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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
c642471f-9118-4fcb-995f-ac6de101941f | the-mvtec-3d-ad-dataset-for-unsupervised-3d | 2112.09045 | null | https://arxiv.org/abs/2112.09045v1 | https://arxiv.org/pdf/2112.09045v1.pdf | The MVTec 3D-AD Dataset for Unsupervised 3D Anomaly Detection and Localization | We introduce the first comprehensive 3D dataset for the task of unsupervised anomaly detection and localization. It is inspired by real-world visual inspection scenarios in which a model has to detect various types of defects on manufactured products, even if it is trained only on anomaly-free data. There are defects that manifest themselves as anomalies in the geometric structure of an object. These cause significant deviations in a 3D representation of the data. We employed a high-resolution industrial 3D sensor to acquire depth scans of 10 different object categories. For all object categories, we present a training and validation set, each of which solely consists of scans of anomaly-free samples. The corresponding test sets contain samples showing various defects such as scratches, dents, holes, contaminations, or deformations. Precise ground-truth annotations are provided for every anomalous test sample. An initial benchmark of 3D anomaly detection methods on our dataset indicates a considerable room for improvement. | ['Carsten Steger', 'David Sattlegger', 'Xin Jin', 'Paul Bergmann'] | 2021-12-16 | null | null | null | null | ['rgb-3d-anomaly-detection-and-segmentation', '3d-anomaly-detection-and-segmentation', 'depth-anomaly-detection-and-segmentation', 'rgb-depth-anomaly-detection-and-segmentation'] | ['methodology', 'methodology', 'methodology', 'methodology'] | [ 5.25125802e-01 3.03019613e-01 6.49573863e-01 -3.74965280e-01
-3.81286860e-01 -3.92000675e-01 3.75439584e-01 6.37115836e-01
4.31958467e-01 -1.42013833e-01 -4.94549632e-01 -1.25360116e-01
-1.10937409e-01 -6.08869255e-01 -6.12874925e-01 -6.46815896e-01
-3.34939867e-01 6.07932329e-01 8.71764719e-01 -1.72669426e-01
5.60136855e-01 9.60206926e-01 -2.03226686e+00 3.05539697e-01
4.55115676e-01 1.28347588e+00 -1.83777884e-02 5.06919265e-01
-1.64778531e-01 4.56083305e-02 -7.45465875e-01 -4.23031412e-02
5.02222180e-01 -8.71542469e-02 -5.72903037e-01 8.98834407e-01
7.20048666e-01 -3.02283019e-01 1.16064467e-01 1.25505817e+00
5.71634509e-02 -2.65047133e-01 9.36816156e-01 -1.34754515e+00
-3.33365291e-01 -4.38422382e-01 -7.90578067e-01 8.04485977e-02
6.14484012e-01 2.45038390e-01 6.82836473e-01 -9.45716143e-01
7.55505860e-01 1.13251066e+00 5.85076094e-01 4.56309289e-01
-1.15622246e+00 1.17077425e-01 4.00971435e-02 1.67070571e-02
-7.72141159e-01 -1.38854653e-01 1.22111583e+00 -8.90977979e-01
8.70903552e-01 1.07768446e-01 5.00826538e-01 8.94406021e-01
5.73455274e-01 6.45439744e-01 5.45636475e-01 -4.73725289e-01
5.86529672e-01 -2.60915458e-01 1.99625090e-01 7.08294094e-01
5.31690300e-01 3.36866044e-02 -2.58007646e-01 -2.55773723e-01
7.28509009e-01 3.13587099e-01 4.83478382e-02 -1.05080485e+00
-9.62551296e-01 5.30795932e-01 2.29115054e-01 2.07314789e-02
-6.87184155e-01 -4.24325734e-01 5.52449644e-01 6.22614920e-01
6.00491464e-01 5.24013638e-01 -7.23650098e-01 2.05028951e-01
-2.93675303e-01 1.52779952e-01 4.23067272e-01 6.64479077e-01
8.61572564e-01 7.87243843e-02 3.66751134e-01 8.34055960e-01
3.43240768e-01 3.65638733e-01 1.95273727e-01 -7.65008569e-01
1.84711039e-01 1.20932484e+00 -7.68755451e-02 -1.02775156e+00
-3.00325990e-01 7.23448321e-02 -5.79363763e-01 9.35306609e-01
4.42011029e-01 4.22143131e-01 -1.41230011e+00 7.54863918e-01
5.83756149e-01 -1.36391714e-01 -2.56521851e-01 6.86402559e-01
5.19759059e-01 9.60980132e-02 -6.51597023e-01 1.19331636e-01
1.00160921e+00 -3.93488318e-01 -4.88306612e-01 -2.16021076e-01
7.06825376e-01 -6.59334540e-01 1.15124714e+00 9.08544481e-01
-8.43689680e-01 -3.93381476e-01 -1.10383439e+00 5.14120042e-01
-5.69166064e-01 -2.73454219e-01 2.19771653e-01 3.38309079e-01
-4.48473334e-01 6.59621477e-01 -1.02997017e+00 -5.14185131e-01
5.84326982e-01 -3.47585790e-02 -8.21768463e-01 -3.56429011e-01
-2.61148363e-01 6.69849813e-01 9.85658467e-02 2.43044257e-01
-1.12649035e+00 -5.20256221e-01 -1.25173390e+00 -7.37119615e-01
4.73054796e-01 2.94047952e-01 9.92583573e-01 -5.45405030e-01
-9.70527112e-01 1.48932660e+00 1.95073411e-01 -4.63665463e-02
4.27039564e-01 -4.07730460e-01 -6.56105995e-01 6.26640692e-02
2.41205603e-01 -1.23552114e-01 1.05952263e+00 -1.63127303e+00
-4.98890400e-01 -7.99526572e-01 -2.16228709e-01 -6.51475787e-01
2.68195689e-01 -1.03583485e-01 -2.13492304e-01 -4.79661167e-01
1.08470857e+00 -5.86018622e-01 -3.77290189e-01 1.94855094e-01
-9.67706859e-01 -9.32888016e-02 1.41678989e+00 -4.66159403e-01
6.27958715e-01 -2.30293775e+00 -2.95191169e-01 6.02387547e-01
1.11513563e-01 5.69708347e-02 -1.26001403e-01 3.85989964e-01
-2.51090795e-01 -8.56034681e-02 -7.95866787e-01 -1.51978120e-01
-8.22673440e-02 5.56871653e-01 -1.39929771e-01 8.57995033e-01
6.95901394e-01 2.35885784e-01 -8.05919647e-01 -1.66611403e-01
4.09100413e-01 -1.13710605e-01 -4.93970245e-01 2.79083908e-01
-3.18758458e-01 4.83637363e-01 -5.44812560e-01 1.54921699e+00
7.83028483e-01 9.65052173e-02 -6.66921556e-01 4.21723016e-02
5.03318720e-02 1.34387556e-02 -1.20395315e+00 1.72892070e+00
-7.21445680e-02 4.59809810e-01 1.69040143e-01 -1.15519977e+00
1.30192065e+00 1.48613647e-01 5.42057991e-01 -5.57951629e-01
-1.27282396e-01 4.26031113e-01 -6.18231222e-02 -8.24658036e-01
2.69735724e-01 1.93683207e-01 -8.00846368e-02 3.82343352e-01
-7.60249794e-03 -6.23123586e-01 1.35194421e-01 -6.82996884e-02
1.64985168e+00 -2.77462378e-02 3.37246768e-02 -1.87772930e-01
4.22257751e-01 1.32748798e-01 4.71157253e-01 5.66398084e-01
-1.30761370e-01 1.02158916e+00 7.28695393e-01 -7.51827836e-01
-1.24306381e+00 -1.51060593e+00 -3.78441274e-01 3.08639616e-01
3.20563704e-01 -7.68605992e-02 -3.46080482e-01 -1.14922857e+00
3.68562073e-01 4.54600066e-01 -7.68026054e-01 -2.42949173e-01
-4.62833703e-01 -4.81825799e-01 -7.64388666e-02 2.53221840e-01
1.03997193e-01 -1.23811185e+00 -9.70181167e-01 -5.07747717e-02
5.24367630e-01 -9.41915035e-01 2.25188255e-01 3.63009661e-01
-1.28754294e+00 -1.75484240e+00 -6.90730587e-02 -6.68437243e-01
1.32477844e+00 -3.09152547e-02 1.35552466e+00 4.38191742e-01
-1.03353882e+00 7.52038062e-01 -5.28532386e-01 -7.70024836e-01
-8.35413396e-01 -6.77765369e-01 1.13291301e-01 -4.77594957e-02
3.61081839e-01 -4.74118114e-01 -2.78765857e-01 6.34901643e-01
-1.01864338e+00 -8.29376280e-01 4.24658239e-01 8.05931032e-01
9.35910761e-01 4.26384002e-01 1.61981523e-01 -7.64566958e-01
3.81278276e-01 -5.27534068e-01 -5.91816604e-01 -1.99287415e-01
-3.89246494e-01 -1.29294828e-01 1.79387569e-01 -2.59113789e-01
-7.29608595e-01 8.23199302e-02 -1.60877347e-01 -5.47605813e-01
-1.18344009e+00 2.21353278e-01 -3.64910156e-01 2.38661915e-01
8.31882119e-01 -1.02660596e-01 1.61161527e-01 -1.01004446e+00
-2.30969310e-01 4.10767257e-01 9.13949013e-01 -4.20566708e-01
1.12215686e+00 5.67070961e-01 1.86594650e-01 -1.32152474e+00
-6.95449829e-01 -7.04198420e-01 -1.10576487e+00 -2.79071838e-01
5.46835124e-01 -2.96289831e-01 -1.15212850e-01 8.69156599e-01
-1.00933719e+00 -1.93672061e-01 -8.93994987e-01 1.14400297e-01
-4.98003006e-01 5.02194107e-01 -3.97271186e-01 -8.53732586e-01
2.52698720e-01 -9.48602498e-01 1.46037555e+00 -2.61818498e-01
-9.72496122e-02 -8.87820601e-01 4.31922674e-02 -5.52171208e-02
-6.02780990e-02 9.02115285e-01 1.21850622e+00 -9.11288798e-01
-2.00633079e-01 -7.79772580e-01 2.99647659e-01 8.67800832e-01
6.81235909e-01 3.57696503e-01 -1.01739264e+00 -2.01346681e-01
1.90726131e-01 -2.81354606e-01 5.84654093e-01 1.10593744e-01
1.37542593e+00 3.62994194e-01 -3.98989409e-01 1.63447130e-02
1.22685325e+00 3.17852914e-01 7.06317425e-01 3.07754844e-01
6.98998213e-01 8.39582503e-01 1.03470802e+00 4.57944274e-01
-3.29461098e-01 4.94674593e-01 1.37527871e+00 -1.13005586e-01
2.67729551e-01 3.42967361e-02 2.06183210e-01 3.53881508e-01
-6.10257089e-02 1.57794803e-02 -1.18551469e+00 8.79634380e-01
-1.27166343e+00 -6.30059719e-01 -6.02128685e-01 2.33566785e+00
1.50223881e-01 5.69096982e-01 -1.66159049e-02 9.52300191e-01
6.21291518e-01 -1.63649574e-01 -9.18969691e-01 -3.70933503e-01
3.89698036e-02 2.19171703e-01 1.50126860e-01 7.98930079e-02
-1.28080666e+00 3.08439463e-01 6.72209883e+00 4.07813676e-02
-7.53060162e-01 -5.09950042e-01 2.06815183e-01 2.79013067e-01
-1.30952343e-01 -3.98891866e-01 -2.25056842e-01 2.87408441e-01
4.33015198e-01 5.59172630e-01 -3.90616447e-01 1.02982318e+00
-1.10900134e-01 -4.53773379e-01 -1.29648364e+00 7.35773981e-01
2.72424947e-02 -7.71685898e-01 -1.76927552e-01 2.28376240e-01
6.68508589e-01 1.44280046e-02 -2.08038509e-01 -1.52865320e-01
1.48136526e-01 -7.11816669e-01 6.26751363e-01 4.81127650e-01
4.55854952e-01 -5.24159253e-01 9.75891292e-01 2.20992431e-01
-6.80570245e-01 -8.22374299e-02 -3.51268858e-01 6.63891435e-02
1.94639564e-01 1.15353382e+00 -7.63672531e-01 4.80422407e-01
1.03676498e+00 7.92799234e-01 -8.40117931e-01 1.17904484e+00
-2.51182884e-01 4.65147406e-01 -4.18666780e-01 7.00366199e-01
4.23617437e-02 -1.30764201e-01 8.56118023e-01 6.23595715e-01
5.51605046e-01 -4.70949858e-01 1.97069705e-01 8.18723083e-01
5.00344157e-01 -3.40271890e-01 -1.22505546e+00 2.68005908e-01
1.58087790e-01 9.60007191e-01 -8.96664262e-01 2.59540588e-01
-4.48450238e-01 1.07578301e+00 -2.87180156e-01 6.91607296e-02
-2.91844130e-01 -4.35730428e-01 1.02916896e+00 5.43430567e-01
4.30303574e-01 -3.30963880e-01 -1.65980548e-01 -6.83532536e-01
3.18924844e-01 -5.64442396e-01 2.67375082e-01 -7.38924384e-01
-1.80643010e+00 2.18721062e-01 -6.96861790e-03 -1.48735666e+00
-4.06919807e-01 -1.16173470e+00 -9.39989448e-01 3.55316013e-01
-1.03050661e+00 -6.99343503e-01 -6.16078436e-01 5.63735962e-01
7.37213790e-01 -3.98102909e-01 8.23843002e-01 5.81080327e-03
-4.16991055e-01 -2.26078276e-03 -1.36025578e-01 8.90457928e-02
2.82045245e-01 -1.52400911e+00 8.15603912e-01 1.07237947e+00
1.04237288e-01 -5.65295853e-02 9.70915437e-01 -7.53671467e-01
-1.18497992e+00 -1.13591635e+00 1.94252938e-01 -8.39740574e-01
5.30653238e-01 -4.41259623e-01 -1.31216121e+00 5.88688493e-01
-4.85058010e-01 6.53236926e-01 2.95096010e-01 -3.20154220e-01
-1.23378776e-01 1.90185785e-01 -1.64283872e+00 4.18548174e-02
1.20413995e+00 -3.72932971e-01 -7.79744327e-01 2.58385390e-01
3.92054319e-01 -5.62702894e-01 -7.24059880e-01 6.76367640e-01
1.13640174e-01 -1.34241092e+00 9.18117285e-01 -6.09346330e-01
5.41664958e-01 -4.70383346e-01 -2.18644112e-01 -1.39999628e+00
1.14343986e-01 -1.46761294e-02 -3.46289694e-01 9.12140608e-01
9.67176557e-02 -6.15340471e-01 9.03435111e-01 1.72187597e-01
-7.98160315e-01 -5.95544577e-01 -8.44316423e-01 -1.02313519e+00
-2.51033753e-01 -8.21585596e-01 5.31663120e-01 8.06678653e-01
-4.82617617e-01 -2.78906077e-01 4.28169847e-01 6.34853184e-01
8.99143755e-01 4.19537425e-02 8.06719244e-01 -2.05909562e+00
2.18856886e-01 -7.79717639e-02 -1.24731302e+00 -6.05601907e-01
-1.19327739e-01 -4.16918993e-01 3.73361975e-01 -1.44605553e+00
-6.15404785e-01 -5.05782127e-01 -1.84610635e-01 5.22361338e-01
3.37796450e-01 4.49224591e-01 -6.59922183e-01 7.40436837e-02
-3.31647366e-01 2.84729809e-01 1.15815473e+00 -2.90483415e-01
-1.16948396e-01 4.34054703e-01 -8.73206481e-02 1.10609913e+00
7.73656547e-01 -3.47841442e-01 -4.22169924e-01 -2.15466812e-01
1.56332195e-01 -3.84493113e-01 6.67073309e-01 -1.16435468e+00
-3.29274893e-01 3.55531611e-02 2.86821783e-01 -1.09630334e+00
1.43151715e-01 -1.22103596e+00 -2.10137248e-01 5.17630219e-01
2.31842458e-01 7.34134018e-02 2.01953888e-01 6.65501654e-01
-4.08337116e-01 -3.72895211e-01 7.77993321e-01 -2.85999447e-01
-9.23803747e-01 4.06966329e-01 -3.20898175e-01 -2.12182909e-01
1.29915309e+00 -7.08133280e-01 -9.19813216e-02 1.99282616e-01
-9.07633603e-01 -5.55865765e-02 9.25623715e-01 6.11740947e-01
1.05113780e+00 -1.28021967e+00 -5.24013638e-01 1.06309521e+00
6.93580985e-01 7.78256714e-01 2.03631684e-01 4.59591657e-01
-5.52852035e-01 -1.05336949e-01 -4.20541584e-01 -1.24809730e+00
-1.20646477e+00 4.61582899e-01 4.16849405e-01 2.16065884e-01
-9.33477998e-01 6.87448084e-01 -2.05763686e-03 -5.32182813e-01
1.04252577e-01 -7.75351048e-01 9.00176470e-04 -2.65756965e-01
3.73888046e-01 4.91677463e-01 7.21424818e-01 -4.35621887e-01
-4.53684419e-01 7.07136750e-01 9.12673771e-02 5.00592172e-01
1.41500640e+00 1.31307498e-01 -1.51924536e-01 7.87317991e-01
8.49491000e-01 2.98445113e-02 -1.32690418e+00 -1.52312428e-01
4.28063273e-01 -7.83397198e-01 -2.20789820e-01 -5.83971918e-01
-1.26682603e+00 9.31020260e-01 8.76934886e-01 8.11819077e-01
1.03369927e+00 5.98965824e-01 3.83020014e-01 4.00349617e-01
6.54787242e-01 -9.12436366e-01 5.08357465e-01 5.00818431e-01
1.07829285e+00 -1.54565144e+00 4.38332036e-02 -4.95908827e-01
-2.61724919e-01 1.11072290e+00 7.33345449e-01 -3.49840909e-01
7.38907814e-01 2.28270352e-01 1.08901434e-01 -1.05223942e+00
-3.93338710e-01 3.57226022e-02 2.40884244e-01 1.17046750e+00
-3.24558951e-02 -2.83763468e-01 5.07532537e-01 2.14732900e-01
-6.18455699e-03 -7.24191844e-01 6.25007987e-01 1.31075203e+00
-4.82970178e-01 -8.44762325e-01 -5.84310234e-01 6.95731819e-01
-2.44869933e-01 7.20225692e-01 -3.93386930e-01 9.29842889e-01
3.01600754e-01 7.26982892e-01 7.93311238e-01 -3.34647775e-01
9.95904982e-01 2.98245829e-02 4.50770229e-01 -7.94450402e-01
-7.11168647e-02 -3.15669358e-01 -1.16487369e-01 -9.71229434e-01
-1.65858984e-01 -8.71037662e-01 -1.23624718e+00 2.80361801e-01
-3.06063443e-01 -1.93154693e-01 8.68661940e-01 8.09344590e-01
6.14213608e-02 5.97814798e-01 8.74416232e-01 -9.78129387e-01
-2.45657414e-01 -9.29253936e-01 -1.21019793e+00 8.83812726e-01
7.14482129e-01 -1.07648337e+00 -7.59682059e-01 1.94718480e-01] | [7.524821758270264, 2.030536651611328] |
d15d0c4a-16df-4a7c-a83c-e578d1520d44 | unified-optimal-transport-framework-for | 2210.17067 | null | https://arxiv.org/abs/2210.17067v2 | https://arxiv.org/pdf/2210.17067v2.pdf | Unified Optimal Transport Framework for Universal Domain Adaptation | Universal Domain Adaptation (UniDA) aims to transfer knowledge from a source domain to a target domain without any constraints on label sets. Since both domains may hold private classes, identifying target common samples for domain alignment is an essential issue in UniDA. Most existing methods require manually specified or hand-tuned threshold values to detect common samples thus they are hard to extend to more realistic UniDA because of the diverse ratios of common classes. Moreover, they cannot recognize different categories among target-private samples as these private samples are treated as a whole. In this paper, we propose to use Optimal Transport (OT) to handle these issues under a unified framework, namely UniOT. First, an OT-based partial alignment with adaptive filling is designed to detect common classes without any predefined threshold values for realistic UniDA. It can automatically discover the intrinsic difference between common and private classes based on the statistical information of the assignment matrix obtained from OT. Second, we propose an OT-based target representation learning that encourages both global discrimination and local consistency of samples to avoid the over-reliance on the source. Notably, UniOT is the first method with the capability to automatically discover and recognize private categories in the target domain for UniDA. Accordingly, we introduce a new metric H^3-score to evaluate the performance in terms of both accuracy of common samples and clustering performance of private ones. Extensive experiments clearly demonstrate the advantages of UniOT over a wide range of state-of-the-art methods in UniDA. | ['Jingya Wang', 'Hoang Duong Tuan', 'Ye Shi', 'Wanxing Chang'] | 2022-10-31 | null | null | null | null | ['universal-domain-adaptation'] | ['computer-vision'] | [ 2.93078214e-01 -4.59151566e-01 -3.81179929e-01 -3.05081993e-01
-9.16985571e-01 -6.54102027e-01 5.14578938e-01 3.29827927e-02
-1.89309895e-01 7.60989130e-01 -3.67862403e-01 2.71026909e-01
-3.08237821e-01 -8.36807072e-01 -4.61340219e-01 -1.19081926e+00
3.11289042e-01 9.68972266e-01 4.54537690e-01 1.17955450e-02
4.10926677e-02 4.51452464e-01 -1.62523770e+00 3.57264787e-01
1.23747659e+00 1.09405148e+00 4.06564951e-01 1.43654430e-02
-2.01490730e-01 5.96902408e-02 -7.06789017e-01 -1.54299974e-01
3.80923629e-01 -5.25449336e-01 -7.44044960e-01 3.16661149e-01
4.99889851e-01 -1.94198452e-02 1.64226741e-01 1.31042719e+00
4.77254719e-01 8.03057551e-02 1.06614757e+00 -1.42744982e+00
-5.39488018e-01 2.99723834e-01 -5.74321449e-01 8.00597370e-02
1.14954181e-01 7.83251897e-02 1.17693150e+00 -7.37894833e-01
6.34412110e-01 1.04091442e+00 6.14339709e-01 3.61620069e-01
-1.59805310e+00 -7.13222742e-01 2.77997553e-01 3.49435836e-01
-1.65039670e+00 -2.94882119e-01 8.02725911e-01 -6.83647513e-01
3.90472353e-01 2.78069645e-01 1.59864977e-01 1.10167348e+00
-1.26638070e-01 6.22795522e-01 1.25799036e+00 -3.98551345e-01
4.27108407e-01 4.55569386e-01 3.57098490e-01 1.49747819e-01
2.00595960e-01 -2.12338939e-01 -4.68739450e-01 -2.90472299e-01
5.60223162e-01 -2.06340700e-01 -2.19013706e-01 -1.05396450e+00
-1.26696980e+00 7.13256717e-01 8.31718594e-02 3.10071409e-01
-1.59764260e-01 -5.59274852e-01 4.53827500e-01 3.09550077e-01
3.61370534e-01 3.56668055e-01 -6.00926638e-01 2.55390912e-01
-6.93165243e-01 3.07219982e-01 4.40063983e-01 1.11191332e+00
9.92845893e-01 -3.50892395e-01 -1.62820101e-01 1.26946020e+00
-3.66975134e-03 6.17460251e-01 6.70613229e-01 -5.65224349e-01
4.79846597e-01 8.43729198e-01 1.30425602e-01 -9.49061692e-01
-1.68311879e-01 -6.04887843e-01 -7.73944318e-01 6.55821860e-02
7.35329509e-01 2.17784643e-01 -7.50440419e-01 1.91184497e+00
6.41698182e-01 2.34811172e-01 2.50378072e-01 8.33418906e-01
5.03664672e-01 4.79928702e-01 -2.48663366e-01 -3.06685627e-01
1.31004190e+00 -6.11354113e-01 -4.79776829e-01 -2.06172973e-01
6.71824694e-01 -7.02208936e-01 1.08575213e+00 3.15663785e-01
-5.21769285e-01 -6.81923211e-01 -1.03555906e+00 2.87537396e-01
-4.21768486e-01 4.08008456e-01 2.58467764e-01 6.27380550e-01
-4.12842780e-01 3.15603733e-01 -4.63701218e-01 -4.92059231e-01
2.65956879e-01 4.20809954e-01 -4.66061831e-01 -2.13544115e-01
-1.06812572e+00 6.25939548e-01 5.99076271e-01 -1.40894040e-01
-6.93891346e-01 -5.88413596e-01 -5.69318056e-01 -6.92002848e-02
5.55770099e-01 -4.62155908e-01 8.98915052e-01 -1.09473729e+00
-1.30299854e+00 1.00963616e+00 -2.40646660e-01 -1.32109910e-01
4.49107975e-01 1.35738239e-01 -5.66328764e-01 3.69938836e-02
4.73658413e-01 4.14575309e-01 6.87131882e-01 -1.43047667e+00
-9.77356493e-01 -4.76421565e-01 -2.54327029e-01 1.07881695e-01
-4.38269883e-01 -2.20161006e-01 -3.41600120e-01 -6.84042573e-01
4.12936836e-01 -8.99605095e-01 1.22592255e-01 -1.25363529e-01
-2.68542022e-01 -4.93601084e-01 8.75689447e-01 -2.87861854e-01
1.15410614e+00 -2.20072150e+00 3.62763613e-01 5.10413349e-01
1.12415113e-01 1.95907414e-01 -2.05141619e-01 1.31578952e-01
-1.47160605e-01 -3.92324090e-01 -5.77926457e-01 4.91968654e-02
1.06818542e-01 3.54649127e-01 -2.84400165e-01 5.16519368e-01
2.07114115e-01 1.29050553e-01 -9.45824027e-01 -4.60288405e-01
1.03079505e-01 5.37349796e-03 -5.03986776e-01 2.17952058e-01
-3.25382054e-01 5.11425555e-01 -3.15449774e-01 6.03174567e-01
1.01566350e+00 -2.97694683e-01 5.82419753e-01 -2.45757297e-01
1.95751250e-01 9.37955007e-02 -1.52575243e+00 1.43556225e+00
-1.87944397e-01 1.60034105e-01 -1.03367187e-01 -1.17957473e+00
1.38997126e+00 8.52391124e-02 6.49540663e-01 -5.19406557e-01
1.92679819e-02 6.80852532e-01 1.80226520e-01 -2.10354954e-01
2.92170018e-01 -9.15909335e-02 -1.83655262e-01 3.20335716e-01
1.39615491e-01 2.76494533e-01 2.39766911e-01 -1.60882607e-01
8.00845206e-01 3.97835299e-02 3.69009376e-01 -4.06211883e-01
8.63270402e-01 5.92234544e-02 1.08624029e+00 6.49312317e-01
-3.14369619e-01 7.34027267e-01 3.95110995e-01 -2.27971479e-01
-1.02697444e+00 -1.22104681e+00 -5.21469057e-01 1.06536984e+00
4.78234082e-01 -1.95957478e-02 -5.32188058e-01 -1.03205514e+00
8.45474899e-02 5.83725035e-01 -4.64019179e-01 -2.18310937e-01
-3.67418200e-01 -9.62021530e-01 3.84465277e-01 2.83800453e-01
5.35445511e-01 -5.82632899e-01 -6.68379664e-02 1.41926453e-01
-4.83794898e-01 -1.24965119e+00 -4.77705628e-01 3.47802818e-01
-6.48856759e-01 -1.29399896e+00 -8.28988791e-01 -8.94697309e-01
7.03365207e-01 3.37802380e-01 7.47550964e-01 -3.87120754e-01
1.51970997e-01 1.19032852e-01 -3.84016305e-01 6.49349615e-02
-5.10253608e-01 2.58366317e-01 3.36500347e-01 5.64241469e-01
8.84188533e-01 -6.84650362e-01 -7.32496530e-02 1.03652275e+00
-7.70777106e-01 -2.50136644e-01 4.45489764e-01 8.76576364e-01
7.59598017e-01 1.92408964e-01 8.02539110e-01 -8.07644784e-01
3.53246242e-01 -7.63758361e-01 -6.56354666e-01 5.42912185e-01
-4.48840380e-01 -2.22417731e-02 6.28932357e-01 -6.58768117e-01
-8.73198330e-01 1.94226027e-01 2.21238017e-01 -4.91634429e-01
-4.24037993e-01 3.26456726e-01 -7.06373870e-01 1.99812323e-01
8.17972004e-01 4.19654757e-01 5.32998815e-02 -6.49489522e-01
-3.52907926e-02 1.00310934e+00 6.81246102e-01 -1.02738774e+00
7.27806509e-01 4.11521882e-01 -1.27861366e-01 -6.88973665e-01
-8.58385265e-01 -8.27089548e-01 -1.21162856e+00 1.63116813e-01
6.49334610e-01 -1.04549599e+00 -1.96702644e-01 6.72740042e-01
-9.09896612e-01 -2.08548501e-01 -2.10083902e-01 4.78606582e-01
-5.11904538e-01 7.42117465e-01 1.09890342e-01 -4.32419628e-01
3.87997553e-02 -1.27102327e+00 9.25794423e-01 -5.50307408e-02
-1.99713871e-01 -8.50706398e-01 4.40806188e-02 2.68820584e-01
-7.89583027e-02 2.03421235e-01 9.78127599e-01 -1.02048397e+00
-3.73546243e-01 -2.37872321e-02 -2.58830786e-01 4.95793551e-01
6.29695833e-01 -1.42859101e-01 -9.53092754e-01 -4.35461611e-01
-2.65456825e-01 -1.63667008e-01 5.06960154e-01 1.61735758e-01
1.05492425e+00 -1.21844467e-02 -4.98251975e-01 4.38152730e-01
1.23987067e+00 3.16367567e-01 4.69332665e-01 5.54266155e-01
6.20691180e-01 6.77335083e-01 1.01933777e+00 4.28821236e-01
3.25162679e-01 1.22861791e+00 8.39823857e-02 2.69164592e-01
-1.00616142e-01 5.32695092e-02 4.94854659e-01 7.02433825e-01
2.58338511e-01 -1.53051838e-01 -1.00391579e+00 8.30042362e-01
-1.99756992e+00 -8.20878327e-01 -1.59323022e-01 2.54729795e+00
1.01173401e+00 4.79493290e-02 3.80119562e-01 1.86168030e-01
1.17677426e+00 -3.11946541e-01 -5.88981211e-01 5.08062951e-02
-3.35694730e-01 1.27950050e-02 3.24136198e-01 1.52208343e-01
-1.24018323e+00 7.51108468e-01 5.48276138e+00 1.33599246e+00
-1.00175953e+00 5.18848263e-02 2.63241172e-01 2.04935864e-01
-1.32655174e-01 -1.69823334e-01 -1.08228004e+00 7.44530559e-01
6.08865857e-01 -8.01434293e-02 1.00057319e-01 9.41807747e-01
-1.99370503e-01 -9.56512056e-03 -1.23193192e+00 9.72540259e-01
-1.66149646e-01 -9.87474382e-01 1.57838866e-01 2.07449734e-01
8.40829492e-01 -2.62717426e-01 -3.46445851e-02 1.44652814e-01
4.20054018e-01 -5.32972276e-01 5.87038457e-01 1.50036588e-01
9.52458680e-01 -8.04884613e-01 8.69039655e-01 4.25594300e-01
-1.13237953e+00 -8.57214704e-02 -6.52333081e-01 3.16317141e-01
-2.59387642e-01 7.20272422e-01 -8.60298038e-01 8.61138344e-01
6.40892446e-01 7.92980969e-01 -5.10810673e-01 1.11078429e+00
4.86126095e-02 3.95793170e-01 -3.97031903e-01 4.07332629e-01
1.51977882e-01 -3.77263993e-01 6.58666790e-01 9.57737625e-01
5.39112866e-01 -3.40129018e-01 5.47048807e-01 5.84169030e-01
1.82993591e-01 2.50989497e-01 -4.40199345e-01 2.23782435e-01
6.68499410e-01 9.32494223e-01 -6.52721345e-01 -3.25982243e-01
-2.34671310e-01 9.30896223e-01 2.51432240e-01 1.90764114e-01
-8.78400743e-01 -3.26767474e-01 1.02516556e+00 7.31194317e-02
4.68591899e-01 -1.42187451e-03 -2.99417734e-01 -1.32133973e+00
1.92098215e-01 -1.08605790e+00 8.26223195e-01 -3.14019918e-01
-1.75289559e+00 4.40172583e-01 1.32091925e-01 -1.69837379e+00
-5.97185455e-02 -5.74586451e-01 -1.97649032e-01 8.12024653e-01
-1.48486507e+00 -1.19102550e+00 -3.35260481e-01 8.39408875e-01
4.35622603e-01 -5.55493712e-01 7.39196658e-01 5.31231940e-01
-6.83233500e-01 9.45660233e-01 6.12054884e-01 7.71543086e-02
1.13660610e+00 -1.09262800e+00 -1.61357462e-01 9.00809765e-01
-7.72961080e-02 5.28672338e-01 5.43714881e-01 -5.44948816e-01
-8.86841714e-01 -1.18821311e+00 6.76333964e-01 -3.96554232e-01
5.34291148e-01 -4.09297049e-01 -1.22481382e+00 6.54450536e-01
-3.63343149e-01 4.51318286e-02 8.21246803e-01 2.63502091e-01
-6.91078484e-01 -4.41165417e-01 -1.26127851e+00 2.84624755e-01
9.30255890e-01 -4.52235222e-01 -5.30783951e-01 3.30441028e-01
4.34483588e-01 -2.11918190e-01 -1.10421622e+00 3.94120455e-01
3.86325747e-01 -8.22618484e-01 9.28299487e-01 -3.12518895e-01
1.64760649e-01 -8.44851673e-01 -4.79250640e-01 -1.22651601e+00
-4.44745183e-01 -7.16197267e-02 3.03029090e-01 1.59192860e+00
1.51440874e-01 -9.90971684e-01 6.09501660e-01 1.63428649e-01
-1.52758062e-01 -1.95221767e-01 -1.20594919e+00 -1.22402620e+00
1.00295678e-01 -1.46277189e-01 8.62990677e-01 1.38286245e+00
-2.35810727e-01 1.97666809e-01 -4.00374264e-01 6.12392604e-01
6.83592975e-01 3.97092909e-01 1.08369315e+00 -1.60400617e+00
-2.49399379e-01 -3.60994458e-01 -5.17879546e-01 -1.03403914e+00
2.06129327e-01 -1.13375008e+00 9.48010106e-03 -1.00593317e+00
2.94613659e-01 -1.14899886e+00 -4.78396595e-01 4.99811888e-01
-4.63083871e-02 2.01910526e-01 -7.14528933e-02 6.77957118e-01
-5.28407097e-01 6.10548675e-01 1.10361230e+00 -1.39293596e-01
-2.93157101e-01 6.10786080e-02 -7.07915604e-01 4.51048493e-01
7.43266165e-01 -5.44325233e-01 -3.60980213e-01 -1.85261920e-01
-2.53281057e-01 -4.38971847e-01 3.52214754e-01 -1.18014860e+00
-3.05456109e-02 -4.72838908e-01 2.19748974e-01 -6.37422323e-01
1.31066337e-01 -9.46595252e-01 2.97481418e-01 1.81179509e-01
-8.41291547e-02 -3.65374833e-01 4.09991480e-03 6.27545714e-01
-3.13090682e-01 -2.78064936e-01 1.05707562e+00 1.07007645e-01
-1.00313711e+00 1.08653568e-01 -1.29924223e-01 -4.69385646e-02
1.27130020e+00 -4.67818916e-01 -3.08898091e-01 1.46358788e-01
-5.93417048e-01 1.77432090e-01 6.52100027e-01 3.02185923e-01
2.29018047e-01 -1.45081007e+00 -6.01927102e-01 2.23877817e-01
5.46680450e-01 1.94748461e-01 2.86313862e-01 7.62002051e-01
-1.05674051e-01 3.02099556e-01 -4.67038184e-01 -1.06460381e+00
-1.41067398e+00 5.67497790e-01 3.90469760e-01 -4.05974746e-01
-3.79146963e-01 5.68261743e-01 5.95714927e-01 -9.95503187e-01
4.27340604e-02 -5.50532788e-02 -2.52666682e-01 3.11938465e-01
3.74998212e-01 2.52605557e-01 1.18759640e-01 -7.55803227e-01
-4.21771586e-01 6.55338049e-01 -2.03173265e-01 4.44383889e-01
1.19047701e+00 -2.70290226e-01 -1.53467357e-01 3.37029368e-01
1.05274355e+00 -1.77574586e-02 -1.18088043e+00 -6.80185854e-01
1.71678275e-01 -6.93782866e-01 -3.46861482e-01 -7.01507866e-01
-8.85578156e-01 7.90551007e-01 7.52561629e-01 -3.03212292e-02
1.10979176e+00 -4.71079499e-02 6.75561666e-01 2.03046933e-01
3.55888903e-01 -1.25253391e+00 1.02698840e-01 4.70202118e-01
5.61058879e-01 -1.21053433e+00 -5.18039502e-02 -7.08586454e-01
-6.39731109e-01 1.15672672e+00 8.69371593e-01 2.13537529e-01
1.89648345e-01 -1.88190669e-01 3.54754478e-02 3.55658308e-02
-3.49773794e-01 -3.71642798e-01 3.36310863e-01 9.57103670e-01
-2.05351785e-02 9.41757187e-02 -8.54639858e-02 5.49289703e-01
3.66720110e-02 -2.31247559e-01 4.03839558e-01 5.82023442e-01
-4.50173914e-01 -1.54199493e+00 -6.90327108e-01 2.08869025e-01
-8.98268744e-02 3.62308264e-01 -3.01692754e-01 7.06620991e-01
4.12776172e-01 7.28761375e-01 9.51044485e-02 -3.96615833e-01
2.51140028e-01 3.06627661e-01 3.24297011e-01 -6.48805559e-01
-3.15116554e-01 1.56217217e-01 -1.00571811e-01 -2.07928140e-02
-5.51268220e-01 -9.13048029e-01 -1.05337918e+00 -1.49316698e-01
-3.69768292e-01 5.00556417e-02 2.19197974e-01 8.78677309e-01
4.82738465e-01 1.92433715e-01 7.45677590e-01 -4.23045069e-01
-6.78837478e-01 -7.63195395e-01 -8.25579703e-01 5.72400987e-01
1.67991981e-01 -1.19442725e+00 -2.21181601e-01 -9.75104496e-02] | [10.335369110107422, 3.121565341949463] |
0d1f7962-eab8-4433-ba06-0f7e2401052a | panoptic-3d-scene-reconstruction-from-a | 2111.02444 | null | https://arxiv.org/abs/2111.02444v2 | https://arxiv.org/pdf/2111.02444v2.pdf | Panoptic 3D Scene Reconstruction From a Single RGB Image | Understanding 3D scenes from a single image is fundamental to a wide variety of tasks, such as for robotics, motion planning, or augmented reality. Existing works in 3D perception from a single RGB image tend to focus on geometric reconstruction only, or geometric reconstruction with semantic segmentation or instance segmentation. Inspired by 2D panoptic segmentation, we propose to unify the tasks of geometric reconstruction, 3D semantic segmentation, and 3D instance segmentation into the task of panoptic 3D scene reconstruction - from a single RGB image, predicting the complete geometric reconstruction of the scene in the camera frustum of the image, along with semantic and instance segmentations. We thus propose a new approach for holistic 3D scene understanding from a single RGB image which learns to lift and propagate 2D features from an input image to a 3D volumetric scene representation. We demonstrate that this holistic view of joint scene reconstruction, semantic, and instance segmentation is beneficial over treating the tasks independently, thus outperforming alternative approaches. | ['Angela Dai', 'Matthias Nießner', 'Ji Hou', 'Manuel Dahnert'] | 2021-11-03 | null | http://proceedings.neurips.cc/paper/2021/hash/46031b3d04dc90994ca317a7c55c4289-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/46031b3d04dc90994ca317a7c55c4289-Paper.pdf | neurips-2021-12 | ['3d-instance-segmentation-1', '3d-scene-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 6.92094207e-01 3.58325183e-01 -2.44403491e-03 -4.75127220e-01
-5.40489793e-01 -6.64941072e-01 6.94701791e-01 1.80102691e-01
-9.92187858e-02 7.09370673e-02 -6.05198033e-02 -3.98774862e-01
4.42670770e-02 -8.65174711e-01 -9.39671874e-01 -3.87065709e-01
4.17261958e-01 7.38675237e-01 2.13498071e-01 -1.15496339e-02
2.96915025e-01 7.73369133e-01 -1.58745909e+00 -3.00272219e-02
4.74297166e-01 1.21636927e+00 5.59073210e-01 7.76050925e-01
-4.23762918e-01 4.16544378e-01 7.41645768e-02 6.95913881e-02
3.85356456e-01 -1.98060289e-01 -9.46571708e-01 9.32582796e-01
3.59736949e-01 -4.44823533e-01 -5.04408777e-02 9.28265810e-01
-1.04515880e-01 1.95422634e-01 6.81341767e-01 -9.95916724e-01
-6.78607672e-02 -1.70860887e-01 -4.32250202e-01 -5.56744516e-01
7.43813932e-01 -1.44773200e-01 8.51021588e-01 -7.15789616e-01
7.57411361e-01 1.31806767e+00 5.21618366e-01 3.08581620e-01
-1.30543804e+00 1.50382504e-01 3.96067649e-01 -3.58095556e-01
-1.15385473e+00 -7.24883378e-02 1.03983915e+00 -6.65608525e-01
8.20724308e-01 1.33046001e-01 8.45177233e-01 4.94438559e-01
-1.00144483e-01 7.44646668e-01 1.22079360e+00 -3.39946508e-01
4.84344900e-01 -6.27967715e-03 1.43830791e-01 8.21139395e-01
3.59635763e-02 -1.18434452e-01 -3.13498050e-01 3.15701902e-01
1.06355274e+00 4.08856988e-01 -2.37585634e-01 -8.62815678e-01
-1.00032067e+00 7.12588787e-01 5.46435177e-01 -1.16018709e-02
-4.52197194e-01 3.93438190e-01 -5.02586924e-02 -6.13149703e-02
6.65963233e-01 3.13811392e-01 -6.79151833e-01 1.90400690e-01
-7.58130550e-01 2.10081056e-01 6.81684554e-01 9.57577884e-01
1.28366268e+00 -1.93952143e-01 6.22623265e-01 2.82447904e-01
5.46202004e-01 6.39893055e-01 -5.85740767e-02 -1.36594021e+00
2.00133204e-01 8.51691365e-01 2.20159143e-01 -7.32104063e-01
-6.44545376e-01 -2.68050805e-02 -3.92234713e-01 2.86814868e-01
2.64089823e-01 3.82071942e-01 -1.28796220e+00 1.33916473e+00
7.14024782e-01 -9.23420116e-02 5.90270795e-02 9.50738966e-01
6.66693270e-01 6.40461206e-01 -2.10764110e-01 3.63066904e-02
1.12669766e+00 -7.44599164e-01 -1.75511520e-02 -5.98901451e-01
5.42156041e-01 -4.11923289e-01 8.35322857e-01 5.02420068e-01
-1.12381399e+00 -5.83642900e-01 -9.42877829e-01 -6.27914131e-01
-4.11884606e-01 -1.01961926e-01 5.80278754e-01 3.09835136e-01
-1.12535822e+00 4.22503740e-01 -8.81035924e-01 -5.69604516e-01
3.50193620e-01 2.45863453e-01 -6.73171520e-01 -5.58100641e-01
-3.61017138e-01 8.79229903e-01 5.20865023e-01 -3.80029045e-02
-9.33228970e-01 -5.49528837e-01 -1.09341776e+00 -2.73271799e-01
5.89695454e-01 -1.02538741e+00 1.18610036e+00 -7.13077366e-01
-1.26548755e+00 1.18620980e+00 -1.95096061e-01 -2.95323849e-01
2.80988336e-01 -3.97264570e-01 5.91992557e-01 2.85397798e-01
8.10041353e-02 7.57806242e-01 8.73051703e-01 -1.88096511e+00
-3.21692705e-01 -1.03854942e+00 2.63387501e-01 5.97060323e-01
5.90445876e-01 -8.31655502e-01 -6.21967077e-01 -5.59333386e-03
9.62160349e-01 -7.58656323e-01 -7.38501966e-01 9.89111513e-02
-7.53335714e-01 2.45681301e-01 8.18440855e-01 -5.66093147e-01
2.16301844e-01 -1.93686450e+00 6.85368955e-01 1.41151652e-01
9.21941176e-02 -3.81624132e-01 -7.17232451e-02 7.80988485e-02
9.94783416e-02 3.55010927e-02 -6.36068404e-01 -8.85539472e-01
-5.94140403e-02 4.86119539e-01 -4.00849313e-01 6.62219286e-01
-1.98093057e-02 9.22221780e-01 -8.34100723e-01 -1.74975872e-01
8.38850498e-01 6.33084536e-01 -6.19081259e-01 3.97324145e-01
-8.59339356e-01 1.01667309e+00 -8.94505918e-01 5.86347759e-01
7.54175246e-01 -2.31788144e-01 9.64106396e-02 -8.93320963e-02
-2.14631766e-01 2.67471910e-01 -9.75870371e-01 2.60857177e+00
-5.82241595e-01 2.39999086e-01 2.30389848e-01 -1.30806494e+00
7.21777856e-01 2.14542970e-02 7.85790265e-01 -7.20840275e-01
1.45062745e-01 7.17337877e-02 -1.04172122e+00 -4.02674645e-01
6.17109895e-01 -4.27366406e-01 -5.22558928e-01 4.85412449e-01
-2.90171872e-03 -1.36007845e+00 -5.26349425e-01 2.03253612e-01
7.94270158e-01 7.01834857e-01 3.01855296e-01 1.95503652e-01
3.19920361e-01 4.10815746e-01 9.24331546e-02 5.66810548e-01
3.01042229e-01 8.26814890e-01 3.64210039e-01 -3.58511955e-01
-1.11081851e+00 -1.35517609e+00 7.08578303e-02 4.61130649e-01
7.45695055e-01 -1.28401428e-01 -8.61392140e-01 -4.49435890e-01
5.58240041e-02 6.78563535e-01 -5.19011736e-01 1.71706587e-01
-4.33081508e-01 -2.93810099e-01 -2.22128734e-01 1.67274266e-01
3.78658980e-01 -6.92558229e-01 -1.10268068e+00 1.94127355e-02
-1.62451386e-01 -1.61367643e+00 4.75349538e-02 3.47807735e-01
-1.24099767e+00 -1.19737124e+00 -4.66318041e-01 -4.07788992e-01
6.62397921e-01 6.72744751e-01 1.10698605e+00 -9.37860683e-02
-5.11287630e-01 1.23942983e+00 -2.65273243e-01 -1.00670949e-01
-3.02332550e-01 -2.20590606e-01 -1.46239027e-01 1.20135387e-02
-3.52653682e-01 -7.69617081e-01 -4.64766741e-01 -1.50313199e-01
-1.01342773e+00 5.91331184e-01 4.60858256e-01 2.05779418e-01
1.09195089e+00 -1.75787248e-02 -1.19605504e-01 -9.33109283e-01
-3.28132093e-01 -3.78815621e-01 -5.94478667e-01 1.14556417e-01
-2.51194298e-01 -2.17592064e-02 3.61644566e-01 2.22710311e-01
-1.00335133e+00 6.55001402e-01 -1.08200096e-01 -6.37589335e-01
-7.79738605e-01 8.76902863e-02 -5.18438637e-01 1.10926904e-01
2.70892411e-01 4.66938436e-01 -9.75837037e-02 -6.43172443e-01
8.85818124e-01 2.13455558e-01 6.23031557e-01 -3.83183450e-01
6.73450828e-01 9.19608355e-01 3.42069656e-01 -1.01429904e+00
-1.09373355e+00 -8.09704721e-01 -1.24063671e+00 -1.81223303e-01
1.36558747e+00 -1.05178082e+00 -4.43787396e-01 3.18877697e-01
-1.29573381e+00 -6.34153664e-01 -4.46617544e-01 3.78219455e-01
-1.18090391e+00 4.47715342e-01 -1.27908081e-01 -8.16213667e-01
1.14052624e-01 -1.31848049e+00 1.87212229e+00 1.08118325e-01
1.41404256e-01 -1.06074488e+00 -1.30776659e-01 8.00424635e-01
-1.90645754e-01 6.75164461e-01 9.70642984e-01 -6.32091388e-02
-1.09940982e+00 -6.62083402e-02 -3.00275832e-01 2.21226469e-01
4.26781327e-02 -4.63012934e-01 -1.21114767e+00 3.02057654e-01
3.39250565e-01 -3.27370524e-01 8.83121550e-01 5.28288662e-01
9.65262592e-01 1.11959837e-01 -2.99972236e-01 7.94494152e-01
1.86198092e+00 -7.20086321e-02 3.76977146e-01 5.48595656e-03
1.11295164e+00 7.87505805e-01 5.57664275e-01 3.55274588e-01
6.83492422e-01 6.50670171e-01 1.06790972e+00 -1.77538052e-01
-1.81227222e-01 -4.00988489e-01 7.43006915e-02 5.11322260e-01
1.02779903e-01 -2.96425577e-02 -7.68993378e-01 4.07737881e-01
-1.63044095e+00 -2.74731725e-01 -1.62485614e-01 2.15578175e+00
1.88970819e-01 -8.76972228e-02 -3.61892879e-02 2.48327747e-01
2.43118286e-01 2.46837422e-01 -7.84321368e-01 -2.80512094e-01
-6.04258059e-03 7.19012916e-02 5.25267482e-01 8.19398344e-01
-8.97099733e-01 1.16879880e+00 5.53285027e+00 3.76036614e-01
-1.06316555e+00 1.11995660e-01 6.05908155e-01 1.82835594e-01
-6.77471161e-01 3.67248774e-01 -5.88089347e-01 -1.62928328e-01
4.38150823e-01 4.90022570e-01 6.06008768e-01 8.16932619e-01
1.24736443e-01 -6.70689940e-01 -1.35260952e+00 1.06996167e+00
1.70157164e-01 -1.19303393e+00 3.78923953e-01 2.57013172e-01
8.06407094e-01 5.43935187e-02 -1.25653669e-01 -1.74666151e-01
1.90018922e-01 -9.96438980e-01 1.05787504e+00 7.21136034e-01
7.25318849e-01 -3.17530990e-01 7.78042153e-02 7.43894637e-01
-1.16147900e+00 6.03287593e-02 8.47165883e-02 -2.42171367e-03
5.31335771e-01 6.50799692e-01 -6.85761273e-01 9.51481938e-01
3.77430946e-01 8.98360074e-01 -2.75172800e-01 6.74316108e-01
-1.98989138e-01 -1.11864589e-01 -3.57820332e-01 5.23920894e-01
3.09795558e-01 -4.21406567e-01 4.49290961e-01 6.36907399e-01
3.01351130e-01 3.78791362e-01 5.17673969e-01 9.95571911e-01
7.02684149e-02 -1.99008897e-01 -8.88706386e-01 1.49017572e-01
-6.11393750e-02 9.97380078e-01 -1.18833280e+00 -3.07777137e-01
-2.81490028e-01 1.20099247e+00 7.85308778e-02 3.65792543e-01
-4.48451489e-01 2.26295322e-01 4.33551937e-01 1.84964910e-01
3.83961946e-01 -7.79331505e-01 -7.93733656e-01 -1.10427630e+00
-8.27755928e-02 5.31310067e-02 -5.43480888e-02 -1.21438503e+00
-5.61343074e-01 2.61979520e-01 2.16639712e-01 -9.71550822e-01
-1.14560321e-01 -7.47808456e-01 -2.04903856e-01 6.64210081e-01
-1.71622372e+00 -1.28909421e+00 -4.78462398e-01 6.71803653e-01
7.74547935e-01 7.77735472e-01 7.04911470e-01 -3.32715988e-01
-8.92886445e-02 -5.41521192e-01 -2.04910547e-01 -3.82763594e-01
-1.86666787e-01 -1.32869875e+00 2.94299960e-01 6.83922648e-01
1.51725858e-01 9.33834240e-02 4.35061485e-01 -4.84860390e-01
-1.91142762e+00 -1.19700658e+00 4.83458549e-01 -6.08482182e-01
1.54380947e-01 -5.19862235e-01 -4.91956323e-01 6.57858074e-01
-3.81897151e-01 -2.10904647e-02 1.27761349e-01 -3.27042401e-01
-3.25718015e-01 1.17106982e-01 -1.21069109e+00 4.11152035e-01
1.14820063e+00 -7.33693838e-01 -3.65754724e-01 3.92562956e-01
1.01053286e+00 -5.27667344e-01 -8.23208988e-01 3.37989688e-01
4.24491316e-01 -1.00586176e+00 1.41746128e+00 -1.45892367e-01
3.48231047e-01 -4.04893458e-01 -7.80574203e-01 -9.22924042e-01
3.75994951e-01 -3.35521013e-01 8.07653069e-02 5.16308486e-01
8.22234601e-02 -2.55721450e-01 9.35998678e-01 5.94423592e-01
-4.73924518e-01 -5.20459712e-01 -8.90785038e-01 -2.23571569e-01
-2.12609977e-01 -1.02384090e+00 4.56213772e-01 6.32311285e-01
-5.68486392e-01 3.68354887e-01 -5.89563549e-02 4.63535070e-01
7.89157569e-01 7.75677025e-01 9.45002913e-01 -1.41208923e+00
-2.39910662e-01 -3.14421773e-01 -5.05497575e-01 -1.66169679e+00
9.01959389e-02 -9.15294051e-01 4.24473882e-01 -1.98194158e+00
9.54438671e-02 -6.37010634e-01 3.90278161e-01 1.93327338e-01
2.54721522e-01 2.96903431e-01 2.66388774e-01 1.31132036e-01
-8.16094577e-01 6.69103682e-01 1.45227325e+00 5.78028485e-02
-3.52603406e-01 -6.48963526e-02 -5.46005845e-01 8.22319210e-01
3.14355046e-01 -1.92801431e-01 -5.10631025e-01 -6.57656789e-01
1.15914322e-01 6.16580486e-01 6.84648752e-01 -7.29582429e-01
6.39988258e-02 -2.40536198e-01 2.81422317e-01 -7.62499750e-01
1.00614810e+00 -8.89988124e-01 1.70825999e-02 2.02390075e-01
-1.35767031e-02 -4.22772676e-01 3.51317823e-02 8.86325300e-01
1.28478006e-01 -9.00333375e-03 5.40168464e-01 -7.75545061e-01
-1.01719618e+00 4.42176789e-01 -1.36700675e-01 -1.51550919e-01
8.98302376e-01 -5.63668132e-01 2.30026290e-01 -2.04481006e-01
-1.22140384e+00 -6.56721592e-02 9.96384859e-01 1.57007322e-01
9.96099830e-01 -8.48182380e-01 -1.96350709e-01 3.45402598e-01
-4.97492589e-02 9.48199093e-01 4.44092512e-01 6.65323079e-01
-6.18053675e-01 6.63732231e-01 7.07627982e-02 -1.17566454e+00
-8.77295554e-01 4.38118279e-01 4.61269766e-01 -1.03900358e-02
-9.13747013e-01 6.70528114e-01 7.94176936e-01 -9.57331836e-01
1.01093844e-01 -7.68857598e-01 1.49979874e-01 -3.13452154e-01
-5.42892702e-02 -2.93627977e-02 -1.30073763e-02 -1.02566564e+00
-8.82138088e-02 1.30011904e+00 4.33968276e-01 -3.53168935e-01
1.24021924e+00 -5.37341893e-01 -2.64739804e-02 8.87437105e-01
1.18741429e+00 -4.27297056e-01 -1.52183902e+00 -2.83118337e-01
-8.05928409e-02 -3.94993424e-01 2.04093531e-01 -4.53447104e-01
-8.09464574e-01 1.16240001e+00 1.36009291e-01 1.45243093e-01
1.17107737e+00 5.26571572e-01 7.30140448e-01 2.73244411e-01
7.67788529e-01 -6.45724952e-01 2.31117114e-01 6.06448174e-01
6.25688255e-01 -1.16413271e+00 2.27369994e-01 -5.99088073e-01
-3.91109079e-01 1.11577809e+00 9.27215591e-02 -2.32971251e-01
7.60887325e-01 -2.06620708e-01 -3.18449825e-01 -4.47899699e-01
-2.60042727e-01 -4.75860983e-01 4.03765947e-01 5.01217067e-01
-1.53300136e-01 2.01733232e-01 6.23741746e-01 2.20892981e-01
2.73566954e-02 -2.27316514e-01 4.66575861e-01 9.41966116e-01
-5.68887472e-01 -9.62610960e-01 -3.58796686e-01 -2.92851496e-02
1.09484196e-01 1.84005052e-01 -4.65140462e-01 7.38046050e-01
3.16568464e-01 7.19834566e-01 2.85120189e-01 -3.49960566e-01
2.90657669e-01 5.25115505e-02 9.87731993e-01 -1.00260198e+00
8.48380178e-02 2.09033340e-01 -1.88156351e-01 -8.70717943e-01
-6.96948111e-01 -7.98524022e-01 -1.56422818e+00 2.75688827e-01
8.56114104e-02 -1.99432760e-01 1.41615665e+00 1.39241481e+00
-2.58412082e-02 1.87456176e-01 8.28526020e-01 -1.53503919e+00
1.83435723e-01 -4.67298746e-01 -7.37494946e-01 1.17167018e-01
7.23627388e-01 -6.58544004e-01 -2.65838116e-01 2.19222084e-01] | [8.461353302001953, -2.906177520751953] |
270b6475-f14d-4210-937c-8d998ea46558 | sequential-deformation-for-accurate-scene | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/6576_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123740103.pdf | Sequential Deformation for Accurate Scene Text Detection | Scene text detection has been significantly advanced over recent years, especially after the emergence of deep neural network. However, due to high diversity of scene texts in scale, orientation, shape and aspect ratio, as well as the inherent limitation of convolutional neural network for geometric transformations, to achieve accurate scene text detection is still an open problem. In this paper, we propose a novel sequential deformation method to effectively model the line-shape of scene text. An auxiliary character counting supervision is further introduced to guide the sequential offset prediction. The whole network can be easily optimized through an end-to-end multi-task manner. Extensive experiments are conducted on public scene text detection datasets including ICDAR 2017 MLT, ICDAR 2015, Total-text and SCUT-CTW1500. The experimental results demonstrate that the proposed method has outperformed previous state-of-the-art methods. | ['Liangrui Peng', 'Keyu An', 'Jaesik Min', 'Shanyu Xiao', 'Gang Yao', 'Ruijie Yan'] | null | null | null | null | eccv-2020-8 | ['scene-text-detection'] | ['computer-vision'] | [ 2.78340638e-01 -6.93960309e-01 1.72723114e-01 -2.96850830e-01
-6.56616747e-01 -5.16001880e-01 6.38978183e-01 2.47963771e-01
-5.10746419e-01 5.33610806e-02 1.62823513e-01 -1.17022865e-01
2.43794292e-01 -5.78784823e-01 -7.20387220e-01 -4.29372847e-01
5.42005658e-01 5.59874713e-01 6.16750062e-01 -1.24349862e-01
4.88025874e-01 2.71895170e-01 -9.46556032e-01 2.59631991e-01
9.51981723e-01 7.45987952e-01 4.91453797e-01 7.09227800e-01
-4.73197907e-01 7.81730354e-01 -4.54983652e-01 -6.01484001e-01
2.14261159e-01 -4.11625691e-02 -4.59569991e-01 3.58794987e-01
6.81331813e-01 -7.05852628e-01 -7.18869507e-01 1.03662062e+00
7.78802216e-01 1.99780818e-02 5.67954183e-01 -7.46727824e-01
-5.42568207e-01 5.88271379e-01 -1.00972128e+00 8.64086300e-03
1.68105125e-01 1.43322468e-01 8.86360645e-01 -9.77982879e-01
4.39875454e-01 1.27086377e+00 7.05150604e-01 6.88711405e-02
-7.49563813e-01 -6.53052211e-01 4.31663483e-01 1.24536082e-01
-1.42504489e+00 -8.26503560e-02 9.24525738e-01 -4.04245913e-01
9.36006546e-01 -3.65558825e-02 2.09130362e-01 9.82694983e-01
-4.95287478e-02 1.21898901e+00 6.08570874e-01 -3.81680518e-01
-3.03332150e-01 -1.41792536e-01 2.79281568e-03 8.38950217e-01
3.25799584e-01 -5.33889353e-01 -3.72367650e-01 3.78761403e-02
5.94791114e-01 1.30220681e-01 -1.47502437e-01 -3.09382111e-01
-1.30132353e+00 7.09770679e-01 2.92692274e-01 3.35648842e-02
1.06276304e-01 1.19205497e-01 7.67348647e-01 -3.16733450e-01
5.86080074e-01 -1.79232452e-02 -2.81085640e-01 -2.16103464e-01
-8.95794034e-01 3.23328972e-01 4.77869749e-01 1.06053555e+00
3.94900829e-01 2.59312958e-01 -2.96563864e-01 1.14276409e+00
3.42143565e-01 8.20055842e-01 4.52452272e-01 -8.22561681e-02
1.19628763e+00 8.36739063e-01 -1.51804760e-01 -1.17868400e+00
-5.43500543e-01 -3.62386853e-01 -1.06790721e+00 -4.10577804e-01
2.97651500e-01 -7.45161474e-02 -9.54563141e-01 8.91816199e-01
3.75653356e-01 9.74431820e-03 -3.57138634e-01 7.72088528e-01
7.90637493e-01 8.28821361e-01 -1.11225778e-02 3.12058866e-01
1.30650079e+00 -1.18693304e+00 -6.10931337e-01 -5.27517319e-01
8.14215899e-01 -1.25634623e+00 1.13073063e+00 4.21479225e-01
-6.85265005e-01 -5.55728734e-01 -9.72488165e-01 -5.68025649e-01
-2.64172256e-01 8.00606430e-01 4.51220661e-01 5.17736197e-01
-4.02850509e-01 1.25504956e-01 -8.53610635e-01 -4.52871025e-01
7.09745347e-01 2.40294352e-01 1.61687911e-01 -1.17869146e-01
-8.09969544e-01 3.93898964e-01 7.08778799e-01 2.83447742e-01
-5.79133034e-01 -4.26027894e-01 -7.70053267e-01 1.46378130e-01
4.88109857e-01 -4.74507987e-01 1.08602488e+00 -5.35877526e-01
-1.49645412e+00 8.09628248e-01 1.36785835e-01 -2.94613153e-01
1.08217120e+00 -4.55249250e-01 -3.45023245e-01 6.14654236e-02
5.20614944e-02 4.49871689e-01 8.72911155e-01 -1.00615108e+00
-8.28551054e-01 -3.60821605e-01 -3.53286713e-01 3.14985991e-01
-7.85265028e-01 2.96150684e-01 -1.00920808e+00 -9.70842421e-01
8.63820985e-02 -8.53101134e-01 -9.65684578e-02 2.04235211e-01
-8.63821149e-01 -2.08478808e-01 1.28909886e+00 -7.71570146e-01
1.20679414e+00 -2.05943489e+00 -1.05195418e-01 -1.91485405e-01
-6.58192486e-03 3.25312227e-01 -5.06243780e-02 4.07792479e-01
1.44152820e-01 5.57348020e-02 -1.97080061e-01 -7.48701453e-01
4.89979982e-02 -2.59272009e-01 -5.54957271e-01 6.98508918e-01
5.44643775e-02 7.42963195e-01 -4.09998745e-01 -8.73292923e-01
6.86544299e-01 4.65589523e-01 -4.02907193e-01 2.45608427e-02
-4.46103901e-01 5.22804796e-04 -7.12009966e-01 6.21276557e-01
1.03261304e+00 -3.86700273e-01 -2.09508911e-01 -2.75131017e-01
-1.79056004e-01 1.60957724e-02 -1.16700721e+00 1.73165393e+00
-1.45662010e-01 9.26183641e-01 -2.89106280e-01 -8.33035052e-01
8.69646132e-01 -6.66796565e-02 2.48424634e-01 -6.13657415e-01
4.97067332e-01 -3.31177227e-02 -2.36643389e-01 -5.20458341e-01
9.60624576e-01 4.75796551e-01 -2.47981828e-02 1.80572033e-01
-4.70544070e-01 -2.21330807e-01 2.32360736e-01 2.84316152e-01
6.20250225e-01 2.61145085e-01 2.04913300e-02 -1.46208450e-01
6.50509059e-01 2.28362411e-01 3.60713035e-01 6.50527596e-01
2.63332948e-02 1.01875401e+00 3.53177607e-01 -5.21729052e-01
-1.20208788e+00 -5.94547868e-01 -2.40763426e-01 9.11150396e-01
3.85125875e-01 -4.15422827e-01 -7.57607996e-01 -5.91484308e-01
-1.14924714e-01 5.38463891e-01 -4.07436728e-01 2.29911283e-01
-1.00510406e+00 -1.02709043e+00 8.99055183e-01 6.35357320e-01
1.03770685e+00 -8.29556584e-01 -3.03578228e-01 1.91580474e-01
-2.72223383e-01 -1.82809532e+00 -9.10939813e-01 -1.81286812e-01
-8.50424051e-01 -7.17555046e-01 -7.83923805e-01 -9.13146973e-01
6.41218543e-01 5.25867760e-01 6.10112071e-01 1.65273830e-01
-5.40881813e-01 -5.96382134e-02 -4.47352141e-01 -4.66900200e-01
-2.85328835e-01 3.79912168e-01 -4.78376061e-01 1.01156034e-01
1.10921755e-01 7.30214491e-02 -6.80338442e-01 2.30765611e-01
-1.14832199e+00 5.79577506e-01 5.07097960e-01 5.51595867e-01
4.80379343e-01 2.31913164e-01 -4.25139405e-02 -8.42413247e-01
2.29030237e-01 -4.21493240e-02 -7.79476523e-01 2.93370277e-01
-2.94653773e-01 -1.36731595e-01 8.73032570e-01 -4.63785112e-01
-1.27478671e+00 2.50616282e-01 -1.97437689e-01 -2.42546991e-01
-2.00415999e-01 3.75374228e-01 -1.93003073e-01 8.48499537e-02
3.02700251e-01 6.90364599e-01 -6.38494551e-01 -5.48902631e-01
5.92327975e-02 8.47323656e-01 4.94623303e-01 -5.42782307e-01
1.00710702e+00 6.89374208e-01 -1.09062150e-01 -1.20119631e+00
-8.18448663e-01 -4.92314279e-01 -6.99532390e-01 6.04376234e-02
1.02539182e+00 -1.04716909e+00 -5.20975590e-01 1.21386051e+00
-1.38250411e+00 -4.63102698e-01 5.02882361e-01 1.98224962e-01
-4.51160781e-02 9.91387606e-01 -7.87603498e-01 -5.40376365e-01
-7.59480476e-01 -1.15129995e+00 1.58802271e+00 1.29244879e-01
3.83917511e-01 -8.13616812e-01 -1.96561098e-01 5.62188864e-01
1.39836967e-01 3.64843085e-02 9.42000747e-01 -5.44637382e-01
-7.99663246e-01 -4.61519927e-01 -6.63979471e-01 -2.85365209e-02
1.08186016e-02 3.75986427e-01 -6.45733893e-01 -2.62742937e-01
-4.09701169e-01 -1.84407964e-01 9.72664893e-01 2.90771902e-01
1.47355199e+00 -3.08012567e-03 -3.78076941e-01 8.96420419e-01
1.52000904e+00 -3.02901492e-02 5.28838933e-01 6.63072646e-01
1.39804912e+00 2.06159472e-01 6.04719162e-01 6.97083533e-01
4.94002819e-01 6.98165894e-01 3.87118876e-01 -4.10129949e-02
-2.27611303e-01 -2.97800541e-01 1.51837394e-01 7.53630042e-01
3.72047067e-01 -6.86715961e-01 -1.09155881e+00 3.39532465e-01
-1.75338387e+00 -7.85662830e-01 -6.05786085e-01 1.80928171e+00
4.38133031e-01 4.11225110e-01 -7.33685791e-02 -1.78798847e-03
9.75909650e-01 3.32697958e-01 -6.17154717e-01 2.11301316e-02
-4.21559989e-01 -3.31353128e-01 7.17589974e-01 1.23627335e-01
-1.47947598e+00 1.33134425e+00 5.20929480e+00 1.24643528e+00
-1.20674276e+00 -1.70317262e-01 6.99952126e-01 1.58468261e-01
1.24418087e-01 -1.86150715e-01 -1.11154056e+00 5.81932604e-01
6.56308234e-02 1.18788950e-01 2.43695408e-01 8.13260615e-01
2.37952083e-01 4.94351052e-02 -5.95429480e-01 1.25587749e+00
3.13260049e-01 -1.16370749e+00 2.42185652e-01 -1.04508244e-01
8.08329105e-01 3.71749222e-01 1.87190920e-02 2.83178449e-01
-6.42838776e-02 -6.33344591e-01 9.17639971e-01 -6.30212389e-03
8.05741131e-01 -6.99811101e-01 4.84361649e-01 4.67556208e-01
-1.49265301e+00 1.04755200e-01 -5.16388059e-01 3.62994224e-01
6.33467957e-02 5.20747066e-01 -7.63430297e-01 6.63140357e-01
7.14289010e-01 8.63722563e-01 -1.03589976e+00 1.05015862e+00
-6.83262646e-02 7.33640671e-01 -4.58821058e-01 -3.32147896e-01
3.11247170e-01 -1.50466293e-01 4.35966492e-01 1.53299522e+00
2.61910707e-01 -2.36071870e-01 4.16496783e-01 6.30541444e-01
-3.00733179e-01 4.31548178e-01 -2.39313588e-01 -9.56257433e-02
2.35457554e-01 1.41098905e+00 -1.22498906e+00 -3.42413992e-01
-5.12231529e-01 1.30397737e+00 1.07561618e-01 1.46629244e-01
-1.09755671e+00 -4.35691029e-01 3.29858158e-04 -5.60259856e-02
5.61357081e-01 -4.04076874e-01 -4.64639693e-01 -1.42800927e+00
1.79753870e-01 -8.47166359e-01 2.35487580e-01 -8.13119173e-01
-1.07531869e+00 4.83525544e-01 -2.29568139e-01 -1.22246933e+00
3.74955982e-01 -7.46568918e-01 -7.75372803e-01 5.41344166e-01
-1.44289005e+00 -1.61862230e+00 -4.76935357e-01 4.64361578e-01
1.17762780e+00 -8.84169042e-02 3.80017072e-01 4.29519147e-01
-1.06779265e+00 7.10301995e-01 4.89749461e-01 6.61297023e-01
8.25569272e-01 -1.06318104e+00 8.03502858e-01 1.01918113e+00
-5.66565096e-02 1.51721001e-01 6.07885182e-01 -8.47130656e-01
-1.54156172e+00 -1.34825802e+00 3.94672602e-01 -2.73620188e-01
6.77693963e-01 -6.67216003e-01 -8.39323759e-01 5.55592358e-01
2.28817135e-01 -7.79393241e-02 1.35506138e-01 -2.35257640e-01
-3.24582130e-01 1.41143426e-02 -6.91764414e-01 7.96142638e-01
8.38510871e-01 -2.33474359e-01 -2.33973190e-01 6.19664192e-01
6.45378411e-01 -8.73469293e-01 -3.70496064e-01 2.48867840e-01
4.36072558e-01 -7.77411997e-01 8.79237354e-01 2.70585790e-02
5.53316355e-01 -3.17149699e-01 -3.03999577e-02 -6.75960183e-01
-5.59785068e-02 -4.54238802e-01 1.01356000e-01 1.46808136e+00
7.63093084e-02 -4.15828466e-01 9.86284137e-01 1.21207371e-01
-2.41898343e-01 -5.93197048e-01 -6.78177476e-01 -4.71610218e-01
1.13676727e-01 -4.26492602e-01 4.24257278e-01 8.43755007e-01
-5.87624073e-01 4.19360012e-01 -5.55933952e-01 2.15266898e-01
6.57621741e-01 2.59059072e-01 1.11725819e+00 -1.13237667e+00
-2.48712078e-01 -6.26918614e-01 -3.12631726e-01 -1.54861748e+00
-3.96095701e-02 -7.53860056e-01 1.18519086e-02 -1.56031036e+00
4.75657046e-01 -1.52984560e-01 2.94801444e-01 3.14179420e-01
-5.22188842e-01 3.98621820e-02 3.13225955e-01 2.84925520e-01
-7.91888237e-01 7.34579563e-01 1.24246538e+00 -2.76949108e-01
5.16485237e-03 -1.18753359e-01 -3.17226470e-01 9.11786079e-01
7.69502640e-01 -4.39251155e-01 -9.54650864e-02 -9.02519763e-01
3.39190781e-01 -1.33145347e-01 3.03821921e-01 -9.61619079e-01
4.10168916e-01 -5.51912608e-03 6.24035299e-01 -1.45014548e+00
1.59080312e-01 -7.92150974e-01 -4.69385296e-01 3.37599039e-01
-2.23596558e-01 9.74566266e-02 4.29738283e-01 6.96423531e-01
5.08370548e-02 -2.35332668e-01 7.08481848e-01 3.10892016e-01
-6.07949018e-01 5.19926131e-01 -1.92947403e-01 1.39444411e-01
6.87716782e-01 -1.93783253e-01 -6.37041867e-01 6.79119816e-03
2.02301994e-01 3.76479417e-01 4.50230569e-01 6.34841561e-01
5.67719817e-01 -9.47853208e-01 -8.76171231e-01 -8.53704810e-02
2.41135508e-01 5.43264031e-01 6.25950694e-01 6.56749308e-01
-9.78199422e-01 4.10713166e-01 2.85282940e-01 -8.46112251e-01
-1.35578084e+00 4.51780170e-01 2.13812366e-01 -4.11800593e-01
-1.01941156e+00 4.76001292e-01 3.29939425e-01 -4.28149194e-01
2.98703820e-01 -4.31752443e-01 7.78014585e-02 -3.29221547e-01
3.48271310e-01 3.31728101e-01 7.45961145e-02 -6.56230867e-01
-1.89613193e-01 1.00415456e+00 -5.01783490e-01 3.39591235e-01
1.20689046e+00 -2.97907233e-01 9.79945436e-02 1.75302327e-01
1.13473833e+00 1.10500701e-01 -1.20363498e+00 -2.88907319e-01
-1.94153041e-01 -5.07909238e-01 1.06420159e-01 -5.95119774e-01
-9.83279288e-01 1.08822584e+00 5.12106597e-01 -1.00919893e-02
8.40956688e-01 -3.42557758e-01 1.11717355e+00 7.05724657e-01
-4.25200127e-02 -1.25225627e+00 3.20069551e-01 6.05898559e-01
6.57747149e-01 -1.41544127e+00 3.14439386e-01 -6.53859138e-01
-5.29100418e-01 1.40086353e+00 7.71634698e-01 -8.66675228e-02
2.36126527e-01 2.65689820e-01 -5.25364317e-02 -4.87236902e-02
-1.26759022e-01 4.29829136e-02 2.09342048e-01 5.14350273e-02
4.42021430e-01 -2.21871495e-01 9.28574502e-02 1.67567402e-01
2.93849949e-02 -3.48833114e-01 4.73306119e-01 8.23491335e-01
-4.65742290e-01 -9.17067170e-01 -5.53640842e-01 5.99576175e-01
-7.08849788e-01 -2.58113980e-01 -2.95639694e-01 9.35915351e-01
-1.68470472e-01 7.55742908e-01 4.87875380e-03 -2.85002049e-02
3.45775813e-01 -2.93560088e-01 3.07593435e-01 -4.56661850e-01
-3.83940876e-01 4.37225997e-01 -1.77391618e-01 1.03420109e-01
-7.68107846e-02 -7.47687459e-01 -1.50200748e+00 -2.40938872e-01
-7.80591309e-01 -5.21391511e-01 7.91940451e-01 9.11095083e-01
1.73451722e-01 6.60950243e-01 6.13432825e-01 -8.14431071e-01
-5.36913097e-01 -1.15857077e+00 -4.12007034e-01 3.24086130e-01
5.62776215e-02 -3.06472838e-01 -2.15990230e-01 2.29061067e-01] | [12.052302360534668, 2.261716842651367] |
f57eb06a-3a3f-46de-a0a3-91c4e68c7127 | sentiment-analysis-in-finance-from | 2306.03997 | null | https://arxiv.org/abs/2306.03997v1 | https://arxiv.org/pdf/2306.03997v1.pdf | Sentiment Analysis in Finance: From Transformers Back to eXplainable Lexicons (XLex) | Lexicon-based sentiment analysis (SA) in finance leverages specialized, manually annotated lexicons created by human experts to extract sentiment from financial texts. Although lexicon-based methods are simple to implement and fast to operate on textual data, they require considerable manual annotation efforts to create, maintain, and update the lexicons. These methods are also considered inferior to the deep learning-based approaches, such as transformer models, which have become dominant in various NLP tasks due to their remarkable performance. However, transformers require extensive data and computational resources for both training and testing. Additionally, they involve significant prediction times, making them unsuitable for real-time production environments or systems with limited processing capabilities. In this paper, we introduce a novel methodology named eXplainable Lexicons (XLex) that combines the advantages of both lexicon-based methods and transformer models. We propose an approach that utilizes transformers and SHapley Additive exPlanations (SHAP) for explainability to learn financial lexicons. Our study presents four main contributions. Firstly, we demonstrate that transformer-aided explainable lexicons can enhance the vocabulary coverage of the benchmark Loughran-McDonald (LM) lexicon, reducing the human involvement in annotating, maintaining, and updating the lexicons. Secondly, we show that the resulting lexicon outperforms the standard LM lexicon in SA of financial datasets. Thirdly, we illustrate that the lexicon-based approach is significantly more efficient in terms of model speed and size compared to transformers. Lastly, the XLex approach is inherently more interpretable than transformer models as lexicon models rely on predefined rules, allowing for better insights into the results of SA and making the XLex approach a viable tool for financial decision-making. | ['Dimitar Trajanov', 'Milos Jovanovik', 'Kostadin Mishev', 'Hristijan Peshov', 'Maryan Rizinski'] | 2023-06-06 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-1.93889409e-01 2.39349321e-01 -1.13262720e-01 -4.71823186e-01
-6.50422931e-01 -9.19835508e-01 5.18975675e-01 4.81140077e-01
-2.01654151e-01 6.42396033e-01 1.18174041e-02 -7.48378754e-01
-4.38261367e-02 -8.74692082e-01 -6.27495706e-01 -1.86758369e-01
2.32963040e-01 8.29800904e-01 -2.50183851e-01 -5.28979659e-01
2.78800815e-01 3.99301231e-01 -1.60543835e+00 3.31008911e-01
7.41707563e-01 1.23720789e+00 4.84444350e-02 2.17905968e-01
-5.85809469e-01 1.02156758e+00 -6.08209014e-01 -1.10249400e+00
2.48803496e-01 -3.78424406e-01 -5.35664797e-01 1.12992190e-01
-1.83093220e-01 -1.88954100e-02 3.82603765e-01 8.05534184e-01
1.98899046e-01 -1.57921344e-01 6.06381416e-01 -1.14150977e+00
-7.60327637e-01 8.42694223e-01 -1.74248666e-01 -7.30573609e-02
2.49427363e-01 -2.62310296e-01 1.34209394e+00 -1.10728383e+00
3.21491867e-01 8.96293402e-01 9.36803102e-01 2.19795063e-01
-9.16985512e-01 -7.16646075e-01 2.84963787e-01 1.73971713e-01
-9.96032596e-01 -2.95050561e-01 6.77558601e-01 -4.09746110e-01
1.37093818e+00 3.71728867e-01 9.29020643e-01 6.87169671e-01
2.59812146e-01 7.24877596e-01 1.10241640e+00 -7.90992260e-01
2.89999753e-01 6.82770610e-01 1.18585229e-01 7.81729043e-01
6.79266930e-01 -2.57836282e-01 -6.59826875e-01 1.15214780e-01
6.19546473e-01 -1.28939003e-01 -1.42059298e-02 -3.14434737e-01
-8.61173987e-01 1.37504888e+00 -9.97132212e-02 3.49435687e-01
-4.82362956e-01 -1.77776024e-01 5.34952939e-01 4.31633204e-01
8.53966057e-01 9.77689385e-01 -9.89350617e-01 -9.46843848e-02
-1.03093052e+00 1.88651457e-01 1.03055072e+00 8.76855254e-01
6.79828763e-01 1.44606873e-01 4.04823005e-01 4.79044378e-01
3.52664471e-01 3.49132866e-01 8.43960643e-01 -4.55538750e-01
4.94330615e-01 1.10722780e+00 3.44998352e-02 -1.15981448e+00
-5.97545028e-01 -6.17371321e-01 -6.41027451e-01 -1.29160807e-01
3.39486539e-01 6.56333938e-02 -6.78871155e-01 1.39744115e+00
6.05329918e-03 -3.17015201e-01 3.55824083e-01 4.81095850e-01
8.17120910e-01 4.00772095e-01 1.44273475e-01 -2.57526457e-01
1.47869742e+00 -9.34863508e-01 -7.42085278e-01 -4.05913323e-01
8.36893618e-01 -8.70709538e-01 1.32913888e+00 6.35417879e-01
-1.05578816e+00 -4.19301480e-01 -1.07745790e+00 -3.61091495e-02
-7.68257141e-01 1.96828276e-01 1.32672608e+00 1.02753043e+00
-9.72901285e-01 4.10280496e-01 -7.42746770e-01 -1.05103828e-01
2.21685037e-01 6.10158741e-01 -3.45688850e-01 4.58073199e-01
-1.34685791e+00 1.10313118e+00 3.80259931e-01 5.27044795e-02
-1.94284111e-01 -5.02811611e-01 -1.15761304e+00 2.75095433e-01
3.80303591e-01 -4.47114944e-01 1.41918397e+00 -1.12625670e+00
-1.55604315e+00 8.09620500e-01 -1.26206279e-01 -8.36092174e-01
2.68670619e-01 -1.02676071e-01 -2.38816038e-01 -1.31916553e-01
1.43708289e-01 3.05855036e-01 4.45057124e-01 -9.88599479e-01
-5.01351655e-01 6.40776977e-02 3.04178178e-01 8.83576795e-02
-5.69815814e-01 -5.31849265e-02 -4.70695257e-01 -8.91990662e-01
1.32193089e-01 -7.56001890e-01 -2.77083158e-01 -5.87263644e-01
-3.48018825e-01 -1.01311296e-01 1.80795059e-01 -5.97497940e-01
1.29014325e+00 -1.69975805e+00 -1.90311790e-01 3.99241775e-01
2.44084597e-01 2.41635382e-01 1.71714529e-01 3.88255686e-01
-1.99086726e-01 3.58850092e-01 -6.86671883e-02 -6.35891736e-01
3.09097141e-01 3.02036732e-01 -4.02831137e-01 -8.53557885e-02
2.91210920e-01 1.02902961e+00 -5.57034016e-01 -4.15784925e-01
4.92966563e-01 3.80677342e-01 -7.11506128e-01 -1.50603190e-01
-2.47057095e-01 2.88680851e-01 -3.28841627e-01 6.65446401e-01
2.10170493e-01 -3.59948814e-01 4.90410239e-01 -1.65528402e-01
-1.54662346e-02 4.94753599e-01 -1.10750115e+00 1.06216896e+00
-8.73580456e-01 5.53725839e-01 -6.67735994e-01 -1.14337850e+00
1.13764989e+00 3.52923214e-01 3.26250464e-01 -7.79689908e-01
5.13715088e-01 5.65340161e-01 -3.71654123e-01 -2.49802470e-01
6.75648510e-01 -5.07412136e-01 -3.47341329e-01 7.89476693e-01
3.61605696e-02 -4.68936920e-01 5.29355466e-01 3.57886367e-02
6.52935147e-01 -5.52958474e-02 7.44571269e-01 -1.10796764e-01
7.77470171e-01 1.87769264e-01 4.25088525e-01 5.56191266e-01
2.40971848e-01 2.87035286e-01 6.90900922e-01 -8.33962381e-01
-9.80499744e-01 -4.95579660e-01 -1.31813839e-01 8.96474957e-01
-1.49542883e-01 -5.47591865e-01 -8.05828631e-01 -7.53891468e-01
-1.29964590e-01 8.76653135e-01 -4.89751786e-01 6.54808506e-02
-3.87533486e-01 -9.15687561e-01 3.41719121e-01 6.10345840e-01
3.32357943e-01 -1.13776541e+00 -8.02343011e-01 3.48188043e-01
-3.32895935e-01 -1.15518308e+00 -8.80342796e-02 5.63488305e-01
-8.16314518e-01 -1.16938722e+00 -2.97761500e-01 -5.68638444e-01
6.23258889e-01 -2.49622419e-01 1.42569923e+00 7.43349940e-02
1.13735162e-01 2.19793618e-01 -6.72484100e-01 -9.31817234e-01
-4.60467696e-01 3.04712981e-01 7.48446137e-02 -1.34762719e-01
8.59188795e-01 -1.22716300e-01 -1.53853729e-01 1.37656987e-01
-1.00358856e+00 2.90799141e-01 7.04780579e-01 7.85665512e-01
7.59886324e-01 4.39957380e-01 7.68818140e-01 -1.25473976e+00
8.86661232e-01 -3.59587103e-01 -7.17679381e-01 3.99539471e-01
-1.11391735e+00 2.17467993e-01 7.17232823e-01 -1.70244798e-01
-9.07793641e-01 1.35346815e-01 -5.29076569e-02 5.84827848e-02
2.67893612e-01 1.11168063e+00 -5.87716550e-02 -1.54100865e-01
4.87911612e-01 7.84272607e-03 -3.53066288e-02 -5.31328440e-01
2.02914953e-01 6.11455977e-01 2.27043644e-01 -2.60551363e-01
6.65055752e-01 1.90684512e-01 -1.09719865e-01 -2.13925138e-01
-1.13934302e+00 -2.45194077e-01 -7.80125856e-01 -9.41421539e-02
6.25639081e-01 -8.27762127e-01 -5.21813571e-01 2.73937106e-01
-9.56289411e-01 -8.46996084e-02 -4.14489746e-01 4.86092359e-01
-4.91274327e-01 1.88931078e-01 -5.15209973e-01 -8.14547837e-01
-4.66382980e-01 -1.15091217e+00 1.00661278e+00 9.03247222e-02
-6.70851409e-01 -1.41276264e+00 -3.26867998e-01 5.59482813e-01
3.34103018e-01 3.28214288e-01 1.21779048e+00 -1.22436321e+00
-2.04759732e-01 -5.45274854e-01 -7.69884437e-02 4.47480232e-01
1.15141518e-01 -2.41318390e-01 -8.47165823e-01 5.64301983e-02
2.20661446e-01 -3.16113204e-01 4.41802084e-01 3.02567154e-01
8.87249231e-01 -2.73585647e-01 1.65035591e-01 3.33967924e-01
1.31998539e+00 2.47294798e-01 2.69224077e-01 9.19147432e-01
4.21411246e-01 8.07788908e-01 7.86584854e-01 4.63502824e-01
6.93613291e-01 5.99278927e-01 1.86509594e-01 -2.93719769e-01
2.98864067e-01 -1.11471727e-01 4.11960870e-01 1.27211189e+00
-1.43293619e-01 -2.37342387e-01 -9.82788205e-01 4.21281308e-01
-1.82182717e+00 -5.05035758e-01 -2.43331790e-01 2.02899957e+00
8.92279088e-01 4.64280665e-01 -2.04769462e-01 6.04026258e-01
2.83027530e-01 -3.36887330e-01 -3.82523894e-01 -7.83580899e-01
-3.69855672e-01 7.06264615e-01 5.45957029e-01 4.53409255e-01
-1.08884180e+00 1.16686440e+00 6.30652475e+00 5.56558013e-01
-1.01330173e+00 2.08994895e-01 7.10270524e-01 -9.85662192e-02
-5.18771529e-01 -1.33543909e-01 -7.65106320e-01 9.59701315e-02
1.10510302e+00 -3.27556700e-01 2.35144362e-01 1.07427227e+00
2.29355812e-01 3.60109173e-02 -1.12968624e+00 8.64546120e-01
1.47541821e-01 -1.41047812e+00 3.47131103e-01 -9.76823866e-02
6.28457606e-01 -3.85970294e-01 1.92777187e-01 4.81535405e-01
1.96072683e-01 -1.15508127e+00 1.15167451e+00 2.63822913e-01
6.67946398e-01 -1.01071525e+00 1.52640176e+00 9.80664492e-02
-1.24556494e+00 -6.31737560e-02 -4.11196828e-01 -4.08132344e-01
1.44312039e-01 6.37483656e-01 -6.84874773e-01 6.27587020e-01
5.30209184e-01 5.88842213e-01 -7.63933957e-01 4.07467633e-01
-3.88704300e-01 4.55317318e-01 -2.44009998e-02 -1.00595087e-01
4.14350927e-01 -1.37994796e-01 -1.80192031e-02 9.84803677e-01
4.07330245e-01 -1.24824978e-01 1.46535896e-02 5.99640965e-01
2.12929752e-02 5.81766188e-01 -4.13649738e-01 -2.97620535e-01
2.34737337e-01 1.17303360e+00 -1.22404754e+00 -5.16700089e-01
-5.18046319e-01 5.67513764e-01 1.06901921e-01 3.70884500e-02
-8.68487179e-01 -2.39001721e-01 3.01093787e-01 4.45448793e-02
3.80544871e-01 9.88574028e-02 -7.36326337e-01 -1.23412287e+00
1.16287932e-01 -1.25012290e+00 3.03225875e-01 -6.63650393e-01
-1.18281066e+00 1.00318301e+00 2.88037267e-02 -1.21346772e+00
-5.96698165e-01 -9.85374212e-01 -2.03997511e-02 7.99086392e-01
-1.67431319e+00 -1.07992113e+00 -7.42832571e-02 4.76826459e-01
7.00316370e-01 -4.58050638e-01 1.11869764e+00 2.55231172e-01
-4.67894644e-01 5.49858093e-01 -4.68446538e-02 -5.73775060e-02
4.48208094e-01 -1.39207244e+00 5.76157153e-01 6.03931606e-01
5.17535210e-01 8.56809139e-01 7.42793143e-01 -5.36878884e-01
-1.00233459e+00 -9.61700559e-01 1.57233989e+00 -6.22676969e-01
8.60384226e-01 -4.25065219e-01 -7.10820794e-01 8.45485330e-01
-9.20707732e-02 -5.07103145e-01 1.10387886e+00 2.80847877e-01
-1.82196528e-01 8.91543925e-02 -9.45195675e-01 4.45935905e-01
4.51545686e-01 -5.14462411e-01 -7.04609871e-01 4.78894502e-01
5.28393507e-01 -3.85828078e-01 -7.58893907e-01 2.79561043e-01
5.79471588e-01 -1.18280280e+00 6.93921506e-01 -4.06454057e-01
4.68204588e-01 -2.31000520e-02 -1.91089772e-02 -1.08773637e+00
-1.28618826e-03 -4.05398399e-01 6.78606108e-02 1.19307327e+00
9.50865507e-01 -9.74470794e-01 7.82929122e-01 7.98636317e-01
-2.03436632e-02 -1.03029776e+00 -7.37627268e-01 -6.61789596e-01
5.94404191e-02 -9.51666594e-01 1.10229063e+00 1.02956533e+00
1.37875244e-01 3.08001250e-01 -1.90245807e-01 -1.94706023e-01
1.29466951e-01 1.74626425e-01 6.75385177e-01 -1.41620636e+00
-3.45037848e-01 -4.87958908e-01 -3.58258396e-01 -6.32013619e-01
2.59609699e-01 -8.07746112e-01 -1.87626451e-01 -1.46438408e+00
3.37738954e-02 -5.99848211e-01 -3.26061904e-01 7.39924073e-01
3.59153338e-02 4.92326766e-01 1.28611580e-01 4.07063335e-01
-3.89885426e-01 2.20523715e-01 8.90938878e-01 2.13181674e-02
-2.43754402e-01 -7.27031529e-02 -1.20589340e+00 1.08096266e+00
9.14647758e-01 -5.22703826e-01 -6.12908959e-01 -2.72524863e-01
7.72662580e-01 -3.21186930e-01 3.87747772e-02 -6.31644070e-01
5.12191802e-02 1.77057266e-01 2.92473972e-01 -4.33055967e-01
3.00410409e-02 -7.84676671e-01 2.49663487e-01 4.42617625e-01
-1.42179251e-01 5.24506271e-01 4.88797694e-01 2.50075012e-01
-5.54792285e-01 -4.73673165e-01 2.96924233e-01 -3.10171217e-01
-5.73304117e-01 -1.22326426e-01 -5.92536032e-01 -1.25714898e-01
9.84233379e-01 -3.80726129e-01 2.22154129e-02 -5.25827944e-01
-6.03516400e-01 -1.24393120e-01 2.72536993e-01 2.98208475e-01
2.37592340e-01 -1.00677431e+00 -2.98531204e-01 3.25756997e-01
-9.06017236e-03 -5.65285571e-02 -1.99279830e-01 8.41390073e-01
-9.25303102e-01 1.04160094e+00 -4.66553532e-02 -2.04110458e-01
-1.07322824e+00 4.30437386e-01 1.74812973e-01 -1.00449657e+00
-3.65819007e-01 9.57974672e-01 1.98847085e-01 -6.53804660e-01
4.43095155e-02 -6.85333014e-01 -5.71738839e-01 3.13214004e-01
2.98932165e-01 -1.00074612e-01 4.73418951e-01 -7.32819676e-01
-2.57943571e-01 6.20694935e-01 -1.34774357e-01 -5.72505221e-02
1.47646439e+00 -1.16193548e-01 -6.53859749e-02 4.95483071e-01
5.24506271e-01 1.69618860e-01 -6.58798039e-01 -4.74690087e-02
4.01393622e-01 -7.05802217e-02 9.32370052e-02 -8.66482437e-01
-1.10913897e+00 7.75462329e-01 -7.65472790e-03 5.62998712e-01
1.36410463e+00 -1.87617287e-01 8.52203250e-01 3.52709025e-01
4.71435726e-01 -1.03261137e+00 -1.05687954e-01 4.60455298e-01
6.86791062e-01 -1.16425633e+00 1.25101671e-01 -5.31069875e-01
-9.92774427e-01 1.30349672e+00 1.41372845e-01 3.00698608e-01
5.86585522e-01 3.73616993e-01 4.75722939e-01 -4.68486577e-01
-6.27281189e-01 -1.77492514e-01 4.41408604e-01 2.76794463e-01
