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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
51c2fe10-4de3-4c32-b52f-9088febf94f3 | findings-of-the-shared-task-on-troll-meme | null | null | https://aclanthology.org/2021.dravidianlangtech-1.16 | https://aclanthology.org/2021.dravidianlangtech-1.16.pdf | Findings of the Shared Task on Troll Meme Classification in Tamil | The internet has facilitated its user-base with a platform to communicate and express their views without any censorship. On the other hand, this freedom of expression or free speech can be abused by its user or a troll to demean an individual or a group. Demeaning people based on their gender, sexual orientation, religious believes or any other characteristics –trolling– could cause great distress in the online community. Hence, the content posted by a troll needs to be identified and dealt with before causing any more damage. Amongst all the forms of troll content, memes are most prevalent due to their popularity and ability to propagate across cultures. A troll uses a meme to demean, attack or offend its targetted audience. In this shared task, we provide a resource (TamilMemes) that could be used to train a system capable of identifying a troll meme in the Tamil language. In our TamilMemes dataset, each meme has been categorized into either a “troll” or a “not_troll” class. Along with the meme images, we also provided the Latin transcripted text from memes. We received 10 system submissions from the participants which were evaluated using the weighted average F1-score. The system with the weighted average F1-score of 0.55 secured the first rank. | ['Bharathi Raja Chakravarthi', 'Shardul Suryawanshi'] | null | null | null | null | eacl-dravidianlangtech-2021-4 | ['meme-classification'] | ['natural-language-processing'] | [-1.95640102e-01 -1.88634768e-01 2.38491166e-02 -3.85923758e-02
-4.09429073e-01 -1.07810938e+00 9.66336966e-01 2.64912575e-01
-4.06370521e-01 8.17565143e-01 1.31547883e-01 -3.26325566e-01
3.89193386e-01 -7.06765831e-01 -3.05688381e-01 -7.06067979e-01
2.36560270e-01 8.25275257e-02 1.20046996e-01 -1.83078915e-01
7.28190541e-01 4.01268661e-01 -8.33791971e-01 3.24862093e-01
8.72936606e-01 5.79850256e-01 -1.75620183e-01 4.29399163e-01
-2.68414587e-01 6.75983310e-01 -1.23699999e+00 -6.88965619e-01
-4.22495045e-02 -5.49373686e-01 -4.89216417e-01 -1.89467192e-01
6.34733737e-01 -4.16955233e-01 -3.61318529e-01 1.11998963e+00
2.86599785e-01 -1.06028482e-01 6.70433164e-01 -1.00651443e+00
-6.61829174e-01 4.14581329e-01 -5.96176565e-01 2.48943508e-01
6.65810227e-01 -1.37437433e-01 4.28796560e-01 -9.48663235e-01
8.33481967e-01 1.36084867e+00 3.55169088e-01 5.84079325e-01
-9.37609732e-01 -1.08527768e+00 -4.33941662e-01 -1.52240425e-01
-1.30831814e+00 -3.35467041e-01 7.19384968e-01 -6.72612429e-01
6.99632615e-02 4.53164607e-01 5.67888021e-01 1.61685169e+00
6.27238572e-01 5.28463244e-01 1.23436403e+00 -9.71624330e-02
9.96477082e-02 4.21632051e-01 -4.26873006e-02 7.17725158e-01
2.83582568e-01 -4.18757081e-01 -6.76154852e-01 -7.77284026e-01
1.62947714e-01 7.30830580e-02 -1.85416982e-01 5.65431178e-01
-1.03235400e+00 1.03602719e+00 4.57563818e-01 5.48605859e-01
-1.86107326e-02 -1.53927580e-01 5.07393956e-01 4.49274004e-01
5.89182794e-01 6.37945533e-01 1.94571167e-01 -1.79775447e-01
-1.01271498e+00 -6.68483600e-02 1.00177431e+00 2.39180505e-01
5.43154716e-01 -3.71625662e-01 1.36835307e-01 9.60002899e-01
2.01546311e-01 7.07412124e-01 2.68565923e-01 -3.97741556e-01
4.02476132e-01 7.76411831e-01 1.23837069e-01 -1.69913673e+00
-1.22502722e-01 -2.05436110e-01 -8.01324904e-01 1.53855145e-01
4.96105731e-01 -3.78279686e-01 -5.46133697e-01 1.47522748e+00
3.42157573e-01 -2.60507137e-01 -3.82953912e-01 8.02652061e-01
7.51382828e-01 1.09276342e+00 1.30013637e-02 -1.56171516e-01
1.38373101e+00 -4.54529881e-01 -7.91437209e-01 -1.03632465e-01
4.54736590e-01 -1.18314683e+00 8.66942942e-01 3.50556374e-01
-4.86792147e-01 2.09695145e-01 -1.05035245e+00 3.73860180e-01
-7.06403852e-01 -3.64531279e-01 2.18708426e-01 9.68898654e-01
-6.63844287e-01 5.20685911e-01 -4.09542531e-01 -9.23237503e-01
3.06638032e-01 -4.77731936e-02 -5.94817698e-01 1.14569224e-01
-1.47874832e+00 8.43797624e-01 2.13439241e-01 5.25705144e-03
-7.11120546e-01 -3.30367118e-01 -4.53456104e-01 -5.88014066e-01
1.72326207e-01 -6.29111081e-02 5.88340759e-01 -1.06968510e+00
-1.05869269e+00 1.18337774e+00 2.94339329e-01 -8.13128874e-02
8.58942747e-01 -1.54611677e-01 -8.70245337e-01 3.28475296e-01
2.29600921e-01 2.11237162e-01 1.22485924e+00 -9.52332497e-01
-2.89360762e-01 -3.16294849e-01 4.12056968e-02 -2.35476017e-01
-8.36677194e-01 5.63387990e-01 4.57547046e-02 -9.37702358e-01
3.99437174e-02 -1.18731415e+00 4.88111645e-01 -1.04505740e-01
-8.43745768e-01 -7.83506930e-02 1.39423633e+00 -1.00378156e+00
1.49654889e+00 -2.22501111e+00 -2.39721015e-01 1.80009276e-01
2.67531961e-01 4.73290950e-01 2.59050161e-01 1.20875657e+00
4.49827701e-01 8.19704175e-01 -3.97547334e-02 -3.97213362e-03
-2.48939216e-01 -1.41866580e-01 -2.51403332e-01 8.60241354e-01
-1.53095320e-01 2.26559281e-01 -8.64402115e-01 -5.74880838e-01
-2.67214000e-01 4.60984945e-01 -2.03073308e-01 1.49531052e-01
1.65137798e-01 7.61563063e-01 -8.07811379e-01 6.09074414e-01
8.02312076e-01 -4.33077812e-02 2.00784028e-01 2.13235021e-01
-2.03515276e-01 8.58147368e-02 -5.61890244e-01 8.35111499e-01
-3.13554555e-01 1.00371301e+00 3.40500861e-01 9.45850685e-02
1.10536492e+00 3.06248784e-01 1.52738065e-01 -2.84981310e-01
2.88637817e-01 4.84355897e-01 -3.02013129e-01 -5.62825978e-01
5.90002656e-01 7.30431229e-02 -4.58681971e-01 8.78234863e-01
-5.03574073e-01 3.22938740e-01 9.60102677e-02 4.95481580e-01
1.10307348e+00 -3.58270645e-01 2.23581165e-01 -2.98168153e-01
5.95340967e-01 -2.62605458e-01 2.94779539e-01 7.29507565e-01
-3.73967171e-01 3.37478876e-01 6.52645826e-01 -4.36066866e-01
-9.56850290e-01 -1.07046247e+00 -2.54194826e-01 1.24851382e+00
1.48188725e-01 -2.70103812e-01 -8.11855376e-01 -9.38703477e-01
-1.90619633e-01 6.57996595e-01 -4.74940777e-01 -1.05048440e-01
-4.72003698e-01 -5.50442874e-01 1.10590708e+00 -4.14060026e-01
9.72289801e-01 -1.13931501e+00 -4.01415110e-01 2.92338412e-02
-3.89999866e-01 -6.28362060e-01 -7.76740015e-01 -5.43054938e-01
-3.76784056e-01 -1.05121326e+00 -9.47689533e-01 -5.43518543e-01
7.86506653e-01 1.82684921e-02 5.49205601e-01 4.19553548e-01
1.47475719e-01 -9.36802849e-02 -5.49415648e-01 -1.67195916e-01
-8.08143914e-01 2.88562924e-01 2.53976017e-01 4.39623386e-01
1.53093860e-01 -4.84541178e-01 -4.77479219e-01 6.21868670e-01
-1.21210396e+00 -2.22339317e-01 -7.12056234e-02 4.19414252e-01
-5.06201804e-01 -6.87590465e-02 6.11075222e-01 -1.14140522e+00
7.83827841e-01 -9.17813063e-01 -2.49023169e-01 8.04731101e-02
-3.83454114e-02 -5.47605038e-01 6.63665235e-01 -4.93318856e-01
-9.30147529e-01 -3.59025836e-01 2.76701413e-02 8.86303633e-02
-6.72216341e-02 3.93454224e-01 -1.25953570e-01 -1.40423980e-02
5.69436908e-01 5.91627136e-02 -1.57561481e-01 -6.38007224e-01
-5.19458912e-02 1.19795859e+00 1.74597919e-01 -2.98168600e-01
1.07988477e+00 5.66803455e-01 -2.90382981e-01 -1.24137938e+00
-8.09407294e-01 -3.18366021e-01 -3.04414779e-01 -5.98926842e-01
9.48713124e-01 -3.67655337e-01 -7.33990967e-01 9.48898137e-01
-1.33798826e+00 3.00481439e-01 8.58167171e-01 2.03737125e-01
4.11466628e-01 6.28337562e-01 -7.23264694e-01 -9.06081915e-01
-2.93078303e-01 -5.14109910e-01 3.36408496e-01 3.31567317e-01
-5.94405532e-01 -1.10904074e+00 8.62016380e-02 6.24272406e-01
5.45551956e-01 7.16414094e-01 6.95925772e-01 -1.03221047e+00
-8.27379599e-02 -6.14848852e-01 -1.67361289e-01 1.12396374e-01
1.87107101e-01 2.14516565e-01 -8.64888966e-01 -4.98375595e-01
1.10364549e-01 -4.46039528e-01 5.81273317e-01 -3.67964417e-01
5.26632190e-01 -9.69624817e-01 -2.43178919e-01 1.84249148e-01
1.07242715e+00 9.48155150e-02 4.50033784e-01 6.31135106e-01
5.28707325e-01 6.18402600e-01 3.27288389e-01 6.16023481e-01
-2.07693756e-01 4.09110129e-01 4.46660578e-01 2.11300477e-01
2.53549576e-01 -7.24484324e-01 8.61319363e-01 8.65850508e-01
8.85685086e-02 -5.56713104e-01 -1.00916696e+00 2.07461938e-01
-1.44058704e+00 -1.06750786e+00 -4.21993911e-01 2.11085010e+00
5.70411444e-01 7.07186293e-04 1.54178679e-01 -1.20702095e-01
1.28834093e+00 4.76154000e-01 -1.56245723e-01 -7.09843040e-01
-1.73998237e-01 -4.72215503e-01 5.20416677e-01 4.99606222e-01
-1.04319608e+00 8.69740427e-01 5.75358677e+00 9.51856315e-01
-1.28911030e+00 2.19761729e-01 5.08888423e-01 9.33835059e-02
-1.85646251e-01 5.56649640e-03 -7.54974604e-01 9.90038753e-01
8.33307385e-01 -1.57696381e-01 3.97785038e-01 3.46713424e-01
4.51749086e-01 -3.44917238e-01 -6.37830794e-01 7.41223276e-01
3.79578918e-01 -1.16067553e+00 -1.37500349e-03 4.60574776e-01
6.88262761e-01 -1.51032105e-01 1.26067817e-01 -4.13673697e-03
-8.88231397e-02 -9.80468333e-01 7.32406557e-01 3.82586300e-01
7.34027863e-01 -6.64616346e-01 6.80929422e-01 6.73409402e-01
-6.48346484e-01 1.74222931e-01 -1.83238804e-01 -7.92649910e-02
2.10774764e-01 5.18882751e-01 -7.87705243e-01 2.16718316e-01
2.99183160e-01 5.46234250e-01 -8.93264294e-01 9.00604486e-01
-4.77017254e-01 1.09491217e+00 -4.96217787e-01 -4.96025741e-01
4.09850448e-01 -2.33500585e-01 1.14819920e+00 1.30188632e+00
4.14000988e-01 -1.35638773e-01 4.67470177e-02 6.91182315e-01
-1.94643915e-01 2.67085731e-01 -7.89788604e-01 -5.55408478e-01
7.25389540e-01 1.27230847e+00 -7.49911249e-01 5.02675772e-03
-3.56205739e-02 1.23459983e+00 -1.01958416e-01 2.37263247e-01
-8.30658436e-01 -6.79227293e-01 1.82814270e-01 4.80046809e-01
-3.62050712e-01 -3.02152544e-01 6.63656788e-03 -1.21144211e+00
-1.42036289e-01 -7.80367494e-01 5.23597658e-01 -4.24193382e-01
-1.58988738e+00 7.38193929e-01 -2.79946208e-01 -1.06428766e+00
3.66126448e-02 -2.30819270e-01 -7.19783008e-01 6.90295935e-01
-8.47718894e-01 -1.07076716e+00 -7.67640918e-02 4.03773725e-01
2.69621223e-01 -3.67778033e-01 5.80738068e-01 4.58197087e-01
-7.45928347e-01 5.22283494e-01 8.03925991e-02 5.01836598e-01
8.83076727e-01 -6.88319683e-01 -3.55170853e-02 6.77711248e-01
-2.45558843e-01 9.10062671e-01 9.05899644e-01 -9.73741710e-01
-1.13542068e+00 -9.19624388e-01 1.29100811e+00 -7.22451210e-01
1.15648699e+00 -6.40543163e-01 -7.99119771e-01 5.27410448e-01
3.32252979e-01 -5.09323001e-01 8.24217677e-01 -3.68429609e-02
-4.76732314e-01 2.11688206e-01 -1.37598491e+00 6.63283527e-01
6.02736950e-01 -6.95850849e-01 -5.94259202e-01 3.96315396e-01
1.86001241e-01 2.23348707e-01 -6.41224563e-01 -4.94434565e-01
6.90299809e-01 -1.05818570e+00 3.20191354e-01 -6.22181892e-01
5.72756946e-01 -1.82225123e-01 7.86485430e-03 -9.92539346e-01
-5.84070645e-02 -9.97036159e-01 2.53300995e-01 1.86811829e+00
4.19650495e-01 -8.21574807e-01 5.26038587e-01 5.77654004e-01
4.61151063e-01 -1.01264976e-01 -1.03480220e+00 -6.28312945e-01
2.79761374e-01 -8.82117674e-02 1.92677066e-01 1.34669209e+00
8.87571797e-02 3.79704177e-01 -8.89207840e-01 4.20111232e-02
7.65967846e-01 -2.84773439e-01 8.07625234e-01 -1.05082214e+00
2.22082213e-01 -3.12107563e-01 -2.22289443e-01 -5.59401095e-01
3.63761559e-02 -1.04578304e+00 -3.81358713e-01 -1.14463067e+00
3.95292163e-01 -2.94201136e-01 7.76824653e-02 3.83953184e-01
3.26889247e-01 6.15528226e-01 5.14082432e-01 7.39537418e-01
-4.81667876e-01 7.30077624e-02 1.09163415e+00 -9.99565274e-02
-2.14438885e-01 6.02890700e-02 -3.97978514e-01 8.63341689e-01
7.67655194e-01 -8.52090240e-01 2.06700489e-01 6.47461489e-02
7.59704530e-01 -1.79564655e-02 3.30164641e-01 -8.56097639e-01
9.23129544e-03 -3.93876284e-01 2.21236199e-01 -4.99902606e-01
2.13109061e-01 -7.24406242e-01 2.26122350e-01 6.81330979e-01
-1.34398893e-01 -1.49325326e-01 -2.97253460e-01 4.50120032e-01
4.33201380e-02 -3.78483683e-01 1.08857036e+00 3.06516085e-02
-1.63785249e-01 -4.49934378e-02 -9.98715580e-01 1.00003421e-01
1.03268707e+00 -1.91687331e-01 -6.87542677e-01 -7.15595841e-01
-5.63267648e-01 6.16055802e-02 8.21588218e-01 4.43793327e-01
4.08582628e-01 -1.20971251e+00 -9.92313623e-01 -7.02631613e-03
9.85166878e-02 -1.12942648e+00 1.27856925e-01 7.16404676e-01
-8.01484764e-01 -8.14298633e-03 -4.27988410e-01 -1.34796217e-01
-1.29176903e+00 1.14494793e-01 1.73704907e-01 -1.13242250e-02
-4.74026233e-01 4.79161799e-01 4.40401025e-02 -4.17852223e-01
-5.58699258e-02 6.52454376e-01 -3.05852950e-01 5.37778080e-01
8.60455990e-01 6.89325392e-01 -3.10515165e-01 -1.11187100e+00
-6.94696665e-01 1.82186216e-01 -9.87380221e-02 -1.09948315e-01
9.99130070e-01 -1.21305257e-01 -5.49866021e-01 5.96464515e-01
1.42984152e+00 7.60280728e-01 -6.34427011e-01 2.07781300e-01
-3.19915228e-02 -9.16591704e-01 -2.66567498e-01 -8.45319033e-01
-8.27580094e-01 5.21299899e-01 1.26400694e-01 8.94314945e-01
4.56813782e-01 -9.57765058e-03 1.24798989e+00 3.14920336e-01
3.35098296e-01 -1.08517575e+00 2.09005475e-01 7.39348412e-01
1.07052433e+00 -1.06373036e+00 -1.31745160e-01 -2.26621389e-01
-5.43070197e-01 1.20717895e+00 2.65922099e-01 -2.88108494e-02
7.16923416e-01 -1.56996071e-01 4.60888565e-01 -1.71806931e-01
-2.34101921e-01 6.22641802e-01 1.67079076e-01 4.06651646e-01
4.59295124e-01 2.98370831e-02 -9.50862646e-01 1.79213062e-01
-3.38168114e-01 -3.72049063e-01 9.38845754e-01 7.70263791e-01
-6.72698796e-01 -8.39900374e-01 -8.16546500e-01 2.80019969e-01
-9.18237388e-01 2.56196588e-01 -1.10327065e+00 5.76770425e-01
7.64224306e-02 1.37883878e+00 -9.18277353e-02 -5.42869925e-01
-1.89084664e-01 -2.00121608e-02 -2.22022370e-01 -4.17568862e-01
-1.02719903e+00 -1.46996990e-01 3.76326591e-01 -5.68294190e-02
-8.51920247e-02 -5.44759572e-01 -1.10970092e+00 -1.15183091e+00
-2.07052201e-01 2.60221690e-01 7.50129223e-01 7.56734729e-01
-3.80341220e-03 -3.68276715e-01 9.79226530e-01 -4.23862159e-01
-1.49521336e-01 -1.05925560e+00 -8.73982430e-01 5.77692866e-01
9.19906050e-02 -3.42244625e-01 -5.91192126e-01 -2.71323860e-01] | [8.474454879760742, 10.637733459472656] |
eb773ea9-a688-4fbc-8be0-3f4936abf259 | semi-supervised-learning-of-galaxy-morphology | 2011.08714 | null | https://arxiv.org/abs/2011.08714v1 | https://arxiv.org/pdf/2011.08714v1.pdf | Semi-supervised Learning of Galaxy Morphology using Equivariant Transformer Variational Autoencoders | The growth in the number of galaxy images is much faster than the speed at which these galaxies can be labelled by humans. However, by leveraging the information present in the ever growing set of unlabelled images, semi-supervised learning could be an effective way of reducing the required labelling and increasing classification accuracy. We develop a Variational Autoencoder (VAE) with Equivariant Transformer layers with a classifier network from the latent space. We show that this novel architecture leads to improvements in accuracy when used for the galaxy morphology classification task on the Galaxy Zoo data set. In addition we show that pre-training the classifier network as part of the VAE using the unlabelled data leads to higher accuracy with fewer labels compared to exiting approaches. This novel VAE has the potential to automate galaxy morphology classification with reduced human labelling efforts. | ['Yarin Gal', 'Lewis Smith', 'Mizu Nishikawa-Toomey'] | 2020-11-17 | null | null | null | null | ['morphology-classification'] | ['computer-vision'] | [ 2.66966581e-01 3.98089647e-01 4.80357438e-01 -4.53067631e-01
-5.60006261e-01 -8.58199358e-01 9.52951074e-01 -1.98422268e-01
-5.15370131e-01 4.18229431e-01 -1.70842066e-01 -4.14538562e-01
-2.09739525e-02 -8.74266386e-01 -5.17610013e-01 -9.85541582e-01
3.34356010e-01 1.05470300e+00 3.40877771e-01 1.07130222e-01
3.15691344e-02 5.01276374e-01 -1.70450211e+00 2.36857891e-01
3.86533409e-01 9.63877380e-01 4.01136309e-01 9.73914742e-01
1.06140703e-01 1.08135259e+00 -2.28103340e-01 -4.80232418e-01
3.58705610e-01 -3.94552231e-01 -1.17600989e+00 7.03770518e-01
5.44412553e-01 3.25869373e-03 -6.42171726e-02 9.08860803e-01
1.89382777e-01 3.95064503e-01 8.60850930e-01 -9.65515316e-01
-4.00951654e-01 1.28303111e-01 -8.04253668e-02 -1.81730062e-01
-3.80107343e-01 6.47210032e-02 1.13152564e+00 -7.67129421e-01
6.82103813e-01 1.22952545e+00 7.51045227e-01 4.00747836e-01
-1.49213350e+00 5.33220023e-02 -3.91430914e-01 6.19748421e-02
-1.17993438e+00 -4.49226975e-01 7.67915845e-01 -7.29469895e-01
1.17922938e+00 1.12945683e-01 6.84583545e-01 7.46104896e-01
-1.47308081e-01 2.03701362e-01 1.14503479e+00 -6.00032210e-01
2.46360436e-01 3.63196105e-01 -1.58648804e-01 1.06030178e+00
-7.22075254e-02 1.59003899e-01 3.05764452e-02 -9.39469784e-02
1.01628089e+00 -3.79914373e-01 1.07324779e-01 -7.84252346e-01
-1.02253222e+00 1.25200701e+00 4.62814957e-01 2.60250896e-01
-5.99034950e-02 9.91457999e-02 3.78718287e-01 1.88587144e-01
6.90546870e-01 1.11250937e+00 -5.19725502e-01 -3.86810452e-02
-9.82747495e-01 -1.43273532e-01 8.58228028e-01 6.17390394e-01
8.03862274e-01 2.64586300e-01 3.77944171e-01 1.02761865e+00
4.01722312e-01 2.05959678e-01 3.61895293e-01 -1.59025979e+00
-3.34854543e-01 8.17208052e-01 -1.12774178e-01 -6.32315397e-01
-1.78773046e-01 -7.64731824e-01 -5.81156433e-01 6.51907027e-01
7.06094861e-01 3.00383717e-01 -1.04479086e+00 1.49665344e+00
2.67476708e-01 -9.33385715e-02 7.89659023e-02 4.11907285e-01
5.77565253e-01 5.00225067e-01 -2.24193886e-01 6.54645115e-02
1.25456142e+00 -9.50513542e-01 -2.00983688e-01 -2.50000268e-01
7.08467066e-01 -6.94242299e-01 9.91383672e-01 5.51762879e-01
-9.02450740e-01 -7.06534088e-01 -1.02767050e+00 -4.37039614e-01
-5.61888516e-01 2.28451505e-01 9.22544479e-01 6.00424170e-01
-1.39210689e+00 7.50591576e-01 -1.13833129e+00 -5.58416188e-01
5.94605148e-01 7.61242568e-01 -5.29255509e-01 2.42340654e-01
-4.32211429e-01 8.67060363e-01 5.41017830e-01 -1.47931397e-01
-9.51292872e-01 -2.18301430e-01 -8.87491405e-01 1.53913781e-01
1.86669454e-01 -5.18941402e-01 1.43706489e+00 -1.14870322e+00
-1.51466238e+00 1.23498309e+00 1.77306794e-02 -4.71830010e-01
3.73026192e-01 5.58643460e-01 1.29366785e-01 2.60032147e-01
-2.96597220e-02 1.23314285e+00 9.56307888e-01 -1.16547227e+00
-7.94016898e-01 -1.03084072e-01 1.09793127e-01 -7.63536543e-02
-2.31016114e-01 -1.83903903e-01 -1.44705698e-01 -9.02848542e-02
6.27300739e-02 -1.26372218e+00 -5.35193123e-02 -1.32288843e-01
2.01659411e-01 -5.34399927e-01 1.01081419e+00 -5.27060270e-01
3.18502963e-01 -2.25690675e+00 4.44698840e-01 -2.17843920e-01
5.06154180e-01 2.95951515e-01 2.10943017e-02 -2.70085838e-02
-4.99988049e-02 3.58198360e-02 -2.66172320e-01 -5.62632859e-01
-1.27814323e-01 7.46137202e-01 -2.55881883e-02 5.14579117e-01
2.95486510e-01 9.60797191e-01 -9.42904174e-01 -4.34569716e-01
5.97409308e-01 4.31379378e-01 -7.04929650e-01 3.62695068e-01
-2.80380756e-01 6.40215337e-01 2.77223408e-01 2.23803341e-01
3.76252919e-01 -5.91671169e-01 1.88600808e-01 1.22454099e-01
-9.75914821e-02 2.22565144e-01 -6.88620806e-01 1.38707936e+00
-6.17824078e-01 9.97731328e-01 9.20128003e-02 -1.19834423e+00
7.78996766e-01 2.90876508e-01 8.81677121e-02 -3.88908386e-02
3.79893839e-01 2.47365132e-01 2.46255964e-01 -1.16037369e-01
2.21982405e-01 -3.61854315e-01 2.31702160e-02 1.97319582e-01
8.71697903e-01 -4.68415856e-01 2.69493908e-01 1.47853062e-01
8.62303793e-01 2.07134932e-01 2.29402810e-01 -3.43285173e-01
5.59574604e-01 2.26737186e-01 3.70296180e-01 7.05481112e-01
-6.13576248e-02 4.35844868e-01 6.33134305e-01 -5.99342704e-01
-1.56002116e+00 -8.21095765e-01 -2.70511687e-01 1.01087606e+00
-5.41403174e-01 -2.54911572e-01 -8.12848389e-01 -7.90770769e-01
-2.42969289e-01 6.16674304e-01 -7.81259418e-01 -1.29306898e-01
-2.20799312e-01 -7.14534223e-01 4.24608320e-01 4.91696775e-01
6.11710608e-01 -1.04876232e+00 -4.74307209e-01 -3.47138904e-02
6.57011122e-02 -1.00521481e+00 1.58124238e-01 4.94144529e-01
-7.37397075e-01 -9.09918129e-01 -3.32800925e-01 -1.01251900e+00
7.78888702e-01 9.25392956e-02 1.13072968e+00 2.36592684e-02
-1.75510645e-01 3.79687518e-01 -2.88078308e-01 -2.85636276e-01
-7.93130577e-01 3.23006451e-01 -1.48488969e-01 -7.97458962e-02
1.17418677e-01 -5.54809690e-01 -2.44171411e-01 9.57473814e-02
-7.35097587e-01 1.46358326e-01 3.84472758e-01 1.26000762e+00
5.92074275e-01 4.45594013e-01 1.91956609e-01 -1.11893559e+00
-4.15895097e-02 -2.74379730e-01 -9.79163766e-01 -1.62961692e-01
-8.24761152e-01 3.29314947e-01 7.40130842e-01 -4.37720537e-01
-1.28859162e+00 3.55397314e-01 -4.23520505e-01 -3.88619274e-01
-4.35025901e-01 2.24394664e-01 -5.88426813e-02 -4.14685607e-01
5.07290006e-01 -2.25373745e-01 2.13109717e-01 -6.59880400e-01
4.44039941e-01 8.08843076e-01 8.06899488e-01 -4.92977537e-02
6.43937051e-01 6.51855528e-01 4.51716870e-01 -7.70209253e-01
-1.00798917e+00 -6.19072378e-01 -1.15735865e+00 -9.46061164e-02
1.18109274e+00 -6.74814284e-01 -5.26049733e-01 4.89015430e-01
-7.81803131e-01 -4.13842618e-01 -6.40739143e-01 2.83813894e-01
-7.99232364e-01 3.02223206e-01 -6.09737396e-01 -6.57277822e-01
1.34654820e-01 -1.29892063e+00 1.02940464e+00 8.66143405e-02
-1.57186434e-01 -1.29793155e+00 7.18233828e-03 4.63633358e-01
1.02553219e-01 -7.63574662e-03 1.03816640e+00 -8.04928184e-01
-5.08486092e-01 -2.81632662e-01 -1.60321414e-01 5.92564106e-01
2.31132619e-02 -1.08210452e-01 -1.27897525e+00 -3.97998869e-01
2.98701137e-01 -5.62009990e-01 1.03906333e+00 2.46232718e-01
9.20972347e-01 -1.16759213e-02 -9.18025747e-02 8.48705769e-01
1.21467078e+00 1.09340645e-01 4.79181528e-01 3.03443819e-01
8.26191306e-01 7.98408866e-01 2.12589040e-01 -2.12777868e-01
2.95748431e-02 5.48413277e-01 4.91872430e-01 -2.06664696e-01
-3.12429607e-01 -1.67002156e-01 1.23331333e-02 9.66266096e-01
-8.48486200e-02 -4.16527927e-01 -8.67410839e-01 5.37455678e-01
-1.81085682e+00 -7.98655391e-01 -3.61903459e-01 1.97882676e+00
3.98179650e-01 1.25717863e-01 1.27928331e-01 2.86667198e-01
3.74965936e-01 -8.18938538e-02 -1.85763493e-01 -6.10214531e-01
-1.04984008e-01 2.74927288e-01 6.38681233e-01 7.68725455e-01
-1.28394294e+00 1.18361843e+00 7.01947308e+00 7.30713964e-01
-8.52111101e-01 5.09731948e-01 6.50014997e-01 2.68523633e-01
1.13912843e-01 3.57000977e-01 -5.33933580e-01 1.21360600e-01
1.23785925e+00 5.38105249e-01 6.68680906e-01 1.02535522e+00
-2.27375686e-01 -4.78987806e-02 -9.75415528e-01 8.21017027e-01
9.25079957e-02 -1.36318338e+00 -1.36565432e-01 6.08219445e-01
6.05535567e-01 7.38891438e-02 5.39899431e-02 4.13000286e-01
6.53455973e-01 -9.67421412e-01 5.09690404e-01 3.60903025e-01
9.07521725e-01 -9.10145164e-01 9.75880206e-01 5.13187587e-01
-9.02684331e-01 5.09663038e-02 -7.19671547e-01 -3.24522167e-01
-2.28264168e-01 1.57393068e-01 -1.32016683e+00 2.33293712e-01
4.32213068e-01 5.67061126e-01 -1.19769120e+00 7.68611193e-01
-2.96184361e-01 9.49164093e-01 -2.47482255e-01 2.16141343e-01
4.07965332e-01 -3.45722526e-01 4.56736416e-01 1.02540183e+00
6.80754781e-02 -2.57508159e-01 9.70926061e-02 7.50636160e-01
1.08768186e-02 -2.78072894e-01 -7.06360340e-01 -3.46733451e-01
-1.34955049e-01 1.68422830e+00 -1.17691743e+00 -5.22954822e-01
-4.13848966e-01 1.06942654e+00 6.40963256e-01 -1.91733311e-03
-3.87916565e-01 -2.41166994e-01 1.14293350e-02 2.84763481e-02
6.04367435e-01 -1.95906237e-01 8.97165835e-02 -9.19645011e-01
-6.79710388e-01 -6.33360386e-01 2.61583656e-01 -1.06363928e+00
-1.16846442e+00 6.14068866e-01 -1.08416356e-01 -7.67287850e-01
-5.44177175e-01 -1.18979084e+00 -3.46034855e-01 6.21524990e-01
-7.20669150e-01 -1.47664022e+00 -1.94102421e-01 -7.76976198e-02
4.60621595e-01 -4.94803846e-01 1.01554000e+00 -4.23607342e-02
-1.88028499e-01 9.83054191e-02 2.85734951e-01 1.94142848e-01
3.64950150e-01 -1.96830225e+00 3.69566470e-01 7.60141075e-01
6.68787658e-01 2.94413328e-01 6.80928707e-01 -5.08077443e-01
-7.59396791e-01 -9.90506053e-01 7.09723055e-01 -9.18337584e-01
6.34744287e-01 -8.14947248e-01 -9.38443482e-01 9.97164130e-01
3.72846484e-01 9.84256864e-02 6.17501259e-01 1.52966782e-01
-1.77132919e-01 2.60877281e-01 -1.07674193e+00 2.21207485e-01
8.73785496e-01 -8.78402591e-01 -7.43886292e-01 2.81875849e-01
5.70958793e-01 -4.54485007e-02 -9.12987113e-01 2.69220173e-01
3.60905319e-01 -7.02006817e-01 8.87134373e-01 -5.08754015e-01
3.08761686e-01 -3.27720553e-01 -1.02448553e-01 -1.39614224e+00
-7.23622978e-01 -3.48226070e-01 1.65686682e-01 1.39972019e+00
3.08827728e-01 -4.88381326e-01 8.89165998e-01 1.84819415e-01
-3.79484706e-02 -3.36460322e-01 -8.99699628e-01 -6.49104953e-01
2.20994160e-01 -2.05476895e-01 3.27336937e-01 7.42028654e-01
-5.09651363e-01 7.30778635e-01 -2.35235214e-01 -1.61450386e-01
6.01181984e-01 3.61583978e-02 6.22704625e-01 -1.73915422e+00
-7.06151605e-01 -1.92675769e-01 -8.30697477e-01 -6.75847292e-01
2.86739677e-01 -1.43473411e+00 1.13920517e-01 -1.29430485e+00
3.07509422e-01 -2.23896801e-01 -1.30633324e-01 4.11235541e-01
1.63933590e-01 6.47120655e-01 -1.95152372e-01 1.71297014e-01
-7.07958937e-01 2.49921873e-01 9.59531128e-01 -6.92545325e-02
2.27825120e-01 -1.37523904e-01 -3.38055164e-01 1.07763219e+00
6.67324781e-01 -5.52139759e-01 -4.25249368e-01 -1.47310749e-01
2.90893793e-01 -3.77285957e-01 6.15750968e-01 -8.90561044e-01
-8.71643424e-04 3.45329702e-01 7.13478148e-01 -4.21503752e-01
5.20394862e-01 -4.90242004e-01 2.24013075e-01 1.46315143e-01
-1.98567167e-01 -1.96797878e-01 3.11847292e-02 5.75342774e-01
-1.71449661e-01 -7.97484279e-01 1.13177431e+00 -6.12927079e-01
-5.73746026e-01 -1.30671099e-01 -8.27029526e-01 -1.92962423e-01
6.68898880e-01 -2.95927245e-02 -8.13916251e-02 -3.40157479e-01
-9.99455035e-01 -1.99065655e-01 9.69512284e-01 1.11483308e-02
1.22891322e-01 -1.06805217e+00 -2.77250439e-01 3.84963840e-01
1.06782898e-01 1.97655514e-01 -1.36009753e-01 6.08077884e-01
-7.00593710e-01 5.45109689e-01 -3.01375329e-01 -8.99072647e-01
-1.32981491e+00 7.07517922e-01 5.99355042e-01 -4.26101863e-01
-4.50137198e-01 1.19459426e+00 4.02748466e-01 -8.82038772e-01
-1.50107995e-01 1.06806926e-01 -2.71138817e-01 5.52208535e-02
8.52499157e-02 4.10142899e-01 1.62977248e-01 -9.52353179e-01
1.97347440e-02 2.02040315e-01 -5.70727065e-02 -4.16944325e-01
1.51567197e+00 -8.96849632e-02 -2.86410749e-01 6.80802941e-01
1.22788322e+00 2.24906076e-02 -1.51795244e+00 1.00768156e-01
1.78426117e-01 -4.53139395e-01 5.43784380e-01 -5.42084098e-01
-8.48513663e-01 1.12632656e+00 6.70564353e-01 4.24051344e-01
8.45881522e-01 4.43976909e-01 3.07911873e-01 5.17531097e-01
2.85436809e-01 -9.23166096e-01 1.64112914e-02 4.57647175e-01
4.45481718e-01 -1.39655662e+00 -1.89940959e-01 -3.41148764e-01
-5.41611969e-01 9.64076281e-01 3.46664608e-01 -3.17219228e-01
3.44718575e-01 1.25674590e-01 -2.81829871e-02 -5.25712132e-01
-8.61852825e-01 -4.91343141e-01 3.06374699e-01 7.55732656e-01
3.00059378e-01 6.96187168e-02 2.25886703e-01 1.61158919e-01
-3.69185090e-01 -4.01242018e-01 6.08118355e-01 6.32566631e-01
-5.16239047e-01 -1.46696782e+00 -1.49192587e-01 7.31638312e-01
-3.97670597e-01 9.39666927e-02 -8.58269632e-01 6.71868980e-01
5.96635699e-01 9.70165908e-01 3.82418394e-01 -2.50961214e-01
-2.42206529e-01 5.83571970e-01 8.02400410e-01 -9.41452563e-01
-3.96936506e-01 3.41686070e-01 2.63345003e-01 -6.12717085e-02
-4.72191870e-01 -7.89196014e-01 -8.55157852e-01 -6.57716319e-02
-6.30648553e-01 2.52270103e-01 3.96305710e-01 9.59357858e-01
6.78198189e-02 5.76974988e-01 4.30424869e-01 -1.08393347e+00
-2.88422942e-01 -1.10011852e+00 -6.03219807e-01 3.83052021e-01
2.19831198e-01 -9.93207395e-01 -6.65878773e-01 3.24599117e-01] | [8.045433044433594, 2.9098246097564697] |
c9e51d27-4b9b-481e-b745-0d4969179aa2 | greedy-transition-based-dependency-parsing | null | null | https://aclanthology.org/J17-2002 | https://aclanthology.org/J17-2002.pdf | Greedy Transition-Based Dependency Parsing with Stack LSTMs | We introduce a greedy transition-based parser that learns to represent parser states using recurrent neural networks. Our primary innovation that enables us to do this efficiently is a new control structure for sequential neural networks{---}the stack long short-term memory unit (LSTM). Like the conventional stack data structures used in transition-based parsers, elements can be pushed to or popped from the top of the stack in constant time, but, in addition, an LSTM maintains a continuous space embedding of the stack contents. Our model captures three facets of the parser{'}s state: (i) unbounded look-ahead into the buffer of incoming words, (ii) the complete history of transition actions taken by the parser, and (iii) the complete contents of the stack of partially built tree fragments, including their internal structures. In addition, we compare two different word representations: (i) standard word vectors based on look-up tables and (ii) character-based models of words. Although standard word embedding models work well in all languages, the character-based models improve the handling of out-of-vocabulary words, particularly in morphologically rich languages. Finally, we discuss the use of dynamic oracles in training the parser. During training, dynamic oracles alternate between sampling parser states from the training data and from the model as it is being learned, making the model more robust to the kinds of errors that will be made at test time. Training our model with dynamic oracles yields a linear-time greedy parser with very competitive performance. | ['Miguel Ballesteros', 'Noah A. Smith', 'Chris Dyer', 'Yoav Goldberg'] | 2017-06-01 | null | null | null | cl-2017-6 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [ 2.45358825e-01 3.71263236e-01 -2.65100092e-01 -3.21695805e-01
-8.73318434e-01 -7.98959196e-01 1.94612250e-01 4.70820963e-01
-6.27088130e-01 2.61804581e-01 3.86347145e-01 -1.05122876e+00
4.60613757e-01 -1.33000886e+00 -9.69377875e-01 -5.20821750e-01
-3.26603532e-01 5.26162922e-01 6.40388846e-01 -2.25711390e-01
1.27879560e-01 3.59271675e-01 -1.26727331e+00 3.63626093e-01
1.21571764e-01 4.51027542e-01 5.57449460e-01 1.06403995e+00
-9.59413946e-01 8.31241071e-01 -4.35786694e-01 -5.04060924e-01
1.22649185e-01 -1.66666344e-01 -9.67981637e-01 -3.44302237e-01
1.16331048e-01 -3.83410454e-01 -4.62198734e-01 1.08996665e+00
9.07767341e-02 -5.59726469e-02 4.48457226e-02 -5.87627709e-01
-5.37587762e-01 1.33749211e+00 1.78142190e-01 4.43619072e-01
2.56691188e-01 3.47770154e-01 1.34315825e+00 -6.66837633e-01
8.42189431e-01 1.28608036e+00 6.31715536e-01 7.49865174e-01
-1.20332110e+00 -1.35683224e-01 6.54639542e-01 -2.16745600e-01
-7.32688308e-01 -4.72137541e-01 4.65667099e-01 -2.99574286e-01
1.65098429e+00 1.61259890e-01 5.21074057e-01 8.74608517e-01
3.45620334e-01 8.00562680e-01 6.28410101e-01 -7.80854762e-01
3.28823596e-01 -1.94937110e-01 8.15201998e-01 1.09552431e+00
2.38167733e-01 4.55815122e-02 -4.94561881e-01 -4.00425792e-01
6.06614947e-01 -8.84642974e-02 1.03932694e-01 -3.11473638e-01
-8.04112017e-01 9.75523651e-01 1.32483289e-01 3.51815194e-01
-2.09742621e-01 4.74093109e-01 7.69666851e-01 3.60706478e-01
1.56384557e-01 7.07179904e-02 -8.17919016e-01 -2.56391525e-01
-7.28538632e-01 1.69972003e-01 1.01362348e+00 8.26672971e-01
7.32250690e-01 2.17801690e-01 -2.69956887e-02 4.30271029e-01
3.76809895e-01 3.47778946e-01 7.88726747e-01 -5.63488722e-01
8.77123594e-01 4.32246745e-01 -1.26623094e-01 -2.51771480e-01
-2.76250124e-01 -1.56947881e-01 -1.20835155e-01 -1.42974705e-02
3.64637196e-01 -1.99863508e-01 -1.07820201e+00 2.06282139e+00
9.02972743e-02 8.29372182e-02 1.95437327e-01 1.94812536e-01
5.29260337e-01 9.52723205e-01 2.30066925e-01 -1.43561497e-01
1.33840525e+00 -8.27532172e-01 -5.17825007e-01 -7.70252764e-01
1.18283665e+00 -2.20569581e-01 9.62956309e-01 -1.56992167e-01
-1.56498265e+00 -3.07802677e-01 -1.14034784e+00 -1.56103343e-01
-4.89031553e-01 -4.14486915e-01 8.40283990e-01 7.39066422e-01
-1.54384732e+00 8.05296481e-01 -1.39051199e+00 -5.52882887e-02
-1.19973615e-01 5.16434371e-01 -4.77767080e-01 4.35253084e-02
-1.20588267e+00 8.13974857e-01 5.65561295e-01 2.67968684e-01
-6.82845414e-01 -2.77244657e-01 -1.31771398e+00 3.00543875e-01
3.00397843e-01 -4.05196697e-01 1.79394925e+00 -5.92377901e-01
-1.48024642e+00 7.46776879e-01 -5.87213457e-01 -6.55827284e-01
-5.51353619e-02 -4.83659990e-02 -9.78663489e-02 -2.57454157e-01
-7.33228177e-02 2.80255586e-01 3.55145246e-01 -6.31136775e-01
-6.58807755e-01 -4.46714550e-01 2.07231581e-01 -9.62904096e-02
-1.66078731e-01 2.53288060e-01 -4.68802273e-01 -2.53845513e-01
4.58653063e-01 -8.12639952e-01 -3.87648314e-01 -3.23441625e-01
-4.57027763e-01 -2.20087528e-01 2.04600245e-01 -5.78012884e-01
1.52606392e+00 -2.17874575e+00 4.16141450e-02 2.93737091e-02
-1.25860780e-01 5.13477087e-01 -2.88883477e-01 6.31743193e-01
-2.23701864e-01 4.16382045e-01 -3.37898642e-01 -4.11110938e-01
7.22018778e-02 6.51890755e-01 -5.94184339e-01 1.50729984e-01
4.10514921e-01 1.03335118e+00 -8.87382150e-01 -1.51572183e-01
-9.43223834e-02 1.18061855e-01 -5.25779307e-01 2.36821115e-01
-4.48885500e-01 -3.77091408e-01 -3.71692300e-01 3.75300527e-01
3.08832169e-01 8.66791531e-02 6.46299005e-01 5.59009314e-01
-2.63194650e-01 1.25536191e+00 -1.16705430e+00 1.44698322e+00
-4.99737769e-01 3.77483606e-01 1.05251424e-01 -7.62910962e-01
6.59607112e-01 4.29345250e-01 -3.32330436e-01 -5.98614991e-01
-1.44872949e-01 3.09687138e-01 -2.34921239e-02 -2.14262098e-01
7.22353220e-01 -2.72950739e-01 -3.52967113e-01 5.00443518e-01
2.31268108e-01 3.03164035e-01 2.90052772e-01 2.27053583e-01
1.50249743e+00 2.08712712e-01 1.19069703e-01 1.58012509e-02
4.15770501e-01 -1.84302777e-01 8.56390715e-01 9.41916406e-01
2.78124958e-02 2.43009403e-01 9.69312131e-01 -6.71319544e-01
-9.48149979e-01 -1.03098857e+00 1.05352096e-01 1.52798855e+00
-4.59257483e-01 -6.40172839e-01 -9.39476132e-01 -4.87996489e-01
-2.27298692e-01 9.40633297e-01 -5.32387018e-01 -1.87573075e-01
-1.37152910e+00 -4.51466709e-01 6.88441575e-01 1.02713263e+00
-2.40401298e-01 -1.48681033e+00 -8.80580902e-01 7.74083257e-01
2.00475216e-01 -8.77665937e-01 -3.63567680e-01 1.11646843e+00
-1.36193061e+00 -7.64567256e-01 -3.24430838e-02 -1.11833370e+00
6.08065665e-01 -3.30799252e-01 1.15895104e+00 2.26150289e-01
1.27467543e-01 8.92481208e-03 -5.98128252e-02 -1.76838025e-01
-9.27048922e-01 1.73227802e-01 -2.52789944e-01 -5.26302040e-01
3.66395742e-01 -3.33671331e-01 3.03981341e-02 -4.08365816e-01
-8.98762524e-01 -1.42872944e-01 3.14377427e-01 8.32370341e-01
5.22017241e-01 -1.80821076e-01 -2.80678809e-01 -1.25543094e+00
5.49643576e-01 -3.28744113e-01 -8.61218989e-01 1.43880829e-01
-2.66403258e-01 7.49525189e-01 8.35594416e-01 -2.73831546e-01
-7.36972809e-01 2.48725191e-02 -6.11757398e-01 1.14797130e-02
1.40347064e-01 6.38749063e-01 -1.39148504e-01 4.87003893e-01
3.34739059e-01 3.32359701e-01 -3.31538506e-02 -7.92902827e-01
4.30975467e-01 3.27183157e-01 5.42521775e-01 -7.12547004e-01
5.50097585e-01 -5.07988129e-03 -2.33271897e-01 -5.21208823e-01
-2.92491883e-01 -1.02238573e-01 -6.24968708e-01 4.48352605e-01
8.07870567e-01 -6.44658983e-01 -5.39352000e-01 5.17729938e-01
-1.44643450e+00 -7.09917247e-01 -5.81359446e-01 -4.90134172e-02
-3.39484185e-01 2.79231399e-01 -1.33177054e+00 -9.53180373e-01
-4.69443649e-01 -1.36245358e+00 7.78941631e-01 -1.05828168e-02
-3.32538150e-02 -1.14335716e+00 2.74113625e-01 -5.21019876e-01
3.09654951e-01 -2.44136050e-01 1.68069327e+00 -1.05222476e+00
-7.37474680e-01 -3.40369940e-01 3.85327637e-01 4.28661078e-01
-4.44022000e-01 2.11288477e-03 -7.35616207e-01 -1.70120046e-01
-4.54255864e-02 -7.37373829e-02 9.08972502e-01 2.78949201e-01
7.80721843e-01 -2.98714787e-01 -4.67572242e-01 5.09381533e-01
1.40592980e+00 3.00400108e-01 4.58557814e-01 5.08098722e-01
4.14273202e-01 4.63284582e-01 6.00185655e-02 -6.95969984e-02
4.91888165e-01 1.49239004e-01 5.09851038e-01 5.01406908e-01
8.64540115e-02 -6.80043042e-01 9.95843172e-01 1.22885728e+00
4.05678779e-01 -3.24248552e-01 -9.97217238e-01 6.22584045e-01
-1.55654216e+00 -7.43766844e-01 -5.94783463e-02 2.48881102e+00
7.68351436e-01 7.95015395e-01 -1.17691062e-01 1.93647176e-01
6.91444218e-01 4.58588570e-01 -3.40963811e-01 -1.43589425e+00
2.29590908e-01 7.46340692e-01 7.96929479e-01 8.58200550e-01
-9.69451129e-01 1.38836730e+00 7.09556532e+00 1.11845687e-01
-1.22193289e+00 1.90569401e-01 2.30614588e-01 -4.24115770e-02
-6.79651797e-01 4.19904530e-01 -1.25749075e+00 3.44890386e-01
1.75666976e+00 2.40037009e-01 2.99563378e-01 9.20410275e-01
-4.44678873e-01 1.58908039e-01 -1.28710604e+00 2.71127135e-01
-3.12825650e-01 -1.36675608e+00 -1.31106824e-01 -6.24006838e-02
-1.82273000e-01 4.35960472e-01 -1.20284110e-01 6.61876500e-01
7.02667117e-01 -7.45354235e-01 1.11706853e+00 1.36194363e-01
6.54224157e-01 -5.70077002e-01 6.14140213e-01 4.77213383e-01
-1.27087593e+00 -4.30784613e-01 -5.90075374e-01 -2.57660031e-01
3.83416325e-01 3.04845035e-01 -6.02728605e-01 5.43040149e-02
2.66945839e-01 -1.85291294e-03 -3.82481754e-01 4.71595913e-01
-5.80797970e-01 8.70807827e-01 -3.98149192e-01 -3.53812099e-01
6.66798711e-01 -3.25052738e-02 4.74029064e-01 1.39417243e+00
1.09253950e-01 -3.57766114e-02 8.99024978e-02 6.00818515e-01
2.64400188e-02 -1.39106870e-01 -6.38546884e-01 -3.40759516e-01
5.46914816e-01 7.43372142e-01 -6.13063157e-01 -3.74258578e-01
-5.46668589e-01 7.23931909e-01 8.16005945e-01 1.16231807e-01
-3.69291753e-01 -3.54744971e-01 6.84473038e-01 9.43096280e-02
8.34732473e-01 -5.36140800e-01 1.69401406e-03 -9.29107070e-01
2.49018818e-01 -6.54701650e-01 4.19216365e-01 -5.00853121e-01
-6.99454844e-01 7.77722180e-01 -1.80942506e-01 -2.79154599e-01
-5.11718571e-01 -9.63227689e-01 -8.44288409e-01 9.96902585e-01
-1.39755070e+00 -6.61000192e-01 6.39635623e-01 7.35272914e-02
5.67235470e-01 1.00991741e-01 1.35996258e+00 -1.69280544e-01
-7.49352992e-01 6.08716607e-01 3.33924741e-02 4.96964872e-01
-2.77274828e-02 -1.44241786e+00 1.57307732e+00 1.08524215e+00
3.22581083e-01 1.15482616e+00 4.87123430e-01 -6.54319584e-01
-1.81781197e+00 -8.02287102e-01 1.39931822e+00 -3.93619210e-01
9.20488298e-01 -7.91694760e-01 -1.24292088e+00 1.38046384e+00
1.99686200e-03 1.62369013e-01 5.09754360e-01 2.60499299e-01
-6.10463202e-01 1.93397731e-01 -6.59675181e-01 6.22893095e-01
9.04981971e-01 -7.06570566e-01 -9.50657725e-01 1.33901209e-01
1.08125508e+00 -6.35461628e-01 -3.64127666e-01 -1.37395337e-01
5.43373227e-01 -8.24450850e-01 6.56917334e-01 -1.05298924e+00
1.98081315e-01 1.17516972e-01 -2.72774935e-01 -1.02467537e+00
-4.77796763e-01 -6.55775249e-01 -3.40914875e-01 9.85137284e-01
7.28140950e-01 -7.78111577e-01 8.59380066e-01 8.19649100e-01
-3.18611473e-01 -9.33264375e-01 -9.66772974e-01 -5.25518417e-01
5.07480621e-01 -7.79325128e-01 6.55414343e-01 2.65083224e-01
-1.48713328e-02 3.78511041e-01 1.97505131e-01 1.93908766e-01
1.19451538e-01 -1.27665445e-01 2.79659212e-01 -1.07785428e+00
-5.59727252e-01 -3.05674702e-01 -4.03085023e-01 -1.27659261e+00
3.71397555e-01 -1.24857175e+00 3.96707594e-01 -1.45702362e+00
3.40545997e-02 -4.57251430e-01 -4.95267093e-01 7.60724843e-01
-1.70503870e-01 -5.88288486e-01 2.43175536e-01 -1.09270573e-01
-2.11111367e-01 -4.17188480e-02 4.63362098e-01 7.36907423e-02
-1.69970572e-01 -1.13961345e-03 -5.25071084e-01 8.81922841e-01
5.96784234e-01 -8.52655888e-01 -2.46565089e-01 -9.28415895e-01
5.35180092e-01 5.45666695e-01 -6.87713847e-02 -6.19741917e-01
3.64986122e-01 1.36359110e-01 -9.70302373e-02 -5.38327038e-01
2.04456747e-01 -3.53263855e-01 -2.65950769e-01 6.70472622e-01
-5.16390085e-01 7.55749583e-01 2.02020735e-01 5.79954088e-01
1.58578500e-01 -8.66984367e-01 4.89732593e-01 -6.73055172e-01
-6.64947987e-01 2.74659723e-01 -9.05943274e-01 1.64666921e-01
3.32288980e-01 -2.45058089e-01 -1.29126295e-01 1.36485830e-01
-8.81080985e-01 -4.25889380e-02 4.03350443e-01 3.71066958e-01
3.23652476e-01 -7.51969039e-01 -1.72438234e-01 6.33771062e-01
-1.19754769e-01 1.63013265e-01 -2.00543866e-01 2.04805002e-01
-7.36662447e-01 5.24607003e-01 2.03402430e-01 -1.73913851e-01
-1.19604421e+00 6.70362651e-01 2.16687337e-01 -7.32364655e-01
-8.61629009e-01 1.15298593e+00 -1.15075126e-01 -7.13827610e-01
3.20259690e-01 -8.62487793e-01 7.46706948e-02 -1.57897785e-01
7.33893037e-01 -1.61556095e-01 3.28810394e-01 -4.83291328e-01
-4.96811897e-01 1.36431918e-01 -3.87793213e-01 -3.29604506e-01
1.41870272e+00 3.36388081e-01 -4.43930417e-01 9.51632917e-01
1.20126963e+00 4.29933406e-02 -8.79837334e-01 -3.81757289e-01
5.39386928e-01 1.74434468e-01 -7.52806440e-02 -3.76977801e-01
-6.62506938e-01 1.17736900e+00 3.16794515e-01 2.75570273e-01
4.23420608e-01 -1.56829536e-01 1.23594058e+00 7.19147265e-01
5.26671290e-01 -9.87693906e-01 -3.77656013e-01 1.07723749e+00
7.59392753e-02 -4.94071513e-01 -4.21513349e-01 -1.31731719e-01
-2.29486436e-01 1.34535301e+00 2.38069519e-01 -1.52215064e-01
4.65977073e-01 8.71500373e-01 1.22142233e-01 1.33862928e-01
-1.22159922e+00 -2.14316230e-02 -4.77871060e-01 3.23425591e-01
4.37177628e-01 1.23109080e-01 -1.73151761e-01 7.17524529e-01
-2.26764381e-01 -4.11729991e-01 6.80370927e-01 1.57192338e+00
-8.51519346e-01 -1.64893079e+00 -1.97198898e-01 4.55701709e-01
-7.01744735e-01 -5.19695282e-01 -1.57770142e-02 6.50408506e-01
-1.98477358e-01 6.08446658e-01 3.77231717e-01 -1.66971788e-01
4.22532886e-01 8.99100065e-01 3.26803327e-01 -1.20680225e+00
-8.50553095e-01 -2.18481824e-01 2.47499987e-01 -6.42437220e-01
4.15508300e-01 -7.34664857e-01 -1.69985592e+00 -1.20291792e-01
-4.34388965e-01 2.30926856e-01 9.48901832e-01 9.78060365e-01
1.38327107e-01 5.04073203e-01 3.80091369e-01 -5.80886543e-01
-1.01050878e+00 -6.60002649e-01 -4.81906772e-01 -4.29736786e-02
4.53505546e-01 -1.09973408e-01 -3.95426571e-01 -2.22546384e-01] | [10.379966735839844, 9.53250503540039] |
b5d9e025-f4e2-4c00-a389-57ea56ce87d3 | cloudsegnet-a-deep-network-for-nychthemeron | 1904.07979 | null | http://arxiv.org/abs/1904.07979v1 | http://arxiv.org/pdf/1904.07979v1.pdf | CloudSegNet: A Deep Network for Nychthemeron Cloud Image Segmentation | We analyze clouds in the earth's atmosphere using ground-based sky cameras.
An accurate segmentation of clouds in the captured sky/cloud image is
difficult, owing to the fuzzy boundaries of clouds. Several techniques have
been proposed that use color as the discriminatory feature for cloud detection.
In the existing literature, however, analysis of daytime and nighttime images
is considered separately, mainly because of differences in image
characteristics and applications. In this paper, we propose a light-weight
deep-learning architecture called CloudSegNet. It is the first that integrates
daytime and nighttime (also known as nychthemeron) image segmentation in a
single framework, and achieves state-of-the-art results on public databases. | ['Yee Hui Lee', 'Soumyabrata Dev', 'Stefan Winkler', 'Atul Nautiyal'] | 2019-04-16 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [-3.31476569e-01 -1.07844222e+00 1.60103947e-01 -5.02461195e-01
-3.52323622e-01 -9.04571295e-01 4.27450359e-01 -3.26676428e-01
-4.12059754e-01 5.54519296e-01 -6.23174548e-01 -3.60532075e-01
2.63696015e-01 -7.62214899e-01 -4.09954458e-01 -1.10582924e+00
1.79516599e-01 3.34394693e-01 3.01876098e-01 3.93445455e-02
1.18967466e-01 6.11226261e-01 -1.78586829e+00 3.85938704e-01
1.27394891e+00 1.08359981e+00 4.74555641e-01 8.32702935e-01
-7.65452445e-01 5.27995050e-01 -4.77182537e-01 -2.16210693e-01
5.24409115e-01 -4.41428125e-01 -4.88674730e-01 3.63796085e-01
8.32860172e-01 -2.88526088e-01 -1.50717720e-02 1.33125103e+00
2.25203067e-01 -1.64691642e-01 3.69148016e-01 -1.34696329e+00
-5.87827265e-01 2.53371540e-02 -5.48103750e-01 4.80329514e-01
-2.94383377e-01 2.74642054e-02 6.92609251e-01 -9.25563097e-01
7.03684166e-02 6.22387588e-01 4.82717276e-01 2.92129308e-01
-4.33011651e-01 -9.19376135e-01 2.27193326e-01 4.16920036e-01
-1.46160853e+00 -1.97922606e-02 6.03072405e-01 -6.96763873e-01
7.22819149e-01 4.68694657e-01 1.22958755e+00 3.32487643e-01
1.64526641e-01 8.10238361e-01 1.69584429e+00 -4.73042399e-01
2.99200267e-01 8.20741951e-02 1.35269463e-01 2.44883984e-01
6.78638518e-01 -5.27469255e-02 -1.19346172e-01 2.38776788e-01
5.09277582e-01 3.78030777e-01 -3.20467114e-01 -1.46629393e-01
-4.50601816e-01 7.87285984e-01 5.88299930e-01 5.12979031e-01
-1.33251131e-01 5.08536436e-02 -1.26407575e-02 1.72808804e-02
8.28459322e-01 2.71030456e-01 -3.89267474e-01 6.23030700e-02
-1.54060137e+00 3.01801443e-01 5.09727418e-01 8.32265913e-01
7.70698607e-01 3.26991469e-01 3.33570272e-01 4.47990000e-01
3.07531804e-01 6.98963106e-01 3.02057892e-01 -6.67792559e-01
4.85929437e-02 5.11939585e-01 2.90995836e-01 -6.26385629e-01
-3.70396525e-01 -5.20843744e-01 -6.43311381e-01 4.97329831e-01
1.99886456e-01 -1.49061263e-01 -1.41282558e+00 7.16754198e-01
1.98125497e-01 7.09191918e-01 2.12750733e-02 1.42931557e+00
1.28613412e+00 5.22597790e-01 -3.12432736e-01 -3.99049111e-02
1.11345589e+00 -1.26999187e+00 -7.48920083e-01 -5.13611138e-01
2.35001668e-01 -7.67852485e-01 9.14801419e-01 6.48745477e-01
-7.21760213e-01 -3.39731187e-01 -8.80358875e-01 8.92696008e-02
-1.14779317e+00 1.63561746e-01 1.04507399e+00 8.58541369e-01
-1.06178343e+00 1.86058298e-01 -6.23708189e-01 -1.85725316e-01
2.23120347e-01 -1.30282804e-01 1.14900507e-01 -8.47718958e-03
-1.17222798e+00 7.65342057e-01 2.66007423e-01 5.12400210e-01
-8.52590442e-01 -3.74880672e-01 -3.10598254e-01 -5.16867153e-02
3.84127319e-01 -3.68895233e-01 1.15070856e+00 -1.35693169e+00
-1.29448605e+00 1.23098266e+00 -7.18126632e-03 -3.37194890e-01
4.45889920e-01 -2.76305825e-01 -6.52246177e-01 2.11742654e-01
-2.84670472e-01 2.79309094e-01 8.03562582e-01 -1.64581227e+00
-1.04049540e+00 -3.82181555e-01 3.20259631e-01 1.77155748e-01
7.37605244e-02 2.12359354e-01 -1.01095819e+00 -2.85905153e-01
7.35999271e-02 -1.28793478e+00 -2.05246150e-01 -7.53244981e-02
-3.20984572e-01 -7.18653277e-02 1.03315830e+00 -6.08640134e-01
9.78012800e-01 -1.85478437e+00 -3.34214360e-01 -3.90835926e-02
3.45407248e-01 6.10400617e-01 2.44931027e-01 5.90480417e-02
5.11805639e-02 1.83345780e-01 -4.82182413e-01 -3.74389797e-01
-2.52710849e-01 3.80353153e-01 -3.07958513e-01 8.24674606e-01
-2.39561126e-01 6.92063510e-01 -7.28210866e-01 -5.73438823e-01
4.65049386e-01 4.22013640e-01 -5.88945858e-02 3.42031389e-01
-1.75490201e-01 4.06825244e-01 -2.47553751e-01 9.51790452e-01
1.43088508e+00 -1.67667866e-01 -1.87461019e-01 1.52774677e-01
-8.23538303e-01 -2.56664995e-02 -9.47598398e-01 1.43263125e+00
-4.18091357e-01 1.08113742e+00 2.90322810e-01 -5.27813435e-01
4.91728365e-01 2.34341905e-01 1.66173339e-01 -6.19061410e-01
1.70275465e-01 4.05503780e-01 -7.36899525e-02 -6.28989816e-01
6.98884249e-01 -4.39716391e-02 4.15162355e-01 6.43514991e-02
-3.48095119e-01 -8.53251278e-01 2.76963681e-01 -1.17236897e-01
1.41257495e-01 4.59071510e-02 -1.58428282e-01 -1.31798670e-01
3.72873127e-01 2.33760327e-01 5.44340372e-01 6.89661801e-01
-3.66727650e-01 1.23087311e+00 -1.91649720e-01 -5.98904252e-01
-7.65241027e-01 -9.00554359e-01 -3.30799431e-01 8.26958358e-01
4.04744357e-01 1.06494628e-01 -9.39257383e-01 -3.68623346e-01
6.46143928e-02 5.24230540e-01 -3.28879684e-01 4.15508568e-01
-2.59918898e-01 -9.69740272e-01 2.75724441e-01 5.85093558e-01
7.81602263e-01 -8.81321013e-01 -8.91051710e-01 -1.50904477e-01
-1.78704083e-01 -1.45650470e+00 -9.27655026e-02 1.59159660e-01
-7.16908216e-01 -1.19327581e+00 -6.80473149e-01 -4.64468390e-01
4.67946380e-01 1.06628180e+00 1.46521270e+00 3.99702311e-01
-5.32599866e-01 8.54515955e-02 -4.66701478e-01 -7.82624424e-01
2.07276523e-01 -1.49099097e-01 -2.25877851e-01 -1.96326952e-02
6.15896285e-01 -1.00251706e-02 -7.29244173e-01 -1.17666759e-01
-1.06634033e+00 3.48796919e-02 2.99282342e-01 2.57669687e-01
6.66055083e-01 3.03605139e-01 -2.63225108e-01 -7.95022666e-01
2.35876024e-01 -4.14134651e-01 -1.34252393e+00 4.06440496e-01
-6.77913249e-01 -7.04162359e-01 5.07874131e-01 3.64297897e-01
-1.03914392e+00 1.35244071e-01 1.22400716e-01 -8.47103953e-01
-4.24642205e-01 4.73669916e-01 1.74205616e-01 -6.00235045e-01
1.89889967e-01 4.46985751e-01 -6.35376930e-01 -4.73512262e-01
3.73392135e-01 1.01096952e+00 5.60265839e-01 -2.74375647e-01
9.61424291e-01 1.03466856e+00 -4.33605164e-01 -9.68853474e-01
-1.15313637e+00 -1.17124355e+00 -8.36264431e-01 -4.98633027e-01
1.31264317e+00 -1.29956222e+00 -4.72135514e-01 9.04420197e-01
-1.03323567e+00 -1.55103222e-01 2.39171728e-01 5.72161734e-01
-6.93071112e-02 3.50275964e-01 -4.81139719e-01 -1.31016886e+00
-5.45057952e-01 -9.04105484e-01 9.83847141e-01 7.76406884e-01
8.00313115e-01 -8.76494706e-01 6.89597428e-02 5.46593904e-01
4.38482434e-01 3.05818379e-01 3.02925438e-01 -1.88752457e-01
-9.65573311e-01 -2.41832107e-01 -5.99713326e-01 7.32055247e-01
1.82628632e-01 7.31037140e-01 -1.52772760e+00 -2.50477284e-01
5.26224300e-02 -1.30500600e-01 1.25419128e+00 6.25658154e-01
1.48151016e+00 3.17469567e-01 -8.50434303e-02 1.13993013e+00
1.80849278e+00 3.52504253e-01 7.44440734e-01 6.86936915e-01
9.51284885e-01 3.37425321e-01 6.92776561e-01 4.39858288e-01
2.80089289e-01 2.00641170e-01 7.91561663e-01 -6.77575052e-01
2.04533413e-01 6.38300717e-01 1.49427610e-03 8.17268193e-01
-4.32508230e-01 -4.31644261e-01 -1.17390394e+00 7.72954345e-01
-1.51757133e+00 -8.72792661e-01 -6.03782535e-01 2.00193977e+00
6.53645158e-01 2.86224894e-02 -1.37554348e-01 -2.40914568e-01
5.71764529e-01 2.56082863e-01 -4.17849630e-01 -4.49994743e-01
-4.25417155e-01 1.37240112e-01 9.99747992e-01 1.09706514e-01
-1.59148514e+00 1.03731465e+00 6.69742918e+00 5.28888524e-01
-1.66137362e+00 2.08900735e-01 4.86212164e-01 -8.15025717e-02
-1.01128161e-01 -1.18743023e-02 -6.98246717e-01 7.32151330e-01
7.09923685e-01 7.41244778e-02 6.11085534e-01 8.50357234e-01
4.03122753e-01 -2.32060298e-01 -3.88791472e-01 9.81038690e-01
1.77522823e-01 -1.28262544e+00 -2.79660404e-01 -9.94617790e-02
1.03402567e+00 6.88969970e-01 1.86679468e-01 -1.36894849e-03
1.35220811e-01 -9.01208401e-01 6.55378580e-01 6.11233890e-01
6.61005497e-01 -6.58045113e-01 7.86240816e-01 1.36464685e-01
-1.24411535e+00 -3.75028923e-02 -4.86228615e-01 -9.41746011e-02
-2.90410161e-01 8.70845139e-01 -5.87753415e-01 5.86127341e-01
1.17982352e+00 4.17600423e-01 -6.13244295e-01 1.56789494e+00
-2.14044690e-01 7.83516824e-01 -2.22009093e-01 2.71640569e-01
6.80246592e-01 -9.09339666e-01 5.79853058e-02 1.51754940e+00
3.46625656e-01 7.68041238e-02 2.93932825e-01 9.78803098e-01
-2.88816709e-02 -1.56667575e-01 -3.70189726e-01 -2.15754017e-01
3.02523643e-01 1.59035766e+00 -9.79434073e-01 -5.60622811e-01
-8.25132191e-01 8.27678502e-01 -3.62561285e-01 4.67908919e-01
-1.03095591e+00 -3.09828758e-01 8.81377935e-01 -1.16247609e-01
3.67074639e-01 -4.56199914e-01 -3.27391714e-01 -1.35638547e+00
-1.88258328e-02 -6.71820164e-01 2.04063311e-01 -1.25487614e+00
-1.18916547e+00 7.07190335e-01 -1.25731468e-01 -1.50746810e+00
2.50794232e-01 -7.23656356e-01 -8.65139723e-01 1.00045967e+00
-2.36333776e+00 -1.27929211e+00 -1.04259217e+00 5.98103821e-01
6.50021434e-01 4.79914881e-02 6.54224873e-01 4.28514987e-01
-6.50135517e-01 9.02764127e-02 7.97192216e-01 1.70548901e-01
4.89255190e-01 -1.60531843e+00 4.67891365e-01 1.19390428e+00
1.48196951e-01 3.16261858e-01 7.48597980e-01 -3.00264508e-01
-1.14377880e+00 -1.37587404e+00 5.22180676e-01 -1.60002366e-01
4.40095037e-01 -4.35543507e-01 -9.28221524e-01 3.40587586e-01
4.84205693e-01 4.01554227e-01 6.72875226e-01 -1.80713475e-01
-2.21791044e-01 -2.23584980e-01 -9.87431169e-01 3.35998565e-01
3.74320716e-01 -7.89866328e-01 -2.54317969e-01 6.67211056e-01
5.48084438e-01 -6.58776760e-01 -3.99645627e-01 8.21088850e-02
5.23842335e-01 -1.26614845e+00 7.89613783e-01 -2.45322913e-01
2.20971838e-01 -6.36613071e-01 -2.71715559e-02 -1.17025971e+00
-2.65439630e-01 4.20525335e-02 2.97203600e-01 9.23187494e-01
1.41914949e-01 -5.40324330e-01 1.00232863e+00 3.77892494e-01
-4.41902220e-01 -3.62910777e-01 -7.75672257e-01 -8.81150663e-01
1.62589729e-01 -3.48370016e-01 8.18091512e-01 1.33622801e+00
-9.58342433e-01 -5.69370747e-01 -3.08512121e-01 5.48427284e-01
5.57673454e-01 1.02483201e+00 6.30170703e-01 -1.43255484e+00
6.79096580e-02 -2.41920054e-01 -3.75727899e-02 -4.34319228e-01
7.95088559e-02 -5.33077240e-01 3.54832500e-01 -1.81894064e+00
3.36247832e-01 -3.30239743e-01 -5.95350623e-01 8.54632258e-02
-1.91655219e-01 5.42969763e-01 4.72775161e-01 3.15222025e-01
-7.10988224e-01 2.13812947e-01 1.12622929e+00 -1.37944773e-01
8.11414644e-02 8.92060176e-02 -1.60774365e-01 9.93430734e-01
9.89908457e-01 -3.48896831e-01 -3.15176994e-02 -7.55192518e-01
1.08378440e-01 -3.80929142e-01 3.52575004e-01 -1.22340584e+00
2.26755127e-01 -5.92169940e-01 4.35835928e-01 -9.84093606e-01
4.92998004e-01 -8.35506260e-01 -2.85841972e-02 2.47691423e-01
2.76658416e-01 -5.49587570e-02 2.72310197e-01 4.33224678e-01
-5.05689681e-01 -2.66745090e-01 8.93653274e-01 -6.41668618e-01
-1.16719615e+00 3.62196922e-01 -4.49066907e-01 -5.61344549e-02
8.11572552e-01 -4.67050701e-01 -5.50724506e-01 -2.32611716e-01
-2.60140032e-01 1.91032588e-01 7.49617636e-01 4.03374642e-01
5.57737410e-01 -7.66236484e-01 -5.63760459e-01 7.52322376e-02
2.07523242e-01 8.32556561e-02 3.07463706e-01 7.49754786e-01
-1.10845900e+00 6.06511950e-01 -6.56016096e-02 -8.03845823e-01
-1.31022406e+00 3.66299450e-01 7.35999703e-01 1.70710593e-01
-3.22794318e-01 9.73994374e-01 3.13636184e-01 -3.09232205e-01
-1.45683456e-02 -6.37069821e-01 -3.21741402e-01 1.19792886e-01
2.51832306e-01 2.15067998e-01 3.48137438e-01 -5.32540560e-01
-5.54673731e-01 7.18195856e-01 2.40210935e-01 3.64272714e-01
1.23812532e+00 -2.72636294e-01 -3.72487158e-01 7.18614221e-01
9.03449535e-01 -1.39513509e-02 -1.19671631e+00 -2.31947694e-02
-2.67521799e-01 -9.78674531e-01 5.78384638e-01 -6.98386788e-01
-1.72996116e+00 1.08880055e+00 1.14484131e+00 3.88359636e-01
1.38577044e+00 -4.80846792e-01 9.38067019e-01 1.62946194e-01
1.98908001e-01 -1.46505356e+00 -5.02232730e-01 5.20719051e-01
3.32487792e-01 -1.70195866e+00 1.40016034e-01 -3.82369518e-01
-6.10646725e-01 1.17897356e+00 7.75871933e-01 -8.81553963e-02
6.38620853e-01 2.92185247e-01 6.01271510e-01 -3.47921640e-01
-4.50447559e-01 -8.08009386e-01 2.53873110e-01 3.48398328e-01
5.58244824e-01 4.06938851e-01 -2.13621035e-01 3.58716726e-01
-7.48642953e-03 1.48807034e-01 8.27085733e-01 1.03756106e+00
-7.72931993e-01 -4.99253958e-01 -6.97827280e-01 4.11863804e-01
-6.44834816e-01 -3.95681471e-01 -5.48770666e-01 6.42352283e-01
6.10544384e-01 1.27310300e+00 4.89786237e-01 -1.98911414e-01
-1.53317794e-01 -1.39408052e-01 1.58181071e-01 -4.80990857e-01
-6.41773939e-01 2.81351268e-01 -1.97699368e-01 -4.10560489e-01
-7.96994627e-01 -3.04751337e-01 -1.26785493e+00 -2.76593208e-01
-7.05278575e-01 2.94149637e-01 1.14044249e+00 8.98057938e-01
9.31746885e-02 4.41333324e-01 7.45272875e-01 -9.33209658e-01
1.42369658e-01 -6.11103952e-01 -1.20796072e+00 1.31277919e-01
5.78322291e-01 -5.48606277e-01 -6.37976050e-01 2.84856893e-02] | [9.749017715454102, -1.7143698930740356] |
394343e5-0d56-4fe4-b331-da9bffc59552 | to-be-or-not-to-be-a-translation-reception | 2307.02358 | null | https://arxiv.org/abs/2307.02358v1 | https://arxiv.org/pdf/2307.02358v1.pdf | To be or not to be: a translation reception study of a literary text translated into Dutch and Catalan using machine translation | This article presents the results of a study involving the reception of a fictional story by Kurt Vonnegut translated from English into Catalan and Dutch in three conditions: machine-translated (MT), post-edited (PE) and translated from scratch (HT). 223 participants were recruited who rated the reading conditions using three scales: Narrative Engagement, Enjoyment and Translation Reception. The results show that HT presented a higher engagement, enjoyment and translation reception in Catalan if compared to PE and MT. However, the Dutch readers show higher scores in PE than in both HT and MT, and the highest engagement and enjoyments scores are reported when reading the original English version. We hypothesize that when reading a fictional story in translation, not only the condition and the quality of the translations is key to understand its reception, but also the participants reading patterns, reading language, and, perhaps language status in their own societies. | ['Antonio Toral', 'Ana Guerberof Arenas'] | 2023-07-05 | null | null | null | null | ['machine-translation'] | ['natural-language-processing'] | [-1.91368591e-02 3.68149370e-01 -2.86584526e-01 2.29835227e-01
-7.83501744e-01 -7.95475841e-01 9.82414544e-01 2.77228892e-01
-5.64272523e-01 6.64477825e-01 1.26561427e+00 -1.39530182e-01
4.12860326e-02 -4.79176849e-01 -6.54432416e-01 -2.67414123e-01
6.46660149e-01 1.53777227e-01 -4.39857453e-01 -5.10115564e-01
3.78703058e-01 3.21074799e-02 -1.23424840e+00 2.15199247e-01
9.65825438e-01 1.32470489e-01 5.35750926e-01 7.04474032e-01
-9.55578610e-02 8.62363756e-01 -7.43884504e-01 -8.11637998e-01
-3.11763883e-01 -9.88634229e-01 -8.51198256e-01 -9.69157368e-02
-2.15165247e-03 -3.20076525e-01 -1.21538326e-01 5.19609690e-01
5.24537265e-01 1.83478251e-01 4.18748170e-01 -4.31142807e-01
-9.14123535e-01 9.44773853e-01 1.05568536e-01 -1.59040183e-01
1.07571042e+00 2.07283571e-01 6.48265421e-01 -8.26001227e-01
1.12932587e+00 1.14792275e+00 3.89437348e-01 1.42385244e-01
-1.07447994e+00 -2.18492150e-01 -2.95532137e-01 2.69097120e-01
-7.99978733e-01 -7.42148817e-01 4.53132182e-01 -5.53526223e-01
7.36851811e-01 4.15320009e-01 1.40543914e+00 1.59561300e+00
4.94559705e-01 2.30696127e-01 1.34918058e+00 -4.76475924e-01
2.40146741e-01 5.49949467e-01 -2.23133802e-01 7.62644559e-02
2.68176883e-01 -2.68032923e-02 -9.87561166e-01 2.18260288e-01
5.51471591e-01 -5.80285788e-01 -3.43783647e-01 4.13135588e-01
-1.76700890e+00 8.41653109e-01 2.30600592e-03 6.63650274e-01
-6.64137483e-01 -1.01957127e-01 6.30959988e-01 6.57996833e-01
4.23803985e-01 7.63109505e-01 -1.06376231e-01 -1.11739957e+00
-6.68470800e-01 1.46030396e-01 9.43210840e-01 3.76460999e-01
-1.42336130e-01 -2.45315790e-01 -4.81490672e-01 7.68884122e-01
-1.62826076e-01 6.89697444e-01 4.79030579e-01 -5.42263567e-01
4.12275821e-01 1.76126406e-01 2.18234837e-01 -1.03084326e+00
-2.76148945e-01 -6.58823252e-01 -1.06068008e-01 2.71825325e-02
5.22020400e-01 -4.13224488e-01 5.47556719e-03 1.65893328e+00
-1.26054034e-01 -7.65519679e-01 2.36943826e-01 1.00974631e+00
9.22686398e-01 8.06372404e-01 5.32527342e-02 -6.72978938e-01
1.32537353e+00 -8.40836167e-01 -1.11649776e+00 -4.35521632e-01
6.57886922e-01 -1.30292380e+00 1.68023717e+00 2.21380502e-01
-1.65050995e+00 -7.11793125e-01 -1.22638369e+00 -2.18767300e-01
-1.45295531e-01 3.54710102e-01 -5.04446542e-03 7.20984161e-01
-7.78041601e-01 5.13374925e-01 -4.24938023e-01 -8.46919119e-01
1.02142453e-01 -3.94813985e-01 -3.63099545e-01 3.64456959e-02
-9.02360857e-01 1.42449379e+00 1.49078220e-01 -7.35569969e-02
-2.86249816e-01 -1.96830377e-01 -5.61106026e-01 -1.08330444e-01
-5.86001053e-02 -3.41377348e-01 1.02719808e+00 -1.36781037e+00
-1.84396088e+00 9.56161439e-01 -2.36760274e-01 8.92845169e-02
8.69359255e-01 -5.77826619e-01 -5.63400686e-01 2.77709275e-01
3.50358307e-01 2.40420640e-01 3.41916382e-01 -7.95254707e-01
-2.85188764e-01 3.89328711e-02 2.49784850e-02 6.97295964e-01
-8.13878849e-02 3.49618584e-01 2.00817123e-01 -5.89307308e-01
-2.92774159e-02 -5.13407886e-01 4.67193186e-01 -1.09418906e-01
-1.12707131e-01 -2.45221127e-02 2.20835552e-01 -1.35213041e+00
9.36339259e-01 -2.42798686e+00 2.43422523e-01 -7.42030963e-02
1.31053224e-01 -2.77602881e-01 -3.28329243e-02 1.03970885e+00
4.77768838e-01 3.39813858e-01 4.41335171e-01 -2.48952825e-02
8.99720863e-02 1.04944870e-01 4.80615675e-01 5.53251266e-01
-1.76928863e-02 1.01896882e+00 -9.55576599e-01 -2.62117624e-01
-1.76472500e-01 4.92351562e-01 -1.53061256e-01 -2.30634451e-01
7.13844178e-03 5.72414935e-01 -6.30155057e-02 1.98734879e-01
3.29277903e-01 1.73674226e-01 6.15553521e-02 3.34354043e-01
-7.68662453e-01 6.26290619e-01 -6.14417851e-01 1.49714673e+00
-5.68655252e-01 1.35105443e+00 -7.86281303e-02 -3.23143482e-01
9.86685038e-01 5.12212157e-01 -1.96938217e-01 -1.40856063e+00
3.09316784e-01 2.79921949e-01 1.03902176e-01 -9.42753732e-01
9.89568293e-01 -3.29322785e-01 -2.66248167e-01 6.49451911e-01
-7.23026454e-01 -1.55393213e-01 1.55679122e-01 1.95634723e-01
7.68299043e-01 1.98364273e-01 2.10592464e-01 -3.26194078e-01
-8.62673149e-02 3.12194586e-01 1.94894806e-01 5.74613750e-01
-1.68566480e-01 4.23414677e-01 9.54223812e-01 3.36111277e-01
-1.09755778e+00 -1.10670638e+00 -2.62116361e-02 9.62914705e-01
2.41933405e-01 -3.91326696e-01 -8.77560198e-01 1.84433132e-01
-7.82655776e-01 1.42208993e+00 -4.20330733e-01 -3.55295271e-01
-4.24660504e-01 2.08393801e-02 3.13822836e-01 1.50999665e-01
9.22976136e-01 -1.19590104e+00 -1.07544291e+00 1.34894162e-01
-1.11580646e+00 -9.00721431e-01 -5.82457125e-01 -4.26841110e-01
-4.30024266e-01 -5.43936431e-01 -4.62783694e-01 -9.20532525e-01
5.16793616e-02 2.40575001e-02 8.59460652e-01 -6.29738197e-02
2.42037728e-01 3.87428403e-01 -9.17294741e-01 -3.20690662e-01
-8.49696696e-01 6.98354989e-02 -3.92505407e-01 -3.90381664e-01
-6.97633252e-02 -1.14429176e-01 -1.71917230e-01 7.29437545e-02
-5.30312479e-01 5.42434633e-01 7.94760823e-01 5.90357006e-01
1.51478454e-01 -1.32191792e-01 5.70923507e-01 -3.82400751e-01
8.31219494e-01 -4.51616019e-01 4.18854773e-01 6.28132820e-02
-2.59272307e-01 -5.88659525e-01 3.72285366e-01 -6.92042708e-01
-1.28821647e+00 -6.37559354e-01 -5.02606407e-02 7.34742582e-01
1.44908205e-01 8.64516556e-01 -1.41063854e-01 4.66604620e-01
7.68265665e-01 -9.84460022e-03 2.28737757e-01 2.52333079e-02
1.52001902e-01 5.07985830e-01 4.65354949e-01 -3.58086437e-01
6.27423108e-01 -2.70344839e-02 -3.78818929e-01 -1.06403387e+00
-1.23341553e-01 2.43589967e-01 -5.78099191e-01 -7.98912406e-01
1.14228082e+00 -8.89714539e-01 -6.09990895e-01 3.54619563e-01
-7.98127830e-01 -6.16273463e-01 -4.77533072e-01 1.00611222e+00
-6.74252927e-01 -2.64576912e-01 -4.71354842e-01 -6.33789778e-01
6.13183435e-03 -9.30941343e-01 5.54033399e-01 2.43105620e-01
-9.16444838e-01 -9.93301153e-01 -5.89563288e-02 5.80516279e-01
4.50785905e-01 7.82161295e-01 1.09577870e+00 -3.75814050e-01
-1.72279149e-01 -1.44997761e-01 -1.34929389e-01 -1.97790056e-01
1.74579269e-03 -1.86676487e-01 -3.38585436e-01 -6.16951175e-02
2.03119785e-01 -4.91278410e-01 1.14356488e-01 3.00087154e-01
-3.72384757e-01 -5.14801979e-01 1.53718472e-01 -6.24045059e-02
1.43804312e+00 1.51716739e-01 9.19254661e-01 7.24616051e-01
1.14398733e-01 6.81816697e-01 8.76473308e-01 2.46402591e-01
2.66165495e-01 5.96089125e-01 -3.14081728e-01 -5.00662066e-02
-1.35090202e-01 -5.13142169e-01 1.23773956e+00 9.63676274e-01
-1.46988228e-01 -2.27391243e-01 -5.70677638e-01 5.13592958e-01
-1.32569075e+00 -1.21542859e+00 -6.23416901e-01 2.06180906e+00
4.96942282e-01 1.98104143e-01 3.79382849e-01 2.56676972e-01
5.71104109e-01 1.07830130e-01 -1.05457857e-01 -1.02059090e+00
-5.78053117e-01 1.57522336e-01 6.59177154e-02 5.58066070e-01
-7.47237578e-02 6.99428439e-01 6.60654259e+00 4.55905885e-01
-9.62094247e-01 2.62223303e-01 6.40798390e-01 -5.92169464e-01
-3.75832468e-01 3.30590196e-02 3.41268666e-02 3.88084259e-03
1.15333438e+00 -5.61089575e-01 5.16316354e-01 8.35049301e-02
9.91406143e-01 -7.39626169e-01 -8.83965492e-01 5.60664415e-01
2.45185003e-01 -1.07341659e+00 -4.15924639e-01 5.59056178e-02
3.73060286e-01 -3.89969617e-01 7.63122737e-02 2.29896918e-01
-5.06908059e-01 -9.98929024e-01 1.64172924e+00 5.02893865e-01
9.77817655e-01 -7.26080894e-01 8.42793345e-01 4.09428060e-01
-4.94359076e-01 2.31269330e-01 -1.15417972e-01 -8.13519418e-01
3.57574880e-01 1.89994216e-01 -5.05127430e-01 3.76382500e-01
3.51084590e-01 5.97120941e-01 -4.96125191e-01 3.87696445e-01
-2.02436820e-01 1.00370991e+00 1.35278821e-01 -5.24897516e-01
-1.99962687e-02 -4.06730533e-01 8.95814836e-01 9.91179764e-01
5.22916377e-01 1.18677668e-01 -2.49029517e-01 7.38236547e-01
1.30553380e-01 7.35134542e-01 -2.03757733e-01 -6.16505325e-01
4.04278755e-01 7.58576751e-01 -1.04947174e+00 -1.09704666e-01
-4.09094185e-01 1.25615036e+00 3.19483615e-02 2.55501032e-01
-7.47991383e-01 -2.79647678e-01 1.93264350e-01 5.61280370e-01
-1.48962691e-01 -2.57979661e-01 -8.83782506e-01 -6.45903170e-01
2.35526919e-01 -7.69033968e-01 -3.29082757e-01 -8.82447779e-01
-6.65444314e-01 2.15394616e-01 1.13547452e-01 -8.31973732e-01
2.60436475e-01 -1.45614505e-01 -6.20155931e-01 9.47655797e-01
-5.65450609e-01 -8.37925076e-01 -4.02089238e-01 -4.04890180e-02
6.49177253e-01 3.22158366e-01 5.76346815e-01 -5.14723286e-02
-4.36915606e-01 4.40811306e-01 2.95312941e-01 -3.06115061e-01
7.33790576e-01 -6.79164112e-01 8.83050635e-02 5.86536407e-01
-4.19345379e-01 2.19501480e-01 8.50062251e-01 -7.03304052e-01
-1.20891857e+00 -4.34790254e-01 1.42518163e+00 -1.13277316e-01
3.46989840e-01 -1.84891179e-01 -3.13034683e-01 4.16037112e-01
8.73428404e-01 -1.54347599e+00 7.07732856e-01 3.67055610e-02
3.52078289e-01 5.24314284e-01 -1.26892257e+00 1.17521799e+00
1.31357002e+00 -4.57723945e-01 -6.32543981e-01 2.77714401e-01
6.69141531e-01 -1.68700442e-01 -1.09549141e+00 -5.96746266e-01
9.13181961e-01 -1.05192792e+00 5.38651764e-01 1.03294589e-01
1.12015212e+00 2.85012335e-01 -8.12440459e-03 -1.40325832e+00
-5.17740369e-01 -6.63632214e-01 8.61817300e-01 1.34232008e+00
6.57988608e-01 -8.16798389e-01 1.63576365e-01 4.97338206e-01
-3.74784976e-01 -3.09755296e-01 -8.67943108e-01 -3.24400395e-01
4.13140267e-01 -1.27771005e-01 3.30059081e-01 8.38924825e-01
4.64203745e-01 5.08642197e-01 -1.74205348e-01 -5.37961066e-01
-1.72448084e-01 -4.41834360e-01 6.08528316e-01 -1.09724128e+00
-2.90214032e-01 -6.38984382e-01 -2.14247722e-02 -4.23301131e-01
-1.11555479e-01 -8.52967799e-01 -1.31708801e-01 -2.23538685e+00
3.34709197e-01 3.26405972e-01 8.23054433e-01 -1.27864063e-01
-1.48905128e-01 2.31778305e-02 5.23112595e-01 1.58049494e-01
-3.54459025e-02 3.21957171e-01 1.77772820e+00 3.59543175e-01
-7.55382001e-01 -3.21980387e-01 -9.32287693e-01 9.59276706e-02
1.00577009e+00 -1.30032614e-01 -3.24172258e-01 -2.25683466e-01
6.44600689e-01 4.63373363e-01 4.03604776e-01 -7.63277769e-01
-1.30777001e-01 -3.48492086e-01 5.65764129e-01 -4.28985387e-01
2.61541575e-01 -3.67632687e-01 7.50485837e-01 6.54772520e-01
-5.92504263e-01 4.07447398e-01 1.23428456e-01 -1.01906568e-01
1.08292185e-01 -1.79837048e-01 6.69882655e-01 1.27939597e-01
1.33002298e-02 -9.00862813e-01 -1.04289889e+00 1.79694682e-01
9.05113339e-01 -8.37004125e-01 -4.01818335e-01 -9.67063069e-01
-8.49171877e-01 -3.01114082e-01 8.53352249e-01 4.22862530e-01
4.33077246e-01 -1.14435220e+00 -1.18256319e+00 -2.00910866e-02
-6.12464175e-02 -5.50477147e-01 1.73883781e-01 1.32339942e+00
-9.29390788e-01 9.24030021e-02 -4.42480654e-01 9.72775072e-02
-9.41543162e-01 3.71581763e-02 -2.11402819e-01 -1.55429468e-01
-9.10412848e-01 1.88756973e-01 -1.90277711e-01 6.41058683e-02
-2.13908836e-01 1.21008232e-01 -1.73682287e-01 3.46339166e-01
3.59077811e-01 9.65321481e-01 -2.23720402e-01 -1.00707579e+00
2.10991576e-01 5.74123204e-01 2.69283593e-01 -7.96672881e-01
7.67279804e-01 -4.76160467e-01 -3.50265801e-02 9.03460324e-01
1.03371918e+00 5.39067328e-01 -4.89025414e-01 2.58750290e-01
-2.60368317e-01 -6.78662717e-01 -1.27768412e-01 -1.21561778e+00
-2.37797707e-01 2.64359266e-01 2.46446371e-01 3.93912895e-03
1.00109923e+00 1.78141177e-01 8.86817992e-01 -1.68176685e-02
7.34857563e-03 -1.55277741e+00 9.69058648e-02 2.70465940e-01
1.33578861e+00 -4.16576296e-01 -3.42041790e-01 -2.65047729e-01
-1.06521416e+00 9.21963632e-01 3.52408402e-02 2.73342520e-01
2.33168244e-01 -9.94726717e-02 7.32986070e-03 -4.97492626e-02
-5.80252230e-01 5.30454423e-03 9.89752542e-03 7.82603621e-01
7.63618290e-01 3.07149231e-01 -1.31800783e+00 7.35189676e-01
-1.24396932e+00 -1.38934836e-01 9.75031376e-01 5.92862606e-01
-3.99520457e-01 -8.33850861e-01 -6.97255015e-01 3.38824719e-01
-3.07460248e-01 2.25831553e-01 -1.00270176e+00 1.01480842e+00
7.51689672e-02 1.27071953e+00 2.47300401e-01 -3.24815989e-01
6.05045557e-01 3.35572623e-02 4.80382860e-01 -2.24127442e-01
-1.14308023e+00 7.52639845e-02 8.25004518e-01 -1.06897607e-01
-3.06636006e-01 -1.23730946e+00 -1.13322687e+00 -1.10806227e+00
-1.36785030e-01 1.72685802e-01 7.74465561e-01 1.13214493e+00
1.14561468e-01 5.49279511e-01 3.04756254e-01 -4.57073182e-01
4.95275296e-02 -1.26321721e+00 -4.42025244e-01 2.89880514e-01
1.14893042e-01 -1.66648656e-01 -3.36800903e-01 -1.22157380e-01] | [11.65351676940918, 8.797774314880371] |
4e44bb86-11f3-473c-a82b-b0eb5ce3cf19 | emoji-prediction-using-transformer-models | 2307.02054 | null | https://arxiv.org/abs/2307.02054v1 | https://arxiv.org/pdf/2307.02054v1.pdf | Emoji Prediction using Transformer Models | In recent years, the use of emojis in social media has increased dramatically, making them an important element in understanding online communication. However, predicting the meaning of emojis in a given text is a challenging task due to their ambiguous nature. In this study, we propose a transformer-based approach for emoji prediction using BERT, a widely-used pre-trained language model. We fine-tuned BERT on a large corpus of text containing both text and emojis to predict the most appropriate emoji for a given text. Our experimental results demonstrate that our approach outperforms several state-of-the-art models in predicting emojis with an accuracy of over 75 \% This work has potential applications in natural language processing, sentiment analysis, and social media marketing. | ['Mehreen Alam', 'Zeeshan Habib', 'Muhammad Osama Nusrat'] | 2023-07-05 | null | null | null | null | ['marketing', 'sentiment-analysis'] | ['miscellaneous', 'natural-language-processing'] | [-0.27197492 -0.2366489 0.02260845 -0.6162265 -0.06712865 -0.31807554
0.21374594 0.24972008 -0.14340734 0.570944 0.20887032 -0.33056033
0.34799558 -0.8043518 0.03574866 -0.06850619 0.4669145 0.1669895
0.06661758 -0.6556566 0.25767383 0.16546452 -1.3835642 0.5770686
1.029278 1.332853 -0.03771459 0.68091255 -0.9108809 1.2982739
-0.73750526 -1.2765 -0.35273692 -0.64002556 -0.7840361 -0.03773963
-0.06873993 0.14683598 0.03293699 0.8743328 0.26413828 0.17113648
0.6210293 -0.8038132 -0.91353637 0.5048221 -0.32046154 0.15541495
0.43040153 -0.5988201 1.1990924 -0.6001501 0.36854446 1.1744322
0.6171629 0.42608127 -0.80343527 -0.4757122 -0.02656359 0.22373064
-0.91460514 -0.27404794 1.109025 -0.31849188 0.89884853 0.5656026
0.5579739 0.8068 0.31713405 1.023021 1.3849559 -0.6787517
0.09630074 0.6882288 0.3288575 0.9037779 -0.486158 -0.7816006
-0.4637173 -0.43199196 0.14192021 0.06197585 0.37301075 0.48621094
-0.33680302 1.0202036 0.16342841 0.54247904 -0.3883481 -0.16793454
0.67787105 0.5475534 1.2734339 0.72112083 -0.7739157 -0.78976136
-0.28890827 -0.2699737 1.3553056 0.5683768 0.56251216 -0.08885845
0.00673055 1.4728627 0.15893252 0.3337001 0.41992843 -0.46840394
0.14061724 0.88486236 0.10420578 -1.4152749 -0.19293222 -0.14879045
-0.48373586 -0.41387576 0.16004641 -0.42586106 -0.42563146 1.181478
0.38437492 -0.07065115 -0.2548643 0.5282588 1.1821766 0.43926334
0.43759203 0.17973234 1.5114387 -0.9572192 -0.90826786 -0.29974732
0.77606374 -1.1671822 1.1634667 0.20458634 -0.6093409 -0.27162114
-0.46751487 -0.22529298 -0.8128 0.48279968 1.1642915 1.1122116
-0.5374155 0.6764314 -0.29478958 -0.48257586 0.29906583 0.22973073
-0.10301048 0.2247054 -1.3336216 1.0123409 -0.34514242 0.10740164
0.19344327 -0.3056317 -0.9764826 0.16500165 0.1616659 -0.3497207
1.1486236 -1.4237905 -2.0520716 0.9027455 -0.56144196 -0.11023121
-0.05859146 -0.38064378 -0.9984869 -0.10535624 -0.04683623 -0.05061925
0.8598071 -1.0353507 -0.4614229 -0.16229662 0.3221956 -0.06668749
-0.9639939 0.62907225 -0.41995114 -0.5093077 -0.5280067 -0.87791055
-0.16004586 -0.32353166 -0.37462267 -0.37822112 0.7347146 -0.9980455
1.5077614 -2.0951571 -0.12898354 0.10784533 0.07931776 0.32276738
0.08174141 0.56734407 0.36946258 0.2634532 0.23735413 -0.8299972
-0.00924239 0.18519208 -0.49522072 -0.20457189 0.07389082 0.8055996
-0.9130655 -0.58029085 0.35241267 0.5366828 -0.28705883 0.48759103
-0.07695656 0.38216937 -0.6518835 0.47594124 0.19065906 -0.3391819
0.22788192 -0.15583488 0.06417814 0.3719816 -0.46422967 0.9595854
-1.2092448 0.73063964 -0.23592514 -0.57965446 1.2607484 0.27509516
0.3781972 -0.644126 0.8246036 0.11105395 -0.1309133 -0.702352
0.66654015 -0.2698746 -0.2852717 0.3699912 -0.14996497 -0.29386342
0.36721554 0.06454454 1.1544701 -0.10019998 0.21669035 -0.0630797
0.7694543 -0.15440477 0.20536508 0.53120285 -0.10779583 0.2551441
0.5061696 -0.36771503 -0.6179915 -0.2809262 0.01288807 1.3140562
-0.03238755 -0.61006147 -0.65209514 -0.9382618 -0.09694631 0.895408
-0.82987213 0.03702496 -0.25464943 -0.6136708 0.3196014 0.25853494
0.53636897 -1.123008 -0.13287538 0.22222547 -0.30323875 -1.2974494
-0.38409308 -0.04884204 -0.6236782 -0.9067305 -0.4281113 -0.70409733
0.59683114 0.33205876 1.3801757 0.3782111 -0.18285985 0.3957592
-0.78124493 -0.47599098 -0.50165087 0.22088295 -0.4082999 0.45937207
0.8026908 -0.3118325 -0.2110651 0.27624378 -0.7044867 0.02742644
0.05918266 0.72278035 -0.23485452 -0.03150751 0.6097051 -1.3310364
0.9988285 -0.63650674 -0.01944845 0.4071301 -0.28512746 -0.21751875
1.2216684 -0.4342453 -1.39134 -0.32301837 -0.50598663 0.19355914
-0.07709006 0.73804647 0.1345285 -0.3008956 0.32064393 -0.1868381
-0.1387056 -0.70380366 0.27380946 1.0802993 -0.4130716 -0.3669576
0.19659755 0.36365038 -0.19817208 -1.1040453 -1.5575022 -0.6359839
-0.3508883 -0.36598837 0.7443563 -0.47464454 -0.9887736 0.35145974
-1.1252563 -0.111676 0.1711342 0.11066236 0.11503908 0.5934505
-0.9989713 -1.0445367 -0.7097766 -0.8106529 0.8671867 0.36999136
-0.5927868 -1.5302739 -0.03127294 0.81675786 0.7111713 -0.25308043
0.898577 -0.64420485 0.3551428 -0.6150858 -0.37587798 0.49179566
0.51058084 0.03875662 -0.85587525 0.48714897 0.12794152 -0.47675613
0.47705376 -0.11757452 1.3278724 -0.08459814 0.06908444 -0.08433685
1.0298128 0.1352322 0.57150674 0.27684766 0.58206666 0.6776341
0.61467016 0.66139984 0.6473507 0.51791316 0.19276606 -0.03844526
0.08670141 -0.03856333 0.16203076 1.2274289 0.09857337 -0.30307028
-0.42229164 0.42927706 -1.7560089 -0.7248015 -0.2955692 1.6536208
1.0612278 0.05900279 -0.04923572 -0.16881296 0.6363987 0.14521585
-0.15090558 -1.0983045 0.09852345 0.5673908 0.1184554 0.489442
-1.0804642 1.2024134 5.909938 0.8394078 -1.2167474 0.2400069
1.0165644 0.5433717 -0.3132936 -0.114522 -0.70617115 0.64353746
0.886839 0.13537577 0.5175373 1.0297834 0.15650362 -0.29059845
-0.5340623 0.9903284 0.4655546 -0.991769 -0.2944458 -0.34939587
0.57057965 -0.3582739 -0.06646044 0.5805326 0.36189008 -0.7131088
0.31960163 0.41262177 0.26066482 -0.4745837 0.8815599 0.11646044
-1.0205277 0.14837267 -0.26635426 -0.6085692 0.24519734 0.932719
-0.5863633 0.2327707 0.601498 0.7662716 -0.70332974 0.71130127
-0.34716308 0.9536678 -0.21062875 -0.8074036 0.02978061 -0.40163013
-0.05708533 1.3357303 0.09021013 0.18886831 0.07503862 0.35204756
-0.3152222 0.71984774 -0.36156073 -0.4146333 0.10041793 1.6892236
-0.7883304 -0.47665513 -0.5450164 1.4162146 0.6620721 -0.00645013
-0.94967306 -0.60370266 0.62870115 -0.42702726 0.22601058 -0.15468976
-0.64088476 -1.3306336 -0.12625343 -0.55406016 0.19649059 -0.798383
-1.7969291 0.908555 -0.6902172 -1.0014291 0.11779918 -0.85638803
-0.8237716 0.71716326 -1.4099256 -1.1225044 -0.18563884 -0.08804926
0.5819823 0.02817727 1.2323824 0.42675066 -0.61181927 0.4690709
0.2325388 0.41283447 0.68945426 -1.1282607 0.37281615 0.20248455
0.1828047 0.67711675 0.8890775 -0.66371775 -1.3975403 -0.6747682
1.4427158 -0.5411494 1.0793066 -0.33417657 -0.6360519 0.23631455
0.3820553 -0.32906 1.4894844 0.6215311 -0.4133069 -0.05418483
-1.105673 0.82588995 0.664056 -0.70026946 -0.38472763 0.37737322
0.29485735 -0.21452783 -1.1233548 -0.18818061 0.6334151 -0.86489403
0.51779234 -0.8864942 0.8017301 0.28957045 0.16101688 -1.5498213
-0.08244397 -0.38407373 -0.09908333 1.4197446 0.4489943 -0.6316947
0.8315424 1.2952378 0.18346362 -0.91030836 -0.53143835 -0.28194553
-0.40227795 -0.5578816 0.33994254 1.000711 0.4480221 0.80895936
-0.7584316 -0.51098806 0.09706893 0.35433784 0.7138554 -1.1562325
-0.35800686 -0.38536593 -0.19732296 -1.1592379 0.52512556 -0.47210655
-0.19634847 -1.5006237 0.03706531 -0.61590016 -0.00882419 0.4216271
-0.36659023 0.63040876 0.30567425 -0.19344991 -0.7854524 0.8202283
1.1400735 -0.20381218 -0.22416165 0.14033213 -0.6974355 0.92133343
0.9080466 -0.3666085 -0.02178559 -0.25689262 0.71207535 -0.13713439
-0.25527644 -0.45359218 -0.24984024 -0.09714642 0.13912395 -0.01450375
0.87332445 -1.0444453 -0.30885208 0.01731171 -0.43786335 -0.11098854
0.14672048 0.20670456 -0.15922062 -0.631456 0.33813474 -0.09020165
-0.6495051 -0.07895049 -0.59626526 0.07577799 0.57237625 0.18135317
-0.16527353 -0.7225699 -0.60284656 -0.17732164 0.17457344 0.79811144
0.5470405 -0.9390671 -0.16333927 -0.20985813 0.02099996 -1.040471
0.2654672 0.7381178 -0.23659691 -0.02149843 0.13457122 0.11910364
-1.6423001 0.10619158 -0.01084453 -0.29925787 -0.16242337 0.85925025
-0.2764412 -0.5561165 -0.20877947 0.03741648 -0.7293686 -0.09556243
0.6251026 0.32324845 0.09225067 -0.8846962 -0.13454348 0.14981854
-0.05878596 0.1023443 1.2361118 -0.30012912 -0.77892274 0.87837565
1.3045789 0.36647654 -0.35763997 -0.22409856 0.11715408 -0.72682637
-0.00616529 -0.86240554 -0.9741387 0.8726067 0.09037419 0.63158315
0.7826043 0.00609112 1.3396113 0.56723785 0.30176148 -1.2500068
0.23829399 0.967434 0.50578076 -1.4365964 -0.11557043 -0.8328611
-1.1705412 1.30573 0.6266476 -0.02436859 1.1527001 -0.01041715
0.40157574 -0.20677353 -0.55149037 -0.27915835 0.47684065 0.03185759
1.0447444 0.15852845 -0.55102617 0.8936727 -0.08102409 -0.0840645
0.3590623 0.9305098 -0.23106514 -1.354848 -0.04445718 0.9778541
-0.8501649 -0.40983108 -0.7689623 -0.08178559 -0.25645384 1.5564924
-0.29566926 -0.7163657 0.06783404 0.28170693 0.2547374 -0.7332203
-1.1956187 -0.38483953 0.84577054 -0.06691845 -0.41821232 -0.32396984
-1.0605298 -0.5291251 -0.51549286 0.46050292 0.85643333 1.5276548
0.7197816 0.3155774 0.8670161 -0.5338357 -0.1591663 -1.0887768
-0.35493425 0.76725423 -0.4148462 -0.5843301 -0.23081733 -0.11197006] | [11.406001091003418, 6.911930084228516] |
de65e4c5-6785-488e-bf19-382fe629313a | plug-and-play-split-gibbs-sampler-embedding | 2304.11134 | null | https://arxiv.org/abs/2304.11134v1 | https://arxiv.org/pdf/2304.11134v1.pdf | Plug-and-Play split Gibbs sampler: embedding deep generative priors in Bayesian inference | This paper introduces a stochastic plug-and-play (PnP) sampling algorithm that leverages variable splitting to efficiently sample from a posterior distribution. The algorithm based on split Gibbs sampling (SGS) draws inspiration from the alternating direction method of multipliers (ADMM). It divides the challenging task of posterior sampling into two simpler sampling problems. The first problem depends on the likelihood function, while the second is interpreted as a Bayesian denoising problem that can be readily carried out by a deep generative model. Specifically, for an illustrative purpose, the proposed method is implemented in this paper using state-of-the-art diffusion-based generative models. Akin to its deterministic PnP-based counterparts, the proposed method exhibits the great advantage of not requiring an explicit choice of the prior distribution, which is rather encoded into a pre-trained generative model. However, unlike optimization methods (e.g., PnP-ADMM) which generally provide only point estimates, the proposed approach allows conventional Bayesian estimators to be accompanied by confidence intervals at a reasonable additional computational cost. Experiments on commonly studied image processing problems illustrate the efficiency of the proposed sampling strategy. Its performance is compared to recent state-of-the-art optimization and sampling methods. | ['Pierre Chainais', 'Nicolas Dobigeon', 'Florentin Coeurdoux'] | 2023-04-21 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 3.30049127e-01 9.79543775e-02 1.12707149e-02 -1.51391670e-01
-1.16094089e+00 -1.27546966e-01 9.03246701e-01 -5.97891845e-02
-5.09134293e-01 1.05878735e+00 -1.95542544e-01 -1.55055448e-01
-1.88395858e-01 -8.41809571e-01 -6.60014570e-01 -1.30001366e+00
3.50278795e-01 6.60502255e-01 -4.60117124e-02 3.06810439e-01
2.52673388e-01 1.85745046e-01 -1.15215051e+00 -4.42651749e-01
1.05280733e+00 1.11016083e+00 4.99446779e-01 3.80360752e-01
-9.30416286e-02 5.42670965e-01 -3.26535344e-01 -4.10434902e-01
6.17749915e-02 -7.28612542e-01 -4.69509125e-01 5.01474261e-01
-6.17539063e-02 -4.26545143e-01 1.26135170e-01 1.25496161e+00
5.11808038e-01 3.36965293e-01 9.15949941e-01 -9.29482996e-01
-3.36813778e-01 4.28332239e-01 -7.73896933e-01 -6.25355542e-02
3.25884074e-02 1.91633224e-01 8.87702227e-01 -1.08959091e+00
5.72688401e-01 1.15963650e+00 7.21300900e-01 8.14035162e-02
-1.74270189e+00 -2.18005672e-01 1.17236078e-01 -1.89181957e-02
-1.38310802e+00 -3.79045069e-01 9.34804857e-01 -5.94601929e-01
3.08165699e-01 -1.07495403e-02 8.33030105e-01 1.10711086e+00
1.22485809e-01 1.03282857e+00 1.27596605e+00 -4.31550056e-01
8.75234008e-01 6.39725775e-02 -1.20735355e-01 6.05557561e-01
3.53052437e-01 -1.59000248e-01 -6.18092299e-01 -3.68932754e-01
8.75494599e-01 -6.62784949e-02 -4.07150507e-01 -5.85661471e-01
-9.93582129e-01 1.13290918e+00 -1.29016772e-01 -8.12956542e-02
-7.67028570e-01 3.26510221e-01 7.69571438e-02 -1.73673630e-01
8.52543652e-01 -1.22774594e-01 3.14928815e-02 -2.04130292e-01
-1.75092447e+00 2.85326600e-01 8.90120327e-01 8.25097144e-01
8.91358256e-01 3.71132284e-01 -3.15558583e-01 7.17311144e-01
9.33953881e-01 5.23069501e-01 2.47490823e-01 -8.87960494e-01
1.63296580e-01 -9.46824849e-02 4.54010963e-01 -8.15049350e-01
2.07750574e-01 -6.54961348e-01 -9.21233654e-01 3.12947661e-01
5.83965361e-01 -3.06791544e-01 -7.89377749e-01 1.54720664e+00
6.71788156e-01 3.27492267e-01 -2.71416992e-01 7.63594866e-01
2.31017455e-01 6.66581631e-01 -2.20292106e-01 -5.59426188e-01
1.21628213e+00 -7.53790081e-01 -8.66829038e-01 -8.36878568e-02
-2.68316865e-01 -7.72761643e-01 6.21444881e-01 6.81362033e-01
-1.33351922e+00 -3.96653503e-01 -1.07435632e+00 2.21414015e-01
6.80847242e-02 2.07952961e-01 4.60327834e-01 8.42301905e-01
-1.02155375e+00 8.54593098e-01 -1.28578424e+00 -1.91840917e-01
4.94015604e-01 -2.78562792e-02 2.71900475e-01 9.29770619e-02
-8.14020395e-01 6.50343657e-01 9.47565436e-02 4.00568277e-01
-1.24944162e+00 -6.40540719e-01 -6.65272117e-01 -3.96895781e-02
4.49126691e-01 -1.00149477e+00 1.25134814e+00 -7.19396412e-01
-2.25435948e+00 4.03517425e-01 -5.92343509e-01 -5.78598857e-01
1.01027191e+00 -4.38852936e-01 2.52967805e-01 3.72590125e-01
5.11690080e-02 2.85325199e-01 1.63225126e+00 -1.24463451e+00
-2.30587617e-01 -1.07799195e-01 -2.59182602e-01 -1.07443362e-01
3.88572663e-02 -2.19387054e-01 -5.34674346e-01 -8.39769423e-01
2.47560412e-01 -7.51997888e-01 -4.34597880e-01 1.84316501e-01
-6.24481916e-01 -3.27007915e-03 2.96557069e-01 -6.12788916e-01
1.17255783e+00 -1.83647752e+00 3.73373359e-01 2.87607491e-01
1.25725210e-01 -1.13972276e-01 4.36664701e-01 6.88811004e-01
2.05714688e-01 -2.24778682e-01 -8.07420075e-01 -9.12585020e-01
1.81563690e-01 1.57555789e-01 -2.22788259e-01 9.49410081e-01
2.06228450e-01 4.58643019e-01 -1.03337586e+00 -6.02136135e-01
2.89144784e-01 7.86352217e-01 -6.18339658e-01 1.62020877e-01
-2.96896487e-01 5.72554767e-01 -4.04507071e-01 4.64253694e-01
9.04638946e-01 -3.78463328e-01 2.37843856e-01 -1.08004354e-01
-1.80309206e-01 2.09951967e-01 -1.68888021e+00 1.65485108e+00
-4.59832788e-01 4.07366693e-01 4.18358535e-01 -1.27835989e+00
8.63395989e-01 3.53655428e-01 3.79499316e-01 1.66787803e-01
2.69629024e-02 2.97978193e-01 -5.07996082e-01 -1.15169197e-01
4.33994472e-01 -3.88399452e-01 3.99151713e-01 4.64030057e-01
1.94403797e-01 -3.74745905e-01 4.32942748e-01 5.83996102e-02
6.81681991e-01 6.47741854e-01 5.07582366e-01 -5.68614602e-01
4.99779612e-01 -2.25494742e-01 6.99399412e-01 1.17336202e+00
7.09564462e-02 7.01371431e-01 3.70372146e-01 3.52205783e-02
-7.63495743e-01 -1.41522253e+00 -3.89170498e-01 4.83280718e-01
-2.24037673e-02 -4.02366936e-01 -9.81693029e-01 -2.40176931e-01
-3.40801746e-01 7.47962475e-01 -5.54967165e-01 5.18078268e-01
-2.29972750e-01 -1.19205308e+00 1.06398324e-02 1.97465807e-01
5.54700553e-01 -7.19576895e-01 -8.32418561e-01 5.95027149e-01
-1.95527598e-01 -7.68175840e-01 -2.63683021e-01 1.60245314e-01
-1.07870185e+00 -7.31487930e-01 -1.22049963e+00 -3.46425503e-01
6.90605521e-01 -1.36094196e-02 1.22557843e+00 -5.09276748e-01
-6.63998500e-02 6.44030690e-01 -2.18348131e-01 -2.42564842e-01
-3.48227769e-01 -2.35818079e-04 -1.62199333e-01 6.06537998e-01
1.60494864e-01 -9.81911302e-01 -7.62645006e-01 5.73234819e-02
-9.71727788e-01 4.67887558e-02 7.62060106e-01 1.15774441e+00
9.93428290e-01 1.56403080e-01 4.66403842e-01 -8.86338353e-01
5.43368101e-01 -6.26096129e-01 -9.57178056e-01 5.17645217e-02
-7.72632837e-01 1.77012071e-01 4.54606354e-01 -4.23217118e-01
-1.43954790e+00 1.66384757e-01 2.83861626e-02 -3.86219919e-01
5.21974862e-02 6.82326794e-01 2.09074784e-02 1.64038658e-01
1.94617972e-01 5.86329281e-01 9.60746184e-02 -6.96008384e-01
5.18195927e-01 3.55061859e-01 4.42597777e-01 -7.47775137e-01
8.49600494e-01 8.53225231e-01 7.88678005e-02 -1.06274188e+00
-6.33472145e-01 -3.00791115e-01 -4.68958080e-01 -3.22839618e-01
8.98142755e-01 -8.00624430e-01 -4.89903718e-01 5.85525513e-01
-1.11296225e+00 -1.72542438e-01 -5.00749946e-01 5.45661092e-01
-8.64156365e-01 5.25881112e-01 -5.43681324e-01 -1.34970367e+00
-2.15265527e-01 -1.20305192e+00 1.16568148e+00 1.52542144e-01
-1.14177912e-01 -1.14831686e+00 2.16850773e-01 1.09691896e-01
5.48253596e-01 2.86417753e-01 4.52560306e-01 -1.08481281e-01
-8.35394025e-01 -1.46754086e-01 8.92721638e-02 4.33410227e-01
5.00263795e-02 -3.94775579e-03 -8.80012631e-01 -3.73265535e-01
7.04669654e-01 -1.41405657e-01 9.64213669e-01 9.72939014e-01
8.40356052e-01 -2.81394660e-01 -2.42741704e-01 5.88052094e-01
1.62195492e+00 -1.62286803e-01 6.74189508e-01 1.85317218e-01
3.07080984e-01 3.45538139e-01 4.47299838e-01 1.05304289e+00
1.17847636e-01 5.27333140e-01 2.95242667e-01 2.80700754e-02
5.17819449e-02 -2.78243959e-01 2.30458066e-01 8.74832451e-01
-3.66445743e-02 -6.68420345e-02 -5.03846586e-01 6.31032884e-01
-1.92477560e+00 -9.52045798e-01 -6.13370575e-02 2.20164251e+00
1.15835178e+00 7.64429420e-02 -4.06378880e-03 1.33002654e-01
6.53394997e-01 4.74295527e-01 -6.50284588e-01 2.78035197e-02
1.45478591e-01 3.73715937e-01 3.36955130e-01 7.05548465e-01
-1.05545390e+00 3.65156591e-01 5.72977018e+00 1.30113912e+00
-7.18345940e-01 3.32419127e-01 5.58969438e-01 -6.56242818e-02
-3.29521447e-01 2.28461131e-01 -9.37609255e-01 7.07897961e-01
6.83735907e-01 -2.98234913e-02 4.94049817e-01 8.69451761e-01
4.80132639e-01 -4.54409391e-01 -9.61702585e-01 9.58865702e-01
-1.31070003e-01 -1.22094607e+00 -1.71933725e-01 1.68454796e-01
6.85756922e-01 -1.19122088e-01 4.16792855e-02 -9.39693153e-02
4.03996021e-01 -7.32947171e-01 9.39247131e-01 8.36221159e-01
4.56132978e-01 -5.93572319e-01 5.50487041e-01 3.36858422e-01
-9.66354489e-01 3.65707427e-01 -3.66474986e-01 3.45946848e-02
7.19853461e-01 1.50928748e+00 -5.12018859e-01 5.10090947e-01
5.85641801e-01 6.23687983e-01 -8.86756834e-03 1.09587491e+00
-5.29446244e-01 9.63121235e-01 -6.27546728e-01 -2.91998293e-02
1.86733708e-01 -8.55266929e-01 9.18389559e-01 1.27709782e+00
4.00445700e-01 -3.61343741e-01 6.59446120e-02 1.30765522e+00
1.24637172e-01 -1.53476730e-01 -8.31444338e-02 6.08528368e-02
5.45169830e-01 1.19434273e+00 -9.63675320e-01 -4.03127730e-01
-2.87625402e-01 8.04109454e-01 7.79346153e-02 5.99680543e-01
-8.12514007e-01 -2.38052741e-01 2.96521276e-01 6.87068999e-02
1.07425225e+00 -3.34874064e-01 -2.95599401e-01 -1.25401437e+00
8.36481526e-02 -7.85670400e-01 7.58425333e-03 -4.44946408e-01
-1.45879579e+00 3.27852011e-01 2.83078164e-01 -1.18771791e+00
-5.99947214e-01 -1.85365185e-01 -5.50496757e-01 1.09003186e+00
-1.58250749e+00 -9.22118127e-01 -8.55585858e-02 4.05134559e-01
7.31775522e-01 1.84053570e-01 6.37525678e-01 7.47914985e-02
-5.72480261e-01 2.20876813e-01 7.10691392e-01 -2.42306456e-01
3.13474596e-01 -1.35389566e+00 1.68720201e-01 1.16098714e+00
3.51094231e-02 7.49657929e-01 1.03128135e+00 -5.64690173e-01
-1.29261160e+00 -8.26878428e-01 6.39861345e-01 7.81643167e-02
7.57157981e-01 -2.75744647e-01 -5.36370218e-01 4.17266726e-01
3.58157068e-01 -1.51306018e-01 5.72555900e-01 -2.22078726e-01
1.42200217e-01 -3.88790742e-02 -1.16651857e+00 5.22877336e-01
5.47849238e-01 -2.87929684e-01 -4.18063253e-01 3.28492701e-01
9.22046527e-02 -2.04999030e-01 -7.32530892e-01 1.27673540e-02
5.29396117e-01 -1.06904650e+00 1.06134355e+00 2.21337046e-04
4.81624544e-01 -4.99691874e-01 -2.97241539e-01 -1.25858164e+00
-1.17418066e-01 -1.16133225e+00 -4.43180114e-01 1.38255215e+00
1.79272711e-01 -6.47293091e-01 7.24062085e-01 4.19306457e-01
1.52313069e-01 -9.35978174e-01 -1.12950504e+00 -8.49318087e-01
-1.04677528e-01 -5.76945484e-01 2.56326288e-01 5.42892396e-01
-5.30654907e-01 1.26204476e-01 -4.99723613e-01 2.09553257e-01
1.39532638e+00 2.70726502e-01 6.81156993e-01 -1.08658612e+00
-1.00847018e+00 -3.57271045e-01 1.82082616e-02 -1.56988525e+00
-1.38820931e-01 -4.88905340e-01 3.82399946e-01 -1.48356152e+00
1.01728432e-01 -2.64045775e-01 3.69601250e-02 -1.65004060e-01
-1.83712617e-01 1.12920001e-01 2.24194899e-02 3.07874739e-01
-3.14204663e-01 1.06480765e+00 9.29672420e-01 -7.34418184e-02
-6.80321828e-02 3.00812304e-01 -4.72017288e-01 1.00806141e+00
5.16891360e-01 -7.56751418e-01 -5.52200794e-01 -2.85939246e-01
1.70701146e-01 2.70422727e-01 4.02506858e-01 -7.82498121e-01
3.73514235e-01 -1.45513162e-01 1.75559655e-01 -7.88397253e-01
5.61069131e-01 -4.83458281e-01 3.42349589e-01 4.51811582e-01
-5.18273301e-02 -2.95692384e-01 -2.93087274e-01 9.78989005e-01
-2.49609679e-01 -7.70964205e-01 9.64283764e-01 -2.45325983e-01
-9.53386426e-02 2.57744007e-02 -7.43077457e-01 1.19334713e-01
5.79416335e-01 -1.78656012e-01 3.04077178e-01 -5.63646853e-01
-8.12007010e-01 -1.40524492e-01 2.79155791e-01 -5.19213378e-01
6.55547559e-01 -1.15851057e+00 -8.00534546e-01 -1.31376823e-02
-4.77230042e-01 2.84083337e-01 1.33931294e-01 1.13490582e+00
-3.92734021e-01 5.86881563e-02 4.95276183e-01 -8.91332448e-01
-6.12377048e-01 2.59768873e-01 1.08352348e-01 -6.47107422e-01
-6.49937451e-01 8.47479522e-01 6.72852620e-02 5.30168694e-03
1.69221908e-01 -3.14628690e-01 1.22778893e-01 2.01101765e-01
5.00682473e-01 6.89614832e-01 -1.45629019e-01 -1.54536843e-01
-2.01190397e-01 3.24651837e-01 1.65670395e-01 -6.57146871e-01
1.62215602e+00 -3.42712402e-01 -1.70506597e-01 6.51482642e-01
8.56400132e-01 -8.78680148e-04 -1.67201698e+00 -2.67108917e-01
-2.20249951e-01 -5.65811574e-01 4.65185016e-01 -4.21298563e-01
-1.12121284e+00 7.95299888e-01 3.97876054e-01 2.66256630e-01
1.09518528e+00 -2.74629086e-01 7.73653090e-01 1.33226424e-01
5.25432229e-01 -1.16735291e+00 -1.13031246e-01 1.25579789e-01
7.23691642e-01 -1.03897262e+00 4.56550092e-01 -3.28097999e-01
-2.58820504e-01 1.01939940e+00 -2.26010472e-01 -4.07025307e-01
9.72033978e-01 7.92084560e-02 -3.84771407e-01 8.48900899e-03
-4.25617576e-01 -7.12963119e-02 8.98013040e-02 5.69470465e-01
2.42002949e-01 -1.47858411e-01 -6.29494488e-01 4.51417595e-01
1.61946786e-03 2.72968590e-01 5.31884551e-01 9.35594797e-01
-1.67708859e-01 -1.04512560e+00 -4.33660239e-01 3.65930855e-01
-5.43576539e-01 -3.88716757e-01 3.46735775e-01 6.18744254e-01
-1.83923900e-01 1.04484701e+00 -1.48648411e-01 3.71414989e-01
-2.31647581e-01 8.87544267e-03 5.27790308e-01 -4.00203258e-01
-1.48492843e-01 6.38919711e-01 -1.85201168e-01 -4.58570898e-01
-8.61672342e-01 -1.10841882e+00 -6.53811812e-01 -5.39030507e-02
-4.97633606e-01 9.70680937e-02 8.16263855e-01 1.02278805e+00
1.28146529e-01 4.16988969e-01 3.81551236e-01 -1.32256949e+00
-9.90283489e-01 -1.04304409e+00 -9.31001008e-01 7.59614334e-02
3.53289813e-01 -9.02495742e-01 -5.66543579e-01 1.54627949e-01] | [6.971826553344727, 3.8488807678222656] |
432d5027-a580-451c-ae9c-3a4ce02c32a8 | automating-dbscan-via-deep-reinforcement | 2208.04537 | null | https://arxiv.org/abs/2208.04537v1 | https://arxiv.org/pdf/2208.04537v1.pdf | Automating DBSCAN via Deep Reinforcement Learning | DBSCAN is widely used in many scientific and engineering fields because of its simplicity and practicality. However, due to its high sensitivity parameters, the accuracy of the clustering result depends heavily on practical experience. In this paper, we first propose a novel Deep Reinforcement Learning guided automatic DBSCAN parameters search framework, namely DRL-DBSCAN. The framework models the process of adjusting the parameter search direction by perceiving the clustering environment as a Markov decision process, which aims to find the best clustering parameters without manual assistance. DRL-DBSCAN learns the optimal clustering parameter search policy for different feature distributions via interacting with the clusters, using a weakly-supervised reward training policy network. In addition, we also present a recursive search mechanism driven by the scale of the data to efficiently and controllably process large parameter spaces. Extensive experiments are conducted on five artificial and real-world datasets based on the proposed four working modes. The results of offline and online tasks show that the DRL-DBSCAN not only consistently improves DBSCAN clustering accuracy by up to 26% and 25% respectively, but also can stably find the dominant parameters with high computational efficiency. The code is available at https://github.com/RingBDStack/DRL-DBSCAN. | ['Philip S. Yu', 'Jingyi Zhang', 'Qingyun Sun', 'Jia Wu', 'Yingtong Dou', 'Hao Peng', 'Ruitong Zhang'] | 2022-08-09 | null | null | null | null | ['unsupervised-spatial-clustering'] | ['time-series'] | [-5.85694194e-01 -5.08456230e-01 -1.88776851e-01 -3.03879797e-01
-8.53612721e-01 -7.72654712e-01 3.74375105e-01 -1.57493770e-01
-4.31254655e-01 6.56762779e-01 -1.13074802e-01 -1.19647853e-01
-5.49718916e-01 -7.30257094e-01 -5.68834364e-01 -1.31641483e+00
-1.12802155e-01 9.40023661e-01 3.54527980e-01 2.64391392e-01
4.96540338e-01 5.69699764e-01 -1.30247438e+00 -4.65360582e-02
1.01952529e+00 9.90235209e-01 4.37866569e-01 3.81708413e-01
-6.20945403e-03 3.66987705e-01 -4.16621357e-01 5.34487516e-02
4.04821873e-01 -2.09861979e-01 -3.69967848e-01 4.31231111e-02
-4.22315627e-01 -3.09438616e-01 -3.34059238e-01 1.08977365e+00
7.21624434e-01 3.19732159e-01 3.15318912e-01 -1.23197460e+00
-5.54382563e-01 7.99861491e-01 -6.43161058e-01 2.32860714e-01
-2.65448168e-02 6.97054982e-01 7.70643473e-01 -4.47603792e-01
3.20901603e-01 1.22388816e+00 4.06239837e-01 4.17118996e-01
-1.14018881e+00 -8.21421087e-01 2.10418880e-01 4.31309700e-01
-1.74142706e+00 -3.23371381e-01 5.86216152e-01 -3.50920290e-01
7.57195055e-01 1.97653845e-03 7.57094681e-01 8.62248123e-01
-1.34889837e-02 8.82541120e-01 8.43316019e-01 -1.17228590e-02
8.93809676e-01 -3.13561797e-01 -8.48466605e-02 3.60871971e-01
1.74179643e-01 3.58706206e-01 -1.18361942e-01 -2.12046653e-01
9.26028073e-01 1.38757646e-01 -1.20129116e-01 -6.97237194e-01
-1.13698173e+00 8.74559104e-01 4.81136054e-01 1.67205453e-01
-4.30277735e-01 4.23706919e-01 2.79069543e-01 -7.52596781e-02
-9.51406062e-02 3.74477118e-01 -2.80803144e-01 -4.20473993e-01
-6.60575449e-01 3.25265318e-01 4.35815394e-01 7.11593390e-01
6.19751513e-01 4.47987616e-02 -1.98801160e-01 9.29506600e-01
2.89808631e-01 6.54256999e-01 7.05828667e-01 -1.24441648e+00
2.21543476e-01 4.44060832e-01 4.80553091e-01 -8.25606823e-01
-3.85983646e-01 -2.08870381e-01 -9.19448018e-01 1.52047351e-01
2.29538769e-01 -2.81763643e-01 -7.56887376e-01 1.41073561e+00
6.57785833e-01 2.99572587e-01 -1.15492560e-01 1.08788002e+00
8.57022479e-02 7.75710523e-01 -3.49407494e-02 -2.54748613e-01
8.49570870e-01 -6.27866745e-01 -4.97559637e-01 1.10496104e-01
2.89211363e-01 -2.69123375e-01 1.21014857e+00 6.39532864e-01
-1.02132404e+00 -2.53741205e-01 -8.37149441e-01 5.84109366e-01
-2.69419640e-01 2.73091227e-01 4.91558641e-01 2.92604536e-01
-1.13338304e+00 5.65939724e-01 -1.27556050e+00 -3.19937170e-01
6.75481856e-01 5.27561724e-01 2.80384451e-01 6.12405427e-02
-9.37115073e-01 3.60664666e-01 7.59000063e-01 2.81255573e-01
-9.20797169e-01 -4.88108546e-01 -2.81744629e-01 1.20057687e-01
6.14723027e-01 -3.71906430e-01 1.23579347e+00 -4.55204040e-01
-1.93285847e+00 4.00668532e-01 1.34656474e-01 -5.54497302e-01
6.93815053e-01 -2.27565289e-01 -3.14835370e-01 1.13465171e-02
9.84476358e-02 6.29202366e-01 6.59929216e-01 -1.11328769e+00
-7.67518044e-01 -5.42897880e-01 -3.66400689e-01 2.90150672e-01
-2.04788566e-01 -2.44568318e-01 -8.32712710e-01 -4.82708246e-01
-1.52559271e-02 -9.62717116e-01 -4.81256872e-01 -3.10619086e-01
-4.01900053e-01 -4.48229492e-01 7.24255979e-01 -3.54724735e-01
1.31505537e+00 -2.18050885e+00 1.09123155e-01 4.84967321e-01
-6.47520367e-03 2.78326869e-01 1.77079048e-02 3.06128919e-01
3.05702597e-01 3.58418301e-02 -3.58827055e-01 2.03186840e-01
8.29060450e-02 1.97402656e-01 -3.03378016e-01 5.52365124e-01
-1.54583275e-01 7.39550591e-01 -9.97272611e-01 -3.77181917e-01
4.06066954e-01 2.62588531e-01 -4.88209605e-01 2.03672171e-01
-3.55033517e-01 4.32528645e-01 -7.58389115e-01 6.64717555e-01
6.28460586e-01 -3.40489656e-01 2.53375441e-01 3.44229415e-02
-1.40114486e-01 -3.09480429e-01 -1.57367396e+00 1.47573543e+00
-1.82774767e-01 2.28623271e-01 1.96526200e-01 -1.18785465e+00
9.51090395e-01 -1.54592946e-01 5.66395998e-01 -9.27405536e-01
1.50436148e-01 1.28191546e-01 -1.14335835e-01 -4.59154665e-01
1.46041006e-01 1.51344210e-01 -5.38998889e-03 5.29100716e-01
-3.72288346e-01 -9.86892381e-04 3.27419072e-01 -7.59583116e-02
1.12782705e+00 9.82167665e-03 1.21307559e-01 -2.90802330e-01
4.64117706e-01 9.15624499e-02 8.71022046e-01 6.68145478e-01
-2.73304313e-01 1.60406753e-01 3.20982993e-01 -3.47596586e-01
-8.51378858e-01 -1.14381754e+00 -1.21514015e-01 9.97186482e-01
4.47553724e-01 4.51795273e-02 -8.66977334e-01 -3.44124794e-01
1.98911145e-01 1.00014555e+00 -4.29074287e-01 -1.37013748e-01
-6.02612674e-01 -8.54306698e-01 3.52420032e-01 4.31993455e-01
7.00711668e-01 -1.32150793e+00 -6.57621324e-01 2.62527943e-01
2.38237455e-01 -6.38584852e-01 -4.98359501e-01 9.47685391e-02
-9.69348788e-01 -1.17453051e+00 -3.74696672e-01 -6.42758250e-01
4.73136902e-01 1.44037321e-01 6.64473414e-01 -5.22446036e-02
-3.22534889e-01 1.04802512e-01 -2.53565431e-01 -8.22550282e-02
-1.24180607e-01 1.84581429e-01 1.37440383e-01 2.48445533e-02
5.54593205e-01 -5.64184964e-01 -8.43102813e-01 6.49133801e-01
-8.30572605e-01 -2.55067766e-01 6.09894693e-01 5.90946078e-01
1.02042294e+00 5.52923918e-01 6.77512705e-01 -6.18274093e-01
9.73495126e-01 -5.64602077e-01 -1.26075661e+00 2.91584373e-01
-7.86448419e-01 2.58636534e-01 7.79706955e-01 -3.43475670e-01
-8.92753303e-01 2.53450245e-01 2.04487696e-01 -9.07549977e-01
-2.49619246e-01 1.52173162e-01 -2.34457538e-01 4.00849998e-01
4.25706029e-01 5.40341735e-01 1.11125052e-01 -3.44367802e-01
3.25332522e-01 6.04004145e-01 6.02911532e-01 -7.58938789e-01
5.79069138e-01 4.69326228e-01 -3.82949889e-01 -3.84718031e-01
-3.37421507e-01 -4.46978927e-01 -5.27284861e-01 -2.18847305e-01
7.52507627e-01 -7.71157861e-01 -1.25295401e+00 5.53890646e-01
-7.15530634e-01 -6.95420504e-01 -7.57446289e-02 1.88169912e-01
-7.09495008e-01 2.16702044e-01 -4.64367419e-01 -7.48855114e-01
-4.41288888e-01 -1.38176692e+00 8.85835767e-01 2.41375357e-01
1.78110167e-01 -8.73069823e-01 1.60072595e-01 7.47194737e-02
3.40936869e-01 1.50923416e-01 8.37762713e-01 -6.59792423e-01
-7.80390084e-01 -1.69617180e-02 1.33761205e-02 7.28556979e-03
1.17799722e-01 2.56081998e-01 -5.31022549e-01 -4.92898881e-01
-2.38853991e-01 -1.94972426e-01 5.73792815e-01 7.42761195e-01
1.75842071e+00 -3.89218181e-01 -4.67583090e-01 6.96433902e-01
1.46236086e+00 8.05761158e-01 5.15457332e-01 5.13410449e-01
4.65814859e-01 1.75083965e-01 7.47975647e-01 7.30469584e-01
4.08160329e-01 5.19746721e-01 5.77606678e-01 3.89937520e-01
3.54841769e-01 -3.60732853e-01 2.08714798e-01 6.68632865e-01
2.03231961e-01 -1.67377412e-01 -1.15038502e+00 3.89894933e-01
-2.26450419e+00 -9.94867384e-01 3.04226965e-01 2.35767889e+00
7.96752870e-01 1.28009662e-01 4.71141011e-01 -7.35500455e-02
1.14160001e+00 -3.06030303e-01 -1.46886861e+00 -1.05098225e-01
4.32936624e-02 -3.01656693e-01 6.22217834e-01 4.36430454e-01
-9.35156703e-01 1.09907401e+00 5.82133436e+00 1.08638835e+00
-1.18976796e+00 -3.16670626e-01 7.75972724e-01 -4.05431420e-01
4.61248755e-02 -2.34625891e-01 -5.91913402e-01 1.01956773e+00
7.44115412e-01 -3.13190013e-01 1.13348937e+00 1.01002848e+00
6.17305994e-01 -1.30780548e-01 -9.17564213e-01 1.37467706e+00
-5.88313103e-01 -1.59720540e+00 7.04236701e-02 1.39932290e-01
6.36145532e-01 1.81962952e-01 -1.39205456e-02 2.55989492e-01
1.06785560e+00 -9.37879145e-01 6.91641450e-01 5.99310279e-01
5.01568139e-01 -1.20295012e+00 4.86003786e-01 4.85343754e-01
-9.41253901e-01 -5.84648550e-01 -4.37590420e-01 2.99452931e-01
-1.44281685e-01 6.90636575e-01 -9.33780849e-01 6.82838708e-02
1.14317334e+00 7.57428527e-01 -2.72724807e-01 1.24769223e+00
-2.08745286e-01 6.35262191e-01 -4.43856537e-01 -3.19742739e-01
2.87755847e-01 -7.70991564e-01 2.74842352e-01 9.72407877e-01
4.60105956e-01 2.52368063e-01 1.78694233e-01 1.00482547e+00
4.53933291e-02 -1.83296502e-01 -3.14558774e-01 -1.83410235e-02
1.25582325e+00 1.07685435e+00 -8.86919260e-01 -1.87828496e-01
3.74692470e-01 7.54947484e-01 4.19142038e-01 4.78634745e-01
-8.68022621e-01 -3.70478958e-01 5.85455477e-01 -6.99342415e-02
6.70118451e-01 -1.92312256e-01 -3.73994231e-01 -7.48270333e-01
-1.75970182e-01 -8.72857273e-01 5.56551814e-01 -7.03999996e-01
-1.28550255e+00 3.70394677e-01 -1.34117484e-01 -1.04437470e+00
-3.36129725e-01 -1.81175381e-01 -8.19645822e-01 4.04308826e-01
-1.01057267e+00 -3.73521417e-01 -2.99960166e-01 7.46747375e-01
5.64361691e-01 -3.99844557e-01 3.78439367e-01 3.83468978e-02
-1.17825150e+00 4.70041901e-01 9.79066133e-01 -6.54807836e-02
6.28028989e-01 -1.24408579e+00 2.14530289e-01 5.44552922e-01
-3.19408447e-01 4.55095023e-01 6.40859902e-01 -5.02382517e-01
-1.64197350e+00 -1.12392890e+00 -7.24216178e-02 -1.48139209e-01
7.82331467e-01 -2.53742099e-01 -9.59946632e-01 3.49260926e-01
1.52759366e-02 -1.80283219e-01 4.23210710e-01 -2.09371582e-01
7.92909339e-02 -5.13592720e-01 -1.40665996e+00 5.75144768e-01
8.89204741e-01 4.45580669e-02 -1.90859318e-01 3.07010889e-01
6.88776493e-01 -2.30136812e-01 -7.21954167e-01 8.19804445e-02
1.08706661e-01 -1.03187263e+00 9.40177321e-01 -3.14110398e-01
1.11871704e-01 -5.53668261e-01 -1.36206850e-01 -1.54987872e+00
-4.93996859e-01 -6.15976274e-01 -2.38437831e-01 1.14747214e+00
-2.99078170e-02 -6.94692791e-01 8.78404915e-01 5.09915948e-01
9.45405662e-02 -9.42812324e-01 -1.08820903e+00 -7.03825116e-01
1.00786448e-01 -3.25062901e-01 1.11061895e+00 7.41240382e-01
-1.88466132e-01 6.17856830e-02 4.84796949e-02 3.74479473e-01
8.51221979e-01 3.43833596e-01 6.46649599e-01 -1.04298675e+00
-3.49961579e-01 -7.33636320e-01 -1.20423056e-01 -1.00806761e+00
1.23524576e-01 -6.35296226e-01 3.29323649e-01 -1.39936411e+00
9.30329859e-02 -8.35035324e-01 -3.67620707e-01 4.88377780e-01
1.41568668e-02 -3.72175187e-01 -3.59498635e-02 5.72288334e-01
-1.13710296e+00 7.40252912e-01 9.39104319e-01 -1.73579175e-02
-5.50372124e-01 1.35477945e-01 -6.31852627e-01 5.73934793e-01
1.12429667e+00 -2.00355038e-01 -5.67471385e-01 -5.22159100e-01
2.34441236e-02 1.84271961e-01 1.19128212e-01 -9.17764127e-01
5.11845887e-01 -5.62111139e-01 4.20863986e-01 -6.82765782e-01
3.07273380e-02 -6.74043655e-01 3.14514227e-02 5.90497613e-01
-3.79243076e-01 1.64837316e-01 9.72447246e-02 8.92506123e-01
3.60862873e-02 3.35349888e-02 8.55363488e-01 1.99320484e-02
-8.25696468e-01 4.71797049e-01 -3.54597509e-01 3.97168845e-02
1.38066173e+00 -2.13309720e-01 -5.83683215e-02 -3.41378450e-01
-6.70404017e-01 9.29832220e-01 5.38820326e-01 2.99592316e-01
5.47376812e-01 -1.42635798e+00 -3.89160573e-01 1.52151406e-01
-1.93269521e-01 4.09769565e-01 1.86387882e-01 4.22592163e-01
-5.76159179e-01 3.79109025e-01 -6.35975897e-02 -7.97482252e-01
-7.21028209e-01 7.84643352e-01 5.21954715e-01 -9.23340842e-02
-4.39438105e-01 5.31335533e-01 -1.90271839e-01 -6.21748686e-01
6.31985545e-01 -1.96988359e-01 1.06957637e-01 -1.98142409e-01
4.58870143e-01 8.83983254e-01 -6.45665312e-03 -1.42561113e-02
-4.04373705e-01 4.23482805e-01 3.00519988e-02 1.05568431e-01
1.53876913e+00 -2.15992063e-01 -1.74695179e-02 2.54465669e-01
8.41671348e-01 -6.18309796e-01 -1.65724432e+00 -1.23581551e-01
1.67367533e-01 -4.14564699e-01 1.56654835e-01 -9.00332510e-01
-1.31256235e+00 6.03929281e-01 8.74238849e-01 1.59174949e-01
1.06547439e+00 -7.06991330e-02 6.31099522e-01 6.50611341e-01
4.69738305e-01 -1.54401541e+00 9.24604386e-02 2.13367730e-01
7.94600606e-01 -1.17649972e+00 -2.87749469e-01 2.90177584e-01
-8.73961747e-01 1.06790078e+00 8.01475585e-01 -2.93483824e-01
7.27752090e-01 2.76693493e-01 -6.31170198e-02 -1.95640802e-01
-6.17464304e-01 1.78795427e-01 -4.57095027e-01 6.11391902e-01
-1.57428294e-01 2.38906398e-01 5.74408993e-02 5.46767712e-01
-1.25929222e-01 4.76877950e-02 1.44799531e-01 6.36879265e-01
-7.05089629e-01 -8.49116266e-01 -6.41028464e-01 5.51583230e-01
6.68694079e-02 3.02182227e-01 -2.02206060e-01 5.53541780e-01
-1.13067448e-01 9.07790661e-01 3.73770595e-01 -3.26329917e-01
1.77848235e-01 -2.84352005e-01 -4.47329460e-03 -2.32698899e-02
-2.79813737e-01 1.76668480e-01 -6.21462882e-01 -7.33843088e-01
7.22705293e-03 -9.39957619e-01 -1.83815539e+00 -4.94885594e-01
6.25216439e-02 2.61077195e-01 4.91293728e-01 7.46930361e-01
8.43553364e-01 3.03743333e-01 9.13558602e-01 -7.33999372e-01
-9.50594962e-01 -6.85455143e-01 -7.38609910e-01 1.52123094e-01
6.53505996e-02 -6.98399723e-01 -2.14704081e-01 -2.61512697e-01] | [4.093035697937012, 2.245962619781494] |
b989a5ab-191b-430a-ba9a-710bb5f99d6d | triplet-knowledge-distillation | 2305.15975 | null | https://arxiv.org/abs/2305.15975v1 | https://arxiv.org/pdf/2305.15975v1.pdf | Triplet Knowledge Distillation | In Knowledge Distillation, the teacher is generally much larger than the student, making the solution of the teacher likely to be difficult for the student to learn. To ease the mimicking difficulty, we introduce a triplet knowledge distillation mechanism named TriKD. Besides teacher and student, TriKD employs a third role called anchor model. Before distillation begins, the pre-trained anchor model delimits a subspace within the full solution space of the target problem. Solutions within the subspace are expected to be easy targets that the student could mimic well. Distillation then begins in an online manner, and the teacher is only allowed to express solutions within the aforementioned subspace. Surprisingly, benefiting from accurate but easy-to-mimic hints, the student can finally perform well. After the student is well trained, it can be used as the new anchor for new students, forming a curriculum learning strategy. Our experiments on image classification and face recognition with various models clearly demonstrate the effectiveness of our method. Furthermore, the proposed TriKD is also effective in dealing with the overfitting issue. Moreover, our theoretical analysis supports the rationality of our triplet distillation. | ['Shiguang Shan', 'Zhongqin Wu', 'Chunrui Han', 'Meina Kan', 'Dongyang Liu', 'Xijun Wang'] | 2023-05-25 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [-1.57888949e-01 3.61386806e-01 -2.48396412e-01 -1.67065293e-01
-4.80039209e-01 -5.33708513e-01 -3.64669263e-02 -5.56409881e-02
-1.18946642e-01 6.82802022e-01 -4.78228748e-01 -3.39518249e-01
-2.55805790e-01 -8.52245212e-01 -8.31879973e-01 -9.44820940e-01
3.54022980e-01 4.55731332e-01 2.02791747e-02 -1.32057607e-01
2.50096582e-02 4.97384161e-01 -1.41943061e+00 -2.32961372e-01
1.43578386e+00 9.80936944e-01 3.46189111e-01 1.00227654e-01
-1.92238420e-01 6.39406919e-01 -6.58957660e-01 -4.31354523e-01
1.98964253e-01 -2.55078077e-01 -5.32158554e-01 2.50389934e-01
4.63257909e-01 -4.07095283e-01 -4.06356692e-01 1.06673348e+00
4.61329043e-01 2.54805654e-01 4.42273110e-01 -1.15972948e+00
-6.16774976e-01 7.21124172e-01 -5.04181564e-01 -1.59682453e-01
3.19741070e-01 1.44092217e-02 8.69053006e-01 -1.02198648e+00
2.77178645e-01 1.01717782e+00 2.57305831e-01 5.37122846e-01
-1.13342035e+00 -1.02198839e+00 3.92393082e-01 1.81554452e-01
-1.47481048e+00 -2.33302012e-01 1.02825117e+00 -3.94161135e-01
9.34227929e-02 1.00834504e-01 7.60381460e-01 8.25178623e-01
-5.52287400e-01 1.17403889e+00 9.63102281e-01 -3.50294709e-01
2.11724669e-01 6.23422265e-01 2.57368147e-01 8.43977690e-01
1.76321894e-01 1.07513107e-01 -4.12397146e-01 -3.40691283e-02
7.53879726e-01 5.13382107e-02 -6.40287876e-01 -8.15183520e-01
-8.12267303e-01 7.09450245e-01 5.07693529e-01 6.65533841e-02
-2.64814824e-01 -2.71342367e-01 -1.13005176e-01 4.21671271e-01
8.01217742e-03 4.15351152e-01 -4.95507181e-01 -1.00314170e-01
-9.30027604e-01 2.50165146e-02 7.93167353e-01 9.39416170e-01
9.49409008e-01 1.41668975e-01 -3.05575263e-02 7.52570510e-01
2.88593233e-01 2.87069559e-01 3.38742375e-01 -7.79924750e-01
5.13177752e-01 1.06118548e+00 5.83927147e-02 -9.07037258e-01
1.43136978e-01 -9.00610030e-01 -7.88560092e-01 1.02605477e-01
6.10512674e-01 -2.81365663e-01 -7.77843595e-01 1.82150924e+00
7.60363340e-01 7.89222896e-01 1.68324083e-01 8.87859106e-01
7.22363830e-01 7.60846257e-01 -1.18672535e-01 -6.28256440e-01
1.26721215e+00 -8.07233691e-01 -4.59957421e-01 -4.06608433e-02
7.50089586e-01 -5.71415961e-01 8.99529159e-01 8.80055189e-01
-1.05770493e+00 -5.94947040e-01 -9.16239440e-01 2.89755434e-01
1.19057037e-01 3.43673021e-01 4.41021472e-01 5.86210728e-01
-7.15850592e-01 4.03035253e-01 -4.83416528e-01 1.53313577e-01
4.29242700e-01 4.87003356e-01 -2.33330458e-01 -3.84579718e-01
-1.02607834e+00 6.72995925e-01 6.48237288e-01 8.40948373e-02
-1.03922963e+00 -9.90801275e-01 -6.35067701e-01 3.12482119e-01
8.22390378e-01 -4.11379576e-01 1.15931368e+00 -9.85984325e-01
-1.67968857e+00 4.78606910e-01 5.46961129e-02 -4.40844335e-02
7.14119852e-01 -4.75772731e-02 -4.68816757e-02 2.10268497e-01
-2.24712655e-01 5.22108257e-01 9.60333228e-01 -1.36051655e+00
-4.95937228e-01 -3.17099869e-01 3.72995585e-01 4.62217361e-01
-7.70219445e-01 -5.45535922e-01 -5.63531041e-01 -4.08822268e-01
3.14984888e-01 -7.95186460e-01 -1.06189579e-01 -3.21533605e-02
-3.43002826e-01 -6.38214648e-01 8.90689731e-01 -2.90994585e-01
1.50398302e+00 -2.33970737e+00 1.39338106e-01 5.59747815e-01
4.23642397e-01 5.19291759e-01 -2.27550238e-01 -2.16113664e-02
-3.28164190e-01 -5.67757636e-02 8.25703666e-02 -4.00760025e-02
-6.99082166e-02 3.38697165e-01 -3.96777779e-01 1.01552911e-01
3.95273007e-02 5.75101376e-01 -9.54002559e-01 -3.79521132e-01
3.72680128e-02 1.12638161e-01 -7.59444296e-01 5.97481191e-01
-2.89088339e-01 3.78842592e-01 -9.16968822e-01 4.22288209e-01
8.84838045e-01 -3.87513459e-01 2.38944843e-01 -1.62065160e-02
1.34592116e-01 8.96855071e-02 -1.61875618e+00 1.29821086e+00
-3.37814212e-01 1.42243385e-01 1.52075604e-01 -1.45577717e+00
1.21076417e+00 3.01617801e-01 3.85799378e-01 -3.03559870e-01
-2.83344034e-02 1.12170123e-01 1.99864283e-01 -4.23246086e-01
1.73873246e-01 5.69190085e-02 4.15541023e-01 3.20571303e-01
1.37322880e-02 6.51883185e-02 1.98356900e-02 3.06455880e-01
6.60591483e-01 1.19163603e-01 5.47603890e-02 -1.94422621e-02
7.58316219e-01 -2.96819240e-01 8.39989781e-01 5.83429813e-01
-1.01356857e-01 -4.48439158e-02 5.40126204e-01 -2.03076139e-01
-5.55737615e-01 -1.08329165e+00 -4.51304764e-02 1.13586307e+00
1.82603270e-01 -4.08344805e-01 -6.84313536e-01 -9.08217192e-01
-3.04438267e-02 4.24681067e-01 -4.51312274e-01 -2.80389339e-01
-5.70527434e-01 -1.76536649e-01 6.90661445e-02 3.48452538e-01
5.29128969e-01 -6.26186907e-01 -1.44488662e-01 3.07149112e-01
-5.95460162e-02 -8.99459422e-01 -4.71988946e-01 1.84805661e-01
-9.28638995e-01 -1.20112407e+00 -8.77303421e-01 -1.04744852e+00
1.14121830e+00 5.04834235e-01 6.27095997e-01 4.52496946e-01
1.98767018e-02 3.44931304e-01 -1.20299242e-01 -1.59130111e-01
-2.60869414e-01 -1.68876737e-01 2.62844473e-01 1.11528866e-01
5.25980830e-01 -6.51255846e-01 -5.99596024e-01 4.44422424e-01
-5.70019901e-01 1.72132581e-01 4.95678574e-01 9.04777288e-01
4.23752427e-01 3.53276342e-01 7.10202038e-01 -4.97100383e-01
4.24003690e-01 -3.83459181e-01 -9.60087478e-01 3.76519233e-01
-6.52778208e-01 1.43105119e-01 1.02720785e+00 -1.08002996e+00
-8.67816567e-01 1.48386374e-01 2.00286761e-01 -8.93949866e-01
-7.45409951e-02 7.05542326e-01 -3.78569961e-01 -2.90408105e-01
4.95556593e-01 4.42844123e-01 1.88390864e-03 -4.64049399e-01
1.11602552e-01 5.20693660e-01 3.32649857e-01 -1.09961963e+00
1.14866805e+00 -8.14108104e-02 -1.07502624e-01 -7.76301205e-01
-9.80697751e-01 -7.02338070e-02 -3.71733993e-01 -3.13894004e-01
3.35673749e-01 -1.07114542e+00 -1.08715308e+00 4.59865272e-01
-9.00375724e-01 -1.68293059e-01 -1.54102966e-01 6.30256236e-01
3.48592614e-04 4.58140165e-01 -2.78549075e-01 -8.69588494e-01
1.63882390e-01 -1.50126553e+00 2.31689975e-01 7.67504096e-01
2.22135484e-01 -8.90148938e-01 -3.45567346e-01 6.71809256e-01
2.45112996e-03 -2.20389932e-01 1.06535041e+00 -1.00328457e+00
-8.22121739e-01 1.43828653e-02 -1.57073755e-02 3.90034318e-01
-3.62274908e-02 7.72836953e-02 -8.49428117e-01 -4.71448690e-01
-2.50786375e-02 -7.07912326e-01 7.14919806e-01 6.15488142e-02
1.47461486e+00 -4.98921096e-01 -3.47172827e-01 4.84753370e-01
1.03522480e+00 3.33285779e-01 3.46094638e-01 2.65546083e-01
6.28133357e-01 5.69531977e-01 5.02069354e-01 3.37207586e-01
5.54416895e-01 4.98001754e-01 3.01650107e-01 -5.10837417e-03
3.88495803e-01 -4.56756383e-01 4.81431305e-01 1.15403008e+00
1.75217167e-01 8.31990838e-02 -7.50296056e-01 3.64007711e-01
-1.67091477e+00 -6.47554457e-01 8.40203688e-02 2.30307102e+00
1.28086722e+00 6.30751401e-02 5.73603213e-02 1.87693402e-01
6.15681052e-01 -2.78265595e-01 -6.47998214e-01 -1.85767606e-01
2.29283169e-01 1.18491143e-01 -9.81400162e-02 3.58283997e-01
-7.63708353e-01 1.04679537e+00 5.16714478e+00 1.31952024e+00
-1.19856238e+00 -4.26539004e-01 6.37083769e-01 1.83476910e-01
-3.49573791e-01 1.36747852e-01 -1.03847814e+00 3.98394406e-01
5.69251776e-01 -1.87381312e-01 5.43866277e-01 9.85439956e-01
6.07250556e-02 -1.77020565e-01 -1.27603126e+00 9.94253218e-01
-1.56190738e-01 -9.79110301e-01 1.28281534e-01 1.09379359e-01
9.34366465e-01 -7.43414640e-01 2.44169921e-01 6.45400465e-01
2.78721303e-01 -9.15585995e-01 2.55701840e-01 2.63207018e-01
4.87701535e-01 -1.00254273e+00 2.14916691e-01 8.68450761e-01
-9.69776809e-01 -2.09623605e-01 -4.74678576e-01 -4.32882681e-02
-5.79437971e-01 5.70243955e-01 -9.65512514e-01 5.40420353e-01
1.03050530e-01 4.22278374e-01 -4.29074496e-01 1.10881793e+00
-5.94120800e-01 7.05677032e-01 -3.36302727e-01 -1.72149763e-02
1.98365480e-01 -6.06370091e-01 3.95984441e-01 6.23586297e-01
5.53045154e-01 7.23072708e-01 7.55296528e-01 7.22884119e-01
-2.35039979e-01 1.61242157e-01 -3.36258441e-01 -1.85125843e-01
7.98211217e-01 1.12910509e+00 -3.15920413e-01 -3.91540736e-01
-2.81304330e-01 7.12247491e-01 7.47129619e-01 6.63824022e-01
-6.84893548e-01 -3.68374646e-01 4.76409018e-01 -1.79704010e-01
4.71738756e-01 -1.24889694e-01 -9.72513631e-02 -1.33768713e+00
9.18954685e-02 -1.07826424e+00 3.29700679e-01 -7.43655741e-01
-9.91056979e-01 2.24688992e-01 -1.35281563e-01 -1.29475248e+00
7.23964274e-02 -5.86345851e-01 -7.84476161e-01 8.30725193e-01
-1.58606064e+00 -6.09908044e-01 -4.49091613e-01 8.01903129e-01
9.46271494e-02 -2.93831646e-01 6.40172899e-01 1.64050102e-01
-9.54845190e-01 1.11482441e+00 7.08103627e-02 2.57106364e-01
6.35339320e-01 -1.15225112e+00 -6.63680851e-01 6.15735769e-01
3.69340032e-02 8.35686922e-01 4.77918357e-01 -4.29063171e-01
-1.56637180e+00 -8.63884389e-01 4.62840676e-01 9.97748524e-02
6.90129519e-01 -1.25740618e-01 -1.33739316e+00 4.14314836e-01
-2.37036198e-01 -3.49486709e-01 8.28006208e-01 1.21609159e-01
-4.23811495e-01 -4.55965132e-01 -9.23117280e-01 7.20501304e-01
5.24211526e-01 -3.46747130e-01 -7.82766402e-01 4.14014161e-01
5.65084517e-01 -5.36377132e-01 -8.89741421e-01 3.71367961e-01
2.99588174e-01 -5.76894939e-01 9.52980459e-01 -5.80783367e-01
4.60478604e-01 -2.22083658e-01 3.35669607e-01 -1.47661304e+00
-1.70463920e-01 -6.31036878e-01 -5.87795377e-01 1.31327641e+00
1.83182925e-01 -6.14637315e-01 1.09011495e+00 6.76559210e-01
-4.15051915e-02 -1.09535933e+00 -8.88122141e-01 -8.52344394e-01
2.90029347e-01 -3.02868426e-01 6.42393470e-01 1.29506052e+00
1.03429548e-01 2.59768963e-01 -2.29727268e-01 2.96421319e-01
6.98012710e-01 2.83501655e-01 8.92018795e-01 -1.44234455e+00
-4.11072135e-01 -4.89839822e-01 -1.00767948e-02 -1.81889653e+00
2.55222708e-01 -8.63337100e-01 -2.29059651e-01 -1.04898846e+00
1.40819207e-01 -9.74725604e-01 -2.95120776e-01 6.49970293e-01
-5.52603126e-01 -2.79199213e-01 1.72145024e-01 1.25471503e-01
-3.77012044e-01 6.93592310e-01 1.89626765e+00 -4.07869853e-02
-3.89035016e-01 2.65461236e-01 -8.22684646e-01 6.97311819e-01
6.45804882e-01 -3.23613077e-01 -6.62076950e-01 -2.55835265e-01
2.18690094e-02 4.18135405e-01 1.13057271e-01 -5.92577577e-01
4.51009691e-01 -4.33286875e-01 4.99132395e-01 -3.91555995e-01
3.82654965e-01 -1.10336077e+00 -2.14464545e-01 4.21248466e-01
-1.99374199e-01 -4.79339242e-01 1.33619294e-01 3.54003251e-01
3.84433642e-02 -4.88102317e-01 8.40439737e-01 1.30800158e-01
-4.19824988e-01 5.54656744e-01 -1.15882650e-01 7.20616058e-02
1.27472055e+00 -4.13300872e-01 -2.63460398e-01 -3.28094035e-01
-8.73784840e-01 9.00327861e-01 1.91684887e-01 -9.77127906e-03
7.42796779e-01 -1.38522160e+00 -4.57758069e-01 2.94905305e-01
-3.18114925e-03 4.20796633e-01 2.67764837e-01 8.94676805e-01
-4.57660593e-02 2.93295443e-01 2.04303876e-01 -6.68217599e-01
-1.17818403e+00 7.58762538e-01 2.60171235e-01 -2.27460653e-01
-3.70811641e-01 9.94585514e-01 5.73282301e-01 -6.02574229e-01
7.21098781e-01 -2.50166655e-01 -1.53877243e-01 1.42053798e-01
5.83744287e-01 1.98671207e-01 -3.23439300e-01 -5.49221560e-02
-1.79259762e-01 3.77233118e-01 -4.00320709e-01 3.10454339e-01
1.14698994e+00 8.33635479e-02 1.06975369e-01 9.90224555e-02
9.27338302e-01 -4.72828597e-02 -1.22274315e+00 -5.37976146e-01
-2.01665193e-01 -3.53319764e-01 -5.28869731e-03 -8.54999006e-01
-1.09133399e+00 1.02967680e+00 2.19878629e-01 6.91922307e-02
1.27762532e+00 -2.37303630e-01 4.88850832e-01 7.08099902e-01
2.50257581e-01 -9.33264375e-01 4.59936857e-01 5.68745255e-01
5.33203363e-01 -1.21216464e+00 -2.09508806e-01 -4.33607310e-01
-5.00008941e-01 1.32433307e+00 1.30561078e+00 1.65717095e-01
4.70881224e-01 -1.23828381e-01 -6.18818551e-02 1.50040016e-01
-8.09130728e-01 1.49626210e-01 4.55541849e-01 3.36439371e-01
1.73864529e-01 -4.55886982e-02 -1.04073822e-01 7.42988586e-01
-1.48180470e-01 -1.64506450e-01 6.58856511e-01 6.95239007e-01
-7.86392391e-01 -1.19771206e+00 -5.85314274e-01 1.97554365e-01
-2.68904082e-02 4.52581346e-02 -2.13292688e-01 6.05229020e-01
1.93536162e-01 8.74597192e-01 -2.54158229e-01 -3.79768610e-01
2.16803983e-01 1.44919619e-01 5.19071758e-01 -5.68381608e-01
-4.26479965e-01 1.38956770e-01 -3.87430072e-01 -3.00972581e-01
1.22436680e-01 -1.05216920e-01 -1.37616718e+00 -3.08875740e-01
-6.68645263e-01 8.48612964e-01 2.54878551e-01 1.05472088e+00
1.39736086e-01 4.94489431e-01 1.13341129e+00 -2.07883924e-01
-1.00016856e+00 -7.49618351e-01 -4.62106913e-01 -8.44794065e-02
3.55496436e-01 -7.84129977e-01 -4.10257459e-01 -1.80939019e-01] | [9.530455589294434, 3.2976768016815186] |
33cc7f50-d343-4517-bd0b-94feacf48c54 | cnerv-content-adaptive-neural-representation | 2211.10421 | null | https://arxiv.org/abs/2211.10421v1 | https://arxiv.org/pdf/2211.10421v1.pdf | CNeRV: Content-adaptive Neural Representation for Visual Data | Compression and reconstruction of visual data have been widely studied in the computer vision community, even before the popularization of deep learning. More recently, some have used deep learning to improve or refine existing pipelines, while others have proposed end-to-end approaches, including autoencoders and implicit neural representations, such as SIREN and NeRV. In this work, we propose Neural Visual Representation with Content-adaptive Embedding (CNeRV), which combines the generalizability of autoencoders with the simplicity and compactness of implicit representation. We introduce a novel content-adaptive embedding that is unified, concise, and internally (within-video) generalizable, that compliments a powerful decoder with a single-layer encoder. We match the performance of NeRV, a state-of-the-art implicit neural representation, on the reconstruction task for frames seen during training while far surpassing for frames that are skipped during training (unseen images). To achieve similar reconstruction quality on unseen images, NeRV needs 120x more time to overfit per-frame due to its lack of internal generalization. With the same latent code length and similar model size, CNeRV outperforms autoencoders on reconstruction of both seen and unseen images. We also show promising results for visual data compression. More details can be found in the project pagehttps://haochen-rye.github.io/CNeRV/ | ['Abhinav Shrivastava', 'Ser-Nam Lim', 'Bo He', 'Matt Gwilliam', 'Hao Chen'] | 2022-11-18 | null | null | null | null | ['data-compression'] | ['time-series'] | [-1.94544151e-01 1.36239260e-01 -2.17365883e-02 -3.12886268e-01
-2.64005333e-01 -1.51276171e-01 5.45525312e-01 -2.64327943e-01
-5.21905243e-01 2.70187020e-01 6.73830986e-01 -1.20280765e-01
2.20619634e-01 -6.61337376e-01 -9.90546584e-01 -4.72072363e-01
1.85412645e-01 1.49993554e-01 8.84955749e-02 -1.93233620e-02
-1.59034282e-01 3.45226437e-01 -1.73372459e+00 5.91391385e-01
1.64834157e-01 8.90672445e-01 6.09062374e-01 8.82045269e-01
2.03070328e-01 1.00032377e+00 -2.95772672e-01 -5.67305267e-01
1.18957974e-01 -2.54249156e-01 -6.56819284e-01 -8.02679434e-02
6.42837226e-01 -9.91585255e-01 -1.14181578e+00 6.81026876e-01
5.45639277e-01 7.23926201e-02 5.32454729e-01 -9.33641016e-01
-1.19553638e+00 3.55949819e-01 -1.95787236e-01 2.73433596e-01
-7.73776919e-02 2.90534556e-01 9.72764015e-01 -1.17941737e+00
6.12242460e-01 1.24606752e+00 8.58565211e-01 8.17310810e-01
-1.18821156e+00 -5.32194912e-01 -3.12646367e-02 5.16213894e-01
-1.31593311e+00 -8.13311815e-01 4.49820846e-01 -1.76651031e-01
1.43378496e+00 5.29490747e-02 7.87228227e-01 1.36467016e+00
2.73171421e-02 9.92256641e-01 4.35130417e-01 -2.43725866e-01
1.72892585e-01 -1.09883197e-01 -2.31848717e-01 6.35572195e-01
2.32733905e-01 3.76656145e-01 -4.33424592e-01 3.19891751e-01
1.13976550e+00 4.68749583e-01 -4.29734707e-01 -2.20638320e-01
-1.09758008e+00 1.01462328e+00 6.22802973e-01 3.08210135e-01
-4.15113270e-01 6.30909383e-01 6.14603937e-01 2.86797017e-01
2.63348132e-01 -3.56426695e-03 -4.62680161e-01 -2.88369358e-01
-1.12173378e+00 9.09442380e-02 5.92632174e-01 1.07079637e+00
6.24842167e-01 5.73587775e-01 -8.49556178e-03 8.19292426e-01
3.28550577e-01 3.67793679e-01 8.64362776e-01 -1.28076613e+00
8.91666263e-02 1.40146077e-01 -3.69092554e-01 -1.03731334e+00
-1.37879536e-01 -6.15825176e-01 -1.23829234e+00 2.50014067e-01
-6.31933883e-02 1.16162211e-01 -1.00706816e+00 1.57391536e+00
-1.22066341e-01 3.85298938e-01 1.57651335e-01 1.07264483e+00
1.11688924e+00 8.35190356e-01 2.56945137e-02 1.47357523e-01
1.27101231e+00 -1.34954381e+00 -7.30514050e-01 -2.95840323e-01
4.22102928e-01 -5.13935864e-01 8.83712232e-01 4.17351723e-01
-1.30319965e+00 -9.52275932e-01 -1.10525823e+00 -6.93032980e-01
-6.94146454e-02 2.33200982e-01 5.08146942e-01 1.39097035e-01
-1.48337853e+00 7.17286944e-01 -1.12212026e+00 -3.04956794e-01
5.25835931e-01 2.07030997e-01 -5.93493640e-01 -2.37683862e-01
-8.84605646e-01 7.91112185e-01 6.43407226e-01 -1.42187104e-01
-1.43179154e+00 -7.24618196e-01 -1.03205407e+00 4.00975823e-01
8.03271607e-02 -1.01769900e+00 1.36341429e+00 -1.15435636e+00
-1.50732434e+00 6.23944640e-01 -1.23862021e-01 -9.36691403e-01
1.88309491e-01 -6.16076171e-01 -2.77244896e-01 5.00782549e-01
-2.37658203e-01 1.11547196e+00 1.06669199e+00 -1.03862667e+00
-1.62920415e-01 2.27610674e-02 -1.19504584e-02 1.39296018e-02
-5.53164899e-01 -2.38543019e-01 -7.57110357e-01 -9.79137719e-01
-1.16928473e-01 -6.63241565e-01 -1.19777963e-01 4.23279494e-01
2.38541216e-02 -2.14404278e-02 1.04202211e+00 -9.47656691e-01
1.15269375e+00 -2.47245002e+00 3.33375663e-01 -3.62939715e-01
4.62646902e-01 5.51121533e-01 -4.80159402e-01 5.20732641e-01
-3.46870542e-01 7.50232069e-03 -1.88947767e-01 -7.39733815e-01
-1.30199715e-01 5.74478149e-01 -3.89990270e-01 4.47885036e-01
2.25998238e-01 1.10726726e+00 -8.05474043e-01 -2.26871327e-01
4.24503714e-01 1.04790723e+00 -9.57833886e-01 4.74586964e-01
-3.63786481e-02 -2.24245656e-02 6.95862770e-02 3.71523380e-01
5.58914006e-01 -5.38423955e-01 3.53785977e-02 -5.90163350e-01
5.73999770e-02 3.02652329e-01 -8.57838035e-01 1.90098858e+00
-5.50260663e-01 1.03342354e+00 1.13406964e-01 -1.06723487e+00
6.10162616e-01 5.32535553e-01 2.16010362e-01 -7.98611283e-01
1.14084467e-01 7.03829303e-02 -2.46429607e-01 -5.35305858e-01
5.62277853e-01 -7.66897425e-02 5.23366034e-01 2.82307774e-01
6.70590162e-01 1.30782366e-01 7.07069039e-02 3.09616745e-01
1.12807572e+00 2.50020474e-01 3.10797423e-01 2.64799118e-01
1.31724551e-01 -3.30704331e-01 2.54952222e-01 5.52860498e-01
-5.23256883e-02 1.08503962e+00 3.34364139e-02 -6.30374968e-01
-1.60919237e+00 -1.03770375e+00 -1.28887817e-01 9.04648781e-01
-9.56934914e-02 -8.45294952e-01 -6.48985803e-01 -3.16754580e-01
-9.07631889e-02 6.85417533e-01 -6.34322226e-01 -3.75353962e-01
-5.43460250e-01 -1.53293401e-01 7.19936788e-01 7.38346815e-01
6.18264556e-01 -1.11673415e+00 -6.61682904e-01 2.48461291e-01
-1.57386839e-01 -1.13575602e+00 -1.92950219e-01 1.07363284e-01
-1.13932800e+00 -7.40465820e-01 -9.81128275e-01 -7.66914010e-01
4.86994505e-01 4.58604932e-01 1.36456203e+00 3.79151762e-01
-3.21719885e-01 5.29325664e-01 -5.87841570e-01 -4.84659970e-02
-3.92681479e-01 -2.11692706e-01 -5.08722402e-02 -2.61826724e-01
1.82715937e-01 -8.32668066e-01 -8.99773657e-01 5.21716550e-02
-1.21132123e+00 3.74095082e-01 8.71177733e-01 1.04947615e+00
6.57691181e-01 -2.57480174e-01 1.65280968e-01 -4.51204270e-01
1.89287245e-01 -6.30473495e-01 -1.38382196e-01 -1.58033729e-01
-4.80249256e-01 2.03888953e-01 7.19181538e-01 -3.94223213e-01
-7.66099155e-01 -9.40701067e-02 -6.35200381e-01 -1.20177519e+00
-1.78526849e-01 5.08668065e-01 2.89976925e-01 1.44469887e-01
6.42486572e-01 5.50296247e-01 2.35757679e-01 -7.29172826e-01
6.23786211e-01 5.11659503e-01 7.52565265e-01 -1.55984521e-01
6.39646113e-01 4.84587133e-01 -3.73208314e-01 -8.59022856e-01
-5.09397805e-01 -3.57263148e-01 -3.78175855e-01 8.87144431e-02
8.79556775e-01 -1.43943095e+00 -3.99871081e-01 1.95222706e-01
-1.32876205e+00 -4.15027440e-01 -5.36741793e-01 6.70807123e-01
-7.45083392e-01 5.79609692e-01 -1.02044296e+00 -3.34638476e-01
-5.50353050e-01 -1.03213227e+00 1.01003754e+00 -3.16238478e-02
-1.77481592e-01 -9.35839593e-01 1.10889986e-01 1.98066786e-01
6.49280787e-01 -8.62753242e-02 6.02489471e-01 -2.97859162e-01
-6.01789176e-01 -6.53181970e-02 -3.52433532e-01 8.56157660e-01
-2.28096947e-01 -4.99865785e-02 -1.13557792e+00 -6.61471665e-01
4.07704115e-02 -6.15537703e-01 1.33682930e+00 4.88262206e-01
1.51138866e+00 -6.24854147e-01 -3.50651890e-02 1.13761663e+00
1.57380068e+00 -3.18743587e-01 1.10365081e+00 4.55256522e-01
5.16895711e-01 1.90876856e-01 4.44274954e-03 5.64878762e-01
3.77055377e-01 5.01458764e-01 8.44178081e-01 -2.05716029e-01
-5.53475678e-01 -3.66398364e-01 6.41719282e-01 1.20566165e+00
-4.02335763e-01 -2.24084869e-01 -4.30457652e-01 6.50892794e-01
-1.67728007e+00 -1.22498393e+00 2.00471967e-01 1.88633800e+00
8.19861591e-01 -9.93077457e-02 -1.09171970e-02 7.44421482e-02
2.71128148e-01 2.92128026e-01 -4.38104689e-01 -4.54885423e-01
-1.78642794e-01 3.49047691e-01 3.45869422e-01 3.60792965e-01
-9.19993877e-01 7.85992622e-01 6.15944529e+00 8.33740354e-01
-1.11218131e+00 3.41508269e-01 5.44134736e-01 -3.00815761e-01
-3.08104366e-01 -2.94038467e-02 -5.94655156e-01 3.54253858e-01
1.36772108e+00 7.76153281e-02 5.70166111e-01 1.20120752e+00
6.57538027e-02 4.24547970e-01 -1.22965264e+00 1.31466341e+00
2.69054919e-01 -1.71986485e+00 3.42400461e-01 -7.44359717e-02
5.30370951e-01 4.65282530e-01 2.28608191e-01 5.78031182e-01
2.66201571e-02 -1.20886004e+00 8.90949070e-01 4.06905740e-01
1.00537503e+00 -5.10992885e-01 7.40063190e-01 2.77444661e-01
-1.10595977e+00 -1.97571427e-01 -9.42437768e-01 -5.20735048e-02
1.98775038e-01 4.68618423e-01 -5.22625804e-01 4.02243018e-01
1.06222379e+00 1.23296738e+00 -5.41794896e-01 9.51279461e-01
-1.87404528e-01 7.20717072e-01 -1.69687599e-01 4.83543068e-01
3.44801366e-01 5.02324849e-02 4.52563614e-01 1.47216988e+00
4.68922168e-01 1.38463452e-01 -2.73071885e-01 9.39898491e-01
-2.85049200e-01 -1.76108986e-01 -6.77964509e-01 -1.72782734e-01
2.35142156e-01 1.03979838e+00 -5.94072938e-02 -4.61238205e-01
-6.91568553e-01 1.27484834e+00 4.53551114e-01 5.32639980e-01
-8.90489459e-01 -1.01466484e-01 5.55616498e-01 2.01910138e-01
7.83184111e-01 -3.30336213e-01 1.10542588e-01 -1.28265536e+00
-9.33137760e-02 -9.19484913e-01 1.58007354e-01 -1.04630768e+00
-9.95111227e-01 7.62101173e-01 -1.41794153e-03 -1.38190258e+00
-3.38892579e-01 -7.83992827e-01 -4.31439221e-01 5.14995039e-01
-1.49278343e+00 -1.07343912e+00 -4.56496447e-01 6.44759238e-01
9.29209888e-01 -3.16013664e-01 1.00149298e+00 5.75524986e-01
-4.01715100e-01 6.85868084e-01 2.12748811e-01 3.29347879e-01
5.59955060e-01 -8.62665355e-01 4.56467837e-01 7.71070063e-01
3.01123440e-01 7.61230230e-01 5.61587334e-01 -3.56469333e-01
-1.57802236e+00 -1.31417441e+00 5.92489839e-01 -8.58622044e-02
2.85087734e-01 -1.64537802e-01 -9.68672931e-01 9.90353227e-01
5.61459303e-01 2.85880685e-01 5.79936981e-01 -1.75833240e-01
-6.46665156e-01 8.97218287e-03 -7.57438123e-01 6.02775991e-01
9.79948878e-01 -6.65620267e-01 -6.28760517e-01 1.09210491e-01
1.05421472e+00 -2.87507474e-01 -8.46060216e-01 3.32132459e-01
6.73498273e-01 -1.32890439e+00 1.28390133e+00 -4.12707925e-01
9.31685865e-01 -1.01347059e-01 -3.49596739e-01 -1.01355577e+00
-8.33861053e-01 -3.21666449e-01 -7.51101911e-01 8.00726593e-01
8.46039727e-02 -1.77693799e-01 7.94991195e-01 1.59757659e-01
-5.38173974e-01 -9.55978513e-01 -7.96313286e-01 -8.19372296e-01
-1.22252561e-01 -4.44787085e-01 2.02851355e-01 6.24940157e-01
-4.43689078e-01 3.03500503e-01 -7.60272980e-01 6.80565415e-03
3.73607188e-01 -3.13278437e-01 7.26634085e-01 -7.46363401e-01
-7.77588785e-01 -2.23756433e-01 -6.32944584e-01 -1.61313891e+00
-7.03515932e-02 -1.00567818e+00 -2.42825538e-01 -1.63350916e+00
2.75890112e-01 3.70616503e-02 -2.40331188e-01 7.14589059e-01
2.06678659e-01 5.08376241e-01 4.27569628e-01 4.14343208e-01
-5.79395056e-01 1.03796983e+00 1.13885701e+00 -2.22506359e-01
2.04513237e-01 -5.29670894e-01 -5.24587989e-01 6.04198754e-01
6.22228026e-01 -3.77942383e-01 -5.15217841e-01 -1.01863110e+00
-5.13135595e-03 -7.02887401e-02 6.43821478e-01 -1.27633917e+00
6.71995208e-02 3.97778809e-01 7.60101438e-01 -6.22603834e-01
5.87311864e-01 -9.55044091e-01 4.57860976e-01 5.99997342e-01
-2.51290441e-01 3.35946470e-01 4.19793308e-01 7.28516996e-01
-5.01795113e-01 -3.17027867e-01 7.35982955e-01 -2.10087404e-01
-9.55453753e-01 4.87431884e-01 -3.74155492e-01 -1.59991175e-01
6.82550132e-01 -4.21815366e-01 -3.74270141e-01 -6.34236693e-01
-6.64422274e-01 -2.38649219e-01 5.62672853e-01 4.32432950e-01
1.35587168e+00 -1.25828099e+00 -8.75740469e-01 3.09628010e-01
-1.27856150e-01 9.97664686e-03 5.17343819e-01 7.55322337e-01
-8.64891768e-01 3.60123575e-01 -4.26630437e-01 -6.14727318e-01
-1.04519272e+00 7.68014073e-01 2.04426616e-01 -4.84296679e-02
-1.13156104e+00 8.86938691e-01 5.09676635e-01 -8.38186499e-03
3.35466504e-01 -2.93354154e-01 -1.02352038e-01 -2.80634224e-01
8.17803025e-01 2.15997040e-01 -6.61567971e-02 -4.99861062e-01
2.25150697e-02 3.86524826e-01 -2.17748791e-01 1.62021950e-01
1.67891300e+00 -7.07172900e-02 2.93272943e-03 1.73853159e-01
1.55941236e+00 -4.10072327e-01 -1.59843278e+00 -1.90352231e-01
-6.00107491e-01 -4.04004395e-01 3.42560709e-01 -4.32245612e-01
-1.42382169e+00 1.16181123e+00 7.88778484e-01 -1.78480208e-01
1.15518320e+00 -1.11588433e-01 1.00314522e+00 4.23145741e-01
-5.32995723e-03 -7.84244716e-01 4.41259265e-01 6.36749566e-01
1.24844623e+00 -1.13663220e+00 1.38088852e-01 4.23013838e-03
-6.48162007e-01 1.33880472e+00 3.92339259e-01 -3.68902892e-01
5.76903999e-01 2.37519249e-01 -2.43834242e-01 -1.26397133e-01
-1.07722425e+00 -7.37660304e-02 1.85363606e-01 6.25138104e-01
4.68997568e-01 -2.03242108e-01 2.32164428e-01 4.25147802e-01
-1.23620890e-01 1.92975387e-01 3.30216378e-01 6.61637604e-01
-3.35680336e-01 -8.43891144e-01 7.64679676e-03 3.33396643e-01
-3.50684732e-01 -4.68984425e-01 1.32850617e-01 7.80908167e-01
1.31479666e-01 6.33940697e-01 2.82472432e-01 -4.95719224e-01
1.23232819e-01 -3.31042141e-01 3.90800893e-01 -4.71720725e-01
-3.82409006e-01 1.12490781e-01 -2.60455050e-02 -8.10642362e-01
-4.15613353e-01 -3.08497220e-01 -1.15025449e+00 -5.59063077e-01
4.68177199e-02 -1.14644952e-01 5.54606438e-01 5.86874604e-01
6.33856714e-01 5.79027712e-01 1.91139057e-01 -1.34756720e+00
-5.32308102e-01 -9.41437602e-01 -3.23125333e-01 4.88211155e-01
5.61585724e-01 -5.01279294e-01 -3.54840130e-01 4.98754323e-01] | [11.174076080322266, -1.4050374031066895] |
742ad067-1b49-4b25-8bb8-7035b711e346 | a-conditional-point-diffusion-refinement-1 | 2112.03530 | null | https://arxiv.org/abs/2112.03530v4 | https://arxiv.org/pdf/2112.03530v4.pdf | A Conditional Point Diffusion-Refinement Paradigm for 3D Point Cloud Completion | 3D point cloud is an important 3D representation for capturing real world 3D objects. However, real-scanned 3D point clouds are often incomplete, and it is important to recover complete point clouds for downstream applications. Most existing point cloud completion methods use Chamfer Distance (CD) loss for training. The CD loss estimates correspondences between two point clouds by searching nearest neighbors, which does not capture the overall point density distribution on the generated shape, and therefore likely leads to non-uniform point cloud generation. To tackle this problem, we propose a novel Point Diffusion-Refinement (PDR) paradigm for point cloud completion. PDR consists of a Conditional Generation Network (CGNet) and a ReFinement Network (RFNet). The CGNet uses a conditional generative model called the denoising diffusion probabilistic model (DDPM) to generate a coarse completion conditioned on the partial observation. DDPM establishes a one-to-one pointwise mapping between the generated point cloud and the uniform ground truth, and then optimizes the mean squared error loss to realize uniform generation. The RFNet refines the coarse output of the CGNet and further improves quality of the completed point cloud. Furthermore, we develop a novel dual-path architecture for both networks. The architecture can (1) effectively and efficiently extract multi-level features from partially observed point clouds to guide completion, and (2) accurately manipulate spatial locations of 3D points to obtain smooth surfaces and sharp details. Extensive experimental results on various benchmark datasets show that our PDR paradigm outperforms previous state-of-the-art methods for point cloud completion. Remarkably, with the help of the RFNet, we can accelerate the iterative generation process of the DDPM by up to 50 times without much performance drop. | ['Dahua Lin', 'Liang Pan', 'Xudong Xu', 'Zhifeng Kong', 'Zhaoyang Lyu'] | 2021-12-07 | a-conditional-point-diffusion-refinement | https://openreview.net/forum?id=wqD6TfbYkrn | https://openreview.net/pdf?id=wqD6TfbYkrn | iclr-2022-4 | ['point-cloud-completion', 'point-cloud-generation'] | ['computer-vision', 'computer-vision'] | [-4.02547084e-02 2.36184597e-02 1.70027003e-01 -9.22494754e-02
-1.00188696e+00 -4.15882021e-01 6.12301290e-01 3.15090492e-02
-2.10958105e-02 2.54786789e-01 -2.31529474e-01 -1.48298264e-01
-6.03079014e-02 -1.18493843e+00 -1.09229779e+00 -6.01071000e-01
1.47327393e-01 9.80136693e-01 3.43039304e-01 -1.61123332e-02
2.37980917e-01 1.04119658e+00 -1.33355308e+00 -9.85798389e-02
1.10180354e+00 1.01766419e+00 6.56628728e-01 2.88336277e-01
-2.57488251e-01 -3.97908092e-02 -2.94370532e-01 -1.48441255e-01
5.07120848e-01 2.19528466e-01 -3.85880440e-01 2.02532992e-01
2.63451338e-01 -4.15147692e-01 -2.92682499e-01 1.02282727e+00
1.93528786e-01 1.17552385e-01 8.08211088e-01 -1.11192501e+00
-6.06312037e-01 2.96400953e-02 -8.79513979e-01 -5.47308266e-01
2.23426864e-01 2.16896906e-01 7.36959159e-01 -1.39017963e+00
5.64447403e-01 1.35511589e+00 8.29351068e-01 2.77149051e-01
-1.19037199e+00 -9.30001736e-01 2.18532264e-01 -4.51647580e-01
-1.82148063e+00 1.52573176e-02 1.02083015e+00 -3.92781615e-01
5.77005386e-01 3.77756320e-02 7.52264738e-01 5.67229390e-01
3.66465189e-02 6.36344731e-01 6.14261568e-01 -8.18412006e-02
2.31583074e-01 -2.23132029e-01 -3.68067890e-01 5.42552352e-01
2.31148034e-01 1.16943032e-01 -3.39752078e-01 -4.93229359e-01
1.42572665e+00 5.38716435e-01 -4.27938879e-01 -5.55533111e-01
-1.17620707e+00 6.90795004e-01 9.17897701e-01 1.66585166e-02
-7.22040176e-01 2.30160862e-01 -4.10159945e-01 -1.30489245e-01
7.51356483e-01 1.77156538e-01 -1.78022325e-01 2.43649647e-01
-1.04464412e+00 5.51397622e-01 4.96074021e-01 1.32612610e+00
1.30518949e+00 -3.06694239e-01 3.69991548e-02 6.50326967e-01
7.22707689e-01 9.57568944e-01 -2.85790473e-01 -1.09444237e+00
6.42702460e-01 9.03466046e-01 2.15676025e-01 -1.10794747e+00
1.71192009e-02 -3.92006934e-01 -1.08986998e+00 4.24300790e-01
-3.48507514e-04 2.61177272e-01 -1.08194602e+00 1.41988325e+00
6.78206682e-01 6.07737243e-01 -3.23990464e-01 9.00377214e-01
6.01859570e-01 8.70615542e-01 -3.26197863e-01 1.52195841e-01
8.96789551e-01 -3.90942186e-01 -8.64982754e-02 -6.31012097e-02
2.18796909e-01 -7.79474914e-01 8.14611375e-01 1.85359463e-01
-1.35734272e+00 -4.14850354e-01 -8.76681507e-01 -1.68129951e-01
2.84383178e-01 -1.18361972e-01 3.59658450e-01 -6.26966357e-02
-9.89469707e-01 8.11479807e-01 -1.13567972e+00 2.54922450e-01
8.55323553e-01 3.21819693e-01 -2.70492017e-01 -5.71538270e-01
-5.81504345e-01 2.82975018e-01 -5.61437905e-02 1.58440158e-01
-9.64667499e-01 -1.29981732e+00 -8.32732975e-01 1.05315499e-01
1.44302398e-01 -1.19340825e+00 9.36206758e-01 -3.34371924e-01
-1.24906814e+00 6.85366035e-01 -3.15510511e-01 -2.69363681e-03
6.18582428e-01 -1.27800316e-01 8.56872350e-02 2.10774168e-01
3.93336535e-01 8.32558632e-01 9.08313394e-01 -1.82993448e+00
-6.40477896e-01 -5.30869663e-01 -3.43664765e-01 2.56411374e-01
3.49247009e-01 -6.26800537e-01 -7.69224405e-01 -5.83244026e-01
7.61393070e-01 -8.28483284e-01 -4.92374003e-01 4.24710274e-01
-5.17522633e-01 -3.29122275e-01 8.74628842e-01 -3.05186987e-01
6.18579149e-01 -2.30230093e+00 6.17095083e-02 5.92953086e-01
5.00508904e-01 -1.08532645e-01 -1.55585915e-01 2.77617097e-01
5.35261147e-02 1.22054748e-01 -6.75873756e-01 -8.10261846e-01
-1.78759247e-02 2.13861167e-01 -5.54960072e-01 5.49889922e-01
3.08202147e-01 9.11578238e-01 -9.86422122e-01 -1.09124586e-01
3.77023816e-01 8.36896181e-01 -6.57682240e-01 2.12799296e-01
-4.72542822e-01 5.75384140e-01 -7.76250482e-01 6.75183356e-01
1.27554047e+00 -5.24463356e-01 -4.68971580e-01 -2.62972355e-01
-1.46897480e-01 1.10992692e-01 -1.26941133e+00 2.16686583e+00
-4.42338675e-01 9.60500762e-02 1.92813277e-01 -3.77000690e-01
1.23344934e+00 1.96552351e-01 6.93632126e-01 -1.78633228e-01
-6.60520196e-02 3.61541063e-01 -6.20157182e-01 1.63500890e-01
5.19402206e-01 -2.66316026e-01 2.12831095e-01 2.29750887e-01
-3.32008064e-01 -8.07785928e-01 -3.99678111e-01 2.76430696e-01
9.79432046e-01 1.84333086e-01 -4.32301611e-01 5.92277236e-02
3.88680905e-01 1.85888246e-01 7.76977360e-01 5.41361451e-01
4.74555910e-01 1.31772029e+00 1.30030409e-01 -3.06835592e-01
-1.13747120e+00 -1.48895144e+00 -6.05636276e-02 -6.07346967e-02
6.22878551e-01 -1.91454679e-01 -4.80169952e-01 -4.97900456e-01
2.71861374e-01 7.37810612e-01 -1.49577856e-01 -4.46355715e-02
-6.01127446e-01 -3.75758767e-01 6.38491213e-02 3.94360721e-01
4.45513159e-01 -6.88200772e-01 -8.35273713e-02 2.03986794e-01
-1.81281760e-01 -9.51032519e-01 -5.58551311e-01 -3.23826790e-01
-1.35975432e+00 -1.01536155e+00 -8.17325056e-01 -6.87646687e-01
1.09418571e+00 6.52215540e-01 1.20557606e+00 2.64998794e-01
1.28195778e-01 2.54756272e-01 -2.49445379e-01 -3.29931229e-01
-3.12438518e-01 -2.71264724e-02 -3.99807654e-02 1.76282506e-02
2.37174690e-01 -9.97845232e-01 -8.18381310e-01 4.38632727e-01
-1.04352784e+00 2.11457402e-01 7.23798037e-01 5.52762270e-01
1.21855748e+00 2.60359585e-01 9.03772339e-02 -5.42578340e-01
2.13481799e-01 -5.06431222e-01 -8.62079501e-01 -1.95572481e-01
-4.11024123e-01 -1.05798706e-01 1.77731961e-01 -2.01645926e-01
-7.93234825e-01 3.12571615e-01 -4.01576370e-01 -1.25553417e+00
-1.29545787e-02 2.61766732e-01 -2.68565923e-01 -2.03263611e-01
3.33726466e-01 3.27310383e-01 6.58974722e-02 -6.91128016e-01
3.53896767e-01 3.81287307e-01 5.09239376e-01 -5.73687375e-01
1.35918725e+00 1.07070708e+00 2.88342029e-01 -7.42623091e-01
-5.74562252e-01 -5.84597051e-01 -7.16641665e-01 -1.02376245e-01
7.68738031e-01 -1.11633253e+00 -6.40360117e-01 4.12715197e-01
-1.52292800e+00 -3.27345401e-01 -4.53649133e-01 4.73271072e-01
-5.36725223e-01 3.17027718e-01 -3.82052958e-01 -6.19450629e-01
-3.49309921e-01 -1.02727318e+00 1.75397968e+00 3.85160372e-02
1.96035832e-01 -7.11975992e-01 -3.99685791e-03 1.18060596e-01
-1.47385309e-02 2.84541607e-01 7.92730272e-01 2.29106456e-01
-1.44049156e+00 -3.62284750e-01 -3.42245638e-01 2.87198007e-01
9.18847844e-02 -5.90747781e-02 -7.54043698e-01 -3.24140877e-01
2.06334084e-01 1.77862570e-01 7.85502911e-01 6.36372685e-01
1.16557372e+00 1.07597122e-02 -5.84475815e-01 9.32430506e-01
1.46209121e+00 -2.99342424e-01 8.41443002e-01 -1.16155036e-01
1.05114424e+00 3.35387141e-01 6.44810915e-01 4.85998005e-01
5.13202250e-01 4.67863083e-01 8.21682930e-01 -1.07761122e-01
-2.07121015e-01 -8.00517201e-01 1.82522368e-02 8.07312727e-01
-1.80197567e-01 -6.35220716e-03 -9.55155671e-01 5.16407609e-01
-1.72429955e+00 -6.47270799e-01 -4.43616658e-01 2.43089056e+00
4.80580986e-01 -7.74171809e-03 -3.05485606e-01 -1.03860751e-01
7.83744991e-01 3.67778689e-02 -5.77765942e-01 5.81719100e-01
1.17425144e-01 2.64570564e-01 4.83655393e-01 6.14669800e-01
-6.61258638e-01 8.10221553e-01 4.78686857e+00 1.04783785e+00
-9.60896313e-01 5.62228672e-02 3.92154932e-01 3.63314040e-02
-6.70515776e-01 1.55861244e-01 -8.71825278e-01 4.67015803e-01
1.94377691e-01 -5.54111898e-02 2.26062387e-01 9.09104824e-01
2.28794590e-01 1.50042266e-01 -9.25281465e-01 1.25906384e+00
-2.07216397e-01 -1.64373255e+00 3.61713707e-01 3.96811157e-01
9.10368025e-01 4.35933679e-01 -2.45685447e-02 -9.15195793e-02
4.79615122e-01 -8.45054865e-01 8.34337533e-01 8.02457452e-01
8.99614573e-01 -9.43278670e-01 4.40802217e-01 6.82590842e-01
-1.37487197e+00 3.24887127e-01 -7.49524057e-01 2.45200872e-01
5.16912937e-01 1.18994033e+00 -7.34961331e-01 8.44386458e-01
7.49345481e-01 8.98320794e-01 -1.18652128e-01 1.27296054e+00
-4.02028501e-01 1.98346764e-01 -7.40170181e-01 3.63322258e-01
2.22575322e-01 -5.83134055e-01 1.04149961e+00 4.60690588e-01
7.30560839e-01 2.41495445e-01 2.64189065e-01 1.51821578e+00
-2.48880804e-01 -2.21340299e-01 -5.17989218e-01 3.85072201e-01
8.27782094e-01 1.17983449e+00 -7.50790000e-01 -7.71644339e-02
-1.48447484e-01 7.69480884e-01 2.95777261e-01 3.29988003e-01
-3.76460105e-01 -1.42828465e-01 8.57045472e-01 4.95488435e-01
3.61388803e-01 -5.65714538e-01 -4.65156227e-01 -1.03062916e+00
2.49456510e-01 -2.67248333e-01 -7.99636021e-02 -1.03792167e+00
-1.61078310e+00 4.73520339e-01 -9.24379155e-02 -1.58992255e+00
1.35424078e-01 -1.93138599e-01 -7.25500703e-01 1.30410409e+00
-1.67586231e+00 -1.19937241e+00 -7.14878082e-01 5.18273294e-01
4.54827547e-01 3.04922193e-01 2.84904242e-01 2.46988803e-01
-8.01531449e-02 -7.53446072e-02 -7.25147575e-02 -8.84637013e-02
3.56373578e-01 -1.09449577e+00 7.88024843e-01 6.38903916e-01
8.66287854e-03 6.22663915e-01 1.73523083e-01 -1.12608540e+00
-1.42818844e+00 -1.60041189e+00 5.81440032e-01 -6.43930316e-01
1.04766332e-01 -2.93858171e-01 -1.25103927e+00 5.14777660e-01
-5.51132500e-01 5.50242439e-02 -2.26044049e-03 -3.46507847e-01
-1.08750567e-01 -4.52786759e-02 -1.18940258e+00 4.91314352e-01
1.10424817e+00 -2.80784398e-01 -4.56177980e-01 4.68894988e-01
1.01466548e+00 -7.13849783e-01 -8.47209096e-01 5.81771314e-01
7.51290843e-02 -8.10389876e-01 1.25745153e+00 9.62535292e-02
5.86079061e-01 -6.54737175e-01 -7.55334571e-02 -1.34093070e+00
-4.25422311e-01 -6.87515676e-01 -9.79266241e-02 1.12220395e+00
2.47541592e-01 -4.74869788e-01 1.14972448e+00 5.27961254e-01
-6.16759121e-01 -8.59034896e-01 -9.12157714e-01 -7.92656600e-01
2.38975853e-01 -7.35965371e-01 1.20694470e+00 7.69660473e-01
-8.43951881e-01 1.05469316e-01 2.09563419e-01 7.45930493e-01
1.02380013e+00 3.30490291e-01 1.02886438e+00 -1.46632707e+00
2.47904006e-02 -4.71400380e-01 -3.29310417e-01 -1.71054935e+00
-3.26136388e-02 -9.91568863e-01 1.61388904e-01 -1.82027316e+00
-1.94703668e-01 -1.19296634e+00 2.66615510e-01 3.63977961e-02
-1.96144432e-01 2.10146651e-01 7.73416609e-02 6.66683793e-01
-1.30942583e-01 1.01264572e+00 1.50399864e+00 3.19629833e-02
-4.81415063e-01 3.47502410e-01 -5.11585295e-01 8.94402981e-01
4.17444646e-01 -6.72730803e-01 -3.04559380e-01 -7.79092729e-01
2.71182716e-01 -1.64236445e-02 6.60005689e-01 -1.00333297e+00
4.51832652e-01 -2.14936003e-01 5.02304375e-01 -1.33150196e+00
6.03000224e-01 -1.01217890e+00 3.88402909e-01 1.41938627e-01
3.24618161e-01 -1.86304361e-01 -7.48933628e-02 8.50306153e-01
-5.69117293e-02 -1.80850159e-02 7.84011006e-01 -1.70100778e-01
-1.41210303e-01 1.17168248e+00 3.90256435e-01 -1.70230806e-01
8.48531365e-01 -2.60037005e-01 1.48330644e-01 -1.70943499e-01
-3.66603881e-01 5.00163019e-01 1.01548707e+00 2.57055253e-01
1.08328915e+00 -1.50557804e+00 -7.94919968e-01 4.67372268e-01
7.12036043e-02 1.39545941e+00 2.84195721e-01 5.09556055e-01
-7.50064313e-01 6.47410899e-02 4.41894352e-01 -1.13926327e+00
-6.96611941e-01 2.74432003e-01 2.71310538e-01 1.51999006e-02
-1.16805220e+00 9.02281821e-01 5.39364219e-01 -7.30053425e-01
-1.13507599e-01 -4.45695788e-01 3.92107636e-01 -4.94940221e-01
3.75508815e-01 3.61768395e-01 9.86275300e-02 -5.28949082e-01
-8.12965482e-02 8.99208844e-01 2.98206061e-02 -1.46703720e-01
1.58129954e+00 2.45342916e-03 -9.88189057e-02 1.59814447e-01
1.02759635e+00 -7.36360699e-02 -1.70405948e+00 -3.92328590e-01
-4.44517404e-01 -8.86642575e-01 3.44221234e-01 -1.82140335e-01
-1.21347368e+00 8.15611720e-01 8.35300907e-02 8.69160332e-03
8.45188916e-01 2.33226672e-01 1.10649788e+00 7.84516260e-02
6.83783770e-01 -5.59641361e-01 -8.77594873e-02 3.62626523e-01
1.08499038e+00 -9.00088072e-01 1.64109901e-01 -8.77603650e-01
-2.58715570e-01 9.19957101e-01 3.90737981e-01 -5.08896470e-01
8.80835474e-01 1.51007306e-02 -3.09928060e-01 -6.05201066e-01
-4.06115592e-01 1.11320168e-01 3.02647650e-01 7.18692541e-01
-3.20600241e-01 -6.70947433e-02 3.87704819e-01 4.52848434e-01
-2.58854538e-01 9.53154191e-02 6.67253435e-02 8.14537644e-01
-3.94966632e-01 -1.03254700e+00 -6.85227871e-01 4.74101871e-01
2.17815056e-01 2.46537011e-02 -6.27948940e-02 5.86101055e-01
-3.25556565e-03 5.52014291e-01 4.45894718e-01 -3.08948129e-01
5.01013935e-01 -4.72272336e-01 3.11209351e-01 -8.17067385e-01
-4.73072603e-02 1.81560650e-01 -7.00902462e-01 -6.28624916e-01
-1.00934401e-01 -6.60382211e-01 -1.39854705e+00 -4.35122639e-01
-3.49887431e-01 1.80589631e-01 8.32420647e-01 6.88493192e-01
7.60798633e-01 2.19001099e-01 9.17751014e-01 -1.50449276e+00
-4.23461437e-01 -6.91689968e-01 -6.65734887e-01 4.01826650e-01
1.98605046e-01 -7.58539796e-01 -5.29234827e-01 -1.41471729e-01] | [8.406919479370117, -3.5717623233795166] |
dc12ad10-8575-4380-8b61-1d7400c89cdc | estimating-risk-and-uncertainty-in-deep | 1905.09638 | null | https://arxiv.org/abs/1905.09638v5 | https://arxiv.org/pdf/1905.09638v5.pdf | Estimating Risk and Uncertainty in Deep Reinforcement Learning | Reinforcement learning agents are faced with two types of uncertainty. Epistemic uncertainty stems from limited data and is useful for exploration, whereas aleatoric uncertainty arises from stochastic environments and must be accounted for in risk-sensitive applications. We highlight the challenges involved in simultaneously estimating both of them, and propose a framework for disentangling and estimating these uncertainties on learned Q-values. We derive unbiased estimators of these uncertainties and introduce an uncertainty-aware DQN algorithm, which we show exhibits safe learning behavior and outperforms other DQN variants on the MinAtar testbed. | ['Sébastien Toth', 'Benoît-Marie Robaglia', 'Reda Bahi Slaoui', 'Bastien Van Delft', 'William R. Clements'] | 2019-05-23 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-2.40641832e-01 4.23132807e-01 -2.01619357e-01 -3.68162841e-01
-1.26347864e+00 -7.70888865e-01 6.11822963e-01 2.29843616e-01
-9.59707379e-01 1.67647541e+00 1.83303014e-03 -3.06419224e-01
-7.92602837e-01 -8.44701469e-01 -8.01370621e-01 -7.61142135e-01
-7.75594175e-01 7.45854378e-01 1.71340361e-01 8.35398287e-02
1.32111475e-01 2.66511887e-01 -1.45836794e+00 -5.89940131e-01
1.05015516e+00 1.19184434e+00 -2.28274316e-01 6.31288826e-01
3.49041522e-01 9.70308423e-01 -9.40350950e-01 -3.14150184e-01
4.35438335e-01 -2.42971972e-01 -3.98564011e-01 -5.33220410e-01
-1.53494865e-01 -8.96876931e-01 -2.36339122e-01 1.26016736e+00
5.01478672e-01 3.37636560e-01 8.67224395e-01 -1.65249693e+00
1.00374185e-01 1.24669802e+00 -3.70194018e-01 2.78578252e-01
2.25776047e-01 2.27362588e-01 8.70267749e-01 -2.66828504e-03
1.98795334e-01 1.39672971e+00 3.97202551e-01 4.99854773e-01
-1.09096289e+00 -7.68227160e-01 3.46502632e-01 1.33093432e-01
-1.14450932e+00 -1.64326146e-01 3.32551539e-01 -2.56727248e-01
5.67431927e-01 -4.65628579e-02 6.67563140e-01 1.39274442e+00
4.70781803e-01 7.77319610e-01 1.40093708e+00 -1.20006680e-01
1.04103303e+00 -3.48843518e-03 -3.62000763e-01 2.53888547e-01
8.17702889e-01 1.14109087e+00 -7.37451553e-01 -3.83131206e-01
7.41389453e-01 -5.85345387e-01 6.18032820e-04 -6.58518791e-01
-8.85666668e-01 8.70771110e-01 -1.54535603e-02 -5.67814708e-01
-3.89254183e-01 8.88789654e-01 2.76127309e-01 3.96753639e-01
-2.24168729e-02 5.03181875e-01 -3.80141407e-01 -6.55549347e-01
-4.04491127e-01 7.32919395e-01 8.91575277e-01 1.04861510e+00
3.12291235e-01 4.02445376e-01 -1.14413127e-01 1.79720536e-01
6.42368257e-01 7.82768130e-01 8.13331455e-02 -1.38271320e+00
5.56174278e-01 -3.77113193e-01 9.85580981e-01 -2.13578731e-01
-3.70303422e-01 -3.03515136e-01 -1.17705436e-02 7.68746376e-01
6.96193337e-01 -9.36504662e-01 -8.45066547e-01 2.14962697e+00
3.65676194e-01 2.46127218e-01 4.59532320e-01 6.83350146e-01
1.38847813e-01 2.81570494e-01 3.20106715e-01 -2.68219620e-01
9.88583148e-01 -3.07150036e-01 -8.91812980e-01 -3.27510893e-01
5.50371818e-02 1.10236563e-01 4.01109040e-01 6.71615183e-01
-1.13995624e+00 3.10629457e-01 -1.36926675e+00 6.34676814e-01
-1.29825890e-01 -7.84845471e-01 7.36901283e-01 1.05616105e+00
-6.03777885e-01 7.49938965e-01 -1.02885175e+00 1.41490474e-01
5.58286071e-01 2.14391440e-01 1.10635005e-01 4.05959278e-01
-1.71717536e+00 1.36655939e+00 6.04446352e-01 9.05067995e-02
-1.85873449e+00 -6.45990193e-01 -8.17809701e-01 -5.51818870e-02
9.11965787e-01 -1.44119218e-01 1.72993684e+00 -2.21671179e-01
-1.88528442e+00 -1.34102941e-01 7.51879394e-01 -7.84857929e-01
1.08758461e+00 -3.72134894e-01 -2.72760630e-01 1.44006863e-01
-1.15757667e-01 3.97435814e-01 7.65954673e-01 -1.19935691e+00
-7.26138651e-01 -1.50262311e-01 3.59515995e-01 5.83436966e-01
2.82822073e-01 -2.55444914e-01 5.50111353e-01 -3.68281901e-01
-3.18176121e-01 -8.22516739e-01 -6.10656738e-01 6.31748373e-03
-3.69527340e-01 2.75241882e-01 3.23290527e-01 -2.00141743e-01
7.21857727e-01 -1.46876276e+00 -3.84586789e-02 5.81347048e-01
-6.56431839e-02 -4.08338100e-01 1.59878358e-01 4.14924294e-01
4.03630793e-01 1.19604297e-01 -2.94936240e-01 1.00864276e-01
6.31790698e-01 4.38778669e-01 -4.71552759e-01 4.44073230e-01
6.74785599e-02 5.83433688e-01 -1.24301398e+00 -4.08257902e-01
5.35074472e-02 9.24784467e-02 -1.69816598e-01 5.00477254e-01
-4.79564369e-01 2.80172825e-01 -9.23134148e-01 5.69939315e-01
4.50036317e-01 5.35539269e-01 4.19109166e-01 4.48901445e-01
-3.58634293e-02 1.36495158e-01 -1.46635985e+00 1.32013679e+00
-3.52377445e-01 8.65039080e-02 1.85020804e-01 -5.73471010e-01
4.83618706e-01 2.96968013e-01 2.39466220e-01 -4.19779778e-01
4.15281832e-01 2.70629525e-01 -1.04439810e-01 -2.14536637e-01
4.04468477e-01 -4.42633301e-01 -5.41547477e-01 7.75971949e-01
2.54867166e-01 -7.24208832e-01 -1.40503515e-02 1.10990994e-01
1.08217573e+00 4.10550386e-01 4.50674802e-01 -6.45680785e-01
-1.52877763e-01 -3.08817387e-01 9.01642501e-01 1.19879985e+00
-7.32173502e-01 -1.38010770e-01 1.07770526e+00 -1.23281851e-01
-4.06739771e-01 -1.62701035e+00 -2.78275043e-01 5.38076401e-01
3.76920134e-01 5.73255159e-02 -6.38348162e-01 -9.89847720e-01
5.14831543e-01 1.33755589e+00 -8.62812400e-01 -4.13861811e-01
9.46575329e-02 -7.43210495e-01 4.81860250e-01 7.96445191e-01
1.30572036e-01 -5.35348773e-01 -1.17508864e+00 2.13613346e-01
1.80707693e-01 -5.85408092e-01 -1.14137255e-01 8.90260279e-01
-6.13800585e-01 -1.18693745e+00 -2.84729302e-01 3.85052711e-01
2.37704113e-01 -3.29538554e-01 1.04178774e+00 -6.53450429e-01
1.48790348e-02 7.65121162e-01 3.47164497e-02 -1.01092112e+00
-3.05793047e-01 -5.79467118e-01 4.98764843e-01 -6.46866262e-01
1.58126697e-01 -5.96873820e-01 -4.43042338e-01 3.55435878e-01
-7.29687095e-01 -6.81717575e-01 2.00582370e-01 8.27652395e-01
3.66565913e-01 5.15231013e-01 9.17387009e-01 -7.71509111e-01
9.55575645e-01 -6.88245893e-01 -1.39979541e+00 3.58396828e-01
-7.27223933e-01 5.83059728e-01 1.35097235e-01 -4.40014869e-01
-1.51382744e+00 -1.74189672e-01 3.19863111e-01 -1.37462988e-01
5.60400225e-02 2.45364338e-01 -3.40812206e-01 -2.36951232e-01
6.58694923e-01 -5.15701234e-01 -5.04502915e-02 -3.25364433e-02
4.40668970e-01 3.50648105e-01 2.71757036e-01 -1.41083062e+00
7.87118971e-01 1.05408423e-01 2.40341499e-01 -2.36806601e-01
-7.43351758e-01 6.29687190e-01 -1.12242512e-01 -4.76747215e-01
4.62890565e-01 -9.97210026e-01 -1.06219113e+00 2.80614197e-01
-5.88065147e-01 -5.40419877e-01 -9.30688322e-01 8.54547918e-01
-1.21722162e+00 3.97351682e-02 -4.63197708e-01 -1.44338834e+00
3.13506663e-01 -1.25683808e+00 4.48106706e-01 5.29908478e-01
-1.52199948e-02 -9.43544507e-01 3.26265067e-01 -1.82283774e-01
3.53678286e-01 3.50678861e-01 6.78653419e-01 -6.95899427e-01
-6.83091104e-01 8.84294137e-02 1.70741156e-01 -7.46983066e-02
-1.87558904e-01 -1.21148460e-01 -9.88127351e-01 -2.91768879e-01
4.01952863e-02 -1.01840746e+00 5.72681308e-01 5.04511178e-01
7.96314776e-01 -4.07948285e-01 1.03555046e-01 9.77692455e-02
1.23657644e+00 3.97195488e-01 1.67426005e-01 5.13203204e-01
-1.06251076e-01 7.52412736e-01 9.64299083e-01 1.18726897e+00
5.48914135e-01 3.15724909e-01 1.03787565e+00 9.50743139e-01
8.42733979e-01 -5.30012071e-01 3.95853996e-01 -1.67799160e-01
-2.03621928e-02 -2.83962518e-01 -6.90480053e-01 3.09993654e-01
-2.05907798e+00 -8.99612486e-01 5.53487659e-01 2.38749194e+00
1.10050321e+00 4.34826106e-01 2.96461910e-01 -1.32558197e-01
4.51834857e-01 9.64145511e-02 -1.29502296e+00 -3.35427135e-01
9.77640599e-03 -1.04821645e-01 9.21446264e-01 6.32312417e-01
-9.11234796e-01 5.62168181e-01 8.36357307e+00 7.80932128e-01
-1.79528818e-01 -6.98693618e-02 4.47147280e-01 -4.60218906e-01
-7.14488804e-01 9.58438590e-02 -5.58582366e-01 3.83255690e-01
1.19769478e+00 -7.02515304e-01 4.75383312e-01 9.14431453e-01
-6.95988983e-02 -8.21953177e-01 -1.23926234e+00 4.30358887e-01
-5.94335258e-01 -7.25648701e-01 -4.12938446e-01 -5.69227450e-02
7.25347400e-01 -1.97850801e-02 1.97193533e-01 4.85799938e-01
1.61851084e+00 -9.59846973e-01 1.31072152e+00 5.90182006e-01
4.33614999e-01 -1.33946824e+00 8.13204765e-01 2.41041705e-01
-6.55411363e-01 -3.80964428e-01 -2.96713501e-01 -2.74636708e-02
1.67830020e-01 6.24818683e-01 -4.32556301e-01 3.45563561e-01
5.37143588e-01 1.14551648e-01 -1.54032167e-02 1.04364145e+00
-9.09297884e-01 4.68056142e-01 -5.66632450e-01 -1.81705222e-01
3.00504029e-01 -2.01334357e-01 5.65612257e-01 4.96228784e-01
5.48390090e-01 1.65148869e-01 4.12557162e-02 1.14396846e+00
1.78699359e-01 -6.99257493e-01 -3.79664212e-01 -1.43325567e-01
1.09764922e+00 8.42678189e-01 -3.49321455e-01 1.37357563e-02
1.48063138e-01 2.22282037e-01 1.63278192e-01 5.14083028e-01
-1.12161791e+00 -4.62736130e-01 9.57885087e-01 -6.95534706e-01
-8.52428190e-03 -1.90567285e-01 4.44325916e-02 -1.07895732e+00
-2.78611243e-01 -7.00768590e-01 6.79173172e-01 -4.02688742e-01
-1.21957290e+00 3.37570250e-01 6.16158426e-01 -1.13530231e+00
-7.93543935e-01 -5.29547155e-01 -4.88659859e-01 6.86814904e-01
-1.71234834e+00 -5.25816858e-01 2.12090850e-01 3.58835965e-01
-5.29075004e-02 -3.33813913e-02 6.23777926e-01 -4.38686877e-01
-6.48217738e-01 3.96247298e-01 2.82964885e-01 -5.02873600e-01
6.13820374e-01 -1.63511586e+00 5.61071597e-02 6.62394106e-01
-3.91169339e-01 2.04007030e-01 1.23274112e+00 -6.61482930e-01
-1.37455261e+00 -7.73106575e-01 -1.64584517e-01 -5.75793922e-01
1.04718304e+00 -1.08564161e-01 -3.64652932e-01 7.11734414e-01
1.32034972e-01 1.02915335e-02 5.35954595e-01 1.96155608e-02
-4.99047786e-01 1.65285505e-02 -1.62109005e+00 5.41093349e-01
7.42073953e-01 -3.76521915e-01 -6.63709342e-01 -8.05900693e-02
8.13271403e-01 -6.16549194e-01 -9.40990806e-01 2.12854594e-01
6.37687266e-01 -1.01450515e+00 7.29036689e-01 -7.43908763e-01
-9.75506678e-02 -2.05085590e-01 -2.59940058e-01 -1.83556879e+00
2.86182463e-01 -1.00401556e+00 -3.70090991e-01 1.06405127e+00
3.72854859e-01 -7.61222959e-01 7.36650348e-01 1.17723119e+00
2.93035328e-01 -2.35866860e-01 -1.48630464e+00 -1.25081289e+00
4.93035585e-01 -6.82190537e-01 8.69891465e-01 4.91991192e-01
3.97971690e-01 -4.74893421e-01 -5.14620304e-01 3.44002992e-01
1.56068408e+00 -1.92075327e-01 6.91272765e-02 -1.05445993e+00
-1.99669853e-01 -2.31940299e-01 -1.70445308e-01 -4.65647519e-01
2.75316685e-01 -1.39264446e-02 6.02807999e-01 -1.03895676e+00
-1.25053540e-01 -5.54651141e-01 -5.55212319e-01 2.80996293e-01
1.08748870e-02 -5.40841222e-01 -5.11128381e-02 -4.57805961e-01
-8.85983646e-01 1.16067815e+00 8.39314342e-01 -8.11868832e-02
7.32921287e-02 2.07116812e-01 -6.74468994e-01 8.05976450e-01
7.55105019e-01 -5.07393837e-01 -8.31412017e-01 -1.84512645e-01
6.02419734e-01 6.33337319e-01 1.91495866e-01 -9.87583756e-01
1.98058262e-01 -9.32323635e-01 3.26013833e-01 -4.03036058e-01
2.44673356e-01 -9.32635188e-01 1.23075113e-01 4.62131709e-01
-8.01756918e-01 -2.68793285e-01 3.47508430e-01 1.09136999e+00
9.74582583e-02 -4.64935809e-01 7.84995914e-01 -1.39543638e-01
-4.77122307e-01 1.49853587e-01 -9.53340709e-01 4.46555912e-01
1.22518992e+00 2.24250764e-01 -5.03838539e-01 -9.11441267e-01
-5.19211054e-01 9.33621883e-01 1.45921499e-01 1.06478132e-01
5.49217880e-01 -1.29643214e+00 -4.30531174e-01 -1.99459910e-01
-7.60271167e-03 -7.70804733e-02 2.49694020e-01 2.76714176e-01
-1.25735968e-01 1.06074110e-01 -3.84272099e-01 2.84138359e-02
-3.48462999e-01 3.26427251e-01 6.08522296e-01 -1.62189007e-01
1.01715513e-01 9.45203900e-01 -1.54868677e-01 -4.22922760e-01
7.44711995e-01 -2.43479982e-01 -6.25646487e-02 1.44297346e-01
6.17931902e-01 6.99771047e-01 -2.70992965e-01 1.02876432e-01
-2.93064058e-01 9.62897092e-02 8.19788054e-02 -7.96816945e-01
1.18459511e+00 -3.76799732e-01 3.82710129e-01 6.46246493e-01
1.87650889e-01 -2.40300551e-01 -2.07321191e+00 -2.31006548e-01
2.91839838e-01 -5.09763300e-01 2.44725063e-01 -1.18011355e+00
-6.91156507e-01 7.20690787e-01 4.72457230e-01 3.03707607e-02
7.30882049e-01 -3.93657953e-01 1.28094628e-01 6.75081372e-01
1.11978614e+00 -1.60743546e+00 -1.42934754e-01 2.95800269e-01
7.10432887e-01 -1.07483828e+00 2.68081039e-01 1.72874063e-01
-9.44225967e-01 9.48780179e-01 7.36746490e-01 9.86171886e-02
7.26505578e-01 7.58284688e-01 -5.58505245e-02 2.14254141e-01
-1.16122723e+00 -2.10175589e-01 -2.83895493e-01 1.02016711e+00
-3.38325649e-01 2.54546881e-01 5.93794696e-02 8.62948120e-01
-6.54452145e-02 -2.25265175e-01 1.00374913e+00 1.28752768e+00
-6.69430375e-01 -1.07351100e+00 -4.71270084e-01 4.18653309e-01
-2.80711591e-01 3.15134853e-01 -6.04651868e-05 7.42900550e-01
-3.07757050e-01 1.09367955e+00 -1.66264743e-01 -2.01560006e-01
2.75137126e-01 -3.75873893e-01 6.38212085e-01 -1.67580098e-01
-1.81788072e-01 -2.38534182e-01 5.38271904e-01 -1.05185807e+00
-1.83028266e-01 -8.48008096e-01 -1.16832459e+00 -7.20221326e-02
-3.19230050e-01 6.90578282e-01 6.89848602e-01 7.94985473e-01
-1.12137683e-01 4.34645593e-01 7.08499491e-01 -8.23193252e-01
-1.66103864e+00 -4.59506482e-01 -1.23270881e+00 -4.97821003e-01
4.72063124e-01 -1.25097251e+00 -9.13076937e-01 -7.52927721e-01] | [4.252823352813721, 2.5113062858581543] |
50357ad8-6865-4388-829f-6fda23374bf2 | learning-the-effects-of-physical-actions-in-a | 2301.11845 | null | https://arxiv.org/abs/2301.11845v2 | https://arxiv.org/pdf/2301.11845v2.pdf | Learning the Effects of Physical Actions in a Multi-modal Environment | Large Language Models (LLMs) handle physical commonsense information inadequately. As a result of being trained in a disembodied setting, LLMs often fail to predict an action's outcome in a given environment. However, predicting the effects of an action before it is executed is crucial in planning, where coherent sequences of actions are often needed to achieve a goal. Therefore, we introduce the multi-modal task of predicting the outcomes of actions solely from realistic sensory inputs (images and text). Next, we extend an LLM to model latent representations of objects to better predict action outcomes in an environment. We show that multi-modal models can capture physical commonsense when augmented with visual information. Finally, we evaluate our model's performance on novel actions and objects and find that combining modalities help models to generalize and learn physical commonsense reasoning better. | ['Alex Lascarides', 'Frank Keller', 'Gautier Dagan'] | 2023-01-27 | null | null | null | null | ['physical-commonsense-reasoning'] | ['reasoning'] | [ 6.75609112e-01 2.58199368e-02 -1.59992665e-01 -2.00455934e-01
-5.16402423e-01 -3.82198632e-01 1.02508533e+00 1.53090999e-01
-3.54435265e-01 7.21913099e-01 8.06947231e-01 -8.89290646e-02
-6.73812404e-02 -8.42368305e-01 -8.82843614e-01 -4.62358505e-01
1.94114730e-01 3.13119650e-01 -7.88509324e-02 -6.95457906e-02
4.71378595e-01 6.19927526e-01 -1.43786442e+00 8.09281707e-01
7.52693892e-01 6.34205937e-01 6.23122036e-01 7.30737269e-01
1.09630730e-02 1.68135619e+00 -3.48511368e-01 -8.32938850e-02
-2.57641803e-02 -6.66947305e-01 -1.00462782e+00 6.45186901e-02
4.25348073e-01 -7.28196263e-01 -7.22647011e-01 9.91458118e-01
7.49758780e-02 5.19763470e-01 9.95657623e-01 -1.33273649e+00
-1.30953264e+00 6.31746054e-01 -3.05656722e-04 7.78392795e-03
7.58024991e-01 6.97695076e-01 1.08454037e+00 -4.44265515e-01
5.91697037e-01 1.73575497e+00 2.48769388e-01 7.37228632e-01
-1.59477603e+00 -2.29337335e-01 1.43963009e-01 7.33837008e-01
-8.41480911e-01 -6.13800824e-01 8.63225222e-01 -3.19698542e-01
1.29566395e+00 1.41844213e-01 6.89696372e-01 1.62550986e+00
5.45664847e-01 1.06727886e+00 1.39645350e+00 -4.28139269e-01
3.73511493e-01 -3.79869610e-01 -2.08952516e-01 6.13278747e-01
1.34937763e-01 3.53520423e-01 -9.00142550e-01 5.99255823e-02
7.26073623e-01 2.13066980e-01 9.94234085e-02 -2.20144466e-01
-1.47860038e+00 5.14849246e-01 6.07712150e-01 2.79653788e-01
-7.27519333e-01 8.26948345e-01 1.05082281e-01 2.08921880e-02
8.30396488e-02 7.10978031e-01 -1.20688118e-01 -3.39183927e-01
-5.31285405e-01 5.40553272e-01 3.34883213e-01 6.83543146e-01
5.02271295e-01 1.90872196e-02 -2.26324886e-01 5.50451696e-01
1.52665779e-01 7.97775686e-01 5.12370825e-01 -1.35695469e+00
3.96575212e-01 5.05318165e-01 3.42252851e-01 -9.33748960e-01
-3.94620001e-01 3.00524831e-01 -4.08993244e-01 3.43678892e-01
3.52023274e-01 3.41673702e-01 -1.06276858e+00 1.97891641e+00
-2.10053325e-01 2.82744914e-01 2.51870513e-01 7.73603082e-01
6.95679069e-01 6.14029825e-01 6.95289552e-01 -1.19962908e-01
8.88808668e-01 -6.18029118e-01 -6.75882518e-01 -8.31750095e-01
8.69951248e-01 -1.93275720e-01 1.49845433e+00 2.70352691e-01
-1.17347491e+00 -4.91116941e-01 -6.56969786e-01 -3.71902585e-01
-4.40869838e-01 -2.58132935e-01 7.92220533e-01 8.42944253e-03
-7.51498580e-01 8.13525856e-01 -9.91825163e-01 -2.84898937e-01
7.27610350e-01 -4.96146157e-02 -4.45470870e-01 -3.76993477e-01
-1.04981351e+00 1.51732481e+00 4.88875866e-01 -1.38042256e-01
-1.33493257e+00 -4.28234756e-01 -1.16768742e+00 5.10576665e-02
4.27329659e-01 -7.65915275e-01 1.19963825e+00 -7.83114374e-01
-1.39829803e+00 8.54402125e-01 -2.66227812e-01 -4.05945778e-01
1.20084502e-01 -2.93821216e-01 -2.82837063e-01 1.87199339e-01
1.28702492e-01 8.29220653e-01 7.01126039e-01 -1.31923854e+00
-1.65755004e-01 -5.42408824e-01 5.08209944e-01 3.68933678e-01
5.70871644e-02 -1.80269390e-01 3.44197273e-01 -3.40985507e-01
6.06392398e-02 -9.73514616e-01 -2.40683541e-01 -8.40613022e-02
-5.39532065e-01 -1.96912453e-01 3.17192405e-01 -5.49762249e-01
7.46941566e-01 -2.15850401e+00 5.61925650e-01 -2.99635887e-01
3.29423174e-02 -1.50672644e-01 -3.38343710e-01 5.55448830e-01
2.18528770e-02 1.55826524e-01 2.86025070e-02 -3.91177922e-01
4.63587880e-01 3.80243331e-01 -5.69477856e-01 2.25521937e-01
1.10760994e-01 1.41713965e+00 -1.04944420e+00 -7.10546255e-01
5.74871182e-01 1.57430977e-01 -5.27051449e-01 4.70942035e-02
-6.85331166e-01 3.75651091e-01 -4.39474702e-01 4.66701686e-01
9.05561820e-02 -3.06505978e-01 9.54446271e-02 -1.55075798e-02
2.70816594e-01 6.09353602e-01 -7.09303737e-01 1.94748569e+00
-7.21108079e-01 5.22159874e-01 -6.13335550e-01 -6.75225198e-01
1.92090437e-01 1.96334615e-01 1.79421738e-01 -9.12489951e-01
1.87093089e-03 -1.34892195e-01 1.39211923e-01 -8.86685848e-01
3.07692558e-01 -9.45000112e-01 -3.52203906e-01 8.09543133e-01
-7.37061799e-02 -5.83961248e-01 4.21120971e-02 3.83638889e-01
1.20643568e+00 4.73381609e-01 4.99610633e-01 4.56017762e-01
4.12904061e-02 2.36966446e-01 3.29727322e-01 1.04581189e+00
-2.84953862e-01 2.41893411e-01 2.96102971e-01 -3.83381337e-01
-9.19421494e-01 -1.36929429e+00 1.45262420e-01 1.19016063e+00
2.69927025e-01 -2.37654254e-01 -2.59098470e-01 -5.76022446e-01
-3.13367024e-02 1.71906757e+00 -7.96391428e-01 -6.69214249e-01
-2.62162477e-01 -3.48215133e-01 4.38415617e-01 9.08648431e-01
3.43095243e-01 -1.76680183e+00 -1.21226251e+00 -2.01168861e-02
-4.43275452e-01 -1.05102861e+00 -2.70345975e-02 4.93961535e-02
-7.60498047e-01 -9.22720015e-01 1.70571283e-02 -1.96789473e-01
4.47913170e-01 1.10894270e-01 1.05372322e+00 -5.86085655e-02
-3.13762307e-01 6.92659140e-01 -2.63120174e-01 -4.04774189e-01
-7.50278711e-01 -7.75272131e-01 2.35361785e-01 -3.67363632e-01
4.63939786e-01 -6.49937868e-01 -2.08093628e-01 -3.08167428e-01
-1.07094038e+00 6.87243938e-01 6.47148550e-01 5.79647005e-01
2.87463099e-01 -1.42574310e-01 4.09677267e-01 -4.38859344e-01
6.46864653e-01 -3.79729360e-01 3.28795016e-02 4.29603040e-01
-4.66553681e-02 3.05552870e-01 5.77938676e-01 -7.60144472e-01
-1.54122376e+00 -8.94217268e-02 2.96761245e-01 -3.17017913e-01
-5.97535610e-01 5.41608393e-01 -2.35904232e-01 5.25883973e-01
8.39305639e-01 4.73032802e-01 -2.77347893e-01 -2.33319908e-01
7.92448401e-01 1.74663410e-01 3.56781662e-01 -8.38997424e-01
3.66280228e-01 6.66975439e-01 3.66286218e-01 -5.76601923e-01
-1.13135314e+00 -9.46221128e-03 -7.63505757e-01 -2.56992579e-01
1.02248085e+00 -7.24673927e-01 -8.50816250e-01 4.01491880e-01
-1.21237075e+00 -7.39967883e-01 -5.10148704e-01 7.07536221e-01
-1.23479283e+00 2.47622833e-01 -6.05290651e-01 -8.72601748e-01
4.47476149e-01 -9.11017835e-01 1.06841826e+00 -1.61809847e-02
-5.65317810e-01 -1.10063291e+00 6.42730594e-02 4.42952067e-01
7.18724802e-02 3.67264181e-01 1.19931805e+00 -3.98569226e-01
-6.04133964e-01 -3.72690819e-02 -1.21065229e-01 2.28107363e-01
2.55589098e-01 -3.88850331e-01 -9.53911066e-01 2.71519095e-01
-5.70940636e-02 -1.04312778e+00 9.27392781e-01 3.46253127e-01
1.18556464e+00 -4.46110696e-01 -2.85187721e-01 1.95482031e-01
1.34148932e+00 1.62056372e-01 9.32939053e-01 2.02497393e-01
5.53512812e-01 4.63636339e-01 6.09912813e-01 4.26048726e-01
3.43200237e-01 4.40661550e-01 6.52989507e-01 4.52090561e-01
-2.98680395e-01 -6.20684743e-01 6.85286939e-01 -4.82440703e-02
-3.66228253e-01 -1.48599476e-01 -7.90477872e-01 5.02942741e-01
-1.75503623e+00 -1.75787568e+00 7.36007020e-02 1.80738115e+00
1.02559078e+00 9.06820893e-02 -1.24329403e-01 -1.92098722e-01
1.03476621e-01 3.77500206e-01 -9.42668140e-01 -3.42141330e-01
-5.46421222e-02 1.09967656e-01 4.27884795e-02 4.90212351e-01
-6.48096919e-01 1.21875036e+00 7.00301790e+00 6.11552894e-01
-6.98773861e-01 1.04529403e-01 2.72486031e-01 -4.41354126e-01
-5.62158883e-01 4.08922061e-02 -2.39608824e-01 1.73660338e-01
9.26939845e-01 -6.24360442e-02 7.66930223e-01 6.70809329e-01
3.09693694e-01 -5.01529753e-01 -1.58837342e+00 9.70263243e-01
2.26332873e-01 -1.28487909e+00 2.88866967e-01 4.73384969e-02
5.23544371e-01 -1.84552416e-01 1.77278534e-01 5.20982087e-01
7.64860988e-01 -1.41658688e+00 9.42937016e-01 8.28732729e-01
4.92057413e-01 -3.89706433e-01 1.82529017e-02 8.27626824e-01
-5.33346057e-01 -3.26137513e-01 -4.05675292e-01 -7.60048687e-01
3.19997013e-01 -2.32281741e-02 -6.59476697e-01 -5.53871281e-02
2.16067120e-01 7.21927762e-01 -5.49634099e-01 4.45560426e-01
-6.47385061e-01 2.90530056e-01 -4.60925587e-02 -1.99992388e-01
6.18047491e-02 9.21081901e-02 4.61356550e-01 7.73122966e-01
1.51074514e-01 3.98316920e-01 5.25760800e-02 1.44931817e+00
2.54479051e-01 -2.28313670e-01 -1.00734866e+00 -4.27135646e-01
2.36472428e-01 6.99952781e-01 -3.31925213e-01 -4.67240691e-01
-2.57977396e-01 1.18314075e+00 5.86292505e-01 5.29130876e-01
-9.22655642e-01 3.31937641e-01 6.35880351e-01 -1.74941331e-01
-2.85759419e-01 -3.52428824e-01 -5.19676208e-01 -1.27272141e+00
-1.83321655e-01 -6.80754840e-01 7.14153498e-02 -1.71340132e+00
-1.18663275e+00 -7.63917938e-02 3.18377197e-01 -8.13126743e-01
-4.45787817e-01 -7.69050002e-01 -4.46290672e-01 7.03165770e-01
-1.04842365e+00 -1.56894553e+00 -9.06177461e-02 5.31995773e-01
6.40809000e-01 2.48827919e-01 1.07235992e+00 -6.03494287e-01
-3.15457880e-02 -2.58460373e-01 -2.40060434e-01 -8.79368782e-02
4.31591123e-01 -1.22172081e+00 4.53981981e-02 7.22849607e-01
4.08639044e-01 8.67252946e-01 9.00704384e-01 -1.04030633e+00
-1.43043065e+00 -6.32127702e-01 8.40041697e-01 -9.29108739e-01
6.95405126e-01 -2.46093441e-02 -6.88348949e-01 1.26619959e+00
2.39751637e-02 -6.05687760e-02 7.82020032e-01 2.92975664e-01
-6.33049607e-01 4.74492252e-01 -1.13125300e+00 1.06174266e+00
1.48331535e+00 -9.68035400e-01 -1.27012241e+00 4.46815521e-01
6.49330556e-01 -2.10428298e-01 -3.85850489e-01 6.72662538e-03
5.12771010e-01 -1.00918663e+00 1.19206822e+00 -1.43606305e+00
1.16400993e+00 -5.09843454e-02 -5.50433576e-01 -1.61268878e+00
-5.86555481e-01 2.34394409e-02 -2.73221195e-01 5.82430899e-01
6.22877255e-02 -2.53805697e-01 2.98718601e-01 1.00149751e+00
8.98516551e-02 -3.34760338e-01 -7.26852834e-01 -7.79902160e-01
3.88715304e-02 -9.20166016e-01 4.68368530e-01 7.03766525e-01
5.04642367e-01 4.00130361e-01 -4.21883643e-01 3.80440727e-02
5.44982374e-01 3.38416010e-01 5.67707479e-01 -8.86771381e-01
-4.56566125e-01 -3.23370695e-01 -4.11794037e-01 -9.28719699e-01
7.13566542e-01 -8.71712446e-01 1.92314610e-01 -1.91502047e+00
7.86241174e-01 -1.81207508e-01 -3.18803549e-01 8.00167739e-01
-1.67019159e-01 1.11748509e-01 6.55086100e-01 8.08511302e-02
-7.98585057e-01 5.48125446e-01 1.48364806e+00 -2.26246238e-01
-9.41473022e-02 -3.77201915e-01 -5.84490478e-01 1.05720067e+00
6.97492063e-01 -3.64273101e-01 -4.88192409e-01 -3.06975543e-01
5.61370134e-01 2.69608587e-01 1.11430669e+00 -7.22617507e-01
-1.71298668e-01 -1.11414921e+00 7.26518273e-01 -3.63957018e-01
8.15328300e-01 -7.69916415e-01 -2.81408746e-02 4.66937542e-01
-9.53463674e-01 -2.63418883e-01 1.72785014e-01 7.54062951e-01
2.43501663e-01 -3.44896734e-01 5.27166426e-01 -5.36634803e-01
-1.06396580e+00 -5.02539165e-02 -4.38025475e-01 2.43688505e-02
1.06675208e+00 -1.92905843e-01 -3.54962349e-01 -6.14401698e-01
-1.18370306e+00 -2.05576912e-01 7.09333420e-01 4.57872242e-01
8.96777332e-01 -1.45438373e+00 -3.90075326e-01 -1.92673370e-01
2.30220467e-01 -4.13794845e-01 5.96032679e-01 6.94017887e-01
-1.53299272e-01 4.81807500e-01 -4.71154720e-01 -2.00870410e-01
-1.08878505e+00 8.75746608e-01 1.51967064e-01 -7.12319165e-02
-6.28534973e-01 6.60299301e-01 5.87287068e-01 -1.40787080e-01
-9.04665440e-02 -5.25165975e-01 5.35482913e-03 -4.39383507e-01
6.09476566e-01 2.70238727e-01 -7.34777510e-01 -7.09288001e-01
-3.43605042e-01 3.47636119e-02 3.03890795e-01 -4.94670302e-01
1.49238873e+00 -1.25491604e-01 -4.74678576e-02 1.04437745e+00
7.59548485e-01 -1.27098545e-01 -1.49934769e+00 -1.69193000e-01
-2.90444136e-01 -6.30640686e-01 -9.18550044e-02 -1.24598348e+00
-2.04182416e-01 1.01245070e+00 3.45017686e-02 -1.29464343e-01
9.48064983e-01 3.99487108e-01 5.98955810e-01 6.54922426e-01
5.79262853e-01 -1.15019679e+00 7.66653121e-01 4.01300102e-01
1.18310821e+00 -1.22631991e+00 1.22441556e-02 5.11256345e-02
-1.00334167e+00 8.52988958e-01 5.31903088e-01 7.24235922e-02
1.03873909e-01 3.71173187e-03 -4.52239722e-01 -2.93323636e-01
-9.58551586e-01 -3.23645920e-01 2.43049309e-01 7.74928629e-01
1.80813909e-01 4.35469002e-01 3.35512936e-01 3.37832689e-01
-1.32206917e-01 1.58506364e-01 6.49150729e-01 1.05574107e+00
-5.46045303e-01 -6.23929262e-01 -4.40126181e-01 4.68468755e-01
-1.22115344e-01 -3.24488699e-01 -5.90129018e-01 6.20930970e-01
2.56749064e-01 9.05013621e-01 -8.04318637e-02 -1.59935907e-01
9.93735567e-02 5.65330327e-01 1.21662164e+00 -8.94161701e-01
3.81430872e-02 -5.17882288e-01 1.09706067e-01 -9.75955665e-01
-6.01772070e-01 -9.94421422e-01 -1.66861522e+00 -2.35922933e-01
1.53255612e-01 -6.72909856e-01 4.70350325e-01 1.47180033e+00
1.17888637e-01 6.25742614e-01 -5.58519103e-02 -8.49057257e-01
-8.48946750e-01 -9.42781746e-01 -6.12255752e-01 1.07551920e+00
3.72274697e-01 -8.51004124e-01 -3.87154251e-01 4.51354206e-01] | [4.5295610427856445, 0.8791645765304565] |
9cbf85b8-f864-4d1c-bf7f-06dc89656dc0 | dc3dcd-unsupervised-learning-for-multiclass | 2305.05421 | null | https://arxiv.org/abs/2305.05421v1 | https://arxiv.org/pdf/2305.05421v1.pdf | DC3DCD: unsupervised learning for multiclass 3D point cloud change detection | In a constant evolving world, change detection is of prime importance to keep updated maps. To better sense areas with complex geometry (urban areas in particular), considering 3D data appears to be an interesting alternative to classical 2D images. In this context, 3D point clouds (PCs) obtained by LiDAR or photogrammetry are very interesting. While recent studies showed the considerable benefit of using deep learning-based methods to detect and characterize changes into raw 3D PCs, these studies rely on large annotated training data to obtain accurate results. The collection of these annotations are tricky and time-consuming. The availability of unsupervised or weakly supervised approaches is then of prime interest. In this paper, we propose an unsupervised method, called DeepCluster 3D Change Detection (DC3DCD), to detect and categorize multiclass changes at point level. We classify our approach in the unsupervised family given the fact that we extract in a completely unsupervised way a number of clusters associated with potential changes. Let us precise that in the end of the process, the user has only to assign a label to each of these clusters to derive the final change map. Our method builds upon the DeepCluster approach, originally designed for image classification, to handle complex raw 3D PCs and perform change segmentation task. An assessment of the method on both simulated and real public dataset is provided. The proposed method allows to outperform fully-supervised traditional machine learning algorithm and to be competitive with fully-supervised deep learning networks applied on rasterization of 3D PCs with a mean of IoU over classes of change of 57.06% and 66.69% for the simulated and the real datasets, respectively. | ['Thomas Corpetti', 'Sébastien Lefèvre', 'Iris de Gélis'] | 2023-05-09 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 1.89488336e-01 -2.17923760e-01 7.38843307e-02 -3.81911337e-01
-4.81125325e-01 -6.31329656e-01 1.03576076e+00 6.37400150e-01
-5.89244306e-01 5.50161302e-01 -2.49049932e-01 -2.78661072e-01
-2.73245871e-01 -1.18069589e+00 -6.69719040e-01 -7.46488869e-01
-4.95127916e-01 8.27848315e-01 3.92914712e-01 -2.12717921e-01
9.64751244e-02 1.16838121e+00 -2.04421353e+00 -8.36952031e-02
8.86845350e-01 9.54388380e-01 1.72028854e-01 4.09229547e-01
-3.80671144e-01 -2.98600085e-02 -3.32525104e-01 6.33452013e-02
5.17458260e-01 6.42472655e-02 -5.51057160e-01 2.72642225e-01
4.90164489e-01 -6.58011585e-02 2.63241976e-01 1.05304825e+00
5.17935276e-01 4.84136641e-02 9.23797309e-01 -9.97006834e-01
2.39698365e-01 3.34942758e-01 -6.60435975e-01 3.65320563e-01
1.61861151e-01 -3.88793019e-03 8.09091210e-01 -8.35486829e-01
7.62361228e-01 9.46821690e-01 8.40501249e-01 -2.01757163e-01
-1.39237654e+00 -4.06716198e-01 3.14726740e-01 2.74623781e-01
-1.47356927e+00 -1.97913378e-01 9.30802405e-01 -8.23571622e-01
9.96584058e-01 4.16821063e-01 1.08043611e+00 6.36633813e-01
-6.08945563e-02 5.55259347e-01 1.25689030e+00 -3.52088690e-01
5.43706954e-01 8.24379995e-02 -1.62248667e-02 1.93586826e-01
4.08020467e-01 2.56316438e-02 8.60741809e-02 9.56281349e-02
4.41426992e-01 8.22052732e-02 -2.01891765e-01 -8.08059990e-01
-9.99249518e-01 6.53261542e-01 7.02008545e-01 7.10904002e-01
-5.10665178e-01 1.06737753e-02 3.03571373e-01 1.64287731e-01
9.93157506e-01 1.60222515e-01 -4.05798376e-01 -1.91896781e-01
-1.22954631e+00 2.54491806e-01 4.64126140e-01 5.64869583e-01
1.11246932e+00 -2.37800673e-01 3.89900655e-01 7.59576201e-01
1.60583198e-01 5.38001597e-01 3.25228244e-01 -5.54064810e-01
2.67381757e-01 1.08869576e+00 1.26584932e-01 -1.20017838e+00
-7.08179057e-01 -6.51528656e-01 -1.18152571e+00 6.07430041e-01
1.87162742e-01 2.30475336e-01 -1.02074218e+00 1.28633165e+00
5.82310617e-01 4.70280349e-02 -2.66817361e-01 6.65697753e-01
4.05362755e-01 5.60213983e-01 -2.26438299e-01 -1.34328753e-01
8.65415931e-01 -2.37608880e-01 -3.78249764e-01 1.31472498e-01
5.71996808e-01 -3.96711260e-01 1.00422454e+00 3.32003653e-01
-6.03724837e-01 -6.42867208e-01 -9.79488909e-01 3.77411664e-01
-9.58186984e-01 1.46657992e-02 3.98738861e-01 6.41553044e-01
-1.24406123e+00 7.47593224e-01 -1.11527014e+00 -8.05244803e-01
6.44381225e-01 4.80542868e-01 -3.69002879e-01 4.82697710e-02
-8.64280581e-01 8.30130875e-01 4.61198002e-01 1.99157834e-01
-6.08314991e-01 -5.29960334e-01 -6.31889880e-01 -7.64236227e-02
1.02402642e-01 -3.34258497e-01 7.06391990e-01 -1.04600763e+00
-1.12003040e+00 1.14911783e+00 5.98420091e-02 -6.34719193e-01
7.81362057e-01 -1.00165168e-02 -2.31474787e-01 1.33812129e-01
6.77827671e-02 7.23045409e-01 8.09729278e-01 -1.73863661e+00
-1.00515163e+00 -4.69149679e-01 5.34014814e-02 1.47099629e-01
-9.31301489e-02 -4.72994119e-01 -3.91881704e-01 -3.84561896e-01
4.34824139e-01 -9.64677989e-01 -3.53285193e-01 -1.67078272e-01
-2.33562306e-01 -3.11590046e-01 8.46362889e-01 -4.99136508e-01
9.76304352e-01 -1.91545904e+00 1.04256356e-02 5.60247004e-01
2.61859000e-01 3.54845792e-01 2.59207100e-01 3.12185556e-01
-2.38162398e-01 2.78805345e-01 -9.48544800e-01 -3.14074963e-01
7.71633908e-02 3.00926924e-01 1.62317920e-02 7.57187843e-01
3.25776905e-01 4.94941354e-01 -8.29121888e-01 -3.37706715e-01
7.32057571e-01 3.08562130e-01 -5.07016659e-01 -1.49854213e-01
-1.72949582e-01 5.34753621e-01 -2.38186106e-01 5.66665649e-01
1.04661882e+00 1.84244394e-01 -9.33330730e-02 -3.40688527e-02
-5.92264950e-01 -8.90282914e-02 -1.49965787e+00 1.45511699e+00
-2.93139219e-01 7.00313449e-01 -1.73852265e-01 -1.32103443e+00
1.12848794e+00 5.42378984e-03 8.76499295e-01 -5.89547276e-01
-1.62396170e-02 3.79408598e-01 -1.02002837e-01 -4.01930034e-01
5.53876936e-01 -9.01982561e-02 -2.17492152e-02 2.33305335e-01
-2.58909076e-01 -4.67873842e-01 1.08857028e-01 -1.68673858e-01
1.02370894e+00 1.35596782e-01 3.73867095e-01 -3.80266935e-01
6.31934047e-01 2.43132323e-01 2.86605120e-01 5.76071620e-01
-2.06491604e-01 5.70526242e-01 2.76060998e-01 -6.64375603e-01
-1.00267196e+00 -9.99555886e-01 -3.48795831e-01 3.64205211e-01
1.36993214e-01 3.82832363e-02 -4.92026091e-01 -5.47794998e-01
2.23669529e-01 5.47434449e-01 -5.19496560e-01 1.43567175e-01
-5.20588994e-01 -9.84904885e-01 1.91287562e-01 7.86399245e-02
6.20066941e-01 -7.64660716e-01 -6.72731578e-01 1.79090470e-01
7.18128830e-02 -7.53503799e-01 4.83533084e-01 2.91200310e-01
-1.07549262e+00 -1.04302835e+00 -4.88268435e-01 -6.47486389e-01
7.10637867e-01 2.18825296e-01 9.74648416e-01 -1.74202278e-01
-1.56142756e-01 5.30846357e-01 -5.87923348e-01 -2.70702869e-01
-3.68897170e-01 3.86633068e-01 1.01215271e-02 1.45551205e-01
3.91929209e-01 -9.85600948e-01 -4.76333708e-01 7.58336484e-02
-1.04117703e+00 -1.58750579e-01 4.89180237e-01 2.60769576e-01
7.79040992e-01 5.70218980e-01 2.80438364e-01 -7.42706180e-01
2.07920328e-01 -5.11658013e-01 -7.42068946e-01 -1.17217124e-01
-7.17631936e-01 -2.06671745e-01 4.13602144e-01 -6.24482110e-02
-8.84162188e-01 3.92164141e-01 -3.91006917e-01 -2.02628404e-01
-8.84658098e-01 6.55790627e-01 -3.00432652e-01 -9.77630168e-02
6.81044042e-01 8.61877650e-02 -3.38396698e-01 -7.15010941e-01
4.12141949e-01 7.41887689e-01 4.51296479e-01 -2.16388971e-01
1.24510515e+00 9.52047229e-01 2.51684785e-02 -9.63137507e-01
-1.19394757e-01 -7.81500697e-01 -1.37585402e+00 -5.67068100e-01
6.73196316e-01 -8.41699123e-01 -1.10676326e-01 6.65875375e-01
-1.01138365e+00 -3.59255552e-01 -5.58321297e-01 3.01365435e-01
-4.91706252e-01 4.52887386e-01 -2.31365804e-02 -7.37785697e-01
-5.36240153e-02 -9.04351354e-01 1.16487670e+00 8.01610276e-02
7.57585838e-02 -9.96219099e-01 3.75796318e-01 -9.71269608e-02
2.54411280e-01 8.98113966e-01 9.72079933e-01 -5.32532275e-01
-4.88879144e-01 -4.32762951e-01 -1.41469747e-01 2.81756401e-01
5.37895083e-01 1.58707544e-01 -9.66270983e-01 -2.10409433e-01
-1.61716014e-01 3.23270947e-01 9.07939553e-01 4.44076180e-01
9.77344990e-01 1.61971092e-01 -5.58039010e-01 4.75873500e-01
1.73787236e+00 1.77719176e-01 5.44991732e-01 6.48300827e-01
6.36130512e-01 6.46531224e-01 5.81551611e-01 6.10925794e-01
4.59807217e-01 8.34366679e-01 7.71281481e-01 -4.02554601e-01
-1.03346474e-01 1.15592137e-01 1.67811602e-01 6.93650484e-01
-2.72622108e-01 -4.48102653e-02 -1.37417543e+00 8.03061604e-01
-1.64712214e+00 -8.80302548e-01 -5.73573351e-01 2.40066814e+00
5.91078699e-01 4.10095394e-01 8.95595774e-02 6.02544963e-01
8.25478613e-01 1.70105621e-01 -4.55622345e-01 -1.21748902e-01
-3.08724254e-01 2.93324232e-01 6.47319317e-01 5.12293935e-01
-1.46328604e+00 8.81091177e-01 4.97726774e+00 6.46069407e-01
-1.24904501e+00 7.27309287e-02 3.40095937e-01 5.67813478e-02
-1.36739761e-01 -8.97163674e-02 -6.25759661e-01 4.30779040e-01
6.88226521e-01 1.19972624e-01 1.30814642e-01 6.56407654e-01
5.53976715e-01 -3.98481697e-01 -8.38246644e-01 9.95087206e-01
-9.90951881e-02 -1.09991097e+00 1.35260180e-01 2.25516543e-01
9.79777455e-01 4.19024110e-01 -2.91675985e-01 1.13350041e-01
1.43634066e-01 -6.67793572e-01 6.58562183e-01 6.75837815e-01
6.07348502e-01 -7.66870975e-01 8.18215668e-01 3.76631379e-01
-1.35035980e+00 -4.10380922e-02 -3.42337191e-01 -2.35708207e-01
1.52507171e-01 1.05275226e+00 -9.38118100e-01 8.34457397e-01
9.55734909e-01 1.08880365e+00 -8.51459801e-01 1.34358394e+00
-6.08621649e-02 5.88838577e-01 -7.54185736e-01 2.84276813e-01
3.61864805e-01 -3.45692337e-01 7.54530847e-01 1.40639520e+00
4.58818167e-01 -2.47172862e-01 1.65732935e-01 7.21375585e-01
2.22321868e-01 8.01073313e-02 -7.75708139e-01 2.98504591e-01
2.69869655e-01 1.30252445e+00 -1.17416167e+00 -3.00496429e-01
-9.38635394e-02 8.25835466e-01 8.35048631e-02 2.20010012e-01
-5.66461325e-01 -3.45166028e-01 6.85658693e-01 3.26129854e-01
5.52125812e-01 -6.27452374e-01 -3.50088298e-01 -8.87065709e-01
5.08715436e-02 -4.22554702e-01 2.37188607e-01 -4.56771761e-01
-1.05848968e+00 4.64419514e-01 3.07222217e-01 -1.60572994e+00
-2.29496002e-01 -4.35219318e-01 -5.65927625e-01 5.60861528e-01
-1.96532655e+00 -1.03498507e+00 -6.78761780e-01 5.36732674e-01
4.96241331e-01 1.95168972e-01 5.33734620e-01 4.99362499e-01
-4.62698787e-01 -1.06375918e-01 5.47800779e-01 -9.06771943e-02
4.13003594e-01 -1.62029588e+00 5.20015359e-01 1.05135906e+00
1.18815288e-01 -2.68056933e-02 6.74914837e-01 -6.50946200e-01
-8.51732075e-01 -1.34937489e+00 8.53260756e-01 -2.47774303e-01
4.57161307e-01 -3.45014751e-01 -9.13841963e-01 3.70091856e-01
-1.51075587e-01 -1.68685690e-01 4.09115374e-01 -6.14095144e-02
1.33215308e-01 -4.43081349e-01 -1.10488868e+00 4.04160053e-01
1.19385314e+00 -2.79982358e-01 -5.42012691e-01 1.81579828e-01
3.87120754e-01 -2.00226828e-01 -9.24218118e-01 5.75435698e-01
7.82127529e-02 -1.16185939e+00 8.07181120e-01 -1.58223044e-03
1.35995254e-01 -6.67275429e-01 -9.22252089e-02 -1.47600687e+00
-1.68394700e-01 -8.91225487e-02 3.14840227e-01 1.22927225e+00
1.55702278e-01 -6.30041122e-01 7.70519614e-01 1.12244070e-01
-3.60192120e-01 -3.57899725e-01 -1.31285083e+00 -7.69572318e-01
1.23210542e-01 -8.27701211e-01 7.64411569e-01 1.13701189e+00
-5.44875383e-01 -7.06053227e-02 2.36784190e-01 5.27384818e-01
5.00773132e-01 2.52401978e-01 1.04746342e+00 -1.92221522e+00
3.22995156e-01 -7.73523033e-01 -8.71375382e-01 -6.63862288e-01
3.73165943e-02 -1.07560730e+00 4.48835874e-03 -1.80341339e+00
-2.88423121e-01 -9.53187823e-01 -1.66673839e-01 3.84522885e-01
1.90844536e-01 2.45132193e-01 -5.74077815e-02 5.48536062e-01
-2.41019800e-01 6.59135461e-01 6.90284550e-01 -4.50893134e-01
-5.44188738e-01 2.84571797e-01 -4.53743860e-02 7.17644095e-01
9.10026729e-01 -4.41353470e-01 -1.28252193e-01 -3.01993728e-01
2.93893874e-01 -8.08395565e-01 4.28116649e-01 -1.42090189e+00
8.83437246e-02 7.36239254e-02 2.51961559e-01 -1.11677992e+00
1.66041255e-01 -1.30130005e+00 3.67038876e-01 5.54253161e-01
1.12598360e-01 -1.23263612e-01 2.46230468e-01 5.16488671e-01
-3.09141189e-01 -1.04797885e-01 7.63677299e-01 -1.73668846e-01
-9.74214435e-01 3.14132303e-01 -6.10575080e-01 -4.79757369e-01
9.92168069e-01 -4.99642015e-01 1.49828136e-01 -2.25011259e-01
-7.86749005e-01 9.28078443e-02 7.00530589e-01 3.25147539e-01
3.57612461e-01 -1.28914583e+00 -5.55182755e-01 1.78028554e-01
2.62072623e-01 4.32185501e-01 1.72254816e-01 8.01910222e-01
-7.55797803e-01 3.20385069e-01 -2.67447978e-01 -1.13228965e+00
-1.07546830e+00 4.12083000e-01 5.06754398e-01 -2.97168225e-01
-7.28671908e-01 2.65961081e-01 -2.59212285e-01 -8.73645902e-01
9.14304033e-02 -8.88708591e-01 -5.00368953e-01 6.50483012e-01
-1.36815533e-02 4.71847713e-01 5.11637032e-01 -8.23174238e-01
-3.87481511e-01 8.69487941e-01 4.37264204e-01 6.88505471e-02
1.70764470e+00 -3.06248277e-01 -1.52304620e-01 6.24852479e-01
1.15570998e+00 -1.60021618e-01 -1.26029992e+00 -2.32590362e-01
4.17549402e-01 -3.93128842e-01 2.05586165e-01 -6.52909756e-01
-8.88993382e-01 9.04551208e-01 1.27258170e+00 4.65176523e-01
1.22628307e+00 -1.12901432e-02 3.12831014e-01 4.39680696e-01
4.49912697e-01 -1.20750368e+00 -4.18906480e-01 3.43680114e-01
8.45064640e-01 -1.36831319e+00 1.96108550e-01 -3.22536230e-01
-1.34206275e-02 1.08452022e+00 2.53408104e-01 -1.53426260e-01
8.70237231e-01 -6.94664866e-02 -1.53679341e-01 -2.88830489e-01
-1.05308786e-01 -7.76394963e-01 4.58560623e-02 7.97279179e-01
-1.02927275e-01 2.82408446e-01 -2.55431831e-01 5.73639199e-02
-1.80935696e-01 -1.62383527e-01 4.18490261e-01 9.36238766e-01
-7.01738715e-01 -1.08184743e+00 -6.05934262e-01 4.65591758e-01
1.56927899e-01 2.36904964e-01 -5.06551445e-01 1.10629904e+00
6.61903501e-01 8.22187543e-01 5.16546130e-01 -3.73886198e-01
6.11420572e-01 2.13316046e-02 1.96037039e-01 -4.18842226e-01
-5.20783305e-01 1.06750177e-02 -1.24523386e-01 -3.88553619e-01
-9.91828561e-01 -1.05669987e+00 -1.05627728e+00 -8.58354419e-02
-1.27180308e-01 -1.02960825e-01 9.71295238e-01 9.35495257e-01
2.56035864e-01 1.75811082e-01 9.95751858e-01 -1.33068013e+00
2.67898552e-02 -9.09289420e-01 -6.60431206e-01 4.17787552e-01
3.22291434e-01 -8.38689983e-01 -3.99845302e-01 1.88886121e-01] | [8.516683578491211, -2.4198503494262695] |
9cb750b0-6dae-467e-a61e-d9b86ca7ff94 | evolutionary-n-level-hypergraph-partitioning | 1803.09258 | null | http://arxiv.org/abs/1803.09258v3 | http://arxiv.org/pdf/1803.09258v3.pdf | Evolutionary n-level Hypergraph Partitioning with Adaptive Coarsening | Hypergraph partitioning is an NP-hard problem that occurs in many computer
science applications where it is necessary to reduce large problems into a
number of smaller, computationally tractable sub-problems. Current techniques
use a multilevel approach wherein an initial partitioning is performed after
compressing the hypergraph to a predetermined level. This level is typically
chosen to produce very coarse hypergraphs in which heuristic algorithms are
fast and effective. This article presents a novel memetic algorithm which
remains effective on larger initial hypergraphs. This enables the exploitation
of information that can be lost during coarsening and results in improved final
solution quality. We use this algorithm to present an empirical analysis of the
space of possible initial hypergraphs in terms of its searchability at
different levels of coarsening. We find that the best results arise at
coarsening levels unique to each hypergraph. Based on this, we introduce an
adaptive scheme that stops coarsening when the rate of information loss in a
hypergraph becomes non-linear and show that this produces further improvements.
