
-1.27130449e+00 -3.16190481e-01 8.70522201e-01 1.21169472e+00
4.28431660e-01 2.07825363e-01 -2.98502684e-01 1.10891283e+00
8.89177144e-01 8.80554318e-01 8.15700173e-01 -8.83973300e-01
6.38117373e-01 -7.61884600e-02 2.09083073e-02 -5.14923334e-01
4.31857593e-02 6.27807319e-01 -4.03373390e-01 3.40669036e-01
4.05045331e-01 -1.19037237e-02 -1.20090926e+00 1.74800515e+00
4.67937738e-02 2.01866269e-01 4.17402864e-01 8.09448540e-01
1.29604340e+00 7.97816992e-01 1.06912270e-01 -2.87172943e-01
1.37249088e+00 -1.22039688e+00 -9.25120234e-01 -1.43543154e-01
-2.62620300e-02 -1.06948352e+00 1.30797207e+00 9.26964208e-02
-1.51445651e+00 -7.05917001e-01 -8.35565984e-01 -4.77633297e-01
-3.48769993e-01 3.56879860e-01 -9.48005766e-02 9.18828547e-01
-1.97811067e+00 -7.32807145e-02 -5.13647795e-01 -6.28920615e-01
4.96591717e-01 4.30964440e-01 -4.50056940e-01 6.19383641e-02
-6.87079668e-01 6.73146963e-01 -9.20416340e-02 -2.72661716e-01
-1.33543074e+00 -5.35952270e-01 -1.21342123e+00 1.75636277e-01
1.12303860e-01 -2.67297328e-01 1.56592679e+00 -9.91576612e-01
-2.12716484e+00 1.21184015e+00 -8.16425622e-01 -2.94563949e-01
3.71243685e-01 -2.79044099e-02 -5.32634795e-01 5.56246281e-01
-1.47206277e-01 1.23881900e+00 1.33886003e+00 -8.46761465e-01
-5.44906735e-01 -1.00992627e-01 -3.42482120e-01 3.09517503e-01
-3.90924603e-01 5.71942389e-01 -8.37693095e-01 -2.76487947e-01
-4.94704962e-01 -4.96442914e-01 7.48679221e-01 3.96198004e-01
-3.14748764e-01 -4.12295580e-01 9.39030349e-01 -1.08200049e+00
8.18076789e-01 -2.27265668e+00 -2.59923398e-01 -3.89183074e-01
1.20131977e-01 6.60946846e-01 -4.12489563e-01 2.51097739e-01
9.35429782e-02 1.70016035e-01 2.23829329e-01 -8.35697234e-01
1.43315792e-01 -3.05841327e-01 -1.41435891e-01 5.39620817e-01
-1.21072441e-01 9.76572573e-01 -5.09726286e-01 -6.45338476e-01
5.08784890e-01 9.20250177e-01 -3.78367484e-01 4.09759462e-01
-1.24836937e-01 -2.40552962e-01 4.42829907e-01 9.44543898e-01
7.01887429e-01 -5.22375200e-03 7.41817802e-02 -4.25011553e-02
-8.40007514e-02 5.99189520e-01 -7.38979042e-01 1.82867730e+00
-6.14770055e-01 1.52392006e+00 5.27253568e-01 -6.26193762e-01
8.55345607e-01 8.78980756e-01 1.25898421e-01 -7.01059997e-01
8.33395272e-02 8.39461088e-02 -5.26140749e-01 -9.88737226e-01
4.62953120e-01 1.73932225e-01 4.56384540e-01 3.52408350e-01
3.37003917e-01 -1.39505789e-03 -8.33259970e-02 -9.88784526e-03
7.69259632e-01 -6.46001548e-02 1.03538290e-01 -6.90063462e-02
8.69889498e-01 -6.62759244e-01 -8.56238380e-02 4.23122317e-01
-8.01239789e-01 7.03499079e-01 4.33463812e-01 1.82108898e-02
-1.34751368e+00 -1.33410215e+00 -2.44313618e-03 1.19043028e+00
-6.76974878e-02 -3.78156900e-01 -1.09310806e+00 -1.88026905e-01
-2.75039136e-01 5.08937776e-01 -1.70687318e-01 2.11784124e-01
-1.46748826e-01 2.08413497e-01 8.85520160e-01 4.25137252e-01
5.21700323e-01 -1.46066117e+00 -2.64058709e-01 -2.78350174e-01
-4.50891286e-01 -1.40553963e+00 -1.01101816e+00 -4.06017900e-01
2.01680250e-02 -9.25619066e-01 -1.08299053e+00 -1.28881097e+00
3.17601353e-01 3.20338428e-01 8.24286282e-01 -1.86259642e-01
-2.07416698e-01 4.43219334e-01 1.13440547e-02 -2.55181253e-01
-7.74422288e-01 -1.66270778e-01 4.00478423e-01 2.37951368e-01
5.31298161e-01 -2.17875987e-01 -5.98101735e-01 2.90457815e-01
-5.77635109e-01 -1.59742795e-02 2.98773766e-01 6.56925440e-01
2.12899581e-01 -7.54645228e-01 5.62391996e-01 4.18492496e-01
5.94219744e-01 -2.12863699e-01 -6.80197239e-01 4.67177451e-01
-3.14127803e-01 -3.27310145e-01 3.08166444e-01 -6.47086918e-01
-1.04508781e+00 9.28020403e-02 -3.59884024e-01 -8.69144261e-01
-5.36783874e-01 -3.70390892e-01 -3.59928668e-01 -1.26114339e-01
2.90002435e-01 7.05340505e-01 5.85024774e-01 -3.93888175e-01
3.91435176e-01 1.84304380e+00 8.99876416e-01 1.56543463e-01
7.27515146e-02 2.33168900e-01 -6.17490709e-01 -1.49980628e+00
9.20147449e-02 -6.66433275e-01 -3.31992090e-01 -4.19828862e-01
9.71148193e-01 -1.15869892e+00 -1.25189638e+00 1.11890125e+00
-1.20836234e+00 -5.57943523e-01 6.32446706e-02 3.95527601e-01
-1.03033257e+00 5.00571966e-01 -5.58124304e-01 -1.06387305e+00
-4.17927653e-01 -1.43286860e+00 1.28397989e+00 2.93060362e-01
-2.79883165e-02 -4.82495844e-01 -4.25251126e-02 6.67999923e-01
5.40491402e-01 -5.73314011e-01 2.43455440e-01 -5.18497646e-01
-4.14261252e-01 1.54667139e-01 -5.38263738e-01 6.42492652e-01
1.54736802e-01 1.54457897e-01 -1.55487549e+00 -3.44696194e-01
-4.65485990e-01 -9.09968913e-01 9.13359284e-01 6.53349459e-01
1.06077051e+00 -5.53448975e-01 -3.53664726e-01 7.01073647e-01
1.13960469e+00 2.67798841e-01 8.79311919e-01 -1.87401444e-01
3.73517096e-01 5.67910969e-01 -5.10036089e-02 3.30939204e-01
4.74035800e-01 9.33924735e-01 2.79302895e-01 1.80517733e-01
-8.69067371e-01 -5.57431042e-01 9.26344156e-01 6.94183826e-01
2.87028700e-01 -6.48962557e-01 -9.28302288e-01 8.14323783e-01
-1.30545449e+00 -1.12261331e+00 4.28311735e-01 2.16301680e+00
8.33060205e-01 -2.93227762e-01 3.20650846e-01 1.61297873e-01
1.10542190e+00 2.86590397e-01 -5.95786154e-01 -5.94837666e-01
-2.02392344e-03 1.22861110e-01 3.76130581e-01 9.47474003e-01
-9.66780782e-01 1.27572870e+00 6.91306067e+00 8.90062034e-01
-1.36048925e+00 1.03052601e-01 4.38294470e-01 -4.36734974e-01
3.24697271e-02 -6.78016305e-01 -8.72089565e-01 6.26407444e-01
1.53153145e+00 9.46162920e-03 7.54223943e-01 7.74762690e-01
5.06388605e-01 -5.00587039e-02 -9.83597696e-01 1.56634712e+00
7.74211049e-01 -1.34116149e+00 -2.84990668e-02 -2.72557195e-02
2.63808399e-01 4.06074852e-01 2.61960298e-01 1.14654161e-01
-1.16963498e-01 -1.35839069e+00 9.30420816e-01 5.34705222e-01
1.74318457e+00 -4.09336329e-01 1.32904664e-01 2.14436188e-01
-1.26553524e+00 9.38083678e-02 -2.99786568e-01 3.17344487e-01
2.03657299e-01 -2.36610860e-01 -1.25664234e+00 -1.78724721e-01
7.41091311e-01 5.33220291e-01 -1.30718008e-01 8.32097709e-01
-1.50947481e-01 5.68887115e-01 -3.66821647e-01 -2.61176884e-01
-1.17523335e-01 5.29053509e-01 4.50171769e-01 1.33462286e+00
1.74313694e-01 -2.74109215e-01 -4.10116374e-01 5.22569776e-01
-4.32255447e-01 2.11941794e-01 -7.74614215e-01 -1.25826523e-01
6.45519078e-01 9.62877393e-01 -1.68794349e-01 -5.11813819e-01
-6.79082334e-01 9.55279350e-01 1.58025384e-01 5.50434470e-01
-7.59203196e-01 -5.88589907e-01 1.14493287e+00 -1.01399414e-01
5.36386371e-01 -9.62179378e-02 1.52967215e-01 -1.21356618e+00
1.86670348e-01 -1.09133458e+00 -1.94256783e-01 -1.33934629e+00
-7.51123428e-01 5.53631663e-01 -5.32040775e-01 -7.53177404e-01
-4.99776244e-01 -8.05287659e-01 -4.21687365e-01 9.61053967e-01
-1.68765831e+00 -1.05235350e+00 -4.54783529e-01 9.85107481e-01
1.21977019e+00 -5.89032590e-01 7.16799736e-01 1.76249146e-01
-4.28786248e-01 9.71281588e-01 1.22929089e-01 2.77878374e-01
1.01492250e+00 -6.58574760e-01 5.94403088e-01 7.30541348e-01
1.01454593e-01 2.14116901e-01 3.08032334e-01 -3.25112134e-01
-1.44187570e+00 -8.29927325e-01 1.41322792e+00 -2.59758085e-01
3.72698426e-01 -7.52066970e-01 -5.17358482e-01 5.94476759e-01
7.03554332e-01 3.62489745e-02 4.79908705e-01 -5.91091394e-01
-4.87811267e-01 -1.71404868e-01 -1.24946785e+00 6.75356805e-01
9.88216043e-01 -1.09605086e+00 -4.61595446e-01 2.56785959e-01
7.62445033e-01 -3.12168419e-01 -4.04385000e-01 5.26202237e-03
9.94758487e-01 -1.03165770e+00 1.03939474e+00 -4.23408687e-01
2.84103036e-01 4.50363941e-02 -4.91733223e-01 -8.82698298e-01
3.87372226e-01 -1.02477586e+00 -1.07388884e-01 1.36381876e+00
3.78823996e-01 -5.40374696e-01 5.19666612e-01 4.63402361e-01
-4.04740721e-02 -4.53558445e-01 -1.07413781e+00 -6.47296906e-01
-1.24990687e-01 -6.16891921e-01 4.56640512e-01 2.85694659e-01
3.68549168e-01 -4.94788662e-02 -3.89264673e-01 7.82006457e-02
5.33497334e-01 -4.95446354e-01 8.40126097e-01 -5.01824975e-01
2.43096262e-01 -4.77233082e-01 -4.65830892e-01 -1.28932369e+00
6.34700179e-01 -7.80625939e-01 3.57433498e-01 -1.37764204e+00
1.01746649e-01 1.38601348e-01 -6.44672364e-02 4.82234091e-01
2.77679443e-01 4.07859683e-01 2.30137527e-01 2.01021045e-01
-5.22686601e-01 4.94396180e-01 7.17449248e-01 -3.63240689e-01
-1.71113133e-01 -1.83969155e-01 -2.70642787e-01 5.52439034e-01
7.53287673e-01 1.26959234e-01 -4.20320898e-01 -5.06794631e-01
-4.32302117e-01 3.40852499e-01 4.61714715e-01 -8.41793180e-01
5.96397519e-01 1.02618195e-01 2.76260704e-01 -6.12941504e-01
8.45699430e-01 -3.53186607e-01 -4.03554499e-01 1.23321779e-01
-6.20486319e-01 -2.30493367e-01 3.65808666e-01 3.29473704e-01
-1.84194922e-01 3.49260382e-02 9.48347330e-01 2.08785787e-01
-8.52944374e-01 2.28270233e-01 -7.00534761e-01 1.21810488e-01
1.10016584e+00 -4.77718890e-01 -5.83633482e-01 -7.71741807e-01
-6.32022083e-01 -5.29444367e-02 5.61564505e-01 6.43584847e-01
1.07230628e+00 -1.00602138e+00 -6.87027752e-01 6.48649871e-01
-5.95686063e-02 -5.34523129e-01 1.54016748e-01 5.98523974e-01
-4.54699695e-01 9.01931643e-01 -2.96030343e-01 -7.54689634e-01
-1.75463009e+00 5.33138871e-01 4.32462573e-01 6.33754909e-01
-5.43384612e-01 9.94856298e-01 -5.03510647e-02 6.79869577e-02
8.88257563e-01 -1.75688669e-01 -9.79821458e-02 1.00968428e-01
8.79724801e-01 3.68908256e-01 1.06439367e-01 -9.24490869e-01
-6.23964787e-01 5.93811154e-01 2.15427488e-01 -5.75031042e-01
8.57404113e-01 -5.40864766e-01 3.26610893e-01 2.18504608e-01
1.47804630e+00 1.78928897e-01 -1.42333388e+00 1.04173511e-01
-5.97087383e-01 -1.93197817e-01 1.15476444e-01 -8.14040482e-01
-9.19789076e-01 1.17836285e+00 6.16149604e-01 -1.45565391e-01
1.03705144e+00 2.56472439e-01 8.03454518e-01 1.98620752e-01
-6.44777492e-02 -9.95506346e-01 -1.17808785e-02 2.61270732e-01
8.52480352e-01 -1.40213394e+00 -6.85473561e-01 -1.15892731e-01
-6.04274690e-01 1.01747143e+00 3.27474982e-01 4.20204908e-01
4.75156814e-01 4.71967340e-01 4.24919277e-01 4.06900942e-01
-1.11050165e+00 -2.91319102e-01 1.85363367e-01 1.10443246e+00
1.91423953e-01 -1.08708143e-01 3.59009773e-01 3.12581845e-02
-2.07651526e-01 2.53068417e-01 2.94222623e-01 2.57136077e-01
-5.38406074e-01 -4.80100214e-01 -2.55337566e-01 -5.46250194e-02
-3.92263055e-01 -5.08410156e-01 -2.61579692e-01 5.60539126e-01
-2.94755906e-01 1.36074698e+00 4.02988732e-01 -2.06620410e-01
-1.42604589e-01 6.32500350e-01 1.89496115e-01 -2.42936090e-01
2.97983103e-02 -4.83938269e-02 2.16867596e-01 -8.51161778e-01
-2.37357855e-01 -7.18874276e-01 -8.12206626e-01 -4.12039071e-01
-4.90373224e-02 -3.82588923e-01 1.21323776e+00 6.41107440e-01
4.56524879e-01 -6.33012876e-02 4.62772608e-01 -9.39752936e-01
-5.47773302e-01 -1.10451722e+00 -3.83383602e-01 9.84897688e-02
9.51492071e-01 -3.38607788e-01 -6.49819314e-01 4.54246372e-01]
|
[14.3401517868042, 5.008251190185547]
|
7b1ba5b0-08bf-4701-b17c-a92be2b9ee44
|
an-ensemble-of-convolution-based-methods-for
|
2305.05532
| null |
https://arxiv.org/abs/2305.05532v1
|
https://arxiv.org/pdf/2305.05532v1.pdf
|
An ensemble of convolution-based methods for fault detection using vibration signals
|
This paper focuses on solving a fault detection problem using multivariate time series of vibration signals collected from planetary gearboxes in a test rig. Various traditional machine learning and deep learning methods have been proposed for multivariate time-series classification, including distance-based, functional data-oriented, feature-driven, and convolution kernel-based methods. Recent studies have shown using convolution kernel-based methods like ROCKET, and 1D convolutional neural networks with ResNet and FCN, have robust performance for multivariate time-series data classification. We propose an ensemble of three convolution kernel-based methods and show its efficacy on this fault detection problem by outperforming other approaches and achieving an accuracy of more than 98.8\%.
|
['Chetan Gupta', 'Ahmed Farahat', 'Aniruddha Rajendra Rao', 'Lasitha Vidyaratne', 'Aman Kumar', 'Xian Yeow Lee']
|
2023-05-05
| null | null | null | null |
['fault-detection', 'time-series-classification']
|
['miscellaneous', 'time-series']
|
[-2.88782954e-01 -7.02782214e-01 4.11845297e-01 -1.55433184e-02
-1.31044075e-01 7.17654638e-03 1.31512448e-01 -1.75586328e-01
-2.62494028e-01 5.15853643e-01 -2.23942861e-01 -3.07856768e-01
-7.32480884e-01 -8.19369018e-01 -4.27777499e-01 -5.46413958e-01
-9.48369741e-01 2.95946091e-01 3.52427334e-01 -5.80414236e-01
5.85918307e-01 9.41743314e-01 -1.96737993e+00 1.41486153e-01
5.35151303e-01 1.54149127e+00 -8.25082958e-02 6.21653795e-01
2.54908830e-01 7.75512516e-01 -1.26779318e+00 3.82971823e-01
7.81467780e-02 1.03158623e-01 -8.52747083e-01 -2.19361648e-01
-2.05483735e-01 -2.47942626e-01 -7.30014801e-01 5.58879256e-01
5.60180545e-01 5.65371573e-01 6.71766996e-01 -1.27616870e+00
-7.62071788e-01 2.85616666e-01 -1.02824450e-01 7.97277987e-01
8.55196640e-02 1.63143367e-01 2.43312120e-01 -7.99069822e-01
1.69566795e-02 9.62184668e-01 1.26022863e+00 3.03213328e-01
-8.28345656e-01 -6.91582084e-01 -7.19103277e-01 6.50305152e-01
-1.16109335e+00 3.20595413e-01 1.01075029e+00 -5.33218741e-01
1.64310980e+00 2.55405277e-01 6.12103224e-01 8.62162828e-01
9.81726110e-01 3.57842445e-01 9.77122307e-01 -2.94167638e-01
2.66080916e-01 -8.12158823e-01 2.20463932e-01 3.85970950e-01
1.43621579e-01 4.23088402e-01 -4.33627397e-01 -2.49736145e-01
1.12382424e+00 4.34864193e-01 -2.34083071e-01 5.46314955e-01
-1.24787593e+00 7.46699810e-01 3.91025692e-01 6.73958123e-01
-6.56473815e-01 2.80028701e-01 9.48282361e-01 1.00724137e+00
8.49331021e-01 7.54181504e-01 -8.63973916e-01 -3.64616752e-01
-7.13043213e-01 3.96772742e-01 5.52987814e-01 3.86786997e-01
3.85373980e-01 7.08727837e-01 -1.57999713e-02 7.76837766e-01
-9.84322131e-02 1.78933054e-01 1.16391778e+00 -3.79381001e-01
3.33301574e-02 6.80751503e-01 -4.05561365e-02 -9.55876648e-01
-7.02156603e-01 -1.97608173e-01 -7.96723127e-01 6.78776085e-01
3.74038145e-02 -4.02989268e-01 -9.76907492e-01 8.38495135e-01
1.07090198e-01 4.73231524e-01 1.21818490e-01 9.31379318e-01
5.36723137e-01 3.98434788e-01 -4.00701940e-01 2.55292505e-01
1.28039885e+00 -4.60984081e-01 -8.47491086e-01 3.82586122e-01
4.99672681e-01 -7.59804964e-01 8.53229284e-01 8.20470095e-01
-5.90755999e-01 -8.80662501e-01 -1.39621902e+00 4.95927960e-01
-7.48823464e-01 2.02244177e-01 6.36499643e-01 4.24923748e-01
-9.42819178e-01 1.41214156e+00 -1.08750510e+00 -6.21927716e-02
2.90348142e-01 5.18538058e-01 -6.10200047e-01 2.04688013e-01
-1.38042843e+00 1.11115527e+00 4.26856458e-01 2.45486438e-01
-1.11954355e+00 -6.73492014e-01 -7.28549778e-01 -3.83100398e-02
2.58407146e-02 -6.38461411e-02 1.25290549e+00 -3.99375677e-01
-1.30836403e+00 -2.07663886e-02 4.94692057e-01 -6.03638113e-01
1.24121085e-01 -5.08854210e-01 -9.84562874e-01 2.60116994e-01
-2.38244161e-01 -4.24399018e-01 1.02144623e+00 -2.58481234e-01
-4.11119580e-01 -9.36996043e-02 -1.03271179e-01 -5.05624712e-01
-5.32354653e-01 3.53795648e-01 5.09727240e-01 -9.42076802e-01
1.30466938e-01 -4.52869207e-01 8.01263228e-02 -3.62911910e-01
-1.43880844e-01 -9.64723587e-01 1.57892001e+00 -8.93517792e-01
9.53099787e-01 -1.99358618e+00 -1.54553309e-01 7.60404617e-02
2.10070655e-01 2.46868670e-01 3.55965972e-01 7.80974507e-01
-6.27065361e-01 -2.60822088e-01 -2.70060211e-01 1.96472093e-01
-2.16223687e-01 2.10688010e-01 -1.06630176e-01 7.39117503e-01
9.36987877e-01 5.96709669e-01 -6.10624313e-01 1.13822065e-01
5.97184420e-01 6.00392997e-01 -5.31564727e-02 3.88289869e-01
1.71825096e-01 1.17696680e-01 -2.32845411e-01 9.68532562e-01
3.94477576e-01 1.00245431e-01 -3.76658648e-01 -5.36488891e-01
-3.65250319e-01 -1.08914301e-01 -1.00287962e+00 1.42614985e+00
-3.16700131e-01 1.07063079e+00 -2.21575007e-01 -1.63302553e+00
1.25105488e+00 7.75246918e-01 7.90480852e-01 -5.45582891e-01
2.29781121e-01 3.73659551e-01 4.81021889e-02 -1.03944731e+00
4.09801543e-01 -3.17136884e-01 -3.61602753e-02 4.27578777e-01
5.42047203e-01 1.45507038e-01 -1.87014163e-01 -6.39173627e-01
1.49330378e+00 1.89190602e-03 -5.41307509e-01 -6.28142416e-01
2.81826526e-01 1.00156769e-01 4.65720028e-01 2.85130650e-01
-7.32793510e-02 5.78161478e-01 4.66576248e-01 -9.31630254e-01
-1.01153672e+00 -6.19730532e-01 -2.47152090e-01 6.78485870e-01
-2.13369533e-01 -2.39908695e-06 -5.67731142e-01 -3.12765837e-01
5.38785279e-01 2.64281452e-01 -9.92360532e-01 -6.73154473e-01
-8.52914870e-01 -7.87932277e-01 1.04054093e+00 1.15858674e+00
5.26500285e-01 -1.16218400e+00 -8.06451142e-01 5.94165564e-01
3.25357020e-01 -5.39718390e-01 2.12661214e-02 7.45559454e-01
-1.07246625e+00 -1.43612337e+00 -5.66093326e-01 -8.78419340e-01
3.66088629e-01 -1.71118394e-01 8.09401929e-01 1.87899515e-01
-1.00847983e+00 1.89999640e-01 -5.97113490e-01 -6.68703258e-01
4.39948142e-02 -4.71492648e-01 4.73246783e-01 -2.41788309e-02
2.93942094e-01 -6.04345560e-01 -5.85455775e-01 3.78500968e-01
-1.08736444e+00 -7.80495286e-01 4.55338538e-01 1.14978015e+00
1.61793262e-01 7.38866091e-01 9.56768870e-01 -6.59589320e-02
1.22504294e+00 -7.23310590e-01 -2.02673137e-01 -1.18868619e-01
-7.80947566e-01 -1.13961592e-01 6.85985327e-01 -7.14075387e-01
-6.67864919e-01 -5.66615105e-01 -2.62993753e-01 -9.71264541e-01
-2.82939851e-01 8.27041626e-01 4.24669296e-01 -3.52196634e-01
8.10325742e-01 3.15919876e-01 6.79756045e-01 -7.75096655e-01
-2.51531124e-01 7.80769885e-01 7.97504842e-01 -7.38568604e-01
6.80040121e-01 3.18437099e-01 -1.10384673e-01 -8.72117758e-01
2.04079688e-01 -3.37860912e-01 -7.01827168e-01 -6.52832568e-01
9.05572474e-01 -6.27346337e-01 -8.87017071e-01 1.20284033e+00
-8.71908426e-01 -3.09042335e-01 -3.69799942e-01 7.11804271e-01
-5.09645224e-01 3.12415749e-01 -9.32779074e-01 -8.02006841e-01
-5.90383291e-01 -9.01449323e-01 9.50476110e-01 1.24978334e-01
5.69461174e-02 -1.07138169e+00 -1.01313248e-01 -3.30300719e-01
1.07929242e+00 1.06655455e+00 8.10766220e-01 -1.11055815e+00
3.53682935e-01 -9.86750722e-01 1.33994803e-01 9.50113714e-01
3.38441342e-01 5.01838475e-02 -7.75414944e-01 -3.58609319e-01
2.82886982e-01 -3.73638093e-01 5.80388069e-01 1.54492557e-01
1.37297010e+00 2.37659827e-01 -1.70057014e-01 2.38216236e-01
1.25011754e+00 4.24060225e-01 6.21097028e-01 4.29021090e-01
4.76993918e-01 3.54426265e-01 6.00202203e-01 4.86687124e-01
-8.93140435e-02 6.58710375e-02 5.98961115e-01 -5.75511269e-02
1.57672465e-01 5.66681147e-01 2.05098331e-01 8.23597312e-01
-4.92857397e-01 1.25276253e-01 -1.06694114e+00 6.03517473e-01
-1.74928451e+00 -1.02611291e+00 -2.37550154e-01 1.72050261e+00
4.43293929e-01 3.59473705e-01 1.12492003e-01 9.87383187e-01
7.54440069e-01 -2.65708357e-01 -6.66582823e-01 -3.05973411e-01
-2.51155607e-02 7.73683786e-01 3.58193994e-01 -6.39641583e-01
-1.35876775e+00 2.40981102e-01 7.31638908e+00 5.29994309e-01
-1.49856043e+00 7.67976940e-02 -1.03926048e-01 2.31047630e-01
3.09917510e-01 -5.44372022e-01 1.06581561e-01 3.27477783e-01
1.34363139e+00 -8.61918256e-02 3.17230105e-01 8.50512087e-01
4.89464588e-03 1.55879810e-01 -8.34495783e-01 9.06867325e-01
-1.05339952e-01 -1.41247094e+00 -5.54076493e-01 -1.28698051e-01
2.95916706e-01 3.30653310e-01 -2.83194005e-01 5.12227416e-01
-8.94518867e-02 -1.13792682e+00 6.16750121e-01 8.54357421e-01
5.94817638e-01 -1.07806575e+00 1.16587174e+00 -1.76582634e-01
-1.31702316e+00 -5.18461704e-01 -3.89015257e-01 -5.65375984e-01
-1.58149570e-01 6.95139647e-01 -6.37431026e-01 9.85114217e-01
1.35202932e+00 8.32198858e-01 -2.15008840e-01 1.13350761e+00
3.70299190e-01 9.91887808e-01 -2.58893818e-01 -9.59263742e-02
2.22765714e-01 2.87488699e-01 3.60619962e-01 8.29393983e-01
4.54464227e-01 -3.15790921e-01 6.02706000e-02 8.04663301e-01
3.69605184e-01 -3.21541309e-01 -5.95568120e-01 -2.63905168e-01
3.01376194e-01 1.18702829e+00 -7.18458354e-01 -2.87665367e-01
-5.18095970e-01 6.33782983e-01 -1.66751966e-01 1.65789887e-01
-7.83348501e-01 -1.31879258e+00 1.19163370e+00 -1.42281771e-01
4.68260884e-01 -5.86436033e-01 -1.04662478e-02 -7.79262781e-01
1.05245158e-01 -5.33644915e-01 4.83654022e-01 -8.01121652e-01
-1.72263288e+00 7.05575287e-01 1.52376503e-01 -1.57845640e+00
-1.87272176e-01 -1.38786650e+00 -1.25366926e+00 1.04111803e+00
-1.31269920e+00 -6.42037809e-01 -2.47644052e-01 9.56131816e-01
6.16815329e-01 -5.31105459e-01 8.53480756e-01 4.91323680e-01
-5.66057622e-01 3.10039401e-01 4.22943562e-01 5.37935019e-01
3.84638846e-01 -1.46632147e+00 5.56506813e-01 6.50045693e-01
-7.61418879e-01 5.09434938e-01 6.21017992e-01 -7.24986374e-01
-2.03737164e+00 -1.11711693e+00 1.39219016e-01 -1.99001998e-01
8.21534753e-01 1.10449031e-01 -1.24419487e+00 3.37008834e-01
2.64966458e-01 3.96873534e-01 6.48727119e-01 -2.46245041e-01
-6.95239827e-02 -6.72274008e-02 -1.50989807e+00 -1.52478665e-01
4.46242183e-01 -6.90161884e-01 -1.05481482e+00 4.71350878e-01
5.81930041e-01 -2.80746937e-01 -1.85955977e+00 8.38537276e-01
4.33549196e-01 -5.25456309e-01 9.25927699e-01 -9.34319913e-01
2.57792234e-01 -4.29703236e-01 4.94648665e-02 -1.46700609e+00
-4.00859684e-01 -2.95620054e-01 -3.56463164e-01 6.83067203e-01
-1.61074579e-01 -1.08383429e+00 2.17590868e-01 4.59310189e-02
-9.75482106e-01 -8.62384677e-01 -1.27965748e+00 -1.16598368e+00
8.41171443e-02 -7.33105421e-01 8.81834030e-01 1.15021646e+00
1.13106340e-01 -4.06653523e-01 -1.93717882e-01 2.36492932e-01
2.79607058e-01 -6.27468852e-03 2.50839442e-01 -1.72635257e+00
1.83281526e-01 -3.78859341e-01 -1.12178624e+00 1.51477993e-01
8.30424353e-02 -6.02420926e-01 1.68725085e-02 -1.28050470e+00
-8.33757281e-01 1.63471326e-02 -8.00129116e-01 7.71218300e-01
2.12413982e-01 1.98086321e-01 -7.24204004e-01 2.38550216e-01
1.70999497e-01 4.87811834e-01 9.63079453e-01 -1.55868575e-01
2.88076758e-01 -2.63838917e-01 7.70338774e-02 2.42947683e-01
9.95513499e-01 -8.83131698e-02 -1.07490949e-01 -2.11432353e-01
-3.00119251e-01 6.69071302e-02 6.12780988e-01 -1.59811568e+00
1.03122458e-01 6.80429637e-02 7.72821188e-01 -7.25656092e-01
4.10374776e-02 -7.58488357e-01 2.52712727e-01 8.83304477e-01
1.86867818e-01 6.91050649e-01 4.51940268e-01 5.18600225e-01
-5.80045223e-01 3.30420397e-02 2.29788840e-01 4.29627113e-03
-1.20792413e+00 1.11751474e-01 -6.71120405e-01 -5.11861444e-01
1.08633423e+00 -3.31480145e-01 -4.82264727e-01 3.38110715e-01
-7.09808350e-01 -8.67155120e-02 -7.27821812e-02 8.72370481e-01
1.14604056e+00 -1.59451115e+00 -6.87219620e-01 6.61529601e-01
1.43666536e-01 -1.26905650e-01 2.64719337e-01 1.09392416e+00
-9.32564020e-01 1.59046710e-01 -8.08463871e-01 -6.06300652e-01
-8.65932703e-01 3.44291866e-01 5.67642868e-01 4.36966717e-01
-1.06490755e+00 7.51654327e-01 -8.92095447e-01 -4.17341590e-01
-6.51366934e-02 -7.72790015e-01 -3.03129762e-01 -9.45776701e-02
4.84001428e-01 8.08589220e-01 7.32238293e-01 -3.70655894e-01
-4.57489848e-01 2.50992149e-01 3.49878848e-01 3.91296744e-01
1.66163158e+00 6.84268236e-01 -2.68712312e-01 6.48239374e-01
1.32114553e+00 -9.63676155e-01 -8.53764534e-01 7.76236653e-02
2.75183201e-01 -2.09125072e-01 3.00659090e-01 -6.18024290e-01
-1.16068792e+00 6.81233525e-01 9.82600093e-01 8.09294343e-01
1.33548117e+00 -2.04997212e-01 8.12492192e-01 5.44307709e-01
4.11567301e-01 -1.25867081e+00 2.68252373e-01 7.01510191e-01
9.91851389e-01 -8.67641985e-01 -3.63236219e-01 3.63861710e-01
-7.37470239e-02 2.00024843e+00 7.60068178e-01 -7.55800486e-01
1.12099445e+00 5.26216567e-01 -1.24455638e-01 -7.67577529e-01
-6.89319730e-01 -7.07571805e-02 4.06693578e-01 7.28067100e-01
6.05996668e-01 4.30414788e-02 -7.74172843e-02 7.63703763e-01
-1.05259925e-01 3.34210396e-02 2.12450475e-01 1.57137978e+00
-5.77157259e-01 -5.97020388e-01 -7.68634796e-01 9.29239690e-01
-6.77998662e-01 3.13669115e-01 6.34534750e-03 1.02900612e+00
4.02976424e-02 9.20411825e-01 4.89240795e-01 -1.12974501e+00
7.26882100e-01 4.15666014e-01 1.90137222e-01 -1.97125614e-01
-8.66044760e-01 -2.88242519e-01 -9.37310979e-03 -5.14192104e-01
-3.44908118e-01 -4.03305948e-01 -1.46836877e+00 -2.39010498e-01
-7.69884586e-01 3.58681083e-01 7.82978594e-01 1.07807326e+00
3.45749587e-01 1.28892934e+00 8.90261412e-01 -1.17868042e+00
-7.11726487e-01 -1.66573548e+00 -9.12580907e-01 3.44677567e-01
5.58495164e-01 -1.19625664e+00 -4.91696686e-01 -1.76304877e-01]
|
[6.832170486450195, 2.372307777404785]
|
718384b7-5230-4bf8-9fe7-fb09a8d3fbd6
|
self-supervised-point-cloud-completion-on
|
2203.10569
| null |
https://arxiv.org/abs/2203.10569v1
|
https://arxiv.org/pdf/2203.10569v1.pdf
|
Self-supervised Point Cloud Completion on Real Traffic Scenes via Scene-concerned Bottom-up Mechanism
|
Real scans always miss partial geometries of objects due to the self-occlusions, external-occlusions, and limited sensor resolutions. Point cloud completion aims to refer the complete shapes for incomplete 3D scans of objects. Current deep learning-based approaches rely on large-scale complete shapes in the training process, which are usually obtained from synthetic datasets. It is not applicable for real-world scans due to the domain gap. In this paper, we propose a self-supervised point cloud completion method (TraPCC) for vehicles in real traffic scenes without any complete data. Based on the symmetry and similarity of vehicles, we make use of consecutive point cloud frames to construct vehicle memory bank as reference. We design a bottom-up mechanism to focus on both local geometry details and global shape features of inputs. In addition, we design a scene-graph in the network to pay attention to the missing parts by the aid of neighboring vehicles. Experiments show that TraPCC achieve good performance for real-scan completion on KITTI and nuScenes traffic datasets even without any complete data in training. We also show a downstream application of 3D detection, which benefits from our completion approach.
|
['Yuexin Ma', 'Xinge Zhu', 'Peishan Cong', 'Yiming Ren']
|
2022-03-20
| null | null | null | null |
['point-cloud-completion']
|
['computer-vision']
|
[-2.08257630e-01 -3.12209390e-02 1.41829193e-01 -6.90827429e-01
-6.61234081e-01 -3.84476930e-01 4.78420436e-01 -3.08345854e-01
-2.76102647e-02 2.52076149e-01 -2.82482475e-01 -3.00203949e-01
1.51044950e-01 -1.06670403e+00 -1.21300399e+00 -3.55931699e-01
1.06400348e-01 9.41100955e-01 7.30480850e-01 -2.68879116e-01
4.44504991e-02 8.54975820e-01 -1.57210445e+00 2.06137046e-01
8.33829045e-01 8.49431515e-01 4.97416019e-01 9.41952467e-02
-6.28175259e-01 8.41896832e-02 -7.12116286e-02 -3.32775533e-01
6.01896465e-01 3.63537610e-01 -2.42586792e-01 4.49448377e-01
6.48172915e-01 -6.87469244e-01 -6.61511600e-01 9.46490824e-01
2.33487397e-01 -4.63000908e-02 5.04299700e-01 -1.37480879e+00
-1.39249757e-01 1.12833083e-01 -9.00413036e-01 -2.39498869e-01
2.71722451e-02 3.30243886e-01 6.23381674e-01 -1.33554399e+00
6.08518779e-01 1.46131253e+00 8.06990206e-01 4.78788912e-01
-9.17645693e-01 -9.38140333e-01 2.66509891e-01 2.71624833e-01
-1.49945509e+00 -5.29650807e-01 1.20334160e+00 -3.24136943e-01
7.08796024e-01 7.86577910e-02 6.20651722e-01 7.64198124e-01
-3.48206401e-01 7.36568391e-01 5.43996990e-01 -3.27969925e-03
2.78313249e-01 -1.09118603e-01 -6.97025359e-02 4.74564672e-01
4.36803341e-01 3.32459867e-01 -1.24003701e-01 -1.62775852e-02
9.94320154e-01 3.72609794e-01 1.64471507e-01 -9.91883934e-01
-1.07052362e+00 6.85202718e-01 7.16353893e-01 -1.27639577e-01
-3.75212014e-01 2.28696808e-01 1.93364099e-01 6.79155886e-02
4.35702503e-01 -2.48714402e-01 -3.43678564e-01 3.46880019e-01
-8.00539255e-01 3.31122667e-01 4.16759878e-01 1.52361619e+00
1.13975441e+00 2.07581580e-01 2.08539397e-01 7.02509582e-01
2.92360663e-01 9.23769712e-01 -2.51585603e-01 -1.02385855e+00
9.95163023e-01 7.68555105e-01 2.51295149e-01 -1.06986189e+00
-2.71751136e-01 -4.44335908e-01 -7.79128969e-01 3.02515626e-01
1.54876903e-01 8.62739086e-02 -1.05150926e+00 1.39782476e+00
7.60908723e-01 5.57074130e-01 -1.10424370e-01 1.15553916e+00
1.01086187e+00 6.35814369e-01 -1.97132230e-01 2.18021378e-01
9.24015760e-01 -6.79185390e-01 -2.65981942e-01 -3.73080462e-01
5.66175878e-01 -4.96486545e-01 8.62317681e-01 9.50271860e-02
-8.06420922e-01 -8.04109216e-01 -8.48882794e-01 -6.82950094e-02
-2.88786173e-01 1.70419708e-01 4.57944304e-01 3.58073413e-01
-7.51078010e-01 2.33049542e-01 -8.93834531e-01 -1.67515516e-01
7.27081478e-01 3.22092384e-01 -3.76144499e-01 -5.28517663e-01
-6.23272121e-01 6.17800295e-01 1.10178612e-01 2.70237774e-01
-9.27087903e-01 -7.20452249e-01 -8.61334562e-01 -4.74831350e-02
4.43769217e-01 -5.88505387e-01 9.46121454e-01 -5.70599079e-01
-8.06595564e-01 7.57092953e-01 -2.70649791e-01 -2.98476189e-01
7.31163144e-01 -8.59828293e-02 -2.58345395e-01 9.65132341e-02
2.25159734e-01 1.10314691e+00 7.83412635e-01 -1.60849845e+00
-6.46912038e-01 -5.82906544e-01 4.04673629e-02 8.63263384e-02
1.27756998e-01 -5.90115547e-01 -8.23250771e-01 -2.01594636e-01
6.78146422e-01 -9.04259264e-01 -4.29106176e-01 3.68921101e-01
-2.52954185e-01 -2.32848436e-01 1.28979397e+00 -3.25670511e-01
4.64425713e-01 -2.19568849e+00 -4.62603867e-01 3.54933918e-01
2.26602167e-01 3.35076332e-01 -5.52020788e-01 3.01494986e-01
-4.45196498e-03 -1.62707761e-01 -6.86056912e-02 -5.97319543e-01
-5.68280071e-02 5.29996395e-01 -4.76117581e-01 5.26764095e-01
4.62293804e-01 9.81796205e-01 -7.10720420e-01 -5.83651185e-01
6.81262732e-01 4.72697139e-01 -5.42931855e-01 -7.52349868e-02
-5.02068102e-01 3.98910254e-01 -9.45491970e-01 6.90505564e-01
1.56117570e+00 4.30354625e-02 -2.88409173e-01 -3.87718916e-01
-3.09564739e-01 1.50386795e-01 -1.26720357e+00 1.85578585e+00
-2.94603556e-01 3.71484309e-01 2.27678344e-01 -1.06854963e+00
1.17819250e+00 2.64592771e-03 5.75156868e-01 -8.45500588e-01
-7.72624537e-02 3.46388608e-01 -3.50901634e-01 -5.15323222e-01
2.96615332e-01 2.50079155e-01 2.24321887e-01 2.18689144e-01
-4.36264426e-01 -4.38067824e-01 3.01398747e-02 2.33906165e-01
9.34788764e-01 1.95392489e-01 -3.64146262e-01 2.33290829e-02
4.23750103e-01 3.64834279e-01 6.65158808e-01 5.65069139e-01
7.69959614e-02 9.74675953e-01 1.31213754e-01 -8.21837962e-01
-1.40561581e+00 -9.83083546e-01 -1.89303771e-01 4.44299608e-01
5.64078331e-01 -8.07740092e-02 -4.29201454e-01 -5.21971703e-01
2.02658698e-01 4.47905630e-01 -2.96593845e-01 1.20855972e-01
-9.74145234e-01 -2.79787242e-01 1.91698372e-01 6.42681003e-01
7.06180096e-01 -6.99875355e-01 -5.54564536e-01 2.53151059e-01
-1.26883015e-01 -1.49602163e+00 -3.54103684e-01 -2.79394925e-01
-1.15512621e+00 -1.07235324e+00 -4.73215848e-01 -7.74363935e-01
7.02104568e-01 1.01957130e+00 8.93605709e-01 3.07442516e-01
2.67661810e-02 7.77976662e-02 -2.09370583e-01 -4.03678596e-01
-2.21232563e-01 -2.56480984e-02 -6.71643987e-02 4.18401174e-02
5.43132007e-01 -8.76098096e-01 -6.50161147e-01 5.33399105e-01
-6.66657686e-01 3.45885545e-01 7.37738431e-01 3.85612398e-01
7.71265984e-01 -1.27462685e-01 3.10273767e-01 -4.51668143e-01
-1.63318142e-01 -4.31572676e-01 -8.96665514e-01 -6.82736859e-02
-7.57166976e-03 -1.45085722e-01 2.66971022e-01 -2.88684219e-01
-8.15339684e-01 6.08523726e-01 -3.01708311e-01 -1.14194596e+00
-3.60647917e-01 1.34109080e-01 -4.19373065e-01 -1.71293020e-01
4.52550739e-01 2.47971639e-01 3.87351103e-02 -7.03203559e-01
2.02305511e-01 5.66973746e-01 5.16550541e-01 -4.85306740e-01
1.23809993e+00 9.47972894e-01 2.54881710e-01 -8.09253454e-01
-5.06321669e-01 -5.76514065e-01 -7.64659464e-01 -2.52328247e-01
5.23546040e-01 -1.19790936e+00 -5.54073751e-01 9.35913920e-02
-1.47780228e+00 -3.59091163e-02 -2.30662584e-01 4.95005518e-01
-4.41456318e-01 5.19335032e-01 -2.16713473e-01 -6.53593242e-01
-1.40863940e-01 -1.08980346e+00 1.43620861e+00 -1.80435374e-01
5.30353785e-01 -4.68306243e-01 -2.49789253e-01 1.87732175e-01
1.76751509e-01 2.65414923e-01 5.73727071e-01 -3.87738109e-01
-1.25801563e+00 -3.13539147e-01 -5.32512307e-01 1.51302874e-01
-1.10769518e-01 -2.29264379e-01 -9.57088649e-01 -1.35881111e-01
1.53864753e-02 -6.15281202e-02 9.12244320e-01 3.78912002e-01
1.18646562e+00 8.20979401e-02 -7.70266652e-01 6.47029102e-01
1.35331905e+00 -1.27141371e-01 6.16169512e-01 1.08974688e-02
9.75096524e-01 7.67219663e-01 7.97352910e-01 3.32059950e-01
5.43448091e-01 6.13098025e-01 9.68326807e-01 -2.90381551e-01
-3.07334095e-01 -6.36559904e-01 -2.94027776e-02 5.09035349e-01
1.15858309e-01 -2.76597477e-02 -1.07628334e+00 6.46015942e-01
-1.84524918e+00 -7.96138346e-01 -7.65517235e-01 2.09791088e+00
1.99314624e-01 3.28511447e-01 -1.35371983e-01 -1.34661853e-01
8.84280145e-01 3.02921627e-02 -7.73591518e-01 3.01833540e-01
-7.05167279e-02 -1.87265605e-01 6.80340230e-01 3.73999983e-01
-9.85975921e-01 9.68167126e-01 5.34495115e+00 9.82450724e-01
-9.58912849e-01 1.90734670e-01 3.24605793e-01 4.65657413e-02
-5.19265950e-01 1.59204781e-01 -8.36365163e-01 2.68009037e-01
2.05185160e-01 3.91043007e-01 2.11031333e-01 9.17427540e-01
3.92772943e-01 1.08517535e-01 -1.16395867e+00 1.31503868e+00
-1.92269698e-01 -1.51114762e+00 1.96726188e-01 1.22888178e-01
7.92008042e-01 7.49147058e-01 -2.64918655e-01 2.19157502e-01
1.72042340e-01 -6.35284543e-01 8.54232490e-01 2.37344265e-01
8.81055057e-01 -6.77707076e-01 4.78262156e-01 7.94702888e-01
-1.44340670e+00 4.70630527e-02 -9.22797263e-01 5.72226346e-02
2.78507352e-01 7.57178247e-01 -8.56394291e-01 5.87127924e-01
5.64502120e-01 6.87943399e-01 -4.60750759e-01 1.22690749e+00
1.05468959e-01 3.71137351e-01 -8.41826499e-01 2.98895687e-01
3.35393876e-01 -2.75553405e-01 6.81704879e-01 9.20945108e-01
4.65986848e-01 2.15548873e-01 3.76395136e-01 1.20245492e+00
1.17530204e-01 -2.51142681e-01 -9.54802155e-01 5.43516576e-01
6.95344329e-01 1.18476260e+00 -6.21897578e-01 -3.93915176e-01
-6.58216000e-01 3.77869040e-01 1.42952457e-01 4.19071138e-01
-7.90690780e-01 3.08768498e-03 5.68292379e-01 5.47211587e-01
5.63381374e-01 -6.37396932e-01 -4.07436103e-01 -1.08946717e+00
3.76080900e-01 -3.24538529e-01 -1.72176346e-01 -8.96352947e-01
-1.16744852e+00 4.02086675e-01 1.72505692e-01 -1.81153822e+00
-8.41191337e-02 -2.86940753e-01 -6.99984431e-01 6.71681285e-01
-1.72398937e+00 -1.41037464e+00 -6.28685236e-01 7.46085584e-01
7.20759332e-01 6.13070652e-03 1.25480160e-01 4.62434620e-01
-1.77567273e-01 7.93378279e-02 -9.78759676e-03 1.99892297e-01
2.24384636e-01 -4.77402896e-01 8.94315183e-01 6.92683399e-01
6.36123260e-03 2.49419704e-01 5.48810005e-01 -8.89837146e-01
-1.55337346e+00 -1.43034160e+00 7.08725154e-01 -3.96405071e-01
9.54299867e-02 -6.69596493e-01 -9.87560391e-01 6.45537436e-01
-3.52632493e-01 3.61170590e-01 -1.68247238e-01 -2.23582089e-01
-2.70495176e-01 -4.60152030e-01 -1.06145477e+00 4.21017438e-01
1.49858892e+00 -2.36026973e-01 -4.77390796e-01 4.27187473e-01
9.00235057e-01 -5.57855427e-01 -1.43368006e-01 6.89122796e-01
3.14110488e-01 -7.80384600e-01 1.18914092e+00 -3.36217403e-01
2.07332298e-01 -6.52213156e-01 -2.91663319e-01 -9.73047256e-01
-2.15091556e-01 -1.67245716e-01 1.88261256e-01 1.06983054e+00
2.35056430e-02 -4.39937830e-01 1.13680220e+00 6.48628116e-01
-5.75548351e-01 -3.59733671e-01 -1.07660544e+00 -8.13691378e-01
-8.80097374e-02 -1.01010466e+00 1.28495657e+00 7.44513929e-01
-6.45449281e-01 2.45694891e-01 -1.27241746e-01 5.83649933e-01
8.48139584e-01 4.48953152e-01 1.33610654e+00 -1.49151921e+00
2.44459763e-01 -1.75959303e-03 -4.03559238e-01 -1.57874811e+00
1.01857871e-01 -7.04708278e-01 -3.82991042e-03 -1.51607871e+00
2.58464683e-02 -1.02887130e+00 2.56331146e-01 2.77262330e-01
3.38854522e-01 2.88371891e-01 1.44462377e-01 4.61565673e-01
-5.98188221e-01 8.11134756e-01 1.44852531e+00 -2.61540532e-01
-2.35322922e-01 2.21336782e-01 -1.58342734e-01 7.49274433e-01
7.71170497e-01 -5.41422665e-01 -5.17387748e-01 -1.07584596e+00
1.34366408e-01 1.94764435e-01 8.46136451e-01 -1.08483720e+00
3.52335006e-01 -3.54143977e-01 4.01757061e-01 -1.54867256e+00
7.55135417e-01 -1.21225584e+00 3.64946544e-01 2.53463507e-01
1.98074624e-01 -2.12959927e-02 1.14018388e-01 6.88556731e-01
-5.95235452e-02 -1.10255942e-01 4.37590331e-01 -3.16040754e-01
-7.67925739e-01 8.20884168e-01 2.00563505e-01 -2.30555296e-01
9.70769763e-01 -6.50228024e-01 -5.78518100e-02 -2.21927345e-01
-5.14198601e-01 6.32705867e-01 5.41518986e-01 5.48417389e-01
1.01573801e+00 -1.48825896e+00 -8.22108686e-01 6.54411614e-01
2.20657691e-01 8.21446896e-01 4.45858181e-01 6.01520061e-01
-5.62885284e-01 5.59211612e-01 -6.65749386e-02 -1.15009224e+00
-9.86211419e-01 6.19463146e-01 2.47046664e-01 3.80698591e-01
-9.07962084e-01 3.35813522e-01 6.87843084e-01 -6.49411976e-01
1.69197500e-01 -5.94451785e-01 8.97009820e-02 -3.63700509e-01
2.86281377e-01 1.34907752e-01 3.24510932e-01 -8.23269486e-01
-3.63444239e-01 9.40533340e-01 1.27349645e-01 6.56245127e-02
1.43929553e+00 -2.88081527e-01 3.40976655e-01 -3.10095102e-02
1.12353647e+00 -1.62825435e-01 -1.65068746e+00 -4.75099832e-01
-4.09210473e-01 -7.71889031e-01 2.85054892e-02 -1.07891537e-01
-1.44528329e+00 1.24227428e+00 5.98530471e-01 -2.55801648e-01
6.86313689e-01 1.61047399e-01 8.92672420e-01 6.91877127e-01
7.00929403e-01 -8.93458486e-01 -1.39687210e-01 4.90990311e-01
8.84350955e-01 -1.26769054e+00 -1.22420937e-01 -7.47818828e-01
-2.69245356e-01 1.09553576e+00 7.94715106e-01 -2.86063582e-01
6.96290612e-01 8.14135298e-02 -2.00067908e-01 -3.57982874e-01
-6.72021747e-01 -3.33231330e-01 1.13660078e-02 9.50956821e-01
-4.76737112e-01 -2.18929201e-02 2.90129315e-02 2.38963470e-01
-2.72252620e-03 -5.93168475e-02 2.73530483e-01 7.56249130e-01
-6.60739064e-01 -9.24798131e-01 -5.56783080e-01 4.46697801e-01
2.95519561e-01 2.82709092e-01 -6.16528131e-02 8.75417650e-01
3.36559504e-01 7.57200718e-01 4.00863379e-01 -3.81825477e-01
5.84995866e-01 -3.64562750e-01 2.65242875e-01 -4.70335215e-01
1.43840715e-01 2.14649051e-01 -8.74695927e-02 -6.17231786e-01
-3.31384867e-01 -7.55459547e-01 -1.42348814e+00 -4.53907549e-01
-3.48399132e-01 -1.47535026e-01 8.53539467e-01 8.32199931e-01
6.30668879e-01 -1.23272268e-02 7.49187052e-01 -1.28849781e+00
-5.08362234e-01 -7.86845863e-01 -3.19431484e-01 4.45144385e-01
3.71141136e-01 -8.14837754e-01 -8.59905258e-02 -1.83057025e-01]
|
[8.16415786743164, -3.08120059967041]
|
873f515f-37a1-4fd6-b17c-b1700a49b38f
|
computing-the-ensemble-spread-from
|
2205.09182
| null |
https://arxiv.org/abs/2205.09182v1
|
https://arxiv.org/pdf/2205.09182v1.pdf
|
Computing the ensemble spread from deterministic weather predictions using conditional generative adversarial networks
|
Ensemble prediction systems are an invaluable tool for weather forecasting. Practically, ensemble predictions are obtained by running several perturbations of the deterministic control forecast. However, ensemble prediction is associated with a high computational cost and often involves statistical post-processing steps to improve its quality. Here we propose to use deep-learning-based algorithms to learn the statistical properties of an ensemble prediction system, the ensemble spread, given only the deterministic control forecast. Thus, once trained, the costly ensemble prediction system will not be needed anymore to obtain future ensemble forecasts, and the statistical properties of the ensemble can be derived from a single deterministic forecast. We adapt the classical pix2pix architecture to a three-dimensional model and also experiment with a shared latent space encoder-decoder model, and train them against several years of operational (ensemble) weather forecasts for the 500 hPa geopotential height. The results demonstrate that the trained models indeed allow obtaining a highly accurate ensemble spread from the control forecast only.
|
['Alex Bihlo', 'Rüdiger Brecht']
|
2022-05-18
| null | null | null | null |
['weather-forecasting']
|
['miscellaneous']
|
[ 6.90571815e-02 -1.80662990e-01 4.17186171e-01 -7.73929477e-01
-7.90990949e-01 -7.70117342e-01 8.97095382e-01 -4.33436073e-02
-9.20515433e-02 1.07577169e+00 1.08133078e-01 -6.57220721e-01
-1.88422240e-02 -9.54065204e-01 -5.92485905e-01 -1.14765465e+00
-2.50007480e-01 4.67878103e-01 -3.74977559e-01 -4.49806333e-01
1.78301618e-01 4.08842146e-01 -1.85904014e+00 2.45364234e-01
1.12214065e+00 1.18570030e+00 -2.00378578e-02 1.04321992e+00
3.22102569e-02 6.94894433e-01 -5.47724485e-01 -7.71824718e-02
1.62832603e-01 -5.59785128e-01 -4.33769077e-01 -3.60591382e-01
3.20178479e-01 -5.18553913e-01 1.50545686e-01 7.64911890e-01
5.26761770e-01 3.46919715e-01 7.78757989e-01 -8.93370211e-01
-7.27459863e-02 5.40109813e-01 -9.01734829e-02 7.39470348e-02
-1.62545979e-01 1.26864791e-01 9.40983891e-01 -6.95854723e-01
2.08181649e-01 8.93196285e-01 8.51232052e-01 1.39072046e-01
-1.66465139e+00 -6.83528900e-01 1.03977263e-01 -6.12516440e-02
-1.19335055e+00 -8.97947848e-01 4.71601486e-01 -7.14270711e-01
1.06429553e+00 3.34703833e-01 3.77593517e-01 1.05742502e+00
6.49572670e-01 2.78766692e-01 1.23932123e+00 -2.67712891e-01
5.22800505e-01 1.34860158e-01 4.69353721e-02 3.41611743e-01
-2.22837612e-01 7.88709641e-01 -2.71149009e-01 -9.10011977e-02
1.75596982e-01 -1.80144519e-01 -4.73526150e-01 1.23561688e-01
-1.05460370e+00 6.73721492e-01 3.38128150e-01 2.22968444e-01
-7.18795717e-01 1.81019485e-01 3.00281942e-01 6.34574354e-01
1.12775636e+00 4.88037318e-01 -8.72167706e-01 -2.64133930e-01
-1.53347683e+00 5.57480812e-01 9.33309734e-01 2.78010577e-01
6.80510998e-01 5.01120448e-01 7.29221553e-02 3.46891880e-01
2.59686053e-01 1.12720537e+00 3.32114071e-01 -6.47685409e-01
2.06637710e-01 -5.23890257e-02 5.23916006e-01 -1.08679116e+00
-6.03743494e-01 -7.84606695e-01 -1.30698788e+00 6.11858845e-01
2.57662922e-01 -8.87424231e-01 -8.91527891e-01 1.43588364e+00
1.03920735e-01 4.66040969e-01 2.94266880e-01 6.74860835e-01
3.52083445e-01 1.17074513e+00 -1.77347153e-01 -1.42983288e-01
6.23181343e-01 -6.45167708e-01 -7.27340400e-01 -1.83989584e-01
9.24967885e-01 -5.33472240e-01 3.19679469e-01 4.10578787e-01
-6.80385828e-01 -7.44183481e-01 -1.05319738e+00 4.05069292e-01
-5.73971629e-01 3.81659776e-01 7.42779076e-02 4.32135940e-01
-1.11776876e+00 1.10734987e+00 -9.71697092e-01 1.62120253e-01
-1.64067268e-01 1.90004781e-01 -2.81599790e-01 3.81794721e-01
-1.42492521e+00 1.27714360e+00 4.63994622e-01 5.59757590e-01
-7.29829550e-01 -6.22945547e-01 -6.98673546e-01 3.56166095e-01
-2.31496155e-01 -7.00578153e-01 1.24617910e+00 -1.06245494e+00
-1.79992974e+00 6.64563924e-02 -5.59909940e-01 -5.26329994e-01
1.82592228e-01 -6.04270339e-01 -6.55822337e-01 -5.32131076e-01
-2.74023324e-01 8.74349400e-02 1.03481734e+00 -1.01387274e+00
-7.14035749e-01 -1.42735019e-01 -3.95844698e-01 6.43935502e-02
3.20706576e-01 -3.35883588e-01 5.40699720e-01 -6.09898448e-01
1.07894480e-01 -1.00500762e+00 -5.23839951e-01 -6.98712945e-01
-5.05066454e-01 3.20281714e-01 4.66282696e-01 -9.58625138e-01
1.14917898e+00 -1.99715698e+00 1.90883145e-01 4.46234912e-01
-1.29965572e-02 6.93466980e-04 -4.59029526e-02 5.79370975e-01
-2.60625064e-01 3.56809348e-02 -5.40972114e-01 -5.29372334e-01
2.00379062e-02 1.42201990e-01 -1.14184701e+00 3.32504004e-01
2.25524619e-01 6.92253232e-01 -6.04200006e-01 2.18629748e-01
4.52888608e-01 5.60005069e-01 -2.67110288e-01 5.07834554e-01
-5.09636045e-01 1.16502833e+00 -2.52259552e-01 -1.12073109e-01
7.67923236e-01 -1.17498942e-01 2.06228286e-01 8.47897381e-02
-5.56509674e-01 4.42568064e-01 -1.02743578e+00 1.21926045e+00
-8.08903992e-01 6.82112515e-01 -1.62834048e-01 -1.07981360e+00
1.20218921e+00 6.41661704e-01 1.92695424e-01 -3.11775208e-01
6.79867268e-02 3.18985760e-01 4.43559326e-02 -1.36586085e-01
4.52243865e-01 -4.52532113e-01 7.94235021e-02 7.59472966e-01
8.66623446e-02 -2.71242321e-01 -3.86254340e-01 -4.08775091e-01
5.83268821e-01 3.53288561e-01 1.85959965e-01 -3.83735090e-01
4.09527481e-01 -1.65262297e-01 3.61166537e-01 8.72954130e-01
5.39046109e-01 5.74761331e-01 3.91623467e-01 -8.88194203e-01
-1.19262779e+00 -6.74631715e-01 -3.95518392e-01 1.02836621e+00
-4.98153448e-01 -2.25033671e-01 -4.57917869e-01 -3.75646710e-01
-5.11554554e-02 1.33921385e+00 -5.60437143e-01 1.30528882e-01
-3.05536807e-01 -1.03708959e+00 4.44259167e-01 3.78098279e-01
3.64926755e-01 -6.96175039e-01 -4.99785274e-01 5.98603427e-01
-1.11929171e-01 -6.90584898e-01 1.94161862e-01 2.85466194e-01
-8.88952851e-01 -4.25271630e-01 -6.48692489e-01 2.01128200e-01
2.26220608e-01 -4.46064770e-01 1.29543483e+00 -4.08259965e-02
5.67558289e-01 -2.47692779e-01 -1.84106663e-01 -5.12184918e-01
-4.65006620e-01 1.78349078e-01 3.34769130e-01 3.53731453e-01
8.42805803e-02 -1.00114310e+00 -3.23892653e-01 5.43318093e-02
-6.17606819e-01 2.80914724e-01 1.44706264e-01 1.05140555e+00
3.50955904e-01 2.41017103e-01 7.17413068e-01 -6.91856265e-01
3.14059287e-01 -5.15407681e-01 -1.17608047e+00 2.50436783e-01
-8.39837134e-01 4.17973012e-01 8.32265377e-01 1.84812263e-01
-1.28293502e+00 1.63311064e-01 -3.78767669e-01 -1.56240001e-01
-3.67050737e-01 1.02080190e+00 1.34804130e-01 3.46793681e-01
4.98374104e-01 5.21619439e-01 -1.20444894e-01 -7.19854772e-01
5.55687726e-01 5.79664767e-01 5.31243145e-01 -2.98801214e-01
6.19591415e-01 2.42967337e-01 3.53846024e-03 -7.15385079e-01
-9.37581122e-01 5.22513576e-02 -9.51091528e-01 -2.74305910e-01
7.02454150e-01 -1.10479987e+00 -4.96189177e-01 6.59400821e-01
-1.43279731e+00 -3.46469015e-01 -2.23541111e-02 7.53220856e-01
-3.68193567e-01 -3.17577690e-01 -1.84346199e-01 -1.13711846e+00
-4.75480437e-01 -9.21630085e-01 1.19798481e+00 -3.65308560e-02
-3.06302011e-01 -1.26045203e+00 7.12237716e-01 -5.37898302e-01
7.77121246e-01 4.94808525e-01 8.63241017e-01 -6.68636262e-01
-4.65868473e-01 -5.09218454e-01 4.68228534e-02 4.67145175e-01
3.44091840e-02 4.26192313e-01 -1.44891930e+00 -1.85827062e-01
2.36553684e-01 1.51780143e-01 1.03892267e+00 7.54537642e-01
1.01269913e+00 -3.98169398e-01 -3.40914071e-01 8.51873815e-01
1.26443839e+00 2.89652627e-02 4.61986005e-01 4.90134098e-02
4.42080408e-01 5.99172413e-01 8.97435024e-02 5.27902246e-01
1.55510351e-01 5.22864163e-01 2.30264619e-01 6.87683895e-02
7.70973802e-01 5.50800888e-03 4.29018140e-01 9.74375665e-01
-6.36601925e-01 -4.90307480e-01 -1.23976922e+00 2.69782573e-01
-1.80064070e+00 -1.22846007e+00 -2.09715813e-01 2.38995957e+00
5.27014911e-01 -9.53667015e-02 -4.33052212e-01 4.02752236e-02
2.72504210e-01 5.43144584e-01 -4.83801305e-01 -5.01337409e-01
-2.82814413e-01 3.29755396e-01 4.68687654e-01 7.58864462e-01
-1.37462747e+00 5.77733278e-01 7.08653450e+00 3.73057038e-01
-1.56202626e+00 -6.10139370e-02 8.25153410e-01 -1.24823123e-01
-2.50714988e-01 -4.26481618e-03 -8.23453903e-01 6.10637367e-01
1.80080163e+00 -3.34385365e-01 3.16925257e-01 7.06226110e-01
3.54340464e-01 -1.09693766e-01 -1.25318706e+00 6.77638531e-01
-5.49487472e-01 -1.58595979e+00 -3.06681544e-01 9.27379653e-02
9.99809802e-01 3.71492624e-01 1.29200280e-01 3.88032377e-01
5.02953529e-01 -1.35736680e+00 3.63437355e-01 1.27712882e+00
7.24360704e-01 -8.22619975e-01 1.24265218e+00 7.86895931e-01
-8.90717208e-01 1.31554320e-01 -3.06529462e-01 -4.31264579e-01
4.16699529e-01 1.16068292e+00 -5.04330516e-01 7.22706378e-01
6.61669970e-01 8.88414383e-01 -9.13566425e-02 5.41656375e-01
-5.28075173e-02 9.34470952e-01 -5.34284592e-01 2.82253534e-01
1.65560573e-01 -3.30849499e-01 5.48956454e-01 9.82022464e-01
9.35714483e-01 3.76041710e-01 -7.88439512e-02 7.13327527e-01
2.17765138e-01 -2.27545410e-01 -8.78890038e-01 1.27814617e-02
1.63224369e-01 9.00033593e-01 -1.61134839e-01 -6.16657794e-01
-4.64626774e-03 7.61045158e-01 6.88785613e-02 5.71870327e-01
-6.00344479e-01 -1.31114453e-01 8.82914901e-01 -4.12852466e-01
5.02829492e-01 -3.19971830e-01 -5.76202095e-01 -1.24315798e+00
-3.48950624e-01 -8.27471495e-01 -2.52528578e-01 -8.37506294e-01
-1.26385736e+00 1.13553190e+00 -2.20788822e-01 -1.43369675e+00
-1.01061654e+00 -4.81450289e-01 -9.98995900e-01 1.65338910e+00
-1.22059357e+00 -6.53107643e-01 -3.38033549e-02 -6.59254938e-02
5.07660396e-02 -2.58141905e-01 1.50359666e+00 -1.10112503e-01
-5.09419799e-01 1.50341079e-01 1.14564943e+00 -1.88203335e-01
5.82425117e-01 -1.43175673e+00 7.41645515e-01 6.80639684e-01
-5.03432704e-03 2.30208635e-01 1.11824381e+00 -5.19385159e-01
-8.79923105e-01 -1.22195280e+00 1.15585971e+00 -6.77997410e-01
5.30003309e-01 -1.12350620e-01 -1.26795614e+00 7.24703550e-01
3.15569460e-01 -1.82806719e-02 7.23284364e-01 4.35684592e-01
-1.26122665e-02 -1.44841149e-01 -7.01203287e-01 1.97211385e-01
2.57620782e-01 -6.87003195e-01 -4.90129054e-01 2.04846591e-01
4.12935019e-01 -3.79389644e-01 -9.44688261e-01 5.04578590e-01
7.94325888e-01 -1.15512192e+00 5.52133799e-01 -7.96814501e-01
7.83933163e-01 -4.04382050e-01 -4.04340267e-01 -2.16012788e+00
-2.97182024e-01 -3.16808522e-01 -1.64904743e-01 9.53085721e-01
7.49647379e-01 -9.10779655e-01 3.17715317e-01 6.85186684e-01
6.14865273e-02 -6.48550570e-01 -9.52870250e-01 -5.57577252e-01
4.11062896e-01 -7.41520166e-01 1.18534303e+00 1.00676990e+00
-3.89102280e-01 3.50903928e-01 -7.51303554e-01 8.16421092e-01
5.23097456e-01 5.88659167e-01 8.21928322e-01 -1.41267860e+00
-2.83938050e-01 -2.61861056e-01 -9.84695703e-02 -1.05622351e+00
1.86288550e-01 -5.38917184e-01 3.36168289e-01 -1.01454449e+00
-6.04331315e-01 -4.74487215e-01 -4.43172634e-01 2.17481345e-01
-1.01055898e-01 -9.92019624e-02 -1.28395215e-01 2.22360879e-01
9.77097675e-02 9.62755263e-01 6.84509456e-01 1.51619166e-01
-1.68848380e-01 3.84089470e-01 -3.71518470e-02 7.88803160e-01
8.14122736e-01 -4.45838213e-01 -1.75659969e-01 -4.95739013e-01
3.88658464e-01 3.83725524e-01 3.67333531e-01 -1.11785531e+00
1.66866779e-01 1.56174842e-02 6.67672932e-01 -9.07783449e-01
2.51935869e-01 -6.41816854e-01 5.85492969e-01 -2.55428068e-03
-3.68014395e-01 1.59646526e-01 4.25710559e-01 4.29980487e-01
-4.90898043e-01 7.23297074e-02 4.33189154e-01 3.42888832e-01
-4.77537483e-01 1.96153238e-01 -7.02312648e-01 -5.31602323e-01
6.05478644e-01 1.78329781e-01 3.03611159e-02 -7.36165762e-01
-1.01755500e+00 2.52108395e-01 2.40743130e-01 7.62772635e-02
2.61873066e-01 -1.05749083e+00 -1.28538036e+00 6.04387581e-01
-2.92520940e-01 -6.50951713e-02 5.16499400e-01 8.38051736e-01
-3.36719453e-01 6.18078589e-01 -5.12512103e-02 -6.67955875e-01
-8.28142107e-01 2.42264077e-01 8.06113422e-01 -3.29204232e-01
-5.78814685e-01 8.36511314e-01 1.71412349e-01 -8.46165001e-01
-3.36035609e-01 -5.03134191e-01 -2.90416330e-01 1.52880251e-01
8.30212951e-01 1.14697360e-01 1.93290785e-01 -5.91525316e-01
-3.29970643e-02 4.49853450e-01 4.15523022e-01 -4.01871741e-01
1.49159348e+00 -3.42424929e-01 -1.54602170e-01 1.00458694e+00
1.01780295e+00 -2.03609958e-01 -1.27476251e+00 -3.27613920e-01
1.06985435e-01 -1.83788553e-01 5.75085878e-01 -9.48853791e-01
-9.26527560e-01 1.25522709e+00 6.77691996e-01 4.26592022e-01
1.17124176e+00 -5.02625406e-01 4.87641573e-01 5.75611353e-01
1.37650073e-01 -7.73238182e-01 -1.26253176e+00 1.02651095e+00
1.04840410e+00 -1.34303176e+00 -8.42649043e-02 4.51748848e-01
-5.99309802e-01 1.38436103e+00 -2.09352002e-02 -1.82253927e-01
1.32027030e+00 2.92575866e-01 2.19237417e-01 4.63705659e-02
-1.33092403e+00 4.51320596e-02 6.08309269e-01 2.38331065e-01
5.92707396e-01 2.97064364e-01 5.41311681e-01 6.28704965e-01
-6.17672563e-01 1.57732051e-02 1.37118399e-01 3.44806373e-01
-3.84538591e-01 -7.54362106e-01 -4.81334716e-01 7.38083541e-01
-3.22341830e-01 -4.61374879e-01 1.99072689e-01 1.84756041e-01
9.85653400e-02 7.74797142e-01 4.84406948e-01 -6.71690941e-01
7.75656849e-02 6.99628353e-01 -6.19718134e-02 -3.24620426e-01
-3.93997997e-01 -9.84241515e-02 2.07806036e-01 -4.80721831e-01
-2.75390238e-01 -6.00083053e-01 -6.73776805e-01 -7.97456801e-01
-2.68457264e-01 3.70232224e-01 8.22391152e-01 1.35292399e+00
4.84273404e-01 6.19113922e-01 9.83085692e-01 -1.39205611e+00
-6.47975206e-01 -1.24383783e+00 -7.12972522e-01 -1.68910205e-01
9.40282464e-01 -5.00199556e-01 -5.58021605e-01 2.75267631e-01]
|
[6.580016613006592, 3.0071358680725098]
|
424df873-513a-4f4e-a501-ef2d97430311
|
learning-deep-context-aware-features-over
|
1710.06555
| null |
http://arxiv.org/abs/1710.06555v1
|
http://arxiv.org/pdf/1710.06555v1.pdf
|
Learning Deep Context-aware Features over Body and Latent Parts for Person Re-identification
|
Person Re-identification (ReID) is to identify the same person across
different cameras. It is a challenging task due to the large variations in
person pose, occlusion, background clutter, etc How to extract powerful
features is a fundamental problem in ReID and is still an open problem today.
In this paper, we design a Multi-Scale Context-Aware Network (MSCAN) to learn
powerful features over full body and body parts, which can well capture the
local context knowledge by stacking multi-scale convolutions in each layer.
Moreover, instead of using predefined rigid parts, we propose to learn and
localize deformable pedestrian parts using Spatial Transformer Networks (STN)
with novel spatial constraints. The learned body parts can release some
difficulties, eg pose variations and background clutters, in part-based
representation. Finally, we integrate the representation learning processes of
full body and body parts into a unified framework for person ReID through
multi-class person identification tasks. Extensive evaluations on current
challenging large-scale person ReID datasets, including the image-based
Market1501, CUHK03 and sequence-based MARS datasets, show that the proposed
method achieves the state-of-the-art results.
|
['Zhang Zhang', 'Xiaotang Chen', 'Kaiqi Huang', 'Dangwei Li']
|
2017-10-18
|
learning-deep-context-aware-features-over-1
|
http://openaccess.thecvf.com/content_cvpr_2017/html/Li_Learning_Deep_Context-Aware_CVPR_2017_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2017/papers/Li_Learning_Deep_Context-Aware_CVPR_2017_paper.pdf
|
cvpr-2017-7
|
['person-identification']
|
['computer-vision']
|
[-3.11670661e-01 -6.86008215e-01 3.12222809e-01 -6.06006742e-01
-3.74843568e-01 -4.33798552e-01 5.12466967e-01 -3.96634191e-01
-6.88404977e-01 7.42714167e-01 4.76106405e-01 5.21610618e-01
1.06072584e-02 -4.84046966e-01 -8.06015253e-01 -7.22479582e-01
1.21988140e-01 4.20555919e-01 1.96549252e-01 -2.70992100e-01
-2.10303307e-01 5.10686696e-01 -1.68741858e+00 9.21219364e-02
6.93674743e-01 6.66782081e-01 -6.89952075e-02 3.90769988e-01
9.82176065e-02 2.71862507e-01 -7.54211783e-01 -8.59443605e-01
5.19505322e-01 5.16080251e-03 -6.93461478e-01 -2.38999333e-02
1.12550092e+00 -4.13574338e-01 -6.45350993e-01 1.09028769e+00
9.76052582e-01 4.38206315e-01 5.08985579e-01 -1.15357482e+00
-7.46946573e-01 2.17608526e-01 -8.20702672e-01 3.02075088e-01
6.42885387e-01 9.74364281e-02 8.65407661e-02 -7.14577019e-01
4.28218573e-01 1.84307468e+00 1.01352501e+00 9.20778036e-01
-1.03803396e+00 -9.63635564e-01 6.51004136e-01 4.91616488e-01
-1.71738434e+00 -5.14204860e-01 5.28217971e-01 -3.74362200e-01
4.71237332e-01 1.93958044e-01 5.59052765e-01 1.59443748e+00
-1.90538317e-01 8.65526438e-01 9.50236440e-01 -1.54083535e-01
-3.63067746e-01 6.32134303e-02 4.12093610e-01 5.71876705e-01
5.46650350e-01 1.73979521e-01 -3.52634490e-01 3.63913993e-03
8.28790963e-01 4.30827647e-01 -2.82928467e-01 -3.86387140e-01
-1.14990377e+00 3.50943059e-01 7.36201048e-01 -1.66805331e-02
-2.37140618e-02 5.95943965e-02 6.59137070e-01 1.30723007e-02
9.27233174e-02 -3.07746649e-01 -3.94081265e-01 2.67029464e-01
-5.44660807e-01 5.09344518e-01 5.29093385e-01 1.30241585e+00
4.74630564e-01 -9.78697762e-02 -5.76244235e-01 1.09495676e+00
2.75172323e-01 7.82234788e-01 5.56948483e-01 -2.92514205e-01
4.98416364e-01 5.43423057e-01 2.81029284e-01 -9.00682449e-01
-5.22388458e-01 -6.32202744e-01 -1.27259612e+00 -8.91626179e-02
4.47510421e-01 -2.18867466e-01 -1.06578374e+00 1.99561262e+00
6.68785632e-01 4.53020632e-01 -2.36337349e-01 1.21470594e+00
1.21797383e+00 1.95066914e-01 4.68008578e-01 1.97692737e-01
1.74638808e+00 -1.12917840e+00 -4.96437699e-01 -3.69147629e-01
4.50239964e-02 -5.33079684e-01 2.50515580e-01 -8.15543681e-02
-7.78644323e-01 -1.23350716e+00 -8.89693081e-01 -1.24515213e-01
-6.10978305e-01 4.10126895e-01 4.91344631e-01 7.67954409e-01
-9.36126351e-01 3.19179654e-01 -3.65860969e-01 -7.81351984e-01
3.18661273e-01 6.20370150e-01 -9.08050299e-01 -3.98504704e-01
-1.06804752e+00 8.61564100e-01 3.03523064e-01 5.46661854e-01
-6.39212847e-01 -4.01551396e-01 -1.11132908e+00 -1.66580528e-01
1.56895146e-01 -1.39113343e+00 8.66907418e-01 -7.66460001e-01
-1.28256834e+00 9.59150136e-01 -3.82454395e-01 -2.67616212e-01
7.13108599e-01 -5.58789194e-01 -7.06602931e-01 -2.13811427e-01
1.77994058e-01 6.74542904e-01 8.56401443e-01 -1.22959220e+00
-6.70966148e-01 -8.41436982e-01 -5.49048409e-02 2.69845694e-01
-1.70669481e-01 4.99448985e-01 -7.58173108e-01 -7.08351016e-01
-1.58963636e-01 -1.08178449e+00 -8.94038454e-02 -2.59550177e-02
-3.47474724e-01 -3.47171515e-01 5.88503957e-01 -1.02320445e+00
6.42744601e-01 -1.88483047e+00 2.27063969e-01 -5.77185750e-02
6.44387584e-03 3.56753439e-01 -1.84954748e-01 -1.76937014e-01
-2.31087178e-01 -3.49112779e-01 1.76976010e-01 -6.70932889e-01
3.72458696e-02 1.31072551e-02 1.70782968e-01 7.82129228e-01
-8.47559348e-02 1.10373044e+00 -6.29446268e-01 -5.70321143e-01
4.67104077e-01 7.75124490e-01 -9.40846726e-02 1.98820829e-01
6.79492414e-01 7.10561991e-01 -2.99095839e-01 7.98993289e-01
1.13121426e+00 1.83327496e-01 -3.29215288e-01 -4.56543416e-01
-2.30446365e-02 -3.79144013e-01 -1.47458470e+00 1.92503393e+00
-3.80896144e-02 2.20022470e-01 1.92328140e-01 -8.74170065e-01
7.15465069e-01 2.20089108e-01 2.14123189e-01 -6.85269654e-01
2.08857417e-01 -2.05373242e-01 -1.87645152e-01 -3.67321104e-01
3.97984117e-01 1.87503010e-01 -1.81477368e-01 -7.62836784e-02
1.73748419e-01 9.26896989e-01 5.66806197e-02 -1.98151469e-01
5.66141784e-01 2.87479699e-01 1.39241830e-01 -2.98320830e-01
1.01972008e+00 -5.25456131e-01 9.00739074e-01 8.01024139e-01
-6.25938833e-01 7.04169035e-01 -4.62263495e-01 -9.10539687e-01
-9.21124697e-01 -1.09555888e+00 -2.56246388e-01 1.17957187e+00
5.35177171e-01 -5.94436564e-02 -7.02776313e-01 -6.02925897e-01
2.54904777e-01 -1.29477635e-01 -7.95528948e-01 -1.25633761e-01
-9.53714788e-01 -9.34680521e-01 5.72058737e-01 9.27215934e-01
1.10812926e+00 -6.68742657e-01 -2.74482612e-02 1.99772611e-01
-3.60265702e-01 -1.35608888e+00 -9.42568779e-01 -5.75449169e-01
-4.56189424e-01 -1.06445348e+00 -1.46197534e+00 -9.74891245e-01
7.80396104e-01 7.12742567e-01 8.34110320e-01 -1.03125788e-01
-6.42515242e-01 5.49669087e-01 -1.61425278e-01 -3.16815525e-01
3.71718198e-01 -9.04779211e-02 6.74083591e-01 2.49483645e-01
5.01473367e-01 -4.13016796e-01 -9.57424879e-01 8.03505003e-01
-1.76993430e-01 -1.46021023e-01 5.63645780e-01 8.04382861e-01
4.64654654e-01 -1.38897672e-01 4.49866772e-01 -2.88079500e-01
2.93733865e-01 -2.31588468e-01 -3.03540558e-01 5.17914236e-01
-1.11506637e-02 -2.66954839e-01 3.84341538e-01 -6.12115681e-01
-1.29832482e+00 1.47671178e-01 -1.99848525e-02 -4.23943728e-01
-4.99694705e-01 -3.52120101e-01 -6.70763791e-01 -3.95498663e-01
4.87970740e-01 4.41667974e-01 -3.04843366e-01 -7.46454656e-01
3.18111688e-01 5.37802100e-01 1.07973683e+00 -9.12865877e-01
1.13744700e+00 5.02480268e-01 -1.00742884e-01 -4.70641136e-01
-7.74815917e-01 -9.10670698e-01 -1.11401212e+00 -1.39231905e-01
1.10987437e+00 -1.60240793e+00 -9.01989460e-01 9.07859981e-01
-1.19057691e+00 2.99925864e-01 2.21646260e-02 3.88558000e-01
-4.96751405e-02 6.68035328e-01 -5.40573418e-01 -7.71942079e-01
-7.67322421e-01 -8.95136178e-01 1.22476482e+00 8.51020634e-01
1.98534712e-01 -5.34620106e-01 -5.80952466e-02 4.79330391e-01
2.88899153e-01 3.39063168e-01 8.74769241e-02 -4.41142797e-01
-3.98490757e-01 -4.71116334e-01 -5.71787775e-01 3.11706178e-02
2.33106643e-01 -5.25365174e-01 -1.16094148e+00 -7.60250032e-01
-3.79591972e-01 -9.64367762e-02 9.18669999e-01 4.63026673e-01
1.14719081e+00 -2.33732909e-01 -7.62821496e-01 1.02261317e+00
9.83756542e-01 -1.51072204e-01 5.44910669e-01 4.11114752e-01
1.19032860e+00 6.23657882e-01 3.30627561e-01 3.31215203e-01
7.84113348e-01 1.00268245e+00 -1.18083656e-02 -4.00589481e-02
-3.29596609e-01 -3.42065036e-01 2.78295517e-01 3.07006925e-01
-6.25175774e-01 1.56290963e-01 -4.27570671e-01 4.11462724e-01
-2.03637958e+00 -1.27692425e+00 1.88792101e-03 2.35381126e+00
3.48969340e-01 -3.11093777e-01 6.14583611e-01 -2.87724674e-01
1.31095099e+00 -2.01121226e-01 -5.83147407e-01 3.13793063e-01
-3.40412974e-01 -2.06160218e-01 5.63043416e-01 1.64453331e-02
-1.60139990e+00 8.05424035e-01 5.77193451e+00 6.06366396e-01
-6.90623462e-01 2.12416783e-01 3.18930060e-01 -5.49082868e-02
5.09807289e-01 -4.89226907e-01 -1.33674598e+00 7.61836469e-01
5.82637727e-01 9.39662755e-03 2.79749513e-01 9.08760667e-01
-1.06426515e-01 2.99916297e-01 -1.17048788e+00 1.71703935e+00
4.20414418e-01 -7.11312354e-01 4.03775871e-02 -1.20864131e-01
5.76950192e-01 -2.75859386e-01 -1.91184562e-02 5.98275363e-01
2.03930095e-01 -1.11919332e+00 6.73432231e-01 6.97472215e-01
7.37660229e-01 -7.29446709e-01 1.01084995e+00 1.61566198e-01
-1.90930331e+00 -2.85616726e-01 -7.19606578e-01 1.11805692e-01
1.98291540e-01 1.90185551e-02 -1.93658099e-01 8.74647439e-01
1.30520570e+00 7.39406824e-01 -7.93986797e-01 1.54591107e+00
-2.34643836e-02 -1.89269021e-01 -2.83206850e-01 4.48875964e-01
-4.24972415e-01 4.07502390e-02 5.04560113e-01 1.48639417e+00
6.99671805e-02 1.51047602e-01 5.38819194e-01 7.56137252e-01
-1.40928984e-01 -1.59789011e-01 -2.38647714e-01 8.68934512e-01
2.99005359e-01 1.24460018e+00 -1.61708206e-01 -3.82909805e-01
-4.39726561e-01 1.34324956e+00 3.30556393e-01 5.09103060e-01
-8.52421999e-01 1.74457917e-03 1.06749475e+00 -8.64034221e-02
1.55777454e-01 -1.88504606e-01 1.99239388e-01 -1.57305717e+00
2.78486282e-01 -7.34662294e-01 7.68765271e-01 -3.89129817e-01
-1.85198033e+00 3.78962457e-01 1.45718470e-01 -1.13589549e+00
1.38581276e-01 -6.61890328e-01 -6.63582444e-01 1.30913532e+00
-1.54722631e+00 -1.83071828e+00 -8.49423409e-01 1.21782804e+00
6.09803736e-01 -2.83700675e-01 6.31137729e-01 7.58068740e-01
-1.09699154e+00 1.25283825e+00 -6.74652606e-02 7.28839219e-01
1.11355007e+00 -1.13564718e+00 6.27966642e-01 9.80587304e-01
-3.72003287e-01 1.04004347e+00 3.49951357e-01 -7.55542219e-01
-1.45711458e+00 -1.32950914e+00 4.11463946e-01 -6.90263271e-01
-9.56917107e-02 -4.77534682e-01 -7.44290233e-01 8.01815510e-01
-1.23375475e-01 4.85744029e-01 6.53160334e-01 2.75441349e-01
-6.32735372e-01 -3.01058739e-01 -1.37639296e+00 3.77856493e-01
1.69884658e+00 -3.17740083e-01 -7.60415852e-01 1.93151772e-01
5.01641572e-01 -5.76917708e-01 -7.72201359e-01 4.40051883e-01
8.81658912e-01 -6.71688557e-01 1.71519506e+00 -6.71099722e-01
-3.74443203e-01 -6.36899710e-01 -2.70861126e-02 -1.06052935e+00
-8.21120083e-01 -3.19159687e-01 -4.98597324e-02 1.58321404e+00
-3.34031612e-01 -7.17533886e-01 6.11857533e-01 1.08032644e+00
4.71378639e-02 -6.58765510e-02 -1.07216406e+00 -9.46421921e-01
-8.88055284e-03 2.31806248e-01 9.51177657e-01 6.85545444e-01
-4.16439563e-01 1.29734114e-01 -7.42009282e-01 5.58763981e-01
1.12567651e+00 2.50133388e-02 1.23888206e+00 -1.49507475e+00
-1.59709305e-02 -2.43470564e-01 -9.06788349e-01 -1.07713532e+00
2.82517254e-01 -6.07268035e-01 -1.14759728e-01 -1.36548698e+00
7.47913480e-01 -3.22522163e-01 -4.96088207e-01 4.18073505e-01
-5.40567994e-01 6.08220994e-02 3.14515620e-01 2.63940394e-01
-7.18350351e-01 7.39378333e-01 1.07348669e+00 -6.03725076e-01
1.44487698e-04 4.70794477e-02 -7.30125189e-01 6.77818358e-01
2.64645249e-01 -2.38054097e-02 -1.22101195e-01 -6.32250190e-01
-4.92742479e-01 -4.00757819e-01 9.66297925e-01 -1.39975190e+00
5.57448745e-01 5.08593731e-02 1.40845835e+00 -6.48253381e-01
5.79344094e-01 -7.85959780e-01 2.92161137e-01 3.01252037e-01
1.93798050e-01 1.43327132e-01 3.89024377e-01 7.03577638e-01
-1.39353645e-03 2.39652127e-01 8.57654274e-01 -3.34074199e-01
-9.79821682e-01 8.56086791e-01 3.63499671e-01 -2.65271068e-01
9.08685744e-01 -4.42436963e-01 -3.52658212e-01 4.11321372e-02
-5.96992016e-01 4.66483682e-01 3.83221269e-01 8.26917410e-01
5.59063315e-01 -1.52522576e+00 -9.46864247e-01 1.74326062e-01
2.89242834e-01 5.06158657e-02 9.38770950e-01 5.26776195e-01
6.58614039e-02 4.10337031e-01 -4.97814685e-01 -6.63525403e-01
-1.59885144e+00 7.35858619e-01 6.90544546e-01 6.98532462e-02
-8.70989442e-01 1.08901858e+00 6.53834343e-01 -6.09252036e-01
2.69507825e-01 1.35989875e-01 -5.22831857e-01 -1.05559997e-01
1.01932836e+00 4.74383622e-01 -2.94164091e-01 -1.24052620e+00
-7.18625605e-01 1.12507761e+00 -3.12858224e-01 5.06832540e-01
1.01002240e+00 -4.66416836e-01 9.77849513e-02 -2.89189309e-01
9.71412599e-01 -3.88677597e-01 -1.19869876e+00 -4.04973656e-01
-4.20426846e-01 -6.82776988e-01 -5.16234457e-01 -7.09551871e-01
-9.62951362e-01 6.65396988e-01 1.20476139e+00 -6.67409778e-01
7.14980364e-01 -1.94575310e-01 9.40839946e-01 4.39884871e-01
7.94818223e-01 -1.05343711e+00 -2.80412555e-01 2.96143889e-01
9.37527537e-01 -1.52380741e+00 -5.28569845e-03 -3.81185830e-01
-3.43984663e-01 8.29571009e-01 1.02591085e+00 4.44160402e-03
4.17370975e-01 -2.53361464e-01 -3.34738716e-02 2.18233392e-01
1.21647574e-01 -4.13989276e-01 5.09496987e-01 1.06356466e+00
1.44593030e-01 2.24551260e-01 9.47475657e-02 1.01315415e+00
-2.73576915e-01 -5.94088994e-02 -6.00327402e-02 6.44540370e-01
-1.23951688e-01 -1.09701216e+00 -9.40761566e-01 2.10708827e-01
-2.65669554e-01 2.98483223e-01 -2.13413119e-01 6.82522416e-01
5.53169668e-01 9.05173242e-01 -2.84116209e-01 -3.99222165e-01
5.83012044e-01 5.69420569e-02 6.50457740e-01 -3.54720145e-01
-5.99268615e-01 -3.42124671e-01 -3.56383771e-02 -4.68865275e-01
-5.78829050e-01 -7.80758321e-01 -7.59581745e-01 -5.17215669e-01
-1.67373404e-01 -2.21590221e-01 4.96502161e-01 1.03657603e+00
3.83016348e-01 3.87489021e-01 2.56579310e-01 -1.21892154e+00
-5.50376832e-01 -1.05389857e+00 -4.32592720e-01 7.59864330e-01
3.31715494e-01 -1.05448496e+00 1.31716713e-01 1.14367731e-01]
|
[14.66459846496582, 0.8996713757514954]
|
2aab9a82-518e-44fc-9b90-70ce1919e1c7
|
improving-speech-related-facial-action-unit
|
1706.10197
| null |
http://arxiv.org/abs/1706.10197v1
|
http://arxiv.org/pdf/1706.10197v1.pdf
|
Improving Speech Related Facial Action Unit Recognition by Audiovisual Information Fusion
|
It is challenging to recognize facial action unit (AU) from spontaneous
facial displays, especially when they are accompanied by speech. The major
reason is that the information is extracted from a single source, i.e., the
visual channel, in the current practice. However, facial activity is highly
correlated with voice in natural human communications.
Instead of solely improving visual observations, this paper presents a novel
audiovisual fusion framework, which makes the best use of visual and acoustic
cues in recognizing speech-related facial AUs. In particular, a dynamic
Bayesian network (DBN) is employed to explicitly model the semantic and dynamic
physiological relationships between AUs and phonemes as well as measurement
uncertainty. A pilot audiovisual AU-coded database has been collected to
evaluate the proposed framework, which consists of a "clean" subset containing
frontal faces under well controlled circumstances and a challenging subset with
large head movements and occlusions. Experiments on this database have
demonstrated that the proposed framework yields significant improvement in
recognizing speech-related AUs compared to the state-of-the-art visual-based
methods especially for those AUs whose visual observations are impaired during
speech, and more importantly also outperforms feature-level fusion methods by
explicitly modeling and exploiting physiological relationships between AUs and
phonemes.
|
['Yan Tong', 'Zibo Meng', 'Ping Liu', 'Shizhong Han']
|
2017-06-29
| null | null | null | null |
['facial-action-unit-detection']
|
['computer-vision']
|
[ 1.31542355e-01 -1.07816290e-02 2.27497548e-01 -3.01027238e-01
-5.76221645e-01 -1.10206150e-01 4.34564948e-01 -3.48000646e-01
-2.43358254e-01 8.37886155e-01 3.12783927e-01 4.32918280e-01
2.22279951e-01 4.84566838e-02 -5.13468802e-01 -1.09615123e+00
2.10259557e-01 -2.70432651e-01 -2.01044083e-01 1.51643142e-01
-8.19767639e-02 5.32941103e-01 -2.24207282e+00 8.84518474e-02
7.42059469e-01 1.41495395e+00 2.47405276e-01 3.63249511e-01
1.21483438e-01 3.59216005e-01 -5.86031497e-01 -2.83689618e-01
5.32459579e-02 -4.79619116e-01 -1.06009170e-02 4.04094934e-01
3.91553074e-01 -2.80258685e-01 -2.75538951e-01 1.12988341e+00
8.65863681e-01 1.79044873e-01 6.89182997e-01 -1.36999917e+00
-2.10252181e-01 -1.44727528e-01 -6.73914671e-01 2.97260851e-01
4.85745400e-01 9.21450555e-02 5.77675462e-01 -1.10367632e+00
2.68624902e-01 1.47418809e+00 1.94097683e-01 6.64350867e-01
-1.10106218e+00 -1.01453364e+00 2.25333020e-01 3.98858398e-01
-1.71162832e+00 -1.26063657e+00 9.11238968e-01 -4.82049435e-01
8.03048551e-01 1.26916111e-01 6.01290524e-01 1.26535439e+00
1.98131442e-01 5.73381424e-01 1.27627015e+00 -2.01556340e-01
1.83618203e-01 3.92331839e-01 -5.19502312e-02 4.29619402e-01
-7.10964501e-02 3.17420304e-01 -8.78531814e-01 -3.08835149e-01
4.40043241e-01 -2.54609764e-01 -6.06637895e-01 -5.39197139e-02
-8.08144331e-01 3.67256910e-01 -8.82441923e-02 3.34913850e-01
-8.43246996e-01 -6.95513263e-02 2.06331909e-01 -1.51323378e-01
3.29747051e-01 -3.59157592e-01 -8.94674435e-02 -2.13182464e-01
-9.28964794e-01 -1.02067053e-01 5.96652746e-01 5.52953303e-01
4.75212097e-01 5.62928259e-01 -9.13124830e-02 1.04231966e+00
6.98236406e-01 8.52300584e-01 5.88138938e-01 -8.45044315e-01
6.30321950e-02 6.33657575e-02 5.13618141e-02 -9.73641336e-01
-3.51773679e-01 -2.55097330e-01 -6.20605826e-01 4.61410582e-01
1.09229967e-01 -2.02929318e-01 -8.46462727e-01 2.18858790e+00
4.93039519e-01 5.28418899e-01 2.44587123e-01 9.91921961e-01
1.14100897e+00 5.10730505e-01 1.26400054e-01 -1.16520607e+00
1.41788924e+00 -2.16205314e-01 -1.46171474e+00 -1.46186158e-01
-7.66511858e-02 -7.61468470e-01 4.78541017e-01 6.21680856e-01
-1.03064752e+00 -7.27799892e-01 -9.58078146e-01 4.18003321e-01
-1.79365762e-02 3.32799911e-01 3.81046683e-01 7.63351619e-01
-9.85250175e-01 3.42432633e-02 -7.41389215e-01 -4.54832494e-01
2.29263276e-01 3.85671705e-01 -6.28754497e-01 1.81177974e-01
-1.11261821e+00 7.63945341e-01 -4.92865667e-02 6.27126515e-01
-9.97281611e-01 -1.53572693e-01 -9.15667117e-01 -7.48801604e-02
2.83346891e-01 -2.46471927e-01 1.09292185e+00 -1.23443258e+00
-1.62399995e+00 4.98704672e-01 -7.82949865e-01 3.20965634e-03
6.12921715e-02 -4.99542393e-02 -7.46764541e-01 3.88587445e-01
-3.56846541e-01 5.75373471e-01 1.35369718e+00 -1.33320272e+00
-2.69684643e-01 -7.28597403e-01 -4.35042620e-01 4.73769993e-01
-2.14964524e-01 2.92868078e-01 -2.31422484e-01 -3.12831163e-01
3.06963492e-02 -7.58330464e-01 4.71456110e-01 1.06015995e-01
-1.61292031e-02 -2.80476183e-01 7.96634316e-01 -8.28123271e-01
1.06528711e+00 -2.40065336e+00 3.04852203e-02 1.51969388e-01
1.72609732e-01 4.11356896e-01 -4.26576026e-02 1.18355468e-01
-2.11916566e-01 -3.79683167e-01 9.16490257e-02 -5.90451837e-01
-1.78054735e-01 2.39883363e-01 2.78251432e-02 7.90477157e-01
9.01032612e-03 4.58914757e-01 -3.61940145e-01 -6.47628069e-01
2.69160092e-01 8.24468970e-01 -2.19837710e-01 3.35186213e-01
2.64569342e-01 6.16584897e-01 -2.50821710e-01 9.74033833e-01
8.00481260e-01 2.89058238e-01 -2.78468784e-02 -5.92921197e-01
5.04614487e-02 -2.49772236e-01 -1.34013331e+00 1.60612428e+00
-2.00619712e-01 5.81096411e-01 6.83233559e-01 -8.41311693e-01
1.04063737e+00 9.42524791e-01 5.38130701e-01 -7.28709161e-01
4.60826367e-01 5.66407517e-02 2.17312604e-01 -7.80082822e-01
-7.74286836e-02 -2.21745133e-01 4.25390452e-01 -3.19020972e-02
4.18507606e-01 2.15062171e-01 -1.81962460e-01 -1.32019252e-01
7.05429912e-01 2.25130264e-02 5.43597698e-01 -1.62209421e-02
7.65457332e-01 -9.14284348e-01 9.77135539e-01 1.54391378e-01
-6.49916589e-01 3.35478336e-01 3.72889549e-01 3.32269937e-01
-3.23464334e-01 -1.20798755e+00 -3.24130416e-01 6.36496723e-01
4.05068882e-02 -1.63636386e-01 -8.53053749e-01 -2.52665401e-01
-4.52147387e-02 5.19422472e-01 -5.37755907e-01 -1.91916987e-01
1.09962374e-01 -6.40271723e-01 4.54529315e-01 4.33298975e-01
3.98993462e-01 -1.00047708e+00 -3.78611207e-01 6.46006390e-02
-2.68173784e-01 -1.32898784e+00 -2.80430257e-01 -1.25438139e-01
-4.01385397e-01 -1.09560883e+00 -5.76006472e-01 -4.04302895e-01
4.83946413e-01 2.62327969e-01 3.88223946e-01 -3.86224389e-01
-4.67916071e-01 7.44879067e-01 -1.70374930e-01 -7.67669022e-01
-1.03491910e-01 -9.13446248e-01 4.82904464e-01 9.19368804e-01
5.09349644e-01 -7.51444995e-01 -4.76930618e-01 3.57994199e-01
-4.98304397e-01 -3.08672667e-01 5.62843621e-01 7.35061049e-01
4.13532674e-01 3.02820429e-02 8.80799413e-01 -9.50259939e-02
6.11689508e-01 -4.13368344e-01 -3.49348992e-01 1.14842787e-01
-3.14820826e-01 -4.42901522e-01 1.75333574e-01 -5.73317111e-01
-1.62387002e+00 1.50241047e-01 -2.19889451e-02 -7.94750094e-01
-6.14877760e-01 2.34741271e-01 -6.33135915e-01 -7.00438395e-02
4.43142802e-01 3.09235811e-01 4.78397787e-01 -3.26550752e-01
-6.95887655e-02 1.32055223e+00 7.94068515e-01 -3.11559677e-01
3.38618428e-01 4.38665003e-01 5.90419546e-02 -1.45568550e+00
-1.76751226e-01 -4.85731930e-01 -4.84977067e-01 -6.82548046e-01
9.18055654e-01 -1.24361670e+00 -1.08478308e+00 7.34446168e-01
-1.17519283e+00 2.83436388e-01 3.27118695e-01 1.08163619e+00
-5.83495796e-01 3.87761980e-01 -3.01034540e-01 -1.56795037e+00
-2.70232230e-01 -1.22823060e+00 1.03938651e+00 3.89731258e-01
-1.86170489e-01 -4.98251945e-01 -1.43574283e-01 4.95156556e-01
2.23277390e-01 2.20100850e-01 4.74816352e-01 -4.52072024e-01
-2.44856089e-01 -1.72464252e-01 -5.00135422e-02 6.79788589e-01
4.15149093e-01 1.33483678e-01 -1.72105277e+00 -1.27893433e-01
2.97346979e-01 -4.06168789e-01 3.27626348e-01 6.56480134e-01
6.53006136e-01 6.69446867e-03 -1.91168472e-01 2.61821896e-01
1.03482378e+00 7.51021922e-01 5.33137679e-01 -6.81651950e-01
3.51433754e-01 8.55267704e-01 6.22604012e-01 7.26687074e-01
1.44412845e-01 8.24859560e-01 7.00332582e-01 -8.23169760e-03
-1.04199171e-01 7.77261704e-02 4.99007165e-01 7.03496039e-01
-1.34687081e-01 -2.98083752e-01 -3.50126892e-01 2.82569051e-01
-1.66890943e+00 -7.98402369e-01 1.46843374e-01 2.29543829e+00
5.44217527e-01 -3.59534532e-01 -3.91745195e-02 2.87258059e-01
7.91102111e-01 1.48621341e-02 -6.05160117e-01 -2.96462327e-01
-1.30191535e-01 1.76892072e-01 -1.38709068e-01 3.99346948e-01
-7.47958720e-01 6.01287603e-01 5.93037701e+00 8.23522508e-01
-1.32527304e+00 1.30321309e-01 1.54837564e-01 -3.66696030e-01
1.25703871e-01 -4.74659920e-01 -6.76978350e-01 5.10533810e-01
1.02430141e+00 2.22785566e-02 5.09294152e-01 5.39089680e-01
7.82934606e-01 -4.41103548e-01 -9.74629819e-01 1.59579885e+00
5.36049604e-01 -4.44298595e-01 -3.66454154e-01 2.88631499e-01
9.28183869e-02 -3.95065248e-01 1.01640269e-01 -8.12130645e-02
-4.50102210e-01 -1.05542243e+00 5.70122123e-01 8.95423949e-01
7.66332150e-01 -6.39842272e-01 6.92600250e-01 3.12607616e-01
-1.17462659e+00 -5.46090342e-02 -1.90308943e-01 1.10341400e-01
8.91215727e-02 2.96154708e-01 -5.46974540e-01 3.27390492e-01
7.27306664e-01 6.00830555e-01 -1.41944930e-01 8.24708164e-01
-1.28040463e-01 4.95807946e-01 -4.36333060e-01 1.31773248e-01
-3.69243741e-01 -8.95287767e-02 6.16956532e-01 8.87156069e-01
5.61610341e-01 3.53696018e-01 -2.11795077e-01 7.34962881e-01
2.92586535e-01 3.04993957e-01 -9.15336132e-01 -1.15862370e-01
5.61360776e-01 1.43382490e+00 -1.81208923e-01 -2.37777345e-02
-6.56204820e-01 7.11533964e-01 -5.66775985e-02 6.51218295e-01
-7.63685822e-01 7.10150925e-03 1.02928460e+00 -2.21755981e-01
1.94394425e-01 -9.02484134e-02 2.73769170e-01 -9.14410651e-01
1.71545237e-01 -8.49667847e-01 7.87679572e-03 -1.18615305e+00
-9.00644600e-01 7.10945666e-01 7.95175135e-02 -1.26778245e+00
-4.34901655e-01 -5.18282592e-01 -4.33149248e-01 1.14870405e+00
-1.32532918e+00 -9.74332869e-01 -5.06658673e-01 1.02167916e+00
3.89400423e-01 -2.11177200e-01 8.36311340e-01 3.86449516e-01
-7.69253612e-01 7.77486026e-01 -1.42011374e-01 -3.44385952e-01
8.34147274e-01 -6.04601979e-01 -7.09068120e-01 7.37473905e-01
1.80341769e-02 4.32314634e-01 7.75695622e-01 -5.21142304e-01
-1.36393642e+00 -5.86945415e-01 5.59588552e-01 1.54266506e-01
3.36207181e-01 -4.93874729e-01 -8.79279971e-01 2.86331475e-01
2.74600565e-01 1.92102730e-01 8.69161963e-01 -3.84758770e-01
-6.77893534e-02 -4.78061855e-01 -1.27779663e+00 3.23922008e-01
8.39924514e-01 -7.09776640e-01 -5.94475567e-01 -4.03019525e-02
2.02926695e-01 -1.44179493e-01 -7.76760995e-01 5.83332241e-01
8.07650268e-01 -1.03699565e+00 5.13799667e-01 -1.80245519e-01
-4.38993961e-01 -3.86068106e-01 -4.23150003e-01 -1.41326904e+00
1.67838708e-01 -6.89243674e-01 -2.14009985e-01 1.68086016e+00
-2.47199852e-02 -7.29068637e-01 5.16742349e-01 5.35044193e-01
1.33793394e-03 -3.65986615e-01 -1.29003310e+00 -6.19027615e-01
-7.53927290e-01 -4.96310025e-01 1.65699780e-01 4.51654196e-01
8.59237015e-02 2.35375270e-01 -5.59656382e-01 2.82488883e-01
6.23475313e-01 -4.42202121e-01 6.29206538e-01 -1.29199731e+00
-5.49907312e-02 1.51548267e-03 -7.35381663e-01 -5.52468956e-01
5.19993305e-01 -4.42469776e-01 3.62648427e-01 -1.06175530e+00
-8.20607692e-03 2.89811373e-01 -2.89227396e-01 1.66055501e-01
5.72454296e-02 6.09376058e-02 3.04793153e-04 1.95829757e-02
2.34163981e-02 8.94202590e-01 1.06680739e+00 1.65827662e-01
-2.23294958e-01 1.86318476e-02 -5.24042428e-01 8.49323034e-01
3.52861494e-01 -1.71868697e-01 -6.87110722e-01 1.49604946e-01
-4.79996145e-01 5.80102026e-01 4.29277360e-01 -1.03544307e+00
1.83092490e-01 3.31385992e-02 4.40151215e-01 -4.81491327e-01
1.15334654e+00 -1.01227725e+00 2.29802534e-01 4.22135256e-02
-2.41778437e-02 -2.12258622e-01 3.41422707e-01 8.20135891e-01
-4.97622877e-01 8.70769620e-02 8.56769562e-01 5.23376428e-02
-5.18257260e-01 2.83653557e-01 -5.60273349e-01 -5.02240002e-01
1.07598293e+00 -4.61302608e-01 -2.38409787e-01 -7.49381959e-01
-1.08939135e+00 -9.57771391e-02 -1.19855545e-01 3.49326462e-01
7.96118975e-01 -1.30819416e+00 -5.15238225e-01 6.08713210e-01
7.95113072e-02 -5.14547110e-01 6.51035666e-01 1.33661795e+00
3.15053046e-01 4.58859056e-01 -3.81393492e-01 -7.09264517e-01
-1.80701220e+00 4.61525947e-01 4.07475859e-01 7.25070477e-01
-1.02456838e-01 6.83967590e-01 6.48688912e-01 3.74878854e-01
6.63656950e-01 -9.21293274e-02 -5.96600056e-01 4.38827515e-01
6.00737929e-01 3.37628156e-01 9.35275033e-02 -1.17114127e+00
-4.49028879e-01 6.02709591e-01 2.87003845e-01 -3.01160574e-01
7.46967256e-01 -6.01919353e-01 1.05830222e-01 6.02293909e-01
1.13380289e+00 -7.96088874e-02 -1.15828085e+00 -4.32993293e-01
-5.62319338e-01 -5.07490814e-01 2.46997148e-01 -6.50094092e-01
-1.12498057e+00 1.19611537e+00 1.06906569e+00 -2.91009158e-01
1.39485919e+00 -1.13505363e-01 3.43464762e-01 1.78590775e-01
4.27607805e-01 -1.04138827e+00 6.61677867e-02 -2.06199005e-01
1.18748379e+00 -1.14277554e+00 -3.51609528e-01 -5.39235175e-01
-7.60225177e-01 1.00865495e+00 6.66548371e-01 2.46538907e-01
9.77321208e-01 2.50396222e-01 1.59672827e-01 1.65187810e-02
-8.88177931e-01 -4.92190331e-01 4.04001623e-01 8.03292811e-01
2.92251408e-01 -9.45003629e-02 -1.54537871e-01 6.96825683e-01
7.29767606e-02 9.84546989e-02 2.63462931e-01 5.98473847e-01
-4.72048938e-01 -5.21969378e-01 -6.67363048e-01 2.71102726e-01
-6.28937423e-01 7.39063993e-02 -2.49043331e-01 7.81626999e-01
2.26753131e-01 1.41978061e+00 -1.13408063e-02 -4.49014783e-01
3.35375458e-01 5.08797467e-01 5.48880935e-01 -4.11387861e-01
-1.43973261e-01 7.73503661e-01 1.39319226e-01 -8.75908434e-01
-7.18110263e-01 -9.88500059e-01 -1.14777303e+00 1.00788936e-01
-3.32474470e-01 9.70401242e-03 9.03803766e-01 1.06258559e+00
2.63883859e-01 4.21081781e-01 6.56716406e-01 -9.73898470e-01
-3.30622345e-01 -1.31198132e+00 -1.18727362e+00 2.08785906e-01
4.78614062e-01 -1.25738323e+00 -7.74574518e-01 4.69815582e-02]
|
[14.349449157714844, 4.975132942199707]
|
b237c53e-d20c-4394-a0f7-2fe43ff1402b
|
semi-supervised-anomaly-detection-using
|
2001.03674
| null |
https://arxiv.org/abs/2001.03674v1
|
https://arxiv.org/pdf/2001.03674v1.pdf
|
Semi-supervised Anomaly Detection using AutoEncoders
|
Anomaly detection refers to the task of finding unusual instances that stand out from the normal data. In several applications, these outliers or anomalous instances are of greater interest compared to the normal ones. Specifically in the case of industrial optical inspection and infrastructure asset management, finding these defects (anomalous regions) is of extreme importance. Traditionally and even today this process has been carried out manually. Humans rely on the saliency of the defects in comparison to the normal texture to detect the defects. However, manual inspection is slow, tedious, subjective and susceptible to human biases. Therefore, the automation of defect detection is desirable. But for defect detection lack of availability of a large number of anomalous instances and labelled data is a problem. In this paper, we present a convolutional auto-encoder architecture for anomaly detection that is trained only on the defect-free (normal) instances. For the test images, residual masks that are obtained by subtracting the original image from the auto-encoder output are thresholded to obtain the defect segmentation masks. The approach was tested on two data-sets and achieved an impressive average F1 score of 0.885. The network learnt to detect the actual shape of the defects even though no defected images were used during the training.
|
['Manpreet Singh Minhas', 'John Zelek']
|
2020-01-06
|
semi-supervised-anomaly-detection-using-1
|
https://openjournals.uwaterloo.ca/index.php/vsl/article/view/1654
|
https://openjournals.uwaterloo.ca/index.php/vsl/article/view/1654/2021
|
journal-of-computational-vision-and-imaging
|
['supervised-anomaly-detection', 'semi-supervised-anomaly-detection']
|
['computer-vision', 'computer-vision']
|
[ 6.05640769e-01 2.72527754e-01 6.97934628e-01 -3.60804021e-01
-4.33255643e-01 -1.59761146e-01 3.10784757e-01 6.15077019e-01
-1.47584200e-01 3.25702637e-01 -5.51101685e-01 -1.58235565e-01
-1.94315817e-02 -6.45068228e-01 -6.67962909e-01 -9.78937268e-01
-3.29306036e-01 4.08541024e-01 4.43545491e-01 -1.86745122e-01
4.89755720e-01 6.12803698e-01 -1.84762335e+00 4.16016608e-01
9.42280352e-01 1.23968792e+00 2.72188663e-01 6.21740341e-01
-9.92706791e-02 4.23646182e-01 -9.18568373e-01 -7.92257302e-03
4.89815444e-01 -6.18961930e-01 -5.94324410e-01 8.22732151e-01
5.19696653e-01 -3.23758841e-01 2.11405724e-01 1.22571206e+00
1.10268310e-01 5.15490882e-02 5.52189946e-01 -8.59222054e-01
-2.86883861e-01 -2.74999768e-01 -6.32983387e-01 4.63720262e-01
2.40910456e-01 2.70575494e-01 9.43877578e-01 -9.24905062e-01
4.77235734e-01 7.63806880e-01 2.82474130e-01 3.38944644e-01
-1.05614483e+00 -7.38503737e-03 6.02945266e-03 8.55759680e-02
-1.01254094e+00 -2.92154551e-01 8.75274837e-01 -6.06151462e-01
8.85005414e-01 1.82168886e-01 5.49150765e-01 7.61048317e-01
3.89299959e-01 6.50631726e-01 7.04577267e-01 -5.89184940e-01
3.72410208e-01 1.98262513e-01 -1.30382299e-01 5.25312066e-01
5.28334975e-01 -2.71429103e-02 -1.38323635e-01 7.63688385e-02
6.03483140e-01 2.46950835e-01 -3.26588809e-01 -3.89271170e-01
-7.74813294e-01 6.59536421e-01 4.26953375e-01 6.54898405e-01
-8.70525897e-01 -3.94928962e-01 6.09160542e-01 6.08717382e-01
5.56433618e-01 6.94317222e-01 -4.07571971e-01 8.15694407e-02
-9.83529806e-01 -1.55675709e-02 4.97460395e-01 3.78164381e-01
7.04041421e-01 2.50059724e-01 1.88127264e-01 7.36363173e-01
-5.63168786e-02 6.82121841e-03 4.21246707e-01 -3.69153678e-01
1.67020634e-01 1.23996460e+00 -4.25958782e-02 -9.97402906e-01
-2.11943641e-01 -4.83879447e-01 -8.74943018e-01 6.40457749e-01
5.63401461e-01 5.00697233e-02 -1.33082235e+00 8.25013041e-01
2.48252138e-01 -1.36124238e-01 6.98372796e-02 8.89404655e-01
2.78921545e-01 4.27822530e-01 -3.01772088e-01 6.52590469e-02
9.35165644e-01 -4.55202460e-01 -4.59904611e-01 -4.20845777e-01
6.27093256e-01 -8.17334235e-01 1.00722134e+00 7.30813742e-01
-8.95971179e-01 -4.77682501e-01 -1.04666805e+00 5.36973000e-01
-4.12888348e-01 3.65036756e-01 2.35639244e-01 2.94869065e-01
-7.60117173e-01 7.50442088e-01 -8.20530474e-01 -5.27997077e-01
4.78360295e-01 3.38716090e-01 -6.94716275e-01 -2.68208921e-01
-6.67710841e-01 8.19343150e-01 5.45211613e-01 4.24966246e-01
-7.72513986e-01 -2.44146883e-01 -8.41688752e-01 -5.56685328e-02
5.30407965e-01 1.39398932e-01 8.40591788e-01 -1.46273851e+00
-7.84605026e-01 9.19691682e-01 1.35081843e-01 -5.43066800e-01
6.18168592e-01 -1.66247010e-01 -4.94702399e-01 2.02811524e-01
1.83616772e-01 1.06218137e-01 1.28235638e+00 -1.29542100e+00
-9.90941048e-01 -2.76054740e-01 -1.38077050e-01 -2.60823905e-01
9.76533145e-02 -2.37877071e-01 -4.18310240e-02 -4.41770673e-01
6.70387030e-01 -6.17650449e-01 -1.74421459e-01 -1.90758958e-01
-4.77094978e-01 -6.01672977e-02 1.26024842e+00 -8.91381621e-01
7.09909260e-01 -2.60712218e+00 -4.10496920e-01 5.87871134e-01
1.23041958e-01 2.82185435e-01 1.70073137e-01 2.41359413e-01
-4.92614448e-01 -1.88409522e-01 -6.55622661e-01 1.83920607e-01
-4.09785450e-01 2.12983951e-01 -4.57904041e-02 6.86441541e-01
8.82256985e-01 2.43665501e-01 -7.19965696e-01 -2.83191353e-01
3.73123825e-01 2.20384188e-02 -4.12379771e-01 4.80082422e-01
-1.29340902e-01 7.49081194e-01 -1.44785821e-01 9.38503742e-01
5.16660035e-01 1.00563280e-01 -2.31224671e-01 -1.80178564e-02
9.53982472e-02 -4.38082032e-02 -1.18708372e+00 1.18534613e+00
-1.92498535e-01 6.05140746e-01 3.29595022e-02 -1.35611606e+00
1.21121120e+00 4.56844836e-01 6.21802866e-01 -6.80475235e-01
-7.53650218e-02 7.46757209e-01 4.61473882e-01 -8.77043307e-01
3.83868158e-01 -1.04751177e-01 1.62753150e-01 2.34383836e-01
8.71702060e-02 -1.45147175e-01 4.22185600e-01 -8.99480879e-02
1.27093065e+00 -1.09345749e-01 3.29378963e-01 -8.35292041e-02
6.48809195e-01 -8.25437903e-03 6.28112614e-01 4.08450872e-01
-9.28666368e-02 8.10578048e-01 8.28294456e-01 -8.64496827e-01
-1.13849509e+00 -8.94030750e-01 -9.86211300e-02 4.24953789e-01
-2.08990723e-02 1.64655134e-01 -6.20154023e-01 -1.02593446e+00
-2.70379484e-02 7.49745607e-01 -6.46004498e-01 -4.15972948e-01
-5.35233498e-01 -6.29277289e-01 2.68448535e-02 2.97304332e-01
4.71638054e-01 -1.35049045e+00 -1.07855296e+00 2.59506524e-01
1.83784634e-01 -7.97273695e-01 5.04514091e-02 4.23673898e-01
-8.74251962e-01 -1.40208209e+00 -4.01128531e-01 -6.99013889e-01
1.26548386e+00 -8.61219764e-02 1.04360747e+00 5.31682968e-01
-7.77520120e-01 1.22563280e-01 -5.24423480e-01 -6.34363890e-01
-6.46801353e-01 -3.51238966e-01 -1.39643744e-01 3.02347094e-01
5.07565260e-01 -3.11082095e-01 -4.36417222e-01 2.31160551e-01
-1.15325904e+00 -5.87827742e-01 8.90201509e-01 1.01440990e+00
5.85180461e-01 7.09995925e-01 5.84693253e-01 -9.36260819e-01
4.22837526e-01 -3.59455407e-01 -5.24295866e-01 -7.06245825e-02
-4.48804140e-01 -1.38215408e-01 5.94184220e-01 -1.40970230e-01
-9.27417457e-01 1.89846024e-01 -3.01561415e-01 -4.21085328e-01
-8.63878310e-01 1.93234831e-01 -5.83890714e-02 1.16367124e-01
6.09291136e-01 1.34886965e-01 9.86988097e-02 -4.35343057e-01
-4.84092236e-01 6.05461180e-01 5.47014236e-01 -1.20490104e-01
6.15096152e-01 3.72843146e-01 -3.75689045e-02 -1.15443599e+00
-6.79067492e-01 -7.03378141e-01 -9.07621443e-01 -3.94985974e-01
6.05722427e-01 -3.12604100e-01 7.25133792e-02 4.12182063e-01
-1.04884052e+00 1.64673030e-01 -8.37124884e-01 2.62619823e-01
-3.47235233e-01 3.06161493e-01 -1.62215143e-01 -1.04878151e+00
-1.41146272e-01 -1.18375957e+00 1.02401757e+00 7.37178847e-02
-2.45737731e-01 -7.66849220e-01 -4.46559250e-01 2.65120994e-02
3.14155281e-01 6.33500993e-01 1.11297286e+00 -1.19786692e+00
-4.22305465e-01 -9.65014875e-01 9.26610380e-02 8.02085340e-01
6.74168289e-01 2.37157553e-01 -1.03952980e+00 -3.05122584e-01
2.47319311e-01 -2.78121866e-02 7.72658348e-01 3.28377277e-01
1.04726982e+00 -9.33550671e-03 -1.64054379e-01 -5.12015186e-02
1.34105468e+00 4.79741365e-01 5.78489721e-01 2.90963531e-01
5.10302603e-01 9.29097891e-01 9.08654809e-01 3.86139035e-01
-4.52669978e-01 3.42592120e-01 9.03357327e-01 -4.14342582e-01
3.15953970e-01 2.77027994e-01 2.62961000e-01 2.62395442e-01
-1.14265397e-01 -8.65356252e-03 -9.25573707e-01 1.00419927e+00
-1.39467537e+00 -7.41958737e-01 -2.73079038e-01 2.41945553e+00
1.96004733e-01 6.09907150e-01 1.32017545e-02 8.11458230e-01
8.20459664e-01 -1.65540069e-01 -6.07644856e-01 -6.88460946e-01
-5.59005253e-02 7.25500062e-02 1.31624028e-01 -1.58202983e-02
-1.17032456e+00 4.80230272e-01 5.20960426e+00 4.22636986e-01
-1.03233027e+00 -2.64740735e-01 6.63394034e-01 3.55255641e-02
2.92386293e-01 -2.17355475e-01 -1.96153805e-01 4.50382233e-01
6.75111830e-01 4.36543435e-01 -1.62266403e-01 9.31657732e-01
8.39000121e-02 -5.97001076e-01 -1.17241168e+00 7.07767308e-01
9.16736014e-03 -6.69318914e-01 -1.72248110e-01 9.43055302e-02
6.72047436e-01 -3.42247486e-01 -8.03747848e-02 -1.82439759e-02
-3.06578308e-01 -9.12734330e-01 4.84356374e-01 3.48991811e-01
4.12611693e-01 -9.12163556e-01 1.47662771e+00 2.51346529e-01
-7.68620193e-01 -3.70404005e-01 -4.75460172e-01 6.87458515e-02
4.98838387e-02 1.17410004e+00 -1.08490312e+00 4.38490808e-01
8.84411812e-01 5.73971331e-01 -6.12462163e-01 1.11976337e+00
-9.28819552e-02 5.44703364e-01 -2.83130050e-01 4.73758668e-01
3.91635209e-01 -2.07401037e-01 6.11814678e-01 9.27894354e-01
4.64307725e-01 -4.86877084e-01 5.42027429e-02 8.48796368e-01
3.33269089e-01 1.41857509e-02 -1.01318514e+00 2.57042907e-02
-2.57218212e-01 1.12436128e+00 -1.12359846e+00 -2.85681300e-02
-4.60572958e-01 1.17495847e+00 -1.03605330e-01 8.13824609e-02
-3.03580016e-01 -6.09864831e-01 5.76891959e-01 2.72043765e-01
4.73044813e-01 1.06607527e-01 -1.76671326e-01 -5.81350863e-01
4.05636042e-01 -9.17943358e-01 4.36895430e-01 -3.53778332e-01
-1.33260036e+00 7.38931537e-01 -3.37619275e-01 -1.44780922e+00
-4.40947413e-01 -7.74924159e-01 -1.03141069e+00 7.45622873e-01
-1.08509934e+00 -6.11852348e-01 -3.94539803e-01 1.62087306e-01
8.80508900e-01 -2.54944593e-01 5.49084485e-01 1.80282921e-01
-5.01360595e-01 1.83405012e-01 -1.07071653e-01 1.94939598e-01
3.88179958e-01 -1.52573216e+00 4.55493778e-01 1.28858519e+00
8.77799764e-02 9.67851579e-02 1.04285491e+00 -7.97932267e-01
-5.36127448e-01 -9.78228033e-01 6.83103025e-01 -2.27374956e-01
3.54059875e-01 -8.45975652e-02 -1.46761620e+00 4.96184796e-01
3.82857360e-02 2.28041917e-01 2.78486937e-01 -3.94275159e-01
2.88568318e-01 2.57484317e-02 -1.39483106e+00 2.03353450e-01
4.31445926e-01 -1.80933535e-01 -5.20546198e-01 2.15443730e-01
1.30366400e-01 -3.28723937e-01 -5.36857426e-01 4.18371767e-01
1.27920672e-01 -1.42338383e+00 5.67617178e-01 -4.04054016e-01
4.65314478e-01 -3.63402933e-01 5.82319358e-03 -1.46652174e+00
1.13944918e-01 4.59636375e-02 1.66410089e-01 1.01264560e+00
3.81795794e-01 -5.50547898e-01 8.26621950e-01 3.00798178e-01
-2.79654622e-01 -9.14020002e-01 -8.85850489e-01 -6.44180655e-01
-4.96494323e-01 -3.42934400e-01 4.80342716e-01 6.29434884e-01
-4.02585447e-01 -2.14660466e-01 -1.44561296e-02 3.99859816e-01
4.24358159e-01 4.08802032e-02 4.67714280e-01 -1.55336785e+00
7.17848018e-02 -6.26255497e-02 -1.04369473e+00 -3.05423826e-01
-2.14674279e-01 -3.92050117e-01 3.90816152e-01 -1.36028075e+00
-4.52462673e-01 -2.85752207e-01 -3.39652508e-01 3.87255222e-01
-1.36634484e-01 2.42986441e-01 -3.71464640e-01 1.58185273e-01
-2.44109333e-01 2.21065775e-01 1.02512038e+00 -2.76447445e-01
-2.45256439e-01 5.45999527e-01 -2.20460668e-01 9.14174438e-01
1.07893574e+00 -4.60826784e-01 -2.11482614e-01 -7.21796826e-02
5.78086711e-02 -3.24720263e-01 4.35743839e-01 -1.28354073e+00
-4.40538041e-02 2.06982732e-01 5.10810375e-01 -6.93162382e-01
5.41610941e-02 -1.24686909e+00 -2.63858050e-01 5.93106985e-01
-3.75083444e-04 2.33980864e-01 6.35675341e-02 6.80985630e-01
-6.25477672e-01 -6.82184637e-01 7.39257514e-01 -3.74599427e-01
-8.57248545e-01 2.55685803e-02 -5.32060206e-01 -2.88741142e-01
1.22322476e+00 -6.76329374e-01 2.40439519e-01 -2.34636948e-01
-8.11284900e-01 -2.04954043e-01 4.81747538e-01 2.78507680e-01
9.17424917e-01 -8.52471828e-01 -7.17412472e-01 8.25609148e-01
3.91026825e-01 5.33946276e-01 2.62916267e-01 9.85975146e-01
-6.28303051e-01 1.99355602e-01 -3.19052249e-01 -8.83730292e-01
-1.17517948e+00 3.92247200e-01 4.75025952e-01 -9.30122808e-02
-9.57390249e-01 6.80525780e-01 2.12630983e-02 -6.52004033e-02
6.20067380e-02 -3.95535588e-01 -3.08889776e-01 -5.48628233e-02
3.69923800e-01 4.54177558e-01 5.97414434e-01 -5.84364235e-01
-1.93726331e-01 3.28192383e-01 -1.50604635e-01 4.75958735e-01
1.40950489e+00 1.03393212e-01 -2.89947867e-01 5.71057320e-01
9.06183243e-01 -1.27660319e-01 -1.24284410e+00 -1.19228419e-02
6.37939334e-01 -7.51860976e-01 -9.89191905e-02 -4.99756962e-01
-1.43163168e+00 9.39873815e-01 8.84853184e-01 7.25273311e-01
1.30683208e+00 -2.58689076e-02 4.41618472e-01 3.80034804e-01
2.31229991e-01 -1.21319234e+00 1.76092476e-01 2.57627100e-01
8.88159633e-01 -1.60393286e+00 -1.71511754e-01 -2.81401128e-01
-7.57019162e-01 1.25118160e+00 8.05262625e-01 -3.08358997e-01
4.43106800e-01 1.47920400e-01 1.51488140e-01 -6.08020663e-01
-4.73538220e-01 -1.22604884e-01 3.36953670e-01 7.28653610e-01
4.18206364e-01 -2.14317769e-01 -8.58981721e-03 8.84606019e-02
8.79902318e-02 -4.86968935e-01 6.47271514e-01 1.15959287e+00
-6.36264861e-01 -7.04086661e-01 -5.69973648e-01 1.10895753e+00
-7.56064355e-01 2.88080573e-01 -4.32347715e-01 7.43473351e-01
5.00845551e-01 9.28332984e-01 5.11657596e-01 -1.81880131e-01
6.33236468e-01 2.04057679e-01 5.91284409e-02 -7.28004038e-01
-4.92267579e-01 -6.18291311e-02 -1.64909303e-01 -5.28383791e-01
-1.51658848e-01 -6.82420731e-01 -1.21624875e+00 2.79784739e-01
-5.93384266e-01 2.12946355e-01 6.87454164e-01 1.01477337e+00
1.62288286e-02 7.98434258e-01 6.80993021e-01 -7.44468868e-01
-3.26393127e-01 -9.86181617e-01 -1.03231716e+00 7.92042136e-01
6.39286101e-01 -6.92973018e-01 -6.42375290e-01 1.41838983e-01]
|
[7.491888999938965, 2.0817506313323975]
|
82e3c928-d9ac-4833-a965-5db8c8d260f8
|
streaming-submodular-maximization-under-a-k-1
| null | null |
https://proceedings.icml.cc/static/paper_files/icml/2020/1126-Paper.pdf
|
https://proceedings.icml.cc/static/paper_files/icml/2020/1126-Paper.pdf
|
Streaming Submodular Maximization under a k-Set System Constraint
|
In this paper, we propose a novel framework that converts streaming algorithms for monotone submodular maximization into streaming algorithms for non-monotone submodular maximization. This reduction readily leads to the currently tightest deterministic approximation ratio for submodular maximization subject to a $k$-matchoid constraint. Moreover, we propose the first streaming algorithms for monotone submodular maximization subject to $k$-extendible and $k$-system constraints. Together with our proposed reduction, we obtain $O(k\log k)$ and $O(k^2\log k)$ approximation ratio for submodular maximization subject to the above constraints, respectively. We extensively evaluate the empirical performance of our algorithm against the existing work in a series of experiments including finding the maximum independent set in randomly generated graphs, maximizing linear functions over social networks, movie recommendation, Yelp location summarization, and Twitter data summarization.
|
['Amin Karbasi', 'Moran Feldman', 'Ran Haba', 'Ehsan Kazemi']
| null | null |
https://proceedings.icml.cc/static/paper_files/icml/2020/1126-Paper.pdf
|
https://proceedings.icml.cc/static/paper_files/icml/2020/1126-Paper.pdf
|
icml-2020-1
|
['movie-recommendation', 'data-summarization']
|
['miscellaneous', 'miscellaneous']
|
[ 1.30683750e-01 5.20406187e-01 -5.86611032e-01 -3.57498944e-01
-9.15114880e-01 -9.13211644e-01 -3.27735811e-01 4.44299787e-01
-2.26755321e-01 8.25328231e-01 3.17837536e-01 5.42791337e-02
-7.97680199e-01 -9.67562139e-01 -7.86276042e-01 -5.85525751e-01
-6.30194902e-01 7.77420759e-01 -2.34863162e-02 -3.98371965e-01
4.40624058e-01 2.32025683e-01 -1.18055463e+00 -1.58965304e-01
9.88720953e-01 1.10414922e+00 2.35339835e-01 6.76252961e-01
-8.06080364e-03 7.16135740e-01 -3.29368681e-01 -3.79791766e-01
8.03845584e-01 -3.69027019e-01 -9.28915918e-01 6.01379037e-01
6.12850964e-01 -4.94916439e-01 -6.62063301e-01 1.14627433e+00
4.22455400e-01 1.04932785e-01 3.41002405e-01 -1.43583751e+00
-4.17752385e-01 1.36993515e+00 -9.29465830e-01 3.80962074e-01
4.69757766e-01 -3.17058653e-01 1.54398143e+00 -3.02717388e-01
9.36771214e-01 1.03066242e+00 1.64303169e-01 1.90023810e-01
-9.19575572e-01 -2.86705106e-01 2.68662840e-01 -2.59303115e-02
-1.33665717e+00 -4.14065361e-01 7.50263035e-01 3.20516318e-01
1.00188482e+00 8.02060366e-01 7.00439513e-01 -3.53522837e-01
-1.38073459e-01 1.00678456e+00 3.43625069e-01 -1.45333365e-01
3.22473705e-01 2.24054351e-01 2.77172595e-01 6.67205989e-01
9.69746888e-01 -7.26374984e-01 -7.23000288e-01 -3.99905622e-01
1.96843237e-01 -4.03391942e-02 -4.13967758e-01 -4.12094206e-01
-8.27382565e-01 8.67219269e-01 9.46487486e-02 -1.71829328e-01
-3.50081325e-01 5.55785418e-01 2.16322988e-01 7.79691279e-01
4.70718354e-01 5.77537060e-01 -4.88053471e-01 7.23381862e-02
-1.34835207e+00 6.93562150e-01 1.16493499e+00 1.70095742e+00
6.90960526e-01 8.24831985e-03 1.05809815e-01 5.32175481e-01
2.22871140e-01 8.24065030e-01 -7.73204491e-02 -1.46565938e+00
1.05410480e+00 7.87688196e-01 1.75257206e-01 -1.26833403e+00
-3.81914377e-01 -1.02738574e-01 -6.01703286e-01 -6.15564346e-01
2.20105320e-01 -5.01987636e-01 -2.11037949e-01 1.71332014e+00
5.00453353e-01 -5.03823400e-01 5.01026027e-02 6.10153794e-01
7.47785032e-01 9.17720318e-01 -8.52537930e-01 -1.25352180e+00
7.30196416e-01 -8.52821767e-01 -7.55745530e-01 7.55847013e-03
9.41700518e-01 -2.28940740e-01 3.79686981e-01 6.27913296e-01
-1.98077822e+00 4.41449374e-01 -8.66699636e-01 6.64640497e-03
2.52887607e-01 -5.76763570e-01 8.28575730e-01 8.16234410e-01
-1.35255182e+00 2.84088016e-01 -2.86111772e-01 -3.36932421e-01
3.43399972e-01 5.97466350e-01 -1.06119327e-02 -2.75152475e-01
-6.05535269e-01 1.55245632e-01 3.21135700e-01 -3.28980833e-01
-7.90149808e-01 -7.35670209e-01 -7.93126762e-01 2.38766789e-01
9.01538730e-01 -6.93194151e-01 9.57077742e-01 -2.26632103e-01
-9.34925258e-01 8.16275477e-01 -3.55987638e-01 -6.98936939e-01
3.33827555e-01 2.46568307e-01 2.17891529e-01 6.15137637e-01
-1.33465394e-01 4.37734962e-01 5.18622458e-01 -1.12534356e+00
-7.72211552e-01 -7.45773673e-01 5.44687033e-01 6.63058758e-01
-9.42700922e-01 4.52561416e-02 -3.19798708e-01 -1.32438049e-01
3.43411863e-01 -6.91626370e-01 -7.38170326e-01 -3.17766011e-01
-5.97995698e-01 -4.43738252e-01 4.26624984e-01 -5.09399056e-01
1.80101180e+00 -1.82391179e+00 6.82394147e-01 4.98249590e-01
5.02468288e-01 -5.66234410e-01 -8.97341073e-02 8.86213720e-01
5.87067366e-01 2.42321298e-01 -3.98383260e-01 -4.55567390e-01
2.00353608e-01 6.23967312e-02 -4.02707271e-02 7.75507092e-01
-8.25969994e-01 7.78804064e-01 -6.35876715e-01 -4.61822748e-01
-5.15963018e-01 -5.71935833e-01 -8.21984708e-01 -2.58044779e-01
-7.25335658e-01 -6.27257943e-01 -3.42824876e-01 9.17258561e-01
1.10473859e+00 -3.99335176e-01 6.88255727e-01 5.91518223e-01
1.73135325e-01 -1.18212268e-01 -1.69375610e+00 1.82826626e+00
5.89083228e-03 2.11901084e-01 7.74644673e-01 -1.31800592e+00
8.52334738e-01 -1.76427718e-02 1.02143788e+00 -2.65741974e-01
1.47412479e-01 3.94687891e-01 -5.50768614e-01 -4.95824426e-01
1.04281473e+00 6.25486020e-04 -4.05157924e-01 1.18995762e+00
-2.53620148e-01 -3.79999131e-01 6.28307939e-01 1.15732169e+00
1.10907698e+00 -1.08636272e+00 9.15205777e-02 -6.17638648e-01
2.82292217e-01 2.25655183e-01 4.33021367e-01 8.25818121e-01
1.28831659e-02 5.30785263e-01 9.87532079e-01 -1.61758274e-01
-8.77681673e-01 -7.55221605e-01 2.09190965e-01 1.02367675e+00
5.92475057e-01 -4.62963343e-01 -7.90998042e-01 -5.10689139e-01
4.31679517e-01 6.38327539e-01 -4.08974528e-01 2.96239763e-01
-4.81942713e-01 -8.73812914e-01 2.41249502e-02 1.09688178e-01
1.97051510e-01 -6.18698597e-01 7.76339741e-03 1.86519310e-01
-1.36746764e-01 -1.08710146e+00 -1.13329661e+00 -1.99507296e-01
-1.30613041e+00 -8.36815596e-01 -6.40078008e-01 -9.57160234e-01
9.84073520e-01 9.59162176e-01 7.76900887e-01 5.78833334e-02
-7.41226673e-02 7.40095556e-01 -1.83044881e-01 -5.03287852e-01
6.58452362e-02 2.48880446e-01 1.14138268e-01 -1.48097977e-01
1.36453629e-01 -7.43632495e-01 -7.23885953e-01 2.02392563e-02
-1.24508417e+00 -6.28743291e-01 1.00345656e-01 2.57556830e-02
8.02608669e-01 3.78655702e-01 8.23624074e-01 -1.08644986e+00
8.26584518e-01 -1.05825043e+00 -8.41349125e-01 3.03348213e-01
-9.25036669e-01 -1.28824100e-01 6.59559786e-01 -1.28220215e-01
-6.26922965e-01 -3.91633511e-02 5.14464498e-01 -5.22627383e-02
7.24778771e-01 8.57305229e-01 -1.68976009e-01 3.41788009e-02
5.84247947e-01 6.25314772e-01 1.64002985e-01 -2.17179433e-01
6.09008849e-01 6.54618025e-01 1.17907204e-01 -2.90052444e-01
6.63552999e-01 1.13574004e+00 2.02506766e-01 -9.09784973e-01
-6.07085764e-01 -7.79862583e-01 -8.85330737e-02 -8.26887190e-02
1.16879620e-01 -8.51504564e-01 -7.85518527e-01 4.41737175e-01
-8.60316098e-01 2.19639149e-02 -6.93556726e-01 4.37956825e-02
-7.60833383e-01 9.00645196e-01 -3.49816322e-01 -9.79662180e-01
-7.55608976e-01 -4.45648432e-01 5.37703037e-01 1.95848778e-01
1.58341378e-01 -7.53640771e-01 1.10416032e-01 7.52265573e-01
3.67189318e-01 1.71741799e-01 5.48352897e-01 -5.06928563e-01
-1.10805511e+00 -3.88536930e-01 -1.21133119e-01 1.19676284e-01
-2.05956101e-01 -4.02050883e-01 1.49182361e-02 -7.97379017e-01
-6.31400868e-02 -4.53504831e-01 9.14864302e-01 7.55240738e-01
1.08266020e+00 -1.01859498e+00 -2.34554410e-01 8.49791646e-01
1.59934986e+00 -2.08881706e-01 5.43889582e-01 4.77723852e-02
2.37025991e-01 6.66215956e-01 6.63370311e-01 1.39922762e+00
7.65512466e-01 1.01507396e-01 6.80671632e-01 6.23298168e-01
5.82172513e-01 -3.27063054e-02 4.53326225e-01 9.19216216e-01
1.29193351e-01 -9.12212849e-01 -1.96512535e-01 1.22822165e+00
-1.88296044e+00 -1.10043132e+00 -2.22544268e-01 2.10106826e+00
6.12332582e-01 -2.31822491e-01 6.67421162e-01 1.05713554e-01
5.78328133e-01 4.65024859e-01 -7.61341214e-01 -7.53364325e-01
-5.42294562e-01 -9.91291106e-02 1.02074420e+00 6.57309651e-01
-3.97763580e-01 6.46978796e-01 6.25488281e+00 8.73978317e-01
-4.85959262e-01 -5.55339921e-03 4.87082034e-01 -1.15633595e+00
-1.26425576e+00 -9.31080878e-02 -1.00385034e+00 3.26546758e-01
7.48379707e-01 -1.10896790e+00 8.75156045e-01 8.97054613e-01
4.01933372e-01 -3.35452735e-01 -8.11504245e-01 1.10886145e+00
4.64310348e-01 -1.68662810e+00 4.35160734e-02 4.49121296e-01
1.40173507e+00 -1.25004351e-01 8.77056047e-02 -3.66840005e-01
2.48226285e-01 -8.54187369e-01 3.81984591e-01 -1.31739751e-01
6.90552354e-01 -1.08191526e+00 2.83899933e-01 7.34121084e-01
-9.97500777e-01 -4.02747780e-01 -6.64820313e-01 2.48742662e-02
8.22574973e-01 7.24933565e-01 -5.78678846e-01 6.28537774e-01
6.06529832e-01 3.87596548e-01 2.32285693e-01 9.20406520e-01
2.57242471e-01 4.01167333e-01 -1.04471600e+00 -4.51748431e-01
1.63147047e-01 -3.04242790e-01 1.02769947e+00 9.61422920e-01
3.96566987e-01 5.39348125e-01 -1.21755395e-02 4.58315194e-01
-7.00062394e-01 6.01989686e-01 -5.84838927e-01 -2.99855292e-01
5.99516809e-01 1.17900074e+00 -6.59628093e-01 -2.04110458e-01
-9.27113742e-02 7.12471843e-01 2.10820675e-01 1.41489044e-01
-7.61957824e-01 -5.34545064e-01 4.53722268e-01 5.80869734e-01
5.73316336e-01 -4.61627334e-01 -5.96261084e-01 -1.10968840e+00
4.56427455e-01 -5.98018825e-01 1.00919986e+00 2.45533474e-02
-8.23009372e-01 -2.35448442e-02 2.78338850e-01 -5.83105028e-01
-9.65252705e-03 1.01409361e-01 -5.07139981e-01 2.98949093e-01
-1.51981950e+00 -6.05055034e-01 -6.67091608e-02 6.40774012e-01
2.93582380e-01 -6.76508397e-02 1.93922371e-02 1.00137383e-01
-1.71992347e-01 7.82028258e-01 4.65384990e-01 -6.97256744e-01
6.82246834e-02 -1.07438171e+00 -2.12618001e-02 8.73866737e-01
-1.52846247e-01 2.92989433e-01 8.01287889e-01 -4.14661616e-01
-2.08789515e+00 -8.23388398e-01 1.13146222e+00 -1.20552413e-01
5.74337363e-01 -1.47941828e-01 -3.27921093e-01 7.31640458e-01
1.38062611e-01 -2.52983928e-01 6.87738299e-01 -1.87559947e-01
4.48910184e-02 -4.35853004e-01 -1.70969594e+00 2.17113659e-01
1.55044258e+00 2.33503237e-01 5.07909246e-03 8.55739117e-01
1.15234721e+00 -4.21071112e-01 -9.98942077e-01 1.89250439e-01
2.65174001e-01 -5.31034589e-01 7.74393737e-01 -7.23222792e-01
5.57303071e-01 1.10921882e-01 -4.98285204e-01 -8.16262841e-01
-8.57971013e-02 -1.42098868e+00 -5.84877074e-01 8.76309216e-01
8.30501080e-01 -5.40423214e-01 1.47645044e+00 5.62644899e-01
2.51091681e-02 -9.65281665e-01 -1.00230932e+00 -8.57733607e-01
1.15427405e-01 -2.21208230e-01 5.86628079e-01 8.37345839e-01
5.27016699e-01 3.32517130e-03 -6.77485943e-01 2.53340364e-01
1.06732666e+00 5.47365069e-01 9.44699764e-01 -5.95150292e-01
-7.31779993e-01 -1.42432988e-01 4.30543870e-02 -1.77230179e+00
-2.91406691e-01 -1.12750351e+00 -3.82043809e-01 -1.82271874e+00
7.95945346e-01 -3.28120381e-01 8.29308853e-02 -1.63576990e-01
2.93953001e-01 -1.38482064e-01 3.75643730e-01 -9.02356282e-02
-1.37172425e+00 5.40565789e-01 1.30713725e+00 -1.10824853e-01
-6.01767838e-01 1.86202690e-01 -1.49441171e+00 1.60224900e-01
6.91168666e-01 -4.53508943e-01 -6.40400290e-01 -7.02713907e-01
8.19643438e-01 5.97442448e-01 -5.28871775e-01 -1.45775616e-01
4.94738191e-01 -6.23237550e-01 -7.69555509e-01 -8.02930593e-01
5.72679974e-02 -5.78282654e-01 -8.15411955e-02 4.61982906e-01
-3.15407813e-01 -7.32250605e-03 -1.62700415e-01 7.72166848e-01
-3.84259075e-02 -5.94592452e-01 6.57640219e-01 -2.13325784e-01
-2.94307083e-01 7.27939665e-01 -1.20239131e-01 5.49029052e-01
1.50576353e+00 -4.10152406e-01 -4.51359451e-01 -1.14074671e+00
-5.59613526e-01 1.17855132e+00 2.58300811e-01 -2.22517282e-01
8.08801711e-01 -7.48023808e-01 -1.13799381e+00 -4.94143128e-01
-1.67345926e-01 4.97691870e-01 6.41400695e-01 9.42616105e-01
-6.95553541e-01 4.35458720e-01 4.60072398e-01 -2.97118068e-01
-1.45479989e+00 6.62755728e-01 -2.76698247e-02 -4.80173498e-01
-9.28163528e-02 1.22831321e+00 -4.12291586e-01 -3.63880605e-01
4.05814469e-01 -3.57509963e-02 2.50025034e-01 3.16027194e-01
4.17775303e-01 1.03989899e+00 -3.57801795e-01 -6.28666952e-02
-1.81883112e-01 2.17465281e-01 -3.10149968e-01 -2.45288521e-01
1.68993974e+00 -5.76657236e-01 -6.26712024e-01 -1.81351200e-01
1.24976945e+00 2.94640243e-01 -5.50160050e-01 -4.41743553e-01
-3.60492706e-01 -5.34419477e-01 -2.51125485e-01 -4.85304333e-02
-1.27348471e+00 4.84943837e-02 -4.06545997e-01 7.47196972e-01
1.28180993e+00 4.19084519e-01 1.41621268e+00 6.00450099e-01
6.46649897e-01 -1.38610840e+00 -7.45152384e-02 2.02579752e-01
7.54769027e-01 -9.55652356e-01 6.17486835e-01 -7.04204082e-01
-4.16552663e-01 8.33639145e-01 2.63919502e-01 -3.50928098e-01
8.24414074e-01 2.38930121e-01 -6.10687196e-01 -4.66995656e-01
-1.03694832e+00 1.67475745e-01 -2.55511463e-01 5.62499836e-02
-2.31635496e-01 1.06430918e-01 -1.02653635e+00 6.16162837e-01
-3.49240959e-01 -2.19063424e-02 1.06476498e+00 8.77297819e-01
-1.17283463e+00 -8.23144674e-01 -2.29537353e-01 8.70987952e-01
-5.77432454e-01 -7.47556984e-02 -4.77351904e-01 3.70088398e-01
-5.63000202e-01 1.16909051e+00 -5.43306507e-02 -1.43164173e-01
1.31855100e-01 -6.21614218e-01 5.95726371e-01 -5.77010751e-01
-4.07607466e-01 -1.53169304e-01 1.67443112e-01 -3.59286964e-01
-2.35303923e-01 -1.04313684e+00 -1.37317121e+00 -9.37509358e-01
-4.38644648e-01 2.59559602e-01 4.65598613e-01 5.09270966e-01
3.75903666e-01 -3.09696317e-01 1.14837134e+00 -1.00627579e-01
-1.26158440e+00 -6.79190993e-01 -1.25487685e+00 3.13103884e-01
-1.03287406e-01 2.80928254e-01 -7.12005436e-01 -2.66491890e-01]
|
[6.603572845458984, 4.949862480163574]
|
c19f007c-b511-459b-a56a-fab2dd5631b5
|
design-implementation-and-evaluation-of-an
|
2305.04226
| null |
https://arxiv.org/abs/2305.04226v1
|
https://arxiv.org/pdf/2305.04226v1.pdf
|
Design, Implementation and Evaluation of an External Pose-Tracking System for Underwater Cameras
|
In order to advance underwater computer vision and robotics from lab environments and clear water scenarios to the deep dark ocean or murky coastal waters, representative benchmarks and realistic datasets with ground truth information are required. In particular, determining the camera pose is essential for many underwater robotic or photogrammetric applications and known ground truth is mandatory to evaluate the performance of e.g., simultaneous localization and mapping approaches in such extreme environments. This paper presents the conception, calibration and implementation of an external reference system for determining the underwater camera pose in real-time. The approach, based on an HTC Vive tracking system in air, calculates the underwater camera pose by fusing the poses of two controllers tracked above the water surface of a tank. It is shown that the mean deviation of this approach to an optical marker based reference in air is less than 3 mm and 0.3{\deg}. Finally, the usability of the system for underwater applications is demonstrated.
|
['Kevin Köser', 'Felix Woelk', 'David Nakath', 'Birger Winkel']
|
2023-05-07
| null | null | null | null |
['pose-tracking', 'simultaneous-localization-and-mapping']
|
['computer-vision', 'computer-vision']
|
[ 8.47873464e-02 1.50605798e-01 1.19964659e+00 -3.81925553e-01
-4.23045754e-01 -8.43904018e-01 2.46707007e-01 -1.11391641e-01
-1.09264290e+00 7.22909868e-01 -2.72689253e-01 7.02579916e-02
-2.78664321e-01 -6.33164704e-01 -8.11084688e-01 -7.54581213e-01
-2.26864934e-01 5.20600915e-01 4.21062231e-01 -4.26220328e-01
4.72763568e-01 6.59141541e-01 -1.56633615e+00 -8.28214586e-01
6.48043096e-01 6.47653043e-01 5.28212607e-01 8.56840849e-01
1.96465388e-01 -1.39495507e-01 -5.87055624e-01 -2.44498268e-01
5.95724583e-01 -1.68949291e-01 -1.25983283e-01 -1.01778738e-01
9.39008415e-01 -9.24251556e-01 1.92478731e-01 1.19687283e+00
7.90680408e-01 6.37670532e-02 4.26432937e-01 -6.10508025e-01
5.60391843e-01 3.56812268e-01 -9.19715092e-02 -2.91968971e-01
4.77087170e-01 -1.11967968e-02 4.59123760e-01 -7.79106915e-01
6.28314734e-01 7.72509336e-01 1.19103944e+00 3.36377054e-01
-7.59953499e-01 -3.09782714e-01 -3.52589220e-01 -2.68942833e-01
-1.35532618e+00 -4.92286652e-01 3.67141485e-01 -8.08449447e-01
5.27126551e-01 -2.24835780e-02 9.64574158e-01 4.35995668e-01
4.07095671e-01 -3.61895293e-01 9.75876033e-01 -5.97297132e-01
4.74328279e-01 -8.29655901e-02 -7.00139329e-02 4.20450598e-01
1.01121950e+00 9.16703939e-02 -4.52334434e-01 -8.22295472e-02
6.98242724e-01 -3.82945575e-02 -8.93233120e-01 -4.77209568e-01
-6.15286112e-01 5.62452495e-01 1.62022531e-01 -2.35725418e-01
-1.06061541e-01 4.34349656e-01 5.79861514e-02 1.73902050e-01
-2.29291439e-01 6.68878794e-01 -3.63619328e-01 -3.31857592e-01
-5.27532995e-01 9.30525064e-02 1.13365638e+00 1.28436077e+00
8.91427040e-01 2.24273041e-01 9.73252952e-01 3.00629109e-01
8.13703537e-01 1.14384365e+00 1.90070748e-01 -1.08417809e+00
2.39565253e-01 3.13294590e-01 6.05861485e-01 -7.40058005e-01
-7.55757511e-01 2.81823259e-02 -1.33633120e-02 7.49256134e-01
3.89618129e-01 -8.25290263e-01 -6.25557482e-01 9.26872730e-01
1.72783092e-01 1.25234714e-02 7.18852758e-01 1.13598096e+00
8.31103146e-01 2.48388842e-01 -7.08852828e-01 -1.61840141e-01
1.03987730e+00 1.25926375e-01 -7.17573643e-01 -5.00973821e-01
5.12557626e-01 -6.52936518e-01 4.19636846e-01 2.90544838e-01
-6.99208975e-01 -9.94031653e-02 -1.42703950e+00 2.84469992e-01
-5.72057478e-02 1.35893986e-01 -8.84865690e-03 5.17367542e-01
-1.13228321e+00 4.34502959e-01 -1.13764954e+00 -5.65236866e-01
-4.69824702e-01 5.52606761e-01 -5.80518186e-01 -5.20795165e-03
-1.01453269e+00 1.33781278e+00 3.49634066e-02 7.25833654e-01
-9.80608642e-01 -5.69498003e-01 -1.13278186e+00 -4.09693033e-01
-3.40481043e-01 -1.15273654e-01 1.37935543e+00 -4.54739183e-01
-1.83791256e+00 4.30322260e-01 3.84454161e-01 -3.70866746e-01
8.44359875e-01 -7.69303977e-01 2.86636502e-01 3.48921508e-01
-5.14369756e-02 1.48032010e-01 4.43413049e-01 -1.57482290e+00
-7.59728432e-01 -4.59580630e-01 1.80969298e-01 4.43130583e-01
-8.22037980e-02 -5.29052198e-01 -4.09106821e-01 1.88554868e-01
6.77559853e-01 -1.13827538e+00 -2.98233092e-01 4.94691402e-01
8.44934210e-02 6.93565845e-01 8.28060627e-01 -4.15527582e-01
2.84461051e-01 -1.90971184e+00 -4.09236513e-02 2.79648423e-01
-3.86355042e-01 2.50970870e-02 4.18498278e-01 8.06857586e-01
6.09435320e-01 -5.24113715e-01 -1.23277076e-01 -2.02135816e-01
-2.14436695e-01 4.16538686e-01 1.46748722e-01 1.24602759e+00
-7.46250808e-01 -6.32609800e-02 -8.21453750e-01 -1.13724232e-01
2.98371911e-01 5.38253725e-01 -2.87511468e-01 2.61536270e-01
3.45779330e-01 4.40551490e-01 -3.63477804e-02 5.36030114e-01
8.80436718e-01 8.33588123e-01 4.99927759e-01 -2.91790932e-01
-9.51073587e-01 -2.79830247e-01 -1.53335273e+00 1.56112707e+00
-5.87166369e-01 1.05511606e+00 9.00934935e-01 -4.52044196e-02
1.47988570e+00 3.22623938e-01 2.17487529e-01 -2.07992837e-01
3.07215780e-01 4.91494805e-01 -1.39232621e-01 -9.21408474e-01
7.96074569e-01 -4.07489926e-01 2.07092121e-01 -8.36553499e-02
-1.96346775e-01 -6.02969885e-01 -1.89524397e-01 -2.21100673e-01
6.74663365e-01 4.41533327e-01 7.27906898e-02 -8.01254094e-01
2.41056487e-01 2.31266856e-01 5.68916321e-01 4.17206913e-01
2.34276518e-01 7.30313540e-01 -1.17057689e-01 -4.55659628e-01
-9.19843197e-01 -6.80613697e-01 -4.31579709e-01 3.21186751e-01
1.04471624e+00 2.12627247e-01 -6.98754013e-01 2.30925396e-01
2.85228670e-01 6.71932027e-02 -4.70862895e-01 2.00968668e-01
-6.53176129e-01 -3.71663690e-01 5.25143564e-01 3.64359707e-01
3.21349561e-01 -2.55116224e-01 -1.40534389e+00 3.91031027e-01
2.92311877e-01 -1.27432835e+00 2.72422075e-01 5.96320890e-02
-1.03396833e+00 -1.28055954e+00 -5.78484833e-01 -6.92784488e-01
8.36209297e-01 3.31053108e-01 5.11183679e-01 1.02794088e-01
-3.26375850e-02 5.26505470e-01 -7.68964767e-01 -5.68200171e-01
-2.40234420e-01 -6.76853240e-01 4.28548425e-01 -2.32348248e-01
-9.48251188e-02 -3.85801286e-01 -6.64796114e-01 5.61907768e-01
-5.19208908e-01 -4.68829006e-01 2.91991264e-01 3.60981554e-01
8.51787105e-02 -6.23479903e-01 -6.98432997e-02 -3.49638671e-01
-8.67620781e-02 -5.64986281e-02 -1.49978578e+00 -1.46864101e-01
-2.01070666e-01 -2.84414500e-01 2.82978982e-01 3.05983680e-03
-8.36622715e-01 7.34416485e-01 -7.70197585e-02 -9.99765694e-02
4.42894883e-02 7.35215843e-01 5.33945523e-02 -6.80823147e-01
7.83363879e-01 1.89804003e-01 4.69189316e-01 -5.46166837e-01
-2.29503989e-01 8.72899234e-01 6.36086643e-01 -2.31144205e-01
1.00377905e+00 9.14223015e-01 2.10734650e-01 -1.34314346e+00
-1.67353883e-01 -5.97244143e-01 -8.80906045e-01 -6.62065864e-01
8.08382452e-01 -1.33543432e+00 -7.62098134e-01 6.54256165e-01
-1.09535933e+00 -4.44102317e-01 3.39804202e-01 9.55851138e-01
-2.20733404e-01 7.18580306e-01 -2.80930996e-01 -9.64519322e-01
-4.41238403e-01 -1.43909585e+00 1.02998579e+00 5.71514785e-01
2.30650067e-01 -1.05275834e+00 4.03510690e-01 -6.56245127e-02
2.58428782e-01 4.20365810e-01 -5.06483138e-01 6.07854612e-02
-4.90057319e-01 -6.83729827e-01 3.72041494e-01 -9.69895944e-02
1.27171412e-01 4.22552913e-01 -9.15921450e-01 -7.49296188e-01
-6.93322942e-02 -1.05521716e-01 3.87923360e-01 1.41902208e-01
-5.02022088e-01 3.91253382e-02 -1.89217642e-01 8.16706836e-01
1.96147740e+00 1.90408185e-01 6.29193664e-01 8.55480254e-01
4.11381036e-01 6.77330315e-01 8.67836952e-01 7.19479978e-01
4.33005005e-01 7.78306186e-01 1.07794917e+00 2.49638394e-01
3.53267431e-01 2.68725961e-01 5.64312756e-01 4.41233605e-01
-5.46782017e-01 1.36566579e-01 -9.97415841e-01 5.59783518e-01
-1.30680835e+00 -5.34971714e-01 -7.36640394e-01 2.46645069e+00
1.69637278e-01 -2.29742572e-01 -7.42424786e-01 8.75482429e-03
6.22663319e-01 -5.60088575e-01 -1.30592793e-01 -1.81935117e-01
1.78695038e-01 -3.32057685e-01 1.45260465e+00 1.02263129e+00
-8.10200691e-01 6.28863335e-01 5.32142115e+00 -7.30339348e-01
-1.29718208e+00 -4.08144087e-01 -9.24900591e-01 5.96071780e-01
-2.52563860e-02 2.21986130e-01 -1.31122589e+00 9.22898278e-02
6.47449732e-01 1.39045373e-01 -1.94869205e-01 9.44978058e-01
3.75457853e-01 -4.62760061e-01 -6.96262896e-01 6.90690577e-01
2.20686033e-01 -1.13014638e+00 -2.54899293e-01 2.26982847e-01
6.72826469e-01 3.32031220e-01 -7.69196510e-01 -4.13116425e-01
2.85883099e-01 -1.63224176e-01 1.21030760e+00 7.50225127e-01
7.91016400e-01 -1.99976325e-01 1.58083856e+00 3.26177150e-01
-1.16497982e+00 1.34615868e-01 -6.56961501e-01 -5.90342939e-01
4.85583395e-01 3.50571573e-02 -1.17110610e+00 3.83794695e-01
9.04698849e-01 2.83473730e-01 -1.89998463e-01 1.73545897e+00
-3.30394477e-01 1.31838024e-01 -7.65548408e-01 -3.65676999e-01
2.58659750e-01 -5.60579062e-01 5.53469181e-01 8.65655422e-01
9.19004619e-01 3.15000921e-01 -1.41485825e-01 4.37187962e-02
3.40503573e-01 -4.36677746e-02 -6.78460419e-01 7.96857178e-01
5.37057459e-01 1.15214884e+00 -2.91249156e-01 2.14749143e-01
-2.63420403e-01 4.41227108e-01 -3.94878238e-01 9.19667725e-03
-4.11589742e-01 -8.01260650e-01 8.66570652e-01 -1.56209162e-02
1.55017719e-01 -7.01587141e-01 -3.84820215e-02 -6.39091551e-01
-8.18648264e-02 -5.44791073e-02 -3.09738547e-01 -7.33644724e-01
-5.07487118e-01 6.27318919e-01 -7.50955418e-02 -2.13418078e+00
-8.51516426e-02 -9.83160496e-01 -5.26308715e-01 8.06660175e-01
-1.66040003e+00 -7.07387745e-01 -1.24948597e+00 -1.25666456e-02
1.53872788e-01 1.33421078e-01 7.40169764e-01 1.05237804e-01
-1.29218102e-01 3.38610858e-01 6.25068009e-01 3.56062770e-01
5.50250530e-01 -1.29223192e+00 -8.26717094e-02 1.00982642e+00
-4.50543016e-01 5.47415018e-01 1.40187013e+00 -7.38936841e-01
-2.21279120e+00 -7.58602202e-01 2.08529830e-01 -1.94465593e-01
7.45057642e-01 -1.62020832e-01 -7.68710136e-01 7.16166437e-01
-4.89197113e-02 1.21446885e-01 3.06792051e-01 -5.99360883e-01
3.23907107e-01 -5.10779619e-01 -1.05792952e+00 3.12962741e-01
3.27239543e-01 7.91142136e-02 -5.19429624e-01 1.20413929e-01
1.27961636e-01 -1.04421556e+00 -1.11407757e+00 4.48548079e-01
1.06820369e+00 -1.08143413e+00 4.49410528e-01 4.15232062e-01
-1.43983409e-01 -8.11681747e-01 -3.11096340e-01 -1.17751932e+00
4.43815321e-01 -7.56964147e-01 9.55696106e-01 8.00803900e-01
4.85988528e-01 -7.49672592e-01 9.72647667e-01 3.58825475e-01
-8.74224961e-01 1.99024245e-01 -1.34751546e+00 -6.62875116e-01
1.80831905e-02 -4.00182903e-02 -7.47273564e-02 3.47889662e-01
3.37269045e-02 -6.99034631e-02 -2.99396038e-01 1.41398239e+00
9.45726871e-01 1.01905428e-01 1.36644804e+00 -1.59765148e+00
1.46237642e-01 2.46897057e-01 -1.00284755e+00 -8.55619729e-01
-2.95411140e-01 1.21351548e-01 1.00968516e+00 -1.90979755e+00
-5.68422079e-01 -2.56712079e-01 7.47649968e-01 3.18868607e-02
3.63348752e-01 3.81451875e-01 -1.22083798e-02 2.69861549e-01
1.21646589e-02 4.58628178e-01 8.62770319e-01 3.93735498e-01
-1.23241924e-01 1.09064460e-01 2.35114500e-01 1.02183092e+00
4.89110559e-01 -3.51937383e-01 5.47571220e-02 -9.60685730e-01
3.20447564e-01 3.13124686e-01 1.13538086e-01 -1.53435242e+00
8.29506099e-01 1.08512364e-01 -1.61629975e-01 -1.64064005e-01
6.42495394e-01 -1.28177738e+00 4.93431449e-01 9.01537418e-01
4.06003952e-01 -6.22700863e-02 1.04169078e-01 4.95863199e-01
-3.19022685e-01 -9.95773613e-01 9.67437267e-01 -2.54746109e-01
-1.07715034e+00 -2.02465340e-01 -3.29582483e-01 -3.73179346e-01
1.07042062e+00 -7.59947062e-01 -5.93749523e-01 -3.96310240e-01
-4.38027471e-01 2.47600988e-01 8.84296119e-01 -3.65445286e-01
1.00502396e+00 -2.29203418e-01 -7.69912243e-01 1.57194898e-01
2.95311809e-01 2.68387556e-01 8.60232711e-02 6.17697299e-01
-2.03233719e+00 -2.95338720e-01 -1.65486142e-01 -8.35717022e-01
-1.33856595e+00 -4.51763660e-01 9.61540461e-01 1.02845395e+00
-8.32797050e-01 8.68089974e-01 -4.29557413e-01 -3.75456572e-01
2.90887654e-02 -2.55664259e-01 -3.79849225e-01 4.69633304e-02
6.22207463e-01 5.05829155e-01 1.17619179e-01 -7.58557796e-01
-3.85844946e-01 1.52327335e+00 5.97183347e-01 -2.47659087e-01
1.49116886e+00 -5.05164385e-01 -4.32150178e-02 1.88659310e-01
7.71203339e-01 4.22261387e-01 -1.93232870e+00 3.16439033e-01
-4.98555414e-03 -4.36422914e-01 7.13707432e-02 -2.04891548e-01
-8.85532081e-01 7.40806878e-01 8.74851882e-01 -4.92151380e-02
6.51495397e-01 -2.36040533e-01 1.48108199e-01 8.89919698e-01
9.31414902e-01 -9.30326164e-01 -5.12752831e-01 7.78694212e-01
9.47604418e-01 -1.01915371e+00 5.12133539e-01 -3.16242963e-01
-4.20188576e-01 1.70373189e+00 5.24937212e-01 -2.39919469e-01
5.39204955e-01 8.19509327e-01 1.17348683e+00 2.63973381e-02
1.12877540e-01 -1.97168589e-01 -5.81506193e-01 5.88824511e-01
-6.61880418e-04 -2.37141013e-01 -3.90993357e-01 -8.20469856e-02
-3.94483656e-01 -3.82993490e-01 1.53059733e+00 1.31440020e+00
-1.11984718e+00 -4.82448310e-01 -8.19438696e-01 -3.32801819e-01
-2.96765119e-01 2.97177881e-01 6.14080578e-02 1.10017586e+00
1.75137501e-02 7.22261906e-01 2.39332244e-01 -3.16109955e-01
8.44093859e-01 -4.83950049e-01 1.25439599e-01 -4.81324762e-01
-3.57307434e-01 1.31564483e-01 2.66960204e-01 -7.92048499e-02
-6.48916781e-01 -9.12760794e-01 -1.66477716e+00 4.51474667e-01
-5.38402736e-01 7.88753510e-01 1.25974953e+00 6.56174481e-01
-1.57193661e-01 -1.54507160e-01 6.33904338e-01 -1.46363986e+00
-5.34036458e-01 -1.30691767e+00 -7.92360127e-01 -2.84432501e-01
4.07885045e-01 -8.50444019e-01 -9.96419609e-01 1.10608406e-01]
|
[7.502319812774658, -1.7634952068328857]
|
62cbecdb-252f-4b29-b856-3678e4061e00
|
cdf-transform-shift-an-effective-way-to-deal
|
1810.02897
| null |
https://arxiv.org/abs/1810.02897v3
|
https://arxiv.org/pdf/1810.02897v3.pdf
|
CDF Transform-and-Shift: An effective way to deal with datasets of inhomogeneous cluster densities
|
The problem of inhomogeneous cluster densities has been a long-standing issue for distance-based and density-based algorithms in clustering and anomaly detection. These algorithms implicitly assume that all clusters have approximately the same density. As a result, they often exhibit a bias towards dense clusters in the presence of sparse clusters. Many remedies have been suggested; yet, we show that they are partial solutions which do not address the issue satisfactorily. To match the implicit assumption, we propose to transform a given dataset such that the transformed clusters have approximately the same density while all regions of locally low density become globally low density -- homogenising cluster density while preserving the cluster structure of the dataset. We show that this can be achieved by using a new multi-dimensional Cumulative Distribution Function in a transform-and-shift method. The method can be applied to every dataset, before the dataset is used in many existing algorithms to match their implicit assumption without algorithmic modification. We show that the proposed method performs better than existing remedies.