6.17356122e-01 -2.05492042e-02 -3.39019358e-01 1.28910506e+00
-1.02849913e+00 -9.84625295e-02 4.62600112e-01 7.21135855e-01
-1.59733221e-01 -1.45569289e+00 -2.47453481e-01 5.51145196e-01
-6.94292724e-01 -5.47612846e-01 -6.90189421e-01 1.10083091e+00
2.52135158e-01 1.17239606e+00 -2.64486875e-02 -3.81029814e-01
4.26120371e-01 3.28807831e-01 1.09389536e-01 -8.56740952e-01
-9.81691539e-01 -3.40269245e-02 2.02337563e-01 -3.09340447e-01
-5.53718030e-01 -6.88552320e-01 -1.14882326e+00 -4.05113250e-01
-6.81694388e-01 4.86447066e-01 7.38535464e-01 1.45395064e+00
1.85251340e-01 5.04495561e-01 3.75452012e-01 -5.68816245e-01
-2.94716567e-01 -8.02240491e-01 -5.79806864e-01 1.76880226e-01
-1.45743176e-01 -7.74032652e-01 -4.91284192e-01 3.03935260e-01] | [11.1276273727417, 6.9837541580200195] |
46d42675-55b0-4ec0-82e5-9be159638a4c | frustratingly-easy-cross-lingual-transfer-for | null | null | https://aclanthology.org/N16-1121 | https://aclanthology.org/N16-1121.pdf | Frustratingly Easy Cross-Lingual Transfer for Transition-Based Dependency Parsing | null | ['Fran{\\c{c}}ois Yvon', "Oph{\\'e}lie Lacroix", 'Lauriane Aufrant', 'Guillaume Wisniewski'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['transition-based-dependency-parsing'] | ['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] | [-7.338669776916504, 3.7057290077209473] |
6eb7cccf-6668-4d51-9e82-aa173f26489f | turing-an-accurate-and-interpretable-multi | 2106.04559 | null | https://arxiv.org/abs/2106.04559v1 | https://arxiv.org/pdf/2106.04559v1.pdf | Turing: an Accurate and Interpretable Multi-Hypothesis Cross-Domain Natural Language Database Interface | A natural language database interface (NLDB) can democratize data-driven insights for non-technical users. However, existing Text-to-SQL semantic parsers cannot achieve high enough accuracy in the cross-database setting to allow good usability in practice. This work presents Turing, a NLDB system toward bridging this gap. The cross-domain semantic parser of Turing with our novel value prediction method achieves $75.1\%$ execution accuracy, and $78.3\%$ top-5 beam execution accuracy on the Spider validation set. To benefit from the higher beam accuracy, we design an interactive system where the SQL hypotheses in the beam are explained step-by-step in natural language, with their differences highlighted. The user can then compare and judge the hypotheses to select which one reflects their intention if any. The English explanations of SQL queries in Turing are produced by our high-precision natural language generation system based on synchronous grammars. | ['Yanshuai Cao', 'Meidan Alon', 'Harsh Barot', 'Jawad Ateeq', 'Wei Yang', 'Keyi Tang', 'Ákos Kádár', 'Hamidreza Shahidi', 'Wenjie Zi', 'Peng Xu'] | 2021-06-08 | null | https://aclanthology.org/2021.acl-demo.36 | https://aclanthology.org/2021.acl-demo.36.pdf | acl-2021-5 | ['value-prediction'] | ['computer-code'] | [-2.16936588e-01 7.17507184e-01 -1.82994887e-01 -7.94499695e-01
-1.14185774e+00 -8.00470173e-01 3.02913010e-01 1.21841729e-01
1.12754941e-01 5.45282841e-01 1.32948741e-01 -8.02865028e-01
-1.33416653e-01 -1.14305890e+00 -6.51333749e-01 2.44253322e-01
1.50297508e-01 9.91627216e-01 4.05701071e-01 -5.03311574e-01
3.69472831e-01 1.34637520e-01 -1.81761026e+00 8.12052906e-01
8.84837806e-01 8.28433037e-01 2.64191747e-01 6.34628177e-01
-9.67744589e-01 8.50865364e-01 -6.54979765e-01 -7.26784408e-01
2.24955752e-01 -4.24169749e-01 -9.70504522e-01 -5.77153921e-01
1.18489601e-01 -1.45779148e-01 4.64520127e-01 7.37031639e-01
2.48482987e-01 -4.71720308e-01 2.62123104e-02 -1.50968826e+00
-4.67519969e-01 9.72615898e-01 2.40173545e-02 -2.57643431e-01
1.01039231e+00 1.65572658e-01 1.43698418e+00 -8.23741853e-01
1.12826288e+00 1.31340861e+00 4.17578310e-01 6.41440213e-01
-1.39316785e+00 -4.48734373e-01 -2.48226359e-01 -2.66510770e-02
-1.16045916e+00 -2.68657655e-01 5.86332262e-01 -3.32133502e-01
1.41952670e+00 5.46005487e-01 4.67523336e-01 8.22357833e-01
-3.66964452e-02 2.93168038e-01 9.55187261e-01 -6.99134350e-01
2.28859112e-01 6.87728167e-01 2.58680910e-01 9.60821152e-01
2.28452936e-01 4.63732071e-02 -1.00439227e+00 -2.56538033e-01
4.82990503e-01 -5.45326412e-01 1.80910289e-01 -4.62193638e-01
-8.44019890e-01 8.34509850e-01 -1.83369499e-02 1.09762318e-01
-2.27676526e-01 -1.50004879e-01 3.26632768e-01 4.90109682e-01
-2.91292858e-03 8.23589504e-01 -9.68321085e-01 -6.33090317e-01
-5.37625968e-01 7.53602684e-01 1.40163994e+00 1.60417950e+00
6.79069221e-01 -2.94112802e-01 1.37541100e-01 7.58500278e-01
3.63737136e-01 5.28342128e-01 2.67490000e-01 -1.01805305e+00
4.97335047e-01 1.12237561e+00 1.97772384e-01 -8.36241722e-01
-3.89432907e-01 -1.19888872e-01 4.55583520e-02 2.65707374e-01
5.83397925e-01 2.30635211e-01 -3.91882986e-01 1.69326246e+00
3.24102461e-01 -1.04550397e+00 4.94413614e-01 8.80669415e-01
7.95049727e-01 4.71765906e-01 2.11679459e-01 8.75486434e-02
1.69349670e+00 -1.96303681e-01 -4.60172802e-01 -3.48021269e-01
1.04101682e+00 -6.02104962e-01 1.85502589e+00 4.46974009e-01
-1.20004833e+00 -5.93170643e-01 -1.10938990e+00 -2.95603722e-01
-5.47142625e-01 -2.20108151e-01 5.80788970e-01 6.63490593e-01
-7.82790244e-01 3.47159147e-01 -7.33830631e-01 -5.99870622e-01
-1.07684337e-01 1.48479715e-01 -8.80007520e-02 7.67889693e-02
-1.19716084e+00 8.85363817e-01 3.89318615e-01 -6.58424079e-01
-1.03847325e-01 -9.98025358e-01 -8.04103136e-01 1.60615683e-01
6.20878160e-01 -7.38414705e-01 1.72686338e+00 -4.00512546e-01
-1.23469555e+00 1.07984507e+00 -1.59472719e-01 -5.73030174e-01
4.73052293e-01 -1.55323595e-01 -3.82447571e-01 -2.35731676e-02
5.72621167e-01 8.19829941e-01 -1.41252980e-01 -1.10373580e+00
-8.28119934e-01 -6.62677765e-01 2.36280784e-01 -8.87953416e-02
1.12816699e-01 8.84827599e-02 -4.28518027e-01 -2.30312154e-01
2.76207775e-01 -5.16064882e-01 2.02886358e-01 -2.10653603e-01
-3.96854043e-01 -2.97564656e-01 4.21614498e-01 -6.79981172e-01
1.36652720e+00 -1.95974290e+00 -3.73782396e-01 2.48954371e-01
4.79169786e-02 -3.78950447e-01 2.01639190e-01 6.86744928e-01
8.47086608e-02 5.43919563e-01 1.01885200e-01 2.88828909e-01
4.93999094e-01 8.59542266e-02 -5.16856909e-01 -5.87069452e-01
4.28565741e-01 6.12921536e-01 -6.03084862e-01 -7.99034178e-01
-1.05073154e-01 -1.66808799e-01 -7.98495054e-01 5.58758199e-01
-8.44989121e-01 -2.27063581e-01 -5.53211987e-01 7.81007707e-01
4.01300937e-01 -1.83236316e-01 6.36497915e-01 -1.24376512e-03
-2.16698229e-01 8.66927624e-01 -9.88712072e-01 1.82967818e+00
-6.42920136e-01 2.95767516e-01 -1.35093197e-01 -4.95576859e-01
1.50811291e+00 -7.77405128e-02 7.49664977e-02 -1.20862544e+00
-5.37105620e-01 4.91638750e-01 -4.92923968e-02 -8.39172184e-01
4.86330301e-01 -2.80088365e-01 -6.13260090e-01 5.36596596e-01
-2.19258308e-01 -4.85847563e-01 3.94578487e-01 4.43562388e-01
1.23150945e+00 4.47245508e-01 3.95682067e-01 -4.88788038e-01
1.78901315e-01 8.91659260e-01 4.90179032e-01 7.92939484e-01
2.87072450e-01 3.89871001e-01 1.07219446e+00 -6.33941710e-01
-1.21106994e+00 -1.06771052e+00 5.65346181e-02 1.23026276e+00
1.21016622e-01 -8.15119982e-01 -8.57001483e-01 -5.93313873e-01
8.79894868e-02 1.56190336e+00 -1.69037104e-01 1.84832271e-02
-3.77000898e-01 1.79731566e-02 5.40214956e-01 5.35156548e-01
2.75532216e-01 -1.12707889e+00 -1.19406188e+00 3.22292119e-01
-3.12800676e-01 -1.04111481e+00 2.15932727e-01 1.97014391e-01
-6.96623623e-01 -1.16037118e+00 5.82733929e-01 -4.76277232e-01
1.12593569e-01 -2.72690773e-01 1.67220557e+00 1.11639328e-01
-3.18337172e-01 1.20851777e-01 -4.58085209e-01 -6.61096394e-01
-9.73832548e-01 1.90160140e-01 -3.88785839e-01 -9.67661560e-01
9.90618050e-01 -3.33206207e-01 -2.00380817e-01 4.10705417e-01
-7.58354843e-01 5.67353725e-01 3.02879721e-01 5.88007390e-01
4.61278290e-01 -2.51909584e-01 5.00986993e-01 -1.16990781e+00
7.68118560e-01 -2.81767428e-01 -7.26800740e-01 5.10306180e-01
-1.02391303e+00 5.48116505e-01 7.10659325e-01 2.15837017e-01
-1.03734100e+00 -8.41657519e-02 -1.24909632e-01 3.45051140e-01
-3.50835055e-01 8.09154510e-01 -5.10206521e-01 7.63251781e-01
1.07978904e+00 -9.98376310e-03 9.56561044e-02 -4.74545866e-01
4.16647881e-01 5.64094067e-01 6.77628458e-01 -9.38900530e-01
4.55675274e-01 -7.99732506e-02 -3.31393093e-01 -2.18793988e-01
-3.34464222e-01 6.77254573e-02 -3.77072662e-01 1.19225889e-01
6.47510052e-01 -5.15870154e-01 -9.00290787e-01 -2.66804039e-01
-1.19122803e+00 -3.58774453e-01 -4.71533775e-01 2.12472794e-03
-8.35125446e-01 -1.64944932e-01 -2.52829671e-01 -9.08268273e-01
-3.50812227e-01 -1.01217377e+00 1.04850602e+00 9.96007621e-02
-9.37067449e-01 -4.03063685e-01 -1.77762195e-01 3.98669332e-01
3.47063780e-01 1.18296497e-01 1.68073177e+00 -1.17832208e+00
-6.42377794e-01 -2.70523936e-01 -1.99387625e-01 -2.92467386e-01
-2.32697919e-01 2.62736708e-01 -5.81886470e-01 4.54233021e-01
-2.35493481e-01 -3.67358834e-01 -2.74436355e-01 -1.63056597e-01
9.52551901e-01 -3.02031010e-01 -3.84504557e-01 1.90517783e-01
1.38435328e+00 5.25957406e-01 5.80843985e-01 6.98984385e-01
3.23509090e-02 8.67238224e-01 9.39514160e-01 4.42646354e-01
7.00773716e-01 8.77725899e-01 2.70973414e-01 2.84118801e-01
1.62427068e-01 -7.41932869e-01 8.82412866e-02 1.11325778e-01
5.62804520e-01 -6.00612052e-02 -1.54230344e+00 4.27644998e-01
-1.56655538e+00 -6.95286095e-01 -1.97411060e-01 1.99259615e+00
9.19113100e-01 7.03157425e-01 1.44771203e-01 -9.57374051e-02
2.87206650e-01 -2.81137615e-01 -4.76935476e-01 -9.12003815e-01
-5.84975965e-02 3.58534962e-01 -9.58939374e-04 4.79941607e-01
-2.67839789e-01 1.00911188e+00 5.57276535e+00 4.09222335e-01
-7.82863796e-01 -1.72717974e-01 3.37857127e-01 -1.31767020e-01
-7.22925127e-01 3.58702272e-01 -8.71958256e-01 1.55833900e-01
1.35783958e+00 -4.87209886e-01 3.31010818e-01 1.26522124e+00
2.73374707e-01 -1.99776947e-01 -1.47386229e+00 7.48617351e-01
-3.07844430e-01 -1.55620813e+00 6.11062488e-03 -1.22437902e-01
-2.60081202e-01 -4.06511307e-01 -3.58708709e-01 4.76238966e-01
4.27448750e-01 -9.43141460e-01 9.64200020e-01 6.02117777e-01
7.90620148e-01 -6.63314879e-01 5.23712158e-01 6.16495430e-01
-7.84120083e-01 -1.75969720e-01 -1.26291111e-01 -8.41028988e-03
6.81996420e-02 3.37803930e-01 -1.34982812e+00 5.90531886e-01
9.18543339e-01 -2.91669164e-02 -6.20892465e-01 1.38122112e-01
-1.71568111e-01 3.44202995e-01 -4.02634531e-01 -5.80093443e-01
-3.53254199e-01 -7.79395476e-02 2.35363647e-01 1.22519839e+00
5.10342121e-01 3.27856362e-01 -2.12644506e-02 1.33381689e+00
1.50180236e-01 2.95372188e-01 -7.50048339e-01 -2.60161668e-01
8.69899213e-01 7.86860049e-01 -3.24446797e-01 -5.71985066e-01
-3.69126529e-01 5.66123068e-01 2.74737239e-01 -2.36467440e-02
-6.52272403e-01 -6.39431059e-01 5.54662466e-01 5.80886185e-01
-6.41167685e-02 -4.91670556e-02 -1.00793779e+00 -7.92269945e-01
5.56608200e-01 -1.41585255e+00 5.32425165e-01 -1.29287386e+00
-9.75336254e-01 7.65370846e-01 2.24312380e-01 -7.64164448e-01
-9.11344349e-01 -5.93175590e-01 -7.66737685e-02 9.43684757e-01
-5.77523768e-01 -9.46980536e-01 -3.66535574e-01 5.23681045e-02
5.10782480e-01 -1.83202937e-01 1.18215191e+00 -9.07137692e-02
-9.62211713e-02 4.96119261e-01 -6.26818419e-01 1.03474781e-02
5.05162597e-01 -1.36797130e+00 8.41529131e-01 6.28130496e-01
3.88463475e-02 1.04049766e+00 1.02853513e+00 -7.82594860e-01
-1.60206831e+00 -6.26284361e-01 1.42806756e+00 -6.64383829e-01
5.61658800e-01 -7.10017502e-01 -1.01612198e+00 4.84012753e-01
2.91439723e-02 -5.26662290e-01 6.22270226e-01 2.54166722e-01
-6.92693710e-01 -2.72986054e-01 -1.30598450e+00 6.61999404e-01
1.17418194e+00 -7.07830846e-01 -8.76020491e-01 2.97565609e-01
1.10829365e+00 -5.48013628e-01 -1.01610672e+00 2.97115296e-01
6.83289289e-01 -1.29853344e+00 6.43597305e-01 -9.44205463e-01
8.53104770e-01 -1.54434085e-01 -5.21567404e-01 -7.45562971e-01
2.71555930e-01 -5.48783541e-01 2.28672981e-01 1.26992595e+00
8.90037835e-01 -6.24615014e-01 9.99978364e-01 1.33007658e+00
-9.83699039e-02 -6.74100816e-01 -5.32859981e-01 -6.27549648e-01
-6.79920539e-02 -1.01068580e+00 1.11584079e+00 5.96408546e-01
7.37908185e-01 3.33669186e-01 3.89729410e-01 1.58283359e-03
3.05554003e-01 5.78027666e-01 1.13677609e+00 -1.25999022e+00
-3.56670141e-01 -3.02718699e-01 -1.98619083e-01 -8.80570173e-01
-1.76453352e-01 -7.86628485e-01 8.48166570e-02 -1.42757046e+00
1.06258750e-01 -6.52142107e-01 5.76280355e-01 6.24521732e-01
2.29426265e-01 -5.17236292e-01 1.09488048e-01 4.68686819e-02
-1.66962638e-01 3.03414632e-02 3.89580876e-01 2.50052541e-01
-3.66409928e-01 -4.17120934e-01 -1.11686444e+00 4.44694906e-01
8.23819160e-01 -6.43834472e-01 -4.75639910e-01 -4.87094790e-01
5.24132729e-01 4.47518915e-01 3.90276343e-01 -8.31856489e-01
1.73406884e-01 -3.83319467e-01 -4.03657742e-03 -5.23058712e-01
-6.95785880e-02 -5.58987081e-01 4.08457190e-01 3.94025892e-01
-6.59591496e-01 4.49550301e-01 2.58717746e-01 -5.11572845e-02
-1.94501445e-01 -2.83345342e-01 4.00816113e-01 -2.53489375e-01
-6.34605825e-01 -3.28116626e-01 -2.35639498e-01 3.08565468e-01
7.50224113e-01 -2.35581949e-01 -6.85836017e-01 -3.00814718e-01
-6.60588801e-01 1.99131921e-01 6.10457718e-01 4.97298241e-01
4.85457689e-01 -9.31161821e-01 -3.27405840e-01 4.47115093e-01
5.81199765e-01 -3.62853669e-02 -2.97220439e-01 3.08950216e-01
-8.17253411e-01 6.99704349e-01 -2.88157672e-01 -4.79321241e-01
-1.08614838e+00 4.46078539e-01 2.48574659e-01 -2.33471259e-01
-3.86650890e-01 6.30688906e-01 -2.13225171e-01 -7.37941921e-01
8.05100985e-03 -4.85108882e-01 2.51022696e-01 -3.44064564e-01
3.67815644e-01 7.72489384e-02 3.08784157e-01 1.28395483e-01
-6.16481006e-01 6.26866519e-02 1.69404551e-01 -7.13239908e-01
1.31745160e+00 9.07394886e-02 -3.77527885e-02 4.26889658e-01
8.79337013e-01 4.47410978e-02 -6.76978052e-01 -5.40186884e-03
6.94820583e-01 -4.39717770e-01 -5.61827004e-01 -1.21139765e+00
-3.41117173e-01 8.13986480e-01 2.79679865e-01 7.36962974e-01
9.75373209e-01 3.31601977e-01 5.09546816e-01 5.04779875e-01
6.88701928e-01 -1.14309978e+00 -1.92107067e-01 1.60025328e-01
1.04978955e+00 -1.05889010e+00 -2.57469594e-01 -5.81062376e-01
-7.42154121e-01 1.37443125e+00 1.00329638e+00 3.60692948e-01
5.07360846e-02 6.44012749e-01 3.75026941e-01 -6.17205441e-01
-1.20789969e+00 7.67310560e-02 -2.67413169e-01 7.81887054e-01
6.75002396e-01 1.01839840e-01 -5.05105436e-01 1.14018786e+00
-9.79201496e-01 8.74593630e-02 4.13928628e-01 1.13136351e+00
-5.25257528e-01 -1.41601503e+00 -2.82473922e-01 3.25529695e-01
-1.77462623e-01 -1.63420096e-01 -7.59821415e-01 1.17294288e+00
-1.51509985e-01 1.11862028e+00 1.26548484e-01 -5.11001587e-01
9.33231831e-01 6.65332258e-01 2.74020523e-01 -7.71780968e-01
-6.71513259e-01 -2.79216081e-01 7.34063148e-01 -8.07921410e-01
3.53625566e-01 -7.68513680e-01 -1.89948308e+00 -3.89333963e-01
1.60518050e-01 4.33779329e-01 9.59764719e-01 7.43490934e-01
7.77098656e-01 8.35110769e-02 1.45697296e-01 2.12816432e-01
-8.21254253e-01 -7.42690384e-01 -2.38275722e-01 5.36604345e-01
-4.42487627e-01 -2.42878735e-01 7.70583525e-02 2.88208753e-01] | [9.826875686645508, 7.83422327041626] |
e9718359-3be1-4db8-9de0-d3daa1c774b9 | identifying-subgroups-of-icu-patients-using | 2306.02121 | null | https://arxiv.org/abs/2306.02121v1 | https://arxiv.org/pdf/2306.02121v1.pdf | Identifying Subgroups of ICU Patients Using End-to-End Multivariate Time-Series Clustering Algorithm Based on Real-World Vital Signs Data | This study employed the MIMIC-IV database as data source to investigate the use of dynamic, high-frequency, multivariate time-series vital signs data, including temperature, heart rate, mean blood pressure, respiratory rate, and SpO2, monitored first 8 hours data in the ICU stay. Various clustering algorithms were compared, and an end-to-end multivariate time series clustering system called Time2Feat, combined with K-Means, was chosen as the most effective method to cluster patients in the ICU. In clustering analysis, data of 8,080 patients admitted between 2008 and 2016 was used for model development and 2,038 patients admitted between 2017 and 2019 for model validation. By analyzing the differences in clinical mortality prognosis among different categories, varying risks of ICU mortality and hospital mortality were found between different subgroups. Furthermore, the study visualized the trajectory of vital signs changes. The findings of this study provide valuable insights into the potential use of multivariate time-series clustering systems in patient management and monitoring in the ICU setting. | ['Guilan Kong', 'Huiying Zhao', 'Shuai Jin', 'Jianguo Hao', 'Junhua Fang', 'Wentie Liu', 'Zhilong Zhang', 'Tongyue Shi'] | 2023-06-03 | null | null | null | null | ['icu-mortality', 'time-series-clustering'] | ['medical', 'time-series'] | [-4.69796062e-01 -6.62593484e-01 1.34249389e-01 -1.28183261e-01
-2.22808525e-01 -4.55139816e-01 -1.74386874e-01 1.05865681e+00
-5.99238455e-01 3.63706738e-01 3.29119682e-01 -5.97581923e-01
-9.66399014e-01 -5.09519100e-01 2.06817523e-01 -8.31666946e-01
-6.62488341e-01 6.16392791e-01 -2.30791181e-01 3.07664812e-01
1.85698524e-01 5.75584471e-01 -1.04229093e+00 3.05834055e-01
6.37380242e-01 1.00752091e+00 -6.50648400e-02 7.08135068e-01
1.46183938e-01 7.37371325e-01 -6.82948589e-01 3.79059941e-01
2.36618429e-01 -6.94269240e-01 -5.03203645e-02 -1.84512615e-01
-6.27037942e-01 -5.07181399e-02 -2.84699470e-01 1.22082211e-01
8.20881665e-01 2.17336789e-01 6.75176382e-01 -1.26167870e+00
4.10897672e-01 6.53156519e-01 -3.61754335e-02 4.45236951e-01
-1.19691432e-01 3.68690461e-01 8.16624239e-02 -2.86381006e-01
-9.74032804e-02 6.82261646e-01 9.58276451e-01 3.36569160e-01
-9.91710007e-01 -3.29792559e-01 -5.31011105e-01 2.83457786e-01
-1.56577551e+00 6.84275031e-02 4.91491497e-01 -7.49474764e-01
5.26404500e-01 2.53311366e-01 9.12673414e-01 4.76010591e-01
5.75431406e-01 -4.48949814e-01 1.04215622e+00 -3.05371940e-01
1.77138552e-01 -1.74301378e-02 2.71245837e-01 2.54435897e-01
3.81801546e-01 1.09237939e-01 -2.35105038e-01 -8.28850210e-01
4.60733205e-01 9.28086340e-01 -4.05684620e-01 1.49609661e-02
-1.66875172e+00 5.57327807e-01 -1.06139712e-01 3.45403522e-01
-8.41690004e-01 -2.91173279e-01 7.60853887e-01 2.97619641e-01
1.15242466e-01 5.02652109e-01 -9.51925993e-01 -5.13878584e-01
-6.57276690e-01 -6.65718615e-01 6.17906451e-01 5.97626030e-01
1.49612963e-01 -3.03123504e-01 -1.45429984e-01 6.20355189e-01
2.41185501e-01 3.29646170e-01 8.30604672e-01 -1.16181934e+00
2.09807009e-01 6.74333930e-01 1.22260302e-01 -9.42769527e-01
-1.00284064e+00 3.16373259e-02 -1.10663879e+00 -3.12588274e-01
3.02341342e-01 -5.80015063e-01 -3.08947235e-01 1.13173068e+00
8.11785534e-02 6.77266270e-02 2.44279251e-01 4.87081200e-01
2.57967293e-01 6.63605556e-02 3.26367706e-01 -8.43868017e-01
1.21230710e+00 -4.80235964e-01 -5.87798834e-01 7.40845561e-01
1.02577484e+00 -3.37742388e-01 3.34689558e-01 4.09520030e-01
-9.84467864e-01 -5.44062436e-01 -2.78682828e-01 9.13637936e-01
-2.13451922e-01 -1.85527593e-01 6.48611933e-02 4.03675467e-01
-8.07850122e-01 1.08484471e+00 -1.23209965e+00 -7.15093374e-01
1.58684775e-01 2.17187479e-01 -1.87829852e-01 1.80032760e-01
-9.22268093e-01 8.48999977e-01 4.65075165e-01 -1.02510557e-01
-3.80413324e-01 -9.62685287e-01 -5.51844478e-01 1.60364993e-02
-3.61876041e-01 -1.01933146e+00 5.05893290e-01 -2.09095597e-01
-9.70172167e-01 4.01976794e-01 -2.15789720e-01 -3.23063612e-01
4.34155583e-01 4.68653031e-02 -4.68688458e-01 5.43847680e-01
-3.63339484e-01 -3.52634162e-01 2.81053782e-01 -8.50673497e-01
-5.81983387e-01 -5.65492272e-01 -8.65777493e-01 1.34321332e-01
-3.00709754e-01 1.95928529e-01 1.67838722e-01 -2.86644816e-01
1.69517815e-01 -8.94573092e-01 -5.76538622e-01 -6.05468512e-01
1.74850315e-01 5.93003854e-02 7.95949757e-01 -5.93463600e-01
1.62537372e+00 -2.50118065e+00 -6.30067587e-02 5.01860201e-01
4.98369545e-01 -1.76330194e-01 6.29577518e-01 7.27266490e-01
-3.91212881e-01 3.88711542e-01 -3.05917740e-01 -2.15980142e-01
-5.64468801e-01 2.07842067e-01 2.75525823e-02 8.42656076e-01
-2.94976950e-01 6.76457405e-01 -8.64265859e-01 -6.72192574e-01
1.00268018e+00 2.98460096e-01 -1.93080068e-01 7.66775668e-01
7.94714868e-01 1.15422130e+00 -2.10021242e-01 3.14519703e-01
1.34779125e-01 -1.96918279e-01 2.07837239e-01 1.21795073e-01
-4.02928472e-01 -2.57322878e-01 -5.15248179e-01 1.26775944e+00
-4.34692390e-02 5.09817421e-01 -1.51292920e-01 -9.22849059e-01
1.05188072e+00 8.79612803e-01 1.61462414e+00 -2.66579151e-01
5.12149930e-01 -2.98788287e-02 4.10936654e-01 -9.45574760e-01
-4.30496067e-01 -4.18583930e-01 1.18660398e-01 5.77967703e-01
-5.09324193e-01 -1.76987916e-01 -1.30595799e-04 -1.82237282e-01
1.15315187e+00 -5.38162887e-01 2.69638419e-01 -5.61064780e-01
3.24379921e-01 2.41090998e-01 5.25215268e-01 4.71839219e-01
-5.53550482e-01 8.12872410e-01 3.40295464e-01 -4.14832443e-01
-7.72496760e-01 -9.95719552e-01 -3.29342157e-01 4.10901874e-01
-2.89524376e-01 -3.81404132e-01 -1.92871630e-01 -7.40223899e-02
5.75985461e-02 6.54226542e-01 -6.54030025e-01 -4.29785162e-01
-6.88576519e-01 -1.00903618e+00 2.40058959e-01 3.92492384e-01
-1.24119729e-01 -1.04522896e+00 -1.39502835e+00 7.06185162e-01
-2.99137056e-01 -4.22646642e-01 -1.28692791e-01 4.30816650e-01
-1.51116645e+00 -1.53808844e+00 -6.12165987e-01 -4.23681170e-01
6.81342959e-01 1.93842426e-01 1.02103817e+00 3.41849208e-01
-5.95157444e-01 7.76614130e-01 -3.69451255e-01 -6.75484419e-01
-5.56842327e-01 -3.25971514e-01 4.98739779e-01 -1.34887218e-01
5.09978533e-01 -6.79657221e-01 -1.14463007e+00 4.45120484e-01
-8.30915987e-01 -4.54294294e-01 4.22955193e-02 5.45880914e-01
5.62385798e-01 1.55936465e-01 5.91720581e-01 -4.38341320e-01
6.58512592e-01 -9.94264483e-01 -2.08977088e-02 1.51371375e-01
-1.25561976e+00 -5.79135418e-01 9.36564863e-01 -3.23275208e-01
-1.99077323e-01 -2.78700352e-01 5.76350808e-01 -8.03932309e-01
-4.43683535e-01 6.58591807e-01 5.62544763e-01 7.56696165e-01
5.25437176e-01 2.68150955e-01 7.07130075e-01 -3.06709111e-01
-3.89435858e-01 7.82165289e-01 3.86643708e-01 -2.69268334e-01
3.85769010e-01 4.15572226e-01 2.92097181e-01 -5.91819346e-01
6.34237053e-03 -1.13485849e+00 -1.23906016e+00 -3.63569021e-01
9.83561099e-01 -8.93596947e-01 -1.18266714e+00 5.37431479e-01
-4.98774529e-01 -5.38841486e-01 -4.45147902e-01 9.75596428e-01
-4.68512893e-01 5.76593935e-01 -4.35261905e-01 -7.84922421e-01
-5.52866399e-01 -7.12329865e-01 2.33138636e-01 -1.52532673e-02
-6.63150668e-01 -1.39996254e+00 3.99814844e-01 -4.05141488e-02
5.90032339e-01 1.02173102e+00 1.27790689e+00 -8.18910778e-01
4.07085478e-01 -3.64533573e-01 3.19282025e-01 3.23535562e-01
9.55186844e-01 5.35466433e-01 -8.68692324e-02 -4.47680414e-01
3.76969039e-01 4.15322125e-01 6.42782524e-02 7.12267458e-01
1.41343331e+00 -4.06949297e-02 -5.17910779e-01 7.52145529e-01
1.44391453e+00 9.35400903e-01 4.05315459e-01 2.74637938e-01
7.11836278e-01 7.89878666e-01 4.45755899e-01 1.16418827e+00
4.28319514e-01 -9.41099413e-03 2.36855850e-01 -5.85368425e-02
6.75165415e-01 4.56958264e-01 1.69647008e-01 1.31197166e+00
-2.95579463e-01 1.39807919e-02 -1.31802571e+00 5.06512642e-01
-1.60615575e+00 -9.97447312e-01 -7.61412680e-01 2.52266788e+00
4.45939213e-01 -4.51889038e-01 4.95155156e-01 3.39934796e-01
7.94017434e-01 -4.65443075e-01 -4.34305102e-01 -3.51777107e-01
4.21342300e-03 4.73229922e-02 4.91466075e-01 -1.01039000e-01
-6.00893617e-01 -3.71461660e-02 7.09229040e+00 -4.18015271e-01
-1.27952480e+00 -2.00756162e-01 8.15173507e-01 -2.45012030e-01
5.49449623e-01 -1.56602204e-01 2.52386868e-01 8.53776217e-01
1.80471504e+00 -4.03207958e-01 2.80255169e-01 7.17626214e-01
1.12487364e+00 -6.59060627e-02 -9.90726650e-01 1.05143762e+00
-2.46394649e-01 -9.28577542e-01 -6.82413638e-01 -2.05745369e-01
2.55290955e-01 2.87594497e-02 -3.66644859e-01 -2.92171258e-02
-7.91811943e-02 -1.03984964e+00 1.18102334e-01 7.30373800e-01
1.06695449e+00 -5.95578969e-01 9.21217501e-01 8.23421925e-02
-1.00947309e+00 -4.66938734e-01 1.12315618e-01 3.64536233e-02
3.22676480e-01 4.31759387e-01 -8.96698356e-01 5.36894441e-01
9.86929357e-01 8.04830194e-01 -4.35059011e-01 1.11279082e+00
3.37542534e-01 9.24543142e-01 -4.89136785e-01 3.64698291e-01
-2.07836062e-01 -4.61774737e-01 2.99368888e-01 1.03058994e+00
4.79419827e-01 6.29083633e-01 -9.62110683e-02 2.14861020e-01
6.05154812e-01 1.84675589e-01 -6.47863269e-01 8.45835805e-02
4.89492297e-01 9.88040507e-01 -8.73934090e-01 -6.29041970e-01
-3.03579301e-01 4.60056484e-01 -4.21925873e-01 2.80817360e-01
-8.17487359e-01 -3.56927633e-01 5.56254983e-01 4.50917423e-01
-3.08396995e-01 -4.93268698e-01 -8.56705844e-01 -8.77167106e-01
-4.03504521e-01 -7.29831398e-01 9.34408605e-01 -4.68888283e-01
-1.13673508e+00 3.62813801e-01 2.47429371e-01 -1.52409887e+00
-4.36932951e-01 -2.19567508e-01 -9.11490679e-01 1.13598275e+00
-9.72746253e-01 1.59344703e-01 -6.81277931e-01 1.11196411e+00
1.56722769e-01 -3.53858136e-02 1.25215840e+00 1.29299879e-01
-7.47871518e-01 -6.67030141e-02 6.32692933e-01 1.53879866e-01
7.28116095e-01 -1.06759667e+00 -5.62116861e-01 1.45884827e-01
-1.09158230e+00 9.45297956e-01 4.05846298e-01 -6.89134300e-01
-1.16593707e+00 -1.09633482e+00 5.68262100e-01 -6.91232860e-01
6.57942474e-01 2.93004930e-01 -9.33522046e-01 2.16326982e-01
6.89205527e-02 -1.45836338e-01 1.43310094e+00 -3.76900941e-01
4.06068176e-01 1.30475342e-01 -1.29210329e+00 2.98372597e-01
4.08422291e-01 -1.33497447e-01 -6.70604050e-01 2.02837169e-01
3.37475896e-01 -5.91729023e-02 -2.03055477e+00 7.02710569e-01
3.74384344e-01 -9.42877054e-01 8.30791950e-01 -6.32047176e-01
1.67739596e-02 -2.16892332e-01 7.01113865e-02 -1.33378148e+00
-6.50402427e-01 -7.92558908e-01 7.59270191e-02 8.15458655e-01
6.51408583e-02 -1.11309469e+00 3.48027617e-01 1.00288808e+00
-2.33362719e-01 -7.54033983e-01 -9.23822224e-01 -5.24070203e-01
2.55980968e-01 -1.55553371e-01 6.15770042e-01 9.86773729e-01
4.10625577e-01 -1.51808038e-01 4.48955782e-02 1.25321383e-02
6.13973975e-01 -1.01513239e-02 4.48158145e-01 -1.36974657e+00
6.65762462e-03 -5.65301895e-01 -1.64524749e-01 3.08906704e-01
-3.53444576e-01 -6.87081695e-01 -1.88372836e-01 -1.74392200e+00
1.41382486e-01 -6.19464517e-01 -1.00527847e+00 2.83483267e-01
-4.43157405e-01 -4.84894902e-01 1.73366413e-01 7.45615840e-01
-1.89999584e-02 2.62302488e-01 7.34826922e-01 4.16134566e-01
-8.92992377e-01 9.64105427e-02 -2.19702244e-01 2.06077397e-01
1.13158166e+00 -7.96954930e-01 -3.16359520e-01 4.63414043e-02
-7.21147001e-01 6.75908506e-01 -1.19051687e-01 -1.10279262e+00
3.23081940e-01 -5.29354036e-01 6.00559592e-01 -7.02716947e-01
-8.52563530e-02 -1.46009445e+00 7.37269998e-01 1.03773320e+00
-9.53545719e-02 9.37327385e-01 5.12571633e-01 2.22375900e-01
-2.85999715e-01 1.65123910e-01 7.91743040e-01 -1.34821028e-01
-4.68568578e-02 5.29345572e-01 -9.04897809e-01 1.33030877e-01
1.49236798e+00 -3.11800748e-01 -8.78633633e-02 -1.33140221e-01
-7.88471162e-01 2.94841498e-01 4.28132713e-01 3.62391740e-01
5.23317277e-01 -1.10674226e+00 -8.18720877e-01 -5.02799414e-02
3.10357511e-01 1.16399810e-01 5.78033328e-01 1.65102100e+00
-7.86134720e-01 2.52833784e-01 -1.74092531e-01 -7.74068952e-01
-1.24517262e+00 9.39425170e-01 3.52566034e-01 1.39430389e-01
-7.96268463e-01 1.37395993e-01 -2.43108809e-01 1.39976844e-01
1.91633269e-01 -6.51884824e-02 -1.53044313e-01 3.54965657e-01
4.62201595e-01 7.51083672e-01 -1.78123072e-01 -4.31553870e-01
-6.26008391e-01 5.34509480e-01 6.26717746e-01 1.33766100e-01
1.32122314e+00 -4.86112416e-01 -6.36276484e-01 1.09452343e+00
1.35471368e+00 -4.48800802e-01 -8.98589194e-01 2.01037094e-01
2.68949326e-02 -7.30790198e-02 -4.65968430e-01 -4.20427650e-01
-9.86431181e-01 7.13077188e-01 8.31280410e-01 3.84368241e-01
1.38811076e+00 -2.29923561e-01 4.69916105e-01 3.67638879e-02
3.46901596e-01 -8.41816306e-01 -2.14455217e-01 3.03369910e-01
4.34418976e-01 -1.03695703e+00 -2.70676523e-01 3.45056444e-01
-7.83294261e-01 1.50631285e+00 1.44452870e-01 8.29119459e-02
1.09011471e+00 -3.09709460e-02 5.96876204e-01 -1.78019643e-01
-7.74362326e-01 5.02184272e-01 -2.95819134e-01 5.58474362e-01
4.22829509e-01 3.30120623e-01 -3.15865248e-01 3.73271853e-01
-1.46834537e-01 1.59195233e-02 4.99187887e-01 1.09514785e+00
-1.33613974e-01 -6.97779357e-01 -5.77831626e-01 8.14964652e-01
-7.07311392e-01 -1.23057827e-01 -4.05733474e-02 5.87066650e-01
-8.43697116e-02 1.52627110e+00 2.14417577e-01 -4.81200606e-01
4.48973268e-01 6.54549062e-01 -8.10009167e-02 -4.72935796e-01
-1.07750356e+00 -1.55409068e-01 -5.84667742e-01 -2.24345997e-01
-4.09013897e-01 -8.74504626e-01 -1.60784829e+00 -2.93631822e-01
1.22162163e-01 6.48019314e-01 7.96069324e-01 6.18275225e-01
7.29017913e-01 8.70542943e-01 1.20920217e+00 -3.95694762e-01
-3.06473285e-01 -1.18584538e+00 -6.68340445e-01 4.91318226e-01
5.40155768e-01 -2.46140555e-01 -1.10269845e+00 2.78638870e-01] | [7.976280212402344, 6.113954067230225] |
bd25fc08-e32e-4979-89c1-58924a691b90 | exploring-semantic-variations-in-gan-latent | 2305.14551 | null | https://arxiv.org/abs/2305.14551v1 | https://arxiv.org/pdf/2305.14551v1.pdf | Exploring Semantic Variations in GAN Latent Spaces via Matrix Factorization | Controlled data generation with GANs is desirable but challenging due to the nonlinearity and high dimensionality of their latent spaces. In this work, we explore image manipulations learned by GANSpace, a state-of-the-art method based on PCA. Through quantitative and qualitative assessments we show: (a) GANSpace produces a wide range of high-quality image manipulations, but they can be highly entangled, limiting potential use cases; (b) Replacing PCA with ICA improves the quality and disentanglement of manipulations; (c) The quality of the generated images can be sensitive to the size of GANs, but regardless of their complexity, fundamental controlling directions can be observed in their latent spaces. | ['Adil Khan', 'Rustam A. Lukmanov', 'Andrey Palaev'] | 2023-05-23 | null | null | null | null | ['disentanglement'] | ['methodology'] | [ 3.43762875e-01 -1.49778731e-03 -1.91770032e-01 2.71582484e-01
-6.66115105e-01 -1.11596572e+00 1.09482813e+00 -6.17085934e-01
9.67379510e-02 6.81763589e-01 6.45531058e-01 -1.02116756e-01
-1.80530325e-01 -5.89151025e-01 -7.53204942e-01 -1.01061153e+00
1.52290836e-01 1.88375086e-01 -3.55757892e-01 -9.60679948e-02
-2.20675897e-02 5.27492583e-01 -1.40282857e+00 7.34328702e-02
7.40039825e-01 7.88586140e-01 -2.31468081e-02 8.51622641e-01
2.02555731e-01 6.50520682e-01 -6.54903829e-01 -2.02281982e-01
6.02167964e-01 -8.46200943e-01 -3.07333291e-01 1.98189110e-01
4.32950437e-01 -3.77043933e-01 -3.10500264e-01 8.89534950e-01
2.10587859e-01 -2.29582906e-01 8.40454042e-01 -1.53668034e+00
-9.48758960e-01 1.49622008e-01 -4.82512027e-01 -2.80181885e-01
2.64631301e-01 8.28080237e-01 9.56903100e-01 -4.49012429e-01
8.41245830e-01 1.27481818e+00 2.60008782e-01 6.91103756e-01
-1.78975439e+00 -8.23170066e-01 -2.27969036e-01 -3.75846535e-01
-9.17635858e-01 -7.38100350e-01 7.29192436e-01 -6.54790699e-01
7.39058673e-01 3.67393106e-01 8.51646066e-01 1.39821267e+00
1.88216731e-01 3.86170030e-01 1.42600024e+00 -4.46947038e-01
2.46730179e-01 -7.96662197e-02 -4.01325583e-01 5.91920018e-01
2.41032436e-01 4.22376543e-01 -7.45829940e-01 -3.48967053e-02
1.08892334e+00 -1.51800275e-01 -4.07470822e-01 -6.81452990e-01
-1.52373207e+00 8.21031928e-01 4.15202856e-01 2.30876192e-01
-4.45353895e-01 5.84535778e-01 1.07383408e-01 3.11793774e-01
6.13292605e-02 1.16930747e+00 -1.61485210e-01 -5.69727361e-01
-8.68035972e-01 4.50048923e-01 6.07128382e-01 1.09125090e+00
5.41379988e-01 3.17066491e-01 -3.42795014e-01 2.96349168e-01
2.19660491e-01 6.56157255e-01 2.60674238e-01 -1.26141310e+00
5.40977955e-01 4.89530861e-01 1.25121117e-01 -6.88138247e-01
-7.12678283e-02 -2.22939759e-01 -7.67831564e-01 6.97862148e-01
5.90032578e-01 -2.20379889e-01 -1.21856320e+00 1.93609798e+00
-1.91202667e-02 -1.82248518e-01 -4.31208946e-02 4.91395712e-01
3.24598908e-01 5.98012090e-01 2.25423835e-02 -1.85508966e-01
1.12643325e+00 -1.01206326e+00 -1.07938921e+00 -4.16272044e-01
1.33893311e-01 -7.38138855e-01 1.28492987e+00 2.76417255e-01
-1.42108834e+00 -3.56273353e-01 -1.10572100e+00 -1.11627840e-01
-1.51017502e-01 -4.10614273e-04 7.84652770e-01 9.33668852e-01
-1.23188186e+00 4.63884383e-01 -1.07560718e+00 -2.21567258e-01
6.70938134e-01 3.47299427e-01 -5.76730490e-01 -1.56184003e-01
-5.60154498e-01 8.67684662e-01 6.84227124e-02 -7.69620910e-02
-1.19002557e+00 -8.18067491e-01 -7.07517922e-01 2.60176003e-01
1.67512208e-01 -9.38178539e-01 7.93170452e-01 -5.19127786e-01
-1.89938724e+00 5.37224114e-01 -8.08691010e-02 -1.06736682e-02
6.15310192e-01 -1.49388388e-01 -7.63258263e-02 6.04648851e-02
-5.17895594e-02 9.94935155e-01 1.18939853e+00 -1.39967310e+00
-7.37768412e-02 -2.40717813e-01 6.03686795e-02 1.88120067e-01
-3.84216160e-01 -1.16787285e-01 -2.30756208e-01 -7.45375931e-01
-4.16499600e-02 -1.27664816e+00 -1.51088387e-01 1.61251768e-01
-5.50005555e-01 6.42784476e-01 9.97045517e-01 -5.17219961e-01
8.91785622e-01 -2.29850340e+00 6.24195516e-01 -1.62938371e-01
3.91471952e-01 1.58071935e-01 -3.30793381e-01 5.38386881e-01
-4.47018594e-02 5.90753675e-01 -5.99123500e-02 -4.28214639e-01
2.24238485e-01 1.35468721e-01 -4.41686511e-01 3.81398797e-01
2.02024519e-01 1.25109231e+00 -5.84540844e-01 -5.56581803e-02
3.92418444e-01 5.48809230e-01 -6.16727114e-01 4.13547009e-01
-1.05883084e-01 9.06096637e-01 -1.00584775e-01 6.05149388e-01
4.11635339e-01 -1.53664604e-01 4.22706157e-02 -4.06476557e-01
1.24194041e-01 2.54115492e-01 -7.03884184e-01 1.64355588e+00
-3.43019307e-01 9.56940353e-01 1.44665390e-01 -4.16911572e-01
5.10535717e-01 3.77768993e-01 3.03249866e-01 -4.67119485e-01
-7.25960918e-03 3.89143266e-02 2.22004160e-01 -2.07691297e-01
3.93822044e-01 -2.86547482e-01 -6.50413036e-02 5.81800342e-01
2.02577770e-01 -7.91654706e-01 3.21228147e-01 1.33103818e-01
1.23386228e+00 1.60695881e-01 2.43842110e-01 -2.64217466e-01
-2.84509599e-01 -2.80448407e-01 3.51922989e-01 7.93339372e-01
-5.64762726e-02 7.22497284e-01 9.23096240e-01 -1.00928031e-01
-1.36461258e+00 -1.15818679e+00 -5.21560125e-02 5.58510482e-01
-2.53584296e-01 -5.81094861e-01 -7.22810864e-01 -5.05830824e-01
-1.14870302e-01 6.88210428e-01 -6.55935168e-01 -2.27370024e-01
-3.91017348e-01 -6.55043662e-01 4.37214345e-01 4.95836616e-01
3.81891936e-01 -8.80951345e-01 -6.07827902e-01 -2.26266608e-01
1.36180088e-01 -1.35793781e+00 -4.22860801e-01 2.12507620e-01
-7.52497911e-01 -8.34758162e-01 -5.04649401e-01 -1.87308371e-01
8.44966352e-01 1.17372148e-01 1.02670503e+00 -2.79216319e-01
-5.58554521e-03 4.42077428e-01 -1.03458643e-01 -3.58358502e-01
-7.33218253e-01 -2.05832180e-02 4.91374619e-02 -3.82299542e-01
-2.94237554e-01 -1.09096384e+00 -5.07610321e-01 3.04216832e-01
-8.76896501e-01 3.71958375e-01 6.38932347e-01 9.05838370e-01
3.31904918e-01 2.32880130e-01 2.33152620e-02 -7.74707139e-01
8.21866393e-01 -7.55654573e-02 -7.34105647e-01 8.94327536e-02
-5.92916906e-01 2.46955648e-01 4.56420511e-01 -5.17582357e-01
-8.95939946e-01 -7.93020651e-02 3.93190384e-01 -3.97086114e-01
-1.31186247e-01 1.80428803e-01 -4.25832897e-01 -1.92114830e-01
6.67356133e-01 -4.59061936e-02 1.61795989e-01 -2.19728693e-01
9.12004054e-01 1.06980927e-01 4.46903020e-01 -5.86269736e-01
1.14510655e+00 4.86362875e-01 2.82668382e-01 -6.12989724e-01
-2.90296227e-01 4.69118208e-01 -5.62270343e-01 -4.56179343e-02
1.16941249e+00 -7.97086596e-01 -4.86944199e-01 5.77177465e-01
-8.65487576e-01 -7.09599376e-01 -5.89821160e-01 3.35805804e-01
-6.55335426e-01 -1.36502564e-01 -6.87902629e-01 -5.72606862e-01
-4.08569649e-02 -1.59013915e+00 1.08184338e+00 1.13606729e-01
-6.02104664e-01 -8.57719600e-01 -1.65074855e-01 4.17652488e-01
7.30980933e-01 5.81506729e-01 1.19653094e+00 1.17647890e-02
-9.81861651e-01 -1.59103259e-01 -7.39327669e-02 2.46180862e-01
5.87668180e-01 1.64474845e-01 -9.10442710e-01 -4.66344297e-01
2.80253515e-02 -3.51708144e-01 4.93434727e-01 4.59853798e-01
8.26813161e-01 -6.24954045e-01 -1.54532954e-01 9.25617397e-01
1.15308738e+00 2.66354173e-01 8.51806521e-01 1.47391647e-01
8.85056615e-01 3.75306487e-01 7.15344995e-02 1.03575431e-01
-1.08229399e-01 7.51195550e-01 3.28160256e-01 -2.97482103e-01
-4.56179261e-01 -6.12462640e-01 4.30088371e-01 7.18522608e-01
4.05409001e-03 -3.54673564e-01 -7.01782644e-01 1.40790552e-01
-1.54920626e+00 -8.46785009e-01 2.17427686e-01 2.07003903e+00
7.64909327e-01 2.26554945e-01 3.78104597e-02 6.75684661e-02
3.18462998e-01 4.40772027e-01 -5.39399803e-01 -2.06790388e-01
-1.89089432e-01 2.61111200e-01 5.37936628e-01 2.42084309e-01
-6.79253638e-01 7.25402236e-01 8.29733753e+00 7.81013787e-01
-1.23443401e+00 1.61388308e-01 5.80688536e-01 -4.74727333e-01
-6.63684249e-01 1.49868011e-01 -4.31039214e-01 5.91392636e-01
5.24703979e-01 -1.03102483e-01 7.69683421e-01 5.61173975e-01
-1.50486529e-02 -3.44153829e-02 -1.33415687e+00 8.83689165e-01
-5.57263233e-02 -1.41640711e+00 1.34365678e-01 4.91015166e-01
9.99172211e-01 -3.54149133e-01 4.30740744e-01 9.16125700e-02
4.39393401e-01 -1.31902897e+00 7.97384322e-01 2.92717159e-01
1.15609109e+00 -4.97497618e-01 5.89921959e-02 4.44699340e-02
-6.77543998e-01 -1.40630081e-01 7.55160004e-02 -1.36694655e-01
2.89132684e-01 5.43700039e-01 -3.37032080e-01 4.80573140e-02
4.67478156e-01 4.98462111e-01 -5.81660330e-01 4.08543974e-01
-8.54342520e-01 7.66500413e-01 -2.55258918e-01 2.61538684e-01
-8.45857561e-02 -5.60662806e-01 6.66937888e-01 7.49863446e-01
4.58213061e-01 1.28858071e-02 -3.00975025e-01 1.39710534e+00
-1.45407900e-01 -6.22991979e-01 -9.07441497e-01 -6.75491750e-01
4.35836464e-01 1.15243351e+00 -7.12156653e-01 -2.39069834e-02
-1.66719332e-02 8.85753870e-01 1.62686571e-01 7.08203912e-01
-7.09282577e-01 1.01008177e-01 8.69774342e-01 -5.33942953e-02
3.37847173e-01 -6.41186953e-01 -6.78380132e-01 -1.22579336e+00
8.09313357e-02 -1.23431075e+00 -2.84758538e-01 -9.83010471e-01
-1.03639364e+00 3.31019014e-01 3.64971929e-03 -1.05061114e+00
-3.13948661e-01 -4.20701116e-01 -5.53152740e-01 7.91051865e-01
-9.27086890e-01 -1.39929569e+00 -3.56025100e-01 2.72545874e-01
2.10047424e-01 -3.92076254e-01 9.51559722e-01 -2.13771816e-02
-5.54915726e-01 5.12610555e-01 1.16102912e-01 -1.87960625e-01
4.66197610e-01 -1.15539253e+00 5.78893661e-01 1.04066765e+00
2.57595032e-01 7.72982955e-01 8.91722798e-01 -4.75616932e-01
-1.92328680e+00 -7.17670083e-01 2.07773000e-01 -7.38643229e-01
4.99812573e-01 -7.41140544e-01 -4.79531795e-01 8.67961228e-01
3.54412794e-01 6.00699848e-03 5.62373400e-01 1.56615451e-01
-4.67475086e-01 1.05281500e-03 -1.18911052e+00 8.99301291e-01
1.02927494e+00 -8.31573904e-01 -6.11286983e-02 1.86179310e-01
8.95332158e-01 -4.57972646e-01 -7.31195092e-01 2.71639496e-01
7.18986750e-01 -9.79212344e-01 8.44751060e-01 -3.67681235e-01
8.56710851e-01 -2.65291035e-01 -2.86149345e-02 -1.57031012e+00
-5.27976096e-01 -8.59731555e-01 -1.89643919e-01 1.50597143e+00
2.96604306e-01 -5.78550220e-01 6.57598019e-01 1.10099041e+00
2.52442718e-01 -4.92861599e-01 -7.88160801e-01 -8.60783875e-01
-3.61875282e-03 -1.16210610e-01 7.54374683e-01 9.27214682e-01
-2.15891376e-02 3.70858401e-01 -4.69230652e-01 -1.80882826e-01
3.64994138e-01 -9.20084044e-02 1.05194485e+00 -6.23777390e-01
-5.38389444e-01 -6.46826684e-01 -5.39166689e-01 -8.94764483e-01
-4.73189838e-02 -5.38526595e-01 -2.24330798e-01 -1.24114430e+00
9.56510678e-02 -5.01594245e-01 1.41657338e-01 5.33812702e-01
-1.63319007e-01 2.55221039e-01 6.47322655e-01 3.69611502e-01
-6.19398616e-02 7.95513093e-01 1.27470613e+00 2.63727959e-02
-2.67655164e-01 -4.57207650e-01 -8.83074343e-01 4.29768711e-01
5.11986315e-01 -3.19982111e-01 -6.22025192e-01 -5.95314622e-01
3.06996077e-01 1.19686395e-01 3.01665127e-01 -1.04762566e+00
-1.44511580e-01 -2.64990449e-01 4.08170074e-01 8.38918760e-02
5.76444149e-01 -7.62845039e-01 7.24818349e-01 3.42408359e-01
-3.54043990e-01 6.69719055e-02 3.55695337e-01 5.14136851e-01
-3.63658220e-02 1.83144748e-01 6.94718957e-01 -8.88711214e-02
9.83833522e-02 1.11466683e-01 -3.79519671e-01 -7.07703084e-02
9.12408471e-01 -7.91560933e-02 -4.83283252e-01 -7.55059242e-01
-1.29045397e-01 -3.07241261e-01 1.02276552e+00 2.90393651e-01
9.43428576e-02 -1.58723557e+00 -3.28572154e-01 2.83766627e-01
-1.33320913e-01 2.52515450e-02 -2.40412578e-02 6.13662064e-01
-3.73973817e-01 2.08679542e-01 -4.71381426e-01 -6.69329345e-01
-1.06281054e+00 6.42334163e-01 6.23510592e-02 -3.05531710e-01
-2.59874672e-01 6.48894310e-01 5.03056645e-01 -5.30959405e-02
-1.50436997e-01 -1.62508085e-01 2.59626657e-01 -1.24846786e-01
2.90433973e-01 2.57097244e-01 -2.19655082e-01 -2.96264321e-01
8.32792744e-02 5.01990914e-01 1.94228783e-01 -3.06475133e-01
1.38613832e+00 5.70057929e-02 -2.79731214e-01 2.03602955e-01
1.03434217e+00 1.23959504e-01 -1.67284107e+00 2.97268182e-01
-6.52037621e-01 -8.00619364e-01 1.03986919e-01 -7.11086512e-01
-1.21321118e+00 8.14514399e-01 5.33968568e-01 2.14083403e-01
1.23058677e+00 -1.78267926e-01 5.64478874e-01 -8.06759372e-02
2.17349097e-01 -6.26790285e-01 1.88406363e-01 1.23118842e-03
9.64177430e-01 -8.93874168e-01 2.18466848e-01 -4.79204863e-01
-5.86539984e-01 8.79923165e-01 3.40213478e-01 4.07486632e-02
3.89425188e-01 5.90286553e-01 -1.31454393e-01 -3.41471910e-01
-6.67508185e-01 1.78460807e-01 2.55006433e-01 7.72894740e-01
1.71611592e-01 8.98092613e-02 -1.18180916e-01 5.27775772e-02
-5.82776010e-01 -1.49786010e-01 6.09008670e-01 8.70650351e-01
2.40284666e-01 -1.32371271e+00 -3.27709228e-01 3.65654141e-01
-1.60130143e-01 -1.52301313e-02 -2.47769102e-01 7.80912757e-01
7.65832886e-02 8.71234238e-01 -1.78296164e-01 -2.37863034e-01
2.87245326e-02 -3.32620442e-02 7.79707253e-01 -3.98645937e-01
-3.49354118e-01 7.00175762e-02 -1.62048843e-02 -6.80764377e-01
-2.66784549e-01 -6.96454287e-01 -5.99739015e-01 -2.73371011e-01
-3.87639135e-01 -2.46491611e-01 8.53883684e-01 9.46209848e-01
6.84434652e-01 5.00575066e-01 4.20823991e-01 -1.09428942e+00
-3.93419474e-01 -7.74248898e-01 -5.32470167e-01 7.14170039e-01
4.60186630e-01 -6.48478270e-01 -6.04668260e-01 3.29738528e-01] | [11.650444984436035, -0.3575461506843567] |
a0eb48ad-af43-4e70-a337-40f92b2a860f | multi-task-deep-morphological-analyzer | 1811.08619 | null | https://arxiv.org/abs/1811.08619v2 | https://arxiv.org/pdf/1811.08619v2.pdf | Multi Task Deep Morphological Analyzer: Context Aware Joint Morphological Tagging and Lemma Prediction | The ambiguities introduced by the recombination of morphemes constructing several possible inflections for a word makes the prediction of syntactic traits in Morphologically Rich Languages (MRLs) a notoriously complicated task. We propose the Multi Task Deep Morphological analyzer (MT-DMA), a character-level neural morphological analyzer based on multitask learning of word-level tag markers for Hindi and Urdu. MT-DMA predicts a set of six morphological tags for words of Indo-Aryan languages: Parts-of-speech (POS), Gender (G), Number (N), Person (P), Case (C), Tense-Aspect-Modality (TAM) marker as well as the Lemma (L) by jointly learning all these in one trainable framework. We show the effectiveness of training of such deep neural networks by the simultaneous optimization of multiple loss functions and sharing of initial parameters for context-aware morphological analysis. Exploiting character-level features in phonological space optimized for each tag using multi-objective genetic algorithm, our model establishes a new state-of-the-art accuracy score upon all seven of the tasks for both the languages. MT-DMA is publicly accessible: code, models and data are available at https://github.com/Saurav0074/morph_analyzer. | ['Anil Kumar Singh', 'Akhilesh Sudhakar', 'Saurav Jha'] | 2018-11-21 | null | null | null | null | ['morphological-tagging'] | ['natural-language-processing'] | [-4.11246456e-02 -2.31957972e-01 1.80177748e-01 -5.41902065e-01
-1.22962844e+00 -8.56851459e-01 2.18476653e-01 5.42686343e-01
-7.29549706e-01 6.72781646e-01 1.64696455e-01 -7.09674001e-01
-6.17353246e-02 -6.46202087e-01 -5.83734035e-01 -7.28366673e-01
6.77613192e-04 7.98930943e-01 -8.88545364e-02 -1.58331737e-01
1.23364292e-01 4.20455247e-01 -1.15775752e+00 3.84443671e-01
1.02416658e+00 6.26139879e-01 3.90493482e-01 5.21030366e-01
-2.76658237e-01 1.10224552e-01 -4.69601244e-01 -1.05834532e+00
2.81776488e-01 -1.27629265e-01 -7.56456375e-01 -4.09631193e-01
4.90120232e-01 2.16449931e-01 3.17393154e-01 1.37718201e+00
6.54615045e-01 8.84756148e-02 6.28790021e-01 -6.66114330e-01
-9.44092453e-01 1.34089577e+00 -6.80535376e-01 5.22192121e-01
-6.67852536e-02 5.02810776e-02 1.67597222e+00 -9.45126772e-01
4.88506645e-01 1.53217208e+00 7.05342293e-01 3.29218864e-01
-1.21235573e+00 -4.70200211e-01 -8.91670659e-02 2.98328578e-01
-1.28705037e+00 -4.33190346e-01 7.31739998e-01 -3.97965968e-01
1.32529604e+00 1.07401721e-01 2.28459418e-01 7.35130906e-01
3.95355046e-01 7.41957426e-01 1.26949167e+00 -7.34975576e-01
-2.94493213e-02 -2.96092659e-01 2.46315598e-01 9.51172411e-01
2.32817382e-01 -1.81755066e-01 -5.26236236e-01 -3.75561975e-03
4.67319787e-01 -7.70403028e-01 4.00909454e-01 2.77266681e-01
-1.01823270e+00 9.89334106e-01 -2.26034090e-01 4.80245918e-01
-3.17244053e-01 -1.93719044e-01 5.82000077e-01 2.86113918e-01
3.88873637e-01 3.12357426e-01 -9.36326981e-01 -6.14320487e-02
-6.16347790e-01 1.18179947e-01 4.83687550e-01 5.03792942e-01
7.57986248e-01 3.88084769e-01 -4.37997319e-02 1.34340775e+00
3.31835210e-01 4.95423317e-01 8.71129632e-01 -5.68516433e-01
6.94588721e-01 4.53085542e-01 -5.95217407e-01 -4.22986776e-01
-6.56386971e-01 -2.83249646e-01 -6.05910301e-01 8.26174319e-02
6.12176001e-01 -5.31327426e-01 -8.72685730e-01 2.17256641e+00
4.79068071e-01 -4.30727273e-01 6.63561150e-02 3.90369117e-01
7.73468375e-01 8.28824639e-01 5.67431331e-01 -1.69129163e-01
1.73335314e+00 -7.42408395e-01 -6.25008523e-01 -4.20742542e-01
7.30555058e-01 -8.26951444e-01 1.14740896e+00 3.56914103e-01
-1.59305859e+00 -5.86038411e-01 -8.79983127e-01 -3.55958611e-01
-6.95175529e-01 4.63724464e-01 7.21260905e-01 9.14326251e-01
-9.54868615e-01 3.81270081e-01 -8.23810279e-01 -7.37867802e-02
5.74564077e-02 6.43086612e-01 -5.15428901e-01 4.72375005e-01
-1.03544259e+00 8.31695318e-01 7.85601020e-01 2.58062035e-01
-4.22247946e-01 -5.76677203e-01 -1.07850611e+00 -1.91914931e-01
-9.04538333e-02 -2.40530863e-01 9.31536317e-01 -8.91279101e-01
-1.49200618e+00 1.52474880e+00 5.28627150e-02 -1.87822402e-01
-5.04499339e-02 -1.84594557e-01 -6.07416928e-01 -1.49871334e-01
1.50310144e-01 6.21027708e-01 4.63659644e-01 -8.48634660e-01
-8.74573112e-01 -6.11394644e-01 -3.99523139e-01 1.17619894e-01
-8.76840726e-02 1.02538264e+00 1.60422578e-01 -7.98866928e-01
1.19908340e-02 -7.00445473e-01 -7.79637620e-02 -7.28130758e-01
-5.30877650e-01 -5.99358857e-01 2.65722811e-01 -1.49602473e+00
1.17010868e+00 -1.94378912e+00 2.41450518e-01 -3.49487402e-02
-2.79020280e-01 5.92524171e-01 -3.12886178e-01 1.03252135e-01
-1.38520062e-01 1.65810779e-01 -4.25838470e-01 -4.60344285e-01
1.35607332e-01 1.26526996e-01 3.64996046e-01 4.85720515e-01
5.93236089e-01 1.15684378e+00 -6.71900153e-01 -4.33082491e-01
5.39113283e-02 2.99945772e-01 -4.41319436e-01 -1.06873205e-02
-1.89429268e-01 3.70853245e-01 -9.40360874e-02 9.02113557e-01
6.89290285e-01 4.52261686e-01 3.78203958e-01 2.55960792e-01
-3.15667868e-01 8.07606101e-01 -1.23000550e+00 1.41247690e+00
-5.97575188e-01 4.96748136e-03 2.03492641e-01 -8.85331452e-01
9.80609715e-01 1.28485203e-01 1.36661336e-01 -7.03020036e-01
3.12339842e-01 6.40362203e-01 6.71780288e-01 -1.52837858e-01
4.17745888e-01 -3.41421813e-01 -5.21439254e-01 2.84864753e-01
5.55423975e-01 1.43854201e-01 2.92644352e-01 -5.61204374e-01
9.26681280e-01 1.82518169e-01 7.50763714e-01 -6.52649045e-01
7.10875392e-01 -4.41045374e-01 1.06596708e+00 3.42274219e-01
-1.16867140e-01 3.10889572e-01 3.50697428e-01 -3.49058390e-01
-1.05118978e+00 -1.45306528e+00 -3.37798685e-01 1.64091790e+00
-7.63658345e-01 -5.38327843e-02 -7.53559291e-01 -5.08874953e-01
1.38152940e-02 1.14095330e+00 -3.79600018e-01 1.13920443e-01
-1.30975461e+00 -1.19500339e+00 7.80338943e-01 4.17638987e-01
2.12095841e-03 -1.75081766e+00 -2.87559241e-01 4.38735843e-01
-1.15615530e-02 -1.11702669e+00 -3.28609645e-01 7.69425929e-01
-5.82734883e-01 -5.54124057e-01 -2.23395243e-01 -1.30306911e+00
2.52290070e-01 -5.91525495e-01 1.06205392e+00 -1.02609381e-01
-2.05370337e-01 -1.85267165e-01 -3.00364345e-01 -5.93743086e-01
-6.88602388e-01 1.83865443e-01 9.75279585e-02 8.16610530e-02
3.80560517e-01 -6.21813059e-01 1.02823265e-01 -4.05002654e-01
-6.40673637e-01 -3.53191555e-01 4.71288532e-01 5.53663671e-01
8.00027430e-01 -2.36335665e-01 5.58345258e-01 -9.08844590e-01
2.83463836e-01 -5.67366779e-01 -7.84772635e-01 2.86075234e-01
-9.61239710e-02 1.20623901e-01 6.94029272e-01 -4.19807881e-01
-1.09376657e+00 1.33547395e-01 -6.16253614e-01 3.58436704e-01
-3.11665535e-01 6.51812315e-01 -8.41905177e-01 3.06047171e-01
2.26849288e-01 1.78971216e-02 -3.64695609e-01 -8.83184075e-01
4.71501172e-01 5.51600873e-01 8.32124174e-01 -9.04421747e-01
6.13600433e-01 -8.38621482e-02 9.46264435e-03 -9.08898652e-01
-8.30786943e-01 -3.11711818e-01 -1.27871323e+00 3.13454300e-01
1.11072147e+00 -8.14775050e-01 -4.20224011e-01 8.37767422e-01
-1.15482271e+00 -4.99479204e-01 -4.89728376e-02 2.74708986e-01
-4.03538823e-01 3.42617720e-01 -1.02336073e+00 -5.51331997e-01
-7.62683451e-01 -9.79223430e-01 9.43172991e-01 -1.21311210e-02
-1.85563579e-01 -1.25725067e+00 1.70417562e-01 3.23325694e-01
-9.73264128e-02 1.33502945e-01 1.90513086e+00 -1.24245512e+00
-6.07706457e-02 2.72479057e-01 -3.51237543e-02 4.79852140e-01
1.18562728e-01 1.92493722e-02 -6.02855444e-01 -9.88584682e-02
-7.06438497e-02 -4.38634932e-01 8.39242518e-01 5.80134451e-01
7.92227685e-01 -4.02476400e-01 3.76992524e-01 9.36787963e-01
1.57251358e+00 2.21117601e-01 4.06181216e-01 4.89256561e-01
8.30490053e-01 6.49613082e-01 5.19199073e-01 3.22939336e-01
7.13433802e-01 5.74441135e-01 4.50128354e-02 2.71580458e-01
-2.49999687e-01 5.34823388e-02 8.08663428e-01 1.28013933e+00
1.77822076e-02 -2.54240036e-01 -1.14993918e+00 9.69916523e-01
-1.29346001e+00 -5.86878061e-01 -3.54298025e-01 2.16340470e+00
1.27015758e+00 3.70334908e-02 3.21247637e-01 1.54417351e-01
9.36680853e-01 8.25939775e-02 -5.06809540e-02 -1.15569258e+00
-6.43962443e-01 9.25158143e-01 4.14157838e-01 7.29887426e-01
-1.24616110e+00 1.67792070e+00 5.01574945e+00 1.09670794e+00
-9.18409169e-01 3.52554560e-01 5.94202161e-01 2.10789874e-01
-3.17365855e-01 -8.93192887e-02 -1.41297078e+00 4.01867121e-01
1.32154191e+00 6.96997717e-02 4.41110045e-01 4.57423657e-01
6.72937483e-02 9.11300536e-03 -8.28765631e-01 6.53339922e-01
9.00351629e-03 -8.82529557e-01 2.72893310e-01 3.76626551e-02
4.78373528e-01 2.62887746e-01 1.16107523e-01 2.04583347e-01
5.68540096e-01 -8.25741112e-01 1.24209726e+00 -2.58793645e-02
8.51500928e-01 -1.00809145e+00 5.77710450e-01 1.24402940e-01
-1.13200748e+00 -1.66144863e-01 -5.50174236e-01 6.20606588e-03
4.01009619e-01 4.22048539e-01 -5.62347293e-01 3.46903265e-01
3.18612337e-01 3.60410325e-02 -8.76428425e-01 8.02979529e-01
-5.70783377e-01 1.09311068e+00 -3.75955045e-01 4.60263416e-02
3.04719388e-01 -3.44004691e-01 7.65184462e-01 1.65726638e+00
4.46337223e-01 3.01844254e-02 1.38217553e-01 6.24022663e-01
5.74501157e-02 8.11438322e-01 -9.17791128e-02 -5.40171973e-02
4.32330698e-01 1.51212370e+00 -8.85188520e-01 1.41613483e-01
-2.19828963e-01 7.42301464e-01 1.05227840e+00 -2.63314217e-01
-4.83249277e-01 -4.07262176e-01 7.57613778e-01 -2.50188917e-01
5.77651024e-01 -4.97793972e-01 -5.91805518e-01 -6.02477014e-01
-4.80361329e-03 -8.36950779e-01 8.09311867e-01 -1.93814471e-01
-1.42146027e+00 6.23266399e-01 -2.21746892e-01 -3.09783340e-01
-2.44202539e-01 -1.11457276e+00 -6.86476469e-01 1.02405083e+00
-1.53835452e+00 -1.63603878e+00 7.21450031e-01 4.30160522e-01
5.81302822e-01 -6.32357061e-01 9.90113199e-01 3.34001064e-01
-8.04747760e-01 9.46516633e-01 -3.59000526e-02 5.66096187e-01
5.49454749e-01 -1.73820269e+00 7.67550111e-01 1.03933144e+00
4.96793658e-01 3.27782243e-01 3.78643781e-01 -8.03927064e-01
-1.08028281e+00 -1.14220798e+00 1.57011390e+00 -3.10398430e-01
1.08947432e+00 -6.51829720e-01 -9.20906067e-01 8.77280593e-01
2.04234749e-01 -2.85039216e-01 9.22269940e-01 2.36676171e-01
-2.77802110e-01 1.06868088e-01 -9.66825366e-01 5.88157594e-01
9.08812642e-01 -2.89342940e-01 -7.38125384e-01 3.24891359e-01
7.37706125e-01 -1.30276263e-01 -9.00495708e-01 6.54962286e-02
2.44238839e-01 -7.91898727e-01 7.54546225e-01 -6.35727525e-01
2.48199522e-01 7.00895339e-02 -3.85807842e-01 -1.43520319e+00
-6.62014902e-01 -7.30991662e-01 2.01707795e-01 1.71261716e+00
8.68360043e-01 -4.97868538e-01 3.34090173e-01 -1.38095066e-01
-6.75152004e-01 -4.15209681e-01 -1.33782482e+00 -7.37596750e-01
6.71458304e-01 -5.38150132e-01 5.88110268e-01 8.62659693e-01
-1.75506771e-01 5.80524504e-01 -3.05816159e-02 2.68560410e-01
5.71307421e-01 -1.80113003e-01 -1.25760242e-01 -1.35054612e+00
-4.19802666e-01 -6.80503190e-01 -3.86294037e-01 -6.90097958e-02
8.34815145e-01 -1.50876260e+00 -5.96096441e-02 -1.18425989e+00
-1.11796567e-02 -5.29993296e-01 -4.99931335e-01 9.99662817e-01
-4.36243594e-01 1.54142946e-01 2.55754858e-01 -4.35942918e-01
-3.46397400e-01 1.72986522e-01 6.87763631e-01 2.31237382e-01
-2.49114841e-01 -4.77505066e-02 -6.19671404e-01 1.02034199e+00
1.07469606e+00 -6.18117213e-01 3.60644311e-01 -7.43042350e-01
5.27382910e-01 -6.75067231e-02 -1.79443657e-01 -7.91542411e-01
-1.11693658e-01 -1.70865387e-01 3.44030023e-01 -5.86175501e-01
3.73327643e-01 -2.18033597e-01 -2.66290069e-01 3.62250924e-01
-4.33654822e-02 7.88062155e-01 3.26671094e-01 -2.86097407e-01
2.39440724e-02 -6.09443069e-01 9.50116515e-01 -2.56832421e-01
-7.08249211e-01 1.65996790e-01 -4.78954643e-01 3.88605773e-01
6.79712117e-01 -1.25441164e-01 -7.80338868e-02 4.27336216e-01
-5.30559242e-01 -9.64958817e-02 -6.18427992e-02 4.23760742e-01
3.33378792e-01 -1.12249625e+00 -1.32007027e+00 3.23689759e-01
-1.74921989e-01 -4.30537790e-01 1.23647116e-02 5.06146312e-01
-5.62261105e-01 1.23646453e-01 -5.53391576e-01 1.02874808e-01
-1.33376348e+00 4.25723732e-01 5.08862287e-02 -6.21653855e-01
-1.71758145e-01 1.16852808e+00 -1.73046160e-03 -1.05171680e+00
-1.07621476e-01 -9.84867886e-02 -4.91978198e-01 3.30547869e-01
2.74644256e-01 4.91317987e-01 2.97662437e-01 -1.29901886e+00
-4.03148383e-01 4.24582183e-01 -1.14439197e-01 -2.56692022e-01
1.65383613e+00 -8.89117718e-02 -6.78108215e-01 6.11938715e-01
1.02073610e+00 6.11735880e-01 -5.99454820e-01 -8.85934681e-02
4.80949402e-01 1.78126812e-01 -1.34677410e-01 -1.08740067e+00
-7.91844368e-01 9.37078655e-01 7.99841583e-02 -3.20106119e-01
8.74538958e-01 3.91281769e-02 1.10586953e+00 1.79135188e-01
1.26874685e-01 -1.46988392e+00 -4.25716728e-01 1.16244936e+00
6.58765137e-01 -9.30907190e-01 -4.44456011e-01 -4.62290525e-01
-7.51454473e-01 9.43854690e-01 3.94544750e-01 -1.04315512e-01
6.71086729e-01 4.36708093e-01 3.49013627e-01 -3.34178470e-02
-5.65809369e-01 -4.53234315e-01 4.73315984e-01 6.05888188e-01
7.33143687e-01 5.42868793e-01 -7.22169340e-01 9.16225493e-01
-8.75150561e-01 -1.16341722e+00 2.97886193e-01 5.93618274e-01
-4.95837420e-01 -1.48996007e+00 -2.18462572e-01 4.41382349e-01
-1.25688612e+00 -5.78462958e-01 -1.90105155e-01 5.76857030e-01
4.88600552e-01 8.27922881e-01 1.86051384e-01 2.60091834e-02
-2.19218135e-02 4.61022288e-01 7.09001005e-01 -9.43707824e-01
-1.08215511e+00 1.43297791e-01 4.42660302e-01 4.83979322e-02
-1.23176575e-02 -1.30356038e+00 -1.44089818e+00 4.71180826e-02
1.25445826e-02 2.06096321e-02 6.92658901e-01 1.04777002e+00
-2.04198927e-01 1.10677600e-01 3.26991409e-01 -7.23385513e-01
-5.26492655e-01 -1.01442897e+00 -9.22361910e-01 2.86991864e-01
-8.00281912e-02 -4.35380250e-01 -1.79819018e-01 -1.22105382e-01] | [10.445755958557129, 10.080602645874023] |
3ee0dbca-a6fc-482c-aa9b-8006a4bb10d2 | predicting-protein-secondary-structure-with | 1809.09210 | null | https://arxiv.org/abs/1809.09210v2 | https://arxiv.org/pdf/1809.09210v2.pdf | Predicting protein secondary structure with Neural Machine Translation | We present analysis of a novel tool for protein secondary structure prediction using the recently-investigated Neural Machine Translation framework. The tool provides a fast and accurate folding prediction based on primary structure with subsecond prediction time even for batched inputs. We hypothesize that Neural Machine Translation can improve upon current predictive accuracy by better encoding complex relationships between nearby but non-adjacent amino acids. We overview our modifications to the framework in order to improve accuracy on protein sequences. We report 65.9% Q3 accuracy and analyze the strengths and weaknesses of our predictive model. | ['Ian Bulovic', 'Evan Weissburg'] | 2018-09-24 | null | null | null | null | ['protein-secondary-structure-prediction'] | ['medical'] | [ 6.59923851e-01 1.00917175e-01 -3.27343851e-01 -7.53525078e-01
-6.65713787e-01 -5.76628447e-01 1.05440594e-01 2.12371439e-01
-4.02321517e-01 1.43344259e+00 1.32471606e-01 -1.13336146e+00
2.39850059e-01 -4.96917725e-01 -1.11061037e+00 -7.02969313e-01
-1.13121152e-01 6.02525055e-01 2.95248311e-02 -4.81421858e-01
3.02213162e-01 7.14502871e-01 -1.26572013e+00 9.58870471e-01
8.99213433e-01 5.95607698e-01 2.75373429e-01 8.07805717e-01
-1.08056866e-01 5.40272593e-01 -4.05465990e-01 -2.79121250e-01
6.68337271e-02 -6.99076712e-01 -1.05101788e+00 -6.25277102e-01
2.84884363e-01 -8.39844346e-02 5.23370467e-02 4.66997981e-01
4.88160998e-01 -1.44243121e-01 5.42095482e-01 -5.16985595e-01
-9.87772465e-01 3.46341610e-01 1.34879127e-01 4.35080379e-01
4.38978136e-01 3.80322695e-01 8.84598494e-01 -9.14627969e-01
1.00496614e+00 8.08826387e-01 1.06248891e+00 5.90328157e-01
-1.59685373e+00 -3.46987218e-01 -3.19249928e-01 5.82821846e-01
-9.36222851e-01 -3.95033032e-01 1.37368590e-01 -1.59346506e-01
2.37058854e+00 1.47864267e-01 6.97821736e-01 1.02059031e+00
1.17746472e+00 2.72134423e-01 1.15908337e+00 -4.60349888e-01
1.43711835e-01 -5.07549822e-01 2.85319418e-01 8.84334505e-01
-1.14981502e-01 3.47223431e-01 -7.32549489e-01 -6.42698526e-01
4.33924854e-01 -3.35596204e-02 8.82858504e-03 -5.43416291e-02
-1.36440516e+00 7.58706093e-01 4.47863638e-01 1.62264213e-01
-6.66561186e-01 2.00596498e-03 2.28808835e-01 5.32110989e-01
5.59756517e-01 6.84595168e-01 -1.31675828e+00 -4.72027242e-01
-9.88982320e-01 1.38334408e-01 9.86584306e-01 6.49256945e-01
7.26400614e-01 -1.37594953e-01 2.13138014e-01 4.89978939e-01
-1.89711489e-02 2.19443172e-01 5.80406129e-01 -7.78953671e-01
6.23753294e-02 3.69004577e-01 5.73329404e-02 -3.70945543e-01
-5.83882391e-01 -1.75113991e-01 -5.79466999e-01 1.21004589e-01
2.64574498e-01 -7.64726475e-02 -9.76288199e-01 1.40143466e+00
-9.20594186e-02 -2.28045553e-01 2.57625580e-01 7.12559223e-01
5.12770593e-01 7.86488950e-01 3.04357827e-01 -7.51153350e-01
9.02898192e-01 -1.01367879e+00 -6.53028250e-01 2.87776917e-01
8.03873956e-01 -8.44076693e-01 7.28617609e-01 4.60234523e-01
-1.02159870e+00 -3.15962911e-01 -1.20185268e+00 -4.13755238e-01
-4.36623663e-01 1.17615864e-01 7.83374846e-01 3.86947900e-01
-1.09602857e+00 1.36630940e+00 -1.02750552e+00 -4.43138599e-01
3.22491489e-02 9.38265860e-01 -4.52219725e-01 4.48729873e-01
-1.22519934e+00 1.35033381e+00 6.27989948e-01 -2.14326471e-01
-3.47980231e-01 -5.09610116e-01 -2.64905632e-01 -8.38817284e-02
-2.31740326e-01 -7.85881460e-01 1.46563303e+00 -1.33547008e+00
-1.74796200e+00 7.61048913e-01 -7.66330004e-01 -8.74961853e-01
3.79046611e-02 -2.22945645e-01 -3.13090384e-01 2.00434327e-01
-3.35278332e-01 7.42130458e-01 1.29617378e-01 -3.84243250e-01
-3.25345069e-01 -3.87388259e-01 -4.84801441e-01 -2.30781604e-02
3.51138383e-01 2.45349988e-01 4.82576400e-01 -5.46190321e-01
2.20673159e-01 -1.03958511e+00 -4.54155326e-01 -3.51171717e-02
3.46169531e-01 -1.58902019e-01 5.75412750e-01 -1.03119135e+00
9.33249533e-01 -1.31204236e+00 2.59346992e-01 -4.90554683e-02
1.03459477e-01 4.45564091e-01 -1.76714599e-01 8.64601552e-01
-5.57485640e-01 -1.73851788e-01 -2.40851298e-01 3.60902339e-01
-4.49712038e-01 4.40092057e-01 -4.16714907e-01 3.69006336e-01
5.71144342e-01 1.42143083e+00 -5.93298733e-01 -8.61767158e-02
-1.62489433e-02 5.10983944e-01 -5.11664033e-01 7.15252459e-02
-5.65425277e-01 2.41011485e-01 -1.30128369e-01 8.71966302e-01
3.92765731e-01 -4.76136893e-01 6.40364766e-01 -9.96123925e-02
-1.79524884e-01 6.96691811e-01 2.67446321e-02 1.65140295e+00
8.84948000e-02 5.40831506e-01 -2.90575206e-01 -8.89908791e-01
1.20302522e+00 2.31875002e-01 4.22918290e-01 -5.24951220e-01
-2.11178243e-01 3.45174104e-01 4.25042242e-01 -3.97931457e-01
3.25558841e-01 -5.79935133e-01 6.26753390e-01 4.20217812e-01
3.53325039e-01 4.51706916e-01 -8.56244862e-02 -7.78188631e-02
1.22232592e+00 1.09027445e+00 8.20752203e-01 -4.04299080e-01
2.36155704e-01 4.78260398e-01 6.55352592e-01 2.49727428e-01
-4.06069160e-01 3.25452000e-01 3.40288639e-01 -1.19771612e+00
-1.68101978e+00 -6.53067172e-01 -4.19142544e-02 1.55413175e+00
-3.95753354e-01 -6.54513419e-01 -7.14491010e-01 -6.96870565e-01
-2.10244328e-01 5.39504409e-01 -4.63553280e-01 -1.66168302e-01
-9.00817752e-01 -1.11736786e+00 6.55498683e-01 5.88935494e-01
-3.42498511e-01 -1.35204291e+00 -5.48872828e-01 7.88322568e-01
-2.37882420e-01 -6.12370729e-01 -5.42542040e-02 1.19331169e+00
-1.39995480e+00 -8.71269763e-01 -5.06152272e-01 -8.98852885e-01
2.65791386e-01 -3.53662781e-02 1.35068405e+00 2.32470766e-01
-1.36596158e-01 -5.45892596e-01 -3.41891170e-01 -3.21085691e-01
-8.32433522e-01 2.24441394e-01 4.65315938e-01 -7.84288228e-01
1.05453563e+00 -7.64351487e-01 -3.94915909e-01 2.58755416e-01
-5.52325070e-01 4.27083313e-01 8.51162851e-01 1.00971913e+00
7.07472980e-01 -8.99622738e-01 6.83817625e-01 -7.21871674e-01
6.39803231e-01 -3.68488729e-01 -3.14649761e-01 4.39623564e-01
-1.12028539e+00 6.35569930e-01 8.70499194e-01 -3.16959888e-01
-7.22918987e-01 6.87945366e-01 -6.07841969e-01 1.17257118e-01
-1.99773863e-01 3.67779613e-01 2.12223291e-01 -2.70361394e-01
9.26138461e-01 6.48010552e-01 3.61262172e-01 -4.88926500e-01
2.66588211e-01 5.27257204e-01 2.29583353e-01 -6.32502198e-01
2.03097388e-02 -3.36678997e-02 2.42342129e-01 -5.46746373e-01
-2.47965857e-01 -1.61670089e-01 -1.07167995e+00 2.98930436e-01
5.48396051e-01 -5.58217525e-01 -8.62537384e-01 8.53104666e-02
-1.33099711e+00 -1.96477398e-01 2.17954755e-01 4.48195934e-01
-1.18831801e+00 7.48910785e-01 -1.02407730e+00 -2.01707602e-01
-8.32538545e-01 -1.20540679e+00 8.13448846e-01 -1.64030492e-01
-5.82556844e-01 -6.53092265e-01 3.53930295e-01 2.41491765e-01
3.47391367e-01 5.98898456e-02 1.18759596e+00 -8.91531765e-01
-3.31135422e-01 4.90851790e-01 -1.09339662e-01 1.40420198e-01
1.89434998e-02 2.03828663e-01 -4.01075512e-01 -1.45472437e-01
-2.98944235e-01 -4.24360484e-01 9.37846243e-01 2.49199331e-01
6.92227185e-01 -3.78852665e-01 -3.61593008e-01 6.11064851e-01
1.26757824e+00 4.48576361e-01 5.18319607e-01 6.21962130e-01
1.06045119e-01 4.93924081e-01 6.15490675e-01 1.75436363e-02
-2.42867038e-01 5.40132999e-01 3.70918065e-01 2.49026995e-02
3.14012676e-01 -1.56950951e-01 5.96890628e-01 8.76139700e-01
-5.18474698e-01 1.86059605e-02 -1.08888948e+00 1.63349718e-01
-1.93994319e+00 -8.98800492e-01 -1.81071922e-01 1.58622074e+00
1.42845321e+00 2.04795226e-02 1.78664014e-01 -3.31621021e-01
4.10332084e-01 -4.08810765e-01 -8.55921268e-01 -1.16342628e+00
-3.68646711e-01 6.61283851e-01 7.54612446e-01 5.64706087e-01
-6.73010886e-01 1.22177374e+00 8.72758389e+00 4.20443237e-01
-1.17012131e+00 -1.58342570e-01 4.88535225e-01 -2.15648845e-01
-1.37881180e-02 1.39578715e-01 -7.78209865e-01 3.16459924e-01
1.84155166e+00 -9.76834521e-02 5.55967331e-01 6.68701053e-01
4.45468098e-01 4.45765406e-01 -1.05046988e+00 3.68585497e-01
-1.05411962e-01 -1.94245768e+00 1.69598222e-01 1.68690458e-01
4.09414202e-01 4.61243987e-01 -2.41306230e-01 1.46576926e-01
3.13042343e-01 -1.24766970e+00 1.63048461e-01 6.15452707e-01
7.13091791e-01 -9.63039219e-01 6.77907228e-01 4.84906554e-01
-5.37778914e-01 1.54049620e-01 -6.91393495e-01 -4.12701011e-01
4.97378185e-02 3.73022527e-01 -1.16034496e+00 3.06346208e-01
4.45431590e-01 6.67953789e-01 -4.27909344e-01 4.09872293e-01
-3.23030986e-02 8.57185781e-01 -1.93241328e-01 -3.35806459e-01
3.30815792e-01 -4.62700725e-01 1.28173023e-01 1.36623180e+00
1.23141341e-01 5.69660664e-01 -3.29848453e-02 5.47954857e-01
9.73572806e-02 3.55366796e-01 -3.46513450e-01 -2.75703847e-01
-9.72291231e-02 7.21974850e-01 -3.92867565e-01 -5.59156716e-01
-1.54336512e-01 1.15790093e+00 7.28073001e-01 2.66052306e-01
-8.71179521e-01 -2.88259804e-01 9.35001731e-01 -4.79787104e-02
3.30830246e-01 -5.11325538e-01 -1.62910685e-01 -1.00650704e+00
-8.08335394e-02 -1.23836088e+00 1.11692597e-03 -9.25557196e-01
-1.03521991e+00 7.11669147e-01 -6.03497028e-01 -7.30938733e-01
-5.89738369e-01 -1.14007580e+00 -6.84541166e-02 1.06535327e+00
-1.11805832e+00 -1.25216341e+00 7.81178415e-01 -6.45903349e-02
4.79087323e-01 -4.27459240e-01 1.42339087e+00 -2.40467802e-01
-1.33513749e-01 3.45257521e-01 7.93443441e-01 -5.23342013e-01
8.79643798e-01 -1.24734664e+00 9.74104583e-01 5.14786065e-01
-3.74023989e-02 1.00920248e+00 1.04498208e+00 -8.34322989e-01
-1.33070600e+00 -1.01759851e+00 1.55189419e+00 -6.05227172e-01
4.35140878e-01 -5.79284132e-02 -1.18024623e+00 7.22091258e-01
2.49274537e-01 -5.54535389e-01 1.22003055e+00 1.55857652e-01
-3.98527384e-01 4.90680277e-01 -1.10383618e+00 3.29095751e-01
1.01032674e+00 -6.31107926e-01 -1.09388065e+00 5.54346204e-01
1.01905370e+00 -1.46076918e-01 -9.97331381e-01 5.75317502e-01
9.16481853e-01 -9.84989166e-01 9.50735688e-01 -1.50694585e+00
7.20244527e-01 -1.76891536e-01 -1.46680206e-01 -1.01665437e+00
-7.56176770e-01 -7.84782350e-01 -5.32930553e-01 7.82330409e-02
8.40900481e-01 -5.81745327e-01 1.14014542e+00 2.89361984e-01
-1.92307830e-01 -1.20013869e+00 -1.02887046e+00 -5.85461080e-01
5.40297151e-01 -1.14859398e-02 4.45510238e-01 7.74035811e-01
4.97055948e-01 6.48802161e-01 -5.23564756e-01 -2.41007701e-01
-3.25414538e-02 3.44521195e-01 1.67255297e-01 -9.30654228e-01
-5.79346657e-01 -6.68534487e-02 -3.33836794e-01 -1.09932363e+00
2.70638734e-01 -1.01347530e+00 -6.68929741e-02 -1.21801448e+00
3.70376438e-01 4.38707769e-01 -4.74388808e-01 8.73847425e-01
-7.36868680e-02 5.13808787e-01 -2.29156420e-01 3.37461203e-01
-3.50547224e-01 3.04883748e-01 8.77522290e-01 -2.86996574e-03
-4.91148159e-02 -2.35759005e-01 -3.88839245e-01 3.64888340e-01
1.18319952e+00 -5.00884891e-01 1.17904097e-01 -2.10708305e-01
4.62864518e-01 1.82735957e-02 -5.23992889e-02 -6.82972729e-01
-1.67638153e-01 -4.39665139e-01 7.53319919e-01 -8.07061374e-01
2.34102055e-01 -4.87666607e-01 2.73277551e-01 9.73311841e-01
-5.19440413e-01 7.64653087e-01 3.29830080e-01 4.31762010e-01
1.29012138e-01 1.27740964e-01 8.15442085e-01 -3.27834100e-01
-4.93879288e-01 -2.39780005e-02 -1.00893307e+00 -7.87072480e-01
8.52006793e-01 -3.42605025e-01 -2.32437536e-01 2.26897770e-03
-1.27720356e+00 -3.49709839e-01 7.95253932e-01 3.63983750e-01
6.40361488e-01 -8.59589875e-01 -5.49036086e-01 3.22892308e-01
1.60459965e-01 -1.16925991e+00 -3.45821232e-01 5.18276513e-01
-1.15039408e+00 1.06590426e+00 -9.03026402e-01 -3.29998225e-01
-1.87090206e+00 8.88483644e-01 4.50569451e-01 -9.61688161e-02
-4.25217688e-01 5.14884830e-01 -5.69632351e-01 -6.08110487e-01
-2.63429433e-01 -5.63941598e-01 -1.21443029e-02 -4.97289926e-01
4.12175536e-01 1.08125463e-01 4.03443575e-01 -7.40654349e-01
-3.48256350e-01 2.84206271e-01 -2.61285007e-01 2.92983353e-01
1.48113215e+00 1.31338656e-01 -6.22498512e-01 3.34034383e-01
1.12690878e+00 -5.20045936e-01 -8.91644537e-01 -1.13615230e-01
2.07113013e-01 1.53774336e-01 -4.02988136e-01 -1.44898105e+00
-1.10136485e-02 8.12013566e-01 8.85477960e-01 -4.86382872e-01
9.15274858e-01 -3.07963341e-01 1.14782536e+00 1.20248997e+00
4.16108817e-01 -9.72132027e-01 -5.59651852e-01 9.79241133e-01
5.62173843e-01 -1.02295852e+00 1.48214579e-01 -1.41276181e-01
-2.21793056e-01 1.57937622e+00 3.43750268e-01 2.27825586e-02
2.28047878e-01 1.78325519e-01 3.43478084e-01 1.03841797e-02
-1.52156568e+00 2.73681164e-01 6.12095632e-02 7.18029618e-01
1.08789790e+00 1.67640641e-01 -1.08884239e+00 3.72775018e-01
-2.47024879e-01 2.80470848e-01 2.68405020e-01 1.12362266e+00
-7.60434091e-01 -1.58568168e+00 -1.87223360e-01 3.69445443e-01
-8.51105750e-01 -5.57288766e-01 -1.13470292e+00 6.32640123e-02
1.55946031e-01 6.34157717e-01 -3.07464898e-01 -4.15137410e-01
-9.52634662e-02 7.00265706e-01 7.22691476e-01 -3.38879913e-01
-9.94940996e-01 -2.42604166e-01 2.45988563e-01 -6.89353585e-01
-4.65066940e-01 -3.67955953e-01 -1.44594383e+00 -3.69884104e-01
-2.08594829e-01 4.10616547e-01 8.37030590e-01 8.37561965e-01
1.04016542e+00 1.63362429e-01 1.41955033e-01 -7.58723855e-01
-8.36434543e-01 -1.05857098e+00 2.00331733e-02 3.14944834e-02
3.25471789e-01 -1.47458300e-01 1.31348714e-01 3.81804556e-01] | [4.72514009475708, 5.600780963897705] |