The results show that we have identified a valuable role for evolutionary
algorithms within the current state-of-the-art hypergraph partitioning
framework. | ['Jim Smith', 'Richard J. Preen'] | 2018-03-25 | null | null | null | null | ['hypergraph-partitioning'] | ['graphs'] | [ 4.29589093e-01 4.84238327e-01 -1.21921651e-01 1.30495399e-01
-4.23504859e-01 -4.74714786e-01 1.31508291e-01 4.10011172e-01
-1.79712281e-01 1.14993382e+00 -1.54902428e-01 -1.15542017e-01
-7.63230085e-01 -1.24340701e+00 -3.95902961e-01 -8.19542527e-01
-1.97883010e-01 1.13017011e+00 5.63392758e-01 -2.29431316e-01
5.62991619e-01 5.74240267e-01 -2.10434055e+00 2.59890318e-01
9.98324871e-01 2.72126257e-01 1.97338358e-01 5.90748727e-01
-2.47276738e-01 -6.40302375e-02 -8.39585364e-01 -3.03213149e-01
5.39919794e-01 -6.13790929e-01 -1.29686487e+00 3.30622107e-01
-2.41689295e-01 4.96366441e-01 7.51586780e-02 1.20891666e+00
3.16114932e-01 1.42487153e-01 3.54602069e-01 -1.44011676e+00
-6.56872615e-02 8.95416141e-01 -8.74767601e-01 2.83955544e-01
5.56327760e-01 -1.60967618e-01 8.74053299e-01 -6.51278496e-02
7.80048430e-01 1.15646243e+00 5.27293444e-01 1.81086324e-02
-1.58909810e+00 -5.50108016e-01 -4.10430245e-02 2.58466035e-01
-1.79891980e+00 7.52348825e-02 5.07784605e-01 -1.48866456e-02
1.22976243e+00 8.55650485e-01 1.09399509e+00 -7.48359486e-02
4.38767076e-01 1.04213051e-01 1.20549309e+00 -9.04758394e-01
5.68976402e-01 3.44683707e-01 2.31001675e-01 3.69258106e-01
9.37807560e-01 2.30203480e-01 -4.40470949e-02 -4.52078193e-01
2.39684746e-01 -5.65295517e-01 -2.77897060e-01 -5.26408553e-01
-6.42615020e-01 9.52576995e-01 1.97178572e-01 6.11001790e-01
-4.94266957e-01 -1.23219684e-01 2.42810890e-01 4.41337675e-01
1.49926782e-01 8.56122375e-01 -2.16356441e-01 -7.17580169e-02
-1.11915541e+00 3.31206471e-01 1.24177039e+00 7.80680299e-01
8.56057644e-01 -3.57186794e-01 1.72797263e-01 6.32055998e-01
-4.78209108e-02 -1.05578326e-01 2.02595294e-01 -6.90691173e-01
1.02729149e-01 1.05627441e+00 -1.95050538e-01 -1.04950953e+00
-3.93528759e-01 -4.00270939e-01 -5.67662537e-01 3.13497782e-01
4.74950559e-02 1.29760981e-01 -9.83173668e-01 1.42250884e+00
6.57492757e-01 -1.00446999e-01 -9.84674692e-02 4.02148396e-01
3.76984686e-01 8.67536008e-01 -1.06683254e-01 -8.06639135e-01
1.18277299e+00 -7.21182287e-01 -7.30596125e-01 3.34196299e-01
4.56108451e-01 -6.36636913e-01 6.64525807e-01 7.26136863e-01
-1.37122250e+00 -2.29490086e-01 -1.26708579e+00 5.57837009e-01
-6.78212285e-01 -5.81477404e-01 5.70058405e-01 1.39705491e+00
-1.40459728e+00 5.92928171e-01 -4.93252963e-01 -5.44658184e-01
-1.08366750e-01 1.03614712e+00 -6.37063086e-02 -4.38282378e-02
-9.88736212e-01 1.05845737e+00 1.16321254e+00 -1.85013488e-01
-1.56866610e-01 -5.24840057e-01 -4.85282928e-01 3.31381828e-01
6.83284998e-01 -6.39657497e-01 5.28561175e-01 -7.14865088e-01
-1.15461648e+00 7.81759441e-01 2.89160181e-02 -4.25030082e-01
3.23624134e-01 8.23598444e-01 -1.78656906e-01 2.61233836e-01
-2.54893035e-01 5.49483657e-01 5.67340970e-01 -1.61000025e+00
-8.81392956e-01 -2.03699440e-01 5.40589243e-02 3.20237666e-01
-3.08844358e-01 -1.40004471e-01 -4.53066111e-01 -2.38825232e-01
2.04500571e-01 -1.27530372e+00 -5.30627847e-01 -1.22313702e+00
-4.49932724e-01 -2.34875411e-01 6.50240242e-01 -1.93614304e-01
1.98979068e+00 -1.48091626e+00 6.19798362e-01 9.67271030e-01
2.56176203e-01 1.18080162e-01 6.31556809e-02 9.01601076e-01
-1.32034123e-01 4.68115896e-01 -4.59738314e-01 3.40825438e-01
9.34823975e-02 3.67504984e-01 2.70990700e-01 4.45903122e-01
-3.79966736e-01 5.11206746e-01 -5.09381711e-01 -5.81609845e-01
2.19261289e-01 1.51089907e-01 -7.71982312e-01 -7.74249732e-02
-1.12762049e-01 -2.01661453e-01 -3.21302593e-01 4.93751645e-01
9.82024550e-01 -1.32623967e-02 5.24634719e-01 3.98931429e-02
-2.98025727e-01 -2.30952457e-01 -1.61283720e+00 1.13522434e+00
9.88599472e-03 2.07179785e-01 1.07093327e-01 -1.03309536e+00
7.71774113e-01 2.44382501e-01 7.67805994e-01 -3.23357403e-01
3.38529527e-01 3.91665176e-02 2.85272986e-01 -1.95978314e-01
7.81380475e-01 -1.89288542e-01 -6.25841394e-02 6.40749216e-01
-3.79602671e-01 -5.24121881e-01 8.57034743e-01 1.16592698e-01
1.24827874e+00 -4.18657839e-01 4.70989764e-01 -8.77464771e-01
6.82843208e-01 4.60554540e-01 4.49528843e-01 7.04372823e-01
3.36260907e-02 3.69188190e-01 6.13601625e-01 -2.98092514e-01
-1.25522316e+00 -4.04104441e-01 -2.93978691e-01 7.03983068e-01
3.82565200e-01 -5.44500530e-01 -1.12601256e+00 -2.33741626e-01
-1.76253885e-01 7.45033443e-01 -7.83022881e-01 -2.45222032e-01
-5.55732250e-01 -1.28873050e+00 1.15354117e-02 -9.57863480e-02
1.68906167e-01 -1.12504375e+00 -9.56159174e-01 4.11672920e-01
1.89221591e-01 -5.34410655e-01 1.22798368e-01 4.52297926e-01
-1.00714910e+00 -1.05857182e+00 -2.11692467e-01 -7.10149169e-01
9.16089594e-01 2.05852374e-01 1.21523237e+00 5.66795290e-01
-4.95457858e-01 1.75105929e-01 -6.28890097e-01 -2.78060973e-01
-5.65395057e-01 5.22558331e-01 -3.17744613e-01 -5.54175138e-01
3.48021954e-01 -5.88216126e-01 -1.58204541e-01 3.13317448e-01
-1.09779835e+00 -1.85682938e-01 5.28130352e-01 7.29484916e-01
7.26017058e-01 1.34899795e+00 3.46020430e-01 -1.16076624e+00
9.28314745e-01 -5.84146678e-01 -8.54347825e-01 5.29806674e-01
-9.00495887e-01 2.72709429e-01 4.24000502e-01 -2.15997532e-01
-9.84071612e-01 1.38434634e-01 8.15171003e-02 -2.76083481e-02
-1.11121526e-02 5.71541846e-01 -1.94778159e-01 -5.26846588e-01
4.18638349e-01 -4.53419238e-02 -3.23571324e-01 -1.16319150e-01
1.20968714e-01 3.00794035e-01 1.50857404e-01 -5.16246080e-01
7.86894441e-01 1.46485001e-01 4.46355373e-01 -6.91187382e-01
-2.23724976e-01 -5.87710679e-01 -6.27443671e-01 -2.94312596e-01
5.24513364e-01 -3.72654535e-02 -8.31035256e-01 -1.96203828e-01
-5.85316479e-01 5.64799011e-02 -3.83208930e-01 4.51299585e-02
-4.08334643e-01 3.31560880e-01 -1.52795643e-01 -8.66332054e-01
-8.42865705e-02 -1.30280340e+00 6.34630799e-01 3.70176286e-01
-4.40001488e-01 -9.18465674e-01 3.66697431e-01 1.28764018e-01
1.87633216e-01 4.10419285e-01 1.36725116e+00 -6.05502427e-01
-6.33853078e-01 -9.72334370e-02 1.20817833e-01 -5.02707660e-01
-1.75120216e-02 1.81485668e-01 -4.02363479e-01 -6.45485878e-01
-5.09189721e-03 -3.69255245e-02 6.90978050e-01 5.54689407e-01
9.63413835e-01 -2.62324274e-01 -7.58444846e-01 4.35033083e-01
1.95849705e+00 7.05951333e-01 9.02343869e-01 7.80777574e-01
2.66512215e-01 8.58572066e-01 7.07921743e-01 4.83646214e-01
7.20405653e-02 4.15614933e-01 3.85566235e-01 -9.22849923e-02
1.79229558e-01 1.99638337e-01 -4.75815088e-01 6.28457010e-01
-1.56552091e-01 -7.06780910e-01 -8.70109200e-01 6.70637548e-01
-1.75070393e+00 -9.45251465e-01 -1.93053320e-01 2.33153987e+00
7.66618609e-01 2.95115650e-01 3.61237884e-01 7.09920108e-01
1.18427956e+00 -3.24724942e-01 -9.88432169e-02 -1.11773825e+00
8.97978246e-02 3.04751366e-01 7.16014147e-01 6.37669861e-01
-7.42660403e-01 8.16889346e-01 6.88690424e+00 8.64751160e-01
-6.28124654e-01 -2.57158846e-01 5.12231231e-01 -9.86672118e-02
-5.79482913e-01 1.73714355e-01 -6.74618363e-01 4.48482722e-01
1.14165854e+00 -7.22913921e-01 5.86284041e-01 4.42073017e-01
-1.40922144e-01 -6.76315665e-01 -8.41817915e-01 5.58525145e-01
-1.15306601e-01 -1.18533099e+00 8.10529515e-02 5.00091970e-01
1.17871439e+00 -6.79955363e-01 -1.72227800e-01 5.40258847e-02
3.94526571e-01 -1.09235799e+00 2.47160450e-01 -1.60007477e-01
4.06831175e-01 -1.59145010e+00 7.87473083e-01 2.98587978e-01
-1.28635752e+00 -2.97989011e-01 -5.73907971e-01 9.82341170e-02
2.53899872e-01 3.75792235e-01 -8.35119784e-01 8.39070141e-01
6.35007739e-01 -2.75588214e-01 -4.45861697e-01 1.61904562e+00
2.68455207e-01 2.01215848e-01 -5.37595868e-01 -7.43638948e-02
3.00677240e-01 -5.05315304e-01 7.20215797e-01 9.44649637e-01
3.43418449e-01 5.57256341e-01 2.59020478e-01 6.34674072e-01
2.47014448e-01 2.50847757e-01 -5.34941316e-01 6.83949664e-02
7.05267608e-01 1.04238963e+00 -1.61521411e+00 -2.16541573e-01
5.55467568e-02 5.46172082e-01 4.85337945e-03 -6.04820959e-02
-7.42476523e-01 -6.08099282e-01 -6.24110587e-02 1.39150441e-01
3.82725298e-01 2.37536401e-01 -5.06932139e-01 -4.38497692e-01
-3.76841456e-01 -9.09965336e-01 7.95549691e-01 -1.86168522e-01
-6.43910885e-01 9.95396376e-01 5.86765945e-01 -6.64581835e-01
-2.97233254e-01 -3.66628207e-02 -4.48073864e-01 6.44788206e-01
-9.59392846e-01 -7.36972749e-01 -1.38980716e-01 2.57488549e-01
4.75007594e-01 2.64385883e-02 5.78211486e-01 9.35914516e-02
-3.48295659e-01 4.15960550e-01 -5.32357916e-02 -9.95506227e-01
-1.28531046e-02 -1.33991385e+00 -5.33053689e-02 1.04049408e+00
-2.48175621e-01 4.38749433e-01 1.25412774e+00 -1.00001645e+00
-1.28740418e+00 -5.84817469e-01 9.26084816e-01 -1.66865308e-02
3.25745761e-01 -6.92272708e-02 -9.78843033e-01 2.40400299e-01
5.55573940e-01 -7.12573826e-01 5.78819633e-01 1.26771629e-01
4.99203652e-01 1.19126052e-01 -1.60288310e+00 4.10458893e-01
8.32239389e-01 1.87661678e-01 -5.18637538e-01 1.81028411e-01
4.10384476e-01 -2.07571641e-01 -8.27625930e-01 5.09457111e-01
2.93477952e-01 -1.21329689e+00 8.35741818e-01 -2.07190946e-01
-5.62915802e-02 -2.97166437e-01 2.62510478e-01 -1.31899667e+00
-6.50831342e-01 -8.97718191e-01 1.28625080e-01 1.04857028e+00
4.55047518e-01 -6.62625134e-01 1.14248931e+00 8.58829260e-01
5.23386337e-02 -8.00904870e-01 -9.01391029e-01 -7.64355004e-01
-4.19587009e-02 2.09688917e-01 8.71540189e-01 9.09598649e-01
3.57414275e-01 2.65409555e-02 -6.31812960e-03 6.42912313e-02
7.75362909e-01 5.60954392e-01 3.82131875e-01 -1.50225246e+00
-1.25744149e-01 -6.80471599e-01 -4.36204135e-01 9.37966555e-02
4.12785225e-02 -6.53424740e-01 -7.80580863e-02 -1.57595730e+00
5.65412879e-01 -6.34363592e-01 -1.38794467e-01 2.54941225e-01
-8.86630490e-02 2.26625398e-01 5.12024835e-02 -1.10767834e-01
-4.12145764e-01 4.25806344e-02 1.08534670e+00 9.36375111e-02
-7.70827889e-01 -4.72676344e-02 -6.11099839e-01 3.86170447e-01
7.94923186e-01 -7.28361845e-01 -6.95185781e-01 2.47517049e-01
3.47962618e-01 1.97760731e-01 -4.38757896e-01 -8.25120270e-01
2.61907816e-01 -2.79337674e-01 3.68516967e-02 -8.65258574e-01
3.51902321e-02 -1.04046595e+00 9.72989798e-01 8.95059645e-01
-1.70373276e-01 3.27454120e-01 4.14265573e-01 4.24154162e-01
-1.59382671e-01 -7.43578732e-01 9.88312960e-01 -2.13462323e-01
-5.83305240e-01 -3.62108126e-02 -5.00687182e-01 -2.43125290e-01
1.73834944e+00 -9.22446311e-01 3.15224379e-02 -9.63152573e-02
-7.14961529e-01 3.41166288e-01 8.26997995e-01 -6.91475570e-02
2.95236021e-01 -8.57155740e-01 -5.09341300e-01 3.62669900e-02
-2.65031159e-01 -3.31462204e-01 3.00044388e-01 6.18830144e-01
-8.88943076e-01 5.72610140e-01 -5.39834797e-01 -3.45735043e-01
-1.73425031e+00 9.86375451e-01 1.70455519e-02 -4.86819059e-01
-6.36706829e-01 8.58531356e-01 -8.64931941e-02 7.07577094e-02
1.49949968e-01 -5.01078740e-02 -3.70849133e-01 2.04550251e-01
1.59977645e-01 9.14767325e-01 2.56464422e-01 -5.13151586e-01
-3.20847154e-01 5.17552257e-01 9.86756384e-03 -4.97542582e-02
1.44779396e+00 -4.04246420e-01 -6.66997194e-01 -1.93652302e-01
9.52970505e-01 8.27498809e-02 -5.67534089e-01 4.05368358e-01
1.45561591e-01 -7.36390293e-01 1.78091124e-01 -6.73703253e-01
-9.96513963e-01 8.23957399e-02 3.36142719e-01 9.52102184e-01
1.71719801e+00 -4.60089110e-02 5.39173365e-01 2.17688218e-01
5.31785131e-01 -1.22154880e+00 -5.46336293e-01 1.25061706e-01
6.50151968e-01 -6.72864497e-01 5.76366246e-01 -8.94121885e-01
-3.25312972e-01 9.92915869e-01 5.22741318e-01 -4.64625331e-03
4.05538648e-01 5.86435497e-01 -7.00684786e-01 -3.68599206e-01
-9.56272840e-01 -2.55060464e-01 8.67958292e-02 4.71520096e-01
-9.40319523e-02 1.03103824e-01 -1.13071489e+00 2.10642323e-01
-5.08087039e-01 -4.16349262e-01 8.46158981e-01 1.18683910e+00
-7.95239091e-01 -1.60573053e+00 -7.55410135e-01 5.11923075e-01
-4.49407905e-01 1.28354296e-01 -7.38623142e-01 1.09586847e+00
3.22703719e-01 9.55226302e-01 -1.44514292e-01 -2.81901777e-01
8.15444961e-02 -3.46741199e-01 8.14838052e-01 -5.33813000e-01
-9.20628726e-01 2.77502358e-01 3.49858522e-01 -4.50755000e-01
-2.80222803e-01 -5.23148954e-01 -1.28930712e+00 -6.84508085e-01
-6.95975006e-01 9.23997521e-01 5.11716247e-01 6.19497657e-01
6.57544807e-02 6.88051999e-01 5.08727789e-01 -8.12427342e-01
-9.51896980e-02 -3.21153939e-01 -7.58926451e-01 -2.92347316e-02
-2.30161607e-01 -7.50740170e-01 -2.18554080e-01 -1.25309721e-01] | [5.853832721710205, 3.90830135345459] |
c85e8351-3d56-4e25-93b3-c52b6a748e59 | wakeword-detection-under-distribution-shifts | 2207.06423 | null | https://arxiv.org/abs/2207.06423v1 | https://arxiv.org/pdf/2207.06423v1.pdf | Wakeword Detection under Distribution Shifts | We propose a novel approach for semi-supervised learning (SSL) designed to overcome distribution shifts between training and real-world data arising in the keyword spotting (KWS) task. Shifts from training data distribution are a key challenge for real-world KWS tasks: when a new model is deployed on device, the gating of the accepted data undergoes a shift in distribution, making the problem of timely updates via subsequent deployments hard. Despite the shift, we assume that the marginal distributions on labels do not change. We utilize a modified teacher/student training framework, where labeled training data is augmented with unlabeled data. Note that the teacher does not have access to the new distribution as well. To train effectively with a mix of human and teacher labeled data, we develop a teacher labeling strategy based on confidence heuristics to reduce entropy on the label distribution from the teacher model; the data is then sampled to match the marginal distribution on the labels. Large scale experimental results show that a convolutional neural network (CNN) trained on far-field audio, and evaluated on far-field audio drawn from a different distribution, obtains a 14.3% relative improvement in false discovery rate (FDR) at equal false reject rate (FRR), while yielding a 5% improvement in FDR under no distribution shift. Under a more severe distribution shift from far-field to near-field audio with a smaller fully connected network (FCN) our approach achieves a 52% relative improvement in FDR at equal FRR, while yielding a 20% relative improvement in FDR on the original distribution. | ['Joseph Wang', 'Christin Jose', 'Lu Zeng', 'Sree Hari Krishnan Parthasarathi'] | 2022-07-13 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 4.32210147e-01 3.75110596e-01 -2.25797936e-01 -5.82997262e-01
-1.33494723e+00 -5.65761983e-01 1.02614075e-01 3.65809500e-01
-4.96761441e-01 8.52678061e-01 -1.62241757e-01 -3.17650110e-01
-7.61270821e-02 -4.17585969e-01 -1.12871015e+00 -6.72516465e-01
-1.49833784e-02 4.50616896e-01 3.43734026e-01 3.94811541e-01
-2.29361713e-01 4.82165039e-01 -1.70054913e+00 2.23543048e-01
4.85411733e-01 1.23152244e+00 7.27634132e-03 7.89577425e-01
1.27071142e-01 6.29135072e-01 -1.19504952e+00 -7.67856985e-02
2.26832166e-01 -2.22198486e-01 -8.29817295e-01 -2.70367768e-02
7.86898971e-01 -4.07428265e-01 -2.63126731e-01 1.12168109e+00
9.61584866e-01 2.45097261e-02 5.25022447e-01 -1.36207545e+00
-2.88690031e-01 8.75650287e-01 -6.01298094e-01 2.77327806e-01
4.75753732e-02 6.77677542e-02 8.33428442e-01 -7.25055039e-01
1.60870641e-01 8.65637600e-01 8.17551792e-01 4.06652927e-01
-1.47677696e+00 -1.07119393e+00 1.84058458e-01 -3.51891965e-02
-1.70345891e+00 -4.17301923e-01 2.55930573e-01 -2.12713718e-01
7.03697085e-01 2.18955874e-01 4.22909647e-01 9.75985587e-01
-3.10087968e-02 7.52643585e-01 8.16709101e-01 -4.04160619e-01
6.96530700e-01 4.41974849e-01 -3.53378542e-02 2.97782868e-01
7.70425498e-02 -1.54595166e-01 -9.79724228e-01 -4.45758104e-01
1.85097292e-01 -1.50545314e-01 -4.95886028e-01 -1.36379167e-01
-7.32299328e-01 3.52068365e-01 -3.19077596e-02 7.49451891e-02
-6.54194430e-02 2.37467766e-01 3.34594786e-01 3.64192992e-01
5.94857275e-01 3.13622475e-01 -9.75537121e-01 -2.61173934e-01
-1.25962710e+00 1.74210250e-01 8.05110395e-01 1.11966407e+00
7.78888226e-01 6.80883154e-02 -4.37247634e-01 7.24071145e-01
1.32232279e-01 7.28902936e-01 4.08568174e-01 -6.26857996e-01
2.28084549e-01 2.84222011e-02 9.06966403e-02 -4.58233058e-01
-1.49996713e-01 -7.92491436e-01 -4.38913167e-01 -1.98285282e-01
5.00345170e-01 -2.78948337e-01 -1.09918761e+00 2.00571823e+00
4.75848585e-01 5.05288720e-01 -1.81860164e-01 5.04210949e-01
3.64885598e-01 5.11622787e-01 -9.24379155e-02 -4.05553013e-01
1.17520738e+00 -2.81628311e-01 -5.83336830e-01 -1.25599299e-02
6.38244212e-01 -7.23633707e-01 1.03256822e+00 7.74283171e-01
-8.47553194e-01 -5.74963689e-01 -1.16493165e+00 3.12223852e-01
-8.16366300e-02 1.41882882e-01 6.76439479e-02 6.85934424e-01
-1.07895494e+00 5.77050567e-01 -7.98055470e-01 -2.07565308e-01
6.86632037e-01 5.74192703e-01 -2.07560882e-01 -1.63013548e-01
-1.16812527e+00 3.46874624e-01 1.08682133e-01 -1.51318863e-01
-1.29295111e+00 -1.44624436e+00 -5.67763627e-01 2.91932255e-01
6.16163850e-01 -1.03744313e-01 1.66908658e+00 -6.84563458e-01
-1.36765528e+00 5.41054547e-01 6.31552637e-02 -5.65871656e-01
3.42207432e-01 -1.39939159e-01 -5.26429176e-01 -8.06926563e-02
8.32830444e-02 5.24737835e-01 1.15005577e+00 -8.24805558e-01
-9.61278021e-01 -2.23975390e-01 -2.87505656e-01 -1.95322618e-01
-5.30328155e-01 -2.65468001e-01 -2.85067856e-01 -5.29016197e-01
-2.48346195e-01 -9.10650611e-01 1.68692231e-01 -3.67849320e-02
-4.73090559e-01 -3.38764548e-01 9.13839579e-01 -4.62543309e-01
1.28497279e+00 -2.44220734e+00 -6.75891280e-01 3.25116992e-01
3.33017200e-01 3.82558703e-01 -5.91489039e-02 1.34853646e-01
-2.41868079e-01 -2.57596578e-02 -1.08499099e-02 -4.37657833e-01
-1.15689062e-01 1.01536354e-02 -4.35503364e-01 5.39923191e-01
3.02239180e-01 3.06998640e-01 -1.02602589e+00 3.75378244e-02
-3.08329999e-01 2.76628762e-01 -5.75275838e-01 4.30548400e-01
-1.55374095e-01 2.84665257e-01 -2.36133579e-02 4.96989608e-01
8.34351182e-01 -2.13656574e-01 1.75439894e-01 -3.54169726e-01
2.18137547e-01 3.80603284e-01 -1.37256718e+00 1.54141355e+00
-5.60151815e-01 6.41341925e-01 1.63072541e-01 -1.16845894e+00
9.01215494e-01 1.94524899e-01 2.58525074e-01 -4.33415204e-01
-6.39736233e-03 1.27021059e-01 -2.66558714e-02 -2.51370758e-01
2.82260895e-01 -1.02030970e-01 -8.10761601e-02 4.88544971e-01
5.06888568e-01 -1.73940867e-01 -2.75758624e-01 3.86599004e-01
1.41794908e+00 -3.22482884e-01 -8.50963891e-02 -3.64400417e-01
-1.34235933e-01 -3.39760244e-01 6.89981103e-01 1.04781151e+00
-7.96744004e-02 4.68785167e-01 6.27193332e-01 -2.03582034e-01
-7.96468973e-01 -1.25640845e+00 -4.94540602e-01 1.12181938e+00
-3.66294116e-01 -3.39893788e-01 -6.93115175e-01 -8.27136159e-01
2.08968922e-01 7.06021667e-01 -4.10138756e-01 -4.66349781e-01
1.95388566e-03 -6.35313630e-01 9.26957190e-01 4.35882866e-01
1.35380730e-01 -4.99958098e-01 -4.82715905e-01 1.55878678e-01
9.43502411e-02 -1.12689340e+00 -5.29064953e-01 8.64646673e-01
-3.76139164e-01 -9.14237678e-01 -4.84031469e-01 -4.20641035e-01
7.35427022e-01 -1.22904353e-01 7.82332242e-01 -4.23454642e-01
-5.57817996e-01 4.89379257e-01 -1.59106463e-01 -7.52625346e-01
-4.29539084e-01 1.24620952e-01 4.04361874e-01 1.01618521e-01
2.44092405e-01 -5.57347596e-01 -4.42812204e-01 1.85278729e-01
-9.11425769e-01 -5.29849946e-01 2.40302727e-01 1.06537426e+00
5.06775081e-01 4.85282928e-01 9.11829948e-01 -9.29553509e-01
2.69594520e-01 -5.06809771e-01 -8.34605217e-01 1.20309122e-01
-8.17277431e-01 8.91266614e-02 5.55821180e-01 -9.08083081e-01
-7.47045994e-01 -5.20560145e-02 -6.89778943e-03 -6.22441292e-01
-1.59117132e-01 2.93717921e-01 -3.18772942e-01 2.28847921e-01
8.13377917e-01 -1.52503356e-01 -1.10237800e-01 -5.52531719e-01
1.09546937e-01 1.06724739e+00 7.30124891e-01 -5.39408028e-01
6.18036926e-01 2.86724538e-01 -3.73078406e-01 -5.79091847e-01
-1.09904575e+00 -3.82615149e-01 -3.01323563e-01 6.63908124e-02
3.42437595e-01 -1.22829914e+00 -6.67872012e-01 5.27384222e-01
-7.03925073e-01 -6.91512227e-01 -7.64143109e-01 6.82487547e-01
-9.20500755e-02 -9.95484740e-03 -3.15311700e-01 -8.38801682e-01
-2.09476575e-01 -8.64969492e-01 1.20735812e+00 1.91463947e-01
-3.06636214e-01 -6.49664938e-01 -1.99092716e-01 4.88910684e-03
4.27992523e-01 -1.31419763e-01 1.00921786e+00 -1.26897287e+00
-1.31228179e-01 -3.29931438e-01 3.74710150e-02 8.13807726e-01
4.28726673e-01 -1.48738608e-01 -1.42054892e+00 -6.62520230e-01
-1.11151412e-01 -5.26744723e-01 5.49443066e-01 3.30134362e-01
1.58348584e+00 -1.55315042e-01 -3.74310702e-01 2.80897558e-01
8.87725174e-01 -1.10256607e-02 1.45677909e-01 -3.63283694e-01
4.67223525e-01 3.45715672e-01 8.08467925e-01 8.07674527e-01
1.99248809e-02 4.90622669e-01 2.10663736e-01 -1.06459484e-01
-1.88233420e-01 -4.98463333e-01 3.83787036e-01 4.06289965e-01
9.90839124e-01 -4.29776341e-01 -8.64138663e-01 7.37191617e-01
-1.40558672e+00 -4.93413240e-01 3.27896357e-01 2.73701763e+00
1.37909186e+00 3.00165087e-01 -4.30788510e-02 2.64312148e-01
7.99238443e-01 -3.36139947e-01 -8.81289124e-01 -1.72922850e-01
3.34904760e-01 6.67216659e-01 7.66857803e-01 3.17315668e-01
-8.61708462e-01 6.19455576e-01 6.16040897e+00 1.31498325e+00
-1.47724283e+00 1.65751159e-01 9.45953250e-01 -4.94705439e-01
-2.25389421e-01 -2.42074445e-01 -1.11481667e+00 7.05487549e-01
1.45396459e+00 -1.68549549e-02 3.05147827e-01 8.32554579e-01
2.02337191e-01 -3.43335599e-01 -1.44913340e+00 1.12393475e+00
-1.13413930e-01 -1.14081442e+00 -3.97249550e-01 3.06425784e-02
7.44823635e-01 1.61715075e-01 2.13035122e-01 5.33310056e-01
4.16520953e-01 -8.81672978e-01 8.73892367e-01 3.48019540e-01
1.44237268e+00 -9.05507684e-01 7.57496655e-01 5.48963249e-01
-8.40425730e-01 1.50481880e-01 -2.69725919e-01 9.45550576e-02
-2.56636858e-01 1.28154302e+00 -1.41053045e+00 1.45094946e-01
1.01480126e+00 1.83570072e-01 -3.82436693e-01 7.12189317e-01
-8.76837596e-02 1.26739681e+00 -6.37229443e-01 1.15958996e-01
-2.89758503e-01 6.95163846e-01 2.49782771e-01 1.20362592e+00
3.69373202e-01 -2.33280122e-01 3.48374963e-01 6.68918729e-01
-4.54751819e-01 -1.93546087e-01 -2.57322669e-01 9.31707677e-03
1.04592919e+00 1.22514153e+00 -6.18729651e-01 -4.55285400e-01
-8.45992044e-02 7.47872472e-01 1.87030703e-01 4.63851005e-01
-9.30472434e-01 -4.49520499e-01 5.84794939e-01 3.24881762e-01
4.52117056e-01 4.74309832e-01 3.03822696e-01 -6.70717597e-01
-2.01359484e-03 -7.71532059e-01 2.95261383e-01 -6.36625111e-01
-1.42116058e+00 2.89878041e-01 2.43685886e-01 -9.79225397e-01
-2.40109637e-01 -2.10670039e-01 -2.77518511e-01 1.04545152e+00
-1.19172597e+00 -5.87408721e-01 5.29321097e-02 4.11265820e-01
2.93839842e-01 2.70308293e-02 8.11412036e-01 6.60887480e-01
-2.46064827e-01 1.21289003e+00 8.52464214e-02 -4.91366684e-02
1.21279013e+00 -1.32797778e+00 1.34369940e-01 5.04156828e-01
1.29625902e-01 4.47043628e-01 7.16465950e-01 -5.23869514e-01
-1.11764050e+00 -1.51039147e+00 5.98705947e-01 -3.31165969e-01
6.33671939e-01 -7.95243621e-01 -9.94190991e-01 5.12448609e-01
-1.75490975e-01 6.52264237e-01 9.53065157e-01 8.85073841e-02
-4.26769912e-01 -4.50671732e-01 -1.40494418e+00 -2.13621221e-02
7.91868687e-01 -8.06982040e-01 -9.23274979e-02 3.69755358e-01
8.59683871e-01 -5.92250466e-01 -8.24443340e-01 5.28418660e-01
3.43060493e-01 -4.99098808e-01 3.45733702e-01 -5.41724324e-01
-2.89897531e-01 -5.71661174e-01 -2.14484349e-01 -1.52413213e+00
1.01516217e-01 -8.66341352e-01 -1.00714751e-01 1.65375745e+00
4.30430442e-01 -4.36157674e-01 6.76549137e-01 5.91353655e-01
-1.21879235e-01 -8.95271838e-01 -1.18546939e+00 -8.02711129e-01
-2.82249570e-01 -7.74107993e-01 5.20633936e-01 7.98642278e-01
-1.47842675e-01 3.90199721e-01 -2.64032781e-01 6.32902920e-01
3.02977651e-01 -3.44213605e-01 6.76808000e-01 -1.07954240e+00
-5.87426543e-01 1.42568260e-01 -3.67424965e-01 -1.03323972e+00
1.02888353e-01 -6.99132740e-01 3.99075329e-01 -8.34342241e-01
1.93995699e-01 -6.06188297e-01 -4.36154842e-01 7.69791782e-01
-5.36938794e-02 3.13253969e-01 -2.83058286e-01 -2.14635789e-01
-5.73381424e-01 4.16359544e-01 4.05706227e-01 -2.96013057e-01
-1.77712888e-01 2.53957421e-01 -6.92714214e-01 5.53679526e-01
5.62171936e-01 -9.04954195e-01 -6.26265645e-01 -7.31074512e-02
1.41011268e-01 -2.43629575e-01 2.14470029e-01 -1.07143164e+00
2.78273076e-01 2.19576254e-01 5.01054883e-01 -5.38975358e-01
-7.37080351e-03 -7.25228846e-01 3.37186120e-02 2.90540338e-01
-6.35393679e-01 -2.33134001e-01 5.15961111e-01 7.88890958e-01
9.53620858e-03 -1.65196016e-01 8.13847423e-01 4.26578879e-01
-2.07934424e-01 2.92926222e-01 -4.53050494e-01 2.55051106e-01
7.96022177e-01 3.54784101e-01 -1.24489218e-01 -4.99916971e-01
-8.76622558e-01 2.36877784e-01 9.45656933e-03 2.60526031e-01
2.15737015e-01 -1.05189896e+00 -4.29537773e-01 3.94982636e-01
2.18524918e-01 3.74696314e-01 2.71470577e-01 7.66365170e-01
2.12298587e-01 -1.87820438e-02 5.22356808e-01 -9.12977040e-01
-1.24294662e+00 2.29299203e-01 3.30076575e-01 -1.03698075e-01
-1.93470731e-01 1.39929390e+00 1.44532725e-01 -4.88362670e-01
6.92400277e-01 -8.00436497e-01 4.06047106e-01 1.50145665e-01
6.73503697e-01 2.61203289e-01 6.55352712e-01 5.50560839e-02
-3.00165713e-01 -6.87936917e-02 -2.80138493e-01 -3.34126651e-02
1.14304888e+00 5.02962023e-02 1.89568684e-01 7.96789467e-01
1.23637497e+00 3.77084404e-01 -1.44000793e+00 -4.61625934e-01
-2.13543847e-01 -4.06429380e-01 2.46174932e-01 -1.11305380e+00
-8.10049057e-01 7.74959266e-01 1.02321577e+00 1.75764740e-01
1.05118906e+00 2.78400093e-01 5.87222338e-01 1.64532319e-01
3.87300968e-01 -1.17183316e+00 1.42248541e-01 1.83215067e-01
2.31338546e-01 -9.33143258e-01 -1.09391376e-01 -1.25884652e-01
-3.77736688e-01 7.82788813e-01 6.82832778e-01 2.86828458e-01
8.78943622e-01 8.52661550e-01 7.13680033e-03 -1.28206477e-01
-8.58878970e-01 1.14050344e-01 9.65048745e-02 6.45961106e-01
1.10368893e-01 6.77372366e-02 4.79507208e-01 7.76654541e-01
-1.51316881e-01 -5.51411398e-02 3.43461931e-01 6.98201418e-01
-3.29343468e-01 -8.44110012e-01 -2.48960435e-01 8.11399639e-01
-5.17953157e-01 -1.53402135e-01 -1.52206104e-02 3.99086893e-01
4.37568933e-01 8.93810868e-01 4.47350055e-01 -4.00531501e-01
3.59198928e-01 2.26120159e-01 2.35100135e-01 -1.08270144e+00
-2.22334892e-01 3.47022921e-01 -8.61760974e-02 -4.85190630e-01
-6.56298473e-02 -6.55033827e-01 -1.41157281e+00 -5.87171353e-02
-6.55557036e-01 4.13880080e-01 4.70033973e-01 9.91310537e-01
5.15940785e-01 7.94142544e-01 7.76153207e-01 -4.13477480e-01
-9.49697137e-01 -1.09700096e+00 -1.02407682e+00 -1.01284429e-01
5.25811374e-01 -5.69888353e-01 -5.94774008e-01 6.19213283e-02] | [14.290728569030762, 5.861223220825195] |
73a78b09-a21c-48a4-8023-73674d3b82a7 | a-bayesian-approach-to-multimodal-visual | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Irie_A_Bayesian_Approach_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Irie_A_Bayesian_Approach_2013_CVPR_paper.pdf | A Bayesian Approach to Multimodal Visual Dictionary Learning | nary learning methods rely on image descriptors alone or together with class labels. However, Web images are often associated with text data which may carry substantial information regarding image semantics, and may be exploited for visual dictionary learning. This paper explores this idea by leveraging relational information between image descriptors and textual words via co-clustering, in addition to information of image descriptors. Existing co-clustering methods are not optimal for this problem because they ignore the structure of image descriptors in the continuous space, which is crucial for capturing visual characteristics of images. We propose a novel Bayesian co-clustering model to jointly estimate the underlying distributions of the continuous image descriptors as well as the relationship between such distributions and the textual words through a unified Bayesian inference. Extensive experiments on image categorization and retrieval have validated the substantial value of the proposed joint modeling in improving visual dictionary learning, where our model shows superior performance over several recent methods. | ['Shih-Fu Chang', 'Zhenguo Li', 'Go Irie', 'Dong Liu'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['image-categorization'] | ['computer-vision'] | [-1.24719724e-01 -5.73316455e-01 -7.25080013e-01 -4.45048153e-01
-6.29351735e-01 -6.31881654e-01 7.95856416e-01 6.15776658e-01
-4.91338462e-01 1.71626195e-01 2.35326603e-01 6.37048334e-02
-2.84744889e-01 -6.94595873e-01 -5.15478194e-01 -9.50572610e-01
3.16764712e-01 2.70028144e-01 2.92921271e-02 7.71355331e-02
3.30056131e-01 1.64931044e-01 -1.55481935e+00 1.57370657e-01
5.71838319e-01 9.84380662e-01 5.19261181e-01 1.44366756e-01
-5.05662262e-01 7.83786833e-01 -3.53234738e-01 -3.09649616e-01
1.02156736e-02 -1.18981190e-01 -3.97627980e-01 4.62103784e-01
4.60428536e-01 -2.61347771e-01 -5.77597797e-01 1.47785115e+00
3.87638211e-02 6.85172752e-02 1.04469955e+00 -1.24009144e+00
-9.69876111e-01 4.40246671e-01 -5.78151286e-01 -3.16917114e-02
1.73435107e-01 -3.59333247e-01 1.42047834e+00 -8.69284809e-01
5.84415972e-01 1.20604455e+00 4.84646261e-01 -9.71653536e-02
-1.20320296e+00 -5.98465145e-01 2.35890687e-01 4.97540057e-01
-2.02027917e+00 -2.18270451e-01 1.15097046e+00 -5.31138420e-01
4.48748231e-01 -2.76232567e-02 6.66251183e-01 8.83101046e-01
7.73612037e-02 9.43062365e-01 1.12029541e+00 -5.07554889e-01
-2.70245373e-02 6.05397284e-01 2.01871023e-01 8.47347319e-01
2.09996879e-01 -1.64576426e-01 -5.34000635e-01 -2.77959764e-01
7.27083325e-01 6.19565725e-01 -9.09734964e-02 -8.23672712e-01
-1.05967951e+00 1.09673989e+00 2.76626974e-01 4.50849324e-01
-4.43613052e-01 1.13084197e-01 2.35277563e-01 1.11612240e-02
3.69316518e-01 -1.62446395e-01 1.06824487e-02 2.16264039e-01
-7.99812973e-01 -4.02567871e-02 6.02197289e-01 1.32286108e+00
1.03536177e+00 -1.69588313e-01 -7.15835243e-02 1.07927084e+00
8.06556940e-01 8.13817322e-01 4.51793998e-01 -6.56955361e-01
2.19248161e-01 4.57247496e-01 -4.70043123e-02 -1.62664592e+00
2.08664536e-01 -9.12487879e-02 -6.51409686e-01 -3.79942626e-01
2.23961934e-01 5.38156509e-01 -8.06631982e-01 1.60620749e+00
1.51923388e-01 6.99991658e-02 -1.24903210e-01 6.59683466e-01
8.44560981e-01 5.56397855e-01 2.85347611e-01 -2.34162673e-01
1.39059460e+00 -5.20578921e-01 -9.24472809e-01 1.58322036e-01
3.41827869e-01 -9.34481144e-01 9.02073145e-01 3.24205399e-01
-6.41343236e-01 -6.22082174e-01 -9.94861007e-01 -1.56643748e-01
-4.19412076e-01 3.04054350e-01 6.39272332e-01 5.93002856e-01
-7.43316650e-01 9.77651402e-02 -6.99686944e-01 -3.28744262e-01
4.57008362e-01 6.48436323e-02 -3.24584693e-01 -4.62071896e-01
-9.70462024e-01 5.26468992e-01 3.93562227e-01 -2.12352186e-01
-8.52814794e-01 -2.37521365e-01 -9.55354989e-01 -1.92497093e-02
4.39967513e-01 -2.04338759e-01 6.66232467e-01 -7.29240060e-01
-1.01096237e+00 7.71226823e-01 -3.38125020e-01 -1.24177866e-01
1.07245870e-01 4.14206833e-02 -3.75789374e-01 5.30965567e-01
1.68589950e-01 7.18448460e-01 1.23715842e+00 -1.59523439e+00
-5.90166926e-01 -3.92868966e-01 3.88068892e-02 1.71076611e-01
-8.08158040e-01 -3.67431223e-01 -8.78838241e-01 -8.36122632e-01
2.52377629e-01 -8.15932810e-01 6.24615252e-02 1.87689483e-01
-2.46638656e-01 -5.25107265e-01 7.78202116e-01 -3.73126388e-01
1.25571120e+00 -2.37352800e+00 4.67093624e-02 5.99886835e-01
2.14905843e-01 -1.00240365e-01 -1.79361209e-01 5.88175058e-01
2.89937347e-01 -6.08051494e-02 1.11982353e-01 -3.53851616e-01
1.33743197e-01 6.17491901e-01 -4.45216864e-01 6.27433240e-01
-2.20219776e-01 7.78828442e-01 -9.78057325e-01 -1.06873655e+00
2.71616876e-01 6.23621047e-01 -3.67833376e-01 1.34954020e-01
-1.34861171e-01 1.43646225e-01 -7.02649593e-01 6.77421033e-01
5.19647539e-01 -2.99427181e-01 3.93902659e-01 -6.29403174e-01
1.37113184e-01 -1.22654386e-01 -1.20551825e+00 1.77843356e+00
-3.64782870e-01 3.85624737e-01 -1.30246356e-01 -1.38084853e+00
9.69775558e-01 2.16855496e-01 8.69952917e-01 -6.34502769e-01
5.58048524e-02 -1.92754030e-01 -4.24269944e-01 -5.42349577e-01
4.59857434e-01 -2.71569580e-01 2.88854819e-02 3.74793291e-01
2.50092924e-01 -8.90504271e-02 2.17030033e-01 5.60183942e-01
5.29808700e-01 -1.40069649e-01 2.32627079e-01 -2.38501593e-01
4.84119713e-01 -9.11130458e-02 3.60346466e-01 8.36492181e-01
3.07332389e-02 3.29264909e-01 3.03782403e-01 -6.50737286e-02
-1.06676435e+00 -1.03379512e+00 -4.43954557e-01 8.32801402e-01
5.53318799e-01 -7.71049321e-01 -2.91246742e-01 -6.79870903e-01
1.87307522e-01 4.18099523e-01 -6.78517342e-01 -2.28336781e-01
1.51547894e-01 -5.34802735e-01 3.06238294e-01 4.80212241e-01
2.51850247e-01 -5.59682012e-01 2.23702509e-02 1.14182629e-04
-3.48683178e-01 -1.27974510e+00 -6.94627702e-01 7.31285661e-02
-6.87969446e-01 -1.07146561e+00 -5.42719007e-01 -9.40618217e-01
7.77309775e-01 9.23169136e-01 8.67570162e-01 2.00550541e-01
-3.77657533e-01 1.05846870e+00 -5.71169853e-01 -1.80170864e-01
-1.43081576e-01 -2.73474514e-01 -5.91406934e-02 3.66074204e-01
7.71367490e-01 -4.69847709e-01 -4.25807118e-01 3.15356135e-01
-1.27642047e+00 -3.14940423e-01 6.71826065e-01 1.03110540e+00
8.89209986e-01 3.90039355e-01 1.64589092e-01 -9.31016088e-01
3.57348293e-01 -6.12143278e-01 -5.15008390e-01 5.03477752e-01
-7.81418681e-01 1.75692081e-01 1.81636795e-01 -6.77497506e-01
-9.93670344e-01 1.81381330e-01 3.22655052e-01 -7.53101885e-01
-1.80316463e-01 7.56014347e-01 -2.51673460e-01 -3.46839912e-02
1.18685290e-01 4.90101963e-01 1.36387914e-01 -5.61026990e-01
5.80309629e-01 7.31926978e-01 2.92123258e-01 -7.27704048e-01
7.48632789e-01 8.86107862e-01 6.83823824e-02 -1.03760409e+00
-9.29627836e-01 -1.15616572e+00 -8.77984285e-01 3.72219756e-02
9.27704394e-01 -1.22395968e+00 -4.40412223e-01 1.66824073e-01
-8.28898787e-01 3.28780323e-01 -2.00988259e-02 7.43594170e-01
-3.77437890e-01 9.69458342e-01 -5.21674752e-01 -5.70546448e-01
3.26073885e-01 -1.14701903e+00 1.15345955e+00 -1.73673093e-01
1.39263704e-01 -1.41046250e+00 -5.63848130e-02 4.76695836e-01
4.28310502e-03 -2.36895502e-01 1.22436130e+00 -6.16277277e-01
-5.10862112e-01 -4.03930247e-01 -4.10947323e-01 5.17057836e-01
2.96274185e-01 -1.83527410e-01 -6.04631186e-01 -2.98393577e-01
6.16702763e-03 -5.05548716e-01 8.82278025e-01 2.55757153e-01
1.22499609e+00 -1.14991844e-01 -4.18391496e-01 3.84107500e-01
1.78869796e+00 2.20914111e-02 4.54781502e-01 3.15577313e-02
9.71384585e-01 4.64072764e-01 6.43924773e-01 7.79376388e-01
6.94294393e-01 7.26751387e-01 1.95430845e-01 8.07843432e-02
-1.63541827e-02 -4.23947483e-01 2.17626065e-01 1.27229118e+00
2.86759198e-01 -2.63065491e-02 -7.66837895e-01 6.72855973e-01
-1.72702038e+00 -9.73589003e-01 -1.31757319e-01 1.94277227e+00
9.80833411e-01 -2.35061347e-01 -2.69082282e-02 -2.12549628e-03
7.31457293e-01 2.07356557e-01 -2.86651164e-01 3.34384471e-01
-2.58392632e-01 -2.48274729e-02 5.26754379e-01 2.80869961e-01
-1.16489375e+00 9.05113041e-01 6.44337511e+00 1.20643139e+00
-7.91456759e-01 1.84639573e-01 2.25343909e-02 4.11859900e-01
-5.15690088e-01 3.20231408e-01 -8.06098104e-01 4.14189279e-01
3.76050651e-01 -1.27729207e-01 1.59691870e-01 8.67904902e-01
-1.85849726e-01 -1.97461531e-01 -1.00442529e+00 1.24761903e+00
6.05815291e-01 -9.70790267e-01 4.92354125e-01 2.34215543e-01
8.56878340e-01 -3.16408455e-01 2.53487021e-01 1.18633315e-01
4.64067638e-01 -6.75370276e-01 6.57257915e-01 7.11707532e-01
5.12517393e-01 -6.79587483e-01 5.45237720e-01 2.37586319e-01
-1.23525405e+00 1.07024446e-01 -6.84768677e-01 3.96718860e-01
-1.80573553e-01 4.67104852e-01 -5.66714823e-01 4.30168539e-01
5.96467912e-01 1.40981162e+00 -6.99307323e-01 8.73253524e-01
-1.49779662e-01 4.47564274e-01 -2.11751804e-01 7.21154083e-03
3.17196578e-01 -3.63446176e-01 2.58106351e-01 1.19152951e+00
9.15837809e-02 -1.61195863e-02 7.13665962e-01 7.48537719e-01
1.88323315e-02 4.74914640e-01 -6.35494769e-01 -2.48094007e-01
6.57534838e-01 1.16658115e+00 -9.69831169e-01 -5.70428193e-01
-8.91423345e-01 7.21002042e-01 1.48240328e-01 4.54784870e-01
-6.73865259e-01 -2.50449568e-01 3.48206252e-01 -2.05066383e-01
6.95634544e-01 -6.38408303e-01 -2.86325645e-02 -1.33839524e+00
-8.66348445e-02 -7.25149453e-01 4.49353367e-01 -6.12655520e-01
-1.57052886e+00 1.36836708e-01 3.57566833e-01 -1.45596993e+00
1.34733126e-01 -5.72240472e-01 9.89297256e-02 4.05808270e-01
-1.69156611e+00 -1.55486381e+00 -2.76900321e-01 1.25759339e+00
3.74426782e-01 -3.01928371e-01 6.85239315e-01 4.31121826e-01
-1.64861262e-01 4.87703115e-01 6.25400603e-01 3.24190766e-01
8.49894226e-01 -1.19478321e+00 -5.88442087e-01 4.02993888e-01
8.97804081e-01 8.37639689e-01 3.58027816e-01 -5.99785447e-01
-1.64409912e+00 -8.67372632e-01 5.40381968e-01 -3.83238167e-01
9.20942962e-01 -2.66530186e-01 -8.99854898e-01 4.53006744e-01
-2.23998539e-02 5.52720800e-02 1.07555282e+00 1.88719943e-01
-8.37608516e-01 -1.16288140e-01 -6.46989584e-01 2.90489197e-01
6.34239554e-01 -1.17063904e+00 -5.70891678e-01 3.87069583e-01
3.53454709e-01 2.81023115e-01 -1.07791054e+00 2.86128409e-02
5.40622354e-01 -4.97528404e-01 1.16278112e+00 -3.07386518e-01
2.24285141e-01 -3.24651450e-01 -8.38046253e-01 -1.01355422e+00
-2.52360404e-01 1.60069093e-01 8.21890458e-02 1.56191814e+00
-5.29952273e-02 -7.85076767e-02 4.41266388e-01 2.33040512e-01
3.67608309e-01 -2.40668043e-01 -6.84337080e-01 -8.55914116e-01
4.72093858e-02 -5.87302089e-01 2.21650317e-01 1.21137154e+00
3.58092375e-02 3.24515224e-01 -6.15526736e-01 1.44582734e-01
1.06262863e+00 2.63890713e-01 6.15427673e-01 -1.11987233e+00
-1.68051586e-01 -3.20981294e-01 -6.95636809e-01 -1.25512481e+00
5.61262071e-01 -1.13727260e+00 1.23695053e-01 -1.23161459e+00
7.21438766e-01 -6.63605809e-01 -4.70416695e-01 3.50356966e-01
-1.87738210e-01 2.67190665e-01 1.30675435e-01 8.76670659e-01
-9.76338506e-01 7.08981514e-01 9.95771289e-01 -4.81561035e-01
2.79609501e-01 -3.55964124e-01 -4.64416832e-01 6.98163807e-01
2.80171067e-01 -6.33904278e-01 -6.65224016e-01 -4.39807266e-01
1.01958305e-01 -2.14584023e-02 4.74305898e-01 -5.80298126e-01
5.29249251e-01 -2.77629375e-01 2.46116444e-01 -6.16014421e-01
3.50898951e-01 -1.28183222e+00 -1.45724073e-01 6.64434731e-02
-5.18669486e-01 -3.13054882e-02 -3.59573692e-01 1.27362752e+00
-6.52784884e-01 -3.44415247e-01 6.33963823e-01 -1.55355901e-01
-9.41841900e-01 3.88390094e-01 -2.98945457e-01 -1.84685260e-01
8.63724291e-01 -5.23263626e-02 2.16526285e-01 -6.23623550e-01
-8.82072031e-01 5.60587347e-02 4.74914551e-01 4.70716357e-01
8.37495446e-01 -1.56689894e+00 -2.57116139e-01 1.48860723e-01
6.31294191e-01 -4.66036946e-01 1.51819229e-01 7.21320570e-01
-6.80172071e-02 4.67528582e-01 -7.88379014e-02 -9.46444035e-01
-1.34667563e+00 9.81577754e-01 -2.86027879e-01 -1.89081894e-03
-4.82316762e-01 4.33871627e-01 2.28063077e-01 9.23632309e-02
4.51655149e-01 -5.96025772e-02 -4.38957036e-01 3.51805955e-01
2.99470961e-01 -4.12185602e-02 -2.40326673e-01 -1.01462436e+00
-1.01650596e-01 8.01789522e-01 -3.08751792e-01 -1.69014782e-01
1.10389543e+00 -6.33297920e-01 -2.79115260e-01 7.78889477e-01
1.55297816e+00 -6.72284514e-02 -9.70727384e-01 -1.10941827e+00
1.36114806e-01 -8.05668652e-01 4.50350821e-01 -2.03522667e-01
-1.15688324e+00 1.01418483e+00 5.35449326e-01 8.56394097e-02
9.11287606e-01 2.31908292e-01 6.06866896e-01 4.68158722e-01
5.59361100e-01 -1.07980919e+00 5.63128471e-01 1.79166496e-01
4.21504468e-01 -1.56034696e+00 4.07272041e-01 -4.26623285e-01
-7.49767721e-01 1.05971837e+00 1.80256039e-01 -2.34704971e-01
1.25530124e+00 -1.73807189e-01 -1.95594300e-02 -3.91819775e-01
-5.91485322e-01 -4.13693309e-01 5.57640731e-01 5.99667966e-01
2.33963549e-01 -3.32192518e-02 -2.34717026e-01 3.90018791e-01
3.76980722e-01 -4.37275738e-01 1.31427348e-01 1.01756740e+00
-3.53280514e-01 -1.40105915e+00 -2.91393757e-01 2.57436991e-01
-4.59630251e-01 -1.97777733e-01 -1.34294614e-01 6.41815126e-01
1.78333178e-01 9.40149248e-01 9.61921886e-02 -2.71137565e-01
-8.35802034e-02 -1.37390196e-01 4.52706069e-01 -6.22573018e-01
1.58236772e-01 5.25706291e-01 -4.93909031e-01 -1.83746427e-01
-9.19006348e-01 -8.80563378e-01 -8.96884859e-01 -5.01967594e-02
-4.76373553e-01 3.69697750e-01 8.04472446e-01 1.00183511e+00
4.33208421e-02 1.51663804e-02 7.38991439e-01 -5.12500226e-01
-6.42502069e-01 -6.95738018e-01 -1.02439511e+00 8.80125761e-01
1.32883862e-01 -9.93313611e-01 -1.82376996e-01 5.48812032e-01] | [10.403409957885742, 1.1571249961853027] |
fe8679f2-4639-4e56-98aa-ebd41a88ba7b | non-parametric-probabilistic-time-series | 2306.03782 | null | https://arxiv.org/abs/2306.03782v1 | https://arxiv.org/pdf/2306.03782v1.pdf | Non-parametric Probabilistic Time Series Forecasting via Innovations Representation | Probabilistic time series forecasting predicts the conditional probability distributions of the time series at a future time given past realizations. Such techniques are critical in risk-based decision-making and planning under uncertainties. Existing approaches are primarily based on parametric or semi-parametric time-series models that are restrictive, difficult to validate, and challenging to adapt to varying conditions. This paper proposes a nonparametric method based on the classic notion of {\em innovations} pioneered by Norbert Wiener and Gopinath Kallianpur that causally transforms a nonparametric random process to an independent and identical uniformly distributed {\em innovations process}. We present a machine-learning architecture and a learning algorithm that circumvent two limitations of the original Wiener-Kallianpur innovations representation: (i) the need for known probability distributions of the time series and (ii) the existence of a causal decoder that reproduces the original time series from the innovations representation. We develop a deep-learning approach and a Monte Carlo sampling technique to obtain a generative model for the predicted conditional probability distribution of the time series based on a weak notion of Wiener-Kallianpur innovations representation. The efficacy of the proposed probabilistic forecasting technique is demonstrated on a variety of electricity price datasets, showing marked improvement over leading benchmarks of probabilistic forecasting techniques. | ['Lang Tong', 'Qing Zhao', 'Meijen Lee', 'Xinyi Wang'] | 2023-06-05 | null | null | null | null | ['probabilistic-time-series-forecasting'] | ['time-series'] | [-1.38697866e-02 -2.12228537e-01 3.51268947e-01 -3.58399779e-01
-7.21640170e-01 -6.91083133e-01 1.15573490e+00 -6.80742413e-02
1.06289513e-01 8.56385410e-01 3.46059889e-01 -7.34322786e-01
-4.37193334e-01 -1.07718277e+00 -8.28090549e-01 -8.83145332e-01
-3.47161800e-01 6.80895209e-01 -5.66974394e-02 1.77386831e-02
1.75991938e-01 5.74695766e-01 -1.31301606e+00 -2.64809936e-01
6.38272762e-01 1.12293863e+00 -3.59668732e-02 6.91854775e-01
-3.06979954e-01 6.16517663e-01 -7.76458204e-01 -2.47253641e-01
2.49668106e-01 -4.73435760e-01 -1.59043521e-01 -5.25760293e-01
-5.25479555e-01 -3.43097538e-01 -4.00762290e-01 8.50628078e-01
3.09443593e-01 2.16011629e-01 1.29291189e+00 -1.35830510e+00
-7.27917016e-01 1.00826418e+00 -5.20268798e-01 3.85692596e-01
-5.92981353e-02 6.02180921e-02 7.15919375e-01 -5.31690180e-01
-2.55780518e-02 1.12078249e+00 1.00340223e+00 -2.06906665e-02
-1.24684751e+00 -6.68619633e-01 1.93313193e-02 5.40964603e-02
-1.36994362e+00 8.82648677e-02 1.07852471e+00 -7.74398625e-01
9.79131579e-01 -5.41784242e-02 4.66579437e-01 1.28728545e+00
9.39146280e-01 4.14103299e-01 1.16913891e+00 -3.46038103e-01
7.34938622e-01 -4.60192233e-01 -2.22259343e-01 -5.35897240e-02
-1.50311634e-01 7.12737381e-01 -2.55045235e-01 -4.86126661e-01
7.75274336e-01 3.46222937e-01 8.75864029e-02 -1.38563126e-01
-9.57842648e-01 1.03083932e+00 1.25308082e-01 4.66634899e-01
-8.15696955e-01 6.68138921e-01 2.48377383e-01 1.95117429e-01
7.14406967e-01 -1.80025309e-01 -7.07155824e-01 -2.40609542e-01
-1.28429592e+00 3.01926225e-01 1.11010075e+00 6.45193636e-01
3.02717805e-01 7.40102470e-01 -2.42659122e-01 7.85780102e-02
5.56682825e-01 1.01507020e+00 4.98901010e-01 -4.39174563e-01
4.19099927e-02 -2.93245763e-01 4.85371560e-01 -8.35520089e-01
-4.04273629e-01 -5.21529496e-01 -9.33228910e-01 1.69213682e-01
4.23207462e-01 -4.48794663e-01 -1.10747838e+00 1.60102403e+00
-1.41004845e-01 7.53705561e-01 4.32168841e-02 2.46988609e-01
6.42689541e-02 1.15296197e+00 1.23566151e-01 -1.91472873e-01
7.09820211e-01 -2.90317148e-01 -5.51311433e-01 1.88761160e-01
2.46817563e-02 -6.81066751e-01 4.50318098e-01 1.82027206e-01
-6.71430290e-01 -4.46553856e-01 -7.16248035e-01 6.72741234e-01
-4.77562428e-01 -2.80747712e-01 5.36246777e-01 8.85235071e-01
-9.63865221e-01 8.46242905e-01 -9.44019258e-01 1.29443303e-01
1.08808398e-01 -1.36163533e-01 3.02189648e-01 4.27174240e-01
-1.31160855e+00 1.00363529e+00 5.30313671e-01 2.80357987e-01
-1.15623009e+00 -8.62993062e-01 -5.42222857e-01 3.63092870e-01
-1.66479543e-01 -4.66134042e-01 1.31706953e+00 -7.53521860e-01
-1.75598538e+00 -9.60683078e-02 4.36801761e-02 -9.83487844e-01
6.39594674e-01 -1.59388050e-01 -8.81166279e-01 -2.97504902e-01
-1.26149297e-01 -9.15629864e-02 1.26214194e+00 -9.25827682e-01
-4.38030243e-01 2.45968755e-02 -5.53200960e-01 -3.57220888e-01
1.90650612e-01 -2.60029346e-01 4.21546549e-01 -1.25815344e+00
1.87126786e-01 -7.55068123e-01 -3.08220774e-01 -5.56404591e-01
-2.54980683e-01 -2.77183682e-01 5.30575335e-01 -8.82036090e-01
9.98472452e-01 -1.78892517e+00 -1.72043934e-01 5.75113416e-01
-1.99058115e-01 -7.94907287e-02 6.67505935e-02 9.20093596e-01
-4.87096906e-01 4.52221744e-02 -4.55079228e-01 -2.90726945e-02
2.50871241e-01 2.28624105e-01 -1.29774392e+00 5.10724545e-01
1.83297038e-01 9.69142377e-01 -9.97919083e-01 3.72570068e-01
4.61776406e-01 6.79405212e-01 1.16692506e-01 9.68538374e-02
-4.92642820e-01 7.21489191e-01 -2.77705103e-01 3.49054903e-01
6.43805087e-01 -6.27124682e-02 -2.30780914e-01 5.24987243e-02
-1.61640778e-01 1.68733969e-01 -1.06433010e+00 1.09729266e+00
-5.00435710e-01 3.11579973e-01 -7.79196441e-01 -1.18582141e+00
1.26795650e+00 5.57577610e-01 5.36443472e-01 -2.68072397e-01
1.38971418e-01 3.44466329e-01 -2.38658831e-01 -5.49052097e-02
1.34358153e-01 -7.03645825e-01 -2.80310661e-01 8.89550984e-01
1.62634969e-01 -4.58360314e-01 -1.70036495e-01 -2.31890455e-01
1.06734431e+00 4.59800631e-01 2.93641776e-01 -1.91871569e-01
3.06654274e-01 -4.22490031e-01 5.14695585e-01 9.35253859e-01
1.39662653e-01 4.67693418e-01 5.46067059e-01 -7.29983687e-01
-1.18659616e+00 -1.77051008e+00 -2.10646182e-01 7.37285137e-01
-5.66016078e-01 1.46887064e-01 -6.00287914e-02 -4.35428232e-01
2.97223240e-01 1.84525585e+00 -6.84249938e-01 -1.65170282e-01
-1.46706909e-01 -9.23161507e-01 4.81348664e-01 6.78638995e-01
-4.20020074e-02 -9.05794382e-01 -3.66976321e-01 6.98502719e-01
2.35142335e-01 -5.00216246e-01 -7.26212859e-02 3.15121472e-01
-7.81666100e-01 -5.14862537e-01 -9.65101302e-01 -1.53168261e-01
2.54469216e-01 -3.34389299e-01 1.04050195e+00 -8.77614617e-01
1.52203754e-01 5.95717728e-01 -2.32428387e-01 -9.37836766e-01
-6.64102614e-01 -5.41892767e-01 3.39709744e-02 2.78390422e-02
3.12010080e-01 -1.09427798e+00 -3.69204372e-01 4.47012484e-02
-8.38342130e-01 -3.00680429e-01 4.90789831e-01 7.65944600e-01
4.63122755e-01 7.10525155e-01 1.02656901e+00 -3.50879490e-01
5.68768203e-01 -8.81341696e-01 -9.14480150e-01 2.00307205e-01
-6.40265226e-01 2.17788443e-01 7.38715410e-01 -4.21930820e-01
-1.20729327e+00 -1.04703777e-01 -6.08217567e-02 -5.03786325e-01
3.09100915e-02 7.97509313e-01 3.15280259e-01 3.70428115e-01
2.63065368e-01 6.33416891e-01 -2.49699607e-01 -4.94754285e-01
6.12179399e-01 3.03862393e-01 7.37706780e-01 -6.12493873e-01
9.83052254e-01 6.01780951e-01 1.40339181e-01 -4.32533711e-01
-5.90037644e-01 -1.30856737e-01 -6.45734608e-01 -2.21287221e-01
5.25078535e-01 -8.71042192e-01 -4.04073328e-01 6.81720018e-01
-1.17995536e+00 5.32601774e-02 -6.56156719e-01 6.88659489e-01
-7.94411838e-01 1.65554240e-01 -3.67873877e-01 -1.27329743e+00
-3.28933716e-01 -5.29380798e-01 9.13265467e-01 1.11162432e-01
-1.71310797e-01 -1.46336913e+00 4.37025636e-01 -4.28125441e-01
7.93847740e-01 4.49652016e-01 1.09754646e+00 -8.52854669e-01
-2.52364159e-01 -7.35295773e-01 5.14415987e-02 5.24137855e-01
1.11196138e-01 1.68170348e-01 -7.89243698e-01 -9.36070457e-02
2.51173973e-01 4.19972509e-01 7.60129929e-01 7.33771682e-01
1.00262439e+00 -1.78064689e-01 -2.38820150e-01 3.85947853e-01
1.37918437e+00 6.20214403e-01 4.68559325e-01 4.64304313e-02
3.56515348e-01 2.81860769e-01 -6.96995333e-02 8.96633804e-01
4.53964382e-01 -1.47859707e-01 2.44440153e-01 5.77894032e-01
4.62448627e-01 -6.16293371e-01 4.49273229e-01 8.41162026e-01
-1.48526430e-01 -4.56442624e-01 -1.19879317e+00 7.27837443e-01
-1.70319986e+00 -1.20735729e+00 1.34845739e-02 2.23106766e+00
5.17685413e-01 2.43390694e-01 -2.43618280e-01 5.54556847e-02
7.09914148e-01 2.96692431e-01 -6.77715838e-01 -3.21515560e-01
-1.66649014e-01 5.75387418e-01 5.61500788e-01 2.26279706e-01
-1.16696966e+00 4.46382344e-01 6.84531927e+00 6.99246168e-01
-1.11715782e+00 1.19573414e-01 6.15218341e-01 2.94670314e-01
-7.03820467e-01 1.16282322e-01 -5.02750397e-01 8.31457436e-01
1.51352274e+00 -8.57807934e-01 2.60961235e-01 5.62344551e-01
3.53844136e-01 7.56082535e-02 -1.10081112e+00 7.93330669e-01
-2.73662299e-01 -1.30612886e+00 1.20627254e-01 -2.73982018e-01
8.86249542e-01 2.58644670e-01 2.51317441e-01 4.11042094e-01
9.36073422e-01 -1.14501917e+00 9.51270938e-01 1.30876040e+00
3.00276995e-01 -8.95158708e-01 8.84430587e-01 3.88396621e-01
-1.06477630e+00 -2.11314321e-01 -2.10352227e-01 -1.69389248e-01
5.91884613e-01 1.14564550e+00 -7.55094767e-01 7.06955254e-01
6.70370877e-01 5.16915977e-01 1.33742705e-01 9.45568025e-01
-4.40391183e-01 1.10031176e+00 -6.03170812e-01 4.68650721e-02
2.20509022e-01 -2.54144400e-01 6.60603940e-01 9.14773881e-01
1.03933418e+00 -2.11724773e-01 -7.97200948e-02 1.10548794e+00
3.47022980e-01 -4.53044623e-01 -4.91596848e-01 -4.50440526e-01
5.83753407e-01 7.39222288e-01 -8.02115023e-01 -2.26132467e-01
-3.77636194e-01 5.30296624e-01 -3.74795049e-01 6.41207457e-01
-9.81076658e-01 -8.87392983e-02 1.50037512e-01 -1.20935291e-01
7.71360040e-01 -5.63437879e-01 -3.16771716e-01 -8.03962588e-01
-1.29991189e-01 -2.60373533e-01 4.34460849e-01 -8.69205177e-01
-2.01505518e+00 5.45487165e-01 2.59788454e-01 -1.26300263e+00
-9.50501442e-01 -2.67651856e-01 -1.08365548e+00 1.59172022e+00
-1.45532370e+00 -1.12674248e+00 3.61413300e-01 4.70870584e-01
2.02990785e-01 -3.86636138e-01 9.92391706e-01 -3.83898169e-01
-8.79715234e-02 -9.17334929e-02 9.09058571e-01 -2.40525752e-01
2.42297441e-01 -1.27930546e+00 9.32619214e-01 9.18753684e-01
-4.15762654e-03 3.60686183e-01 9.77129221e-01 -8.62326145e-01
-1.14317012e+00 -1.17363107e+00 7.59275436e-01 -3.06460559e-01
1.31482399e+00 -8.01157951e-02 -9.88113880e-01 9.27383780e-01
1.19108625e-01 -1.68253407e-01 6.21530771e-01 -1.68704107e-01
-3.50895435e-01 -6.26455769e-02 -1.13496113e+00 2.91083008e-01
1.53833926e-01 -5.74041128e-01 -1.04466617e+00 3.56888592e-01
5.00417948e-01 4.41112230e-03 -8.66800547e-01 2.95353293e-01
5.48515916e-01 -7.46846199e-01 9.10809636e-01 -3.50706697e-01
2.15415314e-01 -4.90240067e-01 -1.88038871e-01 -1.65620077e+00
-3.58210564e-01 -9.05627549e-01 -4.56376553e-01 1.18473566e+00
1.94545895e-01 -1.08638680e+00 3.15969676e-01 4.82385337e-01
1.47368744e-01 -3.91346544e-01 -1.38562369e+00 -1.01365674e+00
4.96545047e-01 -9.01322842e-01 1.01661682e+00 7.26048172e-01
-4.92065072e-01 -2.36598894e-01 -2.89486468e-01 4.83689636e-01
7.88646877e-01 2.29825392e-01 3.26642007e-01 -1.25238633e+00
-3.79953057e-01 -5.20611286e-01 -3.25358957e-01 -6.53382182e-01
1.46772385e-01 -6.50472581e-01 1.52753890e-01 -1.41876352e+00
-3.63013536e-01 -4.02906001e-01 -6.96140170e-01 2.28429303e-01
1.18050106e-01 -3.22567105e-01 -8.06510374e-02 3.13280970e-01
1.66321501e-01 9.76293564e-01 5.32228589e-01 4.32850309e-02
-3.40777338e-02 5.45031071e-01 -3.25963587e-01 8.82958531e-01
8.39230359e-01 -4.53457117e-01 -5.52478492e-01 -1.20121755e-01
3.37112993e-01 3.97502154e-01 5.48647225e-01 -9.80730832e-01
2.48768970e-01 -1.57433778e-01 6.01182401e-01 -1.15813255e+00
9.36710611e-02 -6.79694653e-01 7.61042118e-01 4.48885739e-01
-6.85847551e-02 2.64490336e-01 2.12111548e-01 1.09914911e+00
-8.04804638e-02 -9.37710255e-02 3.49856228e-01 3.65555054e-04
-7.59687483e-01 1.59524545e-01 -6.48704469e-01 -3.28068525e-01
1.08091533e+00 2.05232397e-01 -9.62454155e-02 -8.17436218e-01
-8.14460754e-01 2.70251539e-02 -1.00695439e-01 4.89055991e-01
3.91832262e-01 -1.16471612e+00 -8.87262702e-01 1.94746554e-01
-3.27949077e-01 -4.56387520e-01 2.70447791e-01 5.17229140e-01
-3.23380709e-01 5.86165726e-01 -1.28383994e-01 -5.59326172e-01
-6.07973300e-02 6.47180676e-01 3.00265372e-01 -2.59054869e-01
-7.18517661e-01 7.07573771e-01 -5.42813614e-02 -4.86190140e-01
-1.84398338e-01 -5.15437007e-01 8.93168896e-03 1.12536967e-01
5.14422119e-01 3.32175583e-01 6.59764698e-03 -5.89826524e-01
-1.91988662e-01 2.64514595e-01 2.27352276e-01 -4.07197505e-01
1.64829195e+00 -1.17486820e-01 -1.26257420e-01 1.12785983e+00
7.99123466e-01 -3.67266774e-01 -1.51154363e+00 -3.10587496e-01
2.17127964e-01 -1.52169764e-01 1.12184346e-01 -1.04692233e+00
-7.13326752e-01 8.30859661e-01 5.56309700e-01 5.67362189e-01
9.13529456e-01 -2.05370963e-01 7.90923297e-01 6.08661026e-02
5.22014916e-01 -7.82812953e-01 -5.37020683e-01 6.50181353e-01
1.06131446e+00 -7.12025762e-01 -1.46192700e-01 2.00921044e-01
-3.72865319e-01 1.42696714e+00 -4.41786170e-01 -3.37072551e-01
1.44673705e+00 5.42252481e-01 -1.38645709e-01 1.71025127e-01
-7.06387579e-01 -6.30874634e-02 3.51161152e-01 7.24266410e-01
2.84171671e-01 4.69901502e-01 -6.99370215e-03 9.78081346e-01
-4.87541050e-01 1.37138531e-01 3.86328995e-01 9.94661629e-01
-1.47065550e-01 -7.40014195e-01 -5.94850659e-01 6.16514802e-01
-2.67017752e-01 -2.31365323e-01 4.55552965e-01 3.68748486e-01
-2.04175636e-01 7.70465314e-01 4.62958604e-01 -8.56687576e-02
4.83460836e-02 4.75635976e-01 1.29153013e-01 -1.57475188e-01
-5.24735041e-02 3.33374023e-01 -4.40892637e-01 -9.92227718e-02
-1.75142273e-01 -1.00116336e+00 -1.08029139e+00 -3.19339126e-01
6.05129413e-02 1.14314906e-01 8.42690527e-01 1.19635129e+00
2.55418450e-01 7.39922404e-01 7.96592772e-01 -1.09192014e+00
-1.07478642e+00 -9.64298904e-01 -8.96066368e-01 -7.40392134e-02
2.04727322e-01 -6.40971959e-01 -6.43310010e-01 1.25681356e-01] | [6.863265514373779, 3.2465429306030273] |
7e1180a3-bf40-4622-b68d-64fc2c76fbec | object-wake-up-3-d-object-reconstruction | 2108.02708 | null | https://arxiv.org/abs/2108.02708v3 | https://arxiv.org/pdf/2108.02708v3.pdf | Object Wake-up: 3D Object Rigging from a Single Image | Given a single image of a general object such as a chair, could we also restore its articulated 3D shape similar to human modeling, so as to animate its plausible articulations and diverse motions? This is an interesting new question that may have numerous downstream augmented reality and virtual reality applications. Comparing with previous efforts on object manipulation, our work goes beyond 2D manipulation and rigid deformation, and involves articulated manipulation. To achieve this goal, we propose an automated approach to build such 3D generic objects from single images and embed articulated skeletons in them. Specifically, our framework starts by reconstructing the 3D object from an input image. Afterwards, to extract skeletons for generic 3D objects, we develop a novel skeleton prediction method with a multi-head structure for skeleton probability field estimation by utilizing the deep implicit functions. A dataset of generic 3D objects with ground-truth annotated skeletons is collected. Empirically our approach is demonstrated with satisfactory performance on public datasets as well as our in-house dataset; our results surpass those of the state-of-the-arts by a noticeable margin on both 3D reconstruction and skeleton prediction. | ['Minglun Gong', 'Xingyu Li', 'Zhenbo Yu', 'Xinxin Zuo', 'Ji Yang', 'Li Cheng', 'Bingbing Ni', 'Sen Wang'] | 2021-08-05 | null | null | null | null | ['3d-object-reconstruction', 'object-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 4.21419382e-01 5.40956199e-01 -7.41551444e-02 -9.01208445e-02
-7.21064508e-01 -4.60567802e-01 4.99743640e-01 -6.11515582e-01
5.86636662e-02 3.55514586e-01 2.76629746e-01 -4.97725047e-02
4.73626107e-02 -5.47560692e-01 -1.00225890e+00 -3.36747199e-01
1.62847817e-01 9.60726321e-01 4.75351125e-01 -1.51792958e-01
1.30791003e-02 8.62099648e-01 -1.54081798e+00 1.05819451e-02
6.04703903e-01 6.58488929e-01 2.60475397e-01 6.72385454e-01
8.53253230e-02 3.32923353e-01 -2.00421706e-01 -7.87011385e-01
4.32237059e-01 -1.84791237e-01 -8.21038127e-01 7.53430903e-01
7.89609969e-01 -5.23027539e-01 -4.61018860e-01 7.16486931e-01
5.76167881e-01 1.11960307e-01 7.22581804e-01 -8.30133617e-01
-4.87609982e-01 4.46032882e-01 -8.56335163e-01 -4.51171935e-01
7.32815802e-01 3.95469129e-01 8.39570165e-01 -1.02013540e+00
9.77759480e-01 1.48834944e+00 6.78723812e-01 9.63050663e-01
-1.25224638e+00 -3.37255806e-01 2.68552661e-01 -3.16230178e-01
-1.14264488e+00 -4.66640979e-01 9.88264918e-01 -5.52074075e-01
6.04189992e-01 2.92622298e-01 1.00333846e+00 1.26310933e+00
-5.79067431e-02 1.07340181e+00 7.34139860e-01 -3.86807323e-01
-2.54493803e-02 -4.87849027e-01 -1.68896288e-01 1.07574344e+00
1.60287529e-01 -9.70663428e-02 -4.48693871e-01 -2.51706958e-01
1.24384379e+00 -6.15387447e-02 -3.25726718e-01 -9.58420455e-01
-1.55029273e+00 3.42401892e-01 1.95790365e-01 -2.97289968e-01
-5.95065296e-01 6.21153116e-01 4.91352305e-02 -3.99980307e-01
3.45032066e-01 5.61008137e-03 -5.51935375e-01 6.39200881e-02
-9.21999872e-01 6.10505223e-01 6.35485053e-01 1.08954656e+00
4.54498082e-01 1.04425170e-01 7.96516687e-02 6.22231066e-01
6.03770494e-01 4.43476349e-01 1.56402990e-01 -1.43699431e+00
3.20696980e-01 5.35298705e-01 1.67971998e-01 -5.38139701e-01
-2.99487442e-01 -3.07802737e-01 -6.22275531e-01 4.50238049e-01
5.50125599e-01 2.85524935e-01 -1.19294381e+00 1.45408511e+00
9.76827860e-01 4.54535663e-01 -2.42363244e-01 9.91894603e-01
9.40138996e-01 2.76861936e-01 -1.65891245e-01 1.59074947e-01
1.43190908e+00 -1.00415230e+00 -4.19128716e-01 -1.88896880e-01
1.61708802e-01 -7.98473835e-01 1.20408154e+00 5.27603507e-01
-1.63480818e+00 -4.98266339e-01 -7.88228035e-01 -3.25576186e-01
3.39135408e-01 -1.61795076e-02 7.73145974e-01 6.22806072e-01
-7.10593581e-01 8.08290720e-01 -1.19216335e+00 -2.96928734e-01
5.97430825e-01 3.35890144e-01 -5.89270711e-01 -6.30424693e-02
-4.86888468e-01 8.66454840e-01 1.08137816e-01 1.48799002e-01
-1.00223839e+00 -7.74333835e-01 -9.28511441e-01 -2.80226439e-01
4.59633738e-01 -1.58777511e+00 1.32640278e+00 -3.88914555e-01
-1.79213917e+00 1.29170275e+00 1.67610809e-01 -1.43256947e-01
9.06402171e-01 -4.02560711e-01 2.21575961e-01 2.04381153e-01
-4.85464446e-02 7.61030972e-01 8.94282043e-01 -1.64782238e+00
-8.33405033e-02 -7.17968762e-01 1.75793752e-01 3.33193779e-01
2.50092477e-01 -5.67907766e-02 -1.01850581e+00 -8.74682844e-01
6.74158275e-01 -1.12902856e+00 -5.21074831e-01 7.45226264e-01
-6.84832036e-01 2.07013652e-01 7.34969020e-01 -9.39172149e-01
6.04049444e-01 -1.78041673e+00 5.91461539e-01 1.91925019e-02
2.30289489e-01 -3.97319067e-03 4.56480682e-03 -1.05371006e-01
-1.77782811e-02 -5.49840853e-02 -5.40075362e-01 -6.97099864e-01
1.07945949e-01 4.04274642e-01 -9.63436067e-02 6.27455890e-01
1.26823589e-01 1.05411303e+00 -8.43770981e-01 -7.58205652e-01
2.79542655e-01 7.03811228e-01 -9.88297582e-01 2.38672286e-01
-4.46818799e-01 8.64914715e-01 -5.88064551e-01 1.08105791e+00
6.10101879e-01 -2.30127722e-01 1.15388773e-01 -2.88719863e-01
2.77972341e-01 4.50990424e-02 -1.39001775e+00 2.25554585e+00
-2.72077680e-01 9.28089209e-03 2.73274004e-01 -7.31712520e-01
7.96752572e-01 3.37598383e-01 8.26617658e-01 5.47383986e-02
-2.11798716e-02 6.79582879e-02 -3.16317976e-01 -5.69385290e-01
4.12416190e-01 -2.26041973e-01 4.71525975e-02 2.73367345e-01
-1.52711319e-02 -8.80240977e-01 -4.32895303e-01 6.62153810e-02
1.03239703e+00 1.09591770e+00 1.56888172e-01 1.40962169e-01
2.18711182e-01 -3.13238725e-02 2.66278803e-01 2.39120841e-01
1.94154054e-01 1.11051905e+00 5.36598302e-02 -4.15258139e-01
-1.10590613e+00 -1.40593541e+00 -6.77747130e-02 4.69212264e-01
3.31082195e-01 -2.13324368e-01 -8.67830753e-01 -5.61419368e-01
2.05571771e-01 5.66666245e-01 -4.65671510e-01 7.99942613e-02
-8.00058365e-01 -5.20426989e-01 3.73474389e-01 5.52381694e-01
2.70856947e-01 -9.43846226e-01 -7.44292676e-01 1.60277292e-01
-2.40459636e-01 -1.23001719e+00 -4.10750955e-01 -1.56777263e-01
-1.26630998e+00 -1.12381029e+00 -9.55979347e-01 -5.71944535e-01
7.04007924e-01 2.51498580e-01 1.22098088e+00 1.38263270e-01
-4.60254103e-01 6.63810790e-01 2.41787676e-02 -2.35853106e-01
-3.93896669e-01 -1.88216776e-01 -5.00285588e-02 -1.20767318e-01
-4.01351988e-01 -8.98060739e-01 -7.08870947e-01 3.89090866e-01
-8.11510503e-01 3.21420431e-01 4.85844523e-01 5.57757616e-01
8.78790736e-01 -3.67316663e-01 4.95649204e-02 -6.56559050e-01
-4.53991853e-02 -2.62308389e-01 -3.53004366e-01 -2.14443426e-03
3.88066657e-02 -4.77691041e-03 -1.83137152e-02 -6.51766181e-01
-1.22254229e+00 7.34099567e-01 -3.86514157e-01 -7.93567538e-01
-1.12181738e-01 7.19906762e-02 -5.20195484e-01 9.06064510e-02
5.96214235e-01 -5.39759584e-02 7.91356862e-02 -9.93102252e-01
6.48006797e-01 2.45147794e-01 8.78647566e-01 -1.01583600e+00
1.14502966e+00 8.50814283e-01 1.70113519e-01 -7.56764293e-01
-8.62820506e-01 -1.46355212e-01 -1.13864672e+00 -2.62512863e-01
9.06214058e-01 -9.08317924e-01 -7.17790663e-01 4.16312844e-01
-1.48192000e+00 -5.28912842e-01 -5.85091114e-01 4.89463687e-01
-1.06231594e+00 6.44657850e-01 -5.04945397e-01 -7.14083672e-01
-1.77909762e-01 -1.23162043e+00 1.82035291e+00 -1.88521296e-01
-3.34751755e-01 -6.46945536e-01 2.45221518e-02 9.45999146e-01
-2.63607383e-01 8.52435708e-01 8.02573025e-01 6.54819310e-02
-9.39417720e-01 -1.46781370e-01 2.53542602e-01 1.27517343e-01
-2.72646602e-02 3.29466969e-01 -8.23316276e-01 -6.03542943e-03
-2.25775152e-01 -1.61083356e-01 5.91582835e-01 4.23312277e-01
1.35559297e+00 -2.07211211e-01 -4.05445397e-01 8.18250835e-01
9.62753296e-01 -2.79375166e-01 7.11373508e-01 1.22438289e-01
1.04660702e+00 6.27861857e-01 6.11163437e-01 5.83185196e-01
5.20462155e-01 1.08122981e+00 9.60365176e-01 -5.33348098e-02
-6.08615220e-01 -4.32850122e-01 3.45203966e-01 7.23814726e-01
-7.08634853e-01 -2.57933270e-02 -8.01421046e-01 4.40507323e-01
-1.69011891e+00 -8.20697308e-01 -4.84577864e-01 2.11459041e+00
7.42454290e-01 4.78726998e-02 2.07754046e-01 1.35806113e-01
6.09450340e-01 9.92069114e-03 -6.88815773e-01 1.74946696e-01
8.51965547e-02 3.96304399e-01 8.69831145e-02 4.27186728e-01
-8.75044584e-01 9.95622933e-01 6.24069929e+00 5.01617193e-01
-7.91603506e-01 6.45572022e-02 1.64078429e-01 -5.15812747e-02
-4.06464815e-01 8.44084695e-02 -7.26185381e-01 8.25442672e-02
4.23862100e-01 1.23549394e-01 2.47692347e-01 9.57406878e-01
1.81906298e-01 1.29347056e-01 -1.16693068e+00 9.76569295e-01
4.27847579e-02 -1.26548934e+00 2.66206324e-01 2.56149709e-01
7.60896027e-01 -2.72590369e-01 -4.94897850e-02 -1.17099509e-02
4.00062770e-01 -9.44966257e-01 1.24323225e+00 8.54078770e-01
7.38376975e-01 -4.00376469e-01 8.64146128e-02 5.91800988e-01
-1.05886412e+00 3.18780452e-01 -2.03036964e-01 1.60859331e-01
5.68630993e-01 4.40408379e-01 -6.40030265e-01 6.51602030e-01
5.55388331e-01 6.01733506e-01 -3.19925934e-01 1.10959697e+00
-3.04023087e-01 2.96402872e-01 -3.83651495e-01 4.87205714e-01
-2.00856492e-01 -1.67943344e-01 6.80335522e-01 9.06789899e-01
3.52041900e-01 2.74880499e-01 5.70704117e-02 8.41457725e-01
-3.08155298e-01 -1.21613275e-02 -6.08408511e-01 3.55430335e-01
7.67183229e-02 1.15799212e+00 -8.16461802e-01 -3.50107402e-01
-7.90946707e-02 1.25430977e+00 1.24084577e-01 1.15648068e-01
-1.13814056e+00 2.33676225e-01 6.24947906e-01 6.59697413e-01
3.18681538e-01 -5.00073969e-01 -2.87755877e-01 -1.41657877e+00
1.95704088e-01 -7.54143119e-01 -1.70051828e-02 -1.09629941e+00
-1.01886737e+00 2.12179631e-01 1.79401979e-01 -1.25841725e+00
-6.15983047e-02 -4.87687528e-01 -3.53494436e-01 5.45466840e-01
-1.14137125e+00 -1.59995782e+00 -5.27734160e-01 5.97901583e-01
7.76290298e-01 2.72276163e-01 7.08593190e-01 2.43757427e-01
-2.40721241e-01 1.45741269e-01 -3.90531331e-01 -7.59668350e-02
3.72883618e-01 -1.03428364e+00 8.30691814e-01 7.14487553e-01
3.05096716e-01 4.69261706e-01 6.18280590e-01 -7.75104582e-01
-1.87552214e+00 -1.01255298e+00 2.54266590e-01 -1.12505114e+00
4.00279313e-01 -3.48459274e-01 -8.49694431e-01 1.05946362e+00
-3.83067012e-01 4.35708940e-01 1.75318062e-01 -2.98324645e-01
-1.96058750e-01 5.07334113e-01 -1.23545003e+00 8.49808931e-01
1.96715045e+00 -6.43121377e-02 -8.35387230e-01 5.18062413e-01
6.84980214e-01 -1.10763705e+00 -1.09953117e+00 6.17730319e-01
9.80795681e-01 -7.89923966e-01 1.65526843e+00 -7.62954593e-01
6.57491267e-01 -5.20977676e-01 -2.92091489e-01 -9.75909233e-01
-5.05746752e-02 -9.21822846e-01 -6.77551806e-01 8.17609310e-01
-6.70891628e-02 6.13458417e-02 1.23979867e+00 6.87958181e-01
-4.60139185e-01 -8.55497956e-01 -8.68105948e-01 -7.63614237e-01
-1.47630438e-01 -6.52767122e-01 5.52487314e-01 7.35444427e-01
-5.32337010e-01 1.12842068e-01 -3.66648734e-01 3.54398787e-01
9.50840354e-01 1.97730154e-01 1.46974742e+00 -1.46610880e+00
-4.51717138e-01 -1.64937064e-01 -5.51717997e-01 -1.65035677e+00
4.29008812e-01 -9.41020131e-01 -2.30581909e-02 -1.70073962e+00
1.59703001e-01 -4.41361189e-01 5.75173497e-01 4.54341531e-01
-7.58585036e-02 4.75088239e-01 2.59611070e-01 2.12562621e-01
-2.99271762e-01 7.73840666e-01 1.95359027e+00 -1.09362632e-01
-1.85295075e-01 3.27848107e-01 -4.54259962e-01 1.21363556e+00
1.69518813e-01 -4.84677434e-01 -1.56944990e-01 -4.77089018e-01
-1.81018546e-01 3.34434152e-01 7.59897888e-01 -7.41116464e-01
-1.64671227e-01 -2.19413787e-01 4.08388674e-01 -7.97548115e-01
1.01928270e+00 -1.01976860e+00 7.44709969e-01 5.53156495e-01
-1.35283673e-03 -1.61024183e-01 -1.39914698e-03 6.51952446e-01
3.61549854e-01 -2.15583175e-01 8.44398320e-01 -4.48832303e-01
-3.34245771e-01 5.01623750e-01 6.35706447e-03 -4.04095761e-02
9.85139132e-01 -6.71474040e-01 2.21893042e-01 -1.79609254e-01
-1.21395421e+00 -1.44492328e-01 9.47963476e-01 3.75222445e-01
8.96608233e-01 -1.27100563e+00 -7.48322725e-01 -8.87301639e-02
-1.00536399e-01 7.06341267e-01 9.99005586e-02 6.31091535e-01
-7.46034324e-01 -2.26732895e-01 -1.42667159e-01 -8.73530328e-01
-1.32520962e+00 4.48154867e-01 4.29139167e-01 -2.47137975e-02
-1.11358082e+00 5.75488567e-01 3.80159199e-01 -8.38841975e-01
2.07112238e-01 -5.77674806e-01 2.43517444e-01 -3.48634750e-01
5.87766655e-02 3.75327945e-01 -1.51374005e-03 -8.41442287e-01
-2.38079205e-01 1.15777719e+00 4.42887634e-01 -2.15476632e-01
1.53909385e+00 5.34425080e-02 1.79149389e-01 1.63014814e-01
8.52494657e-01 2.72848845e-01 -1.64063776e+00 -1.27137005e-01
-1.89632773e-01 -7.30942369e-01 -2.94878274e-01 -5.09018004e-01
-1.22064078e+00 7.81547606e-01 4.86462377e-02 -4.13556814e-01
6.11209154e-01 4.44249719e-01 8.35571945e-01 2.33732700e-01
6.14555955e-01 -6.22279882e-01 5.42860866e-01 2.59627014e-01
1.27009070e+00 -9.67312992e-01 3.18327039e-01 -8.49669993e-01
-6.10133350e-01 1.06552815e+00 4.55175191e-01 -2.49924198e-01
6.17599905e-01 1.81643009e-01 -2.28834599e-01 -4.45739955e-01
-2.40561947e-01 -6.84674308e-02 4.00231063e-01 8.01331758e-01
2.22401857e-01 -1.99714191e-02 1.00556962e-01 3.94689292e-01
-3.86444926e-01 -4.07110713e-02 5.61919630e-01 9.55955684e-01
-1.82585046e-01 -1.00335789e+00 -6.61332309e-01 8.89407769e-02
-3.63461763e-01 2.68504143e-01 -2.64633298e-01 9.69407916e-01
-4.34034280e-02 3.09907258e-01 -1.40306056e-01 -3.84726822e-02
7.13763356e-01 -2.20670272e-02 1.01082230e+00 -9.31414247e-01
-4.10450935e-01 1.82402924e-01 -9.72901238e-04 -7.21622050e-01
-6.43637240e-01 -9.12825108e-01 -1.40195262e+00 -8.77232477e-02
-3.42279196e-01 -5.44898331e-01 9.54770803e-01 6.76241517e-01
1.71828225e-01 4.61122394e-01 2.48028815e-01 -1.73075855e+00
-6.03502870e-01 -5.16747475e-01 -4.33815420e-01 4.74807620e-01
7.99642652e-02 -9.77777898e-01 -1.49220243e-01 5.19671679e-01] | [7.02018404006958, -1.350393533706665] |
57b1ca4e-f450-466b-a869-48a333ae53ea | image-synthesis-with-disentangled-attributes | 2207.09389 | null | https://arxiv.org/abs/2207.09389v1 | https://arxiv.org/pdf/2207.09389v1.pdf | Image Synthesis with Disentangled Attributes for Chest X-Ray Nodule Augmentation and Detection | Lung nodule detection in chest X-ray (CXR) images is common to early screening of lung cancers. Deep-learning-based Computer-Assisted Diagnosis (CAD) systems can support radiologists for nodule screening in CXR. However, it requires large-scale and diverse medical data with high-quality annotations to train such robust and accurate CADs. To alleviate the limited availability of such datasets, lung nodule synthesis methods are proposed for the sake of data augmentation. Nevertheless, previous methods lack the ability to generate nodules that are realistic with the size attribute desired by the detector. To address this issue, we introduce a novel lung nodule synthesis framework in this paper, which decomposes nodule attributes into three main aspects including shape, size, and texture, respectively. A GAN-based Shape Generator firstly models nodule shapes by generating diverse shape masks. The following Size Modulation then enables quantitative control on the diameters of the generated nodule shapes in pixel-level granularity. A coarse-to-fine gated convolutional Texture Generator finally synthesizes visually plausible nodule textures conditioned on the modulated shape masks. Moreover, we propose to synthesize nodule CXR images by controlling the disentangled nodule attributes for data augmentation, in order to better compensate for the nodules that are easily missed in the detection task. Our experiments demonstrate the enhanced image quality, diversity, and controllability of the proposed lung nodule synthesis framework. We also validate the effectiveness of our data augmentation on greatly improving nodule detection performance. | ['Dinggang Shen', 'Qian Wang', 'Jie-Zhi Cheng', 'Bin Xiao', 'Xi Ouyang', 'Zhenrong Shen'] | 2022-07-19 | null | null | null | null | ['lung-nodule-detection'] | ['medical'] | [ 4.16654140e-01 5.79516053e-01 -8.02193731e-02 -1.48630023e-01
-8.75626445e-01 -4.79573280e-01 4.42877293e-01 -2.47011438e-01
1.67338610e-01 4.59998935e-01 2.07954645e-02 -5.65562546e-01
5.50261065e-02 -1.07358289e+00 -6.57987177e-01 -7.64092624e-01
5.40656745e-01 6.64856553e-01 4.48947728e-01 6.74339458e-02
-4.09179479e-01 6.54227793e-01 -1.20616412e+00 4.05164063e-01
9.82520998e-01 9.45811510e-01 4.97288942e-01 8.55830967e-01
1.61139131e-01 7.63609648e-01 -3.00354391e-01 -2.52240241e-01
4.17089760e-01 -9.20212030e-01 -3.12178969e-01 6.57776117e-01
1.50678441e-01 -5.61586201e-01 -5.25123835e-01 7.34362721e-01
4.79630083e-01 -4.35254127e-01 8.15860331e-01 -7.16880143e-01
-6.95724726e-01 4.18702185e-01 -4.62094903e-01 1.02736207e-03
-2.28821903e-01 6.20078862e-01 8.11381817e-01 -6.86833560e-01
4.27733094e-01 7.80492306e-01 3.82311970e-01 9.21019852e-01
-1.09690905e+00 -3.94212067e-01 -3.02156270e-01 -4.71760124e-01
-1.06037223e+00 -9.74326134e-02 4.84392047e-01 -6.08625293e-01
1.07204519e-01 7.78324366e-01 8.97963941e-01 9.49308515e-01
3.48355472e-02 5.88925838e-01 9.61461782e-01 -3.23731929e-01
-5.57945343e-03 2.33301938e-01 -7.50692487e-01 1.11804807e+00
6.43370211e-01 2.96473563e-01 1.36471808e-01 -8.52698982e-02
1.76579165e+00 2.76400536e-01 -3.48920643e-01 -5.16891479e-01
-1.58024395e+00 7.41994917e-01 8.79053831e-01 4.62999165e-01
-3.57354641e-01 2.47936010e-01 2.90729497e-02 -3.00498098e-01
9.81728733e-02 7.16519475e-01 6.98822513e-02 4.76226926e-01
-7.70030975e-01 -1.07842498e-01 4.33285773e-01 8.34173858e-01
2.93533117e-01 1.86094031e-01 -1.04106438e+00 5.16808927e-01
2.61951447e-01 7.35075772e-01 4.55752045e-01 -5.13332784e-01
2.21225709e-01 8.47514212e-01 6.63222279e-03 -6.32989585e-01
-2.89335877e-01 -6.27655983e-01 -1.24484193e+00 -2.41116509e-02
3.83732229e-01 -1.65427234e-02 -1.38718784e+00 1.30223703e+00
4.21121806e-01 6.05032891e-02 -2.99423903e-01 1.09373212e+00
9.02503490e-01 2.50230908e-01 -5.92345335e-02 2.35349406e-02
1.67039895e+00 -9.36128914e-01 -4.96388495e-01 -2.00884357e-01
6.18670940e-01 -6.63313806e-01 1.42343473e+00 -1.71695173e-01
-1.20989251e+00 -4.93971735e-01 -9.54122007e-01 1.09251216e-01
4.30965751e-01 9.88635063e-01 4.29491997e-01 8.27163935e-01
-7.00550616e-01 6.39577433e-02 -1.16176701e+00 4.64082733e-02
7.38288641e-01 4.01414335e-01 7.74109270e-03 9.24631879e-02
-7.77180910e-01 6.10423446e-01 1.07221797e-01 6.35700524e-02
-8.86695206e-01 -9.18339074e-01 -6.05156124e-01 2.16748551e-01
5.87958097e-01 -1.14900458e+00 1.29500484e+00 -7.36216486e-01
-1.40938139e+00 9.22189653e-01 2.69844353e-01 -2.74604023e-01
9.22949433e-01 4.48281288e-01 5.22527806e-02 2.20448747e-01
-9.60578173e-02 7.25268245e-01 1.04850721e+00 -1.13867545e+00
-4.48681384e-01 -1.84641734e-01 -2.85736620e-01 1.12215176e-01
-2.53975093e-01 -5.22522211e-01 -6.81068182e-01 -8.10855746e-01
1.03863865e-01 -9.80987549e-01 -6.99272394e-01 4.83346343e-01
-6.72351599e-01 4.09819305e-01 7.00742543e-01 -6.24065638e-01
1.08725631e+00 -2.07090592e+00 -3.06779742e-01 3.65722895e-01
5.83702683e-01 2.38187402e-01 9.73099668e-04 -4.61620778e-01
2.47346446e-01 3.30208689e-01 -2.74828464e-01 9.29467604e-02
-2.20507458e-01 1.11347117e-01 8.68712291e-02 1.43916994e-01
6.73457682e-01 1.47455871e+00 -6.96385562e-01 -8.13144684e-01
4.62486386e-01 5.92724323e-01 -6.27263665e-01 4.73076135e-01
-4.75340426e-01 9.59950209e-01 -8.97282481e-01 6.80989504e-01
4.23851699e-01 -7.71923900e-01 2.57978201e-01 -4.42874670e-01
1.43935531e-01 1.52242593e-02 -7.65606761e-01 1.07256639e+00
-6.15930855e-01 2.09902644e-01 3.03729381e-02 -4.00732040e-01
8.93019915e-01 6.41419709e-01 6.08189166e-01 -4.27214116e-01
2.45241255e-01 5.19188166e-01 5.93564749e-01 -4.76136506e-01
-5.14369644e-02 -2.03480840e-01 2.27578267e-01 3.92172545e-01
-5.10008633e-01 -6.47505462e-01 6.53888136e-02 -6.86874315e-02
1.15709186e+00 -4.38003391e-01 3.70946109e-01 -9.09770057e-02
5.69288671e-01 3.87681760e-02 1.34577110e-01 6.90851629e-01
1.29001457e-02 1.20889461e+00 5.18534780e-01 -2.43694276e-01
-1.44233572e+00 -9.52125788e-01 -1.20503262e-01 4.15317208e-01
-1.32716954e-01 3.10720235e-01 -6.84656382e-01 -9.39769030e-01
-2.90011704e-01 2.31803134e-01 -8.22965086e-01 2.79445872e-02
-7.09600329e-01 -8.83549571e-01 5.34847617e-01 7.53929555e-01
4.83625799e-01 -1.22731256e+00 -8.38127375e-01 8.58461410e-02
-1.98280334e-01 -1.04724216e+00 -6.77417219e-01 -1.84249338e-02
-8.66920650e-01 -1.20938849e+00 -1.07042825e+00 -8.41461360e-01
1.20533252e+00 7.15074837e-02 1.08994937e+00 4.54431564e-01
-9.39431250e-01 1.07582502e-01 -1.47016838e-01 -1.81279734e-01
-1.01528716e+00 1.00499406e-01 -5.53379476e-01 5.82173727e-02
-3.81616741e-01 -1.88580871e-01 -1.04263616e+00 5.62609076e-01
-1.15925920e+00 5.31817317e-01 1.40362656e+00 1.10943317e+00
1.08352530e+00 -1.78412557e-01 3.21900189e-01 -1.19048309e+00
2.72049934e-01 -1.12065174e-01 -6.37814641e-01 4.82788235e-01
-9.87744555e-02 1.35312110e-01 4.62439388e-01 -5.89581609e-01
-1.15130424e+00 4.63235527e-01 -1.60165742e-01 -4.70055938e-01
-9.68260020e-02 1.45008549e-01 -3.36771347e-02 -1.59995303e-01
1.01042783e+00 3.77742380e-01 2.92986538e-02 3.55320498e-02
3.06272388e-01 3.88877988e-01 4.81475204e-01 -3.37325573e-01
1.10157382e+00 6.68822229e-01 2.27908850e-01 -6.28947616e-01
-7.77673602e-01 -2.87329316e-01 -5.03551722e-01 -7.70335346e-02
1.04273283e+00 -7.89551914e-01 -3.24843138e-01 -4.66407314e-02
-8.00470293e-01 -3.24244231e-01 -7.47640729e-01 5.36853671e-01
-4.91759092e-01 2.97144558e-02 -5.64046085e-01 -4.51327056e-01
-4.79538441e-01 -1.36102426e+00 1.22813320e+00 1.28049448e-01
1.10547006e-01 -7.35907614e-01 -1.89926818e-01 3.22388530e-01
6.30630851e-01 5.56007385e-01 1.15202653e+00 -4.51836258e-01
-1.10566747e+00 -2.26227418e-01 -5.76231420e-01 6.98011667e-02
6.05857253e-01 6.29818365e-02 -6.97479367e-01 -1.36557013e-01
4.46119159e-02 -2.43029445e-01 7.19006360e-01 6.72419488e-01
1.47662711e+00 -2.03223631e-01 -4.36927110e-01 6.24279201e-01
1.14972043e+00 2.48839222e-02 5.07463336e-01 -8.06053206e-02
9.40355837e-01 2.34136745e-01 2.92608589e-01 4.02446121e-01
-1.91249341e-01 4.73614961e-01 4.99491632e-01 -7.28036284e-01
-6.90993607e-01 -2.70606726e-01 -3.09810489e-01 4.14636761e-01
-1.67081162e-01 -2.57908165e-01 -8.94235134e-01 3.51899862e-01
-1.38115907e+00 -4.36351210e-01 -3.12705874e-01 2.02864242e+00
8.81700695e-01 1.16664819e-01 -3.13869864e-02 -1.23933412e-01
6.79570794e-01 -2.59563863e-01 -7.28926957e-01 3.81868482e-01
1.58003792e-01 2.79962838e-01 4.45334375e-01 1.62517667e-01
-9.90270555e-01 4.10345078e-01 5.18470478e+00 8.26582313e-01
-1.31925261e+00 -6.06514513e-02 1.18289542e+00 1.53277293e-01
-5.81914127e-01 -3.27632368e-01 -4.34017658e-01 1.51103601e-01
7.23390654e-02 1.66658859e-03 4.41115201e-02 8.35277140e-01
1.49031147e-01 2.09990844e-01 -1.15341544e+00 6.80130720e-01
-2.90235281e-01 -1.60355020e+00 3.13119650e-01 2.25549102e-01
1.01043749e+00 -3.73922795e-01 4.30243909e-01 -1.16305714e-02
-2.06666943e-02 -1.24284530e+00 3.66122752e-01 4.50155854e-01
1.27846146e+00 -3.95016223e-01 6.42505407e-01 2.69558340e-01
-1.16354680e+00 1.44784808e-01 -3.32479954e-01 7.31117368e-01
-1.47482812e-01 7.20540345e-01 -1.64232194e+00 3.99277478e-01
1.61787704e-01 8.09304118e-02 -1.00642705e+00 1.00620449e+00
-1.76864475e-01 5.61845124e-01 -2.49792084e-01 -3.59616876e-02
-3.04172877e-02 -2.50882506e-02 3.03944319e-01 8.92505109e-01
6.17914796e-01 2.37269565e-01 4.45807390e-02 1.39242661e+00
-2.59391367e-01 1.02195725e-01 -3.04810464e-01 -2.39701346e-01
4.39691752e-01 1.40624511e+00 -1.05814123e+00 -3.89853895e-01
-3.62785637e-01 6.32971823e-01 -1.41137645e-01 1.90647971e-02
-1.01033783e+00 2.37314001e-01 -1.63638234e-01 8.20148945e-01
4.10552800e-01 2.73431927e-01 -5.70384920e-01 -1.15542758e+00
5.28554469e-02 -9.74902749e-01 6.69971630e-02 -4.71404493e-01
-1.25291967e+00 9.19540524e-01 -3.82389367e-01 -1.48535144e+00
-3.66753995e-01 -4.90177751e-01 -7.33254313e-01 5.92520058e-01
-1.20666397e+00 -1.42496681e+00 -7.06335425e-01 4.45456684e-01
3.75602394e-01 -5.08345105e-02 6.38073862e-01 1.76372260e-01
-3.66572767e-01 6.49383307e-01 -2.33307153e-01 2.11508706e-01
3.72376382e-01 -1.27820981e+00 5.12490571e-01 5.74211180e-01
6.39754161e-02 4.54711206e-02 2.25990236e-01 -6.87693715e-01
-1.07788479e+00 -1.51632452e+00 2.14542776e-01 -4.38679576e-01
3.64254504e-01 -1.63904011e-01 -8.04292023e-01 4.21703994e-01
-3.88707459e-01 5.49530566e-01 2.16139719e-01 -8.09110105e-01
1.04524404e-01 2.22322404e-01 -1.15334141e+00 6.98905051e-01
6.55105591e-01 -1.45805910e-01 1.94905903e-02 5.44520497e-01
4.38825190e-01 -7.05933630e-01 -6.46002710e-01 7.96675384e-01
3.18916589e-01 -9.42570448e-01 1.08091855e+00 -2.60802895e-01
6.10246420e-01 -2.33834922e-01 2.92495608e-01 -8.54989886e-01
-2.72116989e-01 -1.84124991e-01 6.71124011e-02 6.41817629e-01
6.04182720e-01 -2.86816299e-01 1.30905616e+00 4.97292042e-01
-1.25250801e-01 -1.06149697e+00 -5.88260472e-01 -4.15414631e-01
-1.00841351e-01 -5.55930175e-02 5.81173480e-01 5.65959930e-01
-6.91591144e-01 -3.20340022e-02 -3.71762142e-02 8.60540047e-02
3.66462648e-01 2.17331961e-01 6.64852262e-01 -1.03263950e+00
-7.32694447e-01 -5.45257866e-01 -1.57782093e-01 -9.43387985e-01
-7.06138551e-01 -8.02783012e-01 1.56386971e-01 -1.60263300e+00
4.65847403e-01 -6.43470407e-01 7.87362233e-02 3.79713476e-01
-4.67670530e-01 4.83522832e-01 6.27140701e-02 1.99876904e-01
-5.43343425e-02 4.32612121e-01 2.37653041e+00 -1.17290154e-01
-2.12939605e-01 5.90816498e-01 -5.46776950e-01 6.96480870e-01
6.23449028e-01 -3.93455446e-01 -4.50674742e-01 -2.00071931e-01
-9.74994451e-02 5.33151984e-01 6.40891850e-01 -9.27200198e-01
-9.69278812e-02 -6.07154854e-02 7.53319025e-01 -4.49767321e-01
2.57405311e-01 -7.38653541e-01 1.55359417e-01 1.00103521e+00
-2.38959789e-01 -4.72478628e-01 -9.38114524e-02 5.87173223e-01
-1.29331529e-01 7.75878802e-02 1.10576379e+00 -3.15587103e-01
1.85198233e-01 6.11892343e-01 -2.76757270e-01 5.35467863e-02
1.13580096e+00 -1.58163607e-01 8.45205262e-02 -2.57900178e-01
-8.70001793e-01 -2.19682064e-02 3.89229715e-01 5.24402410e-03
7.51578927e-01 -1.39249074e+00 -1.04629374e+00 5.23223758e-01
1.37422174e-01 5.08400977e-01 2.41833493e-01 1.01302791e+00
-8.76881480e-01 5.17623365e-01 -2.05802191e-02 -8.08633924e-01
-1.09774554e+00 3.40562075e-01 6.75301731e-01 -7.92420566e-01
-6.40053153e-01 7.94019520e-01 9.03602421e-01 -4.01800245e-01
-6.95050741e-03 -8.03370714e-01 1.63828984e-01 -5.71164191e-01
1.04651958e-01 -1.01648040e-01 2.33061984e-01 -1.51841924e-01
-2.29481757e-02 3.57558668e-01 -2.10909452e-02 9.69871879e-02
9.80636179e-01 2.12463111e-01 2.72288144e-01 -7.24458396e-02
6.02028787e-01 5.11709042e-02 -1.36527979e+00 -4.72605936e-02
-3.23796421e-01 -4.47757632e-01 -7.92500749e-03 -7.91059196e-01
-1.33934236e+00 8.37782204e-01 6.51230097e-01 1.51582256e-01
1.08333862e+00 1.83029622e-01 7.30361283e-01 2.99157858e-01
-1.27847731e-01 -2.44520888e-01 5.98892629e-01 -1.35328323e-01
9.32746828e-01 -1.26771176e+00 -2.81703100e-02 -9.25656021e-01
-7.79722214e-01 1.22481668e+00 9.28355455e-01 -3.62394713e-02
3.35510075e-01 4.55913991e-01 1.86032966e-01 -3.13774437e-01
-6.08436227e-01 -2.26728603e-01 8.24092627e-01 4.96509314e-01
4.67860460e-01 2.82532990e-01 -2.09103376e-02 6.23863578e-01
-3.43877450e-02 -1.96489766e-01 4.32128459e-01 5.70338488e-01
-4.75217938e-01 -8.91140878e-01 -6.04081333e-01 8.28934252e-01
-2.97833741e-01 -6.96533965e-03 -4.03290391e-01 1.03663468e+00
9.64975283e-02 3.44982535e-01 -2.09090821e-02 4.37585004e-02
3.48866493e-01 -5.19698977e-01 6.33783877e-01 -1.02748728e+00
-5.66309392e-01 3.55837762e-01 -2.99584895e-01 -1.50928214e-01
1.92442667e-02 -2.52985418e-01 -1.16002095e+00 1.87342435e-01
-6.00209057e-01 -1.94524720e-01 3.20784032e-01 6.75411999e-01
1.06786108e-02 1.10926485e+00 7.68167973e-01 -5.18930793e-01
-6.94135070e-01 -7.78652728e-01 -4.08164918e-01 3.65807801e-01
2.20514715e-01 -2.81709015e-01 -2.28240758e-01 2.07863286e-01] | [15.336407661437988, -2.1384620666503906] |
638e18eb-4633-4b25-b327-1505d161b718 | static-hand-gesture-recognition-for-american | 2207.12559 | null | https://arxiv.org/abs/2207.12559v3 | https://arxiv.org/pdf/2207.12559v3.pdf | Static Hand Gesture Recognition for American Sign Language using Neuromorphic Hardware | In this paper, we develop four spiking neural network (SNN) models for two static American Sign Language (ASL) hand gesture classification tasks, i.e., the ASL Alphabet and ASL Digits. The SNN models are deployed on Intel's neuromorphic platform, Loihi, and then compared against equivalent deep neural network (DNN) models deployed on an edge computing device, the Intel Neural Compute Stick 2 (NCS2). We perform a comprehensive comparison between the two systems in terms of accuracy, latency, power consumption, and energy. The best DNN model achieves an accuracy of 99.93% on the ASL Alphabet dataset, whereas the best performing SNN model has an accuracy of 99.30%. For the ASL-Digits dataset, the best DNN model achieves an accuracy of 99.76% accuracy while the SNN achieves 99.03%. Moreover, our obtained experimental results show that the Loihi neuromorphic hardware implementations achieve up to 20.64x and 4.10x reduction in power consumption and energy, respectively, when compared to NCS2. | ['Mohammadreza Mohammadi', 'Ramtin Zand', 'Sara Hendrix', 'James Seekings', 'Peyton Chandarana'] | 2022-07-25 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.15027973e-02 -4.19846922e-01 -6.45253900e-03 1.68015450e-01
1.81431994e-01 -3.92415762e-01 3.78813326e-01 -4.25590724e-01
-9.85720038e-01 5.20999491e-01 -3.37308973e-01 -2.14245170e-01
5.83031774e-02 -5.83323836e-01 -7.25044787e-01 -9.01504576e-01
1.84676722e-01 2.38077193e-01 5.54075003e-01 1.27894372e-01
2.25131825e-01 5.99161088e-01 -1.88131940e+00 2.73608007e-02
4.81674999e-01 1.60093927e+00 6.34633899e-02 6.47075057e-01
7.17202723e-02 8.15646529e-01 -7.62430370e-01 -7.07795992e-02
3.64277810e-01 6.60483784e-04 -3.27158198e-02 -1.06296504e+00
3.55716974e-01 -3.40611786e-01 -1.00313127e+00 1.27118909e+00
8.43360901e-01 -1.48827299e-01 4.37535256e-01 -1.59174085e+00
-4.70223308e-01 5.42152166e-01 -1.58524334e-01 2.40320638e-01
-2.59648383e-01 4.11754906e-01 2.36578763e-01 -5.42881072e-01
6.72878802e-01 9.45251465e-01 8.95750463e-01 1.05665827e+00
-6.16805911e-01 -1.17797673e+00 -2.70534754e-01 2.66703457e-01
-1.53788197e+00 -4.44408089e-01 4.21252668e-01 -1.01121828e-01
1.37208152e+00 -1.05576016e-01 1.25256038e+00 1.40879810e+00
6.66972458e-01 1.03247094e+00 1.28037333e+00 -7.50548318e-02
9.47393179e-01 -7.05308974e-01 7.64655113e-01 6.74230099e-01
9.76321459e-01 2.63858676e-01 -1.10344493e+00 6.59663528e-02
1.03481865e+00 -7.16987997e-02 -4.29222435e-02 4.62587178e-01
-7.89814711e-01 -1.37441143e-01 5.73569298e-01 3.33134234e-01
-5.76524854e-01 9.01624858e-01 5.38571835e-01 -9.73941945e-03
-4.67687100e-01 -2.28333883e-02 -1.74308002e-01 -7.46528625e-01
-7.09744692e-01 2.63137519e-01 9.22654688e-01 1.17133355e+00
7.51772746e-02 4.65203077e-01 7.96933845e-02 5.41512549e-01
2.75149822e-01 8.67717385e-01 8.76379848e-01 -1.00417411e+00
-6.42688572e-02 6.46268010e-01 -2.04686165e-01 -3.36399406e-01
-6.36859894e-01 -1.28993273e-01 -1.01611352e+00 7.13815510e-01
6.57248318e-01 -7.11156949e-02 -1.54309881e+00 1.69005322e+00
-4.33037847e-01 1.86265349e-01 3.02229941e-01 1.12167466e+00
1.21114266e+00 3.39530140e-01 2.69041121e-01 4.84675497e-01
1.48257816e+00 -7.21554279e-01 -4.95609820e-01 -2.02754274e-01
1.09732307e-01 -1.28133923e-01 7.76001096e-01 4.36887652e-01
-9.25565422e-01 -4.56524968e-01 -1.25216639e+00 1.41200321e-02
-4.41844463e-01 3.57327312e-01 7.88892925e-01 7.92452216e-01
-1.25159550e+00 5.02787650e-01 -1.48455405e+00 -6.74982846e-01
4.45679903e-01 6.57630384e-01 8.47576559e-02 2.30362579e-01
-8.34510803e-01 5.78758001e-01 4.34193820e-01 -3.94860804e-02
-6.37871861e-01 -4.80073303e-01 -1.52550280e-01 1.01981513e-01
-4.77612019e-01 -4.96987224e-01 1.30184126e+00 -7.36159325e-01
-1.83648634e+00 8.16579998e-01 -5.14768716e-03 -5.16600847e-01
2.75979191e-01 1.67811900e-01 -4.42317665e-01 8.47327188e-02
-3.79890233e-01 9.74990726e-01 3.30799632e-02 -5.89530289e-01
-6.88132524e-01 -7.05551565e-01 -7.40891919e-02 -2.92654932e-01
-2.51910627e-01 -1.80718958e-01 -4.89165336e-01 -4.34479654e-01
3.56875032e-01 -1.30287600e+00 1.35031745e-01 5.70420742e-01
-1.45080477e-01 -3.56336653e-01 7.95792818e-01 -6.02363169e-01
9.88250494e-01 -2.06030869e+00 -3.35968673e-01 3.50670129e-01
1.85016498e-01 6.89056635e-01 -1.91629440e-01 -2.99498171e-01
4.92199510e-01 -3.92464489e-01 -1.70029067e-02 6.02813214e-02
-9.55954790e-02 5.38972676e-01 1.33333817e-01 1.56190813e-01
-1.85575813e-01 1.17479885e+00 -6.96825504e-01 -8.81646946e-02
-2.51102075e-02 5.35396814e-01 -3.10208350e-01 -2.33481318e-01
7.87989572e-02 1.95267588e-01 -2.20072791e-01 1.40818691e+00
4.01779890e-01 -1.58633888e-01 3.18848670e-01 -1.91252872e-01
-2.95597404e-01 1.40265837e-01 -7.44728625e-01 1.96033216e+00
-1.11059740e-01 1.06790960e+00 1.03396498e-01 -3.70797068e-01
1.20286715e+00 2.64978588e-01 3.32219094e-01 -1.18693864e+00
6.59579635e-01 8.50669384e-01 6.49484038e-01 -2.69401491e-01
6.83213100e-02 3.97310555e-01 1.58523824e-02 3.16178769e-01
3.83264750e-01 6.02975905e-01 2.57849425e-01 -2.46461704e-01
1.60984945e+00 1.34422585e-01 -1.03279307e-01 -5.96252859e-01
-1.83312505e-01 -9.81122479e-02 6.36618555e-01 9.50819433e-01
-8.20344210e-01 6.85712993e-02 2.86057889e-01 -5.37218511e-01
-1.00288963e+00 -1.28209913e+00 -2.05970675e-01 5.81811488e-01
4.66171324e-01 4.00066637e-02 -1.07850385e+00 1.23683929e-01
2.08131988e-02 1.95703179e-01 -2.30746433e-01 -1.25925452e-01
-8.01390111e-01 -7.48972952e-01 1.40296197e+00 1.15864766e+00
1.38636553e+00 -1.27730680e+00 -1.61907554e+00 1.95910275e-01
2.81490445e-01 -1.13662481e+00 6.11012354e-02 3.82571280e-01
-7.98610747e-01 -9.19501901e-01 -4.90795404e-01 -9.72485781e-01
2.77044952e-01 -4.03170139e-01 6.97492480e-01 -7.29713440e-02
-5.83707154e-01 1.46997750e-01 -2.40242612e-02 -7.15946615e-01
1.47329792e-01 8.63919258e-02 4.96340424e-01 -5.00769854e-01
1.00525010e+00 -8.53362501e-01 -7.34638095e-01 -1.52441323e-01
-5.53928316e-01 3.46497586e-03 5.22969127e-01 7.27672398e-01
6.27241194e-01 -5.47891140e-01 3.89471084e-01 -1.46110228e-03
4.11273777e-01 -1.37617350e-01 -6.72659636e-01 3.22208643e-01
-6.83476806e-01 2.95460075e-01 5.76913953e-01 -6.06326699e-01
-7.00161040e-01 2.42630541e-01 -2.17090040e-01 -4.13687289e-01
-6.80307150e-02 6.15088157e-02 -1.23043798e-01 -5.26425123e-01
5.93875945e-01 4.08851743e-01 -4.36132662e-02 -3.45973998e-01
-4.64079946e-01 1.04013693e+00 1.06397581e+00 -4.91461456e-01
-2.30224475e-01 4.23878193e-01 1.35650128e-01 -8.40440869e-01
1.66106954e-01 8.74165669e-02 -1.81686506e-01 -5.49289823e-01
5.56291401e-01 -7.21006572e-01 -1.33581471e+00 1.38019276e+00
-1.17291021e+00 -6.35613799e-01 -1.39149532e-01 5.53866982e-01
-5.88820457e-01 -2.35121503e-01 -9.36192751e-01 -1.01022017e+00
-9.64431107e-01 -9.46439505e-01 9.68970597e-01 7.83923090e-01
-3.61532122e-01 -4.98959482e-01 -4.45719883e-02 -3.12675148e-01
5.76902270e-01 2.84740925e-01 7.27620780e-01 -4.72036809e-01
-5.49606442e-01 -2.22079739e-01 -4.73028928e-01 -1.14173487e-01
-1.62792251e-01 3.49882990e-02 -1.21147776e+00 -2.38474563e-01
-3.19280684e-01 -3.65310431e-01 9.23553467e-01 6.01114690e-01
1.10007727e+00 1.30861700e-01 -4.90306348e-01 8.99516463e-01
1.47813821e+00 9.81536508e-01 8.89596224e-01 2.71879494e-01
3.57080817e-01 -9.79807973e-02 -1.80975497e-01 2.90300816e-01
1.30081654e-01 5.63637018e-01 3.84046286e-01 4.44532573e-01
-6.60139501e-01 -2.21589178e-01 1.62941128e-01 8.76731336e-01
-5.91796517e-01 -4.16522592e-01 -1.16214120e+00 2.70599693e-01
-1.84839189e+00 -6.45102561e-01 9.56378691e-03 1.84201396e+00
4.78100419e-01 3.18109654e-02 1.25826061e-01 2.83791453e-01
7.17252314e-01 -2.46182263e-01 -1.37658584e+00 -3.64476413e-01
-4.39576238e-01 8.40949237e-01 8.26871037e-01 -1.39162183e-01
-7.75364518e-01 8.34662616e-01 5.96210146e+00 5.28441668e-01
-1.51355350e+00 1.27262585e-02 1.87903762e-01 -5.96293807e-01
3.26234788e-01 -5.34454167e-01 -7.59519458e-01 1.11259544e+00
1.31030703e+00 3.13733518e-02 7.88473070e-01 9.14242148e-01
-3.24938416e-01 -6.08573630e-02 -8.25850189e-01 1.55289602e+00
-2.45368168e-01 -1.37963974e+00 -2.59165443e-03 2.00301055e-02
3.60226691e-01 4.35923606e-01 -8.07885006e-02 2.87128985e-01
1.24159276e-01 -9.34948564e-01 9.44451272e-01 7.03222334e-01
1.19780266e+00 -7.38594592e-01 7.62971938e-01 2.66740084e-01
-1.33706439e+00 -2.10059211e-01 -3.59763443e-01 -3.46844614e-01
-1.87998995e-01 7.23532960e-02 -1.71480067e-02 -2.86211461e-01
1.35888100e+00 5.06395102e-01 -1.00397095e-01 1.08002460e+00
-8.37673396e-02 6.06934667e-01 -7.46461987e-01 -9.13878858e-01
1.82621628e-01 -2.69559808e-02 3.75617236e-01 1.14923394e+00
4.56118464e-01 5.54603696e-01 -6.29881084e-01 8.33645642e-01
-3.62284541e-01 -6.23444676e-01 -3.39794338e-01 -1.94434524e-01
9.85485673e-01 8.38470280e-01 -8.16007972e-01 -5.51678181e-01
6.57694563e-02 1.11385298e+00 6.28863126e-02 3.44968081e-01
-7.14789748e-01 -8.19523752e-01 1.09342766e+00 -3.85732621e-01
-1.25249341e-01 -2.30071217e-01 -7.93785214e-01 -1.05204725e+00
3.10087442e-01 -4.37974572e-01 -2.30475277e-01 -8.49196434e-01
-9.12532210e-01 7.94134438e-01 -7.71631718e-01 -9.94967163e-01
-2.19409421e-01 -1.56382906e+00 -5.20172000e-01 4.89698857e-01
-1.05097175e+00 -1.01605737e+00 -8.45292866e-01 3.62896472e-01
1.24686010e-01 -5.27329981e-01 8.78013134e-01 5.61549425e-01
-7.24165082e-01 9.97745693e-01 4.06373143e-01 5.27059793e-01
-5.19582257e-02 -7.56318688e-01 8.14371169e-01 7.77894557e-01
-2.74585158e-01 5.21627128e-01 1.75387561e-01 -5.89268625e-01
-1.71930647e+00 -1.00369346e+00 5.93240678e-01 1.26381973e-02
4.68714893e-01 -2.77768075e-01 -6.51806772e-01 5.53575039e-01
-1.95562989e-01 1.28418028e-01 3.82730126e-01 -7.04234421e-01
-3.63119364e-01 -5.43439984e-02 -1.41576397e+00 8.05189371e-01
1.84533787e+00 -3.86405736e-01 -4.94714268e-02 -1.38838574e-01
2.85910303e-03 -4.83667731e-01 -7.01536238e-01 4.69093382e-01
1.51111591e+00 -7.89366961e-01 5.21667600e-01 -4.80278760e-01
1.11233711e-01 -3.63218397e-01 -4.18605477e-01 -8.23850036e-01
-1.78485170e-01 -1.76794812e-01 -4.83912408e-01 6.72349453e-01
-1.17956161e-01 -9.80192065e-01 1.16047215e+00 8.65872681e-01
-2.49894366e-01 -8.83300662e-01 -1.63559163e+00 -1.29618764e+00
-3.31997246e-01 -4.45318937e-01 4.95138764e-01 2.01920465e-01
5.79381250e-02 -2.96133548e-01 7.39910379e-02 -1.42708123e-01
8.87114167e-01 -9.08376276e-03 2.59388179e-01 -1.30341423e+00
1.38435692e-01 -7.18259871e-01 -1.28595960e+00 -9.32051659e-01
4.25453223e-02 -7.94155598e-01 2.03402750e-02 -1.35218012e+00
2.05179989e-01 -3.09736967e-01 -6.48129523e-01 7.41417050e-01
5.81616163e-01 5.12646079e-01 1.57989562e-01 2.44600460e-01
-2.58446872e-01 1.52057707e-01 5.77191949e-01 -2.34185770e-01
-7.57665783e-02 -1.91273466e-01 -1.27058059e-01 8.59163225e-01
9.36560631e-01 -2.23831445e-01 1.08039036e-01 -6.89287841e-01
-4.33636367e-01 -2.80920595e-01 5.27335584e-01 -1.85591280e+00
8.35552573e-01 3.18143755e-01 5.35370648e-01 -3.23285669e-01
4.76096362e-01 -7.09734976e-01 2.68565118e-01 1.34453785e+00
2.49237847e-02 -5.82430474e-02 4.44776505e-01 2.11146176e-01
8.09350610e-02 8.06779414e-03 8.66320789e-01 1.18485600e-01
-1.17696571e+00 1.90771297e-01 -7.29145110e-01 -4.61283356e-01
7.04630911e-01 -7.94275403e-01 -7.19953120e-01 1.38414845e-01
-5.01622438e-01 -7.51239285e-02 4.22034293e-01 3.67203087e-01
8.29505503e-01 -1.45916617e+00 -1.66263610e-01 6.13874197e-01
7.82308057e-02 -4.52283949e-01 1.10859521e-01 4.37320471e-01
-8.81095469e-01 7.92170227e-01 -1.14077532e+00 -5.86551607e-01
-1.14347661e+00 -1.65227622e-01 6.30975664e-01 2.06332117e-01
-4.56809789e-01 8.11035097e-01 -5.56573749e-01 -2.79042155e-01
6.42529786e-01 -6.56470239e-01 3.71467412e-01 -5.59061587e-01
5.00101447e-01 9.76709843e-01 8.39920193e-02 -1.57257602e-01
-7.45139956e-01 6.87005937e-01 5.24924755e-01 -1.68003380e-01
1.15729022e+00 7.21836030e-01 -2.28617061e-02 2.57498085e-01
7.44203806e-01 -9.04046893e-01 -1.16972387e+00 6.97304234e-02
3.33708413e-02 1.53516501e-01 -9.86072123e-02 -1.18877041e+00
-1.18036509e+00 8.94967735e-01 1.49238944e+00 -6.50818348e-01
1.45173085e+00 -3.49566758e-01 1.22666323e+00 8.80156398e-01
1.00399387e+00 -1.17395580e+00 -2.31996760e-01 7.87544847e-01
5.12446463e-01 -5.86274266e-01 -6.67201936e-01 2.71904647e-01
-3.64042707e-02 1.31104136e+00 1.10503840e+00 -5.58171749e-01
6.60897493e-01 8.70228350e-01 -4.33936268e-02 -1.76862314e-01
-4.96100396e-01 1.17158995e-03 -2.47254282e-01 5.42161942e-01
6.33401051e-02 3.76734376e-01 -3.32149714e-01 8.32442284e-01
-2.97065914e-01 7.01462686e-01 1.67225584e-01 1.15430188e+00
-4.19387817e-01 -4.02380198e-01 9.80547722e-03 6.36793077e-01
4.20842245e-02 3.05388384e-02 -5.92518091e-01 7.02231050e-01
2.26978704e-01 5.09429693e-01 7.44640708e-01 -9.14033055e-01
4.06205744e-01 2.59962648e-01 7.66668499e-01 2.18217745e-01
-6.90269113e-01 -3.75590652e-01 -2.42911667e-01 -7.04901636e-01
-5.15097566e-02 -2.09607214e-01 -1.88592243e+00 -6.51158690e-01
1.15836591e-01 -5.42795360e-01 1.23990238e+00 7.78696716e-01
7.05248654e-01 5.73990822e-01 -3.81868392e-01 -8.53021562e-01
-2.84553736e-01 -8.99977863e-01 -9.07374442e-01 -2.08679959e-02
-2.59939164e-01 -6.23498082e-01 6.19669221e-02 -2.86792815e-01] | [8.24474811553955, 2.4696450233459473] |
f5887485-d648-4ef6-a1a1-1b8069513969 | autoqgs-auto-prompt-for-low-resource | 2208.12461 | null | https://arxiv.org/abs/2208.12461v1 | https://arxiv.org/pdf/2208.12461v1.pdf | AutoQGS: Auto-Prompt for Low-Resource Knowledge-based Question Generation from SPARQL | This study investigates the task of knowledge-based question generation (KBQG). Conventional KBQG works generated questions from fact triples in the knowledge graph, which could not express complex operations like aggregation and comparison in SPARQL. Moreover, due to the costly annotation of large-scale SPARQL-question pairs, KBQG from SPARQL under low-resource scenarios urgently needs to be explored. Recently, since the generative pre-trained language models (PLMs) typically trained in natural language (NL)-to-NL paradigm have been proven effective for low-resource generation, e.g., T5 and BART, how to effectively utilize them to generate NL-question from non-NL SPARQL is challenging. To address these challenges, AutoQGS, an auto-prompt approach for low-resource KBQG from SPARQL, is proposed. Firstly, we put forward to generate questions directly from SPARQL for the KBQG task to handle complex operations. Secondly, we propose an auto-prompter trained on large-scale unsupervised data to rephrase SPARQL into NL description, smoothing the low-resource transformation from non-NL SPARQL to NL question with PLMs. Experimental results on the WebQuestionsSP, ComlexWebQuestions 1.1, and PathQuestions show that our model achieves state-of-the-art performance, especially in low-resource settings. Furthermore, a corpus of 330k factoid complex question-SPARQL pairs is generated for further KBQG research. | ['Xiaodong He', 'Youzheng Wu', 'Wen Zhao', 'Junwei Bao', 'Guanming Xiong'] | 2022-08-26 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [-3.57400537e-01 6.70867503e-01 -1.42303901e-02 -3.39508653e-01
-1.38754427e+00 -7.42991745e-01 6.07289732e-01 1.81769773e-01
-2.15745896e-01 1.18678951e+00 4.11683410e-01 -5.54828227e-01
-1.80157423e-01 -1.58051729e+00 -1.04303586e+00 2.84728408e-01
3.89673859e-01 8.59085083e-01 4.17086661e-01 -6.31459475e-01
-2.67795354e-01 3.44654173e-02 -1.51110888e+00 6.80051625e-01
1.67414725e+00 8.43985915e-01 8.30341950e-02 6.45025492e-01
-1.10635185e+00 1.26702571e+00 -6.14932418e-01 -1.06611514e+00
9.56259966e-02 -4.78919387e-01 -1.20469880e+00 -4.41484451e-01
7.43822634e-01 -7.56522566e-02 -2.53804117e-01 8.82714272e-01
5.11247694e-01 6.01378605e-02 3.18031937e-01 -1.23532081e+00
-9.14143503e-01 9.16735351e-01 3.17770720e-01 1.02833696e-01
9.21383202e-01 2.02429056e-01 1.46405852e+00 -1.04759192e+00
7.81143606e-01 1.34900999e+00 3.71975452e-01 4.76587147e-01
-9.11783576e-01 -4.98635918e-01 -3.41071516e-01 4.33926702e-01
-1.31978083e+00 -1.87883645e-01 2.80510426e-01 -1.17285140e-01
1.35796642e+00 6.23004377e-01 6.32534206e-01 7.58246899e-01
4.44015041e-02 5.70253968e-01 1.02089596e+00 -4.03309733e-01
4.02068719e-02 4.98472840e-01 5.54675311e-02 8.81345510e-01
4.15433675e-01 -3.11980367e-01 -5.90994537e-01 -4.29237574e-01
4.83181596e-01 -6.24676645e-01 -2.84011424e-01 1.05150245e-01
-1.16746497e+00 9.68346834e-01 3.03442210e-01 2.02443507e-02
-4.86168921e-01 -8.70867968e-02 1.79999322e-01 4.44549471e-01
3.23324800e-01 1.04323447e+00 -8.64787877e-01 -1.24312565e-02
-5.79568446e-01 7.75491297e-01 1.41790521e+00 1.69058657e+00
1.04998505e+00 -3.26203287e-01 -6.91653073e-01 6.25241339e-01
5.48478104e-02 6.53093457e-01 3.57707918e-01 -8.17287087e-01
9.93887782e-01 1.16623008e+00 2.16468513e-01 -6.99648440e-01
-2.28092462e-01 -1.73782390e-02 -5.13417840e-01 -7.29761958e-01
1.90949023e-01 -1.70858368e-01 -5.83987713e-01 1.14387870e+00
6.78348243e-01 -2.38757789e-01 6.78637207e-01 6.86860085e-01
1.69385529e+00 9.77957010e-01 2.52681345e-01 -4.48090062e-02
2.01716661e+00 -9.50613081e-01 -7.76311278e-01 -1.58405870e-01
7.75016129e-01 -5.85941434e-01 1.89999723e+00 3.95713449e-02
-9.46206033e-01 -5.61617613e-01 -4.74790901e-01 -3.89724523e-01
-6.35880888e-01 7.15107694e-02 7.53877699e-01 5.10481894e-01
-8.35903823e-01 -8.59940872e-02 -1.84780270e-01 -6.63242519e-01
1.21391274e-01 -3.02173167e-01 -2.13793591e-01 -3.56264174e-01
-1.88022828e+00 6.45621181e-01 1.15511906e+00 -3.51038218e-01
-8.10592413e-01 -1.27658260e+00 -9.36517179e-01 2.03933299e-01
1.25461066e+00 -1.17890739e+00 1.33871186e+00 -1.09048612e-01
-1.45099616e+00 5.34502566e-01 -2.19516531e-02 -3.72733563e-01
5.71706370e-02 -2.11953893e-01 -7.48752773e-01 1.14968583e-01
4.01691258e-01 3.99158865e-01 3.75407398e-01 -9.11559939e-01
-5.60097396e-01 3.25625800e-02 6.62026405e-01 2.39829630e-01
-1.99662447e-01 2.41187382e-02 -5.31208634e-01 -4.53775197e-01
-2.46353105e-01 -5.73763371e-01 -1.00840017e-01 -6.90808296e-01
-7.08742559e-01 -8.53890419e-01 6.47030398e-02 -8.36571693e-01
1.16357267e+00 -1.44219267e+00 -5.50317109e-01 -7.05690496e-03
-3.93960662e-02 3.03828716e-01 -3.61411422e-01 9.97097015e-01
5.13769537e-02 3.12170923e-01 -4.41853814e-02 5.40623367e-01
3.73880804e-01 4.10732806e-01 -7.38799572e-01 -6.33284926e-01
5.79829395e-01 1.57813740e+00 -1.15891099e+00 -9.06281292e-01
-3.91796350e-01 -2.69869834e-01 -8.09444904e-01 8.18418622e-01
-1.14343286e+00 3.95764485e-02 -7.29732573e-01 7.18554199e-01
4.94128287e-01 -4.23353046e-01 7.81749412e-02 -5.99046767e-01
3.30322891e-01 8.00647855e-01 -9.21735048e-01 1.93982208e+00
-6.62076056e-01 -2.14485541e-01 -5.23535132e-01 -4.13958520e-01
9.66919422e-01 4.32595283e-01 7.21243843e-02 -8.42290521e-01
-5.23649573e-01 2.77259052e-01 -5.34508049e-01 -1.18599641e+00
8.46819580e-01 -2.51459897e-01 -4.42200780e-01 3.28704625e-01
5.37871361e-01 -7.75709271e-01 9.06447768e-01 8.06762516e-01
1.28928459e+00 1.73340112e-01 4.29251492e-01 -7.49763399e-02
6.90884292e-01 5.13241231e-01 4.05697167e-01 7.34055817e-01
5.73261917e-01 2.75850236e-01 5.80336750e-01 -5.59900738e-02
-5.53501427e-01 -1.27857316e+00 1.42478332e-01 9.77300525e-01
-2.30700374e-01 -9.33856666e-01 -5.64463496e-01 -1.39267480e+00
1.36155844e-01 1.36169803e+00 -2.93200351e-02 5.25273606e-02
-5.69275856e-01 -6.58738434e-01 8.07423234e-01 2.48864129e-01
5.54706275e-01 -1.51258731e+00 1.45016629e-02 5.15545070e-01
-6.11931086e-01 -1.65227687e+00 -4.09776270e-01 -4.07023758e-01
-4.16431874e-01 -1.42300797e+00 -1.98769838e-01 -4.17501390e-01
4.96977746e-01 -1.41878650e-01 2.01447940e+00 -2.44551614e-01
-2.34184220e-01 5.75914800e-01 -7.12353408e-01 -3.15728664e-01
-5.68903267e-01 3.08594763e-01 -4.80074555e-01 -4.39567119e-01
4.56349999e-01 -3.35741103e-01 -4.24946427e-01 1.07073627e-01
-1.36673105e+00 5.59090972e-02 6.46691442e-01 4.63510185e-01
8.20478737e-01 2.44752653e-02 1.08230591e+00 -1.32554388e+00
1.12124252e+00 -6.63783789e-01 -7.09821343e-01 9.05595124e-01
-6.79680109e-01 2.24471584e-01 1.11297297e+00 3.02712303e-02
-1.43192947e+00 -5.56099534e-01 -2.31962934e-01 1.75971895e-01
-5.82950376e-03 1.03918612e+00 -7.57241964e-01 2.42801517e-01
8.12908709e-01 4.18349832e-01 -6.43591881e-01 -4.17822331e-01
1.25047553e+00 4.67886716e-01 5.01909494e-01 -1.01489449e+00
1.11562502e+00 -2.23672241e-02 -3.45686972e-01 -4.55920249e-01
-1.21147835e+00 -4.26680982e-01 -2.25946568e-02 1.12621918e-01
8.27016592e-01 -9.76910710e-01 -6.21878624e-01 -2.52793372e-01
-1.32259560e+00 -1.58058673e-01 -9.32812631e-01 2.60583788e-01
-5.24642348e-01 3.39142054e-01 -3.58691216e-01 -4.88748163e-01
-8.12627137e-01 -4.95382607e-01 9.29941058e-01 3.73844028e-01
4.38343622e-02 -9.08671081e-01 3.84957314e-01 8.88829648e-01
1.74883872e-01 -9.59420204e-02 1.33924520e+00 -9.39785898e-01
-1.05770850e+00 -7.54238889e-02 -5.28623164e-01 4.69921887e-01
1.95572004e-01 -4.64858472e-01 -3.42978150e-01 1.86483786e-02
-2.60268211e-01 -8.11601162e-01 2.57136792e-01 -6.26482248e-01
1.04519331e+00 -9.62187171e-01 2.36010045e-01 5.48289716e-01
1.50825965e+00 -3.28808635e-01 6.95784569e-01 1.65604994e-01
6.99237168e-01 7.82938123e-01 9.29453731e-01 2.96326220e-01
1.35915613e+00 4.28173065e-01 1.55760899e-01 3.23850632e-01
-3.42077255e-01 -8.81519735e-01 2.97757268e-01 1.27519345e+00
4.42330569e-01 -3.78343761e-01 -8.90573144e-01 6.42154932e-01
-1.67952728e+00 -6.03104353e-01 -3.81397009e-01 1.74564421e+00
1.36964869e+00 -2.85350770e-01 -2.74802327e-01 -5.63922822e-01
5.59088401e-02 5.42359725e-02 -4.45484251e-01 -9.95596647e-02
-3.59357983e-01 5.64814270e-01 1.68494552e-01 2.60338277e-01
-4.75043714e-01 1.30412757e+00 4.55578709e+00 9.95424509e-01
-3.29048365e-01 2.24671930e-01 -5.12441508e-02 1.30214453e-01
-9.87008393e-01 4.32019204e-01 -1.23289037e+00 4.54043180e-01
9.90319967e-01 -7.69067466e-01 4.51277852e-01 6.80310786e-01
-2.26297677e-01 2.19985913e-03 -9.01555002e-01 7.81218827e-01
2.71907419e-01 -1.37536776e+00 8.09964478e-01 -4.12157685e-01
7.11926401e-01 -3.19421560e-01 -5.79752803e-01 1.12519658e+00
6.60595596e-01 -9.42790806e-01 2.67351359e-01 9.27941859e-01
9.14302945e-01 -6.14512503e-01 7.18271255e-01 5.44358969e-01
-1.29486787e+00 1.34621322e-01 -7.14705467e-01 5.03691196e-01
3.27400446e-01 1.06352758e+00 -1.47038019e+00 1.50599384e+00
6.88389659e-01 2.74676718e-02 -7.91929185e-01 6.40990436e-01
-7.62549400e-01 5.75776100e-01 -2.17703134e-01 -2.01164350e-01
1.21217020e-01 -1.35422304e-01 2.67363101e-01 9.99894977e-01
6.02814376e-01 4.34516698e-01 1.50872469e-01 1.17332757e+00
-4.24593270e-01 7.62530029e-01 -3.36769730e-01 -5.01520157e-01
4.17156428e-01 1.65149283e+00 2.85226442e-02 -6.41524851e-01
-5.14242411e-01 6.71012163e-01 5.70277572e-01 5.13062894e-01
-6.68355823e-01 -5.97140908e-01 2.69138962e-01 4.72612321e-01
6.71525225e-02 7.17492998e-02 2.72905022e-01 -1.54918051e+00
3.94352674e-01 -1.31467438e+00 9.60140526e-01 -1.09692681e+00
-1.53174400e+00 7.14912474e-01 1.13124907e-01 -7.09988058e-01
-5.73159039e-01 -1.27637759e-01 -3.22143734e-01 1.12732291e+00
-1.93330741e+00 -1.32329917e+00 -6.09447658e-01 7.66932189e-01
2.96579123e-01 -1.54808253e-01 8.74150574e-01 2.60063708e-01
-1.20920844e-01 5.29314399e-01 -5.50123572e-01 -6.54330477e-03
8.58743131e-01 -1.51095486e+00 6.62489653e-01 6.44145072e-01
3.59003395e-01 6.38271213e-01 2.98201472e-01 -1.04225779e+00
-1.88851202e+00 -1.69649112e+00 1.31956840e+00 -8.10592771e-01
8.10399950e-01 -4.85610723e-01 -1.25832677e+00 5.57652116e-01
2.81164616e-01 1.17314734e-01 8.25617373e-01 -5.38181476e-02
-3.33227158e-01 -4.05034423e-01 -9.62950885e-01 4.78470594e-01
9.91931736e-01 -6.71114028e-01 -1.03592420e+00 7.33431995e-01
1.27064598e+00 -2.81953126e-01 -1.23492897e+00 5.98732471e-01
-1.51903316e-01 -6.20962501e-01 8.92948031e-01 -1.10173869e+00
3.05777758e-01 -4.93548900e-01 -1.21647604e-01 -1.33951950e+00
7.62676820e-02 -7.15831757e-01 -4.43720251e-01 1.57307255e+00
6.53853416e-01 -7.70451188e-01 8.29991162e-01 5.38534999e-01
-1.18485548e-01 -6.64780378e-01 -8.98002028e-01 -8.73819768e-01
-1.08231179e-01 -3.55079651e-01 1.19092715e+00 8.47226560e-01
-1.19790725e-01 6.97263002e-01 7.63925985e-02 2.00225383e-01
3.85312498e-01 3.79609436e-01 1.30328858e+00 -7.87914753e-01
-3.53107780e-01 2.97210991e-01 1.86484188e-01 -9.86828923e-01
2.44629696e-01 -1.31546474e+00 -9.83607844e-02 -2.12642527e+00
1.02883764e-01 -5.72965384e-01 2.49776229e-01 4.87125546e-01
-8.23198974e-01 -3.59549344e-01 3.93354110e-02 -1.34119615e-01
-7.80986071e-01 9.58917975e-01 1.57162833e+00 -7.49434233e-02
9.23679397e-02 -2.83665985e-01 -8.57263207e-01 1.78070560e-01
4.55636442e-01 -4.99181300e-01 -7.93311894e-01 -1.68881074e-01
1.02796197e+00 4.49620724e-01 2.89117336e-01 -7.88324833e-01
3.87713492e-01 -3.81206036e-01 -3.58532280e-01 -5.69295228e-01
5.22135086e-02 -4.11198676e-01 8.33857507e-02 -8.08795691e-02
-1.27333179e-01 1.51912361e-01 4.60908376e-02 3.41788203e-01
-5.20380557e-01 -4.59524304e-01 7.32961670e-03 -5.18763185e-01
-6.80811346e-01 4.43693399e-01 1.55119628e-01 1.06899071e+00
6.60221756e-01 4.23447579e-01 -7.69112825e-01 -4.45581436e-01
-2.95021057e-01 5.48902273e-01 -1.67183522e-02 3.61047804e-01
6.57316148e-01 -1.43004060e+00 -1.14594519e+00 2.53647696e-02
6.89365864e-01 3.47853094e-01 4.45444643e-01 4.72309023e-01
-5.44789314e-01 9.28498089e-01 3.33402425e-01 3.69705260e-02
-6.44310832e-01 5.85655987e-01 6.93606809e-02 -9.39871371e-01
-3.14276665e-01 8.00728500e-01 1.45898312e-01 -1.23435235e+00
-3.95588011e-01 -5.28281987e-01 -2.65719920e-01 1.33819237e-01
4.29435045e-01 1.79455727e-01 3.72273773e-01 -1.94358379e-02
-1.14761405e-01 7.31936693e-02 -1.84923168e-02 9.64698941e-02
8.59324753e-01 1.97535023e-01 -5.00772059e-01 2.26231874e-03
8.22095037e-01 2.55197227e-01 -4.43167746e-01 -5.19525647e-01
2.70804465e-01 -2.86067903e-01 -5.29409945e-01 -1.13562286e+00
-6.13940716e-01 3.75235617e-01 -1.32564381e-01 4.50262129e-01
9.87378240e-01 3.97267997e-01 1.22607172e+00 8.74066472e-01
5.82348824e-01 -8.23518634e-01 2.10102379e-01 8.55461895e-01
1.32227361e+00 -1.03848517e+00 -1.84136391e-01 -7.25654542e-01
-5.93112528e-01 8.82782459e-01 1.07126296e+00 2.65068054e-01
1.68491513e-01 -7.85771161e-02 1.86604530e-01 -5.16226053e-01
-1.07168519e+00 -4.03601706e-01 4.91849273e-01 4.22229677e-01
2.81569511e-01 -8.96528438e-02 -2.81451672e-01 1.04289675e+00
-8.52542400e-01 2.00364277e-01 4.58416432e-01 7.62433112e-01
-2.01718271e-01 -1.19802547e+00 -1.83062851e-01 7.88460970e-01
-3.39350343e-01 -6.84638023e-01 -2.73687094e-01 7.88505971e-01
-1.97543278e-02 8.58564913e-01 -3.28861028e-01 -1.23319574e-01
9.11877573e-01 3.82243991e-01 4.15750980e-01 -1.12953317e+00
-8.05493176e-01 -7.42706597e-01 6.94984555e-01 -6.52003884e-01
4.39592451e-02 -3.44118141e-02 -1.22599852e+00 5.98203205e-02
-2.31903374e-01 9.07416046e-01 4.04999316e-01 6.91899419e-01
6.07393205e-01 4.87055272e-01 3.22922200e-01 6.07220650e-01
-8.27894568e-01 -1.08675861e+00 -5.14175594e-01 4.12043840e-01
-3.89791995e-01 -1.24578709e-02 -8.99950862e-02 -3.80935594e-02] | [10.788351058959961, 7.928165435791016] |
ecf619a3-0bba-4e61-be95-f9511e81f7c2 | precision-anti-cancer-drug-selection-via | 2306.17771 | null | https://arxiv.org/abs/2306.17771v1 | https://arxiv.org/pdf/2306.17771v1.pdf | Precision Anti-Cancer Drug Selection via Neural Ranking | Personalized cancer treatment requires a thorough understanding of complex interactions between drugs and cancer cell lines in varying genetic and molecular contexts. To address this, high-throughput screening has been used to generate large-scale drug response data, facilitating data-driven computational models. Such models can capture complex drug-cell line interactions across various contexts in a fully data-driven manner. However, accurately prioritizing the most sensitive drugs for each cell line still remains a significant challenge. To address this, we developed neural ranking approaches that leverage large-scale drug response data across multiple cell lines from diverse cancer types. Unlike existing approaches that primarily utilize regression and classification techniques for drug response prediction, we formulated the objective of drug selection and prioritization as a drug ranking problem. In this work, we proposed two neural listwise ranking methods that learn latent representations of drugs and cell lines, and then use those representations to score drugs in each cell line via a learnable scoring function. Specifically, we developed a neural listwise ranking method, List-One, on top of the existing method ListNet. Additionally, we proposed a novel listwise ranking method, List-All, that focuses on all the sensitive drugs instead of the top sensitive drug, unlike List-One. Our results demonstrate that List-All outperforms the best baseline with significant improvements of as much as 8.6% in hit@20 across 50% test cell lines. Furthermore, our analyses suggest that the learned latent spaces from our proposed methods demonstrate informative clustering structures and capture relevant underlying biological features. Moreover, our comprehensive empirical evaluation provides a thorough and objective comparison of the performance of different methods (including our proposed ones). | ['Xia Ning', 'Vishal Dey'] | 2023-06-30 | null | null | null | null | ['drug-response-prediction'] | ['medical'] | [ 4.21106309e-01 -5.75558782e-01 -7.53748894e-01 -2.70630836e-01
-1.39173388e+00 -5.00630558e-01 5.23576140e-01 5.34995198e-01
-1.07647441e-01 1.24591660e+00 4.34882075e-01 -1.49052203e-01
-6.91856861e-01 -7.59600103e-01 -5.50017595e-01 -1.01486373e+00
1.16494358e-01 6.71723902e-01 -1.86266646e-01 -4.45112698e-02
3.46184999e-01 6.74444497e-01 -1.08076406e+00 4.74510074e-01
1.09371436e+00 6.15778387e-01 1.85451172e-02 3.80923808e-01
2.76966840e-01 7.81376660e-01 -4.14411008e-01 2.66089141e-02
-1.84252128e-01 -3.15727353e-01 -6.14527464e-01 -3.30424964e-01
2.64441758e-01 -1.75394461e-01 -5.04341781e-01 8.05146158e-01
9.15268064e-01 -9.43828747e-02 9.58155036e-01 -9.08166826e-01
-7.56629348e-01 5.06076574e-01 -4.27668929e-01 -2.36991465e-01
2.52013594e-01 9.15164500e-02 1.15468061e+00 -9.80582893e-01
6.05448902e-01 8.73111308e-01 5.07427633e-01 8.82599533e-01
-1.40192187e+00 -7.26320088e-01 -2.06050519e-02 -1.59110334e-02
-1.59444559e+00 -4.16813225e-01 4.44738418e-01 -6.37103796e-01
9.00423288e-01 5.15492022e-01 1.58604443e-01 1.10989571e+00
2.39452258e-01 7.16367364e-01 9.69574988e-01 -8.04933545e-04
3.12709987e-01 -7.17311278e-02 2.81976968e-01 5.58548331e-01
3.31872493e-01 -2.60440409e-02 -3.82356197e-01 -6.88066542e-01
5.16452968e-01 6.75824702e-01 -5.81555426e-01 -1.38698891e-01
-1.13465369e+00 8.86755645e-01 3.28792751e-01 2.84753650e-01
-3.29437912e-01 2.00234517e-01 4.10890788e-01 -2.87200987e-01
4.19405192e-01 7.82685995e-01 -8.08207095e-01 1.97988763e-01
-8.40916395e-01 4.73154001e-02 4.83564198e-01 4.74799067e-01
4.81981903e-01 -3.16672236e-01 -7.21804440e-01 1.12035728e+00
1.98899508e-01 8.91060010e-02 6.43569529e-01 -3.56983840e-01
3.00899535e-01 9.70328867e-01 7.16563016e-02 -9.65626776e-01
-6.53289139e-01 -6.91496909e-01 -1.03325748e+00 -2.80221760e-01
9.88023654e-02 7.13306740e-02 -9.32097495e-01 1.85706651e+00
3.18948291e-02 5.09387791e-01 2.27146804e-01 5.44073462e-01
9.07498360e-01 5.54183424e-01 3.63442361e-01 -5.61734498e-01
1.03759515e+00 -7.22116172e-01 -5.37594497e-01 3.73420835e-01
1.03349614e+00 -4.77215260e-01 9.73730743e-01 3.02929282e-01
-7.14751720e-01 -7.53928255e-03 -9.24099565e-01 1.12694077e-01
-4.21944737e-01 3.76847476e-01 8.56623650e-01 3.40407759e-01
-8.51519644e-01 8.13399136e-01 -5.53531408e-01 -1.71749681e-01
8.02531004e-01 9.26924527e-01 -3.37404698e-01 -2.99371958e-01
-1.28058410e+00 5.48395872e-01 1.71906874e-01 -7.93311372e-02
-1.18514979e+00 -9.86897171e-01 -6.26962066e-01 1.03614807e-01
2.24293843e-01 -9.09533799e-01 7.35500515e-01 -2.37837270e-01
-1.69131958e+00 5.37384748e-01 -2.75548726e-01 -1.58081740e-01
3.01328301e-03 -2.76133090e-01 -2.60344446e-01 -1.39424041e-01
3.01895663e-02 3.54608297e-01 -1.18946292e-01 -1.08159256e+00
-3.86532694e-01 -4.85627413e-01 -1.78021312e-01 2.20386267e-01
-5.62548041e-01 -1.63753778e-01 -5.37019730e-01 -5.64680696e-01
-2.96840787e-01 -8.76387417e-01 -6.05181754e-01 -4.50566262e-01
-7.60736585e-01 -3.65557581e-01 4.83092010e-01 -1.42799780e-01
1.52181244e+00 -1.78738081e+00 2.65086323e-01 1.45764768e-01
5.94799340e-01 2.45884329e-01 -2.75249898e-01 5.91784358e-01
-4.35271949e-01 4.57545161e-01 -1.91987127e-01 -9.05737504e-02
-2.44256750e-01 -3.56819779e-01 -2.35554516e-01 5.98577380e-01
2.39475727e-01 9.37999785e-01 -1.07331800e+00 -2.21596941e-01
-1.88004207e-02 7.32162654e-01 -5.46035230e-01 7.99872205e-02
-3.93944740e-01 4.10961241e-01 -8.31346750e-01 1.05460906e+00
4.53448802e-01 -6.79266036e-01 2.65404195e-01 -1.73126698e-01
1.79951906e-01 1.32714257e-01 -4.45215791e-01 1.55051291e+00
-1.57843351e-01 6.71434924e-02 -7.02415884e-01 -7.72353411e-01
8.50460351e-01 4.41527396e-01 1.15696454e+00 -5.69504380e-01
5.48980683e-02 5.31007409e-01 -7.24067837e-02 -7.15664402e-02
-2.19699256e-02 4.23363671e-02 -8.88782367e-02 5.55847846e-02
-3.48262995e-01 2.82470137e-01 1.05935603e-01 1.91596106e-01
1.48692584e+00 -1.66274741e-01 4.66732472e-01 -8.65917429e-02
6.29991829e-01 4.10677604e-02 7.87480950e-01 4.63988423e-01
-2.50556748e-02 4.76545721e-01 5.89429080e-01 -3.55908036e-01
-5.66550910e-01 -5.13051629e-01 -3.23931932e-01 8.09937656e-01
8.59476775e-02 -5.03776908e-01 -4.76502329e-01 -8.45003128e-01
-1.59684122e-02 3.65822762e-01 -6.79155231e-01 -2.92061687e-01
-2.63784498e-01 -1.38474870e+00 6.03964150e-01 2.36911416e-01
6.46739081e-02 -6.51337445e-01 3.43427688e-01 2.79201210e-01
2.07483247e-01 -5.98774612e-01 -5.86258411e-01 3.82623762e-01
-7.79210091e-01 -1.49521089e+00 -6.91052675e-01 -6.86815977e-01
8.08861494e-01 8.58879983e-02 8.46194625e-01 -1.00401966e-02
-2.89848030e-01 -2.77185112e-01 -9.67517644e-02 -1.51652887e-01
-5.82782775e-02 1.06907643e-01 2.65470952e-01 -6.20849580e-02
4.13143843e-01 -2.89665312e-01 -8.12675536e-01 4.46047455e-01
-8.49585950e-01 -7.41506591e-02 7.40020454e-01 1.18667746e+00
1.19961035e+00 1.07863873e-01 1.02087963e+00 -1.47869635e+00
8.38472962e-01 -7.91632235e-01 -5.35832226e-01 4.08523798e-01
-8.01585197e-01 7.54993260e-02 9.74232078e-01 -3.82750005e-01
-4.84704971e-01 4.34486747e-01 -3.01124215e-01 -2.93891758e-01
-1.70198251e-02 7.69534826e-01 -4.54871684e-01 1.95488092e-02
7.11338460e-01 2.53583819e-01 -3.07398349e-01 -2.24549025e-01
8.49442631e-02 6.41515553e-01 1.13528736e-01 -4.63196129e-01
3.94690752e-01 2.58041739e-01 3.22263002e-01 -1.81858674e-01
-7.74797142e-01 -5.34821272e-01 -2.33449608e-01 3.44041198e-01
6.32837117e-01 -9.71012175e-01 -9.94876206e-01 5.03736615e-01
-9.14253354e-01 -3.27856243e-01 3.14322226e-02 6.15369081e-01
-4.80660141e-01 1.73935264e-01 -8.52676272e-01 -4.04573113e-01
-4.22966808e-01 -1.47418904e+00 1.14843261e+00 1.18916489e-01
-2.73900151e-01 -1.16559029e+00 5.38154423e-01 2.76895702e-01
1.97174966e-01 5.96009552e-01 1.31536198e+00 -1.09396815e+00
-6.46402836e-01 -3.68288130e-01 -2.99673080e-01 -3.55043620e-01
7.18614936e-01 7.34025687e-02 -9.88117516e-01 -3.39507550e-01
-4.73003298e-01 -4.59981948e-01 1.03646874e+00 7.93796718e-01
1.61263919e+00 -3.92823279e-01 -9.46544886e-01 7.85901487e-01
1.63646793e+00 5.39243460e-01 6.85602009e-01 8.27294309e-03
7.94940233e-01 1.75652400e-01 4.90824640e-01 4.73907322e-01
1.65584862e-01 8.37847769e-01 3.53266060e-01 -5.59699178e-01
1.29348338e-01 -1.83430195e-01 1.87654719e-01 5.82708657e-01
2.01064229e-01 -4.93943334e-01 -7.74038911e-01 2.82637596e-01
-1.89928925e+00 -7.76980698e-01 -8.76881331e-02 2.40043187e+00
1.47128010e+00 -1.43449441e-01 7.01316446e-02 -1.59399852e-01
6.11796618e-01 -3.25085491e-01 -1.02408016e+00 -9.62519553e-03
-2.47172832e-01 1.68109298e-01 5.37169099e-01 5.61656654e-01
-1.13718820e+00 9.17706192e-01 6.66567039e+00 1.31184816e+00
-1.24001527e+00 -3.77024770e-01 1.33949304e+00 -7.67895654e-02
-5.39662898e-01 -1.99599966e-01 -1.02396095e+00 5.06011784e-01
8.77698898e-01 -2.55105883e-01 1.69297829e-01 6.00087285e-01
5.53149164e-01 3.64275962e-01 -1.47627521e+00 9.34394479e-01
-1.17341295e-01 -1.70618522e+00 3.10905576e-01 4.76693481e-01
9.70448732e-01 -1.19363800e-01 2.47016653e-01 2.09400490e-01
6.79082513e-01 -1.53814113e+00 -3.74812931e-01 7.85462797e-01
1.12622309e+00 -7.37665474e-01 7.04951406e-01 1.03298292e-01
-7.79145777e-01 -1.05915971e-01 -2.04385862e-01 3.71117234e-01
-2.98264146e-01 8.54762673e-01 -1.13887107e+00 6.16154909e-01
1.31281018e-01 1.14622700e+00 -5.56937158e-01 1.20553005e+00
2.11720034e-01 6.02491140e-01 1.15994684e-01 -5.63872866e-02
-7.40714148e-02 1.34534225e-01 4.09484804e-02 1.06073070e+00
4.00614560e-01 4.13949005e-02 2.81073570e-01 5.19793928e-01
-4.24802482e-01 4.45566595e-01 -6.96065545e-01 -3.82412463e-01
5.91518581e-01 1.19885027e+00 -4.46822107e-01 -2.13328287e-01
-1.46465987e-01 6.33036435e-01 3.02619815e-01 3.87878269e-01
-8.55302989e-01 -3.02772433e-01 7.68053412e-01 -2.42775455e-02
-3.41139048e-01 2.35226676e-01 -3.55861664e-01 -9.48948681e-01
-7.61301994e-01 -8.77406538e-01 6.25291586e-01 -3.46715361e-01
-1.62397993e+00 5.80311775e-01 -3.70455444e-01 -1.41711473e+00
1.99426357e-02 -7.32618451e-01 -3.30577016e-01 7.93801665e-01
-1.58801305e+00 -9.15178478e-01 -5.80998920e-02 5.91953576e-01
5.78912675e-01 -4.11787421e-01 1.18249512e+00 5.44736266e-01
-1.11986554e+00 7.29448557e-01 6.41880095e-01 -1.43977284e-01
9.66983318e-01 -1.08994114e+00 -3.20283979e-01 1.63659692e-01
-1.30920991e-01 9.61502850e-01 2.84619123e-01 -8.10559392e-01
-1.31972206e+00 -1.37771535e+00 6.76072598e-01 -4.49395508e-01
5.38329184e-01 -9.52796042e-02 -7.44055867e-01 9.41651538e-02
-1.05864108e-01 -1.13843523e-01 1.47946835e+00 2.06048816e-01
-2.41434336e-01 -1.46995068e-01 -9.24313068e-01 7.62839973e-01
8.47306371e-01 -3.66169423e-01 2.46437550e-01 8.56508911e-01
7.81024456e-01 -2.20605135e-01 -9.87090051e-01 6.65820241e-01
4.08117324e-01 -4.06138480e-01 1.08107984e+00 -8.23610544e-01
6.02756262e-01 -5.48368156e-01 -2.08837003e-01 -1.30414188e+00
-7.66733050e-01 -4.21558946e-01 -4.86413613e-02 1.04723835e+00
7.33642519e-01 -5.28458178e-01 9.58689868e-01 6.00171447e-01
-1.61633432e-01 -1.43619156e+00 -6.09894514e-01 -3.79451454e-01
1.79929525e-01 1.95685804e-01 6.14539921e-01 1.17116594e+00
1.13819391e-01 5.70360184e-01 -6.30757987e-01 1.24824569e-01
2.61060894e-01 2.64654364e-02 5.50030768e-01 -1.31730163e+00
-5.04447997e-01 -7.30512798e-01 -3.10646415e-01 -7.20299482e-01
1.94845200e-01 -1.12935817e+00 -1.10976249e-01 -1.68303263e+00
8.11655283e-01 -5.18171310e-01 -1.05086648e+00 7.24280059e-01
-5.13884306e-01 1.69135153e-01 -4.70925033e-01 4.24272895e-01
-6.63395822e-01 4.73215938e-01 1.09569514e+00 -5.55489302e-01
-3.17358524e-01 6.40954496e-03 -1.29956603e+00 4.25874412e-01
7.22392499e-01 -5.51170111e-01 -5.98440230e-01 -1.05167791e-01
3.54085028e-01 1.18250497e-01 -4.82173972e-02 -7.79347420e-01
2.77899563e-01 -7.37237871e-01 5.35626650e-01 -4.31037873e-01
3.62800181e-01 -6.55434072e-01 3.00002486e-01 5.37261367e-01
-6.36999786e-01 -3.49122345e-01 1.11143671e-01 7.65759885e-01
-2.99519181e-01 1.31189257e-01 6.70022905e-01 1.47886381e-01
-3.21914166e-01 8.48539352e-01 -1.38328671e-01 -3.52102727e-01
1.22720075e+00 -1.36650011e-01 -6.18315935e-01 -8.38004425e-02
-5.34547925e-01 4.00064647e-01 4.02372807e-01 -1.21757491e-02
7.82057941e-01 -1.52484870e+00 -8.34904671e-01 -1.30621105e-01
4.29037780e-01 -8.36633444e-02 1.87016204e-01 8.14313829e-01
-3.55604708e-01 7.61238635e-01 2.24639758e-01 -3.71757865e-01
-1.27133036e+00 4.92953211e-01 4.61308509e-01 -7.97988772e-01
9.91583169e-02 9.31559861e-01 7.32653201e-01 -4.22785193e-01
7.82945305e-02 -2.97361892e-03 -6.49692178e-01 -9.33950916e-02
4.57662940e-01 6.27285615e-02 2.47646570e-02 -4.31885779e-01
-5.79013824e-01 6.70132935e-01 -4.43918973e-01 5.19140959e-01
1.55895591e+00 4.92440343e-01 -8.43083858e-02 1.85540840e-01
1.35149169e+00 2.71133721e-01 -9.76492167e-01 -8.37871432e-03
6.21228293e-02 -4.51350510e-01 -2.69114040e-02 -1.14309084e+00
-9.06778932e-01 5.28668344e-01 5.13408542e-01 -4.00739789e-01
1.28617549e+00 -1.46015495e-01 6.87356830e-01 4.36310977e-01
1.60829991e-01 -7.29500294e-01 1.59195125e-01 3.94807935e-01
6.49005830e-01 -1.10109115e+00 2.71930426e-01 -5.54593146e-01
-5.73793709e-01 8.74555409e-01 5.89294255e-01 1.43432200e-01
4.60599124e-01 1.30238710e-02 -8.46092850e-02 -4.16537732e-01
-1.11354160e+00 6.10446110e-02 3.47301155e-01 2.72265702e-01
8.97611320e-01 1.24231055e-01 -5.06829143e-01 7.76734114e-01
4.82621700e-01 3.12226862e-01 2.12036014e-01 5.98118246e-01
-2.58053452e-01 -1.62562728e+00 -1.62103474e-02 7.26952493e-01
-6.50884092e-01 -3.78811359e-01 -6.89326704e-01 3.27054888e-01
-9.35893357e-02 7.72079527e-01 -5.37424982e-01 -7.43993938e-01
2.14316636e-01 -1.41284555e-01 3.10486704e-02 -8.75377655e-01
-3.88521492e-01 3.30888599e-01 -1.78716928e-01 -3.16863179e-01
-1.93032220e-01 -3.87486935e-01 -1.39067948e+00 8.70367289e-02
-5.62914252e-01 2.28808269e-01 3.16763610e-01 7.11456954e-01
7.75066853e-01 7.73984790e-01 1.00065088e+00 -3.64535689e-01
-3.47410917e-01 -6.57775998e-01 -5.77564061e-01 3.68000448e-01
1.96839273e-01 -7.01822281e-01 -4.91556935e-02 -9.66506675e-02] | [5.662940979003906, 5.7301130294799805] |
a8cec014-e3df-4043-bcfe-5fb994f8b9f2 | solving-cosine-similarity-underestimation | 2305.10610 | null | https://arxiv.org/abs/2305.10610v1 | https://arxiv.org/pdf/2305.10610v1.pdf | Solving Cosine Similarity Underestimation between High Frequency Words by L2 Norm Discounting | Cosine similarity between two words, computed using their contextualised token embeddings obtained from masked language models (MLMs) such as BERT has shown to underestimate the actual similarity between those words (Zhou et al., 2022). This similarity underestimation problem is particularly severe for highly frequent words. Although this problem has been noted in prior work, no solution has been proposed thus far. We observe that the L2 norm of contextualised embeddings of a word correlates with its log-frequency in the pretraining corpus. Consequently, the larger L2 norms associated with the highly frequent words reduce the cosine similarity values measured between them, thus underestimating the similarity scores. To solve this issue, we propose a method to discount the L2 norm of a contextualised word embedding by the frequency of that word in a corpus when measuring the cosine similarities between words. We show that the so called stop words behave differently from the rest of the words, which require special consideration during their discounting process. Experimental results on a contextualised word similarity dataset show that our proposed discounting method accurately solves the similarity underestimation problem. | ['Danushka Bollegala', 'Yi Zhou', 'Saeth Wannasuphoprasit'] | 2023-05-17 | null | null | null | null | ['word-similarity'] | ['natural-language-processing'] | [-7.76382862e-03 9.33247805e-02 -9.21374410e-02 -5.49219370e-01
-4.45421070e-01 -3.10975373e-01 6.95496023e-01 8.72429848e-01
-1.15215647e+00 3.46037328e-01 3.97196263e-01 -1.72169924e-01
1.10992201e-01 -6.49634302e-01 -2.59816438e-01 -8.26860547e-01
-1.23855390e-01 2.11784810e-01 1.61388993e-01 -1.88260168e-01
5.87633610e-01 8.49372670e-02 -1.68619609e+00 9.06759202e-02
8.11486065e-01 8.03994179e-01 5.70255756e-01 2.92285353e-01
-3.47298652e-01 1.54493541e-01 -8.92839670e-01 -5.18328846e-01
2.30243579e-01 -5.22186875e-01 -3.71328950e-01 -2.52241850e-01
6.50302708e-01 7.02401176e-02 1.16701804e-01 1.40681648e+00
4.76428062e-01 5.15477598e-01 6.45801723e-01 -9.77567911e-01
-9.56875801e-01 8.45146477e-01 -3.17858130e-01 5.90873420e-01
6.12126887e-02 -2.61312932e-01 1.47859550e+00 -1.27984679e+00
2.76873350e-01 1.02998424e+00 7.84127533e-01 4.83228236e-01
-1.06995356e+00 -4.72548604e-01 3.88482630e-01 2.52965152e-01
-1.76523018e+00 -2.80176699e-02 5.17387390e-01 -3.27182382e-01
1.01646388e+00 9.25988257e-02 5.84194064e-01 8.24018180e-01
2.79291213e-01 3.23100835e-01 9.92237449e-01 -6.79165900e-01
4.12185520e-01 5.36948144e-01 1.83572307e-01 3.07692856e-01
4.42121923e-01 1.89237972e-03 -4.79536831e-01 -1.46947265e-01
4.43359278e-02 -7.21185952e-02 -9.48387478e-03 -1.95368424e-01
-1.00904930e+00 1.06174064e+00 1.24287173e-01 8.11925113e-01
-2.71158606e-01 1.20888814e-01 4.53604251e-01 1.49087042e-01
6.84859514e-01 4.67420429e-01 -2.44360834e-01 -1.32482514e-01
-9.31224227e-01 3.42403203e-01 2.88954645e-01 8.25497448e-01
7.56262839e-01 -7.65505508e-02 -8.79832730e-02 9.39257979e-01
3.55201364e-01 2.58126795e-01 1.09116948e+00 -4.23406005e-01
2.92597413e-01 2.31967330e-01 5.96401468e-02 -1.35592282e+00
-1.83297068e-01 -3.57356757e-01 -2.43847236e-01 -1.57907873e-01
5.45905530e-01 6.27877861e-02 -3.87767196e-01 2.24367094e+00
1.73298120e-01 2.34461278e-01 -3.47720012e-02 9.72805023e-01
2.35538349e-01 5.55018187e-01 5.69228172e-01 -2.97472090e-01
1.52205861e+00 -6.72641397e-01 -8.89368176e-01 -6.30460918e-01
9.58326280e-01 -9.47205186e-01 1.34143472e+00 1.45548955e-01
-9.48642790e-01 -5.74678540e-01 -1.29868388e+00 -3.20240706e-02
-5.92782915e-01 -2.78356612e-01 4.04298484e-01 9.41679955e-01
-6.80999875e-01 7.86895633e-01 -4.61531073e-01 -3.66530567e-01
-1.27423778e-01 -1.13861691e-02 -2.22772464e-01 1.42355785e-01
-1.66449559e+00 1.36627424e+00 4.47066426e-01 2.90761925e-02
-3.47366095e-01 -6.38620257e-01 -9.10347939e-01 1.52045995e-01
3.94956879e-02 -2.86610946e-02 9.83310342e-01 -9.05072272e-01
-7.16295540e-01 8.17654967e-01 -2.28449494e-01 -5.57324827e-01
1.01853542e-01 -1.81861117e-01 -6.85316622e-01 -3.32332253e-01
1.30408719e-01 2.85210043e-01 9.28678036e-01 -9.46504295e-01
-4.95957911e-01 -3.03668916e-01 -2.36418098e-01 2.95715690e-01
-7.58380711e-01 -1.96393095e-02 3.14646959e-01 -9.50891137e-01
5.92538230e-02 -7.94823527e-01 -2.33751222e-01 -1.21401288e-01
1.63743898e-01 -5.56137145e-01 2.31339127e-01 -2.87962973e-01
1.59251750e+00 -2.41765690e+00 -3.08902264e-01 1.12815924e-01
7.23359659e-02 3.72411430e-01 -3.46630603e-01 6.36680484e-01
-3.12207252e-01 8.03045928e-03 -3.23771358e-01 -4.06608164e-01
1.91475987e-01 3.30685556e-01 -3.34819973e-01 6.81204855e-01
2.16365889e-01 3.40897232e-01 -1.25693798e+00 -4.10852402e-01
-7.37111270e-02 4.34793651e-01 -5.70497692e-01 1.29138440e-01
3.87185141e-02 -4.55822468e-01 1.45152390e-01 -1.52780622e-01
8.50249529e-01 3.89041960e-01 3.32521051e-01 -1.81066334e-01
-2.19495252e-01 6.98707998e-01 -1.22799826e+00 1.38253021e+00
-7.05815792e-01 7.20794141e-01 -5.09265602e-01 -1.08656836e+00
1.16868830e+00 3.72024328e-01 2.41495552e-03 -7.72879541e-01
5.49353138e-02 2.85410136e-01 2.79771239e-01 -3.54278862e-01
1.04415727e+00 -8.79137576e-01 -9.90491584e-02 4.71732944e-01
-2.14348450e-01 -5.75449653e-02 2.97315776e-01 1.70523673e-01
7.03463376e-01 -3.31206024e-01 5.01934171e-01 -4.78208572e-01
5.78525603e-01 -4.30758148e-01 5.63953221e-01 5.93007028e-01
-4.93009329e-01 5.12408495e-01 3.20348650e-01 -3.49091180e-02
-1.07052624e+00 -1.21121407e+00 -3.79159570e-01 1.17998433e+00
1.31325275e-01 -7.42880940e-01 -7.13734686e-01 -4.54121232e-01
1.06696323e-01 1.18305564e+00 -7.75216341e-01 -6.29882753e-01
-5.17660618e-01 -6.18683577e-01 3.89805317e-01 5.62670827e-01
-1.36035994e-01 -7.15730131e-01 -6.34172320e-01 3.10168028e-01
-9.79388356e-02 -9.67940748e-01 -8.81633162e-01 2.58658230e-01
-7.94192016e-01 -6.26341641e-01 -5.63496649e-01 -8.73252988e-01
6.91726804e-01 3.84429097e-01 9.93030190e-01 4.63269614e-02
-1.59545124e-01 2.29031481e-02 -4.51722890e-01 -3.69966060e-01
-4.31571126e-01 -3.87695253e-01 5.16762137e-01 -2.41352301e-02
1.06303978e+00 -3.60035211e-01 -5.77992439e-01 1.96836606e-01
-1.11030269e+00 -5.39233029e-01 3.37739438e-01 9.06196713e-01
3.76435727e-01 9.30840522e-02 5.27258992e-01 -6.86686754e-01
1.02861500e+00 -5.46168745e-01 -2.33861059e-01 6.86774999e-02
-1.04055882e+00 3.59050065e-01 7.51799822e-01 -9.58012402e-01
-8.01250398e-01 -4.85827893e-01 -4.73825354e-03 1.63704768e-01
-7.04022963e-03 4.83713746e-01 1.46530673e-01 3.49934220e-01
6.47670269e-01 2.64934450e-01 -2.00232521e-01 -3.62670094e-01
4.86317247e-01 7.71370828e-01 1.45436570e-01 -4.23370838e-01
5.43353379e-01 1.98981062e-01 -2.83914328e-01 -8.55817318e-01
-8.78075242e-01 -5.60407162e-01 -3.52699488e-01 1.51130602e-01
6.72930241e-01 -7.41146803e-01 -3.24120551e-01 -1.91013232e-01
-1.16785288e+00 3.51622969e-01 -4.14578348e-01 8.61641824e-01
-2.88774163e-01 7.81484425e-01 -1.68067947e-01 -1.00321138e+00
-7.56517500e-02 -8.26924801e-01 5.82438290e-01 -3.54648083e-02
-8.68596733e-01 -1.24978459e+00 3.26241642e-01 -2.43259016e-02
5.61704457e-01 -2.59913772e-01 1.28768015e+00 -1.03729701e+00
2.20599294e-01 -4.03239995e-01 -2.15890352e-03 8.09563518e-01
2.53382921e-01 -1.81523204e-01 -8.28076541e-01 -1.97094455e-01
4.55680251e-01 5.35503589e-02 7.44162381e-01 4.04162146e-02
6.21978760e-01 -2.17798412e-01 4.49712053e-02 -1.10418700e-01
1.40508842e+00 4.70616706e-02 3.96042764e-01 1.29149765e-01
2.60856062e-01 7.28030682e-01 9.23781812e-01 5.14818311e-01
1.21474616e-01 6.28480613e-01 2.65381724e-01 3.90567362e-01
9.62758660e-02 -3.67197692e-01 5.67720473e-01 1.35345936e+00
3.69138628e-01 -1.07923657e-01 -7.68985748e-01 1.01187456e+00
-1.37668931e+00 -7.84357369e-01 -2.34027684e-01 2.62148714e+00
1.04513061e+00 2.79622734e-01 -1.54654741e-01 3.33562136e-01
8.29637289e-01 4.37195331e-01 -1.17412642e-01 -1.03906584e+00
-1.29821926e-01 2.73114443e-01 3.81450444e-01 7.25002646e-01
-5.23162663e-01 9.32640791e-01 6.13043642e+00 9.71534908e-01
-8.82944167e-01 2.68793494e-01 8.08458477e-02 -2.57982701e-01
-5.19042492e-01 1.54847279e-01 -7.22745299e-01 7.06656277e-01
1.20016313e+00 -7.86564469e-01 7.25419372e-02 8.58981550e-01
2.66020358e-01 -4.58801210e-01 -1.26722240e+00 8.05434883e-01
3.38537902e-01 -6.96197629e-01 1.32516548e-01 -1.65432215e-01
5.18935621e-01 -1.86431676e-01 3.43386561e-01 3.22864026e-01
-2.95409948e-01 -8.62294018e-01 8.15623701e-01 2.50497669e-01
3.39561045e-01 -1.01687253e+00 8.16542864e-01 3.04492682e-01
-1.06465232e+00 1.65255994e-01 -9.20029640e-01 -3.10508639e-01
1.87306833e-02 1.09954178e+00 -8.46865833e-01 -1.06821820e-01
2.19367012e-01 2.64835656e-01 -6.02899611e-01 6.50828063e-01
-2.24751785e-01 4.59317565e-01 -8.58302861e-02 -3.98381889e-01
2.80449212e-01 -2.83237249e-01 4.07104582e-01 1.29460239e+00
3.57714087e-01 -2.11387545e-01 -4.11413968e-01 8.96194339e-01
1.16609022e-01 4.80199665e-01 -4.90328699e-01 -3.64129663e-01
6.07413113e-01 1.02679169e+00 -4.02177244e-01 -4.78899777e-01
-5.57901144e-01 8.44698071e-01 4.41171408e-01 4.14731763e-02
-7.91031003e-01 -6.09415650e-01 1.15269864e+00 1.11994654e-01
2.63112932e-01 -4.18472022e-01 -2.67536789e-01 -7.86574543e-01
1.22300677e-01 -5.35606027e-01 1.18805945e-01 -3.69538844e-01
-1.49410653e+00 7.27947295e-01 -4.08375151e-02 -1.13987577e+00
-1.98416904e-01 -6.02007508e-01 -3.93622100e-01 1.12087071e+00
-1.36044872e+00 -8.48366320e-02 1.79006994e-01 1.86436638e-01
4.72842902e-01 1.25285968e-01 8.88333738e-01 3.93178225e-01
-1.70359418e-01 7.57049561e-01 6.46505505e-02 -1.41419545e-01
9.37041819e-01 -1.17825902e+00 5.88312447e-01 7.11843729e-01
3.19075465e-01 1.16452086e+00 1.16405642e+00 -6.67664766e-01
-8.11125815e-01 -7.47198403e-01 1.76769543e+00 -3.70481253e-01
9.12530780e-01 -3.94360304e-01 -1.14172614e+00 2.24631459e-01
1.00706272e-01 -1.10603711e-02 1.13601387e+00 1.77955609e-02
-7.50403643e-01 -6.87482860e-03 -1.02152050e+00 8.32172751e-01
9.32555199e-01 -8.66985023e-01 -1.02903450e+00 3.02659541e-01
7.97415912e-01 2.89374441e-01 -6.94867909e-01 3.30093652e-02
5.32793522e-01 -1.16656685e+00 7.56538928e-01 -5.09405494e-01
3.02680492e-01 -1.51821315e-01 -3.92923594e-01 -1.61027277e+00
-1.50044531e-01 -2.75760144e-02 1.33669853e-01 1.03139198e+00
3.22407037e-01 -6.10574901e-01 5.49833357e-01 4.96482491e-01
2.29477867e-01 -6.59480095e-01 -1.14138913e+00 -1.27423060e+00
4.42914665e-01 -6.37274742e-01 4.98115599e-01 9.08237994e-01
3.87009770e-01 9.11945626e-02 -9.59359929e-02 3.27072889e-02
5.20701051e-01 -2.56184876e-01 -5.80045208e-02 -1.01226318e+00
5.84482290e-02 -4.15430039e-01 -6.86842918e-01 -9.29996133e-01
1.83588371e-01 -1.04116571e+00 2.71764368e-01 -9.57563639e-01
-5.54572940e-02 -3.84599507e-01 -4.00462061e-01 -9.19499844e-02
-4.46815789e-01 1.28854334e-01 1.07903637e-01 -1.34342223e-01
-1.54353812e-01 6.93834901e-01 7.90947258e-01 5.74406981e-02
1.86080441e-01 -4.03840423e-01 -5.78340948e-01 6.48496449e-01
8.66956651e-01 -8.98696542e-01 -4.22994643e-01 -4.36329752e-01
4.60668355e-01 -7.17128158e-01 8.64245296e-02 -9.37226713e-01
1.29473865e-01 -5.90023026e-02 -1.30676240e-01 -3.65589917e-01
2.50800818e-01 -1.03328633e+00 -3.10898513e-01 5.33832192e-01
-7.27365375e-01 4.49632198e-01 9.01422352e-02 5.34387052e-01
-2.93064892e-01 -8.29959452e-01 7.83450484e-01 -8.60022102e-03
-4.90347296e-01 -1.71792805e-01 -6.98350966e-01 4.05515969e-01
9.20123279e-01 -2.49143273e-01 7.75131434e-02 -1.66851386e-01
-5.17810583e-01 -2.71690935e-01 3.19921970e-01 6.86252415e-01
6.67687178e-01 -1.48596179e+00 -5.72462738e-01 2.71858692e-01
4.42249715e-01 -6.20933652e-01 1.28099173e-01 7.03969300e-01
-1.75675392e-01 3.75154585e-01 5.93881831e-02 -6.60551563e-02
-1.29818201e+00 7.73283541e-01 1.15504488e-01 8.63569975e-02
-2.27918237e-01 9.01576459e-01 7.44303092e-02 -1.00986674e-01
1.80147618e-01 -5.64303577e-01 -1.26698809e-02 5.21808803e-01
7.25875676e-01 2.91883916e-01 2.32120425e-01 -7.56493628e-01
-5.93514442e-01 5.44341505e-01 -3.39817017e-01 -2.64126003e-01
1.03742158e+00 -2.60192096e-01 -1.23635948e-01 8.23554873e-01
1.52901995e+00 2.85524398e-01 -6.86000645e-01 -5.38073540e-01
5.85850358e-01 -7.54979491e-01 5.34082688e-02 -2.20241010e-01
-6.43363416e-01 9.44842756e-01 6.78057253e-01 3.09369177e-01
6.74532115e-01 -1.19879514e-01 8.97861362e-01 1.55922741e-01
1.91399217e-01 -1.64914715e+00 -1.63068268e-02 4.99844223e-01
5.62020421e-01 -1.01155877e+00 8.40349868e-02 -2.21660435e-01
-7.82341957e-01 9.34347391e-01 3.61070663e-01 -3.16771120e-01
5.50426662e-01 3.92501652e-02 1.94923759e-01 2.12034881e-01
-7.70973742e-01 -1.88466892e-01 8.44667777e-02 6.53274953e-01
7.74160862e-01 8.38506967e-02 -1.18835282e+00 4.94353831e-01
-4.09967512e-01 -6.85254395e-01 6.90520108e-01 8.30240726e-01
-6.25860095e-01 -1.08837259e+00 -2.74664402e-01 5.34818135e-02
-1.96832001e-01 -7.58968890e-01 -7.65757784e-02 4.51570421e-01
3.77884597e-01 9.05874908e-01 4.80450690e-01 -2.95265973e-01
2.82244116e-01 5.03896475e-01 4.36317801e-01 -7.85574436e-01
-6.25820518e-01 -5.41120827e-01 -6.89750239e-02 -3.14618200e-01
-2.31843650e-01 -6.48483694e-01 -1.28161287e+00 -2.44670630e-01
-4.97707367e-01 4.12840992e-01 1.08568764e+00 9.61852431e-01
4.61324789e-02 1.58971906e-01 6.28743947e-01 -3.06494147e-01
-9.50528920e-01 -9.27601635e-01 -8.84367645e-01 7.89528728e-01
2.07446069e-01 -7.27869511e-01 -9.52726245e-01 -4.07337964e-01] | [10.484764099121094, 8.866049766540527] |
5d6666e3-75d1-42b7-8336-d68878dee00a | exploiting-large-neuroimaging-datasets-to | 2305.17300 | null | https://arxiv.org/abs/2305.17300v1 | https://arxiv.org/pdf/2305.17300v1.pdf | Exploiting Large Neuroimaging Datasets to Create Connectome-Constrained Approaches for more Robust, Efficient, and Adaptable Artificial Intelligence | Despite the progress in deep learning networks, efficient learning at the edge (enabling adaptable, low-complexity machine learning solutions) remains a critical need for defense and commercial applications. We envision a pipeline to utilize large neuroimaging datasets, including maps of the brain which capture neuron and synapse connectivity, to improve machine learning approaches. We have pursued different approaches within this pipeline structure. First, as a demonstration of data-driven discovery, the team has developed a technique for discovery of repeated subcircuits, or motifs. These were incorporated into a neural architecture search approach to evolve network architectures. Second, we have conducted analysis of the heading direction circuit in the fruit fly, which performs fusion of visual and angular velocity features, to explore augmenting existing computational models with new insight. Our team discovered a novel pattern of connectivity, implemented a new model, and demonstrated sensor fusion on a robotic platform. Third, the team analyzed circuitry for memory formation in the fruit fly connectome, enabling the design of a novel generative replay approach. Finally, the team has begun analysis of connectivity in mammalian cortex to explore potential improvements to transformer networks. These constraints increased network robustness on the most challenging examples in the CIFAR-10-C computer vision robustness benchmark task, while reducing learnable attention parameters by over an order of magnitude. Taken together, these results demonstrate multiple potential approaches to utilize insight from neural systems for developing robust and efficient machine learning techniques. | ['Joan A. Hoffmann', 'William R. Gray-Roncal', 'Brock A. Wester', 'I-Jeng Wang', 'Matthew J. Roos', 'Nathan Drenkow', 'Elizabeth P. Reilly', 'Patricia K. Rivlin', 'Lindsey Kitchell', 'Arun V. Reddy', 'Michael S. Robinette', 'Martha Cervantes', 'Marisel Villafañe-Delgado', 'Isaac Western', 'Raphael Norman-Tenazas', 'Jordan K. Matelsky', 'Justin Joyce', 'Gautam K. Vallabha', 'Brian S. Robinson', 'Erik C. Johnson'] | 2023-05-26 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 2.80398726e-01 1.72175586e-01 4.00276572e-01 -2.59931177e-01
-3.04063827e-01 -8.44951212e-01 5.80604255e-01 -1.35112122e-01
-5.30405164e-01 5.43024361e-01 -1.58521030e-02 -1.81281745e-01
-4.39929247e-01 -3.58478040e-01 -9.42075908e-01 -6.08949780e-01
-5.20526886e-01 3.52041900e-01 3.32987681e-02 -2.19289511e-01
4.35457021e-01 7.71884024e-01 -1.28194392e+00 3.38511020e-01
4.12066221e-01 9.40281153e-01 3.72732967e-01 5.01966476e-01
4.96143550e-01 5.61602890e-01 -3.70608866e-01 -1.33172899e-01
4.53619480e-01 -3.86521667e-01 -6.95010066e-01 -4.34313893e-01
4.10446435e-01 -9.94976312e-02 -2.33636633e-01 7.50197470e-01
6.93749666e-01 -2.48281229e-02 3.81539643e-01 -1.33104527e+00
-3.03035647e-01 8.13224852e-01 -2.10964605e-01 6.47117853e-01
-1.56551883e-01 6.00202680e-01 9.17515159e-01 -4.82024491e-01
8.58119071e-01 1.03523636e+00 8.99861574e-01 6.41298890e-01
-1.64393198e+00 -8.77844572e-01 -4.65188734e-02 1.62928239e-01
-1.06293154e+00 -5.76848328e-01 4.98131454e-01 -3.68543446e-01
1.44554234e+00 -1.51186779e-01 1.25429702e+00 1.24162138e+00
5.29139340e-01 5.16813338e-01 9.08346176e-01 9.20620039e-02
4.73519355e-01 -3.10540706e-01 1.45519106e-02 9.54439938e-01
9.19980332e-02 2.74500549e-01 -5.72947264e-01 -1.61479667e-01
7.69249320e-01 -1.12322733e-01 -1.62482232e-01 -1.74432516e-01
-1.23631656e+00 5.89107811e-01 7.16797531e-01 2.28148401e-01
-2.87449628e-01 4.05553997e-01 2.87604362e-01 1.92447066e-01
-9.38892141e-02 1.06870294e+00 -6.38889670e-01 -1.49826244e-01
-8.31270158e-01 2.82913804e-01 7.29970217e-01 4.21132028e-01
5.98512352e-01 2.48036116e-01 1.38062105e-01 5.22306502e-01
3.70333642e-01 1.92677841e-01 4.88401622e-01 -1.44444442e+00
3.34900580e-02 7.23446131e-01 -5.72548747e-01 -9.90160942e-01
-1.14109266e+00 -7.15372264e-01 -5.98407745e-01 6.10958278e-01
3.02270502e-01 -2.72219211e-01 -6.93463802e-01 1.97080040e+00
8.87857303e-02 1.45009398e-01 -2.06119522e-01 8.25995862e-01
3.42600286e-01 2.93220878e-01 -1.24929339e-01 2.86558270e-01
1.16460705e+00 -5.99879444e-01 -8.43015220e-03 -2.39427432e-01
5.94758093e-01 -9.69985798e-02 6.11609161e-01 7.34868526e-01
-1.09444201e+00 -3.07168961e-01 -1.31365597e+00 1.68720424e-01
-6.14387512e-01 -2.09532440e-01 8.95063102e-01 3.63484114e-01
-1.28897429e+00 9.19201791e-01 -1.09509826e+00 -5.43045819e-01
8.60752046e-01 8.34975421e-01 -3.50233436e-01 8.69707912e-02
-7.47494578e-01 9.40442085e-01 2.81605244e-01 -8.55636969e-02
-1.43832660e+00 -9.98694301e-01 -3.97164702e-01 3.09228748e-01
-1.37639448e-01 -1.15446997e+00 8.62015188e-01 -8.13254118e-01
-1.39481580e+00 5.86584806e-01 2.54168272e-01 -8.60119283e-01
-1.39800489e-01 1.82015449e-01 -1.03829779e-01 2.52910614e-01
-2.64335692e-01 1.39933491e+00 6.72365844e-01 -7.36423552e-01
-3.02351028e-01 -5.44312179e-01 -8.93707573e-03 9.37159136e-02
-3.30021024e-01 -7.69540593e-02 4.41089794e-02 -5.64430952e-01
9.30088572e-03 -1.10921979e+00 -2.40244493e-01 2.66599089e-01
-6.64788708e-02 1.42838910e-01 7.65293598e-01 -4.88221139e-01
4.88751262e-01 -1.92948151e+00 6.27227128e-01 3.54968458e-01
5.02443552e-01 1.66577637e-01 -4.85247582e-01 1.19917341e-01
-3.72935623e-01 2.44752660e-01 -2.31540561e-01 -1.40016945e-02
-2.45853141e-01 -6.03217669e-02 -1.63288951e-01 3.56911331e-01
5.63074529e-01 1.14278877e+00 -5.77131450e-01 1.37220278e-01
-1.55316338e-01 5.79491198e-01 -8.58790338e-01 -1.61089435e-01
-2.91420519e-01 4.54397857e-01 -1.11608930e-01 6.22096896e-01
2.07960397e-01 -2.36764610e-01 8.30336660e-02 -4.05888379e-01
3.78187327e-03 1.41375989e-01 -6.18540764e-01 2.01892209e+00
-3.45935076e-01 8.74185979e-01 2.54957944e-01 -1.15930581e+00
7.27166653e-01 1.48599856e-02 7.35666573e-01 -7.61219978e-01
4.69642252e-01 -7.45451376e-02 6.51733935e-01 -3.04847300e-01
-2.81961650e-01 3.58404696e-01 2.24268094e-01 4.59058791e-01
5.23325801e-01 -1.55089483e-01 6.54496476e-02 2.09125072e-01
1.97614956e+00 1.86107963e-01 -2.68471807e-01 -3.87057602e-01
-1.70624375e-01 3.18788975e-01 4.40450639e-01 4.50893909e-01
-2.46099770e-01 5.68594754e-01 4.73708719e-01 -3.73626232e-01
-1.07626784e+00 -1.13565254e+00 -5.71769662e-02 9.06157076e-01
-5.15990734e-01 -4.53510374e-01 -9.70679998e-01 -2.35931933e-01
-1.32160231e-01 3.18947732e-01 -6.72967613e-01 -6.22379124e-01
-3.93238217e-01 -8.67518187e-01 1.16694438e+00 3.95258754e-01
5.27207375e-01 -9.78034735e-01 -1.20471931e+00 2.87878871e-01
2.58273035e-01 -7.87351966e-01 9.68946740e-02 7.22783446e-01
-9.76840556e-01 -1.23597431e+00 -4.68352050e-01 -6.64800525e-01
5.97135425e-01 4.53973003e-02 6.98970735e-01 -2.53743678e-02
-1.08968484e+00 5.34078598e-01 6.23135231e-02 -4.03934658e-01
2.15359647e-02 1.27223790e-01 2.66737968e-01 -6.22047186e-01
5.70879206e-02 -9.85016406e-01 -5.83274126e-01 2.28806660e-01
-8.56861830e-01 -4.04514037e-02 7.45211661e-01 9.99434948e-01
5.58842599e-01 7.38137076e-03 7.80403793e-01 -3.69076461e-01
7.11029172e-01 -7.49398649e-01 -7.72878468e-01 9.70191136e-02
-6.94323301e-01 3.49559397e-01 5.91636300e-01 -5.35771251e-01
-7.04954863e-01 3.57380062e-01 -2.84483396e-02 -4.41696107e-01
-5.23960330e-02 5.32499492e-01 1.16900206e-02 -4.92738962e-01
8.89812350e-01 -6.17995672e-02 3.99181575e-01 -8.54165331e-02
3.35547924e-01 -6.73350692e-02 5.33059239e-01 -4.50409889e-01
5.17995954e-01 3.88581425e-01 2.75946945e-01 -7.25522220e-01
-1.64771840e-01 6.70014098e-02 -5.95008373e-01 -1.27447397e-01
8.09162557e-01 -8.19570720e-01 -8.62023115e-01 4.52225596e-01
-9.67972398e-01 -6.63498223e-01 -1.69526249e-01 5.27049065e-01
-6.53812170e-01 -1.08029976e-01 -6.54840112e-01 -1.71056882e-01
-4.26835328e-01 -1.16916931e+00 6.71883404e-01 4.72591847e-01
-4.89176065e-01 -8.23959291e-01 2.86260992e-01 2.78307199e-01
6.92723453e-01 1.25903398e-01 1.20184922e+00 -7.25583255e-01
-6.80554748e-01 1.14896394e-01 -1.79439574e-01 4.03942242e-02
-1.85646474e-01 6.54478669e-02 -9.62754011e-01 -3.08113784e-01
-1.64888933e-01 -5.71772158e-01 8.59756291e-01 3.76964033e-01
1.21189427e+00 -1.17020376e-01 -5.62301993e-01 9.80703890e-01
9.73091424e-01 2.52072573e-01 4.53460962e-01 4.23986703e-01
2.74709523e-01 5.92586398e-01 -1.20116889e-01 1.14144199e-01
9.26101506e-02 4.66812611e-01 5.95458329e-01 4.39518206e-02
-2.67598212e-01 9.71981958e-02 4.70536947e-01 4.23410803e-01
5.07303253e-02 -1.13172280e-02 -9.91490543e-01 2.60320634e-01
-1.68186760e+00 -1.05770612e+00 4.18238074e-01 1.75898528e+00
6.16035402e-01 2.72605181e-01 -3.08202840e-02 -2.08799154e-01
3.31077933e-01 -4.00763184e-01 -1.28361201e+00 -2.21642211e-01
-2.13465795e-01 4.31379616e-01 4.30184513e-01 1.15207151e-01
-6.63424551e-01 8.72969031e-01 7.35857058e+00 4.07000154e-01
-1.23848271e+00 -4.65272926e-02 6.59498155e-01 -6.84191108e-01
-1.24335892e-01 -4.79774475e-02 -7.47020721e-01 2.50066698e-01
1.42362154e+00 -2.41435990e-02 1.20879865e+00 6.07870340e-01
7.20610023e-02 -1.60854310e-02 -1.20054340e+00 8.68869245e-01
8.54428932e-02 -1.79972029e+00 -3.76076519e-01 2.33184382e-01
5.24930179e-01 8.25191021e-01 2.61607081e-01 8.92155096e-02
4.54982728e-01 -1.27026713e+00 3.05619299e-01 7.17878163e-01
4.07047004e-01 -6.86624348e-01 2.20094457e-01 2.97829986e-01
-8.37501943e-01 -5.26882887e-01 -3.48715574e-01 2.61949208e-02
-1.27776936e-01 4.18137759e-01 -9.46058333e-01 -1.59491822e-01
1.03760505e+00 7.99060464e-01 -9.12152946e-01 1.27538204e+00
1.30967334e-01 5.14337480e-01 -6.36599064e-01 -5.78157008e-02
-5.57732768e-02 1.46793589e-01 6.20265484e-01 9.33890343e-01
2.37235725e-01 -1.16886934e-02 -3.53018135e-01 1.39692962e+00
-2.76275039e-01 -5.57959855e-01 -9.07947004e-01 -2.84449399e-01
5.25742173e-01 1.74565303e+00 -9.82487202e-01 1.62806571e-01
1.15299374e-01 8.28024030e-01 5.55838943e-01 2.75681943e-01
-8.38438332e-01 -3.38126302e-01 6.69564009e-01 -1.04207307e-01
2.10499704e-01 -3.67203444e-01 -3.37939292e-01 -7.60916531e-01
-3.41241270e-01 -1.02967894e+00 2.59112976e-02 -1.02097404e+00
-1.04841709e+00 4.97718722e-01 -5.96571527e-02 -6.09879255e-01
-3.99139881e-01 -7.97924459e-01 -7.82469630e-01 6.44535720e-01
-8.66493821e-01 -9.76825178e-01 -1.30764186e-01 6.02381885e-01
3.98075581e-01 -6.12178862e-01 7.85146236e-01 2.22094670e-01
-8.18516314e-01 4.48982358e-01 -2.21281894e-03 -2.22339541e-01
4.15289998e-01 -7.87023187e-01 4.90776986e-01 6.55305803e-01
4.15911108e-01 9.22988534e-01 2.99590409e-01 -6.27562940e-01
-1.69427514e+00 -9.39979970e-01 -5.55137135e-02 -4.04885083e-01
6.99796200e-01 -4.61225480e-01 -7.54876196e-01 6.04525924e-01
3.11297596e-01 -1.96943223e-01 6.25631094e-01 -7.80553445e-02
-3.91898751e-01 -1.54043421e-01 -1.40404129e+00 6.59473479e-01
1.19239247e+00 -3.73552501e-01 -4.74245042e-01 1.02910802e-01
6.92265689e-01 2.03014575e-02 -8.63715053e-01 3.10851991e-01
8.23422313e-01 -6.93790138e-01 9.17942822e-01 -6.86161637e-01
2.51353860e-01 -2.48165682e-01 -4.81244884e-02 -1.77396929e+00
-5.56735814e-01 -6.16511047e-01 -1.89328283e-01 1.16771114e+00
6.69755161e-01 -5.30843794e-01 8.46765339e-01 5.82699776e-01
-3.68967980e-01 -8.85342538e-01 -1.01191962e+00 -5.90131998e-01
1.86279371e-01 -1.49163768e-01 6.26993030e-02 7.03155279e-01
2.71248639e-01 5.13805509e-01 2.47259036e-01 -1.75232366e-01
4.28588301e-01 -5.44512093e-01 3.42418790e-01 -1.30984771e+00
-5.71890414e-01 -8.86187077e-01 -5.42513907e-01 -4.96333778e-01
2.96893001e-01 -1.26508915e+00 -1.58273235e-01 -1.32229447e+00
6.71909824e-02 -3.65925461e-01 -3.12222898e-01 8.65264595e-01
4.84875709e-01 2.04666078e-01 1.35388419e-01 8.26126635e-02
-2.54148781e-01 2.47763038e-01 6.88034415e-01 -3.33223581e-01
8.67681578e-03 -4.64726031e-01 -9.80752110e-01 7.49498010e-01
1.17756045e+00 -5.16811728e-01 -5.29856443e-01 -7.17632711e-01
4.60448146e-01 -4.05599385e-01 6.49837852e-01 -1.66483879e+00
4.96943861e-01 3.29706520e-01 1.00204003e+00 -1.43672556e-01
4.79470193e-01 -8.95342946e-01 2.10554630e-01 7.90093005e-01
-4.74853724e-01 3.40020090e-01 7.26197481e-01 4.09817398e-01
3.30010504e-01 3.99924815e-02 8.35940480e-01 -2.64658928e-01
-6.03360713e-01 1.90963715e-01 -7.25840569e-01 -6.25451729e-02
1.04407620e+00 -9.90299061e-02 -8.19970250e-01 7.11048022e-02
-6.95955098e-01 2.97980368e-01 2.79499590e-01 4.40167874e-01
6.38522208e-01 -9.06409204e-01 -3.96038324e-01 3.68522555e-01
-3.42740536e-01 -3.49346966e-01 3.99996415e-02 7.53310382e-01
-3.97515446e-01 2.72338390e-01 -1.01504743e+00 -5.83714187e-01
-1.03608978e+00 4.46059495e-01 6.22562826e-01 4.87832911e-02
-2.17753112e-01 1.07363963e+00 -1.46210268e-01 -4.72440481e-01
2.27998704e-01 -2.31140733e-01 -8.31616018e-03 1.77451689e-02
3.70325446e-01 3.32943797e-01 1.41393989e-01 -5.12353480e-02
-3.27262610e-01 4.56017733e-01 -3.44011225e-02 -2.57883042e-01
1.74057198e+00 1.03882246e-01 -1.33748695e-01 1.97096124e-01
1.23268855e+00 -4.61518407e-01 -1.42120659e+00 2.96862423e-01
-4.69658431e-03 3.72004539e-01 1.59513205e-01 -1.27857387e+00
-1.28786826e+00 1.00095797e+00 1.00451505e+00 -1.41865775e-01
1.18745506e+00 -4.48467471e-02 5.11728168e-01 1.02963519e+00
5.20208478e-01 -1.02585304e+00 3.53189290e-01 5.97105622e-01
8.24934721e-01 -1.00804830e+00 -2.65767351e-02 3.05902869e-01
-2.71848470e-01 1.23757601e+00 9.28314090e-01 -2.03833550e-01
6.73648238e-01 6.65504217e-01 -3.79609615e-01 -5.53457260e-01
-1.22110367e+00 1.84526727e-01 1.00607291e-01 8.82478416e-01
2.64971346e-01 -5.00554264e-01 1.84157431e-01 5.40693581e-01
-2.57101119e-01 1.90170463e-02 3.45741361e-01 8.36020887e-01
-6.15575850e-01 -9.46223736e-01 -4.36108634e-02 6.58840120e-01
-3.53822708e-01 -2.12573186e-01 -6.36050344e-01 5.87597430e-01
1.45261779e-01 6.35862231e-01 5.17886989e-02 -7.00974524e-01
8.86179879e-02 1.29162431e-01 7.27898955e-01 -4.87878203e-01
-7.87070751e-01 -2.07125485e-01 -9.77506414e-02 -8.50426018e-01
4.90997992e-02 -6.55650377e-01 -1.54763019e+00 -1.09648891e-01
-1.40875757e-01 -2.22021207e-01 1.09804285e+00 8.14264774e-01
9.00158048e-01 8.06115150e-01 2.64114618e-01 -1.21092379e+00
-2.39868224e-01 -7.37993836e-01 -1.90821931e-01 -2.00566128e-01
3.91798355e-02 -5.51331758e-01 -1.70647398e-01 5.42495586e-02] | [9.558963775634766, 2.4881625175476074] |
4cf8cbd5-e55f-4b25-a031-aba096de3b21 | towards-real-time-single-channel-speech | 2303.07569 | null | https://arxiv.org/abs/2303.07569v2 | https://arxiv.org/pdf/2303.07569v2.pdf | Towards Real-Time Single-Channel Speech Separation in Noisy and Reverberant Environments | Real-time single-channel speech separation aims to unmix an audio stream captured from a single microphone that contains multiple people talking at once, environmental noise, and reverberation into multiple de-reverberated and noise-free speech tracks, each track containing only one talker. While large state-of-the-art DNNs can achieve excellent separation from anechoic mixtures of speech, the main challenge is to create compact and causal models that can separate reverberant mixtures at inference time. In this paper, we explore low-complexity, resource-efficient, causal DNN architectures for real-time separation of two or more simultaneous speakers. A cascade of three neural network modules are trained to sequentially perform noise-suppression, separation, and de-reverberation. For comparison, a larger end-to-end model is trained to output two anechoic speech signals directly from noisy reverberant speech mixtures. We propose an efficient single-decoder architecture with subtractive separation for real-time recursive speech separation for two or more speakers. Evaluation on real monophonic recordings of speech mixtures, according to speech separation measures like SI-SDR, perceptual measures like DNS-MOS, and a novel proposed channel separation metric, show that these compact causal models can separate speech mixtures with low latency, and perform on par with large offline state-of-the-art models like SepFormer. | ['Sebastian Braun', 'Julian Neri'] | 2023-03-14 | null | null | null | null | ['speech-separation'] | ['speech'] | [-1.77439023e-03 -3.78907889e-01 4.80911642e-01 4.10005152e-02
-1.21190703e+00 -6.25802219e-01 3.43840361e-01 -2.72353917e-01
-1.77178174e-01 5.07702768e-01 6.86439514e-01 -5.34056723e-01
5.96067868e-02 7.40959942e-02 -3.99029374e-01 -7.62304008e-01
-2.04559863e-01 2.39781007e-01 2.78210100e-02 3.16199102e-02
-6.71606362e-01 4.00703520e-01 -1.59876120e+00 3.17566246e-01
7.32947052e-01 1.00518000e+00 3.78329605e-01 1.61639619e+00
3.02402407e-01 6.20437622e-01 -1.13958251e+00 1.44635355e-02
1.37161985e-01 -7.82073498e-01 1.69867739e-01 -1.52171001e-01
3.79727244e-01 -3.51281494e-01 -6.89999044e-01 8.42440188e-01
1.24249876e+00 1.44732982e-01 6.02176368e-01 -6.43847585e-01
-1.22158103e-01 1.06036055e+00 -8.01077560e-02 4.38612849e-01
3.49957913e-01 1.29805923e-01 6.76490664e-01 -9.85085011e-01
-4.20789570e-01 1.38663733e+00 6.68192685e-01 5.49954057e-01
-1.28980982e+00 -9.89661038e-01 1.49991617e-01 -9.92535874e-02
-1.21785975e+00 -1.36272120e+00 6.39744222e-01 -1.07627869e-01
1.31800139e+00 6.89015746e-01 4.44930613e-01 1.57430995e+00
-1.76860064e-01 7.06868529e-01 7.37115204e-01 -3.89338881e-01
2.41042286e-01 -2.11616442e-01 1.46945015e-01 -1.69409260e-01
-1.07398786e-01 4.03297901e-01 -8.11922371e-01 -2.85546243e-01
4.70197439e-01 -2.28676304e-01 -7.41214395e-01 4.30431098e-01
-9.69168603e-01 1.51510239e-01 -8.46055895e-02 4.21799153e-01
-4.53906924e-01 2.92977184e-01 1.70409098e-01 4.18664306e-01
6.03294969e-01 1.62530869e-01 -4.46590990e-01 -4.24501061e-01
-1.27498603e+00 1.98531058e-02 1.38641834e+00 7.20322728e-01
-4.59732711e-02 9.72241521e-01 8.65432546e-02 1.26999080e+00
4.72022444e-01 1.18877077e+00 4.85567570e-01 -6.87194407e-01
4.99288619e-01 -8.17322195e-01 1.38399042e-02 -3.88718277e-01
-3.80704373e-01 -1.27595818e+00 -1.23412549e+00 3.07332836e-02
3.59112471e-02 -6.55439377e-01 -1.12441278e+00 1.76153135e+00
1.06441036e-01 7.61225760e-01 5.39279997e-01 9.69011486e-01
7.92674184e-01 9.52124834e-01 -4.97871548e-01 -6.00551009e-01
1.18616641e+00 -1.07154143e+00 -1.06813765e+00 -6.61452413e-01
-3.35662991e-01 -1.12592411e+00 3.32690567e-01 8.88200402e-01
-1.44079292e+00 -7.69190013e-01 -1.22525728e+00 3.21549028e-01
1.00594901e-01 3.87120247e-02 -2.26313233e-01 1.16625607e+00
-1.00421488e+00 1.99559093e-01 -8.93301904e-01 2.69128710e-01
-2.57511348e-01 1.40122488e-01 -9.40797925e-02 8.16712752e-02
-1.24634862e+00 3.72623712e-01 -1.67937756e-01 4.33637977e-01
-1.47202194e+00 -7.96682000e-01 -7.25347161e-01 3.71834278e-01
1.03318132e-01 -7.92282164e-01 1.57543910e+00 -8.79787385e-01
-1.80329216e+00 5.55658108e-03 -5.19411683e-01 -6.52128816e-01
2.30272591e-01 -6.92162812e-01 -1.40376616e+00 -3.18021663e-02
-2.85409957e-01 -1.80549100e-01 1.38313687e+00 -1.43977189e+00
-6.23435795e-01 -8.93600509e-02 -6.34761333e-01 4.63846117e-01
-3.47773850e-01 2.26364106e-01 -4.51574206e-01 -9.69047010e-01
1.80319443e-01 -6.80434704e-01 -3.95922482e-01 -8.76908898e-01
-5.79274833e-01 5.52281976e-01 5.69501460e-01 -1.22229719e+00
1.33036244e+00 -2.57490206e+00 1.10186227e-01 7.09871873e-02
2.00250059e-01 6.81739867e-01 -2.36958236e-01 1.87799379e-01
-3.76162112e-01 -1.86387941e-01 -9.66517180e-02 -1.08412790e+00
1.09852878e-02 -1.85883000e-01 -5.56883931e-01 4.04809475e-01
-1.22389048e-01 1.84385091e-01 -7.14749396e-01 1.60202056e-01
2.05052763e-01 1.02681386e+00 -3.56505334e-01 5.61215341e-01
3.35793078e-01 4.58793461e-01 4.78871375e-01 1.16626471e-01
7.59701431e-01 4.72266972e-01 1.75531849e-01 -1.06856087e-03
3.07793524e-02 1.04757273e+00 -1.57894337e+00 1.59629059e+00
-8.82773221e-01 8.44643176e-01 1.13215041e+00 -4.96957183e-01
6.82855248e-01 1.11093879e+00 -3.13829482e-02 -6.26980722e-01
5.22713736e-02 6.76301479e-01 1.75023675e-01 -1.41427949e-01
2.13879257e-01 -5.27063370e-01 2.09872901e-01 2.92827964e-01
3.20650667e-01 -7.57741630e-02 -1.59761891e-01 -1.20741412e-01
1.40036809e+00 -6.62383437e-01 1.23898638e-02 1.48905024e-01
3.09157878e-01 -1.05509496e+00 4.70842183e-01 8.44127417e-01
1.72919594e-02 9.32552040e-01 -1.80108890e-01 4.86077726e-01
-6.21189535e-01 -1.75299191e+00 3.28789473e-01 1.17694056e+00
-2.49032229e-01 -3.15974504e-01 -7.77338624e-01 1.47165269e-01
-2.87902981e-01 1.07706547e+00 1.43581241e-01 -3.16509753e-02
-8.43689680e-01 -5.56290567e-01 1.10708451e+00 3.60839307e-01
5.22599295e-02 -5.40423632e-01 6.49947301e-02 6.93941772e-01
-4.40588593e-01 -1.14605796e+00 -6.32979810e-01 8.62066507e-01
-4.01520699e-01 -1.61415130e-01 -1.01579058e+00 -6.66016936e-01
1.97086856e-02 6.12162530e-01 1.07678139e+00 -7.38413572e-01
3.07719558e-02 3.03346008e-01 -8.49975869e-02 -5.61191797e-01
-8.35937738e-01 -2.96852201e-01 7.15578914e-01 1.93961740e-01
-9.24563874e-03 -1.23599982e+00 -6.08324468e-01 4.77042794e-01
-7.27567613e-01 -1.88651860e-01 5.28125942e-01 4.68530029e-01
2.44244207e-02 6.29393876e-01 5.93651950e-01 1.15940407e-01
9.77324486e-01 -3.65346640e-01 -2.42647424e-01 -7.86639899e-02
4.39458825e-02 -3.37515831e-01 9.40370083e-01 -7.76155293e-01
-1.46292281e+00 -2.97127306e-01 -6.81917965e-01 -7.63024867e-01
-3.26770008e-01 3.27990919e-01 -5.45352161e-01 7.45704114e-01
8.25295031e-01 4.60030824e-01 -3.38239104e-01 -8.35005283e-01
5.30259728e-01 1.21558559e+00 1.01810086e+00 -1.00372486e-01
7.62124658e-01 1.24409907e-01 -5.37325919e-01 -1.14565074e+00
-2.77107209e-01 -9.53815699e-01 -9.74347517e-02 1.47052407e-01
5.51349938e-01 -1.45660532e+00 -4.86077905e-01 8.26073766e-01
-1.36702490e+00 -3.60099614e-01 -6.46090284e-02 9.92643416e-01
-1.88398525e-01 1.14184834e-01 -8.74151766e-01 -1.44217336e+00
-3.81180674e-01 -9.24778581e-01 6.57128632e-01 -3.32638472e-02
-2.92826563e-01 -7.12196052e-01 2.46286914e-01 3.21909070e-01
6.90943122e-01 -6.49109423e-01 1.80094361e-01 -8.44716370e-01
-1.26018554e-01 -1.50347039e-01 4.49944884e-01 1.01759851e+00
4.64054823e-01 -4.40562248e-01 -1.58530259e+00 -2.75392085e-01
6.32583976e-01 2.34089106e-01 9.39930379e-01 8.29379916e-01
2.61363983e-01 -5.21618843e-01 -3.14200699e-01 5.46807528e-01
7.81588078e-01 6.35083079e-01 4.96646196e-01 -6.77354574e-01
7.31139779e-01 3.38625371e-01 -1.62936345e-01 2.98464924e-01
-1.98748723e-01 3.73402476e-01 6.53230911e-03 -3.60778093e-01
-8.09107184e-01 4.75913621e-02 8.04195642e-01 1.68337548e+00
2.21941531e-01 -9.01830196e-01 -7.43346632e-01 5.65537930e-01
-1.39069104e+00 -1.06667924e+00 -2.04761952e-01 2.36558795e+00
7.94333279e-01 2.21942618e-01 1.00888662e-01 4.76998806e-01
6.59030795e-01 2.66634256e-01 -4.21290100e-01 -2.65837163e-01
-4.27714735e-01 4.15901154e-01 5.02113283e-01 9.89164650e-01
-9.37365234e-01 3.36973220e-01 5.97361660e+00 1.11244500e+00
-1.34676027e+00 4.00286615e-01 5.23374081e-01 -7.26872087e-01
-1.41511053e-01 -5.98519146e-01 -5.79256594e-01 1.94589496e-01
1.72954404e+00 1.81900814e-01 7.67739475e-01 3.50249261e-01
4.76654589e-01 2.68834978e-01 -1.40447474e+00 1.33564162e+00
1.64889947e-01 -8.24204147e-01 -4.20653284e-01 -1.91729933e-01
4.15353954e-01 3.34899068e-01 3.79368097e-01 1.53653532e-01
4.39146221e-01 -1.19598818e+00 1.09902155e+00 2.98296064e-01
6.91254973e-01 -6.53121889e-01 2.69014567e-01 4.34765667e-01
-1.39627945e+00 -3.44749629e-01 1.22922324e-01 -1.18197277e-01
4.62829739e-01 1.18553317e+00 -9.77982044e-01 6.54093921e-01
5.57019234e-01 1.11073919e-01 2.34544039e-01 1.15479553e+00
-3.17712635e-01 1.24259007e+00 -5.32111704e-01 3.29666853e-01
-2.22898293e-02 2.06624031e-01 1.07857203e+00 1.79283714e+00
5.52518547e-01 1.71135947e-01 -2.21314713e-01 5.19433618e-01
4.48478945e-02 -4.67109174e-01 -2.95784444e-01 1.45663157e-01
6.03111446e-01 8.54590118e-01 -2.99416751e-01 -3.67598832e-01
-1.34893104e-01 1.08824170e+00 -4.13483262e-01 9.92377937e-01
-9.89200652e-01 -4.37540501e-01 1.09720099e+00 -1.49138778e-01
5.13461828e-01 -4.38465387e-01 -1.73470005e-01 -1.02266383e+00
2.21190006e-02 -1.32028830e+00 -2.73859531e-01 -6.83864772e-01
-1.31414640e+00 1.02213776e+00 -5.15573323e-01 -1.22458327e+00
-3.99625897e-01 -4.26232457e-01 -5.50931573e-01 1.47984111e+00
-1.03896821e+00 -5.96123099e-01 3.51612210e-01 5.76232612e-01
8.56422246e-01 -1.74687356e-01 9.44049418e-01 8.36305261e-01
-6.06285036e-01 4.82602924e-01 5.24761558e-01 6.47457764e-02
7.24123061e-01 -1.21294665e+00 8.22671175e-01 1.29658949e+00
2.69723415e-01 7.14332998e-01 1.23504972e+00 -4.39095587e-01
-1.10464370e+00 -9.78036284e-01 6.90173268e-01 -6.55251592e-02
4.93994981e-01 -9.68629658e-01 -7.79700935e-01 2.63297737e-01
4.65697378e-01 -1.65635437e-01 9.08441424e-01 1.50944173e-01
-4.38793272e-01 -4.66135889e-01 -3.73076588e-01 6.37164950e-01
8.48470926e-01 -7.38442242e-01 -7.41297007e-01 -2.16809642e-02
9.96640921e-01 -2.77217776e-01 -1.25633061e-01 -6.20332407e-03
5.55936456e-01 -1.12531459e+00 1.31464052e+00 -3.62033881e-02
-1.34518910e-02 -2.94686109e-01 -3.34711522e-01 -1.93729258e+00
-2.95435518e-01 -1.44254434e+00 -3.26498359e-01 1.36524558e+00
7.56236732e-01 -7.88794219e-01 2.38912717e-01 2.48151422e-01
-6.07212603e-01 -7.11058546e-03 -1.11654520e+00 -1.20813286e+00
-1.74911574e-01 -9.85899389e-01 2.52446383e-01 3.80262673e-01
-1.68061510e-01 8.22540700e-01 -6.09800279e-01 8.56655002e-01
7.99990594e-01 -4.95685905e-01 5.82038760e-01 -8.84352028e-01
-8.78040493e-01 -5.99671721e-01 1.28875285e-01 -1.77618003e+00
-4.99049574e-02 -3.23619545e-01 6.32224917e-01 -1.56364393e+00
-6.03443027e-01 -1.53510496e-01 -4.36311752e-01 -2.57302910e-01
-3.43994498e-01 -2.34157041e-01 1.86849028e-01 -9.41438973e-02
-2.67541289e-01 7.08189487e-01 5.22232711e-01 -2.14755610e-01
-4.47171152e-01 4.91586328e-01 -5.90221703e-01 8.44577134e-01
3.85431498e-01 -5.88071585e-01 -6.11804128e-01 -4.73380744e-01
-2.33114243e-01 6.23848438e-01 2.17100933e-01 -1.58323610e+00
5.49357474e-01 4.57149982e-01 1.83522373e-01 -5.64879477e-01
1.01195216e+00 -9.04251814e-01 5.06961048e-01 3.01383734e-01
-2.24957898e-01 -5.98371983e-01 4.07764643e-01 6.70031548e-01
-3.88150960e-01 2.87445545e-01 5.39316475e-01 7.72309899e-02
1.91857889e-01 -1.72085121e-01 -8.59642565e-01 -7.02322051e-02
2.51310140e-01 9.75675955e-02 -1.36045441e-01 -8.26851547e-01
-9.27307129e-01 -2.44364277e-01 -5.37276804e-01 3.09775889e-01
6.05683267e-01 -1.08166754e+00 -1.12922335e+00 2.46793225e-01
-6.78920388e-01 -7.71062002e-02 5.66514194e-01 6.91702902e-01
-5.05278856e-02 7.26140201e-01 3.40435892e-01 -5.24776101e-01
-1.49304497e+00 5.69213867e-01 7.59063184e-01 -1.82212502e-01
-2.93646008e-01 1.38139164e+00 6.65458202e-01 -1.60682186e-01
6.37500226e-01 -4.76799518e-01 1.66132405e-01 1.20435758e-02
7.72324443e-01 7.03488111e-01 3.82929176e-01 -4.52129245e-01
-3.25876564e-01 7.25061074e-02 3.88150007e-01 -9.47730184e-01
1.00248802e+00 -5.03612459e-01 2.83401944e-02 8.90415668e-01
1.24128389e+00 6.77865624e-01 -1.14475667e+00 -2.73253411e-01
-6.07805252e-01 -9.62054804e-02 1.64984405e-01 -1.10386956e+00
-8.41779411e-01 1.07497489e+00 7.66911268e-01 6.52015805e-01
1.52947748e+00 -2.90809542e-01 9.63416219e-01 1.83011979e-01
-1.37488237e-02 -6.25912964e-01 1.08575366e-01 5.38748980e-01
9.57453966e-01 -5.29412806e-01 -5.99839628e-01 -1.90726236e-01
-1.91234544e-01 8.70808065e-01 1.14812814e-01 2.69542098e-01
9.59427774e-01 9.96730208e-01 5.45487940e-01 2.94924051e-01
-8.30399156e-01 -1.27925977e-01 1.74247608e-01 6.00193679e-01
3.90887231e-01 3.35955143e-01 6.04152918e-01 1.00313222e+00
-6.14611387e-01 -6.67146742e-01 2.98583657e-01 3.94055575e-01
-5.18535495e-01 -8.79398227e-01 -1.19938374e+00 1.27172887e-01
-6.11262143e-01 -7.09996879e-01 -2.29708850e-01 -1.49217114e-01
-2.44759232e-01 1.88741839e+00 2.41691947e-01 -3.82355154e-01
4.82094049e-01 2.00822398e-01 9.18451250e-02 -4.30942327e-01
-8.23850095e-01 1.18389511e+00 4.96261656e-01 -1.25071600e-01
-3.02897170e-02 -5.15162706e-01 -8.58288646e-01 -1.65990010e-01
-3.59958053e-01 2.62863457e-01 8.16617370e-01 8.13337684e-01
2.94705629e-01 1.37789834e+00 6.84555948e-01 -1.12943268e+00
-4.69819367e-01 -1.28220868e+00 -8.04389775e-01 4.67931665e-02
1.21357965e+00 1.40958503e-01 -8.53722274e-01 2.03124598e-01] | [15.004122734069824, 5.869960308074951] |
169ed8d4-9ae8-4b04-aa98-d18f3ff6e909 | language-identification-and-named-entity | null | null | https://aclanthology.org/P18-3008 | https://aclanthology.org/P18-3008.pdf | Language Identification and Named Entity Recognition in Hinglish Code Mixed Tweets | While growing code-mixed content on Online Social Networks(OSN) provides a fertile ground for studying various aspects of code-mixing, the lack of automated text analysis tools render such studies challenging. To meet this challenge, a family of tools for analyzing code-mixed data such as language identifiers, parts-of-speech (POS) taggers, chunkers have been developed. Named Entity Recognition (NER) is an important text analysis task which is not only informative by itself, but is also needed for downstream NLP tasks such as semantic role labeling. In this work, we present an exploration of automatic NER of code-mixed data. We compare our method with existing off-the-shelf NER tools for social media content,and find that our systems outperforms the best baseline by 33.18 {\%} (F1 score). | ['Ponnurangam Kumaraguru', 'Kushagra Singh', 'Indira Sen'] | 2018-07-01 | null | null | null | acl-2018-7 | ['abuse-detection'] | ['natural-language-processing'] | [-1.40304714e-01 1.48638710e-01 -3.69633943e-01 -1.52623475e-01
-7.95884788e-01 -8.29579115e-01 4.11383033e-01 7.48293519e-01
-4.62284833e-01 5.95798731e-01 6.15270674e-01 -4.34386492e-01
7.28652775e-02 -4.90900010e-01 -3.04404467e-01 4.49247509e-02
-7.52035305e-02 2.11983442e-01 5.14167547e-01 -4.51176912e-01
5.39413631e-01 6.58377036e-02 -1.52565491e+00 5.03608584e-01
1.27848339e+00 3.67295682e-01 3.38471949e-01 6.22117817e-01
-1.13956940e+00 1.32871819e+00 -5.50011218e-01 -7.02925742e-01
-1.83329195e-01 -1.81431130e-01 -1.20165968e+00 -5.40458024e-01
1.48116022e-01 2.69624412e-01 -7.85080343e-03 1.47301090e+00
2.19333634e-01 -8.23877081e-02 2.94017643e-01 -1.36417341e+00
-2.78236032e-01 1.33820045e+00 -4.30540174e-01 4.41507787e-01
6.67297602e-01 -4.07830358e-01 1.20509255e+00 -5.98903120e-01
1.03542089e+00 7.49136925e-01 9.98923123e-01 6.18730307e-01
-9.62988675e-01 -5.67508996e-01 -4.54811305e-01 -2.33596176e-01
-1.06012177e+00 -5.05681515e-01 4.80220139e-01 -9.03461099e-01
1.21551049e+00 3.20066065e-02 2.21573740e-01 8.78466606e-01
1.68522432e-01 6.27618134e-01 6.97725594e-01 -7.03769982e-01
8.74736160e-02 2.00254247e-01 4.12336737e-01 7.79381752e-01
2.14260608e-01 -8.37598503e-01 -5.20283639e-01 -5.92849314e-01
2.99487919e-01 1.76251028e-02 -1.64061394e-02 1.14889905e-01
-1.14868498e+00 6.53351068e-01 -9.60427672e-02 8.35026860e-01
-1.79005697e-01 2.30802730e-01 7.40111649e-01 2.53884435e-01
6.69826627e-01 7.89059818e-01 -7.34600961e-01 -1.00168431e+00
-9.78628218e-01 -1.19105816e-01 1.35726106e+00 1.20988369e+00