|
['Maia Angelova', 'Ye Zhu', 'Mark Carman', 'Kai Ming Ting']
|
2018-10-05
| null | null | null | null |
['clustering-algorithms-evaluation']
|
['methodology']
|
[-1.86533958e-01 -4.13477607e-02 -9.16893873e-03 -2.86799759e-01
-3.75224292e-01 -5.52583575e-01 4.98952597e-01 3.90119880e-01
-3.10149103e-01 4.93167371e-01 -4.07295637e-02 -1.34760380e-01
-3.55926216e-01 -8.83645594e-01 -3.28788280e-01 -1.00854301e+00
-3.21809612e-02 7.81815529e-01 7.49962747e-01 6.28342256e-02
4.53378379e-01 6.46420062e-01 -1.55878913e+00 -7.48870149e-02
9.42286670e-01 5.28952599e-01 -6.10788800e-02 6.03076816e-01
-3.46059203e-01 5.48397601e-01 -7.26997435e-01 -3.96252386e-02
3.90902132e-01 -6.01887226e-01 -7.43278325e-01 2.99956173e-01
1.91268325e-01 1.26874313e-01 -1.13454640e-01 1.24528491e+00
3.50740522e-01 8.89770463e-02 1.06422269e+00 -1.44931388e+00
-3.10380191e-01 2.97725946e-01 -1.10036707e+00 3.23438138e-01
-2.01764181e-02 -4.54201043e-01 5.35097718e-01 -5.63349485e-01
2.58412570e-01 1.05615985e+00 7.71382987e-01 3.54198366e-01
-1.39259589e+00 -3.97756517e-01 -2.53510941e-02 1.32062539e-01
-1.88257051e+00 -2.31715709e-01 7.07409501e-01 -5.37284315e-01
4.47141737e-01 3.80383670e-01 4.76857930e-01 2.68554837e-01
-1.06823906e-01 2.55733073e-01 7.21190929e-01 -4.03454274e-01
5.47710538e-01 2.94478029e-01 -5.27943820e-02 3.98047686e-01
7.85743773e-01 -5.25088131e-01 4.96710874e-02 -5.32125115e-01
5.53171396e-01 1.60121411e-01 -1.67556614e-01 -9.49332237e-01
-8.20441484e-01 1.01390219e+00 2.26779163e-01 8.79587114e-01
-1.88013390e-01 -7.95536637e-02 2.00521350e-01 2.97083035e-02
6.27673447e-01 3.41171712e-01 -7.14912117e-02 4.28090291e-03
-1.39790964e+00 4.24737126e-01 8.62959087e-01 7.36249328e-01
1.04826903e+00 -1.66736841e-01 4.29948717e-01 7.26137340e-01
1.85330942e-01 3.52579892e-01 5.22197843e-01 -9.31847334e-01
2.28124410e-02 8.33891034e-01 5.05488962e-02 -1.35476339e+00
-1.51973844e-01 -2.32748434e-01 -1.15559077e+00 9.76983234e-02
6.88910723e-01 -7.25583360e-03 -5.86541116e-01 1.59048676e+00
6.35144889e-01 4.50092524e-01 -1.79096878e-01 5.48014760e-01
-3.95805482e-03 6.73418641e-01 -1.04034685e-01 -4.37695652e-01
7.25789487e-01 -2.38415152e-01 -5.90554774e-01 2.23784581e-01
8.02937925e-01 -6.60597324e-01 6.62325025e-01 4.02994663e-01
-9.60582256e-01 -2.80950546e-01 -9.73122120e-01 4.81592149e-01
-3.55344266e-01 -3.23283374e-01 1.10346377e-01 1.02366996e+00
-1.19396806e+00 5.67122996e-01 -7.96067476e-01 -8.06604683e-01
4.93562259e-02 4.00983453e-01 -1.35639578e-01 4.33902591e-02
-5.97313821e-01 6.54210329e-01 5.44054210e-01 -2.07490399e-01
-2.60556966e-01 -6.57360077e-01 -6.32787943e-01 5.49521782e-02
2.11340219e-01 -4.18568790e-01 7.71645844e-01 -9.51474071e-01
-9.31466997e-01 8.30078185e-01 -3.20294976e-01 -4.04375881e-01
3.55930418e-01 1.74336925e-01 -4.32541877e-01 1.72506735e-01
2.49527678e-01 2.46909752e-01 6.25877440e-01 -1.55864573e+00
-8.11266363e-01 -4.24220860e-01 -6.49514616e-01 1.53777838e-01
-5.46037257e-01 -1.83745101e-01 -4.57258254e-01 -3.71019334e-01
3.85008603e-01 -8.18061113e-01 -3.83242071e-01 -2.61301756e-01
-3.03173393e-01 -2.13328809e-01 1.30643630e+00 -1.72448084e-01
1.56725562e+00 -2.42747974e+00 4.38258164e-02 8.23179364e-01
3.96470100e-01 2.68014610e-01 2.85461396e-01 3.72468501e-01
-1.18143953e-01 2.03035548e-01 -7.18925536e-01 -3.74453098e-01
-1.31237954e-01 4.50088531e-01 -4.62444201e-02 1.06343544e+00
2.61701718e-02 1.59290612e-01 -9.15567636e-01 -8.17519307e-01
3.80053222e-01 3.93266022e-01 -7.14396238e-01 1.31051615e-01
1.75089374e-01 2.62766510e-01 -2.07616642e-01 1.96351513e-01
1.10835242e+00 -9.19423625e-02 2.56488204e-01 5.21096401e-02
-1.96398094e-01 -2.61410594e-01 -1.62600088e+00 1.10686123e+00
1.70907855e-01 4.04276550e-01 2.77026564e-01 -1.44839239e+00
1.11642790e+00 1.88898966e-01 1.17810965e+00 -2.02832118e-01
-5.61957918e-02 3.74580920e-01 8.16760585e-02 -1.65917158e-01
5.38515449e-01 -3.86131853e-01 -1.21088028e-01 4.34942991e-01
-2.63586968e-01 -1.52749047e-01 2.56406695e-01 3.86572510e-01
1.21322298e+00 -4.32894677e-01 3.77769589e-01 -6.17862523e-01
4.86862302e-01 7.16981515e-02 5.22943437e-01 7.35500515e-01
-2.98844695e-01 9.04350519e-01 5.06063104e-01 -1.27460495e-01
-1.32054150e+00 -1.22918999e+00 -2.38270089e-01 5.89593232e-01
1.95832178e-01 -3.20184439e-01 -1.01677215e+00 -4.78977710e-01
2.62169540e-02 3.86620432e-01 -6.05942905e-01 -2.08883271e-01
-5.45255899e-01 -1.15396929e+00 5.49014091e-01 8.85962229e-03
4.50726897e-01 -5.69660008e-01 -4.05478805e-01 2.52564877e-01
-1.05234623e-01 -7.12248683e-01 -3.42584401e-01 1.76623359e-01
-9.17512774e-01 -1.14632332e+00 -8.97853315e-01 -7.26841390e-01
9.16287899e-01 3.70178521e-01 9.66428101e-01 4.25834626e-01
-1.74827576e-01 2.99895257e-01 -5.11460423e-01 -5.40179312e-02
-5.55217683e-01 2.43887246e-01 3.24840307e-01 1.47251487e-01
6.85189784e-01 -8.15015376e-01 -4.04877245e-01 4.84013557e-01
-1.05621624e+00 -7.15665340e-01 2.34196290e-01 3.12441826e-01
4.23711628e-01 5.85522294e-01 7.49580503e-01 -9.74849164e-01
6.11255169e-01 -8.99704933e-01 -4.29946214e-01 -9.54509825e-02
-7.11517096e-01 -3.97073962e-02 6.13976598e-01 -3.45704317e-01
-6.91264808e-01 1.96374789e-01 7.54239177e-03 -5.39560854e-01
-5.73071539e-01 7.78947994e-02 -3.44099134e-01 1.12588026e-01
6.26312494e-01 1.79885447e-01 1.72151446e-01 -5.07927239e-01
2.24033639e-01 8.00584376e-01 6.48418486e-01 -4.70974356e-01
9.93882954e-01 7.35740721e-01 9.68501866e-02 -1.03693676e+00
-4.89688367e-01 -1.09797597e+00 -8.70797753e-01 -6.92534670e-02
6.89564943e-01 -4.31898057e-01 -2.17150837e-01 4.55027401e-01
-7.37291932e-01 -1.17671564e-02 -4.46382761e-01 3.01613510e-01
-4.50703591e-01 9.08352554e-01 -1.79263279e-01 -1.02756548e+00
1.22192338e-01 -7.71456838e-01 7.02959180e-01 8.77160206e-02
-2.89348960e-01 -1.25255644e+00 4.79391128e-01 -3.83435071e-01
4.51930314e-01 3.64491284e-01 9.47343826e-01 -7.79824018e-01
4.34025936e-02 -5.64836301e-02 -8.30840394e-02 1.42761871e-01
6.41180933e-01 3.37165743e-01 -6.27026737e-01 -5.06822526e-01
1.96770996e-01 4.25233215e-01 6.21139288e-01 4.76512998e-01
1.02284515e+00 -3.10310245e-01 -5.48279703e-01 5.01460731e-01
1.52086866e+00 3.46594006e-01 8.20309818e-01 2.82302529e-01
6.79788768e-01 7.18104482e-01 2.96780378e-01 5.37408769e-01
1.86971396e-01 5.82814336e-01 3.08315158e-01 -3.15945566e-01
2.18230069e-01 1.03848524e-01 1.10722773e-01 8.36938202e-01
2.35664740e-01 -1.83347926e-01 -1.10320175e+00 1.06237137e+00
-1.94990432e+00 -1.28364599e+00 -6.20501101e-01 2.36574221e+00
5.95292807e-01 5.65280393e-02 7.79379308e-01 5.98422289e-01
1.19319952e+00 -1.93263575e-01 -3.13466460e-01 -3.52950007e-01
3.08419969e-02 -5.98055162e-02 5.32794416e-01 4.87270862e-01
-9.58892286e-01 4.97346938e-01 6.85532665e+00 7.34343112e-01
-9.26487863e-01 -2.42687706e-02 4.42740977e-01 -2.21999417e-05
-1.55996367e-01 1.02467671e-01 -4.83001828e-01 7.61521518e-01
9.27804828e-01 -4.19843107e-01 -4.70477119e-02 9.01693702e-01
1.42741367e-01 -4.02314842e-01 -8.35991025e-01 8.69711876e-01
2.67324317e-03 -8.48888576e-01 -2.08830014e-01 5.08503556e-01
6.18872225e-01 -3.26301575e-01 -4.11845557e-02 4.05233838e-02
3.32718611e-01 -8.50148976e-01 2.82838613e-01 3.49959910e-01
4.96125400e-01 -1.19526815e+00 7.43047535e-01 4.84737664e-01
-1.38752890e+00 1.18335336e-01 -6.09530985e-01 8.85775164e-02
1.16836339e-01 1.02669120e+00 -8.35478365e-01 5.41494727e-01
8.62155855e-01 4.91250187e-01 -7.40423918e-01 1.39954662e+00
6.65874362e-01 7.39269555e-01 -7.31547236e-01 1.09295227e-01
2.40799025e-01 -4.93282974e-01 4.96398747e-01 1.23727858e+00
6.10766232e-01 -2.50695407e-01 7.51431212e-02 6.87389672e-01
3.80266398e-01 1.99684590e-01 -8.82671475e-01 3.62035215e-01
5.82519591e-01 9.74669337e-01 -1.10813260e+00 -3.79543841e-01
-3.69187236e-01 8.59196246e-01 1.72159851e-01 1.39371499e-01
-6.45059526e-01 -3.43108892e-01 4.93020654e-01 7.33845830e-01
3.67927372e-01 -2.58966863e-01 -1.24865621e-01 -8.90008211e-01
-1.69413865e-01 -6.70355797e-01 5.55423439e-01 -1.91660583e-01
-1.60585546e+00 3.59688580e-01 4.05767202e-01 -1.12660658e+00
-3.43805999e-01 -2.66029984e-01 -7.31659412e-01 5.69850087e-01
-9.60749686e-01 -5.38339674e-01 -1.62468180e-01 9.74931359e-01
6.72777044e-03 -9.79555994e-02 6.19790912e-01 6.81637228e-01
-4.10329670e-01 3.03708851e-01 4.28825110e-01 3.60914250e-03
8.14557135e-01 -1.53323197e+00 -3.62205021e-02 1.01532793e+00
1.47439232e-02 5.61250269e-01 9.94502306e-01 -7.39616036e-01
-7.28933752e-01 -1.14160395e+00 6.14817441e-01 -3.83425176e-01
4.90861297e-01 -3.76708567e-01 -1.35025036e+00 3.81753623e-01
9.18478817e-02 -2.70893332e-02 6.47303700e-01 -2.12500338e-02
4.99570630e-02 -2.64073536e-03 -1.50549221e+00 2.01003239e-01
6.46537066e-01 -5.02252802e-02 -4.54413801e-01 9.32526886e-02
2.40482971e-01 1.90041974e-01 -7.37239420e-01 1.93154246e-01
-9.19183940e-02 -1.31959832e+00 7.54104674e-01 -5.18737175e-02
-3.37259829e-01 -7.47374713e-01 -1.44249976e-01 -1.19687760e+00
-4.58468765e-01 -4.18543220e-01 -9.45346355e-02 1.42511559e+00
1.32465214e-01 -7.01295018e-01 9.59321558e-01 3.48903775e-01
1.26196072e-01 -3.07362556e-01 -1.19479609e+00 -8.77997041e-01
3.33455235e-01 -4.80144396e-02 5.92794120e-01 1.16479838e+00
1.12729959e-01 1.16706438e-01 -1.93980873e-01 2.01801166e-01
8.20150316e-01 -2.48733878e-01 9.25992906e-01 -1.61717808e+00
5.74826002e-02 -5.10739565e-01 -6.95653021e-01 -7.16157556e-01
-1.01339273e-01 -5.41793048e-01 1.18136797e-02 -1.38430858e+00
2.96589673e-01 -7.56776333e-01 6.52582198e-03 1.06962360e-01
-8.46773908e-02 2.42561281e-01 -1.42547294e-01 5.43764174e-01
-5.72181284e-01 2.09056646e-01 6.35781288e-01 3.04158032e-01
-2.14622498e-01 -1.28050921e-02 -6.16656959e-01 7.58113563e-01
1.04203272e+00 -8.50203931e-01 -4.23904538e-01 2.06154242e-01
3.34097780e-02 -5.11301517e-01 1.39455840e-01 -1.41183662e+00
3.71372342e-01 -1.62381962e-01 1.81390718e-01 -7.47801661e-01
6.52728081e-02 -1.09986663e+00 3.63868147e-01 4.24786538e-01
1.42616004e-01 4.31116343e-01 -5.14908740e-03 6.47759676e-01
-2.50237495e-01 -3.70038331e-01 1.17117739e+00 -2.23048739e-02
-3.82758766e-01 2.02563718e-01 -6.34336412e-01 -2.16966216e-02
1.45798731e+00 -4.92364079e-01 8.90339613e-02 -5.10097206e-01
-5.35162032e-01 6.55685887e-02 1.00967991e+00 -3.04512363e-02
3.44128758e-01 -1.26581860e+00 -6.34184897e-01 6.35198951e-02
-1.13732278e-01 2.38809749e-01 -8.23492371e-03 9.50215995e-01
-7.02663779e-01 -7.30804726e-02 -1.33927269e-02 -9.36369479e-01
-1.15397620e+00 8.15818131e-01 6.77332640e-01 -1.49527520e-01
-5.66191792e-01 3.20884675e-01 1.76609516e-01 -4.12844211e-01
5.64972535e-02 -2.46239323e-02 -8.88199285e-02 -4.45008744e-03
3.95516574e-01 4.06867027e-01 7.83248022e-02 -8.05678487e-01
-4.63371992e-01 7.15689063e-01 2.95776222e-02 -1.64417580e-01
1.33444226e+00 -3.91505390e-01 -2.77577937e-01 5.08033574e-01
1.12203443e+00 2.98700780e-01 -8.68882954e-01 -2.19989591e-03
2.52838761e-01 -5.55997193e-01 -2.67562419e-01 -2.63967048e-02
-1.09908128e+00 6.79183662e-01 6.43302023e-01 9.69945729e-01
1.10263216e+00 2.22085625e-01 3.79556954e-01 2.16686036e-02
2.51603156e-01 -1.18716335e+00 -3.72268967e-02 1.84087574e-01
1.87218383e-01 -8.55410576e-01 1.61826774e-01 -4.04927701e-01
-3.60656589e-01 8.12544882e-01 6.65933490e-01 -4.73502159e-01
8.55121434e-01 4.57921565e-01 -2.03151181e-01 -2.51808196e-01
-2.16192082e-01 -2.31376708e-01 -3.32635008e-02 8.89706612e-01
3.57809871e-01 -1.16322860e-01 -2.51438946e-01 4.67995405e-02
-1.04716823e-01 -3.96918863e-01 6.45887196e-01 9.12126541e-01
-7.47190118e-01 -1.12460268e+00 -8.33754539e-01 5.27849078e-01
-4.02890295e-01 2.01690525e-01 -5.86660743e-01 9.84498143e-01
2.69113421e-01 9.26585853e-01 5.93513846e-01 -1.97959557e-01
2.58905679e-01 4.90490384e-02 2.46345073e-01 -5.64963460e-01
-3.52271259e-01 2.07893223e-01 -4.05120313e-01 -2.03102604e-01
-6.31423593e-01 -7.97277749e-01 -1.20437443e+00 -8.43362749e-01
-3.85094821e-01 6.75896764e-01 3.30635041e-01 6.97220743e-01
2.63005763e-01 1.80212215e-01 6.82926357e-01 -6.57581925e-01
-1.39983356e-01 -7.49451578e-01 -1.10103178e+00 5.52335620e-01
2.11376622e-01 -5.98693192e-01 -8.41033936e-01 1.17058419e-01]
|
[7.546825885772705, 4.571656227111816]
|
81fd8e97-4c0a-42eb-94c8-975b34e5c3c1
|
avoid-overfitting-user-specific-information
|
2206.08864
| null |
https://arxiv.org/abs/2206.08864v1
|
https://arxiv.org/pdf/2206.08864v1.pdf
|
Avoid Overfitting User Specific Information in Federated Keyword Spotting
|
Keyword spotting (KWS) aims to discriminate a specific wake-up word from other signals precisely and efficiently for different users. Recent works utilize various deep networks to train KWS models with all users' speech data centralized without considering data privacy. Federated KWS (FedKWS) could serve as a solution without directly sharing users' data. However, the small amount of data, different user habits, and various accents could lead to fatal problems, e.g., overfitting or weight divergence. Hence, we propose several strategies to encourage the model not to overfit user-specific information in FedKWS. Specifically, we first propose an adversarial learning strategy, which updates the downloaded global model against an overfitted local model and explicitly encourages the global model to capture user-invariant information. Furthermore, we propose an adaptive local training strategy, letting clients with more training data and more uniform class distributions undertake more local update steps. Equivalently, this strategy could weaken the negative impacts of those users whose data is less qualified. Our proposed FedKWS-UI could explicitly and implicitly learn user-invariant information in FedKWS. Abundant experimental results on federated Google Speech Commands verify the effectiveness of FedKWS-UI.
|
['De-Chuan Zhan', 'Le Gan', 'Yunfeng Shao', 'Yinchuan Li', 'Bingshuai Li', 'Shaoming Song', 'Jin-Lin Tang', 'Xin-Chun Li']
|
2022-06-17
| null | null | null | null |
['keyword-spotting']
|
['speech']
|
[-3.14697176e-01 1.99431833e-02 -1.88427597e-01 -6.69153333e-01
-9.26194310e-01 -5.44788539e-01 2.70561606e-01 -6.07548654e-01
-3.82371932e-01 6.74894929e-01 3.43914747e-01 -3.35249901e-01
1.16508221e-02 -5.14880002e-01 -5.29253840e-01 -8.14861655e-01
2.23381653e-01 7.50759467e-02 4.43851613e-02 -1.62487209e-01
-2.30037376e-01 2.94693649e-01 -8.83705020e-01 1.00226916e-01
8.68566751e-01 9.53229785e-01 1.84954837e-01 4.45654839e-01
-1.41890675e-01 6.78274453e-01 -8.88218880e-01 -5.23298144e-01
4.46150929e-01 -1.37506232e-01 -4.99154866e-01 -3.76330346e-01
2.59697825e-01 -8.11140835e-01 -7.50810206e-01 1.06887436e+00
9.49176908e-01 2.20354125e-01 1.10084563e-01 -1.48972440e+00
-8.57059658e-01 8.01772833e-01 -3.05860490e-01 1.39977351e-01
1.79551039e-02 3.91360283e-01 7.92414784e-01 -6.24121070e-01
-3.33596882e-03 9.06899810e-01 6.33639395e-01 7.82760501e-01
-1.05965447e+00 -1.36565340e+00 5.35992205e-01 1.58776984e-01
-1.71524096e+00 -7.19836235e-01 8.07431102e-01 1.89944685e-01
3.69302958e-01 8.35998774e-01 3.88728172e-01 1.49943793e+00
-4.63113666e-01 1.08628833e+00 9.46099281e-01 1.18083090e-01
3.39923799e-01 6.48574531e-01 2.32636184e-01 2.62280822e-01
-1.02190860e-01 -4.91778143e-02 -6.81496143e-01 -8.16932857e-01
1.83622569e-01 3.00539553e-01 -7.40220129e-01 -1.41446695e-01
-7.27786124e-01 6.08933270e-01 -2.11283900e-02 1.89225644e-01
-2.09421784e-01 -5.46189137e-02 3.44260305e-01 3.31569284e-01
5.78741729e-01 1.74099132e-02 -9.64169741e-01 -1.69279650e-01
-8.71745169e-01 1.29899606e-01 8.30423534e-01 1.07426643e+00
9.41069782e-01 1.99863374e-01 -3.92299503e-01 8.37998092e-01
2.55999297e-01 7.77228534e-01 1.04944372e+00 -5.43487310e-01
5.58437407e-01 1.56756684e-01 2.17124098e-03 -9.88157928e-01
8.21617916e-02 -3.80006641e-01 -8.04324985e-01 -4.03443903e-01
1.30741775e-01 -6.32625520e-01 -6.96530461e-01 2.27351856e+00
2.43532762e-01 5.15586317e-01 -8.66392255e-02 8.29361737e-01
5.63024461e-01 6.10522807e-01 2.35555544e-02 -1.92317367e-01
1.06005538e+00 -6.85784578e-01 -9.06977355e-01 -9.47287977e-02
4.14984465e-01 -3.92327368e-01 1.32221353e+00 6.42216653e-02
-6.10273778e-01 -1.38256922e-01 -7.67581403e-01 3.93629730e-01
-6.06558502e-01 7.82529649e-04 5.64124942e-01 9.65318084e-01
-1.03242254e+00 3.12870681e-01 -4.54420418e-01 -7.52446381e-03
5.42458355e-01 6.00080013e-01 -2.10949525e-01 2.36315250e-01
-1.79538107e+00 2.99426407e-01 5.01425639e-02 -1.31360143e-01
-8.00898433e-01 -8.20656896e-01 -6.05328619e-01 2.21262112e-01
5.75201869e-01 -4.53268647e-01 1.65835178e+00 -9.10976827e-01
-1.61075091e+00 2.76773363e-01 -2.36978997e-02 -5.42435288e-01
5.32958210e-01 -3.06551773e-02 -8.23885977e-01 -2.63640821e-01
-1.89863607e-01 6.78647170e-03 1.31351781e+00 -1.19975817e+00
-3.26451898e-01 -4.48493242e-01 -4.10217568e-02 3.07992280e-01
-1.07360828e+00 -1.01289384e-01 -6.34344161e-01 -8.62005413e-01
-3.21519703e-01 -5.75015008e-01 -4.66549620e-02 -9.53643918e-02
-7.14711308e-01 -1.15368031e-01 1.12669957e+00 -8.06291461e-01
1.59448409e+00 -2.37131143e+00 -3.82460147e-01 5.32780588e-01
3.76792192e-01 5.04598439e-01 -1.06537946e-01 2.06116393e-01
1.60840824e-02 3.52865398e-01 4.25908528e-02 -5.96314669e-01
1.81122407e-01 5.31262040e-01 -4.88234580e-01 3.75485063e-01
-3.41566831e-01 7.25056171e-01 -7.79913366e-01 -2.01286867e-01
-2.83739734e-02 4.18762386e-01 -5.93553543e-01 5.97851038e-01
1.12288981e-03 5.34069836e-02 -6.29909217e-01 5.03699839e-01
9.71995950e-01 -6.03271462e-03 7.74631277e-02 -2.32100919e-01
1.24225833e-01 2.93304354e-01 -1.18454981e+00 1.27412903e+00
-6.15880489e-01 1.60538748e-01 6.63828731e-01 -9.18791652e-01
6.58550739e-01 5.06016374e-01 2.89055139e-01 -7.26704359e-01
1.53275818e-01 4.68099639e-02 -4.64903384e-01 -4.74616587e-01
4.79639202e-01 1.31469652e-01 3.07011288e-02 5.99850476e-01
3.33397612e-02 5.31332910e-01 -8.14440012e-01 4.10020053e-01
9.54544187e-01 -5.32987475e-01 -1.64019922e-03 -1.43697366e-01
3.93554062e-01 -9.18886840e-01 6.82215214e-01 1.11907852e+00
-4.84967470e-01 5.25729299e-01 2.31465492e-05 -6.73580542e-02
-4.46419239e-01 -8.12967360e-01 1.31582677e-01 1.51604176e+00
2.57290184e-01 -4.61599350e-01 -7.52951682e-01 -9.90879118e-01
8.28338787e-02 7.48683929e-01 -4.51063395e-01 -6.54865265e-01
-2.56336361e-01 -6.90344810e-01 1.03053463e+00 2.25906581e-01
6.51541054e-01 -7.03382730e-01 8.59060884e-02 4.76885773e-02
-2.93046862e-01 -7.66222596e-01 -1.30175948e+00 3.08699846e-01
-1.01833493e-01 -7.41080463e-01 -8.95996034e-01 -3.52782965e-01
5.80011725e-01 6.02417350e-01 5.48212230e-01 -1.56542838e-01
1.00692913e-01 5.13069391e-01 -2.50567436e-01 -2.42039293e-01
-3.13681960e-01 2.42402241e-01 4.99077052e-01 5.98129749e-01
4.57397342e-01 -7.61204422e-01 -7.30504870e-01 4.21733707e-01
-1.00623083e+00 -3.13623995e-01 2.29281902e-01 8.07300091e-01
-5.99126779e-02 8.52133036e-02 7.86193728e-01 -8.21910977e-01
1.11913860e+00 -7.78099656e-01 -2.95981225e-02 4.01870698e-01
-7.29368329e-01 -2.37544975e-03 9.28889155e-01 -8.62959862e-01
-1.00546181e+00 -2.64775574e-01 -4.44556564e-01 -5.80883265e-01
-6.49844781e-02 3.25317942e-02 -7.26264775e-01 -1.34805724e-01
3.74996692e-01 6.61995590e-01 -9.67934951e-02 -6.97943389e-01
3.76420051e-01 1.37145221e+00 5.46316504e-01 -5.68667710e-01
9.51910615e-01 2.14725390e-01 -1.18724239e+00 -6.88233554e-01
-3.69342387e-01 -5.55511236e-01 2.32670784e-01 5.70296459e-02
4.38494980e-01 -1.00528657e+00 -5.41070700e-01 7.64930606e-01
-1.01192677e+00 -1.62026659e-01 -2.47325227e-01 1.96948960e-01
-1.57735776e-02 4.60900605e-01 -3.07406098e-01 -1.02703500e+00
-7.51511037e-01 -9.38336730e-01 7.90633917e-01 4.14085001e-01
-1.61238760e-01 -7.82139361e-01 3.43496189e-03 2.66942501e-01
8.93483520e-01 -3.87539595e-01 4.93426502e-01 -1.40729785e+00
-2.02595696e-01 -4.06067938e-01 8.52132495e-03 6.33272886e-01
5.82374513e-01 -2.75149286e-01 -1.29794037e+00 -3.32601786e-01
1.76464096e-01 -2.65860766e-01 2.83972949e-01 -7.25273937e-02
1.71174288e+00 -1.08645248e+00 -2.31672436e-01 8.38699698e-01
9.35989201e-01 2.20683783e-01 2.57713705e-01 6.71040118e-02
6.89609826e-01 4.44827639e-02 1.59532428e-01 6.90413356e-01
3.57239753e-01 6.52719676e-01 3.64815295e-01 -3.66165563e-02
3.94219339e-01 -5.21915495e-01 5.87081850e-01 9.25110698e-01
4.80174839e-01 -4.41019952e-01 -2.20720813e-01 4.75419372e-01
-1.71260619e+00 -9.42039192e-01 5.57596385e-01 2.24691343e+00
1.17137361e+00 -1.05085231e-01 1.02597341e-01 -8.37116465e-02
7.50003159e-01 4.06380862e-01 -8.02323043e-01 -2.26220727e-01
-1.54349878e-01 2.03444093e-01 8.89296710e-01 2.79515296e-01
-8.87607396e-01 6.94129467e-01 5.22989368e+00 1.20945203e+00
-1.26247048e+00 4.56114739e-01 7.73981810e-01 -3.54141176e-01
-7.66808748e-01 -2.27269128e-01 -6.91208363e-01 1.14299273e+00
9.35507596e-01 -4.87408131e-01 8.54752243e-01 1.04296517e+00
3.19024175e-01 5.24261355e-01 -7.30789483e-01 9.58030522e-01
-2.87279673e-02 -1.10472286e+00 -1.46177843e-01 1.40432298e-01
2.72523552e-01 2.49112785e-01 3.12114686e-01 6.89382315e-01
6.34261191e-01 -5.94069004e-01 5.89920998e-01 4.49147314e-01
7.55231023e-01 -8.28881562e-01 3.88497382e-01 6.92206919e-01
-8.93125594e-01 -4.27959561e-02 -1.87891439e-01 4.82081443e-01
-1.22235253e-01 3.16804856e-01 -7.35220850e-01 4.75311875e-01
9.15936947e-01 -1.41085060e-02 -3.55537921e-01 6.32811844e-01
1.78540289e-01 8.54165912e-01 -7.00241745e-01 -7.57111013e-02
1.97763264e-01 4.46947105e-02 5.98995686e-01 1.16992307e+00
2.40908563e-01 7.54763708e-02 2.11602509e-01 5.92573225e-01
-4.19233739e-01 2.55785137e-01 -6.22699142e-01 1.81533992e-02
9.13961351e-01 1.15665901e+00 1.81652755e-01 -3.12485844e-01
-4.31773216e-01 1.31609821e+00 3.31864268e-01 6.72679543e-01
-9.09480095e-01 -4.27562445e-01 1.24191737e+00 1.21724539e-01
1.01506814e-01 1.03722803e-01 3.34889889e-01 -1.27076340e+00
4.34814300e-03 -1.06323242e+00 4.03171659e-01 -6.07079387e-01
-1.51543987e+00 5.30178428e-01 -2.39824131e-01 -9.45721149e-01
-9.97861177e-02 -4.94149774e-02 -9.15238738e-01 9.57226872e-01
-1.30690861e+00 -8.74687791e-01 -6.65774569e-02 1.22489774e+00
3.14992785e-01 -3.72340888e-01 9.09155846e-01 6.29011929e-01
-8.12163472e-01 1.54936159e+00 4.69588816e-01 3.69740248e-01
9.63587344e-01 -1.00593257e+00 2.19666183e-01 7.09208190e-01
1.38154179e-01 1.03004730e+00 4.71623242e-01 -4.44642901e-01
-1.27230787e+00 -1.19599140e+00 5.29722035e-01 -1.96237177e-01
6.16435885e-01 -5.61733365e-01 -1.14638448e+00 6.32561266e-01
1.97814777e-01 2.96026826e-01 9.58086073e-01 -7.94238970e-02
-4.98463422e-01 -4.36617851e-01 -1.41043103e+00 7.02198744e-01
7.08767831e-01 -9.20604706e-01 -2.08627075e-01 3.33267927e-01
1.15686989e+00 -3.43670249e-01 -4.24717307e-01 2.96783745e-02
4.16736394e-01 -7.58051872e-01 8.56263459e-01 -8.42110813e-01
-6.32032514e-01 -1.49313912e-01 -2.77685165e-01 -1.42160773e+00
-1.22516930e-01 -1.14948499e+00 -4.85663861e-01 1.73877716e+00
1.62441939e-01 -1.00182033e+00 8.10000837e-01 1.14504600e+00
1.00633688e-01 -5.47037721e-01 -1.10733306e+00 -6.05108261e-01
-3.15093458e-01 -6.23374462e-01 1.19673395e+00 1.14670694e+00
-8.78267512e-02 -6.56197071e-02 -1.01897669e+00 5.43111861e-01
3.72602582e-01 -2.26508915e-01 8.05908561e-01 -6.39913380e-01
-4.54548508e-01 -1.34882838e-01 -3.10122613e-02 -1.24509001e+00
2.72448421e-01 -7.81689048e-01 -1.41541496e-01 -7.26351321e-01
-1.26182258e-01 -5.25790334e-01 -6.29498959e-01 8.98419380e-01
-2.86932111e-01 -2.08101645e-02 1.00277297e-01 1.46680772e-01
-5.80788612e-01 7.81987369e-01 7.94369042e-01 -2.67309010e-01
-2.90522128e-01 5.46091378e-01 -1.27100480e+00 5.02540767e-01
1.07425678e+00 -4.82524157e-01 -6.57371104e-01 -2.56782591e-01
-3.21503729e-01 -3.74310136e-01 2.15560481e-01 -6.39506161e-01
3.55001628e-01 -2.81883329e-01 1.05133526e-01 -2.15968102e-01
2.00676933e-01 -1.17326593e+00 6.10809624e-02 4.71498966e-02
-2.46293098e-01 -3.25197637e-01 1.84571166e-02 8.31916213e-01
6.42010272e-02 -9.22472998e-02 4.99415994e-01 -1.01089656e-01
-4.68688130e-01 7.20291793e-01 -4.83209670e-01 8.45943093e-02
8.72479141e-01 5.95442727e-02 -1.24849752e-01 -8.76712024e-01
-6.38862729e-01 5.38106561e-01 7.34171644e-02 6.73489690e-01
4.06700552e-01 -1.39150548e+00 -2.98727095e-01 5.41699946e-01
5.75632788e-02 -6.76653385e-02 3.15093994e-01 3.21025729e-01
2.08574712e-01 8.56423154e-02 5.09295106e-01 -2.51683414e-01
-1.37883365e+00 3.94252449e-01 5.06206572e-01 2.83724982e-02
-3.96275759e-01 9.73451912e-01 1.05349280e-01 -8.48945022e-01
6.93255126e-01 -1.63421318e-01 2.24874020e-01 -8.65225494e-02
6.91614509e-01 1.23274714e-01 2.18095034e-01 -3.94615859e-01
-5.35528779e-01 -1.25509083e-01 -4.31690782e-01 9.63009521e-02
9.71334338e-01 -4.10958648e-01 3.23470265e-01 8.16524327e-02
1.49485004e+00 5.18756807e-01 -1.25879085e+00 -7.49307394e-01
-4.13775444e-01 -5.37225962e-01 8.98342654e-02 -7.10823834e-01
-1.41343462e+00 4.10501599e-01 6.61104679e-01 3.22472572e-01
1.17162311e+00 -1.38270915e-01 1.04063022e+00 2.60495871e-01
3.59974235e-01 -1.18133569e+00 -1.74870163e-01 2.08929971e-01
4.35153723e-01 -1.12601674e+00 -4.38881874e-01 2.01251507e-01
-7.59763598e-01 6.84157073e-01 7.48038292e-01 5.27599454e-01
9.04759109e-01 4.59160149e-01 3.38303804e-01 2.23345846e-01
-5.94735682e-01 1.13629863e-01 3.63478400e-02 5.93628168e-01
-1.43420070e-01 7.53563941e-02 7.38272220e-02 1.37357652e+00
-1.01044260e-01 -1.64437160e-01 1.47115842e-01 7.74192512e-01
-1.21551983e-01 -1.08153880e+00 -2.98233271e-01 5.90596974e-01
-5.60789347e-01 -2.61444420e-01 -4.29134548e-01 2.90332288e-01
1.16994016e-01 7.71916866e-01 -2.31538713e-01 -8.59655499e-01
3.69702190e-01 3.34911019e-01 -3.70828688e-01 -3.27690065e-01
-7.65685081e-01 -1.01916902e-01 -2.85158992e-01 -5.30078471e-01
2.07980406e-02 -3.89838278e-01 -1.04997826e+00 -5.39396882e-01
-5.45567393e-01 4.58077133e-01 4.61096168e-01 8.85114789e-01
6.35827065e-01 3.59069467e-01 1.25573826e+00 -3.63932699e-01
-1.13588083e+00 -7.55758166e-01 -6.81526363e-01 2.70996898e-01
5.23701072e-01 -1.00748144e-01 -7.16811299e-01 -4.81697232e-01]
|
[5.804357528686523, 6.387543201446533]
|
040b5bbf-de78-4e4f-8114-25b9725a9f43
|
spelling-error-correction-with-soft-masked
|
2005.07421
| null |
https://arxiv.org/abs/2005.07421v1
|
https://arxiv.org/pdf/2005.07421v1.pdf
|
Spelling Error Correction with Soft-Masked BERT
|
Spelling error correction is an important yet challenging task because a satisfactory solution of it essentially needs human-level language understanding ability. Without loss of generality we consider Chinese spelling error correction (CSC) in this paper. A state-of-the-art method for the task selects a character from a list of candidates for correction (including non-correction) at each position of the sentence on the basis of BERT, the language representation model. The accuracy of the method can be sub-optimal, however, because BERT does not have sufficient capability to detect whether there is an error at each position, apparently due to the way of pre-training it using mask language modeling. In this work, we propose a novel neural architecture to address the aforementioned issue, which consists of a network for error detection and a network for error correction based on BERT, with the former being connected to the latter with what we call soft-masking technique. Our method of using `Soft-Masked BERT' is general, and it may be employed in other language detection-correction problems. Experimental results on two datasets demonstrate that the performance of our proposed method is significantly better than the baselines including the one solely based on BERT.
|
['Shaohua Zhang', 'Hang Li', 'Haoran Huang', 'Jicong Liu']
|
2020-05-15
|
spelling-error-correction-with-soft-masked-1
|
https://aclanthology.org/2020.acl-main.82
|
https://aclanthology.org/2020.acl-main.82.pdf
|
acl-2020-6
|
['csc']
|
['natural-language-processing']
|
[ 5.01014411e-01 -3.19326997e-01 2.43498400e-01 -1.79457992e-01
-7.50849068e-01 -1.91281885e-01 3.26977760e-01 4.97833282e-01
-8.21915090e-01 8.71279716e-01 5.57514839e-02 -5.62606871e-01
3.49797189e-01 -6.36963367e-01 -6.93915486e-01 -5.93037128e-01
3.98373634e-01 1.95249781e-01 6.21536434e-01 -4.13113922e-01
7.43820012e-01 2.30927065e-01 -1.36831987e+00 4.83914733e-01
1.35568333e+00 6.36765778e-01 8.01141679e-01 6.28662884e-01
-4.86838639e-01 7.79589474e-01 -8.48694861e-01 -3.49866927e-01
-1.22070439e-01 -6.89860046e-01 -6.91514313e-01 -3.08493644e-01
1.30624801e-01 -2.15488719e-03 2.23959416e-01 1.39093816e+00
4.49668616e-01 1.27889797e-01 6.18976772e-01 -5.18937826e-01
-7.07308650e-01 5.38483918e-01 -5.03073037e-01 6.05007708e-01
3.05170357e-01 -3.22763234e-01 7.50224650e-01 -1.03672767e+00
2.73182631e-01 8.43777716e-01 7.81038225e-01 8.11771452e-01
-7.05466330e-01 -5.88773549e-01 1.38508439e-01 2.94598371e-01
-1.53228283e+00 -3.02552223e-01 4.23232406e-01 -4.61947471e-01
1.12506354e+00 2.62883484e-01 2.42106155e-01 4.85556334e-01
3.52870375e-01 6.08815968e-01 1.10982466e+00 -1.10345328e+00
9.59719345e-02 2.96931863e-01 3.84280473e-01 5.64367473e-01
3.50030273e-01 -1.12349808e-01 -2.93700039e-01 9.92952660e-02
4.31052804e-01 -1.41448811e-01 -4.96228993e-01 2.63372838e-01
-8.40313733e-01 4.92377639e-01 2.86969572e-01 7.88235664e-01
-2.68153965e-01 2.31506433e-02 3.28896850e-01 2.44169474e-01
4.47287738e-01 3.51092607e-01 -4.31902170e-01 -4.79342155e-02
-1.24801600e+00 1.54652447e-01 5.08766532e-01 8.91747117e-01
5.76688945e-01 1.83206916e-01 -3.71959597e-01 7.09995747e-01
1.85248286e-01 6.76287785e-02 9.32492554e-01 9.68063250e-02
7.73261428e-01 7.89368451e-01 3.37522835e-01 -8.80826116e-01
-3.70726526e-01 -5.85885108e-01 -8.04085672e-01 1.33128464e-01
4.07779783e-01 -1.34667650e-01 -9.82077777e-01 1.50588810e+00
-1.11617289e-01 2.40919665e-01 2.11220477e-02 9.50136423e-01
6.63995445e-01 8.10612381e-01 1.82752654e-01 -4.08757031e-01
1.25627124e+00 -1.11566257e+00 -1.05345571e+00 -3.82779539e-01
8.45868587e-01 -9.08028305e-01 1.03951645e+00 3.83232385e-01
-9.76169765e-01 -5.47458291e-01 -1.19568837e+00 -5.02402447e-02
-4.77435738e-01 8.29571128e-01 1.77084357e-01 7.58235216e-01
-1.02976799e+00 6.48382723e-01 -5.36492229e-01 -3.12517464e-01
-2.07626760e-01 2.71676958e-01 -1.65944532e-01 1.15952358e-01
-1.25342441e+00 1.13833904e+00 6.85961008e-01 3.78920943e-01
-4.74865258e-01 -4.64468040e-02 -5.73109329e-01 2.05759436e-01
3.23880553e-01 -1.44093573e-01 1.31183624e+00 -1.42057025e+00
-1.23357975e+00 8.46984208e-01 -5.94105661e-01 -5.34399569e-01
5.52943110e-01 -3.33778739e-01 -6.70145690e-01 -1.58554703e-01
-7.63813872e-03 7.70372301e-02 7.99764395e-01 -1.10228682e+00
-8.15421879e-01 -2.45352849e-01 -3.10679853e-01 3.68562229e-02
-1.20528787e-01 6.33335650e-01 -4.31595176e-01 -9.91286039e-01
1.68548033e-01 -6.85022652e-01 -1.18456826e-01 -3.44397157e-01
-2.98508018e-01 -3.56860340e-01 4.66627657e-01 -1.11124349e+00
2.03202295e+00 -2.12340355e+00 -1.08315103e-01 3.31957340e-02
-1.54985622e-01 9.61576462e-01 6.02226332e-02 5.55969656e-01
-2.24628687e-01 2.81928420e-01 -5.37023365e-01 -6.08110070e-01
-4.16925937e-01 -1.53124675e-01 -2.87700415e-01 4.05943125e-01
2.61441350e-01 4.80909169e-01 -8.38211060e-01 -2.73112029e-01
-1.70156196e-01 2.07175449e-01 -4.40491468e-01 3.45687360e-01
-1.23783097e-01 3.12956572e-01 -9.26974863e-02 3.98443103e-01
8.64568532e-01 -3.02634505e-03 2.44200397e-02 3.02647829e-01
-5.09685040e-01 6.58205152e-01 -1.48304725e+00 1.29345632e+00
-1.66271403e-01 4.75363016e-01 -7.52892271e-02 -8.84082019e-01
1.10775566e+00 3.62966776e-01 -4.04372483e-01 -4.79837537e-01
2.11001173e-01 6.23136580e-01 1.68725058e-01 -5.07222056e-01
9.26192880e-01 -1.22951366e-01 1.68093771e-01 4.10420179e-01
-3.03017765e-01 3.81285787e-01 2.95694292e-01 5.07730059e-02
7.60411739e-01 -6.64106160e-02 5.50195932e-01 -2.42156237e-01
1.08483112e+00 -1.60513610e-01 7.55484998e-01 9.88087237e-01
-3.54766190e-01 6.75828040e-01 1.27526954e-01 -1.44447669e-01
-8.24812889e-01 -5.63744307e-01 9.33864415e-02 9.25394833e-01
1.56884804e-01 -3.51747394e-01 -1.00362217e+00 -5.61232090e-01
-3.63104284e-01 9.37193632e-01 -3.93710226e-01 -1.37372077e-01
-8.21899474e-01 -8.26200187e-01 6.31011069e-01 5.51083684e-01
7.00015366e-01 -1.16481531e+00 -3.99750948e-01 4.27785665e-01
-2.85574436e-01 -8.27520013e-01 -6.62170053e-01 3.97303194e-01
-6.81791067e-01 -8.57721150e-01 -6.58745527e-01 -1.20386982e+00
8.41642678e-01 4.79247957e-01 6.57555461e-01 9.21664476e-01
2.21511900e-01 -2.87782907e-01 -7.66214728e-01 -5.45222878e-01
-7.88470745e-01 -2.90828515e-02 -1.66248307e-02 1.70352966e-01
6.41569793e-01 2.71208417e-02 -2.33244717e-01 -4.09495384e-02
-8.94087672e-01 4.05214317e-02 5.26866674e-01 8.45517635e-01
3.47598553e-01 1.34930745e-01 5.29818058e-01 -1.04196668e+00
8.75834107e-01 -2.30474710e-01 -4.74054635e-01 3.58709574e-01
-6.20355308e-01 -4.07488197e-02 7.35274196e-01 -1.48015693e-01
-1.06404674e+00 -5.41881211e-02 -4.55235243e-01 2.79686749e-01
-1.65738672e-01 5.95731497e-01 -2.60005705e-02 -1.52618870e-01
6.05544806e-01 6.04375064e-01 -3.64129871e-01 -8.87633920e-01
-1.70458913e-01 1.10703456e+00 5.40238976e-01 -1.56793237e-01
4.73974615e-01 -9.38127860e-02 -3.91530186e-01 -5.47371209e-01
-5.65712631e-01 -6.36051774e-01 -7.21289337e-01 -2.35107131e-02
8.01008821e-01 -8.09633613e-01 -2.76083142e-01 7.11804807e-01
-1.67802072e+00 6.86835721e-02 3.25754642e-01 5.06679714e-01
-4.96757291e-02 7.26215005e-01 -7.85694838e-01 -1.13917732e+00
-4.56566632e-01 -1.14135671e+00 6.90287769e-01 2.89315879e-01
7.58514851e-02 -6.77300513e-01 -7.94255733e-02 1.43676952e-01
4.44792122e-01 -2.56505162e-01 1.01577413e+00 -8.38446081e-01
-4.38135713e-01 -3.48366201e-01 -2.93830663e-01 6.74351215e-01
-8.01121071e-02 -2.42614821e-01 -8.64266038e-01 -3.38506639e-01
1.55850008e-01 1.48881420e-01 1.13695967e+00 3.99713181e-02
1.22898662e+00 -2.17467159e-01 -2.05235094e-01 1.81706876e-01
1.66130078e+00 3.32841784e-01 8.33703458e-01 5.14632165e-01
5.41354537e-01 3.82479578e-01 5.59885740e-01 3.55226815e-01
2.53408670e-01 8.11580479e-01 2.81351626e-01 4.49840315e-02
-1.32696480e-01 -1.95313901e-01 3.94142151e-01 1.12196577e+00
-1.95487440e-01 -5.51124096e-01 -9.61020648e-01 5.21930754e-01
-1.76495135e+00 -9.29044545e-01 -6.53048038e-01 2.43596315e+00
9.81972337e-01 2.67761171e-01 -1.95933864e-01 5.79114020e-01
1.13951945e+00 -2.42442444e-01 1.45640776e-01 -9.29908395e-01
-2.39577398e-01 7.41811097e-02 4.41115379e-01 6.67369962e-01
-1.04406214e+00 1.10889888e+00 5.74502993e+00 9.68112350e-01
-1.25583565e+00 2.77801961e-01 2.94302613e-01 5.36781073e-01
-1.76924184e-01 -1.44885276e-02 -1.18792641e+00 7.13781834e-01
8.11220765e-01 2.34774396e-01 4.72787678e-01 4.69577312e-01
4.40059334e-01 -4.54279989e-01 -6.96483850e-01 7.81932473e-01
5.04460394e-01 -1.04196787e+00 9.77029949e-02 -4.59878117e-01
6.22041702e-01 -2.99108297e-01 -3.88552666e-01 3.23478192e-01
-1.92378700e-01 -8.49106908e-01 1.05761826e+00 4.56619829e-01
6.13036573e-01 -6.56159639e-01 9.46067154e-01 8.40651512e-01
-8.68720531e-01 -4.91049252e-02 -5.20018220e-01 -5.26507914e-01
-2.13418659e-02 5.17128885e-01 -7.13643312e-01 3.76509935e-01
4.08413559e-01 3.30005854e-01 -8.09080303e-01 1.47851765e+00
-5.35021365e-01 7.35682905e-01 1.08633615e-01 -4.82580006e-01
1.75277397e-01 -5.51221631e-02 5.26537001e-01 1.75090992e+00
5.38479090e-01 8.16474631e-02 -2.80148499e-02 6.93145514e-01
1.03619441e-01 5.04212797e-01 -2.10796177e-01 1.37421668e-01
4.47460890e-01 6.94777489e-01 -6.82839930e-01 -4.32970643e-01
-4.02307540e-01 1.36626291e+00 6.44462407e-01 1.84421659e-01
-5.96158683e-01 -7.03312099e-01 1.76315114e-01 -6.10568095e-03
4.73955035e-01 -2.81996548e-01 -5.27306437e-01 -1.24712813e+00
1.67211726e-01 -8.22052836e-01 1.87617406e-01 -6.98156655e-01
-8.96809518e-01 7.42147326e-01 -4.38178897e-01 -1.21921694e+00
4.67574820e-02 -7.29167759e-01 -6.38341963e-01 1.21549213e+00
-1.88182890e+00 -7.24697769e-01 -1.54294282e-01 3.46182585e-01
6.83423698e-01 -5.45122325e-02 7.63861001e-01 5.89058518e-01
-7.31041551e-01 9.20152605e-01 2.25583941e-01 1.27403513e-01
7.81947851e-01 -1.25920928e+00 2.39026323e-01 1.57147658e+00
3.90860662e-02 8.14340174e-01 8.15292180e-01 -9.50975657e-01
-8.62565637e-01 -1.06464124e+00 1.87793493e+00 -1.79486319e-01
2.15474576e-01 -3.05539906e-01 -1.25944090e+00 4.65627193e-01
1.45938918e-01 -3.24389368e-01 5.01819670e-01 -2.65398264e-01
-1.25168087e-02 1.52971521e-01 -1.08766985e+00 5.40141284e-01
7.25672245e-01 -3.21645528e-01 -9.57835376e-01 2.36573994e-01
6.16893828e-01 -5.46286464e-01 -3.51180322e-02 3.39485765e-01
-1.57858636e-02 -9.68935490e-01 4.03318226e-01 -3.95882010e-01
3.53004783e-01 -7.43640184e-01 -8.47265124e-02 -1.32210612e+00
-5.56718409e-01 -3.75442177e-01 2.60278314e-01 1.39710498e+00
4.42504376e-01 -3.71337205e-01 3.04168671e-01 1.65562093e-01
-4.30301249e-01 -5.17045438e-01 -9.05927420e-01 -7.03122914e-01
3.49199143e-03 -5.26591182e-01 6.33521974e-01 7.48164058e-01
-3.36452685e-02 1.17323838e-01 -6.36850715e-01 3.85751605e-01
1.73572253e-03 -2.99292684e-01 3.25527817e-01 -8.87225151e-01
-4.14285421e-01 -5.26186883e-01 -5.28668910e-02 -1.22301149e+00
-7.05412105e-02 -8.87905777e-01 5.40111899e-01 -1.37384570e+00
1.27554581e-01 -4.48691577e-01 -3.50333333e-01 2.53619879e-01
-6.53806627e-01 2.69677877e-01 3.22987378e-01 2.75749058e-01
-3.35672647e-01 3.20683658e-01 9.24665928e-01 1.16992086e-01
-2.60857075e-01 1.90630570e-01 -6.06571257e-01 6.92602336e-01
7.15653658e-01 -6.70654655e-01 2.22165078e-01 -7.88689613e-01
2.86055386e-01 -1.70058072e-01 1.86642855e-01 -1.07131469e+00
4.01960611e-01 -1.47602363e-02 1.71110839e-01 -5.64824998e-01
-7.18022063e-02 -6.77787006e-01 -2.25394949e-01 6.33056283e-01
-3.75902176e-01 4.75217253e-01 2.99475461e-01 5.33720553e-01
-2.26958692e-01 -9.99673367e-01 8.88760209e-01 -2.57546246e-01
-9.51131761e-01 -3.72735560e-01 -7.12647021e-01 -1.62344351e-01
7.77992427e-01 -3.03823113e-01 -2.14265406e-01 2.13687457e-02
-3.73257488e-01 -9.35256109e-02 2.77232021e-01 3.48364234e-01
6.19736791e-01 -1.04400730e+00 -6.69638216e-01 4.25125390e-01
2.10377350e-01 -3.91775072e-01 -4.30029817e-02 7.85106897e-01
-8.67508352e-01 5.40436924e-01 8.17024633e-02 -1.45354122e-01
-1.34992611e+00 6.06922805e-01 3.44799191e-01 -4.01724517e-01
-3.55738193e-01 9.73267376e-01 -1.56581208e-01 -3.29266071e-01
4.59146917e-01 -3.22935045e-01 -6.61532938e-01 -2.91674078e-01
9.63500202e-01 3.08599532e-01 4.67859715e-01 -8.05129111e-01
-4.86891568e-01 3.62660408e-01 -1.05848663e-01 4.53230292e-02
8.72680426e-01 -2.98688769e-01 -4.16551679e-01 4.73144531e-01
6.40513361e-01 3.75971973e-01 -5.03335297e-01 -2.82154977e-01
2.51022279e-01 -4.99119461e-01 9.36335176e-02 -1.06917477e+00
-6.06416941e-01 1.11856318e+00 6.39789760e-01 -1.12148426e-01
1.19529712e+00 -6.27728641e-01 7.71805108e-01 2.13683590e-01
3.65332633e-01 -1.23779142e+00 -3.03971976e-01 8.88608396e-01
7.88873911e-01 -1.22779930e+00 -3.15908551e-01 -4.95901197e-01
-5.05180299e-01 1.28125906e+00 7.62093186e-01 -3.29068244e-01
4.04134035e-01 5.27991578e-02 2.81907115e-02 3.75941157e-01
-5.35819471e-01 -1.69954941e-01 3.26781064e-01 3.28671455e-01
8.54932368e-01 6.30083382e-02 -1.29901278e+00 8.34267557e-01
7.84826204e-02 4.04479168e-03 8.24607432e-01 9.26012218e-01
-1.02994621e+00 -1.24797916e+00 -5.97092688e-01 3.09674710e-01
-6.99781895e-01 -5.27400017e-01 -2.73367733e-01 3.13498765e-01
5.80542147e-01 1.08975840e+00 -1.34683382e-02 -3.63860071e-01
2.31976062e-01 1.64583847e-02 2.13871747e-01 -9.11023736e-01
-1.07912862e+00 -2.47211736e-02 -1.76943783e-02 -3.00261620e-02
-2.02608466e-01 -6.68999791e-01 -1.37846518e+00 -1.75107941e-01
-7.54424036e-01 2.79556900e-01 6.76789701e-01 1.25181091e+00
1.41272843e-01 3.75926912e-01 2.66527802e-01 -3.78349215e-01
-6.33972943e-01 -1.22062516e+00 -4.92241710e-01 2.36922130e-01
4.54857379e-01 -4.47311044e-01 -2.41359845e-01 -1.00679725e-01]
|
[10.956887245178223, 10.80453109741211]
|
a175efa5-c9da-4f11-b2c3-1a976ec7cb8f
|
gradient-surgery-for-one-shot-unlearning-on
|
2307.04550
| null |
https://arxiv.org/abs/2307.04550v1
|
https://arxiv.org/pdf/2307.04550v1.pdf
|
Gradient Surgery for One-shot Unlearning on Generative Model
|
Recent regulation on right-to-be-forgotten emerges tons of interest in unlearning pre-trained machine learning models. While approximating a straightforward yet expensive approach of retrain-from-scratch, recent machine unlearning methods unlearn a sample by updating weights to remove its influence on the weight parameters. In this paper, we introduce a simple yet effective approach to remove a data influence on the deep generative model. Inspired by works in multi-task learning, we propose to manipulate gradients to regularize the interplay of influence among samples by projecting gradients onto the normal plane of the gradients to be retained. Our work is agnostic to statistics of the removal samples, outperforming existing baselines while providing theoretical analysis for the first time in unlearning a generative model.
|
['Woohyung Lim', 'Hyemin Jung', 'Seoyoon Kim', 'Seohui Bae']
|
2023-07-10
| null | null | null | null |
['multi-task-learning']
|
['methodology']
|
[ 4.85541701e-01 3.81204069e-01 -2.79310226e-01 -3.14766914e-01
-5.23058712e-01 -5.49917758e-01 7.80539274e-01 -2.85235584e-01
-7.22076058e-01 8.39404941e-01 5.66975057e-01 -6.30060881e-02
-2.44512245e-01 -5.35895050e-01 -1.07929909e+00 -9.38813806e-01
3.44501823e-01 4.33566719e-01 -4.12887335e-02 -5.06867953e-02
1.94867268e-01 4.22329038e-01 -1.32837439e+00 1.69171974e-01
3.71461570e-01 4.20388162e-01 -1.68333739e-01 7.65621483e-01
2.86009818e-01 5.29789925e-01 -4.89196777e-01 -4.99109596e-01
4.40009922e-01 -5.42816341e-01 -4.35237199e-01 3.51065546e-02
7.29740739e-01 -3.61554950e-01 -4.51609194e-01 9.24712479e-01
6.23958945e-01 2.49502718e-01 8.88938129e-01 -1.00769401e+00
-8.85703743e-01 6.94663107e-01 -4.83552545e-01 4.03932422e-01
-2.88527548e-01 -8.16926174e-03 1.24381542e+00 -1.11700010e+00
6.83367610e-01 1.25391066e+00 5.02432585e-01 8.03822756e-01
-1.59918654e+00 -5.03220081e-01 3.57988864e-01 -1.06778648e-02
-1.04930270e+00 -6.00738347e-01 8.26881289e-01 -3.94246876e-01
8.89575303e-01 3.14656287e-01 6.72243059e-01 1.44792199e+00
1.32464379e-01 8.73191059e-01 1.07621181e+00 -6.44144773e-01
6.59709424e-03 5.50045855e-02 -7.50118820e-03 6.99836433e-01
4.76244062e-01 2.33232528e-01 -6.75282240e-01 -1.57008171e-01
9.23619866e-01 6.97609857e-02 -7.43112713e-02 -6.97746217e-01
-1.12550461e+00 8.53345037e-01 3.37044001e-01 3.51215243e-01
-1.43116921e-01 4.14923459e-01 4.58726957e-02 2.74817973e-01
5.93500555e-01 4.34849769e-01 -6.93534136e-01 -5.38141318e-02
-9.82580125e-01 3.16206157e-01 6.84171557e-01 5.55969059e-01
1.05578864e+00 1.83554143e-01 -5.37778974e-01 7.82489181e-01
1.07434466e-01 4.01372433e-01 5.93384206e-01 -6.14615023e-01
6.93466812e-02 3.15649360e-01 -2.68165916e-02 -6.86365068e-01
-2.74193943e-01 -6.96103036e-01 -7.91519642e-01 3.10188085e-01
5.11812985e-01 -3.60092014e-01 -1.09577215e+00 1.88769495e+00
1.92382619e-01 2.98038691e-01 -2.73276806e-01 8.42180490e-01
9.00003761e-02 1.77337065e-01 -1.31459892e-01 4.94093299e-02
7.32297957e-01 -1.08108783e+00 -4.31187570e-01 -4.56956863e-01
3.46403182e-01 -6.69044912e-01 9.52883184e-01 6.52830958e-01
-8.91901553e-01 -3.12116295e-01 -1.06561244e+00 -3.73576343e-01
-3.35919112e-01 1.85956329e-01 7.26152062e-01 4.94031578e-01
-1.09442818e+00 9.71961021e-01 -8.26417983e-01 -9.82697159e-02
6.26676738e-01 3.16729426e-01 -3.36958826e-01 -2.59868614e-02
-9.14221764e-01 9.73207533e-01 2.72578299e-01 1.41159771e-02
-1.28661215e+00 -1.00097823e+00 -6.31484210e-01 1.26827002e-01
5.35930514e-01 -9.25835252e-01 1.00679719e+00 -1.02135491e+00
-1.59614038e+00 7.19255328e-01 -2.57253051e-01 -3.35201770e-01
7.03686416e-01 -6.19967222e-01 -6.08605891e-03 -4.11285400e-01
-8.15029740e-02 5.85755944e-01 1.54444492e+00 -1.18242037e+00
-3.43949556e-01 -3.57186258e-01 -2.97193438e-01 1.83312416e-01
-3.78824025e-01 -3.29744220e-01 -4.61971372e-01 -8.88673365e-01
-3.75220254e-02 -9.14582074e-01 -4.25516576e-01 -1.53204024e-01
-4.51983392e-01 6.38757795e-02 7.72696018e-01 -3.07490975e-01
1.18838847e+00 -2.11927199e+00 4.36818838e-01 2.34006375e-01
5.58269620e-01 2.99038947e-01 -3.90710741e-01 2.63085395e-01
-1.39116794e-01 5.64998128e-02 1.24242818e-02 -5.15066028e-01
3.06472719e-01 4.35162663e-01 -5.43453634e-01 6.45548582e-01
1.57385379e-01 1.08765423e+00 -1.02405763e+00 -3.39157158e-03
1.73202798e-01 6.83191717e-01 -9.35055077e-01 1.23516902e-01
-3.43942791e-01 2.01593906e-01 -2.94169068e-01 2.26033583e-01
4.59381402e-01 -1.29607663e-01 8.25371519e-02 5.44538014e-02
2.21261531e-01 3.27337116e-01 -1.07083011e+00 1.46530735e+00
-4.25571352e-01 4.78489399e-01 1.20702595e-01 -1.06957877e+00
6.63521409e-01 2.47914363e-02 3.52881759e-01 -3.05352986e-01
3.99487577e-02 8.81888419e-02 1.56456336e-01 -2.34824076e-01
4.08331156e-01 -3.88004988e-01 9.26144198e-02 5.78310966e-01
5.65990150e-01 -1.26742646e-01 1.24623857e-01 1.18867651e-01
9.49252188e-01 3.64405662e-01 3.61662924e-01 -4.00812507e-01
7.02961674e-03 -5.68352461e-01 4.53669041e-01 1.00130057e+00
-1.14362938e-02 5.53377032e-01 4.92255390e-01 -4.59522724e-01
-1.18727756e+00 -1.18675292e+00 -4.61795405e-02 1.60959804e+00
-5.22525370e-01 -3.72907817e-01 -7.81705976e-01 -7.15064645e-01
2.85925239e-01 8.12519848e-01 -1.04123843e+00 -3.04143012e-01
-5.77184796e-01 -1.08251226e+00 4.69255775e-01 3.22690904e-01
-7.79946670e-02 -7.95288205e-01 -2.16741607e-01 1.60294980e-01
2.12975726e-01 -7.24280238e-01 -6.30147994e-01 5.04717231e-01
-9.38083470e-01 -8.98612976e-01 -6.57233000e-01 -2.43917391e-01
9.60172236e-01 3.71174552e-02 1.05254495e+00 -1.56401433e-02
-2.68305212e-01 3.62133801e-01 1.73324123e-01 -6.90609157e-01
-1.92661822e-01 4.80137765e-01 -1.22074252e-02 1.69467658e-01
3.93934786e-01 -8.46818566e-01 -5.34255862e-01 1.52375713e-01
-9.06064749e-01 5.74121773e-02 8.51157546e-01 9.13770735e-01
4.97949153e-01 -4.53995973e-01 6.02290988e-01 -1.15362334e+00
6.67280853e-01 -4.29135740e-01 -3.51010799e-01 8.61210972e-02
-8.30808878e-01 5.63872218e-01 5.88683486e-01 -6.66200757e-01
-9.68986034e-01 -1.73357241e-02 6.18184470e-02 -6.88418567e-01
-8.03558379e-02 1.46188661e-01 -3.65676805e-02 1.08995251e-01
6.84957087e-01 8.02178010e-02 -2.58830130e-01 -6.19458318e-01
7.85946846e-01 -1.63015813e-01 2.56394744e-01 -6.91105664e-01
8.56205642e-01 5.54350317e-01 4.69854623e-02 -6.33166313e-01
-1.28441131e+00 -3.09501439e-01 -8.29670727e-01 -1.40570819e-01
5.95755875e-01 -5.85063636e-01 -3.41237009e-01 3.37205678e-01
-7.48121679e-01 -5.63075066e-01 -4.84065473e-01 4.37384486e-01
-3.46484661e-01 6.58610389e-02 -4.54690486e-01 -5.92974365e-01
-1.84627756e-01 -7.71482587e-01 1.06208229e+00 -7.00775953e-03
-3.24191570e-01 -1.23140168e+00 4.10328299e-01 9.48165357e-03
7.20835686e-01 -8.42397511e-02 8.35477769e-01 -7.35774398e-01
-6.43147171e-01 -1.87705949e-01 4.53236997e-02 4.10561621e-01
9.35955718e-02 3.49093787e-02 -1.21049583e+00 -3.90296608e-01
4.12831306e-02 -1.86965629e-01 1.46611464e+00 4.14480954e-01
1.11501586e+00 -5.97733319e-01 -3.18791866e-01 7.94189632e-01
1.11423349e+00 -5.40065467e-01 6.20002985e-01 1.36820450e-01
8.58309627e-01 2.37832859e-01 -9.15432721e-02 4.31100249e-01
9.53619368e-03 2.89259404e-01 2.64152884e-01 -1.30722195e-01
-1.28939867e-01 -5.83167434e-01 4.03383195e-01 7.11775780e-01
-2.77889550e-01 -1.54638320e-01 -3.32273871e-01 5.47926605e-01
-1.79676509e+00 -1.01072419e+00 3.02168339e-01 2.09099722e+00
1.09004211e+00 4.18255538e-01 -1.69174194e-01 -1.26028791e-01
3.77176404e-01 4.92325902e-01 -6.20824158e-01 -2.60780036e-01
-1.74263641e-01 5.49761057e-01 6.33458555e-01 9.24786031e-01
-1.06485546e+00 1.16146576e+00 7.26131105e+00 8.92224848e-01
-1.23687553e+00 1.64733648e-01 4.57622081e-01 -7.93443322e-01
-6.74177647e-01 1.82485536e-01 -1.05305243e+00 4.72756699e-02
6.80940509e-01 -2.53798831e-02 7.32996047e-01 8.40066552e-01
1.42811775e-01 -2.58119255e-02 -1.22619355e+00 4.95573103e-01
3.33994001e-01 -1.16833627e+00 2.83164561e-01 1.39822930e-01
9.24100041e-01 1.40412107e-01 3.04304034e-01 4.27430421e-01
6.96505249e-01 -1.09221375e+00 6.06507838e-01 8.73579144e-01
4.32810575e-01 -4.83753890e-01 1.81447119e-01 3.42821389e-01
-4.44503874e-01 5.11729829e-02 -5.27794540e-01 -2.86848336e-01
-7.64246136e-02 1.06404626e+00 -9.63669479e-01 9.98167545e-02
3.32928717e-01 5.40632606e-01 -6.85364783e-01 5.50519109e-01
-6.16235137e-01 9.12362695e-01 -3.31290603e-01 -1.36459107e-02
1.71445191e-01 -2.75977671e-01 6.92392468e-01 1.29907227e+00
3.11906725e-01 -2.73642808e-01 -3.63749228e-02 9.88627911e-01
-3.76457572e-01 -1.73444003e-02 -7.29278445e-01 -1.78305302e-02
-1.34004891e-01 1.19684947e+00 -4.14330989e-01 -2.80476689e-01
-2.72968084e-01 1.16159213e+00 6.23379886e-01 6.01585329e-01
-6.27213240e-01 -6.63299859e-02 5.69301605e-01 2.55978554e-01
6.06565952e-01 -3.35574657e-01 -3.62601310e-01 -1.29123437e+00
8.60479660e-03 -6.55401349e-01 3.23513359e-01 -4.71732318e-01
-1.23903954e+00 2.22116619e-01 -3.81025337e-02 -6.53083384e-01
-2.75558323e-01 -6.52214646e-01 -5.70303082e-01 9.04107690e-01
-1.43094921e+00 -1.11363494e+00 2.44705230e-01 4.73773628e-01
3.41632664e-01 -1.24700986e-01 7.48048663e-01 1.40631363e-01
-5.91839612e-01 5.86231709e-01 2.09576041e-01 -4.77260277e-02
9.08514977e-01 -1.36920416e+00 3.56970340e-01 1.13079572e+00
5.35702884e-01 1.07922935e+00 7.68133819e-01 -6.50455952e-01
-1.23486996e+00 -8.47830832e-01 8.55372787e-01 -6.94357455e-01
8.22255075e-01 -7.56904483e-01 -8.12833607e-01 1.07613075e+00
3.31243664e-01 2.26364821e-01 6.43789649e-01 4.35376197e-01
-6.37947202e-01 -8.46872181e-02 -7.11097896e-01 6.90826595e-01
1.03018570e+00 -4.80147034e-01 -5.54873943e-01 8.19741488e-02
3.61478269e-01 -3.62841934e-01 -3.74075800e-01 2.97620684e-01
7.02126205e-01 -8.55723858e-01 9.87461507e-01 -1.08061063e+00
2.44784385e-01 5.68870381e-02 4.63099852e-02 -1.69739711e+00
-3.96610856e-01 -9.73950624e-01 -3.32259536e-01 8.64716351e-01
5.08820057e-01 -4.81077731e-01 1.01292551e+00 4.54761982e-01
-4.78608161e-02 -7.67546952e-01 -8.51204693e-01 -3.86552840e-01
3.73552114e-01 -3.23414683e-01 1.24328367e-01 8.68004918e-01
-3.19856167e-01 7.46623218e-01 -7.73649633e-01 -1.90195277e-01
7.09455073e-01 -2.33298615e-01 8.30063820e-01 -1.18276525e+00
-6.27094448e-01 -6.04550242e-01 5.66863269e-02 -1.07070851e+00
9.89555940e-02 -1.14847243e+00 -7.86637738e-02 -1.27967870e+00
4.27838534e-01 -1.12903200e-01 -5.37852943e-01 5.80910206e-01
-4.51726615e-01 1.13261715e-01 8.73903260e-02 4.53684703e-02
-5.73472738e-01 6.49950445e-01 1.45507658e+00 1.14814341e-01
-1.81870945e-02 1.03800721e-01 -8.61085951e-01 8.63465726e-01
7.50435710e-01 -7.64213443e-01 -5.15389323e-01 -4.90480334e-01
4.95592743e-01 -6.41748190e-01 3.28818411e-01 -6.31160319e-01
9.04431790e-02 -1.33937523e-01 6.24640763e-01 -4.32695836e-01
1.74418479e-01 -6.02627635e-01 -1.33527189e-01 1.14735283e-01
-5.18707037e-01 8.14752281e-03 1.53221935e-01 6.58078969e-01
2.23621592e-01 -4.85400528e-01 7.76221275e-01 -2.55904168e-01
-3.77378345e-01 2.75112003e-01 -4.80435491e-01 3.31716627e-01
5.60808063e-01 3.19476515e-01 -3.08610618e-01 -3.49875391e-01
-9.31353629e-01 -1.82697415e-01 2.35704467e-01 4.00482655e-01
4.30598915e-01 -1.12358677e+00 -6.58582270e-01 4.90402907e-01
-2.36290440e-01 -9.61143076e-02 1.49068862e-01 6.87883914e-01
-1.29861040e-02 4.30667877e-01 -3.95520851e-02 -1.46249548e-01
-8.50469947e-01 7.32286274e-01 3.55062932e-01 -5.44338822e-01
-4.44308400e-01 9.52762723e-01 2.52063185e-01 -4.81516123e-01
2.86453933e-01 -4.47077721e-01 2.31101841e-01 2.60929048e-01
3.33623588e-01 2.80798852e-01 2.13504747e-01 -2.45534942e-01
-3.84103470e-02 2.79752582e-01 -4.36977684e-01 -3.06826770e-01
1.44945168e+00 7.97143206e-02 -1.44849360e-01 6.24307871e-01
1.30691528e+00 4.37486798e-01 -1.41755712e+00 -4.04393941e-01
1.02964595e-01 -3.77213091e-01 9.66987237e-02 -8.28925669e-01
-9.38981652e-01 9.33134794e-01 1.87759534e-01 -1.53432950e-01
6.67108893e-01 -2.41855290e-02 4.46033269e-01 6.02789998e-01
-2.37936825e-02 -1.26734650e+00 4.72367913e-01 6.43422484e-01
9.26141798e-01 -1.12277412e+00 2.16987416e-01 2.07601339e-02
-4.75454658e-01 1.06370795e+00 5.59267581e-01 -3.17707688e-01
9.48159218e-01 5.33936918e-01 1.71316881e-02 -1.37284517e-01
-8.17126811e-01 -6.88062087e-02 4.86744165e-01 4.39509749e-01
3.62075090e-01 -6.45787418e-02 -9.29838642e-02 4.99965906e-01
-1.78626150e-01 1.20204024e-01 4.45617378e-01 8.21313024e-01
-4.12585765e-01 -1.18186545e+00 -1.34241134e-01 6.08564556e-01
-6.47507608e-01 -4.25896019e-01 -4.69464034e-01 7.70482481e-01
6.40785471e-02 3.01402837e-01 -4.60847467e-02 -1.57879755e-01
2.67943650e-01 3.43365878e-01 7.74524927e-01 -7.70461380e-01
-3.53105903e-01 3.97633821e-01 -7.05927834e-02 -3.27441365e-01
-3.39413851e-01 -7.99831569e-01 -7.72350788e-01 -1.06949471e-02
-1.89881608e-01 -2.05011949e-01 3.26112837e-01 1.02464330e+00
2.84890503e-01 5.24081230e-01 2.74470001e-01 -1.03162944e+00
-8.19283128e-01 -1.10709536e+00 -5.66028535e-01 4.80714947e-01
5.15553176e-01 -7.13917434e-01 -6.43371224e-01 3.03799808e-02]
|
[8.398063659667969, 3.590689182281494]
|
2766a0bd-36ff-4dbf-8ca7-c09fed1bf715
|
human-machine-knowledge-hybrid-augmentation
|
2304.13963
| null |
https://arxiv.org/abs/2304.13963v2
|
https://arxiv.org/pdf/2304.13963v2.pdf
|
Human-machine knowledge hybrid augmentation method for surface defect detection based few-data learning
|
Visual-based defect detection is a crucial but challenging task in industrial quality control. Most mainstream methods rely on large amounts of existing or related domain data as auxiliary information. However, in actual industrial production, there are often multi-batch, low-volume manufacturing scenarios with rapidly changing task demands, making it difficult to obtain sufficient and diverse defect data. This paper proposes a parallel solution that uses a human-machine knowledge hybrid augmentation method to help the model extract unknown important features. Specifically, by incorporating experts' knowledge of abnormality to create data with rich features, positions, sizes, and backgrounds, we can quickly accumulate an amount of data from scratch and provide it to the model as prior knowledge for few-data learning. The proposed method was evaluated on the magnetic tile dataset and achieved F1-scores of 60.73%, 70.82%, 77.09%, and 82.81% when using 2, 5, 10, and 15 training images, respectively. Compared to the traditional augmentation method's F1-score of 64.59%, the proposed method achieved an 18.22% increase in the best result, demonstrating its feasibility and effectiveness in few-data industrial defect detection.
|
['Xiaoqiao Wang', 'Yu Gong', 'ChiChun Zhou']
|
2023-04-27
| null | null | null | null |
['defect-detection']
|
['computer-vision']
|
[ 2.74828583e-01 -1.50245249e-01 1.15140177e-01 3.92714189e-03
-5.45930505e-01 -1.31274208e-01 6.38313890e-02 4.97144401e-01
-1.39445558e-01 5.60956240e-01 -4.49990958e-01 5.33505082e-02
8.70828703e-02 -6.20033026e-01 -4.42332566e-01 -8.01764071e-01
3.75160456e-01 4.24631476e-01 3.57861727e-01 -2.30489448e-01
5.63359439e-01 3.08962733e-01 -1.65130413e+00 2.69389510e-01
1.01410747e+00 1.26274240e+00 5.84903419e-01 3.85413677e-01
-2.57337019e-02 5.69682240e-01 -8.80420506e-01 -8.12113378e-03
2.98187464e-01 -1.75962776e-01 -2.08664924e-01 6.26706600e-01
1.37180507e-01 -1.40241295e-01 -1.06026478e-01 1.00921369e+00
5.55168331e-01 1.90060511e-02 5.35261631e-01 -1.14091444e+00
-5.84308982e-01 -1.63079783e-01 -1.04852057e+00 1.53123245e-01
8.62579197e-02 3.95233274e-01 4.34297800e-01 -9.85984981e-01
4.32198018e-01 9.25238609e-01 4.43527341e-01 5.17299712e-01
-9.29385602e-01 -6.87551618e-01 1.47283703e-01 9.04520974e-02
-1.21299160e+00 -1.09141260e-01 9.82970059e-01 -6.83061361e-01
6.15251005e-01 1.93925144e-03 5.83700120e-01 6.80072367e-01
2.41158187e-01 6.54485881e-01 1.19425225e+00 -5.49667656e-01
1.80198014e-01 3.05984676e-01 4.16841209e-02 9.46162701e-01
4.64131743e-01 8.65397155e-02 -4.79357481e-01 2.09245831e-01
1.09184301e+00 3.24641585e-01 -2.09124506e-01 -3.07162821e-01
-1.07107198e+00 6.83921754e-01 2.94051975e-01 2.12123081e-01
-3.46558213e-01 -4.05872881e-01 4.45636779e-01 2.58719981e-01
4.85421330e-01 5.55420876e-01 -5.54719388e-01 -5.47862016e-02
-4.74185228e-01 -5.20127155e-02 3.14242721e-01 9.42645967e-01
8.38657618e-01 2.61462569e-01 -9.81701091e-02 1.00936711e+00
2.42986947e-01 5.63576639e-01 4.26750183e-01 -7.55172312e-01
5.56068897e-01 8.48258495e-01 1.34318128e-01 -1.15005970e+00
-2.26296723e-01 -6.29397810e-01 -1.05821800e+00 4.28877920e-01
3.23129147e-01 5.65019734e-02 -1.32411659e+00 1.03728843e+00
3.56084585e-01 -1.29660115e-01 -1.65842161e-01 9.65887249e-01
5.94051182e-01 6.64055943e-01 -9.02321413e-02 -4.04436201e-01
1.03136384e+00 -9.79406416e-01 -8.80783975e-01 -3.75584662e-01
3.01296532e-01 -1.08555317e+00 1.31363261e+00 9.51425552e-01
-7.35199451e-01 -1.00368977e+00 -1.14674783e+00 3.02299976e-01
-8.01251009e-02 6.40038848e-01 6.08519018e-01 4.89763945e-01
-4.13545161e-01 5.20351768e-01 -7.13973522e-01 -1.30432397e-01
6.42274916e-01 3.25053602e-01 -2.39580035e-01 -5.49456835e-01
-5.64956188e-01 4.53587681e-01 3.41539651e-01 1.71869546e-01
-1.16917229e+00 -6.79011464e-01 -7.01703668e-01 -3.42086673e-01
6.60126090e-01 -3.82870495e-01 9.89824414e-01 -5.87125659e-01
-1.42809701e+00 4.90561098e-01 8.74203891e-02 2.37334326e-01
2.99516797e-01 -5.01587212e-01 -4.12233859e-01 1.51084468e-01
5.27680032e-02 3.00711066e-01 9.38584507e-01 -1.56703925e+00
-8.75360787e-01 -6.05932057e-01 -1.26531338e-02 1.55223593e-01
-4.28712934e-01 -2.60049880e-01 -5.42491317e-01 -6.79104745e-01
3.40241611e-01 -6.45585537e-01 -3.21773231e-01 1.96762204e-01
-3.49614680e-01 -5.73500106e-03 9.53940392e-01 -9.53898549e-01
1.04707432e+00 -2.30702853e+00 -1.41279161e-01 1.90760627e-01
2.52947688e-01 5.11627257e-01 -1.84828602e-02 1.12315990e-01
1.12934619e-01 -2.00383291e-01 -2.08104268e-01 -2.27034334e-02
-3.04626137e-01 -2.35832781e-02 4.46860403e-01 2.81709969e-01
4.54625994e-01 4.92348909e-01 -8.73083472e-01 -4.26915258e-01
5.91546297e-01 7.90054873e-02 -4.56989646e-01 2.91473150e-01
-1.27929196e-01 7.23576546e-01 -5.49913406e-01 1.10441732e+00
6.62859321e-01 -1.50753900e-01 -1.16641238e-01 -3.13272178e-01
5.39001301e-02 -5.29977620e-01 -1.31332552e+00 1.68867350e+00
-6.22233868e-01 4.55996275e-01 1.88061565e-01 -1.17466903e+00
1.36596859e+00 1.89355403e-01 4.81471509e-01 -8.90025139e-01
2.61913478e-01 2.25603864e-01 -1.50030293e-02 -8.39721322e-01
1.87556654e-01 -3.27956498e-01 -6.96971593e-03 6.80313027e-03
-5.65308332e-02 -3.67054194e-01 2.35106364e-01 -1.26632750e-01
1.08845425e+00 -1.70031130e-01 9.19793174e-02 6.72409236e-02
5.09490907e-01 -3.27626541e-02 7.79902875e-01 3.18888485e-01
-1.82632893e-01 7.30842769e-01 6.21537492e-02 -5.20450115e-01
-1.15500319e+00 -9.55587983e-01 1.46602824e-01 4.91373152e-01
3.44017565e-01 -1.27319857e-01 -2.80119628e-01 -7.76667416e-01
1.10825919e-01 3.04432005e-01 -3.43489081e-01 -4.38530713e-01
-3.92108381e-01 -6.25633955e-01 -2.64532655e-01 5.56838512e-01
6.38601243e-01 -1.01609707e+00 -2.84818530e-01 1.71971411e-01
4.17189859e-03 -8.73472452e-01 -1.52271688e-01 2.19980460e-02
-1.01995873e+00 -1.21977651e+00 -6.80942118e-01 -1.06363535e+00
1.18155003e+00 2.57389098e-01 9.48892593e-01 2.95666218e-01
-7.90434122e-01 -2.11771354e-02 -3.86877835e-01 -5.99857569e-01
-2.82342821e-01 -4.33217287e-01 -1.95875149e-02 -1.65175080e-01
7.59128407e-02 -2.35230312e-01 -5.22894919e-01 4.64558721e-01
-7.53060699e-01 -1.37320995e-01 1.09214246e+00 1.11288023e+00
7.34582245e-01 6.74133956e-01 7.02578604e-01 -8.16142142e-01
4.63826776e-01 -1.93160385e-01 -5.52786529e-01 5.44547774e-02
-7.26828516e-01 -2.97742099e-01 7.59707510e-01 -4.68444139e-01
-1.25593305e+00 7.82932341e-02 -3.34803984e-02 -6.05474472e-01
-4.11390007e-01 3.45331252e-01 -3.75634253e-01 -2.94076931e-02
5.49940050e-01 1.12567618e-01 1.66527387e-02 -5.79042375e-01
-1.20030314e-01 8.30489218e-01 3.07422608e-01 -5.01377225e-01
9.39263523e-01 1.55180782e-01 -1.24044806e-01 -9.51404691e-01
-7.54717052e-01 -7.79438972e-01 -5.73847473e-01 -3.38617295e-01
5.80889761e-01 -8.78382027e-01 -5.08959711e-01 8.85595560e-01
-8.50207865e-01 -6.39672726e-02 -4.73369062e-01 6.03457510e-01
-1.25310212e-01 5.65875411e-01 -5.92899024e-01 -8.39573860e-01
-2.66370118e-01 -1.00268245e+00 9.85687077e-01 2.18397215e-01
2.95345187e-01 -6.16949856e-01 -4.44410622e-01 6.79824352e-01
2.65886813e-01 3.23785096e-01 1.10433781e+00 -2.77166575e-01
-4.99245346e-01 -5.23860335e-01 -2.48597398e-01 7.81351864e-01
6.15873158e-01 -1.19755007e-01 -6.87799513e-01 -4.23785150e-01
1.74560830e-01 -5.16589165e-01 5.70014536e-01 1.84949979e-01
1.27278697e+00 1.15129603e-02 -3.02386314e-01 -6.00786246e-02
1.48708606e+00 6.37412071e-01 6.02838814e-01 5.39909713e-02
8.58914554e-01 5.15583992e-01 1.32049537e+00 7.34801888e-01
-5.96305057e-02 4.21603233e-01 5.58816910e-01 -2.70356148e-01
-1.64923728e-01 4.39150073e-02 7.66487122e-02 1.24702644e+00
-1.19025394e-01 -1.06127277e-01 -7.82229960e-01 6.38153970e-01
-1.55686748e+00 -5.55563927e-01 -1.57131642e-01 2.11039591e+00
7.78003454e-01 5.93859911e-01 -1.54503301e-01 5.67353964e-01
9.39245939e-01 -3.01793724e-01 -7.33806252e-01 -2.44468655e-02
2.53094614e-01 2.87676543e-01 1.58213869e-01 -1.03585050e-01
-1.07175362e+00 5.38716733e-01 5.29898643e+00 9.47332382e-01
-9.75913048e-01 -2.62544155e-02 6.80588543e-01 -2.53776424e-02
1.06269188e-01 -2.91431993e-01 -5.69818914e-01 4.35796946e-01
3.89847368e-01 2.13636994e-01 2.84327120e-01 8.37883234e-01
1.37270361e-01 -4.10774857e-01 -7.99521804e-01 1.15674734e+00
1.47961721e-01 -8.91872168e-01 -2.75391191e-02 -4.39680889e-02
1.06821132e+00 -6.47886276e-01 1.15870595e-01 2.03636393e-01
1.92775458e-01 -8.01406562e-01 3.61833692e-01 4.79921639e-01
7.03106523e-01 -8.85517597e-01 1.13007712e+00 4.28924441e-01
-1.01754963e+00 -4.62247163e-01 -4.71738338e-01 -1.85966432e-01
4.92655486e-02 1.14623022e+00 -8.23062360e-01 7.84086466e-01
6.70105100e-01 6.51258051e-01 -5.36141276e-01 1.12198317e+00
-2.00180560e-01 5.72093129e-01 -9.98056456e-02 9.51981023e-02
-1.53083056e-01 1.27443401e-02 7.61727169e-02 4.73207384e-01
5.01192093e-01 -6.94032907e-02 5.70832431e-01 5.15646338e-01
7.34148026e-02 1.16566308e-01 -4.30966765e-01 -1.22179314e-01
2.89223462e-01 1.23000157e+00 -7.46350765e-01 -1.58141181e-01
-5.94946504e-01 8.63886297e-01 1.02770299e-01 6.05670922e-02
-7.51985550e-01 -6.11952603e-01 2.68028408e-01 2.16034353e-01
3.30107689e-01 -2.71163195e-01 -1.80943415e-01 -7.23696709e-01
2.00471207e-01 -1.13838601e+00 1.69188514e-01 -7.73051620e-01
-1.44703281e+00 2.88396716e-01 -3.89134139e-01 -1.34647071e+00
3.23752344e-01 -9.85074639e-01 -4.23437893e-01 7.27794945e-01
-1.23800004e+00 -1.14333475e+00 -6.90882385e-01 2.87529171e-01
1.00484931e+00 -4.17025626e-01 4.68667150e-01 4.85828847e-01
-8.06269825e-01 4.22773480e-01 2.27297217e-01 -9.76826400e-02
6.77766383e-01 -1.11736357e+00 -3.46102528e-02 7.11538970e-01
-6.58498555e-02 2.40375549e-01 5.03009081e-01 -8.25777769e-01
-1.36455214e+00 -1.15815175e+00 1.10691346e-01 -1.33463606e-01
2.19176799e-01 -2.08521634e-01 -8.12333047e-01 3.05630028e-01
-9.63624865e-02 2.30524376e-01 3.46932977e-01 -5.95270991e-02
3.87882888e-01 -3.00878793e-01 -1.34789371e+00 2.39923671e-01
8.69298637e-01 -2.01333493e-01 -2.49740973e-01 3.02135140e-01
4.63858604e-01 -4.43600804e-01 -1.05722880e+00 7.36743271e-01
2.89349735e-01 -7.50239491e-01 7.08833694e-01 -2.01149553e-01
4.93615627e-01 -5.64708352e-01 -3.85065898e-02 -1.35977209e+00
-3.22427273e-01 -2.35168934e-02 -9.89603810e-04 1.32276976e+00
3.57211173e-01 -3.97322029e-01 8.82112503e-01 1.87546164e-01
-5.47396362e-01 -9.63021517e-01 -4.88687992e-01 -6.51606143e-01
-3.28418881e-01 -1.79742783e-01 2.95638114e-01 8.54090929e-01
-2.96115160e-01 2.59558946e-01 -2.88878202e-01 3.65871638e-01
4.44643945e-01 7.84499571e-02 7.58714795e-01 -1.54235494e+00
-6.63885176e-02 2.00841695e-01 -5.52510321e-01 -7.97693372e-01
-3.08181137e-01 -4.47543889e-01 2.81197816e-01 -1.68635213e+00
1.66435525e-01 -6.78071856e-01 -5.37038505e-01 4.30203319e-01
-3.16034138e-01 3.25394392e-01 -1.21732764e-01 1.02063678e-01
-4.84220892e-01 6.67984724e-01 1.86057031e+00 -4.76287067e-01
-1.06616579e-01 1.22367129e-01 -5.14949977e-01 6.50198817e-01
1.00812399e+00 -2.27811828e-01 -5.90401292e-01 -3.31162035e-01
-1.30160049e-01 -1.35678098e-01 2.47013435e-01 -1.32420743e+00
-5.80027059e-04 -1.90431908e-01 8.07560980e-01 -5.26774943e-01
3.27409595e-01 -9.00670648e-01 -5.51595800e-02 5.79097509e-01
2.03280389e-01 -8.91704410e-02 3.51433575e-01 6.63873136e-01
-4.25413966e-01 -4.03172493e-01 6.02597058e-01 -3.78122509e-01
-1.07193184e+00 2.38607183e-01 -7.23500997e-02 -1.34840786e-01
1.35442448e+00 -3.83179039e-01 -1.39345348e-01 1.07106112e-01
-8.48089695e-01 1.14511363e-01 4.98622805e-01 5.39573312e-01
9.88086760e-01 -1.33428252e+00 -4.29823637e-01 4.99514103e-01
3.28096449e-01 3.90000880e-01 6.06413364e-01 6.27508163e-01
-4.25787210e-01 1.26921296e-01 -4.81577277e-01 -7.16838539e-01
-1.15867996e+00 6.91619396e-01 -2.05818444e-01 -2.46771395e-01
-4.38487083e-01 7.54230559e-01 9.16240290e-02 -2.44110733e-01
-1.86737254e-02 -3.82211447e-01 -9.76470932e-02 -1.67781085e-01
1.88839048e-01 6.55163050e-01 4.06732380e-01 -2.01008886e-01
-8.39223266e-02 8.43975604e-01 -2.13577732e-01 3.49752188e-01
1.21764612e+00 -1.71826221e-02 2.22808614e-01 4.61058676e-01
8.52303267e-01 9.79829058e-02 -1.32878947e+00 -1.18868634e-01
-1.55435130e-01 -7.69835055e-01 5.72339445e-02 -1.03213203e+00
-1.47112942e+00 1.08011019e+00 9.78042006e-01 2.28217188e-02
1.29191375e+00 -9.44211781e-02 7.94011652e-01 2.04712108e-01
5.48578203e-01 -1.32154131e+00 7.23186076e-01 6.67547882e-02
9.14873302e-01 -1.49738264e+00 1.50326878e-01 -7.59087920e-01
-6.84597433e-01 1.00370562e+00 1.12750626e+00 1.22613721e-01
4.68095332e-01 2.16706902e-01 2.70537376e-01 -4.52614397e-01
-4.44181651e-01 7.23018795e-02 1.58715025e-01 8.14059794e-01
1.53609082e-01 -2.53152966e-01 -3.55911665e-02 5.47846496e-01
2.83786267e-01 2.73941997e-02 2.05096409e-01 1.35572565e+00
-7.36807704e-01 -1.02688897e+00 -4.20567751e-01 8.32865298e-01
-3.14208597e-01 3.11093241e-01 1.24495722e-01 7.96828985e-01
5.04809439e-01 1.17324460e+00 -6.89984858e-02 -4.95895982e-01
6.23273611e-01 -2.42412120e-01 5.63662052e-01 -9.25588608e-01
-1.16675198e-01 7.15309307e-02 -2.36754511e-02 -3.78711849e-01
-2.80564368e-01 -4.24912125e-01 -1.30761361e+00 1.42318504e-02
-8.00067723e-01 3.61810401e-02 6.30588889e-01 8.87099087e-01
2.02761412e-01 9.92516160e-01 8.40385497e-01 -6.90454662e-01
-4.39169884e-01 -1.13877583e+00 -8.17413807e-01 5.86376607e-01
1.19076937e-01 -1.02164626e+00 -1.83147341e-01 3.13819796e-01]
|
[7.392986297607422, 1.917964220046997]
|
58626b85-a44f-457c-99dd-02409f97eb82
|
the-dots-have-their-values-exploiting-the
| null | null |
https://aclanthology.org/2020.findings-emnlp.409
|
https://aclanthology.org/2020.findings-emnlp.409.pdf
|
The Dots Have Their Values: Exploiting the Node-Edge Connections in Graph-based Neural Models for Document-level Relation Extraction
|
The goal of Document-level Relation Extraction (DRE) is to recognize the relations between entity mentions that can span beyond sentence boundary. The current state-of-the-art method for this problem has involved the graph-based edge-oriented model where the entity mentions, entities, and sentences in the documents are used as the nodes of the document graphs for representation learning. However, this model does not capture the representations for the nodes in the graphs, thus preventing it from effectively encoding the specific and relevant information of the nodes for DRE. To address this issue, we propose to explicitly compute the representations for the nodes in the graph-based edge-oriented model for DRE. These node representations allow us to introduce two novel representation regularization mechanisms to improve the representation vectors for DRE. The experiments show that our model achieves state-of-the-art performance on two benchmark datasets.
|
['Thien Huu Nguyen', 'Minh Trung Nguyen', 'Hieu Minh Tran']
|
2020-11-01
| null | null | null |
findings-of-the-association-for-computational
|
['document-level-relation-extraction']
|
['natural-language-processing']
|
[ 5.30414470e-02 4.62296993e-01 -4.85173345e-01 -4.29497331e-01
-4.37583566e-01 -3.08550566e-01 6.13597989e-01 7.04198360e-01
-3.22398990e-02 3.60795438e-01 4.94382799e-01 -4.17584032e-01
-1.12362340e-01 -1.15553296e+00 -5.03179610e-01 -4.12709743e-01
-3.79488647e-01 1.54582754e-01 2.28636459e-01 -2.91487157e-01
-1.22134574e-02 3.84889483e-01 -1.08331180e+00 4.38709229e-01
9.80094731e-01 8.93451750e-01 -1.33583248e-01 2.24068716e-01
-5.54613829e-01 1.03109980e+00 -6.18209898e-01 -5.55612087e-01
-1.35327399e-01 -2.95508772e-01 -9.12038743e-01 2.80806720e-02
2.64529973e-01 4.55061868e-02 -9.75651622e-01 1.09011459e+00
9.10850614e-03 3.31511885e-01 9.72656608e-01 -1.11422443e+00
-1.00668883e+00 8.13015223e-01 -9.17733252e-01 4.71798062e-01
3.42052639e-01 -5.81074715e-01 1.78919876e+00 -8.06276917e-01
7.73991644e-01 1.33397722e+00 3.38900298e-01 3.31943035e-01
-8.68446887e-01 -3.80897015e-01 6.40680969e-01 3.65898401e-01
-1.51713240e+00 -2.38797352e-01 9.92886126e-01 -3.84659022e-01
1.29710984e+00 2.15599149e-01 2.57513613e-01 8.46462607e-01
3.59055921e-02 8.02115798e-01 4.61935371e-01 -4.95275140e-01
-6.42702356e-02 -1.72493868e-02 1.02621257e+00 8.45670879e-01
6.90993667e-01 -2.93127209e-01 -3.29532921e-01 -1.49200037e-01
6.07205749e-01 -1.47150934e-01 -3.77378315e-01 -2.92823076e-01
-5.04295886e-01 9.51113939e-01 8.37533355e-01 5.32966375e-01
-4.61611301e-01 -7.50560313e-02 5.03792703e-01 -6.07227795e-02
7.30682194e-01 3.75014573e-01 -5.06839931e-01 4.47752267e-01
-5.24333775e-01 1.53880604e-02 7.08277047e-01 1.02256024e+00
7.03243971e-01 -1.58389106e-01 -6.15829289e-01 7.72241652e-01
5.24353623e-01 -2.43356213e-01 2.18237817e-01 -1.93544939e-01
1.16224563e+00 1.31790984e+00 -3.23969603e-01 -1.42459297e+00
-4.45867807e-01 -6.94718838e-01 -8.33633482e-01 -5.76726139e-01
5.45424223e-02 -1.96910128e-01 -1.00070751e+00 1.56481242e+00
4.00361061e-01 2.36416981e-01 2.96996206e-01 5.68162024e-01
1.42344940e+00 8.83822381e-01 2.70499378e-01 -1.55288070e-01
1.65783250e+00 -1.04279447e+00 -9.47567344e-01 -5.36512852e-01
1.01667440e+00 -1.21848859e-01 7.27657020e-01 -4.83862370e-01
-5.48769414e-01 -2.12372959e-01 -1.20515525e+00 -3.17391187e-01
-5.21192074e-01 3.04112434e-01 8.87685299e-01 2.19945326e-01
-4.54617947e-01 4.86786574e-01 -8.00441444e-01 -1.23310141e-01
3.81745487e-01 1.56094179e-01 -4.46111232e-01 -2.26236492e-01
-1.55251610e+00 1.03752840e+00 5.62913775e-01 3.25269729e-01
-3.17056000e-01 -4.61979002e-01 -1.48652327e+00 6.75311089e-01
4.88764822e-01 -5.36245644e-01 7.22235203e-01 -4.55508471e-01
-5.69367468e-01 8.82485449e-01 -4.79179412e-01 -2.95004159e-01
-4.05146897e-01 -2.01860785e-01 -5.92462361e-01 1.45232052e-01
-5.00930250e-02 9.03069135e-03 4.32220191e-01 -1.49737132e+00
-2.83256412e-01 -4.87948269e-01 3.19820493e-01 1.36443704e-01
-5.41467369e-01 1.78323910e-01 -7.71196365e-01 -7.03684747e-01
3.33216965e-01 -6.64296448e-01 -2.96929181e-01 -5.63472450e-01
-8.31204593e-01 -7.36000597e-01 6.54998481e-01 -8.31781983e-01
1.73868740e+00 -2.12885094e+00 5.83213195e-02 3.25504780e-01
3.42951566e-01 2.75984496e-01 -3.29646796e-01 5.99215567e-01
-2.86472321e-01 3.65080416e-01 -9.92046744e-02 -3.29397112e-01
-1.38551325e-01 2.61899382e-01 -2.56134033e-01 2.16762498e-01
5.39980710e-01 9.84283268e-01 -9.29323614e-01 -5.78862667e-01
-3.20646048e-01 7.27084398e-01 -2.51645982e-01 3.60412717e-01
-5.20089306e-02 3.54235843e-02 -9.29961979e-01 3.69801730e-01
5.59106648e-01 -5.42136014e-01 6.47282839e-01 -4.51061726e-01
5.09946883e-01 1.06484258e+00 -9.72885132e-01 1.26630056e+00
-2.33453125e-01 4.91647780e-01 -3.30984324e-01 -1.10040140e+00
1.18558741e+00 1.81913659e-01 4.49379236e-01 -4.62294668e-01
1.04942331e-02 -2.61479229e-01 6.13328144e-02 -4.56182808e-01
5.31264901e-01 1.52117983e-01 -1.30110070e-01 9.67287943e-02
-3.17012449e-03 3.89370918e-01 4.73326534e-01 6.09543383e-01
1.15497208e+00 -1.59544513e-01 3.56979579e-01 5.62780686e-02
6.12483442e-01 -2.65327483e-01 8.48784864e-01 2.82533139e-01
3.86474252e-01 4.58868742e-01 9.30266142e-01 -1.47358075e-01
-4.41422164e-01 -7.41197228e-01 8.27178061e-02 7.92261958e-01
-4.82828766e-02 -1.09288931e+00 -5.04804909e-01 -1.16035235e+00
-2.04058997e-02 8.73089194e-01 -7.21339822e-01 -5.14949143e-01
-7.37265825e-01 -7.02746809e-01 1.33226410e-01 8.46324563e-01
2.40670457e-01 -7.59066522e-01 7.46058449e-02 1.90106511e-01
-3.58263671e-01 -1.40169179e+00 -3.84322345e-01 2.67077208e-01
-8.35223615e-01 -1.29896295e+00 -2.41825014e-01 -9.62500989e-01
1.04394209e+00 1.47175401e-01 1.24177980e+00 3.88788968e-01
-1.97755024e-02 5.21280095e-02 -5.70869684e-01 1.67909544e-02
-6.70367032e-02 3.62641722e-01 -3.66060585e-01 6.00482263e-02
6.15276992e-01 -3.76195848e-01 -2.61641473e-01 -1.02369986e-01
-7.43862867e-01 -2.40763232e-01 3.49695385e-01 7.08340228e-01
6.46710873e-01 3.00782830e-01 6.16684139e-01 -1.45473421e+00
8.53793025e-01 -5.74297726e-01 -3.46217066e-01 7.35004127e-01
-5.80110133e-01 2.93877244e-01 5.50709367e-01 -3.07948411e-01
-8.82697105e-01 -2.40952559e-02 -6.27142414e-02 -1.55868724e-01
2.49815241e-01 1.24604237e+00 -4.17242616e-01 3.90529662e-01
3.53931725e-01 -3.76695022e-02 -8.00491869e-01 -5.98798096e-01
4.23757434e-01 5.47551572e-01 2.52896160e-01 -5.79293072e-01
6.89862967e-01 -1.87476173e-01 1.67516306e-01 -6.98240757e-01
-1.28144908e+00 -6.43014729e-01 -6.66581154e-01 3.70099872e-01
8.51211190e-01 -9.77412820e-01 -2.83015013e-01 -1.98402256e-01
-1.34684134e+00 2.43644163e-01 -1.90657929e-01 2.83251107e-01
1.10457152e-01 4.97211546e-01 -8.46455753e-01 -8.88301611e-01
-3.77401531e-01 -8.91151011e-01 8.57938945e-01 3.22031379e-01
-2.06611976e-01 -1.11784279e+00 -1.46484859e-02 2.72376031e-01
-3.32562625e-02 2.80367821e-01 1.56834149e+00 -9.61356103e-01
-4.50901836e-01 -5.26571453e-01 -5.20922840e-01 2.09414601e-01
2.76599288e-01 3.12985666e-02 -7.95863092e-01 -1.67116866e-01
-2.73910820e-01 5.59159219e-02 1.23719013e+00 1.05912447e-01
1.16310334e+00 -4.64977354e-01 -6.56613052e-01 5.21306574e-01
1.26510727e+00 1.39055960e-02 6.95524156e-01 6.44062161e-02
1.00373363e+00 7.64355361e-01 5.56330979e-01 2.24226743e-01
7.67250240e-01 7.59254038e-01 3.34626377e-01 2.24258262e-03
-1.08743131e-01 -6.74865723e-01 2.19472602e-01 9.74993229e-01
-9.05303955e-02 -5.52244306e-01 -9.27504659e-01 6.04848981e-01
-2.02193165e+00 -6.21737421e-01 -5.63263357e-01 1.82112682e+00
7.84138739e-01 2.20209226e-01 -2.81243473e-01 6.15944751e-02
7.54915714e-01 5.70119798e-01 -2.03325287e-01 -3.14610571e-01
5.95706031e-02 8.07722807e-02 2.19070092e-01 2.06375420e-01
-1.38741803e+00 1.05084062e+00 6.01390076e+00 4.22220677e-01
-6.06263816e-01 -1.80261225e-01 4.29997563e-01 4.72903490e-01
-4.74638820e-01 2.08348483e-01 -1.18708813e+00 9.30921286e-02
9.37952876e-01 -3.21338654e-01 -8.89713168e-02 7.49020159e-01
-8.17089006e-02 2.52651244e-01 -1.17753446e+00 6.23554289e-01
1.32927433e-01 -1.28571463e+00 2.80391127e-01 5.37804738e-02
5.96217811e-01 -2.80394346e-01 -2.25335792e-01 5.58196664e-01
2.71673262e-01 -1.04884541e+00 4.70570922e-02 3.23288828e-01
6.09952033e-01 -7.97478437e-01 8.86734843e-01 2.15882033e-01
-1.72408545e+00 6.78484663e-02 -6.05003595e-01 3.56771387e-02
6.71829805e-02 7.14000642e-01 -7.43716300e-01 9.75733876e-01
2.97099739e-01 8.82744849e-01 -7.38029242e-01 7.87721097e-01
-8.17647934e-01 7.24059701e-01 1.24660339e-02 -3.90293635e-02
2.45894492e-01 -2.59726584e-01 2.60215372e-01 1.44383979e+00
1.15413003e-01 2.49662608e-01 2.07776695e-01 7.90384889e-01
-5.91417074e-01 3.31735641e-01 -7.08862960e-01 -3.82095158e-01
5.16055346e-01 1.21900594e+00 -5.61160803e-01 -2.12668136e-01
-7.49700069e-01 6.44176722e-01 9.61021185e-01 5.82400739e-01
-4.87977713e-01 -6.44052565e-01 6.88020349e-01 8.67077336e-02
5.06022155e-01 -4.62248802e-01 -2.44187042e-01 -1.34402311e+00
1.76406130e-01 -5.56278169e-01 8.45531166e-01 -3.80966008e-01
-1.59303713e+00 6.40228212e-01 2.29646992e-02 -8.17068875e-01
-1.72989696e-01 -4.86408710e-01 -7.63226449e-01 9.71507192e-01
-1.62926888e+00 -1.15785336e+00 -1.24758631e-02 4.44991678e-01
1.34837404e-01 -6.64429367e-02 9.47386861e-01 2.57811010e-01
-9.64889526e-01 7.88369834e-01 -1.62678793e-01 9.37116802e-01
6.62221014e-02 -1.13087046e+00 7.52936304e-01 1.00192559e+00
5.72146058e-01 9.99286771e-01 3.40050220e-01 -7.94199228e-01
-1.26393545e+00 -1.13715589e+00 1.47050333e+00 -2.03633174e-01
7.01049805e-01 -4.92137253e-01 -1.36008036e+00 9.43685472e-01
-1.71937328e-02 3.39640707e-01 7.24756300e-01 7.21076369e-01
-7.59473979e-01 2.21799128e-02 -7.48712838e-01 4.11230832e-01
9.75282907e-01 -7.96885312e-01 -8.29946280e-01 4.14256245e-01
7.89209843e-01 -3.00354332e-01 -8.98265719e-01 3.98090065e-01
-6.59386814e-02 -1.13888808e-01 1.11702132e+00 -1.12041497e+00
6.40640020e-01 -1.26714259e-01 2.39891466e-02 -1.40479493e+00
-4.29934591e-01 -2.15810031e-01 -8.94452393e-01 1.78616893e+00
8.09038222e-01 -4.02165413e-01 6.88212335e-01 6.29102468e-01
-5.65494597e-02 -9.58256483e-01 -7.46539712e-01 -7.46655941e-01
1.08554915e-01 -6.58792928e-02 6.40921414e-01 1.06366313e+00
3.62386346e-01 1.11264479e+00 -7.18008652e-02 4.21320766e-01
3.14576596e-01 3.45222920e-01 2.58981198e-01 -1.37979984e+00
-1.05067477e-01 -2.23165452e-01 -4.50647593e-01 -1.30486357e+00
6.18988633e-01 -1.36547315e+00 -1.66623935e-01 -2.29667330e+00
3.54026020e-01 -4.66639102e-01 -5.06236613e-01 6.62607312e-01
-6.91660643e-01 -5.50083876e-01 1.51120961e-01 4.20056097e-02
-3.87978286e-01 6.45952165e-01 1.27882433e+00 -4.81435061e-01
-9.55293626e-02 -2.64358707e-02 -1.07925797e+00 5.57476342e-01
5.48972249e-01 -6.67153060e-01 -5.33134282e-01 -3.75529677e-01
1.83232248e-01 -9.66408998e-02 -2.73075774e-02 -3.64357769e-01
2.76526988e-01 -1.09057128e-01 2.14822620e-01 -5.40264845e-01
2.82269895e-01 -7.54520178e-01 -2.71080017e-01 -1.23493131e-02
-5.68046868e-01 -6.98310062e-02 -1.34480834e-01 7.15008080e-01
-5.03652751e-01 -4.32049364e-01 4.37468201e-01 5.34059964e-02
-5.77018678e-01 4.55163330e-01 2.14599054e-02 3.48935694e-01
8.57289433e-01 3.01945418e-01 -7.06262767e-01 -2.98315585e-01
-6.03447974e-01 4.37356442e-01 1.06132114e-02 5.12075245e-01
7.75276899e-01 -1.40866947e+00 -7.08539963e-01 -2.49914750e-02
3.44569594e-01 2.59657651e-01 -4.41628322e-02 4.05085862e-01
-5.61079495e-02 3.69912982e-01 3.53507191e-01 3.85854505e-02
-1.41114879e+00 8.02176833e-01 2.29444146e-01 -1.01168931e+00
-8.41385603e-01 8.24780166e-01 3.33057135e-01 -1.35337412e-01
2.92205840e-01 -3.59257311e-01 -9.81985033e-01 1.10042766e-01
4.34302509e-01 4.82252203e-02 1.78391442e-01 -8.42675388e-01
-6.08178139e-01 3.15293789e-01 -4.87005472e-01 3.95932585e-01
1.50451279e+00 7.89066181e-02 -2.55586743e-01 3.33481640e-01
1.33142436e+00 -1.00865297e-01 -8.21728826e-01 -4.37739074e-01
6.51271462e-01 -2.70983696e-01 2.34023839e-01 -4.46655691e-01
-1.27715504e+00 9.75350976e-01 -1.58977553e-01 2.42541581e-01
8.12929928e-01 3.08805615e-01 8.33202600e-01 4.17475611e-01
1.95456445e-01 -8.39365482e-01 -1.85945958e-01 8.69117796e-01
9.04192746e-01 -8.47077608e-01 3.24532837e-01 -1.26670766e+00
-7.66862392e-01 1.10508025e+00 6.65479779e-01 -1.14573598e-01
7.32647419e-01 2.14386329e-01 -2.18379587e-01 -5.33494473e-01
-8.52797627e-01 -4.08130854e-01 8.86121511e-01 7.65169144e-01
8.95383000e-01 7.47257322e-02 -5.05946279e-01 1.04490387e+00
2.30653897e-01 -5.66037059e-01 3.24594885e-01 9.33927894e-01
-2.14265287e-01 -1.28091025e+00 1.65018439e-01 6.09870553e-01
-4.64794189e-01 -2.65150607e-01 -7.22252488e-01 5.97124517e-01
-3.18479508e-01 1.09448230e+00 -8.09795335e-02 -4.04680312e-01
6.63560927e-01 2.62005717e-01 2.84745127e-01 -1.02203417e+00
-6.83356106e-01 -3.32053959e-01 5.92118859e-01 -1.96061894e-01
-2.53947258e-01 -4.19130683e-01 -1.68919253e+00 4.89141382e-02
-6.29525125e-01 4.26069975e-01 2.65590936e-01 9.38299537e-01
4.48483765e-01 9.03457999e-01 6.52729392e-01 -1.76917255e-01
-3.62873435e-01 -1.01135564e+00 -7.02606082e-01 6.00787997e-01
5.88049516e-02 -8.06915462e-01 -1.69016272e-01 -3.69345963e-01]
|
[9.280449867248535, 8.614015579223633]
|
3b83031e-c08e-4e99-bff9-0ecb0d5c91d5
|
lip-listening-mixing-senses-to-understand
|
2207.05692
| null |
https://arxiv.org/abs/2207.05692v1
|
https://arxiv.org/pdf/2207.05692v1.pdf
|
Lip-Listening: Mixing Senses to Understand Lips using Cross Modality Knowledge Distillation for Word-Based Models
|
In this work, we propose a technique to transfer speech recognition capabilities from audio speech recognition systems to visual speech recognizers, where our goal is to utilize audio data during lipreading model training. Impressive progress in the domain of speech recognition has been exhibited by audio and audio-visual systems. Nevertheless, there is still much to be explored with regards to visual speech recognition systems due to the visual ambiguity of some phonemes. To this end, the development of visual speech recognition models is crucial given the instability of audio models. The main contributions of this work are i) building on recent state-of-the-art word-based lipreading models by integrating sequence-level and frame-level Knowledge Distillation (KD) to their systems; ii) leveraging audio data during training visual models, a feat which has not been utilized in prior word-based work; iii) proposing the Gaussian-shaped averaging in frame-level KD, as an efficient technique that aids the model in distilling knowledge at the sequence model encoder. This work proposes a novel and competitive architecture for lip-reading, as we demonstrate a noticeable improvement in performance, setting a new benchmark equals to 88.64% on the LRW dataset.
|
['Hesham M. Eraqi', 'Nourhan Sakr', 'Omar Abugabal', 'Hadeel Mabrouk']
|
2022-06-05
| null | null | null | null |
['lipreading']
|
['computer-vision']
|
[ 3.90809834e-01 1.70302033e-01 -3.51076484e-01 -3.03915180e-02
-9.37610686e-01 -2.94500887e-01 7.00187445e-01 -1.95185691e-01
-4.30305839e-01 5.41567743e-01 3.97279829e-01 -5.92474163e-01
3.68031234e-01 -2.45456155e-02 -7.20521629e-01 -6.85613990e-01
3.12946737e-01 1.64934248e-01 3.83527994e-01 -7.79350102e-02
3.43570620e-01 5.02122104e-01 -1.96182108e+00 5.95860898e-01
4.86717224e-01 9.74538922e-01 4.48924065e-01 1.32820380e+00
-3.53636861e-01 9.02403414e-01 -6.21518373e-01 -2.97686070e-01
-1.56504616e-01 -4.24600661e-01 -7.56147325e-01 1.07651219e-01
5.78074396e-01 -2.88055867e-01 -4.98345047e-01 7.35688508e-01
1.02445531e+00 1.21187046e-01 8.44003856e-01 -1.16410935e+00
-8.66597891e-01 3.92211556e-01 -2.85440505e-01 3.38539720e-01
4.59795058e-01 4.25663888e-01 8.48366737e-01 -1.35156000e+00
3.13068122e-01 1.33358908e+00 3.28811109e-01 9.62323666e-01
-9.53638792e-01 -4.50780660e-01 4.88407072e-03 1.00386298e+00
-1.52516675e+00 -1.35881555e+00 8.76143456e-01 -2.94114023e-01
1.55043685e+00 1.66760296e-01 6.54253960e-01 1.22259569e+00
-2.41666660e-01 1.23781812e+00 9.47423875e-01 -9.96081889e-01
1.84624240e-01 4.02033597e-01 -2.21890330e-01 4.76240277e-01
-4.70207959e-01 3.93680632e-01 -1.23126256e+00 3.44663531e-01
3.17315936e-01 -7.43840933e-01 -4.18008327e-01 -2.91881651e-01
-8.02704275e-01 6.22159481e-01 -3.36740240e-02 2.69198149e-01
-1.62854895e-01 1.36321947e-01 5.67431211e-01 3.09998989e-01
4.43191081e-01 -2.57167816e-01 -2.07587421e-01 -5.40305912e-01
-1.41452456e+00 -1.73858523e-01 6.84784412e-01 8.06451559e-01
2.36397982e-01 5.47994912e-01 -4.05133635e-01 1.24505937e+00
8.54368746e-01 6.10721886e-01 8.27450395e-01 -6.26927078e-01
4.42737490e-01 -1.18791303e-02 -2.04083949e-01 -3.31512898e-01
6.68783262e-02 -1.28989011e-01 -4.31390554e-01 4.37591732e-01
2.56559879e-01 3.64479691e-01 -1.19107282e+00 1.48376429e+00
2.97979806e-02 5.36635756e-01 4.21289086e-01 7.28082299e-01
1.15579319e+00 6.72195315e-01 1.33634120e-01 -2.78514832e-01
1.37117028e+00 -1.08915353e+00 -9.51795995e-01 -1.97525192e-02
2.00927049e-01 -1.06275773e+00 1.15263557e+00 4.67562526e-01
-1.52012146e+00 -8.92128408e-01 -1.14885211e+00 -2.33900800e-01
-4.85105813e-01 3.71027172e-01 6.31247759e-02 1.06560397e+00
-1.68475819e+00 -1.29065990e-01 -4.50777292e-01 -4.48148280e-01
4.33346063e-01 2.35621363e-01 -1.70334086e-01 6.61946610e-02
-1.22075677e+00 1.05346382e+00 3.14920187e-01 -8.64457786e-02
-1.14677775e+00 -6.56082153e-01 -9.77683365e-01 1.45435646e-01
2.86909223e-01 -5.66655695e-01 1.53653252e+00 -7.79322207e-01
-2.04333329e+00 1.00622213e+00 -6.28297210e-01 -6.96341276e-01
5.11143088e-01 9.28800553e-02 -6.78503394e-01 4.37422514e-01
-5.33335865e-01 1.11197615e+00 1.43334234e+00 -1.32359505e+00
-4.23642367e-01 3.63553204e-02 -3.44459683e-01 2.65494943e-01
-2.69649088e-01 2.50093788e-01 -6.37680352e-01 -7.25827873e-01
-3.88260335e-01 -5.13763130e-01 6.10861778e-01 1.80588752e-01
-2.25190237e-01 -4.95895833e-01 8.96382630e-01 -8.91280115e-01
1.21993887e+00 -2.26035619e+00 2.39713024e-02 -1.99480265e-01
-6.62935004e-02 1.08028996e+00 -2.88440615e-01 5.30604124e-01
-1.08013675e-01 -1.21660409e-02 -6.06563315e-02 -9.08205509e-01
2.35648766e-01 1.16197273e-01 -6.48352504e-01 1.38694078e-01
4.02947813e-01 1.13180017e+00 -4.84422028e-01 -7.60249555e-01
6.88742280e-01 9.01975811e-01 -4.07765895e-01 2.59246290e-01
-1.25478759e-01 -9.16113183e-02 3.35517228e-01 8.13912451e-01
6.11169994e-01 1.78049624e-01 -2.43915841e-01 -9.89082363e-03
-1.38348132e-01 5.18003941e-01 -8.74515831e-01 1.73217273e+00
-4.28810418e-01 1.13661027e+00 1.28767058e-01 -1.02958596e+00
9.02433693e-01 9.11611617e-01 1.78688079e-01 -7.18890071e-01
1.59461331e-02 1.52797878e-01 -1.57412186e-01 -7.09809363e-01
4.96531904e-01 -2.69853354e-01 7.03519046e-01 7.98142701e-02
4.31575567e-01 -3.25660408e-01 -1.88023299e-02 2.65472494e-02
5.14592648e-01 -4.25725803e-02 2.51409113e-01 2.04507172e-01
1.06926131e+00 -4.45326030e-01 -2.43034825e-01 6.39537215e-01
-6.05382919e-01 6.90304160e-01 1.78476065e-01 1.69237405e-01
-9.50128317e-01 -1.27615452e+00 -1.64838389e-01 1.18111336e+00
-2.78908819e-01 -2.68590242e-01 -9.01660442e-01 -3.75535548e-01
-1.24897890e-01 7.19141126e-01 -3.59270394e-01 -2.22926170e-01
-2.65032083e-01 -3.29964720e-02 1.14841270e+00 5.90173364e-01
3.31978947e-01 -1.21100116e+00 -2.65044212e-01 1.49982303e-01
-1.03893369e-01 -1.24433434e+00 -5.95369518e-01 -2.43395176e-02
-4.54295427e-01 -7.71094620e-01 -1.40741611e+00 -1.03456295e+00
2.87399478e-02 2.89362580e-01 8.12490821e-01 -2.07427144e-01
-3.74312788e-01 8.50295365e-01 -3.65031183e-01 -7.01639712e-01
-7.94101536e-01 -8.20954666e-02 2.05779567e-01 -1.37609886e-02
6.03435397e-01 -2.86789805e-01 -2.93540835e-01 2.34714866e-01
-8.67991269e-01 -1.17583729e-01 6.06667161e-01 8.97102296e-01
3.44483733e-01 -5.11900842e-01 8.20994198e-01 2.97809273e-01
7.78363228e-01 -9.80883986e-02 -3.04147840e-01 4.24419552e-01
-4.15829480e-01 -5.69692701e-02 2.64427543e-01 -6.91084266e-01
-9.32079434e-01 -3.89874838e-02 -6.60516858e-01 -7.45302916e-01
-3.87110233e-01 1.45607173e-01 -1.61108404e-01 -2.93861032e-01
4.14885789e-01 7.25834310e-01 4.85293120e-01 -6.18890643e-01
5.85180998e-01 1.27727830e+00 7.49671519e-01 -2.42425963e-01
3.25699300e-01 1.48487687e-01 -2.46545061e-01 -1.47019160e+00
-9.12262592e-03 -8.10457051e-01 -4.91745293e-01 -3.15710574e-01
7.29712844e-01 -9.61388350e-01 -9.66732681e-01 7.74793744e-01
-1.31368840e+00 -3.04251760e-01 -4.13564414e-01 5.09010673e-01
-9.28176284e-01 7.74974525e-01 -3.33947808e-01 -1.16654384e+00
-1.26193300e-01 -1.36516964e+00 1.11665511e+00 7.13087097e-02
-2.22530030e-02 -8.75457346e-01 1.67330772e-01 6.74054086e-01
6.34796798e-01 -8.64519417e-01 7.82347381e-01 -6.13743603e-01
-3.97938669e-01 2.40759373e-01 -3.10072541e-01 8.03468645e-01
-7.11130723e-02 -8.97794683e-03 -1.66959512e+00 -2.14715794e-01
-2.50629604e-01 -6.72213137e-01 1.11991227e+00 5.33216655e-01
9.12482142e-01 -4.62391265e-02 -1.48742601e-01 3.71788353e-01
7.62978137e-01 3.30070436e-01 8.27793300e-01 -1.58085957e-01
2.80096710e-01 6.44262075e-01 3.46304804e-01 2.43863121e-01
4.03055578e-01 1.20421922e+00 2.53579438e-01 -1.68147579e-01
-1.21726286e+00 -5.61295688e-01 7.40470827e-01 9.96209025e-01
5.82903102e-02 -3.58947814e-01 -8.91691387e-01 6.79781675e-01
-1.46254563e+00 -1.08021891e+00 4.17669088e-01 1.97228467e+00
9.27579939e-01 -6.26494139e-02 3.17033321e-01 6.20097339e-01
6.68127000e-01 2.88132966e-01 -4.39568460e-01 -7.20505238e-01
-2.19970271e-01 3.77032161e-01 1.03636116e-01 1.01763237e+00
-8.32159460e-01 1.20931733e+00 6.52872515e+00 1.33672523e+00
-1.43969083e+00 1.80058226e-01 1.91457987e-01 -1.07387312e-01
1.84487682e-02 -5.28108358e-01 -1.01732707e+00 3.86818260e-01
1.21062958e+00 6.56840578e-02 4.42878962e-01 6.13079786e-01
4.38696384e-01 -9.16968584e-02 -8.94318759e-01 1.40281272e+00
7.40533650e-01 -1.19173229e+00 1.82581931e-01 -1.18380278e-01
1.51726469e-01 3.90180200e-02 5.39746940e-01 3.83744955e-01
-2.97187656e-01 -1.38602638e+00 8.64403069e-01 4.73308146e-01
1.19247520e+00 -5.96493721e-01 2.83045739e-01 1.66622952e-01
-1.42123044e+00 9.00460407e-02 -2.86993772e-01 2.99778461e-01
2.15136766e-01 -3.64032835e-02 -1.70117533e+00 2.21549898e-01
4.52535361e-01 4.44690853e-01 -4.26127851e-01 1.21292901e+00
-3.07483733e-01 8.51985514e-01 -1.37634695e-01 -1.55827776e-01
4.72624153e-02 5.50748050e-01 6.38799429e-01 1.58524704e+00
9.80607271e-02 -4.03671890e-01 -3.05675775e-01 6.61698580e-01
4.94644828e-02 1.97449103e-01 -6.97673023e-01 -7.49939233e-02
3.45986396e-01 6.02523804e-01 -1.34242490e-01 -2.31069833e-01
-5.19871652e-01 8.77684832e-01 5.86838499e-02 5.92106521e-01
-7.44237244e-01 -3.33011299e-01 7.05120802e-01 2.59688105e-02
5.62673092e-01 -3.72957915e-01 -8.68140757e-02 -8.46668124e-01
1.10156476e-01 -9.69638705e-01 -4.97561321e-02 -1.06879807e+00
-7.93063164e-01 5.20292580e-01 -1.51198164e-01 -1.04554033e+00
-6.60559654e-01 -9.01977301e-01 -2.32274219e-01 1.05630028e+00
-2.21510983e+00 -1.27216291e+00 9.13509578e-02 8.69452894e-01
1.08220005e+00 -5.83674550e-01 7.60362864e-01 3.24149102e-01
1.10843740e-02 1.12876403e+00 -7.26605132e-02 -8.53568986e-02
9.40813601e-01 -8.49679172e-01 3.96899581e-01 6.66555464e-01
5.82338810e-01 2.06260711e-01 5.61243951e-01 -2.84081459e-01
-1.26013362e+00 -5.70764840e-01 1.32588291e+00 -2.99000680e-01
5.14493644e-01 -4.67189431e-01 -9.28572536e-01 1.63819388e-01
4.53940004e-01 -5.08463420e-02 7.42944241e-01 -2.88240075e-01
-4.66213286e-01 -3.20798270e-02 -1.04220259e+00 5.03042758e-01
8.19273293e-01 -1.23869812e+00 -8.55409086e-01 -7.06922486e-02
7.89560437e-01 -1.31053090e-01 -4.16372150e-01 2.64272809e-01
6.56963468e-01 -8.51222575e-01 1.37050712e+00 -5.77425003e-01
-4.89750504e-02 -3.15641731e-01 -2.69351035e-01 -1.24072516e+00
3.06436360e-01 -7.46297956e-01 -4.99532938e-01 1.38073480e+00
3.70467782e-01 -4.02585834e-01 7.27730751e-01 7.28567271e-03
-2.95182437e-01 -5.94312012e-01 -1.51433241e+00 -9.07523155e-01
8.85639042e-02 -9.21012938e-01 2.12961316e-01 4.53974277e-01
2.70060897e-01 7.23207667e-02 -6.89180434e-01 -9.41274911e-02
2.44831294e-01 -6.53358877e-01 6.47327721e-01 -7.15060830e-01
-4.57353294e-01 -6.12740636e-01 -6.62474096e-01 -1.51956451e+00
3.77588511e-01 -8.11098158e-01 1.95965216e-01 -1.32253158e+00
-1.43840745e-01 1.60594389e-01 -2.13357329e-01 4.84833926e-01
6.55317828e-02 2.97725588e-01 5.71899593e-01 3.01304981e-02
-3.70071769e-01 7.68944979e-01 1.10714281e+00 -4.51743931e-01
-1.69092134e-01 8.98817629e-02 -2.15857372e-01 3.95570844e-01
5.73822379e-01 -1.81434713e-02 -6.15165710e-01 -2.66678154e-01
-3.43201488e-01 1.72138870e-01 6.02998793e-01 -1.04337966e+00
6.42728865e-01 3.49402398e-01 5.35935573e-02 -7.78850138e-01
1.06216013e+00 -6.07500911e-01 -4.47888702e-01 2.83418983e-01
-5.26008248e-01 -2.82831520e-01 5.59699714e-01 3.88595879e-01
-5.46789408e-01 -3.23576957e-01 9.55247760e-01 2.93224543e-01
-1.01153231e+00 -1.25417560e-01 -5.95826328e-01 1.64039329e-01
7.97727644e-01 -4.41143870e-01 -3.45953882e-01 -5.85529864e-01
-9.21205997e-01 -2.07993299e-01 2.50694543e-01 6.60333514e-01
1.06834543e+00 -1.06093657e+00 -8.06326270e-01 5.59250772e-01
2.71240473e-01 -6.89350605e-01 2.52293974e-01 8.73756111e-01
-7.37475008e-02 9.59421039e-01 -2.07682729e-01 -7.91255236e-01
-1.69620931e+00 5.64140677e-01 2.98500687e-01 2.62194723e-01
-3.25457424e-01 1.13724554e+00 -1.21747985e-01 5.71789742e-02
1.01952350e+00 -2.91727215e-01 -3.30648303e-01 2.20122755e-01
7.94386268e-01 4.45523262e-01 3.07669342e-01 -8.56296599e-01
-3.42505008e-01 6.57722294e-01 -4.79325242e-02 -6.07977748e-01
6.69041574e-01 -3.45916122e-01 6.03128016e-01 4.86705750e-01
1.10595608e+00 -2.23096795e-02 -1.19387865e+00 -2.08661526e-01
-2.73388684e-01 -3.60402495e-01 1.07995741e-01 -1.03582716e+00
-5.72324455e-01 1.73855150e+00 7.94163883e-01 8.14219788e-02
1.03282201e+00 1.38625279e-01 4.93450254e-01 2.13432625e-01
-2.66484637e-02 -1.19065320e+00 2.72708416e-01 6.52013540e-01
9.65500534e-01 -1.28190851e+00 -6.03385210e-01 -2.38023028e-01
-7.99393117e-01 1.14347899e+00 6.97735325e-02 4.80075359e-01
4.51840103e-01 3.64468366e-01 2.27958620e-01 3.49856883e-01
-8.15164983e-01 -6.22877896e-01 5.93813121e-01 1.13827741e+00
3.95599306e-01 -2.21251890e-01 1.04394265e-01 2.44369984e-01
-3.75815555e-02 3.75322819e-01 1.61697656e-01 5.32303810e-01
-6.82880938e-01 -1.08163726e+00 -4.49523926e-01 -2.80066133e-01
-4.12289947e-01 -6.17985010e-01 -5.01846850e-01 6.65810168e-01
-1.00239903e-01 1.32854831e+00 -1.41612291e-01 -1.88682795e-01
1.40966788e-01 6.75251663e-01 5.17363608e-01 -4.02310431e-01
-3.11701536e-01 7.25700259e-02 9.14355218e-02 -2.87083268e-01
-3.14748883e-01 -4.96734440e-01 -9.53866601e-01 2.94780862e-02
-2.28824556e-01 -1.85645316e-02 1.07754719e+00 8.10130417e-01
3.21056843e-01 5.34714997e-01 2.49737546e-01 -1.03021789e+00
-6.87295675e-01 -9.81671691e-01 -4.27343637e-01 -1.02642901e-01
7.84260213e-01 -7.43482947e-01 -5.17466724e-01 2.29474097e-01]
|
[14.324458122253418, 5.021787166595459]
|
1c9e2410-fe57-4b5f-ba4e-1b6f279810d5
|
nerv-neural-representations-for-videos
|
2110.13903
| null |
https://arxiv.org/abs/2110.13903v1
|
https://arxiv.org/pdf/2110.13903v1.pdf
|
NeRV: Neural Representations for Videos
|
We propose a novel neural representation for videos (NeRV) which encodes videos in neural networks. Unlike conventional representations that treat videos as frame sequences, we represent videos as neural networks taking frame index as input. Given a frame index, NeRV outputs the corresponding RGB image. Video encoding in NeRV is simply fitting a neural network to video frames and decoding process is a simple feedforward operation. As an image-wise implicit representation, NeRV output the whole image and shows great efficiency compared to pixel-wise implicit representation, improving the encoding speed by 25x to 70x, the decoding speed by 38x to 132x, while achieving better video quality. With such a representation, we can treat videos as neural networks, simplifying several video-related tasks. For example, conventional video compression methods are restricted by a long and complex pipeline, specifically designed for the task. In contrast, with NeRV, we can use any neural network compression method as a proxy for video compression, and achieve comparable performance to traditional frame-based video compression approaches (H.264, HEVC \etc). Besides compression, we demonstrate the generalization of NeRV for video denoising. The source code and pre-trained model can be found at https://github.com/haochen-rye/NeRV.git.
|
['Abhinav Shrivastava', 'Ser-Nam Lim', 'Yixuan Ren', 'Hanyu Wang', 'Bo He', 'Hao Chen']
|
2021-10-26
| null |
http://proceedings.neurips.cc/paper/2021/hash/b44182379bf9fae976e6ae5996e13cd8-Abstract.html
|
http://proceedings.neurips.cc/paper/2021/file/b44182379bf9fae976e6ae5996e13cd8-Paper.pdf
|
neurips-2021-12
|
['video-denoising', 'neural-network-compression', 'neural-network-compression']
|
['computer-vision', 'methodology', 'miscellaneous']
|
[ 4.97822464e-01 8.24481994e-02 -1.68379828e-01 -3.16319346e-01
-2.67376125e-01 -1.06792659e-01 2.72442847e-01 -3.87054086e-01
-5.18321812e-01 4.55553085e-01 1.75222337e-01 -2.39745900e-01
2.25818396e-01 -1.05149651e+00 -1.20651066e+00 -6.59967184e-01
-3.07506998e-04 -2.31258914e-01 1.91864781e-02 5.43646924e-02
1.68606147e-01 3.24444592e-01 -1.56244504e+00 4.59272891e-01
2.94364870e-01 1.34107220e+00 4.44573611e-01 8.98156881e-01
5.06424308e-02 1.32766855e+00 -5.41314900e-01 -3.55330139e-01
3.08835357e-01 -4.93582040e-01 -6.37006879e-01 -5.54182902e-02
3.76638979e-01 -8.93996716e-01 -1.26113737e+00 1.21115470e+00
2.23320156e-01 2.58007824e-01 3.61411601e-01 -1.01598406e+00
-9.23348844e-01 6.03168607e-01 -2.50832826e-01 -1.04864404e-01
3.59944165e-01 1.27003610e-01 5.49122095e-01 -6.93491936e-01
5.04847646e-01 1.26032054e+00 6.91772580e-01 7.43261695e-01
-8.71856809e-01 -4.82583493e-01 -1.94074482e-01 5.49233913e-01
-1.29020309e+00 -6.57394052e-01 5.75036883e-01 -7.47901723e-02
1.01532614e+00 3.59269321e-01 1.03414583e+00 9.42476332e-01
1.96978021e-02 8.34581196e-01 2.47288227e-01 -2.76779711e-01
2.23913193e-01 -6.03390098e-01 -1.13604784e-01 5.89325488e-01
2.06585154e-02 5.85797317e-02 -4.55193669e-01 5.20338178e-01
1.25045466e+00 6.77552819e-01 -7.94535160e-01 1.86024934e-01
-9.62982118e-01 5.75889468e-01 7.66733766e-01 8.97256583e-02
-4.89464521e-01 9.64590728e-01 5.86709738e-01 4.68503088e-01
3.29011679e-02 -9.73249376e-02 -9.23970267e-02 -3.58455956e-01
-1.32332313e+00 7.62319416e-02 6.45309448e-01 1.07932830e+00
7.80014932e-01 4.59932715e-01 -1.08070277e-01 6.89445853e-01
1.19177744e-01 3.47689301e-01 7.18031943e-01 -1.74193990e+00
2.31992111e-01 1.38020560e-01 -3.61973912e-01 -8.33566308e-01
6.11794814e-02 -4.42629009e-02 -1.39127803e+00 6.24874830e-02
1.36054782e-02 4.83626276e-02 -1.03900182e+00 1.51426113e+00
-2.47613296e-01 4.64434773e-01 8.80295336e-02 9.85550284e-01
1.01657498e+00 1.17373765e+00 -1.22273080e-01 -3.44146639e-01
1.15231359e+00 -1.19216478e+00 -8.81238878e-01 8.37048143e-02
3.51032197e-01 -4.67469841e-01 7.90993035e-01 6.38714552e-01
-1.55429459e+00 -7.35717773e-01 -1.03743660e+00 -5.81303358e-01
-1.29880086e-01 -1.50795607e-02 4.58593279e-01 1.90620750e-01
-1.44908917e+00 9.64276612e-01 -8.89360070e-01 -1.62520424e-01
4.43161935e-01 5.23701727e-01 -4.34944868e-01 -4.11845714e-01
-1.10406375e+00 4.15130556e-01 7.14889050e-01 2.41899982e-01
-9.73545372e-01 -5.43756485e-01 -1.05339706e+00 4.14759308e-01
2.39177346e-01 -5.66242695e-01 1.38832021e+00 -1.24613118e+00
-1.57000887e+00 3.48615199e-01 -2.73209572e-01 -8.51079583e-01
2.89406598e-01 -4.50968206e-01 -1.66588008e-01 6.52373374e-01
-4.63720769e-01 1.20262218e+00 1.01366580e+00 -9.00007129e-01
-3.09193432e-01 1.25879183e-01 2.75677800e-01 2.84354426e-02
-4.58079070e-01 -7.99917653e-02 -1.04970324e+00 -7.85331488e-01
6.71033561e-02 -4.77200240e-01 -9.64066610e-02 4.45814401e-01
-8.49279389e-02 1.25772208e-01 1.19408381e+00 -1.01251793e+00
1.28071237e+00 -2.30862355e+00 2.63884455e-01 -1.54580593e-01
3.69959891e-01 4.94168609e-01 -2.22061709e-01 1.59394115e-01
-6.78733364e-02 1.42255425e-01 -2.92038828e-01 -3.44379663e-01
-1.83737099e-01 5.42008638e-01 -1.26611695e-01 3.06000292e-01
1.52965367e-01 9.25593078e-01 -6.85410142e-01 -3.32337081e-01
3.94541293e-01 1.04842031e+00 -9.16024446e-01 4.27462995e-01
-1.24438167e-01 -1.44108934e-02 -1.65492091e-02 6.53126121e-01
7.48967290e-01 -2.22774327e-01 2.24207804e-01 -5.67926407e-01
1.44347232e-02 6.30786419e-02 -7.51281977e-01 1.84925079e+00
-3.77404779e-01 1.09943950e+00 2.37867355e-01 -1.34853804e+00
7.26255417e-01 5.53422093e-01 4.74323273e-01 -8.13789010e-01
4.32742476e-01 1.43214660e-02 -2.42741764e-01 -6.74861729e-01
5.68535388e-01 1.57460690e-01 4.00912195e-01 5.18104620e-02
3.49933237e-01 8.03106278e-02 4.38118070e-01 1.68640986e-01
1.26638329e+00 4.54830706e-01 1.35180831e-01 3.58545929e-01
3.97741914e-01 -4.36904341e-01 5.32853544e-01 6.26601338e-01
-1.23162903e-01 8.97831261e-01 4.32133675e-01 -5.20793378e-01
-1.21368456e+00 -9.22025621e-01 2.66864663e-03 8.02497745e-01
9.27205384e-02 -6.96965933e-01 -1.10316825e+00 -3.73321027e-02
-2.85784394e-01 4.30929244e-01 -4.97550488e-01 -2.14269817e-01
-7.77556479e-01 -1.76107571e-01 5.56326568e-01 6.17957294e-01
8.21923614e-01 -1.25581932e+00 -8.26408863e-01 2.07941175e-01
-2.64884800e-01 -1.12011480e+00 -3.43866736e-01 3.72469634e-01
-1.11379158e+00 -9.09861922e-01 -1.01664865e+00 -9.68131423e-01
5.88833392e-01 3.34711790e-01 1.08296180e+00 5.09096503e-01
-1.99752703e-01 1.51356176e-01 -4.91083384e-01 1.25912666e-01
-1.93816692e-01 -5.10675430e-01 -1.22347422e-01 -3.37578088e-01
2.52165765e-01 -7.26125717e-01 -9.33618546e-01 -1.37136042e-01
-1.37638855e+00 3.28462034e-01 5.35948217e-01 8.10189486e-01
6.89893663e-01 3.81450132e-02 1.64198149e-02 -6.10029161e-01
2.68868238e-01 -3.58993977e-01 -3.87439221e-01 6.01551346e-02
-6.97441250e-02 -2.02331692e-01 8.93961787e-01 -4.02849823e-01
-7.57079661e-01 4.76119481e-02 -5.07955194e-01 -1.08248580e+00
-2.26293802e-01 5.49415886e-01 3.73391062e-02 -1.30617265e-02
4.00084436e-01 4.03634012e-01 1.39579222e-01 -4.87416536e-01
4.31442082e-01 5.92265368e-01 9.69770372e-01 -2.84370303e-01
4.44367975e-01 1.94675997e-01 -1.52114451e-01 -9.19838786e-01
-2.26679325e-01 -7.46040419e-02 -5.02081871e-01 -4.73056078e-01
9.75301087e-01 -1.09507620e+00 -7.61418164e-01 2.84926474e-01
-1.30735219e+00 -4.70452726e-01 -4.18503404e-01 5.59334874e-01
-7.41334975e-01 4.88227367e-01 -1.24605882e+00 -4.79751468e-01
-4.26192075e-01 -1.33708096e+00 7.00974405e-01 1.81089953e-01
-3.49585805e-03 -7.94393778e-01 -4.62220311e-01 -6.16975036e-03
5.86798191e-01 1.43725768e-01 6.70533776e-01 9.28112343e-02
-7.43768454e-01 -1.11882783e-01 -4.00115788e-01 7.99857616e-01
-1.59268916e-01 2.03877717e-01 -9.19270217e-01 -2.69320399e-01
2.18767509e-01 -4.58907127e-01 1.15809608e+00 7.44209886e-01
1.90743530e+00 -6.30189359e-01 1.02807559e-01 1.26564026e+00
1.73174250e+00 4.68528152e-01 1.20835876e+00 2.93049634e-01
7.00498700e-01 2.53371596e-01 3.44355963e-02 4.33703631e-01
5.56782521e-02 2.61547685e-01 7.41145074e-01 -1.95796579e-01
-3.85550171e-01 -1.61003068e-01 6.02809250e-01 1.19416761e+00
-5.35019875e-01 -4.29739743e-01 -5.02372265e-01 4.70285639e-02
-1.81182909e+00 -1.16823900e+00 5.31721264e-02 1.89402533e+00
6.40741825e-01 -3.01379617e-02 -2.56893933e-01 3.74629080e-01
6.11121178e-01 9.01775956e-02 -5.09711266e-01 -3.92101645e-01
-3.89081500e-02 3.26058328e-01 5.20264804e-01 2.36008555e-01
-9.45769072e-01 7.39310861e-01 6.29364443e+00 9.29421306e-01
-1.34823847e+00 2.96917520e-02 7.12549925e-01 -4.85620439e-01
-7.23496778e-03 -2.20534340e-01 -1.87900126e-01 4.99310315e-01
1.27348685e+00 3.63082513e-02 8.93452466e-01 7.41924226e-01
3.52311701e-01 5.51078580e-02 -1.18688655e+00 1.53329682e+00
1.07164368e-01 -1.67864633e+00 4.16616946e-01 -9.21917632e-02
4.02278602e-01 -9.13004801e-02 -1.88212302e-02 2.80356407e-01
-2.45382592e-01 -1.21782506e+00 9.95093107e-01 6.49475515e-01
1.18134570e+00 -6.12897038e-01 6.21692121e-01 -5.33985533e-03
-1.23169160e+00 -1.67602152e-01 -7.67167091e-01 -1.83812648e-01
3.95663321e-01 3.69764030e-01 6.31385148e-02 3.45831901e-01
1.11749530e+00 1.19834232e+00 -2.34745696e-01 1.13680863e+00
-9.54712480e-02 6.88428283e-01 -8.79586414e-02 4.73534942e-01
2.66188443e-01 -3.87561500e-01 7.14025199e-02 1.31581259e+00
7.41093278e-01 2.91741818e-01 -2.39789054e-01 6.34448409e-01
-5.40811241e-01 -2.64902502e-01 -6.07940197e-01 -1.84240267e-01
2.54364640e-01 9.63272214e-01 -4.89010960e-01 -6.50217056e-01
-6.63847327e-01 1.27743363e+00 1.14011861e-01 6.66794538e-01
-1.01558387e+00 -5.82138717e-01 5.02823591e-01 5.63447811e-02
5.15234590e-01 -1.44499183e-01 1.31515950e-01 -1.12978685e+00
-8.84260163e-02 -8.18852484e-01 -1.09256446e-01 -1.18605018e+00
-5.98120093e-01 6.24119878e-01 -2.36755252e-01 -1.44554138e+00
-2.23886251e-01 -9.32785988e-01 -4.60283369e-01 4.22354847e-01
-1.40070450e+00 -6.25545263e-01 -5.13204575e-01 6.81626916e-01
7.50111997e-01 -1.01475462e-01 6.45402908e-01 7.12642372e-01
-7.33040214e-01 4.08104986e-01 3.05439293e-01 3.19960177e-01
4.81314540e-01 -7.83246100e-01 2.48629823e-01 9.34553444e-01
-1.71982840e-01 6.60656035e-01 4.52088535e-01 -3.53501558e-01
-1.80094349e+00 -1.30164492e+00 3.71858090e-01 5.38340271e-01
3.61811697e-01 -1.39044255e-01 -1.04507446e+00 8.40938568e-01
6.47861362e-01 9.06996354e-02 4.84146982e-01 -5.93629718e-01
-1.74819469e-01 -1.14675529e-01 -1.05436611e+00 5.48894286e-01
1.16040957e+00 -6.17724597e-01 -1.97487727e-01 2.99770355e-01
1.03091812e+00 -3.92561138e-01 -1.04690969e+00 2.88298190e-01
6.44258320e-01 -1.12567401e+00 1.03375947e+00 -1.13588482e-01
1.12636125e+00 -2.76020169e-01 -5.41463077e-01 -8.50741684e-01
-5.49087644e-01 -3.79811227e-01 -6.83667183e-01 8.79921675e-01
-7.05591664e-02 -1.35669261e-01 8.51779282e-01 3.09955776e-01
-3.28150749e-01 -9.11094606e-01 -6.95526004e-01 -5.61176956e-01
-3.13833743e-01 -6.32486999e-01 5.02819836e-01 5.29642105e-01
-2.71233141e-01 -1.74683332e-01 -6.61207199e-01 -1.00237302e-01
5.20829201e-01 -4.12873894e-01 3.62206727e-01 -6.69601023e-01
-4.42669272e-01 -5.47408223e-01 -5.05726397e-01 -1.64430308e+00
-1.85486535e-03 -7.29677558e-01 1.38345420e-01 -1.70594108e+00
1.15517311e-01 2.67705368e-03 -2.33209997e-01 5.56838512e-01
3.54035199e-01 6.93918407e-01 5.33389568e-01 3.91827971e-01
-5.38723648e-01 6.25828922e-01 1.23912489e+00 -3.09431165e-01
2.29558185e-01 -5.33449054e-01 -3.82865489e-01 7.38706172e-01
7.58801043e-01 -9.09527540e-02 -4.04790252e-01 -9.66507554e-01
-7.30082095e-02 5.03587365e-01 4.39666092e-01 -1.33990109e+00
3.32977235e-01 1.19768627e-01 8.06215167e-01 -4.69975471e-01
6.54048026e-01 -8.91918361e-01 5.37498355e-01 6.97660685e-01
-3.34057957e-01 1.20253697e-01 -1.86790917e-02 3.16296995e-01
-7.48453438e-01 -5.29328942e-01 5.04558384e-01 -4.40713435e-01
-9.94380176e-01 5.23486555e-01 -5.57973862e-01 -4.32829887e-01
7.42286801e-01 -5.72762847e-01 -3.88792187e-01 -6.02504790e-01
-7.88780093e-01 -1.80221558e-01 5.50828934e-01 1.21507347e-01
1.16386199e+00 -1.34547234e+00 -2.77808100e-01 3.72511089e-01
-5.37015498e-01 2.57160425e-01 3.53463471e-01 6.62326157e-01
-1.23591864e+00 3.13056022e-01 -4.59275901e-01 -5.55517793e-01
-1.10059071e+00 6.09989524e-01 3.28526109e-01 3.63041550e-01
-8.33929658e-01 8.04562151e-01 2.39915520e-01 2.20855519e-01
6.39529407e-01 -4.70349222e-01 -1.90535411e-01 -3.17474604e-01
8.13022912e-01 3.37416440e-01 -1.39160678e-01 -4.96059597e-01
1.30569056e-01 4.59993988e-01 2.37066537e-01 1.74870566e-01
1.52814782e+00 -5.21063022e-02 -3.29295456e-01 3.92126702e-02
1.58230209e+00 -6.73605382e-01 -1.58716559e+00 4.78360662e-03
-6.10067308e-01 -4.41956043e-01 2.74983704e-01 -2.59685814e-01
-1.75547051e+00 1.07279122e+00 6.25881135e-01 8.60865340e-02
1.70142508e+00 -4.06549901e-01 9.87702012e-01 4.60090965e-01
1.43957987e-01 -8.71889353e-01 4.98915203e-02 5.51608860e-01
9.38190520e-01 -8.29457581e-01 -2.95511950e-02 -3.50886762e-01
-2.22227231e-01 1.67319465e+00 4.84347701e-01 -3.91131252e-01
5.81747234e-01 4.81535584e-01 -1.64685115e-01 8.73265192e-02
-7.94865310e-01 1.43714845e-01 -1.12189263e-01 4.50399369e-01
6.99009776e-01 -1.09406263e-01 -1.55030236e-01 1.68730631e-01
-2.22488239e-01 5.01715541e-01 5.79781592e-01 9.89023626e-01
-5.54288149e-01 -9.00543213e-01 -3.40298384e-01 6.08237505e-01
-4.26875591e-01 -2.38498360e-01 4.09353733e-01 4.75012749e-01
2.46159017e-01 6.63338602e-01 5.27934015e-01 -5.91248572e-01
4.99544963e-02 -1.57449991e-01 5.02014220e-01 -1.29437923e-01
-3.49579662e-01 2.37719804e-01 -2.49467522e-01 -8.69110346e-01
-6.43121600e-01 -2.63452500e-01 -1.26002204e+00 -6.51953995e-01
2.37064540e-01 -1.48666993e-01 5.64733744e-01 6.16598427e-01
1.17172733e-01 8.73293161e-01 2.16391429e-01 -1.34056914e+00
-2.90124267e-02 -7.87096024e-01 -3.91737998e-01 5.11533737e-01
5.18152297e-01 -2.63520479e-01 -1.22553617e-01 7.20676303e-01]
|
[11.299532890319824, -1.543127417564392]
|
8658c786-f26d-44bc-92e1-f0b32f0b6df5
|
unsupervised-domain-adaptation-for-semantic-3
|
2112.03241
| null |
https://arxiv.org/abs/2112.03241v1
|
https://arxiv.org/pdf/2112.03241v1.pdf
|
Unsupervised Domain Adaptation for Semantic Image Segmentation: a Comprehensive Survey
|
Semantic segmentation plays a fundamental role in a broad variety of computer vision applications, providing key information for the global understanding of an image. Yet, the state-of-the-art models rely on large amount of annotated samples, which are more expensive to obtain than in tasks such as image classification. Since unlabelled data is instead significantly cheaper to obtain, it is not surprising that Unsupervised Domain Adaptation reached a broad success within the semantic segmentation community. This survey is an effort to summarize five years of this incredibly rapidly growing field, which embraces the importance of semantic segmentation itself and a critical need of adapting segmentation models to new environments. We present the most important semantic segmentation methods; we provide a comprehensive survey on domain adaptation techniques for semantic segmentation; we unveil newer trends such as multi-domain learning, domain generalization, test-time adaptation or source-free domain adaptation; we conclude this survey by describing datasets and benchmarks most widely used in semantic segmentation research. We hope that this survey will provide researchers across academia and industry with a comprehensive reference guide and will help them in fostering new research directions in the field.