2eaad1ce-78df-4120-a2e7-05d17b259811 | real-time-online-multi-object-tracking-in | 2204.02081 | null | https://arxiv.org/abs/2204.02081v1 | https://arxiv.org/pdf/2204.02081v1.pdf | Real-time Online Multi-Object Tracking in Compressed Domain | Recent online Multi-Object Tracking (MOT) methods have achieved desirable tracking performance. However, the tracking speed of most existing methods is rather slow. Inspired from the fact that the adjacent frames are highly relevant and redundant, we divide the frames into key and non-key frames respectively and track objects in the compressed domain. For the key frames, the RGB images are restored for detection and data association. To make data association more reliable, an appearance Convolutional Neural Network (CNN) which can be jointly trained with the detector is proposed. For the non-key frames, the objects are directly propagated by a tracking CNN based on the motion information provided in the compressed domain. Compared with the state-of-the-art online MOT methods,our tracker is about 6x faster while maintaining a comparable tracking performance. | ['Nenghai Yu', 'Weihai Li', 'Yue Wu', 'Bin Liu', 'Qiankun Liu'] | 2022-04-05 | null | null | null | null | ['online-multi-object-tracking'] | ['computer-vision'] | [-1.60974577e-01 -3.70047003e-01 -3.33546460e-01 1.84089869e-01
-3.72289658e-01 -4.11533177e-01 1.61289215e-01 1.14896119e-01
-6.08101189e-01 5.50823867e-01 -3.08268249e-01 1.90638512e-01
1.89169958e-01 -4.88407642e-01 -7.11011827e-01 -7.97714591e-01
7.97831416e-02 1.98084071e-01 1.08004928e+00 3.21499586e-01
-9.94140580e-02 4.30589467e-01 -1.57915211e+00 -1.73450261e-01
5.13184607e-01 1.33978522e+00 4.45603371e-01 6.91048741e-01
-1.17945792e-02 8.45501184e-01 -4.63283062e-01 -7.35421851e-02
2.20989421e-01 -2.60746896e-01 -2.60161132e-01 1.15049198e-01
8.10094357e-01 -6.22402787e-01 -8.29276204e-01 1.36027777e+00
2.92917311e-01 6.36664778e-02 1.35607004e-01 -1.47342348e+00
-6.36891425e-01 5.57413511e-02 -7.53043711e-01 3.52042019e-01
2.04968631e-01 9.22544748e-02 5.22796154e-01 -6.93737626e-01
6.77823842e-01 1.21984804e+00 6.88998759e-01 7.11270154e-01
-7.31305182e-01 -1.01492870e+00 4.30373102e-01 2.98607051e-01
-1.29458880e+00 -4.10791665e-01 6.56672657e-01 -2.39327133e-01
4.77737278e-01 -8.77979472e-02 9.96865451e-01 6.41246676e-01
3.55563760e-01 1.00873935e+00 5.11737406e-01 -1.12535998e-01
-1.67847835e-02 -1.47137716e-01 -1.19685866e-01 8.33872080e-01
5.06051540e-01 4.22474980e-01 -6.53559506e-01 1.35575756e-02
1.07639134e+00 4.83612686e-01 -1.45872846e-01 -4.19220477e-01
-1.54652178e+00 4.17528778e-01 6.92029893e-01 3.08747053e-01
-5.26638269e-01 7.31352985e-01 3.08194935e-01 1.08652115e-01
1.81987524e-01 -2.84259796e-01 -2.80226827e-01 -2.72205323e-02
-1.02800798e+00 1.13112040e-01 3.58065635e-01 1.17833459e+00
6.99192166e-01 1.77773982e-01 -4.13343638e-01 2.12621316e-01
8.54692161e-01 8.98184299e-01 2.26158023e-01 -9.69902396e-01
1.58694297e-01 5.77867746e-01 3.78174394e-01 -1.06661475e+00
-2.23235518e-01 -3.96091133e-01 -7.17951834e-01 3.77362430e-01
5.18691897e-01 -1.35997772e-01 -6.68799996e-01 1.49840856e+00
8.49502742e-01 4.97138649e-01 -8.79876986e-02 1.02872217e+00
6.69690728e-01 4.21154916e-01 1.15247041e-01 -4.69290644e-01
1.30753338e+00 -1.16650999e+00 -1.00804591e+00 -3.82497385e-02
1.96580127e-01 -8.29138696e-01 1.95225310e-02 3.18436563e-01
-1.13169527e+00 -1.09361362e+00 -1.07952464e+00 1.30476817e-01
-1.89796656e-01 2.96643376e-01 6.34732485e-01 3.56624812e-01
-9.54885483e-01 4.43616509e-01 -1.20089531e+00 -3.64158452e-01
4.45105463e-01 4.84327346e-01 -9.96812657e-02 -6.99383393e-02
-7.84773529e-01 6.79008782e-01 4.92337734e-01 1.27776340e-01
-1.00301194e+00 -3.62923175e-01 -7.19000101e-01 -3.82742554e-01
2.46011600e-01 -6.52055442e-01 1.13809133e+00 -8.88540506e-01
-1.32793379e+00 4.51526314e-01 -3.11276048e-01 -4.69581038e-01
5.96106648e-01 -3.89987946e-01 -4.92485881e-01 2.29866594e-01
7.11753070e-02 8.65740180e-01 1.06283081e+00 -8.99832547e-01
-1.27116549e+00 -2.35974774e-01 -6.25714809e-02 7.73335472e-02
-5.55118501e-01 3.54966104e-01 -1.02554166e+00 -5.22064328e-01
2.22687989e-01 -9.90428984e-01 -4.85093659e-03 1.00814664e+00
-1.95168242e-01 -3.11305970e-01 1.71449935e+00 -4.59867597e-01
1.11324441e+00 -2.34164023e+00 9.33454782e-02 -5.12433350e-01
4.85886127e-01 3.21573973e-01 3.74885984e-02 -1.75568476e-01
4.21678275e-01 -7.19903767e-01 4.48395669e-01 -5.13737082e-01
-2.35820606e-01 1.42158583e-01 -3.03527564e-02 9.07997787e-01
-1.35006979e-01 9.01623011e-01 -1.22957814e+00 -8.75288963e-01
6.79457307e-01 5.77299595e-01 -1.94441035e-01 1.42431989e-01
-2.68676490e-01 7.62324750e-01 -6.81041181e-01 9.62114334e-01
8.32793653e-01 -5.28103232e-01 -1.31252483e-01 -5.32138050e-01
-5.16549885e-01 -2.23795250e-01 -1.36396229e+00 1.68927908e+00
-9.70979873e-03 8.66743863e-01 2.28952602e-01 -5.03891826e-01
6.89195275e-01 3.50142449e-01 9.79162514e-01 -5.60717165e-01
2.95801520e-01 5.42683937e-02 -7.66839981e-02 -2.90293664e-01
7.53808022e-01 3.07652771e-01 3.37185621e-01 1.53318703e-01
-1.44772828e-02 5.32624602e-01 1.33479089e-01 8.15012753e-02
1.03890300e+00 4.67566311e-01 9.01144743e-02 9.79286358e-02
5.52890897e-01 9.72260833e-02 9.35699582e-01 6.39233530e-01
-6.85655594e-01 3.81468892e-01 -5.85283041e-01 -6.26551986e-01
-9.58804190e-01 -9.74810004e-01 -1.97258387e-02 8.22049201e-01
9.93143559e-01 -3.83480519e-01 -2.41711318e-01 -4.40735370e-01
2.26478785e-01 -7.65307620e-02 -4.38219696e-01 -1.30312353e-01
-7.21040666e-01 -3.17186862e-01 3.07059824e-01 8.68726730e-01
7.76044667e-01 -9.51429009e-01 -1.07677734e+00 6.87397778e-01
-7.58706778e-02 -1.36726224e+00 -6.25211000e-01 -9.59760323e-02
-9.46797669e-01 -1.17038620e+00 -8.40819955e-01 -7.03484595e-01
5.51963806e-01 8.25457931e-01 5.26605785e-01 4.69164908e-01
-1.63771540e-01 3.10632586e-01 -3.06454092e-01 -4.14011061e-01
-1.26719728e-01 -4.63558704e-01 3.45054686e-01 5.21787480e-02
3.22320342e-01 -2.26778924e-01 -8.50704551e-01 3.44104439e-01
-6.06611848e-01 7.37111494e-02 5.23009717e-01 5.44123888e-01
8.22917223e-01 3.49278264e-02 1.30143836e-01 3.99383903e-02
-4.39422339e-01 -1.21176966e-01 -9.91475046e-01 2.21044898e-01
-3.73496205e-01 -2.45874971e-01 4.34835255e-01 -7.94894576e-01
-7.16896057e-01 6.55160964e-01 3.30499232e-01 -1.01478028e+00
8.85925367e-02 -2.85147965e-01 2.57806510e-01 -5.82654417e-01
6.54377043e-02 3.21669996e-01 -1.38443425e-01 -3.41982245e-01
2.65322596e-01 5.00910580e-01 7.73694098e-01 -1.59885779e-01
9.96763825e-01 8.57194662e-01 -1.77084822e-02 -5.77407956e-01
-7.44645238e-01 -7.34737813e-01 -7.46606171e-01 -7.74832428e-01
1.15711057e+00 -1.15823817e+00 -9.03319955e-01 6.14726305e-01
-1.36560607e+00 -7.39751756e-02 -8.81652460e-02 8.62804294e-01
-3.42767596e-01 4.88765091e-01 -7.36390889e-01 -1.06376779e+00
-4.21669424e-01 -1.04509270e+00 1.26740599e+00 7.87655056e-01
3.35365623e-01 -7.40871191e-01 -5.24664000e-02 -1.57474205e-01
5.87309241e-01 1.90324992e-01 -3.23173255e-02 -1.01171292e-01
-1.23585546e+00 -2.58777708e-01 -4.29489613e-01 -2.69893587e-01
2.12177187e-01 2.17822045e-01 -8.38251173e-01 -6.19752884e-01
-2.96540499e-01 -1.16901197e-01 8.44886601e-01 6.49290204e-01
8.56618345e-01 8.36182535e-02 -1.08414817e+00 7.25539029e-01
1.50254428e+00 3.93132538e-01 3.05479169e-01 5.16259968e-01
8.28300357e-01 -1.29961029e-01 1.00290656e+00 3.91080678e-01
3.99333686e-01 9.07644570e-01 6.83785379e-01 -4.46231328e-02
-4.91224110e-01 -6.37757555e-02 5.54291785e-01 7.00141609e-01
3.69678736e-02 -1.94172233e-01 -2.51525700e-01 5.31799018e-01
-2.20308733e+00 -9.09900725e-01 -2.29104310e-01 2.11224675e+00
5.13308227e-01 9.53617021e-02 1.75104916e-01 -3.40909302e-01
9.18332160e-01 2.21846968e-01 -8.04223537e-01 5.72168708e-01
-1.94196671e-01 -3.23011905e-01 7.57343173e-01 6.15067082e-03
-1.35971856e+00 7.00754702e-01 6.07162952e+00 4.81323630e-01
-9.82326925e-01 2.91808128e-01 5.98967355e-03 -1.77924842e-01
4.16044176e-01 1.19895868e-01 -1.30452192e+00 6.90028250e-01
6.30505621e-01 -3.26152518e-02 9.59053263e-02 9.55242395e-01
-2.18663774e-02 -2.50252694e-01 -9.77783203e-01 1.18604171e+00
-6.59552589e-02 -1.28824127e+00 -4.42407101e-01 1.12123303e-01
7.67435610e-01 3.16510230e-01 -5.76745532e-03 1.20252088e-01
3.50940645e-01 -2.89461315e-01 1.04865754e+00 7.45516062e-01
5.97050488e-01 -4.91454601e-01 5.36578357e-01 2.32255116e-01
-2.12419438e+00 -1.90170318e-01 -7.66963720e-01 1.45775825e-01
3.70583564e-01 4.14740950e-01 3.59992310e-02 6.70470536e-01
1.12508321e+00 1.21985281e+00 -5.50812006e-01 1.56038618e+00
1.22696005e-01 1.42662004e-01 -5.46117306e-01 2.86722183e-02
1.73968911e-01 -3.14201415e-02 6.54766023e-01 8.53990078e-01
5.45970857e-01 -4.72204313e-02 6.17551148e-01 7.56288528e-01
4.67062071e-02 -3.78765315e-01 -4.04236466e-01 2.21673891e-01
7.73552179e-01 1.54400074e+00 -8.45942736e-01 -6.38351917e-01
-8.46382022e-01 1.13885224e+00 7.57606700e-02 5.83393648e-02
-1.12687492e+00 -1.11988448e-01 5.39418936e-01 -2.06741139e-01
8.09597611e-01 -3.59814495e-01 4.12715852e-01 -1.03418529e+00
6.40601367e-02 -4.01229382e-01 3.90899420e-01 -8.30921233e-01
-1.05030501e+00 4.47839797e-01 -3.36339891e-01 -1.83104265e+00
1.33426726e-01 -5.56296229e-01 -3.82111818e-01 4.42064822e-01
-1.73818231e+00 -1.37067437e+00 -7.32057691e-01 6.74093544e-01
4.33948666e-01 9.77724716e-02 4.51214671e-01 6.44978225e-01
-6.91857815e-01 5.18135607e-01 3.63238573e-01 3.41523498e-01
7.90452540e-01 -9.37851489e-01 2.36788273e-01 1.02324724e+00
2.29557622e-02 4.42102790e-01 4.28973407e-01 -9.55460906e-01
-1.75464129e+00 -1.38263810e+00 4.35319066e-01 -2.36755863e-01
6.66474760e-01 -1.45208957e-02 -7.38037765e-01 6.33101106e-01
1.60093501e-01 6.65808618e-01 4.42791395e-02 -6.06975675e-01
-1.57395448e-03 -9.60133672e-02 -8.41820240e-01 3.12868953e-01
1.07671583e+00 -2.96295583e-01 -2.88455009e-01 3.75718623e-01
7.32288897e-01 -7.81563282e-01 -8.15607965e-01 2.04584002e-01
7.35596418e-01 -5.20580947e-01 1.08052754e+00 -3.07583809e-02
-2.76935518e-01 -1.02337193e+00 -8.80934149e-02 -6.98786914e-01
-4.63586777e-01 -5.84295988e-01 -7.73543954e-01 1.03729439e+00
-3.81761372e-01 -3.70258898e-01 8.55357111e-01 2.07049340e-01
-5.65699115e-02 -2.76608020e-01 -1.19410598e+00 -1.02646184e+00
-6.02087915e-01 -1.20734230e-01 2.80070305e-01 6.63267732e-01
-4.24183518e-01 -2.39835400e-02 -6.35750175e-01 4.59316671e-01
1.23861897e+00 4.86599773e-01 7.37429321e-01 -1.45278597e+00
-2.97711164e-01 -1.26721367e-01 -5.58294773e-01 -1.57967389e+00
-2.38600105e-01 -4.81770277e-01 2.46039972e-01 -1.32341027e+00
3.87932122e-01 -4.63591903e-01 -5.26995838e-01 4.72589850e-01
-3.21882248e-01 4.53232765e-01 4.23078746e-01 4.95631754e-01
-1.34187293e+00 6.03523433e-01 1.33073986e+00 -2.29603171e-01
1.03551514e-01 -1.42844856e-01 -6.11567274e-02 5.89619339e-01
3.24450940e-01 -6.35949194e-01 2.30854765e-01 -5.34488380e-01
-2.90489525e-01 1.62326589e-01 6.36877716e-01 -1.55402386e+00
1.09966445e+00 1.18342504e-01 7.76278913e-01 -1.41822553e+00
5.49002767e-01 -1.11227000e+00 4.13974375e-01 9.11484718e-01
1.36041269e-01 3.22668791e-01 3.08690637e-01 9.05446649e-01
-1.54985800e-01 2.15990260e-01 7.85636544e-01 7.54445642e-02
-9.75603759e-01 8.36096406e-01 -1.94214955e-01 -4.55944449e-01
1.23863578e+00 -5.18776238e-01 -4.13545310e-01 -1.15911342e-01
-5.21401882e-01 3.62365901e-01 6.31757677e-01 5.29306650e-01
6.92601323e-01 -1.70844150e+00 -3.70068163e-01 1.03940228e-02
6.64423928e-02 5.46413809e-02 3.00097644e-01 1.30959201e+00
-1.73526257e-01 3.96163523e-01 -3.30982685e-01 -9.33542907e-01
-1.42230320e+00 9.44629550e-01 3.89894783e-01 1.10076480e-01
-1.02147198e+00 6.34777963e-01 1.23797618e-01 3.54409546e-01
4.88532186e-01 -2.87700236e-01 -5.95030487e-02 -1.63553655e-01
9.60919619e-01 3.95374805e-01 -4.91150737e-01 -8.14491928e-01
-6.07585192e-01 8.39715004e-01 -4.23891991e-02 2.24825233e-01
1.10313070e+00 -3.03977966e-01 4.57313508e-02 1.89965501e-01
9.51993465e-01 -2.09220186e-01 -1.81734776e+00 -5.96808791e-01
-2.71831602e-01 -6.68363035e-01 2.54867226e-01 -1.60183564e-01
-1.49669361e+00 5.36146581e-01 1.07486951e+00 -7.32177263e-03
1.05252719e+00 1.58924814e-02 9.25124347e-01 2.19707116e-01
7.22183406e-01 -1.03515112e+00 2.07882687e-01 3.92854929e-01
2.96268493e-01 -1.26393199e+00 2.33329162e-01 -2.61693448e-01
-2.55668480e-02 1.31044233e+00 8.35308015e-01 -9.74291712e-02
5.14701962e-01 4.76187378e-01 8.20420962e-03 -2.51792997e-01
-4.76401567e-01 -3.51086229e-01 -1.83966011e-02 7.44790435e-01
5.92303686e-02 -3.33677500e-01 1.00803278e-01 -7.76043534e-02
5.19757748e-01 1.45737752e-01 1.02178594e-02 9.56828296e-01
-6.00998938e-01 -9.27978694e-01 -6.33079410e-01 1.61484569e-01
-2.80295283e-01 2.37890035e-01 6.21246584e-02 8.53508770e-01
2.94960529e-01 8.99802327e-01 2.75918037e-01 -3.80361915e-01
1.21446766e-01 -4.56527293e-01 6.20617270e-01 -2.50607342e-01
-5.00015080e-01 4.13190663e-01 -3.17683548e-01 -9.35218632e-01
-9.04557109e-01 -8.95274758e-01 -1.42271066e+00 -2.84554601e-01
-8.17419350e-01 -2.19177529e-01 4.43110764e-01 8.65373194e-01
2.54211098e-01 8.45607042e-01 4.45873886e-01 -1.12848604e+00
-1.60320252e-01 -7.94321895e-01 -4.82240736e-01 1.56028986e-01
6.36154652e-01 -1.06646204e+00 -2.69459728e-02 9.15080383e-02] | [6.422514915466309, -2.1157805919647217] |
5d1ae32b-5a52-4af8-9aeb-2a53432acdc4 | explaining-dialogue-evaluation-metrics-using | null | null | https://aclanthology.org/2022.naacl-main.430 | https://aclanthology.org/2022.naacl-main.430.pdf | Explaining Dialogue Evaluation Metrics using Adversarial Behavioral Analysis | There is an increasing trend in using neural methods for dialogue model evaluation. Lack of a framework to investigate these metrics can cause dialogue models to reflect their biases and cause unforeseen problems during interactions. In this work, we propose an adversarial test-suite which generates problematic variations of various dialogue aspects, e.g. logical entailment, using automatic heuristics. We show that dialogue metrics for both open-domain and task-oriented settings are biased in their assessments of different conversation behaviors and fail to properly penalize problematic conversations, by analyzing their assessments of these problematic examples. We conclude that variability in training methodologies and data-induced biases are some of the main causes of these problems. We also conduct an investigation into the metric behaviors using a black-box interpretability model which corroborates our findings and provides evidence that metrics pay attention to the problematic conversational constructs signaling a misunderstanding of different conversation semantics. | ['Sungjin Lee', 'Baber Khalid'] | null | null | null | null | naacl-2022-7 | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 2.61353076e-01 5.95596254e-01 1.39269128e-01 -7.16657400e-01
-6.10570908e-01 -8.36119473e-01 1.03131449e+00 1.22845225e-01
-3.95234555e-01 1.06323004e+00 7.70024657e-01 -6.47132099e-01
-1.09473117e-01 -6.84425414e-01 -2.90870905e-01 -1.78284734e-01
1.21036157e-01 7.32529521e-01 -2.10579857e-01 -6.35494590e-01
4.68638599e-01 -1.00494444e-01 -1.13044238e+00 6.45805299e-01
1.04317963e+00 4.42164451e-01 -6.00673676e-01 9.56914544e-01
-2.34898448e-01 1.14797628e+00 -1.39115059e+00 -9.24725413e-01
4.82529290e-02 -6.69279575e-01 -1.32807958e+00 2.43199952e-02
4.17016476e-01 -5.02369165e-01 -3.26018557e-02 9.32549834e-01
6.15604579e-01 4.40350398e-02 8.90182674e-01 -1.44584608e+00
-2.85764992e-01 1.08722532e+00 5.46573550e-02 2.26856813e-01
6.94442689e-01 5.67985594e-01 1.17373347e+00 -3.20681244e-01
7.02074111e-01 1.63397264e+00 8.58989358e-01 9.22599971e-01
-1.36136258e+00 -4.01939809e-01 -3.17230821e-03 -6.17038384e-02
-6.44029975e-01 -5.83134115e-01 6.44571304e-01 -4.63190824e-01
9.24575150e-01 6.44108415e-01 4.64846730e-01 1.84033275e+00
7.19894245e-02 4.58854914e-01 1.51976144e+00 -2.60876775e-01
1.19378857e-01 6.47170067e-01 5.30771613e-01 3.00430983e-01
3.28992188e-01 8.78000483e-02 -4.29998726e-01 -6.23555124e-01
2.82121360e-01 -6.83963180e-01 -3.44643056e-01 9.81154889e-02
-9.33771312e-01 1.14747369e+00 -3.02755064e-03 3.36872131e-01
-1.34329811e-01 -2.13900343e-01 8.46763134e-01 7.37419128e-01
5.05493402e-01 1.49205005e+00 -6.12556398e-01 -7.30792165e-01
-3.98362279e-01 6.15281880e-01 1.65994334e+00 6.55909836e-01
3.31789106e-01 -5.32396091e-03 -5.30147076e-01 1.06723821e+00
5.41425683e-02 -1.44120527e-03 5.61483800e-01 -1.21759403e+00
6.20424867e-01 5.82380056e-01 3.21061313e-01 -9.92380202e-01
-5.79114676e-01 -3.35149989e-02 -3.34980994e-01 1.31536022e-01
1.10650182e+00 -6.10398769e-01 -1.15504898e-01 1.82507324e+00
-4.50809225e-02 -7.21749187e-01 3.83306414e-01 7.18460917e-01
8.98966849e-01 2.33560517e-01 2.95210391e-01 -1.30676374e-01
1.31124842e+00 -7.54733205e-01 -7.21001267e-01 -2.67230362e-01
9.35203195e-01 -6.54882014e-01 1.69713986e+00 3.38741124e-01
-1.22540438e+00 -1.60339296e-01 -9.10179973e-01 -1.32311257e-02
-2.64522046e-01 -5.38863003e-01 6.47096336e-01 8.90443027e-01
-7.38299072e-01 7.47846544e-01 -1.44231156e-01 -4.18550819e-01
-6.20853081e-02 3.98022160e-02 5.81887066e-02 3.93584758e-01
-1.38466609e+00 1.63293159e+00 1.09019734e-01 -8.72830525e-02
-6.15531981e-01 -4.76752996e-01 -7.80449927e-01 2.71736700e-02
2.46578127e-01 -4.10815030e-01 1.59628689e+00 -1.25802290e+00
-1.55942440e+00 8.94722223e-01 1.90109029e-01 -4.72483873e-01
9.22139764e-01 -1.46943107e-01 -3.23763251e-01 -3.54772359e-01
-8.15795735e-02 3.93435389e-01 3.71186286e-01 -1.37621391e+00
-2.58344889e-01 -6.99293613e-02 5.70972383e-01 2.90166318e-01
-2.97494888e-01 -6.42880946e-02 5.75305581e-01 -4.90109980e-01
-4.02543366e-01 -8.49501431e-01 -1.08599000e-01 -4.98663932e-01
-7.50086308e-01 -3.47751528e-01 2.43185669e-01 -4.88220543e-01
1.35903466e+00 -1.44551682e+00 4.05392684e-02 1.28133520e-01
3.29988927e-01 2.01131016e-01 -6.83797598e-02 6.23554587e-01
1.54038548e-01 6.29496932e-01 -1.68214068e-01 -2.95548946e-01
3.53453428e-01 2.10830405e-01 -3.19187045e-01 1.39410108e-01
2.33787626e-01 6.91818595e-01 -8.75227809e-01 -3.48301858e-01
7.41402879e-02 1.03344440e-01 -7.19510138e-01 5.03824472e-01
-4.56580371e-01 4.98547345e-01 -1.45622402e-01 1.97360620e-01
3.78090948e-01 5.94570488e-02 2.78510273e-01 5.07916473e-02
1.54637560e-01 1.04892051e+00 -5.61046004e-01 1.01121128e+00
-5.98273396e-01 9.72173631e-01 -1.03797704e-01 -5.16166210e-01
8.68543208e-01 3.15952271e-01 -2.35773578e-01 -4.97031957e-01
3.48265201e-01 1.87874526e-01 6.50714934e-01 -7.85344005e-01
8.64944935e-01 -3.80750090e-01 -1.66604340e-01 9.04389739e-01
5.00696898e-03 -4.13571328e-01 9.86377224e-02 1.35219619e-01
1.13628542e+00 -3.47091973e-01 2.07973495e-01 -4.02843475e-01
4.30084258e-01 1.86388299e-01 3.99840027e-01 1.12553024e+00
-4.52189952e-01 4.51511592e-01 1.19156325e+00 -4.39459503e-01
-1.10659528e+00 -7.16047287e-01 -1.62454367e-01 1.10131907e+00
-4.03081149e-01 -3.77131075e-01 -9.87055719e-01 -9.91712511e-01
-9.06483606e-02 1.47483909e+00 -7.47759223e-01 -4.12209511e-01
-3.77213359e-01 -6.96864307e-01 1.15902221e+00 1.32838354e-01
2.77707517e-01 -1.23432291e+00 -5.99771976e-01 1.19522274e-01
-4.92850274e-01 -8.73503804e-01 -1.16599821e-01 1.23471491e-01
-6.53780639e-01 -1.26751244e+00 -2.59866804e-01 -2.14085579e-01
4.88233119e-01 -2.22037584e-01 1.74888194e+00 3.32550138e-01
3.60669136e-01 3.75741929e-01 -3.65975142e-01 -6.25117660e-01
-1.42343807e+00 2.35400692e-01 -1.82066336e-01 -4.51962262e-01
6.63238525e-01 -3.27231377e-01 -3.59225035e-01 5.65863788e-01
-6.94875300e-01 -2.02325195e-01 8.83389562e-02 1.02048481e+00
-6.43904150e-01 -5.09126842e-01 7.65044034e-01 -1.45744622e+00
1.76187599e+00 -6.84129179e-01 -1.33152157e-01 1.57783523e-01
-8.33881199e-01 2.32244238e-01 6.28955364e-01 -4.42444801e-01
-1.15103900e+00 -8.29302013e-01 -1.61505431e-01 1.88488796e-01
-4.00545746e-01 3.40228260e-01 -6.69995463e-03 1.38682052e-01
1.17869365e+00 -2.10462287e-01 3.24725598e-01 -7.31482506e-02
3.30268368e-02 7.73604572e-01 -5.77671304e-02 -9.89046037e-01
4.84896123e-01 -3.41628566e-02 -5.21337152e-01 -7.71636128e-01
-8.50026608e-01 1.13775201e-01 -5.86912930e-02 -3.46521288e-01
5.75688362e-01 -4.33259130e-01 -8.74357820e-01 2.72106141e-01
-1.41675687e+00 -6.04970574e-01 -1.29441693e-01 1.45128876e-01
-5.75432897e-01 4.02079791e-01 -8.79497528e-01 -9.78642583e-01
-2.97567576e-01 -1.33267522e+00 3.67316693e-01 1.74767748e-01
-1.42257559e+00 -1.25297523e+00 1.91126242e-01 7.45106101e-01
6.73186421e-01 2.46995091e-01 1.11399972e+00 -1.46058774e+00
1.45689607e-01 2.63964813e-02 2.81395968e-02 4.55895394e-01
3.92401172e-03 2.33725801e-01 -1.25837505e+00 3.69104967e-02
3.30330729e-01 -9.32418346e-01 2.96844721e-01 -5.44310957e-02
6.95917487e-01 -8.50279748e-01 3.12484831e-01 1.73911098e-02
7.31913388e-01 1.34024873e-01 5.81379235e-01 4.51944113e-01
3.46255124e-01 1.38918090e+00 4.68724221e-01 3.09151471e-01
4.86543745e-01 5.53024888e-01 1.70333222e-01 2.05728740e-01
3.41700584e-01 -1.56449422e-01 4.50859696e-01 4.22216445e-01
1.00905970e-01 -4.74584371e-01 -9.86714542e-01 3.90426874e-01
-1.52077937e+00 -1.05255687e+00 -2.84939468e-01 2.11411953e+00
1.22213888e+00 5.72529018e-01 3.06522101e-01 6.63721338e-02
6.53337300e-01 3.56579125e-01 -3.51496905e-01 -1.23036456e+00
-4.19174060e-02 -1.01094432e-01 1.68526545e-01 8.68985534e-01
-5.64166605e-01 6.90925419e-01 7.04485607e+00 2.72869587e-01
-7.34732926e-01 -1.38319507e-01 9.45005417e-01 3.61589752e-02
-7.84828842e-01 -1.45984396e-01 -3.35568041e-01 4.45859611e-01
1.12850392e+00 -6.00211248e-02 5.10114849e-01 6.16610467e-01
2.77370721e-01 -6.04651794e-02 -1.49928999e+00 1.97137833e-01
5.75335547e-02 -9.61724639e-01 1.53325006e-01 6.66577592e-02
4.23488289e-01 -3.59654337e-01 -1.59227252e-01 6.49124980e-01
7.77337253e-01 -1.22967362e+00 5.49981475e-01 1.54406175e-01
1.94171309e-01 -6.95100188e-01 9.06215847e-01 2.39260316e-01
1.24819010e-01 3.61199081e-02 -2.28585452e-01 -5.55866599e-01
-1.35456011e-01 1.83289543e-01 -1.43417490e+00 -1.97003648e-01
3.19436163e-01 3.44091430e-02 -4.51023936e-01 2.85376936e-01
-1.87228218e-01 8.52528036e-01 1.43886805e-02 -4.96309370e-01
3.01198632e-01 -5.58261871e-02 7.34601140e-01 1.43128240e+00
-3.47227484e-01 -1.88762084e-01 -3.34123433e-01 1.10943294e+00
6.02820218e-02 -5.16226999e-02 -9.31946933e-01 -1.46044105e-01
5.60947001e-01 1.18731606e+00 -3.77549678e-01 -1.86754212e-01
-3.34299535e-01 5.85080504e-01 4.45109069e-01 2.13041052e-01
-8.57747614e-01 -2.25706890e-01 8.83666575e-01 -1.88560128e-01
-3.41319740e-01 1.58776268e-01 -8.68920207e-01 -9.82050717e-01
-7.87606686e-02 -1.58827937e+00 2.63313025e-01 -4.71483797e-01
-1.58077073e+00 6.56150281e-01 4.11329530e-02 -8.19067597e-01
-5.18799961e-01 -4.61991519e-01 -1.10442138e+00 8.47388208e-01
-1.08127928e+00 -5.68619847e-01 -3.34577322e-01 2.54371524e-01
6.66273713e-01 -2.14170828e-01 9.89039063e-01 -1.66310638e-01
-5.59602141e-01 8.62381577e-01 -3.46791506e-01 2.56392986e-01
9.93209660e-01 -1.49537528e+00 5.23514688e-01 2.08048493e-01
-2.55548239e-01 7.52248824e-01 1.32497239e+00 -4.39121157e-01
-7.68886566e-01 -6.02139533e-01 9.32263434e-01 -9.24795985e-01
7.60611236e-01 -2.17833310e-01 -1.02432621e+00 5.85363805e-01
6.49498343e-01 -1.01097643e+00 1.02131045e+00 5.65914512e-01
-4.16718245e-01 3.91523123e-01 -1.47788954e+00 9.71375585e-01
1.02072728e+00 -5.29443264e-01 -9.30745482e-01 3.25563133e-01
6.94044650e-01 -5.02674580e-01 -8.65158975e-01 1.03162475e-01
5.09400427e-01 -1.39507747e+00 5.82812786e-01 -1.14560223e+00
1.05278587e+00 4.56515282e-01 4.04647365e-02 -1.73034167e+00
2.05211744e-01 -7.49265075e-01 2.54630148e-01 1.46783280e+00
8.85175526e-01 -8.71633768e-01 5.87651074e-01 1.28332222e+00
-5.89886010e-02 -5.72047651e-01 -6.07520580e-01 -2.71351010e-01
5.48171520e-01 -4.06995445e-01 5.42687058e-01 1.25318921e+00
6.07750535e-01 7.03182638e-01 -2.41363883e-01 -3.45205337e-01
2.47314453e-01 -3.11894327e-01 1.02431822e+00 -9.95981455e-01
-2.55297899e-01 -6.84071004e-01 -7.34523162e-02 -5.63948989e-01
3.94088835e-01 -5.48334599e-01 1.38119072e-01 -1.14376903e+00
-1.13480143e-01 -4.71121162e-01 3.08826268e-01 6.31791577e-02
-2.48787940e-01 -2.04743549e-01 1.73627660e-01 -7.45833889e-02
-3.02027464e-01 4.21714365e-01 1.16055000e+00 -1.05247416e-01
-1.49416134e-01 8.44818130e-02 -1.13511264e+00 8.59270513e-01
1.03174794e+00 -3.29088211e-01 -4.90853608e-01 -3.50516111e-01
2.88209647e-01 6.80453330e-02 3.28194082e-01 -8.05090368e-01
-1.41543925e-01 -2.91311830e-01 -9.24998615e-03 2.13862881e-01
3.35676014e-01 -6.15319550e-01 -3.67768377e-01 3.71124297e-01
-1.20489252e+00 2.37434506e-01 2.07977176e-01 2.97107041e-01
-1.59149114e-02 -4.75294024e-01 5.25278687e-01 -3.82681459e-01
4.34756838e-02 -5.73307812e-01 -8.60685289e-01 7.68529415e-01
7.75997341e-01 -2.60218441e-01 -6.88472211e-01 -8.99549842e-01
-4.61579859e-01 4.05791670e-01 4.71465021e-01 5.23412287e-01
1.44562557e-01 -8.00433517e-01 -9.06848431e-01 -1.27279237e-01
1.22550406e-01 -2.98011422e-01 -1.12950057e-01 6.05102062e-01
-6.20220542e-01 4.12077248e-01 -2.65423059e-01 -4.01257396e-01
-1.23735380e+00 -1.09191895e-01 6.83096170e-01 -4.29128081e-01
-1.32260442e-01 7.01706946e-01 -3.10771726e-02 -8.75829101e-01
4.40248072e-01 -3.97141337e-01 -2.84975171e-01 1.84402511e-01
2.67679006e-01 5.53038895e-01 -2.87610628e-02 -3.14419776e-01
5.42027913e-02 -2.94363737e-01 -2.46675283e-01 -3.19258243e-01
8.57137024e-01 -6.02043979e-03 4.04986255e-02 7.21035063e-01
1.00312817e+00 8.72451514e-02 -9.29996490e-01 2.80869938e-02
7.02804774e-02 -3.98893356e-01 -5.23551166e-01 -1.19814956e+00
-4.53789532e-01 7.06661701e-01 -1.40751656e-02 8.32617939e-01
2.82550812e-01 -3.76731962e-01 3.79006147e-01 7.09906399e-01
2.57477134e-01 -1.28131485e+00 2.94314306e-02 8.62350404e-01
1.16614449e+00 -1.43576896e+00 -1.60472512e-01 -2.79529035e-01
-1.23038554e+00 1.14029157e+00 1.23045552e+00 6.59705326e-02
2.43936256e-02 7.87201896e-02 5.62048495e-01 -3.03245157e-01
-1.11598802e+00 3.42155904e-01 -1.97862715e-01 5.58587193e-01
1.00931597e+00 1.15139455e-01 -6.47903204e-01 5.30965388e-01
-8.32423329e-01 -4.70501035e-01 1.08098435e+00 6.73123598e-01
4.75781485e-02 -9.87179399e-01 -3.74701530e-01 6.79919899e-01
-5.14901876e-01 -9.67036411e-02 -1.23644948e+00 9.88129556e-01
-5.26852489e-01 1.33746529e+00 7.69468322e-02 -6.98269367e-01
4.15388882e-01 4.23758805e-01 2.13728741e-01 -6.31837249e-01
-1.44906390e+00 -8.05119038e-01 1.20212138e+00 -3.62948626e-01
-1.97448626e-01 -6.13038540e-01 -7.87914455e-01 -8.13458920e-01
-4.88689333e-01 5.71758687e-01 2.94028372e-01 9.64348316e-01
1.66621521e-01 3.74169886e-01 5.48575759e-01 -5.36489725e-01
-1.17276204e+00 -1.40342343e+00 -5.77545948e-02 9.22699153e-01
2.01460898e-01 -5.27031243e-01 -7.68222809e-01 -5.55309594e-01] | [12.689497947692871, 8.116058349609375] |
db3dc823-c42f-4da0-ac7a-ccf4a5a30620 | rethinking-symmetric-matrix-factorization-a | 2209.02528 | null | https://arxiv.org/abs/2209.02528v2 | https://arxiv.org/pdf/2209.02528v2.pdf | Rethinking Symmetric Matrix Factorization: A More General and Better Clustering Perspective | Nonnegative matrix factorization (NMF) is widely used for clustering with strong interpretability. Among general NMF problems, symmetric NMF is a special one that plays an important role in graph clustering where each element measures the similarity between data points. Most existing symmetric NMF algorithms require factor matrices to be nonnegative, and only focus on minimizing the gap between similarity matrix and its approximation for clustering, without giving a consideration to other potential regularization terms which can yield better clustering. In this paper, we explore factorizing a symmetric matrix that does not have to be nonnegative, presenting an efficient factorization algorithm with a regularization term to boost the clustering performance. Moreover, a more general framework is proposed to solve symmetric matrix factorization problems with different constraints on the factor matrices. | ['Kai Liu', 'Mengyuan Zhang'] | 2022-09-06 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.80807654e-02 -1.99519143e-01 -3.15668434e-01 -1.24593481e-01
-1.92126572e-01 -6.85548365e-01 1.13791451e-01 5.43897115e-02
-2.52114117e-01 3.56739193e-01 1.45877168e-01 -3.49146217e-01
-6.63126230e-01 -6.70855105e-01 -2.72240102e-01 -8.53568494e-01
6.87800422e-02 5.52342355e-01 -1.68097720e-01 -2.58182585e-01
2.17735067e-01 3.67728651e-01 -1.13321412e+00 1.40098318e-01
1.20820260e+00 4.61264759e-01 4.12864953e-01 3.76716793e-01
-9.84502435e-02 2.78630704e-01 -5.99432051e-01 -3.97024542e-01
4.15913343e-01 -3.23198557e-01 -6.70769632e-01 4.88955498e-01
6.44201413e-02 4.45226282e-01 -1.04901850e-01 1.31778491e+00
8.63554478e-02 4.43579197e-01 7.72012353e-01 -1.72547507e+00
-7.61452615e-01 6.42390311e-01 -9.71386433e-01 -4.50237282e-03
1.97571769e-01 -3.35707992e-01 1.15815508e+00 -6.73591614e-01
4.30337936e-01 1.22838569e+00 3.60378623e-01 4.61204082e-01
-1.36941516e+00 -6.28530443e-01 4.62171435e-01 2.64650375e-01
-1.41341066e+00 4.91785929e-02 6.99544907e-01 -5.19867420e-01
5.17496943e-01 5.16204536e-01 7.87680566e-01 4.78149861e-01
-1.14382312e-01 6.75587177e-01 7.47502029e-01 -2.70399779e-01
2.34203979e-01 -1.78560913e-01 3.13477337e-01 3.16236585e-01
4.42306727e-01 -4.85104740e-01 -8.52921233e-02 -4.02135879e-01
6.07711196e-01 4.28243905e-01 -4.16282296e-01 -7.21538723e-01
-1.44984066e+00 9.90038633e-01 2.09165782e-01 4.88215953e-01
-3.28904837e-01 -3.33851506e-03 2.59059191e-01 2.18798622e-01
2.12994710e-01 4.79837537e-01 -2.05046028e-01 7.27880895e-02
-8.00426126e-01 -6.24430515e-02 3.76054496e-01 6.80911481e-01
9.42203581e-01 -3.36741097e-02 -1.92929879e-02 9.41974282e-01
4.25612181e-01 6.19161785e-01 4.38168526e-01 -1.02756178e+00
3.67710561e-01 1.07239878e+00 2.71526491e-03 -1.47480285e+00
-5.21539211e-01 -3.32966566e-01 -1.16785443e+00 -1.94860175e-01
1.95510328e-01 3.86010371e-02 -6.11367404e-01 1.78911805e+00
3.08993787e-01 4.16109353e-01 -1.05924129e-01 1.08068538e+00
5.41282952e-01 5.69453835e-01 -4.12345052e-01 -5.81890643e-01
1.05851829e+00 -8.61438513e-01 -8.26438725e-01 -1.14060692e-01
5.70650816e-01 -9.51767981e-01 1.12646258e+00 5.02072036e-01
-7.90908396e-01 -1.74495563e-01 -8.09215963e-01 1.75107479e-01
-1.04161978e-01 2.11492479e-01 1.01459718e+00 5.57073116e-01
-1.04587817e+00 2.20351443e-01 -7.73381233e-01 -2.97367156e-01
-2.52559662e-01 5.67784607e-01 -6.16834819e-01 -5.55153899e-02
-1.09451699e+00 4.26260412e-01 4.49061692e-01 3.60993057e-01
-9.14070159e-02 -5.18434763e-01 -7.00169086e-01 6.41697198e-02
4.73170787e-01 -8.60552669e-01 7.30738878e-01 -8.46912205e-01
-1.11840117e+00 5.45984387e-01 -4.88952905e-01 -2.19949856e-01
3.07744831e-01 8.89844000e-02 -5.03389299e-01 1.07461728e-01
2.57814765e-01 3.14626813e-01 1.03966367e+00 -1.37293196e+00
-2.83675730e-01 -4.32622373e-01 2.17715099e-01 1.81117475e-01
-7.83063293e-01 -2.77155280e-01 -7.60592937e-01 -7.69510925e-01
6.15065217e-01 -1.15880287e+00 -8.32834899e-01 -6.39394760e-01
-4.97586757e-01 -3.38262290e-01 7.83659697e-01 -3.08408052e-01
1.72537577e+00 -2.26528430e+00 6.40721679e-01 7.08791792e-01
5.05038679e-01 1.21667840e-01 -2.67315209e-01 6.05507433e-01
-2.91791081e-01 1.31376967e-01 -2.61888772e-01 -1.99106634e-01
-1.32991359e-01 3.91129136e-01 -3.74909900e-02 4.72863168e-01
-2.35748589e-01 5.99531054e-01 -9.88934636e-01 -2.29315087e-01
3.01785827e-01 3.78142238e-01 -7.39512444e-01 -1.65525794e-01
4.50387225e-02 2.56105185e-01 -2.90662587e-01 7.22565353e-02
9.07335103e-01 -4.68070179e-01 5.51483333e-01 -4.84077752e-01
-2.85715684e-02 -4.08041060e-01 -1.80910909e+00 1.67265773e+00
-3.45736027e-01 2.03334272e-01 5.11992693e-01 -1.17246342e+00
8.05919468e-01 2.06483707e-01 1.07679081e+00 -1.34908166e-02
1.65750310e-01 -2.38946117e-02 3.45382541e-01 -3.09046954e-01
5.98523259e-01 -2.14652970e-01 2.29289576e-01 7.17213511e-01
-3.12844634e-01 2.02039704e-01 9.16803360e-01 8.88896465e-01
7.72812188e-01 -7.35992730e-01 1.60595775e-01 -5.34914017e-01
8.67369413e-01 -2.69365758e-01 8.74156177e-01 2.26096109e-01
1.92179441e-01 5.12202024e-01 4.01525289e-01 -2.01720163e-01
-6.60096228e-01 -8.87469888e-01 5.17469533e-02 7.69283772e-01
3.40784341e-01 -1.10822606e+00 -6.87154233e-01 -5.61457813e-01
2.87075005e-02 1.96965486e-01 -7.94104218e-01 -2.59661347e-01
-2.00949728e-01 -1.03182840e+00 -1.54986247e-01 4.98577058e-01
5.06017394e-02 -3.55118275e-01 6.66656345e-02 -1.27524078e-01
-5.09282708e-01 -7.68292964e-01 -8.58006299e-01 1.67331278e-01
-1.01331091e+00 -1.40438962e+00 -5.18302023e-01 -7.03810751e-01
1.24080980e+00 1.22318923e+00 8.75985026e-01 3.57789576e-01
1.22163236e-01 5.07774234e-01 -5.47888756e-01 -6.85096439e-03
-2.43762359e-01 6.19987249e-02 3.61552238e-01 3.29011291e-01
2.96700358e-01 -6.29319906e-01 -4.98069972e-01 6.58475161e-01
-1.34433997e+00 1.27300888e-01 5.07279456e-01 9.35637414e-01
6.32190108e-01 4.39160317e-01 4.09842253e-01 -1.08552790e+00
8.74424219e-01 -3.68017972e-01 -1.87927142e-01 4.24988419e-01
-5.92336893e-01 9.10316259e-02 8.51815581e-01 -4.59434211e-01
-7.49227166e-01 3.07829291e-01 1.99451804e-01 -6.97272837e-01
3.79390717e-01 5.80235958e-01 -6.22784078e-01 -5.52172726e-03
5.50271213e-01 1.37460455e-01 9.19605345e-02 -3.36623639e-01
6.54997647e-01 4.29391682e-01 2.03982130e-01 -6.01307094e-01
9.97383118e-01 7.56637514e-01 2.75509894e-01 -7.68224180e-01
-7.27633655e-01 -1.06375229e+00 -9.64123964e-01 -1.46182552e-01
6.73717082e-01 -7.04009712e-01 -9.34021175e-01 -1.82731315e-01
-7.73310244e-01 2.76488423e-01 -1.01981387e-02 6.51276529e-01
-3.23769182e-01 8.45626414e-01 -4.87662107e-01 -5.82043350e-01
-7.75828511e-02 -1.01014566e+00 6.67018533e-01 -1.24299265e-01
-1.74659044e-01 -1.12994182e+00 1.96776420e-01 5.41064262e-01
3.67382802e-02 -4.33668979e-02 9.35119450e-01 -3.31054240e-01
-8.85521844e-02 -4.90266085e-02 -1.60515592e-01 4.02093351e-01
4.80759472e-01 2.76320279e-01 -3.17160040e-01 -6.20542705e-01
8.31885487e-02 3.51738542e-01 9.76007700e-01 5.13530374e-01
8.35809231e-01 -2.29418308e-01 -4.79168206e-01 4.79884744e-01
1.21834350e+00 2.51489431e-01 2.49238938e-01 1.38599038e-01
9.45314050e-01 4.23368275e-01 6.37131929e-01 5.77835202e-01
1.62706882e-01 4.08729255e-01 3.85014445e-01 -1.97606564e-01
4.77804005e-01 -4.19237427e-02 1.92194551e-01 1.19325769e+00
-1.46087006e-01 -9.48040932e-02 -7.38452077e-01 2.98056245e-01
-2.34231114e+00 -8.54826689e-01 -6.42980278e-01 2.24466419e+00
2.58332253e-01 -2.23991007e-01 4.07229751e-01 4.97969210e-01
1.24415159e+00 9.50422138e-02 -4.23415631e-01 -2.24775031e-01
-2.27004960e-01 -2.55908132e-01 2.74928540e-01 5.96799016e-01
-1.04605556e+00 6.67487264e-01 6.58866358e+00 7.99025595e-01
-1.01858532e+00 -7.78464079e-02 1.08187802e-01 -1.18890174e-01
-6.99149847e-01 8.98131430e-02 -3.16037834e-01 4.67324317e-01
4.16056573e-01 -3.06746364e-01 5.51851273e-01 7.10432053e-01
5.93706489e-01 9.43839550e-02 -7.47005939e-01 1.23090720e+00
3.69828716e-02 -9.99691784e-01 3.87331337e-01 2.62845546e-01
1.01541972e+00 -4.88638818e-01 1.18295521e-01 -2.34551847e-01
2.95908511e-01 -7.60323524e-01 2.58575827e-01 2.15020344e-01
3.34471524e-01 -1.11841643e+00 6.07607245e-01 3.90850455e-01
-1.36445451e+00 -9.79615375e-02 -6.99280620e-01 -2.06765473e-01
3.75344902e-01 1.05325091e+00 -5.72535872e-01 7.99488544e-01
4.66335893e-01 1.11547911e+00 -5.54096758e-01 8.88968527e-01
-2.71901816e-01 3.74362677e-01 -2.66557455e-01 2.51063466e-01
7.76999816e-02 -1.01848757e+00 5.05884171e-01 8.61544371e-01
2.43488163e-01 1.56191671e-02 5.40788889e-01 4.03726667e-01
1.27144337e-01 6.09890938e-01 -5.38519204e-01 -1.72329083e-01
3.33444059e-01 1.55400264e+00 -1.31651914e+00 -1.50183693e-01
-6.09293520e-01 1.10566413e+00 1.89906418e-01 3.37591618e-01
-6.78656816e-01 -2.27292385e-02 8.51200342e-01 2.28959903e-01
-8.35447535e-02 -5.42906344e-01 -1.46001026e-01 -1.65163898e+00
-4.30710427e-02 -7.54830599e-01 6.51456237e-01 -2.69984245e-01
-1.40141463e+00 3.25806528e-01 -3.16609859e-01 -1.43020999e+00
-3.05065252e-02 -4.10691172e-01 -5.54390609e-01 5.59477031e-01
-8.96149397e-01 -9.35059607e-01 -1.02337286e-01 1.14910924e+00
6.87360317e-02 4.61572856e-02 6.11718953e-01 6.35555565e-01
-8.94133508e-01 3.80848765e-01 4.07643944e-01 -8.45688507e-02
8.05820704e-01 -1.32274699e+00 -6.98883310e-02 1.05061996e+00
4.09697413e-01 1.40598750e+00 6.68561637e-01 -6.44728839e-01
-1.46304321e+00 -9.45077538e-01 6.01813436e-01 -1.51611477e-01
8.03894639e-01 -3.16229433e-01 -7.62177706e-01 5.65650046e-01
3.52476537e-02 -1.66992828e-01 1.28520930e+00 6.09719574e-01
-3.89529735e-01 3.19248922e-02 -9.20803607e-01 5.43023348e-01
1.08733726e+00 -2.60665029e-01 -2.09142506e-01 4.00312722e-01
5.86390257e-01 1.34111613e-01 -8.94831419e-01 4.26326811e-01
2.94344008e-01 -9.60499287e-01 8.89844656e-01 -8.12457442e-01
9.87210125e-03 -1.02065384e+00 -1.58358887e-01 -1.51905441e+00
-9.72956061e-01 -6.66386306e-01 -1.66536286e-01 1.20794499e+00
2.94456214e-01 -4.67426240e-01 1.18565619e+00 3.88268292e-01
2.16362569e-02 -6.23425484e-01 -5.93714535e-01 -1.01542318e+00
-2.30868682e-01 -4.02162611e-01 4.94286299e-01 1.37236524e+00
3.37187797e-01 7.57660329e-01 -6.01205945e-01 1.20808803e-01
3.88238102e-01 5.19990206e-01 7.86875188e-01 -1.62556660e+00
-2.09894344e-01 -3.71281236e-01 -5.07839262e-01 -9.86159444e-01
3.98484439e-01 -1.18304062e+00 -2.96154529e-01 -1.67779207e+00
4.92541820e-01 -1.40884176e-01 -2.58917689e-01 2.14405969e-01
-3.64575744e-01 2.92911947e-01 4.33917224e-01 4.29950356e-01
-7.54135132e-01 5.16178370e-01 1.37638676e+00 -8.48455429e-02
-3.63415152e-01 1.77333862e-01 -9.54729438e-01 6.33843124e-01
6.11384153e-01 -2.26943105e-01 -8.71629655e-01 -3.35749000e-01
6.04255080e-01 -9.33962539e-02 -2.69725055e-01 -6.66267991e-01
2.68878222e-01 -4.30193901e-01 4.69066799e-02 -5.16312540e-01
8.32690969e-02 -1.13699710e+00 5.95252454e-01 4.59207147e-01
9.11770854e-03 4.68971908e-01 -2.16815799e-01 7.08600163e-01
-5.35098374e-01 -2.78208226e-01 7.32313931e-01 -1.16940804e-01
-5.17029703e-01 4.72118527e-01 -4.84931588e-01 -2.29689255e-01
1.06122196e+00 -1.59946427e-01 1.24724023e-01 -3.68959993e-01
-1.08339190e+00 4.90284055e-01 7.20310748e-01 4.13986802e-01
5.50369442e-01 -1.67011786e+00 -4.29453343e-01 -2.16244534e-03
4.00108732e-02 -6.52533844e-02 5.28412759e-01 9.35670555e-01
-1.33737966e-01 4.92992669e-01 2.36263592e-02 -5.08249938e-01
-1.45432103e+00 1.09769773e+00 -1.66101813e-01 -3.23133767e-01
-2.62748212e-01 5.56454599e-01 4.19041038e-01 -5.37352979e-01
-8.81133005e-02 -6.89811185e-02 -2.07341790e-01 4.01556969e-01
5.28235853e-01 6.61617517e-01 1.85428690e-02 -8.75716150e-01
-3.50162297e-01 6.74577355e-01 7.71184936e-02 6.33088648e-02
1.17109370e+00 -3.49310011e-01 -6.89590573e-01 3.86604309e-01
1.14260566e+00 2.96191841e-01 -7.05423474e-01 -1.27875030e-01
2.60721296e-02 -5.97150505e-01 -1.80123195e-01 -1.51647672e-01
-1.56205320e+00 7.42888927e-01 9.33129787e-02 4.27439690e-01
1.30608678e+00 -1.00252777e-01 8.21788251e-01 3.74422103e-01
3.70653510e-01 -9.96523619e-01 1.16167732e-01 4.90779310e-01
5.28789937e-01 -1.19278097e+00 5.54834306e-02 -9.37733948e-01
-5.29064953e-01 1.14143848e+00 6.21979296e-01 -3.66032682e-02
8.56645703e-01 6.85853139e-02 -5.03613278e-02 -1.26484200e-01
-4.32773024e-01 -2.77137458e-01 3.50550443e-01 5.04554570e-01
4.97177273e-01 3.09518725e-01 -8.80230904e-01 7.28972971e-01
-9.31509882e-02 -4.23031837e-01 3.98711026e-01 4.60937440e-01
-3.81863087e-01 -1.52608728e+00 -5.41720152e-01 6.11585319e-01
-2.41957009e-01 -2.47392878e-02 -6.74596846e-01 3.69660825e-01
1.56559035e-01 1.18075180e+00 -1.66798249e-01 -6.32035553e-01
2.07132310e-01 -2.87119508e-01 2.78435171e-01 -8.12663376e-01
-4.95009571e-01 2.83168226e-01 -5.10012805e-01 -4.74684447e-01
-7.28967667e-01 -8.38238418e-01 -1.41594112e+00 -4.25855726e-01
-5.66942751e-01 9.60581958e-01 3.16309422e-01 6.55917048e-01
3.72597307e-01 1.80487543e-01 7.92692482e-01 -4.04129624e-01
7.32777044e-02 -6.25866473e-01 -1.02267540e+00 7.22222567e-01
-3.53758447e-02 -7.00689435e-01 -4.82136339e-01 -1.37170792e-01] | [7.429959297180176, 4.836345672607422] |
3f7c4b99-f325-4f36-bc4b-aa1c4ec17666 | hierarchical-memory-networks | 1605.07427 | null | http://arxiv.org/abs/1605.07427v1 | http://arxiv.org/pdf/1605.07427v1.pdf | Hierarchical Memory Networks | Memory networks are neural networks with an explicit memory component that
can be both read and written to by the network. The memory is often addressed
in a soft way using a softmax function, making end-to-end training with
backpropagation possible. However, this is not computationally scalable for
applications which require the network to read from extremely large memories.
On the other hand, it is well known that hard attention mechanisms based on
reinforcement learning are challenging to train successfully. In this paper, we
explore a form of hierarchical memory network, which can be considered as a
hybrid between hard and soft attention memory networks. The memory is organized
in a hierarchical structure such that reading from it is done with less
computation than soft attention over a flat memory, while also being easier to
train than hard attention over a flat memory. Specifically, we propose to
incorporate Maximum Inner Product Search (MIPS) in the training and inference
procedures for our hierarchical memory network. We explore the use of various
state-of-the art approximate MIPS techniques and report results on
SimpleQuestions, a challenging large scale factoid question answering task. | ['Hugo Larochelle', 'Pascal Vincent', 'Yoshua Bengio', 'Gerald Tesauro', 'Sungjin Ahn', 'Sarath Chandar'] | 2016-05-24 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 9.82890949e-02 4.10364240e-01 -1.25480786e-01 -6.63961470e-01
-5.96095324e-01 -3.68597627e-01 3.90300274e-01 2.46078968e-01
-7.24065125e-01 9.06770468e-01 -6.77904934e-02 -4.09062624e-01
-7.86181763e-02 -1.36063170e+00 -1.48179913e+00 -4.24440503e-01
3.36889178e-01 9.65508103e-01 3.01983684e-01 1.04714446e-02
2.66730666e-01 3.53585333e-01 -1.44630039e+00 5.62833369e-01
8.86240482e-01 1.54399371e+00 4.79227990e-01 5.74188709e-01
-3.72817248e-01 1.36154175e+00 -6.77277863e-01 -5.27120769e-01
-4.77025248e-02 -1.15821168e-01 -1.42811048e+00 -5.38000703e-01
4.77682739e-01 -5.75318336e-01 -4.01976228e-01 9.48466718e-01
2.75205404e-01 5.68889201e-01 3.77617329e-01 -6.45752668e-01
-9.68272746e-01 7.70385325e-01 -3.49844322e-02 2.22698301e-01
2.22710624e-01 6.49844334e-02 1.17589188e+00 -9.56532478e-01
1.83858022e-01 1.36789083e+00 2.74768472e-01 3.31702024e-01
-9.99612629e-01 -4.58565116e-01 1.32697806e-01 7.03080714e-01
-1.19219685e+00 -3.19038182e-01 3.30264986e-01 1.08659588e-01
1.73030460e+00 2.57703274e-01 4.22895491e-01 6.39582574e-01
1.03930257e-01 1.02532840e+00 7.95947254e-01 -3.65793049e-01
4.89253908e-01 8.36566091e-02 3.94602031e-01 9.00095880e-01
-1.90077022e-01 6.46449700e-02 -5.49852729e-01 -6.20396510e-02
5.35616696e-01 1.17847718e-01 -3.30297083e-01 7.27086142e-02
-7.42251277e-01 1.02648079e+00 1.13973570e+00 1.28772840e-01
-3.60783130e-01 3.80295962e-01 2.37263784e-01 5.04281700e-01
2.10779399e-01 5.52655518e-01 -4.60531533e-01 1.76052842e-02
-1.14079535e+00 5.76640479e-03 1.12561476e+00 8.04684877e-01
8.43848109e-01 -2.46326979e-02 -1.38915405e-01 8.39228034e-01
1.89129919e-01 2.81797945e-01 5.89124322e-01 -1.03042471e+00
5.95002055e-01 5.05437195e-01 -5.85981347e-02 -7.36957610e-01
-4.82175618e-01 -2.43717059e-01 -1.02280927e+00 4.79234569e-02
3.22775185e-01 1.31778017e-01 -1.03180480e+00 1.78562272e+00
-4.75381930e-05 2.01211333e-01 -1.19747989e-01 9.16195035e-01
8.94961953e-01 1.13689363e+00 -6.21271506e-02 -1.19600348e-01
1.31287944e+00 -1.49159932e+00 -6.81562066e-01 -5.28286040e-01
5.86805165e-01 -3.89804900e-01 1.05899620e+00 4.13226634e-01
-1.86952233e+00 -7.38827288e-01 -1.14338326e+00 -6.15979493e-01
-7.90814757e-01 -3.50560099e-01 5.14062464e-01 1.35548949e-01
-1.13568640e+00 8.70282948e-01 -8.97487342e-01 2.60726474e-02
3.04632157e-01 6.83716297e-01 -1.38460308e-01 -1.21010110e-01
-1.59796119e+00 1.36217260e+00 8.94915462e-01 3.61800104e-01
-6.15805566e-01 -2.83971578e-01 -9.67001438e-01 5.67364037e-01
5.66223085e-01 -9.55930054e-01 1.44686532e+00 -9.20978248e-01
-1.77429295e+00 6.13652706e-01 -1.69098377e-01 -9.20029581e-01
-4.88384999e-02 -3.86173010e-01 3.09901848e-03 2.73248971e-01
-3.78536046e-01 9.09000695e-01 8.24576318e-01 -6.21407449e-01
-2.67734528e-01 -2.13013276e-01 4.55171198e-01 3.71276401e-02
-4.22386110e-01 -2.88437724e-01 -3.44756156e-01 -5.73559284e-01
-1.64739266e-01 -7.85745323e-01 1.10305168e-01 4.49394286e-02
-2.90794998e-01 -4.78133202e-01 5.48193038e-01 -6.76440537e-01
1.43977976e+00 -1.88869607e+00 4.94405448e-01 2.94714361e-01
1.63860468e-03 4.50509548e-01 -1.54526696e-01 5.14605418e-02
2.05740049e-01 -4.28306088e-02 -4.47307765e-01 -2.81838685e-01
3.37005883e-01 7.83559322e-01 -5.75385571e-01 -8.88572559e-02
2.60318846e-01 1.25899792e+00 -7.17890024e-01 -3.64103585e-01
-1.09850414e-01 1.93317518e-01 -8.23512495e-01 6.76573336e-01
-7.03925908e-01 -3.37132096e-01 -4.64642383e-02 5.72899580e-01
5.36754847e-01 -6.29325330e-01 1.38374129e-02 -1.73942119e-01
2.21810386e-01 7.77453601e-01 -1.08365214e+00 1.53716195e+00
-6.78462744e-01 3.12983543e-01 1.26533145e-02 -1.31145227e+00
6.42798960e-01 1.22491457e-01 -3.98194492e-01 -1.08822107e+00
7.75671601e-02 2.35774279e-01 -4.65116231e-03 -3.16840619e-01
6.80227041e-01 -5.10747842e-02 7.20337257e-02 5.74415803e-01
3.44706833e-01 -3.04428823e-02 2.42823347e-01 -3.16412002e-02
8.78669202e-01 -5.86308055e-02 9.59383845e-02 2.43390407e-02
5.16962469e-01 -3.48488688e-01 2.96937078e-01 9.96146798e-01
2.81440377e-01 4.19943869e-01 2.32282862e-01 -4.85882193e-01
-6.82729125e-01 -1.19890749e+00 -5.60950581e-03 1.47503412e+00
6.65038219e-03 -3.07753146e-01 -8.46146345e-01 -4.61064547e-01
-7.68836513e-02 7.37328470e-01 -5.24877131e-01 -4.80452865e-01
-8.91617775e-01 -5.96987486e-01 6.10235751e-01 1.06716168e+00
8.82913947e-01 -1.55188179e+00 -7.80240417e-01 4.18261081e-01
-1.18989721e-01 -9.07395184e-01 -5.31434774e-01 7.80793726e-01
-9.13112044e-01 -8.44922841e-01 -6.67297065e-01 -1.00392544e+00
5.08305669e-01 -2.68297940e-01 1.51294661e+00 3.89200360e-01
3.83598022e-02 -8.29161704e-03 -6.85785860e-02 -1.15080208e-01
-8.32531303e-02 5.10692358e-01 -1.76188901e-01 -1.38842300e-01
2.35631824e-01 -4.44164306e-01 -5.73975801e-01 1.71683595e-01
-1.17676115e+00 -8.17690119e-02 8.09002459e-01 1.20265090e+00
8.42199981e-01 -6.20406168e-03 5.32575130e-01 -8.64063859e-01
5.96682847e-01 -5.84427357e-01 -6.55940175e-01 5.08304536e-01
-3.81607711e-01 5.77257872e-01 8.20327640e-01 -3.85719925e-01
-8.28561306e-01 -9.99365970e-02 -5.34255803e-01 -3.45040858e-01
1.16726391e-01 8.24792147e-01 -1.48347365e-02 -1.06298774e-01
3.41381758e-01 3.37489277e-01 -1.81957290e-01 -3.98246139e-01
3.51781756e-01 4.53477889e-01 5.88680983e-01 -6.88920319e-01
4.02863175e-01 -1.75556634e-02 -2.03725621e-01 -5.33423603e-01
-1.20886087e+00 7.21009374e-02 -3.32956344e-01 4.16551679e-01
9.26403463e-01 -4.07145828e-01 -1.01195931e+00 4.12405342e-01
-1.09769368e+00 -8.68195474e-01 -1.71541542e-01 2.14399070e-01
-5.97586989e-01 6.28338903e-02 -1.16447604e+00 -4.97029275e-01
-6.76871061e-01 -1.07842374e+00 7.38365352e-01 2.75584191e-01
-3.80308568e-01 -1.11511588e+00 -2.61017442e-01 3.66266847e-01
8.82123232e-01 -3.01064253e-01 1.22779107e+00 -6.06682301e-01
-9.30270612e-01 6.69620633e-02 -4.93055761e-01 4.19861197e-01
-5.21402955e-01 -4.10737693e-01 -1.00364649e+00 -2.03245699e-01
1.53142363e-01 -9.08124626e-01 1.27820420e+00 2.98513710e-01
1.70072103e+00 -7.27256060e-01 3.57504636e-02 7.17524886e-01
1.26052427e+00 3.04145627e-02 8.08051109e-01 4.50032830e-01
3.35747182e-01 2.69395322e-01 3.26180339e-01 8.54944661e-02
6.84897840e-01 4.82956827e-01 6.51795268e-01 1.24193214e-01
2.40623668e-01 -1.68060392e-01 2.58497655e-01 8.25281978e-01
1.61607698e-01 -2.27683365e-01 -7.67496109e-01 2.76219487e-01
-1.84490192e+00 -8.99419606e-01 3.55767548e-01 2.13507104e+00
1.34688485e+00 5.79032123e-01 -2.33096048e-01 -3.03205978e-02
2.80839503e-01 1.60306975e-01 -6.75511420e-01 -9.31438386e-01
-4.57470082e-02 7.98106611e-01 2.34654516e-01 8.04837108e-01
-9.77342963e-01 7.91812658e-01 5.76949787e+00 7.65545249e-01
-1.10578823e+00 1.06176861e-01 7.66761243e-01 -3.93908054e-01
-9.33408272e-03 -2.47985289e-01 -8.32236648e-01 5.06412327e-01
1.41775405e+00 3.34061027e-01 7.98142195e-01 7.11792171e-01
-5.26576400e-01 -3.72400016e-01 -1.46401596e+00 9.99355435e-01
-1.32488012e-01 -1.44533885e+00 1.45890832e-01 -2.62881070e-01
3.82039636e-01 3.84369753e-02 1.48573518e-01 8.16971838e-01
2.34759022e-02 -1.40769315e+00 7.23976672e-01 7.86622584e-01
4.11815196e-01 -8.76786113e-01 8.85013282e-01 7.37758219e-01
-1.07670140e+00 -3.31879556e-01 -6.81874275e-01 -2.54189521e-01
1.96479723e-01 4.12788212e-01 -2.44096562e-01 7.54537731e-02
8.28849196e-01 1.71524152e-01 -4.78732705e-01 9.81449842e-01
-2.71793336e-01 4.82689440e-01 -5.26668251e-01 -2.08711371e-01
4.48816121e-01 9.24641490e-02 -1.84352726e-01 1.18524647e+00
2.91790307e-01 3.08263242e-01 2.25844979e-01 1.02526259e+00
-6.15448296e-01 -3.10220510e-01 -2.06307665e-01 -4.34890054e-02
4.79363859e-01 8.91631246e-01 -3.49454522e-01 -6.01516187e-01
-3.96706909e-01 1.07941878e+00 1.05747414e+00 1.76831260e-01
-9.66460228e-01 -7.10016191e-01 1.69102728e-01 9.57524776e-02
5.43367445e-01 -5.88025600e-02 -2.88094103e-01 -9.99139845e-01
1.51654705e-01 -5.77414930e-01 6.06928229e-01 -9.90141034e-01
-1.17004132e+00 7.00505316e-01 -2.59693116e-01 -4.89505023e-01
-4.49776143e-01 -7.19199657e-01 -6.23680174e-01 1.02132750e+00
-1.90065241e+00 -7.01609969e-01 -1.35103807e-01 6.56313241e-01
3.45926493e-01 2.52349108e-01 1.08582222e+00 3.85976225e-01
-4.67898816e-01 1.01997387e+00 -4.71373722e-02 8.21916461e-02
4.66243684e-01 -1.32463956e+00 1.78056672e-01 2.60920048e-01
-3.70821767e-02 8.74007642e-01 2.76775181e-01 -2.01424167e-01
-1.51302278e+00 -1.03984571e+00 1.07066607e+00 -1.20130554e-01
7.03282535e-01 -2.74468601e-01 -1.55166364e+00 7.83938825e-01
5.85069954e-01 9.87014398e-02 5.59548616e-01 5.19630089e-02
-4.03575242e-01 -9.58398879e-02 -1.06247377e+00 2.06372783e-01
6.26122832e-01 -7.86673903e-01 -1.08698285e+00 3.31415057e-01
6.33288383e-01 -6.89260304e-01 -9.56027210e-01 2.49195725e-01
3.15546334e-01 -7.45693743e-01 1.00738168e+00 -4.52662319e-01
5.37244976e-01 -1.22769624e-01 -1.19986936e-01 -1.24270689e+00
-2.86820740e-01 -1.99597329e-01 -1.13695264e+00 8.62260699e-01
5.49996316e-01 -7.93011010e-01 7.44336784e-01 8.35157812e-01
-2.04423711e-01 -1.23291993e+00 -8.89635801e-01 -6.73277080e-01
3.64307016e-01 -1.92931015e-02 3.98561627e-01 5.17889678e-01
7.02045858e-02 5.62078178e-01 -2.37529710e-01 1.30411386e-01
2.79629081e-01 5.25036871e-01 1.80204645e-01 -9.92113411e-01
-7.43925989e-01 -5.64275146e-01 1.06307685e-01 -1.75841129e+00
3.53716552e-01 -9.85443115e-01 1.68892011e-01 -1.59565544e+00
1.15046762e-01 -1.59346595e-01 -3.08006316e-01 6.39832735e-01
-1.18026733e-01 2.17451528e-01 1.45073071e-01 -7.35775977e-02
-8.08592141e-01 5.44669390e-01 1.03498912e+00 -3.85552168e-01
6.03583641e-02 -3.36378813e-02 -4.92129564e-01 7.24101067e-01
8.26846063e-01 -3.19201022e-01 -2.94652283e-01 -9.04631793e-01
5.38757265e-01 3.58040124e-01 3.66843343e-01 -9.73079741e-01
7.95537591e-01 8.24323446e-02 5.16768396e-01 -7.26916492e-01
6.72340751e-01 -7.22871661e-01 -3.18073958e-01 4.29814875e-01
-5.11139333e-01 2.00620875e-01 3.01275134e-01 2.28709206e-01
-5.79219162e-01 -6.29279912e-01 8.93202782e-01 -4.96284276e-01
-5.64615428e-01 2.99325198e-01 -1.75227642e-01 2.29632854e-01
6.16588175e-01 5.20925298e-02 -3.72023016e-01 -3.33355248e-01
-9.67333019e-01 5.26948452e-01 -4.61502150e-02 7.49540180e-02
7.83908784e-01 -1.27551818e+00 -2.82483190e-01 2.37759694e-01
-3.62648755e-01 3.35577518e-01 6.94723353e-02 6.52333736e-01
-5.04694700e-01 6.66297853e-01 -7.56084695e-02 -1.90727785e-01
-7.78503001e-01 8.93537104e-01 4.98928994e-01 -6.19794965e-01
-2.38324478e-01 1.00397062e+00 7.34608993e-02 -5.75243235e-01
6.69946015e-01 -5.83890200e-01 6.98874239e-03 8.79533067e-02
8.94686162e-01 3.28044295e-01 2.04013020e-01 -1.36185348e-01
-1.74053714e-01 3.81684154e-01 -2.69278765e-01 2.40390971e-01
1.32535088e+00 1.28467754e-01 -4.65210855e-01 2.78102636e-01
1.31114626e+00 -6.56281114e-01 -9.25235808e-01 -4.75321561e-01
1.35520846e-01 -4.78202626e-02 5.67408800e-02 -8.90806198e-01
-1.19820452e+00 1.32064652e+00 3.23619455e-01 4.60388452e-01
1.22353172e+00 -2.09547505e-01 1.34212840e+00 1.15300953e+00
1.19125411e-01 -1.45322716e+00 1.81248993e-01 1.13338041e+00
9.80465233e-01 -1.23511064e+00 -3.25144112e-01 1.63373455e-01
-9.71835256e-02 1.40043068e+00 9.17313039e-01 -2.90242821e-01
7.63168931e-01 3.73245478e-01 -6.25630379e-01 4.25964594e-02
-9.77633774e-01 -7.23771974e-02 3.02166820e-01 -7.87469298e-02
2.83337802e-01 -1.47636965e-01 9.22442228e-02 7.83376873e-01
-1.96479931e-01 1.59857541e-01 1.08260557e-01 1.08401144e+00
-8.93836200e-01 -8.63601685e-01 -2.48401612e-01 6.93909287e-01
-5.03068984e-01 -5.15439570e-01 -1.55503765e-01 3.32788140e-01
-1.37848780e-01 6.88256025e-01 2.70324767e-01 -8.48111361e-02
3.15783828e-01 4.31817353e-01 5.02198517e-01 -5.62139928e-01
-7.81315148e-01 -5.56556344e-01 -3.14151905e-02 -6.15465641e-01
-3.30358371e-02 1.42856225e-01 -1.55369461e+00 -5.59327245e-01
-3.94317567e-01 6.07444346e-02 2.34987602e-01 1.19338524e+00
2.57228255e-01 6.32425964e-01 6.55847341e-02 -9.33139443e-01
-8.00612211e-01 -8.91848683e-01 -2.71247327e-01 9.46794823e-02
3.87607723e-01 -5.43376029e-01 -1.12400629e-01 -2.79235303e-01] | [11.001983642578125, 7.686696529388428] |
f0ad48c4-cf14-487c-aa68-b8e29f2d02e6 | encoding-structure-texture-relation-with-p | 2008.03632 | null | https://arxiv.org/abs/2008.03632v1 | https://arxiv.org/pdf/2008.03632v1.pdf | Encoding Structure-Texture Relation with P-Net for Anomaly Detection in Retinal Images | Anomaly detection in retinal image refers to the identification of abnormality caused by various retinal diseases/lesions, by only leveraging normal images in training phase. Normal images from healthy subjects often have regular structures (e.g., the structured blood vessels in the fundus image, or structured anatomy in optical coherence tomography image). On the contrary, the diseases and lesions often destroy these structures. Motivated by this, we propose to leverage the relation between the image texture and structure to design a deep neural network for anomaly detection. Specifically, we first extract the structure of the retinal images, then we combine both the structure features and the last layer features extracted from original health image to reconstruct the original input healthy image. The image feature provides the texture information and guarantees the uniqueness of the image recovered from the structure. In the end, we further utilize the reconstructed image to extract the structure and measure the difference between structure extracted from original and the reconstructed image. On the one hand, minimizing the reconstruction difference behaves like a regularizer to guarantee that the image is corrected reconstructed. On the other hand, such structure difference can also be used as a metric for normality measurement. The whole network is termed as P-Net because it has a ``P'' shape. Extensive experiments on RESC dataset and iSee dataset validate the effectiveness of our approach for anomaly detection in retinal images. Further, our method also generalizes well to novel class discovery in retinal images and anomaly detection in real-world images. | ['Zaiwang Gu', 'Jianlong Yang', 'Wen Liu', 'Jiang Liu', 'Yuting Xiao', 'Shenghua Gao', 'Kang Zhou', 'Weixin Luo', 'Jun Cheng'] | 2020-08-09 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3484_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123650358.pdf | eccv-2020-8 | ['novel-class-discovery', 'novel-class-discovery'] | ['computer-vision', 'methodology'] | [ 1.97999641e-01 8.91268998e-02 1.53945595e-01 -3.15404326e-01
5.40326722e-02 -1.96459800e-01 2.02221163e-02 -1.24213621e-01
-1.43464223e-01 5.24187207e-01 1.51150987e-01 -1.49593025e-01
-1.08972616e-01 -6.92955315e-01 -6.86302006e-01 -1.11530697e+00
1.12638347e-01 -3.60471815e-01 1.25249147e-01 1.26593351e-01
3.17872494e-01 6.02632523e-01 -1.47435796e+00 1.01874456e-01
1.16645610e+00 1.49829841e+00 1.63111500e-02 2.30413899e-01
-3.18161637e-01 7.99858928e-01 -1.57441258e-01 1.52858847e-03
5.69775343e-01 -5.98967195e-01 -6.65587604e-01 5.67918599e-01
6.23636305e-01 -6.28642917e-01 -4.10634488e-01 1.66987622e+00
5.26741743e-01 -1.27662525e-01 5.29672146e-01 -7.15694487e-01
-5.41684449e-01 -1.83884889e-01 -8.24969828e-01 5.42581975e-01
-1.85665544e-02 3.78679961e-01 6.49282813e-01 -7.55690217e-01
3.48545164e-01 8.79401743e-01 4.14564997e-01 3.90077174e-01
-1.01635730e+00 -3.66301537e-01 8.34691301e-02 1.91017538e-01
-9.72999036e-01 -6.35092735e-01 6.88771367e-01 -5.43551445e-01
3.61843497e-01 2.45900080e-01 8.96906495e-01 6.09270215e-01
3.30780894e-01 7.68034458e-01 1.13905251e+00 -1.89903572e-01
-2.50133276e-01 -3.03430766e-01 -5.22056744e-02 9.84351397e-01
5.82291722e-01 2.25165904e-01 1.61439791e-01 7.79366270e-02
9.79071975e-01 5.58248535e-02 -7.45125949e-01 -1.43458605e-01
-1.13454854e+00 2.41822392e-01 4.16245431e-01 1.54474769e-02
-4.64790136e-01 -2.53377318e-01 3.94403517e-01 4.87085044e-01
1.21517323e-01 3.41168165e-01 -7.81909302e-02 2.43659556e-01
-3.87746900e-01 -3.71420607e-02 3.14782262e-01 3.65832835e-01
6.06160879e-01 1.99391007e-01 -3.36562186e-01 8.09108555e-01
4.07391191e-01 1.80349633e-01 7.60049880e-01 -9.11256731e-01
1.23132989e-01 1.03014767e+00 1.47810385e-01 -1.19343317e+00
-3.27815652e-01 -6.20681942e-01 -1.30035424e+00 4.17085201e-01
6.92445278e-01 2.40426958e-02 -1.04086637e+00 1.39966726e+00
4.31962639e-01 4.82886642e-01 -8.38351920e-02 1.29292762e+00
8.12118948e-01 9.04109553e-02 -4.56495434e-01 -4.07963395e-01
1.17313147e+00 -6.76644921e-01 -6.58633649e-01 -8.86605978e-02
5.22541285e-01 -6.81106210e-01 9.98997033e-01 3.35578710e-01
-9.98372436e-01 -4.30832177e-01 -9.91702914e-01 4.56478521e-02
2.24115759e-01 5.58394969e-01 4.50762630e-01 2.84811974e-01
-8.06875288e-01 3.81176710e-01 -8.66557360e-01 -1.90096930e-01
7.98371971e-01 2.26039007e-01 -4.32096690e-01 -1.11756504e-01
-6.09332263e-01 4.90028292e-01 3.13474745e-01 4.24849331e-01
-4.91507977e-01 -5.04211605e-01 -6.43167138e-01 -1.55880392e-01
2.76031196e-01 -8.32209527e-01 7.40148425e-01 -1.29719365e+00
-1.31159270e+00 9.80577290e-01 -2.05142096e-01 -3.76560122e-01
4.77654338e-01 2.65495423e-02 -6.51278853e-01 3.83508056e-01
7.40669370e-02 2.72243563e-02 9.15194035e-01 -9.29768622e-01
-7.74542451e-01 -4.29810435e-01 -3.88634726e-02 -1.30046487e-01
-1.42131954e-01 -1.46316037e-01 -1.86804131e-01 -6.40222788e-01
6.69100463e-01 -4.84305471e-01 -1.38823465e-01 4.49890643e-01
-8.65914822e-01 1.39529901e-02 6.23036385e-01 -7.66355276e-01
1.05922925e+00 -2.36352897e+00 -1.84702352e-01 5.30117512e-01
7.06197083e-01 3.73444885e-01 -2.59527862e-01 -3.85692328e-01
-4.12749171e-01 -4.80761789e-02 -2.61864007e-01 2.24403799e-01
-5.18479943e-01 1.57939345e-01 -2.09729657e-01 8.54268432e-01
4.34120387e-01 8.57615173e-01 -6.68295562e-01 -5.39413393e-01
-3.87521018e-03 3.33398096e-02 -4.12784874e-01 1.86158717e-01
-3.11408052e-03 8.68178725e-01 -6.39608622e-01 7.76674986e-01
7.59660065e-01 -4.60720420e-01 -1.17284000e-01 -3.74397159e-01
-4.84476201e-02 -7.55822882e-02 -1.06860280e+00 1.21133685e+00
1.31947726e-01 5.62786579e-01 1.23311849e-02 -1.35391390e+00
8.58443916e-01 1.58113331e-01 6.39164150e-01 -9.26453650e-01
2.75488794e-01 4.45881188e-01 6.16502583e-01 -1.05356836e+00
-2.22455561e-01 1.53251275e-01 8.24042857e-01 3.58455032e-01
-3.46325487e-01 6.78617120e-01 1.83944535e-02 -2.86135435e-01
7.95866132e-01 -1.15463518e-01 3.48054677e-01 3.01231109e-02
8.24195445e-01 -4.71479565e-01 8.75452638e-01 5.61470211e-01
-3.93040776e-01 6.54242158e-01 8.35626781e-01 -7.82594085e-01
-1.05876470e+00 -9.65621233e-01 -4.89304155e-01 1.70831352e-01
3.89763564e-01 9.57424566e-02 -6.13481343e-01 -8.80288303e-01
-5.24056517e-02 -1.70806915e-01 -5.97875535e-01 -3.87801200e-01
-4.82615501e-01 -8.23186159e-01 4.21371043e-01 3.65330577e-01
9.72858965e-01 -7.08711028e-01 -4.15979862e-01 -5.78496419e-03
-1.16716310e-01 -9.74634409e-01 -4.84074682e-01 -6.74483538e-01
-9.68608677e-01 -1.53525972e+00 -5.82047462e-01 -8.32502782e-01
1.11291754e+00 2.68446624e-01 7.86671817e-01 5.83305240e-01
-6.38039172e-01 -3.58139537e-02 -1.07739791e-01 -1.11782923e-01
-1.14955902e-01 -7.70117283e-01 1.14603922e-01 8.72495174e-01
2.28755668e-01 -8.29536915e-01 -1.16557932e+00 2.95061201e-01
-8.27891707e-01 -2.59721667e-01 1.09770668e+00 1.07758355e+00
1.13524175e+00 4.95700836e-01 3.16654503e-01 -6.74308836e-01
3.96838605e-01 -2.49305576e-01 -6.78513885e-01 1.99802801e-01
-5.43283641e-01 -9.24225077e-02 4.90615934e-01 -3.11350167e-01
-7.81047821e-01 -7.16539621e-02 -4.60057659e-03 -5.59945107e-01
-3.05460930e-01 6.00988984e-01 -4.13457006e-01 -3.72302830e-01
5.14831603e-01 5.97715855e-01 5.28259099e-01 -5.45720875e-01
-1.86729953e-01 4.86714900e-01 7.47109056e-01 -3.09936166e-01
6.23961210e-01 8.35745335e-01 3.04675043e-01 -7.86111355e-01
-1.04048550e+00 -4.31671441e-01 -3.31967086e-01 -7.16049075e-02
6.78759336e-01 -5.69405437e-01 -9.27186906e-01 8.39606881e-01
-9.39714611e-01 6.29247800e-02 -3.09774905e-01 7.15960443e-01
-3.13973993e-01 9.19580638e-01 -3.68212909e-01 -6.23823225e-01
-2.35266894e-01 -9.85182285e-01 7.28669345e-01 4.51052904e-01
3.98197770e-01 -8.10735822e-01 -2.04601169e-01 -5.91400685e-03
1.30210489e-01 4.50387299e-01 1.42691922e+00 -4.32882398e-01
-7.96658397e-01 -2.37541661e-01 -7.93889403e-01 8.61875653e-01
3.25957090e-01 1.42595157e-01 -8.45322788e-01 -2.41098866e-01
3.25227827e-01 3.31082158e-02 1.01461482e+00 8.12269568e-01
1.56710017e+00 -5.08917451e-01 -3.61028314e-02 1.01905596e+00
1.30375791e+00 7.69125000e-02 1.04238367e+00 4.00020570e-01
7.52018273e-01 7.06656992e-01 2.84593910e-01 3.26061517e-01
-4.32512872e-02 3.73844743e-01 5.68305135e-01 -4.09139603e-01
-3.06515008e-01 -2.01235246e-02 1.45257622e-01 5.86073518e-01
-4.18869317e-01 2.28956893e-01 -6.80634737e-01 3.68768483e-01
-1.63581800e+00 -7.13500798e-01 -3.03192347e-01 2.41361308e+00
7.65822053e-01 -7.22724050e-02 -4.61735912e-02 -1.33666158e-01
7.79020250e-01 -2.30890378e-01 -9.56586659e-01 2.38416493e-01
-4.68177170e-01 8.53344649e-02 2.44829535e-01 -2.33336687e-02
-9.74724412e-01 4.22729284e-01 5.55857658e+00 6.37309611e-01
-1.43049932e+00 -3.60653937e-01 7.54849076e-01 -8.19906220e-02
-6.52551418e-03 -5.87320887e-03 -3.04704279e-01 7.27320552e-01
1.22984394e-01 -1.54854715e-01 3.03200483e-01 3.68329912e-01
4.05311197e-01 3.46580856e-02 -8.90905023e-01 1.06568205e+00
-1.92253500e-01 -1.05490077e+00 2.65374929e-01 3.55744183e-01
5.21382630e-01 -1.49443597e-01 2.73656905e-01 -3.31663132e-01
-3.30678165e-01 -1.16843784e+00 8.55966657e-02 1.12509525e+00
9.35812891e-01 -5.91282308e-01 9.95416164e-01 7.59984851e-02
-8.57056379e-01 -1.13880225e-01 -6.83930755e-01 6.25341162e-02
-3.22102249e-01 8.15765083e-01 -4.81245816e-01 5.70169032e-01
7.70735741e-01 1.25149059e+00 -7.48125732e-01 1.65366614e+00
-1.02552325e-01 4.86877650e-01 -7.03667328e-02 7.17077494e-01
1.95821486e-02 -7.69191027e-01 9.94105577e-01 5.29226959e-01
4.78535533e-01 2.34694779e-02 1.64855957e-01 1.09054661e+00
1.18822932e-01 2.18512610e-01 -6.12691402e-01 -1.05671979e-01
1.54091120e-01 1.15276790e+00 -4.40458715e-01 7.31309056e-02
-6.17741466e-01 7.27811873e-01 7.54704420e-03 7.01409519e-01
-2.22748041e-01 -4.25187886e-01 7.64951885e-01 3.03503394e-01
6.50450736e-02 2.06166744e-01 -1.94247141e-01 -1.32951880e+00
5.64110637e-01 -8.81736755e-01 1.61650568e-01 -6.23899400e-01
-1.69385135e+00 5.94124258e-01 -6.79310441e-01 -1.71402454e+00
1.95038497e-01 -7.85878181e-01 -9.35226858e-01 9.68601644e-01
-1.83865619e+00 -8.56162071e-01 -7.08465695e-01 7.90302455e-01
-1.48734108e-01 -5.78708529e-01 4.03544247e-01 1.75833553e-01
-9.98136163e-01 6.81957066e-01 9.15505588e-02 5.59542060e-01
5.88150799e-01 -1.26628447e+00 3.29313353e-02 1.14539278e+00
-3.57344329e-01 7.42454410e-01 2.04557106e-01 -6.04799509e-01
-1.00642681e+00 -1.20575070e+00 1.76448226e-01 -5.51817678e-02
6.84358418e-01 3.64383787e-01 -1.20174861e+00 4.75805223e-01
-1.23987444e-01 6.23652339e-01 4.30720419e-01 -2.70047605e-01
-2.58844852e-01 -2.48658672e-01 -9.14076507e-01 5.62492311e-01
9.43563223e-01 -3.32320690e-01 -3.97024810e-01 3.79050255e-01
3.97098601e-01 -5.82043886e-01 -7.92276204e-01 5.80354929e-01
5.51173449e-01 -1.24912059e+00 8.82698238e-01 -1.05192220e+00
5.77522874e-01 -4.53151792e-01 2.68000402e-02 -1.02676713e+00
-1.96958542e-01 -5.77269316e-01 -1.50485501e-01 8.55556309e-01
1.15372837e-01 -1.12133205e+00 6.90437078e-01 3.09275746e-01
-4.25213367e-01 -1.07387233e+00 -8.39011192e-01 -6.56351626e-01
-1.32059351e-01 4.99298573e-02 5.38784504e-01 8.47374618e-01
-7.33651042e-01 -1.09747097e-01 -1.88223362e-01 4.09545690e-01
7.27193356e-01 2.54929632e-01 5.67807376e-01 -1.35540271e+00
-1.84669733e-01 -6.31303072e-01 -9.04946804e-01 -1.03553581e+00
2.24350497e-01 -9.35271204e-01 -2.15089172e-01 -1.34180331e+00
1.61862463e-01 -5.31604826e-01 -4.59454507e-01 3.27350885e-01
-2.66579509e-01 1.85656309e-01 -4.87100750e-01 6.12070560e-01
-2.43275568e-01 5.26266694e-01 1.93360651e+00 -8.52654949e-02
-2.95987636e-01 2.90952533e-01 -8.76811266e-01 1.01655304e+00
6.40352428e-01 -6.74006017e-03 -2.46795326e-01 -3.13089073e-01
2.80273438e-01 -3.04369658e-01 8.62390280e-01 -9.19633090e-01
1.80243850e-01 -7.25454539e-02 4.93416131e-01 -1.04828909e-01
-2.12546229e-01 -7.77963102e-01 -4.19673532e-01 2.58841544e-01
-4.01271619e-02 -3.80725592e-01 -5.39995953e-02 8.85960877e-01
-5.99457800e-01 -4.57408726e-02 1.06061661e+00 -1.50108516e-01
-5.28151870e-01 7.83488512e-01 1.37794003e-01 2.36788049e-01
7.81183958e-01 -6.31123900e-01 -6.26197219e-01 2.29468849e-02
-6.72863722e-01 1.52323201e-01 4.77143973e-01 -3.20728086e-02
1.01694608e+00 -1.32228327e+00 -7.94748247e-01 6.94796503e-01
2.25896344e-01 3.04615736e-01 4.45714474e-01 1.69823778e+00
-6.46045685e-01 4.06969488e-02 -5.31774223e-01 -8.47300172e-01
-1.02693403e+00 2.04807803e-01 1.08696830e+00 2.11812869e-01
-1.10284412e+00 6.46300733e-01 6.47262454e-01 2.42050141e-01
8.95722881e-02 -5.35031796e-01 -4.57261980e-01 -2.73680001e-01
7.45259643e-01 1.25090957e-01 1.29013583e-01 -5.17929256e-01
7.52830133e-02 6.96528018e-01 -3.95627141e-01 6.02710664e-01
1.07838249e+00 -3.81156147e-01 -7.45150566e-01 1.55356571e-01
1.09341812e+00 1.06685348e-01 -1.27731287e+00 -5.60988545e-01
-1.35420904e-01 -9.19648468e-01 3.18988383e-01 -5.42809904e-01
-1.56236517e+00 7.89734483e-01 9.57029998e-01 2.74176836e-01
1.37935722e+00 -3.47354442e-01 1.00521374e+00 4.36538905e-01
-1.02405295e-01 -6.34622514e-01 1.73547596e-01 9.99833196e-02
6.98490858e-01 -1.25806832e+00 -2.24670157e-01 -4.25540149e-01
-4.38686728e-01 1.20369756e+00 7.64451921e-01 -1.05942577e-01
6.12222075e-01 -2.41049200e-01 2.27651000e-01 -4.07905430e-01
-1.82559788e-01 -4.19796139e-01 5.61231554e-01 6.95948303e-01
1.70343846e-01 -1.96290061e-01 -6.12265244e-02 6.56838775e-01
2.40389146e-02 -2.54986007e-02 6.82791412e-01 3.86476606e-01
-5.78039646e-01 -7.78707087e-01 1.09703345e-02 9.92062449e-01
-6.69609189e-01 -1.76231176e-01 -3.13754290e-01 5.69065511e-01
2.27004588e-01 5.99225640e-01 2.38335118e-01 -1.42833352e-01
3.85688841e-01 -1.85974225e-01 4.81117606e-01 -6.43903136e-01
1.48688164e-02 2.65898794e-01 -2.20885843e-01 -8.41439605e-01
-4.03753191e-01 -3.59903097e-01 -1.06167936e+00 -9.53508839e-02
-1.38832644e-01 -1.35550588e-01 6.40665293e-02 8.93805206e-01
4.34836507e-01 4.50905889e-01 8.03107142e-01 -9.96386930e-02
-3.79338950e-01 -8.24652672e-01 -1.07958329e+00 4.99677360e-01
8.60006928e-01 -7.43609548e-01 -4.10017490e-01 6.83966056e-02] | [15.815592765808105, -3.9953906536102295] |