7.66105771e-01 -4.33345377e-01 6.66931793e-02 1.25994503e+00
5.00352979e-01 5.70200272e-02 6.47615850e-01 -9.26908612e-01
9.96588886e-01 1.08531702e+00 -9.38125849e-02 -6.61503315e-01
-3.79706502e-01 -1.15274720e-01 -3.61519933e-01 -2.54337639e-01
6.08711958e-01 -3.11994821e-01 -3.12279761e-01 1.29403794e+00
2.06886306e-01 -2.92653352e-01 -2.25346640e-01 7.14216754e-02
8.78702283e-01 4.16387469e-01 3.84493649e-01 1.77286059e-01
1.49178922e+00 -1.02470267e+00 -5.79901278e-01 -3.24190021e-01
1.02673781e+00 -1.07894528e+00 8.01191628e-01 -1.28955215e-01
-7.97617853e-01 -4.47886661e-02 -5.23606777e-01 -1.99189544e-01
-5.36972225e-01 -7.70439431e-02 6.84351981e-01 9.02548611e-01
-1.08834219e+00 6.75966084e-01 -8.70849013e-01 -7.44178653e-01
4.89997685e-01 1.46149427e-01 -6.09707117e-01 9.72881392e-02
-8.74733567e-01 5.54200530e-01 2.36631647e-01 -5.55410981e-01
-5.22854567e-01 -7.50661373e-01 -1.00506604e+00 1.64029971e-01
4.79383588e-01 -1.57475561e-01 1.42304993e+00 -9.05689061e-01
-1.27438903e+00 9.66972113e-01 -1.68547720e-01 -2.74380684e-01
3.15580308e-01 -3.29590589e-01 -6.57991648e-01 6.93081394e-02
6.36206150e-01 2.63250291e-01 4.37931180e-01 -8.41288447e-01
-7.02396214e-01 -5.21316864e-02 1.41901091e-01 -2.59298176e-01
-5.86313367e-01 1.13116467e+00 -1.77911237e-01 -6.61492288e-01
-3.12026143e-01 -8.76066864e-01 -3.57404083e-01 -2.87117600e-01
-4.17598516e-01 -4.29833174e-01 3.74956638e-01 -1.00718915e+00
1.78353798e+00 -2.27159500e+00 -2.55152047e-01 -2.75312159e-02
4.67536092e-01 1.66438833e-01 -9.13374722e-02 9.29691792e-01
-2.16737807e-01 8.03039014e-01 -4.59577590e-01 -5.70633471e-01
6.00351915e-02 1.94328167e-02 -1.66456997e-01 3.12303722e-01
6.39268905e-02 7.49268651e-01 -1.26910245e+00 -8.12076330e-01
-3.26656461e-01 1.28394842e-01 -7.49290764e-01 1.64128706e-01
-3.20376605e-01 8.73242840e-02 -4.85714734e-01 7.03367412e-01
2.82880872e-01 -2.96461105e-01 1.78926617e-01 6.87534332e-01
-6.53796792e-01 7.15053737e-01 -9.65489447e-01 1.87539482e+00
-5.64793348e-01 8.73410046e-01 2.56780088e-01 -5.48115551e-01
6.14043057e-01 3.80872399e-01 6.55874908e-01 -7.32563809e-02
7.46514648e-02 5.18050194e-01 -1.12110876e-01 -9.28587794e-01
8.91406298e-01 1.56849459e-01 -5.21862209e-01 5.65338254e-01
2.17892185e-01 3.00513506e-01 5.34174144e-01 4.86163676e-01
1.86585808e+00 1.39945373e-02 5.71637094e-01 -5.86812854e-01
6.06142640e-01 3.05162072e-01 6.66838348e-01 5.25410414e-01
-2.59174168e-01 4.81029749e-01 1.01948869e+00 -1.08495317e-02
-1.13069701e+00 -4.11280841e-01 -1.23749293e-01 1.42045808e+00
-1.79809406e-01 -9.86798108e-01 -1.03264523e+00 -1.12065506e+00
-1.63758382e-01 4.87467080e-01 -2.98500538e-01 4.68022585e-01
-6.83270037e-01 -4.34779018e-01 1.02791631e+00 4.30495590e-01
3.84682059e-01 -1.03336763e+00 -3.18082497e-02 4.72605646e-01
-4.73799288e-01 -1.17885911e+00 -5.32778025e-01 1.49455652e-01
-6.78156376e-01 -1.20715570e+00 -3.52433503e-01 -8.44537139e-01
5.78736722e-01 1.59920275e-01 1.40864873e+00 5.53821385e-01
3.45684811e-02 2.30514213e-01 -7.95721948e-01 -2.85989225e-01
-9.84485090e-01 6.46558344e-01 -3.74059528e-01 -4.49791163e-01
4.74997371e-01 -6.71680629e-01 -1.28062934e-01 3.05030942e-01
-8.71270597e-01 -2.55166292e-01 4.92989302e-01 4.51446801e-01
-8.92755911e-02 -9.26852450e-02 6.15811408e-01 -1.42568171e+00
6.30260050e-01 -1.10855353e+00 -4.27521765e-01 2.20499024e-01
-3.80182892e-01 8.70443732e-02 8.24353039e-01 -1.65798292e-01
-1.17908227e+00 -8.16840827e-02 -4.72439468e-01 3.64770949e-01
-3.17993015e-01 7.82547057e-01 3.08663193e-02 -6.97569028e-02
8.81970346e-01 -9.50583518e-02 -1.55738637e-01 -8.93644094e-01
6.01252951e-02 1.19631422e+00 2.06823602e-01 -6.25604749e-01
6.54263675e-01 2.64678508e-01 -5.33207834e-01 -1.02679551e+00
-7.23161399e-01 -1.18827224e+00 -6.54728472e-01 2.76793037e-02
1.00223196e+00 -9.13434327e-01 -2.35129818e-01 3.53109032e-01
-1.25718069e+00 -3.81256282e-01 -2.11109780e-02 2.51850873e-01
-9.64200646e-02 7.28330553e-01 -9.79889989e-01 -5.39433062e-01
-2.17159331e-01 -9.13358986e-01 7.75641263e-01 1.46473348e-01
-5.89099467e-01 -9.79975402e-01 4.41415608e-01 8.66238713e-01
4.76156950e-01 -9.18845013e-02 9.32883203e-01 -1.17844641e+00
-2.56433278e-01 -1.23916171e-01 -1.66555464e-01 1.68334961e-01
-2.69845396e-01 3.37462217e-01 -7.81545699e-01 1.56172484e-01
-3.57071429e-01 2.57724468e-02 5.56801438e-01 -1.87159538e-01
1.01988149e+00 -2.59147495e-01 -4.35192794e-01 2.62798876e-01
1.39223468e+00 -1.12327849e-02 7.33600616e-01 5.84698915e-01
7.67889619e-01 5.91347516e-01 2.89841890e-01 6.25102997e-01
6.99260414e-01 2.03303128e-01 3.40732962e-01 4.68443900e-01
-1.23646930e-01 -2.98674405e-01 5.66164315e-01 1.24483430e+00
-8.62804800e-02 -2.46109098e-01 -1.59949565e+00 8.03682029e-01
-1.81648028e+00 -1.03585935e+00 -8.60901117e-01 1.70128500e+00
1.11596346e+00 -1.94427036e-02 9.49888825e-02 -1.42431453e-01
1.05730128e+00 4.75090221e-02 1.60938978e-01 -2.85938919e-01
2.24453747e-01 1.26857400e-01 4.78107929e-01 1.62642300e-01
-1.15439105e+00 7.55923986e-01 5.78749895e+00 9.49950635e-01
-5.36010206e-01 3.97706151e-01 -3.11906096e-02 6.42323434e-01
-3.54921788e-01 4.91497904e-01 -9.15972054e-01 7.75996089e-01
1.18070817e+00 -2.17365369e-01 6.08270466e-01 1.23552787e+00
-1.21499605e-01 -3.61920059e-01 -1.06484020e+00 7.64059603e-01
-4.15654145e-02 -1.37675774e+00 -6.04665101e-01 1.38416439e-01
9.18883741e-01 6.98674679e-01 -6.78589642e-01 5.81175864e-01
6.99570477e-01 -6.71696782e-01 8.08906317e-01 1.57414407e-01
5.77729166e-01 -3.15199524e-01 8.84869695e-01 5.42239189e-01
-1.42631376e+00 -1.91981673e-01 -1.71266012e-02 -1.14331879e-01
-4.30786423e-02 9.05869305e-01 -7.52267420e-01 3.55049372e-01
6.62675500e-01 8.49227726e-01 -8.25654566e-01 1.51974475e+00
-4.61566061e-01 7.37204134e-01 -1.46662667e-01 -2.24072129e-01
-4.57455888e-02 1.70789719e-01 6.30172551e-01 1.62998521e+00
2.61973053e-01 -1.60175562e-01 1.42638341e-01 8.08704495e-01
-6.80870295e-01 4.17005628e-01 -4.93044317e-01 -5.36366999e-01
6.76152110e-01 1.56137490e+00 -1.17248368e+00 -2.31319696e-01
-6.71080828e-01 7.63068557e-01 5.05482614e-01 -1.30356923e-01
-6.49166822e-01 -8.96735668e-01 6.85398638e-01 2.47319177e-01
1.06005080e-03 -4.31619614e-01 -5.90098202e-02 -1.56442308e+00
2.44178958e-02 -7.16811240e-01 5.25671244e-01 -5.31079710e-01
-1.41431320e+00 4.82930511e-01 -1.48751304e-01 -1.21932971e+00
-1.47678211e-01 -3.74881327e-01 -6.63007975e-01 4.69605118e-01
-1.42610741e+00 -7.72250354e-01 -2.47327596e-01 2.48196959e-01
5.43648601e-01 -7.16895610e-02 6.30750537e-01 8.23431194e-01
-6.09865963e-01 3.08548778e-01 3.71210665e-01 8.65734100e-01
7.43687034e-01 -1.34176791e+00 5.79899907e-01 1.00184298e+00
1.18231714e-01 9.10937011e-01 5.04630566e-01 -8.87704432e-01
-1.16266429e+00 -9.23579514e-01 1.47983801e+00 -8.19827616e-01
1.32014298e+00 -6.71200156e-01 -8.39126825e-01 5.72102487e-01
3.20387691e-01 5.65334335e-02 1.09111571e+00 2.74838418e-01
-5.24735451e-01 4.27981108e-01 -1.09000540e+00 3.60676497e-01
1.20995617e+00 -9.73101437e-01 -7.75486469e-01 1.03999190e-01
6.64016724e-01 -2.09304750e-01 -1.21878052e+00 -2.96071947e-01
1.75248489e-01 -9.42157030e-01 4.91987079e-01 -4.52541769e-01
9.07274544e-01 -2.29556546e-01 7.03015104e-02 -1.06784070e+00
3.00096888e-02 -8.97713184e-01 2.91390449e-01 1.89137900e+00
7.65919089e-01 -5.68558872e-01 5.24364650e-01 9.52667475e-01
-5.50184369e-01 -1.37012348e-01 -7.66435325e-01 -6.71902359e-01
-5.86034469e-02 -6.65291727e-01 4.04775083e-01 1.38493955e+00
7.59216368e-01 1.63688987e-01 1.91099465e-01 -1.25130102e-01
3.59780610e-01 -2.09768146e-01 5.25985003e-01 -1.51002371e+00
-1.65173486e-01 -5.28649390e-01 -3.38468105e-01 -6.36886358e-01
5.79373658e-01 -1.18433797e+00 5.18163256e-02 -1.66429973e+00
5.97310781e-01 -6.83585107e-01 2.63139725e-01 8.15218210e-01
-4.45176177e-02 -1.65026546e-01 7.56452158e-02 4.17357683e-01
-1.03862357e+00 9.03077349e-02 3.95217538e-01 1.93012103e-01
-1.10936172e-01 -1.75450593e-01 -6.45371854e-01 9.64397430e-01
7.56966054e-01 -1.15662801e+00 1.41981706e-01 -3.33205283e-01
1.06806779e+00 7.57403448e-02 -1.33460447e-01 -9.68810320e-01
4.55959380e-01 -6.86089694e-03 -4.31044370e-01 -1.60703197e-01
-4.78499919e-01 -6.83224082e-01 -3.03818434e-02 2.78835714e-01
-3.33395094e-01 7.88468346e-02 -2.66107768e-02 5.02042651e-01
-3.25734377e-01 -1.08419144e+00 4.33323652e-01 -3.71605814e-01
-7.55523860e-01 -1.74533445e-02 -7.82433569e-01 8.13874364e-01
8.15861940e-01 -2.49320474e-02 -8.05825233e-01 6.11197874e-02
-3.18688452e-01 2.12022126e-01 8.42293978e-01 5.70040226e-01
-5.48629165e-02 -8.70645463e-01 -4.71047431e-01 -2.26447776e-01
3.96849781e-01 -1.88934594e-01 -3.47312493e-03 8.79476368e-01
-7.85745382e-01 1.11189615e-02 2.24384926e-02 -1.40799945e-02
-1.35645783e+00 3.16728503e-01 -2.84508877e-02 -4.14687306e-01
-4.16265488e-01 7.65644431e-01 -5.44476628e-01 -6.79612875e-01
-1.41796708e-01 -3.13306272e-01 -5.12988687e-01 4.01998907e-01
5.00167489e-01 5.92336893e-01 1.16482429e-01 -6.32665873e-01
-4.59023178e-01 1.55179381e-01 2.74307966e-01 1.60507306e-01
1.74419057e+00 -1.29472022e-03 -9.29289103e-01 4.23294336e-01
1.29605615e+00 8.07597458e-01 -6.13891423e-01 -5.93760572e-02
8.54742765e-01 -5.14208078e-01 -2.78908551e-01 -4.96510118e-01
-5.95884621e-01 5.33599436e-01 -3.14717084e-01 8.95295441e-01
4.06297386e-01 3.69455665e-01 7.29780138e-01 2.73978800e-01
6.48037672e-01 -1.22707236e+00 -1.47423953e-01 1.00375557e+00
3.87331277e-01 -1.22653806e+00 -3.22533667e-01 -6.69204116e-01
-3.56320232e-01 1.22167826e+00 3.80773574e-01 1.92318916e-01
8.08994770e-01 5.81569016e-01 -1.48139596e-01 -2.67210066e-01
-5.64853072e-01 -2.58530736e-01 -7.71280453e-02 2.37283871e-01
9.39134538e-01 -2.71207601e-01 -3.87867033e-01 8.61383259e-01
-1.37398183e-01 -1.02765344e-01 9.24328923e-01 1.17717099e+00
-5.07680774e-01 -1.50031912e+00 -2.81509370e-01 6.97265863e-01
-1.35713923e+00 -5.13481557e-01 -2.68383324e-01 5.43821216e-01
1.41872928e-01 1.03998935e+00 -1.03305705e-01 -2.28556588e-01
4.06963825e-02 2.80889153e-01 -5.47430143e-02 -1.32963014e+00
-1.33243954e+00 -4.73020941e-01 7.76866198e-01 -4.32308614e-01
-4.67647314e-01 -9.78802085e-01 -1.51800609e+00 -3.80422145e-01
-5.16843200e-01 4.26772267e-01 8.87833178e-01 8.86226535e-01
5.38485050e-01 3.17964971e-01 2.84504354e-01 -4.58059072e-01
-1.93555668e-01 -8.86796713e-01 -4.27922338e-01 4.15771365e-01
-4.49037999e-02 -3.21345121e-01 -6.00527942e-01 2.20039472e-01] | [9.691502571105957, 9.875917434692383] |
8bcc7e01-3c90-44da-9140-3c7e2842a84e | deep-learning-methods-for-small-molecule-drug | 2303.00313 | null | https://arxiv.org/abs/2303.00313v2 | https://arxiv.org/pdf/2303.00313v2.pdf | Deep Learning Methods for Small Molecule Drug Discovery: A Survey | With the development of computer-assisted techniques, research communities including biochemistry and deep learning have been devoted into the drug discovery field for over a decade. Various applications of deep learning have drawn great attention in drug discovery, such as molecule generation, molecular property prediction, retrosynthesis prediction, and reaction prediction. While most existing surveys only focus on one of the applications, limiting the view of researchers in the community. In this paper, we present a comprehensive review on the aforementioned four aspects, and discuss the relationships among different applications. The latest literature and classical benchmarks are presented for better understanding the development of variety of approaches. We commence by summarizing the molecule representation format in these works, followed by an introduction of recent proposed approaches for each of the four tasks. Furthermore, we review a variety of commonly used datasets and evaluation metrics and compare the performance of deep learning-based models. Finally, we conclude by identifying remaining challenges and discussing the future trend for deep learning methods in drug discovery. | ['Gaoang Wang', 'Hongwei Wang', 'Hangyue Chen', 'Wenhao Chai', 'Xuanyu Chen', 'Yingying Liu', 'Wenhao Hu'] | 2023-03-01 | null | null | null | null | ['drug-discovery', 'retrosynthesis', 'molecular-property-prediction'] | ['medical', 'medical', 'miscellaneous'] | [ 3.24197710e-01 -2.28521332e-01 -8.86551440e-01 -3.61069553e-02
-4.54642087e-01 -6.05024755e-01 4.72196311e-01 6.52282536e-01
-1.37493581e-01 1.17206192e+00 1.29833907e-01 -4.43071395e-01
-1.78434387e-01 -6.38920069e-01 -4.60050762e-01 -1.01242185e+00
-2.55739748e-01 2.33683243e-01 -1.69125527e-01 -1.78162992e-01
4.31793213e-01 8.83343697e-01 -9.21198368e-01 4.13894147e-01
6.41741514e-01 9.07769382e-01 -4.96309921e-02 2.38116205e-01
-1.04410209e-01 9.74924624e-01 -6.48891628e-01 -6.38394713e-01
-1.48527026e-01 -5.44254899e-01 -1.04792106e+00 -3.33765328e-01
-1.93053242e-02 -6.65801689e-02 -4.79471356e-01 7.73292065e-01
1.12997711e+00 1.41818240e-01 9.29810405e-01 -7.28545129e-01
-7.55301535e-01 5.45231998e-01 -2.39899248e-01 3.50053310e-01
5.09651363e-01 1.36894910e-02 9.08239365e-01 -1.02588570e+00
7.07441628e-01 8.96235585e-01 5.84123552e-01 6.60199761e-01
-9.66808021e-01 -7.62024999e-01 2.45044865e-02 6.02458298e-01
-1.22649097e+00 -4.64570880e-01 6.06381357e-01 -5.76965570e-01
1.44789600e+00 1.33456895e-02 4.65437710e-01 1.24992418e+00
6.14878953e-01 8.03266346e-01 7.62507379e-01 -1.18364751e-01
2.56410778e-01 -1.44590810e-01 1.53401131e-02 6.77343905e-01
1.22110493e-01 3.81535329e-02 -6.13350272e-01 -3.97068709e-01
5.21002829e-01 1.37515599e-02 -1.63898066e-01 -3.78484070e-01
-1.25799608e+00 1.05203032e+00 4.66280162e-01 3.43905985e-01
-4.75145996e-01 -3.13489027e-02 7.55398810e-01 3.85488421e-02
4.74074513e-01 8.16000044e-01 -7.74274349e-01 1.97320640e-01
-6.14821553e-01 4.92464006e-01 7.06034541e-01 6.68262303e-01
6.30107149e-02 2.58960575e-01 -1.35934740e-01 7.90603280e-01
4.72958498e-02 -3.09059292e-01 3.81001532e-01 -3.08268130e-01
3.26028951e-02 1.72944486e-01 1.05883507e-03 -7.55648077e-01
-7.39233494e-01 -5.12110293e-01 -9.81173754e-01 -1.84554338e-01
1.55373991e-01 -1.45583048e-01 -7.23942876e-01 1.14821494e+00
3.62079680e-01 2.02260509e-01 1.04299806e-01 4.80516732e-01
1.52684104e+00 7.35042512e-01 6.69505179e-01 -5.11059046e-01
1.17017078e+00 -1.08586049e+00 -8.48134518e-01 7.26724923e-01
7.28273988e-01 -9.68899667e-01 1.18584864e-01 5.50072551e-01
-1.01531994e+00 -3.56472254e-01 -9.79333222e-01 -2.86694914e-01
-9.23104227e-01 -1.72732517e-01 1.31505001e+00 7.59247184e-01
-4.51020032e-01 1.15046835e+00 -7.47634351e-01 -2.92922169e-01
7.44661272e-01 8.79188955e-01 -3.81851345e-01 1.85319692e-01
-1.34210610e+00 9.85148072e-01 2.84617990e-01 -8.42320472e-02
-1.29243660e+00 -1.00603878e+00 -5.39287806e-01 -1.52000248e-01
3.35239135e-02 -8.56256485e-01 1.21537173e+00 -3.27295512e-01
-1.70630229e+00 7.25709081e-01 8.59236158e-03 -5.35832584e-01
2.73856759e-01 9.26102549e-02 -4.84178483e-01 -1.36722445e-01
-1.64003640e-01 5.96327007e-01 1.28967479e-01 -7.12505341e-01
-5.73867798e-01 -3.35702956e-01 1.06642976e-01 1.54791623e-01
-1.23989493e-01 3.17245543e-01 3.21520977e-02 -6.91409171e-01
-5.05236924e-01 -5.08992553e-01 -5.68566382e-01 -2.99840510e-01
-6.23946011e-01 -6.85693562e-01 3.96116138e-01 -8.64401087e-02
1.29899836e+00 -1.51044905e+00 4.55312520e-01 -2.92763114e-01
4.38897461e-01 3.73210073e-01 -2.33974710e-01 1.16544616e+00
-7.60112643e-01 2.52530158e-01 4.78197820e-02 1.06074408e-01
-3.75221997e-01 -2.96510518e-01 -3.74955386e-01 8.10409188e-01
3.10944207e-02 1.01484513e+00 -1.10236931e+00 -1.79304685e-02
3.13696891e-01 6.77349210e-01 -2.72655696e-01 -9.93021578e-02
-4.87701386e-01 7.90705979e-01 -7.76973426e-01 1.04470789e+00
4.21966791e-01 -3.77060235e-01 3.56749415e-01 -3.64886820e-01
-3.85875225e-01 6.25196576e-01 -4.67566490e-01 1.63649976e+00
-7.63237402e-02 2.04408213e-01 -6.40577674e-01 -1.30692029e+00
7.61651278e-01 6.79692209e-01 9.85697925e-01 -5.43129265e-01
2.45912239e-01 4.61612135e-01 4.54306543e-01 -3.83079618e-01
3.46115604e-02 -3.25022966e-01 4.79116619e-01 1.50712550e-01
1.57879144e-01 3.89122516e-01 1.32105291e-01 -1.94292888e-01
8.52489114e-01 2.91710645e-01 9.73192453e-01 9.63067636e-04
8.90744925e-01 -8.75932500e-02 2.93460101e-01 2.49779955e-01
-3.18859726e-01 -1.73019208e-02 4.58227098e-01 -1.19009662e+00
-9.60054994e-01 -4.30703074e-01 -5.80062568e-01 1.13329983e+00
-3.87452878e-02 -6.79431736e-01 -5.74254751e-01 -6.87309682e-01
-1.38054848e-01 3.95567492e-02 -7.89522171e-01 -9.44424942e-02
-3.88432086e-01 -1.59626949e+00 6.61512434e-01 3.51093441e-01
3.70269775e-01 -1.06050980e+00 1.19551040e-01 5.75946927e-01
1.95042640e-01 -5.89654148e-01 1.00546941e-01 4.63838607e-01
-9.78795230e-01 -1.26059020e+00 -8.75270009e-01 -1.01896501e+00
-4.20001661e-03 1.73679158e-01 9.23023045e-01 -1.24001913e-01
-5.84238648e-01 -2.74533510e-01 -2.35219076e-01 -7.06337690e-01
-3.39406580e-01 2.61984348e-01 4.99863699e-02 -3.50875705e-01
5.85496068e-01 -5.11492670e-01 -9.03308153e-01 -7.69713670e-02
-5.28479159e-01 -2.23557144e-01 5.31034410e-01 8.23649704e-01
7.47448146e-01 -2.95371234e-01 1.08809900e+00 -9.80399966e-01
6.51286423e-01 -8.28660071e-01 -2.86204159e-01 1.08488932e-01
-6.24627292e-01 2.55839480e-03 6.39741659e-01 -3.63162428e-01
-5.37517190e-01 1.53709084e-01 -6.46244466e-01 2.84357607e-01
-7.62055516e-02 8.08915794e-01 -1.02656156e-01 -5.36740124e-01
6.89399242e-01 2.17953309e-01 -3.30309629e-01 -6.58618748e-01
3.14230412e-01 4.18781340e-01 -1.36630997e-01 -3.70462805e-01
1.35013819e-01 1.98382542e-01 4.02799428e-01 -8.51650357e-01
-7.87117481e-01 -3.97871137e-01 -3.77707779e-01 2.93019265e-01
8.13244522e-01 -6.94565773e-01 -1.16728354e+00 4.60536331e-01
-1.27313983e+00 -1.08050369e-01 1.96005866e-01 4.76116270e-01
-6.11039758e-01 5.83110392e-01 -9.65303779e-01 -3.10776144e-01
-7.63326883e-01 -1.55751991e+00 8.03414762e-01 -1.32004498e-02
-1.59375131e-01 -1.29316878e+00 5.50031066e-01 3.93500715e-01
1.97230093e-02 6.16551638e-01 1.38601387e+00 -9.29522514e-01
-5.13074756e-01 -2.54896551e-01 2.88459975e-02 -1.69943973e-01
4.91123438e-01 -1.48259521e-01 -1.13147306e+00 -8.96509066e-02
-3.90044391e-01 -4.49489921e-01 1.02962208e+00 7.63093293e-01
1.58227348e+00 -1.10472642e-01 -9.08292532e-01 8.08673203e-01
1.32027006e+00 8.47242355e-01 6.63057745e-01 3.89705598e-01
8.09356213e-01 3.52365136e-01 4.18476075e-01 3.81865412e-01
9.05221049e-03 5.43580592e-01 5.76638639e-01 -3.29222441e-01
6.13585711e-02 1.41047230e-02 -1.02748573e-02 4.61582512e-01
-6.85997844e-01 -5.86723030e-01 -6.45956457e-01 1.08468458e-01
-1.69588876e+00 -8.60429168e-01 5.03051514e-03 1.95313227e+00
1.25735247e+00 -4.27740812e-01 5.13431787e-01 -4.81605306e-02
4.31277364e-01 2.04125345e-01 -8.21702957e-01 -4.49731559e-01
-4.39118370e-02 8.28248858e-01 4.54846144e-01 4.35493469e-01
-1.49460602e+00 1.28614116e+00 7.71720600e+00 1.10814404e+00
-1.40361118e+00 -2.51144171e-01 1.05957747e+00 2.81289190e-01
-1.84455421e-02 -3.93776685e-01 -8.01827431e-01 1.92496479e-01
1.11478555e+00 -2.89425850e-01 6.90018907e-02 7.66219616e-01
4.38678592e-01 4.39711720e-01 -1.43145120e+00 1.01288378e+00
-1.68583840e-01 -2.19849610e+00 4.79784131e-01 2.15713903e-01
8.11462998e-01 6.43325150e-02 3.57003272e-01 8.51873234e-02
-1.20037682e-01 -1.60431564e+00 -6.44092709e-02 1.14875980e-01
7.74537086e-01 -1.01339340e+00 5.35315514e-01 -1.03911072e-01
-1.09083807e+00 2.20293775e-02 -4.82610703e-01 -1.45651564e-01
-2.81139135e-01 3.69995773e-01 -8.22069645e-01 8.97301972e-01
1.74532279e-01 1.37023425e+00 -1.33053169e-01 1.25771332e+00
2.20508575e-02 3.30995202e-01 3.35607141e-01 -3.66633445e-01
4.42957312e-01 -3.12499642e-01 -4.91115672e-04 1.38841438e+00
-1.68556929e-01 2.13178590e-01 3.41906637e-01 5.71945429e-01
-3.90735090e-01 5.39440393e-01 -5.29302776e-01 -5.52328646e-01
3.34135443e-01 1.06504142e+00 -6.42998576e-01 -3.28793287e-01
-6.49574935e-01 6.26421690e-01 -4.17186581e-02 3.12075078e-01
-8.68520498e-01 -4.98846024e-01 9.67039108e-01 -1.84790984e-01
-2.19535595e-03 -7.97702894e-02 -2.64374733e-01 -8.30091119e-01
-8.00755143e-01 -1.13318765e+00 3.69230360e-01 -7.46028051e-02
-1.12196279e+00 4.21984136e-01 -3.09269220e-01 -9.95834589e-01
4.10139561e-01 -1.08742154e+00 -2.81657070e-01 7.42057383e-01
-1.54151905e+00 -8.28035474e-01 3.57247859e-01 2.48187929e-01
8.93616498e-01 -4.55413163e-01 1.25264823e+00 5.98912656e-01
-9.43801522e-01 4.55066711e-01 5.59453845e-01 -1.64502978e-01
9.26979661e-01 -1.11854887e+00 5.11078596e-01 -6.08198531e-02
-7.78057352e-02 1.08782613e+00 5.81363678e-01 -5.04970074e-01
-1.23678708e+00 -1.18312824e+00 7.92280495e-01 -3.48042160e-01
6.59786344e-01 -1.59621730e-01 -5.52121460e-01 3.70407403e-01
2.73858160e-01 -2.21472770e-01 1.46594608e+00 -9.63642895e-02
-2.36714885e-01 1.70639500e-01 -9.12305057e-01 5.62931478e-01
8.77768278e-01 -2.44760737e-01 1.69007316e-01 9.83721256e-01
4.41365808e-01 -4.09332454e-01 -1.23419690e+00 5.11263371e-01
7.50300765e-01 -6.05990887e-01 1.34427607e+00 -1.54829466e+00
5.89233398e-01 -2.10300818e-01 2.66610324e-01 -9.31550324e-01
-7.48745382e-01 -6.54935420e-01 -2.29083285e-01 3.20586085e-01
5.02116501e-01 -2.10979670e-01 9.01622951e-01 -9.01876613e-02
-3.54010910e-01 -1.49501002e+00 -5.50758421e-01 -3.80950123e-01
6.36269391e-01 3.50238383e-02 4.47487682e-01 1.14787519e+00
3.02669883e-01 9.56389248e-01 -5.56483567e-01 -2.44971961e-01
2.86378056e-01 1.95651099e-01 3.44572037e-01 -1.10069919e+00
-1.88272446e-01 -5.79734981e-01 -6.14658117e-01 -1.06324697e+00
7.89747760e-02 -1.06048667e+00 -6.16969049e-01 -1.64328611e+00
7.14540899e-01 1.46306753e-01 -4.94215459e-01 3.34639966e-01
-5.31572923e-02 2.23504260e-01 -5.54055452e-01 6.05336204e-02
-5.28836489e-01 4.07940865e-01 1.38845980e+00 -5.07393837e-01
-2.11032614e-01 -4.29815166e-02 -8.18233192e-01 5.35166681e-01
8.46355438e-01 -2.68586278e-01 -4.51869220e-01 -4.59076688e-02
3.38060111e-01 -2.43856013e-01 -1.62506416e-01 -5.14709651e-01
-4.81342599e-02 -3.98512721e-01 7.62163043e-01 -5.49598813e-01
2.35811993e-01 -3.75581801e-01 -9.85471765e-04 7.48652279e-01
-4.87640649e-01 -4.16189820e-01 1.79391235e-01 7.35248327e-01
-8.30059126e-02 9.28826034e-02 9.90280390e-01 -3.70527625e-01
-5.24654031e-01 1.04534137e+00 -7.01951206e-01 -6.18493855e-01
1.13903487e+00 -1.86685354e-01 -1.21516369e-01 2.80184820e-02
-9.28655624e-01 -3.88665758e-02 -1.49928719e-01 2.68691450e-01
5.13770461e-01 -1.04452682e+00 -4.35493648e-01 -4.62834716e-01
2.51132935e-01 -3.82636368e-01 2.77130175e-02 9.59807396e-01
-8.76745641e-01 1.11066592e+00 -1.63270146e-01 -5.54604717e-02
-1.33222556e+00 1.06702197e+00 8.21078777e-01 -2.32249230e-01
-1.49159178e-01 8.12198877e-01 6.74770474e-01 1.29580628e-02
6.26499116e-01 -2.61664212e-01 -7.02598631e-01 -3.58210467e-02
6.34659052e-01 4.26986426e-01 3.36116880e-01 -2.63264179e-01
-6.42214656e-01 4.45055783e-01 -5.27670681e-01 8.09893429e-01
1.62236094e+00 5.03085852e-01 -1.27912834e-01 6.39465153e-02
1.19327569e+00 -4.03917968e-01 -5.01056790e-01 7.10744262e-02
1.11678712e-01 2.75481015e-01 1.23565197e-01 -9.89137888e-01
-8.03955793e-01 1.04132175e+00 5.45194447e-01 -1.97902948e-01
7.48194873e-01 -3.39386463e-02 8.95505965e-01 7.64710724e-01
1.59931019e-01 -6.45563602e-01 2.85495013e-01 5.02525330e-01
6.18799448e-01 -1.06459546e+00 4.31626886e-01 -5.26797116e-01
-2.20554337e-01 1.45517266e+00 1.62524715e-01 1.74248405e-02
6.75961435e-01 -1.56344965e-01 -2.94034749e-01 -4.93071347e-01
-6.33755922e-01 -6.86063766e-02 8.74335766e-02 7.50079453e-01
1.58758759e+00 -9.58974361e-02 -7.33417630e-01 3.98464918e-01
2.21927583e-01 4.34897281e-02 2.00795054e-01 8.43002260e-01
-5.79665840e-01 -1.77508235e+00 7.26115406e-02 9.71716195e-02
-9.08808112e-01 -4.73937988e-01 -1.13370597e+00 4.70307916e-01
2.22916707e-01 8.84407640e-01 -4.67830062e-01 -1.88956514e-01
1.52157381e-01 -3.09702114e-06 9.42273736e-01 -7.97754765e-01
-5.09018779e-01 2.66029716e-01 1.46952584e-01 -3.34349394e-01
-6.87361956e-01 -2.66782641e-01 -1.11032891e+00 -4.95116651e-01
-3.71824920e-01 2.85592079e-01 4.70731318e-01 8.77367496e-01
4.10316825e-01 4.69080389e-01 5.74459672e-01 -6.24338925e-01
-2.10634708e-01 -8.82476985e-01 -5.06326795e-01 -3.57054919e-01
4.21950340e-01 -7.55817711e-01 3.11627120e-01 2.28872538e-01] | [5.06015682220459, 5.8448920249938965] |
af346a13-b199-432a-898f-b72b32826b97 | impacts-and-risk-of-generative-ai-technology | 2306.13033 | null | https://arxiv.org/abs/2306.13033v1 | https://arxiv.org/pdf/2306.13033v1.pdf | Impacts and Risk of Generative AI Technology on Cyber Defense | Generative Artificial Intelligence (GenAI) has emerged as a powerful technology capable of autonomously producing highly realistic content in various domains, such as text, images, audio, and videos. With its potential for positive applications in creative arts, content generation, virtual assistants, and data synthesis, GenAI has garnered significant attention and adoption. However, the increasing adoption of GenAI raises concerns about its potential misuse for crafting convincing phishing emails, generating disinformation through deepfake videos, and spreading misinformation via authentic-looking social media posts, posing a new set of challenges and risks in the realm of cybersecurity. To combat the threats posed by GenAI, we propose leveraging the Cyber Kill Chain (CKC) to understand the lifecycle of cyberattacks, as a foundational model for cyber defense. This paper aims to provide a comprehensive analysis of the risk areas introduced by the offensive use of GenAI techniques in each phase of the CKC framework. We also analyze the strategies employed by threat actors and examine their utilization throughout different phases of the CKC, highlighting the implications for cyber defense. Additionally, we propose GenAI-enabled defense strategies that are both attack-aware and adaptive. These strategies encompass various techniques such as detection, deception, and adversarial training, among others, aiming to effectively mitigate the risks posed by GenAI-induced cyber threats. | ['Shahram Rahimi', 'Sudip Mittal', 'Ivan A. Fernandez', 'Subash Neupane'] | 2023-06-22 | null | null | null | null | ['face-swapping', 'misinformation'] | ['computer-vision', 'miscellaneous'] | [ 5.86108863e-01 3.28469425e-01 5.44717163e-02 6.53565407e-01
-5.16848683e-01 -1.16551554e+00 1.27700746e+00 8.97690654e-03
2.23081727e-02 3.80494356e-01 4.75415617e-01 -6.10702634e-01
-3.28904688e-01 -8.95370901e-01 -5.90647340e-01 -5.06995440e-01
5.64983673e-02 -6.07991703e-02 -2.71092385e-01 -4.21852142e-01
7.20113397e-01 7.25401163e-01 -1.20006633e+00 3.22393715e-01
6.01870358e-01 9.41478908e-01 -3.68897706e-01 7.52583683e-01
1.95312593e-02 1.00772679e+00 -1.47110760e+00 -1.07749403e+00
1.87179074e-01 -5.25359809e-01 -5.41315436e-01 -1.72027290e-01
4.48605977e-02 -5.20441651e-01 -5.56280077e-01 1.12981761e+00
4.17814106e-01 -1.70843646e-01 4.20183271e-01 -1.66945362e+00
-7.57278264e-01 7.05640316e-01 -7.20892772e-02 1.61204532e-01
7.91126072e-01 5.33038497e-01 3.90238523e-01 -3.42962682e-01
9.56452072e-01 1.36661148e+00 4.65218782e-01 6.57698214e-01
-7.72247970e-01 -9.86386478e-01 -1.75882578e-01 5.86370751e-02
-9.97421682e-01 -9.04448852e-02 9.59806204e-01 -4.81626928e-01
5.75625896e-01 4.74421382e-01 7.51091838e-01 2.10098743e+00
7.09580004e-01 7.11228132e-01 9.35620189e-01 -3.46069664e-01
4.57040757e-01 2.85297185e-01 -4.64199334e-01 2.31099039e-01
4.46369708e-01 4.87427145e-01 -6.30480409e-01 -6.71853721e-01
5.71228325e-01 -2.43877575e-01 1.51645048e-02 2.97906429e-01
-9.68610764e-01 1.00366342e+00 2.15214878e-01 3.99900466e-01
-5.93315661e-01 1.48615524e-01 6.91449344e-01 -2.52761971e-02
4.53480244e-01 1.02534974e+00 1.73913881e-01 -4.71497655e-01
-2.35877797e-01 3.49125147e-01 8.33476245e-01 6.18778288e-01
-1.97876301e-02 5.17951131e-01 9.99379531e-02 1.40430868e-01
3.33376169e-01 4.89053100e-01 1.32702351e-01 -9.75336611e-01
2.77280122e-01 3.45119387e-01 -7.32999742e-02 -1.65789950e+00
1.13952182e-01 -4.98276979e-01 -4.23794806e-01 2.28973925e-01
8.70698169e-02 -5.75309098e-01 -5.98848343e-01 1.32280731e+00
3.65837216e-01 3.04039389e-01 1.55700386e-01 5.94630420e-01
2.35955998e-01 8.80754232e-01 3.95467073e-01 -2.04660222e-02
1.03357089e+00 -1.72724962e-01 -7.51970589e-01 -1.28531978e-01
3.64367515e-01 -5.41040361e-01 5.51840246e-01 8.56868982e-01
-7.73460448e-01 -6.47060052e-02 -9.69889760e-01 5.98085761e-01
-6.21077001e-01 -9.09409225e-01 5.73695242e-01 1.44588912e+00
-4.57755834e-01 2.83997983e-01 -2.39113733e-01 -7.42432848e-02
4.58494693e-01 -2.31746301e-01 7.43170008e-02 -9.92019549e-02
-1.46062052e+00 7.47111380e-01 4.88644332e-01 2.90090032e-02
-1.15875137e+00 -6.53199315e-01 -7.51253784e-01 -1.36571541e-01
6.89481735e-01 -3.40938032e-01 8.14103484e-01 -9.26315725e-01
-1.36567748e+00 1.85328439e-01 9.60355401e-01 -6.16578817e-01
4.57516134e-01 -4.55636144e-01 -8.23181808e-01 3.22574615e-01
-2.11370274e-01 3.88761222e-01 1.04339707e+00 -1.37100708e+00
-1.71341509e-01 -1.96470097e-01 2.27855712e-01 -2.01976169e-02
-5.69900334e-01 4.73658323e-01 3.65582317e-01 -9.14402902e-01
-5.03985226e-01 -9.58127499e-01 -1.09553270e-01 -5.74494779e-01
-7.58765101e-01 2.58076400e-01 1.38367867e+00 -9.75164056e-01
1.15455222e+00 -2.29982972e+00 2.59900000e-02 4.85486627e-01
1.73731834e-01 6.94841862e-01 1.25281453e-01 1.08529580e+00
3.17241371e-01 7.41169274e-01 2.09625140e-01 1.79049581e-01
-2.35350300e-02 -2.79236794e-03 -7.63511121e-01 7.65121877e-02
1.92692131e-01 6.94570065e-01 -9.37079430e-01 2.29649842e-02
1.43705070e-01 6.45984054e-01 -3.79412115e-01 -9.03006122e-02
-3.79551440e-01 5.10768652e-01 -5.69486558e-01 8.66688013e-01
3.54815483e-01 3.32671046e-01 2.44186863e-01 2.04415649e-01
-1.07087173e-01 -1.32063618e-02 -5.84397733e-01 9.06411409e-01
-2.18874797e-01 8.30185711e-01 2.27362558e-01 -3.37622136e-01
7.44430602e-01 3.68714601e-01 2.03893229e-01 -5.22497296e-01
6.36814117e-01 -1.01203799e-01 -8.96706656e-02 -5.60911477e-01
5.36381602e-01 1.41686536e-02 -4.99735057e-01 8.02850187e-01
-1.27111048e-01 -1.58705130e-01 -2.71507323e-01 6.48478329e-01
1.27969944e+00 -1.11725979e-01 -6.66522011e-02 4.43073899e-01
2.23615468e-01 2.03433722e-01 1.02509603e-01 7.77254343e-01
-2.10212246e-01 6.59070723e-03 6.13250673e-01 -1.75039142e-01
-8.64882052e-01 -8.86907339e-01 4.92488325e-01 7.44396269e-01
9.91066843e-02 -5.06338060e-01 -1.12965524e+00 -9.38520074e-01
-9.69700441e-02 1.36684227e+00 -4.41478878e-01 -6.53467596e-01
-3.02643746e-01 -3.58163059e-01 1.21427441e+00 8.12445059e-02
5.01622915e-01 -1.33360016e+00 -6.78695798e-01 2.46940151e-01
-2.55049169e-01 -1.23840654e+00 -1.14247203e-01 -5.49178541e-01
-1.89537019e-01 -1.10281408e+00 -1.67149067e-01 1.41464442e-01
2.59649783e-01 3.32320243e-01 2.97621757e-01 -5.15788123e-02
-5.63072681e-01 9.26060021e-01 -7.24553168e-01 -9.63787019e-01
-1.19360578e+00 -3.34787995e-01 1.26944751e-01 1.63106218e-01
-1.15769900e-01 -3.07657540e-01 -2.00972870e-01 1.51101112e-01
-1.46394515e+00 -2.36415759e-01 5.38039386e-01 3.61094773e-01
-3.17145556e-01 4.26280111e-01 7.79553831e-01 -7.63035417e-01
1.20173645e+00 -1.01155913e+00 -1.96621209e-01 5.97425662e-02
-4.49765354e-01 -5.96823573e-01 4.01832432e-01 -8.89212847e-01
-1.31111860e+00 -6.29422605e-01 -6.94651976e-02 -2.97057271e-01
-2.83890784e-01 6.73016846e-01 -2.66915381e-01 -3.54719341e-01
1.02273774e+00 1.56066015e-01 4.02835697e-01 -3.31212580e-02
4.28160906e-01 6.59530282e-01 3.19065511e-01 -2.77248979e-01
1.16112232e+00 3.51745456e-01 -1.58740252e-01 -9.42711174e-01
-5.96423626e-01 9.86164138e-02 1.08817577e-01 -9.94804502e-01
8.20187390e-01 -3.13466936e-01 -8.38254213e-01 7.35887825e-01
-1.16574740e+00 -2.42533218e-02 8.75352696e-02 1.11440293e-01
-2.29321092e-01 5.52295446e-01 -7.06294775e-01 -1.13420582e+00
-4.64678258e-01 -8.36663127e-01 5.15135169e-01 2.10812911e-01
-5.41516721e-01 -8.36288929e-01 -1.90181041e-03 9.65600967e-01
6.57791436e-01 7.87911832e-01 9.38958585e-01 -1.05084062e+00
-7.20055044e-01 -7.81778812e-01 4.24665175e-02 4.67588723e-01
-1.69253528e-01 1.25076354e-01 -6.86909735e-01 6.19633384e-02
1.33610606e-01 -3.54948670e-01 -1.39739320e-01 -2.95367122e-01
6.81505442e-01 -9.43031430e-01 -2.56391205e-02 1.41139179e-01
1.12030029e+00 9.16928589e-01 9.53350902e-01 5.31991780e-01
5.84893465e-01 8.96559477e-01 6.73663914e-01 5.75037658e-01
-2.89314866e-01 3.40449899e-01 9.45883095e-01 4.36341465e-01
1.11967035e-01 -6.13044322e-01 5.86570382e-01 2.34886914e-01
1.20117131e-03 -5.52221954e-01 -8.68617833e-01 -5.56367226e-02
-1.28102863e+00 -1.36629903e+00 3.85297649e-02 1.79788017e+00
1.90672725e-01 4.49237794e-01 1.53672295e-02 3.14540535e-01
6.99482441e-01 1.70510918e-01 -2.36678839e-01 -7.85985649e-01
1.44998627e-02 5.13311364e-02 4.33213949e-01 1.17309369e-01
-7.34673023e-01 9.24887538e-01 6.70029831e+00 8.50277960e-01
-8.65110815e-01 1.19807616e-01 5.94633520e-01 9.78256389e-02
-3.88534158e-01 -6.14939770e-03 -4.55625713e-01 5.99110007e-01
1.21608567e+00 -1.86781377e-01 5.60511172e-01 9.35797870e-01
2.24379152e-01 -3.43556106e-02 -2.71986157e-01 3.22802722e-01
6.05031073e-01 -1.65343022e+00 2.14364275e-01 4.69079524e-01
5.65218508e-01 -5.93762279e-01 3.86224359e-01 2.23956361e-01
2.90479511e-01 -7.58736670e-01 9.22803402e-01 1.17512554e-01
3.33635092e-01 -1.05428779e+00 5.84034264e-01 4.40904319e-01
-2.76824802e-01 -5.16585648e-01 4.36292291e-01 -1.00055799e-01
4.81416464e-01 1.28187642e-01 -9.83093798e-01 4.31624562e-01
2.97227174e-01 -1.24194980e-01 -2.71595597e-01 6.45245016e-01
-3.99265319e-01 8.35486054e-01 9.85923558e-02 -1.33689076e-01
3.40556026e-01 1.35152549e-01 1.03655565e+00 1.03633249e+00
3.42785269e-01 2.01725125e-01 -1.36582077e-01 7.58839011e-01
3.07834893e-01 -3.39842469e-01 -1.02513397e+00 -8.38180244e-01
6.65992856e-01 1.09672868e+00 -6.22760475e-01 5.82716577e-02
1.49609357e-01 9.80411172e-01 -4.88405138e-01 2.62802333e-01
-1.01436377e+00 -5.26329219e-01 5.99099219e-01 2.53875941e-01
-3.11585456e-01 -4.64029670e-01 -4.98063743e-01 -5.59979737e-01
-5.55299878e-01 -1.43709826e+00 1.89629540e-01 -9.01117802e-01
-1.01052082e+00 5.86172819e-01 3.11878949e-01 -1.00651300e+00
-6.30915225e-01 -4.44283515e-01 -7.71106064e-01 3.03768009e-01
-5.46534777e-01 -1.28881156e+00 -5.77803850e-02 3.72578055e-01
3.81267160e-01 -4.58273619e-01 5.67388475e-01 -4.62663546e-02
-4.33784097e-01 3.68284583e-01 -4.52942908e-01 -9.36679021e-02
1.96583778e-01 -3.83427173e-01 5.83150148e-01 1.12319684e+00
-4.84907515e-02 6.80109560e-01 8.59357178e-01 -1.32798493e+00
-1.76749170e+00 -7.38820791e-01 3.61628085e-01 -4.54890251e-01
1.01729560e+00 -5.78969717e-01 -5.83197176e-01 5.11720598e-01
1.81348428e-01 -1.07933104e+00 8.83040011e-01 -3.67717892e-01
-5.88237643e-01 4.53784943e-01 -1.52156544e+00 9.91151214e-01
9.57076788e-01 -6.06155932e-01 -1.56729117e-01 3.54649812e-01
1.00538790e+00 -2.26516295e-02 -5.48693061e-01 -1.11136250e-01
3.78022462e-01 -8.77653241e-01 9.98235583e-01 -4.84408647e-01
5.85560441e-01 2.22902969e-01 7.60082304e-02 -1.12535357e+00
-1.92373589e-01 -1.25458932e+00 -3.45967174e-01 1.37350214e+00
1.50919542e-01 -7.54241347e-01 6.90638304e-01 7.38106012e-01
-1.34976804e-01 -1.99266776e-01 -8.28707576e-01 -7.69933939e-01
-3.90088320e-01 -9.15437937e-01 2.87666351e-01 1.13567078e+00
2.50935465e-01 -1.26667358e-02 -7.34887421e-01 3.64166051e-01
7.68281400e-01 -9.78372931e-01 8.91382754e-01 -9.47277427e-01
-2.17973366e-01 -3.46136421e-01 -6.16827726e-01 -4.76071313e-02
-1.24653749e-01 -4.68451440e-01 -3.75413448e-01 -1.12416053e+00
-2.13098675e-01 -1.09051026e-01 2.77465075e-01 2.01581851e-01
1.97349072e-01 3.34698796e-01 7.32238114e-01 1.27021223e-01
2.50885859e-02 1.20689593e-01 1.07453704e+00 -3.18351351e-02
-8.99502784e-02 -5.21429107e-02 -1.02700615e+00 6.31723642e-01
8.03148031e-01 -4.21924859e-01 -4.83432889e-01 1.57438755e-01
6.12268746e-01 1.21442340e-01 6.59731686e-01 -8.64493370e-01
7.81468302e-02 -5.28539717e-01 1.44581079e-01 -1.65885717e-01
4.70047206e-01 -7.49694467e-01 5.40889919e-01 7.29135513e-01
-1.80388629e-01 -3.42845172e-02 3.28097373e-01 8.03528130e-01
1.24115311e-01 -2.97812343e-01 3.18164140e-01 9.15711746e-02
-3.95411193e-01 -3.00779521e-01 -1.25749481e+00 -2.50969321e-01
1.51887238e+00 -3.59191746e-01 -7.31731892e-01 -8.93765450e-01
-4.56594318e-01 -3.52021486e-01 1.80564687e-01 7.64703453e-01
7.84654200e-01 -1.01795411e+00 -3.94848794e-01 1.94205403e-01
-2.10739598e-01 -6.90287113e-01 5.94208300e-01 1.67395011e-01
-5.25343478e-01 1.04826950e-01 -3.31232131e-01 1.70229957e-01
-1.20888448e+00 8.81408751e-01 -1.94010675e-01 -3.60909328e-02
-4.60243285e-01 5.64358592e-01 -1.07069984e-01 8.77443925e-02
2.18755737e-01 6.67515159e-01 -6.58286810e-02 -1.11833937e-01
8.21662426e-01 7.48664260e-01 -4.76060778e-01 -5.67260861e-01
-1.31769389e-01 -3.58761400e-01 -2.37000585e-01 -4.56454009e-01
8.46249700e-01 1.31982744e-01 4.28312607e-02 -1.06988810e-01
6.28480792e-01 2.75136679e-01 -9.06145751e-01 3.03964913e-01
5.29635809e-02 -6.72596395e-01 -1.34207085e-01 -1.31406999e+00
-6.17995918e-01 4.22034591e-01 5.88781247e-03 7.86840022e-01
9.41351891e-01 -1.34518459e-01 1.03588569e+00 -8.09109956e-02
5.11610687e-01 -9.93587434e-01 7.48060644e-01 3.48199666e-01
1.09392333e+00 -4.18779314e-01 -2.01317042e-01 -6.23599946e-01
-1.08556235e+00 8.60213399e-01 4.35038745e-01 6.26687557e-02
3.14946622e-01 3.76652896e-01 -7.14591369e-02 -2.11407185e-01
-4.39328939e-01 5.13189316e-01 5.15684336e-02 1.07537854e+00
-2.34069884e-01 -4.70142625e-03 -2.50250280e-01 4.09204662e-01
-8.95260554e-03 -3.63803089e-01 9.57641482e-01 1.18934858e+00
-2.69906968e-01 -1.12554228e+00 -9.50149298e-01 6.00367412e-02
-8.03693116e-01 9.63193625e-02 -1.23081756e+00 8.04078579e-01
9.49341953e-02 1.23122692e+00 -3.36445242e-01 -1.05288160e+00
-5.59692830e-02 6.06486686e-02 5.14973560e-03 -1.60600424e-01
-1.09343314e+00 -1.26594171e-01 4.56530333e-01 -3.31347674e-01
1.82148784e-01 -6.77787185e-01 -7.32156992e-01 -7.16709197e-01
-1.42809883e-01 -1.19940437e-01 1.30636430e+00 8.35797012e-01
6.20957136e-01 3.85780275e-01 5.96557975e-01 -8.03113103e-01
-5.22409439e-01 -7.21899569e-01 -5.72066493e-02 2.58590788e-01
-3.39893103e-01 -4.75354373e-01 -4.04989034e-01 -3.53232652e-01] | [6.217748641967773, 7.777961254119873] |
589cc696-2046-4303-aeac-f7dd467b5510 | deep-recursive-hdri-inverse-tone-mapping | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Siyeong_Lee_Deep_Recursive_HDRI_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Siyeong_Lee_Deep_Recursive_HDRI_ECCV_2018_paper.pdf | Deep Recursive HDRI: Inverse Tone Mapping using Generative Adversarial Networks | High dynamic range images contain luminance information of the physical world and provide more realistic experience than conventional low dynamic range images. Because most images have a low dynamic range, recovering the lost dynamic range from a single low dynamic range image is still prevalent. We propose a novel method for restoring the lost dynamic range from a single low dynamic range image through a deep neural network. The proposed method is the first framework to create high dynamic range images based on the estimated multi-exposure stack using the conditional generative adversarial network structure. In this architecture, we train the network by setting an objective function that is a combination of L1 loss and generative adversarial network loss. In addition, this architecture has a simplified structure than the existing networks. In the experimental results, the proposed network generated a multi-exposure stack consisting of realistic images with varying exposure values while avoiding artifacts on public benchmarks, compared with the existing methods. In addition, both the multi-exposure stacks and high dynamic range images estimated by the proposed method are significantly similar to the ground truth than other state-of-the-art algorithms. | ['Suk-Ju Kang', 'Siyeong Lee', 'Gwon Hwan An'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['tone-mapping', 'inverse-tone-mapping'] | ['computer-vision', 'computer-vision'] | [ 4.94820774e-01 -4.46526408e-01 5.05169392e-01 -4.09489602e-01
-8.41739833e-01 -4.67709869e-01 2.99780786e-01 -7.11848855e-01
-4.24552739e-01 8.47003222e-01 1.12691633e-01 -3.62082459e-02
-1.02231830e-01 -1.16816163e+00 -1.24079812e+00 -9.58036423e-01
3.16180319e-01 -1.59700677e-01 3.02623600e-01 -3.08664501e-01
1.94351897e-01 4.89704013e-01 -1.43633664e+00 1.71007976e-01
9.99065876e-01 1.25154305e+00 3.23355973e-01 7.16068387e-01
3.27410012e-01 7.58784533e-01 -8.49142730e-01 -3.26898009e-01
6.06203496e-01 -5.08233607e-01 -1.09888308e-01 -7.57266805e-02
7.17390358e-01 -8.32434297e-01 -7.10244298e-01 1.26975310e+00
7.13305950e-01 3.12931091e-01 2.79492974e-01 -1.06907141e+00
-8.83891046e-01 2.77197838e-01 -7.29905486e-01 1.24089807e-01
3.16894382e-01 4.37165648e-01 1.03031397e-01 -4.14118737e-01
5.16377330e-01 1.20978236e+00 7.25587666e-01 3.20309967e-01
-1.07738614e+00 -1.03736293e+00 -1.61148980e-02 2.06129089e-01
-1.23684275e+00 -1.45355314e-01 8.91614556e-01 -1.10211939e-01
5.48303962e-01 1.39527708e-01 6.43279672e-01 1.09818697e+00
4.97938722e-01 2.12147787e-01 1.63338602e+00 -1.37665227e-01
-2.24475130e-01 -1.33694604e-01 -3.25429976e-01 3.99354607e-01
2.29337215e-01 5.33754528e-01 -1.46390393e-01 2.31612518e-01
8.34192097e-01 2.02728406e-01 -8.20100546e-01 -1.26984879e-01
-1.29131246e+00 2.09500253e-01 7.84695804e-01 -8.69692478e-04
-1.24160148e-01 3.20028663e-01 9.31862965e-02 3.61770898e-01
2.78767109e-01 1.27754048e-01 -8.13445002e-02 9.76867601e-02
-8.40038121e-01 1.37846887e-01 4.71468091e-01 6.63246095e-01
7.27484941e-01 2.64184266e-01 -1.54313877e-01 6.19383276e-01
3.19773555e-01 9.07831967e-01 4.61985141e-01 -8.46977711e-01
6.03233576e-01 1.13510899e-01 1.81861535e-01 -1.08864737e+00
-1.37441546e-01 -3.97015989e-01 -1.21791852e+00 6.57415569e-01
3.71795386e-01 -3.75007018e-02 -1.16369867e+00 1.94493651e+00
1.45240515e-01 4.34351087e-01 3.75132024e-01 1.06506824e+00
7.42104650e-01 1.11797726e+00 -1.78712904e-01 -3.04077089e-01
9.82743621e-01 -8.72415960e-01 -1.08782578e+00 -1.90515846e-01
-4.56202388e-01 -9.44842160e-01 1.30545831e+00 5.64260721e-01
-1.03132606e+00 -8.96494865e-01 -1.48592472e+00 -1.28360331e-01
-3.07430208e-01 -2.29520053e-01 2.91118890e-01 7.65678167e-01
-9.69578624e-01 5.62374353e-01 -3.91055256e-01 2.50900954e-01
1.64063737e-01 -6.05383003e-03 -3.76264602e-01 -3.33355933e-01
-1.53266001e+00 6.64839983e-01 6.61618710e-01 4.70746666e-01
-9.48009133e-01 -9.00989890e-01 -9.20095623e-01 -4.98616248e-02
3.03208649e-01 -5.96390724e-01 7.16482341e-01 -9.88172889e-01
-1.72497666e+00 4.78574336e-01 3.21321785e-01 -2.25695893e-01
8.24380577e-01 -3.30166996e-01 -6.66336775e-01 1.26955971e-01
-1.70957759e-01 2.77705461e-01 8.65059316e-01 -1.53162313e+00
-2.31782898e-01 -6.16864637e-02 1.48617802e-02 8.33876878e-02
1.32338986e-01 -2.11983293e-01 -5.72024643e-01 -7.73763895e-01
1.94764182e-01 -8.52957070e-01 1.19684786e-01 1.55782238e-01
-5.10355353e-01 6.53725088e-01 1.03131485e+00 -7.95108199e-01
8.59269857e-01 -2.04814935e+00 -2.54662693e-01 3.44873033e-02
2.41618250e-02 3.56009811e-01 -7.47142136e-02 -5.93730770e-02
-2.74231136e-01 -7.09606102e-03 -4.81951118e-01 2.00351002e-03
-1.37816891e-01 -2.29866236e-01 -7.00504780e-01 5.36117971e-01
-2.04030246e-01 8.81768584e-01 -8.43079150e-01 -2.32295856e-01
7.23372340e-01 9.82451081e-01 -1.02434993e-01 4.86512661e-01
-7.29817227e-02 6.42762482e-01 -3.82566527e-02 5.31493247e-01
1.41306746e+00 4.25426231e-04 -1.90963984e-01 -6.28755271e-01
-4.39277887e-02 -4.49394792e-01 -1.16858077e+00 1.63307428e+00
-6.80328906e-01 7.88163602e-01 -2.16518581e-01 -3.31250459e-01
1.04610646e+00 -1.42195567e-01 1.98408455e-01 -1.15171087e+00
1.72280967e-01 7.80198276e-02 -1.79392830e-01 -5.31191766e-01
4.35729831e-01 -3.31010491e-01 -1.12601310e-01 3.63208234e-01
-3.70393187e-01 -2.71979570e-01 -6.35854825e-02 -5.97985834e-02
7.42399454e-01 4.51200306e-01 -1.49545103e-01 4.41499472e-01
7.51902640e-01 -5.92151880e-01 7.23133922e-01 8.08221042e-01
-1.05837956e-01 1.13019383e+00 2.48422951e-01 -2.54175514e-01
-1.30361795e+00 -1.46768212e+00 -2.15444282e-01 3.46265167e-01
7.93778419e-01 2.06210047e-01 -7.19256103e-01 -3.51271600e-01
-2.72949040e-01 4.51954901e-01 -4.43658561e-01 -3.98018360e-01
-7.01394200e-01 -8.61459911e-01 7.09303319e-01 3.35816771e-01
1.48149216e+00 -1.04972732e+00 -3.03768188e-01 -6.03300296e-02
-4.37783927e-01 -1.34139931e+00 -6.27724886e-01 -3.21528375e-01
-4.80290949e-01 -1.08692062e+00 -8.94198895e-01 -5.64253151e-01
4.64217514e-01 1.70365408e-01 1.15457737e+00 -8.13742876e-02
-3.05574656e-01 -8.89736786e-02 -2.38970175e-01 -2.61422426e-01
-4.83787775e-01 -3.95148546e-01 -2.69668430e-01 3.21278363e-01
-1.62205756e-01 -5.73535204e-01 -1.14380574e+00 3.99264723e-01
-1.31311893e+00 2.02786267e-01 7.87645042e-01 7.48103261e-01
1.02762020e+00 4.26250666e-01 5.00273526e-01 -8.50384474e-01
4.11206841e-01 -3.24664742e-01 -9.14723456e-01 2.31772214e-01
-6.82179570e-01 -6.25830814e-02 9.54534113e-01 -6.68590069e-01
-1.70204210e+00 -8.36586356e-02 -3.56581807e-01 -4.68881696e-01
-4.88863438e-02 -2.28851795e-01 -6.37770772e-01 -3.26908320e-01
4.83978212e-01 4.94756609e-01 -1.05572045e-01 -2.35874906e-01
3.11041594e-01 5.50576448e-01 1.06693089e+00 -3.63741279e-01
1.11450481e+00 6.87159002e-01 1.54389620e-01 -2.60260731e-01
-8.12889099e-01 1.30398571e-01 -9.58058760e-02 -4.53142077e-01
9.18345034e-01 -1.05829751e+00 -8.35928142e-01 1.25439203e+00
-9.08419251e-01 -4.90298271e-01 -2.01397508e-01 5.07547975e-01
-7.25717664e-01 4.72271144e-01 -6.35280907e-01 -4.36524183e-01
-4.49196070e-01 -1.25215089e+00 9.86611307e-01 7.19353497e-01
7.61655211e-01 -7.15136766e-01 3.97902057e-02 3.04761052e-01
7.82871008e-01 1.04994071e+00 6.16846621e-01 2.82869816e-01
-9.91159201e-01 -1.17301278e-01 -4.63223726e-01 7.77440190e-01
7.52899423e-02 -2.10201129e-01 -1.06178105e+00 -3.86344433e-01
3.57443511e-01 -2.73992091e-01 8.78795505e-01 4.99081820e-01
1.60081136e+00 -6.66017532e-02 3.11147273e-02 1.38612604e+00
2.09896803e+00 3.47929716e-01 1.41149163e+00 6.37033224e-01
8.72720122e-01 1.69949323e-01 7.69594133e-01 1.18442178e-01
1.58049408e-02 5.93084514e-01 6.41584933e-01 -3.89521182e-01
-4.32504356e-01 -2.74493217e-01 3.67951810e-01 2.33291134e-01
2.52966076e-01 -5.54586828e-01 -4.70027894e-01 2.20406607e-01
-1.45907354e+00 -1.22678626e+00 -1.65578440e-01 2.37039661e+00
7.97124207e-01 2.25597844e-01 -4.44214433e-01 1.57005005e-02
9.45415676e-01 5.42627811e-01 -8.20392847e-01 -1.82214543e-01
-5.09733200e-01 -1.07685402e-02 7.73944736e-01 3.90856266e-01
-8.97652447e-01 4.93415296e-01 6.01893997e+00 8.22247803e-01
-1.36613762e+00 3.10203042e-02 7.86953986e-01 -1.81823671e-01
-3.97684783e-01 -1.73155457e-01 -6.24691308e-01 1.01474488e+00
8.33390176e-01 -7.02423602e-02 4.96477962e-01 5.30820370e-01
2.82422841e-01 -1.69620737e-01 -6.71258152e-01 1.10131764e+00
3.99549574e-01 -1.04761636e+00 -8.04513246e-02 5.73182143e-02
9.96310592e-01 -2.82347232e-01 5.06906688e-01 9.78573561e-02
7.18970597e-02 -1.08613193e+00 7.07002044e-01 1.02615416e+00
1.33891141e+00 -1.02143860e+00 7.65315354e-01 1.57489985e-01
-1.06572616e+00 -7.76856989e-02 -4.02609020e-01 5.03850579e-01
5.93767583e-01 7.99465358e-01 -7.09057897e-02 6.42200828e-01
1.00129712e+00 5.13307869e-01 -6.82063997e-01 1.02570844e+00
-3.61704499e-01 2.56758094e-01 -2.24645883e-01 4.79512542e-01
-7.26477578e-02 -4.22892690e-01 5.08976877e-01 1.07260871e+00
4.91815269e-01 -4.39067371e-02 -1.47760808e-01 1.06946146e+00
-2.54235148e-01 -3.77209842e-01 -7.13843346e-01 5.13621867e-01
2.66005427e-01 1.13288355e+00 -3.30132723e-01 -2.43570343e-01
-2.89325744e-01 1.00825274e+00 -2.12794125e-01 5.31004667e-01
-1.21847987e+00 -9.70144570e-01 3.72804791e-01 2.52169311e-01
8.66663679e-02 1.01071976e-01 2.14426275e-02 -1.22485125e+00
4.32947516e-01 -8.90015781e-01 5.35685942e-02 -1.41176212e+00
-1.18017828e+00 9.12891567e-01 7.63314217e-02 -1.48600757e+00
-2.00722769e-01 -4.43825245e-01 -5.87034464e-01 1.18802214e+00
-2.00246167e+00 -1.16400731e+00 -1.18164563e+00 7.24186897e-01
2.91000962e-01 -9.53904390e-02 2.63818651e-01 5.16207159e-01
-3.39309305e-01 5.24560094e-01 3.81920516e-01 3.82454060e-02
1.09999871e+00 -1.20396471e+00 3.44496757e-01 1.33750129e+00
-5.34768581e-01 3.69954437e-01 6.31208241e-01 -6.23292983e-01
-1.14646220e+00 -1.24662685e+00 -8.93240422e-02 -3.66542861e-02
1.58548698e-01 -2.23695144e-01 -1.16522396e+00 5.56251168e-01
2.11895898e-01 2.56434560e-01 2.78345287e-01 -8.56423378e-01
-3.80620271e-01 -6.35965109e-01 -1.60949326e+00 3.60079378e-01
8.81178975e-01 -4.29082334e-01 -2.67745316e-01 7.34758675e-02
1.07938623e+00 -8.52345884e-01 -9.16226983e-01 6.11392617e-01
5.22083938e-01 -1.30105364e+00 1.37813377e+00 4.99210089e-01
6.08862221e-01 -7.81439900e-01 -9.72909033e-02 -1.17289591e+00
1.15481675e-01 -6.38589263e-01 -1.52508050e-01 1.35160565e+00
-1.21334389e-01 -8.14779580e-01 4.62970734e-01 3.25766057e-01
-7.06214011e-02 -5.72004497e-01 -5.31047523e-01 -8.43058586e-01
-1.91383027e-02 -1.99765891e-01 8.85565817e-01 5.75926542e-01
-1.29336143e+00 -1.44060358e-01 -6.76580191e-01 3.74628484e-01
1.25897372e+00 2.02487528e-01 8.28798473e-01 -6.22869372e-01
-4.35340613e-01 2.24645380e-02 -3.90987873e-01 -9.68371868e-01
2.60778159e-01 -4.92508143e-01 2.86153674e-01 -1.36965144e+00
2.61730313e-01 -4.12631124e-01 -4.25185561e-01 1.27092551e-03
-2.79995084e-01 7.50760138e-01 8.53622258e-02 -1.32220583e-02
-2.03927264e-01 5.49569428e-01 1.66011500e+00 -1.70026302e-01
-1.27324328e-01 -3.13103706e-01 -6.69464231e-01 5.70491910e-01
7.19280958e-01 -3.60592633e-01 -6.71423376e-01 -4.37560111e-01
2.22740710e-01 1.67598844e-01 6.02844357e-01 -1.41759276e+00
1.65613681e-01 -1.99207485e-01 9.69898582e-01 -7.07826555e-01
2.75066495e-01 -9.02530551e-01 7.06835151e-01 2.99236268e-01
-1.54874831e-01 -7.24883452e-02 2.67374277e-01 7.32028127e-01
-2.83675104e-01 -1.69136319e-02 1.31210339e+00 -1.44783288e-01
-9.49279070e-01 3.45078975e-01 2.64058262e-01 1.90534666e-02
1.10848689e+00 -3.36893499e-01 -6.48367941e-01 -4.51019436e-01
-2.41396010e-01 -1.70316324e-01 7.25472271e-01 4.21639234e-01
9.42284942e-01 -1.43459606e+00 -5.97764432e-01 3.43757242e-01
-9.53035951e-02 1.94989413e-01 7.79353082e-01 2.90824503e-01
-9.95197058e-01 -3.00527275e-01 -6.78515255e-01 -4.54706758e-01
-1.02063394e+00 7.87685156e-01 7.55634487e-01 -3.15883398e-01
-9.42969203e-01 3.32761437e-01 3.12883556e-01 -2.39480942e-01
-8.66744816e-02 3.79236042e-02 -2.36129318e-03 -4.89755094e-01
7.55775571e-01 2.62856513e-01 -1.45337075e-01 -5.98541319e-01
-5.96301928e-02 1.05772960e+00 1.15866579e-01 -2.26235181e-01
1.31259739e+00 -4.26373094e-01 -1.63523443e-02 2.55270332e-01
1.52319312e+00 5.46493270e-02 -1.64361644e+00 -1.86701361e-02
-9.67399836e-01 -1.04908693e+00 9.07603949e-02 -9.45262074e-01
-1.38178051e+00 6.14088058e-01 1.25044990e+00 -1.06203280e-01
1.80505037e+00 -6.34266853e-01 1.13110864e+00 1.49925098e-01
2.69251645e-01 -8.99196386e-01 2.49752045e-01 9.13979933e-02
8.32986832e-01 -1.33257627e+00 1.18856676e-01 -3.47347915e-01
-3.14647347e-01 9.83761013e-01 9.04119372e-01 -3.55381280e-01
3.82249475e-01 4.02987450e-01 2.45282963e-01 1.78317785e-01
-3.42691123e-01 6.70828223e-02 7.74483904e-02 7.53156543e-01
-2.33647563e-02 -4.49631423e-01 -1.12713866e-01 1.64394766e-01
-1.11161418e-01 1.54030725e-01 9.25864398e-01 5.49727499e-01
-1.73709452e-01 -8.12040925e-01 -6.21361852e-01 1.07391797e-01
-6.56810403e-01 -3.03216409e-02 3.32317382e-01 8.31439793e-01
2.92894870e-01 8.62270236e-01 1.35867447e-01 -3.89064610e-01
5.92447519e-01 -3.96770716e-01 3.93932998e-01 6.02570735e-02
-3.48581731e-01 -9.10591781e-02 -5.00330389e-01 -7.69601107e-01
-3.33839417e-01 -1.76357906e-02 -1.05599082e+00 -4.93191153e-01
-1.47644356e-01 -3.20377678e-01 9.00005937e-01 5.27702034e-01
1.19887888e-01 9.57641125e-01 9.04024184e-01 -8.44272017e-01
-3.48710775e-01 -7.01963782e-01 -6.03518903e-01 7.74531960e-01
5.51936269e-01 -4.37579870e-01 -6.51043534e-01 1.42275095e-01] | [10.899666786193848, -2.2002456188201904] |
c7086fb6-6e08-480b-b99e-62743cfc57c7 | hifi-hierarchical-feature-integration-for | null | null | http://kaizhao.net/hifi | https://arxiv.org/abs/1801.01849 | Hifi: Hierarchical feature integration for skeleton detection | In natural images, the scales (thickness) of object skeletons may dramatically vary among objects and object parts. Thus, robust skeleton detection requires powerful multi-scale feature integration ability. To address this issue, we present a new convolutional neural network (CNN) architecture by introducing a novel hierarchical feature integration mechanism, named Hi-Fi, to address the object skeleton detection problem. The proposed CNN-based approach intrinsically captures high-level semantics from deeper layers, as well as low-level details from shallower layers. By hierarchically integrating different CNN feature levels with bidirectional guidance, our approach (1) enables mutual refinement across features of different levels, and (2) possesses the strong ability to capture both rich object context and high-resolution details. Experimental results show that our method significantly outperforms the state-of-the-art methods in terms of effectively fusing features from very different scales, as evidenced by a considerable performance improvement on several benchmarks. | ['Ming-Ming Cheng', 'Dandan Li', 'ShangHua Gao', 'Wei Shen', 'Kai Zhao'] | 2018-07-01 | null | null | null | null | ['object-skeleton-detection'] | ['computer-vision'] | [ 1.08615063e-01 -1.67651847e-01 -1.99541718e-01 -2.88997352e-01
-5.63430309e-01 -7.65684769e-02 3.60308409e-01 9.78726000e-02
-2.87650168e-01 4.01140958e-01 2.62134790e-01 4.44623291e-01
-1.14562780e-01 -1.04145312e+00 -5.82154751e-01 -5.53133130e-01
-1.34972066e-01 5.56151830e-02 1.12708473e+00 -1.89497411e-01
4.80284207e-02 7.65403032e-01 -1.67369032e+00 4.73057806e-01
5.60325205e-01 1.33856332e+00 1.11663856e-01 2.70117432e-01
-2.42484599e-01 5.29683650e-01 -1.74923271e-01 -2.86408573e-01
3.33680212e-01 -1.56655923e-01 -7.37510443e-01 2.07602143e-01
7.23826706e-01 -3.28801274e-01 -2.40596965e-01 1.09164047e+00
2.93801308e-01 -1.82978779e-01 3.92348886e-01 -8.18404257e-01
-4.81825829e-01 4.94054466e-01 -9.83824253e-01 1.15752511e-01
1.79182906e-02 1.47248641e-01 1.13540173e+00 -1.02267051e+00
4.91410911e-01 1.38043249e+00 6.88287675e-01 4.48667198e-01
-1.14278853e+00 -7.10578382e-01 3.37510526e-01 5.78528158e-02
-1.47195864e+00 -1.09061643e-01 1.01869071e+00 -3.47602129e-01
5.02332568e-01 -2.26704646e-02 9.25544918e-01 7.61736929e-01
1.26425549e-01 8.61504555e-01 9.63303924e-01 -5.15720882e-02
-5.24130873e-02 -4.52374220e-01 2.88935304e-02 1.08771443e+00
3.83445144e-01 -1.27759308e-01 -7.73174226e-01 -1.30786419e-01
1.37866271e+00 2.35366493e-01 -1.81505859e-01 -5.24857998e-01
-1.55471647e+00 7.27372348e-01 8.97329092e-01 4.44074780e-01
-4.31820154e-01 3.92191947e-01 4.26389098e-01 -1.15092501e-01
8.20365772e-02 1.61702260e-01 -7.15099156e-01 3.44283402e-01
-7.49075949e-01 1.38489962e-01 3.70374769e-01 7.77779281e-01
9.40147877e-01 3.15893628e-02 -1.34101182e-01 8.98638248e-01
3.01437020e-01 5.69282733e-02 5.22314131e-01 -1.01106000e+00
3.73095512e-01 1.04245079e+00 -2.51005441e-01 -1.04215300e+00
-5.84798872e-01 -6.31892562e-01 -1.12229764e+00 4.33217824e-01
3.89135569e-01 2.24047571e-01 -1.02818227e+00 1.70542061e+00
7.23852992e-01 2.25246251e-02 -1.22174792e-01 9.60568547e-01
1.20034599e+00 2.88365930e-01 1.85966771e-02 1.53741106e-01
1.73937774e+00 -1.01891756e+00 -3.17094833e-01 -1.24622114e-01
1.44942582e-01 -7.04585493e-01 1.05290389e+00 4.75641452e-02
-1.30970144e+00 -8.71793807e-01 -1.16594303e+00 -3.13903779e-01
-2.41170898e-01 2.57344544e-01 7.96468794e-01 1.25260085e-01
-9.28730547e-01 6.74816787e-01 -1.03525007e+00 -2.93638408e-01
8.10972273e-01 5.68640947e-01 -6.01965904e-01 -9.89505369e-03
-8.76746893e-01 2.77752876e-01 4.85766798e-01 1.12798803e-01
-6.50286436e-01 -5.83609045e-01 -1.04725492e+00 1.82632565e-01
3.95882607e-01 -1.16470897e+00 1.04158819e+00 -8.54342997e-01
-1.34815097e+00 7.02271879e-01 5.05128354e-02 6.93597039e-03
5.75941324e-01 -3.95813376e-01 7.49266818e-02 4.53747451e-01
2.23658189e-01 8.79034758e-01 7.78677344e-01 -1.22470987e+00
-9.29143965e-01 -6.57442093e-01 1.94078550e-01 1.09370038e-01
-6.32463515e-01 1.78761095e-01 -7.87063777e-01 -1.01146901e+00
5.48547029e-01 -4.62611377e-01 -3.92319292e-01 6.57206535e-01
-3.42353910e-01 -3.11209351e-01 9.85088646e-01 -1.75155476e-01
1.02090561e+00 -2.04597306e+00 2.55624831e-01 -1.52845532e-02
4.96685475e-01 8.15196112e-02 -9.49678496e-02 -1.84195135e-02
1.70177698e-01 -1.49734681e-02 -3.88597220e-01 -2.48864144e-01
-2.85870641e-01 2.54022449e-01 3.35761636e-01 4.07522976e-01
5.93552709e-01 1.03940403e+00 -8.77245307e-01 -9.52920854e-01
2.17187822e-01 6.47312820e-01 -5.46963930e-01 7.28702024e-02
3.57312970e-02 3.53548348e-01 -7.59513855e-01 1.08481550e+00
4.56502885e-01 -6.28220260e-01 -8.86118263e-02 -6.51986778e-01
-9.96883214e-02 -1.40223175e-01 -1.28845811e+00 1.85390913e+00
-2.21227869e-01 3.51097494e-01 2.84421474e-01 -6.79777920e-01
7.26017535e-01 1.29070044e-01 6.87016547e-01 -4.69624460e-01
9.69728678e-02 2.06077412e-01 -2.85609476e-02 -2.63564587e-01
2.05441102e-01 -5.05195931e-03 4.53669392e-02 6.51550889e-02
1.71631798e-01 1.43104225e-01 7.60056749e-02 -6.82565523e-03
1.06893945e+00 4.78067845e-02 5.32742560e-01 -3.33325565e-01
7.01400459e-01 -4.67347711e-01 9.21032786e-01 5.40845752e-01
-2.73591787e-01 8.00859451e-01 3.05011839e-01 -6.62614942e-01
-7.80831456e-01 -1.03116703e+00 -2.47053608e-01 1.04411399e+00
4.82630223e-01 -5.42571306e-01 -5.29520869e-01 -7.92160571e-01
1.02506489e-01 -5.13316989e-01 -1.04626262e+00 5.91864139e-02
-9.74901676e-01 -5.51034391e-01 3.12748730e-01 1.09546041e+00
8.66590142e-01 -1.11957836e+00 -7.61750162e-01 2.63626218e-01
-8.90871063e-02 -1.32680261e+00 -4.31082100e-01 1.84731692e-01
-1.23757052e+00 -1.05070913e+00 -9.51970041e-01 -9.57422674e-01
5.72168767e-01 3.65046382e-01 1.14680398e+00 4.93644267e-01
-6.58477008e-01 1.98963359e-02 -3.68216336e-01 -7.61971697e-02
1.26784787e-01 2.05241516e-01 -1.76326990e-01 1.57967016e-01
-1.93838045e-01 -8.06691229e-01 -9.01831806e-01 4.88511503e-01
-1.17310047e+00 2.08265826e-01 9.63576496e-01 7.61119783e-01
8.11730981e-01 1.42491326e-01 5.38717270e-01 -6.12420619e-01
4.09649797e-02 -9.58088338e-02 -3.90414923e-01 1.55717343e-01
-1.08507387e-01 1.45606063e-02 5.09274602e-01 -4.10530001e-01
-9.11972702e-01 3.30592394e-01 -1.49979860e-01 -2.96316803e-01
-3.76363248e-01 2.59936899e-01 -4.26449478e-01 -3.98955822e-01
4.07241166e-01 1.11939222e-01 -1.97607636e-01 -7.15722680e-01
2.88340449e-01 2.17181563e-01 7.92646229e-01 -7.77909100e-01
9.44466054e-01 8.68195236e-01 2.72911519e-01 -7.55431354e-01
-1.04715121e+00 -5.55776775e-01 -1.15763497e+00 -1.79709420e-01
9.29648042e-01 -9.14979100e-01 -3.61305624e-01 7.43755460e-01
-1.14718759e+00 -1.07162774e-01 -3.56054097e-01 3.36810172e-01
-5.38469732e-01 4.74867016e-01 -1.03467298e+00 -2.00778574e-01
-3.62657011e-01 -1.24373984e+00 1.45928478e+00 5.26440501e-01
1.03222549e-01 -7.40455687e-01 -1.70515120e-01 3.47283065e-01
2.28932828e-01 6.93323731e-01 8.81963730e-01 -2.64194876e-01
-7.94986427e-01 -1.14805803e-01 -6.97742820e-01 1.02952667e-01
3.37745279e-01 1.83504328e-01 -7.41864562e-01 -2.50638932e-01
-2.64857352e-01 -4.60139334e-01 1.13201308e+00 3.26350570e-01
1.37709939e+00 -1.69129837e-02 -3.16260487e-01 7.57846534e-01
1.38784420e+00 -4.71085310e-01 4.42368031e-01 4.33119535e-01
1.00875962e+00 5.86103201e-01 4.58800226e-01 4.15567428e-01
3.32969666e-01 7.59100676e-01 7.42870688e-01 -5.66689491e-01
-4.67205197e-01 1.00532874e-01 1.29513755e-01 6.85268939e-01
-3.57358336e-01 5.29349208e-01 -5.94514012e-01 4.60790306e-01
-1.94842005e+00 -6.45657003e-01 -2.73428708e-01 1.79447889e+00
9.22980487e-01 3.27019602e-01 3.40446919e-01 1.40976778e-03
6.42138779e-01 2.39202723e-01 -6.49915934e-01 2.85020798e-01
-1.80296451e-01 1.87796608e-01 2.45220780e-01 -6.49955794e-02
-1.34725320e+00 9.10448968e-01 6.35222435e+00 9.09247875e-01
-8.72862875e-01 -2.38204300e-02 2.78396577e-01 1.17061779e-01
1.40881896e-01 -2.66279548e-01 -8.03465188e-01 1.52176887e-01
3.98109704e-02 1.34390905e-01 -2.84480065e-01 9.31589365e-01
-1.81290418e-01 2.03727022e-01 -1.01346791e+00 7.44538426e-01
-7.67117040e-03 -1.36684668e+00 2.53180146e-01 1.67549983e-01
8.07567239e-01 4.54054363e-02 -1.81804135e-01 1.27907582e-02
1.88762546e-01 -8.19132268e-01 8.93110335e-01 3.07655096e-01
6.14222705e-01 -8.34031940e-01 7.67064393e-01 2.12526381e-01
-1.96221769e+00 -2.42590293e-01 -4.57720429e-01 1.66368082e-01
-1.94269959e-02 6.17945671e-01 -4.03309539e-02 7.47795045e-01
1.04530001e+00 1.00657654e+00 -7.82115579e-01 1.19119358e+00
-8.04734975e-02 2.88390398e-01 -3.32187712e-01 3.56832802e-01
2.46534213e-01 2.36900598e-01 2.41563693e-01 1.37405622e+00
1.11856923e-01 1.45560130e-01 3.44004005e-01 7.07780719e-01
-2.16380998e-01 2.18445390e-01 -2.94174522e-01 3.28823686e-01
1.17919162e-01 1.55960405e+00 -1.18977225e+00 -4.41057265e-01
-7.10668206e-01 8.32884133e-01 5.54975152e-01 1.15696132e-01
-7.30259001e-01 -3.12835246e-01 8.14664960e-01 1.63082227e-01
5.53119481e-01 -3.23609948e-01 -3.12664956e-01 -1.24602354e+00
2.13734403e-01 -5.08068919e-01 6.49762154e-01 -4.95351851e-01
-1.35840821e+00 5.77394664e-01 -1.96568981e-01 -1.35192358e+00
3.54852289e-01 -5.94112515e-01 -4.82002348e-01 4.19278204e-01
-1.73080492e+00 -1.57379985e+00 -5.57772160e-01 8.39439392e-01
8.20570111e-01 1.21209718e-01 4.91614252e-01 3.52001399e-01
-4.82843220e-01 4.87070680e-01 -3.95815641e-01 4.67236578e-01
4.60275590e-01 -1.19915998e+00 1.59220278e-01 6.14852488e-01
1.67510122e-01 6.98139787e-01 1.06032930e-01 -5.39359331e-01
-1.01527190e+00 -1.25299942e+00 3.25888366e-01 -1.31361365e-01
6.25712931e-01 -5.85341193e-02 -1.08300591e+00 2.74816364e-01
-2.24369109e-01 5.54171085e-01 6.16951227e-01 -1.72876213e-02
-6.21469319e-01 -1.13926880e-01 -8.46073687e-01 4.51263994e-01
1.19903743e+00 -3.25265974e-01 -6.24722064e-01 6.22246191e-02
7.70003796e-01 -3.35119426e-01 -1.17167878e+00 1.03876758e+00
9.54967141e-01 -1.32285631e+00 1.30565023e+00 -5.26865721e-01
6.85231447e-01 -4.78979409e-01 -2.12149173e-01 -5.39062083e-01
-7.19287515e-01 -1.74240500e-01 -4.44918782e-01 1.07359087e+00
5.69184087e-02 -1.92881137e-01 8.76416922e-01 1.29676372e-01
-1.44917339e-01 -1.30153704e+00 -9.04859841e-01 -6.74216986e-01
-9.10726667e-04 -2.16407657e-01 5.94174325e-01 7.45691478e-01
-5.23192883e-01 1.75674722e-01 -1.10635333e-01 1.99001193e-01
7.29041159e-01 6.25420332e-01 7.99265087e-01 -1.58155513e+00
-1.90320864e-01 -8.54704499e-01 -8.72638106e-01 -1.17605925e+00
-1.26616314e-01 -6.79006219e-01 -2.07037680e-04 -1.56362498e+00
6.08717799e-01 -1.17069744e-01 -5.25507808e-01 6.81156814e-01
-4.36910778e-01 8.12398493e-01 7.17148036e-02 3.47980052e-01
-7.45021760e-01 6.68643177e-01 1.60675991e+00 9.60042141e-03
6.14261068e-02 -3.07668578e-02 -7.01288104e-01 1.33531189e+00
5.71293473e-01 -3.40347767e-01 4.06006537e-02 -4.03256148e-01
-1.07335776e-01 -2.58589506e-01 6.31014824e-01 -1.30607736e+00
2.01453418e-01 -2.56996214e-01 8.49935293e-01 -6.38710320e-01
3.40354443e-01 -8.21250379e-01 -1.20265074e-01 3.95991087e-01
-2.34387770e-01 -2.13168502e-01 -3.80557217e-03 7.01768100e-01
-3.47349316e-01 2.16243908e-01 1.10234165e+00 -4.11479026e-01
-6.01604104e-01 7.08222568e-01 6.59881392e-03 -1.59177467e-01
8.32072437e-01 -4.54108626e-01 -2.44200788e-02 2.28136718e-01
-7.84427822e-01 1.84405595e-01 7.39096403e-01 4.82161909e-01
7.83405602e-01 -1.50609708e+00 -5.11965215e-01 1.97321013e-01
3.85840356e-01 5.36855340e-01 1.33745149e-01 8.47452641e-01
-3.87486041e-01 3.33435200e-02 -3.99677992e-01 -8.67459357e-01
-1.26268542e+00 1.39516875e-01 2.87842244e-01 -3.53697360e-01
-1.00035131e+00 9.94988859e-01 6.77184880e-01 -2.14961246e-01
3.20453256e-01 -5.39088428e-01 -2.81130850e-01 -3.58363278e-02
5.32823086e-01 2.10926294e-01 -6.88405782e-02 -9.15283024e-01
-3.57339889e-01 1.22264850e+00 -1.06307380e-01 2.70486593e-01
1.35222232e+00 -1.36927683e-02 -2.03119412e-01 2.94813126e-01
1.23872161e+00 -1.04738675e-01 -1.50524485e+00 -6.24192595e-01
-4.89395037e-02 -5.93809247e-01 1.89065654e-02 -3.84819955e-01
-1.55002344e+00 8.45055819e-01 4.22830611e-01 -2.81472574e-03
1.18124795e+00 4.23300326e-01 1.14044523e+00 3.00987095e-01
4.40742731e-01 -9.72436190e-01 6.61185741e-01 3.13074619e-01
8.77806008e-01 -1.38248193e+00 2.63979554e-01 -5.89396477e-01
-2.21161485e-01 1.39254761e+00 1.10722911e+00 -1.57969892e-01
7.16602385e-01 2.80652374e-01 -2.62634392e-04 -6.61237061e-01
-4.54680204e-01 -6.08689725e-01 5.75258732e-01 4.33326066e-01
4.26047117e-01 -1.10001050e-01 -1.77005231e-01 7.22140849e-01
1.35286063e-01 -1.09411500e-01 -3.68803553e-02 1.11486852e+00
-6.99964285e-01 -9.84824002e-01 -4.07773256e-01 4.07031000e-01
-6.58412158e-01 1.69799805e-01 -4.90864158e-01 8.99479687e-01
4.11089212e-01 4.36479867e-01 -1.22211076e-01 -2.91638583e-01
3.13273191e-01 -4.57487255e-01 3.88787806e-01 -7.52542436e-01
-6.04232609e-01 3.38582665e-01 -3.94961655e-01 -1.03786802e+00
-7.66069114e-01 -5.65678716e-01 -1.46390057e+00 3.07489365e-01
-5.08464515e-01 -2.70799190e-01 3.98925304e-01 8.80274713e-01
1.85977921e-01 8.23032498e-01 4.58475441e-01 -1.25142848e+00
-2.46330261e-01 -7.53922820e-01 -5.76800585e-01 3.20200235e-01
4.29848790e-01 -1.06463933e+00 -2.02360496e-01 8.88510421e-02] | [9.5739107131958, -0.6423471570014954] |
0e37c28a-407a-4445-8bdc-8dd9fe38a3b5 | a-novel-patent-similarity-measurement | 2303.16767 | null | https://arxiv.org/abs/2303.16767v1 | https://arxiv.org/pdf/2303.16767v1.pdf | A Novel Patent Similarity Measurement Methodology: Semantic Distance and Technological Distance | Measuring similarity between patents is an essential step to ensure novelty of innovation. However, a large number of methods of measuring the similarity between patents still rely on manual classification of patents by experts. Another body of research has proposed automated methods; nevertheless, most of it solely focuses on the semantic similarity of patents. In order to tackle these limitations, we propose a hybrid method for automatically measuring the similarity between patents, considering both semantic and technological similarities. We measure the semantic similarity based on patent texts using BERT, calculate the technological similarity with IPC codes using Jaccard similarity, and perform hybridization by assigning weights to the two similarity methods. Our evaluation result demonstrates that the proposed method outperforms the baseline that considers the semantic similarity only. | ['Deaho Seo', 'Zachary Schimke', 'Junwon Lee', 'Sanguk Gim', 'Cheonkam Jeong', 'Yongmin Yoo'] | 2023-03-23 | null | null | null | null | ['semantic-textual-similarity', 'semantic-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [ 8.93047825e-02 -2.34602898e-01 -4.29855496e-01 -3.78491431e-02
-5.18712401e-01 -1.21209192e+00 8.16421747e-01 3.28636467e-01
-1.48133725e-01 4.54600453e-01 9.74516571e-02 -6.23580694e-01
-6.28422379e-01 -8.56705546e-01 -1.39898106e-01 -1.76173031e-01
6.03276312e-01 4.88346927e-02 1.42778590e-01 1.81834940e-02
1.38944829e+00 4.76068884e-01 -1.79728496e+00 1.94006950e-01
1.44827068e+00 1.23221970e+00 4.79040831e-01 3.82230207e-02
-1.03496647e+00 4.78734463e-01 -8.47443819e-01 -4.97389674e-01
4.10478652e-01 -4.94538844e-01 -8.69002521e-01 -6.42820358e-01
4.05715317e-01 6.78640485e-01 2.96432704e-01 1.52173305e+00
3.42060387e-01 -2.12607667e-01 8.40770125e-01 -1.11494493e+00
-8.05424035e-01 5.49855590e-01 -1.19524680e-01 2.91525364e-01
6.62412226e-01 -4.53593969e-01 1.13204551e+00 -5.29195189e-01
5.58354199e-01 9.69634593e-01 6.89500988e-01 -1.92375124e-01
-7.81018674e-01 -9.06723440e-01 4.02408764e-02 2.18736932e-01
-1.30901587e+00 -1.27842382e-01 8.75442088e-01 -7.49736786e-01
7.54853308e-01 1.56933561e-01 4.58308190e-01 7.44730830e-01
1.42341688e-01 3.44875492e-02 1.47990108e+00 -3.68100613e-01
4.75105345e-01 2.53979355e-01 1.75988734e-01 7.01801628e-02
4.78389353e-01 -2.03036189e-01 -6.56289384e-02 -3.08222026e-01
5.84183514e-01 2.48077691e-01 -8.26761127e-02 -1.49232134e-01
-1.40505528e+00 7.26362348e-01 5.87578230e-02 1.33503020e+00
-3.48937571e-01 -3.93159360e-01 2.40313500e-01 7.82958627e-01
4.14936483e-01 1.44316459e+00 -6.03938699e-01 -3.78706068e-01
-1.18628657e+00 -1.82995498e-01 8.97025347e-01 1.07217586e+00
6.43593967e-01 -5.87139249e-01 -1.46235675e-01 3.51980090e-01
1.59695715e-01 3.49703789e-01 8.89285505e-01 -9.37259674e-01
3.41804415e-01 1.00011039e+00 3.24605480e-02 -1.43440998e+00
1.09195903e-01 -6.36495233e-01 -2.81554639e-01 1.44927442e-01
1.43609166e-01 1.13558777e-01 -6.64525270e-01 1.13784194e+00
-1.45424321e-01 3.59307557e-01 1.79056421e-01 2.72567272e-01
9.51241016e-01 3.39146227e-01 1.40451699e-01 -1.75582528e-01
1.15449321e+00 -7.10368335e-01 -7.51977682e-01 4.54227149e-01
6.90459251e-01 -1.53142440e+00 9.87907350e-01 2.94600546e-01
-8.35450292e-01 -8.06593597e-01 -9.27013159e-01 4.99221683e-01
-9.75628436e-01 1.13271512e-01 5.34285665e-01 8.04545283e-01
-8.62931550e-01 1.05832696e+00 -4.45592180e-02 -5.00775397e-01
2.92537808e-01 2.40238935e-01 -1.34375721e-01 2.74989694e-01
-1.34196758e+00 1.07690978e+00 6.12338066e-01 -8.24489474e-01
1.31594867e-01 -7.82768488e-01 -4.23867613e-01 3.33399445e-01
1.56486660e-01 -6.13641083e-01 7.18991220e-01 -1.16956520e+00
-1.63595808e+00 4.68258560e-01 9.27977636e-02 -3.72769423e-02
2.22657025e-01 1.84048578e-01 -9.51601803e-01 -7.72255659e-02
4.15746361e-01 3.90561283e-01 5.81720412e-01 -9.93034005e-01
-9.13942993e-01 -2.97217131e-01 -1.33456546e-03 4.08035032e-02
-6.62348688e-01 8.88174921e-02 -2.22129554e-01 -9.37331915e-01
1.61057055e-01 -6.42003298e-01 3.50696668e-02 -6.26333535e-01
-8.45786780e-02 -6.81459844e-01 5.72572529e-01 -4.41227347e-01
1.48817670e+00 -2.17070985e+00 -1.97362140e-01 6.79130912e-01
-3.21956575e-02 5.34636438e-01 -4.48532581e-01 4.11242813e-01
-2.04025745e-01 2.70899802e-01 -1.49175838e-01 4.26176131e-01
1.12983607e-01 -4.45902884e-01 -5.50637901e-01 -2.49696687e-01
-5.36779389e-02 9.50672925e-01 -1.04375732e+00 -5.46243072e-01
2.68460631e-01 2.40480363e-01 -1.14496641e-01 -5.04175052e-02
7.83965886e-02 3.68991017e-01 -9.17239130e-01 8.10729742e-01
5.43805182e-01 -2.16900140e-01 1.97763890e-01 -3.19433928e-01
-3.55244458e-01 4.07528847e-01 -1.12513793e+00 2.04898477e+00
-4.41112608e-01 2.70864904e-01 -6.67081594e-01 -1.11681569e+00
1.12662315e+00 2.64000207e-01 6.78205490e-01 -7.54648387e-01
2.16106102e-01 6.34692132e-01 1.93512551e-02 -4.36242968e-01
3.67890686e-01 3.05045396e-02 1.53758749e-01 2.54293740e-01
-1.78209379e-01 -5.06252825e-01 1.63150504e-01 -6.96128746e-03
1.22560704e+00 1.79049090e-01 4.86096263e-01 -2.59435564e-01
9.10752714e-01 7.46739134e-02 1.88675240e-01 5.75724125e-01
-1.70055971e-01 1.98700219e-01 2.44321778e-01 -9.93595272e-02
-7.34873474e-01 -8.19194853e-01 -1.50224730e-01 4.64212388e-01
4.28183287e-01 -4.70470130e-01 -8.45862448e-01 -1.07582021e+00
3.45951319e-01 5.51602364e-01 -2.70343125e-01 4.34738621e-02
1.53256193e-01 6.62777275e-02 3.50183189e-01 1.77892208e-01
4.26537097e-01 -7.11187363e-01 -1.43041655e-01 5.21965474e-02
-2.77558248e-03 -8.73805463e-01 -2.75043130e-01 -4.74691331e-01
-1.08482099e+00 -1.15154493e+00 -8.86414826e-01 -1.02323627e+00
4.41456914e-01 6.79861486e-01 1.02558160e+00 -9.03309658e-02
-4.22852859e-02 2.66302943e-01 -5.28638363e-01 -5.45835137e-01
-2.56473303e-01 3.82466167e-01 -3.73277277e-01 -3.61095488e-01
7.15780377e-01 -5.43759346e-01 -5.42470098e-01 5.32371759e-01
-7.84889758e-01 -5.33195257e-01 1.08125234e+00 1.69275090e-01
6.06185615e-01 5.99784613e-01 6.97340548e-01 -5.00177681e-01
1.17024601e+00 -4.38689649e-01 -5.84383249e-01 5.65444946e-01
-1.11939418e+00 2.75026053e-01 4.47018206e-01 -4.21886772e-01
-8.92598689e-01 -3.19358557e-01 1.64986804e-01 -2.21671853e-02
-3.63811046e-01 6.67833328e-01 -1.79769963e-01 -5.87753415e-01
1.45470187e-01 -1.03285603e-01 -1.37151435e-01 -5.18002570e-01
5.88608503e-01 1.16568184e+00 2.29858235e-01 -4.52382654e-01
7.02702820e-01 1.86518177e-01 1.38865277e-01 -3.95862848e-01
-7.12001622e-01 -8.08105826e-01 -5.03431678e-01 1.44008100e-01
6.54291868e-01 -5.49287081e-01 -8.08316529e-01 -1.59243047e-01
-1.17103267e+00 6.24029398e-01 -3.44637245e-01 1.00174201e+00
-2.79431373e-01 9.21984136e-01 1.46202296e-01 -4.30663913e-01
-3.36356729e-01 -1.03869331e+00 9.31184113e-01 9.53677669e-03
-5.40685892e-01 -1.04186606e+00 4.41287100e-01 5.14269233e-01
6.12799227e-01 1.19042648e-02 1.02707231e+00 -8.99342120e-01
-1.47570759e-01 -4.20356452e-01 -3.69948924e-01 3.99045706e-01
7.67944634e-01 -1.84404194e-01 -5.89199364e-01 2.55195741e-02
1.67106107e-01 4.60683852e-01 6.50222301e-01 3.09074730e-01
1.21873379e+00 1.10228799e-01 -5.57554960e-01 6.70784041e-02
1.40971708e+00 6.24906600e-01 8.17681372e-01 6.54993653e-01
5.72234988e-01 9.77024972e-01 8.69910598e-01 6.62234128e-02
9.10511985e-02 5.03314793e-01 1.18838675e-01 2.59072751e-01
1.00904569e-01 -8.01111758e-03 3.88490371e-02 1.10133410e+00
-4.42358047e-01 1.99101325e-02 -8.16248059e-01 2.25691974e-01
-1.76089132e+00 -1.04462838e+00 -7.92942196e-02 2.44115138e+00
5.35694897e-01 2.39158183e-01 1.93197392e-02 3.13576102e-01
1.11406469e+00 -2.80900478e-01 -1.60946026e-01 -2.39912868e-01
-2.17055678e-01 5.82939982e-01 5.76010883e-01 1.18058905e-01
-8.38883638e-01 1.04077792e+00 7.03812790e+00 1.16038501e+00
-1.01878846e+00 -1.37514770e-01 -1.22987650e-01 2.86404192e-01
-6.75955296e-01 3.56835157e-01 -2.65906751e-01 9.74637985e-01
8.10405731e-01 -3.17797989e-01 2.10511237e-01 6.66637301e-01
-1.62154779e-01 2.03604192e-01 -7.36636579e-01 1.12420762e+00
2.06953198e-01 -1.39799690e+00 4.07854348e-01 2.18422875e-01
8.06192040e-01 -4.58100438e-01 1.98425859e-01 -1.46238995e-03
1.85079530e-01 -7.62287736e-01 2.64765769e-01 8.06700051e-01
4.29437339e-01 -6.64800465e-01 8.84145975e-01 -2.34355107e-01
-1.35580981e+00 -5.49966358e-02 -3.26442420e-01 -2.43506446e-01
-2.51665890e-01 7.86798418e-01 -5.14330804e-01 1.19265091e+00
4.58701015e-01 1.12406707e+00 -6.21235371e-01 1.12120414e+00
-1.99037448e-01 1.81174994e-01 3.40227187e-01 -8.53211898e-03
2.50005275e-01 -6.37582004e-01 3.25688750e-01 8.13144267e-01
1.20274878e+00 -3.21454704e-01 -1.93464175e-01 8.26804101e-01
5.79815079e-03 6.38220191e-01 -8.48071635e-01 -8.01256955e-01
8.84450614e-01 1.13858008e+00 -7.80021191e-01 -5.26604950e-01
-6.47333443e-01 9.37658787e-01 -2.92050570e-01 -1.12103075e-02
-6.29771948e-01 -1.27237809e+00 6.03878617e-01 -1.27053291e-01
3.87679130e-01 -1.78744003e-01 -5.95430076e-01 -1.00708497e+00
6.44036680e-02 -6.15494072e-01 3.61786723e-01 -7.92061746e-01
-1.43058729e+00 3.86370093e-01 -4.79330748e-01 -1.77357733e+00
3.74386400e-01 -6.78555489e-01 -8.02843451e-01 9.01556492e-01
-1.59810841e+00 -6.20895088e-01 -3.53253335e-01 3.27530324e-01
2.24058375e-01 -5.33285737e-01 6.95181489e-01 3.55535656e-01
-1.28659680e-01 2.65202403e-01 2.28802532e-01 -4.54463720e-01
1.05652225e+00 -1.29057205e+00 4.72974509e-01 2.37187043e-01
3.40641141e-02 8.31426263e-01 4.75869119e-01 -9.41375852e-01
-9.07649338e-01 -7.45421290e-01 1.30593371e+00 -6.81899786e-01
1.09285426e+00 5.61045706e-01 -7.25439966e-01 -3.57349366e-02
4.71431911e-01 -7.56892920e-01 1.29137468e+00 -5.29664122e-02
-4.07843053e-01 -2.72889063e-02 -1.17803645e+00 3.40527624e-01
1.06661582e+00 -5.75953782e-01 -1.19196236e+00 1.87506050e-01
8.33988547e-01 4.25482392e-01 -1.67898941e+00 4.33143020e-01
7.12023556e-01 -6.43545806e-01 1.03221083e+00 -5.48358500e-01
2.61190295e-01 -7.90389836e-01 -5.15565053e-02 -1.32391524e+00
-5.69932818e-01 -3.47698182e-01 4.38667059e-01 1.42829406e+00
3.45676810e-01 -8.55654001e-01 5.79427958e-01 2.18726248e-01
-1.06173297e-02 -2.52084792e-01 -5.56406379e-01 -1.42684507e+00
5.88531904e-02 -9.99463424e-02 1.33825064e+00 1.58163798e+00
3.16362470e-01 3.00796717e-01 3.39109093e-01 -3.14058632e-01
3.82096201e-01 4.49699789e-01 4.16577846e-01 -2.15677619e+00
2.50949144e-01 -1.28188467e+00 -9.94403124e-01 -2.57753104e-01
4.63711292e-01 -1.13376963e+00 -5.21490216e-01 -1.73040497e+00
1.09710827e-01 -3.67519945e-01 -9.68445897e-01 1.30830323e-02
3.22477473e-03 -6.47426322e-02 -1.17612584e-02 6.01605475e-01
-5.69529474e-01 1.94553480e-01 1.22159398e+00 -2.43105382e-01
-2.46293366e-01 -1.22853994e-01 -1.23214161e+00 7.25958228e-01
1.01438141e+00 -3.75617653e-01 -4.23169911e-01 -1.24389753e-01
5.10742664e-01 -6.87697589e-01 -1.70810923e-01 -1.18803823e+00
3.29208642e-01 -4.09436703e-01 9.62100402e-02 -2.71530300e-01
-4.34343696e-01 -1.05383170e+00 3.48525882e-01 4.81370628e-01
-3.31732690e-01 2.63466775e-01 -1.53358012e-01 5.61126411e-01
-6.90204024e-01 -3.68469238e-01 1.62795916e-01 -1.80347219e-01
-5.85578561e-01 -2.36554705e-02 -1.27088889e-01 -1.79884568e-01
1.10788941e+00 -5.98161161e-01 -2.56222159e-01 1.88161314e-01
-2.80374885e-01 -1.00718893e-01 8.02123547e-01 6.91615939e-01
2.84466952e-01 -1.49491119e+00 -3.32657874e-01 -2.96563148e-01
5.11917353e-01 -7.98455715e-01 -3.18018287e-01 8.14677656e-01
-1.86738238e-01 9.15478349e-01 -3.69922638e-01 5.54344319e-02
-1.31463730e+00 8.51046562e-01 -1.03221744e-01 -6.90446571e-02
6.32978752e-02 2.54770815e-01 -1.65135320e-02 -2.93203980e-01
2.54596084e-01 -4.54726309e-01 -9.00620699e-01 3.91723901e-01
2.47918993e-01 6.97963059e-01 1.82981655e-01 -5.17107427e-01
-4.81176108e-01 1.37989712e+00 2.50239283e-01 -3.78808789e-02
9.75127518e-01 1.91367060e-01 -4.78538662e-01 1.31909132e-01
9.68849003e-01 1.65076420e-01 -1.11657880e-01 -1.92141041e-01
7.04088748e-01 -7.29582667e-01 8.21617022e-02 -8.84775877e-01
-7.99816728e-01 7.18845189e-01 7.21047103e-01 6.24792218e-01
9.39877808e-01 -1.56289726e-01 7.66479075e-01 7.12622821e-01
4.12516654e-01 -1.40826368e+00 -3.24570225e-03 2.90478855e-01
5.01721442e-01 -1.07791388e+00 -2.41082683e-01 -7.53586829e-01
-4.33473557e-01 1.18473077e+00 6.31563589e-02 1.40042931e-01
7.06397235e-01 -3.78785342e-01 -7.52181709e-02 -3.26738209e-01
1.36892647e-01 -4.41841215e-01 8.20752978e-01 4.62877065e-01
9.55299318e-01 4.11024094e-02 -1.04908609e+00 4.79625881e-01
-2.10532710e-01 1.06735423e-01 -1.32815884e-02 1.18884492e+00
-8.12844574e-01 -1.71577382e+00 -3.11771482e-01 4.38072473e-01
-5.06288230e-01 -3.94346386e-01 -1.19677842e+00 2.46641889e-01
4.90541786e-01 1.25090361e+00 -6.01277053e-02 -5.46318293e-01
6.23808563e-01 -9.81196165e-02 5.93610585e-01 -5.69078624e-01
-7.78320789e-01 -1.11584991e-01 -1.56186491e-01 -4.45299596e-01
-8.56341720e-01 -4.19890225e-01 -8.41349006e-01 -2.03349158e-01
-4.81655031e-01 8.52371752e-01 1.18264425e+00 1.04472876e+00
1.05005717e+00 5.83843827e-01 7.83327103e-01 -2.51234263e-01
-3.48773837e-01 -8.88193011e-01 -5.18713474e-01 7.20062375e-01
-3.91977876e-01 -8.35209668e-01 -5.44660389e-01 -2.54120409e-01] | [9.72999095916748, 8.370230674743652] |
fe0f46de-882d-4c3a-9a6c-4534ba5cbed7 | an-integrated-semantic-web-service-discovery | 1502.02840 | null | http://arxiv.org/abs/1502.02840v1 | http://arxiv.org/pdf/1502.02840v1.pdf | An Integrated Semantic Web Service Discovery and Composition Framework | In this paper we present a theoretical analysis of graph-based service
composition in terms of its dependency with service discovery. Driven by this
analysis we define a composition framework by means of integration with
fine-grained I/O service discovery that enables the generation of a graph-based
composition which contains the set of services that are semantically relevant
for an input-output request. The proposed framework also includes an optimal
composition search algorithm to extract the best composition from the graph
minimising the length and the number of services, and different graph
optimisations to improve the scalability of the system. A practical
implementation used for the empirical analysis is also provided. This analysis
proves the scalability and flexibility of our proposal and provides insights on
how integrated composition systems can be designed in order to achieve good
performance in real scenarios for the Web. | ['Pablo Rodriguez-Mier', 'Manuel Mucientes', 'Manuel Lama', 'Carlos Pedrinaci'] | 2015-02-10 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [ 6.69236630e-02 1.86420694e-01 1.40977979e-01 -4.07556891e-01
-1.35743082e-01 -6.75856769e-01 7.00834572e-01 2.89043576e-01
8.07134956e-02 1.64475635e-01 2.49673799e-01 -4.55954134e-01
-8.26606452e-01 -1.07726502e+00 -2.44067818e-01 -6.68714345e-01
-3.52108181e-01 7.20806658e-01 6.35391116e-01 -5.37720263e-01
7.75509998e-02 8.34980071e-01 -1.92166650e+00 2.86731541e-01
4.90655780e-01 1.20048261e+00 2.93116063e-01 4.51201499e-01
-5.86294830e-01 -4.89321388e-02 -5.13224244e-01 -3.81254315e-01
4.19772863e-01 -7.52568781e-01 -6.95751846e-01 4.85599756e-01
-3.37865353e-01 1.38521671e-01 3.95136982e-01 9.20366049e-01
4.06498373e-01 9.21214297e-02 3.41702580e-01 -1.62494051e+00
4.27298062e-03 8.97059321e-01 2.61928141e-01 -1.16517752e-01
6.84769213e-01 8.44853744e-02 1.32309520e+00 -4.23390508e-01
9.31267262e-01 9.68295991e-01 -7.72036538e-02 4.82369781e-01
-1.29642117e+00 -1.12051807e-01 2.07141399e-01 4.17332083e-01
-1.30922842e+00 -4.88950104e-01 4.86613899e-01 -6.42959177e-02
9.79941666e-01 8.22547734e-01 9.56149936e-01 4.26119804e-01
-9.21356454e-02 1.12889446e-01 8.10658395e-01 -7.53035247e-01
6.48804963e-01 9.62485820e-02 -5.98223470e-02 4.22931731e-01
3.85796964e-01 -1.61677659e-01 -1.73393071e-01 -4.56223011e-01
2.69567341e-01 -1.44996107e-01 -2.19342038e-01 -8.09293568e-01
-6.71376705e-01 4.84151006e-01 1.62323758e-01 8.55703354e-01
-4.83137250e-01 -6.39188453e-04 4.40528303e-01 5.18796682e-01
-9.84454378e-02 1.75940394e-01 -4.18986440e-01 -4.68906127e-02
-5.52437603e-01 3.72211725e-01 1.62157393e+00 1.05954599e+00
5.64291298e-01 3.85631882e-02 2.98897892e-01 4.34899688e-01
5.23447871e-01 2.40053043e-01 1.93895236e-01 -5.50016999e-01
-1.36680007e-01 1.12308252e+00 8.39187503e-02 -6.24233663e-01
-4.69056010e-01 -3.79640281e-01 -3.07658873e-02 2.96555161e-01
2.19777286e-01 2.35174552e-01 -2.80758590e-01 1.30791187e+00
6.38429701e-01 -3.80275697e-02 1.50577828e-01 1.05444288e+00
3.13142091e-01 2.51614809e-01 -5.41560911e-02 -3.31124723e-01
1.58149815e+00 -6.29991353e-01 -5.68980932e-01 6.65695012e-01
4.40601170e-01 -5.54033756e-01 9.65602458e-01 2.54367560e-01
-9.51592803e-01 1.28261238e-01 -9.98265505e-01 6.71046674e-01
-3.59880328e-01 -5.56378663e-01 4.06500369e-01 8.48013043e-01
-1.14367628e+00 4.26857740e-01 -4.87133175e-01 -8.20427418e-01
-3.56008649e-01 5.55367649e-01 -3.60349476e-01 1.18952505e-01
-7.48017430e-01 6.95084929e-01 6.02396071e-01 -1.10318489e-01
-8.80784571e-01 1.00244969e-01 -5.65395236e-01 6.54834867e-01
8.40819180e-01 -7.71431625e-01 1.20663702e+00 -1.50658345e+00
-1.44409013e+00 4.70736712e-01 9.15554259e-03 -6.23978376e-01
3.15516680e-01 7.09997058e-01 -8.05056810e-01 2.27003738e-01
-2.60640562e-01 -3.56738448e-01 5.29482126e-01 -1.32949865e+00
-1.08137763e+00 -4.67787564e-01 3.56666833e-01 1.03201762e-01
-2.50198632e-01 3.49815756e-01 -5.16543806e-01 2.18627788e-03
2.92534858e-01 -7.57485926e-01 -4.03437823e-01 -6.55887783e-01
6.58311509e-03 -2.11657226e-01 3.77527922e-01 -3.63625973e-01
1.27715504e+00 -1.82889605e+00 3.82863969e-01 9.68791246e-01
-4.87614982e-02 -1.61354557e-01 -1.91390291e-01 1.24581563e+00
1.90731406e-01 1.55725136e-01 6.07710369e-02 1.87426746e-01
4.48751003e-01 5.46385109e-01 1.89008787e-01 2.82767087e-01
-3.26240867e-01 4.62014228e-01 -5.94630122e-01 -1.77331656e-01
7.16792420e-02 2.73702353e-01 -4.84919935e-01 4.46975321e-01
-7.69273758e-01 3.25245410e-01 -5.16832411e-01 5.88862300e-01
3.48629594e-01 -1.10328019e-01 1.01184678e+00 -5.00240624e-02
-2.88237631e-01 3.68586898e-01 -1.52474582e+00 1.39349151e+00
-6.82710230e-01 -4.11450982e-01 4.77359444e-01 -1.00764322e+00
1.13946152e+00 8.02819610e-01 8.58524084e-01 -7.23310292e-01
2.78610080e-01 5.12319982e-01 3.21063846e-01 -4.84360456e-01
-9.27836224e-02 -6.03234023e-03 4.79216054e-02 6.91485643e-01
6.63542300e-02 2.10868582e-01 6.02160394e-01 1.67079736e-02
1.08056569e+00 3.19183469e-01 4.20118362e-01 -5.74939489e-01
1.16445625e+00 -2.05911323e-01 3.00630063e-01 1.87144384e-01
2.27801710e-01 -1.76369935e-01 6.38581872e-01 -4.80957508e-01
-1.00520575e+00 -6.89865768e-01 3.95439863e-01 1.07143199e+00
2.77014315e-01 -7.94416070e-01 -9.09428000e-01 -6.34942472e-01
-8.11131820e-02 6.36813760e-01 2.40161475e-02 2.34852731e-01
-4.47860330e-01 -4.16073591e-01 -2.14958303e-02 -3.13604653e-01
-2.10675418e-01 -1.16434407e+00 -7.80301213e-01 4.16500568e-01
1.33224800e-01 -1.18164384e+00 -1.46262208e-02 -2.30511017e-02
-6.76070273e-01 -1.21626019e+00 2.90658921e-02 -4.29280549e-01
5.77835917e-01 -2.31599156e-02 1.00693369e+00 4.73374546e-01
6.56631961e-02 5.06867409e-01 -7.86088526e-01 1.43406287e-01
-8.24428618e-01 1.76401109e-01 -1.98227018e-01 5.93707561e-01
8.82441998e-02 -1.01455629e+00 -5.14970124e-01 4.65550512e-01
-1.30031919e+00 4.31496277e-03 3.07938248e-01 2.87216276e-01
4.57266539e-01 3.81967485e-01 6.35960817e-01 -6.26558959e-01
8.30464661e-01 -5.61668754e-01 -1.09889245e+00 3.18610579e-01
-1.23157465e+00 3.41591209e-01 8.15113127e-01 4.94158491e-02
-8.64736497e-01 -2.87275631e-02 -5.16999841e-01 4.15587813e-01
-2.18075037e-01 3.90539616e-01 -7.10381925e-01 -2.54949123e-01
2.81207860e-01 1.17325418e-01 -2.44772527e-02 -9.35213149e-01
5.01166224e-01 5.60104489e-01 -8.56022462e-02 -7.11022198e-01
6.32391393e-01 5.50055444e-01 2.82592356e-01 -4.31446344e-01
3.06187809e-01 -5.91000795e-01 -2.66815126e-01 -3.48982304e-01
2.62777269e-01 -1.49971262e-01 -1.29013515e+00 -3.46411288e-01
-6.40396118e-01 9.94655490e-02 -3.68321270e-01 3.80246155e-02
-7.69421160e-01 1.98976398e-01 -1.28963664e-01 -1.02173340e+00
-5.85822463e-01 -1.34730339e+00 7.15677202e-01 -8.08565225e-03
-1.67801678e-01 -6.83772802e-01 -2.02260524e-01 4.91349772e-03
5.97868741e-01 4.69099618e-02 1.00360894e+00 -1.20093548e+00
-9.54529762e-01 -1.35576785e-01 4.06987742e-02 5.49652837e-02
-7.37725431e-03 -5.35121597e-02 -1.29400447e-01 -2.39002794e-01
-2.14674860e-01 7.69606113e-01 -9.73003507e-02 -2.22904071e-01
5.06534398e-01 -5.74377179e-01 -2.76782244e-01 4.60783958e-01
2.09282732e+00 4.82690513e-01 4.28873599e-01 5.08139789e-01
3.02276045e-01 7.98511863e-01 4.76842046e-01 5.78176618e-01
2.55771399e-01 1.30013025e+00 8.32549095e-01 4.03033406e-01
-1.37885600e-01 3.08001135e-02 2.58263558e-01 6.76775515e-01
-1.68044269e-01 -2.94699579e-01 -8.52844417e-01 5.47856927e-01
-1.99117029e+00 -6.53954864e-01 -1.78153053e-01 2.33352900e+00
4.55764271e-02 4.30476815e-02 6.35011256e-01 4.08835530e-01
5.72083056e-01 -2.63585299e-01 9.58927870e-02 -9.92338836e-01
1.91443756e-01 4.12154675e-01 5.12625337e-01 6.54062927e-01
-1.13320246e-01 5.77451110e-01 5.76316023e+00 2.90591866e-01
-8.83084059e-01 4.79831606e-01 -1.78968936e-01 9.80117545e-02
-8.45425725e-01 4.44532692e-01 -6.79803848e-01 3.61895800e-01
1.38971806e+00 -6.59644306e-01 8.44106197e-01 6.81522310e-01
4.80418444e-01 1.38884231e-01 -8.32059860e-01 3.15042555e-01
-1.65992260e-01 -1.22622120e+00 1.93499655e-01 1.84872583e-01
4.41777825e-01 -6.39441758e-02 -8.21609497e-01 -4.47036564e-01
1.40216663e-01 -4.61898983e-01 8.87769163e-01 3.33042055e-01
2.95806974e-01 -8.55657339e-01 6.69822156e-01 3.24419528e-01
-1.21493638e+00 -4.39140052e-01 3.62797230e-01 2.87969373e-02
4.66230333e-01 3.95109385e-01 -9.99485731e-01 1.11236477e+00
4.11776274e-01 -2.01737627e-01 -2.33375639e-01 1.08483195e+00
-1.83588892e-01 1.89227715e-01 -3.18940312e-01 -3.27589244e-01
8.60807523e-02 -6.73665285e-01 8.11665118e-01 1.19418204e+00
5.32741666e-01 -1.92926422e-01 2.56185889e-01 4.77810323e-01
2.61193424e-01 7.90179968e-01 -3.71397436e-01 -1.08877979e-01
3.56329650e-01 1.19876552e+00 -9.70653653e-01 -4.16442305e-01
-5.00215113e-01 6.93420470e-01 -1.07415803e-01 4.91736799e-01
-4.05797035e-01 -2.56038308e-02 9.68098462e-01 6.01312697e-01
4.11602825e-01 -2.14699641e-01 1.65517047e-01 -9.47529197e-01
1.55407280e-01 -1.12161195e+00 5.89209855e-01 -3.06582510e-01
-8.38971734e-01 9.29657936e-01 -1.05035029e-01 -8.02954614e-01
-5.59928775e-01 -1.31833941e-01 -8.82689729e-02 8.76368582e-01
-1.17496145e+00 -1.13378966e+00 -5.20784073e-02 6.16192937e-01
1.04534633e-01 -6.83618560e-02 1.10042667e+00 4.11380053e-01
-3.17944139e-02 5.70722558e-02 -2.02525780e-01 -5.76864243e-01
-1.33303821e-01 -1.05660534e+00 3.82241726e-01 9.97842669e-01
2.15688109e-01 3.30656171e-01 9.97876763e-01 -4.54635859e-01
-1.82396805e+00 -3.50072235e-01 1.09850419e+00 2.26956710e-01
8.06785464e-01 -3.58781844e-01 -4.95096415e-01 5.34391284e-01
6.60647899e-02 -2.25503102e-01 5.03408790e-01 -5.60744852e-02
-1.96594849e-01 -7.35734820e-01 -1.40537894e+00 6.99363053e-01
1.26929164e+00 -2.37164110e-01 -1.01590931e-01 1.95996553e-01
7.39767671e-01 1.52824432e-01 -1.10679781e+00 2.94908106e-01
6.59248173e-01 -1.22202635e+00 6.29310250e-01 -4.97861028e-01
-3.80671084e-01 -4.81870025e-01 -3.08215350e-01 -9.67957854e-01
-1.46461532e-01 -9.06237483e-01 -1.09812275e-01 1.12095797e+00
2.44790658e-01 -8.92454743e-01 4.70109254e-01 3.48021388e-01
9.93760899e-02 -5.89352667e-01 -9.93137002e-01 -8.31669748e-01
-8.46642375e-01 -3.40510756e-01 1.33415580e+00 3.25670034e-01
1.84665993e-01 6.01887926e-02 1.97421893e-01 3.42891335e-01
5.07755697e-01 5.57815194e-01 6.35379672e-01 -1.22903216e+00
-5.73254168e-01 -7.18420148e-01 -7.89897442e-01 -1.77476108e-01
-3.15467417e-02 -1.00684237e+00 -3.44832003e-01 -1.79528296e+00
-3.78154904e-01 -5.95887482e-01 -3.42879981e-01 1.82906285e-01
7.24166036e-01 -8.78087506e-02 4.44956750e-01 8.72130990e-02
-3.61683458e-01 -1.98326319e-01 9.27493274e-01 3.86751205e-01
-1.81915179e-01 3.58582467e-01 -5.72050869e-01 1.96025148e-01
6.03241801e-01 -4.98277694e-01 -4.34250861e-01 1.08337164e-01
3.48812282e-01 3.78148407e-01 -1.00180849e-01 -6.32447422e-01
-7.48246759e-02 -4.39324051e-01 -6.42955720e-01 3.36694926e-01
1.13627702e-01 -1.77631819e+00 1.14923906e+00 6.37149870e-01
-2.71432519e-01 1.76730692e-01 -3.93283308e-01 1.65298373e-01
-1.93400774e-02 -5.95934272e-01 3.20203274e-01 -1.15441121e-01
-8.16369712e-01 2.93644637e-01 -2.17063293e-01 -7.81587183e-01
1.17848599e+00 1.15074180e-01 1.54529795e-01 -3.25011939e-01
-1.16838408e+00 -1.03628255e-01 8.02440286e-01 5.32110512e-01
1.31784067e-01 -1.10248935e+00 -5.10106087e-01 2.33995840e-01
2.58556992e-01 -9.61043358e-01 -2.41432518e-01 6.00765347e-01
-6.33185148e-01 3.14075291e-01 -4.45961475e-01 -1.83654219e-01
-1.60306370e+00 8.80361497e-01 4.03910369e-01 -1.95323274e-01
-7.41401315e-01 -1.03141218e-01 -4.76949513e-01 -1.57014593e-01
2.03889325e-01 -1.28931791e-01 -1.69441476e-01 -1.72257677e-01
5.33544362e-01 2.35070080e-01 4.36602592e-01 -7.87424505e-01
-5.70974886e-01 1.24375522e-01 8.03861439e-01 -6.06785655e-01
1.26802063e+00 -3.95413190e-01 -7.09241688e-01 1.11277007e-01
8.28961968e-01 2.44692534e-01 -7.08305895e-01 -2.94777602e-02
6.64022684e-01 -5.37164032e-01 -2.86899388e-01 -8.01406085e-01
-1.02147424e+00 1.91789657e-01 3.85467976e-01 1.10240984e+00
1.46332383e+00 6.45960495e-02 4.70864952e-01 -3.32992971e-02
7.50572026e-01 -6.79380715e-01 -6.83016956e-01 -1.00784890e-01
7.66359985e-01 -2.46515587e-01 -2.75124758e-01 -6.83926225e-01
-2.37034038e-01 1.25721574e+00 -1.97872538e-02 1.30810719e-02
5.68589509e-01 5.52684009e-01 -1.30857497e-01 -4.25837398e-01
-9.35827374e-01 -8.90930057e-01 -5.64016402e-02 6.48663580e-01
-8.10320675e-03 4.89339560e-01 -1.47578061e+00 6.73013866e-01
1.56481758e-01 -8.14348757e-02 3.66695911e-01 7.96578705e-01
-4.65519190e-01 -2.00202990e+00 -2.79681176e-01 -9.64488760e-02
-3.75511110e-01 2.59649307e-01 -1.80509537e-01 4.81610656e-01
2.37575322e-01 1.07214546e+00 -2.07448706e-01 -1.93902642e-01
7.17631280e-01 3.93568605e-01 5.25947690e-01 -5.47416806e-01
-1.04288232e+00 2.88104534e-01 6.23634994e-01 -7.87930310e-01
-2.80092865e-01 -4.68311161e-01 -1.47048020e+00 -2.26662785e-01
-2.54010975e-01 7.01531947e-01 1.42425406e+00 1.01037288e+00
5.56348979e-01 6.53083503e-01 6.83237314e-01 -2.32888579e-01
-3.70216548e-01 -5.46073377e-01 -7.19953954e-01 6.40828550e-01
-2.53126860e-01 -2.83333331e-01 -1.84813529e-01 -5.86079843e-02] | [8.628315925598145, 6.960301876068115] |
9a55348c-11f7-4dd4-8d8d-15f9fa0714be | leveraging-explicit-lexico-logical-alignments | null | null | https://aclanthology.org/2022.acl-short.31 | https://aclanthology.org/2022.acl-short.31.pdf | Leveraging Explicit Lexico-logical Alignments in Text-to-SQL Parsing | Text-to-SQL aims to parse natural language questions into SQL queries, which is valuable in providing an easy interface to access large databases. Previous work has observed that leveraging lexico-logical alignments is very helpful to improve parsing performance. However, current attention-based approaches can only model such alignments at the token level and have unsatisfactory generalization capability. In this paper, we propose a new approach to leveraging explicit lexico-logical alignments. It first identifies possible phrase-level alignments and injects them as additional contexts to guide the parsing procedure. Experimental results on \textsc{Squall} show that our approach can make better use of such alignments and obtains an absolute improvement of 3.4% compared with the current state-of-the-art. | ['Kang Liu', 'Jun Zhao', 'Jinlong Li', 'Yaohan He', 'Chong Zhu', 'Shizhu He', 'Runxin Sun'] | null | null | null | null | acl-2022-5 | ['text-to-sql'] | ['computer-code'] | [-1.45244792e-01 5.48591204e-02 -4.84930426e-01 -6.96515381e-01
-1.18323648e+00 -6.22147560e-01 2.65344918e-01 5.92533410e-01
-3.60721648e-01 5.70857942e-01 2.84979433e-01 -1.01151288e+00
1.75875127e-01 -1.21932459e+00 -7.82188535e-01 1.41702339e-01
1.09801516e-01 7.53654480e-01 6.63182676e-01 -3.98420751e-01
3.19142371e-01 3.32685351e-01 -1.34906280e+00 7.82799423e-01
1.13535213e+00 7.89042830e-01 9.32132676e-02 5.12810349e-01
-1.05228972e+00 9.41213071e-01 -4.54356939e-01 -7.53609359e-01
-3.96425724e-02 -1.69069543e-01 -1.34046626e+00 -3.74371946e-01
2.13213801e-01 -1.84104115e-01 2.33046085e-01 9.59752977e-01
1.08852811e-01 -2.77678460e-01 -1.08508967e-01 -8.22531402e-01
-4.46227938e-01 9.31964993e-01 -3.81730616e-01 2.85057545e-01
7.61761665e-01 1.50547931e-02 1.58936858e+00 -7.29238093e-01
4.91558403e-01 1.18484402e+00 4.67033088e-01 3.09517384e-01
-1.07250273e+00 -2.54491597e-01 1.80788711e-01 3.16646516e-01
-1.07092857e+00 -2.71375924e-01 5.18845022e-01 5.23312464e-02
1.66921258e+00 6.67352915e-01 3.20404828e-01 4.53824252e-01
4.38225083e-02 8.67290795e-01 1.09212625e+00 -8.53306413e-01
6.65160120e-02 1.37728184e-01 6.60818040e-01 7.29950547e-01
1.59562320e-01 -4.00877655e-01 -6.06769264e-01 -3.14582825e-01
4.01186317e-01 -3.06445539e-01 1.80728510e-01 4.87551764e-02
-9.42248821e-01 8.98126841e-01 1.71032757e-01 3.55328053e-01
-4.86183554e-01 -3.10952812e-02 4.23490852e-01 2.72952467e-01
1.19317457e-01 7.24077106e-01 -6.67762935e-01 -5.21014512e-01
-5.88216424e-01 3.39605749e-01 1.19554293e+00 1.21172440e+00
7.24604785e-01 -2.74034977e-01 -5.55928163e-02 8.85999024e-01
1.40059248e-01 2.74229497e-01 3.25732440e-01 -9.92111385e-01
8.74981582e-01 9.88876343e-01 5.13751879e-02 -5.81745803e-01
-2.48417005e-01 -1.48836404e-01 -1.59788057e-01 -3.51390272e-01
3.88054460e-01 2.81578600e-02 -7.02565372e-01 1.32474887e+00
2.81785935e-01 -4.66553122e-01 1.18626475e-01 4.90350276e-01
5.35949826e-01 5.97405374e-01 7.57984444e-02 -6.95300326e-02
1.75690579e+00 -8.88242483e-01 -7.45927930e-01 -4.60999817e-01
7.90034592e-01 -8.63818467e-01 1.65649319e+00 3.37986708e-01
-1.40343773e+00 -3.48030031e-01 -8.06489825e-01 -1.31688118e-01
-4.11034226e-01 -1.93826109e-01 8.53908062e-01 8.76402318e-01
-8.03751886e-01 2.52011418e-01 -9.52909827e-01 -4.81626600e-01
4.38130386e-02 2.90291458e-01 -1.08565696e-01 -7.36966208e-02
-1.11027074e+00 7.73779631e-01 5.91100812e-01 -2.00588107e-01
-4.20943042e-03 -4.85461414e-01 -7.06722260e-01 1.51932105e-01
9.93511856e-01 -4.23641145e-01 1.41921949e+00 -1.55467212e-01
-1.45394373e+00 7.09753692e-01 -7.72802413e-01 -6.05464518e-01
4.73945402e-02 -4.63181376e-01 -1.77750424e-01 3.94527130e-02
1.72238424e-01 4.31775123e-01 -9.24499109e-02 -9.24820364e-01
-8.62822890e-01 -3.85384113e-01 5.05245686e-01 -2.50114240e-02
-4.43279564e-01 5.26863754e-01 -1.09919155e+00 -4.40203041e-01
2.87035078e-01 -6.59038007e-01 -3.70984018e-01 -5.67028224e-01
-4.58521962e-01 -5.80810845e-01 2.47257605e-01 -6.71879649e-01
1.76838040e+00 -1.65826499e+00 -2.97401726e-01 2.29932442e-01
-1.88387811e-01 3.05558264e-01 4.72263694e-02 8.26276898e-01
1.66569069e-01 5.77651680e-01 -2.33073041e-01 -1.53362066e-01
7.24751651e-02 5.63481331e-01 -4.47575212e-01 -5.34483552e-01
5.14109969e-01 1.14975631e+00 -6.24459445e-01 -7.99077690e-01
-5.64748496e-02 -2.16221020e-01 -9.63527799e-01 4.00610983e-01
-8.49408686e-01 -1.13208733e-01 -6.81428432e-01 8.81314695e-01
5.68720460e-01 -2.10654825e-01 5.81166267e-01 4.67507280e-02
-1.64918065e-01 1.01423419e+00 -1.00770926e+00 1.36857247e+00
-5.19686639e-01 -4.73880302e-03 -3.07756513e-01 -1.10503376e+00
9.57483947e-01 -3.94916944e-02 3.04422885e-01 -8.46326411e-01
-2.89309829e-01 3.11256021e-01 -1.18983112e-01 -8.29673648e-01
5.85313737e-01 5.45118786e-02 -3.46233755e-01 3.19653988e-01
-2.35331759e-01 -6.86532632e-02 4.11405027e-01 1.67806521e-01
1.14561760e+00 1.33278698e-01 4.56718832e-01 -2.06089094e-01
7.90219724e-01 3.80605221e-01 5.99865913e-01 8.86191905e-01
2.43498698e-01 2.44843066e-01 7.64704168e-01 -4.89756107e-01
-8.35512698e-01 -7.92472541e-01 -2.59295087e-02 1.32916760e+00
-2.90217280e-01 -8.80854666e-01 -8.39780450e-01 -9.51963007e-01
-8.23460221e-02 1.12638950e+00 -1.65487349e-01 2.21010670e-01
-1.06599057e+00 -7.60483682e-01 6.45801306e-01 9.11434174e-01
4.16253656e-01 -1.14669871e+00 -5.98190725e-01 5.54544449e-01
-5.41146815e-01 -1.44430351e+00 -1.13846071e-01 4.11332309e-01
-9.97648954e-01 -9.86321449e-01 1.80292353e-01 -6.83100104e-01
2.86899656e-01 -9.11967456e-02 1.43147147e+00 2.77866960e-01
-9.63814929e-02 -4.15434726e-02 -6.07083321e-01 -6.27960265e-01
-4.95969832e-01 5.21484077e-01 -5.81449747e-01 -4.45992500e-01
1.08725464e+00 -3.27794045e-01 -8.93595293e-02 2.37826735e-01
-9.62890446e-01 -1.79021910e-01 5.58118701e-01 8.32044184e-01
6.16358519e-01 -1.39116123e-01 6.19878650e-01 -1.45656943e+00
6.59515798e-01 -2.93784857e-01 -6.45473003e-01 6.46074474e-01
-9.16454971e-01 5.57546496e-01 9.37507451e-01 1.32348478e-01
-1.05382180e+00 -4.74017411e-02 -7.07100391e-01 2.52909452e-01
-3.64955455e-01 1.08928502e+00 -3.85250151e-01 2.12165534e-01
4.06545609e-01 1.36236519e-01 -2.07270652e-01 -6.29751861e-01
3.28897268e-01 6.78069293e-01 5.28501630e-01 -1.20547354e+00
5.95525265e-01 1.54902220e-01 -2.05413818e-01 -5.55571258e-01
-8.51332366e-01 -6.49585843e-01 -5.53147256e-01 2.87688851e-01
7.49548197e-01 -3.89968276e-01 -8.66277039e-01 6.44444972e-02
-1.19630456e+00 -1.37425318e-01 -1.65810600e-01 1.37486130e-01
-5.05718231e-01 6.05104506e-01 -6.50581062e-01 -7.12019742e-01
-3.41585547e-01 -1.10217988e+00 9.34402466e-01 1.75042786e-02
-3.68185431e-01 -8.10711324e-01 -1.10621624e-01 6.97476089e-01
4.89340365e-01 -9.68516320e-02 1.49965882e+00 -1.14039457e+00
-8.51192355e-01 -2.46511802e-01 -1.03381120e-01 1.28217921e-01
-3.26734297e-02 -9.44916904e-02 -5.79734206e-01 2.65388668e-01
2.70670354e-02 -3.49548191e-01 5.37736952e-01 -2.16692805e-01
1.50406837e+00 -4.49805737e-01 -1.42458797e-01 3.17105204e-01
1.59124994e+00 2.35084429e-01 8.46023679e-01 6.30560040e-01
2.56860435e-01 6.70792282e-01 9.34097290e-01 4.12703723e-01
8.88673723e-01 8.35660577e-01 5.20555735e-01 8.54545906e-02
2.98149258e-01 -3.64207238e-01 1.64903119e-01 9.16299462e-01
9.35179368e-02 -1.28581271e-01 -1.25543332e+00 6.98288620e-01
-1.75006437e+00 -6.53168142e-01 -3.06364059e-01 2.14899492e+00
1.23826587e+00 5.88736117e-01 6.31620502e-03 5.25104217e-02
2.93854833e-01 3.23539227e-02 -2.34621763e-01 -9.14582849e-01
4.79043461e-02 7.89066017e-01 4.93887156e-01 5.57512224e-01
-9.17486727e-01 1.33969963e+00 6.65948629e+00 6.84962928e-01
-9.04283583e-01 -2.16084450e-01 2.47593164e-01 3.14281076e-01
-5.52223265e-01 2.54368722e-01 -1.28061557e+00 2.35113591e-01
1.22322404e+00 -1.33134097e-01 3.03556710e-01 8.43189895e-01
4.99894358e-02 -2.14931250e-01 -1.09513831e+00 2.82402724e-01
-1.74079716e-01 -1.32168174e+00 2.54314303e-01 -6.29237890e-02
1.50335148e-01 -1.86958894e-01 -3.34198296e-01 3.96085143e-01
4.48592842e-01 -9.05371428e-01 4.26763862e-01 4.59841669e-01
3.35046172e-01 -7.86888421e-01 1.05591321e+00 5.09560585e-01
-1.17934811e+00 2.46345196e-02 -3.57484132e-01 2.07986087e-02
2.13787034e-01 3.95921260e-01 -1.01201224e+00 9.32396352e-01
8.75930309e-01 1.79561347e-01 -7.76549339e-01 7.23170340e-01
-1.94763735e-01 9.72054899e-01 -3.98855388e-01 -3.72695625e-01
2.85516530e-01 -1.19502597e-01 2.08734930e-01 1.43475449e+00
1.49394140e-01 2.33316123e-01 3.14925104e-01 6.96619451e-01
-3.75636220e-02 4.81660604e-01 -3.05814981e-01 -1.08440667e-01
6.37400985e-01 8.07325304e-01 -5.27880430e-01 -5.87962449e-01
-8.52147400e-01 6.61938369e-01 6.04325175e-01 5.34702390e-02
-7.81056702e-01 -5.90382576e-01 3.84312660e-01 2.34223336e-01
5.18407583e-01 -2.61018813e-01 -6.72308743e-01 -1.06673217e+00
6.98436618e-01 -1.31885684e+00 6.41792953e-01 -5.25214970e-01
-1.08159733e+00 5.52528858e-01 9.27559808e-02 -6.01963937e-01
-4.97432560e-01 -5.23792863e-01 -5.28447390e-01 1.00148976e+00
-1.56292152e+00 -1.11283326e+00 -1.36781812e-01 2.64127105e-01
6.07196033e-01 9.36509073e-02 1.15311980e+00 2.47442394e-01
-4.11625236e-01 7.85491288e-01 -2.38909841e-01 3.00314903e-01
5.69039822e-01 -1.48180509e+00 7.76971877e-01 9.67101872e-01
3.07917267e-01 1.16475797e+00 5.75767696e-01 -4.98278439e-01
-1.73757041e+00 -8.42961252e-01 1.55037975e+00 -4.25760329e-01
7.57985115e-01 -1.85330629e-01 -1.31597066e+00 7.69320369e-01
3.55337054e-01 -2.54041493e-01 8.72045398e-01 3.88747960e-01
-5.05229652e-01 -2.93888718e-01 -8.37339818e-01 5.14108598e-01
7.84599364e-01 -6.25018656e-01 -9.51723456e-01 2.07030490e-01
9.93809640e-01 -4.30920631e-01 -1.19099545e+00 5.29704392e-01
4.52862084e-01 -1.05079627e+00 9.52279925e-01 -1.03564394e+00
4.42915022e-01 -5.94694018e-02 -3.79300624e-01 -6.30116880e-01
1.31106675e-01 -5.62753379e-01 -2.97858901e-02 1.39919317e+00
7.21367776e-01 -7.97667265e-01 1.08775663e+00 9.06398356e-01
-1.88144878e-01 -1.12296629e+00 -6.39908791e-01 -8.24259818e-01
2.62625605e-01 -7.81639457e-01 9.74133790e-01 7.43133128e-01
3.35749686e-01 2.84041554e-01 2.30176132e-02 1.74170792e-01
2.71513999e-01 3.83651763e-01 8.52675021e-01 -9.82563198e-01
-4.11746293e-01 -4.40853268e-01 -1.09608965e-02 -1.46306241e+00
4.81485240e-02 -7.85615385e-01 9.97317024e-03 -1.59671879e+00
9.85106602e-02 -7.07344890e-01 -1.17882214e-01 6.94289088e-01
-4.61030275e-01 -2.70311236e-01 2.46174499e-01 -1.00722313e-01
-5.76635599e-01 -4.00462709e-02 6.13157928e-01 8.30720961e-02
-1.12139687e-01 1.85194388e-01 -9.82563913e-01 5.11006832e-01
8.83159935e-01 -4.94270086e-01 -4.48344648e-02 -6.14538431e-01
3.38748485e-01 3.04244757e-01 -6.17913529e-02 -7.32104063e-01
4.12403852e-01 -4.01390404e-01 -3.80096972e-01 -8.63986015e-01
6.57765754e-03 -6.94512367e-01 -2.64422178e-01 3.28485131e-01
-5.12347400e-01 5.22046387e-01 3.19239885e-01 1.97270170e-01
-6.66242003e-01 -6.84829116e-01 2.72619337e-01 -4.04798031e-01
-4.67588633e-01 -9.39022973e-02 -3.55349332e-01 4.30105835e-01
5.80326080e-01 1.12604260e-01 -3.67898911e-01 -2.40978211e-01
-4.30972844e-01 4.38418567e-01 3.29149157e-01 2.76675940e-01
4.94485646e-01 -9.03355300e-01 -5.78866422e-01 2.46018067e-01
1.95658997e-01 2.51836982e-02 -4.50441957e-01 5.21756589e-01
-7.68797576e-01 9.77840006e-01 4.68357950e-02 -4.67347205e-01
-1.34264219e+00 5.47019243e-01 8.55743289e-02 -9.07443404e-01
-2.45502934e-01 6.64357007e-01 -4.27923352e-01 -5.95844567e-01
3.21897209e-01 -7.72179484e-01 -1.24734558e-01 -2.70343274e-01
2.58644700e-01 -1.62185077e-02 4.28303987e-01 -9.40125734e-02
-4.86423403e-01 5.67406237e-01 -2.43483931e-01 -8.50404501e-02
1.22713554e+00 9.33162645e-02 -5.08341670e-01 2.66254634e-01
9.80088949e-01 4.52257991e-01 -3.82864535e-01 -5.16917288e-01
9.31059599e-01 -4.85170066e-01 -5.22653759e-01 -8.52342486e-01
-6.61260188e-01 1.06775999e+00 -1.58877105e-01 6.56933129e-01
1.18201900e+00 1.39323130e-01 1.07828379e+00 8.71585667e-01
5.56844175e-01 -9.31968093e-01 -4.02990431e-02 7.79159307e-01
5.88007927e-01 -1.19793296e+00 -9.94847864e-02 -7.95441389e-01
-5.86587489e-01 1.28742468e+00 7.41494179e-01 2.70643145e-01
3.15402925e-01 4.38155383e-01 2.12855004e-02 -1.50261432e-01
-9.34545159e-01 -3.85067493e-01 1.28616989e-01 2.34682053e-01
7.17331111e-01 -1.33341864e-01 -5.98080099e-01 7.55969584e-01
-4.14668590e-01 -7.90532306e-02 4.50595081e-01 1.40834558e+00
-6.78571284e-01 -1.90668643e+00 -2.91731358e-01 4.78879869e-01
-9.41413283e-01 -3.95502150e-01 -3.52497637e-01 8.20603192e-01
-2.31948927e-01 1.07905948e+00 -5.75435907e-02 -2.48253852e-01
6.91565394e-01 5.41348279e-01 2.20624000e-01 -8.42933178e-01
-9.13587332e-01 5.85337058e-02 5.98607779e-01 -7.50470698e-01
-2.74565905e-01 -6.40436172e-01 -1.43686521e+00 -1.64828226e-01
-2.35459551e-01 4.38054055e-01 4.40048307e-01 7.45989501e-01
3.96918625e-01 3.23239446e-01 4.60452586e-01 3.07682276e-01
-8.93418550e-01 -6.56855643e-01 1.55837327e-01 2.74358630e-01
-1.12979397e-01 -8.09547231e-02 8.82875770e-02 2.94377506e-02] | [9.95201587677002, 7.780074596405029] |
da00adea-1b2c-4f62-bb85-ed57c73152e8 | towards-improved-room-impulse-response | 2211.04473 | null | https://arxiv.org/abs/2211.04473v2 | https://arxiv.org/pdf/2211.04473v2.pdf | Towards Improved Room Impulse Response Estimation for Speech Recognition | We propose a novel approach for blind room impulse response (RIR) estimation systems in the context of a downstream application scenario, far-field automatic speech recognition (ASR). We first draw the connection between improved RIR estimation and improved ASR performance, as a means of evaluating neural RIR estimators. We then propose a generative adversarial network (GAN) based architecture that encodes RIR features from reverberant speech and constructs an RIR from the encoded features, and uses a novel energy decay relief loss to optimize for capturing energy-based properties of the input reverberant speech. We show that our model outperforms the state-of-the-art baselines on acoustic benchmarks (by 17\% on the energy decay relief and 22\% on an early-reflection energy metric), as well as in an ASR evaluation task (by 6.9\% in word error rate). | ['Paul Calamia', 'Dinesh Manocha', 'Pablo Hoffmann', 'Vamsi Krishna Ithapu', 'Ishwarya Ananthabhotla', 'Anton Ratnarajah'] | 2022-11-08 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 3.39620024e-01 -2.53621936e-01 6.47913992e-01 -2.45324045e-01
-1.95731282e+00 -4.83085185e-01 5.56609869e-01 -4.77847457e-01
-2.69336611e-01 4.65588301e-01 9.93343711e-01 -4.50505883e-01
1.87448263e-02 -3.13340187e-01 -7.45188475e-01 -8.70318711e-01
-1.93366840e-01 -2.39195704e-01 -3.93846393e-01 -2.86265492e-01
-1.85696661e-01 4.07559693e-01 -1.24535728e+00 2.40488335e-01
6.74895585e-01 1.17132115e+00 9.80036780e-02 1.23323071e+00
6.41310394e-01 1.00146103e+00 -1.01045203e+00 5.53423688e-02
1.86895624e-01 -6.19023204e-01 -3.84838700e-01 -5.19104481e-01
2.63429582e-01 -2.97619790e-01 -9.41430688e-01 6.82003558e-01
1.30054617e+00 8.13123405e-01 9.46885049e-01 -2.23040119e-01
-8.27700734e-01 5.95890582e-01 -5.10100760e-02 3.25560123e-01
4.55431193e-01 2.18058690e-01 9.57527339e-01 -1.14349723e+00
-1.79168642e-01 1.12369609e+00 7.78025031e-01 8.57580543e-01
-1.08081961e+00 -6.04323685e-01 -4.00941372e-02 8.81577805e-02
-1.37327886e+00 -1.21150255e+00 6.82786942e-01 -1.16163284e-01
1.40534139e+00 5.56384981e-01 -3.13970372e-02 1.47172737e+00
-2.34528109e-01 6.86418951e-01 1.08524489e+00 -6.58316910e-01
3.48189741e-01 -2.82901138e-01 3.99490669e-02 2.81270653e-01
-5.47269821e-01 6.89953804e-01 -5.77509344e-01 -5.60295908e-03
3.90153885e-01 -6.24021351e-01 -8.75267804e-01 6.74116135e-01
-9.20329809e-01 5.06083906e-01 7.01154232e-01 2.20200300e-01
-2.01965913e-01 5.79791963e-01 2.17993837e-02 2.52650917e-01
6.96609557e-01 5.26171386e-01 -2.06861004e-01 -2.17600867e-01
-8.71175885e-01 -1.04582578e-01 6.12053931e-01 5.24123788e-01
2.01492265e-01 8.42456162e-01 -6.44935548e-01 1.63447344e+00
7.08020866e-01 1.08265436e+00 3.50651175e-01 -7.35732615e-01
4.69843358e-01 -7.25744784e-01 2.64006592e-02 -2.06413791e-01
-7.00739622e-02 -7.41486251e-01 -8.78797472e-01 1.32617474e-01
2.82068793e-02 -1.85850501e-01 -1.34670305e+00 1.80259049e+00
-3.23794961e-01 2.67099768e-01 3.53289455e-01 1.02605247e+00
8.93995285e-01 1.01848066e+00 -2.07728043e-01 -8.39407519e-02
1.11157811e+00 -1.01940107e+00 -6.99303389e-01 -4.98450905e-01
7.06249699e-02 -8.88569951e-01 1.10342979e+00 4.20589387e-01
-1.26445138e+00 -5.61161041e-01 -1.12715161e+00 2.13348791e-02
1.84239656e-01 1.99924350e-01 -1.61825567e-02 1.05225503e+00
-1.32889748e+00 2.99763590e-01 -7.09967971e-01 2.96472937e-01
1.55869916e-01 -2.17197314e-02 6.96861371e-02 -2.60373950e-01
-1.05008996e+00 6.94917083e-01 -4.77425873e-01 2.80408114e-01
-1.54058576e+00 -1.03040659e+00 -8.68229687e-01 1.04261264e-01
-1.78552449e-01 -5.91568410e-01 1.65300989e+00 -6.10969603e-01
-1.97234547e+00 4.31969523e-01 -2.96640545e-01 -6.41624629e-01
2.73440957e-01 -5.42731524e-01 -1.10899270e+00 -6.81944117e-02
-6.63021684e-01 -2.98151076e-02 1.09785712e+00 -1.45179284e+00
4.92795110e-02 -8.80480036e-02 -2.86254853e-01 1.61271945e-01
4.12216000e-02 1.71431944e-01 8.14015493e-02 -1.13287437e+00
-3.71729396e-02 -6.11747146e-01 -1.54060796e-01 -6.46931946e-01
-3.60903531e-01 8.22805241e-03 2.27645382e-01 -1.32029092e+00
1.11822045e+00 -2.34754133e+00 -2.79671699e-01 1.77757114e-01
2.88329064e-03 4.72736984e-01 -2.97532409e-01 1.36078402e-01
-3.88712525e-01 -4.18836288e-02 -5.23927987e-01 -6.18618965e-01
1.50512308e-01 -4.88976389e-01 -9.09455001e-01 5.42546034e-01
1.22744814e-01 5.80122471e-01 -5.90436101e-01 2.92204052e-01
3.37858677e-01 1.23062921e+00 -6.33480310e-01 6.80132866e-01
2.27271393e-01 5.08043170e-01 1.34906113e-01 3.27652842e-01
5.23692250e-01 3.04544777e-01 -3.11576813e-01 -2.12241337e-01
1.00177117e-01 9.03338790e-01 -5.94038844e-01 1.63629711e+00
-1.49111855e+00 9.86564517e-01 2.69580662e-01 -6.78500533e-01
1.03594589e+00 7.07234800e-01 -5.26972115e-02 -9.00734425e-01
-1.43523425e-01 5.56436300e-01 -3.94654460e-02 8.51843320e-03
3.05471718e-01 -7.32524469e-02 2.94271082e-01 4.20265645e-01
1.74313128e-01 -2.15037882e-01 -8.41461360e-01 -1.37843564e-01
1.47793722e+00 -2.31283039e-01 -9.99731123e-02 -3.51605155e-02
5.18314421e-01 -9.62694585e-01 -8.71229619e-02 8.94805908e-01
6.02922551e-02 1.07252860e+00 -4.89747077e-01 9.43724737e-02
-9.66125309e-01 -1.61550188e+00 -2.64734864e-01 1.10929668e+00
-4.72324431e-01 -3.80439609e-02 -8.33980918e-01 -2.93673038e-01
-4.70500648e-01 1.36274731e+00 -2.69125670e-01 -3.14491630e-01
-9.06658292e-01 -7.94430315e-01 8.58529091e-01 7.69488931e-01
2.54190773e-01 -8.97913277e-01 5.15818223e-02 1.11739233e-01
-3.95599663e-01 -9.95959461e-01 -7.75111854e-01 5.38349569e-01
-3.66550505e-01 -1.62442490e-01 -1.02497482e+00 -6.22302830e-01
3.87054622e-01 3.21275204e-01 1.36121929e+00 -3.10497195e-01
-3.72646600e-01 7.49500811e-01 -3.15156281e-01 -3.09806406e-01
-7.66580641e-01 -4.16433066e-01 1.74568832e-01 -7.89931715e-02
-1.32044002e-01 -7.40248501e-01 -1.03116107e+00 4.17627126e-01
-5.18570006e-01 -5.52209914e-01 3.21215957e-01 8.40244949e-01
3.78154993e-01 -2.60411292e-01 8.96592438e-01 -1.43805631e-02
6.52343035e-01 -2.04109997e-01 -4.36822951e-01 2.04228252e-01
-4.83099699e-01 2.99639791e-01 8.14931929e-01 -2.51492709e-01
-1.40433979e+00 -4.85646605e-01 -9.36419010e-01 -3.14545721e-01
1.43497288e-02 -1.12175666e-01 -2.72145003e-01 5.84726296e-02
1.19212675e+00 3.64844918e-01 -5.63986182e-01 -5.32139242e-01
5.15841246e-01 1.21909440e+00 8.16979349e-01 -4.60588694e-01
8.90092015e-01 3.62916179e-02 -3.02827239e-01 -9.56293881e-01
-5.58186233e-01 -5.74841201e-01 4.11997616e-01 9.69218183e-03
6.84665024e-01 -1.29753041e+00 -5.76136768e-01 4.21456486e-01
-1.11157024e+00 -6.98647678e-01 -3.02171648e-01 9.37764704e-01
-7.77747571e-01 -2.61091962e-02 -6.96000695e-01 -1.43539143e+00
-8.52448642e-01 -1.08102405e+00 1.11394525e+00 -1.69520348e-01
5.24947718e-02 -7.35359788e-01 3.57745439e-01 4.59499061e-01
9.66798544e-01 -4.49433774e-01 7.39754677e-01 -4.55337435e-01
-3.90355617e-01 -1.15308300e-01 7.95337092e-03 9.56829727e-01
-4.70615178e-02 -5.61565638e-01 -1.98284125e+00 -3.71595800e-01
4.37139928e-01 -1.40100032e-01 1.27579176e+00 7.85408914e-01
1.32546604e+00 -4.98315275e-01 4.79526222e-02 8.53058994e-01
1.17670524e+00 3.37997109e-01 1.21216595e+00 -2.24858135e-01
4.93035972e-01 7.15443566e-02 -6.78334832e-02 2.10097879e-01
-2.25465730e-01 7.40639269e-01 1.72298387e-01 -1.40418574e-01
-8.69305909e-01 -2.73789287e-01 7.71688521e-01 1.13729572e+00
-1.78503349e-01 -8.49027574e-01 -7.08950818e-01 6.07956290e-01
-1.15850568e+00 -1.05313182e+00 3.33765239e-01 2.51905560e+00
7.50564098e-01 -3.38283509e-01 -1.36673659e-01 2.71487862e-01
5.08477449e-01 4.16934699e-01 -3.46711576e-01 -5.85064173e-01
-1.60340577e-01 8.34083378e-01 4.86146599e-01 9.23581779e-01
-7.12864697e-01 3.02132636e-01 6.80359554e+00 8.29368770e-01
-9.39747274e-01 3.10887873e-01 8.56462359e-01 -1.45836100e-01
-5.15003860e-01 -6.71357989e-01 -4.34828550e-01 7.36510381e-02
1.55198944e+00 2.14950815e-01 1.30500090e+00 6.27751589e-01
1.79745749e-01 3.49869549e-01 -1.24196959e+00 1.29734647e+00
3.44231486e-01 -9.09859717e-01 -3.65340322e-01 2.52544507e-02
5.20604789e-01 4.20264661e-01 6.47689342e-01 3.15215707e-01
4.73746717e-01 -1.73346376e+00 6.99482024e-01 5.31174600e-01
1.16076434e+00 -8.04824829e-01 4.34639335e-01 -9.08596665e-02
-1.24637103e+00 -1.51430756e-01 -1.61057070e-01 4.24006194e-01
1.41342133e-01 6.67973757e-01 -1.22063875e+00 3.81907374e-01
6.33310139e-01 5.12774289e-02 3.66193405e-03 9.31947768e-01
-3.81870002e-01 1.36731839e+00 -1.84883371e-01 2.76408255e-01
-8.76130462e-02 2.22849712e-01 9.26459789e-01 1.52354288e+00
5.32241166e-01 1.43265337e-01 -5.14643669e-01 8.25087905e-01
-5.67058742e-01 -1.88234568e-01 -3.27692688e-01 2.01589346e-01
5.60627818e-01 8.82545829e-01 1.32838160e-01 1.68718994e-01
-1.31571248e-01 1.04922259e+00 -1.20863140e-01 9.49077070e-01
-9.47022974e-01 -6.66592598e-01 8.92351151e-01 -2.40849778e-02
3.88930202e-01 -1.59617871e-01 2.12162510e-02 -9.57591295e-01
2.32969895e-01 -1.13525689e+00 -2.10438952e-01 -1.07518244e+00
-1.27587628e+00 1.04428434e+00 -4.40887123e-01 -1.08532321e+00
-3.83198231e-01 -6.60129607e-01 -6.80366218e-01 1.38252139e+00
-1.68009317e+00 -7.73826718e-01 -1.46256551e-01 5.38063407e-01
7.06954062e-01 -3.17650616e-01 1.18932450e+00 4.51403141e-01
-3.16508681e-01 1.10273778e+00 4.66797113e-01 2.75842905e-01
5.69184065e-01 -1.22559881e+00 1.11686015e+00 1.14036548e+00
4.46789354e-01 6.31008148e-01 7.60884941e-01 1.20810002e-01
-1.38021755e+00 -1.21652555e+00 3.18030328e-01 -6.81419611e-01
3.91113907e-01 -6.08902574e-01 -7.87773192e-01 4.73065257e-01
2.37765014e-01 3.85026634e-01 7.47735858e-01 6.92168847e-02
-9.23676968e-01 -2.64703453e-01 -1.11194670e+00 5.45602918e-01
1.18951428e+00 -1.11055267e+00 -5.02736747e-01 3.78072292e-01
1.03295696e+00 -2.75726557e-01 -6.52104378e-01 2.55188435e-01
3.56605560e-01 -6.90412223e-01 1.65286338e+00 -2.11649552e-01
1.65771902e-01 -2.02407092e-01 -9.21653867e-01 -1.82328439e+00
-2.27798402e-01 -1.17909300e+00 -3.29572648e-01 1.18412900e+00
6.32697821e-01 -5.94624758e-01 5.54861315e-02 1.66998222e-01
-7.84489751e-01 -3.67375404e-01 -1.05632937e+00 -1.00435936e+00
7.61468038e-02 -9.88804996e-01 3.57716233e-01 -4.31025177e-02
-5.55535972e-01 3.98073196e-01 -4.37176675e-01 4.79827434e-01
7.46612191e-01 -4.97932792e-01 2.73607790e-01 -5.73895097e-01
-6.98492408e-01 -3.33360553e-01 -1.02870598e-01 -1.39582586e+00
1.50642529e-01 -7.95027792e-01 8.74398410e-01 -1.65096831e+00
-4.59962100e-01 -4.12930250e-01 -6.75444067e-01 -5.85997440e-02
-1.06088161e-01 2.64203221e-01 8.85891262e-03 -1.94123223e-01
-2.59289950e-01 8.46431673e-01 6.33882761e-01 -2.78595179e-01
-2.47059092e-01 3.17429483e-01 -6.19537294e-01 4.54763293e-01
4.92761642e-01 -3.51654083e-01 -3.21000278e-01 -6.57573283e-01
-5.10416999e-02 1.62015364e-01 5.92621267e-01 -1.44688272e+00
2.04213499e-03 4.07218218e-01 3.37130457e-01 -2.09323883e-01
7.79467881e-01 -5.20588338e-01 -4.37377505e-02 -4.53252383e-02
-7.16007531e-01 -4.34966952e-01 2.31843397e-01 6.24791026e-01
-5.53658456e-02 2.81366795e-01 1.04386163e+00 2.19740167e-01
-1.13879032e-01 -5.95441610e-02 -2.13398874e-01 4.62413549e-01
1.04019478e-01 2.20935404e-01 -4.70301032e-01 -6.88505650e-01
-3.25389713e-01 -5.67670643e-01 -1.76597282e-01 2.81061590e-01
8.53359401e-01 -1.15866971e+00 -1.21305680e+00 2.43195876e-01
5.89932315e-02 -3.15718889e-01 4.66286153e-01 3.97304893e-01
-1.06188782e-01 4.29592162e-01 4.95850891e-01 -3.28747690e-01
-9.17738318e-01 3.31559956e-01 9.07325387e-01 2.93695536e-02
-8.09536517e-01 1.38164091e+00 5.28743744e-01 -2.46814340e-01
4.70458001e-01 -4.30328906e-01 9.51660275e-02 -6.28739417e-01
9.32811499e-01 6.32560134e-01 7.56244242e-01 -5.90991914e-01
-5.00150561e-01 4.04606581e-01 4.32599694e-01 -7.24668741e-01
1.29037917e+00 -2.31455743e-01 4.05818045e-01 3.33990127e-01
1.59123564e+00 6.28284693e-01 -1.20345259e+00 -3.10626298e-01
-6.32911861e-01 -3.84319335e-01 6.55357540e-01 -1.40104175e+00
-8.42467368e-01 9.22798336e-01 1.09337354e+00 5.73672950e-02
1.56780934e+00 5.16770147e-02 8.97307038e-01 4.48429734e-01
1.39137849e-01 -7.69279838e-01 2.10666373e-01 5.81639826e-01
1.23773694e+00 -9.10095692e-01 -4.64445621e-01 9.48353931e-02
-4.47627187e-01 7.97201633e-01 -6.27979711e-02 -1.53428828e-02
6.31148994e-01 5.60790241e-01 2.97753155e-01 3.41180831e-01
-5.90415239e-01 -1.64566934e-02 6.82967126e-01 9.83014226e-01
6.85230136e-01 1.74061775e-01 5.08741677e-01 6.27172709e-01
-4.93855745e-01 -6.30166411e-01 2.04082262e-02 1.58612624e-01
-4.19980407e-01 -6.86909139e-01 -6.26163542e-01 1.52999803e-01
-6.52945459e-01 -7.21111059e-01 5.30328080e-02 -1.27152875e-01
-6.74511492e-01 1.56474113e+00 -8.27767029e-02 -4.59397674e-01
5.08472860e-01 9.33302790e-02 4.15908724e-01 -3.12403202e-01
-6.59427285e-01 1.09338082e-01 2.65782654e-01 -3.66683871e-01
-9.46336612e-02 -2.83287168e-01 -9.60991263e-01 -2.21057839e-04
-3.63440484e-01 1.93513855e-01 1.11434519e+00 5.99724710e-01
2.34353229e-01 1.27845871e+00 1.16262102e+00 -7.24372745e-01
-8.84105444e-01 -1.15470862e+00 -3.46537024e-01 -2.24380940e-02
9.80227411e-01 -1.10190086e-01 -8.71026099e-01 -9.59342048e-02] | [15.093616485595703, 5.939564228057861] |
3e5e62ef-d224-4fa5-855f-3f05af6864cf | learning-warped-guidance-for-blind-face | 1804.04829 | null | http://arxiv.org/abs/1804.04829v2 | http://arxiv.org/pdf/1804.04829v2.pdf | Learning Warped Guidance for Blind Face Restoration | This paper studies the problem of blind face restoration from an
unconstrained blurry, noisy, low-resolution, or compressed image (i.e.,
degraded observation). For better recovery of fine facial details, we modify
the problem setting by taking both the degraded observation and a high-quality
guided image of the same identity as input to our guided face restoration
network (GFRNet). However, the degraded observation and guided image generally
are different in pose, illumination and expression, thereby making plain CNNs
(e.g., U-Net) fail to recover fine and identity-aware facial details. To tackle
this issue, our GFRNet model includes both a warping subnetwork (WarpNet) and a
reconstruction subnetwork (RecNet). The WarpNet is introduced to predict flow
field for warping the guided image to correct pose and expression (i.e., warped
guidance), while the RecNet takes the degraded observation and warped guidance
as input to produce the restoration result. Due to that the ground-truth flow
field is unavailable, landmark loss together with total variation
regularization are incorporated to guide the learning of WarpNet. Furthermore,
to make the model applicable to blind restoration, our GFRNet is trained on the
synthetic data with versatile settings on blur kernel, noise level,
downsampling scale factor, and JPEG quality factor. Experiments show that our
GFRNet not only performs favorably against the state-of-the-art image and face
restoration methods, but also generates visually photo-realistic results on
real degraded facial images. | ['WangMeng Zuo', 'Yuting Ye', 'Ruigang Yang', 'Liang Lin', 'Xiaoming Li', 'Ming Liu'] | 2018-04-13 | learning-warped-guidance-for-blind-face-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Xiaoming_Li_Learning_Warped_Guidance_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Xiaoming_Li_Learning_Warped_Guidance_ECCV_2018_paper.pdf | eccv-2018-9 | ['blind-face-restoration'] | ['computer-vision'] | [ 3.53574872e-01 -9.05720964e-02 1.06015347e-01 -2.24762723e-01
-3.05740297e-01 -3.29636574e-01 4.58768189e-01 -9.19093430e-01
-9.37835574e-02 8.23132575e-01 4.69649673e-01 -1.66912064e-01
2.18948741e-02 -7.93348253e-01 -7.51671314e-01 -9.54787791e-01
3.42869371e-01 -1.90863132e-01 -2.89174557e-01 -1.95799708e-01
8.25856104e-02 6.07451260e-01 -1.43594790e+00 7.51539916e-02
1.12247503e+00 9.71434116e-01 1.06023051e-01 3.98133755e-01
1.14084788e-01 6.97810054e-01 -3.29318285e-01 -4.55518186e-01
6.32467449e-01 -6.72565341e-01 -5.88378012e-01 3.45826209e-01
9.56898808e-01 -9.73765671e-01 -8.11469495e-01 1.35444081e+00
5.89241445e-01 3.79067481e-01 3.76413941e-01 -1.08742392e+00
-1.15270746e+00 2.36059306e-03 -6.58638716e-01 1.48050174e-01
3.58036608e-01 5.23630679e-01 3.22069317e-01 -1.04988921e+00
6.13282979e-01 1.62264562e+00 4.45373237e-01 8.56298387e-01
-1.34938681e+00 -8.62018585e-01 1.99220963e-02 2.19465092e-01
-1.16912997e+00 -9.98946369e-01 7.84049571e-01 -1.98655948e-01
2.77558506e-01 -1.28155788e-02 5.98846436e-01 1.23095989e+00
1.66397214e-01 2.00931579e-01 1.12852490e+00 -6.84725493e-02
2.01490968e-01 -2.78902531e-01 -4.20070082e-01 8.25478733e-01
1.96229354e-01 6.13851428e-01 -4.78615731e-01 9.00649428e-02
1.36388791e+00 1.58335388e-01 -1.07668829e+00 -2.52716005e-01
-9.43499923e-01 5.36020279e-01 7.70958006e-01 -2.03358904e-02
-4.60264176e-01 5.60673997e-02 -8.98303390e-02 3.61493915e-01
5.34782529e-01 1.24418899e-01 -1.47492424e-01 2.76488274e-01
-1.04187489e+00 9.22828093e-02 5.50261021e-01 5.67308366e-01
8.93655002e-01 5.11451602e-01 -3.39044571e-01 1.04502821e+00
3.98999959e-01 5.74226201e-01 4.17576045e-01 -1.25853682e+00
4.00927991e-01 1.91239879e-01 2.58287817e-01 -1.06821108e+00
2.07822584e-02 -4.04474795e-01 -1.20478952e+00 6.24593854e-01
6.50616586e-01 -1.03833012e-01 -1.41637897e+00 1.82908094e+00
2.37911686e-01 5.67025125e-01 6.22893423e-02 1.65789664e+00
9.37959909e-01 6.07774198e-01 -1.66793123e-01 -4.67848897e-01
1.15966630e+00 -1.04271340e+00 -9.24174905e-01 -3.45807165e-01
-2.19550282e-01 -9.53339815e-01 8.90929699e-01 1.37616619e-01
-1.26540411e+00 -7.46771157e-01 -9.16104138e-01 -2.81047225e-01
1.97278753e-01 7.90172294e-02 2.31726140e-01 5.37773490e-01
-1.35840762e+00 8.24922323e-01 -6.52658403e-01 -3.44360799e-01
6.99329078e-01 1.23645648e-01 -7.10055292e-01 -8.93571496e-01
-9.70036387e-01 7.11959064e-01 -1.91640630e-01 5.15518010e-01
-1.34465206e+00 -8.67155433e-01 -9.43700910e-01 4.19749245e-02
6.95937127e-02 -1.02862513e+00 7.24024415e-01 -1.12009573e+00
-1.62404585e+00 6.21851802e-01 -2.35352024e-01 3.38099226e-02
6.86938286e-01 -6.41799271e-02 -3.93820196e-01 2.57963151e-01
8.41470212e-02 5.43635249e-01 1.42040586e+00 -1.53716815e+00
-1.53539494e-01 -3.59760284e-01 8.97773821e-03 2.74906784e-01
-5.35650812e-02 -6.73798621e-02 -4.63269860e-01 -8.02750826e-01
2.62139440e-01 -4.98605043e-01 4.33758833e-02 4.53030348e-01
-4.67184842e-01 6.26539886e-01 9.51569021e-01 -1.27645779e+00
6.82804167e-01 -2.27001214e+00 1.66977435e-01 -9.39326547e-03
2.21258268e-01 2.66084045e-01 -7.10251033e-01 -1.21123880e-01
-4.68014091e-01 -1.41148930e-02 -4.59343582e-01 -5.25553048e-01
-5.31623304e-01 3.32203865e-01 -3.11875373e-01 8.01519454e-01
4.82686386e-02 7.83109546e-01 -1.08161509e+00 -1.86575562e-01
3.87314320e-01 1.07564342e+00 -5.59017599e-01 4.72686917e-01
1.69434577e-01 1.02292824e+00 -2.00795174e-01 8.71660471e-01
1.20789647e+00 -4.21682838e-03 -1.34992436e-01 -6.52533054e-01
1.76407583e-02 -2.29224890e-01 -1.04437089e+00 1.62066722e+00
-5.65000594e-01 5.18786192e-01 6.09802365e-01 -5.72268367e-01
9.44967449e-01 4.01260465e-01 3.05858642e-01 -7.09226012e-01
1.33471757e-01 1.17998451e-01 -1.26242369e-01 -5.54302573e-01
2.14777365e-01 -4.08375740e-01 8.15748692e-01 2.75518000e-01
1.50777489e-01 4.49985042e-02 -2.20706329e-01 -1.34387434e-01
9.50801015e-01 2.35104725e-01 -2.63266116e-01 -9.64920036e-03
6.87054217e-01 -5.97496152e-01 7.80977964e-01 2.59376138e-01
-2.31665105e-01 1.19314086e+00 2.78217643e-01 -4.44046468e-01
-1.07917523e+00 -1.18027067e+00 -9.29463282e-02 6.91972673e-01
4.90501851e-01 1.83904335e-01 -8.06874871e-01 -4.93179083e-01
-5.98190501e-02 3.83012891e-01 -5.40583253e-01 -3.37432474e-01
-6.66809678e-01 -6.97578788e-01 2.70463675e-01 1.95480898e-01
1.02055860e+00 -1.07485652e+00 8.11580718e-02 1.09960712e-01
-4.43140954e-01 -1.08932245e+00 -9.39569056e-01 -5.69997430e-01
-8.05014610e-01 -1.31424201e+00 -9.32102978e-01 -8.00915539e-01
1.19209146e+00 4.88797218e-01 6.91582203e-01 2.79262573e-01
-2.05536425e-01 1.05276227e-01 -2.73933560e-02 5.03602087e-01
-3.14397454e-01 -7.33947158e-01 3.28016020e-02 6.04503751e-01
-4.94083077e-01 -9.05413568e-01 -1.03701389e+00 4.22720462e-01
-1.01176906e+00 3.55030410e-02 4.02902186e-01 1.22553718e+00
3.63897562e-01 1.17704622e-01 2.90762275e-01 -3.21365714e-01
6.06994331e-01 -1.89491495e-01 -5.27212977e-01 6.12036213e-02
-5.91996968e-01 -1.86143592e-01 7.75176942e-01 -4.96572226e-01
-1.42941654e+00 -1.78292200e-01 -3.98119278e-02 -9.58911598e-01
-2.88674571e-02 3.41106020e-02 -5.47301412e-01 -5.63451290e-01
6.26657188e-01 2.61405796e-01 4.89387661e-01 -7.06393600e-01
3.65770847e-01 3.75082314e-01 9.67027128e-01 -2.91143358e-01
1.06646860e+00 6.31351292e-01 3.69778275e-02 -5.66200376e-01
-5.49380958e-01 7.86423236e-02 -3.37147266e-01 -3.33347589e-01
6.11157656e-01 -1.00671303e+00 -5.99995792e-01 7.72283196e-01
-1.15354443e+00 -4.14692104e-01 -2.29049787e-01 4.28829134e-01
-4.61758077e-01 5.12443662e-01 -9.03420806e-01 -5.14372885e-01
-3.29320431e-01 -1.29399502e+00 8.74203324e-01 6.20369971e-01
4.81274396e-01 -8.96687508e-01 -4.04106379e-01 6.33997560e-01
6.79170191e-01 2.92507678e-01 7.67883718e-01 3.06554615e-01
-8.70784879e-01 1.13351643e-01 -5.69135308e-01 7.53400326e-01
4.54481661e-01 -1.59309089e-01 -1.12266719e+00 -6.58064604e-01
1.58818752e-01 3.72233661e-03 1.01475298e+00 6.04034305e-01
9.83395636e-01 -6.08775377e-01 9.48405918e-03 1.25913441e+00
1.32487726e+00 3.12844776e-02 1.06208110e+00 1.66958366e-02
8.56609643e-01 5.30611575e-01 2.80067027e-01 4.70798053e-02
1.43599033e-01 6.75644338e-01 7.15943038e-01 -3.99928272e-01
-8.85102093e-01 -3.72129679e-01 4.94942725e-01 2.03173235e-01
-4.09486771e-01 -1.31173268e-01 -4.35286552e-01 5.05666733e-01
-1.63932455e+00 -8.90090823e-01 2.54548281e-01 2.31907940e+00
8.17857802e-01 -5.48580408e-01 -4.30822581e-01 -1.12316869e-01
8.99297416e-01 3.38018477e-01 -6.36347532e-01 9.25589427e-02
-2.82863110e-01 1.32717013e-01 3.61422360e-01 8.60886514e-01
-7.78111756e-01 9.58238363e-01 5.17485189e+00 6.74260437e-01
-1.31898141e+00 1.90052226e-01 8.02226841e-01 -2.47191312e-03
-1.89447537e-01 -5.62533848e-02 -2.79248089e-01 6.32637143e-01
4.41344351e-01 -2.18359120e-02 1.15623093e+00 3.46561521e-01
5.64196527e-01 1.00584917e-01 -7.82484114e-01 1.21383953e+00
2.14374125e-01 -1.09048188e+00 1.50663540e-01 1.27756402e-01
7.94703066e-01 -3.10128361e-01 1.48597583e-01 -2.33327299e-01
2.82430023e-01 -1.34956074e+00 4.35584933e-01 8.73823106e-01
1.22233236e+00 -6.23535037e-01 8.02258432e-01 -4.86410186e-02
-7.70581126e-01 -1.22436404e-01 -4.57750946e-01 2.32476130e-01
3.74660224e-01 6.83439791e-01 -2.44472966e-01 6.37076676e-01
6.45899892e-01 9.09271061e-01 -2.96905279e-01 1.10210323e+00
-5.12370706e-01 3.53014469e-01 1.16639838e-01 1.02278042e+00
-1.84448898e-01 -5.08122027e-01 7.25012839e-01 5.84353983e-01
5.33517063e-01 1.95802271e-01 2.54262518e-02 1.08077013e+00
-2.99865991e-01 -2.62330264e-01 -3.30901891e-01 4.09543663e-01
2.62259454e-01 1.36247981e+00 -2.87700444e-01 -7.71476105e-02
-1.75621793e-01 1.14505315e+00 -3.46151963e-02 1.02883363e+00
-5.54202914e-01 7.25160688e-02 1.26184893e+00 2.25948796e-01
1.21425860e-01 1.13806486e-01 -1.47733524e-01 -1.31203628e+00
-4.21181833e-03 -7.65049517e-01 -2.08502542e-02 -1.09276211e+00
-1.44148326e+00 8.75634432e-01 -3.75665933e-01 -1.19059730e+00
2.41110213e-02 -4.06123459e-01 -8.19673479e-01 1.39867616e+00
-1.94963431e+00 -9.99738634e-01 -7.80046523e-01 8.78968656e-01
2.92811275e-01 -2.77687192e-01 4.14394796e-01 4.37967330e-01
-7.31083751e-01 5.40806055e-01 -8.11583400e-02 2.15849474e-01
9.85801935e-01 -6.96364939e-01 1.80254653e-01 1.15110421e+00
-4.91855025e-01 5.07033825e-01 6.32962763e-01 -6.61038756e-01
-1.38760293e+00 -1.49893773e+00 4.96903688e-01 5.33495061e-02
2.33986825e-01 3.87304164e-02 -1.03744543e+00 4.79426384e-01
2.47609783e-02 7.30764568e-01 -1.35736344e-02 -6.43369317e-01
-4.42572713e-01 -4.26889896e-01 -1.65604007e+00 6.36423409e-01
1.30045462e+00 -5.43463588e-01 -1.88241333e-01 2.08794177e-01
6.20813608e-01 -5.76286018e-01 -6.44520342e-01 4.21301097e-01
5.01969039e-01 -1.11721253e+00 1.23497880e+00 -2.09230870e-01
6.35341048e-01 -5.70608199e-01 -3.82267386e-02 -1.68586802e+00
-3.23306590e-01 -7.16203034e-01 -1.48829669e-01 1.17657351e+00
-4.43154126e-02 -8.40998948e-01 6.92478538e-01 5.21582127e-01
-1.21504389e-01 -5.38823247e-01 -1.02583027e+00 -6.45700455e-01
-1.66667596e-01 8.47256184e-02 6.93391442e-01 9.32148755e-01
-5.61767697e-01 1.84909292e-02 -7.46369600e-01 2.43933409e-01
9.26926255e-01 2.38189194e-03 5.03690720e-01 -8.89754593e-01
5.94629534e-02 -2.43856132e-01 -2.71245241e-01 -1.13979173e+00
2.03723013e-01 -7.07804859e-01 7.04713911e-02 -1.50802028e+00
-5.63442660e-03 -1.39257520e-01 4.49979007e-02 5.61763704e-01
-1.66390046e-01 6.07918561e-01 2.62124747e-01 3.17661494e-01
3.38701487e-01 7.90734887e-01 2.06773663e+00 -1.62950769e-01
-2.56827831e-01 -1.56032518e-01 -8.96625876e-01 4.22657371e-01
4.90154624e-01 -1.04317680e-01 -3.94938439e-01 -6.26510203e-01
-4.50298816e-01 3.71412158e-01 8.96205544e-01 -8.60042870e-01
3.03294182e-01 -1.35869548e-01 7.93068588e-01 2.25661352e-01
4.91601825e-01 -7.48240292e-01 2.21304238e-01 1.98891357e-01
-7.05428272e-02 -3.48563164e-01 -4.42531407e-02 5.58573782e-01
-2.99827248e-01 2.40308478e-01 1.18610299e+00 -3.06463391e-02
-3.64505440e-01 7.81947374e-01 2.43114978e-01 -2.06068218e-01
5.16905010e-01 -5.37206829e-01 -7.39738584e-01 -6.56151414e-01
-7.47128725e-01 -5.17386757e-03 7.74506509e-01 5.13081908e-01
9.64959621e-01 -1.35867620e+00 -8.53144348e-01 7.84690619e-01
-3.22100222e-01 -3.61888274e-03 5.63944042e-01 7.79175937e-01
-5.25649250e-01 -8.80385861e-02 -5.32087266e-01 -2.48486847e-01
-9.97836769e-01 5.08255661e-01 7.52175808e-01 1.99031293e-01
-8.78509521e-01 6.85617089e-01 4.45299685e-01 -3.75250846e-01
2.18133911e-01 1.66874267e-02 -1.77920446e-01 -2.85594940e-01
7.46194720e-01 3.12908828e-01 -2.85495762e-02 -8.71690094e-01
-1.35397300e-01 7.57522643e-01 1.59192607e-01 -1.67011307e-03
1.21539116e+00 -4.40910190e-01 -3.95245463e-01 -4.81317312e-01
1.16558158e+00 -9.64545086e-02 -1.84581387e+00 -2.18534067e-01
-7.72588015e-01 -1.01684701e+00 4.52618599e-01 -8.77816141e-01
-1.93570077e+00 6.09385133e-01 7.79829800e-01 -3.96634281e-01
1.43635416e+00 -3.98659617e-01 7.48276949e-01 -3.37139130e-01
2.43344501e-01 -4.61875081e-01 7.70170391e-02 2.67639518e-01
1.28458107e+00 -1.07849216e+00 -8.75495821e-02 -5.59411407e-01
-2.65704662e-01 1.21043682e+00 7.85578251e-01 -1.31218627e-01
6.26588583e-01 1.05943764e-02 2.35455513e-01 1.27968848e-01
-3.31808358e-01 1.41565083e-02 2.31158093e-01 8.52003574e-01
1.11994773e-01 -2.55051881e-01 2.75744908e-02 1.61958620e-01
-2.53003687e-01 2.41750434e-01 6.31820083e-01 3.43849957e-01
-1.36873454e-01 -7.16790617e-01 -7.37575829e-01 2.08183974e-01
-1.42744765e-01 -1.33447379e-01 6.95195198e-02 4.59381878e-01
3.25122774e-01 1.20990479e+00 6.84117945e-03 -3.24238330e-01
3.92530978e-01 -2.95401692e-01 4.87642229e-01 -3.22077960e-01
-2.19650075e-01 3.43306772e-02 -2.03648195e-01 -9.35521007e-01
-2.27136642e-01 -2.11899027e-01 -8.97839010e-01 -7.39606261e-01
3.84296849e-02 -1.50950477e-01 5.19293547e-01 7.77836025e-01
2.89476007e-01 5.08151472e-01 8.64174545e-01 -1.14975095e+00
-2.00144827e-01 -9.51761901e-01 -8.16955745e-01 4.52077687e-01
1.01719797e+00 -6.73298061e-01 -7.57002652e-01 1.99148670e-01] | [12.793498039245605, -0.03583632782101631] |
5ef737b3-701f-4e00-9dce-196a299c49e6 | neural-factor-graph-models-for-cross-lingual | 1805.04570 | null | http://arxiv.org/abs/1805.04570v3 | http://arxiv.org/pdf/1805.04570v3.pdf | Neural Factor Graph Models for Cross-lingual Morphological Tagging | Morphological analysis involves predicting the syntactic traits of a word
(e.g. {POS: Noun, Case: Acc, Gender: Fem}). Previous work in morphological
tagging improves performance for low-resource languages (LRLs) through
cross-lingual training with a high-resource language (HRL) from the same
family, but is limited by the strict, often false, assumption that tag sets
exactly overlap between the HRL and LRL. In this paper we propose a method for
cross-lingual morphological tagging that aims to improve information sharing
between languages by relaxing this assumption. The proposed model uses
factorial conditional random fields with neural network potentials, making it
possible to (1) utilize the expressive power of neural network representations
to smooth over superficial differences in the surface forms, (2) model pairwise
and transitive relationships between tags, and (3) accurately generate tag sets
that are unseen or rare in the training data. Experiments on four languages
from the Universal Dependencies Treebank demonstrate superior tagging
accuracies over existing cross-lingual approaches. | ['Matthew R. Gormley', 'Graham Neubig', 'Chaitanya Malaviya'] | 2018-05-11 | neural-factor-graph-models-for-cross-lingual-1 | https://aclanthology.org/P18-1247 | https://aclanthology.org/P18-1247.pdf | acl-2018-7 | ['morphological-tagging'] | ['natural-language-processing'] | [ 6.94710156e-03 2.38024756e-01 -3.35813522e-01 -6.05426490e-01
-8.07388783e-01 -9.53861356e-01 1.81606740e-01 3.82447809e-01
-6.24091268e-01 9.23569739e-01 3.77700001e-01 -6.09087348e-01
1.82800099e-01 -7.13807821e-01 -5.26353896e-01 -4.24585342e-01
-2.01082170e-01 5.46903908e-01 7.86008760e-02 -3.56932655e-02
-1.41383752e-01 4.08048958e-01 -1.02071106e+00 4.65481341e-01
9.47294354e-01 5.39159894e-01 3.61945987e-01 1.55255310e-02
-3.43333483e-01 5.30272424e-01 -4.35086578e-01 -9.32772517e-01
1.45745441e-01 -1.51150733e-01 -8.77517819e-01 -2.16075644e-01
4.48424250e-01 2.56594568e-01 1.49279967e-01 1.34571695e+00
2.03317687e-01 -1.58931632e-02 5.56535304e-01 -6.09477818e-01
-1.09787691e+00 1.37755787e+00 -5.27539492e-01 1.94037229e-01
4.60404418e-02 -3.40939552e-01 1.56764889e+00 -8.76895070e-01
7.45130837e-01 1.37128925e+00 1.01410747e+00 7.73956299e-01
-1.40034938e+00 -6.10704422e-01 2.22737297e-01 -2.58783430e-01
-1.72713816e+00 -4.50186729e-01 5.47197342e-01 -4.17143703e-01
1.13259101e+00 -6.97291270e-03 3.54763061e-01 8.43068063e-01
1.75537035e-01 6.10603094e-01 1.26992893e+00 -7.41715729e-01
-7.74798468e-02 2.03602538e-01 1.82788759e-01 7.75028586e-01
4.36416447e-01 -1.49885833e-01 -4.37179059e-01 -1.91561908e-01
5.25296569e-01 -4.99430388e-01 -6.81289751e-03 -3.11818346e-02
-1.02533853e+00 8.96865606e-01 9.44344252e-02 8.59727919e-01
-9.95021462e-02 -2.29684785e-01 3.83231163e-01 1.21006947e-02
5.59666812e-01 5.57521045e-01 -9.91198480e-01 2.22550303e-01
-7.65940368e-01 -1.22589849e-01 8.17507923e-01 1.01504230e+00
9.48497832e-01 1.71353519e-01 3.32192555e-02 1.22484660e+00
3.13015342e-01 3.70561332e-01 8.28013241e-01 -5.34043610e-01
4.84232038e-01 4.13486689e-01 1.26295993e-02 -5.64875901e-01
-5.21702766e-01 -3.16237241e-01 -3.97257447e-01 -3.86388212e-01
4.86237437e-01 -3.82306308e-01 -9.00292635e-01 2.27430296e+00
2.05466196e-01 -3.62105608e-01 4.28792387e-01 2.28992894e-01
7.44229496e-01 5.28978944e-01 7.80150771e-01 -1.41630769e-01
1.61665297e+00 -4.13642555e-01 -6.97185814e-01 -6.15722537e-01
1.13776398e+00 -8.62555742e-01 9.61344182e-01 -8.59589502e-02
-1.12020421e+00 -6.29969418e-01 -8.46943200e-01 -2.09027365e-01
-7.39935398e-01 4.88053352e-01 9.57379937e-01 9.60917950e-01
-1.10998821e+00 5.76359510e-01 -8.87391865e-01 -3.47663999e-01
1.01900667e-01 5.90264261e-01 -5.70576310e-01 6.75913393e-02
-1.34540522e+00 9.38224852e-01 9.30341125e-01 6.64279913e-04
-1.00112028e-01 -6.04926288e-01 -1.24235010e+00 -1.45386815e-01
-5.09429984e-02 -7.70980045e-02 9.59367871e-01 -1.30190825e+00
-1.10109949e+00 1.37216687e+00 -4.01757360e-02 -2.72806615e-01
-2.89979726e-01 -2.60116547e-01 -7.69335151e-01 -3.72089803e-01
4.23773080e-01 8.68130267e-01 2.47028153e-02 -1.14435101e+00
-8.02433610e-01 -3.57577413e-01 -5.66142984e-02 1.50009349e-01
-3.78891289e-01 5.76679885e-01 -1.51754871e-01 -7.86581397e-01
5.99347241e-02 -9.35879946e-01 -2.18179047e-01 -7.82704890e-01
-3.29999268e-01 -7.16577053e-01 2.06634715e-01 -9.96126950e-01
1.23713636e+00 -2.10903311e+00 -3.07311714e-01 1.59519881e-01
-4.20164883e-01 4.07524437e-01 2.43217554e-02 3.81653845e-01
-1.84092462e-01 4.31761205e-01 -8.06014091e-02 -3.85150135e-01
-6.05580909e-03 7.31527627e-01 -8.07087496e-02 4.37941641e-01
5.29472291e-01 8.10264528e-01 -7.36237884e-01 -7.31059015e-01
-2.43112043e-01 2.84343690e-01 -5.96529067e-01 -1.78737100e-03
3.46406386e-03 2.27559865e-01 -1.68588355e-01 6.92965567e-01
4.98047203e-01 2.31761709e-01 9.31023479e-01 -2.12136909e-01
-4.74643975e-01 9.18288589e-01 -1.06048262e+00 1.44959593e+00
-6.38708413e-01 8.67002830e-02 1.94467567e-02 -7.39377737e-01
1.00638068e+00 3.19113642e-01 3.57635349e-01 -4.43613142e-01
1.48941308e-01 5.05332410e-01 3.97754550e-01 -1.01340204e-01
6.50359511e-01 -4.87114906e-01 -6.15628362e-01 1.37810946e-01
4.84497368e-01 3.27757299e-01 2.90650457e-01 -3.09536427e-01
7.26529598e-01 4.55071688e-01 7.58994997e-01 -6.79277837e-01
3.23696584e-01 -7.91273192e-02 1.31510603e+00 5.08297324e-01
-4.29973528e-02 8.80597755e-02 2.11227864e-01 -3.93576324e-01
-9.66279984e-01 -1.12495196e+00 -6.07049286e-01 1.44899225e+00
-1.50536492e-01 -3.45488042e-01 -6.12062335e-01 -8.88406038e-01
-9.31940451e-02 9.27928329e-01 -6.01934552e-01 1.50466606e-01
-1.03904891e+00 -9.37350154e-01 8.98675740e-01 7.28734612e-01
6.77216798e-02 -1.28823435e+00 9.82716214e-04 5.35214782e-01
-8.22153986e-02 -1.34766662e+00 -3.32187444e-01 7.90434778e-01
-5.19293487e-01 -5.71019053e-01 -8.41668472e-02 -1.35621583e+00
6.23594046e-01 -2.72906303e-01 1.42367160e+00 4.95356061e-02
1.43298566e-01 -1.82726488e-01 -4.84365523e-01 -3.00854862e-01
-6.38372898e-01 2.42853403e-01 2.58396685e-01 -7.79836345e-03
7.84920216e-01 -3.90292704e-01 1.64784238e-01 3.53375413e-02
-7.35672653e-01 -2.38954276e-01 4.19132382e-01 7.16212988e-01
6.71373546e-01 -4.77351695e-02 5.62289178e-01 -1.31577015e+00
-4.59494479e-02 -5.02854228e-01 -5.08353710e-01 4.11590010e-01
-2.83914536e-01 2.21396044e-01 6.26795173e-01 -5.38522720e-01
-1.35246015e+00 2.28899390e-01 -5.24327338e-01 3.54023695e-01
-3.06410044e-01 5.89129090e-01 -6.58267915e-01 1.47464439e-01
4.15657163e-01 -1.51939541e-01 -6.84154689e-01 -7.34767675e-01
2.86259085e-01 5.44610322e-01 5.88453650e-01 -9.15656209e-01
5.88532448e-01 7.97590390e-02 -2.34319791e-01 -7.06788599e-01
-1.22418940e+00 -3.73929143e-01 -1.29483211e+00 4.14183557e-01
1.07468808e+00 -1.11244226e+00 -2.01305225e-01 7.64508992e-02
-1.03963709e+00 -3.89888525e-01 -1.80152819e-01 5.88672817e-01
-1.17912717e-01 3.20883960e-01 -1.03274846e+00 -6.22548044e-01
-1.60505176e-01 -6.05608106e-01 9.02716637e-01 1.68403119e-01
-4.63731170e-01 -1.49649811e+00 4.01527621e-02 -1.20307975e-01
-7.49368295e-02 -3.14988717e-02 1.26172709e+00 -1.22496402e+00
-2.57685059e-03 -9.30758491e-02 6.31426200e-02 4.07231539e-01
3.60664338e-01 -8.76296964e-03 -7.29355812e-01 -2.20349446e-01
-4.29862291e-01 -3.72373223e-01 5.71530461e-01 2.36757934e-01
5.33243597e-01 -3.85165095e-01 -3.62365276e-01 6.61186039e-01
1.49896026e+00 1.83059350e-01 4.35611129e-01 1.63848266e-01
7.75838912e-01 8.53795052e-01 5.35649657e-01 5.10558859e-03
6.43356025e-01 5.64040661e-01 -2.47054607e-01 -1.70118034e-01
-1.76014110e-01 -5.24203956e-01 7.69187331e-01 1.34886050e+00
1.14021048e-01 -1.27896935e-01 -9.43491220e-01 9.29818034e-01
-1.41788805e+00 -6.68918967e-01 -1.55492261e-01 2.15717244e+00
1.36812556e+00 1.11779071e-01 -1.74934447e-01 -3.89813751e-01
9.47131932e-01 -7.44580701e-02 1.75842479e-01 -8.17456305e-01
-4.30144757e-01 6.17739797e-01 6.86949372e-01 6.08818829e-01
-1.38177371e+00 1.66255414e+00 5.79018354e+00 9.54800665e-01
-8.49094510e-01 3.55231375e-01 3.76071811e-01 6.13574445e-01
-3.83661270e-01 1.55563354e-01 -1.54931211e+00 3.11690211e-01
9.29320872e-01 1.60182670e-01 6.16228383e-04 7.11818337e-01
-3.09131712e-01 1.42053276e-01 -1.05431938e+00 3.70796919e-01
2.80156005e-02 -8.16105068e-01 3.62513848e-02 1.84000656e-01
7.47169256e-01 1.46562546e-01 -1.71005189e-01 4.67768669e-01
1.02468753e+00 -8.69415700e-01 8.97770464e-01 -1.29220905e-02
1.07579541e+00 -7.65628934e-01 9.41921115e-01 2.09496230e-01
-1.40724206e+00 2.31157780e-01 -7.20637023e-01 1.58892497e-02
1.51970908e-01 2.64807492e-01 -8.78289998e-01 4.88866478e-01
5.09302199e-01 4.45774376e-01 -6.97697937e-01 6.66693091e-01
-5.31116605e-01 9.08176303e-01 -2.79032946e-01 -3.24242450e-02
2.70273447e-01 -3.99269909e-02 2.67642677e-01 1.71863604e+00
3.58686775e-01 -6.75268620e-02 6.52037144e-01 5.71717441e-01
-6.79830834e-02 6.95852876e-01 -6.35856092e-01 -1.01833060e-01
5.87075830e-01 1.25306404e+00 -8.68046045e-01 -9.27379355e-02
-7.41357028e-01 7.14332163e-01 9.77326155e-01 -1.99169517e-02
-3.49238247e-01 -1.07260413e-01 7.13103592e-01 -8.92829150e-02
4.86996472e-01 -4.73752707e-01 -2.91372627e-01 -1.07610536e+00
-1.79325089e-01 -4.42373335e-01 6.89048946e-01 -3.35335314e-01
-1.70663440e+00 7.11620152e-01 -1.98419094e-01 -8.08585346e-01
-2.74524420e-01 -9.44003880e-01 -1.02853730e-01 1.14081442e+00
-1.39486825e+00 -1.71131921e+00 6.79812908e-01 3.93650085e-01
2.00603202e-01 -1.20363407e-01 1.30387151e+00 4.42663461e-01
-4.32249755e-01 9.31682110e-01 -8.16357359e-02 6.79644108e-01
7.92077422e-01 -1.46906662e+00 4.87492859e-01 9.38343883e-01
4.50779170e-01 9.04501677e-01 3.41932118e-01 -9.76513684e-01
-7.20835149e-01 -1.30711508e+00 1.79737759e+00 -4.77403432e-01
9.50086236e-01 -5.79193532e-01 -1.11519873e+00 1.09375393e+00
7.52389133e-02 6.66317865e-02 1.23382318e+00 6.63668215e-01
-5.73107600e-01 1.11291498e-01 -9.97233212e-01 5.59236288e-01
9.71984625e-01 -5.85070848e-01 -8.38281333e-01 1.26069516e-01
6.38795316e-01 -1.53920591e-01 -1.06692886e+00 2.68337876e-01
3.89970332e-01 -4.78385359e-01 6.17240310e-01 -9.28070068e-01
1.20908007e-01 -1.38373375e-01 -4.25055325e-01 -1.37563217e+00
-6.89790428e-01 -5.02264798e-01 5.67371964e-01 1.93104160e+00
7.48081923e-01 -3.68789792e-01 4.02838171e-01 4.14624244e-01
-5.31964481e-01 -2.99945742e-01 -9.86655056e-01 -8.29149127e-01
5.81063807e-01 -3.04035962e-01 6.07324421e-01 1.30102515e+00
7.04806298e-02 4.33647633e-01 -2.69853503e-01 4.49663341e-01
3.17159504e-01 -3.52601893e-02 -2.72351876e-03 -1.39193881e+00
-4.38512623e-01 -1.53988004e-01 -4.24693942e-01 -6.48983061e-01
7.68235326e-01 -1.39100885e+00 3.23033601e-01 -1.18188202e+00
1.30540267e-01 -1.02163303e+00 -4.60847825e-01 1.05785167e+00
-3.25595289e-01 4.34724927e-01 -1.24995671e-02 1.33096026e-02
-4.40115392e-01 7.56749138e-02 6.90812409e-01 2.69668370e-01
-8.26595277e-02 -2.51939863e-01 -8.06130707e-01 1.05551648e+00
7.79180229e-01 -7.70079911e-01 3.02129090e-01 -5.76840818e-01
4.40160483e-01 -1.05182827e-01 -1.67258546e-01 -6.58435166e-01
-1.29172474e-01 -2.01410607e-01 5.46489716e-01 -2.61690885e-01
9.55183953e-02 -7.09946692e-01 -4.93625179e-02 1.99451417e-01
-2.42281497e-01 3.46914977e-01 2.33079717e-01 -5.46194874e-02
-1.64051175e-01 -7.02728748e-01 7.61144400e-01 -3.00730914e-01
-8.13142776e-01 1.83118597e-01 -3.39686692e-01 3.60104740e-01
5.47474325e-01 -2.35092100e-02 -6.24883845e-02 3.48356783e-01
-7.32706368e-01 -2.75829047e-01 5.22872329e-01 4.29984719e-01
-9.53764468e-02 -1.42668688e+00 -6.96218610e-01 2.86964357e-01
2.90673256e-01 -4.89718258e-01 -1.78378567e-01 4.06994998e-01
-1.39369339e-01 1.92943513e-01 -2.70429820e-01 1.62524916e-03
-1.16753805e+00 4.48677301e-01 1.38774619e-01 -5.93399286e-01
-1.97538227e-01 1.06840575e+00 2.80768722e-01 -8.29774559e-01
-2.02556580e-01 6.84691146e-02 -4.84691590e-01 8.69873911e-02
1.43977672e-01 -2.59057909e-01 1.37062788e-01 -1.36140251e+00
-5.68717003e-01 3.90329003e-01 -1.21177539e-01 -1.25594791e-02
1.24530101e+00 -7.67727494e-02 -2.99945891e-01 6.51482463e-01
1.05057287e+00 8.34603667e-01 -8.42388690e-01 -4.19988185e-01
6.01825476e-01 -2.32237950e-02 -2.41281733e-01 -7.61434138e-01
-6.89693451e-01 6.76123857e-01 1.61687106e-01 1.12926051e-01
5.99409640e-01 3.38380992e-01 7.63405323e-01 1.14843830e-01
4.59092200e-01 -1.39437771e+00 -5.82801759e-01 9.36771572e-01
2.57963449e-01 -9.37372029e-01 -1.64061561e-01 -6.59461439e-01
-7.15908229e-01 9.78806317e-01 5.52099645e-01 -3.63941640e-02
5.95431805e-01 6.32452190e-01 2.19753206e-01 1.21640041e-02
-4.56257612e-01 -5.54401577e-01 3.97858620e-01 7.07571328e-01
1.18205750e+00 6.01229250e-01 -7.01730013e-01 9.50328231e-01
-5.93214452e-01 -6.90332055e-01 2.01529324e-01 7.71971762e-01
-2.98989743e-01 -1.63995004e+00 -1.10481046e-01 2.60847807e-01
-1.25258064e+00 -6.90139830e-01 -3.70418757e-01 9.89663780e-01
8.17918360e-01 8.05568874e-01 3.30347389e-01 -7.80224949e-02
-3.37058865e-02 4.96016473e-01 6.28382266e-01 -1.14836967e+00
-8.41761827e-01 1.34179279e-01 6.80679381e-01 -3.22130769e-02
-4.59356457e-01 -1.01559055e+00 -1.29422557e+00 1.89535752e-01
-4.03035462e-01 2.89353341e-01 4.83439952e-01 1.13114917e+00
-5.27380146e-02 6.77197054e-02 3.02273214e-01 -4.16493952e-01
-2.92756200e-01 -1.04150486e+00 -9.26479876e-01 5.74603975e-01
-2.99869478e-01 -7.18923032e-01 -1.03759199e-01 2.88970828e-01] | [10.434612274169922, 10.02780818939209] |
1b2ef3a5-249d-4ff1-8cce-4523508b77cd | gan-q-learning | 1805.04874 | null | http://arxiv.org/abs/1805.04874v3 | http://arxiv.org/pdf/1805.04874v3.pdf | GAN Q-learning | Distributional reinforcement learning (distributional RL) has seen empirical
success in complex Markov Decision Processes (MDPs) in the setting of nonlinear
function approximation. However, there are many different ways in which one can
leverage the distributional approach to reinforcement learning. In this paper,
we propose GAN Q-learning, a novel distributional RL method based on generative
adversarial networks (GANs) and analyze its performance in simple tabular
environments, as well as OpenAI Gym. We empirically show that our algorithm
leverages the flexibility and blackbox approach of deep learning models while
providing a viable alternative to traditional methods. | ['Thang Doan', 'Clare Lyle', 'Bogdan Mazoure'] | 2018-05-13 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-1.51177347e-01 2.63186604e-01 -2.27667972e-01 -1.64043397e-01
-1.11265600e+00 -8.06828260e-01 8.16426396e-01 -3.70226622e-01
-3.65146816e-01 1.25415349e+00 3.73914570e-01 -7.42275476e-01
2.32496276e-03 -9.94272053e-01 -5.63242018e-01 -8.36449265e-01
8.82568806e-02 7.49962926e-01 -2.90591300e-01 -2.58611530e-01
8.82447585e-02 5.22258401e-01 -9.15199518e-01 -1.23950653e-01
7.32790112e-01 8.45334470e-01 -4.87429768e-01 9.32801962e-01
-2.70060062e-01 1.39278376e+00 -9.13929045e-01 -5.81394017e-01
3.80934387e-01 -5.11182666e-01 -5.49398124e-01 -4.75850403e-01
1.77219316e-01 -6.39587760e-01 -5.70862055e-01 8.19332302e-01
8.23456526e-01 3.56200367e-01 1.06935525e+00 -1.43237150e+00
-8.53793204e-01 5.99443555e-01 -4.28879887e-01 1.58738017e-01
2.37275586e-01 7.50898421e-01 9.84220326e-01 -3.53594512e-01
3.08856279e-01 1.60162926e+00 7.03621566e-01 8.24250996e-01
-1.49437892e+00 -7.31780350e-01 -5.28357439e-02 -2.69330949e-01
-7.89796233e-01 -1.03712585e-02 5.09302497e-01 -3.82221937e-01
8.85825872e-01 -3.04102153e-01 4.90274787e-01 1.52843833e+00
5.14017105e-01 1.04974592e+00 1.72046375e+00 -2.47543976e-01
7.36469388e-01 -2.55942017e-01 -4.01348025e-01 3.82232100e-01
-1.71540976e-01 7.16116786e-01 -1.59471393e-01 -4.18971837e-01
1.03140807e+00 9.51586589e-02 2.55158007e-01 -5.08204937e-01
-7.68596828e-01 1.34846616e+00 1.71623260e-01 -2.36271143e-01
-4.71040159e-01 8.46598029e-01 4.07837927e-01 5.24056911e-01
4.60204840e-01 3.68948042e-01 -3.92888546e-01 -6.68221116e-01
-7.96464145e-01 7.53621340e-01 9.89796460e-01 9.11891699e-01
5.60045540e-01 7.99844205e-01 -6.49714530e-01 5.17241836e-01
2.71063149e-01 8.03774774e-01 3.84307534e-01 -1.17503810e+00
3.78831536e-01 -1.00241587e-01 3.27304274e-01 -2.94246167e-01
-1.16398416e-01 -2.03138366e-01 -7.60820389e-01 6.87681139e-01
5.42988896e-01 -7.69950926e-01 -1.00705063e+00 1.76791358e+00
1.94306746e-01 -2.71628201e-02 3.28628361e-01 3.70496452e-01
2.77425617e-01 5.98135054e-01 3.34239602e-01 1.33117870e-03
6.34455442e-01 -7.84874678e-01 -4.90047604e-01 2.09164470e-01
1.70260280e-01 -4.46568012e-01 1.36629224e+00 4.22448993e-01
-1.18628621e+00 -2.74963498e-01 -7.91917324e-01 1.75829306e-01
-2.04582170e-01 -6.05820477e-01 6.89431012e-01 9.55994606e-01
-1.19494474e+00 7.86247313e-01 -8.55624795e-01 1.36299685e-01
7.31519401e-01 3.17335308e-01 2.10978046e-01 -1.22476548e-01
-1.17993832e+00 8.91778827e-01 6.76873997e-02 -3.93494248e-01
-1.43246293e+00 -6.13758385e-01 -6.63006842e-01 5.35563305e-02
4.68457073e-01 -8.45731199e-01 1.79818773e+00 -9.39890862e-01
-2.17756414e+00 1.18643418e-01 2.37149402e-01 -8.91420245e-01
9.54624236e-01 -2.80970812e-01 -4.04260233e-02 6.44409135e-02
-2.63042778e-01 7.01768756e-01 1.16915560e+00 -8.90549541e-01
-3.95825356e-01 -1.55176759e-01 2.78695315e-01 1.38565987e-01
1.74910039e-01 -1.62557811e-01 4.85405058e-01 -8.71586502e-01
-8.06200683e-01 -9.10809577e-01 -5.80869675e-01 -4.37468529e-01
-2.64809221e-01 -4.96409237e-01 5.11928678e-01 -4.18050528e-01
9.65414345e-01 -2.03758860e+00 5.48262708e-02 2.35361889e-01
2.22200885e-01 1.04134366e-01 -2.33415943e-02 8.40551198e-01
2.40825146e-01 2.96143264e-01 -3.36089551e-01 -2.67468482e-01
6.07750416e-01 5.36704421e-01 -6.40601635e-01 2.44864851e-01
2.76042879e-01 1.46401000e+00 -9.83838916e-01 -2.66963363e-01
3.14120539e-02 1.74658895e-01 -5.51239252e-01 3.62300068e-01
-6.78198278e-01 4.59799886e-01 -2.75834978e-01 6.72990322e-01
3.35392535e-01 1.33083090e-01 1.01012781e-01 6.23359203e-01
1.51109204e-01 1.43615872e-01 -9.76184249e-01 1.31215537e+00
-4.57734942e-01 4.41308349e-01 -3.07485878e-01 -8.14356387e-01
7.84974813e-01 2.30889603e-01 4.35019463e-01 -5.91410160e-01
-1.75201464e-02 4.49511688e-03 -1.38381077e-03 -1.18329488e-01
2.50397831e-01 -8.22307169e-01 -1.82558089e-01 8.00060511e-01
2.19584361e-01 -3.17464232e-01 3.94062623e-02 -2.80083641e-02
1.37333894e+00 6.83289289e-01 4.89170104e-01 -1.61312610e-01
-2.20825225e-01 -1.42843023e-01 2.97055691e-01 1.14544499e+00
-3.18085939e-01 2.60002196e-01 9.07658577e-01 -4.12675202e-01
-1.21858096e+00 -1.57243800e+00 2.44702950e-01 1.04386628e+00
-5.54700553e-01 -2.49426275e-01 -6.63067877e-01 -1.05612028e+00
3.79210591e-01 1.01516199e+00 -8.98514509e-01 -1.84702829e-01
-5.67718387e-01 -5.02623916e-01 1.02235019e+00 1.06330824e+00
2.30284825e-01 -1.48159659e+00 -6.26539707e-01 4.07325745e-01
4.52277154e-01 -5.84947884e-01 -3.40319306e-01 5.64402223e-01
-8.54650438e-01 -7.42892206e-01 -8.77911627e-01 -2.06053883e-01
7.80664161e-02 -3.25414628e-01 1.46549296e+00 -8.32108676e-01
1.29981831e-01 8.40154946e-01 -1.47578299e-01 -7.05610573e-01
-9.09320891e-01 1.58470958e-01 5.77555457e-03 -4.56072539e-01
2.88900912e-01 -7.19671667e-01 -6.04254723e-01 -7.53306597e-02
-1.10117447e+00 -4.71496582e-01 7.01243699e-01 1.16122949e+00
5.39421737e-01 -8.82096365e-02 9.87983763e-01 -1.05173588e+00
1.20976472e+00 -6.98064744e-01 -7.73643136e-01 -2.60457937e-02
-8.39927971e-01 4.53966707e-01 8.74724209e-01 -4.72199053e-01
-1.09829092e+00 -4.52595294e-01 -3.88588011e-01 -6.12146199e-01
-3.10970426e-01 3.72605085e-01 9.15202796e-02 2.78120995e-01
6.15615308e-01 1.52257264e-01 2.69722879e-01 -2.07371339e-01
7.26698279e-01 2.76369661e-01 2.70486265e-01 -8.55814815e-01
1.00245094e+00 2.42648110e-01 2.44605854e-01 -3.72211456e-01
-6.62629962e-01 1.72399908e-01 -1.57578930e-01 -2.40992662e-02
7.85083294e-01 -7.54052043e-01 -8.65896046e-01 2.71167904e-01
-5.18103123e-01 -1.08854401e+00 -9.99416947e-01 2.67118126e-01
-1.24701965e+00 -5.83543219e-02 -7.28599668e-01 -1.01930964e+00
-3.88387144e-01 -8.99638116e-01 8.51832867e-01 2.05556899e-01
-1.53083965e-01 -1.35096514e+00 4.77497697e-01 -2.45939001e-01
6.71537101e-01 5.96292078e-01 1.14001691e+00 -6.89402044e-01
-3.43562722e-01 2.38822937e-01 2.42883638e-01 8.09850216e-01
2.10062750e-02 2.90062781e-02 -8.52651238e-01 -3.44742149e-01
-1.05284579e-01 -7.53628433e-01 7.16701150e-01 6.46821737e-01
1.02933216e+00 -5.27966678e-01 2.87231892e-01 5.50304234e-01
1.50525618e+00 2.95250446e-01 7.80895054e-01 3.05744767e-01
4.90226656e-01 2.01386347e-01 4.47636187e-01 6.08876467e-01
2.39910588e-01 1.09151624e-01 3.55506331e-01 -1.06684335e-01
1.87816039e-01 -7.58130014e-01 7.07340896e-01 3.52402657e-01
1.18258908e-01 -1.26929522e-01 -8.64660680e-01 2.01973125e-01
-1.68129241e+00 -1.17692375e+00 5.70649087e-01 1.99581647e+00
9.83765125e-01 3.03554058e-01 7.00140059e-01 -2.06494302e-01
1.79406404e-01 7.37785175e-02 -9.60368335e-01 -1.00125921e+00
-1.12498373e-01 9.53255296e-01 5.68548024e-01 1.73811853e-01
-9.52196658e-01 9.72643733e-01 7.71967030e+00 9.43181276e-01
-7.55664527e-01 1.57289162e-01 8.03943932e-01 -2.18664736e-01
-4.78119701e-01 -2.60572553e-01 -6.32221580e-01 4.57128257e-01
1.18104506e+00 -1.47669226e-01 7.38954604e-01 9.73322511e-01
5.04982211e-02 1.06256902e-01 -1.22094047e+00 7.84142971e-01
-3.94830197e-01 -1.23829818e+00 2.97524277e-02 2.77637899e-01
1.13933945e+00 8.92004445e-02 6.21474504e-01 9.45038080e-01
1.36746502e+00 -1.61804390e+00 5.85374594e-01 5.33075154e-01
8.62426877e-01 -1.36713052e+00 6.68614984e-01 3.40623528e-01
-5.28424680e-01 -3.41078609e-01 -2.80668706e-01 -8.41278732e-02
-6.33464009e-02 8.56124908e-02 -7.92207241e-01 2.98197061e-01
4.54467237e-01 4.45906788e-01 -1.00865178e-01 5.61138272e-01
-4.08617586e-01 1.16136134e+00 -1.47997171e-01 -7.89738521e-02
7.24572182e-01 -2.74153918e-01 3.39933664e-01 1.05262196e+00
1.32066056e-01 -2.21316144e-01 1.97643787e-01 9.64374006e-01
-1.33403212e-01 -4.67685722e-02 -9.97872353e-01 -2.92421967e-01
1.67560905e-01 8.63003910e-01 -2.49442190e-01 -2.41224900e-01
-4.54912782e-01 6.75927401e-01 4.27584857e-01 6.55876637e-01
-1.01441467e+00 -1.03734367e-01 8.74765575e-01 -6.02080673e-02
5.97113490e-01 -4.00724977e-01 -1.32654116e-01 -9.92954791e-01
-3.07832509e-01 -1.28777421e+00 3.35734278e-01 -4.89691883e-01
-1.82933021e+00 2.24194482e-01 2.36486048e-02 -1.02121687e+00
-9.43155348e-01 -5.04310071e-01 -4.77416426e-01 9.32562292e-01
-1.53165233e+00 -1.17837477e+00 2.79515117e-01 7.76655376e-01
4.32693928e-01 -4.43917215e-01 8.93812180e-01 -2.78403342e-01
-1.68160394e-01 7.02323437e-01 7.85561442e-01 7.72102922e-02
6.15994513e-01 -1.96229231e+00 5.65333664e-01 3.29270273e-01
2.53939033e-01 3.78043383e-01 6.40148818e-01 -3.91663790e-01
-1.35860646e+00 -9.97470617e-01 -2.02306449e-01 -4.44060951e-01
8.19184661e-01 -2.38840744e-01 -4.77118701e-01 8.69123995e-01
5.18407285e-01 -3.35693330e-01 8.53203773e-01 -7.86048174e-02
-3.18281293e-01 5.24797514e-02 -1.29381633e+00 7.92938173e-01
6.18659437e-01 -4.90040153e-01 -4.91394281e-01 6.86811060e-02
7.86707759e-01 -3.32978845e-01 -1.05466568e+00 9.30167586e-02
4.35172558e-01 -8.51652980e-01 7.45675504e-01 -1.02410316e+00
6.75205886e-01 6.94955066e-02 -2.90259749e-01 -1.91062307e+00
-5.41031994e-02 -1.21604979e+00 -6.22578323e-01 1.09136450e+00
8.57802182e-02 -9.40648079e-01 7.12290406e-01 4.30661023e-01
1.68637380e-01 -9.74590421e-01 -9.50882733e-01 -1.01886868e+00
9.96611655e-01 -3.25771004e-01 7.76149452e-01 4.44339275e-01
-4.20020044e-01 1.44329339e-01 -3.77815634e-01 -4.97234941e-01
7.11806715e-01 5.04304357e-02 1.01913738e+00 -9.01088059e-01
-7.76530981e-01 -5.88916540e-01 -8.19776207e-02 -1.04102302e+00
4.61344987e-01 -7.30557680e-01 -7.31229857e-02 -1.36896300e+00
-2.12389622e-02 -4.61088389e-01 -4.90441144e-01 4.10552919e-01
-1.36100709e-01 -2.49888487e-02 3.07093620e-01 5.47797717e-02
-4.06590164e-01 9.54144597e-01 1.26000404e+00 -1.02518529e-01
-4.93811443e-02 3.14128131e-01 -7.87429810e-01 5.78829646e-01
1.15648615e+00 -5.09757876e-01 -5.87147534e-01 -5.11399582e-02
2.58601099e-01 2.41808340e-01 3.06198895e-01 -8.19774806e-01
-3.07839781e-01 -4.60611999e-01 4.82730508e-01 -4.25587833e-01
5.41989543e-02 -5.18645525e-01 -1.94655374e-01 4.29364055e-01
-5.67927063e-01 5.74273467e-01 2.24594250e-01 8.80224645e-01
-7.03825280e-02 -1.20929427e-01 7.95467913e-01 -2.86792815e-01
-3.59766126e-01 4.47414279e-01 -5.92736423e-01 6.72778070e-01
1.09477913e+00 9.86422524e-02 -2.27423862e-01 -8.67614627e-01
-6.49519086e-01 1.38597339e-01 4.84776914e-01 6.20797575e-02
5.61754942e-01 -1.33561730e+00 -7.31109798e-01 -6.97435364e-02
-2.55751401e-01 -6.26854897e-02 -8.78636166e-02 5.63842535e-01
-3.77181441e-01 2.36939505e-01 -4.12581056e-01 -2.79524297e-01
-4.26657945e-01 8.03966641e-01 3.98957253e-01 -7.96993434e-01
-5.29309690e-01 4.63228852e-01 1.27959281e-01 -3.80616367e-01
1.46325022e-01 -3.76411736e-01 1.86600879e-01 -1.08939104e-01
2.87903935e-01 5.21213710e-01 -3.77026021e-01 4.12868932e-02
1.95834383e-01 -1.08293436e-01 6.37174621e-02 -6.03960931e-01
1.34797096e+00 2.20257565e-01 3.51076931e-01 7.90518105e-01
7.83085287e-01 -5.32318279e-02 -2.04914594e+00 -5.14592417e-02
-5.02651855e-02 -2.03929096e-01 -2.51229763e-01 -1.03015721e+00
-8.17265868e-01 1.09088933e+00 4.97091979e-01 3.16098720e-01
8.83755505e-01 -2.55527258e-01 7.90483236e-01 3.37182373e-01
4.38135803e-01 -1.03944075e+00 4.25602376e-01 6.54456258e-01
5.96762419e-01 -1.05527365e+00 -2.47930676e-01 5.55560648e-01
-1.07570171e+00 9.89750624e-01 4.78251874e-01 -5.48181713e-01
5.35999238e-01 6.54785037e-01 1.82794761e-02 1.81916475e-01
-9.40492392e-01 -2.98845261e-01 -2.09804252e-01 1.04734814e+00
1.41652986e-01 3.22356999e-01 2.21820042e-01 3.21807772e-01
-2.74635166e-01 -1.32198976e-02 4.65629846e-01 1.09352911e+00
-8.21092725e-02 -1.31205404e+00 -1.14443682e-01 4.84628648e-01
-8.13270748e-01 -3.82460773e-01 -1.35104567e-01 9.77189720e-01
-4.06858236e-01 7.72018194e-01 -8.09889883e-02 -5.85313328e-02
1.35850888e-02 3.25116366e-01 6.81184649e-01 -4.91663337e-01
-8.40088546e-01 3.09416484e-02 -1.74839810e-01 -4.67551708e-01
-1.40841171e-01 -7.54027188e-01 -1.01589167e+00 -5.23971438e-01
1.96449891e-01 -7.86106959e-02 3.84863794e-01 8.52779865e-01
1.77434102e-01 6.13227367e-01 6.41930997e-01 -5.14654577e-01
-1.50106013e+00 -9.98216391e-01 -7.71513045e-01 2.34882176e-01
2.88468510e-01 -7.60336518e-01 -2.39464924e-01 -3.17597419e-01] | [4.0774617195129395, 2.564365863800049] |
57ee820c-5f72-4ab2-a431-3a95bee53acc | self-supervised-few-shot-learning-on-point | 2009.14168 | null | https://arxiv.org/abs/2009.14168v1 | https://arxiv.org/pdf/2009.14168v1.pdf | Self-Supervised Few-Shot Learning on Point Clouds | The increased availability of massive point clouds coupled with their utility in a wide variety of applications such as robotics, shape synthesis, and self-driving cars has attracted increased attention from both industry and academia. Recently, deep neural networks operating on labeled point clouds have shown promising results on supervised learning tasks like classification and segmentation. However, supervised learning leads to the cumbersome task of annotating the point clouds. To combat this problem, we propose two novel self-supervised pre-training tasks that encode a hierarchical partitioning of the point clouds using a cover-tree, where point cloud subsets lie within balls of varying radii at each level of the cover-tree. Furthermore, our self-supervised learning network is restricted to pre-train on the support set (comprising of scarce training examples) used to train the downstream network in a few-shot learning (FSL) setting. Finally, the fully-trained self-supervised network's point embeddings are input to the downstream task's network. We present a comprehensive empirical evaluation of our method on both downstream classification and segmentation tasks and show that supervised methods pre-trained with our self-supervised learning method significantly improve the accuracy of state-of-the-art methods. Additionally, our method also outperforms previous unsupervised methods in downstream classification tasks. | ['Charu Sharma', 'Manohar Kaul'] | 2020-09-29 | null | http://proceedings.neurips.cc/paper/2020/hash/50c1f44e426560f3f2cdcb3e19e39903-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/50c1f44e426560f3f2cdcb3e19e39903-Paper.pdf | neurips-2020-12 | ['few-shot-3d-point-cloud-classification'] | ['computer-vision'] | [ 1.07191376e-01 1.63353652e-01 -2.69274414e-01 -5.50138772e-01
-7.73911297e-01 -4.95736659e-01 5.97251296e-01 3.33636671e-01
-3.56931239e-01 2.95075893e-01 -2.65655190e-01 -2.52710611e-01
1.01186074e-01 -9.45982814e-01 -9.34335291e-01 -5.84108949e-01
-2.62322351e-02 8.67402315e-01 5.77405691e-01 -1.09992102e-01
1.60414994e-01 6.70576036e-01 -1.90008569e+00 2.26413235e-02
9.92074311e-01 1.18674529e+00 2.94986010e-01 3.55262429e-01
-3.92499119e-01 1.93115890e-01 -3.88265848e-01 -3.18864822e-01
4.67380017e-01 1.03414759e-01 -7.64332116e-01 3.41507107e-01
3.05030823e-01 -8.92911330e-02 -2.31640220e-01 1.01287377e+00
9.12923440e-02 4.37158763e-01 7.99873412e-01 -1.50223482e+00
-6.58026636e-01 1.74008340e-01 -4.69478786e-01 3.52244265e-02
-3.84720236e-01 3.39693308e-01 1.04547036e+00 -1.15276301e+00
3.78950983e-01 9.11201894e-01 8.26495528e-01 3.53038698e-01
-1.15455163e+00 -6.70069456e-01 7.56267905e-02 1.18203521e-01
-1.41829801e+00 -2.75021881e-01 8.34013820e-01 -5.51899314e-01
9.85194504e-01 -1.53152674e-01 7.78222680e-01 6.57093406e-01
-2.73157805e-01 8.12848687e-01 4.89663839e-01 -1.52858719e-01
6.47346675e-01 1.84803158e-01 1.15579568e-01 6.54002368e-01
1.33412644e-01 1.30313620e-01 -1.27703801e-01 -2.41860896e-01
7.65562713e-01 3.04412007e-01 3.06664646e-01 -8.41191947e-01
-9.51606035e-01 1.14016259e+00 7.87446976e-01 1.06626779e-01
-2.08058327e-01 2.58751452e-01 4.16248351e-01 -7.33872876e-02
9.15501297e-01 4.41382200e-01 -5.19315422e-01 2.45626252e-02
-1.04287004e+00 4.41788137e-02 6.01511419e-01 1.44478309e+00
1.13223875e+00 -1.67424396e-01 2.09412396e-01 8.83958578e-01
3.81216466e-01 3.12386245e-01 2.22724110e-01 -8.58303130e-01
3.77656013e-01 8.39283168e-01 8.05242062e-02 -6.20885491e-01
-2.38784820e-01 -4.56705362e-01 -7.26059198e-01 2.93342292e-01
1.72714353e-01 6.40427768e-02 -1.18942356e+00 1.21911561e+00
5.95980823e-01 6.13276780e-01 1.35054231e-01 8.99781227e-01
6.55565619e-01 8.10256720e-01 1.10938966e-01 1.90634966e-01
9.47813690e-01 -1.19235396e+00 -3.17016132e-02 -3.06855023e-01
9.64145958e-01 -3.56580347e-01 9.64236796e-01 9.66184363e-02
-7.11701274e-01 -7.16422975e-01 -1.09585953e+00 -1.47943541e-01
-5.75325549e-01 1.67853862e-01 7.41305709e-01 2.63335735e-01
-8.95202935e-01 8.44425797e-01 -1.05644071e+00 -3.24578464e-01
1.17222917e+00 4.67204779e-01 -2.58445740e-01 -1.99902073e-01
-7.78485239e-01 6.82745039e-01 2.63154209e-01 -8.19771513e-02
-7.63759613e-01 -8.74480367e-01 -1.06195331e+00 1.29506722e-01
2.76592880e-01 -5.47807574e-01 1.18841684e+00 -7.43770301e-01
-1.29917037e+00 1.00388026e+00 -1.41911417e-01 -6.10399723e-01
2.98550546e-01 -1.87904894e-01 -4.19232361e-02 1.88100085e-01
4.89698768e-01 1.07142115e+00 8.88053298e-01 -1.38870597e+00
-8.11051130e-01 -4.05339897e-01 -1.18415676e-01 1.56785741e-01
-1.03262447e-01 -4.54245269e-01 -6.06285334e-01 -2.69340068e-01
1.44687861e-01 -1.01390886e+00 -5.96758842e-01 2.49947682e-01
-4.58620846e-01 -7.69841790e-01 1.14893806e+00 1.82074547e-01
4.98220116e-01 -2.38315368e+00 -1.29398733e-01 6.73509315e-02
2.97085375e-01 3.53814423e-01 -2.04647809e-01 3.00262809e-01
-8.09921920e-02 3.72084230e-02 -5.68895102e-01 -7.62327552e-01
-4.44295891e-02 3.87577921e-01 -4.65416551e-01 7.68070519e-01
6.65220916e-01 1.03540945e+00 -1.10144997e+00 -4.03944135e-01
6.51376069e-01 2.32945606e-01 -4.76315469e-01 2.24001572e-01
-5.23852587e-01 5.02799928e-01 -3.49187642e-01 6.14750564e-01
6.30668879e-01 -5.24557173e-01 -6.04005456e-01 7.80408233e-02
-7.44449720e-03 2.18173772e-01 -7.75718153e-01 2.01324344e+00
-4.94196385e-01 5.74962974e-01 -3.26343119e-01 -1.19103575e+00
1.08148026e+00 -3.67310569e-02 7.22964764e-01 -2.71447062e-01
1.11647867e-01 1.93573743e-01 -1.90440565e-01 -3.35758239e-01
2.81641603e-01 -3.41450334e-01 -1.60915419e-01 2.78428376e-01
2.21198946e-01 -6.27450883e-01 -1.93284694e-02 1.42563894e-01
9.47506905e-01 1.44844934e-01 -9.76577848e-02 -4.60231043e-02
2.42224693e-01 5.72931111e-01 4.52045232e-01 4.85259533e-01
-2.31152505e-01 8.36140811e-01 2.37091646e-01 -3.70661587e-01
-1.27628911e+00 -9.76642370e-01 -3.03842366e-01 1.08118010e+00
4.70540375e-01 -2.30055094e-01 -5.04076660e-01 -8.31002295e-01
4.20313865e-01 7.65874982e-01 -5.54046273e-01 -2.87875772e-01
-1.86547816e-01 -2.31931195e-01 3.86751980e-01 8.89214635e-01
2.70349503e-01 -9.13169861e-01 -4.84627724e-01 1.19229466e-01
4.68147278e-01 -1.32100213e+00 -1.35218799e-01 3.60497713e-01
-1.15044653e+00 -1.02640080e+00 -4.93155241e-01 -9.05816138e-01
8.70696545e-01 5.79270303e-01 1.01500893e+00 1.70377895e-01
-2.17856675e-01 2.46417165e-01 -4.00296628e-01 -7.10361183e-01
4.78252806e-02 3.42796296e-01 3.75758037e-02 -2.72318888e-02
8.68969917e-01 -6.35255814e-01 -4.57074791e-01 2.36450762e-01
-7.48266220e-01 -5.58028854e-02 4.29102689e-01 7.15568125e-01
8.05637240e-01 -4.06525098e-02 5.86501837e-01 -9.74817336e-01
3.51462699e-02 -7.25168943e-01 -8.07156324e-01 -3.25219065e-01
-2.40033239e-01 -1.57353953e-02 5.56371391e-01 -2.57481396e-01
-5.50856888e-01 4.24181908e-01 -2.97476470e-01 -1.17358923e+00
-4.04392868e-01 2.00894296e-01 2.40608696e-02 -1.78157482e-02
6.14291608e-01 -2.95185298e-02 1.33313477e-01 -4.08582777e-01
6.38075769e-01 7.86321104e-01 4.97212619e-01 -3.61508608e-01
1.11844909e+00 8.30267429e-01 4.75786477e-02 -9.29312229e-01
-1.24961531e+00 -9.53328192e-01 -1.14303148e+00 -4.14713053e-03
9.59799707e-01 -1.00507307e+00 -4.78585333e-01 2.43677244e-01
-1.21556377e+00 -5.00733376e-01 -6.85217083e-01 2.99002111e-01
-7.43682861e-01 1.75871849e-01 -2.56386548e-01 -8.20188224e-01
-8.14464912e-02 -1.07269645e+00 1.54937959e+00 1.06070571e-01
-1.30972303e-02 -1.04022491e+00 -5.03905229e-02 2.56415904e-01
4.35086004e-02 2.22355440e-01 8.69209707e-01 -1.04625964e+00
-6.00469053e-01 -3.86118025e-01 -2.83479035e-01 5.03551364e-01
1.45936683e-01 -1.88649222e-01 -1.06126678e+00 -1.45033374e-01
-1.63601756e-01 -5.55392206e-01 9.41938043e-01 2.46700197e-01
1.43236768e+00 8.76758322e-02 -6.99767888e-01 7.40129232e-01
1.39554119e+00 -1.11219630e-01 3.37004632e-01 8.92776996e-02
8.95224154e-01 7.90251732e-01 8.55213821e-01 1.59209237e-01
3.09297830e-01 3.48016411e-01 6.72818780e-01 -2.71078378e-01
2.07602963e-01 -3.59541416e-01 -2.66134501e-01 5.27405977e-01
2.63750702e-01 6.58762679e-02 -1.13389874e+00 7.95924604e-01
-2.07237244e+00 -6.13762200e-01 -1.14785686e-01 2.15134168e+00
6.19253218e-01 3.34878594e-01 1.12932667e-01 2.87470818e-02