|
['Boris Chidlovskii', 'Riccardo Volpi', 'Gabriela Csurka']
|
2021-12-06
| null | null | null | null |
['source-free-domain-adaptation']
|
['computer-vision']
|
[ 6.62560821e-01 -2.75859684e-02 -5.69668531e-01 -6.50714219e-01
-7.25932419e-01 -8.13357472e-01 2.18525201e-01 4.70284373e-02
-5.36105692e-01 7.03134835e-01 -2.41668582e-01 -1.03925079e-01
8.00443515e-02 -6.15715683e-01 -4.83678758e-01 -7.13790476e-01
2.48633534e-01 8.55687976e-01 7.35844016e-01 -1.34381726e-01
5.21744907e-01 3.39609444e-01 -1.65270674e+00 6.46112561e-02
9.71239626e-01 9.97666895e-01 4.74641174e-01 5.20989120e-01
-5.06620646e-01 1.49237305e-01 -7.71786094e-01 -1.20243393e-01
1.14838280e-01 -5.20670772e-01 -1.33769047e+00 5.12436390e-01
3.19451451e-01 1.13560975e-01 2.43641645e-01 1.20114708e+00
3.41586709e-01 3.56128335e-01 6.95176363e-01 -1.17992544e+00
-6.65797472e-01 4.01496559e-01 -4.52697456e-01 2.59307832e-01
9.42164063e-02 -3.12195301e-01 7.79900074e-01 -4.72256958e-01
8.53189111e-01 9.30086553e-01 6.17913365e-01 8.98471117e-01
-8.59082758e-01 -4.43520069e-01 5.17192543e-01 4.88533378e-01
-9.74554300e-01 -3.14245485e-02 8.88119578e-01 -4.42250341e-01
6.41376615e-01 7.69449249e-02 5.54350674e-01 1.03140843e+00
-2.17426687e-01 1.21097589e+00 1.13116205e+00 -5.71520984e-01
4.84377801e-01 1.70015484e-01 2.90832698e-01 2.41547853e-01
-2.71399543e-02 -2.69288570e-01 -3.14502299e-01 1.65502161e-01
9.01306689e-01 -1.92800879e-01 1.97541863e-02 -8.48708987e-01
-1.17929423e+00 9.99958098e-01 2.70318329e-01 4.03503627e-01
-4.15374488e-02 -4.25675184e-01 6.40615046e-01 3.12029213e-01
4.17095095e-01 4.48845834e-01 -8.58518362e-01 -1.88882142e-01
-9.37448978e-01 2.67296135e-01 6.95479155e-01 1.25169516e+00
7.91114211e-01 2.62591634e-02 3.39341551e-01 1.34108996e+00
1.43499942e-02 4.34409589e-01 7.07394719e-01 -1.14790487e+00
-6.52586995e-03 4.81516838e-01 -6.26296923e-02 -4.22051430e-01
-3.28186631e-01 -1.80795211e-02 -5.46227515e-01 2.36829128e-02
5.02846777e-01 2.91770324e-02 -1.29950190e+00 1.29605198e+00
4.01820868e-01 2.01884001e-01 -3.29189934e-02 8.46300960e-01
8.60595763e-01 3.28272253e-01 4.34054792e-01 -7.17486665e-02
1.26823735e+00 -1.31166101e+00 -4.94988948e-01 -6.29419267e-01
2.51703680e-01 -9.81019437e-01 1.06391108e+00 3.97769570e-01
-7.67558813e-01 -6.74258769e-01 -9.19362426e-01 -1.64267272e-02
-7.50927150e-01 -4.45253491e-01 9.17592764e-01 8.13481271e-01
-7.16227829e-01 5.12682021e-01 -8.88520896e-01 -1.09060681e+00
8.17688644e-01 3.34262133e-01 -2.56850183e-01 -1.54117852e-01
-9.67108846e-01 7.68882096e-01 6.35674238e-01 -4.28387046e-01
-6.89267337e-01 -5.31582475e-01 -8.51703763e-01 -6.75445616e-01
5.56397974e-01 -4.48082179e-01 1.58943152e+00 -1.32235587e+00
-1.48797464e+00 1.44868410e+00 -2.23854899e-01 -3.68166298e-01
4.09164011e-01 -1.24818079e-01 -5.24570882e-01 2.97784120e-01
4.21559632e-01 9.25673068e-01 6.76176608e-01 -1.34057331e+00
-1.11296666e+00 -4.88983363e-01 -1.94939837e-01 4.26179945e-01
-1.32763177e-01 1.24684699e-01 -6.17972970e-01 -7.11091280e-01
3.21461648e-01 -7.96846390e-01 -4.85549271e-01 -1.26097530e-01
-2.12796882e-01 -3.93514514e-01 1.13956475e+00 -5.44606090e-01
8.34452152e-01 -2.20311880e+00 3.02201752e-02 -2.12559715e-01
-6.25224710e-02 3.83070946e-01 -4.97352034e-02 2.20845431e-01
5.14890365e-02 -1.71064083e-02 -7.63912261e-01 -5.86943664e-02
-1.73468903e-01 4.59004223e-01 -2.15811759e-01 3.09554011e-01
7.56176189e-02 8.41738045e-01 -1.05395401e+00 -5.84609866e-01
4.95107800e-01 1.25606433e-01 -2.85120815e-01 5.82790896e-02
-5.04106462e-01 8.96917641e-01 -7.74432003e-01 9.64535952e-01
6.58665061e-01 -3.65809143e-01 -1.09741405e-01 2.53039390e-01
-1.76375341e-02 -8.02600458e-02 -1.13556552e+00 2.11959910e+00
-1.32480100e-01 4.49441433e-01 1.75926760e-01 -1.66663218e+00
1.01264203e+00 1.77350193e-01 6.38143897e-01 -6.42314255e-01
8.63831118e-02 3.62891823e-01 -2.90741563e-01 -4.28471982e-01
4.62098628e-01 -2.47544721e-01 -3.43254060e-01 1.33642748e-01
2.27704316e-01 -5.75288892e-01 3.89060885e-01 -1.28972426e-01
6.07333899e-01 3.08051199e-01 5.55989146e-01 -2.68882066e-01
7.02453732e-01 5.83066881e-01 7.07858324e-01 6.85801566e-01
-8.32379520e-01 7.63344467e-01 7.01788440e-02 -6.44811153e-01
-1.00899553e+00 -1.25387704e+00 -3.17176729e-01 1.47091866e+00
5.74783325e-01 2.95341760e-01 -1.07175291e+00 -8.52400541e-01
-2.26120874e-01 3.18860888e-01 -5.20150542e-01 1.73228472e-01
-5.50992429e-01 -8.09131324e-01 3.93473536e-01 8.43563497e-01
1.03800523e+00 -1.37641394e+00 -4.76100743e-01 2.39380926e-01
-3.35021973e-01 -1.35884166e+00 -8.82795751e-02 2.64673203e-01
-1.36465776e+00 -1.25853848e+00 -1.06539631e+00 -1.54739594e+00
6.72361434e-01 3.15422267e-01 1.20709872e+00 -1.67035520e-01
-3.27535480e-01 4.97418165e-01 -6.23112381e-01 -4.87524599e-01
-3.07260811e-01 3.26007694e-01 -3.31816345e-01 -3.63477349e-01
1.02953815e+00 -3.70501757e-01 -4.30664986e-01 6.50478542e-01
-9.54410851e-01 -2.75423348e-01 4.22506213e-01 7.03602195e-01
9.50239956e-01 -2.66840272e-02 8.12865257e-01 -1.48870707e+00
3.97007436e-01 -2.66235948e-01 -6.19015872e-01 2.58161455e-01
-4.61716384e-01 -3.53539348e-01 4.35572386e-01 -2.77310848e-01
-1.25133169e+00 2.90613502e-01 -3.47937584e-01 2.67612953e-02
-8.23559403e-01 1.08962364e-01 -2.28896961e-01 -1.58453405e-01
7.43714571e-01 2.90144652e-01 -1.37191117e-01 -6.26002073e-01
5.52141964e-01 8.31273198e-01 7.75043786e-01 -6.92683220e-01
4.12210345e-01 7.09881961e-01 -2.69581169e-01 -1.15258574e+00
-9.65139449e-01 -1.11847854e+00 -1.09543467e+00 -2.54515465e-02
1.09783745e+00 -5.87063670e-01 6.36425838e-02 8.85836899e-01
-8.36942017e-01 -5.61706364e-01 -3.83227587e-01 3.48088332e-02
-8.98210108e-01 5.68674803e-01 -3.95610631e-01 -2.51553625e-01
-5.35305403e-03 -1.26857746e+00 1.01649582e+00 6.83965504e-01
-1.61803707e-01 -1.56719327e+00 3.58322030e-03 8.40184748e-01
2.62793064e-01 1.72032818e-01 9.29895639e-01 -1.04344308e+00
-5.42064250e-01 -6.84439167e-02 -2.15600699e-01 6.08106136e-01
2.55088270e-01 -1.80977210e-01 -1.04653931e+00 9.83377360e-03
-1.29617721e-01 -4.05786097e-01 7.61288762e-01 7.37048805e-01
1.40182364e+00 5.51022232e-01 -6.09652519e-01 5.96918166e-01
1.45016992e+00 4.81720299e-01 4.50096369e-01 6.28390551e-01
6.21257544e-01 6.22755408e-01 9.53462303e-01 5.73936962e-02
3.68250072e-01 4.18103307e-01 6.00172058e-02 -2.01704741e-01
-2.97448069e-01 4.44738492e-02 -3.18009645e-01 6.19543612e-01
-1.56049311e-01 -2.34085191e-02 -1.03955364e+00 8.19302082e-01
-1.79452050e+00 -5.94938576e-01 -5.89355268e-02 2.11538625e+00
8.67502689e-01 5.10188527e-02 3.05906504e-01 1.64874762e-01
9.32455420e-01 -5.30302003e-02 -7.83603489e-01 -3.40296656e-01
-2.23793909e-01 4.16723818e-01 5.86741507e-01 4.67792332e-01
-1.60466325e+00 1.49782395e+00 7.56154871e+00 8.89117897e-01
-1.09770286e+00 2.34521031e-01 5.70467651e-01 6.00729704e-01
1.67428106e-01 -1.11888491e-01 -7.66881227e-01 3.14786822e-01
5.14349043e-01 -6.86720982e-02 2.09043220e-01 1.34027565e+00
-8.74021724e-02 -3.55569780e-01 -9.10172343e-01 9.70600963e-01
1.60676077e-01 -1.12266409e+00 -2.00526312e-01 -2.05674633e-01
1.11513031e+00 2.56868184e-01 -8.44380558e-02 1.47769779e-01
2.80576855e-01 -8.01578283e-01 2.22673252e-01 -7.63239339e-02
5.47150731e-01 -6.05055392e-01 6.32734716e-01 2.22372994e-01
-1.03359854e+00 1.18755018e-02 -5.05225122e-01 2.81877816e-02
3.03824723e-01 5.20547092e-01 -6.49067402e-01 3.99311095e-01
9.07607496e-01 9.07398820e-01 -4.17355686e-01 1.25339985e+00
-2.28105798e-01 4.93428648e-01 -7.16199949e-02 2.39942729e-01
3.15704525e-01 -3.26999664e-01 3.52824867e-01 1.35573387e+00
-1.46575287e-01 -4.18107435e-02 5.23098767e-01 4.13948566e-01
4.92677055e-02 2.30395094e-01 -4.54639673e-01 -8.62738565e-02
4.35653538e-01 1.05973148e+00 -1.48341405e+00 -5.13714314e-01
-6.74855709e-01 1.28136909e+00 -6.81795776e-02 5.15105069e-01
-6.44369364e-01 -4.96162236e-01 6.83457911e-01 -1.31043494e-01
3.71451855e-01 -1.81690946e-01 -8.03333521e-01 -8.80336583e-01
-2.29132056e-01 -8.19145799e-01 6.63186789e-01 -4.45085973e-01
-1.35377085e+00 4.26707625e-01 2.02710420e-01 -9.77789402e-01
-1.08850695e-01 -7.92891800e-01 -1.56028748e-01 7.16658652e-01
-1.74799776e+00 -1.07397628e+00 -4.41159666e-01 5.92375338e-01
1.18378985e+00 -2.45078713e-01 8.68614137e-01 2.97197580e-01
-2.97023803e-01 2.94819653e-01 2.31393218e-01 3.61952394e-01
9.39024329e-01 -1.29279137e+00 5.67683995e-01 6.61501706e-01
1.64427474e-01 3.28472853e-01 5.63534081e-01 -6.12476468e-01
-7.06081867e-01 -1.03628528e+00 6.40184939e-01 -6.03168309e-01
4.82328564e-01 -1.56436011e-01 -1.01353240e+00 8.13749850e-01
2.77661663e-02 -9.18601826e-02 7.99460649e-01 8.46448243e-02
-1.45705655e-01 -2.35710517e-02 -1.34408057e+00 3.47885936e-01
9.88596618e-01 -1.49674505e-01 -7.91169286e-01 3.17443788e-01
4.95812565e-01 -5.41824698e-01 -6.86951339e-01 5.15379429e-01
2.30244696e-01 -9.27226424e-01 9.73860145e-01 -4.14101332e-01
1.76913470e-01 -1.18037455e-01 -7.32168928e-02 -1.19627476e+00
-2.26392135e-01 -2.73946971e-01 3.97987843e-01 1.13743126e+00
1.92248359e-01 -6.68955445e-01 1.26369452e+00 3.72180909e-01
-2.96112865e-01 -4.12142545e-01 -7.94166207e-01 -9.81522560e-01
4.32247937e-01 -3.59753162e-01 3.51726264e-01 1.07827926e+00
-2.49594972e-01 1.93815723e-01 2.59986430e-01 -9.31590945e-02
7.72622168e-01 1.78827375e-01 5.61760843e-01 -1.51457143e+00
2.04446882e-01 -5.96131444e-01 -6.05981052e-01 -1.46721113e+00
2.07530513e-01 -6.09968603e-01 2.47160450e-01 -1.75061035e+00
1.80119157e-01 -5.28516710e-01 -2.47782826e-01 3.29682827e-01
-6.86888471e-02 6.55696511e-01 -1.43333912e-01 1.61788002e-01
-8.15039754e-01 1.13494135e-01 1.60574925e+00 -1.68746978e-01
-2.06227034e-01 3.06545049e-01 -8.40029836e-01 1.09271717e+00
8.12376499e-01 -2.85150886e-01 -7.31251836e-01 -5.66768289e-01
-4.05788481e-01 -4.65454638e-01 2.05066483e-02 -9.56226587e-01
6.00769408e-02 -4.82900560e-01 2.52767980e-01 -5.02361953e-01
1.06049247e-01 -7.88115501e-01 -3.93619448e-01 2.33850144e-02
2.02228129e-02 -3.51881117e-01 2.77539730e-01 5.50410330e-01
-6.04915261e-01 -4.68256295e-01 1.31511545e+00 -6.82481885e-01
-1.96057558e+00 3.00515980e-01 -1.61061138e-01 5.54442227e-01
1.35380435e+00 -8.22427750e-01 1.60355046e-01 -1.49538768e-02
-9.94342923e-01 2.81687498e-01 7.16675758e-01 6.36895418e-01
4.00879145e-01 -9.09687996e-01 -3.39367688e-01 1.92310840e-01
3.77215683e-01 4.58979368e-01 3.42924535e-01 3.47324312e-01
-6.79051757e-01 3.71228367e-01 -3.61341000e-01 -8.97975147e-01
-1.24476254e+00 4.64327991e-01 2.34258950e-01 2.72132773e-02
-5.51599622e-01 1.23241413e+00 2.40144178e-01 -6.68832004e-01
2.58470505e-01 -1.09739423e-01 -2.32911944e-01 -1.12312771e-01
3.56755078e-01 3.51961792e-01 1.11199044e-01 -4.86360610e-01
-4.89010155e-01 8.37106466e-01 -1.71111882e-01 -2.52942066e-03
1.19908273e+00 -2.44139746e-01 -5.50792962e-02 5.19566894e-01
7.91622758e-01 -5.79496920e-01 -1.51723826e+00 -2.34702602e-01
2.25535661e-01 -4.08331484e-01 -2.30264708e-01 -1.04002237e+00
-9.28029180e-01 1.04312623e+00 5.95754862e-01 1.67160407e-01
1.41084981e+00 4.66080815e-01 9.29392934e-01 2.12055027e-01
6.18411720e-01 -1.78718197e+00 1.95648834e-01 6.18145704e-01
3.89076293e-01 -1.64415348e+00 -8.40506852e-02 -7.47185171e-01
-8.54054928e-01 9.77952242e-01 6.11141264e-01 -1.84787765e-01
7.40134180e-01 -2.69724755e-03 5.36332905e-01 1.14662252e-01
1.55441493e-01 -3.60640258e-01 3.30452658e-02 1.40969849e+00
3.83859187e-01 5.50571159e-02 -8.58433023e-02 3.87168735e-01
-1.39592469e-01 9.68759283e-02 2.20918924e-01 1.00421524e+00
-8.42731297e-01 -1.73174453e+00 -3.65294755e-01 2.16411799e-01
-3.45984876e-01 1.92526996e-01 -4.74611074e-01 9.30263579e-01
2.83637524e-01 9.26629245e-01 7.17337951e-02 2.47554496e-01
3.28887373e-01 2.82633364e-01 7.34059036e-01 -1.01105690e+00
-1.05736710e-01 7.61861280e-02 -1.68938920e-01 -3.90674382e-01
-9.12040234e-01 -7.84551859e-01 -1.48016036e+00 2.02259347e-02
-2.46802680e-02 1.13505810e-01 7.49661744e-01 1.14801323e+00
1.78550612e-02 3.26007009e-01 1.20523855e-01 -6.50069296e-01
-1.50139377e-01 -8.00269783e-01 -7.84593284e-01 6.21458054e-01
1.18144497e-01 -6.90439403e-01 -4.86020520e-02 7.28479028e-01]
|
[9.644550323486328, 1.2861918210983276]
|
03e2a412-d73c-4305-8ad7-efa7bb0cb2b7
|
nesy4vrd-a-multifaceted-resource-for
|
2305.13258
| null |
https://arxiv.org/abs/2305.13258v1
|
https://arxiv.org/pdf/2305.13258v1.pdf
|
NeSy4VRD: A Multifaceted Resource for Neurosymbolic AI Research using Knowledge Graphs in Visual Relationship Detection
|
NeSy4VRD is a multifaceted resource designed to support the development of neurosymbolic AI (NeSy) research. NeSy4VRD re-establishes public access to the images of the VRD dataset and couples them with an extensively revised, quality-improved version of the VRD visual relationship annotations. Crucially, NeSy4VRD provides a well-aligned, companion OWL ontology that describes the dataset domain.It comes with open source infrastructure that provides comprehensive support for extensibility of the annotations (which, in turn, facilitates extensibility of the ontology), and open source code for loading the annotations to/from a knowledge graph. We are contributing NeSy4VRD to the computer vision, NeSy and Semantic Web communities to help foster more NeSy research using OWL-based knowledge graphs.
|
['Tillman Weyde', 'Giacomo Tarroni', 'Ernesto Jiménez-Ruiz', 'David Herron']
|
2023-05-22
| null | null | null | null |
['visual-relationship-detection']
|
['computer-vision']
|
[-3.14863533e-01 5.23521304e-01 -5.91196775e-01 -3.86850983e-01
8.02805871e-02 -4.77286100e-01 5.32255471e-01 3.32774788e-01
-1.84897944e-01 4.79381412e-01 5.68239927e-01 -9.35866609e-02
-4.92168933e-01 -7.34417617e-01 -3.73008937e-01 7.19189271e-02
-1.63105130e-01 8.19462299e-01 5.09323418e-01 -5.19191444e-01
-6.02664314e-02 5.54345131e-01 -1.68868291e+00 1.50095746e-01
6.02326870e-01 9.13204074e-01 2.10066661e-01 2.94874579e-01
1.20476313e-01 7.03975856e-01 -1.34121537e-01 -4.74475265e-01
1.00593999e-01 9.23680421e-03 -1.15015280e+00 -1.97872594e-01
5.55914283e-01 -2.24475026e-01 -4.67345089e-01 8.32324624e-01
2.54768521e-01 1.69278145e-01 4.91859585e-01 -1.85565734e+00
-1.06387496e+00 2.89461136e-01 -5.84502518e-02 3.94432157e-01
5.23424745e-01 -7.28255659e-02 1.11383772e+00 -5.00151217e-01
1.78564906e+00 1.22694361e+00 4.24070179e-01 5.49814343e-01
-7.33609855e-01 -6.91053092e-01 -6.01933897e-02 5.74962735e-01
-1.16804802e+00 -4.53973949e-01 2.41538405e-01 -3.81225437e-01
1.24329782e+00 2.86168367e-01 1.24664474e+00 1.31160760e+00
-1.30699247e-01 5.18065572e-01 8.63357842e-01 -3.81410509e-01
3.18337977e-01 -9.69804730e-03 1.55240506e-01 9.94614124e-01
3.74681443e-01 7.06280693e-02 -6.37518406e-01 -8.91690701e-02
1.09587729e+00 -4.27336007e-01 1.17664814e-01 -8.37989628e-01
-8.35188448e-01 5.78130364e-01 5.86330295e-01 5.11149585e-01
-2.26901606e-01 6.09684410e-03 5.45761287e-01 1.95384119e-02
6.31735101e-02 4.93531734e-01 -2.33514234e-01 -2.04771906e-01
-3.08936477e-01 4.66044933e-01 7.83396304e-01 1.34722722e+00
2.76532173e-01 -1.28525332e-01 2.66583592e-01 1.04476070e+00
2.54419804e-01 2.17301250e-01 2.89682835e-01 -1.52357888e+00
-3.20999473e-02 8.92773569e-01 -3.10472529e-02 -1.28123081e+00
-6.80622816e-01 -1.80941984e-01 -4.74447906e-02 5.07131696e-01
6.54701814e-02 6.66798115e-01 -9.19331491e-01 1.62748826e+00
3.63962710e-01 1.07727870e-02 2.08508760e-01 1.08916891e+00
1.39514482e+00 -1.94706827e-01 4.80068982e-01 4.30161923e-01
1.45788789e+00 -8.20085049e-01 -8.51141453e-01 -5.99415079e-02
5.44447482e-01 -2.88557649e-01 9.21183705e-01 1.68941185e-01
-1.04806709e+00 -9.32964459e-02 -1.33963799e+00 -5.38242340e-01
-1.10630441e+00 -1.63225576e-01 8.84419858e-01 3.10802221e-01
-1.01901150e+00 1.72848791e-01 -7.44318187e-01 -1.09671676e+00
8.85704875e-01 2.03108147e-01 -9.01274562e-01 -3.67681950e-01
-1.29431534e+00 1.84557521e+00 7.13407695e-01 -4.82892662e-01
-9.20618057e-01 -7.28184581e-01 -9.87625957e-01 -4.02669489e-01
5.97699940e-01 -8.99935782e-01 8.81420672e-01 -6.11888528e-01
-8.07414353e-01 1.51155257e+00 3.13497305e-01 -4.21143085e-01
1.66929722e-01 1.03712924e-01 -8.92374873e-01 5.92388630e-01
2.67200619e-01 9.75219011e-01 2.75348336e-01 -1.00803578e+00
-5.13952434e-01 -6.45160019e-01 5.84882081e-01 4.01583672e-01
-2.72716790e-01 1.77800879e-01 -7.04181671e-01 -6.14395320e-01
-2.01131225e-01 -7.21843362e-01 5.02135120e-02 3.93583983e-01
-1.88066196e-02 -3.05258662e-01 9.76630986e-01 -8.35635662e-01
9.53214526e-01 -2.18751645e+00 4.64189261e-01 1.85383528e-01
9.26495314e-01 1.39062747e-01 -1.37604192e-01 3.97227198e-01
-2.33670637e-01 6.02352060e-02 8.27665925e-02 1.82595372e-01
2.20932826e-01 7.34265804e-01 3.73954587e-02 3.26039255e-01
-1.04136266e-01 7.75641501e-01 -1.13996553e+00 -5.07907212e-01
4.98543978e-01 5.00469804e-01 -5.03070354e-01 -3.29176724e-01
-2.09414035e-01 2.25593388e-01 -4.11040813e-01 7.50428677e-01
3.45906526e-01 -1.79343030e-01 2.94588089e-01 -4.52320188e-01
-1.68716714e-01 -2.00695604e-01 -8.73026729e-01 2.05570841e+00
-2.71770149e-01 5.45004547e-01 -6.64945990e-02 -6.15121365e-01
7.78158069e-01 3.41114253e-01 6.69712782e-01 -8.98388624e-01
3.47795188e-01 2.37715289e-01 -2.69971132e-01 -7.29751229e-01
5.26348352e-01 -2.40227301e-03 7.83191472e-02 1.80172309e-01
6.64542198e-01 -1.19226992e-01 4.07705843e-01 5.63763022e-01
1.16082883e+00 6.67256713e-01 7.46152759e-01 -3.05959523e-01
1.86131492e-01 3.29656154e-01 2.99427271e-01 2.04830438e-01
-3.28507751e-01 -5.46335317e-02 2.07486331e-01 -4.55527723e-01
-1.36542571e+00 -1.34945571e+00 -5.97951353e-01 7.50208378e-01
2.32503936e-01 -9.68109846e-01 -4.94251311e-01 -5.01208484e-01
9.19766873e-02 6.96551442e-01 -6.93188369e-01 -1.86561167e-01
6.68517575e-02 -5.72679602e-02 9.24942732e-01 6.21038079e-01
3.49506766e-01 -1.16884339e+00 -8.56482744e-01 -1.30467827e-03
-6.97620958e-02 -1.54642701e+00 1.82211637e-01 -2.45955333e-01
-3.47611636e-01 -1.62508714e+00 -3.44150394e-01 -7.48885453e-01
5.07694602e-01 -3.05030763e-01 8.69978428e-01 2.52970546e-01
-8.47927988e-01 8.73183548e-01 -5.90361536e-01 -5.70086598e-01
-3.03226948e-01 -2.42804125e-01 1.75684586e-01 -9.05217111e-01
5.73770881e-01 -6.74678326e-01 -1.23009145e-01 1.67248115e-01
-9.46172357e-01 2.66558528e-01 -1.54749185e-01 2.78497249e-01
5.17662108e-01 -2.82383561e-02 6.34897530e-01 -6.65373206e-01
4.87733662e-01 -4.14499521e-01 -5.00602722e-01 1.87663436e-01
-6.45722449e-01 -3.41417164e-01 3.90943959e-02 8.89185369e-02
-8.45944881e-01 -3.34264606e-01 -1.84789449e-01 -5.56464493e-01
-3.21652234e-01 3.39678615e-01 -9.37571898e-02 -7.41871074e-02
6.38127446e-01 -4.40021783e-01 -7.73115158e-02 -4.10359025e-01
8.35070193e-01 6.85449302e-01 9.59558666e-01 -4.28531408e-01
6.32043004e-01 7.55544782e-01 -1.05137913e-03 -9.50139999e-01
-5.87782204e-01 -2.42243081e-01 -6.16494715e-01 -3.43319982e-01
1.24756217e+00 -6.92132950e-01 -7.78926313e-01 1.45654790e-02
-5.83526969e-01 -2.81621039e-01 -4.68308717e-01 3.92636061e-01
-8.76191020e-01 6.15145527e-02 -3.24340999e-01 -4.02311057e-01
1.05268672e-01 -9.14043307e-01 6.78751826e-01 2.46728212e-01
-6.68680847e-01 -1.22402978e+00 5.34772985e-02 6.35788262e-01
1.23447247e-01 6.22919917e-01 1.17111015e+00 -6.64058864e-01
-1.90894440e-01 4.04277630e-02 -3.60616684e-01 -7.85034820e-02
-1.50253594e-01 2.22766176e-02 -7.58446217e-01 1.63198724e-01
-7.57743895e-01 -4.83035624e-01 2.39631623e-01 9.07419100e-02
7.29309559e-01 -7.63641521e-02 -5.14284492e-01 4.81200159e-01
1.33744705e+00 4.17214453e-01 6.98083162e-01 8.52227092e-01
6.37789249e-01 5.61501861e-01 6.21314824e-01 3.56993347e-01
8.95534694e-01 9.14159238e-01 5.85772812e-01 -1.29958943e-01
-5.13460040e-01 -3.10821891e-01 -3.08733582e-01 4.68526840e-01
-5.42766392e-01 6.11622520e-02 -9.57141757e-01 5.32050729e-01
-2.00296068e+00 -1.00685024e+00 -2.03454450e-01 1.69426763e+00
6.44538999e-01 -2.33899206e-01 1.25587538e-01 -8.11541080e-02
4.43330139e-01 -2.90998966e-02 -5.65428078e-01 -5.50050855e-01
-1.92232132e-01 4.17428881e-01 2.49554813e-01 4.14118767e-01
-7.38778412e-01 1.33424520e+00 7.52097940e+00 5.36884189e-01
-4.37034070e-01 4.03574079e-01 -4.59096551e-01 9.59510580e-02
-3.51553053e-01 9.13973451e-02 -5.97111523e-01 -1.14131318e-02
8.72813821e-01 -5.99939048e-01 8.64838541e-01 7.25741625e-01
1.19308367e-01 -5.85855357e-02 -8.68881881e-01 8.54996979e-01
2.78559983e-01 -1.90366077e+00 -1.33816585e-01 1.78633764e-01
4.12491024e-01 2.73197234e-01 -3.39565575e-01 -8.10558572e-02
6.87628567e-01 -9.26283538e-01 7.69349873e-01 5.41573107e-01
1.01941466e+00 -6.35866046e-01 2.81343520e-01 -1.67680591e-01
-1.14521348e+00 -6.85610250e-02 -3.51023227e-01 1.33800477e-01
3.11517175e-02 9.56361070e-02 -5.90348005e-01 7.30967820e-01
1.16655517e+00 1.25894010e+00 -7.88092494e-01 9.17552054e-01
-3.79582912e-01 -1.32630989e-01 -8.52027722e-03 4.16608274e-01
-1.06117226e-01 -6.54275641e-02 8.30453992e-01 7.39122868e-01
-1.99612930e-01 1.60590872e-01 2.81173289e-01 7.08698690e-01
-8.85591656e-03 5.31741306e-02 -9.35253024e-01 -5.10012388e-01
6.32449746e-01 1.14279044e+00 -7.12253869e-01 -3.33640814e-01
-3.88923019e-01 8.04719627e-01 4.16375339e-01 2.28842959e-01
-8.56515765e-01 -3.54684025e-01 8.71526539e-01 4.59843576e-02
1.59450352e-01 -2.38394916e-01 -6.98444024e-02 -8.74213994e-01
-3.97384375e-01 -7.35202909e-01 8.71588409e-01 -1.61471963e+00
-1.17480350e+00 7.12844372e-01 3.65659267e-01 -9.40049827e-01
-2.89139291e-03 -6.65023804e-01 1.91989630e-01 4.84687090e-01
-1.24409115e+00 -1.57955241e+00 -4.95860696e-01 7.51521528e-01
5.19322157e-02 -2.54230261e-01 1.38730276e+00 4.16640133e-01
-4.99417245e-01 1.05431013e-01 -5.18385649e-01 -3.87666188e-03
5.95069766e-01 -9.09543872e-01 2.51541343e-02 3.78782094e-01
-2.92355102e-02 6.28091872e-01 7.18465686e-01 -9.60062563e-01
-1.02856827e+00 -8.22024465e-01 6.37724340e-01 -7.16714025e-01
1.02520728e+00 6.67570010e-02 -4.30815369e-01 1.34057403e+00
2.98359901e-01 -3.42634246e-02 8.88074338e-01 2.66739465e-02
-5.45537949e-01 3.43591332e-01 -1.35834765e+00 8.07369053e-01
1.69100630e+00 -4.63789225e-01 -1.23818171e+00 3.75866413e-01
7.16489494e-01 -7.18554199e-01 -1.56942427e+00 1.75964549e-01
8.37001443e-01 -7.61595905e-01 1.09859562e+00 -9.02718663e-01
5.26264668e-01 -4.25532669e-01 -4.32243437e-01 -1.07324290e+00
-2.68370122e-01 3.48226689e-02 -1.49070382e-01 8.66383851e-01
1.34120733e-01 -6.41192317e-01 4.27622616e-01 5.93794048e-01
-5.33327222e-01 -6.48979664e-01 -9.26839352e-01 -8.69210005e-01
-3.49983901e-01 -7.07616687e-01 7.59357750e-01 1.41065383e+00
4.77565855e-01 -6.01026453e-02 -6.44423142e-02 2.51669846e-02
5.85912824e-01 -3.38488638e-01 5.14858246e-01 -1.51784372e+00
7.22456872e-02 -3.24291140e-01 -1.16712368e+00 4.16058376e-02
1.44652501e-01 -1.48596513e+00 -6.69353902e-01 -2.16676879e+00
4.69666086e-02 -4.79867488e-01 1.25095872e-02 1.19031000e+00
6.49066091e-01 6.88972354e-01 1.72549322e-01 1.03926204e-01
-4.71039534e-01 2.40636349e-01 1.47255790e+00 1.54773265e-01
-1.61322095e-02 -8.68274212e-01 -9.20228660e-01 1.11869526e+00
5.39661348e-01 -3.22648704e-01 -6.20637476e-01 -3.52475554e-01
2.84305722e-01 -3.35779428e-01 8.33525598e-01 -1.00316334e+00
-4.47933748e-02 -2.59666860e-01 5.29423416e-01 -4.05617118e-01
6.05723143e-01 -1.00243294e+00 3.35263163e-01 2.36579120e-01
-2.21969157e-01 -3.62333357e-02 3.59972149e-01 1.01641953e-01
1.40080586e-01 -2.14875430e-01 6.79320812e-01 -2.47997403e-01
-1.36735344e+00 4.16397274e-01 -1.64869994e-01 2.91223764e-01
1.38406813e+00 -5.92371285e-01 -8.43951106e-01 -2.45888010e-02
-1.31192410e+00 3.85664493e-01 8.49255800e-01 1.00071621e+00
6.41606927e-01 -1.47565031e+00 -3.80046032e-02 5.79824448e-02
6.53361678e-01 -5.31992495e-01 3.48840296e-01 7.54784107e-01
-6.69390202e-01 3.15697968e-01 -1.08104122e+00 -1.36650980e-01
-1.28618932e+00 4.56134707e-01 1.53195590e-01 1.45181209e-01
-1.46494293e+00 3.47249389e-01 -3.90166968e-01 -5.41630447e-01
3.42610329e-01 1.00901283e-01 -5.33994853e-01 -9.81597677e-02
6.50193214e-01 5.69446087e-01 -1.22999735e-01 -8.55942786e-01
-6.06099963e-01 3.24795574e-01 2.46382311e-01 -2.41871566e-01
1.67526388e+00 -5.53050116e-02 -1.43280074e-01 4.19191182e-01
7.81652749e-01 -4.53599691e-02 -6.98911130e-01 2.70214111e-01
8.05279016e-02 -5.33045471e-01 1.00620158e-01 -1.22349823e+00
-8.73018444e-01 1.61314532e-01 5.46877146e-01 -1.33812934e-01
9.63847041e-01 4.92762148e-01 6.74744904e-01 1.55452892e-01
5.77944517e-01 -1.23325825e+00 -2.48535037e-01 2.70598263e-01
1.09976864e+00 -5.48862278e-01 3.44708294e-01 -6.03385746e-01
-8.92612934e-01 9.81445670e-01 7.17226148e-01 -1.24217495e-01
5.33112824e-01 2.85115927e-01 1.12978764e-01 -8.52572381e-01
-5.37415087e-01 -3.93815905e-01 1.88298851e-01 1.39606822e+00
1.46467716e-01 -1.70069013e-03 -5.67795515e-01 5.54213762e-01
-3.88151616e-01 5.39512515e-01 3.95822376e-01 9.29512382e-01
-2.30527267e-01 -1.28679216e+00 3.64284217e-02 3.50278914e-01
1.47024706e-01 1.55959614e-02 -4.38299447e-01 1.13707948e+00
5.92799485e-01 5.74744165e-01 3.21999453e-02 -1.21866979e-01
6.36637390e-01 -7.44687114e-03 9.12762046e-01 -5.20052791e-01
-3.17468703e-01 -3.35961401e-01 6.88660920e-01 -9.03701067e-01
-6.11447632e-01 -3.54352236e-01 -1.99074793e+00 -2.10356742e-01
2.17596784e-01 -1.15188122e-01 8.38061631e-01 9.44406688e-01
5.53762138e-01 7.49589324e-01 -3.87962401e-01 -4.62090045e-01
4.92616206e-01 -2.73076504e-01 -8.14666927e-01 3.86013478e-01
-5.01697659e-01 -1.31921637e+00 2.65421212e-01 -5.34817949e-02]
|
[9.03248119354248, 7.928737163543701]
|
906a3215-4cba-48ad-9cca-948225f17c82
|
purepos-20-a-hybrid-tool-for-morphological
| null | null |
https://aclanthology.org/R13-1071
|
https://aclanthology.org/R13-1071.pdf
|
PurePos 2.0: a hybrid tool for morphological disambiguation
| null |
["Attila Nov{\\'a}k", 'Gy{\\"o}rgy Orosz']
|
2013-09-01
|
purepos-20-a-hybrid-tool-for-morphological-1
|
https://aclanthology.org/R13-1071
|
https://aclanthology.org/R13-1071.pdf
|
ranlp-2013-9
|
['morphological-disambiguation']
|
['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]
|
[-7.259118556976318, 3.5106022357940674]
|
5f005039-bda1-411a-8693-357306d43f6d
|
revisiting-the-roles-of-text-in-text-games-1
|
2210.08384
| null |
https://arxiv.org/abs/2210.08384v1
|
https://arxiv.org/pdf/2210.08384v1.pdf
|
Revisiting the Roles of "Text" in Text Games
|
Text games present opportunities for natural language understanding (NLU) methods to tackle reinforcement learning (RL) challenges. However, recent work has questioned the necessity of NLU by showing random text hashes could perform decently. In this paper, we pursue a fine-grained investigation into the roles of text in the face of different RL challenges, and reconcile that semantic and non-semantic language representations could be complementary rather than contrasting. Concretely, we propose a simple scheme to extract relevant contextual information into an approximate state hash as extra input for an RNN-based text agent. Such a lightweight plug-in achieves competitive performance with state-of-the-art text agents using advanced NLU techniques such as knowledge graph and passage retrieval, suggesting non-NLU methods might suffice to tackle the challenge of partial observability. However, if we remove RNN encoders and use approximate or even ground-truth state hash alone, the model performs miserably, which confirms the importance of semantic function approximation to tackle the challenge of combinatorially large observation and action spaces. Our findings and analysis provide new insights for designing better text game task setups and agents.
|
['Mo Yu', 'Joshua B. Tenenbaum', 'Chuang Gan', 'Shunyu Yao', 'Yi Gu']
|
2022-10-15
| null | null | null | null |
['passage-retrieval']
|
['natural-language-processing']
|
[ 1.40335724e-01 6.64339483e-01 -1.50708109e-01 2.04265624e-01
-1.09612596e+00 -8.51990283e-01 9.79229510e-01 1.66781858e-01
-8.11351001e-01 8.17272544e-01 5.53254306e-01 -5.77590287e-01
-1.03379376e-01 -7.91105032e-01 -8.22831690e-01 -5.95607221e-01
1.09658107e-01 8.98881257e-01 1.38035208e-01 -5.66333473e-01
1.76641807e-01 2.94766156e-03 -1.53160727e+00 1.14268810e-01
4.79287893e-01 5.34444571e-01 2.60579944e-01 1.01511848e+00
-1.79063082e-01 1.43942428e+00 -6.62993610e-01 -5.51611722e-01
3.19072753e-01 -6.76187754e-01 -1.14102030e+00 -2.59449065e-01
3.22450668e-01 -7.99581528e-01 -8.21003616e-01 9.85999882e-01
6.23125732e-01 4.28456932e-01 4.17684942e-01 -1.23777211e+00
-6.15320265e-01 9.42255974e-01 1.02029525e-01 3.18560749e-02
4.49295908e-01 5.53363383e-01 1.51032937e+00 1.00104297e-02
7.96363533e-01 1.43767631e+00 5.04045844e-01 8.61009061e-01
-9.90077794e-01 -3.39475572e-01 1.15144059e-01 3.99073929e-01
-8.35720003e-01 -5.42348504e-01 4.51831937e-01 1.25658019e-02
1.32461298e+00 1.57077998e-01 6.12917304e-01 1.38032997e+00
4.40505184e-02 1.35015714e+00 1.17899370e+00 -4.05847073e-01
3.35289031e-01 -1.16878442e-01 -1.74920931e-01 1.10897827e+00
2.86537588e-01 4.37071770e-01 -9.05654967e-01 -1.97236449e-01
5.85011899e-01 -2.80740052e-01 -1.73512354e-01 -2.94054508e-01
-1.26028121e+00 1.18524885e+00 1.25212833e-01 1.65270761e-01
-1.28860593e-01 6.80104315e-01 7.55826175e-01 6.29405916e-01
1.45618513e-01 9.31854367e-01 -6.75099254e-01 -7.00156569e-01
-5.33977091e-01 4.55380529e-01 1.12347770e+00 8.14511418e-01
5.28591871e-01 4.47324030e-02 -1.67959183e-01 2.17713371e-01
1.84634358e-01 4.71751422e-01 5.59895217e-01 -1.32952166e+00
4.68817294e-01 2.54734099e-01 2.82329082e-01 -5.59362113e-01
-6.16278768e-01 -2.63133645e-01 -3.05559367e-01 1.19503036e-01
9.16729510e-01 -2.28646904e-01 -6.03616059e-01 1.91772759e+00
2.83207566e-01 1.36532918e-01 4.81688946e-01 8.58263254e-01
4.13499415e-01 6.29595399e-01 7.76935220e-02 -2.58412026e-02
1.71968806e+00 -9.87808585e-01 -6.61180198e-01 -4.43846941e-01
1.01347232e+00 -2.70253241e-01 1.40017176e+00 3.79991770e-01
-1.22659767e+00 -8.31057280e-02 -9.85954702e-01 -4.57590193e-01
-2.83292353e-01 -3.91117781e-01 8.18791926e-01 6.18077815e-01
-1.16576707e+00 5.67473352e-01 -8.52797091e-01 -5.03658533e-01
2.61656046e-01 3.13235760e-01 -3.41986008e-02 -1.75352916e-01
-1.54146194e+00 1.09800172e+00 4.86432105e-01 -1.40788272e-01
-1.14662910e+00 -2.91446805e-01 -1.05872488e+00 2.09103048e-01
9.82131898e-01 -9.22179163e-01 1.70027411e+00 -7.20785022e-01
-1.87731576e+00 5.19799232e-01 8.76508802e-02 -8.35806012e-01
5.58408499e-01 -6.57551661e-02 2.33833104e-01 4.02172595e-01
-4.92819101e-02 6.70316279e-01 7.89146781e-01 -8.85094225e-01
-5.83293796e-01 -3.64969254e-01 6.94306850e-01 5.47445834e-01
-3.09722740e-02 -2.44731158e-01 2.35555246e-02 -3.80813688e-01
-3.02877188e-01 -9.65186417e-01 -1.55951262e-01 -2.36637384e-01
-2.60240227e-01 -4.58963037e-01 4.52611864e-01 -5.13886809e-01
7.52256989e-01 -1.81853235e+00 2.59477556e-01 -2.31330588e-01
3.61757308e-01 2.63108462e-01 -3.85806561e-01 7.89427698e-01
5.07807791e-01 9.11696777e-02 1.00955494e-01 -4.86668199e-01
5.79109311e-01 6.26171470e-01 -4.69998956e-01 4.03817266e-01
-6.67378902e-02 1.44432378e+00 -1.33714628e+00 -3.31525445e-01
1.70129925e-01 1.58268020e-01 -6.54003799e-01 1.12872742e-01
-9.73617017e-01 2.93189287e-01 -7.26840615e-01 2.26672053e-01
-1.14503279e-01 -4.20265138e-01 4.76191878e-01 2.69757867e-01
3.48990321e-01 8.55739951e-01 -9.70615804e-01 1.98325706e+00
-3.71810734e-01 7.60013640e-01 -1.75390567e-03 -1.00519705e+00
2.35692620e-01 3.17705929e-01 1.51425861e-02 -9.54182804e-01
2.56251961e-01 5.87236360e-02 -4.05859090e-02 -5.25807083e-01
7.93436944e-01 -4.26220983e-01 -3.13464910e-01 9.23084736e-01
1.78615719e-01 -1.68136925e-01 5.85189424e-02 4.37455475e-01
1.41649806e+00 3.87782037e-01 2.95134306e-01 -1.33571206e-02
1.61135048e-01 2.01461524e-01 7.75905848e-02 1.45622313e+00
-5.49802184e-01 1.50563613e-01 7.06112146e-01 -5.22044957e-01
-1.04634035e+00 -6.79486215e-01 4.17196482e-01 1.65293825e+00
8.11049938e-02 -7.17676103e-01 -8.41578722e-01 -9.45202827e-01
-9.43367258e-02 9.58984315e-01 -5.49875379e-01 -4.02575403e-01
-5.92277825e-01 -5.00696421e-01 1.12956250e+00 3.91064733e-01
4.14608717e-01 -1.31577396e+00 -1.00142026e+00 3.31107795e-01
-4.73287880e-01 -1.33395195e+00 -3.22616249e-01 3.73143822e-01
-5.89029193e-01 -1.05732536e+00 -2.18357131e-01 -4.92667317e-01
1.76991478e-01 1.97254300e-01 1.08921897e+00 2.19470695e-01
4.16657180e-02 7.96519101e-01 -5.47909021e-01 -1.28292874e-01
-8.66139233e-01 2.15523481e-01 2.22524777e-02 -6.68400228e-01
2.36705184e-01 -4.77990657e-01 -5.17719209e-01 -1.19425230e-01
-9.67413008e-01 1.00963704e-01 4.34522927e-01 9.25259590e-01
-4.55047786e-02 1.03117533e-01 4.50967610e-01 -7.86495924e-01
8.75982761e-01 -2.63313353e-01 -6.70022905e-01 2.25599557e-01
-4.39285129e-01 8.12575042e-01 9.85288262e-01 -3.50840002e-01
-9.53833222e-01 -2.91864514e-01 -1.92419160e-02 -5.94999716e-02
-6.37121648e-02 3.30453664e-01 9.59214494e-02 1.10992230e-01
7.60772765e-01 6.31754816e-01 1.86821029e-01 -7.80184269e-02
8.74542356e-01 4.68972236e-01 2.50093788e-01 -1.00105870e+00
6.54900789e-01 5.21296740e-01 2.37451550e-02 -6.70028985e-01
-1.04557908e+00 -3.05309623e-01 -1.15206979e-01 2.12733045e-01
8.52618575e-01 -8.58502805e-01 -1.28713465e+00 1.68194711e-01
-1.13079262e+00 -9.09570575e-01 -5.75332582e-01 3.30975682e-01
-1.21652699e+00 6.42132878e-01 -9.86707032e-01 -9.75814044e-01
-3.31564754e-01 -1.11120701e+00 1.35271108e+00 -9.83701125e-02
-2.01317310e-01 -8.71470749e-01 1.25902072e-01 6.97670877e-01
1.85039923e-01 -3.34839046e-01 1.02537000e+00 -1.07069886e+00
-8.97700727e-01 8.60039517e-02 -6.04758272e-03 -9.91047025e-02
-2.14341432e-01 -7.11435556e-01 -1.16146338e+00 -3.98936272e-01
1.11721143e-01 -1.06366837e+00 8.14674199e-01 6.94010183e-02
5.42003512e-01 -8.70872259e-01 1.57367557e-01 3.22289646e-01
1.19620097e+00 2.71861367e-02 5.72092056e-01 5.40527463e-01
4.83834118e-01 5.17804921e-01 4.09031481e-01 5.75700700e-01
8.60036612e-01 5.20188332e-01 3.60749215e-01 4.76387441e-01
-1.13506811e-02 -8.01985860e-01 6.53257072e-01 5.83733320e-01
2.42070541e-01 -4.92654681e-01 -6.46977127e-01 3.27338934e-01
-2.03436542e+00 -1.09844482e+00 4.64206636e-01 1.86354053e+00
9.55023348e-01 2.02057973e-01 1.79505557e-01 -1.33580536e-01
1.04483724e-01 4.82118160e-01 -5.87252855e-01 -3.29696268e-01
2.77034137e-02 8.77639651e-02 5.93009472e-01 7.65365839e-01
-7.45626569e-01 1.44686699e+00 5.87523127e+00 1.05042934e+00
-6.59668803e-01 2.32306495e-01 2.12865189e-01 -1.13109894e-01
-4.46146250e-01 1.07674532e-01 -5.97588837e-01 1.55331075e-01
1.14984453e+00 1.14182383e-01 1.11311233e+00 5.18839121e-01
2.73509286e-02 -2.18555182e-01 -1.15843070e+00 7.50028431e-01
6.26007766e-02 -1.33198667e+00 1.15808561e-01 1.84067863e-03
3.68499845e-01 3.81399751e-01 -9.52003822e-02 7.62234867e-01
1.02946270e+00 -8.62409890e-01 8.03194880e-01 1.32421166e-01
4.00222808e-01 -5.20410240e-01 4.34670955e-01 7.77795792e-01
-9.01248455e-01 -1.65066242e-01 -4.32674348e-01 -4.39185411e-01
-4.59290259e-02 -3.21681380e-01 -1.00461638e+00 4.77006942e-01
2.48068213e-01 2.32436746e-01 -3.40270549e-01 3.44730467e-01
-4.10825551e-01 4.23579723e-01 -5.08695841e-01 -5.06125033e-01
6.99931681e-01 6.39585927e-02 6.41287446e-01 8.20579052e-01
-9.67321470e-02 2.50508845e-01 2.66157985e-01 7.80804813e-01
-1.69753775e-01 -2.09557250e-01 -7.35348642e-01 -4.57155794e-01
3.27691108e-01 8.20219398e-01 -6.67152762e-01 -4.45903301e-01
-3.92296702e-01 1.21165740e+00 7.25274205e-01 4.36367065e-01
-6.00777626e-01 5.12029715e-02 5.45698106e-01 -4.02745724e-01
3.24965656e-01 -3.71180773e-01 4.58555557e-02 -1.43501019e+00
-7.90364891e-02 -1.27720749e+00 5.84418416e-01 -7.98605621e-01
-7.35864401e-01 2.15816975e-01 -2.51032919e-01 -6.32701099e-01
-6.85246825e-01 -5.96562386e-01 -2.32471511e-01 1.93988115e-01
-1.67810953e+00 -1.11832654e+00 3.14743310e-01 5.33754766e-01
7.82130480e-01 9.36066434e-02 8.52602303e-01 -3.95195276e-01
-1.95221871e-01 5.02449811e-01 2.15747848e-01 6.29433319e-02
2.72112042e-01 -1.42645526e+00 7.18260705e-01 5.89924514e-01
5.49081445e-01 4.53066438e-01 9.15218353e-01 -5.94012558e-01
-2.09665203e+00 -7.12217450e-01 8.25399041e-01 -8.83019626e-01
8.99128735e-01 -5.38987339e-01 -5.63924015e-01 9.23549354e-01
2.56186247e-01 -2.47515708e-01 2.30197445e-01 -4.96412395e-03
-5.51210165e-01 4.73623097e-01 -9.99905944e-01 9.79580283e-01
1.14647031e+00 -9.75655496e-01 -8.43524396e-01 4.38076496e-01
1.28275526e+00 -3.27298641e-01 -4.33953196e-01 -1.96137905e-01
4.57632303e-01 -7.90739655e-01 8.39127898e-01 -9.80599642e-01
3.37578237e-01 -1.37256309e-01 -1.60802558e-01 -1.22603226e+00
1.20261461e-01 -1.18644643e+00 -3.60192329e-01 5.84542692e-01
2.49946639e-01 -6.34374261e-01 9.71750021e-01 5.16555250e-01
7.64577985e-02 -5.27715683e-01 -1.10588026e+00 -7.23650992e-01
1.64816186e-01 -7.43148863e-01 2.82219589e-01 8.38412762e-01
6.12308443e-01 6.76209867e-01 -4.68588293e-01 9.48908832e-03
4.87290651e-01 -3.99515107e-02 7.50966072e-01 -8.45589399e-01
-6.42452598e-01 -4.63136762e-01 -1.44595370e-01 -1.38069499e+00
6.05499148e-01 -8.75871122e-01 1.45628631e-01 -1.48632157e+00
9.30334032e-02 -1.06853649e-01 4.77694608e-02 5.25330365e-01
-1.55671656e-01 -1.79752633e-01 5.32776952e-01 8.60979259e-02
-1.25421584e+00 9.08152223e-01 1.31323397e+00 -1.47796884e-01
-7.46892616e-02 -2.62456089e-01 -8.87554705e-01 5.92767358e-01
7.51036704e-01 -4.31686819e-01 -6.66050255e-01 -5.27156532e-01
8.15129876e-01 3.45024824e-01 4.51951057e-01 -6.06579244e-01
6.27462387e-01 -2.57794708e-02 -2.34514132e-01 -1.63099188e-02
4.36581433e-01 -8.39179039e-01 -5.79865396e-01 5.40483952e-01
-7.44624555e-01 -1.23928137e-01 1.72431365e-01 8.64321887e-01
3.01708043e-01 -4.18194950e-01 3.07288468e-01 -5.83634794e-01
-5.72095990e-01 2.10483521e-02 -7.34908998e-01 5.99233568e-01
5.77880800e-01 7.96398427e-03 -6.01866841e-01 -9.04219925e-01
-5.73843539e-01 2.62134194e-01 4.01101053e-01 1.33240849e-01
4.19774920e-01 -7.76285887e-01 -5.58542848e-01 -2.05170177e-02
6.28645793e-02 -9.23662111e-02 1.26211062e-01 5.52906513e-01
-2.70262241e-01 7.18453288e-01 1.67241096e-01 7.92639032e-02
-7.07237065e-01 6.78662419e-01 4.70396012e-01 -6.70442641e-01
-8.11663866e-01 6.54547095e-01 3.27409893e-01 -6.98193669e-01
4.11517322e-01 -4.95969176e-01 1.25556663e-01 -3.97883132e-02
5.01544476e-01 2.69300908e-01 -1.32299379e-01 -2.49543265e-01
3.05568036e-02 2.16368139e-01 -1.72647849e-01 -4.66325790e-01
1.07505369e+00 -3.47770602e-01 3.45736980e-01 3.06895524e-01
1.00705814e+00 -2.23317057e-01 -1.40723014e+00 -5.00001252e-01
6.26842007e-02 -2.98936456e-03 -4.98432517e-02 -8.82407784e-01
-4.92766291e-01 7.66787708e-01 8.60992670e-02 3.73683751e-01
6.37211263e-01 2.67295241e-01 1.11356843e+00 1.19829082e+00
5.55264890e-01 -1.08410227e+00 2.43242636e-01 8.62325788e-01
4.12600219e-01 -1.15610218e+00 -1.44677326e-01 2.51193225e-01
-8.11509013e-01 9.33958530e-01 4.01687205e-01 5.69752976e-02
-5.03678210e-02 -2.24120878e-02 -1.76658392e-01 -4.26141173e-01
-1.18569672e+00 -6.38185143e-01 -2.05448806e-01 5.73550940e-01
-1.58905566e-01 5.83066465e-03 1.10712118e-01 3.51926565e-01
-5.35033882e-01 -1.20166034e-01 5.91058671e-01 1.14636159e+00
-5.98440230e-01 -1.02309215e+00 -1.65245265e-01 2.30670080e-01
-5.10861337e-01 -4.59408820e-01 -3.58506173e-01 9.47774589e-01
-4.59772974e-01 9.87700522e-01 -1.32854849e-01 -2.78118908e-01
4.45130169e-02 3.40422153e-01 7.92199135e-01 -3.75537038e-01
-7.25826323e-01 -1.86909437e-01 4.41016376e-01 -7.15714753e-01
-1.39356732e-01 -4.08968002e-01 -1.50860047e+00 -5.04524350e-01
-1.87801152e-01 1.31494150e-01 4.35568005e-01 1.42197084e+00
3.43104780e-01 2.63260335e-01 1.63764015e-01 -5.94438553e-01
-1.11592448e+00 -6.44560099e-01 -2.81267345e-01 3.40261728e-01
7.08328366e-01 -4.05750215e-01 -5.95954657e-01 -2.34559774e-01]
|
[3.8234121799468994, 1.38559091091156]
|
69b23e40-0cc4-4ddc-a812-4b8de38c39c3
|
extracting-and-modeling-durations-for-habits
| null | null |
https://aclanthology.org/P12-2044
|
https://aclanthology.org/P12-2044.pdf
|
Extracting and modeling durations for habits and events from Twitter
| null |
['Graham Katz', 'Jennifer Williams']
|
2012-07-01
| null | null | null |
acl-2012-7
|
['game-of-chess']
|
['playing-games']
|
[-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]
|
[-7.466550350189209, 3.7846498489379883]
|
8f60f843-373b-4b85-89f1-7a8294640011
|
srp-efficient-class-aware-embedding-learning
|
1811.03166
| null |
http://arxiv.org/abs/1811.03166v1
|
http://arxiv.org/pdf/1811.03166v1.pdf
|
SRP: Efficient class-aware embedding learning for large-scale data via supervised random projections
|
Supervised dimensionality reduction strategies have been of great interest.
However, current supervised dimensionality reduction approaches are difficult
to scale for situations characterized by large datasets given the high
computational complexities associated with such methods. While stochastic
approximation strategies have been explored for unsupervised dimensionality
reduction to tackle this challenge, such approaches are not well-suited for
accelerating computational speed for supervised dimensionality reduction.
Motivated to tackle this challenge, in this study we explore a novel direction
of directly learning optimal class-aware embeddings in a supervised manner via
the notion of supervised random projections (SRP). The key idea behind SRP is
that, rather than performing spectral decomposition (or approximations thereof)
which are computationally prohibitive for large-scale data, we instead perform
a direct decomposition by leveraging kernel approximation theory and the
symmetry of the Hilbert-Schmidt Independence Criterion (HSIC) measure of
dependence between the embedded data and the labels. Experimental results on
five different synthetic and real-world datasets demonstrate that the proposed
SRP strategy for class-aware embedding learning can be very promising in
producing embeddings that are highly competitive with existing supervised
dimensionality reduction methods (e.g., SPCA and KSPCA) while achieving 1-2
orders of magnitude better computational performance. As such, such an
efficient approach to learning embeddings for dimensionality reduction can be a
powerful tool for large-scale data analysis and visualization.
|
['Ali Ghodsi', 'Amir-Hossein Karimi', 'Alexander Wong']
|
2018-11-07
| null | null | null | null |
['supervised-dimensionality-reduction']
|
['computer-vision']
|
[ 2.30994359e-01 -1.96417391e-01 9.00228880e-03 -1.71266437e-01
-6.68727398e-01 -4.76786822e-01 7.32211947e-01 2.72599105e-02
-3.49831074e-01 4.43204582e-01 2.48956829e-01 -2.29041219e-01
-5.59707582e-01 -7.16582417e-01 -1.44846290e-01 -1.19596159e+00
-6.84095128e-03 4.01801199e-01 -8.34122449e-02 1.15878791e-01
2.78501391e-01 7.27790952e-01 -1.78881109e+00 -2.45557100e-01
7.93184519e-01 8.47156107e-01 -8.45324993e-02 4.55217391e-01
-1.99522376e-01 2.81267583e-01 -1.83523595e-02 -1.09063750e-02
4.38590616e-01 -4.62855846e-01 -5.43458581e-01 2.65642136e-01
8.81763324e-02 -8.69583059e-03 -3.12718511e-01 1.15952528e+00
4.71334577e-01 3.28757554e-01 9.12475944e-01 -1.11577272e+00
-4.76906627e-01 1.14357129e-01 -5.21633804e-01 7.76625276e-02
3.53533626e-02 -1.72178581e-01 1.11989462e+00 -1.19838381e+00
4.06804085e-01 1.09991431e+00 5.03471494e-01 4.32787001e-01
-1.60962117e+00 -2.97772527e-01 -2.20701680e-01 1.59905598e-01
-1.54464591e+00 -4.86981630e-01 1.26928580e+00 -5.48198760e-01
6.06055379e-01 3.32413465e-01 3.90313685e-01 8.45766544e-01
-1.65957138e-01 5.75419426e-01 1.29951358e+00 -5.36158025e-01
6.69038534e-01 4.87863004e-01 2.57232487e-01 5.07368028e-01
3.98165315e-01 -3.68260965e-02 -3.77525270e-01 -3.79185319e-01
4.37362999e-01 3.31530511e-01 -9.42793563e-02 -1.20422113e+00
-1.39467764e+00 1.15698218e+00 1.41493350e-01 1.58209413e-01
-4.32060838e-01 -7.42612928e-02 5.92705429e-01 1.45753086e-01
5.16564071e-01 4.30270374e-01 -5.38655892e-02 -3.09590667e-01
-8.59596670e-01 1.93068668e-01 6.77152634e-01 5.20829380e-01
8.54122818e-01 1.00722156e-01 3.53489399e-01 8.21932673e-01
2.82215983e-01 4.71075028e-01 7.63629913e-01 -8.78804922e-01
4.32625532e-01 7.10113168e-01 1.83577210e-01 -1.33320391e+00
-4.80956644e-01 -9.96954441e-02 -1.04977775e+00 2.11705327e-01
5.60072720e-01 1.55522555e-01 -4.46908802e-01 1.33196914e+00
6.24309123e-01 3.66981998e-02 3.64736378e-01 9.02389586e-01
1.44100815e-01 6.16919816e-01 -2.31773242e-01 -3.85310382e-01
1.31025350e+00 -6.03030562e-01 -6.91393971e-01 1.15070812e-01
9.50028121e-01 -4.16349053e-01 1.23479831e+00 4.29168940e-01
-5.23111343e-01 -2.62402087e-01 -1.10271597e+00 6.91631362e-02
-4.27525580e-01 1.44568831e-01 8.34538877e-01 8.29389691e-01
-6.14099324e-01 5.81823170e-01 -1.13329101e+00 -5.34273148e-01
4.38160390e-01 2.82519370e-01 -5.41247427e-01 -2.01160207e-01
-7.97218978e-01 5.17236531e-01 4.11430091e-01 2.14559242e-01
-4.26265538e-01 -5.03832281e-01 -8.99402201e-01 9.37578976e-02
2.74066985e-01 -1.44564249e-02 5.26927412e-01 -2.89657116e-01
-1.51577103e+00 6.14435911e-01 -1.36259854e-01 -4.54771340e-01
1.66435003e-01 -1.74309209e-01 -2.39993989e-01 3.34454745e-01
-1.03527695e-01 1.83223292e-01 1.13991022e+00 -1.01817477e+00
-2.02525407e-01 -6.02234483e-01 -2.79489368e-01 1.72371566e-01
-1.24701095e+00 -7.87953809e-02 7.11575523e-02 -5.65444112e-01
4.49375391e-01 -1.18638563e+00 -3.59057426e-01 1.68852419e-01
-1.32893249e-01 -1.52829751e-01 1.18311894e+00 -4.11177009e-01
1.31874955e+00 -2.36529255e+00 3.84157211e-01 2.57158309e-01
2.74727613e-01 5.42991221e-01 9.63792205e-02 6.87736928e-01
-2.45518014e-01 -6.02627955e-02 -5.11955917e-01 -4.29531425e-01
1.41388595e-01 1.63232461e-01 -4.81593013e-01 8.16392899e-01
1.78281486e-01 6.33589804e-01 -9.07933474e-01 -2.55674452e-01
4.37483311e-01 6.53875411e-01 -5.39163828e-01 1.48148596e-01
2.17069566e-01 2.77556717e-01 -4.04934108e-01 2.99242198e-01
5.37076771e-01 -2.06120983e-01 1.91224366e-01 -1.66529283e-01
-1.05400115e-01 1.28963098e-01 -1.65120924e+00 1.39059949e+00
-5.39994001e-01 6.24847054e-01 -6.80419803e-02 -1.55533934e+00
1.08626485e+00 2.16692388e-01 6.32343590e-01 -1.93962455e-01
-7.61431232e-02 2.85469383e-01 -1.18862994e-01 -3.13194096e-01
2.43298888e-01 -4.06435907e-01 -1.04124270e-01 6.46245480e-01
-5.85077144e-02 1.51876509e-02 2.99344093e-01 9.99205858e-02
1.03639543e+00 -1.25734553e-01 5.39038002e-01 -3.52582663e-01
7.03317583e-01 -6.62464201e-02 5.31112909e-01 2.17220604e-01
-2.45574102e-01 6.21301591e-01 6.21671677e-01 -5.76120913e-01
-1.09768963e+00 -9.37157154e-01 -2.08794460e-01 6.36448562e-01
-1.54754922e-01 -5.26249111e-01 -6.45267129e-01 -6.88733578e-01
-7.04417676e-02 4.31233108e-01 -4.65040386e-01 -2.40509197e-01
-4.29372072e-01 -1.03861880e+00 2.20388576e-01 5.26788414e-01
3.04086208e-01 -6.39988065e-01 -4.69542027e-01 1.36863425e-01
2.43942201e-01 -1.12465310e+00 -1.32558405e-01 2.37324625e-01
-1.14044189e+00 -9.75395799e-01 -6.49812579e-01 -4.72685367e-01
8.58218014e-01 5.71309209e-01 2.85018206e-01 -4.40507501e-01
-4.39159900e-01 4.74398643e-01 -4.09731865e-01 -7.64727667e-02
-2.34175488e-01 -5.80489747e-02 6.82134628e-01 5.39643943e-01
7.33769774e-01 -8.39134991e-01 -5.14408827e-01 3.01174104e-01
-1.15904319e+00 -2.68777311e-01 5.04353940e-01 1.02858448e+00
6.64952636e-01 5.73770523e-01 4.86108571e-01 -8.30863178e-01
7.58169055e-01 -4.04329836e-01 -7.92490721e-01 2.56254594e-03
-1.02452219e+00 2.29737833e-01 1.08878219e+00 -4.77283061e-01
-8.45380306e-01 2.58295268e-01 2.49877140e-01 -5.93631208e-01
-1.67588636e-01 4.11022514e-01 -1.79820374e-01 -1.11953907e-01
6.62109196e-01 5.25495768e-01 2.94915944e-01 -7.26453006e-01
4.56788868e-01 9.44146574e-01 1.13513179e-01 -3.27821493e-01
1.26672935e+00 8.52965832e-01 3.71181697e-01 -1.18635976e+00
-6.00001752e-01 -8.05102348e-01 -9.32733536e-01 3.03430021e-01
8.76415968e-01 -7.10628808e-01 -2.65782416e-01 1.51268333e-01
-6.40797853e-01 3.33142757e-01 -5.32291472e-01 7.90534675e-01
-5.64482629e-01 6.91015244e-01 -2.13604927e-01 -1.00718713e+00
-2.27633063e-02 -9.63306725e-01 8.45025718e-01 -6.20211624e-02
-1.94535464e-01 -1.16978049e+00 3.14828396e-01 3.00414979e-01
3.73961568e-01 2.00326592e-01 9.72289681e-01 -8.49121630e-01
-1.83387220e-01 -5.22831023e-01 -2.59783268e-01 6.77537918e-01
3.42417687e-01 -1.82203263e-01 -8.99197996e-01 -4.70966011e-01
3.23981911e-01 -2.67235726e-01 6.49650931e-01 -1.05072744e-02
1.04274929e+00 -1.57245100e-01 -2.18264654e-01 6.09091461e-01
1.39336991e+00 -1.90358087e-01 3.61272365e-01 2.04552710e-01
8.90478253e-01 7.30551481e-01 6.50981486e-01 6.11805260e-01
1.41975194e-01 8.33606541e-01 7.29308724e-02 1.91929370e-01
6.24296628e-03 -3.20437998e-01 3.31618279e-01 1.18953180e+00
2.09496059e-02 2.96213835e-01 -1.05597901e+00 6.03306651e-01
-1.72923434e+00 -7.17634976e-01 -1.48449078e-01 2.53019881e+00
4.40205246e-01 -5.43943532e-02 2.87143290e-01 8.00784409e-01
5.67948878e-01 2.50988662e-01 -3.69522005e-01 -3.98002863e-01
-2.07334459e-02 -7.66456947e-02 2.84439206e-01 2.42550805e-01
-1.01296842e+00 6.43629551e-01 5.56141520e+00 7.14653075e-01
-9.94482100e-01 6.74110502e-02 1.18024394e-01 1.16800833e-02
-1.08460218e-01 1.87056974e-01 -8.93736184e-01 5.84030807e-01
1.07377863e+00 -2.65327208e-02 4.80176657e-01 1.00899863e+00
3.38433683e-01 1.14792734e-01 -1.05392075e+00 1.31054103e+00
3.98107804e-02 -1.19352818e+00 7.22796768e-02 4.17503297e-01
6.54385030e-01 -2.89733618e-01 1.31800890e-01 2.66999137e-02
-1.37097493e-01 -6.91770732e-01 4.61618751e-02 2.56362081e-01
5.93099892e-01 -7.86553264e-01 5.29114008e-01 3.67519379e-01
-1.12506437e+00 -2.72571504e-01 -7.27024496e-01 -2.22183540e-01
3.43959145e-02 9.55115616e-01 -5.38400829e-01 2.03360528e-01
4.76998925e-01 8.76732349e-01 -5.33867478e-01 7.46016979e-01
-1.01271719e-01 6.75065219e-01 -4.90974754e-01 -1.34015009e-02
1.80026457e-01 -6.81182027e-01 6.79258645e-01 8.75944436e-01
2.76605457e-01 -3.44187543e-02 -1.10315904e-01 6.66011751e-01
2.10170448e-01 2.02393010e-01 -9.04992998e-01 -6.52647614e-01
3.68881673e-01 1.32782924e+00 -1.01810813e+00 -1.17480285e-01
-6.75408840e-01 9.50074375e-01 4.54631001e-01 3.68175805e-01
-4.13057387e-01 -6.55485332e-01 8.84851336e-01 1.20933905e-01
6.04375362e-01 -7.41627216e-01 -1.77146524e-01 -1.32667232e+00
2.63713717e-01 -7.01043069e-01 2.64604837e-01 -6.70714155e-02
-1.15161180e+00 3.86079758e-01 1.54673159e-02 -1.67470098e+00
-2.90194452e-01 -7.86978185e-01 -3.68377239e-01 5.04750729e-01
-1.59271610e+00 -9.32295144e-01 8.52296967e-03 5.05532444e-01
3.01605344e-01 -2.40440279e-01 1.14241219e+00 1.75097004e-01
-6.87000394e-01 3.07840317e-01 8.47749829e-01 -1.05650134e-01
4.68448877e-01 -1.28921211e+00 2.88980510e-02 8.74681234e-01
4.76880878e-01 6.72230422e-01 6.20535374e-01 -2.34543115e-01
-1.86452854e+00 -8.60317290e-01 6.16630077e-01 -3.35398346e-01
1.00918376e+00 -7.54361272e-01 -9.62060392e-01 1.55585736e-01
-4.98340845e-01 5.22368729e-01 1.06296074e+00 1.17546424e-01
-3.95016670e-01 -2.86295474e-01 -9.54871178e-01 4.86015260e-01
8.61475945e-01 -6.78723991e-01 -4.45095390e-01 4.17910039e-01
3.88855994e-01 2.45936573e-01 -1.07640159e+00 1.17073931e-01
3.62193197e-01 -9.21948075e-01 9.43272293e-01 -5.24670005e-01
2.60532379e-01 -3.61399680e-01 -3.46710891e-01 -1.22916472e+00
-1.56218678e-01 -6.73477650e-01 -5.22544682e-01 1.30887854e+00
-7.56549016e-02 -8.52248549e-01 9.63844955e-01 6.85265541e-01
4.52446342e-01 -7.98213243e-01 -1.02509499e+00 -1.17777371e+00
4.23888974e-02 -4.07710105e-01 3.09260130e-01 9.44069803e-01
2.31104285e-01 1.42602101e-01 -3.93650532e-01 2.33499423e-01
9.41331387e-01 2.99221605e-01 8.54111731e-01 -1.35014617e+00
-2.91228443e-01 -1.85440853e-01 -9.40326393e-01 -8.41202378e-01
2.15352386e-01 -8.78056288e-01 -1.97805002e-01 -1.19421816e+00
1.85361550e-05 -5.41496158e-01 -3.50561172e-01 1.08884022e-01
-1.13617571e-04 2.64162272e-01 8.32414255e-02 5.40379524e-01
-3.91601890e-01 8.79643261e-01 6.89360678e-01 9.53411981e-02
-3.55329663e-01 -6.16688728e-02 -6.19580746e-01 7.46611297e-01
5.00999928e-01 -6.26494408e-01 -6.45763099e-01 -8.07015523e-02
1.54956043e-01 -2.24326506e-01 2.02113077e-01 -9.91381645e-01
2.05531016e-01 -4.71003465e-02 -4.77301702e-02 -3.05123389e-01
3.28674376e-01 -9.57273126e-01 -9.54119116e-02 3.16759437e-01
-2.46780246e-01 -2.32419103e-01 -2.09677607e-01 9.79441166e-01
-4.49065119e-01 -2.34933123e-01 7.42309928e-01 1.92996681e-01
-5.69237411e-01 1.09020308e-01 -3.18806112e-01 -6.28020391e-02
1.12503028e+00 -2.96475798e-01 2.56398261e-01 -1.41448230e-01
-5.55385053e-01 -1.35666773e-01 5.39371789e-01 3.14050406e-01
7.75464773e-01 -1.54607308e+00 -4.63572383e-01 3.91981483e-01
3.18611085e-01 -1.01676621e-01 2.11304333e-02 1.00946534e+00
-2.64725059e-01 6.27047837e-01 1.32353464e-02 -6.11199260e-01
-1.15998435e+00 9.07746732e-01 -1.91997781e-01 -3.73101652e-01
-9.10208941e-01 4.85127360e-01 1.58554390e-01 -5.47277331e-01
9.28877220e-02 -9.40107778e-02 -1.87614605e-01 2.79569089e-01
7.35872984e-01 6.45776689e-01 4.70283441e-02 -6.52998269e-01
-3.23175758e-01 7.18545020e-01 -1.01015717e-01 5.33377528e-02
1.61315715e+00 -3.26622576e-01 -6.59438446e-02 5.51086783e-01
1.68651712e+00 -5.43450527e-02 -1.28077066e+00 -4.24473315e-01
2.26972729e-01 -6.86436594e-01 3.20262492e-01 1.23053662e-01
-8.53981674e-01 1.18985438e+00 6.22716904e-01 4.15060192e-01
1.07521808e+00 -1.49146885e-01 7.05402434e-01 6.95380211e-01
3.45730692e-01 -1.16328716e+00 5.74732386e-02 6.57221228e-02
5.48450530e-01 -1.35853589e+00 2.60510743e-01 -4.21527863e-01
-4.48736608e-01 1.18802321e+00 4.03129645e-02 -3.06182474e-01
8.42457891e-01 -1.13248311e-01 -2.24188671e-01 -2.36217469e-01
-3.97274077e-01 -1.17760547e-01 1.85119212e-01 5.76333463e-01
9.34758484e-02 -4.01495723e-03 -4.38583821e-01 2.47926965e-01
5.33424132e-02 -2.55183548e-01 5.73444307e-01 9.59552944e-01
-1.82595223e-01 -1.26235831e+00 -3.78865600e-01 4.39527303e-01
-1.20250814e-01 1.20332949e-01 -1.19264178e-01 5.58743179e-01
-4.73424584e-01 5.85899115e-01 -1.91062421e-01 -2.85133570e-01
6.89776093e-02 2.56647080e-01 1.47197381e-01 -7.45772898e-01
2.50468552e-01 -4.97668758e-02 -2.10684270e-01 -5.97731650e-01
-4.72854435e-01 -1.04558969e+00 -1.08712566e+00 9.64152813e-02
-3.12132597e-01 1.76754475e-01 8.66091609e-01 9.07874882e-01
4.16817099e-01 -1.30140215e-01 1.00039685e+00 -9.26543951e-01
-1.00443220e+00 -6.48338854e-01 -1.02558374e+00 6.36011064e-01
1.47631958e-01 -9.48048294e-01 -8.49103212e-01 -8.64082500e-02]
|
[7.8506855964660645, 4.176077365875244]
|
b4233166-0ca2-4fae-ba6a-9b0ca6623e29
|
using-reinforcement-learning-to-learn-how-to
|
1801.01999
| null |
http://arxiv.org/abs/1801.01999v1
|
http://arxiv.org/pdf/1801.01999v1.pdf
|
Using reinforcement learning to learn how to play text-based games
|
The ability to learn optimal control policies in systems where action space
is defined by sentences in natural language would allow many interesting
real-world applications such as automatic optimisation of dialogue systems.
Text-based games with multiple endings and rewards are a promising platform for
this task, since their feedback allows us to employ reinforcement learning
techniques to jointly learn text representations and control policies. We
present a general text game playing agent, testing its generalisation and
transfer learning performance and showing its ability to play multiple games at
once. We also present pyfiction, an open-source library for universal access to
different text games that could, together with our agent that implements its
interface, serve as a baseline for future research.
|
['Mikuláš Zelinka']
|
2018-01-06
| null | null | null | null |
['text-based-games']
|
['playing-games']
|
[-1.62741859e-02 4.42687660e-01 -2.67964542e-01 -2.74720013e-01
-5.15642524e-01 -8.66341650e-01 1.18435800e+00 -5.88595010e-02
-7.97657490e-01 1.08804488e+00 4.00070488e-01 -6.50568128e-01
-6.18615039e-02 -8.75496089e-01 -2.60999173e-01 -3.57382327e-01
-1.92187771e-01 9.76565301e-01 5.32241046e-01 -1.02416837e+00
3.45644027e-01 2.18511552e-01 -1.51431739e+00 3.34568679e-01
5.39309204e-01 3.22573155e-01 5.34175873e-01 1.13453734e+00
-1.35614410e-01 1.19972610e+00 -7.60851622e-01 -2.14435443e-01
3.01685780e-01 -4.07217920e-01 -1.15142083e+00 -1.74768373e-01
-1.97272658e-01 -3.89187694e-01 -3.47533643e-01 7.71662056e-01
6.05765998e-01 7.44916022e-01 4.84007865e-01 -9.72217739e-01
-2.41630316e-01 6.32016778e-01 4.08191532e-02 1.67684779e-01
6.58256531e-01 5.75472832e-01 1.36625886e+00 4.10380885e-02
6.80478573e-01 1.53137887e+00 2.00536307e-02 1.12974107e+00
-1.11640048e+00 -2.42997304e-01 1.26726236e-02 5.99151775e-02
-4.60452467e-01 -4.45719957e-01 2.97248095e-01 -6.30156472e-02
1.73425913e+00 3.75738263e-01 7.43589342e-01 1.26643682e+00
4.48253810e-01 1.25267029e+00 1.17124903e+00 -6.02108777e-01
2.91513354e-01 2.03024540e-02 -4.02651906e-01 8.53693545e-01
-4.19336259e-01 5.21930039e-01 -2.37337485e-01 -2.20840096e-01
9.04400468e-01 -5.06383538e-01 7.10355043e-02 -3.75758886e-01
-1.02119112e+00 1.44668841e+00 1.30562454e-01 2.26548344e-01
-3.22669983e-01 2.96950638e-01 7.73486197e-01 8.81054461e-01
1.41605809e-01 1.20275402e+00 -8.34884703e-01 -8.56499732e-01
-2.45014414e-01 9.14007306e-01 1.28556705e+00 5.74298203e-01
4.06389266e-01 1.32412910e-01 -4.43554342e-01 1.08110404e+00
1.83585584e-01 1.82398736e-01 1.01125908e+00 -1.20774126e+00
4.23322082e-01 4.38985109e-01 2.43491203e-01 -3.12352538e-01
-7.54766345e-01 2.65201956e-01 -2.66569436e-01 5.72552502e-01
5.57321846e-01 -7.11482465e-01 -4.12455618e-01 1.64368439e+00
2.00705484e-01 -1.32890806e-01 4.21720535e-01 6.72678113e-01
4.95984614e-01 7.13637471e-01 7.36578405e-02 -1.11274980e-01
1.37367904e+00 -8.80568326e-01 -4.12344366e-01 -4.98896718e-01
9.61237907e-01 -4.41437960e-01 1.23405755e+00 4.67811853e-01
-1.07904720e+00 -3.09873614e-02 -7.87835896e-01 1.81880042e-01
-4.93499219e-01 -3.66904885e-01 8.41940105e-01 6.17224932e-01
-1.18582487e+00 7.73568690e-01 -8.35819840e-01 -4.55558509e-01
-5.71625270e-02 7.34546602e-01 -1.29380152e-01 3.76682341e-01
-1.47607291e+00 1.45014191e+00 7.62491882e-01 -4.94349658e-01
-5.99893749e-01 3.09891123e-02 -1.18846786e+00 -3.36596221e-02
8.96374285e-01 -4.71831679e-01 2.10750580e+00 -6.39172196e-01
-2.41277504e+00 8.07420015e-01 4.18748409e-01 -7.60874331e-01
5.33850133e-01 -9.13086087e-02 -8.41366798e-02 6.61352091e-03
8.60021263e-02 7.70690084e-01 6.29054189e-01 -3.75011861e-01
-9.23599243e-01 -1.33961082e-01 6.91736579e-01 7.83596516e-01
-2.09634930e-01 2.62410134e-01 -1.86435327e-01 -3.84779751e-01
-7.43288457e-01 -8.46539021e-01 -5.29296696e-01 -3.71216089e-01
-2.43909031e-01 -7.02181160e-01 4.72676963e-01 -2.78831244e-01
9.74767387e-01 -1.48486936e+00 3.56797338e-01 -7.81306028e-02
-8.51691328e-03 4.28294092e-01 -2.76230603e-01 7.70941854e-01
1.19669028e-01 1.14164934e-01 1.05774209e-01 8.20143670e-02
3.31045151e-01 5.61810911e-01 9.93549377e-02 8.34815279e-02
4.63796735e-01 1.07951295e+00 -1.21873784e+00 -3.55118215e-01
6.36505485e-01 -1.85737103e-01 -5.90574205e-01 4.86403972e-01
-9.14450228e-01 3.99256051e-01 -7.16917455e-01 -1.75140232e-01
-3.25024962e-01 3.82116586e-02 4.21407729e-01 7.88929045e-01
-1.00207049e-02 6.98926210e-01 -1.17918360e+00 1.61227477e+00
-5.81991017e-01 5.51039338e-01 1.00527205e-01 -9.11424458e-01
7.49190509e-01 3.70905221e-01 4.26226974e-01 -1.00420666e+00
1.75521865e-01 -2.88140383e-02 2.93800890e-01 -6.10005498e-01
9.52691138e-01 -1.59050271e-01 -3.91887218e-01 8.49967122e-01
2.56576985e-01 -6.08491838e-01 6.16046607e-01 1.07342631e-01
1.24829102e+00 4.12731409e-01 7.13893354e-01 -1.24175578e-01
5.04005790e-01 1.10822305e-01 5.80887822e-03 1.12687075e+00
-1.86770767e-01 -1.84467230e-02 8.06341171e-01 -5.81829488e-01
-1.07543361e+00 -5.84790409e-01 1.56255364e-01 1.93916571e+00
-3.11706632e-01 -4.69388545e-01 -5.42435765e-01 -8.24417412e-01
-1.85935706e-01 8.63773227e-01 -5.44192076e-01 -1.89623445e-01
-6.42633617e-01 -5.99389374e-01 6.14469588e-01 3.80525529e-01
4.00475770e-01 -1.91807258e+00 -8.82505894e-01 6.17939353e-01
1.11018524e-01 -8.05136204e-01 -3.16417515e-01 5.19340813e-01
-6.45924270e-01 -1.21025383e+00 -3.60119134e-01 -6.00764692e-01
-1.81272432e-01 -2.18041241e-01 1.37035096e+00 2.30919108e-01
-9.41502154e-02 6.20856643e-01 -5.23909986e-01 -4.96674120e-01
-9.93872881e-01 4.17645425e-01 -9.40755233e-02 -6.28582895e-01
2.58691847e-01 -1.07218340e-01 -2.68738121e-01 1.90278187e-01
-8.43582690e-01 1.04186997e-01 1.60578310e-01 1.27401638e+00
-3.69017839e-01 -2.50004560e-01 5.05876362e-01 -1.12684131e+00
1.62436354e+00 -3.29537421e-01 -8.75595868e-01 2.50997450e-02
-2.95357943e-01 5.70606709e-01 6.99906826e-01 -1.36569813e-01
-8.73613834e-01 -3.40069109e-03 -4.28289384e-01 3.26583415e-01
-3.48601788e-01 4.45341229e-01 1.30886003e-01 2.68574715e-01
1.07335734e+00 2.85306931e-01 4.62272912e-01 -8.16721767e-02
6.00846291e-01 6.85885310e-01 2.37912446e-01 -8.33446145e-01
3.06658268e-01 -2.94059753e-01 -3.76687169e-01 -8.79733443e-01
-3.73560220e-01 -5.35729349e-01 -2.08449438e-01 -3.08951512e-02
8.09573472e-01 -5.53419530e-01 -1.25086451e+00 1.78498566e-01
-9.49626386e-01 -1.05456340e+00 -2.28447929e-01 2.84686416e-01
-1.35137439e+00 1.60750836e-01 -7.22204983e-01 -9.93504763e-01
-2.72321343e-01 -1.19335175e+00 1.08374131e+00 4.62309182e-01
-5.31937599e-01 -1.39646268e+00 3.18853259e-01 9.12622828e-03
2.83567518e-01 -1.48519427e-01 6.27748311e-01 -1.27331245e+00
4.16230410e-02 -1.41126156e-01 2.61808187e-01 3.89075764e-02
9.93200466e-02 -2.37074852e-01 -7.84447610e-01 -4.29901332e-01
-2.72424668e-01 -1.00992429e+00 5.63272119e-01 3.52064133e-01
7.72084415e-01 -4.29966241e-01 7.79441893e-02 9.03794169e-02
1.06662917e+00 1.72649607e-01 5.24980128e-01 7.89490640e-01
1.29641086e-01 7.56061435e-01 7.01258242e-01 7.53097773e-01
3.37595582e-01 8.90838504e-01 5.08332372e-01 6.07995167e-02
4.91259843e-01 -3.03075798e-02 5.66609383e-01 3.19039732e-01
-9.31708440e-02 -4.28604841e-01 -7.67104208e-01 1.99652269e-01
-2.17815137e+00 -1.19318712e+00 3.28082144e-01 1.89262795e+00
1.05077755e+00 3.35934788e-01 6.49637759e-01 -2.01842159e-01
4.25935864e-01 2.41973013e-01 -6.13519251e-01 -1.13536716e+00
2.97833532e-01 6.95171177e-01 4.08053309e-01 7.53633559e-01
-1.21039295e+00 1.43377924e+00 7.29668808e+00 7.65587687e-01
-8.81235957e-01 -1.65922433e-01 3.23462427e-01 9.18594897e-02
1.53758138e-01 -3.42443049e-01 -5.27258813e-01 5.62307946e-02
1.17010415e+00 -4.16284233e-01 9.02947068e-01 7.73779035e-01
5.20222902e-01 -2.45546132e-01 -8.91880929e-01 3.38968903e-01
-5.00640631e-01 -1.40172529e+00 -2.29026690e-01 1.13872781e-01
4.19429958e-01 4.01580840e-01 -2.03398868e-01 9.87644255e-01
1.54038346e+00 -1.18249381e+00 1.32528886e-01 1.38151729e-02
6.28611982e-01 -7.84632206e-01 5.95242918e-01 7.45237470e-01
-8.37678432e-01 -2.57531166e-01 -4.62632567e-01 -5.28122425e-01
-2.07024112e-01 -6.55185699e-01 -1.44134915e+00 4.60065901e-01
1.03390366e-01 5.75434446e-01 -4.44151640e-01 8.82531226e-01
-3.59436929e-01 4.38470662e-01 -3.01327497e-01 -9.02027547e-01
6.25591874e-01 -2.81187370e-02 4.91711825e-01 1.16315281e+00
-2.02492729e-01 -4.17973325e-02 6.51431978e-01 5.80855966e-01
2.07030594e-01 3.78681064e-01 -1.08187497e+00 -2.16992855e-01
1.10284172e-01 1.06735384e+00 -6.80987656e-01 -2.27027819e-01
-4.22568709e-01 9.72815096e-01 5.74270725e-01 9.92767364e-02
-4.65445727e-01 -4.96164918e-01 8.83841753e-01 -4.00503159e-01
1.16354369e-01 -3.68839383e-01 1.10749088e-01 -1.21210504e+00
-4.87343967e-01 -1.34554648e+00 4.08543050e-01 -5.57762146e-01
-1.05652404e+00 5.88686109e-01 1.62156224e-01 -8.72886837e-01
-1.40258849e+00 -1.06297231e+00 -6.75686121e-01 8.32446516e-01
-1.09930289e+00 -7.06238031e-01 4.66639429e-01 7.59165764e-01
9.97819483e-01 -6.74358785e-01 1.31008232e+00 -5.78998685e-01
-3.52139592e-01 4.26395237e-01 2.72775680e-01 3.02960783e-01
4.62324142e-01 -1.92388332e+00 7.62791336e-01 1.28557786e-01
3.97385120e-01 2.62768686e-01 9.13419008e-01 -4.33973551e-01
-1.34049499e+00 -7.44014859e-01 1.76402912e-01 -6.02662623e-01
9.29573655e-01 -4.05590206e-01 -6.60496950e-01 6.06511056e-01
6.20456219e-01 -4.49987620e-01 4.08877909e-01 3.30719382e-01
2.00363487e-01 4.60916519e-01 -1.00679898e+00 9.80059683e-01
5.32164633e-01 -4.19137478e-01 -7.43999839e-01 5.30518770e-01
7.06374288e-01 -9.43286240e-01 -6.83300138e-01 -1.30980313e-01
4.16718602e-01 -1.04336727e+00 6.68566465e-01 -1.00662279e+00
5.36353171e-01 3.42901200e-01 2.86995411e-01 -1.69162548e+00
-1.71200067e-01 -1.10767341e+00 2.65165083e-02 5.93123615e-01
5.23607969e-01 -7.74714828e-01 9.36038315e-01 5.84857583e-01
1.00923985e-01 -6.85918272e-01 -8.11970353e-01 -6.02993548e-01
4.99443769e-01 -4.88624960e-01 4.63636994e-01 6.11356497e-01
5.10800004e-01 7.56739438e-01 -5.88330865e-01 -3.97103697e-01
-1.71489149e-01 -3.66465509e-01 9.27774608e-01 -1.12946916e+00
-7.86041439e-01 -8.31783652e-01 -2.09895879e-01 -1.23993635e+00
4.20257151e-01 -8.48104954e-01 3.42072934e-01 -1.27316809e+00
-3.85215700e-01 -8.94344449e-02 1.55471042e-01 6.86618924e-01
-1.90903872e-01 -1.27394497e-01 4.51251537e-01 -2.29754806e-01
-8.19344938e-01 6.48161769e-01 1.38680446e+00 -1.65604129e-01
-5.04198313e-01 3.58448893e-01 -7.16125190e-01 5.56806922e-01
1.04919744e+00 -3.35235417e-01 -5.39655089e-01 3.90099734e-02
1.16190523e-01 4.14681405e-01 -9.09113735e-02 -7.22263396e-01
1.41667739e-01 -6.73851490e-01 -3.59897874e-02 3.19191217e-01
4.47844118e-01 -3.53617221e-01 -8.04174662e-01 4.24069703e-01
-8.25287163e-01 2.08298609e-01 4.27955836e-01 3.92319709e-01
5.53147420e-02 -7.90534735e-01 4.93225574e-01 -7.45261967e-01
-8.43083799e-01 4.22566459e-02 -1.27591658e+00 4.21398312e-01
1.03740239e+00 -7.53843933e-02 -1.80401444e-01 -8.02366376e-01
-7.91700125e-01 7.98787415e-01 2.38486975e-01 6.64667904e-01
4.48371202e-01 -8.22951198e-01 -9.17844057e-01 1.30700618e-01
1.46095306e-01 -3.20315003e-01 -4.52311099e-01 1.74483061e-02
-5.45554519e-01 3.46951962e-01 -4.94100481e-01 -3.52451146e-01
-1.24636471e+00 1.41903162e-01 6.66469038e-01 -9.36840951e-01
-7.22946823e-01 5.47076762e-01 -1.28525943e-01 -1.03640509e+00
1.08597882e-01 -7.14729279e-02 -6.88592553e-01 -2.63450801e-01
6.26015604e-01 -2.13115424e-01 6.41044555e-03 -8.04097727e-02
2.28825375e-01 -2.82632917e-01 -2.88328737e-01 -6.40213132e-01
1.46104217e+00 1.16587013e-01 2.15304136e-01 5.25189936e-01
5.35890400e-01 -4.27620083e-01 -1.24193847e+00 -1.21693648e-01
3.69777262e-01 -2.11085588e-01 -2.33697176e-01 -8.90436411e-01
-4.80583996e-01 5.55919111e-01 3.69153559e-01 7.72131622e-01
7.52773583e-01 -1.81178734e-01 2.51692206e-01 1.12030137e+00
4.73036647e-01 -1.32566261e+00 2.76399761e-01 1.23101306e+00
8.77841055e-01 -1.29857874e+00 -2.39154622e-01 2.38886997e-01
-1.15243137e+00 1.40133035e+00 7.90281713e-01 -3.99745494e-01
6.64217547e-02 2.38800213e-01 1.65090457e-01 -2.17868924e-01
-1.53712404e+00 -6.47689939e-01 -4.12650891e-02 1.07672179e+00
6.47247851e-01 9.92931575e-02 -3.16953182e-01 1.06948227e-01
-4.64180738e-01 -2.56202877e-01 8.57792079e-01 1.03307950e+00
-5.15729785e-01 -1.77131283e+00 -2.77507424e-01 8.29073071e-01
-5.17049909e-01 1.50032248e-02 -6.06884181e-01 9.39157069e-01
-5.88514626e-01 1.07747006e+00 -4.41874564e-02 -2.75126368e-01
4.14769828e-01 2.26472721e-01 6.62053347e-01 -1.15656257e+00
-1.24611378e+00 -8.41421857e-02 6.47675335e-01 -5.55671751e-01
-1.28643438e-01 -5.47798395e-01 -1.34515083e+00 -3.04501534e-01
-3.41127008e-01 2.88026750e-01 3.92103195e-01 1.03675103e+00
-4.30811234e-02 5.53202987e-01 5.53770244e-01 -9.48519886e-01
-9.79311347e-01 -1.18020797e+00 -5.40351272e-01 3.06707889e-01
1.23282045e-01 -6.37523890e-01 1.84168652e-01 -3.97814393e-01]
|
[3.7843902111053467, 1.4621535539627075]
|
6655c3b2-1fba-4dc5-8e89-f203e01d3a06
|
a-simple-generative-model-of-logical
|
2305.11098
| null |
https://arxiv.org/abs/2305.11098v1
|
https://arxiv.org/pdf/2305.11098v1.pdf
|
A Simple Generative Model of Logical Reasoning and Statistical Learning
|
Statistical learning and logical reasoning are two major fields of AI expected to be unified for human-like machine intelligence. Most existing work considers how to combine existing logical and statistical systems. However, there is no theory of inference so far explaining how basic approaches to statistical learning and logical reasoning stem from a common principle. Inspired by the fact that much empirical work in neuroscience suggests Bayesian (or probabilistic generative) approaches to brain function including learning and reasoning, we here propose a simple Bayesian model of logical reasoning and statistical learning. The theory is statistically correct as it satisfies Kolmogorov's axioms, is consistent with both Fenstad's representation theorem and maximum likelihood estimation and performs exact Bayesian inference with a linear-time complexity. The theory is logically correct as it is a data-driven generalisation of uncertain reasoning from consistency, possibility, inconsistency and impossibility. The theory is correct in terms of machine learning as its solution to generation and prediction tasks on the MNIST dataset is not only empirically reasonable but also theoretically correct against the K nearest neighbour method. We simply model how data causes symbolic knowledge in terms of its satisfiability in formal logic. Symbolic reasoning emerges as a result of the process of going the causality forwards and backwards. The forward and backward processes correspond to an interpretation and inverse interpretation in formal logic, respectively. The inverse interpretation differentiates our work from the mainstream often referred to as inverse entailment, inverse deduction or inverse resolution. The perspective gives new insights into learning and reasoning towards human-like machine intelligence.
|
['Hiroyuki Kido']
|
2023-05-18
| null | null | null | null |
['bayesian-inference', 'logical-reasoning', 'formal-logic']
|
['methodology', 'reasoning', 'reasoning']
|
[ 4.24757041e-02 7.58712888e-01 -2.30499059e-01 -7.02811062e-01
-2.35370591e-01 -3.53357941e-01 1.11253619e+00 1.05116628e-01
-4.33272749e-01 9.49730098e-01 1.57057300e-01 -7.53156483e-01
-9.65582848e-01 -8.79154146e-01 -8.60250413e-01 -5.92974544e-01
7.72527456e-02 8.94378185e-01 2.89967299e-01 -8.02547857e-02
4.94564742e-01 3.21785510e-01 -1.38371515e+00 4.53734457e-01
7.55036175e-01 9.39732909e-01 6.08752854e-03 5.08084059e-01
-3.19957435e-01 1.44790161e+00 -7.12628514e-02 -6.76060259e-01
-1.54661417e-01 -6.60342872e-01 -1.28355861e+00 -4.48436171e-01
7.29306787e-02 -9.39344019e-02 -1.13801092e-01 1.13317692e+00
-9.50797871e-02 -1.87337473e-01 9.30839717e-01 -1.36290431e+00
-1.02717912e+00 1.03137851e+00 -3.08756772e-02 1.03503965e-01
4.94326413e-01 -1.43205747e-01 1.10355508e+00 -6.33336127e-01
3.53824437e-01 1.82921290e+00 7.00396359e-01 4.53031182e-01
-1.47153187e+00 -2.21650675e-01 1.30768612e-01 8.83509278e-01
-1.22426236e+00 -1.32662624e-01 4.44276512e-01 -7.03652442e-01
8.18506658e-01 2.12587118e-01 8.70414793e-01 9.19916451e-01
5.06560981e-01 8.22966397e-01 1.62312257e+00 -8.92481565e-01
8.00378859e-01 3.53126794e-01 3.38321596e-01 5.82093537e-01
5.49979746e-01 3.34694147e-01 -7.91702211e-01 -1.21566430e-01
4.50018615e-01 -2.82403678e-01 5.76339327e-02 -1.95611104e-01
-1.20794678e+00 8.86491954e-01 6.51951283e-02 3.33515167e-01
-3.83465290e-01 4.49490249e-01 1.61028981e-01 1.87441617e-01
-1.60438538e-01 3.21925133e-01 -9.12576497e-01 1.83542714e-01
-9.94373798e-01 5.05056322e-01 1.07403564e+00 4.64649558e-01
3.89984876e-01 -7.17874914e-02 1.48231506e-01 2.78914034e-01
1.15632868e+00 6.09890938e-01 5.52486658e-01 -1.31824601e+00
-2.96019286e-01 4.29604411e-01 -1.81499928e-01 -8.64097476e-01
-2.69698650e-01 3.40183564e-02 -4.83637750e-01 4.23819274e-01
7.46960163e-01 1.48868650e-01 -5.71168005e-01 1.90439093e+00
-1.71343777e-02 1.54113203e-01 2.85112262e-01 5.45098662e-01
7.82512724e-01 2.44468063e-01 1.79875866e-01 -3.10672730e-01
1.51064968e+00 -1.09192468e-02 -9.27859247e-01 -2.04484954e-01
4.56317753e-01 -2.38326371e-01 6.92126274e-01 1.03799474e+00
-1.04903984e+00 -1.77808106e-01 -1.03280807e+00 -9.16649252e-02
-4.42206055e-01 -3.09592426e-01 1.06539786e+00 7.79855013e-01
-1.10536730e+00 5.40145874e-01 -7.88807690e-01 -2.94663191e-01
5.54360449e-01 1.90830946e-01 -2.12516353e-01 1.08558219e-02
-1.47850096e+00 1.39057159e+00 9.17473018e-01 3.50677729e-01
-6.53661370e-01 -3.38446856e-01 -7.97849596e-01 -2.35769033e-01
5.20851076e-01 -8.78778338e-01 1.18439579e+00 -8.76896739e-01
-1.48370779e+00 8.32678556e-01 -4.20100302e-01 -6.94663167e-01
4.21563923e-01 -9.40647125e-02 -2.28736401e-01 6.59765676e-02
-1.67521089e-01 5.23397028e-01 4.29642081e-01 -1.24383807e+00
-3.71549159e-01 -6.81746483e-01 3.40507478e-02 -1.55773088e-01
3.37431967e-01 -1.97176635e-01 4.03452247e-01 -1.08143158e-01
8.66028249e-01 -7.14074135e-01 2.05678996e-02 -1.49791792e-01
-4.34374064e-01 -6.45981073e-01 2.52505362e-01 -3.77672404e-01
8.83846283e-01 -1.67718756e+00 -4.08212058e-02 4.67593849e-01
3.00138682e-01 -3.69960785e-01 3.43998373e-01 1.28376216e-01
-3.01703602e-01 3.59404653e-01 -1.92371607e-01 4.69779432e-01
5.63639939e-01 6.18650854e-01 -8.30432832e-01 4.31559801e-01
9.43793207e-02 1.12196088e+00 -1.05123675e+00 -7.57782698e-01
1.86461225e-01 3.16052139e-01 -6.40268743e-01 -2.50366271e-01
-4.27386403e-01 -5.48001789e-02 -4.64505047e-01 4.64638889e-01
5.39639413e-01 -1.15272745e-01 7.20414579e-01 -7.24712387e-02
-5.35526723e-02 7.04247475e-01 -1.30320048e+00 1.21368778e+00
1.17239796e-01 6.28061652e-01 -5.95271349e-01 -1.46552980e+00
8.09305072e-01 4.14062411e-01 -1.92635030e-01 -4.47008491e-01
2.54183769e-01 3.17033261e-01 4.09781545e-01 -7.37867117e-01
-2.86204815e-01 -7.06447303e-01 1.71979636e-01 4.34163868e-01
2.33313501e-01 -6.15017414e-01 3.02255806e-02 2.38239750e-01
8.52531791e-01 7.80304670e-01 6.95917845e-01 -4.84892577e-01
5.49342394e-01 -1.16560318e-01 5.92558622e-01 9.46447909e-01
4.35334072e-02 -3.56446318e-02 8.92836750e-01 -5.40512562e-01
-4.42526639e-01 -1.66303301e+00 -5.10307074e-01 8.19584131e-01
-1.39847919e-01 -2.68526971e-01 -6.28415465e-01 -3.27861726e-01
-5.10265864e-02 1.39557052e+00 -6.30563378e-01 -1.45445272e-01
5.23762107e-02 -8.12387466e-01 5.02186239e-01 2.43652523e-01
4.13356930e-01 -1.30520213e+00 -7.56273687e-01 -7.66862929e-02
-3.74071784e-02 -8.60051215e-01 9.33860481e-01 4.66356426e-01
-1.06899869e+00 -1.07182062e+00 2.77781516e-01 -1.61054552e-01
4.18000251e-01 -4.87159848e-01 1.01575756e+00 1.36565953e-01
-7.14723440e-03 2.62276888e-01 -5.94153702e-02 -7.81975091e-01
-6.71314895e-01 -6.86088443e-01 3.66705984e-01 -4.53440040e-01
1.00825906e+00 -8.12361658e-01 -8.65125284e-02 -3.06330949e-01
-9.82950270e-01 -8.02368298e-02 6.55954480e-01 7.94126630e-01
2.38101065e-01 4.64079291e-01 4.61715043e-01 -7.52782822e-01
4.57918376e-01 -4.31283087e-01 -4.88725007e-01 4.23534483e-01
-9.12303329e-01 5.04601955e-01 2.35964403e-01 -7.89756924e-02
-1.29779661e+00 -2.03051344e-01 1.53460488e-01 2.21941799e-01
-5.81439614e-01 8.78316224e-01 -2.55937248e-01 4.69543755e-01
7.34043717e-01 3.81754160e-01 -9.44746509e-02 -4.81755249e-02
5.02313137e-01 4.59913999e-01 4.87134874e-01 -1.00264549e+00
7.09842503e-01 6.46801829e-01 3.21732849e-01 -7.14251399e-01
-1.17039144e+00 3.23958099e-01 -1.07953036e+00 -2.51568109e-01
9.04951513e-01 -4.73239601e-01 -1.40820086e+00 4.02711816e-02
-1.30562055e+00 -9.19700339e-02 -3.30521643e-01 9.73494947e-01
-1.04588771e+00 4.26789582e-01 -2.09231913e-01 -1.45402563e+00
2.65488863e-01 -8.45877945e-01 4.71924424e-01 -6.45673797e-02
-7.86836863e-01 -1.26200962e+00 -1.53984539e-02 3.19888502e-01
-1.35264128e-01 1.38429418e-01 1.29179919e+00 -9.31315899e-01
-5.01915753e-01 -1.74576286e-02 3.75365727e-02 3.33721131e-01
-2.61465281e-01 1.17572926e-01 -1.05742466e+00 5.58145344e-01
3.93503010e-01 -5.02452254e-01 9.04775918e-01 4.67345685e-01
8.15632164e-01 -4.26088631e-01 -1.43053606e-01 -1.59115363e-02
1.44470763e+00 1.68962240e-01 8.16834033e-01 3.53434741e-01
1.15781955e-01 1.09755361e+00 2.24147677e-01 1.59510374e-01
6.15594923e-01 1.70803949e-01 3.21954936e-01 7.08492458e-01
1.67361602e-01 -3.36069047e-01 5.11322916e-01 5.28906286e-01
-2.83085823e-01 3.11570764e-01 -1.22345972e+00 3.52832496e-01
-2.07204580e+00 -1.76580524e+00 -4.83884603e-01 2.25487304e+00
1.40575683e+00 5.70048451e-01 -2.80596048e-01 4.35121000e-01
2.36780524e-01 -5.24190784e-01 -2.33038619e-01 -7.21381307e-01
-1.71303242e-01 1.85029402e-01 5.75558692e-02 7.17886031e-01
-4.68110681e-01 8.60477269e-01 7.36135387e+00 7.56878436e-01
-4.47238266e-01 -5.79696056e-03 4.05273944e-01 2.23684609e-01
-7.27761626e-01 4.20912474e-01 -6.96751833e-01 2.28392497e-01
1.27225173e+00 -2.06746589e-02 6.00990951e-01 5.97123325e-01
2.08900094e-01 -8.29927266e-01 -1.60055709e+00 6.46069944e-01
7.77823031e-02 -1.39284110e+00 1.57290146e-01 1.21964335e-01
3.77283096e-01 -2.08447412e-01 -2.15240270e-01 2.28359893e-01
8.07956219e-01 -1.35093915e+00 1.33853543e+00 1.06457973e+00
1.02354556e-01 -4.98245746e-01 9.24755156e-01 8.11940134e-01
-4.23163801e-01 -1.32329479e-01 -1.94021285e-01 -5.54705083e-01
-4.47948575e-02 9.00479853e-01 -6.29542351e-01 1.85539618e-01
6.90105379e-01 3.51540953e-01 -5.23914039e-01 5.68652153e-01
-8.91252458e-01 6.39779747e-01 -3.45150977e-01 -1.91188440e-01
-4.67345305e-02 -2.96545208e-01 2.78869689e-01 1.08666515e+00
-5.10130599e-02 1.43446237e-01 -3.07900518e-01 1.44899857e+00
7.23259449e-01 -3.87929797e-01 -5.01010537e-01 -4.90038134e-02
5.74080467e-01 7.56028175e-01 -9.21150506e-01 -3.82453233e-01
-2.52950907e-01 3.43782932e-01 1.83115676e-01 1.90484002e-01
-6.42254531e-01 3.36748287e-02 1.16500601e-01 -1.58991203e-01
2.67149732e-02 -7.25795850e-02 -7.37322986e-01 -8.44592512e-01
-1.07947573e-01 -6.99762464e-01 2.41278842e-01 -8.15937996e-01
-1.35585141e+00 8.12353492e-02 4.87676650e-01 -3.87152642e-01
-5.30917704e-01 -1.11105835e+00 -3.40271890e-01 7.00389922e-01
-1.28005230e+00 -9.76000488e-01 3.11356604e-01 2.94678330e-01
-6.42000213e-02 -1.92613658e-02 1.11337709e+00 -5.40765464e-01
-5.53429350e-02 4.51987348e-02 -1.35878265e-01 -3.00165378e-02
3.81110936e-01 -1.59797764e+00 -1.80322319e-01 5.24846256e-01
2.43698522e-01 1.08389091e+00 1.04806960e+00 -5.90120077e-01
-1.30559576e+00 -3.34043682e-01 1.27870703e+00 -8.83540273e-01
9.58873153e-01 -2.33441547e-01 -8.34191322e-01 7.65379071e-01
2.40929630e-02 -1.83697015e-01 8.96165073e-01 4.11611766e-01
-7.41916358e-01 3.82342152e-02 -1.20063877e+00 6.79974973e-01
6.25290394e-01 -5.66913188e-01 -1.58580363e+00 4.48405147e-01
2.97451973e-01 1.67137071e-01 -6.31660819e-01 3.04591656e-01
9.38742459e-01 -1.25784791e+00 8.44666719e-01 -7.42623866e-01
5.00142992e-01 -5.14690459e-01 -4.97731328e-01 -6.20869756e-01
-2.95572758e-01 -3.72415215e-01 -2.19388559e-01 8.67052853e-01
4.27250326e-01 -7.13882864e-01 4.60781991e-01 7.37349570e-01
2.44523585e-01 -7.17521071e-01 -1.14370859e+00 -8.17535818e-01
4.35392410e-01 -1.22422791e+00 2.93372035e-01 8.94056857e-01
5.77415943e-01 3.01064759e-01 3.00813228e-01 2.18012314e-02
8.87183249e-01 -9.48798656e-02 2.83855021e-01 -1.82326591e+00
-3.29368621e-01 -6.81023419e-01 -6.39436722e-01 -5.42704880e-01
6.09871864e-01 -1.19611263e+00 1.89477324e-01 -1.73555446e+00
4.47739512e-01 2.54652835e-02 -1.17794476e-01 5.45236766e-01
2.29097828e-01 2.83076763e-02 1.67774521e-02 1.26917124e-01
-4.47368562e-01 9.14208218e-02 7.07922220e-01 1.69877455e-01
1.17873892e-01 -2.85080969e-01 -9.14047599e-01 1.60595870e+00
7.37026632e-01 -4.83615994e-01 -4.84133363e-01 3.88860367e-02
1.30082297e+00 -2.10543007e-01 9.97819841e-01 -6.77935243e-01
2.88436443e-01 -5.32107532e-01 4.92131799e-01 -4.80227470e-01
7.26307333e-02 -6.98453188e-01 -2.44687535e-02 7.22726643e-01
-5.51448524e-01 -4.31199998e-01 -6.22968785e-02 4.64904219e-01
1.27822340e-01 -6.01438284e-01 7.50764310e-01 -3.41658592e-01
-7.81113625e-01 -4.52068806e-01 -7.39539623e-01 9.61776823e-02
8.25471997e-01 -2.76801497e-01 -2.55315691e-01 -2.78970391e-01
-9.81248915e-01 2.76354458e-02 -1.32837325e-01 -6.24652915e-02
8.55622828e-01 -1.15312719e+00 -7.25825667e-01 -1.80114165e-01
-3.22514713e-01 -2.74566203e-01 -1.28465435e-02 1.23903203e+00
-3.33338529e-01 8.13760638e-01 6.90623000e-02 -5.37850738e-01
-7.07647800e-01 4.42832410e-01 2.47657001e-01 1.31345928e-01
-2.76118696e-01 8.31219554e-01 -4.74948948e-03 -5.79353273e-01
1.77033946e-01 -4.27596986e-01 -2.27348700e-01 -1.88247245e-02
6.39817119e-01 3.74219626e-01 -3.72265279e-01 -3.32098693e-01
-6.43359125e-01 3.40087056e-01 1.95228577e-01 -4.45562631e-01
1.20714259e+00 -7.18240812e-02 -8.74541819e-01 1.18178487e+00
3.86381596e-01 -1.62941262e-01 -5.82790613e-01 -2.39059150e-01
6.28951907e-01 4.00103070e-02 1.55453324e-01 -1.23676670e+00
-1.04511745e-01 8.54604363e-01 1.80215240e-01 4.05236781e-01
5.25429785e-01 5.63436449e-01 -2.17589852e-03 8.72838676e-01
2.78986543e-01 -1.05405211e+00 -2.11764276e-01 6.42499685e-01
9.80540335e-01 -1.19185865e+00 4.09800559e-01 -1.56062439e-01
-3.25550199e-01 1.43214345e+00 3.05290490e-01 -5.14407232e-02
8.59315455e-01 4.44687456e-01 -3.07493806e-01 -4.06204462e-01
-1.09043860e+00 3.16217281e-02 3.46352190e-01 6.73804402e-01
7.29272604e-01 1.94202363e-01 -5.50630271e-01 8.77997160e-01
-7.74133801e-01 4.54884797e-01 4.19742942e-01 5.81319511e-01
-7.27157533e-01 -8.27995121e-01 -6.50921822e-01 2.46660888e-01
-3.79501700e-01 -2.64910847e-01 -3.81186366e-01 8.88954163e-01
6.00964725e-01 1.05303109e+00 -6.77784085e-02 -4.75034192e-02
-3.02075505e-01 5.29786050e-01 9.85336244e-01 -5.44775307e-01
2.13682994e-01 -1.95096254e-01 -2.28458613e-01 -5.13202727e-01
-7.98297524e-01 -9.43194985e-01 -1.83087718e+00 -3.68874341e-01
-1.59353673e-01 1.80577591e-01 6.70149922e-01 1.75487196e+00
-4.24877614e-01 4.04285371e-01 -1.69471785e-01 -4.14910406e-01
-8.81368220e-01 -7.92103589e-01 -6.46210015e-01 -2.52954960e-01
9.19286981e-02 -5.94575047e-01 -7.28001952e-01 2.71538615e-01]
|
[8.66307544708252, 6.58110237121582]
|
79708990-a2d4-4b74-bfdf-83517f946f51
|
multi-modal-learning-for-au-detection-based
|
2203.11441
| null |
https://arxiv.org/abs/2203.11441v1
|
https://arxiv.org/pdf/2203.11441v1.pdf
|
Multi-Modal Learning for AU Detection Based on Multi-Head Fused Transformers
|
Multi-modal learning has been intensified in recent years, especially for applications in facial analysis and action unit detection whilst there still exist two main challenges in terms of 1) relevant feature learning for representation and 2) efficient fusion for multi-modalities. Recently, there are a number of works have shown the effectiveness in utilizing the attention mechanism for AU detection, however, most of them are binding the region of interest (ROI) with features but rarely apply attention between features of each AU. On the other hand, the transformer, which utilizes a more efficient self-attention mechanism, has been widely used in natural language processing and computer vision tasks but is not fully explored in AU detection tasks. In this paper, we propose a novel end-to-end Multi-Head Fused Transformer (MFT) method for AU detection, which learns AU encoding features representation from different modalities by transformer encoder and fuses modalities by another fusion transformer module. Multi-head fusion attention is designed in the fusion transformer module for the effective fusion of multiple modalities. Our approach is evaluated on two public multi-modal AU databases, BP4D, and BP4D+, and the results are superior to the state-of-the-art algorithms and baseline models. We further analyze the performance of AU detection from different modalities.
|
['Lijun Yin', 'Xiang Zhang']
|
2022-03-22
| null | null | null | null |
['action-unit-detection']
|
['computer-vision']
|
[ 3.22594255e-01 -1.80274129e-01 -1.24141708e-01 7.21382443e-03
-1.24497426e+00 3.28231230e-02 5.52038014e-01 -1.03019148e-01
-4.22509074e-01 2.15763018e-01 4.26764488e-01 5.49445033e-01
2.95491368e-01 -8.02600324e-01 -6.01141334e-01 -8.66944790e-01
2.99204826e-01 7.24751363e-03 3.64421457e-01 -2.83996940e-01
-8.99062902e-02 2.00447112e-01 -2.10112333e+00 7.01041162e-01
4.61491019e-01 1.57577062e+00 4.00773548e-02 3.64616424e-01
1.66922495e-01 7.67741501e-01 -1.75938800e-01 -5.23034573e-01
2.23367259e-01 -5.16740739e-01 -6.67116702e-01 4.51779254e-02
4.55762446e-01 -6.30163312e-01 -5.06718576e-01 8.31735253e-01
1.02714097e+00 1.83836594e-02 7.31581390e-01 -1.26799965e+00
-6.82159960e-01 1.87179014e-01 -1.23930109e+00 1.56726331e-01
5.63841820e-01 5.91351688e-02 9.97985601e-01 -1.13161540e+00
2.48987243e-01 1.59507906e+00 3.87087703e-01 7.35145330e-01
-6.12014174e-01 -8.18679512e-01 -1.70946345e-01 4.51226026e-01
-1.58167136e+00 -6.30645275e-01 6.09490097e-01 -2.55889416e-01
1.13697922e+00 -5.40962070e-03 4.06121105e-01 1.02379453e+00
3.21726143e-01 1.31481874e+00 8.92111003e-01 -6.47285223e-01
-2.97192484e-01 -3.72819863e-02 -1.54574484e-01 1.02779710e+00
-4.24604982e-01 1.85559630e-01 -7.13162005e-01 -9.23615694e-02
4.98768657e-01 2.64187694e-01 -2.22355023e-01 9.49241128e-03
-8.06080580e-01 8.89569521e-01 4.03453261e-01 2.89972305e-01
-5.13641894e-01 3.72191072e-02 6.75111592e-01 1.04504287e-01
5.91780126e-01 -3.52481425e-01 -1.06701680e-01 1.97912417e-02
-4.18623179e-01 1.70735940e-01 -5.64821111e-03 8.06403041e-01
7.47306943e-01 -2.92242467e-01 -6.48634493e-01 1.11942661e+00
5.83405912e-01 6.33593798e-01 7.12239981e-01 -4.98438686e-01
5.11277735e-01 9.26963270e-01 -3.21433097e-01 -8.05762529e-01
-3.85551751e-01 1.43413052e-01 -7.84891665e-01 2.88126379e-01
5.05993776e-02 -1.27563432e-01 -8.59434903e-01 1.86337209e+00
4.83790547e-01 4.61476058e-01 2.87251443e-01 9.96894896e-01
1.34051919e+00 4.86378938e-01 1.86822936e-01 -3.55990306e-02
1.74090183e+00 -1.09338939e+00 -5.94602525e-01 -9.13567543e-02
6.73292696e-01 -7.93169737e-01 7.01725185e-01 1.21534921e-01
-1.19331491e+00 -5.87094665e-01 -7.62286484e-01 -2.47620478e-01
-3.21320802e-01 4.25476968e-01 4.59387332e-01 4.16228890e-01
-8.65926802e-01 -3.93985584e-02 -4.75459278e-01 -6.44528747e-01
7.15154290e-01 3.49256963e-01 -8.02589417e-01 -2.22498193e-01
-1.35099816e+00 1.03879309e+00 2.18681484e-01 1.21697098e-01
-1.22146070e+00 -3.11218232e-01 -1.23540509e+00 -2.04498768e-02
2.93723792e-01 -5.32847643e-01 1.15044892e+00 -8.92982841e-01
-1.43617260e+00 1.02106392e+00 -1.47180751e-01 -2.42263094e-01
1.61683023e-01 -2.50876218e-01 -3.99079323e-01 3.10227036e-01
2.97119487e-02 9.00163293e-01 1.08683574e+00 -8.03770959e-01
-1.10447621e+00 -6.39429927e-01 1.97915018e-01 5.25513053e-01
-5.81440866e-01 5.28513610e-01 -6.60353959e-01 -3.51271182e-01
-2.25283980e-01 -7.86554456e-01 2.87591577e-01 1.77161515e-01
4.12025899e-02 -7.74138808e-01 1.09393072e+00 -4.97405231e-01
1.21797419e+00 -2.23167682e+00 3.40839535e-01 -2.57592916e-01
1.67933613e-01 4.04669374e-01 -3.30372959e-01 2.42262289e-01
2.97906399e-02 -3.93092126e-01 -1.85394697e-02 -7.34048545e-01
-1.35738254e-01 9.21186283e-02 6.13613501e-02 6.29297614e-01
4.35865045e-01 9.00328755e-01 -7.24529982e-01 -8.75903606e-01
4.27230060e-01 5.53902805e-01 -3.81538928e-01 4.86060590e-01
2.57686168e-01 9.83137935e-02 -3.56996626e-01 1.33469987e+00
5.43691933e-01 9.75220054e-02 -5.96442342e-01 -5.51475465e-01
1.23004438e-02 -5.11480749e-01 -9.41038489e-01 1.97655880e+00
-3.94634247e-01 2.52996147e-01 -2.40272842e-02 -6.83829784e-01
7.24334478e-01 7.06326902e-01 5.63945770e-01 -8.68795693e-01
6.56431496e-01 -7.03121722e-02 -2.13013232e-01 -7.10545242e-01
2.59779125e-01 -3.38442594e-01 -6.57918230e-02 4.15786475e-01
3.53438169e-01 2.53377259e-01 1.19759679e-01 4.18129116e-02
1.00092912e+00 1.95567280e-01 3.91214192e-01 3.30191821e-01
8.63638580e-01 -4.91758734e-01 5.15874982e-01 3.79226148e-01
-5.12890220e-01 7.47489870e-01 1.46091044e-01 -4.81423475e-02
-4.34277415e-01 -7.97530591e-01 -1.70371830e-01 1.44513202e+00
1.49006158e-01 -4.82563615e-01 -6.96168065e-01 -9.77746964e-01
4.63834442e-02 1.03804253e-01 -1.06311226e+00 -5.29756606e-01
-5.45020550e-02 -7.11623311e-01 8.32588077e-01 8.39016736e-01
7.40557909e-01 -1.32404315e+00 -9.42762613e-01 -5.12328856e-02
-2.60978848e-01 -1.21574008e+00 -6.10813498e-01 -2.07378045e-01
-3.26593578e-01 -1.16201913e+00 -1.02576017e+00 -6.76269293e-01
3.96701515e-01 5.17623484e-01 5.96916914e-01 -8.37537721e-02
-5.76367795e-01 7.62395024e-01 -6.70496941e-01 -7.17944145e-01
-1.27629936e-01 -2.53287435e-01 1.16396293e-01 6.05747283e-01
6.14137292e-01 -1.02177180e-01 -5.69783270e-01 3.41535985e-01
-1.02793610e+00 -7.27423728e-02 9.45090592e-01 9.06671643e-01
5.41629255e-01 -3.71893853e-01 5.17208993e-01 -3.35165918e-01
4.74844575e-01 -4.95925426e-01 -1.12921178e-01 4.04984146e-01
-2.25900002e-02 -2.88274020e-01 1.38586640e-01 -2.28871331e-01
-1.15817058e+00 2.19679788e-01 -3.89133602e-01 -9.40167367e-01
-2.40935817e-01 3.75729650e-01 -2.86652029e-01 -3.26072186e-01
5.02161980e-01 2.02355728e-01 4.36790675e-01 -2.74521738e-01
2.40775868e-01 1.10747397e+00 3.06591988e-01 -3.75140160e-01
2.92135477e-01 5.50251782e-01 -4.28227708e-03 -7.97221661e-01
-9.50691402e-01 -6.77691579e-01 -5.25867820e-01 -3.88358057e-01
1.09138763e+00 -1.17670870e+00 -7.32489347e-01 9.62397993e-01
-1.02299714e+00 1.15676679e-01 -1.38509301e-02 3.86485964e-01
-5.87815762e-01 5.53702116e-01 -6.73454463e-01 -8.87947083e-01
-8.47108483e-01 -1.38409841e+00 1.72444570e+00 4.72285837e-01
1.16829492e-01 -5.62399924e-01 4.46953550e-02 4.94745374e-01
1.75021738e-01 2.34253600e-01 4.53371465e-01 -1.23162337e-01
-5.45883253e-02 -4.18261766e-01 -6.05005443e-01 4.41337436e-01
2.56421387e-01 -3.10209781e-01 -1.38523126e+00 -2.94195801e-01
-3.23462665e-01 -7.73926497e-01 9.71787274e-01 1.70753673e-01
9.27085221e-01 9.87304449e-02 -4.62671041e-01 2.84614086e-01
1.19175625e+00 5.17968014e-02 6.89141870e-01 -6.59027621e-02
6.78925097e-01 5.37417352e-01 9.95692015e-01 5.64282298e-01
3.92177910e-01 8.89576256e-01 8.16530526e-01 -1.92234233e-01
-1.47123918e-01 -1.60324238e-02 6.63297236e-01 8.56437683e-02
-2.54472345e-01 -8.85782689e-02 -5.28461456e-01 6.45473421e-01
-2.04447293e+00 -1.09870124e+00 4.22500670e-01 1.95563769e+00
4.96574551e-01 -2.61344701e-01 5.33545256e-01 1.70758888e-02
6.72961831e-01 4.93962765e-02 -4.51444656e-01 -2.59718537e-01
-1.75859258e-01 1.25923187e-01 1.07063614e-01 3.80008966e-02
-1.40372515e+00 9.57054079e-01 5.67275906e+00 9.89403009e-01
-1.00270379e+00 4.89790469e-01 3.72836053e-01 -4.40137610e-02
3.52161050e-01 -4.95799810e-01 -9.88705456e-01 3.52358222e-01
5.29903948e-01 6.68972209e-02 -8.77628550e-02 7.42225707e-01
-9.66702178e-02 -1.94006920e-01 -9.66160893e-01 1.45687079e+00
7.42212653e-01 -8.95578444e-01 1.29722476e-01 -3.95013876e-02
3.74370813e-01 5.06615341e-02 3.20120268e-02 4.77361739e-01
-2.16800526e-01 -9.12784636e-01 6.02046192e-01 5.21013916e-01
8.87623549e-01 -9.22528684e-01 1.01722121e+00 1.79292992e-01
-1.56286359e+00 -3.72095704e-01 -3.22369993e-01 5.17277718e-02
7.48983175e-02 -1.13045521e-01 -4.50205207e-01 5.50567627e-01
9.95490670e-01 9.56332803e-01 -6.15763545e-01 8.61788273e-01
-1.82988644e-02 1.62834525e-01 -2.45468080e-01 1.21243417e-01
3.62471014e-01 1.26867276e-02 2.98636526e-01 1.15032792e+00
4.02873486e-01 2.66356438e-01 1.09699160e-01 4.83163953e-01
-7.89506733e-02 3.28423053e-01 -6.13079309e-01 1.56332642e-01
-1.29017413e-01 1.45931065e+00 -1.43155217e-01 -2.26175804e-02
-1.01230359e+00 1.09256303e+00 4.04779077e-01 -1.32116124e-01
-1.07355416e+00 -2.14436084e-01 7.37300813e-01 -1.51454702e-01
2.66796499e-01 3.57316941e-01 4.82962430e-01 -9.42899525e-01
-6.53511509e-02 -8.25182378e-01 1.00183201e+00 -8.39215517e-01
-1.42398393e+00 6.39696956e-01 5.36487103e-02 -1.39172602e+00
-1.44845605e-01 -6.47473991e-01 -4.73095119e-01 7.55615830e-01
-1.58828080e+00 -1.82726669e+00 -4.04922247e-01 1.06691146e+00
6.14284992e-01 -2.65550375e-01 9.29890811e-01 6.34519041e-01
-7.31031179e-01 1.00375080e+00 -3.19609970e-01 2.73807168e-01
8.02426159e-01 -7.92598963e-01 -3.44031513e-01 7.36613572e-01
-4.95594442e-02 3.11905146e-02 2.41711840e-01 -2.56048918e-01
-1.50427413e+00 -1.10943961e+00 5.05569518e-01 -3.23669195e-01
3.96269709e-01 -9.93726626e-02 -5.98539054e-01 5.96986413e-01
4.19988573e-01 3.03439885e-01 6.86318815e-01 -1.68452993e-01
-2.56313086e-01 -1.35440454e-01 -1.22467554e+00 3.78565997e-01
9.63035047e-01 -3.66681308e-01 -4.59238619e-01 8.28112587e-02
2.81446725e-01 -5.59438050e-01 -8.35678458e-01 6.75442159e-01
6.69482470e-01 -9.11943853e-01 7.81698942e-01 -4.69458103e-01
4.49393779e-01 -3.01078171e-01 -3.99498105e-01 -1.07820129e+00
-1.33864820e-01 -2.82659590e-01 -3.92031491e-01 1.51050973e+00
4.07897215e-03 -4.92114246e-01 5.76223850e-01 3.31323385e-01
-1.16078407e-01 -1.06984770e+00 -1.13905466e+00 -2.12951288e-01
-3.03022981e-01 -3.67656916e-01 4.09014612e-01 4.44620460e-01
9.64000151e-02 5.45945704e-01 -6.81401730e-01 1.59722399e-02
4.84588146e-01 2.14023426e-01 7.29688644e-01 -8.94912481e-01
-1.57454103e-01 -4.16795015e-01 -8.84237409e-01 -7.99815297e-01
2.09722877e-01 -6.74639821e-01 1.16375305e-01 -1.57430148e+00
6.44392371e-01 -2.66399272e-02 -4.13359582e-01 9.26613510e-01
-2.41861656e-01 6.62259042e-01 2.05211386e-01 -5.99768236e-02
-8.41441095e-01 8.99159014e-01 1.03018522e+00 -2.51385510e-01
-2.63967216e-02 -8.06065425e-02 -6.93071246e-01 7.53441334e-01
3.73247564e-01 -1.34512514e-01 -1.39844492e-01 -2.57616937e-01
-2.20688730e-01 -1.44609185e-02 3.68410498e-01 -1.15312135e+00
1.78644270e-01 2.28427932e-01 3.54336202e-01 -7.46023834e-01
7.21385360e-01 -9.35953140e-01 -4.60415184e-01 1.86436683e-01
4.29672115e-02 -5.29330149e-02 4.25770134e-01 5.88924348e-01
-3.86791766e-01 -5.13246059e-02 9.99778330e-01 -9.70479473e-02
-1.19188452e+00 7.39727259e-01 -3.09044749e-01 -2.01332524e-01
1.51972628e+00 -1.93561137e-01 -1.52591139e-01 -2.67336786e-01
-5.36292315e-01 3.92506868e-01 3.53687331e-02 8.70401680e-01
9.90065038e-01 -1.59967172e+00 -1.01276457e+00 2.89932191e-01
6.55057490e-01 -1.94634601e-01 6.61275685e-01 1.07828271e+00
-1.46246022e-02 2.63200194e-01 -4.25476879e-01 -7.64258027e-01
-1.91870630e+00 5.86075604e-01 4.57049847e-01 -4.14472707e-02
-3.97738039e-01 9.80883598e-01 5.86969733e-01 -1.07409470e-02
2.66302466e-01 1.45866096e-01 -6.09925866e-01 4.04249310e-01
1.02858973e+00 2.98244894e-01 2.79627107e-02 -1.41384637e+00
-4.23801303e-01 9.68813956e-01 -1.80667728e-01 2.00149462e-01
1.12340689e+00 -2.86551744e-01 -4.53578904e-02 1.11859478e-01
1.28907537e+00 -4.17611539e-01 -1.07033682e+00 -6.12421036e-01
-5.51423490e-01 -5.36578357e-01 2.06520945e-01 -4.83168513e-01
-1.29856598e+00 1.11990523e+00 1.07793057e+00 -2.49441966e-01
1.54022169e+00 3.61711860e-01 9.94831562e-01 -7.90756419e-02
4.26627427e-01 -8.48630726e-01 2.80388266e-01 3.09165210e-01
9.96677220e-01 -1.42160487e+00 -3.04039776e-01 -3.25315297e-01
-9.25912201e-01 8.97422314e-01 1.07158089e+00 5.52100912e-02
6.20883644e-01 2.24250332e-01 1.00776076e-01 -2.73478270e-01
-6.01504266e-01 -9.35534358e-01 4.43437427e-01 4.83100206e-01
3.44519079e-01 -3.08813304e-01 -5.01619019e-02 6.10912204e-01
6.28485680e-01 9.76925120e-02 -1.02217264e-01 1.00245285e+00
-4.22739476e-01 -9.15315688e-01 -4.77200508e-01 4.97429639e-01
-5.89505374e-01 5.04665375e-02 -3.98967624e-01 7.30796933e-01
4.63715315e-01 1.04294538e+00 -5.01248911e-02 -5.56139946e-01
5.26966333e-01 6.30849823e-02 7.85781443e-01 -6.07841611e-01
-6.05080664e-01 9.49781686e-02 -2.33989522e-01 -6.94432557e-01
-7.18513489e-01 -7.94574499e-01 -1.14837062e+00 1.09143458e-01
-6.36385441e-01 -1.30491033e-01 1.41513959e-01 1.10882592e+00
4.20034498e-01 3.37078750e-01 5.37306488e-01 -9.95493710e-01
-2.66284138e-01 -1.22227907e+00 -6.10308290e-01 4.42187339e-01
2.13042051e-01 -1.22174692e+00 -1.47386014e-01 -3.68600905e-01]
|
[13.605030059814453, 1.6230545043945312]
|
c224c7a6-b746-4b95-88aa-2bed1457370a
|
visual-entailment-task-for-visually-grounded
|
1811.10582
| null |
http://arxiv.org/abs/1811.10582v2
|
http://arxiv.org/pdf/1811.10582v2.pdf
|
Visual Entailment Task for Visually-Grounded Language Learning
|
We introduce a new inference task - Visual Entailment (VE) - which differs
from traditional Textual Entailment (TE) tasks whereby a premise is defined by
an image, rather than a natural language sentence as in TE tasks. A novel
dataset SNLI-VE (publicly available at https://github.com/necla-ml/SNLI-VE) is
proposed for VE tasks based on the Stanford Natural Language Inference corpus
and Flickr30k. We introduce a differentiable architecture called the
Explainable Visual Entailment model (EVE) to tackle the VE problem. EVE and
several other state-of-the-art visual question answering (VQA) based models are
evaluated on the SNLI-VE dataset, facilitating grounded language understanding
and providing insights on how modern VQA based models perform.
|
['Ning Xie', 'Derek Doran', 'Farley Lai', 'Asim Kadav']
|
2018-11-26
| null | null | null | null |
['grounded-language-learning', 'visual-entailment']
|
['natural-language-processing', 'reasoning']
|
[ 3.98931792e-03 2.73528963e-01 -4.58579212e-02 -6.76532149e-01
-6.86725199e-01 -7.85236955e-01 9.47440565e-01 -1.52405009e-01
-2.88225085e-01 4.30372953e-01 4.38924253e-01 -9.48727131e-01
3.59244823e-01 -7.35132277e-01 -1.22499883e+00 7.61784315e-02
3.28377247e-01 6.12971187e-01 -2.29515716e-01 -2.03981206e-01
3.19958925e-01 -3.04097906e-02 -1.34172082e+00 1.09733188e+00
6.98318183e-01 8.99259925e-01 2.33698368e-01 1.15150714e+00
-3.15644383e-01 1.75673127e+00 -4.38499957e-01 -9.61921096e-01
-2.14840472e-01 -5.81167877e-01 -1.54930866e+00 -1.78692251e-01
1.04712725e+00 -6.65223300e-01 -4.18790877e-01 7.70217419e-01
-2.17779696e-01 7.19971880e-02 1.00731742e+00 -1.70625389e+00
-1.79489708e+00 5.59366703e-01 -3.46505463e-01 2.63922393e-01
6.84200943e-01 4.58995670e-01 1.78122854e+00 -1.41985118e+00
8.87015760e-01 1.49686015e+00 4.77919549e-01 6.48579240e-01
-1.20266771e+00 -1.08388655e-01 3.47559601e-01 9.29421782e-01
-1.28831291e+00 -4.00971323e-01 6.94828391e-01 -4.07002956e-01
1.51712132e+00 5.95785022e-01 7.08372056e-01 1.27783203e+00
9.52755585e-02 1.38015378e+00 8.93188000e-01 -4.17034626e-01
-6.76949322e-02 -2.79836953e-01 2.94254631e-01 1.33937073e+00
-1.65980905e-01 -9.76933315e-02 -5.47168553e-01 3.58035475e-01
6.17903471e-01 -2.63180673e-01 -3.19847316e-01 -3.63094658e-01
-1.36845005e+00 1.03456008e+00 1.01500082e+00 1.98623203e-02
-2.25538939e-01 8.47809315e-01 5.63345313e-01 4.81966972e-01
3.90418917e-01 4.42095309e-01 -1.30634308e-01 2.50008792e-01
-7.84553945e-01 3.86306077e-01 5.97415745e-01 1.06789148e+00
8.06428671e-01 -1.22599257e-02 -5.00251114e-01 1.35652855e-01
6.85857892e-01 4.79626149e-01 -1.70903847e-01 -1.18454123e+00
7.23397851e-01 6.77831113e-01 1.12458706e-01 -7.35176206e-01
3.86725552e-02 1.00763716e-01 -6.59599364e-01 2.68458039e-01
4.62218136e-01 3.86146605e-01 -1.06166422e+00 1.54049182e+00
1.13641568e-01 3.09161812e-01 2.19732895e-01 9.89839315e-01
1.69012189e+00 9.83359516e-01 3.33948493e-01 3.55118483e-01
1.57664037e+00 -1.20349181e+00 -7.47215688e-01 -5.78332782e-01
6.77486956e-01 -4.48160887e-01 1.92243731e+00 3.15064818e-01
-1.09378731e+00 -7.13239610e-01 -1.02747917e+00 -9.46133912e-01
-5.47160208e-01 1.35453701e-01 5.32808423e-01 1.16158038e-01
-1.18651807e+00 -1.82529479e-01 -3.64753962e-01 -2.88808011e-02
8.04400265e-01 -3.41619968e-01 -3.10609341e-01 -5.48312962e-01
-1.28125715e+00 1.12055743e+00 9.41406488e-02 5.15959501e-01
-1.58230722e+00 -9.35065091e-01 -1.52953756e+00 2.84115188e-02
4.25586671e-01 -1.24403691e+00 1.52764547e+00 -1.17262459e+00
-9.91743267e-01 1.50147915e+00 -5.99820912e-01 -5.91865957e-01
5.31520069e-01 -5.00883460e-01 -1.82714850e-01 4.07768488e-01
1.36305094e-01 7.86926270e-01 9.35908258e-01 -1.44290125e+00
-7.63219548e-03 -1.77893147e-01 8.70844066e-01 5.24791852e-02
5.63068748e-01 -5.93008213e-02 -3.51893306e-01 -3.41776848e-01
-5.93322933e-01 -4.72646594e-01 1.61806464e-01 4.12486553e-01
-6.96502566e-01 -6.70210600e-01 7.11718082e-01 -8.35451186e-01
9.13644314e-01 -1.86319709e+00 4.10790890e-01 -1.70619085e-01
7.43975818e-01 2.65282020e-02 -4.19839114e-01 4.64446187e-01
-1.09844923e-01 1.22563660e-01 -3.46555501e-01 -5.95834792e-01
6.48660362e-01 7.48163700e-01 -6.34144664e-01 3.21215957e-01
7.34195769e-01 1.88722706e+00 -9.84234452e-01 -7.63220251e-01
5.53668141e-01 4.11369383e-01 -5.77248335e-01 3.80848974e-01
-8.87649536e-01 8.47942084e-02 -2.09142819e-01 7.01061845e-01
4.53006178e-01 -7.71532476e-01 -9.90277603e-02 -6.58444881e-01
8.12314600e-02 1.78372264e-01 -5.49803376e-01 1.81428480e+00
-6.83021307e-01 1.24595785e+00 -1.43859312e-01 -1.04427409e+00
4.45481926e-01 2.72047818e-01 -4.60467815e-01 -9.36069012e-01
1.31185889e-01 -4.93276894e-01 -4.17764038e-01 -9.32292700e-01
5.93294919e-01 -2.04387590e-01 -1.66596726e-01 2.91666538e-01
2.23910481e-01 -2.97536641e-01 2.40043312e-01 9.66504276e-01
6.73975885e-01 4.72258508e-01 4.90045249e-01 -7.55201802e-02
4.57263172e-01 4.63313699e-01 -7.84815252e-02 8.16881776e-01
-4.94799435e-01 4.52747554e-01 7.10368872e-01 -5.89916766e-01
-1.04157305e+00 -1.64061284e+00 3.24770994e-02 1.21547949e+00
2.67026395e-01 -7.53514588e-01 -5.88484347e-01 -9.47078526e-01
4.67318855e-02 1.28715301e+00 -1.11687410e+00 1.35567814e-01
-2.92398244e-01 2.30308399e-01 9.33102429e-01 7.52736568e-01
5.89099407e-01 -1.44442403e+00 -4.82080579e-01 -3.85907531e-01
-8.00654948e-01 -1.33697116e+00 -2.79769301e-01 -4.21243101e-01
-6.60943091e-02 -1.24635768e+00 -2.07629770e-01 -7.67452896e-01
5.92660546e-01 -5.55898026e-02 2.19430232e+00 4.96470779e-01
-3.41650575e-01 1.01099312e+00 -4.44992155e-01 -5.95753491e-01
-4.83939618e-01 -6.21831059e-01 -5.91055930e-01 -1.73078254e-01
5.44762671e-01 2.65002578e-01 -6.24701381e-01 -1.20404333e-01
-9.84071136e-01 5.14358580e-01 2.86065012e-01 8.58253360e-01
8.37264538e-01 -6.50649488e-01 3.78390819e-01 -9.15019572e-01
5.26606500e-01 -4.26946431e-01 -3.37164581e-01 6.81418240e-01
-1.29266664e-01 1.12960979e-01 6.05156839e-01 -6.98427716e-03
-1.38384986e+00 -3.32241327e-01 -2.94034928e-01 -4.81251717e-01
-2.71067947e-01 7.02310324e-01 -9.99685600e-02 3.60003412e-01
6.93433166e-01 1.72919586e-01 -2.89530382e-02 1.34261325e-01
1.39219379e+00 3.45601052e-01 1.04161656e+00 -5.61214685e-01
6.35390639e-01 6.51793122e-01 -2.02961136e-02 -9.17769194e-01
-1.37645781e+00 -3.23345840e-01 -7.18537211e-01 -3.88949275e-01
1.41727448e+00 -1.02666891e+00 -9.56036925e-01 -5.91913052e-02
-1.54638565e+00 -6.92349434e-01 -2.84050137e-01 -1.39910400e-01
-7.55776763e-01 5.53611636e-01 -7.54635453e-01 -6.26139522e-01
-6.00162029e-01 -9.56631362e-01 1.09505653e+00 -1.30009204e-01
-6.18432045e-01 -1.20064640e+00 7.65899345e-02 8.50671172e-01
1.10727914e-01 3.42532486e-01 9.75347519e-01 -5.74461520e-02
-7.89932787e-01 3.14512193e-01 -7.62566745e-01 2.54319638e-01
-4.14033473e-01 2.20096916e-01 -9.59596217e-01 -1.45339360e-02
-4.28069502e-01 -1.17723370e+00 1.05922151e+00 3.60773712e-01
1.28097570e+00 -4.43865955e-01 3.21907550e-02 4.22078460e-01
1.54587758e+00 -6.11832857e-01 1.04058087e+00 1.17246881e-01
1.05328095e+00 3.64919454e-01 7.08236158e-01 1.09240212e-01
9.35307741e-01 1.65240467e-01 1.12408912e+00 -3.56934577e-01
-4.16783452e-01 -4.54681665e-01 3.30072314e-01 3.86069328e-01
-1.99582949e-01 -4.30846989e-01 -9.92769361e-01 8.46680045e-01
-1.88582242e+00 -1.37106323e+00 -6.08973444e-01 1.22424757e+00
6.38879597e-01 -2.81467289e-01 -1.48332030e-01 -2.88038969e-01
1.29633233e-01 4.08977360e-01 -4.87382740e-01 -8.16372871e-01
-2.02255845e-01 3.79402578e-01 -3.87511700e-01 9.05900061e-01
-1.00717473e+00 1.09054708e+00 6.04838371e+00 5.40839493e-01
-5.79066455e-01 6.78815320e-02 5.41385651e-01 -1.74669810e-02
-8.51527989e-01 7.27002546e-02 -2.20098883e-01 -2.19862834e-01
5.27799487e-01 2.24201083e-01 4.30274189e-01 7.74973452e-01
-1.49273261e-01 -3.74407172e-02 -1.44851816e+00 9.79154527e-01
4.86555606e-01 -1.90459645e+00 6.02902472e-01 -6.01218283e-01
3.31165344e-01 2.90825665e-01 1.73943102e-01 5.60676455e-01
4.40934420e-01 -1.68357563e+00 1.05061424e+00 6.47547424e-01
1.15065014e+00 -6.14213407e-01 5.48473835e-01 -1.04757007e-02
-1.52032757e+00 3.21134061e-01 -4.54945832e-01 -2.94241875e-01
3.67315769e-01 7.43110552e-02 -6.69909418e-01 6.38148189e-01
7.79823422e-01 9.34525847e-01 -6.34826481e-01 3.19114357e-01
-1.00462604e+00 7.49580264e-01 2.99977988e-01 -1.55476108e-01
6.12738132e-01 -3.29670347e-02 4.85084116e-01 1.35973561e+00
-2.52289683e-01 1.15135591e-02 -3.29901040e-01 1.71794105e+00
-3.31098229e-01 -3.71226996e-01 -8.24321926e-01 -9.25958604e-02
1.67792767e-01 1.14920962e+00 -7.43990913e-02 -5.67676544e-01
-8.73162448e-01 1.44882596e+00 8.88170838e-01 6.65503919e-01
-1.25557899e+00 -2.13951543e-01 6.23463273e-01 -2.42742866e-01
4.74984199e-01 -1.43886894e-01 -3.58553231e-02 -1.31454301e+00
-9.61951613e-02 -9.75888133e-01 7.88433492e-01 -1.71069503e+00
-1.53586876e+00 4.15746897e-01 -9.70733166e-02 -7.81362057e-01
-2.24254534e-01 -1.14439654e+00 -7.22138584e-01 6.13492250e-01
-1.81584752e+00 -1.84930789e+00 -5.98202050e-01 9.46436763e-01
7.26219952e-01 2.98002154e-01 8.51420999e-01 -2.20653504e-01
-2.45138034e-01 3.75148833e-01 -6.11247838e-01 3.56291741e-01
4.52553570e-01 -1.62876773e+00 8.22133541e-01 9.22394156e-01
9.12843645e-01 5.24540007e-01 7.43008316e-01 -2.24687248e-01
-1.55176926e+00 -1.19840121e+00 1.00987649e+00 -1.38602519e+00
7.71868587e-01 -4.39482659e-01 -9.12546694e-01 1.56895709e+00
9.24124897e-01 1.80151641e-01 6.48082733e-01 1.53004369e-02
-9.39385355e-01 4.54924017e-01 -8.85061562e-01 8.46061170e-01
1.28911805e+00 -1.10461092e+00 -1.23924506e+00 1.49399489e-01
9.82022643e-01 -3.86954188e-01 -6.79690063e-01 1.46888554e-01
5.11427462e-01 -8.05468261e-01 1.34259999e+00 -1.27547109e+00
1.31091332e+00 -4.93843853e-01 -3.65416288e-01 -1.31369889e+00
-4.30470228e-01 -7.84457773e-02 -4.17373598e-01 9.16348159e-01
5.16486645e-01 -1.16146542e-01 5.03517926e-01 4.58080649e-01
-2.84007370e-01 -6.54000878e-01 -8.00854504e-01 -3.95003259e-01
1.08960502e-01 -8.54272008e-01 3.65797758e-01 1.02070308e+00
-3.52551341e-02 8.82282317e-01 -3.42456251e-01 2.87611812e-01
8.04943681e-01 5.61100841e-01 9.40226614e-01 -6.31130040e-01
-4.61074084e-01 -2.21949071e-01 -1.72153756e-01 -1.40310645e+00
4.40768719e-01 -1.23972929e+00 6.76010847e-02 -2.28589988e+00
3.51083666e-01 3.49787444e-01 -2.74838191e-02 6.31993592e-01
-3.12330067e-01 5.24417281e-01 2.24967167e-01 -3.45151007e-01
-1.24687314e+00 7.32931554e-01 1.50988328e+00 -6.40494406e-01
6.05312109e-01 -7.31736183e-01 -5.34269512e-01 7.87128806e-01
4.85096514e-01 1.06611058e-01 -5.90107918e-01 -1.11853206e+00
7.73253500e-01 -1.15162604e-01 1.18656409e+00 -1.14624083e-01
-1.30764037e-01 -1.87190339e-01 5.55297315e-01 -8.78255785e-01
2.21530288e-01 -6.05482340e-01 -2.81789511e-01 2.84458280e-01
-3.59545887e-01 3.70311767e-01 2.47674242e-01 5.68513811e-01
-3.72155190e-01 -5.69694303e-02 2.56428480e-01 -3.01490158e-01
-1.56537712e+00 2.48606637e-01 -3.47016484e-01 6.37969196e-01
9.28652287e-01 6.88808560e-02 -1.00467050e+00 -6.88865721e-01
-6.45441890e-01 6.65635288e-01 1.64426312e-01 2.35108212e-01
1.32283700e+00 -1.30715036e+00 -1.07410681e+00 -2.71015018e-01
7.75172651e-01 1.08873539e-01 4.77160931e-01 6.83598101e-01
-7.61058748e-01 4.16286677e-01 -1.14406653e-01 -6.55849695e-01
-1.52719915e+00 7.58089244e-01 3.35044235e-01 -1.26090214e-01
-6.20661855e-01 1.31906128e+00 6.64487422e-01 -5.23729563e-01
-1.11370631e-01 -6.77735209e-01 -2.08126396e-01 -4.23671663e-01
6.89845681e-01 -1.92305632e-02 -3.37179333e-01 -7.07967937e-01
-5.09895623e-01 2.88719416e-01 1.52049169e-01 3.38929482e-02
9.84620452e-01 -3.38906676e-01 -3.10301453e-01 4.55996245e-01
1.07537067e+00 -3.39517087e-01 -1.03752697e+00 -2.40311965e-01
-2.99519777e-01 -2.53505319e-01 -1.15403853e-01 -7.82809734e-01
-5.61614096e-01 1.16943860e+00 -1.04487047e-01 1.58947129e-02
8.37506652e-01 7.45782673e-01 4.36249942e-01 5.90903461e-01
-7.48566389e-02 -5.96189559e-01 5.67217648e-01 6.45140946e-01
1.66384745e+00 -1.46083152e+00 -8.58351514e-02 -2.25858495e-01
-1.13775063e+00 8.46608460e-01 7.45307148e-01 -1.40153319e-01
2.44346127e-01 -4.22441959e-02 1.14322059e-01 -8.16622555e-01
-8.59779060e-01 -4.60590154e-01 8.48254740e-01 6.48482442e-01
5.44632912e-01 2.17245206e-01 3.87000471e-01 6.04023933e-01
-3.15883875e-01 -1.02416493e-01 2.11758241e-01 6.24171317e-01
-5.41581586e-02 -3.69447827e-01 -3.99582721e-02 2.38575771e-01
-1.36146575e-01 -5.90766311e-01 -8.56041014e-01 1.10362327e+00
-6.40260652e-02 1.07699859e+00 4.07323480e-01 -5.19693382e-02
2.15613484e-01 1.28054649e-01 1.13906622e+00 -3.11702460e-01
-1.61485434e-01 -8.33527088e-01 6.77850246e-01 -8.90919447e-01
-4.53088433e-01 -1.48063794e-01 -1.54511881e+00 -6.02728248e-01
7.91854411e-02 -2.44786486e-01 1.23926409e-01 1.23972631e+00
1.11101963e-01 6.80070102e-01 -1.50075555e-01 -4.51899499e-01
-1.14476547e-01 -6.79954946e-01 -1.33668572e-01 9.85335767e-01
6.20800853e-01 -2.93717742e-01 -2.90611625e-01 4.45097268e-01]
|
[10.845710754394531, 1.746693730354309]
|
a8a83837-0b99-4fc3-a2e4-c8879f086656
|
block-wise-partitioning-for-extreme-multi
|
1811.01305
| null |
http://arxiv.org/abs/1811.01305v1
|
http://arxiv.org/pdf/1811.01305v1.pdf
|
Block-wise Partitioning for Extreme Multi-label Classification
|
Extreme multi-label classification aims to learn a classifier that annotates
an instance with a relevant subset of labels from an extremely large label set.