9ce4cb46-b028-4b91-a42b-21a28e71a4a3 | uor-universal-backdoor-attacks-on-pre-trained | 2305.09574 | null | https://arxiv.org/abs/2305.09574v1 | https://arxiv.org/pdf/2305.09574v1.pdf | UOR: Universal Backdoor Attacks on Pre-trained Language Models | Backdoors implanted in pre-trained language models (PLMs) can be transferred to various downstream tasks, which exposes a severe security threat. However, most existing backdoor attacks against PLMs are un-targeted and task-specific. Few targeted and task-agnostic methods use manually pre-defined triggers and output representations, which prevent the attacks from being more effective and general. In this paper, we first summarize the requirements that a more threatening backdoor attack against PLMs should satisfy, and then propose a new backdoor attack method called UOR, which breaks the bottleneck of the previous approach by turning manual selection into automatic optimization. Specifically, we define poisoned supervised contrastive learning which can automatically learn the more uniform and universal output representations of triggers for various PLMs. Moreover, we use gradient search to select appropriate trigger words which can be adaptive to different PLMs and vocabularies. Experiments show that our method can achieve better attack performance on various text classification tasks compared to manual methods. Further, we tested our method on PLMs with different architectures, different usage paradigms, and more difficult tasks, which demonstrated the universality of our method. | ['Gongshen Liu', 'Haodong Zhao', 'Boqun Li', 'Peixuan Li', 'Wei Du'] | 2023-05-16 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [-5.38259260e-02 -4.22344297e-01 -6.12565756e-01 -4.92671989e-02
-7.37426579e-01 -1.24548972e+00 8.62781584e-01 5.18419631e-02
-5.39355516e-01 4.88578826e-01 2.49433532e-01 -7.17037380e-01
1.76496282e-01 -7.39960968e-01 -4.74689066e-01 -6.47709668e-01
5.87900616e-02 5.43692261e-02 4.79561180e-01 -5.22530556e-01
2.44722158e-01 5.38420022e-01 -7.18051016e-01 3.91234100e-01
5.76042593e-01 6.13964021e-01 5.88550158e-02 4.61368173e-01
-2.09072769e-01 3.51935148e-01 -9.71941829e-01 -5.27313769e-01
2.64457077e-01 -5.08929417e-02 -6.28449678e-01 -5.89002430e-01
2.61593819e-01 -2.18778878e-01 -5.88346720e-01 1.22460008e+00
5.88975549e-01 -4.82251160e-02 6.77499115e-01 -1.18993461e+00
-7.04986751e-01 1.14397597e+00 -4.85205263e-01 4.48770434e-01
7.51298293e-02 3.23293626e-01 1.18677688e+00 -7.08368182e-01
7.19319731e-02 1.43964064e+00 2.95508593e-01 9.09746051e-01
-1.08555925e+00 -1.43988836e+00 6.21713102e-01 3.90687259e-03
-1.25237465e+00 -3.44707251e-01 7.76364207e-01 -3.00178915e-01
1.04462051e+00 4.50706422e-01 2.70166732e-02 1.83982575e+00
4.23303068e-01 8.33195388e-01 6.50794387e-01 -3.77394468e-01
6.74547628e-02 4.30920899e-01 2.87449688e-01 5.77308059e-01
5.25975585e-01 3.90752733e-01 -5.27737200e-01 -7.03573525e-01
1.77156657e-01 -3.05309775e-03 -4.77848977e-01 -1.24408424e-01
-1.03258538e+00 1.12887478e+00 2.39150882e-01 5.16952455e-01
3.30742031e-01 -1.07377190e-02 7.44733930e-01 2.96522588e-01
2.94950694e-01 7.60683656e-01 -8.02182376e-01 5.05096436e-01
-6.00742757e-01 1.05849370e-01 7.23606586e-01 6.88125551e-01
4.56304580e-01 1.10323131e-01 -4.62490231e-01 4.69166905e-01
4.09098208e-01 5.22112250e-01 8.07936549e-01 6.57555787e-03
8.23040128e-01 3.21645558e-01 -1.10991441e-01 -1.07116091e+00
-4.59665745e-01 -5.89461505e-01 -5.73883533e-01 -3.58897209e-01
2.25801423e-01 -4.23092842e-01 -8.30968380e-01 2.06342077e+00
1.79842114e-01 1.54324263e-01 9.85971093e-02 6.15070760e-01
3.88976276e-01 8.18739057e-01 3.07149827e-01 -1.83653638e-01
1.37771130e+00 -7.32169807e-01 -8.39718401e-01 -4.56541002e-01
1.11202252e+00 -5.83422959e-01 1.52911496e+00 4.73217279e-01
-3.98112684e-01 -2.96577346e-02 -1.47171938e+00 2.39888862e-01
-6.49830520e-01 -1.41419858e-01 5.86475968e-01 1.17869890e+00
-4.08097595e-01 2.76925921e-01 -5.19476891e-01 1.07189529e-02
2.54914284e-01 4.18522686e-01 -1.49143368e-01 3.43699068e-01
-1.96899855e+00 7.86127865e-01 7.43716538e-01 -2.10156083e-01
-1.35537553e+00 -5.28127313e-01 -8.02231073e-01 5.30195720e-02
6.16629720e-01 -2.07169324e-01 1.08704925e+00 -2.95598716e-01
-1.66927969e+00 6.74569547e-01 2.16886535e-01 -5.75046480e-01
1.43883765e-01 -5.58439851e-01 -5.40846288e-01 -1.22874312e-01
-1.30578220e-01 1.71513319e-01 1.21673226e+00 -1.02390194e+00
-4.51202095e-01 -1.45727396e-01 1.91816345e-01 2.68022139e-02
-1.25595844e+00 4.18900430e-01 -2.93540955e-01 -9.32250559e-01
-3.73626381e-01 -6.16918921e-01 -3.36044461e-01 -6.83181107e-01
-7.72948563e-01 -3.36784095e-01 9.66386020e-01 -2.54652947e-01
1.82638657e+00 -2.28167200e+00 3.32922548e-01 3.26173812e-01
3.65205556e-01 7.52661705e-01 -2.61205703e-01 5.58105767e-01
-6.77914768e-02 6.20654225e-01 -2.42200419e-02 -1.83329463e-01
9.01983455e-02 3.62650789e-02 -1.26879179e+00 5.05647421e-01
-1.73680946e-01 7.03599572e-01 -7.69363225e-01 -3.22062075e-01
3.07348780e-02 2.29611974e-02 -6.00484550e-01 2.42795438e-01
-5.10723174e-01 2.22739100e-01 -7.74392664e-01 4.35185283e-01
4.66731340e-01 6.37558252e-02 1.01770841e-01 -5.71704935e-03
1.48886114e-01 6.08568609e-01 -7.78182089e-01 1.28045213e+00
-5.48763156e-01 3.39970082e-01 -4.66730632e-02 -8.06535006e-01
9.18262839e-01 3.89886647e-01 -4.26140688e-02 -2.73050189e-01
3.64828020e-01 1.94126755e-01 2.34799329e-02 -2.30736017e-01
1.55508190e-01 2.01168973e-02 -5.64551055e-01 6.03002429e-01
1.93515886e-02 1.42337731e-03 -2.59501129e-01 3.87464732e-01
9.90693748e-01 -4.67544824e-01 3.47172022e-01 -3.78793240e-01
7.51535416e-01 -3.26114565e-01 4.66834575e-01 1.08841527e+00
-3.21006984e-01 -1.16732500e-01 6.04110301e-01 -3.18464488e-01
-4.33341384e-01 -9.77740586e-01 -6.56247959e-02 1.51519513e+00
1.52729362e-01 -9.15156186e-01 -7.81026483e-01 -1.25686824e+00
-7.38639906e-02 8.85191083e-01 -3.77223074e-01 -8.15735936e-01
-6.56023979e-01 -9.10365701e-01 1.24200976e+00 2.05523014e-01
3.80447090e-01 -9.68142509e-01 -2.14063928e-01 5.69807962e-02
8.09186101e-02 -1.11543572e+00 -9.43745792e-01 3.36478025e-01
-6.03457868e-01 -9.86910939e-01 -2.68765479e-01 -6.56870127e-01
3.81863981e-01 2.59856045e-01 6.19997740e-01 6.41498715e-02
-1.67236999e-02 -1.03860319e-01 -2.02801749e-01 -5.42057991e-01
-3.64746869e-01 6.02385581e-01 6.05568409e-01 4.53057550e-02
5.13680398e-01 -2.40805119e-01 -1.98330551e-01 4.00867820e-01
-1.05674958e+00 -5.00908971e-01 3.73392999e-01 7.98845649e-01
5.59733845e-02 1.02630801e-01 7.43567884e-01 -9.90982056e-01
1.18275464e+00 -4.99886453e-01 -8.28741610e-01 3.72614354e-01
-4.15835083e-01 2.16232002e-01 1.04414415e+00 -9.90120769e-01
-6.57481551e-01 -4.98464443e-02 -1.32786170e-01 -5.37368655e-01
1.83900520e-01 2.13136658e-01 -6.52263343e-01 -1.41542079e-02
1.11457837e+00 2.58275092e-01 -5.83116531e-01 -4.11949545e-01
6.82465136e-01 7.26513863e-01 1.45737737e-01 -8.88381422e-01
1.27401137e+00 3.29437315e-01 -4.45055723e-01 -8.34801197e-01
-9.64752078e-01 -1.75743327e-01 -1.48227751e-01 3.29100728e-01
7.34181345e-01 -6.43853843e-01 -7.45944917e-01 4.32267487e-01
-1.31223142e+00 -3.17931026e-01 4.02527779e-01 3.36704552e-01
1.88531876e-02 6.04355514e-01 -7.39633620e-01 -7.44160593e-01
-6.15771949e-01 -1.25135696e+00 7.77078986e-01 5.57949720e-03
-2.34872505e-01 -9.37993288e-01 -2.83291359e-02 -2.95687541e-02
3.24234217e-01 -3.20200920e-01 1.28521812e+00 -1.38626122e+00
-2.77772337e-01 -3.74556333e-01 1.55784443e-01 2.72015393e-01
1.50754005e-01 -3.38117748e-01 -1.02130020e+00 -5.66262245e-01
2.82101303e-01 -4.39854294e-01 9.70758677e-01 -1.16876356e-01
1.43541348e+00 -8.18435967e-01 -8.18233788e-01 7.94214547e-01
8.98594022e-01 2.15744644e-01 2.49469861e-01 4.18743730e-01
7.34001637e-01 5.11651874e-01 4.62600350e-01 8.91375840e-02
-2.46562600e-01 7.23756075e-01 4.25781101e-01 4.37860116e-02
6.28956914e-01 -6.03103638e-01 8.77942085e-01 5.75986803e-01
7.98893392e-01 -4.53929782e-01 -8.06539536e-01 2.80124515e-01
-1.63272119e+00 -6.42488897e-01 3.08744580e-01 2.16574621e+00
1.18683875e+00 6.69088483e-01 2.01808661e-02 7.72684887e-02
7.08416045e-01 3.79840761e-01 -5.30749261e-01 -4.20952916e-01
8.86727720e-02 2.61438906e-01 7.14243293e-01 6.00906909e-01
-1.34913266e+00 1.66846406e+00 6.49054432e+00 1.38918364e+00
-1.35763109e+00 3.08785409e-01 3.21104854e-01 -2.76234567e-01
-4.65106517e-01 -2.51655169e-02 -1.40298176e+00 3.52585882e-01
7.49908864e-01 -2.59759814e-01 3.76157671e-01 8.05347681e-01
-1.33748308e-01 9.23907816e-01 -1.07408750e+00 8.44225347e-01
-1.13300294e-01 -1.25986147e+00 3.59176904e-01 9.77641717e-02
2.87532657e-01 7.16781523e-03 3.50974679e-01 5.88275433e-01
5.79169095e-01 -1.07224762e+00 4.85535026e-01 -3.16741914e-01
6.72416627e-01 -9.17103231e-01 2.57627398e-01 6.97069347e-01
-9.56777871e-01 -4.22312438e-01 -4.75973636e-01 2.36273259e-01
-2.33907923e-01 2.85630435e-01 -7.56523073e-01 2.73393273e-01
5.70621848e-01 2.25949660e-01 -2.83691615e-01 3.46152097e-01
-7.69822717e-01 8.93823504e-01 -3.78247261e-01 -5.18659890e-01
4.25419331e-01 3.23495120e-01 8.03578377e-01 1.47380376e+00
-1.60467744e-01 -1.72190577e-01 4.79309678e-01 5.99260449e-01
-1.95109516e-01 3.47263753e-01 -1.03284287e+00 -3.09044838e-01
8.23880732e-01 1.23923039e+00 -3.36342067e-01 -7.59762004e-02
-1.70932382e-01 7.00718343e-01 3.66861016e-01 4.81779873e-01
-1.00564694e+00 -5.72809100e-01 8.96474361e-01 -9.80814397e-02
-1.87170014e-01 -4.43508297e-01 -1.14735872e-01 -1.44066417e+00
-3.16358835e-01 -1.30845582e+00 8.00064921e-01 2.42184214e-02
-1.39378071e+00 7.11711407e-01 1.56398833e-01 -8.51796865e-01
4.42222226e-03 -9.46870089e-01 -8.07567477e-01 7.89627850e-01
-1.19190502e+00 -9.36473489e-01 6.13369286e-01 7.74179041e-01
5.53541183e-01 -7.08337247e-01 8.76656175e-01 1.85544640e-01
-1.04913437e+00 1.18677270e+00 -2.71691203e-01 4.07044858e-01
9.41687167e-01 -8.11908782e-01 5.69419980e-01 1.05282545e+00
4.24983025e-01 1.20600164e+00 5.45451283e-01 -6.59648776e-01
-1.25805163e+00 -1.16692531e+00 6.97502851e-01 -6.11450136e-01
1.00621092e+00 -1.09072793e+00 -8.01129282e-01 8.52660477e-01
1.13693118e-01 -3.37841660e-01 8.50612342e-01 3.03967059e-01
-8.65291655e-01 8.37667435e-02 -9.03044701e-01 1.24413311e+00
1.00936532e+00 -6.97215140e-01 -6.35850549e-01 6.40404820e-01
1.42895770e+00 -2.47833952e-01 -1.38115868e-01 3.91150236e-01
1.84216455e-01 -3.00308883e-01 1.02537286e+00 -1.19018948e+00
-1.43934146e-01 -1.23329163e-01 -3.09017926e-01 -1.23447442e+00
-2.46340603e-01 -1.00874746e+00 -2.56947696e-01 1.13770056e+00
6.69665217e-01 -1.11773467e+00 3.72143984e-01 2.24585626e-02
1.07776709e-01 -7.18019664e-01 -9.09107268e-01 -7.84125626e-01
4.33131039e-01 -6.57816648e-01 7.17776895e-01 1.13410211e+00
2.31763363e-01 7.50525713e-01 -6.88228488e-01 6.30009115e-01
4.49918419e-01 -3.34558189e-01 7.11896300e-01 -8.48282814e-01
-5.25205493e-01 -6.96379542e-01 5.79394139e-02 -1.13199854e+00
4.49178994e-01 -1.25464618e+00 -3.06331605e-01 -6.25447094e-01
-1.36062488e-01 -4.07778293e-01 -5.64862311e-01 8.04158688e-01
-3.09001029e-01 -1.06702395e-01 1.18131727e-01 2.60754198e-01
-2.34248072e-01 5.49688876e-01 8.07726622e-01 -4.70055401e-01
-3.89142752e-01 2.12090477e-01 -1.12522054e+00 8.13456178e-01
1.00689816e+00 -8.61985028e-01 -6.55972302e-01 -3.88873130e-01
3.09227079e-01 -4.55087960e-01 -4.29496281e-02 -5.07819653e-01
1.87905207e-01 -2.61297971e-01 4.74244989e-02 -3.06755662e-01
7.09089190e-02 -6.13656044e-01 -7.11375177e-01 7.05444634e-01
-5.89437544e-01 -1.27964318e-01 5.32359242e-01 6.25660360e-01
2.83358842e-01 -3.71845841e-01 6.84856057e-01 8.42301641e-03
-3.10327500e-01 5.21869779e-01 -3.86922777e-01 6.56824932e-02
1.01854551e+00 4.43556637e-01 -3.67997795e-01 -1.91595390e-01
-3.35949183e-01 4.35712427e-01 -1.49391919e-01 8.53779674e-01
6.87747061e-01 -1.08397603e+00 -6.11456931e-01 3.63523543e-01
1.28895596e-01 -4.09736961e-01 -3.07341754e-01 2.63697475e-01
-7.56386146e-02 7.31747150e-01 3.84042561e-01 -3.78859818e-01
-1.17284608e+00 1.13768482e+00 5.83136916e-01 -5.07106245e-01
-5.99251509e-01 9.15896416e-01 7.93394446e-01 -5.89029968e-01
5.52747726e-01 -3.52297015e-02 -3.02871794e-01 -1.31981775e-01
6.90255344e-01 -2.08146811e-01 -5.07345945e-02 -8.92077535e-02
-5.46637654e-01 2.14093566e-01 -6.64943099e-01 -1.53824344e-01
6.34878457e-01 3.62007141e-01 5.64189255e-02 1.10772550e-01
1.10954142e+00 5.37955344e-01 -7.18073845e-01 -4.46055740e-01
2.25071475e-01 -3.14845800e-01 2.53954232e-02 -6.18115246e-01
-8.87628436e-01 1.01163113e+00 2.46400252e-01 3.29012096e-01
8.82936716e-01 -8.75656754e-02 1.03695631e+00 6.59952939e-01
4.18374091e-01 -7.72663236e-01 2.71239877e-01 6.70468211e-01
7.96483994e-01 -7.48173654e-01 -2.75502831e-01 -3.24010551e-01
-5.39885819e-01 7.96733677e-01 8.78103912e-01 -2.06709486e-02
7.25643754e-01 4.36109155e-01 1.31163314e-01 -2.06621196e-02
-8.75954449e-01 2.13178918e-01 2.14859501e-01 2.78496802e-01
1.57208890e-01 1.76996049e-02 -4.74115878e-01 9.68126416e-01
-1.98521361e-01 -7.52883494e-01 9.10531580e-02 6.02654755e-01
-6.01114392e-01 -1.45985246e+00 -5.40523052e-01 9.78347063e-02
-8.40318263e-01 -3.21208328e-01 -7.20448494e-01 6.37605667e-01
-1.12313211e-01 1.09714472e+00 -5.70727885e-01 -9.39752698e-01
2.18667626e-01 2.34458566e-01 1.59001827e-01 -9.79203820e-01
-8.11556280e-01 -5.15378974e-02 -1.49002463e-01 -3.42861146e-01
6.04754269e-01 -8.07986781e-02 -1.12476027e+00 -2.17954181e-02
-6.34371638e-01 3.18081439e-01 4.01584327e-01 1.02629101e+00
2.03625187e-01 3.30596983e-01 8.87948155e-01 -4.30421799e-01
-1.25367320e+00 -8.59063983e-01 -2.08488077e-01 7.13804364e-02
2.69016057e-01 -6.87682807e-01 -5.86669505e-01 -5.66333115e-01] | [6.086513042449951, 7.950823783874512] |
eff4cc88-d702-4e7c-9aa1-3659984309e6 | wild-devs-at-semeval-2017-task-2-using-neural | null | null | https://aclanthology.org/S17-2042 | https://aclanthology.org/S17-2042.pdf | Wild Devs' at SemEval-2017 Task 2: Using Neural Networks to Discover Word Similarity | This paper presents Wild Devs{'} participation in the SemEval-2017 Task 2 {``}Multi-lingual and Cross-lingual Semantic Word Similarity{''}, which tries to automatically measure the semantic similarity between two words. The system was build using neural networks, having as input a collection of word pairs, whereas the output consists of a list of scores, from 0 to 4, corresponding to the degree of similarity between the word pairs. | ['Diana ab{\\u{a}}{\\textcommabelow{t}}', 'Tr', 'Alina Beatrice Loren{\\c{t}}', 'Mihaela Pl{\\u{a}}mad{\\u{a}}-Onofrei', '{\\textcommabelow{S}}tefan Oprea', 'Ionu{\\textcommabelow{t}} Hulub', 'R{\\u{a}}zvan-Gabriel Rotari', 'Adrian Iftene', 'Raluca Preisler'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['multilingual-word-embeddings'] | ['methodology'] | [-1.44261867e-02 4.77673346e-03 2.36507609e-01 -7.41667628e-01
-7.82028377e-01 -7.20921516e-01 7.85402298e-01 8.44334960e-01
-1.12868166e+00 2.69828200e-01 3.28887075e-01 -1.70309678e-01
2.10750476e-02 -5.82888663e-01 -4.36830580e-01 -3.19095284e-01
3.10789913e-01 7.81178176e-01 2.55979747e-01 -5.66644788e-01
2.18582079e-01 -2.70614117e-01 -1.31299877e+00 3.43774885e-01
6.74319148e-01 1.28132689e+00 2.84445345e-01 2.80811101e-01
-3.43263209e-01 1.55370668e-01 -5.89615166e-01 -9.43210304e-01
7.37400576e-02 -5.07899761e-01 -9.21456873e-01 -7.14143097e-01
7.35447228e-01 3.45948756e-01 4.08694595e-02 1.34660447e+00
6.25785530e-01 4.41754371e-01 7.50806034e-01 -9.71150815e-01
-1.02911425e+00 1.00011086e+00 1.51643068e-01 6.94433331e-01
5.27831316e-01 -8.42106044e-02 1.55787706e+00 -1.23982155e+00
6.44186974e-01 1.26477492e+00 6.54313922e-01 5.02807200e-01
-1.12800157e+00 -5.26239574e-01 -3.27829123e-01 5.60312271e-01
-1.13921273e+00 -2.58903056e-01 4.97444630e-01 -3.50556433e-01
1.47627842e+00 -2.74240971e-01 5.13875246e-01 1.22428882e+00
-4.10515815e-01 5.99672735e-01 9.92336810e-01 -5.80738723e-01
3.25151443e-01 1.77761748e-01 5.29095531e-01 3.61883223e-01
1.42414451e-01 -4.39405926e-02 -6.39331281e-01 8.60263109e-02
1.49053231e-01 -4.78859186e-01 -6.43698499e-02 -1.69316661e-02
-1.14754176e+00 9.00133133e-01 3.43691558e-01 7.90216208e-01
-3.04415643e-01 -6.84091374e-02 9.04528141e-01 4.93811846e-01
5.79236031e-01 8.63362551e-01 -6.05430603e-01 -2.30706140e-01
-5.54179192e-01 4.29895580e-01 5.30486822e-01 6.66845500e-01
7.61037290e-01 -2.64410116e-02 -2.88231164e-01 1.17827415e+00
3.04479688e-01 3.82923871e-01 1.16665363e+00 -7.53388643e-01
4.51803684e-01 4.57453817e-01 -3.94340783e-01 -7.92054951e-01
-8.73729438e-02 5.21714613e-02 -3.81953061e-01 -2.92994916e-01
1.68348685e-01 -7.89954513e-02 -5.59981167e-01 2.22360992e+00
1.89915821e-02 1.50891781e-01 3.65007997e-01 6.41687870e-01
1.46721959e+00 3.23672295e-01 3.86216819e-01 7.41847008e-02
1.31806767e+00 -7.32558548e-01 -7.17261493e-01 -4.08617944e-01
8.56376231e-01 -8.53683650e-01 1.31544340e+00 -1.48917243e-01
-1.32292974e+00 -7.29320407e-01 -9.30727482e-01 -2.14249253e-01
-1.10695112e+00 -2.46603504e-01 1.76170468e-02 3.63904804e-01
-1.39330602e+00 8.55815828e-01 -2.32406974e-01 -5.57670295e-01
1.29148483e-01 -4.66946885e-02 -5.07702291e-01 -4.20371965e-02
-1.92726040e+00 1.60624886e+00 1.09814477e+00 -4.71552372e-01
-3.45108271e-01 -9.99921203e-01 -1.17134523e+00 -1.53768640e-02
-1.92325547e-01 -4.70278472e-01 1.14889991e+00 -1.04524863e+00
-1.03737569e+00 1.72588420e+00 -1.23180831e-02 -5.88830113e-01
1.76527515e-01 -1.73508748e-01 -7.45469451e-01 -1.46908537e-01
5.12519836e-01 8.51875722e-01 2.24402234e-01 -7.25000918e-01
-5.34639239e-01 -3.56822520e-01 -1.13356307e-01 3.02165747e-01
-5.14308035e-01 4.62309361e-01 -9.44168642e-02 -9.44436371e-01
-2.56421387e-01 -6.31839156e-01 4.33998778e-02 -4.63950604e-01
-1.12793297e-01 -1.07593119e+00 3.21292996e-01 -8.35542202e-01
8.90666902e-01 -2.22928619e+00 3.03002477e-01 1.03883661e-01
-2.67799854e-01 5.05468726e-01 -5.83830297e-01 4.74539727e-01
-4.22609389e-01 -6.06682673e-02 -3.14060956e-01 -4.75908071e-01
2.66925097e-01 5.15176132e-02 6.65419474e-02 6.04924746e-02
5.21023534e-02 8.52129102e-01 -1.18479335e+00 -1.97352454e-01
1.61614686e-01 4.17234361e-01 -2.21881509e-01 3.14725786e-01
-1.95863068e-01 -2.64744848e-01 -5.68986088e-02 -1.11274309e-01
1.59004718e-01 3.40597302e-01 1.56487331e-01 -9.99407843e-02
1.87650800e-01 7.48303592e-01 -6.81300819e-01 2.15209103e+00
-5.35750151e-01 5.08665383e-01 -5.39807498e-01 -1.07020330e+00
9.64635789e-01 4.03657615e-01 1.76583514e-01 -9.21704292e-01
5.68163514e-01 5.31591415e-01 9.00484156e-03 -4.75256771e-01
5.76876462e-01 -2.50118345e-01 -4.89797622e-01 4.80227530e-01
4.39477265e-01 -3.55549127e-01 5.12539804e-01 -5.90382516e-03
1.00913525e+00 2.63698660e-02 2.85915345e-01 -7.01685727e-01
7.30856895e-01 -3.34183335e-01 -4.60799709e-02 1.37204319e-01
-1.28620997e-01 3.48611295e-01 2.29042605e-01 -4.63646084e-01
-9.70788836e-01 -1.38401401e+00 3.34987752e-02 1.30186200e+00
-1.98153127e-02 -5.24872303e-01 -9.60227489e-01 -5.03427386e-01
3.03518251e-02 1.27161527e+00 -8.74523997e-01 -4.45832074e-01
-4.97364402e-01 -2.93156326e-01 7.43616462e-01 7.22884953e-01
1.96253359e-01 -1.23104358e+00 -4.10073429e-01 2.14010440e-02
-4.22513515e-01 -1.28093481e+00 -5.64238966e-01 2.60684699e-01
-4.01848227e-01 -8.20746362e-01 -6.31424725e-01 -1.23117590e+00
1.78791910e-01 -1.04357295e-01 1.68780422e+00 -1.48077264e-01
-3.41527164e-01 6.32917136e-02 -4.95839655e-01 -3.98674518e-01
-4.24492747e-01 3.95008028e-02 3.49870145e-01 -3.41074139e-01
9.16865230e-01 -5.65768898e-01 -1.96826458e-01 -2.20338106e-01
-7.10941613e-01 -3.05825233e-01 1.48087725e-01 6.06188059e-01
6.03406668e-01 -5.36103725e-01 6.11122787e-01 -6.59827471e-01
9.53718364e-01 -6.53816402e-01 -2.08470538e-01 3.31965744e-01
-5.40351927e-01 2.16599002e-01 8.64257038e-01 -2.19221681e-01
-6.88313127e-01 -5.38642943e-01 -3.32572460e-01 -1.97924584e-01
-4.84402061e-01 4.15842474e-01 -3.27662706e-01 4.34278816e-01
8.08506131e-01 7.84960613e-02 -3.25180709e-01 -5.77114046e-01
9.06275034e-01 5.61464310e-01 8.56717288e-01 -6.59220755e-01
3.12267125e-01 -3.00063074e-01 -2.96375006e-01 -3.84589911e-01
-1.12257886e+00 -6.01023734e-01 -6.99766815e-01 1.15673378e-01
1.13598394e+00 -6.53425574e-01 -3.43579203e-01 4.11827415e-01
-1.42536569e+00 -9.93875340e-02 -4.68780398e-01 6.11887991e-01
-4.77892965e-01 3.36003192e-02 -3.83446932e-01 1.33930624e-01
-5.82340240e-01 -7.21599460e-01 8.07797790e-01 -5.18114790e-02
-8.04160297e-01 -1.45631993e+00 5.59730351e-01 4.16695863e-01
5.35615027e-01 4.89653982e-02 1.43007040e+00 -1.52138805e+00
4.85003799e-01 -3.49263191e-01 -1.82136163e-01 6.91248953e-01
8.08962286e-02 -3.60259295e-01 -7.52069235e-01 -3.19105417e-01
-3.42150569e-01 -6.55277908e-01 7.32273400e-01 1.75941080e-01
9.22154784e-01 -1.68184578e-01 -9.55193490e-02 2.70765603e-01
1.36487949e+00 -3.26385386e-02 2.60564506e-01 3.53751063e-01
6.00298762e-01 7.45598018e-01 2.75755256e-01 2.33009178e-03
5.84135771e-01 1.01437867e+00 5.95301762e-02 2.23450989e-01
-3.22583437e-01 -1.87265381e-01 3.79945934e-01 1.23533702e+00
1.18373027e-02 -2.76170015e-01 -9.35034215e-01 8.52292717e-01
-1.47922873e+00 -7.53775537e-01 -1.00821383e-01 2.24692154e+00
9.14323926e-01 -1.27490208e-01 8.10322315e-02 3.24928686e-02
8.84159386e-01 3.29477340e-01 -2.31877252e-01 -9.66631413e-01
-8.24756399e-02 8.47373068e-01 8.91627371e-02 3.97879124e-01
-9.51445878e-01 1.44879782e+00 5.93112707e+00 1.04856241e+00
-1.07047260e+00 2.61364013e-01 1.31001323e-01 -2.58788979e-03
-5.40813565e-01 -3.11894685e-01 -5.06475091e-01 6.81454360e-01
1.40788412e+00 -4.83572096e-01 1.56822443e-01 5.71866572e-01
-3.43860030e-01 1.55223444e-01 -1.25673592e+00 8.04413855e-01
6.67457104e-01 -9.64781463e-01 3.00238192e-01 -7.31652617e-01
6.02547109e-01 4.23431247e-01 -3.19848731e-02 2.90475398e-01
5.41752100e-01 -9.90048766e-01 7.44865954e-01 2.02663437e-01
8.82874727e-01 -8.71193826e-01 8.71887028e-01 9.99004096e-02
-9.47594106e-01 5.46470046e-01 -1.46338984e-01 1.06981859e-01
1.58782780e-01 3.19397241e-01 -6.90382183e-01 2.81803250e-01
8.81833792e-01 8.24461639e-01 -7.23675907e-01 5.63997507e-01
-6.58120096e-01 3.19380879e-01 3.70865203e-02 -2.22438216e-01
3.32305551e-01 9.20896512e-03 3.14933330e-01 1.50306666e+00
3.26727629e-01 -3.69602442e-02 3.28850299e-02 5.75376511e-01
-3.81593913e-01 6.58188939e-01 -6.14103794e-01 -1.86070800e-02
6.32601202e-01 1.05723631e+00 -4.75205719e-01 -5.22177756e-01
-2.30784059e-01 1.11370170e+00 8.56952965e-01 -2.42486894e-01
-4.80367213e-01 -6.30526125e-01 8.20215106e-01 -1.62649602e-01
5.61967976e-02 4.04163785e-02 -2.22649172e-01 -8.09852660e-01
-1.36910558e-01 -3.63786668e-01 5.79589367e-01 -8.09182405e-01
-1.82187569e+00 9.68218327e-01 -1.91805825e-01 -5.78050137e-01
-4.54317927e-01 -8.94804835e-01 -6.33483231e-01 1.03828454e+00
-1.26705170e+00 -8.54085207e-01 -3.02973599e-03 6.13645434e-01
6.47702098e-01 -5.60746372e-01 1.27814066e+00 3.15867007e-01
-4.64734547e-02 8.98954034e-01 -8.17274302e-02 3.23119760e-01
8.72622907e-01 -1.40980232e+00 1.05157340e+00 5.82317650e-01
3.34735900e-01 4.26363438e-01 7.60622799e-01 -3.95607769e-01
-4.71595138e-01 -1.02420509e+00 1.80753314e+00 -6.46455765e-01
8.97560656e-01 -7.61625767e-02 -1.03112745e+00 4.06826347e-01
5.63786864e-01 8.28347579e-02 1.03178620e+00 7.87284225e-02
-8.47579956e-01 1.73517123e-01 -1.04806256e+00 4.15368617e-01
1.11527038e+00 -8.46793175e-01 -1.27263653e+00 2.62876838e-01
9.38854277e-01 -1.10978402e-01 -1.01839578e+00 3.72578412e-01
3.71125221e-01 -7.89413929e-01 1.07719457e+00 -1.05878162e+00
7.35173643e-01 1.39308482e-01 -5.47121465e-01 -1.97378600e+00
-3.54322284e-01 5.67434356e-04 4.76885438e-01 1.16169035e+00
6.23034596e-01 -2.36752346e-01 6.28616884e-02 1.46921650e-01
-3.60248834e-01 -4.73102123e-01 -1.19132543e+00 -9.32217598e-01
4.57490981e-01 -6.29349649e-01 4.62561697e-01 1.34686470e+00
9.57228094e-02 7.15402484e-01 3.65906954e-01 -3.40431184e-01
2.35257134e-01 -3.63222182e-01 -1.35290742e-01 -1.36478794e+00
-8.76546279e-02 -9.08515990e-01 -6.51935160e-01 -1.57073706e-01
9.26571786e-01 -1.60382652e+00 2.79234201e-01 -1.32113349e+00
5.00369109e-02 -2.22760469e-01 -6.28478348e-01 5.40221930e-01
-2.92893946e-01 5.21197855e-01 1.98988408e-01 -3.02755296e-01
-7.51532197e-01 5.01119018e-01 5.56539953e-01 6.47708178e-02
3.37215394e-01 -2.55932063e-01 -4.38589513e-01 9.12722588e-01
7.13702321e-01 -3.60201091e-01 -2.31053010e-01 -1.06928027e+00
4.39987272e-01 -3.09763789e-01 2.91403204e-01 -7.44156361e-01
-7.67407715e-02 1.51081994e-01 -9.00713913e-03 3.98273468e-02
3.38171393e-01 -5.17092109e-01 -2.55889744e-01 3.51127237e-01
-7.64128208e-01 6.11396253e-01 1.23424351e-01 8.11234340e-02
-5.13843358e-01 -4.44631517e-01 1.00339913e+00 -2.35171288e-01
-7.98790395e-01 1.84327960e-01 -3.73711549e-02 7.85898089e-01
9.38344359e-01 4.78323214e-02 -1.78142816e-01 9.88032743e-02
-6.23923361e-01 3.04349232e-02 4.21615273e-01 9.75999355e-01
3.85748386e-01 -1.86745274e+00 -9.83160138e-01 -5.81173711e-02
6.46527171e-01 -4.62467581e-01 9.11164656e-02 2.71666706e-01
-1.18914656e-01 3.22701216e-01 -3.41000617e-01 2.83715259e-02
-1.40522218e+00 3.39583337e-01 2.93422967e-01 -4.96835113e-02
-2.61157036e-01 1.38402593e+00 -4.21692222e-01 -5.82539260e-01
2.60081589e-02 -6.28681704e-02 -5.53608298e-01 4.97774571e-01
4.69505548e-01 4.53569859e-01 3.13584596e-01 -1.20646751e+00
-6.13482475e-01 7.04230547e-01 -9.32320207e-03 -3.76822889e-01
1.27651179e+00 1.57648876e-01 -3.88386607e-01 6.31459117e-01
1.71746099e+00 -5.90689301e-01 -4.44069296e-01 -7.11905122e-01
4.45707142e-01 -2.28179663e-01 -5.37974127e-02 -8.99336815e-01
-9.18121397e-01 1.01870739e+00 6.08558953e-01 -6.73424304e-02
5.40062904e-01 3.72228801e-01 1.16486728e+00 4.21633303e-01
1.24510556e-01 -1.45912468e+00 3.95034179e-02 1.04636991e+00
6.47854328e-01 -1.27239490e+00 -6.00836754e-01 8.09122622e-02
-5.55875659e-01 8.97061825e-01 4.78096485e-01 -1.92061752e-01
3.95632327e-01 -2.38844365e-01 1.09392755e-01 -2.65756130e-01
-6.56046271e-01 -4.23089445e-01 6.88868582e-01 5.95597327e-01
8.23064744e-01 2.30502218e-01 -7.92660594e-01 8.33273172e-01
-8.16862464e-01 -3.29291582e-01 -3.93807627e-02 2.54976690e-01
-4.01883304e-01 -1.15501618e+00 2.19891652e-01 1.81474417e-01
-4.66114551e-01 -5.76679111e-01 -5.69899380e-01 1.94425806e-01
6.25549853e-01 6.95926130e-01 3.72897834e-01 -1.61881343e-01
4.27907795e-01 3.78357887e-01 4.15941775e-01 -7.21831799e-01
-7.65459180e-01 -8.75873268e-01 2.23958477e-01 -5.55236518e-01
-1.91646263e-01 -7.33005345e-01 -1.35354900e+00 1.25847131e-01
1.15177624e-01 2.17701629e-01 8.83131981e-01 1.48345065e+00
-1.58346355e-01 4.04804379e-01 3.81699324e-01 -3.78117740e-01
-2.59046763e-01 -1.13765121e+00 -4.46494728e-01 1.17658985e+00
-3.03003430e-01 -2.85329372e-01 -3.49528193e-01 -2.61467323e-03] | [10.761832237243652, 9.615560531616211] |
7bd74741-59db-4c7d-9045-1c595c0b4722 | analyzing-coreference-and-bridging-in-product | null | null | https://aclanthology.org/2022.crac-1.3 | https://aclanthology.org/2022.crac-1.3.pdf | Analyzing Coreference and Bridging in Product Reviews | Product reviews may have complex discourse including coreference and bridging relations to a main product, competing products, and interacting products. Current approaches to aspect-based sentiment analysis (ABSA) and opinion summarization largely ignore this complexity. On the other hand, existing systems for coreference and bridging were trained in a different domain. We collect mention type annotations relevant to coreference and bridging for 498 product reviews. Using these annotations, we show that a state-of-the-art factuality score fails to catch coreference errors in product reviews, and that a state-of-the-art coreference system trained on OntoNotes does not perform nearly as well on product mentions. As our dataset grows, we expect it to help ABSA and opinion summarization systems to avoid entity reference errors. | ['Christopher Malon', 'Hideo Kobayashi'] | null | null | null | null | coling-crac-2022-10 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 2.70916730e-01 1.04558432e+00 -6.36213183e-01 -4.49528098e-01
-1.29976392e+00 -1.00175619e+00 6.53785825e-01 7.55475581e-01
8.18700939e-02 6.54345512e-01 1.01018083e+00 -1.76602185e-01
1.51841253e-01 -4.24415231e-01 -4.09716547e-01 -7.70685524e-02
4.92872566e-01 9.34900582e-01 1.71854272e-01 -8.32168460e-01
5.02989113e-01 -5.14497124e-02 -1.17741442e+00 8.54363978e-01
1.09296751e+00 4.10029441e-01 -3.58376443e-01 5.49078941e-01
-5.03761828e-01 5.59931934e-01 -1.00972080e+00 -1.31063414e+00
-1.78939044e-01 -2.69537866e-01 -1.28824377e+00 -1.56980425e-01
8.49466622e-01 4.54954386e-01 3.18925172e-01 1.17092943e+00
4.81203645e-01 -1.30526379e-01 3.57035756e-01 -1.03254735e+00
-7.31494784e-01 1.31082118e+00 -7.60155022e-01 1.40269399e-01
9.63415205e-01 -3.83070797e-01 1.74362576e+00 -6.29449546e-01
1.05109417e+00 1.30392754e+00 8.66518438e-01 6.33911371e-01
-1.07589316e+00 -3.84329051e-01 4.93107498e-01 -2.14828774e-01
-7.00978816e-01 -6.31299675e-01 5.35858154e-01 -2.41103664e-01
1.63055491e+00 3.46620828e-01 5.17687559e-01 8.98123205e-01
1.62740454e-01 8.58698487e-01 3.44305038e-01 -5.43545783e-01
-8.60877261e-02 3.15766603e-01 8.11293066e-01 1.26604572e-01
8.63650799e-01 -6.77852631e-01 -7.58828819e-01 -5.17111599e-01
-2.65915334e-01 -6.04303956e-01 -5.10924280e-01 -2.02815514e-02
-9.17104125e-01 8.31516087e-01 -4.79856804e-02 3.20098281e-01
-4.49999541e-01 -4.64616030e-01 6.90743625e-01 1.14024676e-01
5.98214149e-01 1.32417285e+00 -9.39910412e-01 -1.75406039e-01
-7.08255172e-01 4.65311468e-01 1.61832559e+00 1.29084980e+00
5.23009717e-01 -5.09450316e-01 1.51648030e-01 7.85467207e-01
1.05663858e-01 4.40701365e-01 3.86561453e-01 -1.28348243e+00
7.71074831e-01 8.91349137e-01 2.80815482e-01 -1.02691686e+00
-4.94172484e-01 -2.69260406e-01 -9.33751836e-02 -1.98356926e-01
7.50362501e-02 -3.97431314e-01 -4.83249843e-01 1.31970417e+00
2.05689833e-01 -4.23693210e-01 7.06442595e-01 3.71465623e-01
1.34351206e+00 4.63533670e-01 8.27455372e-02 -7.34981418e-01
1.84982705e+00 -1.02917445e+00 -1.19076598e+00 -4.76418108e-01
1.01926827e+00 -1.43686104e+00 5.29841840e-01 3.62414092e-01
-1.43550384e+00 -1.43608913e-01 -1.26913416e+00 -2.11737603e-01
-3.52239639e-01 -2.01346219e-01 8.05083096e-01 3.14153254e-01
-5.20675778e-01 4.88924712e-01 -4.59971279e-01 -6.74002349e-01
1.11578919e-01 4.39778030e-01 -4.72055912e-01 2.13665515e-01
-1.18697691e+00 1.27319133e+00 3.83844972e-02 -2.18866676e-01
7.70364627e-02 -1.16663611e+00 -1.27121115e+00 -8.31827745e-02
4.05622065e-01 -8.45039904e-01 1.90408063e+00 -1.10683382e+00
-1.15778077e+00 9.36383307e-01 -6.85557127e-01 -3.92451555e-01
-3.99228722e-01 -7.90090799e-01 -8.06494296e-01 3.06749567e-02
5.82727492e-01 3.82029086e-01 3.53922844e-01 -1.36000311e+00
-9.68539715e-01 -3.43489587e-01 3.58412147e-01 5.79318106e-01
1.44412413e-01 3.71095508e-01 -6.00788817e-02 -2.61725754e-01
-4.40151654e-02 -7.59786427e-01 -4.61347736e-02 -8.61572981e-01
-6.41073465e-01 -5.69085777e-01 6.59150898e-01 -3.90444309e-01
1.36136985e+00 -1.86869013e+00 1.25738680e-01 -2.49120712e-01
1.30105745e-02 4.69231308e-01 -2.99564838e-01 6.07852399e-01
-3.75635952e-01 4.05100614e-01 1.12234615e-01 -3.67217004e-01
-4.16525193e-02 4.61963713e-02 -5.75998008e-01 -4.72665578e-02
3.66585851e-01 8.73868525e-01 -1.02396083e+00 -4.73978639e-01
-2.94759631e-01 2.65406817e-01 -4.78390396e-01 -2.11950213e-01
-4.44462001e-01 -1.82361811e-01 -4.55953091e-01 6.68115556e-01
5.91861248e-01 -2.14960381e-01 4.49594438e-01 -6.61807656e-01
-1.54266506e-01 1.08813167e+00 -8.28363717e-01 1.52813613e+00
-3.81713927e-01 6.66227043e-01 1.10386126e-01 -3.25991899e-01
7.26225197e-01 5.58248341e-01 3.83561939e-01 -4.14181411e-01
1.91836625e-01 2.85479397e-01 -2.69566812e-02 -5.02727628e-01
1.22588718e+00 -6.71905652e-02 -4.16946560e-01 4.36828613e-01
2.51676559e-01 -4.15300548e-01 5.72222054e-01 7.34782100e-01
9.35941279e-01 -1.33005962e-01 5.95627844e-01 -2.57923037e-01
6.69827759e-01 8.18924487e-01 7.58493006e-01 3.06391686e-01
-3.45190056e-02 6.74499810e-01 7.20234871e-01 -2.27783903e-01
-7.29122460e-01 -5.80399156e-01 -1.64787456e-01 1.04318559e+00
2.26923913e-01 -1.09813929e+00 -8.23378205e-01 -1.02663410e+00
-8.03646352e-03 1.25222039e+00 -6.34302437e-01 9.12279710e-02
-8.34164321e-01 -6.44098103e-01 2.30946571e-01 3.28106970e-01
-8.08853954e-02 -1.02257121e+00 -9.66162086e-02 3.92691225e-01
-5.77008784e-01 -1.17276597e+00 -6.53901875e-01 6.62305057e-02
-5.51966429e-01 -1.49755466e+00 -3.21317434e-01 -8.95311296e-01
4.15268183e-01 2.93831974e-01 1.70239615e+00 -3.15153390e-01
2.90161550e-01 5.59346914e-01 -5.14010847e-01 -8.61046195e-01
-7.18145669e-01 4.66536134e-01 -1.95425868e-01 -5.88218212e-01
1.17556155e+00 -3.07635039e-01 -3.13168973e-01 4.41119745e-02
-3.37998599e-01 -4.14134711e-01 2.95589358e-01 5.67095339e-01
2.99601346e-01 -5.28861105e-01 6.77714467e-01 -1.74704778e+00
1.02510512e+00 -2.82342345e-01 -1.18001156e-01 6.11977994e-01
-6.80477023e-01 -1.28823847e-01 1.22610480e-01 -3.30586910e-01
-1.48456812e+00 -2.47910097e-01 -1.87116772e-01 3.45875651e-01
8.42557847e-03 7.39870310e-01 -2.04805285e-01 4.23959643e-01
8.87148976e-01 -6.26148939e-01 -1.71177700e-01 -2.81406462e-01
6.58056498e-01 6.84792876e-01 5.99624336e-01 -2.93311179e-01
2.60475338e-01 3.13041776e-01 -5.12534142e-01 -6.54337049e-01
-1.61520791e+00 -9.31684613e-01 -5.87055385e-01 3.11875641e-01
7.35237479e-01 -1.04395294e+00 -6.59738958e-01 -1.03936404e-01
-1.80020475e+00 4.08830971e-01 -8.65164757e-01 3.06905180e-01
-2.79855907e-01 4.47200537e-01 -5.03014326e-01 -3.54999065e-01
-9.50066566e-01 -7.44045675e-01 1.12235975e+00 5.99907458e-01
-1.20428574e+00 -9.40108120e-01 5.70503592e-01 7.49372363e-01
8.20472091e-02 -1.28287980e-02 6.47212684e-01 -1.17806053e+00
1.14491321e-01 -3.71209949e-01 1.53278485e-01 1.52566349e-02
3.07735115e-01 2.50457942e-01 -6.95558667e-01 1.80861920e-01
8.93765688e-02 -2.09924519e-01 7.58459389e-01 4.49178189e-01
-4.42339510e-01 -4.33155864e-01 -6.66677773e-01 -8.68163258e-02
9.56514478e-01 2.16102093e-01 4.94355917e-01 4.51947749e-01
5.59151232e-01 1.00640595e+00 1.15388560e+00 1.49412021e-01
6.09733284e-01 4.29483652e-01 1.95884213e-01 2.28223011e-01
-1.74049035e-01 -1.08590037e-01 3.21722478e-01 1.07211208e+00
7.74642527e-02 -3.33126396e-01 -6.14001930e-01 1.08957171e+00
-2.01675129e+00 -1.09936130e+00 -7.19105124e-01 1.60534179e+00
1.22696483e+00 2.83859909e-01 -2.50997305e-01 -4.55126822e-01
8.70341063e-01 3.42810154e-01 -3.08908045e-01 -8.99711609e-01
-4.61444527e-01 1.81029558e-01 3.24803799e-01 8.68256152e-01
-1.02292705e+00 1.14301634e+00 7.02043772e+00 2.57889837e-01
-5.26027501e-01 1.25413835e-01 4.49994653e-02 3.71404253e-02
-7.83239901e-01 3.53159249e-01 -1.41346359e+00 -2.30228841e-01
9.08332646e-01 -4.55307066e-01 -3.14193845e-01 9.34489071e-01
-2.36106098e-01 -1.59091115e-01 -1.11513054e+00 5.23671210e-01
5.52240729e-01 -1.48656869e+00 1.51442036e-01 -2.04336420e-01
1.00402558e+00 3.70443147e-03 -2.55679816e-01 2.99309224e-01
7.19338596e-01 -6.33351922e-01 2.79917955e-01 2.46688798e-02
2.43734866e-01 -7.31932104e-01 1.38042271e+00 -2.31193826e-01
-1.08244109e+00 2.48773485e-01 -3.92193347e-01 1.39825404e-01
9.61393774e-01 5.12522280e-01 -5.11746824e-01 7.89777637e-01
3.91468734e-01 1.02157223e+00 -3.60814452e-01 7.09702730e-01
-4.66867089e-01 2.04479545e-01 -8.01806748e-02 -4.45918031e-02
1.84342335e-03 1.62691489e-01 8.36903453e-01 1.32390022e+00
2.61128768e-02 4.64892656e-01 -2.80465364e-01 2.95474023e-01
-3.31428289e-01 2.72154540e-01 -6.10321581e-01 -7.90517181e-02
6.40403926e-01 1.41702998e+00 -2.68212199e-01 -5.44694960e-01
-6.67234302e-01 6.82359099e-01 -9.04353987e-03 1.82725981e-01
-3.17778021e-01 -5.88024199e-01 1.02917576e+00 -1.59354180e-01
6.54859543e-01 3.75991136e-01 -2.43565306e-01 -1.25492358e+00
-9.71344337e-02 -1.22979903e+00 7.44544029e-01 -7.28408515e-01
-1.55836296e+00 7.20744133e-01 -1.58337623e-01 -1.01622665e+00
-2.26591781e-01 -3.12364429e-01 -8.43185067e-01 5.49379408e-01
-1.47812259e+00 -1.15928078e+00 1.84546769e-01 1.59575462e-01
8.22516680e-01 -8.50909129e-02 1.07318282e+00 -1.29557237e-01
-2.61318207e-01 4.18799669e-01 -4.66530412e-01 4.98780757e-02
1.31751084e+00 -1.38209915e+00 7.00779140e-01 5.49545050e-01
1.48762628e-01 1.06322396e+00 1.26636767e+00 -8.83510053e-01
-1.05093551e+00 -7.18650341e-01 1.78603792e+00 -9.79105115e-01
9.32572603e-01 2.15262458e-01 -8.87900174e-01 9.65675890e-01
1.26002014e+00 -5.74508011e-01 1.09760988e+00 1.02929366e+00
-6.81128740e-01 3.66616659e-02 -8.59441519e-01 4.24905539e-01
7.58047819e-01 -4.78049755e-01 -1.62309408e+00 4.52271253e-01
1.15926325e+00 -5.79624057e-01 -1.19342279e+00 5.15464962e-01
2.90398002e-01 -5.85186005e-01 5.59586942e-01 -8.03677320e-01
5.12199700e-01 -2.49684170e-01 3.23250107e-02 -1.33049083e+00
-1.20083250e-01 -9.40252185e-01 -2.03273535e-01 1.89981222e+00
1.17832625e+00 -5.78499377e-01 5.42206347e-01 9.05585229e-01
-4.64566171e-01 -3.26171368e-01 -5.11168718e-01 -1.73694164e-01
1.49492905e-01 -8.70490074e-02 4.33884889e-01 1.13172555e+00
1.06718826e+00 1.33984721e+00 2.03750238e-01 2.24523664e-01
3.06495279e-01 4.85440791e-01 6.14500165e-01 -1.49140632e+00
2.60279123e-02 -3.12663674e-01 -3.53101920e-03 -9.29815412e-01
3.43895644e-01 -4.48620528e-01 5.21098822e-03 -1.77256107e+00
2.32190475e-01 -7.31344074e-02 1.91689745e-01 3.67704988e-01
-1.61178142e-01 7.88449422e-02 7.99287409e-02 2.65902549e-01
-1.06191528e+00 1.32007897e-01 1.03454459e+00 -3.45484138e-01
-4.36934412e-01 8.78286734e-02 -1.65080667e+00 1.07201505e+00
6.46195769e-01 -6.45489275e-01 -2.69309342e-01 -1.85115814e-01
8.06642234e-01 -1.41174391e-01 -5.11481762e-01 -2.65067220e-01
6.11598551e-01 2.20942289e-01 -2.32331142e-01 -8.36134017e-01
1.73260972e-01 -3.60159069e-01 -9.97247025e-02 -1.64302625e-02
-4.00769502e-01 1.60992160e-01 3.82165313e-01 2.83169061e-01
-5.57840466e-01 -6.70754492e-01 1.35357574e-01 -3.58982861e-01
-2.58908808e-01 -4.76512551e-01 -3.07547688e-01 6.98689759e-01
6.88629866e-01 5.34556981e-04 -1.12542486e+00 -3.14868987e-01
-6.72462344e-01 3.73138070e-01 5.04731417e-01 5.91449976e-01
1.86123788e-01 -8.17551970e-01 -7.48751640e-01 -3.54114115e-01
2.46146783e-01 9.08361673e-02 1.37421833e-02 7.58511662e-01
-7.80701190e-02 7.42424965e-01 2.54903495e-01 -2.67987102e-01
-1.69475722e+00 5.30761719e-01 2.04033643e-01 -6.77880168e-01
-2.25698799e-01 8.39619219e-01 3.76583892e-03 -4.23428893e-01
1.39242992e-01 -1.37838945e-01 -7.53034770e-01 8.35694849e-01
8.06317389e-01 9.06305462e-02 3.41150701e-01 -9.47499812e-01
-5.14546812e-01 8.47554624e-01 -8.39940608e-01 -8.15762803e-02
1.20379961e+00 -4.17692333e-01 -2.45624304e-01 4.12797034e-01
8.32345903e-01 5.29851615e-01 -5.12438715e-01 -5.67982420e-02
2.25175068e-01 -7.40313297e-03 -2.90222734e-01 -8.85428846e-01
-7.09260523e-01 3.88936818e-01 -2.52556264e-01 6.81643903e-01
6.79971159e-01 5.37542045e-01 8.45000386e-01 4.66291100e-01
1.19686604e-01 -1.15042043e+00 -2.17242748e-01 7.31289148e-01
1.03554749e+00 -1.31223667e+00 4.82322037e-01 -9.17697072e-01
-1.25344062e+00 1.05775988e+00 6.25138521e-01 7.45919272e-02
5.98234534e-01 4.59683478e-01 4.44456100e-01 -7.60673225e-01
-1.10656929e+00 -2.00994655e-01 2.65620828e-01 6.90042615e-01
8.78840864e-01 -3.02289456e-01 -5.90121865e-01 9.03685689e-01
-3.82832140e-01 -3.88874888e-01 8.99549723e-01 7.60082364e-01
-2.98590541e-01 -1.31730747e+00 4.63797851e-03 2.59838253e-01
-1.03617346e+00 -4.30277348e-01 -8.95213246e-01 8.49820554e-01
-1.03571564e-01 1.31243014e+00 6.75644651e-02 1.01376651e-02
9.89746630e-01 -4.09306102e-02 2.32627615e-01 -1.06768167e+00
-1.30100906e+00 9.15902033e-02 1.09795117e+00 -5.74475646e-01
-1.14706492e+00 -1.10459960e+00 -1.37700331e+00 -3.38735610e-01
-9.69904244e-01 7.15408981e-01 4.37369287e-01 9.51330602e-01
6.47101998e-01 3.63962501e-01 9.97486785e-02 -5.33085525e-01
-1.95087031e-01 -1.05322969e+00 -3.44001293e-01 5.27101219e-01
2.65810013e-01 -2.25342453e-01 -4.86296535e-01 6.55607134e-02] | [11.156558990478516, 7.052066802978516] |
bdfb1ae1-42a7-43e9-bc9c-70f17d8aeff6 | co-attention-network-with-label-embedding-for | null | null | https://mqianliu.github.io/files/CNLE_Neurocomputing22.pdf | https://mqianliu.github.io/files/CNLE_Neurocomputing22.pdf | Co-attention network with label embedding for text classification | Most existing methods for text classification focus on extracting a highly discriminative text representation, which, however, is typically computationally inefficient. To alleviate this issue, label embedding frameworks are proposed to adopt the label-to-text attention that directly uses label information to construct the text representation for more efficient text classification. Although these label embedding methods have achieved promising results, there is still much space for exploring how to se the label information more effectively. In this paper, we seek to exploit the label information by further constructing the text-attended label representation with text-to-label attention. To this end, we propose a Coattention Network with Label Embedding (CNLE) that jointly encodes the text and labels into their mutually attended representations. In this way, the model is able to attend to the relevant parts of both. Experiments show that our approach achieves competitive results compared with previous state-ofthe-art methods on 7 multi-class classification benchmarks and 2 multi-label classification benchmarks. | ['Qing Du', 'Junyi Cao', 'Lizhao Liu', 'Minqian Liu'] | 2021-11-04 | null | null | null | neurocomputing-2021-11 | ['multi-label-text-classification', 'multi-label-text-classification'] | ['methodology', 'natural-language-processing'] | [ 4.89078790e-01 1.10870205e-01 -6.35751486e-01 -7.01226950e-01
-8.46858382e-01 -4.30116326e-01 6.99087858e-01 2.21859485e-01
-3.94143134e-01 4.08404946e-01 3.36843669e-01 -1.11337304e-01
1.93669543e-01 -5.43974400e-01 -1.04978383e-01 -8.38538587e-01
5.99590242e-01 4.32754040e-01 -1.57486275e-01 1.86705917e-01
3.87386590e-01 1.57813400e-01 -1.46026814e+00 4.63186204e-01
5.45584679e-01 9.94096875e-01 -3.51595273e-03 4.35523599e-01
-4.75481629e-01 1.08838236e+00 -2.27367818e-01 -3.03167373e-01
-1.17369413e-01 -5.46216011e-01 -1.02938855e+00 1.95080191e-01
5.36744833e-01 -2.56623745e-01 -1.89185470e-01 9.92489994e-01
3.32783818e-01 1.95659444e-01 1.12395215e+00 -1.27500856e+00
-7.03341126e-01 4.87350255e-01 -7.68402100e-01 -1.13355167e-01
1.47073358e-01 -3.93936604e-01 1.66123033e+00 -1.06200743e+00
4.05082971e-01 1.30688608e+00 4.94888067e-01 5.63715637e-01
-1.43024564e+00 -6.55716002e-01 4.78623033e-01 4.73396517e-02
-1.27017784e+00 -2.22542077e-01 9.62639630e-01 -5.05212545e-01
8.33945572e-01 1.81619376e-01 2.73878366e-01 1.02404571e+00
3.11040580e-02 1.20842242e+00 1.10946298e+00 -8.75705838e-01
-1.74520180e-01 3.17441255e-01 6.06596649e-01 8.42183232e-01
-6.25613108e-02 -2.09159598e-01 -4.12036866e-01 -1.71074301e-01
1.83883861e-01 3.41926783e-01 -2.30859429e-01 -3.96906585e-01
-1.03747940e+00 1.20832932e+00 4.68117595e-01 4.48994339e-01
-1.07813783e-01 4.07057762e-01 6.12680614e-01 2.45716467e-01
1.05512130e+00 4.66120183e-01 -4.38390911e-01 2.14692339e-01
-9.40574586e-01 5.22896722e-02 5.90712786e-01 8.19698215e-01
8.88441443e-01 -2.75004268e-01 -3.84428054e-01 1.12864292e+00
6.71040952e-01 1.03080198e-02 4.85621035e-01 -4.87905055e-01
5.24922073e-01 8.96956623e-01 -2.29402721e-01 -8.59758973e-01
-6.53684258e-01 -2.68659651e-01 -6.15616679e-01 4.54795770e-02
7.27594271e-02 1.86729759e-01 -9.94758606e-01 1.62162292e+00
2.14609891e-01 6.41911924e-02 4.70529199e-02 6.55434787e-01
7.58896112e-01 7.06037164e-01 4.61810052e-01 9.70120952e-02
1.58140659e+00 -1.59111321e+00 -9.63548183e-01 -3.74811083e-01
1.37699628e+00 -6.60664797e-01 1.04967332e+00 -1.17580056e-01
-6.17163837e-01 -3.84754539e-01 -8.80307674e-01 -5.28084576e-01
-6.41281426e-01 4.49359924e-01 5.44137418e-01 5.32735705e-01
-9.00099456e-01 2.66011566e-01 -6.24681473e-01 -4.59016263e-01
4.94856149e-01 3.14159632e-01 -3.22663128e-01 -3.17021847e-01
-1.09719002e+00 7.41352499e-01 6.07086062e-01 -1.74678147e-01
-4.19032902e-01 -2.57942110e-01 -1.08220887e+00 3.14740986e-01
2.77076364e-01 -4.36653465e-01 1.16048539e+00 -9.19930279e-01
-1.30365038e+00 1.12434447e+00 -3.02851379e-01 -9.54437852e-02
-3.82244028e-02 -4.04107049e-02 -1.10304981e-01 2.15105817e-01
2.72111833e-01 1.08227074e+00 8.41975272e-01 -1.29149401e+00
-8.17293167e-01 -3.19169700e-01 1.10002637e-01 3.38259548e-01
-9.65482891e-01 2.25087807e-01 -3.55469465e-01 -8.77438903e-01
2.50505745e-01 -1.13259530e+00 2.68045943e-02 9.28005129e-02
-4.52889979e-01 -9.07163858e-01 9.88097072e-01 -3.81836325e-01
1.18063736e+00 -2.24500918e+00 1.96923450e-01 -1.70890883e-01
4.85236973e-01 -1.32342568e-02 -3.06025326e-01 5.95731735e-01
-2.09683001e-01 2.40326956e-01 -3.87713090e-02 -1.02142966e+00
2.33000040e-01 1.09295346e-01 -5.10631740e-01 6.05723143e-01
2.63383567e-01 1.10159290e+00 -8.38597715e-01 -7.64120042e-01
9.52317342e-02 4.47604150e-01 -5.59826970e-01 1.93034634e-01
-2.03003183e-01 3.10851157e-01 -7.19040990e-01 7.08251476e-01
3.64380211e-01 -5.26925802e-01 2.76931316e-01 -2.35948145e-01
1.32048070e-01 4.05273765e-01 -6.30323291e-01 1.57192683e+00
-5.25880098e-01 7.62448967e-01 -1.67569667e-01 -1.19469857e+00
7.24399865e-01 4.02068585e-01 6.34281099e-01 -4.35799032e-01
4.66004372e-01 1.00172192e-01 -4.21584576e-01 -3.62722576e-01
4.64258373e-01 -2.99483508e-01 -1.17310956e-01 9.95212018e-01
6.48155902e-03 3.08528125e-01 4.64171963e-03 2.29217172e-01
8.21296751e-01 6.47823960e-02 1.74447224e-01 -2.41810635e-01
4.58042681e-01 -2.56507754e-01 2.11715043e-01 4.63244170e-01
-2.02587888e-01 4.34450209e-01 5.38422465e-01 -3.84960890e-01
-6.39419675e-01 -3.97000015e-01 -2.04255760e-01 1.73274755e+00
5.80914877e-02 -6.68292165e-01 -5.67887783e-01 -1.29602158e+00
-7.28000849e-02 6.65212572e-01 -9.69779730e-01 -3.18313718e-01
-3.75747174e-01 -5.29195905e-01 4.84382629e-01 7.36078560e-01
2.60542512e-01 -1.08813190e+00 -1.92566738e-01 2.28367776e-01
-2.85080284e-01 -9.33815241e-01 -6.35300517e-01 6.57333970e-01
-7.57821023e-01 -8.67265999e-01 -7.62982488e-01 -1.20664918e+00
9.21251297e-01 4.63239759e-01 9.27715182e-01 2.70057291e-01
-3.66404466e-02 3.57509077e-01 -6.25956595e-01 -1.54161096e-01
-1.40193701e-01 5.29969573e-01 -3.32368672e-01 3.00277740e-01
6.46138132e-01 -1.32837191e-01 -3.80175561e-01 2.28039294e-01
-1.09877658e+00 3.21456641e-01 3.47346991e-01 1.26950383e+00
3.97784710e-01 1.87055208e-02 7.83347309e-01 -1.13612354e+00
4.71257210e-01 -4.55636799e-01 -2.59745568e-01 3.30534279e-01
-8.47333312e-01 2.50477314e-01 6.98896766e-01 -3.65557194e-01
-8.51616800e-01 2.36198694e-01 -1.70221344e-01 -1.94413811e-01
-1.51125118e-01 6.66610539e-01 -2.82820556e-02 -5.24421707e-02
2.02567026e-01 1.09648667e-01 -2.33623058e-01 -6.51010275e-01
4.65910226e-01 1.02121425e+00 -1.61022112e-01 -5.51431417e-01
4.63374197e-01 4.56568450e-01 -5.83306560e-03 -3.53395849e-01
-1.49708652e+00 -8.53559732e-01 -6.49241626e-01 3.74791250e-02
1.09707105e+00 -8.16958487e-01 -4.11189586e-01 2.78786033e-01
-1.00683665e+00 -1.42593578e-01 1.99176129e-02 4.04470474e-01
-4.62345362e-01 4.62867022e-01 -8.82945657e-01 -6.26749814e-01
-2.33190209e-01 -1.40567231e+00 1.56657445e+00 -8.32990259e-02
-1.88967735e-01 -1.45829415e+00 1.31417587e-01 4.44307625e-01
1.84358656e-01 -5.59508614e-02 1.20235872e+00 -9.30238485e-01
-2.17828497e-01 -5.13218522e-01 -7.13716030e-01 1.65876284e-01
2.63890445e-01 -4.50440973e-01 -1.30535340e+00 -5.61704516e-01
-1.42087772e-01 -8.18469167e-01 1.34038043e+00 1.09335415e-01
1.18991518e+00 -1.49117127e-01 -7.81908631e-01 4.48736280e-01
1.23314691e+00 -1.02993265e-01 2.82430977e-01 3.36831748e-01
1.14379013e+00 8.42945993e-01 6.28265321e-01 1.91318259e-01
5.71135700e-01 8.78746927e-01 3.22799206e-01 -3.20163786e-01
-1.40696228e-01 -1.95708856e-01 1.08520530e-01 8.15283179e-01
5.57902634e-01 -7.15119302e-01 -8.19515586e-01 3.42769980e-01
-1.90912747e+00 -5.53007305e-01 -7.90337026e-02 1.61049724e+00
8.73232722e-01 -8.64597782e-02 -1.87437996e-01 1.94647759e-01
7.15062797e-01 4.50174719e-01 -3.59735340e-01 -2.68437117e-01
2.40709007e-01 -4.59373593e-02 2.60030329e-01 3.05213839e-01
-1.53049231e+00 1.13822830e+00 6.21975565e+00 8.47650409e-01
-9.53221560e-01 2.32591644e-01 5.97700179e-01 1.15662299e-01
-2.01973632e-01 1.21704660e-01 -1.11837816e+00 3.40903372e-01
8.60512316e-01 2.33335659e-01 8.81632417e-02 8.46290171e-01
-2.75837421e-01 2.04386115e-01 -1.29536378e+00 8.90981853e-01
3.45181823e-01 -8.66840303e-01 1.12772577e-01 3.04732442e-01
7.86384344e-01 -3.30971032e-01 7.39240870e-02 6.78783119e-01
1.51506737e-01 -1.01233244e+00 5.25855780e-01 2.59667277e-01
1.05625999e+00 -6.66233599e-01 8.20332348e-01 3.36882234e-01
-1.40288270e+00 -1.05709195e-01 -3.86088699e-01 1.94659550e-02
-1.79378223e-02 3.15393925e-01 -6.23199761e-01 4.03190583e-01
2.27654323e-01 1.12792540e+00 -7.46444166e-01 5.84809780e-01
-3.26152861e-01 6.46170557e-01 3.00010182e-02 -9.75003652e-03
6.19336963e-01 -5.40192844e-03 -1.15895517e-01 1.24857187e+00
2.13189349e-02 -1.62562966e-01 6.47765696e-01 6.32166564e-01
-5.29744625e-01 3.91293108e-01 -5.81603467e-01 -3.44466805e-01
4.81771789e-02 1.45688212e+00 -1.08968627e+00 -4.54095274e-01
-7.68618345e-01 1.06159496e+00 7.73717344e-01 3.14424247e-01
-7.58096755e-01 -4.18190658e-01 3.16493750e-01 -2.71102905e-01
2.27566913e-01 -1.07147247e-02 -2.01238766e-01 -1.25274038e+00
-1.41960621e-01 -5.63282371e-01 5.34224570e-01 -5.69875836e-01
-1.39846718e+00 4.52856213e-01 -2.24155620e-01 -1.01736844e+00
-4.38760594e-02 -6.94303036e-01 -1.92105025e-01 8.26608300e-01
-1.81752253e+00 -1.69362712e+00 -1.24184519e-01 2.03197032e-01
8.14745069e-01 -6.99639842e-02 1.07092631e+00 4.10584599e-01
-6.03418231e-01 9.74673092e-01 3.27305496e-01 2.73812354e-01
1.02815282e+00 -1.36170816e+00 1.84672013e-01 3.48938704e-01
4.86521542e-01 4.16126758e-01 8.22753236e-02 -5.16101539e-01
-9.41858292e-01 -1.21459305e+00 1.14366460e+00 -5.24985731e-01
5.78765154e-01 -5.96354246e-01 -9.71683919e-01 8.50306690e-01
2.87601501e-01 1.06113896e-01 1.11356676e+00 3.22485000e-01
-6.96011186e-01 2.83282816e-01 -7.32456625e-01 5.66745222e-01
7.64657557e-01 -8.44864428e-01 -4.52218294e-01 6.03466630e-01
7.21660912e-01 1.22819282e-01 -7.39479005e-01 1.49894908e-01
4.40215349e-01 -3.76716077e-01 7.64174938e-01 -6.37052417e-01
4.73840415e-01 -1.44565046e-01 -2.18197033e-01 -1.14019465e+00
-4.43691790e-01 2.87883747e-02 -5.48634492e-02 1.47454965e+00
2.92466700e-01 -7.44893074e-01 7.41692603e-01 4.52833593e-01
-2.03609213e-01 -9.70171571e-01 -6.40941739e-01 -5.03597856e-01
2.45851129e-01 -1.51283070e-01 3.70751143e-01 1.40902078e+00
1.08683497e-01 8.83408487e-01 -4.65542197e-01 -1.51826605e-01
4.92408842e-01 3.03274304e-01 3.67868096e-01 -1.55233014e+00
-7.12233186e-02 -6.69594884e-01 -2.20735922e-01 -1.33680618e+00
9.95395005e-01 -1.43657911e+00 1.59287095e-01 -1.71321321e+00
6.35611832e-01 -5.70616841e-01 -6.38399124e-01 9.10611987e-01
-5.84991693e-01 5.40545642e-01 1.35149226e-01 4.30530578e-01
-9.10608649e-01 7.61588573e-01 1.14210641e+00 -4.88670737e-01
1.85158521e-01 -3.23641002e-01 -8.69013906e-01 5.27115047e-01
7.45560408e-01 -9.56078589e-01 -3.20097357e-01 -4.16434616e-01
2.20679179e-01 -1.59450918e-01 7.64056742e-02 -6.35942519e-01
1.86367467e-01 3.63650285e-02 1.29411042e-01 -6.42928302e-01
3.95946771e-01 -9.88178611e-01 -5.73385000e-01 1.58496052e-01
-1.00318408e+00 -3.21546733e-01 -1.00345924e-01 7.25575686e-01
-4.21221375e-01 -6.89230204e-01 7.70209849e-01 2.32475951e-01
-4.40996885e-01 2.62917936e-01 -5.41967034e-01 -1.30562454e-01
9.01743531e-01 7.79199898e-02 -2.80036330e-01 -2.24738806e-01
-5.66586375e-01 4.19247985e-01 5.00847399e-01 4.66520488e-01
3.20114106e-01 -1.62687111e+00 -5.16396999e-01 1.89250514e-01
6.52120829e-01 -3.67173016e-01 -2.64131185e-02 6.98269725e-01
-9.29009765e-02 7.30703950e-01 2.87592977e-01 -4.80053425e-01
-1.27055275e+00 7.46552467e-01 1.51127726e-01 -7.49339223e-01
-6.31047845e-01 7.25113153e-01 6.71583056e-01 -5.41193962e-01
3.39566529e-01 -1.40382692e-01 -5.61349630e-01 4.06050801e-01
5.64689398e-01 1.08144078e-02 -1.32254496e-01 -8.23895574e-01
-1.92947134e-01 9.23892796e-01 -5.39121747e-01 1.21208623e-01
9.90182757e-01 -3.00710142e-01 -1.57482833e-01 6.24488056e-01
1.80210614e+00 -2.58711010e-01 -1.12175786e+00 -5.03230453e-01
3.32696378e-01 -3.53873223e-01 3.85913402e-01 -5.45466483e-01
-1.10956323e+00 1.15909135e+00 5.02201438e-01 2.85905868e-01
1.00406873e+00 9.98935699e-02 8.44457507e-01 5.10929048e-01
-1.23821292e-02 -9.31131840e-01 4.01754647e-01 7.00464964e-01
4.43407387e-01 -1.56911743e+00 -1.41080413e-02 -4.98577237e-01
-4.23050046e-01 1.16183817e+00 5.42015970e-01 1.96188584e-01
6.76310658e-01 -4.29057181e-02 1.36983156e-01 -4.12419885e-01
-7.17881978e-01 -3.90426725e-01 4.64913845e-01 1.58063397e-01
8.95205915e-01 -1.09444030e-01 -2.76438355e-01 3.57534945e-01
4.92952883e-01 -3.99589539e-01 8.58783647e-02 1.18990231e+00
-5.06225526e-01 -1.41966772e+00 -1.14953600e-01 5.93200028e-01
-6.09954715e-01 -2.27671087e-01 -4.85689700e-01 5.10622025e-01
-1.50901780e-01 9.92979705e-01 -7.42172748e-02 -2.82562047e-01
-8.25889856e-02 6.26378536e-01 1.49186909e-01 -1.05582809e+00
-5.97172618e-01 2.56072342e-01 -3.47970374e-04 -1.65200576e-01
-5.63503683e-01 -3.93740475e-01 -1.17382455e+00 9.00194645e-02
-7.34366179e-01 5.44064306e-02 7.02763796e-01 9.73873019e-01
2.16872007e-01 6.70768797e-01 6.87836766e-01 -8.37109685e-01
-3.43990594e-01 -1.14833415e+00 -6.19633913e-01 4.75270450e-01
1.68514937e-01 -9.50017154e-01 -4.94180948e-01 4.40900661e-02] | [9.70422649383545, 4.448497295379639] |
757a0473-3223-4785-9b98-45eff70d197a | learning-implicit-text-generation-via-feature | 2005.03588 | null | https://arxiv.org/abs/2005.03588v2 | https://arxiv.org/pdf/2005.03588v2.pdf | Learning Implicit Text Generation via Feature Matching | Generative feature matching network (GFMN) is an approach for training implicit generative models for images by performing moment matching on features from pre-trained neural networks. In this paper, we present new GFMN formulations that are effective for sequential data. Our experimental results show the effectiveness of the proposed method, SeqGFMN, for three distinct generation tasks in English: unconditional text generation, class-conditional text generation, and unsupervised text style transfer. SeqGFMN is stable to train and outperforms various adversarial approaches for text generation and text style transfer. | ['Cicero Nogueira dos santos', 'Inkit Padhi', 'Youssef Mroueh', 'Pierre Dognin', 'Payel Das', 'Vijil Chenthamarakshan', 'Ke Bai'] | 2020-05-07 | learning-implicit-text-generation-via-feature-1 | https://aclanthology.org/2020.acl-main.354 | https://aclanthology.org/2020.acl-main.354.pdf | acl-2020-6 | ['conditional-text-generation'] | ['natural-language-processing'] | [ 6.93800032e-01 1.71191618e-01 4.29239981e-02 -5.46570301e-01
-8.42619598e-01 -3.32905382e-01 1.27373159e+00 -6.54576600e-01
-1.27684161e-01 1.01472831e+00 3.95111710e-01 -9.49937403e-02
4.15245742e-01 -8.93195152e-01 -6.54134870e-01 -5.91635525e-01
5.80528796e-01 7.40948021e-01 -2.26778522e-01 -2.79608369e-01
3.08897167e-01 4.81358051e-01 -1.00847983e+00 3.84873569e-01
8.87857616e-01 7.11377084e-01 2.36590892e-01 9.51868653e-01
-1.33688942e-01 8.56596649e-01 -8.06299746e-01 -9.15169597e-01
1.30122781e-01 -8.77533793e-01 -9.05649722e-01 1.63795605e-01
6.46855831e-01 -5.54323792e-01 -5.46087980e-01 8.51449311e-01
8.20266008e-01 2.43098214e-01 1.43883753e+00 -1.37379277e+00
-1.47469211e+00 4.25243735e-01 -4.24077250e-02 -5.19195870e-02
4.07920837e-01 -7.21136108e-03 7.61485159e-01 -1.31267142e+00
9.09241557e-01 1.50218451e+00 6.12586021e-01 1.19290745e+00
-1.28660083e+00 -6.96549714e-01 -3.19874108e-01 -3.42279524e-01
-9.93203044e-01 -5.16637504e-01 7.27154076e-01 -5.02521574e-01
9.14069176e-01 5.28231375e-02 2.77356207e-01 1.72593188e+00
6.56207621e-01 1.01814604e+00 1.22101581e+00 -5.93284488e-01
-5.20626642e-03 1.46025583e-01 -4.86687511e-01 6.56739235e-01
8.19500815e-03 4.41321045e-01 -7.69484878e-01 -8.00222456e-02
1.27048814e+00 -1.65188044e-01 5.80868423e-02 -4.21612188e-02
-1.16009891e+00 1.35814178e+00 2.51487434e-01 -1.43250292e-02
-1.55190036e-01 3.20531756e-01 5.04339516e-01 5.57741761e-01
8.86476159e-01 2.03250125e-01 7.15320557e-02 1.29763737e-01
-1.07932675e+00 3.34151983e-01 8.40601861e-01 1.59783244e+00
7.86379099e-01 6.98561668e-01 -6.21541381e-01 7.00847566e-01
2.00077891e-01 9.77224469e-01 9.77486789e-01 -4.35876042e-01
4.82780248e-01 1.03672706e-01 -1.46471456e-01 -7.77896285e-01
2.23577306e-01 1.04267851e-01 -1.06273568e+00 1.12116188e-01
-2.87726596e-02 -5.51200807e-01 -1.36717141e+00 1.67680585e+00
-1.21230274e-01 -6.44440278e-02 4.31381941e-01 3.11329573e-01
1.21770453e+00 7.93731809e-01 3.60186040e-01 2.15002403e-01
6.90428436e-01 -9.68621433e-01 -7.39704609e-01 -3.28472137e-01
2.55883783e-01 -1.25088787e+00 9.40401971e-01 -1.97130591e-01
-9.76410091e-01 -8.20142150e-01 -6.45171165e-01 -4.23807576e-02
-5.19807339e-01 4.16340560e-01 6.73867464e-01 6.35842085e-01
-1.31577897e+00 7.63710141e-01 -6.49657190e-01 -4.55522448e-01
4.32786494e-01 4.03671443e-01 -3.31916600e-01 1.03920899e-01
-1.27858388e+00 7.93377042e-01 6.65916443e-01 8.56667850e-03
-1.18920755e+00 -4.00786251e-01 -1.06294000e+00 -2.28659123e-01
-3.97304267e-01 -1.15805936e+00 1.24647868e+00 -1.52002978e+00
-1.85084629e+00 1.11128581e+00 -2.42286444e-01 -5.51327288e-01
7.02295482e-01 -3.41444999e-01 -2.84083903e-01 1.54046178e-01
2.17068136e-01 1.15392733e+00 1.69692171e+00 -1.18970382e+00
-2.05222353e-01 1.15382247e-01 -4.26280171e-01 3.60091060e-01
-2.47458324e-01 2.33099461e-02 2.70533293e-01 -1.40384746e+00
-3.16052377e-01 -9.28370118e-01 -1.13328084e-01 -3.00275505e-01
-6.46095395e-01 -3.78501534e-01 1.02563536e+00 -8.07173848e-01
6.32124186e-01 -1.79174614e+00 2.90832579e-01 -7.05777779e-02
-2.01264426e-01 2.37825111e-01 -2.89926291e-01 6.00073099e-01
-1.27003014e-01 7.56654292e-02 -1.78472757e-01 -6.47117734e-01
2.33011767e-01 4.00284231e-01 -7.63526022e-01 -4.40718494e-02
5.28091609e-01 1.49283755e+00 -7.21902370e-01 -6.82259023e-01
1.70363650e-01 5.25122881e-01 -4.70714331e-01 5.14302671e-01
-2.78641045e-01 3.64998996e-01 -2.84209281e-01 4.97487783e-01
3.54517132e-01 -1.14365928e-01 -2.00494722e-01 -4.10968363e-02
3.83847952e-01 5.66293709e-02 -6.67266428e-01 1.85947704e+00
-3.33516985e-01 8.09052408e-01 -7.04517424e-01 -7.72608280e-01
1.05308330e+00 5.06222785e-01 -1.45617887e-01 -2.91014999e-01
2.80872166e-01 2.23565191e-01 -5.67039311e-01 -2.94401735e-01
7.81884074e-01 -5.18474519e-01 -2.18252420e-01 6.61562204e-01
8.98657918e-01 -6.31313145e-01 2.20442265e-01 4.07706171e-01
4.69254732e-01 4.89303827e-01 9.62036103e-02 -2.88276017e-01
4.68426943e-01 -4.35108364e-01 -5.01109054e-03 8.84252906e-01
2.15541795e-01 8.35684776e-01 6.45882636e-02 -3.63971293e-01
-1.28630495e+00 -1.53753221e+00 2.02066079e-01 1.09605598e+00
-2.14279518e-01 -1.12860411e-01 -1.00911236e+00 -9.96923327e-01
-2.14245021e-01 8.00164938e-01 -8.98513019e-01 -3.20349425e-01
-5.35270870e-01 -5.58428705e-01 6.57690823e-01 9.00041640e-01
6.81422710e-01 -1.69321966e+00 1.39106840e-01 1.87020496e-01
-3.46235901e-01 -9.45988536e-01 -9.88684475e-01 -1.41553074e-01
-9.97600138e-01 -5.71786284e-01 -1.14150441e+00 -1.17631137e+00
1.04164398e+00 -2.80824184e-01 1.31739199e+00 -2.38441229e-01
-2.82820106e-01 4.94862109e-01 -3.76579434e-01 -5.45755804e-01
-1.05518115e+00 3.32912534e-01 -1.25127971e-01 -2.86280666e-03
1.08689526e-02 -4.59561795e-01 -3.62501591e-01 1.28685966e-01
-1.15817702e+00 3.43665093e-01 7.40956783e-01 1.26232183e+00
4.95765746e-01 -5.49723685e-01 8.72359872e-01 -1.09732556e+00
9.54290926e-01 -2.22134721e-02 -2.43524626e-01 2.84175098e-01
-6.37705147e-01 2.38628536e-01 6.83772087e-01 -6.58066690e-01
-1.49340296e+00 1.45184277e-02 -1.99914336e-01 -4.09899384e-01
-3.55151147e-01 1.66973457e-01 -4.48475666e-02 -1.55273780e-01
7.62904227e-01 7.88643301e-01 -7.07060099e-02 -1.55079290e-01
5.65093815e-01 5.32990754e-01 5.68043828e-01 -6.47611320e-01
1.12806094e+00 3.07329983e-01 -5.72403567e-03 -5.96304595e-01
-9.02087271e-01 2.02517733e-01 -8.14890623e-01 -3.10992673e-02
1.05674386e+00 -9.07811999e-01 2.14850098e-01 8.34750712e-01
-1.21399808e+00 -5.13750196e-01 -5.80196440e-01 2.89454818e-01
-1.17973006e+00 1.95055082e-01 -8.07465792e-01 -4.23410118e-01
-1.07403708e+00 -7.05356956e-01 1.28136683e+00 2.63664305e-01
-2.46659204e-01 -1.67604983e+00 2.00382456e-01 7.20978826e-02
6.27987385e-01 4.00524288e-01 6.86783433e-01 -7.66796291e-01
-4.52446550e-01 -2.14657456e-01 -8.78402218e-02 7.29462147e-01
4.73612696e-01 -7.08331987e-02 -8.78368080e-01 -5.26509583e-01
-2.49124944e-01 -6.67217314e-01 8.15593541e-01 3.32956821e-01
7.94813931e-01 -5.57152629e-01 -1.80729464e-01 7.41381764e-01
1.31812656e+00 7.99982399e-02 1.00286067e+00 -2.24224515e-02
7.36285567e-01 1.44224152e-01 4.59210485e-01 1.68150991e-01
-1.22737795e-01 2.33084977e-01 -1.82469282e-02 -3.97715181e-01
-4.16409642e-01 -5.27116537e-01 6.43363953e-01 7.07985699e-01
-1.72224224e-01 -5.82786560e-01 -3.04798067e-01 4.98996586e-01
-1.54890859e+00 -1.24180639e+00 2.25007176e-01 1.71823144e+00
1.08002877e+00 8.88892263e-02 -1.94023699e-01 -3.20974737e-01
9.16347980e-01 -3.53350341e-02 -3.36836636e-01 -6.12707734e-01
-1.80506155e-01 9.01978254e-01 1.55652955e-01 3.28424871e-01
-1.29936290e+00 1.30557072e+00 7.58208418e+00 1.14066827e+00
-1.11525095e+00 1.19402975e-01 6.53066993e-01 2.80136973e-01
-2.70999849e-01 -1.01123638e-01 -9.76213217e-01 4.14195389e-01
8.36233377e-01 -6.23164356e-01 2.05107868e-01 7.98754632e-01
-3.36796612e-01 4.20284212e-01 -1.32115149e+00 7.62873828e-01
5.94836891e-01 -1.34913909e+00 9.27612305e-01 -2.71252275e-01
1.33992863e+00 -1.86142281e-01 4.02380973e-01 3.22675437e-01
6.43496692e-01 -1.21990478e+00 6.47411287e-01 6.38960660e-01
1.23676097e+00 -8.52237284e-01 6.23780549e-01 3.68770063e-02
-1.02497065e+00 5.33677518e-01 -5.00436127e-01 4.35137838e-01
2.77081400e-01 1.11536093e-01 -9.15174186e-01 7.52816141e-01
1.04758382e-01 7.35670447e-01 -7.80312300e-01 5.16937554e-01
-5.00857532e-01 5.51889718e-01 1.33656681e-01 7.64377136e-03
2.47321382e-01 1.02863619e-02 4.61114913e-01 1.34826005e+00
6.00155234e-01 -3.56183439e-01 -7.51483766e-03 1.09099257e+00
-3.59160691e-01 1.57331392e-01 -8.75216544e-01 -2.47664779e-01
-1.08672492e-01 1.25393403e+00 -6.27857268e-01 -5.71600318e-01
-1.08862132e-01 1.61163402e+00 1.27459183e-01 4.69477981e-01
-8.51234436e-01 -7.44843245e-01 1.38806209e-01 -1.27521887e-01
2.80552953e-01 1.58024265e-03 1.65825859e-02 -1.28568780e+00
-4.50861365e-01 -6.80850089e-01 1.69489712e-01 -1.10000789e+00
-1.72156000e+00 9.58009601e-01 3.83402295e-02 -1.19544649e+00
-9.68368828e-01 -6.72957838e-01 -9.80250418e-01 1.14762986e+00
-1.14920056e+00 -1.71443450e+00 -2.28413284e-01 1.09563601e+00
8.69201601e-01 -5.94052792e-01 1.14672887e+00 7.07533062e-02
-4.75766547e-02 9.41364765e-01 1.09310605e-01 5.24138153e-01
9.80068624e-01 -1.34379303e+00 9.84005988e-01 9.22386348e-01
2.78338194e-01 5.54028332e-01 6.13188386e-01 -9.37570453e-01
-8.00845325e-01 -1.61480463e+00 1.05665052e+00 -3.98347765e-01
3.88408184e-01 -5.41474223e-01 -4.98745441e-01 1.25518131e+00
7.40323126e-01 -3.30019087e-01 7.04719841e-01 -5.60197711e-01
-1.55832425e-01 3.04118067e-01 -1.25435042e+00 6.43492281e-01
9.95659411e-01 -6.26357973e-01 -7.30145037e-01 5.18977463e-01
7.11481869e-01 -5.39605916e-01 -7.57678270e-01 1.54100239e-01
3.07730943e-01 -4.85356629e-01 9.16961372e-01 -8.01929772e-01
7.67988861e-01 3.02455127e-01 -1.76371988e-02 -1.48769057e+00
-1.42170027e-01 -1.02051151e+00 4.76022474e-02 1.46887481e+00
3.72327775e-01 -7.56411910e-01 6.99214995e-01 1.21590912e-01
-1.86705247e-01 -2.26270631e-01 -6.04083419e-01 -8.41762483e-01
4.59802449e-01 1.97582871e-01 4.33366776e-01 9.52359557e-01
-6.25997663e-01 5.33898771e-01 -7.05829442e-01 -4.45796907e-01
6.51370108e-01 1.03462704e-01 8.51266086e-01 -9.98823404e-01
-1.59997910e-01 -1.67504400e-01 -4.05455977e-01 -9.65786934e-01
5.30219138e-01 -1.19339478e+00 7.89905116e-02 -1.61837530e+00
2.91361094e-01 -4.32695225e-02 1.64281800e-01 4.55354035e-01
-2.50586957e-01 5.56706905e-01 9.80381221e-02 2.58383881e-02
-1.72800153e-01 8.97619188e-01 1.52805650e+00 -1.59586713e-01
2.37234086e-01 1.46862224e-01 -4.13937300e-01 6.69914544e-01
9.45714712e-01 -4.71123844e-01 -6.25200927e-01 -3.24643016e-01
-9.42295566e-02 -1.33718744e-01 4.77195263e-01 -9.08297956e-01
-1.34180367e-01 -1.70395702e-01 9.50817823e-01 -7.52419233e-01
7.97717720e-02 -3.07551771e-01 2.83455461e-01 3.80052090e-01
-5.43365419e-01 8.20379406e-02 2.09026322e-01 3.81359667e-01
-2.13098273e-01 -4.92430151e-01 8.25660884e-01 -3.45557258e-02
-4.40773845e-01 5.42119443e-01 -5.19097984e-01 3.08579683e-01
6.92564249e-01 -1.75667986e-01 -2.43933886e-01 -5.82649112e-01
-1.15920734e+00 -3.09341341e-01 2.03051612e-01 5.66850543e-01
1.04266179e+00 -1.92591274e+00 -9.36638355e-01 4.98417735e-01
-1.55350909e-01 -3.26237768e-01 -3.18017602e-02 3.04105937e-01
-4.57443804e-01 4.39318687e-01 -4.85797673e-01 -3.54390383e-01
-1.17531097e+00 2.99055934e-01 3.19756776e-01 -5.25468588e-01
-4.13769245e-01 9.12366390e-01 6.13747478e-01 -5.88025331e-01
-1.96142852e-01 1.70701016e-02 1.07084446e-01 -2.91080266e-01
2.12730587e-01 1.87372655e-01 -2.32361183e-01 -6.77179873e-01
3.19260061e-01 3.76947016e-01 -2.38228977e-01 -3.28975439e-01
1.01624835e+00 -5.40418513e-02 -1.64868198e-02 1.99457765e-01
1.08352685e+00 -3.88932377e-01 -1.23563731e+00 -1.99471906e-01
-5.78157723e-01 -2.05217555e-01 -4.07146007e-01 -7.05472589e-01
-1.12488985e+00 8.93253744e-01 6.15389645e-01 -4.76451777e-02
1.12548757e+00 -1.34411946e-01 8.29046071e-01 3.92416805e-01
6.69135824e-02 -1.05036271e+00 6.29740715e-01 6.81132019e-01
1.24421120e+00 -1.09267545e+00 -2.25571096e-01 -1.87746584e-01
-5.47418296e-01 1.20293713e+00 7.88188159e-01 -4.97411907e-01
5.76427281e-01 4.57643196e-02 -5.56733347e-02 6.26180833e-03
-5.57376683e-01 2.24604309e-02 7.05041707e-01 8.55408669e-01
5.43766916e-01 -9.88453403e-02 -2.20223859e-01 1.58753857e-01
-6.06424630e-01 2.91125588e-02 2.41323501e-01 9.45842206e-01
-1.61948711e-01 -1.35123158e+00 -1.80131674e-01 4.76198047e-01
-5.50315142e-01 -5.92986226e-01 -6.46351814e-01 8.21052253e-01
-1.24313504e-01 4.53087896e-01 1.31129146e-01 -1.78834215e-01
6.87362179e-02 3.79440576e-01 9.99046862e-01 -7.09765375e-01
-5.18096030e-01 8.53701159e-02 -7.03296214e-02 -1.29705861e-01
-5.89916348e-01 -6.87757552e-01 -9.84402299e-01 -3.02601546e-01
-4.62544709e-01 -9.98454615e-02 4.85173672e-01 9.35754001e-01
2.16988578e-01 5.06334364e-01 7.18275428e-01 -1.10190058e+00
-5.15774250e-01 -1.34148252e+00 -4.64257330e-01 7.13782012e-01
-8.67018327e-02 -4.64207411e-01 -3.26322973e-01 8.75991762e-01] | [11.858040809631348, 9.346879959106445] |
a061597b-9f15-4654-ad47-ec85feee1e5c | bijective-mapping-network-for-shadow-removal | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhu_Bijective_Mapping_Network_for_Shadow_Removal_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhu_Bijective_Mapping_Network_for_Shadow_Removal_CVPR_2022_paper.pdf | Bijective Mapping Network for Shadow Removal | Shadow removal, which aims to restore the background in the shadow regions, is challenging due to the highly ill-posed nature. Most existing deep learning-based methods individually remove the shadow by only considering the content of the matched paired images, barely taking into account the auxiliary supervision of shadow generation on shadow removal procedure. In this work, we argue that shadow removal and generation are interrelated and could provide useful informative supervision for each other. Specifically, we propose a new Bijective Mapping Network (BMNet), which couples the learning procedures of shadow removal and shadow generation in a unified parameter-shared framework. With consistent two-way constraints and synchronous optimization of the two procedures, BMNet could effectively recover the underlying background contents during the forward shadow removal procedure. In addition, through statistic analysis on real-world datasets, we observe and verify that shadow appearances under different color spectrums are inconsistent. This motivates us to design a Shadow-Invariant Color Guidance Module (SICGM), which can explicitly utilize the learned shadow-invariant color information to guide network color restoration, thereby further reducing color-bias effects. Experiments on the representative ISTD, ISTD+ and SRD benchmarks show that our proposed network outperforms the state-of-the-art method [??] in de-shadowing performance, while only using its 0.25% network parameters and 6.25% floating point operations (FLOPs). | ['Zheng-Jun Zha', 'Qibin Sun', 'Feng Zhao', 'Xueyang Fu', 'Jie Huang', 'Yurui Zhu'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['shadow-removal'] | ['computer-vision'] | [ 3.15003067e-01 -2.44491383e-01 2.31366143e-01 -3.06936473e-01