7.28892982e-01 1.53364584e-01 -9.10834432e-01 -2.31145788e-03
3.33611757e-01 2.79450357e-01 6.32430434e-01 1.90379471e-01
-1.39920962e+00 1.32566285e+00 5.77349901e+00 7.03526199e-01
-1.18082213e+00 2.58367389e-01 5.62360227e-01 -1.24212883e-01
-1.42335996e-01 6.38665166e-04 -7.90608346e-01 5.23322999e-01
9.41338122e-01 2.37494800e-02 -6.93759788e-03 1.23677719e+00
1.18582752e-02 -1.61137544e-02 -1.37432528e+00 1.03027189e+00
-1.77015021e-01 -1.68728805e+00 -2.46415421e-01 1.47346050e-01
8.63134921e-01 7.22446203e-01 7.17274239e-03 5.89175045e-01
4.65947360e-01 -1.07634723e+00 5.30044794e-01 1.85326919e-01
1.09612095e+00 -9.22762990e-01 6.91173255e-01 7.27834404e-01
-1.16071689e+00 4.04808968e-02 -7.88874686e-01 -6.16044328e-02
7.54895508e-02 7.82991290e-01 -8.91479611e-01 3.26242715e-01
6.25306606e-01 1.00335598e+00 -2.96099514e-01 1.09810102e+00
-1.11041218e-01 5.76130450e-01 -6.02228045e-01 5.89385703e-02
6.18655682e-01 -4.04370636e-01 4.49833781e-01 1.06764495e+00
1.00435056e-01 2.55455971e-01 4.17270094e-01 1.13034582e+00
-1.45014212e-01 -1.77758485e-01 -1.00250995e+00 -1.02944367e-01
6.39134705e-01 1.18567002e+00 -8.67148638e-01 -4.64567661e-01
-4.43825066e-01 6.41556203e-01 5.76258123e-01 1.92913309e-01
-5.78272402e-01 -5.04038215e-01 8.80128384e-01 2.40538299e-01
6.24075294e-01 -4.82471526e-01 -6.14234388e-01 -9.01391685e-01
-1.96183279e-01 -4.78819907e-02 -2.72406191e-02 -7.46398866e-01
-1.58221388e+00 4.47116524e-01 6.56144619e-02 -1.51743126e+00
-1.42812401e-01 -5.75216770e-01 -8.74256909e-01 7.96356261e-01
-1.60477710e+00 -1.15410709e+00 -3.29644799e-01 3.59707206e-01
7.22566128e-01 -1.06482409e-01 5.34607410e-01 1.99981928e-02
-4.01177555e-01 1.61574662e-01 9.92995948e-02 3.08440298e-01
3.01690906e-01 -1.32679796e+00 6.80683255e-01 6.76189423e-01
1.70276701e-01 5.14194250e-01 3.21194828e-01 -6.89504981e-01
-1.32432640e+00 -1.74793935e+00 4.91508603e-01 -5.41572154e-01
6.03237867e-01 -6.03861570e-01 -1.08038938e+00 6.91098928e-01
-3.28986347e-01 6.74502730e-01 6.89404368e-01 9.42942202e-02
-2.40817025e-01 -1.07339293e-01 -1.14684868e+00 2.51280814e-01
1.14101160e+00 -5.00685275e-01 -7.33008802e-01 6.77811742e-01
1.03134692e+00 -4.39966649e-01 -6.27430737e-01 4.32234973e-01
-1.46686614e-01 -6.59352720e-01 8.80264580e-01 -6.98486567e-01
6.63105428e-01 -1.85553327e-01 5.67637533e-02 -1.38364148e+00
-2.33017713e-01 -2.79154271e-01 -1.06000707e-01 9.62575614e-01
2.19863817e-01 -5.70095003e-01 1.30186474e+00 4.44297850e-01
-5.91526091e-01 -1.08916593e+00 -9.89645898e-01 -7.45011449e-01
1.63102090e-01 -7.24102139e-01 6.10818684e-01 8.18169713e-01
-2.73457378e-01 4.78802711e-01 3.31094265e-01 3.58951688e-01
7.78748155e-01 3.62695396e-01 9.55749094e-01 -1.66073453e+00
1.27728656e-01 -2.33598709e-01 -6.78917587e-01 -1.28755438e+00
4.01451916e-01 -1.15921664e+00 3.72672796e-01 -1.73868251e+00
-1.67357296e-01 -1.00401807e+00 6.81223348e-02 5.58358729e-01
2.76159626e-02 2.66914994e-01 1.02122873e-01 5.30770719e-01
-7.06794620e-01 8.37620497e-01 9.43893909e-01 -2.78388917e-01
-2.97242135e-01 3.39096129e-01 -4.35858548e-01 7.66245544e-01
5.96457660e-01 -6.77009642e-01 -4.22122091e-01 -4.25466239e-01
-5.46028242e-02 -2.62098849e-01 3.92113507e-01 -1.05958140e+00
3.83892268e-01 -6.77206963e-02 2.52487510e-01 -9.88943458e-01
5.57720125e-01 -9.55082357e-01 -3.68905067e-01 2.11963236e-01
-7.23260343e-02 -3.09070498e-01 2.63772786e-01 8.98349583e-01
-1.75769031e-01 -3.67668927e-01 9.26880240e-01 -7.95605853e-02
-7.76980519e-01 7.86000311e-01 4.55169193e-02 -1.32811621e-01
1.38139343e+00 -4.98894274e-01 1.40158311e-01 -3.89860608e-02
-6.52880847e-01 5.38923979e-01 5.88968158e-01 4.26865250e-01
6.65003538e-01 -1.29349387e+00 -3.94441575e-01 2.52393335e-01
4.87933964e-01 1.14353788e+00 -1.15280755e-01 6.09282970e-01
-4.89661992e-01 4.04718161e-01 1.21243253e-01 -1.24614656e+00
-6.77155614e-01 7.18171120e-01 1.01611376e-01 1.21117584e-01
-9.30889606e-01 9.99801040e-01 2.09870562e-01 -6.83940232e-01
2.27437571e-01 -4.38242942e-01 -8.06375220e-02 -6.58559725e-02
6.10404238e-02 2.32809290e-01 1.37972504e-01 -6.20051265e-01
-3.09748650e-01 4.25072819e-01 -3.72542702e-02 5.04653752e-02
1.57657409e+00 1.37217820e-01 1.11433826e-01 7.64830947e-01
1.56161499e+00 -4.39966351e-01 -1.51623249e+00 -5.13461292e-01
1.66958064e-01 -4.15033430e-01 1.34489432e-01 -1.00808509e-01
-9.87839699e-01 1.11952019e+00 2.35100418e-01 2.46417850e-01
6.38815403e-01 3.58064055e-01 9.42152023e-01 5.50835133e-01
5.96069276e-01 -1.02461326e+00 1.82569530e-02 6.22468829e-01
4.98133272e-01 -1.57650030e+00 -1.10812224e-01 -6.00532830e-01
-6.43455684e-01 1.01235569e+00 6.51690543e-01 -6.79220498e-01
8.75630140e-01 4.78910618e-02 -1.39647380e-01 -4.36710060e-01
-7.69777715e-01 -5.15422404e-01 3.34978431e-01 7.16443300e-01
-1.21553436e-01 -9.70378667e-02 3.62054735e-01 4.98516828e-01
-2.65690535e-01 -2.79945545e-02 2.15878502e-01 1.09369254e+00
-7.33065009e-01 -6.30572677e-01 -8.41761231e-02 8.07587445e-01
1.49056762e-01 1.19134411e-01 -3.31079930e-01 6.39471114e-01
2.29603827e-01 8.20566297e-01 7.46924877e-01 -3.64450753e-01
2.25271389e-01 -1.16766365e-02 1.17326841e-01 -1.42418635e+00
-3.14688683e-01 -1.54206842e-01 -2.66812503e-01 -4.44181323e-01
-2.98518419e-01 -5.26843786e-01 -1.51785672e+00 -9.08430964e-02
-5.65658391e-01 2.79573709e-01 7.42698431e-01 1.26172936e+00
5.51047325e-01 2.99532264e-01 8.41326773e-01 -1.44434667e+00
-6.18502796e-01 -9.65614974e-01 -5.33207774e-01 5.16984820e-01
4.53967899e-01 -8.31482708e-01 -3.86741847e-01 -8.60303491e-02] | [8.022809028625488, -3.234128713607788] |
b1e21df8-2c81-4120-9e2e-835cfcedf626 | an-empirical-evaluation-of-doc2vec-with | 1607.05368 | null | http://arxiv.org/abs/1607.05368v1 | http://arxiv.org/pdf/1607.05368v1.pdf | An Empirical Evaluation of doc2vec with Practical Insights into Document Embedding Generation | Recently, Le and Mikolov (2014) proposed doc2vec as an extension to word2vec
(Mikolov et al., 2013a) to learn document-level embeddings. Despite promising
results in the original paper, others have struggled to reproduce those
results. This paper presents a rigorous empirical evaluation of doc2vec over
two tasks. We compare doc2vec to two baselines and two state-of-the-art
document embedding methodologies. We found that doc2vec performs robustly when
using models trained on large external corpora, and can be further improved by
using pre-trained word embeddings. We also provide recommendations on
hyper-parameter settings for general purpose applications, and release source
code to induce document embeddings using our trained doc2vec models. | ['Timothy Baldwin', 'Jey Han Lau'] | 2016-07-19 | an-empirical-evaluation-of-doc2vec-with-1 | https://aclanthology.org/W16-1609 | https://aclanthology.org/W16-1609.pdf | ws-2016-8 | ['document-embedding'] | ['methodology'] | [-4.88084704e-01 -1.14165649e-01 -3.63653094e-01 -2.26399198e-01
-8.08858812e-01 -8.30698669e-01 1.13213825e+00 3.37647259e-01
-7.36906111e-01 3.57067943e-01 8.20734978e-01 -6.40114188e-01
6.93964809e-02 -6.32226586e-01 -3.43280844e-02 -2.73092598e-01
-1.97591946e-01 3.52241993e-01 6.85717957e-03 -2.73775637e-01
4.97920513e-01 3.60380650e-01 -1.29545057e+00 1.19799785e-01
4.91261125e-01 3.79255921e-01 7.94077069e-02 9.68599856e-01
-4.65668499e-01 1.50496244e-01 -7.33966768e-01 -6.99378252e-01
4.89968136e-02 2.81012058e-03 -7.23182917e-01 -2.16084108e-01
5.56084394e-01 -2.81953216e-01 -6.83707774e-01 8.09412479e-01
7.28427291e-01 2.70398140e-01 1.16409481e+00 -1.04935765e+00
-1.61040747e+00 4.30431277e-01 -1.67901546e-01 3.61854434e-01
-4.30904125e-04 1.04929253e-01 1.54842305e+00 -1.21302605e+00
7.32376993e-01 1.17443383e+00 6.32084548e-01 6.01142585e-01
-1.24253237e+00 -3.74471724e-01 6.87192008e-02 2.90528923e-01
-1.36629879e+00 -4.43157107e-02 5.51906288e-01 -5.99336207e-01
1.54250443e+00 -2.41545606e-02 3.52679133e-01 1.39746284e+00
3.18863124e-01 7.21654475e-01 6.85903609e-01 -8.06957901e-01
-6.75757751e-02 5.11213660e-01 5.67933500e-01 3.10961246e-01
7.13105738e-01 -4.79030535e-02 1.35884166e-01 -3.93021107e-01
6.89255476e-01 -5.43604009e-02 -3.40827950e-03 -3.47231776e-01
-7.99101889e-01 1.40475953e+00 1.22949772e-01 8.31847966e-01
-2.34316796e-01 4.93583769e-01 7.80790865e-01 2.62668908e-01
7.47041225e-01 9.63523686e-01 -6.81870401e-01 -5.36813796e-01
-6.86499953e-01 3.41281652e-01 6.39760375e-01 9.52756047e-01
4.10019368e-01 1.37675732e-01 -4.04613197e-01 1.22935891e+00
4.43663806e-01 4.04750139e-01 1.00593972e+00 -5.62809885e-01
3.04586768e-01 5.73045611e-02 9.25079882e-02 -1.09711611e+00
-7.63600925e-03 -2.62596190e-01 -1.52571991e-01 -1.49830922e-01
-1.37298763e-01 -2.49524876e-01 -8.94179881e-01 1.21082604e+00
-3.33610535e-01 -1.05861686e-01 1.82423010e-01 6.45969748e-01
8.21432233e-01 8.75797391e-01 8.06769505e-02 3.61385196e-01
1.30841517e+00 -1.33451474e+00 -9.59380329e-01 -1.44991606e-01
1.03477800e+00 -9.42874610e-01 1.20967531e+00 2.34726980e-01
-6.98195040e-01 -6.80454731e-01 -1.14239800e+00 -2.46608585e-01
-1.05779660e+00 7.56239742e-02 9.59227920e-01 1.00430357e+00
-1.17533910e+00 3.52487564e-01 -7.54740179e-01 -6.76586807e-01
7.30508417e-02 -5.93812391e-02 -3.60422879e-01 -3.60977858e-01
-1.26117003e+00 1.06683838e+00 1.06065899e-01 -4.69355762e-01
-7.98940599e-01 -5.12209475e-01 -1.08586776e+00 1.94227998e-03
-3.45162809e-01 -2.29205891e-01 1.11497903e+00 -1.63544610e-01
-9.47754502e-01 5.98126709e-01 -9.12007540e-02 -2.37479091e-01
-2.37875909e-01 -4.46486831e-01 -8.60181153e-01 -4.99846833e-03
-2.27587344e-03 5.71613014e-01 3.74139667e-01 -1.08977103e+00
-2.08680004e-01 -6.12491891e-02 1.51761785e-01 -3.39990146e-02
-1.17821860e+00 1.84394136e-01 -8.53184998e-01 -7.84852624e-01
-5.14628410e-01 -6.94315195e-01 -2.56898493e-01 -1.49082407e-01
-5.29257469e-02 -6.50415659e-01 6.92253768e-01 -5.44778407e-01
1.53796494e+00 -2.16620255e+00 -9.95115191e-02 -1.04185544e-01
1.26552403e-01 6.54757082e-01 -8.74865115e-01 1.09185696e+00
4.76270206e-02 7.02808082e-01 1.79358542e-01 -5.65831780e-01
3.92155200e-01 2.37420842e-01 -3.61644179e-01 4.13025677e-01
1.25828996e-01 9.51874435e-01 -1.03456175e+00 -3.98958832e-01
2.57559210e-01 1.02522361e+00 -7.21411586e-01 1.60580873e-01
4.37215604e-02 -5.41741788e-01 -2.81109869e-01 3.68187428e-01
6.75495684e-01 3.19687203e-02 3.56400579e-01 -5.26477583e-03
-9.36556160e-02 5.13741434e-01 -9.67139244e-01 1.37820601e+00
-7.69208372e-01 1.29709601e+00 -3.98209512e-01 -8.03369284e-01
9.18695092e-01 4.20299381e-01 3.02526474e-01 -4.71137524e-01
2.19849959e-01 8.04775506e-02 -2.38806963e-01 -6.69223011e-01
1.08869255e+00 8.82696956e-02 -1.74939319e-01 4.90630716e-01
5.48213542e-01 -3.92454773e-01 3.80746305e-01 3.88882399e-01
1.09029293e+00 -2.54050314e-01 1.24771677e-01 -3.23676616e-01
2.53193140e-01 -1.24676555e-01 -1.59277380e-01 6.82298243e-01
-2.49049425e-01 7.73328424e-01 3.49892616e-01 -2.67838985e-01
-1.19312263e+00 -9.56993282e-01 -5.82390368e-01 1.09494936e+00
-3.13976973e-01 -1.05256462e+00 -3.31464171e-01 -9.57462132e-01
3.67402285e-01 1.04160714e+00 -1.01459253e+00 2.93655153e-02
-1.91951245e-01 -3.99649560e-01 8.59146297e-01 8.71831834e-01
-3.70625526e-01 -8.28562438e-01 -8.21867026e-03 2.70998538e-01
2.97187299e-01 -7.39873946e-01 -6.28369153e-01 1.86424762e-01
-7.20282018e-01 -9.55762208e-01 -8.17871094e-01 -8.96489859e-01
4.60924447e-01 5.66666901e-01 1.14455390e+00 1.35095671e-01
-3.80348831e-01 8.21512580e-01 -9.34547603e-01 -4.56767410e-01
-4.04289365e-02 1.22700766e-01 1.50186956e-01 -6.41839087e-01
1.14853597e+00 -1.27864555e-01 -2.97092646e-01 1.37911057e-02
-1.14339113e+00 -9.15380597e-01 1.95564643e-01 1.00384891e+00
3.44047427e-01 -4.21529077e-02 3.88857216e-01 -7.54370987e-01
1.36535180e+00 -5.92862427e-01 -1.88365400e-01 1.94054935e-02
-1.14636195e+00 1.53557420e-01 5.04315913e-01 -4.13982153e-01
-4.72028434e-01 -8.86246324e-01 -4.13906693e-01 -4.52839106e-01
9.75449905e-02 7.09336340e-01 3.62198442e-01 1.90569714e-01
5.76643348e-01 5.86633757e-02 -3.16214532e-01 -6.74031794e-01
9.54122126e-01 1.15781152e+00 -1.92135140e-01 -5.22791743e-01
8.18803310e-01 1.13241576e-01 -7.52883375e-01 -1.02386451e+00
-4.08849627e-01 -7.95839608e-01 -7.32308507e-01 2.47841984e-01
9.37014878e-01 -9.42072928e-01 2.85914958e-01 -1.13302216e-01
-1.35364747e+00 -9.71264318e-02 -2.06243321e-01 7.34947860e-01
-8.59083831e-02 4.35522705e-01 -5.81720948e-01 -6.15578353e-01
-1.66113019e-01 -9.26575661e-01 9.11008120e-01 -2.29461640e-01
-3.65795583e-01 -1.85478449e+00 8.96749258e-01 2.10848376e-02
6.59284830e-01 -1.83114022e-01 7.90656090e-01 -8.72166455e-01
2.41205186e-01 -6.38970017e-01 -2.31728092e-01 7.83592641e-01
3.03900003e-01 3.58070612e-01 -9.33304489e-01 -5.42908967e-01
-6.65946066e-01 -4.24746394e-01 1.02108574e+00 2.49913335e-01
1.08548689e+00 -2.55646497e-01 -4.03065413e-01 4.59694296e-01
1.62480855e+00 5.16862758e-02 6.59125388e-01 5.58615983e-01
5.82218885e-01 2.74857700e-01 4.97082651e-01 3.28649998e-01
4.07845020e-01 6.52619600e-01 4.34723087e-02 1.33643031e-01
-3.93102288e-01 -4.69082654e-01 3.65671635e-01 1.37928677e+00
5.80794066e-02 -6.82466745e-01 -7.67302215e-01 1.18547297e+00
-1.29544711e+00 -8.57352257e-01 -2.55106062e-01 1.63512778e+00
6.11226976e-01 -1.55567210e-02 -1.89404190e-01 -1.44094124e-01
3.46691906e-01 6.75365746e-01 2.97971498e-02 -1.21837711e+00
-9.21252295e-02 6.93021357e-01 5.60701430e-01 6.21670187e-01
-1.07442009e+00 1.17737341e+00 7.53572655e+00 6.84871554e-01
-8.84447753e-01 3.90212923e-01 -5.67881018e-02 -7.34218433e-02
-8.91428053e-01 -2.00601853e-02 -9.08971906e-01 3.02886963e-01
1.11473048e+00 -4.08838660e-01 6.86433762e-02 1.02207446e+00
-2.18084916e-01 6.01292908e-01 -9.02790248e-01 7.90964663e-01
6.09391928e-01 -1.20576072e+00 1.27262950e-01 2.82775491e-01
9.62329984e-01 3.38257611e-01 3.28104854e-01 8.01553845e-01
5.52936494e-01 -1.15949476e+00 9.18561146e-02 -7.79862255e-02
9.05694246e-01 -8.06058824e-01 1.05177021e+00 -4.17701930e-01
-1.20222497e+00 1.29695728e-01 -1.06645906e+00 1.43266600e-02
7.55059272e-02 4.36807096e-01 -7.06022561e-01 3.33124459e-01
3.97355080e-01 1.00401604e+00 -8.41994762e-01 8.18225265e-01
-4.44863349e-01 9.60887611e-01 2.62294143e-01 -4.62680668e-01
7.13424146e-01 -3.80529203e-02 1.46351174e-01 1.78748286e+00
2.60027379e-01 -2.95229614e-01 -1.77247450e-01 4.30480182e-01
-7.91316628e-02 4.56560284e-01 -1.04754472e+00 -7.80000448e-01
5.85416436e-01 1.28384769e+00 -2.55763918e-01 -3.31897587e-01
-7.85105884e-01 1.01450825e+00 4.77634221e-01 4.92639005e-01
-6.60370529e-01 -1.16137719e+00 1.28719687e+00 -1.61061794e-01
9.43234742e-01 -7.35859096e-01 -1.61760300e-01 -1.20423412e+00
-3.34162533e-01 -4.96694863e-01 2.74589598e-01 -5.88303983e-01
-1.65797615e+00 7.34413564e-01 -6.68685362e-02 -1.19809914e+00
-2.30247766e-01 -1.09156048e+00 -7.43258476e-01 8.46895099e-01
-1.51557350e+00 -9.10413861e-01 1.90725133e-01 1.40042290e-01
6.61519110e-01 -3.46196234e-01 1.31431770e+00 3.66019994e-01
-3.17311525e-01 1.08212173e+00 9.52016115e-01 4.04787749e-01
9.99029398e-01 -1.31773484e+00 8.83498669e-01 4.89719391e-01
5.65831065e-01 1.04471135e+00 7.47051001e-01 -4.87158954e-01
-1.51305473e+00 -1.12288606e+00 1.36343777e+00 -8.17490101e-01
1.09065247e+00 -5.68347633e-01 -7.10697889e-01 8.36075723e-01
9.95357215e-01 -3.01506668e-02 1.24383473e+00 5.08056223e-01
-8.73881817e-01 3.72779191e-01 -8.49099159e-01 6.35486305e-01
6.46133840e-01 -7.88839042e-01 -9.35051143e-01 4.02306587e-01
1.02386379e+00 3.69524002e-01 -1.08141613e+00 -1.81966081e-01
6.84561133e-01 -4.50909585e-01 1.02073526e+00 -1.04944789e+00
6.61960781e-01 9.81447473e-02 -6.52759850e-01 -1.80431700e+00
-7.01191485e-01 2.02454580e-03 -2.64632702e-01 1.27366173e+00
4.71747756e-01 -5.81304073e-01 1.42370537e-01 2.58109838e-01
-2.37359509e-01 -6.80624723e-01 -4.89139110e-01 -1.32679260e+00
9.28507090e-01 -6.89329922e-01 4.14693683e-01 1.14575922e+00
3.62428576e-01 1.75396502e-01 -2.16702461e-01 -1.06299914e-01
2.80559301e-01 -4.07468766e-01 8.56640756e-01 -9.65817153e-01
-2.53944576e-01 -5.39338529e-01 -7.70146787e-01 -1.14094257e+00
3.36869240e-01 -1.12938440e+00 -1.36700511e-01 -2.02531552e+00
7.70074278e-02 -2.28192985e-01 -6.58940017e-01 3.28016967e-01
-2.23224133e-01 6.05999112e-01 1.34261861e-01 -3.49488147e-02
-3.28729540e-01 7.06966221e-01 1.00937068e+00 -4.06436265e-01
6.96514100e-02 -7.86524951e-01 -1.04422617e+00 1.67180784e-02
7.55350113e-01 -4.20487344e-01 -5.77401876e-01 -8.60339463e-01
-4.86862697e-02 -7.74534881e-01 -1.55738637e-01 -6.37510598e-01
-1.35767564e-01 -5.51674962e-02 3.18775177e-01 -5.09512484e-01
2.35798225e-01 -4.22350407e-01 -7.07690239e-01 1.12716012e-01
-5.57328761e-01 3.80231827e-01 5.53740859e-01 4.48172927e-01
-3.34471524e-01 -7.09018946e-01 3.81493986e-01 1.27204612e-01
-7.23489225e-01 2.84133166e-01 -7.88703084e-01 2.33852431e-01
8.13431680e-01 -3.48833166e-02 -3.97107631e-01 -3.87868166e-01
-2.38454670e-01 1.21369921e-01 4.86757725e-01 8.97855103e-01
7.36223876e-01 -1.63172829e+00 -6.97281182e-01 2.41918534e-01
6.80594206e-01 -9.04055655e-01 -1.17127024e-01 2.67046511e-01
-4.06558037e-01 9.59878802e-01 1.17893955e-02 2.85299011e-02
-1.13418651e+00 6.87368274e-01 -2.34889060e-01 -3.31043184e-01
-3.66144329e-01 8.35659802e-01 -1.41933143e-01 -5.67966223e-01
2.13429287e-01 -4.73171882e-02 -5.26589692e-01 2.89602757e-01
8.45575094e-01 2.08477125e-01 -2.43755654e-02 -4.42665696e-01
-3.71924311e-01 4.99233991e-01 -4.43904191e-01 -4.69176024e-01
1.51237547e+00 8.35312083e-02 2.18290776e-01 5.24905264e-01
1.98881590e+00 2.14964747e-01 -5.02995789e-01 -4.19021323e-02
-2.38103569e-01 -7.61942863e-01 3.90769392e-01 -4.81654912e-01
-9.07642245e-01 1.26213288e+00 5.90125620e-01 4.16433960e-01
4.75404948e-01 4.53685969e-02 7.92641938e-01 4.53579634e-01
9.99389440e-02 -1.21748471e+00 2.33735248e-01 9.31980789e-01
7.20073283e-01 -9.36010122e-01 1.63935989e-01 2.19371632e-01
-7.60091364e-01 1.29346645e+00 5.42885303e-01 -4.16263193e-01
1.08486176e+00 3.32397148e-02 2.50335485e-01 -3.41818303e-01
-8.69971037e-01 -2.72367537e-01 4.28204060e-01 9.73268628e-01
1.03431308e+00 1.78493470e-01 -8.23314548e-01 5.26463687e-01
-1.00330129e-01 -4.37951803e-01 7.33695030e-01 1.02496588e+00
-3.62338543e-01 -1.64282262e+00 -5.17828539e-02 6.57689929e-01
-3.49196076e-01 -5.79480290e-01 -4.08909023e-01 8.24641526e-01
-1.34296775e-01 1.10015762e+00 3.52713645e-01 -7.19977915e-01
2.69161761e-01 2.47617230e-01 3.37784797e-01 -9.91323054e-01
-5.28093100e-01 -1.10498205e-01 2.60839283e-01 -2.05191493e-01
-1.15406744e-01 -7.13613808e-01 -7.59103835e-01 -2.53952175e-01
-4.80189681e-01 5.05035520e-01 9.71944571e-01 4.83877242e-01
4.99116659e-01 5.13141096e-01 6.95299268e-01 -5.88299692e-01
-5.87890029e-01 -1.47228491e+00 -7.91609764e-01 4.41201568e-01
8.84712189e-02 -7.00490713e-01 -6.27107859e-01 -1.20676063e-01] | [10.490885734558105, 8.568870544433594] |
390c0165-6d5e-4172-941f-9e416a623b57 | a-vector-quantized-masked-autoencoder-for | 2304.11117 | null | https://arxiv.org/abs/2304.11117v1 | https://arxiv.org/pdf/2304.11117v1.pdf | A vector quantized masked autoencoder for speech emotion recognition | Recent years have seen remarkable progress in speech emotion recognition (SER), thanks to advances in deep learning techniques. However, the limited availability of labeled data remains a significant challenge in the field. Self-supervised learning has recently emerged as a promising solution to address this challenge. In this paper, we propose the vector quantized masked autoencoder for speech (VQ-MAE-S), a self-supervised model that is fine-tuned to recognize emotions from speech signals. The VQ-MAE-S model is based on a masked autoencoder (MAE) that operates in the discrete latent space of a vector-quantized variational autoencoder. Experimental results show that the proposed VQ-MAE-S model, pre-trained on the VoxCeleb2 dataset and fine-tuned on emotional speech data, outperforms an MAE working on the raw spectrogram representation and other state-of-the-art methods in SER. | ['Renaud Séguier', 'Simon Leglaive', 'Samir Sadok'] | 2023-04-21 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [-1.72916681e-01 3.60901356e-01 1.63509861e-01 -4.27919567e-01
-7.28474975e-01 -1.03520021e-01 4.20919329e-01 -7.30497167e-02
-1.65241361e-01 4.42976207e-01 4.21134561e-01 1.21913143e-01
3.46490055e-01 -4.55711901e-01 -5.95011830e-01 -7.59465456e-01
-3.16997059e-02 1.68940261e-01 -2.08195284e-01 -4.05680597e-01
-3.66770625e-01 3.58866006e-02 -1.96909297e+00 4.86302137e-01
7.09553361e-01 1.28659666e+00 1.26292080e-01 7.48213768e-01
-2.15509236e-01 1.11576533e+00 -6.33474052e-01 -3.43854576e-01
-3.36746722e-01 -7.40725160e-01 -6.74111068e-01 2.78223366e-01
-5.45887910e-02 -3.22457589e-02 -3.19027334e-01 8.67040575e-01
4.59040761e-01 5.48557043e-01 8.04707646e-01 -1.38953960e+00
-6.34195268e-01 6.25234425e-01 1.64435431e-01 2.25332797e-01
-1.31693920e-02 -3.82069170e-01 1.20206976e+00 -1.00519729e+00
4.46109980e-01 1.17399418e+00 4.66261625e-01 8.77919316e-01
-1.13792682e+00 -3.88151854e-01 -2.12349012e-01 4.23945129e-01
-1.34696031e+00 -6.59113050e-01 1.20518100e+00 -5.04038334e-01
1.48105681e+00 1.21396400e-01 6.66109204e-01 1.61851418e+00
-9.94497258e-03 1.06285512e+00 9.80104268e-01 -5.18045425e-01
9.38715935e-01 5.92201412e-01 -2.16322005e-01 7.64708042e-01
-9.39018011e-01 3.55649859e-01 -9.83240366e-01 -2.90538996e-01
5.69277585e-01 -3.55156481e-01 -1.75020978e-01 -2.90938139e-01
-7.56286204e-01 1.26656091e+00 2.48385463e-02 6.94894552e-01
-7.93970466e-01 -1.86905026e-01 6.95199013e-01 5.47986984e-01
1.00579369e+00 1.08790495e-01 -6.09264433e-01 -5.19594908e-01
-1.26522303e+00 -3.00392121e-01 7.99724758e-01 3.00079554e-01
5.47767937e-01 9.37742174e-01 9.98146832e-02 1.17629421e+00
2.74857968e-01 1.30709082e-01 1.04697478e+00 -7.20808625e-01
-1.11352369e-01 3.19450289e-01 -2.33852983e-01 -7.10399330e-01
-1.49912238e-01 -4.42777961e-01 -1.01286495e+00 3.30199242e-01
-4.38549280e-01 -3.11515659e-01 -5.94504595e-01 1.66285694e+00
3.75513256e-01 4.12997216e-01 7.84950674e-01 8.93395960e-01
1.07503664e+00 1.30769944e+00 1.20720886e-01 -4.78997529e-01
1.03091431e+00 -1.04019237e+00 -1.29142821e+00 -1.11936137e-01
2.00706124e-01 -3.49588305e-01 8.78464282e-01 6.07467890e-01
-8.72713387e-01 -8.04054797e-01 -1.08786535e+00 1.57119766e-01
-5.06850958e-01 2.95090199e-01 2.90587157e-01 8.52168262e-01
-1.15614009e+00 5.65455198e-01 -9.97744858e-01 -2.54412353e-01
1.43000379e-01 9.51436460e-02 -4.91530269e-01 7.98231721e-01
-1.48185146e+00 6.61936343e-01 4.98276949e-01 -6.55464381e-02
-1.08631861e+00 -3.82320434e-01 -1.07667601e+00 4.77662623e-01
1.56512633e-01 -2.32960179e-01 1.26645255e+00 -1.37673521e+00
-2.40287161e+00 6.92320883e-01 -3.93760830e-01 -6.90544784e-01
5.62690645e-02 9.92888436e-02 -8.51749003e-01 4.37971056e-01
-3.23620826e-01 5.61483979e-01 1.40670395e+00 -1.14894879e+00
-2.52627254e-01 -1.54983655e-01 -6.68015301e-01 -1.05260992e-02
-7.67668962e-01 4.93608601e-02 2.95183867e-01 -7.54911900e-01
-2.17865780e-01 -4.68585134e-01 7.84065798e-02 -4.93585557e-01
-9.23865214e-02 -8.64508092e-01 1.01388490e+00 -7.81303942e-01
1.08846414e+00 -2.56085181e+00 4.35632259e-01 -1.81787625e-01
-7.30047449e-02 5.10037780e-01 -2.53741294e-02 6.19587064e-01
-4.49485928e-01 -1.31481096e-01 -3.44722748e-01 -9.49570775e-01
3.03866148e-01 5.20632863e-01 -5.35422802e-01 2.46947244e-01
5.77958047e-01 6.74944878e-01 -8.42973709e-01 -5.09075344e-01
5.38195610e-01 9.59633827e-01 -2.94179380e-01 6.88956559e-01
-3.29661280e-01 3.71739894e-01 1.55324023e-02 2.65551209e-01
2.67135948e-01 -1.12909496e-01 4.09304611e-02 -1.94518697e-02
-2.06240892e-01 2.13861465e-02 -9.46595132e-01 1.57571399e+00
-5.83474755e-01 8.49937201e-01 2.45337695e-01 -1.43527472e+00
1.20045710e+00 1.16115785e+00 5.69078445e-01 -5.40393710e-01
4.53838319e-01 5.61185181e-02 -5.45695364e-01 -7.28854179e-01
4.56419170e-01 -6.26766145e-01 8.34212974e-02 1.92486808e-01
1.03281033e+00 2.75964681e-02 -2.93389171e-01 1.35951087e-01
8.43146563e-01 -1.28492966e-01 4.30372089e-01 8.27186927e-02
6.88058257e-01 -2.58311808e-01 5.99949718e-01 2.73082942e-01
-5.36657929e-01 4.10403192e-01 2.51589954e-01 -2.18976840e-01
-1.02507114e+00 -9.69104707e-01 -8.29580650e-02 1.07899928e+00
-4.83520746e-01 -5.40660143e-01 -1.09102237e+00 -4.89624172e-01
-4.07744825e-01 8.77302408e-01 -7.47090578e-01 -2.17864007e-01
7.32552856e-02 -2.50477403e-01 5.17469823e-01 3.96925807e-01
2.38350078e-01 -1.47103310e+00 -5.51394343e-01 5.58500648e-01
-4.08163875e-01 -1.32608581e+00 2.35803910e-02 4.29213166e-01
-5.18335104e-01 -3.26726645e-01 -8.26680124e-01 -8.26039732e-01
6.28244057e-02 -2.98223704e-01 9.59232628e-01 -3.95571917e-01
-1.02034993e-01 7.33308911e-01 -9.03964162e-01 -5.54308295e-01
-8.44461262e-01 -1.65647253e-01 2.83254892e-01 8.41111720e-01
3.67906094e-01 -7.10701704e-01 -4.15393300e-02 -3.41428779e-02
-9.76619422e-01 -3.54374141e-01 1.78562030e-01 1.18630874e+00
6.07763946e-01 4.17101562e-01 8.43947113e-01 -2.51622200e-01
6.23131454e-01 -6.65271342e-01 -3.37076366e-01 7.82186463e-02
-2.18165994e-01 3.03536564e-01 8.05530012e-01 -5.52575707e-01
-1.27019703e+00 1.36350632e-01 -7.19183683e-01 -1.09199345e+00
-4.26433116e-01 5.13414264e-01 -1.20040148e-01 2.60754466e-01
5.82891941e-01 4.17291433e-01 -2.87481993e-02 -3.96120042e-01
3.86360079e-01 1.30901706e+00 6.03335023e-01 -1.50687844e-01
2.03756988e-01 3.75536174e-01 -5.28328419e-01 -1.51520956e+00
-7.41683185e-01 -4.92614746e-01 -3.51896852e-01 -4.47414756e-01
1.26815641e+00 -1.04949582e+00 -5.47511518e-01 2.96397567e-01
-1.09817123e+00 -4.47370797e-01 -4.20615911e-01 6.23201132e-01
-7.80378222e-01 3.32966119e-01 -6.43693805e-01 -1.46368659e+00
-4.80866432e-01 -8.97365212e-01 1.21724594e+00 7.37577677e-02
-2.65609860e-01 -1.05850458e+00 4.27254319e-01 3.77426624e-01
3.48448366e-01 1.30000159e-01 6.40561521e-01 -8.12300861e-01
1.95365176e-01 1.30463868e-01 5.26615798e-01 1.09968603e+00
1.07415831e-02 -8.23104680e-02 -1.46414196e+00 -1.20225146e-01
3.52376014e-01 -8.59333694e-01 8.06200504e-01 4.43227619e-01
1.01989484e+00 -3.13969105e-01 3.36870253e-01 3.76108229e-01
1.04436278e+00 1.77679121e-01 4.17892367e-01 -1.71310812e-01
1.47021919e-01 5.80820262e-01 3.56806487e-01 6.02258325e-01
3.54673773e-01 4.32785034e-01 3.93195570e-01 -1.21209390e-01
6.47026747e-02 -3.08292150e-01 6.17661774e-01 1.58209050e+00
1.71324715e-01 -2.67741919e-01 -5.05258620e-01 7.17600226e-01
-1.84775126e+00 -1.06823289e+00 2.87690669e-01 1.60802746e+00
5.43115020e-01 -2.82794178e-01 9.02164429e-02 5.54439425e-01
4.94305015e-01 4.22328234e-01 -3.82239908e-01 -8.12902689e-01
-3.88046503e-01 6.63242161e-01 -5.86676598e-01 4.98034686e-01
-1.11440349e+00 1.10700750e+00 5.63792753e+00 8.95863116e-01
-1.42159545e+00 6.01474822e-01 3.01770478e-01 2.04816386e-01
4.21324949e-04 -4.67816293e-01 -2.44717792e-01 4.04308528e-01
1.54599822e+00 1.20887049e-01 6.23790324e-01 1.24902678e+00
1.21092528e-01 2.36628458e-01 -8.72632623e-01 1.29633021e+00
4.57686901e-01 -1.09492612e+00 -4.05766547e-01 -2.19458550e-01
6.60391867e-01 -6.67719729e-03 3.40326279e-01 7.10226476e-01
2.46624015e-02 -8.58654678e-01 7.85063446e-01 2.39750266e-01
5.84917903e-01 -8.67444634e-01 7.00652301e-01 5.35023510e-01
-1.04192662e+00 1.35394912e-02 -4.81199414e-01 8.19743797e-02
1.91661656e-01 5.05659163e-01 -8.51807415e-01 3.93762410e-01
9.98678446e-01 5.28869629e-01 1.94544449e-01 1.14452615e-01
-2.37285689e-01 1.25779474e+00 -1.00545257e-01 -2.36325607e-01
2.68686682e-01 -7.08150566e-02 6.66734755e-01 1.16830444e+00
2.26344511e-01 1.63433895e-01 -1.92628264e-01 8.22866440e-01
8.85006040e-02 4.08070058e-01 -2.64456451e-01 -6.52726412e-01
1.45613313e-01 1.16990948e+00 -3.40379894e-01 -6.00306273e-01
-2.62556493e-01 1.51169276e+00 5.75566649e-01 4.15656924e-01
-5.65547764e-01 -1.17026635e-01 6.89651966e-01 -5.83804250e-01
9.12241518e-01 -8.53469893e-02 4.95574921e-01 -1.40309131e+00
-2.34086111e-01 -9.55982983e-01 2.39481583e-01 -1.10965848e+00
-1.37801933e+00 1.25082040e+00 -4.29184556e-01 -1.04999912e+00
-7.26421714e-01 -6.22334599e-01 -4.29382861e-01 5.19550443e-01
-1.33335984e+00 -9.24640238e-01 1.32226748e-02 8.56261730e-01
8.97671521e-01 -7.55599260e-01 1.40511775e+00 6.66650236e-02
-4.53925371e-01 4.92511243e-01 2.93168128e-01 1.07433163e-01
2.77446687e-01 -1.21439326e+00 8.30330998e-02 4.89124507e-01
7.17440605e-01 1.29390910e-01 8.26535404e-01 -2.23014593e-01
-1.12803888e+00 -8.98931623e-01 9.78828728e-01 -1.02510944e-01
6.06111050e-01 -6.93843067e-01 -1.09019303e+00 5.60559213e-01
6.78357244e-01 2.58315802e-01 1.10002387e+00 -6.13419488e-02
-3.45869213e-01 5.78433610e-02 -1.03525996e+00 7.25094154e-02
3.66899401e-01 -1.15188563e+00 -8.92275989e-01 5.22563681e-02
8.24193180e-01 -1.42690286e-01 -1.16157401e+00 1.68243170e-01
1.80311769e-01 -1.19471562e+00 7.49680459e-01 -5.98594248e-01
3.39061618e-01 1.35664940e-01 -4.49107528e-01 -1.92946434e+00
-8.45889524e-02 -7.29521751e-01 -5.24337888e-01 1.25660992e+00
1.52967691e-01 -3.57980192e-01 6.44896805e-01 -1.34964317e-01
-1.98255241e-01 -6.48945928e-01 -1.42096901e+00 -7.16586113e-01
-1.05332151e-01 -6.45792365e-01 4.17020023e-01 1.15837932e+00
2.64256060e-01 5.10026813e-01 -7.39499748e-01 2.94554025e-01
5.12535632e-01 -1.92594945e-01 4.84358311e-01 -1.21061754e+00
-4.97590691e-01 -1.16336688e-01 -5.98147333e-01 -5.88493228e-01
8.39190900e-01 -6.73547029e-01 1.48793638e-01 -1.19340861e+00
-3.77396166e-01 4.21894908e-01 -4.67954040e-01 4.11481649e-01
2.42355675e-03 -9.21883136e-02 6.25374867e-03 -3.21627617e-01
-4.72618461e-01 1.39362395e+00 3.97517145e-01 -1.58358693e-01
-3.26954424e-01 -3.07207912e-01 -4.21905927e-02 5.53272128e-01
7.58258462e-01 -4.60969359e-01 -2.78885365e-01 2.11677939e-01
1.21511696e-02 5.36444545e-01 3.23896199e-01 -1.08756137e+00
2.41777927e-01 2.96742767e-01 1.96654126e-01 -6.34585500e-01
9.09566522e-01 -7.14087367e-01 -1.93227418e-02 5.05434647e-02
-3.20856154e-01 -4.03395146e-01 6.19254559e-02 5.29749155e-01
-9.31932807e-01 -9.68607292e-02 6.03686631e-01 -4.44810800e-02
-8.59405339e-01 8.79951939e-02 -9.49759722e-01 -2.47229055e-01
7.12750256e-01 1.41034201e-01 2.69821644e-01 -7.68102646e-01
-1.13215768e+00 -2.38380685e-01 -1.90312028e-01 5.48112154e-01
8.26097012e-01 -1.32567096e+00 -6.71004474e-01 4.01691288e-01
2.31177568e-01 -5.21383464e-01 8.29086244e-01 3.39815438e-01
2.82610148e-01 3.59271526e-01 -5.23863286e-02 -5.93399346e-01
-1.22495162e+00 7.14181185e-01 4.02963221e-01 -1.41825810e-01
-4.63341981e-01 9.86298561e-01 -1.39965266e-01 -6.60766184e-01
4.37113971e-01 -8.68814290e-02 -4.48923022e-01 1.35164887e-01
4.47413504e-01 2.52385259e-01 1.58389062e-01 -1.08094549e+00
-1.14854969e-01 1.30611464e-01 4.45830673e-01 -6.61846936e-01
1.67824280e+00 -8.47674981e-02 1.15474515e-01 1.04016054e+00
1.44965732e+00 -1.91346318e-01 -1.16818035e+00 -3.55933040e-01
-3.61579984e-01 1.88726500e-01 7.31579661e-01 -5.64465761e-01
-9.56396282e-01 1.26312959e+00 7.87612438e-01 5.74530184e-01
1.13370895e+00 1.65048793e-01 8.01198244e-01 2.42109597e-01
1.24318235e-01 -1.41499567e+00 4.15817767e-01 4.18740839e-01
9.56152558e-01 -1.30830801e+00 -6.33490443e-01 -1.39691606e-01
-1.11411047e+00 1.07080996e+00 2.07012728e-01 -5.27137518e-02
9.17998970e-01 1.31182805e-01 4.12284702e-01 -8.04755837e-02
-9.76240158e-01 -3.33586812e-01 2.78129458e-01 5.46715617e-01
3.68016779e-01 2.36143291e-01 1.92027554e-01 1.08020306e+00
-4.88880754e-01 -6.53892010e-02 1.21131591e-01 7.37334311e-01
-2.67721087e-01 -1.03895116e+00 -4.06687945e-01 -2.10641436e-02
-3.11504483e-01 1.03159323e-01 -4.69403654e-01 3.09421927e-01
-1.77921704e-03 1.43610561e+00 2.55943775e-01 -6.21191502e-01
1.25695392e-01 5.95173001e-01 2.61048555e-01 -3.23085934e-01
-6.20285273e-01 3.77461225e-01 -5.05220816e-02 -4.67626750e-01
-7.31682241e-01 -4.62448508e-01 -1.16650581e+00 3.83654058e-01
-1.93970352e-01 6.40390515e-01 1.01503468e+00 9.71029937e-01
4.06205088e-01 6.79760039e-01 9.89207447e-01 -7.58600116e-01
-6.35225296e-01 -1.06766129e+00 -8.96780372e-01 3.26516986e-01
7.49683201e-01 -6.45170689e-01 -6.30244136e-01 1.78698823e-01] | [14.005913734436035, 5.66741418838501] |
d35683a2-4153-42d8-bf66-7b1e18de228e | automatic-feature-learning-for-robust-shadow | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Khan_Automatic_Feature_Learning_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Khan_Automatic_Feature_Learning_2014_CVPR_paper.pdf | Automatic Feature Learning for Robust Shadow Detection | We present a practical framework to automatically detect shadows in real world scenes from a single photograph. Previous works on shadow detection put a lot of effort in designing shadow variant and invariant hand-crafted features. In contrast, our framework automatically learns the most relevant features in a supervised manner using multiple convolutional deep neural networks (ConvNets). The 7-layer network architecture of each ConvNet consists of alternating convolution and sub-sampling layers. The proposed framework learns features at the super-pixel level and along the object boundaries. In both cases, features are extracted using a context aware window centered at interest points. The predicted posteriors based on the learned features are fed to a conditional random field model to generate smooth shadow contours. Our proposed framework consistently performed better than the state-of-the-art on all major shadow databases collected under a variety of conditions. | ['Salman Hameed Khan', 'Mohammed Bennamoun', 'Ferdous Sohel', 'Roberto Togneri'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['shadow-detection'] | ['computer-vision'] | [ 7.71050274e-01 5.82405962e-02 1.68579161e-01 -8.46912086e-01
-4.70770478e-01 -1.59966275e-01 5.37584245e-01 -5.25399983e-01
-2.83906996e-01 8.52964044e-01 1.35953188e-01 -2.13696137e-01
2.23890066e-01 -8.86306226e-01 -7.17214823e-01 -8.18646729e-01
9.45941918e-03 2.28154585e-01 8.84166837e-01 1.52153587e-02
3.48197222e-01 7.97023356e-01 -1.59153032e+00 2.72355974e-01
7.91262031e-01 1.23009109e+00 6.28164828e-01 8.55757952e-01
-2.04554889e-02 5.68426490e-01 -6.86684012e-01 -5.97789772e-02
4.65421140e-01 -9.97348428e-02 -3.24573725e-01 2.34946683e-01
8.58211458e-01 -6.46498382e-01 -3.63477409e-01 5.72275937e-01
3.26347172e-01 -3.59488628e-03 6.01234138e-01 -1.02419794e+00
-4.40434217e-01 -4.68733981e-02 -4.83456671e-01 3.47349606e-02
1.88694820e-02 5.10791019e-02 9.61590707e-01 -1.00944555e+00
6.30013585e-01 1.04078162e+00 9.30725157e-01 1.96179375e-02
-9.50947404e-01 -5.66056371e-01 1.30126268e-01 1.76450625e-01
-1.10073471e+00 -1.59347758e-01 1.05975223e+00 -1.02255873e-01
6.55094683e-01 1.08911604e-01 7.72024393e-01 9.60519731e-01
6.98889911e-01 9.80065703e-01 1.41578162e+00 -3.16337436e-01
1.55301288e-01 7.97816887e-02 -1.26467859e-02 1.16966569e+00
1.48380771e-01 2.32585549e-01 -7.87042737e-01 -3.25092614e-01
7.22614825e-01 2.12074265e-01 -3.99868280e-01 -4.28398490e-01
-8.76557291e-01 7.52077460e-01 5.92041373e-01 9.43451934e-03
-5.14714479e-01 4.00600225e-01 -4.34149466e-02 -3.25631768e-01
4.25893337e-01 8.76220118e-04 -7.44058549e-01 4.16685998e-01
-1.34089100e+00 1.28419176e-01 7.79445350e-01 7.50113368e-01
1.05994987e+00 1.14113688e-01 -4.14438546e-01 5.03577113e-01
2.27595434e-01 7.61933267e-01 1.20390486e-02 -9.56116319e-01
-3.98769416e-02 3.13558191e-01 2.90063828e-01 -8.56003225e-01
-3.86039197e-01 -4.40854311e-01 -3.62130612e-01 7.43757308e-01
3.22212696e-01 -3.15103650e-01 -1.40632772e+00 1.22548831e+00
2.97375232e-01 4.18375641e-01 -1.51794199e-02 8.72162402e-01
7.06003606e-01 6.34815693e-01 -2.55246133e-01 6.36623725e-02
1.13654542e+00 -1.01825583e+00 -5.87419033e-01 -4.51718569e-01
-4.22896802e-01 -1.13330591e+00 1.09254444e+00 1.98502377e-01
-4.55213994e-01 -7.22698331e-01 -1.18797314e+00 -4.74504866e-02
-4.78057325e-01 7.19336629e-01 9.31178451e-01 5.19848526e-01
-9.14745748e-01 5.90833127e-01 -5.66473186e-01 -2.34571084e-01
8.18400741e-01 2.93798149e-01 5.07123210e-02 3.97147164e-02
-6.96505308e-01 6.33958638e-01 -1.02916062e-01 2.55527556e-01
-1.27427793e+00 -5.96505105e-01 -8.04995477e-01 2.17423052e-01
4.35606539e-01 -4.90152121e-01 1.18249142e+00 -1.22134614e+00
-1.72798717e+00 4.76350814e-01 -4.34163243e-01 -5.82729340e-01
4.69379961e-01 -7.81511307e-01 -5.05674630e-02 2.84756087e-02
-3.41094099e-02 5.35229623e-01 1.21735835e+00 -1.63439751e+00
-8.70486200e-01 2.91306130e-03 6.86206818e-02 1.57112569e-01
-8.85509327e-02 -1.80777907e-01 -2.35720485e-01 -5.78403056e-01
-8.52351710e-02 -9.97122049e-01 -1.84191868e-01 3.95633161e-01
-4.65837240e-01 -5.79297505e-02 1.70318019e+00 -6.22839212e-01
6.45599604e-01 -1.80442023e+00 -4.80681658e-01 9.96788517e-02
-2.44162679e-02 4.01546270e-01 2.71869212e-01 2.87191689e-01
5.09452403e-01 -5.66717148e-01 -6.69627488e-01 -3.95222723e-01
-1.60801351e-01 2.55620837e-01 -6.21403039e-01 6.13576293e-01
9.09391567e-02 8.22982609e-01 -7.46320367e-01 -6.61062598e-01
5.67447901e-01 7.73780107e-01 -7.01665878e-02 4.69599187e-01
-4.56126809e-01 9.03164074e-02 -4.57461953e-01 9.11366284e-01
9.44020748e-01 -1.71358302e-01 1.27173871e-01 -1.64912209e-01
-2.83800423e-01 2.16068000e-01 -8.49023402e-01 1.48334455e+00
-8.17290187e-01 1.42164767e+00 8.89396574e-03 -4.87230510e-01
9.76795077e-01 3.84709984e-02 2.64275640e-01 -1.00440681e-01
1.25826523e-01 -5.01468927e-02 -2.31557831e-01 -3.72791320e-01
5.43588161e-01 2.70427018e-02 1.73042178e-01 3.98133814e-01
8.01590532e-02 -1.94672808e-01 -4.50175285e-01 -1.21965617e-01
1.12272882e+00 4.94897991e-01 2.59776443e-01 -3.58045459e-01
3.48114878e-01 -1.01250120e-01 7.16940999e-01 8.29413772e-01
-1.45376027e-01 8.45240653e-01 1.47417620e-01 -6.76366329e-01
-6.83983803e-01 -1.26290417e+00 -1.62113175e-01 1.20747983e+00
3.41088861e-01 -1.63703889e-01 -5.79612970e-01 -1.03821146e+00
1.76355064e-01 6.94729924e-01 -8.31999421e-01 4.45061594e-01
-6.27935886e-01 -5.89813292e-01 2.78957337e-01 7.32885778e-01
9.91303742e-01 -1.58478391e+00 -1.57576990e+00 3.34115960e-02
2.84259826e-01 -1.10026944e+00 -2.91269749e-01 6.47536695e-01
-6.17523253e-01 -1.10608423e+00 -6.96495891e-01 -7.28015661e-01
6.71954870e-01 5.74169993e-01 1.05445278e+00 1.22288037e-02
-7.66311407e-01 1.77033767e-01 -1.27254903e-01 -8.43016863e-01
1.44499436e-01 -2.48777822e-01 -5.06973863e-01 3.27655524e-01
9.70589072e-02 -6.69279277e-01 -8.69471610e-01 9.05289426e-02
-6.84470892e-01 2.97120333e-01 1.06010783e+00 9.06974018e-01
3.70417058e-01 -1.17966369e-01 2.40110536e-03 -1.13726580e+00
4.63222295e-01 4.09415625e-02 -9.21632290e-01 3.09228003e-01
-5.88657081e-01 1.51553512e-01 7.19541728e-01 -9.53240022e-02
-1.74383724e+00 6.98120892e-01 2.67456323e-01 -1.47002518e-01
-5.01986027e-01 -2.09628746e-01 -2.36965846e-02 -3.07835013e-01
5.49647391e-01 2.73333669e-01 -6.54033422e-01 -3.14739257e-01
3.66334647e-01 5.67971826e-01 5.10944784e-01 -4.27696139e-01
1.06411457e+00 1.09412444e+00 8.04457907e-03 -1.02271438e+00
-1.17384684e+00 -3.89099598e-01 -1.04994893e+00 -4.53582048e-01
8.15709054e-01 -5.36254048e-01 -3.16980988e-01 7.01141119e-01
-1.13612914e+00 -7.35168517e-01 -1.90824613e-01 -5.46494965e-03
-5.05373716e-01 1.56659961e-01 -1.35701984e-01 -9.96173739e-01
-4.88674998e-01 -8.74513149e-01 1.60891306e+00 8.03053081e-01
2.26926491e-01 -8.13432515e-01 5.15760761e-03 4.35696840e-02
5.09685993e-01 4.78006661e-01 6.55936122e-01 -1.50462501e-02
-9.89810824e-01 -2.11657122e-01 -5.16625702e-01 4.33249772e-01
3.09134066e-01 1.39160752e-01 -1.62759733e+00 1.84622839e-01
-1.88570261e-01 -3.48739803e-01 1.27364707e+00 7.94115067e-01
1.40292323e+00 4.06588800e-02 -5.66330314e-01 7.61332393e-01
1.65082455e+00 1.18662246e-01 4.35914397e-01 -2.86783297e-02
6.28455758e-01 2.47646928e-01 8.26085746e-01 5.77723563e-01
1.67058423e-01 3.51629406e-01 4.42788303e-01 -5.26379704e-01
-4.78603363e-01 -2.02027839e-02 1.53704330e-01 -2.60429353e-01
-2.58405536e-01 -1.38956532e-01 -7.26441205e-01 5.24470747e-01
-1.70647049e+00 -7.39811838e-01 1.19699739e-01 1.80071318e+00
5.57989478e-01 3.94019306e-01 -4.22253102e-01 -1.67870224e-01
4.36877728e-01 7.57149160e-01 -5.29014945e-01 -2.22020075e-01
-1.22288585e-01 6.20015740e-01 8.83100331e-01 6.78233624e-01
-1.29223204e+00 1.42884111e+00 6.26570082e+00 5.52501321e-01
-1.25707793e+00 -2.47008875e-02 4.30853307e-01 3.26490104e-01
-1.41081333e-01 3.55656862e-01 -7.48085558e-01 7.81549662e-02
4.44981098e-01 2.97082394e-01 2.19737574e-01 1.28966868e+00
8.40106010e-02 -7.59525895e-01 -4.71231222e-01 5.48781633e-01
2.54079789e-01 -1.54685640e+00 -5.14395535e-01 1.43028926e-02
1.03944755e+00 1.50291130e-01 1.69196371e-02 1.06350714e-02
5.42160571e-01 -9.52906966e-01 5.58496475e-01 7.55683362e-01
7.68418610e-01 -5.85115790e-01 7.26725399e-01 6.84458017e-02
-1.37543571e+00 -1.02222092e-01 -3.96697909e-01 -7.07510114e-02
1.10831641e-01 8.24423015e-01 -1.62390733e+00 1.15926370e-01
8.40110719e-01 3.10916334e-01 -5.72468221e-01 9.14250493e-01
-6.54742122e-01 7.77644277e-01 -2.94880062e-01 -8.62776935e-02
2.49898627e-01 4.67038639e-02 1.54155910e-01 1.65075266e+00
-3.12821940e-02 -9.93106961e-02 2.01840624e-01 6.65665984e-01
-5.97336777e-02 -4.36948448e-01 -7.15269744e-01 3.88882160e-01
2.08807096e-01 1.71366203e+00 -1.25074577e+00 -5.01307011e-01
-4.15888727e-01 1.29543483e+00 2.15437397e-01 4.38187748e-01
-8.71406317e-01 -7.65469253e-01 4.42599386e-01 8.29645768e-02
6.83172286e-01 -3.28283072e-01 -4.90876436e-01 -6.08459234e-01
-2.82491203e-02 -2.39611477e-01 2.98631173e-02 -9.53741431e-01
-9.63507533e-01 6.42951071e-01 -7.40719140e-02 -7.98434138e-01
3.49313319e-02 -8.14369261e-01 -1.05121350e+00 7.58470654e-01
-1.83346558e+00 -1.49921250e+00 -1.06793702e+00 5.84727407e-01
9.53832507e-01 -1.04006939e-02 8.59672546e-01 -3.60100418e-01
1.50614521e-02 1.51753277e-01 9.14298519e-02 2.29598165e-01
8.31225336e-01 -1.47077537e+00 6.01511300e-01 9.31664646e-01
1.43514052e-01 2.29497537e-01 7.58223951e-01 -7.75528491e-01
-8.90105665e-01 -1.11796856e+00 7.22328424e-01 -1.64286166e-01
1.94090962e-01 -5.79982281e-01 -5.87187052e-01 4.58211869e-01
3.93044561e-01 4.89602923e-01 3.01335722e-01 -1.37847401e-02
-8.78235474e-02 -3.09962034e-01 -1.08651161e+00 3.56818050e-01
1.08165896e+00 -4.57515985e-01 -4.23326164e-01 5.02269864e-01
3.86148006e-01 -5.39536238e-01 -2.88201690e-01 5.14970422e-01
9.06489074e-01 -1.53609145e+00 8.76883745e-01 -1.89082071e-01
3.38261634e-01 -3.22476894e-01 -3.43609661e-01 -1.02941084e+00
-1.05776317e-01 -6.72799468e-01 -1.53135911e-01 7.67208397e-01
2.94896275e-01 -3.00276041e-01 1.17182112e+00 9.99493450e-02
-4.60334331e-01 -1.16048014e+00 -6.20420694e-01 -2.94201255e-01
-7.19772696e-01 -3.34133595e-01 3.77513975e-01 1.14619553e-01
-8.63313198e-01 1.19759217e-01 -3.88520062e-01 5.16514301e-01
7.47836232e-01 1.03163064e+00 9.49391663e-01 -1.28095531e+00
-2.07922786e-01 4.05192524e-02 -3.39406319e-02 -9.78736937e-01
3.83136481e-01 -3.15614730e-01 6.64239883e-01 -1.73070550e+00
2.26291090e-01 -5.54600060e-01 -1.34415090e-01 4.58701313e-01
-2.94528753e-01 2.55941153e-01 1.49692282e-01 -9.68457907e-02
-4.04933572e-01 7.34081149e-01 1.14844179e+00 8.02442953e-02
-2.30962023e-01 4.42074746e-01 -1.75290465e-01 1.14909518e+00
9.05408144e-01 -4.48903590e-01 -3.95168424e-01 -1.49323538e-01
-6.05877757e-01 -2.06854045e-01 7.24721491e-01 -1.27243757e+00
1.37042612e-01 -5.56174338e-01 1.06843746e+00 -9.83333290e-01
9.11484540e-01 -8.13367128e-01 -3.17696959e-01 3.18977356e-01
-9.64360386e-02 -4.48909014e-01 3.30500640e-02 6.49092913e-01
2.78874874e-01 -5.40956743e-02 8.04862738e-01 -1.55727014e-01
-7.99738228e-01 1.90662056e-01 -3.45797211e-01 -3.55295062e-01
9.49717999e-01 -1.46340549e-01 -1.76822051e-01 -4.35784131e-01
-1.81402773e-01 -3.40597123e-01 3.57635111e-01 2.80408084e-01
9.69673753e-01 -7.94684052e-01 -2.90886998e-01 4.07341957e-01
-2.22226620e-01 5.95855713e-02 -4.26238626e-02 3.17692995e-01
-5.24549484e-01 4.85143900e-01 -4.67269450e-01 -6.23973727e-01
-1.34979475e+00 3.25209759e-02 3.57963771e-01 -1.22405149e-01
-9.29800272e-01 1.02588880e+00 4.37032253e-01 -1.34606570e-01
4.15728331e-01 -3.59169275e-01 2.16067508e-01 -3.23269337e-01
1.39785871e-01 1.38454422e-01 -1.61268979e-01 -4.91164804e-01
-3.13294291e-01 4.73468035e-01 3.00858289e-01 -2.15974167e-01
1.51938713e+00 4.01799530e-01 3.02731693e-01 3.13855857e-01
9.45069253e-01 3.33750427e-01 -2.27762151e+00 -2.96454191e-01
-1.72362238e-01 -7.21858740e-01 2.61451602e-01 -9.34437871e-01
-1.15920103e+00 7.81847715e-01 7.96165407e-01 -2.86166459e-01
1.28202260e+00 -8.70364383e-02 8.46906841e-01 5.81536293e-01
2.95922995e-01 -1.22009861e+00 2.03245953e-01 3.88901472e-01
7.96070874e-01 -1.37053370e+00 3.34345043e-01 -4.91960555e-01
-6.54094219e-01 1.42690635e+00 6.20068371e-01 -6.60565972e-01
9.71524715e-01 7.03435898e-01 2.94614404e-01 -2.88412541e-01
-4.02633011e-01 -3.59734207e-01 4.08143669e-01 8.36064041e-01
2.79791534e-01 2.62042016e-01 1.36023378e-02 3.00817430e-01
-1.42416790e-01 -6.25222027e-02 3.61606866e-01 1.16353106e+00
-9.06783819e-01 -7.66694844e-01 -5.43898761e-01 5.53147614e-01
-1.00998476e-01 -1.00601412e-01 -6.52729034e-01 8.47480834e-01
4.19230402e-01 6.99334443e-01 -1.76640287e-01 -3.77970368e-01
-7.91154876e-02 7.35559985e-02 4.47917581e-01 -5.52665710e-01
-2.46760756e-01 2.25667432e-02 3.04690786e-02 -6.99854076e-01
-4.04378116e-01 -7.51637399e-01 -1.34254503e+00 2.20982656e-01
-4.45925206e-01 -2.84066767e-01 7.48187244e-01 1.04534483e+00
2.12645873e-01 7.10551143e-01 5.74655116e-01 -1.40782475e+00
-8.06306750e-02 -1.03363323e+00 -4.36483324e-01 -9.47265178e-02
5.23048997e-01 -8.87110591e-01 -5.84783442e-02 3.38807195e-01] | [10.851943016052246, -4.117950916290283] |
1fe1c481-c468-4e70-9c24-42655490ec59 | simultaneous-translation-and-paraphrase-for | null | null | https://aclanthology.org/2020.ngt-1.28 | https://aclanthology.org/2020.ngt-1.28.pdf | Simultaneous Translation and Paraphrase for Language Education | We present the task of Simultaneous Translation and Paraphrasing for Language Education (STAPLE). Given a prompt in one language, the goal is to generate a diverse set of correct translations that language learners are likely to produce. This is motivated by the need to create and maintain large, high-quality sets of acceptable translations for exercises in a language-learning application, and synthesizes work spanning machine translation, MT evaluation, automatic paraphrasing, and language education technology. We developed a novel corpus with unique properties for five languages (Hungarian, Japanese, Korean, Portuguese, and Vietnamese), and report on the results of a shared task challenge which attracted 20 teams to solve the task. In our meta-analysis, we focus on three aspects of the resulting systems: external training corpus selection, model architecture and training decisions, and decoding and filtering strategies. We find that strong systems start with a large amount of generic training data, and then fine-tune with in-domain data, sampled according to our provided learner response frequencies. | ['Burr Settles', 'Will Monroe', 'Bill McDowell', 'Klinton Bicknell', 'Stephen Mayhew', 'Chris Brust'] | 2020-07-01 | null | null | null | ws-2020-7 | ['multilingual-nlp'] | ['natural-language-processing'] | [ 3.65774870e-01 -7.06544593e-02 -1.37118727e-01 -3.57303560e-01
-1.61019099e+00 -1.05791414e+00 6.10405564e-01 1.62864700e-01
-6.65179551e-01 9.93028283e-01 4.87407774e-01 -9.08245265e-01
2.52425224e-01 -3.87088537e-01 -9.53270137e-01 2.27433905e-01
7.51455307e-01 9.78683531e-01 8.95491149e-03 -7.82525361e-01
2.20221385e-01 1.93466693e-02 -1.20882916e+00 8.68612707e-01
1.64615381e+00 2.50921790e-02 5.24148107e-01 6.30115330e-01
-1.51445761e-01 6.66286349e-01 -7.59448647e-01 -7.78353631e-01
-3.66980024e-02 -7.36187220e-01 -1.17829216e+00 -2.79294550e-01
7.40037262e-01 1.08135544e-01 2.72854298e-01 7.14437664e-01
8.35045338e-01 4.50371020e-02 5.76128483e-01 -7.79455066e-01
-1.04318094e+00 1.13561749e+00 1.04462549e-01 3.20889473e-01
7.54125714e-01 3.49363774e-01 8.15468669e-01 -1.12936735e+00
1.01833034e+00 9.74108636e-01 5.37860513e-01 9.48911548e-01
-1.36101437e+00 -5.19546568e-01 -7.60069117e-02 -3.57826464e-02
-9.53450501e-01 -7.97868371e-01 2.11451262e-01 -4.84081179e-01
1.30694318e+00 3.52974594e-01 5.72663009e-01 1.67609751e+00
1.96693093e-01 8.81866336e-01 1.18249738e+00 -9.46815550e-01
-7.87451044e-02 6.68178380e-01 -1.76781788e-01 4.08645838e-01
-6.46243170e-02 -1.71630755e-01 -9.51776624e-01 6.99766353e-02
2.26773500e-01 -7.94544399e-01 -4.14586067e-01 -1.31498531e-01
-1.47567701e+00 6.27182961e-01 -2.66912341e-01 2.90965527e-01
1.94972187e-01 -3.59239101e-01 3.36510539e-01 1.14353693e+00
2.66420215e-01 1.15478969e+00 -8.49884033e-01 -6.91403627e-01
-8.16072762e-01 3.65017474e-01 1.03255975e+00 1.30730283e+00
3.05142462e-01 -1.82552874e-01 -2.20505670e-01 1.22321415e+00
2.26243120e-02 6.55841887e-01 9.72951353e-01 -7.38358438e-01
1.25520349e+00 3.81722897e-01 -6.24676794e-02 -7.06329867e-02
8.13926831e-02 -6.85180068e-01 -1.16889298e-01 -1.77250892e-01
6.45545900e-01 -5.10276258e-01 -5.73274612e-01 1.94582534e+00
1.26474509e-02 -4.44534749e-01 3.56551111e-01 4.90582764e-01
8.74159813e-01 6.00399911e-01 -5.01461104e-02 -2.39198625e-01
9.58891392e-01 -1.19508064e+00 -2.80299604e-01 -6.15699291e-01
1.05250561e+00 -1.37805641e+00 1.70685327e+00 4.21346635e-01
-1.68429708e+00 -6.26546025e-01 -8.28315318e-01 -3.93966466e-01
-1.93811193e-01 1.77990362e-01 7.39011243e-02 5.37394404e-01
-1.19481194e+00 5.08981764e-01 -5.32185018e-01 -4.85122323e-01
-2.30150118e-01 1.85244516e-01 -2.19572008e-01 -1.76437050e-01
-1.09771585e+00 1.49599326e+00 1.44736737e-01 -4.65280533e-01
-6.60470843e-01 -6.98716462e-01 -6.21075928e-01 -1.54834285e-01
-1.96907029e-01 -8.17239225e-01 1.84100199e+00 -1.21794879e+00
-1.83861113e+00 1.04508460e+00 9.34779420e-02 -4.93790805e-02
6.35174453e-01 -3.16369534e-01 -1.75142258e-01 -4.50945169e-01
2.52157927e-01 5.45925498e-01 3.01030070e-01 -6.38374686e-01
-6.83231652e-01 -4.49174456e-02 -2.64318407e-01 8.10047030e-01
-4.57298309e-01 3.44606280e-01 -2.73063451e-01 -5.21656036e-01
1.38437813e-02 -8.69343877e-01 -1.81129232e-01 -8.87478173e-01
7.78920129e-02 -2.72078842e-01 -1.26556531e-01 -8.83947492e-01
1.21528137e+00 -1.80404735e+00 5.09375215e-01 -5.14043160e-02
-4.06852335e-01 6.58445880e-02 -6.43784344e-01 6.39515638e-01
1.14687406e-01 2.09900558e-01 4.11582254e-02 -3.17059517e-01
8.94400664e-03 -1.58245578e-01 -4.54110116e-01 -6.50059059e-02
1.88971713e-01 8.95344794e-01 -1.12909520e+00 -5.54597497e-01
-2.49719247e-01 -8.06515440e-02 -7.13066697e-01 3.77080590e-01
-4.32820052e-01 4.73123103e-01 -2.99913198e-01 4.47899193e-01
-8.60511437e-02 2.42269002e-02 1.93267345e-01 4.65908647e-01
-2.93330044e-01 1.21878624e+00 -8.57686579e-01 2.03118014e+00
-1.02117169e+00 6.17606878e-01 -1.21817432e-01 -4.37990218e-01
9.14853394e-01 3.87887567e-01 2.63382122e-02 -9.59579051e-01
-5.23878746e-02 8.76614213e-01 1.99861810e-01 -5.32791793e-01
7.30696082e-01 1.12798080e-01 -1.29821002e-01 9.02282715e-01
3.78010958e-01 -5.92144966e-01 5.53341508e-01 1.55134618e-01
9.90881085e-01 4.61455941e-01 6.47890493e-02 -3.95458311e-01
3.65848899e-01 4.85146016e-01 2.41981938e-01 6.48959219e-01
9.33320634e-03 5.74787796e-01 8.40675682e-02 -6.91572502e-02
-1.12645221e+00 -9.81301188e-01 -2.88388599e-02 1.61039114e+00
-3.91868621e-01 -5.05594432e-01 -9.76633966e-01 -7.04522550e-01
-3.26936215e-01 9.82406378e-01 2.02385802e-02 -1.84950173e-01
-1.00624514e+00 -3.92413139e-01 4.89529669e-01 2.40982190e-01
-8.19112435e-02 -1.15376651e+00 -3.29975843e-01 4.48416799e-01
-7.17478633e-01 -8.83490801e-01 -8.51552427e-01 2.62104154e-01
-8.52176368e-01 -5.89377522e-01 -5.93982637e-01 -1.31984699e+00
4.11654472e-01 -4.06850837e-02 1.82149756e+00 1.84538979e-02
2.80788034e-01 1.55060843e-01 -2.32155830e-01 -4.82623458e-01
-1.08443677e+00 7.33483434e-01 4.91837114e-02 -7.96842456e-01
5.04462123e-01 -4.79123175e-01 5.38079925e-02 2.04943717e-01
-3.93501729e-01 4.66617316e-01 7.93814480e-01 9.67734098e-01
3.47739339e-01 -7.52651989e-01 6.22089267e-01 -8.90937328e-01
1.18951929e+00 -3.83583218e-01 -4.01827037e-01 8.21746171e-01
-5.00798762e-01 1.08623005e-01 8.18814337e-01 -6.38219535e-01
-9.58362460e-01 -9.13134068e-02 -2.65748233e-01 1.46190703e-01
-3.93183017e-03 5.70936501e-01 -1.51530385e-01 -3.29699810e-03
1.28142118e+00 4.53380495e-01 -2.34751001e-01 -5.31980097e-01
5.24810076e-01 8.51015389e-01 4.16939408e-01 -1.15681124e+00
5.75217366e-01 -7.62446940e-01 -7.49763966e-01 -6.33349717e-01
-6.94842339e-01 -1.68668389e-01 -5.30516684e-01 5.00737801e-02
3.83461505e-01 -1.17118418e+00 -3.77871655e-02 1.33533478e-01
-1.24467099e+00 -8.09343100e-01 -4.06872988e-01 6.72742069e-01
-6.39995933e-01 -1.14571378e-01 -8.15254629e-01 -2.23392949e-01
-4.00951028e-01 -1.53859890e+00 7.74950266e-01 7.66639188e-02
-7.33891666e-01 -1.01986766e+00 5.54972887e-01 7.51048744e-01
6.04380608e-01 -4.23463255e-01 1.11332428e+00 -1.00180304e+00
-4.59957212e-01 4.19347197e-01 3.43001455e-01 2.87564278e-01
-2.95086652e-01 2.23479606e-02 -5.84047139e-01 -4.37621564e-01
-1.52822807e-01 -1.00851786e+00 3.50666821e-01 -9.84243304e-02
5.69158971e-01 -3.80866945e-01 1.46628055e-03 5.59995890e-01
9.42260861e-01 -1.35444999e-01 3.65164578e-02 6.06927693e-01
4.19521362e-01 7.04228222e-01 4.08887476e-01 -1.93400040e-01
5.36095381e-01 8.01565945e-01 -3.96504492e-01 2.18691066e-01
-1.60830408e-01 -6.65970862e-01 9.45248306e-01 1.58402789e+00
2.68571049e-01 -2.68812090e-01 -1.09568095e+00 8.83137167e-01
-1.50019038e+00 -7.25031316e-01 -4.37354483e-03 2.26733541e+00
1.68886411e+00 1.62006915e-01 1.92014515e-01 -5.11320353e-01
3.10820520e-01 -3.91437650e-01 -1.91510975e-01 -9.10962820e-01
-1.55855631e-02 6.58959329e-01 3.53998363e-01 7.80835152e-01
-4.53124702e-01 1.31882215e+00 6.78856277e+00 9.90829229e-01
-1.11365199e+00 1.76494211e-01 4.67342466e-01 -1.86700732e-01
-8.69913280e-01 -1.12034485e-01 -1.01753914e+00 4.22862440e-01
1.55499876e+00 -5.70864737e-01 7.15517461e-01 5.77933013e-01
4.15707901e-02 4.46861386e-01 -1.44992983e+00 5.25542259e-01
2.02347785e-02 -1.20361614e+00 2.09556982e-01 -2.99648196e-01
1.15415621e+00 4.55280662e-01 5.88997193e-02 7.48443544e-01
8.05473089e-01 -1.20800769e+00 8.78886521e-01 3.73896286e-02
9.06478167e-01 -6.16064131e-01 3.71099472e-01 7.80633390e-01
-5.15472054e-01 1.24775499e-01 -3.70644420e-01 -2.18965307e-01
-1.77358568e-01 1.47169800e-02 -1.15430748e+00 2.59162456e-01
2.48798758e-01 3.79895449e-01 -8.04103136e-01 7.55278111e-01
-5.02366126e-01 9.00255203e-01 -2.24128902e-01 -4.81223583e-01
8.82384181e-02 -2.46069700e-01 3.42666775e-01 1.46485889e+00
4.17729855e-01 -2.23278105e-01 3.19078356e-01 8.11723232e-01
-3.07083637e-01 6.42810702e-01 -6.33401632e-01 -8.34323764e-02
7.31628418e-01 1.00569785e+00 -3.46381366e-02 -2.73060918e-01
-4.70620602e-01 8.30573857e-01 6.43079817e-01 2.96747774e-01
-3.06592733e-01 -1.76415861e-01 4.73824471e-01 7.04902038e-02
-3.47766578e-01 -2.76045740e-01 -6.35564864e-01 -1.46026492e+00
6.01667427e-02 -1.87102365e+00 2.22399622e-01 -5.58140695e-01
-1.08792973e+00 6.89416170e-01 -2.12715685e-01 -1.17966223e+00
-6.93209767e-01 -5.08295596e-01 -6.37187302e-01 1.42341948e+00
-1.07935846e+00 -8.97857547e-01 2.25753054e-01 3.34104270e-01
1.08988822e+00 -5.92346728e-01 8.23706448e-01 3.46429765e-01
-3.70554984e-01 9.43571329e-01 3.21402878e-01 -1.11364655e-01
1.13663638e+00 -1.38961577e+00 1.03110468e+00 8.88737440e-01
5.00337660e-01 8.81982565e-01 7.33282804e-01 -6.26428545e-01
-1.39540410e+00 -8.73234093e-01 1.79338574e+00 -1.02622545e+00
6.98149979e-01 -3.77319127e-01 -6.33443058e-01 7.37970114e-01
4.80890036e-01 -6.69827104e-01 6.55372202e-01 3.21626693e-01
2.94042919e-02 2.40345255e-01 -8.30999374e-01 8.66279542e-01
1.15959144e+00 -5.45758069e-01 -9.28760529e-01 6.96587205e-01
7.68040121e-01 -1.01912892e+00 -7.09811628e-01 2.79511064e-02
5.00949562e-01 -4.56108361e-01 7.07873762e-01 -1.13228321e+00
9.39227879e-01 1.62375599e-01 4.88993265e-02 -1.97026992e+00
-1.65183723e-01 -9.03602004e-01 4.08531070e-01 1.20174503e+00
1.08615935e+00 -3.21619362e-01 6.96668327e-01 5.30833960e-01
-4.51421261e-01 -8.06286514e-01 -6.45500958e-01 -6.70970380e-01
7.52762377e-01 -1.77145183e-01 5.17875731e-01 1.03049088e+00
3.87192398e-01 1.01580536e+00 1.18436895e-01 -3.05907905e-01
3.37031305e-01 8.96695033e-02 1.00661170e+00 -8.26284111e-01
-5.58906078e-01 -6.38644814e-01 5.18337071e-01 -1.24990964e+00
4.14917395e-02 -1.34219754e+00 1.22990839e-01 -1.47263491e+00
8.32878202e-02 -4.26779598e-01 5.62787801e-02 3.63225192e-01
-3.90010625e-01 1.00416243e-02 6.53661638e-02 1.11159779e-01
-4.83855397e-01 3.21253210e-01 1.30539250e+00 9.00630504e-02
-4.46957529e-01 1.98135883e-01 -8.70105386e-01 3.67784649e-01
6.68693125e-01 -2.86681801e-01 -5.06183267e-01 -1.20032752e+00
4.02846575e-01 2.68416762e-01 -3.11780721e-01 -8.18472028e-01
3.05435508e-01 -6.43294036e-01 3.01125199e-01 -2.09351063e-01
-1.47313565e-01 -3.01874518e-01 -1.46017581e-01 2.63839334e-01
-9.39859927e-01 7.41184056e-01 2.88849920e-01 -4.88498926e-01
-1.43576011e-01 -4.17373449e-01 7.05121279e-01 -3.68707240e-01
-3.22659492e-01 -2.72648931e-01 -3.46246958e-01 8.09957862e-01
5.18737078e-01 -6.29684702e-02 -5.54580808e-01 -3.09607536e-01
-3.45452189e-01 4.77865309e-01 5.74248195e-01 7.26883292e-01
2.88308680e-01 -1.25125659e+00 -1.34043419e+00 2.92655796e-01
2.44939059e-01 -1.30753443e-01 -3.67250502e-01 5.20873249e-01
-5.24477124e-01 5.33229649e-01 -4.10123348e-01 -3.91791135e-01
-1.24583709e+00 -7.96018913e-02 3.62127274e-01 -5.81512094e-01
-1.34166598e-01 1.19685972e+00 -4.31364149e-01 -1.08798802e+00
1.04929604e-01 -2.58834571e-01 -9.34285447e-02 -1.62814498e-01
3.74911070e-01 1.38839364e-01 5.19198298e-01 -2.94060081e-01
-3.93745005e-02 2.44306177e-01 -1.55629531e-01 -6.21903539e-01
1.07640254e+00 1.02818469e-02 1.93987995e-01 4.27738965e-01
8.56407583e-01 4.92416769e-01 -6.51864350e-01 -3.71161222e-01
2.53227115e-01 -1.73109353e-01 -4.22673464e-01 -1.31786406e+00
-2.88007826e-01 8.42013419e-01 7.69138485e-02 -2.17064008e-01
7.55411744e-01 -2.72368938e-01 9.03061569e-01 6.55393660e-01
4.25821126e-01 -1.32972920e+00 6.70986623e-02 1.14217210e+00
9.29742694e-01 -1.02437818e+00 -3.00439388e-01 -8.17210078e-02
-6.74890995e-01 1.06368434e+00 1.03559601e+00 4.10887688e-01
-2.96700329e-01 9.05557722e-02 4.28470731e-01 3.37665528e-01
-1.33018649e+00 1.07487828e-01 5.28541684e-01 4.16112036e-01
1.24182081e+00 -2.01615915e-02 -7.07332909e-01 6.46522999e-01
-9.92508709e-01 -3.96730341e-02 5.98047793e-01 9.08938885e-01
-5.76861501e-01 -1.69503939e+00 -2.61847049e-01 1.79182068e-01
-4.29906905e-01 -5.13296127e-01 -8.40909362e-01 6.05200469e-01
-5.65744974e-02 9.66176808e-01 -2.60891169e-01 -3.44564080e-01
7.07124412e-01 4.24009293e-01 7.16555238e-01 -1.18794298e+00
-1.24062073e+00 -2.85484254e-01 4.50914264e-01 -1.62443608e-01
4.46460545e-01 -7.96555340e-01 -7.88300991e-01 -1.81014851e-01
-1.05755053e-01 6.28789604e-01 6.97558761e-01 1.06402576e+00
2.29574800e-01 3.61272782e-01 4.31917340e-01 -2.48005047e-01
-1.10709059e+00 -1.19789815e+00 2.48736516e-01 3.98098737e-01
7.78736696e-02 4.42627072e-02 -6.33096173e-02 5.44714406e-02] | [11.618091583251953, 10.29741096496582] |
1e63d14f-8a32-4314-83c0-bac87f0cb174 | applications-of-deep-learning-for-top-view | 2304.08193 | null | https://arxiv.org/abs/2304.08193v1 | https://arxiv.org/pdf/2304.08193v1.pdf | Applications of Deep Learning for Top-View Omnidirectional Imaging: A Survey | A large field-of-view fisheye camera allows for capturing a large area with minimal numbers of cameras when they are mounted on a high position facing downwards. This top-view omnidirectional setup greatly reduces the work and cost for deployment compared to traditional solutions with multiple perspective cameras. In recent years, deep learning has been widely employed for vision related tasks, including for such omnidirectional settings. In this survey, we look at the application of deep learning in combination with omnidirectional top-view cameras, including the available datasets, human and object detection, human pose estimation, activity recognition and other miscellaneous applications. | ['Gangolf Hirtz', 'Ana Cecilia Perez Grassi', 'Jingrui Yu'] | 2023-04-17 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-2.15784431e-01 -5.98847985e-01 -4.88184243e-02 -2.62492567e-01
1.68441057e-01 -6.14880264e-01 6.83750451e-01 -7.10104406e-01
-7.80651271e-01 4.86882925e-01 2.02577665e-01 -2.98804464e-03
1.04863137e-01 -4.32536572e-01 -6.51628733e-01 -6.72302961e-01
2.03461587e-01 3.30973417e-01 8.30031559e-02 1.76714987e-01
2.37396270e-01 8.28678429e-01 -1.47724581e+00 -2.80648395e-02
-3.84575963e-01 8.14507782e-01 5.24481475e-01 9.56196666e-01
9.35234010e-01 7.22641230e-01 -4.29458350e-01 -3.61683756e-01
5.71783423e-01 4.10693586e-01 -2.75660783e-01 5.49744487e-01
1.02604151e+00 -1.01852012e+00 -6.01663649e-01 7.48288274e-01
7.73799002e-01 1.50880078e-02 2.82020539e-01 -1.03011096e+00
-3.25783938e-01 -3.75669718e-01 -9.17216420e-01 5.90789139e-01
9.65188682e-01 1.57071307e-01 5.27198613e-01 -1.24539804e+00
4.74790186e-01 1.10412669e+00 7.34163821e-01 4.07860756e-01
-2.27975368e-01 -1.40221119e-01 1.71349436e-01 3.19088817e-01
-1.05471408e+00 -4.54480201e-01 5.30626655e-01 -5.06295502e-01
1.16977727e+00 -8.53383914e-02 1.18614805e+00 1.50687015e+00
4.57669944e-01 9.51225638e-01 7.38968670e-01 -2.08967328e-01
-7.56506845e-02 -4.70326953e-02 -1.19161271e-01 7.13207841e-01
6.51055932e-01 1.60478391e-02 -3.79002959e-01 3.10349852e-01
1.11138082e+00 9.20989096e-01 -3.67559493e-01 -1.20979774e+00
-1.35504878e+00 6.41823351e-01 2.45369032e-01 -1.76066607e-01
-2.05159381e-01 1.99967176e-01 6.01067841e-01 -1.49890296e-02
3.30845982e-01 3.49420041e-01 -7.49629796e-01 -2.11262956e-01
-3.06898773e-01 1.21389382e-01 2.49413192e-01 1.35112596e+00
-7.56325014e-03 3.15950781e-01 4.82550621e-01 8.05279732e-01
2.59866476e-01 7.21132696e-01 3.01118016e-01 -1.08874035e+00
6.37278140e-01 4.54844326e-01 3.09970349e-01 -8.24189365e-01
-8.76902819e-01 -5.62703550e-01 -7.97805130e-01 3.10488760e-01
3.59107435e-01 -5.34650922e-01 -3.25897217e-01 9.57491159e-01
3.34145188e-01 -1.82137609e-01 4.33367789e-02 1.26222837e+00
6.79698408e-01 4.64851677e-01 -5.92470229e-01 2.17595741e-01
1.56182516e+00 -1.38522458e+00 -4.51422572e-01 -7.70917475e-01
3.39892030e-01 -4.63960916e-01 9.26575005e-01 1.11741757e+00
-9.49790537e-01 -5.35957873e-01 -9.82402146e-01 -1.82448253e-01
-3.24411035e-01 6.71702564e-01 8.26040208e-01 7.31873691e-01
-1.09637094e+00 -2.02141749e-03 -8.38133097e-01 -6.51212752e-01
1.72560751e-01 3.78756076e-01 -9.23929572e-01 -4.45663571e-01
-6.22771859e-01 1.16304266e+00 -1.26111045e-01 3.83296281e-01
-1.22517371e+00 -2.82873541e-01 -1.01863825e+00 -5.70884570e-02
6.47121072e-01 -8.76186430e-01 1.15109062e+00 -2.03629956e-01
-1.63141775e+00 8.38235736e-01 2.76942760e-01 -3.10667902e-01
8.15273464e-01 -9.77010071e-01 -1.63980022e-01 7.84799010e-02
-9.29847062e-02 3.49458098e-01 7.11140931e-01 -6.38638735e-01
-8.46338511e-01 -1.09458911e+00 3.58417541e-01 8.94916058e-01
-2.89061755e-01 3.17587823e-01 -4.79086161e-01 -9.30577144e-03
-2.48271212e-01 -1.18365681e+00 -1.20738767e-01 3.16720814e-01
-1.59906164e-01 -1.07066363e-01 1.37728322e+00 -4.12868470e-01
1.98200285e-01 -1.64154315e+00 1.71327069e-01 -6.53735101e-01
1.89801157e-01 3.04434687e-01 6.04963541e-01 3.59731913e-01
8.13651308e-02 -7.72210181e-01 5.53091586e-01 -3.06272119e-01
-6.03435397e-01 2.15673819e-02 2.71442801e-01 1.08000076e+00
-5.16761959e-01 6.57494724e-01 -1.05300331e+00 -4.76514548e-02
9.97381389e-01 4.47589427e-01 -5.35587490e-01 4.57387269e-01
7.05802381e-01 5.86109936e-01 5.96818402e-02 6.45550549e-01
7.47325420e-01 -1.80656567e-01 1.44093469e-01 9.60351378e-02
-3.22933972e-01 7.99167380e-02 -1.14230347e+00 1.60086679e+00
-8.63931715e-01 1.12381828e+00 2.55370706e-01 -8.93849373e-01
6.55788720e-01 2.92711854e-01 4.47846830e-01 -4.74500149e-01
4.81946021e-02 -2.63300359e-01 -3.91986161e-01 -6.02141201e-01
4.92852449e-01 1.32469296e-01 -1.03894703e-01 -4.75090444e-02
2.54679710e-01 1.94363534e-01 -9.57691148e-02 -1.16143651e-01
8.43576968e-01 2.90230185e-01 9.61591363e-01 -3.73880118e-02
6.55819893e-01 -4.21096563e-01 2.40530506e-01 4.81821299e-01
-1.91440269e-01 8.07049632e-01 -9.47605595e-02 -1.34924841e+00
-1.13738823e+00 -9.55681145e-01 7.35884309e-02 1.12969053e+00
2.55342394e-01 -1.90933511e-01 -5.14868796e-01 -6.45985484e-01
-2.55785912e-01 1.08585313e-01 -2.48301372e-01 1.31801382e-01
-8.60268712e-01 -2.99997181e-01 2.04163894e-01 1.16611874e+00
8.79942894e-01 -7.39675462e-01 -1.65268517e+00 -4.15321320e-01
-5.95606416e-02 -1.41644549e+00 -3.60515267e-01 7.22172335e-02
-8.38934541e-01 -1.23146892e+00 -1.18435240e+00 -4.66553450e-01
4.22984511e-01 1.32701862e+00 1.13931751e+00 -6.98459148e-01
-2.05630362e-01 8.77674162e-01 3.57085802e-02 -5.13998687e-01
7.42619038e-01 -2.30160266e-01 5.48467338e-01 -1.61564797e-01
5.28189123e-01 -2.74675459e-01 -9.35439944e-01 5.78405201e-01
-1.83255941e-01 -6.42565638e-02 6.95958436e-01 7.02771187e-01
-2.67786384e-02 -2.64130741e-01 -4.25579637e-01 -4.45140719e-01
1.55249655e-01 -3.18634272e-01 -1.04829061e+00 -4.01343927e-02
-1.21773623e-01 -7.51430869e-01 9.09404397e-01 -2.07311049e-01
-1.04929745e+00 4.00499970e-01 1.07599460e-01 -5.13679743e-01
-5.12756884e-01 -8.18213969e-02 -3.22486460e-01 -1.11459546e-01
3.94778997e-01 2.66236126e-01 -2.40596414e-01 -4.75876361e-01
-4.61270940e-03 7.46020913e-01 4.78262484e-01 3.14110965e-01
3.22719842e-01 9.43554223e-01 7.84125850e-02 -1.41410959e+00
-7.20266223e-01 -9.24692690e-01 -1.25533330e+00 -6.10551834e-01
8.96983922e-01 -1.49666905e+00 -8.12327623e-01 8.69861841e-01
-1.33542418e+00 1.09045841e-01 2.25593358e-01 9.38599706e-01
-5.80079794e-01 2.66994953e-01 -3.19908887e-01 -8.30716789e-01
-7.25884363e-02 -1.08430839e+00 1.58018625e+00 3.83976430e-01
2.86008362e-02 -1.14496839e+00 2.54696663e-02 7.40237176e-01
8.86200294e-02 -4.48080339e-02 -1.50000632e-01 -2.91031096e-02
-3.56572300e-01 -3.30592960e-01 7.22838789e-02 3.55541915e-01
6.76808730e-02 -3.85246217e-01 -9.90165293e-01 -5.86272240e-01
1.23155482e-01 -4.22718316e-01 3.44976783e-01 9.33378577e-01
1.01499736e+00 -2.23161817e-01 -4.80401337e-01 9.98890936e-01
1.24208224e+00 4.21973854e-01 5.87732911e-01 4.56644952e-01
1.03788686e+00 2.32888684e-01 5.46481133e-01 7.46310711e-01
2.43327051e-01 1.11092269e+00 6.34704590e-01 -1.48822844e-01
2.58105904e-01 -1.31164312e-01 5.20421028e-01 3.08338135e-01
-5.53941190e-01 -4.37010437e-01 -9.09143567e-01 4.81816322e-01
-1.45448220e+00 -1.22781014e+00 -9.19088945e-02 2.15844154e+00
-2.61654466e-01 -1.19912066e-01 3.86008352e-01 5.87612251e-03
6.49445117e-01 2.63433367e-01 -4.53243762e-01 -2.06342548e-01
4.49115843e-01 -7.10559487e-01 8.97838414e-01 3.76513511e-01
-1.31237495e+00 5.68415344e-01 7.39980650e+00 -2.72330828e-02
-1.25416267e+00 4.67622466e-02 -1.48353502e-02 -4.31143999e-01
6.08534932e-01 -4.29353297e-01 -1.27640092e+00 3.88993561e-01
5.56425393e-01 6.56743586e-01 2.69074976e-01 1.66891026e+00
2.71474272e-01 -4.65116143e-01 -1.20687020e+00 1.60918689e+00
5.55576265e-01 -9.62832332e-01 -3.18307042e-01 4.41152751e-01
5.46403468e-01 4.10697073e-01 2.33897448e-01 3.70849669e-02
9.32218581e-02 -6.39916182e-01 5.08485913e-01 9.96146798e-02
6.50264561e-01 -6.76111877e-01 9.14720356e-01 7.61023343e-01
-1.02676535e+00 -6.68821454e-01 -8.70722294e-01 -6.80697083e-01
3.58117282e-01 3.33736628e-01 -1.04646051e+00 1.85783625e-01
1.25746465e+00 1.05178237e+00 -5.60194314e-01 9.04603958e-01
-1.98993891e-01 -1.52915478e-01 -2.24338084e-01 -7.45918229e-02
2.50150263e-01 -5.39097369e-01 5.70463002e-01 1.23795164e+00
2.19150022e-01 -1.32690817e-01 1.75274998e-01 1.55823585e-02
5.68937249e-02 -5.37889719e-01 -1.61777318e+00 8.08234990e-01
-5.88971637e-02 1.37195325e+00 -5.67676842e-01 -2.14082375e-01
-9.15028274e-01 1.08916199e+00 2.30286315e-01 -1.89889129e-03
-8.24329793e-01 -2.45162442e-01 5.30304730e-01 3.55322868e-01
4.49917138e-01 -8.39296103e-01 2.07728982e-01 -1.57897615e+00
2.60848701e-01 -6.31342471e-01 2.45803297e-01 -1.33052433e+00
-5.97441912e-01 3.51282269e-01 3.27898532e-01 -1.53797603e+00
-2.82107979e-01 -1.26976526e+00 -3.70589912e-01 1.82403445e-01
-1.30205297e+00 -1.11464810e+00 -9.31965232e-01 8.92583489e-01
1.22881877e+00 -5.91018438e-01 4.76258159e-01 3.36873561e-01
-4.33221698e-01 2.35862449e-01 2.07403287e-01 3.18099558e-01
5.46347260e-01 -1.36106956e+00 4.18490499e-01 1.02483606e+00
1.68100104e-01 8.07507992e-01 4.65738684e-01 -5.66730928e-03
-1.77748263e+00 -7.82088280e-01 3.68667305e-01 -9.53901410e-01
-1.04030579e-01 -7.25115657e-01 2.04441752e-02 1.06541312e+00
4.94887710e-01 5.56966186e-01 4.60593104e-01 1.98272150e-02
1.07216276e-01 -3.29928905e-01 -9.53153253e-01 5.42596638e-01
1.28311622e+00 -1.64233029e-01 -4.04978245e-01 4.84530121e-01
-2.28894781e-02 -6.33778512e-01 -3.53706509e-01 1.32536590e-01
9.33640718e-01 -1.32476425e+00 1.18202770e+00 -4.91900235e-01
2.43363142e-01 -1.87642828e-01 -1.97505653e-01 -1.33152604e+00
-5.74039876e-01 -4.39788073e-01 -1.61387250e-01 3.62414300e-01
-1.85541376e-01 -4.41157997e-01 1.01511824e+00 1.21301990e-02
-3.27466726e-01 -6.27056599e-01 -9.05375421e-01 -7.96684563e-01
-6.29431546e-01 1.13984354e-01 -3.62918258e-01 3.40765893e-01
-1.27422377e-01 7.46486902e-01 -1.13197815e+00 2.34784886e-01
5.15240610e-01 -3.69812906e-01 1.22924352e+00 -1.30821013e+00
-1.86352745e-01 2.58087873e-01 -6.91411555e-01 -1.60305595e+00
-3.39948833e-01 -8.47897828e-02 -1.85270146e-01 -1.55471253e+00
2.95129985e-01 3.05208832e-01 2.34738320e-01 -1.16140634e-01
3.05895776e-01 2.22522408e-01 2.19071563e-03 -3.13504599e-02
-9.06897724e-01 2.98689216e-01 1.05179572e+00 -5.08312993e-02
3.03777814e-01 6.07174039e-01 -5.46014071e-01 1.37950015e+00
5.77680767e-01 -2.96090215e-01 -6.63169384e-01 -9.72247124e-01
2.91972786e-01 9.47607607e-02 4.81938094e-01 -1.38397801e+00
3.41308534e-01 -1.14042811e-01 9.54463601e-01 -8.32979321e-01
7.89643824e-01 -1.16773808e+00 -1.04261182e-01 7.25321829e-01
1.01280421e-01 8.57139528e-01 -1.56074017e-01 6.28763020e-01
6.34905919e-02 -1.48495197e-01 1.04992294e+00 -8.42285275e-01
-9.21402812e-01 1.78762093e-01 -7.66638994e-01 -1.39378563e-01
1.33075726e+00 -6.61149800e-01 -1.14202119e-01 -7.30163455e-01
-1.54958323e-01 -1.82017624e-01 5.43559909e-01 6.60685122e-01
9.70180333e-01 -1.11264920e+00 -2.61613190e-01 4.11616534e-01
-5.05253440e-03 8.95824358e-02 1.38635725e-01 8.44765365e-01
-8.34877729e-01 1.27419031e+00 -7.85558403e-01 -1.00357652e+00
-1.76723182e+00 8.43211353e-01 4.08547372e-01 -6.81742728e-02
-1.00174129e+00 7.00909257e-01 4.29798096e-01 -7.10047781e-01
2.25317851e-01 -2.46423811e-01 -5.44964075e-01 -3.88683707e-01
6.84529960e-01 8.71906340e-01 1.92543298e-01 -5.77350914e-01
-7.18787968e-01 1.06524444e+00 7.89199844e-02 -1.28252938e-01
1.47945499e+00 -3.13181013e-01 4.32854265e-01 2.03851178e-01
1.09999788e+00 1.96252346e-01 -1.75821447e+00 2.55622864e-01
-4.57784981e-01 -1.00613558e+00 7.64878765e-02 -3.13789010e-01
-1.13487506e+00 1.38913763e+00 7.57477582e-01 -4.05002624e-01
7.82354891e-01 -1.76154882e-01 3.70697260e-01 1.00453722e+00
8.98415029e-01 -1.15191543e+00 5.06475747e-01 6.50117517e-01
7.75743544e-01 -1.29053843e+00 3.78082633e-01 -1.28718585e-01
-8.29512835e-01 1.30991292e+00 1.12667322e+00 -2.36438021e-01
3.04516733e-01 4.16223347e-01 -4.77734506e-02 -2.74247646e-01
-6.34219825e-01 3.29918146e-01 -1.09027773e-01 1.20709944e+00
2.66698241e-01 -1.74572736e-01 2.27105796e-01 -2.51858741e-01
9.91392061e-02 -2.03886285e-01 1.00966632e+00 1.11968362e+00
-4.54275042e-01 -3.17230374e-01 -4.74709064e-01 1.28760864e-03
-5.59627235e-01 3.80108207e-01 -3.18148285e-01 8.85747254e-01
6.40974715e-02 8.19824696e-01 2.03926221e-01 -1.85448661e-01
5.37067294e-01 -4.09814119e-01 8.18892956e-01 -6.06081486e-01
-5.71312495e-02 -1.57619923e-01 -4.25058864e-02 -7.72593796e-01
-4.45375323e-01 -8.31218362e-01 -2.58595735e-01 -1.06975950e-01
-3.04145664e-01 -4.33104873e-01 6.91931367e-01 8.90993893e-01
3.08769584e-01 6.96404502e-02 6.18121147e-01 -1.51876116e+00
-5.22287130e-01 -8.49753320e-01 -5.72371185e-01 -9.76459458e-02
4.08970535e-01 -1.08545244e+00 -8.66523907e-02 -9.35563594e-02] | [7.272392749786377, -1.0471357107162476] |
4cb5520f-cb82-4c12-9599-c42d168233ad | moglow-probabilistic-and-controllable-motion | 1905.06598 | null | https://arxiv.org/abs/1905.06598v3 | https://arxiv.org/pdf/1905.06598v3.pdf | MoGlow: Probabilistic and controllable motion synthesis using normalising flows | Data-driven modelling and synthesis of motion is an active research area with applications that include animation, games, and social robotics. This paper introduces a new class of probabilistic, generative, and controllable motion-data models based on normalising flows. Models of this kind can describe highly complex distributions, yet can be trained efficiently using exact maximum likelihood, unlike GANs or VAEs. Our proposed model is autoregressive and uses LSTMs to enable arbitrarily long time-dependencies. Importantly, is is also causal, meaning that each pose in the output sequence is generated without access to poses or control inputs from future time steps; this absence of algorithmic latency is important for interactive applications with real-time motion control. The approach can in principle be applied to any type of motion since it does not make restrictive, task-specific assumptions regarding the motion or the character morphology. We evaluate the models on motion-capture datasets of human and quadruped locomotion. Objective and subjective results show that randomly-sampled motion from the proposed method outperforms task-agnostic baselines and attains a motion quality close to recorded motion capture. | ['Gustav Eje Henter', 'Simon Alexanderson', 'Jonas Beskow'] | 2019-05-16 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [ 1.60744309e-01 1.69151947e-01 -1.67285204e-01 -6.75658435e-02
-3.23123217e-01 -5.92522025e-01 9.65566576e-01 -4.74262595e-01
-4.16128069e-01 9.25065279e-01 3.89577150e-01 9.69767496e-02
4.51330096e-02 -9.17645156e-01 -7.57141411e-01 -9.32057977e-01
-2.65841752e-01 6.69756353e-01 5.14732122e-01 -3.40600222e-01
8.25827420e-02 5.42868495e-01 -1.62927580e+00 -5.24011673e-03
5.03681600e-01 4.18644369e-01 4.59136903e-01 1.51417804e+00
1.96258992e-01 1.02501082e+00 -5.68712115e-01 -8.00290480e-02
-7.87568316e-02 -7.27648079e-01 -5.36617815e-01 1.80876240e-01
1.06888674e-01 -3.73957723e-01 -4.63849396e-01 5.61249793e-01
6.67886078e-01 4.89720732e-01 1.00860763e+00 -1.27585053e+00
-3.52744311e-01 3.15382034e-01 -1.42027408e-01 -4.77039404e-02
3.38839442e-01 5.95019341e-01 6.54541731e-01 -5.34785032e-01
9.22078192e-01 1.24813831e+00 4.75790292e-01 1.07632816e+00
-1.37600386e+00 -1.28437519e-01 -4.26347591e-02 1.65251687e-01
-1.00943995e+00 -2.97686219e-01 8.01588237e-01 -7.19694555e-01
8.43910396e-01 8.72682184e-02 9.70727742e-01 1.69913769e+00
2.87186384e-01 9.40673470e-01 4.92839724e-01 -1.42575011e-01
6.65933490e-01 -4.41029847e-01 -6.62094653e-01 6.32816613e-01
1.12236897e-02 1.89596385e-01 -6.50672555e-01 -6.73563480e-02
1.19970512e+00 -3.78596604e-01 -2.68753946e-01 -7.62623429e-01
-1.62636352e+00 8.63389492e-01 9.09996778e-02 1.62558347e-01
-7.02755213e-01 8.59844983e-01 3.57319802e-01 -1.07793324e-01
1.27490669e-01 3.39552462e-01 -2.47301683e-01 -7.20284939e-01
-1.14023697e+00 6.27894223e-01 8.40677083e-01 1.04116762e+00
3.28166366e-01 7.12175369e-01 -1.24634750e-01 3.65522861e-01
2.57528037e-01 7.01857686e-01 6.83088005e-01 -1.44352531e+00
1.36872306e-01 4.19224836e-02 2.33198509e-01 -1.03616846e+00
-3.84590000e-01 7.51015395e-02 -9.90932584e-01 3.78984064e-01
4.45944935e-01 -4.62615490e-01 -1.00524688e+00 1.92311299e+00
3.04738343e-01 2.14632362e-01 1.76182911e-02 8.95649791e-01
4.85172808e-01 9.34764981e-01 2.34703884e-01 -6.05894998e-02
9.63432670e-01 -6.65682554e-01 -7.56509483e-01 -3.41558069e-01
3.82953107e-01 -3.94569039e-01 8.25760841e-01 3.35451037e-01
-1.14727068e+00 -6.18922591e-01 -8.99572372e-01 3.42849195e-02
1.64327342e-02 -4.42315377e-02 5.28368056e-01 4.60093319e-01
-1.14926791e+00 8.21532428e-01 -1.29197168e+00 -5.54353654e-01
1.32417336e-01 2.49495909e-01 -2.76157081e-01 4.15320814e-01
-1.02690554e+00 8.58466387e-01 3.67872149e-01 7.71683678e-02
-1.35921669e+00 -3.64615917e-01 -1.08063447e+00 -1.96006745e-01
2.46389192e-02 -1.16625130e+00 1.31617200e+00 -7.33925641e-01
-2.11439586e+00 4.41221416e-01 -1.96470901e-01 -6.19239807e-01
6.83323741e-01 -5.10289252e-01 8.97267461e-02 2.45984629e-01
-4.74298969e-02 1.12656808e+00 1.02227688e+00 -1.12208128e+00
-3.74184817e-01 1.31112963e-01 -8.28955621e-02 3.18552852e-01
-5.41856028e-02 -4.93782759e-01 -3.07472676e-01 -7.93395460e-01
-2.94577479e-01 -1.26158798e+00 -5.63491762e-01 2.49249324e-01
-4.32981789e-01 9.55972448e-02 9.03910756e-01 -5.19484401e-01
9.81639564e-01 -1.78826118e+00 8.42283249e-01 -2.12737843e-01
-2.79813170e-01 2.72395372e-01 -5.70185259e-02 6.60264015e-01
3.87021095e-01 3.62225473e-02 -4.94862527e-01 -2.31996953e-01
1.63288698e-01 5.29234350e-01 -3.19698721e-01 3.64986181e-01
3.96763206e-01 1.10391343e+00 -1.03282237e+00 -5.98446310e-01
6.55474842e-01 6.20798588e-01 -6.74150050e-01 4.27355796e-01
-7.04854310e-01 9.34642136e-01 -4.04642254e-01 1.56030893e-01
1.25619741e-02 -5.69282658e-02 1.49943873e-01 1.14558846e-01
-3.92723503e-03 -1.30220037e-02 -9.79866982e-01 2.07284784e+00
-4.71443713e-01 8.50351155e-01 -2.90890366e-01 -7.58888304e-01
8.46133649e-01 4.92500991e-01 5.54350674e-01 -1.83054432e-01
-3.53604406e-02 5.03758937e-02 6.88487962e-02 -7.05226541e-01
6.23373091e-01 -6.80671930e-02 -2.99381346e-01 2.96836317e-01
1.58028960e-01 -6.03293896e-01 2.34292164e-01 1.04416095e-01
1.07258022e+00 1.03533590e+00 3.67503494e-01 -7.31429905e-02
3.42988163e-01 1.72403798e-01 5.78350425e-01 5.41437566e-01
7.14904862e-03 7.91927993e-01 2.79916108e-01 -3.58174980e-01
-1.30701733e+00 -1.06978810e+00 4.22065586e-01 1.00138581e+00
3.72874774e-02 -1.60408363e-01 -7.54739583e-01 -2.43518636e-01
-4.30801421e-01 8.17503631e-01 -6.13910973e-01 -1.28314227e-01
-1.05344057e+00 -6.10254347e-01 4.63201493e-01 7.92182624e-01
2.67184556e-01 -1.61115646e+00 -1.26679456e+00 6.12569869e-01
-3.09722692e-01 -1.04552603e+00 -4.10509586e-01 -1.22825243e-01
-9.83482480e-01 -7.90127516e-01 -1.15367818e+00 -6.81890607e-01
3.23727012e-01 -3.54803264e-01 1.14410639e+00 -3.34489852e-01
-1.83791310e-01 4.69092071e-01 -2.28476658e-01 -1.30543381e-01
-6.17470205e-01 1.00748099e-01 -1.01804338e-01 -7.45028555e-02
-1.25544086e-01 -7.88342118e-01 -6.85730577e-01 1.75227895e-01
-8.42249751e-01 2.45088562e-01 4.40165848e-01 9.35013354e-01
5.18646240e-01 -3.40944469e-01 4.79950577e-01 -5.43045759e-01
4.19525295e-01 -5.02915680e-01 -2.72050798e-01 -1.77121371e-01
1.70762807e-01 3.49122822e-01 6.15769923e-01 -8.62397254e-01
-1.06501079e+00 5.26721179e-01 -1.48421928e-01 -5.05784690e-01
-4.83666301e-01 2.00767100e-01 -2.27499634e-01 3.32689196e-01
7.18735814e-01 5.57960391e-01 -1.09582804e-01 -1.01159945e-01
6.82580292e-01 1.00540720e-01 8.75442982e-01 -5.39292872e-01
7.98195481e-01 6.62587881e-01 3.24092478e-01 -1.35185146e+00
5.18750250e-02 -5.65001443e-02 -7.54161656e-01 -4.95940894e-01
1.16727483e+00 -6.93624556e-01 -6.91635489e-01 6.45545781e-01
-1.38537717e+00 -8.24956834e-01 -4.33996856e-01 5.65872490e-01
-1.39867699e+00 3.09931576e-01 -7.29572475e-01 -1.08266461e+00
-6.48495331e-02 -9.44296539e-01 1.13566673e+00 2.19119102e-01
-8.04179728e-01 -1.25487816e+00 3.14398199e-01 -2.82706708e-01
3.25799733e-01 9.20086026e-01 6.44211769e-01 -2.25098625e-01
-5.80098748e-01 -7.86728933e-02 6.68007612e-01 -5.08323163e-02
1.00729838e-01 4.52378064e-01 -7.50875473e-01 -2.12967500e-01
-2.69224137e-01 -3.81025285e-01 5.81698239e-01 7.00705886e-01
6.58935010e-01 -3.66791040e-01 -2.11129427e-01 4.21991497e-01
1.05169845e+00 2.72829264e-01 7.21668899e-01 1.29033610e-01
9.08464015e-01 8.72621834e-01 5.59914470e-01 5.33437014e-01
4.33551848e-01 8.38406265e-01 5.69302499e-01 1.53375939e-01
-1.63573191e-01 -6.39370143e-01 5.54900587e-01 8.35390687e-01
-3.10930133e-01 -6.08507216e-01 -7.24645257e-01 9.54193652e-01
-2.04115987e+00 -1.24429929e+00 -3.61270398e-01 2.09790611e+00
5.46807408e-01 -4.41576466e-02 3.36977422e-01 1.25272393e-01
7.18833804e-01 2.65633315e-01 -6.26213908e-01 -3.64866853e-01
-2.53031850e-02 1.85336694e-01 2.89285988e-01 4.76226002e-01
-1.07323539e+00 9.75497603e-01 6.50942230e+00 6.28530025e-01
-1.06034243e+00 -1.81936190e-01 1.96979001e-01 -8.16071332e-02
-2.62629271e-01 -8.33904967e-02 -4.02966350e-01 6.01054966e-01
9.87666845e-01 -4.96302173e-02 1.24975927e-01 8.70660841e-01
6.22192502e-01 -4.49255258e-02 -1.21083188e+00 6.82690978e-01
-2.70986885e-01 -1.31863046e+00 9.25366133e-02 1.03570364e-01
6.60663664e-01 -3.34672540e-01 3.91160361e-02 8.09316114e-02
6.69937968e-01 -9.87380445e-01 9.11109567e-01 8.25034618e-01
5.98252058e-01 -8.44994783e-01 3.51593971e-01 6.95809901e-01
-1.19240642e+00 2.06806853e-01 -1.44981042e-01 -7.74606913e-02
7.62337208e-01 3.38436365e-02 -6.44893944e-01 3.38890225e-01
3.42538476e-01 7.50929952e-01 -1.80645302e-01 9.55199361e-01
-4.90523100e-01 6.53893590e-01 -3.42276424e-01 -3.64226341e-01
1.79447919e-01 8.68118927e-02 8.73963654e-01 1.32656443e+00
4.74098951e-01 -1.86433330e-01 -2.68159434e-02 7.23383784e-01
5.07852256e-01 -2.26918042e-01 -9.37575698e-01 -5.67985773e-02
3.82415026e-01 8.93875837e-01 -7.29867637e-01 -3.23520958e-01
4.02374342e-02 1.23145187e+00 -7.89953843e-02 4.02286947e-01
-9.53714848e-01 -1.01786152e-01 5.96865058e-01 2.48438101e-02
6.58547819e-01 -7.39573717e-01 2.80811414e-02 -9.89680648e-01
-2.78620511e-01 -5.14985025e-01 1.32779563e-02 -6.70220435e-01
-1.07407749e+00 5.55585086e-01 2.21425712e-01 -1.34229922e+00
-1.28168225e+00 -3.58234853e-01 -6.64870679e-01 5.96925497e-01
-8.77541959e-01 -1.16632342e+00 -1.58497617e-01 4.57334638e-01
7.81346858e-01 4.32292335e-02 8.78453076e-01 -1.43586863e-02
-1.08842142e-01 2.59036701e-02 -2.21474499e-01 5.21352217e-02
4.02425140e-01 -1.30742836e+00 7.96291709e-01 9.83751655e-01
2.43713826e-01 1.88469559e-01 1.16592526e+00 -7.75018156e-01
-1.32010686e+00 -1.24895716e+00 6.42888069e-01 -5.60871542e-01
4.36613888e-01 -2.82431036e-01 -8.27807426e-01 5.54840326e-01
2.64585704e-01 -4.73313443e-02 2.96698779e-01 -5.97854495e-01
3.54427934e-01 4.48682576e-01 -9.20981348e-01 7.67695189e-01
1.09854877e+00 -1.17249936e-01 -4.61712778e-01 3.92326266e-02
5.15860915e-01 -3.79053473e-01 -6.50678158e-01 3.31227332e-01
6.48201048e-01 -9.65519607e-01 8.37330937e-01 -5.51980436e-01
6.07149303e-01 -4.34583068e-01 -1.21989530e-02 -1.33051980e+00
-3.65368277e-01 -9.12177563e-01 -4.85885680e-01 1.05281425e+00
1.82183325e-01 -2.91192140e-02 1.06403363e+00 5.58785498e-01
2.06894353e-02 -4.20774251e-01 -7.90283740e-01 -7.82692432e-01
1.79688290e-01 -4.29127604e-01 2.65799791e-01 6.14526629e-01
-3.38082045e-01 1.51931614e-01 -9.55180824e-01 1.12925828e-01
5.78741252e-01 -1.25802368e-01 1.20066011e+00 -8.71892691e-01
-5.47182143e-01 -3.37632924e-01 -7.15897381e-01 -1.25190187e+00
1.69437453e-01 -2.75964320e-01 6.15313947e-01 -1.68277776e+00
-2.80862093e-01 9.84886065e-02 5.43584347e-01 1.20778471e-01
-9.18551534e-03 1.47624880e-01 2.41583645e-01 2.14368254e-01
-3.01124990e-01 9.67411339e-01 1.42309105e+00 1.26263127e-01
-2.99613744e-01 5.82783744e-02 1.42288312e-01 1.00705469e+00
7.51997888e-01 -4.01094109e-01 -7.39633620e-01 -3.78758371e-01
-9.76922140e-02 4.80504841e-01 5.65178752e-01 -1.23195851e+00
3.85934770e-01 -4.09573525e-01 2.96521217e-01 -4.41928267e-01
7.00171947e-01 -5.13112187e-01 6.23855472e-01 7.48627305e-01
-3.85794073e-01 1.77359506e-01 -2.46175770e-02 8.75368297e-01
-1.31668914e-02 -2.25498434e-02 7.51026869e-01 -3.68472457e-01
-9.26856518e-01 2.50796825e-01 -1.00532079e+00 6.37386367e-02
1.18375444e+00 -3.83953065e-01 1.53794616e-01 -9.11977708e-01
-8.89389455e-01 -3.18399817e-02 6.09200954e-01 4.65002209e-01
6.41835690e-01 -1.42468178e+00 -7.34490693e-01 -1.59768298e-01
-1.73092589e-01 3.43908846e-01 2.73313195e-01 3.82870674e-01
-8.49374592e-01 2.37028375e-01 -4.25507218e-01 -8.71790290e-01
-8.07537794e-01 4.59514439e-01 1.57614738e-01 -1.29212499e-01
-7.14082897e-01 5.31682968e-01 2.84793854e-01 -2.45727345e-01
-1.59076646e-01 -1.23478264e-01 -2.70970225e-01 -2.29009584e-01
2.78721780e-01 4.57061946e-01 -5.56114733e-01 -9.35554147e-01
-2.04513997e-01 6.30597651e-01 5.11386573e-01 -6.52929485e-01
1.15671980e+00 -1.01278290e-01 4.21088487e-01 8.45352471e-01
6.74658835e-01 -3.48153561e-01 -1.84289181e+00 3.28689992e-01
5.75931296e-02 -1.91777050e-01 -4.06840324e-01 -2.52367377e-01
-8.82699668e-01 1.10772383e+00 1.78841889e-01 2.50904504e-02
7.91811824e-01 -2.29115397e-01 6.63200855e-01 2.20931873e-01
5.95031619e-01 -8.45907629e-01 3.60048920e-01 4.01904821e-01
8.81087184e-01 -7.51534879e-01 -4.17369664e-01 -1.18645623e-01
-9.89312828e-01 1.04952550e+00 7.02649653e-01 -6.05962157e-01
2.76702374e-01 3.52574825e-01 -9.95680690e-02 1.57316059e-01
-1.02634382e+00 -2.24955961e-01 1.75783828e-01 1.18729138e+00
3.36796999e-01 -1.14212103e-01 -2.15527236e-01 7.51351789e-02
-3.84317011e-01 6.20815344e-02 6.70499861e-01 1.02570045e+00
-2.76382476e-01 -1.02465403e+00 -1.59550071e-01 -6.05290271e-02
-1.87584475e-01 2.03993931e-01 -2.59501457e-01 1.03067660e+00
-8.83354321e-02 7.27473676e-01 1.84615791e-01 -3.14705163e-01
9.12180617e-02 2.41235849e-02 5.65293431e-01 -5.11032999e-01
-1.00407228e-01 1.23101033e-01 8.87093842e-02 -6.03973806e-01
-6.92693472e-01 -8.03529024e-01 -1.29820991e+00 -1.88341394e-01
8.19148123e-02 -2.84536313e-02 5.89853764e-01 7.08267331e-01
3.58543754e-01 5.35590053e-01 2.97751337e-01 -1.56092799e+00
-8.76838118e-02 -1.01385009e+00 -2.55233169e-01 4.37493026e-01
4.23702627e-01 -6.65124714e-01 -1.63769752e-01 6.65417075e-01] | [7.29831075668335, -0.16431424021720886] |
1866ea6c-fb5c-4815-9687-4f34d560bc4d | generalized-time-warping-invariant-dictionary | 2306.17690 | null | https://arxiv.org/abs/2306.17690v1 | https://arxiv.org/pdf/2306.17690v1.pdf | Generalized Time Warping Invariant Dictionary Learning for Time Series Classification and Clustering | Dictionary learning is an effective tool for pattern recognition and classification of time series data. Among various dictionary learning techniques, the dynamic time warping (DTW) is commonly used for dealing with temporal delays, scaling, transformation, and many other kinds of temporal misalignments issues. However, the DTW suffers overfitting or information loss due to its discrete nature in aligning time series data. To address this issue, we propose a generalized time warping invariant dictionary learning algorithm in this paper. Our approach features a generalized time warping operator, which consists of linear combinations of continuous basis functions for facilitating continuous temporal warping. The integration of the proposed operator and the dictionary learning is formulated as an optimization problem, where the block coordinate descent method is employed to jointly optimize warping paths, dictionaries, and sparseness coefficients. The optimized results are then used as hyperspace distance measures to feed classification and clustering algorithms. The superiority of the proposed method in terms of dictionary learning, classification, and clustering is validated through ten sets of public datasets in comparing with various benchmark methods. | ['Jianguo Wu', 'Yongxiang Li', 'Chao Wang', 'Ruiyu Xu'] | 2023-06-30 | null | null | null | null | ['clustering', 'dictionary-learning', 'time-series-classification', 'dynamic-time-warping'] | ['methodology', 'methodology', 'time-series', 'time-series'] | [ 1.76604763e-01 -8.90419662e-01 -1.55959010e-01 -1.07771322e-01
-3.89813215e-01 -5.23603499e-01 4.93966073e-01 2.99099535e-01
-4.79633033e-01 4.49983031e-01 3.38528991e-01 -1.06674068e-01
-5.87189376e-01 -5.12118995e-01 -3.29734683e-02 -1.11684906e+00
-8.13819394e-02 2.27876112e-01 3.13535356e-03 -2.99811214e-01
2.52822489e-01 5.82719088e-01 -1.13705277e+00 -3.27447474e-01
8.08525145e-01 8.79091978e-01 1.14222646e-01 3.91902626e-02
4.81170267e-02 3.18479657e-01 -2.68424004e-01 1.68437526e-01
5.22579193e-01 -3.86085212e-01 -1.70001686e-01 6.61842898e-02
-1.91970974e-01 3.99950519e-02 -5.79679191e-01 1.00629091e+00
5.84672868e-01 7.17049658e-01 5.07561386e-01 -1.20827305e+00
-4.14511204e-01 2.21701682e-01 -7.37505615e-01 6.16920114e-01
1.05492063e-01 -3.06323946e-01 8.68833303e-01 -9.20082510e-01
5.16946316e-01 9.21995223e-01 6.11687481e-01 -7.51828700e-02
-1.14309764e+00 -7.63693213e-01 -9.75083709e-02 5.77531636e-01
-1.68812716e+00 -2.54864573e-01 1.37832296e+00 -5.30120552e-01
6.66945696e-01 2.72268683e-01 7.80214131e-01 6.27179325e-01
4.01287317e-01 3.38373601e-01 1.09521210e+00 -4.37922925e-01
2.32926756e-01 -1.87081024e-01 3.99350040e-02 4.25867677e-01
-6.03776686e-02 2.76252091e-01 -3.35512549e-01 -1.82556570e-01
6.13907874e-01 3.86376321e-01 -3.45887721e-01 -3.87705535e-01
-1.49095547e+00 9.36083555e-01 2.63393015e-01 5.38260877e-01
-4.26815629e-01 -4.89721268e-01 6.04854524e-01 4.43433166e-01
4.45580095e-01 3.06780010e-01 -1.06829740e-01 -2.01450497e-01
-9.36645091e-01 9.02290866e-02 4.50289041e-01 6.15255713e-01
3.52905065e-01 4.91250604e-01 9.61870775e-02 9.53551948e-01
2.28487030e-01 5.61607540e-01 9.20418262e-01 -4.30869550e-01
4.20873433e-01 5.33865213e-01 -1.44796655e-01 -1.78236294e+00
-5.75486898e-01 -3.59624118e-01 -1.04532254e+00 -3.28048348e-01
1.49234250e-01 -2.61707772e-02 -5.79150736e-01 1.42201281e+00
5.21843910e-01 6.76327169e-01 -8.55523199e-02 1.08325696e+00
3.23303014e-01 9.15139198e-01 -1.24149747e-01 -8.69977176e-01
1.03732467e+00 -5.90416729e-01 -1.04784548e+00 1.89353794e-01
3.64385486e-01 -1.00529861e+00 6.39342844e-01 3.31397355e-01
-4.84841347e-01 -5.21573484e-01 -1.22369230e+00 1.82161301e-01
-2.58581221e-01 -1.08037286e-01 4.78709072e-01 2.36335471e-01
-5.13196945e-01 4.67127442e-01 -9.40926313e-01 -4.28583026e-01
-1.07048608e-01 1.88518196e-01 -3.08759302e-01 -6.74278662e-02
-1.18914878e+00 7.32881486e-01 4.40021455e-01 5.45793734e-02
-4.77489233e-01 -5.47658741e-01 -7.55486906e-01 -2.91595489e-01
5.71454614e-02 -2.10383274e-02 4.91439521e-01 -5.84800124e-01
-1.23675847e+00 5.68107307e-01 -1.09425820e-02 -4.04935807e-01
1.39666155e-01 2.06157148e-01 -9.92579818e-01 -3.65521424e-02
2.61416435e-02 -1.12589478e-01 1.10450876e+00 -6.61783993e-01
-3.84237081e-01 -4.30129290e-01 -3.09187323e-01 4.14960921e-01
-5.53665996e-01 -1.55733198e-01 -2.24202931e-01 -1.39991856e+00
6.53584301e-01 -9.07581866e-01 -1.10035956e-01 -1.60656065e-01
4.78364760e-03 3.32536772e-02 1.23638940e+00 -7.77641535e-01
1.62198138e+00 -2.44415474e+00 6.07777834e-01 5.36871552e-01
-5.18353619e-02 8.52266103e-02 -1.36545971e-01 7.53683925e-01
-3.48240584e-01 -5.55469096e-01 -3.60492021e-01 -4.77642789e-02
-2.77512044e-01 3.69037390e-01 -4.09506530e-01 8.01313400e-01
-1.80747956e-01 2.05756888e-01 -9.37156439e-01 -4.50685650e-01
4.96806175e-01 6.21604264e-01 -2.48876870e-01 1.17473438e-01
3.07080954e-01 7.23365545e-01 -3.02642584e-01 7.09698021e-01
4.88835722e-01 2.30969802e-01 1.04506820e-01 -6.45363867e-01
-4.08231109e-01 -1.02293618e-01 -1.42810202e+00 1.93801022e+00
-3.12645465e-01 5.02480567e-01 -1.11606307e-01 -1.52357721e+00
1.31735909e+00 4.21229720e-01 1.13055122e+00 -6.73732936e-01
4.56547886e-01 2.74769247e-01 1.84349120e-01 -5.44826746e-01
2.03422114e-01 -2.73902714e-01 1.13216400e-01 4.63755816e-01
-4.22450267e-02 -1.10694416e-01 8.53232443e-02 -3.09064478e-01
8.01192403e-01 -1.71999544e-01 4.70417440e-01 -4.46471244e-01
8.48100901e-01 5.79248890e-02 7.30225742e-01 1.36240767e-02
-1.02934435e-01 3.16625834e-01 1.03648461e-01 -6.98097169e-01
-1.13457119e+00 -5.89089930e-01 -3.77335578e-01 7.09081650e-01
3.45663935e-01 -3.03768516e-01 -1.36071473e-01 -2.60492623e-01
5.02642058e-02 3.92704040e-01 -4.73374069e-01 -3.87696028e-01
-6.41677260e-01 -8.33564758e-01 2.25438654e-01 1.92433253e-01
5.26794732e-01 -4.66264099e-01 -2.78350860e-01 4.70871031e-01
-8.50514993e-02 -8.42795372e-01 -9.10903096e-01 1.35968417e-01
-1.07023001e+00 -9.53818917e-01 -6.12684608e-01 -9.62061226e-01
5.18642664e-01 7.24443495e-01 2.00280875e-01 -5.15984856e-02
-3.55320007e-01 2.11644322e-01 -6.99327052e-01 -8.66033733e-02
-3.42267267e-02 -2.23505184e-01 4.53847498e-01 4.30980861e-01
3.25674891e-01 -9.65834856e-01 -5.77737570e-01 4.84849602e-01
-1.05503857e+00 -2.95021892e-01 1.34621143e-01 1.21004820e+00
8.69271040e-01 5.55935740e-01 4.36738044e-01 -3.66964549e-01
7.92600691e-01 -6.70383155e-01 -6.32375360e-01 1.24881372e-01
-6.40311241e-01 2.07512677e-02 6.94304824e-01 -8.03406358e-01
-6.10059142e-01 -8.85288194e-02 8.69309604e-02 -7.21788406e-01
4.83761579e-01 8.74872208e-01 7.15139136e-02 -3.95781785e-01
5.50127804e-01 6.28221154e-01 1.58215776e-01 -5.49519658e-01
1.36704519e-01 5.60712636e-01 2.77309448e-01 -5.92174649e-01
9.46703255e-01 5.63444376e-01 1.22709490e-01 -8.94681871e-01
-4.26392525e-01 -6.12812638e-01 -7.04348564e-01 -2.45843962e-01
6.87106669e-01 -7.58264124e-01 -2.19321564e-01 3.53712142e-01
-7.53384292e-01 3.35858285e-01 -1.87638164e-01 7.84603894e-01
-1.69818088e-01 5.31355500e-01 -2.65595734e-01 -4.85325813e-01
-2.63702184e-01 -1.06134772e+00 5.82568705e-01 -1.02299452e-01
-3.62911038e-02 -1.31344545e+00 3.05136055e-01 2.01013796e-02
3.74107897e-01 6.58470571e-01 9.45175827e-01 -6.60735309e-01
5.86634763e-02 -6.17620289e-01 2.99136519e-01 3.20420086e-01
4.19132173e-01 -1.78178504e-01 -2.86415040e-01 -6.73725545e-01
5.56649268e-01 1.29852295e-01 3.54868025e-01 1.23666212e-01
1.00854397e+00 -4.05974925e-01 -2.15426788e-01 8.66204679e-01
1.56945884e+00 8.45975816e-01 2.07140848e-01 3.19126725e-01
6.89505756e-01 4.00071323e-01 8.78948152e-01 8.43993783e-01
2.12713003e-01 8.40634644e-01 7.90478066e-02 1.75428718e-01
3.57998282e-01 -1.11831114e-01 1.63768768e-01 1.62404597e+00
-1.94796622e-01 2.94322371e-01 -1.11980307e+00 5.51239669e-01
-1.82458329e+00 -1.05159974e+00 1.10289581e-01 2.21873522e+00
7.08364666e-01 -8.65586922e-02 -8.93013179e-03 6.24758720e-01
7.06611574e-01 4.94509548e-01 -6.36265457e-01 -2.08730459e-01
-1.84317201e-01 3.14802647e-01 3.98687780e-01 3.34245056e-01
-1.21586561e+00 4.14130121e-01 5.44609547e+00 8.61874938e-01
-1.65578282e+00 3.43681753e-01 -2.61744130e-02 1.56050250e-01
-1.69562012e-01 2.30803359e-02 -9.03245807e-02 6.45883858e-01
4.19621110e-01 -4.71504360e-01 8.48851442e-01 4.53759879e-01
5.69103003e-01 3.17384094e-01 -6.49094224e-01 1.47404385e+00
1.70575425e-01 -1.00640893e+00 -5.92048131e-02 -1.63366303e-01
6.88551307e-01 -1.87133223e-01 1.10368624e-01 8.15532729e-02
-2.86611199e-01 -7.69941926e-01 5.05407155e-01 3.81104589e-01
5.87360740e-01 -8.17976475e-01 5.80239773e-01 2.64601320e-01
-1.49745905e+00 -1.95415378e-01 -1.61613733e-01 -1.32652864e-01
2.38070890e-01 5.46814740e-01 -3.16024065e-01 7.12035477e-01
3.14584851e-01 1.34738278e+00 -1.12180494e-01 1.00671625e+00
2.44891286e-01 3.51593971e-01 -4.83793765e-01 3.41543049e-01
1.86304003e-01 -7.47014582e-01 7.48921216e-01 8.50023448e-01
6.10414922e-01 5.61009347e-01 3.35047573e-01 2.47046530e-01
3.59003037e-01 4.01696324e-01 -6.50061429e-01 -1.25429779e-01
9.09608245e-01 1.23780215e+00 -7.46143639e-01 3.27535876e-04
-4.55022871e-01 6.88399076e-01 -1.92913041e-01 2.43804678e-01
-9.12843347e-01 -4.52653050e-01 5.10199904e-01 -1.68418791e-02
6.12244420e-02 -7.76593387e-01 -8.86664912e-02 -1.40141511e+00
-3.47088948e-02 -9.69186723e-01 5.69457889e-01 -2.79647887e-01
-1.28532541e+00 5.65219283e-01 3.56230326e-02 -1.94189095e+00
2.11939380e-01 -3.29035878e-01 -6.37692988e-01 6.70864999e-01
-1.29274940e+00 -9.29630518e-01 -3.96077037e-01 1.20111442e+00
5.85383236e-01 -5.16999483e-01 7.00978637e-01 8.33663642e-01
-6.46964610e-01 5.02435029e-01 4.95880365e-01 -1.90323237e-02
7.04289138e-01 -7.56336808e-01 -3.85699451e-01 9.58058894e-01
7.83261061e-02 6.76241815e-01 8.33848119e-01 -4.63967532e-01
-1.68109059e+00 -1.07937467e+00 4.95964885e-01 3.24371010e-01
9.12267506e-01 1.12987570e-02 -8.77835572e-01 4.31369632e-01
-1.28758714e-01 1.25695616e-01 8.17804515e-01 -2.52845079e-01
-3.74097735e-01 -4.69612360e-01 -1.06940985e+00 3.24638009e-01
7.18701780e-01 -5.62331080e-01 -6.80980146e-01 4.60075170e-01
6.34362400e-01 -4.71574724e-01 -1.34290683e+00 3.73089939e-01
5.91446877e-01 -4.44598049e-01 9.40811157e-01 -3.34483683e-01
2.45002899e-02 -5.60290098e-01 -4.05881643e-01 -1.33725035e+00
-5.41182816e-01 -5.72821736e-01 -1.32333422e-02 1.14696920e+00
-1.62672117e-01 -7.40381956e-01 2.97164589e-01 -5.21342130e-03
-7.56064951e-02 -6.10154212e-01 -1.24196517e+00 -7.85795212e-01
-2.20642090e-01 -1.33170351e-01 5.11059463e-01 1.67029583e+00
2.04529613e-01 1.98246762e-01 -5.88392377e-01 2.36021895e-02
6.97885513e-01 1.60837963e-01 2.91962713e-01 -9.88829911e-01
-1.41630873e-01 -2.22815573e-01 -5.33666670e-01 -7.66598523e-01
1.98755581e-02 -1.04722738e+00 -2.13642240e-01 -9.08735454e-01
-4.56091583e-01 -6.98348999e-01 -5.26635885e-01 2.71279246e-01
7.76076391e-02 -1.13450646e-01 1.12042487e-01 7.23275602e-01
-1.12888857e-03 7.52884924e-01 1.31182218e+00 -3.64741802e-01
-4.08110678e-01 -1.99680716e-01 -1.93927377e-01 3.95449758e-01
7.32542813e-01 -4.75958347e-01 -6.38788462e-01 -5.51669836e-01
-1.26631513e-01 4.57482100e-01 -3.15250158e-02 -1.20913112e+00
4.63269174e-01 -3.80868286e-01 3.95607278e-02 -4.76314217e-01
3.75320435e-01 -1.15104258e+00 5.67080617e-01 4.99647915e-01
-8.29943195e-02 5.60741842e-01 2.70916242e-02 7.04241037e-01
-6.89804673e-01 1.21379174e-01 7.30171680e-01 2.32522324e-01
-7.76946306e-01 4.67202693e-01 -3.50534432e-02 -1.34542555e-01
1.14418089e+00 -2.43927643e-01 6.80455714e-02 -7.08266497e-02
-7.20828950e-01 1.81628972e-01 2.84809053e-01 5.56669354e-01
7.35647619e-01 -1.76841950e+00 -6.00387573e-01 3.65386575e-01
1.53674737e-01 -2.65145212e-01 2.60065317e-01 1.24123359e+00
-5.88206768e-01 3.07096511e-01 -5.12740374e-01 -5.02944112e-01
-1.08475888e+00 6.56222820e-01 1.74747109e-01 -1.59301788e-01
-5.99164724e-01 4.39908385e-01 -4.70109016e-01 -7.55925775e-02
2.00254381e-01 -1.40566960e-01 -3.79201502e-01 5.31801641e-01
2.62441069e-01 3.98161858e-01 6.48172796e-02 -9.22723532e-01
-4.52289551e-01 9.83846784e-01 2.10573643e-01 -8.49860460e-02
1.38179183e+00 -1.87736660e-01 -4.07010108e-01 6.02323472e-01
1.41243505e+00 -1.16403438e-01 -8.20480704e-01 -6.34365737e-01
5.39108505e-03 -5.95086455e-01 1.00632302e-01 -1.70155585e-01
-1.08754480e+00 7.67695487e-01 7.52039313e-01 2.34880865e-01
1.46371150e+00 -6.88946009e-01 9.99727011e-01 1.29833400e-01
5.62957644e-01 -9.97541785e-01 -5.68199977e-02 5.81224620e-01
9.17064071e-01 -1.11858869e+00 2.46906251e-01 -3.78559045e-02
-4.72779959e-01 1.24391508e+00 2.48439252e-01 -4.46109653e-01
1.07214558e+00 5.57184219e-03 -5.69816791e-02 1.23112435e-02
-4.36816394e-01 5.61239384e-02 3.41881931e-01 4.30583596e-01
1.51029289e-01 1.58392899e-02 -9.28492665e-01 3.47406358e-01
-2.12553948e-01 -3.21082860e-01 8.01212117e-02 9.46296453e-01
-2.61707425e-01 -1.15783787e+00 -5.23922443e-01 2.75746882e-01
-9.16605070e-02 1.19885251e-01 1.71298191e-01 5.95196426e-01
3.13119441e-01 8.70658517e-01 5.73322512e-02 -7.27314770e-01
4.39189941e-01 -1.08864158e-01 2.15065897e-01 -4.31704223e-01
-3.26035678e-01 2.24983960e-01 -4.28994745e-01 -4.26747352e-01
-6.74386442e-01 -7.87009239e-01 -1.06856596e+00 -2.34750450e-01
-3.60905439e-01 3.11662257e-01 6.48206472e-01 8.16141486e-01
7.49504045e-02 3.53778481e-01 1.17111850e+00 -6.94287658e-01
-5.83512664e-01 -6.14568830e-01 -8.61938357e-01 6.89960778e-01
4.32110667e-01 -9.28023279e-01 -4.57208514e-01 1.56998664e-01] | [7.4920830726623535, 3.285782814025879] |
145e6cc9-f9ed-4e9b-a78c-46d9b4a6561c | confident-head-circumference-measurement-from | 1908.02582 | null | https://arxiv.org/abs/1908.02582v1 | https://arxiv.org/pdf/1908.02582v1.pdf | Confident Head Circumference Measurement from Ultrasound with Real-time Feedback for Sonographers | Manual estimation of fetal Head Circumference (HC) from Ultrasound (US) is a key biometric for monitoring the healthy development of fetuses. Unfortunately, such measurements are subject to large inter-observer variability, resulting in low early-detection rates of fetal abnormalities. To address this issue, we propose a novel probabilistic Deep Learning approach for real-time automated estimation of fetal HC. This system feeds back statistics on measurement robustness to inform users how confident a deep neural network is in evaluating suitable views acquired during free-hand ultrasound examination. In real-time scenarios, this approach may be exploited to guide operators to scan planes that are as close as possible to the underlying distribution of training images, for the purpose of improving inter-operator consistency. We train on free-hand ultrasound data from over 2000 subjects (2848 training/540 test) and show that our method is able to predict HC measurements within 1.81$\pm$1.65mm deviation from the ground truth, with 50% of the test images fully contained within the predicted confidence margins, and an average of 1.82$\pm$1.78mm deviation from the margin for the remaining cases that are not fully contained. | ['Alberto Gomez', 'Matthew Sinclair', 'Emma Robinson', 'David Lloyd', 'Nicolas Toussaint', 'Jacqueline Matthew', 'Bernhard Kainz', 'Samuel Budd', 'Bishesh Khanal'] | 2019-08-07 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 1.05335638e-01 9.17926729e-01 2.88471967e-01 -7.68985331e-01
-9.76733088e-01 -6.47983253e-01 -1.51991069e-01 3.15012366e-01
-2.60181516e-01 3.54680717e-01 -3.24102044e-01 -4.29504871e-01
-3.06091666e-01 -6.91296220e-01 -1.08780968e+00 -6.91470861e-01
-2.75348455e-01 6.89102352e-01 7.71312267e-02 5.81649125e-01
-9.30272192e-02 5.49870551e-01 -1.13317025e+00 2.54148841e-01
7.15181768e-01 1.11568677e+00 -1.26365617e-01 9.52404976e-01
8.81332681e-02 3.09532195e-01 -5.40899217e-01 -6.22184098e-01
1.33045390e-01 -2.74437428e-01 -3.96388918e-01 -1.29155979e-01
6.82687283e-01 -5.91242492e-01 -1.86404102e-02 9.12618876e-01
9.08126831e-01 -2.23002389e-01 8.81786168e-01 -5.79261661e-01
-1.51485011e-01 6.59149051e-01 -7.73963690e-01 1.53581798e-01
2.29366571e-01 -2.51252770e-01 4.58478838e-01 -7.09897280e-01
4.11476284e-01 4.60669994e-01 9.65597391e-01 4.52366918e-01
-9.88435388e-01 -8.45910609e-01 -5.06013155e-01 -2.55186826e-01
-1.50942838e+00 -6.73048079e-01 4.40128505e-01 -8.06776762e-01
4.08533841e-01 1.72296450e-01 5.67595541e-01 6.61558151e-01
3.82773161e-01 3.64150047e-01 6.23875678e-01 -4.39947724e-01
6.06164671e-02 1.40711412e-01 -4.92693007e-01 1.06295896e+00
3.50057095e-01 4.77967300e-02 -2.15761289e-01 -1.31376818e-01
1.01922512e+00 -4.20348108e-01 -2.07649365e-01 -6.09352171e-01
-6.06508493e-01 7.33322561e-01 7.79202208e-02 3.87140006e-01
-3.95332307e-01 -1.06524035e-01 1.96065381e-01 -1.65890515e-01
2.96862572e-01 5.15770674e-01 -3.56142372e-01 -2.44052559e-01
-1.20530307e+00 -1.89069677e-02 7.07878828e-01 9.82943773e-01
4.98811096e-01 -1.09859437e-01 -1.72116324e-01 9.03854072e-01
2.76843816e-01 6.78441763e-01 1.19839214e-01 -1.20300770e+00
4.08913434e-01 2.49344707e-01 6.47887290e-02 -1.06264925e+00
-8.26707244e-01 -5.91888964e-01 -9.35941458e-01 9.77473482e-02
5.85407674e-01 -6.35269880e-01 -9.37580943e-01 1.48980558e+00
3.56960535e-01 7.88397342e-02 -5.07749975e-01 6.85505271e-01
9.15368736e-01 2.65687197e-01 -2.96338707e-01 -3.50233525e-01
1.02124560e+00 -2.03568339e-01 -3.69192660e-01 -6.59244359e-02
4.93447453e-01 -6.58536077e-01 3.79035532e-01 5.15296400e-01
-1.38770092e+00 -3.93380821e-01 -9.58029151e-01 5.60513616e-01
1.73793077e-01 5.03526330e-01 1.47309333e-01 1.12149107e+00
-1.01521254e+00 7.82143950e-01 -1.39051139e+00 1.83610186e-01
7.04480112e-01 3.06565553e-01 -5.40831387e-01 -6.04508491e-03
-7.98561275e-01 9.02403235e-01 1.78705066e-01 3.31287771e-01
-5.58156908e-01 -9.96745884e-01 -1.11385047e+00 2.87831485e-01
2.27684408e-01 6.44580200e-02 1.31152678e+00 -4.62341577e-01
-1.32367074e+00 8.60512733e-01 -7.35033816e-03 -2.31304780e-01
8.66109312e-01 -2.61440158e-01 -2.12024495e-01 4.64643985e-01
8.37434903e-02 2.97153801e-01 7.59818733e-01 -1.11733496e+00
-3.89384627e-01 -4.51144606e-01 -4.69469547e-01 -2.23908409e-01
1.52852600e-02 2.30050180e-02 -6.26638651e-01 -2.83210844e-01
6.10391676e-01 -1.05110359e+00 4.89517674e-02 2.91907072e-01
-2.12744355e-01 1.94817513e-01 1.20537154e-01 -1.07067907e+00
8.93942118e-01 -1.89829183e+00 -4.35058594e-01 7.26451218e-01
4.31282640e-01 4.22673374e-01 2.00465664e-01 -6.19101673e-02
-1.88298523e-01 -2.63612028e-02 -1.68255061e-01 -2.01190740e-01
-2.59777218e-01 -3.71492747e-03 3.46525550e-01 9.49774742e-01
1.47162572e-01 6.74266875e-01 -9.25055623e-01 -6.16691411e-01
3.20465922e-01 4.52566892e-01 -3.32048357e-01 6.05633318e-01
3.63977492e-01 6.46489084e-01 -2.61935174e-01 5.37103832e-01
8.90073657e-01 -9.13723335e-02 3.58067662e-01 -1.42138392e-01
8.67406055e-02 -1.74617007e-01 -1.11330855e+00 1.54527259e+00
-5.40714264e-01 9.94000971e-01 1.83182403e-01 -9.28881705e-01
8.34250093e-01 6.52049005e-01 5.80735743e-01 -5.10921597e-01
3.22268635e-01 1.95326045e-01 4.28234160e-01 -7.83920467e-01
-2.33430982e-01 -7.74171948e-02 4.37305570e-01 4.64206785e-01
1.89453989e-01 -1.97740242e-01 1.34095192e-01 -2.80638635e-01
1.01719236e+00 -1.00027874e-01 2.98505008e-01 -1.94185540e-01
1.54002205e-01 -1.02656615e+00 5.21304488e-01 9.94873941e-01
-1.42230853e-01 1.42559934e+00 8.48616898e-01 -5.08474648e-01
-8.97002578e-01 -1.01887131e+00 -5.37054062e-01 6.79502845e-01
-2.26603299e-01 1.10835455e-01 -7.73323834e-01 -9.77172494e-01
7.98462480e-02 5.41697741e-01 -1.13935208e+00 6.80433214e-02
-5.43125689e-01 -4.12289053e-01 6.98508799e-01 7.37056732e-01
-1.19087934e-01 -8.99777353e-01 -9.84239221e-01 5.17499596e-02
1.29716769e-01 -8.73863935e-01 -8.72305334e-02 -7.55149201e-02
-4.73352641e-01 -1.21661842e+00 -1.25698614e+00 -3.11536789e-01
9.61948156e-01 -5.20811260e-01 1.09398317e+00 1.07461192e-01
-3.67142737e-01 1.01069748e-01 -4.13713902e-01 -5.01629889e-01
-8.49911988e-01 -8.08829144e-02 -4.71053012e-02 -2.40985714e-02
-4.08248119e-02 -4.98171479e-01 -8.36729527e-01 2.63920397e-01
-5.46094954e-01 -1.33350566e-01 7.05622613e-01 6.76872075e-01
2.61383802e-01 -4.12114263e-01 3.21821600e-01 -1.16665864e+00
-4.05624025e-02 -5.27677894e-01 -9.31183219e-01 4.06305909e-01
-6.91679060e-01 -1.19947642e-01 3.42103153e-01 -3.00732225e-01
-8.64188552e-01 -3.63404937e-02 -4.63662624e-01 -5.72837412e-01
-3.63073379e-01 5.52375436e-01 2.68865615e-01 -2.68510789e-01
7.86998689e-01 1.50702829e-02 -5.22612548e-03 -3.31330955e-01
-1.68250903e-01 6.16675615e-01 4.21726167e-01 -4.62391198e-01
4.01913106e-01 1.15255877e-01 1.60611764e-01 -7.15554953e-01
-7.32435822e-01 -2.77033836e-01 -5.97880006e-01 -4.14355993e-01
6.60186231e-01 -3.83447737e-01 -6.41149044e-01 4.54679541e-02
-1.05186689e+00 -1.60069495e-01 2.17436820e-01 7.60178387e-01
-5.05416870e-01 4.20203984e-01 -4.35399383e-01 -9.89995778e-01
-4.37374443e-01 -1.27552164e+00 1.07930124e+00 3.25549066e-01
-2.48125687e-01 -6.68944657e-01 -6.65885583e-02 2.03057244e-01
4.11982208e-01 4.28541899e-01 7.00562298e-01 -7.59808898e-01
3.14380713e-02 -8.49540055e-01 -4.07655597e-01 5.35193682e-01
1.26161307e-01 2.20590860e-01 -1.14721525e+00 -3.70956361e-01
-1.23143628e-01 -3.65890213e-04 4.16440755e-01 9.99607205e-01
1.47440016e+00 8.84579793e-02 -2.76848793e-01 6.56640828e-01
1.01651573e+00 3.71986300e-01 4.23002988e-01 -3.25844169e-01
4.29840416e-01 6.31142199e-01 5.02382219e-01 7.33547509e-01
1.89020008e-01 2.67802596e-01 4.65900064e-01 -1.83760703e-01
1.46193981e-01 5.39588109e-02 -5.14150679e-01 2.96036333e-01
-3.08369339e-01 -1.73511922e-01 -1.32829225e+00 6.26677632e-01
-1.37896192e+00 -4.22927082e-01 2.13718876e-01 2.46032214e+00
7.36391127e-01 2.56033570e-01 -4.27492261e-02 -1.32836979e-02
7.36655653e-01 -8.71823654e-02 -3.48080188e-01 -3.92264038e-01
4.80610490e-01 2.67150611e-01 5.41524172e-01 1.89551756e-01
-1.09759045e+00 1.33401632e-01 5.80306482e+00 4.06357646e-01
-1.11602437e+00 -4.33691144e-02 1.02108932e+00 6.73375204e-02
1.73261479e-01 -4.95503962e-01 -3.24479848e-01 5.99749327e-01
8.36286247e-01 3.34010959e-01 -7.92551860e-02 7.62937903e-01
-7.18983561e-02 -2.90721059e-01 -1.31772578e+00 9.51994717e-01
3.35552692e-02 -1.44721842e+00 -6.68126702e-01 -2.13326231e-01
8.54761481e-01 -1.55334339e-01 -8.52169096e-02 -2.01703757e-02
-4.91379231e-01 -1.25699568e+00 5.12059391e-01 6.74109876e-01
1.54804957e+00 -7.66488731e-01 1.31117141e+00 4.87329513e-01
-6.54548049e-01 3.91979694e-01 -3.07615131e-01 4.10126150e-01
-4.94790934e-02 6.57518983e-01 -1.39644158e+00 3.59667033e-01
8.35190773e-01 6.31252602e-02 -2.19180018e-01 1.19201100e+00
-1.69941336e-01 7.52029479e-01 -4.13675219e-01 2.33954996e-01
-1.81397453e-01 2.45430395e-01 3.46000552e-01 1.51332569e+00
6.39526486e-01 1.60498947e-01 -4.48009551e-01 7.79699862e-01
4.31402540e-03 5.90990037e-02 -4.19018656e-01 1.54036939e-01
5.68391025e-01 1.01826084e+00 -7.17526972e-01 -5.88087663e-02
-5.24530411e-01 4.06123400e-01 3.37234050e-01 1.22387521e-01
-7.44571328e-01 -5.65916419e-01 -8.28270167e-02 3.66217226e-01
3.81398857e-01 2.33951464e-01 -3.62319678e-01 -5.94422638e-01
2.13017777e-01 -5.48815250e-01 2.93783307e-01 -5.53926587e-01
-7.83599138e-01 5.48921824e-01 -1.08084209e-01 -1.19051766e+00
-5.63782275e-01 -6.12809420e-01 -6.34741366e-01 9.17570353e-01
-9.79897738e-01 -8.05526078e-01 -3.55087668e-01 -4.14404958e-01
1.38531640e-01 -1.71723142e-01 9.20419753e-01 4.00259852e-01
-3.18820596e-01 1.16522288e+00 -4.26550731e-02 4.82672423e-01
5.55228651e-01 -1.39898252e+00 2.00617358e-01 7.74909019e-01
-1.93299688e-02 6.07223332e-01 8.01276147e-01 -5.31126916e-01
-8.68141353e-01 -8.45326364e-01 4.58337963e-01 -3.70746255e-01
1.21326912e-02 1.76612679e-02 -1.02072358e+00 6.34936571e-01
-2.74847448e-01 5.65746844e-01 7.60835052e-01 4.92805094e-02
-5.23868203e-03 -1.22663997e-01 -1.41887915e+00 1.59461215e-01
4.28892970e-01 1.80648901e-02 -1.75829172e-01 1.00112118e-01
5.87095506e-02 -1.21298230e+00 -1.16376102e+00 8.90411139e-01
1.21489549e+00 -1.30164826e+00 5.70664048e-01 -2.80631214e-01
5.85521281e-01 2.16550559e-01 2.51196355e-01 -1.06272757e+00
-1.35053083e-01 -6.20047987e-01 -2.96333104e-01 9.69381154e-01
7.21767306e-01 -4.04435903e-01 1.25393665e+00 9.05292153e-01
-7.07734525e-02 -1.00043678e+00 -1.14837599e+00 -1.73711821e-01
1.26885712e-01 -7.45282710e-01 3.68495524e-01 7.16043890e-01
-2.44916841e-01 -5.12573063e-01 -7.04374686e-02 6.87614501e-01
6.30516410e-01 4.44562249e-02 3.47855628e-01 -9.82228816e-01
-3.64424318e-01 -4.04354960e-01 -6.05401158e-01 -6.52499914e-01
-1.49776280e-01 -3.77869666e-01 4.43722248e-01 -1.28479302e+00
2.52789974e-01 -5.62023342e-01 -3.26421916e-01 2.23988801e-01
-1.82086289e-01 3.37233156e-01 -2.95444280e-01 -5.78716636e-01
-3.15667421e-01 -6.49712533e-02 1.05239010e+00 3.67175639e-01
4.70328853e-02 9.18171108e-01 -3.20301652e-01 1.00900412e+00
6.09145522e-01 -5.22089839e-01 -1.24902561e-01 -6.05909228e-01
2.26432666e-01 7.39263058e-01 -1.60267293e-01 -8.85831594e-01
1.02819741e-01 6.71192929e-02 9.82004285e-01 -7.74631858e-01
9.01622549e-02 -6.35447264e-01 -6.81527108e-02 3.43376309e-01
-3.41877073e-01 -3.63212198e-01 2.96996891e-01 1.40415087e-01
1.78219199e-01 -6.47826910e-01 1.01875854e+00 9.67334732e-02
1.01560943e-01 3.26990545e-01 -2.98977733e-01 1.80272181e-02
6.94545090e-01 -2.02496409e-01 4.54690844e-01 -5.77551723e-01
-1.08634758e+00 2.30230033e-01 6.07321262e-02 9.90522206e-02
7.68766642e-01 -6.33583426e-01 -1.07713640e+00 5.57600319e-01
7.10184127e-02 3.30071479e-01 6.60505354e-01 9.48622406e-01
-1.08048689e+00 4.03210282e-01 2.01206714e-01 -9.39858079e-01
-1.20921421e+00 8.81108418e-02 6.42120779e-01 2.12133572e-01
-4.48133230e-01 1.40967321e+00 1.72535837e-01 -5.24152994e-01
3.81785184e-01 -3.42991471e-01 -2.14486703e-01 -2.18951225e-01
5.66803277e-01 5.32948434e-01 4.63088870e-01 -3.73489201e-01
-2.99521089e-01 6.86630845e-01 -4.87580970e-02 7.93462917e-02
1.12447333e+00 -5.03806956e-02 1.51213750e-01 1.78278133e-01
1.10340965e+00 3.24992418e-01 -1.27863872e+00 -1.07926138e-01
-1.83890343e-01 -6.96978390e-01 -5.95629811e-02 -8.85410309e-01
-1.38393271e+00 1.16398132e+00 9.83916163e-01 3.11575860e-01
7.96365499e-01 2.12690219e-01 1.86174408e-01 6.53525069e-02
1.03158355e-01 -7.51521409e-01 -4.58544910e-01 -8.52335319e-02
9.13120568e-01 -1.51987302e+00 1.21014871e-01 -3.14034730e-01
-5.37784338e-01 1.34231007e+00 6.37782633e-01 -4.11149561e-02
6.27534091e-01 3.59235555e-01 2.93233126e-01 -1.37253866e-01
-3.76857460e-01 4.29724604e-01 7.89268196e-01 6.87085807e-01
6.53347254e-01 2.85798669e-01 -3.38490084e-02 7.15158522e-01
-1.37306854e-01 -2.53180981e-01 5.01419425e-01 7.75886893e-01
-2.89189965e-01 -3.09821814e-01 -1.87636480e-01 9.81797993e-01
-1.03471398e+00 1.94473043e-01 3.81150812e-01 6.13470018e-01
3.20152253e-01 6.25636339e-01 5.87875955e-02 2.22101323e-02
3.11738223e-01 -7.29428828e-02 6.77341461e-01 -8.73073816e-01
-3.39613482e-02 2.89064676e-01 4.15799133e-02 -6.10360265e-01
2.51091808e-01 -9.47619379e-01 -1.12747514e+00 2.10004956e-01
-4.92897421e-01 9.38505903e-02 8.16285729e-01 1.01096940e+00
5.17936088e-02 4.70969290e-01 7.10310698e-01 -7.76323020e-01
-4.63692188e-01 -1.11981285e+00 -6.67370915e-01 -1.98273987e-01
6.68972135e-01 -5.76216638e-01 -2.33283997e-01 5.82496962e-03] | [14.2352294921875, -2.4078328609466553] |
9864dd77-892a-48b4-be51-41e9bc7f82d0 | ensemble-diffusion-for-retrieval | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Bai_Ensemble_Diffusion_for_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Bai_Ensemble_Diffusion_for_ICCV_2017_paper.pdf | Ensemble Diffusion for Retrieval | As a postprocessing procedure, diffusion process has demonstrated its ability of substantially improving the performance of various visual retrieval systems. Whereas, great efforts are also devoted to similarity (or metric) fusion, seeing that only one individual type of similarity cannot fully reveal the intrinsic relationship between objects. This stimulates a great research interest of considering similarity fusion in the framework of diffusion process (i.e., fusion with diffusion) for robust retrieval. In this paper, we firstly revisit representative methods about fusion with diffusion, and provide new insights which are ignored by previous researchers. Then, observing that existing algorithms are susceptible to noisy similarities, the proposed Regularized Ensemble Diffusion (RED) is bundled with an automatic weight learning paradigm, so that the negative impacts of noisy similarities are suppressed. At last, we integrate several recently-proposed similarities with the proposed framework. The experimental results suggest that we can achieve new state-of-the-art performances on various retrieval tasks, including 3D shape retrieval on ModelNet dataset, and image retrieval on Holidays and Ukbench dataset.
| ['Zhichao Zhou', 'Qi Tian', 'Song Bai', 'Longin Jan Latecki', 'Jingdong Wang', 'Xiang Bai'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['3d-shape-retrieval'] | ['computer-vision'] | [ 9.99491438e-02 -7.74425924e-01 -1.91821344e-02 -2.63168633e-01
-7.12603748e-01 -4.32298511e-01 1.10020316e+00 2.90855825e-01
-3.95315468e-01 1.76255807e-01 2.46790081e-01 1.85861111e-01
-5.63794911e-01 -6.12865746e-01 -4.88902107e-02 -1.09608579e+00
7.17614070e-02 1.02002241e-01 2.63970912e-01 -3.82552594e-01
6.45428538e-01 6.42365336e-01 -1.80484402e+00 1.50944486e-01
9.44102943e-01 1.18847370e+00 2.30268925e-01 3.01113248e-01
-2.76726305e-01 5.25682271e-01 -5.43301284e-01 -4.15596604e-01
2.90164471e-01 -2.56333619e-01 -4.46570873e-01 9.50113833e-02
3.82981330e-01 -9.51083079e-02 -4.13045108e-01 1.31758869e+00
8.83787870e-01 2.84277916e-01 9.27731931e-01 -1.10061562e+00
-1.31110299e+00 2.14114532e-01 -8.26602519e-01 4.53508615e-01
4.53271925e-01 2.38772412e-03 1.07590580e+00 -1.19484186e+00
6.04645491e-01 1.35053062e+00 3.38810265e-01 9.95059162e-02
-8.58380914e-01 -5.68033397e-01 1.61993459e-01 4.12672609e-01
-1.69961858e+00 -4.30510134e-01 9.88920987e-01 -2.82296211e-01
7.99976468e-01 3.40269119e-01 5.44571996e-01 8.31150949e-01
2.64347196e-01 1.11914897e+00 1.19659436e+00 -3.44373196e-01
-7.42825493e-02 1.06967680e-01 3.07850003e-01 4.90518808e-01
2.69990683e-01 2.13401746e-02 -2.98157811e-01 -2.02603847e-01
3.52037579e-01 3.24356914e-01 -3.12254220e-01 -6.42317891e-01
-1.30294001e+00 5.60728908e-01 5.04503846e-01 6.82606220e-01
-3.77091587e-01 -3.48373562e-01 3.69937360e-01 6.20481968e-01
7.13727295e-01 9.27878320e-02 2.35558525e-01 2.76427656e-01
-9.75298345e-01 2.04972342e-01 3.56123954e-01 8.45775008e-01
7.06843495e-01 -1.05533209e-02 -3.69775444e-01 1.22052288e+00
4.92851466e-01 8.68783712e-01 7.08028197e-01 -5.57247162e-01
8.99276584e-02 6.39829695e-01 -3.47980745e-02 -1.55901825e+00
-2.63287097e-01 -5.29623508e-01 -1.15788794e+00 9.55965966e-02
6.19199127e-03 3.16507280e-01 -8.28059554e-01 1.32293475e+00
3.01500291e-01 8.58643055e-02 2.71344423e-01 1.12345576e+00
1.15517688e+00 5.37878931e-01 -1.16781853e-01 -2.87596583e-01
1.08682644e+00 -1.13006902e+00 -8.48993838e-01 2.78601170e-01
4.22774404e-01 -1.36814070e+00 6.28229320e-01 3.85273010e-01
-1.14436567e+00 -8.91388535e-01 -1.08806157e+00 -5.88985533e-02
-5.38193941e-01 -7.41494372e-02 6.20594323e-01 4.05706972e-01
-1.11029375e+00 6.15506053e-01 -4.50715452e-01 -6.43740773e-01
1.18222795e-01 7.18365312e-02 -2.78617561e-01 -2.81016350e-01
-1.12062311e+00 1.12021172e+00 1.89291164e-01 2.32101902e-01
-6.67060971e-01 -2.52093613e-01 -5.42021930e-01 -1.17338583e-01
3.39959681e-01 -7.94350326e-01 7.13902116e-01 -6.50377512e-01
-1.05449688e+00 8.56421053e-01 -9.57737342e-02 -2.36018926e-01
5.24955809e-01 -2.80366063e-01 -7.45278478e-01 1.70398891e-01
-1.71678215e-01 5.12506604e-01 1.13813043e+00 -1.51751125e+00
-6.45189226e-01 -5.51869154e-01 -1.96120486e-01 6.06448174e-01
-6.60944939e-01 1.30542308e-01 -7.69395590e-01 -8.20074558e-01
3.95532191e-01 -8.30716312e-01 -1.66347623e-02 1.90375298e-01
6.16511367e-02 -5.15621006e-01 9.80300307e-01 -3.36241096e-01
1.40928984e+00 -2.20231819e+00 3.85079086e-01 4.79732275e-01
1.98251098e-01 5.14796615e-01 -4.40262705e-01 6.63508892e-01
1.15228780e-01 -1.14421755e-01 -2.42581531e-01 -5.13166785e-01
-2.63323970e-02 1.53600127e-01 -3.09409946e-01 5.64777613e-01
-1.43413886e-01 8.38573933e-01 -8.42308044e-01 -6.48065329e-01
3.08845907e-01 6.93624914e-01 -1.37558617e-02 1.78766266e-01
4.06148523e-01 2.43398994e-01 -6.22318327e-01 1.00169802e+00
9.43602800e-01 -1.28867164e-01 -3.20943534e-01 -4.11784291e-01
-1.30017042e-01 -3.44184101e-01 -1.35555720e+00 1.92293799e+00
-1.34985730e-01 3.28601301e-01 -1.01303205e-01 -9.68355954e-01
1.24636710e+00 5.02234399e-02 6.85534179e-01 -7.96476960e-01
1.26317114e-01 3.43966335e-01 1.11752838e-01 -5.50574243e-01
9.62883651e-01 2.24623084e-01 3.95476073e-01 5.62685192e-01
6.86940625e-02 -2.97386765e-01 3.23274463e-01 2.73832262e-01
7.21011579e-01 -1.02707982e-01 2.76420414e-01 -2.83444256e-01
8.98349106e-01 -2.59534895e-01 8.30281526e-02 7.75141776e-01
-3.32059711e-01 7.82798827e-01 -2.22472295e-01 -8.13153982e-02
-9.83700633e-01 -1.02439225e+00 -1.51229754e-01 7.81707168e-01
6.08074129e-01 -3.27039808e-01 -2.35679477e-01 -4.85836804e-01
2.29177594e-01 1.93586975e-01 -6.28112376e-01 -3.38971943e-01
-2.33818591e-01 -9.67900813e-01 5.20255744e-01 2.62047589e-01
7.69053519e-01 -1.06953335e+00 -2.81648561e-02 1.79208200e-02
3.76873463e-02 -7.44409382e-01 -4.82297599e-01 -3.54724139e-01
-8.75653088e-01 -9.49084699e-01 -1.38591421e+00 -8.87376726e-01
4.42117959e-01 1.19612277e+00 7.24756420e-01 3.91848922e-01
-1.21744610e-01 7.40513861e-01 -6.15987420e-01 -2.98834234e-01
-4.86432053e-02 -1.62962988e-01 1.05222262e-01 3.31530541e-01
4.68067139e-01 -5.20136535e-01 -8.03293824e-01 3.34734350e-01
-1.05357850e+00 -4.67210501e-01 7.42862046e-01 8.11676860e-01
6.60778284e-01 8.39731991e-02 4.35471296e-01 -4.14560914e-01
1.17299116e+00 -3.88500988e-01 -2.10375354e-01 6.49377108e-01
-9.59908485e-01 -8.38618800e-02 1.12637550e-01 -5.48863530e-01
-1.22954476e+00 -5.40192068e-01 3.81401069e-02 -5.99730313e-01
9.62284058e-02 4.20728981e-01 2.29490176e-01 -3.87981623e-01
3.89750391e-01 6.05416477e-01 9.89813581e-02 -5.99189341e-01
5.24495482e-01 7.06259489e-01 2.42785871e-01 -4.39974993e-01
7.52258599e-01 7.72270083e-01 -6.92239217e-03 -8.29378486e-01
-5.47972500e-01 -8.82628560e-01 -5.51980019e-01 -2.65567899e-01
5.22714496e-01 -7.94178426e-01 -4.16740477e-01 7.47094274e-01
-9.85030055e-01 5.15072703e-01 -1.17610224e-01 4.50123131e-01
-2.80968845e-01 9.78351057e-01 -4.03046578e-01 -8.39985669e-01
-5.97354054e-01 -1.09273422e+00 9.57366109e-01 2.98547596e-01
1.26568004e-01 -9.99608994e-01 1.52653307e-01 1.95811570e-01
7.05431223e-01 -2.90535629e-01 6.93426132e-01 -8.14335883e-01
-3.85668427e-01 -2.75945872e-01 -5.33828080e-01 5.89980602e-01
2.47367591e-01 2.00570524e-01 -8.47300172e-01 -4.18528557e-01
2.92478234e-01 -1.62209883e-01 1.19047558e+00 1.91729933e-01
8.37050796e-01 2.68950582e-01 -4.22471672e-01 4.13361490e-01
1.36118293e+00 3.49740952e-01 5.88534951e-01 2.37903595e-01
4.16582018e-01 5.58710694e-01 7.09961951e-01 4.05232519e-01
1.83465376e-01 5.41869402e-01 1.31030515e-01 -2.64780194e-01
-4.13741291e-01 5.82308769e-02 4.69850525e-02 1.31344068e+00
-3.57113808e-01 -4.45164621e-01 -6.81741476e-01 5.52961409e-01
-1.88306344e+00 -1.05892754e+00 -1.49233058e-01 2.15837884e+00
3.73914063e-01 1.62772797e-02 -2.52696872e-01 1.92448586e-01
7.94658363e-01 6.67748451e-01 -3.59215766e-01 -9.72150564e-02
-7.39513636e-01 2.43269261e-02 1.22653022e-01 3.01431149e-01
-1.04011750e+00 8.13516796e-01 6.04911804e+00 1.29067576e+00
-1.00457025e+00 9.30105895e-03 9.80114788e-02 1.87313557e-01
-4.93693471e-01 -5.82956597e-02 -4.61373776e-01 2.29254186e-01
7.55309165e-02 -4.08127546e-01 2.50504464e-01 4.31196511e-01
-1.83886103e-02 -2.42726281e-01 -6.58547163e-01 1.19982541e+00
5.18543839e-01 -8.15794408e-01 4.62258667e-01 -1.23873353e-01
1.02300179e+00 -9.06638056e-02 4.43193883e-01 1.72178999e-01
7.65757039e-02 -6.59277141e-01 2.71330118e-01 1.09586501e+00
2.16838300e-01 -7.78774083e-01 9.72137153e-01 2.73784637e-01
-1.21999264e+00 9.19765905e-02 -6.32223606e-01 2.42945239e-01
1.92653760e-01 7.14252949e-01 -2.53196418e-01 1.01733744e+00
5.55268407e-01 1.16023469e+00 -8.38647783e-01 1.45692432e+00
-1.50379464e-02 4.92129996e-02 -2.33955190e-01 -1.39638156e-01
3.67080092e-01 -3.25404733e-01 7.57391930e-01 1.16355217e+00
5.33877671e-01 2.42428005e-01 5.75455502e-02 4.96466190e-01
7.63384402e-02 4.56375539e-01 -7.92891264e-01 3.85210104e-03
2.67420709e-01 1.37486768e+00 -6.25016928e-01 -5.66247165e-01
-4.32577997e-01 1.01430523e+00 -6.41820729e-02 4.48759586e-01
-4.66678739e-01 -3.16804111e-01 4.83783215e-01 -4.87923384e-01
2.70485908e-01 -2.60108262e-01 -1.57841295e-01 -1.15721667e+00
1.83701161e-02 -7.12384045e-01 4.15665179e-01 -9.88927245e-01
-1.61751616e+00 6.57811344e-01 1.54539365e-02 -1.53545082e+00
1.54368788e-01 -1.33318394e-01 -4.39026058e-01 8.74721706e-01
-1.84888577e+00 -1.05839252e+00 -2.77431190e-01 7.49875128e-01
5.41387320e-01 -2.40867332e-01 5.92268229e-01 5.00295877e-01
-2.64467835e-01 4.40359235e-01 5.58471084e-01 -3.62627357e-01
1.16452360e+00 -9.05072033e-01 -2.88047809e-02 6.61011159e-01
4.83765513e-01 7.53186405e-01 5.49293220e-01 -6.10615432e-01
-1.31044590e+00 -6.18369997e-01 8.94100606e-01 -1.59698665e-01
7.66849995e-01 2.35765517e-01 -1.18531787e+00 -4.60935831e-02
6.13774836e-01 -8.45668614e-02 4.98534054e-01 -1.65271431e-01
-2.73673981e-01 -2.27014482e-01 -1.04732287e+00 6.04083121e-01
1.22912824e+00 -5.45741737e-01 -9.02487993e-01 2.59144396e-01
4.70928729e-01 -1.27391517e-01 -8.82492065e-01 6.25853121e-01
6.81038260e-01 -1.09701777e+00 1.14680862e+00 -3.43446761e-01
1.58259422e-01 -4.15080279e-01 -3.57087046e-01 -1.35302138e+00
-4.40860689e-01 -2.52845734e-01 -4.25385721e-02 1.36206865e+00
-3.17789875e-02 -4.49808091e-01 2.23861784e-01 1.19021691e-01
-6.34231493e-02 -6.88175321e-01 -6.06938601e-01 -8.60670567e-01
-8.93725157e-02 -2.95050174e-01 4.74853992e-01 9.91142929e-01
-2.91921020e-01 1.17527686e-01 -4.39214677e-01 2.61509735e-02
7.39704072e-01 3.57594132e-01 6.48651063e-01 -1.21661294e+00
7.40114525e-02 -7.88772583e-01 -4.80778843e-01 -1.25472879e+00
-3.57871987e-02 -9.26451504e-01 -2.70565629e-01 -1.54407990e+00
3.92969370e-01 -3.36910367e-01 -8.70269537e-01 -3.41420770e-02
-2.42622226e-01 5.09619892e-01 3.36349696e-01 7.59503245e-01
-8.48834813e-01 8.81001294e-01 1.67119193e+00 -4.42917943e-01
6.55109435e-02 -3.68188396e-02 -5.66712618e-01 6.96780503e-01
4.57466424e-01 -3.02988708e-01 -4.59926724e-01 -5.78874171e-01
-1.54709946e-02 -2.18345270e-01 1.82424873e-01 -9.28690672e-01
7.18510270e-01 5.93915917e-02 2.72405028e-01 -9.20875251e-01
4.62540984e-01 -9.31826651e-01 -4.75724824e-02 2.57614732e-01
-3.42270583e-01 2.42602229e-01 -5.97044118e-02 7.08741307e-01
-5.97607195e-01 -2.88617253e-01 5.33594072e-01 -6.81553269e-03
-9.59193587e-01 1.99216262e-01 -3.93279195e-02 -2.47692376e-01
8.85645866e-01 -3.89150769e-01 -1.96406960e-01 -3.47031921e-01
-7.19017029e-01 3.20123017e-01 3.88749689e-01 7.77542710e-01
8.72250021e-01 -1.50971520e+00 -6.87678874e-01 2.14338422e-01
3.87077153e-01 -3.88391763e-01 7.03332663e-01 9.26071882e-01
-2.10525990e-02 4.34558094e-01 -2.01822534e-01 -7.07388043e-01
-1.59304225e+00 9.18550014e-01 -2.91038100e-02 -1.56461269e-01
-5.31708479e-01 7.51930654e-01 6.14061058e-02 -6.38684928e-02
2.52243370e-01 9.28907748e-03 -5.38949072e-01 5.24807930e-01
5.23402631e-01 4.44771886e-01 1.74970821e-01 -8.36741805e-01
-3.21299672e-01 1.14003348e+00 -3.81516427e-01 2.20229160e-02
1.28507900e+00 -4.46471751e-01 -1.91872463e-01 4.20871496e-01
1.23011506e+00 3.99871171e-02 -7.32186913e-01 -6.96404159e-01
-2.77120415e-02 -5.60941160e-01 1.55064166e-01 -6.66549683e-01
-1.20602036e+00 1.00011790e+00 9.10998702e-01 3.73348743e-01
1.33043373e+00 -2.32595861e-01 8.22898507e-01 5.40704548e-01
3.87297928e-01 -1.01469433e+00 6.95515648e-02 2.69175410e-01
1.04322445e+00 -1.44576597e+00 3.41285706e-01 -1.53974831e-01
-6.47779524e-01 9.56742465e-01 2.54510909e-01 -3.88166338e-01
8.04258704e-01 -1.35236517e-01 2.76131765e-03 -2.68127441e-01
-7.41844952e-01 -5.93147159e-01 7.68120468e-01 4.72148895e-01
2.18483597e-01 -2.39176601e-01 -8.91273856e-01 1.44524917e-01
3.78202558e-01 -1.84477061e-01 4.13535647e-02 9.95829642e-01
-7.55484402e-01 -1.16730976e+00 -5.22927344e-01 3.23474169e-01
-2.03852355e-01 -2.31937796e-01 -6.04896247e-01 8.01614761e-01
2.55824383e-02 9.41062987e-01 -1.12978831e-01 -4.64588165e-01
3.45063388e-01 -6.59487545e-02 4.75272566e-01 -1.29528955e-01
-5.99387884e-01 2.49577388e-01 -3.71770054e-01 -2.78781146e-01
-1.06210816e+00 -6.20318532e-01 -7.22439349e-01 -2.04654232e-01
-5.14842033e-01 2.02428356e-01 4.66012001e-01 8.00365269e-01
3.09047818e-01 2.93722659e-01 6.60101175e-01 -7.04798281e-01
-6.45473182e-01 -1.14595115e+00 -6.87218666e-01 5.24332941e-01
8.46542343e-02 -9.57300603e-01 -5.00940025e-01 -3.18720073e-01] | [10.994397163391113, 0.9314542412757874] |
d35d7bbb-90ff-486c-b7df-96ec0d5af7cc | sunet-scale-aware-unified-network-for | 2209.02877 | null | https://arxiv.org/abs/2209.02877v1 | https://arxiv.org/pdf/2209.02877v1.pdf | SUNet: Scale-aware Unified Network for Panoptic Segmentation | Panoptic segmentation combines the advantages of semantic and instance segmentation, which can provide both pixel-level and instance-level environmental perception information for intelligent vehicles. However, it is challenged with segmenting objects of various scales, especially on extremely large and small ones. In this work, we propose two lightweight modules to mitigate this problem. First, Pixel-relation Block is designed to model global context information for large-scale things, which is based on a query-independent formulation and brings small parameter increments. Then, Convectional Network is constructed to collect extra high-resolution information for small-scale stuff, supplying more appropriate semantic features for the downstream segmentation branches. Based on these two modules, we present an end-to-end Scale-aware Unified Network (SUNet), which is more adaptable to multi-scale objects. Extensive experiments on Cityscapes and COCO demonstrate the effectiveness of the proposed methods. | ['Ming Yang', 'Chunxiang Wang', 'Yeqiang Qian', 'Weihao Yan'] | 2022-09-07 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 3.79493237e-02 -2.12774351e-01 -1.42393291e-01 -6.80701852e-01
-3.96558702e-01 -4.99684423e-01 2.90010393e-01 1.20452158e-02
-3.04345757e-01 4.16420192e-01 -5.68375960e-02 -2.48123016e-02
-9.13006365e-02 -1.36082637e+00 -3.25311929e-01 -6.58459663e-01
1.25959143e-01 3.30046445e-01 1.07803118e+00 -2.54532158e-01
-1.85329132e-02 6.42657816e-01 -1.56205726e+00 1.67486578e-01
1.12231612e+00 1.30687714e+00 5.58300495e-01 3.43533456e-01
-6.05224669e-01 2.38642558e-01 -2.16816261e-01 5.26017584e-02
3.78797859e-01 -1.61268830e-01 -5.92381120e-01 3.20185363e-01
4.23909128e-01 -1.46211639e-01 4.87265885e-02 1.20312881e+00
3.20214480e-01 2.56215721e-01 3.82353842e-01 -1.08637047e+00
-1.60150573e-01 3.75068098e-01 -6.87621236e-01 3.35954040e-01
-3.59853476e-01 3.33593398e-01 1.01304328e+00 -4.06602144e-01
5.34754515e-01 1.54444206e+00 6.34667039e-01 4.22212481e-02
-8.62675011e-01 -6.90147996e-01 4.65612590e-01 4.17691506e-02
-1.21064401e+00 -1.50983423e-01 8.46439779e-01 -6.24099486e-02
4.71293896e-01 2.52009511e-01 7.68529773e-01 3.86454314e-01
-1.09111726e-01 7.59410977e-01 1.17520535e+00 2.28114665e-01
2.89791673e-01 1.61162525e-01 5.47867119e-02 3.88510048e-01
2.64232397e-01 -2.60877937e-01 -4.60799746e-02 3.87905419e-01
8.02366257e-01 1.62163034e-01 7.34758750e-02 -2.95223534e-01
-1.12453115e+00 6.39891088e-01 1.05433667e+00 4.15280938e-01
-4.57472563e-01 8.71045515e-02 2.22730115e-01 -1.50994271e-01
7.22041786e-01 8.01005363e-02 -7.05961823e-01 3.18969101e-01
-1.11419821e+00 -1.04956470e-01 5.39827049e-01 9.45898950e-01
1.15179622e+00 -2.40898933e-02 -1.36106536e-01 9.58746314e-01
4.43572044e-01 7.95245945e-01 8.82379264e-02 -9.50129807e-01
4.51331466e-01 9.07650113e-01 3.94859947e-02 -9.40689445e-01
-8.60724211e-01 -6.37894690e-01 -8.07562888e-01 1.30356550e-01
2.35890485e-02 -1.72073632e-01 -1.25078571e+00 1.57704282e+00
9.31051910e-01 3.47180516e-01 8.69165361e-02 1.20191777e+00
1.11476743e+00 8.55472207e-01 3.87422949e-01 -7.96876326e-02
1.63803744e+00 -1.24476850e+00 -3.97764385e-01 -4.25691128e-01
2.44879887e-01 -6.51436269e-01 9.87155914e-01 -1.62697613e-01
-6.57922089e-01 -7.91056812e-01 -8.72123778e-01 -1.28400460e-01
-7.36580193e-01 -7.99747184e-02 7.21483648e-01 5.45588672e-01
-8.67428064e-01 1.94887936e-01 -4.47034687e-01 -6.78907514e-01
6.00269973e-01 2.31416686e-03 2.27372810e-01 6.85736388e-02
-1.30157971e+00 4.10381258e-01 6.62005067e-01 3.71987879e-01
-5.78511894e-01 -5.90226114e-01 -7.78389931e-01 2.20557675e-01
6.64075136e-01 -6.54764116e-01 8.56824458e-01 -1.00920546e+00
-1.39015269e+00 5.91495275e-01 -8.56751800e-02 -2.50759929e-01
4.95326608e-01 8.63759667e-02 -5.98602831e-01 5.06434202e-01
4.37252015e-01 1.13384461e+00 6.07603729e-01 -1.27476037e+00
-1.15180075e+00 -5.22106946e-01 7.86738321e-02 6.17437065e-01
-9.90240425e-02 -9.92956385e-02 -1.02878428e+00 -4.34337169e-01
4.13591117e-01 -6.54185474e-01 -6.19079769e-01 1.95885636e-02
-3.49478841e-01 -1.96276471e-01 9.18040276e-01 -4.76033151e-01
9.68720794e-01 -2.15946770e+00 -2.36685514e-01 3.75487119e-01
1.03701368e-01 2.11611554e-01 -2.43211165e-01 -6.48750439e-02
3.65538806e-01 1.25080168e-01 -4.73068953e-01 2.48141736e-02
-1.23059392e-01 2.01488927e-01 1.89210057e-01 1.35622874e-01
1.21268980e-01 9.83977795e-01 -8.00523162e-01 -9.01322007e-01
4.56783861e-01 1.87739968e-01 -2.87577838e-01 1.81334801e-02
-6.32465899e-01 4.78299767e-01 -1.11849713e+00 9.21237350e-01
1.22642922e+00 -1.49966016e-01 -1.15959294e-01 -4.28356349e-01
-5.39400399e-01 -1.82084799e-01 -1.43752873e+00 1.57741630e+00
-4.97923523e-01 1.79590538e-01 5.21102428e-01 -7.35019505e-01
8.85939658e-01 -2.43812665e-01 4.51315790e-01 -9.20871198e-01
2.39519313e-01 1.70267344e-01 -3.79523873e-01 -5.40065527e-01
4.57753778e-01 1.77422836e-02 -1.20111518e-01 -1.64145723e-01
-4.04122591e-01 -4.33994591e-01 2.98696309e-01 3.11300661e-02
4.72419173e-01 1.54720739e-01 -7.47416541e-02 -4.71554577e-01
7.47826397e-01 1.80050686e-01 8.14797223e-01 4.90484864e-01
-3.44034731e-01 4.96351540e-01 2.24485472e-01 -3.22576493e-01
-8.32942307e-01 -1.01342213e+00 -3.58047903e-01 8.93528521e-01
9.61278498e-01 -1.16261259e-01 -8.63233387e-01 -5.97499371e-01
2.48953048e-03 4.95273858e-01 -3.91271025e-01 3.64427239e-01
-4.13572311e-01 -9.02488351e-01 1.15076527e-01 5.66129625e-01
1.27212703e+00 -1.02507675e+00 -5.41362345e-01 3.66371930e-01
-3.43061805e-01 -1.30849528e+00 -4.31460291e-01 -1.37776047e-01
-8.63367856e-01 -8.05967450e-01 -5.04495323e-01 -7.92300344e-01
4.71321404e-01 7.32529283e-01 9.09389496e-01 -9.48393941e-02
-1.63177237e-01 -5.88531904e-02 -2.00995743e-01 -2.39128932e-01
-6.81038667e-03 2.95278609e-01 -4.36478019e-01 3.72291684e-01
1.20175943e-01 -6.09804213e-01 -9.18767631e-01 6.47199750e-01
-1.10929644e+00 3.15926999e-01 7.65884817e-01 2.54134357e-01
9.80394006e-01 3.60726982e-01 6.16300464e-01 -8.18061948e-01
9.42027047e-02 -5.49318552e-01 -8.84090960e-01 3.03796172e-01
-4.70819503e-01 -2.25284934e-01 5.72843075e-01 -1.96682602e-01
-1.41066444e+00 6.19741790e-02 -2.92397857e-01 -9.28695872e-02
-3.84582758e-01 2.35837162e-01 -6.50884986e-01 1.53297549e-02
3.08078498e-01 9.79503319e-02 -3.55539262e-01 -6.47221029e-01
7.32907832e-01 8.67066741e-01 4.48890805e-01 -4.35620576e-01
7.85230875e-01 8.35615098e-01 6.95454553e-02 -9.16636109e-01
-9.76260722e-01 -8.64327610e-01 -7.28121281e-01 -3.19302708e-01
1.23918295e+00 -1.11130154e+00 -3.21363240e-01 5.57905793e-01
-7.27481306e-01 -2.78984606e-01 -3.81928980e-01 3.36824685e-01
-3.35713446e-01 9.25253406e-02 -3.65692496e-01 -5.29400706e-01
-4.13708448e-01 -1.07824326e+00 1.23121870e+00 9.25689518e-01
5.22354603e-01 -7.79986560e-01 -2.77330995e-01 4.95781809e-01
5.21478117e-01 2.32055113e-01 5.11521339e-01 -3.53609234e-01
-1.03871620e+00 1.72874197e-01 -9.45247531e-01 3.32270205e-01
6.99597448e-02 -7.73537457e-02 -9.00412738e-01 1.09404467e-01
-1.10234916e-01 -1.02290079e-01 1.12995231e+00 4.91602391e-01
1.13970423e+00 1.59629788e-02 -4.55742925e-01 8.24674726e-01
1.64948750e+00 1.86491720e-02 4.69584405e-01 3.04424912e-01
8.42195392e-01 7.10401773e-01 9.22267258e-01 1.27944797e-01
6.18132055e-01 4.98198211e-01 6.21602952e-01 -5.40245771e-01
-3.52761269e-01 -2.86517609e-02 -8.24609101e-02 4.97644544e-01
1.15986384e-01 -2.31966048e-01 -5.42551219e-01 7.17190444e-01
-1.84947181e+00 -7.33429432e-01 -3.96053553e-01 1.72348738e+00
5.72451949e-01 2.73934513e-01 1.23623937e-01 -5.21584034e-01
8.29793096e-01 4.84743029e-01 -9.23481047e-01 -2.40163822e-02
-3.40331405e-01 -1.53995186e-01 7.56972611e-01 3.55966181e-01
-1.27159238e+00 1.41489518e+00 5.46333551e+00 1.33775187e+00
-1.14237607e+00 3.12483191e-01 8.67089331e-01 1.48000345e-01
-3.61039609e-01 9.37356353e-02 -8.75419796e-01 5.54801881e-01
5.92661023e-01 2.48124689e-01 2.83403903e-01 8.79732847e-01
4.32595581e-01 -4.54947114e-01 -2.93460578e-01 6.25643730e-01
-3.48389298e-01 -1.26225805e+00 -1.35215431e-01 -1.65629685e-01
9.89478171e-01 5.87736189e-01 -2.61342138e-01 1.24207824e-01
2.34068006e-01 -5.03482938e-01 6.80742681e-01 3.13328296e-01
5.61264455e-01 -6.36559665e-01 6.94316983e-01 3.32203627e-01
-1.72960877e+00 -6.45384640e-02 -4.96650517e-01 3.62687647e-01
1.99075744e-01 9.36710238e-01 -4.20290560e-01 7.98403263e-01
8.66554558e-01 7.02837467e-01 -6.78022563e-01 1.14109385e+00
-1.72283396e-01 6.01363957e-01 -8.43970060e-01 6.50924817e-02
6.62069380e-01 -4.98542011e-01 4.70180064e-01 1.44694853e+00
2.27959499e-01 2.62942195e-01 5.33090532e-01 8.87700260e-01
6.20560050e-02 3.06581318e-01 -2.56585002e-01 5.07052600e-01
4.53879803e-01 1.99943578e+00 -1.49458563e+00 -5.86040258e-01
-3.76848370e-01 8.37722182e-01 -2.76867766e-03 4.97094482e-01
-9.03133810e-01 -3.45104188e-01 6.10843778e-01 4.58937585e-02
6.18411481e-01 -1.75443724e-01 -4.45179254e-01 -1.08425558e+00
-2.64352728e-02 -2.95548886e-01 3.65030020e-01 -9.05368805e-01
-9.80110765e-01 4.14607018e-01 -4.89694299e-03 -1.23388839e+00
4.72588986e-01 -2.57902801e-01 -7.09999859e-01 6.91666484e-01
-2.05158997e+00 -1.53278840e+00 -6.34952784e-01 4.26503688e-01
7.23578036e-01 3.52385938e-01 3.39916423e-02 3.22420686e-01
-6.23629749e-01 -1.42480612e-01 2.63551444e-01 -8.84164050e-02
3.40378553e-01 -1.06158495e+00 2.05678120e-01 1.00098181e+00
-1.69508532e-01 1.12590089e-01 3.54298681e-01 -7.07396984e-01
-7.80224085e-01 -1.64582598e+00 5.20258605e-01 -1.43213347e-01
5.58053493e-01 -1.76024809e-01 -8.03812742e-01 1.23235106e-01
-9.20298174e-02 1.79366440e-01 1.43317580e-01 -1.29781410e-01
-9.36111584e-02 -7.68359482e-01 -1.31823254e+00 4.60067660e-01
1.18976688e+00 -2.13392839e-01 -2.39399016e-01 4.11463022e-01
1.17864692e+00 -3.36107075e-01 -7.80082405e-01 6.50924742e-01
2.41279885e-01 -1.07439375e+00 1.06053436e+00 -7.84929097e-02
2.25899071e-01 -7.41810620e-01 -9.20389295e-02 -1.13451922e+00
-4.89226133e-01 -2.03480244e-01 4.40117419e-01 1.49030542e+00
3.19780469e-01 -6.49401128e-01 6.14346921e-01 1.27949849e-01
-3.06012958e-01 -5.37285268e-01 -9.15065765e-01 -8.14127684e-01
-1.81903362e-01 -4.82084394e-01 1.08932865e+00 7.57270217e-01
-7.80182898e-01 5.16379893e-01 5.30487150e-02 5.11235774e-01
5.97586334e-01 8.39722276e-01 6.04665041e-01 -1.32510412e+00
1.89837560e-01 -5.98751247e-01 -3.35728049e-01 -1.27401316e+00
-1.71107650e-01 -7.43233323e-01 1.94567353e-01 -1.87536120e+00
1.16287850e-01 -7.07317352e-01 -4.85874027e-01 2.20470428e-01
-2.96331525e-01 4.03439283e-01 1.63416907e-01 1.99621707e-01
-9.62579489e-01 5.93130112e-01 1.66368937e+00 -2.06276312e-01
-3.29415470e-01 1.48030728e-01 -6.43814564e-01 8.12503278e-01
1.05549872e+00 -1.97210893e-01 -4.84228790e-01 -3.23832124e-01
2.22442159e-03 -1.78875238e-01 3.09564739e-01 -1.16998088e+00
5.98866642e-02 -6.44735336e-01 2.83942997e-01 -1.03967941e+00
1.17211565e-01 -9.16432559e-01 2.39091981e-02 1.25679880e-01
1.42132966e-02 -4.92750287e-01 1.85422674e-01 5.94155431e-01
-2.20231459e-01 3.85463722e-02 1.05107450e+00 -1.93473771e-01
-1.16301608e+00 6.68347716e-01 1.99161563e-02 1.31197184e-01
1.15284586e+00 -2.28413641e-01 -1.43409535e-01 1.57400705e-02
-5.30843198e-01 9.12219226e-01 4.07222271e-01 4.24870133e-01
2.40705028e-01 -1.13642573e+00 -5.40378809e-01 1.93658903e-01
2.97314584e-01 6.22463584e-01 5.71450233e-01 7.84911752e-01
-6.94994152e-01 3.13439578e-01 -8.29977393e-02 -7.53925741e-01
-1.05320168e+00 5.17984152e-01 3.42035353e-01 -1.28261387e-01
-5.98905742e-01 6.38484538e-01 6.02142930e-01 -5.64357340e-01
-1.70815423e-01 -7.43465960e-01 -3.78282070e-01 3.93992156e-01
2.90293813e-01 1.85937703e-01 -1.87989131e-01 -8.10754538e-01
-2.40722701e-01 8.62035215e-01 3.14886272e-01 6.41174540e-02
1.23121393e+00 -7.29256392e-01 -3.04696321e-01 2.70926863e-01
9.76079524e-01 -2.36771926e-02 -1.61840963e+00 -5.26925445e-01
-2.98935860e-01 -2.98518151e-01 3.16782415e-01 -7.79132724e-01
-1.43722785e+00 8.38827789e-01 6.61622465e-01 3.72914135e-01
1.37219036e+00 1.44295990e-01 1.18798029e+00 2.25339159e-01
2.23429814e-01 -1.39425516e+00 -3.48792404e-01 3.61198962e-01
4.01898474e-01 -1.32965565e+00 1.03439856e-02 -7.89333105e-01
-4.66326833e-01 9.75579858e-01 8.85422587e-01 2.66285628e-01
7.36452281e-01 -3.40706185e-02 1.03691854e-01 -2.97078222e-01
-1.69571087e-01 -7.62199759e-01 2.44083509e-01 4.61388350e-01
-3.20339561e-01 3.41486037e-01 -1.84857711e-01 4.69911724e-01
1.88073978e-01 1.37218088e-02 -8.45232699e-03 5.05887389e-01
-1.08713901e+00 -6.91877902e-01 -4.52094525e-01 3.75657111e-01
3.50332819e-02 -2.46211022e-01 -4.92661772e-03 6.00578308e-01
7.57350504e-01 9.19434309e-01 2.49449015e-01 -1.70932189e-01
2.15641990e-01 -3.17211211e-01 -2.91320961e-02 -3.52047890e-01
-3.35044891e-01 3.63193810e-01 1.06495835e-01 -7.19085217e-01
-7.44933724e-01 -6.00427091e-01 -1.49939752e+00 -3.08832321e-02
-2.17698902e-01 -4.16193791e-02 7.20993102e-01 8.41509342e-01
4.32861894e-01 5.78136384e-01 8.63264024e-01 -9.30436254e-01
-3.62366140e-02 -8.60172093e-01 -6.67184293e-01 3.63029510e-01
1.57423407e-01 -3.52250725e-01 -1.12263694e-01 -6.10809103e-02] | [9.511253356933594, -0.35330837965011597] |
533e0ac6-29c9-4b08-bcc3-041cc07f09aa | parameter-efficient-fine-tuning-with-layer | 2305.08285 | null | https://arxiv.org/abs/2305.08285v3 | https://arxiv.org/pdf/2305.08285v3.pdf | Parameter-Efficient Fine-Tuning with Layer Pruning on Free-Text Sequence-to-Sequence Modeling | The increasing size of language models raises great research interests in parameter-efficient fine-tuning such as LoRA that freezes the pre-trained model, and injects small-scale trainable parameters for multiple downstream tasks (e.g., summarization, question answering and translation). To further enhance the efficiency of fine-tuning, we propose a framework that integrates LoRA and structured layer pruning. The integrated framework is validated on two created deidentified medical report summarization datasets based on MIMIC-IV-Note and two public medical dialogue datasets. By tuning 0.6% parameters of the original model and pruning over 30% Transformer-layers, our framework can reduce 50% of GPU memory usage and speed up 100% of the training phase, while preserving over 92% generation qualities on free-text sequence-to-sequence tasks. | ['Wensheng Zhang', 'Yuanyuan Wu', 'Xuebing Yang', 'Yunqi Zhu'] | 2023-05-15 | null | null | null | null | ['dialogue-generation', 'dialogue-generation'] | ['natural-language-processing', 'speech'] | [ 4.56885844e-01 5.57607174e-01 -2.64761806e-01 -4.66262281e-01
-1.39352727e+00 -6.82269990e-01 2.82207251e-01 5.85058570e-01
-6.26488566e-01 9.48609948e-01 7.10875630e-01 -6.00947499e-01
1.98667213e-01 -6.12129629e-01 -7.48258173e-01 -2.42751911e-01
1.50044233e-01 6.12018168e-01 1.10970937e-01 -7.60608315e-02
8.19263011e-02 7.21287131e-02 -8.35324466e-01 8.96719694e-01
1.04282331e+00 5.42732358e-01 5.47343530e-02 9.70836163e-01
-1.51224613e-01 7.43469059e-01 -9.48163807e-01 -6.23323858e-01
-2.81960964e-01 -3.80678952e-01 -1.14807463e+00 -1.74039960e-01
3.66021961e-01 -4.23945785e-01 -4.45565432e-02 5.79612911e-01
1.00961685e+00 -4.88752238e-02 4.04145539e-01 -2.63961643e-01
-2.78795213e-01 1.00224066e+00 -3.79811138e-01 2.34763354e-01
1.33003235e-01 2.88510054e-01 1.02425039e+00 -4.33630198e-01
6.34135783e-01 1.16972983e+00 8.35249841e-01 8.16280961e-01
-1.29827034e+00 -4.49816704e-01 -2.06580549e-01 -2.06144765e-01
-1.01974690e+00 -6.40476525e-01 1.28029827e-02 -1.84328914e-01
1.62647188e+00 5.08423209e-01 2.97719210e-01 1.35129094e+00
4.69188154e-01 7.00955212e-01 7.20669806e-01 -1.20830022e-01
2.44544938e-01 -6.93953261e-02 2.50416726e-01 8.80815208e-01
3.41217786e-01 -4.33300793e-01 -4.59214926e-01 -5.80774128e-01
2.50825644e-01 -5.09473205e-01 -1.24002658e-01 3.59196484e-01
-1.18159032e+00 9.02138412e-01 -2.08452176e-02 5.16458750e-02
-2.27967501e-01 1.13298856e-01 9.64582920e-01 -2.36333683e-02
6.59165323e-01 8.44856560e-01 -6.54119253e-01 -4.72065657e-01
-1.23529232e+00 1.57240123e-01 9.62934136e-01 8.69997263e-01
4.32504326e-01 -1.08991913e-01 -9.59562242e-01 1.03572083e+00
-1.42695546e-01 3.29436123e-01 6.29361868e-01 -8.64577532e-01
8.57350230e-01 6.25220776e-01 -2.79595643e-01 -3.87025625e-01
-8.03153157e-01 -7.08572507e-01 -9.82677221e-01 -7.15369821e-01
1.12411737e-01 -5.83512604e-01 -8.86745751e-01 1.73739052e+00
3.53631645e-01 3.83435860e-02 1.28174812e-01 4.69027489e-01
1.01269627e+00 6.72737002e-01 3.20873782e-02 -3.39626431e-01
1.63522375e+00 -1.20831132e+00 -6.98126376e-01 -3.77557904e-01
1.07870352e+00 -7.43359864e-01 1.16201627e+00 2.34686911e-01
-1.39592302e+00 -2.78770775e-01 -9.94048417e-01 -5.46618819e-01
1.21532425e-01 3.83731991e-01 5.84432602e-01 5.73062062e-01
-1.08089685e+00 6.91192508e-01 -1.04858780e+00 -3.10420007e-01
4.80895460e-01 4.94226873e-01 -1.77635506e-01 1.18477739e-01
-1.10411429e+00 7.37730145e-01 5.85940301e-01 -4.61651720e-02
-8.27850461e-01 -1.21836782e+00 -6.95732772e-01 3.66345584e-01
1.82487756e-01 -1.44549167e+00 1.45889568e+00 -1.48138434e-01
-1.62980306e+00 8.43046546e-01 -2.26414338e-01 -8.94148886e-01
6.65898204e-01 -3.95968437e-01 -9.99136344e-02 1.21406049e-01
-7.08101988e-02 5.77360570e-01 5.09454787e-01 -4.10993427e-01
-3.44445050e-01 -1.02332287e-01 -6.04283288e-02 1.11824080e-01
-2.98348784e-01 8.35364964e-03 -5.61539352e-01 -6.39080465e-01
-4.86254424e-01 -9.15551782e-01 -4.36874628e-01 -6.80969596e-01
-7.34901309e-01 2.34005898e-02 2.23101690e-01 -1.05714869e+00
1.52386367e+00 -1.86000085e+00 -3.83877009e-02 -1.96129024e-01
3.44012320e-01 5.47100186e-01 -3.81604820e-01 5.01452863e-01
5.81056997e-02 1.88791096e-01 -5.63494802e-01 -5.36291599e-01
-1.93768933e-01 9.82494354e-02 -3.04587007e-01 1.43737346e-01
2.30870038e-01 1.22090721e+00 -8.30312908e-01 -5.33412457e-01
-1.37105361e-01 3.58171403e-01 -1.15567708e+00 3.08282793e-01
-4.54541206e-01 3.97696733e-01 -5.24620533e-01 2.35620290e-01
5.23937106e-01 -4.99367654e-01 4.09767151e-01 -9.00635868e-02
1.84193850e-01 8.73402953e-01 -5.54519117e-01 2.11069584e+00
-5.63438654e-01 3.86930227e-01 -1.14827551e-01 -6.28093064e-01
6.78154647e-01 2.42481768e-01 2.45719299e-01 -7.10742116e-01
1.09926291e-01 6.98320717e-02 -2.05602068e-02 -6.54363632e-01
8.41656327e-01 4.91990857e-02 -4.34094429e-01 6.82033241e-01
1.25286937e-01 -4.15852480e-02 3.15246165e-01 4.03444618e-01
1.43809617e+00 -7.54431561e-02 5.10258377e-01 -3.32538307e-01
6.33584976e-01 -2.35537197e-02 5.07508337e-01 8.82584810e-01
2.56208807e-01 5.37689328e-01 7.50816226e-01 -1.19877256e-01
-1.15537846e+00 -6.78885639e-01 -1.28810763e-01 1.35131168e+00
-7.03360677e-01 -9.07797754e-01 -1.29093146e+00 -8.22451770e-01
-1.51823536e-01 9.61025655e-01 -5.70330620e-01 -4.09597993e-01
-8.93685043e-01 -1.27616036e+00 1.34862506e+00 2.94233203e-01
3.25132102e-01 -9.04897332e-01 -4.23748434e-01 5.56000412e-01
-5.63886404e-01 -1.17659032e+00 -7.07318127e-01 3.10907681e-02
-1.12989783e+00 -7.94626236e-01 -5.71287751e-01 -3.31440419e-01
4.16835546e-01 -2.84087539e-01 1.43403685e+00 -3.18938754e-02
-3.97120923e-01 -7.13763386e-02 -1.78208962e-01 -3.22101444e-01
-9.98854339e-01 9.83178198e-01 -3.29135776e-01 -2.87739724e-01
2.07774676e-02 -4.61394548e-01 -6.04741514e-01 -2.01664999e-01
-7.66703784e-01 3.81952852e-01 7.28575945e-01 9.66747761e-01
5.43593705e-01 -6.34954631e-01 7.91920125e-01 -1.29859149e+00
1.03863287e+00 -3.73163104e-01 -3.31509620e-01 5.18799126e-01
-4.76479113e-01 5.72390556e-01 6.69592738e-01 -2.71047264e-01
-1.16466653e+00 -3.34174633e-01 -4.95917559e-01 1.25004739e-01
1.70252770e-01 3.74811679e-01 4.22233194e-02 5.72664261e-01
8.58520210e-01 3.10492784e-01 -1.07278161e-01 -6.02661014e-01
5.79958558e-01 7.34651566e-01 6.27639949e-01 -5.32164156e-01
2.41284311e-01 1.66802242e-01 -3.09505522e-01 -5.76057255e-01
-1.17095411e+00 -2.16884926e-01 -2.55391687e-01 5.84601521e-01
9.17471111e-01 -9.45671260e-01 -9.57680047e-01 1.46051720e-01
-1.29642761e+00 -4.76122200e-01 -3.35102409e-01 2.37098739e-01
-4.13949221e-01 4.23517525e-01 -9.75052416e-01 -3.19182634e-01
-1.31808817e+00 -1.10265374e+00 1.35087538e+00 -8.25800970e-02
-6.70930028e-01 -7.67544210e-01 3.68653357e-01 7.91398883e-01
3.36364508e-01 -8.92181993e-02 1.25832140e+00 -7.91304111e-01
-2.88079023e-01 1.79036483e-01 -1.43682003e-01 3.17826450e-01
-1.39120772e-01 -3.44025403e-01 -1.04956460e+00 -4.69926596e-01
-4.74759229e-02 -4.23014373e-01 1.14331222e+00 3.77415299e-01
1.28482533e+00 -6.92694545e-01 -3.96836668e-01 8.60272586e-01
9.05792773e-01 -1.70062795e-01 4.60690945e-01 8.69337916e-02
5.88331282e-01 3.41143191e-01 4.01681870e-01 6.34973407e-01
3.63839567e-01 4.59470689e-01 -1.89447969e-01 -1.15529168e-02
-2.48470485e-01 -3.84140879e-01 3.35723341e-01 1.17697942e+00
2.67661482e-01 -3.71844858e-01 -7.39600182e-01 4.26629424e-01
-1.77891409e+00 -7.15279639e-01 2.57649105e-02 1.91918588e+00
1.45526099e+00 3.42478491e-02 -7.06942677e-02 -3.92532945e-01
4.33728576e-01 1.96134076e-01 -7.92301774e-01 -8.42269719e-01
-6.51601329e-03 4.34347838e-01 5.32871366e-01 5.89350522e-01
-9.26538289e-01 1.18001831e+00 6.63185024e+00 1.04798245e+00
-1.08505607e+00 4.13298666e-01 8.97057116e-01 -5.53896546e-01
-3.20279777e-01 -2.36290872e-01 -1.01019502e+00 3.74175102e-01
1.54537141e+00 -1.74658373e-01 2.52859265e-01 6.03487194e-01
5.22842944e-01 9.81098488e-02 -1.21417701e+00 6.78421140e-01
8.01270530e-02 -1.81987202e+00 2.35659882e-01 5.18176518e-02
6.46932542e-01 3.54796052e-01 -1.92936122e-01 6.41097486e-01
3.86651248e-01 -1.08683205e+00 3.89087319e-01 3.20433885e-01
1.09655130e+00 -7.85749018e-01 7.55580068e-01 5.07811129e-01
-4.66656953e-01 8.02670270e-02 -3.63622814e-01 2.68203139e-01
3.91380727e-01 7.96711981e-01 -1.55270064e+00 6.18936360e-01
3.58256429e-01 3.95998567e-01 -7.29330242e-01 6.10838592e-01
-1.59468994e-01 1.05406332e+00 -2.19905227e-01 6.02437295e-02
2.69033849e-01 -4.83696262e-04 3.82398456e-01 1.90346575e+00
1.11219049e-01 2.59072836e-02 -1.54889271e-01 7.55196512e-01
-4.32843834e-01 1.86251298e-01 -6.02728166e-02 -2.85075195e-02
3.70573074e-01 1.30486810e+00 -3.10193509e-01 -5.33393919e-01
-2.39156978e-03 1.03206599e+00 6.00380421e-01 5.61189950e-02
-1.07297230e+00 -3.08692724e-01 5.06331682e-01 -5.71963415e-02
1.80842534e-01 4.30441871e-02 -5.33756316e-01 -1.14333570e+00
-3.47334966e-02 -1.29333341e+00 5.97646654e-01 -3.15922558e-01
-8.38032901e-01 7.79646754e-01 -3.03575277e-01 -6.63688779e-01
-6.03249252e-01 -2.08337568e-02 -3.74429137e-01 8.13234031e-01
-1.25321937e+00 -1.06731653e+00 -1.65260900e-02 3.08833331e-01
5.51053703e-01 -3.69826816e-02 9.36460853e-01 3.71266961e-01
-8.09348166e-01 1.11485565e+00 2.24657372e-01 1.87615007e-02
8.72472584e-01 -1.04017675e+00 8.22613180e-01 5.31856537e-01
-2.65554279e-01 9.47212696e-01 6.14342391e-01 -6.30883157e-01
-1.15155745e+00 -1.45481706e+00 1.21578264e+00 -4.52591300e-01
5.13944626e-01 -7.01174557e-01 -9.49624658e-01 5.96378028e-01
1.86980188e-01 -5.56192756e-01 8.36413205e-01 2.03255638e-01
-1.81616902e-01 -3.24421562e-02 -1.13842416e+00 5.96525848e-01
1.05332649e+00 -4.94128764e-01 -7.65520096e-01 5.97478092e-01
1.11115742e+00 -7.47025430e-01 -1.04529488e+00 3.29848915e-01
3.55180681e-01 -7.24356413e-01 8.14406276e-01 -8.44785988e-01
5.79823971e-01 2.13016599e-01 2.36577243e-01 -1.22351360e+00
-2.70125002e-01 -1.07704723e+00 -3.09514761e-01 1.32056916e+00
7.64820158e-01 -5.64157128e-01 7.12436736e-01 2.86288768e-01
-4.61801827e-01 -1.05836868e+00 -8.28916490e-01 -3.42050433e-01
2.16860563e-01 -2.32119530e-01 6.24152541e-01 5.86814106e-01
6.80160895e-02 9.47528064e-01 -2.40569174e-01 -1.09146141e-01
2.39123181e-01 -7.87490830e-02 6.73478782e-01 -7.18868732e-01
-6.67396009e-01 -3.59142721e-01 3.28715146e-01 -9.97882426e-01
1.93798780e-01 -1.20756567e+00 -1.65262252e-01 -1.62636864e+00
5.81977963e-01 -1.86388940e-01 4.00167191e-03 7.50267744e-01
-4.95758057e-01 7.80024973e-04 -3.88886854e-02 -3.64163518e-03
-5.99134624e-01 4.96652395e-01 1.02680552e+00 -1.72881201e-01
-2.38056719e-01 -9.64099020e-02 -1.05720031e+00 4.84797448e-01
7.39782989e-01 -6.57222867e-01 -4.00304943e-01 -6.58143520e-01
2.51825780e-01 3.62998337e-01 -1.50735423e-01 -6.43994391e-01
-2.48582847e-02 9.80704725e-02 9.69933644e-02 -5.86705923e-01
1.42881706e-01 9.43784714e-02 -1.11657180e-01 6.52627289e-01
-8.39126885e-01 4.40224679e-03 5.46263099e-01 3.31763744e-01
1.15353405e-01 -1.49969101e-01 8.09235811e-01 -1.28131405e-01
5.79548441e-02 6.63447976e-02 -3.56029838e-01 4.66094613e-01
5.55575430e-01 2.20490873e-01 -6.22420430e-01 -1.08581297e-01
-5.34158051e-01 3.39715093e-01 7.44188204e-02 4.14523572e-01
1.74644351e-01 -6.66781545e-01 -1.13360941e+00 4.06278521e-02
1.95845515e-02 1.53721645e-01 7.69297898e-01 8.75529945e-01
-7.36936212e-01 9.67637539e-01 2.30895683e-01 -4.98533219e-01
-1.38120282e+00 3.43065470e-01 3.13065499e-01 -1.15870702e+00
-8.15846145e-01 9.24128115e-01 3.66438866e-01 -7.35421658e-01
-5.93836345e-02 -5.67990065e-01 9.75257345e-03 -3.20767835e-02
7.32757628e-01 4.29459482e-01 6.13633037e-01 -2.49472782e-02
-4.47852045e-01 -4.23787609e-02 -5.65226078e-01 -8.97956043e-02
1.32856226e+00 6.68533295e-02 -2.46673048e-01 -5.93487807e-02
1.21332335e+00 2.04874605e-01 -1.04499841e+00 -1.04461133e-01
6.74327537e-02 3.02390635e-01 9.14911330e-02 -1.18187010e+00
-6.07371807e-01 8.54839265e-01 1.71474576e-01 -4.38284576e-01
1.01746607e+00 -1.27689525e-01 1.17033172e+00 7.35779405e-01
-4.88634333e-02 -1.00068378e+00 -1.58803761e-01 6.82435870e-01
7.31034398e-01 -6.68487430e-01 1.99546292e-01 -2.04523921e-01
-7.19671071e-01 7.45491922e-01 3.44286740e-01 1.65648073e-01
-1.42252624e-01 4.84706551e-01 -1.19298272e-01 1.10752601e-02
-1.27451813e+00 2.77978957e-01 3.69898528e-01 1.56780034e-01
8.29419494e-01 1.24757849e-01 -6.09867930e-01 9.01075721e-01
-6.65025055e-01 1.19650178e-01 5.67154706e-01 6.42987132e-01
-1.76873192e-01 -1.03893149e+00 -1.48275476e-02 7.63333321e-01
-9.56301332e-01 -6.10389233e-01 -1.77741006e-01 2.74875700e-01
-1.13793060e-01 9.03890073e-01 2.09151469e-02 -1.08288892e-01
3.42307150e-01 1.44861519e-01 4.86309201e-01 -8.46895278e-01
-1.29771841e+00 1.03850193e-01 6.89395547e-01 -6.71741009e-01
1.30973652e-01 -4.67570931e-01 -1.32995152e+00 -3.74980181e-01
-6.46075085e-02 2.92200208e-01 5.49794972e-01 9.01473463e-01
1.03237069e+00 9.24481630e-01 2.08696753e-01 -3.14696729e-01
-9.76073325e-01 -1.33334386e+00 -5.84914610e-02 7.19005615e-02
2.86634028e-01 9.68937576e-02 8.83923694e-02 -1.11931777e-02] | [12.058844566345215, 9.231460571289062] |
b6d4ae24-6fad-4c22-b0a4-a86e93cf4117 | pricing-european-options-with-google-automl | 2307.00476 | null | https://arxiv.org/abs/2307.00476v1 | https://arxiv.org/pdf/2307.00476v1.pdf | Pricing European Options with Google AutoML, TensorFlow, and XGBoost | Researchers have been using Neural Networks and other related machine-learning techniques to price options since the early 1990s. After three decades of improvements in machine learning techniques, computational processing power, cloud computing, and data availability, this paper is able to provide a comparison of using Google Cloud's AutoML Regressor, TensorFlow Neural Networks, and XGBoost Gradient Boosting Decision Trees for pricing European Options. All three types of models were able to outperform the Black Scholes Model in terms of mean absolute error. These results showcase the potential of using historical data from an option's underlying asset for pricing European options, especially when using machine learning algorithms that learn complex patterns that traditional parametric models do not take into account. | ['Juan Esteban Berger'] | 2023-07-02 | null | null | null | null | ['automl'] | ['methodology'] | [-7.59239376e-01 -7.10558593e-01 -4.55721110e-01 -4.96595651e-01
-1.25869408e-01 -4.48394507e-01 6.73459947e-01 -1.41698495e-01
-5.37900388e-01 8.65341246e-01 -3.09819996e-01 -1.43727624e+00
-4.01005387e-01 -1.28962672e+00 -1.10602453e-01 -1.78968415e-01
-4.30286288e-01 6.02027833e-01 -1.79963559e-01 -3.45452011e-01
6.43961489e-01 7.18580604e-01 -1.26688218e+00 1.69399410e-01
6.19463384e-01 1.62407267e+00 -4.35251206e-01 5.14401793e-01
-6.44049048e-01 9.64031160e-01 -2.35172659e-01 -7.50477910e-01
8.96084070e-01 3.67928922e-01 -2.11001843e-01 -8.17565441e-01
1.76038340e-01 -5.34447849e-01 -3.95589098e-02 8.16166699e-01
3.49615991e-01 1.23179339e-01 4.97339398e-01 -1.24480307e+00
-6.23479366e-01 5.32401383e-01 -2.77003855e-01 8.12904000e-01
-1.25549257e-01 1.63156629e-01 1.15020895e+00 -7.30647206e-01
2.34957524e-02 1.15850091e+00 1.23865485e+00 2.56102800e-01
-1.24334705e+00 -8.57617319e-01 -2.54223108e-01 2.63744801e-01
-7.35851586e-01 1.09542318e-01 8.53757679e-01 -7.04964876e-01
1.48958993e+00 4.94360626e-01 8.11897039e-01 5.34582496e-01
4.74181712e-01 5.22343576e-01 1.72533274e+00 -5.30432761e-01
6.55753732e-01 3.58839035e-01 1.81968525e-01 -7.55011961e-02
3.19556624e-01 8.74902964e-01 8.53633583e-02 -7.95417011e-01
9.32876050e-01 2.74587572e-01 4.37829852e-01 1.23460561e-01
-5.75978160e-01 1.47420442e+00 3.47870082e-01 2.88973659e-01
-7.89550781e-01 -1.18508294e-01 5.45972705e-01 7.68928587e-01
8.59820843e-01 6.07203364e-01 -1.09740412e+00 -3.52542341e-01
-1.45801091e+00 5.56761682e-01 1.17437470e+00 3.39523047e-01
4.51116145e-01 7.83324957e-01 3.61435711e-01 5.67990720e-01
2.53018320e-01 1.16370603e-01 8.14694822e-01 -9.01959419e-01
3.07171196e-01 3.19415659e-01 2.61383474e-01 -7.18737245e-01
-3.61748964e-01 -3.20528507e-01 -4.78947550e-01 1.05137682e+00
4.91839916e-01 -6.13015354e-01 -5.39646268e-01 7.93079257e-01
-4.09989320e-02 2.61783488e-02 -2.31876187e-02 6.09530032e-01
1.65395841e-01 4.21188384e-01 2.88048297e-01 -5.64555712e-02
7.92311966e-01 -7.56515622e-01 -4.15841788e-01 2.33553946e-01
4.38728988e-01 -7.95139611e-01 6.21025205e-01 6.43703818e-01
-9.55052674e-01 -5.52761614e-01 -9.05757904e-01 5.85232079e-01
-6.07310355e-01 -3.48845184e-01 1.37148607e+00 1.16119206e+00
-8.17609668e-01 1.41241622e+00 -6.45588696e-01 4.50944901e-01
4.83715832e-01 2.31101677e-01 3.72454017e-01 5.82554042e-01
-1.36686027e+00 1.24178195e+00 4.23007101e-01 -1.31388515e-01
-4.88614254e-02 -9.28344607e-01 -2.05753222e-02 9.86683965e-02
-1.59475043e-01 -4.71990138e-01 1.19411206e+00 -1.36993098e+00
-1.66680956e+00 1.66515961e-01 7.00917542e-01 -1.14873397e+00
8.20881784e-01 -1.22741163e-01 -9.36097324e-01 -2.18218103e-01
-4.21529204e-01 7.13072810e-03 6.25002325e-01 -9.26777124e-02
-9.73392129e-01 -4.13248003e-01 -1.74747989e-01 -3.52320015e-01
-3.15213315e-02 4.86247301e-01 7.98979104e-01 -1.10404110e+00
-1.50306940e-01 -6.83676124e-01 -4.19991106e-01 -2.40131319e-01
2.79923201e-01 -3.22234482e-01 5.86388648e-01 -1.05857313e+00
9.66642797e-01 -1.61627531e+00 -8.85558009e-01 5.49270749e-01
-4.27466214e-01 2.43787676e-01 5.47266006e-01 3.43349367e-01
-5.99474728e-01 5.18449128e-01 3.81310940e-01 3.03481281e-01
1.89923331e-01 9.30219218e-02 -7.01956034e-01 1.30049109e-01
2.63892356e-02 8.29939604e-01 -1.80647939e-01 1.07124083e-01
3.83267045e-01 1.31931037e-01 -3.02596897e-01 -1.50046244e-01
-3.07370424e-01 6.93204552e-02 -2.49011621e-01 6.88996434e-01
7.95551896e-01 1.16941489e-01 -2.59270906e-01 4.63593960e-01
-4.02718037e-01 4.43211526e-01 -1.16020799e+00 7.29923844e-01
-3.99227232e-01 6.80615842e-01 -4.27067608e-01 -1.16702926e+00
1.24581599e+00 3.75623465e-01 4.52225059e-01 -8.14364851e-01
-1.47218421e-01 7.44451106e-01 1.41270623e-01 -1.81222096e-01
1.95385441e-01 -6.45665824e-01 4.85453337e-01 4.21840787e-01
-2.49319360e-01 2.14841038e-01 -5.96075505e-02 -5.98372757e-01
6.81781292e-01 2.73367465e-01 2.13174429e-02 -4.85990435e-01
2.51858771e-01 1.52132452e-01 4.82885063e-01 6.36822820e-01
1.64555058e-01 2.64135338e-02 2.58302361e-01 -1.27685463e+00
-1.33776605e+00 -8.03566039e-01 -5.25535285e-01 1.18207908e+00
-1.03899765e+00 2.40848809e-01 -2.44373083e-01 -2.12367296e-01
7.69654036e-01 1.29141557e+00 -5.61068892e-01 5.75981021e-01
-3.07249933e-01 -1.00433242e+00 2.83818603e-01 6.30590737e-01
3.59176278e-01 -1.12609947e+00 -8.57767403e-01 5.31630635e-01
8.47873807e-01 -4.49104786e-01 4.37688321e-01 3.38329345e-01
-1.73461235e+00 -8.35995793e-01 -8.45774710e-01 2.85365313e-01
-2.52109677e-01 -3.89095902e-01 1.23534775e+00 -7.62606263e-02
-1.30670935e-01 7.29183555e-02 -1.84427612e-02 -9.85885501e-01
-6.11280240e-02 -1.54963762e-01 -1.83360383e-01 -3.00907999e-01
7.86339283e-01 -8.39728296e-01 -4.57370311e-01 9.66698974e-02
-4.58610684e-01 -4.39392507e-01 3.52841079e-01 8.48065615e-01
1.25561133e-01 2.41839856e-01 7.85129368e-01 -6.71445847e-01
8.03050101e-01 -8.62825871e-01 -1.23413312e+00 5.78630120e-02
-1.67660773e+00 -1.47470683e-01 3.31037343e-01 -3.68559986e-01
-1.06389618e+00 -6.40851855e-01 1.40565727e-02 -5.62392712e-01
2.51040608e-02 8.57058108e-01 6.00479662e-01 -1.53243691e-01
5.85188925e-01 -9.28566381e-02 1.10033773e-01 -9.26230192e-01
-8.21516737e-02 6.81793451e-01 1.77094713e-01 -5.49693882e-01
4.96236354e-01 1.62608594e-01 8.61415789e-02 -2.79760152e-01
-2.50771463e-01 -1.20610259e-01 -4.24283832e-01 -8.41487870e-02
3.20217252e-01 -8.05924714e-01 -5.13715506e-01 6.19937003e-01
-6.72324657e-01 -2.71190971e-01 -4.56763476e-01 8.52723300e-01
-5.37386417e-01 -1.60416484e-01 -6.54618800e-01 -1.41798735e+00
-6.52786076e-01 -3.65067244e-01 -1.64839908e-01 3.34565073e-01
-6.46630004e-02 -1.57980216e+00 2.39160895e-01 2.06164598e-01
1.19765270e+00 4.28992957e-01 9.16878343e-01 -1.48398554e+00
-1.47600874e-01 -7.64283895e-01 -9.46725681e-02 6.22423232e-01
-1.84618272e-02 3.17097962e-01 -7.58256912e-01 -2.00018629e-01
3.35782111e-01 5.74621595e-02 6.08664095e-01 4.92710203e-01
9.48331833e-01 -5.94466567e-01 2.52181292e-01 6.39963627e-01
1.56124580e+00 6.35719478e-01 5.78435123e-01 1.30628407e+00
-2.25437731e-02 5.51005423e-01 3.95519286e-01 5.26795387e-01
-3.69281992e-02 3.23601097e-01 3.04817677e-01 7.67743215e-02
7.63727725e-01 1.10762911e-02 3.90027948e-02 2.15828851e-01
-6.93709195e-01 5.98535359e-01 -1.21588087e+00 7.36161247e-02
-1.68870389e+00 -1.20452118e+00 -1.29439160e-01 2.32395053e+00
6.21551752e-01 5.98038793e-01 6.67180240e-01 -8.74800701e-03
5.57495415e-01 -2.21740618e-01 -7.95733929e-01 -9.38855529e-01
-1.83095872e-01 6.24386489e-01 9.67321575e-01 -2.45752670e-02
-1.12281311e+00 3.76621246e-01 7.41647959e+00 7.12038457e-01
-1.45965314e+00 -3.20118875e-03 9.89173234e-01 -2.47513607e-01
-4.36132967e-01 2.26147205e-01 -5.49386621e-01 8.78021896e-01
1.55708897e+00 -5.38970709e-01 5.69847465e-01 1.33029807e+00
8.90316889e-02 6.98340088e-02 -6.25078261e-01 6.80638433e-01
-6.30479693e-01 -1.59661460e+00 -2.27838233e-01 2.24261284e-01
6.84645474e-01 4.48434234e-01 2.60662407e-01 6.11387134e-01
5.40881217e-01 -1.22887611e+00 6.32359147e-01 1.04491842e+00
5.59401661e-02 -1.09585071e+00 9.91665781e-01 4.08253819e-01
-4.19661254e-01 -3.97698402e-01 -5.84101856e-01 -9.34561670e-01
-1.44159988e-01 8.63084555e-01 -5.19129992e-01 4.30734426e-01
1.11207867e+00 2.86900342e-01 -3.79845679e-01 1.29406321e+00
2.51881123e-01 7.88973987e-01 -6.64942086e-01 -1.89093247e-01
4.76324260e-01 -5.24723351e-01 1.42773882e-01 9.50314522e-01
5.40727317e-01 -5.73985651e-02 -1.98950753e-01 8.72064233e-01
6.00176573e-01 3.83864731e-01 -2.95045108e-01 7.38701150e-02
2.87734538e-01 8.80218446e-01 -4.78573829e-01 -3.98520082e-01
-8.57964694e-01 1.89960480e-01 -2.45981053e-01 2.95849860e-01
-4.76830095e-01 -2.61171907e-01 5.92882752e-01 3.06062549e-01
4.93165910e-01 -1.97151780e-01 -9.62688804e-01 -1.01341152e+00
-2.43457817e-02 -6.81922674e-01 5.58770657e-01 -7.16503143e-01
-1.85362172e+00 4.54085946e-01 -5.32803200e-02 -1.01314652e+00
-6.82471037e-01 -1.16100609e+00 -1.18363392e+00 1.43832612e+00
-1.54485238e+00 -6.73702776e-01 5.00465155e-01 3.59856427e-01
9.94219556e-02 -1.04507709e+00 9.28729057e-01 -8.80569220e-02
-1.60944968e-01 1.01055950e-01 7.15882003e-01 1.47642940e-01
2.68522024e-01 -1.41714346e+00 9.53270853e-01 1.82537541e-01
2.34196216e-01 4.43157315e-01 5.85751176e-01 -8.69894028e-01
-8.18596184e-01 -5.99608719e-01 5.06011724e-01 -6.08199537e-02
1.26570177e+00 4.00088668e-01 -9.34662819e-01 7.50171840e-01
2.88490981e-01 -6.89263269e-02 9.49231923e-01 2.95159489e-01
-3.40251327e-01 -2.36403465e-01 -1.26103437e+00 1.92423612e-01
-7.64838532e-02 -2.97046095e-01 -9.52770352e-01 1.85440570e-01
2.03406569e-02 -6.71621636e-02 -1.20118952e+00 4.36675340e-01
9.05663729e-01 -1.34934008e+00 1.05458832e+00 -1.31314826e+00
2.09638789e-01 5.06477535e-01 -1.40426576e-01 -9.43313479e-01
-2.81688899e-01 -8.37681234e-01 -1.52934581e-01 1.02027214e+00
6.58530593e-01 -1.23579276e+00 9.21252728e-01 1.35112858e+00
5.14256001e-01 -9.29513454e-01 -1.35118139e+00 -1.04408610e+00
8.76601040e-01 -8.70705426e-01 1.10601532e+00 1.20514953e+00
-2.01620832e-01 -4.17618752e-01 -3.84648770e-01 -5.43048263e-01
6.80732846e-01 3.51840824e-01 4.14608389e-01 -1.79309940e+00
-3.72886807e-01 -8.81093919e-01 -5.49239695e-01 -1.99216545e-01
1.85868382e-01 -5.55605650e-01 -1.12185061e+00 -8.40872705e-01
-2.83035159e-01 -7.70393014e-01 -8.49616647e-01 3.59660029e-01
1.58347845e-01 -2.03576192e-01 3.03623170e-01 5.22932947e-01
3.94994795e-01 4.85297889e-02 4.90669996e-01 -9.15823970e-03
-5.41403191e-03 7.97143281e-01 -3.98183256e-01 1.01241839e+00
1.19168174e+00 -3.17694724e-01 1.04308940e-01 -2.41522882e-02
4.51794326e-01 1.20273061e-01 3.94375473e-01 -7.63053298e-01
2.88973838e-01 -4.85127687e-01 8.52381110e-01 -5.86610317e-01
1.06107444e-01 -7.82570183e-01 5.83858967e-01 6.16954625e-01
-1.73957095e-01 4.93481666e-01 4.80925977e-01 1.82242379e-01
-1.87955573e-01 -6.99976504e-01 5.03581047e-01 -4.32957292e-01
-7.07862914e-01 1.73366398e-01 -2.74569601e-01 -2.01364711e-01
9.77017939e-01 -3.99738461e-01 -1.33296758e-01 -4.12908554e-01
-1.03258073e+00 1.49554415e-02 4.18696731e-01 2.53927618e-01
2.52948642e-01 -1.44017637e+00 -9.03756380e-01 1.87150285e-01
-7.25215256e-01 -7.53664136e-01 -6.62576109e-02 5.33033907e-01
-8.68623078e-01 6.60484076e-01 -7.67175078e-01 -7.42340833e-02
-6.53621018e-01 6.82646155e-01 6.22703075e-01 -3.92363191e-01
-5.16023874e-01 5.45745015e-01 -8.82519662e-01 -2.89006412e-01
-1.07634574e-01 -7.54340068e-02 -1.44385576e-01 3.12285155e-01
5.34990966e-01 9.03092563e-01 1.08525604e-01 9.78268124e-03
2.81492304e-02 2.08909005e-01 4.50108573e-02 -3.35655987e-01
1.69662642e+00 4.08205658e-01 -2.09143087e-01 8.15784097e-01
7.49977946e-01 -3.23897928e-01 -1.05049586e+00 -7.12543875e-02
8.53084445e-01 -6.53957963e-01 5.20232558e-01 -1.02258825e+00
-1.20414066e+00 1.00317323e+00 8.68532181e-01 6.41280472e-01
8.90381932e-01 -7.61639118e-01 5.21425903e-01 4.50902492e-01
3.66459072e-01 -1.50811350e+00 -6.97976947e-01 3.02793831e-01
7.68792450e-01 -1.11483729e+00 3.23127687e-01 3.27289283e-01
-6.01372600e-01 1.81544209e+00 -1.00484779e-02 -6.64388120e-01
1.23491573e+00 4.06721950e-01 1.99670643e-01 2.04415545e-01
-1.04151595e+00 3.25459838e-01 6.43018112e-02 3.64038616e-01
3.22443008e-01 4.13359940e-01 -2.29455799e-01 8.27242553e-01
-5.42826116e-01 5.48122466e-01 2.24518612e-01 8.27537537e-01
-3.73530954e-01 -1.12656164e+00 -6.35422647e-01 1.02315259e+00
-1.18744445e+00 -5.05045950e-01 2.02072449e-02 1.02011597e+00
-2.29426891e-01 4.63029385e-01 5.73584855e-01 -9.59835425e-02
1.85091972e-01 6.81259453e-01 -6.36035651e-02 3.40575799e-02
-8.74724448e-01 1.79572299e-01 1.59893155e-01 -1.79631487e-01
-1.39373526e-01 -1.12642670e+00 -6.23378634e-01 -8.49784434e-01
-3.69484693e-01 2.55525768e-01 1.13746321e+00 7.39475727e-01
4.23843674e-02 2.24973589e-01 1.03671968e+00 -8.95674706e-01
-1.49417031e+00 -1.12369359e+00 -9.94114101e-01 3.78828309e-02
-1.11454591e-01 -6.12859905e-01 -7.61951447e-01 -3.41489196e-01] | [4.4819440841674805, 4.222935199737549] |
2e801adb-72ca-4731-b213-5d503b51de39 | understanding-humans-in-crowded-scenes-deep | 1804.03287 | null | http://arxiv.org/abs/1804.03287v3 | http://arxiv.org/pdf/1804.03287v3.pdf | Understanding Humans in Crowded Scenes: Deep Nested Adversarial Learning and A New Benchmark for Multi-Human Parsing | Despite the noticeable progress in perceptual tasks like detection, instance
segmentation and human parsing, computers still perform unsatisfactorily on
visually understanding humans in crowded scenes, such as group behavior
analysis, person re-identification and autonomous driving, etc. To this end,
models need to comprehensively perceive the semantic information and the
differences between instances in a multi-human image, which is recently defined
as the multi-human parsing task. In this paper, we present a new large-scale
database "Multi-Human Parsing (MHP)" for algorithm development and evaluation,
and advances the state-of-the-art in understanding humans in crowded scenes.