Many existing solutions embed the label matrix to a low-dimensional linear
subspace, or examine the relevance of a test instance to every label via a
linear scan. In practice, however, those approaches can be computationally
exorbitant. To alleviate this drawback, we propose a Block-wise Partitioning
(BP) pretreatment that divides all instances into disjoint clusters, to each of
which the most frequently tagged label subset is attached. One multi-label
classifier is trained on one pair of instance and label clusters, and the label
set of a test instance is predicted by first delivering it to the most
appropriate instance cluster. Experiments on benchmark multi-label data sets
reveal that BP pretreatment significantly reduces prediction time, and retains
almost the same level of prediction accuracy.
|
['Thomas C. M. Lee', 'Cho-Jui Hsieh', 'Yuefeng Liang']
|
2018-11-04
| null | null | null | null |
['extreme-multi-label-classification']
|
['methodology']
|
[ 5.98462045e-01 4.13170271e-02 -6.87958479e-01 -6.63564086e-01
-1.13631499e+00 -9.92855966e-01 5.77373058e-02 3.97529334e-01
-3.49731185e-02 5.61885238e-01 -4.15575296e-01 -9.78922322e-02
-4.40490454e-01 -5.92382073e-01 -1.94425732e-01 -1.12960875e+00
2.11807713e-01 9.90965426e-01 -1.83403686e-01 6.00644171e-01
2.46187925e-01 3.91594797e-01 -1.62806666e+00 7.36002684e-01
4.97843295e-01 1.24630702e+00 -2.12424189e-01 2.61736363e-01
-6.17607012e-02 7.67806113e-01 -3.32942039e-01 -2.53337115e-01
4.03770447e-01 -4.87846971e-01 -1.19244337e+00 3.71281236e-01
5.85969687e-01 3.47244561e-01 3.62836152e-01 1.09288704e+00
2.24601880e-01 -1.58702508e-01 9.74709749e-01 -1.59042943e+00
-8.98472443e-02 5.22870839e-01 -9.91903543e-01 -2.00032488e-01
1.97189108e-01 -4.62353885e-01 1.16648185e+00 -9.05238748e-01
5.43252885e-01 7.70408809e-01 7.77461648e-01 4.52173263e-01
-1.68571484e+00 -8.50739300e-01 -3.74786486e-03 1.19967178e-01
-1.57645762e+00 -1.45336181e-01 8.01752031e-01 -4.95978802e-01
2.28671059e-01 5.86378276e-01 2.78502166e-01 4.82365966e-01
4.57867607e-02 4.30523664e-01 1.47753572e+00 -4.62504953e-01
3.84417087e-01 4.91599441e-01 6.66707277e-01 6.51722848e-01
-3.82477790e-02 -1.48142666e-01 -3.21319312e-01 -6.25891805e-01
-1.60735935e-01 2.00133711e-01 5.15854359e-03 -6.87607884e-01
-1.05484569e+00 8.56112719e-01 2.11412534e-01 1.84376523e-01
-3.94624978e-01 -3.17971736e-01 3.95352542e-01 4.02485579e-01
6.80827081e-01 2.98339635e-01 -7.24542379e-01 4.10542935e-01
-1.07003903e+00 -1.29683554e-01 7.59160638e-01 7.29081213e-01
1.30239713e+00 -9.08027112e-01 -4.80318889e-02 9.36038971e-01
3.31845015e-01 1.79069221e-01 3.83884072e-01 -9.24299300e-01
1.93987608e-01 9.40311670e-01 -2.25806817e-01 -7.41594076e-01
-6.91025317e-01 -5.29581130e-01 -8.15363705e-01 5.25621213e-02
2.39240870e-01 -4.66989279e-02 -4.56967533e-01 1.54810333e+00
8.41375172e-01 3.05764496e-01 7.53905475e-02 6.52946949e-01
5.35120964e-01 5.74890852e-01 3.00621867e-01 -7.19010115e-01
1.29483163e+00 -1.03659439e+00 -2.61666507e-01 -1.85646057e-01
1.18370438e+00 -6.52474761e-01 6.59010291e-01 3.70658308e-01
-6.96035445e-01 -4.51280951e-01 -6.50616884e-01 4.29297894e-01
-2.00508460e-01 3.67843121e-01 5.13247848e-01 5.77947855e-01
-7.16440916e-01 4.48066264e-01 -2.21400127e-01 -1.61509346e-02
2.26654232e-01 6.04379833e-01 -6.62167728e-01 -5.32088459e-01
-8.41920793e-01 4.74512547e-01 6.25023365e-01 -3.46421301e-01
-5.58788717e-01 -6.51038587e-01 -7.89205313e-01 9.55611020e-02
4.39572930e-01 -1.97621688e-01 8.97668898e-01 -1.06990802e+00
-7.88468540e-01 1.38278973e+00 -3.63095462e-01 2.72906125e-02
1.36903197e-01 3.56138974e-01 -4.22699422e-01 1.69571549e-01
5.40231168e-01 8.47134650e-01 8.76009345e-01 -1.45536005e+00
-1.15361512e+00 -6.11937761e-01 -1.21147268e-01 1.17422126e-01
-5.22633314e-01 1.34738386e-01 -2.22157434e-01 -1.63327366e-01
7.93324411e-01 -1.48193467e+00 -2.41491660e-01 -5.65437913e-01
-5.71665049e-01 -6.19549215e-01 9.36078668e-01 -1.41552791e-01
1.32145739e+00 -2.29984760e+00 1.45113885e-01 4.94019151e-01
3.02766204e-01 -1.34237915e-01 -1.03950270e-01 1.56007081e-01
-6.99310839e-01 1.44141410e-02 -2.06476465e-01 -3.48113686e-01
-7.20267668e-02 -2.74172053e-02 -3.75870436e-01 7.78123438e-01
-1.56545177e-01 6.52826130e-01 -7.77218819e-01 -8.04217279e-01
4.25175205e-02 6.34571314e-02 -5.81384122e-01 1.07076950e-01
-7.96713904e-02 4.82704431e-01 -3.77946258e-01 7.13840485e-01
6.43384337e-01 -7.03652859e-01 5.81537902e-01 -2.83378512e-01
2.78218448e-01 -1.77833751e-01 -1.46107340e+00 1.13355350e+00
-1.09883875e-01 7.27071539e-02 -2.43387565e-01 -1.14463758e+00
8.67402434e-01 4.33378041e-01 1.08694077e+00 -2.68383145e-01
-9.00475681e-02 1.88952267e-01 -4.68803972e-01 -3.72510999e-01
-2.18985579e-03 -3.04902196e-01 -3.51868987e-01 9.51755702e-01
-2.38341242e-01 3.70562285e-01 2.34796852e-01 1.22524150e-01
1.05113780e+00 -1.19132593e-01 4.82933998e-01 -1.96366727e-01
6.86729491e-01 1.45613821e-02 8.24360430e-01 3.98663342e-01
-1.91379651e-01 5.08096278e-01 5.81456959e-01 -6.69742882e-01
-6.98118627e-01 -6.49497390e-01 -5.25210619e-01 1.53602564e+00
1.30579188e-01 -4.32354242e-01 -6.73111975e-01 -1.52035892e+00
1.33542538e-01 5.34673095e-01 -6.98233664e-01 -2.13813484e-01
-3.24177474e-01 -8.92886579e-01 1.15363717e-01 6.83063045e-02
4.16231193e-02 -9.75934982e-01 -1.92173496e-01 1.04046844e-01
-3.15554798e-01 -7.35403955e-01 -4.49311525e-01 8.53240430e-01
-6.53543890e-01 -1.36385739e+00 -2.93532372e-01 -1.22981763e+00
1.15282822e+00 3.25338602e-01 1.07307351e+00 7.58320764e-02
1.68534555e-02 5.07533997e-02 -3.92758816e-01 1.90649614e-01
-4.91232991e-01 3.72227579e-01 1.64782032e-01 5.15814364e-01
6.75792456e-01 -2.21715212e-01 -1.63931906e-01 7.98708320e-01
-6.07270241e-01 -9.76611450e-02 3.83584887e-01 8.29986811e-01
1.12088704e+00 4.89420384e-01 8.37666273e-01 -1.37977099e+00
1.23360544e-01 -8.23509395e-01 -4.54898149e-01 4.82063472e-01
-1.07696521e+00 -1.84055045e-01 5.96687257e-01 -4.71648991e-01
-5.27847230e-01 8.75344694e-01 3.31641167e-01 -2.85195410e-01
-5.24727881e-01 5.30775011e-01 -1.10546954e-01 -5.21287695e-02
6.59867048e-01 5.71796410e-02 -3.51894498e-01 -3.30303758e-01
2.44572029e-01 9.14372385e-01 2.82035589e-01 -4.33220416e-01
6.58103347e-01 1.82772815e-01 3.45094562e-01 -6.11717962e-02
-1.51377726e+00 -9.70600307e-01 -1.16851270e+00 -2.91145653e-01
5.87627172e-01 -7.42346346e-01 -7.59154856e-01 6.80304691e-02
-5.68399608e-01 -1.30656421e-01 -3.72941196e-01 3.72791260e-01
-4.03976530e-01 2.97866464e-01 -5.35780549e-01 -3.14948827e-01
1.45264426e-02 -1.17784238e+00 1.12182093e+00 2.26138309e-02
-4.08392042e-01 -9.42595065e-01 2.37148538e-01 7.19114721e-01
-3.28275412e-01 1.45117134e-01 1.26817274e+00 -1.02045321e+00
-5.10223061e-02 -7.63475358e-01 -1.66510403e-01 1.12122767e-01
1.43042371e-01 -3.67477447e-01 -1.06147969e+00 -6.83814466e-01
1.25874415e-01 -7.03750610e-01 6.86952770e-01 2.38713086e-01
1.22912872e+00 -2.86885481e-02 -9.61787820e-01 4.45262998e-01
1.41449881e+00 1.97218940e-01 -9.00752768e-02 2.27187887e-01
7.53039360e-01 7.45414019e-01 1.10334301e+00 3.62705976e-01
2.86499977e-01 7.20031679e-01 3.15889150e-01 -6.07413538e-02
1.42326146e-01 2.29875326e-01 5.12976646e-02 7.00014889e-01
5.00607967e-01 -1.99270472e-01 -9.98296559e-01 1.25672206e-01
-1.64161837e+00 -7.24474907e-01 -2.35418126e-01 2.15079403e+00
9.56676424e-01 -3.37932333e-02 1.01176649e-01 5.28488398e-01
1.09958541e+00 -1.54307917e-01 -6.55595839e-01 4.33556139e-02
1.12209231e-01 -2.51823723e-01 2.51975268e-01 2.60282248e-01
-1.50978673e+00 6.88836277e-01 6.32270002e+00 8.13620329e-01
-1.01278615e+00 1.35738522e-01 9.99168158e-01 -2.26689324e-01
2.04358269e-02 1.10557802e-01 -1.11636031e+00 4.98992413e-01
1.14364100e+00 3.58613096e-02 2.81260937e-01 1.02429426e+00
-2.37842098e-01 -6.18369095e-02 -1.51251042e+00 9.84265387e-01
2.54025191e-01 -9.54450250e-01 -2.97949135e-01 4.24098462e-01
1.01682115e+00 -3.50678921e-01 2.02833980e-01 3.81933302e-01
1.94271803e-01 -8.64149630e-01 4.66180712e-01 2.01223269e-01
1.11242878e+00 -1.05579901e+00 6.45403326e-01 6.95206523e-01
-1.24584317e+00 -4.48363870e-01 -3.54351312e-01 1.21650524e-01
-1.69184640e-01 8.24590921e-01 -8.74839842e-01 2.60752231e-01
4.72758949e-01 6.78073049e-01 -8.07065785e-01 8.76124382e-01
-6.30208710e-03 7.67012358e-01 -5.43001741e-02 5.32822669e-01
2.10686639e-01 -3.00521880e-01 -6.11221902e-02 8.96893084e-01
1.00766867e-01 1.52066946e-01 8.67185414e-01 2.67638594e-01
-1.74202204e-01 3.99253666e-01 -5.87996840e-01 3.45608026e-01
6.00693703e-01 1.75709879e+00 -1.15051866e+00 -4.56565946e-01
-4.61538672e-01 9.14442778e-01 5.86571455e-01 5.61785474e-02
-7.09285736e-01 1.91602968e-02 2.76440263e-01 -1.85859144e-01
1.01257138e-01 5.54338634e-01 -4.50002283e-01 -7.52976000e-01
-1.24762058e-01 -7.98544824e-01 8.89442265e-01 -4.77452457e-01
-1.24255419e+00 3.98901731e-01 -3.53124171e-01 -1.55748451e+00
-2.88607031e-01 -2.38803566e-01 -5.61267398e-02 6.88926697e-01
-1.22464931e+00 -1.13383365e+00 -1.42928585e-01 3.91787440e-01
2.26062194e-01 -2.32787788e-01 1.25501907e+00 4.70640868e-01
-7.21978009e-01 7.91028082e-01 3.10696125e-01 -9.13802609e-02
1.02120709e+00 -1.31122959e+00 -3.24104041e-01 1.39188290e-01
4.00052369e-01 2.36039013e-01 2.40438461e-01 -4.90415573e-01
-8.81703436e-01 -1.50137019e+00 1.19493020e+00 -6.80612206e-01
1.77108556e-01 -2.37848591e-02 -9.91234243e-01 8.58819067e-01
-2.52020478e-01 4.37512428e-01 1.48365474e+00 3.17862213e-01
-4.13239479e-01 -1.81614220e-01 -1.43014157e+00 1.50135413e-01
5.60021937e-01 -6.39821768e-01 -2.93124825e-01 8.69072855e-01
4.31144804e-01 -1.16817974e-01 -1.26177204e+00 3.64071071e-01
4.09397811e-01 -7.39710033e-01 7.53937960e-01 -6.69199288e-01
3.63393575e-01 -4.64093387e-01 -1.57737359e-01 -1.28714132e+00
-7.95583189e-01 3.84778678e-02 -1.24721760e-02 1.39939082e+00
5.25800288e-01 -3.69830489e-01 1.15382183e+00 7.05053747e-01
1.03040412e-01 -9.58442509e-01 -7.98420012e-01 -5.06802022e-01
-4.21190001e-02 -2.16687217e-01 5.80243468e-01 1.62952220e+00
3.69072229e-01 5.97542524e-01 -2.60567099e-01 4.22265232e-01
8.45276654e-01 7.78687298e-01 4.36325848e-01 -1.77862024e+00
-8.86751115e-02 -1.58083692e-01 -3.12469006e-01 -3.47712070e-01
7.28831232e-01 -1.50294256e+00 -3.40469554e-02 -1.14361966e+00
7.68446982e-01 -9.38575149e-01 -6.96537852e-01 9.13017511e-01
-3.14858466e-01 8.75122249e-01 -1.35348186e-01 6.14025295e-01
-1.00854027e+00 -2.26012960e-01 7.26248324e-01 -3.15481454e-01
-1.09101102e-01 1.97194204e-01 -6.34900868e-01 7.67519176e-01
7.46549785e-01 -1.14536870e+00 -3.83774459e-01 2.27787539e-01
3.10178518e-01 2.59001434e-01 -8.25089291e-02 -9.59134102e-01
3.79931420e-01 -2.65006810e-01 4.91034687e-01 -6.46562755e-01
1.59312170e-02 -1.14122951e+00 4.44821626e-01 4.56866264e-01
-8.77341211e-01 -2.88683027e-01 -1.19018033e-01 4.80139703e-01
-1.73872814e-01 -5.57603478e-01 1.09428859e+00 1.57159448e-01
-5.41998684e-01 4.32541937e-01 -4.47357371e-02 -1.42708734e-01
1.65089500e+00 -1.08557463e-01 -5.55839203e-02 2.29812816e-01
-1.00170362e+00 2.92204857e-01 6.29731476e-01 6.45824447e-02
2.65812427e-01 -1.61768413e+00 -5.99867046e-01 4.20341045e-01
4.69278336e-01 -7.78207555e-02 2.57788390e-01 6.70876980e-01
8.69816765e-02 5.08426905e-01 1.88036218e-01 -8.16200614e-01
-1.61726499e+00 1.07816613e+00 1.15174346e-01 -4.32888299e-01
-4.97879475e-01 8.53789568e-01 3.30939561e-01 -7.28727162e-01
4.30571809e-02 4.10180479e-01 -4.69403148e-01 5.71265757e-01
5.37331998e-01 3.67974281e-01 1.62926674e-01 -9.81243849e-01
-4.26196665e-01 9.46046650e-01 -3.23302835e-01 3.25547636e-01
1.11216426e+00 -1.00369498e-01 -4.17786777e-01 5.20635009e-01
1.46949494e+00 -1.58714667e-01 -1.14409602e+00 -5.11214674e-01
3.23456049e-01 -2.66872972e-01 -6.31693229e-02 -8.76471579e-01
-1.24839783e+00 3.89931470e-01 6.28418744e-01 4.46888894e-01
1.18487895e+00 3.19233119e-01 4.03608978e-01 3.54641169e-01
5.08944094e-01 -9.58040178e-01 -4.13178690e-02 3.19949597e-01
3.59544545e-01 -1.44047463e+00 -1.02348998e-01 -5.25114596e-01
-5.96004725e-01 8.94424200e-01 5.98430276e-01 2.74118304e-01
7.27561057e-01 2.43059650e-01 1.80365622e-01 -3.95745188e-01
-8.37275088e-01 1.73825845e-01 2.34486729e-01 2.38452479e-01
3.10414493e-01 2.98281312e-01 -1.52258620e-01 3.39918852e-01
7.56620690e-02 -2.09885940e-01 6.98615760e-02 6.75882936e-01
-4.82300192e-01 -1.19456661e+00 -5.77120245e-01 7.33508170e-01
-4.05760288e-01 3.72226611e-02 -2.63558149e-01 1.29934505e-01
4.62407976e-01 1.05268729e+00 -5.98524092e-03 -6.42518282e-01
9.52447355e-02 7.18999684e-01 3.44964042e-02 -9.86610770e-01
-6.46989226e-01 4.89936844e-02 -3.24554414e-01 -3.92585635e-01
-2.40973562e-01 -8.83059084e-01 -1.36694896e+00 -6.93018362e-02
-5.16098201e-01 4.48175162e-01 4.44442570e-01 1.00080097e+00
2.43196711e-01 2.42940918e-01 1.24858236e+00 -6.07914329e-01
-5.33648729e-01 -7.82966793e-01 -1.05197537e+00 5.50788581e-01
9.14829448e-02 -6.10767663e-01 -5.36652148e-01 2.71992296e-01]
|
[9.46960735321045, 4.31952428817749]
|
44f32ed0-73ce-4fae-a648-1b3d269fb598
|
massive-a-1m-example-multilingual-natural
|
2204.08582
| null |
https://arxiv.org/abs/2204.08582v2
|
https://arxiv.org/pdf/2204.08582v2.pdf
|
MASSIVE: A 1M-Example Multilingual Natural Language Understanding Dataset with 51 Typologically-Diverse Languages
|
We present the MASSIVE dataset--Multilingual Amazon Slu resource package (SLURP) for Slot-filling, Intent classification, and Virtual assistant Evaluation. MASSIVE contains 1M realistic, parallel, labeled virtual assistant utterances spanning 51 languages, 18 domains, 60 intents, and 55 slots. MASSIVE was created by tasking professional translators to localize the English-only SLURP dataset into 50 typologically diverse languages from 29 genera. We also present modeling results on XLM-R and mT5, including exact match accuracy, intent classification accuracy, and slot-filling F1 score. We have released our dataset, modeling code, and models publicly.
|
['Prem Natarajan', 'Gokhan Tur', 'Wouter Leeuwis', 'Misha Britan', 'Laurie Crist', 'Swetha Ranganath', 'Richa Singh', 'Vishesh Kakarala', 'Liam Urbach', 'Aaron Nash', 'Ana Sanchez', 'Kay Rottmann', 'Scott Mackie', 'Charith Peris', 'Christopher Hench', 'Jack FitzGerald']
|
2022-04-18
| null | null | null | null |
['zero-shot-slot-filling', 'xlm-r', 'slot-filling']
|
['natural-language-processing', 'natural-language-processing', 'natural-language-processing']
|
[-1.27579682e-02 1.03541188e-01 -8.54034722e-01 -4.30208087e-01
-1.20582652e+00 -9.66663241e-01 4.84445274e-01 1.35583058e-01
-7.60588109e-01 9.06333327e-01 5.55877209e-01 -8.61581147e-01
3.44548076e-01 -9.56324711e-02 -5.00904441e-01 4.26359653e-01
4.18217212e-01 1.43549347e+00 -1.96075186e-01 -3.76618862e-01
-6.26159534e-02 -3.22272956e-01 -8.98570657e-01 2.97700614e-01
1.24999404e+00 7.41130948e-01 4.67308134e-01 5.38015068e-01
-4.20622677e-01 9.62882400e-01 -5.27344704e-01 -8.69331598e-01
-1.10266849e-01 3.64603363e-02 -1.11977577e+00 -3.56401026e-01
3.28694403e-01 -2.37813324e-01 -3.14011484e-01 5.32046616e-01
3.99976820e-01 1.76956952e-01 7.28246689e-01 -1.46707225e+00
-1.13392830e+00 1.14487553e+00 -2.04546809e-01 1.80847105e-02
9.90246415e-01 6.69548213e-02 1.25991702e+00 -1.24542785e+00
1.54748940e+00 1.29517829e+00 8.85178864e-01 4.75959092e-01
-1.17309058e+00 -5.86284518e-01 -2.20301479e-01 1.43540621e-01
-1.53601003e+00 -6.00892544e-01 8.11108947e-02 -4.22847867e-01
1.66540205e+00 2.91997731e-01 3.21540147e-01 1.73510945e+00
4.70495634e-02 1.29518032e+00 1.01183200e+00 -3.79328698e-01
9.69518498e-02 5.96817017e-01 5.11442065e-01 7.51987457e-01
-2.28799522e-01 -3.80879454e-03 -8.38468492e-01 -5.44181466e-01
3.44380826e-01 -5.59735715e-01 4.29091677e-02 -9.72637460e-02
-1.24849308e+00 8.62244248e-01 -1.22625686e-01 -5.99123351e-02
-2.70256490e-01 -4.45390821e-01 8.57027352e-01 4.75620508e-01
5.28396010e-01 6.67164743e-01 -9.21481252e-01 -6.09091103e-01
-4.65656102e-01 3.76672864e-01 9.63606656e-01 1.59531844e+00
5.66107571e-01 -3.69942561e-02 -3.59157503e-01 1.48645544e+00
1.58291295e-01 8.45195591e-01 7.59136438e-01 -1.02365136e+00
5.61346054e-01 5.85003197e-01 2.06832483e-01 -4.47409391e-01
-7.68470347e-01 -2.77938277e-01 -2.43670583e-01 -9.28233683e-01
1.36339366e-01 -1.56396434e-01 -8.29556465e-01 1.82081044e+00
1.13471463e-01 -5.83938360e-01 2.14988649e-01 7.34655321e-01
1.19515979e+00 5.31950772e-01 6.57814503e-01 -1.00511305e-01
1.55132318e+00 -1.11166477e+00 -9.11843419e-01 -6.92553222e-01
1.08453858e+00 -1.02157831e+00 1.53797746e+00 -5.98792844e-02
-7.67336965e-01 -4.71737415e-01 -4.90464747e-01 -7.46469140e-01
-4.87195909e-01 3.91325682e-01 9.61699605e-01 5.57671726e-01
-1.01100898e+00 -2.25526080e-01 -4.48224455e-01 -7.01949477e-01
-1.58059597e-01 -1.80524111e-01 -6.35024428e-01 -2.83279926e-01
-1.53534925e+00 1.20404243e+00 3.73089135e-01 -6.14961445e-01
-6.97212815e-01 -7.44159639e-01 -1.40028930e+00 -4.77164656e-01
1.79080725e-01 -4.91465658e-01 1.68881667e+00 -2.43640378e-01
-1.00303316e+00 1.41893470e+00 -4.81456578e-01 -4.25075591e-01
3.93398911e-01 -7.03915805e-02 -5.36368191e-01 -5.93621850e-01
7.01727688e-01 8.28851879e-01 7.84748495e-02 -6.96249068e-01
-6.31655395e-01 -4.79648739e-01 -3.01434606e-01 5.48229754e-01
-2.88248118e-02 2.79148579e-01 -4.48448211e-01 -5.23057401e-01
3.61175351e-02 -9.40264463e-01 -2.43866771e-01 -5.45479655e-01
-4.48310614e-01 -3.13366175e-01 2.36531571e-01 -1.39454019e+00
1.17456067e+00 -2.18789911e+00 9.46474522e-02 -3.35783154e-01
-8.42847899e-02 -1.57891989e-01 -6.44388497e-01 5.11791646e-01
3.40000659e-01 -1.00968629e-02 3.34518701e-02 -9.67043996e-01
4.00751978e-01 3.64829957e-01 -4.63770896e-01 9.77626741e-02
-2.36395642e-01 1.27835047e+00 -8.21586013e-01 -6.15661979e-01
2.26863727e-01 4.10148464e-02 -6.39917970e-01 1.96047083e-01
-3.48056197e-01 1.49191648e-01 2.05845281e-01 1.00837219e+00
4.64177907e-01 -5.98679818e-02 5.49818099e-01 1.77258849e-01
2.96792928e-02 8.02594602e-01 -6.15678370e-01 2.12775183e+00
-8.91747475e-01 5.82468331e-01 1.38452500e-01 -1.75048202e-01
9.60763454e-01 1.09648585e-01 3.02004367e-01 -1.03596210e+00
-8.31528306e-02 6.97996438e-01 -5.14892042e-01 -7.24182054e-02
1.73253798e+00 3.57224941e-01 -9.72572327e-01 4.44906920e-01
6.46110296e-01 -1.96617797e-01 1.04963787e-01 3.89013797e-01
1.08024085e+00 6.96207136e-02 6.89404309e-01 -1.19715437e-01
-3.05472314e-02 6.51233971e-01 5.32727003e-01 9.33997989e-01
-3.99686396e-01 -1.16671398e-01 2.68259794e-01 -3.12877595e-01
-1.09950304e+00 -1.26787829e+00 -4.21084732e-01 1.82836425e+00
4.68321554e-02 -6.85119092e-01 -3.48715901e-01 -6.03881598e-01
2.06429884e-01 1.16630065e+00 -1.81315169e-01 3.88513096e-02
-2.28514060e-01 -1.51895478e-01 8.92243683e-01 4.31444705e-01
9.74534750e-02 -1.16036737e+00 1.21365488e-01 2.40011141e-01
-9.47632849e-01 -1.66196024e+00 -7.51738787e-01 4.32804912e-01
-2.02671051e-01 -7.94119418e-01 -3.97985816e-01 -1.06368792e+00
1.23030823e-02 2.34423261e-02 1.63310468e+00 -5.04328430e-01
-2.72112370e-01 3.70299876e-01 -3.72123092e-01 -5.28884307e-02
-6.43423617e-01 6.55617893e-01 6.38439357e-01 -1.09191966e+00
8.26567888e-01 -5.20718396e-02 3.91157091e-01 4.28022146e-01
-2.31560856e-01 2.41047278e-01 1.80186734e-01 9.89183724e-01
6.52463436e-01 -7.52882600e-01 2.40896508e-01 -9.16726112e-01
7.02544570e-01 -7.39393175e-01 -4.09120947e-01 5.37483275e-01
-3.06312203e-01 -2.72752583e-01 3.93655598e-01 -3.13273609e-01
-8.94926310e-01 -1.30137235e-01 -4.88079071e-01 -1.05037853e-01
3.67644764e-02 4.86027390e-01 -4.98042889e-02 3.54248792e-01
9.33811724e-01 3.85487437e-01 -2.92075306e-01 -4.32778329e-01
8.13269854e-01 1.22645402e+00 8.98684323e-01 -8.72227550e-01
1.13473199e-01 -4.36582774e-01 -1.00979006e+00 -9.59980130e-01
-5.59814274e-01 -6.28680825e-01 -3.73690695e-01 1.11492433e-01
5.36641657e-01 -1.25178802e+00 -6.14083767e-01 4.71902609e-01
-1.29541802e+00 -6.68128252e-01 -2.63126373e-01 2.42817447e-01
-6.06058657e-01 2.98722219e-02 -1.15614295e+00 -8.32602441e-01
-5.38348317e-01 -1.17353082e+00 1.22935092e+00 3.15331630e-02
-1.13884199e+00 -9.28268373e-01 1.71093330e-01 7.44521499e-01
5.00746965e-01 -7.32574940e-01 9.78540003e-01 -1.11426723e+00
-8.36482737e-03 1.85777113e-01 -4.08264518e-01 -3.07223290e-01
-1.30316734e-01 -2.80397475e-01 -6.08153462e-01 -6.79067522e-02
-6.52272999e-01 -1.04792988e+00 2.10089579e-01 2.66547859e-01
4.97523695e-01 -2.50806451e-01 -4.14508581e-01 5.49454331e-01
9.14250076e-01 2.24240810e-01 1.20875083e-01 3.15370798e-01
4.80918676e-01 4.29876566e-01 1.02196145e+00 5.08615375e-01
1.00750160e+00 9.43760455e-01 -2.70803034e-01 2.30643168e-01
1.12284534e-01 -9.67319489e-01 4.70017165e-01 1.01892328e+00
6.79030836e-01 -2.30854079e-01 -1.15706146e+00 6.17356598e-01
-1.60568702e+00 -3.37296873e-01 2.78430820e-01 1.63322246e+00
1.01822376e+00 -1.43648699e-01 1.47336602e-01 -6.77178204e-01
3.69528294e-01 2.15519130e-01 -6.03583753e-01 -5.63535631e-01
-3.04443985e-01 6.59210160e-02 4.60586250e-01 9.74193454e-01
-8.80652845e-01 1.79013216e+00 7.47865534e+00 9.85030651e-01
-6.78733706e-01 5.05744457e-01 5.16863227e-01 1.87941551e-01
-8.41382384e-01 -1.80050239e-01 -1.06515455e+00 4.09593582e-01
1.14926541e+00 -4.29283887e-01 7.82401085e-01 1.58001804e+00
-4.84729201e-01 -9.36069787e-02 -9.02218103e-01 1.00335896e+00
-5.46962805e-02 -1.09217298e+00 -1.00242287e-01 -2.26223275e-01
3.70545417e-01 5.89682281e-01 1.35049999e-01 1.33463001e+00
9.91623878e-01 -8.90667081e-01 8.51953030e-01 -4.49892879e-03
1.48567891e+00 -4.35796916e-01 5.66627741e-01 3.32341790e-01
-1.10260081e+00 6.95691258e-02 -3.95912617e-01 -3.27649377e-02
4.11554158e-01 3.99431773e-02 -1.23406637e+00 3.35782140e-01
4.19357419e-01 4.51929569e-01 -5.22246957e-01 1.46360740e-01
-1.28975406e-01 2.69823998e-01 -2.67661482e-01 -1.86143264e-01
-1.48685044e-02 -1.05198815e-01 4.89525259e-01 1.28652000e+00
-3.06386799e-02 -2.38026083e-01 5.48137486e-01 6.41380608e-01
-2.59121180e-01 5.29249132e-01 -8.46558571e-01 -5.81771195e-01
1.17887914e+00 1.12983036e+00 -3.41157854e-01 -2.83125371e-01
-2.03344032e-01 1.47957444e+00 6.35195315e-01 2.68693030e-01
-6.99370980e-01 -4.23713811e-02 1.12766314e+00 -3.65468562e-01
-6.19727731e-01 -3.82012904e-01 -4.35397476e-01 -1.20437562e+00
-1.52033404e-01 -1.13964248e+00 5.53415120e-01 -8.46533060e-01
-1.36453855e+00 8.66128385e-01 -3.11923802e-01 -5.38794219e-01
-8.62080038e-01 -8.01925898e-01 3.36377621e-01 1.02170098e+00
-8.37352395e-01 -1.36901963e+00 6.93121254e-02 3.42404515e-01
9.67393696e-01 -6.01945460e-01 1.23990512e+00 4.65381801e-01
-4.85772908e-01 1.01353848e+00 -1.09021187e-01 1.70440376e-01
7.59011447e-01 -1.20912051e+00 1.27073348e+00 4.49248016e-01
1.57035455e-01 7.72361457e-01 7.81519830e-01 -9.59973752e-01
-1.27635419e+00 -9.46003079e-01 1.52520585e+00 -9.27943707e-01
9.00102377e-01 -6.97570324e-01 -5.31295240e-01 1.27576900e+00
7.43519664e-02 -4.29113865e-01 9.02296722e-01 7.85748780e-01
-5.38908660e-01 6.86370969e-01 -1.16923332e+00 7.08237886e-01
1.48365879e+00 -1.22632527e+00 -5.13743877e-01 5.77209055e-01
1.32318091e+00 -1.01144314e+00 -8.43303263e-01 2.41171703e-01
6.44987166e-01 -2.86125153e-01 7.03254104e-01 -9.27523136e-01
3.55486900e-01 4.04300004e-01 -8.03893089e-01 -1.39660347e+00
-1.86443701e-01 -2.47084364e-01 -5.46018258e-02 1.16862762e+00
8.25802088e-01 -6.78937197e-01 4.62192029e-01 6.40349567e-01
-3.18991661e-01 -4.67379898e-01 -1.17277420e+00 -6.80155218e-01
-1.09077148e-01 -9.45040524e-01 5.30037761e-01 1.24579322e+00
4.51150328e-01 7.10056782e-01 -4.86608565e-01 -2.18997493e-01
3.91588122e-01 -1.85877427e-01 7.76126623e-01 -8.55302811e-01
-4.05333608e-01 -1.17947817e-01 6.35589659e-02 -1.01815677e+00
6.91178381e-01 -1.15248084e+00 -1.01819746e-02 -1.42408264e+00
2.71998376e-01 -1.02667642e+00 3.03153217e-01 9.29484367e-01
-1.86009720e-01 1.02466367e-01 -5.50559238e-02 1.42663255e-01
-8.52451682e-01 4.85913754e-01 5.82170248e-01 -2.16634035e-01
-1.79195717e-01 -3.75211746e-01 -5.93931139e-01 2.93407917e-01
4.31008548e-01 -1.89096600e-01 -3.45619738e-01 -5.71402490e-01
8.83868188e-02 5.03643930e-01 -2.49822035e-01 -6.38009965e-01
-9.06154215e-02 -1.93762958e-01 1.98839232e-01 -6.87330067e-01
5.85343659e-01 -4.00760949e-01 1.31917313e-01 4.22938079e-01
-5.22428811e-01 4.54196572e-01 4.84362394e-01 -2.39785969e-01
1.28409177e-01 8.37824121e-03 2.13444471e-01 -1.19108662e-01
-1.25058508e+00 2.16727167e-01 -6.19036257e-01 5.11640251e-01
5.34096837e-01 4.28350031e-01 -6.53581083e-01 -4.89890635e-01
-6.61010623e-01 5.15099823e-01 8.55703175e-01 1.16524827e+00
2.66351581e-01 -1.42549765e+00 -6.09621584e-01 4.05919373e-01
7.68654823e-01 -6.67189837e-01 4.34684932e-01 3.87571037e-01
-3.32684129e-01 9.36311662e-01 -2.54957438e-01 -4.74011958e-01
-1.12678087e+00 4.25759941e-01 1.22192524e-01 -4.29934889e-01
-1.66267142e-01 1.03493381e+00 -3.45709473e-01 -1.81387305e+00
1.88586414e-01 5.41421212e-02 1.75909936e-01 3.35281417e-02
4.44332629e-01 2.75256515e-01 -1.02256916e-01 -8.08763444e-01
-6.52282357e-01 -2.71546692e-01 -3.12470585e-01 -6.86476052e-01
5.26022375e-01 -2.78587401e-01 1.57244578e-02 6.43249094e-01
1.02947998e+00 1.98924005e-01 -1.66307524e-01 -6.29984200e-01
2.54547060e-01 -1.84309721e-01 -5.38120985e-01 -1.15183830e+00
-2.33243778e-01 2.28317037e-01 4.09026116e-01 -4.10497099e-01
4.96069938e-01 1.81064934e-01 9.83598650e-01 3.84195983e-01
1.04078209e+00 -1.33537602e+00 -2.54173905e-01 1.36041903e+00
6.95196092e-01 -1.43177176e+00 -5.95968068e-01 -1.70146301e-01
-1.23308659e+00 1.50279716e-01 9.97212470e-01 8.22809696e-01
1.02426901e-01 3.34858418e-01 5.39181888e-01 1.28361732e-01
-8.09214711e-01 6.96292743e-02 7.12547265e-03 8.46741676e-01
7.77593672e-01 7.40396738e-01 -4.30334389e-01 1.04940796e+00
-9.11897421e-01 -8.97746608e-02 1.65287271e-01 7.04619110e-01
-6.44040778e-02 -1.11852646e+00 2.50754622e-03 4.63543802e-01
-1.90770011e-02 -4.98011738e-01 -4.14823800e-01 6.31103456e-01
-2.54887372e-01 8.03129613e-01 1.96331158e-01 -9.02132571e-01
2.76266098e-01 4.43182856e-01 1.28170714e-01 -6.62006021e-01
-3.06410879e-01 -7.02878714e-01 7.39271104e-01 -6.51678205e-01
5.94421864e-01 -7.27578580e-01 -1.09488523e+00 -5.48200548e-01
1.34931818e-01 3.29312801e-01 8.81417632e-01 7.67049253e-01
4.66883689e-01 9.66275260e-02 2.70067543e-01 -3.21483046e-01
-4.32268977e-01 -1.48395753e+00 -4.89889771e-01 1.04902089e-01
-9.38967392e-02 -5.40702939e-01 -7.89423510e-02 -6.88070118e-01]
|
[12.192140579223633, 8.619315147399902]
|
dd0b2e5d-a542-45b6-821b-dc8a1c4d19aa
|
unsupervised-contrastive-learning-based
|
2205.00122
| null |
https://arxiv.org/abs/2205.00122v1
|
https://arxiv.org/pdf/2205.00122v1.pdf
|
Unsupervised Contrastive Learning based Transformer for Lung Nodule Detection
|
Early detection of lung nodules with computed tomography (CT) is critical for the longer survival of lung cancer patients and better quality of life. Computer-aided detection/diagnosis (CAD) is proven valuable as a second or concurrent reader in this context. However, accurate detection of lung nodules remains a challenge for such CAD systems and even radiologists due to not only the variability in size, location, and appearance of lung nodules but also the complexity of lung structures. This leads to a high false-positive rate with CAD, compromising its clinical efficacy. Motivated by recent computer vision techniques, here we present a self-supervised region-based 3D transformer model to identify lung nodules among a set of candidate regions. Specifically, a 3D vision transformer (ViT) is developed that divides a CT image volume into a sequence of non-overlap cubes, extracts embedding features from each cube with an embedding layer, and analyzes all embedding features with a self-attention mechanism for the prediction. To effectively train the transformer model on a relatively small dataset, the region-based contrastive learning method is used to boost the performance by pre-training the 3D transformer with public CT images. Our experiments show that the proposed method can significantly improve the performance of lung nodule screening in comparison with the commonly used 3D convolutional neural networks.
|
['Ge Wang', 'Chuang Niu']
|
2022-04-30
| null | null | null | null |
['lung-nodule-detection']
|
['medical']
|
[ 1.50453463e-01 1.06348895e-01 -2.38545388e-01 -2.00274084e-02
-9.76393104e-01 -4.04588223e-01 3.52147639e-01 1.05798647e-01
-2.78398544e-01 8.43259878e-03 1.63196400e-01 -7.63111591e-01
2.62705889e-02 -9.56304729e-01 -5.01529455e-01 -6.01146638e-01
-1.38931200e-01 7.90015578e-01 9.03387904e-01 2.90919393e-01
-3.38704228e-01 8.20433140e-01 -1.07311726e+00 5.06369472e-01
6.67107821e-01 1.18814898e+00 5.07929087e-01 6.63393676e-01
-2.35697865e-01 9.29156661e-01 3.68444435e-02 -1.53508782e-01
3.45799357e-01 -5.73280573e-01 -7.52084076e-01 2.08816007e-01
1.04108632e-01 -6.75671041e-01 -4.27292407e-01 8.18265498e-01
4.32155401e-01 -3.74327153e-01 9.33995783e-01 -7.32618451e-01
-2.32018188e-01 3.47818613e-01 -4.32758421e-01 3.54386926e-01
-2.20736548e-01 2.29414001e-01 8.61719847e-01 -1.15980268e+00
2.22525135e-01 7.29803801e-01 6.94149911e-01 5.87570846e-01
-5.00687599e-01 -5.00617504e-01 -2.66323298e-01 5.16383946e-02
-1.13412356e+00 2.09319428e-01 6.04356110e-01 -5.82499564e-01
6.05537593e-01 3.24757755e-01 1.12889004e+00 6.81764305e-01
3.99982125e-01 6.51764750e-01 6.92276478e-01 -2.13733256e-01
9.19464529e-02 2.84928888e-01 -3.15194637e-01 1.29334712e+00
4.27497953e-01 2.21368417e-01 2.80263722e-01 -2.67498851e-01
1.18353319e+00 4.72290158e-01 -1.50779843e-01 -7.63409436e-01
-1.37254179e+00 9.18361664e-01 1.20770681e+00 4.77235883e-01
-4.24356580e-01 7.64958262e-02 3.36054087e-01 -3.80678535e-01
3.01996738e-01 2.71181017e-01 -8.08901805e-03 5.53437293e-01
-7.53644943e-01 -1.78130090e-01 5.67703903e-01 4.26119298e-01
2.43852556e-01 -2.24249274e-01 -5.88695645e-01 5.56111991e-01
3.84946555e-01 4.41473693e-01 7.38363743e-01 -2.48382568e-01
3.58938098e-01 1.05004501e+00 -2.82335132e-01 -6.19512260e-01
-4.18317527e-01 -6.98729873e-01 -1.18397808e+00 7.08090290e-02
1.86406836e-01 2.79828817e-01 -1.02042878e+00 1.08667660e+00
4.70320255e-01 1.55683905e-01 -3.32220376e-01 9.71512735e-01
8.70480418e-01 3.01674038e-01 7.62465894e-02 1.03359289e-01
1.62079859e+00 -1.12620854e+00 -8.94842967e-02 -3.59929681e-01
8.31700563e-01 -4.13178712e-01 8.29294503e-01 -3.25219005e-01
-9.11391556e-01 -4.22003686e-01 -8.15700829e-01 3.12380260e-03
-1.10182166e-01 5.73689103e-01 2.16554955e-01 5.45004904e-01
-8.36487770e-01 2.60899574e-01 -1.24832189e+00 -3.66986066e-01
8.99045408e-01 4.02661979e-01 -2.21795291e-01 -2.77211368e-01
-8.34280729e-01 9.81772363e-01 3.58353376e-01 -4.30931151e-02
-1.01598346e+00 -7.55791843e-01 -5.90268433e-01 2.82133579e-01
4.08924609e-01 -8.20961356e-01 1.35276186e+00 -3.86352360e-01
-9.82875049e-01 9.65070486e-01 8.16942602e-02 -4.53925222e-01
6.76917195e-01 1.78673312e-01 8.41241777e-02 4.03070331e-01
7.87708834e-02 5.70420086e-01 7.96840131e-01 -8.50264430e-01
-6.96552038e-01 -2.85073966e-01 -3.44576776e-01 3.05749863e-01
-4.22366351e-01 -1.28129005e-01 -7.02058196e-01 -3.85027438e-01
2.94916898e-01 -1.02286267e+00 -6.99708760e-01 5.83141565e-01
-3.31873864e-01 -1.93674892e-01 9.42720413e-01 -5.43656349e-01
9.72920954e-01 -2.02327275e+00 -2.65958160e-01 3.12650621e-01
7.92840242e-01 1.43097252e-01 2.34336883e-01 -2.31318399e-01
-7.73874745e-02 1.19274713e-01 -2.95963615e-01 2.70413965e-01
-4.14593160e-01 -5.50224781e-02 2.57401109e-01 3.64145070e-01
4.81653273e-01 1.31868422e+00 -8.10589671e-01 -1.13565660e+00
6.14904523e-01 4.60218847e-01 -5.09308517e-01 4.88180339e-01
-6.90466240e-02 4.19790357e-01 -8.66252780e-01 7.36310124e-01
3.89396757e-01 -9.51388121e-01 9.07812491e-02 -2.77056098e-01
-5.72963059e-03 1.58094063e-01 -5.70773065e-01 1.04732823e+00
-3.10279518e-01 2.07789853e-01 -2.60649979e-01 -7.45479286e-01
6.05925739e-01 5.43971241e-01 8.76595199e-01 -3.97757024e-01
2.68183559e-01 4.38428462e-01 4.00136918e-01 -5.72541237e-01
-3.24399412e-01 -1.72092125e-01 1.57467589e-01 5.92409432e-01
-3.36658925e-01 -4.15140152e-01 -2.82976300e-01 1.35576621e-01
1.54656029e+00 -5.40855348e-01 8.04263175e-01 6.84304610e-02
5.68542898e-01 3.44567060e-01 1.65249124e-01 4.88305211e-01
-4.70307738e-01 7.36702979e-01 3.17632377e-01 -3.74656439e-01
-1.14759326e+00 -1.17388844e+00 -2.21837252e-01 4.54515308e-01
-1.40290484e-01 2.42144063e-01 -4.03572977e-01 -1.27595913e+00
-2.62521822e-02 1.80815503e-01 -7.53792942e-01 -2.92746484e-01
-7.48687387e-01 -5.94855905e-01 3.04325342e-01 9.53166366e-01
6.63594782e-01 -9.05570626e-01 -8.71155381e-01 3.14257890e-02
-2.01252595e-01 -9.28316891e-01 -5.40452421e-01 5.69652796e-01
-1.16799164e+00 -1.37146533e+00 -1.04818869e+00 -1.11643028e+00
9.82137382e-01 6.63665652e-01 1.01196277e+00 3.29891443e-01
-5.93134522e-01 2.30486751e-01 -1.10292554e-01 -4.19520229e-01
-6.52131259e-01 8.82096961e-02 -4.27264124e-01 -3.53409588e-01
1.84697583e-01 1.45319104e-02 -9.32113826e-01 3.50103825e-01
-8.05761397e-01 3.97214517e-02 1.32980049e+00 9.03445661e-01
8.32224905e-01 2.38450453e-01 7.98348337e-02 -9.28427875e-01
1.89896584e-01 -6.69106781e-01 -4.82847124e-01 1.50271624e-01
-3.15641731e-01 -9.71124619e-02 5.95319748e-01 -2.72156537e-01
-6.74953163e-01 4.54970837e-01 7.39416257e-02 -6.20421410e-01
1.05575845e-01 3.85852665e-01 2.10017651e-01 -2.95331508e-01
7.44224727e-01 3.54499370e-01 1.43618450e-01 -9.44931358e-02
-7.87672549e-02 4.95086491e-01 3.93891841e-01 3.06358606e-01
1.21038628e+00 4.22370076e-01 2.09032163e-01 -4.68416452e-01
-9.88749683e-01 -8.68859410e-01 -8.08993161e-01 -2.88197905e-01
8.91340613e-01 -1.00349820e+00 -7.27863237e-02 -4.07933407e-02
-8.77533197e-01 -5.79695329e-02 -3.10810357e-01 7.95774043e-01
-2.03269675e-01 2.74805278e-01 -4.92370903e-01 -4.55615252e-01
-4.72562402e-01 -1.20453930e+00 1.01463461e+00 2.93544456e-02
3.40077886e-03 -9.86224055e-01 1.40518614e-03 3.32170755e-01
5.40091217e-01 -1.28943231e-02 1.27221537e+00 -9.50534523e-01
-8.97184730e-01 -6.67559087e-01 -7.17853546e-01 3.73691171e-01
3.44832301e-01 -1.87552363e-01 -7.80440032e-01 -1.53551996e-01
1.33670792e-01 -2.90255785e-01 8.83409083e-01 5.32428741e-01
1.49241209e+00 -6.92958161e-02 -8.76246452e-01 4.08067077e-01
1.40104401e+00 1.55168280e-01 1.79213181e-01 9.09546018e-02
8.90777886e-01 2.20812634e-01 4.08726037e-01 2.32649654e-01
1.28925323e-01 1.32294253e-01 8.44284952e-01 -5.19781470e-01
-2.42636904e-01 -2.19553292e-01 -7.94636980e-02 7.35259831e-01
-2.92644836e-03 -1.23780437e-01 -1.11187625e+00 5.76081872e-01
-1.26916766e+00 -7.55660474e-01 -7.04262704e-02 2.11591387e+00
5.78915894e-01 1.58645779e-01 6.30321726e-02 5.65210842e-02
6.90285623e-01 -7.14960173e-02 -6.62832022e-01 3.33066285e-01
4.57380533e-01 2.71482557e-01 5.66416442e-01 -3.21101248e-02
-1.44507957e+00 3.68354470e-01 6.10011435e+00 6.60600424e-01
-1.31763113e+00 1.22758329e-01 7.54104912e-01 1.32817760e-01
-1.06293529e-01 -3.77451926e-01 -5.03926933e-01 1.57840192e-01
4.46003646e-01 1.72379008e-03 -1.84218124e-01 1.14278066e+00
9.31896120e-02 9.57764536e-02 -1.24982059e+00 7.05242872e-01
3.02933846e-02 -1.35229635e+00 9.41444486e-02 2.48196974e-01
5.46945274e-01 2.70976931e-01 1.92346334e-01 2.73268312e-01
3.55859622e-02 -1.18011665e+00 2.04879686e-01 9.09122080e-02
9.46377099e-01 -3.73896122e-01 1.13726163e+00 4.97316152e-01
-1.44100678e+00 -1.75761625e-01 -3.36334705e-01 6.74368203e-01
-3.28473151e-01 5.94236434e-01 -1.74881899e+00 2.95975387e-01
6.83442891e-01 4.13298965e-01 -8.11697245e-01 1.44336247e+00
1.78987309e-01 6.96861982e-01 -4.15187091e-01 -2.57525682e-01
3.11591536e-01 1.90918699e-01 2.36088395e-01 9.23589230e-01
6.81187451e-01 2.01207146e-01 2.73635626e-01 8.73319626e-01
-1.55235052e-01 2.21906275e-01 -7.40781486e-01 -5.72696589e-02
4.87208605e-01 1.39789736e+00 -1.04699934e+00 -3.09389263e-01
-6.93195760e-01 8.66441190e-01 2.01857999e-01 -3.02507013e-01
-1.05705905e+00 7.48386309e-02 -1.83737189e-01 6.13411844e-01
5.03146112e-01 2.20304474e-01 -2.61568338e-01 -8.65169525e-01
-3.48907709e-02 -5.59971869e-01 5.09752035e-01 -5.04284263e-01
-1.20466888e+00 6.94457293e-01 -2.05943376e-01 -1.76016402e+00
-1.95656151e-01 -8.06345284e-01 -8.82638037e-01 5.61440170e-01
-1.49661911e+00 -1.25222349e+00 -7.29197860e-01 4.40456539e-01
4.18461353e-01 -1.21228464e-01 7.95587063e-01 5.38827963e-02
-4.51631516e-01 3.15124363e-01 -1.95295438e-02 5.09969175e-01
4.69026417e-01 -1.31337845e+00 1.85640186e-01 4.50928092e-01
-1.28394842e-01 -1.22443199e-01 -1.14431947e-01 -5.73619723e-01
-1.23520100e+00 -1.78849471e+00 7.50484288e-01 -5.82703829e-01
4.05472130e-01 -3.29721645e-02 -8.91180754e-01 5.45656383e-01
-3.11396092e-01 6.44651234e-01 6.38390124e-01 -6.84046924e-01
-9.89378393e-02 8.83234218e-02 -1.11591768e+00 5.29418945e-01
7.35914111e-01 -3.16681087e-01 -4.04694915e-01 5.09659350e-01
5.90776205e-01 -3.84479135e-01 -7.23397672e-01 6.45461500e-01
3.88169795e-01 -1.01046205e+00 1.20582426e+00 -2.34897628e-01
5.12471080e-01 -8.59645680e-02 1.48621082e-01 -9.09879208e-01
-7.05011129e-01 4.10222709e-01 1.22305773e-01 3.31843287e-01
4.51384634e-01 -3.01819623e-01 1.23994279e+00 3.65970254e-01
-1.37487352e-01 -1.11594081e+00 -8.39392483e-01 -4.99699086e-01
2.08901942e-01 -2.16939345e-01 2.53576100e-01 5.55079818e-01
-4.72721010e-01 5.90311410e-03 3.81861240e-01 2.89618194e-01
5.07779479e-01 3.24711129e-02 3.40462536e-01 -1.26506758e+00
-2.93474048e-01 -5.88590384e-01 -4.17772204e-01 -9.07949328e-01
-4.78938103e-01 -1.28763878e+00 5.80301173e-02 -1.71692085e+00
7.72075653e-01 -4.24395591e-01 -4.32197660e-01 3.88902813e-01
-3.25940758e-01 3.25243652e-01 -1.90451711e-01 4.37687367e-01
-4.05547142e-01 4.25536066e-01 1.74890482e+00 -2.56542057e-01
6.78885402e-03 5.53849518e-01 -5.18518806e-01 8.08593154e-01
5.16332328e-01 -6.45423234e-01 -3.60882103e-01 -1.64469391e-01
-1.91848308e-01 3.23299766e-01 7.83936322e-01 -1.23777938e+00
2.59405106e-01 -3.21450047e-02 8.67119789e-01 -1.17497635e+00
2.54152536e-01 -1.14078212e+00 -3.12333047e-01 1.14893520e+00
-1.59895137e-01 -2.83248484e-01 -4.63191755e-02 6.73028350e-01
-2.36620709e-01 -1.21448085e-01 9.59513903e-01 -5.35597742e-01
-2.57066458e-01 7.44008601e-01 -5.05034268e-01 -1.71873853e-01
1.30944335e+00 -3.34280252e-01 1.16186745e-01 -1.17965378e-01
-3.70975941e-01 1.54498041e-01 1.57091975e-01 -9.49186683e-02
8.91118288e-01 -1.47300649e+00 -8.17509890e-01 3.00207078e-01
2.80831218e-01 6.43637538e-01 1.17894106e-01 1.08415163e+00
-8.57500732e-01 6.63798034e-01 9.04233828e-02 -8.97889316e-01
-1.26659894e+00 4.55565810e-01 7.00711906e-01 -8.31206203e-01
-7.24155962e-01 9.41293418e-01 5.62803626e-01 -4.01882559e-01
2.25510240e-01 -6.64849460e-01 -2.91273773e-01 -4.39580530e-01
1.32825270e-01 3.35404947e-02 1.68681040e-01 -1.44991770e-01
-4.65050310e-01 5.06552398e-01 -2.43357569e-01 2.18128860e-01
9.98361886e-01 3.45474660e-01 2.72199232e-02 1.98285952e-01
1.18943918e+00 -1.41524151e-01 -8.70168388e-01 -4.23418552e-01
-1.26257420e-01 -2.24718615e-01 2.98439205e-01 -6.21025443e-01
-1.16088140e+00 9.93600547e-01 8.09825540e-01 7.60892183e-02
1.06384230e+00 4.87748593e-01 7.92788804e-01 4.76926416e-01
-1.20203122e-01 -3.29118460e-01 6.27985120e-01 1.87475502e-01
6.76284730e-01 -1.71080589e+00 1.16102897e-01 -5.93463659e-01
-6.49232268e-01 1.24550140e+00 8.78818572e-01 -1.18302971e-01
9.57322419e-01 1.79586023e-01 8.10252503e-02 -3.54008585e-01
-7.75671661e-01 -2.88302600e-01 5.66576719e-01 4.67605501e-01
4.73982155e-01 1.10328272e-01 2.63915241e-01 6.81633711e-01
1.30707562e-01 -6.17798716e-02 4.69319196e-03 9.64838445e-01
-8.23199868e-01 -6.65843368e-01 -5.47851503e-01 9.58432734e-01
-2.47830674e-01 -1.28837237e-02 -6.10136509e-01 1.04864621e+00
4.74336669e-02 3.20791960e-01 2.11199939e-01 -2.98436075e-01
1.78616703e-01 -8.13834295e-02 3.47039163e-01 -9.18401539e-01
-6.50073647e-01 3.31835836e-01 -3.25091183e-01 -2.66466975e-01
-5.88744618e-02 -5.42673409e-01 -1.39243078e+00 1.89278558e-01
-5.10067761e-01 -1.56238163e-02 3.76372188e-01 7.75290668e-01
8.17978084e-02 7.47011423e-01 1.09149957e+00 -4.49656039e-01
-9.55798388e-01 -9.06057060e-01 -3.09198439e-01 5.19502955e-03
2.52098918e-01 -4.35365766e-01 -3.56563270e-01 -1.49966717e-01]
|
[15.400002479553223, -2.1304943561553955]
|
7b250b60-dfd9-4f05-abaf-deab4468c724
|
a-cascaded-approach-for-ultraly-high
|
2306.16036
| null |
https://arxiv.org/abs/2306.16036v1
|
https://arxiv.org/pdf/2306.16036v1.pdf
|
A Cascaded Approach for ultraly High Performance Lesion Detection and False Positive Removal in Liver CT Scans
|
Liver cancer has high morbidity and mortality rates in the world. Multi-phase CT is a main medical imaging modality for detecting/identifying and diagnosing liver tumors. Automatically detecting and classifying liver lesions in CT images have the potential to improve the clinical workflow. This task remains challenging due to liver lesions' large variations in size, appearance, image contrast, and the complexities of tumor types or subtypes. In this work, we customize a multi-object labeling tool for multi-phase CT images, which is used to curate a large-scale dataset containing 1,631 patients with four-phase CT images, multi-organ masks, and multi-lesion (six major types of liver lesions confirmed by pathology) masks. We develop a two-stage liver lesion detection pipeline, where the high-sensitivity detecting algorithms in the first stage discover as many lesion proposals as possible, and the lesion-reclassification algorithms in the second stage remove as many false alarms as possible. The multi-sensitivity lesion detection algorithm maximizes the information utilization of the individual probability maps of segmentation, and the lesion-shuffle augmentation effectively explores the texture contrast between lesions and the liver. Independently tested on 331 patient cases, the proposed model achieves high sensitivity and specificity for malignancy classification in the multi-phase contrast-enhanced CT (99.2%, 97.1%, diagnosis setting) and in the noncontrast CT (97.3%, 95.7%, screening setting).
|
['Ling Zhang', 'Chien-Hung Liao', 'Le Lu', 'Min Wu', 'Ke Yan', 'Chien-Wei Peng', 'Chi-Tung Cheng', 'Fakai Wang']
|
2023-06-28
| null | null | null | null |
['specificity']
|
['natural-language-processing']
|
[-1.21862572e-02 -2.31124699e-01 -1.65775567e-01 -5.02598137e-02
-1.03401685e+00 -7.31453776e-01 4.18571234e-01 4.81232315e-01
-1.97174907e-01 1.78921476e-01 1.10229701e-01 -5.84219694e-01
-1.09506719e-01 -5.26533306e-01 -1.62966847e-01 -1.03628862e+00
-5.29860914e-01 9.34233069e-01 5.84100425e-01 6.30807161e-01
-1.23688884e-01 7.34921515e-01 -6.25081539e-01 5.00524700e-01
9.10135746e-01 7.65738666e-01 4.65894848e-01 9.50899363e-01
1.10804074e-01 5.59241354e-01 -1.74324885e-01 5.56009077e-02
3.64698529e-01 -6.39876902e-01 -6.60620451e-01 4.52984095e-01
7.53285140e-02 -9.48591828e-02 -1.25095129e-01 1.14764321e+00
6.13827229e-01 -4.52744007e-01 8.06178093e-01 -7.11461306e-01
-2.25604892e-01 6.41279817e-01 -7.16318190e-01 7.39403009e-01
-9.47585702e-02 7.09578276e-01 3.58873159e-01 -1.11135602e+00
2.68781602e-01 6.34596884e-01 6.62557304e-01 3.14270347e-01
-1.09229350e+00 -4.42616701e-01 -3.72660100e-01 -1.23702101e-02
-1.65192544e+00 4.57183234e-02 3.35212909e-02 -7.45658755e-01
5.70610762e-01 6.97845638e-01 1.03060281e+00 2.11904511e-01
4.89949524e-01 6.05560422e-01 1.34629357e+00 -3.41740251e-01
-1.03439853e-01 2.08253235e-01 -1.73973992e-01 1.22543216e+00
5.36382675e-01 1.02503844e-01 2.34380484e-01 -4.57263499e-01
8.89022470e-01 3.40092391e-01 -6.15192831e-01 -5.68137705e-01
-1.70740783e+00 6.55159712e-01 6.06494844e-01 4.59264010e-01
-7.22060800e-01 -2.83643961e-01 1.92537710e-01 -2.61238486e-01
8.89425725e-03 2.90224224e-01 -2.55576402e-01 5.20573080e-01
-9.31163013e-01 -5.16882539e-01 7.56970108e-01 6.16612613e-01
9.89416391e-02 -2.99944133e-01 -5.20424426e-01 6.45494699e-01
4.19387192e-01 5.41537702e-01 9.04879928e-01 4.31830203e-03
-1.59858257e-01 7.26954758e-01 -3.38480949e-01 -4.37928796e-01
-8.66753519e-01 -6.86325729e-01 -1.24162209e+00 1.13243833e-02
4.41763252e-01 -9.15042683e-02 -1.38404799e+00 8.99293244e-01
4.84718442e-01 1.30320668e-01 -1.75856307e-01 1.16899824e+00
8.95480275e-01 3.04578155e-01 5.27955890e-01 -4.40236032e-01
2.04587793e+00 -1.00114477e+00 -2.94707686e-01 -4.83932532e-02
5.11480212e-01 -8.90687287e-01 7.55693853e-01 -9.02410597e-02
-9.59594727e-01 -3.49353850e-02 -7.63155580e-01 7.17633545e-01
1.37961311e-02 4.76277292e-01 5.43182313e-01 7.72258162e-01
-8.22113276e-01 5.47162965e-02 -1.20005131e+00 -4.62607592e-01
6.04475677e-01 4.58789796e-01 -1.63250908e-01 -3.02480072e-01
-5.88900089e-01 1.26560819e+00 5.42064786e-01 -6.10733554e-02
-1.15031385e+00 -1.01630342e+00 -7.08815634e-01 1.29588842e-01
4.94247079e-01 -8.56746912e-01 1.05220222e+00 -4.93926585e-01
-1.05113626e+00 9.65969205e-01 -4.60261712e-03 -2.13350981e-01
6.42936409e-01 6.76445663e-01 -2.68886566e-01 5.43483615e-01
2.14743670e-02 4.83940214e-01 5.10203540e-01 -1.05489957e+00
-8.65873992e-01 -1.75437629e-01 -6.84709847e-01 3.64854395e-01
3.36187571e-01 1.73754752e-01 -4.78052676e-01 -5.95782459e-01
4.71113712e-01 -1.09690475e+00 -5.95945120e-01 2.47466154e-02
-2.18383119e-01 1.10291757e-01 6.82349443e-01 -8.62990379e-01
8.67165029e-01 -1.99018013e+00 -2.09883586e-01 4.87534463e-01
4.10229713e-01 5.38879186e-02 2.50469416e-01 -5.26585042e-01
-8.64609629e-02 2.95977980e-01 -3.16797435e-01 2.73373693e-01
-3.41200739e-01 5.38217500e-02 4.51574087e-01 8.24656367e-01
2.05524489e-01 1.28458226e+00 -8.87315035e-01 -8.50654125e-01
6.25241578e-01 3.58799309e-01 -1.82898447e-01 1.12743571e-01
2.71443278e-01 8.18973124e-01 -2.22241387e-01 9.71931875e-01
5.86785257e-01 -9.32528317e-01 5.15847921e-01 -4.43343759e-01
-2.20683292e-01 -4.90498953e-02 -1.04259968e+00 1.38758671e+00
-2.95269608e-01 1.68666512e-01 3.86200808e-02 -3.33271712e-01
2.81072378e-01 7.13901758e-01 9.43401396e-01 -3.69853646e-01
7.08968714e-02 3.70282590e-01 7.01902211e-01 -7.05970407e-01
-2.87721843e-01 -3.36741567e-01 2.12482139e-01 4.97847319e-01
-1.25911042e-01 -3.70693505e-01 3.53244334e-01 -5.41885048e-02
9.99081075e-01 -6.16746783e-01 7.39714146e-01 -6.40979826e-01
5.38365245e-01 4.19502616e-01 4.81428444e-01 7.62723267e-01
-5.54695487e-01 7.23373771e-01 2.55449831e-01 -6.33980870e-01
-8.09541464e-01 -1.12803662e+00 -4.89190996e-01 5.78954041e-01
1.11663312e-01 1.19698152e-01 -3.03567290e-01 -1.07156312e+00
-7.11584240e-02 2.26867720e-01 -5.33134460e-01 8.20468292e-02
-6.70086920e-01 -1.73067653e+00 2.15175256e-01 3.16622019e-01
4.10409540e-01 -6.69649065e-01 -1.02688718e+00 1.71924263e-01
-2.99291074e-01 -8.73432338e-01 -5.31848669e-01 3.98932964e-01
-8.52323830e-01 -1.61045456e+00 -1.05013180e+00 -1.07862091e+00
1.07889020e+00 2.48038799e-01 1.22235084e+00 4.23686892e-01
-9.76969182e-01 1.61457762e-01 -6.14590272e-02 -2.73869280e-02
-6.02153838e-01 -2.09196389e-01 -2.64981866e-01 -1.89439878e-01
3.95564437e-02 2.24917158e-01 -8.46857727e-01 3.73552918e-01
-6.58891857e-01 2.94711411e-01 1.26737392e+00 1.07563996e+00
7.80566931e-01 1.53696880e-01 8.95477831e-02 -6.85384631e-01
7.05719888e-02 -5.35382867e-01 -5.93134284e-01 5.71002483e-01
-5.61190307e-01 -3.00898969e-01 1.00644298e-01 -6.07403457e-01
-7.65889406e-01 5.11743486e-01 1.75867870e-01 -2.13056952e-01
-2.31639519e-01 5.05029917e-01 4.02112335e-01 -5.39128065e-01
6.38380349e-01 4.83906418e-01 -5.84905148e-02 -1.08597595e-02
-1.32175609e-01 1.90944567e-01 3.56045634e-01 -5.98813668e-02
6.01700723e-01 3.69690746e-01 2.42548436e-01 -6.55070126e-01
-2.86812782e-01 -8.44491720e-01 -8.48101795e-01 -1.72878981e-01
8.91728580e-01 -7.78429747e-01 -4.53996778e-01 1.88893467e-01
-5.84979475e-01 -2.80935526e-01 -3.52544159e-01 9.36251700e-01
4.91062179e-02 4.01755393e-01 -8.97248685e-01 -3.98875237e-01
-6.23804331e-01 -1.85242033e+00 6.90437734e-01 4.16802913e-01
5.48215061e-02 -9.94937181e-01 -4.04746830e-01 -1.68275923e-01
7.35198557e-01 2.38476247e-01 1.18483806e+00 -8.82855833e-01
-8.58823776e-01 -2.51809973e-02 -5.40618598e-01 -2.26916149e-01
3.01368117e-01 -1.36900187e-01 -4.71485347e-01 -3.89862001e-01
-8.07469934e-02 1.82751268e-01 7.83574700e-01 9.60913718e-01
1.05252850e+00 2.07037367e-02 -4.48046774e-01 7.17878222e-01
1.45858359e+00 3.92194629e-01 2.23023351e-02 5.93361035e-02
3.98463070e-01 1.54392391e-01 7.61899874e-02 3.03103209e-01
2.19879672e-01 3.11728150e-01 5.17548203e-01 -6.53388500e-01
-4.40081835e-01 2.37791777e-01 -1.98997840e-01 5.55162907e-01
1.28044531e-01 1.39170870e-01 -1.60097134e+00 7.09157050e-01
-1.26563108e+00 -5.89930177e-01 -4.93394017e-01 2.01361966e+00
6.46515906e-01 -2.07111016e-01 6.82021677e-02 -2.72981167e-01
9.03296530e-01 -4.55543488e-01 -2.84032971e-01 5.42002738e-01
-1.20162912e-01 1.63735032e-01 7.56136596e-01 4.16364223e-01
-1.42251551e+00 2.74311036e-01 6.42171764e+00 4.57697839e-01
-1.22724319e+00 3.33899915e-01 9.34786856e-01 1.88395873e-01
1.26265004e-01 -1.51252985e-01 -4.33396697e-01 3.09972227e-01
3.68185490e-01 -1.88411176e-01 6.21382752e-03 5.31296909e-01
8.20047930e-02 -5.85943341e-01 -9.55206275e-01 8.37376893e-01
7.36624524e-02 -1.39323771e+00 -7.23654255e-02 -7.17503428e-02
6.72360063e-01 2.60561794e-01 9.05971322e-03 5.02525233e-02
1.74416438e-01 -9.89998698e-01 3.74141455e-01 1.74399689e-01
1.12590623e+00 -2.01687023e-01 9.76839721e-01 3.42940390e-01
-1.38533568e+00 2.50480659e-02 1.85572967e-01 7.02754736e-01
1.14799127e-01 6.53189659e-01 -1.57581580e+00 3.90527397e-01
6.01924658e-01 2.97879487e-01 -8.27602863e-01 1.77664936e+00
-2.32291758e-01 5.41984141e-01 -4.69603151e-01 1.46625012e-01
1.54077467e-02 -4.23785932e-02 4.75549310e-01 1.69912624e+00
2.66577691e-01 4.84550238e-01 8.09445202e-01 5.97476363e-01
2.40312174e-01 1.70005307e-01 1.43459365e-01 3.31364065e-01
2.29737520e-01 1.77149367e+00 -1.56426501e+00 -7.74137914e-01
-4.30298388e-01 7.18274236e-01 -4.05620009e-01 3.18300612e-02
-9.00013328e-01 2.87779152e-01 -1.44578457e-01 -6.15604222e-03
-3.97614911e-02 7.28116557e-02 -4.49866086e-01 -1.20571911e+00
-4.35685486e-01 -6.94009840e-01 9.57471848e-01 -2.49068961e-01
-1.16140735e+00 5.93659580e-01 -1.78660393e-01 -1.05285239e+00
6.82785083e-03 -5.27819872e-01 -5.99855721e-01 1.05062830e+00
-1.71486962e+00 -1.25774634e+00 -7.81483710e-01 3.56387854e-01
6.13302231e-01 2.02729926e-01 8.44672561e-01 3.66029322e-01
-4.57728922e-01 2.02154234e-01 -3.20773691e-01 3.64578873e-01
5.08736432e-01 -1.44846094e+00 -2.49924630e-01 8.21209669e-01
-3.72042567e-01 2.52100319e-01 1.59949228e-01 -6.49514139e-01
-1.35811460e+00 -1.14988768e+00 5.83824337e-01 -1.72087610e-01
3.90574932e-01 2.74350166e-01 -8.50125849e-01 6.62737429e-01
-4.07939777e-02 7.69380033e-01 9.70376849e-01 -5.83099902e-01
1.27469689e-01 4.87374783e-01 -1.58041036e+00 2.70182610e-01
2.98610210e-01 4.17119190e-02 -4.19458091e-01 6.14824831e-01
-8.68291259e-02 -1.00209725e+00 -1.21817565e+00 7.66107976e-01
5.44494629e-01 -5.80227613e-01 1.10400665e+00 -2.65282005e-01
-8.03598985e-02 -4.78096038e-01 2.98908710e-01 -1.25598657e+00
-8.82414818e-01 3.69105896e-04 1.02772214e-01 2.44834900e-01
5.70515156e-01 -5.44591725e-01 7.05922246e-01 5.29669225e-01
-3.39518428e-01 -6.68862164e-01 -7.80210674e-01 -1.99737832e-01
-4.99125645e-02 2.91806698e-01 3.25780421e-01 1.03153515e+00
-8.17705095e-02 -2.58725643e-01 4.18455541e-01 6.15663171e-01
7.28479624e-01 4.08732086e-01 1.00793667e-01 -1.11160266e+00
-2.78084874e-01 -8.11873436e-01 -1.78419530e-01 -5.76381087e-01
-5.44988275e-01 -1.31332481e+00 -3.30729820e-02 -1.72869766e+00
1.05422568e+00 -8.69364917e-01 -3.54873180e-01 5.98267019e-01
-5.08098662e-01 4.10457700e-01 1.47170499e-01 5.51887929e-01
-3.67160410e-01 -2.12881535e-01 1.49600196e+00 -2.72012115e-01
-3.92873660e-02 1.95873067e-01 -1.79299459e-01 8.67099583e-01
6.55030966e-01 -4.94000345e-01 2.91588515e-01 -1.10072792e-01
-5.46220660e-01 3.89828086e-01 6.61976218e-01 -1.02163088e+00
4.49076921e-01 -3.16904813e-01 1.09454286e+00 -6.53606653e-01
-1.05846316e-01 -9.65129256e-01 4.72752154e-01 1.52523351e+00
8.14788342e-02 1.62925765e-01 2.84525990e-01 1.91462710e-01
-5.99786155e-02 -1.84177354e-01 1.41533291e+00 -7.62502015e-01
-6.68394506e-01 4.85864162e-01 -6.39390230e-01 -2.15371832e-01
1.44156027e+00 6.68097511e-02 -2.08373606e-01 2.54390389e-01
-9.89816427e-01 3.63311172e-01 2.13945776e-01 -7.83536881e-02
6.47681594e-01 -1.04074752e+00 -1.06497657e+00 5.03325999e-01
1.05330698e-01 2.38949686e-01 3.50663424e-01 1.73225248e+00
-8.65005434e-01 4.65699136e-01 -6.66448399e-02 -1.22901487e+00
-1.36946285e+00 2.73833543e-01 9.83295619e-01 -7.18629360e-01
-8.15799356e-01 7.01009214e-01 3.10416311e-01 -1.45381391e-01
-1.80584952e-01 -4.71377522e-01 1.30557179e-01 -1.94973171e-01
4.79556650e-01 1.32763341e-01 4.09244895e-01 -5.71264386e-01
-5.50431788e-01 3.61890376e-01 -1.28990680e-01 2.19219893e-01
9.12727177e-01 1.23652639e-02 -2.15833589e-01 -3.21654677e-02
6.00241423e-01 -9.07643363e-02 -7.85518289e-01 -2.20796645e-01
1.25724688e-01 -4.45339799e-01 2.05807388e-01 -1.35863888e+00
-1.11096895e+00 5.66973448e-01 8.72993290e-01 1.44767717e-01
1.08765900e+00 1.83544442e-01 2.33561054e-01 -4.05966073e-01
2.19650760e-01 -1.79481268e-01 -1.56057671e-01 3.09922457e-01
5.30398190e-01 -1.53691638e+00 2.41740182e-01 -7.44796395e-01
-6.98998094e-01 1.22688556e+00 5.14484346e-01 3.50516051e-01
5.63600898e-01 5.66407204e-01 2.03005642e-01 -3.20059359e-01
-6.38116837e-01 -2.59675920e-01 4.41136867e-01 4.06435788e-01
4.51330870e-01 5.76534271e-01 -1.57761678e-01 2.92301714e-01
4.18096095e-01 -3.15885544e-01 4.18125033e-01 8.55508089e-01
-6.79816902e-01 -5.64194620e-01 -7.07460046e-01 7.62760580e-01
-5.99699974e-01 -1.69487834e-01 1.17124885e-01 9.07319486e-01
1.31516531e-01 5.47294378e-01 -1.14284540e-02 2.03696132e-01
1.82150885e-01 1.34266034e-01 4.79740977e-01 -7.40636468e-01
-1.06687570e+00 4.74342465e-01 -3.72422040e-01 -1.27020985e-01
-1.51935384e-01 -7.87709296e-01 -1.46947312e+00 1.68583259e-01
-5.17500520e-01 2.45243445e-01 8.74895096e-01 6.38727248e-01
1.41722709e-01 7.57494628e-01 5.55682421e-01 -6.53108954e-01
-7.19478607e-01 -9.64911819e-01 -3.12795937e-01 3.69170874e-01
2.96712130e-01 -4.64312434e-01 -3.25171828e-01 2.19615608e-01]
|
[14.553059577941895, -2.6781444549560547]
|
89c07b29-c0f2-4c4e-8316-57e28e486c60
|
toponym-detection-in-the-bio-medical-domain-a
| null | null |
https://aclanthology.org/R19-1106
|
https://aclanthology.org/R19-1106.pdf
|
Toponym Detection in the Bio-Medical Domain: A Hybrid Approach with Deep Learning
|
This paper compares how different machine learning classifiers can be used together with simple string matching and named entity recognition to detect locations in texts. We compare five different state-of-the-art machine learning classifiers in order to predict whether a sentence contains a location or not. Following this classification task, we use a string matching algorithm with a gazetteer to identify the exact index of a toponym within the sentence. We evaluate different approaches in terms of machine learning classifiers, text pre-processing and location extraction on the SemEval-2019 Task 12 dataset, compiled for toponym resolution in the bio-medical domain. Finally, we compare the results with our system that was previously submitted to the SemEval-2019 task evaluation.
|
['Tharindu Ranasinghe', 'Alistair Plum', 'Constantin Orasan']
|
2019-09-01
| null | null | null |
ranlp-2019-9
|
['toponym-resolution']
|
['natural-language-processing']
|
[ 3.22241604e-01 2.83405147e-02 -2.26494476e-01 -3.69388372e-01
-8.33750486e-01 -6.30142808e-01 6.00951135e-01 1.36796141e+00
-1.18489277e+00 8.87824416e-01 3.12525064e-01 -3.51833493e-01
-1.38511389e-01 -7.24791050e-01 -4.70599174e-01 -3.42363924e-01
1.81571245e-01 8.97615969e-01 2.76081979e-01 -2.49626517e-01
5.08725762e-01 4.53796744e-01 -1.12692344e+00 1.21724641e+00
5.69606543e-01 6.70394599e-01 1.32475048e-02 7.99412012e-01
-3.72486204e-01 4.92392212e-01 -1.10466385e+00 -5.03372312e-01
-2.12411746e-01 -3.75604719e-01 -1.12834680e+00 -8.02017093e-01
5.09511352e-01 5.63566387e-01 -4.78480533e-02 7.53402531e-01
7.06870139e-01 -2.70492315e-01 8.63248169e-01 -6.36637688e-01
-1.04016088e-01 9.40558672e-01 -3.09124500e-01 6.63922250e-01
1.03780043e+00 -4.48415220e-01 1.18192029e+00 -7.79175520e-01
1.21492565e+00 8.20592642e-01 9.67652559e-01 2.51852483e-01
-9.75560486e-01 -4.40606028e-01 -6.03536069e-01 4.03369665e-01
-1.70682132e+00 -1.49803117e-01 -4.03895117e-02 -5.61984062e-01
1.54242492e+00 6.26434624e-01 2.55580097e-01 7.76210904e-01
4.64655966e-01 5.05105317e-01 9.13588643e-01 -1.02947700e+00
3.93553883e-01 4.60312665e-02 4.98645782e-01 5.47085047e-01
2.41283268e-01 -2.63499558e-01 -7.33948946e-01 -6.70726359e-01
-2.87330419e-01 -1.16603903e-01 -6.49428293e-02 1.51053205e-01
-1.28874266e+00 5.29269397e-01 2.59108335e-01 7.86894202e-01
-4.10309613e-01 -5.42401552e-01 1.01610529e+00 3.60784382e-01
4.04262602e-01 9.43182290e-01 -8.79878521e-01 -2.80236751e-02
-1.31300306e+00 1.07429452e-01 1.10502172e+00 8.09038877e-01
3.22029322e-01 -1.20247591e+00 -4.64934707e-01 8.34863305e-01
9.12435260e-03 5.05198725e-02 9.68785226e-01 -1.05121523e-01
8.05069685e-01 8.00628662e-01 6.42402619e-02 -9.26551819e-01
-1.12411988e+00 -2.48328194e-01 -4.64812368e-01 -2.52841353e-01
5.56999028e-01 -1.72356293e-01 -5.65228164e-01 1.20480692e+00
2.05742404e-01 2.46912479e-01 2.73409754e-01 2.85325855e-01
1.30626285e+00 4.04608548e-01 2.48210683e-01 -3.76908220e-02
1.76173723e+00 -4.06754553e-01 -6.83610022e-01 1.95946783e-01
1.52535343e+00 -8.78910422e-01 2.28987813e-01 2.10802466e-01
-6.11620665e-01 -3.27336103e-01 -1.10699737e+00 3.28002423e-02
-1.22841918e+00 3.38487029e-01 -1.55020282e-02 7.94990420e-01
-5.55557966e-01 7.69916952e-01 -4.06102091e-01 -8.39259028e-01
1.22747727e-01 2.06268027e-01 -5.84275484e-01 1.89517930e-01
-1.55614460e+00 1.43861389e+00 8.61345410e-01 -4.23706293e-01
9.79164243e-02 -9.64119077e-01 -8.23930264e-01 -1.28363550e-01
-1.76449999e-01 -8.00718665e-01 1.09418702e+00 -3.66147906e-01
-6.74262822e-01 1.67686605e+00 -2.49755934e-01 -8.60929072e-01
4.08985913e-01 1.28038406e-01 -8.86514068e-01 1.45877451e-01
4.73720402e-01 2.36682758e-01 1.84414372e-01 -1.73412144e-01
-1.10181928e+00 -4.63625073e-01 -5.10451078e-01 1.44763123e-02
-1.03184901e-01 5.69583237e-01 -1.85741991e-01 -6.95149481e-01
-1.42506659e-01 -8.15617681e-01 -8.00748840e-02 -3.98865670e-01
-6.66107893e-01 -5.94302297e-01 2.98391789e-01 -6.81766450e-01
1.70868862e+00 -2.03619766e+00 -4.80663478e-02 2.65594274e-01
5.95909804e-02 1.97485298e-01 5.96493930e-02 7.61148453e-01
-5.46524048e-01 1.40738577e-01 -1.05997235e-01 -2.10730135e-01
-1.75193578e-01 -1.32171765e-01 -2.08599329e-01 6.00232124e-01
-1.50370434e-01 6.45520508e-01 -8.87145102e-01 -8.23429883e-01
2.89403144e-02 2.53162161e-02 -6.91768005e-02 -2.34141946e-01
1.57844066e-03 -1.06789554e-02 -3.10408890e-01 4.71153259e-01
4.09024954e-01 -5.00668809e-02 4.38962489e-01 -1.52725220e-01
-4.48637002e-04 5.51273108e-01 -9.45105970e-01 1.83079493e+00
-4.68953907e-01 7.33028412e-01 -5.09765446e-01 -1.02109551e+00
8.61405730e-01 4.69714552e-01 5.21063209e-01 -5.99829733e-01
-6.09001666e-02 5.35891831e-01 -2.01120958e-01 -8.43196869e-01
4.48311061e-01 2.35372290e-01 -4.98542190e-01 1.37350664e-01
1.40601933e-01 4.73712862e-01 6.13265812e-01 5.11801653e-02
1.77747250e+00 -1.44935742e-01 1.01973581e+00 -2.86236912e-01
8.10911477e-01 2.39590690e-01 4.01047975e-01 8.49024534e-01
6.78205267e-02 6.36086643e-01 4.23498452e-01 -7.68439710e-01
-9.20274377e-01 -5.81510603e-01 -6.40529037e-01 1.17505074e+00
-3.21056157e-01 -8.37830782e-01 -8.13740432e-01 -9.04507220e-01
2.00303704e-01 7.82950461e-01 -7.58225679e-01 3.43400463e-02
-8.61984789e-01 -9.58364248e-01 1.14617312e+00 2.33708858e-01
-2.83707101e-02 -1.13059306e+00 -9.63130891e-01 2.11504668e-01
-3.22103322e-01 -1.20171773e+00 -1.87108055e-01 3.38038176e-01
-4.13901508e-01 -1.42696202e+00 -7.88881302e-01 -8.66106689e-01
4.98460591e-01 -5.65615058e-01 1.39785314e+00 8.02553594e-02
-9.07617569e-01 -7.73852542e-02 -5.08676350e-01 -6.67018950e-01
-5.48201323e-01 6.12873793e-01 -1.64005876e-01 -2.69106299e-01
9.92694914e-01 -7.30078220e-02 -3.75170678e-01 1.15213394e-01
-5.90304792e-01 -4.48838204e-01 5.24842203e-01 6.69906378e-01
4.34262484e-01 -2.46221364e-01 3.04259986e-01 -1.34714687e+00
3.83385032e-01 -6.95013821e-01 -3.18865150e-01 4.66945052e-01
-4.14240092e-01 -3.89335956e-03 5.73018789e-01 9.72801261e-03
-5.40491045e-01 5.94473898e-01 -5.73645532e-01 3.19920599e-01
-4.71711844e-01 8.24722171e-01 1.26888663e-01 1.85434431e-01
1.16758931e+00 1.97319850e-01 -5.68593860e-01 -7.10112631e-01
3.47253948e-01 1.09837818e+00 5.68560123e-01 1.70225371e-02
2.49535680e-01 1.78357810e-01 1.60683617e-01 -9.08357978e-01
-7.86509931e-01 -1.01160872e+00 -1.15599513e+00 4.58799124e-01
1.00205135e+00 -7.09531963e-01 -4.59992886e-01 2.97066450e-01
-1.38933825e+00 3.86255443e-01 -3.21504846e-02 3.43531102e-01
-4.41899508e-01 3.54561657e-01 -3.92667323e-01 -2.02641249e-01
-6.22897744e-01 -5.16977012e-01 1.04168868e+00 7.93085769e-02
-8.94148946e-01 -7.66554952e-01 6.46327436e-01 2.30621442e-01
-8.17076862e-02 3.53917331e-01 1.04356074e+00 -1.63945460e+00
5.34176111e-01 -7.59048998e-01 1.86596215e-01 -7.15779722e-01
-5.47066480e-02 -2.28370458e-01 -5.78815818e-01 4.74421680e-02
-5.01069665e-01 2.17578068e-01 1.06195438e+00 4.91899997e-02
1.03744721e+00 -2.88276225e-01 -1.01244879e+00 4.70167339e-01
1.22881961e+00 8.33299905e-02 4.99771118e-01 9.00610745e-01
1.57846913e-01 6.52341843e-01 6.99158132e-01 4.23304945e-01
-3.81023437e-03 1.12580574e+00 -2.74201836e-02 -3.62065644e-03
4.50285338e-02 -3.04959416e-02 -2.99298346e-01 2.62931664e-03
4.90640819e-01 -3.48564446e-01 -1.29452741e+00 5.66769838e-01
-1.78179204e+00 -9.98678565e-01 -2.27893263e-01 1.94687259e+00
1.07047141e+00 1.06607474e-01 2.82600343e-01 2.61413097e-01
9.49688613e-01 -2.69740611e-01 9.58407447e-02 -5.45294881e-01
-1.53786093e-01 3.39069992e-01 7.87814260e-01 1.46342600e-02
-1.46416438e+00 6.14909589e-01 6.79662275e+00 8.95568132e-01
-9.73326325e-01 1.72500625e-01 4.05705214e-01 -1.24300420e-02
4.80077624e-01 -4.55761254e-01 -1.21958888e+00 8.31099272e-01
1.42527735e+00 9.20619220e-02 -3.59083772e-01 6.33306801e-01
-3.12398076e-02 -1.34368643e-01 -1.30676377e+00 1.16116762e+00
3.55557024e-01 -1.80597246e+00 -2.47598007e-01 -3.42754513e-01
1.90183818e-01 3.18723619e-01 -4.37903851e-01 -8.71019810e-03
-1.44007698e-01 -9.41338301e-01 6.86069489e-01 7.04907358e-01
5.91337085e-01 -5.42065620e-01 1.05831134e+00 3.51276487e-01
-9.98553991e-01 -1.34689078e-01 -2.32732981e-01 3.23267549e-01
8.51067714e-04 7.45840788e-01 -1.15798593e+00 6.70262456e-01
8.54285300e-01 6.86892748e-01 -9.85212386e-01 1.72384906e+00
-3.17665160e-01 4.05350506e-01 -4.89445865e-01 -4.42766011e-01
-1.10571697e-01 5.26539028e-01 6.59638643e-01 1.95597863e+00
1.96581692e-01 -1.12536572e-01 -1.49085000e-01 3.82564723e-01
-1.01640180e-01 8.26309264e-01 -4.38415974e-01 4.39743042e-01
7.44875908e-01 1.09526527e+00 -8.63700867e-01 -5.24438918e-01
-2.61913449e-01 9.67215300e-01 2.84048021e-01 -1.72267348e-01
-6.06485963e-01 -1.09123552e+00 1.62020087e-01 -4.17306907e-02
3.91089529e-01 3.44172180e-01 -4.38736022e-01 -9.66144204e-01
-1.30445093e-01 -6.60634220e-01 1.21246767e+00 -8.01305413e-01
-1.22044575e+00 5.98334253e-01 -1.85404941e-02 -1.08872616e+00
-2.91989535e-01 -8.63827288e-01 -6.18692756e-01 8.19730163e-01
-9.82952476e-01 -1.01882482e+00 2.77786016e-01 2.09328324e-01
2.88714617e-01 -4.02683854e-01 1.15694273e+00 4.55753416e-01
-5.91723144e-01 8.63851726e-01 4.61959332e-01 6.66106164e-01
1.16105032e+00 -1.40285599e+00 5.41146100e-01 3.82128745e-01
4.77912039e-01 6.73548281e-01 7.40475476e-01 -7.47812271e-01
-7.67680764e-01 -1.09414375e+00 1.84679282e+00 -9.08379495e-01
5.59066832e-01 -4.30518091e-01 -6.11738324e-01 3.34166259e-01
3.10570836e-01 -1.68811902e-01 1.00138104e+00 2.45149180e-01
-4.03633833e-01 2.16934308e-01 -1.29398310e+00 -3.52862850e-02
5.80324233e-01 -4.76401716e-01 -1.07993770e+00 7.39028394e-01
3.14317167e-01 -5.35100222e-01 -1.03649235e+00 2.63778120e-01
4.76772755e-01 -2.54157305e-01 1.10327446e+00 -9.15710628e-01
2.94185519e-01 -2.62529701e-01 -4.84529883e-03 -1.18624651e+00
-2.72929706e-02 -4.46294159e-01 1.98604107e-01 1.26970029e+00
1.23236167e+00 -5.37578166e-01 6.56395495e-01 1.37232795e-01
2.57485900e-02 -4.84912604e-01 -1.47746313e+00 -6.40800595e-01
2.87318826e-01 -2.49694120e-02 5.65510571e-01 1.10175037e+00
7.06752598e-01 4.67858523e-01 9.31509957e-02 -1.02983415e-01
7.18523785e-02 -8.03586915e-02 2.35973507e-01 -1.25967562e+00
-5.82540780e-02 -5.08915842e-01 -9.41688240e-01 -1.69474944e-01
1.77412927e-01 -1.28918135e+00 -1.22563109e-01 -1.40980506e+00
3.02615464e-01 -1.57609865e-01 -5.04447401e-01 7.59971619e-01
2.56786868e-02 1.45246401e-01 -8.95574864e-04 2.14986116e-01
-7.18410552e-01 -4.96872663e-01 -5.46169728e-02 -2.61527687e-01
-1.26488298e-01 1.03644803e-01 -1.84232429e-01 5.76285720e-01
6.10455513e-01 -9.97570395e-01 4.78949785e-01 -4.11407351e-02
7.01833367e-01 -1.17467932e-01 -2.06382703e-02 -8.95651162e-01
4.47271734e-01 3.10934097e-01 6.64538682e-01 -1.07806194e+00
-2.47910187e-01 -6.48727775e-01 3.51793505e-02 8.87021482e-01
-7.39130437e-01 2.93115884e-01 2.04955667e-01 4.52700406e-01
-6.43505752e-02 -9.15291011e-01 6.97287679e-01 -2.75352746e-01
-7.21108139e-01 -3.81753832e-01 -7.81369448e-01 2.49477923e-02
1.03462887e+00 9.82795060e-02 -5.40926874e-01 3.86242479e-01
-1.07414758e+00 -2.13604234e-02 3.42619985e-01 3.28829080e-01
2.17722669e-01 -9.46562409e-01 -8.66053283e-01 -2.68366396e-01
6.14618659e-01 -8.51363063e-01 3.16253714e-02 8.20473015e-01
-9.10145044e-01 9.31101263e-01 -1.80596799e-01 -4.78235930e-01
-1.78142023e+00 8.69929016e-01 4.60555613e-01 -5.41244268e-01
-7.02001929e-01 7.00059593e-01 -5.16019285e-01 -6.36608660e-01
1.87471613e-01 -2.50234008e-01 -7.59746909e-01 6.06753290e-01
8.68245065e-01 3.29231769e-01 8.71655285e-01 -6.59124374e-01
-9.83799517e-01 5.50766587e-01 -7.70454928e-02 4.33660075e-02
1.04832983e+00 1.88368127e-01 -3.16896170e-01 1.26624063e-01
1.45663214e+00 -1.00282803e-01 4.48296756e-01 -3.92259389e-01
8.12464058e-01 1.64651647e-02 -2.13389173e-01 -1.18542457e+00
-3.58391404e-01 5.02803385e-01 8.37465107e-01 2.03931481e-01
8.59110594e-01 1.04160234e-01 5.33817470e-01 7.49787569e-01
5.74330837e-02 -1.32281709e+00 -5.96420467e-01 6.46595776e-01
5.52887142e-01 -1.05051303e+00 1.63210824e-01 -3.94909024e-01
-2.75484055e-01 1.32417369e+00 -7.86830764e-03 1.84758291e-01
7.66430140e-01 2.24720359e-01 -5.89337423e-02 -2.67270565e-01
-6.21110559e-01 -1.32526755e-01 5.13826311e-01 4.52935696e-01
7.59695530e-01 -4.30216081e-02 -9.16524470e-01 6.02690279e-01
-3.73356193e-01 -8.80602151e-02 3.04111063e-01 1.13667750e+00
-4.16129619e-01 -1.14658248e+00 -3.77682477e-01 6.16497517e-01
-1.05467415e+00 -4.01102245e-01 -7.90706873e-01 5.25956631e-01
3.62186164e-01 8.06591809e-01 5.86700514e-02 -8.84702951e-02
5.31702876e-01 4.05683160e-01 2.46330053e-01 -9.26363349e-01
-1.21001124e+00 -3.02314579e-01 5.99993765e-01 -3.55584830e-01
-2.65328377e-01 -1.03844976e+00 -1.33877051e+00 2.25726798e-01
-2.69806355e-01 3.95641387e-01 6.84894979e-01 1.15161252e+00
4.43170518e-01 3.57991576e-01 2.02129945e-01 -1.11761689e-01
-2.64085621e-01 -9.73532021e-01 -3.46893877e-01 5.61785579e-01
-3.71896699e-02 -2.27729753e-01 2.08050944e-02 -1.49515306e-03]
|
[8.642224311828613, 8.908021926879883]
|
4f37c18a-c8df-47a3-abcf-22f3694c3047
|
a-unified-object-counting-network-with-object
|
2212.14193
| null |
https://arxiv.org/abs/2212.14193v3
|
https://arxiv.org/pdf/2212.14193v3.pdf
|
A Unified Object Counting Network with Object Occupation Prior
|
The counting task, which plays a fundamental role in numerous applications (e.g., crowd counting, traffic statistics), aims to predict the number of objects with various densities. Existing object counting tasks are designed for a single object class. However, it is inevitable to encounter newly coming data with new classes in our real world. We name this scenario as \textit{evolving object counting}. In this paper, we build the first evolving object counting dataset and propose a unified object counting network as the first attempt to address this task. The proposed model consists of two key components: a class-agnostic mask module and a class-incremental module. The class-agnostic mask module learns generic object occupation prior via predicting a class-agnostic binary mask (e.g., 1 denotes there exists an object at the considering position in an image and 0 otherwise). The class-incremental module is used to handle new coming classes and provides discriminative class guidance for density map prediction. The combined outputs of class-agnostic mask module and image feature extractor are used to predict the final density map. When new classes come, we first add new neural nodes into the last regression and classification layers of class-incremental module. Then, instead of retraining the model from scratch, we utilize knowledge distillation to help the model remember what have already learned about previous object classes. We also employ a support sample bank to store a small number of typical training samples of each class, which are used to prevent the model from forgetting key information of old data. With this design, our model can efficiently and effectively adapt to new coming classes while keeping good performance on already seen data without large-scale retraining. Extensive experiments on the collected dataset demonstrate the favorable performance.
|
['Qingshan Liu', 'Yuankai Qi', 'Fengna Cheng', 'Qing Wang', 'Shengqin Jiang']
|
2022-12-29
| null | null | null | null |
['object-counting']
|
['computer-vision']
|
[ 1.52113780e-01 -3.50094259e-01 -2.72903562e-01 -6.25817418e-01
-1.20946437e-01 -1.75464779e-01 4.41318631e-01 1.48796946e-01
-9.02152598e-01 9.66031373e-01 -1.90390840e-01 -1.25505686e-01
2.24506557e-01 -1.28001666e+00 -6.24521255e-01 -6.75293744e-01
5.24924472e-02 8.71667087e-01 8.67910326e-01 1.83016181e-01
2.41653144e-01 5.88622868e-01 -1.75611413e+00 1.82335436e-01
8.61291587e-01 1.02549469e+00 6.42787695e-01 7.50333607e-01
-4.55998451e-01 1.00975776e+00 -7.50382304e-01 -3.82188857e-01
4.70285043e-02 -7.79053047e-02 -4.96056110e-01 9.14720222e-02
3.67098391e-01 -8.00624549e-01 -6.72261298e-01 1.06863832e+00
3.24715614e-01 3.05596739e-01 7.27703333e-01 -1.03486156e+00
-5.52798152e-01 3.98841351e-01 -7.74530709e-01 1.06931353e+00
-2.39581227e-01 3.46265048e-01 3.74309659e-01 -9.87785757e-01
1.03189595e-01 1.07185197e+00 3.93446922e-01 7.02158928e-01
-7.87539899e-01 -8.99468064e-01 5.88897645e-01 3.79057884e-01
-1.46117175e+00 -3.94375503e-01 6.49671972e-01 -3.75724524e-01
7.49874175e-01 2.71539148e-02 8.12393606e-01 6.28565133e-01
-8.00078213e-02 1.02667010e+00 7.20002770e-01 1.54221924e-02
3.32809240e-01 3.88620019e-01 3.95823151e-01 7.07796693e-01
5.21595120e-01 -2.63549715e-01 -2.33341426e-01 8.41636360e-02
7.40329862e-01 5.65757692e-01 1.07568830e-01 -1.25712365e-01
-7.95946598e-01 7.01445103e-01 8.25067043e-01 3.34079444e-01
-2.99830496e-01 1.96954742e-01 2.14104965e-01 -1.81463689e-01
5.35144091e-01 -2.66882896e-01 -3.33475590e-01 4.71701249e-02
-9.75472510e-01 1.77986920e-01 5.17470777e-01 8.26285005e-01
1.13583028e+00 4.89631444e-02 -4.73718286e-01 9.95309532e-01
8.73663723e-02 6.78965867e-01 5.47280669e-01 -4.85926330e-01
7.51017451e-01 9.03927684e-01 8.44360292e-02 -8.32542658e-01
-3.88567209e-01 -5.77523291e-01 -8.39851201e-01 -8.48180354e-02
3.91432852e-01 5.24009904e-03 -1.25482190e+00 1.60049546e+00
4.89956766e-01 5.79113245e-01 -2.30078727e-01 4.75067228e-01
6.98499680e-01 8.84893537e-01 2.28251293e-01 -2.35675573e-01
1.32200694e+00 -8.86896610e-01 -2.43156642e-01 -6.20485902e-01
2.52588719e-01 -1.75217167e-01 8.82516384e-01 -2.36269012e-02
-8.38210881e-01 -7.76545644e-01 -1.08835435e+00 3.08847055e-02
-6.69234514e-01 1.23273633e-01 6.89083040e-01 5.14191568e-01
-6.22907698e-01 3.05533499e-01 -8.09916794e-01 -1.29936472e-01
1.11878824e+00 3.94965082e-01 1.07842229e-01 -2.80060977e-01
-7.87081957e-01 7.34162569e-01 7.13954091e-01 6.04361184e-02
-8.08349788e-01 -4.89534110e-01 -6.99910045e-01 1.64726391e-01
5.01869261e-01 -4.59410936e-01 1.26416159e+00 -6.40606284e-01
-1.03540194e+00 6.96177423e-01 -5.14911294e-01 -4.15112585e-01
4.74627972e-01 -7.36711128e-03 -4.44982648e-01 -4.34633568e-02
4.10587102e-01 7.60661244e-01 9.13979948e-01 -1.34370983e+00
-1.33665097e+00 -3.90866876e-01 -1.14591941e-01 1.06291093e-01
-5.96996963e-01 -4.15636808e-01 -6.44595265e-01 -4.89603460e-01
1.29698208e-02 -4.76671785e-01 -2.90674176e-02 1.43754492e-02
-7.51337633e-02 -3.26199383e-01 1.07381761e+00 -3.42869401e-01
1.35760152e+00 -1.87846363e+00 -3.13213825e-01 8.66563320e-02
5.01641870e-01 5.30981958e-01 1.16438814e-03 -4.19034809e-01
3.03905964e-01 -2.91551560e-01 -2.57181525e-01 -4.06433493e-01
-4.14513499e-01 4.79570210e-01 -2.58156329e-01 1.90100804e-01
5.28847933e-01 1.01590550e+00 -1.14803648e+00 -7.39363790e-01
2.71023661e-01 2.17102572e-01 -5.25912344e-01 2.16842785e-01
-4.02638078e-01 4.89716530e-01 -4.80025351e-01 7.85875380e-01
9.07732010e-01 -3.76523465e-01 -1.73630774e-01 -3.89518254e-02
-6.11293018e-02 -1.72126532e-01 -1.21202195e+00 1.08814144e+00
-4.08919632e-01 1.87495649e-01 -4.26108241e-01 -9.85567987e-01
8.90394568e-01 -2.90576279e-01 2.23058686e-01 -7.21325994e-01
3.03566724e-01 1.63966000e-01 1.29436851e-01 -4.79119837e-01
6.28658891e-01 -8.67089927e-02 -4.05453891e-02 4.35097545e-01
1.61121801e-01 2.17388585e-01 5.30685902e-01 1.22615345e-01
1.02709973e+00 -3.06869507e-01 3.62306416e-01 1.75743848e-01
7.22470760e-01 -1.33692697e-01 6.08513653e-01 9.08049703e-01
-4.44403738e-01 4.14838731e-01 1.57410011e-01 -7.45380878e-01
-9.62946713e-01 -1.20394266e+00 -2.64737874e-01 1.26018178e+00
3.24283242e-01 1.94206491e-01 -5.20908296e-01 -9.40784097e-01
1.21385455e-01 5.66685081e-01 -7.78879404e-01 -3.77715558e-01
-9.76150155e-01 -1.06063449e+00 1.72351971e-01 8.82636964e-01
1.03113258e+00 -1.09029174e+00 -5.77230990e-01 3.35563749e-01
-6.44257874e-04 -9.87003803e-01 -5.85445046e-01 1.48706794e-01
-9.32634056e-01 -1.05908465e+00 -8.05341065e-01 -8.08369339e-01
8.38740826e-01 5.02493143e-01 9.55866694e-01 4.83025104e-01
-3.87956768e-01 1.99072525e-01 -3.86359282e-02 -5.94529450e-01
-1.87862605e-01 3.04128587e-01 5.14547490e-02 2.70320028e-01
8.07623506e-01 -6.58380866e-01 -7.32469380e-01 2.15359718e-01
-8.18135381e-01 -1.49047151e-01 5.94527960e-01 7.57248938e-01
5.79779029e-01 3.41472119e-01 8.13274741e-01 -8.74940395e-01
3.22869450e-01 -7.49572039e-01 -5.89124024e-01 2.32143879e-01
-2.80533284e-01 -3.24304187e-04 6.75487816e-01 -8.54545832e-01
-1.09132147e+00 -6.23491369e-02 -1.26355002e-02 -3.25303435e-01
1.46505293e-02 7.88489804e-02 -2.39645839e-01 2.53839463e-01
4.26372916e-01 4.65916634e-01 -3.57671767e-01 -4.00897264e-01
3.49585302e-02 6.99039578e-01 8.54983151e-01 -5.10722876e-01
9.19763744e-01 5.81423402e-01 -2.51895428e-01 -5.82211554e-01
-1.14154720e+00 -5.22263288e-01 -8.11282098e-01 -4.01541799e-01
7.83832312e-01 -8.74647498e-01 -7.06339359e-01 7.59148777e-01
-1.08659971e+00 -2.53362000e-01 -5.16327441e-01 2.64543593e-01
-1.36605620e-01 4.77609374e-02 -3.57526898e-01 -1.02474284e+00
-2.35176578e-01 -7.50323713e-01 9.37629461e-01 8.93756151e-01
3.78018230e-01 -7.84002900e-01 2.09684819e-02 2.18753085e-01
4.24200565e-01 -2.47905597e-01 8.72802377e-01 -7.43081927e-01
-7.37490892e-01 -4.75789011e-01 -4.92992789e-01 4.51053232e-01
1.58389181e-01 -2.66036242e-01 -1.13452685e+00 -2.08282217e-01
-5.93133122e-02 -2.46483192e-01 1.28353071e+00 1.88322499e-01
1.66331363e+00 -3.53340060e-01 -5.97827852e-01 4.34851050e-01
1.21963072e+00 3.06693107e-01 5.17647743e-01 1.46140769e-01
7.20049500e-01 1.51089996e-01 4.40190643e-01 4.82102215e-01
6.46788716e-01 2.93356210e-01 3.66475910e-01 1.49641231e-01
-1.97561353e-01 -5.56619644e-01 -1.41457599e-02 6.29253149e-01
-1.06391646e-01 -1.94504380e-01 -9.56409574e-01 5.01473486e-01
-1.82798302e+00 -1.14892220e+00 4.24728811e-01 2.28647733e+00
7.32640743e-01 5.11871159e-01 3.66774410e-01 6.47438541e-02
1.01445389e+00 7.09966719e-02 -1.02363205e+00 2.57899135e-01
4.23843041e-02 2.77745932e-01 4.44441199e-01 3.90876770e-01
-1.13287210e+00 9.76584613e-01 5.04887295e+00 9.67227221e-01
-9.90323365e-01 2.13323772e-01 9.09462750e-01 -2.10429266e-01
2.24240899e-01 -1.78471625e-01 -1.19734347e+00 8.13079476e-01
5.17712951e-01 -4.36592624e-02 2.85704076e-01 9.49981153e-01
-2.67556101e-01 -5.31478584e-01 -8.98436129e-01 9.78437245e-01
4.65505458e-02 -1.26437449e+00 1.66532427e-01 -1.37329355e-01
5.40135920e-01 -4.36423384e-02 1.21722490e-01 8.43220413e-01
2.29847431e-01 -7.90097535e-01 7.41360247e-01 8.37806642e-01
6.59298658e-01 -7.71163225e-01 7.96571732e-01 8.05786729e-01
-1.45350754e+00 -5.14913738e-01 -7.17321932e-01 -1.98058099e-01
1.47602558e-01 6.09242797e-01 -7.27497399e-01 7.83859566e-02
6.97306633e-01 7.03202903e-01 -8.36163580e-01 1.35265219e+00
4.69874591e-02 4.54948336e-01 -4.41128165e-01 -7.04534575e-02
7.87818804e-02 2.14017034e-02 2.53986835e-01 1.01688671e+00
2.15096310e-01 2.68411756e-01 3.64265561e-01 6.59549415e-01
-2.97625780e-01 -2.43615463e-01 -4.13961947e-01 3.74857455e-01
7.07574606e-01 1.23649955e+00 -1.03598058e+00 -8.09608877e-01
-4.08541232e-01 8.63398194e-01 8.02849948e-01 2.25177571e-01
-9.55309391e-01 -1.22611322e-01 2.28460327e-01 3.05511564e-01
5.62768340e-01 2.37288494e-02 -1.59777850e-01 -1.24114406e+00
4.91863899e-02 -2.19634026e-01 5.67518771e-01 -4.88258362e-01
-1.44844484e+00 4.59795952e-01 2.08512589e-01 -1.04601395e+00
-1.08079519e-02 -5.79791129e-01 -9.75886166e-01 6.17517173e-01
-1.63927674e+00 -9.95044470e-01 -5.87279916e-01 5.71547151e-01
6.22695029e-01 -2.90992290e-01 2.90676385e-01 5.45722187e-01
-8.50666881e-01 6.11239672e-01 -2.17661168e-02 4.61128920e-01
3.15154910e-01 -1.01031148e+00 2.29416832e-01 7.61989295e-01
-9.32911783e-03 3.89296800e-01 1.65520176e-01 -8.50326359e-01
-7.31631994e-01 -1.53718364e+00 5.74449539e-01 -5.88203728e-01
4.63937521e-01 -4.88848597e-01 -1.11858189e+00 5.85388243e-01
-6.07773840e-01 5.01894414e-01 2.37620771e-01 -2.53835112e-01
-2.56596357e-01 -4.29176152e-01 -1.34082925e+00 3.64452153e-01
1.21485007e+00 -2.03389943e-01 -5.90632498e-01 2.59132206e-01
5.06261349e-01 -2.77375996e-01 -2.90135533e-01 4.07557189e-01
3.69890451e-01 -7.88330317e-01 9.38145995e-01 -5.22362649e-01
1.27515405e-01 -4.52089846e-01 -7.75485858e-02 -8.45590472e-01
-4.53644216e-01 1.15998439e-01 -7.43272245e-01 1.27378356e+00
1.68123633e-01 -7.09432423e-01 1.09268630e+00 4.44595248e-01
3.63237001e-02 -7.90921688e-01 -1.21865726e+00 -7.69404888e-01
5.64909130e-02 -1.88922107e-01 9.12877262e-01 5.80558002e-01
-5.54446936e-01 5.53152263e-01 -2.23287389e-01 1.68130502e-01
5.76399326e-01 7.05818757e-02 9.41139817e-01 -1.26037347e+00
-1.83948368e-01 -3.94572020e-01 -6.48217261e-01 -1.40360665e+00
-6.19970262e-03 -8.98524940e-01 1.01787567e-01 -1.35983908e+00
8.55357468e-01 -8.56293082e-01 -3.44418466e-01 5.18446982e-01
-7.60979414e-01 2.67989546e-01 2.21414506e-01 3.29820544e-01
-8.39783251e-01 6.73019052e-01 1.19097161e+00 -4.06649828e-01
-2.76673079e-01 4.51899529e-01 -5.80643475e-01 8.53511691e-01
8.93125534e-01 -6.08300745e-01 -4.95411903e-01 -4.78583425e-01
-1.84362426e-01 -3.41055363e-01 5.86243391e-01 -1.41974640e+00
4.18403804e-01 -2.49441773e-01 9.45937693e-01 -8.82333279e-01
3.55958700e-01 -7.59510279e-01 -2.96247721e-01 5.88474333e-01
1.88555166e-01 -1.73817083e-01 2.14868203e-01 8.02455366e-01
1.25366613e-01 -5.04623830e-01 9.34748590e-01 -2.82297164e-01
-9.18873847e-01 8.76112342e-01 -1.43225743e-02 1.95947945e-01
1.11002803e+00 -5.21304846e-01 -5.28773546e-01 -2.52670073e-03
-5.92784643e-01 4.90674317e-01 2.46957585e-01 3.82719547e-01
6.14552975e-01 -1.28538752e+00 -5.44177473e-01 2.81370342e-01
1.09698348e-01 4.56124544e-01 4.18120235e-01 3.30830574e-01
-1.81902260e-01 1.38548791e-01 -6.99574873e-02 -8.53155077e-01
-8.23462009e-01 8.31274807e-01 4.60817695e-01 -4.38139111e-01
-4.99902159e-01 8.96336615e-01 3.26523930e-01 -3.40714991e-01
2.24249035e-01 -2.41372168e-01 -3.66700500e-01 4.81185801e-02
9.42328274e-01 3.72083783e-01 -1.23136006e-01 -5.83247900e-01
-3.38484049e-01 3.72360528e-01 -5.96424818e-01 2.04929113e-01
1.47067606e+00 -1.61502464e-03 1.00695468e-01 7.65514255e-01
9.71288979e-01 -3.44251215e-01 -1.57899988e+00 -6.25516117e-01
-1.68205708e-01 -6.52186990e-01 -2.45058164e-01 -5.99906385e-01
-1.19898164e+00 9.18320060e-01 7.81967521e-01 -1.17598906e-01
9.53489602e-01 1.89910263e-01 7.90163696e-01 5.96267879e-01
5.05327642e-01 -1.13103068e+00 3.81073266e-01 6.61045790e-01
3.86437237e-01 -1.39667737e+00 8.43639299e-02 -1.56011209e-02
-3.16278845e-01 7.71809340e-01 1.29892969e+00 -4.66581210e-02
7.77566791e-01 1.75951749e-01 -3.71480882e-01 -1.39566407e-01
-4.62539315e-01 -3.93831372e-01 1.01224534e-01 6.75655663e-01
-2.01020911e-01 2.61549503e-02 2.39912823e-01 7.40934193e-01
-7.46084154e-02 1.90193310e-01 1.86615348e-01 1.15660799e+00
-1.13173246e+00 -6.89375639e-01 -4.16950345e-01 1.13066101e+00
-3.77064422e-02 3.65762748e-02 1.39036313e-01 6.40624464e-01
5.36968350e-01 5.59854984e-01 5.98418355e-01 -2.28781089e-01
3.66696209e-01 -1.62694429e-03 5.61585486e-01 -7.69204974e-01
-2.85719693e-01 -4.91151661e-01 -4.78158981e-01 1.69375725e-02
-3.54985625e-01 -5.60120940e-01 -1.25224793e+00 -4.27059472e-01
-4.06433523e-01 -1.03295319e-01 1.96012780e-01 9.87417698e-01
-1.07521296e-01 6.01268709e-01 5.82292199e-01 -9.20711040e-01
-4.10097688e-01 -1.05417013e+00 -4.62212116e-01 1.23202421e-01
3.72126043e-01 -9.69236374e-01 -1.63785927e-02 -1.50239710e-02]
|
[9.037168502807617, 0.4983205795288086]
|
a9ead0e6-0634-4713-9cc5-9b0327bf8b02
|
blind-identification-of-ambisonic-reduced
|
2305.03558
| null |
https://arxiv.org/abs/2305.03558v2
|
https://arxiv.org/pdf/2305.03558v2.pdf
|
Blind identification of Ambisonic reduced room impulse response
|
Recently proposed Generalized Time-domain Velocity Vector (GTVV) is a generalization of relative room impulse response in spherical harmonic (aka Ambisonic) domain that allows for blind estimation of early-echo parameters: the directions and relative delays of individual reflections. However, the derived closed-form expression of GTVV mandates few assumptions to hold, most important being that the impulse response of the reference signal needs to be a minimum-phase filter. In practice, the reference is obtained by spatial filtering towards the Direction-of-Arrival of the source, and the aforementioned condition is bounded by the performance of the applied beamformer (and thus, by the Ambisonic array order). In the present work, we suggest to circumvent this problem by properly modelling the GTVV time series, which permits not only to relax the initial assumptions, but also to extract the information therein is a more consistent and efficient manner, entering the realm of blind system identification. Experiments using measured room impulse responses confirm the effectiveness of the proposed approach.
|
['Jérôme Daniel', 'Srđan Kitić']
|
2023-05-05
| null | null | null | null |
['room-impulse-response']
|
['audio']
|
[ 0.25711167 -0.21046227 0.6912693 0.03600211 -0.36007532 -0.73130494
0.43300608 -0.07471137 -0.52347636 0.6447013 0.28385648 -0.29970434
-0.75560933 -0.5651467 -0.24226633 -1.095391 -0.05163706 -0.21074061
-0.20340395 -0.14116812 0.20823732 0.6840656 -1.5122162 -0.46901593
0.9433666 0.7684253 0.06911528 0.77823335 0.12963356 0.44267696
-0.63026035 0.01260667 0.24759407 -0.54855007 -0.2274658 -0.14377472
-0.00986274 -0.08751784 -0.17205374 0.95539767 0.56323874 0.5097843
0.8682555 -0.59193295 -0.1803919 0.14528862 -0.02628077 0.24187762
0.54054993 -0.17630094 0.5692657 -0.9367408 0.3441743 0.6219007
0.6280827 0.15231094 -0.84676105 -0.40631267 -0.18725766 0.03048559
-1.5017507 -0.77670115 0.8958967 -0.46044055 0.39745802 0.66459036
0.3855019 0.87348866 0.0158592 0.02326223 1.1759661 -0.77694213
0.4783289 0.09587716 0.33316284 0.40150148 0.29801753 0.2070785
-0.23686315 -0.3775532 0.55768895 -0.4937617 -1.0876292 -0.44103068
-1.09253 0.41766092 0.15818739 0.9748161 -0.63999456 -0.30095446
-0.09663663 0.0607712 0.24164537 0.5174726 0.09386447 0.08188384
-1.0816102 -0.03483283 0.9954728 0.53265285 0.40928254 0.51264995
-0.11052727 0.47363344 0.5502664 0.8265366 0.01954 -0.4180243
0.20542249 -0.20552243 0.5220761 -1.0883871 -0.46780506 -1.1172471
-0.8090582 -0.10302888 0.8583134 -0.43686664 -0.6020387 1.6270032
0.39447296 0.33666494 0.34333593 1.0328069 0.5808481 0.71612203
-0.22325836 -0.714862 1.0739003 -0.22636427 -1.0334711 -0.14863202
0.23336555 -0.89061123 0.337301 0.4549239 -0.69566596 -0.51629007
-0.98579246 0.80191797 0.07564759 0.22826819 0.04985197 1.0539204
-1.0415308 0.12790634 -0.48008516 -0.18850482 -0.75158423 0.01317041
-0.16939771 -0.03114022 -1.0057402 0.7221177 -0.27946416 0.83777905
-0.325586 -0.7351861 -0.5694349 0.03759988 -0.06609809 -0.4508888
0.8091164 -0.68544704 -1.3547618 0.26371342 -0.41449964 -0.1678218
0.5911321 -0.07611094 -0.90274966 0.18754226 -0.14839396 -0.6280015
1.1780567 -1.4410726 -0.11157527 -0.2339582 -0.25528064 0.03783485
-0.22410658 -0.23384313 0.06651405 -0.3795164 0.6087116 -0.7412577
-0.24354951 -0.5094551 -0.36460233 0.19058016 0.12500152 -0.86354136
1.2062738 -2.427912 0.08372203 0.8240518 -0.08891673 0.17191988
0.12046577 0.65138113 -0.38798353 -0.5783241 -0.376927 -0.16444208
-0.30964664 -0.34048843 -0.5033307 1.1230676 -0.3120866 0.04083855
-0.70291287 -0.09609649 0.40139753 0.8198758 -0.34344128 0.32336828
0.52323794 0.96093875 -0.63057214 0.15525438 0.9138618 0.39547426
-0.02059685 -0.47504276 -0.8407511 -0.13470681 -1.6393199 1.0306249
-0.64150304 0.46645585 0.5394435 -0.93158895 1.2441188 0.51611257
0.51454794 -0.78316313 0.25169027 0.49485257 0.02253423 -0.50636715
0.2873846 -0.00989695 0.28936136 0.18862863 -0.2106203 0.11932751
-0.06719836 -0.31838223 0.7996171 -0.10508109 0.42323408 -0.49490887
1.1217754 -0.37705022 0.14235783 0.59572047 0.09732118 0.4740273
-0.0862775 0.01926092 -0.6616565 -0.8105319 -0.2470166 0.4123187
0.13485734 0.06257146 -0.6469573 0.18626557 -0.28206366 1.0125395
-0.31256184 0.03784931 -0.83988017 -0.62820596 0.2758046 -0.18871726
0.2265547 -0.5251134 -0.6551489 0.36479262 -0.3199919 -0.98680454
-0.11903107 0.07166017 -0.673479 -1.0119997 -0.9457606 -0.3901138
0.64247227 0.45752445 0.58811027 -0.07850101 -0.16298327 1.0611731
-0.4645968 -0.01605975 -0.25819868 -0.5803111 0.19195683 0.69562876
-0.24819872 -0.71400654 -0.62662375 0.3471612 -0.5536743 -0.41853267
0.3704868 0.34469235 0.06785985 0.20689088 0.46938264 -0.11075462
0.6296014 -0.18989049 -0.8220658 0.23200715 -0.39403576 -0.00944136
0.80975026 -0.09054582 -1.3003309 -0.16901308 -0.37329775 0.13589735
-0.38722733 0.38932535 -0.09105351 -0.49976918 0.62930745 0.68454975
-0.3357184 -0.4084682 0.16215956 0.40679562 0.6703373 -0.37871367
1.2106704 0.40669966 0.32299876 -1.5598971 -0.14708407 -0.6604474
-0.30311605 -0.46576703 0.66728157 -0.5721277 -0.87594396 0.40014204
-1.2066101 0.2007894 -0.0188189 1.0937846 -0.20124198 0.59304315
-0.05842316 -1.5922899 -0.17906652 -0.8339543 0.24916936 0.04212164
-0.21486749 -0.92570406 0.18593504 -0.03496922 0.6320709 0.12714998
0.76749104 -0.25718936 -0.45852682 -0.37932158 0.21685134 0.13804609
0.12983091 -0.2907205 -1.0787654 -0.3104636 0.65501046 0.6236045
0.4356689 0.9237354 0.34002072 -0.08209747 -0.21704416 0.7722861
1.6189862 0.44341686 0.59036535 0.22371928 0.29851976 0.6551468
0.5474046 0.6967388 -0.00782798 0.56295323 0.3537079 -0.05350067
0.05466954 0.21426879 0.07463347 0.8869192 -0.5834837 -0.58530337
-0.66851807 0.5078128 -1.2302506 -0.83625156 -0.5959203 2.7308717
0.03915397 -0.33650327 -0.29776034 0.5324913 0.5563197 0.29956412
-0.07336362 -0.24024375 -0.14060457 0.25970793 0.6897744 0.9226021
-0.7500513 -0.11023479 5.4207206 0.18173487 -1.1045789 -0.07536653
-0.25042692 0.45039088 -0.6601795 -0.12585664 -0.43504858 0.20492716
0.71240944 -0.13070455 0.47656953 0.262987 0.32625425 -0.21972534
-0.80598044 0.9503125 0.10905399 -0.53632253 -0.5413315 0.00639184
0.14464994 -0.6570807 0.29183644 -0.10678916 -0.530024 -0.67721826
0.71580446 0.81781155 0.49566036 -0.7609812 0.5274851 0.456802
-1.3352296 -0.13193682 -0.08308056 0.17783149 0.30545747 0.7374744
-0.87927455 0.90803856 0.31880757 0.08716296 -0.06974965 1.3758589
-0.05807377 0.7491019 -0.28553873 -0.04934932 0.1710602 -0.53950495
1.1006728 1.1272447 0.9188715 0.4089192 -0.35747686 0.7500688
0.5417072 0.18577087 -0.41122925 0.09758041 0.41461706 1.0088909
-0.5990098 0.42503774 -0.37038663 0.6790806 -0.58055973 1.0313005
-0.62012047 -0.42229787 0.36684266 0.13195668 0.3341546 -0.6927246
0.03576725 -0.71458167 0.09681462 -0.5572676 -0.05781135 -0.5195946
-0.6551001 0.83627707 -0.09826099 -1.5715424 -0.3520934 -0.40346873
-0.40809435 1.2852038 -1.5044777 -0.6755033 -0.24439721 0.8100109
-0.08585534 0.05698279 1.023957 0.3203426 -0.1616674 0.39634898
0.48277807 -0.12129872 0.5663336 -0.78005254 -0.15413041 1.2396543
-0.20346332 0.97729105 1.4710065 -0.44797474 -1.5498847 -0.5672882
0.8800558 0.01160716 0.43033656 -0.22364973 -0.7574747 0.16074839
0.06949122 -0.20811751 0.6525009 -0.08283471 -0.06632219 -0.19810115
-1.0465385 0.30366132 0.66160166 -0.45716405 -0.5364377 0.19787987
0.17079994 -0.10706035 -0.59420353 0.33316043 0.47259775 -1.1870453
1.2756227 0.12086418 -0.30701792 -0.45982838 -0.3672194 -1.2642866
-0.39988324 -0.5755332 0.21887991 1.1651329 0.11057111 -1.0419496
0.27984378 0.48355287 -0.20363525 -0.15678859 -1.1046996 -0.6647878
-0.5118886 -0.46866167 0.21189873 0.75930095 -0.33707842 -0.04823573
-0.5237296 1.0487797 1.0848839 0.18428846 0.29507437 -1.2978551
-0.37500608 -0.05628489 -0.16626129 -0.98493737 0.02076158 -0.31851605
0.26772594 -1.4722326 -0.7946417 -0.5785162 -0.20174271 -0.49668935
0.1374338 -0.0171548 -0.10337627 0.06718068 0.23307785 0.46564433
1.0216252 0.17877908 -0.338982 0.6559751 -0.23596384 0.583363
0.27849177 -0.15566216 -0.48085454 -0.36213467 0.28945908 0.46070006
0.39494902 -1.1563971 0.7494165 0.17488761 0.28285825 -0.52379483
0.5730747 -1.0897934 0.56254464 0.35308757 0.08517151 -0.23863699
0.01901777 0.56782436 -0.04502431 -0.37128723 0.70204496 0.1672293
-0.60050577 -0.27072588 -0.67561585 -0.43264616 0.7253782 -0.40617177
-0.09995585 -0.65204823 -0.52239007 -0.4433305 0.00681022 -0.1173785
0.5520452 -0.8851983 -0.67115265 0.4375563 -0.054296 -0.5424561
0.73328906 1.0825554 -0.256698 0.7243559 0.03697896 -0.37544233
-1.2137758 0.44980145 0.5195066 0.3712518 -0.5058969 0.8152606
0.09540376 -0.06104271 0.08977579 -0.11218647 -0.73890835 0.0601712
0.57032776 0.5220062 0.24978615 -0.97503823 -0.7138497 0.9854385
0.7157476 -0.27576432 1.2030742 -0.46030393 -0.26678503 0.28520525
1.022979 0.8782244 -0.62660223 -0.18982719 -0.26951078 -0.43376943
0.15741381 -0.45456758 -0.55796564 0.5076363 0.8164669 0.31074286
1.2796812 -0.38025376 0.21283248 0.24758849 0.6874369 -0.41970637
-0.4996658 0.22333147 0.81017756 -0.4809268 -0.11920571 -0.6122507
0.10915684 1.34608 -0.03259132 -0.05089731 0.6883827 0.23066375
0.19692738 0.14030035 0.2539849 -0.09533013 0.49212146 0.650501
0.6500045 0.06523665 -0.59479415 0.47604594 -0.15862527 -0.2729549
0.5261184 0.55651075 -0.47050238 -0.8208732 -1.152478 -0.1874935
-0.20396863 -0.09558622 0.15673584 0.5221998 -0.16142619 1.3867294
-0.17392778 -0.02480332 0.78318256 -0.07249745 0.4659764 0.11737953
-0.08748947 0.3031734 0.05802534 -0.1982167 -0.5545553 -0.54688543
-0.8660004 -0.08127931 -0.34250346 0.7449035 0.979707 0.8649356
-0.12962243 0.7360862 0.87681276 -0.7550819 -0.5341975 -0.64286923
-0.74756664 -0.1315385 0.79966396 -0.3957558 -0.87309235 -0.3814642 ]
|
[15.173714637756348, 5.7023725509643555]
|
f5ad2924-f031-4926-9453-ab41d74355fb
|
svldl-improved-speaker-age-estimation-using
|
2210.09524
| null |
https://arxiv.org/abs/2210.09524v2
|
https://arxiv.org/pdf/2210.09524v2.pdf
|
SVLDL: Improved Speaker Age Estimation Using Selective Variance Label Distribution Learning
|
Estimating age from a single speech is a classic and challenging topic. Although Label Distribution Learning (LDL) can represent adjacent indistinguishable ages well, the uncertainty of the age estimate for each utterance varies from person to person, i.e., the variance of the age distribution is different. To address this issue, we propose selective variance label distribution learning (SVLDL) method to adapt the variance of different age distributions. Furthermore, the model uses WavLM as the speech feature extractor and adds the auxiliary task of gender recognition to further improve the performance. Two tricks are applied on the loss function to enhance the robustness of the age estimation and improve the quality of the fitted age distribution. Extensive experiments show that the model achieves state-of-the-art performance on all aspects of the NIST SRE08-10 and a real-world datasets.
|
['Jing Xiao', 'Junqing Peng', 'Jianzong Wang', 'Zuheng Kang']
|
2022-10-18
| null | null | null | null |
['age-estimation', 'age-estimation']
|
['computer-vision', 'miscellaneous']
|
[-3.60420704e-01 -4.44448441e-02 -1.85394138e-01 -9.41066742e-01
-9.18104649e-01 -3.35584521e-01 4.77476835e-01 9.94743630e-02
-5.45732081e-01 6.31401181e-01 3.35611224e-01 -6.70595933e-03
2.63587505e-01 -4.01408345e-01 -2.54500955e-01 -8.78671885e-01
3.54222625e-01 3.58537465e-01 9.95163918e-02 2.07967490e-01
9.33253914e-02 -1.05140328e-01 -1.49951494e+00 -1.97024941e-01
1.13224745e+00 1.18979561e+00 -1.19770840e-01 5.28502524e-01
-2.41603136e-01 4.25522834e-01 -8.93983603e-01 -6.05995238e-01
-6.58327565e-02 -1.01054370e-01 -5.59370399e-01 -2.03271970e-01
5.20589888e-01 -4.37217623e-01 -3.51305664e-01 1.11848497e+00
7.31050372e-01 9.58861187e-02 1.06460142e+00 -1.37659883e+00
-5.73372602e-01 8.44008446e-01 -8.88612688e-01 -1.18194513e-01
2.70078868e-01 -4.17295933e-01 6.05491817e-01 -6.36690974e-01
-1.73159182e-01 1.57611310e+00 6.46717489e-01 8.39745343e-01
-7.84161210e-01 -1.17149317e+00 2.98136711e-01 2.45635673e-01
-1.59368289e+00 -5.85215032e-01 6.69277012e-01 -3.99724483e-01
-2.55833883e-02 7.76797011e-02 2.32211098e-01 1.31380320e+00
-1.65999830e-02 7.66888976e-01 1.32129681e+00 -2.20094323e-01
2.69828171e-01 1.28119171e-01 4.10636902e-01 6.04497135e-01
7.00151175e-02 -4.45896760e-02 -6.91635132e-01 -1.37900531e-01
2.78551847e-01 -2.16360599e-01 2.63216440e-02 -9.79039818e-02
-8.11567426e-01 6.50509655e-01 -1.05909012e-01 -1.29261583e-01
2.37930238e-01 2.59149149e-02 4.35175925e-01 1.88199371e-01
8.37515652e-01 -1.83096483e-01 -4.42435861e-01 -4.45055544e-01
-8.67643476e-01 3.19138944e-01 6.59174263e-01 6.96290493e-01
4.96001840e-01 -2.17098314e-02 -2.85799682e-01 1.13587296e+00
4.64032114e-01 8.19010019e-01 4.33720559e-01 -9.27572012e-01
4.75530237e-01 2.85749793e-01 -4.29840796e-02 -4.78254765e-01
-3.39702278e-01 -2.45557010e-01 -8.47290277e-01 9.06486437e-02
9.90989327e-01 -1.97187185e-01 -1.01010787e+00 2.10465789e+00
4.31973487e-01 6.24813586e-02 -2.44320065e-01 6.13660753e-01
7.72296846e-01 5.74289262e-01 4.48123664e-01 -4.49005127e-01
1.36614931e+00 -6.67452455e-01 -7.79591143e-01 -4.75296915e-01
2.57227957e-01 -6.06061161e-01 9.96702313e-01 3.13567221e-01
-8.91773462e-01 -6.21330619e-01 -1.02124965e+00 -6.73766434e-02
-1.75952628e-01 3.43968093e-01 4.84492242e-01 1.10484815e+00
-4.97732520e-01 3.92323792e-01 -7.53521383e-01 -8.61034393e-02
-2.47354377e-02 4.95381504e-01 -1.96028620e-01 1.01336524e-01
-1.27580082e+00 4.23835188e-01 2.95232058e-01 -2.89311290e-01
-4.05623674e-01 -8.18679452e-01 -8.91606808e-01 2.16508191e-02
1.60241365e-01 -4.62514162e-01 1.43916154e+00 -6.87472939e-01
-1.54341125e+00 8.97365332e-01 -3.20100963e-01 -1.66858777e-01
5.05988181e-01 -2.81031519e-01 -5.19091606e-01 -3.33067924e-01
-9.63785574e-02 4.12917286e-01 1.15714896e+00 -8.99835825e-01
-6.10031188e-01 -7.70448744e-01 -2.36886442e-01 1.29027233e-01
-5.97735107e-01 1.33418187e-01 -4.05980945e-01 -7.60997355e-01
3.56657833e-01 -9.41965699e-01 2.45677769e-01 -2.29792863e-01
-7.65718520e-02 -5.37100315e-01 6.25568211e-01 -1.13323498e+00
1.64067149e+00 -2.46043658e+00 1.06082698e-02 1.40958980e-01
1.47177070e-01 -3.80083695e-02 2.04088748e-01 -7.17547014e-02
1.61613494e-01 -5.76971807e-02 3.45155969e-02 -7.12314963e-01
2.22429723e-01 -1.92160774e-02 5.64223900e-02 4.81851876e-01
-3.18866074e-01 1.31925061e-01 -6.63844228e-01 -6.21434212e-01
-7.25336969e-02 3.73302996e-01 -2.31252715e-01 5.22332907e-01
2.96558529e-01 5.83156824e-01 -1.94998592e-01 4.20522302e-01
9.36905086e-01 3.28456908e-01 -1.60008054e-02 -2.89371610e-01
2.49518212e-02 1.86820760e-01 -1.22389138e+00 1.47320521e+00
-5.41452527e-01 7.72669613e-02 8.47764760e-02 -6.07929289e-01
1.14766741e+00 1.27513751e-01 2.33199403e-01 -2.95263857e-01
3.02826405e-01 2.06662178e-01 4.20519486e-02 -1.44876391e-01
3.89927357e-01 2.32401118e-02 -4.42962945e-01 5.33110023e-01
9.75370780e-03 -7.09392205e-02 3.32955783e-03 -8.82958695e-02
5.03986418e-01 -3.12546492e-01 6.33447990e-02 -1.71792373e-01
8.62219095e-01 -1.22665179e+00 9.62408543e-01 5.80498517e-01
-5.92492819e-01 6.77119076e-01 5.61446548e-01 -8.40099603e-02
-1.01690757e+00 -1.54404223e+00 -1.20193958e-01 1.36570740e+00
8.73903092e-03 -3.68521363e-01 -9.79033172e-01 -1.01675105e+00
-5.91501184e-02 6.44714296e-01 -5.59249938e-01 -3.94546509e-01
-3.47709119e-01 -6.59137547e-01 5.14975250e-01 6.83853567e-01
5.30705273e-01 -4.49820578e-01 2.80663967e-01 -2.77255982e-01
-2.88187057e-01 -1.14240515e+00 -9.01922286e-01 -1.74379706e-01
-4.30963069e-01 -8.18451464e-01 -8.15201700e-01 -7.78875470e-01
6.92197859e-01 -2.99268126e-01 8.13387156e-01 -3.23575675e-01
3.04477632e-01 2.47146115e-01 -2.31787965e-01 -6.22228205e-01
-5.88322997e-01 3.88893425e-01 5.49030364e-01 5.97459339e-02
5.36400974e-01 -4.63449657e-01 -5.16580641e-01 2.85675615e-01
-2.07515985e-01 -3.35257202e-01 1.68452173e-01 7.06042826e-01
1.09005772e-01 1.19672701e-01 9.29889441e-01 -5.28954864e-01
7.03758836e-01 -1.82882637e-01 -3.85949552e-01 3.00255746e-01
-9.30982888e-01 3.12486917e-01 4.05960351e-01 -7.08696306e-01
-1.23086309e+00 -7.36456439e-02 -3.98205370e-01 -3.11165936e-02
7.04517290e-02 5.29446267e-02 -4.47229683e-01 2.81114310e-01
1.04622185e-01 7.32131526e-02 2.40014225e-01 -5.87877929e-01
2.13523909e-01 1.32521939e+00 8.38258445e-01 -7.08043456e-01
7.27527261e-01 -7.82030169e-03 -1.31265596e-01 -6.02377236e-01
-1.10013008e+00 -3.50408763e-01 -2.68844247e-01 -1.55754656e-01
6.90388322e-01 -1.09438515e+00 -8.84854257e-01 1.28547812e+00
-9.02803123e-01 -7.91818351e-02 1.23449735e-01 5.84703207e-01
-4.37816888e-01 5.87933362e-01 -8.10067296e-01 -1.19383228e+00
-6.02124512e-01 -1.06867588e+00 1.13953507e+00 5.89604914e-01
-2.87514627e-01 -7.61487007e-01 -1.10721156e-01 4.55365062e-01
-9.49315447e-03 -3.67285848e-01 9.24161792e-01 -6.19809449e-01
3.07637066e-01 -2.71139592e-01 -1.58934906e-01 6.37268424e-01
2.15011925e-01 8.74799490e-02 -1.10447371e+00 -4.55321193e-01
-1.42935095e-02 -3.23920220e-01 8.14020038e-01 4.45698231e-01
1.59370708e+00 2.38679890e-02 -7.64414668e-02 3.55679184e-01
6.05392098e-01 2.21272230e-01 5.28683305e-01 -4.21259291e-02
6.68758452e-01 7.12473214e-01 7.70061433e-01 6.10953927e-01
6.44393384e-01 6.68367922e-01 -1.30185947e-01 3.69526297e-01
-1.17755920e-01 -3.54389191e-01 5.88809371e-01 1.11526620e+00
3.94143499e-02 6.00592718e-02 -7.69675255e-01 2.39751503e-01
-1.64586592e+00 -6.79267287e-01 3.13268244e-01 2.66642380e+00
1.12736607e+00 2.67638862e-01 5.24057448e-01 3.17886293e-01
9.58253205e-01 4.71603572e-01 -4.69431818e-01 -4.25476164e-01
8.92354101e-02 -1.49281397e-01 5.46580672e-01 5.38847685e-01
-1.19605756e+00 5.52884102e-01 6.64704847e+00 1.03183937e+00
-9.86931562e-01 6.87454492e-02 8.60863090e-01 4.66272011e-02
2.31893044e-02 -4.69929487e-01 -1.05234671e+00 8.63535583e-01
1.00145185e+00 -3.99405271e-01 3.80314678e-01 8.14218640e-01
1.01326279e-01 -1.06222153e-01 -9.71394718e-01 1.36534524e+00
3.00811417e-02 -3.60723972e-01 -5.37555516e-01 -6.00046888e-02
4.86261398e-01 -5.55232048e-01 4.56676036e-01 5.84066272e-01
3.77682224e-02 -7.09755182e-01 8.37502062e-01 4.71274167e-01
9.00976121e-01 -1.21686780e+00 7.84339726e-01 3.05793554e-01
-1.14574313e+00 -3.68414223e-01 -2.56488144e-01 3.94761004e-02
8.05257335e-02 8.02363813e-01 -4.33443069e-01 2.11164385e-01
7.52029359e-01 1.55173391e-01 -7.53598511e-01 6.90361261e-01
-3.03381920e-01 6.72239363e-01 -1.93842396e-01 7.95476735e-02
-3.78305703e-01 -7.89241567e-02 2.71664232e-01 8.38311732e-01
5.14792502e-01 -2.95852363e-01 1.54746473e-01 3.51085216e-01
-1.33666396e-01 4.37380709e-02 -6.58395737e-02 -5.90337925e-02
9.99002695e-01 1.07354474e+00 -1.86665341e-01 -2.99849231e-02
-4.32775855e-01 8.08263719e-01 3.56852382e-01 2.00196326e-01
-8.04810882e-01 -3.35115731e-01 8.78765523e-01 1.27118468e-01
1.03749186e-01 -2.82742143e-01 -3.39739233e-01 -9.14368808e-01
-1.46567211e-01 -9.20089662e-01 4.44066167e-01 -3.98574442e-01
-1.53549266e+00 8.96156356e-02 4.01855893e-02 -8.15871835e-01
-3.83285612e-01 -3.93585205e-01 -6.17376745e-01 8.55195165e-01
-9.34758544e-01 -1.15459013e+00 -3.01249534e-01 2.15674385e-01
6.27408147e-01 -4.08832669e-01 4.95599538e-01 3.86063516e-01
-6.80372477e-01 1.33402216e+00 1.21176668e-01 1.48438647e-01
1.29032946e+00 -1.36056530e+00 3.47328812e-01 5.85291326e-01
-3.73143047e-01 5.08240342e-01 8.25326800e-01 -5.98845303e-01
-9.11783993e-01 -8.58836889e-01 9.55207586e-01 -2.78302848e-01
5.03848612e-01 -6.15149498e-01 -7.09909797e-01 3.23367000e-01
-4.51009385e-02 -3.89316738e-01 6.91484213e-01 6.08204067e-01
-6.67674839e-01 -4.10864532e-01 -1.18166041e+00 5.50342619e-01
7.82135427e-01 -6.75043046e-01 -4.27608401e-01 -1.25097424e-01
7.60746181e-01 -4.10621643e-01 -1.03860056e+00 5.22398651e-01
9.20390487e-01 -8.85819018e-01 9.63643789e-01 -1.74800590e-01
1.15464777e-01 -1.28329322e-01 -1.83190525e-01 -1.39096403e+00
-1.54730946e-01 -4.30998951e-01 -6.67688787e-01 1.89490557e+00
1.33343965e-01 -5.33294737e-01 8.12407911e-01 8.60761523e-01
1.01978920e-01 -7.39789903e-01 -8.09349418e-01 -8.69272470e-01
2.21538126e-01 -2.77607769e-01 9.88570929e-01 4.13995743e-01
-1.30533710e-01 4.69100058e-01 -5.05533993e-01 2.22550720e-01
8.91835392e-01 -4.40777451e-01 6.47922337e-01 -1.33555436e+00
-2.20143959e-01 -2.88596511e-01 -3.37826610e-01 -8.91578674e-01
6.40263736e-01 -2.44316280e-01 1.34601131e-01 -1.06012321e+00
3.25174719e-01 -4.01823789e-01 -4.07129496e-01 6.45358488e-02
-6.44905090e-01 -1.43212467e-01 7.59434924e-02 -2.08219886e-01
-4.90155160e-01 6.08874977e-01 8.39813411e-01 -3.78364384e-01
-2.04525907e-02 5.35539865e-01 -6.43971264e-01 8.90099466e-01
8.74619246e-01 -3.94330859e-01 -6.34165466e-01 1.61680263e-02
-3.35383229e-02 -2.39658758e-01 -2.14506239e-01 -1.04032159e+00
1.04871020e-01 -1.24159612e-01 3.75721276e-01 -7.33461976e-01
3.82316947e-01 -3.20281267e-01 -2.03792170e-01 2.68236607e-01
-2.41321921e-01 -7.66307786e-02 -2.76267946e-01 4.95567322e-01
-1.49007067e-01 -3.52771491e-01 9.05709028e-01 2.29941085e-01
-4.65861648e-01 5.21214068e-01 2.00851094e-02 2.14944392e-01
5.63108861e-01 6.88546002e-02 -5.53553939e-01 -6.17533684e-01
-4.42465574e-01 3.29479247e-01 4.73052144e-01 5.51391780e-01
1.82300970e-01 -1.48189759e+00 -6.97036684e-01 1.46481887e-01
1.32352114e-01 -2.49684706e-01 4.37406331e-01 7.26967692e-01
1.10244282e-01 -1.21236347e-01 1.42636582e-01 -3.58698100e-01
-1.55237174e+00 4.80892181e-01 1.85123473e-01 -2.51985192e-01
1.56718880e-01 1.04887295e+00 3.23139250e-01 -5.80508113e-01
6.89248741e-01 9.35137197e-02 -3.19094002e-01 2.51897037e-01
7.90884495e-01 6.66617274e-01 -2.21216097e-01 -4.50798780e-01
-4.08416301e-01 5.93231022e-01 -3.48900825e-01 -2.30937555e-01
8.05021405e-01 -6.29060626e-01 -3.88097726e-02 8.94789219e-01
1.02836788e+00 2.37048790e-01 -1.29251814e+00 -1.59088269e-01
-1.46238238e-01 -3.76357943e-01 -1.04042277e-01 -6.09096766e-01
-9.49709713e-01 7.08543181e-01 8.80244255e-01 5.28124087e-02
1.00889015e+00 2.56436840e-02 9.60129678e-01 -1.43626064e-01
1.89397410e-01 -1.60024667e+00 1.90252170e-01 5.41758478e-01
5.51186919e-01 -1.29396605e+00 1.07983269e-01 -5.67643344e-01
-6.35047555e-01 8.78894389e-01 8.58831882e-01 5.43273628e-01
7.05636382e-01 1.88594729e-01 3.34690213e-01 4.68250841e-01
-2.87376493e-01 1.10962383e-01 4.77888286e-01 8.52805495e-01
6.35707617e-01 2.43114233e-01 -8.16858888e-01 1.11677563e+00
-4.36449021e-01 -4.22925740e-01 2.28702143e-01 3.34612727e-01
-4.50973988e-01 -1.36613178e+00 -3.47197741e-01 4.00115162e-01
-6.59136534e-01 1.25498325e-01 -1.91115201e-01 2.28888020e-01
1.40554100e-01 1.06046903e+00 8.35124925e-02 -6.68940187e-01
1.65246595e-02 3.68416607e-01 6.86445355e-01 -3.81493419e-01
-2.33478174e-01 -1.18669868e-01 2.17117444e-01 -1.52667984e-01
-1.66603029e-01 -9.39113021e-01 -1.15084875e+00 -5.00197768e-01
-2.33259931e-01 2.35560179e-01 8.43366385e-01 8.97911847e-01
-7.24223182e-02 3.11259598e-01 1.11061394e+00 -3.69628936e-01
-1.00668705e+00 -1.10708296e+00 -1.13886213e+00 3.70910347e-01
1.16137840e-01 -8.22111726e-01 -5.89714587e-01 -1.67477190e-01]
|
[14.15589427947998, 6.023719787597656]
|
53f8d7f3-9b23-41be-bdba-91227e6a5139
|
euler-detecting-network-lateral-movement-via
| null | null |
https://www.ndss-symposium.org/ndss-paper/auto-draft-227/
|
https://www.ndss-symposium.org/wp-content/uploads/2022-107A-paper.pdf
|
Euler: Detecting Network Lateral Movement via Scalable Temporal Link Prediction
|
Lateral movement is a key stage of system compromise used by advanced persistent threats. Detecting it is no
simple task. When network host logs are abstracted into discrete temporal graphs, the problem can be reframed as anomalous edge detection in an evolving network. Research in modern deep graph learning techniques has produced many creative and complicated
models for this task. However, as is the case in many machine learning fields, the generality of models is of paramount importance for accuracy and scalability during training and inference. In this paper, we propose a formalized approach to this problem with a framework we call EULER. It consists of a model-agnostic graph neural network stacked upon a model-agnostic sequence encoding layer such as a recurrent neural network. Models built according to the EULER framework can easily distribute their graph convolutional layers across multiple machines for large performance improvements. Additionally, we demonstrate that EULER-based models are competitive, or better than many state-of-the-art approaches to anomalous link detection and prediction. As anomaly-based intrusion detection systems, EULER models can efficiently identify anomalous connections between entities with high precision and outperform
other unsupervised techniques for anomalous lateral movement detection.
|
['H. Howie Huang', 'Isaiah J. King']
|
2022-04-24
| null | null | null |
ndss-2022-4
|
['edge-detection', 'dynamic-link-prediction']
|
['computer-vision', 'graphs']
|
[ 1.65137619e-01 4.02413681e-02 -1.90751523e-01 -3.95443849e-02
2.56047137e-02 -5.66233277e-01 5.54792583e-01 6.27537608e-01
-1.58617407e-01 1.23228543e-01 -2.90105700e-01 -1.04569852e+00
-1.75197557e-01 -9.74372566e-01 -5.52737296e-01 -2.45449081e-01
-9.13931966e-01 7.34375000e-01 7.36878574e-01 -4.59796488e-01
3.45348239e-01 1.02267325e+00 -7.51639962e-01 1.08985804e-01
2.21284360e-01 1.05097759e+00 -9.00157273e-01 1.15098608e+00
-2.04745337e-01 1.14071512e+00 -7.64840066e-01 -5.81084192e-01
1.78494215e-01 -2.31564149e-01 -1.05439341e+00 -1.91552773e-01
2.40438968e-01 -1.95414111e-01 -6.49269581e-01 9.82269406e-01
2.75490999e-01 -4.73684855e-02 4.91966695e-01 -1.76997745e+00
-2.55938560e-01 7.38835275e-01 -6.28608346e-01 9.48038638e-01
2.52719134e-01 9.68406573e-02 1.23742676e+00 -3.00597072e-01
5.21289885e-01 9.28237855e-01 1.00208020e+00 3.70352685e-01
-1.27917981e+00 -5.84860981e-01 6.38883173e-01 3.95698100e-01
-1.04376578e+00 -2.37188593e-01 8.92116964e-01 -3.81869942e-01
1.53332734e+00 2.60291576e-01 6.48945212e-01 1.27020514e+00
5.40134728e-01 6.12230718e-01 1.57673791e-01 -6.90371543e-02
1.31999806e-01 -2.72335202e-01 3.78399849e-01 9.23101723e-01
4.67694759e-01 -2.70168850e-04 -2.49234468e-01 -5.58999300e-01
2.92425960e-01 4.40696239e-01 -7.94835016e-02 -3.27606112e-01
-6.87215626e-01 1.03158319e+00 7.01778114e-01 4.29662824e-01
-2.92833328e-01 4.41593558e-01 9.34770226e-01 8.33326638e-01
7.07198143e-01 7.73789167e-01 -4.28103775e-01 -6.90421015e-02
-8.98676157e-01 -6.67874590e-02 1.28478873e+00 4.63473201e-01
2.75089681e-01 5.81333458e-01 2.08199695e-01 1.68263301e-01
3.01489294e-01 -8.15677047e-02 2.78639406e-01 -1.31389543e-01
7.95825720e-02 9.39683378e-01 -6.67789757e-01 -1.34487903e+00
-8.82181466e-01 -7.97488928e-01 -8.47990096e-01 9.76931304e-02
1.14586659e-01 -1.52975991e-01 -8.96245182e-01 1.61879265e+00
2.23441184e-01 5.77984273e-01 -1.71906427e-01 2.50129879e-01
2.65893817e-01 4.74496514e-01 -2.10784934e-03 -1.78919360e-02
9.28181469e-01 -8.96804273e-01 -3.76136184e-01 -5.04723907e-01
1.09221029e+00 -2.60490924e-01 4.73537266e-01 4.12849724e-01
-7.57009327e-01 2.06016570e-01 -1.25309443e+00 4.88558710e-01
-9.71857488e-01 -9.54562604e-01 7.95358241e-01 8.20272267e-01
-1.30323935e+00 8.79722655e-01 -1.03221822e+00 -6.88397288e-01
5.56608081e-01 5.64223230e-01 -3.11734885e-01 4.08356965e-01
-1.30524886e+00 7.97253728e-01 5.22271633e-01 -9.57380384e-02
-7.49236286e-01 -4.15392965e-01 -7.48316348e-01 2.57810175e-01
5.87585509e-01 -4.81516749e-01 1.19118142e+00 -7.24098861e-01
-1.07742000e+00 6.78160429e-01 2.80641645e-01 -1.26840615e+00
3.84548366e-01 -7.95657635e-02 -1.07023478e+00 1.12883121e-01
-1.98404372e-01 -2.50789493e-01 1.12021077e+00 -7.37585604e-01
-4.05129313e-01 -2.44136125e-01 9.57394391e-02 -3.60157043e-01
-6.92184985e-01 2.80693680e-01 -1.70357656e-02 -5.40388644e-01
-2.96385325e-02 -8.49822640e-01 -4.92353410e-01 -2.73597777e-01
-7.68123448e-01 -2.30739072e-01 1.37965679e+00 -5.25450289e-01
1.75889385e+00 -1.81405723e+00 -4.30096723e-02 8.35803568e-01
8.52314651e-01 5.83346426e-01 4.58231419e-02 8.71939659e-01
-3.53403121e-01 3.96893710e-01 -3.44613314e-01 -2.04263151e-01
-6.48779124e-02 2.41650287e-02 -5.48318326e-01 5.07555544e-01
3.76867205e-01 1.04331481e+00 -9.60944474e-01 -1.38037637e-01
-9.01677161e-02 2.12940186e-01 -5.10968208e-01 1.42254815e-01
-3.18316668e-01 1.15523778e-01 -3.47812980e-01 6.21784866e-01
2.27985263e-01 -7.57907689e-01 3.14330578e-01 2.78176665e-01
5.45525372e-01 3.38019013e-01 -7.69125938e-01 1.15626287e+00
-2.27621973e-01 8.21079314e-01 -2.15258375e-01 -1.25830710e+00
8.00457656e-01 2.52005965e-01 5.69821179e-01 -5.50984621e-01
5.55556156e-02 1.00779429e-01 3.28600228e-01 -2.02439323e-01
3.68143082e-01 2.23762512e-01 -1.24910258e-01 8.61611307e-01
6.93511367e-02 3.78901809e-01 1.65482029e-01 7.72924900e-01
1.94691181e+00 -6.97309196e-01 4.59037095e-01 1.14126660e-01
4.99178797e-01 -3.81126814e-02 3.01678151e-01 9.19184864e-01
-1.60571456e-01 4.36368510e-02 1.03640771e+00 -8.42313051e-01
-8.11818361e-01 -1.12221074e+00 2.85690635e-01 1.11758399e+00
-2.85406977e-01 -9.17267919e-01 -4.60742056e-01 -1.30906582e+00
1.39998585e-01 4.47314471e-01 -5.27793944e-01 -8.05240035e-01
-7.41607130e-01 -8.36196899e-01 1.01329660e+00 6.14332736e-01
1.52982026e-01 -1.08436012e+00 -2.04643428e-01 4.62448180e-01
3.81378293e-01 -1.23376763e+00 -2.81051576e-01 3.81852925e-01
-6.73578382e-01 -1.39561021e+00 1.97230384e-01 -6.07618928e-01
4.86099511e-01 8.62081572e-02 1.48822021e+00 6.46830201e-01
-5.05330920e-01 5.16003609e-01 -3.24073106e-01 -4.01443154e-01
-6.60244346e-01 4.32439536e-01 3.99233013e-01 1.62494987e-01
5.06337106e-01 -1.07302475e+00 -4.95523930e-01 2.84515116e-02
-1.04360783e+00 -6.61237240e-01 4.36522931e-01 6.05286598e-01
1.84987730e-03 2.17282414e-01 6.37255132e-01 -1.45924032e+00
1.00720704e+00 -8.98204803e-01 -4.83539820e-01 1.30979540e-02
-9.90904212e-01 3.97574762e-03 9.72584307e-01 -3.41899097e-01
-1.28001630e-01 -4.32305962e-01 -2.33240843e-01 -6.16241813e-01
-3.63280135e-03 6.51230454e-01 1.72265381e-01 -4.56681907e-01
7.77315021e-01 2.54445314e-01 -5.74670136e-02 -7.61934519e-02
1.52992815e-01 3.13313305e-01 3.67022485e-01 -1.52843028e-01
1.15827537e+00 3.37084532e-01 3.29452634e-01 -9.49132204e-01
-5.92054486e-01 -5.08786201e-01 -5.31523526e-01 -1.63537398e-01
5.36017954e-01 -3.28203499e-01 -7.84829617e-01 4.68294114e-01
-1.02372086e+00 -3.07578683e-01 -1.83786288e-01 -1.01830564e-01
-1.91532180e-01 5.82609832e-01 -9.49847162e-01 -7.23161340e-01
-6.45888805e-01 -6.44972384e-01 5.61063170e-01 -7.98903629e-02
-4.28514242e-01 -1.71313846e+00 3.50662261e-01 -2.94603378e-01
7.32081592e-01 4.94390160e-01 1.06657946e+00 -1.71248925e+00
-3.54086787e-01 -9.17477667e-01 -1.09899662e-01 1.64229661e-01
7.02740848e-02 1.84431911e-01 -7.12275565e-01 -6.50995553e-01
-2.69019276e-01 6.47458732e-02 8.56828511e-01 -7.91283324e-02
1.20532119e+00 -5.91709733e-01 -6.44134879e-01 7.23241150e-01
1.10055506e+00 1.85660720e-01 3.49475324e-01 2.65014172e-01
1.02500963e+00 2.32012972e-01 -1.78963602e-01 2.58826882e-01
1.87612742e-01 4.29207295e-01 8.11052144e-01 -1.10657789e-01
3.39168429e-01 -2.10847646e-01 4.29219276e-01 6.46459043e-01
2.35504553e-01 -5.86073279e-01 -1.42084730e+00 3.19953948e-01
-1.83289671e+00 -1.14320314e+00 -1.99084550e-01 1.82634914e+00
8.06789845e-02 8.61793220e-01 4.19592977e-01 4.58391935e-01
5.75843632e-01 5.83915889e-01 -6.06041253e-01 -9.25061941e-01
1.47774741e-01 2.22926274e-01 4.79391575e-01 4.21464622e-01
-1.24790752e+00 9.72505867e-01 6.22962141e+00 5.03686845e-01
-1.24998188e+00 -1.38317853e-01 3.77175987e-01 8.79875273e-02
6.42558783e-02 -1.11838803e-02 -4.71921891e-01 3.12540650e-01
1.53586769e+00 -3.55923504e-01 2.83039600e-01 8.69774163e-01
-3.00445080e-01 6.29138112e-01 -1.32742989e+00 6.61405087e-01
1.51271015e-01 -1.43851507e+00 2.34775692e-01 3.32273453e-01
3.24892581e-01 3.37580174e-01 1.78427711e-01 3.29520345e-01
7.14071751e-01 -1.19139302e+00 3.01260930e-02 2.60200143e-01
3.00188333e-01 -9.34432089e-01 7.14354515e-01 1.13723047e-01
-1.30508912e+00 -3.13670874e-01 8.60519800e-03 2.15779599e-02
8.80615860e-02 4.64284718e-01 -1.20541048e+00 7.30450451e-01
5.72929561e-01 1.04809928e+00 -8.58771980e-01 9.62579012e-01
-1.77355111e-01 8.41909111e-01 -3.78652602e-01 1.00498199e-01
5.49748659e-01 1.55207232e-01 9.42257285e-01 1.23664796e+00
-5.93811907e-02 -4.72078323e-01 3.72900724e-01 3.77402484e-01
-3.25624824e-01 -2.65731663e-01 -1.20881081e+00 -6.23266935e-01
4.22804445e-01 1.30350387e+00 -9.79678214e-01 -1.47697538e-01
-3.76838118e-01 1.01235306e+00 5.15349329e-01 2.13590190e-01
-6.20826662e-01 -5.08241177e-01 7.15465963e-01 3.08707565e-01
1.47957921e-01 -2.23860934e-01 1.90268233e-01 -1.08347285e+00
-1.38475969e-01 -9.47574735e-01 1.05931664e+00 -5.09282127e-02
-1.55660963e+00 1.10120881e+00 -5.02618194e-01 -1.24219656e+00
-5.72774291e-01 -7.71299779e-01 -1.31817865e+00 3.93349320e-01
-1.08197141e+00 -1.02274477e+00 2.26044171e-02 7.71683216e-01
1.92261294e-01 -3.91112387e-01 9.69017208e-01 2.19565809e-01
-9.70288217e-01 7.28973389e-01 -2.07234278e-01 7.29163706e-01
3.72283190e-01 -1.44464135e+00 1.37795067e+00 1.38529837e+00
5.14055550e-01 4.48892742e-01 6.36775136e-01 -7.96122432e-01
-1.26576781e+00 -1.31634343e+00 5.43647647e-01 -6.25139117e-01
1.43226159e+00 -4.83586282e-01 -1.15654075e+00 1.35697806e+00
-8.48006234e-02 3.39349806e-01 7.32822955e-01 4.13440019e-01
-7.50586212e-01 1.31429076e-01 -7.94462264e-01 6.79199755e-01
1.24140012e+00 -8.05894077e-01 -1.67691022e-01 5.37513673e-01
7.96116769e-01 -9.80299488e-02 -7.46059418e-01 3.30595911e-01
7.72433579e-02 -9.17664886e-01 9.20318961e-01 -1.47708631e+00
-2.12457746e-01 -1.63879618e-02 2.65281737e-01 -1.37834764e+00
-4.03412461e-01 -9.97831523e-01 -1.15744114e+00 7.38596261e-01
5.77101052e-01 -1.10911441e+00 1.04159284e+00 2.41977274e-02
-8.21566656e-02 -8.53014112e-01 -7.76294768e-01 -9.70484495e-01
-9.33328494e-02 -6.00810707e-01 5.68558872e-01 1.17660260e+00
2.50265539e-01 6.01397097e-01 -3.10784340e-01 5.58808029e-01
6.16417527e-01 -1.44613357e-02 6.26897037e-01 -1.81865978e+00
-2.27699161e-01 -8.99427414e-01 -1.12574649e+00 -5.40183067e-01
2.81736553e-01 -1.15277195e+00 -6.89610898e-01 -1.31840169e+00
-3.92891765e-01 -1.46987453e-01 -6.01848900e-01 4.42160100e-01
3.80578861e-02 1.32487804e-01 -1.75245583e-01 2.59412229e-02
-8.02913666e-01 1.26078457e-01 3.94104362e-01 -3.20902228e-01
-1.88846543e-01 3.40239078e-01 -4.80241090e-01 7.29096711e-01
1.04227543e+00 -5.57949066e-01 -4.34630096e-01 -3.65820155e-02
7.30408669e-01 -4.14980687e-02 3.78125489e-01 -1.15393090e+00
6.06103897e-01 2.41761848e-01 2.17537865e-01 -3.32388073e-01
1.53254896e-01 -8.28894734e-01 -1.88565716e-01 8.85149717e-01
-1.18233301e-01 7.20277190e-01 1.08382925e-01 1.04919231e+00
-1.76302388e-01 3.62438500e-01 5.07886887e-01 8.25493131e-03
-7.80169308e-01 9.77152944e-01 -5.26961267e-01 2.51099974e-01
1.16285062e+00 -9.09848586e-02 -5.31790674e-01 -7.37160742e-01
-8.99752557e-01 3.82947206e-01 2.78657198e-01 6.80012941e-01
7.25961566e-01 -1.07895052e+00 -4.67268974e-01 3.20585459e-01
2.77542889e-01 -4.19607013e-01 1.20464237e-02 1.02315366e+00
-6.12063587e-01 2.58978009e-01 -1.47866324e-01 -5.41592300e-01
-1.14860320e+00 9.32886124e-01 6.08307779e-01 -8.13302934e-01
-8.31710041e-01 7.58985758e-01 -2.38490477e-01 -3.12382251e-01
2.06877261e-01 1.89216569e-01 -8.82866532e-02 -2.62861270e-02
4.73257840e-01 2.48859853e-01 3.97237390e-01 -4.40487713e-01
-6.62671626e-01 -2.04304270e-02 -6.80858970e-01 4.73534137e-01
1.15454888e+00 2.48459592e-01 -2.52098471e-01 4.74403352e-01
1.15407634e+00 -1.40878618e-01 -5.86832285e-01 -4.39219505e-01
6.59401059e-01 9.57342908e-02 -5.59437871e-02 -3.90020043e-01
-1.20597577e+00 8.43253434e-01 2.94398397e-01 1.17929113e+00
9.65926766e-01 -1.62637949e-01 8.32806468e-01 7.52723396e-01
1.44816339e-01 -6.12031758e-01 3.65519792e-01 8.02563608e-01
2.46192515e-01 -1.10851479e+00 -1.85112849e-01 -2.44579196e-01
-2.13988721e-01 1.21691430e+00 7.01020598e-01 -3.74862999e-01
1.02229822e+00 3.76875192e-01 -1.92288920e-01 -6.83244467e-01
-1.02273905e+00 -1.22867962e-02 2.27776289e-01 5.49450994e-01
1.82009846e-01 -1.90803483e-01 3.57625693e-01 2.11650193e-01
-1.37819335e-01 -8.13774407e-01 6.87563419e-01 1.05910814e+00
-3.94576550e-01 -9.99725342e-01 2.29014114e-01 8.26981664e-01
-7.14980543e-01 -2.50605494e-01 -8.20842922e-01 8.38308454e-01
-7.56073117e-01 8.75506938e-01 2.27357909e-01 -8.86608362e-01
2.61319757e-01 4.02741551e-01 -9.73484591e-02 -6.27877235e-01
-8.22458327e-01 -7.65191078e-01 2.45616019e-01 -9.01415586e-01
2.95301735e-01 -1.06916718e-01 -1.09730661e+00 -8.69827271e-01
-2.73391277e-01 3.70737404e-01 2.43124023e-01 8.27265739e-01
4.86148447e-01 9.28254783e-01 8.88214052e-01 -4.06176805e-01
-4.14523393e-01 -6.80096745e-01 -5.98253131e-01 2.48950347e-01
6.43449783e-01 -3.51127684e-01 -6.94087148e-01 -6.52465045e-01]
|
[6.596451282501221, 5.969913959503174]
|
33b9a1b6-7a8a-4cb9-8f6b-434c9fe9836f
|
discovering-human-object-interaction-concepts
|
2203.14272
| null |
https://arxiv.org/abs/2203.14272v2
|
https://arxiv.org/pdf/2203.14272v2.pdf
|
Discovering Human-Object Interaction Concepts via Self-Compositional Learning
|
A comprehensive understanding of human-object interaction (HOI) requires detecting not only a small portion of predefined HOI concepts (or categories) but also other reasonable HOI concepts, while current approaches usually fail to explore a huge portion of unknown HOI concepts (i.e., unknown but reasonable combinations of verbs and objects). In this paper, 1) we introduce a novel and challenging task for a comprehensive HOI understanding, which is termed as HOI Concept Discovery; and 2) we devise a self-compositional learning framework (or SCL) for HOI concept discovery. Specifically, we maintain an online updated concept confidence matrix during training: 1) we assign pseudo-labels for all composite HOI instances according to the concept confidence matrix for self-training; and 2) we update the concept confidence matrix using the predictions of all composite HOI instances. Therefore, the proposed method enables the learning on both known and unknown HOI concepts. We perform extensive experiments on several popular HOI datasets to demonstrate the effectiveness of the proposed method for HOI concept discovery, object affordance recognition and HOI detection. For example, the proposed self-compositional learning framework significantly improves the performance of 1) HOI concept discovery by over 10% on HICO-DET and over 3% on V-COCO, respectively; 2) object affordance recognition by over 9% mAP on MS-COCO and HICO-DET; and 3) rare-first and non-rare-first unknown HOI detection relatively over 30% and 20%, respectively. Code is publicly available at https://github.com/zhihou7/HOI-CL.