-2.61475742e-01 -3.31540674e-01 2.63685584e-01 -5.27155876e-01
-1.96808532e-01 8.46566260e-01 9.05552134e-03 -5.34049094e-01
1.90908298e-01 -6.45170212e-01 -6.97433174e-01 -1.17282748e+00
1.64100841e-01 -2.32894178e-02 4.83558714e-01 -2.29233593e-01
4.96791154e-02 5.34510314e-01 -1.26141715e+00 -6.64450079e-02
1.34657943e+00 1.02038419e+00 4.66943055e-01 3.61033946e-01
-2.24348113e-01 7.34242558e-01 -6.84509933e-01 -2.92222351e-01
4.93474901e-01 -4.42157507e-01 -1.49164930e-01 1.61881000e-01
5.91949761e-01 -7.39111245e-01 -7.03191161e-01 1.01825631e+00
6.18859172e-01 2.61449456e-01 2.60711610e-01 -1.42796218e+00
-7.07227826e-01 -1.15600638e-02 -9.26679730e-01 1.36939391e-01
-1.07492432e-01 2.91093528e-01 6.85404181e-01 -8.16766322e-01
2.38304332e-01 1.24511099e+00 5.68807244e-01 4.01362181e-01
-1.08572257e+00 -8.47509623e-01 5.37376881e-01 2.68634617e-01
-1.28322589e+00 -3.54427725e-01 1.06947029e+00 4.40325104e-02
2.25906104e-01 3.54916066e-01 6.23927653e-01 1.00875306e+00
1.45531580e-01 9.40225542e-01 1.46233785e+00 -2.28847161e-01
1.22575916e-01 7.13372678e-02 1.02692731e-01 9.86827075e-01
5.18008649e-01 2.83308387e-01 -7.15964258e-01 6.13646284e-02
7.76572704e-01 1.23319864e-01 -7.45895982e-01 -4.10066873e-01
-9.60309625e-01 5.20196855e-01 6.52930796e-01 -3.02155584e-01
-7.39956945e-02 1.65145501e-01 1.26345247e-01 -2.65419465e-02
5.96514106e-01 -2.60923415e-01 -4.51312095e-01 3.44936490e-01
-7.27882624e-01 -8.17556828e-02 7.91446388e-01 9.52425122e-01
9.40110445e-01 4.07584459e-01 -3.60961497e-01 6.33932650e-01
4.45490003e-01 9.12012696e-01 -1.42659917e-01 -8.49514544e-01
3.67589712e-01 5.00594854e-01 2.64766246e-01 -1.27933931e+00
-2.39220142e-01 -5.61164379e-01 -1.22386444e+00 4.87344801e-01
3.73293191e-01 -1.57969564e-01 -1.21694088e+00 1.71509147e+00
5.51073909e-01 7.98976123e-01 9.32811126e-02 1.14899170e+00
8.66064250e-01 5.70132613e-01 -3.07262093e-01 -3.48717988e-01
1.13298857e+00 -1.11334419e+00 -6.86281741e-01 -4.30648804e-01
-4.09956276e-02 -8.09401035e-01 1.31096494e+00 3.30698788e-01
-7.32795238e-01 -4.63864088e-01 -1.27721167e+00 -4.14922163e-02
-1.43940508e-01 2.88531750e-01 8.50264192e-01 7.14105487e-01
-9.68432009e-01 3.65513027e-01 -7.14066446e-01 -7.00809062e-03
5.18604577e-01 2.07183346e-01 2.52901949e-02 -3.66084278e-01
-1.06106031e+00 6.07530475e-01 -5.68334274e-02 7.93571830e-01
-1.04269922e+00 -7.48786747e-01 -5.49314260e-01 -8.18156824e-02
8.16170394e-01 -5.44723213e-01 6.78032339e-01 -9.60293233e-01
-1.57457924e+00 5.59835136e-01 -2.27226570e-01 -9.72108468e-02
7.03146100e-01 -3.64489108e-01 -4.12006140e-01 3.29027735e-02
-8.29730853e-02 1.99523062e-01 9.08728957e-01 -1.96146286e+00
-4.86209035e-01 -3.50456238e-01 1.90501347e-01 4.69039798e-01
-2.81771570e-01 -3.39408636e-01 -1.08828008e+00 -8.00211132e-01
1.57564282e-01 -9.95255291e-01 -7.25825131e-02 5.66578507e-01
-6.63976252e-01 3.83931130e-01 1.16562319e+00 -9.54977393e-01
9.19591725e-01 -2.13613558e+00 -2.35102009e-02 2.46098369e-01
3.16810280e-01 3.31894457e-01 -1.47340208e-01 -1.18129782e-01
2.84491986e-01 -2.93728828e-01 -5.84312677e-01 -4.52716649e-01
4.18772846e-02 4.78721142e-01 -5.98496318e-01 7.60733247e-01
-1.57202765e-01 6.49189532e-01 -8.05978417e-01 -5.41335583e-01
2.85323471e-01 8.39986920e-01 -1.42546743e-01 4.09348011e-01
-7.53444210e-02 2.60117471e-01 -3.74993086e-01 9.16589081e-01
1.34270680e+00 -4.29203510e-02 2.58219212e-01 -6.10148728e-01
1.86818391e-02 -1.30598590e-01 -1.28701377e+00 1.44959748e+00
-5.61315238e-01 7.94556856e-01 5.73917747e-01 -6.43500507e-01
8.88569891e-01 -1.88155636e-01 2.98583269e-01 -1.03157651e+00
5.83449751e-02 8.82723480e-02 -2.57404417e-01 -1.89391136e-01
2.21617714e-01 9.50652733e-02 3.40420306e-01 4.27639633e-01
-5.29108524e-01 6.16248623e-02 -1.97785154e-01 2.65207469e-01
9.16092753e-01 2.73275584e-01 -1.99912205e-01 -2.77667433e-01
6.43543780e-01 -4.35981631e-01 9.51945245e-01 7.35188305e-01
-2.94114053e-01 5.52356184e-01 3.09764028e-01 -2.77169138e-01
-4.20515329e-01 -1.20016921e+00 4.87787575e-02 1.13262069e+00
9.72787082e-01 1.75590906e-02 -5.64724326e-01 -7.60440052e-01
1.26401424e-01 5.85355461e-01 -7.03988910e-01 -1.69184089e-01
-7.30605662e-01 -1.18529618e+00 3.42165679e-01 3.12527239e-01
8.53294611e-01 -8.47424090e-01 -5.11997998e-01 -9.96324122e-02
-1.85401887e-01 -1.06453764e+00 -4.87033218e-01 2.83434361e-01
-5.87213337e-01 -1.20484006e+00 -7.19745040e-01 -5.56931734e-01
8.00255299e-01 1.10556889e+00 1.04489744e+00 5.66264391e-01
-3.18224609e-01 2.11698472e-01 -2.18619794e-01 -2.66648144e-01
1.53440014e-02 -4.02742952e-01 -2.31846422e-01 2.83580124e-01
-1.47184432e-01 -8.61009955e-01 -9.27443385e-01 4.51535106e-01
-9.40595090e-01 5.02565503e-01 7.98535168e-01 9.02979732e-01
4.41284031e-01 1.15023859e-01 5.88642992e-02 -9.19276714e-01
3.51434201e-01 -2.54218876e-01 -8.59526813e-01 4.35343713e-01
-7.48171747e-01 -1.11405656e-01 7.05096662e-01 -1.88747644e-01
-1.70256436e+00 -7.55711347e-02 3.77109468e-01 -5.47939658e-01
5.00716418e-02 -9.07279551e-02 -6.83021307e-01 -4.12693143e-01
2.40087390e-01 4.96521473e-01 -2.22774893e-01 -4.18449372e-01
3.51989686e-01 1.48056611e-01 7.29197800e-01 -6.14340961e-01
1.39937520e+00 9.96395409e-01 1.86407655e-01 -6.08102143e-01
-1.01891410e+00 -2.14788482e-01 -4.27249104e-01 -3.79735440e-01
5.72400570e-01 -8.57901394e-01 -7.66482294e-01 8.42035234e-01
-1.12994838e+00 -7.61915326e-01 2.62140363e-01 -4.67584506e-02
-1.64078437e-02 5.83743870e-01 -4.27944452e-01 -8.28825653e-01
-3.28473240e-01 -9.87272382e-01 1.04827237e+00 4.41143274e-01
7.47440577e-01 -9.28117335e-01 -2.78839499e-01 3.05079430e-01
3.70872289e-01 3.25061142e-01 7.51334250e-01 8.43711793e-02
-1.06597579e+00 3.35974276e-01 -9.37026739e-01 4.58831042e-01
3.58881265e-01 -1.28325820e-01 -1.11843741e+00 -2.92153358e-01
-8.36501736e-03 1.00267917e-01 1.26381719e+00 3.08041722e-01
1.32649076e+00 -1.51164711e-01 -2.40249529e-01 1.09335268e+00
1.63121796e+00 1.56619623e-01 6.73010767e-01 3.39565516e-01
1.19628596e+00 4.96952951e-01 7.19111204e-01 4.62813705e-01
4.01128680e-01 5.57596803e-01 6.53351367e-01 -6.96398020e-01
-6.37278020e-01 1.12613320e-01 3.99039686e-01 5.22851467e-01
-1.26649803e-02 -5.00954449e-01 -6.29798293e-01 1.11361161e-01
-1.88174784e+00 -5.60919940e-01 -3.30690980e-01 2.02199793e+00
8.43410313e-01 1.64891943e-01 -4.61632311e-01 -9.15104002e-02
7.04060435e-01 5.84073663e-01 -8.35476458e-01 3.18859458e-01
-4.71293092e-01 1.33253187e-02 8.04486811e-01 7.01917946e-01
-9.44659054e-01 9.66357052e-01 5.17749977e+00 7.85381019e-01
-1.09045649e+00 -1.65917818e-02 6.75324798e-01 3.28431614e-02
-4.81735349e-01 8.68252814e-02 -5.18114686e-01 6.41748786e-01
1.90680847e-01 2.85500914e-01 8.26838195e-01 3.24035972e-01
3.90132517e-01 -4.86831874e-01 -5.75150132e-01 8.34515929e-01
3.56848508e-01 -1.21759129e+00 -1.41085356e-01 -1.61326751e-02
8.06767166e-01 -1.80741176e-01 3.73029448e-02 5.08170575e-02
3.97205949e-01 -6.56785905e-01 6.26319766e-01 7.36868024e-01
6.31853640e-01 -6.56328619e-01 7.08016336e-01 -1.19901642e-01
-1.33967388e+00 -3.75598036e-02 -2.04615831e-01 1.79220557e-01
2.07790941e-01 1.08020747e+00 -5.87189972e-01 7.89204836e-01
9.01641428e-01 6.11998022e-01 -6.43246770e-01 8.85855436e-01
-4.76207286e-01 6.13241315e-01 -1.69843346e-01 3.59839737e-01
-2.43610889e-02 -5.61388910e-01 4.71467078e-01 1.14676392e+00
1.14029385e-01 1.75098732e-01 2.30868578e-01 8.51325572e-01
-2.65279040e-02 -3.88559431e-01 -7.84006417e-02 5.55739403e-01
5.15280545e-01 1.31426442e+00 -9.42265451e-01 -2.89658487e-01
-4.18812186e-01 1.34138298e+00 1.37200430e-02 8.67272437e-01
-1.21905649e+00 -3.08481187e-01 7.97558784e-01 -2.00698808e-01
1.98298067e-01 -2.90611386e-01 -4.69592214e-01 -1.05829728e+00
3.09042156e-01 -8.51069808e-01 6.98749050e-02 -9.58172560e-01
-1.18624568e+00 4.17717546e-01 -2.80090332e-01 -1.10260129e+00
6.83655143e-01 -6.05235279e-01 -8.05532157e-01 8.25172365e-01
-2.26533484e+00 -1.28406358e+00 -1.11486542e+00 8.25736523e-01
3.18047971e-01 -3.55900936e-02 2.95569509e-01 4.76916552e-01
-9.99710917e-01 4.88823801e-01 2.12547854e-01 -2.11374536e-02
1.00209165e+00 -1.27529311e+00 2.44176716e-01 1.22018445e+00
-2.75558144e-01 4.60467249e-01 7.42618680e-01 -6.38449609e-01
-1.71148038e+00 -1.23183715e+00 4.50174324e-02 9.55034271e-02
5.12234330e-01 -3.63835096e-01 -1.06249678e+00 1.37587041e-01
8.97166282e-02 3.40412259e-01 2.81273156e-01 -2.11593494e-01
-4.74238098e-01 -7.85589337e-01 -9.00866449e-01 8.34548116e-01
1.31814194e+00 -4.65768784e-01 -4.17924225e-02 4.79345798e-01
9.00816798e-01 -5.06922901e-01 -1.36667028e-01 3.93235892e-01
3.25617820e-01 -1.38177133e+00 1.30676258e+00 5.70079833e-02
1.90369427e-01 -7.23265648e-01 -2.63040245e-01 -1.09665668e+00
6.12952374e-02 -7.25119948e-01 -3.92731309e-01 1.37509406e+00
-6.73355386e-02 -7.86929786e-01 7.64666378e-01 4.33213174e-01
-5.29208183e-01 -8.39344442e-01 -7.00065374e-01 -6.52944088e-01
-4.90945667e-01 -2.07984626e-01 5.98209202e-01 7.32361674e-01
-9.40387666e-01 -2.28613913e-02 -7.42410958e-01 6.19278193e-01
9.65573370e-01 4.70403284e-01 9.48705912e-01 -1.10963070e+00
-2.89039761e-01 -2.76329935e-01 2.09582210e-01 -1.04743409e+00
3.27493548e-01 -4.25277650e-01 5.05948067e-01 -1.60893905e+00
2.97595441e-01 -6.80308104e-01 -4.03490931e-01 6.26118660e-01
-5.39643526e-01 3.70652348e-01 2.45985523e-01 1.02205202e-01
-5.58857381e-01 9.04505491e-01 1.52721143e+00 -1.51472315e-01
-2.00237438e-01 1.49269020e-02 -7.71614969e-01 7.05474079e-01
6.73993409e-01 -2.34992504e-01 -5.85545123e-01 -6.86465442e-01
-1.02899015e-01 -1.42879322e-01 8.30705523e-01 -8.73764753e-01
1.79943755e-01 -4.83776331e-01 4.86773133e-01 -6.17610395e-01
6.55221701e-01 -9.39863861e-01 -2.02313825e-01 3.16767067e-01
1.49483934e-01 -2.86934793e-01 2.73305357e-01 9.13695514e-01
7.71171302e-02 3.26953262e-01 7.30427802e-01 8.95034522e-02
-9.25136209e-01 3.85853261e-01 9.68717504e-03 3.38556468e-02
8.56354356e-01 -2.88676560e-01 -6.06531739e-01 -3.20619434e-01
-1.24727942e-01 2.49969050e-01 5.11159778e-01 1.05292939e-01
7.12689638e-01 -1.15043092e+00 -5.35272419e-01 2.14252785e-01
-3.30860883e-01 1.31142497e-01 4.85388577e-01 8.61858904e-01
-5.20570993e-01 -8.46204460e-02 -9.37183350e-02 -3.43716294e-01
-1.24506438e+00 2.72337735e-01 4.31165516e-01 -2.26063747e-02
-7.86310554e-01 9.08317506e-01 7.25736380e-01 -3.56094152e-01
5.96776903e-01 -1.76415578e-01 2.48152241e-01 -1.23964362e-01
3.59249413e-01 5.89995027e-01 -5.93830198e-02 -2.53768891e-01
-4.25056368e-01 4.24555063e-01 2.13016316e-01 7.51085505e-02
1.37095284e+00 -4.00014162e-01 -3.33476931e-01 8.88164192e-02
9.78603005e-01 1.40739322e-01 -1.92227769e+00 -3.54023218e-01
-4.12858248e-01 -8.41165304e-01 4.11021888e-01 -1.01241362e+00
-1.72728193e+00 8.48587275e-01 7.34061539e-01 -7.71927312e-02
1.53281915e+00 -4.33781803e-01 9.83940482e-01 3.48556399e-01
1.08515345e-01 -1.05034268e+00 2.54579037e-01 4.00084019e-01
8.70651543e-01 -1.31004024e+00 4.00705397e-01 -6.72442019e-01
-4.34855819e-01 9.82187867e-01 1.00680828e+00 3.52188125e-02
4.99589652e-01 4.23286825e-01 1.98381618e-01 2.45533208e-03
-3.52289617e-01 -1.27054676e-01 2.78722972e-01 6.55471504e-01
-3.88818756e-02 8.77132826e-03 6.76188543e-02 3.42949718e-01
2.49267444e-01 -4.14807111e-01 4.16690737e-01 8.20982277e-01
-4.08188075e-01 -8.21943879e-01 -6.41775370e-01 2.21380249e-01
2.00069591e-01 -3.92328978e-01 -5.10945439e-01 8.06824267e-01
1.97162732e-01 9.34331536e-01 -1.82441875e-01 -2.76188254e-01
1.76054329e-01 -5.41034102e-01 2.88567662e-01 -1.00521930e-01
-1.01774983e-01 1.87790290e-01 -5.85322492e-02 -8.61182988e-01
-4.88074869e-01 -4.20993179e-01 -1.17785037e+00 -4.83242393e-01
-3.01678807e-01 -2.01211810e-01 4.55739915e-01 9.05170441e-01
2.76712149e-01 8.15297723e-01 7.87902594e-01 -1.12537003e+00
-2.69214928e-01 -4.24200565e-01 -6.56416059e-01 1.03755705e-01
6.25263155e-01 -8.59002411e-01 -6.35322094e-01 -9.15863514e-02] | [10.859941482543945, -4.038919925689697] |
50c72e76-e376-4f2b-9bbc-3f085c7f628d | trustworthy-reinforcement-learning-for | 2302.11694 | null | https://arxiv.org/abs/2302.11694v1 | https://arxiv.org/pdf/2302.11694v1.pdf | Trustworthy Reinforcement Learning for Quadrotor UAV Tracking Control Systems | Simultaneously accurate and reliable tracking control for quadrotors in complex dynamic environments is challenging. As aerodynamics derived from drag forces and moment variations are chaotic and difficult to precisely identify, most current quadrotor tracking systems treat them as simple `disturbances' in conventional control approaches. We propose a novel, interpretable trajectory tracker integrating a Distributional Reinforcement Learning disturbance estimator for unknown aerodynamic effects with a Stochastic Model Predictive Controller (SMPC). The proposed estimator `Constrained Distributional Reinforced disturbance estimator' (ConsDRED) accurately identifies uncertainties between true and estimated values of aerodynamic effects. Simplified Affine Disturbance Feedback is used for control parameterization to guarantee convexity, which we then integrate with a SMPC. We theoretically guarantee that ConsDRED achieves at least an optimal global convergence rate and a certain sublinear rate if constraints are violated with an error decreases as the width and the layer of neural network increase. To demonstrate practicality, we show convergent training in simulation and real-world experiments, and empirically verify that ConsDRED is less sensitive to hyperparameter settings compared with canonical constrained RL approaches. We demonstrate our system improves accumulative tracking errors by at least 62% compared with the recent art. Importantly, the proposed framework, ConsDRED-SMPC, balances the tradeoff between pursuing high performance and obeying conservative constraints for practical implementations | ['David Boyle', 'Yanran Wang'] | 2023-02-22 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-2.13731766e-01 -8.08138102e-02 -3.55948240e-01 5.35834849e-01
-4.08094019e-01 -1.06001616e+00 4.04911786e-01 -9.77560654e-02
-2.69899875e-01 1.14788270e+00 -3.25940967e-01 -3.18634897e-01
-4.73545820e-01 -3.16162735e-01 -9.39144135e-01 -1.03591967e+00
-7.50386044e-02 5.23289852e-02 -3.46214534e-03 -2.93130726e-01
-1.39785334e-01 4.93860036e-01 -1.24866366e+00 -9.65112925e-01
1.33546436e+00 9.95804191e-01 2.14516163e-01 6.81826293e-01
9.35160756e-01 3.56661260e-01 -6.23167992e-01 2.19812822e-02
7.39832163e-01 -1.08024217e-02 2.77713567e-01 7.41232326e-03
3.10366839e-01 -2.23409846e-01 -2.14119121e-01 1.26766300e+00
7.46542752e-01 5.49594223e-01 6.46608531e-01 -1.09623313e+00
-6.19723618e-01 2.33584628e-01 -4.86576915e-01 8.20446312e-02
-5.94012029e-02 5.40870726e-01 6.18532717e-01 -7.26978302e-01
3.35815430e-01 1.00157011e+00 9.04738188e-01 5.69316685e-01
-1.01950932e+00 -7.22118437e-01 6.87991858e-01 -4.43356127e-01
-1.37979341e+00 -1.38286084e-01 5.35484314e-01 -5.92034280e-01
5.75916409e-01 2.79433548e-01 7.70860493e-01 9.17189419e-01
5.88355303e-01 4.29811209e-01 5.99770129e-01 1.97222143e-01
2.87243098e-01 2.54000723e-01 -3.54165643e-01 7.24335968e-01
9.43812668e-01 7.85765231e-01 1.95872381e-01 -2.84362561e-03
1.13864756e+00 -2.62787670e-01 -6.70195818e-01 -5.93500495e-01
-1.01670504e+00 7.43810952e-01 4.65816528e-01 -2.19164357e-01
-3.71989727e-01 3.02603394e-01 2.76627630e-01 4.16132063e-01
3.86495978e-01 1.02911222e+00 -5.80515862e-01 -6.88280836e-02
-3.85660112e-01 6.01344943e-01 7.35677898e-01 1.42294657e+00
-1.38027683e-01 1.11723459e+00 7.93293770e-03 5.82515538e-01
2.19148204e-01 1.16070795e+00 3.50953788e-01 -9.20314312e-01
4.11569685e-01 3.90914112e-01 7.35318661e-01 -1.14414692e+00
-4.27971482e-01 -9.98430312e-01 -1.03057027e+00 4.17386651e-01
3.61452311e-01 -1.08573914e+00 -3.89396995e-01 1.83704984e+00
6.93636119e-01 -4.27323692e-02 8.22630748e-02 1.12454665e+00
-2.63748970e-02 7.70865500e-01 -3.45669538e-01 -7.46637821e-01
9.17741060e-01 -7.32936919e-01 -1.09542227e+00 -6.34671301e-02
2.13775411e-01 -4.20978636e-01 9.77655292e-01 5.92628539e-01
-1.02479303e+00 -4.64760035e-01 -1.34492266e+00 6.40105665e-01
-2.07607299e-01 7.88485527e-01 5.84266223e-02 5.80792964e-01
-7.13727117e-01 8.12099397e-01 -8.67373645e-01 1.40589312e-01
-1.46799237e-01 5.58718204e-01 1.68669060e-01 9.87986803e-01
-1.10891068e+00 1.05026734e+00 4.71161246e-01 2.47608244e-01
-9.16669905e-01 -1.08579886e+00 -7.00098395e-01 -2.61901826e-01
8.05674493e-01 -8.25432956e-01 1.20653033e+00 -7.50449836e-01
-2.26258183e+00 -1.95463896e-01 3.26985329e-01 -6.17637634e-01
6.40734494e-01 -7.36657977e-01 -2.03835040e-01 -2.94851232e-02
-3.36517960e-01 5.85820265e-02 1.34617090e+00 -1.12431157e+00
-6.86608255e-01 6.48262575e-02 7.71363899e-02 4.52968657e-01
-3.64674509e-01 -5.15506506e-01 1.73454076e-01 -7.55474567e-01
-3.91479790e-01 -1.12706935e+00 -2.99178541e-01 2.09793031e-01
-2.03315079e-01 -8.56489837e-02 9.71538246e-01 -3.08434844e-01
1.42005742e+00 -1.83016050e+00 2.90161133e-01 -1.40554190e-01
5.93312234e-02 8.08743715e-01 2.34364897e-01 3.56284559e-01
7.05169514e-02 -9.07556862e-02 1.73390340e-02 1.56914860e-01
2.09000275e-01 -1.15647793e-01 -8.92856061e-01 7.46555269e-01
4.51778919e-01 6.22225225e-01 -1.23511457e+00 1.92310035e-01
4.50533956e-01 4.24120873e-01 -5.43218970e-01 2.02565596e-01
-4.12581474e-01 4.19617623e-01 -6.71698213e-01 3.45954776e-01
5.72026730e-01 -5.44777103e-02 1.16966635e-01 -9.22486559e-02
-5.59342146e-01 -2.17873514e-01 -1.34594429e+00 8.34638774e-01
-5.21723866e-01 5.99288225e-01 7.54583836e-01 -1.00491953e+00
1.04259598e+00 1.78208217e-01 2.46420860e-01 1.00889923e-02
5.98120332e-01 1.51646152e-01 -2.47185215e-01 -1.56627700e-01
3.69394928e-01 -1.11396715e-01 8.73042271e-02 -3.20296645e-01
-8.94648284e-02 -6.00591719e-01 -7.28136897e-02 -3.64434987e-01
5.39636612e-01 3.16513658e-01 5.61056376e-01 -6.69352353e-01
8.34742725e-01 3.87435071e-02 8.68204296e-01 4.10586685e-01
-3.45660329e-01 -4.39226925e-02 3.57166499e-01 -9.13215354e-02
-9.83220935e-01 -8.89662743e-01 -1.12957351e-01 6.05489016e-01
3.98440123e-01 -1.42397860e-03 -4.59972024e-01 -3.12560171e-01
4.30861950e-01 7.05066204e-01 -2.08778262e-01 -5.04225612e-01
-4.50877607e-01 -5.09175777e-01 2.07578346e-01 5.75581431e-01
2.23522991e-01 -1.44741032e-02 -6.42175853e-01 2.76480973e-01
8.22620511e-01 -1.02145529e+00 -7.50747204e-01 1.85438961e-01
-5.24005175e-01 -1.10250759e+00 -5.01128733e-01 -6.31911576e-01
7.00972974e-01 1.52574079e-02 4.55708116e-01 -4.31353331e-01
2.32414398e-02 4.15378451e-01 2.04152707e-02 -6.61333263e-01
-8.82607922e-02 -1.17710322e-01 9.28236425e-01 -1.54121026e-01
-6.51018083e-01 -3.65592510e-01 -4.43011791e-01 4.94443834e-01
-4.57877517e-01 -2.98608124e-01 3.09834898e-01 1.17038631e+00
6.17219746e-01 3.13654244e-01 7.00535417e-01 -3.97391111e-01
9.53767121e-01 -3.28904599e-01 -1.78255403e+00 -8.45731199e-02
-8.26951683e-01 -1.01604417e-01 1.71449208e+00 -9.60972428e-01
-8.65240872e-01 5.72143607e-02 2.63781726e-01 -1.15913010e+00
4.43564296e-01 1.58192903e-01 6.77823275e-02 -3.51782471e-01
4.71018553e-01 9.93928835e-02 3.05209547e-01 1.33809149e-01
3.24986696e-01 4.07718331e-01 4.82035100e-01 -6.35651827e-01
1.34398878e+00 6.43705204e-02 3.93686563e-01 -8.77134144e-01
-7.55832255e-01 -2.15985566e-01 -2.80891299e-01 -3.26689214e-01
3.36021960e-01 -1.17314374e+00 -1.34975600e+00 2.54369229e-01
-8.17576587e-01 -3.23216677e-01 -5.46563447e-01 7.36630201e-01
-7.68870831e-01 -8.39984342e-02 -3.43650401e-01 -1.41855717e+00
-4.46697652e-01 -9.92554069e-01 6.69860363e-01 4.84274656e-01
2.61284441e-01 -1.36431742e+00 7.04188086e-03 -4.67877656e-01
4.47424740e-01 8.58541250e-01 3.43897343e-01 -1.81494012e-01
-2.75559217e-01 -2.28729963e-01 2.49559537e-01 5.70110500e-01
1.79602355e-01 2.31723413e-01 -3.50769550e-01 -9.52772260e-01
2.95522451e-01 -1.97544590e-01 1.85492605e-01 7.43315041e-01
6.89378798e-01 -7.77702153e-01 -4.72392470e-01 7.24672496e-01
1.71034443e+00 5.91050625e-01 -3.42177063e-01 3.25607844e-02
7.61392593e-01 1.29387915e-01 9.65404928e-01 7.87387073e-01
-3.00291702e-02 3.38302225e-01 6.08300328e-01 2.36504540e-01
4.26654994e-01 -3.57181072e-01 7.75811791e-01 1.03381002e+00
7.26720765e-02 -3.26152772e-01 -4.10194516e-01 6.07634902e-01
-1.86900640e+00 -6.31592631e-01 5.67324832e-02 2.50996208e+00
7.91569531e-01 9.16267335e-02 9.30823386e-02 -1.27446637e-01
9.18501318e-01 1.00355521e-02 -1.04720056e+00 -4.96496528e-01
8.39593187e-02 -2.30969355e-01 9.96908784e-01 6.02127790e-01
-1.28445804e+00 6.51284695e-01 5.58198261e+00 8.99608254e-01
-1.40021169e+00 -3.97062272e-01 1.17423246e-03 -2.87962019e-01
9.92330909e-02 -3.80473316e-01 -1.08939433e+00 5.32187998e-01
9.31730926e-01 -7.31876373e-01 7.29010403e-01 1.18872678e+00
6.81708634e-01 4.12643194e-01 -7.20963717e-01 5.96272647e-01
-2.31903389e-01 -9.94925559e-01 -1.79259390e-01 -1.25088483e-01
1.10703421e+00 -2.53174871e-01 5.27050078e-01 5.01946747e-01
5.10131955e-01 -5.81002831e-01 7.11503088e-01 7.39367485e-01
7.91922092e-01 -1.02380311e+00 4.43942636e-01 5.42733133e-01
-1.26771486e+00 -6.56069696e-01 -5.81191540e-01 -1.56606808e-01
2.72396617e-02 3.20795894e-01 -8.76387954e-01 8.30452025e-01
-1.43469617e-01 1.10986853e+00 1.29852921e-01 1.11538899e+00
-4.49669182e-01 5.04449248e-01 -6.47535443e-01 -8.06771874e-01
4.21697944e-01 -6.26358390e-01 1.34586811e+00 6.80376410e-01
3.18444997e-01 -3.05160712e-02 4.59122807e-01 8.60561132e-01
3.10711354e-01 -3.35089266e-02 -8.72551382e-01 -2.19651714e-01
7.49355853e-01 1.12849760e+00 -3.67844515e-02 -8.82027894e-02
1.50962815e-01 1.43797308e-01 6.78782165e-02 3.03319544e-01
-1.21604359e+00 -7.23466396e-01 9.52992439e-01 -1.87338799e-01
4.11437929e-01 -4.74381089e-01 2.61709094e-02 -1.01684666e+00
2.18649413e-02 -8.71239662e-01 -1.91253647e-02 -3.73615801e-01
-1.22850776e+00 4.80386406e-01 -1.54621989e-01 -1.99088621e+00
-5.09364963e-01 -8.31240356e-01 -4.50744718e-01 5.67058325e-01
-1.43156683e+00 -6.91908419e-01 1.11981206e-01 2.43534490e-01
5.75983942e-01 -2.71445334e-01 4.84513879e-01 -1.22054704e-02
-1.01665306e+00 4.87979978e-01 8.54441404e-01 -3.99221927e-01
6.02946281e-01 -1.51639545e+00 -9.03311148e-02 9.20892179e-01
-7.33682632e-01 7.02168345e-01 1.06781471e+00 -6.32121563e-01
-1.97969818e+00 -1.67126143e+00 -1.43982694e-01 -1.93471432e-01
1.13859379e+00 -2.27795601e-01 -5.42118013e-01 5.64980924e-01
6.62343502e-02 1.91244602e-01 -2.92437792e-01 -5.41060686e-01
4.18383181e-01 -5.53659916e-01 -9.36180294e-01 7.14454472e-01
7.85588980e-01 2.42351517e-02 -4.01756823e-01 3.39234650e-01
1.10967672e+00 -9.82137799e-01 -1.17091167e+00 7.15680301e-01
3.86703908e-01 -9.72016305e-02 7.75210500e-01 -5.50477028e-01
-2.22158171e-02 -6.76229596e-01 1.99842870e-01 -1.96416724e+00
-6.74779788e-02 -1.47772980e+00 -6.11321151e-01 8.92689466e-01
1.87799349e-01 -9.85664070e-01 4.30955678e-01 -1.46942779e-01
-3.25495958e-01 -1.16091728e+00 -8.36690187e-01 -1.41777813e+00
4.50701296e-01 -8.41010511e-02 2.70626366e-01 9.75059867e-01
3.68996114e-02 2.99129844e-01 -4.39857751e-01 1.06905377e+00
5.96626818e-01 -8.12612176e-02 7.69717276e-01 -1.00879371e+00
-1.20725609e-01 -5.07391691e-01 9.24629495e-02 -1.28875625e+00
1.68400213e-01 -3.78222644e-01 2.16131970e-01 -1.05549061e+00
-8.61670554e-01 -3.35619569e-01 -5.34791648e-02 -8.24430659e-02
-3.47137481e-01 -4.28130627e-01 3.13209221e-02 -1.01308040e-01
-4.26100045e-01 1.07336700e+00 1.54724777e+00 -9.47050899e-02
-5.06886125e-01 4.56155777e-01 -2.88491040e-01 7.76895165e-01
7.25258768e-01 -4.47217748e-03 -1.06454146e+00 -2.89201170e-01
-5.87095693e-02 3.78427565e-01 6.04915470e-02 -1.39111984e+00
3.86967182e-01 -4.02160466e-01 2.55929619e-01 -4.99778330e-01
1.01374827e-01 -9.16310072e-01 1.47990793e-01 6.97109520e-01
-1.57145604e-01 1.14569426e-01 6.16609395e-01 8.53169858e-01
1.46804424e-03 2.27241382e-01 1.02670336e+00 4.61491585e-01
-1.80106997e-01 4.03653830e-01 -4.45324957e-01 1.62930012e-01
1.26498067e+00 3.85735407e-02 -3.39032739e-01 -3.97113442e-01
-3.32632452e-01 8.78059208e-01 2.25173801e-01 3.86029899e-01
4.59438801e-01 -1.10666966e+00 -2.25377902e-01 1.86375976e-01
-5.84665656e-01 2.18266202e-03 -9.77537557e-02 8.33105743e-01
-2.07451165e-01 8.58293116e-01 5.72852790e-02 -5.12959480e-01
-8.24864507e-01 7.23813713e-01 6.89011037e-01 -1.37311205e-01
-4.35522348e-01 6.00143194e-01 1.17967434e-01 -4.79964614e-01
1.98991463e-01 -7.85623133e-01 -4.19035181e-02 -2.53857166e-01
1.44477174e-01 4.37614799e-01 -1.71494752e-01 -1.10143982e-01
-1.85941271e-02 7.32427299e-01 2.94647962e-01 3.98665279e-01
1.11235428e+00 -1.65342361e-01 5.76972008e-01 3.29524606e-01
6.64184988e-01 4.61829454e-02 -1.99642313e+00 2.19371945e-01
-3.85389179e-01 -1.67009458e-01 2.53109366e-01 -7.81675637e-01
-1.08057940e+00 2.31460243e-01 4.73386198e-01 3.29962462e-01
1.08402371e+00 -9.71479893e-01 5.39048195e-01 5.13791502e-01
3.81580919e-01 -1.25616264e+00 -2.28277043e-01 8.87239039e-01
1.03390110e+00 -7.59586334e-01 3.11381161e-01 -3.60231727e-01
-7.79557824e-01 1.07992911e+00 1.05133891e+00 -9.08018887e-01
4.71837193e-01 4.34855402e-01 -2.17702642e-01 3.74349535e-01
-1.21781731e+00 1.00115299e-01 4.77662116e-01 4.67857063e-01
-1.13527767e-01 -1.65452305e-02 -5.63660026e-01 6.85613632e-01
-2.08192915e-01 -4.36976075e-01 5.36333561e-01 7.05862701e-01
-5.08738756e-01 -5.59011579e-01 -4.06436712e-01 -2.44321097e-02
-2.87862659e-01 3.89681697e-01 6.32405132e-02 1.39085221e+00
-3.55776952e-04 6.73302472e-01 4.34393398e-02 -1.72968581e-01
7.10820138e-01 -6.80389047e-01 3.59843433e-01 -3.55920017e-01
-4.56786394e-01 2.26201177e-01 -4.50628065e-02 -4.49351043e-01
5.99885546e-02 -4.99763608e-01 -1.12413359e+00 2.66427826e-02
-9.84470069e-01 3.71659338e-01 4.75047410e-01 6.49273813e-01
3.78049940e-01 7.75747955e-01 1.04051125e+00 -6.11116409e-01
-1.49346220e+00 -7.20074236e-01 -5.67756355e-01 -3.57333094e-01
9.33619857e-01 -1.23936129e+00 -7.93804824e-01 -1.88315764e-01] | [5.015445709228516, 2.3321259021759033] |
c69e72ed-0190-40fb-a0d9-b80aff708b2f | a-novel-deep-arrhythmia-diagnosis-network-for | null | null | https://doi.org/10.1109/ACCESS.2019.2918792 | https://ieeexplore.ieee.org/ielx7/6287639/8600701/08721643.pdf | A Novel Deep Arrhythmia-Diagnosis Network for Atrial Fibrillation Classification Using Electrocardiogram Signals | Atrial fibrillation (AF), a common abnormal heartbeat rhythm, is a life-threatening recurrent disease that affects older adults. Automatic classification is one of the most valuable topics in medical sciences and bioinformatics, especially the detection of atrial fibrillation. However, it is difficult to accurately explain the local characteristics of electrocardiogram (ECG) signals by manual analysis, due to their small amplitude and short duration, coupled with the complexity and non-linearity. Hence, in this paper, we propose a novel deep arrhythmia-diagnosis method, named deep CNN-BLSTM network model, to automatically detect the AF heartbeats using the ECG signals. The model mainly consists of four convolution layers: two BLSTM layers and two fully connected layers. The datasets of RR intervals (called set A) and heartbeat sequences (P-QRS-T waves, called set B) are fed into the above-mentioned model. Most importantly, our proposed approach achieved favorable performances with an accuracy of 99.94% and 98.63% in the training and validation set of set A, respectively. In the testing set (unseen data sets), we obtained an accuracy of 96.59%, a sensitivity of 99.93%, and a specificity of 97.03%. To the best of our knowledge, the algorithm we proposed has shown excellent results compared to many state-of-art researches, which provides a new solution for the AF automatic detection. | ['Hao Dang', 'Xingqun Qi', 'Xiaoguang Zhou', 'Guanhong Zhang', 'Qing Chang', 'Muyi Sun'] | 2019-05-24 | null | null | null | ieee-access-2019-5 | ['arrhythmia-detection', 'atrial-fibrillation-detection', 'electrocardiography-ecg'] | ['medical', 'medical', 'methodology'] | [ 1.14473850e-01 -3.84773165e-01 1.32600412e-01 -7.18847513e-02
-2.97807217e-01 -1.15306549e-01 -1.72960088e-01 1.53995112e-01
-3.36094469e-01 1.09660470e+00 -3.57708007e-01 -5.18857956e-01
-4.20374572e-02 -7.96120405e-01 -1.12242542e-01 -8.34916353e-01
-2.21867591e-01 9.88818407e-02 -2.57061332e-01 1.51795313e-01
-6.13939688e-02 4.61002648e-01 -1.29435122e+00 2.20497668e-01
1.25179636e+00 1.19006944e+00 -1.69754684e-01 6.60393655e-01
1.45280123e-01 4.17504400e-01 -8.57706249e-01 -9.97750163e-02
-4.66839969e-01 -8.11503172e-01 -2.44176179e-01 -4.11025316e-01
-3.02439719e-01 -2.58616328e-01 -1.70878261e-01 7.62274623e-01
1.10730064e+00 -5.14129937e-01 4.44914401e-01 -7.46072412e-01
-4.12884623e-01 4.95850146e-01 -5.44547141e-01 7.97489405e-01
-6.17901832e-02 5.04543632e-02 3.73367041e-01 -7.64689803e-01
-3.24265100e-02 5.35026550e-01 1.14312875e+00 4.93219823e-01
-9.24673855e-01 -1.00954664e+00 -1.88488901e-01 2.62765348e-01
-1.58194864e+00 -1.88920453e-01 8.13057780e-01 -5.26841819e-01
6.96475625e-01 2.17186898e-01 1.24754679e+00 1.02375829e+00
6.68251395e-01 4.69492882e-01 9.38635409e-01 -2.49977216e-01
-8.27687308e-02 -1.96021453e-01 3.33662599e-01 5.38512409e-01
6.18488431e-01 1.09072775e-01 -6.68060333e-02 -4.08812165e-01
9.91325200e-01 2.62246907e-01 -6.77907646e-01 2.49577537e-01
-1.44412851e+00 5.20321190e-01 4.16413993e-01 8.49526763e-01
-6.24918878e-01 -7.81959742e-02 5.01233637e-01 1.34425759e-01
2.54209965e-01 1.06235102e-01 -4.76116419e-01 -4.01200473e-01
-8.30269575e-01 2.08007470e-02 5.52782536e-01 3.35805506e-01
1.78760052e-01 4.29253876e-01 -1.39056340e-01 5.84601462e-01
2.23827511e-01 6.76280856e-01 9.57219183e-01 -2.55354881e-01
2.17651665e-01 5.37552953e-01 1.43899933e-01 -1.17691088e+00
-7.63628364e-01 -1.25575399e+00 -1.59695613e+00 -1.79719076e-01
2.84243613e-01 -5.12744129e-01 -7.77977407e-01 1.72745562e+00
-9.11974907e-02 3.81748557e-01 2.18913838e-01 8.88401151e-01
1.02212751e+00 4.46182191e-01 1.78229257e-01 -7.79458523e-01
1.37082934e+00 -2.91166037e-01 -1.06193995e+00 -4.24951464e-02
4.34887201e-01 -5.12807369e-01 8.59764338e-01 5.87169290e-01
-7.91970670e-01 -8.18890750e-01 -1.18697810e+00 3.66533577e-01
6.67964295e-02 8.54554892e-01 5.56097031e-01 7.59127617e-01
-8.51273894e-01 4.30829167e-01 -8.99647534e-01 -2.84231216e-01
6.28158569e-01 3.65291417e-01 -4.09829319e-02 3.49794120e-01
-1.52555215e+00 6.28769159e-01 3.45290452e-01 6.69781744e-01
-6.36709750e-01 -4.08713728e-01 -5.39620459e-01 2.33531073e-02
-3.24339926e-01 -9.10982847e-01 9.85792875e-01 -8.33206952e-01
-1.12531292e+00 7.47919321e-01 -2.57417291e-01 -5.20823598e-01
4.15913731e-01 -5.27278185e-01 -8.10800552e-01 -3.79231796e-02
-8.84286165e-02 -1.22875951e-01 6.84414208e-01 -7.62529612e-01
-3.98344666e-01 -6.30605817e-01 -5.24290085e-01 -3.24454755e-02
-3.85936171e-01 -2.05739006e-01 1.66592464e-01 -8.82755995e-01
2.29572594e-01 -5.95693707e-01 -1.66450098e-01 -4.25316900e-01
-2.01520815e-01 -6.20451681e-02 4.89235729e-01 -1.12213719e+00
1.88185048e+00 -2.21423554e+00 -1.71261787e-01 2.01497138e-01
5.36351919e-01 7.44755268e-01 4.27855074e-01 2.99629197e-02
-2.55393773e-01 2.10527256e-01 -5.66791058e-01 2.15660527e-01
-7.10595906e-01 -1.19208999e-01 -2.06556067e-01 3.07090849e-01
7.71523938e-02 9.67404485e-01 -7.88932025e-01 -2.97705412e-01
9.95713472e-02 8.24857414e-01 2.13514596e-01 1.40427962e-01
2.62182117e-01 5.96019149e-01 -6.25671923e-01 4.32698548e-01
6.67897701e-01 -2.40030408e-01 1.04030721e-01 -2.48196453e-01
-2.20169947e-02 -1.90422330e-02 -8.62165272e-01 1.46752918e+00
-1.45168245e-01 4.39928055e-01 -4.20765817e-01 -1.05932808e+00
1.35953987e+00 8.04086983e-01 5.08558035e-01 -6.16678715e-01
4.00943696e-01 6.23073816e-01 5.86298168e-01 -6.38494432e-01
-4.74318564e-01 -1.74338609e-01 3.25257152e-01 4.23317105e-01
-3.39174479e-01 4.94053602e-01 -1.32789582e-01 -3.28112274e-01
8.67839754e-01 -1.54408887e-01 4.82455194e-01 -2.16030449e-01
7.77187228e-01 -3.88786197e-01 7.21534908e-01 6.01885915e-01
-2.58113772e-01 7.04893529e-01 2.50937343e-01 -1.07406414e+00
-5.79781532e-01 -7.44137049e-01 -3.71504575e-01 -1.41917109e-01
-1.06582053e-01 -3.24147642e-01 -7.40143478e-01 -4.07761037e-01
-2.36696824e-01 1.52454183e-01 -4.77769375e-01 -5.17155409e-01
-8.16419661e-01 -1.24331248e+00 1.00239730e+00 7.33178318e-01
9.90990639e-01 -1.49454963e+00 -1.01397264e+00 5.52619040e-01
-4.54214752e-01 -7.46312201e-01 -1.68781877e-01 -8.50801077e-03
-1.26483130e+00 -1.19315171e+00 -1.12345362e+00 -9.59306002e-01
2.78657377e-01 -1.94176868e-01 1.03423202e+00 3.99113894e-01
-5.25621116e-01 -3.33589435e-01 -8.94119367e-02 -7.86208510e-01
-1.05063781e-01 1.47671029e-01 1.35847226e-01 3.30971688e-01
5.06703019e-01 -8.19512069e-01 -9.44742084e-01 1.09622471e-01
-2.76800305e-01 -7.58821657e-03 8.98012996e-01 7.69889295e-01
7.44990587e-01 -1.19323574e-01 1.15178752e+00 -6.67291522e-01
8.98696423e-01 -1.97692409e-01 -1.35984868e-01 9.91152450e-02
-8.02331388e-01 -4.98832732e-01 7.85690784e-01 -3.89321655e-01
-6.66054845e-01 1.98202208e-02 -2.76982099e-01 -2.06182823e-01
-2.71069854e-01 6.10514104e-01 -1.79835215e-01 2.11771086e-01
7.08344579e-01 5.58333457e-01 1.31570533e-01 -3.26043427e-01
-2.95565963e-01 9.10237372e-01 7.23573565e-01 -2.10220560e-01
3.47791046e-01 1.53181717e-01 -2.07135692e-01 -7.17841089e-01
-6.85925663e-01 -2.25796282e-01 -5.03310621e-01 -9.05752182e-02
9.06657159e-01 -1.02343893e+00 -6.49870098e-01 9.25129950e-01
-1.24022675e+00 2.49475632e-02 -2.04381812e-02 7.60729313e-01
-2.89480627e-01 4.69443023e-01 -6.03148103e-01 -1.07754672e+00
-1.28181911e+00 -7.49335706e-01 5.61003923e-01 3.77013505e-01
-3.43888462e-01 -7.86427438e-01 1.96870174e-02 -1.15342930e-01
6.07547402e-01 8.48521650e-01 7.80217588e-01 -5.86301982e-01
1.24706283e-01 -2.83602744e-01 -1.89969435e-01 4.53394443e-01
5.26336908e-01 -6.82453811e-02 -1.05862248e+00 -3.07883441e-01
4.37558711e-01 2.07644239e-01 7.79397607e-01 6.30185366e-01
1.31942797e+00 -2.46902183e-02 -5.07265210e-01 6.43311262e-01
1.00415874e+00 8.53947103e-01 8.09168994e-01 1.85250655e-01
6.39248252e-01 -1.98485225e-01 4.68799740e-01 4.83530551e-01
8.40230286e-02 2.58573472e-01 1.77942961e-01 -5.59134960e-01
1.00918166e-01 1.51703611e-01 1.60914198e-01 8.30770552e-01
-5.82916677e-01 -7.27826804e-02 -1.11031759e+00 4.70930785e-01
-1.89221704e+00 -9.16313827e-01 -7.33824313e-01 2.33758640e+00
9.47878838e-01 2.68470347e-01 1.15859851e-01 7.96622813e-01
9.17304099e-01 -1.42039850e-01 -6.40154779e-01 -2.26911873e-01
-3.14965576e-01 5.96427500e-01 -2.31777936e-01 9.15751606e-02
-1.26399767e+00 1.90867856e-01 5.39537430e+00 3.46980214e-01
-1.45445633e+00 -3.57980877e-02 8.83877039e-01 3.86238426e-01
6.55297004e-03 -4.34794009e-01 -4.14703190e-01 7.94894814e-01
9.70094800e-01 3.64410831e-03 -6.09112829e-02 3.67412329e-01
4.31651771e-01 5.21760404e-01 -8.63010764e-01 1.37258685e+00
1.21339262e-01 -1.01947856e+00 6.50498085e-03 -1.78025290e-01
1.48915783e-01 -4.36245501e-01 -2.04636708e-01 8.89305845e-02
-6.24398112e-01 -1.27619863e+00 2.84653395e-01 8.19234669e-01
1.17499793e+00 -8.62980902e-01 1.30287361e+00 3.76387984e-01
-1.23433673e+00 -1.16480380e-01 -1.05934054e-01 -3.18575829e-01
-2.61128368e-03 1.17980409e+00 -3.70311975e-01 6.53872669e-01
9.58459795e-01 1.08584249e+00 -3.40015829e-01 1.25544643e+00
-3.73065680e-01 9.57683325e-01 -2.35440314e-01 -2.34116465e-02
-3.76342714e-01 -1.32746860e-01 3.29900146e-01 1.03942287e+00
5.67325354e-01 3.35668266e-01 -8.35135877e-02 8.47417057e-01
1.81472197e-01 1.78184196e-01 -4.53482091e-01 1.22805402e-01
3.71521682e-01 1.11722946e+00 -5.22062063e-01 -4.11295027e-01
-7.49286190e-02 6.27655447e-01 -1.62608966e-01 4.17567044e-01
-8.85422528e-01 -8.77629578e-01 6.14976659e-02 5.89606129e-02
-1.66716367e-01 1.35022417e-01 -7.42571235e-01 -1.20193028e+00
3.37643743e-01 -9.86345768e-01 5.67335606e-01 -5.26322722e-01
-9.18325126e-01 8.83002281e-01 -5.05621552e-01 -1.37903571e+00
-3.40848118e-02 -2.37314925e-01 -9.64269280e-01 1.24409890e+00
-1.30443907e+00 -7.92267382e-01 -9.05738294e-01 5.07213116e-01
8.99256766e-02 -2.83395112e-01 1.30601776e+00 5.94542205e-01
-8.01303566e-01 6.76323831e-01 -1.79355904e-01 3.50614697e-01
5.12465835e-01 -1.18509614e+00 3.71055484e-01 9.50630188e-01
-9.85853150e-02 7.53094196e-01 2.74418026e-01 -5.47981858e-01
-6.21286929e-01 -1.00438547e+00 1.17228067e+00 -7.54382461e-02
1.67835191e-01 -1.23033784e-02 -1.04443097e+00 3.90512973e-01
-1.20866559e-01 5.17849475e-02 7.80292094e-01 -1.57155082e-01
2.44832814e-01 -3.60553294e-01 -8.28533351e-01 2.47335389e-01
8.15785706e-01 -2.13710636e-01 -5.57835698e-01 1.50787160e-02
3.48050803e-01 -3.55844051e-01 -1.04871082e+00 9.65138137e-01
9.50492442e-01 -9.12284851e-01 9.26373243e-01 -2.80427814e-01
1.34575397e-01 -4.23891544e-01 5.44045091e-01 -1.08971727e+00
-4.82314587e-01 -5.47267079e-01 -4.56032246e-01 9.68186140e-01
2.22074464e-01 -9.38243210e-01 4.85826105e-01 -7.21849650e-02
-2.64749348e-01 -1.19022119e+00 -7.24748671e-01 -3.47549856e-01
-3.05007100e-01 -6.23493716e-02 5.00862956e-01 8.02618325e-01
-2.34541282e-01 4.75086898e-01 -2.61719704e-01 7.77889639e-02
4.49169427e-01 3.29997569e-01 2.91005790e-01 -1.73577118e+00
-7.50775859e-02 -3.40635061e-01 -3.26824427e-01 -3.82784158e-01
-4.01979268e-01 -5.86310804e-01 -3.51833999e-01 -1.66951489e+00
3.65565196e-02 -5.18158734e-01 -8.92241359e-01 5.63393235e-01
-4.82917339e-01 4.31679428e-01 -4.37820792e-01 3.46363157e-01
1.14685982e-01 3.76510561e-01 1.08402431e+00 -1.41401350e-01
-6.11893356e-01 5.87746084e-01 -7.32468843e-01 7.44454384e-01
1.26540363e+00 -1.56338319e-01 -3.88749659e-01 -2.42832810e-01
-4.18077281e-04 1.62483484e-01 2.64560044e-01 -1.22422028e+00
-1.85972139e-01 3.85999620e-01 9.93225932e-01 -5.38434565e-01
-3.94259058e-02 -5.09461164e-01 3.77273977e-01 1.25556219e+00
-3.16408053e-02 2.78191537e-01 2.12931097e-01 3.15211773e-01
-4.26278383e-01 2.30162993e-01 4.83629912e-01 -1.16427699e-02
-1.95798174e-01 5.23686767e-01 -3.55373323e-01 -5.48247844e-02
9.04911458e-01 -2.80943781e-01 -1.27531394e-01 -1.49848431e-01
-9.55503285e-01 -6.53497428e-02 -1.50637835e-01 3.26357871e-01
1.05249918e+00 -1.43212938e+00 -9.45273936e-01 4.73812550e-01
7.24620968e-02 2.52836496e-01 4.35420394e-01 1.23187661e+00
-7.00607777e-01 2.44252220e-01 -3.80881935e-01 -6.63392007e-01
-1.27058506e+00 2.45766744e-01 7.74561763e-01 -1.99208289e-01
-9.10079181e-01 4.98674750e-01 -4.74680215e-02 3.27499181e-01
3.61851662e-01 -5.24590909e-01 -7.84170270e-01 2.46806368e-02
6.76972747e-01 2.42804661e-01 3.10483783e-01 -3.45687598e-01
-6.61313534e-01 7.90112019e-01 1.33243337e-01 4.23661470e-01
1.05016696e+00 3.78932238e-01 -2.41039053e-01 6.00487828e-01
6.30649686e-01 -1.78520307e-01 -4.84730452e-01 6.36777356e-02
-2.89478660e-01 2.07730997e-02 -1.76477462e-01 -9.82407749e-01
-1.13688493e+00 1.26501167e+00 1.13447225e+00 1.15517706e-01
1.27359557e+00 -5.19141614e-01 1.12594700e+00 1.55419216e-01
2.06844002e-01 -4.38401312e-01 -2.57005155e-01 1.87932793e-02
7.78134167e-01 -7.22895026e-01 -2.14817330e-01 -7.92238712e-02
-1.66890055e-01 1.41977715e+00 5.47681749e-01 -2.71998011e-02
8.03634703e-01 6.77433759e-02 4.00101513e-01 -5.70487380e-02
-2.14978024e-01 2.04942614e-01 8.07820708e-02 6.11658990e-01
6.83297873e-01 1.56537309e-01 -9.64572370e-01 1.31528544e+00
-2.18550384e-01 5.86877227e-01 3.08120698e-01 6.76812530e-01
-4.02026743e-01 -8.83301139e-01 -1.84546784e-01 7.33547986e-01
-7.52327442e-01 -1.02824479e-01 -2.19776258e-01 6.09430790e-01
4.18280303e-01 9.28987384e-01 -1.38351515e-01 -5.47414422e-01
2.12600231e-01 1.85672879e-01 1.48779616e-01 -3.59909117e-01
-5.92124283e-01 4.21063639e-02 -1.69368461e-01 -1.12415180e-01
-4.62518394e-01 -4.47570056e-01 -1.50710297e+00 2.34781757e-01
-2.10305527e-01 4.32343841e-01 2.38021061e-01 8.42920721e-01
5.42140543e-01 8.59542489e-01 4.89262879e-01 -1.67324781e-01
-2.83603907e-01 -1.28286552e+00 -5.81787825e-01 2.45693266e-01
6.32160902e-01 -3.66629571e-01 -2.64239371e-01 1.16862819e-01] | [14.249869346618652, 3.295222043991089] |
e8c564d6-6638-4b7e-92be-acba30c4ce78 | frequency-of-interest-based-noise-attenuation | 2210.11068 | null | https://arxiv.org/abs/2210.11068v3 | https://arxiv.org/pdf/2210.11068v3.pdf | Frequency of Interest-based Noise Attenuation Method to Improve Anomaly Detection Performance | Accurately extracting driving events is the way to maximize computational efficiency and anomaly detection performance in the tire frictional nose-based anomaly detection task. This study proposes a concise and highly useful method for improving the precision of the event extraction that is hindered by extra noise such as wind noise, which is difficult to characterize clearly due to its randomness. The core of the proposed method is based on the identification of the road friction sound corresponding to the frequency of interest and removing the opposite characteristics with several frequency filters. Our method enables precision maximization of driving event extraction while improving anomaly detection performance by an average of 8.506%. Therefore, we conclude our method is a practical solution suitable for road surface anomaly detection purposes in outdoor edge computing environments. | ['Won Seok Park', 'Myung Jin Kim', 'YeongHyeon Park'] | 2022-10-20 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 2.34725326e-01 -4.44490999e-01 2.06537634e-01 7.88823590e-02
-4.16229934e-01 -1.95263222e-01 4.32660758e-01 2.39873856e-01
-7.12536156e-01 4.25829351e-01 -2.29298413e-01 -3.30658168e-01
-3.76629204e-01 -1.11033452e+00 -4.06698793e-01 -9.36006308e-01
6.37771562e-02 -1.56336099e-01 5.50528705e-01 -3.20697635e-01
4.49130386e-01 8.34284782e-01 -2.32244587e+00 -1.55821323e-01
1.11498451e+00 1.17662013e+00 -1.99698377e-02 7.22513020e-01
-2.11793154e-01 3.37531477e-01 -6.04254663e-01 -1.08192883e-01
3.93624566e-02 -3.45744472e-03 -1.27091721e-01 -2.51619458e-01
1.34301648e-01 -3.10317308e-01 -2.13712946e-01 9.31887329e-01
4.77588445e-01 5.29555142e-01 8.09954643e-01 -1.25765169e+00
2.75475115e-01 -1.03210136e-01 -6.90028787e-01 6.91394687e-01
-4.60021496e-02 -5.85984737e-02 5.33043444e-01 -9.64174211e-01
1.65327638e-01 4.67527866e-01 6.98058844e-01 2.12006599e-01
-6.77047849e-01 -5.99417150e-01 -2.01232567e-01 7.31049836e-01
-1.64451563e+00 -5.70184529e-01 1.06290853e+00 -1.87966600e-01
9.39590275e-01 5.08680344e-01 4.53643084e-01 7.37648726e-01
3.47523630e-01 5.05817235e-01 5.71784854e-01 -2.35293418e-01
1.81107655e-01 -2.20725879e-01 3.83950204e-01 2.39517167e-01
6.57238245e-01 2.53963053e-01 -3.42303991e-01 -2.43476838e-01
1.88981712e-01 1.18775936e-02 4.11858894e-02 1.72458261e-01
-5.92753112e-01 5.06919980e-01 -6.64361492e-02 1.74567550e-01
-6.28792346e-01 9.66562480e-02 7.30186284e-01 5.32837138e-02
4.52861786e-01 1.02154039e-01 -2.16522366e-01 -5.62312663e-01
-9.54346955e-01 4.59713131e-01 5.61462224e-01 5.79633355e-01
5.61942637e-01 5.61791420e-01 3.77923548e-02 7.06766367e-01
1.05957665e-01 1.01194334e+00 2.23745972e-01 -7.02623725e-01
1.77887917e-01 3.10187519e-01 3.20815831e-01 -1.23963368e+00
-4.57336009e-01 -4.86154616e-01 -7.29668617e-01 2.78296679e-01
5.02059817e-01 -3.26438427e-01 -6.63187683e-01 1.07896864e+00
4.58981037e-01 6.64850771e-01 1.65294670e-02 6.31204665e-01
6.83329463e-01 6.07849658e-01 1.44362301e-01 -1.75560713e-01
1.50302434e+00 -4.37834561e-01 -1.15537775e+00 -1.03038125e-01
6.93738043e-01 -9.83518839e-01 7.18728483e-01 4.35138792e-01
-5.99843681e-01 -5.08166850e-01 -9.85582709e-01 2.90014356e-01
-5.21808982e-01 3.73203814e-01 6.54200554e-01 7.89987981e-01
-3.91258717e-01 2.90924996e-01 -8.42486143e-01 -3.12655866e-02
2.45074838e-01 1.62316039e-01 -1.32852584e-01 3.13408941e-01
-1.30328393e+00 8.46304476e-01 1.29924208e-01 5.06399512e-01
-1.76391914e-01 -5.37428975e-01 -9.50493038e-01 7.06650019e-02
4.23968703e-01 -2.00954199e-01 9.62435544e-01 -3.51699531e-01
-1.37142444e+00 4.38831538e-01 -7.36965358e-01 -7.93076932e-01
4.40618008e-01 -5.17415226e-01 -1.10422087e+00 1.37659073e-01
1.30802765e-02 -3.08690518e-01 1.24176562e+00 -8.64451885e-01
-1.01685441e+00 -3.13768655e-01 -5.55299222e-01 -1.65187977e-02
-2.22725481e-01 -1.72646627e-01 -7.42981061e-02 -7.26358473e-01
1.10815629e-01 -7.71246910e-01 -2.11400747e-01 -5.72739244e-01
1.77071393e-02 -3.16803426e-01 1.23912954e+00 -6.71160340e-01
1.69097340e+00 -2.41806149e+00 -9.15620744e-01 6.64746225e-01
1.07573532e-01 5.23788393e-01 3.38041842e-01 3.40553999e-01
1.60708100e-01 -3.12236518e-01 -1.79875329e-01 -4.29694541e-02
-3.18388462e-01 2.06936702e-01 -5.70403337e-01 6.03969216e-01
4.12762463e-01 6.32717729e-01 -8.84892941e-01 -3.36206913e-01
7.39978135e-01 4.20665234e-01 -3.28061402e-01 -9.91963223e-02
5.48527181e-01 1.18429907e-01 -6.55315399e-01 5.74577153e-01
1.00293052e+00 5.81395328e-01 -6.23685002e-01 -2.86393225e-01
-3.88057739e-01 3.25133890e-01 -1.68216395e+00 9.85006511e-01
-5.37261009e-01 9.21044946e-01 3.02491579e-02 -9.64659631e-01
1.18607283e+00 2.02240646e-01 5.79636157e-01 -7.94882655e-01
2.02820927e-01 2.12025210e-01 -2.04849452e-01 -7.47599304e-01
9.59301710e-01 1.94975749e-01 6.21808805e-02 -1.99776143e-02
-4.89277542e-01 1.63892321e-02 8.82623643e-02 -2.08974972e-01
9.36300516e-01 -2.18163207e-01 2.18877479e-01 -4.05473292e-01
7.17809856e-01 -1.95449904e-01 5.67775667e-01 8.12120914e-01
-5.14383078e-01 1.96910203e-01 4.02172692e-02 -3.66370320e-01
-8.05536211e-01 -1.07302976e+00 -5.94082832e-01 6.29198074e-01
3.83279979e-01 -4.10764873e-01 -5.76572537e-01 -3.07525784e-01
1.24893829e-01 1.02158952e+00 -3.82741630e-01 -3.73056412e-01
-5.93886912e-01 -9.29931819e-01 8.17656279e-01 5.06641924e-01
6.68996572e-01 -7.89898932e-01 -7.13716924e-01 2.40079880e-01
-5.55972278e-01 -1.25924718e+00 -1.51012391e-01 2.46013954e-01
-8.15435469e-01 -9.21611488e-01 -5.58023900e-02 -3.48178178e-01
5.28055668e-01 6.01804018e-01 7.96947062e-01 3.99735011e-02
-5.82577884e-01 5.51213399e-02 -1.43678784e-01 -8.28098238e-01
-3.73662189e-02 -1.28145278e-01 2.12547332e-02 3.45311075e-01
8.85024309e-01 -5.21600068e-01 -6.00408316e-01 4.39356685e-01
-7.22065032e-01 -5.74821472e-01 3.56556058e-01 5.81628561e-01
6.68610513e-01 7.91303039e-01 7.34161139e-01 -7.50623763e-01
6.06969357e-01 -5.64233005e-01 -6.03201568e-01 -4.35662359e-01
-5.59839725e-01 -1.39600053e-01 7.30520070e-01 -1.24679036e-01
-1.37918544e+00 7.99153075e-02 -4.92532879e-01 -7.68280998e-02
-6.53518498e-01 -3.12752202e-02 -3.64127867e-02 -1.40508577e-01
7.52426624e-01 2.83293128e-01 -1.14769988e-01 -2.37743482e-01
1.42616481e-01 8.24426830e-01 8.08464229e-01 -3.56208503e-01
8.76297593e-01 8.41132879e-01 3.95286381e-01 -1.69613171e+00
-4.14343119e-01 -1.04756677e+00 -3.50221545e-01 -3.66773874e-01
7.22164631e-01 -8.75025451e-01 -9.86906230e-01 7.78194249e-01
-8.47786248e-01 3.12295347e-01 -2.66699880e-01 5.83218932e-01
-1.44146085e-01 7.26880848e-01 -2.89007485e-01 -1.42511857e+00
-3.81574273e-01 -6.79455221e-01 9.83847976e-01 2.86173463e-01
-3.24204296e-01 -6.86025262e-01 -2.29272708e-01 1.47785053e-01
6.07753336e-01 4.12339836e-01 3.32004964e-01 -4.46075231e-01
-2.21068308e-01 -4.75922555e-01 -1.42985985e-01 2.87490755e-01
1.53671235e-01 3.35243136e-01 -1.21779287e+00 2.38300890e-01
2.37145394e-01 6.84285462e-01 1.09002435e+00 5.73709249e-01
1.26536262e+00 2.29743242e-01 -3.98127526e-01 5.43775380e-01
1.34494090e+00 3.43414575e-01 1.06096745e+00 3.33895028e-01
6.20125234e-01 4.96746570e-01 9.37806010e-01 5.26518345e-01
1.14506580e-01 5.04482865e-01 4.67563331e-01 -9.83935073e-02
1.49175674e-01 9.78006646e-02 2.85454512e-01 6.34176075e-01
-3.59780192e-01 -1.38410062e-01 -6.78124011e-01 9.50948536e-01
-1.82007706e+00 -1.31131482e+00 -8.77936304e-01 2.31380439e+00
1.35338083e-01 4.39734161e-01 -5.82307465e-02 9.70233083e-01
7.18127131e-01 1.14700705e-01 -1.40590265e-01 -8.11012745e-01
5.07771894e-02 3.66240412e-01 9.78019714e-01 4.71786648e-01
-1.30917716e+00 5.82197249e-01 6.06792736e+00 1.25858676e+00
-8.01671445e-01 -3.48594904e-01 1.35909930e-01 1.72384664e-01
-8.12436342e-02 -3.46624434e-01 -1.07781839e+00 7.28861451e-01
1.10021091e+00 -4.17482145e-02 9.38187018e-02 8.91704857e-01
6.96737587e-01 -6.01449013e-01 -2.49273866e-01 9.94541049e-01
-1.36956438e-01 -9.41589057e-01 -2.69245170e-02 -9.92160365e-02
3.50254685e-01 -1.98959067e-01 -1.06696919e-01 4.06898074e-02
-3.32233340e-01 -7.67624319e-01 2.44015455e-01 6.72191918e-01
3.90221208e-01 -1.33783329e+00 9.41992283e-01 2.28040949e-01
-1.55614269e+00 1.06156744e-01 -2.10714966e-01 -3.78922433e-01
2.57421762e-01 1.26320481e+00 -7.82526612e-01 6.62356198e-01
7.16032922e-01 5.23086250e-01 -1.94836229e-01 1.10789645e+00
-6.61931410e-02 1.00903285e+00 -7.82457590e-01 1.33058848e-02
9.19112638e-02 -3.01135808e-01 9.69569981e-01 1.41909730e+00
4.71528530e-01 -9.69841424e-03 -1.88651413e-01 4.55108941e-01
3.99128526e-01 8.88957903e-02 -9.80438411e-01 3.13045084e-01
5.61772764e-01 1.28331769e+00 -8.01490366e-01 -4.77310494e-02
-5.15214145e-01 6.69983864e-01 -3.26863974e-01 3.52820545e-01
-1.02501798e+00 -1.10387301e+00 9.74229753e-01 3.08446735e-01
4.12229955e-01 -3.83804888e-01 -6.98616743e-01 -7.72505045e-01
2.71453828e-01 -2.36557990e-01 2.50180990e-01 -2.79668540e-01
-1.00517130e+00 3.63787532e-01 -2.64949083e-01 -1.40690196e+00
-1.87308073e-01 -5.56235969e-01 -9.24321055e-01 9.24520850e-01
-1.82519782e+00 -7.41300583e-01 -5.38444877e-01 6.73514128e-01
3.52273583e-01 1.23768575e-01 6.36597514e-01 7.19673812e-01
-5.77773452e-01 5.96052647e-01 2.26390347e-01 3.10898349e-02
4.59251404e-01 -1.04118538e+00 6.23592377e-01 1.30835187e+00
-9.18684229e-02 3.55075866e-01 8.86927605e-01 -7.15138972e-01
-1.16770065e+00 -1.12769318e+00 1.18654442e+00 -4.46802557e-01
6.84380174e-01 -7.50117525e-02 -1.02966285e+00 -5.91937713e-02
-4.19875234e-01 5.96685782e-02 3.58605772e-01 -1.13117062e-01
2.92309314e-01 -2.76606530e-01 -1.07030916e+00 6.20344341e-01
9.23459351e-01 -6.67954028e-01 -6.56958401e-01 -1.34999037e-01
8.59406739e-02 -4.14909393e-01 -5.71026564e-01 5.88992953e-01
4.41796958e-01 -7.20982075e-01 1.07875848e+00 -4.01058309e-02
-1.99549031e-02 -7.05631554e-01 1.87102377e-01 -8.06423962e-01
-5.13106361e-02 -6.47200227e-01 -3.10108662e-01 1.13991618e+00
7.41558075e-02 -9.53168809e-01 7.23802865e-01 2.88905382e-01
-2.39116669e-01 -5.13537109e-01 -1.13522768e+00 -6.67311728e-01
-7.44996250e-01 -1.18069577e+00 3.92037153e-01 5.55072546e-01
-4.49556917e-01 -2.73170080e-02 -2.84841597e-01 5.97406209e-01
7.42672980e-01 -7.13492259e-02 7.16425657e-01 -1.40516043e+00
2.90466011e-01 -2.56733835e-01 -6.41552269e-01 -8.14823329e-01
-5.74421771e-02 -3.69544894e-01 1.36328727e-01 -1.00671220e+00
-6.54986739e-01 -2.47960582e-01 -4.13272947e-01 -9.41707790e-02
-3.22324455e-01 3.18060786e-01 -4.96054828e-01 -1.98063806e-01
-2.21669838e-01 4.94840890e-01 8.41018975e-01 2.31369838e-01
-2.86260366e-01 5.26353776e-01 -1.70825705e-01 1.06832743e+00
7.68561125e-01 -4.07232761e-01 -3.90942097e-01 8.82785097e-02
2.59094089e-01 -3.77486497e-01 5.42962968e-01 -1.16638637e+00
2.09227994e-01 -5.23946248e-03 1.70917019e-01 -9.37795877e-01
-2.83690169e-02 -1.10043895e+00 -5.69746755e-02 3.57280791e-01
3.48401457e-01 5.73793873e-02 6.18457437e-01 7.95486152e-01
-4.38612819e-01 -2.76207328e-01 5.43904722e-01 4.81307358e-01
-1.30436110e+00 -2.86910273e-02 -8.60813439e-01 -2.05745567e-02
1.10083115e+00 -4.85012889e-01 -2.16829732e-01 -3.06491017e-01
-5.92354417e-01 7.67392069e-02 1.12675540e-02 2.75468737e-01
4.70409304e-01 -1.15007329e+00 -5.46320856e-01 5.53921223e-01
1.20732464e-01 -2.11294964e-01 5.42138815e-01 8.78928185e-01
-5.47755182e-01 2.19917864e-01 -9.15391669e-02 -6.35308206e-01
-1.27987635e+00 2.12092966e-01 1.80387929e-01 -7.51633644e-02
-8.11258972e-01 4.67762828e-01 -2.35325485e-01 3.12160641e-01
4.86917868e-02 -4.91802216e-01 -6.30920053e-01 9.61491168e-02
5.77707887e-01 1.10595489e+00 6.17451787e-01 -6.89967930e-01
-5.82605124e-01 6.02244318e-01 2.27165923e-01 3.16792578e-01
7.58693755e-01 -4.40946370e-01 2.18950242e-01 2.85349071e-01
8.84022295e-01 5.49944282e-01 -8.96371067e-01 4.76802737e-02
-1.15990853e-02 -4.53310668e-01 3.98467094e-01 -1.77417293e-01
-8.48016500e-01 7.02993572e-01 6.88428581e-01 3.77951413e-01
1.39407468e+00 -7.07293570e-01 1.28210962e+00 4.73103195e-01
1.17995143e-01 -1.47654068e+00 -6.78815901e-01 5.96828401e-01
3.11128825e-01 -1.21606338e+00 -2.48293281e-01 -6.91962540e-01
-2.99386114e-01 1.08703256e+00 3.50926727e-01 -3.56817126e-01
9.52141643e-01 6.45387888e-01 2.44817371e-03 -1.43364698e-01
-2.04948664e-01 -6.66303039e-01 5.43185711e-01 5.36927462e-01
-9.96428262e-03 1.08463094e-01 -6.51929200e-01 6.52620673e-01
-7.60342851e-02 -2.42767751e-01 4.04422969e-01 7.42075443e-01
-5.52151322e-01 -6.44562066e-01 -3.03644538e-01 6.75433993e-01
-7.99829543e-01 -5.29033169e-02 3.40465724e-01 6.91029131e-01
2.50707328e-01 1.32833493e+00 3.70383322e-01 -3.45862448e-01
7.93868840e-01 2.17692703e-01 -2.10613504e-01 7.86276758e-02
-3.46367359e-01 -9.11579803e-02 3.80870104e-01 -7.42824435e-01
-2.32550517e-01 -6.00008607e-01 -1.52363694e+00 -3.95774126e-01
-5.79878032e-01 2.88616270e-01 7.65366554e-01 1.08539534e+00
5.44871151e-01 7.63265014e-01 6.51686609e-01 -4.25002754e-01
-5.87070361e-02 -5.13065934e-01 -6.89851046e-01 6.12184703e-01
5.39366364e-01 -8.47876906e-01 -8.15373898e-01 -4.26094420e-02] | [8.00756549835205, -0.8205987811088562] |
146d8e57-8801-4097-b31f-33f08714ae53 | deep-supervision-for-pancreatic-cyst | 1706.07346 | null | http://arxiv.org/abs/1706.07346v1 | http://arxiv.org/pdf/1706.07346v1.pdf | Deep Supervision for Pancreatic Cyst Segmentation in Abdominal CT Scans | Automatic segmentation of an organ and its cystic region is a prerequisite of
computer-aided diagnosis. In this paper, we focus on pancreatic cyst
segmentation in abdominal CT scan. This task is important and very useful in
clinical practice yet challenging due to the low contrast in boundary, the
variability in location, shape and the different stages of the pancreatic
cancer. Inspired by the high relevance between the location of a pancreas and
its cystic region, we introduce extra deep supervision into the segmentation
network, so that cyst segmentation can be improved with the help of relatively
easier pancreas segmentation. Under a reasonable transformation function, our
approach can be factorized into two stages, and each stage can be efficiently
optimized via gradient back-propagation throughout the deep networks. We
collect a new dataset with 131 pathological samples, which, to the best of our
knowledge, is the largest set for pancreatic cyst segmentation. Without human
assistance, our approach reports a 63.44% average accuracy, measured by the
Dice-S{\o}rensen coefficient (DSC), which is higher than the number (60.46%)
without deep supervision. | ['Elliot K. Fishman', 'Yuyin Zhou', 'Lingxi Xie', 'Alan L. Yuille'] | 2017-06-22 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [-1.02770068e-01 3.28948051e-01 -1.79800466e-01 -3.96619737e-01
-3.54039311e-01 -4.46489125e-01 1.47087917e-01 4.36454773e-01
-6.58389091e-01 4.22558188e-01 3.61820422e-02 -2.01009676e-01
-9.34533700e-02 -4.76513892e-01 -6.32922888e-01 -8.58774543e-01
-2.92248368e-01 6.60151839e-01 3.40894610e-01 2.90778458e-01
-6.61497712e-02 4.79070157e-01 -7.23169088e-01 -5.46377897e-02
1.36783719e+00 8.74583185e-01 4.85341996e-01 3.69034767e-01
-2.97510713e-01 4.94653821e-01 -3.08561236e-01 -3.53045076e-01
2.61770874e-01 -6.69832647e-01 -7.16479182e-01 2.00472638e-01
9.54270214e-02 -2.95487374e-01 -2.19752073e-01 1.30576444e+00
3.35837722e-01 -2.22411260e-01 6.57832086e-01 -6.76094949e-01
-3.43631119e-01 8.93174231e-01 -3.42143536e-01 5.24875037e-02
-1.58971012e-01 8.65759701e-02 4.69261795e-01 -3.48060191e-01
6.55700803e-01 5.87018967e-01 8.40738893e-01 4.15819943e-01
-9.85164464e-01 -5.30650795e-01 1.10050760e-01 -1.97113663e-01
-1.07058442e+00 1.27782971e-01 5.14913499e-01 -4.50034142e-01
2.26251081e-01 -4.87443693e-02 1.17961490e+00 6.32772923e-01
3.48473430e-01 8.22732270e-01 9.89675999e-01 -3.20222735e-01
1.23090163e-01 1.73895821e-01 1.38884470e-01 7.70657897e-01
5.23998559e-01 -1.43243209e-01 2.94761896e-01 7.69263059e-02
9.94463146e-01 3.76035124e-01 -5.84596157e-01 -6.51981533e-01
-1.41358161e+00 8.34528029e-01 9.03127193e-01 3.77548218e-01
-4.14128840e-01 -1.16674669e-01 3.22140247e-01 2.58587729e-02
2.72031724e-02 4.35808748e-01 -2.60717213e-01 1.48792297e-01
-9.80619431e-01 -2.80972451e-01 1.23616874e+00 7.78388917e-01
-2.96216868e-02 -2.72775173e-01 1.30081952e-01 6.40971899e-01
6.29151165e-01 3.04460943e-01 7.54294455e-01 -7.58426309e-01
1.98289126e-01 5.79509437e-01 -2.52796173e-01 -4.94870424e-01
-8.28077435e-01 -6.09386384e-01 -1.08552229e+00 7.18720257e-02
9.68198895e-01 -3.86202842e-01 -1.14173555e+00 1.31400907e+00
3.27471763e-01 -2.84901611e-03 -7.76089877e-02 1.27785289e+00
9.97615755e-01 2.14903235e-01 2.26836473e-01 -2.42302999e-01
1.47762978e+00 -1.08022881e+00 -6.76441133e-01 -1.98317524e-02
8.43786418e-01 -6.21204138e-01 6.78585291e-01 4.35904920e-01
-8.95747006e-01 7.93303922e-02 -9.27409828e-01 2.70339549e-01
-3.78436148e-02 2.21051008e-01 8.75382185e-01 5.42528152e-01
-8.84305239e-01 5.79007506e-01 -1.25301051e+00 -5.51184356e-01
6.42494798e-01 5.20002842e-01 -3.50220859e-01 2.74645705e-02
-8.89025688e-01 1.10757589e+00 4.60656643e-01 1.04591317e-01
-4.84749317e-01 -6.68686748e-01 -8.01681578e-01 1.18754273e-02
3.98073494e-01 -7.70029008e-01 1.15917420e+00 -9.02730286e-01
-1.68145204e+00 8.72424662e-01 1.95835829e-01 -5.75536788e-01
9.55854475e-01 1.12177119e-01 -2.50111409e-02 3.43066990e-01
-2.23105729e-01 6.99045956e-01 3.86497319e-01 -1.22790360e+00
-4.96841371e-01 -4.82085496e-01 -3.19991320e-01 1.54491499e-01
-5.96749745e-02 -6.49310946e-02 -6.82105124e-01 -6.31989241e-01
6.10647798e-01 -1.19664609e+00 -6.35818481e-01 4.90103245e-01
-2.81888574e-01 1.20749645e-01 4.03085798e-01 -9.76819813e-01
8.31855476e-01 -2.04771495e+00 7.29574170e-03 3.61654133e-01
3.55674684e-01 3.17333877e-01 3.33102316e-01 -2.08703205e-01
1.86025739e-01 4.53022718e-02 -4.67272401e-01 -1.18499026e-01
-1.81630999e-01 2.78430790e-01 2.94391721e-01 8.60915422e-01
-1.45085916e-01 1.04320395e+00 -8.94119859e-01 -7.48649657e-01
2.48487771e-01 4.12234336e-01 -3.83781910e-01 9.81853679e-02
-5.80496751e-02 5.94902635e-01 -5.25643885e-01 5.89145064e-01
7.21922576e-01 -5.62269449e-01 4.62755710e-01 -7.33671188e-02
-3.33566740e-02 -5.23723736e-02 -9.61284876e-01 1.83101535e+00
-2.39390805e-01 3.32557738e-01 4.00937974e-01 -9.04383004e-01
9.16101515e-01 3.55502218e-01 8.30340803e-01 -4.93215412e-01
3.06324095e-01 5.78102469e-01 5.22407949e-01 -4.55647051e-01
2.20172927e-02 -2.40472317e-01 2.10318625e-01 3.39424074e-01
4.05083299e-02 -2.75013626e-01 3.32386911e-01 -8.54457468e-02
7.25530922e-01 -6.56786337e-02 4.26538855e-01 -5.19497871e-01
3.35471421e-01 1.75087065e-01 5.70285499e-01 5.54375112e-01
-4.58545089e-01 8.91786098e-01 6.67862833e-01 -4.12807465e-01
-8.08537960e-01 -8.92522871e-01 -6.39180362e-01 3.17057222e-01
5.30148983e-01 1.62133634e-01 -8.10945272e-01 -9.00850773e-01
1.12256028e-01 2.29948357e-01 -4.99620497e-01 5.62872961e-02
-8.35230529e-01 -1.04548025e+00 4.55250263e-01 5.40322840e-01
4.25473005e-01 -1.06090808e+00 -6.01138413e-01 2.63728708e-01
2.63901576e-02 -1.08629322e+00 -3.89735430e-01 5.81934869e-01
-1.43799126e+00 -1.25139034e+00 -1.24464810e+00 -1.20372045e+00
1.05986953e+00 -1.17274106e-01 9.57209706e-01 1.14228889e-01
-2.22450241e-01 -2.35552579e-01 -2.86245733e-01 -1.99187994e-01
-5.59329689e-01 1.00749396e-01 -4.20870543e-01 -5.57667792e-01
1.28120497e-01 -2.92725265e-01 -7.92456985e-01 3.78623664e-01
-7.41005957e-01 -1.95876323e-03 8.47538054e-01 9.22332287e-01
6.92362726e-01 -8.84745345e-02 7.47557431e-02 -8.60428035e-01
2.98874319e-01 -5.54000914e-01 -6.56718373e-01 9.32767913e-02
-5.79033732e-01 -6.35308623e-02 7.06995845e-01 -5.65239310e-01
-7.82387972e-01 2.96381861e-01 -1.26222655e-01 -1.21029235e-01
-3.24994624e-01 3.80298853e-01 1.33908197e-01 -3.48144054e-01
2.38308370e-01 1.54046252e-01 3.82379889e-01 -5.01499236e-01
9.34900194e-02 3.54743212e-01 4.29485828e-01 -1.25284299e-01
2.93585896e-01 4.74045783e-01 8.73827487e-02 -3.98218304e-01
-5.22139549e-01 -4.27528679e-01 -8.32245350e-01 -3.83165814e-02
8.70431423e-01 -6.11506104e-01 -6.74817085e-01 4.65066165e-01
-7.56310582e-01 -4.98860091e-01 -2.80223995e-01 1.02718747e+00
-3.02221209e-01 6.63410783e-01 -1.02446103e+00 -1.50222406e-01
-5.97306371e-01 -1.40758300e+00 5.73812664e-01 4.27335232e-01
-7.68494681e-02 -1.26035511e+00 -1.74773514e-01 1.08505204e-01
4.54141647e-01 4.45747018e-01 6.88222528e-01 -9.20541048e-01
-3.80812705e-01 -3.21471214e-01 -4.17635322e-01 1.56352878e-01
1.77143142e-01 -1.10152870e-01 -4.12056029e-01 -1.79048091e-01
1.02820382e-01 7.78395161e-02 9.30746496e-01 9.35364366e-01
1.14591682e+00 3.25102359e-02 -4.21320438e-01 1.19654524e+00
1.39609766e+00 3.40526849e-01 2.67862380e-01 5.81869662e-01
6.96195185e-01 2.79785335e-01 4.35519308e-01 3.73132735e-01
3.56542170e-01 3.62297267e-01 6.49075091e-01 -3.74980539e-01
-2.14564309e-01 6.91056401e-02 6.69697076e-02 8.65163922e-01
7.20893294e-02 -2.02154264e-01 -9.77422357e-01 4.67080593e-01
-1.60946047e+00 -3.74951273e-01 -4.63971496e-01 1.96841371e+00
7.04191267e-01 -2.47561596e-02 5.45503162e-02 -3.41421664e-01
7.79610515e-01 -1.83811545e-01 -5.59647202e-01 -2.41899505e-01
9.44526568e-02 -9.90540609e-02 8.10419023e-01 3.46831739e-01
-1.25984716e+00 5.37442923e-01 6.24451351e+00 5.17179012e-01
-1.35413420e+00 -1.28375322e-01 4.97086823e-01 2.98606828e-02
3.61386687e-02 -1.48863316e-01 -6.06009483e-01 8.59951973e-01
4.92505640e-01 1.22261070e-01 2.26125225e-01 6.45401418e-01
3.51389907e-02 -2.52242386e-01 -1.13368118e+00 6.50410354e-01
5.92690110e-02 -9.76141453e-01 -3.06785434e-01 1.60911798e-01
6.39711857e-01 2.19989002e-01 -1.78530216e-01 1.01861991e-01
2.15458155e-01 -9.87868190e-01 3.34414005e-01 4.71142501e-01
5.26663005e-01 -5.01325607e-01 1.28345335e+00 4.01229858e-01
-8.74659240e-01 1.07602760e-01 -1.32906273e-01 4.74161863e-01
2.78004974e-01 7.60109127e-01 -1.02979958e+00 1.90125197e-01
6.09017611e-01 5.67401946e-01 -2.42921829e-01 1.59920084e+00
-2.70069242e-01 5.15269816e-01 -6.49438381e-01 -1.46136925e-01
2.53470749e-01 -5.22760451e-01 4.66183901e-01 1.31099463e+00
4.79454994e-01 -8.22379161e-03 2.83921570e-01 9.47268903e-01
-1.50195599e-01 3.38601530e-01 -1.45716995e-01 2.05478474e-01
1.15183897e-01 1.39619708e+00 -1.13833320e+00 -2.34767705e-01
-2.87256986e-01 8.17696095e-01 -3.12674753e-02 4.83398251e-02
-8.53867650e-01 -2.96604902e-01 8.53695422e-02 -3.52832228e-02
3.85178149e-01 4.72695567e-02 -5.26134431e-01 -1.13844299e+00
1.75286144e-01 -5.41868925e-01 4.73365724e-01 -1.95149913e-01
-1.12927890e+00 5.25595248e-01 -3.56683791e-01 -1.29833424e+00
-3.99806462e-02 -7.52186358e-01 -6.79852724e-01 6.42291427e-01
-1.58532000e+00 -8.52288365e-01 -6.03045225e-01 2.40851387e-01
1.39168605e-01 1.65627792e-01 7.12668598e-01 2.86219835e-01
-4.63356167e-01 5.04378855e-01 2.82762587e-01 4.00346845e-01
7.20041871e-01 -1.50582743e+00 -2.34154806e-01 5.79570234e-01
-4.98391092e-01 4.74648774e-01 3.97664189e-01 -6.28672898e-01
-1.15390313e+00 -1.05255592e+00 6.71274126e-01 9.98497009e-02
8.15074801e-01 5.91865778e-02 -1.02447760e+00 6.27586663e-01
2.14987025e-02 3.78746986e-01 7.07460284e-01 -3.27000052e-01
2.52024919e-01 7.16794655e-02 -1.44156969e+00 5.89705408e-01
6.37021065e-01 3.12781543e-01 -4.64693457e-01 3.32250386e-01
5.42989910e-01 -8.65153193e-01 -1.48616564e+00 5.66700697e-01
5.41428030e-01 -8.37485492e-01 7.87624598e-01 -1.03351017e-02
3.95050406e-01 -1.13234766e-01 4.12534088e-01 -1.37173736e+00
-2.20104322e-01 -2.38771349e-01 -2.02694517e-02 7.70029843e-01
4.23117071e-01 -6.79706633e-01 9.77449894e-01 7.30552554e-01
-3.92615646e-01 -8.97163212e-01 -6.55780613e-01 -4.26242858e-01
4.47779804e-01 8.16286951e-02 3.53108943e-01 1.07904780e+00
2.16344625e-01 -5.65128148e-01 7.96435326e-02 1.06627315e-01
7.44146287e-01 3.85567605e-01 3.66848856e-01 -1.30209005e+00
-9.28299874e-02 -7.46864796e-01 -2.20811799e-01 -1.06683910e+00
-2.15686485e-01 -1.19492424e+00 1.02647662e-01 -1.69335151e+00
4.74428445e-01 -6.49608314e-01 -2.45533168e-01 3.70724589e-01
-2.16545805e-01 8.12534466e-02 3.01748395e-01 4.06320512e-01
-5.52334011e-01 3.29771072e-01 1.95839822e+00 4.69026566e-02
-3.54408324e-01 1.70935988e-01 -5.66326320e-01 1.01983809e+00
9.94249403e-01 -5.76912463e-01 2.29623109e-01 -3.07362884e-01
-3.67711693e-01 3.37518513e-01 1.63134560e-01 -7.27029979e-01
4.50165451e-01 1.52222037e-01 5.38345873e-01 -4.51444119e-01
-2.39199623e-02 -1.07200873e+00 3.73524167e-02 1.09991431e+00
-2.00664550e-01 -4.38489109e-01 8.86900797e-02 2.75489271e-01
-3.56639147e-01 -5.74099898e-01 9.18489218e-01 -4.84309644e-01
-4.28191751e-01 3.69891673e-01 -4.14250076e-01 -6.05803728e-02
1.13438606e+00 -2.92085260e-01 -1.38630539e-01 -4.68401685e-02
-9.71903741e-01 4.82362807e-01 5.85331202e-01 -7.53994659e-02
4.11816955e-01 -9.80407834e-01 -6.31044865e-01 2.56992131e-01
-1.25693843e-01 5.14195025e-01 1.16916634e-01 1.57900846e+00
-1.12960362e+00 3.53778422e-01 -1.28025055e-01 -8.80911767e-01
-1.07148576e+00 2.38170981e-01 6.04710937e-01 -5.75709105e-01
-1.16539264e+00 5.69018364e-01 1.46362763e-02 -5.76642096e-01
3.71677935e-01 -8.00529122e-01 -3.58908296e-01 -8.07952508e-02
1.15997270e-01 1.91724136e-01 -5.62952347e-02 -4.22964364e-01
-2.99008071e-01 5.81465483e-01 -1.41085163e-01 5.14032841e-01
1.31513691e+00 5.84483035e-02 -2.08089903e-01 -1.09988011e-01
9.62264895e-01 -1.38564706e-01 -1.28409421e+00 -1.99429110e-01
2.46930700e-02 -1.76178321e-01 1.81861430e-01 -9.12813962e-01
-1.53362930e+00 6.39538050e-01 6.75102174e-01 1.11485757e-01
8.80913138e-01 -1.35007259e-02 1.04404616e+00 1.77451953e-01
1.77762300e-01 -7.23140001e-01 -3.78400505e-01 5.06693244e-01
5.05008042e-01 -1.61513925e+00 8.68220478e-02 -5.13070583e-01
-7.33751953e-01 1.32603085e+00 4.98634994e-01 -4.31988031e-01
6.79945409e-01 5.23155868e-01 4.32307392e-01 8.86790268e-03
-1.87531754e-01 4.79992628e-02 2.24119678e-01 3.84941101e-01
5.56734860e-01 2.05671355e-01 -5.63494384e-01 6.60600722e-01
-8.28433335e-02 1.46060988e-01 3.97952437e-01 7.08046138e-01
-5.24500608e-01 -8.03246021e-01 -3.39605123e-01 6.14548206e-01
-7.81989098e-01 1.89226512e-02 -5.10300286e-02 1.11968923e+00
3.67623530e-02 4.61750507e-01 1.06286854e-01 2.44391426e-01
4.12805192e-02 -2.04748720e-01 5.10080695e-01 -3.57307494e-01
-8.22507143e-01 3.94467115e-01 -2.16944411e-01 -3.81993175e-01
-3.31196547e-01 -7.89925516e-01 -1.80639541e+00 1.01705112e-01
-4.26923662e-01 1.65791824e-01 8.74398172e-01 7.66921043e-01
-9.65477824e-02 5.99300504e-01 3.78939509e-01 -7.37286806e-01
-6.58993125e-01 -8.98116648e-01 -7.85368383e-01 5.31461477e-01
1.41461462e-01 -3.37246120e-01 -5.46741784e-01 -2.03753226e-02] | [14.585638999938965, -2.7163445949554443] |
bdb17e0f-095b-415c-a762-acbcdfc96c67 | fairface-face-attribute-dataset-for-balanced | 1908.04913 | null | https://arxiv.org/abs/1908.04913v1 | https://arxiv.org/pdf/1908.04913v1.pdf | FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age | Existing public face datasets are strongly biased toward Caucasian faces, and other races (e.g., Latino) are significantly underrepresented. This can lead to inconsistent model accuracy, limit the applicability of face analytic systems to non-White race groups, and adversely affect research findings based on such skewed data. To mitigate the race bias in these datasets, we construct a novel face image dataset, containing 108,501 images, with an emphasis of balanced race composition in the dataset. We define 7 race groups: White, Black, Indian, East Asian, Southeast Asian, Middle East, and Latino. Images were collected from the YFCC-100M Flickr dataset and labeled with race, gender, and age groups. Evaluations were performed on existing face attribute datasets as well as novel image datasets to measure generalization performance. We find that the model trained from our dataset is substantially more accurate on novel datasets and the accuracy is consistent between race and gender groups. | ['Kimmo Kärkkäinen', 'Jungseock Joo'] | 2019-08-14 | null | null | null | null | ['facial-attribute-classification'] | ['computer-vision'] | [ 4.40910906e-02 -7.47667253e-02 -6.10224128e-01 -9.20437217e-01
-2.18427703e-01 -3.93623948e-01 4.95001882e-01 -2.39743993e-01
-3.04692835e-01 7.12392330e-01 5.23907423e-01 -2.44910449e-01
1.35174431e-02 -7.71188200e-01 -2.30017468e-01 -3.30575824e-01
2.23826200e-01 2.48457298e-01 -8.83702040e-01 1.68264225e-01