MHP contains 25,403 elaborately annotated images with 58 fine-grained semantic
category labels, involving 2-26 persons per image and captured in real-world
scenes from various viewpoints, poses, occlusion, interactions and background.
We further propose a novel deep Nested Adversarial Network (NAN) model for
multi-human parsing. NAN consists of three Generative Adversarial Network
(GAN)-like sub-nets, respectively performing semantic saliency prediction,
instance-agnostic parsing and instance-aware clustering. These sub-nets form a
nested structure and are carefully designed to learn jointly in an end-to-end
way. NAN consistently outperforms existing state-of-the-art solutions on our
MHP and several other datasets, and serves as a strong baseline to drive the
future research for multi-human parsing. | ['Shuicheng Yan', 'Jianshu Li', 'Li Zhou', 'Jian Zhao', 'Yu Cheng', 'Terence Sim', 'Jiashi Feng'] | 2018-04-10 | null | null | null | null | ['multi-human-parsing', 'human-parsing'] | ['computer-vision', 'computer-vision'] | [ 2.67357409e-01 2.93759368e-02 4.74351257e-01 -7.09379673e-01
-7.10602283e-01 -2.66604185e-01 3.16583812e-01 -1.22438654e-01
-5.40615559e-01 5.52982211e-01 1.05131559e-01 2.34318182e-01
4.39520478e-01 -5.83031416e-01 -9.33784485e-01 -4.34906721e-01
1.12186030e-01 8.69261742e-01 6.77146494e-01 -2.74786949e-01
-1.98290080e-01 3.37550007e-02 -1.66079819e+00 5.35273135e-01
8.34241748e-01 7.56467760e-01 2.57342488e-01 1.18895423e+00
2.53031373e-01 9.32859898e-01 -9.62589979e-01 -9.41847801e-01
1.77931577e-01 -5.00948429e-01 -8.68148565e-01 4.41359967e-01
7.28696048e-01 -1.51751876e-01 -2.36537069e-01 1.25277424e+00
6.93721294e-01 3.60352099e-01 2.92306453e-01 -1.54679787e+00
-7.48886704e-01 2.70318896e-01 -8.44231963e-01 2.81017751e-01
4.29592997e-01 5.24206579e-01 6.45648539e-01 -6.40270710e-01
5.05306780e-01 1.79071975e+00 6.92229569e-01 9.59584773e-01
-6.33391857e-01 -5.58508694e-01 4.92602378e-01 3.21828902e-01
-9.33420181e-01 -9.75316614e-02 6.89393282e-01 -4.72555429e-01
6.77218139e-01 2.05475539e-01 7.13002086e-01 1.38111889e+00
-5.30077517e-02 1.22970796e+00 1.00271332e+00 -2.32400391e-02
8.63320157e-02 1.61632210e-01 2.81096816e-01 9.41915035e-01
3.38913321e-01 -2.45310515e-01 -3.93584996e-01 1.21777289e-01
7.37496972e-01 3.17339748e-02 1.58091068e-01 -2.04936057e-01
-1.09863508e+00 8.44073474e-01 5.84653974e-01 -1.62389532e-01
-2.80988723e-01 1.24817371e-01 4.26228940e-01 -2.79256076e-01
2.69364089e-01 7.93128312e-02 -2.93801248e-01 2.33951226e-01
-6.36688530e-01 7.77747273e-01 4.49892730e-01 1.14514267e+00
6.04046404e-01 -3.40440460e-02 -4.20022607e-01 9.62618113e-01
1.39367342e-01 7.71920025e-01 3.96194816e-01 -1.08110797e+00
6.99837387e-01 8.04117680e-01 2.11319506e-01 -1.09566629e+00
-7.42825806e-01 -4.04336125e-01 -9.67311621e-01 -4.06509191e-02
5.16224563e-01 -5.39053798e-01 -1.18115878e+00 1.72139120e+00
6.47804797e-01 2.67175496e-01 1.86097264e-01 1.25946438e+00
1.29121983e+00 6.42675579e-01 5.39534628e-01 3.43127638e-01
1.86972094e+00 -1.70155573e+00 -3.96859109e-01 -9.96283472e-01
-1.01863056e-01 -3.04624647e-01 1.05592895e+00 1.44832611e-01
-1.16784024e+00 -1.04623675e+00 -7.41094708e-01 -2.93311864e-01
-5.65161109e-01 1.09080404e-01 6.08955443e-01 6.55323684e-01
-9.47418928e-01 8.33617374e-02 -6.25756860e-01 -4.21470910e-01
8.17686558e-01 8.46794769e-02 -2.71832764e-01 -2.29658037e-01
-1.09901011e+00 4.63088095e-01 3.20086867e-01 1.82937652e-01
-1.17803586e+00 -4.06498998e-01 -9.70988214e-01 -2.64389932e-01
5.42797327e-01 -1.24633563e+00 1.19593418e+00 -1.15647984e+00
-9.77804720e-01 1.36613858e+00 -3.33429843e-01 -6.94792688e-01
8.87965918e-01 -3.85060221e-01 -5.44833958e-01 3.21755111e-01
6.46393418e-01 9.75148976e-01 9.34408307e-01 -1.57904446e+00
-1.11172676e+00 -7.72773147e-01 1.48005009e-01 3.20557028e-01
3.66703570e-01 4.94761437e-01 -9.17463481e-01 -6.15904331e-01
-3.16469759e-01 -9.51897085e-01 -6.05570972e-01 -5.08951485e-01
-7.96390355e-01 -3.33232522e-01 7.42835283e-01 -9.75417137e-01
5.47762752e-01 -1.68474007e+00 4.38253522e-01 -3.36877704e-01
4.67448503e-01 4.26441580e-01 -9.38436389e-02 -1.59804836e-01
2.39086449e-01 -2.83571724e-02 -4.78015870e-01 -8.80369663e-01
-8.14337190e-03 1.18053243e-01 1.66629106e-01 2.93203831e-01
3.12159628e-01 1.27014351e+00 -9.67129469e-01 -8.15491557e-01
4.12633091e-01 3.52106959e-01 -3.60605955e-01 6.45004451e-01
-2.81583905e-01 7.39259720e-01 -4.91522521e-01 1.09935105e+00
6.54753864e-01 -3.18111777e-01 -2.66780227e-01 5.37307076e-02
3.12765390e-01 -7.19520569e-01 -1.15592790e+00 1.56582844e+00
1.18503742e-01 5.19247472e-01 1.93007082e-01 -8.68605375e-01
5.80159128e-01 -1.42117783e-01 1.29447415e-01 -6.70007467e-01
2.28151739e-01 -2.25260898e-01 -3.22996080e-01 -6.84595883e-01
6.31922841e-01 3.56557220e-01 -5.19531250e-01 -1.53912336e-01
8.39351043e-02 2.05439731e-01 4.24990356e-01 2.43747726e-01
9.20537114e-01 -2.32624099e-01 1.27524316e-01 -4.35145795e-02
6.39804959e-01 1.82077050e-01 8.66308928e-01 9.10804033e-01
-9.51782048e-01 9.30487275e-01 3.77740622e-01 -7.43134856e-01
-9.49186087e-01 -1.05660272e+00 3.30846846e-01 1.58453286e+00
6.95970714e-01 3.83748785e-02 -1.41267240e+00 -8.04738522e-01
-6.09331802e-02 4.11269009e-01 -7.79714167e-01 1.88851923e-01
-8.00700605e-01 -9.84679997e-01 7.93682992e-01 7.76767492e-01
1.28224778e+00 -1.54179502e+00 -7.53818512e-01 7.93963969e-02
-4.26925182e-01 -1.67725229e+00 -3.11500907e-01 -1.95899636e-01
-7.32425898e-02 -1.39342153e+00 -8.43259156e-01 -1.05328906e+00
4.06765133e-01 4.05707061e-01 1.54515386e+00 -9.00842622e-02
-6.12813294e-01 4.45801467e-01 -2.97790468e-01 -6.61037028e-01
-8.04772601e-02 -2.18785480e-01 -1.60790294e-01 2.60826021e-01
6.70195222e-01 -5.55452369e-02 -8.71669888e-01 5.66506743e-01
-5.17302871e-01 2.00969666e-01 5.31799912e-01 6.80600345e-01
6.99715197e-01 1.11232579e-01 5.13292849e-01 -1.14793766e+00
4.54976112e-01 -4.09250170e-01 -3.33808005e-01 4.22360837e-01
1.54990017e-01 -5.75112045e-01 6.88078880e-01 -1.90289661e-01
-1.25928950e+00 9.46268737e-02 -1.57569543e-01 -3.03900599e-01
-8.30201030e-01 -4.05754566e-01 -4.86302376e-01 5.95205277e-02
4.27003771e-01 1.63468599e-01 -4.35067296e-01 -1.36805534e-01
4.44728583e-01 4.59176689e-01 1.23044717e+00 -4.40672278e-01
6.31738961e-01 4.63826627e-01 -3.58277649e-01 -7.90995240e-01
-1.14547300e+00 -6.53153539e-01 -6.96394980e-01 -3.38653713e-01
1.76451313e+00 -1.32077539e+00 -7.14879870e-01 9.24337387e-01
-1.37373996e+00 -4.84589249e-01 1.34645700e-01 -2.24499434e-01
-5.30315638e-01 2.90120304e-01 -7.48401523e-01 -8.98046434e-01
-3.00484270e-01 -1.27535689e+00 1.56444383e+00 7.27178931e-01
7.99344629e-02 -9.36783075e-01 -2.63573498e-01 1.24004900e+00
7.06796814e-03 6.88409686e-01 3.54763895e-01 -6.18994355e-01
-6.48224652e-01 -6.02468699e-02 -5.39625287e-01 2.13183537e-01
-4.92864788e-01 -4.18859869e-01 -1.05981159e+00 -1.68013260e-01
-1.83681175e-01 -3.92165095e-01 9.90288973e-01 6.34241939e-01
1.45110798e+00 -4.23150063e-02 -4.25630897e-01 7.14131773e-01
1.12074041e+00 1.01213448e-01 5.07429838e-01 2.31369436e-01
1.30752778e+00 9.49701667e-01 6.57089472e-01 3.52214813e-01
9.22756374e-01 4.61661339e-01 5.51686406e-01 -4.29983169e-01
-2.85580039e-01 -3.83860379e-01 2.87111904e-02 1.37705475e-01
-2.51209259e-01 -5.72225928e-01 -9.77338970e-01 5.85645795e-01
-2.09835887e+00 -1.03766537e+00 -2.69392073e-01 1.54141426e+00
2.01565117e-01 1.41856298e-01 6.67037368e-01 -3.61882657e-01
1.24612486e+00 3.43733102e-01 -8.63499463e-01 5.70581928e-02
-3.79632831e-01 -2.75335699e-01 5.60431302e-01 2.67069787e-01
-1.72207391e+00 1.31435728e+00 5.70070219e+00 5.31984925e-01
-4.12920773e-01 3.24430466e-01 9.98131931e-01 1.73817217e-01
3.54464471e-01 -5.82373261e-01 -1.15863144e+00 6.96825743e-01
5.66484988e-01 3.49440902e-01 3.17539841e-01 1.12689245e+00
4.67440952e-03 -4.82258610e-02 -8.52517545e-01 1.37770188e+00
4.35982972e-01 -9.36187744e-01 9.11999941e-02 -4.08060998e-01
8.90269458e-01 -3.48118097e-02 4.47460786e-02 4.81238306e-01
5.26915669e-01 -1.12296116e+00 7.08754539e-01 3.26350987e-01
2.39373773e-01 -8.37051809e-01 9.74240482e-01 3.71814698e-01
-1.41080284e+00 -3.02668005e-01 -4.59027916e-01 9.96470824e-02
4.92918640e-01 2.39576623e-01 -3.50595444e-01 3.26810271e-01
1.26219714e+00 6.45133972e-01 -1.16364992e+00 6.35697722e-01
-1.59777522e-01 2.64642894e-01 1.47333145e-01 -7.07944483e-02
3.60344738e-01 -8.74090940e-02 4.99551326e-01 1.40116644e+00
-3.53137374e-01 2.81766742e-01 6.23456299e-01 8.25513184e-01
4.15214105e-03 -3.10446382e-01 -2.92785913e-01 2.95828789e-01
3.49540621e-01 1.31810784e+00 -9.57349658e-01 -7.22083867e-01
-5.02795100e-01 1.36458242e+00 3.42399240e-01 4.76107746e-01
-1.22247291e+00 -1.47003472e-01 8.28185141e-01 -1.84855405e-02
2.37331480e-01 -1.05378702e-02 -3.02564144e-01 -1.08580470e+00
2.21653104e-01 -1.15066290e+00 6.85851753e-01 -9.00908470e-01
-1.43961966e+00 8.63208771e-01 -2.42613908e-02 -6.11680746e-01
-1.55334994e-01 -7.30141521e-01 -5.40753245e-01 5.72569847e-01
-1.36206651e+00 -1.58609056e+00 -9.31234717e-01 9.32031453e-01
1.19260168e+00 -3.75060976e-01 3.46037060e-01 2.55574316e-01
-9.85234678e-01 6.20744407e-01 -5.21759093e-01 6.07388675e-01
5.25928676e-01 -1.57709360e+00 8.62123251e-01 1.26978695e+00
-1.04989782e-01 1.33842200e-01 7.82036245e-01 -7.81852782e-01
-1.04273474e+00 -1.72785020e+00 3.83953333e-01 -8.58276129e-01
2.51230270e-01 -5.28899252e-01 -7.12634563e-01 8.66702139e-01
2.71019042e-01 1.64870411e-01 3.52912605e-01 -1.37958497e-01
-4.99308109e-02 6.99856281e-02 -1.34591115e+00 5.84911168e-01
1.52621734e+00 -2.79441595e-01 -5.67459702e-01 6.99049532e-01
1.12738395e+00 -6.62238836e-01 -4.23740387e-01 4.01268691e-01
1.11831985e-01 -1.35724306e+00 1.44793630e+00 -7.76802182e-01
5.52370131e-01 -2.81472415e-01 -2.04980701e-01 -1.17256451e+00
-3.99714530e-01 -4.41098183e-01 2.56972425e-02 1.10536134e+00
6.20083883e-02 -4.15864378e-01 1.06725585e+00 7.75048494e-01
-3.17811966e-01 -5.51608741e-01 -5.56148529e-01 -3.47035438e-01
-1.85008734e-01 -3.33333760e-01 7.00281978e-01 4.58073288e-01
-7.81888485e-01 4.75440383e-01 -6.83286309e-01 6.04363680e-01
1.16320145e+00 -1.01545602e-01 1.24924898e+00 -1.13147140e+00
-4.58306253e-01 -3.10151607e-01 -6.49592400e-01 -1.04026437e+00
2.34843701e-01 -2.60810792e-01 1.84887171e-01 -1.49592078e+00
5.57951272e-01 -5.58931641e-02 1.71413608e-02 2.03967497e-01
-7.82205820e-01 3.59144717e-01 3.93050313e-01 -4.04194556e-03
-1.45858860e+00 2.64849752e-01 1.27427566e+00 -4.02842164e-01
-5.23016229e-02 1.18350826e-01 -7.51206636e-01 1.10237753e+00
6.47815764e-01 -1.14019722e-01 -2.05145076e-01 -5.83704054e-01
-3.30075383e-01 1.36518151e-01 1.02262545e+00 -1.40643430e+00
4.42627996e-01 -1.34586290e-01 7.93982744e-01 -8.81175697e-01
5.06638944e-01 -4.86678362e-01 -1.24670319e-01 2.81002969e-01
-2.50859380e-01 2.85240352e-01 -5.28228767e-02 8.99620116e-01
-2.79105484e-01 1.95530862e-01 9.02906001e-01 -6.18044376e-01
-1.55592656e+00 5.40530443e-01 9.43149924e-02 6.45135164e-01
1.31546116e+00 -1.86106771e-01 -5.33106089e-01 -1.77018359e-01
-7.83780813e-01 6.98265851e-01 2.71076947e-01 7.57987738e-01
6.41372263e-01 -9.40116942e-01 -9.26351488e-01 1.35199666e-01
1.54648110e-01 4.91436034e-01 9.30584311e-01 2.55476624e-01
-5.72993040e-01 2.12615192e-01 -3.32116336e-01 -8.86893809e-01
-1.39850593e+00 6.43405795e-01 4.79199320e-01 -3.12781185e-01
-5.32716691e-01 1.21919250e+00 5.48430026e-01 -5.56723177e-01
2.24918917e-01 -3.17723304e-02 -1.97984442e-01 -2.56582916e-01
7.78454781e-01 5.82209289e-01 -3.42783689e-01 -9.57717836e-01
-4.33651268e-01 6.15755796e-01 1.07915573e-01 2.65077084e-01
1.06945193e+00 -2.80821681e-01 8.26180577e-02 2.46316627e-01
9.61087167e-01 -4.68257844e-01 -1.56867826e+00 -1.15327120e-01
-3.77188236e-01 -4.20486748e-01 -6.48891747e-01 -8.26231122e-01
-1.17553437e+00 8.48453343e-01 6.22267306e-01 1.77537903e-01
9.76837575e-01 3.19516420e-01 1.07057524e+00 1.46426365e-01
5.33855081e-01 -1.39128971e+00 4.40393269e-01 2.74740577e-01
7.04850137e-01 -1.85365176e+00 -3.81677151e-01 -6.92650259e-01
-1.27075994e+00 5.02110183e-01 1.16538608e+00 -1.54967874e-01
1.00449927e-01 -1.34154782e-01 2.23878711e-01 -3.04922432e-01
-2.56376475e-01 -5.62476099e-01 2.37254560e-01 1.11072588e+00
-1.07919455e-01 5.33717036e-01 3.50274235e-01 1.04267085e+00
-4.16186631e-01 -5.12740433e-01 1.99843839e-01 4.57997143e-01
-6.06283188e-01 -5.06619453e-01 -6.48663819e-01 2.10238546e-01
-3.30829322e-01 3.16930830e-01 -5.48026919e-01 9.30494606e-01
5.71291745e-01 1.32987440e+00 4.61855195e-02 -3.41024876e-01
5.00445426e-01 -1.94212124e-01 1.14642113e-01 -4.40581203e-01
-5.93994260e-01 -4.02878284e-01 9.85779166e-02 -7.70961881e-01
-6.26487613e-01 -7.06553876e-01 -1.07991600e+00 -2.15955913e-01
3.26735914e-01 -1.99490517e-01 3.86449158e-01 1.02238166e+00
2.76678681e-01 8.88749540e-01 2.01178133e-01 -1.13292742e+00
-7.79014602e-02 -8.44908476e-01 -4.29291070e-01 1.03425169e+00
1.26095235e-01 -7.47228682e-01 -1.08333163e-01 2.02208281e-01] | [8.41385555267334, -0.17422939836978912] |
ceeb4bc9-8e59-4959-99fb-f847bf065a6a | novel-and-fast-algorithm-for-extracting | 1407.6496 | null | http://arxiv.org/abs/1407.6496v2 | http://arxiv.org/pdf/1407.6496v2.pdf | Novel and Fast Algorithm for Extracting License Plate Location Based on Edge Analysis | Nowadays in developing or developed countries, the Intelligent Transportation
System (ITS) technology has attracted so much attention to itself. License
Plate Recognition (LPR) systems have many applications in ITSs, such as the
payment of parking fee, controlling the traffic volume, traffic data
collection, etc. This paper presents a new and fast method for license plate
extraction based on edge analysis. our proposed method consist of four stage,
which are edge detection, non-useable edge and noise removing, edge analysis
and morphology-based license plate extraction. In the result part, the proposed
algorithm is applied on vehicle database and the accuracy rate reached 98%.
From the experimental results it is shown that the proposed method gives fairly
acceptable level of accuracy for practical license plate recognition system. | ['Reza Azad', 'Mohammad Baghdadi'] | 2014-07-24 | null | null | null | null | ['license-plate-recognition'] | ['computer-vision'] | [-1.42405629e-01 -8.97858858e-01 4.62593976e-03 -8.89712665e-03
-4.29820448e-01 -5.65824568e-01 3.24023038e-01 -3.94325703e-01
-5.43107092e-01 6.09972060e-01 -2.13901520e-01 -6.16111994e-01
9.85959396e-02 -9.31587458e-01 -1.86465591e-01 -5.07358491e-01
5.63401341e-01 2.76451558e-01 7.76649296e-01 -1.27347335e-01
1.03421593e+00 9.82455373e-01 -1.27217555e+00 -2.34339297e-01
7.72337556e-01 7.07304418e-01 6.01367205e-02 6.87035501e-01
-2.16368586e-01 5.11162281e-01 -2.01272234e-01 -3.72135252e-01
4.84058619e-01 -1.71145424e-01 -1.92251474e-01 6.32174313e-01
-2.28740811e-01 -3.41604054e-01 -5.86118639e-01 9.97183800e-01
2.72711217e-01 3.73905420e-01 9.61612165e-01 -9.39048231e-01
-4.83259469e-01 -4.53908235e-01 -7.99678802e-01 2.45584324e-01
-1.62125286e-02 -7.23816361e-03 1.53544903e-01 -1.32806480e+00
4.90010977e-01 6.25695229e-01 8.25344026e-01 2.16662630e-01
-3.91857237e-01 -5.02862215e-01 -7.33467579e-01 7.20035315e-01
-1.75074196e+00 -6.33403778e-01 7.07699180e-01 -4.08828110e-01
8.36908281e-01 2.77403027e-01 3.34670782e-01 -2.83920288e-01
4.07895446e-01 6.70149863e-01 1.15817356e+00 -5.96836150e-01
5.71336560e-02 4.15604949e-01 5.02279758e-01 5.40871561e-01
5.20131826e-01 -2.78793275e-01 3.82610053e-01 4.60935414e-01
8.07228446e-01 1.21934369e-01 3.86826247e-02 2.01550141e-01
-5.54223180e-01 4.97271955e-01 -6.61384538e-02 3.98526400e-01
-3.60840529e-01 -3.41346234e-01 1.54289767e-01 -1.63940415e-01
5.67437150e-02 -2.21162930e-01 -9.11070853e-02 -3.77640694e-01
-9.57714081e-01 -7.19278380e-02 4.58792180e-01 1.09119070e+00
5.52594125e-01 2.25057110e-01 2.84246415e-01 9.74394560e-01
5.48311591e-01 8.86726916e-01 3.10651720e-01 -4.73808885e-01
5.68095088e-01 6.02906704e-01 3.26154977e-01 -1.28784263e+00
-1.02060035e-01 2.87746787e-01 -5.45080304e-01 2.90083796e-01
1.97211772e-01 -4.32162315e-01 -1.06297386e+00 3.64989102e-01
-1.30884975e-01 1.57178074e-01 9.47947651e-02 7.01608837e-01
8.03804815e-01 1.20842898e+00 -8.76776651e-02 -4.41532880e-02
1.43360221e+00 -9.65229928e-01 -9.18251574e-01 -1.00805528e-01
2.23127380e-01 -1.14405417e+00 5.92299342e-01 1.67575940e-01
-8.61684680e-01 -3.36046249e-01 -1.03204131e+00 1.58608019e-01
-5.80779254e-01 7.04950750e-01 2.23679185e-01 8.27812076e-01
-7.30908334e-01 -1.29752606e-01 -5.08379459e-01 -5.39435387e-01
2.29162380e-01 6.15190268e-01 -5.65141201e-01 -3.35359454e-01
-6.35001659e-01 1.08402622e+00 2.62636170e-02 4.21997964e-01
2.65455842e-01 2.01554909e-01 -5.91813684e-01 3.49521004e-02
1.89200863e-01 -4.58652005e-02 6.40308976e-01 -5.56073129e-01
-1.52154708e+00 8.24431360e-01 -2.92802662e-01 2.04180032e-01
2.66868830e-01 3.12271625e-01 -9.34574902e-01 4.72304039e-02
-9.95154381e-02 5.56457303e-02 4.94412601e-01 -6.76972985e-01
-9.05400574e-01 -6.55735910e-01 -6.30355954e-01 2.86420166e-01
2.35128418e-01 4.41596448e-01 -7.93348551e-01 -2.61001587e-01
2.01845556e-01 -1.00647366e+00 1.30748391e-01 -3.37978631e-01
-2.16093481e-01 -2.71986336e-01 1.30691981e+00 -8.91123354e-01
1.06819344e+00 -2.49002218e+00 -7.46891618e-01 6.50452316e-01
-3.84957165e-01 6.17355287e-01 3.86165082e-01 3.58765453e-01
1.88208312e-01 8.20056647e-02 -9.09124687e-02 2.53739506e-01
-9.86931399e-02 -2.37863421e-01 1.84836954e-01 5.49246311e-01
1.75880387e-01 4.99946535e-01 2.06667949e-02 -5.41407943e-01
5.96909881e-01 5.15150487e-01 5.45450114e-02 -1.78633198e-01
7.59160578e-01 -2.51354158e-01 -4.97441769e-01 1.06954098e+00
1.12082970e+00 4.09096956e-01 -2.85750210e-01 -1.01911493e-01
-6.96120203e-01 -3.96511406e-01 -1.31988859e+00 2.77797788e-01
1.36707339e-03 1.34373426e+00 1.12658434e-01 -5.53243518e-01
1.28069079e+00 2.93139547e-01 2.37975419e-01 -6.84164405e-01
5.16204953e-01 5.51477969e-01 -6.19884580e-02 -9.69655752e-01
9.10786152e-01 -7.31343850e-02 1.99117377e-01 -1.02298066e-01
-5.90329587e-01 2.20332369e-01 3.48293692e-01 -2.90905684e-01
6.05429769e-01 -2.62775302e-01 1.60397470e-01 -1.00255564e-01
1.12363529e+00 5.07059455e-01 4.61606205e-01 1.02057733e-01
-5.66458941e-01 3.06983501e-01 -1.36811927e-01 -9.80551764e-02
-1.22754443e+00 -6.35586321e-01 -2.81224877e-01 2.37507120e-01
5.57374120e-01 5.87974787e-01 -7.10397243e-01 -4.97626178e-02
-6.87410533e-02 4.84866470e-01 7.09562302e-02 3.04366678e-01
-8.73476923e-01 -4.60281461e-01 5.17893076e-01 3.43623936e-01
1.22384429e+00 -1.07071161e+00 -7.31539279e-02 3.10446359e-02
8.57316181e-02 -1.19161403e+00 -7.03151405e-01 -6.21040106e-01
-7.72174954e-01 -8.92377257e-01 -8.70025218e-01 -1.44172621e+00
9.21109378e-01 7.05485225e-01 2.40782395e-01 4.13174406e-02
-9.43788737e-02 1.77423984e-01 -5.70825562e-02 -3.30821574e-01
-1.77195534e-01 -2.65398085e-01 -1.76269114e-01 1.65323451e-01
9.49797511e-01 -8.66888314e-02 -5.30621529e-01 6.50120616e-01
-6.14181399e-01 -3.67541432e-01 9.77748871e-01 1.00816652e-01
2.21369073e-01 6.59706771e-01 5.28931737e-01 -6.42999768e-01
8.91053855e-01 -2.21384779e-01 -9.93726373e-01 2.17586815e-01
-6.26841903e-01 -5.19790828e-01 4.81023610e-01 2.20790863e-01
-1.10752654e+00 -3.53413299e-02 -3.43665570e-01 3.77820656e-02
-4.17086631e-01 3.10747832e-01 -3.92827988e-01 -4.46718305e-01
-3.18480954e-02 8.51942658e-01 2.05679774e-01 -3.55971605e-01
-3.35155994e-01 1.43630910e+00 7.47479200e-01 2.13765755e-01
8.42040718e-01 2.07705408e-01 1.38345286e-01 -1.27978659e+00
5.73869467e-01 -9.89625216e-01 -3.63602430e-01 -7.86596775e-01
9.17566419e-01 -4.87900585e-01 -1.02220201e+00 1.17580676e+00
-7.93973565e-01 4.21290815e-01 5.46289206e-01 8.52316737e-01
-3.58920656e-02 4.92942065e-01 -5.52475750e-01 -1.02230740e+00
-1.95522800e-01 -1.17488468e+00 3.35599750e-01 7.74360836e-01
4.08993691e-01 -9.57748711e-01 -9.36150774e-02 5.67740142e-01
5.74373305e-01 -1.36326715e-01 6.47687852e-01 -5.12521684e-01
-8.37654769e-01 -8.81906152e-01 -5.59726119e-01 2.06775814e-01
2.13260084e-01 3.99435967e-01 -4.45612788e-01 2.85059988e-01
-1.97752014e-01 3.91101271e-01 6.37114286e-01 6.01633251e-01
3.89084697e-01 -1.21286027e-01 -3.41100514e-01 2.36154288e-01
1.94376266e+00 9.30250466e-01 1.49224520e+00 7.94606924e-01
4.64076906e-01 1.72484130e-01 6.33152306e-01 1.74920589e-01
2.27075338e-01 4.73008901e-01 -1.65723488e-01 3.66748832e-02
6.86025396e-02 1.19289100e-01 5.14458656e-01 8.08529019e-01
-4.51906383e-01 -2.74694711e-01 -9.23172891e-01 3.85938793e-01
-1.42637503e+00 -1.24723780e+00 -7.49816298e-01 2.02252197e+00
1.43848047e-01 2.05744356e-01 8.99041891e-02 3.71476769e-01
1.25237393e+00 -4.66585636e-01 -2.61862755e-01 -7.91753590e-01
3.61635908e-02 -2.29207158e-01 1.03531682e+00 6.53933942e-01
-9.57623720e-01 7.98558950e-01 5.81610060e+00 7.43236244e-01
-1.43285739e+00 -4.01467621e-01 3.76559138e-01 8.28190148e-01
9.35914516e-02 -2.50781357e-01 -8.90563190e-01 6.23661697e-01
5.49262822e-01 -2.60214806e-01 2.09178403e-01 7.08289564e-01
5.33789039e-01 -5.45794010e-01 -1.63835548e-02 1.16475916e+00
6.48192838e-02 -1.35841930e+00 -1.31469622e-01 1.75893977e-01
3.66545498e-01 -1.72724798e-01 8.81448165e-02 1.63799256e-01
-4.71712500e-01 -6.32419288e-01 3.03853422e-01 7.19503880e-01
6.62837088e-01 -1.06574249e+00 1.10772169e+00 3.14297438e-01
-1.19755745e+00 1.04984805e-01 -3.58045578e-01 6.04593977e-02
1.44702628e-01 1.41971990e-01 -1.09492826e+00 1.11589454e-01
2.28437558e-02 4.50242430e-01 -2.98073798e-01 1.61529624e+00
3.89421284e-01 6.21308267e-01 -3.03129047e-01 -3.95489961e-01
4.33737874e-01 -8.53669107e-01 6.30039632e-01 1.43014312e+00
5.08661032e-01 4.79641140e-01 -3.86531115e-01 3.90922934e-01
-1.67062487e-02 3.86969477e-01 -6.28535330e-01 1.07522942e-01
5.65629125e-01 1.12705350e+00 -1.11758029e+00 -2.45610759e-01
-7.20482945e-01 5.67325175e-01 -6.43789530e-01 2.51243740e-01
-8.32677841e-01 -1.25330234e+00 6.64766878e-02 4.29620773e-01
4.63258147e-01 -5.38436592e-01 -4.66683775e-01 -8.38484228e-01
-1.10522108e-02 -3.09738189e-01 2.01575495e-02 -7.62265086e-01
-6.48006558e-01 1.86418265e-01 -3.25171530e-01 -1.47532618e+00
1.13633424e-01 -1.12220323e+00 -9.79531884e-01 8.28660011e-01
-1.34515595e+00 -8.32433641e-01 -8.57986882e-02 6.23473346e-01
8.61401737e-01 -7.07250535e-01 2.03052148e-01 6.31388903e-01
-1.04409325e+00 4.41819727e-01 7.56209791e-01 6.58878505e-01
3.43487829e-01 -6.88933611e-01 -8.28302000e-03 1.29289949e+00
-5.16498685e-01 4.51774687e-01 5.61010897e-01 -6.28149807e-01
-1.46193564e+00 -8.62356544e-01 1.05827975e+00 3.96181084e-02
2.43102133e-01 1.90286785e-01 -6.92593515e-01 4.24773574e-01
1.89887881e-01 -2.12696552e-01 4.97076184e-01 -5.81902683e-01
3.69718164e-01 -2.78586894e-01 -1.51239693e+00 5.74386835e-01
-5.44044713e-04 -4.14275676e-02 -4.22203809e-01 -1.05247520e-01
-3.12716603e-01 -4.22910601e-02 -4.63499933e-01 7.52235577e-02
6.97257340e-01 -7.00778127e-01 5.45027912e-01 -3.67803872e-02
6.94947615e-02 -7.95480371e-01 3.97522971e-02 -5.59063673e-01
-4.95129347e-01 -4.31124568e-02 7.39852309e-01 1.36774135e+00
5.83919346e-01 -7.26173580e-01 8.70340705e-01 1.19171405e+00
-2.93581307e-01 -3.45222175e-01 -7.22248077e-01 -5.15738368e-01
-5.12347698e-01 -1.81445614e-01 9.22679752e-02 6.13558710e-01
9.38171595e-02 1.96289226e-01 -2.14623824e-01 2.67509907e-01
5.55124938e-01 -1.65288776e-01 5.43621659e-01 -8.77780080e-01
3.93177301e-01 -5.20571530e-01 -9.26427722e-01 -8.53125989e-01
-1.29332796e-01 -2.88623452e-01 1.53612673e-01 -1.85045445e+00
1.11489728e-01 -2.08421424e-01 -8.39952081e-02 7.00390339e-02
2.12984681e-02 5.75113535e-01 9.90272835e-02 5.89826107e-01
-4.54640210e-01 -2.98428386e-02 1.05847919e+00 -1.77911714e-01
-2.62386441e-01 4.49552208e-01 -2.27765843e-01 7.30637074e-01
9.21500981e-01 -2.44094044e-01 -1.40840039e-01 -1.77112803e-01
-1.61571130e-01 1.38567537e-01 -1.12462886e-01 -9.10081267e-01
6.39011681e-01 -2.45547250e-01 5.40599048e-01 -8.55019569e-01
1.31857499e-01 -1.09668887e+00 6.91722259e-02 2.89932519e-01
5.08654654e-01 2.19946995e-01 3.72574896e-01 3.05167615e-01
-3.21134776e-01 -5.51499307e-01 8.08656335e-01 1.57387570e-01
-1.41271591e+00 1.41592836e-02 -1.17823839e+00 -4.23969835e-01
1.33253646e+00 -1.12436306e+00 -4.03471172e-01 -1.96099773e-01
-3.70852351e-01 -1.10932693e-01 4.57071453e-01 1.58907250e-01
8.37743640e-01 -1.28904247e+00 -7.40459085e-01 2.04151377e-01
-9.08243433e-02 -5.43881774e-01 2.51557380e-01 8.74094427e-01
-1.42062235e+00 7.50734627e-01 -6.11171246e-01 -2.35816985e-01
-1.52556217e+00 1.30008996e-01 1.33263409e-01 2.79152304e-01
-3.36929381e-01 1.89221725e-01 -6.69135273e-01 6.02375492e-02
-3.59084129e-01 1.04148045e-01 -6.12207174e-01 -3.47648025e-01
3.74819398e-01 1.03599215e+00 1.16878010e-01 -1.15421307e+00
-6.02909386e-01 1.17742360e+00 -3.36889811e-02 -3.27903003e-01
1.01606917e+00 -2.45590165e-01 -2.17061460e-01 3.06710731e-02
1.00662136e+00 5.56488335e-01 -8.75737190e-01 4.20669496e-01
3.57031561e-02 -7.42256701e-01 1.87227905e-01 -5.68964422e-01
-8.82457852e-01 5.47370732e-01 5.02934813e-01 1.82622969e-01
1.14888370e+00 -6.12957060e-01 8.82130563e-01 5.43542385e-01
3.19957703e-01 -1.53890955e+00 -7.67334878e-01 5.69874763e-01
3.20956707e-01 -1.24621248e+00 -1.08547322e-01 -3.57288986e-01
-6.27114415e-01 1.39500594e+00 4.59347606e-01 -4.32170898e-01
7.81363189e-01 4.30053622e-01 5.23756817e-02 2.38700688e-01
-2.62172520e-01 -9.29745287e-02 2.10750610e-01 5.53763986e-01
3.00651789e-01 -3.71599421e-02 -7.74269462e-01 4.40994203e-01
4.31391507e-01 3.09184909e-01 9.01337385e-01 9.58134294e-01
-1.15371549e+00 -7.51976788e-01 -8.94149482e-01 4.79232192e-01
-5.70681870e-01 6.63684458e-02 -2.71872163e-01 1.18800747e+00
3.52896587e-03 1.00398183e+00 2.19006255e-01 -3.63198578e-01
4.83607084e-01 8.31089169e-02 2.58215100e-01 1.66036665e-01
8.67237300e-02 -1.16438884e-03 1.44138217e-01 2.57499605e-01
-2.88067132e-01 -7.39837408e-01 -1.52797151e+00 -7.93139219e-01
-5.60828626e-01 3.99076819e-01 1.07188570e+00 9.55888510e-01
2.07058206e-01 -9.73700266e-03 7.38885403e-01 -3.58163565e-01
1.18286483e-01 -8.12720418e-01 -9.43658590e-01 1.25562713e-01
-8.68798271e-02 -4.28202122e-01 -1.97907016e-01 3.17838639e-01] | [9.798116683959961, -4.993718147277832] |
5b490c77-7685-4164-a960-50649b7971cf | computer-aided-interpretable-features-for | 2106.08077 | null | https://arxiv.org/abs/2106.08077v3 | https://arxiv.org/pdf/2106.08077v3.pdf | Computer-aided Interpretable Features for Leaf Image Classification | Plant species identification is time consuming, costly, and requires lots of efforts, and expertise knowledge. In recent, many researchers use deep learning methods to classify plants directly using plant images. While deep learning models have achieved a great success, the lack of interpretability limit their widespread application. To overcome this, we explore the use of interpretable, measurable and computer-aided features extracted from plant leaf images. Image processing is one of the most challenging, and crucial steps in feature-extraction. The purpose of image processing is to improve the leaf image by removing undesired distortion. The main image processing steps of our algorithm involves: i) Convert original image to RGB (Red-Green-Blue) image, ii) Gray scaling, iii) Gaussian smoothing, iv) Binary thresholding, v) Remove stalk, vi) Closing holes, and vii) Resize image. The next step after image processing is to extract features from plant leaf images. We introduced 52 computationally efficient features to classify plant species. These features are mainly classified into four groups as: i) shape-based features, ii) color-based features, iii) texture-based features, and iv) scagnostic features. Length, width, area, texture correlation, monotonicity and scagnostics are to name few of them. We explore the ability of features to discriminate the classes of interest under supervised learning and unsupervised learning settings. For that, supervised dimensionality reduction technique, Linear Discriminant Analysis (LDA), and unsupervised dimensionality reduction technique, Principal Component Analysis (PCA) are used to convert and visualize the images from digital-image space to feature space. The results show that the features are sufficient to discriminate the classes of interest under both supervised and unsupervised learning settings. | ['Thiyanga S. Talagala', 'Jayani P. G. Lakshika'] | 2021-06-15 | null | null | null | null | ['supervised-dimensionality-reduction'] | ['computer-vision'] | [ 3.29090387e-01 -4.46006060e-01 -2.39397243e-01 -2.67051160e-01
-7.36070424e-02 -1.00915706e+00 3.34593594e-01 3.38684827e-01
-7.99950659e-02 4.60190862e-01 -3.83436978e-01 -5.13942301e-01
-5.74786842e-01 -1.04228687e+00 4.83802594e-02 -1.05586493e+00
-1.04393408e-01 9.29729640e-02 1.84758827e-01 7.77998641e-02
2.19151765e-01 1.22103167e+00 -1.66143346e+00 -7.18788281e-02
9.49741244e-01 1.21966267e+00 5.14417350e-01 7.83407569e-01
-5.33000469e-01 5.18415689e-01 -2.93559998e-01 3.24764639e-01
2.66505927e-01 -4.63152915e-01 -8.60914588e-01 6.50296628e-01
-2.12666661e-01 -3.15276921e-01 1.01960525e-01 1.11975551e+00
3.00660461e-01 -2.67256767e-01 9.91650462e-01 -1.18067443e+00
-1.10361028e+00 2.84679711e-01 -8.56774867e-01 -9.09352228e-02
-9.59976688e-02 1.82070792e-01 6.57744110e-01 -7.96609104e-01
3.11323375e-01 1.03107297e+00 6.15266323e-01 5.91078252e-02
-1.41671467e+00 -9.20296013e-02 -2.84228027e-01 3.26090306e-01
-1.27549827e+00 7.32678641e-03 7.94735014e-01 -4.07008499e-01
3.22795749e-01 4.68913317e-01 7.22900629e-01 2.71406949e-01
3.78439367e-01 5.45057893e-01 1.24203777e+00 -5.95651209e-01
3.79964113e-01 1.03172608e-01 2.44068667e-01 6.70292318e-01
3.98973495e-01 -1.08254543e-02 4.61719036e-01 4.07120958e-02
6.97705686e-01 2.54458904e-01 -1.05077788e-01 -5.33582091e-01
-7.17707157e-01 9.13512111e-01 6.67575538e-01 6.01608038e-01
-5.89790523e-01 -4.77613509e-01 2.81981289e-01 1.73207626e-01
1.42599002e-01 4.69351053e-01 -7.90803730e-01 1.70069799e-01
-7.88484395e-01 -3.90648514e-01 7.35618591e-01 5.41182220e-01
7.80962527e-01 2.69184038e-02 1.22759342e-02 8.36294591e-01
2.99391657e-01 5.86525679e-01 7.15742350e-01 -9.86698568e-01
-4.44300950e-01 1.01270175e+00 -5.94822355e-02 -1.21540976e+00
-5.02993524e-01 1.59203187e-01 -1.19357610e+00 5.39984405e-01
4.63041186e-01 9.12431702e-02 -9.61474717e-01 1.16916537e+00
1.94444373e-01 -5.32715023e-01 1.57464612e-02 7.62971103e-01
9.61858273e-01 6.52678847e-01 2.45176911e-01 -3.09929550e-01
1.56068265e+00 -3.73365104e-01 -6.69650555e-01 5.08533567e-02
6.36818647e-01 -9.60647941e-01 1.12109447e+00 3.71052265e-01
-7.91670859e-01 -6.11353576e-01 -9.47731793e-01 8.59134793e-02
-8.44831467e-01 9.28973854e-01 9.51586306e-01 8.47253382e-01
-7.13493526e-01 8.47654819e-01 -8.41989279e-01 -6.89169347e-01
5.94974458e-01 3.78401935e-01 -6.62574589e-01 7.75511041e-02
-4.20173943e-01 5.96849799e-01 5.99852800e-01 2.94057429e-01
-6.24435842e-01 -2.56339669e-01 -6.21333838e-01 3.60255241e-01
2.94083655e-01 -1.49577484e-01 6.53090179e-01 -1.16680193e+00
-1.84667516e+00 1.14956582e+00 4.14921455e-02 -2.96912994e-02
-1.29336253e-01 2.23812759e-01 -1.48525000e-01 1.13977499e-01
-9.97605622e-02 3.43809664e-01 8.50136340e-01 -1.30490255e+00
-5.70331693e-01 -7.56428957e-01 -1.29617274e-01 -2.48274490e-01
-6.95986509e-01 -3.71934585e-02 1.37635559e-01 -4.34267759e-01
6.49647653e-01 -8.74101818e-01 -3.51847000e-02 6.26313746e-01
-4.56312388e-01 -1.17680319e-01 1.52496147e+00 -8.39822531e-01
7.26320386e-01 -2.22129941e+00 1.08997874e-01 7.85315260e-02
2.55844891e-01 6.88956141e-01 -2.29093879e-01 1.67279765e-01
-2.75321573e-01 3.24073732e-01 -3.19740087e-01 4.27957773e-01
-3.07136804e-01 3.53339016e-01 1.01231299e-01 4.47085857e-01
5.42434156e-01 7.55028009e-01 -5.99302113e-01 -5.37783563e-01
5.92166424e-01 3.09125513e-01 2.73807105e-02 1.21788681e-01
2.15889022e-01 2.23136202e-01 -5.84213197e-01 1.08572805e+00
1.08435512e+00 -8.15944001e-02 1.73153713e-01 -6.52029693e-01
-4.57314074e-01 -2.57418275e-01 -1.12820041e+00 9.51195002e-01
-2.71496207e-01 6.15224659e-01 2.49992371e-01 -1.45211768e+00
1.29286194e+00 -3.06499610e-03 7.01010823e-01 -8.43183249e-02
3.47399294e-01 9.98735875e-02 2.59388685e-01 -6.04631066e-01
-1.78839505e-01 3.02373350e-01 4.70634013e-01 3.36735427e-01
4.72155511e-02 -3.48653764e-01 4.11385030e-01 -2.34093234e-01
7.61695445e-01 1.00464731e-01 6.54046178e-01 -4.72624600e-01
8.53794158e-01 8.96013379e-02 4.26249355e-01 1.98612809e-01
-4.93621737e-01 2.93066561e-01 5.97674131e-01 -5.63236892e-01
-8.38276327e-01 -8.79057586e-01 -1.41893670e-01 9.07991052e-01
-9.58293602e-02 5.13148047e-02 -6.64111972e-01 -5.14033675e-01
1.13234267e-01 3.41240108e-01 -6.19509280e-01 -2.66422898e-01
-1.33579224e-01 -1.02356839e+00 3.78060967e-01 6.55678570e-01
8.60049188e-01 -1.21280169e+00 -8.98513794e-01 -7.86802247e-02
2.19229996e-01 -6.48577690e-01 1.83782429e-01 7.47628808e-01
-1.10743845e+00 -1.17704320e+00 -5.79906881e-01 -9.89728451e-01
7.71936595e-01 5.21244407e-01 7.88765013e-01 7.73636922e-02
-9.71992910e-01 2.14253724e-01 -6.07486784e-01 -3.74704897e-01
-2.38216296e-01 4.56078863e-03 -9.86537412e-02 -1.61484286e-01
5.19648373e-01 -5.13734579e-01 -5.05952597e-01 1.26015112e-01
-9.20922458e-01 -3.90147328e-01 8.61302614e-01 9.75550771e-01
7.08692014e-01 5.05697072e-01 6.57171980e-02 -5.65844238e-01
4.34587359e-01 1.44294456e-01 -7.57946134e-01 4.59178746e-01
-3.16958308e-01 -1.57945961e-01 9.55838621e-01 -4.76680458e-01
-8.65318894e-01 4.54064250e-01 1.99906647e-01 3.51491794e-02
-6.87054098e-01 6.69181764e-01 -7.78983474e-01 -4.06819046e-01
8.77438009e-01 4.21066284e-01 2.97858447e-01 -3.76664519e-01
3.22786659e-01 8.63836467e-01 4.37757283e-01 -2.24333003e-01
1.05639422e+00 6.43577933e-01 3.12727720e-01 -1.41244805e+00
-6.13760352e-01 -3.17018330e-01 -1.24001205e+00 -6.41088039e-02
9.04462099e-01 -1.77804455e-01 -1.05304992e+00 6.68325424e-01
-8.64481330e-01 -2.76735872e-02 -4.90456074e-01 4.53487694e-01
-3.72205347e-01 7.13267505e-01 -5.32425523e-01 -8.41172814e-01
-4.46515709e-01 -9.37953293e-01 7.92929173e-01 7.85458446e-01
2.00257316e-01 -9.91714537e-01 -4.35645103e-01 2.59323586e-02
3.05445343e-01 2.74517357e-01 1.18940437e+00 -3.59008342e-01
-6.35339022e-02 -5.40135205e-01 -8.19497228e-01 5.71324587e-01
9.12103593e-01 8.24057400e-01 -9.23725903e-01 -1.35241568e-01
5.37336543e-02 -3.98943752e-01 5.17122209e-01 8.34443212e-01
1.60845697e+00 8.15814659e-02 -3.96437466e-01 8.50293398e-01
1.44713283e+00 5.95613956e-01 6.76149845e-01 2.76627302e-01
6.37942851e-01 6.56023860e-01 6.32129073e-01 4.05730903e-01
-1.27426580e-01 1.35317966e-01 5.88483274e-01 -4.56016004e-01
9.39961001e-02 2.22019106e-01 -1.74127258e-02 5.90617299e-01
-3.46976846e-01 -6.80148602e-02 -7.23888516e-01 3.08399171e-01
-1.42127144e+00 -8.16897511e-01 -3.05359602e-01 2.04865170e+00
5.15845656e-01 -8.59755352e-02 3.81273367e-02 6.89808667e-01
5.56350768e-01 -1.33635253e-01 -6.65004194e-01 -4.82080221e-01
-3.98258865e-01 1.59154668e-01 6.70114994e-01 3.03676259e-02
-1.48786008e+00 1.01499045e+00 5.93220615e+00 5.35558343e-01
-1.46923232e+00 -3.12406570e-01 7.15534449e-01 8.73295069e-01
2.66975701e-01 2.94257671e-01 -4.57049310e-01 1.47701785e-01
3.64473581e-01 6.72262236e-02 4.83339995e-01 1.03582263e+00
2.49085456e-01 -2.37759292e-01 -5.73133647e-01 7.50842392e-01
-1.71265379e-01 -8.11677933e-01 -4.49570306e-02 5.11864312e-02
1.52249321e-01 -5.55398583e-01 -4.71135322e-03 7.25249872e-02
2.11473659e-01 -8.92826319e-01 6.71774745e-02 2.77173311e-01
6.77831352e-01 -4.52918231e-01 6.25460684e-01 2.07457945e-01
-1.55771208e+00 -3.82039040e-01 -9.11161423e-01 3.05176768e-02
-4.17979062e-01 7.04042792e-01 -4.46242541e-01 5.25436819e-01
8.22389364e-01 7.38174617e-01 -7.88775027e-01 9.59658921e-01
-2.19169155e-01 5.43581665e-01 -4.06924814e-01 -5.07044829e-02
2.11359710e-01 -5.51611125e-01 1.06941961e-01 8.38728428e-01
3.60699534e-01 1.07024245e-01 2.94762433e-01 1.00341284e+00
4.96193379e-01 2.84282386e-01 -4.66390848e-01 -7.29225576e-01
2.67505020e-01 1.83044052e+00 -1.53331470e+00 -1.82347104e-01
-2.54453659e-01 1.15609968e+00 -1.89170465e-01 -5.91899082e-03
-3.80076170e-01 -7.71219194e-01 2.47336030e-01 -1.90029144e-01
3.67581904e-01 -3.14088047e-01 -4.94518071e-01 -7.68354595e-01
-3.08695227e-01 -6.37139559e-01 3.29407781e-01 -7.86946237e-01
-1.22927606e+00 3.24903101e-01 -2.44906232e-01 -1.11402011e+00
1.37245685e-01 -1.21069646e+00 -7.24453151e-01 9.61786628e-01
-1.20001233e+00 -1.34091616e+00 -9.23236191e-01 3.30584854e-01
4.79969233e-01 -3.57643574e-01 1.29560471e+00 -7.34930485e-03
-5.69663823e-01 8.87838751e-02 5.44458389e-01 1.49099082e-01
4.02340144e-01 -1.30510497e+00 -2.30748937e-01 5.96528947e-01
-2.22930424e-02 2.11033970e-01 3.45759958e-01 -4.59299803e-01
-1.32919705e+00 -8.42005849e-01 7.50629127e-01 4.06663381e-02
4.88514870e-01 3.98330092e-02 -8.63110065e-01 3.77291590e-01
-2.05705047e-01 1.61257386e-01 8.25985968e-01 -1.69226721e-01
2.00377665e-02 -2.67294258e-01 -1.31705618e+00 4.87392396e-01
3.33429635e-01 -1.33046985e-01 -1.29478395e-01 4.07965004e-01
-5.76733146e-04 9.74305049e-02 -1.10660052e+00 3.57197821e-01
7.64772058e-01 -7.55157292e-01 1.07549691e+00 -5.19013524e-01
3.29943419e-01 -4.88548666e-01 -8.21153373e-02 -1.18322277e+00
-1.05990815e+00 -1.76521167e-01 4.58495229e-01 1.35896885e+00
4.58833389e-02 -4.80346143e-01 6.70817435e-01 3.35877270e-01
2.28641137e-01 -4.70563591e-01 -1.54136106e-01 -6.94032311e-01
6.48951307e-02 2.64616698e-01 4.58679974e-01 9.48453963e-01
-1.78755656e-01 2.76004933e-02 7.39962831e-02 1.13764979e-01
3.79891932e-01 4.67312038e-01 6.13873243e-01 -1.82571077e+00
4.22384739e-02 -5.18806279e-01 -6.36943221e-01 -4.00296360e-01
2.84082047e-03 -6.85183227e-01 -2.05129772e-01 -1.65021527e+00
2.62318492e-01 -3.40139002e-01 2.12572217e-02 7.40506649e-01
-7.99733177e-02 4.58143232e-03 1.77951500e-01 2.63276994e-01
2.08178163e-01 3.26907992e-01 1.18727219e+00 -3.46483916e-01
-3.90469342e-01 2.26830840e-01 -7.15737820e-01 9.24947917e-01
1.08156717e+00 -1.90627962e-01 -3.42013597e-01 -2.36595586e-01
-3.24321657e-01 -3.69319797e-01 3.83828074e-01 -9.42042887e-01
-1.72402620e-01 -4.19221580e-01 9.53101695e-01 -5.86941421e-01
2.84451902e-01 -1.11169577e+00 -1.37928709e-01 9.41379249e-01
5.22295423e-02 -2.34821320e-01 3.23885649e-01 5.91891110e-02
-5.28991856e-02 -7.22731948e-01 1.09813404e+00 -1.65039167e-01
-9.29825366e-01 1.08650744e-01 -7.25645483e-01 -6.06177747e-01
1.26679873e+00 -5.08880317e-01 -2.38808900e-01 -2.19791427e-01
-8.08206081e-01 -2.43827492e-01 2.40910143e-01 1.18267916e-01
3.93241704e-01 -1.10813081e+00 -5.00417054e-01 3.05308789e-01
-2.99764555e-02 -4.30938363e-01 -3.06476810e-04 7.42816865e-01
-1.13762414e+00 4.46229696e-01 -8.95527005e-01 -8.47846925e-01
-1.56446421e+00 6.73363149e-01 1.59031302e-01 -1.25860199e-01
-3.90259922e-01 5.22690356e-01 3.80967148e-02 -3.52434933e-01
1.19294927e-01 -4.76205379e-01 -7.69135058e-01 1.60129085e-01
2.94090837e-01 4.17479455e-01 -7.51676112e-02 -6.24763727e-01
-3.86800051e-01 8.71878624e-01 1.28821105e-01 5.69815338e-01
1.43497860e+00 2.55198535e-02 -4.04474407e-01 3.40202302e-01
8.66373599e-01 -5.24584472e-01 -8.84569526e-01 1.30453885e-01
1.77607104e-01 -5.28778911e-01 2.92138785e-01 -6.85022354e-01
-1.13622594e+00 1.05575359e+00 1.25740361e+00 9.75027680e-01
1.59027660e+00 -4.62770760e-01 3.51533949e-01 4.50825214e-01
-1.17659010e-01 -1.05531383e+00 -3.41079742e-01 3.42255473e-01
6.17515385e-01 -1.44037855e+00 9.30706859e-02 -7.14973807e-01
-3.04645360e-01 1.54584324e+00 5.41801989e-01 -5.72204702e-02
1.05106223e+00 3.41402352e-01 1.33636445e-01 -2.72664949e-02
-3.61739129e-01 -5.82620203e-01 1.28025949e-01 9.78798807e-01
7.71710873e-01 1.61332563e-01 -4.97072935e-01 5.30877829e-01
8.47055241e-02 -2.90803798e-03 2.31459305e-01 1.14562511e+00
-6.95312679e-01 -8.95526946e-01 -6.45953715e-01 5.74479640e-01
-2.32992902e-01 2.45085642e-01 -8.44589770e-01 7.09500551e-01
6.59462512e-02 8.66000473e-01 -8.46430510e-02 -1.84238523e-01
2.24645555e-01 -7.99260437e-02 5.74825168e-01 -3.50120187e-01
-1.96515530e-01 7.36264884e-02 -5.83715618e-01 -1.01050936e-01
-3.85233819e-01 -4.24584568e-01 -9.49421763e-01 -2.59610623e-01
-6.36727810e-01 -1.89666435e-01 9.81945872e-01 9.43287790e-01
1.08708151e-01 5.22591710e-01 7.65792489e-01 -9.94840145e-01
-4.85932976e-01 -9.88953531e-01 -8.93081903e-01 3.75172287e-01
-1.89914048e-01 -7.06502914e-01 -4.14473772e-01 5.68518639e-01] | [9.167213439941406, -1.5549505949020386] |
79944a1d-de27-4f2b-92fa-0b11bc82fd3a | twenty-five-years-of-advances-in-beamforming | 2211.02165 | null | https://arxiv.org/abs/2211.02165v3 | https://arxiv.org/pdf/2211.02165v3.pdf | Twenty-Five Years of Advances in Beamforming: From Convex and Nonconvex Optimization to Learning Techniques | Beamforming is a signal processing technique to steer, shape, and focus an electromagnetic wave using an array of sensors toward a desired direction. It has been used in several engineering applications such as radar, sonar, acoustics, astronomy, seismology, medical imaging, and communications. With the advances in multi-antenna technologies largely for radar and communications, there has been a great interest on beamformer design mostly relying on convex/nonconvex optimization. Recently, machine learning is being leveraged for obtaining attractive solutions to more complex beamforming problems. This article captures the evolution of beamforming in the last twenty-five years from convex-to-nonconvex optimization and optimization-to-learning approaches. It provides a glimpse of this important signal processing technique into a variety of transmit-receive architectures, propagation zones, paths, and conventional/emerging applications. | ['Robert W. Heath Jr', 'Sergiy A. Vorobyov', 'Kumar Vijay Mishra', 'Ahmet M. Elbir'] | 2022-11-03 | null | null | null | null | ['astronomy'] | ['miscellaneous'] | [ 3.58844131e-01 -4.71494406e-01 3.35163385e-01 -3.57033074e-01
-7.09103167e-01 -4.87250984e-01 2.32439965e-01 -3.77425045e-01
-1.09186739e-01 6.49769545e-01 7.31090009e-01 -3.85486543e-01
-6.38313830e-01 -4.28260416e-01 -1.39535651e-01 -1.17778313e+00
-6.72816694e-01 -1.09139897e-01 -4.03454959e-01 -1.32379308e-01
1.81517661e-01 7.55231440e-01 -7.50467837e-01 1.64898798e-01
2.95014739e-01 1.07824719e+00 5.86951450e-02 7.36592650e-01
2.85485238e-01 2.50692308e-01 -4.40302521e-01 -1.79124713e-01
2.08958626e-01 -2.49132127e-01 3.09463054e-01 -2.63708849e-02
1.32287025e-01 7.00156763e-02 -2.15733349e-01 8.24139774e-01
9.41979468e-01 -1.36569560e-01 5.36800981e-01 -8.77459168e-01
-2.10142255e-01 5.74657023e-01 -9.41368163e-01 3.17411423e-01
4.75964814e-01 -2.37512290e-01 6.33350313e-01 -1.24150455e+00
4.47174162e-02 1.10807085e+00 7.96006203e-01 1.04467824e-01
-6.69915378e-01 -7.17142344e-01 -3.34792674e-01 1.52196273e-01
-1.14171863e+00 -5.84653556e-01 9.48803365e-01 -4.37799275e-01
4.17448580e-01 5.36736190e-01 6.69088602e-01 7.18815506e-01
8.00953448e-01 7.41812527e-01 1.13950944e+00 -5.75380087e-01
2.87452877e-01 -2.42770657e-01 -2.70545751e-01 4.76858050e-01
2.91084200e-01 3.53416950e-01 -8.22828948e-01 -2.40298226e-01
6.36303782e-01 -1.29030481e-01 -7.25417078e-01 -2.12986350e-01
-1.31293678e+00 8.24069619e-01 2.57011801e-01 3.29698682e-01
-6.08940363e-01 6.54967874e-02 -6.14188798e-02 2.16960758e-01
4.49848026e-01 8.44361246e-01 -4.34853971e-01 -5.52227274e-02
-1.23306894e+00 1.33673131e-01 8.73928487e-01 5.25507987e-01
3.57106060e-01 9.23871577e-01 4.04569358e-02 8.28861773e-01
7.97643304e-01 1.16179919e+00 -3.35034425e-03 -5.54539204e-01
5.26431441e-01 -1.61602244e-01 1.61511719e-01 -1.33988357e+00
-9.70083654e-01 -1.34381640e+00 -1.30178738e+00 8.88982490e-02
-1.08834073e-01 -9.83518720e-01 -6.52651131e-01 1.32296133e+00
2.77366936e-01 4.25210387e-01 -7.96104595e-02 1.04617643e+00
7.00250030e-01 8.22069466e-01 -5.60050189e-01 -5.68956375e-01
1.03800046e+00 -5.83893895e-01 -6.38725936e-01 -5.17155647e-01
7.03174993e-02 -1.11878479e+00 9.05494988e-02 8.50849330e-01
-1.16419494e+00 -1.46713868e-01 -9.60349977e-01 7.42710412e-01
2.19721749e-01 -1.39396921e-01 6.19728506e-01 1.02705514e+00
-9.13122535e-01 -1.27185628e-01 -7.36232221e-01 5.39093465e-02
5.16467333e-01 1.96235016e-01 -1.82927605e-02 -3.86381894e-01
-8.10174584e-01 7.40859449e-01 -3.88504565e-01 4.09933716e-01
-5.78611076e-01 -1.07832706e+00 -7.10994124e-01 -4.73396778e-02
-1.48262113e-01 -7.08220303e-01 6.58697844e-01 -3.72949630e-01
-1.54100847e+00 2.81479716e-01 -2.12502033e-01 -2.12606460e-01
-6.52477965e-02 -4.93014336e-01 -8.08355331e-01 -2.22095132e-01
-1.17211036e-01 -6.74565434e-02 1.06135690e+00 -1.10049784e+00
-6.21824563e-01 -1.98138952e-01 -6.09833002e-01 8.85244012e-02
-1.22361355e-01 2.72404999e-01 9.62835252e-02 -9.42189336e-01
6.44576788e-01 -6.63296103e-01 -5.06736279e-01 -3.29397231e-01
-4.89864141e-01 4.09726590e-01 6.09656513e-01 -4.02151912e-01
1.16660416e+00 -1.96484959e+00 5.62157452e-01 5.25732636e-01
1.19186984e-02 -3.85645614e-03 -2.93897558e-02 7.76447117e-01
-1.57421246e-01 -4.34973210e-01 -2.35423312e-01 -8.52607265e-02
-3.93621534e-01 -2.78474629e-01 -4.29157108e-01 1.02603865e+00
-1.57556206e-01 7.73188710e-01 -7.89222598e-01 9.84931961e-02
3.06717277e-01 4.12151307e-01 -7.80478239e-01 2.92993218e-01
4.74801213e-01 9.60812926e-01 -6.23350382e-01 1.13634562e+00
7.07505345e-01 8.12746137e-02 -2.18381822e-01 -4.23486710e-01
-5.35483479e-01 -3.40439945e-01 -1.32730997e+00 1.42354012e+00
-6.32794678e-01 8.60968530e-01 1.04315937e+00 -1.35338736e+00
7.58834243e-01 5.14247596e-01 9.56551194e-01 -4.81543303e-01
1.16501413e-01 3.39066297e-01 2.09050760e-01 -9.73019063e-01
-2.40458082e-02 -3.82401943e-01 1.83144026e-03 3.75334412e-01
-9.26056355e-02 -3.11654121e-01 -1.66259363e-01 -4.16734487e-01
1.32318401e+00 -4.76850599e-01 3.54188353e-01 -4.79243040e-01
5.24827063e-01 -1.15077436e-01 4.33392525e-01 6.38397098e-01
2.78296262e-01 6.26636863e-01 -5.66120863e-01 -2.36412585e-01
-5.77941298e-01 -1.19831157e+00 -2.25058258e-01 1.03426325e+00
-9.14071873e-02 2.86600858e-01 -5.44256568e-02 3.69594783e-01
-4.07765135e-02 4.47007596e-01 -5.21560153e-03 1.40488729e-01
-9.74589467e-01 -1.01211739e+00 4.04358923e-01 1.90183043e-01
2.50556916e-01 -7.71724582e-01 -7.95651078e-01 5.14132380e-01
2.12915853e-01 -8.41359079e-01 -2.23153904e-01 2.19118625e-01
-8.06488931e-01 -6.14287734e-01 -9.23618793e-01 -6.59948289e-01
5.61234057e-01 4.87290770e-01 9.32613075e-01 -4.48057383e-01
-4.54237998e-01 8.39647591e-01 -4.47257727e-01 -8.75541985e-01
2.12202951e-01 -3.35705400e-01 4.04493511e-01 3.55793029e-01
-4.35538232e-01 -8.36957276e-01 -8.18243384e-01 2.30965599e-01
-5.40232778e-01 -1.49471909e-01 1.12288344e+00 7.69335270e-01
6.53402582e-02 3.83797772e-02 8.64655972e-01 -4.22234029e-01
7.59108007e-01 -9.49390054e-01 -4.28645432e-01 5.97718544e-02
-1.66392058e-01 -2.86350369e-01 4.52271253e-01 -1.45048961e-01
-1.01409411e+00 -1.96390077e-01 -4.72661793e-01 -2.38723634e-03
-5.23079932e-02 1.20459926e+00 -1.00662589e-01 -5.31735778e-01
6.18202984e-01 3.92246395e-01 -2.46287569e-01 -3.02761823e-01
3.52477640e-01 6.08110487e-01 4.95140046e-01 -1.75080255e-01
1.21596932e+00 5.62346280e-01 5.19778967e-01 -1.39098394e+00
-9.48925614e-01 -8.25516105e-01 -1.71980217e-01 -5.69026589e-01
4.42331284e-01 -5.60933053e-01 -2.80101895e-01 3.38328123e-01
-1.05531847e+00 1.25893131e-01 2.95707166e-01 1.02590513e+00
-1.04760826e-01 -9.55388532e-04 -4.61698957e-02 -9.60194528e-01
-6.24384582e-01 -9.59960938e-01 7.55001485e-01 2.27253363e-01
-1.24600172e-01 -1.36506712e+00 2.73852259e-01 2.21750244e-01
1.04307830e+00 2.09261134e-01 5.56930721e-01 -2.44234294e-01
-4.25154388e-01 -4.01241452e-01 9.44289267e-02 9.72059369e-02
2.86720634e-01 -6.42062962e-01 -6.47283196e-01 -6.26285315e-01
3.75344247e-01 -1.82037428e-02 5.30118525e-01 1.09901440e+00
7.68523633e-01 -4.10341591e-01 -5.47697842e-01 1.02109134e+00
1.58243227e+00 4.01871353e-01 2.72044986e-01 -4.09904458e-02
3.28411311e-01 3.80141318e-01 2.86485881e-01 6.57052577e-01
-6.15872517e-02 3.48007858e-01 5.84679365e-01 -1.87694788e-01
6.40268624e-02 5.37428856e-01 1.87705636e-01 9.92306530e-01
-3.54003124e-02 -5.89230835e-01 -6.42836213e-01 4.31160092e-01
-1.32050467e+00 -1.02573025e+00 -5.95926978e-02 1.92075098e+00
4.09311086e-01 -1.58842146e-01 -4.71926093e-01 9.88610014e-02
3.79318595e-01 1.25445634e-01 -2.79748797e-01 -1.96805298e-01
-1.14846617e-01 4.60045338e-01 5.86454451e-01 5.06165087e-01
-1.02146327e+00 1.08351380e-01 6.85327864e+00 4.70705569e-01
-1.65575922e+00 -9.73361060e-02 1.17656440e-01 -9.26120132e-02
-5.11011660e-01 -5.97033203e-01 -5.44463277e-01 2.40163699e-01
3.99585485e-01 -4.30894876e-03 4.08528149e-01 4.04787868e-01
6.10931933e-01 -1.10814460e-01 -8.90895009e-01 1.23618782e+00
2.59049386e-01 -1.49266732e+00 -3.71354699e-01 -1.51394919e-01
8.38652670e-01 1.82526901e-01 3.22374195e-01 -8.60490277e-02
-1.25930294e-01 -1.18983769e+00 5.53122640e-01 6.40895486e-01
3.61864448e-01 -3.83813113e-01 7.06569672e-01 4.49138671e-01
-9.35322285e-01 -2.89249450e-01 2.86431126e-02 -2.68043071e-01
8.50236833e-01 1.29105687e+00 -8.84355307e-01 6.39847815e-01
5.04548311e-01 4.37771618e-01 2.86196887e-01 1.87808049e+00
-1.16020404e-01 1.26340735e+00 -4.73913282e-01 -2.09121794e-01
3.08831811e-01 -4.48401302e-01 1.27416277e+00 1.45821905e+00
8.42242241e-01 5.26563585e-01 3.19471747e-01 2.75436759e-01
2.01413274e-01 -1.33576140e-01 -3.92610967e-01 4.22692358e-01
6.46342695e-01 1.52877378e+00 -4.41936821e-01 3.83950830e-01
-6.33903742e-01 1.66067690e-01 -5.12911141e-01 6.21419251e-01
-6.69778585e-01 -4.10599738e-01 5.48698008e-01 1.30510449e-01
1.47971064e-01 -8.00388992e-01 -7.29890049e-01 -7.03956366e-01
-3.67654324e-01 -6.99300528e-01 7.17527196e-02 -6.92586780e-01
-1.13246703e+00 3.90755534e-01 1.20385379e-01 -1.17075372e+00
-1.65113971e-01 -5.85980892e-01 -9.28937972e-01 8.53580773e-01
-1.54931927e+00 -1.01907957e+00 -1.61380991e-01 2.65926301e-01
5.74123561e-01 -7.20366776e-01 5.57152748e-01 5.58651149e-01
-2.25617886e-01 1.19239651e-01 4.80316967e-01 -2.39709169e-02
6.16336286e-01 -8.68396401e-01 -2.63782114e-01 9.51963484e-01
2.02123076e-01 4.90648538e-01 1.33063579e+00 -1.76585793e-01
-2.03587341e+00 -1.02374351e+00 2.77019978e-01 1.04924053e-01
6.23858631e-01 -1.07981429e-01 -2.34062061e-01 3.19326848e-01
5.77299178e-01 -1.16189130e-01 9.42389250e-01 2.24908069e-02
2.84153491e-01 -4.64069039e-01 -9.08270955e-01 4.43511248e-01
4.51562166e-01 3.92808527e-01 -4.25883561e-01 3.67368788e-01
-7.92720169e-02 -6.34327054e-01 -6.69109225e-01 5.92213750e-01
6.30673051e-01 -7.92278469e-01 1.17278147e+00 -2.51687944e-01
1.94826454e-01 -5.53808361e-02 -3.16388726e-01 -1.78738356e+00
-7.55867064e-01 -9.64253843e-01 8.10325798e-03 6.72777593e-01
5.12045324e-01 -6.02170646e-01 8.53802443e-01 -3.06567669e-01
-7.53957629e-01 -1.12467694e+00 -1.20955956e+00 -3.02739322e-01
-1.54920071e-01 -6.64026737e-01 3.34089026e-02 6.59803629e-01
-7.18903542e-02 1.92649350e-01 -6.27781272e-01 6.66852474e-01
1.14679217e+00 2.78602064e-01 4.68785912e-01 -9.84821856e-01
-4.40218240e-01 -5.46122313e-01 -3.10182363e-01 -1.53321350e+00
-3.85564983e-01 -7.73109138e-01 2.50275046e-01 -1.72793591e+00
-3.19770366e-01 -6.18976772e-01 -2.04510823e-01 -1.10011213e-02
1.88475937e-01 4.69381630e-01 -3.11425894e-01 -1.68956518e-02
-2.30490506e-01 2.44990721e-01 1.24091089e+00 -2.21301973e-01
-1.35101646e-01 8.23461175e-01 -7.36713767e-01 8.50132883e-01
7.20654666e-01 -2.10500330e-01 -1.08654208e-01 -9.54748333e-01
6.55557811e-01 2.77039379e-01 1.14339553e-01 -1.08945501e+00
7.91422009e-01 -2.91330963e-01 4.16118056e-01 -6.11908019e-01
4.79249120e-01 -8.55994344e-01 1.66684762e-01 5.65647006e-01
4.46727537e-02 -6.98442683e-02 8.19789171e-02 5.18330455e-01
-3.30105036e-01 4.73192967e-02 8.65287721e-01 6.79118857e-02
-5.87511420e-01 2.05053896e-01 -6.42784953e-01 -3.40529974e-03
8.31962824e-01 -1.82406202e-01 1.87102273e-01 -6.56564295e-01
-5.57686806e-01 3.43035072e-01 -6.29447341e-01 9.01233405e-02
1.01418626e+00 -1.04253066e+00 -1.28917360e+00 3.30970019e-01
-4.03847218e-01 -2.63304025e-01 4.56572585e-02 1.09525931e+00
-5.28178990e-01 8.33342373e-01 5.65541349e-02 -7.47284710e-01
-1.18419838e+00 -1.63842246e-01 2.52657294e-01 1.12237565e-01
-2.01270029e-01 1.26498961e+00 1.49292601e-02 -1.01508737e-01
1.18453033e-01 4.83501732e-04 -2.61185199e-01 -2.69679904e-01
4.63129312e-01 7.92042673e-01 1.24795668e-01 -3.58848512e-01
-5.46675026e-01 9.45381939e-01 5.42310119e-01 -2.42285073e-01
1.71461105e+00 -1.29130319e-01 -3.13008040e-01 2.49227006e-02
9.16192234e-01 7.87549317e-01 -8.00761878e-01 -6.15962520e-02
-7.46904984e-02 -3.20407748e-01 4.36813623e-01 -8.18019867e-01
-1.16521204e+00 7.79873490e-01 3.59677672e-01 4.01639938e-01
1.31340599e+00 1.13128707e-01 5.53486109e-01 4.02659357e-01
7.56041884e-01 -6.21873617e-01 2.56925255e-01 5.05388021e-01
1.23944247e+00 -8.79454792e-01 4.32923168e-01 -2.74620950e-01
-3.68814208e-02 1.25782752e+00 -1.49729505e-01 -1.27326027e-01
1.40718138e+00 1.00162148e+00 2.06720740e-01 -2.00138107e-01
-3.72575730e-01 3.81424665e-01 6.16840482e-01 6.54110372e-01
6.85779333e-01 -3.24555822e-02 -7.40154311e-02 4.22649264e-01
-3.34984839e-01 -2.92964906e-01 2.77811795e-01 9.06469166e-01
-1.00855613e+00 -8.90227556e-01 -9.18844104e-01 7.63637245e-01
-5.77962458e-01 -2.30409950e-01 3.04487079e-01 2.29035899e-01
3.05130221e-02 1.35047102e+00 -2.45059907e-01 -1.14776514e-01
2.77859092e-01 -4.95530248e-01 5.68300426e-01 -7.11406291e-01
-2.39926383e-01 4.61369604e-01 1.87608134e-03 -3.23997676e-01
-5.77075362e-01 -7.89016485e-01 -7.74200022e-01 1.80688381e-01
-5.53847730e-01 3.52000862e-01 7.96184182e-01 1.13826680e+00
2.46937394e-01 7.54056871e-01 9.75636542e-01 -1.08936870e+00
-6.03716731e-01 -8.86192024e-01 -5.80356896e-01 -4.96222615e-01
5.43660760e-01 -5.03416717e-01 -3.77121329e-01 -1.91935718e-01] | [6.489408016204834, 1.340433955192566] |
9c555204-f079-46d4-9e8e-7b0b7197b8bb | ai-ml-nit-patna-trac-2-deep-learning-approach | null | null | https://aclanthology.org/2020.trac-1.18 | https://aclanthology.org/2020.trac-1.18.pdf | AI\_ML\_NIT\_Patna @ TRAC - 2: Deep Learning Approach for Multi-lingual Aggression Identification | This paper describes the details of developed models and results of team AI{\_}ML{\_}NIT{\_}Patna for the shared task of TRAC - 2. The main objective of the said task is to identify the level of aggression and whether the comment is gendered based or not. The aggression level of each comment can be marked as either Overtly aggressive or Covertly aggressive or Non-aggressive. We have proposed two deep learning systems: Convolutional Neural Network and Long Short Term Memory with two different input text representations, FastText and One-hot embeddings. We have found that the LSTM model with FastText embedding is performing better than other models for Hindi and Bangla datasets but for the English dataset, the CNN model with FastText embedding has performed better. We have also found that the performances of One-hot embedding and pre-trained FastText embedding are comparable. Our system got 11th and 10th positions for English Sub-task A and Sub-task B, respectively, 8th and 7th positions, respectively for Hindi Sub-task A and Sub-task B and 7th and 6th positions for Bangla Sub-task A and Sub-task B, respectively among the total submitted systems. | ['Kirti Kumari', 'Jyoti Prakash Singh'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['aggression-identification'] | ['natural-language-processing'] | [-5.96931040e-01 1.25677273e-01 2.42452845e-01 -5.34504950e-01
-4.53966022e-01 -1.61135003e-01 7.69954026e-01 2.39517227e-01
-1.01497030e+00 6.60892069e-01 3.86638612e-01 -1.57743514e-01
-3.52402478e-01 -5.80491304e-01 -8.79193619e-02 -6.15306318e-01
-1.10420994e-01 9.17140901e-01 1.42119627e-03 -6.35438979e-01
5.94262481e-01 2.08253920e-01 -1.14845979e+00 6.57870889e-01
3.01447898e-01 6.77160203e-01 -4.38962132e-01 1.01379287e+00
-6.18839860e-02 1.04523444e+00 -7.76657879e-01 -1.10614681e+00
1.67582110e-01 -2.39576653e-01 -1.31010306e+00 -6.87984169e-01
8.72907817e-01 -4.86152232e-01 -6.02927983e-01 8.66440296e-01
8.24183404e-01 1.52187586e-01 9.78330135e-01 -1.35003972e+00
-4.84921485e-01 1.04345083e+00 -5.68670094e-01 6.30877495e-01
1.46242063e-02 9.61893499e-02 1.12342322e+00 -5.89283049e-01
3.89865488e-01 1.34696364e+00 9.03337717e-01 8.20061803e-01
-7.35435009e-01 -1.00117719e+00 -1.94607556e-01 5.14849544e-01
-1.10737813e+00 -2.76535600e-01 5.82612157e-01 -7.31032491e-01
1.12057996e+00 2.69689798e-01 4.22506571e-01 1.68499947e+00
3.15252662e-01 7.61391819e-01 9.13930118e-01 1.76418856e-01
-1.90099224e-01 2.90255040e-01 6.64996147e-01 3.81226599e-01
-1.74620509e-01 1.85759533e-02 -5.43453574e-01 -2.40820602e-01
2.67779678e-01 -3.44776183e-01 4.09461558e-01 6.87658727e-01
-7.55909622e-01 1.31011736e+00 1.87990844e-01 7.21525431e-01
-1.76243633e-01 1.77318320e-01 1.12881243e+00 3.42735052e-01
4.39542562e-01 5.01038432e-01 -3.19068491e-01 -8.72100949e-01
-1.00580025e+00 7.01478362e-01 8.26223195e-01 4.67465967e-01
2.58926988e-01 1.58563092e-01 -4.99664724e-01 1.43616414e+00
-4.31510992e-02 1.48667797e-01 7.83625245e-01 -6.93797469e-01
7.23957956e-01 5.32010078e-01 -1.40659779e-01 -1.08747089e+00
-9.56186354e-01 -3.16040009e-01 -6.03223741e-01 3.06171268e-01
5.56628108e-01 -3.60536128e-01 -6.55587554e-01 1.75108576e+00
-1.10767201e-01 -2.91877687e-01 -3.46536376e-02 7.32256114e-01
1.08280826e+00 8.54398191e-01 1.46085680e-01 1.29575625e-01
1.12819314e+00 -1.22655582e+00 -6.25797451e-01 -1.38402387e-01
9.78248417e-01 -7.35016167e-01 7.97424018e-01 3.70918304e-01
-1.35681176e+00 -5.19222081e-01 -8.41816247e-01 -3.31785440e-01
-8.09537888e-01 3.49749476e-02 2.60151029e-01 7.16262281e-01
-1.01974845e+00 4.44716305e-01 -2.96279967e-01 -4.26643997e-01
1.23292916e-01 5.99227846e-01 -6.38907969e-01 3.67328286e-01
-1.26675296e+00 1.16561282e+00 3.62480521e-01 6.49809092e-02
-7.65556276e-01 -5.55400908e-01 -6.24555767e-01 -7.50908554e-02
-2.24951878e-01 2.20151786e-02 1.19099057e+00 -6.54696405e-01
-1.43274295e+00 1.24665439e+00 3.20208162e-01 -4.60502028e-01
6.21004403e-01 -3.83021772e-01 -5.92988491e-01 -2.73665041e-01
3.00541162e-01 5.99898577e-01 5.48499048e-01 -6.94115281e-01
-8.27391028e-01 -7.97563672e-01 9.50232819e-02 2.05778733e-01
-6.78289354e-01 6.09111547e-01 7.13908300e-02 -4.97554630e-01
-5.23019016e-01 -9.79631782e-01 1.94966435e-01 -4.46472615e-01
-6.96517348e-01 -7.80779600e-01 1.18331456e+00 -8.09920967e-01
1.60743868e+00 -2.25844884e+00 3.25100660e-01 -1.83383167e-01
3.57996732e-01 7.68795967e-01 4.18939516e-02 7.27907896e-01
-3.86708528e-01 2.31020242e-01 9.82401147e-02 -4.28932071e-01
9.81385931e-02 4.42788303e-01 -4.51692902e-02 4.54909235e-01
-2.45790139e-01 3.38967562e-01 -5.33513308e-01 -3.47510457e-01
-1.14577701e-02 1.69192001e-01 -5.22701740e-01 2.74923861e-01
3.89608443e-01 1.69292420e-01 -3.96800227e-02 9.58727449e-02
5.61839283e-01 6.21052563e-01 -4.61876422e-01 9.48956758e-02
-4.08076108e-01 3.71064723e-01 -8.81201208e-01 1.06473982e+00
-4.86496538e-01 9.00419354e-01 2.78123707e-01 -7.87026584e-01
1.03499997e+00 2.88401812e-01 1.73828334e-01 -6.76599681e-01
5.20069957e-01 3.29819709e-01 5.99897385e-01 -8.92192125e-01
8.57019842e-01 -2.41941378e-01 -3.61369044e-01 6.36093974e-01
2.18498111e-01 1.38356119e-01 3.89142573e-01 2.67057210e-01
1.25207520e+00 -4.84765351e-01 -1.91484079e-01 -2.77323872e-01
5.64185739e-01 -4.31452781e-01 2.00876758e-01 7.46499538e-01
-4.67527837e-01 5.82940161e-01 1.08257556e+00 -8.47633541e-01
-1.13390803e+00 -6.15468442e-01 -2.76753098e-01 1.86783075e+00
-4.65213090e-01 -7.18954146e-01 -6.20564520e-01 -8.13705802e-01
-7.84394592e-02 9.96505499e-01 -1.39221931e+00 -2.99765170e-01
-7.82031417e-01 -5.70869267e-01 1.26208591e+00 6.28480315e-01
5.33227324e-01 -1.22928047e+00 -5.24577856e-01 1.26746714e-01
-5.08036949e-02 -7.21295357e-01 -9.98670012e-02 4.17621881e-01
-1.26558006e-01 -7.15088606e-01 -6.51994705e-01 -5.88738501e-01
-4.12845239e-02 -6.16578341e-01 7.64509678e-01 8.57730806e-02
-2.07666472e-01 -3.76201898e-01 -4.90207881e-01 -4.70042378e-01
-3.31463546e-01 3.83905679e-01 1.43510625e-01 -4.18246239e-02
4.62031901e-01 -3.92829001e-01 -3.06121737e-01 3.44213575e-01
-5.52325010e-01 -5.61107360e-02 2.49562606e-01 8.36103678e-01
-4.87162977e-01 -2.95327663e-01 3.65448207e-01 -1.07611418e+00
1.15378904e+00 -7.30034769e-01 1.83613449e-01 -4.39692765e-01
-6.87886998e-02 -3.40950787e-01 6.20396197e-01 -2.31680438e-01
-7.01078176e-01 -5.98233759e-01 -9.49761987e-01 -1.20102823e-01
-3.76430869e-01 3.31271380e-01 3.75250369e-01 4.09771055e-01
7.35266507e-01 -1.41459882e-01 -2.50175059e-01 -4.29280907e-01
-6.32499158e-02 1.23612678e+00 3.74965578e-01 -3.80892903e-01
5.86755872e-01 6.56914562e-02 -2.60186076e-01 -9.90074694e-01
-8.55905592e-01 -4.16220874e-01 -7.53406405e-01 -3.98307025e-01
1.38812435e+00 -4.65621948e-01 -8.00122857e-01 8.86791348e-01
-1.06831062e+00 -4.93299603e-01 8.40904340e-02 3.06107134e-01
-3.99902463e-01 -1.30911365e-01 -9.46773291e-01 -8.22669923e-01
-5.41685164e-01 -1.22383559e+00 9.54978406e-01 9.38532948e-02
-7.76706517e-01 -1.02551222e+00 5.46637952e-01 6.26103997e-01
4.65710014e-01 8.52081254e-02 1.32750988e+00 -1.55908000e+00
8.96462023e-01 -4.67263371e-01 -2.64649153e-01 5.21029472e-01
-3.91249120e-01 3.36070746e-01 -1.03750944e+00 -1.73439711e-01
-2.54378349e-01 -7.83486307e-01 8.32115710e-01 3.93295102e-02
1.00615120e+00 -4.02015269e-01 1.30997628e-01 3.04326028e-01
7.33631790e-01 2.32493222e-01 7.42253244e-01 5.17273903e-01
8.11666250e-01 7.22962856e-01 3.30132484e-01 5.37746370e-01
4.84982699e-01 9.36035395e-01 3.25401336e-01 1.62527829e-01
1.76148564e-02 -5.14020473e-02 5.84929466e-01 7.38099515e-01
-1.73746571e-01 -2.54906207e-01 -1.17960560e+00 6.49132907e-01
-1.78511727e+00 -1.22705650e+00 -6.90855622e-01 1.78099096e+00
6.89016521e-01 2.16464728e-01 8.58876467e-01 4.24449950e-01
4.98699158e-01 2.97233790e-01 -1.07647590e-01 -1.76746023e+00
2.08674327e-01 1.32395208e-01 8.40822682e-02 4.28941756e-01
-1.09795737e+00 1.00372982e+00 6.13618708e+00 9.54713225e-01
-1.30850720e+00 2.84237802e-01 6.38866901e-01 -4.37566817e-01
4.12772894e-01 -5.75618088e-01 -9.79989648e-01 7.66962111e-01
1.55705369e+00 3.86874191e-02 -1.36445567e-01 8.52579534e-01
-1.09809436e-01 -2.57574897e-02 -1.26112378e+00 8.80143404e-01
3.08282435e-01 -9.06438231e-01 9.43471417e-02 2.86031693e-01
3.66926908e-01 2.21463695e-01 3.86937618e-01 9.83718514e-01
3.48473728e-01 -1.42255723e+00 7.10609436e-01 1.25475958e-01
4.84182894e-01 -1.03625035e+00 1.32278633e+00 3.98925155e-01
-5.44462204e-01 -6.00844860e-01 -4.53015536e-01 -4.68532115e-01
-1.84037596e-01 -1.07224956e-01 -5.53912103e-01 2.86760211e-01
1.12198257e+00 5.98614931e-01 -6.86919570e-01 8.32571208e-01
-2.10001722e-01 8.82046521e-01 -1.70363724e-01 -2.56686866e-01
9.11713362e-01 -7.20556974e-02 7.13017523e-01 1.54843616e+00
-1.23300150e-01 -9.02487561e-02 -2.64119077e-02 3.36090446e-01
3.41167673e-02 3.44994485e-01 -5.73541105e-01 -6.27276450e-02
-1.10992789e-01 1.38027215e+00 -4.00510691e-02 -2.07150221e-01
-1.75754562e-01 7.75578618e-01 7.25822508e-01 -6.81670457e-02
-9.61067617e-01 -7.10578084e-01 6.10351384e-01 1.63168892e-01
-7.66655058e-02 4.23134528e-02 -4.04278576e-01 -5.37475705e-01
-4.77818757e-01 -7.49656260e-01 7.75506437e-01 -6.40724182e-01
-1.44016612e+00 9.41588521e-01 2.00558960e-01 -6.90077841e-01
-4.35532212e-01 -8.96368086e-01 -1.08906949e+00 7.93660879e-01
-5.45304656e-01 -1.46571207e+00 1.22773675e-02 6.77754819e-01
7.34607756e-01 -5.46923459e-01 8.30174208e-01 5.42669117e-01
-9.58701611e-01 1.04952681e+00 1.27610698e-01 6.37875915e-01
7.75613248e-01 -1.29749811e+00 -1.18318960e-01 2.35375643e-01
-1.82951942e-01 3.28579754e-01 9.03837442e-01 -2.98213273e-01
-5.94692707e-01 -9.53336000e-01 1.06527913e+00 -5.01698136e-01
1.03452587e+00 -5.56741655e-01 -8.02280366e-01 7.82565236e-01
5.19920468e-01 -6.21955276e-01 9.86485779e-01 6.44090056e-01
-3.72039497e-01 6.61102459e-02 -1.19816411e+00 5.92522740e-01
6.95252299e-01 -3.63996863e-01 -7.68984199e-01 5.01447737e-01
4.62897085e-02 -4.03858870e-01 -7.96646953e-01 7.35603943e-02
5.95296800e-01 -1.31035435e+00 3.95164132e-01 -1.01424086e+00
1.22906506e+00 4.34286028e-01 -2.24766694e-02 -1.41408825e+00
-5.52216947e-01 -2.67981231e-01 4.96772051e-01 1.16159630e+00
6.96323276e-01 -4.68173951e-01 6.12872601e-01 5.67703485e-01
-3.81552339e-01 -7.46495664e-01 -1.28474844e+00 -5.36629736e-01
9.70645964e-01 -3.37966144e-01 -7.72168264e-02 1.05410016e+00
1.42533585e-01 6.57925308e-01 -7.44240940e-01 -3.01486999e-01
8.70404169e-02 -3.90991956e-01 1.10465467e+00 -1.10513723e+00
-7.63341133e-03 -8.08862269e-01 -8.91791523e-01 -2.58670807e-01
3.65962654e-01 -9.54343557e-01 -4.33662653e-01 -1.48827982e+00
3.44054967e-01 -1.37237310e-01 -1.86679065e-01 7.49266386e-01
1.23883978e-01 5.78409493e-01 8.32538605e-02 -1.60465732e-01
-2.85182983e-01 3.42491150e-01 6.94352090e-01 -1.95251051e-02
7.79704452e-02 9.88349039e-03 -6.07722044e-01 9.09370482e-01
8.93571258e-01 -6.93356872e-01 -4.85543981e-02 -4.98643517e-01
3.82846206e-01 -2.87012637e-01 1.04124777e-01 -1.08611727e+00
1.70037434e-01 -5.21416776e-02 3.07021707e-01 -7.04672992e-01
7.58454084e-01 -2.98791021e-01 -4.69114602e-01 4.09622431e-01
-7.69410789e-01 2.35133529e-01 2.93893069e-01 -1.14496918e-02
-2.60702640e-01 -6.57007456e-01 9.40918505e-01 1.15084499e-02
-3.14971954e-01 2.39461079e-01 -8.88216019e-01 1.77577794e-01
1.09809589e+00 -2.78509021e-01 -4.63788182e-01 -5.99913955e-01
-9.23419297e-01 3.22026342e-01 -1.39000237e-01 6.49446189e-01
3.93807948e-01 -1.06131029e+00 -1.21138489e+00 8.46917927e-02
1.38514757e-01 -7.64662325e-01 4.28809494e-01 9.77018774e-01
-5.26705980e-01 2.88843989e-01 -8.58767033e-01 -1.33043781e-01
-1.65373790e+00 4.27095592e-02 2.67447591e-01 -3.78959537e-01
-2.84933388e-01 1.30124903e+00 -2.05212552e-02 -7.71785796e-01
2.03568563e-01 3.45534801e-01 -5.88969529e-01 4.06481951e-01
7.02856898e-01 7.97689021e-01 7.19915330e-02 -1.41286504e+00
-3.20706308e-01 3.13868642e-01 -8.17527354e-01 -2.07256660e-01
1.57624829e+00 3.61587048e-01 -4.13356334e-01 7.92237759e-01
1.60546446e+00 5.12535200e-02 -2.44477347e-01 3.48239601e-01
-1.75322548e-01 -4.56314147e-01 2.09966980e-04 -1.03548670e+00
-1.06085718e+00 1.37023318e+00 5.34507036e-01 4.91071790e-01
3.83880675e-01 -5.75556122e-02 7.98803627e-01 4.26470876e-01
-1.76098675e-01 -1.51465750e+00 2.53703505e-01 1.11509860e+00
1.29035556e+00 -8.79829228e-01 -1.86322421e-01 3.41316938e-01
-9.84001696e-01 1.28820753e+00 1.23907268e+00 -2.59725630e-01
4.94647890e-01 8.58711973e-02 2.62907326e-01 -5.56998014e-01
-7.62491286e-01 9.75295976e-02 3.90978977e-02 4.66307670e-01
5.06627798e-01 1.14324920e-01 -6.37865543e-01 7.87687719e-01
-9.92504895e-01 -5.76395690e-01 6.24768496e-01 5.66173196e-01
-2.00512424e-01 -8.91061127e-01 -4.97741625e-02 7.21330225e-01
-8.13848794e-01 1.18816443e-01 -9.50186908e-01 1.09916127e+00
4.03724730e-01 9.02455091e-01 2.96546459e-01 -9.94199872e-01
3.06953967e-01 2.10112363e-01 1.09632894e-01 -6.04323626e-01
-1.39817417e+00 -3.28236908e-01 7.36688256e-01 -3.22686017e-01
2.21310169e-01 -5.18193305e-01 -9.85082030e-01 -7.73809075e-01
-1.84136853e-01 2.09582284e-01 6.65025353e-01 9.16255295e-01
-1.49295837e-01 1.99953824e-01 3.52171421e-01 -7.79877543e-01
-5.74425399e-01 -1.58236206e+00 -8.16960156e-01 6.62223995e-01
-1.30219636e-02 -7.74134576e-01 -4.31321263e-01 -5.81395566e-01] | [8.810683250427246, 10.756573677062988] |
b263c0ba-0e2b-4dd6-ba9f-3ec438b4decf | antm-an-aligned-neural-topic-model-for | 2302.01501 | null | https://arxiv.org/abs/2302.01501v2 | https://arxiv.org/pdf/2302.01501v2.pdf | ANTM: An Aligned Neural Topic Model for Exploring Evolving Topics | This paper presents an algorithmic family of dynamic topic models called Aligned Neural Topic Models (ANTM), which combine novel data mining algorithms to provide a modular framework for discovering evolving topics. ANTM maintains the temporal continuity of evolving topics by extracting time-aware features from documents using advanced pre-trained Large Language Models (LLMs) and employing an overlapping sliding window algorithm for sequential document clustering. This overlapping sliding window algorithm identifies a different number of topics within each time frame and aligns semantically similar document clusters across time periods. This process captures emerging and fading trends across different periods and allows for a more interpretable representation of evolving topics. Experiments on four distinct datasets show that ANTM outperforms probabilistic dynamic topic models in terms of topic coherence and diversity metrics. Moreover, it improves the scalability and flexibility of dynamic topic models by being accessible and adaptable to different types of algorithms. Additionally, a Python package is developed for researchers and scientists who wish to study the trends and evolving patterns of topics in large-scale textual data. | ['Bernd Amann', 'Camelia Constantin', 'Hubert Naacke', 'Hamed Rahimi'] | 2023-02-03 | null | null | null | null | ['topic-models', 'dynamic-topic-modeling'] | ['natural-language-processing', 'natural-language-processing'] | [-2.55382061e-01 -4.63617817e-02 -5.26949704e-01 -3.82029802e-01
-5.61920524e-01 -6.38470888e-01 1.20069873e+00 6.96791768e-01
-1.80360302e-02 2.93795586e-01 3.80491734e-01 -1.66360825e-01
-4.96859312e-01 -1.01274407e+00 -1.50950521e-01 -6.06194794e-01
-8.86857986e-01 8.86949956e-01 7.36039340e-01 1.25643089e-01
3.51359278e-01 3.82184647e-02 -1.88341630e+00 4.44839209e-01
7.47388721e-01 6.46794915e-01 5.40687740e-01 5.29445052e-01
-8.85223091e-01 7.17653632e-02 -8.75038028e-01 1.68325201e-01
1.19260196e-02 -1.77988723e-01 -7.17127740e-01 1.43782079e-01
1.65159672e-01 2.13509887e-01 -1.14729464e-01 6.50563180e-01
-9.34468731e-02 2.91363448e-01 6.19637430e-01 -1.57552028e+00
-1.07522927e-01 9.60751712e-01 -6.00523233e-01 6.21915817e-01
2.02782273e-01 -5.09066761e-01 1.03440487e+00 -6.65475786e-01
1.03740168e+00 1.49831641e+00 7.36813724e-01 -1.35819949e-02
-1.14713132e+00 -6.74145460e-01 7.07684696e-01 2.55610585e-01
-1.27009499e+00 1.28617272e-01 8.52777421e-01 -6.13106072e-01
1.00556982e+00 4.05049324e-01 8.40484440e-01 9.96169508e-01
5.99243760e-01 7.99048662e-01 7.36856759e-01 -2.15381026e-01
4.63168561e-01 3.35271746e-01 8.14552069e-01 5.86645976e-02
1.43217281e-01 -4.52397168e-01 -9.40713227e-01 -8.65368009e-01
1.73068911e-01 2.77853072e-01 -1.72211342e-02 -4.65965539e-01
-1.24651623e+00 1.07604444e+00 -2.30407059e-01 7.75710225e-01
-6.13501906e-01 -3.11484098e-01 6.07980609e-01 2.98651874e-01
1.26308870e+00 2.04341665e-01 -5.53121030e-01 -3.30799192e-01
-1.60621166e+00 6.10799789e-01 8.59567106e-01 8.64007831e-01
8.31279635e-01 -1.57699466e-01 -8.63295272e-02 8.39766860e-01
3.76896530e-01 2.33672857e-01 1.10904181e+00 -6.07818425e-01
1.95783958e-01 8.13330412e-01 -2.88734194e-02 -1.23870587e+00
-6.33243978e-01 -2.48045787e-01 -4.53619450e-01 -3.23267430e-01
-3.73425521e-02 1.05331704e-01 -8.57112348e-01 1.53326976e+00
6.31052554e-01 2.25476623e-01 -1.85513496e-01 1.58260033e-01
4.70869243e-01 1.44737089e+00 3.89205843e-01 -5.79259634e-01
1.54087162e+00 -5.72278798e-01 -8.84772718e-01 2.07993463e-01
1.78322777e-01 -4.86513138e-01 7.71906912e-01 4.23736125e-01
-9.72162485e-01 -3.78823489e-01 -6.80473208e-01 3.37904781e-01
-9.73650694e-01 -3.70919883e-01 8.32279027e-01 6.41150773e-01
-1.37434697e+00 4.01835322e-01 -1.34327257e+00 -8.79569769e-01
6.66934699e-02 2.18739212e-01 1.95629671e-01 3.87505293e-01
-1.26386917e+00 4.12283599e-01 7.62472212e-01 -5.80003142e-01
-7.90535927e-01 -1.11447895e+00 -7.18289673e-01 1.97723836e-01
1.25301704e-01 -3.93616498e-01 1.09687364e+00 -4.68113273e-01
-1.15546048e+00 6.26621366e-01 -6.77072823e-01 -9.74811733e-01
2.54269332e-01 -2.39283159e-01 -6.51614964e-01 3.56389225e-01
3.68864775e-01 6.52286947e-01 7.80410826e-01 -1.10579360e+00
-1.11515152e+00 -4.85680252e-01 -6.96875870e-01 1.25051215e-01
-9.48639989e-01 3.72326255e-01 -6.78050220e-01 -9.75161672e-01
2.00622186e-01 -7.82887280e-01 -2.57909834e-01 -4.77810800e-01
-2.07311675e-01 -8.89522254e-01 1.78318083e+00 -5.30137122e-01
1.75896370e+00 -1.77450275e+00 5.89815713e-02 4.14625973e-01
3.82201433e-01 -4.02154833e-01 3.34982663e-01 6.19458973e-01
9.19990800e-03 2.14884616e-03 -1.65936068e-01 -4.25181210e-01
-6.93954825e-02 1.12809695e-01 -8.81724715e-01 2.37458289e-01
-2.36905068e-01 5.27278543e-01 -6.62639916e-01 -5.21586537e-01
1.38227135e-01 2.78003901e-01 -1.54445484e-01 6.33467808e-02
-6.66475594e-01 -8.27346072e-02 -4.15086031e-01 4.04520631e-01
5.24762154e-01 -4.27087188e-01 3.07901978e-01 5.94527423e-01
-5.17564058e-01 3.76177818e-01 -1.17783785e+00 1.69056940e+00
-1.73537731e-01 1.10972464e+00 -2.85317361e-01 -9.34852481e-01
1.15516067e+00 5.64286411e-01 1.20470977e+00 -2.60868967e-01
-1.94816813e-01 -2.56805897e-01 -6.01049542e-01 -7.98853785e-02
1.32450211e+00 1.46443754e-01 -2.63698488e-01 1.06203914e+00
2.06105843e-01 1.60180941e-01 3.37634534e-01 4.79749858e-01
8.81047368e-01 -3.65732491e-01 1.64654985e-01 -7.80401826e-01
-4.72381897e-02 4.49174732e-01 4.42576379e-01 6.64023399e-01
1.46368429e-01 3.11824024e-01 1.68014169e-01 -8.93369079e-01
-1.16590846e+00 -1.21835232e+00 -4.30084467e-01 1.55632663e+00
-6.10929318e-02 -9.24236059e-01 -3.73148173e-01 -1.85699552e-01
-1.23128138e-01 7.36385882e-01 -8.77521396e-01 1.60031140e-01
-4.46396261e-01 -1.17101634e+00 -2.75169853e-02 2.01705143e-01
2.35991506e-03 -9.97968376e-01 -9.56584096e-01 5.38653672e-01
-4.49103057e-01 -5.42911470e-01 -6.17073886e-02 2.47521549e-01
-1.13201249e+00 -7.01817095e-01 -9.04261231e-01 -5.53327978e-01
2.62130052e-01 5.91086268e-01 1.08973610e+00 -5.11958063e-01
-2.40915164e-01 4.88386214e-01 -3.08533967e-01 -7.90134251e-01
-2.78189987e-01 5.28720140e-01 1.53856903e-01 -1.65126383e-01
5.92449963e-01 -6.98054433e-01 -3.43224674e-01 3.57131571e-01
-9.01253104e-01 -1.96817309e-01 -1.83793798e-01 3.70923072e-01
4.05274600e-01 7.67784297e-01 3.76649380e-01 -7.28845954e-01
8.23107481e-01 -1.12887490e+00 -5.92152655e-01 2.31179938e-01
-1.09412515e+00 -2.22805008e-01 5.29640876e-02 -8.55931461e-01
-1.28760695e+00 -5.30142844e-01 7.73391902e-01 -2.63729423e-01
-2.24329591e-01 5.24084747e-01 2.00761884e-01 9.35033023e-01
3.96267593e-01 5.94188571e-01 -4.00610715e-01 -5.15825272e-01
4.24750030e-01 7.39896894e-01 5.05258322e-01 -6.37183249e-01
6.00491226e-01 1.02322388e+00 -7.56749928e-01 -1.13169992e+00
-4.19782281e-01 -1.19051468e+00 -5.40845156e-01 -3.39112014e-01
6.15400374e-01 -9.57065225e-01 -6.80534244e-02 6.23594642e-01
-9.71846461e-01 -1.66648641e-01 -4.19614553e-01 2.81979591e-01
-3.86222452e-01 2.37667024e-01 -4.52832520e-01 -7.93689668e-01
-5.75626016e-01 -4.90802497e-01 1.08001077e+00 3.31859142e-01
-9.60126817e-01 -1.45539057e+00 5.68338752e-01 -3.86559784e-01
5.36451340e-01 2.66103029e-01 9.78513837e-01 -9.52069521e-01
-2.92294621e-01 -1.28166720e-01 2.95264006e-01 -6.76089883e-01
1.92517921e-01 2.21973047e-01 -8.73694837e-01 -4.03095484e-01
7.58217275e-02 3.85646194e-01 9.49798763e-01 8.25964987e-01
9.24195826e-01 -4.65101868e-01 -1.10007048e+00 2.09459990e-01
1.03285277e+00 6.01322651e-01 2.94536680e-01 8.81723464e-01
1.68030486e-01 9.10419464e-01 5.01477122e-01 7.13350475e-01
7.63535678e-01 5.46025336e-01 -9.51737687e-02 1.76183090e-01
4.75633621e-01 -6.03651404e-02 3.34237248e-01 9.02413964e-01
2.11427212e-01 -4.49416071e-01 -1.19823217e+00 1.21176183e+00
-1.91645908e+00 -1.22952342e+00 -2.43087217e-01 1.96873617e+00
7.36468971e-01 3.34826797e-01 4.89360839e-01 9.00380015e-02
9.15843964e-01 4.23224092e-01 -4.79428679e-01 -2.85694838e-01
-2.33373463e-01 -1.33089602e-01 -1.59357246e-02 1.42891452e-01
-1.23797452e+00 9.34749663e-01 7.18417788e+00 7.79188693e-01
-9.75669444e-01 2.53968477e-01 5.60609400e-01 -1.57579795e-01
-4.73081440e-01 1.79717820e-02 -1.06665206e+00 5.91607511e-01
1.62829876e+00 -9.37056422e-01 -4.53569531e-01 1.12757111e+00
3.25501770e-01 -1.78258508e-01 -5.65277696e-01 4.40857232e-01
1.24966078e-01 -1.52608573e+00 2.28343651e-01 2.79235393e-01
8.63317132e-01 2.79411465e-01 1.50717199e-01 1.66868240e-01
7.40725935e-01 -3.19223940e-01 6.49699986e-01 4.44216460e-01
8.87756422e-02 -7.23005652e-01 3.25231761e-01 2.66484052e-01
-1.35080063e+00 -1.26966774e-01 -3.95576060e-01 2.51932979e-01
4.01969045e-01 8.03546190e-01 -8.44321787e-01 4.70975369e-01
1.44642627e+00 8.49458575e-01 -4.18224394e-01 1.18497550e+00
5.38398147e-01 8.26343775e-01 -6.74810052e-01 1.23977400e-01
3.62556577e-01 -8.75655338e-02 9.47563052e-01 1.50583017e+00
5.06719410e-01 -3.34674716e-01 2.38985762e-01 5.66205621e-01
4.69575375e-01 1.29250577e-02 -4.75356609e-01 6.08713180e-02
7.34198868e-01 1.20623910e+00 -1.53748465e+00 -6.30057633e-01
-8.00776184e-02 5.43890715e-01 -2.28066728e-01 1.84291556e-01
-4.80488509e-01 -3.09843004e-01 7.29762912e-01 2.30863973e-01
3.57902944e-01 -5.56310058e-01 -3.45104128e-01 -1.07333612e+00
-2.94312268e-01 -4.13611740e-01 1.06333053e+00 -4.89088953e-01
-1.03600490e+00 8.51060688e-01 6.73527598e-01 -1.16317475e+00
-5.33544183e-01 1.02064066e-01 -1.11356783e+00 3.74090403e-01
-1.08556902e+00 -1.02162755e+00 -1.54019937e-01 6.06211901e-01
1.15412402e+00 -2.40414292e-01 7.82619059e-01 -4.14090991e-01
-2.13512614e-01 -7.44635984e-02 4.90879238e-01 -4.29377466e-01
5.21248519e-01 -1.38089406e+00 7.26858199e-01 7.30047941e-01
3.47170830e-01 8.16173255e-01 1.05727530e+00 -8.66223872e-01
-7.33173192e-01 -1.16579604e+00 1.13360572e+00 -6.02910399e-01
9.93744195e-01 -6.07919693e-01 -1.37232769e+00 6.15330219e-01
5.00227392e-01 -1.00355351e+00 1.11163080e+00 5.97785830e-01
-2.53132612e-01 1.98554054e-01 -7.11124539e-01 2.73900628e-01
3.94323707e-01 -2.32587874e-01 -9.71730053e-01 5.80507576e-01
1.07460189e+00 1.38600357e-02 -9.51901019e-01 -5.08267730e-02
4.11309391e-01 -6.26382709e-01 8.37592006e-01 -5.85574925e-01
6.91982955e-02 -8.61198008e-02 2.68123224e-02 -1.11773992e+00
-3.33272547e-01 -9.87260640e-01 -5.30622840e-01 1.56970453e+00
3.03603113e-01 -3.82706225e-01 8.82888317e-01 4.90397722e-01
1.61992356e-01 -4.09000009e-01 -8.79147708e-01 -8.63412499e-01
1.56756878e-01 -8.43124330e-01 7.19190300e-01 1.26498306e+00
3.77049744e-01 -8.00607502e-02 -1.81877136e-01 1.74602166e-01
5.34951270e-01 6.45303667e-01 5.70388138e-01 -1.88152611e+00
3.20844084e-01 -6.34992242e-01 -3.67875636e-01 -5.93798280e-01
7.68280774e-02 -5.42423487e-01 -2.19248205e-01 -1.44516683e+00
3.49675328e-01 -3.95714760e-01 -3.05413038e-01 3.68416995e-01
-1.90530978e-02 -2.64734954e-01 -3.85151058e-02 8.02140236e-01
-9.63314354e-01 6.11058116e-01 1.27678707e-01 -1.80811182e-01
-7.90077567e-01 3.19023699e-01 -4.89927143e-01 7.15215206e-01
7.61983275e-01 -6.85437739e-01 -4.29413438e-01 3.87574285e-02
-1.07288457e-01 -3.03789735e-01 -9.29749757e-02 -9.17967200e-01
5.57021976e-01 -1.27951995e-01 2.06746504e-01 -1.42345655e+00
1.37438357e-01 -5.66795707e-01 2.71552861e-01 3.77953976e-01
-3.88695240e-01 4.20721799e-01 4.59006488e-01 8.37509453e-01
-3.43754500e-01 3.20337981e-01 4.20432925e-01 -5.88656217e-02
-9.83092904e-01 3.13724518e-01 -9.51196849e-01 -2.25189343e-01
1.29392695e+00 -2.80413955e-01 -1.84861094e-01 -5.10301888e-01
-7.79782295e-01 5.72183609e-01 3.06843847e-01 1.00031662e+00
2.52935559e-01 -1.02993357e+00 -4.61654752e-01 -1.48724079e-01
2.54384190e-01 3.68231721e-02 4.48599368e-01 2.78268248e-01
6.86762184e-02 8.97646844e-01 1.18490659e-01 -1.03879082e+00
-1.37362885e+00 6.00006580e-01 -2.15222791e-01 -5.58327079e-01
-1.05103445e+00 5.20147145e-01 3.82140130e-02 -2.89215803e-01
3.12138230e-01 -3.07487994e-01 -4.31594312e-01 6.72148287e-01
8.47265244e-01 4.69324201e-01 -1.61785871e-01 -4.62477148e-01
-1.89988479e-01 3.69353533e-01 -4.64364558e-01 -3.00230056e-01
1.82587874e+00 -5.22191703e-01 -2.33743146e-01 1.17814207e+00
9.09266412e-01 -4.26713914e-01 -1.18260646e+00 -5.61841369e-01
7.89964676e-01 5.68764396e-02 -9.44693014e-03 -3.08766872e-01
-5.17308712e-01 4.31839287e-01 4.82443452e-01 9.48407829e-01
9.59874213e-01 5.40560842e-01 6.76780581e-01 1.10721052e-01
1.21604010e-01 -1.22780144e+00 -7.06199184e-02 4.30122197e-01
5.37634909e-01 -8.22896600e-01 1.25462741e-01 -1.60407513e-01
-3.79336208e-01 1.19600272e+00 2.08281815e-01 2.48599872e-01
1.18964350e+00 1.93438023e-01 6.66510081e-03 -6.43853843e-01
-1.24409330e+00 2.45297357e-01 4.12798643e-01 6.00950062e-01
1.63696378e-01 -4.45451355e-03 -2.71503448e-01 4.20995712e-01
-7.73580313e-01 -5.55561662e-01 3.55377734e-01 8.87597024e-01
-8.30120087e-01 -9.34745431e-01 -5.88399172e-01 3.32402498e-01
-5.33061445e-01 1.98777676e-01 -2.83765733e-01 9.24325705e-01
-1.84444711e-01 8.35820854e-01 8.16231310e-01 -1.34532928e-01
-3.96959007e-01 4.70105469e-01 -5.04402816e-01 -6.11062288e-01
-2.95158803e-01 4.41712052e-01 -3.38329136e-01 -4.24758822e-01
-4.99147356e-01 -1.16285324e+00 -1.07200539e+00 7.08693042e-02
-2.58365393e-01 7.56309927e-01 1.03095877e+00 7.94769466e-01
6.33408368e-01 3.40912312e-01 6.63484454e-01 -5.93466222e-01
2.34091192e-01 -1.22070849e+00 -5.82156658e-01 1.37063786e-01
-5.51696680e-02 -6.27811253e-01 -9.27176625e-02 4.62858677e-01] | [10.351737022399902, 7.041601181030273] |
8fa37918-3946-489b-b0d1-c1abfe4391cf | clean-text-and-full-body-transformer | 2210.13326 | null | https://arxiv.org/abs/2210.13326v1 | https://arxiv.org/pdf/2210.13326v1.pdf | Clean Text and Full-Body Transformer: Microsoft's Submission to the WMT22 Shared Task on Sign Language Translation | This paper describes Microsoft's submission to the first shared task on sign language translation at WMT 2022, a public competition tackling sign language to spoken language translation for Swiss German sign language. The task is very challenging due to data scarcity and an unprecedented vocabulary size of more than 20k words on the target side. Moreover, the data is taken from real broadcast news, includes native signing and covers scenarios of long videos. Motivated by recent advances in action recognition, we incorporate full body information by extracting features from a pre-trained I3D model and applying a standard transformer network. The accuracy of the system is further improved by applying careful data cleaning on the target text. We obtain BLEU scores of 0.6 and 0.78 on the test and dev set respectively, which is the best score among the participants of the shared task. Also in the human evaluation the submission reaches the first place. The BLEU score is further improved to 1.08 on the dev set by applying features extracted from a lip reading model. | ['Oscar Koller', 'Cyrine Chaabani', 'Abhilash Pal', 'Subhadeep Dey'] | 2022-10-24 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 2.48496026e-01 4.48630929e-01 -2.38761798e-01 -2.34355688e-01
-1.56518054e+00 -4.90934610e-01 8.62008035e-01 -6.49902046e-01
-7.31605709e-01 7.77353287e-01 7.32468367e-01 1.15381800e-01
3.73964727e-01 2.86007166e-01 -7.18165934e-01 -7.05083668e-01
1.77240655e-01 5.21013916e-01 3.14594299e-01 -1.15834810e-01
-6.67205304e-02 1.00456610e-01 -1.56843972e+00 4.77895498e-01
5.42185724e-01 9.34547365e-01 7.16122761e-02 8.27524662e-01
2.42441565e-01 6.57315195e-01 -5.40682077e-01 -5.06824374e-01
4.90573645e-01 -6.50261760e-01 -8.08081508e-01 -7.39174858e-02
1.08733559e+00 -5.21486521e-01 -3.96920919e-01 8.39612603e-01
1.04325426e+00 -9.43875089e-02 5.12580097e-01 -1.12781322e+00
-2.24319488e-01 4.04696822e-01 -3.16919267e-01 -1.86066836e-01
4.59187597e-01 4.89076227e-01 1.08725095e+00 -9.88069236e-01
1.23096800e+00 9.12452519e-01 3.45376998e-01 1.06358445e+00
-5.99118114e-01 -5.24790049e-01 -9.82711790e-04 1.45578861e-01
-1.12893212e+00 -9.97514665e-01 2.96186090e-01 -4.15702701e-01
9.39542353e-01 2.37481460e-01 6.87689185e-01 1.47768235e+00
-2.62619972e-01 1.15953290e+00 1.28368378e+00 -6.47138596e-01
-2.40329280e-02 2.97620445e-02 -2.91845948e-01 5.19051313e-01
-1.72249347e-01 1.47252530e-01 -1.20858479e+00 1.98491186e-01
3.61078769e-01 -7.04210639e-01 -4.32874113e-01 -4.48862106e-01
-1.72037780e+00 3.21836442e-01 -2.62874607e-02 3.91143441e-01
-3.57777655e-01 3.93665045e-01 6.18207455e-01 3.95155370e-01
1.78046852e-01 -8.50213412e-03 -5.77671349e-01 -6.14090979e-01
-9.85728920e-01 3.65583658e-01 8.03443193e-01 9.27620053e-01
-1.02040246e-01 -2.46110931e-01 -4.75533307e-01 7.77943671e-01
4.13295388e-01 1.00512707e+00 6.96174443e-01 -8.53111386e-01
7.89841712e-01 9.86842290e-02 1.11242428e-01 -9.38303843e-02
-9.63154137e-02 -5.02689295e-02 -3.74197066e-01 4.74114627e-01
1.19213665e+00 -2.92326719e-01 -1.40009129e+00 1.86886001e+00
7.67208114e-02 -1.98354140e-01 2.92392820e-01 1.20090830e+00
1.01109421e+00 2.38047749e-01 5.22003174e-02 -5.57737015e-02
1.38223505e+00 -1.04787254e+00 -6.68858707e-01 -1.27681971e-01
6.82020426e-01 -8.91599536e-01 1.07949889e+00 5.44666290e-01
-1.01718533e+00 -1.36510149e-01 -8.79427671e-01 -1.56974033e-01
-9.36291739e-02 4.67717886e-01 3.00400615e-01 5.24431705e-01
-1.06107402e+00 9.88519490e-02 -7.08346725e-01 -7.15252042e-01
4.10080492e-01 4.07639861e-01 -7.21931219e-01 -1.57310784e-01
-9.53998625e-01 1.20326769e+00 -7.26584569e-02 -1.01362459e-01
-7.36610591e-01 -3.59817207e-01 -6.31466210e-01 -5.98971307e-01
3.34507853e-01 -4.10994291e-01 1.49537218e+00 -9.90211725e-01
-2.08477283e+00 1.38202834e+00 -1.41874254e-01 -4.28977668e-01
1.27796376e+00 -3.91121596e-01 -2.73430735e-01 8.47975165e-02
-7.70616308e-02 7.44465351e-01 7.49792755e-01 -6.63423836e-01
-7.88766384e-01 -2.26427466e-01 -1.46020621e-01 2.81129092e-01
1.05324686e-01 3.91473502e-01 -5.46955645e-01 -5.10115504e-01
3.18509229e-02 -1.06850946e+00 6.96195662e-02 1.36960223e-01
-2.00167090e-01 -1.76587760e-01 4.61158484e-01 -1.27168965e+00
6.99749768e-01 -2.10543966e+00 3.81702930e-01 1.38000309e-01
6.43682778e-02 3.05919081e-01 -2.51978397e-01 2.15352654e-01
1.48792118e-01 -3.40083629e-01 -3.81480120e-02 -3.03849369e-01
1.98064491e-01 5.58995605e-02 -2.23446608e-01 4.97342110e-01
-1.89301535e-01 1.11482644e+00 -6.80653095e-01 -4.75822717e-01
2.24102605e-02 3.30981344e-01 -2.66121447e-01 -3.48169990e-02
-2.35119462e-01 6.64797962e-01 -1.18878797e-01 5.89445293e-01
2.98623145e-01 1.43704817e-01 3.98078747e-02 -6.39897510e-02
-1.94690868e-01 4.98759329e-01 -9.54328775e-01 2.19963169e+00
-4.36836511e-01 9.88637030e-01 8.54102150e-02 -6.63720369e-01
5.83462596e-01 6.78943396e-01 5.09043872e-01 -7.39087820e-01
2.53265083e-01 6.09949708e-01 9.44046155e-02 -7.32096553e-01
3.25327292e-02 -1.44352108e-01 -1.75714314e-01 4.48541582e-01
3.52735609e-01 -2.77494162e-01 2.57914275e-01 5.12779625e-05
1.06848705e+00 5.79174399e-01 6.77591041e-02 -2.00409025e-01
5.79708099e-01 -5.41218668e-02 3.69445026e-01 4.38302726e-01
-4.07947451e-01 8.69827807e-01 3.59162837e-01 -1.74659133e-01
-9.39631402e-01 -9.56825972e-01 6.72699213e-02 9.47253048e-01
-2.77684659e-01 -4.10297602e-01 -8.54328334e-01 -8.72349203e-01
-2.37201229e-01 4.53915924e-01 -4.84220654e-01 1.43217310e-01
-8.05568039e-01 -2.13795587e-01 9.20715988e-01 4.91632938e-01
5.61900437e-01 -1.07276952e+00 -6.88618779e-01 -6.48627877e-02
-6.43786848e-01 -1.38069642e+00 -7.30405211e-01 -2.66806215e-01
-3.06851864e-01 -1.01776314e+00 -1.46106458e+00 -8.79235029e-01
3.58239412e-01 -4.30327505e-01 6.67039156e-01 -2.54078180e-01
-1.66425243e-01 4.04779136e-01 -3.71036798e-01 -6.55309737e-01
-5.65349519e-01 -4.30508656e-03 2.25004286e-01 8.59800801e-02
4.34359342e-01 -4.12081517e-02 -2.71769911e-01 3.42243105e-01
-2.91425467e-01 2.01576710e-01 6.75527751e-01 9.09842312e-01
2.62911826e-01 -9.28544462e-01 2.69853860e-01 6.17394559e-02
2.75674015e-01 3.36423010e-01 -5.83794534e-01 2.87480354e-01
-3.26055288e-01 1.90984949e-01 -4.83367108e-02 -6.13044918e-01
-9.08986509e-01 3.08328509e-01 -8.78767744e-02 1.03121035e-01
-3.11912447e-02 -8.22119694e-03 -4.59177494e-01 -3.24385837e-02
4.25366402e-01 2.83658355e-01 3.25880975e-01 -5.01406312e-01
3.60372066e-01 1.09607220e+00 8.03791523e-01 -2.05394939e-01
6.62082314e-01 3.32555711e-01 -1.49123266e-01 -8.79871011e-01
-6.13188148e-01 -3.62867951e-01 -7.19121516e-01 -4.25702363e-01
7.64221489e-01 -9.67221618e-01 -6.77956760e-01 1.01963341e+00
-1.09232807e+00 -5.20313561e-01 -5.32562971e-01 9.66944039e-01
-9.67238784e-01 2.80332714e-01 -2.89065301e-01 -5.74081063e-01
-3.42912883e-01 -1.23128843e+00 1.25908434e+00 -1.40438780e-01
-4.57519829e-01 -4.23734814e-01 2.65617430e-01 7.57970810e-01
3.81515026e-01 2.00428545e-01 2.19982952e-01 -6.24238253e-01
-5.23369849e-01 -4.92964268e-01 -6.50930554e-02 6.75462127e-01
-1.99199114e-02 -3.69350612e-01 -1.09612834e+00 -1.32557958e-01
-4.28195715e-01 -5.78824222e-01 9.41868544e-01 2.78587312e-01
2.06820250e-01 -9.91105568e-03 -2.36349657e-01 1.91922769e-01
8.08212638e-01 -1.81141242e-01 6.67562664e-01 1.09456927e-01
5.26848733e-01 6.99945033e-01 5.92318892e-01 1.35948643e-01
2.94671237e-01 1.16393530e+00 -2.72559542e-02 1.77991718e-01
-8.93755436e-01 -2.57179797e-01 7.54579961e-01 6.92671001e-01
-4.00616527e-01 -1.05863258e-01 -9.37234402e-01 7.43741333e-01
-1.75273216e+00 -8.47848773e-01 -8.21368396e-02 2.30608130e+00
8.56760561e-01 7.08116405e-03 3.38580191e-01 1.69702262e-01
3.67195159e-01 -3.09995711e-02 -1.94747820e-01 -1.86869994e-01
-3.45509320e-01 2.53589064e-01 6.72479451e-01 6.69351995e-01
-1.07227671e+00 1.15811825e+00 6.07183743e+00 7.02227712e-01
-1.43864524e+00 1.66177481e-01 -1.02470763e-01 -5.31743646e-01
2.21964762e-01 -1.57397136e-01 -8.34051609e-01 4.41113621e-01
1.08057916e+00 -2.42132202e-01 2.84713119e-01 3.55409324e-01
2.18000770e-01 -2.73075849e-01 -1.02954662e+00 1.11629713e+00
5.04843652e-01 -8.77187252e-01 -2.22602218e-01 3.75431180e-01
5.94067872e-01 5.89412749e-01 -3.11159361e-02 1.44300997e-01
1.43866539e-01 -9.89050448e-01 9.64021087e-01 7.27735758e-01
1.31069624e+00 -3.47832710e-01 7.35645056e-01 3.78470600e-01
-7.86597371e-01 1.82553157e-01 2.31561720e-01 7.90810734e-02
3.99140567e-01 9.64651704e-02 -9.37681735e-01 2.14915335e-01
4.83539462e-01 5.34121513e-01 -2.18148187e-01 1.11282492e+00
-5.01048744e-01 6.64576948e-01 -7.60312557e-01 -2.04702348e-01
1.27596244e-01 2.90450335e-01 8.48811388e-01 1.03270042e+00
3.82533729e-01 -3.30217928e-01 -2.53684342e-01 -1.77061893e-02
-1.87051177e-01 3.72832656e-01 -6.38533235e-01 1.30354777e-01
-3.45217258e-01 7.42215395e-01 -1.77848503e-01 -3.78105938e-01
-3.49382907e-01 1.31400895e+00 -1.80859581e-01 2.55022049e-01
-6.52404368e-01 -2.89428651e-01 7.09553242e-01 6.45527840e-02
3.86071563e-01 -1.03990152e-01 8.61255452e-02 -1.34232676e+00
6.07393563e-01 -1.02561784e+00 -1.07993275e-01 -6.91961408e-01
-7.38922536e-01 5.27747512e-01 -4.26160656e-02 -1.51322830e+00
-7.33022034e-01 -8.43467593e-01 -1.48940176e-01 8.86821985e-01
-1.19839299e+00 -1.55711925e+00 -7.53546730e-02 4.91875976e-01
5.49585819e-01 -3.88419509e-01 8.74038041e-01 4.13321644e-01
-4.43413481e-02 9.37712908e-01 8.83924291e-02 2.72557616e-01
1.22689128e+00 -9.17472601e-01 5.63036442e-01 8.32034588e-01
3.58829319e-01 7.97350854e-02 7.83778012e-01 -4.90051240e-01
-1.18096209e+00 -5.71042538e-01 1.63826215e+00 -8.71682584e-01
5.83318233e-01 -5.34819841e-01 -2.64858425e-01 5.63693106e-01
2.41658986e-01 1.24090098e-01 3.77354771e-01 -3.26876044e-01
-3.36873204e-01 1.60222977e-01 -1.10771859e+00 5.17027199e-01
1.35095954e+00 -4.75353777e-01 -6.91190779e-01 5.49172044e-01
2.41281986e-01 -7.24419832e-01 -5.84309042e-01 3.29169303e-01
1.29410100e+00 -4.62957650e-01 5.30228019e-01 -8.04136992e-01
2.81769454e-01 -1.58002421e-01 -4.23849225e-01 -1.18334568e+00
4.68946248e-01 -7.97696471e-01 4.45194058e-02 9.83916938e-01
6.02201164e-01 -4.02794331e-01 8.47640455e-01 5.51770031e-01
2.29216620e-01 -4.89702463e-01 -1.45715845e+00 -1.03738081e+00
1.39283553e-01 -5.68042397e-01 1.90643132e-01 2.81970382e-01
1.51487216e-01 2.42493227e-01 -7.42662787e-01 -3.57820570e-01
7.28702664e-01 -2.22479075e-01 1.16338599e+00 -1.13571370e+00
-2.49592394e-01 -2.88089186e-01 -4.91093427e-01 -1.31757581e+00
9.48212147e-02 -9.94104326e-01 3.47083777e-01 -1.45712233e+00
-1.42986150e-02 2.97897190e-01 3.38154994e-02 5.61441123e-01
2.83070624e-01 5.62072337e-01 4.66380149e-01 1.38253897e-01
-4.77530837e-01 3.48182857e-01 1.37574804e+00 -2.28146031e-01
6.82280213e-02 2.46520042e-01 -3.45409423e-01 7.88074195e-01
6.82562709e-01 -2.03813002e-01 1.45966440e-01 -5.01337469e-01
-4.80355993e-02 -8.49557370e-02 4.07811493e-01 -9.27715003e-01
1.48125052e-01 1.64353877e-01 -1.20355554e-01 -4.61458474e-01
4.95123386e-01 -6.64903462e-01 -2.96170801e-01 4.68402117e-01
-4.10711676e-01 -5.16171992e-01 1.23509660e-01 1.70119211e-01
-1.73086464e-01 1.41118169e-01 7.63321161e-01 3.45201679e-02
-7.27466941e-01 4.78816591e-02 -3.78118485e-01 2.66831309e-01
8.82966936e-01 -3.60681355e-01 5.95113412e-02 -5.98636031e-01
-9.97652352e-01 4.94274162e-02 2.31573656e-01 6.73223019e-01
4.18243140e-01 -1.40982068e+00 -1.13263428e+00 2.92712837e-01
4.61466372e-01 -4.71454859e-01 -9.99922901e-02 1.17374349e+00
-3.44615340e-01 5.45525432e-01 -2.40372583e-01 -4.71159637e-01
-1.88535726e+00 -3.41607124e-01 4.01896566e-01 -2.39729956e-01
-8.61781120e-01 8.77468109e-01 -2.66316712e-01 -4.27087039e-01
4.83419687e-01 -4.21656638e-01 9.48097557e-02 1.06633338e-03
6.75140679e-01 3.60452145e-01 1.31292582e-01 -1.15979218e+00
-5.84774673e-01 8.65870595e-01 4.73199114e-02 -7.65672505e-01
1.13870251e+00 -2.76319217e-02 3.55119199e-01 2.67555833e-01
1.14253187e+00 3.26904804e-01 -1.15259755e+00 -2.66179591e-01
-4.96246517e-02 -2.60247529e-01 -9.38241556e-02 -1.36659813e+00
-8.40813279e-01 7.43451715e-01 7.54655778e-01 -7.26966977e-01
8.02471817e-01 1.59119815e-01 9.50073540e-01 5.44376194e-01
4.76434231e-01 -1.33231413e+00 -2.44046688e-01 7.47984946e-01
1.38222051e+00 -1.36169553e+00 -3.09521109e-01 2.77904756e-02
-6.98463380e-01 7.58487523e-01 2.13472649e-01 9.34216082e-02
3.53499353e-01 2.94133276e-01 6.76074326e-01 1.56204239e-01
-4.99296695e-01 -3.40027153e-01 6.11546397e-01 6.72037363e-01
2.57566094e-01 3.65830697e-02 -6.72729075e-01 3.82146657e-01
-4.38996941e-01 4.43891972e-01 2.69448161e-01 7.80984879e-01
-2.08544254e-01 -1.24206650e+00 -2.44695246e-01 1.73905000e-01
-5.74979484e-01 1.72045119e-02 -6.24778330e-01 7.98223495e-01
7.30405897e-02 7.61024594e-01 -4.40128028e-01 -2.27913603e-01
6.44846499e-01 4.63513970e-01 9.02820230e-01 -2.31918395e-01
-3.74388546e-01 -1.79912802e-02 3.27246934e-01 -7.77973056e-01
-4.54038471e-01 -1.08184803e+00 -1.22718632e+00 2.35540479e-01
-1.49659023e-01 -1.36277333e-01 1.05370140e+00 1.14673328e+00
1.10847488e-01 6.98993281e-02 -7.45542571e-02 -7.98912525e-01
-8.36182177e-01 -1.31253946e+00 -3.25072616e-01 3.39330137e-01
5.41051865e-01 -5.07369399e-01 -4.98729080e-01 2.97169119e-01] | [9.2039213180542, -6.532049655914307] |
399ee552-fb85-49e8-a338-097e8c5a54a3 | minimalistic-unsupervised-learning-with-the | 2209.15261 | null | https://arxiv.org/abs/2209.15261v2 | https://arxiv.org/pdf/2209.15261v2.pdf | Minimalistic Unsupervised Learning with the Sparse Manifold Transform | We describe a minimalistic and interpretable method for unsupervised learning, without resorting to data augmentation, hyperparameter tuning, or other engineering designs, that achieves performance close to the SOTA SSL methods. Our approach leverages the sparse manifold transform, which unifies sparse coding, manifold learning, and slow feature analysis. With a one-layer deterministic sparse manifold transform, one can achieve 99.3% KNN top-1 accuracy on MNIST, 81.1% KNN top-1 accuracy on CIFAR-10 and 53.2% on CIFAR-100. With a simple gray-scale augmentation, the model gets 83.2% KNN top-1 accuracy on CIFAR-10 and 57% on CIFAR-100. These results significantly close the gap between simplistic "white-box" methods and the SOTA methods. Additionally, we provide visualization to explain how an unsupervised representation transform is formed. The proposed method is closely connected to latent-embedding self-supervised methods and can be treated as the simplest form of VICReg. Though there remains a small performance gap between our simple constructive model and SOTA methods, the evidence points to this as a promising direction for achieving a principled and white-box approach to unsupervised learning. | ['Yann Lecun', 'Bruno Olshausen', 'Yi Ma', 'Zeyu Yun', 'Yubei Chen'] | 2022-09-30 | null | null | null | null | ['sparse-representation-based-classification', 'unsupervised-image-classification', 'spectral-graph-clustering', 'unsupervised-mnist'] | ['computer-vision', 'computer-vision', 'graphs', 'methodology'] | [-1.04534790e-01 4.53745931e-01 -3.35708201e-01 -2.53882170e-01
-6.35916531e-01 -3.43831867e-01 7.12733388e-01 -9.47570354e-02
-2.53245860e-01 5.51138520e-01 3.43204290e-01 -3.63890886e-01
-3.38346422e-01 -6.50050640e-01 -6.05037332e-01 -8.48924220e-01
-2.16403261e-01 3.85657102e-01 -1.90954208e-01 -7.09868148e-02
1.80183679e-01 3.50962013e-01 -1.47744429e+00 1.01025693e-01
7.67590702e-01 9.01533425e-01 -2.57980794e-01 4.52566713e-01
-3.29411119e-01 6.92412853e-01 -4.04561281e-01 -2.12336965e-02
2.31802881e-01 -3.47251117e-01 -7.21954286e-01 1.71969473e-01
6.98692322e-01 1.11785978e-01 -3.89306396e-01 8.37574065e-01
1.47694603e-01 6.88759461e-02 1.01948428e+00 -1.39612079e+00
-9.57878292e-01 7.44903684e-01 -4.60069478e-01 -5.46580665e-02
-3.06358114e-02 5.31974845e-02 9.38994050e-01 -1.21781516e+00
3.26717854e-01 1.08365798e+00 6.81604624e-01 4.01391566e-01
-1.41457617e+00 -4.79468495e-01 1.59246132e-01 -2.23502874e-01
-1.52167225e+00 -4.42282498e-01 7.98554063e-01 -7.26320922e-01
6.86403453e-01 1.78260937e-01 5.43100834e-01 8.33679676e-01
-1.82607010e-01 7.33998477e-01 1.08766782e+00 -6.74145758e-01
5.16323924e-01 5.27694166e-01 3.67924541e-01 7.41674304e-01
3.51999313e-01 4.00202461e-02 -5.06948411e-01 -2.36033082e-01
9.89221990e-01 2.70967066e-01 -8.34892318e-02 -6.44301653e-01
-1.27119517e+00 9.93998706e-01 5.48310995e-01 5.21680057e-01
8.40201825e-02 2.50119239e-01 9.37537700e-02 4.35457438e-01
4.46648628e-01 6.41799331e-01 -2.53746778e-01 -5.10506257e-02
-1.09966242e+00 -2.12534636e-01 6.39169276e-01 9.09265339e-01
9.01941419e-01 7.07500339e-01 1.91430509e-01 7.26317227e-01
4.70628202e-01 5.77479005e-01 6.30349219e-01 -9.49529648e-01
2.00893104e-01 7.54470229e-01 -1.27752364e-01 -1.13408041e+00
-3.32868159e-01 -9.67369437e-01 -9.64910448e-01 3.55691016e-01
2.21194535e-01 3.10033411e-02 -8.29453886e-01 1.50334287e+00
-4.00403813e-02 2.26467356e-01 2.76164263e-01 7.09607899e-01
7.96089709e-01 6.72928393e-01 -2.07629308e-01 -1.19829483e-01
1.10915971e+00 -1.07873130e+00 -6.06958270e-01 -2.67041493e-02
9.51498330e-01 -4.70349431e-01 1.55789399e+00 4.44566399e-01
-8.23670506e-01 -6.40290499e-01 -1.18845809e+00 1.22590572e-01
-5.01389563e-01 2.86506951e-01 9.34116721e-01 9.12850142e-01
-1.09587932e+00 4.90552396e-01 -7.71560252e-01 -3.35373282e-01
6.81480885e-01 2.74733305e-01 -5.37503541e-01 3.55000496e-02
-6.78789675e-01 6.05044186e-01 2.20167875e-01 -1.66096970e-01
-7.48207390e-01 -7.52003610e-01 -7.46172845e-01 1.57849655e-01
1.54306605e-01 -4.54044968e-01 6.51535869e-01 -7.16550648e-01
-1.59029114e+00 7.03949034e-01 -8.19060206e-02 -7.06096470e-01
9.32760909e-02 -3.73527944e-01 -6.11099780e-01 3.26667160e-01
-1.00654490e-01 7.59979367e-01 1.04616177e+00 -1.29227698e+00
-1.40374407e-01 -1.97072834e-01 -2.24176347e-01 -1.50965005e-01
-1.12694252e+00 -3.59337300e-01 -5.85519224e-02 -9.29505825e-01
4.67253327e-01 -1.03936553e+00 -3.25296223e-01 2.34880112e-02
-4.26945955e-01 -7.72624165e-02 1.07068646e+00 -1.36430025e-01
1.24299097e+00 -2.46348596e+00 -1.44399758e-02 3.10252100e-01
4.25785393e-01 2.03335911e-01 -2.27225482e-01 4.61939007e-01
-3.60350847e-01 3.79762560e-01 -4.57690775e-01 -4.26098287e-01
2.98632812e-02 1.75653817e-03 -5.09525776e-01 4.99806315e-01
2.78510034e-01 9.43525255e-01 -8.37813497e-01 -1.80812851e-01
1.60563275e-01 6.24872088e-01 -5.94857395e-01 -1.14959866e-01
1.44345507e-01 2.78041214e-01 -1.17920138e-01 5.26125371e-01
4.89801794e-01 -4.91439819e-01 -1.11181512e-01 -2.69055009e-01
-1.16517954e-01 8.70348588e-02 -1.19904256e+00 1.90007770e+00
-2.56935418e-01 7.63410687e-01 -3.67530406e-01 -1.25068307e+00
1.16948080e+00 1.98415697e-01 4.21380907e-01 -3.33022952e-01
-1.35921940e-01 1.51957482e-01 -2.16134578e-01 -1.96724877e-01
2.22951487e-01 -4.41859737e-02 6.87533990e-02 5.06028235e-01
3.77114087e-01 -3.69570330e-02 -2.71443240e-02 5.78475296e-01
9.81500745e-01 -1.48559302e-01 1.04448609e-01 -5.98760128e-01
3.06072086e-01 -5.59461638e-02 1.31237134e-01 6.69701815e-01
1.74079999e-01 8.00898671e-01 4.19556797e-01 -5.56259573e-01
-8.70055914e-01 -1.29781353e+00 -2.31754825e-01 9.52940702e-01
-3.26716870e-01 -7.08882689e-01 -6.79028392e-01 -7.43374407e-01
-9.19941217e-02 7.30942667e-01 -6.13931000e-01 -4.03350830e-01
-9.72218364e-02 -6.50796354e-01 5.05302131e-01 4.32709217e-01
4.15692061e-01 -6.39638901e-01 -6.07968718e-02 -1.09013870e-01
2.04888552e-01 -1.00531995e+00 -2.20130458e-01 2.10859492e-01
-1.34147823e+00 -6.91121399e-01 -6.29476130e-01 -9.35334563e-01
8.97129416e-01 3.49923193e-01 9.25432563e-01 -1.34725258e-01
-1.56319603e-01 4.85515386e-01 -3.77240956e-01 -2.53096581e-01
-1.73101231e-01 9.25706476e-02 4.72514182e-01 1.90763339e-01
2.72545159e-01 -9.82688725e-01 -5.03478408e-01 2.68734097e-01
-8.76491964e-01 -6.06391616e-02 5.36581278e-01 7.16714740e-01
7.13618934e-01 2.53724176e-02 7.04363406e-01 -1.03782094e+00
2.83228546e-01 -5.44277132e-01 -4.24212098e-01 3.82774882e-02
-9.06947494e-01 3.23869109e-01 7.14887738e-01 -5.81498206e-01
-5.10853171e-01 3.06782752e-01 2.35657781e-01 -8.18806589e-01
-1.54400125e-01 4.71778601e-01 -9.09125730e-02 -5.39827980e-02
1.22118902e+00 4.57164407e-01 2.26694912e-01 -8.02413285e-01
8.30168307e-01 6.91672981e-01 5.61966658e-01 -3.54020685e-01
1.24926901e+00 8.08647573e-01 -2.42864072e-01 -1.06081665e+00
-9.07314837e-01 -3.68181378e-01 -8.40068579e-01 2.08757475e-01
6.11311972e-01 -1.03241158e+00 -2.86503583e-01 1.04108207e-01
-5.58207929e-01 -3.45804751e-01 -9.64414954e-01 6.47728860e-01
-5.82823634e-01 2.37960875e-01 -3.67752969e-01 -7.22142458e-01
-3.02609921e-01 -6.03758097e-01 7.56937921e-01 -7.73009509e-02
-3.22063804e-01 -1.19547391e+00 2.55265925e-02 2.56550163e-01
6.80173934e-01 5.92542142e-02 8.45192671e-01 -7.19614804e-01
-3.85227263e-01 -3.79513800e-01 -1.37137264e-01 6.09305799e-01
3.04837167e-01 -5.31374775e-02 -1.22369432e+00 -4.96620238e-01
8.44930634e-02 -3.79280835e-01 7.94791937e-01 2.67543644e-01
1.23917305e+00 -5.86089611e-01 -1.23549603e-01 8.16594183e-01
1.30661857e+00 -1.98831365e-01 6.84261441e-01 1.19534783e-01
8.99890065e-01 3.65322113e-01 2.15395048e-01 2.65419483e-01
2.29451403e-01 4.11066145e-01 1.86289042e-01 -2.96394140e-01
-3.08837056e-01 -3.84149641e-01 4.55379725e-01 1.27911210e+00
-2.29815114e-02 3.05677742e-01 -9.68025684e-01 4.31030571e-01
-1.61806071e+00 -8.53969336e-01 -5.84843345e-02 2.21304679e+00
5.29264987e-01 2.64520019e-01 2.67109245e-01 4.97695327e-01
4.18837488e-01 1.27444685e-01 -3.49221140e-01 -2.09445775e-01
-2.60930151e-01 1.55145019e-01 2.43426949e-01 5.13451338e-01
-1.13964045e+00 9.68352497e-01 7.06671715e+00 9.61631477e-01
-1.13879681e+00 1.14435516e-01 7.43080437e-01 -1.38551548e-01
-4.68753844e-01 1.55657366e-01 -6.85343444e-01 2.91256338e-01
1.12539482e+00 -9.44552664e-03 2.30883822e-01 1.02639234e+00
1.57689676e-01 2.51166552e-01 -1.28067911e+00 1.34554923e+00
1.32716641e-01 -1.74863017e+00 3.03382248e-01 1.53340682e-01
9.97120440e-01 3.57350223e-02 3.64001632e-01 3.13094854e-01
9.55833495e-02 -1.43242085e+00 4.18329209e-01 5.45701504e-01
1.00780320e+00 -4.94033664e-01 4.25621420e-01 2.60974228e-01
-1.10779738e+00 -1.60354581e-02 -4.88852412e-01 -1.17974915e-01
-2.43587747e-01 8.04052353e-01 -4.85086411e-01 1.87083051e-01
5.46992540e-01 1.04055917e+00 -6.88800275e-01 1.04578352e+00
-1.27590165e-01 1.06421638e+00 -4.55080748e-01 1.09247252e-01
1.66902423e-01 -2.63760418e-01 5.13091326e-01 1.24831355e+00
2.12967038e-01 -6.81991726e-02 1.30858690e-01 8.61500561e-01
-4.43466939e-02 2.35300347e-01 -8.33525002e-01 -3.93197387e-01
5.44096231e-01 1.15822017e+00 -7.33890653e-01 -4.46014225e-01
-3.27593327e-01 8.35603178e-01 2.41987735e-01 6.73201382e-01
-5.15333831e-01 -4.31327581e-01 4.02970046e-01 2.40871191e-01
3.04454595e-01 -3.91173273e-01 -4.84703451e-01 -1.41828382e+00
-1.20511293e-01 -7.80966699e-01 1.90022886e-01 -5.39861321e-01
-1.39216161e+00 6.60379350e-01 -1.62947565e-01 -1.61569846e+00
-8.88688937e-02 -5.71932793e-01 -5.89780211e-01 3.69022727e-01
-1.34495270e+00 -9.43097889e-01 -1.98023602e-01 6.39652371e-01
2.54162461e-01 -7.16888547e-01 1.20333350e+00 1.58076569e-01
-4.64744717e-01 8.95486712e-01 5.15426397e-01 2.23703861e-01
3.43408376e-01 -1.19567120e+00 2.80852079e-01 6.15994990e-01
7.33787179e-01 7.01614559e-01 4.38707441e-01 -9.65558812e-02
-1.35212255e+00 -1.09943449e+00 6.21858656e-01 -5.30904949e-01
8.12341094e-01 -5.21407068e-01 -1.00205195e+00 6.62397921e-01
-8.82981122e-02 4.00876045e-01 9.86513019e-01 3.63732666e-01
-7.60252595e-01 -2.28624001e-01 -1.03423381e+00 5.62252283e-01
1.06817305e+00 -7.60089517e-01 -6.92661762e-01 4.67391670e-01
1.01436079e+00 1.78276837e-01 -1.14535868e+00 1.17037199e-01
2.80629367e-01 -7.57914364e-01 1.07722116e+00 -7.44215310e-01
2.67713070e-01 -2.16479793e-01 -3.42039883e-01 -1.18681741e+00
-5.65861762e-01 -7.56434321e-01 -5.01487494e-01 1.15936804e+00
5.30039966e-01 -7.95387268e-01 1.08921623e+00 3.07606667e-01
-3.00897975e-02 -7.57317305e-01 -6.73540294e-01 -1.16843164e+00
4.80952650e-01 -5.01321614e-01 3.08126599e-01 1.34579480e+00
1.09400280e-01 4.26730394e-01 -2.36174121e-01 5.66521585e-02
7.82136858e-01 8.94886106e-02 9.49972689e-01 -1.40464318e+00
-2.58766234e-01 -4.04225558e-01 -6.52536392e-01 -1.04148567e+00
-1.14447214e-01 -1.15488994e+00 -5.13362348e-01 -1.30866027e+00
-9.70451161e-03 -7.15178251e-01 -6.68232143e-01 6.02794230e-01
2.43267179e-01 3.80989164e-01 2.27684051e-01 5.07623374e-01
-5.85399151e-01 7.89393842e-01 8.75205994e-01 -1.37410805e-01
-3.83259535e-01 -3.35418701e-01 -1.04706669e+00 8.93642902e-01
9.62794304e-01 -6.61204159e-01 -7.05670774e-01 -3.45319360e-01
-1.01841660e-02 -4.04152304e-01 4.02419209e-01 -1.33348680e+00
2.40575671e-01 7.92489760e-03 2.39290759e-01 -2.54135013e-01
3.63417417e-01 -7.25572526e-01 -1.59138128e-01 4.14232314e-01
-4.81535256e-01 -2.39221022e-01 1.18314050e-01 7.25174665e-01
-3.50553930e-01 -4.50114012e-02 6.49488866e-01 6.10382818e-02
-4.25319105e-01 2.96296597e-01 -3.07903677e-01 2.32159153e-01
7.52984345e-01 -3.90870094e-01 -2.47789785e-01 -4.20562804e-01
-9.27152455e-01 -3.38029377e-02 3.05069745e-01 3.31499249e-01
7.05850899e-01 -1.44095886e+00 -6.61101699e-01 4.79252964e-01
1.14229374e-01 -1.98831800e-02 4.27054092e-02 7.73930013e-01
-2.64216036e-01 4.82462585e-01 -9.19562727e-02 -7.98187137e-01
-5.89703023e-01 4.62691635e-01 6.11740500e-02 3.10677532e-02
-7.94257462e-01 9.80282366e-01 2.52040267e-01 -5.41643202e-01
3.56201172e-01 -2.20742345e-01 -1.87815055e-01 -1.49854440e-02
5.63803256e-01 2.78417081e-01 -5.37949875e-02 -3.92290503e-01
-3.05261642e-01 6.66261137e-01 3.05533670e-02 -1.01470180e-01
1.43884814e+00 5.40187024e-02 -3.75058502e-02 9.24610317e-01
1.44735932e+00 2.56359637e-01 -1.16423917e+00 -4.00790870e-01
5.96718788e-02 -2.73697972e-01 1.57977775e-01 -4.97296691e-01
-1.13273883e+00 1.19473445e+00 8.01493704e-01 2.14884400e-01
8.27774882e-01 5.80037907e-02 4.60525244e-01 7.40007818e-01
2.69336849e-01 -9.35812652e-01 4.79193121e-01 4.40541625e-01
8.37045252e-01 -1.32470775e+00 1.65262014e-01 -4.05267656e-01
-4.91231292e-01 9.51269388e-01 3.84398460e-01 -5.39566457e-01
1.04593194e+00 2.30300561e-01 8.27692971e-02 -2.87995100e-01
-6.16720855e-01 1.17131375e-01 4.79964077e-01 6.24202728e-01
3.61624748e-01 2.94638500e-02 1.35316789e-01 6.24762893e-01
-5.47879517e-01 -1.60949796e-01 2.31836304e-01 6.75964177e-01
-6.68155074e-01 -7.18275249e-01 -3.70175958e-01 3.89284551e-01
-9.43657756e-02 -2.79918730e-01 -2.10678309e-01 8.94102693e-01
-2.98442036e-01 7.80900359e-01 1.90587237e-01 -4.65202808e-01
-1.61053501e-02 3.45949501e-01 1.43519878e-01 -7.84499645e-01
-1.74325779e-01 -3.92744225e-03 -2.70917743e-01 -4.16454226e-01
-3.77265334e-01 -2.32268363e-01 -1.41776276e+00 -1.96754664e-01
-2.40949526e-01 3.15785378e-01 6.16281927e-01 8.62913251e-01
5.93636692e-01 1.93626225e-01 7.62279809e-01 -7.25274205e-01
-3.88138801e-01 -7.75753081e-01 -5.75521886e-01 3.66946697e-01
1.46694854e-01 -6.15671337e-01 -7.00581968e-01 1.47444159e-01] | [9.102083206176758, 3.0067641735076904] |
af65dc06-c6fc-464a-a65f-1610b88e8f50 | composing-structure-aware-batches-for | null | null | https://openreview.net/forum?id=DKVhbjXs9aG | https://openreview.net/pdf?id=DKVhbjXs9aG | Composing Structure-Aware Batches for Pairwise Sentence Classification | Identifying the relation between two sentences requires datasets with pairwise annotations. In many cases, these datasets contain instances that are annotated multiple times as part of different pairs. They constitute a structure that contains additional helpful information about the inter-relatedness of the text instances based on the annotations. This paper investigates how this kind of structural dataset information can be exploited during training.
We propose three batch composition strategies to incorporate such information and measure their performance over 14 heterogeneous pairwise sentence classification tasks. Our results show statistically significant improvements (up to 3.9%) - independent of the pre-trained language model - for most tasks compared to baselines that follow a standard training procedure. Further, we see that even this baseline procedure can profit from having such structural information in a low-resource setting. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['sentence-classification'] | ['natural-language-processing'] | [ 3.26423734e-01 3.49932700e-01 -6.83325380e-02 -8.94271076e-01
-1.18205142e+00 -9.40177381e-01 7.27786839e-01 8.85462821e-01
-6.54971361e-01 8.28508019e-01 7.72288978e-01 -1.75825655e-01
5.99609576e-02 -3.02681893e-01 -7.14008272e-01 -6.83912396e-01
-1.32011101e-01 5.99632382e-01 1.93536937e-01 -3.87894034e-01
2.25324526e-01 -4.95360680e-02 -1.21751177e+00 8.39460015e-01
5.04355848e-01 7.70747900e-01 1.23274170e-01 6.13947749e-01
-1.75618544e-01 8.53316784e-01 -7.83572316e-01 -7.50268817e-01
2.55747229e-01 -1.54938698e-01 -1.25335550e+00 -9.35096219e-02
7.67791271e-01 7.93800950e-02 -8.25423077e-02 6.87022328e-01
5.04706979e-01 3.89842726e-02 5.47664225e-01 -9.35678124e-01
-1.19218521e-01 1.24997425e+00 -4.64796066e-01 3.94737601e-01
5.65988064e-01 -2.28005126e-01 1.63000596e+00 -8.70789051e-01
8.63694489e-01 9.64780629e-01 9.30124700e-01 1.65468797e-01
-1.20575285e+00 -3.19273144e-01 -1.56835802e-02 1.17553391e-01
-1.21212089e+00 -6.81858361e-01 7.00680614e-01 -3.78030956e-01
1.42303407e+00 5.20808518e-01 7.64408186e-02 1.19021595e+00
-1.35188296e-01 7.19118774e-01 8.21267366e-01 -6.92045450e-01
-2.14926124e-01 2.10552335e-01 6.62967682e-01 3.71694565e-01
1.53046638e-01 -6.12835288e-01 -5.38474321e-01 -1.93283260e-01
-2.00823858e-01 -4.61031854e-01 -3.20199370e-01 -2.54509896e-01
-1.25995624e+00 5.04765153e-01 2.98655421e-01 7.89311230e-01
2.54185051e-02 -2.31944963e-01 9.66791630e-01 4.88439441e-01
8.72440338e-01 8.56451213e-01 -8.93620968e-01 -2.62412101e-01
-7.98621893e-01 4.10548858e-02 1.11543941e+00 1.06004977e+00
6.80195153e-01 -6.88912749e-01 -4.17087078e-01 1.01296306e+00
-1.32143572e-01 -6.29767701e-02 5.87336063e-01 -7.10610271e-01
1.26896679e+00 6.44720614e-01 -6.17606863e-02 -8.89249742e-01
-6.23449087e-01 -3.61889601e-01 -7.50757515e-01 -6.90116644e-01
6.64649785e-01 -3.11637640e-01 -1.94765806e-01 1.80489516e+00
3.72545600e-01 -6.63396642e-02 1.83271199e-01 4.92768317e-01
9.71194386e-01 2.92408228e-01 -1.33178830e-01 -1.43118978e-01
1.33017838e+00 -1.08629370e+00 -5.68717599e-01 -2.53851265e-01
1.40204167e+00 -9.03761029e-01 9.60706353e-01 -9.75028127e-02
-9.78203356e-01 -4.36052382e-01 -8.92778933e-01 -3.28776240e-01
-2.84749925e-01 1.39077261e-01 3.79413396e-01 5.26811481e-01
-1.10958135e+00 6.54105723e-01 -4.11405921e-01 -4.12730008e-01
1.83218330e-01 3.28010350e-01 -7.35905111e-01 -7.03849569e-02
-1.24561310e+00 1.07570028e+00 4.26784456e-01 -7.11794049e-02
-1.01733573e-01 -9.06059623e-01 -9.22818899e-01 -1.59206763e-02
3.54923427e-01 -5.37572265e-01 1.21802044e+00 -9.25561428e-01
-8.85272622e-01 1.27007771e+00 -5.26000023e-01 -5.32614648e-01
5.36659300e-01 -3.59899580e-01 2.99549811e-02 3.85115109e-02
1.43490151e-01 5.01596093e-01 3.02363634e-01 -1.10709226e+00
-3.17916512e-01 -3.05101186e-01 3.45086932e-01 3.63807142e-01
-5.42271018e-01 4.78016853e-01 -3.75258207e-01 -7.59898245e-01
-2.55141973e-01 -9.49103773e-01 -8.30456614e-02 -4.91101474e-01
-7.42956936e-01 -4.20944512e-01 5.87595522e-01 -7.31560647e-01
1.11785758e+00 -2.00739431e+00 2.55367935e-01 5.03668450e-02
1.64733812e-01 2.81141788e-01 -3.70079428e-01 8.82529557e-01
-2.34478816e-01 3.91945541e-01 -2.95758635e-01 -9.02373075e-01
3.04086618e-02 1.58787683e-01 -1.31126970e-01 2.67343789e-01
2.55894333e-01 8.48690331e-01 -7.77500391e-01 -5.09571075e-01
-2.18990445e-01 2.65935242e-01 -3.81342769e-01 3.90913069e-01
5.89325503e-02 6.08918786e-01 -1.44040331e-01 1.76187485e-01
3.80199492e-01 -2.24667281e-01 6.22532725e-01 -3.03795069e-01
2.54992932e-01 1.01967525e+00 -7.85090268e-01 1.66391325e+00
-4.63104844e-01 7.16391921e-01 -4.86847721e-02 -1.22613966e+00
7.33536184e-01 2.64465779e-01 4.19305712e-01 -4.33216721e-01
9.00948718e-02 1.09155953e-01 3.41554642e-01 -6.27549767e-01
6.99724436e-01 -5.34922257e-02 -1.88387230e-01 5.58601618e-01
8.49619359e-02 6.49062768e-02 4.94470119e-01 4.54797506e-01
1.42755222e+00 -1.98029861e-01 4.20793802e-01 -4.31575984e-01
7.81875908e-01 -3.25386673e-01 4.66747850e-01 8.54181528e-01
4.17603739e-02 6.48446679e-01 7.90503979e-01 -2.70225376e-01
-1.22165215e+00 -4.16722417e-01 -3.46312195e-01 1.04431653e+00
-2.11426571e-01 -9.36635852e-01 -6.19777083e-01 -1.00563526e+00
-1.46786258e-01 5.26084900e-01 -6.81740522e-01 -1.35461479e-01
-7.90223241e-01 -7.61578679e-01 7.71579623e-01 7.12476313e-01
2.46852770e-01 -7.16332436e-01 -1.49804473e-01 -5.71514182e-02
-5.57045698e-01 -1.62523448e+00 -4.00974840e-01 3.83347839e-01
-7.56836534e-01 -1.03701723e+00 -2.79947340e-01 -7.81761825e-01
6.67070031e-01 1.29758820e-01 1.68055594e+00 3.80943894e-01
-2.14776509e-02 2.65777856e-01 -6.79599583e-01 -1.29844710e-01
-4.75491524e-01 4.45984632e-01 -1.13799572e-01 -1.72534898e-01
4.27496344e-01 -5.52049756e-01 -5.73709868e-02 1.47872165e-01
-4.57910508e-01 -9.42620561e-02 3.91043901e-01 9.45160389e-01
1.60206810e-01 -6.92029670e-02 3.98509026e-01 -1.32980394e+00
4.76830363e-01 -5.07949650e-01 -7.69308433e-02 5.60751021e-01
-7.37213641e-02 2.41090611e-01 6.09953344e-01 -1.97839901e-01
-9.51143205e-01 -1.05701536e-01 -2.31221542e-02 1.39775932e-01
-2.41744384e-01 6.18186712e-01 -3.21225047e-01 2.28508517e-01
3.69592845e-01 -3.22185636e-01 -1.74003333e-01 -5.71540475e-01
2.17948660e-01 7.44201541e-01 9.69816744e-02 -8.51277411e-01
6.13457620e-01 5.77207059e-02 -3.23177967e-03 -6.32029772e-01
-1.29570603e+00 -7.54895091e-01 -1.17731869e+00 2.40048066e-01
7.11760342e-01 -9.29917395e-01 -4.24577534e-01 2.09703699e-01
-1.36577117e+00 -3.81418467e-01 -9.35627967e-02 4.00296211e-01
-1.64812431e-01 5.61819315e-01 -8.31926882e-01 -6.07061684e-01
-1.35683596e-01 -1.04302168e+00 1.12888515e+00 -3.78088742e-01
-6.42728925e-01 -1.24409044e+00 1.79076679e-02 7.65617013e-01
1.57927319e-01 -1.89749300e-01 8.60044122e-01 -1.39117861e+00
-7.27778748e-02 -3.26975286e-01 -8.28059018e-02 3.24423045e-01
1.77219778e-01 -5.74877784e-02 -1.09713840e+00 -2.97651857e-01
-3.55862230e-02 -4.35312778e-01 8.46355498e-01 -7.13199154e-02
1.17654634e+00 -4.60380197e-01 -2.46830821e-01 1.59652621e-01
1.04660094e+00 -3.60708624e-01 3.02569658e-01 2.58121818e-01
9.31126177e-01 1.14254677e+00 5.60978472e-01 2.77225763e-01
6.25423789e-01 1.06713045e+00 1.72680169e-01 -2.65542362e-02
-4.75162826e-02 1.52797159e-02 4.02190059e-01 1.23003948e+00
-4.30142228e-03 -3.97909969e-01 -9.67889786e-01 5.07109702e-01
-1.81519854e+00 -1.02738678e+00 -4.53377128e-01 2.07213402e+00
1.28106034e+00 1.43860295e-01 1.01628743e-01 2.07607239e-01
6.25475764e-01 2.27678999e-01 -6.71257302e-02 -4.96701837e-01
-4.58254129e-01 1.64957315e-01 4.51777488e-01 6.23904109e-01
-1.28829861e+00 6.84793591e-01 6.67703438e+00 5.50707519e-01
-5.45037806e-01 3.26275192e-02 7.22810686e-01 -1.03144810e-01
-3.30243230e-01 1.33660622e-02 -9.79139268e-01 4.56830084e-01
1.15434635e+00 6.54336289e-02 -1.86099976e-01 3.74035031e-01
-4.22626697e-02 -1.46831900e-01 -1.48991323e+00 5.46076834e-01
3.06978464e-01 -1.06536782e+00 -1.68350339e-01 -1.18719898e-01
6.13261342e-01 1.93673506e-01 -3.27510953e-01 2.79888779e-01
3.08263004e-01 -7.88373172e-01 5.71815133e-01 2.20736817e-01
2.76420534e-01 -5.23618400e-01 1.36561561e+00 5.48890889e-01
-1.01506186e+00 5.17858304e-02 -3.35784733e-01 -2.30439052e-01
2.17072368e-01 1.00812161e+00 -9.20124292e-01 9.21126723e-01
6.58194184e-01 9.75224674e-01 -1.11133122e+00 8.20883274e-01
-2.92721599e-01 7.16848612e-01 -2.62859911e-01 -6.65691122e-02
8.75039101e-02 -1.27492165e-02 5.25300860e-01 1.50541461e+00
-5.55123668e-03 -1.44513845e-01 1.11174554e-01 2.23099738e-01
-4.79560107e-01 3.53606075e-01 -7.28681862e-01 7.62295276e-02
4.00537729e-01 1.35394824e+00 -5.40974736e-01 -4.21907395e-01
-6.79135263e-01 8.90725315e-01 8.87899876e-01 -1.44034550e-01
-6.15814328e-01 -2.17846587e-01 4.78611410e-01 -1.57034636e-01
3.17609280e-01 -2.09941298e-01 -2.52856702e-01 -1.30430019e+00
5.11707127e-01 -8.77238810e-01 4.79761422e-01 -6.35679424e-01
-1.74517584e+00 6.62323177e-01 6.45417571e-02 -9.17328954e-01
-2.86073387e-01 -5.90873837e-01 -5.16181529e-01 7.41617203e-01
-1.24801517e+00 -1.26747215e+00 -3.36583657e-03 1.72542423e-01
3.59419078e-01 -3.38192768e-02 9.62037146e-01 3.79681915e-01
-7.47722328e-01 9.81283069e-01 2.47258097e-02 5.36666155e-01
1.12460363e+00 -1.45248342e+00 3.02908480e-01 8.58484745e-01
6.14191949e-01 8.64873469e-01 6.86221004e-01 -4.35429484e-01
-1.00132978e+00 -7.63383210e-01 1.70988631e+00 -1.03581774e+00
9.08328533e-01 -7.57577837e-01 -1.04901016e+00 8.76860499e-01
7.14345455e-01 -1.51432380e-02 1.21773052e+00 9.21973646e-01
-8.26073408e-01 -2.75317654e-02 -8.36181521e-01 3.45807135e-01
1.26367974e+00 -7.87338197e-01 -8.91724825e-01 5.87659359e-01
8.55636120e-01 -4.80463445e-01 -1.08987463e+00 5.36886215e-01
2.50479043e-01 -9.57587540e-01 8.61984670e-01 -1.08511949e+00
7.86855638e-01 1.41261742e-01 -3.18217218e-01 -1.19055772e+00
6.17170036e-02 -4.11891639e-01 2.23010674e-01 1.80197716e+00
8.67845476e-01 -4.90442097e-01 4.27033097e-01 8.77598345e-01
-2.89717704e-01 -5.94091713e-01 -8.80546808e-01 -7.73427188e-01
3.06661159e-01 -3.00164759e-01 4.47621584e-01 1.35375571e+00
5.28768718e-01 8.07685316e-01 -8.23516548e-02 -1.78698972e-01
1.84953585e-01 -2.07817480e-02 7.55332708e-01 -1.10147285e+00
-3.62854004e-01 -4.23008621e-01 -5.64761937e-01 -7.45087743e-01
9.04247463e-01 -1.26514912e+00 1.96924582e-02 -1.27563119e+00
5.36681414e-01 -5.49044788e-01 -2.00266510e-01 7.08044827e-01
-6.39786482e-01 1.38409659e-01 1.13924749e-01 2.47684434e-01
-8.56977880e-01 2.44408786e-01 6.64681911e-01 -1.20606713e-01
1.66787028e-01 -2.05471009e-01 -6.92514539e-01 5.77400386e-01
9.83296037e-01 -4.92303461e-01 -2.92065740e-01 -7.39669442e-01
4.48806226e-01 -2.86972404e-01 1.13632105e-01 -7.11008132e-01
2.27391109e-01 3.48425388e-01 4.10807040e-03 -3.98736089e-01
2.19416171e-01 -8.50300074e-01 -2.12987959e-01 6.79662824e-02
-9.75917399e-01 2.33259901e-01 1.93577766e-01 4.53340113e-01
-4.76980001e-01 -5.56946516e-01 2.92149246e-01 -7.04983473e-02
-3.39062124e-01 -2.69874603e-01 -1.83645800e-01 4.84717876e-01
8.81020606e-01 3.22437853e-01 -5.78056395e-01 -3.14209491e-01
-5.49833536e-01 1.39477089e-01 4.00608122e-01 4.53346610e-01
-9.35913175e-02 -1.05996883e+00 -8.43426108e-01 -1.70415416e-01
2.99859941e-01 -9.11351964e-02 7.40323216e-02 1.23485720e+00
3.65479179e-02 4.34439689e-01 1.34584829e-01 -4.45411831e-01
-1.75571132e+00 3.73551786e-01 1.10916547e-01 -7.51465023e-01
-5.49526751e-01 1.04492736e+00 -1.39470458e-01 -7.75039911e-01
9.42482874e-02 -1.34377167e-01 -2.81503588e-01 4.27691251e-01
3.73349786e-01 -8.45206156e-03 4.33516532e-01 -9.20246065e-01
-5.44406056e-01 3.77216578e-01 -2.80308068e-01 -4.46759053e-02
1.38914442e+00 -2.33784035e-01 -6.28550947e-01 7.93089628e-01
1.55677617e+00 4.45640206e-01 -6.29893243e-01 -4.66065079e-01
5.15080035e-01 -2.96820164e-01 -4.71935064e-01 -6.94705427e-01
-8.61140490e-01 9.17482257e-01 -3.47268820e-01 2.66985714e-01
7.58742571e-01 2.32887149e-01 5.32978356e-01 5.78002810e-01
1.89178273e-01 -8.70778739e-01 -6.77999705e-02 7.52517879e-01
8.09831440e-01 -1.55399573e+00 1.52340248e-01 -8.11098933e-01
-7.76959896e-01 9.73043323e-01 5.26049972e-01 1.68677002e-01
5.50716937e-01 4.42588747e-01 -3.21832895e-02 -2.63424907e-02
-1.01545525e+00 -2.23006785e-01 5.32458842e-01 4.28890020e-01
1.03322637e+00 -1.72823429e-01 -5.19503832e-01 6.03621721e-01
-3.28722805e-01 -7.29483306e-01 3.04986447e-01 8.33596706e-01
-3.60975191e-02 -1.61948729e+00 2.23020941e-01 4.63251919e-01
-6.36314869e-01 -3.93979877e-01 -8.49469781e-01 5.86374342e-01
-1.82077870e-01 1.10952032e+00 1.06547363e-01 -3.76818985e-01
2.20422715e-01 3.36758643e-01 4.87745434e-01 -8.25210690e-01
-1.12570953e+00 -3.76908541e-01 8.97038043e-01 -4.14931655e-01
-8.65381539e-01 -8.62076879e-01 -9.87600982e-01 -7.78086483e-02
-3.88422430e-01 2.84413189e-01 2.51999199e-01 1.25423658e+00
3.71566713e-01 4.37094271e-01 5.62862396e-01 -5.88841200e-01
-4.62915629e-01 -1.27815270e+00 -1.71702251e-01 8.39855731e-01
7.93320164e-02 -3.86367023e-01 -4.79738206e-01 2.62268409e-02] | [9.613173484802246, 8.893842697143555] |
eb5443ac-0bdd-4e82-8b65-7dc64ac2fdb0 | liquid-state-machine-empowered-reflection | 2208.04400 | null | https://arxiv.org/abs/2208.04400v1 | https://arxiv.org/pdf/2208.04400v1.pdf | Liquid State Machine-Empowered Reflection Tracking in RIS-Aided THz Communications | Passive beamforming in reconfigurable intelligent surfaces (RISs) enables a feasible and efficient way of communication when the RIS reflection coefficients are precisely adjusted. In this paper, we present a framework to track the RIS reflection coefficients with the aid of deep learning from a time-series prediction perspective in a terahertz (THz) communication system. The proposed framework achieves a two-step enhancement over the similar learning-driven counterparts. Specifically, in the first step, we train a liquid state machine (LSM) to track the historical RIS reflection coefficients at prior time steps (known as a time-series sequence) and predict their upcoming time steps. We also fine-tune the trained LSM through Xavier initialization technique to decrease the prediction variance, thus resulting in a higher prediction accuracy. In the second step, we use ensemble learning technique which leverages on the prediction power of multiple LSMs to minimize the prediction variance and improve the precision of the first step. It is numerically demonstrated that, in the first step, employing the Xavier initialization technique to fine-tune the LSM results in at most 26% lower LSM prediction variance and as much as 46% achievable spectral efficiency (SE) improvement over the existing counterparts, when an RIS of size 11x11 is deployed. In the second step, under the same computational complexity of training a single LSM, the ensemble learning with multiple LSMs degrades the prediction variance of a single LSM up to 66% and improves the system achievable SE at most 54%. | ['Ekram Hossain', 'Hina Tabassum', 'Mehdi Rasti', 'Mohamad Robat Mili', 'Narges Gholipoor', 'Hosein Zarini'] | 2022-08-08 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 6.31502032e-01 2.71478415e-01 2.55094916e-01 -7.73392096e-02
-7.90852427e-01 -1.94290981e-01 5.11651993e-01 2.19953246e-03
-2.22091079e-01 7.28926897e-01 1.56367287e-01 -5.08898377e-01
-5.59304476e-01 -1.03954065e+00 -6.71891928e-01 -1.25380003e+00
-1.38719277e-02 1.35170966e-01 1.10982224e-01 -3.28781068e-01
1.99205011e-01 5.03237247e-01 -1.42985773e+00 1.60643801e-01
8.21287036e-01 1.44625068e+00 -3.25918905e-02 6.42926931e-01
-5.08764945e-02 5.64134657e-01 -4.92285460e-01 -1.20404854e-01
2.87765205e-01 -3.26703042e-02 -1.23262726e-01 -5.33453345e-01
3.67671698e-02 -1.62714094e-01 -3.95219147e-01 3.84696722e-01
6.95256650e-01 -3.59199792e-02 7.93308318e-01 -5.20334721e-01
1.87179714e-01 5.99795818e-01 -5.03697515e-01 -1.75141856e-01
-2.30205253e-01 -1.91619247e-01 5.51321685e-01 -4.16189402e-01
9.19245258e-02 7.06923306e-01 8.03524792e-01 3.12943071e-01
-1.01459610e+00 -9.54614282e-01 -1.65638223e-01 -5.29836565e-02
-1.16051304e+00 -6.26241326e-01 7.79567003e-01 -2.73998410e-01
7.26682544e-01 3.51306558e-01 6.06375217e-01 7.95235872e-01
5.89981258e-01 -2.50548683e-02 8.56455743e-01 -6.98455155e-01
2.37496018e-01 -5.09618670e-02 1.09317392e-01 7.53670692e-01
2.34420821e-01 3.20889562e-01 -4.10445213e-01 -1.41189680e-01
6.88349009e-01 -2.25981817e-01 -8.76103193e-02 6.21100552e-02
-1.04841793e+00 4.50284451e-01 4.50915128e-01 3.76228422e-01
-3.59164894e-01 4.21107948e-01 5.95117398e-02 3.27671051e-01
2.83552021e-01 5.55629969e-01 -3.32945019e-01 3.65150012e-02
-8.35183740e-01 -3.90792131e-01 7.99206197e-01 6.69944704e-01
5.82607329e-01 2.11288050e-01 -5.26109517e-01 5.44558942e-01
3.99776578e-01 8.21683764e-01 3.38887155e-01 -6.40616417e-01
5.17888188e-01 4.21716809e-01 4.01266515e-01 -6.89331114e-01
-6.63539112e-01 -1.11448932e+00 -1.00595593e+00 6.44312501e-02
1.27548918e-01 -4.46758389e-01 -1.00323761e+00 1.56825507e+00
4.31248605e-01 3.63943279e-01 3.56176853e-01 4.07276422e-01
4.22352284e-01 9.48849261e-01 -2.10438058e-01 -6.90336347e-01
9.92088854e-01 -6.35973692e-01 -3.22802782e-01 1.29262716e-01
3.93792063e-01 -8.05155098e-01 5.14605224e-01 5.00789583e-01
-6.78578734e-01 -4.13176447e-01 -1.47343147e+00 4.34175700e-01
5.77576682e-02 4.69083279e-01 4.31057692e-01 1.06488657e+00
-6.95660710e-01 6.99631512e-01 -1.08808851e+00 -1.56917140e-01
3.06236535e-01 7.19762981e-01 2.92559564e-01 3.10170054e-01
-7.37967074e-01 4.76675570e-01 -4.22265492e-02 1.73678443e-01
-4.80143249e-01 -1.18377423e+00 4.98167891e-03 1.52479172e-01
4.34330804e-03 -7.04528868e-01 1.03601575e+00 -7.02496171e-01
-2.31294227e+00 1.77096948e-01 -2.44972348e-01 -7.74131894e-01
5.06472528e-01 -2.01021321e-02 -6.45815790e-01 2.37431396e-02
-6.12702429e-01 -1.47124812e-01 9.82157707e-01 -1.13544130e+00
-5.91578424e-01 -1.71522230e-01 -2.65282452e-01 -2.03716934e-01
-6.76936924e-01 -7.37225533e-01 -9.47474763e-02 -2.93631494e-01
4.72099721e-01 -8.69485319e-01 -1.86006218e-01 -5.51464379e-01
-3.21837604e-01 2.63891548e-01 7.41688132e-01 -2.70834744e-01
1.31128109e+00 -1.97452247e+00 -2.03524679e-01 6.23887479e-01
1.37396842e-01 2.98704505e-01 1.39989415e-02 6.98665917e-01
1.41694829e-01 -3.10056031e-01 1.24041950e-02 -8.08503330e-02
-4.21977550e-01 -1.70024946e-01 -3.66656005e-01 3.04549068e-01
-9.77028720e-03 4.91184026e-01 -5.70100367e-01 1.20786250e-01
3.43611389e-01 6.60214186e-01 -4.70775574e-01 1.12250082e-01
-1.88407421e-01 9.47716057e-01 -7.02987850e-01 3.90336156e-01
6.63440347e-01 -3.10274780e-01 4.55696374e-01 -8.63466740e-01
-5.74941337e-01 2.05382526e-01 -8.25883746e-01 1.16580808e+00
-1.02182174e+00 5.29813647e-01 -3.44696760e-01 -1.28165185e+00
1.26160645e+00 2.62043536e-01 1.09759891e+00 -1.03115571e+00
2.48243809e-01 3.67129564e-01 5.60651422e-02 -4.45168227e-01
8.24101791e-02 -1.27415746e-01 5.64779863e-02 3.66032988e-01
-7.61147588e-02 3.09637398e-01 -3.84796828e-01 -3.32127869e-01
1.44708598e+00 -1.36548743e-01 4.71172668e-03 -1.64906844e-01
7.92304933e-01 -3.64591241e-01 2.60440350e-01 8.64639640e-01
4.02329743e-01 6.78851455e-02 2.83190124e-02 -3.66464108e-01
-1.06694663e+00 -1.20269191e+00 -1.91626996e-01 7.23258913e-01
1.01262264e-01 -1.15658104e-01 -4.09291089e-01 -1.07728705e-01
-1.09471403e-01 8.34112763e-01 -2.18043387e-01 -3.56636971e-01
-6.40558541e-01 -1.17412269e+00 3.77764195e-01 7.89118558e-02
7.61896312e-01 -2.55156487e-01 -7.75062561e-01 3.72546673e-01
1.78319812e-01 -1.01504302e+00 3.17718297e-01 3.84579629e-01
-8.79175782e-01 -5.28890669e-01 -4.32282686e-01 -3.09004903e-01
5.43444216e-01 5.75373285e-02 5.63304782e-01 -1.73583746e-01
-1.45505771e-01 4.58924532e-01 -3.16785455e-01 -4.80968595e-01
-3.89210135e-01 2.48788804e-01 -1.03569182e-03 2.46051922e-01
-2.67747134e-01 -1.04096997e+00 -9.51763630e-01 1.64406016e-01
-3.72573793e-01 1.93841249e-01 8.53758574e-01 7.65854716e-01
4.80701476e-01 2.10840121e-01 8.01616848e-01 -7.85168171e-01
-7.66616762e-02 -4.25597906e-01 -1.02169025e+00 4.64040697e-01
-7.52038419e-01 3.90365690e-01 8.15888226e-01 -2.87251145e-01
-1.19805586e+00 -1.51609518e-02 -3.41368467e-01 1.42906159e-01
5.18807411e-01 4.24450725e-01 1.37190551e-01 -6.60757542e-01
6.05545759e-01 1.64945647e-01 -2.39101022e-01 -1.08148992e-01
5.10073490e-02 7.52301633e-01 2.79777139e-01 -5.36611080e-01
9.23100293e-01 4.95774031e-01 5.26225090e-01 -1.01009583e+00
-9.34377730e-01 -2.67420620e-01 -2.45058611e-01 -4.20539141e-01
2.68986017e-01 -9.06824529e-01 -1.11916089e+00 2.38761380e-01
-6.91047251e-01 -4.36570764e-01 -4.21888754e-02 6.59492552e-01
-4.32049364e-01 -6.60496578e-02 -1.29903868e-01 -1.07530332e+00
-8.81366849e-01 -6.04245484e-01 9.94648695e-01 4.03124988e-01
1.85284644e-01 -7.00229704e-01 -2.03995705e-01 1.61139414e-01
6.85286224e-01 4.72105742e-01 1.14749694e+00 -3.00192147e-01
-7.74643958e-01 -3.92834842e-01 -4.82312143e-02 -9.21338350e-02
2.34668419e-01 -1.08570352e-01 -1.00541031e+00 -3.93610626e-01
2.42105499e-01 3.95884007e-01 7.36541927e-01 5.83213568e-01
1.04619443e+00 -1.75590917e-01 -5.37934542e-01 6.58188522e-01
1.70026374e+00 4.91449714e-01 7.07449198e-01 1.13429949e-01
3.92008334e-01 1.54809624e-01 5.04039049e-01 7.37500668e-01
1.68294996e-01 5.08094728e-01 3.79445761e-01 2.98894525e-01
-2.05092281e-02 -5.67546906e-03 3.39200526e-01 9.22203481e-01
-1.19242147e-01 -5.32178044e-01 -6.39470637e-01 -9.90911648e-02
-1.68673050e+00 -8.51413190e-01 -1.29095453e-03 2.77361131e+00
3.46290290e-01 1.43368505e-02 -2.51100898e-01 2.10775137e-01
3.79671097e-01 -1.45532608e-01 -8.05070519e-01 5.08564375e-02
3.06759298e-01 5.81894279e-01 8.82788301e-01 3.30967963e-01
-8.06755304e-01 5.66571832e-01 5.11568213e+00 6.86275065e-01
-1.67545080e+00 1.29244059e-01 4.85422790e-01 -2.53984809e-01
-2.85803437e-01 -1.91025138e-01 -7.98064470e-01 4.42438692e-01
1.41780174e+00 2.53879465e-02 4.71440703e-01 1.75141037e-01
4.51002181e-01 -5.32168411e-02 -7.64511764e-01 9.06406879e-01
-3.80142272e-01 -1.50348580e+00 -2.01423571e-01 1.11317910e-01
7.69234300e-01 3.54411528e-02 2.25942641e-01 1.57199159e-01
-1.54299378e-01 -6.83904231e-01 4.56503958e-01 1.01747811e+00
7.34843969e-01 -8.98891151e-01 5.61191022e-01 3.78080338e-01
-9.91323411e-01 -5.63897729e-01 -1.87352046e-01 1.83538765e-01
3.29031534e-02 1.13418245e+00 -9.59844947e-01 9.04214323e-01
4.54525054e-01 2.16919973e-01 8.39422271e-02 9.99004424e-01
2.18410760e-01 9.26853895e-01 -6.37854934e-01 -4.24723834e-01
-8.58380925e-03 -3.82883757e-01 6.19953573e-01 8.28435302e-01
1.02511489e+00 3.47342461e-01 -3.33003700e-01 2.97590107e-01
4.81603146e-02 -8.47195908e-02 -2.51060426e-02 4.87236194e-02
7.85685182e-01 1.26793909e+00 -4.89610314e-01 1.75574049e-02
-1.58933893e-01 5.43419480e-01 3.67805548e-02 4.49174672e-01
-8.59178662e-01 -3.18828434e-01 4.48996663e-01 3.13160419e-01
3.58817428e-01 -5.23008466e-01 -4.40389216e-01 -6.67560220e-01
-1.87329665e-01 -2.15277195e-01 5.81425382e-03 -3.69297206e-01
-9.49609756e-01 4.24314916e-01 -5.65543890e-01 -1.42827487e+00
-2.32177138e-01 -4.02655780e-01 -3.75917524e-01 7.45082796e-01
-1.49259555e+00 -1.16484272e+00 -4.15434182e-01 5.95124476e-02
-4.01809476e-02 -2.56121963e-01 8.87268662e-01 3.06547463e-01
-5.16351700e-01 7.76871860e-01 8.95527542e-01 -3.81578773e-01
5.30962050e-01 -6.53806090e-01 -5.67067116e-02 5.68851411e-01
-1.77707043e-04 5.53188443e-01 8.15411687e-01 -4.29288357e-01
-1.91719794e+00 -1.30883145e+00 2.23740757e-01 -1.98239759e-02
4.84087437e-01 -2.45071411e-01 -2.89913744e-01 1.32940426e-01
-1.52488589e-01 -2.76148468e-01 7.72495687e-01 4.90596592e-02
-2.95131393e-02 -8.69555235e-01 -1.25871587e+00 5.24519205e-01
8.56522560e-01 -1.01568401e-01 2.64728785e-01 4.69638705e-01
5.51654577e-01 -6.09753609e-01 -1.21347642e+00 7.12612033e-01
1.10340345e+00 -8.58616471e-01 1.01080382e+00 -1.98634155e-02
2.23695502e-01 -3.02104652e-01 -3.02846551e-01 -1.01681554e+00
-3.55906129e-01 -8.20582449e-01 -3.54147255e-01 1.22829425e+00
4.39363033e-01 -9.29272175e-01 9.18655217e-01 2.09477022e-01
-3.47710937e-01 -9.80715513e-01 -1.13233590e+00 -6.76661432e-01
-1.83809370e-01 -3.80513012e-01 5.72976291e-01 1.19034715e-01
-4.68059391e-01 2.96087384e-01 -3.59438747e-01 8.09556425e-01
6.83559597e-01 2.65853345e-01 6.62826478e-01 -1.20434284e+00
-4.78209853e-01 -1.77220508e-01 -2.00486585e-01 -1.09657145e+00
-1.52652368e-01 -6.40705943e-01 9.05487686e-02 -1.24652803e+00
-2.55735159e-01 -9.41558838e-01 -6.81085527e-01 6.69055134e-02
2.45375231e-01 1.30749792e-02 -1.75118923e-01 -4.45770770e-02
-1.77730054e-01 5.99934340e-01 7.99056053e-01 -4.25986797e-02
-4.98941153e-01 6.48960829e-01 -5.07661521e-01 2.86595404e-01
8.29447150e-01 -2.17656568e-01 -4.78710175e-01 -3.95849496e-01
1.96871668e-01 2.37641588e-01 2.71954834e-01 -1.69129753e+00
3.25829029e-01 9.55497473e-02 6.42856061e-01 -4.78299558e-01
4.44019049e-01 -9.81571257e-01 5.37831366e-01 6.74986720e-01
-4.52822372e-02 -7.48783231e-01 2.20532700e-01 9.19780791e-01
4.22220111e-01 -7.11779743e-02 8.46584439e-01 4.63664860e-01
-1.36874288e-01 2.40627497e-01 -3.17376107e-01 -7.26887822e-01
8.35422277e-01 4.60771983e-03 -2.80779779e-01 -2.59608775e-01
-2.05903649e-01 -6.91970736e-02 -9.93801653e-02 -5.39769344e-02
1.78181902e-01 -1.01401091e+00 -1.91257268e-01 4.43078950e-02
-6.26535565e-02 -3.56679142e-01 4.62311596e-01 1.01365447e+00
-1.62161812e-01 5.37514269e-01 8.66283700e-02 -7.25232959e-01
-9.17943716e-01 -6.12883382e-02 4.87197131e-01 -3.69032890e-01
-5.54063797e-01 8.26583087e-01 -2.84102082e-01 -4.73644175e-02
-5.80388755e-02 -2.74649233e-01 -1.84813172e-01 -2.40509212e-01
2.37660110e-01 3.87758940e-01 6.16293669e-01 -1.72042266e-01
-2.09012419e-01 8.52418482e-01 1.17441721e-01 -8.14691857e-02
1.66957664e+00 -1.62174359e-01 1.24831952e-01 3.64141524e-01
1.16716349e+00 3.88439089e-01 -1.29523623e+00 -1.65471464e-01
-6.95833787e-02 -4.95882658e-03 3.09025854e-01 -9.67768252e-01
-1.03383386e+00 5.06084323e-01 1.09293830e+00 1.00137167e-01
1.43572760e+00 -4.48768139e-01 9.53406453e-01 6.27345264e-01
6.80966616e-01 -8.15016270e-01 3.20385993e-02 4.76483315e-01
2.99560487e-01 -9.27425623e-01 5.16785942e-02 -4.18714494e-01
5.99690564e-02 1.48679399e+00 1.74095854e-01 -7.01684356e-02
8.11437309e-01 4.50520813e-01 -2.78423429e-01 9.74471448e-04
-7.21489489e-01 1.62008837e-01 2.34618738e-01 3.54420573e-01
5.19134998e-01 1.46509737e-01 -2.41916448e-01 7.74955511e-01
-2.86732882e-01 -1.43583253e-01 2.04230741e-01 7.39784837e-01
-6.22373760e-01 -1.13004851e+00 -1.06030896e-01 8.66086781e-01
-1.84127524e-01 5.97018898e-02 3.09255242e-01 2.75015950e-01
-2.18878146e-02 1.15389347e+00 -1.16795665e-02 -7.75937021e-01
3.37340921e-01 -1.63986161e-01 7.93773592e-01 -2.87055999e-01
-3.21859658e-01 5.02349101e-02 5.33629507e-02 -4.60863143e-01
-3.19326103e-01 -5.04712760e-01 -1.29553246e+00 -1.05665803e-01
-3.34651351e-01 2.43421923e-02 8.55532408e-01 1.27688360e+00
6.16145134e-01 8.97508740e-01 1.09246826e+00 -6.00785434e-01
-5.36171854e-01 -6.54534340e-01 -1.77231312e-01 -6.23907089e-01
4.39398944e-01 -5.88467777e-01 -1.32636964e-01 -2.47134507e-01] | [6.094210147857666, 1.4617098569869995] |
0b637763-d13e-4b4c-9ee5-360ccd3349d5 | joint-face-detection-and-facial-motion | 1902.10744 | null | http://arxiv.org/abs/1902.10744v1 | http://arxiv.org/pdf/1902.10744v1.pdf | Joint Face Detection and Facial Motion Retargeting for Multiple Faces | Facial motion retargeting is an important problem in both computer graphics
and vision, which involves capturing the performance of a human face and
transferring it to another 3D character. Learning 3D morphable model (3DMM)
parameters from 2D face images using convolutional neural networks is common in
2D face alignment, 3D face reconstruction etc. However, existing methods either
require an additional face detection step before retargeting or use a cascade
of separate networks to perform detection followed by retargeting in a
sequence. In this paper, we present a single end-to-end network to jointly
predict the bounding box locations and 3DMM parameters for multiple faces.