|
['DaCheng Tao', 'Baosheng Yu', 'Zhi Hou']
|
2022-03-27
| null | null | null | null |
['affordance-recognition', 'human-object-interaction-concept-discovery']
|
['computer-vision', 'computer-vision']
|
[ 2.44330242e-01 1.72037550e-03 -1.75009489e-01 -2.60899425e-01
-5.43951929e-01 -4.96061176e-01 4.24968690e-01 3.36608350e-01
-2.53239095e-01 6.17559135e-01 -1.47973597e-01 -8.97474438e-02
-2.80860871e-01 -4.55253601e-01 -8.63779366e-01 -2.99610853e-01
-1.85600817e-01 6.64557815e-01 3.56575191e-01 9.38748792e-02
2.34816298e-01 1.11197099e-01 -2.21677351e+00 1.95615023e-01
1.08227098e+00 1.22869706e+00 4.37549025e-01 5.13117552e-01
2.15029940e-01 4.71231729e-01 -2.76108682e-01 1.42741472e-01
1.20709918e-01 -1.19195618e-01 -7.69867837e-01 1.14796408e-01
3.19137603e-01 -1.74107283e-01 1.01187430e-01 9.87553000e-01
3.52340788e-02 2.53445953e-01 7.96726525e-01 -1.70460820e+00
-4.86988187e-01 6.24442637e-01 -4.42547143e-01 -9.64510515e-02
5.77121735e-01 2.66246378e-01 1.30312824e+00 -1.39436579e+00
6.64988041e-01 1.17468989e+00 3.45750928e-01 5.18037498e-01
-9.65309560e-01 -1.02250588e+00 3.27493846e-01 4.53506768e-01
-1.80478990e+00 -1.51939869e-01 6.45293236e-01 -5.15456200e-01
9.67301488e-01 2.94310451e-01 7.77599335e-01 9.29186344e-01
-2.63469666e-01 1.38476777e+00 1.02077913e+00 -3.65188986e-01
1.82094619e-01 2.04345599e-01 5.91317773e-01 6.31622851e-01
4.89490837e-01 2.19942287e-01 -6.13267481e-01 -5.78241460e-02
4.46463257e-01 -9.89434961e-03 -5.80134466e-02 -3.30872983e-01
-1.54257464e+00 7.04073668e-01 3.49539757e-01 2.95698643e-01
-9.77560803e-02 -3.05089504e-01 3.50304276e-01 -3.03272121e-02
-2.06256583e-01 3.92540097e-01 -5.32742083e-01 -4.01387289e-02
-3.40217501e-01 5.85159540e-01 7.62038052e-01 1.53623903e+00
1.02961814e+00 -3.43398780e-01 6.86302707e-02 7.89958239e-01
2.18985036e-01 4.36394006e-01 5.61870933e-01 -4.12551254e-01
5.05177677e-01 7.54219651e-01 2.91152388e-01 -8.08800042e-01
-5.21428287e-01 -2.84988075e-01 -6.46253049e-01 -3.42609882e-02
2.19259545e-01 2.38368571e-01 -9.37544882e-01 1.86567485e+00
6.21503651e-01 9.60845277e-02 1.16464101e-01 8.12904298e-01
9.62708592e-01 5.08154988e-01 2.55781531e-01 -2.43464977e-01
1.22217679e+00 -8.24301660e-01 -5.52320540e-01 -3.42969626e-01
6.51005447e-01 -6.57274067e-01 1.38586342e+00 3.84061664e-01
-5.03763378e-01 -8.21809471e-01 -1.22485769e+00 -3.17486301e-02
-4.33852971e-01 4.61496800e-01 1.17462242e+00 2.64035881e-01
-2.56143957e-01 3.06129575e-01 -5.99924982e-01 -3.86191845e-01
5.65325081e-01 5.50915658e-01 -5.05459905e-01 -3.11316282e-01
-1.01076365e+00 5.22089541e-01 1.10163844e+00 1.71432912e-01
-1.35291243e+00 -5.90042710e-01 -9.78171349e-01 -1.57710258e-02
1.07541788e+00 -3.43981326e-01 9.28656876e-01 -5.97060978e-01
-1.02950597e+00 6.43300295e-01 -1.51154473e-01 -1.86060697e-01
4.18569952e-01 -5.05704463e-01 -4.11560953e-01 -1.49234131e-01
3.69249612e-01 7.26046801e-01 7.57409155e-01 -1.40347004e+00
-1.21684074e+00 -3.61102223e-01 8.48486498e-02 4.71494973e-01
-1.89869568e-01 -1.70112789e-01 -2.90944070e-01 -4.70612377e-01
5.38622379e-01 -1.15560222e+00 3.82736117e-01 7.97059014e-02
-6.64841652e-01 -7.89839923e-01 9.01924729e-01 -4.03913617e-01
1.17697096e+00 -2.16157436e+00 6.04334027e-02 2.19905868e-01
3.57391179e-01 2.27753147e-01 -1.06508464e-01 3.74948978e-02
-1.47288159e-01 1.17940586e-02 -4.08404887e-01 5.96079007e-02
-5.32358959e-02 2.49084383e-01 8.39271396e-03 1.64968580e-01
2.18586549e-01 7.93886065e-01 -1.03160548e+00 -4.07456964e-01
4.78956223e-01 8.87388811e-02 -6.43877685e-01 4.62095529e-01
-2.63371855e-01 3.87102306e-01 -2.76275158e-01 1.09302235e+00
4.21059310e-01 -2.59430408e-01 4.08753276e-01 -3.65859330e-01
-8.38555098e-02 2.04143152e-01 -1.70775473e+00 1.40802574e+00
-1.98416904e-01 1.20088309e-01 -4.19138014e-01 -1.04519665e+00
8.63150954e-01 2.86138803e-01 5.61218679e-01 -1.39464036e-01
5.21838330e-02 5.42239904e-01 2.76229590e-01 -3.95009279e-01
2.19538152e-01 1.04483053e-01 -1.39406532e-01 2.02293664e-01
3.79899055e-01 5.10792248e-02 4.18277413e-01 1.15078576e-01
7.23826289e-01 1.75593317e-01 8.16002369e-01 -2.50122547e-01
5.53826332e-01 -2.86728404e-02 7.45120108e-01 8.77023101e-01
-3.51587772e-01 4.35856432e-01 3.04651350e-01 -3.28575432e-01
-9.09061015e-01 -1.07829475e+00 -2.66260743e-01 1.06697643e+00
4.48178738e-01 -3.79423678e-01 -4.39272463e-01 -7.32223809e-01
2.62651175e-01 7.77341545e-01 -6.51512802e-01 -1.69717550e-01
-3.36547226e-01 -6.52542353e-01 -1.03819773e-01 7.06940770e-01
4.43266839e-01 -1.33166611e+00 -5.27817369e-01 2.34356478e-01
-3.05995077e-01 -9.70952630e-01 -3.84595186e-01 3.17546457e-01
-6.23002052e-01 -1.33372891e+00 -1.80078283e-01 -8.83699536e-01
5.75216770e-01 4.82859820e-01 6.95271611e-01 3.02001417e-01
-5.66231012e-01 1.28549680e-01 -5.86752653e-01 -6.17606699e-01
-1.08789325e-01 9.69570130e-02 4.08232212e-01 9.82775763e-02
4.79860932e-01 -4.16802019e-01 -3.52261513e-01 6.19915724e-01
-6.77150071e-01 1.65016681e-01 4.28980589e-01 1.05806124e+00
7.48196721e-01 1.63633823e-01 5.42377293e-01 -8.23436201e-01
1.26705840e-01 -5.54919064e-01 -5.75170815e-01 4.19374198e-01
-6.78702235e-01 -8.91620144e-02 1.66870445e-01 -9.38333988e-01
-1.00595748e+00 2.65707582e-01 6.71456903e-02 -3.60864639e-01
-3.68056178e-01 5.61748147e-01 -5.28755128e-01 5.51116467e-02
5.74107766e-01 1.06494747e-01 -3.84980679e-01 -4.23174858e-01
4.10230279e-01 6.66249037e-01 5.80143869e-01 -8.48088861e-01
9.09464002e-01 2.96184570e-01 -3.41978043e-01 -7.97747314e-01
-1.02989841e+00 -7.90209413e-01 -8.70982230e-01 -1.20699890e-01
6.97881341e-01 -9.08363402e-01 -1.09316790e+00 4.40064549e-01
-9.57139730e-01 -7.36903474e-02 -5.11321910e-02 6.61849082e-01
-4.86427516e-01 3.16362351e-01 -2.59245276e-01 -1.05182219e+00
-2.37348273e-01 -1.10660386e+00 1.00403118e+00 3.19191277e-01
-3.24358076e-01 -3.81353945e-01 -4.26266938e-01 4.29820240e-01
-2.29420215e-01 1.47570342e-01 1.02529895e+00 -8.95171583e-01
-6.60581112e-01 -2.02537999e-01 -4.43368644e-01 2.83297747e-01
2.10176811e-01 -2.25415185e-01 -9.73704755e-01 -3.60758364e-01
-2.31492266e-01 -7.82159448e-01 5.24595380e-01 2.64164377e-02
1.17363954e+00 -1.98060557e-01 -5.02605140e-01 3.19227874e-01
1.29124951e+00 4.95993227e-01 4.49709833e-01 7.64865130e-02
9.87721384e-01 7.05278218e-01 1.31312227e+00 6.26854241e-01
4.32393759e-01 6.62425995e-01 3.81253600e-01 3.32703769e-01
-1.43239927e-02 -4.92607296e-01 -6.21525384e-02 9.06233609e-01
-1.84496060e-01 -9.18349624e-03 -1.00535214e+00 7.13215232e-01
-2.08034873e+00 -6.19212568e-01 -8.47477466e-02 2.34040999e+00
8.49040627e-01 2.72616148e-01 1.76576108e-01 3.70756328e-01
6.30481124e-01 -3.93287212e-01 -8.42813075e-01 4.96288806e-01
2.29990724e-02 -3.21818441e-02 5.52362464e-02 4.28028852e-01
-1.35642457e+00 1.11291206e+00 5.12592936e+00 7.94544041e-01
-6.67110503e-01 2.54621655e-01 2.09305435e-01 1.22347459e-01
-2.53807958e-02 1.75655276e-01 -1.08746445e+00 3.25973511e-01
2.54504651e-01 -1.43917978e-01 4.81435865e-01 1.39048409e+00
-3.91074032e-01 -2.18844011e-01 -1.33412850e+00 1.21610725e+00
1.82915553e-01 -9.77750182e-01 2.06872642e-01 -7.23506138e-02
5.42929053e-01 -2.77673125e-01 -2.12106720e-01 6.47582710e-01
3.05942018e-02 -8.05631220e-01 9.72687364e-01 2.91598380e-01
8.60268056e-01 -6.18794918e-01 7.15962708e-01 3.62601429e-01
-1.64153481e+00 -3.96152556e-01 -3.79589647e-01 6.68998957e-02
-2.87399083e-01 1.48497224e-01 -7.37236798e-01 5.29356539e-01
8.74074638e-01 7.32923687e-01 -7.11229205e-01 9.68661726e-01
-3.46172929e-01 4.74160165e-01 -5.84409416e-01 -1.36614531e-01
4.33875248e-02 7.61202127e-02 5.15114605e-01 7.71923423e-01
1.29616290e-01 3.11086029e-01 5.68358839e-01 8.46632957e-01
1.60605654e-01 1.82896331e-01 -3.39429229e-01 -8.72718841e-02
6.46213174e-01 9.74425495e-01 -6.61842287e-01 -4.76226628e-01
-4.88807380e-01 8.69142294e-01 5.37557065e-01 2.15802640e-01
-8.43625069e-01 -4.48543876e-01 5.88323057e-01 -1.16740420e-01
2.80584097e-01 -1.56558797e-01 -1.30203381e-01 -1.36046898e+00
1.50958568e-01 -9.15323377e-01 7.45793641e-01 -7.11426258e-01
-1.23938048e+00 3.95772040e-01 3.56889814e-01 -1.34450436e+00
-1.25093400e-01 -9.46371496e-01 -1.92614034e-01 4.59362358e-01
-1.44262099e+00 -1.34935462e+00 -6.74991906e-01 5.15592158e-01
7.27089942e-01 -8.68136883e-02 7.33551919e-01 3.03100199e-01
-5.10019958e-01 5.74536383e-01 -2.11240247e-01 -6.83021992e-02
4.81686413e-01 -1.04164231e+00 5.94223442e-04 4.85402435e-01
4.11755025e-01 7.73762047e-01 6.12840414e-01 -9.05951798e-01
-1.17351317e+00 -9.21423495e-01 9.82981622e-01 -6.63267195e-01
5.62275290e-01 -5.85748315e-01 -1.18822587e+00 8.53606641e-01
-6.09228969e-01 1.69268832e-01 4.76866692e-01 3.61639559e-01
-4.37967628e-01 -1.83491185e-02 -9.21506763e-01 4.45903480e-01
1.44487822e+00 -4.70283806e-01 -7.51541555e-01 6.31988764e-01
7.58073568e-01 -4.69265312e-01 -7.85879016e-01 7.89353371e-01
8.50051343e-01 -5.97511232e-01 8.78404319e-01 -5.92882574e-01
-1.89084671e-02 -6.89924657e-01 -4.42400664e-01 -6.56735718e-01
-2.57572383e-01 -2.24019364e-01 -3.47820252e-01 9.38182831e-01
4.09736186e-01 -4.82862949e-01 5.17009377e-01 5.07056296e-01
-1.08934581e-01 -7.22938418e-01 -8.40451896e-01 -8.81951571e-01
-4.68121231e-01 -8.16251338e-01 4.91176188e-01 8.29692841e-01
1.51130809e-02 3.69359910e-01 -4.27401006e-01 2.57969767e-01
8.18621874e-01 2.91108102e-01 8.02214682e-01 -1.33668649e+00
-2.59589583e-01 -1.31123841e-01 -4.44682598e-01 -8.76708806e-01
-1.13206720e-02 -8.94453168e-01 2.72362471e-01 -1.07904768e+00
6.98786020e-01 -6.62386358e-01 -4.56822813e-01 8.26464891e-01
-4.86799151e-01 3.35679017e-02 2.17230454e-01 5.50808847e-01
-7.16398299e-01 6.07967377e-01 9.00512874e-01 -1.30258664e-01
-3.03475797e-01 -6.90816790e-02 -4.32035446e-01 7.25388467e-01
5.06359816e-01 -4.38775212e-01 -5.56843042e-01 -6.87244385e-02
-7.17357174e-02 -3.21600795e-01 5.41934013e-01 -1.27834427e+00
-5.71904927e-02 -2.07641065e-01 2.27551416e-01 -9.06407654e-01
1.50251091e-01 -6.45819604e-01 1.69138294e-02 3.70691717e-01
-2.46873528e-01 -5.22724271e-01 1.86769083e-01 6.87836289e-01
-4.41279262e-02 -3.06045890e-01 6.08631313e-01 -1.13035299e-01
-1.26554060e+00 3.62406313e-01 -3.08456346e-02 1.80129744e-02
1.10954034e+00 -1.44395739e-01 3.86839956e-02 -7.44593330e-03
-8.60361993e-01 3.43619287e-01 1.37096599e-01 7.50750244e-01
8.02079856e-01 -1.25613248e+00 -3.49475741e-01 3.77670497e-01
8.96250606e-01 3.80756594e-02 2.28939056e-01 5.27430415e-01
3.83798010e-03 4.86335665e-01 -1.95684999e-01 -7.49100864e-01
-1.23104644e+00 8.26820612e-01 -4.05382738e-02 4.60909083e-02
-7.08042800e-01 8.37659717e-01 6.01831436e-01 -7.69094706e-01
4.10148054e-01 -4.20366973e-01 -2.92898268e-01 -2.21768543e-01
3.95392001e-01 1.65337339e-01 -3.60109687e-01 -7.21339583e-01
-6.07415557e-01 3.83702725e-01 -1.11292027e-01 1.59528792e-01
1.05444920e+00 -1.43216783e-02 3.87288518e-02 7.02595055e-01
1.06033003e+00 -6.71016693e-01 -1.10493910e+00 -2.90513813e-01
2.88680315e-01 -4.88326281e-01 -3.57460320e-01 -8.57796550e-01
-4.68936622e-01 7.67038703e-01 7.20564544e-01 -4.29843038e-01
7.69344866e-01 3.39572966e-01 5.27928054e-01 6.32070422e-01
7.34420955e-01 -1.04349673e+00 2.37381473e-01 3.08977634e-01
1.07943404e+00 -1.54157364e+00 6.65498450e-02 -8.22706997e-01
-7.59301066e-01 8.40098798e-01 9.57583070e-01 1.79318294e-01
6.69300616e-01 -2.86969572e-01 -4.55216259e-01 -2.89169282e-01
-5.61599433e-01 -3.94061923e-01 5.19259155e-01 5.76039970e-01
2.57460803e-01 4.67289537e-01 -4.41447824e-01 7.47416794e-01
-6.09036274e-02 -6.81052282e-02 5.62980063e-02 9.11388218e-01
-4.33426023e-01 -6.47885621e-01 -1.51701704e-01 6.31870747e-01
3.10199529e-01 -3.13366298e-03 1.71393827e-02 1.05285680e+00
4.40298170e-01 9.37642932e-01 -2.42554963e-01 -5.71023047e-01
4.19898450e-01 1.29305929e-01 3.07676047e-01 -9.65237439e-01
-4.59242985e-02 6.63875192e-02 1.61324427e-01 -6.92395270e-01
-2.78181881e-01 -7.75689542e-01 -1.41250074e+00 3.07415605e-01
-7.73878217e-01 5.47878910e-03 5.57870269e-01 1.17562401e+00
9.65085849e-02 2.62491137e-01 3.07527244e-01 -7.19826579e-01
-4.98136878e-01 -1.16733825e+00 -6.88288569e-01 8.16599727e-01
-4.50888369e-03 -1.35639083e+00 -4.44738030e-01 -2.05242168e-02]
|
[9.615500450134277, 1.5424422025680542]
|
b35bed02-b912-4024-aaa1-320f9ae14c58
|
seed-self-supervised-distillation-for-visual-1
|
2101.04731
| null |
https://arxiv.org/abs/2101.04731v2
|
https://arxiv.org/pdf/2101.04731v2.pdf
|
SEED: Self-supervised Distillation For Visual Representation
|
This paper is concerned with self-supervised learning for small models. The problem is motivated by our empirical studies that while the widely used contrastive self-supervised learning method has shown great progress on large model training, it does not work well for small models. To address this problem, we propose a new learning paradigm, named SElf-SupErvised Distillation (SEED), where we leverage a larger network (as Teacher) to transfer its representational knowledge into a smaller architecture (as Student) in a self-supervised fashion. Instead of directly learning from unlabeled data, we train a student encoder to mimic the similarity score distribution inferred by a teacher over a set of instances. We show that SEED dramatically boosts the performance of small networks on downstream tasks. Compared with self-supervised baselines, SEED improves the top-1 accuracy from 42.2% to 67.6% on EfficientNet-B0 and from 36.3% to 68.2% on MobileNet-v3-Large on the ImageNet-1k dataset.
|
['Zicheng Liu', 'Yezhou Yang', 'Lei Zhang', 'Lijuan Wang', 'JianFeng Wang', 'Zhiyuan Fang']
|
2021-01-12
|
seed-self-supervised-distillation-for-visual
|
https://openreview.net/forum?id=AHm3dbp7D1D
|
https://openreview.net/pdf?id=AHm3dbp7D1D
|
iclr-2021-1
|
['unsupervised-pre-training']
|
['methodology']
|
[ 2.94170737e-01 5.76176405e-01 -5.89610279e-01 -7.59825349e-01
-7.90314198e-01 -4.54084426e-01 7.43180990e-01 -5.15105091e-02
-7.78864145e-01 7.59603918e-01 2.37651080e-01 -4.83980030e-01
4.16451633e-01 -6.30152106e-01 -1.06368959e+00 -4.40959543e-01
-5.51330075e-02 7.33314395e-01 3.47677946e-01 -1.33779109e-01
-1.41118690e-01 1.17141463e-01 -1.03253222e+00 3.12412143e-01
8.36005569e-01 8.35593998e-01 2.40185261e-01 4.99408007e-01
8.02768618e-02 1.18190944e+00 -3.59697908e-01 -2.56863505e-01
2.41034955e-01 -4.30784613e-01 -1.11947691e+00 -1.31352479e-02
7.09235966e-01 -4.31620151e-01 -4.43688929e-01 9.48261857e-01
3.38995039e-01 7.78835341e-02 7.94649243e-01 -1.12623680e+00
-5.45978904e-01 1.11207044e+00 -6.37913883e-01 2.72112012e-01
-4.65156853e-01 2.63456315e-01 1.18683589e+00 -8.75394583e-01
6.55821025e-01 1.16161811e+00 5.46266139e-01 6.41524494e-01
-1.49854827e+00 -8.83930147e-01 1.89267501e-01 -3.08395028e-02
-1.17584538e+00 -6.34443223e-01 3.82732511e-01 -1.47090733e-01
1.07114136e+00 -2.93977588e-01 3.13605934e-01 1.17170227e+00
-2.87371129e-01 1.03770828e+00 9.93116677e-01 -3.52415532e-01
1.39440596e-01 1.41791597e-01 3.21383715e-01 7.30603576e-01
1.89800754e-01 3.65576982e-01 -2.12925479e-01 9.79797244e-02
6.74697220e-01 -1.29426599e-01 1.08834263e-02 -4.65947956e-01
-1.25347805e+00 7.88133383e-01 9.08106148e-01 2.44706973e-01
-9.31943133e-02 4.82074708e-01 4.99939322e-01 5.95695019e-01
7.20441282e-01 6.43300414e-01 -8.54377449e-01 -6.97201267e-02
-1.01170206e+00 2.56129797e-03 8.32889736e-01 8.88265908e-01
8.97122204e-01 1.78210869e-01 -4.73721586e-02 8.33566368e-01
2.86270618e-01 4.28891182e-01 7.91860878e-01 -8.79814565e-01
6.18542075e-01 4.67531770e-01 -2.78648943e-01 -2.91654080e-01
-2.68951774e-01 -7.76580095e-01 -8.24316978e-01 1.85662955e-01
4.67396736e-01 -4.28146243e-01 -1.11071324e+00 1.93938780e+00
-4.75750864e-02 3.74670804e-01 2.89742768e-01 4.76774424e-01
7.92598903e-01 6.74839675e-01 2.73148417e-01 1.64460361e-01
8.34867954e-01 -1.77283549e+00 -7.66708376e-03 -6.52512610e-01
1.18160248e+00 -3.89696062e-01 1.03644478e+00 1.40044600e-01
-1.09679377e+00 -7.24559188e-01 -9.22728240e-01 6.31962903e-04
-2.66762763e-01 2.88745249e-03 5.03062129e-01 2.87418276e-01
-1.42205119e+00 6.69557571e-01 -8.48077834e-01 -4.71598357e-01
8.32543314e-01 5.28547287e-01 -4.77663308e-01 -1.45672932e-01
-9.31411266e-01 9.79974985e-01 4.27646101e-01 -4.41427648e-01
-1.33480692e+00 -8.47208917e-01 -8.18280101e-01 2.99009055e-01
4.28645372e-01 -6.08407259e-01 1.72773647e+00 -1.30103004e+00
-1.35748410e+00 9.62655902e-01 -4.17964906e-02 -1.07839227e+00
4.25667375e-01 -2.27463469e-01 -1.54729202e-01 8.37608427e-02
1.52905002e-01 1.18199122e+00 7.40753412e-01 -1.22014678e+00
-6.68027163e-01 -1.28873572e-01 8.93583596e-02 2.34677240e-01
-4.29382533e-01 -1.34458318e-01 -4.51858312e-01 -4.59908426e-01
-1.45366177e-01 -1.10661435e+00 -5.49581468e-01 -2.10402712e-01
-4.62625593e-01 -2.93619066e-01 7.55926788e-01 -1.69638991e-01
9.03131008e-01 -2.13354063e+00 -5.05390018e-02 1.68762729e-01
5.18637896e-01 6.04587197e-01 -5.46066940e-01 1.69355839e-01
-3.27494413e-01 1.11148439e-01 -3.64770532e-01 -5.59992373e-01
-2.98961163e-01 2.97042489e-01 -3.23856801e-01 2.15442643e-01
3.27596366e-01 1.29151082e+00 -1.19319546e+00 -3.46892297e-01
-7.19063580e-02 2.24186972e-01 -8.41903329e-01 3.48936319e-01
-4.36320662e-01 2.30470881e-01 -1.52313814e-01 3.55253309e-01
4.12890911e-01 -6.76280320e-01 2.01512843e-01 1.81925762e-02
2.87854671e-01 6.11050487e-01 -5.48082948e-01 1.91410100e+00
-6.35359645e-01 7.73548424e-01 -9.56805050e-02 -1.38798892e+00
7.04791486e-01 1.71705455e-01 2.74656534e-01 -5.68201244e-01
2.23570056e-02 1.00417867e-01 4.18961972e-01 -2.25416407e-01
1.07784532e-01 -9.84498709e-02 2.15889737e-01 6.19052827e-01
5.25465429e-01 8.12259167e-02 1.00858800e-01 5.62575281e-01
1.40061963e+00 1.52008357e-02 1.18907295e-01 -4.22172308e-01
2.44099408e-01 7.92557523e-02 3.07356417e-01 7.73414075e-01
-1.95203215e-01 4.86575216e-01 5.37722230e-01 -3.51491630e-01
-9.75245535e-01 -1.13605833e+00 1.22383542e-01 1.44879901e+00
-1.01801887e-01 -5.73016286e-01 -7.43081450e-01 -1.24751401e+00
2.91577518e-01 5.69341600e-01 -6.56578124e-01 -4.84313667e-01
-6.32914305e-01 -5.27365744e-01 6.05861068e-01 9.08927441e-01
7.17090905e-01 -9.35897350e-01 -5.25306985e-02 2.10398421e-01
1.48185462e-01 -1.14638388e+00 -4.13278162e-01 6.03849173e-01
-1.20996797e+00 -9.05633271e-01 -6.44532263e-01 -1.05259597e+00
8.67992938e-01 4.34364647e-01 1.49604487e+00 6.48069680e-02
2.08519652e-01 1.09066954e-02 -2.96055973e-01 -2.92393863e-01
-5.87784111e-01 7.06380427e-01 2.56609004e-02 -1.42684028e-01
3.35557699e-01 -7.02393889e-01 -5.07970452e-01 3.26341271e-01
-7.44721711e-01 1.00205354e-01 9.24074173e-01 1.03101623e+00
3.97611260e-01 -2.32890919e-01 9.06456232e-01 -1.42138898e+00
3.08839440e-01 -6.94596708e-01 -4.44209009e-01 4.72596521e-03
-8.85953069e-01 2.55291373e-01 8.78798425e-01 -5.38049519e-01
-9.24587965e-01 1.89970046e-01 -1.10392220e-01 -3.97512466e-01
-1.98744103e-01 4.50434297e-01 1.68187954e-02 -1.14146182e-02
1.03271139e+00 3.15923899e-01 2.69285589e-01 -5.47425270e-01
6.45774364e-01 7.94225156e-01 5.63531756e-01 -4.49642599e-01
1.05319536e+00 3.73403817e-01 -2.87646681e-01 -5.09471416e-01
-1.28370655e+00 -4.10023838e-01 -6.78965330e-01 3.68548065e-01
6.92735136e-01 -1.37959445e+00 -2.60826021e-01 2.85972744e-01
-7.31753170e-01 -1.16885459e+00 -5.21864951e-01 3.29083800e-01
-4.89516824e-01 2.57696919e-02 -6.91805780e-01 -1.41123354e-01
-3.59745800e-01 -1.09621787e+00 7.62712955e-01 -4.54086922e-02
-1.95843190e-01 -1.15521491e+00 4.73941537e-03 3.38972479e-01
6.78805947e-01 -3.07341635e-01 7.38662064e-01 -1.10131955e+00
-3.72273833e-01 -1.45323589e-01 -3.68561357e-01 8.13922048e-01
7.26571903e-02 -4.38407421e-01 -1.00880718e+00 -3.91190052e-01
-3.67405474e-01 -9.85352695e-01 1.34386110e+00 1.46884039e-01
1.29980755e+00 -1.71633095e-01 -5.27125180e-01 7.12002158e-01
1.17749202e+00 -1.97169155e-01 4.84184235e-01 1.81674898e-01
7.44548917e-01 1.98608965e-01 2.57352084e-01 -1.21623956e-01
3.92090142e-01 3.39709640e-01 5.12388229e-01 -3.18594366e-01
-4.32219207e-01 -5.72623789e-01 3.19221705e-01 6.39063179e-01
1.48576200e-01 -1.09584287e-01 -9.89492118e-01 5.36079943e-01
-1.76019204e+00 -7.35795021e-01 2.65245259e-01 1.97238755e+00
1.11971319e+00 4.94389027e-01 2.54555911e-01 -1.72433376e-01
4.74118412e-01 2.95026213e-01 -7.16385245e-01 -1.98706128e-02
3.91512960e-02 2.72914588e-01 6.58160031e-01 5.38636923e-01
-1.18373871e+00 1.44934213e+00 6.49395561e+00 8.29549372e-01
-1.14348471e+00 2.87920654e-01 9.46067572e-01 -2.03854963e-01
-9.31201652e-02 1.27743647e-01 -9.20033813e-01 3.77696902e-01
1.29512668e+00 -1.07834712e-01 2.17836529e-01 1.11316800e+00
-2.15390343e-02 5.30942678e-02 -1.43006670e+00 8.07269335e-01
-4.82208468e-02 -1.43219054e+00 9.31054950e-02 5.50697930e-02
1.05888081e+00 8.54746461e-01 1.06720679e-01 7.74501681e-01
8.46913576e-01 -1.24839830e+00 3.45966786e-01 -3.62114646e-02
9.19377923e-01 -5.13791203e-01 5.59087455e-01 6.42959952e-01
-7.81247258e-01 -1.05099022e-01 -5.32241166e-01 -2.11281180e-01
6.26284033e-02 4.87250209e-01 -1.13303554e+00 -1.56225666e-01
3.13242257e-01 9.78155136e-01 -6.90904021e-01 1.01219606e+00
-5.33995152e-01 1.37591839e+00 -3.49822968e-01 1.91933781e-01
6.21407628e-01 1.90170214e-01 5.14266118e-02 1.27265334e+00
-2.67865896e-01 -9.35548469e-02 2.52499521e-01 7.42911398e-01
-8.39294910e-01 2.30112746e-02 -7.40394473e-01 -1.11441791e-01
3.84425372e-01 1.17030656e+00 -4.93340284e-01 -8.60331476e-01
-6.05975628e-01 9.75096405e-01 9.57008958e-01 4.06719267e-01
-5.60537934e-01 -2.01459229e-01 4.59796786e-01 2.39287063e-01
3.49559873e-01 -7.77577981e-02 -1.66278824e-01 -1.25321376e+00
-3.57920170e-01 -8.71174335e-01 2.89545685e-01 -7.41644561e-01
-1.26762962e+00 6.64634407e-01 -1.18526712e-01 -9.84404445e-01
-4.26774591e-01 -5.43575823e-01 -7.71941006e-01 6.84609294e-01
-1.84493291e+00 -1.11226249e+00 -5.32609560e-02 4.52523470e-01
6.73468232e-01 -4.15894121e-01 8.61319721e-01 1.72176525e-01
-4.62020695e-01 8.26921701e-01 3.46927404e-01 5.03501475e-01
9.82885063e-01 -1.42487848e+00 8.54388535e-01 5.97407937e-01
4.82255280e-01 4.80857790e-01 2.61022925e-01 -4.13711995e-01
-8.92056942e-01 -1.32558644e+00 9.80866194e-01 -4.83025342e-01
8.80549550e-01 -3.87090176e-01 -7.75687695e-01 1.15204751e+00
2.48895660e-01 6.61938488e-01 5.90496004e-01 2.99005777e-01
-8.62881839e-01 -2.92780876e-01 -8.96160781e-01 5.39117694e-01
1.28402030e+00 -4.36879784e-01 -6.19527280e-01 4.60760117e-01
1.00693989e+00 -2.01691613e-01 -7.44186997e-01 4.82767433e-01
1.83263555e-01 -6.86309516e-01 9.42965806e-01 -1.13140273e+00
7.96885550e-01 2.02063397e-01 -4.60170805e-02 -1.65232992e+00
-3.39656174e-01 -5.26039958e-01 -3.46277714e-01 8.35267544e-01
8.21402311e-01 -6.30819678e-01 1.11330104e+00 2.64236093e-01
-1.94848403e-01 -9.66800630e-01 -5.41299701e-01 -9.43322480e-01
5.74636340e-01 -2.43029699e-01 2.89528370e-01 9.50882077e-01
5.31874076e-02 8.94550145e-01 -1.82048440e-01 -1.76104829e-01
6.46637261e-01 -3.71499211e-02 9.35530722e-01 -1.13222444e+00
-5.73558390e-01 -3.02544385e-01 -3.44211996e-01 -1.54363632e+00
5.25255203e-01 -1.28902602e+00 -9.94860875e-06 -1.30150425e+00
3.40415239e-01 -7.04012990e-01 -5.47106206e-01 6.73713624e-01
-2.43772894e-01 4.13313746e-01 1.33365542e-01 3.60752255e-01
-9.59455311e-01 4.11632955e-01 1.23113608e+00 -1.39910549e-01
-3.21657956e-02 1.31810442e-01 -1.04560959e+00 7.70794928e-01
9.64783192e-01 -5.37182033e-01 -7.61360466e-01 -7.98356891e-01
-1.42319486e-01 -2.26528764e-01 1.52511686e-01 -9.68501687e-01
2.14605659e-01 1.89430654e-01 2.04219222e-01 -2.77831823e-01
4.23426069e-02 -6.66771173e-01 -6.78971827e-01 5.80135107e-01
-8.83432388e-01 -1.07319318e-01 1.23909004e-02 6.29749477e-01
-2.48709217e-01 -1.68460727e-01 9.37027931e-01 -2.37918749e-01
-7.87389576e-01 4.47662294e-01 -1.44545332e-01 3.98826540e-01
8.09028685e-01 2.66304523e-01 -4.82507795e-01 -4.74260062e-01
-8.58834803e-01 4.29124087e-01 3.42673630e-01 2.61532843e-01
2.99800485e-01 -1.20074391e+00 -7.33088434e-01 2.64984369e-01
2.72440612e-01 3.06557506e-01 -2.29819089e-01 6.07376516e-01
-2.76919097e-01 4.20307010e-01 9.22253430e-02 -5.69926262e-01
-9.30687428e-01 3.78119737e-01 4.48039621e-01 -5.05178630e-01
-4.67448533e-01 1.23877537e+00 3.10880393e-01 -7.11538613e-01
4.73620147e-01 -3.41910630e-01 -3.14884074e-02 -3.22622359e-01
4.71839070e-01 6.75897375e-02 -5.19413799e-02 -5.37553132e-01
-3.73357348e-02 3.99918295e-02 -5.48307717e-01 -4.18208875e-02
1.47845614e+00 1.51991099e-01 1.95953399e-01 2.75694579e-01
1.61943710e+00 -3.77280563e-01 -1.50383770e+00 -8.03384066e-01
1.23484321e-01 1.37440152e-02 1.20406292e-01 -9.06055748e-01
-1.20293581e+00 9.43127692e-01 2.94749856e-01 -1.39531031e-01
7.08420455e-01 2.34936461e-01 8.06515217e-01 8.25508237e-01
1.73654675e-01 -8.99239480e-01 4.06051993e-01 8.59573662e-01
3.70977640e-01 -1.74094796e+00 -3.21414888e-01 -2.37152636e-01
-5.39509535e-01 6.14211559e-01 8.70086908e-01 -4.03542548e-01
8.47602606e-01 4.17263955e-01 2.15745285e-01 -7.74725601e-02
-1.07805300e+00 -3.34361374e-01 9.33086500e-02 5.27565241e-01
6.49567544e-01 -5.76421283e-02 1.91981390e-01 4.29356009e-01
-2.67313153e-01 1.41374320e-01 2.92387068e-01 8.17490995e-01
-6.51020646e-01 -1.12198555e+00 2.34817266e-01 7.61722565e-01
-3.95753980e-01 -5.04529536e-01 -1.77694947e-01 7.23477900e-01
-5.02494574e-02 8.66440773e-01 2.85545439e-01 -4.77269977e-01
4.05741185e-02 1.42032057e-01 3.22426885e-01 -1.22785735e+00
-6.32963717e-01 -8.04649442e-02 2.25803494e-01 -4.64552194e-01
-3.68643761e-01 -2.30862334e-01 -1.24672663e+00 -1.19674429e-01
-2.45306313e-01 2.82600045e-01 4.93549109e-01 9.72061276e-01
3.29466552e-01 3.80245835e-01 6.04559958e-01 -7.64850318e-01
-1.15524995e+00 -1.13243091e+00 -4.37957942e-01 4.25916702e-01
3.04858565e-01 -2.89512604e-01 -4.29358095e-01 8.02933946e-02]
|
[9.46528434753418, 2.7484960556030273]
|
7897b1bc-6ca2-4470-ac5c-fc41acfa27a7
|
retain-an-interpretable-predictive-model-for
|
1608.05745
| null |
http://arxiv.org/abs/1608.05745v4
|
http://arxiv.org/pdf/1608.05745v4.pdf
|
RETAIN: An Interpretable Predictive Model for Healthcare using Reverse Time Attention Mechanism
|
Accuracy and interpretability are two dominant features of successful
predictive models. Typically, a choice must be made in favor of complex black
box models such as recurrent neural networks (RNN) for accuracy versus less
accurate but more interpretable traditional models such as logistic regression.
This tradeoff poses challenges in medicine where both accuracy and
interpretability are important. We addressed this challenge by developing the
REverse Time AttentIoN model (RETAIN) for application to Electronic Health
Records (EHR) data. RETAIN achieves high accuracy while remaining clinically
interpretable and is based on a two-level neural attention model that detects
influential past visits and significant clinical variables within those visits
(e.g. key diagnoses). RETAIN mimics physician practice by attending the EHR
data in a reverse time order so that recent clinical visits are likely to
receive higher attention. RETAIN was tested on a large health system EHR
dataset with 14 million visits completed by 263K patients over an 8 year period
and demonstrated predictive accuracy and computational scalability comparable
to state-of-the-art methods such as RNN, and ease of interpretability
comparable to traditional models.
|
['Walter F. Stewart', 'Joshua A. Kulas', 'Mohammad Taha Bahadori', 'Jimeng Sun', 'Edward Choi', 'Andy Schuetz']
|
2016-08-19
|
retain-an-interpretable-predictive-model-for-1
|
http://papers.nips.cc/paper/6321-retain-an-interpretable-predictive-model-for-healthcare-using-reverse-time-attention-mechanism
|
http://papers.nips.cc/paper/6321-retain-an-interpretable-predictive-model-for-healthcare-using-reverse-time-attention-mechanism.pdf
|
neurips-2016-12
|
['disease-trajectory-forecasting']
|
['medical']
|
[ 3.70138437e-01 5.65663338e-01 -4.71655101e-01 -7.40507782e-01
-7.06642866e-01 -1.16119079e-01 -2.98527945e-02 5.73045254e-01
-2.72498578e-01 5.64076126e-01 8.72020781e-01 -9.20417309e-01
-6.52032256e-01 -4.92622197e-01 -5.54299355e-01 -9.20595080e-02
-2.19398990e-01 9.63411093e-01 -9.32909548e-01 5.79100437e-02
-1.46492675e-01 2.79835284e-01 -9.05783176e-01 3.72891188e-01
9.30294037e-01 6.68533206e-01 -3.86297494e-01 8.46775174e-01
3.69636595e-01 1.43302655e+00 -4.32074904e-01 -2.37863988e-01
1.22636825e-01 -4.53845561e-01 -8.46504092e-01 -4.16428983e-01
7.64270574e-02 -5.80630541e-01 -3.82407337e-01 4.28976029e-01
4.46086198e-01 -2.87215728e-02 6.23299003e-01 -7.17167318e-01
-1.25088799e+00 8.80442977e-01 -1.50229514e-01 5.37831426e-01
2.55249172e-01 2.21419796e-01 1.17122948e+00 -4.80957955e-01
5.68405271e-01 1.05447078e+00 1.26729906e+00 3.97396535e-01
-1.54544854e+00 -2.03855246e-01 2.09914565e-01 -9.93478820e-02
-9.25241172e-01 -3.96335125e-01 4.78923284e-02 -4.58737642e-01
1.30399811e+00 5.99610984e-01 7.76173770e-01 1.14017844e+00
6.29065037e-01 6.28770530e-01 6.21448934e-01 -3.69363874e-02
1.84384987e-01 -4.35819291e-02 6.92790806e-01 3.33282739e-01
4.90743071e-01 3.54047604e-02 -7.24014193e-02 -7.13044763e-01
9.04773653e-01 1.00706327e+00 -2.46448532e-01 5.08918047e-01
-1.21774983e+00 9.19455171e-01 7.80643582e-01 1.93797573e-01
-1.23196566e+00 2.36775264e-01 3.76476705e-01 1.95227325e-01
4.18977052e-01 1.01160264e+00 -6.80562735e-01 -3.50263238e-01
-9.72164571e-01 6.84857965e-02 8.67759585e-01 5.66567719e-01
-7.06460774e-02 -4.81569842e-02 -6.22833788e-01 5.93666315e-01
1.36655718e-01 4.81015891e-01 4.55328763e-01 -8.62343729e-01
1.55641243e-01 8.15474689e-01 -5.59366271e-02 -1.07624352e+00
-7.42284656e-01 -6.95633471e-01 -1.24776983e+00 -4.53242987e-01
2.32574460e-03 -2.98080951e-01 -1.14970183e+00 1.45205712e+00
-1.14616752e-01 -9.69114713e-03 8.82618800e-02 6.87475681e-01
8.02923083e-01 5.34091115e-01 4.49014187e-01 -1.77961707e-01
1.54103911e+00 -6.40815496e-01 -9.35668647e-01 -1.24813557e-01
7.58062840e-01 -3.66228402e-01 8.09676647e-01 1.55406177e-01
-1.17109048e+00 -2.18691736e-01 -4.58911598e-01 -3.10947508e-01
1.72560345e-02 6.63943142e-02 5.94375134e-01 2.07881764e-01
-1.09633064e+00 7.19880879e-01 -1.24021435e+00 -4.34503913e-01
6.62924826e-01 6.13951027e-01 1.11490451e-01 3.91227528e-02
-1.02285063e+00 8.33799839e-01 -1.97608359e-02 2.39795357e-01
-4.51244026e-01 -1.06186402e+00 -7.31723964e-01 4.11324859e-01
-1.38834323e-04 -1.31357467e+00 1.55888033e+00 -1.09721100e+00
-9.01385307e-01 4.35639948e-01 -4.88696724e-01 -9.06155646e-01
4.94314760e-01 -6.57680511e-01 -4.70380008e-01 -2.07385212e-01
1.33436814e-01 2.63548285e-01 3.86041522e-01 -6.61256731e-01
-4.71365094e-01 -5.13471425e-01 -3.97577018e-01 5.11176474e-02
-1.51687443e-01 -2.06721537e-02 -1.09621353e-01 -6.36794388e-01
-8.76711458e-02 -1.06483126e+00 -7.89959610e-01 -2.95136750e-01
-6.07547939e-01 -8.03047419e-02 3.54362249e-01 -1.22326887e+00
1.75429440e+00 -1.80050981e+00 -5.87203130e-02 3.19492668e-01
9.73901212e-01 1.66903779e-01 1.25954434e-01 2.37570524e-01
-3.81041080e-01 4.63156074e-01 2.59055197e-02 -2.29674980e-01
-4.58046079e-01 1.95357144e-01 -1.65047988e-01 2.00166225e-01
6.41073644e-01 1.45820665e+00 -9.70435977e-01 -1.37291312e-01
7.25792572e-02 8.46565604e-01 -7.98807085e-01 2.99702555e-01
3.08719799e-02 3.43698800e-01 -8.32301676e-01 9.05558944e-01
-7.64355287e-02 -1.37639844e+00 3.13775629e-01 -9.86876083e-04
1.49510011e-01 5.82906365e-01 -4.07391042e-01 8.26342344e-01
-3.35742146e-01 5.06669223e-01 -2.49588221e-01 -5.49785078e-01
6.09436691e-01 5.24561882e-01 7.19714165e-01 -4.11944032e-01
1.91216573e-01 3.73308472e-02 3.32269758e-01 -6.30786896e-01
3.63664836e-01 1.69209927e-01 2.01927066e-01 4.71810818e-01
-6.77227974e-01 5.36665201e-01 -4.84884948e-01 1.86952159e-01
1.71640038e+00 -2.05391839e-01 8.16005528e-01 -1.09625347e-01
-2.58821666e-01 2.82192558e-01 6.97796047e-01 1.09326220e+00
-1.93142876e-01 6.20575070e-01 5.04791915e-01 -1.07934630e+00
-1.23178124e+00 -7.04911411e-01 -3.27574193e-01 8.65356147e-01
-6.17725968e-01 -3.42655808e-01 -3.29174846e-01 -3.85452777e-01
2.44476095e-01 7.83851504e-01 -1.02930856e+00 -2.20842376e-01
-6.10615253e-01 -8.33752811e-01 4.64195102e-01 9.35989499e-01
-4.05749232e-02 -1.21695209e+00 -8.86883438e-01 7.73553669e-01
-1.34607643e-01 -6.87929869e-01 -3.83678764e-01 3.67617995e-01
-1.14297378e+00 -1.10648346e+00 -6.40606582e-01 -3.85792673e-01
7.91586876e-01 -1.93657786e-01 1.67191911e+00 1.34834558e-01
-4.74077255e-01 3.16913098e-01 -1.01501308e-01 -7.31288671e-01
-4.49112087e-01 3.11178118e-01 1.35162368e-01 -3.23364109e-01
8.30071807e-01 -2.89982975e-01 -1.04287529e+00 -1.51516452e-01
-8.00566316e-01 5.49055897e-02 8.97883356e-01 1.23060238e+00
6.96296573e-01 -5.02430022e-01 7.50194192e-01 -1.45015919e+00
9.34065998e-01 -8.89515877e-01 -1.02524966e-01 2.17921048e-01
-1.07635081e+00 8.60130712e-02 8.27644587e-01 -2.76386440e-01
-6.35663450e-01 1.18217608e-02 -1.68630123e-01 -4.19075131e-01
6.42589554e-02 8.62489402e-01 8.40877295e-01 5.63818872e-01
7.72837281e-01 -5.39388098e-02 1.93928733e-01 -3.24292630e-01
-1.18866809e-01 8.50896478e-01 3.19206625e-01 -1.13160228e-02
8.70426968e-02 1.32905334e-01 -1.72071964e-01 -5.40316403e-01
-9.22836065e-01 -2.83728331e-01 -1.83613688e-01 2.84784377e-01
8.15979838e-01 -9.30459499e-01 -1.11724925e+00 -2.74605930e-01
-8.97482395e-01 -4.17556353e-02 -4.96844232e-01 5.60505569e-01
-1.56903908e-01 -2.16181561e-01 -1.07093775e+00 -9.20586467e-01
-1.00508022e+00 -1.06447279e+00 1.02178419e+00 7.80515894e-02
-1.23814833e+00 -1.13089871e+00 4.19095904e-02 2.70673722e-01
6.61836147e-01 5.16517937e-01 1.26380801e+00 -1.06899881e+00
-4.22262162e-01 -3.33034396e-01 -3.27053159e-01 -3.01115345e-02
4.65049475e-01 9.87667739e-02 -7.26567984e-01 -1.22762486e-01
-7.84404501e-02 6.92894161e-02 7.22621799e-01 1.18488073e+00
1.36435401e+00 -8.48860264e-01 -5.86253822e-01 7.07418263e-01
1.21887851e+00 4.83508348e-01 5.04493475e-01 7.55316317e-02
7.14690268e-01 1.84724182e-01 9.71944779e-02 5.32381296e-01
5.88397622e-01 1.80302188e-01 8.56468156e-02 -7.03653336e-01
5.47718823e-01 -2.62726843e-01 -1.64205462e-01 8.20691228e-01
-2.82568723e-01 -2.17616037e-01 -1.31278348e+00 5.49840152e-01
-2.11153030e+00 -8.15274000e-01 -1.91779301e-01 2.11083174e+00
7.31595457e-01 -9.75472555e-02 1.15589976e-01 -3.20272624e-01
4.42611247e-01 -3.04232657e-01 -1.02503622e+00 -5.73247135e-01
7.91554600e-02 6.16849773e-02 5.33167243e-01 3.15878302e-01
-8.57398510e-01 2.42698073e-01 7.72698784e+00 -6.87710345e-02
-8.51861179e-01 -7.96367377e-02 1.72671068e+00 -4.88491207e-01
-5.26695907e-01 -5.09611845e-01 -5.99441588e-01 4.31052297e-01
1.53620422e+00 5.05532138e-02 4.16513056e-01 9.92030501e-01
7.41912961e-01 4.29293782e-01 -1.42424524e+00 9.52717602e-01
-2.03985959e-01 -1.53021312e+00 3.20902281e-02 2.55651683e-01
7.26621270e-01 2.92755306e-01 2.90548295e-01 3.42194468e-01
7.58321106e-01 -1.70726216e+00 -1.67954952e-01 9.45164979e-01
8.62419009e-01 -3.88718992e-01 1.02951908e+00 -5.63611984e-02
-6.43919468e-01 -4.18466151e-01 -8.53444042e-04 -7.43964911e-02
1.88878775e-01 6.85601175e-01 -1.23835969e+00 1.60039678e-01
9.20470178e-01 1.10986173e+00 -2.56617993e-01 6.33805335e-01
3.84223282e-01 1.02287292e+00 -3.25487494e-01 -1.02299109e-01
3.75559896e-01 -7.93448389e-02 3.11562091e-01 1.12332821e+00
2.13186935e-01 5.08409977e-01 -1.33306652e-01 1.00817692e+00
-2.35475421e-01 9.73930024e-03 -8.65478575e-01 -4.15938348e-01
3.43179971e-01 9.36753929e-01 -3.46465021e-01 -6.22172415e-01
-2.90639579e-01 6.97191954e-01 3.68169755e-01 4.99600321e-01
-7.88369119e-01 6.10736273e-02 7.80697703e-01 4.17299390e-01
-2.68545151e-02 4.51341331e-01 -6.45709276e-01 -8.54616582e-01
-8.55873376e-02 -1.14970362e+00 6.61016881e-01 -6.09670877e-01
-1.41552901e+00 9.50278997e-01 -6.42523229e-01 -8.94505501e-01
-6.09689534e-01 -3.81779253e-01 -1.90223619e-01 1.01530433e+00
-1.31071663e+00 -8.15071642e-01 -2.79347748e-01 2.14898944e-01
4.60438102e-01 3.45378183e-02 1.14303935e+00 2.52812386e-01
-7.28568494e-01 6.14529252e-01 3.76013488e-01 1.88803256e-01
4.23997730e-01 -1.13754094e+00 7.51250684e-01 2.96092629e-01
-4.77466285e-01 1.25248909e+00 4.52305675e-01 -9.62011576e-01
-1.30603313e+00 -1.36931980e+00 1.28311050e+00 -9.24245119e-01
2.06934437e-01 3.26187491e-01 -1.05131638e+00 1.29220963e+00
-1.78660005e-01 -2.97337383e-01 1.03271890e+00 5.20249188e-01
7.84376040e-02 8.90912563e-02 -1.03242350e+00 6.75970554e-01
9.05339003e-01 -2.92235374e-01 -6.00688875e-01 4.20039594e-01
9.43287253e-01 -2.88476527e-01 -1.23426807e+00 5.78741372e-01
7.11091518e-01 -4.84843016e-01 8.76468778e-01 -1.30471933e+00
8.77380252e-01 1.50001034e-01 2.70852506e-01 -1.20413053e+00
-9.88624036e-01 -7.95863926e-01 -4.38502580e-01 4.83698726e-01
7.17397571e-01 -9.26404417e-01 5.53971410e-01 1.30924273e+00
6.42126501e-02 -1.28390634e+00 -4.78912771e-01 -6.55714199e-02
-5.00066102e-01 -1.00442559e-01 7.49813855e-01 1.12543654e+00
-1.73343986e-01 4.95747149e-01 -5.38826644e-01 8.68154764e-02
1.33855060e-01 -1.51433036e-01 2.68189102e-01 -1.39162576e+00
-4.05921072e-01 -4.13768321e-01 -2.78080940e-01 -8.60414445e-01
-4.45053250e-01 -6.64609134e-01 -3.61476779e-01 -1.82806504e+00
4.28077817e-01 -5.05933762e-01 -6.10337079e-01 8.85279059e-01
-5.26089013e-01 -1.58715189e-01 -2.98491657e-01 6.49904251e-01
-4.58216012e-01 2.16163933e-01 8.49249005e-01 -1.17414661e-01
-7.44564772e-01 1.46955013e-01 -1.29302204e+00 4.72746104e-01
5.70034087e-01 -4.64504749e-01 -1.24752790e-01 -5.74409604e-01
2.65549839e-01 4.73552585e-01 2.54678845e-01 -5.37842512e-01
1.51327103e-01 -6.52600601e-02 8.30131829e-01 -4.66037929e-01
1.34978712e-01 -9.06853199e-01 5.36231279e-01 9.29825246e-01
-9.07052040e-01 7.96811521e-01 1.76285341e-01 6.45400643e-01
-2.69233733e-02 4.86196458e-01 3.28991622e-01 -1.43634737e-01
1.34581467e-02 3.42728913e-01 -4.15157765e-01 -9.86996293e-02
7.15924323e-01 -2.03728348e-01 -7.93146342e-02 -6.88206911e-01
-8.30128729e-01 4.09105003e-01 3.19032222e-02 3.03997189e-01
6.82144701e-01 -9.99153078e-01 -1.05555475e+00 2.13265598e-01
7.67508596e-02 -1.05004534e-02 1.92970932e-01 9.19178724e-01
-7.29090095e-01 7.92647421e-01 2.74220616e-01 -6.68320417e-01
-1.20816207e+00 4.66449112e-01 4.45660710e-01 -6.55430794e-01
-9.13054407e-01 3.35732073e-01 -1.48116611e-02 -4.17586148e-01
1.30686015e-01 -8.39406371e-01 -7.46549591e-02 -3.19135576e-01
7.78843403e-01 3.91556233e-01 -3.12262457e-02 -1.26944989e-01
-3.09581339e-01 5.30451685e-02 -4.96130586e-01 5.21254301e-01
1.78469753e+00 9.81856287e-02 -8.03385898e-02 4.58263129e-01
9.28488851e-01 -5.28265476e-01 -7.97012091e-01 -2.48555377e-01
-8.00695121e-02 -1.63790345e-01 1.15268216e-01 -1.11638367e+00
-7.20876753e-01 3.11351299e-01 5.33954024e-01 3.18738580e-01
1.07189560e+00 -1.33812025e-01 7.69743621e-01 4.96960968e-01
-3.06562841e-01 -7.66407609e-01 -4.61372405e-01 3.34719777e-01
5.69184601e-01 -1.40050507e+00 1.49218999e-02 3.76148224e-02
-7.50072777e-01 9.14612949e-01 3.23257506e-01 -7.01663867e-02
6.29436135e-01 1.03672378e-01 2.73494184e-01 -5.09114623e-01
-1.14197147e+00 4.91708785e-01 4.05722529e-01 1.85859352e-01
9.94320929e-01 4.75490302e-01 1.24171348e-02 5.99348664e-01
-9.40131769e-02 5.35893023e-01 2.27042958e-01 7.96723127e-01
2.79769987e-01 -4.52491015e-01 -2.94651866e-01 1.46995497e+00
-8.80616844e-01 -5.64149082e-01 -2.37817615e-01 6.00062013e-01
-4.88809824e-01 8.25853407e-01 2.78243959e-01 -2.56735027e-01
3.18170369e-01 4.39162441e-02 -3.84380072e-01 -5.25332391e-01
-1.24978983e+00 -3.18855271e-02 1.69338629e-01 -6.18601978e-01
1.22896284e-02 -5.89942336e-01 -1.06185031e+00 -5.15692115e-01
2.76165586e-02 7.60517456e-03 2.39096597e-01 6.06897235e-01
1.22953773e+00 1.18420982e+00 3.50620776e-01 -1.61405087e-01
-8.80613685e-01 -8.89287293e-01 -2.47269034e-01 6.23932540e-01
9.12464201e-01 9.51312557e-02 -1.11890994e-01 2.50435732e-02]
|
[7.971344470977783, 6.274102687835693]
|
86f26a85-bd00-48c7-8631-70fe5117d1fd
|
caco-both-positive-and-negative-samples-are
|
2203.14370
| null |
https://arxiv.org/abs/2203.14370v1
|
https://arxiv.org/pdf/2203.14370v1.pdf
|
CaCo: Both Positive and Negative Samples are Directly Learnable via Cooperative-adversarial Contrastive Learning
|
As a representative self-supervised method, contrastive learning has achieved great successes in unsupervised training of representations. It trains an encoder by distinguishing positive samples from negative ones given query anchors. These positive and negative samples play critical roles in defining the objective to learn the discriminative encoder, avoiding it from learning trivial features. While existing methods heuristically choose these samples, we present a principled method where both positive and negative samples are directly learnable end-to-end with the encoder. We show that the positive and negative samples can be cooperatively and adversarially learned by minimizing and maximizing the contrastive loss, respectively. This yields cooperative positives and adversarial negatives with respect to the encoder, which are updated to continuously track the learned representation of the query anchors over mini-batches. The proposed method achieves 71.3% and 75.3% in top-1 accuracy respectively over 200 and 800 epochs of pre-training ResNet-50 backbone on ImageNet1K without tricks such as multi-crop or stronger augmentations. With Multi-Crop, it can be further boosted into 75.7%. The source code and pre-trained model are released in https://github.com/maple-research-lab/caco.
|
['Guo-Jun Qi', 'Dan Zeng', 'Yuhang Huang', 'Xiao Wang']
|
2022-03-27
| null | null | null | null |
['self-supervised-image-classification']
|
['computer-vision']
|
[ 1.45804822e-01 2.91253865e-01 -4.74020481e-01 -6.50804102e-01
-9.98361766e-01 -6.42070591e-01 5.50340891e-01 -1.43285498e-01
-7.00709641e-01 7.54920244e-01 -5.58498278e-02 9.47259590e-02
3.73232454e-01 -6.55161381e-01 -1.14881706e+00 -6.98994696e-01
-3.93858016e-01 4.44877207e-01 1.89295691e-02 -2.80269533e-01
-9.64223593e-02 1.84670523e-01 -1.28442216e+00 1.78590998e-01
4.89670753e-01 1.22956276e+00 1.85479924e-01 8.22948754e-01
2.38355681e-01 1.10357916e+00 -6.77692711e-01 -4.72622663e-01
5.27576506e-01 -2.17803031e-01 -7.32549608e-01 -6.59746304e-02
5.74922204e-01 -4.33351815e-01 -5.36428154e-01 1.17452455e+00
7.31050253e-01 -9.68989432e-02 5.88954628e-01 -1.45069993e+00
-1.19368660e+00 5.43228567e-01 -7.20342219e-01 3.27445358e-01
-9.91599783e-02 1.42604351e-01 1.18603945e+00 -1.05274975e+00
4.90910381e-01 1.02140462e+00 4.53047991e-01 8.99963617e-01
-1.07583535e+00 -1.08983207e+00 2.10574880e-01 2.38949209e-01
-1.35029316e+00 -6.26749039e-01 8.65842342e-01 -2.38151819e-01
6.26588583e-01 1.75821722e-01 5.03082991e-01 1.26913118e+00
-9.54314843e-02 1.19161034e+00 8.44256997e-01 -3.20027441e-01
4.04517204e-02 3.68704081e-01 1.08831506e-02 6.49465740e-01
-2.18344957e-01 2.59488821e-01 -4.28727984e-01 -3.15723531e-02
6.53902352e-01 2.20233113e-01 -2.45731443e-01 -4.66371506e-01
-1.05817735e+00 9.55731630e-01 9.44527924e-01 2.02164333e-02
-2.61767805e-01 4.13514495e-01 5.62072396e-01 6.39371276e-01
7.16815531e-01 3.32954943e-01 -8.18109274e-01 2.30336189e-01
-7.32631683e-01 -3.16737546e-03 4.91395772e-01 1.22886384e+00
1.03798819e+00 1.49800554e-01 -1.88651830e-01 8.70647013e-01
2.25543573e-01 7.05275178e-01 6.82100177e-01 -6.00701034e-01
4.25968885e-01 4.34400648e-01 -6.23801611e-02 -7.15216339e-01
1.75593812e-02 -6.91930115e-01 -9.84489501e-01 1.69378132e-01
-2.55426168e-02 -4.23273057e-01 -1.30613375e+00 1.90790200e+00
9.16134566e-02 3.66939008e-01 3.17706347e-01 7.75093794e-01
7.76602924e-01 6.89683318e-01 1.65251777e-01 1.53862566e-01
1.08503473e+00 -1.35425854e+00 -4.55037147e-01 -6.07252121e-01
6.27927542e-01 -6.75143003e-01 1.10281396e+00 8.90003368e-02
-9.78977621e-01 -5.64953327e-01 -1.17267764e+00 -1.13196559e-01
-3.73322606e-01 3.18055391e-01 4.29310918e-01 1.52323261e-01
-1.21200061e+00 4.85796183e-01 -7.89196432e-01 8.52282792e-02
8.90454531e-01 6.72504723e-01 -5.04907608e-01 -1.36009008e-01
-1.35736883e+00 7.20377266e-01 2.35807747e-01 -5.98371923e-02
-1.37195599e+00 -6.59686387e-01 -8.59457433e-01 -6.38473108e-02
1.96871147e-01 -3.70749325e-01 1.29007363e+00 -1.65355885e+00
-1.30621350e+00 1.14980376e+00 -2.81426366e-02 -7.55212188e-01
5.67548752e-01 -4.83091623e-01 -3.52758706e-01 1.12329379e-01
3.22620243e-01 1.23023891e+00 1.07033145e+00 -1.18593907e+00
-6.41973615e-01 -1.71784684e-01 1.01695150e-01 2.24495143e-01
-5.96873343e-01 -1.47086427e-01 -5.83716989e-01 -7.29979932e-01
-1.78605303e-01 -9.24692392e-01 -3.48118871e-01 3.08848113e-01
-4.28678006e-01 -2.35247388e-01 8.86075616e-01 -4.76302058e-01
5.67536473e-01 -2.42051840e+00 -1.02662958e-01 -3.90700623e-02
3.23728532e-01 4.20861363e-01 -5.22486269e-01 1.92440853e-01
-2.64044106e-01 4.20048945e-02 -1.34358943e-01 -5.44843554e-01
-1.87474132e-01 2.56046921e-01 -3.79247963e-01 5.58517456e-01
7.05676377e-01 1.03326046e+00 -1.22837436e+00 -3.26884001e-01
1.56057447e-01 7.36047924e-01 -5.56438744e-01 4.36723620e-01
-1.92308888e-01 2.96622604e-01 -2.67084152e-01 5.72981596e-01
7.68038988e-01 -3.87120396e-01 -2.70954948e-02 -5.34348115e-02
3.81324321e-01 2.84489125e-01 -7.41597474e-01 1.59574676e+00
-4.38056409e-01 6.86785877e-01 4.19984832e-02 -1.22300196e+00
1.01861835e+00 2.52386063e-01 2.25254476e-01 -9.05525506e-01
-2.71684583e-02 8.41959417e-02 -2.23329231e-01 -2.32374325e-01
4.97975737e-01 -8.20658207e-02 3.58653143e-02 1.02934368e-01
6.43805325e-01 4.29604769e-01 5.98427057e-02 2.96629906e-01
1.00196075e+00 2.30945032e-02 1.59971550e-01 -2.15747848e-01
4.94420916e-01 -2.13905454e-01 5.87406218e-01 5.96502483e-01
-2.89919198e-01 6.90954626e-01 4.84511524e-01 -2.96641469e-01
-9.30790663e-01 -1.03451514e+00 4.74371165e-02 1.51903081e+00
3.88305523e-02 -1.87651083e-01 -3.71594399e-01 -1.19990230e+00
4.36928049e-02 4.16085541e-01 -1.01048553e+00 -4.80897248e-01
-4.99141425e-01 -5.17090917e-01 5.55844188e-01 5.70594847e-01
5.58452308e-01 -1.14003658e+00 -7.39126429e-02 2.73818225e-02
2.97931097e-02 -7.82326639e-01 -3.64443749e-01 7.58378029e-01
-5.94649673e-01 -1.10172141e+00 -8.49642038e-01 -1.11998498e+00
1.06325090e+00 2.48192549e-01 1.34555829e+00 3.08610704e-02
1.23923179e-02 3.26521337e-01 -4.66762155e-01 -5.52649200e-01
-3.52514386e-01 3.75509799e-01 -1.42127067e-01 6.87883496e-02
3.33828658e-01 -6.07235193e-01 -7.99519122e-01 2.68303066e-01
-8.65378439e-01 -5.79315983e-02 7.88747489e-01 1.03032851e+00
8.90269101e-01 -4.94817525e-01 9.00996149e-01 -1.08473456e+00
1.66125968e-01 -8.89783919e-01 -4.40250605e-01 1.28631383e-01
-4.52503294e-01 8.51148441e-02 9.86326694e-01 -6.21308267e-01
-5.11661947e-01 2.72498786e-01 -2.68145174e-01 -7.41275489e-01
-8.74609780e-03 1.63103893e-01 -1.40092477e-01 -5.39663285e-02
7.20840216e-01 3.53783458e-01 -5.20519074e-03 -2.29766279e-01
5.04848719e-01 6.13907874e-01 5.57323694e-01 -4.34173971e-01
1.09573841e+00 4.16261077e-01 -5.62667727e-01 -4.38268453e-01
-1.11506259e+00 -5.63246846e-01 -3.96086872e-01 -3.84933650e-02
4.73297507e-01 -1.44772434e+00 -1.95032254e-01 3.51610363e-01
-7.18916237e-01 -5.22313952e-01 -6.25091195e-01 2.01734737e-01
-4.08233941e-01 -5.43004125e-02 -5.19182205e-01 -5.16162157e-01
-5.93536377e-01 -1.02086830e+00 1.01085269e+00 1.83569834e-01
1.03127517e-01 -9.46078718e-01 7.86637701e-03 3.62436399e-02
5.94498754e-01 1.02123216e-01 4.20329720e-01 -1.03089845e+00
-3.89256567e-01 -5.48783302e-01 -6.42091781e-02 1.00165558e+00
2.34860718e-01 -1.38055980e-01 -1.28146029e+00 -6.07456088e-01
-2.40207940e-01 -9.30149972e-01 1.00496662e+00 7.17663988e-02
1.18307400e+00 -5.96800685e-01 -1.81756869e-01 8.01138580e-01
1.54419911e+00 -6.66139424e-02 6.37619376e-01 3.82773578e-01
6.34378731e-01 1.36199296e-01 6.21767759e-01 3.31898928e-01
3.04801524e-01 2.47717276e-01 8.94022048e-01 -3.38090479e-01
-9.41570103e-02 -4.29047942e-01 5.13098419e-01 6.76992476e-01
3.28321368e-01 -2.12306336e-01 -6.04466319e-01 7.52634346e-01
-1.69537139e+00 -8.84242117e-01 4.46094543e-01 2.00192928e+00
1.22586536e+00 2.79296517e-01 2.12921277e-02 5.38925231e-02
7.14113891e-01 4.76780981e-01 -8.27669859e-01 1.63418785e-01
-1.72480792e-01 3.33784878e-01 8.72109771e-01 4.57430154e-01
-1.55269122e+00 1.08635354e+00 5.76707268e+00 8.26294422e-01
-1.55389071e+00 1.90576762e-01 9.42093730e-01 -2.84999758e-01
-1.69551730e-01 -1.73677698e-01 -8.59681070e-01 4.66297507e-01
9.55824733e-01 5.22668066e-04 2.54579574e-01 1.40378654e+00
-2.36538738e-01 3.94831181e-01 -1.16923785e+00 6.86996222e-01
-9.41936150e-02 -1.29398811e+00 -4.03328761e-02 -2.41385311e-01
7.77167022e-01 6.62832797e-01 4.01113510e-01 7.20084250e-01
6.04680717e-01 -9.73270297e-01 8.80221486e-01 2.26079941e-01
1.07746923e+00 -7.51327276e-01 7.22861111e-01 2.49454707e-01
-9.01556730e-01 2.03644298e-02 -6.24849617e-01 5.25500067e-02
-3.00559878e-01 5.02298236e-01 -9.58179474e-01 1.76698551e-01
7.59194613e-01 1.08755815e+00 -7.29364336e-01 7.46537328e-01
-4.45565641e-01 9.30961072e-01 -3.33484828e-01 1.07707120e-01
3.21271122e-01 1.92402437e-01 1.21106021e-01 1.18674624e+00
-1.43581361e-01 -4.65971738e-01 8.25539529e-02 5.84243298e-01
-6.98459804e-01 -6.51105419e-02 -5.36723137e-01 7.79709890e-02
5.54726839e-01 1.26490068e+00 -3.40820879e-01 -4.66579735e-01
-3.65000784e-01 1.19818509e+00 8.05751204e-01 5.15449226e-01
-9.63022351e-01 -3.33821684e-01 6.26234412e-01 -4.56883051e-02
4.24923837e-01 1.91590011e-01 7.15891644e-02 -1.09943080e+00
6.06788769e-02 -8.92625153e-01 2.58117706e-01 -5.08700252e-01
-1.53204513e+00 6.43904567e-01 -4.14931834e-01 -1.40846717e+00
-2.35580891e-01 -4.99353558e-01 -5.34866571e-01 8.02453458e-01
-1.90101755e+00 -1.24949193e+00 -2.93678373e-01 7.59633362e-01
3.68417948e-01 -3.09034884e-01 1.06316853e+00 4.53569919e-01
-5.26182771e-01 1.19358194e+00 3.17624807e-01 6.85326576e-01
9.69704628e-01 -1.39039004e+00 5.20121038e-01 6.67986393e-01
1.79410502e-01 3.63985062e-01 4.48887885e-01 -1.43918321e-01
-1.06202877e+00 -1.51556683e+00 5.87396801e-01 -1.79418787e-01
6.68235362e-01 -5.73369682e-01 -8.41936231e-01 1.01290643e+00
3.50545853e-01 7.80873418e-01 7.39986420e-01 -1.14534348e-01
-6.73416495e-01 -3.55156392e-01 -1.25558400e+00 3.41588885e-01
8.49584401e-01 -5.34481049e-01 -1.71638176e-01 6.06031656e-01
9.52084780e-01 -3.94160539e-01 -6.93887830e-01 3.88070703e-01
2.19631851e-01 -7.45261252e-01 1.10330021e+00 -6.81519806e-01
5.72838724e-01 -1.29421949e-01 -1.89838320e-01 -1.29260206e+00
-2.79819161e-01 -4.33886439e-01 -2.58146137e-01 1.13192117e+00
5.20151734e-01 -6.91995800e-01 9.89052355e-01 1.59834884e-02
-1.95931837e-01 -1.12969291e+00 -8.15673649e-01 -6.66364491e-01
2.18866885e-01 -1.52046144e-01 2.64371127e-01 1.11185336e+00
-5.70795417e-01 5.67386806e-01 -5.75171769e-01 2.02030554e-01
5.48360765e-01 -3.02041888e-01 7.86623001e-01 -8.83984327e-01
-2.63254434e-01 1.79227293e-02 -5.04555166e-01 -1.33779109e+00
2.66276270e-01 -1.07294869e+00 6.14188127e-02 -1.16634190e+00
6.90114200e-02 -7.39671469e-01 -7.94869602e-01 7.10564613e-01
-2.17337325e-01 4.56350863e-01 2.85933260e-02 3.31791133e-01
-9.05058503e-01 7.99618363e-01 1.11660242e+00 -3.84174824e-01
4.37128097e-02 8.85901079e-02 -8.73025715e-01 6.13375545e-01
1.12183928e+00 -7.80436158e-01 -6.94025457e-01 -6.26569629e-01
1.50684327e-01 -2.49939561e-01 4.64828670e-01 -8.78378332e-01
3.58260386e-02 1.72884822e-01 6.77084863e-01 -4.64639992e-01
3.59418541e-01 -7.91094422e-01 -4.23390508e-01 3.56094986e-01
-7.83854127e-01 4.73440886e-02 6.52847961e-02 7.19596624e-01
-3.59076858e-01 -2.26576492e-01 1.01480865e+00 -1.24710292e-01
-7.78324068e-01 6.57269597e-01 1.33627310e-01 3.81338507e-01
1.04912698e+00 2.03808755e-01 -4.25822824e-01 -5.06070256e-01
-7.46740103e-01 5.14968216e-01 3.52535486e-01 5.13501883e-01
6.64545059e-01 -1.47701383e+00 -8.13377798e-01 3.35948914e-01
1.83850184e-01 4.42259490e-01 1.26606405e-01 5.75965703e-01
-4.80339766e-01 -5.54305501e-02 -1.53879210e-01 -5.23178935e-01
-1.05917859e+00 5.41380048e-01 4.54033434e-01 -4.29040194e-01
-3.24879140e-01 1.35152423e+00 3.54071915e-01 -4.96921271e-01
5.74853778e-01 8.17775801e-02 -7.07327947e-02 -2.33295318e-02
7.03965724e-01 -5.17633967e-02 -1.76103681e-01 -6.07961595e-01
-3.99771959e-01 4.37391773e-02 -5.86350739e-01 1.01060309e-01
1.56207919e+00 1.08098887e-01 9.83530656e-02 3.20050985e-01
1.84161985e+00 -1.81172729e-01 -1.62944794e+00 -4.94051456e-01
-1.79371029e-01 -1.05711885e-01 -1.40293434e-01 -7.57716358e-01
-1.59121454e+00 8.92987967e-01 8.85969639e-01 2.47981958e-02
1.11721349e+00 1.48501784e-01 6.48644686e-01 3.89763802e-01
8.84792879e-02 -7.75426924e-01 3.68981570e-01 4.45788234e-01
9.52127874e-01 -1.55553544e+00 -2.42393374e-01 -9.99498591e-02
-6.12386286e-01 6.90990806e-01 7.94726610e-01 -8.13212931e-01
7.32320011e-01 2.66236335e-01 2.48042986e-01 4.30094972e-02
-9.52932060e-01 -7.79116526e-02 1.29680008e-01 7.66588986e-01
4.58448052e-01 1.40417978e-01 3.00457388e-01 5.21922350e-01
-2.64187008e-01 -9.18050334e-02 1.94775105e-01 1.00415981e+00
-2.91258097e-01 -8.53600085e-01 6.53853044e-02 5.43648422e-01
-7.08548427e-01 -3.21494699e-01 -3.18200618e-01 7.97314048e-01
8.78697038e-02 5.54390490e-01 2.26217851e-01 -6.97130919e-01
2.04662770e-01 -9.25717726e-02 2.24534854e-01 -6.17653787e-01
-5.70923090e-01 -3.27374965e-01 -6.66602254e-02 -3.84290308e-01
-4.40510064e-01 -3.05870533e-01 -1.21098602e+00 -1.06284380e-01
-3.71595860e-01 1.86123863e-01 3.75567645e-01 5.92599213e-01
4.55405802e-01 5.20062923e-01 1.02876627e+00 -1.00688088e+00
-8.57175589e-01 -1.27266562e+00 -2.66711265e-01 3.82003307e-01
6.86207235e-01 -4.06293392e-01 -5.78157187e-01 1.25178903e-01]
|
[9.551514625549316, 2.572080373764038]
|
624f9b8e-6dd2-45a0-b9b5-6a4a9a8d3809
|
neural-network-extrapolations-with-g
| null | null |
https://openreview.net/forum?id=7t1FcJUWhi3
|
https://openreview.net/pdf?id=7t1FcJUWhi3
|
Neural Network Extrapolations with G-invariances from a Single Environment
|
Despite —or maybe because of— their astonishing capacity to fit data, neural networks are widely believed to be unable to extrapolate beyond training data distribution. This work shows that, for extrapolations based on transformation groups, a model’s inability to extrapolate is unrelated to its capacity. Rather, the shortcoming is inherited from a classical statistical learning hypothesis: Examples not explicitly observed with infinitely many training examples cannot be likely outcomes in the learner’s model. In order to endow neural networks with the ability to extrapolate over group transformations, we introduce a learning framework guided by a new learning hypothesis: Any invariance to transformation groups is mandatory even without evidence, unless the learner deems it inconsistent with the training data. Unlike existing invariance-driven methods for counterfactual inference, this framework allows extrapolations from a single environment. Finally, we introduce sequence and image extrapolation tasks that validate our framework and showcase the shortcomings of traditional approaches.
|
['Bruno Ribeiro', 'S Chandra Mouli']
|
2021-01-01
| null | null | null |
iclr-2021-1
|
['counterfactual-inference']
|
['miscellaneous']
|
[ 6.44009829e-01 5.25106072e-01 -3.71145278e-01 -5.06387651e-01
-3.85503590e-01 -5.42551100e-01 9.39366996e-01 -2.11613238e-01
-6.43682837e-01 1.15247929e+00 1.05036817e-01 -8.06014240e-01
-4.20641840e-01 -7.76229918e-01 -1.38761353e+00 -5.83909452e-01
-9.31333285e-03 1.47427097e-01 5.01192510e-02 -8.58774036e-02
2.55323857e-01 4.71256346e-01 -1.54754448e+00 3.23590666e-01
8.91305983e-01 6.35217488e-01 2.76364572e-02 5.15816092e-01
1.36702895e-01 1.17247498e+00 -2.32591495e-01 -5.43645024e-01
3.12595844e-01 -7.73098707e-01 -6.53653264e-01 -1.16001710e-01
8.36721838e-01 -7.13509619e-01 -4.38416958e-01 1.23324239e+00
-2.33746693e-02 2.48300716e-01 9.39314902e-01 -1.54799569e+00
-1.09496379e+00 8.08797002e-01 1.03915902e-02 7.46736750e-02
2.64920950e-01 2.97201693e-01 8.87268841e-01 -6.83415592e-01
7.04914808e-01 9.84224558e-01 8.63242924e-01 8.13123703e-01
-1.54550517e+00 -5.65224469e-01 3.91672939e-01 1.77176788e-01
-9.32791173e-01 -4.89740908e-01 7.53006458e-01 -4.60434675e-01
6.66910410e-01 4.00584370e-01 7.51028180e-01 1.67048395e+00
2.29911685e-01 7.54395187e-01 1.51112676e+00 -6.89496875e-01
4.84575927e-01 3.40246081e-01 -2.45428875e-01 4.18386132e-01
4.40983117e-01 7.38159955e-01 -7.04548776e-01 8.27378184e-02
9.22309935e-01 -1.54341459e-01 -3.74703616e-01 -7.62508750e-01
-1.20890355e+00 7.94189274e-01 4.20961827e-01 1.62141651e-01
-3.62053126e-01 3.11533511e-01 2.70125389e-01 4.71488804e-01
2.72874117e-01 6.22662842e-01 -5.50205946e-01 2.34826654e-01
-1.19021511e+00 4.71673787e-01 5.90705693e-01 8.33105206e-01
6.20891511e-01 2.41904050e-01 1.17450014e-01 2.38230720e-01
2.38479555e-01 2.21321151e-01 6.93381667e-01 -1.27270901e+00
1.78727493e-01 1.95633978e-01 1.43698096e-01 -5.12477398e-01
-1.06597021e-01 -5.35307705e-01 -6.59722269e-01 5.48861802e-01
8.62935305e-01 -1.18572414e-01 -4.93184596e-01 2.34396338e+00
-4.28379886e-02 2.92463690e-01 1.48624226e-01 5.18374681e-01
7.03552738e-02 1.65063724e-01 6.99835122e-02 -5.28106391e-01
5.71305692e-01 -3.46524358e-01 -3.54237914e-01 -1.30137086e-01
7.98060179e-01 -1.01115420e-01 1.43472350e+00 5.90989769e-01
-1.20789790e+00 -5.33931077e-01 -1.23591602e+00 9.63276476e-02
-2.09689394e-01 -5.84460974e-01 8.72678041e-01 7.02882767e-01
-1.00196481e+00 1.06326330e+00 -6.53852880e-01 -3.35956961e-01
6.90644264e-01 1.63794637e-01 -4.31538761e-01 9.06960815e-02
-1.17887962e+00 1.17090416e+00 8.43572080e-01 -8.24593827e-02
-8.23499441e-01 -1.00794494e+00 -7.28423297e-01 3.92102413e-02
2.65791237e-01 -9.73688662e-01 1.28405988e+00 -1.55497777e+00
-1.12949681e+00 6.86487973e-01 1.89808328e-02 -1.04501307e+00
1.21424758e+00 -1.73825815e-01 -3.04746300e-01 -8.76255259e-02
-9.90303978e-02 8.69646549e-01 7.46963263e-01 -1.41622376e+00
-4.29756284e-01 -2.53939062e-01 1.08310468e-01 1.77435294e-01
-1.92363977e-01 -3.38220477e-01 5.66125631e-01 -6.14445329e-01
-6.93131164e-02 -8.25174570e-01 -2.44452119e-01 2.88973480e-01
-4.77665752e-01 -1.26018777e-01 4.55202311e-01 -2.37073749e-01
9.19681430e-01 -1.97718120e+00 -4.68571365e-01 5.75772822e-02
-5.07995412e-02 -1.21489607e-01 2.02245172e-02 2.50423133e-01
-4.87342030e-01 3.77381414e-01 -4.38757658e-01 1.38085604e-01
4.34864849e-01 2.16872737e-01 -8.39129210e-01 5.01710474e-01
1.21652052e-01 9.47640121e-01 -9.26795423e-01 -4.85708654e-01
2.93713033e-01 1.19070180e-01 -6.46229386e-01 -1.16311297e-01
-3.35139036e-01 4.96288985e-01 -2.38595027e-02 -2.39361264e-02
6.47307932e-01 -4.06285413e-02 2.00608462e-01 1.75836936e-01
-3.11226547e-02 1.21349655e-01 -8.99367630e-01 1.45058000e+00
-3.64309132e-01 6.96742177e-01 -7.39044666e-01 -1.40010524e+00
6.66386247e-01 2.88377702e-01 7.47015253e-02 -4.58416075e-01
-9.98566747e-02 3.91358793e-01 3.68106037e-01 -5.50577998e-01
2.53343165e-01 -8.22293341e-01 2.16243804e-01 6.82276726e-01
2.26018801e-01 -1.13314211e-01 -9.21454728e-02 5.87386489e-02
7.47492611e-01 8.54492426e-01 4.55429703e-01 -2.71130621e-01
3.22157800e-01 2.83148475e-02 5.78863859e-01 1.17773747e+00
-1.20030768e-01 3.92390460e-01 4.66687113e-01 -6.92245781e-01
-1.38553858e+00 -1.43810487e+00 -3.57320219e-01 1.01619411e+00
-2.60479867e-01 2.07125619e-01 -7.29002953e-01 -1.04791880e+00
-9.08011869e-02 1.59563458e+00 -9.28023279e-01 -3.02215248e-01
-3.68438125e-01 -5.34775555e-01 5.85148692e-01 6.60264134e-01
3.09144050e-01 -1.20244694e+00 -8.40684652e-01 3.31721902e-02
4.55539264e-02 -8.08398366e-01 -1.27075523e-01 9.48306769e-02
-1.06850600e+00 -9.29159999e-01 -5.18189371e-01 -3.57084036e-01
8.44390213e-01 -2.11437330e-01 1.01949549e+00 3.09282318e-02
2.46452957e-01 3.37167710e-01 1.71297356e-01 -7.26983845e-01
-8.60300124e-01 -3.43793660e-01 3.90849262e-01 -3.63507003e-01
5.84325850e-01 -1.06449997e+00 -5.33909142e-01 2.78968718e-02
-9.89720464e-01 2.32601270e-01 5.76340377e-01 9.13596272e-01
3.02876025e-01 -2.60506302e-01 9.91979003e-01 -1.01163065e+00
4.84084845e-01 -4.59752709e-01 -5.81236422e-01 3.79232436e-01
-7.61889815e-01 2.44464397e-01 1.04818809e+00 -8.17362368e-01
-1.29017758e+00 8.10090173e-03 1.86066911e-01 -3.47895324e-01
-5.16636848e-01 4.69343185e-01 -2.01301008e-01 2.61188716e-01
1.06431079e+00 5.45656860e-01 8.15223530e-02 -1.26657501e-01
5.75065196e-01 1.13539025e-01 6.71610892e-01 -8.60980153e-01
7.29396343e-01 5.02869129e-01 2.95245260e-01 -5.52473128e-01
-8.76959980e-01 4.90950406e-01 -9.56657469e-01 -2.84253657e-01
4.48905349e-01 -6.42882168e-01 -7.71189988e-01 1.77462082e-02
-8.65995586e-01 -6.22381866e-01 -8.61518323e-01 8.97418261e-01
-1.30260813e+00 3.91513199e-01 -1.52051955e-01 -1.03565288e+00
3.68151039e-01 -6.96625769e-01 1.04797244e-01 -6.15009107e-02
-5.56481183e-01 -1.45552707e+00 -2.62527794e-01 -1.45602286e-01
3.16351771e-01 4.28403676e-01 1.08985686e+00 -1.12184703e+00
-5.58285773e-01 -1.85608983e-01 9.23863053e-02 5.55491865e-01
3.70509597e-03 -7.27185234e-02 -1.13297534e+00 -1.24136582e-01
3.16257626e-01 -5.27571261e-01 8.83286715e-01 4.55989987e-01
1.38780522e+00 -8.98509562e-01 -5.45488223e-02 4.41110343e-01
1.39827394e+00 2.57364642e-02 5.48180223e-01 4.01032209e-01
1.96670070e-01 8.32516432e-01 2.64554292e-01 9.67566371e-02
3.05609275e-02 1.93560034e-01 5.34999192e-01 2.53906220e-01
-3.97019312e-02 -1.03300452e+00 5.12196898e-01 2.91062713e-01
-2.65166670e-01 -5.00536822e-02 -5.91693819e-01 5.92947662e-01
-1.79806411e+00 -1.41807532e+00 -6.74560890e-02 2.71860647e+00
9.98935223e-01 3.97495121e-01 5.20355962e-02 9.09032971e-02
4.72596943e-01 -1.18006743e-01 -9.88847852e-01 -4.31377500e-01
-2.62905151e-01 -1.05968483e-01 4.53397542e-01 5.33573687e-01
-7.47451782e-01 6.22614741e-01 7.29168177e+00 8.58676195e-01
-8.24442267e-01 -1.25934705e-01 7.21697450e-01 3.26289646e-02
-7.69894004e-01 2.29240954e-01 -3.71148169e-01 3.98742914e-01
1.08402097e+00 -4.14192677e-01 4.52307731e-01 7.57056534e-01
8.82352740e-02 -1.10693000e-01 -1.90249014e+00 3.15912187e-01
-5.25603350e-03 -1.45780253e+00 3.10204059e-01 2.65938818e-01
7.89840460e-01 -2.46023923e-01 4.20468003e-01 4.65998173e-01
8.53968680e-01 -1.42634785e+00 1.04391479e+00 6.08253896e-01
6.75173938e-01 -6.44540012e-01 5.18652737e-01 1.06289256e+00
-1.60241425e-01 -5.24218343e-02 -5.23161650e-01 -3.75522316e-01
-1.95020646e-01 1.34176880e-01 -9.74641085e-01 3.20418507e-01
9.01680961e-02 3.46716762e-01 -4.08931881e-01 6.33506238e-01
-4.99517083e-01 7.75120616e-01 -2.20676258e-01 3.09832767e-02
7.55007714e-02 -8.57454464e-02 4.13648993e-01 9.71692145e-01
5.92157423e-01 -1.18685260e-01 -2.81562895e-01 1.20470035e+00
-7.83665013e-03 -2.76953960e-03 -1.14169562e+00 4.57091004e-01
5.28123438e-01 5.81013799e-01 -4.00951356e-01 -4.05348361e-01
-4.93220061e-01 8.00988555e-01 2.75246382e-01 6.23073816e-01
-7.87376702e-01 2.89538670e-02 1.00088805e-01 -6.33369833e-02
5.47045656e-02 -1.17647732e-02 -5.90980291e-01 -1.15792871e+00
5.35809286e-02 -8.06187272e-01 2.57146478e-01 -9.49608266e-01
-1.37958920e+00 1.59954906e-01 2.59674370e-01 -1.15918899e+00
-7.42978692e-01 -7.60893643e-01 -5.82496107e-01 9.52641487e-01
-1.24145305e+00 -1.27088058e+00 3.77980888e-01 5.35696805e-01
3.76591802e-01 -3.68669443e-02 8.65561783e-01 -5.26594639e-01
-1.03323059e-02 6.52201891e-01 -7.30296597e-03 6.52214745e-03
5.66909730e-01 -1.47346723e+00 3.21336776e-01 8.90202284e-01
4.05138999e-01 9.61962938e-01 1.14272404e+00 -4.35883284e-01
-9.14437652e-01 -9.12755251e-01 8.33454549e-01 -7.78541446e-01
7.74845004e-01 -3.00505430e-01 -1.14077985e+00 1.24636149e+00
2.40933597e-01 2.27345571e-01 7.43590534e-01 2.68472791e-01
-7.44769931e-01 8.66303742e-02 -1.16358232e+00 9.24284160e-01
1.19907963e+00 -4.93099540e-01 -1.06164598e+00 2.60932207e-01
5.45011163e-01 -1.13199884e-02 -6.95005000e-01 3.38952601e-01
8.49998653e-01 -1.19924378e+00 9.12063360e-01 -1.35522974e+00
7.79475689e-01 -4.97053638e-02 -4.38445121e-01 -1.36627042e+00
-7.92316496e-02 -4.27059203e-01 -1.90973744e-01 1.02012932e+00
4.99285221e-01 -7.90695667e-01 8.88511598e-01 8.90640676e-01
3.53173018e-02 -5.01659691e-01 -9.98422861e-01 -1.29347003e+00
9.58109498e-01 -9.05951083e-01 7.01792002e-01 1.21516216e+00
1.77702963e-01 1.05289631e-01 -3.22747916e-01 -2.97544189e-02
9.09268439e-01 1.63150337e-02 6.80135787e-01 -1.22846711e+00
-4.47513312e-01 -5.22912204e-01 -2.45538846e-01 -8.73445034e-01
3.91238391e-01 -1.00563478e+00 2.64141876e-02 -1.18476176e+00
3.10917974e-01 -3.79636824e-01 -5.54487288e-01 3.30139399e-01
2.23354176e-02 1.65108025e-01 3.51277858e-01 2.16215461e-01
-2.46465728e-01 3.74236703e-01 1.16743159e+00 3.22044462e-01
8.47058222e-02 1.52269617e-01 -1.01364541e+00 1.17974758e+00
8.69060099e-01 -4.14840698e-01 -7.42303014e-01 -6.48718700e-02
6.55313492e-01 -4.58890796e-02 9.70208108e-01 -8.08682084e-01
2.80028164e-01 -5.05172908e-01 7.79444337e-01 -2.04055503e-01
4.00170833e-02 -7.22426832e-01 2.05137089e-01 6.02482438e-01
-9.68401372e-01 -3.98218989e-01 1.10160910e-01 5.97517431e-01
1.93581760e-01 -4.74764973e-01 5.82372308e-01 -2.99649596e-01
-5.43631136e-01 3.80713344e-02 -3.54057312e-01 1.79466844e-01
7.94566214e-01 -4.41507339e-01 -2.85413295e-01 -4.35546160e-01
-8.92953038e-01 -2.07170576e-01 7.44562745e-01 1.43025771e-01
5.55447638e-01 -1.35029852e+00 -8.09831381e-01 6.19514920e-02
8.99224579e-02 -1.92684948e-01 4.00222868e-01 8.55696619e-01
-6.07164204e-02 3.60900611e-01 -2.96822459e-01 -4.03646886e-01
-6.50579989e-01 1.09003437e+00 3.93686980e-01 1.91741604e-02
-4.95897830e-01 5.80733299e-01 5.89191914e-01 -5.71521521e-01
-7.80105665e-02 -4.06589597e-01 2.04868019e-01 -3.16842258e-01
4.42068756e-01 7.27136359e-02 -2.26064950e-01 -2.82618076e-01
-4.81855273e-02 2.69706044e-02 -1.14963688e-01 -4.26481485e-01
1.30907679e+00 -6.88942075e-02 4.02713954e-01 9.93983150e-01
7.57682383e-01 -9.44524258e-02 -1.80407155e+00 -2.63356239e-01
7.01907426e-02 -5.49964786e-01 -3.70777547e-01 -1.00218654e+00
-2.00030267e-01 1.10318792e+00 1.73258960e-01 1.91673040e-01
9.31928694e-01 -1.79893672e-01 4.23927307e-02 5.53130805e-01
3.82792771e-01 -1.10005641e+00 -2.10340098e-01 1.55750394e-01
1.00811672e+00 -1.18805909e+00 1.38211414e-01 5.37733994e-02
-5.97163558e-01 1.23840022e+00 4.60884064e-01 -2.59077698e-01
2.27656022e-01 3.79933193e-02 -1.57164335e-01 3.08692306e-01
-9.48604167e-01 2.03022778e-01 1.45957515e-01 9.82613981e-01
3.85693699e-01 -2.09276658e-02 -8.10186043e-02 5.33365548e-01
-5.42407572e-01 4.01499718e-01 7.45612979e-01 5.44747472e-01
-3.95966202e-01 -7.37489223e-01 -2.51902401e-01 3.68089199e-01
-4.93905991e-01 -1.73594132e-01 -1.19039088e-01 1.34151471e+00
3.00140768e-01 5.10837138e-01 8.42360482e-02 3.99760753e-02
1.70544516e-02 4.40281659e-01 7.77602255e-01 -3.07789713e-01
1.86297055e-02 -3.94222081e-01 -1.52036399e-01 -3.19297791e-01
-6.24925911e-01 -9.84688818e-01 -8.49633753e-01 -2.75517195e-01
-1.02362268e-01 -1.19023129e-01 3.92909259e-01 1.15072632e+00
-1.66387931e-01 3.35372537e-01 4.99784946e-01 -3.31903428e-01
-1.10633481e+00 -9.37472284e-01 -5.28810740e-01 5.11970162e-01
4.75273818e-01 -3.21948469e-01 -5.66086829e-01 3.50112408e-01]
|
[8.5984468460083, 5.432613849639893]
|
caf4aa7e-b4d1-40df-906c-e7979d2cdaf2
|
beyond-frontal-faces-improving-person
|
1501.05703
| null |
http://arxiv.org/abs/1501.05703v2
|
http://arxiv.org/pdf/1501.05703v2.pdf
|
Beyond Frontal Faces: Improving Person Recognition Using Multiple Cues
|
We explore the task of recognizing peoples' identities in photo albums in an
unconstrained setting. To facilitate this, we introduce the new People In Photo
Albums (PIPA) dataset, consisting of over 60000 instances of 2000 individuals
collected from public Flickr photo albums. With only about half of the person
images containing a frontal face, the recognition task is very challenging due
to the large variations in pose, clothing, camera viewpoint, image resolution
and illumination. We propose the Pose Invariant PErson Recognition (PIPER)
method, which accumulates the cues of poselet-level person recognizers trained
by deep convolutional networks to discount for the pose variations, combined
with a face recognizer and a global recognizer. Experiments on three different
settings confirm that in our unconstrained setup PIPER significantly improves
on the performance of DeepFace, which is one of the best face recognizers as
measured on the LFW dataset.
|
['Rob Fergus', 'Manohar Paluri', 'Yaniv Taigman', 'Ning Zhang', 'Lubomir Bourdev']
|
2015-01-23
|
beyond-frontal-faces-improving-person-1
|
http://openaccess.thecvf.com/content_cvpr_2015/html/Zhang_Beyond_Frontal_Faces_2015_CVPR_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2015/papers/Zhang_Beyond_Frontal_Faces_2015_CVPR_paper.pdf
|
cvpr-2015-6
|
['person-recognition']
|
['computer-vision']
|
[ 1.82267874e-01 -4.53572929e-01 3.53254974e-01 -5.21251202e-01
-5.24258494e-01 -7.35696852e-01 7.94692576e-01 -8.79145682e-01
-3.70915353e-01 4.92450386e-01 3.31916660e-01 7.74571598e-01
2.15334252e-01 -3.80373120e-01 -6.19186759e-01 -7.31279731e-01
-4.17215005e-02 5.07101834e-01 -4.39776182e-01 -1.15785450e-01
-1.46623939e-01 8.12534988e-01 -1.76819944e+00 2.94006050e-01
2.70416766e-01 1.25170946e+00 -5.13423204e-01 7.32369602e-01
4.86296684e-01 3.96983951e-01 -6.26129925e-01 -1.13465035e+00
8.93025160e-01 1.09282248e-01 -4.13580805e-01 5.26683748e-01
1.91083956e+00 -6.58786356e-01 -7.44886696e-01 9.38271523e-01
9.73455548e-01 1.46022782e-01 5.17776549e-01 -1.34046984e+00
-7.96877444e-01 1.55006871e-01 -5.26703298e-01 1.29406959e-01
6.91087604e-01 2.11896047e-01 6.52252436e-01 -1.33973205e+00
4.13820058e-01 1.72492337e+00 1.08347297e+00 7.90111244e-01
-1.09467852e+00 -6.60391271e-01 4.84335348e-02 -6.06556982e-02
-1.90451407e+00 -1.12399399e+00 5.14964640e-01 -3.66834909e-01
6.15429103e-01 2.89828688e-01 4.59278166e-01 1.43957233e+00
-4.98302579e-01 6.46194875e-01 8.52554917e-01 -3.65719020e-01
-1.80238843e-01 -6.80717900e-02 1.05701573e-01 6.20220482e-01
3.76111865e-01 9.58466530e-02 -8.06062222e-01 -2.60848314e-01
8.97603571e-01 4.53886092e-01 -3.34586263e-01 -3.12484503e-01
-9.80289280e-01 3.48263621e-01 5.95054269e-01 1.29051656e-02
-1.78611547e-01 5.41514605e-02 8.95464122e-02 2.46573806e-01
5.48029184e-01 2.45180875e-01 -2.14661449e-01 2.66386151e-01
-8.91309559e-01 4.31550086e-01 6.23276174e-01 1.03792202e+00
5.23747325e-01 4.01189439e-02 -4.00486559e-01 1.00152230e+00
-1.91029662e-03 1.06899703e+00 9.42999572e-02 -7.12126732e-01
4.25540984e-01 6.57537341e-01 4.67303485e-01 -1.12823677e+00
-2.32721955e-01 -3.88202906e-01 -7.78661907e-01 -1.68884490e-02
7.07483053e-01 -2.29207262e-01 -1.01676142e+00 1.51792276e+00
2.75486588e-01 1.43119931e-01 -2.69502431e-01 1.07453215e+00
8.15918565e-01 7.84233138e-02 -1.13057233e-01 1.53146431e-01
1.55377316e+00 -1.00725579e+00 -2.78691798e-01 -5.93428314e-01
-2.01779470e-01 -6.78972423e-01 5.76612711e-01 3.30966830e-01
-1.00384617e+00 -9.16029394e-01 -8.48178089e-01 -1.44259602e-01
-2.88699061e-01 6.49463654e-01 4.31510121e-01 1.10343730e+00
-1.25332606e+00 4.52753723e-01 -3.58512402e-01 -6.48186803e-01
6.92641735e-01 5.02396107e-01 -6.69813633e-01 -5.19953966e-01
-8.32310736e-01 5.36114037e-01 -3.11336756e-01 4.03888673e-01
-1.00323427e+00 -7.04094529e-01 -9.07280743e-01 7.10017467e-03
2.80737698e-01 -5.86378574e-01 8.17550480e-01 -1.40110075e+00
-1.16041028e+00 1.27522635e+00 -2.25565270e-01 -1.25327513e-01
8.54530871e-01 -5.59539318e-01 -5.87513566e-01 1.96055308e-01
-4.12287861e-02 5.71629226e-01 1.47641718e+00 -1.04795849e+00
-2.66396523e-01 -9.05553341e-01 -4.23662513e-02 1.84730694e-01
-4.95730788e-01 5.07287741e-01 -6.01012528e-01 -5.68001866e-01
-3.13402742e-01 -1.06072557e+00 1.77939266e-01 2.19213799e-01
-2.97090948e-01 -9.57265645e-02 4.12115157e-01 -9.26218808e-01
6.02647245e-01 -2.35080218e+00 9.46590975e-02 1.99837282e-01
1.27943590e-01 3.77090126e-01 -3.26503098e-01 2.48704240e-01
-1.96894422e-01 -2.25314289e-01 2.21996576e-01 -7.94321477e-01
1.56090602e-01 -8.44416395e-02 -1.91548526e-01 6.11047506e-01
1.27756149e-01 9.88328040e-01 -5.14883459e-01 -8.71021450e-02
6.11413456e-02 7.27773368e-01 -2.91611463e-01 3.58142883e-01
3.82796615e-01 2.21480474e-01 -1.32447369e-02 1.13508248e+00
1.06794930e+00 2.42210533e-02 1.16969854e-01 -3.08528960e-01
2.29572654e-01 -3.74983400e-01 -1.46312332e+00 1.45026886e+00
2.69891159e-03 6.00889146e-01 2.88695753e-01 -4.40412164e-01
8.58871043e-01 3.14077169e-01 2.44004861e-01 -2.89400846e-01
1.20491823e-02 -2.09164005e-02 -3.30225438e-01 -3.32567900e-01
3.29627663e-01 7.38570392e-02 1.94130033e-01 7.83527568e-02
1.83331951e-01 6.26250625e-01 6.02260605e-02 -1.22258581e-01
1.00735795e+00 -2.52565831e-01 -2.68855272e-03 -2.91454554e-01
6.09039724e-01 -7.19923973e-01 4.30412769e-01 9.17807877e-01
-4.85156387e-01 1.13108575e+00 2.77494248e-02 -1.10670066e+00
-1.09684038e+00 -1.24475193e+00 -2.16504633e-01 1.24257863e+00
-1.61046386e-02 -1.62505031e-01 -9.00889277e-01 -7.19078720e-01
3.31964076e-01 -2.25323930e-01 -7.35732257e-01 2.79839098e-01
-5.74793637e-01 -6.97699964e-01 6.49991095e-01 6.13653600e-01
9.00591791e-01 -7.61981606e-01 -1.38094164e-02 -3.80870134e-01
4.38914122e-03 -1.45405972e+00 -1.06874263e+00 -9.15416062e-01
-6.13901131e-02 -1.24972880e+00 -1.32405591e+00 -7.82898962e-01
1.02636015e+00 5.64287066e-01 1.17234993e+00 2.98359364e-01
-6.67510629e-01 8.18003297e-01 -1.12447031e-01 -2.30616316e-01
1.92527860e-01 -7.18160942e-02 6.02490425e-01 8.33217204e-01
4.10523385e-01 -3.00248623e-01 -9.50802624e-01 5.19844890e-01
-4.47161168e-01 -4.62820411e-01 2.72977769e-01 6.50837064e-01
-1.45801812e-01 -1.36322677e-01 2.06085756e-01 -5.46713650e-01
1.66569814e-01 8.61232877e-02 -5.30792058e-01 4.29214060e-01
4.90751565e-02 -5.13694525e-01 3.43457818e-01 -5.52577972e-01
-1.02627230e+00 4.12974030e-01 1.52916163e-01 -4.62647378e-01
-3.24659437e-01 -3.67511570e-01 -4.70157504e-01 -7.10844457e-01
9.80018377e-01 9.95109081e-02 -6.10835887e-02 -4.57068652e-01
-2.05390621e-02 8.30487847e-01 7.12738872e-01 -5.06396532e-01
1.11220813e+00 6.24522865e-01 -5.16448207e-02 -1.03952241e+00
-8.35911274e-01 -5.88937998e-01 -1.05973482e+00 -4.37982619e-01
5.60629487e-01 -1.51637983e+00 -9.63559330e-01 1.05213642e+00
-1.07131100e+00 9.49908197e-02 8.81634206e-02 2.01281592e-01
1.93702672e-02 3.22067499e-01 -5.65221786e-01 -1.03044116e+00
-3.87510449e-01 -8.46495628e-01 1.50281370e+00 3.34702760e-01
1.32909030e-01 -5.54909408e-01 -2.54119009e-01 7.02209532e-01
5.01280963e-01 4.20684725e-01 -3.16774696e-02 -4.60380346e-01
-6.70707941e-01 -6.55884206e-01 -3.49023014e-01 3.67595553e-01
2.04959765e-01 -2.03976557e-01 -1.48632562e+00 -7.51993835e-01
-3.95381272e-01 -5.02909720e-01 8.61126781e-01 1.73465118e-01
8.75475645e-01 -4.32793260e-01 -2.92677492e-01 7.81868100e-01
1.23027503e+00 -3.38323593e-01 6.36981547e-01 -1.92304164e-01
1.00388086e+00 6.68231189e-01 -5.84857725e-02 5.30702353e-01
3.01794887e-01 8.60562146e-01 6.35649562e-02 -2.36247495e-01
-1.94732472e-01 -2.26845160e-01 4.88377184e-01 -5.16222231e-02
-6.38917148e-01 -1.67114779e-01 -8.35885406e-01 3.17206502e-01
-1.56477582e+00 -1.17497563e+00 3.15851748e-01 2.47548747e+00
3.91248167e-01 -4.93900031e-01 5.27437687e-01 -1.49562284e-01
1.06898284e+00 2.40314543e-01 -4.77693260e-01 3.17503780e-01
-3.21205735e-01 1.16460703e-01 4.92337316e-01 3.58140767e-01
-1.45104194e+00 6.75993562e-01 6.66201115e+00 4.01912332e-01
-8.95015836e-01 -1.44888610e-01 7.08217740e-01 -4.36077476e-01
5.15964508e-01 -7.01909065e-01 -1.14824164e+00 4.87076610e-01
5.20536602e-01 3.08252692e-01 1.02363682e+00 8.93996835e-01
-3.75810951e-01 4.42414552e-01 -1.40379524e+00 1.65298164e+00
8.52664530e-01 -9.53482628e-01 8.53156447e-02 1.21544041e-01
8.39106917e-01 -6.30953535e-02 3.71768266e-01 2.09832296e-01
4.47121263e-03 -1.36722541e+00 4.20324266e-01 6.90358579e-01
9.81763482e-01 -6.91874266e-01 6.14204526e-01 -2.55423617e-02
-1.18708849e+00 -3.01499814e-01 -5.43048918e-01 -1.63800210e-01
-2.99818188e-01 2.34218240e-01 -6.15877330e-01 3.09807688e-01
1.01351881e+00 7.49067128e-01 -8.22013199e-01 7.94380486e-01
-2.84300949e-02 1.09954447e-01 -3.32963854e-01 5.03098130e-01
-4.03057277e-01 2.09855437e-01 4.09394920e-01 1.03266442e+00
2.70037502e-01 8.26528072e-02 3.16949964e-01 5.25400400e-01
-5.87706029e-01 -2.32783407e-01 -4.89435911e-01 1.21572524e-01
5.25673330e-01 1.48004484e+00 -1.77718014e-01 -1.17545478e-01
-3.19021940e-01 1.22381377e+00 3.19465935e-01 6.19543910e-01
-4.97413099e-01 7.94114098e-02 1.12750089e+00 3.30827624e-01
2.58734554e-01 -1.57310486e-01 3.37982565e-01 -1.51694179e+00
3.73800129e-01 -1.13260818e+00 4.03547704e-01 -6.47664011e-01
-1.71389651e+00 6.98894501e-01 -2.81223416e-01 -9.78307426e-01
-9.52219144e-02 -1.00851488e+00 -3.95755887e-01 9.89655375e-01
-1.32898283e+00 -1.69410968e+00 -7.23434687e-01 7.94042766e-01
4.18070674e-01 -5.83035350e-01 7.23027349e-01 6.32454991e-01
-8.00842285e-01 1.23545420e+00 4.20390442e-02 7.08901227e-01
9.97000933e-01 -1.00950921e+00 7.16616631e-01 9.71508205e-01
1.91714168e-01 8.06203246e-01 3.22328895e-01 -3.80433977e-01
-1.70230150e+00 -1.07713807e+00 8.14666152e-01 -1.05819225e+00
2.17503612e-03 -7.72754490e-01 -4.95030522e-01 9.11192596e-01
-6.27946183e-02 6.29126430e-01 5.74527800e-01 2.79784024e-01
-6.64004922e-01 -4.51226085e-01 -1.35610771e+00 4.48245317e-01
1.48676181e+00 -7.32982576e-01 -2.03132838e-01 5.20369232e-01
4.25622538e-02 -4.56713796e-01 -7.68006742e-01 3.11309844e-01
9.11338508e-01 -9.04350400e-01 1.53018403e+00 -5.08650422e-01
7.60071725e-02 -1.77713871e-01 -1.96906954e-01 -9.98664796e-01
-5.61383486e-01 -6.68328285e-01 9.48625132e-02 1.53429699e+00
-1.61958754e-01 -7.42599547e-01 8.11746478e-01 1.18764472e+00
5.04530370e-01 -4.98925932e-02 -9.98777330e-01 -7.91167259e-01
-4.76174295e-01 2.49291092e-01 8.32160890e-01 7.71830499e-01
-5.66619515e-01 1.34676099e-01 -9.27786648e-01 4.06523913e-01
9.07153249e-01 -5.93818799e-02 1.05320382e+00 -1.37256432e+00
-1.86575532e-01 5.50195910e-02 -6.20525599e-01 -7.44421840e-01
1.10453844e-01 -5.34497619e-01 -1.17034070e-01 -7.42512226e-01
4.80936289e-01 -2.16833055e-02 -5.73860295e-02 4.98489171e-01
-3.15068901e-01 8.08331549e-01 4.09387827e-01 3.48206311e-01
-6.61499202e-01 2.26037890e-01 9.18037534e-01 -3.57836396e-01
2.54187644e-01 6.18819036e-02 -6.98775887e-01 8.00976336e-01
2.90970832e-01 -3.03350948e-02 -1.56982448e-02 -7.13532865e-01
1.08802617e-02 -3.53276014e-01 8.01010370e-01 -1.25408983e+00
2.09656432e-01 2.20362812e-01 1.26870048e+00 -1.40691444e-01
7.42284000e-01 -7.45719254e-01 2.15333611e-01 2.66785681e-01
-2.31542856e-01 8.74751881e-02 2.83182897e-02 4.17659998e-01
1.18187897e-01 2.08854973e-01 8.19170535e-01 -3.19985151e-01
-5.61494291e-01 8.09569120e-01 3.12856734e-01 9.03669670e-02
5.82231283e-01 -2.49091551e-01 -5.37942827e-01 -2.92212039e-01
-5.32605648e-01 6.37490898e-02 6.32373810e-01 4.49413747e-01
4.14329737e-01 -1.49123216e+00 -1.07059109e+00 5.57879567e-01
1.96185350e-01 -3.62145096e-01 4.89951372e-01 3.15728724e-01
-4.26414609e-01 2.27294296e-01 -3.10354680e-01 -4.07548815e-01
-1.51632476e+00 4.88920689e-01 6.88897848e-01 2.26896048e-01
-3.29147369e-01 1.15654051e+00 3.18890274e-01 -3.87687862e-01
3.25175494e-01 3.18747550e-01 -1.13716379e-01 7.37263560e-02
1.17904162e+00 4.96387094e-01 7.04326993e-03 -1.14558876e+00
-6.73300743e-01 7.65486062e-01 -1.58914283e-01 2.60645777e-01
1.07682610e+00 -1.36064038e-01 -1.60523485e-02 -2.44402617e-01
1.03455937e+00 1.48927197e-01 -1.61691380e+00 -3.96674514e-01
-4.37512308e-01 -1.09105659e+00 -3.49423170e-01 -7.47977257e-01
-1.13551021e+00 7.27136910e-01 9.76600707e-01 -9.80512947e-02
9.45149124e-01 -2.58277595e-01 6.69547081e-01 3.96146983e-01
5.44333041e-01 -9.14373219e-01 1.91134989e-01 3.62992436e-01
1.25833738e+00 -1.48804963e+00 4.59676199e-02 -3.89007658e-01
-4.52022403e-01 1.05187786e+00 6.64735138e-01 -1.84752390e-01
3.85898530e-01 -6.40685856e-02 2.26192236e-01 1.09007865e-01
-1.65714458e-01 -2.04387873e-01 6.72113001e-01 7.88630009e-01
9.97675657e-02 1.05790913e-01 6.42374098e-01 4.75699276e-01
-3.24325204e-01 -2.17913702e-01 -6.54793382e-02 3.43916863e-01
-5.90422079e-02 -7.24530280e-01 -6.83170438e-01 8.33888799e-02
-4.74037260e-01 -1.02728091e-01 -7.73917675e-01 4.68328655e-01
3.70630682e-01 8.93867254e-01 1.39255837e-01 -2.53015935e-01
4.24514234e-01 1.67855471e-01 7.87604034e-01 -4.71240044e-01
-6.89325511e-01 -2.88241595e-01 5.21724559e-02 -5.78291476e-01
-3.96275401e-01 -8.01796675e-01 -3.05323988e-01 -5.68586051e-01
1.60143644e-01 -4.01860267e-01 2.92299300e-01 9.70214665e-01
3.69591743e-01 -1.42442688e-01 8.63080263e-01 -1.36583161e+00
-6.08031034e-01 -9.26774144e-01 -7.01997519e-01 7.89876282e-01
5.58793783e-01 -6.53310716e-01 -3.19535792e-01 1.43595845e-01]
|
[14.30441951751709, 0.9856342077255249]
|
a2affe0e-7603-4e8e-8068-fa50c0967e98
|
qasr-qcri-aljazeera-speech-resource-a-large
|
2106.13000
| null |
https://arxiv.org/abs/2106.13000v1
|
https://arxiv.org/pdf/2106.13000v1.pdf
|
QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpus
|
We introduce the largest transcribed Arabic speech corpus, QASR, collected from the broadcast domain. This multi-dialect speech dataset contains 2,000 hours of speech sampled at 16kHz crawled from Aljazeera news channel. The dataset is released with lightly supervised transcriptions, aligned with the audio segments. Unlike previous datasets, QASR contains linguistically motivated segmentation, punctuation, speaker information among others. QASR is suitable for training and evaluating speech recognition systems, acoustics- and/or linguistics- based Arabic dialect identification, punctuation restoration, speaker identification, speaker linking, and potentially other NLP modules for spoken data. In addition to QASR transcription, we release a dataset of 130M words to aid in designing and training a better language model. We show that end-to-end automatic speech recognition trained on QASR reports a competitive word error rate compared to the previous MGB-2 corpus. We report baseline results for downstream natural language processing tasks such as named entity recognition using speech transcript. We also report the first baseline for Arabic punctuation restoration. We make the corpus available for the research community.
|
['Ahmed Ali', 'Shammur Absar Chowdhury', 'Amir Hussein', 'Hamdy Mubarak']
|
2021-06-24
| null | null | null | null |
['dialect-identification', 'punctuation-restoration', 'speaker-identification']
|
['natural-language-processing', 'natural-language-processing', 'speech']
|
[ 1.51301488e-01 2.76677907e-01 1.35009795e-01 -7.74584293e-01
-1.69526029e+00 -6.91843510e-01 3.03057432e-01 6.69192374e-02
-3.79202753e-01 3.39824855e-01 7.23991990e-01 -5.85101128e-01
2.33218700e-01 -2.15697512e-01 -5.46874106e-01 -6.29608691e-01
-1.18805356e-01 7.36952603e-01 9.59389210e-02 -6.50598526e-01
1.19688615e-01 2.02215210e-01 -1.10128140e+00 5.68462908e-01
7.99875200e-01 6.66569352e-01 4.33984488e-01 8.36972356e-01
-6.82600914e-03 6.85743988e-01 -9.70297217e-01 -4.76915240e-01
-1.99772626e-01 -4.28019762e-01 -1.30148804e+00 3.74070078e-01
3.16299379e-01 -2.02905715e-01 -1.80250049e-01 6.75989747e-01
7.23455012e-01 9.06274468e-02 4.44207460e-01 -8.22874129e-01
-5.77439904e-01 1.37796247e+00 -5.52254505e-02 3.57805878e-01
5.51049531e-01 -2.03628555e-01 9.43129897e-01 -1.13537931e+00
5.83959699e-01 1.60246289e+00 4.57564294e-01 6.22152030e-01
-9.40443635e-01 -3.52572978e-01 8.93008634e-02 2.61037201e-01
-1.54815423e+00 -1.43285656e+00 4.61183012e-01 3.81413326e-02
1.19310355e+00 3.77690226e-01 -1.49662361e-01 1.31039393e+00
-5.49173415e-01 1.05511284e+00 7.51947284e-01 -9.85496342e-01
1.49255976e-01 -3.45589183e-02 2.49649644e-01 5.34979463e-01
-6.56581521e-01 -4.55807418e-01 -8.68196964e-01 -8.15945566e-02
6.23264313e-02 -1.03706729e+00 -4.82324272e-01 6.19143128e-01
-1.27499831e+00 6.91512704e-01 -3.09495956e-01 2.73723572e-01
-3.05780202e-01 -3.65843117e-01 6.33881569e-01 6.65417016e-01
4.68633562e-01 1.33986743e-02 -7.04639018e-01 -4.34412628e-01
-7.31941164e-01 -2.48521730e-01 8.27308178e-01 1.17133522e+00
3.60070050e-01 5.05591452e-01 1.64357275e-01 1.60566759e+00
6.16944313e-01 9.95811939e-01 8.47194850e-01 -9.72619474e-01
7.44909048e-01 -9.99585763e-02 -1.14134349e-01 -3.13796610e-01
-3.04497033e-01 1.89081937e-01 -2.38712728e-01 -3.89397115e-01
6.22280478e-01 -3.57291728e-01 -1.03475821e+00 1.50363266e+00
2.44856656e-01 -3.77590716e-01 6.24281466e-01 8.20694029e-01
8.94733489e-01 1.33098865e+00 -2.21707642e-01 -2.54322201e-01
1.58412802e+00 -1.10282302e+00 -8.81049216e-01 -4.56237853e-01
5.65666556e-01 -1.35112429e+00 1.35928595e+00 6.79093480e-01
-1.31153297e+00 -3.13454568e-01 -7.55218089e-01 -1.15280248e-01
-3.27839494e-01 3.64693582e-01 -1.10124886e-01 1.26302779e+00
-1.28199041e+00 -2.19584808e-01 -8.36103499e-01 -4.73629951e-01
-1.77055925e-01 -2.11233255e-02 -3.40220183e-01 -1.11318536e-01
-1.35002089e+00 9.14512277e-01 7.21675158e-02 8.60893354e-02
-9.65157211e-01 -8.03325996e-02 -1.18567991e+00 -2.42395088e-01
-7.64149502e-02 4.91454899e-01 1.79643261e+00 -9.38296795e-01
-1.88070107e+00 9.73127365e-01 -5.18694580e-01 -6.87009931e-01
-1.22110732e-01 -3.94970365e-02 -9.59449410e-01 2.49129593e-01
8.41423869e-02 4.92296964e-01 8.51056635e-01 -1.00406659e+00
-8.38575184e-01 -5.13385415e-01 -6.66028976e-01 3.53077173e-01
-1.96266040e-01 8.43344629e-01 -2.90626585e-01 -8.60202134e-01
3.32382590e-01 -7.54967630e-01 7.48402774e-02 -9.69717383e-01
-4.44815427e-01 -2.50611097e-01 8.42296362e-01 -1.46965921e+00
1.21050918e+00 -2.28269887e+00 3.90068926e-02 1.89391255e-01
-8.32093596e-01 1.92606360e-01 -3.83818388e-01 5.78765094e-01
3.23607475e-02 -2.58164164e-02 -4.24194723e-01 -6.76316440e-01
6.28944784e-02 2.72689044e-01 -5.89872479e-01 4.98761117e-01
4.51972514e-01 4.42838639e-01 -5.57543635e-01 -8.55937302e-02
-4.00453955e-02 5.24607301e-01 -6.44574538e-02 5.84267266e-02
3.93538326e-02 4.23902690e-01 2.10418835e-01 1.04213119e+00
5.32007694e-01 7.46527135e-01 1.09752104e-01 2.66927332e-01
-1.95510253e-01 1.36145604e+00 -1.06320643e+00 1.72756159e+00
-7.20422089e-01 6.55791521e-01 6.76731288e-01 -9.08947647e-01
9.13624585e-01 8.61025453e-01 -1.52565628e-01 -5.69353223e-01
1.18450917e-01 5.40191650e-01 -5.35720140e-02 -1.78372741e-01
8.17863584e-01 8.06596652e-02 -4.33610648e-01 4.93243814e-01
3.48687440e-01 -3.58740538e-01 8.67270529e-02 1.56964600e-01
8.51109505e-01 -4.26905125e-01 1.47259101e-01 -2.90013283e-01
7.61764348e-01 2.04120725e-01 1.70956448e-01 5.41927934e-01
-2.95478791e-01 8.85926068e-01 4.92911078e-02 2.29047820e-01
-7.59663880e-01 -1.13356209e+00 -3.56851906e-01 1.70483625e+00
-4.34600502e-01 -4.11922693e-01 -1.27109444e+00 -4.79040653e-01
-4.78639305e-01 1.00428927e+00 6.16769195e-02 2.22782284e-01
-1.10473990e+00 -6.24743283e-01 1.41523468e+00 1.70569226e-01
2.52963364e-01 -1.37512112e+00 4.18060392e-01 5.78445673e-01
-7.86283672e-01 -1.16506612e+00 -9.28580582e-01 3.17853034e-01
-3.60112607e-01 -6.81621075e-01 -7.11090982e-01 -1.60066795e+00
8.28611329e-02 1.29528165e-01 9.17775095e-01 -2.64027804e-01
2.08243191e-01 5.10371506e-01 -7.86870182e-01 -3.41102362e-01
-1.38667560e+00 3.24592918e-01 2.61105865e-01 -1.14769693e-02
3.71847034e-01 -2.04278976e-02 7.46960044e-02 5.41326880e-01
-6.65024161e-01 -6.93705976e-01 2.42777467e-01 7.91756034e-01
3.72961909e-01 -3.14390451e-01 1.14750648e+00 -5.22763312e-01
6.68233693e-01 -3.30599457e-01 -3.78609508e-01 1.64601460e-01
-1.25773717e-02 -2.20849335e-01 5.53765893e-01 -1.97788551e-01
-1.46091270e+00 2.93409199e-01 -1.18178546e+00 5.40523887e-01
-6.00702941e-01 6.63179815e-01 -6.81281567e-01 3.83899480e-01
9.47252810e-01 4.94934738e-01 5.98766878e-02 -7.31404662e-01
5.73269248e-01 1.79397142e+00 9.75380838e-01 -4.08304602e-01
2.25399122e-01 -2.32306160e-02 -1.17827559e+00 -1.56627238e+00
-4.20239300e-01 -7.55846977e-01 -5.82605541e-01 3.07839494e-02
5.58093250e-01 -1.20002985e+00 -3.02332312e-01 9.87082839e-01
-1.27465284e+00 -4.73166734e-01 -1.04194008e-01 1.79834291e-01
-6.01956844e-01 4.85957026e-01 -1.18598402e+00 -8.59366119e-01
-4.15740430e-01 -1.24037576e+00 1.20238328e+00 -3.36656034e-01
-2.51585603e-01 -7.33390391e-01 -3.63225788e-02 7.77379453e-01
2.37820327e-01 -7.30661094e-01 5.58591783e-01 -8.85100126e-01
-3.99531052e-02 1.46706060e-01 3.65284562e-01 6.49509907e-01
3.94447118e-01 -6.64435923e-02 -1.23328626e+00 -2.73575306e-01
-3.29583809e-02 -5.93324423e-01 6.29088283e-01 3.44985038e-01
3.73175681e-01 -5.64612627e-01 2.66555607e-01 2.94491220e-02
4.12255287e-01 3.67208302e-01 6.68859482e-01 2.86520422e-01
3.63376647e-01 8.99380267e-01 7.17056215e-01 2.37614900e-01
8.12086642e-01 4.65384483e-01 9.21837636e-04 2.94180602e-01
-3.65970731e-01 5.75421797e-03 1.28310108e+00 1.69829595e+00
7.21450448e-01 -4.97426599e-01 -1.37125432e+00 1.06992972e+00
-1.31078804e+00 -6.29225552e-01 -3.07436287e-01 2.02851725e+00
1.47985649e+00 -2.28492111e-01 4.03188974e-01 4.36603993e-01
9.39648211e-01 8.99673700e-02 2.45784238e-01 -7.48559415e-01
-4.33382511e-01 2.17036784e-01 3.82605612e-01 1.09670711e+00
-1.17292380e+00 1.52658260e+00 6.26785231e+00 1.04685390e+00
-9.38708842e-01 4.01878327e-01 6.14970386e-01 3.49180877e-01
-1.20749809e-01 -2.28165135e-01 -1.08912444e+00 3.02557081e-01
1.95656431e+00 2.15675682e-01 6.18154705e-01 6.36970580e-01
3.10918808e-01 9.65296756e-03 -6.91159427e-01 7.55609512e-01
3.21428329e-01 -9.30194855e-01 -1.99078154e-02 -5.36674023e-01
3.59766811e-01 5.83935380e-01 -1.52892634e-01 3.07197630e-01
3.28303427e-01 -8.83096576e-01 1.25356019e+00 -2.01308340e-01
7.58044243e-01 -1.22171915e+00 5.93241334e-01 2.94040561e-01
-9.37526882e-01 2.70565927e-01 -1.93183526e-01 3.25281620e-01
5.18725395e-01 1.00387476e-01 -1.46931899e+00 3.40528250e-01
6.56437218e-01 3.65890831e-01 -5.87604523e-01 7.53396630e-01
-3.84428322e-01 1.64701796e+00 -4.26246673e-01 9.10854936e-02
2.29382291e-01 -2.75094472e-02 8.26655507e-01 1.78471577e+00
2.13875815e-01 -9.58859548e-02 7.84286261e-02 -6.45691901e-02
-2.52282321e-02 5.05463362e-01 -1.82639793e-01 -1.53836980e-01
7.78820932e-01 9.58797932e-01 -7.44437218e-01 -2.63294816e-01
-2.58979201e-01 1.18409801e+00 1.32728353e-01 3.77192855e-01
-3.98548245e-01 -8.01891863e-01 7.81807065e-01 -1.60278663e-01
3.35116178e-01 -5.75708628e-01 -1.29954487e-01 -8.62183392e-01
-2.17725374e-02 -1.34432507e+00 3.90402853e-01 -6.33719742e-01
-1.35022402e+00 1.13531172e+00 -5.58246732e-01 -7.36785710e-01
-7.36451149e-01 -7.44890749e-01 -2.02565491e-01 1.02802205e+00
-1.61458576e+00 -1.04922962e+00 2.64781475e-01 5.18011868e-01
1.11830711e+00 -5.15089393e-01 1.28154480e+00 4.49695766e-01
-5.35318851e-01 6.72950327e-01 2.71900862e-01 5.75924456e-01
1.12913001e+00 -1.28294361e+00 9.84915495e-01 9.75921750e-01
4.38280195e-01 2.60780454e-01 6.23833001e-01 -4.11696821e-01
-1.15614140e+00 -1.14347923e+00 1.22566032e+00 -5.68766654e-01
8.71353090e-01 -6.45647824e-01 -1.09740615e+00 8.86866927e-01
6.42421126e-01 -5.18773079e-01 7.13884354e-01 1.65942311e-02
-1.02325134e-01 -1.60105646e-01 -1.04253244e+00 5.38749278e-01
6.78996921e-01 -7.56814778e-01 -8.22832584e-01 3.43153417e-01
1.03525043e+00 -5.64366102e-01 -6.37213588e-01 -1.48791239e-01
3.87099870e-02 -3.22928429e-01 6.98841274e-01 -4.27372187e-01
-2.68815845e-01 -2.16187909e-01 -5.05507708e-01 -1.79043305e+00
2.16496363e-01 -8.90810490e-01 5.44053137e-01 1.74219155e+00
9.84258950e-01 -5.14162421e-01 2.42026851e-01 9.30084847e-03
-7.94847369e-01 2.60137677e-01 -1.27313447e+00 -8.62734377e-01
2.67908096e-01 -8.73626590e-01 4.47957009e-01 9.41524565e-01
4.25999969e-01 4.60135788e-01 -7.23597780e-02 6.30084157e-01
1.92896053e-01 -4.02394772e-01 3.37524801e-01 -5.52797019e-01
-5.05661704e-02 -1.15104526e-01 -2.41576850e-01 -1.13797915e+00
4.17369992e-01 -9.09381330e-01 8.33779037e-01 -1.31647253e+00
-8.27843726e-01 -4.61051494e-01 3.95523936e-01 8.07611406e-01
1.53034359e-01 2.96686262e-01 -1.87931225e-01 -1.45379409e-01
-2.86617041e-01 6.30697310e-01 4.94894981e-01 -2.34989017e-01
-5.04418969e-01 1.56039447e-01 -3.68702441e-01 5.04131019e-01
1.02155125e+00 -5.44615209e-01 -1.21387735e-01 -6.66224301e-01
-4.56328869e-01 3.27054292e-01 -2.10149050e-01 -5.81465185e-01
1.93677112e-01 1.22674637e-01 -1.50692508e-01 -6.96244776e-01
3.64876360e-01 -3.20685387e-01 -5.43145716e-01 6.55791834e-02
-2.11867914e-01 1.02370165e-01 4.39099044e-01 -1.71503320e-01
-5.88733375e-01 -5.21731019e-01 7.47558236e-01 7.57817775e-02
-7.27922320e-01 -2.82125682e-01 -1.56676233e+00 3.84438157e-01
3.43974441e-01 1.20006777e-01 -5.14179885e-01 -7.27089345e-01
-8.70165348e-01 1.03369243e-01 4.86230180e-02 7.82366097e-01
5.55473387e-01 -9.99229848e-01 -1.16325867e+00 5.32129586e-01
1.23324066e-01 -8.60137418e-02 -1.12142764e-01 6.14776969e-01
-6.31407320e-01 4.07055348e-01 1.87330514e-01 -4.82334852e-01
-1.36729598e+00 -1.07485964e-03 2.31714189e-01 6.46090984e-01
-2.27298498e-01 9.52257097e-01 -5.22576869e-01 -1.00036311e+00
4.81659949e-01 -2.77194262e-01 -1.34173185e-01 3.00786972e-01
8.81919920e-01 3.55369031e-01 8.30546856e-01 -1.35817826e+00
-7.01499999e-01 -3.19413543e-01 -1.78579435e-01 -8.97390425e-01
1.11370885e+00 -8.27528119e-01 -1.72511503e-01 5.59191823e-01
9.23940420e-01 7.05059052e-01 -7.87739575e-01 -1.68017358e-01
5.64749658e-01 9.19594318e-02 5.31110317e-02 -1.11126339e+00
-5.03852308e-01 7.74397552e-01 2.75305450e-01 3.87078792e-01
9.55747783e-01 1.25026822e-01 9.41605747e-01 6.98378503e-01
2.27934927e-01 -1.46403921e+00 -2.59800643e-01 1.33844876e+00
1.19132328e+00 -1.25581598e+00 -9.28541899e-01 -4.35089052e-01
-1.01877820e+00 1.01900136e+00 2.67862976e-01 4.16291654e-01
6.67300045e-01 3.91185075e-01 1.02773416e+00 2.88042396e-01
-3.42988074e-01 -2.91162223e-01 2.80700862e-01 9.27477419e-01
6.67299509e-01 2.74517084e-03 -4.03560400e-02 8.31349075e-01
-9.47384238e-01 -1.05597794e+00 8.70426297e-01 8.87612760e-01
-7.40878701e-01 -1.29016614e+00 -7.51676202e-01 -1.57525912e-01
-5.67179203e-01 -5.63515365e-01 -5.70652127e-01 4.00190532e-01
-6.89614058e-01 1.68038309e+00 1.69636175e-01 1.28392220e-01
4.35627967e-01 5.49733341e-01 1.09898321e-01 -7.10224807e-01
-5.35058975e-01 4.65627819e-01 7.46050596e-01 -2.54740804e-01
-1.84395745e-01 -1.06872094e+00 -1.61687219e+00 -6.60287812e-02
-3.58818799e-01 4.07872707e-01 9.89981055e-01 8.93061996e-01
9.61874425e-02 1.30506635e-01 8.18997979e-01 -7.65005827e-01
-4.17650521e-01 -1.39462864e+00 -6.70712590e-01 -1.65615767e-01
5.69798410e-01 1.55489087e-01 -4.61764663e-01 3.76155883e-01]
|
[14.398089408874512, 6.824253559112549]
|
e1836370-5f9e-4593-8c21-9ec30a4013b4
|
uiu-net-u-net-in-u-net-for-infrared-small
|
2212.00968
| null |
https://arxiv.org/abs/2212.00968v1
|
https://arxiv.org/pdf/2212.00968v1.pdf
|
UIU-Net: U-Net in U-Net for Infrared Small Object Detection
|
Learning-based infrared small object detection methods currently rely heavily on the classification backbone network. This tends to result in tiny object loss and feature distinguishability limitations as the network depth increases. Furthermore, small objects in infrared images are frequently emerged bright and dark, posing severe demands for obtaining precise object contrast information. For this reason, we in this paper propose a simple and effective ``U-Net in U-Net'' framework, UIU-Net for short, and detect small objects in infrared images. As the name suggests, UIU-Net embeds a tiny U-Net into a larger U-Net backbone, enabling the multi-level and multi-scale representation learning of objects. Moreover, UIU-Net can be trained from scratch, and the learned features can enhance global and local contrast information effectively. More specifically, the UIU-Net model is divided into two modules: the resolution-maintenance deep supervision (RM-DS) module and the interactive-cross attention (IC-A) module. RM-DS integrates Residual U-blocks into a deep supervision network to generate deep multi-scale resolution-maintenance features while learning global context information. Further, IC-A encodes the local context information between the low-level details and high-level semantic features. Extensive experiments conducted on two infrared single-frame image datasets, i.e., SIRST and Synthetic datasets, show the effectiveness and superiority of the proposed UIU-Net in comparison with several state-of-the-art infrared small object detection methods. The proposed UIU-Net also produces powerful generalization performance for video sequence infrared small object datasets, e.g., ATR ground/air video sequence dataset. The codes of this work are available openly at \url{https://github.com/danfenghong/IEEE_TIP_UIU-Net}.
|
['Jocelyn Chanussot', 'Danfeng Hong', 'Xin Wu']
|
2022-12-02
| null | null | null | null |
['small-object-detection']
|
['computer-vision']
|
[ 3.30206782e-01 -2.43412748e-01 -3.08682412e-01 -1.86999708e-01
-6.07061028e-01 -1.18547574e-01 3.00798696e-02 -4.41236764e-01
-2.99784571e-01 4.84345138e-01 -1.49761230e-01 -1.81187794e-01
6.63986802e-02 -1.02032626e+00 -9.39542174e-01 -9.37313139e-01
8.78133699e-02 -3.41829687e-01 4.02133942e-01 -2.07568407e-01
-2.59076641e-03 3.49678904e-01 -1.51387405e+00 2.90820986e-01
8.98409128e-01 1.49779630e+00 7.79490948e-01 4.62682396e-01
2.31913254e-01 1.09055555e+00 -2.87138253e-01 1.68995857e-01
7.29310870e-01 -2.75235385e-01 -4.29868728e-01 -1.50974482e-01
7.68709302e-01 -9.54719603e-01 -7.46889949e-01 1.31727052e+00
6.90494418e-01 2.35535637e-01 2.05378100e-01 -1.00117540e+00
-8.95607531e-01 3.41600716e-01 -9.29253817e-01 5.92150986e-01
1.60387829e-01 3.35987717e-01 7.93492258e-01 -7.63748705e-01
3.64247590e-01 1.29010773e+00 7.11351991e-01 6.39467180e-01
-9.65818703e-01 -1.13519549e+00 1.66287944e-01 3.90988171e-01
-1.37092912e+00 -2.04749361e-01 8.42362761e-01 -3.07536840e-01
7.24142075e-01 1.81444064e-01 2.60411710e-01 9.88251448e-01
7.47728646e-02 1.03212273e+00 8.49612117e-01 -3.77735376e-01
2.43766792e-02 -1.07407786e-01 9.22734812e-02 8.62049103e-01
4.31369603e-01 3.26158106e-01 -3.30151469e-02 2.82270342e-01
1.20041227e+00 5.09315491e-01 -6.16996467e-01 8.54351148e-02
-1.19557524e+00 5.99475384e-01 1.03152132e+00 2.58529395e-01
-1.95027351e-01 3.61766100e-01 3.47608358e-01 2.82019705e-01
6.67424619e-01 -1.34815142e-01 -5.06244719e-01 5.05040288e-01
-3.83081764e-01 -4.52298857e-02 -4.65977974e-02 1.13841224e+00
9.85745430e-01 5.68249375e-02 -2.25747049e-01 9.52852488e-01
2.22178221e-01 4.35559064e-01 3.05198610e-01 -8.47521544e-01
6.88944340e-01 4.57207263e-01 4.90835980e-02 -6.59587562e-01
-3.96561503e-01 -5.30225396e-01 -1.02209604e+00 5.33402085e-01
1.89003035e-01 -2.67073393e-01 -1.00861871e+00 1.62970722e+00
4.29689109e-01 3.00006866e-01 -2.29878519e-02 1.19756722e+00
1.10481977e+00 8.75832677e-01 7.44005442e-02 -1.78896710e-02
1.31550837e+00 -1.12326825e+00 -2.64749080e-01 -4.06471789e-01
4.91945714e-01 -6.29433393e-01 9.83462214e-01 -2.02478245e-02
-8.03602993e-01 -1.23316932e+00 -9.29898024e-01 -1.35691136e-01
-7.61136413e-02 7.00803757e-01 6.39345109e-01 2.58673608e-01
-9.19897676e-01 4.21018422e-01 -5.53106606e-01 -4.31329429e-01
7.37669289e-01 3.00489575e-01 -3.00451875e-01 -5.80468476e-01
-1.23444784e+00 4.31712389e-01 8.45038533e-01 4.88796443e-01
-7.74379432e-01 -4.74154949e-01 -9.39520180e-01 -9.11597442e-03
6.55830503e-01 -7.61178315e-01 1.00828767e+00 -1.08339143e+00
-1.12100947e+00 6.31428182e-01 1.46848127e-01 -5.54945879e-03
4.25357699e-01 -2.83685207e-01 -3.75063211e-01 3.87451768e-01
2.41954893e-01 6.64252937e-01 8.81591260e-01 -1.24399137e+00
-8.47644210e-01 -3.34685266e-01 5.13528526e-01 2.07671955e-01
-3.77485216e-01 3.10442626e-01 -4.08735871e-01 -9.35013711e-01
3.26954164e-02 -6.42795205e-01 -4.46333643e-03 4.19651389e-01
-3.39614511e-01 -5.33414900e-01 1.16493714e+00 -5.06563544e-01
9.64406073e-01 -2.18694353e+00 -1.10249028e-01 -2.21171886e-01
2.44485036e-01 4.89360988e-01 -4.70441639e-01 -8.66964459e-02
-2.97623634e-01 -9.45934355e-02 6.52911980e-03 1.46093920e-01
-3.17115784e-01 -9.88888070e-02 -3.69174294e-02 6.15900517e-01
1.45916179e-01 8.74637783e-01 -1.14369106e+00 -4.51936424e-01
5.99562645e-01 4.23356742e-01 -3.14679533e-01 3.05302501e-01
-2.15210006e-01 2.80394316e-01 -8.03147674e-01 1.12498307e+00
9.07075405e-01 -8.72940570e-02 -2.87065804e-01 -7.32053518e-01
-3.33659202e-01 -3.17923546e-01 -9.74599123e-01 1.53302181e+00
-4.63597029e-01 5.51713169e-01 1.32246599e-01 -1.07298255e+00
8.54491174e-01 1.70122087e-01 3.92655909e-01 -8.63424480e-01
3.76362711e-01 1.61216691e-01 -1.93012506e-01 -6.57143652e-01
2.05488786e-01 -5.60818687e-02 3.94590914e-01 1.11930870e-01
-5.81475273e-02 3.63543242e-01 1.40696824e-01 -3.85010429e-02
8.11529517e-01 3.73519808e-01 2.77807891e-01 -8.38384330e-02
6.12839043e-01 -4.57600504e-01 8.62898231e-01 8.08505893e-01
-3.17068100e-01 7.97595859e-01 6.33996874e-02 -5.76879084e-01
-8.27619910e-01 -9.33533907e-01 -3.29823017e-01 1.21026504e+00
5.57730854e-01 -7.03053698e-02 -4.43255961e-01 -6.50378108e-01
-1.29311442e-01 3.53125453e-01 -7.25844502e-01 -1.91394269e-01
-5.62199116e-01 -6.78652525e-01 1.03462160e-01 8.38740528e-01
9.19036388e-01 -1.21529806e+00 -5.80911577e-01 2.33199708e-02
-2.69427866e-01 -1.17629671e+00 -5.40166378e-01 5.74194361e-03
-9.22656298e-01 -1.14293444e+00 -8.67119491e-01 -1.09491217e+00
6.03800952e-01 1.08356571e+00 8.10612082e-01 1.64984733e-01
-6.28802180e-01 9.98692736e-02 -5.79326153e-01 -4.40841228e-01
9.87096950e-02 -2.29016438e-01 1.23187592e-02 3.01861744e-02
3.55235428e-01 -3.28957766e-01 -1.14034688e+00 6.07361078e-01
-1.08237183e+00 2.62943566e-01 9.94086027e-01 9.20004368e-01
5.01863480e-01 1.96740761e-01 3.89732540e-01 -2.63961464e-01
-1.23415887e-01 -3.73780698e-01 -7.15922177e-01 2.09880084e-01
-1.71968356e-01 -4.41951066e-01 6.46372378e-01 -3.52252543e-01
-1.13814342e+00 -1.57935992e-02 3.59444506e-02 -8.46823573e-01
-9.16022360e-02 1.17634632e-01 -3.05831581e-01 -3.81969213e-01
6.16672039e-01 3.00957978e-01 -2.26816133e-01 -4.96095121e-01
2.05196232e-01 8.67508292e-01 5.40805519e-01 -4.44224238e-01
9.43361282e-01 5.19730926e-01 -2.77592182e-01 -9.43199337e-01
-1.25440204e+00 -7.82753289e-01 -5.66331923e-01 -3.67921352e-01
1.02461350e+00 -1.40398061e+00 -5.44646561e-01 7.71245420e-01
-1.15755296e+00 -5.38437784e-01 -2.25525737e-01 6.15517616e-01
-2.37996578e-01 4.60906416e-01 -7.85429120e-01 -7.00496793e-01
-5.43238163e-01 -9.51257408e-01 1.33652544e+00 6.73016965e-01
5.24392247e-01 -5.81222594e-01 -4.26138818e-01 6.19248092e-01
2.09067076e-01 2.40672156e-01 5.72263062e-01 1.10514976e-01
-9.56982851e-01 -2.89593667e-01 -1.01329756e+00 7.33572304e-01
2.28158742e-01 -2.26469070e-01 -1.13377011e+00 -5.83734214e-01
4.11625346e-03 -6.34679854e-01 1.26301086e+00 5.98045826e-01
1.64453387e+00 -2.35296458e-01 -4.14817274e-01 9.77106690e-01
1.66999173e+00 1.84423819e-01 7.76947081e-01 6.21226609e-01
1.14763558e+00 2.61985153e-01 9.04636979e-01 2.43938819e-01
-2.06419993e-02 6.29061401e-01 6.43029988e-01 -4.70054418e-01
-2.16441572e-01 2.45264955e-02 5.15728474e-01 1.70696631e-01
-7.50178695e-01 -2.70809270e-02 -4.98425931e-01 1.81318849e-01
-1.78792059e+00 -1.43854272e+00 9.93892476e-02 2.07111073e+00
5.95779300e-01 1.18851056e-02 -1.16925500e-02 -9.79132056e-02
8.94704580e-01 4.48381484e-01 -5.83567798e-01 4.54215974e-01
-1.89100757e-01 -1.21454541e-02 7.45330274e-01 1.24331638e-01
-1.41766250e+00 7.18513370e-01 4.46080875e+00 1.07188833e+00
-1.17132628e+00 2.18063310e-01 6.76602244e-01 -2.21676722e-01
3.01544685e-02 -2.65664369e-01 -8.25281322e-01 4.50191200e-01
3.15126687e-01 2.60489583e-01 1.62457809e-01 9.99001145e-01
3.17296505e-01 -5.86211532e-02 -7.67651796e-01 1.14631927e+00
9.28509459e-02 -1.19133651e+00 1.98779441e-02 -1.95805386e-01
7.05916286e-01 4.85497504e-01 -9.99678150e-02 2.62935817e-01
-1.64458871e-01 -7.30106235e-01 9.36062098e-01 1.23140603e-01
1.24524426e+00 -6.98112130e-01 8.17668617e-01 1.64140463e-01
-1.72654998e+00 -6.52310073e-01 -1.02129042e+00 1.58478662e-01
-8.51763561e-02 5.05739391e-01 -6.82664588e-02 8.29544663e-01
1.12980092e+00 1.04513896e+00 -5.31985223e-01 1.01526070e+00
-2.01530978e-01 4.15425032e-01 -1.19132333e-01 2.18227640e-01
2.88244158e-01 -5.14506638e-01 4.49703515e-01 1.06952715e+00
2.59103000e-01 4.73093510e-01 4.34374183e-01 8.61405492e-01
-2.00148061e-01 -2.76868522e-01 -5.22342980e-01 2.83386886e-01
1.81087911e-01 1.50716865e+00 -5.88692963e-01 -3.59302580e-01
-7.77052402e-01 1.05373788e+00 1.82226241e-01 5.52402496e-01
-9.23639357e-01 -7.15621352e-01 6.57790661e-01 5.10720871e-02
4.80679601e-01 -9.70091224e-02 2.55400717e-01 -1.28691220e+00
2.26430401e-01 -6.65967822e-01 5.27653515e-01 -1.20074499e+00
-1.19642472e+00 7.21651256e-01 -1.22257143e-01 -1.57090831e+00
5.37401557e-01 -8.78066659e-01 -7.97288477e-01 7.70542443e-01
-1.84616125e+00 -1.43599796e+00 -9.55610871e-01 6.56406701e-01
8.79946589e-01 -2.05265898e-02 2.29174361e-01 4.83617187e-01
-8.69348586e-01 4.97853070e-01 3.70951176e-01 5.22987962e-01
5.74791491e-01 -8.62870097e-01 2.77032197e-01 9.64937925e-01
-2.47725785e-01 4.19452429e-01 2.38912627e-01 -5.10840416e-01
-1.32039189e+00 -1.65895498e+00 1.22960381e-01 -2.80423582e-01
5.61255574e-01 -5.24191380e-01 -7.95621514e-01 6.47680819e-01
-1.93753988e-01 6.45536900e-01 1.97015196e-01 -3.68287653e-01
-3.61810088e-01 -4.54850912e-01 -9.64421153e-01 1.96723923e-01
1.32939279e+00 -7.05637753e-01 -3.90651673e-01 5.53825200e-01
8.31721425e-01 -2.30549231e-01 -7.15589404e-01 5.61859787e-01
6.12637043e-01 -1.18517399e+00 1.53828549e+00 -1.79260656e-01
5.29780507e-01 -5.29435217e-01 -1.70701921e-01 -8.05627227e-01
-6.05644226e-01 -3.32429975e-01 -1.03689790e-01 1.04134393e+00
-1.25387475e-01 -7.48766661e-01 5.56790173e-01 1.32253796e-01
-2.89707482e-01 -7.51956224e-01 -6.96203351e-01 -8.54825735e-01
-3.51022065e-01 -3.04120719e-01 1.10469267e-01 8.57271552e-01
-6.71038270e-01 2.51923144e-01 -3.30282867e-01 3.48842204e-01
1.05309904e+00 2.69643724e-01 4.84055817e-01 -1.01214898e+00
-2.07182556e-01 -3.40620816e-01 -5.37477851e-01 -1.21944618e+00
-4.55321409e-02 -7.42377281e-01 1.12425856e-01 -1.72509992e+00
3.45627338e-01 -6.08787656e-01 -6.12105072e-01 5.21657407e-01
-6.26582086e-01 5.33572197e-01 1.21325098e-01 3.64778697e-01
-6.98643982e-01 7.23244429e-01 1.37850010e+00 -4.70116854e-01
-3.09724659e-02 4.09789532e-02 -6.45290673e-01 5.72878182e-01
9.01538432e-01 -1.99489966e-01 -3.57369483e-01 -5.26075602e-01
-2.46139050e-01 -2.00955451e-01 7.28438556e-01 -1.18613267e+00
1.02754369e-01 -2.09705666e-01 8.47214282e-01 -5.09503603e-01
8.26396197e-02 -6.19578302e-01 -3.15604120e-01 5.49842954e-01
4.09994759e-02 -3.88824850e-01 1.88089579e-01 8.07847381e-01
-1.06274612e-01 -3.07861087e-03 1.10342002e+00 -2.61499226e-01
-1.27285361e+00 7.98084319e-01 1.18232809e-01 -1.73490614e-01
1.07705283e+00 -4.56436753e-01 -5.79964161e-01 7.46833906e-02
2.73521314e-03 2.78193444e-01 4.25453693e-01 5.96130908e-01
7.34461665e-01 -1.30308580e+00 -7.55659759e-01 1.68562323e-01
4.45193589e-01 5.47570348e-01 7.10824072e-01 8.75151575e-01
-5.98457634e-01 1.84404835e-01 -2.59924233e-01 -6.52443349e-01
-1.35686827e+00 7.30739474e-01 5.17244279e-01 4.44934487e-01
-9.59867716e-01 1.06399286e+00 8.31635058e-01 -1.66699603e-01
1.76767841e-01 -5.70157945e-01 -1.07223429e-01 -2.55692750e-01
9.23897147e-01 4.19278115e-01 -3.53086501e-01 -4.85762477e-01
-2.31634140e-01 9.01449740e-01 -7.76371509e-02 7.92666912e-01
1.36714971e+00 -2.21062407e-01 -2.06432596e-01 1.03405409e-01
1.27741885e+00 -4.55991775e-01 -1.49340439e+00 -4.33077037e-01
-5.86873829e-01 -7.41694093e-01 3.96565467e-01 -4.60276008e-01
-1.31363463e+00 7.40346432e-01 9.53563213e-01 -2.44405597e-01
1.39022541e+00 1.31137535e-01 7.94105291e-01 5.48632145e-01
3.89846325e-01 -9.10740316e-01 2.31642589e-01 2.20955297e-01
7.41277874e-01 -1.69845331e+00 1.35359824e-01 -3.94048542e-01
-2.42080107e-01 1.19733596e+00 1.08639276e+00 7.33197704e-02
2.93155223e-01 -7.67255053e-02 2.24952519e-01 -4.90209125e-02
-2.66959727e-01 -4.59202081e-01 2.95993030e-01 6.52608633e-01
3.95633698e-01 -1.76469252e-01 1.79873616e-01 2.20003590e-01
5.43468714e-01 2.32311621e-01 2.01287672e-01 9.10367966e-01
-8.09277654e-01 -5.46823800e-01 -5.37515700e-01 5.79892159e-01
-2.34525979e-01 -2.11947978e-01 7.89301246e-02 6.96347117e-01
4.33650941e-01 7.83118486e-01 -1.55526578e-01 -5.28838694e-01
2.17106342e-01 -7.31099606e-01 5.67397296e-01 -4.48602349e-01
-7.16039762e-02 2.20869817e-02 -1.51783213e-01 -8.36695731e-01
-5.80298901e-01 -2.22416416e-01 -1.09156096e+00 -1.53904215e-01
-6.07579947e-01 -6.65926859e-02 3.53072196e-01 5.86416662e-01
-1.06470780e-02 6.62361264e-01 8.43148112e-01 -1.35282743e+00
-6.18820012e-01 -1.10430443e+00 -7.57860720e-01 2.83484191e-01
4.82388794e-01 -6.36777163e-01 -3.81368399e-01 -9.31826308e-02]
|
[9.090545654296875, -0.889919102191925]
|
cb15001f-5da4-4bd0-a673-5141352ec924
|
effect-of-word-embedding-variable-parameters
|
2101.02906
| null |
https://arxiv.org/abs/2101.02906v1
|
https://arxiv.org/pdf/2101.02906v1.pdf
|
Effect of Word Embedding Variable Parameters on Arabic Sentiment Analysis Performance
|
Social media such as Twitter, Facebook, etc. has led to a generated growing number of comments that contains users opinions. Sentiment analysis research deals with these comments to extract opinions which are positive or negative. Arabic language is a rich morphological language; thus, classical techniques of English sentiment analysis cannot be used for Arabic. Word embedding technique can be considered as one of successful methods to gaping the morphological problem of Arabic. Many works have been done for Arabic sentiment analysis based on word embedding, but there is no study focused on variable parameters. This study will discuss three parameters (Window size, Dimension of vector and Negative Sample) for Arabic sentiment analysis using DBOW and DMPV architectures. A large corpus of previous works generated to learn word representations and extract features. Four binary classifiers (Logistic Regression, Decision Tree, Support Vector Machine and Naive Bayes) are used to detect sentiment. The performance of classifiers evaluated based on; Precision, Recall and F1-score.