2.18617201e-01 5.47535181e-01 -1.60813069e+00 2.53875673e-01
7.96449721e-01 1.00116956e+00 -9.91522431e-01 1.87081367e-01
1.45927906e-01 8.09174478e-01 -5.37425637e-01 -6.64431572e-01
4.63580668e-01 -6.19767368e-01 -5.38527608e-01 -1.55475512e-01
1.14650941e+00 -4.77908701e-01 -4.10292238e-01 1.05357313e+00
5.88355660e-01 -1.09894983e-01 9.69907463e-01 -1.68454278e+00
-1.21831071e+00 6.25668705e-01 -1.36067665e+00 8.12781453e-02
3.15793335e-01 -5.40109836e-02 4.87050056e-01 -1.13573897e+00
7.33328402e-01 1.71187305e+00 9.45786417e-01 9.11076188e-01
-1.26443326e+00 -1.60584915e+00 1.24843061e-01 5.76797500e-02
-1.86752963e+00 -1.10390663e+00 6.68159127e-01 -5.39804459e-01
1.85951471e-01 2.05788091e-01 3.71672273e-01 1.23420727e+00
-4.97831441e-02 9.60078761e-02 1.61679661e+00 -1.27206087e-01
1.52511254e-01 1.71503335e-01 8.15383047e-02 5.44648290e-01
3.67758512e-01 2.98279017e-01 -8.68284285e-01 -5.76979995e-01
3.96386027e-01 3.37826763e-03 6.42535761e-02 -3.21373530e-02
-7.67798722e-01 7.06151128e-01 3.05036306e-01 -2.32118711e-01
-1.51280597e-01 -1.89914867e-01 1.31747022e-01 3.90676975e-01
6.99384689e-01 1.74388215e-02 -2.95437157e-01 2.97263302e-02
-8.33995283e-01 2.84505665e-01 6.34647965e-01 8.22869658e-01
8.62685263e-01 2.07576334e-01 -1.35461828e-02 1.26355624e+00
2.58917242e-01 8.57682168e-01 1.68732554e-01 -1.27740562e+00
1.06569141e-01 7.07584202e-01 -1.63404927e-01 -9.90897775e-01
-5.21042168e-01 -1.07699953e-01 -8.92395020e-01 3.70317727e-01
6.51880324e-01 -2.00049266e-01 -1.24520338e+00 2.11533999e+00
4.12103444e-01 -1.03017874e-01 -1.04270048e-01 6.11338496e-01
9.72351134e-01 2.19473746e-02 5.77416003e-01 -1.96945742e-01
1.20302677e+00 -2.68584579e-01 -4.29100603e-01 -1.42887384e-01
3.04341555e-01 -5.93025267e-01 9.36422467e-01 5.25655318e-03
-7.72058964e-01 -4.89523381e-01 -5.11216462e-01 2.16508552e-01
-4.08575416e-01 8.91674906e-02 7.18407452e-01 1.47313929e+00
-1.02193785e+00 1.56860292e-01 -2.36229539e-01 -5.51313281e-01
1.20735061e+00 3.81057024e-01 -7.25075305e-01 -3.40228856e-01
-6.69144809e-01 5.94986081e-01 -1.50620311e-01 -1.50163785e-01
-6.44888878e-01 -1.35159409e+00 -8.90998781e-01 -2.28037030e-01
-2.40677949e-02 -2.85227239e-01 7.74238408e-01 -1.63561606e+00
-8.92519295e-01 1.33717692e+00 -3.99609059e-01 6.48121834e-02
3.92879099e-01 3.01307768e-01 -1.01267338e+00 4.67508938e-03
2.28264958e-01 1.04777777e+00 9.36038792e-01 -1.33258486e+00
-5.82523704e-01 -1.08796823e+00 -2.97252834e-01 -4.21819240e-02
-4.57885742e-01 4.72533375e-01 1.28569797e-01 -6.22662306e-01
-4.24572937e-02 -1.11262214e+00 3.34288239e-01 1.84908569e-01
-1.99356437e-01 -3.10220923e-02 8.87142777e-01 -7.54734516e-01
1.08461654e+00 -2.33256340e+00 -4.50903058e-01 4.17304277e-01
3.11892420e-01 -2.01748967e-01 -2.01940820e-01 -2.93493003e-01
-5.00654995e-01 4.77182388e-01 2.35245042e-02 1.72381446e-01
-3.00208569e-01 -7.85182267e-02 -5.64415976e-02 6.92221463e-01
3.09626341e-01 5.13175547e-01 -4.87848282e-01 -6.66647136e-01
-1.92693844e-01 3.35311025e-01 -7.33227849e-01 -2.68827260e-01
3.43310624e-01 4.53714162e-01 -1.11362688e-01 1.39976203e+00
1.15588093e+00 2.14304477e-01 1.15144640e-01 -1.53958827e-01
5.15869819e-02 -3.92072737e-01 -7.84373760e-01 8.42530549e-01
-6.60629794e-02 5.22932470e-01 2.58613914e-01 -4.04824018e-01
1.16931117e+00 -9.41841081e-02 6.47383630e-01 -6.54366970e-01
2.60964185e-01 9.99804512e-02 4.27809417e-01 -2.06208736e-01
5.05766034e-01 -3.11280757e-01 1.51711002e-01 5.62696695e-01
-2.13803053e-01 3.27454150e-01 1.66312128e-01 -2.77877636e-02
5.74856699e-01 -1.09212063e-01 -1.60083193e-02 -4.70617592e-01
2.70892113e-01 -5.50717674e-02 9.73834574e-01 6.43487930e-01
-7.93767095e-01 7.55019248e-01 6.72092199e-01 -4.43840653e-01
-9.90073740e-01 -1.27587986e+00 -5.61610937e-01 1.39569187e+00
-2.03968108e-01 -1.19702145e-01 -5.79333901e-01 -7.41611481e-01
4.37285364e-01 5.21047473e-01 -9.70491111e-01 -5.14971137e-01
-2.19571605e-01 -1.04183149e+00 9.19463694e-01 6.19816065e-01
6.61793113e-01 -6.59460485e-01 -7.35390233e-03 -5.93587756e-01
1.21963620e-01 -8.99912298e-01 -6.02929294e-01 -6.69780433e-01
-4.42876548e-01 -1.43604529e+00 -8.63098323e-01 -6.77930236e-01
8.76073480e-01 -3.44312824e-02 1.32312071e+00 1.35503158e-01
-4.06451672e-01 6.13286853e-01 -2.02764347e-01 -8.75458062e-01
-3.01054239e-01 -8.67008269e-02 3.80202293e-01 4.37808782e-01
9.33520377e-01 -2.46120960e-01 -6.32757008e-01 6.03633583e-01
-4.17240828e-01 -2.09964871e-01 1.73604652e-01 4.92297232e-01
1.44770294e-01 -1.77235454e-01 1.10018492e+00 -1.19474173e+00
1.40965819e-01 -7.41651833e-01 -1.11112781e-01 3.34700078e-01
-8.53473902e-01 -7.34565854e-01 2.21947432e-01 -5.70773900e-01
-1.53685832e+00 4.91995970e-03 2.48370409e-01 -1.48609400e-01
-2.81724006e-01 2.35843599e-01 -3.06240976e-01 -2.84337759e-01
8.66845965e-01 -3.08841407e-01 3.15781027e-01 -2.67447025e-01
-7.79806972e-02 1.08661389e+00 5.69580734e-01 -8.12519550e-01
6.96729124e-01 8.37213695e-01 -1.34513423e-01 -1.06262481e+00
-5.67690611e-01 8.44602510e-02 -5.42400539e-01 -4.84223217e-01
4.88641858e-01 -1.26331079e+00 -6.57777905e-01 8.67484152e-01
-2.49841616e-01 -1.50713623e-01 -7.44133592e-02 6.33698404e-01
-2.77847387e-02 -8.97739604e-02 -4.52571422e-01 -8.66395473e-01
-2.32274577e-01 -6.31359339e-01 6.87491298e-01 6.21408880e-01
-5.55834591e-01 -4.87815827e-01 -3.20439100e-01 5.66388786e-01
3.82095665e-01 4.06765491e-01 1.13113391e+00 -3.16490144e-01
1.43121764e-01 -1.85187697e-01 -7.80058563e-01 2.13388890e-01
4.25643772e-01 6.45664930e-01 -1.15642405e+00 -3.03940624e-01
-5.41834772e-01 -6.99888289e-01 7.95782268e-01 3.50951433e-01
1.44342291e+00 9.14216638e-02 -3.79990131e-01 6.79029703e-01
1.04063249e+00 3.11221063e-01 4.87611204e-01 -2.75734395e-01
6.19054437e-01 7.66202390e-01 3.18194211e-01 5.12616754e-01
6.07058942e-01 8.00005421e-02 -1.75688975e-02 -1.02810882e-01
-2.67843813e-01 -2.92032391e-01 1.61094204e-01 -9.98022631e-02
-1.64967164e-01 2.52819538e-01 -1.14286530e+00 4.64794964e-01
-1.04874635e+00 -1.12599313e+00 1.41108394e-01 2.22998071e+00
8.56981337e-01 -3.74768734e-01 3.82798731e-01 -1.88669041e-01
9.53581989e-01 1.14983238e-01 -8.25293243e-01 -3.09453070e-01
-4.47613686e-01 3.46250772e-01 4.71421629e-01 6.10164516e-02
-9.69269991e-01 6.85224950e-01 7.50586462e+00 3.65545601e-01
-1.25489700e+00 -8.38064700e-02 1.40961921e+00 -6.51823461e-01
-3.03908050e-01 -1.67498261e-01 -7.08696902e-01 3.60047013e-01
9.11577463e-01 -5.08691907e-01 5.64386606e-01 8.19883049e-01
-6.06846577e-03 -1.82001501e-01 -9.74431634e-01 1.29446888e+00
4.45735723e-01 -9.45468545e-01 5.08821830e-02 2.73627520e-01
1.00678074e+00 -2.03641146e-01 5.29651403e-01 5.08563280e-01
2.58692801e-01 -1.59724772e+00 7.27743804e-01 2.39571869e-01
1.46100831e+00 -9.24465358e-01 2.38921493e-01 -2.47021317e-01
-8.12666893e-01 -5.40503919e-01 -4.29072022e-01 -2.90641069e-01
-5.04104793e-01 2.86506504e-01 -1.48573577e-01 4.60262708e-02
1.16186941e+00 6.94608390e-01 -9.89354730e-01 4.53119814e-01
5.31518936e-01 6.74808621e-01 -2.58566141e-01 4.67324972e-01
-3.38495076e-01 -1.84585795e-01 -1.83994114e-01 4.33939695e-01
3.83426875e-01 2.03855544e-01 -1.29828427e-03 7.98150182e-01
-6.38128102e-01 6.78896382e-02 -6.88619971e-01 -2.19345018e-01
7.39286065e-01 1.31000495e+00 -5.57114422e-01 -2.24980026e-01
-8.08661640e-01 1.52005553e-01 1.72978431e-01 7.01550424e-01
-6.95980072e-01 -4.81578410e-02 1.11407483e+00 2.54585177e-01
-5.27202725e-01 1.00384735e-01 -6.14357114e-01 -9.75609362e-01
-3.54683071e-01 -1.35399389e+00 7.93542683e-01 -6.58521950e-01
-1.52986956e+00 2.42385507e-01 3.17553096e-02 -6.40136421e-01
5.98563217e-02 -6.88138127e-01 -3.54035139e-01 1.08179724e+00
-1.17083192e+00 -1.27665281e+00 -5.73684454e-01 7.32686639e-01
-1.79344192e-01 -6.69413805e-01 6.39518559e-01 6.62218451e-01
-8.45048666e-01 1.11652541e+00 -1.90029711e-01 6.03268743e-01
1.25989878e+00 -6.60628736e-01 1.40136071e-02 3.40821266e-01
-2.01761141e-01 8.55159163e-01 7.44282156e-02 -7.78405547e-01
-1.23406780e+00 -1.11956275e+00 5.06753385e-01 -7.08190620e-01
1.88932702e-01 -5.82680963e-02 -5.85038006e-01 8.90498161e-01
-7.10388646e-02 1.32090479e-01 1.32733631e+00 6.02643549e-01
-1.07001936e+00 -3.39449823e-01 -1.71838129e+00 6.69587910e-01
1.26769543e+00 -6.87005997e-01 -4.23756130e-02 -5.61381541e-02
-3.78439575e-02 -2.27815151e-01 -1.04493213e+00 6.25329196e-01
1.08589816e+00 -1.02732003e+00 8.73029947e-01 -1.00020218e+00
5.24371266e-01 1.29602151e-02 -2.17851430e-01 -1.17528176e+00
-4.15602148e-01 -8.54823664e-02 8.78578946e-02 1.58189750e+00
3.88137490e-01 -8.25319648e-01 1.03622508e+00 1.15905464e+00
3.74772817e-01 -4.27284032e-01 -8.82462025e-01 -6.26184583e-01
6.27416968e-01 -6.65898174e-02 1.20487106e+00 1.23543155e+00
-3.05244505e-01 -8.75478983e-03 -1.80528864e-01 -1.22027166e-01
8.40507209e-01 2.50034869e-01 9.90639508e-01 -1.43297553e+00
6.13328695e-01 -7.07693756e-01 -1.58104181e-01 1.26072466e-01
7.36749232e-01 -8.49870265e-01 -5.32344878e-01 -7.56295383e-01
6.05001509e-01 -7.66676664e-01 -2.60255724e-01 6.28392458e-01
-1.40568972e-01 8.52185249e-01 -1.19324075e-02 8.17709020e-04
2.65397757e-01 2.23497063e-01 9.46291506e-01 -3.32022429e-01
8.81450921e-02 -3.38054031e-01 -1.34750223e+00 8.02007496e-01
7.83279479e-01 -1.93704277e-01 -2.64477849e-01 -2.92335361e-01
3.61369476e-02 -3.72930974e-01 3.21519315e-01 -9.09857571e-01
-1.24681056e-01 -6.18526816e-01 1.34728754e+00 -2.52934307e-01
3.09321582e-01 -6.71813965e-01 2.42164776e-01 2.90755689e-01
-4.25569296e-01 5.45005687e-02 6.44749776e-02 1.50768355e-01
8.19332004e-02 2.79606998e-01 1.20424056e+00 3.78004387e-02
-6.52940154e-01 7.31095731e-01 9.01492909e-02 4.69204485e-01
9.85056520e-01 -3.58602166e-01 -8.00641000e-01 -4.53945249e-01
-3.60078633e-01 1.86084405e-01 9.55695689e-01 7.29469478e-01
2.95379758e-01 -1.63503027e+00 -1.05137002e+00 6.36989117e-01
6.17880225e-01 -8.38353157e-01 4.50838119e-01 6.40873671e-01
-2.46527597e-01 -2.76287168e-01 -7.06090569e-01 -2.39286751e-01
-1.53269410e+00 4.95670110e-01 3.81083131e-01 8.59816015e-01
-6.79322854e-02 6.43614531e-01 2.36947551e-01 -6.24199808e-01
3.37400734e-02 3.62075061e-01 -1.40709788e-01 3.96931499e-01
7.14685619e-01 7.78452516e-01 -4.26192313e-01 -1.26837432e+00
-6.73695326e-01 6.25439584e-01 1.04685007e-02 1.58501878e-01
7.98150599e-01 -5.31900302e-02 -3.18738461e-01 1.05963714e-01
1.16983891e+00 2.80056775e-01 -7.15849698e-01 1.16140872e-01
-2.54381776e-01 -9.55731511e-01 -2.53741652e-01 -7.03631699e-01
-1.40799832e+00 4.88734663e-01 8.06948960e-01 -2.12802470e-01
1.05316532e+00 -1.06838442e-01 2.65382051e-01 -2.10350394e-01
4.31341767e-01 -1.35406721e+00 -2.87727594e-01 2.83631563e-01
5.94413698e-01 -1.46387923e+00 2.43319780e-01 -4.40946668e-01
-5.17169774e-01 7.73815811e-01 1.18610680e+00 3.29632759e-01
1.05888247e+00 3.78436856e-02 5.86766601e-01 5.69549613e-02
-5.17428517e-01 2.80438531e-02 2.28590742e-02 9.62223113e-01
8.00928354e-01 2.74409026e-01 -2.03661904e-01 5.09478688e-01
-4.64600414e-01 2.04662859e-01 5.13725936e-01 6.85249269e-01
1.54361799e-01 -7.10757375e-01 -7.03978777e-01 1.15535557e+00
-6.99922025e-01 2.22272709e-01 -8.58621418e-01 7.95873523e-01
2.18369126e-01 1.12857890e+00 5.75851262e-01 -3.06840211e-01
1.02224417e-01 3.73794347e-01 6.25552416e-01 -4.28043872e-01
-4.47030574e-01 -5.92685759e-01 1.32879943e-01 -4.16011900e-01
-3.60402644e-01 -7.81288266e-01 -9.11591351e-01 -1.02068770e+00
1.65688932e-01 -2.09312931e-01 3.46652955e-01 3.95648688e-01
5.57998538e-01 -1.73994198e-01 7.49271989e-01 -3.84909570e-01
-4.20906872e-01 -1.00301480e+00 -8.66155505e-01 7.46663570e-01
1.30054697e-01 -7.42232740e-01 -3.91887933e-01 5.60167357e-02] | [13.0468168258667, 1.2537099123001099] |
aee0e32b-7f52-45da-a6d7-df53641dec56 | towards-better-understanding-with-uniformity | 2206.07960 | null | https://arxiv.org/abs/2206.07960v1 | https://arxiv.org/pdf/2206.07960v1.pdf | Towards Better Understanding with Uniformity and Explicit Regularization of Embeddings in Embedding-based Neural Topic Models | Embedding-based neural topic models could explicitly represent words and topics by embedding them to a homogeneous feature space, which shows higher interpretability. However, there are no explicit constraints for the training of embeddings, leading to a larger optimization space. Also, a clear description of the changes in embeddings and the impact on model performance is still lacking. In this paper, we propose an embedding regularized neural topic model, which applies the specially designed training constraints on word embedding and topic embedding to reduce the optimization space of parameters. To reveal the changes and roles of embeddings, we introduce \textbf{uniformity} into the embedding-based neural topic model as the evaluation metric of embedding space. On this basis, we describe how embeddings tend to change during training via the changes in the uniformity of embeddings. Furthermore, we demonstrate the impact of changes in embeddings in embedding-based neural topic models through ablation studies. The results of experiments on two mainstream datasets indicate that our model significantly outperforms baseline models in terms of the harmony between topic quality and document modeling. This work is the first attempt to exploit uniformity to explore changes in embeddings of embedding-based neural topic models and their impact on model performance to the best of our knowledge. | ['Linqi Song', 'Shihua Ma', 'Shuqi Liu', 'Lei Huang', 'Wei Shao'] | 2022-06-16 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [-1.62456751e-01 3.31630737e-01 -3.25579673e-01 -4.15029734e-01
-5.09686351e-01 -4.15732741e-01 8.01984966e-01 8.05417150e-02
-3.63997996e-01 1.59287348e-01 5.11487067e-01 -1.40846357e-01
-8.94742906e-02 -8.66779327e-01 -6.01387203e-01 -5.99887788e-01
-9.14550871e-02 7.92917758e-02 -1.28087765e-02 -9.79450494e-02
2.91068673e-01 -9.98480842e-02 -1.40289915e+00 1.43048093e-02
9.80190992e-01 7.91483045e-01 1.51735157e-01 1.30975768e-01
-3.42061371e-01 -2.53176428e-02 -8.00983012e-01 -3.49786609e-01
6.87136129e-02 -9.75924507e-02 -3.00074011e-01 -1.68393478e-02
5.09469092e-01 -4.31505501e-01 -5.57528496e-01 8.08148146e-01
4.81431812e-01 3.32520545e-01 8.69237542e-01 -1.35454488e+00
-1.45574093e+00 8.04505467e-01 -2.26440966e-01 1.05026983e-01
-2.27005586e-01 -7.66883343e-02 1.55483520e+00 -1.02557981e+00
5.37877381e-01 1.28141284e+00 7.17965722e-01 4.94199783e-01
-1.33868921e+00 -6.17167294e-01 5.97446144e-01 1.49829656e-01
-1.43627381e+00 -3.55885252e-02 1.04005504e+00 -4.91742104e-01
1.04569256e+00 1.98413152e-02 7.79701114e-01 1.33825684e+00
2.21018255e-01 6.19252324e-01 5.81239820e-01 -5.24996936e-01
4.76885915e-01 6.27954423e-01 6.60413265e-01 2.05594927e-01
6.01469278e-01 -5.29226474e-02 -6.12308085e-01 -3.35036993e-01
5.11410356e-01 2.22642049e-01 -3.15880686e-01 -6.42738223e-01
-9.61649716e-01 1.37325490e+00 4.88404930e-01 2.99395740e-01
-1.94831818e-01 2.56809920e-01 4.20181990e-01 2.19430998e-02
9.73124623e-01 1.03922737e+00 -4.23266619e-01 -1.34701401e-01
-9.09545779e-01 3.68012547e-01 4.57808316e-01 8.43263090e-01
6.71938658e-01 2.02305257e-01 -4.70817417e-01 1.12598741e+00
4.87403810e-01 2.69997567e-01 8.79813254e-01 -7.17013001e-01
3.05868328e-01 6.08959794e-01 -7.78162852e-02 -1.29217064e+00
-2.06360623e-01 -5.11384726e-01 -3.29765528e-01 -1.61790192e-01
3.61053157e-03 -1.66401684e-01 -9.94664431e-01 1.89709198e+00
-1.88095588e-02 1.04725607e-01 1.00865113e-02 7.91463494e-01
7.77145922e-01 9.35910225e-01 2.30052978e-01 2.94133008e-01
1.57628906e+00 -1.04565978e+00 -1.11604917e+00 -3.35798651e-01
8.25245023e-01 -4.58526492e-01 1.59548342e+00 1.29904095e-02
-7.81441271e-01 -3.61140758e-01 -1.23827851e+00 -1.84196457e-01
-8.37696552e-01 3.79516363e-01 7.88740396e-01 8.81210446e-01
-1.06984699e+00 5.14498651e-01 -8.77142966e-01 -4.71684247e-01
3.57231647e-01 2.23794028e-01 7.86934607e-03 4.43616435e-02
-1.58275235e+00 8.42516065e-01 5.27169287e-01 -1.56786069e-01
-5.60386002e-01 -1.03851151e+00 -1.07040417e+00 4.09087688e-01
9.17342678e-02 -5.06775737e-01 9.79241192e-01 -1.92367077e-01
-1.41279805e+00 3.19371015e-01 -1.03067085e-01 -4.11742061e-01
-1.39292374e-01 -3.61708164e-01 -1.22079127e-01 -1.97621793e-01
-9.33713093e-02 1.06632495e+00 8.32713425e-01 -1.22641957e+00
-2.16900766e-01 -1.38093814e-01 1.59783028e-02 1.27545327e-01
-1.32683110e+00 -2.40898415e-01 -4.19589132e-01 -9.78179038e-01
8.04369748e-02 -9.83190954e-01 2.30653174e-02 1.85946245e-02
-2.76919842e-01 -6.29824877e-01 1.05119395e+00 -4.29097116e-01
1.54495573e+00 -2.38298178e+00 1.05903987e-02 -4.39185463e-02
4.03462768e-01 -2.19179429e-02 -3.16825837e-01 4.48643744e-01
4.68772911e-02 6.85523868e-01 -8.65680203e-02 -7.69152582e-01
3.70083719e-01 3.02773744e-01 -8.70903969e-01 2.33334109e-01
3.64025593e-01 9.00122702e-01 -5.40705860e-01 -2.92266190e-01
-4.18911837e-02 8.52978468e-01 -9.86087024e-01 1.97385535e-01
-1.50128528e-01 -3.57058853e-01 -4.08359349e-01 3.31395775e-01
4.99768645e-01 -1.83384240e-01 -8.16582739e-02 -1.39368087e-01
4.86189164e-02 6.55948818e-01 -7.03145087e-01 1.77569199e+00
-5.31828821e-01 1.09703732e+00 -5.48217118e-01 -7.24419355e-01
1.09303176e+00 3.89676005e-01 2.31089294e-01 -3.69514138e-01
-4.78699915e-02 -1.01598971e-01 1.63668737e-01 -8.53614956e-02
1.08079803e+00 8.80696066e-03 8.17419142e-02 7.98095942e-01
1.39318034e-01 -3.09578270e-01 4.12828662e-02 2.96106786e-01
6.87822878e-01 -2.59046406e-01 -2.96704769e-01 -3.59339446e-01
-3.86302531e-01 -1.81621313e-01 3.66507262e-01 6.29972875e-01
-2.85428204e-02 7.61306643e-01 6.40022039e-01 -2.77006060e-01
-1.16544521e+00 -9.57672477e-01 -5.47445178e-01 1.21132743e+00
-9.70198736e-02 -9.22681749e-01 -6.69758856e-01 -4.50538278e-01
1.17969744e-01 1.12560332e+00 -1.21654272e+00 -6.16773546e-01
-3.06262165e-01 -9.81391072e-01 4.50254381e-01 6.11851513e-01
3.19647938e-02 -8.38825822e-01 -5.75989127e-01 -8.56944248e-02
5.37844263e-02 -8.16427290e-01 -6.23516083e-01 1.93770364e-01
-1.24836791e+00 -3.89486998e-01 -8.88760448e-01 -5.03665745e-01
7.07626879e-01 3.37742776e-01 8.84741008e-01 -1.30388558e-01
6.58319192e-03 1.79196298e-01 -4.34695512e-01 -6.75733805e-01
8.85281488e-02 5.77974021e-01 -6.54325821e-03 -5.05031109e-01
5.06043196e-01 -4.66560096e-01 -5.27941167e-01 2.53453821e-01
-1.13414085e+00 -2.13376492e-01 4.47550446e-01 9.05620694e-01
1.05689526e-01 9.04494226e-02 4.92851973e-01 -8.13944697e-01
1.36569715e+00 -5.47076106e-01 -1.58772647e-01 1.39435068e-01
-1.18780482e+00 2.88722098e-01 2.43474483e-01 -8.16981673e-01
-7.76279867e-01 -7.62062371e-01 3.03710878e-01 -5.61766684e-01
1.67711675e-01 6.60288990e-01 3.10345720e-02 5.35032451e-01
5.82368553e-01 1.12437479e-01 -4.28006351e-02 -5.67060173e-01
7.27915347e-01 5.40479839e-01 -1.11264050e-01 -5.68904400e-01
7.14566827e-01 2.91112691e-01 -6.91356957e-01 -7.11363316e-01
-6.68171346e-01 -2.88532346e-01 -3.52185667e-01 3.12929571e-01
6.14190996e-01 -8.96997929e-01 2.01889828e-01 2.20377952e-01
-1.41708279e+00 -1.47459567e-01 -4.91026819e-01 6.72119319e-01
-2.91260272e-01 1.12664349e-01 -6.13538086e-01 -4.88757908e-01
-3.81632656e-01 -1.35799968e+00 1.12993956e+00 3.75451483e-02
-5.38224578e-01 -1.43396831e+00 3.07743579e-01 -1.50385976e-01
9.33868647e-01 -1.41829729e-01 1.27424252e+00 -8.24002802e-01
-1.18534625e-01 -1.82275698e-01 -2.65876740e-01 2.98241526e-01
2.93004930e-01 1.12172112e-01 -1.13408470e+00 -2.42362127e-01
1.08562388e-01 3.75386617e-06 1.08379495e+00 6.80213034e-01
1.17227340e+00 -3.79292101e-01 -3.92524093e-01 6.33200705e-01
1.10331118e+00 1.04727611e-01 6.59111321e-01 6.46125317e-01
6.27994239e-01 7.29335725e-01 3.25834185e-01 2.93857962e-01
2.29733557e-01 8.79239082e-01 1.79344460e-01 -1.95445195e-01
-1.90680046e-02 -3.94723982e-01 4.57169056e-01 9.86589372e-01
4.29957837e-01 -2.67847836e-01 -8.86260331e-01 6.89932168e-01
-1.55317557e+00 -5.35082281e-01 3.40910494e-01 1.91008794e+00
9.68232930e-01 1.54767960e-01 -1.13833562e-01 -1.91943794e-01
6.36939287e-01 4.80929285e-01 -5.10220587e-01 -5.96759200e-01
1.84838623e-01 -8.94088969e-02 2.19657242e-01 3.12548608e-01
-9.41081464e-01 1.09142375e+00 6.93917131e+00 8.02572727e-01
-1.24282146e+00 2.67860770e-01 4.98880774e-01 -3.53114814e-01
-7.74567306e-01 -5.67400083e-02 -9.29889679e-01 5.35925686e-01
9.12379384e-01 -3.45885187e-01 2.03008335e-02 1.22184467e+00
1.95658654e-01 4.19057757e-01 -1.19539392e+00 6.48792505e-01
2.78423876e-01 -1.32726872e+00 4.30997610e-01 4.45219189e-01
7.64093697e-01 -1.44855142e-01 6.46373868e-01 6.30740821e-01
1.06957316e-01 -1.07638025e+00 6.04658067e-01 1.25497997e-01
5.04480183e-01 -6.37936413e-01 7.97204733e-01 -2.53096014e-01
-7.93603599e-01 3.77595462e-02 -7.20710695e-01 2.04520285e-01
6.01055995e-02 6.25767648e-01 -9.82471526e-01 -1.89025123e-02
6.44528389e-01 7.21043468e-01 -6.80221617e-01 7.61803210e-01
-1.41748637e-01 9.89556611e-01 -2.11486220e-01 -1.47941902e-01
3.47223818e-01 -1.09302804e-01 5.16699195e-01 1.18952823e+00
4.58375961e-01 -4.01507407e-01 -2.33662888e-01 1.50339258e+00
-1.58212245e-01 1.69313788e-01 -7.35074043e-01 -3.26790899e-01
6.97151601e-01 9.58091855e-01 -4.86799300e-01 -2.40069598e-01
-3.40164959e-01 5.74491501e-01 3.47548753e-01 6.23289883e-01
-1.00174057e+00 -5.39063752e-01 1.11566663e+00 2.23826110e-01
4.30896312e-01 -4.08293724e-01 -5.95302880e-01 -1.07871974e+00
1.04058543e-02 -5.19614041e-01 -2.26909731e-04 -5.35564661e-01
-1.26409566e+00 6.85717940e-01 4.16361243e-01 -7.88689673e-01
-1.35444045e-01 -4.71415043e-01 -1.05284369e+00 9.21808958e-01
-1.54312670e+00 -9.13507104e-01 -8.12408105e-02 -7.54154995e-02
6.86345339e-01 -2.98736066e-01 7.13285327e-01 -5.96652888e-02
-7.55106509e-01 8.93771529e-01 3.65923941e-01 -1.93723291e-02
9.12057042e-01 -1.22474301e+00 7.08382547e-01 5.45816183e-01
1.67383164e-01 1.32264721e+00 8.14364314e-01 -4.83483583e-01
-1.06283808e+00 -1.11413264e+00 6.87837183e-01 -5.73987305e-01
7.16990292e-01 -7.89424479e-01 -1.15641356e+00 6.50414109e-01
3.83641571e-01 -4.25357193e-01 1.14428270e+00 6.91025198e-01
-5.77075720e-01 1.59590289e-01 -6.58135772e-01 7.36716092e-01
5.48380613e-01 -5.79864144e-01 -7.45590866e-01 1.84672162e-01
1.63289738e+00 1.76566005e-01 -1.02398753e+00 1.89914688e-01
5.57829916e-01 -4.67925966e-01 1.04948223e+00 -6.12832129e-01
4.80097264e-01 5.57982065e-02 -2.88949251e-01 -1.63930821e+00
-3.83955210e-01 -6.51314929e-02 -3.27481747e-01 1.43091798e+00
6.66618943e-01 -8.27741027e-01 8.37236941e-01 8.54958713e-01
-7.04711601e-02 -1.17318082e+00 -7.41891921e-01 -7.68442810e-01
6.66594684e-01 -3.91427130e-01 7.32428193e-01 1.04219806e+00
1.20097362e-01 1.59036323e-01 -4.56044711e-02 -4.77389991e-02
2.49304131e-01 -1.39475256e-01 5.70717990e-01 -1.12517190e+00
-1.08568273e-01 -6.21915042e-01 -2.68769920e-01 -9.54086244e-01
2.37654120e-01 -7.23288655e-01 -1.34340554e-01 -1.52485693e+00
2.08761498e-01 -6.30184412e-01 -4.08284813e-01 5.04857063e-01
-4.13634181e-01 -6.66869506e-02 2.24906161e-01 5.54378688e-01
-3.79766285e-01 1.25210118e+00 7.40563154e-01 -2.44267344e-01
-3.77830595e-01 -3.14256340e-01 -1.06107986e+00 4.39834088e-01
1.11206901e+00 -6.55423880e-01 -7.35975325e-01 -8.51460040e-01
9.51083899e-02 -8.85790646e-01 -4.12629358e-02 -6.62596583e-01
6.94125891e-02 5.46859466e-02 7.88086951e-02 -4.16031331e-01
5.75892150e-01 -5.74184358e-01 -4.34429675e-01 1.98123604e-01
-8.46415937e-01 1.64538354e-01 4.72372800e-01 6.39804244e-01
-2.25966051e-01 -3.64886850e-01 3.16944271e-01 2.80997992e-01
-4.53101844e-01 1.82394490e-01 -3.63242775e-01 -3.26420255e-02
8.31263483e-01 -2.52016246e-01 -4.15351123e-01 -3.58308405e-01
-3.68597955e-01 7.86851346e-02 5.76238334e-01 8.51083577e-01
5.98688722e-01 -1.49791098e+00 -5.68298995e-01 1.21962331e-01
3.16539198e-01 -4.65173647e-02 1.45958021e-01 4.10890877e-01
8.06235000e-02 8.58977020e-01 3.45452093e-02 -5.67586720e-01
-8.61873150e-01 4.07770038e-01 1.36834979e-01 -3.91340733e-01
-5.17122507e-01 6.33912265e-01 6.21055663e-01 -6.95521057e-01
5.07776976e-01 -7.65779495e-01 -2.87876636e-01 1.92204729e-01
3.46326649e-01 1.05362102e-01 -5.14799505e-02 -5.49998209e-02
-1.18174747e-01 4.61345166e-01 -5.61339796e-01 -2.51147121e-01
1.34406888e+00 -1.27023265e-01 1.21821031e-01 6.25376821e-01
1.49883151e+00 -3.16011906e-01 -1.30771780e+00 -2.09687874e-01
-6.39156550e-02 -4.45707202e-01 3.55245918e-01 -4.74746525e-01
-9.43348169e-01 1.19243610e+00 6.35707915e-01 2.13477552e-01
6.42008722e-01 9.34562385e-02 7.67822266e-01 3.41303498e-01
-1.81077763e-01 -1.10682976e+00 3.09062451e-01 4.77671444e-01
1.00054383e+00 -1.10478365e+00 4.67697904e-02 -2.40946010e-01
-6.67213500e-01 9.26926076e-01 7.12374866e-01 -9.65259522e-02
8.18989873e-01 -1.62669905e-02 2.33812872e-02 -4.10669863e-01
-1.02888083e+00 3.25883836e-01 4.96762246e-01 6.01516664e-01
6.47782922e-01 -7.78037682e-02 -2.91141748e-01 7.33657598e-01
-5.99835575e-01 -5.33255577e-01 3.28160673e-01 6.74150825e-01
-4.30618972e-01 -1.12798464e+00 -2.48066649e-01 4.32049751e-01
-2.13576421e-01 -4.10646856e-01 -2.94935346e-01 8.42164457e-01
-8.89538378e-02 7.55341947e-01 4.10101026e-01 -4.23146933e-01
1.50532871e-01 3.05112451e-01 -1.81227893e-01 -9.31098104e-01
-2.79739529e-01 -2.17130944e-01 -2.23166123e-01 -1.95216343e-01
8.89682546e-02 -5.79758167e-01 -9.28664088e-01 -2.14844737e-02
-8.88373256e-01 3.94269973e-01 1.14231384e+00 5.96613765e-01
7.93454230e-01 7.83058465e-01 4.35313851e-01 -7.34555602e-01
-7.41144896e-01 -1.29246259e+00 -5.22881150e-01 3.26984793e-01
1.48105866e-03 -9.72134829e-01 -6.59876823e-01 -2.21490964e-01] | [10.502030372619629, 7.145901203155518] |
9cba1bed-2a10-4ec4-bd9c-7889e91f9feb | timereplayer-unlocking-the-potential-of-event | 2203.13859 | null | https://arxiv.org/abs/2203.13859v1 | https://arxiv.org/pdf/2203.13859v1.pdf | TimeReplayer: Unlocking the Potential of Event Cameras for Video Interpolation | Recording fast motion in a high FPS (frame-per-second) requires expensive high-speed cameras. As an alternative, interpolating low-FPS videos from commodity cameras has attracted significant attention. If only low-FPS videos are available, motion assumptions (linear or quadratic) are necessary to infer intermediate frames, which fail to model complex motions. Event camera, a new camera with pixels producing events of brightness change at the temporal resolution of $\mu s$ $(10^{-6}$ second $)$, is a game-changing device to enable video interpolation at the presence of arbitrarily complex motion. Since event camera is a novel sensor, its potential has not been fulfilled due to the lack of processing algorithms. The pioneering work Time Lens introduced event cameras to video interpolation by designing optical devices to collect a large amount of paired training data of high-speed frames and events, which is too costly to scale. To fully unlock the potential of event cameras, this paper proposes a novel TimeReplayer algorithm to interpolate videos captured by commodity cameras with events. It is trained in an unsupervised cycle-consistent style, canceling the necessity of high-speed training data and bringing the additional ability of video extrapolation. Its state-of-the-art results and demo videos in supplementary reveal the promising future of event-based vision. | ['Jianxing Liao', 'Yaoyuan Wang', 'Huchuan Lu', 'Wenhui Wang', 'Ziyang Zhang', 'Xu Jia', 'Zhendong Qiao', 'Kaichao You', 'Weihua He'] | 2022-03-25 | null | http://openaccess.thecvf.com//content/CVPR2022/html/He_TimeReplayer_Unlocking_the_Potential_of_Event_Cameras_for_Video_Interpolation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/He_TimeReplayer_Unlocking_the_Potential_of_Event_Cameras_for_Video_Interpolation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['event-based-vision'] | ['computer-vision'] | [ 2.37084419e-01 -1.89841688e-01 -1.59073733e-02 -9.37195495e-02
-3.25456679e-01 -2.44837940e-01 2.67621040e-01 -4.21611905e-01
-5.70318997e-01 6.40419781e-01 -3.74467403e-01 -2.91395485e-01
2.29235202e-01 -7.07782507e-01 -1.02875721e+00 -4.11060482e-01
-1.48271248e-01 -2.06630364e-01 7.84222424e-01 2.93706600e-02
-3.36037911e-02 4.01786357e-01 -1.88834631e+00 3.69516909e-01
6.85000241e-01 1.16104162e+00 4.80014831e-01 9.61703658e-01
-5.72365969e-02 1.13658786e+00 -3.12104434e-01 -5.03934801e-01
5.49274147e-01 -5.45372725e-01 -1.92519411e-01 2.68831849e-01
4.47853208e-01 -8.80057573e-01 -6.50321901e-01 8.23879004e-01
1.26193091e-01 -7.69274682e-02 -1.09657444e-01 -1.22042787e+00
-1.36526018e-01 -7.96666294e-02 -3.94809753e-01 3.55685890e-01
5.96208513e-01 5.64949751e-01 3.63656193e-01 -8.44759107e-01
7.78464913e-01 7.83768892e-01 7.99440742e-01 5.01791179e-01
-9.92817998e-01 -4.99379039e-01 -1.21205382e-01 4.61346537e-01
-1.46355712e+00 -4.27171350e-01 8.22095573e-01 -4.27770078e-01
9.43358839e-01 2.78134346e-01 1.17899013e+00 1.12058473e+00
2.63442606e-01 5.41305959e-01 7.59191096e-01 -3.54080170e-01
3.51642191e-01 -5.53169884e-02 -6.00284457e-01 5.54387748e-01
-5.36759980e-02 3.08371902e-01 -7.91055977e-01 3.02663505e-01
1.34891963e+00 2.71907747e-01 -5.33767343e-01 9.92570817e-02
-1.42172766e+00 4.71415877e-01 1.30665272e-01 8.30556601e-02
-4.40772951e-01 3.68675023e-01 2.19455108e-01 3.47446978e-01
3.38904202e-01 -5.42239733e-02 -3.14394385e-01 -5.50946534e-01
-1.25546277e+00 2.40137875e-02 4.11403954e-01 1.18444467e+00
7.68346727e-01 4.65841442e-01 1.95036620e-01 1.38106104e-02
1.01613335e-01 4.29834336e-01 2.90794134e-01 -1.48964477e+00
3.02444845e-01 5.14593422e-01 4.28141624e-01 -9.44082379e-01
-1.17598563e-01 1.52106047e-01 -8.85730505e-01 4.48960006e-01
7.43217885e-01 -1.65335700e-01 -4.46004331e-01 1.07944381e+00
3.07712883e-01 6.78269088e-01 -2.76893765e-01 1.06816995e+00
3.60723585e-01 9.33783054e-01 -2.22725958e-01 -8.99263918e-01
1.07408047e+00 -5.59947491e-01 -6.59990013e-01 3.11948340e-02
1.61575571e-01 -6.73131168e-01 1.30322731e+00 6.96436346e-01
-1.33658648e+00 -9.70911741e-01 -8.79427314e-01 -5.01837768e-03
-3.20099927e-02 -4.30808775e-02 7.94837117e-01 5.29900551e-01
-1.04674399e+00 6.32255137e-01 -1.04698634e+00 -1.40351638e-01
2.73932993e-01 3.62547696e-01 -2.58883357e-01 -5.57380542e-02
-1.02296305e+00 4.46571559e-01 3.48980308e-01 1.91523850e-01
-7.65508413e-01 -7.46128976e-01 -6.72512650e-01 -1.30396456e-01
3.84546936e-01 -5.42907774e-01 9.73472714e-01 -1.40777898e+00
-1.87481689e+00 5.82593977e-01 -2.01974705e-01 -5.72060585e-01
8.97115827e-01 -1.66680902e-01 -4.54363376e-01 5.84590197e-01
-2.30936721e-01 7.88383365e-01 1.16062760e+00 -6.60642982e-01
-8.19566131e-01 2.62764245e-02 1.90402806e-01 -1.24212973e-01
-3.37164730e-01 1.98112309e-01 -5.46193123e-01 -4.10141766e-01
-1.11636065e-01 -8.35680664e-01 -2.70440698e-01 4.06500936e-01
3.23659867e-01 4.65216711e-02 9.04492855e-01 -5.12664616e-01
1.18055964e+00 -2.17449141e+00 -1.87559247e-01 -4.30231780e-01
-9.11173448e-02 3.53134513e-01 3.28654647e-01 7.54347220e-02
1.80666223e-02 -1.99957564e-01 -4.51397859e-02 -2.05577731e-01
-5.10680318e-01 1.92635417e-01 -3.16499323e-01 3.97047937e-01
2.71633267e-01 5.79912543e-01 -1.03393042e+00 -7.18728065e-01
8.31502676e-01 6.47918582e-01 -6.38456047e-01 1.75344050e-01
-2.95923799e-01 6.73522353e-01 -8.08066651e-02 7.37753570e-01
6.49981201e-01 -2.62796134e-01 -8.27829242e-02 -3.38286668e-01
-5.65925539e-01 -2.52895564e-01 -1.46209049e+00 1.94994175e+00
-1.06289789e-01 9.28037524e-01 -8.17228556e-02 -6.57857597e-01
6.49230480e-01 4.02086377e-01 8.12738061e-01 -5.92518747e-01
4.66866605e-02 1.64699808e-01 -5.45525432e-01 -8.09684575e-01
6.02181077e-01 -1.23594344e-01 2.19691902e-01 -2.84241945e-01
-1.62128527e-02 -2.16600284e-01 3.32767606e-01 -5.15210629e-02
1.12194192e+00 7.31123269e-01 8.83066803e-02 9.76942256e-02
3.41376811e-01 1.20639563e-01 7.22764134e-01 4.31280524e-01
-1.86236992e-01 8.12672734e-01 1.20488651e-01 -7.35483170e-01
-1.22042835e+00 -1.14192772e+00 -1.33827627e-02 7.27215350e-01
2.03961298e-01 -4.16347235e-01 -6.82601750e-01 -1.14431232e-01
-4.86818314e-01 2.88352072e-01 -1.06331162e-01 2.35474080e-01
-7.99280941e-01 -4.94953543e-01 3.46987486e-01 5.24439871e-01
7.61565387e-01 -7.49872088e-01 -1.17369068e+00 3.82740766e-01
-2.09994063e-01 -1.43806231e+00 -5.08058369e-01 -2.04925448e-01
-9.93587732e-01 -1.01525390e+00 -6.45820856e-01 -6.08708918e-01
6.40229642e-01 3.59820962e-01 8.62680495e-01 -2.25701526e-01
-3.15337092e-01 3.44106764e-01 -3.75187576e-01 -2.46326461e-01
-1.14048272e-01 -6.31504476e-01 2.02132478e-01 3.09504300e-01
2.41739824e-01 -6.49599612e-01 -9.49769199e-01 3.15248519e-01
-1.03880835e+00 4.31647390e-01 2.67482787e-01 5.81185520e-01
6.63066149e-01 1.88282840e-02 7.88911134e-02 -1.85583293e-01
-2.18065113e-01 -1.94221839e-01 -9.98484612e-01 -1.28476873e-01
-2.16596112e-01 -4.45578843e-01 1.10770869e+00 -7.73892999e-01
-9.99890089e-01 5.67445219e-01 8.18075240e-02 -1.07165885e+00
-9.44786668e-02 8.48258063e-02 7.63335004e-02 -9.28726345e-02
6.25590801e-01 4.16650146e-01 -7.16316551e-02 -5.15339114e-02
2.46318802e-01 4.76505816e-01 9.51171875e-01 -2.01981559e-01
4.81766701e-01 9.50272679e-01 6.65197670e-02 -1.19294083e+00
-2.05693975e-01 -4.14199442e-01 -5.58403254e-01 -8.09927702e-01
1.01753402e+00 -1.23674166e+00 -1.01363838e+00 5.27648389e-01
-1.35948086e+00 -3.73099744e-01 -5.84817171e-01 9.34002101e-01
-6.60274148e-01 4.62774336e-01 -8.04252744e-01 -8.86681378e-01
6.55015931e-02 -1.01281691e+00 8.23658347e-01 5.16265213e-01
8.75559375e-02 -5.47943711e-01 -2.45315611e-01 3.03369820e-01
1.63922772e-01 3.46148551e-01 -1.24700300e-01 5.53234577e-01
-1.22934330e+00 -2.14772135e-01 -1.10082299e-01 3.30072969e-01
-9.78922993e-02 5.40080070e-01 -8.78703833e-01 -1.26910225e-01
3.05345356e-01 1.07859865e-01 4.53424990e-01 6.61636055e-01
1.25866973e+00 -2.10097209e-01 -5.15562147e-02 9.35167730e-01
1.53995645e+00 4.55083340e-01 1.09755456e+00 1.51952177e-01
6.81884289e-01 2.75099754e-01 6.27186716e-01 7.87573934e-01
3.55184227e-01 7.77426958e-01 4.49745864e-01 6.12656884e-02
-9.34307948e-02 -2.90476114e-01 7.39389956e-01 7.20024049e-01
-7.34167576e-01 -1.21701993e-01 -6.25000000e-01 5.60160875e-01
-1.72080994e+00 -1.37722695e+00 -5.59394360e-01 2.39097548e+00
7.54655421e-01 1.73466727e-01 2.56477863e-01 3.05705160e-01
7.74350464e-01 3.45010236e-02 -3.27985317e-01 -1.89420566e-01
-1.45164356e-01 1.93424702e-01 5.74392736e-01 3.33663672e-01
-9.15044904e-01 7.19958842e-01 5.76277208e+00 7.23639607e-01
-1.37098646e+00 -2.89020278e-02 5.64220548e-01 -3.97448748e-01
-2.63936408e-02 2.32371822e-01 -7.44081497e-01 1.05114496e+00
1.23334205e+00 1.41108334e-01 5.39067805e-01 8.03022206e-01
7.35818505e-01 -3.65096092e-01 -9.57316279e-01 1.41174102e+00
-1.01554506e-02 -1.50742447e+00 -1.32366896e-01 -1.54059976e-01
5.88295519e-01 -2.30609924e-01 -2.01020166e-01 -2.78509576e-02
-2.05076873e-01 -6.05232120e-01 8.85586917e-01 6.74819231e-01
1.14423966e+00 -5.32274663e-01 3.81502539e-01 4.50030506e-01
-1.38049817e+00 -1.61028821e-02 -4.17099237e-01 -6.15323603e-01
6.98414803e-01 6.49313033e-01 -5.75988412e-01 3.41817647e-01
9.51365292e-01 7.15733171e-01 -2.79665142e-01 8.16969097e-01
1.35620728e-01 4.24594969e-01 -5.99677861e-01 2.05218360e-01
-1.11691140e-01 -3.62394512e-01 3.52542967e-01 1.16288805e+00
7.44668841e-01 4.87702757e-01 4.74353805e-02 6.85466290e-01
1.96302578e-01 -4.02826548e-01 -6.27356350e-01 1.37343571e-01
2.32572109e-01 9.93860960e-01 -7.46964991e-01 -3.54089409e-01
-8.45985234e-01 1.41204953e+00 -2.15165421e-01 1.20128900e-01
-1.23109436e+00 -2.17456490e-01 3.56386751e-01 5.45661867e-01
3.41432005e-01 -7.38563061e-01 -8.92863125e-02 -1.41469181e+00
2.50245273e-01 -3.40849459e-01 9.16294679e-02 -1.08586955e+00
-7.48921514e-01 2.78653592e-01 -4.48958762e-02 -1.86117196e+00
-3.93898040e-01 -4.89068627e-01 -3.36764604e-01 3.68732393e-01
-1.33750749e+00 -9.94574130e-01 -6.87075436e-01 1.11678278e+00
8.78161848e-01 7.03461915e-02 4.73255634e-01 5.76824605e-01
-2.85269439e-01 2.25456208e-01 3.09835821e-02 5.07941796e-03
5.23772955e-01 -7.40374804e-01 9.33620334e-02 1.17049539e+00
6.57731807e-03 4.19455543e-02 7.44041145e-01 -5.03942132e-01
-1.86815202e+00 -9.40410674e-01 8.03090096e-01 -3.77795130e-01
5.04462063e-01 -1.67223930e-01 -6.99528039e-01 5.04210532e-01
4.30652779e-03 5.49652457e-01 2.43711874e-01 -9.07824278e-01
2.23956659e-01 -5.89491606e-01 -1.01525354e+00 6.52262688e-01
1.06814134e+00 -5.81448197e-01 -2.17883930e-01 -1.14039099e-03
7.64220417e-01 -5.57227194e-01 -8.16661894e-01 2.19069883e-01
5.73353946e-01 -1.52510643e+00 1.07652354e+00 1.02166533e-01
5.65060258e-01 -6.30202353e-01 -9.79144871e-02 -5.01538277e-01
4.06715795e-02 -1.03313386e+00 -3.23949188e-01 9.98318374e-01
-2.19897255e-01 -2.62972444e-01 1.01027226e+00 7.71972597e-01
-1.61243692e-01 -2.96805799e-01 -1.18639016e+00 -8.78337502e-01
-6.45793617e-01 -7.80522406e-01 1.25225693e-01 8.90195191e-01
5.92620224e-02 -3.58076729e-02 -6.75012290e-01 8.61485228e-02
6.47188187e-01 -2.04137564e-01 7.87668526e-01 -9.14691091e-01
-4.04324770e-01 -7.28599494e-03 -6.30708456e-01 -1.32703805e+00
-2.60443240e-01 -1.55582085e-01 -1.16420701e-01 -9.32186604e-01
-3.26154947e-01 -4.57619652e-02 1.34691790e-01 -2.73881257e-02
-2.65712906e-02 3.55439991e-01 2.84370005e-01 2.89186418e-01
-6.04311407e-01 3.08079809e-01 1.05093837e+00 1.67804122e-01
-3.26939791e-01 -2.15484589e-01 2.20729306e-01 9.74837184e-01
4.25958097e-01 -1.65172577e-01 -5.01606166e-01 -4.97297406e-01
3.29775065e-01 7.31667101e-01 6.69033706e-01 -1.46007681e+00
5.18269300e-01 -3.98757219e-01 4.79002178e-01 -4.45345968e-01
6.34862721e-01 -1.07721841e+00 8.87859404e-01 4.51131880e-01
1.64340049e-01 2.71528423e-01 6.97917938e-02 6.75692499e-01
-2.93264270e-01 -1.88107993e-02 8.65923226e-01 -4.34118807e-01
-1.11715651e+00 3.26253593e-01 -4.54584420e-01 -2.11692750e-01
1.33359456e+00 -8.76661479e-01 4.38739248e-02 -2.86430448e-01
-3.79669070e-01 -2.91406810e-01 7.94802368e-01 9.32035223e-02
8.46376657e-01 -1.18761110e+00 -4.34502900e-01 3.83672774e-01
-3.03363442e-01 2.62885362e-01 4.97704357e-01 7.35380650e-01
-1.15173471e+00 1.10365555e-01 -3.74079138e-01 -9.47667360e-01
-9.26408231e-01 7.72516489e-01 -7.66466651e-03 2.06433415e-01
-7.57695436e-01 7.01215267e-01 -1.71671346e-01 4.98962522e-01
-3.35780047e-02 -5.93425274e-01 3.16128522e-01 -9.32825282e-02
7.51026213e-01 6.06072605e-01 -1.19209133e-01 -3.03207487e-01
-1.68644503e-01 5.61886489e-01 4.59524691e-01 -1.21702120e-01
1.11717343e+00 -3.07391763e-01 2.31333077e-01 4.84713107e-01
8.56364012e-01 -1.34047791e-01 -2.07522941e+00 2.96575010e-01
-4.17500764e-01 -8.46461952e-01 -1.50453756e-02 -3.90608460e-02
-8.90105546e-01 8.51086497e-01 6.45809948e-01 2.20204815e-01
1.44456065e+00 -4.41322505e-01 9.92117941e-01 -1.33033335e-01
8.32868099e-01 -1.16553152e+00 9.65166688e-02 3.04551005e-01
3.89893413e-01 -1.41467035e+00 -1.33484021e-01 -3.86425167e-01
-3.64044100e-01 1.40965903e+00 5.56135595e-01 -1.11909591e-01
3.62160832e-01 4.82869357e-01 -3.74750942e-01 3.36822331e-01
-5.34899414e-01 -7.85341933e-02 -2.76816666e-01 7.06563413e-01
2.22093225e-01 -2.05355868e-01 -3.34664822e-01 2.71094471e-01
1.11708865e-01 8.25174332e-01 7.90636241e-01 8.20103586e-01
-2.23846287e-01 -6.62115514e-01 -4.35605615e-01 1.03807844e-01
-3.70805293e-01 -7.87400305e-02 4.63594675e-01 6.39082193e-01
4.93332028e-01 9.48939979e-01 2.94271380e-01 -4.36053723e-01
2.48058021e-01 -2.22064540e-01 6.03251755e-01 6.92598969e-02
-3.27824175e-01 1.04708448e-01 -2.53395557e-01 -8.84932101e-01
-7.91894913e-01 -6.50987804e-01 -1.12759697e+00 -6.73561156e-01
-9.44466796e-03 -2.46957019e-01 4.90033090e-01 5.83324850e-01
4.20547366e-01 2.21290439e-01 6.15987241e-01 -1.33242965e+00
1.61291301e-01 -4.64463621e-01 -4.04340595e-01 3.96945238e-01
3.10569435e-01 -2.23467574e-01 -4.29018587e-01 1.00509357e+00] | [8.681621551513672, -1.315900444984436] |
08ef8436-aa3a-482f-b7f3-a2fe9056131c | disentangling-monocular-3d-object-detection | 1905.12365 | null | https://arxiv.org/abs/1905.12365v1 | https://arxiv.org/pdf/1905.12365v1.pdf | Disentangling Monocular 3D Object Detection | In this paper we propose an approach for monocular 3D object detection from a single RGB image, which leverages a novel disentangling transformation for 2D and 3D detection losses and a novel, self-supervised confidence score for 3D bounding boxes. Our proposed loss disentanglement has the twofold advantage of simplifying the training dynamics in the presence of losses with complex interactions of parameters, and sidestepping the issue of balancing independent regression terms. Our solution overcomes these issues by isolating the contribution made by groups of parameters to a given loss, without changing its nature. We further apply loss disentanglement to another novel, signed Intersection-over-Union criterion-driven loss for improving 2D detection results. Besides our methodological innovations, we critically review the AP metric used in KITTI3D, which emerged as the most important dataset for comparing 3D detection results. We identify and resolve a flaw in the 11-point interpolated AP metric, affecting all previously published detection results and particularly biases the results of monocular 3D detection. We provide extensive experimental evaluations and ablation studies on the KITTI3D and nuScenes datasets, setting new state-of-the-art results on object category car by large margins. | ['Manuel López-Antequera', 'Samuel Rota Rota Bulò', 'Peter Kontschieder', 'Lorenzo Porzi', 'Andrea Simonelli'] | 2019-05-29 | disentangling-monocular-3d-object-detection-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Simonelli_Disentangling_Monocular_3D_Object_Detection_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Simonelli_Disentangling_Monocular_3D_Object_Detection_ICCV_2019_paper.pdf | iccv-2019-10 | ['3d-object-detection-from-monocular-images'] | ['computer-vision'] | [ 4.77413312e-02 -1.85309827e-01 -2.15348497e-01 -2.39437416e-01
-8.95065427e-01 -8.63384724e-01 6.37804449e-01 -2.49380507e-02
-7.47618794e-01 4.61558282e-01 -2.70160347e-01 -4.56383616e-01
-8.35859755e-05 -4.86719340e-01 -8.94163847e-01 -8.28113079e-01
-1.94793586e-02 3.24389279e-01 6.18703604e-01 2.79308632e-02
1.61377519e-01 9.51911092e-01 -1.82504749e+00 -6.07511811e-02
6.59777880e-01 1.36663151e+00 -1.19860750e-03 5.78046083e-01
2.43996292e-01 5.20014465e-01 -4.37101096e-01 -5.62569261e-01
7.94957876e-01 -5.92136197e-03 -1.99782237e-01 1.93944294e-02
9.97557819e-01 -4.80865538e-01 -2.48016164e-01 8.66644681e-01
6.22594416e-01 -2.77620047e-01 1.04400206e+00 -1.49093521e+00
-4.43903118e-01 -3.46875936e-02 -9.47542012e-01 1.66899428e-01
2.50088096e-01 3.38745862e-01 1.07881057e+00 -1.06539881e+00
4.89090919e-01 1.34007168e+00 9.50279415e-01 3.52929533e-01
-1.31250584e+00 -7.95577347e-01 1.60168350e-01 9.18431431e-02
-1.49949813e+00 -1.59103632e-01 6.91342175e-01 -6.29584253e-01
1.09585702e+00 1.91609547e-01 5.20550191e-01 1.02753627e+00
-1.70946717e-01 9.64957774e-01 1.23720658e+00 -5.72469175e-01
-1.70589909e-01 3.70176852e-01 1.13221787e-01 6.50692761e-01
6.56197906e-01 6.46160722e-01 -2.91888535e-01 -2.68648602e-02
6.47091687e-01 -2.92186081e-01 4.10109423e-02 -1.17640531e+00
-8.92498672e-01 9.15390968e-01 6.94467664e-01 -2.84114540e-01
1.74833521e-01 4.13161814e-02 2.81761885e-01 5.22960983e-02
4.82685179e-01 2.39737868e-01 -3.52833450e-01 1.39720023e-01
-6.29789948e-01 4.07468110e-01 4.40744549e-01 1.14504337e+00
7.86501825e-01 -2.80530959e-01 -1.34864017e-01 5.87523103e-01
2.93079883e-01 7.70109653e-01 -1.00684769e-01 -9.17300165e-01
5.57023883e-01 6.22040987e-01 3.02202046e-01 -8.04356992e-01
-4.59224492e-01 -4.91681725e-01 -3.15203577e-01 9.60235834e-01
6.79200649e-01 9.10970867e-02 -8.91202450e-01 1.53032410e+00
5.09012938e-01 -4.02909577e-01 -3.99606228e-01 1.11842585e+00
8.61014247e-01 3.00306529e-02 -3.90469134e-02 2.79986084e-01
1.20590365e+00 -7.74507999e-01 7.37635195e-02 -2.99777120e-01
4.88194704e-01 -7.01299846e-01 1.01301670e+00 3.99988890e-01
-1.07090962e+00 -4.52990770e-01 -1.23385704e+00 -2.95257390e-01
-4.25164133e-01 3.23146224e-01 5.87895572e-01 8.93999696e-01
-6.80530906e-01 2.54344523e-01 -5.56467950e-01 -3.68640095e-01
6.56105161e-01 2.89277524e-01 -3.56805086e-01 2.32305050e-01
-7.54379094e-01 1.53159833e+00 3.46478671e-01 -1.35115281e-01
-5.80899417e-01 -9.59226847e-01 -9.18661654e-01 -2.22780615e-01
7.27527618e-01 -7.28799701e-01 1.16532874e+00 -3.64332557e-01
-1.13376403e+00 1.46723342e+00 1.60231784e-01 -6.67819202e-01
1.04229808e+00 -3.03368568e-01 4.69085984e-02 -9.85425860e-02
8.67941007e-02 8.68947983e-01 7.71187603e-01 -1.33964682e+00
-8.94405901e-01 -5.34809768e-01 3.97148356e-02 3.16432983e-01
1.34913489e-01 -1.33086830e-01 -4.16390330e-01 -3.25042486e-01
2.31462374e-01 -9.27249134e-01 1.22499734e-01 5.72985888e-01
-3.60836715e-01 -1.20488159e-01 6.39078081e-01 -2.64084935e-01
7.17211783e-01 -2.15319157e+00 6.63957372e-02 -1.95477940e-02
2.62478858e-01 1.36528879e-01 1.69000283e-01 -9.21888277e-02
4.52894438e-03 9.37982798e-02 -3.98307830e-01 -6.94301844e-01
1.74260974e-01 -4.38774303e-02 -2.99718380e-01 7.48606086e-01
5.84077716e-01 8.18825066e-01 -8.54290366e-01 -4.92441237e-01
5.19738972e-01 5.30530095e-01 -5.61779618e-01 -7.40987733e-02
1.41096219e-01 1.65959492e-01 -1.57802150e-01 8.88368964e-01
9.57307339e-01 1.02700934e-01 -5.73562264e-01 -4.30731982e-01
-3.46684754e-01 3.54102284e-01 -1.29509079e+00 1.41724002e+00
-2.79979795e-01 6.18969381e-01 1.52874649e-01 -5.82557797e-01
9.22510445e-01 -3.28820258e-01 3.07878196e-01 -6.67856693e-01
1.91419423e-01 4.20494169e-01 -2.19155788e-01 -1.01084530e-01
4.73947674e-01 -2.12306023e-01 -7.36306906e-02 1.34176584e-02
1.10811315e-01 -5.74588478e-01 -1.86309516e-02 1.73915531e-02
7.64611542e-01 4.62361246e-01 4.35426027e-01 -8.88825357e-02
3.67520720e-01 8.53600428e-02 2.95573026e-01 9.25472200e-01
-5.54692447e-01 1.05848467e+00 5.44139743e-01 -1.18698187e-01
-1.10946214e+00 -1.42411470e+00 -6.34635866e-01 5.12474835e-01
3.76758933e-01 4.49437685e-02 -1.42051965e-01 -9.63902175e-01
9.05193150e-01 6.75650418e-01 -7.85130322e-01 -1.92548126e-01
-4.70320255e-01 -1.00452220e+00 6.94948137e-01 5.61471283e-01
2.72421867e-01 -5.11388481e-01 -9.72136915e-01 -2.66754746e-01
1.21902525e-01 -1.33056927e+00 -1.63461581e-01 6.28585756e-01
-5.89999259e-01 -1.31759560e+00 -6.36958301e-01 -3.54671568e-01
3.59013975e-01 5.57693422e-01 1.17452240e+00 -4.40313637e-01
-6.86140239e-01 4.18402374e-01 -1.83815062e-01 -6.78076804e-01
-1.81282774e-01 1.37971137e-02 9.42502357e-03 -3.19928735e-01
6.07179642e-01 -2.72254467e-01 -6.71402872e-01 5.59292078e-01
-4.68746543e-01 -1.75367787e-01 5.86933017e-01 6.38370454e-01
5.85276127e-01 -4.04864579e-01 2.43035778e-01 -4.09812003e-01
1.96340196e-02 -2.39928409e-01 -1.09733999e+00 -8.45110714e-02
-6.91578329e-01 -7.05451146e-02 -5.67952879e-02 -3.82516265e-01
-9.09909427e-01 2.56862521e-01 1.09395154e-01 -6.48842990e-01
-1.24280311e-01 -3.69284898e-01 -1.68786034e-01 -5.14739275e-01
9.20119226e-01 -1.13863647e-01 8.40601400e-02 -4.95597988e-01
6.89581394e-01 4.79180515e-01 6.31753564e-01 -3.43892157e-01
1.06897604e+00 9.30935383e-01 1.80978253e-01 -6.11607671e-01
-7.56975472e-01 -7.67491341e-01 -8.56463909e-01 -1.67377517e-01
8.40936720e-01 -1.12466717e+00 -8.34941208e-01 5.53892016e-01
-1.21823740e+00 5.16620986e-02 -5.48764765e-01 4.51348215e-01
-5.85092902e-01 4.45704132e-01 -1.97248325e-01 -1.02945471e+00
1.11847728e-01 -1.03106034e+00 1.36555398e+00 -6.45909980e-02
1.89677570e-02 -4.20443863e-01 -5.98732457e-02 2.60431767e-01
1.05836295e-01 5.54080606e-01 5.46160698e-01 -2.59165794e-01
-7.05485523e-01 -4.29380655e-01 -8.13884676e-01 5.90844095e-01
-1.11400858e-01 -1.22513033e-01 -1.37872016e+00 -1.63033336e-01
5.99749349e-02 -4.31507647e-01 1.22998166e+00 3.13326687e-01
7.83008873e-01 3.51935357e-01 -4.39723074e-01 6.99950814e-01
1.39115679e+00 -1.02992103e-01 3.43973339e-01 6.81500196e-01
8.41791928e-01 6.51917219e-01 6.33514643e-01 9.04017091e-02
2.72401243e-01 9.57806945e-01 8.05505514e-01 -1.68279991e-01
-3.88451546e-01 -1.95481241e-01 1.87051192e-01 -1.91666447e-02
8.25724378e-02 1.78633612e-02 -7.51874208e-01 4.29603308e-01
-1.58114779e+00 -6.08149171e-01 -2.41109073e-01 2.39439821e+00
5.14353275e-01 5.78245163e-01 5.84560692e-01 2.92209923e-01
4.01835769e-01 -4.18179631e-02 -7.25947022e-01 -1.29831294e-02
-4.69791263e-01 -1.20918684e-01 1.21586025e+00 6.77695632e-01
-1.44818890e+00 7.45777667e-01 6.08462429e+00 7.32601941e-01
-8.90587628e-01 2.05772281e-01 4.12831545e-01 -5.29236674e-01
1.30276933e-01 -4.64821383e-02 -1.29067135e+00 1.70213193e-01
1.21083215e-01 1.71220317e-01 2.61452165e-03 9.27288890e-01
-5.36254514e-03 -3.41791093e-01 -1.27390873e+00 1.11878157e+00
3.11838329e-01 -8.45099032e-01 -9.92095321e-02 2.51402874e-02
4.29803163e-01 3.74719799e-01 2.75191903e-01 1.51591063e-01
8.71264488e-02 -8.55260491e-01 1.26857555e+00 8.13864768e-02
7.97114015e-01 -5.15803695e-01 3.92107934e-01 1.54798672e-01
-1.03827047e+00 -1.89191476e-01 -2.48757720e-01 -3.34812365e-02
-1.60724930e-02 5.52414596e-01 -9.98357117e-01 4.74637926e-01
8.11428130e-01 6.55626059e-01 -7.86273837e-01 1.33962393e+00
-1.93647146e-01 -6.23379834e-03 -6.14186049e-01 -5.08568101e-02
1.21054932e-01 -4.62973118e-02 9.63844597e-01 1.29200041e+00
4.44510728e-02 -2.95855254e-01 -1.49389565e-01 1.31078660e+00
2.33161509e-01 -2.83611685e-01 -7.37948835e-01 6.96225643e-01
4.39349085e-01 1.11210179e+00 -5.95065475e-01 9.16214287e-02
-5.42220950e-01 6.08468652e-01 2.75484294e-01 1.56695127e-01
-9.66769278e-01 -4.11762834e-01 7.79527664e-01 3.80411267e-01
7.31975317e-01 -3.32169175e-01 -6.88514352e-01 -1.03533411e+00
4.92269903e-01 -4.27668810e-01 2.58889377e-01 -7.29776502e-01
-1.43857372e+00 2.20462933e-01 4.61448789e-01 -1.35954010e+00
-3.84413041e-02 -1.13370121e+00 -1.41027704e-01 8.43140841e-01
-2.00741291e+00 -1.26092720e+00 -3.47352266e-01 2.09363714e-01
9.13509056e-02 1.97594926e-01 3.21853220e-01 5.30088186e-01
-2.96900004e-01 8.23904514e-01 -1.04478031e-01 -1.64824665e-01
8.86789083e-01 -1.34590638e+00 4.60653305e-01 7.84832418e-01
-9.69851762e-02 1.28684700e-01 6.41072750e-01 -3.24625611e-01
-1.14585781e+00 -8.94317746e-01 6.39859200e-01 -1.21533847e+00
6.15442634e-01 -8.36881042e-01 -6.12200499e-01 4.62653965e-01
-5.11450350e-01 5.45206927e-02 7.46917874e-02 -6.87572286e-02
-6.65767372e-01 -2.36634046e-01 -1.38243794e+00 4.48107332e-01
1.41202784e+00 -5.20742178e-01 -4.83034164e-01 9.93578043e-03
8.76375675e-01 -5.88949561e-01 -4.11178112e-01 7.55568624e-01
7.66579926e-01 -1.28870714e+00 1.52526379e+00 -1.76356390e-01
6.41353056e-02 -4.98658746e-01 -3.18864495e-01 -7.47033775e-01
-9.80836004e-02 -3.26972097e-01 -5.91915026e-02 9.73373175e-01
4.24933285e-01 -7.25193918e-01 9.52674568e-01 3.73247594e-01
-1.80966035e-01 -6.08295560e-01 -1.29468989e+00 -1.08394837e+00
3.08426112e-01 -5.98738492e-01 2.60255426e-01 4.11350161e-01
-4.20943320e-01 9.76496190e-02 -1.29133329e-01 1.47116780e-01
9.60424960e-01 -3.63847166e-02 9.70029354e-01 -1.26903796e+00
-2.56472975e-01 -7.63878167e-01 -7.00371563e-01 -1.22813296e+00
-3.00382972e-01 -8.68122995e-01 -3.34962159e-02 -9.27318752e-01
2.01562077e-01 -5.89196503e-01 -7.93825239e-02 1.80416092e-01
3.50189619e-02 5.83776891e-01 3.79053265e-01 1.99349329e-01
-4.28535730e-01 5.50157666e-01 1.05750871e+00 -1.20943606e-01
-1.37032643e-01 -6.70175850e-02 -6.35275364e-01 8.34569931e-01
3.74965549e-01 -4.89718765e-01 -1.94612950e-01 -3.61051083e-01
2.26452723e-01 -5.93375444e-01 1.05794215e+00 -9.90537047e-01
8.68190750e-02 2.63469219e-01 4.43345517e-01 -1.08466566e+00
5.92220604e-01 -8.85790348e-01 -3.72175515e-01 4.09520268e-01
3.12332753e-02 -2.29986042e-01 5.42518377e-01 5.82935333e-01
1.45148903e-01 1.03488173e-02 1.08794534e+00 -6.16827421e-02
-7.46975005e-01 1.26741296e-02 1.01576552e-01 1.47108003e-01
1.10484290e+00 -6.60157561e-01 -3.22843075e-01 6.91708475e-02
-5.42758882e-01 2.36634821e-01 6.52417839e-01 6.41533673e-01
3.99105400e-01 -1.34720755e+00 -7.54227877e-01 4.09336835e-01
4.12070304e-01 2.46541828e-01 -2.03421161e-01 8.09251308e-01
-3.99195433e-01 4.84237492e-01 -7.73627236e-02 -1.00614607e+00
-1.34285641e+00 4.26297158e-01 5.61485469e-01 2.13906132e-02
-5.68514228e-01 9.48847711e-01 3.46074641e-01 -6.13466680e-01
6.70682371e-01 -7.14961588e-01 1.32533371e-01 1.67613834e-01
1.32546619e-01 5.76492369e-01 4.32174206e-01 -6.31555617e-01
-5.62003851e-01 8.08841348e-01 5.90758107e-04 5.48513941e-02
1.05859113e+00 -2.57736623e-01 4.01248097e-01 5.36161065e-01
1.30171418e+00 -1.39507219e-01 -1.58821416e+00 -1.37835994e-01
-2.97061592e-01 -6.00818753e-01 5.45146763e-02 -9.56352472e-01
-5.88059425e-01 7.84194708e-01 9.21857715e-01 1.03037786e-02
7.51111627e-01 3.02673876e-01 4.32918102e-01 1.56922743e-01
3.54883820e-01 -7.46416390e-01 5.56815765e-04 4.67949569e-01
7.94697821e-01 -1.65167379e+00 1.70342624e-01 -5.91803551e-01
-1.92885414e-01 7.22095728e-01 6.58046067e-01 -2.76988268e-01
5.41812956e-01 2.47410029e-01 -1.12984821e-01 -1.53238848e-01
-7.61835650e-02 -3.50050032e-01 4.65718567e-01 7.21485734e-01
3.00192554e-02 -3.31361070e-02 -1.42581478e-01 3.23961347e-01
-1.58949688e-01 -3.20287734e-01 1.43650457e-01 9.22657847e-01
-3.43977809e-01 -8.65186751e-01 -6.07040107e-01 2.05301270e-01
5.01283538e-03 6.51780665e-02 -6.19084597e-01 1.34258354e+00
3.85677904e-01 6.22772515e-01 7.01163337e-02 -3.02431971e-01
8.56946647e-01 -8.45479518e-02 8.91958296e-01 -4.08895433e-01
-2.05796331e-01 -1.84939086e-01 -6.17392920e-02 -4.11517709e-01
-3.53901833e-01 -8.46686840e-01 -8.02784324e-01 -5.33243902e-02
-7.12086082e-01 -4.36982781e-01 8.28240812e-01 6.48576081e-01
2.84222156e-01 2.09326103e-01 5.44328868e-01 -1.29745519e+00
-9.00136650e-01 -6.17505312e-01 -4.36660588e-01 3.17379236e-01
7.25013435e-01 -1.20627105e+00 -8.98017824e-01 -4.98867780e-01] | [7.771485328674316, -2.6213741302490234] |
0ce9c862-49c4-4bff-83f8-7d10ab88410d | online-multi-object-tracking-using-cnn-based | 1708.02843 | null | http://arxiv.org/abs/1708.02843v2 | http://arxiv.org/pdf/1708.02843v2.pdf | Online Multi-Object Tracking Using CNN-based Single Object Tracker with Spatial-Temporal Attention Mechanism | In this paper, we propose a CNN-based framework for online MOT. This
framework utilizes the merits of single object trackers in adapting appearance
models and searching for target in the next frame. Simply applying single
object tracker for MOT will encounter the problem in computational efficiency
and drifted results caused by occlusion. Our framework achieves computational
efficiency by sharing features and using ROI-Pooling to obtain individual
features for each target. Some online learned target-specific CNN layers are
used for adapting the appearance model for each target. In the framework, we
introduce spatial-temporal attention mechanism (STAM) to handle the drift
caused by occlusion and interaction among targets. The visibility map of the
target is learned and used for inferring the spatial attention map. The spatial
attention map is then applied to weight the features. Besides, the occlusion
status can be estimated from the visibility map, which controls the online
updating process via weighted loss on training samples with different occlusion
statuses in different frames. It can be considered as temporal attention
mechanism. The proposed algorithm achieves 34.3% and 46.0% in MOTA on
challenging MOT15 and MOT16 benchmark dataset respectively. | ['Nenghai Yu', 'Wanli Ouyang', 'Hongsheng Li', 'Bin Liu', 'Qi Chu', 'Xiaogang Wang'] | 2017-08-09 | online-multi-object-tracking-using-cnn-based-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Chu_Online_Multi-Object_Tracking_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Chu_Online_Multi-Object_Tracking_ICCV_2017_paper.pdf | iccv-2017-10 | ['online-multi-object-tracking'] | ['computer-vision'] | [-1.78782687e-01 -3.15526068e-01 5.02036558e-03 -2.93477535e-01
-3.10647100e-01 -1.76398501e-01 3.16461295e-01 -1.47146270e-01
-7.10232258e-01 5.72820425e-01 -1.86955258e-02 4.47563112e-01
5.81927560e-02 -4.97301608e-01 -8.93995345e-01 -9.96272862e-01
-1.80628717e-01 3.15910012e-01 8.45423758e-01 8.27650428e-02
1.88584238e-01 4.69589591e-01 -1.65609789e+00 1.33147119e-02
6.66959524e-01 1.33751214e+00 7.39920855e-01 7.14982927e-01
-1.97855696e-01 6.78781748e-01 -9.02399182e-01 -5.75530082e-02
2.45014891e-01 5.37686609e-03 -2.81757027e-01 1.23689156e-02
6.68950677e-01 -4.45508361e-01 -6.14348173e-01 9.40832913e-01
6.44143999e-01 3.63483042e-01 5.19992352e-01 -1.44918215e+00
-5.75116634e-01 3.49990964e-01 -9.77169037e-01 7.91277170e-01
-6.67661726e-02 4.99114931e-01 4.78371203e-01 -7.84581482e-01
3.96470845e-01 1.54005659e+00 4.43367809e-01 6.56573594e-01
-7.77282178e-01 -8.87289643e-01 6.57619119e-01 6.43926144e-01
-1.30137849e+00 -5.63720703e-01 6.66580081e-01 -3.00422013e-01
5.74323773e-01 2.43518427e-01 7.98148215e-01 7.63404846e-01
5.91139555e-01 9.06196117e-01 5.49783528e-01 5.14110699e-02
-3.81031521e-02 2.48863418e-02 -3.35968547e-02 4.95272219e-01
3.67916107e-01 1.75747782e-01 -6.91682518e-01 1.87059417e-01
6.47201896e-01 7.18781576e-02 -2.03927264e-01 -4.01819319e-01
-1.17005718e+00 4.10132647e-01 9.33925390e-01 -4.22282964e-02
-4.77945119e-01 5.86314082e-01 3.61727923e-01 1.13787062e-01
2.09850848e-01 -2.07879573e-01 -5.48709631e-01 1.13046058e-01
-5.69914222e-01 3.73057984e-02 1.17357112e-01 1.13068509e+00
7.91048229e-01 3.16489369e-01 -8.53747010e-01 6.37116134e-01
7.10870087e-01 8.09074521e-01 4.44296122e-01 -8.56325388e-01
2.55285084e-01 5.57470441e-01 4.42766815e-01 -1.09572208e+00
-3.76810730e-01 -6.87099934e-01 -4.37812358e-01 3.13326716e-01
2.65298069e-01 -6.76046163e-02 -1.11693990e+00 1.88398421e+00
1.04003143e+00 5.03012896e-01 -7.54871070e-02 9.46348906e-01
9.20153260e-01 6.61265194e-01 3.42948288e-01 -2.30743974e-01
1.44456244e+00 -1.27808332e+00 -9.57385838e-01 -2.54717946e-01
2.74029374e-01 -8.42439473e-01 5.76070487e-01 2.08494723e-01
-1.01469135e+00 -9.43706870e-01 -1.02685297e+00 8.32264051e-02
-2.79841840e-01 3.66003484e-01 4.58380997e-01 2.29282528e-01
-1.07753408e+00 2.31431827e-01 -9.74428535e-01 -5.33108294e-01
6.18001223e-01 8.08491230e-01 9.28778723e-02 -3.99336405e-02
-8.44408989e-01 7.52054155e-01 3.58485430e-01 5.13758421e-01
-1.57488644e+00 -5.52289069e-01 -5.51632345e-01 -9.98146739e-03
2.91219831e-01 -6.62644207e-01 1.07871497e+00 -1.14655399e+00
-1.41633308e+00 4.20716345e-01 -1.61101729e-01 -5.70458770e-01
4.48152572e-01 -1.97014317e-01 -4.00413632e-01 -2.71201640e-01
1.41499162e-01 1.04896295e+00 1.16715038e+00 -9.96686637e-01
-1.21343839e+00 -3.07719797e-01 5.92208020e-02 5.56453764e-01
-3.32852393e-01 4.31913463e-03 -8.76144826e-01 -3.83710623e-01
-3.68696675e-02 -7.95971930e-01 -8.58845487e-02 3.87548625e-01
1.52855754e-01 -5.76186895e-01 1.25830698e+00 -5.28036237e-01
1.15438616e+00 -2.09395242e+00 1.65241331e-01 -4.43464458e-01
1.88906357e-01 2.39913180e-01 -2.34912008e-01 -1.43775702e-01
4.91422355e-01 -5.68562746e-01 2.72872329e-01 -3.55926007e-01
-1.83723867e-01 6.60426468e-02 -2.05797907e-02 5.80587327e-01
2.13413864e-01 9.58465159e-01 -9.37542975e-01 -8.12984526e-01
4.18718010e-01 6.66897714e-01 -3.56335223e-01 3.82587880e-01
-3.60087067e-01 6.57137275e-01 -5.74370027e-01 6.89830363e-01
1.06031549e+00 -6.90248832e-02 -2.81579614e-01 -4.37926382e-01
-3.35255206e-01 -1.65244117e-01 -1.10278022e+00 1.70425928e+00
-4.38462585e-01 6.52747452e-01 3.27737868e-01 -5.85245907e-01
8.87843072e-01 -8.59208778e-02 4.60733712e-01 -6.81770563e-01
4.81427312e-01 -1.62806451e-01 1.52647674e-01 -6.60628617e-01
5.47147930e-01 7.22628534e-01 2.18220055e-01 -7.40855709e-02
1.24222352e-04 5.98674059e-01 3.63561623e-02 8.52237493e-02
9.83987212e-01 4.01227444e-01 -1.96746290e-01 -2.61070609e-01
7.41974533e-01 7.75252357e-02 8.96233380e-01 7.16156363e-01
-6.82091653e-01 8.48552287e-02 -3.61979038e-01 -7.97055304e-01
-7.11350441e-01 -9.31738198e-01 -1.33072957e-01 1.28790152e+00
7.86166668e-01 1.57097012e-01 -3.58310372e-01 -5.69226623e-01
2.45415531e-02 2.39759773e-01 -8.33272815e-01 -4.07525986e-01
-8.16544056e-01 -6.54533625e-01 5.48255770e-03 4.57934052e-01
6.72269464e-01 -1.41096926e+00 -9.16248798e-01 4.82119173e-01
-1.80344388e-01 -9.05133724e-01 -7.16779888e-01 6.51673451e-02
-7.03172386e-01 -1.03922081e+00 -6.04738593e-01 -6.76556587e-01
6.78256512e-01 6.40567183e-01 7.33574688e-01 1.65458053e-01
-3.55124950e-01 3.08108389e-01 -1.43422067e-01 -5.26350796e-01
1.69269592e-01 -2.33961809e-02 6.10399693e-02 4.68198806e-01
3.46992671e-01 -3.77471149e-01 -1.05655217e+00 4.68825519e-01
-5.78661203e-01 -1.91818133e-01 5.85487843e-01 6.00222290e-01
5.53613424e-01 -2.37685785e-01 3.62061411e-01 -1.45117342e-01
-7.46294186e-02 -5.51729023e-01 -9.67121005e-01 1.75258875e-01
-2.61145711e-01 -5.46256900e-02 3.40975940e-01 -9.05995965e-01
-1.07976162e+00 1.79659620e-01 3.99778605e-01 -7.12715030e-01
2.09768444e-01 -2.83764869e-01 -1.89847365e-01 -4.08960909e-01
4.13572550e-01 2.86642164e-01 -3.37733358e-01 -4.26967680e-01
5.78657277e-02 5.37408471e-01 5.44513166e-01 -4.33010757e-01
9.30974841e-01 7.55700469e-01 -1.16517646e-02 -4.16278332e-01
-9.20181811e-01 -4.55494255e-01 -3.98273140e-01 -7.57280231e-01
1.03829920e+00 -1.27039373e+00 -1.06121099e+00 5.53129852e-01
-1.16303945e+00 -3.29124749e-01 -6.66859746e-02 4.24120694e-01
-2.09002480e-01 1.65979624e-01 -2.66870916e-01 -9.34301317e-01
-4.34170306e-01 -1.11093056e+00 1.16238928e+00 7.97065139e-01
5.79994321e-01 -6.71823978e-01 -5.29227704e-02 -2.17628926e-02
6.44147694e-01 8.04258808e-02 -4.63566780e-02 -1.51821688e-01
-1.24263000e+00 -8.19803029e-02 -2.77798682e-01 -1.88295960e-01
-4.36853394e-02 -2.71818060e-02 -9.69564736e-01 -7.10860312e-01
-1.43953189e-01 -1.19210027e-01 9.63501751e-01 8.60544145e-01
1.11934769e+00 -1.11714564e-01 -7.57645965e-01 9.17164981e-01
1.42630589e+00 4.30664182e-01 5.80439925e-01 3.94616425e-01
8.95828247e-01 2.31863752e-01 1.25110745e+00 4.23802525e-01
4.46232587e-01 1.04858685e+00 8.52945566e-01 2.05267683e-01
-4.87892181e-01 -1.23600006e-01 5.47711253e-01 5.29432952e-01
2.78960198e-01 -3.39443326e-01 -3.57427776e-01 8.03591609e-01
-2.11348915e+00 -6.67694747e-01 -1.65285572e-01 2.30421376e+00
4.04612750e-01 1.06184170e-01 -4.21866812e-02 -6.54581308e-01
9.31539714e-01 5.76292053e-02 -8.54405701e-01 1.83550879e-01
-7.40407929e-02 -2.86086708e-01 6.53869212e-01 5.23931146e-01
-1.13358784e+00 1.02388692e+00 5.30318594e+00 8.81786883e-01
-1.19765592e+00 4.72303838e-01 2.64741391e-01 -3.58932227e-01
2.93371260e-01 8.02079588e-02 -1.25714302e+00 7.28295982e-01
7.02349484e-01 1.61409099e-02 4.14703041e-01 8.23221982e-01
2.72881031e-01 -2.08930701e-01 -7.68923163e-01 1.01116800e+00
6.13776222e-02 -9.38051462e-01 -9.92907137e-02 -2.34995186e-02
6.06861293e-01 3.11850607e-01 2.31615439e-01 4.07700300e-01
4.06079441e-01 -6.53523266e-01 7.04764068e-01 7.45497465e-01
4.50619191e-01 -4.72971976e-01 5.61021984e-01 4.66538928e-02
-1.80874419e+00 -3.14291358e-01 -8.97091985e-01 1.70596346e-01
-1.29759640e-01 1.11238666e-01 -5.40663779e-01 4.57151651e-01
1.11311185e+00 7.75282145e-01 -7.24371433e-01 1.49943542e+00
1.52453572e-01 1.78655595e-01 -6.07305110e-01 -4.53166338e-03
1.05467439e-01 1.48517773e-01 6.13338709e-01 8.76567245e-01
2.37982795e-01 -1.92471653e-01 4.32489485e-01 5.06778836e-01
1.12419173e-01 -4.09930050e-02 -2.99298435e-01 6.07448578e-01
7.14219749e-01 1.53858590e+00 -5.17382383e-01 -4.62642461e-01
-2.01438844e-01 9.45451438e-01 4.40989316e-01 3.82866710e-01
-1.29147232e+00 -1.17099203e-01 5.14955640e-01 -7.13084787e-02
8.52919757e-01 1.06163211e-01 3.67863327e-01 -9.33875382e-01
2.00513676e-02 -3.18555117e-01 5.07431924e-01 -9.23996568e-01
-9.66676056e-01 6.87482357e-01 -6.61245435e-02 -1.45610380e+00
4.75099653e-01 -2.68500775e-01 -7.38473833e-01 6.49113595e-01
-1.59385693e+00 -1.24982738e+00 -8.36277485e-01 5.48274517e-01
8.95778179e-01 -1.66982748e-02 9.03586447e-02 5.70129275e-01
-8.38580608e-01 7.92137682e-01 -8.84011481e-03 -1.56796649e-01
9.00448978e-01 -8.81370425e-01 1.16218269e-01 7.20716357e-01
-2.57983923e-01 2.18109280e-01 6.48784995e-01 -6.70489609e-01
-1.36704934e+00 -1.49172139e+00 5.97143352e-01 -4.82196242e-01
3.85652155e-01 -2.15883851e-01 -7.19748676e-01 7.21614659e-01
2.35511601e-01 4.39414710e-01 -1.37398079e-01 -3.06347132e-01
-4.24321089e-03 -6.40843511e-01 -1.16177213e+00 5.43081105e-01
1.11271358e+00 2.68269300e-01 -7.78357908e-02 3.09345156e-01
9.92307007e-01 -6.56372309e-01 -7.08659232e-01 2.93891549e-01
5.14507294e-01 -5.53919733e-01 9.13868546e-01 -2.99043119e-01
-3.56143087e-01 -1.02746153e+00 -3.32882479e-02 -8.06174695e-01
-6.03018224e-01 -5.83828092e-01 -3.64182532e-01 1.21157753e+00
-3.27742025e-02 -4.59858954e-01 7.04092801e-01 1.82019368e-01
-3.06803256e-01 -6.02873445e-01 -1.22233987e+00 -6.80159867e-01
-4.66679394e-01 2.68505160e-02 6.53957009e-01 5.55367589e-01
-7.34682679e-01 2.13596612e-01 -6.97642565e-01 4.96745050e-01
9.15714383e-01 -1.43447459e-01 1.16015816e+00 -1.08459592e+00
-1.02322929e-01 -1.10618934e-01 -6.23535752e-01 -1.34242320e+00
-9.27818567e-02 -6.19027495e-01 2.79901296e-01 -1.25499713e+00
4.01722997e-01 -5.44130981e-01 -7.80669510e-01 3.48814458e-01
-2.44174019e-01 3.22284102e-01 2.50887662e-01 3.99294615e-01
-1.16659880e+00 8.43531668e-01 1.33611917e+00 -4.94634271e-01
-1.96199477e-01 -5.06337136e-02 -2.79248565e-01 4.21497643e-01
7.30941117e-01 -8.00860465e-01 -2.23340526e-01 -6.89108253e-01
-2.99275756e-01 -2.87663266e-02 5.59026599e-01 -1.47032237e+00
6.36833668e-01 -2.10029602e-01 7.76666939e-01 -1.09914494e+00
5.40391326e-01 -1.11441922e+00 3.59103888e-01 7.25695312e-01
-1.57787669e-02 1.53504033e-02 4.66701835e-01 1.01910484e+00
3.05364400e-01 3.58865708e-02 8.44328523e-01 1.59244150e-01
-7.97257066e-01 7.50051558e-01 -7.68435597e-02 -1.61517113e-01
1.21657789e+00 -2.86924362e-01 -1.26441136e-01 -8.22347179e-02
-6.47135735e-01 8.12339902e-01 3.53804916e-01 6.82971239e-01
4.87254024e-01 -1.57896161e+00 -7.01438427e-01 -2.93352529e-02
2.03525007e-01 7.49070719e-02 7.08374619e-01 1.14410722e+00
-2.60204554e-01 1.63777322e-01 -3.21466923e-01 -1.05513096e+00
-1.38552833e+00 7.26175487e-01 6.26300633e-01 -2.07597002e-01
-4.22932416e-01 1.05438411e+00 7.69448221e-01 1.07999653e-01
6.32737279e-01 -1.46782756e-01 -3.09056491e-01 -4.30699587e-01
6.61192775e-01 1.97346285e-01 -4.29789513e-01 -8.24589491e-01
-6.31648004e-01 8.40350449e-01 -3.91392082e-01 2.09487155e-01
1.04847312e+00 -5.00739217e-01 1.36340812e-01 3.14084291e-01
7.62890995e-01 -2.79296756e-01 -1.98023176e+00 -3.98738474e-01
-3.50599200e-01 -6.68037832e-01 8.74159262e-02 -7.89057136e-01
-1.36589694e+00 6.43492043e-01 1.29053628e+00 -3.28334004e-01
1.11630988e+00 -4.21810448e-02 7.16197193e-01 1.26421511e-01
2.81085730e-01 -1.03089666e+00 1.90014005e-01 3.31684023e-01
8.00221384e-01 -1.25457609e+00 -5.80276437e-02 -3.28226457e-03
-2.97379464e-01 7.63364077e-01 1.32246530e+00 -2.86612719e-01
4.97704923e-01 8.33095163e-02 4.09751870e-02 -2.95648396e-01
-8.94324183e-01 -2.60738760e-01 1.89525351e-01 8.36950958e-01
-1.23574376e-01 -1.18814394e-01 -2.05778524e-01 8.69290680e-02
4.04826134e-01 -2.33051464e-01 1.48512617e-01 6.97452724e-01
-7.88529515e-01 -6.64681971e-01 -6.57987952e-01 2.36035526e-01
-2.76805609e-01 1.36109710e-01 -1.29641350e-02 6.63412988e-01
5.38695455e-01 8.92544270e-01 2.32689172e-01 -4.31059152e-01
1.02843270e-01 -3.90594214e-01 5.68744838e-01 -2.38007024e-01
-6.68910503e-01 2.08497316e-01 -2.65286803e-01 -4.97061968e-01
-4.92918909e-01 -7.78325558e-01 -1.10417640e+00 -1.03697322e-01
-6.44099712e-01 8.83959234e-02 6.17163539e-01 5.75931430e-01
4.73065555e-01 1.01514447e+00 6.58649862e-01 -1.01761091e+00
-2.75795966e-01 -1.17173529e+00 -1.19229071e-01 3.31381708e-01
4.78508919e-01 -1.03486443e+00 -2.27721766e-01 -8.07515010e-02] | [6.3127875328063965, -2.1071486473083496] |
6c5b09dc-1eed-4385-aa72-8ee0231fff71 | musical-instrument-sound-classification-with | 1512.07370 | null | http://arxiv.org/abs/1512.07370v1 | http://arxiv.org/pdf/1512.07370v1.pdf | Musical instrument sound classification with deep convolutional neural network using feature fusion approach | A new musical instrument classification method using convolutional neural
networks (CNNs) is presented in this paper. Unlike the traditional methods, we
investigated a scheme for classifying musical instruments using the learned
features from CNNs. To create the learned features from CNNs, we not only used
a conventional spectrogram image, but also proposed multiresolution recurrence
plots (MRPs) that contain the phase information of a raw input signal.