First, we design a novel multitask learning framework that learns a
disentangled representation of 3DMM parameters for a single face. Then, we
leverage the trained single face model to generate ground truth 3DMM parameters
for multiple faces to train another network that performs joint face detection
and motion retargeting for images with multiple faces. Experimental results
show that our joint detection and retargeting network has high face detection
accuracy and is robust to extreme expressions and poses while being faster than
state-of-the-art methods. | ['Baoyuan Wang', 'Noranart Vesdapunt', 'Bindita Chaudhuri'] | 2019-02-27 | joint-face-detection-and-facial-motion-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Chaudhuri_Joint_Face_Detection_and_Facial_Motion_Retargeting_for_Multiple_Faces_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Chaudhuri_Joint_Face_Detection_and_Facial_Motion_Retargeting_for_Multiple_Faces_CVPR_2019_paper.pdf | cvpr-2019-6 | ['motion-retargeting'] | ['computer-vision'] | [ 1.18415214e-01 8.63738284e-02 -8.98744091e-02 -4.92494375e-01
-5.04594564e-01 -4.84580666e-01 4.32837605e-01 -8.67287934e-01
-3.07066500e-01 1.72429264e-01 -5.02393991e-02 -1.20321047e-02
5.09311318e-01 -1.69669494e-01 -8.22045326e-01 -5.78705251e-01
1.23526782e-01 6.24958336e-01 -3.27431154e-03 2.57303454e-02
2.18247145e-01 9.99667346e-01 -1.49473894e+00 1.97717160e-01
-1.51412889e-01 8.27614486e-01 -2.29078308e-01 9.43903506e-01
1.06475227e-01 4.30631340e-01 -3.32126528e-01 -4.60229993e-01
6.45691812e-01 -2.12362781e-01 -5.88316858e-01 4.55214441e-01
1.26292443e+00 -7.39562452e-01 -1.77848011e-01 8.91752660e-01
7.83631384e-01 7.79738277e-02 7.34516323e-01 -1.39862740e+00
-4.58059430e-01 -7.00097904e-02 -1.43005776e+00 6.21003918e-02
6.01179242e-01 1.81084648e-01 2.41102934e-01 -1.40247893e+00
6.13896191e-01 1.88783216e+00 7.35621393e-01 1.26443887e+00
-1.21129656e+00 -1.25884473e+00 1.29044488e-01 -2.14041069e-01
-1.57049453e+00 -1.10409033e+00 8.56782019e-01 -6.90506041e-01
9.01345015e-01 -6.89266771e-02 4.39919055e-01 1.19263649e+00
1.06946662e-01 3.27012360e-01 6.79191649e-01 -3.53490740e-01
-2.52134055e-01 -1.03602968e-01 -5.09843588e-01 1.33405292e+00
1.18651077e-01 1.37624577e-01 -6.60575211e-01 -5.06326705e-02
1.28421044e+00 -1.07105859e-01 -1.02743069e-02 -5.03879726e-01
-7.29516268e-01 6.66834414e-01 -1.86505504e-02 -2.81692863e-01
-6.73343837e-02 3.04877818e-01 1.36496961e-01 1.51170164e-01
6.08813405e-01 -2.49038920e-01 -4.60554302e-01 3.37282300e-01
-1.07480955e+00 3.18208873e-01 5.97966015e-01 8.38913798e-01
1.08597672e+00 9.82833579e-02 -1.81553617e-01 6.65796638e-01
7.98694432e-01 5.58082640e-01 2.70135432e-01 -9.89067912e-01
2.99511969e-01 5.28195918e-01 2.77642366e-02 -1.13407266e+00
-4.42061156e-01 2.25727320e-01 -5.55164337e-01 6.85118079e-01
3.12592983e-01 -4.07616287e-01 -1.20274246e+00 1.87506032e+00
1.00836074e+00 5.96873701e-01 -2.39917755e-01 1.04412150e+00
9.57728446e-01 2.25835800e-01 -1.90310180e-01 -8.95440280e-02
1.17789197e+00 -9.55981970e-01 -3.48498732e-01 -3.37672055e-01
4.26200271e-01 -1.02140868e+00 5.74246824e-01 -1.25648221e-02
-1.16912889e+00 -5.80338776e-01 -8.16940188e-01 -1.50545940e-01
-3.44108976e-02 3.86493295e-01 2.95358151e-01 8.75668526e-01
-1.26287377e+00 2.53347754e-01 -7.58212745e-01 -4.44159210e-01
7.65876174e-01 8.76132190e-01 -9.65440989e-01 7.51699880e-02
-4.12505209e-01 1.00251734e+00 -9.23136994e-02 3.58897716e-01
-1.35071206e+00 -7.01447248e-01 -1.04695642e+00 -3.75994682e-01
2.69693106e-01 -8.03505063e-01 1.25169265e+00 -1.01381159e+00
-1.91146398e+00 1.55866992e+00 -2.21325025e-01 2.50436038e-01
6.16816700e-01 -3.06161463e-01 -2.08733529e-01 4.30541020e-03
6.49757870e-03 1.05943072e+00 1.63693559e+00 -1.02740753e+00
-3.32419336e-01 -7.27432549e-01 -2.69912302e-01 1.54525623e-01
-1.30857661e-01 6.07927442e-01 -7.11137354e-01 -4.74397033e-01
-5.54908291e-02 -9.94974673e-01 6.99130073e-02 6.28370464e-01
-4.90200549e-01 -7.14252964e-02 1.24509823e+00 -6.31429553e-01
5.37819803e-01 -1.93501997e+00 4.61859077e-01 -2.11511683e-02
3.25181872e-01 4.31269944e-01 -5.39273202e-01 -2.89644808e-01
-2.87223428e-01 -3.92411016e-02 5.70348576e-02 -9.71179962e-01
-1.70731470e-01 -2.09368795e-01 8.63399208e-02 9.39570606e-01
5.89990377e-01 9.46392298e-01 -4.23670173e-01 -5.25974929e-01
1.92103043e-01 7.64324367e-01 -7.10916877e-01 3.87815028e-01
-6.50803670e-02 7.92638063e-01 -1.31261751e-01 8.23523760e-01
1.10412717e+00 3.10825668e-02 7.76153654e-02 -2.64588773e-01
1.42222509e-01 -3.47102314e-01 -1.28951335e+00 1.75229073e+00
-2.85672367e-01 5.94683290e-01 4.04762268e-01 -6.37049854e-01
9.65871990e-01 3.25217873e-01 5.10255516e-01 -1.43279567e-01
5.17120421e-01 3.82732488e-02 -3.33053976e-01 -5.71094692e-01
2.12266902e-03 -2.00677067e-01 2.02995330e-01 6.59837723e-01
2.10377395e-01 -1.31728217e-01 -4.08805430e-01 -3.58192116e-01
6.09132528e-01 5.71676970e-01 1.73434764e-01 -4.04426567e-02
6.21944368e-01 -5.61517358e-01 6.57241583e-01 -2.22671088e-02
-2.95092195e-01 6.64012790e-01 1.97475851e-01 -8.05505753e-01
-1.03604710e+00 -7.49005377e-01 1.93299904e-01 1.17832005e+00
-8.54351521e-02 4.71315309e-02 -9.20452237e-01 -9.01283264e-01
2.61969566e-01 -1.11871548e-01 -9.94528055e-01 -1.62379727e-01
-8.31225038e-01 -5.94835043e-01 7.25440919e-01 4.95934665e-01
4.85970050e-01 -6.52642250e-01 -6.59552753e-01 -2.41935775e-01
2.21043974e-01 -1.24525106e+00 -1.16787720e+00 -3.60617161e-01
-5.06416440e-01 -1.16327488e+00 -6.86059594e-01 -9.04264808e-01
9.79125321e-01 3.19274634e-01 7.57556915e-01 2.08618730e-01
-6.77392602e-01 4.56074744e-01 1.09442517e-01 -3.75882417e-01
-2.62471199e-01 -1.99247614e-01 4.74780619e-01 4.97065604e-01
3.54014724e-01 -3.94029737e-01 -5.44411123e-01 4.86339420e-01
-4.45567936e-01 2.98587773e-02 3.44772726e-01 3.97890151e-01
3.51197183e-01 -5.22002757e-01 2.00556532e-01 -6.62437201e-01
7.43143409e-02 -3.04026064e-02 -8.42034161e-01 2.40294084e-01
-1.54548362e-01 -4.36656103e-02 -5.08642793e-02 -7.45166898e-01
-9.32319283e-01 6.77780688e-01 5.76333180e-02 -1.11448598e+00
-1.82962418e-01 -4.66445029e-01 -4.00857300e-01 -7.93967664e-01
5.65911114e-01 -2.19330370e-01 3.58701259e-01 -4.93915826e-01
4.87104207e-01 4.97152507e-01 5.16187429e-01 -2.90982634e-01
9.82513964e-01 5.42348504e-01 3.19252074e-01 -8.79208744e-01
-7.63236761e-01 -1.53396040e-01 -1.17264748e+00 -4.49095637e-01
1.12452114e+00 -1.03651190e+00 -9.66872156e-01 6.71182394e-01
-1.46063566e+00 -2.33857766e-01 2.21960321e-01 2.67395973e-01
-5.06628215e-01 1.63689166e-01 -2.14128405e-01 -6.58726871e-01
-3.94136757e-01 -1.25079346e+00 1.73486841e+00 2.60548204e-01
-2.08657071e-01 -9.35942829e-01 1.14773512e-01 3.53440881e-01
1.53569132e-01 4.43831563e-01 4.72418338e-01 -2.32929662e-01
-4.12347019e-01 -3.37869436e-01 -1.77713215e-01 3.96102518e-02
3.34601492e-01 2.69653648e-01 -1.21954012e+00 -5.93842089e-01
-1.44806653e-01 -3.82397234e-01 7.09447265e-01 3.88822347e-01
9.18947995e-01 -4.72413629e-01 -5.51081538e-01 1.09130800e+00
9.11448538e-01 -1.15105249e-01 2.32133701e-01 -1.70948967e-01
1.08041811e+00 7.68255115e-01 1.82113782e-01 4.74137247e-01
3.50776553e-01 9.59924459e-01 6.81275547e-01 -3.51735428e-02
-3.26365292e-01 -1.98627308e-01 6.53948963e-01 1.49291605e-01
-2.98942290e-02 2.02529103e-01 -6.87368035e-01 5.72409779e-02
-1.55960464e+00 -6.55170679e-01 3.50375116e-01 2.34379840e+00
6.10356629e-01 -4.60886419e-01 3.95336479e-01 -3.76170844e-01
1.02583921e+00 -1.24922656e-01 -7.34889150e-01 -1.81333601e-01
-7.40783708e-03 1.20523296e-01 2.61803329e-01 7.91376173e-01
-1.28145576e+00 1.36137724e+00 6.34893179e+00 2.77304143e-01
-1.40492737e+00 2.59617209e-01 5.70460439e-01 -4.92071450e-01
9.47664306e-02 -2.16786385e-01 -1.31491959e+00 1.04867131e-01
5.49801886e-01 1.91646919e-01 3.73091906e-01 4.85769153e-01
5.71348444e-02 2.58694530e-01 -1.38468182e+00 1.57645524e+00
6.10209286e-01 -1.19099629e+00 2.49657363e-01 1.48083687e-01
8.08591962e-01 -4.10401076e-01 2.28708014e-01 -3.89185436e-02
1.42515659e-01 -1.38441074e+00 7.37035871e-01 5.20911753e-01
1.07833874e+00 -6.35647178e-01 2.94106692e-01 1.71137437e-01
-1.24200439e+00 9.19962898e-02 -3.71626288e-01 1.15742125e-01
-1.59155391e-02 4.77058860e-03 -1.08583403e+00 1.10741600e-01
5.41606367e-01 5.58129013e-01 -5.65711856e-01 7.36202121e-01
-3.36760059e-02 -1.63473070e-01 -2.04717085e-01 4.40789223e-01
-2.86202997e-01 1.08285822e-01 6.74062848e-01 1.04049253e+00
3.90323251e-01 -8.19601715e-02 2.23279640e-01 7.78487027e-01
-4.99137372e-01 -5.55057451e-02 -7.82349229e-01 1.42198011e-01
3.89649272e-01 1.62794805e+00 -5.85743725e-01 7.73540810e-02
-3.10573250e-01 1.40010560e+00 4.86229271e-01 3.04897308e-01
-8.17303061e-01 1.33401573e-01 1.00538278e+00 1.69342175e-01
3.92004073e-01 -3.14723015e-01 2.26992130e-01 -1.16600871e+00
-1.66650966e-01 -6.32730544e-01 1.20444044e-01 -4.87723827e-01
-1.11921382e+00 6.13259375e-01 -1.26908913e-01 -8.74145329e-01
-4.03013676e-01 -9.11580741e-01 -6.11782432e-01 9.17463481e-01
-1.31686735e+00 -1.65220892e+00 -4.67432320e-01 8.67672622e-01
5.32780170e-01 -4.26650405e-01 7.13828683e-01 2.63213158e-01
-1.01933444e+00 1.05319703e+00 -7.59898424e-01 3.60232174e-01
1.00766897e+00 -7.52281189e-01 5.42538106e-01 7.32979059e-01
9.65217724e-02 3.52603495e-01 3.77054423e-01 -6.13877356e-01
-1.74543738e+00 -1.42442858e+00 4.85296190e-01 -9.47133780e-01
-1.41171617e-02 -5.96684873e-01 -5.81126213e-01 1.04140270e+00
-4.75195199e-02 4.82401162e-01 5.18454552e-01 -3.14870358e-01
-5.47857285e-01 -3.08400150e-02 -1.35830688e+00 6.12522483e-01
1.18733704e+00 -4.76528436e-01 -1.28327534e-01 2.95540452e-01
4.76462752e-01 -7.99951375e-01 -4.32114720e-01 3.18497300e-01
8.65468323e-01 -6.83842361e-01 9.74378228e-01 -9.33572173e-01
-7.20096678e-02 -4.22309428e-01 -2.20580801e-01 -8.54103982e-01
-1.20609142e-01 -1.05383885e+00 -3.61121804e-01 1.09530151e+00
1.66162580e-01 -2.73176581e-01 1.14802611e+00 7.07035840e-01
3.49459082e-01 -6.27924263e-01 -1.17952156e+00 -4.41682637e-01
2.98425630e-02 -1.49827421e-01 6.17032647e-01 8.54628146e-01
-5.25311708e-01 5.24903476e-01 -6.87099636e-01 4.82532442e-01
8.21538091e-01 9.01125278e-03 1.26180696e+00 -1.25789797e+00
1.30132720e-01 -4.09962803e-01 -6.50037408e-01 -9.12284851e-01
6.91150308e-01 -8.93846750e-01 -9.67919454e-02 -8.36122692e-01
3.45974863e-01 1.65515840e-01 4.38495457e-01 8.08180988e-01
-3.58558558e-02 7.23035693e-01 6.91968575e-02 -7.42291007e-03
-2.55994111e-01 5.25171161e-01 1.28253126e+00 -1.83667373e-02
-3.17694455e-01 -6.40168190e-02 -5.58710217e-01 8.78563464e-01
3.60875249e-01 -3.67220670e-01 -9.09288600e-02 -5.91927230e-01
-1.06948234e-01 -3.33901867e-02 6.71266139e-01 -6.23329163e-01
2.85264343e-01 -1.64814636e-01 1.07646775e+00 -3.56596887e-01
6.92040682e-01 -6.44166052e-01 5.49446344e-02 2.12539047e-01
1.19099654e-01 2.24565580e-01 5.46640933e-01 3.81227314e-01
3.82183909e-01 2.12615617e-02 1.23862720e+00 -9.02916566e-02
-6.30539179e-01 9.50455070e-01 -1.00565642e-01 -1.43579230e-01
1.35178006e+00 -4.06730831e-01 1.20363042e-01 -1.82930812e-01
-8.28062654e-01 1.08073177e-02 5.93115509e-01 7.90400088e-01
9.67934787e-01 -1.40886712e+00 -9.76297498e-01 6.73702776e-01
-2.23400742e-01 2.04963740e-02 1.31430328e-01 6.87569797e-01
-4.86754924e-01 3.87752764e-02 -4.10590947e-01 -8.01079631e-01
-1.72965515e+00 5.03260195e-01 8.73891592e-01 5.56778848e-01
-4.47210014e-01 1.37806678e+00 2.98583835e-01 -6.99068606e-01
1.70928955e-01 1.98247999e-01 -1.25807896e-01 5.96425273e-02
6.41451597e-01 3.14148247e-01 -3.93049717e-02 -1.23093486e+00
-6.35179877e-01 1.39178860e+00 -2.25795269e-01 -1.55578136e-01
1.13689303e+00 -1.60180286e-01 -1.75268069e-01 -2.51662564e-02
1.40166998e+00 -2.11244747e-01 -1.65453851e+00 -1.44948512e-01
-3.27194780e-01 -8.14079106e-01 -1.91231482e-02 -3.14865142e-01
-1.40150845e+00 1.00382757e+00 8.27316582e-01 -8.36215198e-01
1.04310906e+00 -3.87307629e-02 3.24486554e-01 2.00761542e-01
2.43393138e-01 -7.34246075e-01 5.48697233e-01 3.09337795e-01
1.10867286e+00 -1.35595953e+00 -1.05572194e-01 -3.98363322e-01
-3.78391773e-01 1.29645967e+00 1.11605763e+00 -5.11360122e-03
7.95508206e-01 3.09810579e-01 2.22234711e-01 -2.78822631e-01
-5.64599276e-01 1.50465742e-01 4.78382796e-01 6.02503419e-01
2.48967364e-01 -2.67387360e-01 6.81678891e-01 9.94004160e-02
-4.10347842e-02 -1.08303718e-01 2.44245663e-01 7.40355253e-01
-3.24423581e-01 -1.17551982e+00 -5.83841264e-01 -1.37471613e-02
-3.97610784e-01 3.22047114e-01 -6.55163467e-01 6.84902191e-01
2.30615124e-01 7.36582041e-01 1.79467618e-01 -4.35336381e-01
2.22459540e-01 1.99852973e-01 1.12328291e+00 -9.92969275e-01
-1.71551988e-01 1.36930928e-01 -4.68849540e-01 -7.17409134e-01
-4.20591235e-01 -7.26362348e-01 -1.00981927e+00 -3.75119478e-01
-2.78212070e-01 -5.20843744e-01 5.43181539e-01 9.51991498e-01
6.17319167e-01 3.03733554e-02 7.78503060e-01 -1.59773910e+00
-3.21994394e-01 -1.10903442e+00 -3.87081981e-01 2.16532294e-02
8.55155647e-01 -1.06077051e+00 -1.44794479e-01 1.83032021e-01] | [13.262887001037598, 0.18482699990272522] |
816d14a5-f7b5-489e-b829-5e52a48e2b16 | metaadapt-domain-adaptive-few-shot | 2305.12692 | null | https://arxiv.org/abs/2305.12692v1 | https://arxiv.org/pdf/2305.12692v1.pdf | MetaAdapt: Domain Adaptive Few-Shot Misinformation Detection via Meta Learning | With emerging topics (e.g., COVID-19) on social media as a source for the spreading misinformation, overcoming the distributional shifts between the original training domain (i.e., source domain) and such target domains remains a non-trivial task for misinformation detection. This presents an elusive challenge for early-stage misinformation detection, where a good amount of data and annotations from the target domain is not available for training. To address the data scarcity issue, we propose MetaAdapt, a meta learning based approach for domain adaptive few-shot misinformation detection. MetaAdapt leverages limited target examples to provide feedback and guide the knowledge transfer from the source to the target domain (i.e., learn to adapt). In particular, we train the initial model with multiple source tasks and compute their similarity scores to the meta task. Based on the similarity scores, we rescale the meta gradients to adaptively learn from the source tasks. As such, MetaAdapt can learn how to adapt the misinformation detection model and exploit the source data for improved performance in the target domain. To demonstrate the efficiency and effectiveness of our method, we perform extensive experiments to compare MetaAdapt with state-of-the-art baselines and large language models (LLMs) such as LLaMA, where MetaAdapt achieves better performance in domain adaptive few-shot misinformation detection with substantially reduced parameters on real-world datasets. | ['Dong Wang', 'Lanyu Shang', 'Yang Zhang', 'Huimin Zeng', 'Zhenrui Yue'] | 2023-05-22 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [ 1.54055834e-01 -1.53798729e-01 -5.06999075e-01 -2.50568420e-01
-1.02297151e+00 -4.44879800e-01 1.00744236e+00 1.81128681e-01
-5.22065222e-01 6.48552775e-01 4.93738383e-01 2.97759417e-02
1.06082276e-01 -6.13075197e-01 -5.16269922e-01 -4.72093284e-01
1.04986981e-01 6.44489408e-01 4.54171658e-01 -4.52800959e-01
4.63675112e-01 -2.97470778e-01 -1.01732779e+00 3.46053809e-01
1.42327964e+00 6.20814860e-01 2.89032668e-01 4.94705617e-01
-3.59768927e-01 4.95063752e-01 -7.69285262e-01 -3.16090494e-01
2.16751527e-02 -4.80708003e-01 -6.52995944e-01 7.13803107e-03
4.25347805e-01 -3.99441868e-01 -3.36418480e-01 1.42914355e+00
5.63239813e-01 2.70156443e-01 1.06683266e+00 -8.80898535e-01
-9.08207715e-01 9.09288406e-01 -7.08702564e-01 7.94906795e-01
-8.48703608e-02 2.41232395e-01 7.12515533e-01 -1.36624575e+00
6.77968025e-01 1.44809258e+00 4.79438126e-01 6.21335089e-01
-1.09390438e+00 -9.97501731e-01 4.66665059e-01 3.43736440e-01
-1.31206834e+00 -5.43555975e-01 7.19700634e-01 -4.96217549e-01
7.51295507e-01 -2.58033931e-01 5.04643433e-02 1.78485417e+00
1.29413217e-01 8.70012879e-01 7.03848362e-01 -8.99123922e-02
5.38819730e-01 4.20053631e-01 2.85114229e-01 3.41428548e-01
1.80507272e-01 9.45020914e-02 -6.15658581e-01 -3.93983841e-01
2.66074061e-01 1.64635509e-01 -3.42543935e-03 2.16108393e-02
-1.08258319e+00 1.05577922e+00 5.28070986e-01 4.63404775e-01
-3.40342045e-01 -2.94271111e-01 4.14735138e-01 3.74522686e-01
1.26818049e+00 5.32720208e-01 -4.16925251e-01 -4.83899973e-02
-8.14219475e-01 2.67705351e-01 6.12468898e-01 8.95730197e-01
9.40020144e-01 1.13859050e-01 -2.66407162e-01 1.29447341e+00
4.22820486e-02 7.61348069e-01 8.36814821e-01 -4.88201261e-01
9.18367445e-01 6.24207199e-01 2.38795042e-01 -9.94012535e-01
-2.88566083e-01 -7.55381405e-01 -8.34228694e-01 -1.50433972e-01
2.36341298e-01 -4.55885947e-01 -1.05511653e+00 1.71864665e+00
4.59899634e-01 6.03454769e-01 4.49866168e-02 8.59875679e-01
4.46655184e-01 9.01891828e-01 1.95101470e-01 -8.91250521e-02
1.02627969e+00 -1.08763838e+00 -3.95493060e-01 -9.70899343e-01
7.47107267e-01 -5.02155304e-01 9.90606189e-01 1.49854720e-01
-6.08290493e-01 -2.08046272e-01 -9.71795738e-01 3.25047940e-01
-5.01350820e-01 -3.36847007e-01 5.88340312e-02 1.34654701e-01
-4.80376214e-01 3.35054725e-01 -3.49242091e-01 -5.24359465e-01
5.06841362e-01 -3.35214764e-01 1.69638172e-01 -3.58707577e-01
-1.75680530e+00 6.81810439e-01 6.60157859e-01 -6.77258193e-01
-1.44310272e+00 -1.00004911e+00 -5.74990869e-01 6.10367246e-02
7.28166819e-01 -6.97239816e-01 1.39221478e+00 -1.28835809e+00
-1.18893611e+00 7.97312558e-01 -5.24170883e-02 -6.67511582e-01
5.02048075e-01 -4.12907690e-01 -6.81251585e-01 -8.05887803e-02
5.50856769e-01 3.37757081e-01 1.42776370e+00 -1.16121364e+00
-1.03102887e+00 -2.58551657e-01 -2.45022357e-01 3.40230912e-01
-7.51780033e-01 -2.72957571e-02 -1.96312502e-01 -9.44012761e-01
-2.75676250e-01 -5.84228218e-01 -8.70254934e-02 -3.76664251e-01
-3.95924598e-01 -2.73008466e-01 7.61123717e-01 -4.14114892e-01
1.42702055e+00 -2.21918941e+00 2.75263414e-02 -2.07908787e-02
5.52840889e-01 7.03025222e-01 -2.74852127e-01 3.33868176e-01
2.80185223e-01 1.23611540e-01 -4.03332561e-01 -3.83257776e-01
-1.93738192e-01 5.31223342e-02 -5.64227581e-01 3.53234082e-01
2.89125014e-02 7.73529232e-01 -1.32107055e+00 -1.92949921e-01
-1.33627236e-01 9.84861478e-02 -4.90487427e-01 4.07288373e-01
-5.01910448e-01 2.14834332e-01 -6.49086714e-01 4.67049450e-01
7.86614001e-01 -5.92278302e-01 -2.13066489e-01 2.90599018e-01
9.69926566e-02 5.38230419e-01 -7.65813708e-01 1.57570648e+00
-4.87325788e-01 4.01489168e-01 1.67208433e-01 -9.45976198e-01
8.06658328e-01 2.55025085e-02 2.76444584e-01 -7.85701394e-01
7.40159303e-02 2.38471538e-01 -2.03809068e-01 -3.84135276e-01
4.67699796e-01 -2.25672901e-01 -2.62785137e-01 7.84549952e-01
1.57411605e-01 4.63623941e-01 -6.72153234e-02 6.44269586e-01
1.15621734e+00 -6.99489236e-01 5.06703436e-01 -1.42133951e-01
2.74241835e-01 2.75831878e-01 5.76822102e-01 1.19383812e+00
-2.79987901e-01 4.07499492e-01 -3.60210873e-02 -1.29160121e-01
-1.06905425e+00 -1.24676895e+00 4.42243256e-02 1.78211403e+00
2.80607343e-01 -2.42276296e-01 -7.18177438e-01 -1.30350924e+00
2.00591549e-01 1.01970613e+00 -6.54864967e-01 -6.45023286e-01
-3.51676017e-01 -9.40565705e-01 3.96827519e-01 1.34350196e-01
6.51536882e-01 -7.29054809e-01 6.02085441e-02 4.46264416e-01
-3.15142632e-01 -8.60278428e-01 -8.74579847e-01 -1.24845199e-01
-7.46544957e-01 -8.82097244e-01 -8.85051191e-01 -5.08978009e-01
6.07135057e-01 8.58234525e-01 1.19397354e+00 -1.81823879e-01
8.34287778e-02 2.52243936e-01 -3.79886627e-01 -3.65515500e-01
-7.98770189e-01 2.74944305e-01 2.37520918e-01 8.67897198e-02
6.05785906e-01 -5.41625023e-01 -6.32109523e-01 4.69135702e-01
-1.02053630e+00 -1.69051543e-01 6.34501874e-01 9.67299402e-01
1.68803073e-02 1.16747515e-02 1.06770158e+00 -1.35217488e+00
1.04685986e+00 -1.37905133e+00 -3.02655220e-01 1.07063994e-01
-8.83364499e-01 1.63519345e-02 5.68985045e-01 -8.02629411e-01
-1.34792864e+00 -6.54824018e-01 1.27500519e-01 -4.85311776e-01
-9.13537294e-02 7.61525393e-01 1.83824852e-01 3.00484568e-01
1.28901184e+00 3.95459741e-01 -3.68089765e-01 -7.20307827e-01
5.09940207e-01 9.87608433e-01 4.43711638e-01 -1.19158477e-01
8.83294404e-01 3.76128942e-01 -9.40649450e-01 -6.72674000e-01
-1.35165524e+00 -9.04759228e-01 -3.66681904e-01 -1.87604073e-02
4.21173513e-01 -1.03765357e+00 2.61105031e-01 5.82277954e-01
-1.17391276e+00 -1.93136796e-01 -5.46344332e-02 2.03007475e-01
-1.77702814e-01 2.63091683e-01 -4.12076831e-01 -6.87027633e-01
-5.21719158e-01 -5.77351987e-01 8.40309322e-01 4.74334098e-02
-1.47654846e-01 -1.47421062e+00 4.26180035e-01 1.09495573e-01
7.57445991e-01 -3.47050339e-01 9.39639330e-01 -1.33544481e+00
-3.03892553e-01 -5.37984557e-02 -4.08887088e-01 2.68937558e-01
3.82424116e-01 -7.07135618e-01 -9.51977432e-01 -3.63716215e-01
1.87937975e-01 -4.85696435e-01 1.40322304e+00 1.04834311e-01
6.44649148e-01 -7.92151630e-01 -5.80238104e-01 3.25805932e-01
1.10252094e+00 -1.95478603e-01 4.79187258e-03 3.93645972e-01
4.69286978e-01 3.71341616e-01 8.06816339e-01 6.07729852e-01
5.31873167e-01 5.40809631e-01 1.54277429e-01 2.89590716e-01
-3.09080750e-01 -6.98581040e-01 8.62538636e-01 9.92059886e-01
7.67396808e-01 -3.62713069e-01 -1.12255621e+00 7.23121643e-01
-1.99595726e+00 -9.73339498e-01 4.89861667e-01 2.14473724e+00
1.15782547e+00 3.06941718e-01 1.99260384e-01 -6.10220730e-01
1.05181777e+00 4.00634050e-01 -1.09627426e+00 1.10644437e-01
2.17646863e-02 -6.09013617e-01 3.24975491e-01 5.52539945e-01
-1.31322765e+00 1.18545151e+00 6.07249832e+00 1.26401746e+00
-1.19797862e+00 6.59154475e-01 4.74250764e-01 -1.83695555e-01
-3.82836819e-01 -2.42054239e-01 -1.14300144e+00 9.43525374e-01
1.05983531e+00 -7.37137556e-01 4.39878404e-01 1.08481181e+00
-4.92695495e-02 9.21879560e-02 -7.42136240e-01 7.05432832e-01
2.94803649e-01 -1.18155193e+00 2.34362394e-01 -2.03347832e-01
1.13903999e+00 7.35620260e-01 2.27906838e-01 7.96359718e-01
7.98377991e-01 -5.24807215e-01 3.65845442e-01 2.52743185e-01
5.00946760e-01 -5.56517005e-01 5.62835753e-01 1.06058931e+00
-7.01306581e-01 -3.48621815e-01 -6.75584793e-01 2.80924529e-01
1.15252353e-01 9.14823651e-01 -1.51734447e+00 1.00489102e-01
3.14224541e-01 1.17637157e+00 -5.83431304e-01 1.01260221e+00
-6.33466318e-02 9.32836533e-01 -2.00758904e-01 -5.48913255e-02
4.18302327e-01 1.56873956e-01 1.19897306e+00 1.46126676e+00
2.87000269e-01 -2.64581323e-01 4.34758067e-01 1.13136613e+00
-4.01880920e-01 -4.26203236e-02 -6.99963987e-01 -2.30030596e-01
7.94852793e-01 9.41267908e-01 7.22337365e-02 -6.36421382e-01
-3.23936194e-01 1.06633127e+00 6.27864361e-01 7.06904709e-01
-5.20581603e-01 -2.19372168e-01 9.74331617e-01 2.41724506e-01
1.10672437e-01 1.51078384e-02 -2.17799082e-01 -1.53888214e+00
-3.78496617e-01 -8.76388609e-01 8.15103710e-01 -1.93250015e-01
-2.06302667e+00 3.68491739e-01 1.52241290e-01 -1.29193664e+00
-5.82632303e-01 -1.10885933e-01 -8.16989601e-01 8.52549374e-01
-1.72070515e+00 -8.38168561e-01 -7.17121288e-02 5.30425549e-01
1.08624625e+00 -5.33983231e-01 4.36615556e-01 1.66459903e-01
-4.88852173e-01 7.77258754e-01 5.83547652e-01 3.93921956e-02
1.15120423e+00 -1.14740360e+00 7.90039361e-01 8.35141003e-01
-7.17033967e-02 5.77692449e-01 6.84038281e-01 -1.01390195e+00
-9.30108309e-01 -1.56062853e+00 7.04873443e-01 -5.58799446e-01
9.23875928e-01 -4.86741364e-01 -1.37081873e+00 5.94781041e-01
1.05486497e-01 -2.64629602e-01 4.60355371e-01 2.94399530e-01
-6.88647568e-01 -9.10614152e-03 -1.09376025e+00 5.12556255e-01
1.07349467e+00 -5.33730567e-01 -8.75865102e-01 5.29141128e-01
8.63685906e-01 -9.66837928e-02 -2.37828985e-01 5.02341287e-03
-5.95316179e-02 -6.87980175e-01 1.05196679e+00 -9.56658781e-01
4.50067163e-01 1.95415139e-01 4.88389693e-02 -2.14678884e+00
-5.64655721e-01 -4.49366331e-01 -3.79742712e-01 1.35125673e+00
5.40884256e-01 -8.26085508e-01 4.78581607e-01 2.41848364e-01
1.65334325e-02 -2.52109528e-01 -1.00964940e+00 -8.31076443e-01
4.18216646e-01 -2.95696676e-01 3.13253284e-01 1.12392092e+00
3.87016088e-02 6.14640594e-01 -6.15482032e-01 1.69190705e-01
9.25130069e-01 7.41782412e-02 6.91187978e-01 -1.10634267e+00
-2.10278407e-01 -3.57299954e-01 1.04836054e-01 -1.42026794e+00
2.89940953e-01 -1.09369755e+00 2.20410883e-01 -1.12150908e+00
5.20741999e-01 -4.28684503e-01 -7.15509295e-01 2.40442082e-01
-7.30585515e-01 -1.76173866e-01 1.68352835e-02 5.75588942e-01
-1.08393395e+00 8.12927663e-01 8.78072798e-01 -6.60415649e-01
-2.39269972e-01 5.69344088e-02 -1.00270581e+00 8.72760415e-01
7.62069345e-01 -1.04064834e+00 -4.33296263e-01 -4.90887940e-01
1.91827580e-01 -2.96263456e-01 2.56936345e-02 -7.58244514e-01
4.42632884e-01 -3.43938351e-01 7.34798387e-02 -2.40906283e-01
8.04630741e-02 -3.38589221e-01 -5.90706706e-01 2.68550664e-01
-7.24932313e-01 -3.67884159e-01 -9.68106240e-02 1.12786674e+00
-6.59713000e-02 -3.45196307e-01 9.26684797e-01 -3.54601771e-01
-1.08770895e+00 5.15100777e-01 -4.93229061e-01 8.49433243e-01
7.60721922e-01 3.69648188e-01 -9.03411448e-01 -5.39355397e-01
-4.31682557e-01 4.72538859e-01 4.36837375e-01 8.86017144e-01
8.10985446e-01 -1.35627675e+00 -1.07431602e+00 1.31233111e-01
6.21824026e-01 -1.64142102e-01 3.69790167e-01 6.57086849e-01
4.27278429e-01 2.57307440e-01 1.37379721e-01 -6.22249842e-01
-8.93180132e-01 6.76275730e-01 3.46951991e-01 -3.30938280e-01
-5.30739963e-01 8.32735538e-01 7.11548448e-01 -6.01680219e-01
1.31439075e-01 2.03717902e-01 -1.34465709e-01 1.58832878e-01
1.15762174e+00 4.94332165e-01 -9.87422615e-02 -2.12903440e-01
-3.26138973e-01 5.69344796e-02 -7.70388842e-01 -2.78486535e-02
1.16284633e+00 -6.78063929e-01 2.31817201e-01 5.89523494e-01
1.27747226e+00 -1.23297445e-01 -1.38033545e+00 -1.20663583e+00
2.11831585e-01 -7.33870029e-01 2.11160421e-01 -1.11578131e+00
-5.68454146e-01 8.77751768e-01 6.15245223e-01 9.14016962e-02
7.57953703e-01 2.96295732e-01 1.12447274e+00 3.23612392e-01
3.00137401e-01 -1.32255137e+00 6.42924130e-01 8.73129249e-01
8.84974360e-01 -1.48446238e+00 -2.76956588e-01 -6.52212873e-02
-8.95042837e-01 6.64132476e-01 7.39108443e-01 -4.17206809e-02
8.05991411e-01 -1.20437341e-02 2.38804787e-01 -1.19166285e-01
-1.06015301e+00 -1.37023672e-01 4.08099204e-01 5.24224102e-01
-1.64459273e-01 -6.62806109e-02 2.27249693e-02 7.77131915e-01
3.58349890e-01 -1.29406646e-01 2.44414717e-01 6.90596163e-01
-1.11916184e+00 -7.45677233e-01 -2.96794653e-01 5.64146936e-01
-3.56753111e-01 -3.79364938e-01 -6.73964262e-01 9.93471146e-02
-1.08032659e-01 9.56552565e-01 -4.22858037e-02 -5.46771288e-01
8.51556137e-02 1.78989097e-01 -3.92809391e-01 -9.07935143e-01
-5.23109913e-01 -7.60838166e-02 -3.68346684e-02 -2.82457411e-01
4.50200289e-02 -4.09583330e-01 -8.92026126e-01 -3.74721318e-01
-8.69496614e-02 7.96218067e-02 2.94738829e-01 1.08531237e+00
8.00561488e-01 2.12878704e-01 8.43222380e-01 -5.48687100e-01
-1.22277868e+00 -1.29902124e+00 -5.05411446e-01 6.28661335e-01
5.62678397e-01 -7.50929892e-01 -6.15685821e-01 -3.44541848e-01] | [10.2212553024292, 3.655172824859619] |
45dab76e-b7ff-4283-b374-a74c2c503aa1 | fine-grained-propaganda-detection-with-fine | null | null | https://aclanthology.org/D19-5011 | https://aclanthology.org/D19-5011.pdf | Fine-Grained Propaganda Detection with Fine-Tuned BERT | This paper presents the winning solution of the Fragment Level Classification (FLC) task in the Fine Grained Propaganda Detection competition at the NLP4IF{'}19 workshop. The goal of the FLC task is to detect and classify textual segments that correspond to one of the 18 given propaganda techniques in a news articles dataset. The main idea of our solution is to perform word-level classification using fine-tuned BERT, a popular pre-trained language model. Besides presenting the model and its evaluation results, we also investigate the attention heads in the model, which provide insights into what the model learns, as well as aspects for potential improvements. | ['Yin Yang', 'Shehel Yoosuf'] | 2019-11-01 | null | null | null | ws-2019-11 | ['propaganda-detection'] | ['natural-language-processing'] | [ 3.01791430e-01 2.12105975e-01 -8.12348902e-01 -2.10849434e-01
-1.13470268e+00 -7.53265202e-01 1.23453724e+00 4.99824643e-01
-5.97331405e-01 4.37363058e-01 1.06386721e+00 -7.42700100e-01
-8.33244547e-02 -5.23926497e-01 -7.59370744e-01 -4.75104153e-01
2.01436765e-02 3.82949710e-01 6.17970265e-02 -3.52906168e-01
1.03261113e+00 4.87279683e-01 -8.81068885e-01 1.19359958e+00
9.57441926e-01 7.28392541e-01 -2.71853149e-01 8.08901370e-01
-2.81674027e-01 1.63480353e+00 -1.27061820e+00 -4.57051456e-01
-1.98484749e-01 -4.46397781e-01 -1.45993078e+00 -2.07763672e-01
6.31668210e-01 -1.19066581e-01 -1.58520237e-01 1.00060260e+00
1.81288585e-01 -5.42526804e-02 1.22041166e+00 -5.17570853e-01
-6.43506944e-01 1.46631634e+00 -7.51758456e-01 9.37602997e-01
3.22389752e-01 -4.05731559e-01 9.42588568e-01 -7.00808287e-01
1.05494118e+00 1.60320354e+00 5.19922674e-01 3.61704290e-01
-1.01948655e+00 -2.41173908e-01 2.69678742e-01 5.34706235e-01
-9.52069163e-01 -3.26458007e-01 7.32743561e-01 -8.01071405e-01
1.08503509e+00 1.19464576e-01 4.20756161e-01 1.49461496e+00
6.64380848e-01 1.35259366e+00 1.11056507e+00 -7.05579042e-01
6.74602687e-02 1.35680631e-01 9.09828842e-01 6.65057123e-01
1.76441222e-02 -2.48358160e-01 -4.18100357e-01 -4.49578464e-01
2.53817260e-01 -6.31879747e-01 -1.54467583e-01 3.07021648e-01
-9.98451173e-01 1.60351908e+00 4.41222221e-01 7.19616413e-01
-3.43648374e-01 2.38309503e-01 7.97191739e-01 9.04089436e-02
1.00623572e+00 8.15603197e-01 -2.94003338e-01 -1.27023399e-01
-1.23068607e+00 6.42804623e-01 6.96054399e-01 4.54774261e-01
1.21319078e-01 -1.02182344e-01 -8.57631028e-01 6.07698619e-01
-3.09966523e-02 1.30262181e-01 2.78080374e-01 -5.02066612e-01
8.27881277e-01 2.93437034e-01 -5.64974472e-02 -1.06824350e+00
-7.13003278e-01 -7.30898678e-01 -3.46576720e-01 -8.69111046e-02
3.25230926e-01 -2.17966273e-01 -1.07316136e+00 1.28564119e+00
8.27592099e-04 -3.31762373e-01 -3.77312511e-01 5.76849580e-01
6.97456658e-01 1.11223853e+00 2.85125196e-01 -4.28491980e-01
1.42967987e+00 -1.37189257e+00 -8.42425108e-01 -1.59245625e-01
1.14230740e+00 -7.69006968e-01 7.98673809e-01 2.83018649e-01
-8.45868409e-01 -3.37867022e-01 -7.80837536e-01 -2.11691439e-01
-4.26372141e-01 -6.57047331e-02 2.70206600e-01 4.21505660e-01
-3.36035401e-01 5.78645051e-01 -5.78147531e-01 -1.97341830e-01
4.84066039e-01 -5.67860186e-01 2.08669215e-01 1.82892308e-01
-1.33964419e+00 1.38048208e+00 6.58490777e-01 -1.33946240e-01
-9.68575716e-01 -8.17124963e-01 -7.52173483e-01 -9.06267669e-03
3.01851362e-01 -1.56895593e-02 1.21831310e+00 -8.58747661e-01
-8.93732190e-01 1.09004402e+00 1.12878554e-01 -9.25545514e-01
3.29052359e-01 -5.73381245e-01 -5.16669869e-01 2.31740996e-01
4.41053063e-01 3.06756288e-01 9.27938044e-01 -1.00911379e+00
-7.10968375e-01 -3.87261882e-02 9.86757949e-02 -2.07702488e-01
2.18925532e-02 8.40101302e-01 1.39278978e-01 -1.03527200e+00
-5.04974842e-01 -5.09941220e-01 -1.13205954e-01 -9.58677292e-01
-7.20992684e-01 -7.78283179e-01 6.50028050e-01 -1.11929381e+00
1.73712683e+00 -1.96269560e+00 2.43437827e-01 2.32164085e-01
1.26195192e-01 4.18977767e-01 -3.99558455e-01 7.72580743e-01
-6.46893531e-02 7.89218783e-01 1.34378940e-01 2.37019971e-01
-1.15589267e-02 -2.61296064e-01 -7.59556234e-01 7.62839437e-01
1.88126609e-01 8.37079048e-01 -7.48674750e-01 -4.29832488e-01
-2.53484190e-01 8.91433656e-02 -4.90755826e-01 -1.06393933e-01
-4.86266911e-01 1.55618966e-01 -5.12783170e-01 2.07249373e-01
3.39445561e-01 2.11939458e-02 -1.59594476e-01 -1.30334124e-01
-4.58017737e-01 1.09099293e+00 -4.04547513e-01 1.31395483e+00
-1.89930648e-01 9.16282833e-01 6.20863475e-02 -1.24408507e+00
4.00752336e-01 7.02682883e-02 9.67226177e-02 -8.12188566e-01
4.92857099e-01 -2.80469954e-02 9.00377408e-02 -5.38992524e-01
3.38634253e-01 -3.71801034e-02 -6.81515574e-01 6.33996844e-01
1.70858577e-01 5.02568111e-02 6.35761261e-01 5.34130573e-01
1.24492288e+00 -1.98809162e-01 7.08038092e-01 -7.67789781e-01
5.82672834e-01 5.59228659e-01 2.71787614e-01 1.39598906e+00
-2.54295915e-01 2.19048679e-01 7.25228667e-01 -6.74811125e-01
-7.91118264e-01 -7.67046809e-01 -3.39493185e-01 1.49449265e+00
-3.99479806e-01 -8.00956845e-01 -6.62900150e-01 -1.25367069e+00
-3.56724203e-01 1.03801823e+00 -1.20119953e+00 4.91457470e-02
-9.93528903e-01 -6.49898231e-01 7.85071254e-01 2.95096308e-01
3.24641854e-01 -1.21643496e+00 -7.95318902e-01 3.57044607e-01
-5.33016920e-01 -6.72109425e-01 -2.14995041e-01 4.94329453e-01
-3.26106489e-01 -1.00732422e+00 -5.12284636e-01 -9.07033622e-01
-6.27624169e-02 -5.08734584e-02 1.06052780e+00 4.71215546e-02
-2.24535093e-01 -1.57202840e-01 -1.02452815e+00 -6.05329573e-01
-8.32692862e-01 5.65932274e-01 -3.94393057e-01 -4.31785285e-01
2.36688986e-01 3.16083610e-01 -8.55324715e-02 -4.38787311e-01
-5.59357464e-01 9.67137888e-02 3.79864067e-01 1.00875461e+00
4.67284322e-02 7.09491447e-02 5.52656889e-01 -1.37290442e+00
1.19154918e+00 -4.30389017e-01 -2.73134261e-01 3.16818565e-01
-2.21280128e-01 8.83778036e-02 6.23209000e-01 -1.99315727e-01
-9.64138031e-01 -5.35130501e-01 -6.29857242e-01 3.72177631e-01
-1.70313701e-01 1.03265595e+00 4.03920799e-01 3.66868585e-01
1.25376511e+00 -2.50565298e-02 -7.34226406e-01 -7.36776233e-01
7.64575422e-01 6.62371755e-01 4.19805139e-01 -4.15510386e-01
6.02478385e-01 7.12892860e-02 -6.97889864e-01 -7.88144469e-01
-1.81264424e+00 -7.33575284e-01 -5.51060855e-01 4.54879329e-02
1.03640366e+00 -5.31754315e-01 -1.44196063e-01 2.01014549e-01
-1.69561172e+00 -3.47950786e-01 -1.69912726e-01 2.09671706e-01
-4.51105148e-01 1.79851815e-01 -1.23938489e+00 -9.40156758e-01
-4.74036574e-01 -7.28835762e-01 1.05467510e+00 -1.90022513e-01
-3.02448124e-01 -1.07397115e+00 1.68190792e-01 5.48385561e-01
1.11743852e-01 3.06682050e-01 1.57955134e+00 -1.12409329e+00
3.69308233e-01 -2.15242058e-01 -2.73118168e-01 9.60557535e-02
-2.90852070e-01 2.61583477e-02 -6.38164878e-01 -2.13215709e-01
2.59658366e-01 -6.14800930e-01 1.47195280e+00 5.51295638e-01
9.34966445e-01 -1.01274669e+00 -3.81602287e-01 9.86329019e-02
1.11807919e+00 1.51487321e-01 5.89505732e-01 7.10608602e-01
4.07000244e-01 7.41231024e-01 6.84569597e-01 1.89937353e-01
-1.81240946e-01 6.81186020e-01 2.43308455e-01 1.27982805e-02
-2.36047387e-01 -3.33847225e-01 7.04129040e-01 5.96497357e-01
1.40763745e-01 -4.96860474e-01 -1.10966885e+00 5.18102109e-01
-1.77979028e+00 -1.72900867e+00 -4.74249542e-01 1.28271854e+00
1.04260552e+00 3.33661824e-01 3.44715416e-01 1.50309831e-01
6.67377949e-01 6.91599250e-01 2.98671871e-01 -1.20066488e+00
-1.90322071e-01 4.66415972e-01 1.96397588e-01 7.37082124e-01
-1.71243584e+00 1.25572383e+00 7.18554735e+00 1.53819609e+00
-1.00106037e+00 3.25976908e-01 5.11723995e-01 -8.26136917e-02
-1.44552633e-01 -2.18363151e-01 -1.04957843e+00 3.29399109e-01
1.18328619e+00 -9.70439464e-02 1.94884002e-01 7.91592121e-01
2.94829011e-01 2.49832729e-03 -7.53312051e-01 2.35366225e-01
5.22040009e-01 -2.07698655e+00 4.13712054e-01 3.57007235e-02
7.34459877e-01 2.56330848e-01 -1.41832516e-01 7.65408397e-01
3.43475848e-01 -1.17517138e+00 1.31520057e+00 -1.09328240e-01
2.25811496e-01 -7.59404898e-01 8.67911458e-01 9.58508492e-01
-5.26679516e-01 -1.41164392e-01 -2.90175289e-01 -2.50111341e-01
4.06278342e-01 6.23771012e-01 -8.37192655e-01 3.90266806e-01
3.94690692e-01 5.69211960e-01 -7.63095498e-01 7.41977453e-01
-6.83694303e-01 1.39689708e+00 6.42854869e-02 -3.44465524e-01
9.02293384e-01 3.80524755e-01 6.71402752e-01 1.94261730e+00
-1.99501604e-01 4.17982675e-02 3.01721573e-01 7.57659376e-01
-2.07129642e-02 3.56686771e-01 -2.43447214e-01 -3.53246689e-01
1.95096985e-01 1.02534616e+00 -7.09291041e-01 -3.77765268e-01
1.54068582e-02 5.94280481e-01 7.92922199e-01 2.51809597e-01
-1.12527788e+00 -4.35110331e-01 9.30877179e-02 2.48192891e-01
2.37316072e-01 -6.05425090e-02 -2.67151415e-01 -1.20494235e+00
-4.77229863e-01 -8.44599843e-01 7.27555037e-01 -2.30222017e-01
-1.28037667e+00 5.42288363e-01 6.34897873e-02 -3.95010114e-01
-3.56403738e-01 -6.52902961e-01 -7.64701903e-01 5.45387447e-01
-1.14443839e+00 -1.14347041e+00 4.38439965e-01 2.95137674e-01
1.00978076e+00 -1.65357385e-02 5.36239862e-01 7.63408318e-02
-3.36345464e-01 3.92596692e-01 1.28845900e-01 5.67388713e-01
4.63836014e-01 -9.90229785e-01 4.01517361e-01 1.10611343e+00
4.63798702e-01 6.58943594e-01 1.05846322e+00 -7.46446192e-01
-7.47320294e-01 -1.19682574e+00 1.58514893e+00 -4.64011908e-01
1.15006411e+00 -5.05731940e-01 -8.45656216e-01 5.35789132e-01
5.14484048e-01 -6.57284260e-01 4.96625066e-01 4.91313398e-01
-8.47965658e-01 7.17286527e-01 -8.37579012e-01 3.68921280e-01
8.55403483e-01 -5.42876780e-01 -1.21237576e+00 8.36816788e-01
5.30457437e-01 -1.76746815e-01 -3.91432196e-01 1.83073506e-01
3.49902004e-01 -5.61009943e-01 7.04729676e-01 -1.31694365e+00
1.04806590e+00 3.40812624e-01 5.56280054e-02 -1.34381175e+00
-9.73794580e-01 -1.42777950e-01 -4.94479127e-02 1.10515916e+00
7.39217877e-01 -8.90708864e-02 4.38255996e-01 -4.33035761e-01
-2.59853393e-01 -6.95005000e-01 -1.01079476e+00 -8.23147535e-01
8.13840330e-01 -4.98237371e-01 -3.22310150e-01 9.63495255e-01
4.51459974e-01 9.50177550e-01 -6.38943374e-01 -1.95650056e-01
3.56561184e-01 2.91844934e-01 3.01901728e-01 -9.43739474e-01
-2.16069058e-01 -8.84149671e-01 -1.36138976e-01 -1.15316069e+00
4.19557333e-01 -1.20952010e+00 2.13011965e-01 -1.91021681e+00
3.32885116e-01 1.00829102e-01 -1.25499517e-01 3.14889699e-01
-8.32768381e-02 1.99880123e-01 4.02933151e-01 1.69631779e-01
-6.11938477e-01 1.78967357e-01 8.10455918e-01 -6.60639226e-01
-7.22511858e-02 -9.70703363e-02 -7.44910955e-01 8.18064690e-01
7.54617929e-01 -7.95583725e-01 9.10917446e-02 -6.83111727e-01
2.15635195e-01 -2.64433622e-01 3.14089566e-01 -4.40499634e-01
4.97407131e-02 -3.93702000e-01 2.22664341e-01 -1.02500570e+00
-1.50401399e-01 -5.63098490e-02 -6.02013767e-01 8.65897059e-01
-8.84150445e-01 -2.19896257e-01 1.35549344e-02 4.15848672e-01
-1.65387630e-01 -6.89461708e-01 9.67499018e-01 -1.93301722e-01
-4.68032449e-01 -3.45465630e-01 -1.13360655e+00 3.78517449e-01
7.15428531e-01 3.92695904e-01 -1.00524771e+00 -6.74876347e-02
-6.36889577e-01 -6.83308169e-02 -8.92380998e-02 5.13053119e-01
9.04387832e-02 -1.14160478e+00 -1.26894295e+00 -3.16658229e-01
2.04065323e-01 -6.09776199e-01 -2.94264196e-03 9.22789454e-01
-6.88283265e-01 8.50448012e-01 7.32889101e-02 -1.46571457e-01
-9.80376482e-01 7.90118694e-01 5.65064177e-02 -8.45797777e-01
-7.14409173e-01 9.42801893e-01 1.59963340e-01 7.57256672e-02
1.25872239e-01 -1.31596074e-01 -5.64930141e-01 3.24012160e-01
9.67949212e-01 4.00319397e-01 1.97463810e-01 -6.82189345e-01
-6.07062101e-01 -5.94214210e-03 -6.11991346e-01 -2.34410971e-01
1.34027696e+00 3.12575817e-01 -2.24600554e-01 3.30614984e-01
1.19204330e+00 3.54286432e-01 -3.88662636e-01 -6.39913604e-02
5.65125167e-01 2.36283541e-02 2.38334596e-01 -1.13209772e+00
-2.83434868e-01 9.34142828e-01 8.29154551e-02 3.58201504e-01
4.29358840e-01 3.41102540e-01 6.18360341e-01 2.90716469e-01
1.25517324e-01 -1.44067585e+00 -6.90833181e-02 1.17460740e+00
1.39637601e+00 -6.86392903e-01 1.22982010e-01 -4.05057132e-01
-2.91333586e-01 1.07901931e+00 6.03116490e-02 -5.68134904e-01
4.68911678e-01 2.56144166e-01 -1.75386183e-02 -5.57550132e-01
-7.53714085e-01 -7.18346937e-03 5.65569937e-01 4.66930002e-01
7.42452145e-01 -1.85475826e-01 -1.26416266e+00 7.12792635e-01
-1.64459839e-01 -6.20246947e-01 4.32369024e-01 8.93237114e-01
-1.01205444e+00 -6.21421993e-01 -2.46807933e-01 5.77223122e-01
-9.37063098e-01 -3.37754309e-01 -9.91906166e-01 6.63502455e-01
1.06781334e-01 1.28238070e+00 -2.28006572e-01 -1.68410823e-01
2.08960682e-01 2.37085763e-02 4.76553500e-01 -9.96789932e-01
-1.24666226e+00 4.33890909e-01 6.67259872e-01 -4.46375519e-01
-3.63959640e-01 -5.30813754e-01 -9.35890257e-01 -3.88346225e-01
-3.90503794e-01 6.64825559e-01 4.07491297e-01 1.16799521e+00
-1.73313797e-01 4.22018498e-01 4.19361949e-01 -6.89294338e-01
-8.60734344e-01 -1.30418408e+00 -3.99053484e-01 4.78000730e-01
2.71842897e-01 -4.54542786e-01 -4.08643246e-01 2.87288148e-02] | [8.475038528442383, 10.670292854309082] |
f45d9186-82d0-4ccf-84b0-6c1a3b0fc99a | skin-lesion-classification-using-class | 1703.01053 | null | http://arxiv.org/abs/1703.01053v1 | http://arxiv.org/pdf/1703.01053v1.pdf | Skin Lesion Classification using Class Activation Map | We proposed a two stage framework with only one network to analyze skin
lesion images, we firstly trained a convolutional network to classify these
images, and cropped the import regions which the network has the maximum
activation value. In the second stage, we retrained this CNN with the image
regions extracted from stage one and output the final probabilities. The two
stage framework achieved a mean AUC of 0.857 in ISIC-2017 skin lesion
validation set and is 0.04 higher than that of the original inputs, 0.821. | ['Xi Jia', 'Linlin Shen'] | 2017-03-03 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 6.33651733e-01 4.08963710e-01 -5.35666049e-01 -2.17084035e-01
-7.12727249e-01 -4.57333058e-01 2.79897660e-01 1.40962765e-01
-8.93855214e-01 6.08049452e-01 -7.84220099e-02 -4.06504810e-01
-3.39575820e-02 -6.91799462e-01 -5.00115931e-01 -8.54566991e-01
-6.05847612e-02 -7.41068348e-02 5.57893813e-01 4.27388430e-01
2.35013574e-01 4.48573798e-01 -1.05953050e+00 8.43993723e-01
7.64185607e-01 1.14508522e+00 -9.84303504e-02 9.19697821e-01
8.15345794e-02 8.19968104e-01 -4.83798385e-01 -3.61953110e-01
-4.12494242e-02 -2.28855103e-01 -1.11270452e+00 5.91657870e-02
4.24094424e-02 -2.28408426e-01 -2.72793740e-01 1.03722441e+00
2.87088990e-01 -2.71696329e-01 6.10684872e-01 -5.96575975e-01
-4.79621649e-01 5.35660684e-01 -7.76951551e-01 3.90053421e-01
-4.30178605e-02 1.64854646e-01 5.14044523e-01 -9.19851363e-01
7.78225660e-01 6.58361197e-01 7.03963459e-01 7.21485019e-01
-9.08795774e-01 -4.98568028e-01 -1.14733867e-01 6.57625422e-02
-1.34330893e+00 -4.83918339e-02 2.43588984e-01 -3.48293155e-01
8.17910910e-01 3.96918803e-01 6.86952531e-01 1.01924360e+00
5.17301500e-01 7.37372816e-01 1.52972889e+00 -4.50794816e-01
3.52727510e-02 2.92228281e-01 1.70293838e-01 5.72501063e-01
-2.38381829e-02 -1.32533431e-01 6.87678233e-02 -1.42824113e-01
6.14128411e-01 -8.26513097e-02 1.51007175e-01 3.31226707e-01
-8.67190301e-01 6.85148180e-01 9.04178321e-01 5.16333461e-01
-7.86813259e-01 -7.04396665e-02 3.58376950e-01 6.70116022e-02
3.73969823e-01 3.83917540e-01 -1.64007574e-01 2.90817261e-01
-8.95215869e-01 -4.99999434e-01 4.27504212e-01 -1.42219746e-02
1.86887398e-01 -4.47656840e-01 -3.51764172e-01 8.56266141e-01
2.21702438e-02 2.34718293e-01 4.03787732e-01 -5.23526073e-01
-4.97326702e-02 8.50322187e-01 -3.75130475e-01 -3.36736530e-01
-5.10037124e-01 -5.09744525e-01 -9.26090300e-01 3.54247421e-01
4.27900344e-01 -2.66330957e-01 -1.80479145e+00 1.26385927e+00
-3.54002528e-02 -1.81828782e-01 1.95823848e-01 9.15716529e-01
8.72588217e-01 3.12988609e-01 6.70572937e-01 1.99235067e-01
1.35759258e+00 -9.23369288e-01 -3.16975325e-01 -2.71544755e-01
5.73457837e-01 -7.04956353e-01 6.37510836e-01 5.54532290e-01
-9.91282880e-01 -1.88763261e-01 -9.38879788e-01 3.50576669e-01
-3.99818987e-01 6.24878228e-01 6.01644218e-01 5.00564456e-01
-1.34627199e+00 5.30252159e-01 -7.32938290e-01 -7.43857086e-01
8.85983050e-01 3.20981234e-01 -6.86217010e-01 -5.38208559e-02
-1.21498394e+00 1.19133604e+00 5.57797670e-01 3.45395297e-01
-9.54896986e-01 -4.86404568e-01 -4.77722168e-01 -4.49622087e-02
4.79905028e-03 -2.75527209e-01 1.00611293e+00 -1.53293097e+00
-1.22085786e+00 1.22231627e+00 -7.91670755e-02 -4.51738954e-01
4.04843777e-01 1.14246882e-01 -5.21106899e-01 5.04066110e-01
1.54166996e-01 7.20675409e-01 5.88342965e-01 -1.02780461e+00
-8.09406996e-01 -2.56911039e-01 5.12797907e-02 1.83636174e-02
-2.51625240e-01 2.72531956e-01 -5.80788314e-01 -1.31528422e-01
-1.16477450e-02 -9.02417779e-01 -6.52909756e-01 2.67021030e-01
-6.12418294e-01 -2.48015821e-01 3.72639209e-01 -8.98515642e-01
1.08404851e+00 -2.23777175e+00 -3.43469828e-01 8.60824049e-01
3.19137365e-01 6.45311475e-01 -2.81026453e-01 -1.09171301e-01
-6.00618303e-01 4.51050311e-01 -2.12329462e-01 1.89858437e-01
-6.64130270e-01 -1.86865613e-01 2.56124616e-01 3.34601283e-01
7.63527334e-01 8.30520213e-01 -6.93598986e-01 -6.70721114e-01
3.37595701e-01 3.68703783e-01 -2.35273913e-01 5.41292913e-02
3.25855047e-01 3.41276862e-02 -1.27568617e-01 8.60766232e-01
7.35017657e-01 -5.56409776e-01 3.84115845e-01 -3.13577026e-01
1.27623836e-02 -3.90075505e-01 -5.87391376e-01 1.33130670e+00
-2.19478384e-01 7.35300422e-01 2.41207089e-02 -5.48437536e-01
6.99663043e-01 3.98779064e-01 6.44616008e-01 -3.56970817e-01
2.41990581e-01 -4.21882235e-02 3.73595625e-01 -8.35761487e-01
-1.30208567e-01 -1.53675452e-01 1.04308397e-01 5.19921705e-02
2.80491471e-01 3.05760056e-01 4.80818897e-02 2.86245078e-01
1.46945405e+00 -1.39838085e-01 3.97062719e-01 -6.49325270e-03
6.57288074e-01 4.26254831e-02 2.87712127e-01 5.00374079e-01
-5.17443836e-01 7.37040341e-01 9.41480994e-01 -5.37860751e-01
-6.99306965e-01 -1.20580208e+00 -5.50238073e-01 6.74249589e-01
-2.02700078e-01 -1.19624950e-01 -8.92929614e-01 -1.17763543e+00
-8.80168602e-02 5.13757646e-01 -1.44781637e+00 -1.80150807e-01
-1.61367580e-01 -7.73507237e-01 7.18230784e-01 6.93379581e-01
4.90935504e-01 -1.25603652e+00 -2.21048996e-01 -6.82010725e-02
7.59475827e-02 -7.56666839e-01 -8.19251537e-02 3.79968137e-01
-5.32191694e-01 -1.48070455e+00 -9.54679787e-01 -8.35606098e-01
1.16285491e+00 -2.06635311e-01 5.57592332e-01 3.30154836e-01
-5.40820539e-01 -3.24057192e-01 -2.52874911e-01 -4.69888955e-01
-4.23795521e-01 2.95854449e-01 -4.99892354e-01 2.04028443e-01
5.43755829e-01 -1.92295178e-03 -8.30248773e-01 -3.39908376e-02
-9.70456183e-01 -1.89208668e-02 1.20839942e+00 8.06287527e-01
5.52269280e-01 -1.44009054e-01 6.99002683e-01 -1.09314787e+00
4.78095263e-01 -5.56343377e-01 -3.10183689e-02 4.55667764e-01
-5.04190028e-01 -5.07055044e-01 5.44756234e-01 -4.32654351e-01
-9.01093066e-01 7.18491614e-01 -3.16055924e-01 -2.62510717e-01
-3.26005191e-01 5.35038352e-01 4.76310521e-01 -1.58811305e-02
8.07465434e-01 -1.06002986e-01 3.72851163e-01 -4.26475555e-02
-1.48416050e-02 8.38449597e-01 5.70020676e-01 7.02693686e-02
4.92568970e-01 2.19828635e-01 -8.92572254e-02 -5.27941108e-01
-8.86726379e-01 -5.43935537e-01 -9.63155210e-01 -5.50056756e-01
1.27124023e+00 -6.45672560e-01 -3.26449394e-01 7.51308739e-01
-1.02522433e+00 -1.77335173e-01 -2.45476186e-01 4.71749812e-01
5.58436960e-02 2.33685836e-01 -5.68105936e-01 -7.23910451e-01
-6.06694102e-01 -9.17686760e-01 6.43415332e-01 8.88033271e-01
-4.50383306e-01 -1.04134607e+00 1.02036245e-01 1.55541515e-02
4.47097212e-01 5.56280375e-01 8.80585611e-01 -7.65961468e-01
3.12595934e-01 -7.81636298e-01 -4.82004762e-01 4.83697534e-01
2.76874959e-01 4.97764558e-01 -1.28028858e+00 -1.01476386e-01
-5.84984243e-01 -3.58295947e-01 1.24410653e+00 5.28735757e-01
1.73310709e+00 2.10644975e-01 -5.08875430e-01 5.14104009e-01
1.57656014e+00 2.38614962e-01 9.16735768e-01 2.07381785e-01
1.56666934e-01 4.86618221e-01 3.07500631e-01 -2.74397314e-01
-8.58730748e-02 -1.20340168e-01 4.83916163e-01 -9.72622812e-01
-2.56701738e-01 -2.67158031e-01 1.58722132e-01 2.09301099e-01
-2.09953114e-01 1.46002501e-01 -8.84722888e-01 8.88061225e-01
-1.37512255e+00 -7.25842535e-01 -1.59621835e-01 1.80492067e+00
8.97778451e-01 4.67246503e-01 2.60232817e-02 -2.79425960e-02
9.47973371e-01 -4.71914969e-02 -6.47328556e-01 -6.83393478e-01
-1.84528336e-01 8.99657905e-01 6.02840960e-01 2.49886855e-01
-1.34414923e+00 1.06729436e+00 7.92548275e+00 5.73296964e-01
-1.37449181e+00 -4.60629864e-03 9.75334466e-01 -2.54441518e-02
8.94350335e-02 -1.86241150e-01 -1.86744466e-01 3.87528390e-01
1.16046321e+00 2.01321632e-01 -2.91018859e-02 7.40681767e-01
-8.19760785e-02 -5.59454024e-01 -5.23382306e-01 3.33674222e-01
1.04385413e-01 -1.19152784e+00 -8.87218118e-02 1.35237724e-01
7.24690855e-01 -1.58434492e-02 1.69662207e-01 1.82611957e-01
2.68510103e-01 -1.51614296e+00 -1.73115209e-02 8.02053690e-01
1.27547228e+00 -4.79530275e-01 1.20014107e+00 -1.54273706e-02
-7.10937917e-01 -8.02650955e-03 -1.79159194e-01 2.55351752e-01
-1.62786201e-01 5.55953979e-01 -1.27923143e+00 3.82724196e-01
8.37931156e-01 4.08771008e-01 -9.67380404e-01 1.25887144e+00
-4.01847482e-01 9.62362528e-01 -6.07492886e-02 -1.98108956e-01
3.64090800e-01 3.59062940e-01 6.25357181e-02 1.33412111e+00
-5.62239140e-02 1.64106414e-02 -3.04194421e-01 4.85726953e-01
3.74871641e-02 1.73330039e-01 -4.10616696e-01 5.57503337e-03
1.22922033e-01 1.88576031e+00 -9.36580956e-01 -4.85677868e-01
-2.25831479e-01 9.37574089e-01 2.58525878e-01 3.28624994e-01
-8.04661036e-01 -8.10144186e-01 1.40290093e-02 4.14161831e-02
-7.65160695e-02 5.96674442e-01 -5.23341119e-01 -7.22971797e-01
-2.00953096e-01 -5.47986686e-01 7.42808819e-01 -8.18480194e-01
-1.34330428e+00 8.25669885e-01 -4.67696130e-01 -7.93137252e-01
1.52302338e-02 -1.01199818e+00 -1.10857177e+00 1.20274210e+00
-1.37334299e+00 -1.28475416e+00 -4.07472551e-01 4.17676300e-01
2.23041952e-01 1.05824452e-02 1.12584257e+00 -3.44654247e-02
-9.13675547e-01 7.26700962e-01 -3.90509427e-01 6.04669571e-01
8.76104712e-01 -1.39435184e+00 2.94582397e-02 7.65224040e-01
-5.20064950e-01 5.84360540e-01 -1.15864933e-01 -6.81093872e-01
-5.76303184e-01 -1.19847023e+00 7.92767704e-01 -2.76436865e-01
6.55468106e-01 2.02061743e-01 -8.22226942e-01 4.51592833e-01
4.57038909e-01 4.43039648e-02 1.06347954e+00 -1.17681604e-02
-1.92918360e-01 4.87395935e-02 -1.49009526e+00 3.45586717e-01
2.53090292e-01 -3.09428602e-01 -3.41061711e-01 3.28188509e-01
3.20605874e-01 -2.75258601e-01 -1.26017177e+00 6.03754520e-01
9.10366893e-01 -4.36681300e-01 7.86506772e-01 -9.56780195e-01
6.07253313e-01 -1.07966423e-01 1.46479636e-01 -1.16349995e+00
-5.50817728e-01 7.53479972e-02 2.53484130e-01 7.41843045e-01
1.04788947e+00 -6.39131129e-01 9.01107848e-01 4.90028560e-01
9.35535133e-02 -1.34163690e+00 -8.08634698e-01 -5.03271632e-02
5.00420094e-01 8.70226324e-02 1.76589131e-01 7.19488800e-01
-1.24126092e-01 -2.04263046e-01 1.41310483e-01 1.34702101e-01
4.11620468e-01 -4.84717160e-01 2.49083176e-01 -9.97621894e-01
3.48731726e-01 -3.48447591e-01 -2.83075571e-01 -1.42871201e-01
-5.99548817e-02 -1.21900117e+00 5.95979244e-02 -1.74841392e+00
6.74871802e-01 -2.41789952e-01 -1.10493422e+00 9.53612626e-01
-4.78776246e-01 6.09697104e-01 -9.63110775e-02 1.77578568e-01
-2.85057813e-01 -3.69646579e-01 1.14811242e+00 -1.84529200e-01
-5.82387485e-03 6.43881559e-02 -1.00614798e+00 6.30512953e-01
1.20840561e+00 -2.32393086e-01 3.54656503e-02 -9.70210508e-02
-2.75616407e-01 -2.78796256e-01 5.64963281e-01 -1.07148767e+00
2.05126926e-01 -3.87403607e-01 1.41918254e+00 -6.17649496e-01
1.94008518e-02 -6.36189520e-01 -6.72901347e-02 9.40036833e-01
-5.42236388e-01 -4.35229331e-01 2.66949743e-01 1.11193784e-01
-4.06688601e-02 -4.21233922e-01 1.18104482e+00 -1.38695622e-02
-6.48288310e-01 2.40478307e-01 -6.82449281e-01 -5.93033552e-01
1.34353614e+00 -4.19245273e-01 -2.84268230e-01 8.51320699e-02
-1.14574409e+00 1.20694317e-01 2.87494451e-01 2.58952826e-01
7.02543318e-01 -1.29402328e+00 -6.57953918e-01 2.27091536e-01
1.38517031e-02 -2.76655555e-01 3.91701519e-01 1.05657792e+00
-6.70525730e-01 1.77485019e-01 -5.20974815e-01 -7.36587107e-01
-1.27168930e+00 1.88659355e-01 6.56175375e-01 -4.92321432e-01
-3.21811080e-01 8.28995943e-01 -9.87529606e-02 -3.92369241e-01
-1.28119513e-02 -6.52806237e-02 -6.32280827e-01 -9.30314958e-02
5.42515576e-01 -5.08409888e-02 1.21286169e-01 -3.94894004e-01
-6.55993164e-01 5.59518814e-01 -4.26973343e-01 -9.74272117e-02
1.19868481e+00 5.56246340e-01 -2.80912042e-01 3.14030014e-02
1.20650876e+00 -2.78434038e-01 -9.61676121e-01 5.33816554e-02
-9.50558260e-02 -1.50079995e-01 1.60586372e-01 -1.48941743e+00
-1.21299517e+00 7.12399781e-01 9.74502504e-01 7.08198324e-02
1.34919131e+00 -1.00998409e-01 4.28934723e-01 -8.23586285e-02
-2.29808733e-01 -1.19453073e+00 -9.56946984e-02 3.56026322e-01
6.37298226e-01 -9.76322293e-01 -3.21028493e-02 -3.87931764e-01
-8.62860501e-01 1.33574283e+00 9.29415941e-01 -4.59525138e-01
7.95329630e-01 3.43985170e-01 3.23358744e-01 -3.18799615e-01
-8.15463185e-01 -2.45817110e-01 4.60697174e-01 7.03489184e-01
5.44782400e-01 2.22522512e-01 -5.68418264e-01 6.55443251e-01
2.21697405e-01 2.95942128e-01 3.12351644e-01 7.23110557e-01
-5.81670046e-01 -8.38336408e-01 -1.16611436e-01 9.50449884e-01
-9.07134473e-01 1.71448931e-01 -1.02160740e+00 8.92439246e-01
6.51050806e-02 7.46895254e-01 2.86675870e-01 -7.46548057e-01
2.67867059e-01 1.52030006e-01 1.75925717e-01 -5.46473444e-01
-9.48016465e-01 8.45546275e-02 1.44413888e-01 -4.93028224e-01
-2.47339606e-01 -4.35206532e-01 -1.28384566e+00 1.79285228e-01
-4.38155621e-01 3.97746973e-02 7.58057833e-01 5.25773406e-01
1.29480094e-01 6.52631342e-01 9.37354028e-01 -2.25172296e-01
-2.92311639e-01 -1.29593205e+00 -2.74739683e-01 2.94535071e-01
1.29675850e-01 -2.33441070e-01 -3.80755484e-01 -1.77078426e-01] | [15.68127155303955, -2.9838926792144775] |
ae433c0d-3f57-4c46-9673-0f6d5493c53d | tog-targeted-adversarial-objectness-gradient | 2004.04320 | null | https://arxiv.org/abs/2004.04320v1 | https://arxiv.org/pdf/2004.04320v1.pdf | TOG: Targeted Adversarial Objectness Gradient Attacks on Real-time Object Detection Systems | The rapid growth of real-time huge data capturing has pushed the deep learning and data analytic computing to the edge systems. Real-time object recognition on the edge is one of the representative deep neural network (DNN) powered edge systems for real-world mission-critical applications, such as autonomous driving and augmented reality. While DNN powered object detection edge systems celebrate many life-enriching opportunities, they also open doors for misuse and abuse. This paper presents three Targeted adversarial Objectness Gradient attacks, coined as TOG, which can cause the state-of-the-art deep object detection networks to suffer from object-vanishing, object-fabrication, and object-mislabeling attacks. We also present a universal objectness gradient attack to use adversarial transferability for black-box attacks, which is effective on any inputs with negligible attack time cost, low human perceptibility, and particularly detrimental to object detection edge systems. We report our experimental measurements using two benchmark datasets (PASCAL VOC and MS COCO) on two state-of-the-art detection algorithms (YOLO and SSD). The results demonstrate serious adversarial vulnerabilities and the compelling need for developing robust object detection systems. | ['Stacey Truex', 'Ka-Ho Chow', 'Mehmet Emre Gursoy', 'Ling Liu', 'Yanzhao Wu', 'Wenqi Wei'] | 2020-04-09 | null | null | null | null | ['robust-object-detection'] | ['computer-vision'] | [-5.11715673e-02 -7.66096590e-03 6.74325526e-02 -1.40726298e-01
-4.77605253e-01 -7.09962308e-01 5.45497000e-01 -3.65297586e-01
-7.04148769e-01 3.18300486e-01 -6.54169321e-01 -6.48801923e-01
2.54741549e-01 -7.40400136e-01 -1.03117764e+00 -5.26864290e-01
-3.62192720e-01 9.69751999e-02 6.69094920e-01 -3.20422262e-01
1.80092789e-02 9.12658334e-01 -1.51694107e+00 2.75234848e-01
2.94533283e-01 1.59802067e+00 -3.48198682e-01 1.24845052e+00
2.94598639e-01 9.65594649e-01 -9.18719292e-01 -7.87183702e-01
8.13786328e-01 1.47220880e-01 -2.00052530e-01 -4.75890577e-01
9.59193647e-01 -5.77028573e-01 -1.06754732e+00 1.65895748e+00
7.67515004e-01 -3.55072558e-01 3.69952083e-01 -2.13160276e+00
-9.50327575e-01 3.45458120e-01 -3.80582124e-01 6.49924874e-01
-2.70250350e-01 4.93803740e-01 3.26394320e-01 -9.28121030e-01
5.21291494e-01 1.17953491e+00 1.00630677e+00 1.02155232e+00
-5.03090918e-01 -1.13118815e+00 -1.25188306e-01 5.26274323e-01
-1.23846662e+00 -4.56508905e-01 5.40106535e-01 -2.24273399e-01
9.10334349e-01 3.95132750e-01 4.61894035e-01 1.34193361e+00
7.48744190e-01 9.32007432e-01 7.39651501e-01 8.47133696e-02
2.97950506e-01 8.85750502e-02 9.09784511e-02 8.98067176e-01
8.05222809e-01 9.73513424e-01 -4.91307616e-01 1.66424811e-02
5.08630991e-01 7.36786574e-02 2.73555666e-02 -1.76029921e-01
-7.65705884e-01 6.82572305e-01 8.25510919e-01 -3.38211238e-01
-1.43351808e-01 9.22725737e-01 7.88273454e-01 3.82541239e-01
3.64498049e-02 9.95202810e-02 -3.29712838e-01 -6.91303015e-02
-4.34964001e-01 1.36796340e-01 6.87039614e-01 1.39193714e+00
3.13120037e-01 6.07875407e-01 -1.58924595e-01 -8.24161842e-02
3.44127327e-01 7.15533137e-01 2.32878983e-01 -5.75363874e-01
2.28008106e-01 3.48059624e-01 -9.72166806e-02 -8.67612660e-01
-3.60198289e-01 -3.32832158e-01 -6.81319952e-01 1.08052313e+00
1.92784384e-01 -3.81757379e-01 -1.38577497e+00 1.43235588e+00
4.01344359e-01 3.42144489e-01 2.01949835e-01 1.01750660e+00
1.17489886e+00 3.21790874e-01 2.47707307e-01 6.07983291e-01
1.60408425e+00 -8.47195566e-01 -4.65740770e-01 -5.88037729e-01
6.97718441e-01 -5.42688966e-01 5.54351568e-01 4.37783778e-01
-5.43846011e-01 -4.03911829e-01 -1.56591320e+00 -8.55259150e-02
-9.27331686e-01 -2.36511618e-01 8.43488753e-01 1.48439085e+00
-6.77856147e-01 2.84784108e-01 -8.67354929e-01 1.70607362e-02
1.19845450e+00 5.77391326e-01 -3.98750573e-01 -4.05677557e-02
-1.10314691e+00 8.95979404e-01 5.64161599e-01 1.11508437e-01
-1.52372015e+00 -7.67499149e-01 -6.67630315e-01 -1.80306602e-02
3.94254297e-01 -3.21223199e-01 1.03889334e+00 -8.19357812e-01
-9.61676121e-01 9.79868412e-01 8.30256939e-01 -9.79067087e-01
8.76281738e-01 -3.50000441e-01 -8.54997754e-01 3.88399959e-02
-2.15440869e-01 9.36440170e-01 1.09465289e+00 -1.03897035e+00
-8.05224895e-01 -4.87346321e-01 2.50112917e-02 -3.35609257e-01
-4.96456623e-01 4.07362968e-01 -6.25546202e-02 -7.24078298e-01
-2.78882295e-01 -1.05601442e+00 -1.26965549e-02 7.61357665e-01
-4.44934517e-01 -1.12996094e-01 1.88802791e+00 -3.20713937e-01
6.50353551e-01 -2.48191714e+00 -6.93532050e-01 -2.14379281e-01
5.39294899e-01 7.96949446e-01 -1.94064870e-01 -1.22052364e-01
8.54566321e-02 8.58410224e-02 2.81510521e-02 -1.74515083e-01
1.13610983e-01 1.32122114e-01 -5.55699527e-01 7.72396803e-01
2.28289932e-01 1.38293409e+00 -9.38786864e-01 -2.11251393e-01
1.83291793e-01 3.80278528e-01 -3.64212245e-01 1.43632572e-02
5.41161150e-02 -2.56239504e-01 -2.08437175e-01 1.12435639e+00
7.18848765e-01 4.30296063e-01 -7.79425561e-01 -2.67764181e-01
3.04637313e-01 -3.29272687e-01 -1.08627427e+00 9.34911311e-01
5.10574207e-02 1.20205462e+00 1.59445584e-01 -6.02267861e-01
8.12569797e-01 2.07252502e-01 -3.06201372e-02 -6.40978575e-01
5.75214565e-01 3.64320964e-01 1.99857339e-01 -4.85665739e-01
6.91589296e-01 2.37198025e-02 -2.68484980e-01 -5.05707972e-02
7.60833845e-02 -9.39466432e-02 -3.27668488e-01 4.50778931e-01
1.34076023e+00 -3.89385700e-01 -1.77372411e-01 -3.00632030e-01
-8.82688835e-02 5.72538115e-02 5.38910866e-01 1.07796466e+00
-9.31292355e-01 3.67246866e-01 2.39838168e-01 -6.30385876e-01
-1.04389167e+00 -1.29349053e+00 -1.98456690e-01 9.13100779e-01
6.09584093e-01 1.20473914e-01 -7.93201387e-01 -9.90135729e-01
4.15326864e-01 5.38672090e-01 -6.52495325e-01 -8.31492007e-01
-2.10429370e-01 -6.29968345e-01 1.45743489e+00 1.01199687e+00
8.82292092e-01 -1.19659531e+00 -9.37954307e-01 4.00677808e-02
5.88081062e-01 -1.39785445e+00 -4.57814425e-01 2.90444911e-01
-4.29499686e-01 -1.12720656e+00 -4.48913246e-01 -8.75957966e-01
5.18461585e-01 3.50720823e-01 9.87533808e-01 -3.28587405e-02
-8.89934003e-01 5.17196655e-01 -1.46972001e-01 -1.34423077e+00
-6.48627639e-01 -4.31920856e-01 5.36535263e-01 -7.09722266e-02
6.29250646e-01 5.63492440e-03 -6.07154131e-01 3.94275129e-01
-9.89844620e-01 -4.88405377e-01 5.08308232e-01 7.88140595e-01
2.05145985e-01 -2.76280772e-02 8.03763092e-01 -7.62915254e-01
3.88210773e-01 -3.64110798e-01 -1.01952386e+00 1.00924328e-01
-5.36703944e-01 -3.79074812e-01 5.46646953e-01 -8.65642250e-01
-6.73870444e-01 5.81578305e-03 -8.28691050e-02 -8.19583058e-01
3.89310569e-02 -2.23132282e-01 -3.72827351e-01 -9.27945137e-01
1.11696815e+00 -7.99023136e-02 -2.56066084e-01 3.66686694e-02
3.88199478e-01 9.45265472e-01 1.19872606e+00 -1.81643635e-01
9.77200866e-01 7.86895633e-01 1.18104942e-01 -8.38936865e-01
-3.68157119e-01 -2.13148564e-01 1.73269913e-01 -5.06097674e-01
9.24930990e-01 -9.44144845e-01 -9.20874000e-01 1.13338912e+00
-1.13974988e+00 -2.96250373e-01 -3.59239966e-01 2.85247982e-01
-2.83369750e-01 1.76066667e-01 -6.13849938e-01 -7.25058675e-01
-5.88651001e-01 -1.08269346e+00 9.54352915e-01 4.29460078e-01
4.85719502e-01 -4.79986310e-01 -5.16011178e-01 8.80451426e-02
5.62911093e-01 4.78459269e-01 1.85752153e-01 -7.54427671e-01
-9.19690073e-01 -9.77485955e-01 -6.66425169e-01 5.76469123e-01
-3.58452857e-01 -1.93492025e-01 -1.24546206e+00 -5.09217381e-01
3.28895785e-02 -4.38115388e-01 8.85248840e-01 -8.30361098e-02
1.12521350e+00 -3.39735001e-01 -4.44834262e-01 1.03374898e+00
1.36195409e+00 3.33990276e-01 8.80324364e-01 3.56735468e-01
1.06194150e+00 -8.77281353e-02 3.67081791e-01 1.12860411e-01
-2.00773105e-01 2.99267262e-01 1.14021671e+00 -1.26874700e-01
-3.87210190e-01 -6.96979165e-02 4.44636673e-01 -1.66534781e-01
5.08053839e-01 -4.84115154e-01 -7.90440798e-01 5.80472112e-01
-1.51061594e+00 -8.82848978e-01 -2.54169166e-01 1.85427999e+00
2.30313554e-01 7.67188489e-01 -1.39289191e-02 -3.11698741e-03
7.99807370e-01 -1.03450157e-01 -1.07150042e+00 -6.33186400e-01
-1.60983458e-01 -5.68704009e-02 1.41045928e+00 -9.82493907e-02
-1.42052698e+00 1.08358121e+00 6.39306450e+00 7.34589994e-01
-1.21385002e+00 2.25110471e-01 7.12878585e-01 -1.01604871e-01
2.85126388e-01 -3.57283562e-01 -1.07567990e+00 4.41774100e-01
8.67719054e-01 -1.79384738e-01 3.01874280e-01 1.54890752e+00
-5.60480237e-01 2.65399456e-01 -1.01993501e+00 1.18025470e+00
3.70732136e-02 -1.46162748e+00 -1.95483729e-01 6.25059381e-02
7.28455424e-01 6.14307046e-01 2.57546425e-01 5.30542910e-01
4.36934799e-01 -9.30234253e-01 1.08466280e+00 -5.49845845e-02
1.03237748e+00 -7.88925886e-01 7.61686504e-01 9.03806984e-02
-1.06087768e+00 -3.36681157e-01 -4.67085898e-01 1.72653794e-01
8.24898556e-02 3.08115929e-01 -5.49087107e-01 -2.25377977e-02
1.00604570e+00 9.50690582e-02 -6.11838996e-01 1.19322598e+00
-7.35637397e-02 4.53361630e-01 -5.36527693e-01 -2.18535125e-01
3.23293120e-01 5.54160953e-01 1.04977691e+00 1.20530438e+00
2.53516492e-02 3.49458270e-02 -9.81128067e-02 8.40884030e-01
-5.90437472e-01 -7.17359364e-01 -9.40370142e-01 -1.71729952e-01
4.48536068e-01 1.34367788e+00 -9.00661290e-01 -3.67976785e-01
-3.62166911e-01 1.08485746e+00 -1.12393603e-01 7.30993971e-02
-1.26388371e+00 -8.67269635e-01 1.04990661e+00 -1.38046339e-01
2.62201130e-01 -3.79249066e-01 -4.87587631e-01 -7.46978164e-01
1.14613511e-02 -8.63646328e-01 3.89459282e-01 -4.62841809e-01
-1.36354721e+00 6.30573153e-01 -2.38042206e-01 -1.16123593e+00
3.63125533e-01 -1.32974172e+00 -7.53302693e-01 2.71209836e-01
-1.32020497e+00 -1.08776379e+00 -5.13946354e-01 4.82185036e-01
2.70803988e-01 -5.30240774e-01 3.68199348e-01 7.53296137e-01
-6.02356315e-01 1.32162082e+00 5.19829094e-02 8.04660499e-01
4.42200631e-01 -9.51535046e-01 1.26031327e+00 1.30049539e+00
3.06114703e-01 -8.28057081e-02 6.08511627e-01 -7.17660546e-01
-1.87348580e+00 -1.37972724e+00 4.29645181e-02 -6.49951577e-01
6.62613750e-01 -6.31410658e-01 -7.30173886e-01 7.82836199e-01
-2.85643727e-01 8.13322723e-01 1.20921098e-02 -7.11275399e-01
-6.45662487e-01 -1.15404472e-01 -1.61129737e+00 6.77416444e-01
1.20745409e+00 -5.88772774e-01 -2.35122651e-01 3.35247457e-01
9.28529084e-01 -6.27059340e-01 -3.17377716e-01 5.70408404e-01
5.64096093e-01 -9.36445713e-01 1.08946323e+00 -9.57578957e-01
-1.24276496e-01 -2.95761228e-01 -1.18289463e-01 -7.37529695e-01
-1.47723451e-01 -9.02227581e-01 -5.83469808e-01 9.19938147e-01
1.04881562e-01 -1.06897175e+00 1.24853790e+00 7.04919815e-01
-6.27012789e-01 -4.75274801e-01 -1.34025812e+00 -1.47125518e+00
-2.17623666e-01 -7.35368490e-01 4.66660410e-01 7.31232405e-01
-4.49978501e-01 -2.12179944e-01 -1.89862758e-01 6.82933927e-01
9.90651071e-01 -6.35508597e-01 8.30424368e-01 -8.55798364e-01
-1.15630865e-01 -5.70815921e-01 -1.55622697e+00 -6.55111969e-01
-4.44549441e-01 -7.19103694e-01 1.84056818e-01 -8.62338781e-01
-3.53189617e-01 -3.14748913e-01 -4.50125664e-01 4.13914651e-01
-1.67920709e-01 9.16062295e-01 4.49136913e-01 -1.69084400e-01
-5.90259016e-01 3.25512558e-01 7.51788259e-01 -4.61740255e-01
1.72805533e-01 -3.23672779e-02 -4.43761736e-01 7.14772880e-01
5.47484100e-01 -1.01799440e+00 -1.11230031e-01 -4.71008748e-01
2.34426618e-01 -5.77470481e-01 9.14921343e-01 -1.66066539e+00
5.89017987e-01 1.95015386e-01 3.35248649e-01 -2.86371976e-01
2.14174882e-01 -1.01892638e+00 -7.21259490e-02 1.04767072e+00
1.67924091e-01 -1.28104970e-01 8.27119172e-01 7.92176306e-01
2.20148504e-01 -1.06157988e-01 1.16991258e+00 2.40062684e-01
-1.36783969e+00 7.13807285e-01 -1.87415525e-01 2.55300969e-01
1.53773856e+00 -5.27026653e-01 -7.41519213e-01 -7.01997951e-02
-2.33156368e-01 2.81222492e-01 1.75805137e-01 9.87174034e-01
8.48416030e-01 -1.51121056e+00 -8.39067221e-01 3.81665736e-01
4.01094884e-01 -1.01099886e-01 8.74998868e-02 2.72803843e-01
-8.85371804e-01 7.53465593e-02 -5.00696659e-01 -5.49047470e-01
-1.21769762e+00 1.03744590e+00 6.74748957e-01 5.24355590e-01
-8.30191612e-01 1.35219073e+00 2.64296502e-01 -5.73030487e-02
5.41599333e-01 -1.39907911e-01 6.09270692e-01 -4.55496877e-01
6.95781827e-01 7.17138529e-01 2.21107259e-01 -3.71911347e-01
-5.14491141e-01 -4.05102596e-02 -2.45880514e-01 2.64459014e-01
7.52911210e-01 4.72859710e-01 4.88973916e-01 -1.78932101e-01
1.21506858e+00 -4.43005770e-01 -1.51839411e+00 4.06592228e-02
-1.23913333e-01 -4.36869740e-01 4.01638746e-01 -7.03896761e-01
-1.19200134e+00 7.73667932e-01 1.32064915e+00 3.74307215e-01
8.06124389e-01 -9.13833678e-02 1.28990531e+00 6.54740393e-01
4.76422220e-01 -9.32995319e-01 1.36681393e-01 3.05199742e-01
6.49217188e-01 -1.48640633e+00 -2.03549296e-01 -3.71725351e-01
-2.70767570e-01 6.77706957e-01 9.25794721e-01 -4.20640886e-01
8.01233709e-01 7.90648341e-01 2.69744366e-01 -3.72965187e-01
-2.81384498e-01 7.31597394e-02 1.09702557e-01 8.55861425e-01
-5.93303144e-01 9.95022058e-02 1.97031930e-01 5.08944094e-01
9.00660753e-02 -1.21731376e-02 5.30083358e-01 1.04325283e+00
-5.70759118e-01 -4.45607126e-01 -5.51575005e-01 5.45712113e-01
-6.72683835e-01 -1.90495588e-02 -4.19645488e-01 7.37809360e-01
1.14130378e-01 6.99531734e-01 2.00660184e-01 -5.82177281e-01
2.63847232e-01 -1.74313888e-01 2.06402302e-01 -1.97876200e-01
-6.71479285e-01 -9.33473706e-01 -7.19166249e-02 -6.32040501e-01
5.14096498e-01 -1.00759760e-01 -1.34088719e+00 -4.52940464e-01
-6.94584370e-01 -4.68215197e-01 1.17950416e+00 6.34637475e-01
5.62621176e-01 5.68766296e-01 3.45654428e-01 -9.94961977e-01
-9.50349867e-01 -5.28472662e-01 -4.52396274e-01 3.32788855e-01
5.16431570e-01 -5.93200922e-01 -5.05464911e-01 -2.11294651e-01] | [5.439186096191406, 7.89467191696167] |
807fce78-b28b-48fc-ab47-6fd8d58125a8 | heterogeneous-graph-neural-networks-for-3 | 2109.04703 | null | https://arxiv.org/abs/2109.04703v1 | https://arxiv.org/pdf/2109.04703v1.pdf | Heterogeneous Graph Neural Networks for Keyphrase Generation | The encoder-decoder framework achieves state-of-the-art results in keyphrase generation (KG) tasks by predicting both present keyphrases that appear in the source document and absent keyphrases that do not. However, relying solely on the source document can result in generating uncontrollable and inaccurate absent keyphrases. To address these problems, we propose a novel graph-based method that can capture explicit knowledge from related references. Our model first retrieves some document-keyphrases pairs similar to the source document from a pre-defined index as references. Then a heterogeneous graph is constructed to capture relationships of different granularities between the source document and its references. To guide the decoding process, a hierarchical attention and copy mechanism is introduced, which directly copies appropriate words from both the source document and its references based on their relevance and significance. The experimental results on multiple KG benchmarks show that the proposed model achieves significant improvements against other baseline models, especially with regard to the absent keyphrase prediction. | ['Qi Zhang', 'Tao Gui', 'Ruijian Cai', 'Jiacheng Ye'] | 2021-09-10 | null | https://aclanthology.org/2021.emnlp-main.213 | https://aclanthology.org/2021.emnlp-main.213.pdf | emnlp-2021-11 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 1.01788670e-01 5.18073253e-02 -4.66858625e-01 3.02620471e-01
-1.05289721e+00 -6.46591127e-01 9.34116900e-01 8.69062424e-01
-1.99157819e-01 8.51532996e-01 7.55001903e-01 -2.12140262e-01
2.98651066e-02 -8.97071064e-01 -1.03031218e+00 -3.84055912e-01
4.30496819e-02 3.47398221e-01 5.49512327e-01 -4.20909673e-01
6.21077001e-01 8.84650275e-02 -1.24985766e+00 4.01950747e-01
8.39073956e-01 8.34204793e-01 4.76842433e-01 5.66333652e-01
-5.28181970e-01 1.10623586e+00 -8.71922731e-01 -5.69733679e-01
1.45009279e-01 -5.68280637e-01 -6.83252037e-01 -9.40937549e-02
7.42139339e-01 -4.55478609e-01 -7.49049902e-01 9.90614176e-01
3.22908729e-01 -1.21960320e-01 5.13284504e-01 -1.21334159e+00
-1.03115416e+00 1.17791450e+00 -6.88602149e-01 2.72353917e-01
4.71564680e-01 -2.40984082e-01 1.48529947e+00 -8.72307956e-01
6.74200714e-01 1.05393898e+00 7.34655038e-02 1.92397669e-01
-1.04220390e+00 -8.03511202e-01 3.81672800e-01 4.74137068e-01
-1.45121276e+00 -2.74369866e-01 8.22297871e-01 -3.59315947e-02
1.03121769e+00 1.63148060e-01 6.72718942e-01 9.68686044e-01
5.81539690e-01 9.81638670e-01 6.69549465e-01 -5.08667171e-01
-1.38844714e-01 -6.32054955e-02 3.47927004e-01 7.47481585e-01
6.44449830e-01 -2.51363337e-01 -8.88469815e-01 -3.20253193e-01
4.12405878e-01 -1.89865083e-01 -5.35741389e-01 -1.94077879e-01
-1.53955781e+00 6.37257218e-01 3.07531565e-01 1.93010256e-01
-6.75174356e-01 4.05777663e-01 3.28681320e-01 2.40443900e-01
2.44569227e-01 5.55782735e-01 -4.45006609e-01 7.86351413e-02
-1.07405698e+00 5.40094435e-01 9.91806507e-01 1.33224463e+00
7.42062807e-01 -3.22205096e-01 -7.43256390e-01 4.76801604e-01
1.42969996e-01 4.56011653e-01 2.63633728e-01 -1.65242925e-01
9.35171425e-01 8.24279130e-01 2.25741744e-01 -1.43603170e+00
1.75438240e-01 -6.20840132e-01 -8.14398885e-01 -6.47829652e-01
-2.35271960e-01 1.30858690e-01 -9.84311521e-01 1.42355824e+00
1.16208784e-01 1.11071087e-01 2.86759976e-02 5.85136294e-01
1.01503098e+00 1.09615636e+00 -2.29055390e-01 -3.87479097e-01
1.17291212e+00 -1.08905959e+00 -1.00048935e+00 -2.01448709e-01
3.66899908e-01 -9.30673182e-01 9.57107961e-01 9.11055654e-02
-9.28547621e-01 -4.79528964e-01 -1.10319304e+00 -2.67365187e-01
-4.01275426e-01 3.12030673e-01 1.89170435e-01 -3.21083739e-02
-9.51680362e-01 3.87680113e-01 -3.59743357e-01 -9.99814942e-02
5.49900085e-02 4.51928191e-02 8.58693197e-03 -2.13103089e-02
-1.56956458e+00 5.47994256e-01 7.99081743e-01 -2.80318230e-01
-1.00958812e+00 -8.50412309e-01 -6.73111022e-01 2.75161624e-01
8.68366778e-01 -8.06095123e-01 1.12160039e+00 -2.55713046e-01
-8.61238003e-01 3.89001817e-01 -3.90926510e-01 -5.34045100e-01
1.43131509e-01 -3.05529982e-01 -6.76663339e-01 2.72639364e-01
3.54504406e-01 4.39133197e-01 1.17746663e+00 -1.15837920e+00
-9.88430798e-01 9.75789875e-03 1.77537560e-01 3.76571268e-01
-4.11990583e-01 -1.24106869e-01 -1.12062323e+00 -1.21156645e+00
-4.00072709e-02 -6.14682138e-01 1.16356276e-01 -4.21856284e-01
-9.68840837e-01 -2.73477316e-01 5.76985538e-01 -9.69939768e-01
1.98805749e+00 -1.73227358e+00 3.21170449e-01 1.58584669e-01
4.25327957e-01 -2.96077458e-04 -6.89926669e-02 1.13192248e+00
2.82941371e-01 2.61438638e-01 2.42325664e-01 2.00945228e-01
2.19354071e-02 -1.31037891e-01 -9.96541977e-01 -1.72075719e-01
-1.02085024e-01 1.13290656e+00 -1.01563835e+00 -5.11281073e-01
-2.60686368e-01 2.03891993e-01 -1.75213531e-01 3.61387700e-01
-6.04292810e-01 -1.86630711e-02 -5.76400161e-01 6.30097806e-01
3.79760951e-01 -6.26603246e-01 -8.81983787e-02 -5.19497275e-01
2.44797781e-01 8.09689760e-01 -1.06329191e+00 1.42069447e+00
-3.83163393e-01 3.32039684e-01 -3.02968651e-01 -3.87512028e-01
9.13525343e-01 3.40326548e-01 8.22285116e-02 -5.76034486e-01
-2.88344890e-01 2.08628073e-01 -2.45940119e-01 3.15258093e-02
1.10054588e+00 4.10781175e-01 -8.88404995e-02 6.44285083e-01
-7.72213936e-02 -2.19908312e-01 6.84854805e-01 1.09758735e+00
1.13682294e+00 -2.99346209e-01 3.91704470e-01 -6.45482391e-02
6.57672703e-01 -1.95905082e-02 1.20814636e-01 1.09087121e+00
4.33762163e-01 2.90522397e-01 7.54478276e-01 -1.68295264e-01
-8.53031754e-01 -8.81663620e-01 4.35750365e-01 9.17700469e-01
3.96256357e-01 -1.16826046e+00 -4.02328134e-01 -9.66471076e-01
1.58178538e-01 8.92524004e-01 -6.62793577e-01 -4.41489726e-01
-4.78100210e-01 -3.06103975e-01 3.99492264e-01 3.46884847e-01
3.23368698e-01 -8.59356046e-01 -1.25047565e-01 5.43258905e-01
-4.28927779e-01 -1.09001637e+00 -9.59769905e-01 1.37160800e-03
-3.32095832e-01 -1.02862287e+00 -1.00689948e+00 -5.68264127e-01
8.18891764e-01 5.13498902e-01 1.26390541e+00 2.85011053e-01
2.13184580e-01 3.96344095e-01 -6.38654590e-01 -2.71537989e-01
-5.93081892e-01 3.78854007e-01 -3.70141119e-01 4.89540026e-03
2.98145920e-01 -1.49134576e-01 -4.40537691e-01 -1.22280411e-01
-9.67623770e-01 3.30836445e-01 7.08420515e-01 8.33852768e-01
7.05915570e-01 2.94077009e-01 4.92305905e-01 -7.52794683e-01
1.21485651e+00 -6.06558800e-01 -6.63679540e-01 6.59216344e-01
-1.10078204e+00 4.46978271e-01 8.29824924e-01 -4.07757699e-01
-6.94642246e-01 -4.84797120e-01 3.82103086e-01 -2.79366791e-01
4.30232048e-01 8.48988473e-01 -4.59352136e-02 2.28396967e-01
2.81871170e-01 9.38233018e-01 -6.05744720e-01 -5.46851158e-01
5.12294114e-01 4.28209424e-01 4.72307205e-01 -5.36213279e-01
1.00473475e+00 4.59976830e-02 -9.60941240e-02 -4.52786356e-01
-7.97992826e-01 -4.88451123e-01 -4.64205831e-01 -8.71505365e-02
3.96469563e-01 -1.04850614e+00 -4.48775887e-02 3.54774803e-01
-1.40357578e+00 2.41248891e-01 -1.51854560e-01 9.13618803e-02
-6.56088218e-02 5.22603691e-01 -5.97969651e-01 -3.80089939e-01
-7.80147910e-01 -1.14441049e+00 1.26311255e+00 7.62940720e-02
-1.78694203e-01 -6.49257302e-01 -1.71331346e-01 -5.00955060e-03
1.65238842e-01 -1.44362345e-01 1.38372445e+00 -7.92815566e-01
-1.08659041e+00 -5.36827505e-01 -3.69002402e-01 1.33002564e-01
5.26993573e-01 -1.11928940e-01 -1.54825762e-01 -3.26422513e-01
-5.48543870e-01 -2.09513575e-01 1.11898553e+00 -1.40758410e-01
1.01959956e+00 -8.53106439e-01 -6.30694211e-01 2.31156737e-01
1.31370080e+00 2.61823952e-01 3.83399457e-01 2.56136954e-01
1.03159881e+00 2.26749614e-01 4.72153962e-01 4.32362854e-01
8.91821265e-01 6.55603349e-01 3.27965528e-01 2.17460543e-01
-3.01256925e-01 -1.00148737e+00 3.01565170e-01 1.01030827e+00
4.63774830e-01 -7.19308138e-01 -6.50305271e-01 8.25316012e-01
-1.67953229e+00 -8.17178071e-01 -2.10296854e-01 2.28029680e+00
1.24181294e+00 3.87913048e-01 -2.60763317e-01 1.49221450e-01
7.21411467e-01 6.28379524e-01 -3.06741536e-01 -2.60355063e-02
-8.45983699e-02 1.46274909e-01 6.26975477e-01 4.41751480e-01
-7.43775010e-01 1.25643706e+00 5.82264709e+00 1.03951955e+00
-9.76240098e-01 -2.83756793e-01 3.00496697e-01 1.10796690e-01
-6.84044302e-01 8.84062052e-02 -1.36545682e+00 6.99388385e-01
6.18696332e-01 -9.35443163e-01 5.55252612e-01 6.23644412e-01
-1.23744868e-02 -9.08858702e-02 -1.10000169e+00 7.18906045e-01
3.45748454e-01 -1.64968252e+00 8.39649081e-01 -8.53603631e-02
7.62176275e-01 -3.55084121e-01 -8.82464051e-02 1.70210525e-01
4.16970342e-01 -4.97421324e-01 8.15151274e-01 5.33022106e-01
5.72708845e-01 -8.80249441e-01 5.05697310e-01 2.29734853e-01
-1.42247295e+00 3.24158370e-01 -4.09788072e-01 3.64993453e-01
3.91011797e-02 7.64196336e-01 -1.12559700e+00 7.47001410e-01
4.72881794e-01 6.86918557e-01 -7.08565354e-01 7.90442288e-01
-8.07209015e-01 5.27364790e-01 -1.62699097e-03 -4.22426462e-01
3.23677897e-01 8.88567045e-02 6.87682211e-01 1.25670433e+00
4.00145441e-01 -8.03893656e-02 1.51259750e-01 7.87502944e-01
-5.21194339e-01 4.35238779e-01 -4.51198995e-01 -5.28717458e-01
8.35965276e-01 1.18225753e+00 -6.86608195e-01 -8.00672948e-01
-4.64619935e-01 1.12564647e+00 1.83256790e-01 5.59702337e-01
-6.32514477e-01 -8.25538933e-01 3.05284411e-01 2.22988501e-01
6.29076421e-01 -2.13884905e-01 1.99703276e-01 -1.30082846e+00
3.35466355e-01 -1.06216490e+00 4.24459130e-01 -9.45601523e-01
-9.77085471e-01 5.97548544e-01 1.82309583e-01 -1.10302413e+00
-4.27115351e-01 5.85761741e-02 -3.19745183e-01 1.06230152e+00
-1.66224980e+00 -1.18777704e+00 -1.88416898e-01 5.05560696e-01
5.31877637e-01 -1.51499048e-01 5.48271477e-01 4.13954211e-03
-1.81631222e-01 6.48124099e-01 1.01340152e-02 2.00910628e-01
7.29049623e-01 -1.22217953e+00 8.63653958e-01 1.12200844e+00
5.22898495e-01 8.76878440e-01 5.25336087e-01 -1.25145125e+00
-1.74406338e+00 -1.21133196e+00 1.44657111e+00 -2.03181952e-01
9.40224230e-01 -3.69589597e-01 -1.02385259e+00 5.19428849e-01
3.94938976e-01 -3.21434498e-01 3.16283524e-01 -2.87598938e-01
-5.34866631e-01 -1.77406162e-01 -3.49367589e-01 9.96375799e-01
8.67897213e-01 -7.38681257e-01 -9.29645836e-01 4.43444490e-01
1.19621015e+00 -5.68524837e-01 -3.91489625e-01 2.41739124e-01
2.45638981e-01 -4.80024755e-01 8.77141714e-01 -4.52067167e-01
6.65107489e-01 -3.90880287e-01 4.19553816e-02 -1.55794322e+00
-4.03533638e-01 -8.18697393e-01 -9.91802692e-01 1.33666611e+00
4.22526091e-01 -3.32619846e-01 4.17409569e-01 6.27497360e-02
3.97941433e-02 -6.83487713e-01 -5.79056680e-01 -8.24095964e-01
-3.66825253e-01 -1.33159652e-01 8.70075226e-01 7.28844583e-01
1.87362414e-02 6.69494629e-01 -6.11534357e-01 2.37877369e-01
5.37986338e-01 4.96189713e-01 8.62981439e-01 -9.53882217e-01
-2.58560419e-01 -2.55420387e-01 -4.83028181e-02 -1.43970311e+00
1.07304342e-01 -1.10931158e+00 -2.05071405e-01 -2.00033116e+00
2.27113321e-01 5.27154543e-02 -3.66558820e-01 5.07636130e-01
-6.03069782e-01 -2.70314217e-01 7.20123276e-02 5.46403408e-01
-6.15072370e-01 5.19741535e-01 1.35371149e+00 -5.40301740e-01
-2.02985033e-01 -2.34252140e-01 -1.04582453e+00 1.71008110e-01
4.24529970e-01 -5.65428376e-01 -6.05482876e-01 -4.19082612e-01
4.46921974e-01 2.18910977e-01 2.26177737e-01 -7.84500062e-01
5.61204791e-01 -1.79046810e-01 1.68397292e-01 -9.70887840e-01
1.46112870e-02 -6.54139519e-01 -1.40859131e-02 4.12627518e-01
-6.91569805e-01 2.58904636e-01 2.40974620e-01 8.37440550e-01
-2.55695552e-01 -1.24997281e-01 8.31854120e-02 -1.83180645e-01
-7.93999493e-01 5.16232252e-01 -2.41976559e-01 2.75863260e-01
6.22417152e-01 1.47846848e-01 -4.58061934e-01 -6.98857725e-01
-4.39595059e-02 2.65796393e-01 2.88196445e-01 8.05202842e-01
8.34732413e-01 -1.30490649e+00 -8.71009707e-01 2.11055726e-01
5.43540895e-01 -2.65942842e-01 -2.92897016e-01 4.21099752e-01
-2.82050282e-01 8.69321406e-01 2.04271674e-01 1.73405726e-02
-1.14637160e+00 8.24249685e-01 -1.34099498e-01 -6.39939010e-01
-5.87087631e-01 7.95864463e-01 8.85302126e-02 2.31686085e-01
3.53110939e-01 -7.19716132e-01 -2.85187572e-01 2.48698086e-01
8.61825943e-01 1.41848862e-01 3.22880834e-01 -5.83492756e-01
-1.73198178e-01 3.69017512e-01 -8.18584144e-01 -1.47803109e-02
8.64595413e-01 -2.41283655e-01 -4.02618378e-01 2.23581254e-01
1.07623279e+00 3.34890932e-01 -8.07912529e-01 -6.88550889e-01
1.86223507e-01 -5.01479864e-01 3.22699368e-01 -1.02317798e+00
-9.43056285e-01 4.87123162e-01 -9.62427184e-02 2.06547126e-01
1.21464694e+00 1.71724543e-01 1.16109538e+00 5.64951956e-01
5.69965601e-01 -9.36993361e-01 2.68319279e-01 5.97594917e-01
1.11891258e+00 -8.53358686e-01 1.73700809e-01 -4.26051289e-01
-4.81384277e-01 1.02682865e+00 4.87037629e-01 1.41455069e-01
3.66649389e-01 9.80430171e-02 -2.77475655e-01 -3.31235707e-01
-1.08021200e+00 -1.49003088e-01 8.32924485e-01 3.45733523e-01
3.52295488e-01 -2.26220876e-01 -6.69739187e-01 5.95926523e-01
-1.70104444e-01 -2.29599439e-02 6.81754947e-01 9.09647584e-01
-4.08408731e-01 -1.14098704e+00 -2.38426536e-01 8.30001116e-01
-6.85680211e-01 -8.34339797e-01 -7.82121241e-01 5.80783963e-01
-3.02993923e-01 8.57173562e-01 -1.13256864e-01 -6.30051970e-01
2.07061172e-01 9.12842229e-02 2.34334275e-01 -7.85367370e-01
-5.30674756e-01 3.24366659e-01 8.34668875e-02 -3.10318708e-01
1.05936639e-01 -1.04163885e-01 -1.15687275e+00 -3.14860404e-01
-4.52276587e-01 5.34435332e-01 3.36398870e-01 6.66174591e-01
6.69722557e-01 6.83988571e-01 6.09340310e-01 -2.04331115e-01
-5.14994621e-01 -8.34968448e-01 -5.72903991e-01 1.98405877e-01
4.09944654e-01 -2.66002029e-01 -2.88518727e-01 2.20425680e-01] | [12.314680099487305, 8.888960838317871] |
fd8bc1b3-ebf5-4470-b3ba-7120512bf674 | efficient-document-image-classification-using | 2106.13802 | null | https://arxiv.org/abs/2106.13802v1 | https://arxiv.org/pdf/2106.13802v1.pdf | Efficient Document Image Classification Using Region-Based Graph Neural Network | Document image classification remains a popular research area because it can be commercialized in many enterprise applications across different industries. Recent advancements in large pre-trained computer vision and language models and graph neural networks has lent document image classification many tools. However using large pre-trained models usually requires substantial computing resources which could defeat the cost-saving advantages of automatic document image classification. In the paper we propose an efficient document image classification framework that uses graph convolution neural networks and incorporates textual, visual and layout information of the document. We have rigorously benchmarked our proposed algorithm against several state-of-art vision and language models on both publicly available dataset and a real-life insurance document classification dataset. Empirical results on both publicly available and real-world data show that our methods achieve near SOTA performance yet require much less computing resources and time for model training and inference. This results in solutions than offer better cost advantages, especially in scalable deployment for enterprise applications. The results showed that our algorithm can achieve classification performance quite close to SOTA. We also provide comprehensive comparisons of computing resources, model sizes, train and inference time between our proposed methods and baselines. In addition we delineate the cost per image using our method and other baselines. | ['Glenn Fung', 'Qian You', 'Eric Bunch', 'Jaya Krishna Mandivarapu'] | 2021-06-25 | null | null | null | null | ['document-image-classification'] | ['computer-vision'] | [ 1.67698503e-01 -4.03542846e-01 -6.48969859e-02 -3.51341397e-01
-3.27433944e-01 -6.79474950e-01 7.01977491e-01 3.40316087e-01
-2.59536058e-01 1.33828819e-01 -2.38349736e-01 -8.04242909e-01
-1.22002825e-01 -1.11288035e+00 -5.21821737e-01 -4.20285374e-01
1.70543179e-01 6.77048981e-01 -9.98376906e-02 1.07359238e-01
6.36979759e-01 5.18477380e-01 -1.24788976e+00 5.41059673e-01
7.61114955e-01 1.27126956e+00 2.64151156e-01 1.19242656e+00
-5.81333399e-01 1.11460090e+00 -7.33835399e-01 -4.73257631e-01
1.96457192e-01 6.97911577e-03 -8.21794331e-01 5.38263202e-01
9.87969100e-01 -2.02802747e-01 -4.04513240e-01 1.02177179e+00
5.56167662e-01 -7.09381849e-02 7.25816548e-01 -1.22693372e+00
-1.49212539e+00 5.33805907e-01 -8.62720430e-01 1.33837029e-01
-3.71760759e-03 -1.24296583e-01 8.58346879e-01 -6.13184273e-01
6.31423056e-01 1.13048196e+00 1.01564109e+00 -1.44910291e-02
-8.80710363e-01 -3.63487750e-01 1.07135959e-01 3.15494239e-01
-1.25122011e+00 1.65544510e-01 8.09506178e-01 -3.74430984e-01
1.45110905e+00 1.45483702e-01 5.68348408e-01 8.10284138e-01
4.37663764e-01 8.01762760e-01 9.92402673e-01 -8.99166346e-01
-1.02542453e-01 2.63774842e-01 8.05983424e-01 1.50416732e+00
4.24544305e-01 -4.54299808e-01 -1.00573801e-01 -2.82378420e-02
7.31148362e-01 3.81061524e-01 1.25490623e-02 -1.12303281e-02
-8.11530709e-01 8.20680201e-01 4.22251523e-01 5.77451766e-01
-2.04172015e-01 4.58565980e-01 7.15785742e-01 2.22492740e-01
3.11557710e-01 -5.43441288e-02 -2.19983414e-01 1.87529653e-01
-9.23037708e-01 -7.96667188e-02 6.62696779e-01 1.18636096e+00
4.24395472e-01 1.19248837e-01 -3.99623245e-01 1.07389665e+00
1.55064061e-01 5.05088985e-01 5.13426006e-01 -5.16576231e-01
7.14531839e-01 1.14327610e+00 -5.03297091e-01 -1.56240821e+00
-4.07016933e-01 -5.80423117e-01 -1.16842794e+00 -3.93805653e-02
3.33835065e-01 2.75377452e-01 -1.32932234e+00 7.44029105e-01
-3.04369807e-01 -2.15411052e-01 -2.16981024e-01 3.93613696e-01
8.42274368e-01 8.37848604e-01 -8.57433230e-02 5.89514151e-02
1.35132444e+00 -1.58857417e+00 -6.66449785e-01 -1.28842250e-01
8.94857466e-01 -9.20216084e-01 1.32686818e+00 6.60329878e-01
-9.14195299e-01 -9.15412068e-01 -9.15610194e-01 -2.47411773e-01
-9.10030305e-01 5.58244288e-01 1.12899792e+00 9.95183408e-01
-1.22170174e+00 3.43861669e-01 -4.63757426e-01 -7.48301327e-01
7.66647458e-01 3.50103140e-01 -3.03462118e-01 -2.89737076e-01
-5.20752788e-01 5.42278707e-01 4.02340978e-01 1.39469638e-01
-5.45010448e-01 -9.22732204e-02 -6.33257985e-01 3.32650393e-01
3.15549552e-01 -5.24979472e-01 1.10878396e+00 -8.60513389e-01
-9.39804971e-01 9.95341122e-01 1.64980277e-01 -5.78956246e-01
3.41064751e-01 1.53252035e-01 -4.59828764e-01 1.13941990e-01
-2.73515522e-01 3.37176472e-01 7.37966895e-01 -1.07920980e+00
-6.41542435e-01 -6.20672643e-01 5.37673533e-02 -1.05081066e-01
-9.37830269e-01 -7.27121020e-03 -9.79875743e-01 -4.49461520e-01
-8.00076276e-02 -8.52127552e-01 7.79165626e-02 1.03985660e-01
-6.48188889e-01 -2.97230005e-01 1.19683945e+00 -6.79066241e-01
1.23046815e+00 -1.54073107e+00 -4.61680859e-01 3.44386488e-01
1.77467272e-01 3.99466604e-01 -1.98075145e-01 4.94251609e-01
3.15851718e-01 2.09950045e-01 6.75440878e-02 -2.89088100e-01
-8.75776410e-02 2.14536384e-01 -1.92648485e-01 3.01559299e-01
-4.29094136e-01 1.18625379e+00 -2.91126192e-01 -8.80828261e-01
3.17012608e-01 5.97700298e-01 -4.33800787e-01 -7.50215426e-02
-2.56356597e-01 -3.94938350e-01 -3.21793228e-01 9.40580666e-01
6.51189804e-01 -7.73207247e-01 3.01293641e-01 -2.40935758e-01
4.12063837e-01 -3.63904655e-01 -1.00619876e+00 1.54448640e+00
-4.89229411e-01 1.02153909e+00 -1.26354083e-01 -1.65384769e+00
9.57376838e-01 -1.09490193e-02 9.18784663e-02 -7.86850750e-01
3.39250207e-01 -1.89589709e-01 -1.90691203e-01 -6.89763606e-01
5.67151248e-01 5.76900661e-01 3.05682451e-01 4.29706633e-01
5.11184186e-02 -9.74311978e-02 4.65085775e-01 3.60782266e-01
8.89934301e-01 -1.90486237e-01 -1.01646958e-02 -2.56883234e-01
6.95487618e-01 2.89730996e-01 -2.70582557e-01 1.20317054e+00
4.04817685e-02 3.37394059e-01 3.73322815e-01 -7.87920654e-01
-1.04919434e+00 -5.94507396e-01 1.80336595e-01 1.25466251e+00
-1.98993161e-01 -3.93611223e-01 -1.03310001e+00 -1.00318468e+00
1.06385298e-01 4.01997656e-01 -6.44741058e-01 3.35730255e-01
-4.72543478e-01 -7.78920472e-01 7.43735254e-01 1.04260302e+00
8.84692073e-01 -9.72619355e-01 -1.23574421e-01 -1.05686761e-01
3.30697179e-01 -1.30390584e+00 -5.55939853e-01 -7.76949525e-02
-9.51836884e-01 -1.13325441e+00 -6.57686591e-01 -1.41982889e+00
9.44935322e-01 4.51876551e-01 1.30714095e+00 7.02717423e-01
-8.45418632e-01 5.87033689e-01 -1.31376758e-01 -7.30472267e-01
-3.33508044e-01 1.16850190e-01 -5.40667593e-01 -2.68486261e-01
6.21040583e-01 1.46322865e-02 -4.35126990e-01 -1.78222254e-01
-9.65838552e-01 1.84800401e-01 3.90395194e-01 1.04550600e+00
2.51532108e-01 5.17733276e-01 1.66748479e-01 -1.32073665e+00
1.29619670e+00 -4.30336483e-02 -1.03164685e+00 7.40817010e-01
-1.35241485e+00 -6.55063838e-02 7.13138342e-01 -1.60745934e-01
-8.15640986e-01 -2.02958867e-01 2.58983999e-01 -1.27928585e-01
-2.30913967e-01 8.90274644e-01 4.16171104e-01 -4.80116047e-02
6.80983782e-01 5.19345105e-01 -7.62040690e-02 -3.86528403e-01
2.60088950e-01 1.23316145e+00 4.31225330e-01 -4.38793272e-01
5.10411441e-01 5.30116916e-01 3.03280149e-02 -7.49115109e-01
-7.46400893e-01 -4.84572053e-01 -6.01235032e-01 -8.42144266e-02
7.53472865e-01 -5.62995076e-01 -8.89735639e-01 7.71039546e-01
-1.28577161e+00 -1.16626762e-01 4.59354818e-01 -8.71281624e-02
-2.08957300e-01 7.33147144e-01 -9.25094962e-01 -7.48061061e-01
-9.71988201e-01 -1.06718791e+00 8.76499295e-01 1.52455509e-01
3.18154693e-01 -1.56218576e+00 -3.14462304e-01 8.02713811e-01
6.56333923e-01 8.43833312e-02 1.19172347e+00 -4.57528621e-01
-5.21865368e-01 -7.75862515e-01 -9.31068599e-01 5.76864183e-01
1.89431563e-01 3.04539621e-01 -8.49038720e-01 -5.64643323e-01
-4.84713346e-01 -2.78258920e-01 9.58558857e-01 5.44242978e-01
1.88378561e+00 -3.14249188e-01 -4.73220527e-01 4.88611162e-01
1.86662829e+00 4.34849858e-01 5.55437386e-01 3.36637199e-01
1.13385141e+00 2.35546842e-01 1.62811354e-01 1.56864986e-01
2.60111094e-01 2.88170218e-01 1.99204877e-01 -3.80604386e-01
-2.26984978e-01 -2.75064278e-02 -5.69299459e-02 9.18953180e-01
-3.35657805e-01 -8.94995153e-01 -1.40525997e+00 5.15106440e-01
-1.98188758e+00 -9.51394498e-01 -1.97847366e-01 1.61115658e+00
4.86011297e-01 4.01961744e-01 -9.41400006e-02 3.21762204e-01
7.46667087e-01 -7.82148764e-02 -2.40341678e-01 -8.45862210e-01
-9.33471844e-02 3.08978081e-01 7.45513678e-01 3.58107597e-01
-1.05621755e+00 9.86671925e-01 6.86139584e+00 8.06703031e-01
-1.28344893e+00 -4.13792357e-02 8.08592081e-01 2.97266632e-01
2.89625257e-01 -2.94995397e-01 -7.93172300e-01 1.65808678e-01
9.73853290e-01 -1.70396835e-01 4.26218450e-01 1.28925836e+00
-1.30992487e-01 4.34997454e-02 -9.61470068e-01 1.22685468e+00
5.36120594e-01 -2.14435267e+00 6.10775113e-01 1.24024205e-01
7.68976986e-01 -9.29185469e-03 2.85054687e-02 2.80196160e-01
3.27197284e-01 -1.26893473e+00 3.95841658e-01 9.47360173e-02
7.25693226e-01 -7.72733808e-01 9.54080522e-01 2.39242986e-01
-1.12100804e+00 -2.40756467e-01 -4.89687592e-01 -1.42681748e-01
-4.69989717e-01 3.45836908e-01 -1.12349451e+00 3.66689444e-01
6.92150176e-01 6.72645867e-01 -1.09330237e+00 7.39172161e-01
2.87591308e-01 6.87445879e-01 9.08619985e-02 -4.78388637e-01
5.92214763e-01 -2.55125225e-01 -3.67909342e-01 1.72867453e+00
2.05704197e-01 -4.48268086e-01 3.94046038e-01 5.48866868e-01
-3.01180631e-01 3.35940152e-01 -7.46388495e-01 -5.76635182e-01
1.59694836e-01 1.27779555e+00 -1.11383700e+00 -8.04901779e-01
-7.51917481e-01 1.01932979e+00 4.23477978e-01 2.81026334e-01
-7.89986134e-01 -7.10821807e-01 -1.72471777e-01 1.13422014e-01
3.15370321e-01 -3.54683548e-01 -3.68456125e-01 -9.28696394e-01
-2.88554244e-02 -8.39606524e-01 6.16367340e-01 -8.60895693e-01
-1.11842811e+00 7.02551186e-01 -3.06104898e-01 -7.11406767e-01
-8.25825408e-02 -1.35959744e+00 -5.55594623e-01 5.98508120e-01
-1.37877727e+00 -1.78334844e+00 -7.79247165e-01 7.05064118e-01
7.81829715e-01 -6.27870739e-01 9.88164604e-01 3.24587882e-01
-7.17512071e-01 7.46823132e-01 3.89909178e-01 7.07790792e-01
3.64565402e-01 -1.24790037e+00 4.46656466e-01 7.38715470e-01
6.74879611e-01 6.88254654e-01 3.12888831e-01 -4.90489841e-01
-1.54605496e+00 -1.07383072e+00 6.49483502e-01 -3.53564620e-01
5.38673759e-01 -4.37784195e-01 -7.11398840e-01 8.22120309e-01
6.75146401e-01 1.41971648e-01 5.58364630e-01 9.82448608e-02
-5.47998190e-01 -1.60155371e-01 -1.11304700e+00 3.83901417e-01
8.52333546e-01 -6.77181721e-01 -7.98302516e-02 8.90795887e-01
2.47078687e-01 -1.61511332e-01 -5.97968936e-01 -1.12424642e-01
4.98592317e-01 -9.36222553e-01 8.82720172e-01 -7.14543402e-01
4.72314328e-01 4.95589226e-02 -1.45112932e-01 -7.13111937e-01
-3.67207259e-01 -1.53825805e-01 -2.29203448e-01 1.08106089e+00
3.02839279e-01 -5.12242496e-01 9.27837372e-01 3.98522764e-01
9.07012001e-02 -7.36258864e-01 -8.04514736e-02 -1.00613368e+00
1.04313292e-01 -5.62066853e-01 3.40999961e-01 1.11514902e+00
-3.12528610e-01 4.41512346e-01 -1.32270128e-01 1.03576425e-02
7.54880130e-01 4.29245234e-01 6.75199628e-01 -1.31249905e+00
-2.74041802e-01 -5.10002494e-01 -5.42367995e-01 -5.71084261e-01
-1.64794154e-03 -1.20777190e+00 -5.03713787e-01 -2.22880411e+00
6.59368634e-01 -2.30873585e-01 -4.50924098e-01 7.14929163e-01
2.44796500e-01 3.71037543e-01 1.71954669e-02 4.73576039e-02
-5.16417146e-01 -2.46415943e-01 1.18178809e+00 -8.26654613e-01
1.13841988e-01 -5.06307900e-01 -6.05235815e-01 7.41541088e-01
7.55489945e-01 -2.00062558e-01 -6.29352510e-01 -8.75105619e-01
1.03998922e-01 -3.12698424e-01 2.70964324e-01 -1.04055274e+00
4.15077239e-01 -9.01325494e-02 7.19024539e-01 -8.20116401e-01
-8.13740566e-02 -8.39556158e-01 -2.92583048e-01 6.25392795e-01
-4.58810091e-01 2.62924999e-01 3.32023442e-01 5.95418215e-01
-1.15024827e-01 -4.47634310e-01 6.41052485e-01 -4.12559271e-01
-7.22357333e-01 3.00322384e-01 -2.45088845e-01 -3.81495297e-01
9.13431466e-01 -5.70482671e-01 -6.98562384e-01 -2.80974716e-01
-4.13999826e-01 -2.87704971e-02 3.42554778e-01 7.31479406e-01
6.04658663e-01 -1.00194824e+00 -5.58377028e-01 2.62950271e-01
1.80670828e-01 -3.07000488e-01 -3.56291942e-02 3.90889555e-01
-1.24799180e+00 1.05216050e+00 -2.55013198e-01 -7.32956827e-01
-1.79827487e+00 8.10211003e-01 1.06944013e-02 -5.77037573e-01
-6.13022387e-01 7.70277858e-01 -2.41816342e-01 -3.65356624e-01
6.13039672e-01 -5.93765914e-01 -2.18676746e-01 -2.54296720e-01
5.28074801e-01 2.78396040e-01 2.66557515e-01 -2.62273937e-01
-1.45161718e-01 6.32376730e-01 -4.63673383e-01 3.73432964e-01
1.20781243e+00 6.46897256e-02 -1.43595174e-01 -1.28881950e-02
1.33665395e+00 -2.52240330e-01 -3.87828767e-01 -1.33760586e-01
7.25794062e-02 -4.55179155e-01 7.03493655e-01 -9.14962471e-01
-1.27460337e+00 1.06219923e+00 9.67808902e-01 6.86212122e-01
1.12625945e+00 -2.71314919e-01 7.83427179e-01 1.11254847e+00
2.53924817e-01 -1.55932403e+00 2.61894405e-01 3.73558670e-01
7.66463101e-01 -1.43309402e+00 2.17042238e-01 -4.05289203e-01
-4.21325803e-01 1.25639713e+00 5.85416973e-01 -2.60376841e-01
6.39858902e-01 2.90430367e-01 3.22476774e-01 -5.72572052e-01
-5.41933358e-01 1.78119317e-01 3.79414350e-01 6.70546234e-01
7.46823430e-01 -1.17138475e-01 -1.18348584e-01 1.36584923e-01
8.22440814e-03 2.29765341e-01 3.94760936e-01 1.16405106e+00
-2.62731940e-01 -9.45826471e-01 -3.08963984e-01 7.48726904e-01
-5.87957561e-01 -2.09118500e-01 -5.10578513e-01 8.43046129e-01
-1.29395455e-01 8.19091678e-01 2.43352100e-01 -1.34478509e-01
1.45936280e-01 1.19227998e-01 6.84017658e-01 -4.56208915e-01
-5.69710195e-01 -1.40577674e-01 6.40556887e-02 -1.72211707e-01
-3.15547258e-01 2.96383165e-02 -1.28256679e+00 -4.51768517e-01
-6.94571197e-01 -1.47197098e-01 1.12855375e+00 6.61065757e-01
4.96511519e-01 6.34723663e-01 1.99165493e-01 -4.67725575e-01
-4.85539585e-01 -1.08009779e+00 -5.86185098e-01 4.94627267e-01
-4.75236066e-02 -1.27626732e-01 1.37763306e-01 5.73060989e-01] | [11.458502769470215, 2.612680673599243] |
7b597170-c605-4b1a-8015-3f05e098950d | offline-rl-with-realistic-datasets | 2211.01052 | null | https://arxiv.org/abs/2211.01052v2 | https://arxiv.org/pdf/2211.01052v2.pdf | Offline RL With Realistic Datasets: Heteroskedasticity and Support Constraints | Offline reinforcement learning (RL) learns policies entirely from static datasets, thereby avoiding the challenges associated with online data collection. Practical applications of offline RL will inevitably require learning from datasets where the variability of demonstrated behaviors changes non-uniformly across the state space. For example, at a red light, nearly all human drivers behave similarly by stopping, but when merging onto a highway, some drivers merge quickly, efficiently, and safely, while many hesitate or merge dangerously. Both theoretically and empirically, we show that typical offline RL methods, which are based on distribution constraints fail to learn from data with such non-uniform variability, due to the requirement to stay close to the behavior policy to the same extent across the state space. Ideally, the learned policy should be free to choose per state how closely to follow the behavior policy to maximize long-term return, as long as the learned policy stays within the support of the behavior policy. To instantiate this principle, we reweight the data distribution in conservative Q-learning (CQL) to obtain an approximate support constraint formulation. The reweighted distribution is a mixture of the current policy and an additional policy trained to mine poor actions that are likely under the behavior policy. Our method, CQL (ReDS), is simple, theoretically motivated, and improves performance across a wide range of offline RL problems in Atari games, navigation, and pixel-based manipulation. | ['Sergey Levine', 'Yevgen Chebotar', 'Quan Vuong', 'Aviral Kumar', 'Anikait Singh'] | 2022-11-02 | null | null | null | null | ['atari-games'] | ['playing-games'] | [-3.37620169e-01 1.03294671e-01 -6.55339837e-01 -2.86134034e-01
-6.75522327e-01 -9.71403956e-01 3.07419151e-01 -1.51617909e-02
-8.46678972e-01 9.96519387e-01 1.33647630e-02 -6.27261400e-01
-4.33929026e-01 -7.27544963e-01 -9.67900157e-01 -6.71120346e-01
-2.66221583e-01 4.55647498e-01 2.49246284e-01 -3.07852477e-01
1.57454699e-01 6.16414726e-01 -1.68947816e+00 -3.29236060e-01
8.10407877e-01 7.57365882e-01 9.84371156e-02 7.30066776e-01
1.16656370e-01 9.25285399e-01 -6.43270135e-01 -1.98921829e-01
6.47888005e-01 -4.10662115e-01 -3.31757337e-01 -1.79641915e-03
2.05867589e-01 -7.28984237e-01 -6.00103617e-01 9.73532498e-01
2.10643813e-01 6.04076326e-01 3.55914384e-01 -1.56150818e+00
-1.85143277e-01 4.84938711e-01 -6.24077678e-01 1.48684308e-01
1.76719110e-02 7.94248998e-01 8.65154564e-01 -1.04246810e-01
5.92049003e-01 9.74330485e-01 3.73168260e-01 6.01996362e-01
-1.21822286e+00 -7.61533916e-01 6.49106741e-01 1.30026340e-01
-8.98004889e-01 -2.76818693e-01 5.72336555e-01 -1.57463193e-01
7.26390958e-01 -3.34343389e-02 8.66219521e-01 1.02666688e+00
2.87694156e-01 9.11582053e-01 1.19152033e+00 1.38494030e-01
6.47873640e-01 -3.48468795e-02 -8.46833810e-02 6.50052547e-01
2.77737826e-01 6.78340614e-01 -4.64096665e-01 -2.18700051e-01
5.57026803e-01 -5.97897060e-02 9.12365839e-02 -7.52751768e-01
-6.34073079e-01 7.73504615e-01 1.33121490e-01 -2.82391727e-01
-3.75708044e-01 5.16697943e-01 2.70751476e-01 5.63299716e-01
-2.35312264e-02 4.29681778e-01 -5.20051301e-01 -7.23332942e-01
-7.54958272e-01 8.58459890e-01 7.45536566e-01 9.97396827e-01
1.00547910e+00 1.16270952e-01 -2.28915915e-01 4.20816600e-01
3.32668833e-02 7.36606538e-01 2.70244479e-01 -1.57378268e+00
6.59903586e-01 2.11933643e-01 8.12404215e-01 -6.19107187e-01
-3.48719239e-01 -2.66757607e-01 -1.69389024e-01 7.63866305e-01
7.98820257e-01 -6.35416627e-01 -7.53082752e-01 2.07502699e+00
3.40033650e-01 1.48261651e-01 -1.26111493e-01 8.86632264e-01
-2.92306304e-01 5.08550763e-01 1.26467705e-01 -3.21206242e-01
6.12793803e-01 -4.64217335e-01 -5.48066914e-01 -3.85940820e-01
5.85914850e-01 -4.28503007e-02 1.27773869e+00 5.19666791e-01
-1.18604612e+00 -3.79794121e-01 -9.81692672e-01 4.08047229e-01
-1.33866847e-01 -3.08477968e-01 5.68243980e-01 5.87159514e-01
-8.67978454e-01 7.34858632e-01 -9.71612692e-01 -4.19919230e-02
5.75963974e-01 2.60580420e-01 -1.12618707e-01 8.94275084e-02
-1.09988308e+00 1.02996373e+00 1.79053321e-01 -6.95993472e-03
-1.42214715e+00 -7.71302700e-01 -7.10318983e-01 5.10643348e-02
9.92439389e-01 -2.25876793e-01 1.61928785e+00 -9.57138121e-01
-1.75082517e+00 2.67499030e-01 1.00566395e-01 -8.35713446e-01
8.92527103e-01 -3.70789796e-01 -1.55093729e-01 -1.87326998e-01
8.15282166e-02 6.50990427e-01 8.43535066e-01 -1.17466807e+00
-8.50272238e-01 -3.42114717e-01 3.30801636e-01 5.20402014e-01
-4.98823859e-02 -4.60100144e-01 -3.04039240e-01 -1.78509504e-01
-6.16638184e-01 -1.13834238e+00 -4.60840940e-01 1.32227927e-01
-2.18658205e-02 -2.97841430e-01 7.47648060e-01 -2.91010499e-01
1.15792680e+00 -2.19085956e+00 -9.16060880e-02 4.61146921e-01
3.43775116e-02 2.16149818e-02 -1.33807212e-01 3.20264399e-01
3.31434280e-01 -2.35332906e-01 -2.50102937e-01 -1.16515346e-02
1.52135283e-01 6.42710626e-01 -6.22168660e-01 7.21831977e-01
-4.91150767e-02 7.94566035e-01 -1.30138028e+00 -1.44022807e-01
2.17315510e-01 -3.71812314e-01 -6.86504543e-01 3.03584814e-01
-5.47081411e-01 3.92513216e-01 -5.71450114e-01 3.21295261e-01
3.68297517e-01 2.73833662e-01 1.51569486e-01 4.71787304e-01
-3.61251116e-01 -2.57031880e-02 -1.26038694e+00 1.37186027e+00
-3.71606290e-01 5.18838942e-01 1.56056464e-01 -1.06929517e+00
6.99416220e-01 -1.07908860e-01 9.19839680e-01 -9.19205427e-01
6.74953684e-02 1.79252513e-02 7.76515752e-02 -6.25997782e-01
4.24478054e-01 -2.08880529e-01 -3.06517661e-01 7.83743620e-01
-1.08463719e-01 -2.28091136e-01 2.20218554e-01 1.06113397e-01
1.22953212e+00 3.69964749e-01 1.32790044e-01 -3.42986658e-02
-8.91811773e-02 2.76690274e-01 7.71976590e-01 1.27619636e+00
-8.30256283e-01 -1.31220266e-01 8.42412710e-01 -1.97890565e-01
-9.70763087e-01 -1.39710534e+00 2.00466767e-01 1.08488488e+00
3.11709255e-01 4.19746861e-02 -4.86816198e-01 -1.01334584e+00
5.74072957e-01 9.20594096e-01 -5.34106016e-01 -4.64541435e-01
-5.51186144e-01 -1.57417446e-01 4.00166690e-01 4.30510134e-01
5.12092531e-01 -9.62703228e-01 -1.00388861e+00 3.25153857e-01
2.11288229e-01 -7.76539922e-01 -4.78447467e-01 2.04379544e-01
-6.69189513e-01 -1.05990636e+00 -3.49593371e-01 -1.30027235e-01
4.23548460e-01 1.15554877e-01 8.60150754e-01 -1.93287328e-01
-1.33830443e-01 7.82095075e-01 -3.77571471e-02 -4.84421194e-01
-1.89029410e-01 -2.17182919e-01 2.86944717e-01 -1.53013498e-01
2.85561174e-01 -3.39454263e-01 -5.93679368e-01 2.53185958e-01
-5.12600422e-01 -4.38109964e-01 4.30380613e-01 6.80791616e-01
5.64763665e-01 2.74013400e-01 7.43819892e-01 -7.01719165e-01
6.67370915e-01 -7.71226943e-01 -1.05767012e+00 1.59484580e-01
-7.32932389e-01 2.25268900e-01 8.66391778e-01 -6.48524225e-01
-9.56504166e-01 1.36981189e-01 1.99498847e-01 -6.60987139e-01
-1.85103014e-01 1.47553027e-01 -1.10958844e-01 1.51412666e-01
7.41694272e-01 1.17406301e-01 4.16399091e-01 -7.91078880e-02
7.36212373e-01 4.48158413e-01 5.14174461e-01 -9.88496482e-01
8.48656833e-01 5.21535933e-01 -9.41750705e-02 -5.62887430e-01
-9.27434802e-01 -2.41206303e-01 -1.59036860e-01 -7.28018403e-01
5.74083745e-01 -6.91957712e-01 -1.12949908e+00 1.76968232e-01
-3.67766857e-01 -1.09060097e+00 -8.27042341e-01 5.11813164e-01
-1.17614400e+00 -4.44135852e-02 -1.64346814e-01 -1.18456459e+00
4.17244375e-01 -8.46847594e-01 5.95623851e-01 4.24249440e-01
6.93650767e-02 -7.72169769e-01 2.47837409e-01 7.76507147e-03
3.78862500e-01 1.70377061e-01 7.57434666e-01 -1.95575789e-01
-5.06905556e-01 5.67588806e-02 2.15630680e-01 3.00203353e-01
1.42579928e-01 -1.73769727e-01 -5.50018609e-01 -5.73400915e-01
-1.54430166e-01 -6.59793913e-01 5.86534142e-01 5.11787236e-01
1.43052101e+00 -3.75564963e-01 -1.10904597e-01 3.40994686e-01
1.20624328e+00 3.63967985e-01 4.09789354e-01 4.30204004e-01
1.69979215e-01 7.39719570e-01 9.59820211e-01 7.63075948e-01
4.10569519e-01 3.78824145e-01 4.66065645e-01 2.20878690e-01
2.99753547e-01 -7.32333422e-01 6.91294312e-01 -1.04876988e-01
2.90747166e-01 -2.38686949e-02 -5.98176658e-01 4.69277203e-01
-2.09911418e+00 -1.19040966e+00 4.96462822e-01 2.54262590e+00
8.43936861e-01 5.45651495e-01 6.30172253e-01 -3.92434537e-01
1.56717613e-01 4.37551476e-02 -1.56164420e+00 -4.96869415e-01
1.36014387e-01 1.42081246e-01 1.14041066e+00 5.11320114e-01
-7.41253257e-01 9.71507788e-01 6.98111534e+00 7.69787133e-01
-9.38449800e-01 2.24110018e-02 6.10158443e-01 -7.08186388e-01
-2.79548198e-01 1.41421780e-01 -7.96494365e-01 5.02436757e-01
1.06144595e+00 -4.16912824e-01 9.53503609e-01 9.81218398e-01
6.63131356e-01 -4.72788572e-01 -9.99331832e-01 7.60102630e-01
-3.87454152e-01 -1.00933826e+00 -5.34669995e-01 1.15813009e-01
7.41263449e-01 1.05635643e-01 2.44801551e-01 8.67027700e-01
1.07634258e+00 -9.08679426e-01 8.87183607e-01 5.11515021e-01
5.80021620e-01 -1.08128917e+00 1.10644370e-01 7.23275781e-01
-8.42015743e-01 -6.99619710e-01 -2.73916036e-01 -8.92669559e-02
2.00795993e-01 2.17114180e-01 -4.32243198e-01 -9.42290202e-03
6.47037983e-01 6.00081384e-01 -2.15280518e-01 8.81729960e-01
-2.64289767e-01 6.62890911e-01 -3.58319491e-01 -1.85248330e-01
5.49798310e-01 -4.04417574e-01 4.47745055e-01 6.41246676e-01
1.65506098e-02 1.14646591e-01 5.22737801e-01 8.25271368e-01
1.33970469e-01 -3.36311758e-01 -9.18751180e-01 -2.03137025e-02
5.62968194e-01 8.95385742e-01 -3.26866925e-01 -2.16704875e-01
-3.23585153e-01 4.89427716e-01 6.22484624e-01 6.60244823e-01
-9.72707450e-01 -1.97234690e-01 1.18744159e+00 7.74688348e-02
1.64195061e-01 -5.76140642e-01 -9.59481448e-02 -7.17394292e-01
2.62235086e-02 -7.68233120e-01 4.11729395e-01 -4.66193467e-01
-1.26206791e+00 -1.68950483e-01 3.36583555e-01 -1.41587925e+00
-4.57147598e-01 -4.01868463e-01 -4.52125698e-01 5.45288026e-01
-1.54037142e+00 -3.68538558e-01 1.65285468e-01 7.22300470e-01
4.93842900e-01 -2.52870530e-01 3.65055203e-02 1.68646500e-02
-5.21384001e-01 6.10280633e-01 1.84701636e-01 -2.57738322e-01
6.28136635e-01 -1.23996687e+00 -1.01097502e-01 6.36424661e-01
-1.25090852e-01 4.75062817e-01 8.76389205e-01 -7.86285698e-01
-1.54480219e+00 -1.06717575e+00 -6.78636283e-02 -4.25394833e-01
8.58661711e-01 -2.74727196e-01 -7.55784214e-01 7.24806130e-01
-1.54884934e-01 1.07580349e-01 2.96912253e-01 7.76041895e-02
-1.05426714e-01 -3.46492797e-01 -1.19968641e+00 9.39809442e-01
1.01411116e+00 -4.81715381e-01 -4.56548303e-01 2.94723660e-01
4.23380435e-01 -4.69368130e-01 -6.07075989e-01 1.08362094e-01
6.19575620e-01 -8.68710995e-01 6.26253068e-01 -1.07795656e+00
-2.47576628e-02 -2.63903290e-01 3.62560898e-02 -1.45759475e+00
-1.10845350e-01 -9.52832758e-01 -4.36046094e-01 7.12531328e-01
2.91974247e-01 -6.19668961e-01 1.09951770e+00 8.68783593e-01
-1.67657994e-02 -9.07467782e-01 -1.10046434e+00 -1.22991729e+00
4.13978219e-01 -5.67559004e-01 5.85923553e-01 4.42159891e-01
4.75302786e-02 -2.88221031e-01 -4.35645521e-01 1.19765773e-01
7.61673868e-01 4.06092368e-02 1.02717984e+00 -5.75175583e-01
-5.94373941e-01 -5.51654637e-01 -4.71502282e-02 -1.46602786e+00
5.35727262e-01 -4.24933910e-01 4.80160981e-01 -1.14451826e+00
-3.92368250e-02 -8.73410821e-01 -2.42595315e-01 4.52467948e-01
-2.46037692e-02 -4.89517063e-01 1.66035175e-01 1.08309723e-01
-6.40953362e-01 7.74327159e-01 1.30076706e+00 1.22755207e-01
-4.81596172e-01 3.24528486e-01 -6.55207574e-01 4.66245621e-01
8.78313363e-01 -4.68586683e-01 -8.25879753e-01 -1.87807217e-01
2.70952851e-01 4.95813698e-01 1.85156718e-01 -7.51123369e-01
2.14025989e-01 -9.27739203e-01 1.30780652e-01 -4.91200358e-01
2.09789306e-01 -7.88480699e-01 -9.82649028e-02 5.83720863e-01
-5.87025821e-01 7.55640864e-02 1.54010639e-01 1.02813303e+00
2.81413615e-01 -7.27805821e-03 8.01604033e-01 -9.05203745e-02
-7.02118278e-01 5.45024753e-01 -6.45535290e-01 4.18572664e-01
1.39111364e+00 -1.84348822e-01 -2.63056606e-01 -7.54169345e-01
-5.88587940e-01 9.07673001e-01 4.48010623e-01 4.51127768e-01
4.64750201e-01 -1.08135402e+00 -2.97787994e-01 2.25572720e-01
1.24318134e-02 -8.18120837e-02 2.28293374e-01 5.93654931e-01
1.21378666e-02 4.91305254e-02 -2.52077073e-01 -4.04179215e-01
-4.11479831e-01 5.91862738e-01 7.24427342e-01 -1.89579502e-01
-7.84196317e-01 4.94416267e-01 -1.12916276e-01 -4.64588612e-01
3.29622358e-01 -2.36253798e-01 6.31645322e-02 -1.80449843e-01
4.50134218e-01 4.18599069e-01 -2.22447529e-01 -1.19668618e-01
-1.45404115e-02 2.70337045e-01 2.62737274e-02 -5.16252935e-01
1.05922830e+00 -1.20636269e-01 5.59862256e-01 6.08345509e-01
7.80804038e-01 -1.85517654e-01 -2.37174225e+00 -7.49705285e-02
4.81181853e-02 -6.80081308e-01 -9.11711752e-02 -7.56245553e-01
-9.36646938e-01 4.17937845e-01 5.96475780e-01 1.09005757e-01
7.81264365e-01 -1.80611476e-01 6.62165105e-01 5.53776860e-01
7.76245713e-01 -1.68479860e+00 2.11949527e-01 4.64064896e-01
4.58889812e-01 -1.21428859e+00 -1.91575140e-02 5.03865719e-01
-1.04789817e+00 7.77011633e-01 9.15789366e-01 -4.83289838e-01
7.15937734e-01 4.31297004e-01 -1.28957957e-01 -2.07217745e-02
-9.84868050e-01 -3.43806684e-01 -2.98059404e-01 7.11518764e-01
-4.37155485e-01 3.18762332e-01 3.22393104e-02 1.88122466e-01
-2.86688179e-01 -4.07657772e-02 5.88745713e-01 1.15018582e+00
-8.42094481e-01 -9.20152724e-01 -1.37834072e-01 7.09156215e-01
7.86480010e-02 6.32129490e-01 -3.05643044e-02 9.04257655e-01
-1.10528897e-02 9.57082391e-01 3.64842325e-01 -1.13518015e-01
4.93165344e-01 -1.67716250e-01 5.32310724e-01 -4.93710250e-01
-2.86892772e-01 -1.94989696e-01 -5.03555462e-02 -1.11693668e+00
-2.15509683e-02 -1.00651956e+00 -1.49725175e+00 -4.29703683e-01
7.07981586e-02 1.87851693e-02 5.07458627e-01 1.16279566e+00
1.37903810e-01 1.89867869e-01 1.05988324e+00 -3.95973295e-01
-1.31908774e+00 -5.06303906e-01 -7.34497726e-01 3.18522811e-01
6.05093479e-01 -9.57570136e-01 -4.14976478e-01 -4.48983997e-01] | [4.138972759246826, 2.196963310241699] |
48d2f864-8d6a-4146-abb0-b25ecd64c206 | llms-as-factual-reasoners-insights-from | 2305.14540 | null | https://arxiv.org/abs/2305.14540v1 | https://arxiv.org/pdf/2305.14540v1.pdf | LLMs as Factual Reasoners: Insights from Existing Benchmarks and Beyond | With the recent appearance of LLMs in practical settings, having methods that can effectively detect factual inconsistencies is crucial to reduce the propagation of misinformation and improve trust in model outputs. When testing on existing factual consistency benchmarks, we find that a few large language models (LLMs) perform competitively on classification benchmarks for factual inconsistency detection compared to traditional non-LLM methods. However, a closer analysis reveals that most LLMs fail on more complex formulations of the task and exposes issues with existing evaluation benchmarks, affecting evaluation precision. To address this, we propose a new protocol for inconsistency detection benchmark creation and implement it in a 10-domain benchmark called SummEdits. This new benchmark is 20 times more cost-effective per sample than previous benchmarks and highly reproducible, as we estimate inter-annotator agreement at about 0.9. Most LLMs struggle on SummEdits, with performance close to random chance. The best-performing model, GPT-4, is still 8\% below estimated human performance, highlighting the gaps in LLMs' ability to reason about facts and detect inconsistencies when they occur. | ['Chien-Sheng Wu', 'Shafiq Joty', 'Caiming Xiong', 'Alexander R. Fabbri', 'Divyansh Agarwal', 'Wojciech Kryściński', 'Philippe Laban'] | 2023-05-23 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [-2.37485781e-01 7.71294773e-01 -6.08002543e-01 -5.42555273e-01
-1.28781450e+00 -5.75247526e-01 9.28231180e-01 6.29866481e-01
-3.01298350e-01 1.07390904e+00 2.82698572e-01 -4.24442589e-01
-1.33545905e-01 -2.85693914e-01 -7.81462133e-01 5.88180497e-02
4.15956713e-02 6.12861276e-01 2.99928665e-01 -1.28192067e-01
6.12923265e-01 -2.35106096e-01 -1.08796728e+00 6.00722134e-01
9.06547725e-01 7.33491361e-01 -7.28140354e-01 2.86797076e-01
-7.05455020e-02 1.50456488e+00 -1.02101326e+00 -1.20938838e+00
-2.84976158e-02 -2.57824481e-01 -1.32209527e+00 -5.14110625e-01
1.02648330e+00 -1.28141627e-01 2.62185961e-01 1.20149112e+00
2.28281662e-01 -3.57576489e-01 4.25333261e-01 -1.63762021e+00
-6.37905955e-01 1.23310125e+00 -3.60839933e-01 5.55968761e-01
7.00191677e-01 -5.69817331e-03 1.20120656e+00 -7.88426757e-01
1.05498552e+00 1.34396720e+00 1.28617680e+00 3.00663859e-01
-1.16326201e+00 -6.28658712e-01 3.45316045e-02 4.26660836e-01
-1.32243168e+00 -7.11944163e-01 3.10534000e-01 -3.43419999e-01
1.39912736e+00 6.45347059e-01 -2.83895675e-02 1.10728526e+00
3.91704857e-01 3.79924446e-01 1.35293734e+00 -4.19767082e-01
3.13990712e-01 3.66587818e-01 3.64411086e-01 6.62709177e-01
9.83349919e-01 -3.67950976e-01 -8.24872375e-01 -7.17415392e-01
6.31915703e-02 -6.48489594e-01 -7.83221722e-02 5.82604557e-02
-1.12557971e+00 6.15821183e-01 3.15733373e-01 4.88228828e-01
-8.54845494e-02 1.31560147e-01 5.07705629e-01 5.46285629e-01
6.66988134e-01 8.81147325e-01 -6.45835996e-01 -3.39880764e-01
-1.07021785e+00 8.02249670e-01 1.19848144e+00 6.84735715e-01
2.58744270e-01 -3.79949987e-01 -1.38245806e-01 9.11246061e-01
2.01746568e-01 3.46542150e-01 3.16109985e-01 -1.51425719e+00
9.02206361e-01 7.64722109e-01 5.43399394e-01 -1.45803058e+00
-4.96027082e-01 -4.44419026e-01 -5.20570219e-01 1.63673803e-01
6.33236885e-01 3.55260223e-01 -2.64102161e-01 1.64945769e+00
7.76481107e-02 -1.77112669e-01 -2.17621606e-02 7.73615777e-01
6.36909306e-01 8.61465707e-02 2.96271741e-01 -1.63770214e-01
1.19012654e+00 -6.99777126e-01 -7.48970449e-01 -6.74153924e-01
8.23352456e-01 -9.28301692e-01 8.77821326e-01 4.59057927e-01
-1.34303439e+00 -2.64084414e-02 -1.11240304e+00 -1.24937899e-01
-3.35089713e-01 -3.11172962e-01 6.37390018e-01 4.92580891e-01
-9.14990902e-01 6.00664020e-01 -4.67200190e-01 -2.86586553e-01
4.80932623e-01 -1.14701785e-01 -5.18965304e-01 -9.83441323e-02
-1.33928275e+00 1.59459698e+00 3.38084549e-01 -7.42199868e-02
-5.24414241e-01 -7.31458426e-01 -8.58041048e-01 -1.91811621e-01
5.85995555e-01 -5.68118572e-01 1.63563883e+00 -7.08683074e-01
-7.76803195e-01 1.28281391e+00 -3.44523609e-01 -6.79977596e-01
9.18183684e-01 -2.02986240e-01 -6.20958686e-01 -3.09371620e-01
7.30617583e-01 2.73274720e-01 3.56438637e-01 -1.40182436e+00
-5.94760597e-01 -3.57001908e-02 1.58973970e-02 -8.06078091e-02
2.98729748e-01 2.49208853e-01 -1.88287824e-01 -3.99344742e-01
2.47388914e-01 -7.29037762e-01 -1.20129272e-01 -1.74897835e-01
-5.67974567e-01 -3.80864352e-01 3.97688627e-01 -7.33090281e-01
1.42116404e+00 -1.67832398e+00 -3.97134751e-01 2.23560989e-01
3.99622679e-01 4.16857213e-01 -2.02210154e-02 4.33980972e-01
-6.48326650e-02 6.89658761e-01 -3.16365689e-01 -5.53510904e-01
1.82307601e-01 4.06268030e-01 -3.56185377e-01 3.27187985e-01
1.00530952e-01 8.82268727e-01 -1.01025999e+00 -6.57115638e-01
-8.31870288e-02 -2.94496596e-01 -4.92374808e-01 -1.14094049e-01
-4.61432815e-01 -7.58442953e-02 8.14887658e-02 6.92558587e-01
6.46009684e-01 -4.97280955e-01 3.74671102e-01 2.60232519e-02
3.56623763e-03 9.00560796e-01 -1.09786224e+00 1.50871897e+00
-2.57412642e-01 6.52899027e-01 -1.25396565e-01 -6.83883190e-01
7.13345051e-01 3.15344393e-01 -8.47159475e-02 -1.07039559e+00
-3.58974546e-01 5.95057964e-01 -5.96687309e-02 -6.21678829e-01
5.13454914e-01 1.68061823e-01 -1.91051468e-01 6.90110207e-01
-1.82118773e-01 -3.09624225e-01 1.04861736e-01 5.47837257e-01
1.20145059e+00 -3.95610690e-01 5.98710120e-01 -4.34715003e-01
1.97573408e-01 4.73475218e-01 7.03962684e-01 1.20024419e+00
-2.43727624e-01 2.66816258e-01 8.45881879e-01 -7.61383355e-01
-1.00237823e+00 -8.61000299e-01 -2.39257276e-01 5.62868118e-01
9.98368338e-02 -8.58975649e-01 -5.66724300e-01 -1.00013614e+00
1.72713548e-01 1.20667398e+00 -5.92982471e-01 -4.92349938e-02
-2.29903027e-01 -8.14129233e-01 9.54610586e-01 2.25323603e-01
6.56073928e-01 -6.94841027e-01 -5.66791713e-01 2.03699008e-01
-7.20361710e-01 -1.27268231e+00 2.20884472e-01 -1.20381825e-01
-5.63899279e-01 -1.43715775e+00 2.29903147e-01 -1.76476553e-01
4.70958024e-01 -1.43287361e-01 1.89992678e+00 5.54701090e-01
-7.09602684e-02 1.13605939e-01 -1.55563757e-01 -4.37132835e-01
-9.78365898e-01 -7.45666698e-02 -8.72191191e-02 -8.20452690e-01
6.20589614e-01 -1.94161683e-01 -2.67069370e-01 4.20992374e-01
-7.14531839e-01 -3.57271582e-02 3.11313570e-01 7.86108017e-01
6.75559491e-02 1.04028098e-01 6.27947986e-01 -1.25449169e+00
8.91195357e-01 -7.26834059e-01 -3.35628688e-01 4.32716429e-01
-1.25085235e+00 3.40130925e-02 3.40985864e-01 1.70547485e-01
-1.06673443e+00 -8.58422756e-01 3.61484326e-02 3.43499660e-01
1.11555755e-01 6.64971948e-01 3.09264600e-01 5.44153117e-02
1.13364422e+00 -5.49784005e-01 -1.97312385e-01 -1.92942783e-01
2.54963636e-01 6.43031538e-01 6.86113060e-01 -7.54666567e-01
4.84309822e-01 3.09048891e-01 -3.73388946e-01 -3.34419683e-02
-1.40044188e+00 -6.59180284e-02 -2.20318064e-01 1.17211595e-01
1.06384568e-01 -8.90244305e-01 -5.27310789e-01 2.28936613e-01
-1.45294404e+00 -3.87309700e-01 9.33124125e-02 1.28976375e-01
-1.52151182e-01 3.66376877e-01 -7.27318764e-01 -6.93206906e-01
-2.16969490e-01 -7.90444195e-01 8.99446666e-01 -3.22451234e-01
-9.94445682e-01 -1.16571772e+00 2.64463335e-01 6.74385011e-01
6.84913278e-01 3.32688808e-01 9.51932907e-01 -1.00316823e+00
-2.86538333e-01 -3.93500209e-01 -3.32210690e-01 3.11487257e-01
-1.14398398e-01 5.00505157e-02 -8.81083310e-01 6.33151531e-02
3.88581678e-02 -4.77453023e-01 5.58633089e-01 -5.11888415e-03
7.81851530e-01 -7.81644762e-01 -3.92515540e-01 -1.84681460e-01
1.30672503e+00 -2.99120009e-01 5.55433750e-01 7.96014488e-01
3.00823152e-01 4.45317090e-01 6.37474000e-01 2.64365494e-01
8.90827119e-01 7.71202147e-01 2.64818996e-01 1.15464136e-01
-2.95933150e-02 -1.84091523e-01 2.09089354e-01 5.27668118e-01
-2.25365739e-02 -5.11743762e-02 -1.44131434e+00 4.90266591e-01
-2.25091362e+00 -1.11387897e+00 -4.89823401e-01 2.09778214e+00
1.24781537e+00 6.53709173e-01 -3.20257783e-01 1.32306591e-01
4.33159649e-01 1.21129304e-01 -1.17894307e-01 -7.46133566e-01
-2.49729499e-01 -1.54028326e-01 3.27884942e-01 8.11376393e-01
-8.19441915e-01 7.99440324e-01 7.94120550e+00 5.62248886e-01
-5.53000033e-01 4.12382275e-01 7.72795796e-01 -2.98660956e-02
-6.42115295e-01 1.54833868e-01 -6.54397726e-01 5.75404942e-01
1.16732121e+00 -1.88612044e-01 2.22754553e-01 8.58559430e-01
5.15764654e-02 -6.46871448e-01 -1.35361183e+00 7.67170072e-01
2.96033740e-01 -1.54817331e+00 -1.67489380e-01 -3.79017711e-01
1.11473536e+00 1.79922864e-01 -3.75912011e-01 3.28539044e-01
6.95738256e-01 -1.20070243e+00 9.63742256e-01 5.08020163e-01
2.28142172e-01 -3.58164787e-01 1.28705978e+00 4.07686859e-01
-4.13547903e-01 8.73606373e-03 -2.93761522e-01 -4.54065710e-01
-3.38448095e-03 1.10406196e+00 -9.47693288e-01 5.41329980e-01
7.67785907e-01 4.96284097e-01 -1.12988555e+00 8.76267374e-01
-4.55992609e-01 5.74272037e-01 -2.64447242e-01 2.09833473e-01
2.68152617e-02 5.54938436e-01 3.26435506e-01 1.57554412e+00
-1.50960356e-01 -3.78176749e-01 3.41899320e-02 1.03398883e+00
-1.98831588e-01 -3.98326755e-01 -6.03025317e-01 2.89844334e-01
8.87065351e-01 9.05647635e-01 -2.80891418e-01 -7.28758335e-01
-3.40770304e-01 6.18362248e-01 6.60248458e-01 1.02177054e-01
-9.34986413e-01 1.43293560e-01 4.11273539e-01 -1.53181046e-01
-4.97470707e-01 4.10449915e-02 -6.17587626e-01 -1.24766850e+00
5.87435603e-01 -1.48733449e+00 7.20119357e-01 -8.80401731e-01
-1.49569595e+00 4.08766448e-01 1.10995449e-01 -5.94358444e-01
-5.13879776e-01 -3.00076187e-01 -3.41146499e-01 6.79451525e-01
-1.46539080e+00 -8.88242722e-01 -3.86982381e-01 2.68514901e-01
1.64519757e-01 2.48832241e-01 9.60633934e-01 2.70202816e-01
-4.83824879e-01 5.86191356e-01 -4.63601828e-01 -4.84065078e-02
1.16246533e+00 -1.33271861e+00 8.10766518e-01 9.50735927e-01
2.75608510e-01 1.00477493e+00 9.88331556e-01 -9.23102617e-01
-8.29751492e-01 -8.14053416e-01 1.55187118e+00 -1.40533137e+00
9.07165229e-01 -1.35275438e-01 -1.19510043e+00 1.03911722e+00
2.75244296e-01 -5.64571694e-02 4.52876389e-01 5.18156648e-01
-1.12401235e+00 6.76926151e-02 -1.22520673e+00 4.81614381e-01
1.15439653e+00 -8.07302058e-01 -1.15346217e+00 5.96981823e-01
2.57053137e-01 -7.98709810e-01 -8.28317463e-01 4.22735304e-01
5.18228650e-01 -1.32107234e+00 6.65608406e-01 -7.90549338e-01
6.55144036e-01 -3.61997277e-01 -1.18884772e-01 -1.26214576e+00
-1.13324635e-01 -3.85300010e-01 1.04000635e-01 1.44244587e+00
7.83698261e-01 -8.16977620e-01 1.95935845e-01 1.26944458e+00
8.74797702e-02 -2.59015918e-01 -1.02404594e+00 -7.88596034e-01
-4.98344516e-03 -9.00669873e-01 3.96004558e-01 1.52027249e+00
4.02976662e-01 -8.98143835e-03 -2.21112773e-01 3.02717477e-01
7.24658310e-01 1.75403766e-02 7.41421521e-01 -1.04085314e+00
-1.64459869e-01 -5.06769121e-01 -3.17026705e-01 -2.61806309e-01
2.01888934e-01 -8.43859673e-01 2.68325079e-02 -1.60062492e+00
6.37528777e-01 -5.94591737e-01 -1.50471032e-01 8.24690998e-01
-3.67839396e-01 3.52763712e-01 -1.68217465e-01 3.95076632e-01
-1.08337390e+00 -2.56220728e-01 5.30292273e-01 -1.65877610e-01
3.26089531e-01 -5.21194100e-01 -8.50484371e-01 9.24158216e-01
6.79593921e-01 -8.17693472e-01 -1.29132405e-01 -6.37137592e-01
8.17815125e-01 -2.33300582e-01 8.20640445e-01 -9.77864325e-01
4.61180687e-01 -2.02813894e-01 1.53216049e-01 -1.25499830e-01
-1.93641588e-01 -4.93753195e-01 4.12542790e-01 3.89681429e-01
-6.91552460e-01 2.62145787e-01 2.90570676e-01 2.03257442e-01
-2.91869432e-01 -3.51234019e-01 3.88907462e-01 -3.61862779e-01
-7.85792768e-01 -4.99449730e-01 -8.09924901e-02 4.94619220e-01
5.86967111e-01 3.97999376e-01 -1.22676301e+00 -1.67150512e-01
-4.10779417e-01 2.00535685e-01 5.62814832e-01 5.16983688e-01
1.05060190e-01 -1.09402919e+00 -8.23545218e-01 -2.33823985e-01
4.61758971e-01 -2.25938857e-01 -1.04970686e-01 1.01285112e+00
-4.12220031e-01 6.33882880e-01 2.10464280e-02 -5.39505601e-01
-9.76174176e-01 1.66074738e-01 5.49611926e-01 -4.16656077e-01
-3.04091811e-01 7.28189409e-01 -5.39295495e-01 -5.27512789e-01
1.84644744e-01 -4.41680491e-01 3.24040920e-01 -2.95252115e-01
6.66543782e-01 4.23074394e-01 5.16721189e-01 -1.26074165e-01
-7.37987220e-01 1.96067333e-01 -5.16175181e-02 -1.00183293e-01
1.04535997e+00 -7.39534572e-02 -6.86308324e-01 5.92120886e-01
6.82871819e-01 4.01184231e-01 -6.69625640e-01 -1.45319805e-01
7.02796340e-01 -6.68549538e-01 -1.62594408e-01 -1.52607954e+00
-4.79648441e-01 1.90801769e-01 -1.06235845e-02 5.13166964e-01
3.34632546e-01 1.02968775e-01 3.19954813e-01 5.14178336e-01
6.85922325e-01 -1.19645011e+00 -1.07505389e-01 3.92247260e-01
1.13918996e+00 -1.72850943e+00 3.74818355e-01 -4.08926785e-01
-6.59750283e-01 8.23763490e-01 7.30432153e-01 1.83791548e-01
8.76146033e-02 4.34843421e-01 3.80145937e-01 -5.94760716e-01
-1.10454035e+00 5.13372779e-01 1.83282062e-01 1.56657755e-01
6.14588916e-01 1.06755286e-01 -5.51944315e-01 5.27480125e-01
-3.31464708e-01 8.62312466e-02 5.29504478e-01 6.47151828e-01
-1.91446662e-01 -9.20625627e-01 -4.08742160e-01 3.54157090e-01
-7.03997374e-01 -1.74999967e-01 -6.34959102e-01 1.01241970e+00
1.73957184e-01 1.31101072e+00 -1.92151628e-02 -1.51547506e-01
3.09308767e-01 1.67335212e-01 2.82934874e-01 -4.94871408e-01
-7.43542552e-01 -7.80438364e-01 9.97272015e-01 -1.15707755e+00
-5.91367066e-01 -6.34005129e-01 -9.59729970e-01 -6.90309405e-01
-1.13352679e-01 3.89115751e-01 5.82226455e-01 1.27778459e+00
3.90074641e-01 1.63749326e-02 1.21813826e-02 -5.82663566e-02
-6.48713112e-01 -8.32819164e-01 -7.96165597e-03 8.19225132e-01
2.52203465e-01 -4.99020875e-01 -6.03144944e-01 -1.18824057e-01] | [9.765311241149902, 8.327980995178223] |
0bf6b095-b4ab-4900-af32-7c69cc310e32 | synfeal-a-data-driven-simulator-for-end-to | 2305.18260 | null | https://arxiv.org/abs/2305.18260v1 | https://arxiv.org/pdf/2305.18260v1.pdf | Synfeal: A Data-Driven Simulator for End-to-End Camera Localization | Collecting real-world data is often considered the bottleneck of Artificial Intelligence, stalling the research progress in several fields, one of which is camera localization. End-to-end camera localization methods are still outperformed by traditional methods, and we argue that the inconsistencies associated with the data collection techniques are restraining the potential of end-to-end methods. Inspired by the recent data-centric paradigm, we propose a framework that synthesizes large localization datasets based on realistic 3D reconstructions of the real world. Our framework, termed Synfeal: Synthetic from Real, is an open-source, data-driven simulator that synthesizes RGB images by moving a virtual camera through a realistic 3D textured mesh, while collecting the corresponding ground-truth camera poses. The results validate that the training of camera localization algorithms on datasets generated by Synfeal leads to better results when compared to datasets generated by state-of-the-art methods. Using Synfeal, we conducted the first analysis of the relationship between the size of the dataset and the performance of camera localization algorithms. Results show that the performance significantly increases with the dataset size. Our results also suggest that when a large localization dataset with high quality is available, training from scratch leads to better performances. Synfeal is publicly available at https://github.com/DanielCoelho112/synfeal. | ['Paulo Dias', 'Miguel Oliveira', 'Daniel Coelho'] | 2023-05-29 | null | null | null | null | ['camera-localization'] | ['computer-vision'] | [-2.90713727e-01 -2.73044020e-01 2.74193566e-02 -3.05074453e-01
-8.97426486e-01 -9.77575779e-01 5.69427729e-01 -9.91969034e-02
-6.40448749e-01 4.75504816e-01 5.95728233e-02 -8.18140730e-02
2.17601471e-02 -6.65489972e-01 -1.10188723e+00 -5.47541797e-01
1.92130700e-01 6.79668665e-01 1.39985174e-01 6.16243705e-02
1.93312958e-01 5.27359068e-01 -1.67145455e+00 -1.08837657e-01
4.67798740e-01 9.41101313e-01 3.46501708e-01 8.02896798e-01
1.27042100e-01 4.93591517e-01 -4.79560822e-01 -1.32104263e-01
5.80034673e-01 -2.73441017e-01 -5.13278365e-01 1.25592917e-01
5.44435322e-01 -1.43869162e-01 -5.14242090e-02 9.35054243e-01
4.87812310e-01 -4.71295379e-02 1.41762188e-02 -1.48926413e+00
-2.07965463e-01 1.00228608e-01 -2.98862010e-01 -2.82115459e-01
8.72412145e-01 2.46095836e-01 5.74970186e-01 -9.42338526e-01
8.88813078e-01 8.26775789e-01 1.17832363e+00 2.46385127e-01
-1.18169606e+00 -4.66228217e-01 -1.48985282e-01 -2.21767932e-01
-1.81316853e+00 -4.48527545e-01 5.87949276e-01 -4.23970997e-01
5.54312944e-01 1.40699133e-01 6.18620396e-01 1.36357319e+00
-1.44574359e-01 3.99385333e-01 1.17830145e+00 -5.59759617e-01
6.04739964e-01 1.73411921e-01 -3.38195592e-01 7.27531016e-01
4.07603413e-01 2.47610509e-01 -6.25657260e-01 -1.40400171e-01
8.37455094e-01 -7.89488927e-02 -3.70551556e-01 -8.48126531e-01
-1.74075997e+00 6.80340469e-01 5.45518219e-01 1.04514107e-01
-3.99933994e-01 3.14941466e-01 7.80589432e-02 5.53200506e-02
3.22421312e-01 5.51950157e-01 -4.15062577e-01 -4.03694093e-01
-1.18840492e+00 2.91954070e-01 8.10919821e-01 1.17179632e+00
9.84131098e-01 -4.64037329e-01 4.80718106e-01 2.47890860e-01
2.70198286e-01 7.56299794e-01 4.66592103e-01 -1.32995665e+00
3.21178228e-01 5.61050892e-01 4.20450926e-01 -9.76952255e-01
-4.06974345e-01 -2.25060046e-01 -3.21401626e-01 1.48122445e-01
7.22297728e-01 -3.55278254e-01 -5.47702670e-01 1.55255103e+00
4.06811059e-01 1.57948375e-01 1.57252893e-01 1.12941051e+00
4.41087127e-01 4.88045841e-01 -4.59151298e-01 2.17645600e-01
9.58782136e-01 -1.01497996e+00 -4.37426865e-01 -3.34769309e-01
6.60900533e-01 -7.52943456e-01 1.04624045e+00 5.03888249e-01
-6.86526239e-01 -6.16500139e-01 -9.94840145e-01 -3.90382558e-02
-5.75969517e-01 3.09182167e-01 7.00844646e-01 5.46066463e-01
-1.28599572e+00 3.61448497e-01 -1.03063464e+00 -8.58369231e-01
1.67350173e-01 3.13477755e-01 -7.77271509e-01 -2.14585707e-01
-6.26141667e-01 7.96942472e-01 4.03715819e-01 -5.56660518e-02
-8.08652818e-01 -5.30525267e-01 -7.32630968e-01 -5.54230034e-01
3.69894534e-01 -6.97734594e-01 1.22355771e+00 -1.00872755e+00
-1.52709079e+00 7.23632514e-01 -1.36066034e-01 -2.93985307e-01
8.50990951e-01 -2.77535707e-01 -1.96119636e-01 3.72074097e-02
3.62121552e-01 7.98088372e-01 4.26729470e-01 -1.49748707e+00
-6.06951654e-01 -3.20753098e-01 -4.98752706e-02 3.02473661e-02
-1.99990030e-02 -3.47529769e-01 -7.68351138e-01 -1.21593378e-01
6.87144250e-02 -1.15212703e+00 -3.41076195e-01 2.52674162e-01
-2.62618303e-01 3.82781774e-01 7.38243341e-01 -2.15038478e-01
6.39252305e-01 -2.16626883e+00 1.12204343e-01 1.10569723e-01
-5.75385913e-02 -1.02687672e-01 -8.78258497e-02 6.84229195e-01
1.17246337e-01 -1.36273757e-01 -5.81424609e-02 -7.71744251e-01
5.62083833e-02 1.84699371e-01 -2.35800236e-01 8.16881239e-01
-3.54329705e-01 5.55898666e-01 -1.11977232e+00 -2.26232395e-01
5.70763230e-01 7.40894973e-01 -4.53095794e-01 2.49679789e-01
-1.55438155e-01 7.63147414e-01 -2.07717836e-01 7.69645035e-01
7.37479985e-01 -3.69872421e-01 -1.01898119e-01 -6.60913885e-02
-3.33109856e-01 -1.20922685e-01 -1.38790989e+00 2.50262642e+00
-4.87016886e-01 6.67175174e-01 2.37982273e-02 -4.05803800e-01
8.55020106e-01 1.05413713e-01 4.65105981e-01 -5.35372078e-01
2.39687905e-01 3.27100486e-01 -5.07705092e-01 -2.73367226e-01
6.04878187e-01 2.93035388e-01 -3.17408234e-01 3.25867474e-01
4.37675714e-02 -4.46155310e-01 1.35240257e-01 8.89399722e-02
1.16537178e+00 6.12751901e-01 3.14111084e-01 -1.88070521e-01
1.84866816e-01 6.10993266e-01 3.59583884e-01 7.31248856e-01
3.07486649e-03 1.06522584e+00 1.42111406e-01 -4.97000486e-01
-1.18574607e+00 -1.05058932e+00 -6.04560487e-02 5.87436855e-01
3.49519551e-01 -6.51887238e-01 -1.23055840e+00 -4.84963745e-01
-1.42853120e-02 4.15544987e-01 -6.26028776e-01 2.70285457e-01
-1.70226425e-01 -6.35458648e-01 7.07025409e-01 4.42843080e-01
6.80812538e-01 -5.45696497e-01 -1.27552140e+00 -7.97665026e-03
-1.45556152e-01 -1.64148951e+00 -1.18783146e-01 1.95362121e-02
-7.21011639e-01 -1.19066226e+00 -4.81669426e-01 -2.31661543e-01
8.79003763e-01 3.93651545e-01 1.11472702e+00 3.91315408e-02
-1.31212607e-01 7.53730297e-01 -5.57962418e-01 -4.34013009e-01
-2.37457544e-01 2.61028409e-01 1.73617736e-01 -5.93596958e-02
3.00501823e-01 -4.12684917e-01 -6.45573616e-01 4.75742340e-01
-8.78147006e-01 1.46202236e-01 4.49130028e-01 2.65786916e-01
8.44520688e-01 -2.12570325e-01 -1.87996134e-01 -6.57777429e-01
3.56349587e-01 -4.54371154e-01 -1.17934537e+00 -3.49856764e-02
-4.14491922e-01 -2.62322165e-02 7.40999103e-01 -3.23678583e-01
-4.60529476e-01 6.67189598e-01 7.71722794e-02 -7.28036463e-01
-5.76023281e-01 3.29236656e-01 2.87660081e-02 -3.46766084e-01
7.60448933e-01 7.80642852e-02 6.32894039e-02 -3.81301671e-01
4.32986975e-01 4.50749993e-01 7.72357762e-01 -5.44276118e-01
8.29079628e-01 8.61933529e-01 1.14444839e-02 -7.58313477e-01
-6.93577826e-01 -4.74440932e-01 -8.84646654e-01 -2.91578561e-01
7.06401646e-01 -1.00986719e+00 -5.76919496e-01 4.02808756e-01
-1.12736130e+00 -5.16591907e-01 -3.50774914e-01 5.98677039e-01
-6.76804304e-01 -4.99467412e-03 1.18161617e-02 -6.09616816e-01
9.20522660e-02 -1.46083581e+00 1.51566052e+00 1.99098289e-01
-1.30672127e-01 -1.05100667e+00 4.75425541e-01 2.53248274e-01
3.36047411e-01 7.32354820e-01 -8.32851008e-02 -2.72164166e-01
-8.62533212e-01 -5.41820407e-01 1.98915787e-02 1.01521641e-01
-1.89185962e-01 1.26850307e-01 -1.12332261e+00 -3.02037179e-01
-2.42247120e-01 -2.02821702e-01 1.71790391e-01 1.29223168e-01
8.59624922e-01 1.18374631e-01 -4.18794364e-01 9.98067737e-01
1.94001961e+00 -2.75509357e-01 4.35868025e-01 6.34211123e-01
6.97989643e-01 3.03730428e-01 7.51856446e-01 4.28007901e-01
6.74395442e-01 8.90071094e-01 7.23042548e-01 -2.97170341e-01
-5.89121953e-02 -5.62296867e-01 3.66622090e-01 6.56207442e-01
-2.03740597e-02 -2.38557264e-01 -1.20812368e+00 5.71186841e-01
-1.93737185e+00 -6.16318107e-01 -3.39891374e-01 2.51107430e+00
2.85673738e-01 -1.57269806e-01 1.75117731e-01 -5.99619783e-02
4.74307895e-01 -5.03201485e-02 -3.29434365e-01 6.77828789e-02
-2.87280940e-02 7.48310797e-03 1.10788298e+00 5.73129714e-01
-8.99823368e-01 1.01012063e+00 6.02166700e+00 2.99552739e-01
-1.38500404e+00 3.07673246e-01 1.22713298e-01 -9.08806249e-02
2.36587301e-01 3.92780721e-01 -7.67249823e-01 7.12230980e-01
1.18056214e+00 1.01298667e-01 6.95396125e-01 1.21334112e+00
2.77839333e-01 -5.89218974e-01 -1.22281551e+00 1.18219042e+00
2.88906842e-01 -1.48327887e+00 -4.43977714e-01 3.52834195e-01
6.88605368e-01 5.85887432e-01 -2.60795146e-01 -1.54613510e-01
4.20753509e-01 -8.63221109e-01 1.17615485e+00 5.35873473e-01
6.55818522e-01 -5.49566329e-01 9.23687696e-01 6.06223822e-01
-1.05248702e+00 2.10087255e-01 -2.72849888e-01 -7.97254816e-02
2.06315905e-01 6.39205933e-01 -1.01586974e+00 6.92444384e-01
7.79221952e-01 5.98494411e-01 -1.08126867e+00 1.21371090e+00
-3.09706748e-01 3.31794679e-01 -6.79517746e-01 9.42240562e-03
2.02013761e-01 -2.01969355e-01 3.48586589e-01 1.03220916e+00
5.25095105e-01 -4.39814925e-01 2.70959824e-01 8.99344146e-01
4.99326773e-02 -2.84254193e-01 -1.02799749e+00 3.28474283e-01
6.85193241e-01 1.32469881e+00 -8.62011254e-01 -1.61623023e-03
-2.18931004e-01 1.02969658e+00 1.94070697e-01 1.30993411e-01
-1.11087847e+00 -2.20119983e-01 3.50552320e-01 2.34756693e-01
1.41735658e-01 -4.84798878e-01 -8.67908522e-02 -1.20689631e+00
1.88263133e-01 -7.69183993e-01 4.93533723e-02 -1.10251904e+00
-9.39371586e-01 8.40958059e-01 1.03275873e-01 -1.39969194e+00
-2.52936214e-01 -6.09360516e-01 -1.88233808e-01 5.13933241e-01
-1.25882375e+00 -1.28532827e+00 -1.02501416e+00 7.04653919e-01
2.49195471e-01 1.66202962e-01 9.33134794e-01 2.41907850e-01
-3.76633763e-01 3.52296919e-01 2.71148950e-01 1.76436260e-01
6.63496614e-01 -1.11211646e+00 4.67446536e-01 1.05509913e+00
5.99971592e-01 5.74706376e-01 7.62949347e-01 -3.12234730e-01
-1.84288895e+00 -1.02063084e+00 4.74293351e-01 -1.12116647e+00
5.53114057e-01 -6.98551357e-01 -2.68633515e-01 8.14818501e-01
9.76796970e-02 5.73788345e-01 3.09685111e-01 -2.69264460e-01
-1.39368594e-01 -2.39449367e-01 -1.21757066e+00 3.40725243e-01
9.90232229e-01 -4.94321138e-01 -9.79267135e-02 3.72821838e-01
4.52806532e-01 -6.82800531e-01 -7.70109534e-01 1.51093811e-01
4.90703374e-01 -1.31098795e+00 8.81672204e-01 2.47216254e-01
1.26511067e-01 -7.00374305e-01 -5.12295842e-01 -1.48427987e+00
2.20974222e-01 -5.87777913e-01 2.30232641e-01 1.00601292e+00
2.70130277e-01 -6.64227664e-01 8.80837679e-01 4.90436375e-01
2.03785896e-01 -2.16626003e-01 -8.48805070e-01 -9.33907807e-01
-2.05300733e-01 -6.08030558e-01 7.63808191e-01 9.35716748e-01
-4.05616462e-01 1.28383428e-01 -2.90003475e-02 4.55604911e-01
7.37798870e-01 -3.41999047e-02 1.55861676e+00 -1.17416179e+00
-1.42998129e-01 3.02428380e-02 -6.85883045e-01 -8.37798119e-01
3.15826610e-02 -5.15373707e-01 1.10087819e-01 -1.41056871e+00
-1.38501421e-01 -6.00888968e-01 3.24966311e-01 3.47372711e-01
3.70690078e-01 3.94008070e-01 2.06950769e-01 2.51259804e-01
-8.00104678e-01 1.05216876e-01 7.54668832e-01 4.19465244e-01
8.84405598e-02 -2.71397918e-01 -3.50192636e-01 8.86150241e-01
6.16099000e-01 -5.45256913e-01 -3.65775436e-01 -7.69008696e-01
1.92088857e-01 -6.36431873e-02 7.38765895e-01 -1.56334162e+00
5.69252133e-01 8.11694488e-02 3.80319625e-01 -4.99929637e-01
4.91869181e-01 -1.29995465e+00 6.34610713e-01 1.76848471e-01
-5.33254305e-03 5.78142047e-01 1.33787647e-01 3.95038188e-01
-1.10138290e-01 -4.70144004e-02 5.68152487e-01 -3.81977290e-01
-8.32580268e-01 1.42736495e-01 1.95878055e-02 1.94725953e-03
1.20953190e+00 -4.30579245e-01 -3.15331936e-01 -3.19268674e-01
-2.52949506e-01 -2.16709316e-01 1.23448968e+00 4.91190910e-01
3.57366204e-01 -1.17192054e+00 -3.71467918e-01 3.55894864e-01
4.00690645e-01 3.90598953e-01 -3.18839461e-01 8.46293628e-01
-1.15303922e+00 3.58239889e-01 -2.76152212e-02 -9.99583423e-01
-8.81174326e-01 5.27905166e-01 3.48784626e-01 1.05132826e-01
-5.19876838e-01 5.35015821e-01 -3.12623441e-01 -8.41490746e-01
2.71757543e-01 -4.79117841e-01 6.22147620e-01 -2.69575298e-01
4.88714129e-01 3.09033424e-01 3.07834923e-01 -7.87605286e-01
-6.05539858e-01 7.46483624e-01 4.87820119e-01 -4.27713394e-01
1.39365983e+00 -1.56479761e-01 -3.06865219e-02 4.57337737e-01
1.22448301e+00 3.28358084e-01 -1.42312872e+00 1.85659796e-01
8.19215029e-02 -9.37028468e-01 2.78921932e-01 -9.08841252e-01
-1.10637653e+00 5.75746596e-01 8.17413628e-01 -4.29495797e-02
8.91845167e-01 -4.49264422e-03 4.25054580e-01 3.57927322e-01
1.29658663e+00 -9.02099311e-01 -2.04940692e-01 2.75039613e-01
5.73911428e-01 -1.38455677e+00 6.38781935e-02 -2.05507562e-01
-5.02232432e-01 9.17589843e-01 4.87176657e-01 -3.16981584e-01
2.97283888e-01 5.67860842e-01 2.77278066e-01 -1.66470751e-01
-2.88921058e-01 -1.48794621e-01 -3.00127149e-01 6.32566333e-01
1.01177715e-01 -4.25759405e-02 3.82535666e-01 2.23756000e-01
-4.05190349e-01 2.32947245e-01 5.53968847e-01 1.03003418e+00
3.95861603e-02 -1.14566243e+00 -8.07470918e-01 -2.93764532e-01
-2.12736130e-02 1.42059892e-01 -6.37187362e-01 1.18402374e+00
3.61233711e-01 9.63666499e-01 -2.81667598e-02 -3.68570060e-01
4.59783792e-01 -3.06706846e-01 4.95801449e-01 -4.21014488e-01
-4.69902664e-01 -2.50745684e-01 -1.55932114e-01 -9.58916187e-01
-5.83258390e-01 -7.34079421e-01 -9.92376387e-01 -4.86474782e-01
-2.78208375e-01 3.05184901e-01 1.20863807e+00 6.39331520e-01
7.84790277e-01 1.25293955e-01 5.89624107e-01 -1.30687153e+00
-1.68507293e-01 -5.89128077e-01 -1.85193419e-01 3.11205506e-01
4.33351398e-01 -6.42116010e-01 -4.17266309e-01 1.39559090e-01] | [7.58809757232666, -2.1733407974243164] |
e5b54665-8299-4a19-9776-8078b18bfaac | ecqed-emotion-cause-quadruple-extraction-in | 2306.03969 | null | https://arxiv.org/abs/2306.03969v2 | https://arxiv.org/pdf/2306.03969v2.pdf | ECQED: Emotion-Cause Quadruple Extraction in Dialogs | The existing emotion-cause pair extraction (ECPE) task, unfortunately, ignores extracting the emotion type and cause type, while these fine-grained meta-information can be practically useful in real-world applications, i.e., chat robots and empathic dialog generation. Also the current ECPE is limited to the scenario of single text piece, while neglecting the studies at dialog level that should have more realistic values. In this paper, we extend the ECPE task with a broader definition and scenario, presenting a new task, Emotion-Cause Quadruple Extraction in Dialogs (ECQED), which requires detecting emotion-cause utterance pairs and emotion and cause types. We present an ECQED model based on a structural and semantic heterogeneous graph as well as a parallel grid tagging scheme, which advances in effectively incorporating the dialog context structure, meanwhile solving the challenging overlapped quadruple issue. Via experiments we show that introducing the fine-grained emotion and cause features evidently helps better dialog generation. Also our proposed ECQED system shows exceptional superiority over baselines on both the emotion-cause quadruple or pair extraction tasks, meanwhile being highly efficient. | ['Chong Teng', 'Bobo Li', 'Jingye Li', 'Shengqiong Wu', 'Hao Fei', 'Fei Li', 'Donghong Ji', 'Li Zheng'] | 2023-06-06 | null | null | null | null | ['emotion-cause-pair-extraction'] | ['natural-language-processing'] | [-8.88087526e-02 3.22144657e-01 8.45860168e-02 -4.37955678e-01
-6.83003604e-01 -6.79900706e-01 7.96446621e-01 -1.74834561e-02
-5.81396222e-02 1.09713423e+00 6.33122623e-01 7.05392063e-02
-1.32024065e-01 -6.36939883e-01 -1.04009993e-01 -6.39416158e-01
1.45803258e-01 6.85302913e-01 -3.08240987e-02 -1.01110780e+00
2.63042301e-01 -3.96923022e-03 -1.34933901e+00 4.48817730e-01
8.21243584e-01 8.50155115e-01 2.81881779e-01 5.29431701e-01
-5.29762745e-01 1.10423052e+00 -9.90391612e-01 -8.56271386e-01
-3.11779261e-01 -6.78521454e-01 -1.27150989e+00 -1.65397123e-01
-1.56287894e-01 -2.03630045e-01 -8.19155872e-02 7.72956669e-01
6.39699876e-01 3.04957747e-01 7.37864912e-01 -1.88181138e+00
-3.85600656e-01 7.19795704e-01 -2.79603124e-01 -2.31997579e-01
9.39698219e-01 -1.37907878e-01 1.33524573e+00 -5.44648945e-01
8.28427553e-01 1.64380360e+00 6.55565798e-01 7.92593181e-01
-4.84908342e-01 -6.04880750e-01 8.89777392e-02 3.94766957e-01
-8.85903299e-01 -1.63577601e-01 1.03637230e+00 -1.98742479e-01
1.19954312e+00 2.79000103e-01 2.21918434e-01 1.27721262e+00
9.63035524e-02 9.19271111e-01 1.19760823e+00 -2.15724707e-01
-3.02659534e-02 1.51737645e-01 3.38599473e-01 6.09401286e-01
-6.28268778e-01 -2.94234425e-01 -5.92631638e-01 -3.32087427e-01
5.53480327e-01 -4.45249915e-01 -2.03658372e-01 9.28355455e-02
-1.37853765e+00 1.04218733e+00 3.59130591e-01 6.24629736e-01
-4.05462921e-01 -3.22186083e-01 7.15915740e-01 5.03543556e-01
3.14489305e-01 6.79683864e-01 -6.75815403e-01 -4.24715400e-01
-2.74934649e-01 5.45897901e-01 1.34519541e+00 1.16034198e+00
5.80478549e-01 -2.28317410e-01 -3.44420940e-01 1.31134176e+00
-7.83373490e-02 1.92007348e-01 4.61869121e-01 -9.15146530e-01
5.44608057e-01 8.69179964e-01 3.08752090e-01 -1.21018136e+00
-9.73715603e-01 3.77943069e-02 -9.15777981e-01 -3.29960346e-01
3.55873913e-01 -6.76962674e-01 -1.18808910e-01 2.07034779e+00
6.98772430e-01 -1.01987749e-01 5.03315151e-01 1.10258031e+00
1.55585551e+00 7.57034898e-01 1.76649809e-01 -3.53869408e-01
1.92422366e+00 -1.13405144e+00 -1.37334192e+00 -1.13805130e-01
8.42139840e-01 -9.13384557e-01 1.13247824e+00 3.44913244e-01
-7.15450466e-01 -2.91655093e-01 -8.49339962e-01 -2.49958575e-01
-7.21638918e-01 1.81004286e-01 1.02932763e+00 3.72715265e-01
-8.12753618e-01 2.67064333e-01 1.50575653e-01 -4.95249778e-01
-4.21447039e-01 4.61920828e-01 -5.47317386e-01 2.80206203e-01
-1.93445957e+00 1.18320572e+00 5.13708293e-01 9.41584110e-02
-2.19785422e-01 -5.07311583e-01 -1.00194108e+00 -3.84485945e-02
7.26178706e-01 -7.47317970e-01 1.37822390e+00 -7.13905871e-01
-1.73628128e+00 6.16213620e-01 -7.22533688e-02 -2.71415919e-01
3.95987257e-02 -2.15004027e-01 -5.28604746e-01 -5.12285195e-02
-4.98589762e-02 8.40012550e-01 4.13388550e-01 -1.25741076e+00
-5.65771103e-01 -3.09762776e-01 4.14915562e-01 7.72672057e-01
-3.06634724e-01 3.70698065e-01 -1.39054567e-01 -4.95320022e-01
-2.85952538e-01 -9.67803061e-01 -1.78781554e-01 -6.13087714e-01
-4.60183591e-01 -8.51946294e-01 1.05003226e+00 -6.99961066e-01
1.24285197e+00 -1.93527782e+00 3.34841341e-01 -3.86499882e-01
3.81553859e-01 -1.26081839e-01 -1.13920629e-01 9.93936598e-01
2.37728357e-01 -1.92111313e-01 -2.05274537e-01 -3.19671512e-01
5.05508184e-01 4.71111715e-01 -2.67936200e-01 -2.22033396e-01
3.46906304e-01 9.53618407e-01 -1.00977826e+00 -7.91371942e-01
2.55640626e-01 3.72391455e-02 -4.13620055e-01 7.31412470e-01
-3.39769214e-01 3.06831121e-01 -5.76155841e-01 2.89489418e-01
6.14923477e-01 -8.25420767e-03 2.91634709e-01 -2.88238406e-01
5.58797605e-02 3.25664282e-01 -9.19727623e-01 1.76148725e+00
-7.76747346e-01 3.13436657e-01 4.51518714e-01 -6.61550701e-01
1.21813762e+00 6.88720942e-01 5.29721618e-01 -4.23961341e-01
3.09680790e-01 -6.80493936e-02 9.07354504e-02 -7.33227968e-01
1.14168811e+00 -3.65111887e-01 -9.26399708e-01 3.97161663e-01
3.26399237e-01 -3.81827444e-01 9.11006182e-02 5.91430783e-01
9.47443724e-01 -4.62636910e-02 5.07271707e-01 -1.96456000e-01
6.38400197e-01 2.02132761e-01 4.18765932e-01 3.69748235e-01
-4.06582832e-01 2.77571112e-01 8.61541748e-01 -1.16259076e-01
-5.70101738e-01 -5.34415543e-01 1.83715761e-01 1.33316028e+00
5.25734425e-01 -5.18344760e-01 -7.72512317e-01 -9.39101875e-01
-3.29776436e-01 7.34898746e-01 -4.33411032e-01 3.57607305e-02
-5.34367323e-01 -9.94452298e-01 7.49470234e-01 1.73718721e-01
8.57284665e-01 -1.49909842e+00 -2.67680675e-01 4.53802168e-01
-8.10582995e-01 -1.46551931e+00 -1.78651527e-01 1.82627484e-01
-1.79285437e-01 -9.63888347e-01 -5.32778144e-01 -9.10041034e-01
-1.18601926e-01 -1.52588924e-02 1.30101991e+00 -1.65478734e-04
-8.94032717e-02 2.97476768e-01 -9.34502542e-01 -3.59633088e-01
-5.06128311e-01 2.02440634e-01 -2.54837781e-01 -1.86434239e-01
3.91915292e-01 -5.63202679e-01 -2.66755849e-01 3.60715568e-01
-5.44496119e-01 2.68518806e-01 3.95624399e-01 1.08928490e+00
-2.55261064e-01 -8.21830109e-02 1.26382947e+00 -9.94963646e-01
1.25293303e+00 -7.06312954e-01 2.62679726e-01 2.78684318e-01
-1.61799714e-01 -6.89281477e-03 8.69106174e-01 -1.84854940e-01
-1.72687030e+00 -7.16209412e-02 -7.03497887e-01 9.71229002e-02
-4.61576849e-01 3.26391667e-01 -5.57801485e-01 2.88637310e-01
3.23569238e-01 5.55675104e-02 -1.16783015e-01 -3.60619664e-01
7.01637030e-01 1.02973747e+00 6.90130115e-01 -7.95725465e-01
2.89024442e-01 -6.35340735e-02 -1.36849105e-01 -7.25636184e-01
-8.21510971e-01 -7.64223456e-01 -6.13189459e-01 -3.58325660e-01
1.12345433e+00 -7.30076492e-01 -1.08749378e+00 4.10315454e-01
-1.62599862e+00 -2.57123172e-01 2.62727916e-01 1.70383438e-01
-5.44753313e-01 4.66721416e-01 -1.00182188e+00 -9.12027478e-01
-4.13966119e-01 -9.02380884e-01 1.41245615e+00 2.25958645e-01
-5.57044744e-01 -1.09082341e+00 -6.35688333e-03 4.64041442e-01
1.69948921e-01 3.54869425e-01 1.11665714e+00 -1.05501497e+00
1.49233025e-02 1.29214182e-01 -3.15921992e-01 -1.36491299e-01
1.01171136e-01 -4.53670651e-01 -9.46646154e-01 2.95139462e-01
1.54369310e-01 -7.69697726e-01 1.99562430e-01 -1.85086623e-01
5.15032649e-01 -4.37347919e-01 -5.17107472e-02 -2.17971191e-01
9.42879856e-01 3.33286017e-01 5.37958741e-01 -4.74178325e-03
6.70961797e-01 1.22864830e+00 1.06288886e+00 6.96896970e-01
8.76113713e-01 1.04688394e+00 3.96725953e-01 -3.13912690e-01
-5.83891943e-02 -1.27753532e-02 1.98954806e-01 1.16288960e+00
8.46519843e-02 -4.69694883e-01 -5.15668392e-01 4.30843323e-01
-2.07919693e+00 -9.06517804e-01 -5.47102153e-01 1.28607070e+00
1.04218829e+00 -4.58602786e-01 3.53006035e-01 -2.43410487e-02
8.62236917e-01 3.53074044e-01 -1.30030587e-01 -8.15508783e-01
-2.83237785e-01 -1.77283123e-01 -2.80334890e-01 4.58699405e-01
-1.08515477e+00 1.18908775e+00 4.95806932e+00 1.11466086e+00
-6.24010444e-01 2.83318043e-01 4.19013709e-01 6.44792378e-01
-9.23633203e-02 -1.97231635e-01 -6.35683656e-01 2.08122566e-01
5.96359372e-01 -5.15816100e-02 3.56746972e-01 8.62823367e-01
-3.77083160e-02 -3.25920619e-02 -9.28301752e-01 1.11883855e+00
1.19863739e-02 -8.76634121e-01 -2.02056691e-01 -2.02449799e-01
3.75032723e-01 -6.36716843e-01 -4.78784889e-01 8.39793384e-01
4.34177876e-01 -7.87742794e-01 3.44035953e-01 8.44647363e-02
3.47286016e-01 -6.63312852e-01 1.24825466e+00 4.47732508e-01
-1.19689524e+00 1.18024953e-01 -3.07230473e-01 -2.68984795e-01
4.63077009e-01 3.16101074e-01 -9.28179145e-01 1.12387848e+00
3.72353017e-01 5.54867029e-01 -2.42441937e-01 4.14779663e-01
-2.32260138e-01 2.66454846e-01 -4.13043797e-02 -6.16017640e-01
3.58251542e-01 -1.09036870e-01 5.72402239e-01 1.58268058e+00
8.53440166e-02 6.30659521e-01 3.34392935e-01 6.16660655e-01
-6.84233522e-03 4.41715568e-01 -5.61569035e-01 2.66037703e-01
5.82375765e-01 1.80288315e+00 -7.31394708e-01 -4.30471152e-01
-2.39690542e-01 1.19672072e+00 3.29949617e-01 -5.22010960e-02
-8.55551839e-01 -6.49499118e-01 4.73922670e-01 -8.06294203e-01
-1.39369816e-01 -1.04885604e-02 -1.09174818e-01 -1.09129131e+00
-1.63598344e-01 -8.78257096e-01 5.76745450e-01 -8.38111341e-01
-1.76656711e+00 9.82095838e-01 1.85793489e-01 -1.00203943e+00
-7.39036620e-01 -7.54137814e-01 -6.98395312e-01 6.30352795e-01
-1.18755949e+00 -1.36096060e+00 -4.43948954e-01 8.36442292e-01
8.43196094e-01 1.41979039e-01 1.15029085e+00 2.07357123e-01
-5.53311288e-01 6.41124129e-01 -5.62903166e-01 1.10176429e-01
1.01408565e+00 -1.52305114e+00 -6.68758824e-02 1.68986052e-01
-8.60477462e-02 4.40134645e-01 9.16122735e-01 -6.64619207e-01
-1.32066238e+00 -7.65423477e-01 1.17171705e+00 -5.20323634e-01
8.24148774e-01 -6.20336890e-01 -8.51054847e-01 2.26468578e-01
8.33128870e-01 -6.78394854e-01 5.13271213e-01 5.31443477e-01
-2.11652964e-01 4.07953739e-01 -1.42527199e+00 6.67824209e-01
1.19788611e+00 -3.78210396e-01 -1.00883639e+00 5.04934072e-01
9.77570236e-01 -5.08099556e-01 -1.15520751e+00 4.97501075e-01
3.58689547e-01 -1.00125754e+00 7.64651835e-01 -6.00175619e-01
6.70509875e-01 6.51048794e-02 -6.96320832e-02 -1.37436950e+00
-3.77955921e-02 -8.20127249e-01 6.75127208e-02 1.90054750e+00
4.03270632e-01 -4.39634353e-01 5.08496404e-01 6.51447356e-01
-4.73328680e-01 -7.49876678e-01 -7.58680582e-01 -5.26407242e-01
-5.33770286e-02 -2.75758326e-01 8.89818311e-01 1.23180342e+00
6.55425012e-01 1.08474886e+00 -7.53235042e-01 -1.55646041e-01
4.45974059e-02 5.62990129e-01 8.21260631e-01 -1.00125861e+00
-3.50267231e-01 -4.33909148e-01 -6.65358678e-02 -1.02901447e+00
5.46732843e-01 -6.11992240e-01 4.30508494e-01 -1.56866980e+00
9.17887613e-02 -4.26463932e-01 9.78605524e-02 3.80472720e-01
-5.90700090e-01 4.73178364e-02 3.36804122e-01 4.82017454e-03
-8.25445175e-01 9.12175894e-01 1.24225950e+00 1.41657442e-01
-1.42339304e-01 -2.49346733e-01 -4.99643147e-01 7.55303621e-01
7.64244318e-01 -2.33266622e-01 -4.78800863e-01 1.89560071e-01
7.70302191e-02 5.81995487e-01 4.00481462e-01 -4.33183372e-01
1.82402685e-01 -2.09071383e-01 -1.51254669e-01 -4.27673459e-01
7.40783572e-01 -6.27149880e-01 -2.04500139e-01 -4.28697653e-02
-3.51209283e-01 2.15828925e-01 1.75388902e-01 3.59580129e-01
-5.16301036e-01 -4.15410131e-01 2.10766494e-01 -1.95584595e-01
-1.02584183e+00 -1.99134484e-01 -3.58867079e-01 1.39008194e-01
7.51354456e-01 6.82484582e-02 -5.30950129e-01 -7.96308517e-01
-5.25592387e-01 3.87654781e-01 3.15187834e-02 6.24192476e-01
2.68589348e-01 -1.33236599e+00 -6.19891465e-01 -4.58419800e-01
1.37585282e-01 -2.87514538e-01 6.31916165e-01 7.77491152e-01
1.62502363e-01 3.18884552e-01 -2.76847929e-01 -1.54730454e-01
-1.41583455e+00 5.68770111e-01 -3.25015821e-02 -8.39103520e-01
-3.36157888e-01 7.34617591e-01 4.82434928e-01 -9.63939369e-01
1.78455949e-01 4.00891714e-02 -7.39915967e-01 2.97579318e-01
1.68653101e-01 1.37491822e-01 -6.58666566e-02 -7.72761583e-01
-3.02684397e-01 1.97454616e-01 9.62765291e-02 -1.79587230e-01
1.02038836e+00 -3.17402720e-01 -4.20359313e-01 4.47693080e-01
1.04149449e+00 -3.45187038e-02 -7.38379538e-01 -9.97452065e-03
2.83984125e-01 7.49857724e-02 -5.77136338e-01 -1.13652217e+00
-5.72757781e-01 8.25992823e-01 -1.91286393e-02 5.49967706e-01
1.19293702e+00 2.52196878e-01 1.23800516e+00 4.64354694e-01
5.06066382e-01 -1.03934407e+00 1.54525816e-01 8.71233881e-01
1.29153240e+00 -1.20454621e+00 -4.29559141e-01 -9.68597651e-01
-1.28901136e+00 1.07886553e+00 9.59286213e-01 1.45296067e-01
2.39030302e-01 4.63988036e-01 2.01593563e-01 -5.49107790e-01
-9.07600582e-01 -3.01184744e-01 1.40974626e-01 4.64565784e-01
6.39721751e-01 1.83910906e-01 -6.52011931e-01 1.29855406e+00
-7.13173985e-01 -4.27803814e-01 4.24570769e-01 7.26928115e-01
-2.51596659e-01 -1.31144357e+00 -2.53791034e-01 3.36841792e-01
-3.68860751e-01 -1.55310854e-01 -1.13160443e+00 1.09847939e+00
1.95600405e-01 1.40735197e+00 -2.63860017e-01 -7.09357083e-01
3.96711111e-01 3.32517236e-01 2.20664486e-01 -4.81304824e-01
-1.26520395e+00 -1.27408311e-01 9.35251713e-01 -3.50926876e-01
-7.86592543e-01 -2.83400297e-01 -1.41894984e+00 -2.85583943e-01
-4.42158729e-01 5.29549599e-01 4.68902677e-01 1.19044757e+00
2.32453436e-01 5.13808370e-01 8.05185199e-01 -6.63857460e-01
-3.97998899e-01 -1.32929170e+00 -5.87964416e-01 7.44227767e-01
-1.49038777e-01 -8.35248947e-01 -2.01249763e-01 -1.12631172e-01] | [12.889557838439941, 6.580177307128906] |
6d9a2994-f410-4aa6-ac2a-b826e5858df9 | random-projection-tree-similarity-metric-for | 2302.13168 | null | https://arxiv.org/abs/2302.13168v1 | https://arxiv.org/pdf/2302.13168v1.pdf | Random projection tree similarity metric for SpectralNet | SpectralNet is a graph clustering method that uses neural network to find an embedding that separates the data. So far it was only used with $k$-nn graphs, which are usually constructed using a distance metric (e.g., Euclidean distance). $k$-nn graphs restrict the points to have a fixed number of neighbors regardless of the local statistics around them. We proposed a new SpectralNet similarity metric based on random projection trees (rpTrees). Our experiments revealed that SpectralNet produces better clustering accuracy using rpTree similarity metric compared to $k$-nn graph with a distance metric. Also, we found out that rpTree parameters do not affect the clustering accuracy. These parameters include the leaf size and the selection of projection direction. It is computationally efficient to keep the leaf size in order of $\log(n)$, and project the points onto a random direction instead of trying to find the direction with the maximum dispersion. | ['Masahiro Takatsuka', 'Adel F. Ahmed', 'John Stavrakakis', 'Mashaan Alshammari'] | 2023-02-25 | null | null | null | null | ['graph-clustering', 'spectral-graph-clustering', 'graph-partitioning'] | ['graphs', 'graphs', 'graphs'] | [-2.70542920e-01 -4.79023084e-02 -8.98588225e-02 -2.78231561e-01
1.41887972e-02 -5.52989841e-01 9.16527361e-02 1.71155468e-01
-5.77796817e-01 2.19213337e-01 -5.53133897e-03 -3.08068782e-01
-7.68662512e-01 -1.20626509e+00 -1.94554999e-01 -7.32816219e-01
-5.04461884e-01 3.70841771e-01 2.78033167e-01 6.52125999e-02
4.83484983e-01 6.31445527e-01 -1.36209941e+00 -1.74134523e-01
6.98052943e-01 8.22455227e-01 2.24417925e-01 4.30675000e-01
-2.65673161e-01 2.52421379e-01 -5.36733806e-01 -8.71713459e-02
5.94897866e-01 -5.01322448e-01 -8.20393920e-01 -2.56398320e-01
1.94923088e-01 4.11981493e-01 -5.53371906e-01 1.28927481e+00
4.22383010e-01 3.19445938e-01 7.95796931e-01 -1.34649503e+00
-6.21827781e-01 7.00278699e-01 -6.54574156e-01 2.61077672e-01
1.13178134e-01 -1.95945367e-01 1.11201000e+00 -6.38731718e-01
7.00330198e-01 9.54997778e-01 1.00226116e+00 1.52999654e-01
-1.32926571e+00 -8.25977683e-01 -2.44781375e-01 3.18850458e-01
-2.01803660e+00 5.05371690e-02 1.24587584e+00 -2.07673579e-01
8.61670852e-01 2.14907229e-01 7.71290541e-01 3.24832469e-01
8.21945220e-02 1.29315183e-02 9.07416642e-01 -5.83765924e-01
4.63756412e-01 -1.42062396e-01 2.09684834e-01 7.51587689e-01
3.60030413e-01 -2.62257457e-01 -1.58696860e-01 -3.26457530e-01
6.50343955e-01 7.02695698e-02 -1.09785452e-01 -8.98020744e-01
-1.14655292e+00 1.08353817e+00 8.16931605e-01 6.49339259e-01
-6.10315651e-02 3.05156201e-01 2.20429569e-01 2.95047134e-01
-9.20998380e-02 6.80499315e-01 -3.36660087e-01 1.11073300e-01
-9.32099998e-01 -3.29748005e-01 6.83015823e-01 7.08779454e-01
1.11651909e+00 -1.35792002e-01 5.14405727e-01 1.02362800e+00
1.64901108e-01 3.00852537e-01 4.49240953e-01 -1.18256402e+00
2.45210811e-01 1.00606668e+00 -5.16328633e-01 -1.84845805e+00
-6.14574254e-01 -1.31988868e-01 -1.20100069e+00 1.32509068e-01
4.13759142e-01 -1.41434073e-01 -6.54154956e-01 1.67744470e+00
1.62118509e-01 1.19217999e-01 -9.25223082e-02 7.00780213e-01
6.72850192e-01 5.71366847e-01 -2.40117326e-01 -1.57448098e-01
8.27296853e-01 -5.49558997e-01 -3.10291946e-01 2.05984682e-01
9.67636526e-01 -5.92452228e-01 1.14377034e+00 1.48312464e-01
-7.79698670e-01 -3.87164116e-01 -1.17617893e+00 4.65013176e-01
-7.29837596e-01 -6.50270656e-02 5.64484179e-01 7.33483374e-01
-1.60692835e+00 9.94671941e-01 -6.62447691e-01 -6.53934836e-01
-7.30819628e-02 7.40058541e-01 -3.99136156e-01 2.56937593e-01
-1.13699889e+00 5.45041978e-01 7.20457733e-01 -2.07233921e-01
4.55442108e-02 -1.81385130e-02 -7.26305842e-01 1.57567650e-01
1.69127747e-01 -2.60784656e-01 2.17451066e-01 -8.04739118e-01
-1.23712337e+00 6.33054912e-01 6.14278018e-02 -3.94110918e-01
-2.05137119e-01 5.82814336e-01 -3.79464626e-01 3.87158096e-01
8.45037699e-02 7.91725278e-01 4.96604770e-01 -1.04733098e+00
-2.06399828e-01 -3.22046697e-01 -1.96022347e-01 2.94034183e-01
-6.40239358e-01 -7.93250054e-02 -3.16672146e-01 -3.63868505e-01
8.87479663e-01 -1.04150891e+00 -4.35818464e-01 -2.76835799e-01
-4.80232924e-01 -4.60859150e-01 1.00214577e+00 -2.08686709e-01
1.38068414e+00 -2.31648278e+00 -3.40051427e-02 1.01385343e+00
4.57053035e-01 2.01435819e-01 -7.20907971e-02 6.70261919e-01
-4.87167031e-01 3.60386312e-01 -2.42197499e-01 3.97207469e-01
-1.55215606e-01 8.77792202e-03 2.82253474e-01 5.65755725e-01
-4.56451386e-01 4.45152014e-01 -7.49152780e-01 -7.59227514e-01
2.34971985e-01 2.46253982e-01 -5.28561294e-01 -2.29376644e-01
3.25752765e-01 -4.35331762e-01 -1.11582130e-01 2.42201611e-01
7.63254642e-01 -4.80125457e-01 4.21320856e-01 -3.78624439e-01
2.69308519e-02 7.19327107e-02 -1.59217775e+00 1.42418575e+00
-6.39191046e-02 6.20243192e-01 -1.94818690e-01 -1.20285046e+00
1.40955317e+00 4.47536930e-02 9.28132057e-01 -3.38292211e-01
5.65319955e-02 1.40940890e-01 3.17141920e-01 -2.16952153e-02
4.57288027e-01 1.71631813e-01 1.88495312e-02 7.59174943e-01
-2.67334253e-01 -1.47895962e-02 3.02065372e-01 3.32703114e-01
1.35644436e+00 -5.90814948e-01 4.37287569e-01 -5.63660681e-01
5.27919412e-01 8.57791975e-02 5.71566939e-01 4.80022758e-01
-2.56349832e-01 5.95555782e-01 5.20103276e-01 -5.29090762e-01
-9.15099800e-01 -1.25142109e+00 -4.78590988e-02 7.96007216e-01
4.16458458e-01 -7.16493905e-01 -8.18369627e-01 -7.60495424e-01
-1.26070857e-01 4.26860124e-01 -5.29035211e-01 -2.81103075e-01
-6.91051841e-01 -7.04062343e-01 5.37735581e-01 2.51554251e-01
5.78491449e-01 -1.20440280e+00 -4.36473668e-01 -4.19976227e-02
1.63623784e-02 -5.12802005e-01 -7.13487208e-01 5.29286265e-01
-1.07789433e+00 -1.24831080e+00 -2.83919424e-01 -1.25572431e+00
1.03696835e+00 5.70434093e-01 8.64258111e-01 8.30056742e-02
-1.55772060e-01 2.07626879e-01 -5.31947136e-01 2.92445868e-01
-8.36917385e-02 2.84167081e-01 1.85435712e-01 -3.70475411e-01
8.18476915e-01 -9.78166342e-01 -7.18635619e-01 5.93334079e-01
-6.33091092e-01 -4.33336616e-01 3.91374826e-01 5.87142169e-01
5.63498855e-01 8.12889755e-01 2.49212041e-01 -4.93848175e-01
9.69359100e-01 -2.61124223e-01 -4.07683760e-01 9.72006619e-02
-7.90323675e-01 1.54521033e-01 9.01032686e-01 -3.84462416e-01
-3.25951055e-02 2.11465359e-01 2.93892354e-01 -5.56367457e-01
-7.91975409e-02 4.40593809e-01 1.20308518e-01 -2.94976473e-01
8.61257970e-01 2.63607234e-01 4.10793871e-02 -1.74915969e-01
4.11290467e-01 7.12738812e-01 1.21960402e-01 -1.92300707e-01
7.93851495e-01 3.83945882e-01 1.68449298e-01 -8.86258960e-01
-3.91945019e-02 -5.59266806e-01 -1.03633595e+00 3.24143358e-02
7.97295392e-01 -2.52932072e-01 -8.42796683e-01 -1.22977262e-02
-8.40741336e-01 -3.39413621e-02 -2.78393239e-01 5.35419524e-01
-4.34194982e-01 6.27413154e-01 -4.07343954e-01 -4.81994003e-01
-4.14633155e-01 -8.63909125e-01 3.03519636e-01 -4.49122377e-02
-4.75499123e-01 -9.16330278e-01 3.14404935e-01 -2.51543790e-01
1.73989087e-01 7.64634013e-02 1.18448675e+00 -6.82886779e-01
-2.57272393e-01 -7.09217936e-02 -4.16966617e-01 4.79760394e-02
3.75901133e-01 1.82878524e-01 -3.44434977e-01 -3.91446292e-01
-1.66281223e-01 3.26096088e-01 6.34930134e-01 6.00017667e-01
1.44130886e+00 -3.18228722e-01 -6.52849555e-01 8.08936477e-01
1.60209668e+00 7.90197194e-01 5.37791371e-01 4.04699534e-01
7.52426922e-01 4.59712356e-01 9.05279666e-02 -1.59840751e-03
1.52715489e-01 4.25448477e-01 3.20008993e-01 -1.67554587e-01
6.66307807e-02 -2.33281583e-01 -1.95947401e-02 1.17894471e+00
1.51631564e-01 -1.56155139e-01 -1.16386855e+00 5.81196666e-01
-1.59204447e+00 -9.97216523e-01 1.78375188e-02 2.34994125e+00
4.86715615e-01 1.52944520e-01 2.80939132e-01 4.83630359e-01
1.13458383e+00 2.33064279e-01 -3.89647841e-01 -6.88460886e-01
-9.79662389e-02 2.10389286e-01 7.89387465e-01 4.02046651e-01
-9.40299571e-01 8.56571734e-01 6.54134846e+00 8.30930710e-01
-1.13522911e+00 -3.29996049e-01 2.84436554e-01 -8.46306235e-02
-1.83372423e-01 7.67960325e-02 -3.97128165e-01 6.57656848e-01
8.33504140e-01 -2.67686937e-02 5.41940033e-01 8.52682292e-01
1.82562560e-01 7.12531060e-02 -8.48509490e-01 1.32986081e+00
-1.10922799e-01 -1.27576756e+00 -6.88281059e-02 2.63868451e-01
5.21357775e-01 1.47412509e-01 -1.13335192e-01 -2.68304348e-01
6.58803463e-01 -1.20431459e+00 6.44671023e-02 5.68790769e-04
6.45148218e-01 -1.15933239e+00 6.25041068e-01 3.36017579e-01
-1.53348958e+00 -2.37878188e-02 -7.29197919e-01 7.70703629e-02
-1.65891424e-01 6.62491322e-01 -9.78622437e-01 1.86235756e-01
1.04752719e+00 6.89114213e-01 -5.24892509e-01 9.61558700e-01
1.16368003e-01 3.35597098e-01 -6.57759488e-01 -3.36094290e-01
4.45113420e-01 -8.65571439e-01 3.79994512e-01 1.08461368e+00
6.71208978e-01 -2.53694393e-02 1.34434402e-01 6.91917479e-01
8.69454667e-02 3.74406278e-01 -1.00252748e+00 -5.55729941e-02
9.51464951e-01 1.03222680e+00 -1.36851573e+00 3.80048156e-02
-1.46990746e-01 6.98569119e-01 4.14535254e-01 1.78283975e-01
-4.45820808e-01 -1.12858045e+00 4.86747414e-01 3.68366182e-01
4.52168047e-01 -6.40479505e-01 -3.73284072e-01 -4.67253804e-01
-1.48016572e-01 -4.95700896e-01 6.17337406e-01 -6.55640662e-01
-1.32907796e+00 7.47496426e-01 8.28119665e-02 -1.24906647e+00
-9.64832008e-02 -5.06515026e-01 -6.16321206e-01 6.67520940e-01
-4.24764842e-01 -3.54822636e-01 -1.18277721e-01 7.35418081e-01
-1.52243778e-01 -1.51080161e-01 7.51784325e-01 5.84970638e-02
-4.19836164e-01 5.99149883e-01 4.95171547e-01 4.76318508e-01
4.78717834e-01 -1.24459493e+00 2.19817802e-01 5.60360432e-01
3.38456929e-01 9.51332629e-01 5.78735054e-01 -4.82249767e-01
-9.33453977e-01 -7.99051523e-01 8.99676681e-01 -6.93780854e-02
5.91499746e-01 -2.37664729e-01 -8.37976336e-01 3.85065109e-01
2.08511591e-01 -1.07256934e-01 8.31714988e-01 2.52112418e-01
-2.80711055e-01 -3.98655623e-01 -1.38469791e+00 8.08560371e-01
1.30998564e+00 -5.53481996e-01 -3.21736306e-01 1.37320369e-01
6.90484166e-01 2.08108068e-01 -9.20572877e-01 3.41724843e-01
3.15471739e-01 -1.18297768e+00 1.07182717e+00 -3.38528231e-02
-1.44762872e-02 -6.44304514e-01 -2.42185161e-01 -1.39086413e+00
-7.67263472e-01 -3.86265695e-01 3.77475560e-01 7.99604356e-01
4.89391178e-01 -9.21877742e-01 1.31504905e+00 6.31707683e-02
3.09152246e-01 -8.46796393e-01 -9.46067631e-01 -9.39661980e-01
5.67673221e-02 -1.98581770e-01 7.20130742e-01 1.24017513e+00
3.76242280e-01 2.79440522e-01 3.77587974e-02 8.91099647e-02
6.37473702e-01 1.23550363e-01 5.11651695e-01 -1.54014134e+00
8.20632502e-02 -5.49774051e-01 -7.84065068e-01 -7.17690587e-01
1.06852546e-01 -1.11978924e+00 -2.02111825e-01 -1.48593271e+00
3.76404747e-02 -9.21371162e-01 -4.29355770e-01 2.92519420e-01
2.72725642e-01 2.73952872e-01 1.38729513e-01 3.99510592e-01
-4.76607531e-01 2.55636066e-01 9.26804602e-01 4.67807241e-02
-7.46118248e-01 -2.30327860e-01 -5.88840723e-01 8.97636354e-01
1.16944146e+00 -6.27140760e-01 -6.80111110e-01 -2.16080904e-01
2.01550692e-01 -6.31518662e-02 -1.77849531e-01 -1.27058351e+00
4.79601949e-01 -2.59923398e-01 3.80661398e-01 -9.62062836e-01
2.68121779e-01 -1.00161850e+00 4.11842704e-01 4.54133213e-01
-2.45657563e-01 5.50817907e-01 -2.43438646e-01 4.80507016e-01
-1.11435428e-01 -3.10235202e-01 9.14158106e-01 -1.61104947e-01
-4.55348551e-01 2.73113877e-01 -2.39566579e-01 -1.89402565e-01
1.20390594e+00 -9.13687289e-01 -3.46909575e-02 -3.65208209e-01
-7.14340746e-01 -1.14712790e-02 6.78978145e-01 -2.60911286e-02
6.14967823e-01 -1.48739624e+00 -1.13723792e-01 1.90858617e-01
-7.81889632e-02 -3.40020619e-02 -3.39228332e-01 5.15375972e-01
-8.76970053e-01 4.76356000e-01 -2.14493945e-01 -6.63688123e-01
-1.54648924e+00 7.43621826e-01 3.19166899e-01 -8.76465961e-02
-6.87723637e-01 6.68625414e-01 -1.50303155e-01 -8.31506968e-01
1.92375615e-01 -1.34857982e-01 -2.84949958e-01 -7.07382336e-02
1.11304052e-01 6.79738283e-01 -1.38074428e-01 -6.14405394e-01
-5.80782890e-01 1.07567739e+00 2.05563635e-01 -7.68116340e-02
1.16031063e+00 -8.72849748e-02 -4.34844494e-01 2.13197485e-01
1.53021812e+00 -9.48775187e-02 -6.97979867e-01 -1.81367442e-01
1.68606192e-02 -4.69474196e-01 -1.26413122e-01 -2.73008108e-01
-1.22136736e+00 6.91773653e-01 7.69832313e-01 7.47760892e-01
1.13745534e+00 3.28037329e-02 6.76162720e-01 6.60941005e-01
4.96003121e-01 -1.30161667e+00 4.91806678e-02 5.56908488e-01
4.11611617e-01 -7.85588086e-01 1.30763188e-01 -4.69092846e-01
-2.21762359e-01 1.08480608e+00 6.65940046e-01 -4.57506686e-01
1.06029701e+00 2.61555947e-02 4.52323891e-02 -4.61826742e-01
-2.39809573e-01 -7.15435222e-02 1.01717398e-01 8.00573468e-01
3.66317987e-01 1.92781642e-01 -4.54506278e-01 -1.75482601e-01
-7.55684078e-01 -3.93407524e-01 3.24168503e-01 6.03765249e-01
-7.35292912e-01 -1.05348337e+00 -3.57079387e-01 6.64341927e-01
-5.79126878e-03 -9.51056033e-02 -6.38496041e-01 8.19124460e-01
3.25963974e-01 1.01161253e+00 4.16692495e-01 -8.42417657e-01
1.17701538e-01 2.19836924e-02 3.07754856e-02 -4.48295832e-01
-3.25916469e-01 -3.04180592e-01 -3.07720095e-01 -5.78877807e-01
-3.62120748e-01 -4.40331608e-01 -1.49851692e+00 -7.78471053e-01
-3.20499241e-01 6.13150001e-01 6.25127077e-01 5.47124624e-01
4.81525630e-01 -6.06598482e-02 8.89735341e-01 -4.54966635e-01
-1.81610480e-01 -7.10770428e-01 -9.63779688e-01 2.66276270e-01
-2.44538665e-01 -4.79876935e-01 -7.70031095e-01 -2.90822804e-01] | [7.397940635681152, 4.857438087463379] |
ba863939-3d9f-4620-a0a5-1525b62849bb | mousai-text-to-music-generation-with-long | 2301.11757 | null | https://arxiv.org/abs/2301.11757v2 | https://arxiv.org/pdf/2301.11757v2.pdf | Moûsai: Text-to-Music Generation with Long-Context Latent Diffusion | The recent surge in popularity of diffusion models for image generation has brought new attention to the potential of these models in other areas of media synthesis. One area that has yet to be fully explored is the application of diffusion models to music generation. Music generation requires to handle multiple aspects, including the temporal dimension, long-term structure, multiple layers of overlapping sounds, and nuances that only trained listeners can detect. In our work, we investigate the potential of diffusion models for text-conditional music generation. We develop a cascading latent diffusion approach that can generate multiple minutes of high-quality stereo music at 48kHz from textual descriptions. For each model, we make an effort to maintain reasonable inference speed, targeting real-time on a single consumer GPU. In addition to trained models, we provide a collection of open-source libraries with the hope of facilitating future work in the field. We open-source the following: Music samples for this paper: https://bit.ly/anonymous-mousai; all music samples for all models: https://bit.ly/audio-diffusion; and codes: https://github.com/archinetai/audio-diffusion-pytorch | ['Bernhard Schölkopf', 'Zhijing Jin', 'Flavio Schneider'] | 2023-01-27 | null | null | null | null | ['text-to-music-generation', 'music-generation', 'music-generation', 'text-to-music-generation'] | ['audio', 'audio', 'music', 'music'] | [-7.49866590e-02 -1.92556635e-01 5.89621663e-02 8.28738138e-03
-1.22198558e+00 -6.63891256e-01 7.22758114e-01 -1.22611612e-01
-7.27941934e-03 4.65052545e-01 7.60944366e-01 -1.74545199e-01
4.51230183e-02 -7.34559774e-01 -4.97967511e-01 -4.99068677e-01
-2.33273014e-01 2.77037710e-01 1.28578991e-01 -7.94873685e-02
2.76170731e-01 -5.77628240e-02 -1.38472676e+00 6.82074547e-01
6.05241418e-01 6.97028160e-01 3.68655294e-01 1.28391457e+00
2.25190565e-01 9.94039297e-01 -5.07341683e-01 -3.68950665e-01
2.09147185e-01 -7.94210851e-01 -4.83109325e-01 -3.18195112e-02
1.11443624e-01 -4.30595309e-01 -3.64038885e-01 1.04070866e+00
9.72674429e-01 7.02969870e-03 6.91042244e-01 -9.83789027e-01
-7.50521421e-01 8.70684266e-01 -5.09254336e-01 3.32960069e-01
3.66461277e-01 5.09665847e-01 1.03810668e+00 -9.51926470e-01
5.36582530e-01 1.31961691e+00 5.95646083e-01 5.53645730e-01
-1.24993455e+00 -1.01306605e+00 -3.57489079e-01 1.00287929e-01
-1.40643632e+00 -9.36629474e-01 7.16483057e-01 -5.18471599e-01
7.10968375e-01 3.28026205e-01 6.27522111e-01 1.53962398e+00
7.43701309e-02 9.55015838e-01 1.08191609e+00 -3.70136678e-01
1.71663016e-01 -1.37662262e-01 -4.02743518e-01 4.04576153e-01
-1.96178749e-01 1.49971187e-01 -9.47796941e-01 -4.21512753e-01
1.06922722e+00 -5.06506026e-01 -2.16689914e-01 1.92684770e-01
-1.49813855e+00 8.95634234e-01 1.46986768e-02 1.85630679e-01
-3.35466146e-01 6.53876662e-01 2.63337195e-01 2.81169146e-01
7.48531103e-01 2.24979520e-01 -1.76894829e-01 -7.01348841e-01
-1.19677234e+00 6.25315547e-01 8.22639763e-01 9.16468322e-01
2.55717576e-01 3.03182721e-01 -1.82595477e-01 1.00493097e+00
3.95719558e-01 5.69248080e-01 4.15298194e-01 -1.36210752e+00
3.46955985e-01 -3.01247507e-01 2.90506911e-02 -7.72232354e-01
4.72367480e-02 -3.15591276e-01 -8.38785946e-01 6.63585737e-02
3.30555260e-01 -5.04925907e-01 -4.57022220e-01 1.64625776e+00
2.72417128e-01 5.88745832e-01 -5.17640002e-02 9.63192344e-01
5.57633758e-01 1.04446375e+00 -2.61019737e-01 -1.26876563e-01
1.12339973e+00 -1.21477091e+00 -5.07433057e-01 1.28897145e-01
2.40901440e-01 -1.58328974e+00 1.06023753e+00 6.62009537e-01
-1.55641794e+00 -5.21215141e-01 -8.41019869e-01 -6.64043054e-03
2.18120739e-01 4.03914116e-02 6.83527052e-01 4.88637686e-01
-1.24597788e+00 6.71208322e-01 -9.44696069e-01 -2.68781602e-01
2.50278294e-01 -8.44521075e-03 2.99896002e-01 1.67444106e-02
-9.41205442e-01 1.40683129e-01 -1.71515375e-01 -1.91058978e-01
-1.16347051e+00 -7.55629420e-01 -3.55601877e-01 -2.01188460e-01
8.95294622e-02 -8.66223931e-01 1.73465860e+00 -1.08644998e+00
-1.71510375e+00 3.06080431e-01 -2.38864556e-01 -3.70197445e-01
6.89714551e-01 -2.79987603e-01 -3.10197383e-01 1.61963880e-01
1.16518728e-01 8.47749531e-01 1.01523650e+00 -1.04965425e+00
-3.82986397e-01 8.69396552e-02 -2.85323262e-01 2.93246537e-01
-4.36835915e-01 1.39908358e-01 -5.63001931e-01 -1.32256973e+00
-3.14907193e-01 -1.23520458e+00 -2.03391746e-01 9.25229304e-03
-5.44016123e-01 1.63383782e-01 4.14519608e-01 -6.18725061e-01
1.19799876e+00 -2.38737798e+00 3.64348702e-02 -9.84675586e-02
-5.94889298e-02 -1.46774486e-01 -3.64726931e-01 8.04965377e-01
1.24699071e-01 1.20229810e-01 -1.47228286e-01 -6.53805494e-01
8.04135054e-02 -2.55772769e-01 -6.25408769e-01 2.12378696e-01
1.46261118e-02 8.37151170e-01 -8.08627009e-01 -3.39451790e-01
-2.37228610e-02 7.44905591e-01 -9.21969891e-01 3.03878617e-02
-5.74862301e-01 6.91247165e-01 -2.47215360e-01 4.58428621e-01
3.67058992e-01 -2.55339444e-01 -1.58847973e-01 1.21628098e-01
-2.75793225e-01 5.54269135e-01 -1.46572077e+00 1.94385684e+00
-2.82651722e-01 7.70727217e-01 1.09158047e-01 -7.87461028e-02
5.49228966e-01 5.59450626e-01 3.96612495e-01 -4.75288004e-01
-1.10714935e-01 4.80471522e-01 4.68138233e-02 -1.90330401e-01
7.41069078e-01 -1.62341937e-01 2.90267318e-02 9.27856445e-01
-7.91500285e-02 -4.72443283e-01 2.04122782e-01 3.57504785e-01
1.03630459e+00 1.47658763e-02 -2.31634930e-01 -7.10385293e-02
-2.21541673e-01 -6.65326789e-02 1.35054803e-02 7.55360484e-01
1.74060896e-01 9.02329624e-01 2.41902694e-01 6.98976144e-02
-1.37940061e+00 -1.20658851e+00 -4.12255861e-02 1.07239580e+00
-3.42731893e-01 -7.14465916e-01 -7.08456933e-01 1.19655296e-01
-1.56185895e-01 8.02634180e-01 -1.56455874e-01 1.96684182e-01
-2.60990739e-01 -8.34851623e-01 9.32519257e-01 2.86047220e-01
1.47273481e-01 -1.14349687e+00 -2.80443549e-01 3.73289376e-01
-2.57381618e-01 -6.60042584e-01 -7.93329418e-01 -2.08514139e-01
-9.52585220e-01 -5.48377931e-01 -1.29609632e+00 -4.92600530e-01
1.32590994e-01 9.28303674e-02 1.07044792e+00 -1.34004399e-01
-2.34639585e-01 5.81224263e-01 -2.79912055e-01 -3.88333082e-01
-6.89120114e-01 1.37652665e-01 3.54108540e-03 -4.17935625e-02
-1.29597276e-01 -8.92225683e-01 -8.99360955e-01 1.04337931e-01
-8.49422157e-01 4.33455050e-01 2.70379692e-01 5.37288010e-01
4.16055471e-01 4.68624756e-04 4.60168421e-01 -6.01112664e-01
1.10181320e+00 -6.96420789e-01 -4.09168571e-01 -4.50021088e-01
-3.62121940e-01 -2.20338956e-01 2.61969000e-01 -7.01737761e-01
-9.29239511e-01 -1.89680621e-01 -4.61414754e-01 -3.85167599e-01
2.57750601e-02 5.59879184e-01 5.03486454e-01 4.27007526e-01
7.73687363e-01 2.78307945e-01 -8.42222050e-02 -5.02369344e-01
7.03014076e-01 7.40603328e-01 3.28353822e-01 -7.81371176e-01
5.40169895e-01 4.32122380e-01 -5.08559167e-01 -9.46696341e-01
-5.06685019e-01 -1.24940738e-01 -4.80205528e-02 -2.90670305e-01
7.21692145e-01 -1.34098196e+00 -3.19602251e-01 7.73086131e-01
-1.04338884e+00 -1.07469034e+00 -5.40156484e-01 4.85748649e-01
-7.97772348e-01 2.54822880e-01 -1.20401323e+00 -8.53769779e-01
-4.18079525e-01 -1.01374841e+00 1.08123839e+00 9.41826552e-02
-6.46233320e-01 -9.31117356e-01 3.90191317e-01 2.75619864e-01
6.32567704e-01 -1.82023823e-01 6.22831345e-01 -1.94999818e-02
-9.38045204e-01 1.06492929e-01 4.94171456e-02 2.54714251e-01
-1.32812541e-02 2.15221509e-01 -1.22626352e+00 -2.65921712e-01
9.11871530e-03 -4.46107745e-01 8.95860136e-01 6.46509707e-01
9.05552566e-01 -3.57491761e-01 4.89175469e-02 5.76683998e-01
1.08662450e+00 -4.52964986e-03 5.64944446e-01 1.02162890e-01
5.21252453e-01 3.10610861e-01 2.48226091e-01 1.00245643e+00
4.92684811e-01 6.54250443e-01 2.35644877e-02 -3.82711031e-02
-6.03996515e-01 -4.09742832e-01 6.17210209e-01 1.50682282e+00
-2.12505326e-01 -3.90333146e-01 -8.86068404e-01 6.69149280e-01
-1.70847559e+00 -1.19193113e+00 -2.48734817e-01 2.03700948e+00
1.16276228e+00 5.39398938e-02 3.07380527e-01 1.69715628e-01
4.92851049e-01 3.65314096e-01 -4.51732039e-01 -2.99671084e-01
-1.27696157e-01 2.59918958e-01 2.38479540e-01 8.43380928e-01
-7.34177947e-01 9.77698207e-01 5.86089897e+00 1.17780304e+00
-1.18175209e+00 3.67218703e-01 8.15659761e-01 -6.08470380e-01
-6.82488918e-01 4.19200957e-02 -6.33349001e-01 5.14025331e-01
1.26317358e+00 -3.61632138e-01 8.70316088e-01 4.84543890e-01
6.00686193e-01 -1.07623965e-01 -7.23007202e-01 9.78955925e-01
-3.93142253e-02 -1.47300065e+00 -1.58406720e-02 2.35080048e-01
1.03274095e+00 4.40716952e-01 5.16878724e-01 -3.66073251e-02
4.86220986e-01 -7.25006521e-01 1.05063260e+00 6.00035012e-01
9.27057147e-01 -6.55838370e-01 4.73148413e-02 3.51633728e-01
-1.18041182e+00 2.29116157e-02 -2.34912336e-01 -2.98494637e-01
4.65777844e-01 7.60097265e-01 -8.61065567e-01 1.05550751e-01
5.75535178e-01 8.93765271e-01 -3.08900237e-01 1.15275168e+00
-9.24020410e-02 1.08344030e+00 -3.66267800e-01 7.15982243e-02
4.38793302e-02 -3.60721871e-02 7.48134077e-01 1.36998940e+00
9.33960140e-01 -1.20611973e-01 -1.57068327e-01 1.01197124e+00
7.97934085e-02 8.73440430e-02 -5.11532247e-01 -3.68260235e-01
3.86491835e-01 9.20831025e-01 -7.19735622e-01 -2.60908127e-01
-1.64760739e-01 1.30672467e+00 -1.39770091e-01 4.40923840e-01
-8.51678133e-01 3.34109254e-02 6.82650566e-01 3.89562696e-01
3.24549347e-01 -7.17699051e-01 -2.92488426e-01 -1.19536722e+00
-2.45904580e-01 -1.21299100e+00 1.59443140e-01 -1.16321409e+00
-1.33855760e+00 4.31882352e-01 -1.70562997e-01 -9.75780666e-01
-5.89785159e-01 -3.82591747e-02 -5.03456354e-01 9.30254221e-01
-1.06431055e+00 -8.57119799e-01 -3.68199386e-02 6.77873790e-01
8.93830836e-01 -1.32916808e-01 9.29179430e-01 5.28091908e-01
-1.11098148e-01 3.24654520e-01 1.78211212e-01 -1.49039671e-01
9.61742699e-01 -9.32048976e-01 1.00146246e+00 5.55171967e-01
5.72216094e-01 5.04854679e-01 8.06106269e-01 -7.22154498e-01
-1.39775646e+00 -9.58838344e-01 8.25029612e-01 -5.58003306e-01
9.31561828e-01 -4.60628361e-01 -5.16171694e-01 6.19435906e-01
5.38066804e-01 -5.24471045e-01 7.35786080e-01 3.04590794e-03
-1.72052354e-01 2.44876519e-01 -4.65618014e-01 9.50442195e-01
8.09997201e-01 -7.06811965e-01 1.32979602e-01 3.32865626e-01
6.84110999e-01 -4.59785998e-01 -7.30514526e-01 -1.02179065e-01
5.90054333e-01 -9.97519195e-01 8.23880613e-01 1.80667818e-01
7.94968724e-01 -3.03123862e-01 -2.46012181e-01 -1.17025149e+00
-2.03950331e-01 -1.10869145e+00 -1.91000596e-01 1.46462071e+00
6.24883354e-01 -3.39521170e-01 6.51282966e-01 2.05404028e-01
-5.31555824e-02 -5.92952490e-01 -7.21647322e-01 -5.16114593e-01
1.28455251e-01 -1.11173117e+00 3.83671701e-01 8.11476231e-01
-1.04062147e-01 4.62171674e-01 -6.73723519e-01 -3.03502027e-02
6.16098404e-01 3.81395817e-02 9.94207025e-01 -5.70976734e-01
-9.28327739e-01 -5.67848146e-01 1.07654698e-01 -1.42095673e+00
-3.16860318e-01 -8.94803524e-01 -8.23945105e-02 -1.50000739e+00
1.89446852e-01 -5.30768037e-01 9.28255394e-02 1.81957141e-01
2.53936321e-01 6.25001729e-01 4.27478433e-01 4.58425224e-01
-4.28758323e-01 6.28203988e-01 1.38948929e+00 -1.94245428e-02
-2.68291503e-01 1.43687427e-01 -6.37659311e-01 6.07410371e-01
1.06706607e+00 -7.60204911e-01 -5.67675710e-01 -7.12038100e-01
4.19966638e-01 3.41698349e-01 3.65410805e-01 -1.18701756e+00
2.75091618e-01 5.51223382e-03 1.86613783e-01 -3.18359971e-01
7.76119232e-01 -1.66438460e-01 4.74086881e-01 2.56333441e-01
-5.41296721e-01 3.16257626e-01 3.76954287e-01 4.23368335e-01
-1.08918868e-01 -3.27207968e-02 5.08584678e-01 -2.33653158e-01
-2.44191065e-01 1.56867802e-01 -7.00637341e-01 2.04137176e-01
5.09556353e-01 1.66269779e-01 -1.97217688e-01 -1.09166467e+00
-6.54632986e-01 -3.31348062e-01 5.55577695e-01 5.82927227e-01
4.67042387e-01 -1.32975388e+00 -9.57955480e-01 6.05835468e-02
-1.78112358e-01 -2.56399691e-01 3.99715662e-01 7.94038713e-01
-5.26395321e-01 5.62461317e-02 1.16237454e-01 -4.59414035e-01
-1.06831133e+00 3.48258734e-01 6.83404040e-03 1.34844452e-01
-6.38428092e-01 1.04147124e+00 1.62124112e-01 6.49107099e-02
1.29066005e-01 -2.01101989e-01 2.94807822e-01 8.02384689e-02
7.16666520e-01 3.10467988e-01 -4.34623927e-01 -5.10906696e-01
1.64588705e-01 2.44878635e-01 2.25466132e-01 -9.94593918e-01
1.23527110e+00 -1.29869759e-01 3.21333036e-02 9.72746909e-01
8.13381910e-01 4.84484643e-01 -1.41844082e+00 4.30346765e-02
-3.52821559e-01 -3.40366006e-01 -3.67480260e-03 -7.79884517e-01
-8.51204932e-01 9.46078658e-01 6.03574693e-01 2.51076221e-01
1.03257418e+00 2.27517486e-02 1.04918337e+00 -2.22785160e-01
2.21969023e-01 -8.60031784e-01 4.38273817e-01 6.32219434e-01
1.04970932e+00 -7.37885714e-01 -1.89755142e-01 -3.80662866e-02
-7.79695451e-01 8.78071487e-01 1.44620258e-02 -2.28757113e-01
6.78967655e-01 6.80611610e-01 2.60895342e-01 6.13614880e-02
-1.21747887e+00 -1.80019550e-02 2.42329940e-01 4.02784467e-01
8.60854387e-01 1.55279741e-01 -5.42360470e-02 5.24230480e-01
-6.80904388e-01 1.41119063e-01 6.24262869e-01 6.02131069e-01
-1.98439389e-01 -1.25022316e+00 -3.91054600e-01 1.73263520e-01
-6.20877981e-01 -6.63360953e-01 -2.39803940e-01 1.50714204e-01
-1.16717875e-01 9.95404959e-01 4.06885892e-02 -2.81625420e-01
-1.09859988e-01 -5.10534309e-02 4.33608025e-01 -6.03576541e-01
-5.03173113e-01 8.06417227e-01 1.70195967e-01 -2.97376096e-01
-1.47844166e-01 -9.60928261e-01 -9.37294900e-01 -6.35964513e-01
-2.88596768e-02 -1.61316022e-02 8.56335521e-01 2.96781659e-01
4.90580231e-01 4.58383054e-01 2.63123065e-01 -1.25105298e+00
-3.72363389e-01 -1.02210724e+00 -6.07282996e-01 8.31575245e-02
2.13374063e-01 -6.08488284e-02 -3.18520576e-01 3.46999615e-01] | [15.622060775756836, 5.71360445022583] |
5fbee0cd-d00a-43cd-ae7a-b60bbc028b15 | s3i-pointhop-so-3-invariant-pointhop-for-3d | 2302.11506 | null | https://arxiv.org/abs/2302.11506v1 | https://arxiv.org/pdf/2302.11506v1.pdf | S3I-PointHop: SO(3)-Invariant PointHop for 3D Point Cloud Classification | Many point cloud classification methods are developed under the assumption that all point clouds in the dataset are well aligned with the canonical axes so that the 3D Cartesian point coordinates can be employed to learn features. When input point clouds are not aligned, the classification performance drops significantly. In this work, we focus on a mathematically transparent point cloud classification method called PointHop, analyze its reason for failure due to pose variations, and solve the problem by replacing its pose dependent modules with rotation invariant counterparts. The proposed method is named SO(3)-Invariant PointHop (or S3I-PointHop in short). We also significantly simplify the PointHop pipeline using only one single hop along with multiple spatial aggregation techniques. The idea of exploiting more spatial information is novel. Experiments on the ModelNet40 dataset demonstrate the superiority of S3I-PointHop over traditional PointHop-like methods. | ['C. -C. Jay Kuo', 'Shan Liu', 'Jintang Xue', 'Min Zhang', 'Hardik Prajapati', 'Pranav Kadam'] | 2023-02-22 | null | null | null | null | ['3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [-2.83984333e-01 -2.66656756e-01 -1.51283056e-01 -2.66160190e-01
-5.13697147e-01 -7.24740684e-01 7.82479465e-01 3.79283041e-01
-1.53298885e-01 4.02842909e-01 -6.24843776e-01 -2.27964118e-01
-3.15422237e-01 -7.88394570e-01 -8.02848756e-01 -5.83051085e-01
-2.61433512e-01 7.39557743e-01 6.45291626e-01 -1.99029878e-01
6.58473492e-01 1.21411514e+00 -1.54267919e+00 -3.03369761e-01
8.07737470e-01 8.02485287e-01 1.56946391e-01 3.37200552e-01
-3.90276968e-01 -1.95126101e-01 -7.17235923e-01 9.40675214e-02
6.96867585e-01 3.07706177e-01 -5.53112090e-01 -1.12760225e-02
4.99423712e-01 -9.15181637e-02 8.84195268e-02 6.93972945e-01
5.80454394e-02 1.03056915e-01 4.97987688e-01 -1.83263886e+00
-7.27101639e-02 5.27612828e-02 -8.47233057e-01 1.89284775e-02
2.51493305e-01 -1.93875253e-01 8.86860728e-01 -1.07424021e+00
5.26842296e-01 1.05128813e+00 7.98961163e-01 3.76195321e-03
-1.11147010e+00 -8.53986740e-01 2.03255937e-01 1.84331447e-01
-1.96981859e+00 8.51686206e-03 1.06336534e+00 -4.17174697e-01
7.31367886e-01 3.24670076e-01 7.10336745e-01 4.94997203e-01
1.20488606e-01 3.14532846e-01 7.73992836e-01 -3.10516328e-01
3.37155849e-01 -5.68027273e-02 3.42645228e-01 5.43229103e-01
5.93370259e-01 -5.17394580e-02 -2.71839291e-01 -4.21187937e-01
9.47246909e-01 5.18760860e-01 -4.54367418e-03 -1.02921629e+00
-1.36871409e+00 6.16451621e-01 9.93103862e-01 2.85636723e-01
-2.79217035e-01 8.77360553e-02 -1.01642117e-01 2.02620223e-01
3.15293849e-01 4.86708820e-01 -4.48731154e-01 2.00704724e-01
-9.77772415e-01 3.89929861e-01 4.55616117e-01 1.58947551e+00
1.27591133e+00 -2.16319486e-01 4.85480070e-01 4.40795481e-01
2.95878857e-01 5.20454407e-01 2.30012508e-03 -8.69289458e-01
4.64976758e-01 8.63189936e-01 2.45539799e-01 -1.21437037e+00
-6.16670310e-01 -5.74576199e-01 -5.97246945e-01 3.42295438e-01
1.63106620e-01 2.57478207e-01 -7.70860493e-01 9.99764621e-01
5.74158728e-01 3.06432486e-01 -1.52308121e-01 8.48749578e-01
4.74249691e-01 4.37324941e-01 -2.31625199e-01 2.23045304e-01
1.08767819e+00 -7.14915574e-01 3.15390266e-02 3.74827273e-02
6.21672094e-01 -7.15050817e-01 7.43781269e-01 3.07094246e-01
-7.73317754e-01 -6.03062391e-01 -1.18564355e+00 1.53425276e-01
-4.78917331e-01 -1.62244052e-01 6.72741890e-01 3.65747184e-01
-9.60645854e-01 6.71421349e-01 -9.12006676e-01 -3.78809392e-01
4.36216891e-01 6.35162711e-01 -4.79103982e-01 5.26983179e-02
-5.86814821e-01 6.01399779e-01 2.65232712e-01 -2.19575480e-01
-2.48502582e-01 -9.27786648e-01 -5.50005734e-01 -8.85196403e-02
2.09574148e-01 -6.64508760e-01 9.44367409e-01 -3.79934490e-01
-1.09432209e+00 6.16393089e-01 -4.63588506e-01 -2.09175661e-01
4.57039237e-01 -1.14964858e-01 -1.42855510e-01 2.77616680e-01
4.01533246e-01 6.79180861e-01 8.68905425e-01 -1.76971626e+00
-9.64546978e-01 -6.28918707e-01 1.44926593e-01 9.62098464e-02
7.75773227e-02 -2.49328077e-01 -5.95460713e-01 -3.22968394e-01
9.35992956e-01 -1.27221835e+00 -3.47884029e-01 2.06112638e-01
-4.92216945e-01 -5.23231149e-01 1.13119280e+00 1.23228244e-01
6.43814266e-01 -2.25464058e+00 -1.51047632e-01 6.88676476e-01
3.51595283e-01 -5.32579841e-03 4.12776396e-02 4.85518217e-01
-1.94501355e-01 3.45593214e-01 -1.66242808e-01 -4.66772139e-01
-3.03686410e-01 4.24972028e-01 -3.03389400e-01 7.84076810e-01
2.09467903e-01 4.39105690e-01 -9.59172368e-01 -5.98670483e-01
5.44347882e-01 3.96326482e-01 -5.93645811e-01 -3.22477430e-01
3.86673138e-02 5.67072868e-01 -7.25483179e-01 1.02490282e+00
1.17801917e+00 -2.15846583e-01 -4.23760742e-01 -2.17501178e-01
-5.13621271e-01 1.47574348e-02 -1.14770782e+00 1.75144911e+00
-3.88651252e-01 3.61676842e-01 -1.32713661e-01 -5.78444481e-01
1.14134252e+00 1.26713127e-01 1.07261109e+00 -7.99905211e-02
-1.07460067e-01 2.92524874e-01 -9.16161463e-02 3.10529679e-01
4.38326150e-01 6.98502138e-02 5.07048592e-02 -5.99029511e-02
-2.38732859e-01 -2.72668988e-01 -2.35660076e-01 1.90280359e-02
9.61720586e-01 3.34714092e-02 3.38342726e-01 -3.22157323e-01
4.88411814e-01 5.72296858e-01 4.06429470e-01 6.79844379e-01
-1.13836586e-01 9.61941242e-01 4.68502305e-02 -3.86914164e-01
-8.74211848e-01 -1.28188992e+00 -4.69884306e-01 5.00690937e-01
7.46146202e-01 -7.02871740e-01 -3.48920286e-01 -5.73960185e-01
4.84669089e-01 3.77261609e-01 -2.87595212e-01 2.33013675e-01
-6.64316773e-01 -2.34836623e-01 9.78651941e-02 4.90859479e-01
4.20740664e-01 -2.16381460e-01 -6.68256760e-01 7.77782500e-02
1.81549251e-01 -1.14338577e+00 5.09623177e-02 4.09233756e-02
-1.16968524e+00 -1.15865314e+00 -5.46007693e-01 -3.89745474e-01
8.32766175e-01 1.13606977e+00 8.68676066e-01 3.92784387e-01
3.85138914e-02 4.19578820e-01 -7.21670926e-01 -6.03712857e-01
2.74822831e-01 3.34788024e-01 2.57085860e-01 -9.39176083e-02
4.89808857e-01 -8.33562791e-01 -6.46437764e-01 6.11859381e-01
-4.74001676e-01 -1.83829218e-01 5.04354894e-01 3.75974238e-01
8.74203622e-01 2.81152755e-01 1.35095358e-01 -6.07740879e-01
-6.57876655e-02 -6.72296345e-01 -6.63226962e-01 -1.74476907e-01
-4.78762001e-01 -2.37469271e-01 5.04883587e-01 4.19898741e-02
-2.98853666e-01 4.53816652e-01 2.07361087e-01 -1.10873520e+00
-1.82659790e-01 1.52239233e-01 -1.64294675e-01 -7.39805758e-01
4.95190471e-01 1.21431179e-01 -1.41809419e-01 -8.00393105e-01
2.57601678e-01 5.97195804e-01 3.16515476e-01 -5.57152331e-01
1.55704486e+00 9.96607304e-01 4.97932345e-01 -1.16450512e+00
-3.68228883e-01 -9.29405451e-01 -1.30141342e+00 -1.24734536e-01
5.37897348e-01 -8.50618184e-01 -6.99226797e-01 -1.43234907e-02
-1.41519022e+00 3.68813902e-01 -9.70983803e-02 4.77267861e-01
-4.49087501e-01 3.33374828e-01 6.62573501e-02 -7.29735017e-01
-5.91054857e-02 -1.13436007e+00 1.45478582e+00 6.41708076e-02
-1.32068768e-01 -7.20854700e-01 -6.71072230e-02 -6.36897832e-02
-5.44596510e-03 4.18405592e-01 6.98629677e-01 -7.43996680e-01
-1.12796855e+00 -3.05159539e-01 -2.06261605e-01 -1.30271286e-01
1.49955273e-01 2.44755924e-01 -7.84029484e-01 -3.83455515e-01
-1.04971252e-01 5.67184806e-01 4.44992781e-01 7.48412386e-02
1.24151194e+00 1.27448559e-01 -7.65274644e-01 9.68627155e-01
1.65229762e+00 3.54275554e-02 2.94714540e-01 6.36334956e-01
9.46362019e-01 4.43541288e-01 7.63739109e-01 3.00626516e-01
5.05928218e-01 8.45370173e-01 7.82386422e-01 -1.93137139e-01
2.10167065e-01 -2.99678594e-01 -1.61626443e-01 7.50335753e-01
-2.44208694e-01 1.48778409e-01 -1.25333655e+00 4.73564804e-01
-1.68603957e+00 -4.79180545e-01 -6.52966321e-01 2.26873660e+00
-4.75329645e-02 1.59238994e-01 1.03157796e-01 4.71232712e-01
6.78106606e-01 -3.13371122e-02 -4.15667713e-01 -6.18426129e-03
1.33763671e-01 2.24017799e-01 1.02570808e+00 3.31058145e-01
-1.18409228e+00 8.16641986e-01 5.85806370e+00 4.87206697e-01
-1.22216296e+00 1.18784956e-03 -1.56988770e-01 6.38442263e-02
9.79359895e-02 4.85759228e-01 -9.33699965e-01 4.17496979e-01
4.90029752e-01 -1.72049448e-01 7.98574388e-02 1.24720407e+00
6.15853211e-03 6.31625652e-02 -1.18070364e+00 1.22870517e+00
-6.93949237e-02 -1.26705086e+00 1.14267305e-01 2.99724996e-01
4.42371488e-01 3.98566693e-01 -5.03650606e-02 7.51276314e-02
-3.54663543e-02 -5.91024280e-01 7.04892695e-01 3.68975133e-01
8.57844055e-01 -6.44739389e-01 3.91701937e-01 4.52880204e-01
-1.51119375e+00 1.33656025e-01 -5.56547225e-01 -2.48333183e-03
-7.11022094e-02 4.27935690e-01 -1.03614521e+00 1.08940125e+00
8.03427219e-01 1.01424730e+00 -8.14395845e-01 1.40693831e+00
1.01903640e-01 2.29222521e-01 -9.23312843e-01 2.87996173e-01
3.82211566e-01 -2.17908338e-01 9.89302635e-01 7.61500835e-01
5.96910179e-01 2.33978909e-02 3.59757662e-01 6.35407448e-01
3.21948886e-01 1.79982465e-02 -8.10464680e-01 6.62019134e-01
8.87557626e-01 1.07721734e+00 -1.03735518e+00 -1.00677788e-01
-5.69421470e-01 6.60752356e-01 -2.59201024e-02 1.97689533e-01
-4.58969533e-01 -4.44592416e-01 9.95145500e-01 3.93212825e-01
4.24675673e-01 -7.60192692e-01 -4.81741101e-01 -9.98492301e-01
1.74722657e-01 -2.86116391e-01 5.63508384e-02 -8.48024368e-01
-1.14756143e+00 4.62204009e-01 4.08769846e-01 -2.01189542e+00
4.33973745e-02 -5.90735555e-01 -7.07043350e-01 8.76173556e-01
-1.94637787e+00 -1.07692122e+00 -5.29882550e-01 8.58135581e-01
4.37487900e-01 4.82889228e-02 6.32285058e-01 1.62581369e-01
-2.32469067e-01 4.23732728e-01 2.31540635e-01 -1.84862837e-01
3.71789128e-01 -1.20926189e+00 4.84340310e-01 5.24013340e-01
2.49761254e-01 9.08557594e-01 6.27558649e-01 -6.48004234e-01
-1.58083570e+00 -1.24679565e+00 6.42303050e-01 -8.12013447e-01
5.13669670e-01 -3.33253205e-01 -9.33672786e-01 6.44042969e-01
-5.46560466e-01 1.50965184e-01 4.78764743e-01 4.85795783e-03
-3.64786029e-01 -4.68771935e-01 -1.27763569e+00 5.27816534e-01
1.12722588e+00 -1.87982112e-01 -5.41167200e-01 4.54169184e-01
8.56210232e-01 -4.22596008e-01 -1.00657427e+00 4.61886108e-01
2.11385474e-01 -8.88816118e-01 1.36937928e+00 -1.22572824e-01
-1.67310759e-01 -7.59305298e-01 -2.44339764e-01 -1.18194103e+00
-4.41436172e-01 -5.09475410e-01 1.83037043e-01 9.40936327e-01
6.85879514e-02 -9.59384024e-01 1.03133452e+00 1.32511884e-01
-2.87769884e-01 -5.45315206e-01 -1.29180133e+00 -1.12982643e+00
2.21201144e-02 -3.71783912e-01 1.14533770e+00 1.04384792e+00
-2.45156884e-01 5.07860184e-02 2.51617074e-01 8.40381324e-01
8.80839825e-01 3.85938674e-01 1.26885235e+00 -1.88548267e+00
2.70519257e-01 -3.23907614e-01 -1.03597569e+00 -1.14922166e+00
-5.05661927e-02 -9.28644359e-01 -2.87481636e-01 -1.18742311e+00
-4.71265465e-01 -1.13771999e+00 -1.50313839e-01 4.38124508e-01
1.65568918e-01 1.55664369e-01 5.06418586e-01 9.96321261e-01
-3.58645648e-01 2.76385665e-01 9.45357442e-01 1.66465729e-01
-2.80603021e-01 2.59325743e-01 -3.31081629e-01 7.07238972e-01
9.91185069e-01 -3.95580769e-01 -3.07220966e-01 -5.80378473e-01
-1.33787692e-01 -1.66671842e-01 5.04379511e-01 -1.59105456e+00
3.97735119e-01 -2.07949907e-01 4.26370680e-01 -1.25148284e+00
6.93888843e-01 -1.54510963e+00 2.80537456e-01 1.99420393e-01
2.65561342e-01 3.34035426e-01 8.27619657e-02 6.86556101e-01
-8.14557299e-02 -5.62329032e-02 6.29521549e-01 -5.56305982e-02
-5.73969245e-01 4.66629714e-01 2.46773466e-01 -6.46878898e-01
1.45722735e+00 -6.34547830e-01 -1.86299160e-01 1.96951479e-02
-4.86975461e-01 3.28346670e-01 1.10143054e+00 3.82176250e-01
6.20913267e-01 -1.31538439e+00 -4.50432867e-01 4.00420725e-01
5.67498982e-01 4.33057755e-01 -1.37239903e-01 6.86616004e-01
-8.79085422e-01 7.24705577e-01 -9.65177566e-02 -1.36704838e+00
-1.13358355e+00 6.31058097e-01 1.10630333e-01 2.53353596e-01
-1.00186110e+00 4.78164792e-01 8.37345608e-03 -4.23229218e-01
-3.12222149e-02 -5.01443863e-01 -2.65829265e-02 -2.26094872e-01
1.29176050e-01 3.86260718e-01 3.93866390e-01 -9.87358868e-01
-6.96888268e-01 1.12036943e+00 4.34852801e-02 1.82070419e-01
1.24180782e+00 -7.42078945e-02 -5.57958661e-03 5.35513520e-01
1.24808121e+00 2.68074311e-02 -9.80193555e-01 -3.02141011e-01
2.08028436e-01 -8.35445285e-01 -9.72391069e-02 -4.09589931e-02
-9.34515953e-01 8.29592764e-01 5.09205937e-01 3.14864129e-01
6.70007288e-01 1.25598222e-01 5.94420195e-01 5.11347294e-01
1.09448969e+00 -4.93495643e-01 -5.43050587e-01 3.81904006e-01
8.50268424e-01 -1.13789916e+00 2.41044462e-01 -8.17231774e-01
-1.28416285e-01 1.07422137e+00 4.47145343e-01 -6.13234580e-01
9.24593389e-01 -1.03018939e-01 -4.23402824e-02 -3.18397105e-01
-2.78202862e-01 -1.26462698e-01 7.23705292e-02 7.84129083e-01
-1.30015895e-01 3.73322377e-03 1.51769035e-02 2.45129034e-01
-7.15518236e-01 -2.47042671e-01 2.27413476e-01 1.24269056e+00
-5.60788572e-01 -1.07180679e+00 -8.96538913e-01 3.78730208e-01
1.35352775e-01 2.88216949e-01 -3.68661702e-01 1.17011952e+00
1.47737309e-01 7.35802650e-01 5.18339157e-01 -6.46701813e-01
5.31302631e-01 -7.86841512e-02 1.56210154e-01 -7.24971712e-01
-3.36285233e-01 -1.91514671e-01 -6.01849139e-01 -7.60832727e-01
-3.63096684e-01 -7.11926639e-01 -1.37575245e+00 -3.69556695e-01
-2.66292274e-01 3.08301508e-01 1.20583963e+00 5.47946632e-01
6.52201593e-01 1.53149918e-01 1.09248531e+00 -1.39062679e+00
-2.87061870e-01 -4.60155606e-01 -5.33735454e-01 1.18132278e-01
5.78732073e-01 -1.05682087e+00 -5.22550583e-01 -3.56081337e-01] | [7.790439128875732, -2.959681749343872] |
08be0153-1293-4695-a9e3-5fd26cd1ef50 | mitigating-backdoor-attack-via-prerequisite | 2306.01983 | null | https://arxiv.org/abs/2306.01983v1 | https://arxiv.org/pdf/2306.01983v1.pdf | Mitigating Backdoor Attack Via Prerequisite Transformation | In recent years, with the successful application of DNN in fields such as NLP and CV, its security has also received widespread attention. (Author) proposed the method of backdoor attack in Badnet. Switch implanted backdoor into the model by poisoning the training samples. The model with backdoor did not exhibit any abnormalities on the normal validation sample set, but in the input with trigger, they were mistakenly classified as the attacker's designated category or randomly classified as a different category from the ground truth, This attack method seriously threatens the normal application of DNN in real life, such as autonomous driving, object detection, etc.This article proposes a new method to combat backdoor attacks. We refer to the features in the area covered by the trigger as trigger features, and the remaining areas as normal features. By introducing prerequisite calculation conditions during the training process, these conditions have little impact on normal features and trigger features, and can complete the training of a standard backdoor model. The model trained under these prerequisite calculation conditions can, In the verification set D'val with the same premise calculation conditions, the performance is consistent with that of the ordinary backdoor model. However, in the verification set Dval without the premise calculation conditions, the verification accuracy decreases very little (7%~12%), while the attack success rate (ASR) decreases from 90% to about 8%.Author call this method Prerequisite Transformation(PT). | ['Han Gao'] | 2023-06-03 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [ 9.35973506e-03 2.30125323e-01 -1.23012532e-02 -3.20926914e-03
1.86537951e-02 -7.61915326e-01 6.25620067e-01 -5.47002964e-02
-4.98014182e-01 8.59368563e-01 -5.50187945e-01 -5.32567203e-01
1.03649579e-01 -1.16349840e+00 -8.96073997e-01 -7.68417299e-01
3.08844298e-01 9.19290558e-02 6.80508077e-01 -2.87664086e-01
1.13918371e-01 7.07618833e-01 -1.63126850e+00 2.44880393e-01
6.21786058e-01 9.00915504e-01 -1.69753149e-01 3.18032086e-01
-1.88414440e-01 5.01618981e-01 -1.23524690e+00 -5.72140574e-01
7.02154040e-01 -2.96226174e-01 -4.87944096e-01 -4.93226767e-01
3.82466972e-01 -4.62055057e-01 -7.33560920e-01 1.70943666e+00
3.24649513e-01 -1.23739511e-01 3.02906930e-01 -1.75676453e+00
-1.05532110e-01 5.41710377e-01 -1.66374668e-01 2.46379793e-01
2.58732289e-01 3.31402272e-01 4.87482876e-01 -4.59417611e-01
4.21531230e-01 1.07791579e+00 5.92525184e-01 7.11395264e-01
-9.56126153e-01 -1.35047519e+00 6.41054362e-02 3.42538387e-01
-1.50012243e+00 -7.38784000e-02 5.31160355e-01 -3.23941022e-01
4.95166689e-01 4.36134040e-01 7.57636487e-01 1.22566199e+00
5.47035992e-01 4.66556340e-01 9.74267602e-01 -2.45366961e-01
-3.85099836e-03 4.49886203e-01 3.78924519e-01 3.58436406e-01
6.41801298e-01 8.40936959e-01 -7.58989947e-03 -1.39420658e-01
4.40369487e-01 9.06063691e-02 -2.50555158e-01 1.28135830e-01
-8.56716037e-01 7.20791280e-01 3.37020338e-01 3.59691828e-01
-8.29074383e-02 1.43923200e-04 4.15046245e-01 4.06128526e-01
-8.32182840e-02 1.57475069e-01 -4.87355262e-01 2.35690624e-01
-7.44807899e-01 4.61646140e-01 6.63892090e-01 9.90192056e-01
7.46377587e-01 1.97664768e-01 -1.40714437e-01 -7.59800104e-03
2.66879380e-01 8.02333117e-01 4.76838201e-01 -3.09809566e-01
6.30527973e-01 6.14578903e-01 -1.51310131e-01 -1.04353380e+00
-1.92283422e-01 -5.14057457e-01 -6.25164866e-01 3.64654899e-01
4.72658008e-01 -4.60457444e-01 -1.16460001e+00 1.77230680e+00
4.50436652e-01 4.04355079e-01 2.48004273e-01 5.97830713e-01
9.19305623e-01 9.28259790e-01 1.08820815e-02 -9.10128802e-02
1.22417045e+00 -2.63705701e-01 -1.06845450e+00 5.10510392e-02
6.98271215e-01 -5.92333078e-01 7.63940096e-01 7.68817782e-01
-3.34441870e-01 -6.55907631e-01 -1.45720458e+00 5.42947114e-01
-7.10405767e-01 -1.20523483e-01 5.32480419e-01 1.17006147e+00
-5.38676560e-01 4.63309288e-01 -4.18155432e-01 -2.08964258e-01
3.52164090e-01 4.66849387e-01 -6.12646103e-01 -5.03973477e-02
-1.84797585e+00 9.99936461e-01 8.15727949e-01 1.95021778e-01
-1.44439614e+00 -6.34195328e-01 -5.84717512e-01 -6.59527257e-02
3.87482166e-01 -1.71435326e-02 8.06542635e-01 -4.85806078e-01
-1.17854774e+00 5.02241135e-01 3.47036153e-01 -6.80732727e-01
8.49344969e-01 -2.46770561e-01 -1.13980746e+00 -2.01234803e-01
1.53031290e-01 3.51736903e-01 7.28775501e-01 -1.15937889e+00
-7.49665737e-01 -1.60841092e-01 1.85932308e-01 -3.86483103e-01
-2.82786071e-01 -1.78956278e-02 6.68372633e-03 -2.70286679e-01
1.21536747e-01 -7.94261992e-01 -1.51335686e-01 -1.43691361e-01
-8.36780250e-01 1.88191291e-02 1.48342371e+00 -2.71822959e-01
1.23018003e+00 -2.55191207e+00 -6.64830089e-01 8.10260415e-01
1.62488624e-01 7.57908642e-01 2.87503779e-01 2.68497437e-01
-3.24594319e-01 2.78922141e-01 -1.21254869e-01 3.79783690e-01
-2.70227462e-01 4.63153332e-01 -8.32142174e-01 5.32092094e-01
-1.53374374e-01 2.89378166e-01 -7.55805612e-01 -2.47719094e-01
2.76373655e-01 1.75324157e-02 -4.57855850e-01 1.44027218e-01
1.17914960e-01 3.86879034e-02 -5.29232919e-01 4.15869236e-01
1.06430483e+00 6.74709558e-01 -2.13783264e-01 -3.03452432e-01
-6.64414167e-02 -1.05906241e-02 -1.34487736e+00 5.42507172e-01
4.12995964e-02 7.43160546e-01 -2.02163219e-01 -8.36481512e-01
1.32535195e+00 4.50265706e-01 -6.10500667e-03 -4.60163951e-01
4.07619029e-01 1.66116372e-01 3.22897434e-01 -5.16820967e-01
2.33524591e-01 -1.88871875e-01 1.13070775e-02 -3.35850343e-02
-5.98201677e-02 4.87448787e-03 -2.51975536e-01 1.94879994e-01
9.11664128e-01 -1.90712601e-01 9.95107219e-02 -2.12953955e-01
6.91666603e-01 2.69305352e-02 8.38494897e-01 7.37743080e-01
-3.18910241e-01 1.64029539e-01 6.48484886e-01 -3.87138754e-01
-6.09436035e-01 -1.13577843e+00 -5.05461931e-01 1.06564373e-01
4.29241747e-01 -3.56657177e-01 -7.89652169e-01 -1.03199768e+00
3.22204143e-01 1.00566804e+00 -4.81781095e-01 -8.49208593e-01
-1.96125075e-01 -5.73772967e-01 1.39043689e+00 1.68010756e-01
9.68222618e-01 -1.05236447e+00 -2.52467632e-01 3.45122144e-02
9.02891159e-02 -1.26700628e+00 4.79044691e-02 2.75773793e-01
-6.47193193e-01 -1.38380909e+00 -2.65892670e-02 -3.99004847e-01
6.90886259e-01 -5.68713024e-02 3.57574373e-01 3.39383245e-01
-1.62252530e-01 -4.59498823e-01 -2.54504830e-01 -6.69368207e-01
-7.34442115e-01 -3.56087595e-01 4.59278703e-01 1.11585416e-01
4.85371828e-01 -3.43696266e-01 2.86253467e-02 5.62344313e-01
-1.01307106e+00 -6.11457705e-01 2.45595738e-01 5.36989331e-01
2.48291686e-01 4.98177856e-01 5.42904377e-01 -9.56882715e-01
4.03466135e-01 -3.41787636e-01 -9.67808604e-01 -1.12528704e-01
-5.52764773e-01 -2.00380132e-01 1.00648320e+00 -7.11658001e-01
-4.93929178e-01 -2.09448293e-01 -4.32945281e-01 -7.07149982e-01
-2.67641604e-01 1.36069223e-01 -8.43548417e-01 -2.88650811e-01
8.32651079e-01 1.73252091e-01 -5.04647680e-02 -3.36703390e-01
-3.53223234e-02 5.46061754e-01 5.65460980e-01 -1.96451589e-01
1.33630538e+00 4.02390778e-01 1.66594684e-01 -6.80219650e-01
-4.34227169e-01 1.12987079e-01 -2.19566241e-01 -3.59669477e-01
7.27850258e-01 -5.73701739e-01 -7.30197906e-01 5.32241762e-01
-1.18540716e+00 2.49032795e-01 -2.08796382e-01 8.37066829e-01
1.03946872e-01 5.25834024e-01 -2.07571596e-01 -7.17362106e-01
2.89282985e-02 -1.23372471e+00 1.82459846e-01 2.29075596e-01
1.81763694e-02 -6.68792546e-01 -2.77942717e-01 -2.37106215e-02
-6.80271685e-02 3.31126809e-01 9.88569677e-01 -1.21449196e+00
-4.46789771e-01 -8.02910149e-01 1.65450037e-01 9.65566635e-01
-7.48493150e-03 3.50735486e-01 -1.26015604e+00 -1.39578402e-01
2.18343347e-01 2.28064641e-01 6.53189123e-01 8.93639252e-02
1.14429140e+00 -1.87929347e-01 -6.16253555e-01 7.62481928e-01
1.17012072e+00 8.67549479e-01 1.09436154e+00 4.61595446e-01
5.40765166e-01 3.48566592e-01 7.38730967e-01 -8.91392957e-03
-3.66384298e-01 4.19774920e-01 8.79056275e-01 -6.14092164e-02
2.13561952e-01 -4.99296755e-01 7.56362855e-01 1.88431218e-01
1.83788225e-01 -3.69916141e-01 -6.45296454e-01 2.79906094e-01
-1.42800868e+00 -1.12282181e+00 -4.58598047e-01 2.35237408e+00
4.69929636e-01 8.43159199e-01 -2.50145495e-01 7.00580239e-01
1.07446063e+00 -1.04631297e-01 -3.51941526e-01 -5.68231165e-01
-2.06029832e-01 -1.24625593e-01 7.75202572e-01 4.99888897e-01
-1.09224761e+00 1.10478163e+00 6.57414055e+00 1.08806562e+00
-1.13251281e+00 -3.03773880e-02 2.75995612e-01 2.10059687e-01
-3.46829951e-01 3.52696657e-01 -1.31782448e+00 6.47599697e-01
6.44054890e-01 -1.06007844e-01 -1.43833561e-02 8.58966351e-01
7.90320039e-02 1.51082218e-01 -1.18446434e+00 8.88280809e-01
-5.51979207e-02 -1.03228366e+00 3.22624296e-01 3.97660077e-01
2.74337143e-01 -2.82735169e-01 -6.26681894e-02 5.81967056e-01
2.11271212e-01 -1.03776157e+00 7.44112492e-01 2.48956293e-01
8.29282165e-01 -9.99977767e-01 1.21309757e+00 6.77306175e-01
-7.93117225e-01 -1.03163250e-01 -5.74501693e-01 -1.05343722e-01
-1.17445722e-01 7.19560862e-01 -8.87116611e-01 7.90004373e-01
5.38363099e-01 2.51059234e-01 -2.54552603e-01 8.71208191e-01
-4.50771540e-01 9.00835454e-01 -6.16499066e-01 -1.55257937e-02
3.14637005e-01 -1.20391212e-01 9.81248975e-01 8.19147348e-01
1.03912607e-01 1.03276344e-02 -7.56720221e-03 9.29359555e-01
1.37733355e-01 -1.17016278e-01 -1.00014007e+00 1.39731839e-01
5.98052323e-01 8.95624220e-01 -4.99380946e-01 -3.68665636e-01
-9.15777385e-02 3.83969963e-01 -3.47823173e-01 2.43620113e-01
-1.25012958e+00 -8.28913987e-01 5.85118175e-01 2.48635143e-01
5.71805015e-02 3.23684186e-01 -1.74990773e-01 -8.39365840e-01
8.19437429e-02 -7.94555664e-01 2.34363288e-01 -5.13375103e-01
-1.02255619e+00 7.35852420e-01 1.70898020e-01 -1.58925676e+00
1.64564043e-01 -7.86661446e-01 -8.81082952e-01 8.76539409e-01
-9.63037312e-01 -6.64866030e-01 -1.04182981e-01 1.04625022e+00
-1.61832333e-01 -6.60797715e-01 6.25731409e-01 5.80034971e-01
-7.94107735e-01 1.00021589e+00 -1.14104420e-01 4.12919432e-01
4.18552786e-01 -4.47195590e-01 2.28735089e-01 1.22884727e+00
2.80313585e-02 5.35847723e-01 9.36806142e-01 -9.10081565e-01
-7.60925114e-01 -1.22788358e+00 8.32106411e-01 -2.98494965e-01
3.94217551e-01 -5.03256142e-01 -7.72729278e-01 7.79507935e-01
-3.73054832e-01 1.66901171e-01 3.33626032e-01 -2.26197913e-01
-3.17715347e-01 -4.79004800e-01 -1.81528604e+00 6.64850950e-01
7.06714928e-01 -4.11521465e-01 -8.09688270e-01 2.25689128e-01
9.20417368e-01 -2.54438728e-01 -4.10926342e-01 4.97188032e-01
4.30355012e-01 -7.75838077e-01 8.18111122e-01 -7.64444411e-01
-3.06561477e-02 -6.42362893e-01 -2.59561270e-01 -8.81710768e-01
-1.04109205e-01 -4.34236467e-01 1.13303825e-01 1.35871291e+00
4.88919646e-01 -1.27000368e+00 7.07158983e-01 1.65295303e-01
-2.21628711e-01 -5.09021699e-01 -1.29494894e+00 -1.08686590e+00
-1.23066425e-01 -6.69061065e-01 1.01509440e+00 1.09650815e+00
-1.84791252e-01 -1.04008891e-01 -3.02420229e-01 6.96678340e-01
4.69008505e-01 -5.43534279e-01 9.27056134e-01 -1.36092770e+00
6.70072883e-02 -3.77370603e-02 -9.95845497e-01 -8.96950126e-01
-4.99883143e-04 -8.55758607e-01 -2.09931448e-01 -1.09605551e+00
-3.94398510e-01 -5.72867990e-01 -4.01328802e-01 6.12660229e-01
4.85275835e-02 -2.54639927e-02 2.87675917e-01 8.45156685e-02
2.58087873e-01 4.06833410e-01 1.21974826e+00 -9.91556719e-02
8.52024835e-03 4.66298252e-01 -4.80654448e-01 7.43723035e-01
6.61330283e-01 -9.76985633e-01 -5.54450989e-01 1.49501458e-01
2.63473094e-02 -8.83444026e-02 5.09646952e-01 -1.26358044e+00
6.49376661e-02 -1.84995443e-01 4.66808677e-01 -8.08118105e-01
2.62829185e-01 -1.42764425e+00 3.25659901e-01 1.03384519e+00
1.18798748e-01 -6.25299588e-02 5.14734566e-01 4.51544404e-01
-2.02430412e-01 -5.10330796e-01 6.09222591e-01 2.20244303e-01
-6.71696603e-01 3.25010896e-01 -5.08451998e-01 -1.77514032e-01
1.41455674e+00 -8.12927544e-01 -2.81576425e-01 -1.49830312e-01
-6.48896992e-01 1.85856655e-01 5.91794029e-02 3.23796481e-01
6.20232761e-01 -1.23869336e+00 -4.58562702e-01 7.41081297e-01
-1.30586728e-01 1.62095670e-02 -2.22733635e-02 4.43641901e-01
-5.93698561e-01 2.05889046e-01 -2.22315177e-01 -4.96740878e-01
-1.22619438e+00 6.30696416e-01 5.32696009e-01 3.26366760e-02
-6.69446230e-01 6.82468712e-01 3.08210522e-01 -4.55275923e-01
1.63492814e-01 -4.87487853e-01 -4.07352328e-01 -6.96254820e-02
4.57849443e-01 2.06137583e-01 1.36118010e-01 -4.70517129e-01
-5.57002664e-01 2.51252890e-01 -1.87929466e-01 1.26192777e-03
7.23156333e-01 4.86039102e-01 5.52624315e-02 2.43180349e-01
1.01576519e+00 3.61467510e-01 -7.99693465e-01 1.86323777e-01
-4.24513161e-01 -5.15362501e-01 9.95185040e-03 -7.11757123e-01
-1.21004820e+00 6.92463279e-01 7.98099637e-01 3.70855182e-01
7.33681738e-01 -3.58806014e-01 7.73807406e-01 4.68813509e-01
4.43617284e-01 -8.26520801e-01 -5.54886639e-01 6.38152778e-01
3.96195710e-01 -6.75500870e-01 -9.62835550e-02 -6.63257122e-01
-4.88986045e-01 1.08251417e+00 8.49053621e-01 -5.02345979e-01
8.79072189e-01 4.19179648e-01 8.93308669e-02 2.28540506e-02
-5.75688660e-01 3.31524253e-01 -1.22426853e-01 7.45279729e-01
-4.45878476e-01 9.00134910e-03 -3.49250108e-01 1.00602496e+00
-6.80701077e-01 -2.60451771e-02 6.58578753e-01 5.90048790e-01
-5.81932366e-01 -8.19425642e-01 -7.49625444e-01 4.66954023e-01
-4.46593881e-01 -4.75349538e-02 -1.98997229e-01 1.26643956e+00
7.33712673e-01 9.44779396e-01 2.65226930e-01 -1.07853651e+00
6.33200824e-01 1.14976116e-01 2.93107890e-02 -3.99462998e-01
-7.56670237e-01 -4.51903582e-01 1.19297400e-01 -4.67570424e-01
2.80427277e-01 -4.22260880e-01 -1.53335667e+00 -6.09652460e-01
-7.17648089e-01 4.20998275e-01 4.47888196e-01 1.09835804e+00
-2.28683859e-01 5.66509664e-01 7.04763949e-01 -3.27532738e-02
-7.20776677e-01 -8.06140482e-01 -6.12379193e-01 2.22353965e-01
3.59010458e-01 -8.82735431e-01 -8.66618335e-01 -3.84766191e-01] | [5.516923427581787, 7.882723808288574] |
be0a8f45-3280-497b-b8a3-8933f925a435 | unist-unified-end-to-end-model-for-streaming | 2109.07368 | null | https://arxiv.org/abs/2109.07368v4 | https://arxiv.org/pdf/2109.07368v4.pdf | Learning When to Translate for Streaming Speech | How to find proper moments to generate partial sentence translation given a streaming speech input? Existing approaches waiting-and-translating for a fixed duration often break the acoustic units in speech, since the boundaries between acoustic units in speech are not even. In this paper, we propose MoSST, a simple yet effective method for translating streaming speech content. Given a usually long speech sequence, we develop an efficient monotonic segmentation module inside an encoder-decoder model to accumulate acoustic information incrementally and detect proper speech unit boundaries for the input in speech translation task. Experiments on multiple translation directions of the MuST-C dataset show that MoSST outperforms existing methods and achieves the best trade-off between translation quality (BLEU) and latency. Our code is available at https://github.com/dqqcasia/mosst. | ['Lei LI', 'Mingxuan Wang', 'Yaoming Zhu', 'Qianqian Dong'] | 2021-09-15 | null | https://aclanthology.org/2022.acl-long.50 | https://aclanthology.org/2022.acl-long.50.pdf | acl-2022-5 | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 4.24493730e-01 -1.00072168e-01 -2.45310903e-01 -4.26458836e-01
-1.58695269e+00 -8.49149704e-01 1.63741648e-01 -1.88102588e-01
-2.66598195e-01 5.20694256e-01 2.97529817e-01 -8.97792041e-01
5.20763993e-01 -3.77725810e-01 -8.98176074e-01 -3.77349436e-01
2.34978139e-01 6.54259264e-01 4.24842834e-01 -1.04604848e-01
1.27253413e-01 1.16616162e-02 -9.79875565e-01 5.68239331e-01
1.02454865e+00 4.36112463e-01 6.17769837e-01 1.29761410e+00
-3.57156694e-01 1.40116975e-01 -9.06022012e-01 -4.10441846e-01
3.10890764e-01 -1.18929863e+00 -9.09496248e-01 -1.85482856e-02
3.60401981e-02 -3.52745593e-01 -4.13579553e-01 9.18051302e-01
8.51931930e-01 2.63923015e-02 2.73925364e-01 -8.06861222e-01
-4.41541284e-01 1.24081194e+00 -7.51841292e-02 6.56209111e-01
4.80771720e-01 1.61787122e-01 9.57761943e-01 -1.05374146e+00
3.83350611e-01 1.18827486e+00 2.32079655e-01 7.91983366e-01
-8.94346654e-01 -3.83470386e-01 -7.66411647e-02 1.11722201e-02
-1.26340258e+00 -1.05543196e+00 2.44197562e-01 6.63974583e-02
1.34307766e+00 6.16627038e-01 6.08002543e-01 1.04292250e+00
3.44076395e-01 9.50903833e-01 8.46393168e-01 -5.06232321e-01
2.60194302e-01 -3.79273176e-01 -1.19036689e-01 2.24025428e-01
-4.69956875e-01 4.55365181e-02 -8.56451035e-01 9.33236927e-02
4.76582617e-01 -4.79943842e-01 -1.46035865e-01 7.12380707e-01
-1.45098734e+00 4.62056249e-01 -2.54468054e-01 2.11583450e-01
-1.75454900e-01 1.86525464e-01 4.97350305e-01 7.18458176e-01
6.12342834e-01 1.23284049e-01 -3.88542026e-01 -9.82026458e-01
-1.20581520e+00 -4.89859246e-02 8.02729964e-01 1.36535347e+00
3.66467983e-01 1.67569786e-01 -3.33856404e-01 8.99867296e-01
7.33922329e-03 9.14820850e-01 4.63856071e-01 -9.81111467e-01
9.48978186e-01 -2.00584844e-01 -1.41640985e-02 -1.02396905e-01
6.57127798e-02 -2.40172148e-01 -3.33644927e-01 -6.26330018e-01
2.98821718e-01 -5.41281343e-01 -9.77696598e-01 1.47020268e+00
2.97800541e-01 1.84818983e-01 1.99528620e-01 1.08615017e+00
5.58140934e-01 1.25652027e+00 -4.57548261e-01 -4.69442070e-01
1.21608102e+00 -1.28737152e+00 -8.39030981e-01 -5.90959609e-01
4.77512866e-01 -1.43418658e+00 1.28810108e+00 6.37049600e-02
-1.60662949e+00 -4.45531368e-01 -8.59401107e-01 -1.17461562e-01
2.70079702e-01 2.44704764e-02 1.39239997e-01 5.97407222e-01
-1.28721690e+00 3.66082579e-01 -9.99696136e-01 -1.74692050e-01
-6.01482391e-02 2.47989029e-01 2.72068083e-01 2.94510946e-02
-1.25801730e+00 5.60525119e-01 -7.41238473e-03 9.64280292e-02
-1.02095914e+00 -2.86933273e-01 -5.64220905e-01 8.16735104e-02
1.81805566e-01 -5.64428627e-01 2.04896402e+00 -8.60223949e-01
-2.09238625e+00 3.24779868e-01 -7.09328830e-01 -4.43365961e-01
4.45331991e-01 -5.26861958e-02 -3.33074272e-01 2.69726813e-01
1.04776889e-01 6.95201218e-01 9.03658211e-01 -6.71571910e-01
-7.25339115e-01 -2.98302863e-02 -3.81666303e-01 5.10774374e-01
1.47283331e-01 4.52389449e-01 -6.45894229e-01 -5.92409074e-01
1.42005876e-01 -1.07513332e+00 -2.07068846e-01 -5.97556710e-01
-4.32345241e-01 -8.16019922e-02 3.60762179e-01 -1.02894080e+00
1.37031984e+00 -2.13499260e+00 1.98130324e-01 -1.79944471e-01
-5.19621253e-01 1.67328849e-01 -3.77873570e-01 6.11873865e-01
3.05911392e-01 1.66138202e-01 -3.69160980e-01 -4.83387172e-01
-6.90603107e-02 3.10210824e-01 -4.93979752e-01 1.61971271e-01
1.22559525e-01 1.04522300e+00 -8.85350049e-01 -4.01144683e-01
-2.18161851e-01 2.99769372e-01 -2.63282418e-01 3.58642042e-01
-3.10883701e-01 4.85975981e-01 -1.78847402e-01 5.91201305e-01
4.22149152e-01 2.06145808e-01 1.41360536e-01 6.66220665e-01
-2.36281112e-01 1.24030209e+00 -6.19493067e-01 2.03218126e+00
-5.56844115e-01 7.42642105e-01 1.21685445e-01 -5.72514176e-01
8.63201618e-01 6.36654496e-01 1.01191454e-01 -8.50856900e-01
1.64787412e-01 6.69665813e-01 1.73829958e-01 -3.85759115e-01
6.36614442e-01 -9.18365493e-02 -1.31741747e-01 7.38368869e-01
-1.49229988e-01 -4.77499813e-01 3.41545105e-01 1.28438964e-01
1.09994733e+00 -8.54559988e-02 -9.33204740e-02 5.43757714e-03
7.69435540e-02 1.67117566e-01 5.68708420e-01 5.16940475e-01
-2.55271524e-01 9.39898968e-01 2.71974176e-01 7.06953406e-02
-1.39253879e+00 -1.10691369e+00 3.06047231e-01 1.13454092e+00
-9.34155565e-03 -4.30187464e-01 -1.30461729e+00 -4.46798116e-01
-5.56761920e-01 7.89819717e-01 1.89614490e-01 -5.19320183e-03
-1.04766989e+00 -3.06926578e-01 7.62628555e-01 3.63652021e-01
-1.52215911e-02 -1.08151591e+00 -4.27482784e-01 5.29322505e-01
-7.99026608e-01 -1.27884328e+00 -1.41559947e+00 2.37041429e-01
-9.57753181e-01 -2.12202877e-01 -7.94252574e-01 -1.04367256e+00
4.82288688e-01 5.68089306e-01 1.01001549e+00 -1.79899465e-02
3.18885893e-02 -2.39149258e-01 -5.66606879e-01 -2.41969734e-01
-1.02025902e+00 3.98094088e-01 2.34908592e-02 -3.96954745e-01
2.62237608e-01 -5.17361999e-01 -5.17490089e-01 4.90629315e-01
-6.07009590e-01 2.03096673e-01 5.39450049e-01 6.42127097e-01
5.97895920e-01 -3.49253267e-01 7.32125819e-01 -2.96592444e-01
8.55365157e-01 -2.72977263e-01 -5.62313855e-01 1.44747585e-01
-4.27216083e-01 -1.02136165e-01 7.51621783e-01 -4.77413774e-01
-7.54468620e-01 -3.03386431e-02 -6.48671687e-01 -1.07651584e-01
-3.79550532e-02 2.80857891e-01 -4.26559970e-02 5.12197316e-01
5.15830696e-01 7.14963377e-01 -7.33568594e-02 -3.53830904e-01
3.29322129e-01 1.19128466e+00 4.52942640e-01 -5.77014208e-01
5.35405874e-01 -1.14865236e-01 -7.41917312e-01 -6.53273880e-01
-2.96167076e-01 -5.15440762e-01 -4.36609477e-01 -1.14704199e-01
4.67612863e-01 -8.85813713e-01 -3.54927510e-01 2.76298374e-01
-1.47396612e+00 -7.73920000e-01 -2.11856216e-01 5.81575155e-01
-7.63304710e-01 4.03361946e-01 -1.05704010e+00 -7.41941571e-01
-8.10506225e-01 -1.38202083e+00 1.31956387e+00 1.73063070e-01
-3.31049114e-01 -4.90242451e-01 8.61652046e-02 4.90901172e-01
4.11036193e-01 -7.82412589e-01 4.59935099e-01 -4.17070776e-01
-7.27476478e-01 1.87993780e-01 1.62314728e-01 3.06805313e-01
9.78895202e-02 4.40073982e-02 -6.66594565e-01 -3.45631748e-01
-9.57898051e-02 4.65869717e-02 6.67096972e-01 4.52499956e-01
6.61438465e-01 -5.92719674e-01 2.79524475e-02 5.10010183e-01
8.93245399e-01 7.22782612e-01 6.11586869e-01 -8.38620514e-02
2.60507822e-01 3.88022453e-01 7.59728789e-01 3.31294149e-01
2.85269529e-01 6.21835172e-01 -8.63102823e-02 1.91423386e-01
-3.50872189e-01 -2.88770318e-01 1.17146432e+00 1.84244311e+00
2.29411140e-01 -6.63638115e-01 -8.45429957e-01 7.36442685e-01
-1.56800604e+00 -9.17549968e-01 -5.04512899e-03 2.09768081e+00
1.24615788e+00 2.78762907e-01 2.39048421e-01 9.25746784e-02
7.39854276e-01 7.81899542e-02 -2.80982882e-01 -1.16723406e+00
3.74560244e-02 3.68127346e-01 5.21687686e-01 1.02257013e+00
-4.75633085e-01 1.29720998e+00 6.86582994e+00 9.75604892e-01
-1.19752884e+00 4.38714683e-01 5.88000119e-01 -4.32298511e-01
-5.33141196e-01 1.17816217e-02 -8.63564372e-01 6.76339924e-01
1.79618537e+00 -3.22139293e-01 8.95228028e-01 3.94625068e-01
7.76699007e-01 8.72920547e-03 -1.00179505e+00 7.09307313e-01
-1.55144572e-01 -1.24353874e+00 -2.51780957e-01 -1.82192221e-01
6.50266707e-01 4.47888285e-01 8.94648880e-02 1.09479837e-01
9.33343843e-02 -6.64932072e-01 1.06341910e+00 -2.72408482e-02
9.46441352e-01 -7.78760076e-01 4.14043248e-01 4.96572882e-01
-1.12849295e+00 2.94215888e-01 -3.91072184e-01 -4.00779620e-02
6.68148935e-01 5.62013924e-01 -1.30204606e+00 5.18213451e-01
3.76751512e-01 2.81189114e-01 -1.31691605e-01 9.73044455e-01
-4.11707878e-01 1.25844324e+00 -3.56000364e-01 -2.10116655e-01
3.34189206e-01 -2.63993084e-01 6.75140917e-01 1.56469965e+00
8.71890128e-01 1.43446296e-01 1.22268565e-01 5.33918560e-01
-1.52130470e-01 2.07174987e-01 -2.58047462e-01 -2.35052541e-01
7.15591252e-01 7.83392012e-01 -8.71555924e-01 -4.09315974e-01
-2.95661122e-01 1.49756002e+00 -1.37283623e-01 4.84780967e-01
-1.02407110e+00 -3.61187428e-01 6.82999372e-01 -1.63103983e-01
2.70346910e-01 -5.98170102e-01 -3.99459004e-01 -1.01175642e+00
3.66851449e-01 -1.18590617e+00 -5.15158772e-02 -6.72683060e-01
-6.58816457e-01 8.55324984e-01 -4.67232198e-01 -1.27662122e+00
-5.00477970e-01 -1.04778059e-01 -5.44673324e-01 1.02378666e+00
-1.30012298e+00 -6.33968651e-01 1.94839001e-01 2.51257569e-01
1.47859251e+00 1.36588126e-01 6.41743660e-01 4.17163163e-01
-5.07704437e-01 6.92811966e-01 1.58280954e-01 4.94146906e-02
7.57125616e-01 -1.01850986e+00 1.48645663e+00 1.24102640e+00
4.09450710e-01 5.33440351e-01 7.95073450e-01 -7.00180352e-01
-1.60843682e+00 -8.01141739e-01 1.33515346e+00 -2.88737744e-01
4.70847934e-01 -6.21948183e-01 -8.06777120e-01 4.61869627e-01
6.93733394e-01 -4.80846226e-01 6.03355050e-01 -3.93277645e-01
-4.01646551e-03 -7.97515363e-02 -6.70594811e-01 7.35440612e-01
1.11329830e+00 -6.15055203e-01 -4.04606313e-01 4.86145467e-01
1.34953845e+00 -8.62431645e-01 -5.67976892e-01 -9.93750244e-02
3.64949614e-01 -5.44510424e-01 5.46936274e-01 -2.03689024e-01
3.33027571e-01 -2.49871820e-01 -2.55222052e-01 -1.55506551e+00
1.94417566e-01 -1.48316431e+00 -1.89582389e-02 1.03995609e+00
1.07120657e+00 -4.37304169e-01 6.39792681e-01 1.24590360e-01
-7.45519638e-01 -6.98599100e-01 -1.17428970e+00 -1.02465653e+00
4.54259478e-02 -5.70019364e-01 7.31905937e-01 3.16536695e-01
2.54870117e-01 5.61234355e-01 -4.05645728e-01 2.60075480e-01
2.41384089e-01 1.07642069e-01 3.90514106e-01 -2.34315276e-01
-4.33361322e-01 -3.41326237e-01 2.47243300e-01 -1.77086210e+00
-8.34737197e-02 -8.91565442e-01 6.59357369e-01 -1.65078759e+00
1.96579043e-02 -2.70077348e-01 1.17696740e-01 3.15166384e-01
-2.36145854e-01 2.09444482e-02 1.97736710e-01 1.56877562e-01
-3.95016879e-01 5.13444901e-01 1.44140804e+00 -1.21663012e-01
-3.66486877e-01 4.28386539e-01 -2.43882924e-01 -4.65568192e-02
1.23666513e+00 -7.22093403e-01 -4.93093193e-01 -9.97369766e-01
4.71975431e-02 6.50263965e-01 -3.24357599e-01 -6.97175562e-01
4.00363088e-01 -2.28144586e-01 -2.17689738e-01 -8.12971652e-01
1.85527682e-01 -4.04834330e-01 1.33427113e-01 4.08569664e-01
-3.59367460e-01 6.32548869e-01 1.09123424e-01 1.34326354e-01
-2.80944407e-01 -2.53220856e-01 5.21479428e-01 9.54438001e-03
-2.47133255e-01 3.25497836e-01 -8.26401651e-01 1.57126099e-01
5.87753773e-01 -3.75278816e-02 -3.21939081e-01 -5.02841473e-01
-6.55008733e-01 2.11761296e-01 3.51418376e-01 6.62704945e-01
7.84036100e-01 -1.11078525e+00 -9.77052093e-01 1.88665569e-01
-3.55736822e-01 -1.30294133e-02 2.45332085e-02 9.60964143e-01
-7.34951913e-01 5.51245630e-01 2.07500026e-01 -7.75199592e-01
-1.32288778e+00 2.62834400e-01 3.08900148e-01 2.06096694e-01
-5.46251297e-01 9.81542289e-01 -3.60297769e-01 -3.61286461e-01
1.30661905e-01 -6.11563265e-01 6.54281914e-01 -3.01270455e-01
6.47114694e-01 2.46064797e-01 3.03025484e-01 -5.49970627e-01
-1.73123896e-01 1.91717014e-01 5.32299243e-02 -8.25431645e-01
7.88512707e-01 -5.77960372e-01 -1.04917228e-01 4.75139856e-01
1.05797601e+00 1.43154874e-01 -1.20993006e+00 -9.66903046e-02
1.02035798e-01 -4.36183512e-01 -3.42661440e-01 -7.42588758e-01
-7.36972034e-01 1.07203054e+00 2.14384377e-01 1.05735138e-01
1.09257758e+00 -3.58492020e-03 1.55212510e+00 8.92944410e-02
3.74282002e-01 -1.18810320e+00 -3.07358205e-02 9.03451562e-01
6.42850041e-01 -8.70922327e-01 -6.55654192e-01 -3.73452812e-01
-6.68912292e-01 1.08683681e+00 3.12044054e-01 3.94139677e-01
8.18628669e-02 7.60850728e-01 5.66748261e-01 4.65311289e-01
-1.07283080e+00 -1.03085712e-01 -7.87242502e-02 3.40916187e-01
6.37975574e-01 4.27683085e-01 -5.67851543e-01 2.93065220e-01
-6.51097298e-01 -3.27762008e-01 6.07850313e-01 7.87240386e-01
-7.82197356e-01 -1.44742560e+00 -4.01995838e-01 -1.87799931e-02
-6.43670499e-01 -5.53603292e-01 -5.42542040e-01 -9.25277919e-03
-3.36212724e-01 1.38224614e+00 7.45065585e-02 -4.62408453e-01
3.52335960e-01 2.26533592e-01 3.88812423e-01 -6.99138343e-01
-6.58450961e-01 7.25108683e-01 3.52913678e-01 -4.94355530e-01
-6.15162440e-02 -8.29186797e-01 -1.63132620e+00 -5.20207584e-01
-2.66488671e-01 4.27002221e-01 8.96133304e-01 7.80348063e-01
5.07864058e-01 6.00531697e-01 9.60680842e-01 -5.55326462e-01
-6.69114232e-01 -1.00675619e+00 -9.63393599e-02 -2.88428754e-01
5.54020464e-01 3.26780111e-01 -3.59067082e-01 2.06039384e-01] | [14.524024963378906, 7.0771942138671875] |
4fa36b25-fddc-4fc6-95a8-eff4abb0816d | time-series-clustering-based-on-the | 1810.11624 | null | http://arxiv.org/abs/1810.11624v1 | http://arxiv.org/pdf/1810.11624v1.pdf | Time series clustering based on the characterisation of segment typologies | Time series clustering is the process of grouping time series with respect to
their similarity or characteristics. Previous approaches usually combine a
specific distance measure for time series and a standard clustering method.