|
['Nursal ARICI', 'Anwar Alnawas']
|
2021-01-08
| null | null | null | null |
['arabic-sentiment-analysis']
|
['natural-language-processing']
|
[-1.84327111e-01 -2.11627722e-01 -2.73080379e-01 -4.77964729e-01
1.57984480e-01 -5.89528620e-01 4.63631600e-01 9.10035789e-01
-6.56593800e-01 5.77915907e-01 4.42717642e-01 -4.37116057e-01
1.71154156e-01 -1.02454937e+00 4.01189178e-01 -6.70695007e-01
-3.76821905e-02 5.68015426e-02 2.46236399e-01 -9.74696398e-01
8.88892889e-01 4.03129607e-01 -1.50912225e+00 5.52449882e-01
5.08221686e-01 8.29316914e-01 -5.48853539e-02 8.95048916e-01
-6.45142078e-01 8.86045277e-01 -8.21961105e-01 -4.52180833e-01
-1.19514562e-01 -3.54471952e-01 -4.92462307e-01 8.06562155e-02
-5.52025318e-01 1.48728177e-01 3.01523417e-01 7.93121040e-01
6.92973614e-01 8.79002213e-02 1.13803625e+00 -1.35389960e+00
-9.35595274e-01 5.79432189e-01 -7.31477439e-01 5.04012585e-01
3.85017872e-01 -5.31437635e-01 6.81201637e-01 -1.19116795e+00
1.61671638e-01 1.20809615e+00 5.78313828e-01 5.89480810e-02
-5.34881532e-01 -4.60441798e-01 -2.99023479e-01 3.23923111e-01
-1.07115662e+00 1.95025489e-01 8.71977150e-01 -3.86074692e-01
1.24186289e+00 1.56519756e-01 8.56049538e-01 5.92497170e-01
7.32110023e-01 4.95404154e-01 1.43432581e+00 -9.66078997e-01
1.70769319e-01 9.68499005e-01 7.52454698e-01 4.38889712e-01
4.10857052e-01 -5.51155210e-01 -5.11012852e-01 -8.09309632e-02
-2.50032544e-01 -2.36568153e-02 3.44991297e-01 2.24484205e-01
-4.95191365e-01 1.49452651e+00 3.40312459e-02 6.51666582e-01
-3.95858794e-01 -3.91950011e-01 8.35855305e-01 7.06555128e-01
6.23298943e-01 3.21882874e-01 -7.44040668e-01 -1.18393086e-01
-4.83655006e-01 2.08549842e-01 8.13910425e-01 6.39452279e-01
4.96117979e-01 2.99571753e-01 3.70178908e-01 8.04294407e-01
9.20549870e-01 7.02596188e-01 1.23258865e+00 2.54660547e-01
1.57321930e-01 1.27170241e+00 -9.17403623e-02 -1.89372122e+00
-6.58977568e-01 2.89757937e-01 -3.13512683e-01 1.88894853e-01
-1.55782431e-01 -6.52889192e-01 -8.16657245e-01 6.37176812e-01
3.12920690e-01 -5.58327496e-01 6.30297303e-01 4.54185635e-01
1.05757701e+00 1.16644943e+00 3.06576658e-02 -2.61557072e-01
1.64330900e+00 -6.98764980e-01 -1.04935348e+00 -1.22310497e-01
8.67994308e-01 -1.45270443e+00 8.50474179e-01 6.70396149e-01
-6.07930183e-01 -2.32830212e-01 -1.27521694e+00 2.06412077e-01
-1.46126807e+00 6.90416545e-02 7.33040154e-01 1.39459145e+00
-6.87365055e-01 1.47638442e-02 -6.07822716e-01 -4.78706241e-01
1.11002885e-01 6.74153745e-01 -4.01244968e-01 3.27365309e-01
-1.12617266e+00 1.32910681e+00 5.77289760e-01 -5.42906336e-02
1.81246519e-01 9.77350473e-02 -9.17854786e-01 -3.15993786e-01
-1.54892996e-01 3.03876966e-01 6.82448447e-01 -1.27410913e+00
-1.26805961e+00 6.39423132e-01 -6.28455952e-02 -2.21213773e-01
-3.82384419e-01 -3.25965136e-03 -8.97391617e-01 1.36578813e-01
-8.03992376e-02 2.75413930e-01 5.70167363e-01 -8.58292818e-01
-6.99820936e-01 -6.76140130e-01 -1.14466347e-01 2.78832048e-01
-1.06155872e+00 6.62128150e-01 2.99580365e-01 -7.65417278e-01
2.31067285e-01 -8.99399102e-01 -1.27573505e-01 -7.81631768e-01
3.74421738e-02 -4.97502208e-01 1.09412062e+00 -6.56548679e-01
1.50031912e+00 -1.83715820e+00 -2.57022172e-01 4.69695181e-01
-3.21081549e-01 5.54730833e-01 2.29982436e-01 9.49862659e-01
-1.40811831e-01 3.87318969e-01 -3.95183228e-02 2.07267985e-01
-1.35446236e-01 2.02034995e-01 -2.60008186e-01 3.13114047e-01
3.06923836e-01 2.33185112e-01 -4.70188260e-01 -7.05225050e-01
2.27271676e-01 7.30510175e-01 -4.49048281e-01 -1.45296007e-01
2.59305120e-01 -2.82986581e-01 -5.67785084e-01 7.83828259e-01
6.02241397e-01 2.98549622e-01 1.85867265e-01 -2.99657702e-01
-2.86047012e-01 -1.08652830e-01 -1.33727491e+00 5.19637525e-01
-4.04360622e-01 7.62226999e-01 -4.83853132e-01 -1.05872858e+00
1.34059346e+00 3.81108016e-01 3.34709764e-01 -3.67431968e-01
8.88944924e-01 1.05293989e-01 4.90698814e-02 -9.79109287e-01
9.57080841e-01 -3.20868760e-01 -1.13620363e-01 5.08772194e-01
2.01382473e-01 -1.50051117e-01 5.89827061e-01 4.02158760e-02
3.27757508e-01 -1.82776690e-01 5.62357247e-01 -2.26438001e-01
9.76641715e-01 3.08046848e-01 4.91749831e-02 -1.23863041e-01
-1.18066885e-01 1.38955310e-01 7.08037615e-01 -4.87124205e-01
-7.56255567e-01 -2.57291049e-01 -2.59995013e-01 1.23664212e+00
-1.65916711e-01 -5.56094468e-01 -5.37241280e-01 -5.18649280e-01
-1.06255352e-01 4.80771810e-01 -5.38097203e-01 -2.27957442e-02
-3.64350915e-01 -1.46049559e+00 2.87157446e-01 2.79622227e-01
8.59645680e-02 -1.31832254e+00 -6.45832479e-01 2.16758728e-01
4.48102802e-01 -5.55916309e-01 3.69480461e-01 2.72670507e-01
-9.66209412e-01 -1.09868038e+00 -4.14447933e-01 -1.18713117e+00
8.52473915e-01 1.46724498e-02 6.79058790e-01 1.61988348e-01
-2.06050262e-01 2.50911210e-02 -1.29231131e+00 -1.26665747e+00
-4.17223185e-01 1.67303324e-01 1.37834445e-01 -6.30389228e-02
1.25306904e+00 -7.02514276e-02 -3.32972050e-01 -7.73384646e-02
-1.25763571e+00 -7.04536855e-01 4.97942597e-01 6.09785259e-01
5.77415712e-02 2.13632002e-01 8.28427494e-01 -1.04483712e+00
1.41175425e+00 -7.21281707e-01 -1.87684774e-01 4.63419408e-02
-9.43607748e-01 -7.20050037e-02 4.20253992e-01 -3.27272207e-01
-6.04387820e-01 -4.07158941e-01 -3.71505827e-01 7.29348600e-01
1.58802241e-01 1.02007544e+00 4.92814213e-01 -4.33664918e-02
7.90825307e-01 4.89626937e-02 2.81828791e-01 -2.36569270e-02
6.30410239e-02 1.46421301e+00 -6.51679516e-01 5.21701761e-02
2.15903834e-01 2.47218743e-01 -1.92249194e-01 -1.25149941e+00
-5.15983820e-01 -6.37056351e-01 -5.37131310e-01 -2.77497679e-01
9.58288133e-01 -4.74274635e-01 -5.74232638e-01 2.81684190e-01
-8.03176641e-01 4.93843228e-01 2.76827425e-01 7.64143646e-01
1.29988894e-01 1.69325471e-01 -4.58702326e-01 -1.22412014e+00
-6.22435510e-01 -1.17054570e+00 3.55330706e-01 4.13669437e-01
-4.64346528e-01 -1.20281529e+00 1.90435931e-01 7.05468878e-02
4.84003156e-01 2.32230186e-01 9.44502532e-01 -9.83779669e-01
6.81911767e-01 -7.22570896e-01 1.54315099e-01 5.96192181e-01
2.85962075e-01 4.24725145e-01 -7.22786725e-01 -8.12722370e-02
1.35375932e-01 -3.38369370e-01 5.82873762e-01 2.11116090e-01
4.28506225e-01 -4.61160094e-01 -4.09329869e-02 -1.84141204e-01
1.66373062e+00 7.30540812e-01 6.39502466e-01 7.28271663e-01
3.08315247e-01 7.46733248e-01 9.06867683e-01 5.16345799e-01
3.53500247e-01 -1.08651541e-01 2.62019366e-01 2.75330931e-01
6.14363432e-01 3.65619063e-01 6.25908792e-01 1.43750191e+00
-3.90524082e-02 -3.98079425e-01 -1.03513706e+00 4.43902254e-01
-1.32704401e+00 -7.51631200e-01 -4.41096544e-01 1.66775119e+00
6.79124534e-01 2.44919464e-01 2.41197631e-01 1.12349558e+00
3.67547899e-01 2.33203530e-01 2.28044346e-01 -1.38727283e+00
-2.68819809e-01 5.55646598e-01 5.26956737e-01 5.39595306e-01
-1.08437204e+00 1.01863968e+00 5.42782784e+00 4.29030806e-01
-1.54549551e+00 -6.11153208e-02 4.05824989e-01 2.70122826e-01
-2.22236395e-01 -1.62750617e-01 -9.49199975e-01 4.71098781e-01
1.09016562e+00 -1.14728391e-01 -1.80647925e-01 8.87003481e-01
2.76533961e-02 -3.33938509e-01 -4.62547578e-02 6.99759185e-01
7.52154529e-01 -9.58334923e-01 2.36463740e-01 -2.49158263e-01
6.59478843e-01 -2.74491400e-01 1.70131817e-01 3.61648083e-01
-7.85135403e-02 -1.03836620e+00 1.57503977e-01 2.38507628e-01
5.37875146e-02 -1.28216505e+00 1.43012035e+00 1.41463587e-02
-8.39838207e-01 -1.98877573e-01 -6.91314876e-01 -3.62683028e-01
2.48629004e-02 1.79145738e-01 -1.05379403e+00 1.43561721e-01
8.10262144e-01 7.27956712e-01 -8.37026834e-01 4.99606222e-01
-4.87184711e-02 6.54475689e-01 -2.29527161e-01 -1.09799683e+00
4.87249821e-01 -4.29830313e-01 7.74824172e-02 1.44849873e+00
2.80704260e-01 4.32810187e-03 -1.01219796e-01 -1.87568411e-01
5.16212106e-01 1.20051587e+00 -7.74989426e-01 -3.90307873e-01
1.64799348e-01 1.29280508e+00 -1.32752144e+00 -1.46720186e-01
-5.86211264e-01 5.62567770e-01 -3.64224046e-01 -9.68005061e-02
-4.18090910e-01 -1.05451334e+00 1.82298422e-01 3.71121354e-02
1.61873460e-01 -2.28158996e-01 -4.22789693e-01 -7.54835665e-01
-3.33635569e-01 -9.88967597e-01 5.06127775e-01 -6.17608011e-01
-1.10892892e+00 8.39492798e-01 -8.27987120e-02 -1.15858662e+00
-5.79584949e-02 -1.26787174e+00 -5.35677612e-01 8.36160004e-01
-1.27156138e+00 -1.05381465e+00 1.48228705e-01 4.07289505e-01
6.09593391e-01 -8.25603187e-01 1.16878521e+00 2.82658786e-01
-3.74478579e-01 4.49243695e-01 1.12482511e-01 2.62031496e-01
7.06983566e-01 -1.21579885e+00 -5.06850481e-01 5.94231308e-01
-1.71412110e-01 6.74535215e-01 9.39888716e-01 -5.77698767e-01
-1.38813257e+00 -3.30515891e-01 1.25438368e+00 -3.90185356e-01
8.00219357e-01 4.28234152e-02 -3.65707397e-01 5.23161173e-01
6.25388861e-01 -3.72948498e-01 1.45131123e+00 -1.96047630e-02
5.12597449e-02 -1.20832704e-01 -1.36761284e+00 6.66999519e-01
-5.03231943e-01 -3.90092991e-02 -8.25892985e-01 3.44269603e-01
3.36884648e-01 1.63351104e-01 -9.31460857e-01 4.76087295e-02
6.61043525e-01 -8.00698280e-01 7.75469840e-01 -8.93816888e-01
6.48503065e-01 -1.54216513e-01 -3.71426731e-01 -1.19649553e+00
3.55071068e-01 -1.26265734e-02 2.52011210e-01 1.28373981e+00
8.92603457e-01 -6.94147766e-01 7.11793661e-01 2.02741846e-01
2.60094851e-01 -1.03362274e+00 -2.01565981e-01 3.73010226e-02
2.23079726e-01 -4.30693477e-01 3.47334743e-01 1.14198136e+00
4.04864550e-01 7.98044920e-01 -1.31380871e-01 -7.47570172e-02
-2.16276884e-01 -2.86904007e-01 4.66044426e-01 -1.19219387e+00
3.87041122e-01 -3.05520952e-01 -8.45050991e-01 -1.33255452e-01
-3.15441221e-01 -6.82189465e-01 -5.82175910e-01 -1.55410957e+00
-4.30844694e-01 -3.57150316e-01 -1.54773265e-01 2.36069426e-01
2.68505514e-02 4.76147890e-01 1.14021160e-01 -2.08877265e-01
5.84214740e-02 1.64193913e-01 8.59428883e-01 -6.43442795e-02
-3.67851108e-01 -1.76670551e-01 -8.79308701e-01 9.65215504e-01
1.18810797e+00 -7.60601044e-01 -5.73042929e-01 1.73848961e-02
1.23495662e+00 -3.36333483e-01 -7.82864273e-01 -5.01807630e-01
-2.68795271e-03 -2.85261542e-01 4.73204166e-01 -8.03720951e-01
2.51108855e-01 -8.75806808e-01 -6.74023807e-01 4.53951448e-01
-1.10508919e-01 9.05418396e-01 2.34665319e-01 2.41692185e-01
-4.66101319e-01 -9.97821987e-01 5.29369533e-01 -1.11007057e-01
-8.10070932e-01 -1.57572374e-01 -1.05758607e+00 -3.46494436e-01
1.29131687e+00 -5.29154480e-01 2.33345497e-02 -1.90884888e-01
-6.87645912e-01 9.34017971e-02 -7.00616278e-03 5.06254256e-01
7.27055550e-01 -1.06124115e+00 -6.82916939e-01 1.58094004e-01
2.85791725e-01 -4.89055157e-01 -1.80189535e-01 7.32794285e-01
-1.44900453e+00 1.92280516e-01 -4.82043415e-01 5.02073057e-02
-1.82548106e+00 4.64595795e-01 -1.73407018e-01 -1.12505935e-01
1.52976541e-02 7.90207386e-01 -9.62515056e-01 -4.28493053e-01
-1.69776931e-01 -1.11595772e-01 -1.73019469e+00 1.09152997e+00
8.24000597e-01 4.96806532e-01 1.79737285e-01 -1.12587726e+00
-3.10910910e-01 6.64301157e-01 -1.32222131e-01 -3.34209442e-01
1.51110280e+00 1.91360757e-01 -5.99221408e-01 8.42244387e-01
1.17764020e+00 4.19138461e-01 1.98919773e-01 3.78865808e-01
2.56542891e-01 -2.89767742e-01 5.50508797e-02 -6.17516041e-01
-6.57804728e-01 8.80010605e-01 9.53626394e-01 8.67351294e-01
9.14251626e-01 -6.08300567e-01 4.67397302e-01 6.32415593e-01
-1.52523160e-01 -1.67968750e+00 7.38063976e-02 8.43969524e-01
6.76828146e-01 -1.60348129e+00 2.23152235e-01 -1.96320698e-01
-9.93029416e-01 1.59012759e+00 2.96634287e-01 -4.70315069e-01
1.48180258e+00 3.79814714e-01 5.31860888e-01 -2.51626670e-01
-3.74725461e-01 -6.15910254e-02 7.19054565e-02 5.27844191e-01
1.19332755e+00 1.01210074e-02 -1.32981491e+00 9.06641185e-01
-6.56634629e-01 -2.79032439e-01 9.35544014e-01 1.55959785e+00
-8.31169128e-01 -1.28586757e+00 -6.40132487e-01 7.74859428e-01
-1.12840903e+00 -1.74739242e-01 -4.55727935e-01 7.03590572e-01
4.44945358e-02 1.47523785e+00 3.10513582e-02 -5.90305746e-01
1.50525287e-01 1.18710399e-01 5.34959733e-02 -7.37846851e-01
-8.98735702e-01 -2.96343118e-01 9.12184119e-02 3.57363373e-01
-7.53662109e-01 -4.37715650e-01 -1.36628783e+00 -4.12418157e-01
-6.54663980e-01 5.41227818e-01 1.44204521e+00 8.29021454e-01
-1.34805143e-01 3.24344486e-01 7.53310919e-01 -5.06270528e-01
-9.55709219e-02 -1.50514400e+00 -5.76183140e-01 2.32500434e-01
2.96500009e-02 -5.27865231e-01 -3.97259504e-01 1.35902062e-01]
|
[11.018199920654297, 6.911370277404785]
|
cf563b3e-ab30-42fc-9c43-79411f2b5119
|
joint-constrained-learning-for-event-event
|
2010.06727
| null |
https://arxiv.org/abs/2010.06727v2
|
https://arxiv.org/pdf/2010.06727v2.pdf
|
Joint Constrained Learning for Event-Event Relation Extraction
|
Understanding natural language involves recognizing how multiple event mentions structurally and temporally interact with each other. In this process, one can induce event complexes that organize multi-granular events with temporal order and membership relations interweaving among them. Due to the lack of jointly labeled data for these relational phenomena and the restriction on the structures they articulate, we propose a joint constrained learning framework for modeling event-event relations. Specifically, the framework enforces logical constraints within and across multiple temporal and subevent relations by converting these constraints into differentiable learning objectives. We show that our joint constrained learning approach effectively compensates for the lack of jointly labeled data, and outperforms SOTA methods on benchmarks for both temporal relation extraction and event hierarchy construction, replacing a commonly used but more expensive global inference process. We also present a promising case study showing the effectiveness of our approach in inducing event complexes on an external corpus.
|
['Dan Roth', 'Hongming Zhang', 'Muhao Chen', 'Haoyu Wang']
|
2020-10-13
| null |
https://aclanthology.org/2020.emnlp-main.51
|
https://aclanthology.org/2020.emnlp-main.51.pdf
|
emnlp-2020-11
|
['temporal-relation-extraction', 'event-relation-extraction']
|
['natural-language-processing', 'natural-language-processing']
|
[ 1.47035256e-01 2.84077555e-01 -4.58147466e-01 -5.70929527e-01
-9.87329960e-01 -7.87640512e-01 9.76547778e-01 7.69072950e-01
-5.76175153e-01 9.31100070e-01 6.19878173e-01 -2.44906694e-01
-3.07818592e-01 -8.18563759e-01 -7.44597077e-01 -1.69793442e-01
-6.07679605e-01 7.50607431e-01 2.44653150e-01 2.85104632e-01
-1.61912769e-01 4.84989375e-01 -1.32695484e+00 5.04825056e-01
3.38385463e-01 6.44674420e-01 -1.09721474e-01 3.50427687e-01
-1.55581430e-01 1.26030517e+00 -5.07780194e-01 -2.02822313e-01
-1.70405641e-01 -3.20024818e-01 -1.09349799e+00 2.50673801e-01
-4.12922092e-02 -7.60930702e-02 -1.73485205e-01 4.03071463e-01
2.57967114e-02 5.39276123e-01 7.45137751e-01 -1.33785200e+00
-1.78005278e-01 1.07083869e+00 -4.33342695e-01 3.56706083e-01
5.21295369e-01 -3.01769286e-01 1.68806076e+00 -6.42339528e-01
8.63250911e-01 1.24012172e+00 4.12636071e-01 1.93460677e-02
-1.51104808e+00 -4.12415922e-01 5.27249634e-01 2.91552395e-01
-1.28457832e+00 -4.82125223e-01 7.02124476e-01 -4.52822983e-01
1.52932823e+00 1.36575952e-01 3.14603835e-01 1.04882395e+00
-1.59590676e-01 7.84256160e-01 7.64663279e-01 -6.56675100e-01
2.54972458e-01 -2.88162768e-01 5.47496855e-01 6.86298668e-01
6.87451987e-03 4.50484231e-02 -8.12474549e-01 -2.90768385e-01
6.12719238e-01 -2.42959574e-01 7.89681375e-02 -3.02507170e-03
-1.28400326e+00 5.71655333e-01 -5.02653643e-02 4.63260859e-01
-2.84898847e-01 2.32925504e-01 5.11059761e-01 1.05139688e-01
6.14724517e-01 3.46524328e-01 -8.63458633e-01 -1.85556665e-01
-8.04510057e-01 2.70285010e-01 9.72697139e-01 8.53981078e-01
6.92639470e-01 -3.76623034e-01 -2.09550798e-01 6.62190676e-01
1.98354140e-01 -2.10643947e-01 9.39408243e-02 -1.07812059e+00
5.09056151e-01 6.59582853e-01 3.67773294e-01 -8.00209403e-01
-6.74951553e-01 2.18550991e-02 -1.97615251e-01 -1.99247032e-01
6.80021226e-01 -1.89430147e-01 -6.01382494e-01 2.12142992e+00
3.06947470e-01 6.89732671e-01 -5.98827526e-02 4.00761157e-01
6.09647512e-01 8.04157019e-01 5.10416865e-01 -7.89329886e-01
1.52905929e+00 -6.00982130e-01 -9.17464197e-01 -1.95461854e-01
7.81951070e-01 -3.64211529e-01 9.58399296e-01 2.97618136e-02
-1.11446750e+00 -1.29038393e-01 -7.60838687e-01 -3.64885688e-01
-2.87471056e-01 -8.53508413e-02 1.22943461e+00 -4.35380898e-02
-3.73618722e-01 3.60207945e-01 -1.30636537e+00 -9.56383422e-02
1.71317309e-01 2.72045493e-01 -3.91629070e-01 3.50227386e-01
-1.35515642e+00 1.02042317e+00 6.88353181e-01 -1.87970370e-01
-7.53571749e-01 -7.44815588e-01 -1.14684701e+00 3.18769813e-01
7.44256794e-01 -2.29281783e-01 1.39135444e+00 -3.17163885e-01
-1.05073166e+00 9.64003801e-01 -4.26937580e-01 -5.95317721e-01
1.24895446e-01 -3.06223035e-01 -5.54712594e-01 -3.38520072e-02
2.80568898e-01 3.44007462e-01 2.28064165e-01 -9.47546363e-01
-8.11169744e-01 -1.24649461e-02 3.83378953e-01 1.77614972e-01
-1.15660273e-01 6.54054880e-01 -5.73482811e-01 -5.73272169e-01
4.78582717e-02 -5.90342820e-01 -1.77157238e-01 -3.37014884e-01
-4.78973478e-01 -8.28144431e-01 5.06343544e-01 -3.62791926e-01
1.40198171e+00 -2.12031412e+00 2.65384078e-01 1.89145938e-01
-2.58656349e-02 -4.05234516e-01 1.31172672e-01 5.00770092e-01
-3.26436639e-01 8.37217048e-02 -3.11475009e-01 -5.25263131e-01
3.84122968e-01 5.66506684e-01 -4.79924053e-01 3.19253206e-01
2.54729569e-01 7.96406686e-01 -1.06312943e+00 -7.87074149e-01
8.82895142e-02 3.21255773e-01 -4.42697793e-01 1.31173655e-01
-6.49836481e-01 2.72855401e-01 -4.36678588e-01 4.58269447e-01
-8.18903744e-02 -4.60295588e-01 6.81409419e-01 -2.44640052e-01
-1.22592479e-01 1.04442918e+00 -1.41974163e+00 1.57061934e+00
-4.88335371e-01 4.38072920e-01 -1.87515348e-01 -1.06693244e+00
5.08057594e-01 7.83939123e-01 7.64710665e-01 -3.03340793e-01
-1.35907724e-01 -7.99314603e-02 -2.38327086e-01 -4.97784644e-01
1.03508875e-01 -4.28576022e-01 -5.07667482e-01 8.38649213e-01
2.11820662e-01 8.52026492e-02 7.40068614e-01 4.16167110e-01
1.18456888e+00 2.95225710e-01 7.15742171e-01 -6.66767173e-03
2.94859737e-01 -1.32069150e-02 1.00189960e+00 5.23229003e-01
1.23076104e-01 1.00691624e-01 9.17668521e-01 -5.86682498e-01
-7.53541231e-01 -1.24444032e+00 -2.31233492e-01 1.28720796e+00
-1.61977664e-01 -7.96662092e-01 -4.20621522e-02 -9.01673555e-01
-1.15649521e-01 9.28221822e-01 -4.61243540e-01 2.75960565e-01
-9.33957398e-01 -7.64577389e-01 6.65211618e-01 8.76528978e-01
-1.07948549e-01 -1.24232399e+00 -5.94798684e-01 5.04025042e-01
-5.02835631e-01 -1.50912476e+00 -4.37486976e-01 5.45804083e-01
-5.41053951e-01 -1.22381425e+00 3.06769192e-01 -8.56758714e-01
3.81463259e-01 -5.84103763e-01 1.46932018e+00 -2.88081229e-01
-1.76748067e-01 1.42297938e-01 -9.99623761e-02 -2.19096705e-01
-2.21556097e-01 -3.34955715e-02 -7.50439018e-02 -1.15479462e-01
4.27906752e-01 -8.31637323e-01 1.43458918e-01 2.46740738e-03
-9.13991690e-01 1.05998039e-01 1.25149950e-01 5.72981179e-01
5.93339741e-01 4.01011139e-01 6.23594582e-01 -1.09186876e+00
4.14431810e-01 -5.18749118e-01 -7.94568002e-01 3.95519376e-01
-2.69131094e-01 2.28163168e-01 2.29443073e-01 -4.81770545e-01
-1.28701866e+00 1.76622406e-01 2.98817813e-01 1.45245284e-01
-4.15141404e-01 9.41770732e-01 -3.57030362e-01 7.87823617e-01
4.96663421e-01 -2.56480455e-01 -6.62486196e-01 -3.29108328e-01
5.22577524e-01 1.41533956e-01 7.08332658e-01 -1.19408476e+00
6.39438570e-01 5.42711318e-01 5.34866415e-02 -4.08699125e-01
-1.22053266e+00 -3.07229877e-01 -8.22987616e-01 9.85236838e-02
9.42630291e-01 -1.02882802e+00 -7.56498456e-01 -1.29160076e-01
-1.31328142e+00 -5.26583493e-01 -4.92477775e-01 8.00801337e-01
-6.50932968e-01 2.33531862e-01 -1.16053843e+00 -6.37141585e-01
2.76972473e-01 -6.23507619e-01 9.79319990e-01 -2.70229936e-01
-7.13036537e-01 -1.24986315e+00 2.13359505e-01 1.48363307e-01
-3.35187346e-01 5.45525253e-01 1.29969525e+00 -8.25561345e-01
-5.54929316e-01 -2.35935543e-02 -1.35935113e-01 -2.61713475e-01
5.27669311e-01 9.29838270e-02 -5.24374783e-01 1.44695640e-01
-2.16012746e-01 -4.70828623e-01 6.74047410e-01 2.47959197e-01
7.28424489e-01 -3.68682951e-01 -5.92345476e-01 2.18441397e-01
1.18799007e+00 2.57159770e-01 1.33495465e-01 1.86141968e-01
6.82728171e-01 8.42970788e-01 5.41310787e-01 4.38460886e-01
6.87829494e-01 7.74094641e-01 -1.75636992e-01 1.04450025e-01
6.83893934e-02 -1.25398934e-01 3.40878576e-01 4.74043250e-01
-2.70265415e-02 -2.67286807e-01 -9.28105414e-01 7.93560505e-01
-2.04074574e+00 -1.12231481e+00 -6.43500239e-02 1.81276810e+00
1.48645246e+00 4.30029094e-01 3.55898701e-02 2.24507228e-02
6.37794495e-01 2.97357142e-01 -2.76058316e-01 -1.91741899e-01
-1.58734828e-01 4.16465342e-01 8.13936293e-02 9.69915330e-01
-1.33313656e+00 9.89492893e-01 6.71895552e+00 4.71297175e-01
-5.40714920e-01 2.00707447e-02 3.04180324e-01 -2.04582751e-01
-3.21842611e-01 2.46121347e-01 -7.31124818e-01 2.14111418e-01
9.54037011e-01 -2.63129383e-01 3.95973563e-01 2.91114241e-01
3.48074079e-01 1.91879347e-02 -1.84437335e+00 5.67122459e-01
-3.63725990e-01 -1.46429050e+00 -1.69203043e-01 1.55548295e-02
5.14643550e-01 -2.29748905e-01 -5.35164595e-01 2.13712946e-01
8.12065184e-01 -9.37526047e-01 6.18944168e-01 2.95248151e-01
4.89785969e-01 -6.30637288e-01 2.65346020e-01 3.01520586e-01
-1.51865685e+00 1.46244571e-01 3.85531098e-01 -4.16680068e-01
7.09842920e-01 8.47131431e-01 -7.64767587e-01 4.88362759e-01
2.51773983e-01 7.52987444e-01 -1.22651130e-01 6.50674641e-01
-6.59667075e-01 8.53027523e-01 -6.53422177e-01 4.53702748e-01
3.26304324e-02 5.88026596e-03 5.13438642e-01 1.52086842e+00
-2.21012592e-01 3.47424179e-01 5.59437037e-01 8.54000807e-01
-9.16261673e-02 -2.05872998e-01 -4.61297721e-01 -1.61708087e-01
7.39030123e-01 1.07683361e+00 -1.00172353e+00 -5.49202859e-01
-6.43016160e-01 3.60873282e-01 5.61129868e-01 4.16168928e-01
-9.77199435e-01 -8.62374716e-03 3.84932429e-01 -1.68910012e-01
2.36616150e-01 -5.01910329e-01 -1.56900778e-01 -1.34899318e+00
1.03071332e-01 -6.23604953e-01 1.06090105e+00 -5.21020889e-01
-1.36959982e+00 2.87501484e-01 5.47695816e-01 -6.33545637e-01
-5.88876963e-01 -2.06721738e-01 -5.75988650e-01 5.53722799e-01
-1.17849898e+00 -1.13471699e+00 2.84016430e-01 6.72471702e-01
5.13283849e-01 3.49420100e-01 9.42694604e-01 4.40352619e-01
-6.04504168e-01 1.30064338e-01 -8.21840942e-01 2.90209949e-01
6.77112937e-01 -1.51454592e+00 2.95569509e-01 9.06597197e-01
5.92112064e-01 5.91871738e-01 6.29040778e-01 -7.31416166e-01
-8.36949706e-01 -9.23790395e-01 1.62758708e+00 -4.43265259e-01
9.90541518e-01 -5.84129870e-01 -1.17029941e+00 1.50572836e+00
2.13719353e-01 7.13216979e-03 8.09090972e-01 7.46466577e-01
-6.13190830e-01 6.71398863e-02 -7.16935575e-01 5.08027434e-01
1.20705068e+00 -8.49832416e-01 -1.12655044e+00 5.44472575e-01
8.88266981e-01 -3.91123056e-01 -9.84668612e-01 5.46996534e-01
1.44467309e-01 -3.65928382e-01 8.90441179e-01 -9.93452787e-01
4.41682011e-01 -4.86599177e-01 -1.95001543e-01 -8.04219782e-01
-2.62347400e-01 -6.46368921e-01 -5.35097420e-01 1.67333949e+00
7.34349430e-01 -4.07316238e-01 5.64634860e-01 9.31932092e-01
-7.22189918e-02 -4.34377849e-01 -8.74569595e-01 -6.22106969e-01
-2.43442103e-01 -7.31457114e-01 3.30475271e-01 1.43378675e+00
4.30692703e-01 6.81546628e-01 -1.95646614e-01 4.18516964e-01
5.26477337e-01 4.29121733e-01 1.42549127e-01 -1.25119209e+00
-6.55381441e-01 -1.82745472e-01 3.10203470e-02 -5.93821645e-01
6.96842790e-01 -9.40453529e-01 8.19692910e-02 -1.54844320e+00
1.14226885e-01 -4.87672001e-01 -3.96362305e-01 1.04574132e+00
-1.89116001e-01 -2.00204700e-01 -1.89402670e-01 1.42638981e-01
-8.00613403e-01 1.93611324e-01 6.72121286e-01 -1.98507495e-02
-4.28456932e-01 -2.14407071e-01 -3.47425699e-01 9.56205964e-01
5.34254789e-01 -6.59732461e-01 -5.71708858e-01 -3.29776973e-01
3.51199716e-01 4.30095285e-01 2.16894820e-01 -5.94606698e-01
3.94771546e-01 -3.70283842e-01 4.99577969e-02 -4.89948988e-01
2.76077062e-01 -5.84842622e-01 2.38351464e-01 -9.82965603e-02
-7.86487341e-01 9.64102708e-03 1.84217408e-01 4.48417574e-01
-4.67550695e-01 1.86210368e-02 2.84429103e-01 -1.02672704e-01
-6.97190940e-01 8.14338177e-02 -5.19894361e-01 3.06779146e-01
9.01677430e-01 4.46563631e-01 -5.81921227e-02 -1.93479925e-01
-1.30102611e+00 2.84472167e-01 -4.47049551e-02 2.97177076e-01
1.86688930e-01 -1.22123635e+00 -6.50175273e-01 -3.01677585e-01
-7.08671883e-02 2.33750716e-01 -2.60650545e-01 6.13085926e-01
1.91722605e-02 2.57511258e-01 9.89859924e-02 -2.80851096e-01
-1.22107744e+00 6.16084218e-01 1.01986423e-01 -8.28979671e-01
-6.95262790e-01 6.47313535e-01 9.77541283e-02 -3.41336280e-01
4.42253202e-01 -5.48544288e-01 -1.56391978e-01 3.90205383e-01
2.29495257e-01 1.55903473e-01 -5.95588349e-02 -4.25607264e-01
-5.70660353e-01 1.06969103e-01 -2.44434066e-02 -5.68342149e-01
1.51380050e+00 -5.25455922e-02 -4.50763881e-01 8.47342908e-01
8.92672777e-01 2.22449228e-01 -1.26678145e+00 -4.36268210e-01
7.51049101e-01 1.24347001e-01 -3.23901474e-01 -6.64684832e-01
-5.60772777e-01 2.35235170e-01 -3.83778334e-01 4.05802578e-01
9.22071278e-01 4.97883141e-01 6.44187570e-01 5.38642347e-01
2.58352995e-01 -9.08694565e-01 6.61414936e-02 6.30476475e-01
5.82933247e-01 -8.23699594e-01 1.71105728e-01 -6.96222723e-01
-4.20275837e-01 7.38798141e-01 5.03340065e-01 -1.00920968e-01
6.38997376e-01 8.10436964e-01 -1.83435574e-01 -3.52004975e-01
-1.21618903e+00 -3.18266153e-01 2.10227102e-01 2.31093094e-01
8.26621234e-01 9.91466194e-02 -4.36765254e-01 5.47779918e-01
-7.77280715e-04 -1.02535963e-01 1.88382328e-01 1.04714489e+00
1.41806679e-03 -1.46396494e+00 3.65970358e-02 1.57867789e-01
-6.55600786e-01 -6.98989406e-02 -2.15369835e-01 9.52715218e-01
3.41793358e-01 9.89007652e-01 3.63135248e-01 2.26891637e-01
2.61997700e-01 3.87251943e-01 5.60125351e-01 -1.02074873e+00
-4.23789620e-01 3.80044430e-01 8.12335730e-01 -5.35103381e-01
-8.05518746e-01 -9.64465022e-01 -1.81421316e+00 3.18677545e-01
-9.54228118e-02 3.87580454e-01 1.18652865e-01 1.51207197e+00
1.34211972e-01 7.02133238e-01 2.95190781e-01 -5.30322433e-01
-3.29090543e-02 -7.19749689e-01 -4.53928947e-01 6.94550931e-01
2.15475276e-01 -6.73082530e-01 -2.50201285e-01 5.20964146e-01]
|
[9.112794876098633, 9.158293724060059]
|
a42c9abd-eba1-4cb0-b5a7-a4043242089c
|
expert-agnostic-ultrasound-image-quality
|
2307.02462
| null |
https://arxiv.org/abs/2307.02462v2
|
https://arxiv.org/pdf/2307.02462v2.pdf
|
Expert-Agnostic Ultrasound Image Quality Assessment using Deep Variational Clustering
|
Ultrasound imaging is a commonly used modality for several diagnostic and therapeutic procedures. However, the diagnosis by ultrasound relies heavily on the quality of images assessed manually by sonographers, which diminishes the objectivity of the diagnosis and makes it operator-dependent. The supervised learning-based methods for automated quality assessment require manually annotated datasets, which are highly labour-intensive to acquire. These ultrasound images are low in quality and suffer from noisy annotations caused by inter-observer perceptual variations, which hampers learning efficiency. We propose an UnSupervised UltraSound image Quality assessment Network, US2QNet, that eliminates the burden and uncertainty of manual annotations. US2QNet uses the variational autoencoder embedded with the three modules, pre-processing, clustering and post-processing, to jointly enhance, extract, cluster and visualize the quality feature representation of ultrasound images. The pre-processing module uses filtering of images to point the network's attention towards salient quality features, rather than getting distracted by noise. Post-processing is proposed for visualizing the clusters of feature representations in 2D space. We validated the proposed framework for quality assessment of the urinary bladder ultrasound images. The proposed framework achieved 78% accuracy and superior performance to state-of-the-art clustering methods.
|
['Subir Kumar Saha', 'Richard Voyles', 'SH Chandrashekhara', 'Dimitrios Ntentia', 'Deepak Raina']
|
2023-07-05
| null | null | null | null |
['image-quality-assessment', 'clustering']
|
['computer-vision', 'methodology']
|
[ 7.20745549e-02 4.49510217e-02 2.55370200e-01 -4.79145110e-01
-7.79128671e-01 -4.23024058e-01 -9.93144140e-02 5.24648130e-01
-5.18323243e-01 1.59581169e-01 5.50283790e-02 -8.87631252e-02
-7.50479698e-01 -6.46162271e-01 -1.52814195e-01 -1.11656260e+00
-3.84870410e-01 3.49647969e-01 -3.27487551e-02 2.63632655e-01
1.34645924e-01 1.90044090e-01 -1.62657189e+00 2.94956625e-01
1.17333019e+00 1.04222059e+00 4.20907527e-01 8.43164802e-01
-5.13793416e-02 7.83070207e-01 -6.26007199e-01 -3.33532132e-02
-1.00627290e-02 -6.17012560e-01 -6.91417992e-01 4.17866737e-01
2.94095185e-03 -1.28274679e-01 -2.66522646e-01 1.46046114e+00
7.64412522e-01 2.90538877e-01 8.59582007e-01 -8.83875370e-01
-7.00709403e-01 4.33329701e-01 -5.30541182e-01 5.07843494e-01
8.45419019e-02 -4.51249927e-02 6.03654027e-01 -6.26145244e-01
5.23022473e-01 1.05055702e+00 5.18403471e-01 3.47487241e-01
-1.02671909e+00 -4.30124432e-01 -5.78540564e-01 3.86553675e-01
-1.19873393e+00 -1.89805776e-01 7.02268004e-01 -7.80578792e-01
3.56470525e-01 4.91262108e-01 7.39652455e-01 5.27409136e-01
3.89226228e-01 5.03259480e-01 1.14253271e+00 -2.19617799e-01
3.93049687e-01 2.42353931e-01 3.99219543e-02 9.45990860e-01
1.63021728e-01 2.04025302e-02 -1.36817530e-01 4.04265448e-02
9.05581594e-01 1.98290065e-01 -3.93154204e-01 -4.49673295e-01
-8.47020090e-01 8.76295388e-01 5.00777245e-01 7.18159735e-01
-6.07202053e-01 -3.18224937e-01 5.58827102e-01 2.43839502e-01
2.60378093e-01 5.91059208e-01 4.47312072e-02 -1.90764442e-01
-1.03786647e+00 -5.55642009e-01 3.73369783e-01 6.13580108e-01
5.59685111e-01 -1.28649756e-01 -2.01220974e-01 9.95245337e-01
3.62550914e-01 3.55781406e-01 8.20873618e-01 -1.14495289e+00
-6.07417896e-02 7.43451297e-01 -1.53616190e-01 -1.43129909e+00
-7.26861835e-01 -4.70940739e-01 -1.15976083e+00 1.86783195e-01
3.80796976e-02 -1.21405810e-01 -1.00616992e+00 9.67900336e-01
3.36480647e-01 -7.12781996e-02 7.34979101e-03 1.22360492e+00
1.29402328e+00 5.26558757e-01 9.93865207e-02 -5.05334973e-01
1.25401378e+00 -8.27123404e-01 -1.06045067e+00 4.81124759e-01
4.96870220e-01 -7.33233035e-01 8.11537027e-01 5.73921442e-01
-1.12402081e+00 -8.02733898e-01 -1.10224211e+00 2.64679313e-01
-2.08507422e-02 2.32323140e-01 4.27950293e-01 8.22508991e-01
-1.04897082e+00 9.10127223e-01 -1.27739811e+00 -1.63298860e-01
4.99410808e-01 3.81928235e-01 -6.28968000e-01 -1.21283745e-02
-9.07717049e-01 9.93473470e-01 5.32704353e-01 4.04265553e-01
-6.59458458e-01 -5.39652646e-01 -1.12589431e+00 1.98417917e-01
2.07484499e-01 -3.00911456e-01 1.00042820e+00 -8.17815721e-01
-1.39028370e+00 6.48847699e-01 1.27563030e-01 9.88255255e-03
4.10764188e-01 1.15338221e-01 -4.17773783e-01 7.29303122e-01
1.10886618e-01 2.28813171e-01 9.15034354e-01 -1.44833732e+00
-5.95676601e-01 -4.61874038e-01 -5.20135522e-01 3.93955231e-01
-4.25477475e-01 -1.97214738e-01 -7.31405973e-01 -3.40967655e-01
5.59558928e-01 -6.26744330e-01 -2.35591784e-01 -1.80462301e-01
-2.65639368e-02 -2.19835609e-01 9.31834519e-01 -8.77377927e-01
1.22228754e+00 -2.31398606e+00 1.21319182e-01 6.10082328e-01
4.56618369e-01 1.99081391e-01 1.56199262e-01 1.10667218e-02
-3.83620709e-02 1.07878268e-01 -2.38547117e-01 -2.31150165e-01
-1.73800752e-01 5.34260333e-01 8.00568044e-01 6.92871094e-01
1.95244983e-01 3.97604913e-01 -1.21400499e+00 -1.19967914e+00
8.10310185e-01 3.59072179e-01 -3.70300144e-01 5.25842130e-01
4.26257282e-01 8.64919841e-01 -3.27764481e-01 5.05553067e-01
7.50524163e-01 -3.62548172e-01 1.84850872e-01 -6.23190045e-01
9.64069217e-02 -3.78684849e-01 -1.44625187e+00 1.87973595e+00
-2.61847585e-01 4.05825287e-01 2.23306477e-01 -1.32071960e+00
7.47985840e-01 5.44960558e-01 8.47038507e-01 -7.15936422e-01
5.34843504e-01 4.69927639e-02 2.47367010e-01 -1.40943933e+00
1.16403259e-01 -1.89838707e-01 1.67943954e-01 3.70013475e-01
5.85935891e-01 -3.42671365e-01 3.43371451e-01 1.36739999e-01
9.71624613e-01 -1.99163258e-01 3.15278500e-01 -3.54141593e-01
3.13704789e-01 -3.40280086e-01 4.66733128e-01 6.60888910e-01
-5.18022835e-01 6.39472604e-01 2.50185430e-01 -2.05544099e-01
-9.30507243e-01 -1.00030315e+00 -4.21862155e-01 7.78125584e-01
4.02502090e-01 -5.02413772e-02 -7.18837202e-01 -6.07395709e-01
-1.42570853e-01 2.56875753e-01 -1.00453758e+00 -2.18379483e-01
-1.37011543e-01 -7.56485820e-01 2.42589369e-01 4.61418211e-01
2.42065772e-01 -1.13722003e+00 -7.27661967e-01 2.39757687e-01
-2.51606315e-01 -6.40314460e-01 -1.52956769e-01 1.77578062e-01
-9.55814242e-01 -1.21924603e+00 -7.12904811e-01 -8.23945761e-01
9.30669010e-01 -1.74715891e-02 9.26477015e-01 3.42314899e-01
-6.58782601e-01 3.13949734e-01 -3.60179543e-01 -2.61517048e-01
-5.72989166e-01 -3.11148584e-01 5.59645668e-02 -1.16591267e-02
9.82752442e-02 -5.38568377e-01 -1.04109740e+00 2.34125450e-01
-1.01833224e+00 -2.23981917e-01 8.62687171e-01 1.03933740e+00
6.68989897e-01 5.82368314e-01 3.24880570e-01 -8.77298295e-01
5.78540802e-01 -3.60462189e-01 -2.56787986e-01 -7.23820329e-02
-4.97415304e-01 -1.48023516e-01 2.86908805e-01 -1.13203801e-01
-1.17015493e+00 -1.56286612e-01 -4.83784750e-02 -6.82843387e-01
-2.00928748e-01 6.95358276e-01 2.89693415e-01 -1.00097373e-01
7.81045437e-01 1.71220109e-01 1.44715056e-01 -2.05770001e-01
7.68220052e-02 9.13913190e-01 6.92585289e-01 -1.22083396e-01
4.02286053e-01 3.45676810e-01 -1.10095508e-01 -9.52938616e-01
-4.94176745e-01 -9.66417670e-01 -6.65367007e-01 -4.99669403e-01
1.07117820e+00 -5.95079303e-01 -8.95749092e-01 3.41783985e-02
-6.95552528e-01 1.25661060e-01 -1.59063160e-01 7.82824814e-01
-3.39898676e-01 6.44441128e-01 -8.95931065e-01 -9.24865961e-01
-5.86599469e-01 -1.32608545e+00 7.57710159e-01 6.39914036e-01
-3.85890529e-02 -9.62120414e-01 -1.29548118e-01 4.63890076e-01
2.68958062e-01 3.03406179e-01 8.31272423e-01 -4.25344020e-01
-1.18577061e-02 -1.87975526e-01 -4.14683282e-01 6.01112306e-01
5.60655892e-01 -1.16444737e-01 -1.03769863e+00 -2.29995131e-01
3.12108845e-01 -3.24131757e-01 5.97718596e-01 7.98038781e-01
1.52780247e+00 1.08121047e-02 -6.62164092e-02 4.26583469e-01
1.44402432e+00 5.29472232e-01 4.66072917e-01 -1.18109278e-01
5.90398490e-01 6.48782969e-01 5.99017203e-01 3.00938547e-01
1.38421699e-01 -4.40883115e-02 5.06405652e-01 -5.41615486e-01
2.93577492e-01 3.56941134e-01 -5.45456231e-01 1.43882191e+00
-5.21291256e-01 2.05208793e-01 -8.30179870e-01 7.14419663e-01
-1.88042486e+00 -8.89304698e-01 -5.58707602e-02 1.91256726e+00
7.81793714e-01 -5.72007261e-02 -1.67937100e-01 4.57669467e-01
6.34012461e-01 -2.61689723e-01 -2.86725610e-01 -3.04265767e-01
3.65306258e-01 2.83822328e-01 2.30852187e-01 3.48454207e-01
-1.29050660e+00 1.69263020e-01 5.87486124e+00 8.05695355e-01
-1.03942800e+00 3.09785783e-01 5.87545335e-01 1.48679242e-01
1.29712954e-01 -7.90660560e-01 3.86884958e-01 4.66351658e-01
5.89191258e-01 2.95937330e-01 5.45591787e-02 8.30742002e-01
4.30995792e-01 -4.69183534e-01 -1.08089316e+00 1.31686258e+00
1.26486897e-01 -9.80024457e-01 -4.35487568e-01 -3.00253093e-01
8.06378007e-01 -2.56041020e-01 7.78687969e-02 1.49055332e-01
3.90484333e-02 -1.22323656e+00 1.06479496e-01 6.37020051e-01
6.74368441e-01 -7.84803629e-01 1.47800136e+00 1.46348938e-01
-8.58799398e-01 -7.22707137e-02 -4.14381772e-01 2.64120221e-01
-2.45216504e-01 5.75358033e-01 -7.54025757e-01 7.60468602e-01
1.10863996e+00 4.01955694e-01 -4.78161007e-01 1.19890916e+00
1.44130558e-01 5.91740429e-01 -1.10871539e-01 -1.20929200e-02
3.91131192e-01 -3.28095257e-01 2.35860869e-01 1.41998482e+00
3.98733765e-01 3.26679766e-01 2.64641225e-01 5.49347281e-01
1.89857349e-01 2.89101690e-01 -3.52741629e-01 4.98050824e-02
2.39141747e-01 1.53526890e+00 -1.08103251e+00 -5.60324788e-01
-1.85398892e-01 9.56290722e-01 -2.44504381e-02 7.03035966e-02
-3.71135831e-01 -5.16622782e-01 -7.76904449e-02 -3.47242385e-01
1.40975088e-01 1.16399623e-01 -3.03732485e-01 -7.64303029e-01
-1.25416443e-01 -8.69133711e-01 5.89406669e-01 -6.11365139e-01
-1.32359898e+00 5.51860631e-01 -2.34169945e-01 -1.31273663e+00
-7.36047626e-02 -3.29783767e-01 -4.19447124e-01 7.47836649e-01
-1.08701074e+00 -5.71413338e-01 -6.68286502e-01 6.11973584e-01
4.91637290e-01 1.49980366e-01 9.67130601e-01 5.92500448e-01
-3.43296409e-01 3.35397065e-01 2.86894649e-01 2.81199247e-01
5.60910523e-01 -1.76371479e+00 -7.69374669e-01 5.02672136e-01
-2.14364693e-01 5.38205743e-01 7.72731602e-01 -4.40920591e-01
-1.17753422e+00 -8.91190171e-01 2.83856541e-01 -8.05580691e-02
3.37184310e-01 1.80414259e-01 -1.08738267e+00 9.29534510e-02
2.56706029e-01 1.47735864e-01 1.15761352e+00 -6.25293851e-02
3.80979657e-01 6.11925498e-02 -1.35957229e+00 2.14549482e-01
5.92052162e-01 -2.40576372e-01 -7.20613360e-01 2.79749542e-01
2.31149480e-01 -5.52186608e-01 -1.56333458e+00 3.64458740e-01
5.28018415e-01 -1.13142884e+00 6.54320955e-01 -1.95576236e-01
5.30108035e-01 -3.32614332e-01 1.79669723e-01 -1.49286568e+00
-6.12831473e-01 -1.43725172e-01 1.27269208e-01 8.38639021e-01
2.71827102e-01 5.97029272e-03 7.04458773e-01 3.51415724e-01
-3.33480775e-01 -6.80335999e-01 -8.12346697e-01 -2.05887169e-01
-4.58519161e-01 -4.09523249e-01 2.76708044e-02 1.17078245e+00
3.61096054e-01 1.27288014e-01 -1.93859920e-01 4.11540806e-01
7.62877524e-01 1.22176334e-02 2.49648809e-01 -1.23205423e+00
-3.14416200e-01 -3.85171682e-01 -7.93748915e-01 -1.55096054e-01
-6.44874513e-01 -7.52429903e-01 3.86749923e-01 -1.86723936e+00
3.01126540e-01 -3.30308706e-01 -6.05412424e-01 1.31730065e-01
-4.43180025e-01 5.55695236e-01 -1.83922686e-02 2.54246652e-01
-8.06433916e-01 4.80044395e-01 1.71831727e+00 -1.93338379e-01
-4.27709430e-01 -1.90503206e-02 -4.08586115e-01 8.16893578e-01
5.24407148e-01 -4.22849923e-01 -4.52876449e-01 -2.68889308e-01
-6.10914454e-02 3.42446834e-01 -1.21029161e-01 -1.06432915e+00
2.90777594e-01 1.35267079e-01 9.26615775e-01 -5.99576652e-01
8.90193507e-02 -1.01886988e+00 -6.37614280e-02 7.23417401e-01
-1.10739082e-01 -2.27557898e-01 -1.11543406e-02 4.58886534e-01
-5.03115475e-01 -5.74114203e-01 8.98059368e-01 -4.76130277e-01
-4.20596033e-01 1.64125234e-01 -6.19185865e-01 -4.02423382e-01
7.81068146e-01 -2.58783728e-01 1.91428140e-01 -4.10216391e-01
-1.29583406e+00 4.42879438e-01 -1.23742558e-01 3.75232100e-02
7.84116805e-01 -1.20493054e+00 -5.58743417e-01 8.81585851e-02
6.81764632e-02 3.22970450e-01 9.55204189e-01 1.17298651e+00
-7.94534385e-01 -4.99263359e-03 -3.23513389e-01 -1.24880373e+00
-1.29028475e+00 5.35388529e-01 2.72472680e-01 -2.90122598e-01
-5.65903664e-01 7.93072999e-01 -1.04403362e-01 -4.89966512e-01
3.94154817e-01 -3.68989587e-01 -7.79950321e-01 1.61601499e-01
4.47165072e-01 6.15048945e-01 3.39591980e-01 -4.63206917e-01
-1.85963646e-01 6.18268728e-01 6.85604066e-02 1.40582398e-01
1.24173093e+00 -2.95664042e-01 -1.00774236e-01 5.26057661e-01
1.18736076e+00 -4.11665589e-01 -7.68930316e-01 -1.89341560e-01
-4.34710383e-02 -3.50106955e-01 4.29668784e-01 -6.43229246e-01
-1.15361154e+00 9.99623120e-01 1.27702987e+00 5.30336261e-01
1.26426089e+00 -1.09185487e-01 3.56979609e-01 2.88108885e-01
-5.17443493e-02 -1.45631552e+00 2.22891450e-01 -1.47265002e-01
7.24100471e-01 -1.66305220e+00 -5.53221665e-02 -2.78708309e-01
-7.50224650e-01 1.10025001e+00 5.12851834e-01 -1.76778704e-01
7.39029527e-01 1.41961917e-01 6.36105716e-01 -5.05961120e-01
-3.55054326e-02 6.25913590e-02 6.42026782e-01 6.95193410e-01
3.69927257e-01 2.29719520e-01 -2.98829257e-01 6.41623795e-01
-1.34433955e-01 -4.18810695e-02 2.84224451e-01 9.60922062e-01
-4.83865052e-01 -3.59826416e-01 -6.17565513e-01 8.23081255e-01
-6.84424877e-01 3.29732895e-01 3.61283183e-01 5.69829643e-01
5.54488420e-01 1.21391404e+00 2.21874908e-01 -3.40279400e-01
2.03809172e-01 -1.68946773e-01 4.61800367e-01 -6.08981371e-01
-4.94937986e-01 6.29423738e-01 -1.44498602e-01 -4.94444311e-01
-8.04820240e-01 -4.98707175e-01 -1.23648024e+00 1.84625253e-01
-6.04156375e-01 4.67620075e-01 7.64459431e-01 9.22069073e-01
-4.02416438e-02 1.10350394e+00 6.80558562e-01 -9.36042547e-01
-3.33495080e-01 -1.42329288e+00 -7.56428182e-01 6.98588312e-01
3.78536075e-01 -6.38862193e-01 -1.73428267e-01 2.47377947e-01]
|
[14.582500457763672, -2.3132903575897217]
|
610dc925-23ff-4cc8-88ef-81112f162367
|
lut-gce-lookup-table-global-curve-estimation
|
2306.07083
| null |
https://arxiv.org/abs/2306.07083v2
|
https://arxiv.org/pdf/2306.07083v2.pdf
|
LUT-GCE: Lookup Table Global Curve Estimation for Fast Low-light Image Enhancement
|
We present an effective and efficient approach for low-light image enhancement, named Lookup Table Global Curve Estimation (LUT-GCE). In contrast to existing curve-based methods with pixel-wise adjustment, we propose to estimate a global curve for the entire image that allows corrections for both under- and over-exposure. Specifically, we develop a novel cubic curve formulation for light enhancement, which enables an image-adaptive and pixel-independent curve for the range adjustment of an image. We then propose a global curve estimation network (GCENet), a very light network with only 25.4k parameters. To further speed up the inference speed, a lookup table method is employed for fast retrieval. In addition, a novel histogram smoothness loss is designed to enable zero-shot learning, which is able to improve the contrast of the image and recover clearer details. Quantitative and qualitative results demonstrate the effectiveness of the proposed approach. Furthermore, our approach outperforms the state of the art in terms of inference speed, especially on high-definition images (e.g., 1080p and 4k).
|
['Jinhui Tang', 'Jiangxin Dong', 'Changguang Wu']
|
2023-06-12
| null | null | null | null |
['image-enhancement', 'low-light-image-enhancement']
|
['computer-vision', 'computer-vision']
|
[ 0.3318722 -0.4908849 0.17607233 -0.2590123 -0.8661816 -0.30246097
0.14384443 0.17313679 -0.680025 0.6483503 -0.24458934 -0.02842506
-0.11889499 -0.8996953 -0.8493633 -0.9305729 0.48734862 -0.06163694
0.54475343 -0.14202502 0.4768046 0.6355823 -1.4945544 -0.19963694
1.1550622 1.2268101 0.3175024 0.6200978 0.01211509 0.5566886
-0.29542914 -0.51596874 0.253951 -0.03251034 -0.2752913 0.02695847
0.9276088 -0.832755 -0.33174366 1.2179414 0.7908558 0.09638347
0.50284594 -0.82930577 -0.4335686 -0.05169563 -0.91107553 0.03199784
-0.11942891 0.385669 0.7587151 -1.0718327 0.6317837 0.7884315
0.84293205 0.12916684 -1.3277981 -0.6847994 -0.1443266 0.48875716
-1.5249931 -0.40909997 1.0021114 -0.11390518 0.5408726 0.01129978
0.6760932 0.37514916 0.46217167 0.31096533 1.5010294 -0.58929926
0.14646275 0.02015029 -0.02089789 0.86597383 0.18029441 0.2512379
-0.52465427 0.1969788 1.1084207 -0.15611903 -0.42701122 -0.3365539
-0.96412444 0.41175422 0.66154903 -0.10680725 -0.27606577 0.37676287
0.23708794 -0.1748007 0.33346683 0.31975302 -0.0726195 0.09916542
-1.0821316 0.10394195 0.47284567 0.8429809 1.1518916 0.09994748
-0.5452366 0.94871557 -0.01178834 0.8337012 -0.11163124 -1.2855152
0.23314045 0.43399385 0.32114494 -1.1503264 -0.37433833 -0.36088452
-1.0341064 0.5213123 0.6059252 0.0758407 -1.0000583 1.4590702
0.4461338 0.2764088 -0.16439943 0.95450455 0.761234 0.74457186
-0.07265563 -0.3997871 1.4551884 -1.0501869 -0.90561855 0.13865454
-0.02489492 -0.75373363 1.1898187 0.51976407 -1.3677087 -0.6678878
-1.1267616 -0.49585888 -0.20159599 0.21355177 0.37721008 0.45069218
-1.0492834 0.5116994 -0.44194058 0.19069311 0.45701352 0.2527529
0.17828046 -0.31781873 -1.1179945 0.7883852 0.3295611 0.1511604
-0.6464165 -1.0402725 -0.5803389 0.11037892 0.58440834 -0.5230149
0.916815 -0.53977025 -1.8093065 0.5844464 -0.17261186 -0.3414783
0.4627352 0.07312926 -0.19088337 0.6970179 -0.33822227 0.7632622
1.3176895 -1.3446512 -0.43118268 -0.10636862 0.08119559 0.23810461
-0.51783836 -0.05618995 -0.929937 -0.6317345 -0.11211369 -0.6537778
-0.15296051 0.7621626 -0.03207231 0.16749318 0.6154375 -0.956485
1.4823166 -2.0837522 -0.27655044 0.26530167 0.43923655 0.33142737
0.04183278 -0.09121964 0.4018785 -0.3199662 -0.5578137 -0.3986765
-0.19236055 0.12669799 0.05160163 0.50154126 0.11819654 0.94434005
-0.66231084 -0.7559021 0.50007206 0.8998596 -0.52034545 -0.1110267
-0.09288906 0.3346273 -0.01658623 0.64467984 1.2609892 -0.27398297
-0.03371914 -0.767128 -0.58267784 -0.4648569 -1.232209 1.5022317
-0.7867846 0.6654038 0.192829 -0.38852528 1.0932097 -0.12255376
0.4006587 -1.0037282 0.26117995 0.24721438 -0.42367342 -0.2568336
0.7335025 -0.0542498 0.31223497 0.1026106 -0.27588245 -0.4111755
0.1575763 0.03810078 0.42760298 0.13172881 0.2054631 -0.2172867
0.78878915 -0.34294778 0.5031223 0.74118495 -0.28364605 0.54856807
0.14237408 -0.21741448 -1.264768 -0.85823095 -0.4347645 0.74456644
0.6671995 -0.26427218 -0.90937614 -0.07549636 -0.09117783 0.50739026
-0.11019015 0.15177852 -0.6635927 -0.7158414 0.17523597 0.46753192
1.067317 -0.7767119 -0.5190865 0.24243556 -0.28184882 -1.398283
-0.7114779 -0.10446206 -0.7715507 -0.8058521 -0.7660768 -0.6526383
0.73364425 0.31423613 0.8417433 0.12020183 -0.5018376 0.3241762
0.09285732 -0.15702924 -0.22253841 -0.30102244 -0.33292136 0.21435826
-0.06899413 -0.37167475 -1.1172477 0.38561028 -0.85902774 0.22971079
0.7037969 0.89182293 1.108969 0.441433 0.2059374 -0.58642673
0.32343185 0.33324638 -1.1575147 0.32673496 -1.1164092 0.04186602
0.765056 -0.4632955 -1.5690566 0.07827985 -0.12958004 -0.48840374
0.16280434 0.02478505 0.07639027 -0.883442 0.57844055 0.2835137
0.0514731 -0.28150746 0.45233062 0.42885554 0.92854816 -0.47026226
0.965811 0.75807875 0.43110335 -1.0098383 -0.6483311 -0.6152972
-0.5216276 -0.63780814 0.9183269 -0.91891444 -1.036427 1.0302682
-1.0429531 -0.47352314 -0.15649249 0.16982554 -0.5457818 0.512825
-0.8712062 -0.6529124 -0.6283604 -1.1288209 1.0132375 0.51677763
0.5170388 -0.9723579 -0.24652894 0.27448624 0.6794884 0.16785496
0.72845274 0.62323767 -1.1006261 0.09867156 -0.9563998 0.61188143
-0.16472985 0.08116743 -0.94903636 -0.34606537 -0.08370335 -0.15427722
0.86135715 0.64785695 1.4747293 -0.16418749 -0.13215429 1.1107985
2.0441353 0.0237242 1.122327 0.3432455 0.813482 0.37290555
0.74935645 0.5219837 0.4310655 0.9049328 0.3750479 -0.55192566
-0.58756495 -0.05983122 0.10420159 0.7894059 -0.3306062 -0.09707043
-0.51481307 0.19391987 -1.4686276 -0.6866447 -0.34494066 2.3112524
1.0078295 0.03392195 -0.27139398 0.00728518 0.8748948 0.2153134
-0.8508709 -0.199429 -0.2382995 0.3319936 0.96112686 0.75113213
-0.8122194 0.8009778 6.275326 1.2926952 -1.1204255 0.16225225
0.6263487 0.21942325 -0.26284578 -0.03439616 -1.0776651 0.4872697
0.5013014 0.05866417 0.71212274 0.40703544 0.30028746 -0.28450578
-0.28375864 1.2042961 0.19855282 -1.3404131 -0.14070271 -0.08123186
0.83251894 -0.29973504 0.04193503 -0.03887356 -0.26715475 -0.31505835
0.5264158 0.9237362 1.3630867 -0.75546354 0.38441592 0.029227
-1.2807279 -0.1030015 -0.5627404 0.42571822 0.30749795 0.76171327
-0.30278763 0.42810124 0.72370905 0.49148238 -0.71610534 1.2158278
-0.34025657 0.32380024 -0.38850895 0.081114 -0.11779498 -0.49380285
0.5024914 1.1597888 0.29547456 0.1307136 -0.02315431 0.843129
-0.2551578 0.16898353 0.06297805 0.4495243 0.40101144 1.6326445
-0.67268014 -0.30302635 -0.36715627 1.0571707 0.2584714 0.49684554
-0.95442426 -0.8328674 0.0557015 0.32998857 0.4143281 -0.15778972
-0.2054768 -0.94072783 0.09002856 -0.5129095 0.08809966 -1.1734564
-1.0509421 0.19114073 -0.03283196 -1.1726489 0.34247684 -0.6488239
-0.56107324 0.75554985 -2.3477905 -1.1996812 -0.8416002 0.6245784
0.18082012 0.38445547 0.26189935 0.51911944 -0.2945576 0.65587926
0.39923686 -0.27974102 1.0236416 -1.0521648 0.02100632 0.95059824
-0.4073863 0.38735083 0.46365386 -0.4052515 -1.348814 -1.0547189
0.47054633 -0.01658204 0.5454884 -0.2711072 -1.0623528 -0.03148723
0.08400678 0.1743963 0.13364874 -0.4946247 -0.22666244 -0.791828
-1.2478224 0.4886365 0.7629349 -0.45409998 0.09765332 0.14848186
0.8067632 -0.70973134 -1.0675418 0.42207038 0.76817125 -0.9658011
1.220978 0.53131926 0.44093707 -0.55512404 0.04746165 -0.9679729
-0.46583787 -0.77399766 -0.38888755 1.2629884 -0.06608272 -0.6934467
0.6287486 0.46331018 -0.11280582 -0.7690772 -0.71566385 -0.6630847
-0.23167378 -0.22252648 0.49882075 0.45603132 -0.6512847 0.03125167
-0.56445205 0.28774023 1.1824723 0.20733562 0.54386276 -1.0207307
-0.26953253 -0.28056365 -0.15955205 -1.2516668 -0.20449695 -0.4875271
0.3370876 -1.4145283 0.33104137 -0.5235343 -0.40805382 0.20914528
-0.32380787 0.7209977 0.14876883 0.13660623 -0.46552408 0.51236016
1.7950782 -0.10291457 -0.30652234 -0.21907176 -0.45453218 0.4718148
0.6879586 -0.14179061 -0.15943034 -0.25561085 0.3443174 -0.05512868
0.64439225 -1.1025321 0.5620298 0.01402807 0.44641855 -0.82779574
0.50139004 -0.8473528 -0.02565628 0.28762797 0.01082066 -0.46761718
0.2898355 0.6557616 -0.10921973 -0.45256084 1.2421774 0.20176895
-0.86747146 0.5559951 0.101545 -0.26564392 0.8428657 -0.3051363
-0.6796261 -0.38470304 -0.26800966 0.1579277 0.5949134 -0.2777872
0.8084781 -1.2763922 -0.51135445 0.24549171 0.24515523 -0.02959524
0.7560215 1.0356618 -0.81802493 0.02576152 -0.15573885 -0.6077442
-1.1618987 0.48450282 0.3575292 -0.18317078 -0.76500076 0.47165164
-0.02272851 -0.05766862 0.133028 -0.29815432 -0.11114769 -0.01131851
0.67058563 0.613892 0.01912853 -0.34929267 0.05099475 1.1722642
-0.01849311 0.00856029 1.2702992 -0.35712942 -0.26463825 0.10025485
1.0484397 0.15286623 -1.7578615 -0.3220388 -0.67944723 -0.71226865
0.61924696 -0.72056705 -1.0354596 0.8536034 0.93870646 -0.16281423
1.5418242 -0.54549456 1.0655967 0.3181678 0.19593522 -1.4446337
0.16553976 0.2512248 0.60722536 -1.1603343 0.3289927 -0.57642365
-0.38133135 1.2340248 0.46999615 0.12270901 0.39388728 0.28223276
-0.15288629 -0.09259778 -0.26939663 -0.1782906 0.3252184 0.47664973
-0.0741193 -0.32536 -0.18879277 -0.20268826 0.36403573 0.24804875
0.5395624 0.5221785 -0.54748315 -0.8104929 -0.4223711 0.28942552
-0.22185674 -0.31483042 0.3313527 0.6489475 0.06383982 0.6034571
-0.03265263 0.07719653 0.28463387 -0.31741348 0.7163053 0.02937496
-0.28654885 0.18770155 -0.30490464 -0.64793175 -0.6371862 -0.26177678
-1.1488476 -0.3760194 -0.7198317 -0.4621244 0.9438602 0.33751717
0.19216166 0.54386467 0.741154 -0.87763155 -0.23463853 -0.5380957
-0.5720548 0.20664567 0.36356515 -0.70383734 -0.50699854 0.10542794]
|
[10.811840057373047, -2.4935972690582275]
|
c3d4344b-69f9-4880-b0a0-7d9837ec7c6f
|
a-unified-framework-for-sparse-relaxed
|
1807.05411
| null |
http://arxiv.org/abs/1807.05411v4
|
http://arxiv.org/pdf/1807.05411v4.pdf
|
A Unified Framework for Sparse Relaxed Regularized Regression: SR3
|
Regularized regression problems are ubiquitous in statistical modeling,
signal processing, and machine learning. Sparse regression in particular has
been instrumental in scientific model discovery, including compressed sensing
applications, variable selection, and high-dimensional analysis. We propose a
broad framework for sparse relaxed regularized regression, called SR3. The key
idea is to solve a relaxation of the regularized problem, which has three
advantages over the state-of-the-art: (1) solutions of the relaxed problem are
superior with respect to errors, false positives, and conditioning, (2)
relaxation allows extremely fast algorithms for both convex and nonconvex
formulations, and (3) the methods apply to composite regularizers such as total
variation (TV) and its nonconvex variants. We demonstrate the advantages of SR3
(computational efficiency, higher accuracy, faster convergence rates, greater
flexibility) across a range of regularized regression problems with synthetic
and real data, including applications in compressed sensing, LASSO, matrix
completion, TV regularization, and group sparsity. To promote reproducible
research, we also provide a companion MATLAB package that implements these
examples.
|
['J. Nathan Kutz', 'Aleksandr Y. Aravkin', 'Travis Askham', 'Peng Zheng', 'Steven L. Brunton']
|
2018-07-14
| null | null | null | null |
['model-discovery']
|
['miscellaneous']
|
[ 4.77965057e-01 -2.58980989e-01 -4.19606179e-01 -2.35970601e-01
-1.13342226e+00 -1.41830519e-01 -6.06874749e-02 -1.06533632e-01
-8.63429829e-02 1.06586385e+00 4.41119462e-01 -6.52005747e-02
-4.13219959e-01 -3.42590809e-01 -7.34816074e-01 -8.71027589e-01
-3.73811066e-01 2.76361555e-01 -6.54770195e-01 -1.01477973e-01
1.36279846e-02 3.39443356e-01 -8.74204338e-01 -2.47454479e-01
1.05079150e+00 9.47157621e-01 1.93382367e-01 1.61908064e-02
3.46074522e-01 4.72810149e-01 -6.19134344e-02 1.49046987e-01
3.64931077e-01 -3.24259549e-01 -2.70081311e-01 4.02211249e-01
1.13516197e-01 2.91543394e-01 -1.50458351e-01 1.09643960e+00
4.52820182e-01 3.96098286e-01 3.49365085e-01 -9.02488172e-01
-7.32751548e-01 5.80228686e-01 -1.41263139e+00 6.29014075e-02
3.18354487e-01 -3.70986998e-01 9.77921963e-01 -1.35406911e+00
4.39118177e-01 1.30675995e+00 9.49365854e-01 1.79344758e-01
-1.62545025e+00 -6.48368597e-01 1.79630816e-01 -3.18683326e-01
-1.61543739e+00 -7.22236454e-01 5.80635011e-01 -5.89861572e-01
1.96307287e-01 5.22670746e-01 4.56080765e-01 8.00527036e-01
-1.53961658e-01 9.87480044e-01 1.18619335e+00 -3.03803205e-01
3.56993705e-01 -3.80194262e-02 2.32311815e-01 5.52579463e-01
4.05714065e-01 -2.19108574e-02 -5.83147943e-01 -5.85884690e-01
7.39363194e-01 1.47773638e-01 -7.24990666e-01 -3.21849674e-01
-1.50214040e+00 1.17593670e+00 1.09293815e-02 2.24215448e-01
-5.48765063e-01 1.70091555e-01 3.50781053e-01 2.51328647e-01
9.73317862e-01 5.13866186e-01 -2.00727731e-01 1.38661310e-01
-1.28677046e+00 3.53984356e-01 6.96605444e-01 8.50725174e-01
6.14040613e-01 7.14218318e-01 -2.57069886e-01 1.32921088e+00
2.04453409e-01 1.05535889e+00 2.89902568e-01 -1.05314469e+00
5.06900609e-01 -8.45267177e-02 -8.72555003e-02 -1.49226284e+00
-5.12820482e-01 -8.75455439e-01 -1.49117768e+00 -3.60213876e-01
-2.62003206e-02 -6.58808127e-02 -5.59494197e-01 1.67477131e+00
2.80017018e-01 4.78636235e-01 -2.10307851e-01 1.05644560e+00
7.57423341e-01 7.08523273e-01 -3.42763782e-01 -8.50682199e-01
8.10390532e-01 -7.36050129e-01 -1.01877964e+00 -4.21426982e-01
4.16062862e-01 -6.31392539e-01 6.75690114e-01 4.88105118e-01
-1.18199217e+00 -2.89286107e-01 -7.65074909e-01 1.03115877e-02
4.01993424e-01 1.35005772e-01 8.96748841e-01 3.43893141e-01
-8.00809681e-01 3.66875410e-01 -9.74001348e-01 3.02989148e-02
5.70439637e-01 3.89378965e-01 -3.41659039e-01 -4.34445471e-01
-7.19247520e-01 3.78447324e-01 -2.99818426e-01 2.61978209e-01
-5.93406558e-01 -1.03699207e+00 -1.10410416e+00 1.25235002e-02
4.06672001e-01 -6.00181401e-01 4.89184707e-01 -8.76191020e-01
-1.14551008e+00 8.82373512e-01 -6.94238603e-01 -3.23130667e-01
3.06371897e-01 -2.98950195e-01 -1.84970111e-01 -1.55755982e-01
4.04243976e-01 -1.02198899e-01 1.11310446e+00 -9.28864300e-01
-2.23942157e-02 -5.75337470e-01 -7.89496124e-01 -7.66811967e-02
-4.49401066e-02 -1.13550417e-01 -3.71153682e-01 -1.31107867e+00
5.70374906e-01 -1.01269090e+00 -8.35366428e-01 -1.35199443e-01
-5.21472573e-01 4.06727374e-01 5.55950999e-01 -8.86824965e-01
1.11034584e+00 -2.51912022e+00 8.08034301e-01 5.78902185e-01
3.67827535e-01 -1.29454643e-01 -1.78440958e-01 -3.44865769e-02
-2.15083331e-01 6.31313473e-02 -8.22384953e-01 -4.34220731e-01
-3.40763927e-01 1.51354045e-01 -3.80982906e-01 1.14549243e+00
-2.84381270e-01 7.61141479e-01 -8.39827597e-01 -1.39873773e-01
5.52307665e-02 5.17627001e-01 -7.88738191e-01 -1.73161924e-01
7.23787099e-02 1.09643114e+00 -7.35437870e-01 9.36687410e-01
7.19509959e-01 -4.62018967e-01 6.04510009e-02 -9.68240425e-02
-5.72092347e-02 -2.29887277e-01 -1.52356040e+00 1.66898263e+00
-4.53601718e-01 7.02256560e-01 9.14811373e-01 -1.51954424e+00
9.56962645e-01 2.36721471e-01 1.02912199e+00 -5.04118919e-01
-2.91407585e-01 3.20366830e-01 -5.36401689e-01 -4.43972200e-01
3.76940548e-01 -1.31979734e-01 1.81488678e-01 3.68144423e-01
-4.57604766e-01 -1.82928443e-01 1.61118716e-01 1.73335135e-01
7.94170260e-01 -1.28308535e-01 4.90128666e-01 -7.25290716e-01
4.83867019e-01 -2.51414925e-01 9.95802701e-01 6.81144416e-01
2.91015714e-01 7.16414094e-01 2.97031164e-01 -2.22168103e-01
-7.20277786e-01 -6.24213874e-01 -5.68095267e-01 1.04769623e+00
1.91540734e-04 -2.96958327e-01 -4.51925509e-02 1.44986317e-01
3.58811706e-01 4.10338670e-01 -5.03179848e-01 2.92567670e-01
-7.97498941e-01 -1.17634988e+00 -1.42074923e-03 2.69502640e-01
1.68344319e-01 -3.83895457e-01 1.67887196e-01 2.15266511e-01
-4.29320872e-01 -1.24995553e+00 -6.94244921e-01 8.11570361e-02
-1.20484781e+00 -9.13022101e-01 -8.65576267e-01 -5.93975723e-01
7.89919317e-01 6.76575005e-01 1.00335419e+00 2.92278957e-02
-3.23599547e-01 5.27674377e-01 -1.73767611e-01 -1.13715760e-01
-1.11369463e-02 -3.84807885e-01 3.20686430e-01 2.69901097e-01
-4.11659688e-01 -7.22621083e-01 -2.33399659e-01 4.60147709e-01
-7.54026830e-01 -2.15946838e-01 4.12151307e-01 1.19859779e+00
1.30932868e+00 -2.82299608e-01 6.75720930e-01 -1.32975042e+00
5.49653471e-01 -8.28777671e-01 -7.11244941e-01 6.34999722e-02
-5.51174998e-01 -3.11208777e-02 4.24679130e-01 -3.19567263e-01
-7.33866274e-01 4.41684462e-02 9.05769914e-02 -5.63351989e-01
6.67759120e-01 1.18452680e+00 2.02050626e-01 -4.02095824e-01
7.64483154e-01 3.00861210e-01 2.71092147e-01 -5.31654835e-01
2.94421315e-01 2.23799899e-01 3.42297852e-01 -8.09242964e-01
7.95624197e-01 6.67740285e-01 4.61091578e-01 -1.42151427e+00
-1.10275829e+00 -7.38469183e-01 -2.12017909e-01 3.09027821e-01
4.45963234e-01 -1.24169838e+00 -4.46845770e-01 1.70589350e-02
-6.55422449e-01 -2.62199879e-01 -5.66450119e-01 6.73373401e-01
-6.46777809e-01 5.20682454e-01 -3.78629506e-01 -7.58150101e-01
-3.55987191e-01 -1.14700651e+00 9.96231437e-01 -1.64259136e-01
-5.12546524e-02 -1.23942578e+00 5.32923751e-02 4.74827647e-01
6.52247012e-01 6.38761640e-01 7.03252614e-01 -2.33144313e-01
-1.85054660e-01 -1.32864192e-01 -5.22988699e-02 3.18433404e-01
1.78329781e-01 -1.62331283e-01 -5.86208642e-01 -6.18422985e-01
3.53495091e-01 -3.51106972e-01 9.53597248e-01 1.22840786e+00
1.59315515e+00 -4.93149459e-01 -4.89938647e-01 1.39986086e+00
1.15367258e+00 -3.98704708e-01 3.30267847e-01 -7.85584971e-02
7.57744849e-01 3.58810574e-01 5.96282780e-01 7.98952222e-01
-1.21567901e-02 6.97598398e-01 2.33991891e-01 -4.69511777e-01
1.35123581e-01 2.63362706e-01 2.55330116e-01 1.15370381e+00
-1.27188668e-01 1.57173723e-01 -5.00713706e-01 5.10968387e-01
-2.00047278e+00 -1.08078301e+00 -5.34991145e-01 2.42252827e+00
6.27216756e-01 -6.43263876e-01 6.46909103e-02 1.77399978e-01
8.24592531e-01 4.99690056e-01 -5.01673281e-01 -5.58471717e-02
-5.70689738e-01 4.04257029e-01 5.98360002e-01 5.93564987e-01
-1.07024264e+00 5.59913397e-01 6.96434212e+00 8.21239471e-01
-1.03475106e+00 2.39039525e-01 8.49726319e-01 -2.52007753e-01
-5.11771560e-01 -1.93557158e-01 -5.26768446e-01 2.12771252e-01
3.04919690e-01 -3.35124165e-01 9.50721145e-01 7.64292181e-01
6.81885064e-01 1.19259313e-01 -1.13244438e+00 1.38511002e+00
1.81117013e-01 -1.43478668e+00 -3.47146004e-01 1.82110026e-01
1.21815896e+00 2.26822317e-01 2.60503978e-01 1.93276554e-01
-1.10366230e-03 -1.34291673e+00 3.48566383e-01 2.83980787e-01
1.13254511e+00 -3.86984915e-01 2.40326568e-01 1.37803569e-01
-9.95415986e-01 9.35549755e-03 -4.72353548e-01 1.38213798e-01
3.32986414e-01 1.40096116e+00 -1.69745162e-01 5.35963833e-01
4.29424107e-01 1.38123798e+00 1.09609164e-01 9.35214698e-01
-2.84732003e-02 8.87611330e-01 -4.64789987e-01 4.18077558e-01
-1.23556249e-01 -8.73266160e-01 1.04770017e+00 1.19921041e+00
4.34342682e-01 2.87662655e-01 4.64369982e-01 9.51735914e-01
-7.58075938e-02 2.35256150e-01 -4.52424675e-01 6.67203292e-02
4.86773252e-01 1.11834967e+00 -3.38581711e-01 1.09547488e-01
-5.21389306e-01 4.20058280e-01 3.67393307e-02 7.75940478e-01
-5.24003327e-01 1.78224534e-01 5.56743979e-01 1.65319294e-01
3.09577286e-01 -4.44868058e-01 -7.25515127e-01 -1.39452696e+00
-1.22373328e-02 -1.34887421e+00 4.54305172e-01 -2.50871748e-01
-1.27223194e+00 2.65485555e-01 -1.82499930e-01 -1.05710876e+00
3.52037232e-03 -2.85915881e-01 -3.55400890e-01 8.24741542e-01
-1.36249661e+00 -7.33409941e-01 -2.93508083e-01 8.07564437e-01
4.87846732e-01 -2.34776676e-01 6.52200282e-01 5.94647288e-01
-7.88134515e-01 3.36909711e-01 6.64969623e-01 -1.60320804e-01
5.14897406e-01 -8.16071391e-01 -1.31087527e-01 9.05882418e-01
3.03555373e-02 6.03518128e-01 7.74881959e-01 -5.45095801e-01
-1.99918556e+00 -1.26206195e+00 3.35060805e-01 1.58797279e-01
6.86147153e-01 1.22407395e-02 -7.77093410e-01 9.13277745e-01
-5.93406558e-01 5.24527133e-01 9.30786014e-01 5.44017553e-01
-1.57550961e-01 -2.35115722e-01 -1.07999718e+00 1.56628877e-01
7.94416487e-01 -5.29792830e-02 6.96256831e-02 9.45066273e-01
3.71085405e-01 -5.69482148e-01 -1.01397586e+00 5.21740437e-01
4.21950996e-01 -4.82624114e-01 1.17935276e+00 -4.31439012e-01
3.66976321e-01 -9.81997326e-02 -6.29521966e-01 -1.29408777e+00
-6.39487863e-01 -9.86832082e-01 -3.92282188e-01 7.18215823e-01
4.52025890e-01 -7.28817046e-01 6.45942509e-01 4.37194884e-01
-3.39752674e-01 -9.34303343e-01 -1.13940585e+00 -7.02655196e-01
-1.80572540e-01 -4.34963256e-01 1.12065293e-01 1.42725718e+00
-7.63285682e-02 1.23583160e-01 -9.39229429e-01 1.07832849e-01
9.89785016e-01 3.54430318e-01 8.61390054e-01 -1.16905081e+00
-6.71407044e-01 -2.47457713e-01 -1.76376820e-01 -1.42932200e+00
2.21874744e-01 -1.09221876e+00 -4.70550694e-02 -1.28716111e+00
2.31654271e-01 -1.10082114e+00 -8.25619921e-02 3.04672658e-01
-3.99282217e-01 2.58871973e-01 9.70738158e-02 5.65148950e-01
-4.02736038e-01 7.52591848e-01 1.36128306e+00 -3.47532034e-01
-5.36357880e-01 2.95665175e-01 -8.98628891e-01 6.74421549e-01
3.27532440e-01 -5.38885653e-01 -7.90228695e-02 -6.27745330e-01
4.38990474e-01 5.74339986e-01 1.19286671e-01 -6.32197738e-01
-4.69035916e-02 -4.58328396e-01 5.56272343e-02 -6.70084432e-02
4.29363400e-01 -5.94844460e-01 3.73212308e-01 2.96651125e-01
-3.60198289e-01 -2.38209575e-01 -5.29944040e-02 7.58097947e-01
-3.54773402e-01 -1.05592459e-01 9.74876106e-01 -6.51264414e-02
-1.40858695e-01 6.50959015e-01 -7.08450451e-02 5.11138380e-01
7.53766179e-01 -2.16315649e-02 8.19291268e-03 -7.29733169e-01
-9.00596380e-01 4.95536625e-01 1.60809129e-01 -1.32354409e-01
7.42434561e-01 -1.24036419e+00 -1.32605672e+00 2.46293500e-01
-2.24933475e-01 2.73830742e-01 1.35213614e-01 1.50084484e+00
-3.28683406e-01 1.93235055e-01 4.88332391e-01 -8.57594848e-01
-9.92176592e-01 4.38230813e-01 1.90704584e-01 -2.53978789e-01
-7.95102894e-01 8.43089342e-01 2.38727584e-01 -3.40263873e-01
2.53115803e-01 -5.81914634e-02 1.50757819e-01 -2.38001317e-01
5.14245510e-01 8.77676308e-01 -1.61732510e-01 -6.07509375e-01
-3.73161256e-01 7.79026091e-01 3.38408411e-01 1.63523197e-01
1.80668712e+00 -4.64848466e-02 -5.65296233e-01 4.41168994e-01
1.12977481e+00 6.04172051e-01 -8.98062050e-01 -5.35188437e-01
-2.19589025e-01 -6.84821308e-01 3.85126531e-01 -1.56126797e-01
-1.52382445e+00 4.08185303e-01 5.05719669e-02 1.93420857e-01
1.13295782e+00 -1.60119444e-01 6.43511713e-01 3.33569914e-01
3.54255229e-01 -7.45484889e-01 -3.21661890e-01 4.09197152e-01
1.24927163e+00 -1.31641603e+00 8.72645676e-01 -6.87467575e-01
-5.21277666e-01 8.00232351e-01 -8.65987912e-02 -3.32238793e-01
7.80647874e-01 3.25362414e-01 -3.58876973e-01 -2.14394376e-01
-3.13045859e-01 2.83406824e-01 3.73428345e-01 4.20740485e-01
6.22761726e-01 1.48229510e-01 -4.37972814e-01 4.35426682e-01
9.29753557e-02 -1.14158772e-01 3.22028428e-01 5.61153233e-01
-2.18897969e-01 -6.80482745e-01 -7.16414809e-01 8.23912859e-01
-6.73410416e-01 -2.81816542e-01 2.26027369e-01 5.22984862e-01
-3.89499038e-01 1.07434261e+00 -4.15758401e-01 2.14529857e-01
-1.29428292e-02 -6.10152304e-01 3.00109029e-01 -6.56558096e-01
-2.16076568e-01 4.45880353e-01 6.51746541e-02 -7.50029147e-01
-7.39074349e-01 -9.12940502e-01 -9.18904305e-01 -2.62109756e-01
-4.46679264e-01 4.06834364e-01 6.74423397e-01 9.43200409e-01
4.50612128e-01 3.74650002e-01 9.02938485e-01 -9.50958371e-01
-6.64619148e-01 -8.07454944e-01 -9.84248936e-01 3.29448134e-01
6.79515719e-01 -5.84790647e-01 -5.85850894e-01 1.54392989e-02]
|
[6.974621772766113, 4.437622547149658]
|
88b8ee25-a5b2-4de5-a62b-6966ef782813
|
practical-stereo-matching-via-cascaded
|
2203.11483
| null |
https://arxiv.org/abs/2203.11483v1
|
https://arxiv.org/pdf/2203.11483v1.pdf
|
Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation
|
With the advent of convolutional neural networks, stereo matching algorithms have recently gained tremendous progress. However, it remains a great challenge to accurately extract disparities from real-world image pairs taken by consumer-level devices like smartphones, due to practical complicating factors such as thin structures, non-ideal rectification, camera module inconsistencies and various hard-case scenes. In this paper, we propose a set of innovative designs to tackle the problem of practical stereo matching: 1) to better recover fine depth details, we design a hierarchical network with recurrent refinement to update disparities in a coarse-to-fine manner, as well as a stacked cascaded architecture for inference; 2) we propose an adaptive group correlation layer to mitigate the impact of erroneous rectification; 3) we introduce a new synthetic dataset with special attention to difficult cases for better generalizing to real-world scenes. Our results not only rank 1st on both Middlebury and ETH3D benchmarks, outperforming existing state-of-the-art methods by a notable margin, but also exhibit high-quality details for real-life photos, which clearly demonstrates the efficacy of our contributions.
|
['Shuaicheng Liu', 'Haoqiang Fan', 'Jiangyu Liu', 'Lei Yang', 'Ziwei Yan', 'Tao Cai', 'Pengfei Xiong', 'Peisen Wang', 'Jiankun Li']
|
2022-03-22
| null |
http://openaccess.thecvf.com//content/CVPR2022/html/Li_Practical_Stereo_Matching_via_Cascaded_Recurrent_Network_With_Adaptive_Correlation_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Practical_Stereo_Matching_via_Cascaded_Recurrent_Network_With_Adaptive_Correlation_CVPR_2022_paper.pdf
|
cvpr-2022-1
|
['stereo-matching-1']
|
['computer-vision']
|
[ 4.64400589e-01 -2.46644884e-01 9.43837687e-02 -4.44752872e-01
-6.65119410e-01 -2.18906417e-01 4.91341293e-01 -2.58135498e-01
-4.57268655e-01 6.62878335e-01 4.05669153e-01 -6.39887080e-02
-5.51551543e-02 -6.55707181e-01 -8.33346963e-01 -3.61581206e-01
2.30427295e-01 8.89292359e-02 4.66541052e-01 -3.59885573e-01
2.75517255e-01 4.29029793e-01 -1.77357650e+00 4.52332757e-02
1.06738150e+00 1.09350991e+00 7.32481480e-02 3.10476989e-01
3.57040912e-01 7.33287156e-01 -2.70242751e-01 -5.21025717e-01
7.78078496e-01 -1.62375838e-01 -3.73202771e-01 2.73575515e-01
1.14970446e+00 -9.01936531e-01 -5.07104397e-01 1.09455359e+00
5.64617932e-01 -1.91171411e-02 2.39065841e-01 -9.62965012e-01
-4.34096426e-01 1.03820600e-01 -9.55095053e-01 -4.80958521e-02
2.59501785e-01 4.05172527e-01 8.30786884e-01 -6.11440837e-01
5.50505638e-01 1.16293573e+00 8.47266674e-01 3.43335629e-01
-1.30150104e+00 -9.95569110e-01 2.69216686e-01 2.70183831e-01
-1.24326098e+00 -6.03146434e-01 6.30061686e-01 -3.33221197e-01
7.88109303e-01 -8.38562027e-02 6.05757892e-01 1.03863692e+00
1.19206019e-01 4.49718237e-01 1.04683328e+00 -2.14081053e-02
7.75405765e-02 -4.66226429e-01 -2.10210830e-01 4.98216778e-01
3.66817325e-01 3.28933060e-01 -5.89516938e-01 1.26381725e-01
1.06896126e+00 1.81260973e-01 -6.04672372e-01 -5.11891901e-01
-1.34347177e+00 4.13716316e-01 7.81217873e-01 6.54114876e-03
-3.75349313e-01 1.21454291e-01 1.23123765e-01 8.10764357e-02
4.20235693e-01 4.45837647e-01 -2.71820992e-01 -4.47710860e-04
-9.05809164e-01 3.20358813e-01 3.85480821e-01 9.64181364e-01
8.83887768e-01 -3.42363715e-02 1.25706226e-01 8.30659330e-01
-6.74656630e-02 3.77134502e-01 3.30831349e-01 -1.19261944e+00
8.39926004e-01 4.99348432e-01 1.43204361e-01 -9.92921829e-01
-3.53694856e-01 -4.71777231e-01 -1.26092291e+00 2.78400838e-01
4.30953503e-01 1.34004682e-01 -8.26114595e-01 1.54534543e+00
1.45921692e-01 3.85388196e-01 -1.73300281e-01 1.17814791e+00
7.87770331e-01 2.92701334e-01 -2.05375627e-01 2.23287344e-01
1.08209002e+00 -9.91142213e-01 -3.51608485e-01 -5.24890840e-01
2.09867701e-01 -7.12458909e-01 9.58227634e-01 4.43886817e-01
-1.20392370e+00 -6.80300653e-01 -1.26653612e+00 -4.86079514e-01
3.91025469e-02 -7.55602792e-02 8.08339596e-01 3.44321907e-01
-1.14912391e+00 6.92780495e-01 -6.23606503e-01 -2.43955895e-01
4.86484170e-01 4.28266913e-01 -5.13069212e-01 -4.65314716e-01
-9.29077804e-01 4.93390977e-01 3.70692052e-02 3.19750041e-01
-4.93531227e-01 -9.71595705e-01 -1.03360355e+00 2.56425496e-02
4.15341616e-01 -1.04475558e+00 1.08054841e+00 -8.69350970e-01
-1.41848826e+00 8.52662504e-01 -1.72935903e-01 -4.23089296e-01
7.86200166e-01 -4.31213439e-01 -8.64360332e-02 -1.06010698e-01
5.26192337e-02 9.18153346e-01 7.06879497e-01 -1.16622698e+00
-8.40073824e-01 -5.25332987e-01 3.37615669e-01 2.45295510e-01
-2.67562836e-01 -4.60755318e-01 -7.47258067e-01 -6.60344779e-01
4.41201091e-01 -8.37143302e-01 -4.94487226e-01 2.56635189e-01
-4.35280263e-01 3.49582165e-01 4.31531429e-01 -5.28776586e-01
8.90390754e-01 -2.28036070e+00 6.66041523e-02 -9.30092484e-02
4.70848441e-01 2.48122439e-01 -1.24157801e-01 -2.19783075e-02
7.38231912e-02 -2.54721850e-01 -3.94164234e-01 -6.92620337e-01
-2.14143202e-01 -3.65038030e-02 -3.02678496e-01 4.08224463e-01
1.38898343e-01 6.71678007e-01 -8.21388066e-01 -2.42804632e-01
4.64619845e-01 5.57505548e-01 -7.93072164e-01 1.48363918e-01
-8.90645608e-02 6.45122170e-01 -2.36184031e-01 4.99356896e-01
9.23611820e-01 -3.52326542e-01 -9.11653563e-02 -5.45198143e-01
-2.67127782e-01 2.98431993e-01 -1.36861956e+00 1.90387583e+00
-5.55009067e-01 7.57572174e-01 1.23423569e-01 -5.62369347e-01
7.65988648e-01 -1.62163764e-01 2.28444844e-01 -1.11113620e+00
4.40147799e-03 2.72741467e-01 -1.06034517e-01 -2.17986971e-01
7.90570199e-01 4.29598019e-02 1.89792544e-01 -5.62707633e-02
-3.42741936e-01 -3.18357617e-01 3.59353386e-02 -4.93437890e-03
1.03762496e+00 1.46438003e-01 1.92314088e-01 -1.32002681e-01
4.74214971e-01 -3.37148160e-01 9.21945393e-01 5.19630253e-01
-3.66800688e-02 1.35307777e+00 3.04882705e-01 -5.31695604e-01
-1.08947051e+00 -9.65235591e-01 -2.26362962e-02 5.19581795e-01
5.46185374e-01 -2.64696747e-01 -5.16836643e-01 -1.84440151e-01
4.04363126e-02 9.21581835e-02 -4.21152234e-01 6.83846474e-02
-6.58608437e-01 -6.38045132e-01 2.38933116e-01 7.26968765e-01
1.22770655e+00 -5.69088280e-01 -6.03420377e-01 1.54662818e-01
-2.20472097e-01 -1.64931309e+00 -5.57010710e-01 -8.87040272e-02
-9.86705840e-01 -1.12895381e+00 -8.08899999e-01 -6.51541173e-01
6.02154374e-01 6.65486634e-01 1.21196640e+00 1.81184933e-01
-2.06588283e-01 -1.32418618e-01 6.95870742e-02 9.00572687e-02
1.18849002e-01 1.32411093e-01 -2.47224525e-01 1.12217711e-02
-4.81400937e-02 -8.84520411e-01 -1.13611805e+00 5.44707179e-01
-9.73503768e-01 3.95572752e-01 6.87559962e-01 7.90661037e-01
5.56578398e-01 -2.03655567e-02 4.26579565e-02 -6.74951196e-01
2.14682639e-01 -6.83442652e-02 -1.02491128e+00 -6.45134971e-02
-4.86428797e-01 8.89389031e-03 6.64628685e-01 -1.46307871e-01
-1.09369242e+00 7.67597333e-02 -2.26590753e-01 -3.78737390e-01
-1.02843001e-01 -1.63500205e-01 -2.86328048e-01 -2.13716775e-01
4.01935518e-01 -4.83817980e-02 -1.35640860e-01 -3.66696566e-01
4.03283350e-02 4.68006223e-01 9.46318924e-01 -4.86011386e-01
8.69493008e-01 9.45663929e-01 2.64152467e-01 -7.33263791e-01
-9.87497985e-01 -3.03969800e-01 -4.89277035e-01 4.20536026e-02
8.28496099e-01 -1.46669257e+00 -8.09327126e-01 8.45988274e-01
-9.68087852e-01 -5.26068032e-01 7.32100904e-02 4.65050966e-01
-3.08499068e-01 5.45738041e-01 -6.73223853e-01 -1.93329036e-01
-3.29390168e-01 -1.39512789e+00 1.33059919e+00 3.81765842e-01
-3.79048549e-02 -6.39334738e-01 -1.72368869e-01 6.83234692e-01
5.12724638e-01 2.31279254e-01 6.42857134e-01 3.04241359e-01
-1.15798330e+00 2.21254319e-01 -6.55781925e-01 4.00780886e-01
1.26659825e-01 -1.01594985e-01 -1.08397281e+00 -3.51987988e-01
-1.45372674e-01 -2.49387369e-01 9.55025673e-01 3.53825361e-01
1.24128568e+00 2.50382513e-01 -1.04009673e-01 1.31290460e+00
1.38956964e+00 -6.74722120e-02 9.29912329e-01 6.17147386e-01
9.19012308e-01 4.94136930e-01 4.10073876e-01 2.66355038e-01
7.23888695e-01 9.30432856e-01 5.89763582e-01 -4.87489372e-01
-3.75727475e-01 -4.00086045e-01 -7.70982578e-02 5.33671677e-01
-8.53951350e-02 -9.55554321e-02 -7.05395043e-01 4.55173045e-01
-1.72754765e+00 -6.71739161e-01 -1.02667391e-01 2.45535374e+00
5.42091072e-01 4.19793338e-01 -1.04425125e-01 1.09032923e-02
5.33147991e-01 4.22666341e-01 -7.36045778e-01 3.36779714e-01
-4.43772763e-01 1.48486510e-01 6.80291355e-01 3.91632289e-01
-1.02455282e+00 7.66481221e-01 5.16778708e+00 6.08933508e-01
-1.36454260e+00 -2.30420589e-01 8.90641332e-01 -1.61155909e-01
-2.61925548e-01 -2.02436447e-02 -8.17048430e-01 4.87133652e-01
1.94759488e-01 3.00282031e-01 4.03939575e-01 4.91872042e-01
4.99498621e-02 -2.55073071e-01 -1.15019441e+00 1.40848875e+00
8.96539912e-02 -1.37493420e+00 -6.31546080e-02 2.41369516e-01
1.15662313e+00 2.32611984e-01 1.16995208e-01 3.30170877e-02
1.24082424e-01 -9.03765917e-01 6.04040980e-01 3.41836840e-01
7.29250610e-01 -6.28497303e-01 7.56847322e-01 1.17011018e-01
-1.19583511e+00 -5.06278202e-02 -4.00835901e-01 -1.93034813e-01
1.83984205e-01 8.46390784e-01 -1.15729891e-01 6.33778512e-01
8.89036655e-01 1.11483276e+00 -6.92502916e-01 1.31194460e+00
-2.39161521e-01 3.74339521e-02 -3.41011137e-01 4.60142076e-01
1.66062430e-01 -2.21680298e-01 1.73569039e-01 7.30209827e-01
3.60657871e-01 1.36230424e-01 -1.73519149e-01 6.96434796e-01
-2.97171682e-01 -2.94601977e-01 -5.04426301e-01 5.55363536e-01
5.01984298e-01 1.02420044e+00 -5.20581543e-01 -1.04489699e-01
-5.77914178e-01 1.08624303e+00 3.59910876e-01 3.21673870e-01
-5.95746696e-01 -2.22588807e-01 1.13298559e+00 3.90804797e-01
2.95294583e-01 -3.05212289e-01 -4.19951439e-01 -1.41557384e+00
3.57006729e-01 -9.12413359e-01 1.54407844e-01 -8.21274340e-01
-1.18816173e+00 6.66582465e-01 -3.24739426e-01 -1.40361738e+00
-8.57000127e-02 -4.70867276e-01 -5.46890616e-01 6.96495891e-01
-1.93199944e+00 -7.46078312e-01 -9.62764204e-01 5.05611539e-01
4.61341351e-01 1.10394724e-01 2.82883614e-01 6.96122706e-01
-4.79266346e-01 5.48933148e-01 6.26198342e-03 8.75168890e-02
9.90194619e-01 -7.77501702e-01 9.02814448e-01 9.63386059e-01
-1.68844298e-01 4.01468784e-01 4.75263476e-01 -3.90160739e-01
-1.34238446e+00 -9.39904928e-01 6.76365793e-01 -1.88953206e-01
3.76486987e-01 -4.32600409e-01 -8.20138097e-01 5.46313941e-01
-2.21119169e-02 1.01024069e-01 4.84275334e-02 2.37393361e-02
-3.88722211e-01 -6.55205905e-01 -1.00540316e+00 9.00396466e-01
1.47681904e+00 -4.33379084e-01 -3.03941637e-01 2.18581241e-02
5.93985856e-01 -8.76427710e-01 -5.79214275e-01 7.08063126e-01
6.42781317e-01 -1.78211749e+00 1.18804419e+00 2.16683000e-01
7.18766749e-01 -4.55527991e-01 -7.80369788e-02 -1.07812619e+00
-1.99661747e-01 -7.03095794e-01 3.26837063e-01 1.11374497e+00
3.41988094e-02 -8.04802895e-01 9.84930158e-01 6.85378432e-01
-7.34041482e-02 -6.89402640e-01 -8.89967382e-01 -6.56535745e-01
-3.37215543e-01 -2.65788794e-01 8.25407028e-01 6.99575067e-01
-5.40400624e-01 3.17895263e-01 -5.68524539e-01 1.82430357e-01
9.17222321e-01 2.52408475e-01 1.13946164e+00 -1.07902813e+00
-4.03871834e-01 -4.76211548e-01 -6.71791852e-01 -1.85986555e+00
-2.82148048e-02 -5.85409105e-02 9.75554809e-02 -1.30360842e+00
2.23852798e-01 -5.23018122e-01 1.11647025e-01 -2.83086803e-02
-2.96184450e-01 6.91597700e-01 2.58518159e-01 1.65240005e-01
-5.25241315e-01 6.51122391e-01 1.31888902e+00 -7.43068114e-04
-7.69000798e-02 -6.43823743e-02 -7.09002793e-01 7.06008434e-01
4.75927711e-01 -5.30140437e-02 -3.63439769e-01 -9.42162991e-01
3.67016733e-01 1.49812713e-01 5.99345148e-01 -1.43289256e+00
3.33998680e-01 1.68789580e-01 4.69901174e-01 -5.41242778e-01
5.41731179e-01 -7.80619323e-01 2.82304823e-01 2.19367966e-01
-1.10585548e-01 1.04560845e-01 1.33883834e-01 4.94403034e-01
-4.06338125e-01 2.26846114e-01 8.97658169e-01 8.24163258e-02
-9.01980162e-01 5.60666978e-01 2.92763561e-01 2.66825795e-01
6.82840765e-01 -4.41239655e-01 -4.55822170e-01 -6.19099915e-01
4.12555337e-02 3.34777713e-01 1.05675507e+00 4.44178075e-01
4.60947007e-01 -1.08465505e+00 -5.85185409e-01 5.08937061e-01
1.30103886e-01 5.52267373e-01 5.39840460e-01 6.85355365e-01
-6.98976576e-01 2.83040047e-01 -3.06334823e-01 -7.07073629e-01
-1.00497842e+00 9.85052586e-02 3.74465406e-01 -1.40392810e-01
-9.38985527e-01 7.37001896e-01 5.40762603e-01 -1.98779017e-01
4.48091030e-01 -5.39217710e-01 1.64409027e-01 -2.70901978e-01
4.60600764e-01 2.47372448e-01 2.40132511e-01 -4.99265015e-01
-1.68428466e-01 1.00728595e+00 -1.14558943e-01 2.36183539e-01
1.28736460e+00 -2.87763596e-01 2.17504546e-01 -4.22006734e-02
1.09288597e+00 -1.13670655e-01 -1.92285132e+00 -4.18794364e-01
-4.00931627e-01 -8.82252097e-01 1.02242179e-01 -4.40308750e-01
-1.36086202e+00 8.55842412e-01 5.91953635e-01 -2.90985435e-01
1.29816580e+00 -2.59959966e-01 1.14883745e+00 3.31034154e-01
4.97956812e-01 -8.93218040e-01 -6.10981137e-04 4.90891665e-01
6.36065841e-01 -1.53895867e+00 1.94825917e-01 -4.31837171e-01
-3.54310006e-01 1.05266082e+00 7.25704312e-01 -1.20198287e-01
4.21088547e-01 2.58284271e-01 1.85711011e-01 4.79807071e-02
-4.62783217e-01 -1.61246985e-01 2.38507390e-01 4.68569875e-01
3.12599838e-01 -3.96557570e-01 1.44820079e-01 1.40873671e-01
-3.29410434e-01 1.15415543e-01 4.15539682e-01 7.10194647e-01
-6.34097978e-02 -8.60959113e-01 -3.06123137e-01 2.74385184e-01
-1.69462785e-01 -2.65545577e-01 -4.36772108e-02 8.89468729e-01
8.31306130e-02 7.91203380e-01 3.72728288e-01 -3.36270362e-01
7.46356905e-01 -6.44955218e-01 5.05298913e-01 -2.79490769e-01
-5.27882755e-01 -1.50428444e-01 -1.02969579e-01 -9.98715997e-01
-4.26477879e-01 -5.21880031e-01 -8.25836420e-01 -5.49197137e-01
2.82583069e-02 -4.84590203e-01 4.53690797e-01 8.44422162e-01
5.74208438e-01 4.62845832e-01 5.17406404e-01 -1.32085133e+00
-3.20887685e-01 -7.82558024e-01 -3.82067531e-01 5.39839625e-01
6.03100419e-01 -5.62231362e-01 -5.18146276e-01 -1.23029120e-01]
|
[8.826929092407227, -2.2811875343322754]
|
32b21582-2e3f-4e5d-a2e4-b9866465d6e4
|
neural-voting-field-for-camera-space-3d-hand
|
2305.04328
| null |
https://arxiv.org/abs/2305.04328v1
|
https://arxiv.org/pdf/2305.04328v1.pdf
|
Neural Voting Field for Camera-Space 3D Hand Pose Estimation
|
We present a unified framework for camera-space 3D hand pose estimation from a single RGB image based on 3D implicit representation. As opposed to recent works, most of which first adopt holistic or pixel-level dense regression to obtain relative 3D hand pose and then follow with complex second-stage operations for 3D global root or scale recovery, we propose a novel unified 3D dense regression scheme to estimate camera-space 3D hand pose via dense 3D point-wise voting in camera frustum. Through direct dense modeling in 3D domain inspired by Pixel-aligned Implicit Functions for 3D detailed reconstruction, our proposed Neural Voting Field (NVF) fully models 3D dense local evidence and hand global geometry, helping to alleviate common 2D-to-3D ambiguities. Specifically, for a 3D query point in camera frustum and its pixel-aligned image feature, NVF, represented by a Multi-Layer Perceptron, regresses: (i) its signed distance to the hand surface; (ii) a set of 4D offset vectors (1D voting weight and 3D directional vector to each hand joint). Following a vote-casting scheme, 4D offset vectors from near-surface points are selected to calculate the 3D hand joint coordinates by a weighted average. Experiments demonstrate that NVF outperforms existing state-of-the-art algorithms on FreiHAND dataset for camera-space 3D hand pose estimation. We also adapt NVF to the classic task of root-relative 3D hand pose estimation, for which NVF also obtains state-of-the-art results on HO3D dataset.
|
['Zicheng Liu', 'Junsong Yuan', 'Lijuan Wang', 'Lin Liang', 'Kevin Lin', 'Chung-Ching Lin', 'Lin Huang']
|
2023-05-07
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Huang_Neural_Voting_Field_for_Camera-Space_3D_Hand_Pose_Estimation_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Huang_Neural_Voting_Field_for_Camera-Space_3D_Hand_Pose_Estimation_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['3d-hand-pose-estimation', 'hand-pose-estimation', '3d-hand-pose-estimation']
|
['computer-vision', 'computer-vision', 'graphs']
|
[-1.48543924e-01 -5.51818490e-01 -4.11354899e-01 -1.56942278e-01
-1.11292315e+00 -4.46480781e-01 1.94421366e-01 -4.55919415e-01
-6.06670320e-01 2.80612230e-01 4.69057828e-01 9.53218266e-02
-2.44413689e-03 -3.40972364e-01 -6.50311887e-01 -6.61716878e-01
3.30229312e-01 9.78110254e-01 -7.19811916e-02 -2.07394548e-02
4.19928938e-01 1.16254282e+00 -1.28261626e+00 -1.23618647e-01
1.45112500e-01 9.89863873e-01 1.30919769e-01 8.19731116e-01
2.33524859e-01 6.20320022e-01 -4.17470843e-01 -6.01778217e-02
5.51686168e-01 2.50081494e-02 -8.91819119e-01 3.03891838e-01
1.05149031e+00 -9.59318340e-01 -5.23350418e-01 6.77840412e-01
1.03390682e+00 -2.13345774e-02 8.75597358e-01 -9.71349776e-01
-3.56400877e-01 -1.38235807e-01 -1.05693114e+00 -4.20965612e-01
8.07348192e-01 3.81583065e-01 9.47277606e-01 -1.38687527e+00
1.10104489e+00 1.46052516e+00 7.78900981e-01 3.73264998e-01
-1.16722381e+00 -4.52786416e-01 2.69711137e-01 6.49095476e-02
-1.79147875e+00 6.84082806e-02 1.20680118e+00 -5.75538337e-01
1.46593165e+00 1.47145897e-01 9.93573606e-01 1.16227853e+00
4.57617044e-02 9.70535755e-01 1.21514189e+00 -6.26423478e-01
-8.59131142e-02 -4.99465346e-01 6.90238401e-02 1.01893902e+00
-9.03417841e-02 1.20902043e-02 -7.16184258e-01 -2.08367497e-01
1.53945470e+00 3.34570587e-01 -3.56862605e-01 -7.92724073e-01
-1.28497970e+00 5.38542807e-01 8.83458614e-01 -5.21909371e-02
-8.54978979e-01 3.65439355e-01 -1.57212287e-01 -2.05449745e-01
2.41874531e-01 -1.23313397e-01 -7.44536042e-01 -5.66976741e-02
-8.18932772e-01 7.46428192e-01 6.20537639e-01 8.88056219e-01
7.92851448e-01 -2.41468728e-01 -2.21854150e-01 5.24730086e-01
8.45855236e-01 9.11823809e-01 -1.59928858e-01 -1.16656995e+00
6.63366914e-01 6.98709786e-01 1.21768139e-01 -7.87800074e-01
-6.20798409e-01 -3.29465657e-01 -6.79772556e-01 5.73141038e-01
6.18389547e-01 2.01142505e-01 -1.10133624e+00 1.40820098e+00
6.03905916e-01 -3.75672907e-01 -7.20491767e-01 1.54505277e+00
7.30988443e-01 1.46674812e-01 -2.11956903e-01 1.00006501e-03
1.25222635e+00 -9.10421371e-01 -1.78294018e-01 -1.39033303e-01
1.25816181e-01 -8.35162997e-01 9.59933162e-01 5.18339396e-01
-1.06095731e+00 -4.16230172e-01 -5.14885724e-01 -4.93589073e-01
-3.39372158e-02 2.72872895e-01 4.81981903e-01 4.37860817e-01
-9.69653606e-01 3.02831292e-01 -8.26632142e-01 -2.23162815e-01
4.30008173e-01 4.20494646e-01 -7.68044233e-01 -8.90848860e-02
-4.99890059e-01 1.04634726e+00 -2.05538884e-01 4.64133739e-01
-7.14421809e-01 -5.39711893e-01 -8.26067328e-01 -5.08004606e-01
4.78455693e-01 -1.08027375e+00 1.11089003e+00 -1.97243124e-01
-1.65893483e+00 1.30393195e+00 -4.90314990e-01 3.34893912e-01
7.28390694e-01 -4.77246821e-01 5.42237759e-01 1.47208199e-01
4.65645604e-02 6.92136049e-01 1.16181064e+00 -1.41103196e+00
-1.90213099e-01 -1.25145471e+00 -5.87419160e-02 4.50517535e-01
2.28143737e-01 -9.64126140e-02 -8.31093252e-01 -8.03534448e-01
6.99530780e-01 -9.57508087e-01 -2.72921354e-01 4.39167112e-01
-6.13375902e-01 -4.52190131e-01 5.39290130e-01 -1.04610538e+00
7.91137159e-01 -1.51334321e+00 8.15832257e-01 4.74697083e-01
3.70304823e-01 -1.37913719e-01 -9.26365554e-02 -1.56879455e-01
1.72409892e-01 -3.92883122e-01 -2.87377298e-01 -7.69986749e-01
-2.31102761e-02 -3.12294718e-02 6.04113974e-02 8.72823834e-01
9.88148898e-02 9.95529592e-01 -7.43481159e-01 -5.50215900e-01
5.92800736e-01 1.21494460e+00 -6.35635853e-01 3.47007900e-01
-5.46120405e-02 5.49859047e-01 -6.27705872e-01 1.23112607e+00
7.31788099e-01 -2.29107827e-01 -3.81655134e-02 -5.94209671e-01
-1.36768013e-01 6.53963983e-02 -1.38209522e+00 2.40778661e+00
-4.59514618e-01 4.95000221e-02 5.69915324e-02 -3.28662485e-01
7.43208230e-01 2.58691281e-01 4.77788568e-01 -2.74485171e-01
4.28766072e-01 2.14419603e-01 -6.86211765e-01 -1.68010384e-01
2.85716295e-01 1.02708535e-02 3.75885936e-03 4.54009086e-01
1.99303448e-01 -5.66549361e-01 -8.45850885e-01 -2.04143405e-01
9.35651064e-01 8.52877021e-01 3.40545595e-01 3.00676674e-01
4.47156489e-01 -2.93908790e-02 1.25272065e-01 5.17728925e-01
-2.96820849e-01 1.16494298e+00 2.83516467e-01 -5.07741749e-01
-1.06658435e+00 -1.18282413e+00 -2.80703213e-02 7.10301578e-01
-6.60418496e-02 -1.38582870e-01 -8.47764194e-01 -7.82375693e-01
5.30645728e-01 -1.91295717e-03 -6.82028174e-01 5.98939717e-01
-1.01256073e+00 -1.45520791e-01 7.31884986e-02 8.91242325e-01
3.93143743e-01 -9.79018569e-01 -8.41249466e-01 -1.25677194e-02
-1.13382921e-01 -7.62637258e-01 -9.38797474e-01 2.08752543e-01
-9.15107965e-01 -1.21624577e+00 -1.34703112e+00 -7.27998137e-01
6.98300660e-01 4.41185385e-01 7.11835861e-01 -2.88991004e-01
-6.24932706e-01 7.80823112e-01 -1.90135539e-01 1.51889920e-01
3.02230686e-01 1.96112186e-01 2.51794100e-01 -2.91548938e-01
2.40787953e-01 -5.93683660e-01 -9.49587524e-01 3.61805558e-01
8.30861777e-02 -2.55492240e-01 5.31917155e-01 7.43153512e-01
1.09774208e+00 -7.48665571e-01 -1.29935965e-01 -7.09385797e-02
1.67375624e-01 2.68043458e-01 -5.00237644e-01 3.06373119e-01
-3.01534265e-01 1.04436100e-01 -6.59819692e-02 -4.51953560e-01
-1.00943696e+00 6.98389530e-01 -3.01928520e-01 -1.00526428e+00
-1.19416185e-01 -2.47310605e-02 -1.44058824e-01 -2.96261311e-01
7.16631711e-01 7.99327046e-02 1.37471452e-01 -7.35267103e-01
6.73553169e-01 8.26405585e-01 6.64917171e-01 -4.97204572e-01
6.64162934e-01 6.87596083e-01 1.41861185e-01 -5.94223440e-01
-6.02507234e-01 -5.26504099e-01 -1.38564348e+00 -4.53269839e-01
1.12018955e+00 -9.53083932e-01 -1.20907557e+00 9.74311709e-01
-1.76119983e+00 -2.63035715e-01 -1.10097915e-01 6.64927602e-01
-7.90168643e-01 4.37496871e-01 -6.00597143e-01 -1.12346983e+00
-6.79146588e-01 -1.36639059e+00 1.94591033e+00 -2.01867297e-01
-4.94972616e-01 -4.69915897e-01 1.85752451e-01 5.83487928e-01
-2.93078572e-01 3.63496393e-01 5.49576521e-01 3.83578748e-01
-7.23667324e-01 -5.42868793e-01 -1.49041548e-01 2.21214637e-01
5.81046194e-03 -2.60438204e-01 -1.20936430e+00 -3.45726967e-01
-6.55119866e-02 -3.44176203e-01 6.25079513e-01 8.79614651e-01
8.55635643e-01 7.66890049e-02 -1.88114837e-01 8.48967016e-01
1.09256887e+00 -4.75724638e-01 2.42005840e-01 3.27099442e-01
1.28926778e+00 5.28326094e-01 4.30988491e-01 6.90639257e-01
6.54983222e-01 1.03886175e+00 6.00930512e-01 1.11892983e-01
-4.57788318e-01 -4.41450685e-01 1.47100091e-01 5.42542398e-01
-1.09562957e+00 4.49057221e-01 -5.67526758e-01 8.04053620e-02
-1.64558244e+00 -4.13784862e-01 -1.83700211e-02 2.31678677e+00
6.89771235e-01 -2.12612510e-01 5.63460708e-01 2.77469039e-01
4.56931949e-01 4.09804612e-01 -8.22944343e-01 2.63061434e-01
-1.55327573e-01 5.51796556e-01 5.49539924e-01 6.90513611e-01
-9.33173478e-01 1.20909917e+00 5.47312927e+00 5.36325097e-01
-9.51120675e-01 3.90325129e-01 -6.97698519e-02 -4.35732633e-01
-2.24904329e-01 -1.10811070e-01 -9.45507526e-01 -3.11750442e-01
-3.84520143e-01 9.90595818e-01 7.41665959e-01 8.14922631e-01
-2.30264943e-02 3.70242707e-02 -1.03845596e+00 1.50170124e+00
3.92137587e-01 -9.98926640e-01 -8.27383325e-02 3.06850523e-01
6.07334197e-01 3.68111990e-02 -7.62163401e-02 -2.41072848e-01
-1.56881437e-01 -8.89993012e-01 1.18716097e+00 6.38524652e-01
1.03093147e+00 -4.74536151e-01 2.69783378e-01 2.74325162e-01
-1.31671929e+00 1.01623401e-01 3.54165807e-02 -3.73970494e-02
2.79386103e-01 2.13522837e-01 -6.24694943e-01 2.04021513e-01
9.85101938e-01 7.26105571e-01 -1.70538098e-01 4.79215145e-01
-7.28111446e-01 -1.49678573e-01 -4.91356969e-01 1.08110845e-01
2.35323496e-02 1.32184982e-01 7.36930490e-01 8.57476652e-01
5.45054302e-02 1.18082084e-01 -5.51526695e-02 9.22607064e-01
9.80251655e-02 -1.26308873e-01 -3.82470489e-01 6.26336992e-01
4.38873142e-01 1.16142941e+00 -5.82219481e-01 7.49758929e-02
-1.16373271e-01 1.48701239e+00 4.86364871e-01 5.00349879e-01
-3.68210793e-01 -1.25884637e-01 7.25630045e-01 1.89661488e-01
4.36263859e-01 -7.35503137e-01 -4.91181612e-01 -1.12408066e+00
4.78271842e-01 -5.56162059e-01 2.37627000e-01 -1.09329152e+00
-1.22017860e+00 4.82281029e-01 -1.76888824e-01 -1.06899011e+00
-3.63301605e-01 -1.08750880e+00 -1.07863382e-01 1.26235318e+00
-1.29463673e+00 -1.58266640e+00 -5.05507469e-01 1.11815894e+00
5.61903000e-01 -7.13505000e-02 9.23264623e-01 -2.11303711e-01
-1.01262614e-01 5.66146374e-01 -5.26940703e-01 2.25710854e-01
6.91019893e-01 -1.16927576e+00 5.22306681e-01 2.26099372e-01
3.09603661e-01 6.78035140e-01 1.46249726e-01 -7.40725875e-01
-2.22925758e+00 -4.80608404e-01 7.74706662e-01 -1.26382709e+00
1.24003366e-01 -3.32609892e-01 -4.19019908e-01 7.82674730e-01
-4.53828543e-01 3.38647693e-01 4.63320203e-02 3.02496135e-01
-8.16480100e-01 -9.32166073e-03 -1.38858569e+00 3.21867853e-01
1.64932978e+00 -9.25889134e-01 -6.19381905e-01 2.52389133e-01
2.56533086e-01 -8.42949212e-01 -1.03741765e+00 2.75228918e-01
1.36968243e+00 -7.30227292e-01 1.40398943e+00 -5.01284659e-01
1.05218090e-01 -4.03215200e-01 -5.41243255e-01 -8.29578757e-01
-3.03278208e-01 -4.52161640e-01 -5.07643163e-01 5.58553159e-01
-2.24910289e-01 -1.93090260e-01 1.21768451e+00 3.71593356e-01
2.83612818e-01 -8.47131312e-01 -1.33160102e+00 -3.30405742e-01
-6.15937412e-02 -7.94167578e-01 3.69675696e-01 3.43055069e-01
-3.54898214e-01 1.58734098e-01 -3.82128358e-01 3.46956551e-01
1.19414568e+00 2.09572181e-01 9.97587919e-01 -1.22637486e+00
-1.22658394e-01 -6.33243024e-01 -5.65771997e-01 -1.59099150e+00
2.50201583e-01 -8.14129591e-01 8.93869624e-02 -1.37621272e+00
2.98614860e-01 -2.98331261e-01 -9.00137238e-03 4.95016754e-01
2.61782091e-02 5.27526915e-01 3.64933312e-01 4.61491942e-01
-2.36270651e-02 3.17591608e-01 1.60265100e+00 -1.85887814e-01
-3.48792404e-01 -1.11057565e-01 -1.95766985e-01 7.48174250e-01
9.41415727e-02 -2.11527094e-01 2.61435211e-01 -6.40904963e-01
-2.16116980e-02 3.15670073e-01 9.59897697e-01 -4.99486476e-01
1.82444334e-01 -1.70780376e-01 7.50674069e-01 -1.20211470e+00
6.53025568e-01 -9.54535425e-01 -5.98982610e-02 4.40519631e-01
4.14057486e-02 -1.79852396e-01 -3.75911534e-01 2.79289007e-01
2.97890872e-01 1.04856186e-01 4.85998392e-01 -2.38984525e-01
-6.11682355e-01 6.79009438e-01 6.32105395e-02 -3.88387829e-01
6.51865184e-01 -5.45395195e-01 2.26428330e-01 -2.41592661e-01
-6.75877869e-01 -7.95404837e-02 5.31377912e-01 4.40566599e-01
1.08029079e+00 -1.39918482e+00 -7.44444489e-01 3.95773619e-01
-3.94579992e-02 5.89146078e-01 3.89969140e-01 8.93564522e-01
-6.16658628e-01 4.45928901e-01 -1.35456383e-01 -1.13837838e+00
-1.36264348e+00 3.05287212e-01 4.70840454e-01 6.59142584e-02
-7.85576522e-01 1.30703843e+00 -1.59170821e-01 -9.63795900e-01
5.83367407e-01 -5.50980568e-01 4.05242920e-01 3.27790380e-02
3.61012548e-01 6.19495511e-01 1.74452245e-01 -1.03525710e+00
-7.58530259e-01 1.69565821e+00 1.94575086e-01 -2.44094118e-01
1.36757898e+00 2.12553784e-01 -6.73727468e-02 1.43817261e-01
1.42398167e+00 -5.07373288e-02 -1.75748217e+00 -3.41737539e-01
-7.13902950e-01 -7.36450315e-01 2.53446788e-01 -9.26684141e-01
-1.20776367e+00 1.16298807e+00 8.82973790e-01 -7.49837816e-01
8.65040481e-01 5.35663486e-01 6.45640075e-01 4.40477341e-01
8.11586857e-01 -8.55147421e-01 6.23637363e-02 4.36636150e-01
1.46690071e+00 -1.20238090e+00 3.41976553e-01 -3.74481142e-01
-4.92767751e-01 1.25166118e+00 4.02720779e-01 -2.11444020e-01
7.08162665e-01 7.27929324e-02 4.61543985e-02 -2.70569831e-01
2.06271291e-01 -1.96483120e-01 5.57868659e-01 8.09044659e-01
2.93620944e-01 2.65089661e-01 2.60685086e-01 3.55504453e-01
-8.35226849e-02 1.16729848e-01 -4.48155910e-01 1.13219643e+00
-8.96402448e-02 -1.03431058e+00 -9.11587298e-01 1.90048553e-02
2.78842330e-01 2.60530025e-01 -4.25099611e-01 8.34807813e-01
1.98296607e-01 4.42662507e-01 7.26296529e-02 -6.30232811e-01
7.79482484e-01 -7.75053203e-02 1.58545578e+00 -4.23836201e-01
-6.16366804e-01 4.40853089e-01 -4.30985153e-01 -7.59826064e-01
-5.28029084e-01 -7.64051735e-01 -1.18222821e+00 -2.73374319e-01
-2.87771791e-01 -7.63350487e-01 8.95046294e-01 9.87124324e-01
4.84267324e-02 -9.24973283e-03 5.52855253e-01 -1.96910870e+00
-7.90825009e-01 -1.04457152e+00 -9.06217456e-01 9.66623873e-02
4.92030501e-01 -1.20484114e+00 -1.81299269e-01 -2.70451069e-01]
|
[6.56567907333374, -0.8334240913391113]
|
b843c5b6-1a02-44a3-948c-37a07a4fb123
|
one-step-knowledge-distillation-and-fine
|
2305.17394
| null |
https://arxiv.org/abs/2305.17394v2
|
https://arxiv.org/pdf/2305.17394v2.pdf
|
One-Step Knowledge Distillation and Fine-Tuning in Using Large Pre-Trained Self-Supervised Learning Models for Speaker Verification
|
The application of speech self-supervised learning (SSL) models has achieved remarkable performance in speaker verification (SV). However, there is a computational cost hurdle in employing them, which makes development and deployment difficult. Several studies have simply compressed SSL models through knowledge distillation (KD) without considering the target task. Consequently, these methods could not extract SV-tailored features. This paper suggests One-Step Knowledge Distillation and Fine-Tuning (OS-KDFT), which incorporates KD and fine-tuning (FT). We optimize a student model for SV during KD training to avert the distillation of inappropriate information for the SV. OS-KDFT could downsize Wav2Vec 2.0 based ECAPA-TDNN size by approximately 76.2%, and reduce the SSL model's inference time by 79% while presenting an EER of 0.98%. The proposed OS-KDFT is validated across VoxCeleb1 and VoxCeleb2 datasets and W2V2 and HuBERT SSL models. Experiments are available on our GitHub.
|
['Ha-Jin Yu', 'Hyun-seo Shin', 'Ju-ho Kim', 'Chan-yeong Lim', 'Jungwoo Heo']
|
2023-05-27
| null | null | null | null |
['speaker-verification']
|
['speech']
|
[-1.04539551e-01 2.09234402e-01 -1.62267819e-01 -7.18043327e-01
-9.78107214e-01 -4.16435301e-01 3.80466491e-01 -6.07894287e-02
-5.53418934e-01 6.72831416e-01 3.42533708e-01 -5.73320627e-01
4.79656868e-02 -3.34875554e-01 -4.69098210e-01 -4.99356061e-01
2.79895157e-01 9.31282714e-02 2.76993960e-02 4.92620375e-03
-1.56121612e-01 3.22764426e-01 -1.30966163e+00 9.39714685e-02
9.23869610e-01 1.14763510e+00 2.34822586e-01 8.08839381e-01
-2.04876140e-01 6.15608752e-01 -7.64867544e-01 -7.99975991e-01
1.76429793e-01 -3.36440951e-01 -6.61188602e-01 -4.42318171e-01
4.62820232e-01 -2.18820870e-01 -5.32461345e-01 1.01230860e+00
9.90840971e-01 3.68150204e-01 4.21415418e-01 -1.26811540e+00
-6.23583257e-01 1.22213638e+00 -2.19928071e-01 2.29931146e-01
-7.29108080e-02 1.01386055e-01 9.37131405e-01 -1.09079504e+00
2.30814978e-01 1.33128691e+00 8.60336661e-01 8.82091165e-01
-9.11471725e-01 -1.15108299e+00 -9.70437080e-02 5.25502503e-01
-1.79953837e+00 -1.08212888e+00 8.38627219e-01 -2.26287246e-01
1.14050305e+00 4.04773116e-01 4.69834924e-01 1.07189798e+00
-5.86092651e-01 1.04729986e+00 9.41280901e-01 -5.52280188e-01
2.19773918e-01 6.83514774e-01 2.71609783e-01 6.33587897e-01
1.61083806e-02 2.64945924e-01 -9.35094714e-01 -1.20672293e-01
2.23507598e-01 -5.53819478e-01 -2.96229988e-01 2.04578906e-01
-7.52273083e-01 8.86712432e-01 5.48850410e-02 9.08135846e-02
-2.82091927e-02 -2.62620628e-01 6.89566612e-01 2.97876179e-01
3.96847099e-01 1.88017726e-01 -7.05882311e-01 -4.09291327e-01
-1.23853803e+00 -4.61994037e-02 7.97948658e-01 1.02512264e+00
5.30288875e-01 7.01830268e-01 -3.28910023e-01 1.29191923e+00
4.08281386e-01 9.32069480e-01 9.54407394e-01 -5.20725429e-01
4.71138328e-01 9.37654153e-02 -5.12789190e-01 -6.29144847e-01
-7.72432014e-02 -6.76908374e-01 -8.19930613e-01 -3.64147305e-01
-1.35789692e-01 -3.78635883e-01 -9.21690047e-01 1.70972049e+00
2.91137308e-01 3.89471918e-01 3.68646383e-01 6.85375452e-01
1.30007350e+00 7.69133329e-01 1.99260250e-01 -3.37118775e-01
1.01907921e+00 -1.10632384e+00 -9.90811646e-01 -2.06053182e-01
6.29546702e-01 -7.38593817e-01 1.01827979e+00 3.81123483e-01
-9.52071488e-01 -6.73798084e-01 -1.08957052e+00 9.34381038e-03
-3.54772687e-01 4.33354765e-01 3.91792297e-01 1.05229378e+00
-9.79786813e-01 3.56318086e-01 -5.34636199e-01 -1.91153541e-01
6.92640483e-01 2.45373324e-01 -1.35543466e-01 3.14766407e-01
-1.43297374e+00 8.27502847e-01 3.85254502e-01 2.64790744e-01
-9.94444668e-01 -1.01700044e+00 -8.89880955e-01 1.57730520e-01
1.42397121e-01 -2.52015829e-01 1.41706049e+00 -5.02475977e-01
-1.93348563e+00 6.00904882e-01 -3.08259398e-01 -7.86822796e-01
5.12877524e-01 -6.29245490e-02 -9.73096907e-01 -3.01330179e-01
-2.53025323e-01 4.63004470e-01 1.11288619e+00 -8.13241661e-01
-5.35590649e-01 -2.10202500e-01 -4.79398370e-01 1.76617086e-01
-7.42294312e-01 -1.88269746e-03 -4.70858335e-01 -7.61717737e-01
-1.27562940e-01 -5.87084115e-01 8.72511268e-02 -3.80327463e-01
-5.88997364e-01 -3.39221507e-01 9.55364704e-01 -1.04740453e+00
1.60648263e+00 -2.35078073e+00 -2.87022114e-01 2.94192255e-01
8.85060132e-02 1.14569712e+00 -1.52498826e-01 6.27134740e-02
6.28007129e-02 -1.57572422e-02 -1.19163848e-01 -7.39699960e-01
1.62247404e-01 1.84625626e-01 -4.51294333e-01 3.47757638e-01
-5.49946204e-02 8.72539043e-01 -5.34822583e-01 -6.85689688e-01
1.84202880e-01 7.20655084e-01 -5.73098898e-01 2.82390207e-01
5.76748960e-02 -4.34640124e-02 -1.20393015e-01 5.56668878e-01
9.64445829e-01 2.43620113e-01 -2.52009481e-02 -4.02777076e-01
1.98297258e-02 6.22794867e-01 -1.17499459e+00 1.50051236e+00
-6.24731123e-01 8.86718392e-01 -1.02795891e-01 -7.84650624e-01
1.17532098e+00 3.40430588e-01 -1.32121265e-01 -5.94869137e-01
2.44626224e-01 1.17132656e-01 -2.54275739e-01 -5.16608477e-01
3.42368186e-01 -2.93319762e-01 1.88065276e-01 1.49355188e-01
4.07002956e-01 -1.92105114e-01 -3.74916285e-01 5.78901134e-02
6.92172527e-01 -3.26851100e-01 1.78295270e-01 -1.12589141e-02
8.17671776e-01 -3.84096771e-01 8.64903510e-01 6.69191480e-01
-5.21829188e-01 2.68560231e-01 -8.25855359e-02 7.08855093e-02
-6.73250377e-01 -1.14386487e+00 -2.43284062e-01 9.29360628e-01
-3.45067799e-01 -4.75567192e-01 -8.21195006e-01 -7.13770211e-01
1.45480007e-01 1.08806753e+00 -2.99502343e-01 -3.67823690e-01
-3.56004626e-01 -4.31625575e-01 1.28282666e+00 3.59868109e-01
7.80613005e-01 -7.46993244e-01 -5.50997593e-02 1.01816125e-01
-1.93587661e-01 -1.29669285e+00 -7.12326288e-01 1.18070155e-01
-5.89178145e-01 -5.22062242e-01 -6.20220661e-01 -8.62243414e-01
3.50929797e-01 8.38974640e-02 5.60284257e-01 -2.77528912e-01
-1.38333172e-01 1.03836611e-01 -3.24753463e-01 -6.44192457e-01
-4.95761096e-01 1.38861284e-01 4.73051131e-01 1.40515953e-01
7.46644616e-01 -2.14284480e-01 -2.37199560e-01 2.40997300e-01
-3.78663182e-01 -1.73824757e-01 3.38887990e-01 1.08697534e+00
4.86250967e-01 1.68335289e-01 8.60370219e-01 -6.50108397e-01
7.23313153e-01 -8.52352977e-02 -6.82726920e-01 4.14249361e-01
-9.32325244e-01 7.49216750e-02 5.05241871e-01 -5.58600843e-01
-1.34208786e+00 -9.19491276e-02 -5.24934113e-01 -7.75924087e-01
3.79131287e-02 3.38069111e-01 -3.38396639e-01 -1.04005896e-01
5.59205651e-01 3.76114428e-01 -6.97420612e-02 -7.37245917e-01
4.19594198e-01 1.28430831e+00 4.71082151e-01 -1.72601983e-01
8.29674482e-01 -9.82204899e-02 -6.67177320e-01 -1.24411464e+00
-8.41822565e-01 -4.29197848e-01 -1.72118172e-01 1.44487098e-02
5.59173822e-01 -1.11437666e+00 -7.44748592e-01 4.81102616e-01
-9.59589124e-01 -2.45525628e-01 -5.14746308e-01 8.18392158e-01
-1.15676329e-01 3.43892395e-01 -4.10880446e-01 -1.04106092e+00
-7.50769019e-01 -9.21223044e-01 5.08764386e-01 2.07135692e-01
-2.37284288e-01 -7.95492768e-01 7.04051554e-02 6.34777665e-01
8.10187399e-01 -6.73209727e-01 5.24891973e-01 -1.01118696e+00
-8.44039209e-03 -1.14726968e-01 -4.92135771e-02 9.88036335e-01
-1.21561522e-02 -1.57625869e-01 -1.66188371e+00 -3.09964269e-01
-3.59096043e-02 -2.42169604e-01 8.91361594e-01 3.60355377e-01
1.22808146e+00 -5.36748707e-01 -4.57190424e-02 9.48385656e-01
9.73322809e-01 2.86900938e-01 1.99146241e-01 1.78473573e-02
5.82920671e-01 3.53947729e-01 5.09983480e-01 4.58445370e-01
3.95112276e-01 7.18852282e-01 -2.15630114e-01 1.36435088e-02
-6.25575125e-01 -7.15141773e-01 4.76621270e-01 1.49107134e+00
2.09110022e-01 -4.11242954e-02 -8.25782239e-01 6.68802202e-01
-1.40988934e+00 -9.21072066e-01 1.77527174e-01 2.10306740e+00
1.06545162e+00 4.42713052e-02 -4.20650281e-02 3.67794126e-01
6.76378906e-01 1.76586986e-01 -6.81300759e-01 -6.20604455e-01
-3.34057659e-01 1.94680631e-01 5.12492299e-01 8.31893265e-01
-9.41294134e-01 1.18491244e+00 5.30134487e+00 1.32501364e+00
-1.35160422e+00 5.24209023e-01 5.25828898e-01 -1.80505082e-01
-2.43744850e-01 -2.63100803e-01 -1.35696387e+00 4.58589256e-01
1.26246142e+00 -4.29898471e-01 4.39045072e-01 1.03683603e+00
2.31703937e-01 3.93855393e-01 -7.66216636e-01 1.57189095e+00
3.41390342e-01 -1.17189991e+00 -4.63430993e-02 -3.34428728e-01
4.46765035e-01 1.16115935e-01 2.13691294e-01 8.19355726e-01
1.56421706e-01 -7.51109302e-01 6.81744516e-01 1.02370963e-01
1.07502306e+00 -7.06004560e-01 7.56129324e-01 1.83838323e-01
-1.02744186e+00 -1.82469234e-01 -5.28010249e-01 4.74585086e-01
2.01799497e-01 6.06298327e-01 -1.52193701e+00 2.52102226e-01
8.75792146e-01 2.92015165e-01 -4.35206622e-01 9.05859888e-01
-3.42797071e-01 1.14021218e+00 -3.49285007e-01 -2.08335906e-01
-9.25773457e-02 2.48279870e-01 4.84748781e-01 1.61862481e+00
2.53917366e-01 -7.91050866e-02 -5.08800566e-01 5.70058703e-01
-1.99516490e-01 2.97996491e-01 -2.54881740e-01 -2.15039581e-01
9.44977582e-01 8.15436661e-01 1.75034061e-01 -5.34287155e-01
-1.03344575e-01 1.00560009e+00 2.80505329e-01 3.70349973e-01
-9.58082080e-01 -6.27534926e-01 9.35679793e-01 -1.63955420e-01
5.21222293e-01 7.11824298e-02 -2.55199403e-01 -1.18955004e+00
-1.37958512e-01 -8.72470796e-01 2.84823358e-01 -5.00004411e-01
-1.17823434e+00 9.89379764e-01 -1.77070647e-01 -1.11134517e+00
-1.63272619e-01 -1.69206589e-01 -2.67381549e-01 9.60285783e-01
-1.80627108e+00 -1.09332418e+00 -9.66231674e-02 7.42068470e-01
7.01779723e-01 -6.64612770e-01 9.00433719e-01 6.29149914e-01
-8.41305137e-01 1.47958803e+00 1.59453914e-01 2.30480701e-01
7.86646783e-01 -8.84795904e-01 3.80159825e-01 8.33129108e-01
1.44042969e-01 5.42353928e-01 5.69976151e-01 -4.57818478e-01
-1.24482548e+00 -1.22369802e+00 1.39816535e+00 -3.92210893e-02
4.45904016e-01 -5.04878938e-01 -9.31936324e-01 4.90192354e-01
4.54554930e-02 4.19475250e-02 8.57017040e-01 6.31079450e-02
-7.05574274e-01 -5.42334378e-01 -1.30042148e+00 4.06347185e-01
8.53639424e-01 -9.61127877e-01 -6.80032551e-01 1.88787133e-02
9.95718420e-01 -2.30109021e-01 -8.59226704e-01 3.29857200e-01
5.33937395e-01 -6.08148873e-01 1.04674017e+00 -5.52592039e-01
-3.07464004e-01 -2.00098261e-01 -3.87211859e-01 -1.48743296e+00
-1.01912953e-01 -6.30849481e-01 -4.57561046e-01 1.80231893e+00
7.12406278e-01 -7.09637702e-01 6.08435094e-01 4.81092840e-01
-2.31712341e-01 -4.23500448e-01 -1.22391474e+00 -1.04109323e+00
-1.76704779e-01 -8.47139239e-01 9.31334794e-01 1.14546156e+00
4.40014340e-02 3.38617861e-01 -4.34185892e-01 2.31536269e-01
7.17727005e-01 -3.45666677e-01 7.03811705e-01 -8.38031948e-01
-3.77392441e-01 -3.87941390e-01 -4.11394954e-01 -1.08758175e+00
2.85122722e-01 -1.02469110e+00 -2.75596268e-02 -1.03033745e+00
-1.29295200e-01 -4.79374647e-01 -4.61262345e-01 6.15569890e-01
-1.23251632e-01 -2.95838844e-02 2.18086004e-01 -8.36031586e-02
-2.58997262e-01 9.00217712e-01 7.71596730e-01 -2.88949400e-01
-3.61398041e-01 6.89006001e-02 -5.48694253e-01 4.67209846e-01
1.17791319e+00 -4.15825784e-01 -5.70933342e-01 -3.65503073e-01
-3.89274269e-01 -1.02220789e-01 -4.61817198e-02 -9.02072966e-01
5.27357101e-01 2.95035392e-01 1.94480225e-01 -6.71737909e-01
6.22918248e-01 -5.76211512e-01 -8.07479173e-02 3.03823441e-01
-4.58733082e-01 -5.24676323e-01 4.64281350e-01 3.82048488e-01
-3.85629565e-01 -2.28862360e-01 8.05351794e-01 3.23621929e-01
-6.97677255e-01 1.79845989e-01 -1.69835851e-01 7.56131709e-02
7.59172678e-01 1.40959732e-02 -2.65658051e-01 -4.55848634e-01
-6.30378783e-01 8.56820494e-02 -2.21741959e-01 4.99355644e-01
9.05064046e-01 -1.30567586e+00 -7.80932307e-01 8.84047508e-01
8.22819322e-02 -2.76535243e-01 6.90624297e-01 6.41101897e-01
-1.23460434e-01 6.49043560e-01 2.99573928e-01 -3.54728520e-01
-1.59807491e+00 2.42323756e-01 2.67282665e-01 4.96007800e-02
-4.55167532e-01 1.48655140e+00 -1.72107160e-01 -7.62887001e-01
8.08010697e-01 -1.79456845e-01 -1.58143044e-01 5.33395521e-02
6.62633955e-01 5.09247780e-01 3.33279520e-01 -6.47928298e-01
-5.03498495e-01 3.97509784e-01 -1.67111412e-01 -1.34429142e-01
1.12821269e+00 -3.38023216e-01 4.59342629e-01 2.33425602e-01
1.38286877e+00 1.13623939e-01 -8.40129495e-01 -6.13753796e-01
-1.68973520e-01 -4.20703411e-01 4.87345785e-01 -9.10970867e-01
-1.16152549e+00 1.10009313e+00 7.39733636e-01 -2.37059921e-01
1.06190193e+00 -3.91880572e-02 1.08511794e+00 2.74386317e-01
5.81081398e-02 -1.35153675e+00 -4.46827263e-01 4.21239406e-01
7.55411625e-01 -1.27078855e+00 -1.43674657e-01 -3.22333038e-01
-9.62504208e-01 7.35181570e-01 5.31707823e-01 6.16072178e-01
9.94160652e-01 3.03443164e-01 2.54887909e-01 2.06886247e-01
-7.20878124e-01 6.98164627e-02 2.68621087e-01 6.02088690e-01
1.68224737e-01 4.08271521e-01 -7.36298459e-03 1.01284695e+00
-6.64185762e-01 -2.75939759e-02 1.54391527e-01 6.38241708e-01
-3.09112936e-01 -7.91653693e-01 -2.08179712e-01 4.74374950e-01
-2.53100574e-01 -3.88510734e-01 -6.35118484e-02 3.26241821e-01
7.09280819e-02 1.21796262e+00 -2.07127601e-01 -8.10878396e-01
4.56176966e-01 2.87300169e-01 4.68857661e-02 -2.80096859e-01
-5.00726581e-01 1.01817682e-01 3.26966614e-01 -2.55747378e-01
-1.94604117e-02 -5.50092936e-01 -9.73198354e-01 -3.04123193e-01
-6.45024121e-01 4.23286229e-01 1.02210414e+00 7.24692106e-01
3.48232716e-01 5.35086572e-01 6.38562143e-01 -1.63457081e-01
-9.33181524e-01 -1.15778387e+00 -6.06532872e-01 1.65782473e-03
5.67338824e-01 -5.60378850e-01 -7.34353065e-01 1.02686293e-01]
|
[14.316469192504883, 6.190550327301025]
|
b01f0c9d-c917-4988-b5f2-7cda9574d3ad
|
fast-and-flexible-indoor-scene-synthesis-via
|
1811.12463
| null |
http://arxiv.org/abs/1811.12463v1
|
http://arxiv.org/pdf/1811.12463v1.pdf
|
Fast and Flexible Indoor Scene Synthesis via Deep Convolutional Generative Models
|
We present a new, fast and flexible pipeline for indoor scene synthesis that
is based on deep convolutional generative models. Our method operates on a
top-down image-based representation, and inserts objects iteratively into the
scene by predicting their category, location, orientation and size with
separate neural network modules. Our pipeline naturally supports automatic
completion of partial scenes, as well as synthesis of complete scenes. Our
method is significantly faster than the previous image-based method and
generates result that outperforms it and other state-of-the-art deep generative
scene models in terms of faithfulness to training data and perceived visual
quality.
|
['Yu-an Lin', 'Kai Wang', 'Daniel Ritchie']
|
2018-11-29
|
fast-and-flexible-indoor-scene-synthesis-via-1
|
http://openaccess.thecvf.com/content_CVPR_2019/html/Ritchie_Fast_and_Flexible_Indoor_Scene_Synthesis_via_Deep_Convolutional_Generative_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/Ritchie_Fast_and_Flexible_Indoor_Scene_Synthesis_via_Deep_Convolutional_Generative_CVPR_2019_paper.pdf
|
cvpr-2019-6
|
['indoor-scene-synthesis']
|
['computer-vision']
|
[ 1.38822034e-01 -4.78791073e-02 6.15288615e-01 -5.59917271e-01
-2.46729180e-01 -6.11801505e-01 7.09969282e-01 -2.42771208e-01
1.60193115e-01 5.40496945e-01 4.57979470e-01 -1.79916501e-01
1.26985952e-01 -1.34728527e+00 -1.12153566e+00 -2.25817278e-01
2.61217654e-01 6.09696329e-01 4.29902762e-01 -1.60652190e-01
1.65249020e-01 7.35109389e-01 -1.68689680e+00 4.44050848e-01
5.83184421e-01 8.58336866e-01 6.68246806e-01 1.23047292e+00
8.08408111e-02 1.17227709e+00 -5.63756824e-01 -2.96144933e-01
2.57088304e-01 -3.12250167e-01 -5.79888999e-01 6.63839221e-01
7.34328449e-01 -6.65388286e-01 -6.94586217e-01 4.03898746e-01
5.12831807e-01 1.79328144e-01 5.36434591e-01 -9.34402108e-01
-1.04357243e+00 4.42822650e-02 -7.15907514e-02 -3.42383593e-01
5.61563849e-01 2.99217880e-01 6.28590465e-01 -1.14361453e+00
7.27500081e-01 1.15794325e+00 7.97325909e-01 2.54131347e-01
-1.52734733e+00 -3.13541263e-01 2.14573443e-01 -2.46844187e-01
-1.20262802e+00 -5.82626581e-01 6.66110575e-01 -6.11099362e-01
1.21321118e+00 2.85781920e-01 8.77380848e-01 9.78640437e-01
2.20593572e-01 5.30856192e-01 8.08960497e-01 -2.93970078e-01
4.23217773e-01 -1.26592517e-01 -6.59100652e-01 8.43123913e-01
-5.61700799e-02 5.52506149e-02 -2.97456712e-01 2.15446621e-01
1.45236719e+00 2.24668294e-01 1.18257269e-01 -7.54775584e-01
-1.33290398e+00 5.48440337e-01 9.06165302e-01 -1.46029457e-01
-4.04784083e-01 6.18950605e-01 -2.05451146e-01 -3.10635239e-01
4.72883642e-01 5.17563283e-01 -2.48963967e-01 1.25417277e-01
-1.10458672e+00 6.87945366e-01 6.41980767e-01 1.36826086e+00
8.58093500e-01 2.68238574e-01 -3.82090300e-01 6.60707891e-01
4.00039256e-01 5.13638735e-01 -1.61656648e-01 -1.40709090e+00
3.28770815e-03 5.72092116e-01 1.74058333e-01 -1.01828980e+00
-3.78604084e-01 -4.21606988e-01 -6.67827070e-01 3.47712219e-01
-2.66894728e-01 3.24285589e-02 -1.32158470e+00 1.29703343e+00
1.81564957e-01 4.10584062e-02 -2.96801925e-01 6.13965034e-01
1.40054440e+00 8.09381306e-01 -7.31760710e-02 3.32434475e-01
8.81570160e-01 -1.34699345e+00 -5.30615151e-01 -6.05941594e-01
-1.39840081e-01 -1.06739402e+00 9.90449488e-01 4.30405855e-01
-1.29786015e+00 -9.80226755e-01 -9.63652551e-01 -4.86375004e-01
-2.57466286e-01 4.15844828e-01 1.19186234e+00 6.97466791e-01
-1.58295250e+00 4.57776636e-01 -6.43939555e-01 -3.80745798e-01
5.80793262e-01 1.96371779e-01 -1.08693592e-01 -3.65155816e-01
-2.19585598e-01 4.25427943e-01 3.70178729e-01 -4.59249876e-02
-1.43913424e+00 -7.74393976e-01 -1.25293791e+00 6.31423248e-03
1.50460536e-02 -1.41231012e+00 1.43154359e+00 -6.06331348e-01
-1.68551087e+00 7.43484020e-01 -1.85143173e-01 -1.09969705e-01
3.85196686e-01 -3.00147563e-01 -5.23266494e-02 -1.97476134e-01
1.97573662e-01 1.09278035e+00 7.29852498e-01 -1.81393778e+00
-5.21883667e-01 1.08622268e-01 2.15517327e-01 2.28988186e-01
2.68248916e-01 -2.41701081e-01 -6.04307950e-01 -5.80075085e-01
1.89967126e-01 -7.54691124e-01 -5.78251958e-01 1.77534565e-01
-6.42346144e-01 3.23646754e-01 7.94751346e-01 -5.52329242e-01
9.05945420e-01 -1.99222457e+00 -9.68376081e-03 -7.67598227e-02
2.34065294e-01 -2.31961980e-01 -9.09985881e-03 4.66915071e-01
1.39557764e-01 -7.99743608e-02 -5.93625233e-02 -9.62427497e-01
7.19070807e-03 2.67544925e-01 -4.89073575e-01 2.05476433e-01
1.57282174e-01 9.63643670e-01 -9.74301815e-01 -2.21609741e-01
9.77861524e-01 6.14535213e-01 -1.02467060e+00 6.65777326e-01
-5.69800079e-01 5.66667795e-01 1.39975205e-01 7.04473794e-01
8.59266162e-01 -3.13852966e-01 9.72251594e-02 -2.80693591e-01
-2.65555143e-01 4.37436223e-01 -1.04897738e+00 2.26895785e+00
-6.85914576e-01 7.82947421e-01 -2.20045269e-01 -3.65048945e-01
9.55093563e-01 -6.37343479e-03 2.98298091e-01 -3.85625452e-01
-1.20193891e-01 -1.26034975e-01 -4.56303269e-01 -3.21695238e-01
1.05236042e+00 2.92459518e-01 -5.94893657e-02 1.18487127e-01
3.23370188e-01 -9.10911918e-01 1.88340068e-01 3.70092243e-01
9.79095101e-01 8.74525130e-01 1.80725455e-01 -9.00381431e-02
-1.43219873e-01 -9.54106525e-02 1.39172167e-01 7.42035151e-01
3.61874789e-01 1.26848459e+00 -8.82077441e-02 -6.84226215e-01
-1.47401786e+00 -1.48869061e+00 6.75832406e-02 1.06661284e+00
1.63059011e-01 -7.08285034e-01 -6.77929521e-01 -1.70208409e-01
-1.83504879e-01 8.42723906e-01 -6.86196804e-01 1.32636309e-01
-3.64139467e-01 -4.09083635e-01 -3.45703424e-03 8.74196589e-01
5.58372498e-01 -1.26457632e+00 -6.38587952e-01 2.77499080e-01
-1.20219372e-01 -1.30288649e+00 -2.01177135e-01 1.16627038e-01
-6.80779219e-01 -7.11934805e-01 -4.40087289e-01 -8.18198204e-01
9.29209888e-01 4.48543489e-01 1.62394118e+00 7.94497039e-03
-5.46456158e-01 4.34323668e-01 -1.10073842e-01 -3.70206296e-01
-3.88220668e-01 -2.13774949e-01 -3.36660296e-01 -2.97985762e-01
-3.45179468e-01 -7.34022796e-01 -8.51398647e-01 1.16734959e-01
-7.89991558e-01 7.52883434e-01 3.76151919e-01 3.81539583e-01
9.25882816e-01 1.30248547e-01 -1.84202105e-01 -7.00056493e-01
3.12545985e-01 -1.97434545e-01 -6.92111671e-01 2.03746054e-02
-1.54252142e-01 -1.68430731e-02 5.78984380e-01 6.96145324e-03
-1.37853408e+00 5.05528808e-01 -3.53534162e-01 -2.83695459e-01
-6.84666634e-01 -1.03514962e-01 -2.42392316e-01 3.63041610e-02
7.45138824e-01 3.99500430e-01 -6.71050012e-01 -3.74640524e-01
6.99274361e-01 1.56216696e-01 9.24691737e-01 -4.76656437e-01
8.62175107e-01 6.87381685e-01 -1.04039321e-02 -7.56038487e-01
-9.38439429e-01 -1.74923629e-01 -9.53058064e-01 -2.95904070e-01
9.11519527e-01 -1.13979995e+00 -4.50563967e-01 6.28152907e-01
-1.42007482e+00 -7.27865577e-01 -4.99171823e-01 2.15172786e-02
-8.94962251e-01 -2.02716634e-01 -5.18878758e-01 -7.15888023e-01
5.23367710e-02 -1.09043121e+00 1.56539047e+00 3.24190557e-01
-1.39449298e-01 -9.06284153e-01 2.28392884e-01 2.26235628e-01
6.23366416e-01 4.73012388e-01 4.12148595e-01 2.31793046e-01
-1.27365088e+00 -2.57588923e-02 -3.37320596e-01 -3.54309231e-02
2.54741818e-01 4.78395253e-01 -1.16538954e+00 -1.28502488e-01
-3.38993251e-01 -2.33385861e-01 7.06209123e-01 8.91110718e-01
1.51952446e+00 -2.12848410e-01 -3.29488218e-01 1.22652376e+00
1.56141102e+00 3.55849713e-02 1.10454953e+00 3.80592197e-01
9.44796681e-01 2.87366182e-01 1.51771262e-01 5.79330444e-01
7.80737102e-01 5.61704516e-01 7.87341595e-01 -4.20963407e-01
-5.77502787e-01 -7.54004300e-01 -4.16416451e-02 3.84853244e-01
-2.78321505e-01 -4.46205914e-01 -4.65070158e-01 5.28168976e-01
-1.77720737e+00 -1.06257689e+00 -2.37173900e-01 1.94691038e+00
3.94312948e-01 4.45916690e-02 9.24263056e-03 -1.83958665e-01
2.70993918e-01 2.49383330e-01 -3.19041759e-01 -5.02647221e-01
-7.62558579e-02 3.19275528e-01 4.11310047e-01 4.00207043e-01
-1.21274674e+00 1.09642470e+00 8.46197796e+00 4.69987273e-01
-7.89905012e-01 -1.34691685e-01 8.84273767e-01 -2.82731764e-02
-4.77589935e-01 -1.59740695e-04 -7.02682734e-01 1.25910476e-01
5.66486776e-01 2.64074296e-01 7.70204365e-01 1.28806388e+00
1.91029087e-01 -2.51091689e-01 -1.15109134e+00 1.06788325e+00
2.23891556e-01 -1.90082943e+00 2.15729997e-02 2.81938184e-02
1.16988730e+00 3.79546992e-02 8.16325694e-02 1.12365223e-01
9.38659608e-01 -1.18277335e+00 1.25988519e+00 7.32554317e-01
8.72834146e-01 -7.34235525e-01 2.19879299e-01 1.88429281e-01
-1.32307899e+00 1.54332995e-01 -5.99312365e-01 -2.64651597e-01
3.07417572e-01 5.31696498e-01 -8.78378451e-01 3.23446363e-01
9.21595395e-01 7.68564045e-01 -8.86602700e-01 1.11335671e+00
-5.76641023e-01 3.25100422e-01 -1.70090288e-01 3.04477185e-01
-1.71234816e-01 7.58048818e-02 1.46889359e-01 1.41600728e+00
5.38900316e-01 -4.20121253e-02 3.63399506e-01 1.47725081e+00
-1.31645784e-01 -3.04099530e-01 -7.75572836e-01 2.85449475e-01
4.61693585e-01 1.37095273e+00 -8.28087032e-01 -4.24363613e-01
9.88937169e-03 1.30580950e+00 3.30350876e-01 3.33254993e-01
-9.69248772e-01 -1.47363156e-01 4.69525069e-01 4.54444766e-01
5.31101644e-01 -5.83309293e-01 -5.70774198e-01 -9.90094423e-01
-2.79834867e-01 -3.37650687e-01 -2.52468914e-01 -1.42450428e+00
-9.85477507e-01 7.71762013e-01 -1.39395803e-01 -1.07848251e+00
-3.13602954e-01 -5.13715446e-01 -7.21273184e-01 6.62709594e-01
-1.18835783e+00 -1.66139841e+00 -8.86420310e-01 6.22530580e-01
8.69620264e-01 9.39308852e-02 1.10190403e+00 1.76745281e-01
-4.18223500e-01 9.20125023e-02 -2.08526954e-01 -2.55885646e-02
3.84863049e-01 -1.41893959e+00 9.95416760e-01 1.15117884e+00
2.44729266e-01 4.55677539e-01 8.09693277e-01 -6.00524843e-01
-1.19081330e+00 -1.37934029e+00 4.51901495e-01 -6.92673683e-01
1.06657594e-01 -7.05201924e-01 -1.42834738e-01 8.91880572e-01
4.24790144e-01 3.76034342e-02 5.16206622e-01 8.56002867e-02
-1.44816801e-01 -1.11573841e-02 -1.04461706e+00 5.87240994e-01
1.39747429e+00 -2.23896042e-01 -1.72288522e-01 5.58153987e-01
9.77945805e-01 -9.89825070e-01 -5.24079800e-01 2.64225483e-01
4.24153596e-01 -1.52323139e+00 1.36601353e+00 -1.37536794e-01
8.88551772e-01 -4.83118594e-01 -3.47256213e-01 -1.32226586e+00
-1.12906754e+00 -4.46829438e-01 -2.18448147e-01 9.31433558e-01
1.70312226e-01 1.49867591e-02 7.40589499e-01 4.99144047e-01
-6.93885684e-01 -4.85700577e-01 -3.22407424e-01 -3.93321544e-01
-5.37110388e-01 -6.06875002e-01 8.01513374e-01 3.43631178e-01
-7.85065711e-01 2.72483051e-01 -5.67954779e-01 1.35756046e-01
6.36254609e-01 4.73109633e-01 1.44066727e+00 -9.23498333e-01
-4.75457996e-01 -3.09532553e-01 -4.12842929e-01 -1.33831835e+00
-3.40887666e-01 -4.61476564e-01 3.72604191e-01 -2.46711540e+00
1.47200271e-01 -3.35760683e-01 2.70275414e-01 4.41754758e-01
1.86857115e-02 6.64514124e-01 9.47883073e-03 -1.51569039e-01
-7.64206350e-01 5.59588015e-01 1.33034456e+00 3.20504904e-02
-2.67621100e-01 -1.59490556e-01 -5.92918396e-01 8.75853777e-01
4.86375391e-01 -3.49009782e-02 -3.91044229e-01 -6.74618542e-01
3.39115709e-01 -4.20906693e-01 6.74749911e-01 -1.36628389e+00
-8.34370125e-03 -3.87707859e-01 1.11933279e+00 -8.83120954e-01
7.39771962e-01 -5.21188259e-01 6.18316531e-01 2.51215935e-01
-1.77251659e-02 -1.77350442e-03 2.97460854e-01 2.71890372e-01
2.06324980e-01 2.83216774e-01 5.57838798e-01 -4.89097714e-01
-8.52721095e-01 3.72189552e-01 -2.38136783e-01 -4.82603639e-01
9.72479165e-01 -4.81790841e-01 -4.38101023e-01 -5.89092374e-01
-6.97517991e-01 -2.21283078e-01 8.82335842e-01 6.08821154e-01
1.00775635e+00 -1.57304382e+00 -6.85415328e-01 4.94440973e-01
1.26841605e-01 5.02053738e-01 3.65682214e-01 4.32947502e-02
-1.17661989e+00 3.32585305e-01 -2.41479754e-01 -8.32301140e-01
-9.31171179e-01 7.00852454e-01 2.93374896e-01 -4.09373716e-02
-5.91811121e-01 1.06187284e+00 6.22977734e-01 -6.94492638e-01
-7.77551085e-02 -5.18344402e-01 2.76524305e-01 -7.32582986e-01
3.49660099e-01 2.10262015e-01 5.86009584e-02 -6.37939453e-01
-2.41949767e-01 3.54845732e-01 4.49461251e-01 5.50197028e-02
1.58741605e+00 -1.59174994e-01 -1.04781970e-01 2.43634611e-01
7.05743134e-01 2.48201028e-01 -1.73890281e+00 8.12088996e-02
-9.14616585e-01 -9.33126211e-01 1.28788576e-01 -9.86108363e-01
-9.35945094e-01 6.52256846e-01 3.44057441e-01 3.95045616e-02
1.07993257e+00 1.32820785e-01 6.35226607e-01 6.25136048e-02
4.81149018e-01 -9.11021471e-01 5.00523567e-01 5.25613129e-01
1.15168166e+00 -1.06615901e+00 2.55063415e-01 -5.36766648e-01
-4.13734883e-01 1.05686951e+00 7.50709653e-01 -4.53643501e-01
5.70623279e-01 6.10504389e-01 -3.10441375e-01 -3.17338854e-01
-5.89789093e-01 -3.53063680e-02 5.76678216e-01 8.94096017e-01
5.76573491e-01 3.41338187e-01 4.62670624e-01 1.77648276e-01
-7.08817124e-01 -4.99917753e-02 4.52389926e-01 9.21686292e-01
-5.40748000e-01 -8.35127950e-01 -4.73864406e-01 1.15504049e-01
1.63924217e-01 -1.93204179e-01 -2.07292527e-01 3.59301209e-01
5.33731520e-01 9.59165633e-01 2.66610056e-01 -4.72529739e-01
4.50920224e-01 -5.29064655e-01 8.53135943e-01 -9.63575542e-01
-3.12289059e-01 1.62126392e-01 -7.29763284e-02 -8.66095722e-01
-3.80974025e-01 -3.52189869e-01 -8.56061101e-01 -4.44923997e-01
-9.99171138e-02 -6.18225157e-01 8.07487488e-01 6.08407557e-01
5.81884563e-01 1.10319245e+00 6.88133478e-01 -1.78264654e+00
4.30802584e-01 -8.67386937e-01 -1.93365872e-01 3.44262183e-01
2.21948430e-01 -5.26214838e-01 1.61122754e-01 4.76295263e-01]
|
[9.204151153564453, -3.0983028411865234]
|
119f8ed2-abbe-4fd5-bf41-a73cc7023efb
|
object-guided-instance-segmentation-for
|
1911.09199
| null |
https://arxiv.org/abs/1911.09199v1
|
https://arxiv.org/pdf/1911.09199v1.pdf
|
Object-Guided Instance Segmentation for Biological Images
|
Instance segmentation of biological images is essential for studying object behaviors and properties. The challenges, such as clustering, occlusion, and adhesion problems of the objects, make instance segmentation a non-trivial task. Current box-free instance segmentation methods typically rely on local pixel-level information. Due to a lack of global object view, these methods are prone to over- or under-segmentation. On the contrary, the box-based instance segmentation methods incorporate object detection into the segmentation, performing better in identifying the individual instances. In this paper, we propose a new box-based instance segmentation method. Mainly, we locate the object bounding boxes from their center points. The object features are subsequently reused in the segmentation branch as a guide to separate the clustered instances within an RoI patch. Along with the instance normalization, the model is able to recover the target object distribution and suppress the distribution of neighboring attached objects. Consequently, the proposed model performs excellently in segmenting the clustered objects while retaining the target object details. The proposed method achieves state-of-the-art performances on three biological datasets: cell nuclei, plant phenotyping dataset, and neural cells.