Consequently, we fed the characteristic timbre of the particular instrument
into a neural network, which cannot be extracted using a phase-blinded
representations such as a spectrogram. By combining our proposed MRPs and
spectrogram images with a multi-column network, the performance of our proposed
classifier system improves over a system that uses only a spectrogram.
Furthermore, the proposed classifier also outperforms the baseline result from
traditional handcrafted features and classifiers. | ['Lee Taejin', 'Park Taejin'] | 2015-12-23 | null | null | null | null | ['sound-classification'] | ['audio'] | [ 3.51010054e-01 -3.90317053e-01 1.29495293e-01 -1.51114285e-01
-5.95149815e-01 -8.43122423e-01 4.27523166e-01 -1.92463044e-02
-4.29269075e-01 5.56557655e-01 1.44316629e-01 1.40161887e-01
-4.47611392e-01 -9.54560935e-01 -5.50238550e-01 -6.63921535e-01
-1.65878966e-01 -4.22241658e-01 -2.00376093e-01 -2.52725512e-01
4.48666573e-01 5.81680059e-01 -1.91570723e+00 4.87655163e-01
3.66103917e-01 1.59186292e+00 1.04824692e-01 8.29723537e-01
-4.94658276e-02 8.30781460e-01 -9.37910676e-01 1.45720571e-01
3.13483596e-01 -5.16808152e-01 -6.01713657e-01 -4.44412768e-01
6.51099861e-01 6.95957616e-02 -3.43319058e-01 9.63472605e-01
4.94992703e-01 2.53931403e-01 4.98306185e-01 -1.03568971e+00
-3.75167668e-01 8.28269184e-01 -3.71390909e-01 8.34727138e-02
3.94787431e-01 -1.77342787e-01 1.31538057e+00 -6.00174189e-01
2.60940552e-01 9.42782760e-01 1.20698261e+00 1.19652063e-01
-1.01172245e+00 -1.03724790e+00 -3.12113911e-01 2.52018631e-01
-1.17543197e+00 -1.20443009e-01 1.30111146e+00 -5.06733298e-01
9.42453921e-01 4.00745422e-01 1.09039342e+00 8.57738137e-01
6.58920705e-02 6.59527242e-01 1.02901316e+00 -5.62853158e-01
-1.21371835e-01 -4.06977057e-01 1.41661808e-01 6.94231093e-01
-1.87215298e-01 1.74489915e-01 -9.47338462e-01 -1.55018628e-01
1.06021655e+00 -2.21543163e-01 -3.59544784e-01 1.12879863e-02
-1.53365266e+00 6.93026483e-01 6.10437095e-01 6.00594521e-01
-4.27062511e-01 4.21629220e-01 5.94808400e-01 3.82326633e-01
1.79405838e-01 1.08204603e+00 -5.93649268e-01 -3.86842340e-01
-1.20662689e+00 3.05656910e-01 5.18211126e-01 3.29773217e-01
8.00614536e-01 4.48214442e-01 -7.24210739e-02 7.60675788e-01
-2.70559400e-01 2.89698809e-01 7.92675555e-01 -1.01492834e+00
3.30071479e-01 4.96477455e-01 -2.83289492e-01 -1.34393990e+00
-7.40022957e-01 -7.60562837e-01 -1.06279206e+00 2.71640092e-01
5.24057508e-01 -1.83140039e-01 -6.14822268e-01 1.64953363e+00
-2.86970139e-01 3.69098812e-01 1.40041903e-01 9.90740955e-01
9.76160228e-01 4.98321831e-01 -6.11638844e-01 -3.71087715e-02
1.27392364e+00 -7.93873012e-01 -8.06582451e-01 2.42512733e-01
2.76761413e-01 -8.92540753e-01 9.52666700e-01 9.77109969e-01
-7.78670311e-01 -1.07397199e+00 -1.64129865e+00 5.73369227e-02
-3.45367849e-01 5.36889136e-01 9.89700377e-01 6.16204143e-01
-8.29033911e-01 1.37229335e+00 -5.68467915e-01 1.07722946e-01
3.22900027e-01 3.75691831e-01 -4.93672282e-01 7.98112452e-01
-1.18881595e+00 4.61787671e-01 6.10395730e-01 1.45504460e-01
-3.71675044e-01 -6.41035080e-01 -7.80895293e-01 2.52703071e-01
-7.89965242e-02 -3.52966040e-01 1.31933987e+00 -1.06421983e+00
-1.98855996e+00 1.85869843e-01 7.49338344e-02 -4.30563420e-01
2.37163641e-02 -5.07408917e-01 -3.53750527e-01 2.14799136e-01
-2.22759932e-01 1.87755033e-01 1.06971335e+00 -7.97695041e-01
-6.10346675e-01 -5.72071299e-02 3.92253287e-02 -1.90417245e-01
-4.48715001e-01 -2.59989202e-01 2.05854982e-01 -1.02601421e+00
4.65990156e-01 -8.22704315e-01 6.49528429e-02 -3.85064065e-01
-6.89916968e-01 -6.41201586e-02 6.76314712e-01 -3.43878150e-01
1.19324839e+00 -2.33451176e+00 3.25142592e-02 3.39077473e-01
4.94727828e-02 1.01791270e-01 -9.79103968e-02 2.61233002e-01
-4.61442649e-01 4.02446724e-02 -6.18846454e-02 1.16997175e-01
-1.67096674e-01 -1.87131450e-01 -5.55613995e-01 3.31371963e-01
2.92842746e-01 6.46851361e-01 -8.70132506e-01 -1.29052579e-01
1.21028028e-01 3.08712959e-01 -5.73676109e-01 3.49801295e-02
4.86065000e-02 4.75098640e-01 -6.94105551e-02 4.90204781e-01
4.42158163e-01 -4.35922630e-02 1.50508583e-01 -8.88928711e-01
-3.07335705e-01 5.94161630e-01 -1.26597321e+00 2.09226227e+00
-6.63318396e-01 9.89381135e-01 -4.41325456e-01 -9.23600197e-01
1.33293700e+00 6.14870906e-01 7.38686740e-01 -4.68349397e-01
1.44959047e-01 1.23661064e-01 1.26224220e-01 -2.55992323e-01
6.02636039e-01 -1.51150435e-01 -2.31609151e-01 4.69190627e-01
6.19221926e-01 -5.71476445e-02 -4.90497947e-02 -3.28829348e-01
1.12211585e+00 4.33226019e-01 2.20188946e-01 -1.88902346e-03
4.81907368e-01 -8.58385339e-02 6.97798669e-01 7.70388305e-01
3.05745453e-01 9.27182913e-01 3.99764478e-01 -6.52281046e-01
-9.50708568e-01 -7.45735765e-01 -1.50333256e-01 9.45428848e-01
-2.16786832e-01 -8.52924526e-01 -3.06060404e-01 -2.36469507e-01
2.49184333e-02 -1.45724744e-01 -6.61130786e-01 2.30884962e-02
-6.46391332e-01 -4.96334672e-01 1.14812875e+00 6.62431479e-01
7.03317583e-01 -1.11977351e+00 -8.55473220e-01 4.52169150e-01
-1.45133018e-01 -8.78234148e-01 -3.37822080e-01 7.36771226e-01
-9.86994863e-01 -1.13221192e+00 -5.47789454e-01 -6.63814485e-01
-2.37062931e-01 -4.87159267e-02 6.64460599e-01 -2.83421576e-01
-3.68563056e-01 4.47198655e-03 -4.43075716e-01 -5.03843606e-01
1.70726210e-01 2.37074688e-01 3.20980310e-01 2.50176847e-01
-8.99107829e-02 -1.18875945e+00 -5.49896777e-01 -3.10702980e-01
-5.24689674e-01 1.39374435e-01 6.57350779e-01 8.65191877e-01
4.39661652e-01 1.16761781e-01 8.57766867e-01 -4.75316375e-01
7.60596037e-01 1.23197965e-01 -2.16976985e-01 -1.50357053e-01
-3.06254864e-01 2.24624231e-01 8.56771529e-01 -8.04404616e-01
-7.06775308e-01 3.66999567e-01 -1.88043833e-01 -5.26782632e-01
2.78424323e-02 6.38135970e-01 2.24996313e-01 -6.22314550e-02
7.73928642e-01 2.50172429e-02 -8.51220563e-02 -7.87175953e-01
3.83417040e-01 7.86478698e-01 1.01476240e+00 -5.74125707e-01
6.92424119e-01 2.57762551e-01 2.55951911e-01 -7.02349186e-01
-8.81367445e-01 -5.37380993e-01 -8.01508307e-01 -2.20536128e-01
6.23495460e-01 -8.65705431e-01 -1.19526827e+00 5.78070939e-01
-1.35092509e+00 1.00384727e-01 -3.67226899e-01 8.84232879e-01
-7.11217701e-01 1.68950543e-01 -7.61180580e-01 -9.68025446e-01
-4.42009002e-01 -5.79734504e-01 9.52261686e-01 2.38478050e-01
-5.26222587e-01 -5.87090611e-01 3.66448164e-01 -8.09261650e-02
3.60398173e-01 5.81075311e-01 9.42981005e-01 -5.31629682e-01
-1.36174366e-01 -1.99516609e-01 -4.71226051e-02 4.42221940e-01
4.34723616e-01 7.28581846e-02 -1.76106274e+00 9.78082139e-03
2.60010604e-02 -4.51115131e-01 9.97366488e-01 2.58065969e-01
1.38984990e+00 -2.17497244e-01 1.64989695e-01 1.00996768e+00
1.34021354e+00 4.03659940e-01 6.17590487e-01 6.28381312e-01
7.76673853e-01 1.15778200e-01 2.77426153e-01 5.65615833e-01
-6.38218969e-02 5.14614463e-01 2.30621710e-01 -5.28926253e-02
-1.19963951e-01 -3.55954707e-01 2.21606880e-01 1.24200773e+00
-7.16727793e-01 4.41068739e-01 -5.58519959e-01 3.56339067e-01
-1.86082971e+00 -1.14067709e+00 -5.03836013e-02 1.89312422e+00
8.23306084e-01 3.61902528e-02 1.53716326e-01 1.05670345e+00
4.37601596e-01 3.19624603e-01 -4.34468687e-01 -2.90833116e-01
-3.56960684e-01 1.05914605e+00 2.73991108e-01 -7.02883154e-02
-1.38513100e+00 5.87879360e-01 6.92667198e+00 6.16421521e-01
-1.64608216e+00 -2.34246105e-01 -3.07003409e-01 -8.04204345e-02
6.14968725e-02 -2.64201432e-01 -9.83913690e-02 -2.43620407e-02
8.44755650e-01 -7.20673287e-03 4.45254415e-01 7.79708087e-01
-9.76254195e-02 2.13341251e-01 -1.06918919e+00 1.52613699e+00
-1.23671725e-01 -1.44636011e+00 6.80111498e-02 -1.29205242e-01
3.79230380e-01 -3.82830739e-01 1.61585972e-01 2.51356095e-01
-1.43190667e-01 -1.15092516e+00 7.92031467e-01 8.17914128e-01
8.03424656e-01 -9.35539186e-01 7.06003904e-01 -4.70439382e-02
-1.58825433e+00 -1.95910856e-01 -9.12470222e-02 -5.89971781e-01
-3.87202203e-01 5.76231539e-01 -6.64291799e-01 8.97377133e-01
8.57202709e-01 1.07272995e+00 -5.20479560e-01 1.17585671e+00
-3.00823227e-02 7.96813548e-01 -1.99272484e-01 5.10059809e-03
1.16287343e-01 -1.55084173e-03 3.06656539e-01 1.16535115e+00
5.90354979e-01 -2.13421240e-01 -2.44193360e-01 7.01531589e-01
-5.98367713e-02 2.61064410e-01 -4.89916354e-01 -3.32742155e-01
2.91121423e-01 1.52809477e+00 -6.37615800e-01 -2.79330701e-01
-1.60540402e-01 6.57470584e-01 1.51413813e-01 1.42894164e-01
-3.61019343e-01 -1.04137528e+00 6.86666369e-01 -2.86018640e-01
3.79674375e-01 -2.89638162e-01 -5.69088042e-01 -1.16934013e+00
-6.20373785e-02 -7.72583723e-01 1.59211338e-01 -9.74258363e-01
-1.21383667e+00 8.58985603e-01 -5.43653905e-01 -1.83920610e+00
-4.03948367e-01 -6.68492436e-01 -7.08206236e-01 9.73374128e-01
-1.32627916e+00 -9.51061308e-01 -3.40489030e-01 6.93690062e-01
2.22128972e-01 -4.95367944e-01 1.33490682e+00 2.12957233e-01
-6.54397458e-02 4.97315347e-01 -6.34676740e-02 5.69496870e-01
6.68685973e-01 -1.41590774e+00 2.05343395e-01 2.77521104e-01
7.99869180e-01 6.08415306e-01 3.93269151e-01 -1.66850716e-01
-1.28830719e+00 -8.30040991e-01 4.14368749e-01 5.27077839e-02
6.65174842e-01 -2.05983460e-01 -8.59064639e-01 3.48311007e-01
2.21805930e-01 -2.01266155e-01 8.94677758e-01 4.68252182e-01
-7.20953286e-01 -3.68855447e-01 -7.13584542e-01 6.46990612e-02
8.38310003e-01 -9.99060810e-01 -8.55064869e-01 -1.07317649e-01
4.60879356e-01 -3.60302627e-01 -1.11517322e+00 4.74902064e-01
1.22847569e+00 -8.59261692e-01 9.73924577e-01 -5.43178618e-01
5.04404545e-01 -4.41862226e-01 -1.92141578e-01 -1.48633933e+00
-5.35034418e-01 -5.74416041e-01 -2.91477554e-02 8.59177768e-01
2.57077545e-01 -3.14345032e-01 5.87268710e-01 -4.26689804e-01
-1.36254475e-01 -4.36871618e-01 -8.88228059e-01 -8.06789458e-01
-2.72227917e-02 -7.50790060e-01 6.94779754e-01 1.20906365e+00
3.19693685e-01 4.19958025e-01 -4.42276806e-01 1.55996099e-01
4.24776763e-01 5.00217617e-01 6.37057900e-01 -1.70683420e+00
-3.70128393e-01 -5.32338560e-01 -8.92040491e-01 -5.81333816e-01
-1.95329487e-02 -1.01389313e+00 5.30009065e-03 -1.19964719e+00
-7.35221431e-02 -2.92624980e-01 -7.40644872e-01 6.13385975e-01
1.96907401e-01 6.92598522e-01 4.76928890e-01 1.50580510e-01
-1.48728698e-01 4.92376536e-01 1.23865545e+00 -2.75319368e-01
-3.21341425e-01 5.95480539e-02 -5.32428145e-01 9.15623903e-01
9.44320619e-01 -3.69623542e-01 -8.59676301e-02 8.69447291e-02
5.40846527e-01 2.99317718e-01 4.42641914e-01 -1.67098391e+00
3.74348700e-01 2.51301438e-01 7.17284203e-01 -7.86279500e-01
7.05946922e-01 -4.40687209e-01 2.60372728e-01 1.99640647e-01
-2.89589852e-01 -3.78880240e-02 3.03095937e-01 3.19928944e-01
-7.11466849e-01 -1.14200667e-01 4.18878019e-01 -7.36577138e-02
-4.45734382e-01 -1.11397550e-01 -3.52854699e-01 -5.19337952e-01
2.38179415e-01 -2.73931146e-01 -1.92967892e-01 -3.80890429e-01
-9.51471150e-01 -5.32930613e-01 -2.08348036e-01 4.02931958e-01
6.60747588e-01 -1.74636960e+00 -3.12946200e-01 1.92397356e-01
-4.45426814e-02 -1.37598813e-01 1.23434938e-01 4.72902358e-01
-4.69691813e-01 4.05527800e-01 -6.97437942e-01 -5.48475683e-01
-1.24666381e+00 3.00358951e-01 5.01992345e-01 6.39626384e-02
-8.18780363e-01 5.74009001e-01 -1.62064433e-01 -5.30391037e-01
2.98320711e-01 -7.83485293e-01 -6.37846768e-01 5.20710707e-01
5.03712833e-01 6.38821125e-02 1.89043388e-01 -6.30656421e-01
-2.49567285e-01 9.09433722e-01 4.84763563e-01 -4.66724187e-01
1.70226729e+00 5.59802055e-01 -2.73964945e-02 1.09952760e+00
1.21368599e+00 1.51795700e-01 -8.50517750e-01 -2.19489187e-01
3.21214721e-02 -3.73961747e-01 2.03093752e-01 -7.03142166e-01
-1.26292241e+00 9.00971055e-01 7.09774196e-01 5.24503946e-01
1.20288706e+00 -5.77012897e-01 6.46081746e-01 7.00975657e-01
3.40101004e-01 -1.00602818e+00 3.67649645e-01 7.82387435e-01
9.58913445e-01 -6.90607905e-01 -3.04309577e-02 -1.79784950e-02
-3.76997292e-01 1.81300449e+00 2.21517324e-01 -4.98465389e-01
8.50051463e-01 3.40531051e-01 3.86616975e-01 -1.93674907e-01
-4.65817183e-01 -4.18754965e-01 6.07859671e-01 4.59425688e-01
7.51256764e-01 -1.65767848e-01 -1.07245287e-02 1.21791339e+00
-1.14517426e+00 1.85141400e-01 5.51805079e-01 8.65563095e-01
-2.60638177e-01 -8.67578328e-01 -5.54001391e-01 5.09373426e-01
-5.03102779e-01 -8.68712664e-02 -3.56424600e-01 5.94124258e-01
5.45095265e-01 7.36133337e-01 2.85985768e-01 -1.19369304e+00
5.03910005e-01 3.76353025e-01 7.02429414e-01 -4.37455863e-01
-1.03419864e+00 1.20811805e-01 -8.99708364e-03 -4.35738891e-01
-9.18756664e-01 -2.45129749e-01 -1.11239874e+00 7.09409565e-02
-4.10145223e-01 1.65180668e-01 8.79807949e-01 7.79125750e-01
-6.40215632e-03 1.13556838e+00 8.28033626e-01 -1.21002901e+00
-1.06437594e-01 -1.36237657e+00 -8.79900277e-01 1.33388460e-01
6.99252307e-01 -6.48807824e-01 -3.15356016e-01 1.03483789e-01] | [15.800642967224121, 5.256643772125244] |
ad284e83-e5c9-4533-986c-b04bc3d4f704 | uncorrelated-semi-paired-subspace-learning | 2011.11124 | null | https://arxiv.org/abs/2011.11124v1 | https://arxiv.org/pdf/2011.11124v1.pdf | Uncorrelated Semi-paired Subspace Learning | Multi-view datasets are increasingly collected in many real-world applications, and we have seen better learning performance by existing multi-view learning methods than by conventional single-view learning methods applied to each view individually. But, most of these multi-view learning methods are built on the assumption that at each instance no view is missing and all data points from all views must be perfectly paired. Hence they cannot handle unpaired data but ignore them completely from their learning process. However, unpaired data can be more abundant in reality than paired ones and simply ignoring all unpaired data incur tremendous waste in resources. In this paper, we focus on learning uncorrelated features by semi-paired subspace learning, motivated by many existing works that show great successes of learning uncorrelated features. Specifically, we propose a generalized uncorrelated multi-view subspace learning framework, which can naturally integrate many proven learning criteria on the semi-paired data. To showcase the flexibility of the framework, we instantiate five new semi-paired models for both unsupervised and semi-supervised learning. We also design a successive alternating approximation (SAA) method to solve the resulting optimization problem and the method can be combined with the powerful Krylov subspace projection technique if needed. Extensive experimental results on multi-view feature extraction and multi-modality classification show that our proposed models perform competitively to or better than the baselines. | ['Ren-cang Li', 'Chungen Shen', 'Lei-Hong Zhang', 'Li Wang'] | 2020-11-22 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [ 4.63127531e-02 -4.11115527e-01 -2.78093129e-01 -2.60497153e-01
-1.00184286e+00 -7.82066345e-01 5.73619008e-01 -5.97329259e-01
-1.63413465e-01 5.82325757e-01 3.61529440e-01 1.29013762e-01
-3.31822723e-01 -3.77616078e-01 -4.53637928e-01 -1.03286481e+00
2.20289052e-01 4.54989523e-01 -1.53822929e-01 5.51667847e-02
-1.43568203e-01 2.48750567e-01 -1.52456391e+00 5.61948895e-01
5.75540721e-01 5.15873551e-01 5.83840646e-02 4.75007504e-01
1.53418511e-01 4.80189174e-01 1.09450161e-01 -2.07261041e-01
4.75041598e-01 -4.14539903e-01 -4.81973171e-01 3.17786336e-01
6.57615781e-01 -2.74151564e-01 -3.05722594e-01 9.36377287e-01
6.07156038e-01 4.44109738e-02 5.88353634e-01 -1.36534417e+00
-3.96427155e-01 2.17379093e-01 -8.06460440e-01 -5.01937754e-02
4.95738715e-01 -4.63275105e-01 1.10176098e+00 -1.57200408e+00
6.90886140e-01 9.44012403e-01 6.57276392e-01 3.12082142e-01
-1.24118841e+00 -4.16889966e-01 2.14654326e-01 4.15843517e-01
-1.22404861e+00 -6.57609761e-01 1.11186004e+00 -4.15433824e-01
4.04004842e-01 4.93866980e-01 5.63019335e-01 1.33760560e+00
-1.35241836e-01 1.22993946e+00 1.56314921e+00 -4.88823175e-01
-4.01913635e-02 2.90403217e-01 1.72102541e-01 6.60900533e-01
1.06544062e-01 6.06214628e-02 -7.63604462e-01 -3.57851535e-01
5.14662266e-01 4.57635790e-01 -4.66134042e-01 -1.31272900e+00
-1.50348973e+00 8.86380315e-01 -2.57391781e-02 2.31899589e-01
-2.54637510e-01 -5.45311153e-01 4.13942158e-01 3.20955396e-01
3.07126135e-01 -3.75583619e-02 -4.61786896e-01 1.68096513e-01
-9.80173767e-01 6.85587153e-02 6.82822406e-01 9.10728872e-01
6.45996571e-01 -8.32318738e-02 2.94051826e-01 8.75379622e-01
2.26115525e-01 6.81910455e-01 5.65095007e-01 -8.48054230e-01
5.02212644e-01 5.05038142e-01 -2.33081207e-02 -8.40523422e-01
-4.34016526e-01 -4.36075240e-01 -1.07101631e+00 4.46136333e-02
3.88283014e-01 1.44867720e-02 -5.53212464e-01 1.60283530e+00
3.97003949e-01 2.89804101e-01 4.25634712e-01 7.50656068e-01
8.90032232e-01 3.14878315e-01 -6.77860796e-01 -7.96869516e-01
9.06578183e-01 -1.15791416e+00 -7.08187461e-01 8.63072649e-02
4.19936031e-01 -7.77130902e-01 8.52006793e-01 8.09601188e-01
-8.46629739e-01 -5.40658176e-01 -9.87756550e-01 2.40979433e-01
-2.19087839e-01 2.32077882e-01 6.64041281e-01 6.07989967e-01
-7.07358897e-01 4.42603707e-01 -9.09346700e-01 -2.91256011e-01
1.91308722e-01 2.66222328e-01 -1.00280488e+00 -5.01364768e-01
-6.06694698e-01 6.02973163e-01 2.66076952e-01 2.46510625e-01
-6.55602336e-01 -5.03184140e-01 -8.66918445e-01 -3.10598403e-01
5.97972155e-01 -9.32359338e-01 6.92140162e-01 -7.40175724e-01
-1.13071835e+00 7.85184920e-01 -4.06659126e-01 1.44025505e-01
5.11467874e-01 -5.50749362e-01 -6.03860378e-01 2.76059836e-01
1.21174149e-01 6.21153601e-03 1.04914021e+00 -1.76809216e+00
-3.78849417e-01 -8.01200986e-01 -2.12235674e-01 5.03630996e-01
-5.09793401e-01 -1.92227319e-01 -5.87113917e-01 -5.29956222e-01
7.48912990e-01 -1.20060718e+00 -2.00247109e-01 -2.71385521e-01
-3.51920724e-01 2.53927231e-01 1.20728004e+00 -5.19506574e-01
9.62399542e-01 -2.20229387e+00 6.70374930e-01 1.72847450e-01
2.82793850e-01 2.14721844e-01 -1.69667974e-01 6.34888351e-01
-4.00118649e-01 -4.59171832e-01 -1.58311784e-01 -4.36407864e-01
-2.21411228e-01 3.58879298e-01 -4.33315694e-01 7.77318716e-01
-4.98314261e-01 6.65614545e-01 -9.05662537e-01 -4.20308083e-01
3.88951689e-01 5.28615654e-01 -4.93233472e-01 1.52887002e-01
6.10557497e-01 7.50552893e-01 -2.75424302e-01 7.40409076e-01
7.55160093e-01 -4.31040585e-01 3.62286508e-01 -5.15736699e-01
5.65270707e-03 -4.65996921e-01 -1.65371001e+00 1.91247094e+00
-5.04748344e-01 1.06601126e-01 1.15829840e-01 -1.12884820e+00
3.91695231e-01 5.68781674e-01 8.90988171e-01 -9.87369269e-02
-2.87471205e-01 2.97627300e-01 -1.93246588e-01 -4.78262097e-01
1.23937294e-01 -4.50158983e-01 1.18442379e-01 4.24184620e-01
3.99070203e-01 1.53713122e-01 -1.58273578e-02 2.32109413e-01
5.39509773e-01 3.20368081e-01 6.26520753e-01 8.54704157e-02
8.74472260e-01 -4.68128562e-01 6.83875203e-01 6.03602886e-01
-1.67589262e-01 8.85505557e-01 1.34142414e-01 -4.38664734e-01
-8.60413313e-01 -1.31005597e+00 -1.52515039e-01 9.80186820e-01
-1.06027666e-02 -5.20517290e-01 -2.82452703e-01 -1.14961195e+00
-1.30719990e-01 2.78149396e-01 -4.57906544e-01 1.18889414e-01
-3.72378200e-01 -8.72282207e-01 1.57676768e-02 6.33051395e-01
2.72412479e-01 -4.92056757e-01 -1.93131521e-01 -1.40114531e-01
-3.55823964e-01 -1.33241272e+00 -2.73576438e-01 2.13905647e-01
-1.02426422e+00 -1.18335211e+00 -7.95176744e-01 -6.04054928e-01
6.66952193e-01 8.40370178e-01 7.58528829e-01 -3.58900517e-01
5.48372567e-02 9.11675215e-01 -5.02539039e-01 -1.82979792e-01
2.37095803e-02 -2.31819004e-01 7.39768088e-01 4.76068407e-01
3.90166909e-01 -9.89128351e-01 -2.77049929e-01 3.71596843e-01
-7.86180794e-01 1.10031441e-01 6.46790922e-01 1.28101218e+00
8.04758847e-01 -1.06436640e-01 5.12377918e-01 -1.13911569e+00
-8.56762007e-03 -3.25860828e-01 -2.50225991e-01 4.16027695e-01
-5.33905089e-01 -5.11399209e-02 8.89090717e-01 -1.41652063e-01
-9.50842321e-01 4.90111738e-01 1.91775486e-01 -9.88999248e-01
-2.14027852e-01 5.29689789e-01 -5.25643468e-01 -1.45129159e-01
2.76000082e-01 6.70341492e-01 3.88309993e-02 -6.12279534e-01
5.30244589e-01 5.54366350e-01 3.73005539e-01 -2.64490962e-01
1.05820870e+00 9.84480858e-01 1.38223052e-01 -1.00802326e+00
-1.11219573e+00 -1.03782725e+00 -1.22056544e+00 -1.76579639e-01
5.56615949e-01 -1.26043510e+00 -3.26961637e-01 4.25098866e-01
-5.73846579e-01 4.78766263e-01 -1.39882535e-01 9.14298594e-01
-7.79831469e-01 9.16269183e-01 -1.74387664e-01 -6.98932052e-01
-7.11381761e-03 -9.56452429e-01 1.06930399e+00 -9.16191563e-02
2.05766752e-01 -1.12009835e+00 3.60789567e-01 6.17474318e-01
-5.80124408e-02 1.81266725e-01 6.03649795e-01 -8.98768246e-01
-1.76818237e-01 -2.51068503e-01 1.06100701e-01 4.80245739e-01
2.35915437e-01 -4.08964843e-01 -1.27682912e+00 -7.88081884e-01
4.24319267e-01 -4.64803189e-01 9.90403414e-01 2.82117486e-01
9.54787731e-01 1.26680732e-02 -4.21900988e-01 8.62580895e-01
1.66183686e+00 -2.28143126e-01 1.98643565e-01 5.92200384e-02
1.09333551e+00 6.34990096e-01 4.42076325e-01 5.40069222e-01
2.85790414e-01 7.38562286e-01 2.61991560e-01 -3.79182845e-02
1.40309870e-01 -2.69717962e-01 5.49444616e-01 1.48302734e+00
-3.33233565e-01 1.23106062e-01 -7.26133704e-01 4.71314549e-01
-2.01754832e+00 -1.32591057e+00 -3.64052981e-01 2.50668788e+00
2.02584177e-01 -2.36211985e-01 3.03358138e-01 2.56063700e-01
4.00193959e-01 3.58040154e-01 -4.74600047e-01 2.49818459e-01
-4.70378160e-01 -1.26038879e-01 1.81954652e-01 2.52946556e-01
-1.42843556e+00 5.08050084e-01 6.36257601e+00 5.01515806e-01
-8.52602363e-01 2.38947973e-01 9.90437344e-02 -3.58116895e-01
-3.26215088e-01 6.80227876e-02 -6.98450148e-01 3.06723663e-03
5.80463052e-01 1.57963522e-02 3.41169834e-01 8.61285448e-01
4.51956727e-02 5.28069660e-02 -1.35710764e+00 1.52725792e+00
6.13996565e-01 -9.27300513e-01 1.66832358e-01 1.76655337e-01
9.89992499e-01 6.16143607e-02 1.54403806e-01 3.13635141e-01
-6.77703470e-02 -5.89457989e-01 2.99446344e-01 5.33209920e-01
6.22089326e-01 -7.00953543e-01 6.38442278e-01 6.26059711e-01
-1.21059370e+00 -2.58013219e-01 -3.21262956e-01 8.95892605e-02
1.82125568e-01 5.89990616e-01 -2.25477532e-01 1.34932220e+00
5.35908043e-01 1.16282916e+00 -6.47669256e-01 8.39782357e-01
3.35422643e-02 3.26919794e-01 -2.80209243e-01 6.30339801e-01
5.24786636e-02 -5.72655618e-01 7.67378151e-01 7.43739903e-01
3.38079542e-01 6.58333451e-02 4.55531865e-01 1.62539676e-01
3.64555448e-01 2.85663873e-01 -1.16004694e+00 2.57872522e-01
7.99032450e-02 1.41549432e+00 -4.55134869e-01 -2.92944103e-01
-1.18293285e+00 1.22578287e+00 3.27387333e-01 4.95369703e-01
-6.25355065e-01 2.40375564e-01 2.66970962e-01 -3.65550697e-01
6.20983124e-01 -2.18617246e-01 -1.03000075e-01 -1.98343289e+00
3.26488316e-01 -1.00972474e+00 6.25047326e-01 -5.22422075e-01
-1.64833713e+00 3.57470602e-01 3.90993953e-02 -1.85143697e+00
-2.60095716e-01 -6.48053765e-01 -3.83610040e-01 5.72528720e-01
-1.43606782e+00 -1.65152252e+00 -1.06244512e-01 1.01855600e+00
4.63165045e-01 -4.95139778e-01 1.03944945e+00 3.34026307e-01
-5.59244454e-01 4.80874211e-01 7.37454653e-01 -4.02720794e-02
9.31938529e-01 -1.26236165e+00 -3.93832982e-01 8.82305861e-01
8.28939140e-01 6.65078938e-01 3.94211262e-01 -3.58075917e-01
-1.88409758e+00 -7.64913619e-01 5.37433624e-01 -8.23603451e-01
5.50772786e-01 -4.21561956e-01 -8.36758196e-01 9.12920117e-01
1.41693696e-01 4.82354313e-01 1.36512148e+00 5.09042203e-01
-5.66474795e-01 -1.08011395e-01 -6.67476296e-01 2.45345429e-01
9.18096066e-01 -7.79010594e-01 -7.82391906e-01 3.72810751e-01
4.37461659e-02 -1.33018583e-01 -8.61593366e-01 6.56011879e-01
7.78937280e-01 -1.38444483e+00 1.19657898e+00 -8.27059090e-01
1.69693217e-01 -4.47625399e-01 -6.97960079e-01 -1.28049684e+00
-3.16092789e-01 -2.93482810e-01 -5.27819276e-01 1.18512499e+00
1.83245540e-01 -4.61219192e-01 8.13756824e-01 2.44386524e-01
-2.73398142e-02 -7.01401472e-01 -8.95232081e-01 -1.06314528e+00
-2.39989981e-02 -2.74439692e-01 7.86851496e-02 1.31751871e+00
2.03896370e-02 5.51613927e-01 -8.58461857e-01 4.52968746e-01
1.05171561e+00 8.01425278e-01 8.97724688e-01 -1.19052219e+00
-6.81670487e-01 -1.06639611e-02 -2.96720386e-01 -7.66352654e-01
1.42724365e-01 -9.44853067e-01 -2.60078579e-01 -1.24135470e+00
8.33334029e-01 -8.85476023e-02 -4.48732942e-01 2.30527952e-01
-4.73181993e-01 2.84503400e-01 3.32115382e-01 5.29534936e-01
-6.89765573e-01 6.53907835e-01 1.06973040e+00 1.82812929e-01
-1.42939791e-01 1.72932729e-01 -6.25219405e-01 1.12370718e+00
1.70161933e-01 -1.54165715e-01 -5.91361225e-01 -1.29006132e-01
1.16482876e-01 2.97585696e-01 3.71952653e-01 -9.85315800e-01
2.43087739e-01 6.30084518e-03 5.07899344e-01 -7.43965209e-01
4.47127163e-01 -1.11707854e+00 2.42624298e-01 6.83632046e-02
9.52795297e-02 -9.54124928e-02 -3.28692794e-01 1.01713431e+00
-5.80554664e-01 -3.95837091e-02 5.63941240e-01 -2.13793695e-01
-8.02921534e-01 3.14276397e-01 1.28575578e-01 -9.90312472e-02
1.03805578e+00 -3.14450264e-01 1.72549844e-01 -3.70188683e-01
-1.24818325e+00 5.67044690e-02 5.61391175e-01 3.48706633e-01
6.43155813e-01 -1.70356965e+00 -6.64774597e-01 5.04027128e-01
5.00370145e-01 -2.80806482e-01 5.55979729e-01 1.30661428e+00
2.25490704e-01 3.70887071e-01 -1.51254922e-01 -8.19164574e-01
-1.45567966e+00 1.03246868e+00 -1.06292881e-01 -3.67157996e-01
-6.08879328e-01 5.90407848e-01 4.47068781e-01 -7.29082763e-01
-9.73531157e-02 3.60712111e-01 -4.59789634e-01 4.50785637e-01
5.63577831e-01 4.83546227e-01 1.45106073e-02 -9.94916856e-01
-3.44903022e-01 8.99404109e-01 -1.49572998e-01 -9.14720669e-02
1.65243292e+00 -2.87566602e-01 -1.36918992e-01 9.88407433e-01
1.44937742e+00 4.04461235e-01 -1.05301118e+00 -4.81444538e-01
-2.24313900e-01 -6.51017427e-01 -4.74535674e-02 -4.15502220e-01
-1.04407585e+00 8.77457857e-01 5.95855117e-01 -9.11141708e-02
1.34027278e+00 -7.87707716e-02 4.91730452e-01 4.83046532e-01
5.58695138e-01 -8.51619482e-01 4.02735136e-02 3.34123492e-01
7.29240894e-01 -1.62806427e+00 3.84709030e-01 -5.63152373e-01
-7.36960471e-01 1.12993121e+00 5.02323091e-01 -9.48134065e-02
7.09458828e-01 -7.00716674e-02 3.54291275e-02 -2.13230953e-01
-5.93467593e-01 -1.71975195e-01 4.87212896e-01 5.66002727e-01
2.59299934e-01 -6.32565990e-02 -1.62763715e-01 6.14372611e-01
1.89362869e-01 -1.81615740e-01 5.24667621e-01 9.41136599e-01
-2.86132377e-02 -1.36023664e+00 -5.86653709e-01 4.88438874e-01
-2.48902261e-01 4.98449281e-02 -3.18687677e-01 7.59822130e-01
-4.09674942e-02 6.04743242e-01 -5.07670701e-01 -4.29012060e-01
2.85995662e-01 4.04503971e-01 6.99763954e-01 -6.30940139e-01
-9.19590145e-02 5.70347726e-01 -1.28854513e-01 -6.43905342e-01
-1.01466322e+00 -1.23780990e+00 -6.38092995e-01 1.68626979e-01
-4.89955455e-01 6.08245246e-02 2.50987619e-01 1.03489792e+00
1.50135979e-01 4.82288599e-02 9.77856874e-01 -8.49378765e-01
-7.40336418e-01 -5.72315395e-01 -9.09419894e-01 5.39220810e-01
4.11023885e-01 -7.96357036e-01 -4.81912285e-01 1.48695692e-01] | [8.363335609436035, 4.551107406616211] |
bb429903-02b6-4c40-ade6-52d83d091d41 | searching-for-discriminative-words-in | 2211.14631 | null | https://arxiv.org/abs/2211.14631v1 | https://arxiv.org/pdf/2211.14631v1.pdf | Searching for Discriminative Words in Multidimensional Continuous Feature Space | Word feature vectors have been proven to improve many NLP tasks. With recent advances in unsupervised learning of these feature vectors, it became possible to train it with much more data, which also resulted in better quality of learned features. Since it learns joint probability of latent features of words, it has the advantage that we can train it without any prior knowledge about the goal task we want to solve. We aim to evaluate the universal applicability property of feature vectors, which has been already proven to hold for many standard NLP tasks like part-of-speech tagging or syntactic parsing. In our case, we want to understand the topical focus of text documents and design an efficient representation suitable for discriminating different topics. The discriminativeness can be evaluated adequately on text categorisation task. We propose a novel method to extract discriminative keywords from documents. We utilise word feature vectors to understand the relations between words better and also understand the latent topics which are discussed in the text and not mentioned directly but inferred logically. We also present a simple way to calculate document feature vectors out of extracted discriminative words. We evaluate our method on the four most popular datasets for text categorisation. We show how different discriminative metrics influence the overall results. We demonstrate the effectiveness of our approach by achieving state-of-the-art results on text categorisation task using just a small number of extracted keywords. We prove that word feature vectors can substantially improve the topical inference of documents' meaning. We conclude that distributed representation of words can be used to build higher levels of abstraction as we demonstrate and build feature vectors of documents. | ['Maria Bielikova', 'Michal Barla', 'Marius Sajgalik'] | 2022-11-26 | null | null | null | null | ['part-of-speech-tagging'] | ['natural-language-processing'] | [ 2.05066532e-01 2.08486468e-01 -4.23631102e-01 -5.94151556e-01
-8.12504709e-01 -7.25728095e-01 1.03654301e+00 5.47607720e-01
-5.52432597e-01 5.33061266e-01 5.78882813e-01 -2.33930126e-01
-4.91963536e-01 -7.30695486e-01 -3.13873053e-01 -7.84265757e-01
-1.57711685e-01 5.80795228e-01 3.16354930e-01 -2.25054830e-01
5.60677528e-01 3.52284700e-01 -1.97171390e+00 4.80932415e-01
7.56938279e-01 8.56281996e-01 5.01109600e-01 5.69931507e-01
-7.02554941e-01 2.96162814e-01 -6.91166818e-01 5.43993264e-02
-2.21447408e-01 -2.37385616e-01 -1.16242802e+00 1.44036487e-01
3.32270890e-01 3.13925147e-01 1.94793344e-01 7.97642589e-01
2.05363989e-01 2.39104941e-01 1.21854222e+00 -9.50618088e-01
-4.14647877e-01 9.86587882e-01 -1.11031801e-01 1.22845732e-01
3.96440595e-01 -5.89415610e-01 1.64079356e+00 -6.22061312e-01
5.79085410e-01 1.30281031e+00 4.02635068e-01 3.90634149e-01
-1.11234403e+00 -3.40272933e-01 3.13583553e-01 2.96728581e-01
-1.19199395e+00 -1.16687156e-01 6.37118280e-01 -5.18917263e-01
1.12446320e+00 3.03112239e-01 2.65502930e-01 1.19101572e+00
2.84116238e-01 8.00206721e-01 9.36779082e-01 -8.54357779e-01
2.82778051e-02 4.29037452e-01 7.51662314e-01 6.20879948e-01
3.14047188e-01 -4.85295564e-01 -3.22609395e-01 -2.11578533e-01
9.74058658e-02 8.44130963e-02 -3.56647044e-01 -2.18711972e-01
-1.12435961e+00 1.16282177e+00 1.36026174e-01 1.07237542e+00
-2.79493600e-01 1.02570437e-01 3.92798126e-01 1.13457851e-01
5.77460229e-01 6.35501802e-01 -8.42277348e-01 -1.07709259e-01
-8.80630970e-01 8.12413841e-02 9.00467813e-01 6.59554124e-01
8.75879407e-01 -4.94687229e-01 -3.86026829e-01 9.74953353e-01
4.13482577e-01 3.91164452e-01 9.55568731e-01 -4.18393821e-01
3.46289337e-01 7.40523040e-01 -1.89828292e-01 -1.05679440e+00
-4.66281980e-01 -2.77405947e-01 -6.46273732e-01 -3.40177715e-01
1.76252276e-01 1.98201627e-01 -1.04183340e+00 1.48950016e+00
-3.90570275e-02 -2.99275786e-01 8.36964697e-02 3.26639891e-01
5.94860494e-01 9.61710870e-01 2.05833390e-01 -2.57251054e-01
1.74177301e+00 -5.38077116e-01 -7.34663665e-01 1.27585148e-02
1.10843062e+00 -7.77135074e-01 1.05005312e+00 5.45106709e-01
-4.39307392e-01 -5.11101961e-01 -7.62400448e-01 -4.38096225e-02
-1.00580645e+00 1.39258534e-01 9.61838722e-01 7.82716930e-01
-8.57725680e-01 6.09718025e-01 -6.01221025e-01 -4.72564131e-01
2.95462489e-01 5.85133731e-01 -5.48441172e-01 1.35004774e-01
-1.29142642e+00 9.02382255e-01 7.36537457e-01 -3.84235531e-01
-4.03093308e-01 -5.14872670e-01 -9.30545688e-01 3.81648839e-01
2.77755260e-01 -5.20285785e-01 1.00412893e+00 -8.29280496e-01
-1.18869734e+00 8.44594240e-01 -4.24272895e-01 -3.30942124e-01
-2.25167528e-01 -2.08593100e-01 -3.18843961e-01 2.10781008e-01
2.86535263e-01 4.92437631e-01 6.77423418e-01 -1.16052389e+00
-8.58103693e-01 -3.15106839e-01 -1.59998864e-01 -1.85581177e-01
-8.81096840e-01 3.89822423e-02 -2.29453996e-01 -5.00941515e-01
7.41997734e-02 -7.06128061e-01 -2.38064211e-02 -5.11438966e-01
-3.72043341e-01 -9.23719585e-01 8.60085487e-01 -3.79233807e-01
1.18144202e+00 -1.95478570e+00 2.14112297e-01 2.84478188e-01
1.64625943e-01 1.28997177e-01 -6.57967329e-02 7.50178933e-01
-4.98843603e-02 4.13277358e-01 -1.12908192e-01 -3.04075986e-01
3.34307015e-01 4.95449692e-01 -6.92144752e-01 2.85307527e-01
1.78797692e-01 8.60349834e-01 -6.88679397e-01 -7.43759811e-01
1.80524319e-01 6.02595270e-01 -4.97401506e-01 6.65255040e-02
-3.41112673e-01 2.72661243e-02 -7.95638084e-01 1.65393442e-01
3.41977179e-01 -1.66537032e-01 1.02642640e-01 -6.28534853e-02
-4.30914573e-02 5.50349236e-01 -1.08620429e+00 1.51962614e+00
-8.78928065e-01 7.09049165e-01 -5.27561128e-01 -1.47049820e+00
8.26456785e-01 3.44190091e-01 3.35527867e-01 -3.60167235e-01
1.71820313e-01 9.45229679e-02 -1.78068712e-01 -5.27475119e-01
4.99879897e-01 -2.35947713e-01 -4.31130856e-01 4.26024228e-01
5.01048028e-01 -9.20838043e-02 2.96984851e-01 2.76600569e-01
1.00282514e+00 -1.80167928e-01 6.41380191e-01 -6.16640389e-01
7.76926339e-01 -1.17985159e-01 -4.77596372e-02 7.44124234e-01
4.92754966e-01 2.32588753e-01 6.72722757e-01 -2.11163685e-01
-5.78157902e-01 -8.09493482e-01 -4.82885420e-01 1.54441595e+00
-1.41297087e-01 -7.38696516e-01 -2.76262879e-01 -1.09032559e+00
-3.74765992e-02 8.47624183e-01 -8.61428201e-01 3.96319777e-02
-4.42319691e-01 -7.02048302e-01 3.61651063e-01 3.92820299e-01
-8.08140114e-02 -9.54655528e-01 -3.26421320e-01 3.11048119e-03
1.01413392e-02 -9.55216229e-01 -5.88180572e-02 7.75980830e-01
-8.28733027e-01 -8.89277279e-01 -6.08987987e-01 -9.20095801e-01
6.50257766e-01 2.31368601e-01 1.22420311e+00 3.71947773e-02
-2.48575330e-01 5.68173349e-01 -8.52423370e-01 -3.67970079e-01
-5.43891072e-01 4.54072922e-01 -5.01192845e-02 -1.32524148e-01
5.26147068e-01 -4.99684006e-01 -2.98592836e-01 1.21640071e-01
-1.30110896e+00 -3.63885760e-01 7.43547499e-01 8.99852097e-01
2.28707060e-01 3.11541438e-01 3.26247483e-01 -1.09133446e+00
6.40230060e-01 -4.38798845e-01 -2.63407826e-01 3.45436126e-01
-6.45888805e-01 8.99778366e-01 6.74526334e-01 -4.66001570e-01
-8.84167612e-01 -1.34553492e-01 -2.72508711e-01 3.50267380e-01
-4.67857182e-01 6.13508463e-01 -2.58570284e-01 2.13736787e-01
4.53729033e-01 2.85194397e-01 -3.74922454e-01 -6.07365489e-01
5.76744556e-01 1.12187576e+00 -1.18035085e-01 -7.72197187e-01
5.99206805e-01 3.12769502e-01 1.44192532e-01 -1.13548744e+00
-1.01880515e+00 -1.18296754e+00 -7.19076276e-01 2.61635631e-01
8.63584697e-01 -5.17914891e-01 -5.89338899e-01 -1.31769210e-01
-1.34654987e+00 1.80883497e-01 -1.64471835e-01 4.52246845e-01
-3.00651997e-01 6.79918826e-01 -6.92809150e-02 -6.94570005e-01
-2.95976698e-01 -9.49545383e-01 1.53517163e+00 -8.98115039e-02
-3.95794094e-01 -1.46650279e+00 2.47524962e-01 -2.02138368e-02
2.36108944e-01 -2.53018737e-01 1.36032128e+00 -1.28175807e+00
-6.81697503e-02 -2.08234236e-01 -6.20053336e-02 4.79309231e-01
4.75829154e-01 -8.51651058e-02 -1.12185121e+00 -1.10281847e-01
-4.59594801e-02 -6.70558214e-02 1.41538537e+00 3.21107566e-01
1.16438806e+00 -3.09506536e-01 -6.56990707e-01 1.58299074e-01
1.48713493e+00 6.45822659e-03 5.98776042e-01 2.65732884e-01
6.14069402e-01 9.81438756e-01 5.80857337e-01 2.84613103e-01
6.74566403e-02 7.84196198e-01 1.87874466e-01 2.50708550e-01
1.62372932e-01 2.70070489e-02 4.18377250e-01 1.01036966e+00
3.44338901e-02 -5.53347647e-01 -9.25718963e-01 6.77490354e-01
-1.62219048e+00 -7.98954248e-01 -1.47040516e-01 1.89480150e+00
8.93031240e-01 1.22521833e-01 -6.07123449e-02 4.33083504e-01
5.63949466e-01 2.10799202e-01 4.18765783e-01 -5.72678745e-01
1.69818535e-01 5.76927066e-01 2.55099326e-01 6.90250874e-01
-1.13975143e+00 1.09201431e+00 5.63238430e+00 1.11644411e+00
-1.04654765e+00 4.07867543e-02 2.25754485e-01 4.63651687e-01
-5.01382053e-01 1.24252714e-01 -1.37939703e+00 2.86710083e-01
1.02266920e+00 -3.25524099e-02 -3.04111958e-01 8.87514710e-01
-1.72960863e-01 -7.20704272e-02 -1.23086894e+00 7.05337405e-01
2.61758238e-01 -1.17122686e+00 5.65892220e-01 2.39518344e-01
4.73771691e-01 -3.59824181e-01 -6.59939572e-02 4.92570221e-01
1.79857817e-02 -1.03828657e+00 1.33086115e-01 3.15513641e-01
3.69134635e-01 -7.61755228e-01 9.51426268e-01 3.34104717e-01
-1.10008824e+00 4.25798222e-02 -6.05029166e-01 -4.44386154e-02
-8.27974826e-02 8.54066074e-01 -1.10488105e+00 7.02467978e-01
2.65673727e-01 6.37825966e-01 -7.16634750e-01 7.24292159e-01
-2.67404586e-01 6.73952341e-01 -2.49528036e-01 -5.78914225e-01
5.28064489e-01 1.67693734e-01 4.28777367e-01 1.67015672e+00
2.77432919e-01 -3.00638646e-01 9.90128592e-02 4.44503963e-01
2.01061949e-01 6.34535849e-01 -6.55308843e-01 -2.71519005e-01
-5.28631324e-04 1.37104034e+00 -1.07254255e+00 -3.97963643e-01
-2.75736988e-01 8.63419175e-01 1.76728547e-01 -1.70278847e-02
-4.85628724e-01 -6.27760589e-01 6.00465834e-01 -8.35304260e-02
7.19639719e-01 -2.96493798e-01 1.71049207e-01 -1.19299853e+00
-1.08785003e-01 -5.66578150e-01 3.83815050e-01 -2.82203197e-01
-1.37313306e+00 7.02179611e-01 3.77734900e-01 -9.32317495e-01
-6.10686302e-01 -1.11296105e+00 -4.26063776e-01 7.18005061e-01
-1.49701750e+00 -1.09071815e+00 -6.16514422e-02 4.46288049e-01
7.42614806e-01 -1.95791945e-01 1.26144254e+00 -1.90930054e-01
1.32535864e-02 3.64322156e-01 1.85215652e-01 1.31169796e-01
7.21809030e-01 -1.51718175e+00 6.02077246e-02 4.98774558e-01
8.35658371e-01 8.48823249e-01 8.37242484e-01 -4.71391559e-01
-1.21173465e+00 -7.53448486e-01 1.44056129e+00 -6.01549923e-01
9.01476920e-01 -6.51873052e-01 -8.55118573e-01 4.38548654e-01
2.43046463e-01 -2.31061041e-01 1.01976573e+00 6.69302642e-01
-4.85628724e-01 1.02985287e-02 -6.85121953e-01 2.47320607e-01
6.33689523e-01 -4.02809530e-01 -1.21485090e+00 3.59762907e-01
8.52902949e-01 4.19862568e-01 -6.66168809e-01 1.54855147e-01
5.16871750e-01 -6.17997706e-01 9.24555838e-01 -8.40351939e-01
3.74020845e-01 7.16491416e-02 -2.59514719e-01 -1.34270048e+00
-2.33431950e-01 -3.46469939e-01 3.51052880e-02 1.61668420e+00
5.13525426e-01 -7.97211349e-01 5.54414332e-01 1.26552358e-02
5.13330102e-02 -6.09270632e-01 -8.21049154e-01 -8.15804899e-01
2.76086628e-01 -5.73519707e-01 2.35094488e-01 8.83221686e-01
2.74696529e-01 6.71296060e-01 -1.69329923e-02 -2.18340904e-02
2.75086552e-01 3.21936786e-01 4.61024076e-01 -1.66433966e+00
-3.65245491e-01 -4.84480739e-01 -8.87899280e-01 -1.22681201e+00
6.66327477e-01 -1.11469841e+00 1.44411221e-01 -1.59056020e+00
2.93124646e-01 -3.05193841e-01 -4.91186202e-01 5.99703848e-01
-2.58616924e-01 -1.13867722e-01 -7.88750350e-02 3.49769369e-02
-6.89434409e-01 5.12562692e-01 9.40435886e-01 -2.48420790e-01
-7.06157684e-02 5.14256991e-02 -7.78489828e-01 6.56898856e-01
7.65303373e-01 -7.60821939e-01 -3.61471593e-01 -9.70625654e-02
3.44423980e-01 -4.62524801e-01 1.58735052e-01 -7.24926949e-01
5.50593846e-02 -4.95489128e-02 3.26619297e-01 -5.49179435e-01
1.03669032e-01 -9.79778707e-01 -5.67680597e-01 3.43405604e-01
-5.47020853e-01 -2.32017487e-01 -4.83056493e-02 4.79985297e-01
-3.82746458e-01 -7.30295658e-01 3.42372596e-01 -7.70973265e-02
-8.17428708e-01 -2.18959332e-01 -4.77903247e-01 3.18394490e-02
7.49385417e-01 3.74927074e-02 -2.01835901e-01 -2.39970148e-01
-5.34377933e-01 -8.00525621e-02 3.35623361e-02 5.40036619e-01
3.46590668e-01 -9.94901896e-01 -4.88806129e-01 5.98800145e-02
2.94026226e-01 -3.80598813e-01 -1.42713234e-01 6.17498159e-01
-1.55340552e-01 1.03183115e+00 1.12118930e-01 -7.39348114e-01
-1.43246305e+00 6.56607509e-01 -2.93508768e-01 -5.43481588e-01
-6.05675638e-01 6.24811172e-01 3.76065880e-01 -1.40660107e-01
2.56944239e-01 -7.75737107e-01 -7.91674674e-01 6.54712796e-01
5.02350748e-01 -1.01501212e-01 1.50706083e-01 -6.00266278e-01
-5.27495980e-01 8.50994468e-01 -3.59117329e-01 -5.84274717e-02
1.49148786e+00 -2.62802951e-02 -1.82640046e-01 6.31950557e-01
1.65464044e+00 3.13417614e-01 -3.64406973e-01 -1.08620688e-01
5.26153088e-01 -1.37309611e-01 1.66064754e-01 -4.84363228e-01
-5.08305609e-01 9.67629075e-01 4.13553923e-01 8.02820623e-01
8.37255657e-01 5.60065031e-01 6.10707045e-01 7.30144382e-01
2.71820217e-01 -8.00131202e-01 -5.73370196e-02 6.12980664e-01
5.99361122e-01 -1.18574393e+00 1.60795555e-01 -5.50417960e-01
-4.08389598e-01 1.53185856e+00 -1.85723558e-01 -7.70670995e-02
7.52501488e-01 1.21690340e-01 -7.78592825e-02 -4.04222280e-01
-7.12746799e-01 -6.88476443e-01 7.59376526e-01 4.68002617e-01
7.44327426e-01 -1.08954221e-01 -5.90999782e-01 6.22163177e-01
-2.92891353e-01 -5.74160397e-01 1.59233838e-01 7.92065859e-01
-8.89760673e-01 -1.67957354e+00 -2.01454014e-01 3.85416538e-01
-7.41255641e-01 -1.95402458e-01 -4.79898244e-01 9.45418715e-01
1.34534150e-01 8.15582871e-01 -8.32662880e-02 -1.72423214e-01
-4.96709161e-02 5.43979526e-01 4.45989281e-01 -1.04147983e+00
-4.23321694e-01 -6.57673031e-02 1.29283935e-01 -3.22895586e-01
-6.89388752e-01 -4.95084941e-01 -1.14185882e+00 3.37674767e-01
-5.56041837e-01 7.83728898e-01 9.78264034e-01 1.27938628e+00
1.33276656e-01 6.16128981e-01 6.29382730e-01 -8.99561465e-01
-4.95469868e-01 -1.15909028e+00 -4.98833090e-01 4.47865248e-01
2.05294713e-01 -8.52137327e-01 -7.01026082e-01 2.30912179e-01] | [10.424077987670898, 8.381789207458496] |
93f3e823-96f4-49fc-8681-abffc1088848 | synthetically-generating-human-like-data-for | 2304.07280 | null | https://arxiv.org/abs/2304.07280v1 | https://arxiv.org/pdf/2304.07280v1.pdf | Synthetically Generating Human-like Data for Sequential Decision Making Tasks via Reward-Shaped Imitation Learning | We consider the problem of synthetically generating data that can closely resemble human decisions made in the context of an interactive human-AI system like a computer game. We propose a novel algorithm that can generate synthetic, human-like, decision making data while starting from a very small set of decision making data collected from humans. Our proposed algorithm integrates the concept of reward shaping with an imitation learning algorithm to generate the synthetic data. We have validated our synthetic data generation technique by using the synthetically generated data as a surrogate for human interaction data to solve three sequential decision making tasks of increasing complexity within a small computer game-like setup. Different empirical and statistical analyses of our results show that the synthetically generated data can substitute the human data and perform the game-playing tasks almost indistinguishably, with very low divergence, from a human performing the same tasks. | ['Prithviraj Dasgupta', 'Bryan Brandt'] | 2023-04-14 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 4.53399926e-01 7.03091979e-01 5.82079351e-01 -7.31689706e-02
-3.03267449e-01 -6.69221580e-01 9.49734688e-01 2.58289482e-02
-9.51298773e-01 9.42696810e-01 -2.12564304e-01 -1.79589570e-01
-6.41949847e-02 -6.82291985e-01 -3.89478266e-01 -5.08875430e-01
1.51876882e-02 1.02622819e+00 2.61947751e-01 -4.91565526e-01
4.14806515e-01 -9.61803645e-02 -1.95014644e+00 2.42638886e-01
1.12130189e+00 6.96678817e-01 3.78033906e-01 1.12396550e+00
4.60956216e-01 9.31277096e-01 -8.05273950e-01 -3.94470304e-01
1.01267374e+00 -6.92256153e-01 -6.11342013e-01 4.97574992e-02
-3.94249320e-01 -3.46417576e-01 7.05659688e-02 7.61239111e-01
6.30354524e-01 3.97198588e-01 9.34845805e-01 -1.44694674e+00
-3.37680340e-01 3.22158784e-01 -3.60206328e-02 -2.42992103e-01
8.51468444e-01 6.63254142e-01 4.90738600e-01 -3.32651109e-01
8.54043603e-01 1.27199805e+00 4.12453413e-01 1.16310751e+00
-1.56992579e+00 -4.69985247e-01 -5.35149634e-01 -2.36121863e-01
-1.20330179e+00 -1.86262339e-01 4.95982438e-01 -6.84166551e-01
5.51549196e-01 4.17307317e-02 9.42877293e-01 1.55163741e+00
2.59547919e-01 7.31211782e-01 1.44174135e+00 -5.28422832e-01
9.23240006e-01 4.97094207e-02 -4.04078305e-01 2.36962929e-01
3.09939355e-01 6.57999754e-01 -2.41844714e-01 -2.41967544e-01
8.86927426e-01 -2.38554969e-01 2.03553036e-01 -7.78116405e-01
-1.47127700e+00 8.93319011e-01 6.81579411e-02 2.55357146e-01
-9.63934422e-01 4.36121188e-02 3.20480198e-01 8.49063396e-01
-1.74750797e-02 1.03072047e+00 -2.76704848e-01 -3.54731202e-01
-5.38715959e-01 1.12429237e+00 1.31383741e+00 8.29039454e-01
4.88202542e-01 1.31018877e-01 -6.44698441e-01 4.63865668e-01
7.90760573e-03 3.07021946e-01 8.44082355e-01 -1.31683517e+00
2.76133090e-01 5.32343090e-01 8.98876071e-01 -8.21885049e-01
-3.67266387e-01 7.79111758e-02 -6.67855382e-01 8.79070222e-01
8.14039886e-01 -7.86970079e-01 -4.85872597e-01 1.69195580e+00
3.27309221e-01 -3.37312728e-01 2.81335860e-01 9.56255555e-01
3.25729884e-02 3.76487225e-01 4.57352623e-02 -1.16200961e-01
9.67647612e-01 -6.39753878e-01 -2.74434388e-01 -9.32959393e-02
5.30324042e-01 -3.19931358e-01 1.27950406e+00 6.51070476e-01
-9.65996265e-01 -7.67219424e-01 -1.03971338e+00 4.84199256e-01
9.15932357e-02 -2.89245188e-01 4.44218189e-01 7.41623402e-01
-1.15173006e+00 8.17909598e-01 -4.21126366e-01 -4.41200197e-01
7.00373128e-02 2.53667027e-01 -8.75917450e-02 2.37382814e-01
-1.27939534e+00 8.47435653e-01 7.02990472e-01 -3.62584293e-01
-1.14361131e+00 -1.46836162e-01 -5.76305807e-01 -2.85164207e-01
6.37353361e-01 -8.18774760e-01 1.75238144e+00 -1.31787050e+00
-1.86510587e+00 1.05049407e+00 4.41864014e-01 -7.40306020e-01
1.19644737e+00 1.26950115e-01 3.20608914e-02 -2.65159696e-01
1.87193796e-01 8.44586909e-01 9.70885754e-01 -1.24249625e+00
-5.76269686e-01 -1.34210616e-01 1.18268661e-01 3.85408551e-01
5.63445151e-01 -2.13042676e-01 3.72726887e-01 -7.59916604e-01
-6.41077101e-01 -1.17387462e+00 -5.93767822e-01 -1.73975691e-01
-2.58558929e-01 -2.28221580e-01 -1.18727852e-02 -1.36484355e-01
7.45306373e-01 -1.85951591e+00 3.61313611e-01 2.57634610e-01
3.99861634e-01 3.19062054e-01 -2.65579581e-01 6.64653897e-01
8.72596353e-02 6.85117021e-02 -3.36507678e-01 -2.47803926e-01
3.12540203e-01 9.54887420e-02 -2.98855901e-02 -9.35224257e-03
-6.98824525e-02 1.08073878e+00 -1.33142912e+00 -1.99947983e-01
-4.08146754e-02 -3.83923829e-01 -6.87334776e-01 8.13410521e-01
-4.56690580e-01 8.58025253e-01 -5.71707547e-01 1.85989146e-03
7.37890005e-02 2.51586854e-01 2.70883590e-01 8.44604790e-01
1.48327827e-01 -1.51291013e-01 -1.09358048e+00 1.64822209e+00
-3.20450336e-01 4.26390201e-01 -8.83445963e-02 -8.97602022e-01
1.07790112e+00 3.55589867e-01 3.20720553e-01 -8.65184546e-01
5.07062137e-01 -1.66247599e-02 8.13306749e-01 -3.54015857e-01
3.71218711e-01 -3.75523657e-01 -5.55125177e-01 1.12910604e+00
-8.11562017e-02 -7.81469285e-01 2.97613680e-01 -6.11627065e-02
1.41314363e+00 1.37104839e-01 7.17467904e-01 -2.07941458e-01
3.06025565e-01 2.78505296e-01 3.32585603e-01 1.33051896e+00
-2.89984882e-01 2.95031190e-01 6.63107157e-01 -5.76750994e-01
-1.51352143e+00 -1.09688854e+00 4.58553255e-01 9.91079271e-01
-6.91195205e-02 -6.45824894e-02 -1.17675471e+00 -2.43408829e-01
8.12251568e-02 1.08437896e+00 -8.89823258e-01 -3.84766042e-01
-3.85991365e-01 -3.81233484e-01 6.48274302e-01 -1.70199312e-02
6.28981888e-01 -1.67495453e+00 -1.26281559e+00 4.39016819e-01
1.78166196e-01 -8.61237288e-01 -2.05330655e-01 -9.11464095e-02
-3.94508958e-01 -1.11616457e+00 -7.64077425e-01 -5.01720548e-01
2.85616875e-01 -1.21288054e-01 1.17483687e+00 -2.44746320e-02
-2.00195059e-01 2.42279410e-01 -5.60828626e-01 -6.34674907e-01
-1.17157006e+00 -2.66181856e-01 5.29419720e-01 9.87862051e-02
2.28035688e-01 -6.64457858e-01 -5.32185793e-01 4.93235141e-01
-1.01445544e+00 5.49774766e-01 3.71399879e-01 9.45163846e-01
-6.97281584e-03 -1.72965422e-01 5.11361003e-01 -8.07172954e-01
1.61680984e+00 -4.79844481e-01 -5.07272780e-01 -8.83074850e-02
-4.83789861e-01 3.47731322e-01 1.03221929e+00 -9.50718522e-01
-9.17936444e-01 3.39890607e-02 4.40830827e-01 -2.23853096e-01
-4.65003967e-01 -5.30710109e-02 1.09235279e-01 1.44568101e-01
1.24643147e+00 2.48630613e-01 2.79971570e-01 -1.48698300e-01
5.27574778e-01 7.18725979e-01 4.88009334e-01 -9.47642446e-01
6.30548120e-01 -4.69267294e-02 -3.09355915e-01 -2.94227064e-01
-2.66688410e-02 2.53666610e-01 -6.09690487e-01 -3.10882330e-01
8.25324357e-01 -4.32391435e-01 -1.28225398e+00 9.34173226e-01
-1.07670033e+00 -9.74067211e-01 -7.93899059e-01 2.29033262e-01
-1.35419297e+00 -3.97846662e-02 -4.04548198e-01 -1.03541398e+00
-6.01473749e-02 -8.94172370e-01 8.30527365e-01 2.95700788e-01
-8.41302693e-01 -4.81240660e-01 5.26166916e-01 1.55877739e-01
3.71370971e-01 7.21236348e-01 6.14546537e-01 -9.74087596e-01
-2.04680771e-01 -2.04782277e-01 2.13614255e-01 1.39404461e-01
7.47908503e-02 -1.96487248e-01 -5.85948169e-01 -2.34374017e-01
2.18649715e-01 -5.76341510e-01 -4.54244800e-02 -1.20513076e-02
5.40004671e-01 -2.75568455e-01 1.19224571e-01 4.33702916e-02
8.55823815e-01 8.11138213e-01 4.19745952e-01 1.66577458e-01
1.73732311e-01 8.73070836e-01 9.12946105e-01 7.85273314e-01
2.41962746e-01 7.91610897e-01 -6.48972169e-02 2.97399491e-01
4.16698396e-01 -6.02043033e-01 2.90887743e-01 2.58475751e-01
-3.98480088e-01 -3.89706105e-01 -1.28790486e+00 4.15507674e-01
-2.23333454e+00 -1.02012181e+00 1.94971114e-01 2.37060308e+00
7.18689263e-01 3.21968973e-01 7.00365782e-01 2.89344907e-01
6.87505662e-01 -2.85174400e-01 -6.90617919e-01 -7.87428558e-01
3.07643533e-01 4.98023629e-01 2.57635951e-01 3.58193934e-01
-7.09788799e-01 9.37663913e-01 6.97050333e+00 6.18596971e-01
-7.62591481e-01 -1.23150937e-01 5.81503153e-01 -1.81712031e-01
5.50014824e-02 -2.89583445e-01 7.73131847e-02 6.84684396e-01
1.11077070e+00 -9.92690861e-01 8.05821836e-01 6.62993014e-01
5.84444940e-01 -4.35773015e-01 -1.50481212e+00 9.47102606e-01
-3.67863387e-01 -9.28465724e-01 3.65184247e-02 2.39504337e-01
6.52684093e-01 -4.32171017e-01 -1.31339267e-01 3.92537296e-01
1.32973874e+00 -1.28195608e+00 8.74692857e-01 6.98926568e-01
7.08735406e-01 -6.47576213e-01 7.27187991e-01 1.18285191e+00
-5.43063641e-01 -2.34370023e-01 1.98185816e-02 -8.45853567e-01
-7.61337653e-02 -8.17563012e-02 -1.21130800e+00 1.94020525e-01
1.10039048e-01 1.21686846e-01 -3.69690716e-01 5.65350354e-01
-1.37577221e-01 2.66818136e-01 -1.84080377e-01 -5.37101269e-01
2.04331383e-01 -4.05116588e-01 4.98097777e-01 5.57633281e-01
2.14030951e-01 3.09378773e-01 4.14757542e-02 1.09992468e+00
4.02780056e-01 2.58639269e-03 -1.11282289e+00 -2.48887315e-01
3.18116397e-01 6.05763137e-01 -6.02790654e-01 -3.12144816e-01
2.70539820e-01 1.31094456e+00 3.63176078e-01 2.36490220e-01
-5.38969278e-01 -3.62188756e-01 5.78660965e-01 6.26451746e-02
-1.18364714e-01 -2.64322132e-01 -3.33726555e-01 -1.07997215e+00
1.04148230e-02 -1.35569489e+00 -1.37817875e-01 -8.37113500e-01
-1.16105330e+00 8.41867626e-01 7.14812651e-02 -1.35844445e+00
-1.30427134e+00 -2.63185352e-01 -6.73732460e-01 8.65431666e-01
-5.92787899e-02 -2.38665596e-01 -2.83653527e-01 4.93923604e-01
5.13858974e-01 -6.08026147e-01 7.86954999e-01 -2.73227274e-01
-2.34133974e-01 3.30181181e-01 3.48533913e-02 8.94433409e-02
2.85978556e-01 -1.27050352e+00 1.01485121e+00 3.59319061e-01
-1.77447543e-01 3.72865826e-01 9.50079501e-01 -6.23667657e-01
-7.80668020e-01 -5.93405068e-01 3.14280361e-01 -5.49368203e-01
4.77539062e-01 -8.29314470e-01 -6.24050915e-01 1.51627839e-01
9.39592998e-03 -5.74831009e-01 3.95771891e-01 -4.56624150e-01
2.67431498e-01 3.34883600e-01 -1.59273148e+00 1.03924632e+00
1.38409626e+00 -3.57494593e-01 -1.12150323e+00 7.22873509e-02
5.09307683e-01 -1.49632394e-01 -3.91870558e-01 -1.17885070e-02
6.43208325e-01 -1.06677985e+00 7.75532067e-01 -8.95246267e-01
4.97534990e-01 -1.77553818e-01 1.07465964e-02 -1.87714410e+00
-2.08869129e-01 -1.17777026e+00 2.24495962e-01 6.67385042e-01
5.24044670e-02 -6.35921717e-01 8.93696785e-01 9.33901608e-01
6.50502741e-01 -4.04541552e-01 -8.82953763e-01 -9.49722946e-01
1.68566689e-01 -3.56988579e-01 5.88241816e-01 4.58976388e-01
1.31609946e-01 1.80079520e-01 -4.66928363e-01 -7.04949141e-01
7.03244388e-01 -2.03140840e-01 1.31250644e+00 -1.46588004e+00
-6.61195815e-01 -3.02205771e-01 -4.85993147e-01 -8.68098736e-01
1.54020935e-01 -5.86649001e-01 3.19470674e-01 -9.48289037e-01
-1.71399266e-01 -3.84520978e-01 2.07709149e-01 -8.92711729e-02
-1.27329841e-01 -1.88990444e-01 6.12094045e-01 2.32863709e-01
-5.12506127e-01 7.04191744e-01 1.47208452e+00 2.72342443e-01
-3.02033067e-01 2.69495189e-01 -7.83164501e-01 7.52767920e-01
6.97161674e-01 -5.70279121e-01 -5.60361028e-01 1.48675382e-01
1.32510751e-01 5.45799732e-01 1.17431678e-01 -1.23050630e+00
1.39511839e-01 -4.26124513e-01 3.55500728e-02 3.30516547e-01
-5.38449958e-02 -6.51164651e-01 3.94090354e-01 8.53609860e-01
-5.53134620e-01 2.57429302e-01 -1.37889981e-01 6.63248658e-01
1.27261460e-01 -1.41727924e-01 5.81471682e-01 -2.94995934e-01
-6.59225106e-01 -1.06515281e-01 -1.11584473e+00 3.68197620e-01
1.63680387e+00 -3.47854525e-01 2.42117986e-01 -8.77566636e-01
-9.45914567e-01 4.31856841e-01 8.10245633e-01 5.31449795e-01
2.76521146e-01 -1.10872996e+00 -9.13599074e-01 2.57386088e-01
5.10573313e-02 -1.94959529e-02 -2.71186620e-01 -2.46739294e-02
-6.47033215e-01 2.09512547e-01 -9.06571269e-01 -1.68908074e-01
-8.19471121e-01 6.56213820e-01 4.71857905e-01 -5.23209512e-01
-2.78494656e-01 3.14798206e-01 1.80976257e-01 -8.86402428e-01
-6.78428113e-02 -3.75635862e-01 -1.99400950e-02 -1.60980090e-01
3.30494076e-01 3.52537394e-01 -3.94582361e-01 -1.04803979e-01
5.28080203e-03 -5.50778173e-02 2.62258768e-01 -7.25494504e-01
9.99496102e-01 1.43840536e-01 5.44465125e-01 6.24100268e-01
2.38300487e-01 -4.70992535e-01 -1.53138900e+00 -1.68732926e-01
1.86091125e-01 -4.85962689e-01 -8.53166580e-01 -7.53657639e-01
-1.53915375e-01 5.03163218e-01 4.82221752e-01 5.87171674e-01
8.18079174e-01 -3.69203717e-01 2.97102571e-01 6.99333012e-01
1.34354103e+00 -1.10474646e+00 3.01106721e-01 4.29409355e-01
1.04715633e+00 -1.23068392e+00 -5.90296030e-01 -4.80812974e-02
-1.15016007e+00 7.30170667e-01 6.86103523e-01 -7.65950441e-01
2.02155665e-01 2.37636358e-01 1.08802490e-01 1.87975943e-01
-1.13384521e+00 -2.84163296e-01 -1.94714487e-01 1.26401579e+00
-6.32997304e-02 4.14365441e-01 -5.71309507e-01 6.82403803e-01
-6.31874144e-01 6.31319225e-01 8.23457241e-01 8.69639516e-01
-7.37990737e-02 -1.26679671e+00 -3.16427708e-01 4.47652698e-01
9.03644711e-02 1.96495458e-01 -8.63250017e-01 9.01453376e-01
2.06773683e-01 1.03909361e+00 1.14420176e-01 -6.68944955e-01
4.65294778e-01 1.34732082e-01 4.93373811e-01 -7.65470982e-01
-7.84432650e-01 -7.93842554e-01 7.73046166e-02 -6.89817369e-01
-7.84341246e-02 -8.07963848e-01 -1.00028002e+00 -6.10402942e-01
5.47194064e-01 2.69882560e-01 2.08432972e-01 9.07593668e-01
2.95848697e-01 2.65606612e-01 6.70396209e-01 -1.18716097e+00
-7.63795674e-01 -1.09361875e+00 -7.11468220e-01 9.63492513e-01
6.28940165e-02 -7.41381943e-01 -1.75067633e-01 -2.05501422e-01] | [3.8855087757110596, 1.6143741607666016] |
a6fd04ae-29a8-404d-9df6-501125a69791 | learning-structural-representations-for | 2110.01209 | null | https://arxiv.org/abs/2110.01209v2 | https://arxiv.org/pdf/2110.01209v2.pdf | Learning Structural Representations for Recipe Generation and Food Retrieval | Food is significant to human daily life. In this paper, we are interested in learning structural representations for lengthy recipes, that can benefit the recipe generation and food cross-modal retrieval tasks. Different from the common vision-language data, here the food images contain mixed ingredients and target recipes are lengthy paragraphs, where we do not have annotations on structure information. To address the above limitations, we propose a novel method to unsupervisedly learn the sentence-level tree structures for the cooking recipes. Our approach brings together several novel ideas in a systematic framework: (1) exploiting an unsupervised learning approach to obtain the sentence-level tree structure labels before training; (2) generating trees of target recipes from images with the supervision of tree structure labels learned from (1); and (3) integrating the learned tree structures into the recipe generation and food cross-modal retrieval procedure. Our proposed model can produce good-quality sentence-level tree structures and coherent recipes. We achieve the state-of-the-art recipe generation and food cross-modal retrieval performance on the benchmark Recipe1M dataset. | ['Chunyan Miao', 'Steven C. H. Hoi', 'Guosheng Lin', 'Hao Wang'] | 2021-10-04 | null | null | null | null | ['recipe-generation'] | ['miscellaneous'] | [ 3.37078989e-01 -2.29043469e-01 -1.50797904e-01 -4.07339096e-01
-1.04610443e+00 -8.84889960e-01 4.19083178e-01 6.17869139e-01
-1.67499948e-02 8.64484534e-02 7.23993957e-01 5.00863083e-02
1.49702337e-02 -1.13101327e+00 -9.42176223e-01 -8.01444769e-01
2.12807164e-01 3.18416893e-01 -2.23157331e-01 -3.34521085e-01
9.14347619e-02 -3.16952944e-01 -1.55862343e+00 8.49938750e-01
1.05563557e+00 7.09261417e-01 7.12378561e-01 7.02976525e-01
-5.85478425e-01 1.03824890e+00 -1.81732893e-01 -3.76517385e-01
7.67415315e-02 -5.80739319e-01 -6.73084080e-01 3.98926020e-01
6.69043899e-01 -4.03812766e-01 -1.15327463e-02 1.30733073e+00
1.89482838e-01 5.82559668e-02 1.03604484e+00 -7.47867346e-01
-1.27874494e+00 1.11085582e+00 -5.06784678e-01 -3.71940047e-01
5.47392666e-01 -2.06058864e-02 1.20307970e+00 -8.90545726e-01
5.75887382e-01 1.39038301e+00 3.78048718e-01 5.35209239e-01
-1.15525699e+00 -3.64136845e-01 1.68846071e-01 7.27495551e-02
-1.29157674e+00 -2.76750416e-01 1.01231086e+00 -5.30401468e-01
7.74000466e-01 -1.64964706e-01 5.56656003e-01 9.73555923e-01
1.71778455e-01 1.14326227e+00 8.95818174e-01 -6.76177323e-01
6.14431463e-02 -2.07432464e-01 4.63777453e-01 9.66798604e-01
1.00459106e-01 9.84326676e-02 -2.69148856e-01 8.67438018e-02
6.69782102e-01 1.65278032e-01 -7.57603198e-02 -5.97036719e-01
-1.59624243e+00 1.25027692e+00 6.09815717e-01 3.71460795e-01
-7.33623803e-01 6.38822541e-02 7.69847155e-01 2.33799145e-01
3.69792551e-01 2.50709563e-01 -2.95706928e-01 6.94396913e-01
-8.70192766e-01 2.68210918e-01 6.56461895e-01 1.21677971e+00
6.88588500e-01 -1.51989296e-01 -3.64202112e-01 1.09347832e+00
5.91435432e-01 9.56867039e-01 1.95096582e-01 -7.31102407e-01
6.87750518e-01 4.93480027e-01 1.47757888e-01 -1.13606882e+00
-2.44517595e-01 2.30448529e-01 -9.63105202e-01 -3.70671690e-01
2.51782417e-01 2.19842508e-01 -9.07679796e-01 1.67356145e+00
2.91427284e-01 -4.30160016e-01 4.18982416e-01 8.54231775e-01
1.58608282e+00 1.18954802e+00 5.34857392e-01 -7.51880258e-02
1.65494919e+00 -1.53499579e+00 -9.41152394e-01 1.87212490e-02
6.74151003e-01 -1.22456670e+00 1.16705573e+00 2.66604096e-01
-9.60343599e-01 -8.44765306e-01 -1.10781300e+00 -3.02596509e-01
-7.37133682e-01 5.76398373e-01 6.09615922e-01 2.72484899e-01
-9.78509188e-01 4.68209714e-01 -4.82300371e-01 -5.89599609e-01
-1.69997830e-02 -1.75813496e-01 -1.85260832e-01 -7.54826963e-01
-1.11667693e+00 7.69285560e-01 9.08721447e-01 1.71564654e-01
-1.00120699e+00 -3.98167849e-01 -1.54322755e+00 4.45005335e-02
5.24889529e-01 -1.10717678e+00 1.10981131e+00 -1.04878068e+00
-1.21465111e+00 9.77632344e-01 2.09122434e-01 -1.21970929e-01
-3.88451666e-01 -3.48540187e-01 -3.12473208e-01 4.94327039e-01
4.47979391e-01 1.26110256e+00 7.55231559e-01 -1.28431189e+00
-3.39196742e-01 -3.73754323e-01 2.36319825e-01 4.43290532e-01
9.08807479e-03 2.61310879e-02 -2.73329735e-01 -9.00096297e-01
8.20081308e-02 -7.17653751e-01 -2.59530723e-01 2.86031887e-02
-3.22878778e-01 -4.80779111e-01 1.27859101e-01 -8.99362624e-01
1.06468606e+00 -2.01576996e+00 2.86315054e-01 -3.95057917e-01
-5.99839538e-02 2.41708178e-02 -7.47710109e-01 5.80803931e-01
1.64480641e-01 -1.85456365e-01 -2.32361272e-01 -1.23973131e-01
4.47758019e-01 2.93763071e-01 -2.68834829e-01 -5.05361520e-02
2.11647570e-01 1.08897734e+00 -1.35603511e+00 -8.46622169e-01
5.95434606e-01 3.72900337e-01 -4.06226099e-01 4.00442451e-01
-7.06665933e-01 1.49870232e-01 -6.86038613e-01 7.23687232e-01
7.06859052e-01 -1.93697602e-01 1.85913652e-01 -8.54951859e-01
-1.54042810e-01 4.21997070e-01 -8.95689368e-01 2.32420158e+00
-4.18682963e-01 -1.66687861e-01 3.13549824e-02 -9.36853290e-01
9.24593866e-01 3.21711302e-01 3.14435810e-01 -7.92265654e-01
4.49593663e-02 1.40292477e-02 -6.06819928e-01 -7.01851666e-01
6.84901476e-01 -1.33312672e-01 -6.00638986e-01 4.81363356e-01
2.74354100e-01 -3.68131012e-01 8.06376576e-01 1.45989597e-01
4.78857189e-01 5.83335936e-01 4.41769689e-01 -4.46600229e-01
4.80393261e-01 4.55371797e-01 1.04953796e-01 5.31408191e-01
-3.36384326e-02 4.76160228e-01 -2.23193809e-01 -5.65520048e-01
-1.28664589e+00 -1.25771558e+00 3.78625631e-01 1.31584692e+00
2.63396204e-02 -6.92189336e-01 -8.43331337e-01 -6.97359979e-01
-3.85737389e-01 7.29997218e-01 -4.57022727e-01 5.32510411e-03
-6.32838845e-01 -5.41674256e-01 2.98283547e-01 3.72797608e-01
7.23622501e-01 -1.41188812e+00 -3.76193464e-01 3.97590667e-01
-8.82739067e-01 -9.19561923e-01 -1.18213451e+00 1.28779616e-02
-7.05123425e-01 -1.20748878e+00 -8.79973352e-01 -1.49757826e+00
5.34794092e-01 6.90814912e-01 1.87070465e+00 -1.24371573e-01
-2.22581133e-01 6.89207315e-01 -8.29518437e-01 -1.20806888e-01
-8.65762830e-01 -1.08701162e-01 -4.61197466e-01 -3.04412514e-01
6.20876789e-01 -9.09600928e-02 -7.33363152e-01 -1.58690110e-01
-1.21258235e+00 5.38856328e-01 5.38804233e-01 8.45643044e-01
9.29385364e-01 1.98897943e-01 4.44509655e-01 -6.29592538e-01
4.40927863e-01 -3.88550401e-01 -4.49692339e-01 7.67396092e-01
-2.58898407e-01 2.63868421e-01 6.60075366e-01 -4.52635258e-01
-1.07162607e+00 3.71072203e-01 -1.96412101e-01 -2.33298689e-01
-3.38972658e-01 6.03331447e-01 -9.46153700e-02 6.37884438e-01
4.60362375e-01 3.95575106e-01 -1.40528098e-01 -6.78387105e-01
1.22272539e+00 4.46341872e-01 4.37739819e-01 -8.98793340e-01
2.92542934e-01 5.25161289e-02 -2.53092468e-01 -6.41313434e-01
-1.16033375e+00 -5.49647927e-01 -7.30284333e-01 -5.67964129e-02
1.41360617e+00 -1.19297647e+00 -3.65261823e-01 3.80175829e-01
-1.31626916e+00 -2.24268883e-01 -1.07657455e-01 3.72555763e-01
-6.69102311e-01 5.53256333e-01 -9.28945720e-01 -4.01287019e-01
-9.31104541e-01 -9.89032745e-01 1.47673678e+00 1.23109445e-01
4.93727773e-02 -9.02104795e-01 2.83952385e-01 4.77698416e-01
8.43064263e-02 3.66828829e-01 1.35986340e+00 -1.37783274e-01
-6.17623985e-01 4.46976960e-01 -3.54008883e-01 2.83814967e-01
4.91890341e-01 -1.58879772e-01 -5.37275732e-01 -2.27162659e-01
5.80267459e-02 -7.93292820e-01 1.00764179e+00 5.68811119e-01
5.37387013e-01 -3.76016885e-01 -3.36218551e-02 2.32886393e-02
1.58491075e+00 6.57309517e-02 4.09871131e-01 -3.58931459e-02
8.16353500e-01 1.02651954e+00 7.71650314e-01 5.25851361e-02
8.31846118e-01 6.65629685e-01 1.69487119e-01 -2.62085766e-01
-5.39705575e-01 -7.55104721e-01 7.43188679e-01 1.49926662e+00
6.53895855e-01 -6.67502657e-02 -3.57117355e-01 9.04641628e-01
-1.89459026e+00 -1.00943089e+00 -1.83036402e-01 1.85157025e+00
1.03147066e+00 -5.59712410e-01 1.01773284e-01 -2.43948221e-01
7.31897056e-01 3.33784074e-01 -2.90404201e-01 -2.80159473e-01
-9.12854970e-02 -3.72305396e-03 1.76956505e-01 2.83960938e-01
-1.56828284e+00 1.01514161e+00 6.07067680e+00 7.81258583e-01
-5.14939129e-01 3.58478837e-02 3.24063599e-01 2.62810022e-01
-4.98114973e-01 -2.93917805e-01 -6.06028736e-01 1.54648557e-01
6.65303528e-01 2.84932464e-01 7.30889440e-01 6.71413481e-01
-6.66121319e-02 1.30433114e-02 -1.34517217e+00 1.08351433e+00
4.56393182e-01 -1.33062732e+00 4.38538045e-01 -4.45595831e-01
7.91836143e-01 -1.26125276e-01 -1.96015656e-01 5.38501382e-01
7.91006148e-01 -7.44022965e-01 9.23141062e-01 4.47454721e-01
5.66651285e-01 -5.61432302e-01 3.27866793e-01 3.08371603e-01
-1.73936117e+00 1.38102248e-01 -6.36602938e-01 3.89503032e-01
2.41633356e-01 6.82878494e-01 -3.16476703e-01 9.73402202e-01
6.79524779e-01 1.12930751e+00 -5.27694643e-01 6.47227466e-01
-3.85578036e-01 2.40471184e-01 -2.78416961e-01 -1.66052952e-01
5.06143808e-01 -5.36682785e-01 5.49884625e-02 1.31093895e+00
3.47800761e-01 4.34631445e-02 6.03485107e-01 1.31314969e+00
1.79597259e-01 5.62919557e-01 -9.75338399e-01 -5.25553226e-01
-2.07109656e-02 1.29128492e+00 -6.94079578e-01 -4.76744354e-01
-6.22215688e-01 1.03982663e+00 1.75075516e-01 2.36811861e-01
-5.58807015e-01 1.07421139e-02 1.74370781e-01 -6.21258855e-01
6.34998679e-01 -1.23923004e-01 1.14995316e-01 -1.36345768e+00
-7.71868378e-02 -1.09456217e+00 7.10950494e-01 -1.01755941e+00
-1.85794306e+00 4.36127722e-01 9.34798196e-02 -1.09539425e+00
-2.35807925e-01 -5.11668026e-01 -1.18086331e-01 5.42748272e-01
-1.78223825e+00 -1.69997859e+00 8.43303949e-02 6.07926667e-01
1.00952780e+00 1.33570805e-01 1.36089945e+00 4.82516140e-01
-1.45731032e-01 5.88174947e-02 1.40557572e-01 2.33942628e-01
6.39932930e-01 -1.22756994e+00 2.09093153e-01 5.29616654e-01
5.69864452e-01 6.82081044e-01 4.55413699e-01 -7.01099634e-01
-1.52298510e+00 -1.33337545e+00 1.17144799e+00 -2.98100621e-01
4.59638596e-01 -4.36212391e-01 -5.46121359e-01 3.13512415e-01
7.36525238e-01 -5.05253971e-01 6.84451818e-01 2.09571198e-02
-6.71239078e-01 -5.59147745e-02 -1.11127150e+00 6.44172490e-01
7.99896359e-01 -6.46219075e-01 -9.86227930e-01 5.74942589e-01
1.03297353e+00 -1.56395718e-01 -9.70165491e-01 3.04492444e-01
4.95601147e-01 -5.55632710e-01 1.22273183e+00 -3.07015985e-01
7.19166338e-01 -4.93752420e-01 -5.32143533e-01 -1.18259239e+00
-3.95084441e-01 -3.61133926e-02 1.48445413e-01 1.38323045e+00
1.57524109e-01 2.24914446e-01 1.95751473e-01 8.39959010e-02
-2.39491209e-01 -3.90200585e-01 -2.41841748e-01 -3.67004126e-01
2.86383957e-01 7.60651976e-02 6.83686256e-01 7.27749109e-01
-1.58420503e-01 8.12813580e-01 -5.05259097e-01 3.88585292e-02
7.28718579e-01 8.63836050e-01 4.92006451e-01 -7.83324182e-01
-2.58077800e-01 -3.32973927e-01 4.40537840e-01 -1.42289269e+00
1.43052533e-01 -1.14054477e+00 5.92024803e-01 -1.96687913e+00
7.45016038e-01 -1.12716518e-01 -2.91821182e-01 5.59625030e-01
-1.80234283e-01 2.84580849e-02 4.39951926e-01 -1.54359937e-02
-9.65210557e-01 6.67835295e-01 1.56921339e+00 -8.03817689e-01
8.67677554e-02 -5.25520504e-01 -6.90006018e-01 4.68993366e-01
5.64777374e-01 -5.58412969e-01 -7.37362742e-01 -7.15000629e-01
2.44227320e-01 1.40456498e-01 3.60630631e-01 -4.90255892e-01
-5.56613393e-02 -1.60988927e-01 1.22520082e-01 -1.08078587e+00
5.25692515e-02 -7.99572885e-01 1.94764182e-01 4.26475137e-01
-4.55997795e-01 5.36064431e-02 7.12698475e-02 4.18789208e-01
-2.29835764e-01 -4.74056870e-01 4.97878075e-01 -7.22950041e-01
-7.65763342e-01 2.58936174e-02 -3.89388174e-01 -2.19052956e-01
4.69003022e-01 3.63937378e-01 -3.97998899e-01 -6.50539473e-02
-7.12613285e-01 3.57020825e-01 4.13724869e-01 7.34465361e-01
5.53381205e-01 -1.82314384e+00 -8.94940495e-01 -4.49813297e-03
5.04773378e-01 -1.68782309e-01 4.29952085e-01 1.68741316e-01
-4.20473933e-01 6.67658389e-01 -9.58387032e-02 -6.36523366e-01
-1.13497663e+00 1.31349134e+00 -1.62035953e-02 -7.83588529e-01
-6.06709421e-01 4.92207587e-01 7.06457257e-01 -8.26767325e-01
1.24819525e-01 -5.69446146e-01 -5.99361658e-01 7.10525662e-02
6.72152340e-01 -2.75082648e-01 -3.19925666e-01 -7.96304524e-01
4.88512293e-02 9.82470810e-01 -1.74243808e-01 3.41617644e-01
1.26728296e+00 -3.94522280e-01 -2.79873788e-01 5.19802034e-01
1.16049016e+00 -6.28201067e-01 -9.78415251e-01 -3.51181030e-01
-3.81221063e-02 -2.64362060e-02 -9.00922418e-02 -9.77194846e-01
-1.03609025e+00 8.97768319e-01 7.75111616e-01 -6.04185835e-03
1.27774704e+00 1.79040000e-01 1.21943605e+00 4.80309576e-01
6.35047078e-01 -9.70141470e-01 3.21969241e-01 3.94109845e-01
9.65825260e-01 -1.33026254e+00 -1.47707343e-01 -5.94678462e-01
-5.21177948e-01 1.03603423e+00 1.46804109e-01 -1.88198000e-01
4.28777516e-01 -1.13002501e-01 1.95088103e-01 -5.95221400e-01
-5.91055512e-01 -4.07695442e-01 4.34459180e-01 7.18832672e-01
6.00290835e-01 3.69070232e-01 -4.05254036e-01 6.76143050e-01
2.15521857e-01 9.11005810e-02 -4.13422473e-02 7.32347786e-01
-5.29313624e-01 -1.38832915e+00 -2.85419941e-01 2.43773051e-02
-1.11318663e-01 -6.07177079e-01 -2.02341348e-01 3.04466069e-01
3.75472844e-01 1.32023120e+00 -2.72285551e-01 -1.56906638e-02
3.48703951e-01 2.72960693e-01 9.43669736e-01 -8.44868958e-01
-6.77729964e-01 4.72617865e-01 1.91433176e-01 -4.21831042e-01
-1.02361822e+00 -4.34870094e-01 -9.71205711e-01 -6.93153590e-02
-2.12324381e-01 1.57618031e-01 4.79674548e-01 8.76841605e-01
1.96404666e-01 4.27684039e-01 5.69858313e-01 -8.18309546e-01
-5.64573348e-01 -9.79663193e-01 -4.72477257e-01 8.31858337e-01
3.30254287e-02 -4.25730050e-01 1.92375422e-01 7.58531749e-01] | [11.516178131103516, 4.4272003173828125] |
52927812-0666-47c9-9adf-2e264a835f80 | talking-detection-in-collaborative-learning | 2110.07646 | null | https://arxiv.org/abs/2110.07646v1 | https://arxiv.org/pdf/2110.07646v1.pdf | Talking Detection In Collaborative Learning Environments | We study the problem of detecting talking activities in collaborative learning videos. Our approach uses head detection and projections of the log-magnitude of optical flow vectors to reduce the problem to a simple classification of small projection images without the need for training complex, 3-D activity classification systems. The small projection images are then easily classified using a simple majority vote of standard classifiers. For talking detection, our proposed approach is shown to significantly outperform single activity systems. We have an overall accuracy of 59% compared to 42% for Temporal Segment Network (TSN) and 45% for Convolutional 3D (C3D). In addition, our method is able to detect multiple talking instances from multiple speakers, while also detecting the speakers themselves. | ['Carlos LópezLeiva', 'Sylvia Celedón-Pattichis', 'Marios S. Pattichis', 'Wenjing Shi'] | 2021-10-14 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [ 1.80320397e-01 3.79886091e-01 -4.83805239e-02 -1.71317965e-01