However, these approaches do not take the similarity of the different
subsequences of each time series into account, which can be used to better
compare the time series objects of the dataset. In this paper, we propose a
novel technique of time series clustering based on two clustering stages. In a
first step, a least squares polynomial segmentation procedure is applied to
each time series, which is based on a growing window technique that returns
different-length segments. Then, all the segments are projected into same
dimensional space, based on the coefficients of the model that approximates the
segment and a set of statistical features. After mapping, a first hierarchical
clustering phase is applied to all mapped segments, returning groups of
segments for each time series. These clusters are used to represent all time
series in the same dimensional space, after defining another specific mapping
process. In a second and final clustering stage, all the time series objects
are grouped. We consider internal clustering quality to automatically adjust
the main parameter of the algorithm, which is an error threshold for the
segmenta- tion. The results obtained on 84 datasets from the UCR Time Series
Classification Archive have been compared against two state-of-the-art methods,
showing that the performance of this methodology is very promising. | ['César Hervás-Martínez', 'Pedro Antonio Gutiérrez', 'Antonio Manuel Durán-Rosal', 'David Guijo-Rubio', 'Alicia Troncoso'] | 2018-10-27 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 2.21568700e-02 -6.71724677e-01 5.89322597e-02 -1.88883856e-01
-3.95156860e-01 -6.41908407e-01 3.73801023e-01 7.78221667e-01
-4.93671268e-01 1.03934735e-01 -1.95938185e-01 -1.77207917e-01
-5.08518994e-01 -8.41657877e-01 -9.18362588e-02 -9.60396349e-01
-5.28240681e-01 5.30248225e-01 4.22942847e-01 5.92651218e-02
3.06821406e-01 7.81556606e-01 -1.65626073e+00 3.25228870e-02
9.11532462e-01 1.07487237e+00 1.13268852e-01 4.15724307e-01
-4.03620154e-01 1.47177726e-01 -7.62150884e-01 4.18346494e-01
3.10377300e-01 -6.94720566e-01 -4.94221449e-01 3.02557826e-01
-4.87540632e-01 2.56691307e-01 7.88706318e-02 9.60539579e-01
8.78655091e-02 6.04343474e-01 6.59365237e-01 -1.29076350e+00
1.32530078e-01 4.91835624e-01 -5.73367298e-01 1.58306405e-01
2.32967660e-01 -2.73359746e-01 7.53635883e-01 -3.85856330e-01
4.87553686e-01 9.16991413e-01 6.27369583e-01 -4.38193965e-04
-1.60169482e+00 -4.61052924e-01 -4.31077518e-02 2.61332542e-01
-1.63666511e+00 1.46470815e-01 1.13066816e+00 -6.31528378e-01
6.92327857e-01 4.91851926e-01 9.67596829e-01 1.52465969e-01
1.42033011e-01 1.20431557e-01 1.02599692e+00 -5.39587259e-01
6.44658387e-01 -1.54045820e-01 4.64542985e-01 -6.29706085e-02
-1.56106308e-01 -2.95258492e-01 3.40619802e-01 -3.99509132e-01
3.18512946e-01 4.09887373e-01 -1.34806022e-01 -3.84180754e-01
-1.05446124e+00 7.96616018e-01 3.48685175e-01 1.10349464e+00
-8.07744861e-01 -4.21313435e-01 5.48592150e-01 4.09376979e-01
6.09835863e-01 2.71614552e-01 -2.68125713e-01 -9.31496453e-03
-1.34692216e+00 9.06770229e-02 8.31132174e-01 1.88679144e-01
8.22580993e-01 -3.78564894e-01 5.71370199e-02 6.91845119e-01
1.67503893e-01 5.64661920e-02 8.69199514e-01 -6.10969901e-01
1.72713399e-01 9.91741061e-01 -1.44803405e-01 -1.33108532e+00
-6.02241457e-01 -1.72485858e-01 -7.87846088e-01 6.86546788e-02
4.24766183e-01 -9.37572271e-02 -7.64156461e-01 1.36691988e+00
5.20986915e-01 3.89015436e-01 3.66613083e-02 7.13350117e-01
2.76069820e-01 1.00228035e+00 4.71553132e-02 -9.06116307e-01
1.21376336e+00 -4.64218467e-01 -5.04313350e-01 5.18543780e-01
5.44527769e-01 -7.47106016e-01 7.02290177e-01 3.90990973e-01
-7.26720393e-01 -7.11072505e-01 -8.24739635e-01 6.22659087e-01
-4.59896624e-01 7.71478638e-02 9.61601436e-02 3.22032362e-01
-8.46308649e-01 8.66319299e-01 -1.03690958e+00 -5.23966789e-01
-2.76169240e-01 2.61467069e-01 -7.42191300e-02 4.91062999e-01
-9.83912826e-01 4.47082639e-01 7.91000545e-01 9.60821733e-02
-2.49754161e-01 -5.02010405e-01 -4.58551943e-01 8.98896605e-02
-4.59696837e-02 -4.41874862e-02 6.79519475e-01 -1.14431608e+00
-1.21426570e+00 7.52091825e-01 -1.94413602e-01 -5.55036247e-01
4.88927305e-01 4.44247365e-01 -8.17244470e-01 2.67848969e-01
4.71664071e-02 1.66116610e-01 8.85080576e-01 -1.10039580e+00
-6.67879462e-01 -4.52338398e-01 -3.53324592e-01 -5.36303967e-03
-3.95865619e-01 2.77715474e-01 -5.67048550e-01 -5.70510030e-01
4.93471175e-01 -1.04472458e+00 -4.67045665e-01 -6.51975691e-01
-1.39443085e-01 -6.09042525e-01 9.28008914e-01 -7.41322577e-01
1.83652568e+00 -2.55389905e+00 3.28886658e-01 8.49298835e-01
7.85104930e-02 1.15219345e-02 1.79781318e-01 6.79824233e-01
-4.97347087e-01 -7.60652944e-02 -4.99725550e-01 -1.88531950e-01
-2.59721249e-01 -2.05095485e-02 -2.73699760e-01 5.64241230e-01
-1.22168332e-01 8.85036960e-02 -8.28899503e-01 -5.07929623e-01
3.64950657e-01 2.64519542e-01 -1.53437555e-01 2.49354705e-01
-1.37950882e-01 5.81724763e-01 -4.23602462e-01 -7.11445734e-02
5.72076261e-01 1.04189485e-01 1.63240790e-01 -1.52422011e-01
-6.59847498e-01 -7.85767473e-03 -1.34064496e+00 1.26057994e+00
3.49254347e-02 3.51967484e-01 -4.69379336e-01 -1.33608663e+00
1.40616202e+00 5.35695970e-01 1.37367475e+00 -2.34468013e-01
1.71453387e-01 3.84013861e-01 3.24917644e-01 -3.52884561e-01
3.29473436e-01 5.55803254e-02 -7.71104917e-02 5.25333703e-01
-3.18601847e-01 -3.06159377e-01 6.71023011e-01 -2.65459478e-01
7.82694578e-01 -9.90396664e-02 2.46637106e-01 -2.35506192e-01
9.00691211e-01 3.42489421e-01 4.01267737e-01 1.43843129e-01
-5.46930619e-02 5.57115197e-01 5.03322303e-01 -5.71541667e-01
-1.28050756e+00 -7.25040913e-01 -2.67617762e-01 6.72652364e-01
-1.17046505e-01 -4.13023412e-01 -9.00603831e-01 -3.26259166e-01
-1.64133996e-01 6.00292087e-01 -6.48295283e-01 -2.86821891e-02
-6.44051015e-01 -8.89490426e-01 5.21511845e-02 -7.03127356e-03
1.69599041e-01 -1.13591516e+00 -8.95400703e-01 4.66280818e-01
-2.34545134e-02 -5.70527852e-01 -3.23139131e-01 2.34378934e-01
-1.24564922e+00 -7.90092468e-01 -7.45543659e-01 -6.98376596e-01
8.11447680e-01 3.40988487e-01 5.06674469e-01 2.90184971e-02
-1.24991592e-02 5.41929156e-02 -6.72025084e-01 -1.27830774e-01
-5.72154582e-01 1.07900970e-01 -1.08245231e-01 5.02398491e-01
5.03977656e-01 -7.53184974e-01 -4.52572614e-01 4.80582505e-01
-9.39503133e-01 -3.28017324e-01 9.86394957e-02 2.50575185e-01
8.09072137e-01 7.01237023e-01 2.58781612e-01 -2.43887261e-01
8.10832083e-01 -6.36186838e-01 -8.08623731e-01 2.18075439e-01
-4.30794775e-01 -2.74365872e-01 1.06894064e+00 -7.69414902e-01
-4.58538949e-01 2.68940955e-01 1.65821135e-01 -7.34344244e-01
-2.00641319e-01 8.55981588e-01 1.66045070e-01 2.62228906e-01
5.27444065e-01 4.92219746e-01 3.83411385e-02 -3.71179283e-01
9.74881947e-02 6.70632899e-01 3.06118727e-01 -1.86707005e-01
7.55015314e-01 3.81520897e-01 -1.09740742e-01 -9.00370240e-01
-8.94408748e-02 -9.13460135e-01 -9.80426908e-01 -4.65965182e-01
9.93762553e-01 -2.41051778e-01 -3.75652522e-01 4.88383532e-01
-8.83198202e-01 -2.94414796e-02 -4.92170572e-01 8.59683573e-01
-4.45400894e-01 4.89983320e-01 -3.06202918e-01 -9.05433655e-01
-1.33341968e-01 -9.19375658e-01 6.49056256e-01 2.67353058e-01
-5.33701837e-01 -1.00440574e+00 4.60302889e-01 -2.92919517e-01
1.15309335e-01 5.67347109e-01 8.71350169e-01 -1.11123896e+00
1.61621034e-01 -5.00852644e-01 2.66007334e-01 2.67702013e-01
3.46240968e-01 4.28904951e-01 -4.98365849e-01 -3.33829999e-01
5.42629898e-01 4.71996307e-01 4.17588711e-01 6.23963356e-01
1.18907809e+00 8.43252018e-02 -5.25234520e-01 4.47893858e-01
1.35770833e+00 1.10788214e+00 4.89059836e-01 4.26256031e-01
4.58790064e-01 8.08356285e-01 7.95907974e-01 4.33975816e-01
7.75296390e-02 6.27585411e-01 -6.02934137e-02 -2.36593291e-01
5.68533063e-01 1.57858163e-01 2.41162300e-01 1.10292208e+00
-3.12218070e-01 3.88428010e-02 -1.09207428e+00 5.32759309e-01
-1.91628850e+00 -1.13994622e+00 -4.82615501e-01 2.52691984e+00
5.04846334e-01 2.44045839e-01 6.46446407e-01 8.21496189e-01
1.04285574e+00 1.79428965e-01 -3.30460042e-01 -4.67914551e-01
6.82102069e-02 6.24158345e-02 1.90655664e-01 1.41551465e-01
-1.20787179e+00 4.73543525e-01 5.74655247e+00 6.08487308e-01
-1.41989386e+00 -2.33438551e-01 4.48118091e-01 2.25224525e-01
6.75859815e-03 1.22066863e-01 -1.80683374e-01 7.07265496e-01
1.10852182e+00 -6.11132801e-01 4.78788674e-01 5.67432642e-01
6.37406588e-01 -4.14932445e-02 -8.12613249e-01 7.47454345e-01
-5.45500219e-02 -6.62978232e-01 -2.80437887e-01 4.24980074e-02
6.20209277e-01 -2.16321260e-01 -2.93633401e-01 -6.90553486e-02
-1.01702496e-01 -5.65481007e-01 6.21875465e-01 6.72505438e-01
2.07800686e-01 -8.92732680e-01 5.92242002e-01 3.71763617e-01
-1.58050728e+00 -1.80214062e-01 -1.99821755e-01 -3.79831754e-02
1.93850890e-01 6.80138886e-01 -5.65566838e-01 8.18750381e-01
7.48270392e-01 6.78209841e-01 -3.22283924e-01 1.37766945e+00
2.26496831e-01 6.09888196e-01 -5.67049503e-01 -6.29162490e-02
2.84708709e-01 -8.39086652e-01 6.04268193e-01 9.75115120e-01
6.94688320e-01 1.90530360e-01 4.27798748e-01 7.42035508e-01
6.01574123e-01 5.18308282e-01 -3.16405237e-01 -6.10977896e-02
6.42822027e-01 1.19518805e+00 -1.42925322e+00 -5.33214033e-01
-2.65336812e-01 7.43817866e-01 -2.17908502e-01 3.05684328e-01
-8.02527845e-01 -5.33861458e-01 2.06799477e-01 2.93300655e-02
2.97207773e-01 -5.23239017e-01 -1.09126240e-01 -8.72922122e-01
1.06325574e-01 -6.83205843e-01 7.10740089e-01 -4.80499595e-01
-1.11756504e+00 8.40009034e-01 1.94781810e-01 -1.91899967e+00
-4.72384065e-01 -2.51577403e-02 -8.00268531e-01 9.21023428e-01
-7.65381515e-01 -5.43520033e-01 -1.94337666e-01 7.58268237e-01
3.38594556e-01 1.05731666e-01 8.07962894e-01 3.00643116e-01
-5.84012210e-01 -3.69616784e-02 4.06274706e-01 1.69822425e-01
4.03585851e-01 -1.07805204e+00 1.76704139e-01 9.08773124e-01
-5.70546128e-02 6.11482203e-01 8.61637175e-01 -7.12328374e-01
-8.90451789e-01 -8.99170160e-01 9.93880093e-01 1.22373410e-01
9.60332334e-01 -2.13629976e-01 -1.27733731e+00 3.98515612e-01
-1.21420763e-01 -3.07263583e-01 8.29507828e-01 -1.79295212e-01
1.55495733e-01 -2.56067395e-01 -1.02061462e+00 4.33544457e-01
3.48017812e-01 -2.91922748e-01 -8.63647401e-01 7.56235421e-02
5.30265450e-01 5.03694527e-02 -1.23094583e+00 1.60232782e-01
4.00832713e-01 -9.06830370e-01 6.95144594e-01 -1.58668339e-01
6.49783090e-02 -6.84071362e-01 8.65438506e-02 -1.34709895e+00
-4.40742075e-01 -5.80544531e-01 4.60567504e-01 1.38716805e+00
2.16820464e-01 -7.18729615e-01 3.71929526e-01 2.66484350e-01
1.28646595e-02 -3.41689348e-01 -8.48650634e-01 -9.32695389e-01
-2.08268270e-01 -4.10166860e-01 8.01280200e-01 1.18674171e+00
2.12685332e-01 -7.47735649e-02 1.95812702e-01 7.80560598e-02
4.98370826e-01 5.77099025e-01 6.19298339e-01 -1.65494859e+00
3.05777173e-02 -7.84993529e-01 -5.85960031e-01 -2.54408598e-01
-2.67076828e-02 -7.32699871e-01 -4.48938385e-02 -1.24440897e+00
-2.47207642e-01 -5.97706079e-01 -4.14967954e-01 1.47883803e-01
-6.47237897e-02 -4.04948443e-02 2.36095220e-01 8.32264185e-01
-1.37371585e-01 1.86338708e-01 6.93804383e-01 1.63519919e-01
-1.01741910e+00 2.75896400e-01 1.27587900e-01 6.23067915e-01
7.89620161e-01 -4.44830775e-01 -3.88295472e-01 1.46049336e-01
-4.62915331e-01 2.42897764e-01 -1.43857319e-02 -1.22193027e+00
4.02935416e-01 -1.47679731e-01 1.34100899e-01 -8.54902804e-01
-5.70087619e-02 -1.31782341e+00 8.19510281e-01 7.51321256e-01
-1.21204779e-01 3.10875058e-01 1.32379368e-01 1.41647756e-01
-4.86504883e-01 -1.99879557e-01 7.19115376e-01 1.80105954e-01
-5.83506703e-01 1.43931627e-01 -4.83087122e-01 -5.53968012e-01
1.46442091e+00 -5.31621158e-01 4.30438876e-01 -1.15121312e-01
-1.01651061e+00 1.21673338e-01 5.49471796e-01 2.92657703e-01
2.45977938e-01 -1.39305568e+00 -6.28051519e-01 2.24455982e-01
1.28630534e-01 -2.19096288e-01 2.55684316e-01 1.08850539e+00
-5.40988207e-01 1.26006141e-01 -2.88460404e-01 -9.99098480e-01
-1.27063096e+00 1.04568970e+00 2.76746124e-01 -2.69571066e-01
-6.63727045e-01 -2.76503023e-02 -3.15455824e-01 -1.22064136e-01
-9.37813427e-03 -5.89997530e-01 -7.34803736e-01 7.00047076e-01
3.92144889e-01 4.21288133e-01 6.37338087e-02 -8.74224126e-01
-2.97747761e-01 1.06281507e+00 4.16278124e-01 -2.42820546e-01
1.51316011e+00 -8.79125521e-02 -5.43366373e-01 1.04011297e+00
1.33035648e+00 -8.11022595e-02 -9.48134363e-01 -9.42285359e-02
5.05305409e-01 -2.96119511e-01 -3.51226330e-01 -7.83861652e-02
-8.56425762e-01 5.00953674e-01 6.80864871e-01 1.11413765e+00
1.69427550e+00 -1.55662507e-01 5.11679351e-01 -2.78284729e-01
2.21139342e-01 -1.01857829e+00 -5.05101800e-01 3.12936932e-01
6.58485413e-01 -5.78478694e-01 -2.96274185e-01 -3.08037758e-01
-3.49564940e-01 1.36259425e+00 -9.71022695e-02 -4.15517539e-01
7.30494440e-01 3.20664048e-02 -3.94323617e-02 1.81550682e-02
-3.08874279e-01 -1.41801029e-01 3.94646585e-01 1.46224812e-01
4.83092993e-01 1.86806262e-01 -1.00511670e+00 3.40098500e-01
-4.22205627e-01 -3.41016124e-03 3.42515111e-02 6.47265851e-01
-5.35121143e-01 -1.01742494e+00 -7.88529158e-01 1.42130956e-01
-2.31414899e-01 4.03079301e-01 -3.87153476e-01 7.17865109e-01
2.28551432e-01 1.13281786e+00 5.36306441e-01 -4.18598413e-01
5.67767620e-01 1.97446525e-01 3.39037143e-02 -3.00548077e-01
-9.00690496e-01 6.09761834e-01 -4.47239339e-01 -3.00238103e-01
-5.40154159e-01 -1.05674183e+00 -1.54269433e+00 -1.80020824e-01
-2.08678201e-01 6.73660576e-01 7.34458983e-01 6.81240678e-01
5.42239733e-02 4.96277511e-01 1.27425170e+00 -9.43125784e-01
-1.40056297e-01 -7.38491297e-01 -6.89658523e-01 6.90634608e-01
6.99065700e-02 -1.96789563e-01 -5.70309699e-01 2.13734522e-01] | [7.238276481628418, 3.383274555206299] |
b31064f8-8956-489b-b59c-0683e140ff56 | autogpart-intermediate-supervision-search-for | 2203.06558 | null | https://arxiv.org/abs/2203.06558v4 | https://arxiv.org/pdf/2203.06558v4.pdf | AutoGPart: Intermediate Supervision Search for Generalizable 3D Part Segmentation | Training a generalizable 3D part segmentation network is quite challenging but of great importance in real-world applications. To tackle this problem, some works design task-specific solutions by translating human understanding of the task to machine's learning process, which faces the risk of missing the optimal strategy since machines do not necessarily understand in the exact human way. Others try to use conventional task-agnostic approaches designed for domain generalization problems with no task prior knowledge considered. To solve the above issues, we propose AutoGPart, a generic method enabling training generalizable 3D part segmentation networks with the task prior considered. AutoGPart builds a supervision space with geometric prior knowledge encoded, and lets the machine to search for the optimal supervisions from the space for a specific segmentation task automatically. Extensive experiments on three generalizable 3D part segmentation tasks are conducted to demonstrate the effectiveness and versatility of AutoGPart. We demonstrate that the performance of segmentation networks using simple backbones can be significantly improved when trained with supervisions searched by our method. | ['Li Yi', 'Chuang Gan', 'Anyi Rao', 'Xiaomeng Xu', 'Xueyi Liu'] | 2022-03-13 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Liu_AutoGPart_Intermediate_Supervision_Search_for_Generalizable_3D_Part_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Liu_AutoGPart_Intermediate_Supervision_Search_for_Generalizable_3D_Part_Segmentation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['3d-part-segmentation'] | ['computer-vision'] | [ 0.3424037 0.5234862 -0.3031513 -0.49691108 -0.34423572 -0.6637228
0.15103869 -0.4290018 -0.17214265 0.34977025 -0.34005108 -0.41969037
0.05740527 -0.66496414 -0.87286246 -0.5101998 0.25076768 1.051043
0.48473665 -0.15090641 0.18278025 0.5975062 -1.2855269 0.03213824
1.1444839 0.9314388 0.5465373 0.29360658 -0.46454996 0.15140091
-0.7084704 -0.23796523 0.66463464 -0.3712981 -1.2667567 0.5997459
0.42549956 -0.07707864 0.19415115 1.0224495 0.15457332 0.01321228
0.70522743 -1.2314774 -0.5475851 0.45105863 -0.47395968 -0.14344831
0.12689677 0.03906167 0.8225936 -0.38132516 0.6968311 1.2520871
0.724341 0.9758665 -1.3484067 -0.3665772 0.5102161 -0.04580524
-1.0935771 0.07229921 0.9910937 -0.3786739 0.89519006 -0.00941994
0.6993764 0.8430072 -0.1749936 1.2280877 0.9674907 -0.47175682
0.20536222 0.26169977 0.32674977 0.6712878 0.3003789 -0.21899185
-0.21057786 0.22982664 1.2200079 -0.01065388 -0.3187645 -0.9732469
-1.1496431 0.6650742 0.5705209 0.29187164 -0.24644284 -0.10725399
0.3615078 0.43304276 0.66499984 0.6866889 -1.10053 0.20348164
-0.90102756 0.13010229 1.0941193 1.3951182 1.1025295 -0.229147
0.12711838 0.85885495 0.28537187 0.29781497 0.34134153 -1.1122558
0.34813496 0.9828552 -0.21463174 -0.41411462 -0.48950458 -0.34818095
-0.6839113 0.18579508 0.5817682 -0.16830972 -1.4507083 1.5611506
0.62288934 -0.17994462 -0.06823108 0.91772753 0.4502985 0.37156352
0.10891367 0.20917846 1.3040838 -1.2372662 -0.25258923 -0.48766363
0.68217313 -0.49808994 1.2400318 0.27129963 -0.8815782 -0.7300813
-0.9307566 -0.11325734 -0.31861168 0.07876917 0.7979366 0.72139573
-1.088185 0.75777364 -0.98562276 -0.5402705 0.656483 0.60652363
-0.36864308 -0.04102537 -0.6872772 0.8280621 0.8657893 0.04732494
-0.79288846 -0.63871986 -0.8008029 -0.06408156 0.8218908 -0.98022795
1.5523129 -1.3407803 -1.6272181 1.1592282 0.02605171 -0.43532354
0.5134023 -0.32907996 0.24896437 0.30638906 0.24401906 0.9656337
0.91626626 -1.3248101 -0.4841808 -0.45842868 0.17259191 0.31632492
-0.09114768 -0.37138623 -0.6781379 -0.45436698 0.53059435 -1.0454364
-0.5205009 0.16009805 -0.54809916 -0.49987778 0.9218059 -0.31705442
0.43253502 -2.012206 0.2602833 0.14671761 0.10504512 0.44854584
-0.17981239 0.1490656 0.12892008 0.24007852 -0.47749984 -0.35850817
0.06343254 0.52979743 -0.07953501 0.1591015 0.18819514 0.944062
-0.66436076 -0.6409137 0.09269678 0.06299612 -0.6843766 0.33141178
-0.8444476 0.5373135 -0.8404112 0.61528605 0.7187422 -0.54555047
0.15013681 -0.01439541 0.26124802 0.28010717 -0.9569714 2.0621488
-0.36509013 0.26945788 0.20816895 -1.4871677 1.004013 0.21476641
0.39157188 -0.3536279 0.03397465 0.20280789 -0.2643122 -0.46064696
0.07514956 -0.35240692 0.02205043 0.5498897 0.39358705 -0.4964401
-0.00661603 -0.02920784 0.91071063 0.5887476 0.16682212 -0.31313324
0.4312709 0.55634946 0.6182482 0.75731504 -0.32336396 0.53903675
0.46144375 -0.60841554 -0.8978255 -0.9615559 -0.02597028 1.0834147
0.41931465 0.03104013 -1.150601 -1.4388644 -0.03920298 0.5844197
-0.6085008 0.03304248 -0.6341258 -0.34708583 0.18897861 0.55378765
1.0293182 -1.2387041 -0.59106517 0.10565773 -0.13895632 -1.3381418
-0.34447843 0.4026439 -1.3714592 -1.0623683 -0.97037935 -1.1293863
1.0918523 0.5382686 1.2794385 0.23645216 -0.18316081 0.46038803
-0.4842703 -0.4181539 -0.5321004 0.5793787 -0.3710773 -0.43114874
0.55517066 -0.5913963 -0.42794102 0.74368507 -0.780745 0.16462259
0.8248737 0.5828798 0.68273544 -0.02052806 0.5599831 -1.1051787
0.26308346 -0.11866824 -0.59146225 0.3729003 -0.37478048 0.22095506
0.67022634 -0.34942788 -1.0555142 0.44091457 -0.11769721 -0.5817526
-0.73761344 0.0437902 -0.50816065 -0.16036847 0.6516656 0.35004008
0.13355999 -0.82577574 0.3450681 0.33160064 0.35095793 -0.86993015
0.89573747 0.4389584 -0.01292024 -0.6872085 -1.1955377 -0.6552529
-1.1335199 0.06315401 1.1691552 -0.70117027 -0.38309234 0.44060263
-1.2431419 -0.68398225 -0.3978484 0.05349058 -0.80808955 0.43132606
-0.24906377 -0.29755405 -0.21801855 -1.1545403 1.1238508 0.22913653
-0.09917178 -1.1080989 -0.17558353 0.54525554 0.11155493 0.04634924
1.081287 -1.0924093 -0.92108643 -0.0709045 -0.3519755 0.62300116
0.2764696 -0.42176473 -0.791904 -0.04226888 0.14349242 -0.2897875
0.64494085 0.4052883 1.3166987 -0.11308552 -0.57318985 0.6747145
1.2926644 0.1310747 0.32576975 0.4261647 0.7369243 0.70913243
0.6316734 -0.13343468 0.27533016 0.55148953 0.48626822 -0.22689408
-0.11836912 -0.3508595 -0.03158094 0.31551602 -0.11841638 -0.15730253
-0.97132236 0.58610964 -1.651943 -0.6342858 0.133721 1.7127771
0.7854201 0.32949573 0.1589583 -0.00630587 0.7150937 -0.19844033
-0.7619367 -0.28775787 0.35626838 0.08263548 0.39110804 0.08965536
-1.220706 1.4333225 6.2649264 0.7453791 -0.97938925 -0.04473281
0.49149093 0.35920462 -0.13942605 0.12882479 -0.8481933 0.03936951
0.40670425 0.23097183 0.18397537 1.3046311 -0.01840086 -0.04662338
-1.344915 0.70557034 -0.09430514 -1.0768287 0.21435869 0.08311143
0.79532987 -0.02134184 -0.24025023 0.32148692 0.4140341 -0.72351617
0.6170469 0.1101851 0.37645316 -0.56820303 0.5426033 0.7232034
-0.8938361 0.20596655 -0.5386594 0.10202438 0.14105429 0.50054896
-1.2968303 0.5341975 0.6518299 0.6292036 -0.4360697 1.0918511
-0.47938263 0.61413616 -0.4207731 -0.09948953 0.3725235 -0.25007114
0.54639196 1.0666257 0.0752326 0.03388257 0.4334895 1.1515754
-0.12220769 -0.03344577 -0.7195906 -0.01189968 0.14172328 1.0505434
-1.2720903 -0.29948425 -0.26414898 1.2375257 0.20273189 0.3964228
-0.44682193 -0.35594922 0.50593626 0.12619688 0.5681238 -0.2993657
-0.5597546 -1.0001105 0.05927332 -0.88098985 0.17587316 -0.59297687
-1.2673231 0.47328365 0.16418198 -1.0729147 -0.05673967 -0.95217836
-0.6020033 0.6153656 -1.5206164 -1.2495105 -0.24846485 0.6323639
0.8503966 -0.05036774 0.5463878 -0.16167493 -0.44012457 0.4360759
-0.36141026 0.25258163 0.5034449 -1.4930143 0.7618824 0.5635432
0.17794356 0.4724702 0.6441469 -0.64248794 -1.12906 -1.0275809
0.6845683 -0.4463041 0.3672245 -0.5525731 -1.024296 0.85299826
-0.00800982 -0.1442569 0.6185594 0.35941714 -0.2681185 0.18405786
-1.2454238 0.48662022 1.4389982 -0.23862979 -0.9277701 0.5467412
1.0073324 -0.45347062 -0.71237564 0.45150334 0.13584465 -0.8800136
0.8929379 -0.57887745 0.2598054 -0.29901102 0.12008183 -1.2643712
0.01749237 -0.456208 0.10972858 1.0155969 0.47665027 -0.6254351
1.2049397 0.5455223 -0.5418165 -0.71472406 -0.7176928 -1.0286138
0.25618255 -0.35396236 0.7636903 0.7725661 -0.3572383 0.28505102
0.13452539 0.31182376 0.6423885 0.52156943 1.044474 -1.4936903
-0.33171466 -0.36252853 -0.40505683 -1.7578183 0.37030715 -0.9918426
0.40920472 -1.7255409 0.07848221 -0.48799688 0.09710761 0.87693614
0.01175496 -0.10866068 0.25709927 0.3543125 -0.6963043 0.4502206
1.896202 -0.15856989 -0.37459922 0.50370365 -0.81934667 0.93806595
0.811749 -0.4317305 -0.775398 -0.6479498 -0.1916558 -0.10702045
0.52251136 -1.0383401 0.17437801 -0.15361959 0.27450615 -0.7443915
0.03363715 -1.0726044 -0.20001325 0.27814284 -0.12478677 -0.37060875
0.26309225 0.504144 -0.11636686 -0.49272564 0.79679877 -0.49367708
-0.9994071 0.42617586 -0.104879 0.12123612 1.0538156 -0.61555344
0.01742516 0.05932685 -0.99919087 0.4632633 0.7366901 0.27339774
0.40972814 -0.82000077 -0.24706706 0.28951967 -0.04818593 0.70031095
0.1129128 0.38226113 -0.5798887 0.4193857 -0.22343348 -0.96101475
-0.9094036 0.60368574 0.5634062 -0.2687242 -0.80123705 0.98490053
0.41522264 -0.8825628 0.19386382 -0.38029826 0.07992792 -0.23007357
0.03068969 -0.02775381 0.00768377 -0.2659251 -0.17510061 0.8313403
-0.30328515 0.1797079 1.3076981 -0.05445028 0.10314941 -0.02484709
1.0782716 -0.62153435 -1.7428268 -0.29240835 0.32398176 -0.26199815
-0.2883533 -0.9023928 -1.2609177 0.97303003 0.30946955 0.43631345
1.1339697 0.37200803 0.8408748 0.75428516 0.5191548 -1.1704749
0.20594689 0.6734834 0.6325214 -1.2635994 -0.20748511 -0.7138237
-0.70253277 1.0973958 0.93451977 -0.26682943 0.73729134 -0.1259686
0.31336957 -0.3736937 -0.35678023 -0.31080475 0.332042 0.8893923
0.02732838 -0.18878675 0.04970867 0.44293168 -0.22019097 -0.134372
0.12123112 1.0128474 -0.7153443 -1.3818446 -0.24645421 0.21268179
-0.21649326 0.4516792 -0.62691665 1.0851331 0.2688471 0.6079548
-0.0782354 -0.04080249 0.45285276 0.34860128 0.47181463 -1.0796604
-0.40558663 0.08259782 -0.13830289 -0.54981947 -0.6089539 -0.574109
-1.4501635 0.1466584 -0.3727276 0.20586541 0.7312743 1.3060516
0.33327225 0.40268013 0.4585083 -1.1321188 -0.55525655 -0.6736857
-0.5050149 0.38904586 0.16446362 -0.6541177 -0.20718592 0.13567403] | [9.320815086364746, 0.4884880781173706] |
7628b838-f161-4dfa-bd6d-e7a5e65d9950 | learning-context-graph-for-person-search | 1904.01830 | null | http://arxiv.org/abs/1904.01830v1 | http://arxiv.org/pdf/1904.01830v1.pdf | Learning Context Graph for Person Search | Person re-identification has achieved great progress with deep convolutional
neural networks. However, most previous methods focus on learning individual
appearance feature embedding, and it is hard for the models to handle difficult
situations with different illumination, large pose variance and occlusion. In
this work, we take a step further and consider employing context information
for person search. For a probe-gallery pair, we first propose a contextual
instance expansion module, which employs a relative attention module to search
and filter useful context information in the scene. We also build a graph
learning framework to effectively employ context pairs to update target
similarity. These two modules are built on top of a joint detection and
instance feature learning framework, which improves the discriminativeness of
the learned features. The proposed framework achieves state-of-the-art
performance on two widely used person search datasets. | ['Bingbing Ni', 'Yichao Yan', 'Minghao Xu', 'Xiaokang Yang', 'Wendong Zhang', 'Qiang Zhang'] | 2019-04-03 | learning-context-graph-for-person-search-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Yan_Learning_Context_Graph_for_Person_Search_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Yan_Learning_Context_Graph_for_Person_Search_CVPR_2019_paper.pdf | cvpr-2019-6 | ['person-search'] | ['computer-vision'] | [ 1.94764454e-02 -5.21090329e-01 6.83692005e-03 -3.82739902e-01
-4.72464949e-01 -3.36529195e-01 6.79795861e-01 1.69689327e-01
-5.73998392e-01 3.82657856e-01 2.82932222e-01 2.15279400e-01
-1.41796917e-01 -6.29935682e-01 -2.59617418e-01 -4.59942192e-01
2.67371207e-01 3.08717966e-01 1.65075928e-01 4.60587367e-02
1.71010211e-01 4.57273275e-01 -1.53365409e+00 -1.01336122e-01
6.48118258e-01 8.51426065e-01 1.11346014e-01 4.79920000e-01
3.69368233e-02 3.59102815e-01 -5.19844949e-01 -7.07144439e-01
2.60318309e-01 -2.35850349e-01 -7.00182676e-01 2.45173171e-01
8.23771477e-01 -4.16871578e-01 -8.42552304e-01 1.10400093e+00
9.46300805e-01 5.10370135e-01 4.64238256e-01 -1.16529143e+00
-8.74913752e-01 3.37858386e-02 -6.20371282e-01 4.73232508e-01
6.32947505e-01 4.34656441e-01 8.95154655e-01 -8.54733825e-01
2.81192213e-01 1.40693319e+00 7.52437055e-01 6.32822394e-01
-1.08332777e+00 -7.11688876e-01 6.23889863e-01 6.65745318e-01
-1.41923773e+00 -3.70065302e-01 9.77500558e-01 -1.70023412e-01
8.17923844e-01 2.19861045e-01 9.22566056e-01 1.13392818e+00
-3.08042735e-01 1.05559957e+00 7.99274564e-01 -4.25497711e-01
-1.36089966e-01 1.70679808e-01 5.26745021e-01 7.26004303e-01
3.88461322e-01 1.85807452e-01 -5.04859507e-01 -1.95350617e-01
6.82594478e-01 5.07598579e-01 -2.77583927e-01 -3.81987870e-01
-8.01784098e-01 6.94772899e-01 8.59655380e-01 3.47363502e-01
-2.04825208e-01 2.61040479e-01 3.61173362e-01 7.50933960e-02
1.64374962e-01 2.44921446e-01 -1.07292108e-01 1.45031720e-01
-7.16464579e-01 3.26787233e-01 6.22583270e-01 8.18735301e-01
7.17824578e-01 -3.58388454e-01 -6.82200015e-01 1.09685564e+00
5.26980579e-01 3.93154919e-01 3.78856212e-01 -4.76692557e-01
3.17401886e-01 8.62417400e-01 3.10326777e-02 -1.21865928e+00
-3.63491297e-01 -7.39565313e-01 -6.60083115e-01 -1.36436582e-01
3.32322031e-01 1.49400249e-01 -1.00069451e+00 1.78385580e+00
4.72328007e-01 4.31561202e-01 -3.57847780e-01 1.08158135e+00
1.04461741e+00 3.98933701e-02 1.96159482e-01 1.38037354e-01
1.46784317e+00 -1.35713708e+00 -5.40843666e-01 -3.65534812e-01
3.33456904e-01 -6.38426840e-01 9.10258949e-01 4.34836559e-02
-8.41826022e-01 -9.82226849e-01 -9.03419614e-01 -2.57836789e-01
-5.29604793e-01 3.15224439e-01 5.71659267e-01 6.72490895e-01
-9.91003394e-01 4.31431144e-01 -5.26856840e-01 -6.99049771e-01
3.71852964e-01 4.67438191e-01 -4.11924601e-01 -5.07368624e-01
-1.08962643e+00 7.38282382e-01 1.48670718e-01 4.95092005e-01
-6.66591227e-01 -3.14722687e-01 -9.05460835e-01 1.46588504e-01
3.40313524e-01 -9.53479826e-01 8.66853654e-01 -7.08577096e-01
-1.25553417e+00 8.73431385e-01 -2.80256420e-01 -1.19863898e-01
4.05465722e-01 -4.24719304e-01 -4.16408211e-01 -1.27962828e-01
1.13668822e-01 3.84788156e-01 8.65870118e-01 -1.30701840e+00
-5.92490494e-01 -7.62740374e-01 2.24787563e-01 3.15360904e-01
-5.42936623e-01 2.18271092e-01 -1.21632838e+00 -6.90510869e-01
-1.11822143e-01 -9.00489688e-01 -3.44661117e-01 -1.21797398e-02
-3.70926917e-01 -5.53353786e-01 7.30689585e-01 -8.93172681e-01
1.29033303e+00 -1.98314250e+00 2.30914786e-01 3.59953433e-01
2.02110842e-01 5.16413152e-01 -1.97357133e-01 2.82854348e-01
5.69284037e-02 -2.26875991e-01 1.37487873e-01 -8.03942740e-01
1.98236942e-01 -1.12400927e-01 1.77603155e-01 3.86763155e-01
-1.77043136e-02 1.05550325e+00 -8.82211685e-01 -3.95757705e-01
2.87561834e-01 6.27035499e-01 -5.54955661e-01 2.20866904e-01
2.41564870e-01 4.91923332e-01 -5.48512697e-01 8.09438288e-01
7.06410170e-01 -3.92056197e-01 6.96173236e-02 -3.20256025e-01
1.66009679e-01 -8.76052827e-02 -1.34900212e+00 1.87746620e+00
-2.75090933e-01 3.82704943e-01 -1.04319282e-01 -9.08542335e-01
7.80880094e-01 -4.05675732e-02 3.02732378e-01 -6.75304592e-01
1.97597921e-01 4.49704845e-03 -1.31030843e-01 -4.95810807e-01
4.94362324e-01 5.92052758e-01 1.04866236e-01 2.57639647e-01
4.90853600e-02 5.19389391e-01 7.24055544e-02 1.72935918e-01
9.79344904e-01 1.13972872e-01 1.52413994e-01 7.10592195e-02
9.71450269e-01 -4.44801062e-01 5.67640722e-01 7.80928850e-01
-6.00214660e-01 4.71137255e-01 -1.74305588e-02 -7.81979501e-01
-7.59190321e-01 -7.67972112e-01 1.35422185e-01 1.12665141e+00
6.17518067e-01 -6.37871385e-01 -6.64471745e-01 -8.78138721e-01
1.99852660e-01 1.91634253e-01 -7.40479589e-01 -3.83968294e-01
-6.38329446e-01 -7.11516023e-01 2.86578149e-01 7.27438688e-01
1.02073252e+00 -9.66045439e-01 -1.56262890e-01 6.36432320e-02
-1.14857905e-01 -1.05745196e+00 -8.85546565e-01 -2.63537943e-01
-3.95303756e-01 -1.23738813e+00 -8.16822469e-01 -1.03861964e+00
7.81717241e-01 4.66453195e-01 8.69803667e-01 3.91659200e-01
-7.48613119e-01 8.43149185e-01 -3.86568934e-01 -1.41834229e-01
3.13169777e-01 4.56366539e-02 1.77047580e-01 3.38477999e-01
7.14560091e-01 -3.70590717e-01 -1.08391595e+00 4.40970182e-01
-5.14563441e-01 -3.55347604e-01 5.45820177e-01 1.12176609e+00
4.81009156e-01 -1.22380806e-02 2.34213084e-01 -2.15450808e-01
6.29048169e-01 -1.00640908e-01 -3.35484177e-01 6.66150928e-01
-6.12240851e-01 -1.15430325e-01 2.50996798e-01 -4.33207273e-01
-9.76492703e-01 1.72478050e-01 -8.93713832e-02 -4.46440101e-01
-2.33475327e-01 1.36749551e-01 -3.57526720e-01 -4.91419822e-01
3.54620337e-01 3.54342550e-01 -3.09665173e-01 -6.92696095e-01
2.15999931e-01 6.35635436e-01 5.60477912e-01 -5.53487599e-01
8.32074106e-01 3.34217072e-01 -1.58330187e-01 -5.63436687e-01
-7.08243310e-01 -8.39224041e-01 -7.68005371e-01 -1.72991380e-01
7.84252048e-01 -9.39691722e-01 -9.19694126e-01 5.52946508e-01
-1.01175201e+00 -6.46917745e-02 2.08576042e-02 3.97312105e-01
-4.94883023e-02 8.02445590e-01 -3.79841506e-01 -7.95949161e-01
-5.27716279e-01 -1.17678881e+00 1.24999809e+00 7.50252366e-01
1.79620788e-01 -9.16436970e-01 1.05435856e-01 4.84634042e-01
3.70256811e-01 -5.78331947e-02 2.68523484e-01 -6.99690521e-01
-6.35289848e-01 -6.20329797e-01 -6.01714253e-01 1.03596352e-01
2.19442844e-01 -5.32642484e-01 -1.00538719e+00 -8.24865222e-01
-3.64224583e-01 -3.03270996e-01 1.23528850e+00 1.89726248e-01
1.34858036e+00 3.55369353e-04 -7.37385631e-01 7.96223819e-01
1.18325877e+00 -8.14023614e-02 4.74639624e-01 4.59132969e-01
8.81818831e-01 3.39578420e-01 3.97283792e-01 4.23031628e-01
6.09229028e-01 1.09212017e+00 1.63036078e-01 -7.07999542e-02
-4.26319629e-01 -3.33156735e-01 -5.18044783e-03 3.08085740e-01
-2.84693390e-01 -9.05571133e-02 -7.63464570e-01 5.48109770e-01
-2.13930345e+00 -1.02744234e+00 3.66040111e-01 2.21242833e+00
5.41605949e-01 -1.24266684e-01 2.28813753e-01 -1.37927160e-01
8.91708195e-01 6.83173239e-02 -6.93682909e-01 1.92093238e-01
7.94194341e-02 2.69650351e-02 1.83707252e-01 4.27511275e-01
-1.37441599e+00 1.03549635e+00 5.88441849e+00 7.93392301e-01
-7.60191798e-01 7.15302974e-02 2.78282821e-01 -4.21459861e-02
-6.13509910e-03 -5.25734574e-02 -1.09470189e+00 4.99032438e-01
1.52000263e-01 1.34427741e-01 4.30897683e-01 9.36182618e-01
-1.28575981e-01 1.06126599e-01 -1.20236135e+00 1.59578514e+00
4.77885753e-01 -8.04846883e-01 1.05072791e-02 2.42565051e-02
4.59975511e-01 -3.59774858e-01 1.53251767e-01 5.10053277e-01
1.83759943e-01 -9.44205582e-01 1.37243852e-01 7.28686869e-01
4.26351845e-01 -6.82084501e-01 7.62491882e-01 2.40127929e-02
-1.55560970e+00 -3.99385810e-01 -4.61214095e-01 6.56574816e-02
2.61952654e-02 1.72633156e-01 -5.67561030e-01 6.63307190e-01
9.74511206e-01 7.98210800e-01 -1.05508041e+00 1.61717832e+00
-1.58958718e-01 1.15621202e-01 -2.68536985e-01 -2.54691858e-03
-4.13600430e-02 2.60763280e-02 4.82863814e-01 1.26563919e+00
-1.57617424e-02 -7.66066089e-02 6.97066903e-01 6.38290644e-01
-7.24448338e-02 1.07177667e-01 -4.06527042e-01 4.05713916e-01
3.84561300e-01 1.31056845e+00 -4.42738414e-01 -2.85638928e-01
-5.66172421e-01 1.47449636e+00 6.24843657e-01 5.36108375e-01
-8.08799088e-01 -3.25363100e-01 6.96216285e-01 -2.14223102e-01
1.73811734e-01 -1.44933179e-01 9.51975659e-02 -1.43619394e+00
2.34071255e-01 -7.93090999e-01 7.51597703e-01 -4.23864990e-01
-1.64136624e+00 3.65376204e-01 -1.73217043e-01 -9.77262735e-01
-1.14384119e-03 -6.25682056e-01 -7.60811627e-01 9.74933326e-01
-1.65420437e+00 -1.70061195e+00 -8.18997681e-01 9.63196158e-01
6.61651254e-01 -3.51154357e-01 6.94927335e-01 5.59354305e-01
-8.99532020e-01 1.25005198e+00 -1.35153458e-01 4.88036543e-01
8.78377259e-01 -1.02861452e+00 3.79506618e-01 9.66579556e-01
1.87669531e-01 1.08766997e+00 2.46189341e-01 -7.24164307e-01
-1.42830038e+00 -9.99987662e-01 7.51685262e-01 -4.24753338e-01
3.62112403e-01 -3.06818157e-01 -6.69335544e-01 5.67698061e-01
1.74886629e-01 2.52063483e-01 7.38004386e-01 4.82973129e-01
-5.14484346e-01 -2.15388581e-01 -9.11413431e-01 6.51550114e-01
1.55693150e+00 -6.49255753e-01 -5.85270405e-01 3.84177417e-01
3.92943412e-01 -2.10856929e-01 -5.58308125e-01 4.05041605e-01
7.14005291e-01 -6.78694487e-01 1.35717022e+00 -7.21221924e-01
-3.58987957e-01 -4.13201749e-01 -1.03468448e-01 -1.15350127e+00
-8.91421974e-01 -4.81084347e-01 -1.86992392e-01 1.30868232e+00
1.86372008e-02 -5.74833691e-01 8.98333728e-01 8.85865986e-01
8.58906060e-02 -5.53563654e-01 -7.79696524e-01 -7.77176261e-01
-3.87123674e-01 2.35451963e-02 6.55182362e-01 7.28880823e-01
-7.46894330e-02 3.44835937e-01 -5.20513356e-01 3.01240444e-01
9.09064114e-01 2.95694292e-01 8.97886455e-01 -1.27177358e+00
-4.83752966e-01 -5.10335982e-01 -6.71791375e-01 -1.21025479e+00
1.13373950e-01 -8.04274499e-01 -2.06810683e-01 -1.58993268e+00
7.12358296e-01 -4.25858647e-01 -6.62342489e-01 4.24559891e-01
-8.03494990e-01 2.51191348e-01 3.10163200e-01 3.14017028e-01
-9.86391008e-01 7.08118320e-01 1.01421225e+00 -5.36969662e-01
-2.78773934e-01 1.09710097e-01 -7.20773101e-01 5.36386847e-01
6.73842490e-01 1.11100599e-01 -2.78712988e-01 -6.63270772e-01
-1.85628220e-01 -5.38883448e-01 9.08934832e-01 -1.11861193e+00
5.74437320e-01 1.85153216e-01 9.60549057e-01 -5.50594807e-01
5.14368176e-01 -8.19534302e-01 -9.59481895e-02 3.88033032e-01
-2.66993135e-01 5.89763522e-02 7.74888173e-02 8.78385603e-01
-1.58837095e-01 -5.80091886e-02 5.02859473e-01 -2.73476839e-01
-1.03199327e+00 8.07795763e-01 4.04521793e-01 -2.61342347e-01
7.91432858e-01 -2.47486189e-01 -1.79928184e-01 -2.17248708e-01
-8.31132472e-01 4.98721391e-01 3.89393270e-01 5.99158347e-01
7.04174340e-01 -1.62402165e+00 -5.22038877e-01 2.31371760e-01
3.02848041e-01 -3.57737035e-01 4.42619145e-01 6.00687146e-01
-4.05157842e-02 3.00088227e-01 -8.95251483e-02 -4.81725842e-01
-1.47188950e+00 9.15061653e-01 4.90209967e-01 -3.55095595e-01
-5.09749234e-01 1.09905267e+00 2.41594568e-01 -3.23937029e-01
5.85160792e-01 2.01303437e-01 -5.34640551e-01 -8.21971968e-02
7.53390014e-01 2.63921797e-01 -2.78993756e-01 -7.67000973e-01
-5.77853739e-01 9.82722998e-01 -3.04628372e-01 1.90149978e-01
1.04047418e+00 -3.09128523e-01 7.10069761e-02 -2.23523617e-01
1.13564348e+00 -1.58664286e-01 -1.12809122e+00 -6.90062940e-01
-1.13289818e-01 -7.74853230e-01 -1.51804425e-02 -7.48356521e-01
-9.85298276e-01 8.74525845e-01 9.78170455e-01 -6.76159486e-02
1.09051716e+00 3.17053869e-02 8.59782755e-01 4.55943108e-01
3.65706444e-01 -1.26346171e+00 3.57359111e-01 3.76875818e-01
9.03764129e-01 -1.54470587e+00 1.05947450e-01 -4.19005990e-01
-2.53638029e-01 9.52352583e-01 9.77321386e-01 4.03082697e-04
6.54890060e-01 -4.25825089e-01 -2.97985345e-01 -1.29955113e-01
-2.01955959e-01 -7.32747853e-01 6.93653882e-01 6.99253261e-01
1.95244089e-01 -1.57309264e-01 -2.65091032e-01 6.97890580e-01
2.32061461e-01 -1.34890035e-01 -3.86241853e-01 7.74349809e-01
-4.35234010e-01 -1.30849099e+00 -3.79197091e-01 2.69634545e-01
-1.97412968e-01 -7.43244812e-02 -5.86859167e-01 6.02875113e-01
2.88060397e-01 9.99714017e-01 -1.80150747e-01 -6.36205912e-01
5.10277808e-01 9.32151526e-02 5.90977013e-01 -4.38905180e-01
-6.21478558e-01 -4.91563827e-02 -8.23684856e-02 -6.85477734e-01
-4.34468836e-01 -7.80541360e-01 -6.29646182e-01 -2.06722885e-01
-4.51666951e-01 -5.37837856e-02 3.54373276e-01 8.77789140e-01
5.13603330e-01 3.81888092e-01 5.14403760e-01 -8.30564737e-01
-4.62634891e-01 -1.05817056e+00 -3.42177093e-01 7.48459458e-01
2.03059182e-01 -9.54792798e-01 -5.50600700e-02 -8.62707198e-02] | [14.728042602539062, 0.8811573386192322] |
692ae69b-786b-4b4a-9f03-ceafb44ba754 | deep-simplex-classifier-for-maximizing-the | 2212.11747 | null | https://arxiv.org/abs/2212.11747v1 | https://arxiv.org/pdf/2212.11747v1.pdf | Deep Simplex Classifier for Maximizing the Margin in Both Euclidean and Angular Spaces | The classification loss functions used in deep neural network classifiers can be grouped into two categories based on maximizing the margin in either Euclidean or angular spaces. Euclidean distances between sample vectors are used during classification for the methods maximizing the margin in Euclidean spaces whereas the Cosine similarity distance is used during the testing stage for the methods maximizing margin in the angular spaces. This paper introduces a novel classification loss that maximizes the margin in both the Euclidean and angular spaces at the same time. This way, the Euclidean and Cosine distances will produce similar and consistent results and complement each other, which will in turn improve the accuracies. The proposed loss function enforces the samples of classes to cluster around the centers that represent them. The centers approximating classes are chosen from the boundary of a hypersphere, and the pairwise distances between class centers are always equivalent. This restriction corresponds to choosing centers from the vertices of a regular simplex. There is not any hyperparameter that must be set by the user in the proposed loss function, therefore the use of the proposed method is extremely easy for classical classification problems. Moreover, since the class samples are compactly clustered around their corresponding means, the proposed classifier is also very suitable for open set recognition problems where test samples can come from the unknown classes that are not seen in the training phase. Experimental studies show that the proposed method achieves the state-of-the-art accuracies on open set recognition despite its simplicity. | ['Hasan Saribas', 'Hakan Cevikalp'] | 2022-12-22 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [-8.92212540e-02 9.56428349e-02 -1.95302457e-01 -5.85594893e-01
-7.79000521e-02 -2.58250654e-01 3.51872057e-01 4.23777670e-01
-7.83984661e-01 8.48946571e-01 -6.55126870e-01 4.75374460e-02
-4.52625543e-01 -1.03427243e+00 -5.82828224e-01 -1.06375492e+00
4.40575331e-02 5.46866477e-01 1.41802743e-01 9.41113681e-02
1.42211780e-01 6.43420696e-01 -1.89477193e+00 -3.69143859e-02
1.24723709e+00 1.42304182e+00 6.00464605e-02 -8.80030096e-02
-3.42816889e-01 2.93683335e-02 -7.25482285e-01 -1.99967071e-01
4.72958356e-01 -2.31321514e-01 -3.15317124e-01 -4.61935811e-02
3.62356842e-01 9.94697064e-02 -1.33470707e-02 1.14033365e+00
3.13327640e-01 4.60816830e-01 1.01153457e+00 -1.36025596e+00
-4.28799212e-01 2.22773507e-01 -3.97784680e-01 -1.43849641e-01
-6.05139174e-02 -5.85411489e-01 9.14596021e-01 -7.75413990e-01
2.29625672e-01 9.14734304e-01 4.99797076e-01 3.95942539e-01
-9.88892794e-01 -7.07829773e-01 1.35966271e-01 4.78081882e-01
-1.76862633e+00 -9.96593684e-02 7.88323522e-01 -2.65469342e-01
5.12293935e-01 4.23875362e-01 7.32396364e-01 6.64629459e-01
2.51636896e-02 6.40364707e-01 7.79580653e-01 -5.15017569e-01
4.48137134e-01 6.07750773e-01 3.76943409e-01 4.79837626e-01
4.51221019e-01 -2.61965275e-01 2.58879721e-01 6.36911020e-02
5.20660281e-01 2.76150167e-01 -6.85456395e-01 -8.33533883e-01
-9.96383369e-01 1.01739049e+00 6.76762700e-01 4.74768877e-01
-2.23070998e-02 -4.98879611e-01 3.29218388e-01 1.55980945e-01
2.61413574e-01 2.50518531e-01 -1.62702218e-01 2.66753167e-01
-7.14523971e-01 6.35737330e-02 7.55357862e-01 6.96482480e-01
5.78958392e-01 -2.06651852e-01 2.24173829e-01 1.17228258e+00
3.49086910e-01 2.09812239e-01 6.24380350e-01 -3.46901864e-01
4.97522652e-01 7.63507366e-01 -2.09964048e-02 -1.16463959e+00
-3.09761316e-01 -5.71201205e-01 -9.20084655e-01 4.11502987e-01
7.04838872e-01 8.09050649e-02 -5.57886422e-01 1.60029960e+00
4.70465630e-01 4.56918553e-02 2.59259582e-01 9.15427148e-01
8.60858381e-01 7.47661054e-01 -4.88346189e-01 -4.89611566e-01
1.07532215e+00 -7.79895782e-01 -6.44945502e-01 3.26922461e-02
4.65539485e-01 -4.12498683e-01 9.96161938e-01 5.57287991e-01
-5.58991134e-01 -5.38387716e-01 -1.47242379e+00 3.56360644e-01
-6.75093293e-01 3.28731954e-01 1.51978552e-01 5.68860233e-01
-4.39490646e-01 7.67647088e-01 -6.39427781e-01 -5.21245524e-02
4.06990111e-01 4.37452525e-01 -4.03970510e-01 1.23597816e-01
-1.12584925e+00 7.45223463e-01 5.36870658e-01 4.79942530e-01
-5.99302314e-02 -1.94295213e-01 -7.56101727e-01 2.22891137e-01
-6.34879693e-02 3.55293117e-02 4.69750911e-01 -9.59771752e-01
-1.17511499e+00 7.43397355e-01 1.96774751e-01 -2.58270413e-01
6.58690512e-01 1.56167567e-01 -2.73791641e-01 -1.37279525e-01
-1.44433469e-01 4.44369823e-01 6.83833480e-01 -1.06268191e+00
-5.32727063e-01 -5.19414604e-01 -5.61622307e-02 3.27546269e-01
-9.33167994e-01 -3.69038135e-01 -2.25227460e-01 -4.74950969e-01
4.80062187e-01 -7.79710829e-01 2.19256103e-01 2.20080927e-01
-4.11522299e-01 -5.75894713e-01 1.03355706e+00 -2.56897718e-01
1.02304995e+00 -2.14515996e+00 2.58051325e-02 5.61940670e-01
2.27357000e-02 3.91822726e-01 2.15748772e-01 1.07298500e-03
-3.19256574e-01 -1.89451694e-01 -3.72070462e-01 -6.21609166e-02
-7.40079656e-02 2.09421232e-01 -1.06669866e-01 7.29718387e-01
-1.51453093e-01 9.77675840e-02 -6.43693089e-01 -4.29936796e-01
2.67796218e-01 5.45662820e-01 -2.23789155e-01 5.40378615e-02
-3.63798589e-02 1.98431145e-02 -3.00818205e-01 8.09220150e-02
8.83664668e-01 1.54677117e-02 -9.89685208e-02 -1.43963352e-01
5.73070757e-02 1.03379390e-03 -1.73237205e+00 1.05953407e+00
-4.40259337e-01 6.49880171e-01 -7.54088834e-02 -1.42979980e+00
1.50867248e+00 2.27738589e-01 3.75522494e-01 -3.12619090e-01
2.91127682e-01 3.33050102e-01 1.73401579e-01 -1.43229380e-01
4.04761620e-02 -1.62172299e-02 4.26377386e-01 2.17597157e-01
-2.28804380e-01 8.75207856e-02 3.03809732e-01 -6.01076424e-01
2.95034140e-01 -2.63236642e-01 2.64104903e-01 -2.28719100e-01
8.46130669e-01 -4.74401712e-01 8.65755320e-01 3.47739846e-01
-8.89116004e-02 6.49349213e-01 4.59537327e-01 -5.86255312e-01
-7.61053979e-01 -9.41600382e-01 -1.05377758e+00 2.92935729e-01
3.62770259e-01 1.19363505e-03 -8.02175760e-01 -8.91159296e-01
-1.80590246e-02 5.66329658e-01 -6.49306476e-01 -2.15133995e-01
-3.72134209e-01 -6.63245797e-01 2.17780441e-01 4.89940941e-01
6.14197433e-01 -8.10332954e-01 -5.46646953e-01 -1.71147779e-01
-1.77805126e-01 -6.58069015e-01 -2.13893384e-01 3.13386530e-01
-9.79751050e-01 -1.30994773e+00 -8.70336771e-01 -9.64294195e-01
1.08229947e+00 4.77030985e-02 3.73722494e-01 -9.00722072e-02
-1.31616518e-01 -3.31039548e-01 -4.77730542e-01 -4.49500114e-01
-1.63003579e-01 -1.24690183e-01 3.13696742e-01 4.69250411e-01
5.92255056e-01 -3.70405078e-01 -4.11185294e-01 7.11478472e-01
-6.48324192e-01 -3.53781790e-01 4.40527946e-02 1.06645179e+00
6.61301255e-01 2.86057323e-01 6.09989703e-01 -3.39194089e-01
4.79209512e-01 -3.86310488e-01 -7.63853848e-01 3.80130976e-01
-6.68057263e-01 -3.20293270e-02 1.10251999e+00 -6.39076531e-01
-5.13351798e-01 -8.32314938e-02 6.30070791e-02 -5.35185635e-01
-1.88065082e-01 2.52317846e-01 -4.16131824e-01 -2.60769665e-01
6.31065190e-01 1.71048433e-01 1.71305239e-01 -3.29877764e-01
-2.97151525e-02 9.84836876e-01 -2.78903246e-02 -3.13393623e-01
4.33934569e-01 3.60792279e-01 1.29282728e-01 -1.04005492e+00
-4.71619308e-01 -2.62667000e-01 -6.26238346e-01 -1.84351608e-01
5.81911743e-01 -3.27073157e-01 -7.69067645e-01 4.57269341e-01
-9.49738204e-01 3.49272013e-01 -4.25175369e-01 7.87222922e-01
-3.74258399e-01 4.82128501e-01 -8.69805440e-02 -7.90916502e-01
-2.72299111e-01 -1.41863132e+00 5.55187404e-01 4.79048371e-01
-3.88402343e-02 -1.03930748e+00 -2.74601251e-01 5.88144399e-02
-7.14733591e-03 3.17604840e-01 1.12042904e+00 -1.29837072e+00
-1.43359914e-01 -7.26205111e-01 -1.73628684e-02 8.01989198e-01
2.74437666e-01 5.87288402e-02 -9.04377043e-01 -4.09170449e-01
1.61879569e-01 -2.49546245e-01 7.18619585e-01 3.72945189e-01
1.32284284e+00 -1.95776075e-01 -4.00110722e-01 7.18083262e-01
1.19146252e+00 6.29225373e-01 5.42347610e-01 5.17766953e-01
3.54049474e-01 7.12116301e-01 6.75042391e-01 3.16071957e-01
-2.38107860e-01 6.14044487e-01 4.90878701e-01 9.74985212e-03
4.53140616e-01 1.63839117e-01 -5.46494834e-02 6.35029912e-01
8.21456164e-02 -2.48948172e-01 -7.25386798e-01 3.25314760e-01
-1.72954023e+00 -9.34357882e-01 -1.49483025e-01 2.89084172e+00
7.57688284e-01 2.42995262e-01 -1.01783037e-01 7.54420996e-01
1.01491904e+00 7.18784928e-02 -5.19710898e-01 -4.92254972e-01
2.72632279e-02 7.42432922e-02 1.68117777e-01 3.88746470e-01
-1.24560094e+00 2.03779146e-01 4.95935869e+00 1.12882292e+00
-1.18337166e+00 -3.28657538e-01 5.64434052e-01 -2.49872521e-01
-1.26257800e-02 -2.11051166e-01 -1.01513028e+00 5.67342103e-01
2.40676448e-01 1.89564273e-01 1.99657559e-01 1.04286134e+00
5.82352914e-02 -3.81036587e-02 -1.30809617e+00 1.13343132e+00
9.32482705e-02 -1.00373220e+00 3.81135760e-04 4.34581675e-02
5.42466700e-01 -2.22414270e-01 6.49675503e-02 1.21304400e-01
-5.10572731e-01 -8.92121971e-01 5.90053022e-01 4.59874153e-01
6.45941556e-01 -1.10238862e+00 9.21401680e-01 6.72999144e-01
-8.97962511e-01 -2.17105255e-01 -7.39293277e-01 1.03948019e-01
-3.51098627e-01 7.40075707e-01 -7.33138144e-01 2.80209690e-01
4.66564417e-01 5.79622090e-01 -2.91004628e-01 1.37379527e+00
3.91421542e-02 1.14018038e-01 -6.78292453e-01 -4.34592724e-01
1.32790238e-01 -9.07271385e-01 5.33287525e-01 6.88037932e-01
2.70592660e-01 -2.51616865e-01 1.94526732e-01 8.37404013e-01
4.05793227e-02 5.02783537e-01 -6.84919357e-01 2.79590875e-01
6.30421638e-01 9.75634336e-01 -8.39992344e-01 -5.00104167e-02
-3.41535568e-01 5.78583658e-01 4.00142312e-01 2.26739809e-01
-8.68501246e-01 -1.15769172e+00 5.87918639e-01 9.57675204e-02
3.28741074e-01 1.40998468e-01 -2.90492058e-01 -9.51192439e-01
5.16313851e-01 -5.65790296e-01 3.98910969e-01 -3.32382351e-01
-1.15328848e+00 7.94183910e-01 1.03089057e-01 -1.51745927e+00
1.79391000e-02 -8.39969695e-01 -7.51319587e-01 7.23802626e-01
-1.21536267e+00 -3.98163706e-01 -4.16386336e-01 5.58709323e-01
1.57891244e-01 -3.06866318e-01 8.98315609e-01 3.15127075e-01
-7.26900339e-01 8.58604372e-01 8.80326748e-01 3.91989321e-01
5.17849505e-01 -1.07473361e+00 -3.06567907e-01 5.00061929e-01
2.07350537e-01 6.68867528e-01 4.20428932e-01 -2.19508067e-01
-6.18973076e-01 -8.79442751e-01 7.12676108e-01 1.46211892e-01
1.48249537e-01 -4.74846810e-01 -1.05135584e+00 2.04920530e-01
-2.41607234e-01 4.11353976e-01 6.59796536e-01 7.85625204e-02
-2.58889467e-01 -5.74454963e-01 -1.32148862e+00 3.75914514e-01
4.38953876e-01 -1.36343107e-01 -5.29791892e-01 5.66716313e-01
1.99299991e-01 -1.34943172e-01 -9.14933026e-01 6.21871054e-01
5.89010358e-01 -9.51706707e-01 7.67746687e-01 -4.90232646e-01
1.21655576e-01 -3.85729641e-01 -1.35209352e-01 -1.39921010e+00
1.45570666e-01 4.68651466e-02 1.70290038e-01 1.02972460e+00
3.52185905e-01 -1.07229054e+00 7.84549356e-01 2.30671659e-01
6.64949045e-02 -1.41096997e+00 -1.19399917e+00 -9.51825321e-01
1.18971884e-01 -1.03108995e-01 3.42193812e-01 1.00314724e+00
1.03312708e-01 7.25993216e-02 2.57334828e-01 -1.92763284e-02
6.90247476e-01 2.68831789e-01 3.26513976e-01 -1.62123120e+00
-1.20969757e-01 -5.29227138e-01 -7.79769957e-01 -8.04454744e-01
2.80752420e-01 -1.01521087e+00 1.82142910e-02 -1.40845680e+00
-1.78516716e-01 -1.06861877e+00 -1.68360963e-01 3.65446866e-01
2.15336800e-01 -3.16752642e-02 -1.59819331e-02 2.70413488e-01
-9.26663727e-02 9.32020247e-01 1.16511250e+00 -1.48598015e-01
-4.40359533e-01 4.63281840e-01 -2.12357298e-01 9.65457797e-01
6.75941408e-01 -3.40406060e-01 -5.65487325e-01 -1.18459105e-01
-1.35496214e-01 -1.46393418e-01 -1.84897799e-02 -1.12810957e+00
-3.04963328e-02 -2.80541442e-02 4.86121863e-01 -3.97873074e-01
5.27003944e-01 -1.12726176e+00 -1.97119713e-01 3.23248416e-01
-2.81535000e-01 -3.86597127e-01 4.05442752e-02 5.36041796e-01
-3.04133266e-01 -7.90416539e-01 1.28210723e+00 3.01377892e-01
-1.01218261e-01 2.66042113e-01 8.32948759e-02 2.69356780e-02
1.49831736e+00 -7.53234565e-01 -8.78494903e-02 -2.30093747e-01
-7.94650912e-01 1.61451191e-01 3.31245273e-01 3.59635532e-01
6.88972175e-01 -1.46711087e+00 -4.63583976e-01 5.30730844e-01
2.83069968e-01 2.17328146e-01 -1.84175625e-01 7.76609063e-01
-6.18004262e-01 4.37401086e-01 -2.11847231e-01 -7.49455988e-01
-1.27516031e+00 5.05897999e-01 7.28047431e-01 8.13096464e-02
-5.65046489e-01 8.08089077e-01 1.01051725e-01 -7.00914204e-01
7.10861504e-01 -3.27348709e-01 -6.04854584e-01 2.08580479e-01
6.01625025e-01 4.54011410e-01 3.40496808e-01 -5.45575202e-01
-2.50055104e-01 5.06657124e-01 -1.60587177e-01 3.39788139e-01
1.18395829e+00 3.70048672e-01 -1.61790162e-01 6.12797379e-01
1.62203896e+00 -2.44786471e-01 -9.33793604e-01 -1.64599314e-01
-4.43136811e-01 -5.06971300e-01 -5.90598583e-02 -3.21634889e-01
-9.77109194e-01 1.00048804e+00 8.32554817e-01 3.00228477e-01
9.87642109e-01 -3.56379360e-01 5.63964546e-01 4.87847626e-01
3.00133139e-01 -1.17550135e+00 -3.41803998e-01 4.87778634e-01
8.50280285e-01 -9.81003225e-01 -1.28880680e-01 -3.19471896e-01
-2.84810156e-01 1.42429054e+00 7.55139947e-01 -2.69316107e-01
7.77018130e-01 -2.16019507e-02 -2.47565255e-01 3.46655756e-01
-2.45816275e-01 2.53943890e-01 3.69658411e-01 5.47056377e-01
3.52102190e-01 2.68950593e-02 -7.28171945e-01 5.95459700e-01
-8.65976512e-02 -5.46463132e-01 3.39376360e-01 4.95577604e-01
-6.73157394e-01 -9.31627631e-01 -6.70806766e-01 5.67499995e-01
2.63494030e-02 2.62424290e-01 -1.54372916e-01 8.50763083e-01
4.01650310e-01 8.15319657e-01 4.30115759e-01 -2.89606303e-01
2.90938884e-01 1.33265808e-01 4.65656698e-01 -2.26127177e-01
-7.01214671e-02 -2.75593698e-01 -1.97146371e-01 -1.56177193e-01
5.61903641e-02 -5.95029593e-01 -1.58587754e+00 -2.47504823e-02
-8.88219297e-01 6.37331128e-01 9.01247561e-01 9.66331542e-01
-1.25237927e-02 1.35600880e-01 1.04716635e+00 -4.73616153e-01
-9.31546390e-01 -9.24562931e-01 -6.98250294e-01 3.01057816e-01
1.07619397e-01 -9.97431755e-01 -6.24423981e-01 -4.04765844e-01] | [8.527623176574707, 3.8928706645965576] |
fdd987a4-5d9f-4829-9069-2623035af53c | temporal-aggregate-representations-for-long | 2006.00830 | null | https://arxiv.org/abs/2006.00830v2 | https://arxiv.org/pdf/2006.00830v2.pdf | Temporal Aggregate Representations for Long-Range Video Understanding | Future prediction, especially in long-range videos, requires reasoning from current and past observations. In this work, we address questions of temporal extent, scaling, and level of semantic abstraction with a flexible multi-granular temporal aggregation framework. We show that it is possible to achieve state of the art in both next action and dense anticipation with simple techniques such as max-pooling and attention. To demonstrate the anticipation capabilities of our model, we conduct experiments on Breakfast, 50Salads, and EPIC-Kitchens datasets, where we achieve state-of-the-art results. With minimal modifications, our model can also be extended for video segmentation and action recognition. | ['Angela Yao', 'Fadime Sener', 'Dipika Singhania'] | 2020-06-01 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2515_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123610154.pdf | eccv-2020-8 | ['action-anticipation'] | ['computer-vision'] | [ 1.79569393e-01 4.73267511e-02 -3.66097957e-01 -5.84792912e-01
-4.82761592e-01 -3.56915623e-01 5.84764063e-01 -8.39354172e-02
-2.67795473e-01 5.56152880e-01 6.14258587e-01 1.10733256e-01
-1.96954817e-01 -5.41520178e-01 -6.90387249e-01 -3.84055257e-01
-6.23936355e-01 5.39468192e-02 5.81196368e-01 -1.14318967e-01
1.63372084e-02 3.49709123e-01 -1.61577296e+00 9.88015771e-01
4.85819906e-01 1.20756876e+00 3.95819545e-02 8.16512346e-01
3.99855822e-01 1.17436850e+00 -2.06096247e-01 -3.32559466e-01
3.72884750e-01 -2.68591881e-01 -1.02264154e+00 3.43923599e-01
3.82243454e-01 -7.17094719e-01 -6.87070489e-01 4.10994649e-01
1.20496690e-01 5.53406358e-01 1.11502469e-01 -1.25189233e+00
-5.87715864e-01 6.69846475e-01 -3.14050227e-01 4.81125385e-01
5.18114090e-01 5.17158568e-01 9.85691905e-01 -3.35619897e-01
8.12232733e-01 1.37285626e+00 7.26885259e-01 4.79054660e-01
-9.75460947e-01 -3.86971831e-01 7.38893092e-01 6.06205106e-01
-8.41560781e-01 -4.15014505e-01 4.86042947e-01 -3.79469126e-01
1.50645113e+00 -1.59770995e-02 1.06350112e+00 1.27550280e+00
2.84438908e-01 1.39453447e+00 9.08212721e-01 3.17586139e-02
2.44651198e-01 -6.87037230e-01 1.31015092e-01 7.01282740e-01
-4.61301565e-01 9.92318317e-02 -7.85590231e-01 2.47924015e-01
6.28299177e-01 2.50313938e-01 2.81530600e-02 -2.13201433e-01
-1.45348728e+00 5.83955348e-01 3.59928966e-01 5.33626154e-02
-4.96178478e-01 6.77068293e-01 4.85294998e-01 1.48312524e-01
6.25823677e-01 4.00218487e-01 -6.11085355e-01 -6.13141775e-01
-1.06983709e+00 3.28335762e-01 6.21315181e-01 8.99153113e-01
3.74737769e-01 -2.57943392e-01 -4.36084598e-01 1.52774721e-01
-9.57125649e-02 -2.72818319e-02 4.17331994e-01 -1.66194975e+00
4.27654296e-01 1.59115046e-01 3.23838919e-01 -6.82690263e-01
-5.59322000e-01 -9.99263451e-02 -4.29119140e-01 6.27536178e-02
3.88830751e-01 -3.21052998e-01 -1.09752631e+00 1.59697270e+00
1.69545889e-01 7.58554280e-01 8.29568505e-02 8.83940995e-01
2.37283319e-01 5.82717359e-01 4.13399488e-01 -2.59554207e-01
9.27594721e-01 -1.36156797e+00 -5.44070423e-01 -4.86772090e-01
6.78580642e-01 -1.94089174e-01 8.03496182e-01 3.09222758e-01
-1.20583141e+00 -8.13398838e-01 -7.60323524e-01 -2.02195942e-01
-1.64416254e-01 3.58480169e-03 1.24263442e+00 9.40574259e-02
-1.04317808e+00 1.17348838e+00 -1.58128607e+00 -6.65014088e-01
6.16907120e-01 2.79396862e-01 -2.75421470e-01 3.15477997e-02
-1.08098686e+00 6.87013984e-01 4.38889444e-01 4.81095128e-02
-1.00395167e+00 -7.24439263e-01 -7.75604486e-01 2.33083233e-01
5.13669670e-01 -6.88435316e-01 1.33918512e+00 -8.96037579e-01
-1.42771220e+00 4.41404611e-01 -1.47839069e-01 -1.04749906e+00
5.10411382e-01 -6.23112321e-01 -3.23199481e-01 4.11487579e-01
4.97020707e-02 1.19754374e+00 4.11595076e-01 -3.18545491e-01
-7.88955092e-01 -3.72385889e-01 4.58882153e-01 2.63637543e-01
-2.07094058e-01 1.19321316e-01 -3.85003209e-01 -5.28319359e-01
5.97149320e-02 -1.11811888e+00 -4.29914564e-01 2.10455596e-01
8.39058235e-02 -2.19058186e-01 6.79934263e-01 -6.81784809e-01
1.08874702e+00 -2.30276966e+00 1.41290382e-01 -4.22777474e-01
-1.80813596e-01 -5.86958649e-03 -2.70547390e-01 3.86541367e-01
6.60128444e-02 9.56063867e-02 1.00375995e-01 -4.36996967e-01
1.93320543e-01 3.21638882e-01 -3.25316936e-01 3.07443023e-01
4.49324042e-01 1.13114965e+00 -1.03959548e+00 -4.50870901e-01
3.21789742e-01 3.38175237e-01 -7.34551013e-01 -6.41175732e-02
-6.36297703e-01 4.81360197e-01 -6.74960792e-01 7.19980240e-01
7.93298781e-02 -5.15663445e-01 2.76290387e-01 4.89121675e-02
-4.11776230e-02 3.66778105e-01 -9.34022069e-01 2.13968325e+00
-2.28740126e-01 8.62312376e-01 -3.77734661e-01 -8.04186404e-01
4.10354853e-01 2.57673025e-01 6.69121981e-01 -7.17089355e-01
-2.84546673e-01 -2.95058697e-01 -1.79766759e-01 -6.56580627e-01
6.77605867e-01 6.46143258e-02 -9.67290476e-02 3.26282889e-01
-6.51721880e-02 2.14329079e-01 4.78547961e-01 1.28101692e-01
1.26143646e+00 8.31847012e-01 9.44144204e-02 3.93943042e-02
-1.07487917e-01 1.20565839e-01 6.43246651e-01 7.57659853e-01
-4.91003990e-01 5.20273030e-01 6.87833548e-01 -7.59697914e-01
-9.62765157e-01 -9.90926445e-01 1.79239184e-01 1.47032917e+00
4.12541218e-02 -4.66156363e-01 -4.29376602e-01 -6.94728851e-01
5.59859164e-03 6.58350527e-01 -7.46960998e-01 1.31676883e-01
-8.62373769e-01 -5.73902965e-01 2.98433214e-01 1.26212132e+00
5.97153485e-01 -1.16380000e+00 -1.11536312e+00 3.62317115e-01
-2.37140343e-01 -1.56998765e+00 -3.34653914e-01 -1.65429547e-01
-1.02619350e+00 -8.24493587e-01 -4.13048983e-01 -2.71280885e-01
-4.59166281e-02 8.65702257e-02 1.16012633e+00 -1.67239487e-01
-4.86891195e-02 4.69544649e-01 -4.60453600e-01 -3.76456305e-02
1.56996459e-01 -8.27785060e-02 -1.32658154e-01 -1.46319836e-01
2.47060925e-01 -6.92668676e-01 -9.13689137e-01 2.01644108e-01
-5.60996056e-01 3.42799217e-01 3.68629396e-01 4.99108315e-01
6.16658866e-01 9.40007623e-03 2.97114670e-01 -4.73984361e-01
1.32717952e-01 -4.40257013e-01 -4.39836830e-01 3.13748449e-01
-1.93771437e-01 -4.88791279e-02 3.09319705e-01 -4.91250843e-01
-1.27014863e+00 2.86402792e-01 -4.03202977e-03 -5.07247627e-01
-3.74447674e-01 4.57779318e-01 2.30537221e-01 3.36103708e-01
3.42842400e-01 -3.46770249e-02 -3.04475278e-01 -4.13163066e-01
6.06265008e-01 2.27562994e-01 6.05867863e-01 -4.29501742e-01
-3.06612439e-02 9.45790648e-01 3.29087935e-02 -4.14766788e-01
-1.19711757e+00 -2.39171341e-01 -7.78210223e-01 -1.90024301e-01
1.28286827e+00 -1.23523676e+00 -8.61770868e-01 4.71034676e-01
-9.46159542e-01 -9.74209905e-01 -4.38281357e-01 6.69256270e-01
-8.69051099e-01 3.46528679e-01 -1.03213000e+00 -8.14148188e-01
4.31973711e-02 -9.46843266e-01 1.17965353e+00 1.64916053e-01
-3.90093207e-01 -8.12786281e-01 -6.71122000e-02 5.51338315e-01
7.38992020e-02 4.67409492e-01 3.85267645e-01 -4.39086080e-01
-1.08257961e+00 7.20088333e-02 -1.25860572e-01 1.56880453e-01
-2.94494569e-01 -3.05289365e-02 -6.80399060e-01 -1.30462423e-01
-3.00617486e-01 -5.48916817e-01 1.31120944e+00 5.86533487e-01
1.46773100e+00 -2.21006528e-01 -3.78672004e-01 7.15490222e-01
1.10249746e+00 2.72975117e-01 8.38914096e-01 3.30922693e-01
4.89115179e-01 5.25900900e-01 1.10559952e+00 7.74013042e-01
6.23817682e-01 5.91722071e-01 3.62821519e-01 2.85897315e-01
-1.28691494e-01 -2.28130236e-01 3.97317678e-01 2.02226534e-01
-2.20712781e-01 -5.34104705e-01 -8.47720265e-01 8.39306951e-01
-2.35664058e+00 -1.34728169e+00 1.79767340e-01 1.59616947e+00
4.36048865e-01 2.93614715e-01 2.10887223e-01 -1.75389245e-01
3.63311827e-01 5.69515526e-01 -6.06727302e-01 -4.24771726e-01
9.42783803e-02 1.52589634e-01 4.13958281e-01 2.89048314e-01
-1.42117417e+00 1.21155083e+00 7.54231071e+00 5.82981884e-01
-8.57486188e-01 5.29277734e-02 8.84590387e-01 -6.74975276e-01
4.89872731e-02 1.42763153e-01 -6.89095080e-01 4.57386732e-01
9.75285530e-01 2.44381040e-01 4.99276042e-01 7.38731921e-01
2.95365423e-01 -1.83890343e-01 -1.41974854e+00 6.26702309e-01
-7.29828477e-02 -1.51413035e+00 -4.10724998e-01 -2.36747414e-01
9.76944745e-01 2.93367594e-01 1.88221072e-03 3.85437697e-01
4.77766871e-01 -8.23127508e-01 6.57361031e-01 7.07065523e-01
3.43378395e-01 -4.93911058e-01 4.09678251e-01 8.05391371e-02
-1.22768831e+00 -6.04097843e-01 -9.47167352e-03 -6.55229807e-01
5.92829764e-01 2.90413409e-01 -4.04541045e-01 4.11740303e-01
9.28498209e-01 1.29720902e+00 -4.64857370e-01 8.22861731e-01
-2.32549533e-01 3.84649485e-01 -4.06129837e-01 1.97749492e-02
4.75233972e-01 1.59621775e-01 1.84115067e-01 1.23093987e+00
4.55491066e-01 4.63984281e-01 4.93323445e-01 4.21290010e-01
1.92464620e-01 -4.77504581e-01 -2.87074387e-01 -2.87954479e-01
5.52952103e-02 7.77695477e-01 -7.99616516e-01 -6.84053302e-01
-4.22126859e-01 1.10956287e+00 3.14544111e-01 4.61542398e-01
-1.17428911e+00 1.61538929e-01 9.89641905e-01 2.28466745e-02
8.56531918e-01 -4.86724347e-01 -2.54649818e-01 -1.27199388e+00
4.66440506e-02 -5.28215170e-01 6.46325767e-01 -1.15571606e+00
-9.63763416e-01 2.52083004e-01 2.14482963e-01 -9.43293154e-01
-6.13334715e-01 -4.43522394e-01 -5.51326036e-01 2.26953954e-01
-1.19008660e+00 -1.23841488e+00 -2.57636845e-01 4.43553329e-01
8.56518686e-01 2.83878624e-01 4.22948480e-01 9.64432210e-02
-3.63323629e-01 3.47136855e-02 -2.18041226e-01 2.19846517e-02
3.38775784e-01 -1.00915420e+00 7.78592050e-01 1.02200770e+00
2.01626867e-01 -9.13954689e-04 7.90550351e-01 -7.11298168e-01
-1.25061715e+00 -1.09287143e+00 8.21922719e-01 -4.54484403e-01
1.01305974e+00 -1.76616713e-01 -6.70082688e-01 1.27827847e+00
1.69923007e-01 2.22886682e-01 4.82385427e-01 3.56514513e-01
-1.90113708e-01 -1.43997436e-02 -8.89437437e-01 6.97430074e-01
1.60339296e+00 -4.37225759e-01 -5.75144470e-01 5.57056129e-01
1.03805470e+00 -4.75713372e-01 -1.06790721e+00 4.59009290e-01
9.33467507e-01 -1.30027819e+00 1.03828025e+00 -8.94475162e-01
6.51091576e-01 4.92266826e-02 -1.48705065e-01 -9.15219903e-01
-4.72658634e-01 -7.47665524e-01 -4.77317065e-01 8.30891371e-01
1.21778645e-01 -2.09819704e-01 1.02622044e+00 8.70055616e-01
-3.28469127e-01 -9.48655307e-01 -8.95400763e-01 -8.63654435e-01
-2.30475292e-01 -7.19871461e-01 6.90361559e-01 5.81335425e-01
1.99076191e-01 -8.62540975e-02 -5.87357104e-01 1.36670843e-01
2.45954871e-01 4.24992234e-01 5.06429017e-01 -6.41941428e-01
-6.50645733e-01 -2.72989273e-01 -4.84252483e-01 -1.43067694e+00
2.60445893e-01 -3.23676795e-01 -1.35307550e-01 -1.67927849e+00
1.45998344e-01 9.32904184e-02 -5.34672499e-01 6.39948726e-01
1.06020823e-01 2.47662917e-01 5.38174331e-01 -8.25628042e-02
-1.45176721e+00 5.76996744e-01 1.37425876e+00 3.78000103e-02
-2.09229901e-01 -2.27424443e-01 -2.81974167e-01 8.22013319e-01
7.73731887e-01 -2.02129379e-01 -4.46114153e-01 -7.66228795e-01
1.00949593e-01 5.00800669e-01 5.98286271e-01 -1.23594010e+00
1.05605125e-01 -4.23344940e-01 5.16383231e-01 -6.69113040e-01
7.76076913e-01 -4.05175716e-01 2.01863825e-01 4.38180894e-01
-6.43359423e-01 4.48673889e-02 2.13202983e-01 7.54986405e-01
-1.17918141e-01 3.52443188e-01 4.32548970e-01 -3.33751947e-01
-1.29657745e+00 4.38747317e-01 -3.33045810e-01 -1.50342444e-02
1.13985777e+00 -2.49254942e-01 -4.03197229e-01 -3.44546348e-01
-1.19108117e+00 6.25733674e-01 3.48634571e-01 5.13332009e-01
4.81066108e-01 -1.21270669e+00 -5.76795757e-01 -1.27876937e-01
-1.13806263e-01 -3.93973917e-01 6.89444423e-01 1.05773973e+00
-3.13396931e-01 4.23511535e-01 -4.01364148e-01 -5.57399035e-01
-1.06216228e+00 8.13251317e-01 4.40301038e-02 -4.64368522e-01
-8.10789406e-01 6.80763245e-01 1.12950332e-01 2.13396028e-01
2.35677391e-01 -7.53623128e-01 -9.15852655e-03 2.89588030e-02
4.49165195e-01 4.84163195e-01 -5.02540529e-01 -2.38236949e-01
-4.69922096e-01 2.76472121e-01 2.24201262e-01 -6.28015846e-02
1.40368903e+00 -2.88520366e-01 2.74316609e-01 3.57770234e-01
9.88559306e-01 -6.18474782e-01 -2.16839194e+00 2.33382329e-01
-1.16595984e-01 -5.43085456e-01 -1.83457747e-01 -7.30923355e-01
-9.63918269e-01 8.67671967e-01 5.59879184e-01 2.39701897e-01
1.35138893e+00 2.48093337e-01 1.10710609e+00 4.74423319e-01
3.16845208e-01 -1.25001717e+00 1.70719653e-01 3.94473314e-01
6.33870244e-01 -1.36054814e+00 1.55975735e-02 -1.21812411e-01
-8.89270127e-01 8.30530584e-01 6.54019952e-01 -3.04078162e-01
6.19954228e-01 1.36909708e-01 -4.30921823e-01 -1.01932384e-01
-1.36556304e+00 -3.86078507e-01 2.55749851e-01 3.92937481e-01
3.51706326e-01 1.41424611e-01 -1.23698749e-01 4.18947190e-01
8.76859576e-02 4.51509595e-01 4.39081758e-01 8.86871874e-01
-3.52290303e-01 -6.77320600e-01 2.27354035e-01 3.62506539e-01
-6.14576221e-01 -8.47851783e-02 -3.55381262e-03 8.32489192e-01
2.33415842e-01 9.19836342e-01 4.75894272e-01 -4.32531804e-01
1.12515114e-01 8.33130330e-02 6.50771737e-01 -4.06627059e-01
-4.48389411e-01 1.61249377e-02 4.57693934e-01 -1.33129227e+00
-6.99416757e-01 -1.04622507e+00 -1.37702906e+00 -1.88549206e-01
1.17266364e-01 -3.83623451e-01 1.91298217e-01 1.04742706e+00
7.13749290e-01 5.97693264e-01 1.61036745e-01 -8.81045401e-01
-3.42108548e-01 -8.38802278e-01 -3.49017560e-01 5.69890201e-01
1.52163774e-01 -5.14843106e-01 -1.20123252e-01 2.58877277e-01] | [8.189238548278809, 0.5144723653793335] |
a1cbb5b2-fa63-493e-99b9-084da12fbbe5 | deep-relation-learning-for-regression-and-its | 2204.06598 | null | https://arxiv.org/abs/2204.06598v1 | https://arxiv.org/pdf/2204.06598v1.pdf | Deep Relation Learning for Regression and Its Application to Brain Age Estimation | Most deep learning models for temporal regression directly output the estimation based on single input images, ignoring the relationships between different images. In this paper, we propose deep relation learning for regression, aiming to learn different relations between a pair of input images. Four non-linear relations are considered: "cumulative relation", "relative relation", "maximal relation" and "minimal relation". These four relations are learned simultaneously from one deep neural network which has two parts: feature extraction and relation regression. We use an efficient convolutional neural network to extract deep features from the pair of input images and apply a Transformer for relation learning. The proposed method is evaluated on a merged dataset with 6,049 subjects with ages of 0-97 years using 5-fold cross-validation for the task of brain age estimation. The experimental results have shown that the proposed method achieved a mean absolute error (MAE) of 2.38 years, which is lower than the MAEs of 8 other state-of-the-art algorithms with statistical significance (p$<$0.05) in paired T-test (two-side). | ['Yangming Ou', 'P. Ellen Grant', 'Yanfang Feng', 'Sheng He'] | 2022-04-13 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [ 5.02933599e-02 1.66848674e-01 -8.05898085e-02 -7.35313356e-01
-2.42551878e-01 1.93233177e-01 4.88377124e-01 4.30019975e-01
-8.42164040e-01 8.25030327e-01 -3.22159141e-01 -5.59877828e-02
-4.79857415e-01 -7.32251704e-01 -6.34494424e-01 -6.12474561e-01
-6.08942986e-01 1.91372871e-01 9.65096727e-02 -1.23734605e-02
1.93388745e-01 2.92644918e-01 -1.58375430e+00 5.09794541e-02
9.00857627e-01 1.33661914e+00 -3.31710815e-01 5.25695741e-01
3.10747325e-01 8.93497825e-01 -5.15098393e-01 -8.17293465e-01
4.60345950e-03 -3.24717730e-01 -7.64236629e-01 -4.11761642e-01
5.91851592e-01 -4.37354743e-01 -4.16939676e-01 8.78432155e-01
5.75428545e-01 5.47414199e-02 1.02722836e+00 -1.28502178e+00
-7.77209282e-01 5.28416395e-01 -8.32417846e-01 6.23498619e-01
7.10344389e-02 -1.73221007e-01 7.28109360e-01 -4.79549736e-01
5.86410798e-02 9.80476677e-01 8.73360395e-01 6.45556092e-01
-1.26211905e+00 -1.19957018e+00 -4.11752798e-02 5.52517593e-01
-1.53823054e+00 -3.16070884e-01 6.07269168e-01 -8.00122023e-01
1.11854875e+00 1.59476642e-02 6.95479214e-01 8.76350820e-01
6.99964702e-01 2.25868851e-01 1.33270705e+00 -4.44823951e-01
-2.82447040e-01 1.24615289e-01 3.60705942e-01 8.72507513e-01
1.73479229e-01 3.67320061e-01 -5.95820308e-01 3.64145100e-01
8.59797835e-01 -3.67451668e-01 1.67538866e-01 -1.06237255e-01
-9.86491621e-01 7.59762645e-01 7.10422277e-01 3.02382290e-01
-3.51901770e-01 1.36669278e-01 3.62192959e-01 5.18076360e-01
7.79833734e-01 -6.12011291e-02 -6.26940608e-01 4.09152657e-01
-6.94862008e-01 2.97426432e-01 4.02264565e-01 5.19616544e-01
2.65800297e-01 -2.75333077e-02 -2.24000067e-02 8.33219826e-01
2.43865564e-01 1.89588502e-01 6.38666749e-01 -4.47989315e-01
4.70335603e-01 5.08859158e-01 -2.94693321e-01 -1.02989376e+00
-8.62330079e-01 -6.51287079e-01 -1.24567270e+00 3.97088617e-01
6.91921890e-01 -3.63393307e-01 -8.68307292e-01 1.96974111e+00
2.57489970e-04 1.19373284e-01 -3.16536240e-02 6.18333340e-01
1.23604608e+00 2.30688393e-01 3.89524847e-01 -5.39745390e-01
1.38259470e+00 -6.66876733e-01 -6.03042960e-01 -1.88157275e-01
4.33922827e-01 -3.98314744e-01 5.51354945e-01 6.09907031e-01
-1.07665765e+00 -7.38195300e-01 -1.25161815e+00 -6.17990121e-02
-4.32858109e-01 5.57227731e-01 7.81380832e-01 5.34551680e-01
-9.93359745e-01 7.94523120e-01 -7.95755863e-01 -1.73952922e-01
6.21360600e-01 9.23542023e-01 -7.11862624e-01 3.19249958e-01
-1.17433763e+00 1.13566124e+00 2.34367371e-01 3.89024705e-01
-5.42277336e-01 -7.71300197e-01 -6.91481113e-01 -9.07897279e-02
-1.73001096e-01 -7.10282385e-01 9.41476941e-01 -9.33531880e-01
-1.08393991e+00 1.35688972e+00 5.05204163e-02 -5.87330997e-01
5.54632723e-01 -6.47359014e-01 -5.43705702e-01 -8.42434913e-02
-3.58065963e-02 6.16664290e-01 7.79223740e-01 -6.82876766e-01
-5.53766251e-01 -7.45880246e-01 -2.64427334e-01 -1.66923609e-02
-2.17069641e-01 4.90378022e-01 2.51013730e-02 -5.69591820e-01
2.44166031e-01 -6.89878225e-01 -1.31004393e-01 9.22057554e-02
-9.61147845e-02 -5.42308569e-01 2.40218014e-01 -1.17939389e+00
1.22945225e+00 -1.84219813e+00 5.56439385e-02 -1.56754889e-02
4.66142684e-01 1.51985571e-01 2.41500959e-01 -1.66504204e-01
-8.45266879e-01 -2.11145431e-01 -9.76670533e-02 -2.33986169e-01
-4.21909571e-01 -3.45836967e-01 2.57450432e-01 6.03746951e-01
2.77413636e-01 5.74779153e-01 -5.39705932e-01 -6.78186893e-01
-9.15715396e-02 4.79099602e-01 -1.47707835e-01 3.10007513e-01
2.74095297e-01 5.63537478e-01 -1.01368070e-01 2.75890589e-01
6.10325456e-01 1.60387412e-01 3.58282519e-03 -3.69598418e-01
4.09236411e-03 -4.89273155e-03 -6.89295053e-01 1.47521234e+00
-3.42246026e-01 8.29484284e-01 -5.42470932e-01 -1.26413393e+00
1.43967855e+00 3.64026070e-01 4.50177580e-01 -1.06187928e+00
5.29857814e-01 8.03246442e-03 5.10277033e-01 -6.89684749e-01
-2.14131027e-01 -1.52328596e-01 1.36331320e-01 1.59443185e-01
3.78771871e-01 3.64464581e-01 2.12189689e-01 -3.19953233e-01
8.79084527e-01 2.22530723e-01 4.50246453e-01 -3.46209228e-01
8.58043730e-01 -7.18541741e-01 7.27116466e-01 2.86698192e-01
-4.47686553e-01 2.72747546e-01 9.82672393e-01 -7.32766986e-01
-7.95640111e-01 -1.24240017e+00 -2.77574092e-01 6.79245591e-01
-2.13864386e-01 -1.57458335e-01 -7.75212705e-01 -7.62719512e-01
-1.30539477e-01 4.81178373e-01 -1.06435239e+00 -2.84513354e-01
-6.98167324e-01 -9.80680764e-01 4.33504194e-01 8.18494439e-01
7.27894902e-01 -1.05610061e+00 -6.81419194e-01 -2.15785876e-01
-2.14573909e-02 -1.13411939e+00 1.31434560e-01 3.47213119e-01
-1.04337263e+00 -1.02451181e+00 -6.25803947e-01 -9.49619114e-01
7.80688465e-01 -7.08234727e-01 1.12473476e+00 -4.19584177e-02
-3.62245321e-01 -1.08114071e-01 -5.61891794e-02 -5.20102739e-01
1.16906583e-01 1.57498285e-01 2.43427917e-01 -1.42320335e-01
5.99967003e-01 -9.21012640e-01 -7.60579407e-01 6.31626621e-02
-1.47407621e-01 4.55755144e-02 8.80200624e-01 7.97854662e-01
3.02584052e-01 1.70374259e-01 7.13607430e-01 -4.96401489e-01
3.76641661e-01 -3.07856292e-01 -6.45372033e-01 1.91647574e-01
-8.97573709e-01 1.80019259e-01 3.29986066e-01 -6.60082579e-01
-9.71138418e-01 7.06840381e-02 7.69533450e-03 -1.44888595e-01
-1.21591851e-01 2.36460283e-01 7.90897086e-02 2.17959642e-01
7.02358603e-01 -9.02509764e-02 6.10225908e-02 -2.79422581e-01
5.91769640e-04 3.65967751e-01 8.88471544e-01 -4.33229625e-01
5.05710483e-01 -1.47206755e-02 4.26696628e-01 -4.18557554e-01
-8.12880278e-01 6.30250608e-04 -1.09618354e+00 -2.56149232e-01
1.17279077e+00 -9.10945177e-01 -9.22930479e-01 9.01076674e-01
-1.00675082e+00 -2.87625194e-01 3.65008175e-01 9.90218520e-01
-5.44818282e-01 -2.51810681e-02 -6.52870774e-01 -5.98666012e-01
-7.48400033e-01 -7.84704685e-01 3.70578527e-01 4.52942431e-01
-2.14785337e-01 -8.43748987e-01 -1.15497336e-01 3.95532489e-01
-8.98859501e-02 5.07299662e-01 1.07297897e+00 -5.76381087e-01
2.36478038e-02 -2.99690753e-01 -6.12818897e-01 5.15834510e-01
2.01857137e-03 2.07503140e-02 -9.30052817e-01 -8.84267911e-02
-9.38465446e-02 -2.20987514e-01 8.05844903e-01 6.71917915e-01
1.17681968e+00 -4.97297868e-02 -1.55733570e-01 5.70088387e-01
1.31412184e+00 4.73668903e-01 9.20666337e-01 3.73080760e-01
6.36783540e-01 9.85960305e-01 5.40663481e-01 3.40467364e-01
2.81255811e-01 4.67794597e-01 4.45709378e-01 -9.67255309e-02
-3.30894776e-02 2.93051124e-01 1.43950090e-01 7.09993720e-01
-5.44951081e-01 3.03553790e-01 -1.00389290e+00 4.97515947e-01
-1.83890152e+00 -7.61089563e-01 -3.73603314e-01 2.34005356e+00
8.87549877e-01 5.41627645e-01 2.06169367e-01 3.18006128e-01
6.74479187e-01 -1.72883555e-01 -4.30585742e-01 -7.02282548e-01
2.29064018e-01 6.42736197e-01 4.48557377e-01 1.46927372e-01
-1.22793269e+00 6.44061923e-01 5.76185656e+00 4.18242007e-01
-1.12555492e+00 4.56264056e-03 9.62248683e-01 1.10626467e-01
3.94629896e-01 -3.95307958e-01 -6.11683607e-01 2.31628522e-01
1.15059733e+00 -1.09086663e-01 1.71132341e-01 5.03177822e-01
-1.19926438e-01 -3.41262549e-01 -1.34679115e+00 1.25760627e+00
8.70468020e-02 -7.73576379e-01 -4.76758420e-01 -2.78649718e-01
4.33349073e-01 -4.21060473e-01 3.56868267e-01 4.14944559e-01
2.40476802e-02 -1.45724189e+00 4.65914428e-01 1.05432367e+00
8.38132620e-01 -1.03232121e+00 8.49569738e-01 1.79868460e-01
-1.13608479e+00 -1.59136832e-01 -8.19453001e-02 -4.53473926e-01
-4.06732231e-01 4.78417277e-01 -5.97766876e-01 5.45830548e-01
1.17306352e+00 6.02285385e-01 -8.80956948e-01 7.49132454e-01
-3.51333946e-01 3.02223027e-01 5.38053550e-03 3.84381157e-04
-2.46718079e-01 -1.72938600e-01 -1.72952116e-01 8.85441422e-01
2.50859648e-01 1.59296021e-01 -6.21588886e-01 8.08386505e-01
1.84029251e-01 2.46642560e-01 -4.20532435e-01 2.29853958e-01
6.20296076e-02 1.09438467e+00 -8.23278368e-01 -5.57281449e-02
-5.82689524e-01 7.25798845e-01 5.25331795e-01 4.93928865e-02
-9.83581960e-01 -5.63952744e-01 2.94557571e-01 1.83037773e-01
2.34221861e-01 -2.02494249e-01 -5.05091786e-01 -7.79907584e-01
4.70685074e-03 -5.44132292e-01 5.06839156e-01 -8.09694648e-01
-1.27376330e+00 5.38905203e-01 3.84084702e-01 -9.28499341e-01
-3.71151686e-01 -7.56995916e-01 -5.33649862e-01 1.08057833e+00
-1.03219688e+00 -1.29918146e+00 -2.86120117e-01 5.99702477e-01
1.74917772e-01 -4.57549244e-01 8.38494658e-01 5.41931152e-01
-7.59503663e-01 8.47684860e-01 -4.50119346e-01 5.05298078e-01
7.46240556e-01 -1.20039642e+00 1.45048901e-01 9.63506341e-01
-1.59889877e-01 5.89908302e-01 6.39548838e-01 -4.49155092e-01
-4.52270716e-01 -7.59653926e-01 1.30246961e+00 -2.69059747e-01
6.27868950e-01 -1.88374370e-01 -7.47154832e-01 7.44510949e-01
1.48743227e-01 3.08391005e-01 6.35809183e-01 3.90148073e-01
-5.76976180e-01 -6.94195449e-01 -1.14587963e+00 5.63148856e-01
9.22095060e-01 -1.90702498e-01 -6.31711364e-01 -1.35357141e-01
5.93284428e-01 -3.31368476e-01 -1.31895065e+00 8.19893062e-01
9.60264921e-01 -1.26977086e+00 1.14486384e+00 -5.25017142e-01
7.58265078e-01 -5.72484396e-02 1.19011275e-01 -9.14945602e-01
-2.81466454e-01 2.43102256e-02 -1.71712399e-01 1.33571053e+00
4.31107074e-01 -4.41661626e-01 6.93442464e-01 6.02232695e-01
1.59191951e-01 -1.11626172e+00 -8.21141422e-01 -6.68285728e-01
3.72619033e-01 -3.67489725e-01 4.08985853e-01 9.04532552e-01
-3.16478312e-01 6.52571917e-01 -3.71281296e-01 2.18626007e-01
7.40381122e-01 -3.97023439e-01 3.32278222e-01 -1.53827679e+00
-1.52864739e-01 -4.50538009e-01 -8.48487914e-01 -1.37786239e-01
3.37710261e-01 -4.85790402e-01 -2.54989117e-01 -1.39842260e+00
3.27696800e-01 -2.91173130e-01 -6.75068080e-01 4.16150600e-01
-9.30097997e-02 5.36922216e-02 -3.37750107e-01 -3.01904202e-01
-1.56749651e-01 4.36096877e-01 1.06580722e+00 -2.05940425e-01
-8.89441520e-02 4.15928900e-01 -4.61329669e-01 8.51259232e-01
9.88802850e-01 -5.07690668e-01 -7.21182764e-01 -2.53010541e-01
1.80104703e-01 1.66126162e-01 3.55566978e-01 -1.39432633e+00
-1.96884898e-03 6.94528073e-02 1.00999308e+00 -6.27556086e-01
2.00637475e-01 -5.78557432e-01 -1.16474200e-02 8.00770104e-01
-5.26881099e-01 2.93404877e-01 1.86820641e-01 7.87218958e-02
3.02973716e-03 -2.17626616e-01 8.58090460e-01 1.91305920e-01
-5.87076068e-01 4.02783155e-01 2.91098002e-02 -1.99303508e-01
1.06221890e+00 1.23487636e-02 -1.72352523e-01 -1.27226561e-01
-1.09716284e+00 -6.68338835e-02 -5.05670488e-01 3.08875471e-01
6.71025336e-01 -1.30091405e+00 -7.80966341e-01 9.73975882e-02
4.96900864e-02 -1.72660723e-01 1.85357317e-01 1.10356367e+00
-4.25700665e-01 1.79162636e-01 -8.12319696e-01 -4.26235259e-01
-1.75112295e+00 5.77213347e-01 4.44614708e-01 -5.08862913e-01
-3.82453084e-01 1.20182037e+00 2.19414026e-01 -1.97803944e-01
3.12546343e-01 -4.50657874e-01 -7.56954372e-01 2.63956845e-01
3.71216029e-01 3.37524056e-01 1.29633486e-01 -6.29260063e-01
-4.24357414e-01 7.18163729e-01 -4.63092953e-01 9.85141750e-03
1.39333999e+00 1.72689646e-01 -3.30560356e-01 6.13256693e-01
1.27604973e+00 -4.52274144e-01 -9.45484996e-01 -3.05326283e-01
6.60158275e-03 -1.52573541e-01 -3.28591503e-02 -8.19048226e-01
-1.28984118e+00 9.31601644e-01 1.28432596e+00 -1.62485898e-01
1.45239019e+00 -9.62633491e-02 4.96760845e-01 1.51212052e-01
-1.92259867e-02 -1.01663280e+00 1.84256002e-01 2.74072170e-01
9.12846863e-01 -1.20985901e+00 5.08453906e-01 -3.82893026e-01
-3.41586888e-01 1.23185503e+00 1.18212724e+00 -2.71151632e-01
8.55265498e-01 1.17119409e-01 -3.13585140e-02 -9.93823484e-02
-7.68426836e-01 -2.44552776e-01 6.72420979e-01 8.19834828e-01
8.29555452e-01 -5.41055314e-02 -7.96353102e-01 8.69609356e-01
-4.34700370e-01 1.78159341e-01 4.87921201e-03 4.96245474e-01
-6.33489564e-02 -9.58075821e-01 -1.32104799e-01 6.63427651e-01
-7.09104359e-01 9.33319330e-03 -2.87290514e-02 1.05672777e+00
6.47115767e-01 7.41681993e-01 4.44297016e-01 -5.96169412e-01
1.70019343e-01 -1.12782657e-01 9.99042273e-01 -1.52755618e-01
-5.18276155e-01 -2.60594577e-01 2.46899128e-01 -4.03075397e-01
-6.03040576e-01 -7.55903721e-01 -1.15442014e+00 -2.30874449e-01
-3.66344564e-02 -3.37477028e-01 5.40434837e-01 1.06392908e+00
-2.74433047e-01 8.37697983e-01 4.06022459e-01 -2.85292357e-01
-1.79019555e-01 -1.14359748e+00 -4.14383084e-01 2.64986426e-01
2.68465817e-01 -9.50624049e-01 4.30545071e-03 2.67413437e-01] | [14.060895919799805, -1.3528903722763062] |
34361398-f7ca-4af4-8301-50259a6bf7a5 | image-to-voxel-model-translation-for-3d-scene | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/156_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123520103.pdf | Image-to-Voxel Model Translation for 3D Scene Reconstruction and Segmentation | Objects class, depth, and shape are instantly reconstructed by a human looking at a 2D image. While modern deep models solve each of these challenging tasks separately, they struggle to perform simultaneous scene 3D reconstruction and segmentation. We propose a single shot image-to-semantic voxel model translation framework. We train a generator adversarially against a discriminator that verifies the object's poses. Furthermore, trapezium-shaped voxels, volumetric residual blocks, and 2D-to-3D skip connections facilitate our model learning explicit reasoning about 3D scene structure. We collected a SemanticVoxels dataset with 116k images, ground-truth semantic voxel models, depth maps, and 6D object poses. Experiments on ShapeNet and our SemanticVoxels datasets demonstrate that our framework achieves and surpasses state-of-the-art in the reconstruction of scenes with multiple non-rigid objects of different classes. We made our model and dataset publicly available | ['Vladimir V. Kniaz', 'Vladimir A. Knyaz', 'Artem Bordodymov', 'Petr Moshkantsev', 'Fabio Remondino'] | null | null | null | null | eccv-2020-8 | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 4.33208197e-01 5.67960262e-01 7.06739053e-02 -5.68140984e-01
-1.03582847e+00 -8.43428910e-01 6.52868867e-01 -3.46813798e-01
-9.72759724e-02 3.40992808e-01 -1.43046856e-01 -1.13632642e-01
3.46400321e-01 -1.09655488e+00 -1.33842134e+00 -4.55328435e-01
2.67985731e-01 1.33938646e+00 5.11365056e-01 1.98790848e-01
1.17804043e-01 7.37302780e-01 -1.16181123e+00 2.29936704e-01
5.42897880e-01 1.13524294e+00 1.42578259e-01 6.63604081e-01
-3.01156223e-01 6.95024371e-01 -1.50360733e-01 -5.75983524e-01
7.81173170e-01 -1.17372356e-01 -9.19967830e-01 3.04417044e-01
1.05152941e+00 -8.04989994e-01 -6.95182264e-01 1.23113096e+00
3.00826758e-01 -2.64807045e-02 7.68724263e-01 -1.16510642e+00
-7.64285862e-01 3.29838127e-01 -5.45709908e-01 -3.23035121e-01
3.73816341e-01 6.11545920e-01 6.49160028e-01 -8.49352360e-01
9.78695750e-01 1.47084379e+00 6.36232376e-01 7.81294703e-01
-1.48346472e+00 -5.10368049e-01 1.17275864e-01 -2.68772572e-01
-1.15998089e+00 -1.66342631e-01 8.73359680e-01 -6.02659881e-01
1.18385494e+00 4.53199930e-02 9.87557650e-01 1.32959735e+00
1.40801277e-02 6.94881499e-01 1.01999485e+00 2.01889515e-01
3.41611266e-01 -3.28096688e-01 -2.99799830e-01 9.05812204e-01
1.01034813e-01 2.75632024e-01 -1.37485802e-01 -9.28584207e-03
1.40558267e+00 -4.43721414e-02 8.48785639e-02 -8.46283376e-01
-1.16737521e+00 5.61665952e-01 6.72859728e-01 -3.60061884e-01
-1.69089288e-01 7.66436636e-01 6.74759299e-02 -7.36747831e-02
4.64951783e-01 3.47892433e-01 -3.84525955e-01 3.13991100e-01
-7.00604439e-01 5.73347211e-01 6.86097741e-01 1.26104891e+00
5.64569116e-01 2.80600786e-01 -1.74585655e-02 2.90416270e-01
2.74115056e-01 8.53762388e-01 -1.05326600e-01 -1.48730755e+00
3.97288203e-01 4.96179044e-01 -1.52240872e-01 -6.86888874e-01
-2.84664333e-01 -1.82587579e-01 -7.17157245e-01 4.08976197e-01
3.82986635e-01 3.12494010e-01 -1.59536731e+00 1.52072775e+00
6.22836709e-01 6.11333966e-01 -6.41523451e-02 1.17909598e+00
1.24851215e+00 3.81221175e-01 4.66478467e-02 4.63283062e-01
9.00516152e-01 -9.96506155e-01 3.03022116e-02 -4.05070662e-01
2.37476304e-02 -4.71094936e-01 1.02812088e+00 1.50000945e-01
-1.52024949e+00 -3.94016117e-01 -6.76604152e-01 -6.25377595e-01
-5.33615351e-02 -6.65482581e-01 7.68414378e-01 3.16834331e-01
-9.46538806e-01 6.35592282e-01 -9.67386603e-01 6.11076355e-02
1.05828059e+00 2.30318949e-01 -5.10032117e-01 -3.17603737e-01
-6.21526659e-01 5.92778683e-01 2.75113404e-01 -3.51161629e-01
-1.85516763e+00 -1.26487315e+00 -1.05411017e+00 -1.55660287e-01
3.00062090e-01 -1.29148757e+00 1.40556598e+00 -5.60837448e-01
-1.28633988e+00 1.71216226e+00 2.98401415e-01 -5.46924651e-01
1.15176475e+00 9.80060995e-02 2.43045270e-01 3.75650227e-01
3.09389859e-01 1.05100298e+00 8.91602278e-01 -1.56410253e+00
-6.88827038e-02 -7.20108032e-01 2.04419121e-01 1.85905889e-01
6.79770768e-01 -5.63405871e-01 -5.38955569e-01 -4.50101644e-01
5.66645384e-01 -7.10277379e-01 -4.08888668e-01 6.03171885e-01
-8.77216160e-01 1.36663035e-01 6.17706299e-01 -5.26193976e-01
-1.56425744e-01 -1.97503567e+00 2.96821237e-01 1.07562035e-01
2.90352702e-01 -1.98540926e-01 -2.82210290e-01 -3.34687233e-01
5.64588085e-02 8.08845386e-02 -6.46143258e-01 -7.02231884e-01
1.81666940e-01 4.96602237e-01 -4.80848700e-01 7.31869280e-01
1.34494513e-01 1.52370000e+00 -1.04809368e+00 -5.10975778e-01
5.78760624e-01 6.72761798e-01 -9.63290036e-01 3.13071012e-01
-7.68666446e-01 8.67929220e-01 -5.08726776e-01 8.98919106e-01
9.90532875e-01 -3.94602120e-01 -2.21368968e-01 -1.45405814e-01
4.55699652e-01 2.36791357e-01 -8.02690983e-01 2.32974124e+00
-2.21899390e-01 2.17387602e-01 1.33688331e-01 -7.62169421e-01
7.66915679e-01 5.77686355e-02 6.61182106e-01 -7.73705959e-01
1.86041147e-01 1.30496055e-01 -4.89345312e-01 -3.68207991e-01
2.55171180e-01 -3.78107667e-01 -9.69565362e-02 2.21026063e-01
1.76945642e-01 -1.46209466e+00 -5.55676639e-01 2.53959268e-01
9.61873770e-01 5.10221124e-01 -1.90971851e-01 -4.56545316e-02
-2.56570932e-02 1.45404398e-01 2.55610585e-01 7.27998734e-01
-2.61662472e-02 1.25370038e+00 4.16900396e-01 -6.12274587e-01
-1.50793397e+00 -1.78153336e+00 -1.63606569e-01 3.44587028e-01
5.89106441e-01 1.61522850e-01 -9.07859743e-01 -5.93703806e-01
3.87871861e-01 9.56229746e-01 -7.57634401e-01 -3.57135236e-02
-4.93731707e-01 -7.59354308e-02 5.84496498e-01 5.73248982e-01
5.72109759e-01 -1.01250815e+00 -6.66276753e-01 1.24682881e-01
-6.86305761e-02 -1.53041685e+00 -3.40107858e-01 9.03886557e-02
-9.55659270e-01 -1.36189055e+00 -5.44027209e-01 -8.04325938e-01
8.36276472e-01 2.07841247e-01 1.59140110e+00 -7.22539350e-02
-5.56906462e-01 6.19732976e-01 2.93860048e-01 -1.97827563e-01
-5.34899712e-01 -1.49915978e-01 -3.41651708e-01 -4.94774997e-01
-1.59922436e-01 -8.63103807e-01 -7.52610803e-01 2.26945430e-01
-8.50878000e-01 4.98668283e-01 -5.08749746e-02 3.80461305e-01
1.23996305e+00 -2.43339345e-01 -5.77022508e-02 -1.05081332e+00
-1.56056687e-01 -2.78765440e-01 -8.12398851e-01 1.90033671e-02
-1.93909407e-01 -1.14239037e-01 3.42531890e-01 -3.53243411e-01
-9.37851429e-01 3.87253582e-01 -3.15401614e-01 -9.40742671e-01
-4.51321155e-01 -4.00747687e-01 -3.11942786e-01 -1.77619815e-01
6.60189927e-01 2.10565552e-01 -1.46166369e-01 -3.76122624e-01
6.30194545e-01 -1.80934474e-01 1.15863299e+00 -9.48039591e-01
1.03129816e+00 9.69788551e-01 2.16481835e-01 -2.12825701e-01
-1.06763422e+00 -1.89358592e-02 -8.18971455e-01 -1.03834406e-01
1.37128687e+00 -1.27306092e+00 -7.90925801e-01 6.62941992e-01
-1.36287320e+00 -9.17774558e-01 -7.11736381e-01 -6.52568489e-02
-1.20557940e+00 3.27470154e-02 -6.39802635e-01 -2.28287593e-01
-3.12397242e-01 -1.26655006e+00 1.65953302e+00 -5.73158562e-02
2.00107992e-02 -8.35891426e-01 -2.45870888e-01 7.11668193e-01
1.44441724e-01 8.70058358e-01 9.99165714e-01 -1.11990057e-01
-1.29287612e+00 2.19920166e-02 -3.27949196e-01 1.53712407e-01
-1.80717722e-01 -2.48263523e-01 -1.10480607e+00 -1.01187401e-01
1.64698511e-02 -7.14253247e-01 7.64515817e-01 4.32652295e-01
1.78359199e+00 -1.89065516e-01 -6.32392615e-02 1.43643010e+00
1.45866644e+00 -2.11836860e-01 7.79135644e-01 -1.36053085e-01
1.16614366e+00 2.42945328e-01 1.37438148e-01 1.29673138e-01
4.88106489e-01 4.33036953e-01 1.17331505e+00 -1.61743522e-01
-6.17663264e-01 -7.95103192e-01 -1.30095094e-01 2.52865165e-01
6.78857230e-03 -9.12232473e-02 -9.20940578e-01 4.13310349e-01
-1.42873359e+00 -7.20467567e-01 -2.11821154e-01 1.75162494e+00
7.80493498e-01 4.27776933e-01 -2.46026069e-01 -4.44853246e-01
2.71707326e-01 1.85044855e-01 -1.29068887e+00 -4.70227934e-02
-2.23962054e-01 4.20344949e-01 7.49858677e-01 7.52520382e-01
-9.27226901e-01 1.42031944e+00 6.29819250e+00 4.93924320e-01
-9.83181953e-01 2.49462187e-01 8.44318688e-01 -2.18713000e-01
-9.37233150e-01 -9.86070335e-02 -4.32440341e-01 1.09059334e-01
2.11419821e-01 -7.96991140e-02 7.10260928e-01 1.06984377e+00
-2.49748915e-01 6.23687170e-02 -1.41160631e+00 1.25466275e+00
8.07398185e-02 -1.54335177e+00 3.03855568e-01 -8.75086337e-03
1.20234573e+00 4.41965789e-01 1.22217976e-01 7.40402415e-02
7.05374956e-01 -1.48106778e+00 1.26489675e+00 5.27326584e-01
1.29374349e+00 -3.78166914e-01 2.44602822e-02 3.25299978e-01
-9.32954788e-01 4.25839186e-01 -3.86778682e-01 2.32534081e-01
3.69566143e-01 2.84436703e-01 -8.00425649e-01 3.64734292e-01
6.20772541e-01 7.18940020e-01 -2.80434549e-01 7.27772355e-01
-5.02901673e-01 5.89823835e-02 -4.18868035e-01 6.47379398e-01
3.48392695e-01 -3.81142274e-02 6.09313726e-01 6.67999446e-01
1.25361636e-01 3.12639922e-01 2.30290383e-01 1.73506904e+00
-4.20888871e-01 -5.40252388e-01 -7.42004633e-01 2.34366551e-01
2.79129863e-01 8.56345415e-01 -1.00734830e+00 -4.37403023e-01
-5.66469785e-03 1.27120376e+00 1.52060002e-01 3.25824291e-01
-1.05777395e+00 4.13290977e-01 9.51271713e-01 4.35593009e-01
2.08958864e-01 -3.88979733e-01 -8.64823461e-01 -1.13741446e+00
-1.29624575e-01 -4.73362356e-01 -1.10342996e-02 -1.21430683e+00
-1.40122032e+00 3.05260032e-01 1.75824631e-02 -8.35852563e-01
-6.23052679e-02 -6.01948261e-01 -3.44208002e-01 5.94175696e-01
-1.33224428e+00 -1.31412816e+00 -5.90329349e-01 8.95171940e-01
7.00592697e-01 2.11232841e-01 8.34376693e-01 1.04723684e-01
1.81202054e-01 3.41473848e-01 -5.47677994e-01 3.31540972e-01
1.74163684e-01 -1.18008685e+00 1.10516012e+00 5.76395988e-01
1.52545601e-01 -4.84763384e-02 4.36277002e-01 -7.82905042e-01
-1.58034778e+00 -1.35534930e+00 1.06705576e-01 -8.88487041e-01
3.60154420e-01 -6.16899073e-01 -7.24776447e-01 9.74546313e-01
-2.85973549e-01 5.18698394e-01 -2.14797147e-02 -6.77097261e-01
-6.33020520e-01 2.81044245e-01 -1.65853190e+00 5.13046086e-01
1.79119790e+00 -7.62929559e-01 -5.96879125e-01 6.36302948e-01
1.18464029e+00 -1.29415345e+00 -7.19939888e-01 5.14292121e-01
4.06537414e-01 -9.16920006e-01 1.71309483e+00 -8.49407017e-01
7.98771083e-01 -1.81092918e-01 -5.77115059e-01 -9.78489697e-01
1.51522057e-02 -2.69280910e-01 -1.45941049e-01 4.58839655e-01
7.59645626e-02 -2.97682345e-01 1.06735563e+00 1.01693344e+00
-4.41033006e-01 -6.58831418e-01 -1.14949393e+00 -6.93982005e-01
3.55173320e-01 -7.56146789e-01 8.46053898e-01 8.87397349e-01
-1.02195334e+00 3.61869298e-02 -5.40679954e-02 2.25043222e-01
1.22778225e+00 4.22456264e-01 1.05293262e+00 -1.03062248e+00
-4.31220800e-01 -4.30418968e-01 -6.10316813e-01 -1.23007500e+00
4.11347866e-01 -1.37963831e+00 -4.86256853e-02 -1.72160292e+00
2.29021922e-01 -4.91245389e-01 1.95018038e-01 6.49767876e-01
3.38296473e-01 5.62386215e-01 1.91969842e-01 -1.12476073e-01
-5.30230939e-01 5.48527718e-01 2.00495911e+00 -6.15006924e-01
8.39109048e-02 -2.58231968e-01 -4.23673481e-01 8.52942705e-01
4.51092362e-01 -5.97178757e-01 -5.30481219e-01 -1.00437713e+00
1.42321631e-01 2.06277177e-01 1.24319470e+00 -7.87210107e-01
4.89107287e-03 -3.73208225e-01 6.81235373e-01 -7.58737147e-01
7.90990293e-01 -8.30362618e-01 4.43105638e-01 3.66160154e-01
-2.51705021e-01 -4.44646508e-01 1.09000728e-01 5.20864844e-01
2.97124654e-01 2.81206042e-01 1.03437042e+00 -8.04483294e-01
-7.38591433e-01 9.55384612e-01 5.29352486e-01 5.19094288e-01
1.05094087e+00 -3.56229007e-01 -1.69191107e-01 -1.19926482e-01
-9.60758984e-01 2.52367049e-01 9.88799870e-01 4.65218455e-01
1.03194368e+00 -1.22740579e+00 -6.38578117e-01 4.34031546e-01
-7.10769147e-02 1.09645951e+00 4.77598757e-01 3.06699306e-01
-8.64063919e-01 3.52123380e-02 -2.68926889e-01 -1.10257256e+00
-7.62325943e-01 4.56849575e-01 7.95260489e-01 2.15900809e-01
-1.26310849e+00 1.24674237e+00 6.36372566e-01 -9.63740408e-01
3.67554575e-01 -5.10480404e-01 8.20040345e-01 -6.67560697e-01
8.22687820e-02 -7.89752230e-02 -1.59653574e-01 -5.29791832e-01
-3.24907869e-01 8.30133319e-01 3.11922103e-01 -5.32985777e-02
1.46427512e+00 1.61873952e-01 -2.55211797e-02 2.40129247e-01
1.28594911e+00 -4.04817104e-01 -1.85777962e+00 -1.51968360e-01
-7.03940034e-01 -7.17449367e-01 -1.30209938e-01 -8.44160497e-01
-1.32422793e+00 9.68622088e-01 3.29949290e-01 -3.66546839e-01
6.01180851e-01 6.36416197e-01 1.04427564e+00 1.03975818e-01
8.75469387e-01 -6.80806398e-01 3.36925536e-01 5.64170361e-01
9.34657156e-01 -1.26518941e+00 -1.67914256e-01 -6.15487456e-01
-3.75308007e-01 8.00941229e-01 8.45263660e-01 -5.84739745e-01
6.87754810e-01 4.27422196e-01 -1.42203659e-01 -4.51195478e-01
-3.21619511e-01 7.41212368e-02 2.15466917e-01 6.75370932e-01
-3.49307269e-01 4.48793948e-01 6.71171844e-01 3.08039129e-01
-6.87802494e-01 -1.95561796e-01 1.67429775e-01 4.87408578e-01
-5.66720292e-02 -6.07472241e-01 -2.57183433e-01 1.84101954e-01
-7.35006705e-02 5.14975935e-02 -3.64117116e-01 7.41325438e-01
1.65768728e-01 3.00476611e-01 4.13153261e-01 -1.85550913e-01
3.91329497e-01 -1.80253237e-01 1.13466585e+00 -8.06058049e-01
-2.22130120e-01 -1.27217725e-01 -2.59430826e-01 -1.11785698e+00
-1.21730864e-01 -3.42491359e-01 -1.74926698e+00 -3.77624184e-01
4.13554907e-01 -5.94177485e-01 7.89222240e-01 7.57712543e-01
2.02256054e-01 5.85725129e-01 4.19819593e-01 -1.20475316e+00
-4.27805752e-01 -4.80037600e-01 -4.44500566e-01 8.49463761e-01
2.81559706e-01 -5.10500669e-01 -1.85155496e-01 4.84262332e-02] | [8.848272323608398, -3.3188180923461914] |
f6150243-e78c-4103-9424-3b6c0b101834 | aligning-language-models-to-user-opinions | 2305.14929 | null | https://arxiv.org/abs/2305.14929v1 | https://arxiv.org/pdf/2305.14929v1.pdf | Aligning Language Models to User Opinions | An important aspect of developing LLMs that interact with humans is to align models' behavior to their users. It is possible to prompt an LLM into behaving as a certain persona, especially a user group or ideological persona the model captured during its pertaining stage. But, how to best align an LLM with a specific user and not a demographic or ideological group remains an open question. Mining public opinion surveys (by Pew Research), we find that the opinions of a user and their demographics and ideologies are not mutual predictors. We use this insight to align LLMs by modeling both user opinions as well as user demographics and ideology, achieving up to 7 points accuracy gains in predicting public opinions from survey questions across a broad set of topics. In addition to the typical approach of prompting LLMs with demographics and ideology, we discover that utilizing the most relevant past opinions from individual users enables the model to predict user opinions more accurately. | ['Niket Tandon', 'Bodhisattwa Prasad Majumder', 'EunJeong Hwang'] | 2023-05-24 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [-2.45290294e-01 3.63682866e-01 -7.91788757e-01 -8.88223827e-01
-3.33772808e-01 -8.08456838e-01 8.56183648e-01 5.85699975e-01
-3.29192102e-01 2.13857457e-01 1.06183231e+00 -7.27054954e-01
3.07561308e-01 -6.12954140e-01 -8.89941379e-02 3.64721976e-02
4.50245917e-01 7.68445849e-01 -2.36427665e-01 -3.65182877e-01
5.88980258e-01 -2.08943244e-02 -1.45194387e+00 7.27646768e-01
1.14660203e+00 7.41301000e-01 -3.24942231e-01 5.88554680e-01
-3.59733462e-01 9.33478653e-01 -5.21662772e-01 -6.67843699e-01
8.34114775e-02 -2.17955709e-01 -9.23853278e-01 1.41502336e-01
9.51003611e-01 -1.95831612e-01 4.51884061e-01 8.29485953e-01
2.22722679e-01 -1.60858233e-03 9.36513484e-01 -1.11667252e+00
-8.73429537e-01 8.18484128e-01 -5.65836191e-01 1.50461853e-01
7.11954474e-01 -2.05535725e-01 1.43992043e+00 -5.10909915e-01
7.14902639e-01 1.62685156e+00 7.29629576e-01 1.56765997e-01
-1.32423580e+00 -5.39160430e-01 6.09331608e-01 -1.77334860e-01
-7.78929234e-01 -4.14530635e-01 7.17394650e-01 -9.96876836e-01
4.50296104e-01 4.40715879e-01 7.38291979e-01 1.20433855e+00
2.19930559e-01 6.14589393e-01 1.49568236e+00 -2.51502961e-01
4.50387180e-01 7.99965739e-01 9.59635496e-01 3.86867672e-01
3.18466067e-01 -5.81634223e-01 -6.17918432e-01 -9.06703472e-01
-1.15690224e-01 2.03025877e-01 1.20313220e-01 -1.03367068e-01
-8.90759945e-01 1.07040167e+00 7.67442882e-02 5.27099431e-01
-5.12587368e-01 -4.71407562e-01 -1.74064990e-02 5.39334893e-01
1.03678453e+00 1.14392424e+00 -8.64268064e-01 -1.13502771e-01
-8.26357901e-01 4.93639618e-01 1.55154848e+00 5.09582162e-01
1.03551126e+00 -5.09597242e-01 9.41339508e-02 8.49644065e-01
5.57421505e-01 5.97671092e-01 6.43696129e-01 -9.98532534e-01
3.89135815e-03 1.23622191e+00 6.19445562e-01 -1.47238898e+00
-4.31924224e-01 -4.81608480e-01 -2.79243648e-01 -1.51068449e-01
4.34833944e-01 -4.84002471e-01 -5.52101016e-01 1.81365883e+00
2.04691947e-01 -5.05229712e-01 -8.09753463e-02 4.28703904e-01
4.67691422e-01 6.89350665e-01 3.68834645e-01 -2.48689845e-01
1.50973225e+00 -3.10581148e-01 -3.84910554e-01 -7.68999159e-01
8.27578127e-01 -1.02268755e+00 1.07919979e+00 3.36407721e-01
-8.62271190e-01 -5.69295645e-01 -6.51381910e-01 7.84419570e-03
-3.91849339e-01 -8.57457593e-02 8.61848354e-01 6.09039485e-01
-1.17202568e+00 3.74003291e-01 -3.75943303e-01 -1.02070773e+00
-1.80787727e-01 2.81281024e-01 5.51128387e-02 3.78220171e-01
-1.00174856e+00 1.09781790e+00 -3.15407604e-01 -6.53831184e-01
8.43458250e-03 -7.80086875e-01 -5.76536953e-01 -2.06094578e-01
-9.57014598e-03 -8.38610411e-01 1.62333941e+00 -1.41776097e+00
-1.15159011e+00 9.94079411e-01 -8.32617223e-01 -1.64138168e-01
3.41113716e-01 -1.64131239e-01 -6.57077372e-01 -5.24657786e-01
4.32844132e-01 3.70483160e-01 4.19388920e-01 -1.28028893e+00
-9.87467229e-01 -6.20855033e-01 2.28877008e-01 2.60585010e-01
-7.84293056e-01 2.40301028e-01 2.23615691e-01 7.77628133e-03
2.99529105e-01 -9.82516766e-01 -4.49242622e-01 -6.49589121e-01
-1.82206884e-01 -5.72192729e-01 7.42205501e-01 -7.11208045e-01
1.38629162e+00 -1.59729886e+00 -6.95583522e-02 5.92262864e-01
4.09817159e-01 -1.09845258e-01 9.23000425e-02 6.92712784e-01
2.06006065e-01 6.37049556e-01 6.47420406e-01 -2.34696865e-01
2.23525137e-01 -2.54201591e-01 -5.81869423e-01 3.73364627e-01
-5.53073287e-01 7.10126162e-01 -9.56786454e-01 -3.48237574e-01
-1.98564515e-01 -1.10843759e-02 -9.33051705e-01 4.39901911e-02
-3.63363922e-01 2.90412277e-01 -8.36635947e-01 5.33394516e-01
2.01732859e-01 -6.37534857e-01 8.14137280e-01 -1.11680977e-01
-3.43408287e-01 5.94027042e-01 -6.48532748e-01 9.00220037e-01
-6.28368020e-01 6.37866378e-01 2.07766056e-01 -4.18748647e-01
1.15175271e+00 2.85514981e-01 6.21281922e-01 -3.96359295e-01
2.85976797e-01 3.00870955e-01 1.87338829e-01 -5.30087411e-01
5.97974956e-01 -4.16464731e-02 -4.38912362e-01 9.01682973e-01
-4.02681738e-01 8.84875432e-02 2.17906848e-01 2.01638892e-01
7.90512264e-01 -4.61827189e-01 7.80482411e-01 -9.27579105e-01
2.53495395e-01 2.62925416e-01 5.69403529e-01 8.20294380e-01
-1.05952516e-01 -1.44672558e-01 3.57090682e-01 -7.40190744e-01
-1.02057874e+00 -4.88685727e-01 6.51237220e-02 1.73084724e+00
-4.34713922e-02 -6.57256365e-01 -4.52281266e-01 -5.64786077e-01
2.24473476e-01 9.73383904e-01 -4.86403316e-01 3.76015455e-01
-2.48363480e-01 -6.64550126e-01 -1.98447257e-01 8.52495208e-02
1.32179722e-01 -6.04749501e-01 -2.05452099e-01 2.30497256e-01
-4.50456768e-01 -7.64826775e-01 -4.42745775e-01 -2.77183920e-01
-6.74157023e-01 -9.13562357e-01 -4.27888781e-01 -6.31059885e-01
8.89634371e-01 2.51058221e-01 1.56028819e+00 2.13183567e-01
6.89490139e-01 3.67263228e-01 -3.26396704e-01 -6.84724152e-01
-7.35551119e-01 5.38947999e-01 4.72595692e-01 2.57746667e-01
1.01214695e+00 -9.53561127e-01 -6.28621757e-01 6.03870392e-01
-3.37984145e-01 2.05785885e-01 3.18910718e-01 7.23382533e-02
-4.78824288e-01 -3.41930717e-01 4.93758291e-01 -1.50770366e+00
1.00219750e+00 -9.05079901e-01 -6.53700083e-02 3.43987823e-01
-9.58067715e-01 -1.33104756e-01 4.74327683e-01 -4.40140605e-01
-1.08848047e+00 -2.39924252e-01 -3.98393646e-02 6.38917148e-01
-2.76285380e-01 6.47301495e-01 1.63365021e-01 4.30613607e-01
1.05817115e+00 -4.42206323e-01 -1.02000281e-01 -5.06042898e-01
2.84429908e-01 1.09192348e+00 9.39794928e-02 -8.46729159e-01
7.15561628e-01 3.95635724e-01 -8.15363765e-01 -8.16032290e-01
-1.29132295e+00 -7.93717444e-01 -3.82839233e-01 -4.30371583e-01
6.77667022e-01 -8.95451307e-01 -9.65652704e-01 1.63671464e-01
-1.12403595e+00 -5.69842607e-02 3.00201386e-01 2.48436347e-01
-1.80061966e-01 1.49457619e-01 -5.45283318e-01 -9.96606171e-01
-2.64504790e-01 -8.41765583e-01 6.98876083e-01 2.53400326e-01
-1.62597334e+00 -1.51454878e+00 2.86921322e-01 7.85280943e-01
5.45141816e-01 3.69820669e-02 1.16109049e+00 -1.23364794e+00
1.51299322e-02 -3.26027185e-01 2.65754700e-01 -3.76738273e-02
3.22377384e-01 2.76240855e-01 -8.10576260e-01 -3.81001085e-01
1.79031849e-01 -2.78593004e-01 3.48932207e-01 2.94588506e-01
3.97133440e-01 -8.33072305e-01 -6.66006923e-01 -2.66492683e-02
1.04100394e+00 2.99152195e-01 -1.87002607e-02 2.55142242e-01
3.53132427e-01 1.08668411e+00 3.64265561e-01 2.75751948e-01
8.03027689e-01 2.57585913e-01 -2.95527548e-01 -2.12242290e-01
7.47008264e-01 -5.49746335e-01 6.69528306e-01 9.05356944e-01
6.50364980e-02 -1.89565629e-01 -1.18317807e+00 3.16005975e-01
-2.03070712e+00 -1.02620006e+00 -5.24828881e-02 1.72882962e+00
6.99680805e-01 1.65134698e-01 2.98202395e-01 -2.50439852e-01
5.20591915e-01 4.40052092e-01 -4.75356489e-01 -7.45258689e-01
1.23193778e-01 -2.61307746e-01 4.03231382e-01 8.80836308e-01
-6.61639631e-01 8.07161152e-01 7.20596933e+00 -5.32393251e-03
-1.04171562e+00 -7.16928765e-02 1.15723228e+00 1.44308686e-01
-9.99351561e-01 3.75693202e-01 -1.08493114e+00 4.26134974e-01
8.65551353e-01 -5.90482295e-01 2.82737523e-01 1.09137917e+00
5.01612425e-01 -1.04536876e-01 -1.27974308e+00 3.73166680e-01
1.65345490e-01 -1.13556886e+00 -2.36175321e-02 4.20932025e-01
1.20394313e+00 -2.77955174e-01 9.94301662e-02 3.92330498e-01
1.02857447e+00 -7.39484072e-01 6.36496365e-01 4.67818975e-01
1.46030307e-01 -4.06724930e-01 5.08273959e-01 7.44223773e-01
-6.67329609e-01 -3.89386594e-01 8.30018055e-03 -6.60177350e-01
9.06977877e-02 6.08459175e-01 -1.25123632e+00 -2.01103464e-01
4.44916368e-01 7.89590001e-01 -6.86275005e-01 1.62748575e-01
9.65717658e-02 8.28462422e-01 -4.46920507e-02 -2.34255314e-01
1.06396727e-01 -3.42595637e-01 1.88637465e-01 1.08986616e+00
4.66587454e-01 1.22295223e-01 6.03917956e-01 5.38997412e-01
1.52499318e-01 4.35882956e-01 -7.15697467e-01 -2.53286660e-01
5.20406365e-01 1.40046215e+00 -4.00560647e-01 -6.62025154e-01
-5.15891492e-01 4.44671780e-01 8.62617493e-02 5.26786268e-01
-4.97485101e-02 3.98044318e-01 1.04708993e+00 7.14288592e-01
-6.84006870e-01 3.02108703e-03 -5.74022532e-01 -1.21056247e+00
-4.05814528e-01 -1.29311311e+00 2.78891295e-01 -7.05837846e-01
-1.52223825e+00 2.73261279e-01 -1.19868442e-01 -7.68666387e-01
-6.18413568e-01 -6.19675398e-01 -8.93708229e-01 8.63368392e-01
-8.04626942e-01 -1.06062579e+00 -1.57073751e-01 -7.67852888e-02
4.06514704e-01 -1.99014738e-01 5.13096273e-01 -1.64706469e-01
-2.43004709e-01 7.24131539e-02 -2.67941058e-01 -9.33338255e-02
1.00948930e+00 -1.33276331e+00 5.99541247e-01 3.56630743e-01
-8.02630708e-02 1.34557760e+00 1.35174942e+00 -8.21540952e-01
-9.72709537e-01 -4.62282777e-01 1.71447706e+00 -1.01337874e+00
8.37227464e-01 -2.53080189e-01 -6.26497269e-01 1.11019790e+00
5.53326964e-01 -1.13117492e+00 1.27818942e+00 1.04486465e+00
-3.14065099e-01 1.26585603e-01 -8.47596347e-01 1.02949834e+00
9.50709343e-01 -7.42384255e-01 -9.27302539e-01 4.24107671e-01
3.18242759e-01 1.75122656e-02 -8.99538517e-01 2.63421237e-01
7.78543413e-01 -1.13200724e+00 6.53794825e-01 -7.51688659e-01
3.89401495e-01 -4.31191400e-02 -1.53997675e-01 -1.66042507e+00
-4.68733668e-01 -6.26461804e-01 1.45270988e-01 1.44034684e+00
6.51476264e-01 -8.85512412e-01 8.87174666e-01 1.34429073e+00
2.68336892e-01 -7.01538563e-01 9.30612385e-02 1.75499216e-01
1.21567525e-01 -1.63049430e-01 6.91313922e-01 1.40383804e+00
4.96766865e-01 8.48504245e-01 -3.23733836e-01 1.08282238e-01
2.80980438e-01 4.28861856e-01 1.12969863e+00 -1.96685147e+00
-9.37281623e-02 -6.45358324e-01 1.50567606e-01 -1.27065229e+00
3.07288110e-01 -8.94682646e-01 -3.30834359e-01 -1.68210673e+00
3.81021887e-01 -2.99463421e-01 -1.80298641e-01 9.51537564e-02
-1.22197807e-01 -2.44791016e-01 -5.72273061e-02 3.46556872e-01
-5.78543603e-01 -2.36607626e-01 8.66580009e-01 5.75110540e-02
-4.35125411e-01 2.13538036e-01 -1.66355813e+00 1.39107847e+00
7.35319614e-01 -2.25121140e-01 -5.14825523e-01 -1.38671070e-01
1.09706414e+00 -3.71948592e-02 -9.05249789e-02 -6.75668061e-01
3.19038004e-01 -7.07027137e-01 5.79024613e-01 -6.58390582e-01
8.54989290e-02 -5.50744712e-01 2.28459999e-01 2.51039863e-01
-7.47765422e-01 4.19200271e-01 -3.64618927e-01 2.24602714e-01
-6.69392422e-02 4.81745526e-02 5.02887607e-01 -4.14935023e-01
-7.29383767e-01 1.61316231e-01 -7.71995604e-01 -7.65909301e-03
5.93423188e-01 -1.13510892e-01 -5.62409163e-01 -1.03450847e+00
-5.39422452e-01 5.15277863e-01 1.14707732e+00 5.58117151e-01
-1.01271771e-01 -9.77334321e-01 -5.88214993e-01 9.43662822e-02
1.24423012e-01 -7.56932855e-01 -1.45593241e-01 5.77781975e-01
1.99610025e-01 3.33417177e-01 -1.27196610e-01 -2.66049623e-01
-1.29838622e+00 -5.02192192e-02 2.22842142e-01 -1.76261306e-01
2.12854832e-01 5.32824159e-01 2.65583098e-01 -7.10412979e-01
-1.54100552e-01 -3.08540821e-01 -5.22270262e-01 7.60955632e-01
4.33969200e-01 3.21406513e-01 -4.93080646e-01 -7.97459960e-01
-1.67311832e-01 4.12291974e-01 -2.87164152e-01 -3.69012833e-01
1.07318997e+00 -2.00949833e-01 -2.51024008e-01 7.61165082e-01
1.20124352e+00 6.92209065e-01 -6.99913919e-01 -4.11183268e-01
3.63948196e-01 -3.79090846e-01 -3.54083389e-01 -8.00476551e-01
-3.86916667e-01 2.88934648e-01 -1.87057443e-02 7.78555274e-01
4.13713485e-01 1.57807410e-01 6.65162027e-01 4.40352291e-01
3.43720824e-01 -1.21864283e+00 5.93184419e-02 6.10799372e-01
6.55687034e-01 -1.34395218e+00 1.41807184e-01 -6.35650679e-02
-8.96154404e-01 7.97476888e-01 7.98202932e-01 1.30008310e-01
8.94864023e-01 -1.55436277e-01 7.03069508e-01 -1.84207663e-01
-1.09891796e+00 1.51931778e-01 5.35095632e-01 8.84534419e-02
1.10373938e+00 4.32999045e-01 -6.21329665e-01 4.73194897e-01
-8.85757923e-01 -1.87771991e-01 4.90590662e-01 7.26028740e-01
-8.99568260e-01 -1.30611372e+00 -2.64276028e-01 8.47398579e-01
-5.54193199e-01 9.38925892e-02 -1.04450762e+00 5.28465569e-01
9.48232338e-02 1.38251054e+00 1.74741954e-01 -7.80882359e-01
2.09345251e-01 4.16326642e-01 -4.57210809e-01 -9.79988933e-01
-8.10155988e-01 -1.74337760e-01 6.54945016e-01 -3.40665758e-01
-4.96982157e-01 -8.65453601e-01 -6.82095528e-01 -8.02817285e-01
1.44460991e-01 4.94243711e-01 5.48868895e-01 1.00673664e+00
4.19563979e-01 -3.82150054e-01 8.48434687e-01 -4.18846577e-01
-5.29731631e-01 -1.08176434e+00 -3.59364718e-01 9.27396178e-01
3.13989520e-01 -2.23101407e-01 -5.26414812e-01 -7.65679404e-02] | [9.133151054382324, 10.094581604003906] |
7ffe3bba-b86c-4e67-bfd1-a5eae879026c | compiler-optimization-for-quantum-computing | 2212.04508 | null | https://arxiv.org/abs/2212.04508v2 | https://arxiv.org/pdf/2212.04508v2.pdf | Compiler Optimization for Quantum Computing Using Reinforcement Learning | Any quantum computing application, once encoded as a quantum circuit, must be compiled before being executable on a quantum computer. Similar to classical compilation, quantum compilation is a sequential process with many compilation steps and numerous possible optimization passes. Despite the similarities, the development of compilers for quantum computing is still in its infancy -- lacking mutual consolidation on the best sequence of passes, compatibility, adaptability, and flexibility. In this work, we take advantage of decades of classical compiler optimization and propose a reinforcement learning framework for developing optimized quantum circuit compilation flows. Through distinct constraints and a unifying interface, the framework supports the combination of techniques from different compilers and optimization tools in a single compilation flow. Experimental evaluations show that the proposed framework -- set up with a selection of compilation passes from IBM's Qiskit and Quantinuum's TKET -- significantly outperforms both individual compilers in 73% of cases regarding the expected fidelity. The framework is available on GitHub (https://github.com/cda-tum/MQTPredictor) as part of the Munich Quantum Toolkit (MQT). | ['Robert Wille', 'Lukas Burgholzer', 'Nils Quetschlich'] | 2022-12-08 | null | null | null | null | ['compiler-optimization'] | ['computer-code'] | [ 9.52541009e-02 -9.74518508e-02 -9.84272286e-02 -2.48838186e-01
-1.11863351e+00 -6.93678379e-01 6.35836661e-01 3.24520826e-01
-3.55768055e-01 8.62793386e-01 -1.60014525e-01 -8.23011577e-01
9.49365422e-02 -1.07172048e+00 -7.76444197e-01 -7.70911813e-01
-7.10549504e-02 6.15918696e-01 -2.21081495e-01 -6.16349876e-01
6.12412572e-01 1.57879382e-01 -1.33237410e+00 6.42675012e-02
1.12504780e+00 5.14647245e-01 9.19644311e-02 1.07028937e+00
-1.17058540e-02 7.13971555e-01 -1.43450156e-01 -6.16566598e-01
4.60962296e-01 -6.74535334e-01 -9.21936274e-01 -5.66046655e-01
6.83526024e-02 -7.54933478e-03 -6.41923428e-01 1.27068162e+00
5.26131809e-01 1.01594012e-02 2.18803570e-01 -8.77375245e-01
-5.01828671e-01 8.79693389e-01 2.72125334e-01 -3.29307048e-03
5.29999077e-01 5.27438521e-01 1.37251699e+00 -2.54615724e-01
7.15815544e-01 8.09752762e-01 4.64273483e-01 4.17278409e-01
-1.37563360e+00 -4.64489400e-01 -7.08473086e-01 3.44454676e-01
-1.48794782e+00 -6.07434273e-01 4.43735570e-01 -2.77886391e-01
1.68305790e+00 1.83890179e-01 7.33269036e-01 7.46879518e-01
7.15377271e-01 3.98167491e-01 1.44506204e+00 -7.97254682e-01
5.25366008e-01 2.24971518e-01 2.06514627e-01 9.79874909e-01
-9.39389020e-02 7.89166629e-01 -4.53474224e-01 -1.03901006e-01
1.79747760e-01 -5.79387069e-01 1.64814606e-01 -2.23716825e-01
-1.32890952e+00 9.31320727e-01 4.03237551e-01 3.60616833e-01
-9.09938961e-02 5.70871770e-01 6.66979313e-01 6.48314893e-01
-3.12962383e-02 6.41238511e-01 -1.30377695e-01 -5.81694841e-01
-8.42177629e-01 4.31363523e-01 1.05773139e+00 1.03719902e+00
1.21256983e+00 -8.62695873e-02 -6.21036179e-02 8.46493989e-02
8.37422907e-02 4.58962113e-01 1.42338321e-01 -9.46444154e-01
1.74228728e-01 1.77907810e-01 -1.55255735e-01 -4.32700396e-01
-4.77654070e-01 -4.45158213e-01 -6.31363034e-01 9.30377245e-02
5.88507056e-02 -2.20135540e-01 -3.89195889e-01 1.43153524e+00
3.58744234e-01 7.55280554e-02 3.41578037e-01 7.58818805e-01
6.01446509e-01 8.40873182e-01 -2.93077320e-01 -1.71560422e-01
1.17920005e+00 -1.02474880e+00 -4.83807892e-01 1.60142615e-01
1.18884230e+00 -9.29431856e-01 8.75014484e-01 4.55707431e-01
-1.11106908e+00 -4.99231517e-01 -1.40693915e+00 -2.64501721e-01
-4.57910269e-01 -1.90604419e-01 1.10717201e+00 1.23082078e+00
-1.29259086e+00 1.17219698e+00 -8.07924747e-01 -2.39066795e-01
-1.16122916e-01 7.59034395e-01 -1.64318338e-01 8.31324086e-02
-1.16420138e+00 1.10404217e+00 7.58452296e-01 -7.75169283e-02
-6.80003881e-01 -4.23496217e-01 -5.10841012e-01 -1.75226226e-01
2.15815321e-01 -1.15489876e+00 1.38345063e+00 -5.08108258e-01
-2.31109953e+00 6.64416969e-01 -3.20869572e-02 -5.92342734e-01
3.53802383e-01 1.62895456e-01 -3.19523394e-01 -9.22717601e-02
-1.61293764e-02 3.43753159e-01 4.76031274e-01 -5.00969172e-01
-5.52695751e-01 -1.20980076e-01 2.99768358e-01 2.08585605e-01
4.02142614e-01 2.94315498e-02 -1.68162227e-01 1.52529031e-01
-9.14220139e-02 -1.27846360e+00 -4.51144516e-01 -1.01266396e+00
-4.47389513e-01 5.67354681e-03 9.22266319e-02 -2.26418570e-01
1.06390297e+00 -1.88307643e+00 5.87713182e-01 1.70018703e-01
-1.71611048e-02 -1.12872846e-01 2.19274610e-01 8.67966831e-01
-1.71630550e-02 -3.42503451e-02 -2.47393712e-01 -1.80231303e-01
5.48651874e-01 1.09405503e-01 -1.05312377e-01 7.57717252e-01
-3.43160182e-02 1.01043379e+00 -1.16654718e+00 -2.65559196e-01
5.50186813e-01 1.62176251e-01 -1.08570254e+00 2.16455170e-04
-4.08677161e-01 7.59491503e-01 -4.56680834e-01 6.70237541e-01
5.95165670e-01 -1.97079778e-02 2.88235426e-01 6.58307150e-02
-6.26820683e-01 7.73717999e-01 -1.22284436e+00 2.21318841e+00
-6.52803481e-01 3.72675508e-01 -5.99586889e-02 -1.08946872e+00
7.11032212e-01 1.47001818e-01 2.46194094e-01 -8.89165223e-01
4.12138641e-01 7.09582865e-01 4.16352183e-01 -2.20605627e-01
1.00468254e+00 -2.12188929e-01 -4.49318677e-01 7.33436882e-01
6.06322587e-01 -9.02042806e-01 7.28753805e-01 3.13578755e-01
1.07658052e+00 4.65067118e-01 4.95750397e-01 -4.64236200e-01
7.14367211e-01 2.26860166e-01 2.49524564e-01 9.88322735e-01
-2.67825931e-01 7.61368722e-02 5.85036635e-01 -3.83298397e-01
-1.38310707e+00 -8.48804533e-01 -4.34729517e-01 9.41528976e-01
9.27325860e-02 -1.10095894e+00 -8.91689479e-01 -1.48174733e-01
-4.00449812e-01 9.78589058e-01 -1.41097128e-01 -1.43689051e-01
-3.53464484e-01 -1.04491591e+00 6.56437039e-01 -3.29014659e-01
3.02485973e-01 -9.82382655e-01 -6.03251755e-01 2.58401066e-01
3.25082600e-01 -1.04927897e+00 -4.16258834e-02 7.37110615e-01
-7.20295489e-01 -7.11493909e-01 1.60017788e-01 -3.72541636e-01
1.97764039e-01 -1.79929748e-01 1.17416871e+00 4.88753011e-03
-3.01154256e-01 5.23670204e-02 -4.86103863e-01 6.53339997e-02
-1.26652884e+00 3.57335091e-01 1.04412258e-01 -5.36026597e-01
6.96739480e-02 -7.16922998e-01 -4.18594092e-01 -4.26599115e-01
-5.87487280e-01 2.44556412e-01 4.52872396e-01 1.12598717e+00
4.63186830e-01 3.53443511e-02 2.06186082e-02 -9.74030077e-01
4.65102941e-01 -1.89667344e-01 -1.08592176e+00 1.46162257e-01
-7.93665409e-01 4.33377951e-01 9.32859838e-01 4.01803285e-01
-6.27130806e-01 1.30149677e-01 -6.06356740e-01 3.08550060e-01
-1.23756491e-01 4.03846115e-01 1.14731327e-01 -5.97495317e-01
8.44033360e-01 4.81417835e-01 -1.87975705e-01 1.68750748e-01
7.24490225e-01 6.18334115e-01 3.19705904e-01 -9.54592884e-01
6.84664071e-01 2.83839162e-02 2.26151243e-01 -6.83756232e-01
-3.02701384e-01 -1.94075450e-01 -7.47879148e-01 2.96386448e-03
6.98107421e-01 -7.93082893e-01 -8.74966979e-01 2.17258334e-01
-1.05512166e+00 -3.37952852e-01 -1.99135289e-01 5.30491114e-01
-7.95759439e-01 5.25616109e-01 -7.11268187e-01 -6.12881660e-01
-5.19478321e-01 -1.95477343e+00 9.69246209e-01 4.01647508e-01
3.44452187e-02 -8.54998589e-01 4.00283277e-01 4.02092725e-01
3.80486816e-01 -1.10238545e-01 8.33846092e-01 -3.45181167e-01
-1.01534808e+00 -2.89747566e-01 1.39670596e-01 2.51093298e-01
-3.56705129e-01 3.04769278e-01 -9.79953229e-01 -2.66331792e-01
-2.53810644e-01 -4.60308880e-01 6.31282389e-01 7.60648847e-02
7.04603434e-01 4.07645032e-02 1.09107986e-01 1.03850925e+00
1.71940923e+00 7.66851828e-02 6.20328426e-01 5.06502271e-01
3.73787761e-01 1.41172251e-02 2.41181836e-01 4.14979160e-01
3.08026522e-01 6.48900151e-01 4.60411012e-01 5.13871014e-01
3.37448210e-01 -2.47198865e-02 6.61125422e-01 1.54180801e+00
-2.69538432e-01 3.17892820e-01 -9.83817279e-01 -1.19015686e-01
-1.57751381e+00 -9.98258889e-01 -3.76563013e-01 2.29814386e+00
7.14525998e-01 1.70591190e-01 -9.97153297e-02 -6.22701012e-02
2.70150751e-01 -6.99408865e-03 -1.84192970e-01 -1.17995596e+00
1.96288288e-01 8.48890960e-01 7.02943623e-01 8.50706697e-01
-9.16238666e-01 1.28321636e+00 5.98595619e+00 9.43990946e-01
-1.22734487e+00 3.99163246e-01 1.93278626e-01 1.44299641e-01
-3.20410073e-01 6.79222584e-01 -6.80410028e-01 3.10319573e-01
1.65566945e+00 -4.64053303e-01 1.14089501e+00 7.43626833e-01
-4.08613756e-02 -2.10552365e-01 -1.21035421e+00 7.80697107e-01
-4.75451440e-01 -1.83149171e+00 -4.65463758e-01 -6.08238298e-03
7.74132311e-01 5.90426207e-01 -3.78253907e-02 7.16653466e-01
3.08736503e-01 -1.02549934e+00 1.03494418e+00 1.91854268e-01
6.99175835e-01 -9.28713322e-01 7.29727924e-01 3.19559515e-01
-8.24069977e-01 4.38210629e-02 -3.93257111e-01 -4.28560555e-01
5.52369398e-04 5.16338348e-01 -7.98563957e-01 1.15812039e+00
3.17312390e-01 5.78623831e-01 -4.51099604e-01 6.94570541e-01
-1.72209263e-01 4.53882366e-01 -9.98550355e-02 -3.65623415e-01
4.12631333e-01 -9.97324109e-01 4.28309947e-01 1.42114937e+00
3.55568200e-01 -7.27552339e-04 1.23406544e-01 1.12937450e+00
4.61241417e-02 2.68805802e-01 -5.71120501e-01 -3.14351171e-01
3.33193481e-01 1.46548188e+00 -6.73687577e-01 -3.99170548e-01
-3.80997509e-01 9.76104975e-01 3.57274383e-01 -1.23345152e-01
-1.08143723e+00 -5.03961086e-01 2.15189785e-01 -6.03037298e-01
6.95393682e-02 -3.87727678e-01 -4.38572943e-01 -1.25520098e+00
-3.27363849e-01 -1.14518225e+00 -8.93629417e-02 -4.17742103e-01
-6.30532265e-01 6.95805669e-01 -3.00651044e-01 -9.59703624e-01
-3.99818063e-01 -6.55200899e-01 -4.78394538e-01 9.46820915e-01
-1.27452111e+00 -8.02814543e-01 2.04902098e-01 1.96829498e-01
-3.78727391e-02 -2.82076210e-01 1.35355949e+00 3.19092333e-01
-7.41740167e-01 5.07718742e-01 6.59115195e-01 -3.87167543e-01
2.83523470e-01 -1.55889380e+00 6.24655187e-01 7.78241038e-01
2.53313363e-01 9.17251647e-01 9.92187679e-01 -1.75893307e-01
-2.34206629e+00 -7.05888271e-01 8.06884527e-01 -2.22963512e-01
1.14158857e+00 -5.41528404e-01 -3.25434983e-01 5.07453442e-01
5.96867323e-01 -2.76327968e-01 6.86260819e-01 4.23921794e-01
-1.48463726e-01 -3.46966274e-02 -7.49965012e-01 4.51496899e-01
6.70489609e-01 -9.73044515e-01 -2.49822974e-01 8.74875546e-01
5.17039895e-01 -8.44672799e-01 -1.03096795e+00 2.66658403e-02
4.63374317e-01 -1.44110596e+00 7.44395196e-01 -4.34706286e-02
5.58794379e-01 -3.61987531e-01 -3.97570908e-01 -1.25818527e+00
-1.32167980e-01 -1.30934572e+00 1.29998907e-01 7.55585670e-01
4.05366331e-01 -7.37617016e-01 4.53615427e-01 1.05832309e-01
-5.44947326e-01 -4.64948744e-01 -1.13329446e+00 -6.34332299e-01
4.10782725e-01 -7.70731091e-01 5.06676137e-01 6.47221446e-01
4.39707369e-01 5.17027020e-01 -2.49628261e-01 2.00814709e-01
6.12681091e-01 3.59254479e-01 9.73187089e-01 -4.45410430e-01
-8.87464523e-01 -6.93559825e-01 -5.08005798e-01 -5.43119729e-01
1.21641509e-01 -1.58485866e+00 7.20609725e-02 -9.44222331e-01
2.11482242e-01 -4.97108877e-01 -1.60152391e-01 7.29232728e-02
1.44772902e-01 6.86393771e-03 2.95653164e-01 1.09991198e-02
-7.01447487e-01 5.52115262e-01 9.77227211e-01 -4.66640294e-03
-2.16747537e-01 -1.43483773e-01 -4.67726976e-01 1.81145325e-01
8.99027884e-01 -6.35567129e-01 -5.53612690e-03 -8.83938670e-02
5.91154695e-01 3.71585995e-01 1.60019085e-01 -1.33679438e+00
3.74658018e-01 1.10866763e-01 -5.06042719e-01 -2.62299895e-01
2.23287269e-01 -9.19912234e-02 3.82007271e-01 7.69298613e-01
5.84961940e-03 1.74040034e-01 5.68155833e-02 1.89783439e-01
-1.86380260e-02 -5.54229498e-01 7.01348126e-01 -2.66126812e-01
-7.07980335e-01 3.61409709e-02 -3.21619332e-01 -1.25034213e-01
8.59238505e-01 1.06311157e-01 -3.32994819e-01 2.07100436e-01
-5.20437658e-01 -4.23889086e-02 8.69060695e-01 -1.41143039e-01
-3.35867628e-02 -8.52626503e-01 -6.67462230e-01 2.34379694e-01
2.28250306e-02 -3.95932823e-01 3.04292947e-01 1.01517200e+00
-1.17140174e+00 8.45375121e-01 -3.64020795e-01 -6.54117584e-01
-7.50407636e-01 4.47183639e-01 4.73938972e-01 -4.49370861e-01
-6.00044131e-01 7.22367346e-01 -2.75285691e-01 -9.66455758e-01
-3.39943230e-01 -3.71888667e-01 6.15742207e-01 -4.72704858e-01
2.34640211e-01 5.15411422e-02 5.53767204e-01 -5.62208116e-01
-2.89136320e-01 3.02618653e-01 1.67030364e-01 -2.44131282e-01
1.21743500e+00 2.10367188e-01 -5.82498014e-01 4.32125360e-01
1.13676918e+00 2.98594800e-03 -5.39727867e-01 1.69260129e-01
-3.20858732e-02 -7.29360655e-02 3.64191324e-01 -5.47581077e-01
-5.66101670e-01 9.45605695e-01 5.27633011e-01 1.14824295e-01
7.29367733e-01 -2.87715495e-01 5.25270998e-01 7.78469861e-01
9.02683794e-01 -1.11117387e+00 -5.03254354e-01 8.47457051e-01
2.32521102e-01 -1.06021428e+00 2.04083756e-01 -8.25552046e-02
-3.92436236e-01 1.36340904e+00 -5.39810844e-02 -3.80209595e-01
2.40257055e-01 2.32819855e-01 -3.50246668e-01 -9.50376168e-02
-8.67827177e-01 -2.24811792e-01 -8.71693194e-02 2.49722376e-01
8.02241981e-01 5.55499852e-01 -5.52621901e-01 2.88460076e-01
-7.29747653e-01 -2.86727920e-02 9.53357637e-01 9.94080544e-01
-3.12472194e-01 -1.69148028e+00 -3.12893361e-01 1.80983707e-01
-2.71743894e-01 -5.23830593e-01 1.28728747e-01 7.22787023e-01
2.46907711e-01 7.11445987e-01 -4.62688774e-01 -6.56359494e-01
-2.75701215e-03 1.88416287e-01 1.11076248e+00 -7.04200923e-01
-1.06880045e+00 -1.47222355e-01 2.57854402e-01 -6.20650649e-01
-3.72445017e-01 -8.31031680e-01 -1.39159560e+00 -9.32198346e-01
-3.99728626e-01 5.24587572e-01 9.18613672e-01 9.94169652e-01
2.61407584e-01 6.19861484e-01 5.88725686e-01 -9.71128345e-01
-8.13383937e-01 -6.03522241e-01 -6.90532446e-01 -3.67942989e-01
1.12705261e-01 -3.60719234e-01 -1.26955668e-02 -2.54013747e-01] | [5.606978893280029, 4.933511257171631] |
cfed94db-002c-4fe5-bf16-f1a8914fa4a4 | sinfusion-training-diffusion-models-on-a | 2211.11743 | null | https://arxiv.org/abs/2211.11743v3 | https://arxiv.org/pdf/2211.11743v3.pdf | SinFusion: Training Diffusion Models on a Single Image or Video | Diffusion models exhibited tremendous progress in image and video generation, exceeding GANs in quality and diversity. However, they are usually trained on very large datasets and are not naturally adapted to manipulate a given input image or video. In this paper we show how this can be resolved by training a diffusion model on a single input image or video. Our image/video-specific diffusion model (SinFusion) learns the appearance and dynamics of the single image or video, while utilizing the conditioning capabilities of diffusion models. It can solve a wide array of image/video-specific manipulation tasks. In particular, our model can learn from few frames the motion and dynamics of a single input video. It can then generate diverse new video samples of the same dynamic scene, extrapolate short videos into long ones (both forward and backward in time) and perform video upsampling. Most of these tasks are not realizable by current video-specific generation methods. | ['Michal Irani', 'Niv Haim', 'Yaniv Nikankin'] | 2022-11-21 | null | null | null | null | ['video-generation', 'image-manipulation'] | ['computer-vision', 'computer-vision'] | [ 4.47629482e-01 1.21667668e-01 -1.71226546e-01 1.32270681e-04
-5.92319071e-01 -6.49104238e-01 7.45100379e-01 -8.17283034e-01
-1.66212440e-01 1.01655281e+00 1.20779857e-01 1.02404198e-02
2.86793858e-01 -7.79381573e-01 -1.02723908e+00 -9.81471121e-01
1.55630350e-01 4.20238972e-01 9.03798789e-02 1.00113340e-02
3.82520934e-03 5.14934659e-01 -1.49846673e+00 3.26217890e-01
5.40746212e-01 9.28163588e-01 5.86035073e-01 1.31769943e+00
7.14629441e-02 1.19615233e+00 -6.64729238e-01 -6.31444007e-02
4.30439949e-01 -1.00952947e+00 -4.89938229e-01 4.47096378e-01
6.00767672e-01 -7.87956953e-01 -8.31444263e-01 9.11868989e-01
4.24104899e-01 1.89946488e-01 5.75103283e-01 -1.32345724e+00
-9.83554959e-01 3.58605534e-01 -5.20483375e-01 2.62997121e-01
3.77024442e-01 6.78828537e-01 4.68104064e-01 -6.95621729e-01
1.14829731e+00 1.09095156e+00 2.37839401e-01 1.14595354e+00
-1.35831499e+00 -5.48269153e-01 1.53733894e-01 -5.52174486e-02
-1.00627613e+00 -5.43971121e-01 7.25888848e-01 -4.92426872e-01
5.79155147e-01 1.06130220e-01 1.07229185e+00 1.60907149e+00
1.70360491e-01 9.93036926e-01 9.23340142e-01 -1.52051017e-01
3.22353512e-01 -1.79915205e-01 -7.57747710e-01 6.39181614e-01
2.87854355e-02 2.90435523e-01 -6.79362833e-01 7.68427551e-02
1.38555348e+00 6.98541254e-02 -6.51274920e-01 -3.84643316e-01
-1.48589277e+00 8.08968842e-01 2.25659966e-01 1.55144572e-01
-4.52346057e-01 6.21821105e-01 -1.95407644e-02 6.71917915e-01
3.11957419e-01 4.16723728e-01 -2.12723821e-01 -2.84671843e-01
-1.32881749e+00 5.21012723e-01 6.63058221e-01 1.08985865e+00
7.89757252e-01 6.12346292e-01 -2.92330891e-01 2.22829938e-01
-3.11280191e-02 8.05508077e-01 6.22135699e-01 -1.49898362e+00
3.08550537e-01 -1.23470962e-01 2.02762201e-01 -7.31820643e-01
3.36509272e-02 1.10526187e-02 -8.94997180e-01 4.11759645e-01
5.35366178e-01 -5.80531716e-01 -1.28049707e+00 1.82601440e+00
2.92417645e-01 3.76918763e-01 -8.82708468e-04 7.76749313e-01
4.99678940e-01 9.47218955e-01 -1.73060194e-01 -5.22269428e-01
6.03985608e-01 -1.16430044e+00 -6.35616302e-01 -2.78528541e-01
2.53684551e-01 -6.16559148e-01 8.49512577e-01 4.76494521e-01
-1.63281107e+00 -6.63693428e-01 -8.56961191e-01 -1.08818874e-01
-1.46579614e-03 -1.80038080e-01 4.20744866e-01 3.98692727e-01
-1.39793384e+00 5.75934350e-01 -8.43191862e-01 -1.46985278e-01
5.49942076e-01 2.70358652e-01 -3.55377167e-01 -4.51823354e-01
-8.64042044e-01 5.93159914e-01 1.13632254e-01 -1.64956123e-01
-1.64114940e+00 -7.70386040e-01 -7.94490695e-01 -1.36609048e-01
3.08451563e-01 -1.30836105e+00 1.24020815e+00 -1.54270661e+00
-1.74349988e+00 5.82583904e-01 -2.32253805e-01 -5.40893734e-01
8.30170572e-01 -9.25336629e-02 -1.66228950e-01 3.89123589e-01
-1.70776062e-02 1.08186722e+00 1.57911754e+00 -1.23532450e+00
-4.98233169e-01 -1.61758021e-01 2.98370570e-01 1.94336534e-01
-2.02521607e-02 -2.67789900e-01 -4.57002729e-01 -8.92402828e-01
-3.43633801e-01 -1.11125863e+00 -4.31984305e-01 3.08723420e-01
-1.69531420e-01 5.07267475e-01 1.11785936e+00 -6.01704955e-01
9.49998021e-01 -2.03225732e+00 8.18190277e-01 -2.61879385e-01
2.61070400e-01 2.61032879e-01 -4.75300521e-01 4.92617846e-01
8.12997147e-02 7.28825927e-02 -1.70587078e-01 -3.13359797e-01
-3.62647444e-01 3.20083767e-01 -3.50801915e-01 2.70785600e-01
1.55884966e-01 1.20624030e+00 -1.14414549e+00 -3.45383018e-01
8.72978270e-02 6.20445192e-01 -8.17030489e-01 3.75939608e-01
-6.09122634e-01 1.07918310e+00 -4.01719928e-01 4.56900924e-01
4.01826024e-01 -2.17443645e-01 3.16258296e-02 -7.87918717e-02
2.65863627e-01 -2.91200399e-01 -8.98791552e-01 1.91931736e+00
-4.79886562e-01 8.34435284e-01 1.15154080e-01 -8.07971299e-01
4.46721494e-01 4.78021502e-01 5.57877183e-01 -3.51749510e-01
-6.33542910e-02 5.89036895e-03 -4.48385924e-02 -8.19370687e-01
4.08684224e-01 -3.72187197e-01 1.56093553e-01 6.24537945e-01
4.00252640e-01 -5.48639655e-01 4.95757073e-01 3.16618681e-01
1.09072709e+00 3.23862463e-01 -6.43970966e-02 1.73330262e-01
1.48952484e-01 -8.00573975e-02 3.67434829e-01 9.18774545e-01
1.10829681e-01 1.02138937e+00 3.64428401e-01 -4.40556824e-01
-1.49261642e+00 -1.22469115e+00 4.65938151e-01 6.10014617e-01
-1.39802918e-01 -3.38169307e-01 -9.45520401e-01 -6.51698947e-01
-2.00885430e-01 4.52929258e-01 -6.99312568e-01 -2.21700326e-01
-6.41646087e-01 -4.76455539e-01 3.38053823e-01 4.11240250e-01
4.03173923e-01 -1.17174721e+00 -5.75230539e-01 1.69798195e-01
-7.95894936e-02 -1.07025313e+00 -6.68437421e-01 -2.93067932e-01
-9.22249854e-01 -9.29820597e-01 -1.31191671e+00 -7.31772542e-01
8.23174477e-01 2.44103864e-01 1.09209728e+00 -1.01284780e-01
-2.54576266e-01 5.82882345e-01 -1.41130075e-01 -5.69300726e-03
-6.05015874e-01 -1.03163041e-01 7.76418373e-02 1.61518499e-01
-3.00799817e-01 -8.40612054e-01 -7.44139671e-01 7.17471838e-02
-1.30435097e+00 1.63965255e-01 4.52236682e-01 1.09046257e+00
6.26843154e-01 1.06823862e-01 4.49368387e-01 -9.10147607e-01
5.74686527e-01 -5.66693366e-01 -4.73784953e-01 4.87529412e-02
-1.80212781e-01 -4.36993502e-02 7.17672229e-01 -8.61387849e-01
-1.12100160e+00 1.09108523e-01 3.36539857e-02 -9.94990408e-01
-7.00412318e-02 3.06992501e-01 -9.13846642e-02 8.12840357e-04
6.31470799e-01 4.73230630e-01 1.50387660e-01 -1.30869493e-01
6.53036714e-01 1.71953633e-01 5.56564033e-01 -5.43282151e-01
8.87580812e-01 6.65202081e-01 -1.11262053e-01 -8.95719171e-01
-5.95159292e-01 1.89016908e-01 -6.41292393e-01 -3.19169402e-01
1.04644787e+00 -9.38934803e-01 -2.78983414e-01 7.57312655e-01
-9.87480938e-01 -8.85493636e-01 -8.48951042e-01 3.18913370e-01
-9.28381801e-01 1.65901169e-01 -7.92034566e-01 -5.57109296e-01
6.31321296e-02 -1.19015384e+00 1.13242769e+00 2.69632518e-01
-1.75372750e-01 -1.09384406e+00 4.07625400e-02 1.30817384e-01
5.49512506e-01 2.92186558e-01 5.75878680e-01 2.27750406e-01
-1.12764895e+00 -5.30615561e-02 1.81579828e-01 5.54392695e-01
2.06110507e-01 2.63602853e-01 -6.39377713e-01 -5.20134985e-01
2.54048914e-01 -4.50412601e-01 9.18660522e-01 6.36091471e-01
1.11774838e+00 -5.05085766e-01 -1.81372330e-01 9.60400641e-01
1.31365526e+00 3.74026060e-01 6.54876769e-01 -2.74126902e-02
7.59910285e-01 1.84512869e-01 2.80783325e-01 2.43274420e-01
4.36870009e-02 5.41700006e-01 2.87362754e-01 -3.89086604e-02
-5.24648905e-01 -5.42622149e-01 6.99492693e-01 6.58161283e-01
-2.89164364e-01 -5.83692253e-01 -5.77286184e-01 5.48767149e-01
-1.68471169e+00 -1.41349077e+00 2.66569346e-01 1.92276752e+00
6.72442079e-01 -1.95445679e-02 1.88345537e-01 -1.76822230e-01
5.04787683e-01 3.14165056e-01 -9.78141844e-01 -2.91460492e-02
-3.59429181e-01 1.26467645e-01 2.89357185e-01 3.88652742e-01
-6.97056592e-01 9.97519612e-01 7.65444708e+00 6.07291400e-01
-1.32271564e+00 5.76189421e-02 7.42759168e-01 -6.31908953e-01
-5.95647871e-01 -9.03400853e-02 -5.66542625e-01 5.06099105e-01
9.48201537e-01 -3.59311670e-01 7.79464126e-01 6.33856654e-01
1.82884768e-01 -1.32616252e-01 -1.23180628e+00 1.07490039e+00
2.06032068e-01 -1.67790079e+00 3.94447297e-01 2.89923638e-01
1.36510968e+00 -1.31108180e-01 4.82083648e-01 1.24998696e-01
4.21143383e-01 -1.14386868e+00 7.85800934e-01 6.53881729e-01
9.60269272e-01 -3.53262842e-01 1.68875992e-01 5.40400803e-01
-5.76975763e-01 -1.78672299e-01 -2.45633498e-01 9.05631855e-03
5.87200701e-01 4.13211197e-01 -2.17676297e-01 1.58082336e-01
5.87938190e-01 8.25714111e-01 -2.10479409e-01 7.06368268e-01
-2.39083827e-01 5.27326703e-01 -2.26279557e-01 4.43674982e-01
3.17702115e-01 -4.63822305e-01 5.52628875e-01 8.04881871e-01
8.15775633e-01 1.02811173e-01 1.20993173e-02 8.51254702e-01
-2.27293923e-01 -4.68376964e-01 -1.02046573e+00 -3.59188944e-01
9.55053642e-02 9.65309262e-01 -4.16057199e-01 -6.49465322e-01
-3.52150321e-01 1.43234205e+00 2.00372905e-01 8.07346523e-01
-9.85933006e-01 1.36852384e-01 7.13109612e-01 1.77508712e-01
6.11999273e-01 -2.81311572e-01 2.39488617e-01 -1.63555455e+00
-1.10886969e-01 -1.13038862e+00 1.00107873e-02 -1.12979543e+00
-1.05661500e+00 6.18962765e-01 -6.51154518e-02 -1.14515793e+00
-6.76939428e-01 -3.08486730e-01 -5.14285743e-01 7.01815605e-01
-1.12220740e+00 -1.06291831e+00 -2.70286202e-01 8.89891744e-01
8.02733481e-01 -2.48812705e-01 6.26470983e-01 2.32189074e-01
-3.49860907e-01 2.01727480e-01 1.92375690e-01 -7.26549402e-02
4.73959118e-01 -9.92673397e-01 4.48230267e-01 8.12305629e-01
2.47572541e-01 3.93915236e-01 6.81252360e-01 -5.80250144e-01
-1.58759594e+00 -9.80995059e-01 3.96621227e-01 -4.21423852e-01
5.90179086e-01 -2.44996443e-01 -5.57259321e-01 9.73052561e-01
3.79719853e-01 2.60950476e-01 2.86448151e-01 -6.41726315e-01
-3.37820649e-02 8.50775465e-02 -1.21397495e+00 8.11928988e-01
1.32107973e+00 -6.79116249e-01 -1.13555491e-01 4.08693433e-01
7.01789021e-01 -6.53274059e-01 -7.47083127e-01 2.95718927e-02
4.92593408e-01 -1.11169481e+00 1.01539004e+00 -7.78009892e-01
8.84461343e-01 -3.08343172e-01 -1.41979396e-01 -1.48799992e+00
-1.56228542e-01 -1.04559505e+00 -6.48202240e-01 1.10463142e+00
2.77945191e-01 -2.32975438e-01 1.00289547e+00 5.45288861e-01
1.41632855e-01 -5.05905569e-01 -5.90747833e-01 -8.06817830e-01
1.56657442e-01 -1.90898836e-01 6.07933760e-01 7.27056980e-01
-3.78788024e-01 2.66022980e-01 -8.59320402e-01 -2.42674887e-01
6.53752685e-01 1.19319633e-01 9.99711931e-01 -5.00699639e-01
-6.62387669e-01 -3.56593281e-01 -4.33640391e-01 -1.40019774e+00
5.93466568e-04 -6.23210728e-01 4.03268728e-03 -1.36360645e+00
6.44979104e-02 -1.64755329e-01 1.85573682e-01 -7.02240318e-02
-5.42595126e-02 3.58217418e-01 5.04109144e-01 2.95392662e-01
-2.28947520e-01 6.54067814e-01 1.79550338e+00 -3.16904813e-01
-2.50085816e-02 8.28090197e-05 -5.24759829e-01 6.51586831e-01
6.36335135e-01 -3.67988139e-01 -8.64711821e-01 -8.19324851e-01
1.62602440e-01 5.94043314e-01 2.25932762e-01 -1.06764162e+00
1.51971444e-01 -3.58799696e-01 6.22924089e-01 -3.81988883e-02
6.15532994e-01 -6.05947316e-01 7.47060061e-01 4.37126786e-01
-4.09645110e-01 1.82500467e-01 -1.18366204e-01 8.06389809e-01
-3.65864217e-01 -1.09469756e-01 7.81777263e-01 -5.94215572e-01
-5.75305402e-01 7.06964612e-01 -5.60075045e-01 1.93159744e-01
1.32557929e+00 -2.79265583e-01 -9.51795503e-02 -8.84219468e-01
-1.04970157e+00 -4.98295240e-02 9.64539945e-01 3.53912681e-01
6.69209719e-01 -1.41281223e+00 -5.56275368e-01 4.03045028e-01
-3.98045987e-01 2.49709070e-01 4.22727048e-01 5.85732222e-01
-6.49922073e-01 -4.24344949e-02 -4.49138135e-01 -5.53157687e-01
-9.12750542e-01 9.44185615e-01 3.59165251e-01 -1.91739619e-01
-6.33970141e-01 9.86231267e-01 5.96545696e-01 1.68157056e-01
-4.54801582e-02 -1.34744912e-01 3.34424198e-01 -8.00505877e-02
7.86080420e-01 8.62073619e-03 -5.76721370e-01 -5.87274253e-01
3.32228273e-01 5.96103191e-01 6.09487221e-02 -4.00651157e-01
1.30629408e+00 -2.28778124e-01 1.89446479e-01 4.47893769e-01
1.15073729e+00 4.06603850e-02 -2.09604383e+00 1.19595677e-01
-5.32856584e-01 -6.54747784e-01 -1.19152181e-01 -4.33305085e-01
-1.60103118e+00 9.56179380e-01 1.53536603e-01 1.84441492e-01
1.13904572e+00 -8.95489380e-02 1.18585885e+00 4.04764712e-02
3.49787086e-01 -8.23990583e-01 3.87134641e-01 2.32149884e-01
7.98360407e-01 -9.56446052e-01 -2.17453167e-01 -6.81480467e-02
-7.90045202e-01 1.00111759e+00 4.73380834e-01 -3.49233419e-01
5.27791977e-01 5.42214274e-01 -1.62263773e-02 -6.95135593e-02
-1.04140663e+00 2.72329301e-02 2.05466691e-02 1.05449712e+00
2.14613557e-01 -3.39943498e-01 1.58048049e-01 -1.77139357e-01
-3.55038680e-02 4.55613583e-01 8.50160003e-01 8.41386259e-01
-1.42932579e-01 -1.07834113e+00 -1.31274834e-01 4.39875305e-01
-4.09455091e-01 -4.73765209e-02 -1.68491840e-01 5.80273449e-01
2.33420916e-02 6.15233481e-01 1.81427430e-02 -1.85682729e-01
-1.52628377e-01 -3.70381549e-02 9.64957356e-01 -5.72486281e-01
-2.52656251e-01 4.56990600e-02 -1.54600427e-01 -7.10380018e-01
-6.73365355e-01 -8.84687483e-01 -8.11393678e-01 -4.79424208e-01
-1.85748115e-02 -6.36542141e-02 4.33650732e-01 6.89756334e-01
3.58520031e-01 5.31093419e-01 4.95859236e-01 -1.38758731e+00
-3.38373154e-01 -6.15084231e-01 -7.04111814e-01 5.84242463e-01
7.50335038e-01 -2.47846916e-01 -3.81854773e-01 6.97596610e-01] | [10.861023902893066, -0.6587976813316345] |
73c04141-acf3-4ab9-bc7f-33f0b427ec8d | neural-headline-generation-on-abstract | null | null | https://aclanthology.org/D16-1112 | https://aclanthology.org/D16-1112.pdf | Neural Headline Generation on Abstract Meaning Representation | null | ['Tsutomu Hirao', 'Sho Takase', 'Masaaki Nagata', 'Jun Suzuki', 'Naoaki Okazaki'] | 2016-11-01 | null | null | null | emnlp-2016-11 | ['video-description', 'headline-generation'] | ['computer-vision', '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] | [-7.194465160369873, 3.620304584503174] |
5f1e87f7-db7b-4c85-80bf-91a650c0caaf | event-guided-person-re-identification-via | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Cao_Event-Guided_Person_Re-Identification_via_Sparse-Dense_Complementary_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Cao_Event-Guided_Person_Re-Identification_via_Sparse-Dense_Complementary_Learning_CVPR_2023_paper.pdf | Event-Guided Person Re-Identification via Sparse-Dense Complementary Learning | Video-based person re-identification (Re-ID) is a prominent computer vision topic due to its wide range of video surveillance applications. Most existing methods utilize spatial and temporal correlations in frame sequences to obtain discriminative person features. However, inevitable degradations, e.g., motion blur contained in frames often cause ambiguity texture noise and temporal disturbance, leading to the loss of identity-discriminating cues. Recently, a new bio-inspired sensor called event camera, which can asynchronously record intensity changes, brings new vitality to the Re-ID task. With the microsecond resolution and low latency, event cameras can accurately capture the movements of pedestrians even in the aforementioned degraded environments. Inspired by the properties of event cameras, in this work, we propose a Sparse-Dense Complementary Learning Framework, which effectively extracts identity features by fully exploiting the complementary information of dense frames and sparse events. Specifically, for frames, we build a CNN-based module to aggregate the dense features of pedestrian appearance step-by-step, while for event streams, we design a bio-inspired spiking neural backbone, which encodes event signals into sparse feature maps in a spiking form, to present the dynamic motion cues of pedestrians. Finally, a cross feature alignment module is constructed to complementarily fuse motion information from events and appearance cues from frames to enhance identity representation learning. Experiments on several benchmarks show that by employing events and SNN into Re-ID, our method significantly outperforms competitive methods. | ['Zheng-Jun Zha', 'Jiebo Luo', 'Kunyu Wang', 'Yukun Huang', 'Hongjian Liu', 'Xueyang Fu', 'Chengzhi Cao'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['person-re-identification'] | ['computer-vision'] | [ 1.59250364e-01 -9.58565176e-01 8.83977413e-02 -3.72701317e-01
-2.22480431e-01 -3.84875327e-01 6.01568520e-01 -2.04536021e-02
-5.84003985e-01 7.50429749e-01 3.98989379e-01 7.25914359e-01
1.55427292e-01 -7.31314778e-01 -6.60012782e-01 -9.35085475e-01
2.25516260e-01 -2.93722332e-01 2.93766022e-01 7.39451721e-02
6.62421435e-02 2.82452226e-01 -1.73069763e+00 2.43354484e-01
4.67148185e-01 1.07469296e+00 2.90481634e-02 2.54714251e-01
7.41528347e-02 7.72656024e-01 -4.07990694e-01 -2.75940567e-01
1.96632177e-01 -5.74067950e-01 -3.13607939e-02 2.44063094e-01
4.00934517e-01 -5.63644588e-01 -1.15476680e+00 1.07564020e+00
5.59170783e-01 3.29304934e-01 1.90592855e-01 -1.35075927e+00
-6.16815507e-01 3.73979472e-02 -6.11022413e-01 7.30357707e-01
7.77709544e-01 6.03537440e-01 3.58052880e-01 -6.96409822e-01
6.13807678e-01 1.28784871e+00 7.40803957e-01 6.03891134e-01
-9.80502248e-01 -8.88130844e-01 3.21390748e-01 7.32837617e-01
-1.45372605e+00 -6.92987680e-01 8.37182283e-01 -3.35782260e-01
6.83744729e-01 1.71243280e-01 1.04590356e+00 1.50457323e+00
6.00039214e-02 8.93256128e-01 8.54963005e-01 2.05749929e-01
5.85979968e-02 -3.83768231e-01 3.59651595e-02 4.45557564e-01
4.35907304e-01 3.19983602e-01 -8.00105214e-01 6.82005510e-02
1.05383086e+00 8.20470691e-01 -5.11395693e-01 1.07908785e-01
-1.51488578e+00 3.17992657e-01 5.12020409e-01 1.94492802e-01
-4.30634439e-01 1.09073520e-01 5.99298894e-01 1.16936736e-01
8.12964961e-02 -2.53479928e-01 1.05145611e-01 -1.83193043e-01
-7.02147007e-01 2.96180129e-01 3.57477725e-01 8.14976454e-01
7.75888562e-01 2.26050653e-02 -4.94740158e-01 4.76151496e-01
1.16470352e-01 7.68599391e-01 6.61521494e-01 -9.12231445e-01
1.98988453e-01 6.25337958e-01 7.78042004e-02 -1.26259005e+00
-2.52759010e-01 -2.34435573e-01 -1.31587517e+00 -2.35214680e-01
4.49282259e-01 3.81799117e-02 -6.34633899e-01 1.85868227e+00
3.31940830e-01 1.14756620e+00 -8.54128599e-02 1.36120009e+00
9.05840576e-01 6.53738797e-01 1.57643542e-01 -4.90651220e-01
1.48775733e+00 -6.07202351e-01 -6.01497293e-01 -1.21780979e-02
-9.24595371e-02 -4.51853752e-01 4.28811371e-01 1.42389432e-01
-8.23697388e-01 -8.39683533e-01 -8.70613098e-01 1.75554097e-01
-1.85601011e-01 -6.95884898e-02 5.40197551e-01 3.84026647e-01
-7.04447746e-01 4.01643068e-01 -8.20590734e-01 -6.44064665e-01
5.65698266e-01 4.41958874e-01 -6.35859072e-01 -3.46653104e-01
-1.17289484e+00 3.60162318e-01 2.88190573e-01 3.13999891e-01
-9.62159872e-01 -4.76771683e-01 -1.03325999e+00 -5.98425977e-02
3.11233718e-02 -9.00552809e-01 7.34890699e-01 -1.04990423e+00
-1.38114822e+00 7.68041193e-01 -6.42099440e-01 -4.97924298e-01
3.46919537e-01 -2.94422661e-03 -6.28580928e-01 3.69462103e-01
2.82077461e-01 5.07317007e-01 8.51866722e-01 -8.33509326e-01
-8.71950209e-01 -5.18774450e-01 -1.58703923e-02 9.70454961e-02
-6.10567331e-01 2.09226713e-01 -6.67635918e-01 -8.09168220e-01
-7.17559829e-02 -7.22627878e-01 -1.48142636e-01 4.51261550e-03
-5.91270439e-02 -1.33974761e-01 8.74281943e-01 -5.50337195e-01
1.08026803e+00 -2.37478805e+00 3.40558775e-02 -1.36271581e-01
4.82258022e-01 2.19616592e-01 -6.39540926e-02 -4.75110626e-03
5.77477217e-02 -4.29480046e-01 -1.05686106e-01 -3.22480530e-01
-3.53565603e-01 1.12092592e-01 -6.77756071e-02 7.37126231e-01
3.75608742e-01 9.32985783e-01 -1.16988552e+00 -5.74294686e-01
5.88448822e-01 6.90254033e-01 -2.20938578e-01 2.09882244e-01
2.75688082e-01 8.81725848e-01 -5.29129326e-01 6.74460113e-01
7.40220845e-01 -3.42202514e-01 -3.58495027e-01 -4.75848913e-01
-2.83236951e-01 -2.06094921e-01 -1.14019418e+00 1.85422981e+00
-8.22350234e-02 6.41952813e-01 -1.04950368e-01 -1.13579774e+00
8.83839548e-01 3.58688414e-01 7.29462743e-01 -8.60134542e-01
1.92517444e-01 5.07920943e-02 -2.45293796e-01 -4.70490754e-01
1.57208532e-01 1.20836988e-01 9.91789158e-03 1.26124427e-01
1.95163667e-01 7.90480673e-01 2.51932353e-01 1.61239989e-02
1.25033319e+00 4.54147719e-02 1.66347235e-01 1.52485088e-01
9.36312318e-01 -2.92414308e-01 1.12054801e+00 4.60976303e-01
-6.66383684e-01 7.41478443e-01 -1.10580958e-01 -9.37447369e-01
-1.00536299e+00 -1.20407104e+00 -7.18942517e-03 7.63509870e-01
7.30887413e-01 -2.23558739e-01 -5.09468496e-01 -1.99844152e-01
-5.48839495e-02 -1.27903193e-01 -6.03731751e-01 -3.10053766e-01
-8.57433856e-01 -9.49531496e-01 4.22092557e-01 5.87712884e-01
1.01237607e+00 -9.53839540e-01 -4.21086103e-01 4.61155474e-01
-4.14232999e-01 -1.40312994e+00 -8.02380919e-01 -4.19977635e-01
-4.80210334e-01 -1.06363630e+00 -9.96469676e-01 -9.39276636e-01
5.34385026e-01 8.00600529e-01 6.32880926e-01 -1.35891512e-02
-5.06172955e-01 5.01010716e-01 -2.94815689e-01 5.27703129e-02
3.43935013e-01 -4.04612303e-01 4.12551522e-01 7.06562281e-01
8.10727239e-01 -8.47059011e-01 -1.06636238e+00 3.62329155e-01
-6.63056016e-01 5.71054742e-02 4.17391926e-01 8.17841172e-01
5.85861981e-01 -1.12019800e-01 5.99147260e-01 -9.17641968e-02
1.47071213e-01 -5.18141031e-01 -3.74197453e-01 1.12217672e-01
1.97771266e-01 -4.20600206e-01 8.12237978e-01 -6.33335650e-01
-1.14013386e+00 3.38577896e-01 2.16135815e-01 -7.10399687e-01
-3.85290682e-01 -2.48119347e-02 -4.38635141e-01 -1.62297219e-01
2.42065847e-01 7.60398090e-01 4.31745313e-03 -1.85855776e-01
-6.25465810e-02 6.16454363e-01 1.13597476e+00 -4.74676937e-01
7.18852341e-01 9.54671800e-01 -1.16274439e-01 -6.95705533e-01
-5.54647446e-01 -7.36822009e-01 -3.72559160e-01 -4.82701600e-01
1.08114696e+00 -1.43736959e+00 -1.11313140e+00 1.05318224e+00
-1.41828489e+00 2.73955584e-01 -1.49148330e-01 6.82042420e-01
-3.76336426e-01 5.09289086e-01 -8.84203196e-01 -6.27210915e-01
-1.61086202e-01 -9.62359309e-01 1.09219813e+00 1.00304425e+00
1.27002046e-01 -5.71875989e-01 -4.27041575e-02 1.53667852e-01
2.93534607e-01 5.86359084e-01 1.40941422e-03 -3.40373665e-01
-7.64537215e-01 -2.44132146e-01 -5.50196230e-01 -5.61794974e-02
4.20925409e-01 -2.75230885e-01 -9.80994284e-01 -3.26018721e-01
1.35372713e-01 4.55245189e-02 1.01322544e+00 4.56281304e-01
1.02146852e+00 -1.06857747e-01 -5.54592848e-01 9.34119999e-01
1.16489732e+00 2.05750391e-01 6.31399870e-01 4.68983084e-01
9.08737779e-01 4.11236018e-01 3.26740742e-01 8.95521700e-01
5.53351581e-01 6.63533390e-01 1.67320013e-01 1.02111429e-01
-1.91680580e-01 -2.50399977e-01 5.09224415e-01 6.56564236e-01
-4.31825936e-01 -3.82026881e-02 -3.89003843e-01 5.10655761e-01
-2.14463758e+00 -1.54882729e+00 -2.03925986e-02 2.28392458e+00
5.57652712e-01 -1.08967200e-01 3.69064033e-01 1.80299729e-02
1.35551381e+00 2.39450380e-01 -7.86621094e-01 5.77379525e-01
-5.93138397e-01 -1.05338454e-01 3.25042516e-01 -8.06414559e-02
-1.31903052e+00 6.48969352e-01 5.03540945e+00 6.74774468e-01
-1.08727777e+00 1.12226903e-01 5.71629643e-01 -3.07575554e-01
1.83205493e-02 -2.51827389e-01 -9.02798116e-01 1.11155677e+00
6.74731433e-01 -4.29854572e-01 3.49595010e-01 4.25795734e-01
5.38159609e-01 1.59424186e-01 -1.10419405e+00 1.70139837e+00
2.31213123e-01 -1.45066392e+00 1.29690453e-01 -1.54733971e-01
6.14783108e-01 -2.96195239e-01 -1.00065857e-01 -2.28102542e-02
9.64267701e-02 -9.10106778e-01 5.61168492e-01 8.86647165e-01
6.55160427e-01 -7.57825255e-01 7.98949659e-01 3.72339524e-02
-1.86824560e+00 -1.71939090e-01 -5.19196272e-01 -3.17776531e-01
4.15945321e-01 5.28824031e-01 3.35394666e-02 5.78255475e-01
1.02023816e+00 1.64438140e+00 -4.57020998e-01 1.32334256e+00
2.11750627e-01 3.19271594e-01 -3.44788790e-01 1.97099641e-01
-6.61315545e-02 -1.98099405e-01 5.55531621e-01 1.29828417e+00
4.05813068e-01 5.02139032e-01 3.94778699e-01 7.05890417e-01
8.68840963e-02 -3.14060241e-01 -5.66039324e-01 2.58771509e-01
5.80363512e-01 1.22144675e+00 -5.27397156e-01 -3.69983435e-01
-7.73835659e-01 1.49163985e+00 -3.35975830e-03 4.68708426e-01
-9.77090418e-01 -3.24848145e-01 1.06227612e+00 -1.31671622e-01
3.37498903e-01 -2.63257325e-01 1.94680810e-01 -1.57496440e+00
2.45921060e-01 -6.98131800e-01 3.85786086e-01 -4.46902215e-01
-1.69941902e+00 3.53643328e-01 -3.77642572e-01 -1.65315592e+00
3.69112603e-02 -4.14666921e-01 -8.74588251e-01 6.51743114e-01
-1.44353509e+00 -1.27513826e+00 -1.01547110e+00 1.18040812e+00
5.69677234e-01 -3.24806213e-01 4.05492395e-01 7.10119724e-01
-1.01406586e+00 6.12609446e-01 1.48702204e-01 6.00021720e-01
8.69898081e-01 -6.40507460e-01 3.43704373e-01 1.04657185e+00
-5.12407273e-02 6.05666399e-01 2.72298902e-01 -6.16355121e-01
-1.63899207e+00 -1.34395921e+00 4.78034496e-01 -1.12529315e-01
4.42519456e-01 -2.25548908e-01 -8.11838627e-01 5.04379809e-01
-1.60093486e-01 5.77625036e-01 6.34785950e-01 -5.37631333e-01
-3.39169979e-01 -4.75211978e-01 -8.69811833e-01 3.68915737e-01
1.48402858e+00 -6.95084274e-01 -5.15990674e-01 1.82017386e-01
6.01986945e-01 -2.21809000e-01 -7.42792487e-01 2.31596172e-01
6.58619106e-01 -1.02270043e+00 1.28090882e+00 -3.95806432e-01
1.92811117e-01 -7.95871258e-01 -2.02377699e-02 -8.41819167e-01
-6.56868398e-01 -6.25636220e-01 -1.79532185e-01 1.50908458e+00
-5.27226865e-01 -6.88689291e-01 7.73863792e-01 4.31109875e-01
1.91248432e-01 -1.62412852e-01 -1.14197564e+00 -8.14742684e-01
-5.68764150e-01 -9.35440809e-02 6.57742798e-01 8.68287385e-01
-2.69373298e-01 9.73489042e-03 -6.17110014e-01 2.55215615e-01
9.13037956e-01 6.13780506e-02 6.45221412e-01 -1.28447676e+00
-7.46873543e-02 -2.64166772e-01 -1.12552917e+00 -1.19669950e+00
2.37307638e-01 -5.25342941e-01 5.86313312e-04 -1.10941303e+00
5.55240929e-01 -8.17524362e-03 -6.44228995e-01 4.14753035e-02
-4.49385822e-01 4.77084339e-01 2.71571428e-01 4.59374368e-01
-1.06036806e+00 7.00456858e-01 1.01701152e+00 -2.99002320e-01
-1.00138895e-02 -2.92120457e-01 -3.78338933e-01 7.82797098e-01
3.75946850e-01 -1.67555436e-01 -9.96503700e-03 -4.85885471e-01
-2.94838935e-01 5.61537482e-02 9.78147864e-01 -1.58048260e+00
7.60596454e-01 -1.37797549e-01 9.89313245e-01 -3.16441625e-01
4.30266619e-01 -6.48040175e-01 3.36127698e-01 3.98821235e-01
-1.55825438e-02 5.45305461e-02 1.05802352e-02 1.00458336e+00
-4.88962114e-01 2.11210176e-01 5.62597990e-01 -2.29940608e-01
-1.13848627e+00 8.44173312e-01 -3.32387298e-01 -5.52587993e-02
1.15161550e+00 -4.68707383e-01 -3.49575818e-01 -2.39678204e-01
-4.69235450e-01 2.84941643e-01 6.57229125e-01 4.79315042e-01
7.93712139e-01 -1.73683751e+00 -8.26620042e-01 2.96648532e-01
1.75101936e-01 -3.99840698e-02 6.88066781e-01 8.64302933e-01
-1.36934325e-01 1.35978028e-01 -6.74320519e-01 -8.14825773e-01
-1.09643090e+00 5.69608212e-01 1.40833795e-01 1.65754080e-01
-7.56013811e-01 7.12962687e-01 4.84353185e-01 3.27961504e-01
4.38276306e-02 6.01376928e-02 -2.95331955e-01 -1.77578013e-02
1.15435088e+00 4.52378213e-01 -3.82966697e-01 -9.68301415e-01
-5.77306926e-01 7.55426943e-01 -4.34562601e-02 6.24494664e-02
1.23139095e+00 -4.37638462e-01 -6.26075640e-02 2.54470110e-01
1.18124998e+00 -3.66933733e-01 -1.69493508e+00 -4.53411102e-01
-2.58800894e-01 -7.32310891e-01 -2.72111595e-01 -3.59477326e-02
-1.15016901e+00 8.82448137e-01 7.23609149e-01 -2.43421003e-01
1.36192226e+00 -2.96844572e-01 1.34564567e+00 2.60046959e-01
4.97711182e-01 -7.77749956e-01 1.94907084e-01 2.70588696e-01
3.59286875e-01 -1.41527057e+00 -2.84853667e-01 -2.74582058e-01
-3.39754373e-01 1.17695189e+00 7.37960219e-01 -3.21422011e-01
3.98755550e-01 1.26536503e-01 -3.08036268e-01 2.00999990e-01
-3.62276018e-01 -3.94654602e-01 -1.39477104e-01 1.00363338e+00
1.22029193e-01 -1.51453257e-01 -8.83384515e-03 9.37156320e-01
2.33940169e-01 3.38866442e-01 1.84072614e-01 6.19159341e-01
-2.69808650e-01 -7.96054959e-01 -5.16438901e-01 3.69140148e-01
-1.34544387e-01 6.79766089e-02 -1.17160892e-02 3.83727670e-01
4.97618645e-01 8.84793222e-01 4.23531353e-01 -6.17731869e-01
2.16861323e-01 -3.55462849e-01 4.21636045e-01 -1.84016913e-01
-5.25842130e-01 -1.12290129e-01 -3.23903531e-01 -7.88320720e-01
-8.18780541e-01 -9.87135649e-01 -1.16023254e+00 -5.37383914e-01
1.03079304e-01 -1.84690133e-01 -3.78316604e-02 9.63766515e-01
5.10236084e-01 3.05185050e-01 7.04223633e-01 -1.09449005e+00
4.68227528e-02 -5.90312600e-01 -4.59093779e-01 8.81187081e-01
4.41471517e-01 -7.66112685e-01 -3.44016790e-01 5.10988295e-01] | [14.625662803649902, 0.9359079003334045] |
5e7941c7-f41a-440e-a998-764357c4be68 | knowledge-of-knowledge-exploring-known | 2305.13712 | null | https://arxiv.org/abs/2305.13712v1 | https://arxiv.org/pdf/2305.13712v1.pdf | Knowledge of Knowledge: Exploring Known-Unknowns Uncertainty with Large Language Models | This paper investigates the capabilities of Large Language Models (LLMs) in the context of understanding their own knowledge and measuring their uncertainty. We argue this is an important feature for mitigating hallucinations. Specifically, we focus on addressing \textit{known-unknown} questions, characterized by high uncertainty due to the absence of definitive answers. To facilitate our study, we collect a dataset with new Known-Unknown Questions (KUQ) and propose a novel categorization scheme to elucidate the sources of uncertainty. Subsequently, we assess the LLMs' ability to differentiate between known and unknown questions and classify them accordingly. Moreover, we evaluate the quality of their answers in an Open-Ended QA setting. To quantify the uncertainty expressed in the answers, we create a semantic evaluation method that measures the model's accuracy in expressing uncertainty between known vs unknown questions. | ['William Wang', 'Wenhu Chen', 'Liangming Pan', 'Alfonso Amayuelas'] | 2023-05-23 | null | null | null | null | ['known-unknowns'] | ['miscellaneous'] | [-3.31363887e-01 5.70005596e-01 4.10985053e-02 -4.05570090e-01