|
['Daniel J. Hoeppner', 'Dimitris N. Metaxas', 'Jingru Yi', 'Wei Fan', 'Lianyi Han', 'Bo Liu', 'Pengxiang Wu', 'Hui Tang']
|
2019-11-20
| null | null | null | null |
['plant-phenotyping']
|
['computer-vision']
|
[ 4.41387177e-01 6.72615916e-02 -1.41944483e-01 -3.00655931e-01
-5.84397972e-01 -3.86291325e-01 2.38088921e-01 6.86171234e-01
-3.68368208e-01 5.47276556e-01 -7.35854566e-01 1.93080425e-01
-3.87337734e-03 -8.13219726e-01 -6.33541584e-01 -1.19827938e+00
3.53656292e-01 6.02912307e-01 7.09878027e-01 2.90447205e-01
5.49272358e-01 7.54564881e-01 -1.49137282e+00 5.69742359e-02
1.08646441e+00 1.13567865e+00 4.99757886e-01 3.22624922e-01
-5.16656458e-01 2.90003181e-01 -5.63815296e-01 1.61960334e-01
-3.29507366e-02 -2.80091137e-01 -6.41387343e-01 6.09298944e-01
1.84185311e-01 2.19724327e-02 2.99237132e-01 1.22420704e+00
2.13664114e-01 -7.37052597e-03 8.27733696e-01 -1.01691031e+00
-1.14985771e-01 1.62042707e-01 -1.00376284e+00 1.13820583e-01
-4.30004239e-01 -6.87781349e-02 8.13734889e-01 -8.10346365e-01
6.66315734e-01 8.26624215e-01 2.57519990e-01 4.31118369e-01
-1.40313697e+00 -4.41318691e-01 4.21546280e-01 6.65118694e-02
-1.44999707e+00 -3.51740628e-01 9.52187002e-01 -4.78924662e-01
2.34978780e-01 3.26923281e-01 4.98229146e-01 2.54993021e-01
-1.36273667e-01 9.78877842e-01 1.05108881e+00 -2.32681811e-01
4.48015273e-01 1.40983164e-01 5.80132127e-01 4.03982610e-01
3.52503508e-01 -4.95216340e-01 7.20546618e-02 4.83595803e-02
6.78689599e-01 3.85630727e-01 -3.29845846e-01 -6.09760821e-01
-1.15508342e+00 3.02639782e-01 5.29193878e-01 5.75614333e-01
-4.12616819e-01 -2.46730119e-01 2.33641669e-01 -6.01465106e-01
5.47829092e-01 1.11271150e-01 -5.57434499e-01 3.44115853e-01
-1.21225631e+00 3.29232246e-01 4.40126657e-01 8.79606068e-01
9.64549482e-01 -3.09721798e-01 -2.73367018e-01 7.80719936e-01
1.77079678e-01 1.65761307e-01 1.55297384e-01 -6.50455534e-01
9.82907638e-02 1.18206584e+00 7.44976476e-02 -1.15994394e+00
-5.35112679e-01 -6.11177444e-01 -8.76442552e-01 1.39021605e-01
8.72583270e-01 2.16786638e-01 -1.17515290e+00 1.36099553e+00
8.77213240e-01 3.60557698e-02 -3.03036809e-01 8.15909088e-01
8.66449058e-01 7.24023819e-01 1.83942392e-01 -4.36920166e-01
1.52868342e+00 -8.48378181e-01 -6.56346083e-01 -1.60150111e-01
5.83437383e-01 -5.62628329e-01 7.80744195e-01 2.37371162e-01
-9.28875208e-01 -4.76314068e-01 -8.30808401e-01 1.29071370e-01
-4.71762359e-01 5.10276258e-01 4.25853759e-01 2.81486899e-01
-3.92770231e-01 5.04241288e-01 -9.58451033e-01 -3.05542648e-01
9.44416523e-01 4.62338656e-01 -2.67233312e-01 7.74507523e-02
-2.25781277e-01 3.15349609e-01 6.07733965e-01 1.32573426e-01
-6.59794688e-01 -5.97184181e-01 -6.09007359e-01 1.85942307e-01
5.72083831e-01 -1.16626076e-01 6.29677296e-01 -8.38942051e-01
-1.23813522e+00 1.00919044e+00 -3.73358577e-01 -1.19981714e-01
3.06199521e-01 2.96225846e-01 1.94631666e-01 2.74374187e-01
8.77484754e-02 7.03501821e-01 7.33689964e-01 -1.57529700e+00
-7.94657946e-01 -9.73142922e-01 -3.00202519e-01 5.73736057e-02
-1.45616665e-01 -2.34154001e-01 -7.20103443e-01 -5.07826924e-01
7.55007803e-01 -6.66843593e-01 -2.65532792e-01 4.82919514e-02
-6.85451865e-01 -1.08463533e-01 1.32865918e+00 -5.30712724e-01
1.03311324e+00 -2.29844022e+00 1.81926057e-01 2.79055923e-01
3.81065041e-01 1.86949715e-01 9.41056162e-02 -1.48339614e-01
4.46867906e-02 1.31002188e-01 -5.55416107e-01 -2.57753700e-01
-3.60735416e-01 6.24546930e-02 2.72349119e-01 7.27226198e-01
2.99633384e-01 6.58928156e-01 -4.96314645e-01 -9.66657817e-01
2.48331964e-01 3.46678883e-01 -3.33488345e-01 1.84691802e-01
-4.05681133e-01 8.38033855e-01 -6.20213509e-01 1.02484894e+00
9.59753036e-01 -2.20565662e-01 1.64226461e-02 -2.65436649e-01
-1.36613205e-01 -2.85789818e-01 -1.28346109e+00 1.25537157e+00
1.65867418e-01 2.25043923e-01 6.04053080e-01 -1.35907435e+00
1.01196003e+00 -3.47229242e-02 8.59574318e-01 -3.43401968e-01
3.21064234e-01 3.72991234e-01 5.36171868e-02 -2.98412651e-01
1.01068102e-01 2.61563323e-02 2.53412575e-01 7.17836320e-02
-9.28752050e-02 -1.84070051e-01 5.44499755e-01 -9.13312882e-02
6.55814946e-01 1.34265825e-01 2.93653607e-01 -4.78520215e-01
7.26198554e-01 1.61153361e-01 8.97301733e-01 3.32017809e-01
-3.47378314e-01 8.25331211e-01 5.86621106e-01 -1.85671270e-01
-8.89203608e-01 -6.04416907e-01 -5.47228277e-01 5.85817039e-01
5.81812084e-01 4.30965703e-03 -1.25086403e+00 -6.79544449e-01
-2.32721210e-01 2.93394417e-01 -5.17120779e-01 8.84011164e-02
-4.84045178e-01 -1.07028186e+00 8.41250941e-02 2.52610326e-01
5.00398219e-01 -9.61462021e-01 -6.98468029e-01 3.69721591e-01
-2.21030727e-01 -1.05818176e+00 -2.90009946e-01 3.89304459e-01
-1.14015508e+00 -1.14950311e+00 -9.87219155e-01 -9.13855135e-01
1.29408145e+00 3.47256869e-01 7.37318337e-01 3.68604630e-01
-5.13851345e-01 -3.04332405e-01 -2.18544602e-01 -3.97663027e-01
5.83448783e-02 1.93619043e-01 -4.20806766e-01 2.71782488e-01
1.10150911e-01 -2.98685759e-01 -7.21399724e-01 6.91245854e-01
-9.49083090e-01 -5.08759059e-02 4.30397958e-01 9.09279227e-01
1.26468825e+00 4.16373014e-01 3.61368924e-01 -1.29219007e+00
-5.59010133e-02 -1.88773572e-01 -7.58995712e-01 1.39235839e-01
-2.79472649e-01 -4.02723730e-01 5.76655090e-01 -5.08622766e-01
-9.17140186e-01 4.04635727e-01 1.97871000e-01 -5.75090498e-02
-5.72716355e-01 2.53672838e-01 -7.25613415e-01 -1.03938416e-01
3.10756326e-01 2.92353600e-01 -1.22250123e-02 -4.02796984e-01
-5.32187335e-02 5.29551566e-01 3.50957155e-01 -5.17232001e-01
5.47511578e-01 8.29081357e-01 1.03788011e-01 -1.00045323e+00
-7.11054981e-01 -6.88277423e-01 -8.74065101e-01 -2.50451177e-01
9.57235456e-01 -4.23918843e-01 -6.24706328e-01 6.65208876e-01
-1.03830874e+00 -2.58331567e-01 -3.07423323e-01 1.33594915e-01
-4.18165535e-01 3.51455361e-01 -4.19882268e-01 -8.41372490e-01
-1.00287244e-01 -1.36526096e+00 1.29007912e+00 7.00319529e-01
1.55950069e-01 -6.66772485e-01 -2.86226630e-01 4.29510385e-01
5.17541170e-02 3.26408058e-01 1.22672677e+00 -7.71886647e-01
-7.99326122e-01 -5.25600374e-01 -3.12199533e-01 1.26858637e-01
2.34771028e-01 4.53782916e-01 -1.00940788e+00 -6.92475513e-02
-8.46269205e-02 1.32152781e-01 7.95257092e-01 9.06378150e-01
1.51264524e+00 2.66913860e-03 -8.43369186e-01 6.60330057e-01
1.42725623e+00 4.08845186e-01 5.94752133e-01 2.14248478e-01
7.01752245e-01 1.04691744e+00 9.29160595e-01 3.51653397e-01
-9.77222249e-02 4.88396525e-01 6.03628695e-01 -4.85671937e-01
1.55054271e-01 2.80173123e-01 -3.16729665e-01 3.24737072e-01
6.79230988e-02 -3.04603875e-01 -8.60971928e-01 7.06113398e-01
-1.75561774e+00 -5.80205381e-01 -4.64037091e-01 2.06662583e+00
6.28593504e-01 2.59850889e-01 2.99862862e-01 3.81106824e-01
1.04878330e+00 -7.96084329e-02 -7.95242488e-01 1.59294963e-01
-1.28851250e-01 -8.25918745e-03 4.78205204e-01 2.01037899e-01
-1.32419646e+00 9.53354418e-01 5.13008451e+00 1.13011909e+00
-8.42206359e-01 -1.59628734e-01 1.17792237e+00 2.25985140e-01
2.23248139e-01 3.60078886e-02 -1.11523628e+00 5.49025536e-01
7.12169632e-02 3.13573897e-01 -1.82945244e-02 6.94025099e-01
3.15280080e-01 -4.56373662e-01 -9.00823832e-01 6.81995034e-01
-2.73933768e-01 -1.00801229e+00 -5.16314730e-02 1.98842391e-01
6.80759311e-01 -5.63628376e-01 -1.46988392e-01 -1.06465444e-01
-3.89697611e-01 -8.78105640e-01 6.41478837e-01 4.40126926e-01
3.46190929e-01 -8.57181370e-01 7.64671504e-01 6.63905799e-01
-1.13055408e+00 7.32133165e-03 -5.93212664e-01 3.40448856e-01
-8.89934152e-02 1.10310400e+00 -5.66903532e-01 3.19145590e-01
5.51169991e-01 5.38600981e-01 -5.43234348e-01 1.35033965e+00
2.67876923e-01 4.62113559e-01 -4.22464520e-01 9.19693336e-02
4.72977422e-02 -5.51545739e-01 5.40551484e-01 9.61008906e-01
2.18005907e-02 2.82324016e-01 4.37978357e-01 1.20574582e+00
1.17797606e-01 4.64377463e-01 -1.51603386e-01 -9.61922631e-02
2.67886788e-01 1.65965199e+00 -1.71217597e+00 -3.06442022e-01
-1.68675825e-01 6.12352967e-01 3.26104283e-01 3.15230340e-01
-7.10307777e-01 -4.72008258e-01 3.56209785e-01 4.33819890e-01
4.53025609e-01 -1.44714087e-01 -6.91579878e-01 -7.22100675e-01
3.91587988e-02 -7.28792250e-01 1.04925834e-01 -3.25260997e-01
-9.15934265e-01 2.01128691e-01 -2.10065246e-01 -1.01230252e+00
5.71681738e-01 -6.71928883e-01 -5.72281718e-01 5.82179725e-01
-1.28971696e+00 -1.02774334e+00 -6.22341633e-01 1.85885921e-01
5.66094220e-01 3.11876506e-01 3.88263881e-01 4.41985011e-01
-9.97203648e-01 6.65405914e-02 1.92414254e-01 8.39029029e-02
3.91910553e-01 -1.16469240e+00 -2.95579374e-01 7.40444660e-01
-1.25114694e-01 6.05989933e-01 5.88205874e-01 -7.22466946e-01
-7.66845107e-01 -1.06470478e+00 3.44254225e-01 -7.93988705e-02
1.57031566e-01 -3.98487151e-01 -1.12450266e+00 1.94396406e-01
-3.37749094e-01 3.25166047e-01 4.00144309e-01 -2.46027038e-01
3.16184670e-01 -2.22954571e-01 -1.28418326e+00 5.86304367e-01
6.30013585e-01 2.07895245e-02 -3.85526638e-03 3.72214288e-01
3.54106009e-01 -4.06926304e-01 -7.57740855e-01 5.90774655e-01
3.07264507e-01 -8.92496884e-01 9.06273067e-01 -3.28643233e-01
4.04697597e-01 -6.73711777e-01 2.12878495e-01 -8.52365673e-01
-3.17885160e-01 -1.28098473e-01 1.74020723e-01 1.69295049e+00
2.97713935e-01 -3.80462915e-01 1.08496523e+00 4.86873031e-01
-1.08023420e-01 -9.97239888e-01 -7.73605227e-01 -4.49164093e-01
-6.96455166e-02 -1.59861259e-02 5.41748405e-01 6.85342312e-01
-4.00196463e-01 -9.09443852e-03 3.01875859e-01 2.16181114e-01
8.99282098e-01 3.95476341e-01 7.72351027e-01 -1.44430339e+00
8.87982473e-02 -6.41893446e-01 -5.01098096e-01 -1.00615299e+00
6.51570261e-02 -6.46886051e-01 2.90183961e-01 -1.53068066e+00
3.93440485e-01 -6.78950608e-01 -1.43084273e-01 2.86624819e-01
-4.80030686e-01 3.04815441e-01 -1.52783737e-01 3.40227544e-01
-5.11780202e-01 2.87742198e-01 1.51229703e+00 -4.00122523e-01
-3.49532843e-01 2.59980321e-01 -4.45396811e-01 8.24702442e-01
8.87018442e-01 -4.79583412e-01 -1.47960559e-01 -1.82123599e-03
-4.65910107e-01 -2.66592681e-01 3.55331451e-01 -9.54132020e-01
2.42065251e-01 -2.56848603e-01 5.46428204e-01 -1.00498426e+00
2.59148926e-01 -1.01141620e+00 2.13576239e-02 2.40019962e-01
-2.33574614e-01 -6.09104455e-01 1.38783991e-01 6.26779020e-01
-1.98565662e-01 -4.48384762e-01 1.41174006e+00 -3.47659111e-01
-3.15910101e-01 4.03669894e-01 -3.92196476e-01 7.62443021e-02
1.45589185e+00 -6.37122631e-01 -2.41599843e-01 2.37129450e-01
-7.38209844e-01 2.30442107e-01 6.22982442e-01 -2.84877151e-01
4.16636735e-01 -7.84932137e-01 -3.72806638e-01 2.34601393e-01
2.02981099e-01 8.30033600e-01 3.64912778e-01 1.13774383e+00
-5.01679480e-01 3.01210880e-01 -1.28968909e-01 -1.06970060e+00
-1.50925469e+00 5.04686534e-01 3.66340548e-01 -1.44228324e-01
-4.72079903e-01 7.87245393e-01 8.46280277e-01 -3.99420291e-01
1.96347445e-01 -5.92442691e-01 -4.98511553e-01 1.21321224e-01
2.84398228e-01 4.14861411e-01 -4.69379574e-02 -7.30651975e-01
-3.55637759e-01 8.39483261e-01 -2.55808711e-01 4.40156490e-01
1.23700035e+00 -1.50908977e-01 -3.16408575e-01 5.09361148e-01
8.71638417e-01 -1.65118501e-01 -1.35556221e+00 -1.02493502e-01
2.05946177e-01 -4.56224859e-01 8.74370411e-02 -4.75504547e-01
-1.40811372e+00 9.46981549e-01 4.53795910e-01 3.07620049e-01
1.08080697e+00 1.02345839e-01 5.86294413e-01 -1.22447357e-01
2.17258424e-01 -1.18356526e+00 -1.90773442e-01 1.64622858e-01
2.61682868e-01 -1.09392405e+00 1.96485475e-01 -1.01509988e+00
-2.60361969e-01 9.85387504e-01 8.87199938e-01 -8.77964124e-02
5.55257857e-01 3.35526586e-01 -1.50200605e-01 -2.58632958e-01
-3.44076991e-01 -3.00129831e-01 1.31463870e-01 5.98227441e-01
2.77903140e-01 7.40927504e-03 -3.21773261e-01 7.32894897e-01
4.57479715e-01 -3.11399519e-01 2.81464666e-01 8.69039655e-01
-7.96761274e-01 -8.25829804e-01 -5.80344975e-01 5.30055821e-01
-5.83187819e-01 2.23312616e-01 -5.20258188e-01 9.06753123e-01
2.80902892e-01 6.85597479e-01 1.22751415e-01 1.39468700e-01
3.40388268e-01 -1.98596328e-01 3.03405553e-01 -6.89935625e-01
-4.22658652e-01 5.53739130e-01 -4.58583623e-01 -3.13108653e-01
-4.40799236e-01 -7.22760975e-01 -1.82532310e+00 2.06533447e-01
-8.12916338e-01 -3.98284197e-02 6.63237095e-01 8.63747060e-01
1.90571427e-01 6.25811398e-01 4.96185601e-01 -9.09269929e-01
-1.15285598e-01 -6.80495858e-01 -9.48673010e-01 3.44717532e-01
8.11262876e-02 -7.66717732e-01 -3.24731082e-01 2.63886929e-01]
|
[9.641822814941406, 0.18241679668426514]
|
ae74e055-9e23-4cc8-9674-5fb526683899
|
clip-nav-using-clip-for-zero-shot-vision-and
|
2211.16649
| null |
https://arxiv.org/abs/2211.16649v1
|
https://arxiv.org/pdf/2211.16649v1.pdf
|
CLIP-Nav: Using CLIP for Zero-Shot Vision-and-Language Navigation
|
Household environments are visually diverse. Embodied agents performing Vision-and-Language Navigation (VLN) in the wild must be able to handle this diversity, while also following arbitrary language instructions. Recently, Vision-Language models like CLIP have shown great performance on the task of zero-shot object recognition. In this work, we ask if these models are also capable of zero-shot language grounding. In particular, we utilize CLIP to tackle the novel problem of zero-shot VLN using natural language referring expressions that describe target objects, in contrast to past work that used simple language templates describing object classes. We examine CLIP's capability in making sequential navigational decisions without any dataset-specific finetuning, and study how it influences the path that an agent takes. Our results on the coarse-grained instruction following task of REVERIE demonstrate the navigational capability of CLIP, surpassing the supervised baseline in terms of both success rate (SR) and success weighted by path length (SPL). More importantly, we quantitatively show that our CLIP-based zero-shot approach generalizes better to show consistent performance across environments when compared to SOTA, fully supervised learning approaches when evaluated via Relative Change in Success (RCS).
|
['Gaurav S. Sukhatme', 'Jesse Thomason', 'Robinson Piramuthu', 'Gunnar Sigurdsson', 'Vishnu Sashank Dorbala']
|
2022-11-30
| null | null | null | null |
['vision-and-language-navigation']
|
['robots']
|
[ 1.31320551e-01 -2.45997578e-01 8.91278535e-02 -3.48598540e-01
-7.16824651e-01 -6.37338221e-01 1.05707884e+00 2.49227211e-02
-8.53253722e-01 6.15048885e-01 3.50507170e-01 -4.12234753e-01
-1.00199521e-01 -6.36910558e-01 -7.65106678e-01 -5.74406803e-01
1.83101613e-02 4.20606673e-01 4.08917189e-01 -7.03978062e-01
4.73949283e-01 3.84845108e-01 -2.04037547e+00 1.11907676e-01
7.91484833e-01 5.03253877e-01 6.76062346e-01 7.85726964e-01
-8.54309723e-02 1.21112955e+00 -3.42962295e-01 -3.18535641e-02
1.27259523e-01 -3.03357750e-01 -5.72846353e-01 -6.21576793e-02
9.76927102e-01 -2.82529056e-01 -3.53621632e-01 9.01358962e-01
5.28679788e-01 8.94232512e-01 8.27758431e-01 -1.26064968e+00
-6.76027119e-01 4.59864348e-01 -8.69072750e-02 2.67966926e-01
8.76128554e-01 6.11853600e-01 8.68938148e-01 -9.43822980e-01
1.03966391e+00 1.40083635e+00 5.16558528e-01 7.52590060e-01
-1.39410365e+00 -3.35036337e-01 4.40476954e-01 3.95907760e-01
-1.18838847e+00 -7.74378955e-01 3.04194003e-01 -6.50220990e-01
1.47373950e+00 -1.66295007e-01 4.08390105e-01 1.46369720e+00
2.51511991e-01 7.30195582e-01 1.17074180e+00 -5.62374949e-01
6.34044588e-01 -2.43032232e-01 4.76675540e-01 9.76198018e-01
1.39083281e-01 4.03699547e-01 -9.28560257e-01 3.04070473e-01
4.46929872e-01 -2.12955967e-01 -4.35292482e-01 -7.71890879e-01
-1.34021413e+00 7.06032693e-01 5.40817201e-01 3.33366662e-01
-3.39576513e-01 2.21687704e-01 3.30395401e-01 2.28784487e-01
-1.71098948e-01 6.04259551e-01 -2.02344671e-01 -4.09681618e-01
-6.49978936e-01 3.47302496e-01 7.77157605e-01 1.25304115e+00
6.90258682e-01 1.83350191e-01 -4.74096060e-01 5.28453350e-01
3.78400274e-02 3.91359955e-01 6.46277189e-01 -1.19048858e+00
2.59931803e-01 3.27328831e-01 2.02783808e-01 -7.87024915e-01
-6.86169863e-01 -2.08146527e-01 -1.71543390e-01 8.16254139e-01
7.50302136e-01 1.50670698e-02 -1.23610568e+00 2.01695895e+00
-8.76112431e-02 -1.44548893e-01 3.67351711e-01 8.58217120e-01
9.82813835e-01 4.24655735e-01 3.29363316e-01 1.16323262e-01
1.31545877e+00 -1.32014573e+00 -6.95386410e-01 -6.43179297e-01
8.68595123e-01 -3.39730382e-01 1.69477999e+00 2.40361139e-01
-6.96532667e-01 -5.48984706e-01 -1.07963955e+00 -3.71615440e-01
-7.95488358e-01 -3.07441503e-01 6.06901228e-01 3.97641212e-01
-1.29826665e+00 2.88029969e-01 -7.22750902e-01 -9.76889491e-01
3.03863347e-01 1.17851228e-01 -5.22860825e-01 -1.31704941e-01
-7.69616485e-01 1.16211927e+00 3.08659852e-01 -3.90951991e-01
-1.32219386e+00 -6.07750058e-01 -1.40458310e+00 -7.93615505e-02
6.70182526e-01 -7.09055185e-01 1.41402543e+00 -4.94077802e-01
-1.60987151e+00 8.79343152e-01 -2.39527583e-01 -5.83437681e-01
5.00441134e-01 -1.57034665e-01 -6.94511011e-02 -8.43145922e-02
4.16186690e-01 1.10299253e+00 3.83087784e-01 -1.37488127e+00
-7.33888626e-01 -3.41525465e-01 5.27585030e-01 2.54210621e-01
2.16210991e-01 -3.90149564e-01 -3.82571518e-01 -3.24742734e-01
-6.51652515e-02 -9.59987640e-01 -2.84033418e-01 1.20253451e-01
-1.48767292e-01 -2.58146048e-01 4.73247260e-01 -3.34896505e-01
7.81332672e-01 -2.15158820e+00 3.15951139e-01 -3.01165760e-01
-3.66066769e-02 -2.86563374e-02 -4.85630363e-01 4.35816705e-01
4.45343316e-01 -3.00174356e-01 -1.85463309e-01 -5.68325996e-01
2.85631686e-01 3.71470183e-01 -1.89515263e-01 2.51993060e-01
-2.09184319e-01 1.07616508e+00 -1.12670600e+00 -3.51382345e-01
4.53382760e-01 4.56807494e-01 -6.59873307e-01 -4.25410010e-02
-4.34451461e-01 3.96655738e-01 -1.08802103e-01 6.11261666e-01
2.79328115e-02 -9.58432630e-02 1.43783295e-03 -8.33334308e-03
-3.55600685e-01 6.67179376e-02 -8.49694133e-01 2.28727460e+00
-7.62597501e-01 1.04677820e+00 -1.09189719e-01 -4.11798388e-01
6.69390202e-01 -9.66899619e-02 -1.15136310e-01 -1.18574727e+00
8.25443268e-02 1.07716165e-01 1.32362489e-02 -6.11804068e-01
7.07456172e-01 8.32057968e-02 -1.28906578e-01 3.54457408e-01
2.06558347e-01 -5.74093051e-02 2.87911117e-01 3.46062005e-01
1.08179641e+00 6.05748355e-01 5.19509017e-01 -3.94222617e-01
2.50345469e-01 4.60252225e-01 2.47819051e-02 1.26305842e+00
-5.81361234e-01 4.08843815e-01 9.87583995e-02 -2.00055405e-01
-8.34265113e-01 -1.10865426e+00 1.14990883e-01 1.70695090e+00
4.90170985e-01 -2.66297281e-01 -8.15841377e-01 -3.85495782e-01
-1.95511803e-01 1.52613795e+00 -7.69151688e-01 -1.74635872e-01
-4.82444495e-01 -2.15943947e-01 3.94610137e-01 5.17961860e-01
3.96329761e-01 -1.35850751e+00 -1.49137056e+00 3.42874900e-02
-1.40103415e-01 -1.35577440e+00 -2.96946973e-01 5.08298397e-01
-3.89492273e-01 -9.68634129e-01 -6.09821260e-01 -9.63309586e-01
4.79361147e-01 5.71775198e-01 1.21400380e+00 -2.16457903e-01
-2.34724939e-01 9.89382982e-01 -5.25523603e-01 -3.34398389e-01
-1.67000905e-01 -1.14835076e-01 1.70239076e-01 -4.81165886e-01
5.38866282e-01 -2.85480767e-01 -3.38422894e-01 1.15557082e-01
-5.42084277e-01 1.03080168e-01 3.36439371e-01 7.38045216e-01
3.64202648e-01 -5.11454940e-01 2.08350584e-01 -4.03552085e-01
8.44299853e-01 -2.51281530e-01 -5.47752142e-01 3.93640190e-01
-4.97669280e-01 2.71615952e-01 2.17196345e-01 -4.96445924e-01
-1.03470743e+00 -6.20111078e-02 1.17737338e-01 -9.24109817e-02
-5.44880986e-01 2.75532275e-01 2.35817675e-02 -3.49749565e-01
8.52818251e-01 4.00721014e-01 -5.09010814e-02 -1.29655629e-01
6.57510817e-01 2.51088619e-01 7.06746340e-01 -6.19723976e-01
4.00021404e-01 4.70727652e-01 -9.71197337e-02 -8.42778325e-01
-6.35034919e-01 -4.37021106e-01 -6.79133356e-01 -2.86067665e-01
1.17588937e+00 -8.51699650e-01 -8.10450137e-01 2.66949415e-01
-1.10854089e+00 -9.38169539e-01 -2.77183771e-01 4.97479081e-01
-1.18119895e+00 2.17024535e-02 -4.02807832e-01 -7.66555846e-01
1.52164817e-01 -1.34124982e+00 1.10950458e+00 3.12590539e-01
-4.33825344e-01 -8.42879295e-01 2.16678947e-01 1.18502073e-01
4.59902227e-01 1.46954656e-01 9.37634289e-01 -6.76232159e-01
-5.77713549e-01 2.42027864e-01 -2.16916353e-01 -3.85561407e-01
-1.31992474e-01 -3.36669803e-01 -9.61551785e-01 -2.47744262e-01
-2.44692475e-01 -5.39560437e-01 9.71772492e-01 3.78005981e-01
4.10033554e-01 7.88639784e-02 -3.19420427e-01 6.83665514e-01
1.57274139e+00 3.66873145e-01 4.97901738e-01 8.05681348e-01
5.74242949e-01 7.44607329e-01 6.81873381e-01 2.16760606e-01
6.42618299e-01 9.76566970e-01 5.63582003e-01 3.30816716e-01
-3.58714610e-01 -3.25018793e-01 6.12172186e-01 2.21809641e-01
-5.43286428e-02 -3.28387737e-01 -1.12570953e+00 5.14567554e-01
-1.95865428e+00 -1.08335412e+00 3.44385654e-01 2.00009847e+00
4.13952619e-01 1.88536972e-01 -3.07092275e-02 -3.38888526e-01
8.63026455e-02 3.14549983e-01 -5.51934004e-01 -4.45747674e-01
-1.55100122e-01 -1.61195830e-01 3.82780552e-01 9.16070461e-01
-7.85187244e-01 1.37533247e+00 6.67637587e+00 6.17502391e-01
-8.52751434e-01 2.07002029e-01 -9.16904658e-02 -2.45850906e-01
-6.05563894e-02 -2.81971961e-01 -8.77846837e-01 2.01935798e-01
7.47450233e-01 -9.70501602e-02 6.11131430e-01 8.16303849e-01
1.45136237e-01 -4.99968618e-01 -1.40362239e+00 1.02951002e+00
4.90595192e-01 -1.21677721e+00 7.58135021e-02 -1.54435441e-01
5.67186117e-01 2.25514308e-01 2.10601315e-01 8.79279196e-01
7.13443279e-01 -1.12655079e+00 1.09254932e+00 7.50278592e-01
3.97746027e-01 -2.56056190e-01 3.53844404e-01 5.14505625e-01
-1.00620174e+00 -1.52715892e-01 -1.88159674e-01 -4.05081183e-01
2.49547780e-01 -3.63066375e-01 -7.84263432e-01 1.84733987e-01
7.10344911e-01 5.11131465e-01 -6.72118664e-01 9.24882531e-01
-1.79919302e-01 -2.70691812e-02 -3.61810736e-02 -3.88257474e-01
7.41296113e-01 -6.31254315e-02 6.82322621e-01 1.22806394e+00
2.24846810e-01 1.37549743e-01 4.40717459e-01 7.11611748e-01
3.82117569e-01 -3.27983424e-02 -9.84648287e-01 2.06960350e-01
4.81542051e-01 6.51053727e-01 -7.77295291e-01 -3.47339898e-01
-3.99842799e-01 1.13038373e+00 5.29617727e-01 8.10415566e-01
-6.86635375e-01 -1.41377717e-01 9.03571844e-01 -1.28534704e-01
4.61765558e-01 -7.73780942e-01 -1.48470193e-01 -8.31152558e-01
-1.65454730e-01 -8.48793089e-01 2.33403444e-02 -1.12323284e+00
-8.81997943e-01 6.90804243e-01 2.57459865e-03 -1.09272397e+00
-3.82074624e-01 -9.31823969e-01 -4.73307967e-01 4.33303356e-01
-1.44910204e+00 -1.21315503e+00 -5.75676978e-01 5.79041123e-01
1.14936304e+00 -2.96435207e-01 1.07470250e+00 -9.55131277e-02
-4.84110236e-01 4.88946199e-01 -7.48403445e-02 -1.82340175e-01
6.82625234e-01 -1.18782973e+00 2.33709604e-01 8.35144341e-01
3.80353183e-01 8.36456299e-01 1.16620278e+00 -4.79374021e-01
-1.35978794e+00 -6.79258049e-01 6.44792318e-01 -8.52539659e-01
5.80658853e-01 -2.82011420e-01 -5.45735657e-01 8.96994174e-01
2.40626037e-01 -1.15785681e-01 5.70854306e-01 2.36697212e-01
-6.61736667e-01 4.48765606e-01 -1.06981003e+00 1.16885626e+00
1.50171292e+00 -7.35178113e-01 -8.25188398e-01 2.37951741e-01
9.53896880e-01 -2.52173334e-01 -1.64629638e-01 1.64257511e-01
6.90981328e-01 -1.19098043e+00 9.85900164e-01 -8.37824643e-01
3.39285612e-01 -3.85919422e-01 -7.75366366e-01 -1.24326885e+00
-5.20785749e-01 -2.32087255e-01 3.94059867e-02 7.70815074e-01
3.99094373e-01 -3.80528986e-01 5.59259474e-01 4.22454387e-01
-2.74684131e-01 -4.89345044e-01 -8.66808474e-01 -1.03073037e+00
-1.44111469e-01 -5.81011117e-01 3.33502829e-01 6.32135034e-01
-1.24726223e-03 4.09936011e-01 -9.13904235e-02 3.98529321e-02
6.24385715e-01 -1.22387506e-01 9.28391576e-01 -9.24359739e-01
-6.59254864e-02 -7.72837162e-01 -6.24611914e-01 -1.02743125e+00
4.54920411e-01 -8.63773167e-01 5.55385053e-01 -1.89771390e+00
3.48217674e-02 -9.02534947e-02 -2.25433439e-01 3.69571835e-01
3.15309577e-02 6.47909241e-03 5.28295696e-01 1.37033507e-01
-1.08080757e+00 5.05978286e-01 1.19868553e+00 -2.90139914e-01
-3.24315131e-01 -5.28890252e-01 -4.57729489e-01 8.46027553e-01
5.57186007e-01 7.34381750e-02 -5.53279042e-01 -7.27698267e-01
-4.38479818e-02 -2.52179444e-01 5.61910450e-01 -1.39482951e+00
6.82569921e-01 -2.02103347e-01 -6.05322607e-02 -4.54360038e-01
6.30497515e-01 -6.38262391e-01 -3.19787681e-01 4.69706565e-01
-5.17148137e-01 6.45072535e-02 3.40136051e-01 7.25041091e-01
1.12266809e-01 -2.83076435e-01 4.61281359e-01 -3.33568275e-01
-1.87369823e+00 -7.50533938e-02 -7.30028629e-01 -3.01049114e-03
1.23602653e+00 -6.54105604e-01 -4.85674053e-01 -3.66786331e-01
-7.38901615e-01 3.04681689e-01 8.74921322e-01 5.82284808e-01
6.32042110e-01 -1.10262823e+00 -4.37034994e-01 1.59288779e-01
6.58301651e-01 -3.39947164e-01 1.24476947e-01 7.62180209e-01
-5.06918013e-01 6.30063474e-01 -5.20089686e-01 -5.81138074e-01
-1.02279270e+00 7.25921631e-01 4.38690186e-01 1.22231133e-01
-7.75063634e-01 1.13918412e+00 3.63777786e-01 -6.48521006e-01
6.75718725e-01 -2.24039480e-01 -3.64157557e-01 1.69420376e-01
5.41031003e-01 3.36698353e-01 -2.66343385e-01 -6.96232557e-01
-4.31827128e-01 7.84357727e-01 -2.54382212e-02 -3.85736763e-01
1.00252438e+00 -3.39896888e-01 2.72641808e-01 8.71266484e-01
7.42359698e-01 -1.39707431e-01 -1.56318915e+00 -2.58768409e-01
2.02589497e-01 -2.47044757e-01 -2.46815924e-02 -9.32049692e-01
-3.29888761e-01 7.73491859e-01 7.99778342e-01 -2.89452434e-01
4.63460773e-01 2.43799575e-02 3.70763391e-01 8.83207321e-01
1.00939012e+00 -8.83691728e-01 3.35811108e-01 1.02194118e+00
8.14642131e-01 -1.67757916e+00 -3.37043732e-01 -8.17210451e-02
-7.35639811e-01 8.60932529e-01 7.68321514e-01 5.95721975e-02
2.10287660e-01 2.24233225e-01 1.81964964e-01 -2.81601131e-01
-8.03945303e-01 -7.39829302e-01 6.03716522e-02 1.14740753e+00
2.32469633e-01 1.15809903e-01 9.75508466e-02 2.93606788e-01
-3.87128323e-01 -5.09736082e-03 4.39958245e-01 1.13866007e+00
-8.81801665e-01 -4.26975131e-01 -1.42368332e-01 7.10215345e-02
3.16387534e-01 -1.83338150e-01 -2.30710506e-01 7.60191441e-01
2.93302983e-02 1.10985529e+00 1.60757035e-01 -3.66057664e-01
4.60646927e-01 1.39639869e-01 6.84519887e-01 -8.42763841e-01
-3.25409710e-01 -4.92993653e-01 1.42299667e-01 -8.85009289e-01
-3.56567889e-01 -6.87335610e-01 -1.48704743e+00 -1.19158186e-01
7.65854642e-02 -3.16391200e-01 5.65346062e-01 1.06505656e+00
1.62913114e-01 7.76099801e-01 -2.86961049e-01 -1.22930539e+00
-3.86785179e-01 -7.13751018e-01 -1.45223260e-01 4.60048765e-01
5.52979410e-01 -1.08935702e+00 -3.27407002e-01 -9.19443835e-03]
|
[4.452682971954346, 0.6675182580947876]
|
281f2c56-6902-4b04-86a3-aa46426c9063
|
imitrob-imitation-learning-dataset-for
|
2209.07976
| null |
https://arxiv.org/abs/2209.07976v3
|
https://arxiv.org/pdf/2209.07976v3.pdf
|
Imitrob: Imitation Learning Dataset for Training and Evaluating 6D Object Pose Estimators
|
This paper introduces a dataset for training and evaluating methods for 6D pose estimation of hand-held tools in task demonstrations captured by a standard RGB camera. Despite the significant progress of 6D pose estimation methods, their performance is usually limited for heavily occluded objects, which is a common case in imitation learning, where the object is typically partially occluded by the manipulating hand. Currently, there is a lack of datasets that would enable the development of robust 6D pose estimation methods for these conditions. To overcome this problem, we collect a new dataset (Imitrob) aimed at 6D pose estimation in imitation learning and other applications where a human holds a tool and performs a task. The dataset contains image sequences of nine different tools and twelve manipulation tasks with two camera viewpoints, four human subjects, and left/right hand. Each image is accompanied by an accurate ground truth measurement of the 6D object pose obtained by the HTC Vive motion tracking device. The use of the dataset is demonstrated by training and evaluating a recent 6D object pose estimation method (DOPE) in various setups.
|
['Matus Tuna', 'Jan K. Behrens', 'Radoslav Skoviera', 'Robert Babuska', 'Josef Sivic', 'Gabriela Sejnova', 'Karla Stepanova', 'Jiri Sedlar']
|
2022-09-16
| null | null | null | null |
['6d-pose-estimation-1', '6d-pose-estimation']
|
['computer-vision', 'computer-vision']
|
[ 1.00935608e-01 -2.60532141e-01 -1.90772936e-01 4.43435926e-03
-4.65671569e-01 -5.36025345e-01 6.57717228e-01 -6.25697494e-01
-4.26939934e-01 5.04742920e-01 -3.24431062e-01 1.10052861e-01
-7.18253031e-02 1.27866030e-01 -8.30118477e-01 -5.15933931e-01
-3.14359018e-03 9.41212893e-01 4.11733419e-01 1.00297704e-01
4.63674158e-01 9.68968630e-01 -1.60125399e+00 -1.28246576e-01
2.97432184e-01 8.49400938e-01 9.39821184e-01 6.22707427e-01
3.37044358e-01 5.94222963e-01 -8.84444237e-01 -7.64204264e-02
6.44451916e-01 -1.69000551e-01 -5.28968990e-01 4.95874494e-01
6.80018067e-01 -7.39620566e-01 -4.45335180e-01 7.00098217e-01
5.68791032e-01 2.58159995e-01 7.63390839e-01 -1.68490326e+00
1.18134782e-01 1.45810559e-01 -3.96299005e-01 -2.52233893e-01
9.29246247e-01 2.58652925e-01 3.94846350e-01 -9.19414699e-01
1.21288967e+00 1.36644971e+00 5.41316509e-01 7.67597735e-01
-8.99874210e-01 -4.97340113e-01 -1.37625083e-01 1.79821193e-01
-1.19363177e+00 -2.11505800e-01 8.05105329e-01 -7.10107982e-01
9.33551192e-01 2.19811946e-02 1.00678027e+00 1.53701341e+00
4.41693664e-01 1.02125323e+00 1.05496693e+00 -5.25811434e-01
-3.35036628e-02 1.28089279e-01 -4.32506204e-01 5.05812585e-01
7.64255151e-02 -5.43416403e-02 -6.22193277e-01 1.83658883e-01
1.43383193e+00 4.46938545e-01 -3.81338537e-01 -1.40795338e+00
-1.62807119e+00 2.95923501e-01 2.92030275e-01 2.79787660e-01
-4.41907525e-01 2.63420582e-01 2.72440076e-01 3.88167292e-01
2.94320192e-02 3.32709759e-01 -4.76424813e-01 -6.28216267e-01
-4.78904903e-01 3.53620946e-01 9.83943880e-01 1.58769250e+00
1.55598536e-01 -2.08568543e-01 7.18039367e-03 3.63409281e-01
2.96177268e-01 3.89152884e-01 2.71880507e-01 -1.18740320e+00
6.72645450e-01 6.11761153e-01 6.81244791e-01 -6.36952579e-01
-3.63815099e-01 1.25316575e-01 -1.44897729e-01 8.81789684e-01
8.09404254e-01 1.86763749e-01 -7.46220648e-01 1.14996088e+00
5.27487040e-01 -3.20804656e-01 -3.27434063e-01 1.20357001e+00
6.34179533e-01 1.35883853e-01 -3.11314136e-01 -8.84747654e-02
1.02358139e+00 -1.00929117e+00 -9.46334660e-01 1.02573829e-02
4.59840775e-01 -9.20670092e-01 1.27287126e+00 7.18198836e-01
-8.48281860e-01 -6.47085905e-01 -9.87069130e-01 -1.38188139e-01
-2.74464041e-01 5.09664893e-01 5.76564312e-01 4.21142161e-01
-3.16793084e-01 7.24280596e-01 -9.49350834e-01 -5.24899960e-01
7.39746541e-02 5.32413960e-01 -8.14726412e-01 -8.00155848e-02
-4.58250016e-01 1.54718435e+00 2.46521801e-01 3.24500144e-01
-1.26486480e+00 -4.14008528e-01 -7.55216837e-01 -6.24068379e-01
6.17556095e-01 -3.93294990e-01 1.16670394e+00 -3.34806144e-01
-1.94335437e+00 1.18184853e+00 4.20203269e-01 1.04398159e-02
1.32697785e+00 -8.68861139e-01 2.58873582e-01 9.17630121e-02
9.38387588e-03 4.90853459e-01 1.22452891e+00 -1.26481986e+00
-1.15578860e-01 -8.02209496e-01 1.24157704e-01 4.32116061e-01
9.32114720e-02 5.86528964e-02 -5.81928611e-01 -4.90456820e-01
2.06531212e-01 -1.36071110e+00 1.34059981e-01 7.07749784e-01
-3.60458374e-01 -2.97871530e-01 1.34099424e+00 -6.42528832e-01
1.06230028e-01 -2.06283998e+00 7.72477567e-01 -1.96482331e-01
-6.18514754e-02 2.77556330e-01 5.45656793e-02 4.56883490e-01
1.73829541e-01 -6.47740006e-01 1.20708525e-01 -6.86567843e-01
5.24250157e-02 3.10795397e-01 7.92324096e-02 8.04795563e-01
-3.33690852e-01 6.66706264e-01 -7.77463436e-01 -5.44832110e-01
8.47637773e-01 7.42576122e-01 -5.13769574e-02 7.28695333e-01
-9.16174799e-02 1.00615859e+00 -4.68377978e-01 8.97360742e-01
3.93285275e-01 1.38497055e-01 -1.25880241e-01 -1.04662307e-01
5.32141849e-02 -2.19212413e-01 -1.42288888e+00 2.32806873e+00
-4.83046055e-01 6.78769529e-01 2.16338202e-01 -5.32140613e-01
9.07337546e-01 5.40906608e-01 5.98993599e-01 -9.31281075e-02
4.09715056e-01 4.25128907e-01 -7.33699799e-02 -8.96653593e-01
2.42593482e-01 1.68051004e-01 1.63810000e-01 2.24523678e-01
2.61686534e-01 -9.54124331e-01 -9.70192552e-02 -2.85574228e-01
1.10187793e+00 1.08119893e+00 1.11467138e-01 1.72792971e-01
2.79369742e-01 2.42897328e-02 -8.54969248e-02 5.11976182e-01
-4.24357265e-01 8.07184756e-01 1.78844422e-01 -3.52679789e-01
-1.15086389e+00 -1.05679011e+00 -6.47672743e-04 6.67999089e-01
2.31806666e-01 -6.01908341e-02 -5.77353299e-01 -4.65384096e-01
3.04954290e-01 3.14961940e-01 -5.89817166e-01 2.40388036e-01
-7.28842854e-01 2.40000322e-01 5.32137193e-02 6.40051544e-01
4.53070611e-01 -1.21962214e+00 -1.36907601e+00 -8.57595205e-02
9.52113569e-02 -1.36833715e+00 -2.23689646e-01 1.57726526e-01
-1.06951106e+00 -1.32804704e+00 -1.22111738e+00 -7.57101715e-01
6.19742155e-01 2.85068452e-01 7.52167463e-01 -4.27282304e-01
-3.95144820e-01 8.78752112e-01 -4.52737987e-01 -4.74207103e-01
-2.96115488e-01 -5.90597950e-02 2.23662913e-01 -5.76708615e-01
-1.17878720e-01 -4.99519497e-01 -4.98048544e-01 6.92821980e-01
-2.77319431e-01 -1.71277002e-01 6.09281659e-01 7.18795002e-01
3.23335350e-01 -5.06413698e-01 -1.71921089e-01 -1.13721281e-01
3.74591410e-01 1.68673545e-01 -6.76056027e-01 4.18816358e-02
1.58023208e-01 -4.86063242e-01 1.12783521e-01 -9.29836988e-01
-1.04949796e+00 6.50083780e-01 1.98793203e-01 -1.00719428e+00
-3.85755897e-01 -8.84799436e-02 -3.70045334e-01 -3.77061248e-01
4.67001557e-01 -1.22308508e-01 3.01407039e-01 -6.56976581e-01
6.85050413e-02 7.60518849e-01 6.34565711e-01 -5.63331902e-01
6.10423684e-01 2.37683415e-01 1.66081175e-01 -8.65252137e-01
-1.45099625e-01 -5.21487713e-01 -1.43999374e+00 -6.54681146e-01
7.14210808e-01 -7.15554416e-01 -1.05030942e+00 6.52565539e-01
-1.39747417e+00 -5.05033076e-01 -2.70248741e-01 9.97296274e-01
-1.25450575e+00 6.91499040e-02 -4.11580175e-01 -7.99972653e-01
1.33561000e-04 -1.61748433e+00 1.66336894e+00 -2.34832019e-01
-4.18987066e-01 -4.78427202e-01 -1.90666720e-01 4.56859022e-01
-5.81275411e-02 7.39915192e-01 3.43550920e-01 -2.48697639e-01
-5.93946338e-01 -7.60284901e-01 3.06615442e-01 2.44137585e-01
2.18941465e-01 -1.86999932e-01 -6.41404510e-01 -3.45564306e-01
1.48572519e-01 -6.05271578e-01 3.96631323e-02 2.22659156e-01
9.05424237e-01 4.16006267e-01 -3.18458408e-01 2.17265233e-01
1.02149117e+00 2.26921260e-01 4.05926734e-01 3.69960159e-01
9.54190850e-01 6.38791442e-01 1.09250188e+00 3.04729611e-01
-2.38157064e-01 1.31872642e+00 7.77188301e-01 5.33590496e-01
-2.20261320e-01 -6.54848740e-02 4.09869581e-01 6.61414564e-01
-7.45432198e-01 -1.10380659e-02 -8.33970368e-01 2.60951966e-01
-1.45460749e+00 -3.53877872e-01 -1.79833159e-01 2.22294807e+00
4.32185948e-01 -2.86208224e-02 1.57401696e-01 5.26199043e-01
4.17505533e-01 -1.35928586e-01 -7.41420448e-01 4.23031226e-02
4.01415676e-01 -1.47679999e-01 1.91166669e-01 2.06245244e-01
-8.49514902e-01 8.88140023e-01 5.80435467e+00 3.45974326e-01
-9.51268196e-01 -6.11953288e-02 -6.34231031e-01 -5.33588193e-02
7.11577833e-01 -2.62878478e-01 -4.28013504e-01 1.90944329e-01
6.51123822e-02 4.29641724e-01 1.78899735e-01 1.20981276e+00
-1.79766998e-01 -4.95623231e-01 -1.66166353e+00 1.27108240e+00
2.10253686e-01 -5.38225114e-01 -5.04289448e-01 -8.55327770e-02
5.55407941e-01 -1.45832449e-01 4.31368500e-03 1.66258141e-01
-2.71604031e-01 -8.23655427e-01 7.44057000e-01 3.94485772e-01
7.50893235e-01 -2.12265745e-01 5.42291939e-01 8.96000862e-01
-7.46979594e-01 3.89729291e-02 -2.66585946e-01 -1.76711231e-01
1.91981286e-01 -9.56657082e-02 -1.02539992e+00 3.21016401e-01
8.25086951e-01 6.36152625e-01 -3.15108478e-01 1.02990580e+00
-4.66497183e-01 -2.48664409e-01 -2.75546879e-01 -1.66940480e-01
-1.04781799e-02 -1.31569371e-01 9.62334454e-01 7.33023465e-01
3.17542255e-01 -2.28766873e-01 1.95628684e-02 5.22976577e-01
6.18711300e-02 -1.17849313e-01 -1.00598156e+00 1.28476396e-01
8.05001482e-02 9.34851766e-01 -9.17618096e-01 6.03943355e-02
-1.21839298e-02 1.53408170e+00 5.26187718e-02 1.78316355e-01
-7.00295031e-01 -3.25616419e-01 3.33325416e-01 -4.72680368e-02
3.09822053e-01 -9.17248905e-01 2.81101882e-01 -1.11668038e+00
5.82037270e-01 -9.43611503e-01 -3.05329472e-01 -1.32045507e+00
-7.44007945e-01 3.10701013e-01 5.34038246e-01 -1.77432013e+00
-4.66128677e-01 -1.15768611e+00 -4.96187247e-02 6.14029050e-01
-7.07775593e-01 -1.13373601e+00 -6.91124678e-01 6.04760349e-01
1.06766367e+00 -1.66236326e-01 8.86397541e-01 -5.31728826e-02
-6.57908022e-02 6.85999766e-02 -8.57830048e-02 -2.43911624e-01
8.12099278e-01 -1.15207088e+00 -4.15716097e-02 5.67333177e-02
3.34923826e-02 5.63481808e-01 8.62565041e-01 -4.94277298e-01
-2.15209770e+00 -2.70432591e-01 2.41102815e-01 -1.30741870e+00
4.01863098e-01 -7.74662495e-01 -6.16048038e-01 1.06028652e+00
-1.04707494e-01 3.86593491e-01 -4.04692218e-02 -4.44167435e-01
9.34169963e-02 1.80646196e-01 -1.38556325e+00 5.16127884e-01
1.36854446e+00 -3.61932814e-01 -1.06595099e+00 5.95497429e-01
1.80250660e-01 -1.19606411e+00 -1.05007946e+00 4.71239656e-01
1.26933706e+00 -6.54043496e-01 9.48115051e-01 -3.12380821e-01
4.04124290e-01 -3.93135488e-01 -1.67890593e-01 -1.25297809e+00
3.05584252e-01 -5.02627969e-01 -4.88312960e-01 6.69250309e-01
-4.07246202e-01 -1.72424912e-01 9.70152795e-01 3.82570863e-01
8.40271190e-02 -5.48889637e-01 -1.19798183e+00 -1.19302177e+00
-2.73647338e-01 -3.11871886e-01 1.92952201e-01 5.27261436e-01
-9.70281586e-02 -2.12709215e-02 -5.46604514e-01 -9.25564170e-02
6.64355755e-01 9.13254991e-02 1.61137724e+00 -1.30422199e+00
4.82596233e-02 -1.00957677e-01 -8.91175449e-01 -1.18554783e+00
4.30354446e-01 -4.31491435e-01 2.47084156e-01 -1.38490629e+00
6.71264576e-03 -3.85104343e-02 4.62738633e-01 1.09439775e-01
3.12009871e-01 5.80269136e-02 4.73314583e-01 5.08353889e-01
-4.28350687e-01 5.60220003e-01 1.74019468e+00 1.38385311e-01
-9.24236327e-02 2.79938817e-01 8.27109039e-01 8.65108430e-01
4.32346106e-01 -5.55851460e-01 -3.29223990e-01 -2.42827773e-01
-1.83106542e-01 4.09611106e-01 7.42519736e-01 -1.14925659e+00
6.10585958e-02 7.36437961e-02 6.48022175e-01 -9.06715810e-01
8.25688422e-01 -1.50066447e+00 5.80094218e-01 8.41616988e-01
-1.17592931e-01 1.31928012e-01 1.70243025e-01 3.91384333e-01
-9.84048285e-03 -3.41449112e-01 3.21296304e-01 -4.82668370e-01
-8.52036834e-01 3.83630395e-02 -2.19192609e-01 -3.72406811e-01
1.24690092e+00 -7.28390217e-01 2.25775436e-01 -2.20738158e-01
-8.62657309e-01 3.30921710e-02 6.83137953e-01 7.80618906e-01
7.82645643e-01 -1.23670137e+00 -4.17982548e-01 1.91893965e-01
1.24988273e-01 2.48623952e-01 6.29813003e-04 1.02101672e+00
-7.23015010e-01 2.84835190e-01 -7.06106544e-01 -1.18981791e+00
-1.53906953e+00 4.86782223e-01 4.02441204e-01 2.50419885e-01
-9.43930209e-01 5.82456410e-01 -2.52271026e-01 -7.03922451e-01
6.95950031e-01 -4.54451829e-01 2.07146425e-02 -1.65256739e-01
5.60502186e-02 7.90928841e-01 2.15421636e-02 -6.73672199e-01
-3.12634349e-01 9.18768466e-01 3.04055154e-01 -3.10091227e-02
1.32662261e+00 1.62013188e-01 9.45590660e-02 8.57485116e-01
1.08533335e+00 -2.12692246e-01 -1.74631870e+00 -1.13191847e-02
-1.93552762e-01 -1.02662075e+00 -5.04541934e-01 -5.35666287e-01
-8.91014397e-01 9.95676160e-01 8.85929585e-01 -3.12873125e-01
5.10988772e-01 2.07252577e-01 2.53845543e-01 9.36882973e-01
1.02385032e+00 -1.10241854e+00 6.30325258e-01 4.13089842e-01
1.78399074e+00 -1.50485408e+00 2.44647145e-01 -5.32879233e-01
-4.76215929e-01 1.14792633e+00 7.29604125e-01 -3.10524553e-01
3.70424449e-01 3.16888422e-01 -1.42856268e-02 -2.47732192e-01
6.09177798e-02 2.26242617e-01 2.91776061e-01 8.15763175e-01
1.78928271e-01 1.04012173e-02 -5.23751043e-02 -7.01669455e-02
-1.17073022e-01 8.56433362e-02 2.92859942e-01 1.50476670e+00
4.66294102e-02 -9.03039813e-01 -5.83500922e-01 4.76705208e-02
-2.52229840e-01 8.73225868e-01 -5.30142784e-01 1.56753933e+00
8.41718465e-02 5.35931110e-01 -2.85311252e-01 -4.33725789e-02
9.20028746e-01 -3.23355123e-02 1.59495878e+00 -6.64861321e-01
-5.25561810e-01 -6.79430589e-02 -2.07494706e-01 -8.22877228e-01
-6.81197584e-01 -9.23608243e-01 -8.38469267e-01 2.31554359e-01
-6.39793932e-01 -3.97709161e-01 1.12005889e+00 8.28185201e-01
-2.08341882e-01 4.37746763e-01 2.06967950e-01 -1.99054837e+00
-7.75885165e-01 -1.28854561e+00 -8.66836548e-01 4.86639827e-01
4.12795603e-01 -1.46864629e+00 -3.07142556e-01 -2.58643087e-02]
|
[6.371726989746094, -0.9351359605789185]
|
b352813b-baf1-4916-8c31-1f92538aca1c
|
automatic-ischemic-stroke-lesion-segmentation
|
2007.03294
| null |
https://arxiv.org/abs/2007.03294v1
|
https://arxiv.org/pdf/2007.03294v1.pdf
|
Automatic Ischemic Stroke Lesion Segmentation from Computed Tomography Perfusion Images by Image Synthesis and Attention-Based Deep Neural Networks
|
Ischemic stroke lesion segmentation from Computed Tomography Perfusion (CTP) images is important for accurate diagnosis of stroke in acute care units. However, it is challenged by low image contrast and resolution of the perfusion parameter maps, in addition to the complex appearance of the lesion. To deal with this problem, we propose a novel framework based on synthesized pseudo Diffusion-Weighted Imaging (DWI) from perfusion parameter maps to obtain better image quality for more accurate segmentation. Our framework consists of three components based on Convolutional Neural Networks (CNNs) and is trained end-to-end. First, a feature extractor is used to obtain both a low-level and high-level compact representation of the raw spatiotemporal Computed Tomography Angiography (CTA) images. Second, a pseudo DWI generator takes as input the concatenation of CTP perfusion parameter maps and our extracted features to obtain the synthesized pseudo DWI. To achieve better synthesis quality, we propose a hybrid loss function that pays more attention to lesion regions and encourages high-level contextual consistency. Finally, we segment the lesion region from the synthesized pseudo DWI, where the segmentation network is based on switchable normalization and channel calibration for better performance. Experimental results showed that our framework achieved the top performance on ISLES 2018 challenge and: 1) our method using synthesized pseudo DWI outperformed methods segmenting the lesion from perfusion parameter maps directly; 2) the feature extractor exploiting additional spatiotemporal CTA images led to better synthesized pseudo DWI quality and higher segmentation accuracy; and 3) the proposed loss functions and network structure improved the pseudo DWI synthesis and lesion segmentation performance.
|
['Ning Huang', 'Tao Song', 'Mei Cui', 'Guotai Wang', 'Qiang Dong', 'Shaoting Zhang']
|
2020-07-07
| null | null | null | null |
['ischemic-stroke-lesion-segmentation']
|
['medical']
|
[ 2.68301070e-01 -2.73687363e-01 -2.60892451e-01 -5.20894229e-01
-1.17794359e+00 -4.95621204e-01 3.42282653e-01 4.29374166e-02
-6.83974504e-01 7.19832599e-01 4.54452366e-01 -2.49679849e-01
-1.41682059e-01 -8.08320701e-01 -5.01452088e-01 -8.20835233e-01
-2.69090474e-01 4.46959645e-01 6.60130143e-01 1.78805977e-01
1.06223606e-01 7.80133247e-01 -7.07029581e-01 5.06703794e-01
1.06443214e+00 1.11546016e+00 4.61905867e-01 6.86579943e-01
-1.72248453e-01 8.65855873e-01 -3.43222201e-01 1.38988867e-01
5.97150862e-01 -7.84689069e-01 -8.31496418e-01 -1.13995217e-01
3.12549025e-01 -7.27755427e-01 -6.90540671e-01 9.58607018e-01
8.91776562e-01 -1.13506243e-01 7.58784771e-01 -9.39019978e-01
-2.46129826e-01 6.50547266e-01 -4.40970600e-01 9.14042830e-01
-3.57793629e-01 6.36881649e-01 4.44119751e-01 -7.33898520e-01
7.32560456e-01 9.26474214e-01 3.98248374e-01 2.72840708e-01
-1.08663774e+00 -5.98666012e-01 2.18585283e-02 6.68818116e-01
-1.11841619e+00 5.61213978e-02 6.57282770e-01 -4.32914108e-01
6.84981823e-01 1.27131805e-01 9.85089302e-01 1.02486169e+00
1.85782820e-01 7.73072600e-01 1.09618926e+00 -1.37990311e-01
7.32319728e-02 -3.74003589e-01 1.03080466e-01 6.42177045e-01
1.18451580e-01 1.17565893e-01 1.73257589e-01 1.46886751e-01
1.20451903e+00 1.11916326e-01 -8.00684154e-01 -4.10496622e-01
-1.49535739e+00 7.05349624e-01 9.32239175e-01 3.51655811e-01
-6.03139281e-01 4.22997586e-02 5.91955841e-01 -2.42090728e-02
-1.22888554e-02 1.89085126e-01 -1.90518871e-01 -1.84560180e-01
-1.01630783e+00 1.43355221e-01 2.83530146e-01 5.65374076e-01
2.54073858e-01 -7.77030066e-02 -7.88749933e-01 8.05972695e-01
-1.19873866e-01 5.41694939e-01 7.39390850e-01 -9.00010109e-01
8.17652524e-01 7.08252728e-01 -9.93304253e-02 -5.61597228e-01
-8.09228599e-01 -6.13501430e-01 -1.07973289e+00 2.46141195e-01
7.56265581e-01 -1.84132278e-01 -1.21704710e+00 1.56869543e+00
9.23274842e-04 2.80907214e-01 -2.06449345e-01 1.41310656e+00
8.92573357e-01 4.41302687e-01 2.47613907e-01 -3.22898254e-02
1.37971950e+00 -1.05775499e+00 -4.24711049e-01 -1.31538883e-01
8.30324352e-01 -3.47993135e-01 1.17116725e+00 -2.83955038e-02
-1.19840026e+00 -3.26250076e-01 -1.09167480e+00 3.58595625e-02
3.73580819e-03 3.04991841e-01 -1.63891856e-02 4.52956766e-01
-8.32176983e-01 5.06209493e-01 -1.07889402e+00 -3.02682109e-02
1.00656259e+00 1.79174393e-01 -3.78147334e-01 -2.75297046e-01
-1.15547884e+00 1.15791202e+00 4.21766847e-01 -3.46372202e-02
-8.25101495e-01 -1.18433523e+00 -7.02900350e-01 1.02723595e-02
3.05900514e-01 -7.94423223e-01 8.40637982e-01 -8.13174367e-01
-1.56069112e+00 6.41639471e-01 6.72662333e-02 -5.46408057e-01
1.06183994e+00 8.25900286e-02 -1.29016250e-01 9.45770383e-01
6.91501647e-02 7.01951921e-01 6.99190974e-01 -8.80017579e-01
-6.41377985e-01 -3.88898492e-01 -2.54491180e-01 4.33393940e-02
2.23052986e-02 -1.83241770e-01 -3.00897330e-01 -7.35456765e-01
3.32496583e-01 -7.03718185e-01 -3.38357627e-01 3.54965776e-01
-2.85902500e-01 3.45882654e-01 7.85954595e-01 -8.80558372e-01
1.00089610e+00 -1.92568588e+00 2.23841313e-02 4.56952989e-01
5.58034241e-01 4.99214828e-01 -2.90675253e-01 -3.61350745e-01
-3.35918874e-01 7.56964982e-02 -7.24966407e-01 2.79252887e-01
-3.92416060e-01 1.10410511e-01 8.50938167e-03 5.43930650e-01
2.29056001e-01 1.29902554e+00 -7.90060163e-01 -6.43087745e-01
3.86367053e-01 5.28718293e-01 -5.60595691e-01 3.23524773e-01
2.94541419e-01 1.01875341e+00 -4.52055514e-01 2.06579879e-01
7.65373349e-01 -1.09469913e-01 -1.29014388e-01 -6.13749146e-01
-1.44101217e-01 -7.89871812e-02 -9.82647240e-01 1.77198064e+00
-4.24611270e-01 3.52862000e-01 -7.40338489e-02 -1.34144306e+00
7.23784447e-01 3.55753869e-01 9.54139113e-01 -1.07246399e+00
4.89536375e-01 3.81895483e-01 4.14193451e-01 -8.25063527e-01
-3.74280244e-01 4.44035009e-02 3.00917953e-01 4.77281094e-01
-2.35852629e-01 -3.09243381e-01 2.81622350e-01 1.03930905e-01
1.12158275e+00 -4.53915931e-02 -7.78724924e-02 -3.07774425e-01
6.90133870e-01 1.42355356e-02 5.20351112e-01 7.61427522e-01
-6.37755990e-01 1.07367682e+00 7.52167344e-01 -6.09824538e-01
-1.21483779e+00 -1.24940634e+00 -4.91546392e-01 1.88197061e-01
2.09859490e-01 4.31040227e-02 -8.85763526e-01 -7.69275308e-01
-2.99355686e-01 2.64111042e-01 -6.73489630e-01 -1.38719887e-01
-1.22435904e+00 -9.87394333e-01 6.04954422e-01 9.73957837e-01
1.08372998e+00 -8.74959767e-01 -1.05747008e+00 4.41485077e-01
-6.47027135e-01 -1.21842170e+00 -7.98415780e-01 2.31295824e-04
-1.15109730e+00 -1.19167924e+00 -1.37974799e+00 -8.21154773e-01
6.05990291e-01 1.04401723e-01 7.00858474e-01 1.02236614e-01
-7.61300325e-01 -8.25675353e-02 -3.93073887e-01 -8.39809328e-02
-3.15802038e-01 4.96451370e-02 -6.44682586e-01 9.13954824e-02
-2.45317027e-01 -6.82865262e-01 -1.28348887e+00 2.78342724e-01
-1.12295604e+00 1.35226324e-01 8.92706692e-01 9.25648093e-01
6.65188015e-01 -2.81309217e-01 7.11190581e-01 -6.58038378e-01
5.22109985e-01 -3.57171297e-01 -3.32269698e-01 3.38628113e-01
-3.79612386e-01 2.47010663e-02 7.05241680e-01 -4.27903473e-01
-9.49211478e-01 1.11449316e-01 -2.22416207e-01 -1.36360601e-01
-9.32092518e-02 3.59861553e-01 1.01966113e-02 -7.09422976e-02
9.12563443e-01 3.60680640e-01 3.10135394e-01 -2.31235310e-01
3.66782665e-01 5.70404530e-01 8.01307321e-01 -5.36541402e-01
3.59064966e-01 7.19029307e-01 1.51585326e-01 -3.77282619e-01
-4.31794018e-01 -1.90724686e-01 -1.02210224e+00 -1.08126879e-01
9.69547570e-01 -5.40572524e-01 -3.05480331e-01 4.89253670e-01
-9.91651893e-01 -5.16349971e-01 -4.90541071e-01 7.88532376e-01
-4.26297367e-01 5.47813356e-01 -7.06303537e-01 8.26689750e-02
-7.48325586e-01 -1.85325360e+00 6.70023739e-01 1.12264611e-01
1.80894703e-01 -7.66888142e-01 -1.35871008e-01 4.58556972e-02
8.88527274e-01 6.54877663e-01 1.45589352e+00 -2.81980425e-01
-5.45299530e-01 -2.78414428e-01 -8.68790805e-01 5.15962243e-01
1.49632722e-01 -2.31678680e-01 -5.31960607e-01 -9.66482162e-02
-1.50202960e-01 -3.69047001e-03 9.14332211e-01 8.90103936e-01
1.29071403e+00 1.08054496e-01 -1.22662023e-01 9.71316159e-01
1.29302561e+00 1.35605514e-01 7.28338063e-01 4.00688738e-01
6.55684948e-01 3.66125762e-01 7.01566935e-02 3.97773087e-01
3.94963235e-01 6.55623913e-01 3.52663130e-01 -5.14858007e-01
-9.10196483e-01 1.78035587e-01 -7.79098272e-02 4.32357013e-01
-1.69801623e-01 1.98581070e-01 -1.00023675e+00 5.27001858e-01
-1.72836149e+00 -6.74769878e-01 -3.09039116e-01 2.08645129e+00
7.71876335e-01 3.33793387e-02 3.85357589e-01 3.94842178e-02
6.11703873e-01 -5.82715385e-02 -7.86668241e-01 7.48791322e-02
-3.26228619e-01 5.70311725e-01 6.06182396e-01 4.03068841e-01
-1.03382695e+00 5.49259484e-01 5.02341747e+00 6.55385673e-01
-1.57551503e+00 3.91453087e-01 7.02395320e-01 -1.29054159e-01
5.08850329e-02 -2.95283377e-01 -1.62445039e-01 5.57921767e-01
4.58114713e-01 -9.63824019e-02 9.28122252e-02 3.81567925e-01
4.42239761e-01 -1.21757649e-01 -8.12022388e-01 9.32685077e-01
-2.47024685e-01 -1.38744473e+00 2.37607747e-01 -2.58896291e-01
6.36902332e-01 3.68058920e-01 -1.10234402e-01 6.13587797e-02
-7.75745511e-02 -9.53258634e-01 7.90775478e-01 4.83203739e-01
9.76650417e-01 -6.20941281e-01 7.62338400e-01 2.29952902e-01
-1.06687284e+00 -1.56266063e-01 -2.36545756e-01 5.31490982e-01
3.34484816e-01 4.89898026e-01 -6.83082461e-01 5.55573821e-01
6.08220577e-01 7.42531240e-01 -6.45966351e-01 1.55483174e+00
-4.00063038e-01 5.73231399e-01 -2.46419907e-01 4.51011717e-01
4.88076746e-01 -2.30529934e-01 5.60064971e-01 1.36773133e+00
1.80488139e-01 3.90426546e-01 3.04762274e-01 9.82876003e-01
8.06044340e-02 3.22183996e-01 -6.24082275e-02 7.46056378e-01
1.77969471e-01 1.24790180e+00 -9.45705414e-01 -7.55718589e-01
-2.88467735e-01 7.95614064e-01 1.95139155e-01 5.78327298e-01
-9.54209507e-01 -3.57915789e-01 9.94070768e-02 3.20930243e-01
7.70416781e-02 -1.94535092e-01 -6.75022304e-01 -1.32775283e+00
6.83775917e-02 -5.36096454e-01 4.83128369e-01 -3.92698646e-01
-1.04007459e+00 1.01381040e+00 9.58270952e-02 -1.26648641e+00
2.34956713e-03 -3.36372942e-01 -8.69344056e-01 1.13905311e+00
-1.95070958e+00 -8.11742783e-01 -8.80966544e-01 7.87460566e-01
3.38235766e-01 1.25531912e-01 5.99259734e-01 5.55044472e-01
-6.85824156e-01 4.51901913e-01 -2.54732788e-01 2.98863322e-01
7.21111774e-01 -1.15612543e+00 1.43829152e-01 9.81761336e-01
-4.38465983e-01 5.01469187e-02 -6.34877160e-02 -5.26372850e-01
-6.54579639e-01 -1.36184084e+00 3.78404051e-01 -2.20497683e-01
3.63962889e-01 2.37383798e-01 -9.63268280e-01 3.33141357e-01
-6.90630078e-02 6.24072552e-01 1.88331932e-01 -1.13168621e+00
-1.44500390e-01 -1.01558842e-01 -1.35650289e+00 5.28480530e-01
8.56504321e-01 -1.75818637e-01 -5.10607481e-01 3.55833560e-01
1.85956314e-01 -5.86573303e-01 -8.77342820e-01 6.37161851e-01
5.63778579e-01 -5.17123103e-01 1.27654934e+00 -5.69813073e-01
6.02022469e-01 -2.19757706e-01 2.36884594e-01 -1.25393176e+00
-4.29260910e-01 -6.05421290e-02 3.99751961e-01 3.98932368e-01
3.30303490e-01 -7.10419178e-01 7.12518692e-01 6.92089081e-01
-4.91330862e-01 -1.02728295e+00 -1.03513169e+00 -7.32877254e-01
6.16446376e-01 -2.98512459e-01 4.81150150e-01 6.46444917e-01
-2.48217300e-01 -2.02920914e-01 1.39073774e-01 5.69018498e-02
7.77338386e-01 1.24273442e-01 1.95035949e-01 -7.91789532e-01
4.76936363e-02 -9.26779151e-01 -2.68504530e-01 -1.03794646e+00
-1.54686481e-01 -1.51079309e+00 -1.15419254e-01 -1.72530937e+00
2.33764738e-01 -5.94566941e-01 -4.66491163e-01 3.78215402e-01
-3.16713363e-01 5.48552334e-01 2.14429572e-01 2.62427509e-01
-3.78134102e-02 2.60400474e-01 2.11253095e+00 -2.29769707e-01
-4.06451315e-01 -1.04585990e-01 -2.56003439e-01 6.50145173e-01
8.22864652e-01 -5.55259764e-01 -3.49422306e-01 -4.99283940e-01
-6.53575897e-01 3.38237762e-01 7.12794960e-01 -9.62841213e-01
2.18620971e-01 9.40035954e-02 5.92744350e-01 -2.85545826e-01
-6.56918138e-02 -7.26060152e-01 -4.05927092e-01 8.43386531e-01
-3.33481967e-01 -1.34182096e-01 -2.43131630e-03 -2.10735556e-02
-2.27073520e-01 -1.22783586e-01 1.40578008e+00 -2.08692357e-01
-4.57577169e-01 7.75166392e-01 -4.66451168e-01 3.87985080e-01
1.04718113e+00 -1.58983514e-01 -2.50407100e-01 -1.16846547e-01
-9.26775932e-01 2.21808985e-01 -1.07781194e-01 2.60810524e-01
8.37045848e-01 -1.20169103e+00 -1.05974889e+00 3.74574512e-01
-1.09048113e-01 2.96455976e-02 5.26161492e-01 1.65456462e+00
-1.12041366e+00 4.20788944e-01 -6.72946632e-01 -8.52002740e-01
-7.98676372e-01 1.09306529e-01 7.11211085e-01 -4.49204832e-01
-1.41732955e+00 6.53680682e-01 1.75247148e-01 6.17325418e-02
3.19837242e-01 -9.49727893e-01 -1.24486707e-01 -1.22843191e-01
7.13178575e-01 2.51016200e-01 2.73068279e-01 -7.47055352e-01
-3.73616397e-01 7.70740390e-01 -1.18729904e-01 -1.81176171e-01
1.27464426e+00 -5.73727004e-02 2.07288116e-01 -5.16458273e-01
1.47281301e+00 -5.74616373e-01 -1.62905979e+00 -4.97308165e-01
-2.18192384e-01 -5.83970010e-01 3.94730389e-01 -1.20190668e+00
-1.59105849e+00 1.23752141e+00 1.07206130e+00 -3.97856116e-01
1.12028372e+00 -3.27074856e-01 1.38643909e+00 -1.52141213e-01
8.60115439e-02 -5.86970627e-01 9.61578041e-02 1.27119541e-01
9.71147716e-01 -1.10061789e+00 -2.75951385e-01 -3.29964966e-01
-7.28731275e-01 1.50134623e+00 4.23445523e-01 -3.18989694e-01
7.45734870e-01 3.59941900e-01 3.77714336e-01 -1.17487267e-01
1.96889862e-02 -1.21481255e-01 3.16439539e-01 5.68326592e-01
-1.64246671e-02 8.22935924e-02 -3.81048143e-01 8.06599796e-01
5.16303293e-02 2.29062006e-01 3.76501530e-01 7.93934166e-01
-5.12109697e-01 -1.13691854e+00 -1.26307845e-01 6.06833279e-01
-3.34632397e-01 -9.95810777e-02 1.62335500e-01 5.86610317e-01
2.66344607e-01 5.38964152e-01 -2.97799975e-01 9.86154899e-02
7.25813270e-01 -8.50619301e-02 6.50113165e-01 -3.71830702e-01
-6.32542312e-01 4.72239917e-03 -3.07871878e-01 -7.20759988e-01
-1.50127336e-01 -4.41289902e-01 -1.64861071e+00 1.56327337e-01
1.46220967e-01 -8.51307586e-02 6.55418813e-01 9.65860486e-01
2.09848255e-01 8.10193658e-01 6.41147017e-01 -7.07236946e-01
-3.58689249e-01 -8.88269603e-01 -4.98668373e-01 6.65023625e-01
3.12465727e-01 -5.62810063e-01 -4.77818623e-02 1.16681360e-01]
|
[14.374393463134766, -2.1077678203582764]
|
53fd165b-04d8-486d-9372-af26515a5918
|
necessary-and-sufficient-polynomial
|
1912.11987
| null |
https://arxiv.org/abs/1912.11987v1
|
https://arxiv.org/pdf/1912.11987v1.pdf
|
Necessary and Sufficient Polynomial Constraints on Compatible Triplets of Essential Matrices
|
The essential matrix incorporates relative rotation and translation parameters of two calibrated cameras. The well-known algebraic characterization of essential matrices, i.e. necessary and sufficient conditions under which an arbitrary matrix (of rank two) becomes essential, consists of a unique matrix equation of degree three. Based on this equation, a number of efficient algorithmic solutions to different relative pose estimation problems have been proposed. In three views, a possible way to describe the geometry of three calibrated cameras comes from considering compatible triplets of essential matrices. The compatibility is meant the correspondence of a triplet to a certain configuration of calibrated cameras. The main goal of this paper is to give an algebraic characterization of compatible triplets of essential matrices. Specifically, we propose necessary and sufficient polynomial constraints on a triplet of real rank-two essential matrices that ensure its compatibility. The constraints are given in the form of six cubic matrix equations, one quartic and one sextic scalar equations. An important advantage of the proposed constraints is their sufficiency even in the case of cameras with collinear centers. The applications of the constraints may include relative camera pose estimation in three and more views, averaging of essential matrices for incremental structure from motion, multiview camera auto-calibration, etc.
|
['E. V. Martyushev']
|
2019-12-15
| null | null | null | null |
['camera-auto-calibration']
|
['computer-vision']
|
[ 5.31776026e-02 -5.59622757e-02 4.94937040e-02 -3.06783170e-01
1.59819354e-03 -7.64550626e-01 6.21360481e-01 -1.58703998e-01
-2.04483867e-01 4.12790805e-01 -1.16196252e-01 -2.19219342e-01
-4.21790481e-01 -8.65688026e-02 -5.97753227e-01 -6.83344483e-01
2.66693980e-01 6.05355442e-01 -1.19756728e-01 -4.73293275e-01
3.93600225e-01 8.69158030e-01 -1.41508901e+00 -5.32158494e-01
7.15832412e-01 7.15854883e-01 1.45200789e-01 6.80672944e-01
2.88624793e-01 3.28924000e-01 -1.03548042e-01 -4.04522181e-01
5.62162340e-01 -4.13459688e-01 -5.90626895e-01 6.94413424e-01
5.92592299e-01 -5.01820333e-02 6.48598894e-02 1.14153910e+00
-1.26957698e-02 -1.49859443e-01 8.77625465e-01 -1.55726433e+00
-4.65082884e-01 9.27157775e-02 -5.73505580e-01 -1.62106380e-01
7.58636713e-01 -2.53864408e-01 9.63546991e-01 -1.11929560e+00
7.99708784e-01 1.06879056e+00 3.27706248e-01 1.80195957e-01
-1.14992547e+00 -8.07065070e-02 -2.41636306e-01 2.43593156e-01
-1.55665064e+00 -4.49490577e-01 7.43371546e-01 -7.13401198e-01
3.49353373e-01 6.39651418e-01 7.89180219e-01 5.38888335e-01
2.91135967e-01 -4.85432856e-02 1.06499898e+00 -7.52107680e-01
-7.23779052e-02 4.22147512e-01 4.90307719e-01 6.44062817e-01
7.86686778e-01 -3.45555395e-01 -1.19088724e-01 -2.00288340e-01
1.00340414e+00 -4.16744426e-02 -3.13175380e-01 -1.11516666e+00
-1.39232874e+00 6.57818139e-01 -1.36480600e-01 3.84200245e-01
-6.72385246e-02 -2.92432457e-01 -7.32774362e-02 7.52326101e-02
-2.14239106e-01 3.34318757e-01 -2.50145979e-02 3.81012022e-01
-2.72393465e-01 -1.04113845e-02 8.95337343e-01 1.34245467e+00
9.96967375e-01 1.05395308e-02 6.61306143e-01 3.56188267e-01
2.79101223e-01 9.30000842e-01 1.12252071e-01 -1.06284201e+00
5.01184940e-01 7.58543074e-01 1.28530785e-01 -1.36112750e+00
-5.05535364e-01 -1.18893094e-01 -9.03783262e-01 4.38571051e-02
4.45261538e-01 2.01946393e-01 -1.85427681e-01 1.53861129e+00
2.81675637e-01 -2.47511864e-01 5.91194220e-02 9.61239219e-01
3.06488782e-01 2.81825036e-01 -7.84491360e-01 -6.01748049e-01
1.33098674e+00 -3.07345569e-01 -6.24542654e-01 1.38517007e-01
1.64750040e-01 -1.23394489e+00 7.34135449e-01 5.09197295e-01
-9.42237496e-01 -4.41588253e-01 -1.18186581e+00 4.97953035e-02
-6.86133057e-02 5.69577336e-01 2.98414469e-01 4.63874429e-01
-1.10509801e+00 3.60011160e-01 -2.78642446e-01 -4.60446984e-01
-8.76658738e-01 5.83810210e-01 -9.52547789e-01 3.18842560e-01
-7.23382413e-01 1.14411438e+00 1.64744049e-01 5.05423486e-01
-1.98451668e-01 -1.91653505e-01 -8.18344533e-01 -1.94359675e-01
3.68440032e-01 -7.30923355e-01 8.73565495e-01 -1.01254880e+00
-1.20908666e+00 1.14823413e+00 -8.92160162e-02 1.09432740e-02
7.47637749e-01 5.69244698e-02 -3.84365320e-01 3.13069791e-01
6.43610582e-02 1.78830951e-01 9.54210877e-01 -1.31482959e+00
-1.43311992e-01 -4.59921420e-01 3.45980525e-01 3.63481462e-01
-1.86644658e-01 -1.75388437e-02 -5.62315047e-01 -1.99828088e-01
8.84670138e-01 -1.36939657e+00 -3.62244755e-01 -5.61320931e-02
-5.74290633e-01 1.38502359e-01 7.30792701e-01 -5.32157183e-01
7.77732074e-01 -1.95873284e+00 8.10676873e-01 4.34000969e-01
1.88767657e-01 -2.05510959e-01 1.55284002e-01 4.45320815e-01
-4.81020778e-01 -2.43873373e-01 -5.89218512e-02 4.23693843e-02
-1.99368805e-01 1.48735508e-01 -1.00134999e-01 9.64441717e-01
1.26228984e-02 4.83660065e-02 -3.64573687e-01 -6.17373645e-01
5.83400488e-01 6.03788495e-01 -3.63851994e-01 9.32037756e-02
2.11448729e-01 6.05482936e-01 -3.55281472e-01 2.52303392e-01
8.59210849e-01 1.69608846e-01 2.22716272e-01 -7.51161873e-01
-5.62478483e-01 -1.65912464e-01 -2.06276417e+00 9.83789146e-01
-1.00156501e-01 2.77133524e-01 4.64538574e-01 -8.54357004e-01
9.32088196e-01 3.35834593e-01 6.77622139e-01 1.02006249e-01
5.30186832e-01 2.26603568e-01 3.63496654e-02 -4.10015762e-01
6.18187129e-01 1.26877859e-01 9.39830989e-02 3.58306020e-01
-2.71146670e-02 -2.58799851e-01 5.96466660e-01 2.85373926e-01
2.02256426e-01 -5.30820824e-02 6.10371768e-01 -6.27968192e-01
1.28467691e+00 -3.85741085e-01 5.90568185e-01 -1.44986704e-01
2.00363249e-01 6.46868527e-01 8.68225694e-01 -5.26045561e-01
-1.38316894e+00 -6.67373359e-01 -1.55185878e-01 -1.42518496e-02
3.51742029e-01 -3.00053656e-01 -7.16550708e-01 3.54635604e-02
-1.47111118e-01 5.77963926e-02 -4.34770167e-01 8.36785808e-02
-7.15233207e-01 -2.01919183e-01 -2.35846460e-01 -6.93523064e-02
5.39805114e-01 -7.55302459e-02 -6.42876863e-01 -3.46024424e-01
-3.03024590e-01 -1.32427490e+00 -4.77481425e-01 -1.15134098e-01
-1.01257896e+00 -1.53605545e+00 -6.65314257e-01 -5.55479944e-01
1.21757197e+00 8.20651770e-01 7.53587186e-01 -4.87962700e-02
-1.40817165e-01 7.66395092e-01 -3.95206884e-02 1.17277540e-01
-4.82105196e-01 -3.07788879e-01 7.84628212e-01 3.93792838e-01
-1.40029445e-01 -4.82067525e-01 -2.66760051e-01 9.40619707e-01
-9.24518645e-01 5.54379448e-02 2.79710948e-01 4.77188915e-01
7.09055603e-01 -3.52089643e-01 -2.02656537e-01 -4.70853060e-01
2.45166510e-01 -3.91891338e-02 -1.09436584e+00 5.35101414e-01
-2.42623791e-01 2.88076907e-01 6.14196181e-01 -3.58781278e-01
-7.76365519e-01 6.13975763e-01 2.02495694e-01 -3.84640336e-01
1.62198871e-01 3.27918768e-01 -6.10135853e-01 -5.01952887e-01
5.09848475e-01 7.26253986e-02 -6.50622994e-02 -4.49099958e-01
3.80690217e-01 4.03287888e-01 6.23486996e-01 -6.26989782e-01
1.13976991e+00 4.40754324e-01 7.77854443e-01 -1.36376321e+00
-3.49738359e-01 -7.72523284e-01 -1.40197659e+00 -4.77035403e-01
7.25947976e-01 -7.30075479e-01 -9.59333777e-01 2.73970544e-01
-1.18374908e+00 5.50279140e-01 -3.43303941e-02 7.37419486e-01
-8.28899384e-01 1.02305973e+00 -2.34082967e-01 -6.50314808e-01
-1.11852717e-02 -1.45107841e+00 7.04856694e-01 1.52183948e-02
-1.53768718e-01 -9.50564384e-01 1.03459977e-01 1.91758007e-01
-1.88547820e-01 2.59271324e-01 5.51772058e-01 -2.21742899e-03
-8.08965981e-01 -3.02115619e-01 1.44067332e-01 3.17628056e-01
7.60178789e-02 6.50818229e-01 -4.84951288e-01 -2.68112183e-01
6.12652600e-01 4.11226839e-01 3.24824840e-01 4.02978212e-01
2.61327952e-01 -3.37385476e-01 -1.89037427e-01 8.35213125e-01
1.62212455e+00 2.22217754e-01 3.00935447e-01 2.37761930e-01
7.74608135e-01 7.64694989e-01 6.36770368e-01 3.53813767e-01
1.70362070e-01 9.77739811e-01 6.40758336e-01 1.93075612e-02
5.81687033e-01 1.98893905e-01 2.91554958e-01 1.17121720e+00
-5.05435407e-01 3.29933167e-01 -6.94037676e-01 2.12191984e-01
-1.58158815e+00 -8.70201051e-01 -8.55080903e-01 2.73818684e+00
8.47310349e-02 -9.71462354e-02 1.96715117e-01 3.80670637e-01
1.07198358e+00 -1.48939148e-01 -2.55060017e-01 -4.55761701e-01
-5.16703129e-01 -6.13286674e-01 6.24231696e-01 9.93254542e-01
-8.57906222e-01 2.89165050e-01 6.34612083e+00 2.78891828e-02
-9.07033086e-01 -2.29933038e-01 -1.08354837e-01 5.18084764e-01
-5.11354446e-01 5.28045535e-01 -1.01407266e+00 1.59473389e-01
3.34778309e-01 -2.34328359e-01 1.20439537e-01 7.63175368e-01
1.13127306e-02 -4.37470913e-01 -1.03075528e+00 1.27793288e+00
4.10752118e-01 -1.15221572e+00 2.42813438e-01 2.95981228e-01
7.43275166e-01 -6.08425856e-01 1.74465105e-01 -5.77203035e-01
-3.00315052e-01 -3.35867643e-01 8.59068155e-01 4.45634305e-01
6.60440326e-01 -7.13924468e-01 5.54343998e-01 3.15652072e-01
-1.28829527e+00 2.00889707e-01 -5.42175055e-01 -1.54423729e-01
3.34118217e-01 4.05563086e-01 -4.57246095e-01 7.90120780e-01
8.63533914e-02 7.20547616e-01 -6.11929357e-01 9.08552170e-01
-4.19792943e-02 -3.27247828e-01 -4.23889458e-01 3.09090674e-01
-2.76452340e-02 -1.29649568e+00 7.22181439e-01 7.15496421e-01
5.90205014e-01 3.96981239e-01 -1.81001455e-01 5.73298454e-01
4.51890945e-01 3.80174279e-01 -9.12548959e-01 3.80937934e-01
1.23217300e-01 1.47561097e+00 -8.04435372e-01 -1.74075320e-01
-4.91161764e-01 8.63528311e-01 -1.42415330e-01 1.39632806e-01
-5.89968324e-01 -7.92108700e-02 5.89806139e-01 1.74688086e-01
-1.34941503e-01 -6.54532731e-01 -1.37199551e-01 -1.61208844e+00
2.86323875e-01 -8.28361332e-01 2.29014680e-01 -8.41669559e-01
-5.05530059e-01 6.39273167e-01 2.84801960e-01 -1.67478693e+00
-3.78676772e-01 -9.15710211e-01 -4.64995146e-01 7.72213042e-01
-7.47569740e-01 -1.14924097e+00 -3.41091007e-01 1.03231192e+00
1.76170334e-01 -2.10425526e-01 6.66028082e-01 1.39759600e-01
-6.26678646e-01 7.07300603e-02 1.52732104e-01 -3.06975663e-01
6.19705617e-01 -1.19568777e+00 -3.89158539e-02 1.17574847e+00
6.75270110e-02 9.05135989e-01 1.10080397e+00 -3.25884998e-01
-1.83539176e+00 -2.56395608e-01 1.09126759e+00 -5.18324733e-01
6.57645762e-01 -1.95659801e-01 -4.94667739e-01 1.07092071e+00
6.99515268e-02 -4.30375606e-01 2.96753138e-01 -1.73136041e-01
-2.04288915e-01 -2.36217141e-01 -6.25822961e-01 5.70377886e-01
5.58576822e-01 -4.45714414e-01 -3.92780125e-01 5.23509026e-01
2.85360038e-01 -4.01952922e-01 -7.99630105e-01 2.53617048e-01
5.47618628e-01 -1.15025687e+00 1.10996282e+00 -1.20263182e-01
-6.52104393e-02 -6.63005352e-01 -3.19706857e-01 -9.20383632e-01
-1.60575032e-01 -8.23196352e-01 5.49493194e-01 1.07913780e+00
-1.04405329e-01 -7.47233033e-01 4.28852379e-01 6.42527461e-01
3.72454785e-02 -1.96587101e-01 -7.46532321e-01 -9.55674767e-01
-3.61785382e-01 1.29927456e-01 1.47144675e-01 1.11981666e+00
2.19475091e-01 5.73398411e-01 -7.08478987e-01 2.76602447e-01
7.87272394e-01 2.42018308e-02 9.34114277e-01 -1.61293483e+00
-9.04314443e-02 -1.39684737e-01 -6.81358218e-01 -7.95422912e-01
3.62103544e-02 -4.97197181e-01 -4.10884291e-01 -1.11990070e+00
1.90005422e-01 -4.72026557e-01 3.01662177e-01 -2.40453988e-01
4.04715091e-01 -5.02387583e-02 5.32219648e-01 5.17352104e-01
-2.15100974e-01 1.08279206e-01 1.01690316e+00 1.95021048e-01
-1.70273304e-01 2.49399364e-01 -3.75256687e-01 1.09621835e+00
5.66660047e-01 -8.79053622e-02 -4.14680719e-01 -2.99114078e-01
6.65015817e-01 5.14905870e-01 2.30915457e-01 -1.00029516e+00
3.28179121e-01 -2.25026593e-01 6.12567971e-03 -8.14913332e-01
5.39162695e-01 -1.15523541e+00 9.88611400e-01 5.55669487e-01
8.60724896e-02 7.49524057e-01 -3.05885077e-01 2.29343012e-01
-3.71947080e-01 -7.34723926e-01 8.18318665e-01 -1.07579187e-01
-5.38307726e-01 3.10877487e-02 -3.08483899e-01 -2.85328060e-01
1.18189001e+00 -5.66078961e-01 -3.41362879e-02 -4.97476995e-01
-8.50865960e-01 -2.61325985e-01 8.45778883e-01 1.21312521e-01
5.42115986e-01 -1.50491011e+00 -4.32826757e-01 3.31292450e-01
1.12578310e-01 -2.58636653e-01 1.76408008e-01 1.27877915e+00
-8.39940310e-01 6.90481186e-01 -7.38942385e-01 -9.36525583e-01
-1.78586757e+00 7.99192309e-01 4.21036512e-01 2.09566206e-01
-1.24019802e-01 1.99351102e-01 2.50936985e-01 -3.31060559e-01
-1.92602407e-02 -2.79406399e-01 -3.63913655e-01 1.46444559e-01
1.71366557e-01 5.96127272e-01 -7.90555626e-02 -1.57387519e+00
-3.27165782e-01 1.53033447e+00 4.08043712e-01 -2.49962941e-01
9.68791723e-01 -3.61359745e-01 -5.75775802e-01 3.82422149e-01
1.33060896e+00 3.96662295e-01 -8.61950338e-01 -8.74714628e-02
-2.90746897e-01 -3.46633613e-01 -6.67326987e-01 1.97955817e-01
-8.09172571e-01 9.09063995e-01 2.00357944e-01 1.50816575e-01
1.04876184e+00 -2.37169519e-01 -7.01796412e-02 5.50583243e-01
5.03433049e-01 -6.61338389e-01 -2.90710151e-01 4.91502553e-01
1.29465735e+00 -9.73072529e-01 2.80350894e-01 -9.17219400e-01
-5.59009194e-01 1.60995543e+00 4.50346172e-01 -3.28327328e-01
5.85312545e-01 -6.04743026e-02 -9.39188078e-02 1.30697861e-01
-2.99172223e-01 -4.06926125e-02 5.05747557e-01 3.95582676e-01
3.54456395e-01 6.22860454e-02 -8.75638247e-01 -1.39762357e-01
-2.98087060e-01 -3.98268610e-01 1.07063961e+00 4.70132172e-01
-4.42964882e-01 -1.27038729e+00 -9.21898007e-01 -2.35573724e-01
-7.87348077e-02 3.30230951e-01 -5.73104799e-01 1.04653776e+00
3.67799439e-02 8.28708887e-01 -1.02159277e-01 -2.78260827e-01
5.88130713e-01 -3.60179156e-01 7.48989642e-01 -2.27457598e-01
-1.30483255e-01 2.72956878e-01 -1.71677828e-01 -5.04446864e-01
-6.32003665e-01 -8.83136511e-01 -7.22326040e-01 -4.14957315e-01
-4.35951024e-01 3.51293296e-01 8.71460140e-01 7.21778810e-01
-1.73833132e-01 -2.09229499e-01 7.32241154e-01 -7.72515655e-01
-7.27508426e-01 -6.21804237e-01 -7.91957736e-01 3.74545544e-01
4.23051655e-01 -6.93278670e-01 -6.43351793e-01 3.51560324e-01]
|
[7.963305473327637, -2.3238656520843506]
|
dc6db2aa-1247-4be9-adf4-6a6b7ad8082b
|
captain-comprehensive-composition-assistance
|
1811.04184
| null |
http://arxiv.org/abs/1811.04184v1
|
http://arxiv.org/pdf/1811.04184v1.pdf
|
CAPTAIN: Comprehensive Composition Assistance for Photo Taking
|
Many people are interested in taking astonishing photos and sharing with
others. Emerging hightech hardware and software facilitate ubiquitousness and
functionality of digital photography. Because composition matters in
photography, researchers have leveraged some common composition techniques to
assess the aesthetic quality of photos computationally. However, composition
techniques developed by professionals are far more diverse than well-documented
techniques can cover. We leverage the vast underexplored innovations in
photography for computational composition assistance. We propose a
comprehensive framework, named CAPTAIN (Composition Assistance for Photo
Taking), containing integrated deep-learned semantic detectors, sub-genre
categorization, artistic pose clustering, personalized aesthetics-based image
retrieval, and style set matching. The framework is backed by a large dataset
crawled from a photo-sharing Website with mostly photography enthusiasts and
professionals. The work proposes a sequence of steps that have not been
explored in the past by researchers. The work addresses personal preferences
for composition through presenting a ranked-list of photographs to the user
based on user-specified weights in the similarity measure. The matching
algorithm recognizes the best shot among a sequence of shots with respect to
the user's preferred style set. We have conducted a number of experiments on
the newly proposed components and reported findings. A user study demonstrates
that the work is useful to those taking photos.
|
['James Z. Wang', 'Mohammad Mahdi Kamani', 'Farshid Farhat']
|
2018-11-10
| null | null | null | null |
['set-matching']
|
['computer-vision']
|
[ 4.33408707e-01 -1.93192199e-01 -1.40445381e-01 -4.68168557e-01
-5.22732794e-01 -6.11702442e-01 4.31655496e-01 -2.01156214e-01
-4.68621626e-02 2.89520193e-02 7.10947514e-01 2.90428549e-01
-2.68391967e-01 -5.01154780e-01 -3.77094507e-01 -4.05776173e-01
3.65136206e-01 7.50129372e-02 4.87513132e-02 -2.98035979e-01
7.52256572e-01 5.58924198e-01 -2.20991611e+00 6.13717854e-01
6.15856826e-01 1.21166325e+00 3.00837606e-01 6.07131839e-01
-3.35266173e-01 6.44529343e-01 -4.94067401e-01 -9.46404517e-01
8.36350262e-01 -5.14924288e-01 -7.42731452e-01 7.11542249e-01
1.25600207e+00 -5.78675032e-01 -2.32493415e-01 1.08961689e+00
6.71785712e-01 2.67070532e-01 4.37389195e-01 -1.40629721e+00
-1.14808583e+00 3.21823180e-01 -3.63831133e-01 4.63777967e-02
6.11661434e-01 3.67980033e-01 1.12547302e+00 -7.15563238e-01
9.63651538e-01 1.28952157e+00 7.66405404e-01 4.88628000e-01
-8.49917710e-01 -5.08562505e-01 -1.64698765e-01 5.20374715e-01
-1.23410416e+00 -6.63820386e-01 1.05064440e+00 -3.15206677e-01
7.18445599e-01 5.44446290e-01 1.15275550e+00 1.04864204e+00
-2.33637452e-01 7.54519463e-01 1.19589663e+00 -5.19350290e-01
4.51349229e-01 5.32892048e-01 -2.01845944e-01 5.54850817e-01
-1.88241810e-01 -3.09632093e-01 -7.01174438e-01 -2.51275301e-03
5.97814560e-01 2.01013997e-01 -2.27688611e-01 -3.75559628e-01
-6.95410550e-01 5.42129755e-01 5.12329757e-01 2.45512262e-01
-3.77339363e-01 9.08257365e-02 3.51988405e-01 4.33563173e-01
5.72280467e-01 7.66078830e-01 1.18226662e-01 -1.36312336e-01
-1.01909745e+00 4.34309989e-03 6.94993377e-01 1.18342042e+00
7.20334291e-01 -3.27827156e-01 -2.37362966e-01 1.02195048e+00
2.71174219e-02 1.73905864e-01 4.50220197e-01 -1.47732341e+00
-1.46910399e-01 9.68216538e-01 5.12908585e-03 -1.25193548e+00
-9.38528497e-03 9.92852375e-02 -4.34112340e-01 6.64364576e-01
1.75713718e-01 1.69260532e-01 -6.21253550e-01 1.05670333e+00
2.91058779e-01 -1.54089540e-01 -4.10139382e-01 1.28311837e+00
7.34782636e-01 3.52868527e-01 1.17219016e-01 1.67053208e-01
1.53041446e+00 -1.02284861e+00 -5.55489242e-01 -7.73483142e-02
5.53969927e-02 -1.20235586e+00 1.60453367e+00 5.62449694e-01
-1.28455162e+00 -6.35677040e-01 -1.14388716e+00 -4.79045868e-01
-4.30782348e-01 1.52751446e-01 5.46518743e-01 8.92043471e-01
-1.42449057e+00 1.08747649e+00 4.54969816e-02 -1.13959813e+00
8.18232656e-01 -8.98152515e-02 -2.16857523e-01 -2.79558539e-01
-6.66458726e-01 9.47224319e-01 -2.88278669e-01 -5.61158955e-01
-5.72070003e-01 -8.61020625e-01 -4.01975513e-01 1.85535979e-02
2.21529633e-01 -9.86212671e-01 1.33476663e+00 -1.90842497e+00
-1.69676995e+00 1.47218704e+00 2.16342732e-01 -1.83439374e-01
5.43277800e-01 -3.99170727e-01 -5.67894995e-01 5.91227055e-01
-4.55421433e-02 7.82951951e-01 1.12277317e+00 -1.13686705e+00
-5.81403255e-01 -2.11360499e-01 5.01268506e-01 6.76138461e-01
-1.05133402e+00 2.69700021e-01 -4.81900126e-01 -5.25303125e-01
-3.13944012e-01 -7.01948941e-01 -1.32122450e-02 8.05594265e-01
-8.80852863e-02 -8.51251185e-02 8.32850575e-01 -5.18270373e-01
1.15473139e+00 -2.20641875e+00 -1.81722760e-01 2.56887935e-02
1.32676780e-01 2.50117302e-01 -1.07371926e-01 1.00186193e+00
6.64113462e-02 1.67139202e-01 1.71107575e-01 -5.50166130e-01
1.72219470e-01 -1.98115841e-01 -5.25671318e-02 4.91043091e-01
-1.82739422e-01 4.81169403e-01 -9.96499360e-01 -4.91901845e-01
5.55301905e-01 3.58560592e-01 -2.90640265e-01 1.99528366e-01
-1.33094355e-01 -4.47925866e-01 -2.42758453e-01 1.04615557e+00
6.36752427e-01 -2.28951246e-01 8.15126896e-02 -6.51984394e-01
-1.78798929e-01 -1.27756715e-01 -1.17487693e+00 1.97613931e+00
-5.32834291e-01 1.00386465e+00 -7.25616217e-02 -1.93019018e-01
8.39239657e-01 -1.34424835e-01 4.46547776e-01 -4.80651855e-01
2.60877669e-01 -1.53249592e-01 -4.99981999e-01 -1.00445998e+00
7.61911750e-01 -7.44531080e-02 2.26650611e-01 7.64857829e-01
-5.96258380e-02 -2.70620674e-01 -2.05031075e-02 2.71240383e-01
1.12600100e+00 1.49911568e-01 6.14424288e-01 -3.26208323e-01
2.98552692e-01 1.67144805e-01 -1.18747398e-01 3.98839682e-01
-4.41468954e-01 9.18132901e-01 -2.48769224e-01 -6.77872181e-01
-1.41998672e+00 -8.78150940e-01 2.31154725e-01 1.36815727e+00
5.20245910e-01 -6.39323592e-01 -9.22770739e-01 -2.23084196e-01
-5.97463287e-02 5.45838416e-01 -5.11491477e-01 6.02377914e-02
-7.88803920e-02 6.68052165e-03 2.48367250e-01 2.59204507e-01
7.54362404e-01 -1.15157032e+00 -9.15007651e-01 -4.26505119e-01
1.18333474e-01 -8.56812298e-01 -8.64622474e-01 -8.89222264e-01
-4.88610685e-01 -1.15344417e+00 -7.82533109e-01 -7.46195197e-01
6.55057073e-01 1.12358356e+00 1.13036418e+00 2.05031425e-01
-8.01750481e-01 1.02534592e+00 -6.42668545e-01 -3.48405123e-01
-1.21560425e-01 -3.38893563e-01 -7.10697426e-03 3.57310534e-01
6.04785144e-01 -7.02376902e-01 -1.27571452e+00 3.21304709e-01
-8.77791882e-01 9.96882319e-02 7.91290224e-01 1.97542280e-01
1.04538247e-01 2.84641329e-02 2.53368844e-03 -6.17624760e-01
6.86718702e-01 -3.04794937e-01 2.19334159e-02 4.20481443e-01
-5.35092056e-01 -5.01662314e-01 5.26241124e-01 -4.30993050e-01
-1.15041363e+00 1.25734106e-01 3.67928356e-01 -7.41248131e-01
-3.06566745e-01 -2.96137005e-01 -2.32652783e-01 -4.68964577e-01
8.56130540e-01 -6.13241494e-02 1.02298625e-01 -4.98861581e-01
7.39866257e-01 8.51381719e-01 5.09178936e-01 -1.52062327e-01
7.33603179e-01 8.17268670e-01 -3.60831261e-01 -9.57924008e-01
-5.84952474e-01 -8.47756207e-01 -3.13521475e-01 -1.01018107e+00
7.10909128e-01 -8.33715916e-01 -8.07305038e-01 4.18537974e-01
-8.30442607e-01 3.45002711e-02 -5.27255714e-01 -3.84294167e-02
-5.11602938e-01 6.83982372e-01 -3.27369690e-01 -9.93062019e-01
-6.62500739e-01 -8.23423028e-01 1.16679919e+00 5.43817639e-01
-6.15728438e-01 -5.93242228e-01 -2.06921399e-01 6.41897619e-01
5.73793769e-01 2.28353247e-01 5.13113201e-01 -7.56726339e-02
-6.91264033e-01 -3.02018583e-01 -3.78450692e-01 3.69343817e-01
-1.03630416e-01 3.02305728e-01 -1.28028011e+00 1.78659350e-01
-4.62051332e-02 -3.06854129e-01 4.79561538e-01 3.17877978e-01
1.29190767e+00 -3.80354911e-01 -7.67980218e-02 6.16781294e-01
1.64562666e+00 3.63569381e-03 8.93181860e-01 5.64081311e-01
4.53613341e-01 7.80815780e-01 4.82784450e-01 7.11219132e-01
3.05918217e-01 6.46390975e-01 2.75907040e-01 -4.01966274e-02
-4.69215453e-01 -4.86509472e-01 5.06822169e-01 2.07109734e-01
-2.17083842e-01 -1.46775246e-01 -5.15754580e-01 4.13975328e-01
-1.52937102e+00 -1.32114768e+00 5.39261103e-02 1.95033395e+00
4.48793441e-01 -2.07535952e-01 4.74156111e-01 1.60953552e-01
8.78081679e-01 9.80169773e-02 -5.70806801e-01 -6.24389410e-01
-2.21995898e-02 5.63068017e-02 6.44820511e-01 8.24543238e-02
-9.31381822e-01 7.71397173e-01 6.29842281e+00 9.50636804e-01
-8.07252944e-01 -8.33605006e-02 6.68335021e-01 -4.13559049e-01
-3.91409516e-01 2.83066798e-02 -2.99289703e-01 4.98218030e-01
4.48617280e-01 -5.47257185e-01 6.51861131e-01 1.33253157e+00
2.47828722e-01 -3.00707161e-01 -1.15320778e+00 1.44922149e+00
7.36908615e-01 -1.64568436e+00 1.08723044e-01 -1.72821864e-01
8.33575249e-01 -4.09873664e-01 3.23826283e-01 -2.06855252e-01
1.42097831e-01 -6.59812331e-01 7.37057269e-01 6.19427085e-01
8.67199361e-01 -4.77191925e-01 1.08428732e-01 -1.92312956e-01
-8.98395956e-01 -4.68541920e-01 -5.32037437e-01 -4.00130570e-01
3.09003532e-01 4.08915251e-01 -7.10898161e-01 2.14131519e-01
1.00444984e+00 9.28309321e-01 -7.25539684e-01 1.36945534e+00
9.32842121e-02 -7.33477622e-02 8.88008028e-02 -2.54670948e-01
9.64862928e-02 -4.75746065e-01 5.02765656e-01 1.12717116e+00
5.61907768e-01 2.91376889e-01 -1.26348406e-01 5.82813263e-01
-2.27918789e-01 3.12985688e-01 -7.30337024e-01 -5.07198311e-02
5.78923941e-01 1.73733020e+00 -8.60502541e-01 -4.00247455e-01
-2.11845756e-01 1.47172165e+00 -2.97067557e-02 8.45161919e-03
-4.51230377e-01 -1.52541950e-01 8.93171966e-01 4.93129104e-01
-1.39569357e-01 1.70892209e-01 -4.48077530e-01 -6.30068183e-01
3.53333098e-03 -8.16927671e-01 4.13416326e-01 -1.45624268e+00
-1.78332937e+00 6.08146071e-01 -1.21100925e-01 -1.57042134e+00
4.39652503e-01 -4.50886130e-01 -1.01136529e+00 4.31350887e-01
-9.81437564e-01 -1.20680928e+00 -9.45674777e-01 4.86099869e-01
1.06940508e+00 -2.58141845e-01 7.06438005e-01 3.81422251e-01
-3.29429150e-01 4.42008168e-01 -1.05876453e-01 -3.59485805e-01
9.68002796e-01 -1.06413019e+00 4.33108777e-01 6.91088974e-01
8.20390061e-02 4.52250183e-01 9.24442530e-01 -3.91300827e-01
-1.46661031e+00 -7.23947585e-01 6.81117058e-01 -4.17560786e-01
5.24739146e-01 -2.50232458e-01 -2.49624103e-01 6.77416548e-02
7.27527082e-01 -6.09997213e-01 1.01114941e+00 -1.80754691e-01
-3.63692224e-01 -3.78892690e-01 -1.37140214e+00 9.35455978e-01
1.58362770e+00 -6.17486656e-01 -4.88572866e-01 5.66835403e-01
4.50185120e-01 8.33306834e-02 -6.31496847e-01 -3.28311920e-01
9.72986460e-01 -1.32765496e+00 1.11738646e+00 -3.26926976e-01
8.06501985e-01 -1.13797061e-01 -1.41590327e-01 -1.20886934e+00
-4.32346284e-01 -9.51857209e-01 3.98914784e-01 1.46466637e+00
-1.92047238e-01 9.02041122e-02 7.74345160e-01 1.18347013e+00
-2.88911045e-01 -3.87790799e-01 -3.44994336e-01 -6.67483330e-01
-8.34299088e-01 -2.86555380e-01 5.19096375e-01 7.84056485e-01
8.15110952e-02 2.13233784e-01 -6.41122818e-01 -3.01820666e-01
6.69634223e-01 2.52179295e-01 9.37723041e-01 -1.06366897e+00
-2.60484576e-01 -8.53548646e-01 -7.13932276e-01 -4.73540395e-01
-2.78080195e-01 -5.88820159e-01 -2.74122506e-01 -1.58091986e+00
2.80188382e-01 -4.69708070e-02 6.33977726e-02 -4.27343249e-02
3.64891440e-02 6.64646208e-01 5.01127422e-01 3.14818919e-01
-8.92753720e-01 2.06458926e-01 1.07385170e+00 -6.05461784e-02
-3.31803747e-02 -1.91591457e-01 -1.09455967e+00 9.32086229e-01
7.22523332e-01 -6.65743649e-02 -6.79666936e-01 -3.47196847e-01
3.01236957e-01 -5.71620882e-01 3.72898787e-01 -1.48497188e+00
3.41417044e-01 -1.05215445e-01 4.82785225e-01 -2.06897154e-01
6.04419172e-01 -9.70274985e-01 5.14432311e-01 9.42039639e-02
-5.64264417e-01 1.24061711e-01 -1.89973772e-01 5.83627820e-01
1.94561481e-01 -2.23043397e-01 7.81612575e-01 -3.85040462e-01
-1.38548613e+00 2.41099343e-01 -1.90235466e-01 -2.60450155e-01
1.34137380e+00 -7.16592908e-01 -3.39431018e-01 -7.53655076e-01
-3.49886537e-01 -1.82842121e-01 1.07584941e+00 5.44999540e-01
9.04264390e-01 -1.45962656e+00 -3.55474561e-01 -1.62908554e-01
4.79790986e-01 -9.05948281e-01 7.12289155e-01 3.57468098e-01
-7.30943084e-01 -3.41532856e-01 -8.23878646e-01 -1.41338974e-01
-1.65942931e+00 7.76461601e-01 -2.02359185e-02 6.55750215e-01
-9.22402084e-01 1.06082785e+00 -7.91052207e-02 3.27358812e-01
3.12965184e-01 7.95008540e-02 -7.55703673e-02 3.24957788e-01
1.03061891e+00 7.94238925e-01 -1.73056304e-01 -4.60481524e-01
-1.91921890e-01 7.79511988e-01 1.71636686e-01 -3.49783711e-03
1.34771550e+00 -6.61317945e-01 1.70998409e-01 2.77268469e-01
1.35995412e+00 -2.81936437e-01 -1.12654567e+00 -9.78801325e-02
-1.80679977e-01 -1.15060651e+00 -8.78875777e-02 -8.34936142e-01
-9.69028175e-01 6.94524348e-01 7.78693914e-01 1.02652892e-01
1.50139630e+00 -1.08522579e-01 9.45847213e-01 1.18421793e-01
3.06907296e-01 -1.66834188e+00 5.02644479e-01 -3.22873294e-01
1.05597818e+00 -1.15461063e+00 1.88601613e-01 -3.58229995e-01
-1.12439668e+00 1.16401684e+00 5.80663681e-01 -5.24700880e-01
4.51593488e-01 -2.05447048e-01 2.19212264e-01 -4.24701691e-01
-5.37137032e-01 -3.88103783e-01 4.67497349e-01 8.68072271e-01
2.21971452e-01 -1.68142524e-02 -3.51005167e-01 3.53774071e-01
-2.53204644e-01 8.41298997e-02 6.03820145e-01 5.34511149e-01
-6.82282209e-01 -8.62834394e-01 -3.10525686e-01 4.38675314e-01
-9.88659784e-02 -5.13275303e-02 -9.86619949e-01 5.05792856e-01
2.68672436e-01 9.60474968e-01 -1.73144527e-02 -7.41167426e-01
2.98405498e-01 -3.05540324e-03 4.61164504e-01 -3.39575708e-01
-7.88499951e-01 -2.65193969e-01 2.23143786e-01 -9.46573019e-01
-6.62045002e-01 -6.32750571e-01 -3.01219672e-01 -6.49859726e-01
3.70067477e-01 -4.71520007e-01 7.43549049e-01 4.51339096e-01
5.14998496e-01 7.14128390e-02 7.10887551e-01 -1.21728659e+00
-1.37359247e-01 -6.33245528e-01 -5.63713133e-01 1.01740217e+00
-2.11391538e-01 -5.06204367e-01 -2.43072942e-01 2.92455077e-01]
|
[11.469855308532715, -0.9798357486724854]
|
379338a6-0ea1-4d32-b35a-1774683c66a9
|
on-graph-based-reentrancy-free-semantic
|
2302.07679
| null |
https://arxiv.org/abs/2302.07679v1
|
https://arxiv.org/pdf/2302.07679v1.pdf
|
On graph-based reentrancy-free semantic parsing
|
We propose a novel graph-based approach for semantic parsing that resolves two problems observed in the literature: (1) seq2seq models fail on compositional generalization tasks; (2) previous work using phrase structure parsers cannot cover all the semantic parses observed in treebanks. We prove that both MAP inference and latent tag anchoring (required for weakly-supervised learning) are NP-hard problems. We propose two optimization algorithms based on constraint smoothing and conditional gradient to approximately solve these inference problems. Experimentally, our approach delivers state-of-the-art results on Geoquery, Scan and Clevr, both for i.i.d. splits and for splits that test for compositional generalization.
|
['Caio Corro', 'Alban Petit']
|
2023-02-15
| null | null | null | null |
['semantic-parsing']
|
['natural-language-processing']
|
[ 4.36279386e-01 6.46631956e-01 -5.24042070e-01 -7.00165033e-01
-1.42631352e+00 -9.77942407e-01 9.52684507e-03 3.74980599e-01
-3.10769141e-01 1.03156877e+00 2.71213770e-01 -8.14617634e-01
-2.87031353e-01 -8.47629726e-01 -8.58650446e-01 -2.99019724e-01
-2.21111819e-01 1.02630556e+00 6.39575481e-01 -1.01121038e-01
2.26226375e-01 1.87931225e-01 -1.31383383e+00 5.00446141e-01
1.06778944e+00 6.09043062e-01 1.55285284e-01 8.77807498e-01
-7.62911975e-01 6.85461819e-01 -5.07217288e-01 -7.04788148e-01
7.42491633e-02 -6.75112545e-01 -1.41984391e+00 -1.71855986e-01
6.62532449e-01 1.82760954e-01 1.41133443e-01 1.28100765e+00
2.86462903e-01 -3.71037982e-02 4.10803914e-01 -1.13298714e+00
-2.69454271e-01 1.22023857e+00 -4.33182120e-01 3.31329823e-01
4.70134914e-01 -3.57296079e-01 1.67718220e+00 -3.21977615e-01
8.21942091e-01 1.56640232e+00 6.16236567e-01 7.69473493e-01
-1.49923801e+00 -4.54390347e-01 3.46407384e-01 9.46974568e-03
-1.11711466e+00 -1.68106392e-01 2.89323211e-01 -1.01694167e-01
1.38471222e+00 3.13905627e-01 1.38366565e-01 1.05447042e+00
-4.16525230e-02 7.80652583e-01 1.08737338e+00 -5.30875981e-01
5.31871557e-01 -2.82739997e-01 3.90118837e-01 9.72305596e-01
6.75140172e-02 -4.66688760e-02 -8.60566258e-01 -4.32072520e-01
4.36470628e-01 -8.36645901e-01 1.14949435e-01 -2.23361880e-01
-7.45148897e-01 1.11264145e+00 -1.51711360e-01 1.56553656e-01
3.53933610e-02 2.89197057e-01 6.14318728e-01 5.00662446e-01
6.13151550e-01 3.21314573e-01 -1.07359803e+00 -3.65495622e-01
-8.01274717e-01 3.27297598e-01 1.19010925e+00 1.22883725e+00
7.92280614e-01 -2.98959792e-01 -2.72504855e-02 7.59555638e-01
1.16654232e-01 4.17828113e-01 2.78908938e-01 -1.01905572e+00
9.33499634e-01 3.03788453e-01 -2.04077929e-01 -3.35583717e-01
-5.95648289e-01 -3.44864696e-01 -2.32989222e-01 -3.73386711e-01
7.16815829e-01 -1.91796526e-01 -9.82787430e-01 2.17167616e+00
3.22498441e-01 2.46540144e-01 2.57796109e-01 5.36144257e-01
5.84054291e-01 4.39033717e-01 5.59171379e-01 -1.91281170e-01
1.52254927e+00 -7.85261691e-01 -6.61005199e-01 -5.49763262e-01
1.12600863e+00 -3.37608814e-01 1.15555596e+00 2.81662613e-01
-1.31617486e+00 -1.93921313e-01 -6.99022233e-01 -1.90992609e-01
-1.44252867e-01 -3.27796310e-01 9.51306343e-01 9.87769723e-01
-1.05172980e+00 6.85495436e-01 -1.08754778e+00 -4.75399315e-01
3.14624637e-01 4.32262391e-01 -3.28977644e-01 -1.21991687e-01
-1.02641976e+00 6.18270636e-01 6.22966051e-01 -2.74244040e-01
-5.82096279e-01 -8.98497283e-01 -9.22624052e-01 3.03949505e-01
6.94968522e-01 -7.63457000e-01 1.39515841e+00 -6.91893339e-01
-1.56976068e+00 1.29598618e+00 -5.17906129e-01 -6.37596607e-01
2.16758832e-01 -3.01442653e-01 -9.22657251e-02 2.28840351e-01
4.62052196e-01 6.88427567e-01 1.71227843e-01 -8.42739999e-01
-9.42657888e-01 -6.77275956e-01 1.21129684e-01 2.15598345e-01
1.04133420e-01 3.03360999e-01 -4.29734766e-01 -4.15487766e-01
4.81258273e-01 -7.77822375e-01 -3.43029499e-01 -5.65876663e-01
-7.23435462e-01 -4.51842129e-01 1.84389055e-01 -7.59895086e-01
1.11942852e+00 -1.84791958e+00 3.82346749e-01 1.07505687e-01
-2.64709771e-01 -6.77494481e-02 -3.24785501e-01 5.16661704e-01
2.74429061e-02 6.01107359e-01 -6.02405965e-01 -4.40169185e-01
2.93707460e-01 6.89076304e-01 -4.24547434e-01 8.57700780e-02
1.86494723e-01 1.03372025e+00 -8.61460567e-01 -7.23291695e-01
-3.17237645e-01 -3.52601022e-01 -8.73943985e-01 2.01449364e-01
-8.34639549e-01 3.53816807e-01 -3.72796923e-01 6.36125922e-01
7.28795946e-01 -2.80060223e-03 9.10714746e-01 1.84569970e-01
2.10119002e-02 9.52590823e-01 -9.38733697e-01 2.40479207e+00
-2.82729387e-01 -7.66376266e-03 2.44962364e-01 -1.38828027e+00
5.80694795e-01 -8.66740197e-03 1.97291765e-02 -5.40756762e-01
-2.77572960e-01 3.19957554e-01 -2.81616807e-01 -6.04896724e-01
3.40976089e-01 -3.94966662e-01 -5.97432196e-01 3.18111062e-01
4.48983580e-01 -1.87390372e-01 3.78494412e-01 2.73096055e-01
1.32697213e+00 5.44987857e-01 6.81092516e-02 -7.16207743e-01
3.25787574e-01 2.60746628e-01 1.07637668e+00 1.06073165e+00
2.48356581e-01 3.40444982e-01 1.25663376e+00 -1.48493379e-01
-8.04472089e-01 -1.46381438e+00 1.85128523e-03 1.60443473e+00
-1.83946770e-02 -5.18685818e-01 -9.59777415e-01 -1.23911893e+00
-2.19747603e-01 1.10170341e+00 -4.00151610e-01 1.21110424e-01
-8.52818847e-01 -8.78030598e-01 9.18863773e-01 8.24415922e-01
1.53634980e-01 -7.55904377e-01 -3.87529463e-01 3.02400678e-01
-3.58891934e-01 -1.53208745e+00 -1.31849349e-01 6.21074796e-01
-1.20877719e+00 -1.14839518e+00 1.17491998e-01 -9.07535553e-01
4.00314808e-01 -3.71944517e-01 1.52994025e+00 -6.01635140e-04
-1.12059362e-01 1.83331624e-01 -4.28808719e-01 -1.22425050e-01
-6.71480060e-01 5.75765729e-01 -5.30589521e-01 -7.41699696e-01
5.37091911e-01 -6.92945838e-01 1.25625879e-01 1.87181666e-01
-7.09708452e-01 4.39104177e-02 3.06403160e-01 9.66518164e-01
6.77893281e-01 -1.34016529e-01 6.28819406e-01 -1.61100161e+00
4.68189895e-01 -4.59108889e-01 -8.37831736e-01 6.28385842e-01
-3.17198217e-01 5.51887631e-01 4.48714793e-01 2.10399792e-01
-1.25904083e+00 7.83311799e-02 -4.48643923e-01 2.23894536e-01
-5.03492534e-01 6.38311982e-01 -3.06053668e-01 3.72989446e-01
4.56193507e-01 -1.99602614e-03 -3.00954789e-01 -6.74032629e-01
4.59844202e-01 3.02608788e-01 7.62671411e-01 -1.21108007e+00
5.56985855e-01 3.35312992e-01 2.44718000e-01 -7.10085094e-01
-1.33961034e+00 -4.81496304e-01 -6.71312630e-01 6.67999566e-01
1.29890490e+00 -8.35316479e-01 -7.23925531e-01 -9.73457620e-02
-1.22794306e+00 -6.19925499e-01 -3.36388707e-01 3.25465262e-01
-8.70516062e-01 6.47589862e-01 -8.65463793e-01 -6.28932893e-01
-2.54483044e-01 -7.56799996e-01 1.27843177e+00 3.30847166e-02
-2.53039420e-01 -1.12682188e+00 2.16669902e-01 4.24313724e-01
-1.19210826e-02 -7.22949132e-02 1.50564218e+00 -9.35119927e-01
-5.25268674e-01 2.67001867e-01 -1.99198350e-01 7.16030151e-02
-3.24517787e-01 -4.08775747e-01 -7.19935298e-01 -3.57204210e-03
-2.67530799e-01 -4.93599087e-01 9.53241527e-01 2.88898587e-01
1.12201631e+00 -3.63203794e-01 -3.21437985e-01 6.89543188e-01
1.39626479e+00 -1.54486030e-01 4.97542650e-01 -3.48895155e-02
3.06491137e-01 8.00319612e-01 7.15502441e-01 2.05252290e-01
4.56515759e-01 5.98020017e-01 3.30915451e-01 2.84499913e-01
1.30588710e-01 -7.40264535e-01 2.44289055e-01 5.31523883e-01
2.16257036e-01 -4.32017654e-01 -9.46484327e-01 6.58350050e-01
-2.05352378e+00 -5.58706522e-01 -1.68594927e-01 2.06871772e+00
1.05568564e+00 1.91354305e-01 9.61551815e-02 -1.65674463e-01
7.05063701e-01 -1.21985614e-01 -2.05952272e-01 -6.95627928e-01
-1.86381310e-01 9.70200539e-01 7.13926554e-01 8.91266823e-01
-1.05582905e+00 1.68215895e+00 6.72551966e+00 7.75317252e-01
-3.63299936e-01 3.27228427e-01 3.54588330e-01 2.25274891e-01
-6.40430391e-01 4.25252914e-01 -1.09218442e+00 1.55438811e-01
1.17987216e+00 1.23541534e-01 4.32367831e-01 6.38127923e-01
-3.78008217e-01 -3.56899083e-01 -1.31524551e+00 3.98377985e-01
-1.61582381e-01 -1.33222711e+00 -1.13386780e-01 -2.46694922e-01
5.93181670e-01 1.79631323e-01 -2.20409513e-01 4.20215547e-01
8.92014742e-01 -1.01036751e+00 6.69880450e-01 -1.94416627e-01
7.52912641e-01 -7.41891921e-01 7.49394357e-01 4.54950601e-01
-9.63613868e-01 9.17274039e-03 -5.13437569e-01 1.12335980e-01
4.68400955e-01 4.77674305e-01 -6.78441703e-01 9.18275416e-01
5.68186104e-01 3.37472826e-01 -3.16946357e-01 5.81612229e-01
-1.08155644e+00 1.06255233e+00 -5.54620683e-01 6.46074638e-02
3.46464902e-01 1.21412994e-02 5.93473673e-01 1.41795397e+00
1.55659661e-01 1.81861565e-01 3.94716114e-01 9.67319489e-01
-2.24213973e-02 7.15238377e-02 -2.19351366e-01 -1.00406475e-01
6.01344764e-01 7.85135329e-01 -8.99019122e-01 -9.05431062e-02
-3.23128819e-01 1.08215392e+00 6.25757575e-01 3.41673464e-01
-5.86082280e-01 -7.32105672e-02 6.15346193e-01 -2.79921860e-01
6.08119607e-01 -3.25475127e-01 -4.04375464e-01 -1.26047969e+00
5.45970760e-02 -7.52772808e-01 1.18827295e+00 -3.43647718e-01
-1.26327705e+00 2.33228728e-01 2.17907459e-01 -1.37225538e-01
-5.04482746e-01 -6.80130064e-01 -5.58144033e-01 8.04149926e-01
-1.62972653e+00 -1.03064704e+00 2.62226671e-01 2.18536898e-01
5.80301464e-01 6.63285702e-02 1.29641795e+00 6.71003088e-02
-3.68766844e-01 5.54379046e-01 -5.03621638e-01 1.12844156e-02
2.03734174e-01 -1.68896008e+00 7.85922766e-01 9.18366313e-01
3.64728212e-01 4.51939285e-01 7.40651309e-01 -6.06606305e-01
-1.41253197e+00 -7.45741069e-01 1.29595971e+00 -2.80911356e-01
6.60177290e-01 -6.51324928e-01 -8.51178586e-01 1.15314937e+00
-2.03579124e-02 -2.08723053e-01 7.69740701e-01 7.67373025e-01
-6.36183858e-01 3.32789302e-01 -9.84310567e-01 1.10727847e-01
1.55245245e+00 -4.82023865e-01 -8.34895074e-01 4.58277076e-01
9.61867094e-01 -6.55366242e-01 -7.02081740e-01 3.21347535e-01
1.71153724e-01 -7.79166043e-01 7.77644634e-01 -1.33618510e+00
4.59384888e-01 1.90903857e-01 -4.12358433e-01 -1.14758897e+00
-7.52137080e-02 -7.64110029e-01 1.32063940e-01 1.19824135e+00
8.50586534e-01 -6.76941574e-01 1.18763685e+00 3.98031116e-01
-3.41625988e-01 -5.19035935e-01 -1.34531498e+00 -1.00399542e+00
4.19777542e-01 -6.56856537e-01 4.58616197e-01 8.45361829e-01
2.35499620e-01 6.37494922e-01 3.99736837e-02 3.76984894e-01
7.42451191e-01 3.65501970e-01 4.78254080e-01 -1.11817145e+00
-5.94990373e-01 -2.42638856e-01 -2.47987792e-01 -1.27279019e+00
8.46125007e-01 -1.17121518e+00 7.14543387e-02 -1.43380928e+00
2.17755273e-01 -6.41883433e-01 1.24137409e-01 6.80704296e-01
-1.44605622e-01 -2.52716184e-01 -1.51227221e-01 -4.64563817e-01
-7.06846893e-01 1.72586247e-01 5.84280372e-01 1.08729348e-01
3.69177870e-02 -6.29123524e-02 -6.76927686e-01 6.10600293e-01
8.60331953e-01 -8.54076445e-01 -3.83689582e-01 -5.81999481e-01
7.58065701e-01 4.99837279e-01 2.26241902e-01 -4.67005789e-01
3.53546031e-02 -2.19222978e-01 -5.24098635e-01 -7.04854548e-01
-4.33399342e-02 -4.39144075e-01 -2.52551258e-01 4.40725476e-01
-6.08785570e-01 -1.99529789e-02 3.86434227e-01 6.39685214e-01
-2.30453864e-01 -7.81895041e-01 3.92125189e-01 -2.97733307e-01
-6.68188870e-01 1.37969732e-01 -3.09569299e-01 9.31938350e-01
6.29860342e-01 3.17501634e-01 -4.09446448e-01 -2.25311071e-01
-9.79937136e-01 4.13454384e-01 1.01710476e-01 1.33434847e-01
1.27024904e-01 -6.74708843e-01 -1.03017950e+00 -3.56914066e-02
-9.08940211e-02 2.19769984e-01 1.30006194e-01 6.43144846e-01
-6.85887814e-01 5.37091970e-01 1.93249837e-01 -5.32989204e-01
-1.35901558e+00 4.17768955e-01 9.42888409e-02 -7.24449337e-01
-4.71187145e-01 1.27934742e+00 1.88313648e-01 -9.37523782e-01
1.77761897e-01 -4.09539700e-01 3.29173416e-01 -2.04269111e-01
4.13198024e-02 1.74657032e-01 5.04303351e-02 -8.87778923e-02
-4.64457572e-01 3.82366300e-01 4.94962968e-02 -3.06597352e-01
1.36286938e+00 -6.61889464e-02 -3.37142944e-01 2.68842250e-01
1.08134961e+00 1.72680262e-02 -6.82589889e-01 -3.30765575e-01
6.54999077e-01 -1.48370564e-01 -3.30764383e-01 -8.55966687e-01
-5.89657962e-01 1.07348096e+00 -5.03164679e-02 1.74094409e-01
8.84941459e-01 5.81443012e-01 8.02011549e-01 5.08002102e-01
4.58437085e-01 -1.24404609e+00 -4.22653198e-01 7.13483930e-01
2.83040345e-01 -7.02919483e-01 -1.36410579e-01 -1.12283456e+00
-3.96604747e-01 9.91471946e-01 3.65266800e-01 2.63375521e-01
2.36122742e-01 5.08255780e-01 -2.47786909e-01 -3.08788419e-01
-9.92371440e-01 -3.52908999e-01 -1.36209920e-01 3.96799594e-01
4.33471382e-01 1.62783399e-01 -6.60607696e-01 8.89208734e-01
-4.32260603e-01 -3.33997130e-01 2.43836388e-01 1.03224313e+00
-5.26225686e-01 -1.44159675e+00 9.35799405e-02 4.80513833e-02
-8.94566655e-01 -3.73158664e-01 -4.12091404e-01 6.46932423e-01
-1.34422362e-01 8.97057593e-01 7.80550316e-02 -1.69556569e-02
1.48114666e-01 4.47748154e-01 8.63196731e-01 -9.82822657e-01
-4.98809487e-01 -1.96839228e-01 8.19444537e-01 -7.83029139e-01
-4.21191037e-01 -7.67310739e-01 -1.63786173e+00 6.13168031e-02
-3.52985561e-01 5.15508294e-01 6.90536737e-01 1.05076754e+00
1.98798090e-01 4.06147182e-01 1.68306440e-01 1.57905389e-02
-9.52449977e-01 -9.18525338e-01 -5.69223464e-01 3.22863519e-01
-4.76046145e-01 -2.11605579e-01 -3.66173804e-01 5.54282498e-03]
|
[10.460500717163086, 9.413471221923828]
|
f178ae46-c012-4814-a556-35d76bd0ad10
|
learning-classifiers-of-prototypes-and
|
2212.08355
| null |
https://arxiv.org/abs/2212.08355v1
|
https://arxiv.org/pdf/2212.08355v1.pdf
|
Learning Classifiers of Prototypes and Reciprocal Points for Universal Domain Adaptation
|
Universal Domain Adaptation aims to transfer the knowledge between the datasets by handling two shifts: domain-shift and category-shift. The main challenge is correctly distinguishing the unknown target samples while adapting the distribution of known class knowledge from source to target. Most existing methods approach this problem by first training the target adapted known classifier and then relying on the single threshold to distinguish unknown target samples. However, this simple threshold-based approach prevents the model from considering the underlying complexities existing between the known and unknown samples in the high-dimensional feature space. In this paper, we propose a new approach in which we use two sets of feature points, namely dual Classifiers for Prototypes and Reciprocals (CPR). Our key idea is to associate each prototype with corresponding known class features while pushing the reciprocals apart from these prototypes to locate them in the potential unknown feature space. The target samples are then classified as unknown if they fall near any reciprocals at test time. To successfully train our framework, we collect the partial, confident target samples that are classified as known or unknown through on our proposed multi-criteria selection. We then additionally apply the entropy loss regularization to them. For further adaptation, we also apply standard consistency regularization that matches the predictions of two different views of the input to make more compact target feature space. We evaluate our proposal, CPR, on three standard benchmarks and achieve comparable or new state-of-the-art results. We also provide extensive ablation experiments to verify our main design choices in our framework.
|
['In So Kweon', 'Sanghyun Woo', 'KwanYong Park', 'Inkyu Shin', 'Sungsu Hur']
|
2022-12-16
| null | null | null | null |
['universal-domain-adaptation']
|
['computer-vision']
|
[ 3.69249433e-01 -1.10308200e-01 -4.98867452e-01 -5.53351760e-01
-1.13122058e+00 -8.13441455e-01 5.58035135e-01 2.24093273e-01
-3.86692822e-01 8.27624917e-01 -1.97240040e-01 7.38144442e-02
-3.65346342e-01 -6.77959800e-01 -7.28235602e-01 -9.00826395e-01
1.41524091e-01 7.47046888e-01 4.76777077e-01 1.77532569e-01
3.07870239e-01 3.99599552e-01 -1.74004638e+00 5.52470982e-01
9.75367725e-01 1.30859327e+00 1.00582931e-02 2.97550112e-01
2.63282582e-02 3.22137475e-02 -6.08817339e-01 -4.11778629e-01
6.55920804e-01 -4.14523661e-01 -6.18624508e-01 9.14763734e-02
5.37179172e-01 -1.67790689e-02 2.68353313e-01 1.05545139e+00
3.17545891e-01 1.76400840e-01 1.03061640e+00 -1.33360207e+00
-7.29054987e-01 3.18739504e-01 -6.03422344e-01 1.37925014e-01
3.45967144e-01 -9.04984027e-02 8.75673294e-01 -1.24505925e+00
4.91563618e-01 9.98579800e-01 6.03705943e-01 6.39575005e-01
-1.24923277e+00 -8.26273680e-01 3.98859113e-01 5.56553960e-01
-1.61319244e+00 -5.00819087e-01 7.88366556e-01 -4.60218310e-01
3.30336362e-01 3.23383361e-01 2.73171157e-01 1.20716786e+00
-8.68729502e-02 7.57970035e-01 1.26598012e+00 -3.97843331e-01
5.17564476e-01 7.78376639e-01 2.75588185e-01 1.83676213e-01
2.76557326e-01 2.49955058e-01 -3.87811780e-01 -3.01695108e-01
1.69298932e-01 1.39685869e-01 -4.62512046e-01 -8.96888435e-01
-1.09636307e+00 8.35627556e-01 3.96570086e-01 2.17740715e-01
-2.49340966e-01 -7.29350924e-01 2.67700016e-01 3.29171747e-01
2.67380476e-01 1.35882273e-01 -7.59499609e-01 3.75289649e-01
-7.52838314e-01 1.02270536e-01 6.90095127e-01 1.00665545e+00
9.46666479e-01 -6.06010020e-01 -2.58304119e-01 8.42849135e-01
1.94093779e-01 4.17927921e-01 7.71128058e-01 -5.20651639e-01
5.82346678e-01 8.19518149e-01 1.53270364e-01 -8.11899722e-01
-6.99773282e-02 -6.55122697e-01 -7.64320791e-01 2.11985961e-01
5.99684894e-01 2.94804703e-02 -1.02553320e+00 1.67559016e+00
6.63884759e-01 3.31990808e-01 3.30716908e-01 8.92940342e-01
3.33010495e-01 3.48309785e-01 -8.60331133e-02 -2.22287551e-01
1.14810836e+00 -6.91005349e-01 -2.68018246e-01 -3.37120503e-01
4.01732534e-01 -5.23566663e-01 1.10455489e+00 3.22268933e-01
-6.44131601e-01 -5.74347854e-01 -1.15896201e+00 4.79111999e-01
-6.63257241e-01 3.96608859e-01 1.35955676e-01 7.16659129e-01
-3.73039961e-01 6.36079788e-01 -5.69979668e-01 -2.51633435e-01
4.66852307e-01 3.35597843e-01 -5.22370875e-01 -2.44776666e-01
-1.13121927e+00 6.88492596e-01 6.13412440e-01 -1.24146879e-01
-7.30966568e-01 -8.11530173e-01 -7.90919483e-01 1.82431855e-03
5.55767596e-01 -5.89064658e-01 9.35858250e-01 -1.18156433e+00
-1.34635699e+00 8.56661201e-01 -2.30221763e-01 -3.47863972e-01
7.27830291e-01 -1.13079594e-02 -5.41940212e-01 -1.33367166e-01
2.15283155e-01 2.42436364e-01 1.04432881e+00 -1.45213985e+00
-1.09611833e+00 -5.49555242e-01 -2.82724410e-01 4.06409353e-01
-4.30167019e-01 -3.22254926e-01 -2.87755370e-01 -4.25865531e-01
3.98288071e-01 -8.43607426e-01 1.00812145e-01 2.91986726e-02
-3.95514369e-01 -1.30362764e-01 9.35708225e-01 -3.20513248e-01
1.08204317e+00 -2.22030473e+00 8.66031125e-02 5.32907605e-01
-4.04325081e-03 2.86000609e-01 -7.32141081e-03 2.80842781e-02
-1.08987138e-01 -7.14031607e-02 -4.20241386e-01 -1.80229902e-01
-6.68351948e-02 1.37230009e-01 -3.88161153e-01 4.66638297e-01
3.03318173e-01 4.55095112e-01 -8.02123487e-01 -2.63300031e-01
9.48407128e-02 2.36377254e-01 -4.58747953e-01 3.03933322e-01
7.42364377e-02 4.27214652e-01 -4.63304967e-01 7.08786309e-01
9.37833250e-01 -1.21744663e-01 2.00378984e-01 -1.98376894e-01
3.31027836e-01 4.05139476e-02 -1.69955707e+00 1.14733577e+00
-2.47615412e-01 1.32041471e-02 -7.43654296e-02 -1.33427799e+00
1.08946025e+00 2.00474970e-02 2.52029151e-01 -4.94533449e-01
-2.06169322e-01 3.77897382e-01 -1.09878397e-02 -1.51943162e-01
1.83399215e-01 -2.70666242e-01 -1.44436464e-01 -2.72054933e-02
1.12643436e-01 3.12978297e-01 -8.07206705e-02 -1.78779781e-01
8.21196198e-01 1.45653397e-01 4.89171356e-01 -1.15595065e-01
8.66662979e-01 -1.18455023e-01 8.30039501e-01 7.92110503e-01
-4.17896479e-01 7.36975431e-01 2.77838141e-01 -2.32366785e-01
-7.43867278e-01 -1.31886566e+00 -3.15967292e-01 1.03238046e+00
3.52020144e-01 1.19985938e-01 -3.68594050e-01 -1.41530526e+00
2.69280136e-01 8.57282698e-01 -8.06437314e-01 -5.28551280e-01
-3.15035403e-01 -6.22537374e-01 9.78950411e-02 4.58905637e-01
3.47680271e-01 -7.33284354e-01 -5.48688591e-01 1.31669953e-01
-9.70813632e-03 -7.59214401e-01 -3.48368227e-01 4.74078268e-01
-6.02963686e-01 -1.25253642e+00 -7.71279991e-01 -6.70961738e-01
7.68250227e-01 1.34936392e-01 8.71732175e-01 -4.13553447e-01
4.05678935e-02 2.68728405e-01 -4.02847558e-01 -2.15663016e-01
-4.33509618e-01 -2.18111221e-02 2.45564446e-01 5.03387451e-01
5.47154307e-01 -3.98074329e-01 -5.95680833e-01 8.18747580e-01
-7.11463392e-01 -4.12971765e-01 7.70591378e-01 1.01808012e+00
8.79179776e-01 3.50448936e-02 8.49003911e-01 -8.80998135e-01
3.85659218e-01 -8.16454589e-01 -3.55454415e-01 5.53948283e-01
-7.12613165e-01 3.66131254e-02 8.65060806e-01 -8.98504257e-01
-8.12495232e-01 3.11519712e-01 3.51854533e-01 -6.93482935e-01
-4.76410002e-01 3.45318288e-01 -5.39887726e-01 1.40128225e-01
8.44311774e-01 2.36045942e-01 -2.12689281e-01 -3.79921049e-01
2.51618385e-01 8.01869929e-01 5.65632343e-01 -5.95314026e-01
8.88333857e-01 3.15476805e-01 -2.84335285e-01 -2.59629607e-01
-8.71811628e-01 -6.74771249e-01 -8.07990849e-01 1.89913765e-01
4.33564752e-01 -7.12830782e-01 -3.31871808e-01 2.08716989e-01
-6.72311604e-01 4.18083929e-03 -5.13394296e-01 5.06731689e-01
-4.16240305e-01 4.15121615e-01 1.54732287e-01 -7.24051118e-01
-1.57460600e-01 -9.94220793e-01 9.86992002e-01 2.21761733e-01
-3.36196125e-02 -8.78578067e-01 1.38634490e-02 1.69048849e-02
1.98899716e-01 9.64224115e-02 8.65236998e-01 -1.45376337e+00
-1.90387383e-01 -5.48486948e-01 -1.33033425e-01 5.78063369e-01
3.97287846e-01 -2.65537977e-01 -9.38347399e-01 -5.10548949e-01
6.76562265e-02 -3.18090707e-01 9.38799441e-01 2.14152858e-01
1.08277774e+00 -2.67850190e-01 -6.47228479e-01 6.22033000e-01
1.39203894e+00 2.87367672e-01 3.14179420e-01 4.01015222e-01
2.95442462e-01 5.23393095e-01 9.59914327e-01 4.33095485e-01
3.50048840e-01 7.94970989e-01 2.48180822e-01 1.59245223e-01
1.65055603e-01 -1.57484964e-01 3.99290919e-01 2.33940259e-01
3.13757032e-01 -2.95295447e-01 -8.71179819e-01 6.36484504e-01
-1.83954835e+00 -8.91273618e-01 3.28345686e-01 2.80378723e+00
8.24914217e-01 2.03689098e-01 3.04016583e-02 1.81346357e-01
1.03761375e+00 -4.25302327e-01 -8.85516346e-01 1.09791644e-02
-1.63474470e-01 9.02684182e-02 4.20666039e-01 2.91828543e-01
-1.49641502e+00 4.85640347e-01 5.21692610e+00 9.40167665e-01
-1.10772741e+00 -1.96272418e-01 6.55215263e-01 3.19637209e-02
2.73957253e-02 7.42324665e-02 -1.17380941e+00 5.21805942e-01
7.76807427e-01 -4.37031358e-01 2.14859799e-01 9.60882187e-01
-3.27283710e-01 -5.76135889e-02 -1.55828619e+00 8.01279902e-01
1.64295614e-01 -9.78942454e-01 7.25716278e-02 -1.16775654e-01
5.92177212e-01 -2.84508973e-01 2.69279569e-01 5.97287416e-01
9.59377959e-02 -6.69154644e-01 6.37938142e-01 4.60559189e-01
8.19191873e-01 -6.32995486e-01 7.03072250e-01 5.67926586e-01
-1.11705017e+00 -3.65979880e-01 -4.30519819e-01 2.62542576e-01
-3.52868766e-01 4.83262330e-01 -9.58873093e-01 8.00290704e-01
7.03955889e-01 6.54622138e-01 -6.91208780e-01 1.21958983e+00
9.30550247e-02 4.15710598e-01 -4.53149319e-01 2.90838808e-01
-1.88722000e-01 -5.66981770e-02 5.38788676e-01 9.58145678e-01
5.14848650e-01 -1.66624323e-01 4.43192959e-01 7.69783914e-01
7.95855969e-02 8.10998008e-02 -4.79996294e-01 4.90305394e-01
6.79908752e-01 1.11143517e+00 -5.41306555e-01 -4.56497401e-01
-3.43061626e-01 1.09468353e+00 4.02737170e-01 3.96598905e-01
-8.28596413e-01 -4.14175004e-01 4.42476183e-01 -3.64437737e-02
6.03106976e-01 2.99767673e-01 -9.19455215e-02 -1.28049028e+00
4.74775225e-01 -8.00992548e-01 8.57885659e-01 -2.00008288e-01
-1.87543964e+00 4.69776958e-01 8.56865197e-02 -1.74112749e+00
-2.93551475e-01 -5.35529971e-01 -4.34174389e-01 9.94295955e-01
-1.45506442e+00 -1.04390430e+00 -2.89049417e-01 8.01856458e-01
4.26207095e-01 -2.48937175e-01 9.16456103e-01 2.23389491e-01
-3.45868111e-01 1.07291305e+00 4.92021829e-01 -6.26942143e-02
1.12963891e+00 -1.23031962e+00 -2.40243208e-02 5.64682126e-01
-8.90091583e-02 4.64871317e-01 3.89091402e-01 -5.80486059e-01
-9.41965938e-01 -1.29825556e+00 4.64172870e-01 -5.19120157e-01
4.73657221e-01 -4.79961246e-01 -1.31270397e+00 6.73704922e-01
-3.70126456e-01 4.75419730e-01 8.00049663e-01 2.93754973e-03
-5.80238879e-01 -2.29619399e-01 -1.61341441e+00 1.91073850e-01
7.59763598e-01 -2.14602247e-01 -7.75831938e-01 1.88664094e-01
3.72760624e-01 -2.86466539e-01 -7.13524580e-01 7.25121140e-01
5.68832099e-01 -9.59077954e-01 9.15525198e-01 -6.56780064e-01
1.44313127e-01 -4.23056126e-01 -4.17210817e-01 -1.50862932e+00
-4.17797804e-01 -7.53289014e-02 -4.46607508e-02 1.40923774e+00
5.24546504e-01 -9.12499607e-01 9.52879667e-01 5.49107790e-01
1.69190988e-02 -8.05267632e-01 -1.18843186e+00 -9.90730286e-01
2.48932093e-01 -1.07348010e-01 5.53119123e-01 1.12437403e+00
-1.66021630e-01 1.75670713e-01 -9.79079083e-02 4.42053437e-01
6.93862975e-01 3.72291416e-01 6.27917230e-01 -1.36875975e+00
-3.36556524e-01 -3.22643816e-01 -5.36777973e-01 -8.09188843e-01
1.37719110e-01 -1.00435710e+00 -1.50260364e-03 -1.01387882e+00
3.33231300e-01 -6.79437816e-01 -6.98474288e-01 4.99994308e-01
-3.58837843e-01 3.37738195e-04 2.55673509e-02 3.20573479e-01
-6.37248933e-01 5.93016207e-01 8.51411939e-01 -1.83382869e-01
-5.21414995e-01 4.21235234e-01 -8.69465768e-01 6.51324511e-01
7.69874871e-01 -5.52281320e-01 -2.87219644e-01 8.33722949e-02
-3.83158684e-01 -1.90028816e-01 4.23022866e-01 -1.18228292e+00
1.36604413e-01 -3.58444661e-01 8.91564727e-01 -5.00673831e-01
1.20660529e-01 -1.21367085e+00 -2.03090664e-02 3.33431602e-01
-4.25671548e-01 -4.56485391e-01 2.38579988e-01 7.76969731e-01
-1.34678453e-01 -2.11059213e-01 1.04545081e+00 2.02929243e-01
-7.23015010e-01 2.55326152e-01 9.76621136e-02 5.67098558e-02
1.44336522e+00 -3.47326934e-01 -2.86960155e-01 -9.41273198e-02
-9.58374202e-01 3.34592491e-01 4.83003855e-01 4.79835421e-01
6.71264231e-01 -1.39353466e+00 -7.16949999e-01 5.60741127e-01
5.55887938e-01 -6.48410767e-02 1.99795559e-01 5.57819426e-01
3.10184419e-01 1.36223614e-01 -1.23928450e-01 -8.27012002e-01
-1.23348975e+00 9.12006855e-01 5.60149312e-01 -2.23115698e-01
-2.95651764e-01 6.59824193e-01 4.26802635e-01 -7.69890010e-01
2.20288292e-01 -1.43954292e-01 -2.45486066e-01 5.38074300e-02
4.71478581e-01 2.71041572e-01 1.83397844e-01 -5.18043220e-01
-5.93515813e-01 6.91282272e-01 -3.55204403e-01 1.29400790e-01
9.81543064e-01 -1.36248827e-01 3.69043082e-01 5.30697823e-01
1.27221215e+00 1.43488785e-02 -1.30309689e+00 -5.84373057e-01
2.25754261e-01 -7.08644390e-01 -4.63813722e-01 -1.10623050e+00
-7.74394214e-01 6.89653933e-01 1.13453591e+00 2.29836442e-02
1.26513708e+00 1.79817930e-01 2.51365751e-01 3.32423210e-01
2.17965260e-01 -1.10853899e+00 -1.37857914e-01 3.66709083e-01
8.48528445e-01 -1.49548244e+00 -9.37647596e-02 -4.03650761e-01
-7.18283117e-01 1.04462361e+00 8.54154646e-01 -6.07340224e-02
4.59804595e-01 -1.16292663e-01 5.11928909e-02 2.38310695e-01
-7.30125606e-01 -4.02786806e-02 6.57493174e-01 7.82092571e-01
-5.48511706e-02 4.13994715e-02 -1.18525773e-01 9.15692449e-01
8.29217881e-02 -1.31822333e-01 2.39538059e-01 7.89358974e-01
-3.93711686e-01 -1.17358518e+00 -5.61981797e-01 6.88342392e-01
-2.71545380e-01 2.35197961e-01 -4.64316189e-01 7.11201012e-01
3.59910727e-01 7.78547287e-01 -2.88009103e-02 -5.23303211e-01
6.62110090e-01 3.44922781e-01 2.52428889e-01 -5.70587218e-01
-2.80787289e-01 -1.20921940e-01 -2.77294338e-01 -2.66907483e-01
-1.33349866e-01 -8.46433997e-01 -1.18470585e+00 2.88345546e-01
-5.01437962e-01 1.88364163e-01 4.51516926e-01 7.77248800e-01
5.37990510e-01 1.61876500e-01 1.01649928e+00 -7.45485783e-01
-1.16891110e+00 -6.77221656e-01 -4.90973502e-01 6.86156929e-01
4.65518743e-01 -9.60998356e-01 -5.92667222e-01 -3.34749408e-02]
|
[10.223665237426758, 3.12904953956604]
|
d23bb5cf-f2bf-4eb9-8f62-a58fbd54c28a
|
data-centric-learning-from-unlabeled-graphs
|
2303.10108
| null |
https://arxiv.org/abs/2303.10108v1
|
https://arxiv.org/pdf/2303.10108v1.pdf
|
Data-Centric Learning from Unlabeled Graphs with Diffusion Model
|
Graph property prediction tasks are important and numerous. While each task offers a small size of labeled examples, unlabeled graphs have been collected from various sources and at a large scale. A conventional approach is training a model with the unlabeled graphs on self-supervised tasks and then fine-tuning the model on the prediction tasks. However, the self-supervised task knowledge could not be aligned or sometimes conflicted with what the predictions needed. In this paper, we propose to extract the knowledge underlying the large set of unlabeled graphs as a specific set of useful data points to augment each property prediction model. We use a diffusion model to fully utilize the unlabeled graphs and design two new objectives to guide the model's denoising process with each task's labeled data to generate task-specific graph examples and their labels. Experiments demonstrate that our data-centric approach performs significantly better than fourteen existing various methods on fifteen tasks. The performance improvement brought by unlabeled data is visible as the generated labeled examples unlike self-supervised learning.
|
['Meng Jiang', 'Tengfei Luo', 'Jiaxin Xu', 'Tong Zhao', 'Eric Inae', 'Gang Liu']
|
2023-03-17
| null | null | null | null |
['graph-property-prediction']
|
['graphs']
|
[ 4.02412295e-01 7.36750066e-01 -5.33342540e-01 -6.06887400e-01
-4.65698987e-01 -4.62637007e-01 4.77929503e-01 1.09537400e-01
1.45643637e-01 9.80208516e-01 2.46172696e-01 -7.28249922e-02
-5.51126786e-02 -8.35940659e-01 -6.32273793e-01 -5.67207277e-01
4.25679283e-03 8.77884150e-01 3.51293772e-01 -2.90004741e-02
3.92914116e-01 2.34210193e-01 -1.31672311e+00 3.57918501e-01
1.24691451e+00 7.87764847e-01 3.97359401e-01 4.12382662e-01
-3.47501993e-01 1.14054978e+00 -3.01866680e-01 -3.48788977e-01
3.57158601e-01 -4.52352345e-01 -7.99250424e-01 7.43006170e-01
3.87573540e-01 4.16155495e-02 -4.14736792e-02 1.21449900e+00
1.71526045e-01 1.91203043e-01 9.15068686e-01 -1.52317941e+00
-1.01870406e+00 8.06836009e-01 -8.42003822e-01 -1.76452566e-02
1.06915504e-01 5.18701673e-02 1.07103562e+00 -7.97111392e-01
9.46915567e-01 9.94398713e-01 6.28142715e-01 5.84803462e-01
-1.17076802e+00 -5.87966144e-01 3.11503828e-01 8.70942622e-02
-1.10288751e+00 -2.19964504e-01 1.34428930e+00 -5.09546041e-01
7.00486362e-01 -1.42825931e-01 5.62292099e-01 1.03084612e+00
-1.02892376e-01 6.35343730e-01 1.09307396e+00 -5.09032726e-01
2.49825895e-01 3.42347026e-01 4.46698159e-01 9.76429701e-01
4.67396379e-01 -2.10304204e-02 -5.02245188e-01 -1.28268763e-01
6.12949073e-01 3.53521667e-02 -1.08161375e-01 -6.05010450e-01
-1.06453788e+00 7.08113611e-01 4.31038439e-01 2.85825282e-01
-5.63559771e-01 -6.43376112e-02 3.35630953e-01 4.91672128e-01
9.23506081e-01 5.08494258e-01 -6.09486699e-01 3.30379814e-01
-7.35552609e-01 -2.27953836e-01 9.15690243e-01 1.27845061e+00
1.35758901e+00 2.21365213e-01 -1.17692433e-01 8.33009660e-01
5.30461490e-01 2.90433824e-01 4.36262518e-01 -6.32837296e-01
6.59328938e-01 1.10121799e+00 -7.72578418e-02 -9.54012454e-01
-2.91916698e-01 -5.45519173e-01 -8.13671529e-01 4.58335737e-03
5.40549517e-01 -3.35500568e-01 -1.25454307e+00 1.48172462e+00
4.28926736e-01 1.15727931e-01 5.13619184e-02 7.26347208e-01
6.93353474e-01 4.65445399e-01 2.78184474e-01 -3.10202748e-01
7.13424981e-01 -1.50424516e+00 -7.01910079e-01 -5.07720888e-01
8.82480741e-01 -5.97836375e-01 1.08191931e+00 3.43623906e-01
-5.85986376e-01 -7.50763834e-01 -8.99707556e-01 1.28557503e-01
-3.35194290e-01 7.53955543e-02 6.01439953e-01 7.04347849e-01
-1.12927032e+00 7.61341810e-01 -4.54320192e-01 -3.09982598e-01
5.11152625e-01 3.31686109e-01 -4.11542118e-01 -1.99458584e-01
-8.18410158e-01 6.98192060e-01 6.03053451e-01 -1.92725733e-01
-9.18582737e-01 -5.49699724e-01 -7.70231009e-01 6.06705584e-02
7.26645947e-01 -5.83764315e-01 9.32121396e-01 -1.39214969e+00
-1.04742301e+00 7.73792148e-01 -1.16825797e-01 -3.63718361e-01
4.19083923e-01 1.47924259e-01 -3.61032486e-01 2.03601979e-02
3.00027639e-01 7.20313609e-01 1.19285405e+00 -1.74070811e+00
-6.27503633e-01 -5.35861373e-01 -3.50694597e-01 3.35873872e-01
-5.11256635e-01 -3.95123541e-01 -3.26622158e-01 -5.99999368e-01
2.45791167e-01 -8.72253895e-01 -4.27747101e-01 -3.90179306e-01
-8.01689863e-01 -3.32666814e-01 9.98619020e-01 -4.21407670e-01
9.69348192e-01 -1.93444955e+00 -1.81003928e-01 4.10644650e-01
7.17930734e-01 4.32550944e-02 -3.75626981e-01 5.85033894e-01
-1.68505311e-01 3.06075215e-01 -6.82182610e-02 -2.92131007e-01
-2.00046256e-01 1.29834905e-01 -1.31852672e-01 2.42374256e-01
2.74019808e-01 9.20624971e-01 -1.18987095e+00 -7.28676081e-01
-2.83785537e-02 8.78857672e-02 -2.01844022e-01 3.57022345e-01
-6.83512688e-01 7.09292889e-01 -1.00252414e+00 6.45817399e-01
5.04876494e-01 -8.22312832e-01 5.73485792e-01 -2.06003919e-01
5.53873479e-01 2.90257987e-02 -9.93465900e-01 1.40969145e+00
-9.82936174e-02 2.78599381e-01 -3.20751697e-01 -1.14269149e+00
1.42769396e+00 1.65055081e-01 4.70931172e-01 -2.93423355e-01
-1.04501806e-01 1.45595327e-01 2.23034644e-03 -5.64605653e-01
3.13675433e-01 -2.85888284e-01 4.17378962e-01 7.93803573e-01
2.93828547e-01 -2.57777963e-02 4.10016418e-01 4.37398732e-01
1.18977571e+00 2.62552410e-01 3.23495120e-01 -2.90885866e-01
3.49307269e-01 5.28825164e-01 4.17088836e-01 5.28586149e-01
-1.87947929e-01 4.48366165e-01 4.30370301e-01 -5.49219966e-01
-1.15998971e+00 -6.80797219e-01 4.63133007e-01 1.21536791e+00
7.77989551e-02 -4.64900881e-01 -5.49005151e-01 -1.53587711e+00
-1.28844693e-01 7.10066617e-01 -8.57349932e-01 -1.01083517e-01
-1.76232755e-02 -3.74526560e-01 -4.02561203e-02 4.21194494e-01
4.89921838e-01 -1.41425931e+00 3.81885022e-01 9.68287066e-02
1.09619498e-02 -7.94984877e-01 -6.03778541e-01 3.46354246e-01
-1.19592953e+00 -1.27553689e+00 -4.07429188e-01 -1.12906623e+00
1.34578812e+00 4.93541628e-01 1.40793002e+00 2.59597123e-01
2.51694947e-01 3.09081703e-01 -5.53080559e-01 -4.70340848e-01
-7.07340598e-01 2.98315495e-01 -2.12072268e-01 1.31695241e-01
5.84987283e-01 -6.41498327e-01 -2.73513436e-01 3.28084648e-01
-7.11054802e-01 2.01246902e-01 6.40342653e-01 7.84184456e-01
7.70858526e-01 9.68970656e-02 1.11868000e+00 -1.83608902e+00
8.90817344e-01 -6.59573972e-01 -3.29648256e-01 5.53142965e-01
-1.22458732e+00 3.13158900e-01 7.96873510e-01 -5.01342714e-01
-1.21100616e+00 5.36879182e-01 4.93808419e-01 -4.87892479e-01
-2.08851486e-01 7.37171054e-01 -8.17491263e-02 -1.35831686e-03
1.11729980e+00 1.33413285e-01 1.00708589e-01 -3.34814906e-01
5.41900754e-01 4.70342785e-01 1.20556252e-02 -5.65932512e-01
1.04290462e+00 9.39847827e-02 -1.23498961e-01 -4.96690273e-01
-1.22245479e+00 -6.20848298e-01 -1.01038659e+00 -4.60719109e-01
5.69163382e-01 -7.57323980e-01 2.88361348e-02 1.05402209e-01
-8.47339988e-01 -5.74268937e-01 -6.74237132e-01 1.90801784e-01
-4.36088413e-01 4.80938286e-01 -3.49308670e-01 -7.51758337e-01
-4.26938176e-01 -8.52912962e-01 8.77719998e-01 1.45318657e-01
-1.36415884e-01 -1.43720937e+00 2.07366258e-01 5.54103136e-01
1.36591271e-01 1.24081500e-01 8.94508600e-01 -1.27919352e+00
-6.06675923e-01 -4.21337008e-01 -3.57167631e-01 4.41230506e-01
5.62182426e-01 -3.36103849e-02 -8.58076096e-01 -1.91030025e-01
-1.22153107e-03 -8.03502202e-01 6.38310671e-01 2.33247936e-01
1.00475323e+00 -2.32404530e-01 -4.32074785e-01 1.86724856e-01
1.33971500e+00 1.26786972e-03 4.27478075e-01 4.49338108e-02
1.13892353e+00 8.46127272e-01 7.94720531e-01 1.07318878e-01
3.41279238e-01 -1.12658702e-01 3.73742640e-01 -1.42555892e-01
-3.00437421e-01 -6.92706943e-01 7.17269480e-02 9.97940898e-01
-2.36677125e-01 -2.65373915e-01 -1.06677020e+00 4.23215955e-01
-2.06540728e+00 -7.85618424e-01 -4.89128143e-01 1.97370815e+00
8.11802030e-01 3.09318125e-01 8.94546732e-02 -8.93102363e-02
1.08613718e+00 1.00653827e-01 -7.82937884e-01 1.99949145e-01
1.73417807e-01 -1.86325788e-01 3.84272188e-01 3.03560585e-01
-9.22444522e-01 1.13925350e+00 6.59774590e+00 8.06330264e-01
-7.84596324e-01 -1.09276846e-01 9.22350347e-01 4.89120603e-01
-5.02725124e-01 1.88810617e-01 -6.66149974e-01 1.30726770e-01
7.43342817e-01 -3.97096127e-01 3.65570307e-01 1.21742964e+00
2.05511257e-01 1.35086581e-01 -1.15153933e+00 6.92569077e-01
5.67075796e-02 -1.35118079e+00 2.34831631e-01 1.77594274e-01
1.26085114e+00 7.20351003e-03 -3.00799161e-01 3.35713893e-01
7.99831629e-01 -8.87567461e-01 2.72526145e-01 4.21647400e-01
6.42311811e-01 -3.14769357e-01 4.85578388e-01 6.47456229e-01
-1.12291801e+00 1.73908666e-01 -5.86588740e-01 -9.65357050e-02
-2.15579435e-01 6.99545741e-01 -1.33653176e+00 5.85183799e-01
1.99241862e-01 1.15000105e+00 -9.51426327e-01 8.15581083e-01
-5.57331264e-01 1.04277635e+00 2.51074079e-02 -1.99401394e-01
1.05365053e-01 -5.95709324e-01 1.42249987e-01 7.87698090e-01
8.28601047e-02 -7.71820918e-02 6.16135836e-01 7.17294276e-01
-3.25183243e-01 5.49005449e-01 -1.06787038e+00 -3.24656427e-01
4.31487799e-01 1.54856026e+00 -9.96012270e-01 -6.17726207e-01
-6.08169794e-01 7.25189269e-01 7.51160145e-01 5.36227226e-01
-4.34345782e-01 6.85168058e-02 -4.35113221e-01 3.26962441e-01
4.52129766e-02 3.17146145e-02 -4.33271796e-01 -1.07713830e+00
-7.36899003e-02 -8.04788709e-01 4.01251733e-01 -8.25003088e-01
-1.95041013e+00 6.22090518e-01 -3.10555607e-01 -1.39802146e+00
-1.64782286e-01 -3.56154263e-01 -7.30802178e-01 7.62310266e-01
-1.37625074e+00 -1.42273068e+00 -5.57440400e-01 7.16925144e-01
5.68959534e-01 -5.37233055e-01 6.56314611e-01 -2.20560413e-02
-3.96209121e-01 1.26103789e-01 -2.81194802e-02 1.16137147e-01
9.52710986e-01 -1.47164178e+00 3.25340539e-01 6.97918236e-01
4.65685010e-01 3.51560056e-01 5.61553478e-01 -1.34254408e+00
-1.25287914e+00 -1.26479244e+00 8.26729715e-01 -3.12726617e-01
9.82940674e-01 -8.14762488e-02 -1.07703269e+00 9.79222238e-01
2.13930666e-01 2.80068368e-01 7.90235400e-01 2.59813458e-01
-3.10055345e-01 -5.65144829e-02 -1.08375704e+00 3.68101597e-01
1.23258567e+00 -2.73098826e-01 -5.70882678e-01 8.10958564e-01
7.01278925e-01 8.80754218e-02 -7.11401939e-01 2.58463889e-01
-9.80477184e-02 -7.31422544e-01 4.90331173e-01 -9.77623403e-01
5.24064839e-01 -1.69461951e-01 1.85848147e-01 -1.49836469e+00
-4.28568363e-01 -4.18018878e-01 -7.08288997e-02 1.22856176e+00
9.56380188e-01 -6.24442935e-01 1.47044444e+00 8.16230595e-01
-1.94009587e-01 -5.06747246e-01 -1.68481320e-01 -6.22658432e-01
-2.93768048e-01 -1.91412404e-01 2.88142085e-01 1.40587950e+00
3.41806002e-02 1.02488101e+00 -4.59915996e-01 1.47842653e-02
8.72209549e-01 3.43547851e-01 1.01493549e+00 -1.55685937e+00
-1.88752651e-01 1.43385619e-01 -8.89774263e-02 -8.57841730e-01
3.03457111e-01 -1.23801816e+00 -6.84345188e-03 -1.85546637e+00
4.09806252e-01 -6.46110296e-01 -2.21224472e-01 8.49163771e-01
-5.61333179e-01 -5.55437319e-02 -1.58443183e-01 5.33541679e-01
-8.03439438e-01 4.76490200e-01 1.69879937e+00 -3.76578122e-01
-3.88194859e-01 9.88446623e-02 -8.10568750e-01 7.10300148e-01
9.54520822e-01 -5.96332490e-01 -1.15572751e+00 -1.28702536e-01
1.04342483e-01 -4.27568369e-02 -1.51183441e-01 -7.52412498e-01
3.25327694e-01 -2.44250551e-01 4.38978195e-01 -4.00407255e-01
-7.41838589e-02 -9.45420146e-01 2.53383189e-01 2.41338775e-01
-3.58185083e-01 -2.64971018e-01 -2.85734981e-01 1.08500826e+00
-2.53673375e-01 -4.27634627e-01 5.37416935e-01 -2.65773237e-01
-8.99395466e-01 6.29512429e-01 -2.51990147e-02 1.78833634e-01
1.23664951e+00 -3.96423310e-01 -5.43482363e-01 -5.13293982e-01
-1.09178340e+00 3.10373902e-01 5.15320003e-01 2.65774310e-01
6.43486142e-01 -1.20327151e+00 -6.24784231e-01 8.43903869e-02
2.41683185e-01 2.37527210e-02 9.95760486e-02 3.89006913e-01
-2.52402008e-01 -1.50853708e-01 -3.09146166e-01 -4.21177953e-01
-1.02251244e+00 1.05865002e+00 -3.98860872e-02 -7.94434965e-01
-5.02792239e-01 5.74046969e-01 2.14181289e-01 -6.14017367e-01
5.75489029e-02 1.11148342e-01 -4.91713643e-01 1.05829770e-02
9.53010246e-02 1.46501169e-01 -1.61882088e-01 -2.21317902e-01
1.92292899e-01 2.68734932e-01 -1.41092628e-01 2.52708882e-01
1.55673492e+00 -5.92348278e-02 -6.76037595e-02 4.02264029e-01
8.68418097e-01 6.40510991e-02 -1.52257514e+00 -5.54914773e-01
4.49905306e-01 -3.44057381e-01 -2.80158460e-01 -6.59803331e-01
-1.41552579e+00 5.56961358e-01 5.66167794e-02 5.91409802e-01
9.82903600e-01 1.31913468e-01 3.44618082e-01 5.69285870e-01
4.66778517e-01 -1.10490131e+00 6.05182707e-01 1.67307481e-01
7.56599844e-01 -1.64539313e+00 1.60836563e-01 -1.17750847e+00
-1.10057139e+00 1.08590496e+00 9.54218149e-01 -2.41069704e-01
7.42512643e-01 1.08313989e-02 1.24681368e-01 -6.21408880e-01
-7.37810671e-01 -2.57328361e-01 5.13998508e-01 1.05545223e+00
5.49294353e-01 -2.57290658e-02 -1.92999810e-01 4.97328192e-01
1.00965530e-01 8.45463425e-02 2.85714686e-01 8.76639605e-01
-7.14572430e-01 -1.42058325e+00 -5.23638949e-02 1.14638102e+00
1.83174461e-02 -1.39188126e-01 -8.15311790e-01 6.06675088e-01
-2.72084147e-01 1.02658534e+00 -4.33942050e-01 -6.85340822e-01
8.04063305e-02 3.27670515e-01 1.92289531e-01 -1.21351635e+00
-4.73092288e-01 1.42866045e-01 4.16350633e-01 -1.56321198e-01
-5.85891187e-01 -2.35779151e-01 -1.20821869e+00 -1.19532095e-02
-5.16310811e-01 2.66936153e-01 2.85328865e-01 8.25793207e-01
3.05710673e-01 5.05187273e-01 7.42222548e-01 -7.09156930e-01
-5.82829654e-01 -1.18552923e+00 -9.54382598e-01 7.50936806e-01
-2.62128055e-01 -5.10848522e-01 -3.01832050e-01 5.45036077e-01]
|
[7.374345302581787, 6.149233818054199]
|
3e8450d7-ad06-462a-9984-2111129d30ec
|
milan-sky-survey-a-dataset-of-raw-deep-sky
| null | null |
https://www.sciencedirect.com/science/article/pii/S2352340923002524
|
https://www.sciencedirect.com/science/article/pii/S2352340923002524/pdfft?md5=64b4209de72fd1893bd8da68fa7fa5cd&pid=1-s2.0-S2352340923002524-main.pdf
|
MILAN Sky Survey, a dataset of raw deep sky images captured during one year with a Stellina automated telescope
|
Modern automated telescopes allow to capture astronomical images in a reproducible way. During the MILAN research project (MachIne Learning for AstroNomy), we have observed deep sky with a Stellina observation station for twelve months from the Luxembourg Greater Region. Thus, we have captured raw images of more than 188 deep sky objects visible from the Northern Hemisphere (galaxies, stars clusters, nebulae, etc.), We have compiled and published this data as the MILAN Sky Survey dataset, allowing interested researchers, industry practitioners and citizens to reuse it.
|
['Benoît Vandame', 'Christophe Destruel', 'Gilles Krebs', 'Pierrick Bruneau', 'Patrik Hitzelberger', 'Olivier Parisot']
|
2023-04-11
| null | null | null |
data-in-brief-2023-4
|
['astronomy']
|
['miscellaneous']
|
[-3.93876940e-01 -9.78167802e-02 1.45191401e-01 -1.18056804e-01
-8.57758150e-02 -9.74367201e-01 1.16115165e+00 -4.92652357e-01
-5.78382671e-01 6.43351316e-01 -1.24477901e-01 -3.05283129e-01
7.70326555e-02 -6.13408446e-01 -2.80113846e-01 -1.02481234e+00
1.08134173e-01 6.15203559e-01 3.35129529e-01 2.73550004e-01
2.89970458e-01 1.00549698e+00 -1.39464235e+00 -1.10302055e-02
3.96579355e-01 7.80886590e-01 6.51347935e-01 8.74043703e-01
3.39304715e-01 5.23086846e-01 -5.18515170e-01 -2.28004903e-01
8.62733305e-01 -4.07803029e-01 -8.74487221e-01 1.77063406e-01
9.20335948e-01 -1.37535304e-01 -4.98549163e-01 1.12695837e+00
1.65759861e-01 1.77153647e-01 6.02856219e-01 -7.66381085e-01
-2.43790701e-01 -2.97239333e-01 -3.46489221e-01 6.45705104e-01
-4.87282649e-02 3.28958899e-01 1.22702181e+00 -7.87153661e-01
1.02073407e+00 9.44709063e-01 2.84452498e-01 -7.83245042e-02
-7.78675675e-01 -9.89230573e-02 -5.76408863e-01 5.60052037e-01
-1.23863733e+00 -2.49247313e-01 2.47286782e-01 -6.08062446e-01
8.45974743e-01 4.56410348e-01 8.65053594e-01 8.77313495e-01
-4.13521416e-02 4.36838925e-01 1.41901255e+00 -5.70735455e-01
2.05650762e-01 5.41093528e-01 -1.32780507e-01 3.51919770e-01
4.18405622e-01 4.43066031e-01 -1.35383084e-01 -2.53366321e-01
1.04203892e+00 6.65088966e-02 -3.90771985e-01 -2.33948752e-01
-1.58182216e+00 9.43117321e-01 7.25386083e-01 5.98527074e-01
-5.26089549e-01 -2.20555454e-01 -1.67972878e-01 5.19735754e-01
1.18356757e-01 8.00208151e-01 -6.76103592e-01 -8.28948393e-02
-6.27677798e-01 3.79364491e-01 7.86506712e-01 6.28195107e-01
7.45225370e-01 -7.14081079e-02 5.60585141e-01 4.62484032e-01
2.70094901e-01 6.91018343e-01 5.02356470e-01 -1.21882260e+00
2.33422481e-02 3.41836870e-01 5.26081383e-01 -6.72960520e-01
-7.11603761e-01 -4.64125752e-01 -6.92781866e-01 6.81036294e-01
6.02832437e-01 -2.05838010e-01 -4.05603349e-01 7.21922100e-01
5.17356515e-01 -1.81865483e-03 1.22294478e-01 1.30585587e+00
9.84441102e-01 4.41410124e-01 -6.90179527e-01 2.73936763e-02
1.47425699e+00 -1.13480854e+00 -1.38629600e-01 7.13849301e-03
4.60191131e-01 -7.96930730e-01 6.70964956e-01 9.71505880e-01
-7.47139037e-01 -7.29404747e-01 -5.75519025e-01 8.01628232e-02
-2.44279370e-01 3.24747622e-01 1.02996111e+00 4.88978982e-01
-8.05915058e-01 5.20018458e-01 -5.05896330e-01 -5.20073116e-01
2.17747182e-01 -2.28989750e-01 -4.45525438e-01 4.71112639e-01
-6.27228320e-01 6.10679567e-01 4.98109251e-01 -2.35460937e-01
-9.98363435e-01 -5.23755074e-01 -1.18568107e-01 1.24222741e-01
1.04989059e-01 -6.96695566e-01 1.36007619e+00 -9.53453898e-01
-1.33468068e+00 1.30006230e+00 1.49511486e-01 -6.48445010e-01
1.17846601e-01 -9.37803313e-02 -7.43840873e-01 2.75115222e-01
-4.44905847e-01 2.72383928e-01 6.15862787e-01 -6.48978293e-01
-1.04929817e+00 -4.65951890e-01 -3.85113582e-02 -1.07587352e-01
1.51230870e-02 4.83476311e-01 -7.57788941e-02 -4.30646509e-01
1.43489882e-01 -1.13997483e+00 -1.13450333e-01 -1.21415146e-01
-6.85483888e-02 -3.74606162e-01 5.08875132e-01 -6.97282076e-01
2.33267054e-01 -2.28543496e+00 1.62838981e-01 -5.61884977e-02
3.36983830e-01 3.66579562e-01 2.04891026e-01 2.85107881e-01
-1.59291431e-01 -4.63823497e-01 1.87656824e-02 5.92796318e-02
-2.83036798e-01 2.35139623e-01 -3.10650170e-01 8.04140925e-01
-6.12326324e-01 6.30868733e-01 -6.40475392e-01 5.57769388e-02
4.88797605e-01 -1.78250283e-01 -1.34918839e-01 2.70406932e-01
-3.98958325e-01 8.95352900e-01 -5.40362000e-01 4.28903252e-01
9.08404648e-01 -3.53762031e-01 -2.13801846e-01 3.02465111e-01
-8.55021715e-01 5.47956109e-01 -8.44304264e-01 1.72970581e+00
-2.75159031e-01 1.19753993e+00 3.24714661e-01 -6.24530971e-01
9.54192221e-01 3.08187574e-01 2.00010344e-01 -6.45910442e-01
1.07868813e-01 1.52387083e-01 1.99128520e-02 -6.14720225e-01
5.36903203e-01 -2.13884443e-01 6.65412128e-01 1.52720034e-01
1.33708149e-01 -5.01272202e-01 7.53531754e-02 3.49336751e-02
1.01692975e+00 -1.00838661e-01 3.23323607e-01 -3.36291343e-01
4.26978678e-01 5.61364830e-01 1.60624653e-01 6.18415058e-01
-8.77835229e-02 9.38439190e-01 7.54290894e-02 -9.22396958e-01
-1.53450656e+00 -1.03982556e+00 -6.19209170e-01 7.12935865e-01
-3.75634670e-01 4.93143685e-02 -3.78435910e-01 -3.48016381e-01
-1.91760540e-01 3.98612052e-01 -2.11017683e-01 3.79667252e-01
-1.17632285e-01 -5.37924767e-01 1.12929285e-01 -4.94434685e-02
7.94090211e-01 -1.34329748e+00 -6.17121816e-01 -3.80073071e-01
2.18421727e-01 -1.02221239e+00 2.43098542e-01 -2.11697280e-01
-7.38086045e-01 -1.08432829e+00 -1.02779448e+00 -4.77570176e-01
8.50975960e-02 7.64157116e-01 1.43715334e+00 -1.98883951e-01
-8.78399730e-01 7.98627138e-02 -3.27263921e-01 -6.86835766e-01
-1.94025591e-01 -5.67243434e-02 6.67486861e-02 -1.44693404e-02
2.33914092e-01 -5.39614439e-01 -5.27265847e-01 1.99677169e-01
-4.17002439e-01 -1.16411209e-01 5.88718295e-01 3.26525241e-01
2.76658028e-01 1.12699546e-01 1.08788088e-01 -8.35987806e-01
-4.48459476e-01 -4.42171901e-01 -1.81118751e+00 -3.72143090e-02
-3.11515629e-01 -3.97791624e-01 4.96759892e-01 2.43503124e-01
-1.31456697e+00 6.95633190e-03 -5.08442856e-02 -3.01095486e-01
-8.10231507e-01 -3.49477120e-02 1.41054362e-01 -4.19228226e-01
9.39403772e-01 1.74419194e-01 -5.44025719e-01 -1.03786552e+00
4.54861343e-01 9.66526330e-01 1.24216533e+00 6.81890473e-02
9.92682099e-01 1.05063760e+00 2.52325654e-01 -1.54546976e+00
-9.65168834e-01 -1.24338996e+00 -7.84104764e-01 -3.81613448e-02
7.96493173e-01 -1.19307625e+00 -6.38538122e-01 4.61127520e-01
-1.02013195e+00 6.38798550e-02 -4.95064378e-01 1.08829331e+00
-1.79745778e-01 2.87667662e-01 4.01140265e-02 -1.00837409e+00
-1.40917227e-01 -1.30285859e-01 9.88065124e-01 6.38248801e-01
9.82408896e-02 -1.17191899e+00 6.79935515e-01 3.02891612e-01
1.11544341e-01 1.93045005e-01 2.14674488e-01 -3.07776570e-01
-1.06459963e+00 -1.23199411e-01 -4.42694932e-01 4.54121739e-01
-9.45806652e-02 2.71615386e-02 -1.29526913e+00 -4.18726176e-01
7.07278967e-01 -1.53006241e-01 1.01974654e+00 3.99553895e-01
1.19612277e+00 -2.80767009e-02 -1.83838442e-01 8.73371840e-01
1.37136412e+00 7.06929266e-02 6.93468809e-01 6.62354887e-01
3.93276721e-01 4.34343070e-01 4.65980500e-01 4.89622772e-01
-2.05384940e-01 5.90038717e-01 6.37515068e-01 -1.85848787e-01
-8.52201506e-02 3.83869350e-01 -2.38138884e-02 2.49888062e-01
-5.08649588e-01 -4.02821563e-02 -6.34636462e-01 9.06933963e-01
-1.39646518e+00 -1.00443196e+00 -1.01996291e+00 2.50773692e+00
1.97495267e-01 -2.29388386e-01 1.40167654e-01 -3.15678358e-01
2.03572243e-01 1.65741652e-01 4.89312876e-03 -8.70904699e-02
-2.60730982e-01 2.11551979e-01 7.12326884e-01 2.12372378e-01
-8.35825324e-01 5.03960848e-01 7.32202387e+00 6.08190179e-01
-1.12915695e+00 1.69283271e-01 -4.83959913e-02 -2.99004674e-01
-3.60488296e-01 2.60176629e-01 -7.35277176e-01 1.65319234e-01
1.10650969e+00 1.61224173e-03 6.68562889e-01 9.74314392e-01
2.80095011e-01 -4.03030246e-01 -4.95087415e-01 8.39296639e-01
-4.10691351e-01 -1.50086164e+00 -7.06801116e-01 5.29324234e-01
1.09098566e+00 1.05985546e+00 -2.87599623e-01 -1.50863692e-01
2.10058123e-01 -7.20763683e-01 4.28744733e-01 8.77244771e-01
3.71347934e-01 -4.13172096e-01 5.18326640e-01 6.75116122e-01
-4.16095287e-01 2.76953075e-02 -1.08972061e+00 -5.01640677e-01
-1.27300739e-01 1.11943519e+00 -1.09220386e+00 7.75324762e-01
1.03152442e+00 5.98010421e-01 -6.32635534e-01 1.52103043e+00
-5.21739662e-01 8.65076959e-01 -5.25509536e-01 2.97005114e-04
4.66134191e-01 -7.17685759e-01 8.32040250e-01 6.71833634e-01
2.62778848e-01 2.14207787e-02 -2.66624689e-01 9.34835196e-01
-1.42274931e-01 -2.37330928e-01 -5.82043827e-01 -7.41657987e-02
-9.09254774e-02 1.89573550e+00 -4.01918501e-01 -3.73296797e-01
-8.23614240e-01 5.99008918e-01 -5.43788746e-02 3.18857372e-01
-2.91514367e-01 -2.23744854e-01 8.20857346e-01 -2.32546851e-02
4.49709207e-01 -4.84902680e-01 -4.75998342e-01 -1.05816972e+00
-2.27934629e-01 -4.49335873e-01 4.10129726e-01 -1.43614495e+00
-9.33209002e-01 6.24683261e-01 -1.97238743e-01 -9.27082241e-01
-1.31580397e-01 -9.27957535e-01 -9.50297594e-01 1.06570792e+00
-1.33733988e+00 -5.81346035e-01 -5.51798820e-01 4.06659991e-01
3.54507923e-01 -8.25535417e-01 7.07536697e-01 -7.13893538e-03
-2.64374733e-01 -5.28497756e-01 9.35599029e-01 -3.22220743e-01
3.85629386e-01 -1.43325019e+00 8.08179140e-01 7.67354012e-01
9.70585525e-01 6.43837005e-02 8.44957113e-01 -1.39774904e-01
-1.04464424e+00 -8.70175421e-01 1.14502609e+00 -6.43518806e-01
8.53491962e-01 -1.28178999e-01 -8.62825155e-01 8.08951855e-01
7.11050570e-01 1.01592895e-02 1.06800146e-01 -1.07888371e-01
-1.40980765e-01 1.37412354e-01 -9.10162866e-01 1.99306905e-01
5.78585446e-01 -4.25043494e-01 -9.10454154e-01 1.05005312e+00
3.92742455e-01 -5.56232110e-02 -7.02502131e-01 2.32805312e-02
2.44194940e-01 -1.32834291e+00 8.91052067e-01 -4.60038871e-01
9.27832723e-03 -3.00120264e-01 3.18470508e-01 -1.28808653e+00
-4.90184814e-01 -8.18120718e-01 1.20945357e-01 6.65371358e-01
1.50208235e-01 -9.18281019e-01 7.18242049e-01 -9.15854350e-02
-2.20231324e-01 3.50058377e-01 -1.05979240e+00 -1.12696362e+00
1.58762038e-02 -1.84034601e-01 5.64744949e-01 7.97823846e-01
-4.98451859e-01 5.08855134e-02 -3.12892437e-01 5.02919257e-01
5.87545693e-01 8.18876684e-01 1.20323920e+00 -1.64597559e+00
-8.00088763e-01 -4.30950791e-01 -3.18396866e-01 -1.13188088e+00
-1.96688160e-01 -7.28650212e-01 -2.00827539e-01 -1.21170294e+00
2.09816679e-01 -1.01115793e-01 8.42801630e-02 1.79145440e-01
2.90865719e-01 4.52629447e-01 -1.47368863e-01 7.38665938e-01
-3.38231832e-01 7.74366707e-02 1.30285668e+00 4.01551247e-01
4.77945149e-01 4.28791881e-01 -1.50826186e-01 8.86633933e-01
8.20954144e-01 -4.78483856e-01 4.40826565e-01 -3.72755647e-01
2.14406952e-01 -2.24090338e-01 7.54476488e-01 -1.19315529e+00
-9.63753015e-02 -3.62098157e-01 3.34453851e-01 -7.29701757e-01
4.00780559e-01 -8.14665258e-01 3.37334573e-01 4.50806201e-01
3.10607731e-01 -6.14425242e-01 -5.30077517e-02 2.24416330e-01
-3.57125163e-01 -8.02157819e-01 9.95779514e-01 -5.14329731e-01
-7.30863571e-01 1.79118127e-01 -4.79733139e-01 -1.18387394e-01
9.01985228e-01 4.02331620e-01 -5.85674524e-01 -2.17236698e-01
-2.65820205e-01 -1.35764211e-01 7.34695137e-01 3.17950070e-01
-1.22810215e-01 -9.58579004e-01 -8.69795918e-01 5.23765624e-01
1.71256498e-01 -6.10487238e-02 1.44258589e-01 8.01726937e-01
-1.08249438e+00 1.00603402e+00 -4.36409265e-01 -5.47069669e-01
-1.28079438e+00 6.07002795e-01 3.27023953e-01 7.96790943e-02
-1.02707815e+00 1.01978672e+00 4.06401694e-01 -4.00113523e-01
-1.44586906e-01 -2.10100263e-01 -2.81395257e-01 -2.27751657e-01
8.73096645e-01 5.53881407e-01 2.92469323e-01 -4.24996555e-01
1.24591403e-02 1.37512475e-01 4.33312982e-01 3.55025977e-02
1.24619961e+00 -1.57951042e-01 -2.67863542e-01 4.01640415e-01
1.00545764e+00 1.36272132e-01 -1.17861056e+00 -4.32534009e-01
4.02869098e-02 -8.61596048e-01 3.12948525e-01 -6.19961441e-01
-1.00969672e+00 8.45894337e-01 6.20240748e-01 8.19379330e-01
1.02244842e+00 4.62778300e-01 5.74849010e-01 8.25541317e-01
4.24156636e-01 -7.63559759e-01 -4.99116302e-01 4.77527589e-01
9.31951642e-01 -1.35700083e+00 1.12625010e-01 -1.13908321e-01
-3.49345982e-01 1.28313923e+00 -1.39056146e-01 -2.67539620e-01
2.29196697e-01 -3.80479544e-01 1.54003695e-01 -5.87942064e-01
-7.87921548e-01 -4.14173156e-01 4.90000814e-01 6.83965087e-01
1.44021168e-01 2.49649897e-01 -3.23652864e-01 -1.51122510e-01
-5.14397860e-01 -1.50979623e-01 7.38369644e-01 3.99626404e-01
-9.47749317e-01 -9.66177523e-01 -9.38027740e-01 2.51711249e-01
-2.98296690e-01 -7.95568526e-02 -7.00569212e-01 6.80963278e-01
9.43186600e-03 7.27919698e-01 3.18771839e-01 4.31747168e-01
-3.01769301e-02 -1.59433663e-01 5.69385946e-01 -5.93065143e-01
-5.25608659e-01 5.65511845e-02 2.16761395e-01 -1.49939045e-01
-3.99369836e-01 -9.75539863e-01 -9.53827381e-01 -2.82489926e-01
-2.14581400e-01 4.18557793e-01 1.14969838e+00 8.91837180e-01
-1.40283719e-01 3.25006098e-01 1.15294242e+00 -1.04815590e+00
-2.45540932e-01 -1.27608955e+00 -1.21696711e+00 2.30308482e-03
3.34625006e-01 -3.67145956e-01 -7.30169594e-01 7.46078342e-02]
|
[7.679409027099609, 3.0558016300201416]
|
26dd6ef3-5198-436c-809e-70940332b7a7
|
unified-interactive-image-matting
|
2205.08324
| null |
https://arxiv.org/abs/2205.08324v2
|
https://arxiv.org/pdf/2205.08324v2.pdf
|
Unified Interactive Image Matting
|
Recent image matting studies are developing towards proposing trimap-free or interactive methods for complete complex image matting tasks. Although avoiding the extensive labors of trimap annotation, existing methods still suffer from two limitations: (1) For the single image with multiple objects, it is essential to provide extra interaction information to help determining the matting target; (2) For transparent objects, the accurate regression of alpha matte from RGB image is much more difficult compared with the opaque ones. In this work, we propose a Unified Interactive image Matting method, named UIM, which solves the limitations and achieves satisfying matting results for any scenario. Specifically, UIM leverages multiple types of user interaction to avoid the ambiguity of multiple matting targets, and we compare the pros and cons of different annotation types in detail. To unify the matting performance for transparent and opaque objects, we decouple image matting into two stages, i.e., foreground segmentation and transparency prediction. Moreover, we design a multi-scale attentive fusion module to alleviate the vagueness in the boundary region. Experimental results demonstrate that UIM achieves state-of-the-art performance on the Composition-1K test set and a synthetic unified dataset. Our code and models will be released soon.
|
['Stephen D. H. Yang', 'Conghui He', 'Yiqi Lin', 'Weijia Li', 'Bin Wang']
|
2022-05-17
| null | null | null | null |
['transparent-objects', 'image-matting', 'foreground-segmentation']
|
['computer-vision', 'computer-vision', 'computer-vision']
|
[ 3.59852552e-01 2.52583846e-02 5.12170978e-02 -3.99399817e-01
-6.04023635e-01 -4.67761427e-01 2.25485906e-01 -5.33672869e-01
-7.24394321e-02 3.41642708e-01 -1.65455624e-01 -3.75603825e-01
1.69805348e-01 -5.83128691e-01 -8.16161036e-01 -8.30659091e-01
5.62004805e-01 4.56609666e-01 5.36871672e-01 -5.71163893e-02
-4.88217324e-02 2.34272126e-02 -1.26006258e+00 6.27866268e-01
1.35331535e+00 1.05973423e+00 5.58028221e-01 4.31225300e-01
-5.16489327e-01 6.17204726e-01 -5.71993411e-01 -6.34918988e-01
5.39518714e-01 -3.68430078e-01 -7.50182509e-01 5.12617111e-01
5.96405149e-01 -5.25628924e-01 -1.49054840e-01 1.08742726e+00
8.48683044e-02 -1.78847179e-01 5.38666010e-01 -1.41920424e+00
-6.04167700e-01 6.74652338e-01 -1.11626232e+00 -1.60090789e-01
1.01962559e-01 3.71850938e-01 6.50475442e-01 -9.47330117e-01
2.48062685e-01 1.21479309e+00 4.95831877e-01 3.83331060e-01
-1.12040126e+00 -6.22891545e-01 5.54897904e-01 6.08494021e-02
-1.26928079e+00 -3.87847543e-01 7.46154666e-01 -4.79137272e-01
2.52943307e-01 5.18792450e-01 7.30805635e-01 8.09125721e-01
7.74132088e-02 1.16925728e+00 1.35637152e+00 -1.43246248e-01
-2.76575852e-02 9.69923735e-02 5.09408750e-02 7.61935294e-01
2.62967080e-01 -4.86229181e-01 -4.68759358e-01 2.78081089e-01
9.03648376e-01 1.54025080e-02 -4.50892359e-01 -3.57678443e-01
-1.40670848e+00 3.75398606e-01 4.34622437e-01 -1.72511395e-02
-5.57322353e-02 5.48270307e-02 3.90424654e-02 3.02418936e-02
6.14328563e-01 9.16949511e-02 -3.86665821e-01 -3.57603282e-03
-1.10470223e+00 -1.72368940e-02 4.45816100e-01 1.21623278e+00
8.76901686e-01 -2.73724161e-02 -2.55059510e-01 7.57240057e-01
4.80556428e-01 5.04632354e-01 1.21289745e-01 -8.86011541e-01
7.93973744e-01 9.57794070e-01 2.45255038e-01 -7.07031965e-01
-1.76954150e-01 -1.46941051e-01 -7.97882557e-01 3.00258279e-01
6.81111991e-01 -4.02483493e-02 -1.39471674e+00 1.23143888e+00
4.73948658e-01 1.17078453e-01 -1.99881077e-01 1.21906769e+00
1.05545616e+00 7.53223598e-01 -8.70161653e-02 -6.18319884e-02
1.45914960e+00 -1.42516434e+00 -9.27219689e-01 -5.38712621e-01
2.63978958e-01 -1.00487447e+00 1.42699087e+00 4.61220264e-01
-1.28654587e+00 -4.87959415e-01 -1.06211877e+00 -2.31299177e-01
-2.43401363e-01 3.28684568e-01 8.62796009e-01 6.98181093e-01
-8.60296369e-01 1.93278819e-01 -9.08381522e-01 -9.11493972e-02
4.58675146e-01 3.19288760e-01 -8.02910849e-02 -6.96899742e-02
-8.14229429e-01 6.99706674e-01 4.02661264e-01 4.85344291e-01
-8.17284942e-01 -6.36864781e-01 -8.12291086e-01 -1.96238652e-01
7.55101979e-01 -6.95295274e-01 1.15379703e+00 -1.29548824e+00
-1.41605031e+00 8.40202153e-01 -1.46503478e-01 -9.55978483e-02
8.57443094e-01 -6.16765916e-01 -1.07744567e-01 -3.70608047e-02
8.39728415e-02 6.78795934e-01 8.97871256e-01 -1.83797026e+00
-4.95402634e-01 -8.65052566e-02 1.91169798e-01 4.02410477e-01
-1.88145831e-01 -1.06289119e-01 -1.16146386e+00 -7.90520310e-01
4.25008953e-01 -8.09234500e-01 -7.65908509e-02 1.11534566e-01
-7.75496125e-01 2.45816261e-01 1.07147455e+00 -7.23117530e-01
1.21133375e+00 -2.04346251e+00 2.94217974e-01 -1.74085684e-02
3.52434963e-01 2.71809042e-01 1.82128102e-02 3.75308632e-03
1.72602177e-01 1.16252713e-01 -4.41687167e-01 -5.86238801e-01
-1.19039975e-01 3.60892594e-01 -4.09727126e-01 2.10123211e-01
1.25787318e-01 1.05271304e+00 -6.67120337e-01 -8.24445307e-01
3.76547694e-01 2.28599012e-01 -2.86810786e-01 3.00755322e-01
-3.76931161e-01 5.36723435e-01 -3.29392195e-01 1.05518234e+00
1.07188463e+00 -3.75517249e-01 -8.75579368e-04 -5.88549793e-01
-1.37548238e-01 1.11784339e-01 -1.32370281e+00 1.79835749e+00
-2.43899062e-01 4.40125227e-01 3.73591423e-01 -4.59529400e-01
6.16098106e-01 3.87231186e-02 4.34515804e-01 -4.16206121e-01
1.60854101e-01 -1.57726798e-02 -3.88772637e-02 -5.28725863e-01
5.97221971e-01 2.41485432e-01 2.22646534e-01 3.98140907e-01
-2.33257040e-01 -1.92051619e-01 4.79989648e-02 2.92129099e-01
7.58883536e-01 5.06338477e-01 -1.85428351e-01 -1.48840010e-01
6.46668077e-02 3.34285870e-02 8.11963081e-01 7.34288871e-01
-1.54129997e-01 1.10275292e+00 3.28335464e-01 -2.54027843e-01
-7.14070559e-01 -1.24979508e+00 -6.28792401e-03 1.03928113e+00
1.00781012e+00 -4.96356666e-01 -1.09473777e+00 -6.60010695e-01
-2.39022896e-01 4.78876680e-01 -6.73680127e-01 1.35243744e-01
-6.18389547e-01 -1.10209441e+00 2.68460453e-01 5.35642862e-01
9.32812393e-01 -8.92457306e-01 -4.13169324e-01 -2.10808977e-01
-5.69416583e-01 -1.26858175e+00 -7.57278562e-01 3.85300629e-02
-8.23677659e-01 -1.08735764e+00 -5.96156418e-01 -5.73366582e-01
7.77509511e-01 7.98625290e-01 1.14198411e+00 2.52175570e-01
-1.24208361e-01 2.36577123e-01 -4.04742628e-01 -3.70100260e-01
-1.99574068e-01 1.60017377e-03 -2.98440248e-01 2.90819973e-01
-8.25003684e-02 -2.34539181e-01 -8.34806561e-01 7.26414382e-01
-1.09367085e+00 9.99997914e-01 7.98028469e-01 5.23641050e-01
6.10078931e-01 3.32238115e-02 6.76925778e-02 -1.07450640e+00
2.11625304e-02 -2.75258034e-01 -4.83900845e-01 5.56606531e-01
-4.96412247e-01 -2.03151792e-01 1.26973838e-01 -5.79376817e-01
-1.40178859e+00 2.20668316e-01 2.37036824e-01 -3.59103918e-01
3.28693315e-02 1.78129792e-01 -7.26435602e-01 -1.25407472e-01
2.25942105e-01 1.79490969e-01 -1.55970320e-01 -5.12823284e-01
5.22721887e-01 5.35788298e-01 5.70267081e-01 -6.53246701e-01
1.06349373e+00 5.59637904e-01 -4.64417607e-01 -5.01132071e-01
-1.06184912e+00 -3.17404360e-01 -7.80182123e-01 -5.30726671e-01
1.18144441e+00 -9.96906817e-01 -5.52476048e-01 7.85117686e-01
-1.07280052e+00 -7.03072309e-01 7.80804157e-02 8.49837139e-02
-2.64779001e-01 5.97683966e-01 -8.87657166e-01 -6.96713388e-01
-3.59630674e-01 -1.61239576e+00 1.24528313e+00 5.06031752e-01
9.45680737e-02 -5.64547598e-01 -5.87355375e-01 1.04345810e+00
2.53126115e-01 1.25108168e-01 7.35147595e-01 1.23885885e-01
-1.10686433e+00 2.02852905e-01 -6.27548993e-01 7.90919960e-02
3.55818927e-01 2.66542405e-01 -1.18066013e+00 2.62016319e-02
-1.36924252e-01 -1.60732940e-01 9.76990819e-01 2.41805926e-01
1.35871303e+00 -1.20617002e-01 -2.66032547e-01 9.17187154e-01
1.01957989e+00 1.59057841e-01 8.87951910e-01 5.49142718e-01
1.34199286e+00 5.60598493e-01 1.02359080e+00 1.49120942e-01
4.74002123e-01 7.43983209e-01 6.10522449e-01 -5.04355311e-01
-2.41716295e-01 -6.13892637e-02 3.83464962e-01 6.65288806e-01
-1.46543697e-01 -3.27021062e-01 -6.46861494e-01 2.08538786e-01
-2.16040921e+00 -4.86653894e-01 -4.95433033e-01 2.03371620e+00
1.03302097e+00 2.69000828e-01 4.53525893e-02 -1.12059943e-01
6.29595935e-01 8.12218934e-02 -5.91339350e-01 1.36735752e-01
-2.63284832e-01 -1.83255747e-01 5.12083411e-01 3.94600868e-01
-1.19950843e+00 1.14283395e+00 5.93415260e+00 9.81958687e-01
-9.11435187e-01 8.90570581e-02 1.00399315e+00 1.99757069e-02
-3.68644416e-01 1.07693084e-01 -6.28037572e-01 7.06339121e-01
1.53533742e-01 4.87971365e-01 4.82363880e-01 6.01640701e-01
1.81132972e-01 -5.29738784e-01 -9.03682530e-01 1.12592745e+00
-7.36014768e-02 -1.06322050e+00 1.75470650e-01 -2.33865753e-01
6.64021313e-01 -2.94226319e-01 9.84131843e-02 1.89631823e-02
9.88444686e-02 -9.89474952e-01 1.15575838e+00 3.91495109e-01
6.43918157e-01 -2.99384475e-01 5.19954264e-01 1.59752280e-01
-1.33659005e+00 2.90479749e-01 -1.60584435e-01 1.15038738e-01
2.05062479e-01 7.99939454e-01 -5.91166317e-01 6.81132019e-01
8.16688776e-01 4.43820417e-01 -8.15444946e-01 1.13124800e+00
-2.70442724e-01 7.12860823e-01 -3.06818515e-01 5.53135693e-01
5.02000228e-02 -6.62670672e-01 3.04154575e-01 1.14812839e+00
9.08120647e-02 -4.60744202e-02 3.43179464e-01 1.15994644e+00
1.33398371e-02 -7.50859901e-02 -1.18144259e-01 1.51629850e-01
3.58943880e-01 1.56467783e+00 -1.28363550e+00 -3.89470398e-01
-3.79347712e-01 1.41005528e+00 1.95145383e-01 6.40143454e-01
-1.24871659e+00 4.36730161e-02 5.54641187e-01 2.01729313e-01
1.35470137e-01 -2.88117051e-01 -9.23499882e-01 -1.38258410e+00
2.77136534e-01 -8.06871474e-01 2.81043500e-01 -1.09148979e+00
-1.02095640e+00 3.93383473e-01 1.05349027e-01 -1.20366776e+00
4.92935061e-01 -5.65278649e-01 -7.51126409e-01 7.10369766e-01
-1.20424592e+00 -1.61388433e+00 -8.16486776e-01 3.72465640e-01
7.67427981e-01 3.75578463e-01 2.56588250e-01 4.39660668e-01
-9.51770842e-01 4.84731764e-01 -2.27245986e-01 2.12316394e-01
8.75352204e-01 -1.42734373e+00 3.61477613e-01 1.01761949e+00
1.27632633e-01 5.33394814e-01 5.91236711e-01 -7.61061549e-01
-1.54300117e+00 -1.11511326e+00 6.80093169e-02 -6.35755062e-01
3.69178146e-01 -7.15746760e-01 -1.01283753e+00 7.50752509e-01
2.49692127e-01 -1.74381018e-01 3.37212473e-01 -2.11291630e-02
-2.63598084e-01 -1.87922955e-01 -8.44063044e-01 9.58319128e-01
1.01094902e+00 -9.61617380e-02 -1.80292666e-01 3.12932789e-01
8.69359851e-01 -8.64459693e-01 -6.50270879e-01 4.89550710e-01
4.84456509e-01 -1.08180404e+00 1.03473067e+00 3.69210690e-02
4.05107588e-01 -9.15532291e-01 8.12815130e-02 -8.69226754e-01
-1.04381077e-01 -6.50816381e-01 -1.50295585e-01 1.61740816e+00
4.20096368e-01 -3.33088219e-01 8.73973429e-01 9.45004284e-01
-3.87862265e-01 -9.25606966e-01 -5.96297026e-01 -3.78805786e-01
-4.64395463e-01 -4.44306642e-01 5.39852440e-01 8.89888942e-01
-4.51817542e-01 2.20798433e-01 -7.58491278e-01 3.58643234e-01
5.79448760e-01 5.21400630e-01 1.09789288e+00 -7.17855453e-01
-4.71831113e-01 -3.45642328e-01 6.76031038e-02 -1.34157622e+00
-3.88420045e-01 -4.60989952e-01 2.63218284e-01 -1.67391241e+00
5.25182843e-01 -6.07783675e-01 8.05683509e-02 6.72589004e-01
-7.44842172e-01 5.54463565e-01 2.93205917e-01 4.25990105e-01
-8.80981326e-01 5.25081635e-01 1.60710001e+00 -2.89714605e-01
-2.67345488e-01 -5.84124699e-02 -7.34868467e-01 8.73886883e-01
6.33030951e-01 -1.74889371e-01 -4.40845728e-01 -7.40087628e-01
1.03899918e-01 -7.98496455e-02 3.66458356e-01 -7.39024460e-01
-1.02476947e-01 -5.51405907e-01 5.70416331e-01 -7.72782803e-01
4.55456883e-01 -7.57237732e-01 2.69137859e-01 8.10188353e-02
1.41196594e-01 -1.11757994e-01 1.94586858e-01 4.81286019e-01
1.47253126e-02 -1.87454194e-01 6.53064549e-01 -1.41401559e-01
-7.61138141e-01 3.63119811e-01 -1.10619359e-01 -1.06742419e-01
1.02708530e+00 -3.32114816e-01 -4.43289131e-01 -2.68913001e-01
-4.22758788e-01 4.32171464e-01 9.25209284e-01 4.11889762e-01
5.38610399e-01 -1.05589998e+00 -3.47807109e-01 8.63295943e-02
5.68070486e-02 6.97626114e-01 2.81991869e-01 1.08196855e+00
-6.95674777e-01 -3.79518062e-01 1.08783409e-01 -7.57718265e-01
-1.38070631e+00 3.86952907e-01 2.91129291e-01 -7.48109829e-04
-7.33478248e-01 8.92653584e-01 9.50183272e-01 -1.01849765e-01
2.02280656e-01 -6.15200460e-01 2.30183139e-01 -1.39759570e-01
4.66479957e-01 1.31454110e-01 2.80841347e-02 -4.29662675e-01
-1.46346107e-01 4.36735511e-01 -3.57891738e-01 1.06838182e-01
9.31258202e-01 -4.08999592e-01 -2.88307905e-01 5.62186241e-01
4.01134104e-01 -1.02449134e-01 -1.64860320e+00 -1.87025927e-02
-2.67357707e-01 -6.92512274e-01 -1.36658862e-01 -9.18295741e-01
-1.38647854e+00 7.76413441e-01 5.09799063e-01 3.43318954e-02
1.22379708e+00 2.33679917e-02 1.02171135e+00 -2.81190723e-02
2.06615970e-01 -1.05250204e+00 1.74146861e-01 1.91219255e-01
7.23450363e-01 -1.51929247e+00 1.96408719e-01 -9.64645624e-01
-8.54266226e-01 9.55995619e-01 1.18327594e+00 3.70235622e-01
1.50526449e-01 4.61116165e-01 5.15454173e-01 -2.31042117e-01
-2.93264449e-01 -1.42404199e-01 5.55319548e-01 3.11484277e-01
4.59431261e-01 1.25313178e-01 -2.00861320e-02 5.87643445e-01
3.26980427e-02 -4.38359648e-01 3.68420362e-01 8.30796480e-01
-4.31278050e-01 -9.99465704e-01 -6.98556781e-01 4.95152980e-01
-3.57632667e-01 -1.50812969e-01 -5.91829598e-01 6.68700218e-01
3.08046967e-01 9.90663707e-01 -1.52692184e-01 -4.97058272e-01
3.35569531e-02 -1.40344381e-01 5.42914450e-01 -4.93953526e-01
-4.88497853e-01 3.48091364e-01 -1.04147449e-01 -5.93642116e-01
-4.62124467e-01 -3.91559541e-01 -1.24712098e+00 -1.88026905e-01
-6.20249093e-01 -1.97169170e-01 4.99548912e-01 1.05989230e+00
1.97449088e-01 6.89660788e-01 1.81792721e-01 -1.15016723e+00
3.93865220e-02 -1.00311041e+00 -4.32371050e-01 5.51914215e-01
1.08786337e-01 -7.73200214e-01 -5.33019416e-02 3.68069470e-01]
|
[10.612481117248535, -0.9170144200325012]
|
e650fae8-6706-4dd7-a55d-801dcb60d109
|
self-supervised-representation-learning-for-5
|
2101.12482
| null |
https://arxiv.org/abs/2101.12482v4
|
https://arxiv.org/pdf/2101.12482v4.pdf
|
Self-Supervised Pretraining for RGB-D Salient Object Detection
|
Existing CNNs-Based RGB-D salient object detection (SOD) networks are all required to be pretrained on the ImageNet to learn the hierarchy features which helps provide a good initialization. However, the collection and annotation of large-scale datasets are time-consuming and expensive. In this paper, we utilize self-supervised representation learning (SSL) to design two pretext tasks: the cross-modal auto-encoder and the depth-contour estimation. Our pretext tasks require only a few and unlabeled RGB-D datasets to perform pretraining, which makes the network capture rich semantic contexts and reduce the gap between two modalities, thereby providing an effective initialization for the downstream task. In addition, for the inherent problem of cross-modal fusion in RGB-D SOD, we propose a consistency-difference aggregation (CDA) module that splits a single feature fusion into multi-path fusion to achieve an adequate perception of consistent and differential information. The CDA module is general and suitable for cross-modal and cross-level feature fusion. Extensive experiments on six benchmark datasets show that our self-supervised pretrained model performs favorably against most state-of-the-art methods pretrained on ImageNet. The source code will be publicly available at \textcolor{red}{\url{https://github.com/Xiaoqi-Zhao-DLUT/SSLSOD}}.
|
['Xiang Ruan', 'Huchuan Lu', 'Lihe Zhang', 'Youwei Pang', 'Xiaoqi Zhao']
|
2021-01-29
| null | null | null | null |
['rgb-d-salient-object-detection']
|
['computer-vision']
|
[ 6.73306882e-02 -2.12896895e-02 -2.07833380e-01 -5.81375718e-01
-8.03883195e-01 -2.62894601e-01 4.52181935e-01 4.30317260e-02
-4.20661062e-01 3.32816213e-01 6.16598055e-02 -1.18505999e-01
1.53483614e-01 -7.95876205e-01 -7.32803762e-01 -7.86478460e-01
3.05305928e-01 -9.77651924e-02 5.84332347e-01 -3.29425991e-01
-8.01121071e-02 5.07081807e-01 -1.79396927e+00 3.15961927e-01
7.76252091e-01 1.52065110e+00 6.05625093e-01 1.57736495e-01
-2.08322093e-01 6.02190018e-01 -1.16938621e-01 -2.80368358e-01
4.08724427e-01 -3.85135889e-01 -7.04857111e-01 1.69908985e-01
4.94766116e-01 -4.55668747e-01 -3.72595191e-01 1.28638291e+00
6.75966322e-01 6.90014735e-02 3.43707651e-01 -1.41916835e+00
-5.62903225e-01 2.52080858e-01 -7.42778838e-01 1.14656411e-01
1.58050880e-01 2.07699537e-01 8.58120382e-01 -1.04977286e+00
4.62261289e-01 1.01795781e+00 5.35738766e-01 6.03518188e-01
-8.70462477e-01 -8.08952928e-01 1.43553987e-01 1.49611443e-01
-1.32601786e+00 -3.54671746e-01 1.23721588e+00 -2.03901708e-01
5.97074747e-01 9.92872790e-02 8.28599811e-01 8.80033851e-01
-2.00123534e-01 1.11716998e+00 1.14155352e+00 -3.05087596e-01
9.34646204e-02 -7.35588223e-02 -9.49334428e-02 1.01676142e+00
2.97414660e-01 1.69998497e-01 -6.51160598e-01 3.12217444e-01
1.15383399e+00 2.65117049e-01 -2.79195786e-01 -6.87871873e-01
-1.11893094e+00 6.86704397e-01 1.07906163e+00 2.98284560e-01
-2.82128185e-01 1.12171270e-01 2.26827934e-01 -6.81832433e-03
3.84988278e-01 -3.63727706e-03 -4.36596811e-01 2.36394867e-01
-7.11151898e-01 -1.06753796e-01 2.04224095e-01 9.02618945e-01
1.19744015e+00 -3.55879031e-02 1.69220135e-01 7.24560976e-01
4.78187352e-01 5.08499801e-01 4.56827283e-01 -8.08355808e-01
4.39083070e-01 9.56688106e-01 -1.51027888e-01 -8.43204379e-01
-5.09570956e-01 -4.57181007e-01 -1.08990002e+00 4.38723207e-01
3.91523451e-01 -2.07036324e-02 -1.17274630e+00 1.60572803e+00
4.71951365e-01 1.53871089e-01 8.19325075e-02 1.28121281e+00
1.27684546e+00 4.75173205e-01 2.15065449e-01 5.69328219e-02
1.26977372e+00 -9.70683813e-01 -4.70356703e-01 -4.64392185e-01
4.75826979e-01 -7.39167094e-01 1.17706239e+00 -4.57527302e-02
-1.05391562e+00 -7.31787324e-01 -1.17916679e+00 -4.80986923e-01
-4.65589732e-01 3.50097537e-01 8.75255406e-01 2.62043715e-01
-9.42141354e-01 2.25302711e-01 -9.25665498e-01 -1.72857136e-01
7.08482981e-01 2.45801747e-01 -5.86288154e-01 -2.11976886e-01
-1.14644325e+00 6.80809855e-01 5.24705231e-01 2.83455014e-01
-7.18193233e-01 -4.86668080e-01 -1.08861113e+00 -1.73293054e-01
3.18394750e-01 -6.36614025e-01 1.06787193e+00 -1.05526948e+00
-1.32595980e+00 1.15126145e+00 -3.34125310e-02 3.86228934e-02
3.78364444e-01 -1.19825378e-01 -2.28497937e-01 3.30541819e-01
1.93981439e-01 8.90011668e-01 7.57634521e-01 -1.49232602e+00
-7.35164106e-01 -5.19336402e-01 1.43469810e-01 4.28242445e-01
-2.92182148e-01 -2.27701753e-01 -8.33878458e-01 -6.44419730e-01
5.12097538e-01 -6.16775632e-01 -8.79482850e-02 2.83138782e-01
-4.43888783e-01 -2.28345215e-01 8.81678998e-01 -3.68920147e-01
8.10862541e-01 -2.28914762e+00 -2.00400781e-02 -1.67163666e-02
1.05388202e-01 2.23042995e-01 -9.41773802e-02 1.44371390e-02
-1.12512842e-01 -2.13192105e-01 -3.16837966e-01 -6.94475532e-01
-1.56736627e-01 1.90884873e-01 -1.99906174e-02 5.96713126e-01
4.32440698e-01 1.05454111e+00 -9.47209299e-01 -7.21364558e-01
5.40109396e-01 5.85048914e-01 -2.73514479e-01 4.52683449e-01
-9.83373448e-02 4.72055167e-01 -5.26669919e-01 1.04049492e+00
8.38293254e-01 -3.78724307e-01 -2.04292879e-01 -5.98207593e-01
-1.56723335e-01 1.47405580e-01 -1.30389810e+00 2.24192286e+00
-3.34330529e-01 3.82949203e-01 -9.48724300e-02 -1.02929330e+00
9.55953658e-01 -9.74847898e-02 5.81578314e-01 -9.50430155e-01
5.47624052e-01 3.01764995e-01 -3.54912221e-01 -3.92137647e-01
2.67734587e-01 -3.23064509e-03 -1.03216209e-01 3.03547293e-01
2.59317935e-01 -2.52118170e-01 5.44446260e-02 1.05475232e-01
5.08631766e-01 3.00480127e-01 1.24743231e-01 -8.36920068e-02
5.81049621e-01 -3.04457527e-02 7.21649587e-01 2.94709384e-01
-2.48328358e-01 9.21883106e-01 1.94610640e-01 -4.42990422e-01
-7.40710557e-01 -1.06644988e+00 -1.87029853e-01 9.27436769e-01
8.28670204e-01 -1.87489778e-01 -4.59809244e-01 -7.56580889e-01
-4.14445512e-02 2.06316486e-01 -6.77842557e-01 -2.41309002e-01
-4.22896713e-01 -6.02062404e-01 2.57233918e-01 7.58898616e-01
1.08433247e+00 -9.92760062e-01 -8.21024001e-01 -1.16481729e-01
-2.60586679e-01 -1.13696742e+00 -2.90970832e-01 5.16536772e-01
-8.11102509e-01 -1.06410062e+00 -7.01284707e-01 -9.98333991e-01
8.65541816e-01 5.16075850e-01 7.92091250e-01 2.34321967e-01
-2.47920558e-01 1.70124382e-01 -4.45153236e-01 -4.68238682e-01
2.29361817e-01 3.52142006e-02 -2.09500223e-01 -1.26762360e-01
4.43143874e-01 -5.13676882e-01 -9.21361327e-01 2.83000886e-01
-1.04386997e+00 4.64153260e-01 8.17636549e-01 7.53056407e-01
8.66222024e-01 -2.65551001e-01 2.70379722e-01 -3.39536160e-01
-2.94681871e-03 -2.79930741e-01 -5.06120265e-01 2.71832764e-01
-3.21129173e-01 -8.71781260e-02 3.09967577e-01 -1.69593439e-01
-1.10647738e+00 4.97964233e-01 -3.41601610e-01 -5.86918354e-01
-3.75583053e-01 3.01765114e-01 -4.53487784e-01 -1.76977932e-01
3.30784202e-01 3.52029294e-01 9.65176076e-02 -5.22470057e-01
4.33304578e-01 5.66404223e-01 6.21061862e-01 -3.55440527e-01
7.96536624e-01 6.65650606e-01 -8.17268044e-02 -4.86994982e-01
-1.25794184e+00 -5.62872291e-01 -8.68505597e-01 -2.20917702e-01
9.26628888e-01 -1.25189126e+00 -4.45977122e-01 7.75975347e-01
-8.98498297e-01 -5.46387136e-01 -3.62757921e-01 3.55176419e-01
-5.53147614e-01 2.48767406e-01 -4.69824195e-01 -4.74557042e-01
-3.29461187e-01 -1.16986811e+00 1.37567484e+00 6.04324043e-01
4.10597980e-01 -8.00371468e-01 -2.14441955e-01 4.04034704e-01
2.67349452e-01 3.23459387e-01 5.14357150e-01 -2.22079232e-01
-7.95333624e-01 -1.62215605e-02 -6.26615644e-01 4.26375866e-01
2.64869094e-01 -1.73035577e-01 -1.12310469e+00 -1.43274769e-01
-2.74832100e-01 -6.36587441e-01 1.04912436e+00 3.24344486e-01
1.33322096e+00 2.21919551e-01 -1.51576623e-01 9.78804290e-01
1.59887624e+00 -2.00158298e-01 5.39747119e-01 5.50870478e-01
9.33636248e-01 4.36391801e-01 8.30650926e-01 4.32407945e-01
6.97773993e-01 4.53064919e-01 7.59709299e-01 -6.60436869e-01
-3.86047781e-01 -3.22388291e-01 9.15943235e-02 5.48873663e-01
-2.22523779e-01 4.80001643e-02 -8.54934037e-01 5.06582081e-01
-1.88715756e+00 -7.34351158e-01 2.99851783e-02 1.82883763e+00
1.06189322e+00 3.67724933e-02 1.19636483e-01 1.52634874e-01
5.04452705e-01 2.43641004e-01 -5.23666382e-01 2.01561689e-01
-3.40060472e-01 1.20855503e-01 5.39705157e-01 2.09745303e-01
-1.39648163e+00 9.51675177e-01 4.41252613e+00 7.44848132e-01
-1.36252201e+00 1.36333227e-01 6.72647417e-01 7.04248622e-02
-1.93237737e-01 -1.28043845e-01 -7.72572875e-01 3.16382557e-01
1.69581681e-01 2.05844998e-01 -7.97794685e-02 8.62918317e-01
-7.88218305e-02 -3.87306631e-01 -7.69108355e-01 1.30407369e+00
-5.01123315e-04 -1.29669571e+00 -9.88392532e-02 -1.88662484e-01
8.59573960e-01 2.98405766e-01 -4.12857570e-02 2.54978649e-02
2.20967278e-01 -6.32304907e-01 8.83779585e-01 3.89346570e-01
8.54757786e-01 -7.12212622e-01 8.40013742e-01 9.84827429e-02
-1.51625574e+00 -1.02623813e-01 -4.44705397e-01 2.37362206e-01
1.22523993e-01 6.93721950e-01 -1.13735735e-01 6.94632232e-01
1.09044659e+00 1.04051161e+00 -6.24778986e-01 9.88210797e-01
-5.13822973e-01 -3.98460310e-03 -3.47335875e-01 1.89889893e-01
2.81215817e-01 2.98616365e-02 1.45523474e-01 1.04420483e+00
1.57138988e-01 2.12432668e-01 2.99181163e-01 6.18653178e-01
-1.35342658e-01 -1.67669205e-03 -2.69658029e-01 2.55199879e-01
3.12812179e-01 1.38802660e+00 -8.88813794e-01 -1.99689433e-01
-5.20432472e-01 1.02887833e+00 4.14200991e-01 3.01630914e-01
-8.04871380e-01 -3.60028446e-01 7.24183083e-01 -1.02871601e-02
4.38555211e-01 -2.92466551e-01 -3.74448180e-01 -1.27087069e+00
3.38621177e-02 -4.77269113e-01 4.88653392e-01 -8.73700142e-01
-1.24795365e+00 6.27202749e-01 -5.21334559e-02 -1.51893198e+00
1.87553734e-01 -6.25349939e-01 -4.43423837e-01 7.39753783e-01
-2.10839200e+00 -1.54887164e+00 -9.04137254e-01 1.06318057e+00
3.93073678e-01 1.06859967e-01 5.05141437e-01 3.55042845e-01
-5.53410769e-01 5.12198329e-01 -2.89261937e-01 3.52965921e-01
6.61292255e-01 -1.14592421e+00 5.48136830e-02 9.49174702e-01
-2.15710476e-02 3.37237418e-01 1.74801186e-01 -4.72433180e-01
-1.37205541e+00 -1.16873705e+00 4.01615441e-01 -7.60898218e-02
3.56022269e-01 -3.13986570e-01 -8.33119512e-01 3.92862737e-01
5.00021130e-02 5.55305541e-01 5.42414606e-01 -3.23338121e-01
-2.90422499e-01 -3.47631931e-01 -9.25930858e-01 2.98936993e-01
1.22246301e+00 -6.85857832e-01 -5.48861861e-01 1.95014983e-01
7.08965838e-01 -6.33848488e-01 -8.49767208e-01 5.98418653e-01
3.56613308e-01 -1.24898112e+00 1.11681104e+00 -1.18975155e-01
5.70176959e-01 -5.68510890e-01 -2.46467292e-01 -8.84525836e-01
-9.44049284e-02 -8.76936615e-02 -3.36776227e-02 1.28258824e+00
1.05210513e-01 -4.19789612e-01 6.73538864e-01 4.49947268e-01
-2.69482851e-01 -1.02303731e+00 -7.40525484e-01 -4.53929007e-01
-2.78327554e-01 -5.48965156e-01 5.85415781e-01 9.85865772e-01
-2.61293650e-01 9.12107825e-02 -1.05861776e-01 2.15748966e-01
8.37107718e-01 5.12930036e-01 6.91127837e-01 -1.15192211e+00
1.24946274e-01 -4.52835619e-01 -5.01439989e-01 -1.13392556e+00
-5.58354221e-02 -8.54385614e-01 1.14820771e-01 -1.70630229e+00
1.21488586e-01 -7.97720432e-01 -4.91671473e-01 8.25920165e-01
-2.32528523e-01 6.13231778e-01 1.98525831e-01 2.24455565e-01
-7.13064253e-01 9.00480390e-01 1.46749890e+00 4.39526746e-03
-2.37969622e-01 -1.19743481e-01 -7.76627243e-01 7.77694046e-01
9.11289036e-01 -2.88757175e-01 -4.55196351e-01 -5.22688806e-01
-2.73932591e-02 -2.78224051e-01 6.87938750e-01 -9.81999099e-01
4.05264467e-01 -2.09696040e-01 7.36399293e-01 -8.59356523e-01
3.86195779e-01 -8.70032907e-01 -2.82864749e-01 3.09236109e-01
-4.58234586e-02 -1.90495357e-01 2.19224453e-01 3.75821710e-01
-4.17197555e-01 9.05739293e-02 9.40712750e-01 -2.39951134e-01
-1.22613156e+00 6.69924498e-01 2.50941724e-01 -5.06667828e-04
1.05146563e+00 -3.26247782e-01 -3.78692001e-01 -1.70106873e-01
-5.06880939e-01 2.35182568e-01 6.49941742e-01 4.78311330e-01
9.07330394e-01 -1.38676655e+00 -2.50927210e-01 4.44695681e-01
2.86089391e-01 6.25802338e-01 4.42297548e-01 8.77540767e-01
-4.64180410e-01 7.88688883e-02 -5.80545306e-01 -8.01456511e-01
-9.91375506e-01 2.82049209e-01 3.51523817e-01 8.03304985e-02
-6.19460106e-01 1.13923442e+00 2.24480957e-01 -3.50835145e-01
3.76548380e-01 -3.24271739e-01 -1.18481800e-01 9.50079933e-02
4.28208947e-01 -1.23313263e-01 1.08100832e-01 -7.73842573e-01
-5.37910938e-01 7.57812619e-01 1.89947531e-01 2.65746057e-01
1.56233907e+00 -3.56173366e-01 -1.19420886e-01 2.66776234e-01
1.43882799e+00 -4.23667371e-01 -1.56307185e+00 -5.53742111e-01
-3.38782579e-01 -5.23463428e-01 3.71523529e-01 -4.84723866e-01
-1.48526347e+00 1.07017481e+00 8.71207476e-01 -1.18019514e-01
1.59782565e+00 2.83851355e-01 8.29237044e-01 1.83883831e-01
3.66331607e-01 -9.95506525e-01 3.13665897e-01 3.12615722e-01
8.35455298e-01 -1.62224114e+00 1.67076692e-01 -5.11922657e-01
-7.75398552e-01 1.06440115e+00 9.28145170e-01 -5.92078082e-02
8.41087461e-01 1.51780084e-01 2.88591444e-01 -3.35838944e-01
-2.06677765e-01 -7.27968037e-01 4.53029722e-01 5.60684323e-01
1.94599718e-01 -1.07808240e-01 8.32478702e-02 5.43440282e-01
-1.52330967e-02 -5.46083190e-02 1.02529131e-01 1.18434179e+00
-3.96384865e-01 -1.03973949e+00 -6.41302615e-02 1.67240381e-01
-1.23714179e-01 2.75474787e-02 -2.22544804e-01 8.82685840e-01
4.64860409e-01 6.38940752e-01 1.36446565e-01 -4.61105615e-01
2.86249459e-01 -1.99362442e-01 4.07518029e-01 -4.40344632e-01
-3.04043710e-01 1.57490343e-01 -1.89735979e-01 -7.60247946e-01
-9.95465279e-01 -4.69791859e-01 -1.45409262e+00 -6.56568408e-02
-3.80415469e-01 -3.31167370e-01 6.63502395e-01 8.51548493e-01
2.88861245e-01 4.60980386e-01 6.52774751e-01 -1.17111552e+00
-5.26934937e-02 -7.20646322e-01 -4.85599875e-01 4.78280276e-01
4.33479786e-01 -1.00666416e+00 -2.08645597e-01 5.94519712e-02]
|
[9.598716735839844, -0.9307129979133606]
|
0698b585-c0e4-4d58-af8d-468509ce6cf1
|
anaphora-resolution-for-machine-translation
| null | null |
https://aclanthology.org/F13-2023
|
https://aclanthology.org/F13-2023.pdf
|
Anaphora Resolution for Machine Translation (R\'esolution d'anaphores et traitement des pronoms en traduction automatique \`a base de r\`egles) [in French]
| null |
["Sharid Lo{\\'a}iciga"]
|
2013-06-01
|
anaphora-resolution-for-machine-translation-1
|
https://aclanthology.org/F13-2023
|
https://aclanthology.org/F13-2023.pdf
|
jeptalnrecital-2013-6
|
['abstract-anaphora-resolution']
|
['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]
|
[-7.372808456420898, 3.732131004333496]
|
1e623bb4-7d44-4397-9de1-83bb72fd0fae
|
automated-evaluation-for-student
|
2205.04083
| null |
https://arxiv.org/abs/2205.04083v1
|
https://arxiv.org/pdf/2205.04083v1.pdf
|
Automated Evaluation for Student Argumentative Writing: A Survey
|
This paper surveys and organizes research works in an under-studied area, which we call automated evaluation for student argumentative writing. Unlike traditional automated writing evaluation that focuses on holistic essay scoring, this field is more specific: it focuses on evaluating argumentative essays and offers specific feedback, including argumentation structures, argument strength trait score, etc. The focused and detailed evaluation is useful for helping students acquire important argumentation skill. In this paper we organize existing works around tasks, data and methods. We further experiment with BERT on representative datasets, aiming to provide up-to-date baselines for this field.
|
['Juneyoung Park', 'Yohan Lee', 'Xinyu Wang']
|
2022-05-09
| null | null | null | null |
['automated-writing-evaluation']
|
['natural-language-processing']
|
[-1.47284448e-01 3.86653841e-01 -6.00793302e-01 -4.17900264e-01
-6.64006472e-01 -8.48640501e-01 6.10613465e-01 9.71397996e-01
-2.93297380e-01 1.10720825e+00 4.49718654e-01 -7.73366094e-01
-5.20575285e-01 -7.39260435e-01 -2.14893743e-01 -1.01422541e-01
6.18864357e-01 5.76751709e-01 4.74459976e-02 -6.65946066e-01
1.22947097e+00 -4.78139222e-02 -1.25189519e+00 4.06649441e-01
1.48077822e+00 2.61337757e-01 -3.45537961e-01 9.20242190e-01
-6.19329929e-01 1.46857691e+00 -1.26530194e+00 -1.10731292e+00
-5.78998387e-01 -6.19069815e-01 -1.34265256e+00 -5.70115983e-01
8.37671995e-01 -2.00957566e-01 4.13844883e-01 8.78802061e-01
5.60361743e-01 1.57964751e-01 9.62894738e-01 -9.88305986e-01
-1.07743382e+00 1.10354865e+00 -4.24240500e-01 3.64176512e-01
8.08646977e-01 -3.58893424e-01 1.33615208e+00 -7.04634488e-01
5.83510756e-01 1.08917952e+00 8.16308618e-01 6.00131154e-01
-6.16456687e-01 -4.07194972e-01 2.28132933e-01 5.37131548e-01
-5.81442425e-03 1.29949346e-01 1.18337560e+00 -5.76401651e-01
3.89924288e-01 2.80334085e-01 8.78817976e-01 1.23736751e+00
-1.50336400e-01 1.24647367e+00 1.42081261e+00 -8.87456357e-01
5.84023818e-03 -1.32569289e-02 1.40364778e+00 5.82428575e-01
6.44354880e-01 -4.37311679e-01 -6.90617681e-01 -2.27258746e-02
7.05830529e-02 -4.89884824e-01 -8.77292380e-02 1.39913755e-02
-1.01733565e+00 9.57099915e-01 -1.34543389e-01 2.73902208e-01
7.16095092e-03 -2.39259660e-01 6.28012002e-01 8.51083279e-01
3.01978230e-01 8.31963718e-01 -3.66807491e-01 -6.29629910e-01
-1.01350737e+00 5.55801868e-01 1.31713665e+00 3.77886921e-01
5.41907735e-02 -7.00497404e-02 -8.58077228e-01 1.04797971e+00
3.62452030e-01 4.08357710e-01 6.82656229e-01 -1.01900375e+00
7.63229132e-01 9.31766689e-01 -1.79076213e-02 -5.42779744e-01
-2.80898541e-01 -5.02387583e-01 -1.66957974e-01 3.54620129e-01
5.29803932e-01 -4.61947769e-01 -1.68839529e-01 1.40651810e+00
1.33381397e-01 -2.06916898e-01 2.31873438e-01 6.25863612e-01
1.76540136e+00 2.58994043e-01 1.89735547e-01 -1.58490002e-01
1.59278297e+00 -1.61907017e+00 -1.10014117e+00 1.89719528e-01
8.75215352e-01 -1.15742433e+00 1.44246364e+00 7.16390014e-01
-1.71064305e+00 -3.21704090e-01 -1.20060658e+00 -4.65276003e-01
-3.32440227e-01 3.98799926e-01 3.28295171e-01 1.00170898e+00
-6.56319737e-01 5.25558710e-01 -3.07555705e-01 4.27078130e-03
4.92221951e-01 -2.21541792e-01 3.19457762e-02 4.52913374e-01
-1.28091109e+00 1.19219542e+00 3.39229293e-02 -5.20273387e-01
-2.78972179e-01 -9.52938616e-01 -6.43612087e-01 1.19200490e-01
1.61410749e-01 -6.19241953e-01 1.92026961e+00 -3.65684152e-01
-2.03714418e+00 9.87224042e-01 1.10578895e-01 -3.36693794e-01
7.01057673e-01 -5.39898992e-01 1.53713256e-01 1.11069903e-02
9.56926495e-02 -1.11308955e-02 1.75953209e-01 -7.53474653e-01
-2.75869906e-01 -2.28875086e-01 2.27205679e-01 4.55519497e-01
-8.49733114e-01 3.22471261e-01 3.76037657e-01 -9.46280062e-01
-2.41302907e-01 -3.84119451e-01 3.25778037e-01 -4.42141205e-01
-3.77165616e-01 -1.06162155e+00 7.68474400e-01 -4.70943332e-01
1.61578536e+00 -1.24759877e+00 -1.51680619e-01 -1.11886196e-01
4.50289875e-01 6.58877790e-01 1.03826545e-01 7.30214596e-01
6.63990229e-02 3.23443830e-01 5.19830687e-03 -3.16333652e-01
3.88603210e-01 -3.01833659e-01 -4.24063832e-01 -1.39025271e-01
-1.18821986e-01 1.12711895e+00 -1.11417365e+00 -8.22226048e-01
-3.89556810e-02 -5.18268682e-02 -1.43034652e-01 3.08755994e-01
-1.68183222e-01 -4.01547141e-02 -7.59413660e-01 7.25221336e-01
4.14927334e-01 -1.72121644e-01 1.18735675e-02 5.33952832e-01
-3.70776325e-01 9.93686736e-01 -8.67101967e-01 1.34660637e+00
-3.53186339e-01 1.01846814e+00 -2.67299026e-01 -1.14927793e+00
1.16589236e+00 3.80014777e-01 1.11055393e-02 -7.20894873e-01
9.70041230e-02 4.49316084e-01 1.66179851e-01 -5.49146771e-01
6.07055426e-01 4.59490925e-01 8.71177092e-02 1.37865388e+00
-1.93515584e-01 -2.81562179e-01 8.51370871e-01 4.32515353e-01
8.44480395e-01 2.27492943e-01 5.18251657e-01 -6.07710242e-01
1.00611007e+00 -3.13168652e-02 -2.46109158e-01 8.11971188e-01
-3.32083613e-01 2.02560619e-01 7.24134862e-01 -4.18735713e-01
-7.54516423e-01 -9.77557421e-01 -3.26771230e-01 1.50391662e+00
-8.23621899e-02 -6.61048293e-01 -1.03992105e+00 -1.00904751e+00
-5.29238544e-02 6.22033834e-01 -8.29426348e-01 1.86318487e-01
-7.22680390e-01 -6.08704090e-01 6.78066373e-01 6.56875253e-01
8.26603174e-01 -1.47820175e+00 -6.93154275e-01 1.59218103e-01
-4.20044452e-01 -4.36023414e-01 -7.97000073e-04 1.16932187e-02
-1.06799483e+00 -1.58615518e+00 -7.41406024e-01 -9.08588469e-01
4.62743491e-01 2.31543839e-01 1.69365251e+00 7.90413320e-01
2.09597304e-01 4.69852149e-01 -5.39835691e-01 -1.16607511e+00
-3.05121511e-01 4.88685161e-01 -5.90099394e-01 -9.89010692e-01
5.34964740e-01 -3.31954837e-01 -3.72890770e-01 4.48792316e-02
-3.40675056e-01 2.39856318e-02 4.28345084e-01 9.81812894e-01
3.07181906e-02 -8.43494058e-01 9.97329235e-01 -1.49684489e+00
1.88880968e+00 -2.09087223e-01 -8.32561553e-02 6.87017143e-01
-1.04602814e+00 1.84085101e-01 5.68830967e-01 -2.10461199e-01
-1.19410002e+00 -1.09359026e+00 -2.89773345e-01 6.92755222e-01
1.14429109e-02 7.53755450e-01 5.92132330e-01 -1.56845838e-01
1.03101909e+00 -2.34549314e-01 -6.49925768e-02 -4.04648662e-01
-2.32256390e-02 4.74877864e-01 3.37245017e-01 -1.55548847e+00
3.79686445e-01 -2.00975642e-01 3.42556760e-02 -6.18866861e-01
-1.61544776e+00 -2.70366728e-01 -4.64425445e-01 -3.65391582e-01
2.45272025e-01 -4.05340731e-01 -1.13510966e+00 1.51627332e-01
-1.16886783e+00 -4.26199049e-01 -3.56913835e-01 4.60229516e-01
-4.37836200e-01 5.55544198e-01 -9.36590314e-01 -8.72603476e-01
-8.75783145e-01 -8.92431974e-01 6.38688624e-01 5.81071794e-01
-8.05211663e-01 -1.56809449e+00 9.58121896e-01 1.05631161e+00
3.91228557e-01 -1.18640428e-02 9.05725718e-01 -9.36570942e-01
2.76984304e-01 4.66447175e-02 2.40814108e-02 2.75777787e-01
-5.58298945e-01 5.01106977e-01 -7.96349049e-01 3.12992543e-01
-1.29391581e-01 -1.02827060e+00 1.03973639e+00 3.49266797e-01
1.26753402e+00 -1.98344991e-01 9.64242443e-02 3.06919720e-02
5.97525656e-01 -2.98587680e-01 3.97897750e-01 9.51027811e-01
3.49597156e-01 8.79958391e-01 8.53803158e-01 2.01304570e-01
7.15575695e-01 4.39647853e-01 1.20388277e-01 2.61079967e-01
-3.49679917e-01 -4.16725837e-02 3.56435210e-01 1.20065188e+00
-5.63452184e-01 -4.09531027e-01 -9.35514867e-01 4.79969501e-01
-2.03490829e+00 -1.21820259e+00 -7.58688688e-01 1.64400017e+00
1.25186908e+00 1.64549816e-02 3.93031985e-01 6.63244724e-01
5.11792302e-01 2.13526547e-01 -2.16856599e-03 -1.17340279e+00
-2.16335952e-01 9.07709658e-01 -4.41054374e-01 8.04522038e-01
-8.76362026e-01 6.71814442e-01 6.96595478e+00 7.02715397e-01
-6.56261861e-01 -7.51351863e-02 5.21452665e-01 2.42798209e-01
-4.31197315e-01 -2.48027429e-01 -8.15840781e-01 3.77079576e-01
8.03087711e-01 -3.76381487e-01 -3.11352074e-01 8.16088676e-01
-9.10250892e-05 6.05929568e-02 -7.87303686e-01 3.79781187e-01
2.44590819e-01 -1.33660316e+00 -4.87424359e-02 -1.76035285e-01
1.14528739e+00 -5.36319137e-01 1.02496088e-01 5.69262743e-01
6.97287738e-01 -1.02576506e+00 5.99455893e-01 5.05815208e-01
7.28024989e-02 -6.18846834e-01 1.11434495e+00 3.38688105e-01
-4.66451377e-01 3.96989360e-02 -1.23787783e-01 -8.88543069e-01
-2.29220465e-01 4.16574419e-01 -5.44352055e-01 3.57977271e-01
4.05501932e-01 7.72150636e-01 -8.18928063e-01 1.04680157e+00
-1.27704155e+00 1.21577978e+00 1.61239505e-01 -9.81114864e-01
2.00873703e-01 -2.97162980e-01 2.34531000e-01 1.24171209e+00
1.94317922e-01 1.72346085e-01 1.54373363e-01 4.84195113e-01
-4.04607356e-01 7.42266059e-01 -2.00915068e-01 3.87272276e-02
5.92697859e-01 1.48920727e+00 -4.58563387e-01 -5.43428600e-01
-7.73415864e-02 1.63967818e-01 7.79539824e-01 -6.90889060e-02
-6.15584373e-01 -3.50874305e-01 9.23789144e-02 -1.42051786e-01
-3.07295620e-01 8.19382146e-02 -1.14853048e+00 -1.08072925e+00
2.22014599e-02 -9.17861164e-01 7.91575611e-01 -6.74920022e-01
-1.40438342e+00 7.33284280e-02 -1.78357393e-01 -8.16970944e-01
-2.87673891e-01 -8.91286254e-01 -1.48159468e+00 7.22272635e-01
-1.76825726e+00 -9.10479605e-01 -7.63593972e-01 2.06214026e-01
7.75201380e-01 -5.02512217e-01 8.04971576e-01 -6.69214204e-02
-6.95264578e-01 9.12323892e-01 -4.87045571e-02 2.12478012e-01
1.21805882e+00 -1.85651493e+00 1.63967028e-01 3.63647074e-01
2.80949801e-01 7.16217756e-01 5.24330854e-01 -5.24882913e-01
-6.15271747e-01 -2.55836397e-01 1.29740787e+00 -1.00264561e+00
7.48390436e-01 4.02313083e-01 -8.75068843e-01 2.46776670e-01
9.19195473e-01 -6.24267519e-01 1.25106132e+00 7.77228415e-01
-4.35573399e-01 4.30164397e-01 -7.77850509e-01 6.45382166e-01
7.50597060e-01 -1.76234454e-01 -1.27711535e+00 4.54508543e-01
-4.24529845e-03 -6.68862164e-01 -8.01915526e-01 4.23245043e-01
8.13230276e-01 -1.03542304e+00 9.83330011e-01 -1.19842529e+00
1.45077789e+00 1.44227326e-01 7.23008573e-01 -1.37100542e+00
-1.50250107e-01 -5.11082828e-01 -5.88264227e-01 1.18453252e+00
4.99009639e-01 -2.67633200e-01 1.07757223e+00 1.53293088e-01
-4.44319546e-01 -1.15158474e+00 -9.92722139e-02 -4.43540215e-01
7.93235660e-01 -4.59034294e-02 5.12062252e-01 1.32675993e+00
5.19760191e-01 8.79071355e-01 1.99679971e-01 -7.00677454e-01
5.64098358e-01 5.85740983e-01 9.24865365e-01 -1.93396032e+00
1.81036845e-01 -1.41029191e+00 2.51854390e-01 -8.02330375e-01
4.49734718e-01 -9.01169896e-01 -4.85586703e-01 -1.88536322e+00
2.46339262e-01 2.31754198e-03 -1.61551297e-01 3.60546559e-01
-7.62319088e-01 3.45061570e-01 -3.25667784e-02 -9.04045328e-02
-1.04767108e+00 2.91764081e-01 1.59415650e+00 -1.46889731e-01
3.07446159e-02 7.29198605e-02 -1.03538978e+00 9.05837655e-01
9.82144475e-01 -1.99048534e-01 -5.42973399e-01 -4.28996176e-01
5.96025646e-01 -1.67224646e-01 1.27085254e-01 -7.12370574e-01
4.67810720e-01 -4.26904708e-01 2.44160905e-01 -6.29931629e-01
-2.20772117e-01 -6.56243041e-03 -1.10684609e+00 3.78582656e-01
-1.04043174e+00 3.93490046e-01 -1.04287662e-01 -1.70137942e-01
-5.34530520e-01 -1.20188332e+00 3.60793680e-01 1.15840919e-02
-6.10934943e-03 -2.44941041e-01 -1.71977997e-01 6.55832946e-01
9.17536736e-01 -2.97881693e-01 -1.12090528e+00 -5.22544324e-01
-4.06247705e-01 5.50381064e-01 5.64836082e-04 1.73510626e-01
3.91248852e-01 -1.23738718e+00 -1.31595445e+00 -4.56449598e-01
6.63252026e-02 -1.65541068e-01 6.16451241e-02 8.75689745e-01
-9.50149179e-01 5.79626858e-01 -5.00800371e-01 -1.43174350e-01
-1.66335285e+00 -1.52683601e-01 6.41910639e-03 -9.14589822e-01
-3.22943598e-01 8.29635501e-01 -4.12816107e-01 -6.63607717e-01
2.26275712e-01 -3.42286602e-02 -1.19509339e+00 4.10514712e-01
7.33430803e-01 9.22424436e-01 2.81783432e-01 5.97934909e-02
2.42677718e-01 4.81683999e-01 -1.38218790e-01 -6.97442964e-02
1.16712034e+00 3.51441920e-01 -1.93124577e-01 6.40984774e-01
3.26322913e-01 5.85162818e-01 -4.20887113e-01 1.03786327e-02
3.31032604e-01 -8.87813121e-02 -2.05070496e-01 -1.40260541e+00
-3.61939490e-01 1.24562752e+00 -1.92385539e-01 5.67212582e-01
4.24685568e-01 -4.28001970e-01 4.89170909e-01 9.87078667e-01
-3.29044908e-01 -1.60658610e+00 5.26067615e-01 1.10712993e+00
8.71125102e-01 -1.28634977e+00 4.97921050e-01 -3.42801303e-01
-4.44077790e-01 1.80049324e+00 8.91718268e-01 -2.30634063e-01
3.66745889e-01 1.36306047e-01 2.49677360e-01 -4.80849952e-01
-9.16793227e-01 1.84266075e-01 7.05822408e-01 2.08794698e-01
1.49492383e+00 4.91290092e-02 -1.29127407e+00 9.70033050e-01
-9.72013354e-01 -2.39174545e-01 7.88311303e-01 8.43039453e-01
-6.88983858e-01 -1.58978581e+00 -5.34438491e-01 4.43606198e-01
-6.70844138e-01 -8.31237733e-02 -1.29511046e+00 7.87176013e-01
-4.48122114e-01 8.75465393e-01 -3.95977437e-01 3.12966824e-01
3.34834009e-01 4.47340667e-01 7.65941560e-01 -6.25742197e-01
-1.27594101e+00 -7.26640284e-01 5.37442029e-01 3.06377560e-01
-8.03912163e-01 -7.79663920e-01 -1.06465971e+00 -6.07143104e-01
-3.19509804e-01 7.51775265e-01 6.39898717e-01 1.13327360e+00
-2.23538250e-01 8.30232084e-01 1.57197997e-01 -3.10682002e-02
-8.86803567e-01 -1.38517296e+00 -3.20844427e-02 3.99030179e-01
-4.45381068e-02 -7.84731567e-01 -1.03956901e-01 -2.50343442e-01]
|
[11.275352478027344, 9.310837745666504]
|
0173552f-d5ac-4015-a7fc-5cbf1f7e6bcb
|
uzbektagger-the-rule-based-pos-tagger-for
|
2301.12711
| null |
https://arxiv.org/abs/2301.12711v2
|
https://arxiv.org/pdf/2301.12711v2.pdf
|
UzbekTagger: The rule-based POS tagger for Uzbek language
|
This research paper presents a part-of-speech (POS) annotated dataset and tagger tool for the low-resource Uzbek language. The dataset includes 12 tags, which were used to develop a rule-based POS-tagger tool. The corpus text used in the annotation process was made sure to be balanced over 20 different fields in order to ensure its representativeness. Uzbek being an agglutinative language so the most of the words in an Uzbek sentence are formed by adding suffixes. This nature of it makes the POS-tagging task difficult to find the stems of words and the right part-of-speech they belong to. The methodology proposed in this research is the stemming of the words with an affix/suffix stripping approach including database of the stem forms of the words in the Uzbek language. The tagger tool was tested on the annotated dataset and showed high accuracy in identifying and tagging parts of speech in Uzbek text. This newly presented dataset and tagger tool can be used for a variety of natural language processing tasks such as language modeling, machine translation, and text-to-speech synthesis. The presented dataset is the first of its kind to be made publicly available for Uzbek, and the POS-tagger tool created can also be used as a pivot to use as a base for other closely-related Turkic languages.
|
['Ogabek Sobirov', 'Ollabergan Yuldashev', 'Elmurod Kuriyozov', 'Maksud Sharipov']
|
2023-01-30
| null | null | null | null |
['text-to-speech-synthesis']
|
['speech']
|
[-6.29684404e-02 7.24152429e-03 -2.66205315e-02 -3.43946099e-01
-4.65206444e-01 -8.97921264e-01 6.45003974e-01 4.63244110e-01
-5.58921814e-01 8.49582911e-01 3.58417451e-01 -7.82846153e-01
-9.84208807e-02 -6.84642911e-01 -1.86963186e-01 -6.40080333e-01
2.52775550e-02 9.93209004e-01 2.87142247e-01 -6.17271245e-01
2.90757537e-01 6.21655822e-01 -1.56797600e+00 3.07709068e-01
7.03368545e-01 1.98033214e-01 7.72869110e-01 3.55644643e-01
-2.17670172e-01 4.84018058e-01 -6.03887618e-01 -1.60187289e-01
2.52515495e-01 -6.26432955e-01 -9.55479562e-01 -2.25609154e-01
1.32636771e-01 8.42368752e-02 2.19429106e-01 1.03454351e+00
3.11264813e-01 4.33743030e-01 6.84884965e-01 -4.30874825e-01
3.32422890e-02 9.80015934e-01 1.39893934e-01 3.52037162e-01
2.25140020e-01 -4.72123563e-01 1.14882183e+00 -7.75620580e-01
8.42662930e-01 1.04869759e+00 1.38267845e-01 4.41811204e-01
-5.45805812e-01 -4.14340466e-01 -4.10107523e-01 3.59776281e-02
-1.18920743e+00 -4.31580067e-01 3.40304434e-01 -7.16097176e-01
1.11408341e+00 2.16305047e-01 6.17164969e-01 5.14832258e-01
2.08112165e-01 2.96535671e-01 1.43578756e+00 -1.14432049e+00
2.51314729e-01 1.14092365e-01 2.03023911e-01 5.89086771e-01
4.37040240e-01 3.75327766e-02 -2.18139946e-01 4.28969525e-02
4.48790580e-01 -4.29829687e-01 5.96966259e-02 3.13635826e-01
-1.15276432e+00 6.90448403e-01 -1.44529730e-01 1.11905634e+00
-4.16133910e-01 -4.03496712e-01 6.02615297e-01 1.81065753e-01
3.75048578e-01 4.41456169e-01 -8.23734820e-01 -2.96261609e-01
-9.16053295e-01 1.61705181e-01 7.49059796e-01 6.42180860e-01
5.40514469e-01 1.27834946e-01 1.22209862e-01 1.07447946e+00
3.92526716e-01 6.19614959e-01 8.79138172e-01 -3.75774473e-01
3.69692028e-01 5.71940601e-01 -1.09799944e-01 -3.45696867e-01
-2.48409748e-01 -6.49583042e-02 -1.53705224e-01 1.09745905e-01
8.03002775e-01 -4.14536357e-01 -1.21330512e+00 1.26588345e+00
4.89246249e-01 -4.57507819e-01 3.24393034e-01 6.55712485e-01
6.07823789e-01 9.99825478e-01 -3.36878784e-02 -4.09215212e-01
1.93741691e+00 -4.55211192e-01 -7.89551437e-01 4.01127934e-02
6.84540331e-01 -1.39787292e+00 6.89933240e-01 3.67323220e-01
-9.25101697e-01 -2.61939853e-01 -1.02940488e+00 1.41812161e-01
-7.21207023e-01 1.76563740e-01 1.50131330e-01 6.47228777e-01
-8.39085221e-01 4.67411190e-01 -6.69619203e-01 -6.63868606e-01
-4.89184231e-01 3.78469855e-01 -4.95506525e-01 3.13657701e-01
-1.18501222e+00 1.16488028e+00 9.97845948e-01 -8.37211236e-02
-3.64502281e-01 3.05823058e-01 -8.25860500e-01 -1.38383612e-01
1.00643307e-01 -6.99895248e-02 1.37692797e+00 -1.04532337e+00
-1.35808563e+00 1.29548693e+00 -1.09926656e-01 -4.18424875e-01
1.29128471e-01 1.02389835e-01 -6.45112872e-01 2.30825096e-02
1.12661764e-01 -2.79382849e-03 5.25889099e-01 -8.49146783e-01
-1.03934276e+00 -4.51843858e-01 -4.87918288e-01 4.48680148e-02
2.04008162e-01 7.21042275e-01 -4.17395011e-02 -8.78521800e-01
2.07537323e-01 -1.01473486e+00 4.28207479e-02 -1.20196176e+00
2.09351689e-01 -3.06530178e-01 4.29254085e-01 -1.16538048e+00
1.21299493e+00 -1.90559685e+00 -2.73414940e-01 2.95412093e-01
-4.45311010e-01 8.21763337e-01 1.67773008e-01 9.81552422e-01
-1.16791241e-01 -2.44064361e-01 -8.31774250e-02 1.83559075e-01
-1.17623612e-01 7.50892937e-01 -3.14523369e-01 3.79180878e-01
5.46282530e-02 4.03441578e-01 -1.04986846e+00 -5.54729998e-01
3.64949584e-01 3.18157077e-01 6.61603212e-02 -1.36988997e-01
-1.13968499e-01 2.67407209e-01 -1.74565285e-01 4.65967178e-01
3.69621813e-01 8.48128438e-01 5.09400308e-01 3.15782279e-01
-6.05504811e-01 8.38734448e-01 -1.00403714e+00 1.04984593e+00
-4.73436594e-01 4.17131782e-01 -1.43284470e-01 -8.52729619e-01
1.47620535e+00 6.44820571e-01 1.62001550e-01 -4.70920920e-01
4.64427084e-01 1.00086403e+00 4.73706782e-01 -4.50110376e-01
9.51551735e-01 -5.18559396e-01 -3.80157940e-02 -7.01868087e-02
3.02937508e-01 -5.03874421e-02 5.76233804e-01 -3.00181657e-01
6.00014269e-01 2.64611274e-01 8.60014319e-01 -6.75708711e-01
7.78091609e-01 3.08073014e-01 6.90907300e-01 6.36117458e-02
6.42990917e-02 2.23427117e-01 2.47463852e-01 -3.89783174e-01
-1.06510091e+00 -5.66967547e-01 -4.96780574e-01 1.08140290e+00
-5.34361839e-01 -2.33299941e-01 -9.45349872e-01 -5.31163752e-01
-4.33882117e-01 8.50166380e-01 -1.61071330e-01 5.19795656e-01
-8.08484614e-01 -3.28087330e-01 6.20110452e-01 -3.40378955e-02
1.00072354e-01 -1.57416856e+00 -3.76245171e-01 6.64843917e-01
-1.79841399e-01 -8.51964176e-01 -5.69039509e-02 5.45942068e-01
-8.23506474e-01 -1.00610280e+00 -5.35482347e-01 -1.29391420e+00
4.62258577e-01 -1.09807670e-01 5.60158074e-01 -1.27801895e-01
3.59665722e-01 -4.15792525e-01 -8.48143339e-01 -6.66774154e-01
-9.84716535e-01 3.97280082e-02 5.06311692e-02 -1.65248588e-01
6.52475059e-01 -1.39146104e-01 1.98872611e-01 9.74607766e-02
-8.58958244e-01 -4.84187156e-01 2.47691572e-01 6.09619677e-01
4.65275139e-01 1.09636344e-01 2.96199918e-01 -9.26720023e-01
3.68773758e-01 -2.85740376e-01 -7.76950002e-01 -1.00237131e-02
-3.71018559e-01 4.01742421e-02 6.96096420e-01 -1.34380445e-01
-8.63144517e-01 1.12704761e-01 -7.77924657e-01 3.87491137e-01
-3.91071469e-01 4.99861628e-01 -4.39814597e-01 2.60762364e-01
4.01582330e-01 1.11058190e-01 3.24269012e-02 -8.55062962e-01
2.65827905e-02 1.21749532e+00 6.21398270e-01 -3.64268333e-01
4.24452692e-01 -4.40340079e-02 1.71784848e-01 -1.19494581e+00
-4.15028423e-01 -9.49964166e-01 -7.54211307e-01 -1.04288593e-01
8.36931050e-01 -6.82690918e-01 -1.60125390e-01 5.28171778e-01
-1.23028874e+00 -1.49210989e-01 -2.56625026e-01 7.27231085e-01
-3.14263493e-01 3.99109662e-01 -5.51546991e-01 -1.03106081e+00
-4.81329292e-01 -7.92834222e-01 6.29132092e-01 8.79168138e-02
-4.96799111e-01 -9.37729478e-01 4.30799514e-01 3.65812212e-01
-2.77603894e-01 -4.40061837e-02 9.63026881e-01 -1.35160530e+00
2.41641819e-01 -2.01649591e-01 3.60574871e-01 6.66976452e-01
1.49082392e-01 2.61851519e-01 -6.43887699e-01 5.76504730e-02
-7.80583322e-02 3.48922938e-01 6.45718694e-01 1.60139471e-01
-1.71528921e-01 -4.03748453e-01 7.72294998e-02 -1.81038491e-02
1.33786285e+00 7.62502193e-01 5.98012865e-01 6.19370103e-01
3.39746922e-01 8.35637987e-01 1.07407701e+00 1.72890306e-01
1.65085062e-01 7.27778912e-01 -3.74357253e-02 3.32570374e-01
7.60074332e-02 -1.89945340e-01 7.52540469e-01 1.28972363e+00
-1.49275362e-01 -2.06166387e-01 -1.32731426e+00 8.71100187e-01
-1.71394622e+00 -8.42588544e-01 -6.87667906e-01 2.36816502e+00
7.36197531e-01 5.69785647e-02 2.08449751e-01 4.44855660e-01
8.35028887e-01 -1.21790171e-01 7.54203260e-01 -1.05023420e+00
-6.90369830e-02 8.79318058e-01 6.25456095e-01 9.65527356e-01
-9.83154655e-01 1.33676159e+00 5.17862797e+00 9.42404211e-01
-1.23782969e+00 1.57491222e-01 -1.12208232e-01 5.05777955e-01
-4.03147601e-02 2.97322452e-01 -1.08851707e+00 5.02967358e-01
1.33989501e+00 6.56588897e-02 2.20653802e-01 5.14679015e-01
7.24044204e-01 -3.59557718e-01 -4.22550708e-01 6.96853936e-01
-5.35028875e-02 -1.02036071e+00 -8.82325694e-03 2.83587724e-01
5.51290989e-01 4.02220398e-01 -7.60959566e-01 7.95755628e-03
3.48102123e-01 -6.10621035e-01 1.05926895e+00 3.24965388e-01
5.27912259e-01 -8.25104535e-01 1.22948122e+00 4.54921216e-01
-1.00628710e+00 2.32668161e-01 -4.54917520e-01 -3.14860016e-01
2.94105232e-01 6.00995600e-01 -1.26107204e+00 6.42912149e-01
1.68682784e-01 3.45551446e-02 -1.70193240e-01 9.75382805e-01
-5.48601151e-01 1.15682316e+00 -4.08533901e-01 -5.30122817e-01
5.74818969e-01 -5.37395537e-01 7.50521243e-01 1.39653468e+00
6.21851981e-01 -2.85610966e-02 1.10129111e-01 -1.60425544e-01
2.40483195e-01 1.05457056e+00 -5.27987182e-01 -2.74280280e-01
6.48894310e-01 1.09055662e+00 -1.11005735e+00 -4.70858306e-01
-3.63055430e-02 6.21769667e-01 4.82014939e-02 -3.09493691e-02
-7.66885206e-02 -5.00134170e-01 7.04678416e-01 5.64251244e-01
4.90767062e-01 -5.51178932e-01 -2.20954672e-01 -6.79513216e-01
-2.44682394e-02 -8.57505977e-01 6.36202276e-01 -6.22730434e-01
-7.74599910e-01 8.63804460e-01 7.54778311e-02 -8.49444211e-01
-6.45610929e-01 -8.50179732e-01 -2.89532483e-01 1.18510926e+00
-1.05164087e+00 -1.07778573e+00 3.44570130e-01 1.14759589e-02
4.01474446e-01 -4.54873085e-01 1.06490374e+00 9.50148925e-02
-2.26516947e-01 -7.35311061e-02 4.36586171e-01 3.88132393e-01
5.51221371e-01 -1.13887417e+00 3.37521285e-01 1.01473284e+00
5.04849434e-01 7.02529728e-01 7.68441141e-01 -9.29482579e-01
-7.50301003e-01 -7.69323826e-01 2.06558824e+00 -1.70812249e-01
7.57860303e-01 -2.78543413e-01 -7.26976871e-01 4.03113037e-01
3.31399411e-01 -4.48169738e-01 7.28269100e-01 -1.06629096e-01
8.33245888e-02 -7.53276423e-02 -1.01321924e+00 3.12951744e-01
4.28440303e-01 -3.26458335e-01 -1.13695896e+00 2.59820372e-01
1.66416258e-01 -2.00325996e-01 -6.04782164e-01 -5.40215336e-02
4.80043501e-01 -6.12798393e-01 2.34584212e-01 -5.45679331e-01
4.66130860e-02 -4.52776462e-01 -1.43352896e-01 -1.31340957e+00
-2.69208342e-01 -9.61236715e-01 5.63273489e-01 1.70894682e+00
5.32235146e-01 -6.87452614e-01 2.62834758e-01 -1.79147795e-01
-3.28820288e-01 -1.49509847e-01 -1.16450870e+00 -9.27233338e-01
1.11644231e-01 -2.71516442e-01 6.22912109e-01 9.50850844e-01
2.78439790e-01 4.48576123e-01 7.67203979e-03 -2.38742650e-01
1.20803058e-01 -3.89622718e-01 2.64896959e-01 -1.36231744e+00
-1.61002681e-01 -3.02254498e-01 -6.02134287e-01 -4.47714418e-01
2.06434891e-01 -1.11967075e+00 2.77824014e-01 -1.45402622e+00
-5.87594390e-01 -5.71034014e-01 -1.17722088e-02 7.09896207e-01
4.90631536e-02 1.87152773e-01 2.01440945e-01 2.04360828e-01
4.79633808e-01 -8.73141363e-03 7.21344531e-01 2.82983512e-01
-3.71764630e-01 2.31175423e-02 -1.42351672e-01 6.67989314e-01
1.09137738e+00 -6.96679771e-01 -2.44211536e-02 1.58410802e-01
2.36882374e-01 -2.11981088e-01 -2.07682177e-01 -7.57985294e-01
-3.11003894e-01 -3.35357606e-01 -1.56084791e-01 -6.56049907e-01
-4.08944823e-02 -8.87788177e-01 2.45865911e-01 5.80515504e-01
3.36326987e-01 2.59580523e-01 1.36028215e-01 -2.77428538e-01
-3.00255865e-01 -1.11350322e+00 9.28601444e-01 -3.78233373e-01
-9.90076363e-01 -1.56078786e-01 -1.00748217e+00 9.41765383e-02
1.06083214e+00 -4.05256718e-01 9.87982228e-02 9.60927363e-03
-4.64542091e-01 -4.43881452e-01 5.54218173e-01 2.60225594e-01
9.86163169e-02 -9.37564671e-01 -8.33337784e-01 3.20210606e-01
2.87841469e-01 -4.38991547e-01 -2.39823386e-01 5.37666023e-01
-1.30611634e+00 6.84751630e-01 -5.01853943e-01 -1.20560187e-04
-1.59540689e+00 4.04570699e-01 -2.24078353e-02 -3.72286767e-01
-4.95089203e-01 3.35798413e-01 -5.23428679e-01 -4.61706609e-01
-2.52120256e-01 -3.54307979e-01 -6.56540632e-01 5.20998240e-01
3.32267344e-01 2.09633380e-01 5.32541215e-01 -1.37629461e+00
-5.17681003e-01 1.57930255e-01 1.55379832e-01 -6.37923121e-01
1.35466635e+00 4.27761162e-03 -4.66220498e-01 5.97091675e-01
6.79792821e-01 7.27018952e-01 -3.01280230e-01 7.24285245e-02
4.63683635e-01 -1.27598077e-01 5.46706729e-02 -7.02116787e-01
-3.04769754e-01 5.01955867e-01 6.99013546e-02 1.28590211e-01
7.76240170e-01 -5.02186865e-02 6.81143463e-01 2.12213874e-01
5.33652008e-01 -1.42871642e+00 -1.04943264e+00 1.18896663e+00
3.99271667e-01 -6.16751134e-01 -4.48768854e-01 -4.50638413e-01
-5.99916160e-01 1.19860947e+00 -1.73099026e-01 -1.06287813e-02
6.74364924e-01 3.95830035e-01 6.16913319e-01 1.58387534e-02
-3.55101794e-01 -7.54102349e-01 3.06613863e-01 6.93609238e-01
9.29866612e-01 5.46652734e-01 -1.49821782e+00 2.75300682e-01
-8.05944920e-01 -2.35775813e-01 6.41390920e-01 1.01587605e+00
-8.61142397e-01 -1.88615561e+00 -6.57372952e-01 2.25474581e-01
-9.55126405e-01 -3.31027567e-01 -5.67030251e-01 7.62273252e-01
5.44178367e-01 1.02664924e+00 4.66446429e-02 -1.58613741e-01
2.63414502e-01 5.11999130e-01 5.35309553e-01 -7.25066781e-01
-1.03426993e+00 2.95058191e-01 7.85018623e-01 1.43018514e-01
-3.33118051e-01 -1.03733635e+00 -1.56168652e+00 -2.29492724e-01
-3.17573398e-01 7.56018102e-01 1.02968764e+00 1.15679383e+00
-3.90239626e-01 6.85368106e-02 4.06764001e-01 -3.81718040e-01
-4.67663914e-01 -1.27592909e+00 -1.08422565e+00 2.58503705e-01
-1.66410618e-02 -4.42276180e-01 -3.17880176e-02 2.35144213e-01]
|
[10.377514839172363, 10.242680549621582]
|
a7b9768c-b427-476e-9e82-08b3a0d90a0a
|
variational-autoencoder-for-anti-cancer-drug
|
2008.09763
| null |
https://arxiv.org/abs/2008.09763v7
|
https://arxiv.org/pdf/2008.09763v7.pdf
|
Variational Autoencoder for Anti-Cancer Drug Response Prediction
|
Cancer is a primary cause of human death, but discovering drugs and tailoring cancer therapies are expensive and time-consuming. We seek to facilitate the discovery of new drugs and treatment strategies for cancer using variational autoencoders (VAEs) and multi-layer perceptrons (MLPs) to predict anti-cancer drug responses. Our model takes as input gene expression data of cancer cell lines and anti-cancer drug molecular data and encodes these data with our {\sc {GeneVae}} model, which is an ordinary VAE model, and a rectified junction tree variational autoencoder ({\sc JTVae}) model, respectively. A multi-layer perceptron processes these encoded features to produce a final prediction. Our tests show our system attains a high average coefficient of determination ($R^{2} = 0.83$) in predicting drug responses for breast cancer cell lines and an average $R^{2} = 0.845$ for pan-cancer cell lines. Additionally, we show that our model can generates effective drug compounds not previously used for specific cancer cell lines.
|
['Jiaqing Xie', 'Zhi Jing', 'Hongyuan Dong', 'Dexin Ren']
|
2020-08-22
| null | null | null | null |
['drug-response-prediction']
|
['medical']
|
[ 5.20880446e-02 -2.21581366e-02 -3.54322970e-01 -1.86302215e-02
-9.16974247e-01 -2.31067017e-01 4.14079696e-01 2.14801669e-01
-4.03542876e-01 1.16197944e+00 -7.16184974e-02 -5.41880846e-01
-1.76359247e-02 -9.57503617e-01 -8.22549343e-01 -1.20233798e+00
2.57358044e-01 4.81696814e-01 -1.39055356e-01 -2.00399667e-01
-8.24560821e-02 6.17058277e-01 -1.07785118e+00 3.53468597e-01
8.93975317e-01 9.13499773e-01 2.62005478e-02 9.53534305e-01
-9.59070176e-02 9.54414308e-01 -4.08376455e-01 -3.99080813e-01
-3.58044863e-01 -4.53750759e-01 -8.37979496e-01 -2.97501683e-01
-2.20980078e-01 -4.51776525e-03 -6.25161290e-01 1.07606494e+00
5.20982563e-01 -5.25991730e-02 1.15052199e+00 -9.48552370e-01
-8.52483153e-01 3.06622356e-01 -3.83860171e-01 -5.73464036e-02
-1.19497113e-01 2.67982274e-01 9.73192215e-01 -7.90376663e-01
5.15200496e-01 8.70298684e-01 5.97490013e-01 1.29094779e+00
-1.36655128e+00 -3.49192858e-01 -2.90121734e-01 3.58138606e-02
-1.34472871e+00 -4.81226623e-01 4.78450000e-01 -6.42203748e-01
1.25912714e+00 3.39530081e-01 5.23494303e-01 1.38366842e+00
9.94295657e-01 8.68651390e-01 6.50149047e-01 -3.89243700e-02
5.64131856e-01 2.94549853e-01 -4.98155914e-02 8.01830173e-01
-2.26786017e-01 1.96600929e-01 -1.13774255e-01 -3.80593330e-01
7.43951559e-01 3.52702796e-01 -4.22811121e-01 3.03769171e-01
-6.19398654e-01 1.29329848e+00 4.78773952e-01 5.21099627e-01
-7.51896381e-01 3.62798989e-01 2.77417928e-01 -3.34705226e-02
2.72983193e-01 4.32719976e-01 -6.70565844e-01 1.51776329e-01
-6.90235078e-01 -1.84923261e-01 6.65929317e-01 3.29940319e-01
3.40272039e-01 2.77437270e-01 -2.17811853e-01 1.02254629e+00
4.46218431e-01 2.31320396e-01 8.00702214e-01 -8.33252072e-01
-5.08674204e-01 6.18854046e-01 -2.01199904e-01 -4.66595709e-01
-5.45021236e-01 -4.74765062e-01 -1.28667879e+00 -8.77988897e-03
8.40240046e-02 -2.41249800e-01 -8.76890182e-01 1.60739636e+00
1.05462216e-01 3.94084871e-01 5.25219381e-01 4.66161758e-01
1.19993973e+00 9.41255867e-01 5.57167411e-01 -7.56882012e-01
1.28926051e+00 -5.87524533e-01 -7.17716873e-01 1.19494997e-01
8.04650724e-01 -1.40495569e-01 4.93194789e-01 3.16627532e-01
-1.11387277e+00 -1.93080470e-01 -7.76991963e-01 -2.32962027e-01
-3.85605931e-01 4.42612134e-02 6.59621775e-01 4.28376585e-01
-1.02044272e+00 8.54769289e-01 -1.01719773e+00 3.75071913e-02
8.35052073e-01 8.01364422e-01 -2.31616497e-01 1.87246650e-01
-1.17356992e+00 8.25618505e-01 -9.10768360e-02 -1.30943850e-01
-1.21056485e+00 -1.00198388e+00 -6.85779631e-01 8.99789408e-02
-2.62266845e-01 -1.04903245e+00 1.06867909e+00 -6.93051279e-01
-1.84662259e+00 7.26498306e-01 -3.30238521e-01 -4.07788754e-01
7.39848183e-04 4.58035439e-01 -5.16673982e-01 -5.32759354e-02
-2.61324048e-01 7.45726049e-01 3.91800463e-01 -9.38051820e-01
-5.26898384e-01 -5.17261982e-01 -5.88752627e-01 -1.67164758e-01
-4.37139571e-01 -2.53305972e-01 -1.26473323e-01 -2.36700282e-01
-4.12341624e-01 -8.86071086e-01 -4.80676532e-01 -1.08719230e-01
-4.18733805e-01 -5.23348391e-01 6.81424677e-01 -7.31983304e-01
1.18238235e+00 -1.91006494e+00 6.80689394e-01 4.21343930e-02
5.26527584e-01 1.65986031e-01 -6.12599477e-02 1.29073471e-01
-1.00302100e-01 3.81179482e-01 -3.80376160e-01 5.58639280e-02
-3.33196044e-01 3.45768273e-01 8.15680027e-02 4.66755211e-01
3.93426597e-01 1.23280108e+00 -8.58682692e-01 -3.78260225e-01
1.15648739e-01 9.96889234e-01 -5.79916596e-01 3.05259507e-02
-7.29616582e-01 3.67893636e-01 -8.00629139e-01 8.07966411e-01
2.98549682e-01 -5.37111223e-01 2.28781641e-01 -3.50732058e-02
2.40332961e-01 -2.99089074e-01 -4.57405061e-01 1.19656670e+00
-2.07086265e-01 5.44111788e-01 -1.54919520e-01 -1.03266346e+00
7.86291063e-01 5.80361962e-01 9.25373316e-01 -5.05603433e-01
5.88748217e-01 1.95316389e-01 -8.70264098e-02 -4.94383484e-01
2.88017150e-02 -5.33651173e-01 5.21230064e-02 -1.95274085e-01
1.99844241e-01 -3.72271775e-03 -1.95494488e-01 -1.50938898e-01
1.40765595e+00 -3.42183143e-01 1.52036205e-01 -1.15171358e-01
6.32377446e-01 1.41613349e-01 7.44039953e-01 2.68674672e-01
-4.86898124e-01 1.38492212e-01 7.13589013e-01 -4.66763109e-01
-9.77409184e-01 -8.54167163e-01 -4.69611913e-01 7.38617897e-01
-3.03210497e-01 3.11402176e-02 -5.67068636e-01 -3.55704904e-01
-4.05684188e-02 9.19863820e-01 -7.70480037e-01 -3.86955261e-01
-7.08368123e-02 -1.31994665e+00 7.54264534e-01 5.21111071e-01
1.58739761e-01 -8.92860472e-01 -9.53255892e-02 4.80609238e-01
1.50590107e-01 -5.99408329e-01 4.07508239e-02 5.49189031e-01
-8.91524076e-01 -1.07529986e+00 -9.73572254e-01 -1.09016788e+00
4.94280756e-01 -6.42250538e-01 8.10019970e-01 -1.15289368e-01
-3.44026536e-01 -1.47816569e-01 1.49962530e-01 -5.75396299e-01
-8.12481046e-01 -3.46256912e-01 2.91291445e-01 -1.31383523e-01
5.82781851e-01 -4.31035072e-01 -6.65230870e-01 -5.50165735e-02
-7.85091817e-01 -2.82742560e-01 4.71393824e-01 1.38370240e+00
1.21588886e+00 2.19604313e-01 5.78042626e-01 -9.29468691e-01
8.14749002e-01 -6.89127445e-01 -5.10770261e-01 2.21791163e-01
-5.21478176e-01 1.63977668e-01 1.01425076e+00 -2.95263350e-01
-8.16095650e-01 2.94845790e-01 -7.47884035e-01 -6.81701362e-01
-1.32909924e-01 6.20583892e-01 1.21228516e-01 9.38612148e-02
8.53390872e-01 4.36149865e-01 -3.54420394e-02 -1.35895655e-01
1.14098825e-01 8.82209659e-01 4.91955161e-01 3.09424754e-03
-8.41135532e-02 2.32416809e-01 3.25821251e-01 -9.69857395e-01
-3.19113582e-01 3.78854647e-02 -1.87847733e-01 2.92187501e-02
1.20039654e+00 -8.25672626e-01 -1.27714574e+00 4.46265489e-01
-1.03833449e+00 -4.11197752e-01 2.37468146e-02 6.05565190e-01
-5.12324095e-01 -2.42168047e-02 -1.27645934e+00 -6.12191558e-01
-5.89124203e-01 -1.54182422e+00 8.09425950e-01 4.37221885e-01
-6.70049116e-02 -1.32649541e+00 4.39195067e-01 2.20706061e-01
1.54522672e-01 4.93814707e-01 1.40025449e+00 -6.70458913e-01
-2.05578864e-01 -3.75341773e-01 -9.82677843e-03 1.74893305e-01
7.00197741e-02 4.04832959e-01 -1.23122168e+00 -1.83913335e-01
-1.74448133e-01 -2.67594606e-01 9.68549609e-01 1.22961855e+00
1.46041954e+00 -2.22826600e-01 -8.28878224e-01 6.89203024e-01
1.51489949e+00 8.32841158e-01 7.51604617e-01 -2.87050158e-01
4.85883683e-01 -1.42812341e-01 -3.13126355e-01 3.75139892e-01
1.48258135e-01 2.00660348e-01 4.96646315e-01 -1.24484636e-01
3.37661266e-01 -3.96432132e-02 2.38876984e-01 6.57100618e-01
3.12320329e-02 -6.46042407e-01 -7.93626070e-01 4.17376190e-01
-1.50719023e+00 -9.20324743e-01 -3.53202522e-01 1.79163551e+00
1.12405288e+00 -2.32969776e-01 -1.30070269e-01 -8.56255218e-02
3.47949892e-01 -3.25051308e-01 -9.76545274e-01 -8.34394276e-01
-1.93861365e-01 5.60857832e-01 4.73708034e-01 6.82450950e-01
-1.00415361e+00 9.01951075e-01 6.71445084e+00 8.86904359e-01
-1.29219842e+00 -1.96686044e-01 1.09359562e+00 8.41086451e-03
-4.51907665e-01 -4.58080649e-01 -3.62260103e-01 4.05960530e-01
1.43742025e+00 4.05218732e-03 3.37930232e-01 6.97723091e-01
5.19259758e-02 2.41611823e-01 -1.22538865e+00 9.47757244e-01
-3.84716302e-01 -1.87495351e+00 -9.05105472e-02 2.65690207e-01
7.47703552e-01 2.15819731e-01 3.09217811e-01 4.54371482e-01
7.99809933e-01 -1.55382836e+00 -3.09877723e-01 8.40772867e-01
8.84398878e-01 -8.87227297e-01 7.52294958e-01 5.41136026e-01
-5.64466059e-01 -1.66028991e-01 -4.35565919e-01 2.68642724e-01
-2.48622119e-01 5.17336667e-01 -8.70176494e-01 2.73424536e-01
5.32378674e-01 7.20139742e-01 1.48593653e-02 5.76950550e-01
5.24028093e-02 6.45687163e-01 -4.61366624e-02 -5.90226889e-01
1.69108436e-01 4.43695523e-02 2.75960386e-01 1.01217306e+00
2.70068258e-01 5.13732851e-01 -1.17315620e-01 8.63557339e-01
-3.54403377e-01 3.25073600e-02 -3.08161855e-01 -3.99397105e-01
2.60631561e-01 9.68105376e-01 -2.46603981e-01 -3.06334257e-01
-1.30430490e-01 1.17364037e+00 5.91830164e-02 4.49958205e-01
-7.90630102e-01 -3.15331668e-01 8.90505254e-01 -4.48200285e-01
8.10156316e-02 4.77071911e-01 -4.25577909e-01 -1.04325008e+00
-8.78951609e-01 -7.60145903e-01 5.10209799e-01 -7.18328536e-01
-1.29538488e+00 5.64889193e-01 -7.12641895e-01 -7.94641793e-01
-2.33975962e-01 -1.04085410e+00 -4.91810888e-01 9.27080929e-01
-1.29752338e+00 -7.50223160e-01 1.69696510e-01 6.35017753e-01
3.99717301e-01 -3.58660609e-01 1.35575962e+00 1.98683515e-01
-1.22002268e+00 6.63613498e-01 6.96915984e-01 1.21721186e-01
3.83020677e-02 -1.15174723e+00 -1.96004212e-01 1.34166539e-01
-1.10134177e-01 1.32087976e-01 7.61246443e-01 -4.76138711e-01
-1.59392905e+00 -1.08634019e+00 9.08891261e-01 -4.54481542e-01
5.05360067e-01 1.99160755e-01 -9.76463139e-01 4.74899441e-01
1.60891309e-01 8.56922641e-02 1.25346458e+00 -1.97183177e-01
1.27847210e-01 1.89698324e-01 -1.43414283e+00 7.04894185e-01
4.06732082e-01 -4.97465789e-01 -1.09571539e-01 4.62811321e-01
6.59678757e-01 -3.97650987e-01 -1.50743032e+00 4.01379198e-01
4.68112141e-01 -4.23035592e-01 9.43105161e-01 -1.01694870e+00
9.47770655e-01 -6.80655092e-02 -2.15574995e-01 -1.52746308e+00
-7.67905712e-01 -2.57339180e-01 -2.50706226e-01 3.88647765e-01
8.48303914e-01 -3.81215602e-01 1.05168545e+00 8.87672603e-01
-1.21469572e-01 -1.19787824e+00 -1.17772925e+00 -1.08507685e-01
6.71732545e-01 -3.51034105e-01 2.00369611e-01 1.06011403e+00
1.45991668e-01 4.69165266e-01 -2.27925792e-01 1.19463772e-01
3.81878167e-01 -3.74266863e-01 1.70133069e-01 -1.29485464e+00
-5.71293652e-01 -7.85855114e-01 -3.30348551e-01 -5.90779543e-01
2.49756157e-01 -1.10019517e+00 -2.75847882e-01 -1.57552934e+00
3.01827520e-01 1.29633665e-03 -7.42079377e-01 6.89876318e-01
3.86036597e-02 -5.72410859e-02 -5.33110678e-01 -7.85857365e-02
-5.71142696e-02 7.35015333e-01 1.05877709e+00 -6.34772718e-01
-5.60668647e-01 6.10556491e-02 -8.77986848e-01 5.54450274e-01
6.81174517e-01 -3.97404253e-01 -7.29553252e-02 -1.67493865e-01
-2.90592629e-02 7.59136796e-01 2.47430727e-01 -6.07988954e-01
1.61912590e-01 -5.31955719e-01 9.40905929e-01 -2.63713688e-01
3.99705470e-01 -4.58856821e-01 3.14659864e-01 8.91337752e-01
-3.44028294e-01 -3.66875559e-01 4.50187862e-01 5.96825957e-01
-1.05501749e-01 -9.90004167e-02 9.07686055e-01 -1.55499548e-01
-4.05898839e-01 6.55380368e-01 -8.61104906e-01 -4.93521839e-01
1.11637306e+00 -1.44024640e-01 -1.85060516e-01 -2.01333493e-01
-8.47472489e-01 4.31451976e-01 9.51939225e-02 4.07217927e-02
8.42645586e-01 -1.11129260e+00 -8.24333370e-01 1.91147283e-01
-7.66721880e-03 -8.63048285e-02 4.43249196e-01 6.93493187e-01
-6.46939278e-01 4.53734875e-01 -5.30949570e-02 -4.83314097e-01
-1.16033316e+00 6.34370089e-01 8.58455062e-01 -2.85269320e-01
-1.35142282e-01 1.41439581e+00 7.22460598e-02 -1.13084964e-01
1.10001555e-02 -6.91327527e-02 -4.41303223e-01 -1.93386734e-01
2.54871815e-01 1.55653223e-01 -1.77416578e-01 -4.93187547e-01
-4.48882461e-01 2.98413247e-01 -2.01669425e-01 1.85006768e-01
1.67217410e+00 6.36225998e-01 -2.01540843e-01 1.10988416e-01
1.49227118e+00 -5.07469356e-01 -9.66928959e-01 7.21563250e-02
-2.84260035e-01 3.04937452e-01 7.44605601e-01 -9.00743186e-01
-1.14295721e+00 8.72897387e-01 6.63624883e-01 -1.02443434e-01
1.21353984e+00 6.12855367e-02 7.83524215e-01 4.24014270e-01
-1.92623302e-01 -9.89531636e-01 -2.20638990e-01 2.20041007e-01
5.90640187e-01 -1.03591621e+00 -6.15156628e-02 5.08908033e-02
-6.22477829e-01 1.08529198e+00 3.16841066e-01 1.47940209e-02
8.89516771e-01 1.83058426e-01 1.56706839e-03 -3.38905603e-01
-1.21141446e+00 2.22134918e-01 1.04255289e-01 4.46885467e-01
6.21048868e-01 3.55615288e-01 -1.95444107e-01 8.80619407e-01
2.60435820e-01 3.98195803e-01 3.20818871e-01 5.84139526e-01
-4.63182211e-01 -1.08807349e+00 -7.00993687e-02 6.89143777e-01
-7.03981519e-01 -2.64443811e-02 -5.01995146e-01 3.65441144e-01
-5.73893264e-02 7.19241500e-01 2.29136989e-01 -6.48993552e-01
3.93365845e-02 3.00590038e-01 3.30503911e-01 -1.02081403e-01
-4.72659826e-01 3.24727029e-01 -3.18595320e-01 -1.18611127e-01
-8.24227482e-02 -3.22059989e-01 -1.52222681e+00 -3.31064701e-01
-2.44904578e-01 1.03277706e-01 6.15633786e-01 7.75668859e-01
5.28467774e-01 9.06103849e-01 7.12394595e-01 -3.04570764e-01
-3.85596752e-01 -6.66025996e-01 -5.91504216e-01 6.35543780e-04
4.20419604e-01 -3.32484633e-01 -3.01576983e-02 1.36437580e-01]
|
[5.918603420257568, 5.732357978820801]
|
c091f07e-57d4-4bca-a643-4a083e3dede0
|
ramp-retrieval-and-attribute-marking-enhanced
|
2305.17131
| null |
https://arxiv.org/abs/2305.17131v1
|
https://arxiv.org/pdf/2305.17131v1.pdf
|
RAMP: Retrieval and Attribute-Marking Enhanced Prompting for Attribute-Controlled Translation
|
Attribute-controlled translation (ACT) is a subtask of machine translation that involves controlling stylistic or linguistic attributes (like formality and gender) of translation outputs. While ACT has garnered attention in recent years due to its usefulness in real-world applications, progress in the task is currently limited by dataset availability, since most prior approaches rely on supervised methods. To address this limitation, we propose Retrieval and Attribute-Marking enhanced Prompting (RAMP), which leverages large multilingual language models to perform ACT in few-shot and zero-shot settings. RAMP improves generation accuracy over the standard prompting approach by (1) incorporating a semantic similarity retrieval component for selecting similar in-context examples, and (2) marking in-context examples with attribute annotations. Our comprehensive experiments show that RAMP is a viable approach in both zero-shot and few-shot settings.
|
['Maria Nadejde', 'Georgiana Dinu', 'Anna Currey', 'Benjamin Hsu', 'Xing Niu', 'Phu Mon Htut', 'Gabriele Sarti']
|
2023-05-26
| null | null | null | null |
['semantic-textual-similarity', 'semantic-similarity']
|
['natural-language-processing', 'natural-language-processing']
|
[ 5.14924884e-01 -9.04652104e-02 -7.46284783e-01 -4.25309598e-01
-1.48189914e+00 -9.14726734e-01 1.34131372e+00 3.48184526e-01
-4.53952521e-01 8.57852757e-01 4.41746205e-01 -5.22248685e-01
1.01538725e-01 -3.95031601e-01 -4.58833724e-01 -2.65248567e-01
3.72558773e-01 8.33586812e-01 -8.06836039e-02 -4.73305166e-01
3.32519919e-01 2.06663832e-01 -1.32886696e+00 3.35346133e-01
1.15619946e+00 4.40065145e-01 1.17243724e-02 4.24869657e-01
-4.34699744e-01 6.63124382e-01 -5.61895967e-01 -8.26226771e-01
2.40664870e-01 -6.72978103e-01 -8.36239219e-01 -8.90415758e-02
4.50691193e-01 -6.67699706e-03 2.51286566e-01 9.46957231e-01
5.70849419e-01 3.39753151e-01 6.12141430e-01 -1.33707714e+00
-9.37365174e-01 6.02776408e-01 -3.58041137e-01 1.56592607e-01
6.82548761e-01 1.99520096e-01 1.27859116e+00 -1.25702369e+00
7.85146534e-01 1.21645546e+00 4.01257277e-01 6.48675859e-01
-1.45886374e+00 -6.49959385e-01 1.89920459e-02 -8.09800327e-02
-1.19375896e+00 -7.06260443e-01 5.17321110e-01 -3.75865102e-01
1.07444465e+00 2.81781822e-01 3.15499038e-01 1.22132623e+00
-8.79206285e-02 9.09807801e-01 1.36411130e+00 -8.01488161e-01
1.67160571e-01 1.30616426e-01 -1.71316326e-01 3.88755053e-01
-9.36501697e-02 -1.52935339e-02 -7.08515525e-01 -3.67755860e-01
5.20117640e-01 -2.02088088e-01 3.57649364e-02 -1.55744469e-02
-1.61067069e+00 9.16204572e-01 -1.08344518e-01 2.69623637e-01
-2.49951512e-01 3.18864398e-02 3.97337049e-01 5.12033224e-01
8.87498081e-01 1.07867885e+00 -4.73339349e-01 -5.89920282e-01
-1.04933333e+00 5.83583057e-01 7.91069090e-01 1.34843862e+00
8.19014013e-01 -1.44272223e-01 -6.48091793e-01 1.02400529e+00
-2.04496413e-01 7.41738439e-01 4.40162778e-01 -8.14677596e-01
7.39280939e-01 7.23279357e-01 3.18089753e-01 -3.36215109e-01
4.69379872e-02 -6.73656389e-02 -1.67914152e-01 -3.18023413e-01
5.24830520e-01 -4.08102199e-02 -1.02659249e+00 1.89810169e+00
2.91425318e-01 -1.56158775e-01 -3.92220402e-03 8.88928175e-01
5.61899364e-01 5.40839732e-01 4.93575752e-01 -1.61889657e-01
1.55949998e+00 -8.90618801e-01 -8.03634226e-01 -4.04926747e-01
8.40548038e-01 -1.30195856e+00 1.67678535e+00 -8.81705508e-02
-8.71452391e-01 -9.07664895e-02 -7.10587561e-01 -3.83551776e-01
-5.46314418e-01 6.43700957e-02 7.90043831e-01 5.41427076e-01
-8.22759092e-01 4.83831108e-01 -5.45981467e-01 -6.55288339e-01
2.14096978e-01 2.80323297e-01 -4.39595580e-01 -2.76095927e-01
-1.42382956e+00 1.05601025e+00 -6.82839304e-02 -3.73184472e-01
-4.46213812e-01 -7.90330768e-01 -1.06624210e+00 -1.50261104e-01
6.24781132e-01 -6.85311556e-01 1.59757054e+00 -1.11134315e+00
-1.54337788e+00 8.87711525e-01 -3.59825492e-01 1.59664650e-03
3.51777256e-01 -3.48783940e-01 -4.42835152e-01 3.70487832e-02
5.29514015e-01 6.97778225e-01 8.14813495e-01 -8.67644310e-01
-5.22821128e-01 -2.49675512e-01 1.40262976e-01 6.16398513e-01
-4.32397604e-01 7.58342922e-01 -4.58413452e-01 -9.21697259e-01
-1.20365836e-01 -1.09468162e+00 -1.74156517e-01 -3.06733251e-01
-3.30100656e-01 -3.65788519e-01 5.85804999e-01 -5.72538376e-01
1.15356803e+00 -1.85554600e+00 -1.72196776e-02 -1.66240916e-01
-1.89136520e-01 2.42895484e-01 -2.33680397e-01 7.88129210e-01
2.87909806e-01 2.01157257e-01 -1.50506273e-01 -5.01334429e-01
4.32113372e-02 8.08676984e-03 -3.88711452e-01 7.91483894e-02
5.42225122e-01 1.08933544e+00 -1.29013836e+00 -7.56493628e-01
-1.38886096e-02 2.64297664e-01 -4.81534243e-01 1.91785485e-01
-4.48104769e-01 3.54725212e-01 -4.09973264e-01 1.03542161e+00
8.09216946e-02 -4.66373339e-02 1.68786854e-01 2.42958203e-01
-9.57320556e-02 7.15628743e-01 -7.32542813e-01 1.82892704e+00
-6.38613105e-01 4.74737048e-01 -1.68881133e-01 -3.45382243e-01
7.79346943e-01 4.90563244e-01 2.93573916e-01 -6.95908546e-01
-6.80973381e-02 5.12230575e-01 -1.56437173e-01 -4.08181220e-01
1.03537416e+00 -3.20833445e-01 -4.37070638e-01 8.76233995e-01
4.52340506e-02 -3.67328197e-01 3.53856295e-01 4.01255935e-01
9.40147877e-01 5.40236771e-01 4.92822587e-01 -2.87083298e-01
1.36592224e-01 4.39860880e-01 6.91771686e-01 5.35670638e-01
-2.47241214e-01 6.50981367e-01 3.03663433e-01 -1.50178850e-01
-1.20028484e+00 -8.70677531e-01 2.44545475e-01 1.55896294e+00
5.75651601e-02 -5.64326882e-01 -7.20900416e-01 -8.78514111e-01
-1.55303569e-03 9.43995953e-01 -4.05789524e-01 -1.10281311e-01
-6.77935123e-01 -4.35363591e-01 6.26345456e-01 6.33355677e-01
1.13527589e-01 -9.67850506e-01 -3.21843594e-01 2.74871796e-01
-4.27560180e-01 -1.20118082e+00 -9.25696075e-01 1.41815469e-01
-6.83586776e-01 -7.11572826e-01 -7.45624065e-01 -7.16421783e-01
5.98043919e-01 2.92375863e-01 1.30944252e+00 -1.51039556e-01
-1.59056485e-02 2.52046078e-01 -5.50659955e-01 -4.84540164e-01
-5.46622813e-01 3.52879256e-01 1.79740265e-01 -1.82389855e-01
7.05433786e-01 -2.81092077e-01 -3.63976270e-01 3.23043972e-01
-6.63941920e-01 1.73253253e-01 7.16385365e-01 8.91543567e-01
4.26427841e-01 -8.63193512e-01 7.58939147e-01 -1.14202893e+00
1.07759273e+00 -3.71332884e-01 -4.11182642e-01 3.86401325e-01
-9.43788767e-01 -4.84987050e-02 5.36403358e-01 -5.49501896e-01
-1.03360653e+00 -7.38733187e-02 2.47967258e-01 -3.44207466e-01
-8.94291326e-02 4.36845422e-01 -7.82124251e-02 1.69128597e-01
6.89833701e-01 6.94101751e-02 2.88058408e-02 -3.69288981e-01
4.90241051e-01 9.39775944e-01 4.26763982e-01 -1.01074386e+00
7.95093179e-01 2.19558999e-02 -2.09199682e-01 -3.50498587e-01
-7.60140955e-01 -4.74963427e-01 -4.81466651e-01 -8.78668129e-02
6.85167491e-01 -1.01411915e+00 -1.79968491e-01 -6.24923445e-02
-9.52045321e-01 -3.80493760e-01 -2.38589138e-01 4.99731034e-01
-6.08120561e-01 -4.18596379e-02 -7.07410574e-01 -7.95363784e-01
-4.14279014e-01 -1.22609043e+00 1.46815157e+00 1.69929322e-02
-8.04771721e-01 -8.45963657e-01 -9.17868316e-02 4.66231257e-01
4.95601684e-01 7.04332814e-02 1.12999880e+00 -1.06068194e+00
-4.79403257e-01 -2.80576020e-01 -9.33251232e-02 -1.39938638e-01
2.46819019e-01 -7.05166534e-02 -8.93783331e-01 -4.04737256e-02
-4.68753189e-01 -5.96450984e-01 3.91357422e-01 -3.34968604e-02
4.43244189e-01 -3.10937554e-01 -1.92570060e-01 1.70772344e-01
1.07514000e+00 1.17706554e-02 3.07714939e-01 4.10145551e-01
5.40328562e-01 6.82788253e-01 1.08572459e+00 6.15834557e-02
5.70483387e-01 1.02767324e+00 -1.22053295e-01 -2.90246814e-01
-3.37995619e-01 -6.05224431e-01 3.85797799e-01 6.35087788e-01
7.34528229e-02 6.41353950e-02 -1.04333735e+00 7.44976163e-01
-1.85852671e+00 -8.78506958e-01 1.77092642e-01 2.17867327e+00
1.24306047e+00 5.22687733e-02 1.43288672e-01 -2.61449963e-01
7.90470779e-01 2.09577009e-02 -2.90301144e-01 -6.83400989e-01
-2.43476689e-01 1.16635680e-01 2.38705203e-01 5.38688719e-01
-8.86550546e-01 1.54823780e+00 6.33162594e+00 9.40150380e-01
-9.01546836e-01 2.08241224e-01 5.76904833e-01 -3.58097740e-02
-6.34307027e-01 3.48576576e-01 -8.57345521e-01 4.40856159e-01
6.04006648e-01 -3.83841515e-01 5.13612151e-01 7.76572108e-01
1.87040001e-01 -9.15682092e-02 -1.44099903e+00 5.78910351e-01
1.96805879e-01 -1.04831541e+00 2.24704802e-01 1.37186721e-02
8.05558205e-01 -2.73261160e-01 1.77189112e-01 5.00869811e-01
5.71267426e-01 -8.93393278e-01 7.16680646e-01 6.95407912e-02
1.42622948e+00 -7.23073244e-01 4.94693130e-01 2.18419477e-01
-9.28364396e-01 8.36872905e-02 2.14732978e-02 -2.34159678e-01
8.41959417e-02 4.09095198e-01 -1.11628437e+00 4.92723942e-01
3.12745452e-01 4.48290259e-01 -4.31655973e-01 5.13886034e-01
-5.15585423e-01 8.63975585e-01 -3.76679525e-02 -1.69004798e-01
3.12018156e-01 -2.86373943e-01 6.38541281e-01 1.34181619e+00
2.83403903e-01 1.25638157e-01 3.32035393e-01 7.71991432e-01
-3.28962415e-01 4.29110557e-01 -6.99622691e-01 -3.87625992e-01
1.01269472e+00 1.23147702e+00 -6.96533382e-01 -4.32061911e-01
-3.27111065e-01 1.02907217e+00 3.20753336e-01 3.88727486e-01
-4.58291233e-01 -5.67456603e-01 8.55289519e-01 8.92260671e-02
-9.12783965e-02 -1.95820868e-01 -4.87572014e-01 -1.11911333e+00
6.52357424e-03 -1.15112162e+00 4.08556312e-01 -7.24533439e-01
-1.16891587e+00 3.95084798e-01 -2.75683161e-02 -1.36322296e+00
-6.67458296e-01 -1.69166952e-01 -3.89812261e-01 1.05263007e+00
-1.44561291e+00 -1.46419835e+00 2.45263249e-01 4.63761330e-01
8.97487223e-01 -2.37749651e-01 1.14640021e+00 1.63540199e-01
-4.15530235e-01 9.02399719e-01 -1.61315441e-01 -9.20666158e-02
1.40365076e+00 -1.40586841e+00 6.58667505e-01 9.65990484e-01
1.98247015e-01 9.68524814e-01 8.72513592e-01 -8.51448178e-01
-1.52119398e+00 -1.05186474e+00 1.82242131e+00 -7.62053490e-01
7.34875262e-01 -7.22642958e-01 -5.81851184e-01 7.37819672e-01
2.06134751e-01 -1.60633579e-01 8.15501869e-01 4.11910921e-01
-4.89567816e-01 1.00207120e-01 -1.11384630e+00 1.07323384e+00
9.96015072e-01 -8.70565891e-01 -6.84478819e-01 5.39880633e-01
9.35676992e-01 -4.50037092e-01 -8.03849995e-01 1.72439128e-01
3.26195657e-01 -2.93138593e-01 5.22967219e-01 -7.51868308e-01
7.45732367e-01 -1.46970421e-01 -1.45013958e-01 -1.57700312e+00
-3.04079920e-01 -1.02483404e+00 1.15213633e-01 1.52026677e+00
7.61159778e-01 -4.63237882e-01 4.46159482e-01 9.27412748e-01
-2.67901838e-01 -8.52534592e-01 -7.61321306e-01 -9.90151286e-01
1.15962483e-01 -1.64218768e-01 9.63480115e-01 1.17072618e+00
3.19772124e-01 7.71590948e-01 -4.08212900e-01 -2.73741633e-01
2.57289231e-01 2.93657809e-01 7.95150161e-01 -9.58784044e-01
-1.33582473e-01 -4.31841224e-01 -9.72758606e-02 -7.97454119e-01
3.05755824e-01 -1.10795605e+00 1.16988122e-01 -1.44740891e+00
3.02478611e-01 -6.10338211e-01 2.43301746e-02 7.57896960e-01
-8.02234888e-01 4.82423425e-01 2.03424275e-01 5.27470648e-01
-6.39119744e-01 4.29497957e-01 1.03027797e+00 6.70835376e-02
-1.50988385e-01 -1.14441879e-01 -9.00246859e-01 2.65574574e-01
5.97693622e-01 -4.89503890e-01 -3.76901656e-01 -5.85317850e-01
2.31677577e-01 -1.07228033e-01 -4.63503376e-02 -3.62971544e-01
9.09195468e-02 -4.59047347e-01 1.45047575e-01 -4.65596206e-02
2.83942431e-01 -4.93033767e-01 -3.09850752e-01 -2.93208007e-02
-7.11090505e-01 5.20990431e-01 2.36868616e-02 3.85967284e-01
-1.76156163e-01 4.16725613e-02 3.64167154e-01 -1.09323941e-01
-5.59807360e-01 1.73608556e-01 -4.46416765e-01 3.80118281e-01
7.59549916e-01 -2.97606915e-01 -3.60999823e-01 -4.46902603e-01
-2.11058229e-01 3.24755281e-01 8.18996131e-01 5.96554279e-01
2.28547290e-01 -1.48465288e+00 -7.56603181e-01 2.39279643e-02
7.93692827e-01 -3.30389082e-01 -5.44753790e-01 8.50415230e-01
1.07745240e-02 3.25516880e-01 1.21724769e-01 -4.20440406e-01
-1.14112651e+00 4.83939230e-01 -1.46252781e-01 -2.97665149e-01
-5.00317752e-01 7.71663666e-01 -2.70580620e-01 -6.56508625e-01
6.09292053e-02 2.88097322e-01 1.63218882e-02 -2.11187992e-02
4.25005376e-01 1.95651636e-01 1.48887381e-01 -6.90356970e-01
-2.70479918e-01 2.14147270e-01 -3.69705677e-01 -6.91747189e-01
9.32331145e-01 -9.17563140e-02 -6.56394809e-02 5.99556923e-01
7.93039203e-01 1.32922694e-01 -9.80109453e-01 -5.23504376e-01
5.06225407e-01 -5.48104405e-01 -1.73391551e-01 -1.15463126e+00
-3.33164752e-01 6.78530931e-01 1.99333876e-02 -2.26158276e-01
8.92989159e-01 -2.94337776e-02 9.69714999e-01 3.71139795e-01
4.53650475e-01 -1.37896371e+00 -2.68309796e-03 6.61457241e-01
5.70140660e-01 -1.39141119e+00 -5.79134822e-02 -5.48759758e-01
-9.64154184e-01 7.17532992e-01 6.35478795e-01 4.63035017e-01
-6.54962063e-02 2.21999213e-01 4.41943675e-01 -7.90920947e-03
-8.84943724e-01 -1.59196064e-01 2.47659162e-01 4.40092444e-01
9.38009977e-01 2.37910897e-01 -6.71191096e-01 3.91943514e-01
-2.71927059e-01 -1.52729899e-01 2.88797587e-01 1.21970224e+00
-1.69925794e-01 -1.44266403e+00 -1.91337198e-01 5.07904232e-01
-5.72968960e-01 -5.64402878e-01 -8.79021406e-01 7.18759239e-01
-1.65364146e-01 1.22969782e+00 -8.24125931e-02 -3.22341323e-01
2.74365902e-01 3.76012892e-01 4.70892340e-01 -1.00741160e+00
-1.02703559e+00 1.06661648e-01 5.26314020e-01 -5.23144066e-01
-8.04944187e-02 -8.23035717e-01 -1.04042017e+00 -7.83218965e-02
-4.26201671e-01 2.98499674e-01 8.24309945e-01 1.10839093e+00
5.41267276e-01 4.22130059e-03 5.99294782e-01 -6.22995496e-01
-8.06091130e-01 -1.10573959e+00 -2.53389537e-01 7.03951716e-01
1.08463690e-01 -7.12015092e-01 -2.11953864e-01 4.81159352e-02]
|
[11.566768646240234, 10.211501121520996]
|
efc983fe-97f3-4dca-9dd4-06dd0130ba4e
|
a-novel-plsa-based-traffic-signs
|
1503.06643
| null |
http://arxiv.org/abs/1503.06643v1
|
http://arxiv.org/pdf/1503.06643v1.pdf
|
A novel pLSA based Traffic Signs Classification System
|
In this work we developed a novel and fast traffic sign recognition system, a
very important part for advanced driver assistance system and for autonomous
driving. Traffic signs play a very vital role in safe driving and avoiding
accident. We have used image processing and topic discovery model pLSA to
tackle this challenging multiclass classification problem. Our algorithm is
consist of two parts, shape classification and sign classification for improved
accuracy. For processing and representation of image we have used bag of
features model with SIFT local descriptor. Where a visual vocabulary of size
300 words are formed using k-means codebook formation algorithm. We exploited
the concept that every image is a collection of visual topics and images having
same topics will belong to same category. Our algorithm is tested on German
traffic sign recognition benchmark (GTSRB) and gives very promising result near
to existing state of the art techniques.
|
['Mrinal Haloi']
|
2015-03-23
| null | null | null | null |
['traffic-sign-recognition']
|
['computer-vision']
|
[ 1.16745643e-01 -4.32437837e-01 -3.41381192e-01 -7.90560901e-01
-4.72895443e-01 -3.98069322e-01 1.02524233e+00 -3.02461892e-01
-6.02550983e-01 5.61490417e-01 4.55089718e-01 -5.86395562e-01
-4.68914628e-01 -5.32591760e-01 -2.59880930e-01 -9.83308792e-01
3.07529628e-01 5.04688561e-01 6.47935569e-01 -2.21131489e-01
7.19582438e-01 8.34858775e-01 -2.37380075e+00 3.83783489e-01
4.02640045e-01 9.52733755e-01 6.33681938e-02 8.55554044e-01
-3.88533831e-01 8.74593556e-01 -2.38588780e-01 4.64281626e-02
6.19600713e-01 -4.21669424e-01 -5.16079783e-01 1.98760018e-01
9.76820886e-01 4.17546369e-02 -1.88897207e-01 7.81872272e-01
3.15026730e-01 3.00985128e-01 1.30899715e+00 -1.55574524e+00
-2.20415637e-01 -1.92563713e-01 -4.58754689e-01 3.53377670e-01
-3.14664721e-01 9.40897502e-03 8.58397007e-01 -8.35212171e-01
7.34956205e-01 1.04471648e+00 2.73588598e-01 5.00471294e-01
-7.50277042e-01 -1.13298309e+00 -1.58538476e-01 1.21576822e+00
-1.03037596e+00 -4.66354966e-01 5.44053376e-01 -5.72318912e-01
8.79754901e-01 2.99672604e-01 6.20417356e-01 3.42581302e-01
3.23743582e-01 9.60279942e-01 1.60282302e+00 -5.39610028e-01
4.93564941e-02 5.91570973e-01 1.07061529e+00 7.00903356e-01
4.38256294e-01 5.14074489e-02 -5.07824838e-01 6.73412234e-02
9.84017178e-02 2.78951135e-04 2.37385556e-01 -5.75349450e-01
-1.03964365e+00 8.57562184e-01 2.46437877e-01 6.76312447e-01
-6.51964188e-01 3.78230184e-01 3.46247017e-01 4.55379546e-01
-3.79310071e-01 -1.69012681e-01 -2.06701532e-01 -2.93520838e-01
-8.92566264e-01 1.99998155e-01 5.42618811e-01 7.27954507e-01
5.88431776e-01 2.92370226e-02 -2.79874474e-01 8.00636172e-01
5.29371798e-01 9.65069652e-01 9.22677636e-01 -5.10105968e-01
-3.84205356e-02 5.78747451e-01 -1.00420550e-01 -1.02054191e+00
-2.89360195e-01 -3.45966406e-02 -5.13455391e-01 7.08715200e-01
3.89052719e-01 4.86322463e-01 -1.52534652e+00 8.74840319e-01
1.57046020e-01 3.21611553e-01 1.12374656e-01 7.09267557e-01
1.02520955e+00 5.59143543e-01 1.87612832e-01 1.99195772e-01
1.81716108e+00 -9.04946744e-01 -6.77723229e-01 1.42956391e-01
5.22928059e-01 -9.79647398e-01 4.36463654e-01 8.70245516e-01
-4.68060046e-01 -9.18782413e-01 -9.36518788e-01 7.14338124e-02
-1.01168966e+00 5.58476150e-01 7.95687675e-01 1.31376600e+00
-1.07161045e+00 -1.43538684e-01 -2.54704654e-01 -8.25872779e-01
6.16582096e-01 4.12565798e-01 -7.18557537e-01 -3.62799913e-01
-4.66896147e-01 1.00267279e+00 5.11373818e-01 -1.02774702e-01
-4.59404439e-01 -1.95617318e-01 -5.14663994e-01 -4.16149288e-01
7.74548054e-02 -2.11957127e-01 8.19993317e-01 -6.15371168e-01
-1.23806262e+00 1.16446555e+00 -6.63409412e-01 -6.59285188e-01
2.12740779e-01 2.69430786e-01 -7.57552385e-01 3.86646204e-02
-2.10413545e-01 7.29031920e-01 1.34725857e+00 -1.07978690e+00
-1.09760880e+00 -2.74366319e-01 -8.82884204e-01 -1.94613025e-01
-1.21988162e-01 3.91381294e-01 -9.97756273e-02 -1.62560791e-01
3.76532704e-01 -7.96061456e-01 2.31850162e-01 -1.49337560e-01
3.81681651e-01 -6.09620512e-01 1.53632259e+00 -5.82970858e-01
8.61499906e-01 -2.02331066e+00 -5.46356320e-01 7.63143182e-01
4.48762067e-02 7.80143261e-01 -1.07825518e-01 1.25707164e-01
-1.00733250e-01 -1.95918247e-01 2.15285286e-01 -1.24626406e-01
1.46646455e-01 4.28514779e-01 -6.27614915e-01 5.75685978e-01
2.72382684e-02 8.57322812e-01 -3.52798879e-01 -7.96893299e-01
6.69621825e-01 2.87862271e-01 -8.67389813e-02 -3.19728553e-01
3.00704271e-01 -1.90674718e-02 -4.33710128e-01 6.69157267e-01
9.04109299e-01 3.99530113e-01 -5.66774130e-01 -1.92852184e-01
-4.08164620e-01 -2.27063209e-01 -1.33405828e+00 8.82893622e-01
-2.34566942e-01 1.29754221e+00 -4.72974092e-01 -1.30398083e+00
1.39877415e+00 1.10500574e-01 4.25702006e-01 -9.12576735e-01
2.79089957e-01 5.12006342e-01 1.84523180e-01 -5.96314073e-01
7.15930581e-01 -1.80697933e-01 4.29497391e-01 4.63420451e-01
1.55242115e-01 -1.31978005e-01 4.10965294e-01 -6.42590644e-03
8.19478273e-01 -2.64587879e-01 3.34806204e-01 -9.24420431e-02
9.90366697e-01 2.73331314e-01 1.29631892e-01 7.89743602e-01
-7.32721150e-01 1.25315994e-01 2.11882532e-01 -3.12798440e-01
-1.07811892e+00 -7.52554715e-01 -4.74191010e-01 6.50831401e-01
-1.02194361e-01 6.73899725e-02 9.73395035e-02 -6.29601955e-01
2.79383659e-01 6.95781171e-01 -6.14850044e-01 1.11915767e-02
-6.20392859e-01 -4.06480849e-01 4.94443119e-01 3.80898416e-01
7.67532706e-01 -1.19964230e+00 -7.31981993e-01 -5.05394042e-02
4.01794255e-01 -1.00165498e+00 -6.69257417e-02 -7.86143392e-02
-5.91827154e-01 -1.26595271e+00 -8.49832356e-01 -1.13418102e+00
5.81465423e-01 7.66053677e-01 3.41571748e-01 -5.04620336e-02
-7.64947712e-01 4.75098521e-01 -6.22158647e-01 -9.34796035e-01
-4.84848201e-01 -5.87715447e-01 -2.56867945e-01 5.55738509e-01
9.62789536e-01 4.58171442e-02 -6.39379621e-01 5.32219589e-01
-5.64639091e-01 -3.61118168e-01 8.73177052e-01 7.99828231e-01
2.71230906e-01 1.84906051e-01 2.22270787e-01 -2.66162455e-01
4.27917629e-01 5.12114353e-02 -6.82694733e-01 2.68209338e-01
-5.36176860e-01 8.46664794e-03 2.87558436e-02 -8.99862722e-02
-8.85355711e-01 2.19347343e-01 1.94096670e-01 -4.98530455e-04
-6.09188199e-01 -7.46053830e-02 1.18790545e-01 -5.65582752e-01
3.83529216e-01 6.75829411e-01 3.60560268e-01 -2.82577544e-01
5.42742729e-01 1.26712704e+00 3.94807160e-01 1.27204061e-01
6.93108857e-01 7.74158597e-01 6.47056580e-01 -1.50729871e+00
1.64443567e-01 -1.67333913e+00 -6.78397238e-01 -5.99687755e-01
8.16495836e-01 -6.17811382e-01 -9.08102214e-01 6.83967233e-01
-9.01603401e-01 1.69070214e-01 -1.37958407e-01 7.76171446e-01
-5.88540912e-01 5.27005017e-01 3.24946493e-01 -1.12116122e+00
-7.84218777e-03 -8.64666045e-01 1.01373601e+00 2.05314234e-01
2.44959086e-01 -5.15672863e-01 8.94531012e-02 6.97295904e-01
6.51176929e-01 -1.54776350e-01 5.93885005e-01 -8.65776718e-01
-7.09621549e-01 -7.16428757e-01 -7.31038630e-01 6.63033605e-01
1.21103778e-01 -4.08524759e-02 -1.03358006e+00 1.76563233e-01
-2.01855510e-01 -2.10003555e-01 1.47542405e+00 3.55945677e-01
5.89336395e-01 2.08274677e-01 -4.89207238e-01 2.20089883e-01
1.40496123e+00 6.31143391e-01 9.76858497e-01 3.97412211e-01
3.44300985e-01 6.66239142e-01 9.43714619e-01 -9.31414682e-03
1.56862915e-01 7.30315387e-01 -1.85068607e-01 1.43811807e-01
-6.68241858e-01 1.48643062e-01 3.09649974e-01 4.94800568e-01
1.31305642e-02 -1.63128152e-01 -1.03686178e+00 7.26810515e-01
-1.79607475e+00 -1.39135373e+00 -8.62101197e-01 2.11222959e+00
4.18409593e-02 -3.56553681e-02 4.51300859e-01 6.16515517e-01
2.22295538e-01 -1.17272623e-01 1.37428641e-01 -7.09461987e-01
-2.77570695e-01 4.69324887e-01 9.14359927e-01 4.62050229e-01
-1.21724916e+00 1.17141032e+00 5.24392366e+00 1.03278148e+00
-1.21316528e+00 2.18362153e-01 7.83787742e-02 5.64023197e-01
4.06594574e-01 -5.48939146e-02 -1.19424748e+00 2.29053631e-01
8.45457554e-01 -1.28962174e-01 -2.79694200e-01 7.54337370e-01
4.22313273e-01 -6.26666725e-01 -2.73762405e-01 1.20366120e+00
7.28109300e-01 -1.07369685e+00 2.39106774e-01 2.68773109e-01
6.59505069e-01 1.29766852e-01 8.71514063e-03 2.72340812e-02
1.58934355e-01 -8.83767605e-01 5.01268029e-01 9.31679606e-01
2.59362668e-01 -4.12908047e-01 8.75713050e-01 6.24013841e-02
-1.22883630e+00 -1.82092935e-01 -4.64669347e-01 2.88838625e-01
2.46959776e-01 -6.42955080e-02 -1.04026818e+00 2.59162098e-01
3.43400627e-01 9.49587464e-01 -6.79655850e-01 1.86736560e+00
1.40830904e-01 7.85864592e-01 -3.39312285e-01 -5.62426507e-01
5.88438213e-01 -3.34021658e-01 4.56947416e-01 1.35198009e+00
2.72944033e-01 -1.43501252e-01 -1.76488906e-01 2.46999174e-01
6.96339786e-01 5.37036657e-01 -8.65629196e-01 1.19210020e-01
-1.69261038e-01 9.69585478e-01 -1.14173210e+00 -7.15961337e-01
-2.34896407e-01 6.60939813e-01 -6.87923551e-01 1.03761472e-01
-4.63412970e-01 -7.37803876e-01 6.18097961e-01 -7.79829845e-02
8.46007109e-01 -4.74904329e-01 -3.33673865e-01 -4.38571334e-01
-9.02210325e-02 -5.36631942e-01 3.35378528e-01 -7.29927242e-01
-7.50090003e-01 2.96069294e-01 3.84670496e-02 -1.55462337e+00
1.09476760e-01 -1.27979004e+00 -5.70106804e-01 7.90272236e-01
-1.83012331e+00 -1.48323333e+00 -6.86258495e-01 7.26899743e-01
5.84994733e-01 -8.41593325e-01 4.76211876e-01 3.42539757e-01
4.17716689e-02 2.66494513e-01 3.92383248e-01 -5.04513867e-02
9.21364665e-01 -9.51649010e-01 2.86945589e-02 8.49634886e-01
3.54647875e-01 1.76967770e-01 5.23235917e-01 -6.04233265e-01
-1.11968255e+00 -9.31367159e-01 1.45166659e+00 -5.00163376e-01
7.96249270e-01 -1.97836701e-02 -5.78765392e-01 1.18611574e-01
2.44268663e-02 -5.51842786e-02 5.84483325e-01 -4.33155686e-01
-3.92523825e-01 -5.00256777e-01 -9.81976926e-01 1.12439573e-01
5.56409180e-01 -2.61884987e-01 -8.35782409e-01 4.08511311e-01
-2.43154958e-01 1.76053569e-01 -1.25822037e-01 -7.78502598e-02
8.61491323e-01 -9.05507863e-01 9.19812799e-01 -8.27214479e-01
-3.47209066e-01 -6.89807296e-01 -1.50501162e-01 -8.11524451e-01
-1.07141696e-01 -2.60423630e-01 4.93532240e-01 8.86350632e-01
2.28273600e-01 -8.30893219e-01 9.89253879e-01 3.21594000e-01
-1.41265199e-01 -4.45719846e-02 -1.05830741e+00 -1.17688024e+00
-1.93446532e-01 -7.16563344e-01 2.96080440e-01 3.68061930e-01
-1.77160442e-01 -9.04937610e-02 -3.43750477e-01 -1.97865054e-01
9.00179327e-01 -8.08105320e-02 1.08899772e+00 -1.31796908e+00
1.17989980e-01 -9.10759866e-01 -1.50155437e+00 -7.20347643e-01
3.85796726e-02 -8.77824962e-01 1.38031527e-01 -1.47013414e+00
7.88385943e-02 -4.94679570e-01 -4.04172391e-01 5.17046034e-01
4.23125416e-01 6.71011806e-01 2.01210737e-01 8.26513991e-02
-4.45003062e-01 1.05952069e-01 9.01468098e-01 -2.84022748e-01
-1.02745503e-01 3.13739449e-01 -9.24398452e-02 3.52481395e-01
7.24511385e-01 -3.34066451e-01 -2.52119720e-01 3.92158836e-01
-1.71980545e-01 -7.32656479e-01 6.93754315e-01 -1.04983461e+00
5.60805440e-01 -9.97654423e-02 -9.55933407e-02 -1.14263785e+00
5.17665565e-01 -1.09352005e+00 -1.23075083e-01 7.35647321e-01
-1.06599011e-01 -3.40055317e-01 1.92504272e-01 6.51012599e-01
-5.33792675e-01 -3.77287388e-01 7.61830211e-01 3.15488964e-01
-1.65866721e+00 -9.04650837e-02 -6.92568719e-01 -7.16612220e-01
1.32788503e+00 -9.09901679e-01 -2.27738768e-01 -3.69263768e-01
-4.30825680e-01 1.50648773e-01 -9.20000449e-02 6.27057433e-01
8.83083761e-01 -1.24406636e+00 -7.19682872e-01 5.36348045e-01
7.00597644e-01 -1.03661573e+00 1.63742095e-01 9.77605820e-01
-7.53282249e-01 1.15134263e+00 -6.26060367e-01 -5.19827425e-01
-2.31581044e+00 1.46680564e-01 3.90216559e-02 2.72142023e-01
-4.85723108e-01 6.35698676e-01 -3.81768167e-01 1.28269508e-01
4.26343590e-01 -6.53977096e-01 -7.88368881e-01 2.71645844e-01
6.72733784e-01 5.45185387e-01 1.80131435e-01 -1.42111135e+00
-4.63697642e-01 1.36907101e+00 -1.92541838e-01 -6.22690320e-02
1.13504708e+00 2.41217449e-01 -2.15102709e-03 2.89331049e-01
1.29243195e+00 -3.26292127e-01 -4.54735786e-01 -1.10340446e-01
1.55916691e-01 -6.68422937e-01 3.15489143e-01 -7.32360363e-01
-5.49016893e-01 8.87803495e-01 1.21698856e+00 -9.29193050e-02
8.93383324e-01 -1.10664189e-01 5.64457178e-01 7.87404656e-01
2.39912599e-01 -1.31495774e+00 -1.48867980e-01 4.69195634e-01
8.66856098e-01 -1.68643606e+00 6.59211725e-02 -1.73987165e-01
-8.97406340e-01 1.34692287e+00 -6.83897883e-02 -1.90536946e-01
1.15943635e+00 -3.32406491e-01 3.11479300e-01 -3.94863904e-01
-4.37123835e-01 -9.55260336e-01 1.03930128e+00 6.90221369e-01
3.90325785e-02 1.71780065e-01 -9.17473376e-01 -6.88821375e-02
-7.89723173e-02 3.19916636e-01 2.82614797e-01 9.87934768e-01
-1.12483335e+00 -1.31915498e+00 -5.84650576e-01 7.44655013e-01
1.90526918e-01 2.33633429e-01 -4.34719890e-01 9.88333642e-01
4.13535267e-01 9.92949009e-01 2.28590122e-03 -4.31869417e-01
2.80467361e-01 4.89542514e-01 4.78255600e-01 -1.94328800e-01
-3.11755657e-01 -2.95893937e-01 2.00476542e-01 -4.93158221e-01
-7.94024587e-01 -1.22478759e+00 -1.03646684e+00 1.42607287e-01
-3.42884660e-01 -4.33512144e-02 1.43853271e+00 8.99366319e-01
1.55175701e-02 1.66460369e-02 5.52385330e-01 -3.17443788e-01
-1.54193401e-01 -8.50274742e-01 -7.86000967e-01 4.69726324e-01
3.64819616e-01 -8.08668792e-01 -1.30508810e-01 3.99401218e-01]
|
[7.995189189910889, -0.8101916909217834]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.