-1.23110592e+00 -8.32027912e-01 5.65901279e-01 5.19752860e-01
-2.70462036e-01 9.02246833e-01 8.42307210e-01 -2.70606041e-01
-7.27748394e-01 -6.52306914e-01 -3.37235540e-01 4.58939234e-03
5.52218020e-01 5.79959810e-01 -1.96897835e-02 -2.18001515e-01
6.15543485e-01 2.61667371e-01 -1.14722884e+00 5.71431756e-01
1.48016572e+00 9.59731340e-01 -4.90504093e-02 1.61611795e-01
-5.49527586e-01 1.31497300e+00 -9.95977044e-01 -8.47390056e-01
-3.22541714e-01 -3.73346835e-01 -1.29867804e+00 -1.42839298e-01
3.42350364e-01 -1.61073565e-01 1.93615463e-02 1.05233097e+00
3.89657408e-01 1.04212053e-01 7.78060734e-01 -1.17462397e+00
-8.55855227e-01 7.01435745e-01 4.21488404e-01 3.63030314e-01
8.58144224e-01 -1.00444779e-01 8.84140134e-01 -6.45356357e-01
4.80138659e-01 1.25740254e+00 5.45671046e-01 5.17421484e-01
-1.03594756e+00 -3.61375242e-01 -1.14954580e-02 6.44462109e-01
-1.33033264e+00 -5.64109087e-01 5.54887772e-01 -6.60240054e-01
6.12564445e-01 3.30174536e-01 -9.45624709e-02 1.14506280e+00
2.65162289e-01 4.38278973e-01 1.52868116e+00 -3.68773341e-01
7.31208801e-01 7.21961617e-01 5.44281483e-01 2.67705798e-01
2.58880784e-03 -8.64074975e-02 -8.37147236e-01 -3.06375176e-01
4.91816737e-02 -5.99889755e-01 -5.89346826e-01 4.97646770e-03
-7.86961794e-01 8.85869920e-01 3.29078764e-01 2.28752404e-01
-3.61139715e-01 -4.54720259e-01 -1.30390793e-01 3.57377797e-01
2.25028008e-01 9.37715411e-01 -3.56236041e-01 -3.48370761e-01
-5.23173690e-01 5.94570711e-02 9.75239277e-01 4.51179862e-01
5.69515526e-01 -3.55698079e-01 -5.91098368e-01 8.13726187e-01
2.19663918e-01 2.27661222e-01 6.62692904e-01 -1.41330791e+00
4.55513597e-01 5.88972926e-01 5.45034409e-01 -9.87510204e-01
-3.67138594e-01 -6.15835726e-01 -3.25230569e-01 -2.56159693e-01
2.08579719e-01 -1.15067020e-01 -1.84295863e-01 1.85764229e+00
-1.31642357e-01 -2.60993212e-01 4.09870476e-01 5.32759905e-01
8.97666276e-01 2.54658163e-01 1.92285657e-01 -2.11924747e-01
1.27581143e+00 -7.51446843e-01 -1.07081282e+00 -3.41024727e-01
6.83511496e-01 -4.22988474e-01 1.31572306e+00 2.37561882e-01
-8.40027690e-01 -3.24622601e-01 -6.91124260e-01 -3.99576984e-02
-3.91385972e-01 3.46938707e-02 2.73890585e-01 7.91146338e-01
-8.64044070e-01 2.87828535e-01 -3.99883613e-02 -6.81841075e-02
1.94097400e-01 -2.18320221e-01 -1.01215854e-01 -3.92724089e-02
-1.63274968e+00 1.43711376e+00 3.20495337e-01 -2.52378315e-01
-7.05738664e-01 -5.17595947e-01 -7.99721837e-01 3.81374776e-01
5.03548205e-01 -8.62851679e-01 1.18215847e+00 -3.51577312e-01
-1.20668721e+00 5.76894403e-01 -3.40056419e-01 -5.02423525e-01
4.06940192e-01 -7.66709819e-02 -6.45983100e-01 2.38876522e-01
4.25061524e-01 5.02986610e-01 6.82595611e-01 -1.35607135e+00
-2.16993541e-01 -5.71086049e-01 5.78181744e-01 1.07598722e-01
-2.78624475e-01 -6.79442484e-04 2.47020409e-01 -3.30339253e-01
2.05390573e-01 -4.06385183e-01 1.54426277e-01 -4.27765220e-01
-3.22310209e-01 -3.90115261e-01 1.90430611e-01 -7.38076627e-01
1.56525755e+00 -1.88353634e+00 -1.98025107e-01 5.39459996e-02
3.42206597e-01 1.51949286e-01 -8.20225254e-02 4.15903777e-01
2.19883263e-01 3.22109491e-01 -3.32332551e-01 -1.58629149e-01
2.92033017e-01 2.82947093e-01 -7.06915081e-01 -3.76482040e-01
3.46931398e-01 1.03035712e+00 -7.17976630e-01 -4.63111460e-01
-2.72181630e-02 -3.13967764e-02 -3.75450909e-01 3.96971494e-01
-4.32381660e-01 3.66365492e-01 -2.80082941e-01 5.11938095e-01
5.48584461e-01 -2.29490668e-01 -1.41685292e-01 -1.01844721e-01
1.38493270e-01 6.51575983e-01 -1.00572526e+00 1.27476537e+00
-5.74200511e-01 2.79806703e-01 -1.75164789e-01 -5.07909775e-01
9.32451606e-01 2.96041965e-01 -2.86713392e-01 -8.19502056e-01
1.25821218e-01 1.61449045e-01 -1.93843365e-01 -8.36803854e-01
4.28146005e-01 -3.69079679e-01 -1.25965327e-01 4.03157145e-01
1.97433196e-02 -4.44219649e-01 6.87486231e-02 4.64734077e-01
9.13908362e-01 -4.63800818e-01 2.32932746e-01 -2.44770989e-01
7.08459139e-01 -1.09963484e-01 2.62242973e-01 9.22828197e-01
-4.60408568e-01 4.47548002e-01 6.97770417e-01 1.89756691e-01
-2.32417256e-01 -1.39904344e+00 -2.26939350e-01 6.92606151e-01
5.88049479e-02 -4.66821522e-01 -7.40160227e-01 -6.89176083e-01
-1.92798451e-01 1.65581858e+00 -5.70590377e-01 -3.79898608e-01
3.86458099e-01 -3.55969787e-01 7.02489138e-01 4.32727754e-01
5.98911941e-01 -7.62565911e-01 -5.32855451e-01 2.25531124e-02
-9.23686743e-01 -1.01381958e+00 -1.70516834e-01 6.51641190e-02
-6.14776909e-01 -1.09965467e+00 -1.52662680e-01 -2.36000463e-01
2.44523838e-01 -1.82265759e-01 1.29952407e+00 -4.03859496e-01
3.88063669e-01 8.92755210e-01 -3.26166213e-01 -4.13783342e-01
-5.39218485e-01 -2.96138138e-01 -1.39809579e-01 -1.79213569e-01
6.72762752e-01 -3.63726199e-01 -2.16108903e-01 3.60626668e-01
-1.23365784e+00 -4.75883037e-01 2.85124063e-01 4.03390825e-01
1.39256507e-01 1.78481311e-01 1.21101582e+00 -5.74088216e-01
1.67389953e+00 -8.86560798e-01 -1.60914138e-01 6.53113782e-01
-6.69612646e-01 4.80065435e-01 3.20231259e-01 -2.00160667e-01
-1.41414499e+00 -7.14926124e-01 -3.22651774e-01 1.59398988e-01
-2.85223246e-01 7.74272799e-01 -2.93273151e-01 -7.72796273e-02
9.25935030e-01 -1.84987541e-02 -2.09686667e-01 -4.02095258e-01
6.03970885e-01 1.02565479e+00 6.16434336e-01 -7.59080231e-01
2.93834776e-01 3.20706367e-01 -5.69597721e-01 -4.90311086e-01
-1.40561616e+00 -4.07290280e-01 -2.23397508e-01 -2.08147138e-01
6.76823735e-01 -6.50606275e-01 -6.32233202e-01 1.04738781e-02
-1.19666231e+00 1.75553396e-01 -3.37210566e-01 4.74257618e-01
-6.27062917e-01 6.08129501e-01 -4.08691227e-01 -9.29175317e-01
-2.41740823e-01 -9.25458372e-01 6.71490252e-01 2.45240644e-01
-7.98960984e-01 -1.07139719e+00 7.13214055e-02 1.03316200e+00
6.45931304e-01 -3.22263211e-01 1.48683679e+00 -9.85496402e-01
-1.78446606e-01 2.09274605e-01 -1.77004978e-01 5.39900303e-01
-6.39266148e-02 -7.11010814e-01 -1.01368105e+00 3.16144884e-01
8.05449069e-01 -8.15271378e-01 6.89096332e-01 -1.65590588e-02
1.00322485e+00 -3.79648447e-01 2.89903164e-01 -1.59890056e-01
1.12626231e+00 -4.26801331e-02 7.47708976e-01 3.05291176e-01
-8.92168134e-02 9.98454809e-01 5.82238436e-01 4.21698838e-01
7.84095287e-01 2.81882465e-01 2.92553544e-01 8.12634885e-01
-3.61702554e-02 -4.71087366e-01 3.45966071e-01 7.40048766e-01
7.01569080e-01 -3.85603607e-01 -8.63504827e-01 6.42801821e-01
-1.38365710e+00 -6.51917040e-01 8.29252079e-02 2.01916957e+00
9.37044919e-01 -2.73928922e-02 -6.25407755e-01 1.11777551e-01
4.96644765e-01 8.92920047e-02 -3.04461747e-01 -5.71745396e-01
-4.00616944e-01 2.28567287e-01 -2.93817967e-01 7.58809149e-01
-4.30885226e-01 6.82834864e-01 6.50614166e+00 9.94888484e-01
-5.22171736e-01 -1.35273980e-02 5.36178470e-01 2.71546304e-01
-1.03230083e+00 9.04816166e-02 -4.27959025e-01 4.91506457e-01
1.10669875e+00 -2.36023933e-01 2.54658520e-01 3.84372205e-01
6.21195026e-02 -4.98794556e-01 -1.11575198e+00 6.93813741e-01
3.46710145e-01 -1.12778711e+00 3.92290801e-01 -4.75364953e-01
5.00787973e-01 -4.48043108e-01 4.76826727e-02 5.60832858e-01
3.66595201e-02 -1.11976051e+00 5.99949539e-01 9.42747772e-01
3.02852720e-01 -5.29408693e-01 8.92589509e-01 8.18527222e-01
-2.51673877e-01 -2.60253787e-01 -2.90494621e-01 -3.27927262e-01
1.22511819e-01 4.88335729e-01 -7.23126948e-01 4.56911504e-01
4.45144802e-01 -2.06455261e-01 -8.98046970e-01 7.50002801e-01
-6.31591082e-01 4.98985559e-01 -3.96060664e-03 5.01422100e-02
1.16724052e-01 -1.17650874e-01 2.82721967e-01 6.70696735e-01
3.36144447e-01 4.10770833e-01 -2.17819929e-01 1.45758986e+00
-1.54551178e-01 3.94272059e-02 -2.85901248e-01 -2.80358136e-01
7.93087482e-01 6.78240895e-01 -1.45497426e-01 -2.35985741e-01
-8.67869332e-02 7.67533422e-01 5.00614762e-01 2.46229425e-01
-4.96179968e-01 -2.71470845e-01 2.34055594e-01 -1.85308367e-01
-4.82644379e-01 5.60682639e-02 -3.75483721e-01 -1.34429955e+00
1.81079686e-01 -9.49404716e-01 5.06119192e-01 -1.40497601e+00
-1.45021594e+00 6.40234470e-01 1.59438662e-02 -5.65920055e-01
-4.47425038e-01 -3.49031121e-01 -2.64751017e-01 8.34964335e-01
-1.62356281e+00 -6.10900044e-01 -2.58854628e-01 4.74153221e-01
2.95880586e-01 1.37721777e-01 1.03438938e+00 2.43682717e-03
-1.19730420e-01 3.38057429e-01 -7.34048383e-03 -1.67188779e-01
8.65687907e-01 -1.00848150e+00 -2.02953339e-01 5.25378525e-01
1.31163076e-02 7.69096375e-01 7.13765383e-01 -6.89480126e-01
-8.04253340e-01 -6.39694512e-01 1.46169126e+00 -9.86556113e-01
7.80992746e-01 1.65190876e-01 -1.22822559e+00 4.68435854e-01
1.86825514e-01 -6.25427604e-01 1.04818094e+00 5.29827997e-02
-5.60070634e-01 4.37658161e-01 -1.63223398e+00 4.21360821e-01
6.38331294e-01 -1.13917363e+00 -1.47202528e+00 2.26059884e-01
9.40347314e-01 2.63132956e-02 -8.88579428e-01 5.26683331e-01
-1.19412746e-02 -1.20480144e+00 6.99184299e-01 -7.76115894e-01
3.98553699e-01 -2.02141404e-02 -4.92444754e-01 -1.47332692e+00
-8.12333673e-02 -2.31685266e-02 -1.42523885e-01 1.20061147e+00
4.51227933e-01 -7.08599448e-01 4.77658570e-01 9.94877160e-01
2.96341199e-02 -5.23710370e-01 -9.64642048e-01 -5.93580544e-01
1.98107913e-01 -7.27939427e-01 5.69803596e-01 9.31319892e-01
1.73037603e-01 4.14295107e-01 -1.15713514e-02 3.87909651e-01
3.24077934e-01 9.78162959e-02 -2.06559841e-02 -1.17621994e+00
-1.78050268e-02 -1.48491964e-01 -1.33866563e-01 -7.66113281e-01
5.02218783e-01 -6.20015085e-01 -3.12733203e-01 -1.59715736e+00
1.75302073e-01 -1.97336689e-01 -9.11759883e-02 4.58832420e-02
-2.24103361e-01 -2.05315068e-01 2.00907946e-01 3.92492255e-03
-7.36314595e-01 8.01826358e-01 7.57744193e-01 -8.98276195e-02
1.52682453e-01 1.18681164e-02 -1.14264452e+00 7.77194023e-01
8.63376558e-01 -3.28569114e-01 -6.35550201e-01 -5.24282932e-01
4.57984388e-01 3.69642705e-01 5.11259079e-01 -1.03120732e+00
3.09368670e-01 -1.12276793e-01 -6.12132140e-02 -3.22779447e-01
4.69024837e-01 -6.37108326e-01 -1.74687639e-01 2.42439672e-01
-8.95396411e-01 -1.88614830e-01 2.07427755e-01 3.52320731e-01
-5.37211955e-01 -8.62150133e-01 3.59405041e-01 -2.12013423e-01
-6.13899231e-01 -3.40062678e-01 -4.54051167e-01 5.93393028e-01
5.98286748e-01 3.08787972e-01 -6.49536431e-01 -9.33002591e-01
-1.01712716e+00 3.85856628e-01 1.91670313e-01 5.28483212e-01
8.62602890e-01 -1.11870968e+00 -6.10012710e-01 -1.97556347e-01
5.09985983e-01 -5.83063364e-01 4.52957213e-01 6.26516700e-01
-5.49427047e-02 9.20151055e-01 -4.86069918e-02 -8.90548676e-02
-8.85978222e-01 4.78882790e-01 4.78683144e-01 -1.44882962e-01
3.81787747e-01 7.78871953e-01 -1.00564174e-01 -7.47251749e-01
3.38726252e-01 -2.31959581e-01 -5.82912743e-01 2.43194699e-01
4.77839470e-01 6.65611207e-01 1.98255196e-01 -3.35343033e-01
-1.99051082e-01 7.62634724e-02 2.90478617e-01 -3.26985240e-01
6.14106834e-01 -5.46890199e-01 -4.09851789e-01 4.97441143e-01
8.41612935e-01 3.95740271e-01 -5.74883878e-01 -4.55289960e-01
5.02108097e-01 -4.74390805e-01 -1.05422169e-01 -1.41292882e+00
-1.58421651e-01 9.56480742e-01 5.50755620e-01 1.60248011e-01
8.74163628e-01 3.12987179e-01 4.88517165e-01 7.45997548e-01
4.61685896e-01 -1.08026505e+00 2.14242682e-01 7.19654143e-01
1.05354738e+00 -1.30588162e+00 -5.17649055e-01 -2.80546188e-01
-8.07062924e-01 8.59503090e-01 5.88792801e-01 7.35308826e-01
2.13723421e-01 -1.64537787e-01 4.71166104e-01 -2.55688846e-01
-7.74613976e-01 -1.81185931e-01 3.78700316e-01 5.24265409e-01
3.50532711e-01 8.94394591e-02 -5.78911304e-01 1.26183736e+00
-4.92660165e-01 1.17635755e-02 7.99133182e-01 6.47738218e-01
-7.37347662e-01 -8.62418652e-01 -4.89276946e-01 4.24748898e-01
-2.37991959e-01 -1.93133235e-01 -8.20646942e-01 5.20554036e-02
1.02798201e-01 1.74319839e+00 -4.25648332e-01 -3.90040666e-01
9.38136131e-02 5.31109989e-01 3.06955576e-01 -5.68140566e-01
-3.98765475e-01 -8.20499003e-01 3.28112364e-01 -3.97061080e-01
-2.40081310e-01 -4.25756685e-02 -1.00317478e+00 8.79553929e-02
-2.97102809e-01 8.07347834e-01 6.74260974e-01 1.26009738e+00
5.52567661e-01 3.99118036e-01 3.58776182e-01 4.20284033e-01
-1.32226694e+00 -9.17038143e-01 -2.67017096e-01 4.51653779e-01
5.68647720e-02 -5.73081493e-01 -7.28157282e-01 -5.12837827e-01] | [11.163863182067871, 8.013045310974121] |
1eca1bb0-3490-4b85-9cbe-a4bb5004d2c6 | sequential-best-arm-identification-with | 2305.11908 | null | https://arxiv.org/abs/2305.11908v1 | https://arxiv.org/pdf/2305.11908v1.pdf | Sequential Best-Arm Identification with Application to Brain-Computer Interface | A brain-computer interface (BCI) is a technology that enables direct communication between the brain and an external device or computer system. It allows individuals to interact with the device using only their thoughts, and holds immense potential for a wide range of applications in medicine, rehabilitation, and human augmentation. An electroencephalogram (EEG) and event-related potential (ERP)-based speller system is a type of BCI that allows users to spell words without using a physical keyboard, but instead by recording and interpreting brain signals under different stimulus presentation paradigms. Conventional non-adaptive paradigms treat each word selection independently, leading to a lengthy learning process. To improve the sampling efficiency, we cast the problem as a sequence of best-arm identification tasks in multi-armed bandits. Leveraging pre-trained large language models (LLMs), we utilize the prior knowledge learned from previous tasks to inform and facilitate subsequent tasks. To do so in a coherent way, we propose a sequential top-two Thompson sampling (STTS) algorithm under the fixed-confidence setting and the fixed-budget setting. We study the theoretical property of the proposed algorithm, and demonstrate its substantial empirical improvement through both synthetic data analysis as well as a P300 BCI speller simulator example. | ['Lexin Li', 'Tor Lattimore', 'Jian Kang', 'Botao Hao', 'Xin Zhou'] | 2023-05-17 | null | null | null | null | ['thompson-sampling', 'eeg', 'multi-armed-bandits', 'eeg'] | ['methodology', 'methodology', 'miscellaneous', 'time-series'] | [ 6.42341375e-01 -4.25854623e-01 -3.50446343e-01 4.15776111e-02
-9.12913620e-01 -5.34930468e-01 4.27697927e-01 -4.96722341e-01
-5.98464251e-01 1.16682434e+00 1.35293916e-01 -6.23919070e-01
-4.38009173e-01 -3.09536576e-01 -5.96032858e-01 -8.93074691e-01
8.67131278e-02 2.81664252e-01 -3.81781608e-01 1.62003636e-01
2.74478227e-01 2.34949797e-01 -1.05680275e+00 6.59703389e-02
1.12466180e+00 1.10976040e+00 6.20487630e-01 5.01724482e-01
2.46985286e-01 2.06989914e-01 -7.97915936e-01 -1.02582768e-01
-1.95759032e-02 -4.74067509e-01 -3.94066334e-01 -3.63821611e-02
-3.68741870e-01 -7.99107924e-02 -2.99255639e-01 1.26069093e+00
7.38010764e-01 1.72848925e-01 6.60368085e-01 -1.04746354e+00
-5.95780075e-01 4.40015197e-01 -6.91116869e-01 3.76642525e-01
3.64457279e-01 1.44498035e-01 5.74888587e-01 -5.68232656e-01
-2.97549963e-01 1.02282166e+00 2.16745779e-01 6.89484358e-01
-1.41566014e+00 -1.17006457e+00 2.32726961e-01 4.65927064e-01
-1.43438387e+00 -5.42729437e-01 7.50122070e-01 -3.69011641e-01
1.00829959e+00 3.39870304e-01 1.08538318e+00 1.64191103e+00
4.68023151e-01 1.04824555e+00 1.31864142e+00 -4.57748890e-01
5.29227436e-01 4.30903211e-02 4.76950347e-01 -2.02845968e-02
4.00732905e-01 1.55583113e-01 -9.38937962e-01 -4.61800307e-01
8.22021067e-01 1.41043872e-01 -7.30704486e-01 -1.35921225e-01
-1.26570404e+00 5.10769963e-01 -1.59286812e-01 6.64917156e-02
-9.11179662e-01 1.75927043e-01 1.99164808e-01 2.88156062e-01
3.35124582e-01 3.40900064e-01 -3.97741884e-01 -4.18677181e-01
-9.25078630e-01 -1.44793503e-02 6.50060833e-01 8.80534053e-01
-2.04029400e-02 9.48911309e-02 -6.23159707e-01 9.30214345e-01
7.30114505e-02 9.31888640e-01 1.01503801e+00 -2.82191187e-01
5.10274112e-01 -5.44613935e-02 4.87896502e-01 -4.31215584e-01
-2.49164239e-01 -4.84312177e-01 -1.00899088e+00 -2.03458712e-01
1.47137716e-01 -4.14477527e-01 -5.81833541e-01 1.77217329e+00
-1.35815129e-01 6.05131447e-01 -1.66563794e-01 7.91346967e-01
-7.63904899e-02 4.84711021e-01 -2.63296142e-02 -8.99927616e-01
1.42040765e+00 -5.00683427e-01 -1.05660009e+00 -5.48864603e-01
6.02756329e-02 -1.17013402e-01 1.22091925e+00 1.15132272e+00
-1.17655551e+00 -5.42416200e-02 -1.08683300e+00 5.39239824e-01
8.55224133e-02 1.41155869e-01 7.19502628e-01 9.70211864e-01
-8.65080714e-01 2.29292408e-01 -8.98270130e-01 -3.70025635e-03
5.47319710e-01 4.87728149e-01 3.80837433e-02 1.33051887e-01
-1.28224909e+00 8.25201809e-01 1.69494435e-01 6.51481152e-02
-9.93795753e-01 -4.90142226e-01 -3.21514666e-01 3.15560669e-01
5.34126043e-01 -6.18691504e-01 1.10297155e+00 -8.47937822e-01
-2.03726101e+00 3.94431412e-01 -4.69170392e-01 -5.64955771e-01
1.85406759e-01 -4.44435924e-01 -3.21880430e-01 -2.24024430e-01
-2.16244519e-01 2.77143326e-02 1.29631495e+00 -5.81358314e-01
-4.51078415e-01 -6.53633177e-01 -3.98547411e-01 2.07253829e-01
-6.91407204e-01 1.76380143e-01 -3.50533605e-01 -8.33936274e-01
-4.18618202e-01 -9.83024716e-01 1.97933152e-01 -7.02131450e-01
-5.59459805e-01 -3.58624905e-01 9.46876872e-03 -7.05546200e-01
1.49042201e+00 -2.10547066e+00 3.14033806e-01 3.14531952e-01
-1.28761053e-01 2.75847197e-01 -1.12504221e-01 -2.39336584e-02
-1.67473823e-01 -1.42090589e-01 -1.55130893e-01 -1.06552385e-01
-2.19544560e-01 -1.94089636e-01 -4.81470615e-01 4.83620137e-01
-2.37572283e-01 9.21516597e-01 -8.17677200e-01 -1.56931654e-02
-3.38538848e-02 2.81132013e-01 -2.71456063e-01 2.09390834e-01
1.89669669e-01 4.14374679e-01 -6.15031183e-01 3.31207514e-01
2.26901218e-01 -3.18964511e-01 7.91223422e-02 2.44229570e-01
2.29873002e-01 2.19984695e-01 -1.06718671e+00 1.65951025e+00
-7.12137282e-01 4.95832443e-01 1.04203954e-01 -1.44502974e+00
5.18851042e-01 6.18221164e-01 2.30926394e-01 -6.53924942e-01
2.77835995e-01 9.04918313e-02 1.15990281e-01 -5.34349382e-01
-3.38857442e-01 -9.76269394e-02 -5.97131765e-03 7.30749130e-01
-7.56220818e-02 -1.35122463e-01 -1.99620351e-01 -1.50533289e-01
1.07673264e+00 -2.40991116e-01 7.18664169e-01 -4.00126845e-01
4.08852786e-01 -5.45155227e-01 3.33785117e-01 1.17432487e+00
-3.14890742e-02 5.21202460e-02 6.52516261e-02 5.55390716e-02
-3.16349328e-01 -1.04336667e+00 -2.93984622e-01 9.70140874e-01
-4.56364127e-03 7.40372315e-02 -7.03334033e-01 -1.23546995e-01
-9.92948338e-02 1.30256057e+00 -4.80228305e-01 -6.22998059e-01
-3.63118500e-02 -8.90780568e-01 1.20836474e-01 3.82739246e-01
3.38896334e-01 -1.27948523e+00 -9.13886905e-01 3.05367202e-01
-2.37408116e-01 -6.90308392e-01 -7.11824775e-01 3.58542919e-01
-8.44987452e-01 -5.92325270e-01 -1.05798841e+00 -5.17476201e-01
1.82429805e-01 3.83088887e-01 7.88456440e-01 -4.78632689e-01
-2.92680115e-02 5.16061306e-01 -2.83324383e-02 -8.35202694e-01
2.93616384e-01 -1.23130821e-01 6.33496821e-01 2.74223626e-01
6.01854146e-01 -7.85836577e-01 -6.29760861e-01 7.83444121e-02
-5.72497845e-01 2.78385371e-01 7.33772516e-01 1.24970460e+00
4.08324838e-01 -1.97312403e-02 9.04032052e-01 -2.79946178e-01
1.47129774e+00 -5.11015296e-01 -4.95339960e-01 6.49085939e-01
-5.13386667e-01 -4.08887342e-02 4.65070307e-01 -1.16810715e+00
-9.82423604e-01 -2.32661322e-01 3.99591058e-01 -3.21099252e-01
7.97895715e-02 5.28770030e-01 -3.93137962e-01 1.85495302e-01
4.98693854e-01 8.32660019e-01 -3.27906489e-01 -2.56886482e-01
5.41957095e-02 1.28155446e+00 6.61135614e-01 -6.67462409e-01
2.00562090e-01 9.39258337e-02 -4.80393261e-01 -9.33551669e-01
-7.79471755e-01 -1.58942953e-01 -1.19317792e-01 -1.59217149e-01
6.03386521e-01 -7.88404882e-01 -1.17500710e+00 4.99269366e-01
-1.25611734e+00 -4.93087828e-01 1.30445287e-01 9.04864728e-01
-7.20097899e-01 -1.01106279e-02 -1.83497742e-01 -1.46257019e+00
-5.51051438e-01 -1.17105949e+00 8.60032260e-01 2.12868005e-01
-3.51277858e-01 -4.13908362e-01 -1.40729904e-01 2.29187876e-01
2.63239592e-01 -4.76622939e-01 9.43871498e-01 -7.62008607e-01
-1.87905610e-01 -3.31053436e-01 1.86068285e-02 3.15040201e-01
3.88286084e-01 -1.10435629e+00 -7.92787850e-01 -6.07638836e-01
6.33052170e-01 -3.85688752e-01 5.15171766e-01 8.04030001e-01
1.73270369e+00 -2.27643713e-01 -5.18533051e-01 5.23935616e-01
8.05353284e-01 7.67794490e-01 3.96597862e-01 1.50333479e-01
3.15039307e-01 2.61226267e-01 9.20949429e-02 7.37093449e-01
1.97159443e-02 6.65106773e-01 -9.15610492e-02 5.73831677e-01
4.23222095e-01 -1.30040608e-02 4.47278589e-01 6.37163699e-01
1.25594437e-01 -4.11736578e-01 -7.77117908e-01 6.46605432e-01
-1.74636972e+00 -6.39508486e-01 5.77655971e-01 2.79659367e+00
1.11593807e+00 1.17859676e-01 -6.50415123e-02 1.16209686e-01
7.67511249e-01 -4.22649384e-01 -1.34365737e+00 -1.85645446e-01
1.15470789e-01 5.36162317e-01 4.07029092e-01 1.40899837e-01
-5.82547665e-01 6.44525588e-01 6.08022785e+00 1.22440410e+00
-1.10174012e+00 2.48743713e-01 5.55695653e-01 -8.02137077e-01
6.15349859e-02 -6.20176136e-01 -5.36569178e-01 7.76213169e-01
1.21668649e+00 -6.49112344e-01 1.37532580e+00 3.74265075e-01
4.81174976e-01 -2.91060448e-01 -1.39025152e+00 1.78844595e+00
8.52359608e-02 -9.26764965e-01 -1.13433659e-01 1.37020767e-01
4.51591223e-01 -1.01923309e-01 1.38103083e-01 2.88250744e-01
1.00667924e-01 -1.13867462e+00 6.95048511e-01 6.85051620e-01
1.09627497e+00 -6.69410706e-01 2.70468384e-01 7.46239722e-01
-5.93596876e-01 -4.73526120e-01 -5.79325370e-02 -2.00419322e-01
1.70885876e-01 4.58107233e-01 -2.85413384e-01 1.87037379e-01
6.14620507e-01 2.86826640e-01 1.46409214e-01 1.19476795e+00
-3.57367426e-01 1.02994204e+00 -3.73552978e-01 -5.94823360e-01
-1.93693787e-01 -3.07946980e-01 6.37125194e-01 8.28392148e-01
6.94612503e-01 5.56157827e-01 -2.04514533e-01 7.67039120e-01
-7.73036405e-02 2.64719054e-02 -4.64982569e-01 -7.94936270e-02
8.58290017e-01 7.80163348e-01 -3.99540156e-01 -3.29143673e-01
-3.47135127e-01 1.22326982e+00 2.49003604e-01 7.02174902e-01
-6.67990804e-01 -4.79909211e-01 5.01351058e-01 -3.95752579e-01
2.40339451e-02 -1.07818156e-01 -5.47731519e-01 -1.27992022e+00
8.39848965e-02 -1.21120739e+00 1.36779428e-01 -9.22574103e-01
-1.27489078e+00 4.09608424e-01 2.07227990e-01 -8.86675656e-01
-2.35079154e-01 -5.62589407e-01 -5.90291321e-01 1.23103833e+00
-1.22669268e+00 -4.14296865e-01 1.84677690e-01 9.52185869e-01
7.90575922e-01 -2.71801293e-01 8.56185198e-01 -7.20967799e-02
-8.01731586e-01 7.26530612e-01 3.88824195e-01 -2.92700946e-01
6.15924418e-01 -8.84438813e-01 2.95940172e-02 5.63467979e-01
1.30779400e-01 8.78412247e-01 6.40089452e-01 -6.49517536e-01
-1.67195034e+00 -6.40691161e-01 3.60249728e-01 -1.98155835e-01
6.75759256e-01 -5.13939738e-01 -7.00512826e-01 5.11631131e-01
2.34250546e-01 -4.57970798e-01 7.10183382e-01 1.06242180e-01
3.53083074e-01 -1.37170002e-01 -8.59978139e-01 8.71584356e-01
1.05873704e+00 -4.95811462e-01 -8.60551178e-01 5.12783766e-01
3.91419888e-01 -6.41230680e-03 -2.87930191e-01 -4.56934199e-02
8.00651491e-01 -2.90902346e-01 9.73133743e-01 -7.45566905e-01
-1.61738560e-01 3.73751909e-01 7.51422048e-02 -1.88474274e+00
-2.56059259e-01 -1.32835925e+00 -5.14576077e-01 7.23354697e-01
2.29712129e-01 -1.03301120e+00 4.73754376e-01 8.69357228e-01
3.41555953e-01 -6.66142106e-01 -1.07137573e+00 -9.06064332e-01
5.52099645e-02 -9.00534809e-01 7.02166736e-01 3.86957854e-01
7.27688789e-01 4.99541044e-01 -7.58609354e-01 6.43047094e-02
7.32762814e-01 -6.59552217e-02 2.90967613e-01 -1.18064117e+00
-5.89111805e-01 -6.49234354e-01 1.74589366e-01 -1.48878992e+00
4.11850154e-01 -8.09963703e-01 2.75735885e-01 -1.04876196e+00
5.42438865e-01 -1.10366821e-01 -9.61092770e-01 3.57673854e-01
-4.07394201e-01 -2.06191689e-01 -1.77418172e-01 1.40225559e-01
-3.82939219e-01 7.88985252e-01 9.31353390e-01 -2.79815704e-01
-4.77759123e-01 4.48747396e-01 -9.21756268e-01 6.53394938e-01
7.17446148e-01 -4.04200017e-01 -7.07214713e-01 -4.46955800e-01
1.61293402e-01 6.34283423e-01 1.35562733e-01 -7.65282154e-01
5.14085054e-01 -2.17628002e-01 5.36620855e-01 -2.50499576e-01
3.45902383e-01 -6.44172788e-01 -2.44432405e-01 4.56413537e-01
-5.62393725e-01 -1.34078950e-01 2.26089463e-01 9.19921041e-01
2.55689412e-01 -1.57088950e-01 4.44527119e-01 5.19478507e-02
-6.89696670e-02 2.57336199e-01 -8.82278681e-01 -3.90672833e-02
1.00988889e+00 -8.18713829e-02 2.28783980e-01 -6.92034662e-01
-6.65261447e-01 3.46937060e-01 -2.93896377e-01 4.71134335e-01
7.18196630e-01 -1.27110159e+00 -5.17859817e-01 5.26200116e-01
-9.50094089e-02 -4.66500789e-01 1.15286194e-01 9.26259935e-01
5.36663353e-01 8.85525346e-01 -8.46773908e-02 -4.05493766e-01
-1.00240314e+00 5.50902486e-01 1.92585945e-01 -9.55280215e-02
-4.24880475e-01 9.00678456e-01 3.90277565e-01 2.81382054e-01
6.30373359e-01 -3.20339978e-01 -3.67622107e-01 -6.87042549e-02
1.02444148e+00 3.33203435e-01 9.24085900e-02 1.15206949e-01
-4.01058286e-01 4.54933234e-02 4.55089770e-02 -5.39828956e-01
1.40319252e+00 -1.68020874e-01 1.00863211e-01 6.14484668e-01
4.91159856e-01 -3.37494135e-01 -1.04973829e+00 -4.46890354e-01
-9.35353562e-02 -4.97444510e-01 5.72612464e-01 -1.05253351e+00
-7.17947066e-01 9.25844431e-01 7.42692113e-01 1.79129049e-01
1.25397038e+00 -3.42198521e-01 7.53692269e-01 7.25214720e-01
9.10652220e-01 -1.09562337e+00 -1.35573763e-02 1.46922514e-01
1.01014233e+00 -8.81536365e-01 -2.33363912e-01 4.12139416e-01
-8.11460614e-01 7.81478822e-01 1.95763960e-01 2.01171245e-02
7.80074179e-01 2.09445521e-01 -6.03481472e-01 1.53998300e-01
-8.44351351e-01 2.18730345e-01 3.94355327e-01 5.81771672e-01
1.95510358e-01 3.30663413e-01 -4.75335270e-01 1.42867315e+00
-7.41895512e-02 4.47519332e-01 2.99477220e-01 6.66667938e-01
-1.81240276e-01 -7.92851567e-01 -4.95989740e-01 9.48027372e-01
-4.46465254e-01 -5.64294159e-01 -1.28656089e-01 1.71959445e-01
-5.04398823e-01 1.27000177e+00 -1.58417970e-01 -2.27160528e-01
2.30559334e-01 4.24450368e-01 6.31453454e-01 -5.07999241e-01
-2.02983946e-01 2.40875006e-01 -3.38466078e-01 -5.18831372e-01
-3.55060771e-02 -8.88125658e-01 -5.52129030e-01 1.73171028e-01
-6.49837255e-01 1.57982767e-01 6.69144273e-01 1.20215952e+00
4.27233756e-01 7.72297561e-01 5.43015003e-01 -1.13714433e+00
-9.99797165e-01 -1.30041790e+00 -8.34172487e-01 4.76089939e-02
3.35579127e-01 -7.77165532e-01 -2.83637524e-01 -2.79639453e-01] | [13.150221824645996, 3.430433511734009] |
10e52217-9ccc-44c6-a499-0fa776116975 | spa-former-transformer-image-shadow-detection | 2206.10910 | null | https://arxiv.org/abs/2206.10910v3 | https://arxiv.org/pdf/2206.10910v3.pdf | SpA-Former: Transformer image shadow detection and removal via spatial attention | In this paper, we propose an end-to-end SpA-Former to recover a shadow-free image from a single shaded image. Unlike traditional methods that require two steps for shadow detection and then shadow removal, the SpA-Former unifies these steps into one, which is a one-stage network capable of directly learning the mapping function between shadows and no shadows, it does not require a separate shadow detection. Thus, SpA-former is adaptable to real image de-shadowing for shadows projected on different semantic regions. SpA-Former consists of transformer layer and a series of joint Fourier transform residual blocks and two-wheel joint spatial attention. The network in this paper is able to handle the task while achieving a very fast processing efficiency. Our code is relased on https://github.com/zhangbaijin/SpA-Former-shadow-removal | ['Shan Ying Zhu', 'Chao Chen Gu', 'Xiao Feng Zhang'] | 2022-06-22 | null | null | null | null | ['shadow-removal', 'shadow-detection-and-removal', 'shadow-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.33644164e-01 2.38230005e-01 5.71223259e-01 -5.71052790e-01
-5.41198969e-01 -2.00919107e-01 3.59137446e-01 -4.83416617e-01
-3.26851279e-01 5.96738219e-01 -3.12365964e-02 -5.57592869e-01
4.03508574e-01 -7.22594380e-01 -8.37126791e-01 -9.00678694e-01
2.33292565e-01 1.83536544e-01 9.50887382e-01 -2.10868195e-02
1.72162786e-01 5.17689884e-01 -1.35021436e+00 3.40733856e-01
8.50625455e-01 1.03124654e+00 8.80037069e-01 7.40275919e-01
3.90832536e-02 6.04372740e-01 -5.14769197e-01 9.79353487e-02
3.57270330e-01 -5.10386646e-01 -3.82343203e-01 -1.22805767e-01
6.70121729e-01 -5.34683764e-01 -3.32632214e-01 7.12043464e-01
5.44958055e-01 2.69060321e-02 4.80735034e-01 -1.12885916e+00
-4.83582884e-01 -2.66599399e-03 -7.37511814e-01 2.11737737e-01
2.91201949e-01 9.86153930e-02 5.84623158e-01 -1.04338014e+00
2.87676632e-01 1.34185076e+00 8.18641484e-01 1.35923818e-01
-1.10390580e+00 -7.73164034e-01 1.86593041e-01 3.24174792e-01
-1.23823190e+00 -3.86834025e-01 8.50036204e-01 -1.03565685e-01
6.71072543e-01 2.86299825e-01 5.50349057e-01 9.99220788e-01
4.50955838e-01 9.05328274e-01 1.64475834e+00 -5.05073249e-01
6.64213672e-02 7.58972578e-03 7.42327273e-02 8.74500573e-01
-1.36048021e-03 2.84269869e-01 -6.17308378e-01 1.19226865e-01
7.41433501e-01 8.16993415e-02 -6.60462260e-01 -4.72034216e-01
-1.07362008e+00 4.26573694e-01 6.17432535e-01 -1.63158905e-02
-1.93064600e-01 2.59945899e-01 -1.37838675e-02 9.05913487e-02
5.53230643e-01 -1.43453270e-01 -5.47367871e-01 4.74884242e-01
-1.06080091e+00 1.01824798e-01 6.12933874e-01 7.27487683e-01
8.26477706e-01 6.33123964e-02 -1.32127002e-01 3.99970919e-01
1.74518257e-01 9.34792876e-01 1.65449604e-01 -6.93442106e-01
1.01784609e-01 1.15531839e-01 2.75304437e-01 -8.50467861e-01
-5.08020282e-01 -3.14927697e-01 -6.73768282e-01 7.10837305e-01
2.58631885e-01 -1.63271233e-01 -1.32214725e+00 1.46178794e+00
5.53428531e-01 6.04828954e-01 -1.33032709e-01 9.29947078e-01
8.66690993e-01 7.84212470e-01 -4.23462629e-01 -6.28392920e-02
1.45649683e+00 -1.33276153e+00 -6.69860661e-01 -6.16578579e-01
-7.37980679e-02 -9.60545778e-01 1.24817073e+00 4.83920336e-01
-1.01821613e+00 -4.96055841e-01 -1.28394902e+00 -4.24383193e-01
-5.49569368e-01 3.18006516e-01 6.16329193e-01 3.46640617e-01
-1.15571892e+00 4.92257625e-01 -7.83130527e-01 -2.18224511e-01
5.17146349e-01 2.00043678e-01 -1.92788959e-01 -3.07615828e-02
-9.07154024e-01 8.57365072e-01 1.69709668e-01 3.68862599e-01
-9.68868077e-01 -6.95166469e-01 -7.52387047e-01 1.71219662e-01
7.03504145e-01 -4.86058980e-01 1.24339581e+00 -1.21334481e+00
-1.68942022e+00 5.65192640e-01 -6.32470310e-01 -3.11640769e-01
6.18716359e-01 -5.77516019e-01 -1.61647633e-01 -6.79579973e-02
1.22516990e-01 1.72050208e-01 1.23466897e+00 -1.68154860e+00
-4.89962757e-01 -3.11875641e-01 -9.47595537e-02 4.83911186e-01
8.57750252e-02 -6.18927404e-02 -6.33712828e-01 -5.87051153e-01
2.06231222e-01 -8.37585270e-01 6.43869191e-02 3.11349332e-01
-5.05167961e-01 2.17726737e-01 1.14491701e+00 -8.05879772e-01
8.60257268e-01 -2.06514692e+00 -1.02757685e-01 -4.21851128e-02
4.71287295e-02 4.31129783e-01 -1.14976548e-01 3.08911264e-01
-7.43310004e-02 -5.68818212e-01 -6.70229912e-01 -7.57052064e-01
-1.75743148e-01 4.87186462e-02 -7.21880555e-01 7.77877212e-01
-6.95927441e-02 7.81864583e-01 -7.16446936e-01 -3.10279250e-01
4.16404843e-01 7.34380186e-01 9.65357572e-02 4.34155822e-01
-1.23492807e-01 2.91379631e-01 -1.66030735e-01 6.11568928e-01
1.30169058e+00 -2.88355071e-02 1.16048001e-01 -3.17508399e-01
-2.24017277e-01 1.09718584e-01 -1.10104120e+00 1.62140822e+00
-6.81184351e-01 8.22974265e-01 5.65466285e-01 -3.98164481e-01
6.56088471e-01 -1.05339758e-01 1.74819648e-01 -9.17017639e-01
1.90632716e-01 1.12223290e-01 -4.44179922e-01 -2.26309001e-01
1.96640700e-01 -1.23216882e-02 3.03930283e-01 4.16881502e-01
-3.00802678e-01 -2.05707118e-01 -3.00512046e-01 6.99962676e-02
9.79533315e-01 5.63581705e-01 9.85116586e-02 -3.59465629e-01
6.58315659e-01 -3.02103251e-01 7.70106733e-01 6.51602209e-01
-1.29739884e-02 9.93770480e-01 1.66284010e-01 -2.68594116e-01
-6.11240387e-01 -1.37177598e+00 -1.28463760e-01 1.18518364e+00
5.95507503e-01 -1.90152511e-01 -6.96543872e-01 -6.99500501e-01
4.39035632e-02 8.70517850e-01 -7.74502456e-01 2.53742514e-03
-5.98673880e-01 -5.11072636e-01 2.73829371e-01 2.93285072e-01
7.93264091e-01 -1.17380917e+00 -1.16332543e+00 -2.12717563e-01
-1.88953757e-01 -9.55975473e-01 -5.23162782e-01 5.95146179e-01
-4.53298420e-01 -1.21681452e+00 -7.60637283e-01 -6.72737658e-01
7.77043402e-01 9.13158476e-01 1.09679687e+00 1.66278183e-01
-4.40026045e-01 6.51427582e-02 -1.97513744e-01 -7.70384908e-01
1.59230858e-01 -2.24929065e-01 -4.27794367e-01 1.32959336e-01
-4.20371406e-02 -8.45874548e-01 -9.98874664e-01 2.68660158e-01
-9.25924778e-01 6.39500082e-01 8.79978597e-01 7.13343740e-01
4.52777207e-01 -6.12606518e-02 -1.64336026e-01 -1.01666582e+00
2.39693105e-01 -1.94624171e-01 -8.39687586e-01 2.78075099e-01
-6.65218174e-01 -1.78427666e-01 6.80146456e-01 -3.26127335e-02
-1.57580447e+00 3.75842392e-01 2.78197825e-02 -3.07051748e-01
-1.61466241e-01 -3.30919743e-01 -2.92552710e-01 -1.47540405e-01
4.61224675e-01 3.94504577e-01 -2.91923761e-01 -4.64541584e-01
4.21233177e-01 3.28117341e-01 6.32671595e-01 -1.61026046e-01
1.19855893e+00 1.01026690e+00 -3.06960046e-02 -6.63939178e-01
-1.23734808e+00 -4.25686121e-01 -8.31883311e-01 -1.93342462e-01
8.71078551e-01 -7.64534056e-01 -4.80906665e-01 8.18722904e-01
-1.10233223e+00 -9.69539106e-01 -1.25260954e-03 -1.70053244e-02
-4.08025354e-01 5.56675017e-01 -3.28259408e-01 -7.57195294e-01
-4.59756404e-01 -8.26227903e-01 1.35472620e+00 3.08145583e-01
3.33470792e-01 -6.63176477e-01 1.29547343e-01 2.12835163e-01
3.22167665e-01 1.71632051e-01 5.99506974e-01 8.94846246e-02
-8.93688083e-01 2.09590495e-01 -6.19520962e-01 3.62184584e-01
9.11375508e-02 -3.97626102e-01 -1.37731397e+00 -4.63570774e-01
8.12563822e-02 -2.39875123e-01 1.33862484e+00 5.55591285e-01
1.23548281e+00 -7.39136711e-02 -3.52289796e-01 1.13906205e+00
1.63607240e+00 1.74624458e-01 7.73628414e-01 3.00058097e-01
8.31017494e-01 3.76284480e-01 7.95221686e-01 8.45269039e-02
3.40405494e-01 6.94091320e-01 6.57732248e-01 -8.22044969e-01
-6.40666544e-01 -3.79754826e-02 5.10205686e-01 1.35545745e-01
9.01232883e-02 -4.80200589e-01 -7.01613784e-01 3.35355580e-01
-1.96447515e+00 -5.98267496e-01 -3.36568266e-01 1.94850004e+00
5.53345442e-01 6.64100498e-02 -3.68817031e-01 -1.79782778e-01
3.37663949e-01 6.10509634e-01 -7.42821336e-01 -1.65977493e-01
-2.59499341e-01 3.24641079e-01 8.10009539e-01 1.06050670e+00
-1.01611650e+00 1.27124274e+00 5.74477482e+00 7.64098287e-01
-1.01085639e+00 2.66715497e-01 4.02467161e-01 2.76254816e-03
-2.72303969e-01 2.47875854e-01 -4.12201762e-01 4.69207734e-01
4.32464838e-01 5.09079874e-01 4.42420512e-01 6.58704758e-01
2.07981020e-01 -7.85434902e-01 -6.16726041e-01 7.49162853e-01
2.81282425e-01 -1.05897748e+00 -4.14097279e-01 -1.92936152e-01
4.56113935e-01 2.28751883e-01 -5.83202802e-02 -5.87612810e-03
2.29552910e-01 -9.73402500e-01 9.30428505e-01 6.11314356e-01
8.09175611e-01 -5.43416917e-01 6.30725324e-01 2.49341011e-01
-1.22427773e+00 -2.53498517e-02 -2.92215526e-01 4.94879000e-02
3.03491861e-01 9.67860758e-01 -7.95622110e-01 6.04111195e-01
9.50881362e-01 2.57560343e-01 -5.64139009e-01 9.17270005e-01
-6.39343619e-01 3.91282886e-01 -1.88519999e-01 3.13213646e-01
-1.40328526e-01 -3.85171056e-01 4.03182238e-01 1.53561032e+00
2.06082225e-01 1.19649053e-01 8.94277915e-02 7.25446939e-01
3.26124042e-01 -3.47953767e-01 -3.53620738e-01 6.28168046e-01
3.10882568e-01 1.44136620e+00 -1.12042940e+00 -3.72353256e-01
-3.04669350e-01 1.79480684e+00 2.62721837e-01 5.55988848e-01
-1.22398734e+00 -5.45042217e-01 3.07895571e-01 2.33648837e-01
3.89798284e-01 -2.65391588e-01 -3.39471549e-01 -9.82120216e-01
1.06479563e-01 -5.61452329e-01 -1.74079072e-02 -1.34390426e+00
-7.71658659e-01 6.47071481e-01 -1.11761622e-01 -6.42457366e-01
4.57687467e-01 -6.89561307e-01 -8.68217289e-01 1.02449238e+00
-2.01261115e+00 -1.41755223e+00 -9.90078807e-01 7.81897366e-01
6.91229045e-01 3.29373896e-01 6.39735460e-01 1.35285929e-01
-4.16147500e-01 2.51051992e-01 1.52578950e-01 -1.61381751e-01
1.00727856e+00 -1.42682898e+00 4.90599245e-01 1.25708354e+00
-8.31682310e-02 1.34825200e-01 9.21348870e-01 -7.25950897e-01
-1.11184025e+00 -9.13831472e-01 6.12574279e-01 -2.97614276e-01
3.05618584e-01 -7.99597979e-01 -9.22327936e-01 6.61435246e-01
3.53460968e-01 8.08445737e-03 5.53925633e-02 -1.65814161e-01
-3.90118420e-01 -4.60525334e-01 -9.37624216e-01 5.70642412e-01
1.07803261e+00 -4.82298464e-01 -3.31232667e-01 6.34668291e-01
5.77660143e-01 -7.38156319e-01 -6.51909858e-02 3.82065475e-01
4.74098414e-01 -1.74469495e+00 1.12692964e+00 2.21346185e-01
1.28476664e-01 -5.89152336e-01 -2.27543768e-02 -1.05334377e+00
-3.60631138e-01 -5.71472704e-01 -1.95799068e-01 8.76078129e-01
1.14628628e-01 -8.81285727e-01 8.17530572e-01 2.53495853e-02
-4.19450521e-01 -8.79945099e-01 -7.61839569e-01 -6.18826091e-01
-3.06394428e-01 -2.40891635e-01 4.03600484e-01 5.07417679e-01
-7.72458613e-01 3.13980937e-01 -4.94466066e-01 6.21773779e-01
8.73139441e-01 5.23453116e-01 8.99353504e-01 -1.07003021e+00
-5.34471214e-01 -1.27572268e-01 3.33430886e-01 -1.16502583e+00
1.01890065e-01 -4.87348706e-01 6.56105220e-01 -1.80704832e+00
2.98452020e-01 -3.56478870e-01 -2.14251027e-01 7.74937510e-01
-2.04233915e-01 3.56230527e-01 2.21469730e-01 1.96602270e-01
-3.63127291e-01 7.49379456e-01 1.16611969e+00 1.98547572e-01
-1.96595177e-01 2.28101224e-01 -4.76950377e-01 9.19257522e-01
9.07682300e-01 -4.55302924e-01 -6.10219598e-01 -4.00189668e-01
-1.15691826e-01 -1.50142342e-01 7.23841488e-01 -1.00591195e+00
-1.60496812e-02 -2.58553118e-01 6.26015484e-01 -8.11056256e-01
7.63077676e-01 -8.44152093e-01 -1.74808688e-02 4.80405182e-01
1.28923222e-01 -2.48329476e-01 1.53213575e-01 6.95662916e-01
2.07237124e-01 3.23869362e-02 9.60007906e-01 -2.26610884e-01
-7.25520253e-01 1.44601837e-01 -2.86474556e-01 -3.42261940e-01
1.02144861e+00 -2.61973172e-01 -4.09794241e-01 -3.07834119e-01
-4.99690026e-01 1.09881565e-01 5.24078846e-01 1.80350095e-01
7.44825542e-01 -8.24475706e-01 -4.98785287e-01 2.40301013e-01
-3.71603042e-01 3.43621105e-01 3.74113202e-01 8.50724161e-01
-7.14842021e-01 2.61810541e-01 -2.67332830e-02 -3.09939355e-01
-1.64113832e+00 3.75165790e-01 4.81404215e-01 -7.08709359e-02
-1.03275871e+00 1.02104247e+00 8.91340554e-01 -3.61551076e-01
3.18158984e-01 -3.35153282e-01 3.64881009e-01 -3.03120852e-01
4.33810383e-01 2.17584953e-01 7.63220116e-02 -5.30114174e-01
-3.69614154e-01 5.28439105e-01 1.88536242e-01 -2.75013745e-01
1.30429113e+00 -2.74427563e-01 -2.85022974e-01 5.19501746e-01
1.01622033e+00 1.89101234e-01 -1.89642739e+00 -2.13352308e-01
-3.70229006e-01 -6.62490904e-01 4.02426720e-01 -1.17768848e+00
-1.13179362e+00 8.12437296e-01 8.65480363e-01 -1.29871666e-01
1.48651826e+00 -1.48604989e-01 9.51314986e-01 2.84003824e-01
1.13933019e-01 -1.01954114e+00 8.34463909e-02 6.07631266e-01
1.12383258e+00 -1.16946900e+00 3.85699749e-01 -7.06333697e-01
-4.13667619e-01 9.95521426e-01 6.37036324e-01 -2.24320516e-01
8.49989116e-01 5.91883302e-01 2.75956750e-01 -1.61209360e-01
-3.74670774e-01 -2.95172364e-01 3.64294916e-01 7.13964760e-01
2.84017287e-02 -8.57026950e-02 4.30756621e-02 1.43806234e-01
-1.23223327e-01 -2.33993262e-01 3.95520896e-01 9.40130115e-01
-6.07248425e-01 -8.39237094e-01 -5.47082007e-01 7.93111920e-02
6.26120716e-02 -4.10523623e-01 -5.56538403e-01 6.80963993e-01
3.97641033e-01 8.12669516e-01 -3.07340734e-02 -2.09468365e-01
2.66325027e-01 -1.09848253e-01 4.40542638e-01 -3.65664631e-01
-3.24310482e-01 1.84301600e-01 -7.90852904e-02 -1.02935946e+00
-2.56248504e-01 -6.13006771e-01 -1.41800344e+00 -6.18850365e-02
-2.96226680e-01 -2.79583007e-01 7.18228877e-01 7.83542752e-01
3.68793339e-01 7.43033230e-01 5.74363351e-01 -1.26353467e+00
2.16003899e-02 -7.22443819e-01 -6.27424121e-01 -7.63002411e-02
7.21315801e-01 -6.47676170e-01 -4.01677728e-01 -1.94124684e-01] | [10.846233367919922, -4.108004093170166] |
e35f441f-9777-4140-b51f-328eb2c7cbea | weakly-supervised-photo-realistic-texture | 2106.08148 | null | https://arxiv.org/abs/2106.08148v1 | https://arxiv.org/pdf/2106.08148v1.pdf | Weakly-Supervised Photo-realistic Texture Generation for 3D Face Reconstruction | Although much progress has been made recently in 3D face reconstruction, most previous work has been devoted to predicting accurate and fine-grained 3D shapes. In contrast, relatively little work has focused on generating high-fidelity face textures. Compared with the prosperity of photo-realistic 2D face image generation, high-fidelity 3D face texture generation has yet to be studied. In this paper, we proposed a novel UV map generation model that predicts the UV map from a single face image. The model consists of a UV sampler and a UV generator. By selectively sampling the input face image's pixels and adjusting their relative locations, the UV sampler generates an incomplete UV map that could faithfully reconstruct the original face. Missing textures in the incomplete UV map are further full-filled by the UV generator. The training is based on pseudo ground truth blended by the 3DMM texture and the input face texture, thus weakly supervised. To deal with the artifacts in the imperfect pseudo UV map, multiple partial UV map discriminators are leveraged. | ['Liming Chen', 'Yunhong Wang', 'Zehua Fu', 'Di Huang', 'Xiangnan Yin'] | 2021-06-14 | null | null | null | null | ['texture-synthesis', '3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.84363782e-01 4.83018816e-01 4.29058298e-02 -4.33024317e-01
-6.63087904e-01 -2.14401141e-01 8.30546737e-01 -5.74749827e-01
3.55308294e-01 6.80091023e-01 -9.98974293e-02 -9.60829109e-03
3.72704685e-01 -1.20979464e+00 -9.80477154e-01 -7.80095994e-01
5.03922164e-01 7.84085512e-01 -1.10240772e-01 3.48822251e-02
1.41363755e-01 4.76997137e-01 -1.97749221e+00 5.33140838e-01
6.20147467e-01 1.08553696e+00 3.09808880e-01 4.88646448e-01
-4.46041644e-01 3.25320929e-01 -3.56804132e-01 -3.78034443e-01
5.32808065e-01 -9.17897582e-01 -4.79553938e-01 4.53159481e-01
8.80263329e-01 -5.99948764e-01 -2.37098336e-01 8.70210767e-01
4.56574708e-01 -7.18236193e-02 9.77134407e-01 -8.80548716e-01
-8.53353500e-01 1.51046365e-01 -7.61697769e-01 -5.14936864e-01
5.29054463e-01 5.96625991e-02 4.97976065e-01 -1.06410718e+00
1.02414954e+00 1.53087652e+00 5.06915867e-01 9.48223948e-01
-1.48522139e+00 -5.83900928e-01 -3.02703321e-01 -2.89180905e-01
-1.32495701e+00 -5.70803761e-01 1.12390792e+00 -4.28105503e-01
5.27127862e-01 9.79468152e-02 6.29101336e-01 1.32823634e+00
1.86469421e-01 4.01909560e-01 1.49148226e+00 -7.00401425e-01
1.76479742e-01 3.23253334e-01 -6.65707111e-01 9.71418262e-01
-7.90353566e-02 4.91762668e-01 -6.66724741e-01 -2.83246934e-01
1.27189791e+00 -2.40891293e-01 -1.66245446e-01 -3.58207852e-01
-6.33710504e-01 4.08988565e-01 1.83286041e-01 -1.63233891e-01
-3.09107929e-01 -6.69277236e-02 -2.07232490e-01 2.79147208e-01
1.09898937e+00 -7.75146559e-02 -7.86018744e-02 1.94250897e-01
-1.12854838e+00 3.88327181e-01 5.12859404e-01 8.51038933e-01
1.17757940e+00 1.70451865e-01 -2.31532529e-01 1.00081444e+00
4.54864293e-01 1.05301082e+00 -1.45794109e-01 -9.11985993e-01
2.15932727e-01 5.42954385e-01 3.47357169e-02 -8.31379533e-01
2.38547996e-01 -5.75472452e-02 -6.84430718e-01 6.40905023e-01
4.80594069e-01 6.29740208e-02 -1.38105392e+00 1.52907157e+00
6.95591211e-01 3.41986984e-01 -3.19187909e-01 1.03440762e+00
7.17577517e-01 6.81977570e-01 -5.38256526e-01 -5.11265509e-02
1.02014720e+00 -4.81473744e-01 -5.35393476e-01 -1.72784925e-01
-1.56911373e-01 -1.02510464e+00 8.88874173e-01 1.10563241e-01
-1.32064915e+00 -6.02788687e-01 -8.48024189e-01 -4.90784384e-02
-1.91213191e-02 1.52381137e-01 3.45436603e-01 8.47793579e-01
-1.10050225e+00 5.52423537e-01 -6.45595074e-01 -1.31793261e-01
8.24518085e-01 1.29510596e-01 -4.39213783e-01 -3.79552394e-01
-8.03114176e-01 6.02767050e-01 -1.15019448e-01 2.27067381e-01
-1.23329306e+00 -7.20084846e-01 -7.45384395e-01 -3.96971673e-01
1.70916736e-01 -8.03534627e-01 9.96302307e-01 -1.04819596e+00
-1.74692523e+00 1.16461539e+00 -4.37885225e-01 1.69025570e-01
7.08768725e-01 4.30156857e-01 -1.35298803e-01 7.62052312e-02
1.06865115e-01 8.24888527e-01 1.47207689e+00 -1.90485287e+00
-3.17873836e-01 -6.43458307e-01 -2.87892818e-01 1.39776599e-02
7.53253996e-02 -3.25751871e-01 -3.68660420e-01 -7.46681392e-01
3.43602717e-01 -6.44300699e-01 2.16134146e-01 4.39679831e-01
-2.29994208e-01 1.88616455e-01 8.73070061e-01 -5.49254954e-01
5.90333939e-01 -2.05258989e+00 1.08232833e-02 3.91221255e-01
-1.55010968e-01 -5.13930479e-03 -2.79681057e-01 5.21775007e-01
1.33304462e-01 -3.61228921e-02 -4.63038176e-01 -7.74977863e-01
-3.31235640e-02 3.42096567e-01 -6.33095264e-01 4.54636782e-01
5.16001105e-01 7.01893032e-01 -8.73579860e-01 -4.64074522e-01
2.27947250e-01 9.60468590e-01 -7.93148458e-01 1.81722835e-01
-5.70718765e-01 7.53053665e-01 -3.84908915e-01 1.06722450e+00
1.20180976e+00 1.39110638e-02 3.14235836e-01 -8.02532732e-02
5.64278066e-02 -2.31154054e-01 -1.05900192e+00 1.64222479e+00
-3.22864562e-01 3.37830096e-01 2.91623771e-01 -6.26329899e-01
1.44233859e+00 3.59721988e-01 3.94107044e-01 -8.86966646e-01
2.05887884e-01 3.85656267e-01 -5.87265730e-01 -2.36380011e-01
4.03171569e-01 -5.00131726e-01 5.17558157e-01 8.12871695e-01
7.03690797e-02 -4.77041155e-01 -2.42910355e-01 -3.04118216e-01
4.65307832e-01 9.48197186e-01 -2.06807733e-01 -2.42162764e-01
4.78377581e-01 -8.52875710e-02 4.24840629e-01 6.38644993e-01
2.13611618e-01 1.21163070e+00 2.95311421e-01 -5.90869725e-01
-1.32095122e+00 -1.37441409e+00 -5.62798738e-01 4.91505593e-01
1.57334715e-01 -9.14690346e-02 -1.30255592e+00 -6.65089369e-01
8.30602869e-02 2.22027019e-01 -9.93756711e-01 3.16954367e-02
-5.32774150e-01 -6.05096400e-01 4.42585140e-01 5.36358133e-02
6.57400966e-01 -1.11960888e+00 -1.69716120e-01 8.39798711e-03
-2.51591355e-02 -8.20777178e-01 -5.56100488e-01 -3.00420672e-01
-8.49336088e-01 -1.17318869e+00 -6.70635760e-01 -7.18179226e-01
1.13339210e+00 2.91923016e-01 1.02981091e+00 4.67575267e-02
-5.20989597e-01 1.73709497e-01 -1.90988764e-01 -3.67446572e-01
-6.96164310e-01 -5.03934443e-01 1.06348477e-01 5.07195115e-01
6.63423017e-02 -3.82091641e-01 -5.90432942e-01 4.23362553e-01
-9.53885913e-01 2.49471471e-01 3.39174598e-01 1.09473813e+00
9.27577615e-01 -1.03150591e-01 4.33635473e-01 -1.14475584e+00
1.73970908e-01 -2.77470380e-01 -6.62899017e-01 3.15992571e-02
-4.39825118e-01 1.92540646e-01 4.05741036e-01 -3.81133318e-01
-1.69738650e+00 1.43536568e-01 -3.33635390e-01 -6.35509968e-01
-2.05483019e-01 -1.11905582e-01 -1.09261118e-01 -2.86495149e-01
7.90616393e-01 2.54949987e-01 4.41381723e-01 -6.44577384e-01
2.91050673e-01 6.76182628e-01 3.85519624e-01 -8.80052626e-01
8.42297494e-01 8.96197915e-01 5.20379245e-02 -9.81489122e-01
-6.89791560e-01 3.24862033e-01 -4.95156080e-01 -4.45890516e-01
5.99212170e-01 -7.00064182e-01 -3.59681189e-01 5.13221502e-01
-9.34730113e-01 -5.57914197e-01 -5.11248112e-01 -6.77933544e-02
-6.83505476e-01 2.26028338e-01 -4.14175242e-01 -1.03852880e+00
-1.39798835e-01 -1.14500487e+00 1.58611226e+00 1.23621225e-01
6.36701584e-02 -7.64297247e-01 7.52501413e-02 4.49835390e-01
1.76340252e-01 2.37183556e-01 8.63615692e-01 6.10015452e-01
-8.57864082e-01 -8.36020932e-02 -1.70789242e-01 3.57897282e-01
3.61635774e-01 7.23588765e-02 -1.46484661e+00 -1.50045365e-01
2.27489471e-02 -3.73117089e-01 6.85412228e-01 3.94090801e-01
9.31492507e-01 -1.06642738e-01 -3.30775261e-01 7.41152823e-01
1.26319110e+00 3.94994915e-02 5.68405092e-01 -9.63856950e-02
5.06106973e-01 8.00216615e-01 4.33985233e-01 4.80393589e-01
1.27589285e-01 7.18814194e-01 4.69511986e-01 -3.19497213e-02
-7.43659914e-01 -8.61584544e-01 3.15382600e-01 8.15109685e-02
-4.92021888e-01 -5.95744513e-02 -6.04727924e-01 1.69140786e-01
-1.43560493e+00 -1.03628087e+00 7.64298439e-02 2.29377651e+00
7.78744161e-01 -1.12517774e-01 -2.49798730e-01 -7.91065544e-02
6.76991045e-01 3.14965993e-01 -4.30184394e-01 -2.49046654e-01
-3.86561304e-01 4.41124141e-01 1.31024763e-01 6.88152313e-01
-6.75985336e-01 9.70886648e-01 6.71176767e+00 7.60949433e-01
-1.02970862e+00 -5.95094003e-02 7.52416015e-01 1.92207471e-02
-8.33017051e-01 8.92055184e-02 -8.08339417e-01 4.23560947e-01
2.78545111e-01 3.17564547e-01 5.32595277e-01 5.38379312e-01
2.98654497e-01 -2.48607025e-01 -7.98624635e-01 1.16116202e+00
3.22555512e-01 -1.43481123e+00 2.42003918e-01 4.79172826e-01
1.26925039e+00 -4.71923977e-01 3.24866742e-01 -2.13512644e-01
1.20318085e-01 -9.82635379e-01 1.05931556e+00 8.17699850e-01
1.43760216e+00 -5.22363126e-01 2.87974656e-01 3.79429579e-01
-9.66771364e-01 1.50806978e-01 -6.08492315e-01 7.64917061e-02
1.03109121e-01 7.97405720e-01 -1.01547301e+00 5.85938036e-01
6.05529845e-01 3.83479178e-01 -2.69610107e-01 4.17927533e-01
-3.09535503e-01 4.33154196e-01 -2.75303036e-01 5.14149725e-01
-1.18350126e-01 -7.06684709e-01 4.09226298e-01 7.07035303e-01
4.72886503e-01 1.13728056e-02 3.67624946e-02 1.29924846e+00
-2.49410525e-01 -7.98299834e-02 -8.48515093e-01 4.47223298e-02
3.13803226e-01 1.04659665e+00 -6.17276728e-01 -2.14781269e-01
-3.22584093e-01 1.03617287e+00 2.87330061e-01 3.84686947e-01
-5.02653062e-01 3.72046441e-01 7.72144496e-01 5.62285066e-01
-1.80233065e-02 -7.56248012e-02 -4.09442097e-01 -1.04745770e+00
6.89726695e-02 -5.20042956e-01 -1.08706340e-01 -8.57391655e-01
-1.42463231e+00 6.98440552e-01 -1.97721377e-01 -1.09545636e+00
-2.70286858e-01 -5.58394670e-01 -4.02800649e-01 1.28618753e+00
-1.45472050e+00 -1.38568032e+00 -3.69696021e-01 3.10861886e-01
3.84959072e-01 -1.53889820e-01 9.66939569e-01 1.18918948e-01
-2.96310544e-01 4.84251857e-01 8.50392804e-02 -1.69955537e-01
7.52556920e-01 -9.08716083e-01 4.38071311e-01 5.20479202e-01
7.98487216e-02 2.86866546e-01 6.20077431e-01 -1.02399278e+00
-1.62991881e+00 -1.01688099e+00 6.58581913e-01 -5.96608162e-01
2.44332780e-03 -6.60593212e-01 -8.37513447e-01 3.67310554e-01
-4.00883444e-02 3.01936537e-01 2.26191774e-01 -4.44909483e-01
-3.98866057e-01 -1.47418320e-01 -1.47265744e+00 5.96467674e-01
1.35598457e+00 -6.98583066e-01 -1.53872356e-01 2.10444987e-01
-7.40590543e-02 -5.75609684e-01 -6.92105114e-01 3.40384692e-01
7.09103465e-01 -1.06865692e+00 1.06030214e+00 -1.79393709e-01
5.45451939e-01 -3.85080129e-01 -1.36448503e-01 -1.22229433e+00
-3.14612091e-02 -3.15668613e-01 -9.65999998e-03 1.11826015e+00
1.77125186e-01 -5.05427122e-01 1.32387853e+00 2.63532072e-01
-1.26798570e-01 -6.74650669e-01 -7.44906604e-01 -5.92838824e-01
-1.16101339e-01 -1.55027375e-01 9.92255270e-01 7.67775834e-01
-5.33613205e-01 -5.07711545e-02 -6.22949421e-01 6.98225498e-02
1.09749377e+00 5.71355581e-01 8.28923106e-01 -1.25054502e+00
-2.46069267e-01 -2.34128043e-01 2.80329343e-02 -9.45597887e-01
2.18251660e-01 -1.01287341e+00 2.04493403e-01 -1.40386128e+00
2.16035113e-01 -6.29424453e-01 4.79722738e-01 3.75341982e-01
-2.01798938e-02 8.44396412e-01 -3.01091999e-01 2.00040415e-01
2.96045154e-01 7.68608510e-01 1.80488241e+00 -1.09425046e-01
-1.26773477e-01 -1.77032650e-01 -4.49844718e-01 5.14331520e-01
5.16341925e-01 -2.06707627e-01 -5.20890951e-01 -5.20706952e-01
1.34691503e-02 2.61805356e-01 5.05682588e-01 -6.14212394e-01
-1.31197199e-01 -2.96470761e-01 8.69946182e-01 -4.94180590e-01
6.42651260e-01 -6.14145279e-01 5.17015934e-01 -1.00527525e-01
1.79000959e-01 -4.38142508e-01 -5.95281236e-02 5.26470482e-01
-1.81374922e-01 -1.03602700e-01 8.74037027e-01 -2.99692482e-01
-3.51615429e-01 7.99127400e-01 -3.40491533e-02 -1.91106990e-01
8.34435761e-01 -6.86264813e-01 -1.94342330e-01 -1.10461950e-01
-5.91889203e-01 -3.66414070e-01 1.06305575e+00 4.54599619e-01
8.42205763e-01 -1.42468488e+00 -7.56187677e-01 1.01626134e+00
6.18336443e-03 2.80591011e-01 5.75344265e-01 1.40344352e-01
-5.65834939e-01 6.00422882e-02 -2.09480554e-01 -6.79642498e-01
-9.82941985e-01 1.92816779e-01 4.23925102e-01 3.64366293e-01
-8.18361759e-01 7.80070901e-01 3.16931188e-01 -6.43430471e-01
-2.44473353e-01 3.70711774e-01 2.14407548e-01 -8.82930830e-02
6.90895677e-01 9.20904949e-02 2.19070926e-01 -8.06209385e-01
7.69210681e-02 9.74895000e-01 3.38179678e-01 -3.62564027e-01
1.28576231e+00 -5.20131662e-02 -2.05351874e-01 1.50764555e-01
9.77816164e-01 6.39210176e-03 -1.86065650e+00 -1.32210895e-01
-4.28123802e-01 -1.00599778e+00 4.35382575e-02 -5.88277102e-01
-1.07014310e+00 9.42895234e-01 5.01576722e-01 -1.53930604e-01
1.06958580e+00 -1.34726521e-02 4.76048440e-01 -1.62680030e-01
8.01369190e-01 -1.02832997e+00 6.97276518e-02 3.48660648e-01
8.08839619e-01 -1.07348335e+00 -1.20620385e-01 -7.13063955e-01
-3.80112410e-01 1.10174227e+00 5.22982955e-01 -9.46184099e-02
5.80322087e-01 4.15967256e-01 1.03187799e-01 -2.55820841e-01
-5.45093000e-01 8.12919140e-02 2.09112376e-01 8.58002961e-01
2.93335259e-01 -5.79486880e-03 1.28421411e-01 8.68722871e-02
-1.99623123e-01 1.66756466e-01 3.19124132e-01 8.50952268e-01
-3.19384098e-01 -1.44678009e+00 -6.53399706e-01 4.10813063e-01
-7.38134012e-02 8.07164088e-02 -3.20218176e-01 3.69322002e-01
2.39232108e-01 6.25009239e-01 2.17084169e-01 -1.55593649e-01
2.79767126e-01 2.80045033e-01 1.17388773e+00 -6.68886602e-01
-1.32380664e-01 1.23549402e-01 -9.09878537e-02 -5.05774379e-01
-1.51261464e-01 -7.22705424e-01 -9.34684277e-01 -3.80010188e-01
-6.21448085e-03 -1.57086745e-01 6.88550055e-01 6.51782930e-01
3.65883529e-01 1.04827993e-01 8.16926777e-01 -1.20891511e+00
-2.20145270e-01 -7.61456966e-01 -9.16459501e-01 4.11963016e-01
2.20336989e-01 -9.46556509e-01 -3.90440941e-01 1.95676759e-01] | [12.713021278381348, -0.3267768323421478] |
3640d994-c92b-4451-8551-6d1bd62eae75 | one-shot-object-detection-with-co-attention-1 | 1911.12529 | null | https://arxiv.org/abs/1911.12529v1 | https://arxiv.org/pdf/1911.12529v1.pdf | One-Shot Object Detection with Co-Attention and Co-Excitation | This paper aims to tackle the challenging problem of one-shot object detection. Given a query image patch whose class label is not included in the training data, the goal of the task is to detect all instances of the same class in a target image. To this end, we develop a novel {\em co-attention and co-excitation} (CoAE) framework that makes contributions in three key technical aspects. First, we propose to use the non-local operation to explore the co-attention embodied in each query-target pair and yield region proposals accounting for the one-shot situation. Second, we formulate a squeeze-and-co-excitation scheme that can adaptively emphasize correlated feature channels to help uncover relevant proposals and eventually the target objects. Third, we design a margin-based ranking loss for implicitly learning a metric to predict the similarity of a region proposal to the underlying query, no matter its class label is seen or unseen in training. The resulting model is therefore a two-stage detector that yields a strong baseline on both VOC and MS-COCO under one-shot setting of detecting objects from both seen and never-seen classes. Codes are available at https://github.com/timy90022/One-Shot-Object-Detection. | ['Tyng-Luh Liu', 'Hwann-Tzong Chen', 'Yi-Chen Lo', 'Ting-I Hsieh'] | 2019-11-28 | one-shot-object-detection-with-co-attention | http://papers.nips.cc/paper/8540-one-shot-object-detection-with-co-attention-and-co-excitation | http://papers.nips.cc/paper/8540-one-shot-object-detection-with-co-attention-and-co-excitation.pdf | neurips-2019-12 | ['one-shot-object-detection'] | ['computer-vision'] | [ 3.71867537e-01 1.00007109e-01 -1.30376041e-01 -2.84071714e-01
-1.23286319e+00 -3.63820642e-01 6.71747923e-01 7.63579085e-02
-4.13043171e-01 2.64856845e-01 -2.14689210e-01 2.15659529e-01
4.63051125e-02 -4.76348072e-01 -8.42543423e-01 -8.95356774e-01
1.92874715e-01 2.57143795e-01 5.88349879e-01 7.84624591e-02
1.51323467e-01 4.60287631e-01 -1.76755965e+00 5.06534159e-01
3.92158061e-01 1.10715461e+00 6.17965400e-01 8.30193698e-01
2.87223816e-01 7.63620436e-01 -4.13368195e-01 -3.57452154e-01
2.52634972e-01 -5.42015314e-01 -6.80744946e-01 4.79077220e-01
7.00580597e-01 -1.50854737e-01 -2.73249537e-01 1.19545698e+00
6.40614033e-01 3.63353163e-01 8.51354718e-01 -1.09999144e+00
-2.65126765e-01 9.88270268e-02 -8.10752451e-01 7.48942554e-01
2.75826938e-02 2.54909158e-01 1.28881788e+00 -1.62873900e+00
6.43421769e-01 1.01331174e+00 8.83254260e-02 5.61736226e-01
-1.15518653e+00 -4.83588964e-01 3.18439573e-01 4.88975376e-01
-1.55411291e+00 -7.22001195e-01 7.43989944e-01 -4.87982631e-01
6.53751373e-01 2.58337736e-01 4.17956710e-01 9.29612875e-01
-1.51255414e-01 1.11885750e+00 7.97318220e-01 -5.23720980e-01
2.91245282e-01 3.90160024e-01 2.27048934e-01 5.58502913e-01
-9.87822264e-02 3.06938231e-01 -5.26712596e-01 -9.97039583e-03
1.74612045e-01 2.60904998e-01 -2.55881995e-01 -5.30404091e-01
-8.94210160e-01 8.04762900e-01 6.54873431e-01 3.84992599e-01
-4.06927019e-01 9.32166204e-02 4.44394648e-02 4.70086634e-02
4.64260221e-01 3.29758167e-01 -2.97893286e-01 2.75637448e-01
-9.55758572e-01 1.96016192e-01 3.18031788e-01 8.16791415e-01
8.90822470e-01 -2.58463413e-01 -6.42385781e-01 8.66438270e-01
1.93516120e-01 3.38588715e-01 8.95056129e-02 -5.67858338e-01
2.84757495e-01 4.00714189e-01 1.90232798e-01 -7.21672773e-01
-1.56208947e-02 -9.60597157e-01 -2.92227924e-01 2.95594543e-01
2.04833016e-01 1.17881879e-01 -1.02663469e+00 1.76489949e+00
7.25382209e-01 2.85581261e-01 -4.96490523e-02 1.04376662e+00
8.33689928e-01 7.00934231e-01 -2.28738822e-02 -1.49504617e-01
1.58255017e+00 -1.33900833e+00 -3.42099607e-01 -3.99691880e-01
4.78945047e-01 -7.03894138e-01 8.59473526e-01 1.86327264e-01
-1.12686801e+00 -7.81720519e-01 -9.45151985e-01 -6.75062537e-02
-4.19768065e-01 4.34360385e-01 1.37552485e-01 2.34619573e-01
-6.82681739e-01 2.79184312e-01 -4.89714980e-01 -4.04263407e-01
7.11957216e-01 1.52855322e-01 -1.61584631e-01 -4.49943364e-01
-8.47262084e-01 7.78365195e-01 4.73970562e-01 5.10982722e-02
-1.49968839e+00 -5.08550048e-01 -6.80170178e-01 2.14876622e-01
8.96310627e-01 -4.30134892e-01 1.27920067e+00 -1.16483068e+00
-8.99369001e-01 1.12987554e+00 -4.00758654e-01 -2.50415683e-01
3.80194843e-01 -1.69095114e-01 -2.63951689e-01 2.29867786e-01
4.03496414e-01 8.15834701e-01 1.10485864e+00 -1.25045574e+00
-1.00899124e+00 -4.77620691e-01 -4.48196642e-02 3.51771384e-01
-6.64261058e-02 3.20793927e-01 -8.82021129e-01 -5.41902363e-01
-5.82830086e-02 -7.28174150e-01 -1.06129892e-01 1.74284190e-01
-4.77840722e-01 -4.04925734e-01 7.93579817e-01 -2.70529360e-01
8.99371266e-01 -2.38572836e+00 1.11043572e-01 -2.11153641e-01
1.81943819e-01 4.01244581e-01 -3.04320186e-01 3.66203070e-01
-5.09933159e-02 -3.14744323e-01 -1.64853930e-01 -5.03374755e-01
-2.92449206e-01 -7.52103105e-02 -4.60046202e-01 6.28395200e-01
7.14364231e-01 1.04935098e+00 -9.16576266e-01 -4.89686400e-01
2.01899558e-01 3.60254526e-01 -3.29508811e-01 4.09663230e-01
-1.76993802e-01 3.34598541e-01 -4.02124912e-01 7.53472924e-01
6.17002785e-01 -3.00816178e-01 -4.07011986e-01 -2.42660880e-01
-6.98851794e-02 -9.17565972e-02 -1.19546866e+00 1.54592764e+00
-1.94521293e-01 5.60285628e-01 1.08754680e-01 -9.20168757e-01
5.80413163e-01 1.35417953e-01 1.38602108e-01 -7.21816957e-01
2.28099629e-01 2.44164795e-01 4.10269164e-02 -5.13399661e-01
3.15931678e-01 -7.17987642e-02 6.07209392e-02 2.66207129e-01
4.92194116e-01 3.99337262e-01 1.55553624e-01 2.73690343e-01
1.04495513e+00 -8.38903114e-02 4.51853931e-01 -7.93933943e-02
5.15981853e-01 -2.21249625e-01 5.38637340e-01 1.00454724e+00
-3.77397329e-01 7.65295565e-01 3.10530275e-01 -2.79527586e-02
-8.51712108e-01 -9.26509798e-01 -1.77384809e-01 1.56460929e+00
3.49508643e-01 4.39213589e-02 -4.82416183e-01 -7.61087596e-01
-1.37276664e-01 7.54643500e-01 -8.88742864e-01 -3.34660619e-01
-1.42109647e-01 -4.69769955e-01 2.58660000e-02 4.22015250e-01
7.19502494e-02 -1.13701272e+00 -9.07282948e-01 6.61421865e-02
4.44122478e-02 -8.97828996e-01 -6.43830717e-01 5.99164724e-01
-4.88000542e-01 -1.14373875e+00 -8.67385209e-01 -9.06205356e-01
6.74324930e-01 6.62346959e-01 9.42813873e-01 -8.53957087e-02
-6.97574079e-01 5.26123166e-01 -3.25432301e-01 -4.03548777e-01
6.77044615e-02 -2.52577484e-01 -3.10987562e-01 5.63704491e-01
4.46385741e-01 -1.57620460e-01 -8.03745210e-01 4.02698725e-01
-7.38005400e-01 -1.04371712e-01 7.42900193e-01 1.02520144e+00
8.75798643e-01 -2.29445219e-01 5.00212669e-01 -6.56154335e-01
1.24548078e-02 -5.73723614e-01 -4.30394530e-01 4.44619089e-01
-2.07073659e-01 -2.42114469e-01 2.11711869e-01 -5.27018249e-01
-9.01370943e-01 3.88997406e-01 1.53156579e-01 -8.17079782e-01
-2.87827313e-01 -3.71724716e-03 -3.06558818e-01 6.28203452e-02
5.54748297e-01 4.00323093e-01 -4.23808396e-01 -4.64522034e-01
4.78516459e-01 5.33098161e-01 6.38063252e-01 -2.87746668e-01
6.57798588e-01 6.56708062e-01 -2.81760454e-01 -8.56655419e-01
-1.23292422e+00 -1.19204557e+00 -6.11064553e-01 -3.54290009e-01
8.50107312e-01 -1.10658813e+00 -2.47473821e-01 2.34302372e-01
-1.19165277e+00 -1.37279376e-01 -5.60815394e-01 2.65940517e-01
-4.28821117e-01 -1.27872378e-02 -1.45216808e-01 -1.15449727e+00
-3.17702085e-01 -1.06722355e+00 1.40352511e+00 3.68075907e-01
2.91375816e-01 -4.64562953e-01 -5.51655218e-02 5.94732724e-02
1.55798987e-01 -3.35718431e-02 4.19901937e-01 -9.85924721e-01
-7.48403490e-01 -5.24714112e-01 -3.88823837e-01 2.26018578e-01
-2.59713084e-01 -1.69901893e-01 -1.36793530e+00 -5.00683606e-01
1.83558792e-01 -6.06553137e-01 1.24475121e+00 3.98601651e-01
1.06849575e+00 -5.87077290e-02 -4.82805282e-01 4.78044599e-01
1.43895161e+00 1.50612727e-01 3.90434712e-01 -2.24611565e-01
5.10320067e-01 8.49638820e-01 9.81192768e-01 3.22137624e-01
-8.31819773e-02 1.01874506e+00 6.02108598e-01 -2.27875784e-01
-2.71838605e-01 -2.26478964e-01 1.91663429e-01 1.09190151e-01
4.15396392e-01 -4.66523141e-01 -6.90877140e-01 9.01281714e-01
-1.78165591e+00 -9.73911822e-01 -1.29337907e-01 2.24548769e+00
3.08413863e-01 5.08951247e-02 1.85400516e-01 -1.84805676e-01
9.58656788e-01 1.76286012e-01 -7.57561803e-01 1.71485871e-01
-9.37991664e-02 1.48914918e-01 2.30795875e-01 3.43238235e-01
-1.35459578e+00 8.63627017e-01 4.77840996e+00 1.05754399e+00
-1.01684523e+00 4.93722200e-01 7.94653952e-01 -3.71213436e-01
2.15746552e-01 -2.70156795e-03 -1.15764654e+00 4.05762970e-01
5.88383257e-01 -1.00313984e-01 5.20253293e-02 1.01905525e+00
4.88577187e-02 -3.19148600e-01 -1.27172327e+00 8.89777005e-01
4.15890574e-01 -1.01051331e+00 -1.22847021e-01 4.43988442e-02
7.15272069e-01 2.00021908e-01 3.44382346e-01 3.84871513e-01
-1.69916630e-01 -7.49254823e-01 9.42260385e-01 4.37921703e-01
7.48575985e-01 -5.23033082e-01 4.54117626e-01 5.38115442e-01
-1.28174543e+00 -3.13124955e-01 -4.66964662e-01 2.04062134e-01
7.62990192e-02 4.11532283e-01 -7.39788771e-01 3.40387851e-01
6.24761462e-01 6.96460247e-01 -6.58930957e-01 1.42273116e+00
-2.82369643e-01 4.75755125e-01 -1.15287431e-01 4.72767651e-02
3.17686051e-01 3.25591564e-01 8.26158643e-01 1.12185109e+00
1.80369854e-01 1.85921103e-01 3.23810220e-01 1.13883746e+00
2.19747573e-02 1.11750886e-01 -4.51790571e-01 1.49094477e-01
3.52844417e-01 1.32275951e+00 -8.59261870e-01 -4.29154128e-01
-4.12205786e-01 1.19381928e+00 4.81265068e-01 5.04109979e-01
-8.55920911e-01 -4.52799946e-01 3.52299660e-01 7.29399994e-02
9.60063934e-01 4.91398722e-01 2.87877858e-01 -1.03962684e+00
1.35691658e-01 -6.15995109e-01 5.12397468e-01 -8.10034633e-01
-1.18928385e+00 5.69926143e-01 -1.53474271e-01 -1.26584947e+00
5.54716922e-02 -6.11051023e-01 -9.33268905e-01 8.49744201e-01
-1.49417770e+00 -1.19882882e+00 -2.61624753e-01 4.90491658e-01
8.91725063e-01 4.53392835e-03 4.82417941e-01 4.59942997e-01
-6.34148538e-01 6.28989756e-01 8.81253630e-02 2.17973739e-02
5.89703918e-01 -1.08964872e+00 4.90352958e-02 1.14399731e+00
5.38844049e-01 2.93699384e-01 5.36355913e-01 -4.35640186e-01
-9.88221407e-01 -1.32623851e+00 7.25925386e-01 -4.45709616e-01
4.23936874e-01 -5.97247958e-01 -1.07125473e+00 4.04349327e-01
-7.42110237e-02 6.70327127e-01 3.00257474e-01 -2.47612908e-01
-3.74796957e-01 -4.35176343e-02 -8.34369123e-01 2.26090446e-01
7.56331384e-01 -6.63087785e-01 -5.17293096e-01 3.89591008e-01
6.53414905e-01 -7.82994330e-02 -2.08843350e-01 3.28415900e-01
2.89293468e-01 -8.63349497e-01 1.00541782e+00 -7.87457585e-01
3.52903515e-01 -4.44388330e-01 -3.01053762e-01 -8.30967069e-01
-3.30296844e-01 -3.82744581e-01 -3.69587034e-01 1.09496081e+00
4.24019784e-01 -2.03377113e-01 6.38667703e-01 6.33154139e-02
-2.42533788e-01 -1.02726376e+00 -9.72188652e-01 -6.42261028e-01
-2.91420937e-01 -3.77560914e-01 2.08045188e-02 5.12444556e-01
-3.95510554e-01 5.71358562e-01 -5.11657357e-01 3.08654487e-01
7.60782421e-01 1.70877159e-01 7.51466274e-01 -9.43636954e-01
-6.73188210e-01 -4.08359557e-01 -4.06596571e-01 -1.17017770e+00
-2.19778359e-01 -8.44049931e-01 5.88937581e-01 -1.16504240e+00
7.09155917e-01 -4.82565686e-02 -6.60754502e-01 3.01814646e-01
-5.52755237e-01 3.42381567e-01 2.10764721e-01 2.98859000e-01
-1.07239258e+00 6.15411341e-01 9.53335702e-01 -1.46941409e-01
-8.30315351e-02 2.89299160e-01 -5.03601730e-01 4.12181795e-01
3.14576864e-01 -7.90023744e-01 -1.88070610e-01 -1.18639275e-01
-1.71449482e-01 1.20374523e-01 8.68339956e-01 -9.47358489e-01
4.27094966e-01 1.23408057e-01 2.92813987e-01 -9.65176880e-01
5.44778883e-01 -6.79125905e-01 -1.98357180e-01 4.21591699e-01
-5.27644098e-01 -3.96430165e-01 -5.39345015e-03 1.11572504e+00
-2.09286153e-01 -4.46586758e-01 1.18968928e+00 -4.38987017e-02
-1.01452172e+00 4.64793473e-01 5.29625006e-02 2.43149459e-01
1.44596541e+00 -2.18457341e-01 -2.30792060e-01 -2.90530026e-02
-9.26584363e-01 3.45591336e-01 3.62894416e-01 2.84288555e-01
6.56439781e-01 -1.14722550e+00 -7.45520592e-01 1.77999496e-01
7.42885768e-01 -1.16780862e-01 4.89475906e-01 1.04832411e+00
1.23186253e-01 3.77212226e-01 1.27074197e-01 -7.95936525e-01
-1.40543842e+00 8.36470366e-01 4.83928770e-01 -9.49549153e-02
-5.28518319e-01 1.53543878e+00 8.36652458e-01 -3.87250446e-02
5.57041466e-01 1.88332751e-01 -1.18600242e-01 2.13139117e-01
7.99274564e-01 2.60236174e-01 -7.73260519e-02 -8.34823966e-01
-3.98118764e-01 4.34304774e-01 -3.47779095e-01 -1.04955867e-01
1.15745008e+00 -1.36801645e-01 2.54051805e-01 6.87457740e-01
1.41587353e+00 -2.31529713e-01 -1.44007587e+00 -5.84762335e-01
-1.30380155e-03 -5.65749824e-01 1.61673650e-01 -6.94276154e-01
-1.03998387e+00 1.14730144e+00 1.01639783e+00 -1.78865626e-01
9.03486371e-01 6.48406446e-01 3.35676521e-01 2.13589981e-01
-1.66211985e-02 -8.47704649e-01 6.44660532e-01 1.86302066e-01
8.06392252e-01 -1.47010100e+00 -3.18807900e-01 -2.50443697e-01
-7.34790087e-01 7.69038916e-01 7.10029244e-01 -1.59373894e-01
5.07353127e-01 -2.38642022e-01 -2.66716927e-01 -5.68410456e-01
-9.32744920e-01 -7.60674298e-01 6.79162979e-01 4.37811583e-01
6.97891936e-02 -1.23189263e-01 1.05854340e-01 3.81879091e-01
5.26774764e-01 -3.16871494e-01 1.59795120e-01 7.99758494e-01
-8.61109018e-01 -5.00239849e-01 -3.53118032e-01 4.97909844e-01
-4.28905547e-01 -4.86460850e-02 -5.15804231e-01 4.91474599e-01
2.75888115e-01 8.98381233e-01 4.67066206e-02 -2.59569526e-01
2.78337508e-01 5.11319675e-02 2.99001694e-01 -1.08393490e+00
-3.43300670e-01 3.90228838e-01 -3.63799810e-01 -4.84292597e-01
-3.03334564e-01 -7.44561553e-01 -7.79779196e-01 5.37633538e-01
-8.00838768e-01 3.99617255e-02 3.69681209e-01 7.95879185e-01
2.99228370e-01 6.29248738e-01 7.06022739e-01 -1.00659895e+00
-6.74331963e-01 -8.80674660e-01 -6.38395309e-01 3.34832698e-01
5.20185471e-01 -9.36107039e-01 -4.83757287e-01 -1.92607373e-01] | [9.474352836608887, 1.4505455493927002] |
85896b7e-3ee6-4e3a-92cf-b3448ad9738d | interpretability-in-linear-brain-decoding | 1606.05672 | null | http://arxiv.org/abs/1606.05672v1 | http://arxiv.org/pdf/1606.05672v1.pdf | Interpretability in Linear Brain Decoding | Improving the interpretability of brain decoding approaches is of primary
interest in many neuroimaging studies. Despite extensive studies of this type,
at present, there is no formal definition for interpretability of brain
decoding models. As a consequence, there is no quantitative measure for
evaluating the interpretability of different brain decoding methods. In this
paper, we present a simple definition for interpretability of linear brain
decoding models. Then, we propose to combine the interpretability and the
performance of the brain decoding into a new multi-objective criterion for
model selection. Our preliminary results on the toy data show that optimizing
the hyper-parameters of the regularized linear classifier based on the proposed
criterion results in more informative linear models. The presented definition
provides the theoretical background for quantitative evaluation of
interpretability in linear brain decoding. | ['Seyed Mostafa Kia', 'Andrea Passerini'] | 2016-06-17 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 3.37330073e-01 2.76091605e-01 1.21225663e-01 -7.83246696e-01
-5.24004698e-01 -4.66320693e-01 5.09104371e-01 1.34614602e-01
-5.64971268e-01 6.68397665e-01 9.43251476e-02 -4.32908267e-01
-5.44657171e-01 -1.95534062e-02 -4.07587171e-01 -7.15812147e-01
7.66029432e-02 6.31108224e-01 -2.48659924e-01 2.14871258e-01
4.37362760e-01 6.08084798e-01 -1.34634554e+00 1.85844719e-01
1.04690361e+00 1.07120728e+00 5.10507882e-01 5.31077325e-01
1.20334603e-01 1.74123719e-01 -5.58526933e-01 -4.47083682e-01
9.22645908e-03 -8.09768081e-01 -7.42879033e-01 3.44744883e-02
2.10384145e-01 2.45998185e-02 -1.13862909e-01 1.17301595e+00
4.16827261e-01 -2.16594692e-02 1.10934544e+00 -1.10907817e+00
-6.01738870e-01 8.74761701e-01 -2.63787899e-02 6.40007794e-01
1.88582450e-01 -1.97950806e-02 1.20784807e+00 -8.34911644e-01
1.29397109e-01 1.36418974e+00 3.17213446e-01 4.26210642e-01
-1.31978476e+00 -2.76224613e-01 1.61890462e-01 5.13562322e-01
-1.24066496e+00 -6.55483961e-01 4.11541879e-01 -5.36931038e-01
7.31908619e-01 6.29601896e-01 9.04207468e-01 8.48177552e-01
5.83541930e-01 5.20811915e-01 1.50138772e+00 -5.42385340e-01
2.16630384e-01 1.63652092e-01 9.56188023e-01 8.53660643e-01
7.01977074e-01 1.57518417e-01 -4.80405807e-01 -5.57797123e-03
5.85369170e-01 -5.10845900e-01 -2.43238851e-01 -8.99846256e-02
-1.04151881e+00 8.54806423e-01 2.78846234e-01 5.06487727e-01
-1.85502172e-01 3.47984195e-01 3.09182853e-01 1.41891703e-01
6.30957365e-01 7.02835381e-01 -5.32282554e-02 2.32710559e-02
-9.07225132e-01 -3.78251337e-02 6.43741310e-01 6.00232661e-01
3.65202695e-01 -4.42902036e-02 -2.70205319e-01 7.48913586e-01
6.64676785e-01 5.82513452e-01 4.80371803e-01 -7.06518829e-01
2.61770904e-01 3.12047958e-01 -4.20237064e-01 -9.87239361e-01
-8.54656637e-01 -5.57767510e-01 -7.07003057e-01 1.56438481e-02
4.88610297e-01 -7.33471438e-02 -6.80488646e-01 1.63846743e+00
-3.22998285e-01 -3.15999329e-01 -7.18572289e-02 1.10642743e+00
8.28830183e-01 4.63365108e-01 7.52620846e-02 -3.40246081e-01
1.36234164e+00 -8.50321949e-01 -9.93284166e-01 -2.85433531e-01
4.41597521e-01 -4.87375140e-01 1.10852575e+00 4.05326128e-01
-1.15265346e+00 -3.08645338e-01 -1.00084293e+00 -6.57158047e-02
-2.37940382e-02 3.43312681e-01 7.33154774e-01 9.17565405e-01
-1.11241579e+00 2.01964095e-01 -7.53373563e-01 -2.60997981e-01
2.03914687e-01 6.32796943e-01 -2.72075444e-01 1.84084892e-01
-1.00046051e+00 1.45773518e+00 7.01219499e-01 4.21346784e-01
-7.97787428e-01 -1.07003145e-01 -6.31680429e-01 1.86463654e-01
7.32496455e-02 -6.45609498e-01 1.06740522e+00 -1.28979814e+00
-1.42039633e+00 1.02759933e+00 -4.66057360e-01 -3.86109114e-01
4.13455933e-01 1.42857552e-01 -2.23232687e-01 5.66261299e-02
-2.77919173e-01 5.42929769e-01 8.13659847e-01 -1.29223287e+00
-1.21264003e-01 -4.21043456e-01 3.18053439e-02 3.08902144e-01
-1.04269564e-01 1.80469930e-01 -3.36439520e-01 -4.95145321e-01
3.22638333e-01 -9.44331586e-01 -6.59044534e-02 -1.39343262e-01
-4.18169737e-01 -1.96817562e-01 -4.72094771e-03 -8.19956899e-01
1.34534681e+00 -2.09937501e+00 6.04742706e-01 5.42853355e-01
6.32626653e-01 -4.30709198e-02 -1.47686332e-01 -2.65049905e-01
-3.18544090e-01 3.35744202e-01 -4.68033284e-01 -3.15950006e-01
8.58696401e-02 2.73436338e-01 8.55581686e-02 6.64481640e-01
1.32674664e-01 8.85960877e-01 -7.67402291e-01 -3.56117874e-01
1.69652849e-01 3.55504066e-01 -6.41087592e-01 7.92846978e-02
2.24883378e-01 4.46311861e-01 -5.11906624e-01 2.23480955e-01
4.12863970e-01 -1.63838655e-01 1.30311415e-01 -8.65694284e-02
3.97882313e-02 2.76903570e-01 -7.32531548e-01 1.17381465e+00
-2.67607361e-01 1.13478661e+00 -2.80805081e-01 -1.19955480e+00
6.92105114e-01 1.57483846e-01 2.24052742e-01 -7.05418706e-01
4.04410243e-01 3.49072963e-01 6.49122953e-01 -5.58949172e-01
1.96949944e-01 -2.68237293e-01 -2.43339874e-03 5.14411032e-01
2.37985522e-01 2.87886633e-05 2.06333935e-01 -3.05239081e-01
7.79162407e-01 -3.24993879e-01 6.22969389e-01 -6.86328471e-01
6.95886970e-01 -3.94556582e-01 1.24077864e-01 8.10544848e-01
-1.53163537e-01 6.47705972e-01 7.92214155e-01 -3.35811973e-01
-9.17743683e-01 -1.05173433e+00 -5.41402876e-01 9.18761551e-01
-1.50930015e-02 -2.98058957e-01 -1.25246906e+00 -3.81108463e-01
-2.74556398e-01 9.59157109e-01 -6.62450135e-01 -2.55344331e-01
-4.68976647e-01 -1.17992437e+00 6.50744319e-01 1.29901946e-01
2.69987762e-01 -8.00994277e-01 -6.28093958e-01 -1.45768419e-01
-3.70906502e-01 -9.21578348e-01 -2.61400938e-01 4.93812233e-01
-1.04501963e+00 -9.54199612e-01 -3.88674527e-01 -6.09599233e-01
9.36047494e-01 5.14358655e-02 1.02923226e+00 3.16897780e-01
4.34651040e-02 3.46530706e-01 -1.99828714e-01 -4.45790470e-01
-6.44491613e-01 1.25275314e-01 1.12985477e-01 -1.07026264e-01
1.50401443e-01 -3.35778594e-02 -1.68142989e-01 4.06035274e-01
-9.00038779e-01 2.72909105e-01 6.72969937e-01 8.16225648e-01
4.60838109e-01 -1.71740696e-01 4.76659119e-01 -6.38561666e-01
1.09886181e+00 -3.82897377e-01 -5.62481821e-01 4.87888217e-01
-8.05746257e-01 6.79928184e-01 4.22162205e-01 -4.34017599e-01
-9.24105883e-01 -1.08160526e-01 -2.69729704e-01 1.33148417e-01
-1.19292967e-01 8.20007145e-01 -2.70750999e-01 -2.15049252e-01
6.56372726e-01 1.73832819e-01 -7.78040886e-02 -4.36911523e-01
2.35383362e-01 5.63551128e-01 1.39531612e-01 -4.53245431e-01
2.97597498e-01 -3.83036509e-02 1.19401878e-02 -1.01913953e+00
-6.02244794e-01 -1.02114782e-01 -7.99397528e-01 -5.19319952e-01
9.94955480e-01 -4.19338405e-01 -5.98112762e-01 1.76223516e-01
-1.41391122e+00 5.37787639e-02 9.32791755e-02 7.25453138e-01
-8.11442316e-01 4.89726126e-01 -4.56764363e-02 -9.70522165e-01
-1.92834929e-01 -1.80960429e+00 9.15322840e-01 -1.43355057e-01
-4.50354069e-01 -1.23674691e+00 -3.07501972e-01 4.51615453e-01
1.35028616e-01 -9.95372385e-02 1.17058301e+00 -8.16515088e-01
-4.09346014e-01 -5.81420958e-02 -2.51242220e-01 3.29087675e-01
-1.99477002e-01 -3.08880121e-01 -1.04410660e+00 2.62008533e-02
3.37943107e-01 1.32338092e-01 9.80708599e-01 6.47348762e-01
1.31704688e+00 -3.18103820e-01 -1.33276612e-01 5.56707501e-01
1.10757101e+00 3.23628068e-01 6.25219524e-01 7.33836144e-02
5.49759090e-01 5.60724556e-01 2.43250757e-01 1.66454524e-01
3.60132121e-02 7.62400031e-01 2.49500543e-01 9.34842676e-02
1.12370923e-01 3.23121876e-01 5.70565879e-01 1.00939441e+00
-4.03212517e-01 -5.19738972e-01 -1.00771463e+00 3.23178545e-02
-1.74650812e+00 -8.18675339e-01 -3.89224023e-01 2.22869396e+00
5.42001724e-01 1.39521942e-01 -4.11905237e-02 5.90691306e-02
8.04601014e-01 -1.94215342e-01 -4.31575179e-01 -6.36690021e-01
-3.01846653e-01 -4.42662686e-02 4.87531036e-01 7.84953833e-01
-7.30728209e-01 6.57281458e-01 7.93102074e+00 7.77513683e-01
-9.82335985e-01 2.41970092e-01 8.10923219e-01 -5.24904281e-02
-4.31491643e-01 -1.52168155e-01 -7.31521428e-01 4.71863389e-01
1.13268733e+00 -4.51045245e-01 7.64001250e-01 6.01933897e-01
5.21753788e-01 -2.51621813e-01 -1.45158136e+00 1.34142256e+00
4.40820009e-01 -9.84005511e-01 8.10993314e-02 2.33616531e-01
2.28197917e-01 -1.51809916e-01 1.35229543e-01 -1.76353350e-01
-4.67735827e-01 -1.29211032e+00 1.04346502e+00 8.50417793e-01
5.28094709e-01 -4.81247723e-01 6.60650074e-01 3.01920652e-01
-6.15661621e-01 -5.34565262e-02 -4.76637572e-01 5.43797724e-02
1.01268575e-01 6.08783126e-01 -6.98352575e-01 7.66279325e-02
1.65770501e-01 4.17212993e-01 -8.44140112e-01 1.21986711e+00
-2.00324848e-01 6.56154335e-01 -3.16796452e-01 -4.21545327e-01
1.89254463e-01 -3.59678537e-01 7.53638029e-01 1.35756898e+00
3.21284115e-01 1.50506333e-01 -3.16660047e-01 1.03438437e+00
1.57933369e-01 3.06457847e-01 -5.23763776e-01 -3.16806674e-01
-1.18776016e-01 1.02067971e+00 -1.04024923e+00 -1.59169197e-01
-2.26724952e-01 7.97386587e-01 4.24862713e-01 1.32878155e-01
-9.87651169e-01 3.36914688e-01 5.51366627e-01 -5.61640412e-02
-3.60559493e-01 -4.44579691e-01 -9.31422532e-01 -1.27962804e+00
-2.62996648e-02 -8.26172948e-01 1.75782993e-01 -6.75107121e-01
-1.03643298e+00 9.13585007e-01 5.74766695e-01 -7.21355498e-01
-2.75051087e-01 -1.03258336e+00 -2.37984583e-01 9.89211380e-01
-1.01247561e+00 -5.93012691e-01 -1.71517417e-01 7.38848746e-02
4.98263419e-01 -3.14333797e-01 8.43493819e-01 -3.54454443e-02
-5.36361694e-01 6.42463803e-01 2.14799523e-01 -3.05398166e-01
1.90365538e-01 -1.16677535e+00 1.53803274e-01 7.58006871e-01
1.52853101e-01 6.93680584e-01 9.90768850e-01 -3.57804269e-01
-1.22692513e+00 -4.78864580e-01 9.41050529e-01 -4.26798135e-01
5.03476501e-01 -3.02372158e-01 -8.67875814e-01 5.05799592e-01
2.37974063e-01 -6.90580606e-01 9.81443763e-01 1.25001222e-01
8.71483237e-02 8.37828498e-03 -9.97871459e-01 6.38010800e-01
9.69965637e-01 -3.59937966e-01 -8.35033059e-01 4.90823209e-01
3.04445267e-01 -5.16641140e-02 -6.87608123e-01 6.91779703e-02
6.72814548e-01 -8.87619317e-01 8.56882393e-01 -6.21629238e-01
2.76274592e-01 -1.12242363e-01 -1.35449916e-01 -1.55268466e+00
-5.01833081e-01 -5.88769838e-02 1.00381672e-01 5.40944695e-01
6.42311096e-01 -7.40196884e-01 2.31201634e-01 8.55624616e-01
-1.94094568e-01 -6.87853992e-01 -1.10024941e+00 -6.50754094e-01
3.42431590e-02 -7.30940878e-01 2.54749089e-01 4.55773264e-01
1.54091701e-01 6.48011506e-01 -2.02824920e-01 3.41333784e-02
7.22171843e-01 -2.97193348e-01 1.36256412e-01 -1.33338439e+00
-3.34694684e-01 -9.77078855e-01 -7.58566439e-01 -1.02657998e+00
5.28589725e-01 -1.15384030e+00 8.12211260e-02 -1.51544988e+00
4.70356196e-01 -2.21156120e-01 -3.36722463e-01 3.02698642e-01
-2.43746564e-01 1.09572425e-01 4.22644645e-01 1.54068381e-01
-4.83988613e-01 3.31347078e-01 1.14701152e+00 -1.79185420e-01
-2.36906875e-02 1.22980718e-02 -7.21257806e-01 8.20778012e-01
1.06299508e+00 -5.40562809e-01 -5.41449308e-01 -6.63486004e-01
3.82965684e-01 5.05042151e-02 3.95203233e-01 -8.39434505e-01
-9.04080570e-02 -1.34317592e-01 3.67186517e-01 -3.12983364e-01
2.83792764e-01 -7.24822104e-01 2.06750229e-01 3.86374772e-01
-6.27874851e-01 2.96751410e-01 1.38292804e-01 3.49795491e-01
-2.47994252e-02 -8.18401456e-01 9.40357566e-01 3.53225246e-02
-6.46086633e-01 -1.31573036e-01 -7.46675014e-01 4.57429811e-02
6.73425913e-01 -3.50947887e-01 8.13794136e-02 -3.01653862e-01
-1.04239476e+00 -3.02550681e-02 3.24266255e-02 3.35701436e-01
7.91805685e-01 -1.04563248e+00 -7.15322971e-01 2.57614195e-01
6.17349818e-02 -6.71279252e-01 -2.08397180e-01 1.01246321e+00
-6.84286714e-01 7.57701993e-01 -4.05104458e-01 -6.58978283e-01
-1.41863739e+00 2.13100120e-01 7.11602032e-01 -1.38551146e-01
-1.74532145e-01 4.24166918e-01 5.12700379e-01 -7.09639639e-02
1.34652674e-01 -6.37155890e-01 -3.66330862e-01 9.99926180e-02
5.35536766e-01 3.91395479e-01 2.09296405e-01 -9.66148436e-01
-4.00049776e-01 2.73917705e-01 3.02299947e-01 -3.81469965e-01
1.25965977e+00 -2.38657326e-01 -3.84356171e-01 6.45679832e-01
1.06170905e+00 -4.61907893e-01 -7.08745003e-01 3.32768530e-01
2.31879264e-01 -3.36845279e-01 2.70781130e-01 -8.99738789e-01
-9.27803755e-01 9.34734583e-01 5.98315597e-01 2.91982919e-01
1.25480044e+00 -7.89017882e-03 -9.68625769e-02 7.25919604e-01
2.67902076e-01 -9.56941664e-01 -3.55632454e-01 7.43028164e-01
1.28741658e+00 -1.09867287e+00 -3.39356326e-02 -4.67163086e-01
-6.70940399e-01 1.38568735e+00 2.63999343e-01 1.11009777e-01
5.61538935e-01 5.07264286e-02 -2.18451887e-01 -1.51092187e-01
-4.01693076e-01 -1.50713727e-01 8.25192869e-01 4.86837000e-01
6.71897352e-01 3.87804687e-01 -1.02902567e+00 7.25891888e-01
-3.57849419e-01 -2.84388006e-01 4.06807721e-01 2.39131734e-01
-5.58807194e-01 -9.09920096e-01 -5.46443403e-01 6.94150627e-01
-2.88696349e-01 -3.77731919e-01 -5.67567885e-01 4.53331769e-01
-8.07348341e-02 9.49584842e-01 -5.66459931e-02 -2.19904050e-01
7.74639547e-02 2.90344536e-01 7.97417879e-01 -4.15295243e-01
-4.45873082e-01 -1.89399421e-02 1.51974574e-01 -2.43539378e-01
-3.37971538e-01 -7.36773908e-01 -1.17778707e+00 -2.35607252e-01
-5.52447617e-01 2.01771572e-01 8.66705418e-01 1.38067770e+00
9.56671685e-03 4.66376632e-01 2.02680767e-01 -8.96750093e-01
-4.32653815e-01 -8.77595365e-01 -4.76327062e-01 1.33376509e-01
1.72044799e-01 -8.79471481e-01 -4.38315868e-01 9.59900320e-02] | [12.607677459716797, 3.4045729637145996] |
246f4662-fc5d-4912-ad07-496af18a928a | density-uncertainty-layers-for-reliable | 2306.12497 | null | https://arxiv.org/abs/2306.12497v1 | https://arxiv.org/pdf/2306.12497v1.pdf | Density Uncertainty Layers for Reliable Uncertainty Estimation | Assessing the predictive uncertainty of deep neural networks is crucial for safety-related applications of deep learning. Although Bayesian deep learning offers a principled framework for estimating model uncertainty, the approaches that are commonly used to approximate the posterior often fail to deliver reliable estimates of predictive uncertainty. In this paper we propose a novel criterion for predictive uncertainty, that a model's predictive variance should be grounded in the empirical density of the input. It should produce higher uncertainty for inputs that are improbable in the training data and lower uncertainty for those inputs that are more probable. To operationalize this criterion, we develop the density uncertainty layer, an architectural element for a stochastic neural network that guarantees that the density uncertain criterion is satisfied. We study neural networks with density uncertainty layers on the CIFAR-10 and CIFAR-100 uncertainty benchmarks. Compared to existing approaches, we find that density uncertainty layers provide reliable uncertainty estimates and robust out-of-distribution detection performance. | ['David M. Blei', 'Yookoon Park'] | 2023-06-21 | null | null | null | null | ['out-of-distribution-detection'] | ['computer-vision'] | [-3.09906781e-01 2.72993743e-01 -3.13690640e-02 -9.31628644e-01
-1.01324964e+00 -4.06750172e-01 6.34212554e-01 9.92816593e-03
-3.92646402e-01 1.01462841e+00 -1.12086898e-02 -5.77041388e-01
-5.04474461e-01 -8.55647385e-01 -1.12890840e+00 -6.41492128e-01
-2.09157720e-01 4.70873713e-01 2.39961028e-01 5.61010897e-01
-1.25065178e-01 5.31605780e-01 -1.35043621e+00 3.03658217e-01
6.75260842e-01 1.71912861e+00 -1.85423762e-01 4.43476379e-01
-2.63307746e-02 9.61340249e-01 -9.00796533e-01 -4.11109924e-01
-2.22002547e-02 -1.69232816e-01 -5.06387293e-01 -9.01130259e-01
3.92711282e-01 -7.47148633e-01 -2.99849808e-01 1.38022316e+00
2.67766684e-01 1.56441271e-01 1.51306927e+00 -1.35480154e+00
-4.19378996e-01 1.32738388e+00 -3.37776333e-01 4.64629799e-01
-5.21187246e-01 -1.30634353e-01 9.24778044e-01 -5.56075394e-01
-1.46347091e-01 1.25848389e+00 7.92291999e-01 7.29768991e-01
-1.22108781e+00 -9.52984154e-01 3.10670316e-01 -1.76658124e-01
-1.51427329e+00 -5.13704360e-01 5.05382419e-01 -5.57473660e-01
8.19531560e-01 -1.31831914e-01 7.66454712e-02 1.23489714e+00
8.88394594e-01 5.35008788e-01 7.50617087e-01 -7.36558363e-02
9.99478936e-01 2.18833447e-01 2.07911521e-01 7.13951215e-02
5.60075641e-01 7.29160428e-01 -4.51536328e-01 -6.12314008e-02
5.48729300e-01 -2.29528964e-01 -4.27008927e-04 4.87375557e-02
-3.82380307e-01 8.53453159e-01 5.39313495e-01 1.62821375e-02
-2.46428981e-01 1.18080103e+00 3.43747348e-01 -6.93600923e-02
5.28461397e-01 1.39932275e-01 -4.38555270e-01 -1.20566256e-01
-1.08324504e+00 3.14437032e-01 7.47342646e-01 9.63082731e-01
3.76670450e-01 4.55708325e-01 -3.39297920e-01 5.57270169e-01
9.09894109e-01 5.11270761e-01 -1.41343653e-01 -1.05980074e+00
2.71277547e-01 -2.60976076e-01 7.22150505e-02 -5.36473632e-01
-2.25484341e-01 -6.95537269e-01 -7.74833262e-01 5.51478267e-01
4.32317287e-01 -5.02702415e-01 -1.02153277e+00 1.91156363e+00
-2.86329359e-01 3.23198944e-01 1.78648517e-01 5.88513851e-01
6.98352873e-01 6.85275435e-01 4.45453167e-01 8.83987397e-02
9.52733576e-01 -1.19747512e-01 -4.65712368e-01 -2.72730172e-01
2.40936965e-01 -1.46610558e-01 4.86409158e-01 3.93026054e-01
-8.09550047e-01 -3.16892952e-01 -1.39500463e+00 3.41876149e-01
-2.27469970e-02 -3.34633559e-01 6.09258235e-01 1.16815042e+00
-8.02341521e-01 9.52094316e-01 -8.79440784e-01 4.51408178e-01
7.82322586e-01 1.17299244e-01 6.17952645e-02 2.17942774e-01
-1.38128948e+00 1.02863657e+00 8.15954089e-01 8.95655602e-02
-1.54876387e+00 -1.03972197e+00 -7.23999083e-01 4.85764146e-01
-1.40128091e-01 -3.90256226e-01 1.65539324e+00 -7.41336286e-01
-1.19565189e+00 8.90192538e-02 4.39491183e-01 -1.04302073e+00
4.14131880e-01 -2.73167610e-01 -2.15088695e-01 -1.21988215e-01
-4.99964833e-01 7.37515390e-01 9.21464801e-01 -1.21199775e+00
-7.48808146e-01 -1.77494049e-01 -1.44594491e-01 -2.50969380e-01
1.91030409e-02 -1.46812454e-01 -6.30228221e-02 -4.28378165e-01
4.22027186e-02 -6.26848698e-01 -1.68511763e-01 -7.63829798e-02
-4.47652757e-01 -3.15512538e-01 3.00868541e-01 -2.22728565e-01
9.51990306e-01 -1.97932589e+00 -5.60096204e-01 5.61127901e-01
2.22345456e-01 -1.39174372e-01 3.30818474e-01 -1.75789475e-01
2.47240495e-02 3.47061843e-01 -1.28987119e-01 -1.96972579e-01
3.84591073e-01 1.74840853e-01 -8.41855586e-01 5.39225996e-01
5.48714101e-01 5.60343742e-01 -6.34857357e-01 -2.68198282e-01
1.03579096e-01 6.38774991e-01 -5.73450685e-01 -2.18632650e-02
-5.67407787e-01 -1.44856378e-01 -3.97473872e-01 2.21670255e-01
7.90410876e-01 3.10919266e-02 -2.38382757e-01 -1.80029735e-01
1.96999475e-01 3.57931286e-01 -1.05630922e+00 1.13445592e+00
-4.40502197e-01 7.65374303e-01 -3.97545904e-01 -5.89149237e-01
8.71840954e-01 2.66520351e-01 1.62195042e-01 -7.99514651e-02
6.58458471e-01 1.62309647e-01 1.50430381e-01 2.70080566e-01
2.77689517e-01 -4.13379818e-01 -1.37846708e-01 3.91376048e-01
5.79599798e-01 -3.11927676e-01 -2.84320593e-01 1.10599034e-01
9.71437335e-01 1.19469419e-01 -7.86615908e-02 -6.58155262e-01
-2.38705441e-01 -4.54224676e-01 6.18155360e-01 1.12214649e+00
-3.98310959e-01 5.51322997e-01 9.63504016e-01 -1.06573023e-01
-9.71146584e-01 -1.78083789e+00 -7.51543701e-01 6.73083484e-01
-3.55092019e-01 1.65117700e-02 -8.61318350e-01 -5.47647417e-01
1.44741580e-01 1.35769045e+00 -6.50852084e-01 -5.27300596e-01
4.61058080e-01 -8.11524749e-01 8.07376087e-01 9.14067090e-01
2.86737353e-01 -5.88872910e-01 -6.10361099e-01 1.42174244e-01
4.39604402e-01 -7.48717189e-01 2.21793819e-02 8.69626105e-01
-7.85525739e-01 -6.44178927e-01 -5.33964634e-01 -2.11202074e-02
3.89640957e-01 -6.60007417e-01 1.25228584e+00 -5.99632621e-01
1.13339223e-01 1.99247301e-01 6.01330400e-02 -1.09729230e+00
-4.70267653e-01 -3.58016580e-01 3.88083130e-01 -6.13554001e-01
4.93642628e-01 -6.25960827e-01 -3.17789733e-01 2.10084897e-02
-8.09171677e-01 -3.98402810e-01 5.13404369e-01 7.52898395e-01
5.94005048e-01 6.85073137e-01 1.01308262e+00 -6.54610634e-01
5.95633924e-01 -6.70671880e-01 -8.70811403e-01 2.28440836e-01
-8.51962686e-01 4.81356978e-01 3.25604916e-01 -3.75881523e-01
-1.32538307e+00 -9.52689722e-02 -3.99760127e-01 -6.66717827e-01
-1.64798379e-01 6.63374007e-01 -1.55432165e-01 2.84604877e-01
1.09897172e+00 -3.10666293e-01 -3.95141512e-01 -2.09471807e-01
1.84116945e-01 5.34008741e-01 5.86039364e-01 -1.07678187e+00
5.06399393e-01 1.69846669e-01 1.11598253e-01 -2.96549976e-01
-1.02944851e+00 1.21528737e-01 -2.17455849e-01 -3.64168495e-01
7.26783216e-01 -9.76447344e-01 -7.63112426e-01 3.38487774e-01
-1.15278435e+00 -1.49461254e-01 -3.32148820e-01 8.50165665e-01
-6.99141026e-01 -2.47287050e-01 -3.99363399e-01 -1.44784546e+00
-3.14629734e-01 -1.13873720e+00 5.59415102e-01 2.86826253e-01
-2.37693727e-01 -9.39915359e-01 -1.80161342e-01 -4.76157337e-01
4.78024572e-01 1.83068022e-01 9.61387694e-01 -9.14366066e-01
-3.24814916e-01 -2.67394394e-01 -4.69583303e-01 7.84036756e-01
-3.39072526e-01 4.47042733e-01 -1.63452065e+00 2.66715258e-01
4.99012955e-02 -3.68202776e-01 1.24857104e+00 1.07604659e+00
1.37136745e+00 -2.49714941e-01 -1.30699486e-01 4.00431693e-01
1.45356894e+00 3.49412620e-01 7.31283963e-01 -1.27755418e-01
1.43971577e-01 6.20956063e-01 1.76382944e-01 5.29627681e-01
-8.53159428e-02 9.40855313e-03 6.59493804e-01 6.64043009e-01
2.34839633e-01 -3.81816506e-01 3.58703047e-01 8.49997774e-02
3.08645159e-01 -3.24997485e-01 -1.12447262e+00 4.29761857e-01
-1.57891059e+00 -1.09193516e+00 2.24595666e-01 2.25926185e+00
1.03789437e+00 8.46168041e-01 -2.27359936e-01 4.60138805e-02
4.71002191e-01 -1.68634653e-01 -7.77783334e-01 -5.37226677e-01
2.48159692e-01 4.81838826e-03 5.76284885e-01 4.58996028e-01
-1.08458531e+00 4.96248871e-01 7.36048412e+00 1.17594445e+00
-7.02865899e-01 -6.98115230e-02 1.29017389e+00 -3.29707772e-01
-7.72621810e-01 -3.77994746e-01 -1.15825856e+00 6.06526732e-01
1.64546299e+00 -2.15271562e-01 1.05527956e-02 1.31609607e+00
-1.05270036e-01 -3.12037647e-01 -1.69190645e+00 7.59420514e-01
-5.72906613e-01 -1.41661179e+00 -8.10270533e-02 -1.36239067e-01
7.28997469e-01 2.65171260e-01 5.90085745e-01 5.86424470e-01
8.84140491e-01 -1.57045710e+00 1.20493090e+00 8.26699853e-01
7.64278531e-01 -1.33951449e+00 1.02702391e+00 3.90723854e-01
-5.12951136e-01 6.18171692e-03 -7.66309261e-01 1.99589655e-01
-8.08070749e-02 1.29032195e+00 -7.07547486e-01 -1.05111361e-01
9.45514560e-01 2.27883816e-01 -5.28787225e-02 1.02098179e+00
-2.30315074e-01 8.78110707e-01 -6.05052412e-01 -2.51291424e-01
1.69018194e-01 1.69376090e-01 2.14545920e-01 1.17610824e+00
4.33251709e-01 -1.45062342e-01 -6.12245083e-01 1.60507023e+00
-2.21794516e-01 -7.02324450e-01 -6.19821191e-01 -2.09834054e-01
8.26818526e-01 6.14340127e-01 -4.67158109e-01 -1.24071896e-01
-1.08147994e-01 1.56322911e-01 9.98439416e-02 5.34136832e-01
-1.11903834e+00 -3.80487770e-01 7.09095180e-01 -4.80747461e-01
2.76644304e-02 1.82873271e-02 -6.85415328e-01 -5.36243439e-01
-3.16321135e-01 -4.29121852e-01 2.54206568e-01 -7.98389256e-01
-1.51970136e+00 7.11090267e-01 5.02228796e-01 -9.33679521e-01
-7.32519031e-01 -1.02385128e+00 -5.95360219e-01 1.32245886e+00
-1.19610238e+00 -6.49993181e-01 3.29466075e-01 1.74925327e-01
5.47245778e-02 -2.64129192e-01 7.85295963e-01 -1.52897332e-02
-3.43406141e-01 7.34674096e-01 3.42552990e-01 8.43568221e-02
3.08016419e-01 -1.17676437e+00 3.16078335e-01 8.56889427e-01
3.78197357e-02 7.02847123e-01 1.18488610e+00 -5.36967337e-01
-7.56816268e-01 -1.16040254e+00 2.24670053e-01 -6.00452304e-01
7.46161997e-01 -8.25117081e-02 -7.73738682e-01 7.43965566e-01
-2.95119911e-01 5.19403853e-02 7.13963151e-01 3.50979298e-01
-7.26155818e-01 -3.13726753e-01 -1.36784184e+00 3.46636266e-01
4.55291837e-01 -5.97513855e-01 -6.42982423e-01 -1.47381052e-01
9.43939984e-01 -2.30520904e-01 -1.02259350e+00 5.69522560e-01
7.55684435e-01 -9.89675581e-01 7.46807396e-01 -6.55906856e-01
5.97440004e-01 -1.20912284e-01 -7.06780553e-01 -1.52980304e+00
-1.96766123e-01 -1.17460877e-01 -4.81092751e-01 1.35775340e+00
7.95657218e-01 -4.29665655e-01 7.34686911e-01 1.17938590e+00
-3.12296510e-01 -7.23687410e-01 -1.34751844e+00 -1.06729782e+00
6.77909791e-01 -1.33014274e+00 6.89082563e-01 1.52024135e-01
-2.69277424e-01 -2.03533038e-01 2.05633510e-02 4.01277810e-01
9.30738330e-01 -6.09934688e-01 -1.14078052e-01 -1.27778339e+00
-2.55008727e-01 -7.01524615e-01 -5.69542587e-01 -5.53008318e-01
3.25861365e-01 -4.02157992e-01 5.48023283e-01 -1.51082790e+00
1.08011074e-01 -6.45496070e-01 -7.87520885e-01 1.97206557e-01
2.81920105e-01 -9.37722623e-02 -2.70693243e-01 -1.57110855e-01
-3.38572532e-01 6.03277624e-01 3.41086566e-01 -2.48156771e-01
2.45344222e-01 2.70663977e-01 -8.39036465e-01 8.66974831e-01
8.30456972e-01 -5.92068613e-01 -7.38663137e-01 -3.47893327e-01
5.37799180e-01 -2.88025260e-01 4.42555875e-01 -1.40077305e+00
2.28299558e-01 -2.44310394e-01 8.32962692e-01 -8.29148233e-01
1.96738437e-01 -1.01434731e+00 2.73566961e-01 3.19461048e-01
-6.40340626e-01 -5.94234645e-01 5.76224208e-01 8.21277201e-01
-2.85846852e-02 -6.64180696e-01 9.65060592e-01 2.29533508e-01
-5.14846325e-01 4.00327027e-01 -5.45242310e-01 1.92441922e-02
8.00697327e-01 3.61327887e-01 -4.96541262e-01 -4.02139366e-01
-3.47398907e-01 1.07345499e-01 -1.15117125e-01 1.76830292e-01
7.45487869e-01 -1.39136052e+00 -6.82445347e-01 -1.78988606e-01
9.84706953e-02 1.17962383e-01 2.49012411e-01 2.78294533e-01
-3.35947633e-01 5.24519444e-01 -6.21780008e-02 -6.77259445e-01
-2.35623375e-01 2.97125578e-01 8.98511410e-01 1.84299275e-01
-7.52307102e-02 1.36241043e+00 3.37690681e-01 -4.10680585e-02
7.59200394e-01 -7.70414352e-01 -3.14064361e-02 -1.45363092e-01
8.69733810e-01 2.55607963e-01 2.02790573e-01 -4.87867706e-02
-3.54621410e-01 -2.88912028e-01 9.66613963e-02 -5.51086009e-01
1.08959305e+00 2.37636909e-01 2.22330362e-01 9.28644657e-01
9.10367131e-01 -4.02597755e-01 -1.91437638e+00 -8.90356675e-02
6.04099073e-02 -1.28756374e-01 7.45692015e-01 -1.01354849e+00
-9.40687776e-01 1.19772887e+00 8.11467111e-01 7.28132844e-04
8.55541825e-01 6.65951148e-02 9.90822837e-02 4.80235040e-01
4.78849083e-01 -1.12977076e+00 -3.52300435e-01 6.36441350e-01
8.52373481e-01 -1.01671100e+00 -1.99351460e-02 2.91718751e-01
-6.45389676e-01 9.92660522e-01 6.50502920e-01 -2.57281377e-03
1.22558534e+00 8.69155586e-01 -2.91826040e-01 1.01873778e-01
-9.53193128e-01 3.51370454e-01 7.27214873e-01 8.42848420e-01
4.34062123e-01 1.77031502e-01 3.37492615e-01 1.34915340e+00
-2.07322672e-01 -7.23112896e-02 2.09961295e-01 3.43209326e-01
-8.06734562e-01 -1.96270972e-01 -2.15454027e-01 9.33850884e-01
-6.87617004e-01 -3.09467316e-01 9.73519757e-02 3.69621128e-01
-1.47938147e-01 8.99513960e-01 4.88439709e-01 -4.76219535e-01
-5.13657667e-02 1.36385560e-01 5.01751781e-01 -4.90507752e-01
6.74867863e-03 -1.93532005e-01 2.53844053e-01 -3.07981163e-01
1.52527615e-01 -4.97031361e-01 -1.54918039e+00 -4.49756354e-01
-3.35770190e-01 1.14203341e-01 9.60069299e-01 9.26658690e-01
4.94996719e-02 8.65908504e-01 2.19541043e-01 -5.99626064e-01
-1.16869044e+00 -1.06216717e+00 -1.06892383e+00 -4.30596560e-01
3.47609460e-01 -8.40707242e-01 -6.99369729e-01 -1.29783288e-01] | [7.393165111541748, 3.813516616821289] |
0d9ed0e1-2419-4fa9-9f5c-9d9d587a97d7 | pose-estimation-of-specular-and-symmetrical | 2011.00372 | null | https://arxiv.org/abs/2011.00372v1 | https://arxiv.org/pdf/2011.00372v1.pdf | Pose Estimation of Specular and Symmetrical Objects | In the robotic industry, specular and textureless metallic components are ubiquitous. The 6D pose estimation of such objects with only a monocular RGB camera is difficult because of the absence of rich texture features. Furthermore, the appearance of specularity heavily depends on the camera viewpoint and environmental light conditions making traditional methods, like template matching, fail. In the last 30 years, pose estimation of the specular object has been a consistent challenge, and most related works require massive knowledge modeling effort for light setups, environment, or the object surface. On the other hand, recent works exhibit the feasibility of 6D pose estimation on a monocular camera with convolutional neural networks(CNNs) however they mostly use opaque objects for evaluation. This paper provides a data-driven solution to estimate the 6D pose of specular objects for grasping them, proposes a cost function for handling symmetry, and demonstrates experimental results showing the system's feasibility. | ['Henrik Christensen', 'Aayush Naik', 'Priyam Parashar', 'Hongyi Ling', 'Jiaming Hu'] | 2020-10-31 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 3.50304954e-02 -4.13947463e-01 6.41295910e-02 -3.77242327e-01
-7.63087720e-02 -5.82465827e-01 2.09065810e-01 -5.02179563e-01
-1.68702845e-02 2.12424934e-01 -3.10197055e-01 3.18772823e-01
-3.51298638e-02 -4.85130191e-01 -8.47939968e-01 -7.69161165e-01
4.12954628e-01 6.37908220e-01 2.32155055e-01 -2.77238768e-02
3.30585718e-01 7.14234829e-01 -1.73283041e+00 1.61963910e-01
2.40377188e-01 1.58105242e+00 5.92985451e-01 2.23320410e-01
-6.80047348e-02 2.08158389e-01 -4.57977265e-01 -3.74068886e-01
7.23128080e-01 3.08206379e-01 -3.29385281e-01 2.96562433e-01
7.80320525e-01 -7.18301654e-01 -4.72891241e-01 1.01268172e+00
4.10502553e-01 -2.48240680e-01 4.78980094e-01 -1.29099596e+00
-5.03710806e-01 1.91611394e-01 -5.40009379e-01 -7.11980522e-01
6.02229714e-01 -3.29889767e-02 4.31814522e-01 -9.47462320e-01
7.55620956e-01 1.19234216e+00 7.10183978e-01 5.83738148e-01
-3.46964598e-01 -5.48869073e-01 3.15060802e-02 3.71273398e-01
-1.00506544e+00 -1.16912752e-01 1.18542838e+00 -2.35482410e-01
5.45970082e-01 1.86116904e-01 6.57032907e-01 1.25998187e+00
2.93866307e-01 7.35033095e-01 1.29185963e+00 -4.46744442e-01
1.91105559e-01 1.50786772e-01 -8.59139487e-02 5.63553810e-01
6.40565932e-01 -2.35186741e-02 -6.00029111e-01 1.81120578e-02
9.45639789e-01 2.68832177e-01 -2.60994136e-01 -1.24167728e+00
-9.53526258e-01 6.21157028e-02 3.67371082e-01 -1.27092972e-01
-1.39124587e-01 2.11610273e-01 1.57439813e-01 1.61005497e-01
1.72323674e-01 4.15762275e-01 -6.17032707e-01 -1.54334143e-01
-2.91872900e-02 7.88152218e-02 1.00418556e+00 1.43346000e+00
5.55446267e-01 -1.41335562e-01 4.62127149e-01 7.68724740e-01
2.62785763e-01 9.66560900e-01 -1.40896574e-01 -9.51237440e-01
5.23536325e-01 6.27216041e-01 5.39067090e-01 -1.12398958e+00
-5.58995903e-01 -3.55936855e-01 -3.95999014e-01 3.24513227e-01
5.31661749e-01 -3.45902890e-02 -6.43157959e-01 1.14858651e+00
5.95585406e-01 -3.80023420e-01 -1.85991526e-01 1.39745378e+00
9.22747433e-01 8.70739147e-02 -7.93720841e-01 1.91166058e-01
1.05201840e+00 -8.56891751e-01 -6.18301570e-01 -2.44992912e-01
1.60181075e-01 -1.20452023e+00 8.35168004e-01 7.54182875e-01
-1.09731817e+00 -3.64120811e-01 -9.29313481e-01 -5.05479705e-03
-2.59274721e-01 5.53176463e-01 8.89029026e-01 6.76793993e-01
-4.78148758e-01 3.62920523e-01 -9.19477344e-01 -3.51999283e-01
-3.02782916e-02 5.54167271e-01 -4.72343773e-01 -5.06173849e-01
-5.43495536e-01 1.25350118e+00 1.38637424e-01 5.36107540e-01
-6.16287529e-01 -5.17270982e-01 -5.24759591e-01 -2.37666905e-01
6.08856738e-01 -4.68716025e-01 1.05815995e+00 -8.16336036e-01
-1.96128368e+00 8.70145619e-01 2.99551427e-01 2.23646998e-01
6.31577611e-01 -7.29313195e-01 1.50864288e-01 2.30647460e-01
-4.47082371e-01 1.67557657e-01 8.86170149e-01 -1.57616770e+00
-1.77109651e-02 -9.24916506e-01 4.39558715e-01 4.10556614e-01
-1.73190966e-01 -8.32048357e-02 -4.95422393e-01 1.96661614e-02
7.85323560e-01 -1.20538044e+00 1.10653915e-01 5.80904543e-01
-2.30114207e-01 1.46497205e-01 1.24285150e+00 -4.27600682e-01
1.24142088e-01 -2.01305938e+00 -1.63603332e-02 -8.11783895e-02
-2.42171854e-01 1.03126630e-01 1.77929886e-02 3.33092123e-01
1.38367996e-01 -7.69421637e-01 3.47534984e-01 -1.65297940e-01
9.30472910e-02 -1.17716402e-01 -9.63492393e-02 8.66353035e-01
-1.44531116e-01 6.10468566e-01 -5.95374227e-01 -6.37368709e-02
4.70197409e-01 6.12766266e-01 -2.57430792e-01 3.70751441e-01
-2.58202612e-01 1.78844661e-01 -4.03028786e-01 1.10138226e+00
1.19691718e+00 3.30943405e-03 2.40388006e-01 -6.75616443e-01
-2.10953608e-01 -5.34769893e-02 -1.48741674e+00 1.70273745e+00
-6.31419003e-01 4.10857588e-01 4.90979850e-01 -5.91247141e-01
1.23225212e+00 1.12550430e-01 5.04340410e-01 -5.90136409e-01
3.91595662e-01 5.68707168e-01 -2.89184541e-01 -6.53409660e-01
3.87595654e-01 1.99357092e-01 2.73850679e-01 1.98895410e-01
-3.01355898e-01 -6.22724235e-01 -3.32313806e-01 -5.34358740e-01
7.84703672e-01 8.62400115e-01 -1.34683803e-01 -6.27789199e-02
1.24316484e-01 -2.26225276e-02 4.10913527e-01 4.92419690e-01
1.46023333e-01 8.98104548e-01 1.60278343e-02 -8.30270410e-01
-1.00082815e+00 -1.04615426e+00 -2.06307888e-01 3.99893463e-01
1.01965189e+00 2.85019755e-01 -6.10805631e-01 -1.68949544e-01
3.98051560e-01 2.00465992e-01 -3.49554360e-01 -6.48548603e-02
-7.66485035e-01 -6.04072630e-01 -1.71762675e-01 5.66261649e-01
5.94234169e-01 -7.77550936e-01 -1.19136465e+00 -1.47132799e-02
-7.97041878e-02 -1.51585221e+00 1.31282955e-02 1.57506928e-01
-8.67422342e-01 -1.41020131e+00 -7.06829429e-01 -6.22352123e-01
7.46019661e-01 9.21119332e-01 9.46241677e-01 -2.11953625e-01
-7.46028662e-01 7.63194144e-01 -3.45842302e-01 -7.16215253e-01
2.41052300e-01 -2.30674341e-01 1.92789942e-01 -1.27639445e-02
3.05083185e-01 -2.89398342e-01 -8.25799048e-01 8.42190564e-01
-4.90528464e-01 7.30182379e-02 4.88813102e-01 5.33127964e-01
2.56754726e-01 -1.59633711e-01 -7.08400607e-02 -3.03059459e-01
-1.41125962e-01 -6.93119913e-02 -1.10833848e+00 1.61199495e-01
3.27288546e-02 -4.43688244e-01 3.12020570e-01 -6.33814096e-01
-1.36104655e+00 2.94686913e-01 5.45686662e-01 -6.77457690e-01
-2.26081863e-01 1.12422109e-01 -2.17445165e-01 -4.44256723e-01
4.93006617e-01 -8.65661353e-02 8.22551250e-02 -5.45318067e-01
-2.29821011e-01 6.57849848e-01 2.85692602e-01 -6.35586083e-01
8.50847781e-01 9.84949350e-01 2.20763654e-01 -9.92213190e-01
-9.62421060e-01 -6.46067798e-01 -7.82897770e-01 -6.10513806e-01
5.99029362e-01 -8.17036688e-01 -1.24899685e+00 7.76209414e-01
-1.45617306e+00 -7.85439983e-02 2.21327558e-01 9.49141324e-01
-7.48339355e-01 3.12434971e-01 -3.81158054e-01 -1.04071510e+00
-2.48335034e-01 -1.25685906e+00 1.41804743e+00 9.75620151e-02
1.44712478e-01 -3.93768311e-01 -3.70752901e-01 7.88088500e-01
3.20611626e-01 2.77431875e-01 5.63592970e-01 4.07849669e-01
-1.29431891e+00 -4.49345827e-01 -4.48936254e-01 1.47248298e-01
3.40590626e-01 -3.51421274e-02 -1.18007684e+00 -2.45701298e-01
2.62759566e-01 -4.28041965e-01 3.11932385e-01 4.74721193e-01
1.13531017e+00 2.26308554e-01 -1.78908914e-01 8.67049098e-01
1.61290717e+00 2.65420854e-01 3.71854037e-01 4.83485907e-01
8.44218254e-01 8.68851125e-01 1.03379261e+00 4.09053266e-01
6.34568557e-02 9.92863238e-01 1.28412795e+00 1.06851086e-01
-4.49281186e-02 3.37355971e-01 3.16693246e-01 6.92944586e-01
-4.12001669e-01 -5.11592776e-02 -7.52078593e-01 1.19823180e-01
-1.47185111e+00 -5.79795420e-01 -2.89651692e-01 2.32814360e+00
3.26763213e-01 -1.07804224e-01 -4.52799469e-01 -6.86457555e-04
5.31482041e-01 -2.63822734e-01 -7.26173580e-01 -8.88471454e-02
-2.06858128e-01 -1.46976039e-01 7.30229855e-01 7.02178404e-02
-8.57129753e-01 8.05805564e-01 5.73102283e+00 3.37047964e-01
-1.36299777e+00 -2.94843972e-01 -1.57051846e-01 -1.73414677e-01
3.08373244e-03 -2.82664716e-01 -9.55140591e-01 1.77188739e-01
-1.06866837e-01 6.09532654e-01 4.10926551e-01 1.24316287e+00
-2.07618713e-01 -4.61595893e-01 -1.05696034e+00 1.34809899e+00
4.44231391e-01 -7.19844103e-01 -3.09084147e-01 -1.61924571e-01
7.98355043e-01 1.49097309e-01 3.01935911e-01 -2.84918159e-01
-1.21930137e-01 -5.48763752e-01 9.38703597e-01 4.44097847e-01
4.84330446e-01 -3.76044422e-01 8.87518167e-01 2.77147591e-01
-7.93105364e-01 -1.40076384e-01 -9.17495668e-01 -1.30252048e-01
-8.15035030e-02 6.76368892e-01 -8.96602392e-01 4.82758820e-01
1.07886755e+00 3.21895868e-01 -1.71136424e-01 1.26821375e+00
-3.30223851e-02 1.19695989e-02 -5.09354293e-01 -2.74506748e-01
3.11881062e-02 -3.75070453e-01 4.98170316e-01 6.81062222e-01
3.92102271e-01 -3.29502970e-02 1.93513095e-01 8.01902413e-01
1.51438102e-01 -2.81879026e-02 -6.40897155e-01 3.39585721e-01
-5.67131341e-02 1.40437663e+00 -7.59889245e-01 4.41179514e-01
-5.31296253e-01 7.78520226e-01 4.41091917e-02 2.03262851e-01
-6.61737919e-01 -2.54009794e-02 4.50085342e-01 1.30167320e-01
3.62811059e-01 -5.73140442e-01 -3.93778652e-01 -1.26653707e+00
8.43996227e-01 -6.58221185e-01 -4.26175952e-01 -1.40186930e+00
-1.19652987e+00 4.03008044e-01 -6.52612597e-02 -1.52174985e+00
3.88697952e-01 -1.37552059e+00 -8.30406994e-02 5.63666463e-01
-1.58680296e+00 -1.38668513e+00 -8.69997680e-01 3.61214012e-01
4.49533969e-01 -4.59006475e-03 7.10029066e-01 8.51421990e-03
-6.21691719e-02 1.23004250e-01 2.13467956e-01 -2.53605276e-01
8.37357402e-01 -8.50660145e-01 -2.58581877e-01 7.37775192e-02
-3.05019170e-01 5.87341130e-01 8.57505679e-01 -4.95282084e-01
-2.51609302e+00 -4.88201737e-01 1.85567662e-01 -6.14267647e-01
4.46902156e-01 -7.45745599e-01 -4.00815725e-01 2.88924873e-01
-1.23456866e-01 6.20091148e-02 3.53912413e-02 9.66548547e-02
-5.47141194e-01 -3.55942756e-01 -1.05436742e+00 5.22725880e-01
1.15264475e+00 -2.99953192e-01 -3.20806324e-01 5.21446943e-01
2.02641919e-01 -1.04398322e+00 -6.99707925e-01 7.72868872e-01
1.24182677e+00 -1.14636016e+00 1.10556269e+00 -8.46513733e-02
3.71321023e-01 -3.18644494e-01 -3.44357610e-01 -9.89146829e-01
1.06564030e-01 -2.82362640e-01 1.84517112e-02 5.77055693e-01
-6.55370727e-02 -5.18004954e-01 1.27174842e+00 6.71441913e-01
-2.67145038e-01 -6.33184731e-01 -7.43048906e-01 -1.01628530e+00
-5.64576328e-01 -1.52597383e-01 4.95518833e-01 6.29114032e-01
-4.26025033e-01 -2.56388158e-01 -3.06155980e-01 5.93132198e-01
7.47842371e-01 9.49248195e-01 1.07751334e+00 -1.54628074e+00
-1.68620586e-01 8.16062540e-02 -4.74922508e-01 -1.01638842e+00
-1.96774676e-02 -3.57650161e-01 2.44471356e-01 -1.30814159e+00
4.84302342e-01 -6.31345868e-01 1.26755655e-01 4.69755232e-02
2.51559168e-01 2.66634464e-01 7.13990405e-02 3.58911157e-02
-4.83579695e-01 4.99686420e-01 1.57815015e+00 -1.05004810e-01
1.22713886e-01 1.81349948e-01 -1.54689066e-02 1.09010279e+00
7.45322883e-01 -2.07722455e-01 -2.35863581e-01 -1.01260638e+00
5.43140173e-01 2.36788280e-02 6.12989962e-01 -1.00626945e+00
9.84060168e-02 -3.26163709e-01 4.66567159e-01 -9.68713522e-01
1.01445067e+00 -1.57530570e+00 2.31461301e-01 4.73184884e-01
1.66828722e-01 -1.36226147e-01 -4.38295724e-03 3.41509610e-01
-7.61075094e-02 -4.61519986e-01 6.12169206e-01 -3.94621760e-01
-4.40870464e-01 3.96400988e-01 2.50457495e-01 -4.30609763e-01
9.09868360e-01 -7.00094879e-01 -3.28853250e-01 -2.57342726e-01
-2.50165641e-01 -2.90628999e-01 8.08741987e-01 5.11261821e-01
7.28574097e-01 -1.10458255e+00 -2.41646945e-01 1.47727445e-01
1.72088549e-01 2.73898244e-01 1.13243267e-01 6.60996139e-01
-9.13816273e-01 3.93327028e-01 -4.42603379e-01 -8.35872412e-01
-1.45555449e+00 4.20763165e-01 2.89768875e-01 4.64912385e-01
-4.51273203e-01 6.85960531e-01 4.03142929e-01 -6.91936016e-01
3.90992135e-01 -2.71412492e-01 2.33053014e-01 -3.93608510e-01
1.85401306e-01 6.30410731e-01 3.35128218e-01 -3.34275961e-01
-1.03421390e-01 1.11631775e+00 1.32433996e-01 2.98192024e-01
1.64738393e+00 -1.40637293e-01 -2.14503288e-01 3.26811463e-01
1.03409600e+00 -1.38930157e-01 -1.45671010e+00 -1.92569748e-01
-2.09442899e-01 -7.07241178e-01 -1.13020234e-01 -6.90822363e-01
-1.26444018e+00 1.01012540e+00 4.69924122e-01 -3.23241055e-01
8.00349712e-01 -1.92501307e-01 5.86403966e-01 1.01552665e+00
1.13766539e+00 -1.36563826e+00 3.38050246e-01 7.30747759e-01
1.17879927e+00 -1.29079413e+00 -2.34464626e-03 -1.07679987e+00
-1.17202334e-01 1.57041514e+00 9.55717564e-01 -1.94594577e-01
6.19282126e-01 5.57261527e-01 2.55713731e-01 -2.09263757e-01
-2.52538741e-01 2.55016655e-01 2.82587200e-01 6.65094733e-01
-7.64749106e-03 -1.05840251e-01 1.42936915e-01 1.13471802e-02
-1.67543113e-01 -1.01727664e-01 4.24958527e-01 1.19238508e+00
-2.97051311e-01 -8.74605536e-01 -6.87454939e-01 3.98342684e-02
-3.47845286e-01 4.06110734e-01 -4.05232996e-01 7.17800796e-01
1.10828578e-01 6.56442046e-01 -1.20786600e-01 -5.06997220e-02
5.29411256e-01 -3.49344641e-01 1.19669247e+00 -5.30630767e-01
-3.02225977e-01 1.84295960e-02 -5.27177788e-02 -7.47998834e-01
-6.53161168e-01 -6.56152368e-01 -7.54656255e-01 1.57568634e-01
-7.78589904e-01 -4.50283766e-01 1.51778209e+00 7.50235498e-01
6.11965358e-02 -1.07835233e-01 7.07418919e-01 -1.12199366e+00
-7.49221385e-01 -7.30104625e-01 -6.26506686e-01 3.96022260e-01
-1.62558146e-02 -1.18153310e+00 -2.64899611e-01 -1.97688043e-01] | [7.021805763244629, -2.0793802738189697] |
aa074d41-5fea-4d1f-8689-9935acb04550 | cerfgan-a-compact-effective-robust-and-fast | 1805.10871 | null | http://arxiv.org/abs/1805.10871v2 | http://arxiv.org/pdf/1805.10871v2.pdf | CerfGAN: A Compact, Effective, Robust, and Fast Model for Unsupervised Multi-Domain Image-to-Image Translation | In this paper, we aim at solving the multi-domain image-to-image translation
problem with a unified model in an unsupervised manner. The most successful
work in this area refers to StarGAN, which works well in tasks like face
attribute modulation. However, StarGAN is unable to match multiple translation
mappings when encountering general translations with very diverse domain
shifts. On the other hand, StarGAN adopts an Encoder-Decoder-Discriminator
(EDD) architecture, where the model is time-consuming and unstable to train. To
this end, we propose a Compact, effective, robust, and fast GAN model, termed
CerfGAN, to solve the above problem. In principle, CerfGAN contains a novel
component, i.e., a multi-class discriminator (MCD), which gives the model an
extremely powerful ability to match multiple translation mappings. To stabilize
the training process, MCD also plays a role of the encoder in CerfGAN, which
saves a lot of computation and memory costs. We perform extensive experiments
to verify the effectiveness of the proposed method. Quantitatively, CerfGAN is
demonstrated to handle a serial of image-to-image translation tasks including
style transfer, season transfer, face hallucination, etc, where the input
images are sampled from diverse domains. The comparisons to several recently
proposed approaches demonstrate the superiority and novelty of the proposed
method. | ['Xiao Liu', 'Hong Liu', 'Shengchuan Zhang', 'Xin Liu', 'Rongrong Ji', 'Cheng Deng'] | 2018-05-28 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 4.67373073e-01 -1.30483195e-01 -7.36045092e-02 -1.95798099e-01
-6.10814810e-01 -2.59432346e-01 6.38372719e-01 -6.59642041e-01
4.29931544e-02 8.36643219e-01 1.12158787e-02 2.40891217e-03
2.49278918e-01 -7.53842115e-01 -7.79220045e-01 -8.82369757e-01
6.88087821e-01 3.60656470e-01 -4.30319943e-02 -3.83864284e-01
-4.59110551e-02 2.23942056e-01 -1.20766342e+00 -8.94226208e-02
9.86581504e-01 9.52531993e-01 3.18961978e-01 2.64986362e-02
-1.04233094e-01 5.08502543e-01 -6.36088431e-01 -6.79126084e-01
2.77773798e-01 -1.16555858e+00 -4.35211629e-01 2.41253197e-01
2.30046451e-01 -3.06378633e-01 -3.42102587e-01 1.20579600e+00
8.51507068e-01 -1.00897677e-01 6.85494065e-01 -1.39132798e+00
-1.01083612e+00 2.88178056e-01 -8.10022831e-01 -9.78562608e-02
1.98008031e-01 1.31933287e-01 4.27138686e-01 -8.32313895e-01
5.05264163e-01 1.48996568e+00 3.59735876e-01 8.80479276e-01
-1.00637078e+00 -1.10187745e+00 -1.54949710e-01 7.98237100e-02
-1.48253369e+00 -6.85687661e-01 8.87211561e-01 -1.51545376e-01
2.70492494e-01 7.20752776e-02 3.91128033e-01 1.23846304e+00
2.43765712e-01 6.30141020e-01 1.40051341e+00 -3.39219987e-01
-3.66781168e-02 2.52072096e-01 -7.44008541e-01 6.22995079e-01
1.04690731e-01 -1.65837444e-02 -4.63378906e-01 1.39312912e-02
1.12016094e+00 -7.25892885e-03 -4.63951051e-01 -1.01806119e-01
-1.02822530e+00 8.09936464e-01 3.06070030e-01 3.26552153e-01
-1.48591518e-01 -2.39239149e-02 2.50809669e-01 6.52122855e-01
5.40470243e-01 4.08326611e-02 -1.27109364e-01 -1.68982166e-04
-7.46568739e-01 -6.99228942e-02 5.12519419e-01 1.23722696e+00
7.22191095e-01 4.22721833e-01 -2.13282168e-01 1.00413668e+00
2.06592605e-01 6.23549759e-01 6.66779637e-01 -5.60324430e-01
5.31424403e-01 2.95991004e-01 -2.34980032e-01 -1.04897702e+00
1.39401108e-01 -5.73341250e-01 -1.37879860e+00 7.49419853e-02
4.01463956e-02 -1.60695463e-01 -8.11533630e-01 2.12671328e+00
3.28486502e-01 3.28279853e-01 2.51283288e-01 8.93630624e-01
6.69889927e-01 8.28282058e-01 -8.01035687e-02 -4.01564777e-01
1.25234771e+00 -1.06313622e+00 -9.10066128e-01 -1.69220313e-01
9.93443057e-02 -1.05302739e+00 1.00898707e+00 1.05597407e-01
-1.11274636e+00 -7.97110796e-01 -1.04402971e+00 1.11635141e-01
5.56553714e-02 1.47998527e-01 4.54067945e-01 5.26033521e-01
-1.03761065e+00 3.01263154e-01 -3.91870409e-01 -4.30998027e-01
5.69662511e-01 4.39041257e-01 -5.10252357e-01 -1.93670690e-01
-1.27430451e+00 8.45386446e-01 3.98126304e-01 5.86401299e-02
-9.29214478e-01 -2.15275437e-01 -8.01866472e-01 1.31343678e-01
2.59980291e-01 -8.86444986e-01 1.02216303e+00 -1.47365332e+00
-1.92494690e+00 7.06650198e-01 -1.77479148e-01 -1.61716193e-01
6.55850530e-01 1.65073723e-01 -5.61126828e-01 1.31100014e-01
1.23789273e-01 6.99323356e-01 1.28012586e+00 -1.17065036e+00
-3.60870302e-01 -3.66870791e-01 -2.11670533e-01 3.81762773e-01
-6.56147003e-01 7.17692748e-02 -5.44615805e-01 -9.28283274e-01
-7.17585310e-02 -9.18411851e-01 1.51303053e-01 -4.65030111e-02
-1.39188319e-01 -1.86038390e-01 1.05157959e+00 -6.03869438e-01
1.07714057e+00 -2.39131451e+00 3.60031724e-01 2.60768496e-02
1.14788458e-01 4.28240389e-01 -2.15043768e-01 4.63308424e-01
5.58008961e-02 -5.50526790e-02 -4.90573317e-01 -3.14431787e-01
-1.91659197e-01 2.72302359e-01 -2.17731252e-01 3.64418298e-01
3.26630354e-01 1.02694833e+00 -6.67464197e-01 -6.61939800e-01
-5.10607436e-02 5.32215834e-01 -3.69421750e-01 4.31309938e-01
-7.00274110e-02 8.68650258e-01 -4.88521397e-01 6.92300320e-01
9.80367124e-01 -2.58227289e-01 8.24609697e-02 -4.56872620e-02
2.28642479e-01 -1.67238981e-01 -9.88628387e-01 1.89684176e+00
-5.07401526e-01 4.43806559e-01 1.55703062e-02 -1.06298518e+00
1.29339254e+00 5.25709093e-01 2.03356102e-01 -8.42008531e-01
1.90891206e-01 4.84154671e-01 -6.20131604e-02 -3.31262231e-01
1.13515854e-01 -5.29984057e-01 5.70214428e-02 3.94999951e-01
3.30092609e-01 -2.21296791e-02 -2.11626559e-01 -1.06426284e-01
7.23289847e-01 1.64565161e-01 3.62883985e-01 -1.80847734e-01
9.01033998e-01 -3.52672517e-01 7.14231193e-01 1.79236248e-01
4.72128652e-02 6.08065069e-01 3.22110146e-01 -5.63661717e-02
-1.14666080e+00 -8.39851737e-01 1.19949162e-01 5.58073223e-01
4.06942636e-01 -1.62474409e-01 -9.49039042e-01 -5.36346018e-01
-3.44406903e-01 3.48756373e-01 -2.69875675e-01 -4.96533185e-01
-5.85384369e-01 -6.06869459e-01 6.78208411e-01 2.76584506e-01
1.22182560e+00 -9.98298943e-01 -6.21474981e-02 1.81803763e-01
-3.17853272e-01 -1.05755448e+00 -8.40038359e-01 -3.21325421e-01
-8.50167453e-01 -6.43021822e-01 -1.07536387e+00 -1.25930917e+00
7.05471098e-01 3.43190104e-01 7.86058962e-01 -9.32160839e-02
2.19704047e-01 -8.42703059e-02 -4.11510825e-01 -3.24768573e-01
-4.62947220e-01 -5.86784154e-05 2.72606611e-02 3.79825920e-01
1.80960819e-01 -8.30823898e-01 -6.01475596e-01 5.69899738e-01
-1.10495484e+00 1.92747936e-01 8.58241796e-01 1.23408997e+00
5.80363870e-01 1.76905513e-01 8.25900853e-01 -9.24503922e-01
6.71138167e-01 -4.62675780e-01 -4.70549941e-01 2.09114581e-01
-5.93203068e-01 5.80177158e-02 1.10281122e+00 -6.56707048e-01
-1.14415455e+00 -2.12219637e-02 -2.17794105e-01 -6.79313600e-01
1.27305269e-01 2.59001464e-01 -6.95637882e-01 -3.69485974e-01
3.44099373e-01 7.81110704e-01 3.29567075e-01 -4.61577922e-01
2.87618071e-01 8.81452501e-01 6.19634807e-01 -5.28695166e-01
1.13997662e+00 2.27259114e-01 7.34606460e-02 -4.99060750e-01
-4.18497235e-01 1.14195511e-01 -2.58747756e-01 3.59066874e-02
6.66776299e-01 -1.10910189e+00 -4.32738006e-01 7.32175410e-01
-1.21141231e+00 -5.16540892e-02 -5.35964295e-02 4.66879040e-01
-6.35481060e-01 2.35361710e-01 -5.65853238e-01 -5.41460454e-01
-5.10164559e-01 -1.08218336e+00 9.75048065e-01 4.77149785e-01
3.15345049e-01 -8.45806837e-01 -1.82641730e-01 2.60099560e-01
5.20711899e-01 3.88201445e-01 8.48418474e-01 -3.62729818e-01
-5.81768572e-01 1.67813286e-01 -2.46385366e-01 5.29049814e-01
4.12400484e-01 -4.43776548e-01 -8.56102347e-01 -5.71384549e-01
3.49315226e-01 -3.94426972e-01 6.81517720e-01 -4.58058678e-02
1.13341939e+00 -4.75910604e-01 -1.69092253e-01 9.15037632e-01
1.50412357e+00 5.47979653e-01 9.34835315e-01 1.31832153e-01
6.82156503e-01 2.93365270e-01 5.99003792e-01 2.46983036e-01
3.09725851e-01 8.31831813e-01 2.86258519e-01 -4.52926457e-01
-2.49747455e-01 -5.10005832e-01 5.11107326e-01 1.14193702e+00
-5.97567372e-02 -4.29347306e-01 -3.15014303e-01 2.19298854e-01
-1.68455684e+00 -9.06117857e-01 2.56172419e-01 1.91240668e+00
9.60557878e-01 -1.37305602e-01 -1.14695489e-01 -5.55790868e-03
9.75148797e-01 1.61448494e-01 -5.52697718e-01 -3.60381633e-01
-3.80535811e-01 5.09437382e-01 2.45719641e-01 1.86162412e-01
-8.15885723e-01 1.13814473e+00 5.47916460e+00 1.25271440e+00
-1.29716539e+00 3.88024807e-01 6.81024194e-01 4.63378906e-01
-3.01707864e-01 1.89721324e-02 -5.18373311e-01 7.75476038e-01
5.83532691e-01 -2.30054051e-01 7.00543165e-01 6.72923446e-01
-5.66368364e-02 2.95252621e-01 -8.94170523e-01 1.26102567e+00
4.30824459e-01 -1.03639507e+00 3.56229484e-01 -4.42605577e-02
7.89479554e-01 -5.33417583e-01 3.19129974e-01 1.19049363e-01
-4.34900708e-02 -1.08681440e+00 5.86699426e-01 2.39650264e-01
1.52914810e+00 -8.70044768e-01 6.98244035e-01 3.71093571e-01
-1.09679127e+00 5.66411391e-02 -6.44498706e-01 1.29880458e-01
-7.85713196e-02 2.55853891e-01 -5.13840497e-01 9.57802176e-01
2.87166446e-01 8.37353587e-01 -3.43203604e-01 6.17374003e-01
-2.89211899e-01 4.41030711e-01 -4.11501192e-02 2.34239787e-01
2.46629864e-01 -4.28629965e-01 3.73300225e-01 8.88318837e-01
8.34202588e-01 2.74213791e-01 -4.99664210e-02 9.19341266e-01
-3.69081169e-01 1.93506360e-01 -8.41883183e-01 5.03282174e-02
5.24900198e-01 1.14230335e+00 -4.29308951e-01 -2.48820469e-01
-4.89381641e-01 1.49366069e+00 2.23056395e-02 2.81359404e-01
-9.46755171e-01 -5.14282286e-01 4.43589300e-01 3.14843133e-02
2.68192649e-01 -8.60089213e-02 1.23279562e-04 -1.28410828e+00
1.30924016e-01 -1.34773660e+00 1.76016852e-01 -7.73535907e-01
-1.21584022e+00 9.26557004e-01 -1.59753099e-01 -1.61638522e+00
-2.36633524e-01 -2.76441723e-01 -5.64588368e-01 1.09655261e+00
-1.64860177e+00 -1.42562091e+00 -4.69247162e-01 1.08377278e+00
5.08859217e-01 -5.98236740e-01 7.45903790e-01 5.80872595e-01
-4.71545756e-01 9.68762755e-01 2.01651230e-01 7.00435638e-02
1.02661157e+00 -7.27550745e-01 2.25763202e-01 8.32629979e-01
-1.79765429e-02 4.87149149e-01 3.81684124e-01 -5.91131270e-01
-1.51792395e+00 -1.28729022e+00 6.62742972e-01 4.13510278e-02
2.61091679e-01 -4.00015861e-01 -8.20858836e-01 6.05614185e-01
5.40387690e-01 -8.28043744e-02 5.36056876e-01 -6.12499535e-01
-2.25419268e-01 -4.35570538e-01 -1.25217593e+00 4.62311387e-01
1.18377531e+00 -4.58987892e-01 -3.71411055e-01 2.10832432e-01
6.79731131e-01 -4.82567668e-01 -6.93248332e-01 3.95864636e-01
3.43188584e-01 -7.83795774e-01 7.67109275e-01 -2.00869739e-01
6.92335129e-01 -4.78844196e-01 -1.08489424e-01 -1.45951283e+00
-3.30136299e-01 -8.91802371e-01 -9.76095535e-03 1.58930409e+00
1.73888877e-02 -9.31449831e-01 4.72041339e-01 -1.84348285e-01
-1.41732886e-01 -6.90134466e-01 -1.07471514e+00 -9.14700866e-01
1.08261950e-01 3.49761128e-01 1.01059306e+00 9.95738626e-01
-5.00634849e-01 6.83551788e-01 -1.00102437e+00 -1.26485363e-01
5.33807218e-01 1.76052064e-01 9.01081204e-01 -1.03532434e+00
-4.24740404e-01 -1.37817293e-01 -4.12038624e-01 -1.23398197e+00
2.38577858e-01 -9.27836239e-01 -1.43223107e-01 -1.12674534e+00
2.36462653e-01 -3.99669260e-01 -2.23183259e-01 2.79040009e-01
-2.95281351e-01 4.05490696e-01 2.56107301e-01 5.10353327e-01
-2.10930139e-01 9.13182855e-01 1.84515905e+00 -2.21799150e-01
1.23619670e-02 -8.07782859e-02 -9.85187352e-01 3.45027566e-01
8.73919606e-01 -4.15617883e-01 -6.79843366e-01 -6.32928967e-01
-2.96446681e-01 3.05829108e-01 2.08487242e-01 -9.22234178e-01
1.41063139e-01 -1.77619904e-01 5.20851970e-01 -1.07696570e-01
4.13050592e-01 -8.60507607e-01 5.30809224e-01 5.25448859e-01
2.33594049e-02 2.09719732e-01 -1.53988916e-02 4.20242727e-01
-6.64250374e-01 8.61370414e-02 1.06866765e+00 -1.20748043e-01
-5.32872260e-01 6.71103477e-01 1.63054898e-01 5.36051877e-02
1.07397509e+00 -1.00050621e-01 -3.18005115e-01 -5.63082457e-01
-2.26329386e-01 7.90437907e-02 5.51366448e-01 4.35441375e-01
7.32154548e-01 -1.84118295e+00 -9.33230817e-01 5.74530959e-01
8.09680074e-02 -7.29591772e-02 1.77835807e-01 7.00622559e-01
-4.08413380e-01 1.44012377e-01 -6.35902762e-01 -5.37823677e-01
-1.05601835e+00 5.82848847e-01 1.65389955e-01 -1.50708318e-01
-4.88521546e-01 8.12418699e-01 4.18687224e-01 -1.89062074e-01
-1.74900353e-01 3.16915661e-01 -1.50424391e-01 -1.61891446e-01
3.31257373e-01 2.14168224e-02 -2.56168276e-01 -8.55839610e-01
-2.22624093e-01 8.79154921e-01 5.37545644e-02 -6.89926594e-02
1.16509616e+00 -2.10507184e-01 -3.37901831e-01 1.16885014e-01
1.18209839e+00 -1.36972681e-01 -1.11384284e+00 -4.18045968e-01
-5.49674571e-01 -7.00812936e-01 -3.15652758e-01 -5.60437024e-01
-1.34566355e+00 9.98943388e-01 6.30867422e-01 -2.03118831e-01
1.61595011e+00 -1.87256962e-01 1.20151806e+00 -5.51849790e-02
5.34132898e-01 -7.72894859e-01 1.66167572e-01 1.63107529e-01
8.86908174e-01 -1.13697159e+00 -2.89068252e-01 -4.48666275e-01
-7.30370343e-01 1.09694946e+00 8.81566226e-01 -1.18173854e-02
3.09574544e-01 1.16608545e-01 5.53669233e-04 2.69384861e-01
-5.23692369e-01 3.60064209e-02 3.62222195e-02 6.42708421e-01
1.03298523e-01 -3.81117724e-02 -5.87637663e-01 4.11105543e-01
-2.09548801e-01 1.26739934e-01 2.98277855e-01 5.85774720e-01
-1.69048414e-01 -1.53695655e+00 -4.05483007e-01 1.23938538e-01
-4.31175411e-01 -7.44999424e-02 -2.79324770e-01 6.89736664e-01
3.32824707e-01 9.06090736e-01 -1.29216507e-01 -6.42190099e-01
1.95370197e-01 -1.04812413e-01 5.59122980e-01 -4.50508356e-01
-3.70732188e-01 3.05586427e-01 -3.55002254e-01 -2.83910811e-01
-5.19378304e-01 -3.44890803e-01 -8.39791298e-01 -4.36774343e-01
-3.05457026e-01 2.40123495e-01 4.22723889e-01 7.45209157e-01
3.49043071e-01 4.85700488e-01 1.14861381e+00 -3.45633090e-01
-7.26774156e-01 -1.02651095e+00 -5.70554256e-01 5.25623322e-01
2.25135744e-01 -6.19996727e-01 -4.27301303e-02 6.25789762e-02] | [11.885010719299316, -0.37508898973464966] |
a40eb5ea-cadd-40b6-bff7-94a0ddabeb5c | mp-senet-a-speech-enhancement-model-with | 2305.13686 | null | https://arxiv.org/abs/2305.13686v1 | https://arxiv.org/pdf/2305.13686v1.pdf | MP-SENet: A Speech Enhancement Model with Parallel Denoising of Magnitude and Phase Spectra | This paper proposes MP-SENet, a novel Speech Enhancement Network which directly denoises Magnitude and Phase spectra in parallel. The proposed MP-SENet adopts a codec architecture in which the encoder and decoder are bridged by convolution-augmented transformers. The encoder aims to encode time-frequency representations from the input noisy magnitude and phase spectra. The decoder is composed of parallel magnitude mask decoder and phase decoder, directly recovering clean magnitude spectra and clean-wrapped phase spectra by incorporating learnable sigmoid activation and parallel phase estimation architecture, respectively. Multi-level losses defined on magnitude spectra, phase spectra, short-time complex spectra, and time-domain waveforms are used to train the MP-SENet model jointly. Experimental results show that our proposed MP-SENet achieves a PESQ of 3.50 on the public VoiceBank+DEMAND dataset and outperforms existing advanced speech enhancement methods. | ['Zhen-Hua Ling', 'Yang Ai', 'Ye-Xin Lu'] | 2023-05-23 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 4.05442148e-01 7.77686685e-02 2.70356387e-01 -2.39601687e-01
-1.18829572e+00 -2.87208736e-01 1.02997638e-01 -3.07466775e-01
-3.17048013e-01 4.44125354e-01 4.15250242e-01 -1.74960971e-01
1.18677929e-01 -3.28538477e-01 -6.59473062e-01 -8.45663548e-01
-2.20471710e-01 -4.46790874e-01 -1.31613851e-01 -4.59933817e-01
-3.91789466e-01 -1.92589872e-02 -1.05139542e+00 1.68014809e-01
9.97126222e-01 1.18541789e+00 5.36250591e-01 1.07372355e+00
5.61220467e-01 9.26800668e-01 -9.51276481e-01 -6.70540810e-01
1.06624082e-01 -4.54619080e-01 -1.93909466e-01 -9.47217196e-02
-2.19207536e-02 -4.66829956e-01 -9.64683354e-01 1.51896954e+00
1.06562424e+00 2.53949910e-01 2.74871707e-01 -7.45960712e-01
-7.79567778e-01 1.03670275e+00 -2.77468413e-01 4.32008475e-01
9.48629305e-02 2.07516745e-01 9.79604900e-01 -1.11189747e+00
1.55819327e-01 1.07878590e+00 1.11330187e+00 3.74422997e-01
-1.16609335e+00 -7.83693790e-01 -2.18629897e-01 6.01209939e-01
-1.04968441e+00 -9.71033335e-01 1.12819767e+00 1.67864203e-01
1.57409322e+00 2.07589522e-01 4.65773255e-01 1.03189254e+00
8.94313026e-03 6.09897554e-01 1.03345144e+00 -5.02897620e-01
-4.68978798e-03 -7.10383579e-02 3.85317430e-02 4.55578923e-01
-5.36509871e-01 6.89080119e-01 -9.24759924e-01 1.79111481e-01
4.69617009e-01 -7.48065114e-01 -9.21133995e-01 4.13975656e-01
-8.11683059e-01 2.48639733e-01 4.27795500e-01 -1.52840856e-02
-7.47834206e-01 1.55945569e-01 5.87353468e-01 5.41818798e-01
6.70789063e-01 5.23635894e-02 -5.06642818e-01 -2.25624412e-01
-1.18438435e+00 -5.86113594e-02 5.75563490e-01 9.70650315e-01
2.34603792e-01 1.05476487e+00 -1.21333264e-01 1.07111025e+00
2.51116335e-01 9.92542386e-01 5.87071955e-01 -8.11777532e-01
6.80470407e-01 -5.39953411e-01 2.29388141e-04 -5.59494853e-01
-2.02819690e-01 -6.57008111e-01 -1.21434641e+00 -8.15154165e-02
-8.89280364e-02 -5.93527913e-01 -9.69578862e-01 1.89855719e+00
1.91102140e-02 3.42889339e-01 4.16174442e-01 1.15334547e+00
6.30325377e-01 1.07719731e+00 -9.27036777e-02 -4.89099085e-01
1.34879076e+00 -1.12274885e+00 -1.69789672e+00 -3.18231463e-01
-3.78490746e-01 -1.06884944e+00 6.69072270e-01 6.44625127e-01
-1.73898399e+00 -7.73406804e-01 -1.41259229e+00 -2.51336932e-01
-6.41762689e-02 4.93874103e-01 -1.40026599e-01 7.54441857e-01
-1.23670399e+00 1.08564007e+00 -7.84243286e-01 7.11679339e-01
1.93938866e-01 2.17124209e-01 -1.03296697e-01 3.97495985e-01
-1.67509651e+00 8.13426852e-01 3.30320001e-01 5.21375358e-01
-8.92219305e-01 -1.03623295e+00 -1.18358362e+00 5.02365470e-01
-1.76950201e-01 -3.38633209e-01 1.73440385e+00 -7.38020837e-01
-1.97550952e+00 3.79079998e-01 -1.84740618e-01 -8.58871996e-01
-1.26254335e-02 -3.13106537e-01 -1.30072129e+00 3.97088140e-01
-3.57075751e-01 2.06260279e-01 1.45203066e+00 -6.28248274e-01
-3.56367439e-01 9.50006545e-02 -6.20321155e-01 2.03430727e-01
-3.68916363e-01 1.64301246e-01 -1.51714817e-01 -1.22262347e+00
1.30721167e-01 -2.06825539e-01 -9.24268067e-02 -3.68843883e-01
-4.59218323e-01 3.25543612e-01 7.70276964e-01 -1.77110684e+00
1.48169661e+00 -2.40741086e+00 1.13162905e-01 -4.06512991e-02
2.76055150e-02 7.04679608e-01 -2.41198972e-01 2.49921024e-01
-5.14050424e-01 -2.79072702e-01 -5.09244144e-01 -5.44664681e-01
5.04753470e-01 -2.30775163e-01 -6.16804898e-01 4.44837153e-01
7.48211265e-01 3.99856359e-01 -8.11670005e-01 -6.55258000e-02
3.10506314e-01 1.16720903e+00 -5.49565017e-01 3.15770775e-01
4.84139509e-02 1.05147183e-01 4.56547141e-01 1.93970904e-01
1.07625258e+00 2.11888492e-01 2.86316484e-01 -6.48001969e-01
-2.10668087e-01 1.23317921e+00 -9.98295307e-01 1.64143479e+00
-5.63370466e-01 1.04536426e+00 8.45887005e-01 -6.91692412e-01
8.76629353e-01 1.09107769e+00 8.76904801e-02 -1.07413960e+00
1.24047227e-01 4.51759815e-01 1.99380834e-02 -3.68693948e-01
5.97908080e-01 -3.41667295e-01 3.40547204e-01 1.18109316e-01
6.16758704e-01 -2.90461242e-01 -1.37233302e-01 -2.06785440e-01
9.09348309e-01 5.77259399e-02 -3.94370146e-02 6.90178201e-02
6.58967853e-01 -5.24599731e-01 7.99211919e-01 7.56649971e-02
-3.71930957e-01 6.14739478e-01 1.36007950e-01 3.09735686e-01
-1.39985728e+00 -1.56303632e+00 -2.14704171e-01 1.01850379e+00
-1.20914966e-01 -3.68240178e-01 -1.07312095e+00 5.69161139e-02
-4.53713059e-01 6.14974022e-01 9.19268951e-02 -3.65201086e-01
-9.73089993e-01 -4.89241779e-01 8.26040447e-01 5.30640364e-01
7.16896772e-01 -8.60698998e-01 -1.07169524e-02 7.06208587e-01
-7.01337993e-01 -1.18318021e+00 -1.18665230e+00 7.07590878e-01
-5.53005636e-01 -5.76488853e-01 -8.44715178e-01 -1.08062780e+00
1.91880971e-01 1.87304299e-02 9.09451246e-01 -4.68198448e-01
4.02162187e-02 -1.65902197e-01 -2.15396702e-01 -3.27607244e-01
-5.99084079e-01 -2.59578049e-01 1.85678720e-01 1.08555533e-01
1.02188155e-01 -1.07305038e+00 -7.17324317e-01 1.29885972e-02
-6.75178826e-01 -1.34310750e-02 6.23228669e-01 1.14908695e+00
4.50789154e-01 1.46541148e-01 9.60159659e-01 -1.30518690e-01
9.27983165e-01 -1.78153455e-01 -6.88141584e-01 9.37392376e-03
-5.32485127e-01 8.21455475e-03 7.83372104e-01 -6.60895765e-01
-1.54269719e+00 -2.77024418e-01 -8.43287945e-01 -6.14341438e-01
4.63929951e-01 2.97552764e-01 -4.32244033e-01 1.68662906e-01
6.03596210e-01 4.93595302e-01 -1.66965619e-01 -7.84600079e-01
3.01832139e-01 1.18615305e+00 1.32366955e+00 -2.30263919e-01
9.06279981e-01 -6.58990592e-02 -6.67687297e-01 -8.74557912e-01
-6.22036338e-01 -3.84072632e-01 -3.54244560e-02 -7.17253387e-02
4.93683726e-01 -1.37883079e+00 -5.79880893e-01 9.21538949e-01
-1.44557393e+00 -4.37721342e-01 -4.81354177e-01 7.19323158e-01
-5.58734834e-01 4.63645428e-01 -1.40481114e+00 -8.16058874e-01
-1.13380194e+00 -1.01224041e+00 9.16618764e-01 2.65475780e-01
8.23372696e-03 -7.95473576e-01 -9.61922184e-02 7.85109922e-02
8.62127662e-01 -2.20141664e-01 5.40768206e-01 -9.31020156e-02
-3.62526149e-01 4.00060005e-02 4.10398431e-02 1.08288503e+00
-1.26946671e-02 -3.84316564e-01 -1.64251709e+00 -3.12870622e-01
6.32819474e-01 2.20837537e-02 6.92732453e-01 5.82886457e-01
9.56136405e-01 -6.13063753e-01 2.60585785e-01 1.05111039e+00
1.08631182e+00 3.25587243e-01 1.02684247e+00 -3.22344035e-01
8.79276469e-02 1.55722752e-01 2.07062349e-01 4.87089097e-01
1.23907231e-01 3.17381948e-01 -4.01039757e-02 -2.11433157e-01
-7.74380267e-01 -2.08440989e-01 9.18793142e-01 2.02870321e+00
2.29046389e-01 -2.70910531e-01 -2.83002883e-01 7.02425182e-01
-1.13967013e+00 -1.19879007e+00 -1.95423037e-01 1.76795363e+00
1.54625881e+00 1.14974290e-01 -1.94093257e-01 7.77925372e-01
7.46468723e-01 4.44147468e-01 -5.01766741e-01 -4.83147174e-01
-4.78098422e-01 1.15311480e+00 4.60698277e-01 6.21388018e-01
-1.04623950e+00 4.43533301e-01 6.06703234e+00 1.07153916e+00
-1.30899835e+00 4.86464113e-01 3.07431519e-01 -4.64767776e-02
-1.34234786e-01 -4.41658020e-01 -2.07562461e-01 3.37018251e-01
1.45827663e+00 -3.00307304e-01 9.04848993e-01 4.30645972e-01
4.28424478e-01 5.57617962e-01 -6.51764572e-01 9.15677190e-01
-3.39151114e-01 -9.98960495e-01 -7.44372070e-01 -5.42609334e-01
5.53863883e-01 1.24561436e-01 4.94585782e-01 4.23343182e-01
1.39168203e-01 -9.13959742e-01 1.21079326e+00 4.05111283e-01
1.41242552e+00 -6.76276922e-01 3.04517567e-01 2.19943210e-01
-1.54222107e+00 -2.55306754e-02 -1.30623773e-01 1.73821718e-01
4.16503012e-01 6.54700100e-01 -8.45222890e-01 5.43657005e-01
6.84551179e-01 5.53869963e-01 5.25825143e-01 9.07093704e-01
-6.76423132e-01 1.15149987e+00 -1.37025893e-01 4.28971887e-01
-8.56682137e-02 -8.66351873e-02 7.68265128e-01 1.61548245e+00
3.42241824e-01 1.77906558e-01 -6.02192819e-01 9.09685791e-01
-2.70676672e-01 -6.08083010e-01 2.26791039e-01 -9.80808735e-02
7.90524423e-01 1.15992022e+00 3.50787908e-01 -1.97728708e-01
-3.33396614e-01 1.03388011e+00 -8.72962028e-02 8.12644958e-01
-1.00406361e+00 -1.03777659e+00 9.77402925e-01 -3.49336267e-01
6.74911261e-01 -1.41320646e-01 -1.57987878e-01 -9.79976892e-01
1.63926914e-01 -1.12365723e+00 -2.73166776e-01 -1.14666736e+00
-1.13714647e+00 8.84714305e-01 -7.60976374e-01 -1.11854196e+00
-3.90696764e-01 -6.69737995e-01 -5.82728684e-01 1.53878248e+00
-2.09403777e+00 -7.08548367e-01 9.13718119e-02 5.57375073e-01
5.26691675e-01 -5.22883143e-03 9.50478375e-01 8.46257746e-01
-5.65742910e-01 7.56899774e-01 2.30919361e-01 2.93237805e-01
5.36151469e-01 -1.43895769e+00 9.28240895e-01 1.11596906e+00
-3.04962635e-01 3.43229860e-01 8.79276395e-01 -3.36942106e-01
-1.14791846e+00 -1.27109790e+00 1.03928077e+00 4.19424087e-01
5.44449568e-01 -4.76259112e-01 -1.14569259e+00 3.85918349e-01
8.84287775e-01 6.27260804e-02 3.44042152e-01 -5.20082295e-01
-4.48052526e-01 -3.78796697e-01 -1.12427866e+00 3.56546164e-01
7.18885660e-01 -1.25788498e+00 -6.62834108e-01 6.52340949e-02
1.16170609e+00 -1.00667357e+00 -1.09408069e+00 2.28215575e-01
6.39734343e-02 -7.53956199e-01 1.21081829e+00 3.02086258e-03
2.21545890e-01 -2.50698626e-01 -2.11107954e-01 -1.94783056e+00
-9.68357176e-02 -1.54871500e+00 -5.38767457e-01 1.24244249e+00
4.37934697e-01 -4.84341949e-01 2.60780454e-01 -1.61090612e-01
-9.19868410e-01 -3.75947684e-01 -9.13824379e-01 -8.32364380e-01
-1.16855517e-01 -6.13093436e-01 7.76937664e-01 6.31145000e-01
1.96107924e-01 2.08956167e-01 -5.31042635e-01 7.86219239e-01
7.99455941e-01 -5.83082378e-01 -1.78419292e-01 -3.79188806e-01
-5.94415486e-01 -3.72072548e-01 2.55465925e-01 -1.31148589e+00
-2.85385642e-02 -6.26503825e-01 6.58653021e-01 -1.05296564e+00
-4.52815652e-01 -3.59265506e-02 -3.54952544e-01 1.13847435e-01
-3.12619448e-01 -7.64217274e-03 4.73491997e-02 -3.67579520e-01
8.41566920e-02 8.89215708e-01 1.10640705e+00 -3.82427007e-01
1.75030977e-02 1.26268407e-02 -2.75148660e-01 6.71177387e-01
7.46741116e-01 -3.97626817e-01 -3.28930318e-01 -5.32020986e-01
-1.16384305e-01 6.07355893e-01 4.85309482e-01 -1.27601171e+00
5.61072052e-01 4.11446095e-01 1.31506175e-01 -6.88805759e-01
9.04779434e-01 -6.19231761e-01 2.19956815e-01 3.40798527e-01
-2.72991687e-01 -5.35553321e-02 4.16873127e-01 3.66816312e-01
-6.63670123e-01 6.28756210e-02 8.78782749e-01 3.77655208e-01
-4.28844720e-01 1.09209139e-02 -3.71096581e-01 2.50997692e-01
1.36934638e-01 1.85783431e-01 -4.51179296e-01 -4.06034142e-01
-8.57790053e-01 -1.11692607e-01 -4.67289448e-01 -9.00234878e-02
9.30089712e-01 -1.41099346e+00 -1.01539540e+00 4.32669014e-01
-6.07135236e-01 -1.83490276e-01 7.24824369e-01 8.16932261e-01
-2.81897008e-01 3.41195554e-01 9.72632319e-02 -1.55340359e-01
-1.03119159e+00 2.27485374e-01 1.00060725e+00 -2.41846144e-01
-6.12234294e-01 1.14035022e+00 -2.16577813e-01 -2.72085458e-01
6.06715679e-01 -5.41243613e-01 2.48726860e-01 -2.68588483e-01
7.24311769e-01 4.13609535e-01 4.92160141e-01 -5.55791855e-01
-9.13984329e-02 1.14516079e-01 2.75259525e-01 -5.02875507e-01
1.55569541e+00 -2.50506788e-01 7.76954070e-02 -8.49001706e-02
1.44516110e+00 1.33417949e-01 -1.57164693e+00 -5.88199735e-01
-2.28343874e-01 7.18168318e-02 4.66729313e-01 -1.13294017e+00
-1.32610035e+00 1.14977396e+00 6.45124912e-01 8.23393539e-02
2.01493382e+00 -7.21665561e-01 1.55882692e+00 -1.58821106e-01
-2.20151991e-01 -1.30115855e+00 2.38283783e-01 6.73084021e-01
9.96476054e-01 -7.24414945e-01 -5.18974125e-01 -4.43478823e-01
-2.82673091e-01 1.14218724e+00 3.98735739e-02 -4.45345836e-03
6.33972704e-01 9.69900489e-01 1.33212656e-01 1.23651743e-01
-8.27813804e-01 -1.71221495e-01 3.40812266e-01 9.86425936e-01
4.89828825e-01 2.83267088e-02 1.97292387e-01 1.17324674e+00
-7.20277846e-01 -2.02176720e-01 3.33004832e-01 4.34582710e-01
-4.34695482e-01 -8.95290077e-01 -7.19081700e-01 -1.02700084e-01
-5.56916475e-01 -5.63218474e-01 2.24444896e-01 2.02456266e-01
-2.16805097e-02 1.19323587e+00 3.09714198e-01 -5.14074087e-01
7.41966307e-01 1.84395611e-01 3.54956448e-01 7.94960856e-02
-1.02698410e+00 5.16351461e-01 4.67927754e-01 -1.54994860e-01
-4.86319028e-02 -4.33837712e-01 -1.37369978e+00 -4.26589280e-01
-5.57650387e-01 3.34604420e-02 8.83868456e-01 4.84447956e-01
1.75097823e-01 1.36052299e+00 9.06591535e-01 -7.46937811e-01
-1.21887290e+00 -1.21316910e+00 -6.30266905e-01 1.20197318e-01
1.24021602e+00 6.20504022e-02 -3.19877028e-01 5.15105546e-01] | [15.020979881286621, 6.002279281616211] |
6fe6d603-5f00-441b-b273-7e5dde596adb | read-reciprocal-attention-discriminator-for | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2135_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123590324.pdf | READ: Reciprocal Attention Discriminator for Image-to-Video Re-Identification | Person re-identification (re-ID) is the problem of visually identifying a person given a database of identities. In this work, we focus on image-to-video re-ID which compares a single query image to videos in the gallery. The main challenge is the asymmetry association of an image and a video, and overcoming the difference caused by the additional temporal dimension. To this end, we propose an attention-aware discriminator architecture. The attention occurs across different modalities, and even different identities to aggregate useful spatio-temporal information for comparison. The information is effectively fused into a united feature, followed by the final prediction of a similarity score. The performance of the method is shown with image-to-video person re-identification benchmarks (DukeMTMC-VideoReID, and MARS). | ['Hsuan-I Ho', 'Minho Shim', 'Dongyoon Wee', 'Jinhyung Kim'] | null | null | null | null | eccv-2020-8 | ['image-to-video-person-re-identification'] | ['computer-vision'] | [ 1.30844027e-01 -6.58883512e-01 2.22604685e-02 -2.37667456e-01
-8.03978264e-01 -6.35313213e-01 8.27526927e-01 -1.56120677e-02
-5.56519091e-01 4.43050027e-01 4.71908301e-01 4.29001749e-01
1.30001962e-01 -1.73384666e-01 -4.34713542e-01 -3.79925549e-01
1.09644800e-01 2.77423054e-01 -3.75693254e-02 2.01609120e-01
2.89405644e-01 3.69680226e-01 -1.89346743e+00 4.00746495e-01
5.07135093e-01 1.14228547e+00 1.35130450e-01 8.02266955e-01
3.43478471e-01 7.22911894e-01 -4.13353473e-01 -9.09571648e-01
2.92062193e-01 -4.98197317e-01 -8.07756960e-01 2.35318899e-01
1.11598158e+00 -6.71354592e-01 -6.17341101e-01 1.18982756e+00
7.34897554e-01 5.47003567e-01 6.25267327e-01 -1.60256588e+00
-7.64935493e-01 2.16631200e-02 -6.28861189e-01 6.27792358e-01
8.86516452e-01 1.10561781e-01 6.12378478e-01 -1.16181099e+00
6.26038790e-01 1.47843707e+00 6.73237979e-01 6.66255832e-01
-1.14239991e+00 -6.18356109e-01 3.83482426e-02 7.33085930e-01
-1.72326255e+00 -8.28335285e-01 4.42276180e-01 -7.54493058e-01
6.93877518e-01 3.05568010e-01 5.24934173e-01 1.21472704e+00
-1.85498789e-01 5.99701047e-01 7.65368462e-01 -1.63981289e-01
-1.62741691e-01 5.66118695e-02 4.19634394e-02 3.97381425e-01
1.75110027e-01 1.65607631e-01 -8.10712278e-01 -1.33234672e-02
5.34248531e-01 3.89730603e-01 -2.28650555e-01 -6.13448094e-04
-1.08936298e+00 3.57355386e-01 2.75684267e-01 2.48623028e-01
-1.99861482e-01 2.91822255e-02 4.87784863e-01 2.23112926e-01
2.48877555e-01 2.12379634e-01 2.84186512e-01 -1.46936536e-01
-8.47224534e-01 3.23766530e-01 2.09927484e-01 7.25444615e-01
6.11094594e-01 -2.43892580e-01 -6.52268112e-01 9.05305445e-01
7.03116832e-03 6.10327899e-01 7.77298748e-01 -8.87740910e-01
5.05488157e-01 3.44101846e-01 3.73345584e-01 -1.29336905e+00
-1.84653521e-01 -1.33893833e-01 -8.23256075e-01 9.87394974e-02
5.12595594e-01 2.82609817e-02 -6.89694583e-01 1.65379298e+00
1.96481943e-01 6.83953941e-01 -8.92070234e-02 1.25817084e+00
8.81006956e-01 2.63113499e-01 2.79679567e-01 8.65541920e-02
1.55397356e+00 -1.06056178e+00 -4.93564427e-01 -2.08829224e-01
1.81359529e-01 -5.64818859e-01 4.13150191e-01 -7.06936941e-02
-1.10468268e+00 -1.11707473e+00 -6.52911663e-01 -2.02164888e-01
-3.69957507e-01 3.24224144e-01 -1.02856703e-01 6.07377708e-01
-1.18472171e+00 5.83933115e-01 -2.44863093e-01 -7.13581204e-01
2.09097683e-01 3.18129867e-01 -9.12714303e-01 -2.00129583e-01
-1.05113220e+00 7.08257496e-01 4.21189249e-01 4.84673446e-03
-8.55161428e-01 -5.62356710e-01 -8.97266924e-01 3.57901183e-04
-4.80885841e-02 -8.07892501e-01 8.88623118e-01 -1.32295799e+00
-1.03758931e+00 1.30792201e+00 -5.97961128e-01 -5.67263663e-01
8.00809085e-01 -1.62662983e-01 -8.55778575e-01 2.71595716e-01
4.63060886e-01 7.20371068e-01 1.04425025e+00 -1.24861670e+00
-1.07817447e+00 -6.07963443e-01 2.00760569e-02 5.21567822e-01
-4.74771827e-01 4.06073362e-01 -8.88084948e-01 -8.02452266e-01
-2.59664774e-01 -1.01429546e+00 2.75660902e-01 -2.74246693e-01
-4.56652164e-01 -2.37007648e-01 5.99670410e-01 -1.29668272e+00
1.07455218e+00 -2.27769327e+00 2.81264901e-01 8.57350007e-02
3.71663243e-01 1.86383367e-01 -3.50588381e-01 2.39260972e-01
-3.37565571e-01 1.43569414e-04 3.42897415e-01 -7.81736791e-01
-3.06786895e-01 -5.31235099e-01 -4.70977537e-02 5.94760895e-01
-2.81229895e-02 9.50146675e-01 -9.99102235e-01 -6.45365059e-01
2.10672095e-01 3.31537098e-01 -2.20183909e-01 4.23573136e-01
6.12647653e-01 6.57701731e-01 -1.51193216e-01 6.63260758e-01
7.00814843e-01 -1.90204486e-01 -1.42490687e-02 -5.87186933e-01
3.79367992e-02 -3.84897560e-01 -1.17302704e+00 1.52937698e+00
-1.10347100e-01 7.84347713e-01 -2.22005561e-01 -6.58833742e-01
6.60707295e-01 2.28462756e-01 6.59759521e-01 -8.47383082e-01
-6.32819459e-02 -8.77382420e-03 -2.46598199e-01 -5.08840322e-01
8.03203106e-01 4.79077637e-01 -7.61766545e-03 5.50688386e-01
3.25405337e-02 8.07868958e-01 2.40926579e-01 1.41275719e-01
6.56088650e-01 -1.85212970e-01 1.19669951e-01 3.16998772e-02
8.45086336e-01 -2.98927873e-01 4.07318383e-01 8.57261121e-01
-5.10124862e-01 8.00307393e-01 -6.64419979e-02 -5.19167662e-01
-1.26622415e+00 -1.07508111e+00 4.72547002e-02 1.20163882e+00
6.01321220e-01 -4.10649270e-01 -7.21680045e-01 -6.10707939e-01
3.46793532e-01 1.47138387e-01 -8.76159489e-01 -7.30374604e-02
-3.93750340e-01 -3.68474782e-01 6.07495904e-01 6.09800398e-01
7.97749400e-01 -5.72997749e-01 -1.61771834e-01 -1.14487253e-01
-6.72441602e-01 -1.33854640e+00 -1.24737120e+00 -8.00316989e-01
-2.75797993e-01 -1.02416801e+00 -1.28648865e+00 -8.09164107e-01
5.72871447e-01 7.90510476e-01 8.86005223e-01 -4.27075587e-02
-3.12472939e-01 8.25820565e-01 -2.08618045e-01 1.49554625e-01
-2.40521990e-02 -3.31445038e-01 5.74149311e-01 9.29880381e-01
4.43631887e-01 -1.25654355e-01 -1.00417984e+00 6.08352780e-01
-4.39133018e-01 1.88400839e-02 1.77212954e-01 1.02657211e+00
4.20117259e-01 -1.79632068e-01 4.81468350e-01 -3.05855311e-02
4.53690976e-01 -3.79306644e-01 -1.11412980e-01 6.41086578e-01
-1.62264854e-01 -3.41280550e-01 2.92264253e-01 -5.94370782e-01
-1.07567692e+00 1.78516313e-01 3.65335345e-01 -7.18731046e-01
-6.79923967e-02 -2.06897948e-02 -1.96700931e-01 -2.91699439e-01
3.52827162e-01 3.41760159e-01 1.33212805e-02 -4.68595862e-01
2.91824341e-01 9.75775182e-01 1.06364298e+00 -2.80410349e-01
6.77744269e-01 6.22263730e-01 -3.02883297e-01 -7.31158912e-01
-3.90084952e-01 -8.13787341e-01 -7.94014931e-01 -5.91292202e-01
1.12155652e+00 -1.30417049e+00 -9.62285876e-01 7.62518466e-01
-1.12158442e+00 1.03081807e-01 -1.12852201e-01 4.92223680e-01
-3.15691471e-01 7.74310052e-01 -4.94200110e-01 -8.16020608e-01
-2.66923994e-01 -9.30133164e-01 1.13936353e+00 5.90159357e-01
-1.52600795e-01 -6.16572380e-01 2.53047589e-02 5.50497770e-01
1.52502164e-01 3.40346843e-02 9.03447866e-02 -6.96417272e-01
-4.55024987e-01 -4.58038896e-01 -7.29078591e-01 1.33185819e-01
2.06344575e-01 -3.88197601e-01 -1.10997009e+00 -6.23862028e-01
-4.04053003e-01 -2.04597756e-01 9.76149023e-01 1.70863017e-01
1.15175056e+00 -3.00450772e-01 -5.51790655e-01 7.92185545e-01
1.16909170e+00 1.43405885e-01 5.49802244e-01 3.67965013e-01
9.96705234e-01 6.44304276e-01 5.02183616e-01 4.86671925e-01
6.00698531e-01 1.27294731e+00 8.17833170e-02 -5.42473719e-02
-3.84651601e-01 -3.41919214e-01 3.77086997e-01 1.75701708e-01
-4.54987884e-01 -3.03561896e-01 -7.77719021e-01 6.10961199e-01
-1.96147645e+00 -1.56932855e+00 1.19170912e-01 2.42813587e+00
1.30280703e-01 -6.47433341e-01 5.71572959e-01 -1.82631671e-01
1.42715454e+00 -2.06792094e-02 -6.11323297e-01 2.67586827e-01
-4.29617167e-01 -5.31828523e-01 5.50834000e-01 2.52385587e-01
-1.37464166e+00 6.61519706e-01 6.27859402e+00 8.53209317e-01
-8.62157941e-01 1.33085147e-01 7.90662885e-01 -1.57990828e-01
1.78168356e-01 -4.65039402e-01 -8.64131808e-01 8.76715302e-01
7.54170775e-01 -5.84544182e-01 5.79103231e-01 5.45228601e-01
1.57022923e-01 -1.34427369e-01 -1.37156820e+00 1.78556955e+00
6.90965056e-01 -1.12427413e+00 1.94907814e-01 -2.95464527e-02
6.21397376e-01 -3.66492867e-01 3.43600065e-01 -1.32744312e-01
3.14483345e-02 -1.13836396e+00 7.83272266e-01 8.29148173e-01
1.10647726e+00 -6.29174769e-01 6.33113563e-01 -3.24447691e-01
-1.58165503e+00 -4.79755312e-01 -1.83353350e-01 2.05119789e-01
2.62804121e-01 -1.89966694e-01 -6.25876665e-01 6.40953541e-01
1.09563839e+00 1.21042502e+00 -9.85011339e-01 1.18460989e+00
5.40573001e-01 -2.36102775e-01 5.88814467e-02 4.76029992e-01
-3.41374010e-01 -1.40170297e-02 5.47979712e-01 1.35950363e+00
6.23455763e-01 8.51208437e-03 2.60669023e-01 5.32507241e-01
-5.13127372e-02 -7.39692897e-02 -5.12174666e-01 2.09739640e-01
4.42851186e-01 1.03461182e+00 -2.54972935e-01 -4.27736104e-01
-5.32529294e-01 1.68475568e+00 3.41541439e-01 5.51008463e-01
-7.11943030e-01 3.66806872e-02 9.76116836e-01 3.01890839e-02
1.41099781e-01 7.86382034e-02 2.38843039e-01 -1.32279468e+00
7.42254928e-02 -7.46050417e-01 8.17028701e-01 -9.07217443e-01
-1.68599737e+00 6.39453351e-01 -6.87153637e-02 -1.47529769e+00
-3.76372188e-01 -3.94413829e-01 -4.63283032e-01 1.10196459e+00
-1.14209878e+00 -1.33166003e+00 -8.76544952e-01 1.01482618e+00
5.20021915e-01 -7.25575626e-01 5.50032258e-01 7.06375957e-01
-8.45150113e-01 1.18472278e+00 2.94929117e-01 6.62394345e-01
1.09898591e+00 -9.86518323e-01 5.77752292e-01 1.05825043e+00
-5.68237389e-03 3.98030400e-01 3.62969965e-01 -8.06479692e-01
-1.18367386e+00 -1.24180925e+00 8.70548546e-01 -6.45705640e-01
3.84055078e-01 -1.29586682e-01 -6.78165317e-01 5.07768810e-01
-2.06191856e-02 1.17666826e-01 6.85246050e-01 -3.40004653e-01
-5.52238345e-01 -2.45213687e-01 -1.03637326e+00 5.35077274e-01
1.33206582e+00 -1.03634012e+00 -2.38801867e-01 1.16848506e-01
2.25313649e-01 -2.04297185e-01 -8.85873199e-01 1.13945212e-02
9.67328429e-01 -9.70895469e-01 1.50097930e+00 -7.70666182e-01
1.93959534e-01 -3.80745441e-01 -1.69505149e-01 -1.11752892e+00
-5.54307938e-01 -4.82030183e-01 9.45774838e-02 1.53248537e+00
-2.20916778e-01 -4.07701641e-01 5.96561968e-01 8.94699812e-01
4.85212833e-01 1.07507214e-01 -1.06296098e+00 -9.22813952e-01
-5.53868711e-01 7.28571266e-02 6.21460140e-01 9.51986432e-01
-6.81805387e-02 7.19590411e-02 -9.76161659e-01 1.79023162e-01
8.75780642e-01 6.39297301e-03 9.55410004e-01 -1.10181451e+00
-2.49239802e-01 -3.69547695e-01 -9.37903643e-01 -8.88694644e-01
1.63823962e-01 -8.23240578e-01 -2.81825662e-01 -1.14370453e+00
7.83358335e-01 -7.07964897e-02 -3.67296845e-01 8.04300234e-02
-4.79037404e-01 6.21326208e-01 5.85400939e-01 7.10563540e-01
-8.35054457e-01 3.31429720e-01 7.63998568e-01 -4.88741726e-01
-5.39600775e-02 -7.45905787e-02 -4.85841513e-01 4.25395817e-01
4.66646254e-01 -6.99978918e-02 -8.16339403e-02 -4.71027106e-01
-4.56791937e-01 1.10338069e-01 1.06567717e+00 -1.24393153e+00
4.53941941e-01 7.05612898e-02 8.51789534e-01 -3.76222849e-01
6.33555055e-01 -5.41182518e-01 5.08948624e-01 2.97117859e-01
-4.14326638e-01 5.67523539e-01 -4.87680323e-02 8.75997245e-01
-3.99967760e-01 -3.58293615e-02 6.68895543e-01 -2.94868983e-02
-1.28557169e+00 5.60049772e-01 5.36746532e-02 4.29134779e-02
1.17094469e+00 -5.29818118e-01 -4.63618875e-01 -5.80377281e-01
-8.39113355e-01 2.52542675e-01 8.13786387e-01 7.05600739e-01
7.28316545e-01 -1.76831543e+00 -1.03772044e+00 1.42314821e-01
5.13913751e-01 -9.19698000e-01 9.20019209e-01 6.61271930e-01
-1.73354134e-01 3.44898373e-01 -6.65099800e-01 -6.24853849e-01
-1.73724341e+00 8.26047003e-01 6.95612550e-01 1.01892427e-01
-4.82243598e-01 6.61835551e-01 4.99114066e-01 3.42468858e-01
2.60266662e-01 4.96585101e-01 -5.88116109e-01 3.28720570e-01
1.09022248e+00 6.16245925e-01 -2.45542616e-01 -1.47977483e+00
-6.06057703e-01 8.03839505e-01 -5.57019897e-02 -2.57137507e-01
6.63051605e-01 -6.76781595e-01 2.06084654e-01 9.23016742e-02
1.34572554e+00 -1.68965071e-01 -1.30380547e+00 -4.88890946e-01
-2.01420754e-01 -1.00421774e+00 -3.54042411e-01 -4.62507844e-01
-9.91706550e-01 6.32320762e-01 1.28188753e+00 2.03621816e-02
9.08611417e-01 -9.23981965e-02 8.48818541e-01 4.15853411e-02
2.56865352e-01 -1.18133199e+00 3.30915928e-01 1.21286392e-01
8.31561685e-01 -1.63569200e+00 -1.39472649e-01 -5.04051074e-02
-9.28038836e-01 8.41580629e-01 6.33138597e-01 1.39501095e-01
3.97362649e-01 -3.61650318e-01 -1.50407016e-01 4.03352380e-01
-9.21828002e-02 -7.14805067e-01 6.40441895e-01 1.08947587e+00
1.75995976e-02 -3.17596234e-02 1.73280835e-01 6.25412881e-01
1.81504428e-01 -6.61252439e-02 1.12698488e-01 2.79751420e-01
-7.07981410e-04 -8.78671587e-01 -6.26748502e-01 2.73187667e-01
-4.49294537e-01 2.35946733e-03 -4.22962278e-01 3.65191877e-01
2.57399023e-01 1.17577851e+00 4.82208610e-01 -7.54256546e-01
1.14618979e-01 -1.93083346e-01 3.37163568e-01 -1.88528851e-01
-5.91420889e-01 -3.02375704e-01 8.42671618e-02 -6.53813362e-01
-5.41665316e-01 -9.41002429e-01 -4.71411258e-01 -4.84670311e-01
2.96953619e-01 -4.53733355e-02 2.66771793e-01 5.86094201e-01
6.55816913e-01 1.40949279e-01 6.93815351e-01 -1.22877467e+00
-2.50055671e-01 -6.33246839e-01 -5.26057005e-01 1.20012569e+00
6.01907194e-01 -8.05131316e-01 -1.62977979e-01 4.34740990e-01] | [14.61725902557373, 0.9784344434738159] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.