-7.22040117e-01 -6.38607085e-01 7.75482178e-01 -3.73822033e-01
-2.45895907e-01 2.71038115e-01 4.53853339e-01 -1.26778170e-01
2.88145781e-01 -1.75454646e-01 -3.84205848e-01 -6.75465822e-01
-2.59571970e-01 2.88460255e-01 3.73498231e-01 4.60600019e-01
1.23930268e-01 7.70337880e-01 -1.58571279e+00 3.16105187e-01
1.32729232e-01 1.34878016e+00 -2.37473533e-01 1.16512775e+00
-1.75088961e-02 1.27840543e+00 -1.00179303e+00 6.63577951e-03
1.82297125e-01 -5.66641510e-01 -6.51861906e-01 6.26309097e-01
1.06383586e+00 -3.15083861e-01 -4.72858787e-01 6.09266222e-01
5.72044253e-01 1.86463580e-01 5.88324547e-01 -1.35824370e+00
8.90875682e-02 1.57950982e-01 -4.06730860e-01 5.43560207e-01
1.06313133e+00 6.58372194e-02 5.41304946e-01 -1.18233454e+00
5.49010277e-01 1.35444009e+00 5.86398184e-01 6.05716825e-01
-9.42714453e-01 -7.05664635e-01 1.21457642e-02 2.15855122e-01
-1.17091250e+00 -8.59650910e-01 6.32529080e-01 -5.56484818e-01
1.22699785e+00 2.61138707e-01 1.14604938e+00 1.29150546e+00
-2.09106430e-01 1.06755471e+00 1.05693901e+00 -1.58705115e-01
2.62578875e-01 2.04735741e-01 9.01532993e-02 1.04507005e+00
-8.51087198e-02 -2.63709694e-01 -1.10773027e+00 -1.13234306e-02
6.69155598e-01 -1.34420127e-01 -4.69779223e-01 -3.48678708e-01
-1.24945533e+00 5.20575404e-01 1.26521543e-01 2.73028255e-01
-2.26220593e-01 6.62807450e-02 3.23345274e-01 2.01045424e-02
5.52128315e-01 -5.25206663e-02 -2.85566505e-02 -5.04315853e-01
-1.10287917e+00 -1.21012568e-01 1.18406987e+00 6.95464313e-01
2.10348010e-01 1.15444392e-01 -1.21625260e-01 6.09438598e-01
3.13920885e-01 4.45943326e-01 5.24458230e-01 -1.57407331e+00
5.08433640e-01 4.54343289e-01 -1.26023591e-01 -5.78555703e-01
-5.24956763e-01 -2.21333697e-01 -4.51363236e-01 3.50593090e-01
6.70915186e-01 -3.68060321e-01 -6.19846523e-01 1.42236149e+00
3.96547705e-01 2.03435421e-01 6.19932078e-02 6.97372198e-01
6.19554520e-01 5.59278190e-01 -3.36643100e-01 -4.67735559e-01
1.09465551e+00 -1.05272436e+00 -6.94331110e-01 -2.92944282e-01
5.91069698e-01 -5.35083771e-01 6.98576629e-01 5.57705164e-01
-1.32185864e+00 -4.56479073e-01 -9.52009439e-01 2.08022315e-02
-2.20689222e-01 1.22465424e-01 5.48833191e-01 1.02041388e+00
-1.26707244e+00 3.35999459e-01 -8.31419408e-01 -6.38724983e-01
7.11515546e-01 5.20607948e-01 -5.73392630e-01 1.12099610e-01
-4.87693161e-01 9.03109848e-01 -2.72165895e-01 1.38063803e-01
-9.95881915e-01 -4.16582704e-01 -8.85786057e-01 -4.53988053e-02
7.15052104e-03 -3.57386023e-01 1.30078328e+00 -9.33445334e-01
-1.79250383e+00 1.02629507e+00 -5.44515312e-01 -5.07658005e-01
1.02383840e+00 -3.47964466e-01 -1.47096008e-01 6.70324206e-01
-1.67429641e-01 6.48582697e-01 7.14811862e-01 -6.06681585e-01
-5.46705186e-01 -3.62249970e-01 -6.52188063e-02 5.20746887e-01
-3.32558423e-01 1.82110563e-01 -4.02201653e-01 -2.19247788e-01
1.64209947e-01 -1.08266973e+00 2.82044023e-01 4.82518643e-01
-4.57436293e-01 -3.54690015e-01 1.34664595e+00 -5.95759690e-01
5.91783643e-01 -2.04798365e+00 -4.78546545e-02 1.18602347e-02
4.61993188e-01 2.20956936e-01 2.27956846e-01 -1.28161684e-01
1.28070369e-01 -3.44916344e-01 1.00168191e-01 -6.46006823e-01
-3.37177664e-01 -2.69923303e-02 8.70057344e-02 7.67616808e-01
-1.96474828e-02 5.97996056e-01 -7.79926300e-01 -7.05548048e-01
6.13433063e-01 7.74655461e-01 -4.42184985e-01 1.52346835e-01
3.35320324e-01 4.75324124e-01 -5.21141384e-03 6.90013230e-01
3.00686300e-01 -2.31130540e-01 3.67924809e-01 5.75290062e-02
1.05614252e-01 5.65458417e-01 -1.23636127e+00 1.39879978e+00
-6.84762418e-01 1.51433098e+00 1.02518246e-01 -1.01547074e+00
7.44846284e-01 6.29148841e-01 7.62319088e-01 -2.26883352e-01
8.33510011e-02 -5.90507835e-02 2.42269542e-02 -5.56103051e-01
-1.99725419e-01 2.80690610e-01 1.73040628e-02 5.95952034e-01
3.96364659e-01 1.64477319e-01 1.18610330e-01 2.87916303e-01
1.29229689e+00 -4.23532305e-03 1.64362013e-01 1.53795481e-02
6.16364777e-01 -4.09815133e-01 3.04585665e-01 6.75761998e-01
-7.41926849e-01 4.12931502e-01 7.09981501e-01 -2.89799571e-01
-5.12954056e-01 -1.18686390e+00 6.83446750e-02 1.06662023e+00
-9.32238530e-03 -2.98497200e-01 -9.71297801e-01 -9.14464653e-01
-1.70724288e-01 1.50066763e-01 -6.25091314e-01 2.00482100e-01
-7.82073438e-01 -4.60334778e-01 7.31986344e-01 8.81441712e-01
6.86581254e-01 -8.29987645e-01 -7.21495688e-01 -2.10903659e-02
-3.50198328e-01 -1.58285463e+00 -5.91863632e-01 1.62081003e-01
-1.01776755e+00 -1.25673103e+00 -8.32791209e-01 -7.74937868e-01
5.99376559e-01 3.67738634e-01 7.24744320e-01 -3.89853954e-01
-3.00744057e-01 7.27145553e-01 1.30359918e-01 -4.18953180e-01
-2.61753172e-01 -3.57298911e-01 3.26310247e-01 2.63435930e-01
5.29692948e-01 -5.63807368e-01 -6.49577498e-01 5.99131644e-01
3.08504105e-02 -2.60607690e-01 3.26700151e-01 2.99181640e-01
5.85383177e-02 -6.87981725e-01 2.25046668e-02 -3.26415926e-01
5.15759766e-01 2.80016735e-02 -3.28566045e-01 5.24050184e-02
-2.54207969e-01 -3.10655922e-01 1.00552715e-01 -9.06821489e-01
-1.07313299e+00 6.21796310e-01 1.23245917e-01 -7.60600030e-01
-3.84244710e-01 -4.94896770e-01 4.55773287e-02 -3.14451247e-01
6.92903996e-01 1.63815096e-01 1.42642945e-01 -2.86258221e-01
2.54213154e-01 6.75968528e-01 5.55414259e-01 1.56864345e-01
2.57845998e-01 7.25349963e-01 1.26081839e-01 -1.35622978e+00
-8.74968827e-01 -7.89853156e-01 -9.73795235e-01 -9.05838013e-01
1.03676116e+00 -1.09996498e+00 -9.30533648e-01 6.12567425e-01
-1.09106672e+00 -2.54392803e-01 -2.03956991e-01 7.72773504e-01
-6.02466345e-01 4.10719305e-01 -8.54238868e-01 -1.17917228e+00
-2.47219458e-01 -1.05113804e+00 1.19105220e+00 8.42308775e-02
-6.15802467e-01 -9.61346745e-01 2.99509428e-02 8.24741244e-01
2.41581485e-01 1.98989749e-01 1.07892506e-01 -6.17687464e-01
-5.10430813e-01 -4.82953310e-01 1.43253490e-01 3.68167460e-01
-2.02293862e-02 9.03622806e-02 -1.60682237e+00 4.89046685e-02
-8.04276243e-02 -2.93201953e-01 7.19156265e-01 6.95618212e-01
9.71387446e-01 -2.40703821e-01 -5.17086387e-01 3.78343821e-01
5.33161819e-01 2.02468082e-01 4.23438191e-01 -2.34740064e-01
5.88862717e-01 7.54345298e-01 1.61170900e-01 3.06140482e-01
7.21591413e-02 7.91618347e-01 1.91805184e-01 1.01025021e-02
-5.51500678e-01 -8.94262716e-02 8.91793251e-01 4.94298130e-01
-4.59420919e-01 -7.44115710e-02 -8.29114079e-01 2.10444674e-01
-1.64362562e+00 -1.25165808e+00 -1.18475981e-01 2.00903058e+00
1.08175904e-01 3.45995933e-01 5.86219251e-01 5.80111802e-01
7.47938633e-01 2.21631169e-01 -3.68967801e-01 -2.67598093e-01
1.91678125e-02 4.41618860e-02 2.84942955e-01 5.52837610e-01
-1.23104107e+00 5.09261429e-01 7.67836666e+00 1.62692785e-01
-1.06992996e+00 3.11962336e-01 3.93146008e-01 -5.83107114e-01
7.42019117e-01 -3.52709472e-01 -8.93137574e-01 4.33252513e-01
1.01663077e+00 1.62053019e-01 1.17338598e-01 9.51127768e-01
4.68404591e-01 -3.71858686e-01 -1.20699012e+00 1.32016706e+00
6.15397930e-01 -1.13769889e+00 -3.94883931e-01 3.87833953e-01
4.78835166e-01 1.17157795e-01 -1.37946531e-01 9.11331177e-02
8.48124623e-02 -7.30270147e-01 6.50873542e-01 4.11357999e-01
4.40565348e-01 -3.01711261e-01 5.09301603e-01 3.81341338e-01
-1.18740571e+00 -3.14895272e-01 1.03418581e-01 -7.92299286e-02
3.14047605e-01 4.15452838e-01 -1.10725749e+00 -4.42282230e-01
9.10907269e-01 6.94608271e-01 -2.59829819e-01 1.03952229e+00
-2.67365903e-01 5.76641321e-01 -6.39073133e-01 -1.91131428e-01
8.97594988e-02 1.15402430e-01 5.15852153e-01 1.42455161e+00
2.48593271e-01 -2.46165678e-01 5.40945008e-02 2.32709989e-01
-3.09939861e-01 -1.67494625e-01 -8.20079982e-01 2.88863599e-01
1.87658653e-01 1.15923774e+00 -6.88405335e-01 -6.60014331e-01
-4.40057009e-01 1.11141503e+00 1.01484597e-01 1.17745362e-01
-5.48341751e-01 -1.84852108e-01 6.78918719e-01 1.95523262e-01
2.75099635e-01 -2.96451807e-01 2.98152622e-02 -1.21938038e+00
1.64478123e-01 -3.22186828e-01 3.06915373e-01 -8.56895864e-01
-6.20217502e-01 2.90622443e-01 -1.10246241e-01 -1.31155264e+00
-4.91071105e-01 -7.11357772e-01 -8.71172845e-01 3.56921464e-01
-9.21942711e-01 -8.13382268e-01 -7.08219111e-01 9.53228474e-01
8.28846812e-01 -8.43866989e-02 6.44176304e-01 1.85818821e-01
-6.09079540e-01 3.70426178e-01 -3.16114128e-01 4.71166521e-01
5.42104363e-01 -1.25510669e+00 2.06824467e-01 8.04286361e-01
4.30353820e-01 1.63468763e-01 4.89377379e-01 -1.32926673e-01
-1.08859968e+00 -7.59613335e-01 1.07416332e+00 -7.72544861e-01
3.95978719e-01 -8.23549509e-01 -4.27079618e-01 7.91069508e-01
1.37761384e-01 3.81171107e-01 8.35302711e-01 -1.13187935e-02
-4.46667135e-01 -1.73131555e-01 -1.14077294e+00 1.04129389e-01
1.34705615e+00 -7.10140884e-01 -5.52078784e-01 7.23116517e-01
-7.19979927e-02 -3.78261745e-01 -7.12041855e-01 1.56659372e-02
8.95127833e-01 -1.14283109e+00 8.97023559e-01 -2.48883724e-01
-3.63282450e-02 1.39999911e-02 3.24117318e-02 -8.39834988e-01
2.23336682e-01 -6.94535494e-01 -8.18161011e-01 6.33714318e-01
3.18067282e-01 -2.91484296e-01 1.29562128e+00 3.00719887e-01
8.47485065e-02 -2.58502960e-01 -1.20597732e+00 -5.97038209e-01
-4.79954720e-01 -6.10849380e-01 -3.66284251e-01 7.05857813e-01
4.33677793e-01 5.41469991e-01 -2.94001460e-01 -7.67616928e-03
7.41310358e-01 -1.02675200e-01 8.17285180e-01 -1.26278818e+00
-1.68730900e-01 -4.71272618e-01 -8.87812495e-01 -1.33111846e+00
8.99844021e-02 -6.21342063e-01 -1.73875317e-01 -1.37121749e+00
1.81094125e-01 3.07534814e-01 -3.82051244e-03 4.34052259e-01
4.62200850e-01 5.79184830e-01 1.75647095e-01 3.34469110e-01
-7.51372159e-01 3.41878049e-02 1.04519403e+00 -1.47984982e-01
-3.10345829e-01 4.09689188e-01 -9.80413407e-02 9.40113008e-01
6.36103570e-01 -1.92528680e-01 -3.29792291e-01 5.07536121e-02
-4.57640439e-01 2.63735145e-01 5.49502254e-01 -1.46808326e+00
3.08367968e-01 3.59512061e-01 8.25687468e-01 -7.25525260e-01
8.98637772e-01 -4.44963485e-01 -4.37758654e-01 6.45639062e-01
-4.22078073e-01 -7.06636608e-02 -9.27730873e-02 5.43536246e-01
-9.15641785e-02 2.96164453e-01 8.38325620e-01 -2.86324263e-01
-5.45671642e-01 -5.12814112e-02 -1.09834480e+00 -2.56594867e-01
1.19530988e+00 -5.32161832e-01 -1.44873887e-01 -7.89547443e-01
-1.23168421e+00 -5.63604534e-02 -5.58186620e-02 4.42743897e-01
4.62288380e-01 -8.91752183e-01 -1.61454737e-01 8.74063373e-02
-6.77865744e-02 -3.25309068e-01 4.26534265e-02 1.12501657e+00
-5.12495160e-01 5.55336475e-01 -1.59496531e-01 -9.95831192e-01
-1.78284240e+00 1.30960941e-01 8.51529241e-01 2.02498958e-01
-6.97647035e-01 9.69338417e-01 4.29946743e-03 -1.64990053e-01
8.63454700e-01 -4.24159288e-01 -3.19285661e-01 4.00140792e-01
6.98054969e-01 6.94869220e-01 8.54363516e-02 -7.35196948e-01
-6.03774905e-01 5.86436450e-01 2.03862667e-01 -3.23323876e-01
9.31903243e-01 -8.27385485e-02 5.53053021e-01 3.16409916e-01
1.37959266e+00 -1.18945152e-01 -1.54088461e+00 -7.74775371e-02
-1.76448166e-01 -2.82799393e-01 1.83537960e-01 -5.00680685e-01
-1.16936135e+00 1.36281157e+00 1.07460177e+00 1.21309251e-01
8.54789317e-01 4.13254946e-01 3.81052285e-01 6.54087543e-01
-1.15133420e-01 -1.10046411e+00 5.50532520e-01 1.85692832e-01
5.02756000e-01 -1.12152493e+00 -1.73258200e-01 -4.89613891e-01
-4.84111100e-01 1.31387079e+00 4.71810013e-01 1.32402971e-01
6.86941087e-01 4.80455071e-01 2.76712328e-01 -2.09752351e-01
-7.56094337e-01 -2.76269704e-01 2.70264223e-02 8.53768051e-01
3.66029292e-01 -1.66974425e-01 3.35990846e-01 -3.50995481e-01
-7.97831118e-02 -2.82300655e-02 5.19017279e-01 1.03103983e+00
-5.29766023e-01 -3.96379828e-01 -4.66510832e-01 4.04280871e-01
-4.83282954e-01 3.99096161e-01 -6.57552421e-01 6.93903029e-01
-3.79950702e-02 1.05763793e+00 5.30033946e-01 -3.15934092e-01
4.09368336e-01 4.46549982e-01 6.17410123e-01 -3.39733332e-01
-3.82882893e-01 2.41590738e-01 3.64191711e-01 -9.15405691e-01
-9.95189726e-01 -1.12684655e+00 -7.03014255e-01 -1.51331782e-01
-1.68068722e-01 -9.27331448e-02 8.27319384e-01 9.57054436e-01
1.78350240e-01 1.87687606e-01 7.70369649e-01 -1.19758129e+00
-9.74644870e-02 -1.11470795e+00 -4.62423652e-01 1.51146531e-01
4.72400934e-01 -7.55428553e-01 -7.82915473e-01 2.34934554e-01] | [14.42126750946045, 5.0834574699401855] |
df3ecdda-cc28-427b-9331-cc332ca82430 | an-investigation-into-pre-training-object | 2302.04419 | null | https://arxiv.org/abs/2302.04419v3 | https://arxiv.org/pdf/2302.04419v3.pdf | An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning | Unsupervised object-centric representation (OCR) learning has recently drawn attention as a new paradigm of visual representation. This is because of its potential of being an effective pre-training technique for various downstream tasks in terms of sample efficiency, systematic generalization, and reasoning. Although image-based reinforcement learning (RL) is one of the most important and thus frequently mentioned such downstream tasks, the benefit in RL has surprisingly not been investigated systematically thus far. Instead, most of the evaluations have focused on rather indirect metrics such as segmentation quality and object property prediction accuracy. In this paper, we investigate the effectiveness of OCR pre-training for image-based reinforcement learning via empirical experiments. For systematic evaluation, we introduce a simple object-centric visual RL benchmark and conduct experiments to answer questions such as ``Does OCR pre-training improve performance on object-centric tasks?'' and ``Can OCR pre-training help with out-of-distribution generalization?''. Our results provide empirical evidence for valuable insights into the effectiveness of OCR pre-training for RL and the potential limitations of its use in certain scenarios. Additionally, this study also examines the critical aspects of incorporating OCR pre-training in RL, including performance in a visually complex environment and the appropriate pooling layer to aggregate the object representations. | ['Sungjin Ahn', 'Heechul Bae', 'Yi-Fu Wu', 'Jaesik Yoon'] | 2023-02-09 | null | null | null | null | ['optical-character-recognition', 'systematic-generalization'] | ['computer-vision', 'reasoning'] | [ 3.20743799e-01 -1.00718774e-01 -3.64805609e-01 -2.09797636e-01
-8.45429957e-01 -4.39990252e-01 5.38125753e-01 3.71627241e-01
-7.93967128e-01 5.28193235e-01 5.77585436e-02 -3.98797810e-01
-1.54854685e-01 -5.99384785e-01 -8.61550212e-01 -6.80530131e-01
-5.95904365e-02 1.69472173e-01 3.16700667e-01 -2.04797342e-01
4.99962181e-01 9.03489470e-01 -1.78029370e+00 4.35962975e-01
8.01944137e-01 9.56979930e-01 4.82375473e-01 6.17139041e-01
3.69477235e-02 8.04121256e-01 -8.85074437e-01 -2.97990561e-01
1.69423133e-01 -4.11641538e-01 -6.50784910e-01 3.00593823e-01
4.64293540e-01 -1.92737564e-01 -2.31982008e-01 8.23516488e-01
6.41575396e-01 3.50419670e-01 8.27792585e-01 -1.13426733e+00
-9.37090218e-01 2.34189272e-01 -7.50916541e-01 5.24289608e-01
8.77258927e-02 7.72643924e-01 1.11906004e+00 -6.02382004e-01
6.35022104e-01 1.14640808e+00 4.80138987e-01 5.17502546e-01
-1.30765951e+00 -4.12193775e-01 1.82289153e-01 2.45066777e-01
-1.07250738e+00 -2.07952619e-01 7.52449155e-01 -4.16000098e-01
1.04410040e+00 1.29722774e-01 5.33968449e-01 9.55076933e-01
9.39770415e-02 1.28591406e+00 1.60967326e+00 -3.79591167e-01
1.95675284e-01 3.31555396e-01 -4.14730087e-02 4.59198803e-01
4.23505872e-01 3.16048950e-01 -5.39840817e-01 3.83202881e-01
9.79601681e-01 -2.76256830e-01 -1.18615329e-01 -1.87621787e-01
-9.03161466e-01 8.39995682e-01 9.51171339e-01 2.95634627e-01
-9.66168121e-02 6.01473927e-01 3.67557168e-01 1.52147666e-01
4.99893486e-01 8.70598495e-01 -3.73477608e-01 6.66382685e-02
-9.62610006e-01 1.47539407e-01 1.55637544e-02 9.18576002e-01
7.98387825e-01 3.42735916e-01 -4.55433577e-01 1.02627766e+00
3.86651546e-01 3.46009403e-01 4.77051914e-01 -8.75718653e-01
3.88524920e-01 6.04339004e-01 -1.17458217e-01 -6.85287952e-01
-5.09327888e-01 -5.75623214e-01 -6.18082106e-01 6.03111625e-01
6.07044101e-01 -7.36284554e-02 -1.19978833e+00 1.58585703e+00
-1.56700715e-01 -3.41217481e-02 4.91892695e-02 1.08692014e+00
1.07503092e+00 4.46345866e-01 5.23715079e-01 9.67013016e-02
1.28415954e+00 -8.85865808e-01 -2.56745458e-01 -4.68117565e-01
6.34134233e-01 -7.29723454e-01 1.42898524e+00 2.60321081e-01
-9.74724829e-01 -7.99819231e-01 -1.08514488e+00 -2.11665407e-01
-7.17516541e-01 3.23538095e-01 8.53764296e-01 7.71892548e-01
-1.09210074e+00 6.36104822e-01 -3.54291707e-01 -3.69447291e-01
8.91919971e-01 2.59244561e-01 -3.41783702e-01 -2.82499403e-01
-8.72395992e-01 1.02837384e+00 3.28710556e-01 -8.12959298e-02
-8.85398507e-01 -7.58515537e-01 -8.17096651e-01 -2.05625650e-02
3.02749306e-01 -3.68716866e-01 1.07266402e+00 -1.06826305e+00
-1.36320233e+00 9.56420898e-01 1.35176271e-01 -5.32227635e-01
4.45286244e-01 -1.05039202e-01 -1.57640800e-01 4.25465912e-01
-1.18354738e-01 1.22328854e+00 9.62230206e-01 -1.64078724e+00
-5.13417006e-01 -2.47154042e-01 2.07773834e-01 4.02468652e-01
2.05859058e-02 -9.08233076e-02 -4.48959291e-01 -9.89034951e-01
-2.79020548e-01 -7.21376598e-01 -2.20162794e-01 6.23708367e-02
-6.37813564e-03 -4.28242207e-01 6.22926772e-01 -5.85156858e-01
8.15569043e-01 -2.38617063e+00 -3.62806439e-01 -8.59941915e-03
6.70923963e-02 4.64339703e-01 -4.73412603e-01 2.71360070e-01
3.06662470e-02 4.72364217e-01 -3.70764136e-01 -1.24220885e-01
-2.70547777e-01 3.15520242e-02 -1.55806884e-01 5.03805816e-01
9.23339069e-01 1.31131697e+00 -9.42555845e-01 -5.94264388e-01
4.46894646e-01 4.77223873e-01 -4.68055367e-01 1.77934691e-02
-3.70913357e-01 3.79151583e-01 -1.72627419e-01 7.15726137e-01
4.10516530e-01 -2.57244796e-01 -1.90872476e-01 -2.30419084e-01
7.95820355e-03 -3.35746855e-02 -7.06101716e-01 1.30053532e+00
-3.89795750e-01 1.17640102e+00 -5.27709544e-01 -1.06533957e+00
9.36758101e-01 6.61592782e-02 3.10284436e-01 -1.12434900e+00
1.54117271e-01 1.33944958e-01 3.40472847e-01 -4.99364048e-01
6.00298822e-01 -1.77846521e-01 1.58463761e-01 3.62854630e-01
1.74225673e-01 -4.16521966e-01 3.34508836e-01 1.55512244e-02
7.54741073e-01 4.63990718e-01 2.54656851e-01 -1.29981905e-01
1.96210876e-01 1.11027651e-01 9.79286060e-02 9.49335992e-01
-4.45848942e-01 8.27431977e-01 2.96821386e-01 8.75973031e-02
-9.32144165e-01 -1.06900370e+00 -2.81502008e-01 1.23214316e+00
3.41717482e-01 -1.02812875e-04 -4.53397244e-01 -8.09023142e-01
2.09893301e-01 7.11420059e-01 -8.06616604e-01 -3.34540308e-01
-3.86863738e-01 -8.47142875e-01 6.06153488e-01 9.01173770e-01
6.30807877e-01 -1.67001367e+00 -9.62667763e-01 -1.60821453e-02
9.03723463e-02 -1.03759778e+00 -4.03594933e-02 4.65212256e-01
-1.12497604e+00 -1.17902565e+00 -1.12163341e+00 -7.18293011e-01
6.84616148e-01 5.06517828e-01 1.03938603e+00 2.90092677e-01
-6.05260789e-01 6.17511272e-01 -5.03487706e-01 -4.64187980e-01
-2.79356629e-01 -1.33537114e-01 -4.47064489e-01 -1.99005798e-01
2.86880374e-01 -1.17995173e-01 -8.99617970e-01 4.19863433e-01
-1.05306375e+00 -1.78156301e-01 1.07502258e+00 6.54037237e-01
5.50807774e-01 -9.50069800e-02 8.20043921e-01 -9.05760705e-01
8.33037734e-01 -5.95278256e-02 -3.23715806e-01 2.62216568e-01
-5.24119735e-01 9.17366967e-02 3.97427291e-01 -4.88400131e-01
-1.06856835e+00 -1.47131428e-01 -1.11521043e-01 -2.49288157e-01
-4.14137095e-01 4.73869652e-01 7.89192989e-02 -1.23706348e-01
7.80966341e-01 2.12906346e-01 4.84269671e-02 -2.63439298e-01
5.45809507e-01 5.54367721e-01 3.55467200e-01 -7.20853209e-01
6.61924124e-01 4.62355167e-01 -1.03104949e-01 -1.10465539e+00
-7.51287818e-01 -7.00052679e-01 -4.75270510e-01 -4.44968462e-01
9.14652824e-01 -8.00077260e-01 -7.62773395e-01 3.04310769e-01
-8.83300245e-01 -8.63117039e-01 -5.83159626e-01 3.78196120e-01
-8.42034340e-01 2.73184061e-01 -3.98153573e-01 -8.87944043e-01
-1.14504710e-01 -1.36645722e+00 8.98315787e-01 5.16929150e-01
-1.84743181e-02 -8.79827738e-01 -2.85793841e-01 4.88174200e-01
4.00619596e-01 1.91013917e-01 9.48415399e-01 -4.67602611e-01
-5.66898406e-01 1.41093237e-02 -8.54218185e-01 6.08818710e-01
-3.49760465e-02 -3.86073589e-02 -1.28108764e+00 -2.99487919e-01
-5.20848334e-01 -5.94838023e-01 1.09934187e+00 5.63793838e-01
1.32560480e+00 3.11914802e-01 -1.14506431e-01 4.15498585e-01
1.65401387e+00 1.93151042e-01 9.09236193e-01 4.23753142e-01
4.40310240e-01 7.62792826e-01 7.71873653e-01 3.47723439e-02
2.96381600e-02 5.17636359e-01 3.84158373e-01 -3.20583552e-01
-6.78715706e-01 -9.74063650e-02 2.42288888e-01 1.84250906e-01
-3.89790267e-01 -1.63690895e-01 -6.83543742e-01 4.55842614e-01
-1.55560040e+00 -8.50511611e-01 3.05053368e-02 2.13282084e+00
6.64771020e-01 1.88540563e-01 3.15450341e-01 1.28844157e-01
5.77353835e-01 1.09713905e-01 -6.15518153e-01 -5.28705835e-01
-3.24040562e-01 3.43414873e-01 5.29933751e-01 8.46354142e-02
-1.00268126e+00 1.14798999e+00 6.07985401e+00 8.44994724e-01
-1.23821509e+00 -2.94403713e-02 8.65498126e-01 3.57727349e-01
-2.46738810e-02 -1.26167521e-01 -8.01624060e-01 2.22946957e-01
5.23217618e-01 2.91877151e-01 3.12244862e-01 8.12454402e-01
2.02649876e-01 -5.69531620e-01 -1.02352870e+00 9.63209271e-01
1.59446120e-01 -1.30904913e+00 2.48883903e-01 7.24068806e-02
7.45167494e-01 1.61413488e-03 4.15860474e-01 5.54297090e-01
9.28624626e-03 -1.29682565e+00 7.02038348e-01 3.28091294e-01
9.40583527e-01 -7.43663788e-01 6.55500114e-01 -7.61840492e-02
-9.28551435e-01 -1.63820758e-01 -4.93165791e-01 2.19447702e-01
-2.02887535e-01 2.16206118e-01 -8.27909470e-01 2.30939984e-01
7.68097281e-01 7.07780302e-01 -1.19129193e+00 1.50307918e+00
-3.97175550e-01 6.25923634e-01 -2.74665598e-02 -2.32686698e-01
4.60868627e-01 1.08436599e-01 1.80683970e-01 1.55481839e+00
-1.75856531e-01 -9.78164747e-02 -1.99380249e-01 7.27560759e-01
-1.40910462e-01 1.53188959e-01 -5.54932535e-01 -1.94434389e-01
2.03338102e-01 1.08781219e+00 -1.25343859e+00 -1.22324541e-01
-3.70258361e-01 7.71445990e-01 5.08099556e-01 7.03869820e-01
-6.14065051e-01 -4.16730642e-01 3.97514015e-01 2.73257971e-01
4.67545241e-01 -1.14583135e-01 -5.38531840e-01 -6.40831769e-01
-3.05681318e-01 -9.06786025e-01 3.72129738e-01 -9.66227353e-01
-1.29646409e+00 2.70146251e-01 7.73167014e-02 -1.17785573e+00
1.41860828e-01 -1.19703150e+00 -4.19831008e-01 5.66427886e-01
-1.87792861e+00 -1.00054252e+00 -3.32462162e-01 4.81812596e-01
7.18472004e-01 -1.09660566e-01 4.14327383e-01 -4.05081734e-02
-3.53741944e-01 7.74014473e-01 -4.69854251e-02 2.53140599e-01
6.74768329e-01 -1.37638104e+00 1.10946624e-02 7.00550258e-01
3.55383098e-01 4.58835453e-01 5.27878940e-01 -4.81214076e-01
-1.12010229e+00 -1.06173944e+00 1.18022271e-01 -3.73680890e-01
2.95516759e-01 -2.54967630e-01 -8.64680052e-01 4.71947044e-01
1.23108640e-01 8.99164751e-02 6.26463115e-01 1.72155648e-02
-3.61983746e-01 -1.26812354e-01 -1.08908069e+00 8.33392441e-01
8.52953136e-01 -4.28732187e-01 -4.43712771e-01 1.86697081e-01
4.82466966e-01 -1.35056719e-01 -8.16204607e-01 5.27179360e-01
3.50067198e-01 -1.02380228e+00 1.21609831e+00 -5.48291981e-01
7.31276929e-01 -2.10036933e-01 -9.55306292e-02 -1.26676023e+00
-2.73331046e-01 -1.31719992e-01 1.46061495e-01 1.15409029e+00
5.39978027e-01 -4.08944339e-01 9.68088686e-01 8.16187784e-02
-1.10636048e-01 -1.07509041e+00 -6.05815351e-01 -7.93765604e-01
2.08782181e-01 -7.03842580e-01 1.10290132e-01 5.25769711e-01
-3.36425245e-01 2.36832783e-01 -4.93849404e-02 -7.91141614e-02
3.81635517e-01 2.58686375e-02 8.07770848e-01 -9.68339562e-01
-2.72282332e-01 -8.53268862e-01 -5.72067678e-01 -8.89277160e-01
-3.52467746e-02 -1.06882048e+00 3.25577080e-01 -1.90804160e+00
1.08541645e-01 -5.59920251e-01 -3.62479568e-01 3.03753793e-01
-3.58491778e-01 6.08894169e-01 5.40977597e-01 1.15645245e-01
-5.58960617e-01 3.85422379e-01 1.68434381e+00 -2.70212114e-01
-2.45010525e-01 1.70333730e-03 -7.55367279e-01 4.42913115e-01
9.92805302e-01 -2.31896222e-01 -5.82641602e-01 -2.68603027e-01
-8.63529891e-02 -2.35752627e-01 5.19706368e-01 -8.92557085e-01
-2.08599642e-01 -1.72591284e-01 7.64227152e-01 -4.76229846e-01
3.19427967e-01 -5.87684393e-01 -6.36923671e-01 3.78782392e-01
-4.48032737e-01 -2.61542369e-02 6.37428939e-01 6.87786341e-01
-1.19638935e-01 -4.67917591e-01 1.06560230e+00 -1.39410585e-01
-1.21080017e+00 4.59007397e-02 -5.43038607e-01 4.44449484e-01
1.01084638e+00 -6.91842794e-01 -4.30187017e-01 -2.85896152e-01
-4.60840851e-01 2.49485951e-02 3.42097491e-01 4.93046224e-01
8.92474115e-01 -9.00668561e-01 -6.42852008e-01 -7.07009137e-02
4.50231731e-01 -1.46807477e-01 5.86418472e-02 7.34514713e-01
-6.27777994e-01 4.02757972e-01 -4.00658429e-01 -7.55577147e-01
-9.95757461e-01 6.80072427e-01 2.03998461e-01 -2.36139610e-01
-6.64184451e-01 9.10489082e-01 3.14200044e-01 -1.44993231e-01
4.05078679e-01 -3.29026878e-01 -2.78217435e-01 7.06674010e-02
2.63327509e-01 2.90816903e-01 -9.77932243e-04 -4.46956575e-01
-2.99832225e-01 5.04093349e-01 -7.97136575e-02 -1.88461587e-01
1.25685620e+00 5.88461980e-02 3.83306235e-01 4.19462919e-01
1.13484061e+00 -3.43990207e-01 -1.53937685e+00 4.35749926e-02
1.59930706e-01 -3.87793839e-01 7.90199963e-04 -1.03241301e+00
-1.16656852e+00 1.19632602e+00 6.98464215e-01 -8.30676630e-02
1.06678891e+00 1.39293700e-01 1.74626902e-01 3.36910248e-01
2.37223983e-01 -1.24682987e+00 7.39893317e-01 2.36811861e-01
1.10673308e+00 -1.46414864e+00 1.81390747e-01 -2.43396252e-01
-8.39174628e-01 9.52906668e-01 8.56123626e-01 -3.57500404e-01
4.15408999e-01 -3.68609428e-02 4.64896522e-02 -1.50104746e-01
-2.31622413e-01 -7.08734632e-01 6.09025598e-01 1.00665009e+00
6.34060502e-01 -8.36587846e-02 -2.09663779e-01 5.36451340e-02
9.85520612e-03 -1.35936841e-01 3.78822625e-01 9.03811038e-01
-5.14455497e-01 -6.91659749e-01 -2.10872605e-01 7.41476893e-01
-2.40700111e-01 -8.97509083e-02 -3.54165554e-01 1.27030778e+00
8.82046148e-02 9.09776092e-01 8.99945349e-02 -1.24606527e-01
2.56600767e-01 -6.64305091e-02 8.19613516e-01 -7.99419403e-01
-7.22895920e-01 -1.00847505e-01 -9.20256898e-02 -3.10514241e-01
-4.64019775e-01 -5.72905481e-01 -1.19615376e+00 1.07128955e-01
-3.89738947e-01 -2.39726037e-01 6.84854925e-01 9.62479830e-01
8.20795372e-02 7.72276223e-01 2.09927008e-01 -8.44185472e-01
-3.68406087e-01 -9.21404123e-01 -4.36039984e-01 6.52281523e-01
1.01510309e-01 -8.15298378e-01 -1.86263815e-01 -9.00927261e-02] | [9.761754989624023, 1.7031868696212769] |
da8ed64d-6679-432e-885c-06aff5b68e28 | towards-unifying-multi-lingual-and-cross | 2305.09220 | null | https://arxiv.org/abs/2305.09220v1 | https://arxiv.org/pdf/2305.09220v1.pdf | Towards Unifying Multi-Lingual and Cross-Lingual Summarization | To adapt text summarization to the multilingual world, previous work proposes multi-lingual summarization (MLS) and cross-lingual summarization (CLS). However, these two tasks have been studied separately due to the different definitions, which limits the compatible and systematic research on both of them. In this paper, we aim to unify MLS and CLS into a more general setting, i.e., many-to-many summarization (M2MS), where a single model could process documents in any language and generate their summaries also in any language. As the first step towards M2MS, we conduct preliminary studies to show that M2MS can better transfer task knowledge across different languages than MLS and CLS. Furthermore, we propose Pisces, a pre-trained M2MS model that learns language modeling, cross-lingual ability and summarization ability via three-stage pre-training. Experimental results indicate that our Pisces significantly outperforms the state-of-the-art baselines, especially in the zero-shot directions, where there is no training data from the source-language documents to the target-language summaries. | ['Jie zhou', 'Jianfeng Qu', 'Zhixu Li', 'Yunlong Liang', 'Duo Zheng', 'Fandong Meng', 'Jiaan Wang'] | 2023-05-16 | null | null | null | null | ['text-summarization'] | ['natural-language-processing'] | [ 5.26569821e-02 2.47312933e-01 -4.50924367e-01 -2.34287411e-01
-1.52024007e+00 -6.05571508e-01 9.03649628e-01 2.38499165e-01
-3.04199517e-01 9.73553300e-01 9.46999729e-01 -2.76759207e-01
5.06947458e-01 -4.30583954e-01 -7.20780313e-01 -2.70835251e-01
2.67182678e-01 5.14575124e-01 2.02631548e-01 -3.67187589e-01
2.86730796e-01 -5.32833263e-02 -1.07069314e+00 6.81496799e-01
1.59840071e+00 1.28508821e-01 4.14432734e-01 7.24746048e-01
-5.98471045e-01 6.37013674e-01 -9.07692850e-01 -6.22236788e-01
-2.21748188e-01 -7.44597673e-01 -8.39007795e-01 -1.18434183e-01
6.30282581e-01 -1.18296362e-01 1.73937883e-02 9.38359678e-01
8.29816520e-01 4.49873619e-02 8.93912137e-01 -8.33485365e-01
-8.83624256e-01 1.19672704e+00 -6.77171886e-01 -1.49574932e-02
6.53899372e-01 -9.82663482e-02 1.04915833e+00 -7.52949655e-01
6.70883298e-01 1.59993291e+00 6.32449567e-01 6.96700335e-01
-9.93000805e-01 -4.25867140e-01 3.49429786e-01 -5.69605976e-02
-9.24686074e-01 -7.32725799e-01 4.82536674e-01 -1.87298551e-01
1.30929577e+00 3.40420544e-01 2.65466183e-01 1.20137811e+00
3.98784965e-01 1.44299066e+00 8.74418855e-01 -5.63966334e-01
3.69582027e-02 1.84478670e-01 2.48271376e-01 4.09407198e-01
6.97827339e-01 -6.93584561e-01 -6.90742612e-01 3.53786140e-03
1.75962493e-01 -5.75555682e-01 -2.16111913e-01 1.79902688e-01
-1.44224548e+00 8.10915232e-01 -9.07075107e-02 5.57927847e-01
-2.64585793e-01 -1.69457465e-01 9.50510323e-01 3.45574200e-01
8.69988501e-01 7.65782297e-01 -3.13568264e-01 -8.37670732e-03
-1.37826073e+00 1.56372964e-01 1.11483765e+00 1.25118911e+00
4.84959990e-01 2.46143311e-01 -5.16513407e-01 9.38794196e-01
1.94554720e-02 6.46644235e-01 9.77857530e-01 -6.30399168e-01
1.08699751e+00 3.60679567e-01 -1.10334493e-01 -3.53082716e-01
-2.32314393e-01 -3.32043201e-01 -1.07068217e+00 -7.78157234e-01
-2.13927969e-01 -4.54639703e-01 -5.11872709e-01 1.68462062e+00
-1.24354720e-01 -1.79329608e-02 8.39588463e-01 2.63787061e-01
1.25385237e+00 1.16800916e+00 -6.44399896e-02 -6.62661850e-01
1.20258522e+00 -1.48632121e+00 -9.73259509e-01 -5.43202460e-01
1.06029975e+00 -1.00090301e+00 1.25881624e+00 1.51512370e-01
-1.41696417e+00 -4.68455315e-01 -1.06090856e+00 -4.57366258e-01
-3.92768949e-01 4.75845337e-01 3.66282761e-01 3.16260695e-01
-1.29113793e+00 3.70006949e-01 -8.34881365e-01 -8.46289814e-01
-1.54530749e-01 -2.61676431e-01 -2.78189391e-01 2.16280948e-02
-1.40966260e+00 1.18967891e+00 7.46772587e-01 -2.77728200e-01
-4.07180399e-01 -4.85055655e-01 -1.15100968e+00 1.45784877e-02
2.35369489e-01 -1.06522357e+00 1.50146627e+00 -5.18701494e-01
-1.69663107e+00 8.04645181e-01 -6.25632167e-01 -5.63142717e-01
4.69238907e-01 -4.93313283e-01 -2.65275478e-01 -1.15351357e-01
5.91761887e-01 4.82800037e-01 3.42025489e-01 -1.32859015e+00
-7.24277258e-01 -4.01402935e-02 -2.70074904e-01 6.28233969e-01
-3.52304071e-01 1.72769353e-01 -6.94399655e-01 -7.81054676e-01
-3.25076640e-01 -5.91586053e-01 -3.81981954e-02 -1.19257462e+00
-9.06284332e-01 -6.76128387e-01 4.77091402e-01 -1.08510101e+00
1.76509666e+00 -1.61486351e+00 3.90641659e-01 -5.56327045e-01
-2.90861905e-01 5.13500452e-01 -5.72198749e-01 1.11738753e+00
3.52583885e-01 3.02730739e-01 -3.65288734e-01 -1.00308883e+00
6.47970140e-02 9.52493399e-02 -5.37558913e-01 -5.79033233e-02
6.56760484e-02 1.26060593e+00 -1.08055389e+00 -7.66470194e-01
-1.33319005e-01 7.64646381e-02 -2.50807315e-01 1.95551485e-01
-2.51925230e-01 3.89576316e-01 -4.62390572e-01 2.58972675e-01
3.84403348e-01 1.05245553e-01 2.24821165e-01 -1.37799367e-01
-3.97118002e-01 6.87091231e-01 -6.69966638e-01 2.10560441e+00
-7.56418884e-01 3.66527468e-01 -2.61121958e-01 -8.91564786e-01
7.32107818e-01 4.40090150e-01 1.19239263e-01 -6.56998396e-01
-1.55232027e-01 5.12407362e-01 -3.22585374e-01 -6.03227198e-01
1.08994663e+00 -9.98571888e-03 -6.58048391e-01 6.56253874e-01
1.44099876e-01 -5.05704939e-01 8.85709763e-01 6.74177706e-01
5.49725592e-01 6.21422045e-02 8.26111794e-01 -3.15915912e-01
6.68824196e-01 -1.08495948e-03 5.45828521e-01 9.38816965e-01
2.60119945e-01 5.09587169e-01 4.32118416e-01 3.07018489e-01
-9.24558401e-01 -9.49644089e-01 1.95854664e-01 1.22958410e+00
1.11264726e-02 -7.55174279e-01 -1.12176538e+00 -7.41229713e-01
-2.77414352e-01 1.50436080e+00 -1.21902607e-01 -1.57417357e-01
-8.13921213e-01 -7.51652002e-01 7.99432099e-01 3.76506627e-01
5.58353066e-01 -1.11305118e+00 -2.52890229e-01 2.64383227e-01
-7.67420053e-01 -1.17673552e+00 -9.52322006e-01 -3.14167023e-01
-8.44605088e-01 -5.90835571e-01 -9.26934361e-01 -1.08674431e+00
2.95547813e-01 3.55223030e-01 1.23051274e+00 -4.99012202e-01
3.89869183e-01 4.09461290e-01 -4.64623809e-01 -5.95186532e-01
-9.29543614e-01 7.90883303e-01 1.45181119e-01 -2.88792759e-01
1.63454279e-01 -3.02887321e-01 -2.53403820e-02 -4.17037398e-01
-8.99320900e-01 5.23745596e-01 9.51085508e-01 4.93192047e-01
4.07945454e-01 -3.01763147e-01 1.16006637e+00 -1.17845809e+00
1.47865164e+00 -3.13745111e-01 3.20120007e-02 9.05903816e-01
-2.49908537e-01 2.19738975e-01 8.23000491e-01 -2.32804105e-01
-1.27063441e+00 -5.01069486e-01 -7.18513802e-02 2.96724945e-01
9.76937115e-02 1.18675482e+00 -2.57934153e-01 6.46340191e-01
3.99171561e-01 7.11215258e-01 -3.34123343e-01 -5.72731256e-01
7.12245584e-01 9.22766387e-01 6.78743362e-01 -5.27485609e-01
4.97897655e-01 9.61282477e-03 -5.59026301e-01 -1.26195884e+00
-1.14162934e+00 -5.47935188e-01 -7.58034110e-01 9.43490341e-02
8.58027279e-01 -1.13746512e+00 6.71201870e-02 5.37881434e-01
-1.68459427e+00 -3.17258418e-01 -2.92915553e-01 4.57354367e-01
-5.89819610e-01 8.08288455e-01 -7.47342050e-01 -5.93596160e-01
-1.07181036e+00 -1.03678238e+00 1.48828065e+00 2.03717142e-01
-4.39518660e-01 -1.35248911e+00 3.20643157e-01 2.39804491e-01
3.52421224e-01 2.41061556e-03 9.74725842e-01 -1.04570055e+00
-1.39391586e-01 -1.92056462e-01 8.65455065e-03 4.68101084e-01
2.84757167e-01 -9.56030637e-02 -5.16820729e-01 -5.06650507e-01
8.45482573e-02 -5.16894162e-01 1.07247436e+00 5.41581750e-01
6.07133090e-01 -7.13620007e-01 -2.33820125e-01 3.31224859e-01
1.05825555e+00 -1.43798888e-01 5.91438055e-01 2.58246064e-01
7.26414859e-01 6.11919403e-01 5.01152873e-01 4.53246906e-02
1.18959582e+00 4.60811019e-01 -3.75120044e-01 -1.89005539e-01
-4.28707272e-01 -5.28363824e-01 1.01142943e+00 2.08789754e+00
2.06517354e-01 -5.02887309e-01 -5.25759518e-01 6.34475112e-01
-2.09092021e+00 -9.66112435e-01 -1.21816352e-01 2.11430097e+00
1.25388420e+00 -1.02966502e-01 -5.93125708e-02 -5.12192488e-01
6.34117544e-01 4.47941184e-01 -4.98078495e-01 -7.40338027e-01
-4.44647193e-01 -1.77866757e-01 2.23469764e-01 7.07456470e-01
-9.13679421e-01 1.53741312e+00 6.06984854e+00 1.10909140e+00
-1.14281046e+00 5.37063666e-02 2.82163262e-01 -3.74946669e-02
-5.17059267e-01 -1.46723464e-01 -1.06571949e+00 3.09384018e-01
9.63668168e-01 -9.38360870e-01 1.50981501e-01 5.12041986e-01
3.85072410e-01 -7.91424140e-02 -1.17945075e+00 6.74300015e-01
6.67137563e-01 -1.16323483e+00 7.82519042e-01 -4.13442791e-01
1.21496618e+00 2.01458886e-01 -2.90767223e-01 7.70680010e-01
4.49598581e-01 -6.39378965e-01 6.90831602e-01 3.94186199e-01
8.14387619e-01 -5.78238189e-01 7.36129165e-01 8.25962782e-01
-1.22006059e+00 5.24926126e-01 -4.32192802e-01 3.08075011e-01
6.50843680e-01 3.71235460e-01 -7.26034105e-01 1.24773669e+00
9.85921547e-02 1.01745617e+00 -7.12561786e-01 7.52238572e-01
-4.49139535e-01 5.77938259e-01 2.50147888e-03 -9.59182605e-02
4.13276583e-01 -3.07793647e-01 6.89855456e-01 1.76797175e+00
5.99430263e-01 -2.71090001e-01 5.11997640e-01 4.37105566e-01
-2.39865988e-01 6.15832150e-01 -6.91012979e-01 -7.94734731e-02
5.22432506e-01 1.08295441e+00 -4.03194278e-01 -9.65297341e-01
-4.71026599e-01 1.12365925e+00 3.71880472e-01 4.66497272e-01
-5.07795036e-01 -6.43728077e-01 1.38094738e-01 -1.93923846e-01
-5.53546213e-02 -2.57684857e-01 -3.37918133e-01 -1.81066370e+00
-5.15482351e-02 -1.08371890e+00 4.12399650e-01 -5.69495022e-01
-1.40170896e+00 7.88516402e-01 2.64057606e-01 -1.05412078e+00
-6.48717880e-01 6.52869642e-02 -1.00489450e+00 8.95913422e-01
-1.81366217e+00 -1.50845182e+00 2.47919977e-01 2.11001158e-01
1.31113183e+00 -3.42350960e-01 7.85961986e-01 -1.39625845e-02
-7.15823710e-01 6.32622540e-01 4.30195630e-01 -1.99779287e-01
1.22132909e+00 -1.28475499e+00 8.74453962e-01 1.28040266e+00
5.35441674e-02 7.80160129e-01 7.56716669e-01 -8.89257967e-01
-1.25705874e+00 -1.41624582e+00 1.72299767e+00 -2.63916284e-01
7.62730658e-01 -3.19024533e-01 -9.66304719e-01 9.94539499e-01
9.61704373e-01 -1.09148002e+00 6.94152057e-01 1.05355762e-01
-6.58809468e-02 2.13963538e-01 -5.77988863e-01 1.02234495e+00
8.28493178e-01 -4.46509361e-01 -1.08813930e+00 6.33902848e-01
1.00853014e+00 -2.31473893e-01 -6.63056493e-01 3.72775763e-01
1.79197162e-01 -7.37944961e-01 6.43698633e-01 -5.62301517e-01
6.34848535e-01 -1.37075454e-01 -9.82278585e-02 -1.89050698e+00
-9.64993238e-03 -6.55017853e-01 -1.45934418e-01 1.77994037e+00
5.55968404e-01 -7.62231946e-01 -3.64336818e-02 -2.14679632e-03
-7.08307385e-01 -4.92806703e-01 -7.10640728e-01 -9.12330389e-01
5.85883558e-01 -1.14959754e-01 4.62727368e-01 8.99667740e-01
2.89875656e-01 1.03891063e+00 -4.97425288e-01 -6.08672574e-02
4.28141862e-01 2.44042933e-01 9.84562576e-01 -9.00335550e-01
-6.44250736e-02 -7.23376989e-01 4.41740662e-01 -1.41408157e+00
5.97582042e-01 -1.38107705e+00 1.19190432e-01 -2.35034275e+00
3.78611505e-01 1.93401501e-01 2.58361787e-01 2.53556669e-01
-6.42032504e-01 -2.22162545e-01 2.38042355e-01 2.58471727e-01
-9.80815113e-01 9.53506708e-01 1.28266037e+00 -1.39040127e-01
-4.77015227e-01 5.24591804e-02 -1.11354041e+00 5.93023956e-01
7.54945993e-01 -1.99853227e-01 -4.94289905e-01 -8.34677517e-01
-1.86268032e-01 1.37260243e-01 -5.03786504e-01 -7.13421404e-01
3.78050268e-01 -4.19006031e-03 -2.16771767e-01 -9.52570200e-01
-3.04741487e-02 1.45955503e-01 -4.06040847e-01 3.79063606e-01
-6.58810973e-01 1.57725781e-01 2.93097764e-01 2.04314455e-01
-4.71541017e-01 -4.17144835e-01 4.02353346e-01 -2.68501729e-01
-5.19947350e-01 -1.87034607e-02 -5.29417932e-01 5.43935657e-01
6.84501648e-01 7.71336555e-02 -5.54060161e-01 -6.02100551e-01
-1.63926914e-01 6.91443264e-01 3.65780830e-01 6.30546451e-01
2.72625536e-01 -1.22635627e+00 -1.35698140e+00 -2.15527594e-01
1.52650505e-01 4.12720218e-02 2.68735945e-01 7.81764328e-01
-2.68069655e-01 9.78561282e-01 2.16724262e-01 -4.03889149e-01
-1.20247340e+00 3.49857748e-01 -1.10697985e-01 -8.40822160e-01
-5.87755859e-01 4.65417594e-01 3.96809846e-01 -7.53212512e-01
1.30936295e-01 -3.62048358e-01 -4.14786160e-01 3.39423567e-01
5.49060106e-01 3.84567410e-01 -4.83571924e-02 -6.40838623e-01
-1.06588453e-01 5.54639459e-01 -3.70758653e-01 -2.49915317e-01
1.10966289e+00 -4.28417772e-01 -4.33720648e-01 8.85097802e-01
9.69877660e-01 2.80004174e-01 -5.90190768e-01 -4.40380484e-01
3.85185421e-01 3.34981859e-01 -3.33037585e-01 -7.16578186e-01
-4.48605359e-01 7.92310297e-01 -4.73599911e-01 1.64944515e-01
9.17094529e-01 9.05538499e-02 1.12691879e+00 5.56987047e-01
1.26286969e-01 -1.29058909e+00 1.64454186e-03 9.92192268e-01
1.29858589e+00 -1.06406236e+00 1.76277414e-01 -1.78726971e-01
-1.23758066e+00 9.64027703e-01 5.10956287e-01 1.06215671e-01
-4.18457426e-02 1.07998718e-02 7.68705085e-02 1.66463152e-01
-9.73311722e-01 -1.33592948e-01 5.17770886e-01 3.90728027e-01
9.12780762e-01 1.59238741e-01 -7.01896369e-01 8.20132434e-01
-6.65859580e-01 -1.53418019e-01 7.52451956e-01 6.58407569e-01
-4.99896020e-01 -1.20848691e+00 -1.21705338e-01 4.10389751e-01
-2.92581439e-01 -4.47120160e-01 -6.08013809e-01 7.63482809e-01
-4.76389974e-01 1.04825830e+00 -1.47424191e-01 -5.87005951e-02
5.00524700e-01 3.02280039e-01 3.59260291e-01 -1.01972234e+00
-5.68223715e-01 2.53392905e-01 4.00648177e-01 -1.22037441e-01
-3.99728477e-01 -7.76485503e-01 -1.05163980e+00 -2.66773701e-01
-2.42128208e-01 4.34351861e-01 6.14183843e-01 1.05822241e+00
4.11074132e-01 6.44938886e-01 4.41388905e-01 -1.02015090e+00
-7.36656189e-01 -1.41590345e+00 -3.30030501e-01 1.37487546e-01
1.40373796e-01 1.46169350e-01 -1.13606244e-01 1.61978260e-01] | [12.423633575439453, 9.54761028289795] |
9737953f-5e58-4ac2-9b6d-3ed67734120d | optimized-eeg-based-mood-detection-with | 2304.01349 | null | https://arxiv.org/abs/2304.01349v1 | https://arxiv.org/pdf/2304.01349v1.pdf | Optimized EEG based mood detection with signal processing and deep neural networks for brain-computer interface | Electroencephalogram (EEG) is a very promising and widely implemented procedure to study brain signals and activities by amplifying and measuring the post-synaptical potential arising from electrical impulses produced by neurons and detected by specialized electrodes attached to specific points in the scalp. It can be studied for detecting brain abnormalities, headaches, and other conditions. However, there are limited studies performed to establish a smart decision-making model to identify EEG's relation with the mood of the subject. In this experiment, EEG signals of 28 healthy human subjects have been observed with consent and attempts have been made to study and recognise moods. Savitzky-Golay band-pass filtering and Independent Component Analysis have been used for data filtration.Different neural network algorithms have been implemented to analyze and classify the EEG data based on the mood of the subject. The model is further optimised by the usage of Blackman window-based Fourier Transformation and extracting the most significant frequencies for each electrode. Using these techniques, up to 96.01% detection accuracy has been obtained. | ['Deepraj Chowdhury', 'Biswajit Saha', 'Kushal Jain', 'Subhrangshu Adhikary'] | 2023-03-30 | null | null | null | null | ['eeg', 'eeg'] | ['methodology', 'time-series'] | [ 3.07294220e-01 -3.45878720e-01 5.13227284e-01 -3.23127091e-01
1.70282841e-01 -2.12403789e-01 3.54080558e-01 3.35134417e-01
-6.02275014e-01 8.12013090e-01 -5.50717190e-02 3.54694091e-02
-2.16329172e-01 -5.31674385e-01 -1.77729860e-01 -8.61166120e-01
-5.07614076e-01 -1.11809708e-01 -1.02732599e-01 5.91717884e-02
6.52225077e-01 4.23734903e-01 -1.53285670e+00 1.71188802e-01
7.94729948e-01 8.30897152e-01 4.15636569e-01 5.50524473e-01
2.40664154e-01 2.67079353e-01 -1.03644049e+00 2.13292301e-01
-1.26308709e-01 -7.36078858e-01 -3.33278269e-01 1.36896390e-02
-7.68431127e-01 1.29542470e-01 2.92699039e-02 1.03060460e+00
7.36095428e-01 -6.38957247e-02 8.06955934e-01 -8.72901738e-01
-3.98282409e-01 2.46368840e-01 -6.32467389e-01 8.52595747e-01
5.63565850e-01 -2.43527189e-01 9.47859418e-03 -7.20892370e-01
1.23991765e-01 5.71365237e-01 4.86912042e-01 1.47476047e-01
-1.15610874e+00 -1.03050315e+00 -4.52388883e-01 9.31033731e-01
-1.51937103e+00 -2.93484390e-01 9.34319675e-01 -4.15417939e-01
1.05964553e+00 2.81532377e-01 1.23305309e+00 8.11612070e-01
8.97240639e-01 1.01285815e-01 1.43715763e+00 -5.40650070e-01
3.12482774e-01 4.18340206e-01 4.29207683e-01 8.15468207e-02
9.61516798e-02 -2.36711636e-01 -6.25842750e-01 -2.40932629e-01
4.56803828e-01 -2.99124002e-01 -5.76918602e-01 2.57696420e-01
-9.10239339e-01 4.01761681e-01 1.69917807e-01 8.15568507e-01
-9.62202489e-01 -2.67146289e-01 3.55168492e-01 3.71271938e-01
4.87014979e-01 5.44006109e-01 -1.15559652e-01 -1.30688369e-01
-9.75292981e-01 -3.00590038e-01 4.13224220e-01 2.04564676e-01
2.58091390e-01 -4.57877479e-02 1.12689629e-01 5.20141721e-01
1.53512970e-01 4.92633849e-01 8.81437480e-01 -3.91422242e-01
-1.97569892e-01 5.91410577e-01 1.52630219e-02 -8.95258188e-01
-5.95230997e-01 -1.28301606e-01 -7.07727790e-01 3.33102793e-01
8.02506283e-02 -2.88641453e-01 -5.11160672e-01 1.14474535e+00
-1.04473479e-01 5.09873688e-01 2.35473644e-02 6.37930989e-01
4.84759659e-01 6.94731534e-01 -7.23993927e-02 -5.16146779e-01
1.58988857e+00 7.94458091e-02 -9.40893114e-01 -7.41417240e-03
-7.89115876e-02 -6.25811040e-01 7.52148330e-01 8.44194472e-01
-9.16572690e-01 -5.42084694e-01 -1.13039982e+00 5.51594615e-01
-3.71764094e-01 1.17966786e-01 4.32455420e-01 8.24497581e-01
-8.37965250e-01 3.88891160e-01 -8.66593778e-01 -2.74889231e-01
1.74678922e-01 5.20794392e-01 -4.62385416e-01 5.47607780e-01
-1.07504475e+00 1.22479069e+00 2.17511386e-01 2.53313392e-01
-2.64211208e-01 -4.24400151e-01 -2.13806257e-01 1.26971960e-01
-5.17417967e-01 -4.20680404e-01 5.38594186e-01 -1.09220731e+00
-1.44029379e+00 8.25959384e-01 -3.83570462e-01 -4.58391547e-01
-7.60673210e-02 1.87015563e-01 -7.27483094e-01 3.31365138e-01
-1.39512494e-01 3.62923145e-02 9.57816601e-01 -4.05931294e-01
-2.91958809e-01 -7.54218221e-01 -7.10555553e-01 4.39086333e-02
-4.89393771e-01 4.17156011e-01 3.63185018e-01 -3.85920227e-01
2.42814660e-01 -5.66994607e-01 2.20487863e-01 -8.15641046e-01
1.03775010e-01 -1.67926431e-01 4.64892268e-01 -8.42239618e-01
1.13206518e+00 -2.29861236e+00 -1.12595886e-01 6.46399915e-01
-1.64650485e-01 9.92505550e-02 1.59900993e-01 3.55500996e-01
-3.97941828e-01 -4.21001732e-01 -2.91089378e-02 3.56413931e-01
-2.83391744e-01 -2.64001608e-01 1.03211720e-02 8.09686482e-01
1.06064640e-01 4.06663209e-01 -3.21712404e-01 -1.64340865e-02
3.99069071e-01 7.59627402e-01 -6.04745634e-02 1.72890395e-01
6.38373494e-01 6.37215853e-01 -1.13530338e-01 2.81146973e-01
5.59459090e-01 2.45909810e-01 -8.25079083e-02 -7.19250366e-02
-3.47468019e-01 4.07942146e-01 -9.48678911e-01 1.26048970e+00
-4.34372663e-01 1.02828419e+00 -6.02713302e-02 -1.24122393e+00
1.18384576e+00 7.13773608e-01 2.89710790e-01 -8.82601917e-01
6.01745546e-01 1.58534482e-01 3.80436182e-01 -9.28976774e-01
-1.40225202e-01 -1.38301402e-01 4.18983340e-01 2.42879063e-01
-2.36942545e-02 -4.28667553e-02 1.98719397e-01 -3.18503141e-01
9.46108639e-01 -2.44504392e-01 4.66063589e-01 -5.33108890e-01
6.93900168e-01 -4.75420773e-01 2.91116778e-02 5.99956699e-02
-4.64744940e-02 1.96538731e-01 3.51462513e-01 -2.70617455e-01
-3.46530139e-01 -4.80074972e-01 -5.38372755e-01 5.89326143e-01
5.88943027e-02 2.01163992e-01 -9.23283458e-01 3.26802850e-01
-3.29417527e-01 7.85052776e-01 -5.89863002e-01 -3.28441083e-01
-1.39016155e-02 -1.27282310e+00 2.70847470e-01 2.26356223e-01
4.55613643e-01 -1.35422003e+00 -1.23551035e+00 5.86596787e-01
9.19776410e-02 -5.44399381e-01 2.88820684e-01 5.65122783e-01
-8.42097282e-01 -9.46979821e-01 -8.02445829e-01 -5.30191004e-01
6.48252130e-01 -1.10666707e-01 4.09148544e-01 -2.19422981e-01
-5.82622349e-01 2.12229818e-01 -1.89537883e-01 -8.86141837e-01
-2.47110799e-02 -2.71064579e-01 9.73220542e-02 3.04026961e-01
9.69675958e-01 -1.00708389e+00 -5.84932446e-01 1.33564368e-01
-4.35031027e-01 -4.31911498e-01 5.75266838e-01 1.24825716e-01
1.50607646e-01 2.90761769e-01 7.97141850e-01 -2.58846879e-01
1.25568700e+00 -3.30242425e-01 -3.80713016e-01 1.99848488e-02
-5.16166866e-01 -2.19886124e-01 3.53011400e-01 -4.99705970e-01
-7.60997117e-01 -1.15591906e-01 -1.02267548e-01 1.62988484e-01
-5.47724545e-01 2.95594662e-01 -1.88738070e-02 -3.68063480e-01
7.71459043e-01 4.00129348e-01 7.26528913e-02 -1.44924819e-01
-4.85034585e-01 9.82579529e-01 8.46458972e-01 1.32051110e-01
7.19666332e-02 1.79269254e-01 -1.24049157e-01 -1.16583538e+00
6.40530605e-03 -6.00718141e-01 -3.50436032e-01 -5.10724664e-01
9.00346875e-01 -7.01665819e-01 -8.06334019e-01 7.44603693e-01
-1.15715218e+00 2.26063341e-01 3.34017217e-01 1.04205167e+00
-2.16303527e-01 2.29798883e-01 -4.31686640e-01 -1.07459378e+00
-8.26892316e-01 -7.67852545e-01 1.84349000e-01 3.67601991e-01
-5.80210209e-01 -5.42090774e-01 9.48920399e-02 -4.51057851e-02
4.12073195e-01 3.97693552e-02 8.21788251e-01 -5.38122594e-01
2.81026721e-01 -5.29283583e-01 2.04758272e-01 3.12802970e-01
1.48185775e-01 -2.93210205e-02 -1.19653511e+00 9.94601846e-02
9.35269833e-01 2.72086143e-01 4.04761016e-01 7.09674537e-01
8.95196140e-01 1.57980159e-01 -2.24278748e-01 3.71033579e-01
1.13442779e+00 1.04387057e+00 1.01127100e+00 3.34246606e-01
-1.50703534e-01 4.37328696e-01 1.88261885e-02 3.59331280e-01
-2.09247932e-01 9.30767953e-02 1.19979337e-01 1.16664164e-01
3.15340191e-01 5.51861525e-01 4.27252710e-01 7.81125009e-01
-5.42990744e-01 -1.52317867e-01 -7.86397517e-01 2.63164133e-01
-1.10611725e+00 -9.61124957e-01 -7.08211005e-01 2.13271260e+00
3.82221758e-01 3.44635755e-01 1.08408532e-03 9.27052438e-01
8.44305634e-01 -5.15107632e-01 -3.82243514e-01 -5.75173140e-01
3.95135954e-02 8.00204456e-01 3.50799620e-01 1.69976115e-01
-5.42203546e-01 -3.27917933e-02 6.60468435e+00 2.42442712e-01
-1.58932292e+00 2.90504489e-02 1.55940399e-01 -2.16454923e-01
3.00696850e-01 -4.49060589e-01 -1.83059037e-01 9.29590523e-01
1.22258914e+00 -2.07824126e-01 5.59331715e-01 4.53064173e-01
4.72755671e-01 -7.99873173e-01 -6.19951963e-01 1.29421258e+00
8.73509645e-02 -7.84588218e-01 -5.49407840e-01 -9.47442651e-02
2.01508403e-01 -1.82109013e-01 -3.97770368e-02 -1.58365875e-01
-8.91858876e-01 -8.51462364e-01 6.27635658e-01 8.58473241e-01
5.26180446e-01 -9.67204452e-01 8.42575848e-01 4.08542603e-01
-8.08534682e-01 -1.84125692e-01 -2.35193580e-01 -6.67783678e-01
-1.11628242e-01 6.37014091e-01 -6.88764811e-01 3.59605737e-02
6.19398952e-01 3.26923132e-01 -4.32646751e-01 1.20811522e+00
-2.38808945e-01 7.60210216e-01 -4.65718240e-01 -3.67825627e-01
-1.64593667e-01 -4.76286471e-01 4.15280432e-01 1.00573957e+00
6.08661950e-01 3.68196458e-01 -9.04398203e-01 9.37027335e-01
4.79350835e-01 2.86094368e-01 -4.63016808e-01 6.29257485e-02
2.81526029e-01 1.17312086e+00 -1.22281480e+00 -1.20481916e-01
-2.00799659e-01 9.91565585e-01 -3.93523514e-01 2.39838809e-01
-5.57382166e-01 -9.01201308e-01 8.37268233e-02 2.35939726e-01
-2.41158143e-01 6.03239692e-04 -6.06501341e-01 -7.57770479e-01
3.62726860e-02 -4.70634252e-01 -8.87682289e-02 -9.01196837e-01
-7.26085484e-01 6.78664207e-01 2.66137049e-02 -7.54076064e-01
-2.11695164e-01 -4.55085427e-01 -1.13235831e+00 1.23212504e+00
-1.05165517e+00 -1.73233196e-01 -1.32737949e-01 8.82478952e-01
2.66295016e-01 -1.95364222e-01 1.10123110e+00 2.35780001e-01
-4.82165515e-01 -1.12744436e-01 -1.09061226e-02 -1.67133808e-01
3.76867652e-01 -8.90788674e-01 -9.22909752e-02 7.61961758e-01
-2.28146855e-02 6.08393967e-01 8.53355289e-01 -4.12180871e-01
-1.00777471e+00 -1.67645514e-01 1.08948839e+00 1.31246105e-01
4.45986569e-01 -5.10702729e-01 -9.56386626e-01 4.72570173e-02
5.61557472e-01 -3.81142348e-01 8.68945301e-01 -2.98547745e-01
4.35119778e-01 -1.15666494e-01 -1.22825813e+00 1.02184333e-01
2.45781064e-01 -4.77735609e-01 -1.08942413e+00 1.02084264e-01
-4.32954401e-01 2.35789582e-01 -8.30842495e-01 6.08337820e-02
5.99584758e-01 -1.31828427e+00 6.23906076e-01 6.19917400e-02
-7.47007877e-02 -5.90510741e-02 5.89090347e-01 -1.42581880e+00
-3.47700506e-01 -5.50365567e-01 4.76975322e-01 9.23977554e-01
2.53288209e-01 -1.00640488e+00 3.14663678e-01 5.29028058e-01
1.18637301e-01 -4.14267510e-01 -7.50808358e-01 -3.55507910e-01
-6.13611162e-01 -3.50114286e-01 3.83666098e-01 5.03753543e-01
7.79924393e-01 4.02494848e-01 4.78799790e-02 4.32774238e-02
3.34274650e-01 -2.02481300e-01 -1.19942762e-02 -1.26682425e+00
9.49219987e-02 -3.70786548e-01 -7.48845756e-01 -7.86311775e-02
7.17366859e-02 -7.64341235e-01 -2.04877526e-01 -1.46717513e+00
1.83675401e-02 1.53635502e-01 -5.75414121e-01 2.34488070e-01
-1.07131209e-02 3.75951439e-01 -2.93334991e-01 -1.17282018e-01
2.36692473e-01 1.20918704e-02 5.14560878e-01 2.28255212e-01
-4.69807416e-01 3.36425751e-01 -3.10629249e-01 7.52885282e-01
1.08071482e+00 -4.91793603e-01 -5.78290224e-01 2.71519553e-03
2.33896017e-01 1.14048779e-01 2.67890841e-01 -1.36438859e+00
2.64628738e-01 3.97005767e-01 8.71989489e-01 -4.92915392e-01
2.94421762e-01 -7.05095410e-01 6.54694259e-01 6.13073409e-01
-1.31638899e-01 1.59362867e-01 1.75922737e-01 2.37183556e-01
-2.11092904e-01 -4.77293104e-01 6.82541132e-01 2.85959877e-02
-5.28338969e-01 -4.63533521e-01 -1.18717110e+00 -5.51743865e-01
1.33046257e+00 -5.16895175e-01 1.43277749e-01 -1.56491011e-01
-6.98850811e-01 -2.67827570e-01 -1.04587656e-02 -2.03540679e-02
7.14046180e-01 -8.65990937e-01 -5.58537543e-01 8.23597789e-01
-1.91987708e-01 -7.90570319e-01 3.45719188e-01 1.22130799e+00
-4.38747168e-01 3.83258581e-01 -9.73085999e-01 -2.82453924e-01
-1.55476594e+00 5.04048169e-01 1.10544197e-01 4.61377323e-01
-5.63348174e-01 8.41300786e-01 -5.49756110e-01 7.35952735e-01
1.96859583e-01 -4.29702193e-01 -8.72798622e-01 6.03913963e-01
7.84912109e-01 5.50377309e-01 5.26959419e-01 -3.84115368e-01
-4.82177079e-01 4.75447625e-01 4.18284982e-01 -1.95730999e-01
1.45019352e+00 1.67722609e-02 -4.16286945e-01 6.21957719e-01
1.04387486e+00 -4.19443697e-02 -6.07665122e-01 5.23519874e-01
2.24578679e-02 -6.62829503e-02 2.18309596e-01 -7.56493866e-01
-7.95076907e-01 9.68682945e-01 8.40218008e-01 5.25360525e-01
1.55391550e+00 -5.21586657e-01 3.88091236e-01 7.99235329e-02
5.32699049e-01 -1.10342455e+00 -4.21774834e-01 -7.08959177e-02
6.78531885e-01 -5.05398571e-01 -1.05490267e-01 4.52177301e-02
-2.74327964e-01 1.41458631e+00 -9.00287926e-02 -5.23922861e-01
7.65031576e-01 4.73204553e-01 -1.62932992e-01 -2.20841497e-01
-5.19531608e-01 1.07041664e-01 2.42692083e-01 6.21447682e-01
5.60126185e-01 1.17893957e-01 -9.18538809e-01 7.94826627e-01
-2.81757444e-01 4.96895194e-01 5.44052720e-01 7.96592176e-01
-5.62469721e-01 -6.17704928e-01 -8.70674729e-01 7.02921808e-01
-6.94573760e-01 3.68310697e-02 -2.78672606e-01 4.38762099e-01
2.11484358e-01 1.08113039e+00 1.56265914e-01 -3.35094541e-01
2.97951877e-01 3.99212241e-01 7.33389318e-01 -3.31787676e-01
-6.59607947e-01 2.35984191e-01 -4.24467474e-01 -2.68929899e-01
-6.27809644e-01 -7.05306590e-01 -1.41652846e+00 1.55956611e-01
-2.90956348e-01 3.85959506e-01 1.27701342e+00 7.48121262e-01
1.64565131e-01 5.66718817e-01 3.76370460e-01 -8.11913788e-01
-1.07855611e-01 -1.29079294e+00 -9.83151853e-01 8.81170928e-02
8.85744616e-02 -4.83456820e-01 -4.89847988e-01 1.23113366e-02] | [13.309934616088867, 3.2971301078796387] |
fe39b6d8-1858-4bbb-96e0-3aabdb474e8f | occupancy-planes-for-single-view-rgb-d-human | 2208.02817 | null | https://arxiv.org/abs/2208.02817v2 | https://arxiv.org/pdf/2208.02817v2.pdf | Occupancy Planes for Single-view RGB-D Human Reconstruction | Single-view RGB-D human reconstruction with implicit functions is often formulated as per-point classification. Specifically, a set of 3D locations within the view-frustum of the camera are first projected independently onto the image and a corresponding feature is subsequently extracted for each 3D location. The feature of each 3D location is then used to classify independently whether the corresponding 3D point is inside or outside the observed object. This procedure leads to sub-optimal results because correlations between predictions for neighboring locations are only taken into account implicitly via the extracted features. For more accurate results we propose the occupancy planes (OPlanes) representation, which enables to formulate single-view RGB-D human reconstruction as occupancy prediction on planes which slice through the camera's view frustum. Such a representation provides more flexibility than voxel grids and enables to better leverage correlations than per-point classification. On the challenging S3D data we observe a simple classifier based on the OPlanes representation to yield compelling results, especially in difficult situations with partial occlusions due to other objects and partial visibility, which haven't been addressed by prior work. | ['Alexander G. Schwing', 'Zhongzheng Ren', 'Yuan-Ting Hu', 'Xiaoming Zhao'] | 2022-08-04 | null | null | null | null | ['3d-human-reconstruction'] | ['computer-vision'] | [ 2.24832416e-01 6.66218325e-02 -8.81062001e-02 -2.38021657e-01
-4.87848014e-01 -5.78131139e-01 7.63276756e-01 2.52794147e-01
-2.01717928e-01 4.82939243e-01 1.65419698e-01 4.94957305e-02
-1.10353880e-01 -6.84535742e-01 -5.82496881e-01 -6.83308184e-01
5.53326793e-02 6.65628314e-01 3.14261287e-01 9.04669017e-02
1.42788723e-01 8.74188960e-01 -1.73254502e+00 1.61887154e-01
1.56253651e-01 1.25277710e+00 1.15240209e-01 5.47070503e-01
1.38651803e-01 3.00496668e-01 -2.90289402e-01 2.17448041e-01
4.78039205e-01 -3.54134329e-02 -4.68135685e-01 5.80047846e-01
5.90944290e-01 -1.65479362e-01 -1.31539911e-01 5.45022130e-01
1.77635163e-01 1.70131341e-01 7.72111237e-01 -1.16048968e+00
1.69759169e-01 -3.63577485e-01 -7.25091636e-01 6.27546832e-02
8.94349635e-01 -2.68622786e-02 1.05180764e+00 -1.00880468e+00
7.97846079e-01 1.12949383e+00 6.85122967e-01 6.95312768e-02
-1.44105697e+00 -1.54379696e-01 3.87653679e-01 -1.92391891e-02
-1.43117917e+00 -2.42700279e-01 8.08111846e-01 -5.87081552e-01
9.18317795e-01 4.70009685e-01 1.10057867e+00 8.58035505e-01
2.32464418e-01 5.48404455e-01 1.26470649e+00 -3.47833097e-01
3.36214691e-01 1.62832841e-01 1.37539551e-01 7.70679712e-01
4.32063825e-02 1.29848510e-01 -7.95864761e-01 -2.38602191e-01
1.02553463e+00 4.62289065e-01 -4.65134114e-01 -1.24663877e+00
-1.33679676e+00 7.63922691e-01 4.75585461e-01 -1.28432989e-01
-4.14479047e-01 2.29259413e-02 -7.73175713e-03 -2.23101914e-01
5.29743373e-01 1.93188667e-01 -4.62734818e-01 -7.18619302e-02
-8.59890580e-01 4.19845611e-01 6.22962356e-01 9.94295835e-01
9.76672471e-01 -5.24449348e-01 2.01772466e-01 4.24301147e-01
2.97418386e-01 3.46655041e-01 -2.13960007e-01 -1.14950824e+00
4.02837664e-01 8.17654133e-01 3.65810275e-01 -9.33260739e-01
-6.94578350e-01 -4.80139464e-01 -6.68548048e-01 5.12569070e-01
3.75975132e-01 3.66628736e-01 -7.99674630e-01 1.35005426e+00
8.33037019e-01 -3.30123231e-02 -1.98050350e-01 1.11879337e+00
3.85262847e-01 2.98241913e-01 -5.95605552e-01 -1.15868963e-01
1.62096465e+00 -6.21669054e-01 -1.60263941e-01 -1.29948601e-01
4.79378045e-01 -5.39776921e-01 9.27005172e-01 6.03149533e-01
-8.96524072e-01 -4.65206176e-01 -9.94334996e-01 -1.18892096e-01
-3.26867372e-01 -4.92650606e-02 4.53815252e-01 5.71186066e-01
-8.14748704e-01 4.31127191e-01 -9.51056361e-01 -3.63038301e-01
2.58007139e-01 1.84763104e-01 -6.45385504e-01 -1.81230336e-01
-5.84096849e-01 8.99362087e-01 -6.87417462e-02 -5.88875152e-02
-5.58372021e-01 -7.55264342e-01 -9.84148026e-01 -1.58817545e-01
6.86817706e-01 -1.00338256e+00 8.08830619e-01 -3.59053820e-01
-1.09103703e+00 9.88348663e-01 -4.00204390e-01 -8.38303864e-02
6.86244488e-01 -2.22596586e-01 1.33237541e-01 2.05806613e-01
2.36388594e-01 3.60279173e-01 8.36272836e-01 -1.63279760e+00
-5.51339746e-01 -1.03754413e+00 3.47574353e-01 4.88807052e-01
3.58520895e-01 -5.69331408e-01 -5.10013938e-01 -2.98109323e-01
9.03337479e-01 -1.05054438e+00 -3.38351518e-01 3.04091483e-01
-5.25402009e-01 7.48324990e-02 7.13618457e-01 -3.16146880e-01
7.83873200e-01 -2.09911609e+00 4.30764019e-01 4.99118388e-01
3.62938046e-01 -5.40303767e-01 4.02284801e-01 3.11061144e-01
-2.00013481e-02 -2.06366047e-01 -1.11744277e-01 -6.48113251e-01
1.82209676e-03 4.58062530e-01 -7.19720796e-02 9.09982145e-01
2.35690903e-02 4.19415325e-01 -7.44293272e-01 -3.03118020e-01
6.73295081e-01 7.17211246e-01 -8.32679987e-01 8.19811076e-02
-1.04830697e-01 7.95652151e-01 -5.11170924e-01 7.01808333e-01
6.82518840e-01 -3.89655381e-01 -1.52567208e-01 -2.79882461e-01
-3.08926135e-01 2.62439549e-01 -1.47180593e+00 1.87867463e+00
-5.70003986e-01 2.06690356e-01 -2.85195392e-02 -5.13990581e-01
6.89127982e-01 1.82208180e-01 7.42765605e-01 -4.02369082e-01
-1.41079843e-01 -1.34500742e-01 -5.90258181e-01 -1.57507375e-01
4.04983640e-01 -2.62333363e-01 -1.63671136e-01 2.06281230e-01
-2.17433020e-01 -2.66692728e-01 -3.29055518e-01 5.01703992e-02
1.23265958e+00 4.25772011e-01 6.04196191e-01 -1.73203245e-01
4.47370946e-01 -6.46178052e-02 3.93042415e-01 5.57427466e-01
-4.71634641e-02 1.08441710e+00 4.86226320e-01 -5.01479089e-01
-9.12783682e-01 -1.41825449e+00 -4.30086493e-01 4.84836906e-01
3.04519534e-01 -6.10725760e-01 -4.45541829e-01 -6.92072868e-01
3.26122701e-01 3.32669228e-01 -8.29618573e-01 3.38868797e-01
-3.40275466e-01 -1.65289015e-01 -1.85486645e-01 4.74425346e-01
8.40733871e-02 -3.70250732e-01 -1.39902675e+00 -7.16096982e-02
-7.80687630e-02 -1.17674863e+00 -8.64304602e-02 5.82976282e-01
-8.94208610e-01 -1.21528792e+00 -6.92732692e-01 -1.04416959e-01
7.51413286e-01 3.58898640e-01 1.00828815e+00 -2.24809900e-01
-3.71695668e-01 9.21378314e-01 -4.31202233e-01 1.21527791e-01
2.92879462e-01 -3.42279106e-01 3.54995370e-01 1.30303754e-02
1.83315620e-01 -7.04100430e-01 -8.29504490e-01 4.37808067e-01
-3.41895849e-01 1.48581490e-01 1.65583238e-01 7.28256464e-01
1.02745187e+00 1.32203922e-02 -2.78650939e-01 -6.52713239e-01
-3.66975307e-01 -5.52461624e-01 -7.34736025e-01 -1.63566731e-02
-9.46121663e-02 -7.45403394e-02 3.82047564e-01 -9.06735659e-02
-6.06950402e-01 4.26419735e-01 3.18280160e-01 -8.07591319e-01
-4.07076150e-01 2.18125582e-01 -2.28546277e-01 7.76555091e-02
2.29637176e-01 5.75298443e-02 -1.92801267e-01 -6.99416637e-01
2.39409700e-01 2.43902117e-01 2.82939464e-01 -5.26773453e-01
5.68446815e-01 9.88743782e-01 4.57331717e-01 -8.67180288e-01
-8.81051004e-01 -7.07274854e-01 -1.21468139e+00 -3.46706480e-01
1.10266435e+00 -8.90173793e-01 -8.70339394e-01 -1.25335470e-01
-1.19247651e+00 -7.58837014e-02 -3.63095284e-01 6.21606112e-01
-8.23792517e-01 3.98345709e-01 -4.66984771e-02 -1.08062172e+00
4.86045837e-01 -1.36078286e+00 1.66377759e+00 -1.70719162e-01
-3.90917182e-01 -9.06863809e-01 -2.31197681e-02 4.61471736e-01
-1.61219016e-01 5.18367350e-01 9.32675540e-01 -1.27125576e-01
-8.30512166e-01 -2.26529986e-01 -1.11814421e-02 1.23080658e-02
1.93306506e-02 -3.02950472e-01 -1.13649607e+00 -1.54592320e-01
1.98886216e-01 7.74098411e-02 6.16786182e-01 4.78044540e-01
1.09725547e+00 -1.87656265e-02 -4.82358903e-01 7.44757533e-01
1.38712418e+00 -2.79578060e-01 3.38314593e-01 2.38279581e-01
7.36897588e-01 8.09977829e-01 7.32042491e-01 8.69628370e-01
4.76361930e-01 1.18310654e+00 7.30973482e-01 -5.70997559e-02
4.56454307e-02 -2.45931864e-01 1.56643428e-02 2.21022755e-01
-2.52720535e-01 -8.69685337e-02 -9.73920524e-01 1.08788237e-01
-1.66756499e+00 -6.45766497e-01 -2.54837900e-01 2.48138714e+00
1.11033008e-01 2.03782227e-02 2.07311258e-01 3.46513242e-01
2.79226214e-01 2.94861585e-01 -5.59939504e-01 2.18053788e-01
5.70807159e-02 1.51367947e-01 5.42130470e-01 6.79462552e-01
-9.07429934e-01 4.86538470e-01 5.61744928e+00 3.57871771e-01
-7.73867786e-01 5.87791093e-02 4.39838469e-01 -4.63455468e-01
-3.08847755e-01 2.40130857e-01 -9.91370857e-01 2.50764012e-01
3.43749732e-01 5.28679132e-01 3.81277829e-01 7.27616370e-01
1.21090055e-01 -7.06653059e-01 -1.43391550e+00 1.25827861e+00
1.95444062e-01 -9.81042504e-01 -6.05207160e-02 6.59073591e-01
4.10675645e-01 -1.88750383e-02 1.27222449e-01 -7.23625049e-02
-2.35310003e-01 -7.81478703e-01 8.98274720e-01 6.30128086e-01
6.23921335e-01 -6.47240043e-01 1.96537927e-01 5.53577602e-01
-1.27401364e+00 6.37501776e-02 -2.85357952e-01 -2.42329165e-01
3.40128660e-01 7.77401984e-01 -7.98985779e-01 5.64330935e-01
8.24632525e-01 7.29465127e-01 -3.39572012e-01 7.90328801e-01
1.49178952e-01 -8.86963531e-02 -5.81103206e-01 4.11750168e-01
2.23691285e-01 -4.13617611e-01 7.48966515e-01 7.49489784e-01
4.39150870e-01 2.25039393e-01 3.52048635e-01 8.53261888e-01
4.58377212e-01 -7.62989074e-02 -8.96871984e-01 6.63914323e-01
1.52885735e-01 1.18075597e+00 -9.93419409e-01 -1.83653399e-01
-5.48727572e-01 1.04375327e+00 2.85830557e-01 4.15232122e-01
-6.49171948e-01 3.70034158e-01 7.82887936e-01 7.51277745e-01
5.47392488e-01 -5.03389120e-01 -3.22840720e-01 -1.20847452e+00
3.07015955e-01 -4.25884306e-01 2.83710092e-01 -9.48883593e-01
-9.89241779e-01 4.79205668e-01 1.73819661e-01 -1.64300168e+00
-9.41300020e-02 -8.51928055e-01 5.35463355e-03 8.26662600e-01
-1.44167745e+00 -1.02455497e+00 -5.36686242e-01 7.05179989e-01
4.17900980e-01 2.10078165e-01 1.03049839e+00 -1.32310241e-01
-3.55222784e-02 4.44156751e-02 -1.92437530e-01 -4.31199342e-01
2.78337538e-01 -1.34240949e+00 -6.99572563e-02 5.38157940e-01
2.55880475e-01 5.76213181e-01 5.88806033e-01 -6.06883347e-01
-1.53872859e+00 -7.80369103e-01 5.76118886e-01 -1.02464437e+00
2.34518960e-01 -6.65482581e-01 -6.02544904e-01 6.64914966e-01
-3.38932991e-01 5.11119306e-01 7.58727372e-01 3.08122784e-01
-5.13787031e-01 8.52435827e-02 -1.17527163e+00 3.36267859e-01
1.26925516e+00 -6.05837047e-01 -4.32413101e-01 2.92114258e-01
2.51846343e-01 -6.98568821e-01 -1.05946577e+00 3.71860355e-01
6.27181292e-01 -1.38842905e+00 1.30601406e+00 -2.47617498e-01
2.91996121e-01 -5.24074852e-01 -7.24288046e-01 -1.09287775e+00
-1.73646919e-02 -3.76736701e-01 -2.61131734e-01 6.14903927e-01
-4.34992686e-02 -6.16279304e-01 9.35513973e-01 5.38023651e-01
-1.18272193e-01 -9.77096438e-01 -1.34149671e+00 -6.57824814e-01
-4.61144894e-01 -5.93892097e-01 4.56564456e-01 6.43968225e-01
-7.27917999e-02 1.25914916e-01 -3.58895399e-02 5.85337877e-01
7.79239058e-01 3.04650694e-01 9.85211432e-01 -1.35880959e+00
-4.53813314e-01 -1.58111334e-01 -8.18065703e-01 -1.50035501e+00
-1.74771976e-02 -7.78860033e-01 -1.19492911e-01 -1.28565180e+00
7.86763951e-02 -7.61868298e-01 -7.22698048e-02 2.28591651e-01
8.58394504e-02 3.59555662e-01 1.28864005e-01 2.61554301e-01
-5.48801363e-01 4.77836430e-01 1.21024156e+00 2.97150999e-01
-5.56650721e-02 1.35744691e-01 -1.85956895e-01 9.12486255e-01
2.52677292e-01 -2.96007872e-01 -2.00777546e-01 -2.10563838e-01
1.85735375e-01 6.37622952e-01 8.39445412e-01 -1.13364565e+00
-5.06073385e-02 -9.16083083e-02 7.19173253e-01 -1.13507509e+00
1.15939867e+00 -1.26175165e+00 3.46233457e-01 2.24940300e-01
6.81782067e-02 1.39524862e-01 -1.79328158e-01 8.28847170e-01
6.99566677e-02 7.78036332e-03 5.98530173e-01 -4.00702417e-01
-4.41250145e-01 4.01997715e-01 2.58833054e-04 -2.42161646e-01
1.10866535e+00 -6.61383629e-01 3.01907271e-01 -4.25822675e-01
-1.09171999e+00 -3.71015891e-02 1.15750670e+00 1.57021865e-01
8.51564646e-01 -1.22955620e+00 -2.77472317e-01 5.96004546e-01
3.25401992e-01 3.90527904e-01 3.36270660e-01 1.00929618e+00
-2.88340420e-01 5.23245513e-01 3.75258513e-02 -1.43025684e+00
-1.28547096e+00 5.68419874e-01 3.26033175e-01 -1.34036765e-01
-1.15007746e+00 6.40454948e-01 5.28642356e-01 -4.55960631e-01
1.63389057e-01 -6.29709959e-01 -1.70705747e-02 5.18778302e-02
3.36443484e-01 3.16024244e-01 1.40157849e-01 -1.09447765e+00
-3.61908585e-01 9.41649914e-01 2.11674675e-01 -2.92383879e-01
1.27400124e+00 -3.52654040e-01 1.19996235e-01 7.83463001e-01
1.25657046e+00 2.59550065e-01 -1.60356414e+00 -1.28091872e-01
-2.62847781e-01 -9.41123843e-01 -1.17487907e-02 -5.29443324e-01
-7.89732099e-01 9.59799707e-01 6.06663346e-01 1.00029811e-01
7.41560102e-01 2.84899592e-01 3.15821797e-01 3.19674574e-02
9.17149544e-01 -5.20754933e-01 5.02487496e-02 3.61229956e-01
8.89955997e-01 -1.08944678e+00 3.15181673e-01 -6.66196883e-01
-5.42324066e-01 1.10593772e+00 2.87423551e-01 -2.02667043e-01
7.34668732e-01 3.87397818e-02 -2.41730437e-01 -4.56095397e-01
-4.09160376e-01 -1.93883464e-01 3.09206873e-01 6.26189530e-01
1.15420528e-01 1.08440906e-01 3.80359530e-01 3.69154215e-01
-4.16309595e-01 -4.35808420e-01 2.78174579e-01 9.83969331e-01
-2.88569540e-01 -8.33274186e-01 -6.34314537e-01 4.35528278e-01
-2.53848564e-02 2.66306221e-01 -4.60824594e-02 9.31838453e-01
3.35635751e-01 6.37712240e-01 4.96912688e-01 -3.06548208e-01
4.49068248e-01 1.67262126e-02 9.39727962e-01 -9.29186940e-01
-2.63034314e-01 3.64779919e-01 -9.43041965e-02 -1.00282502e+00
-5.54412246e-01 -1.04671800e+00 -1.25069666e+00 -2.58520208e-02
-1.75013274e-01 -8.78254920e-02 7.51148760e-01 8.40763688e-01
2.74672240e-01 1.97201669e-01 6.53754592e-01 -1.43113983e+00
-2.89589465e-01 -3.52576375e-01 -8.51479888e-01 3.78295898e-01
6.90591216e-01 -1.23530114e+00 -5.05003750e-01 -2.28187338e-01] | [7.6331787109375, -2.6096267700195312] |
d76c27f7-a7ff-497d-a726-f3f8e0782cde | unsupervised-semantic-correspondence-using | 2305.15581 | null | https://arxiv.org/abs/2305.15581v1 | https://arxiv.org/pdf/2305.15581v1.pdf | Unsupervised Semantic Correspondence Using Stable Diffusion | Text-to-image diffusion models are now capable of generating images that are often indistinguishable from real images. To generate such images, these models must understand the semantics of the objects they are asked to generate. In this work we show that, without any training, one can leverage this semantic knowledge within diffusion models to find semantic correspondences -- locations in multiple images that have the same semantic meaning. Specifically, given an image, we optimize the prompt embeddings of these models for maximum attention on the regions of interest. These optimized embeddings capture semantic information about the location, which can then be transferred to another image. By doing so we obtain results on par with the strongly supervised state of the art on the PF-Willow dataset and significantly outperform (20.9% relative for the SPair-71k dataset) any existing weakly or unsupervised method on PF-Willow, CUB-200 and SPair-71k datasets. | ['Kwang Moo Yi', 'Andrea Tagliasacchi', 'Abhishek Kar', 'Hossam Isack', 'Shweta Mahajan', 'Gopal Sharma', 'Eric Hedlin'] | 2023-05-24 | null | null | null | null | ['semantic-correspondence'] | ['computer-vision'] | [ 3.56008440e-01 4.30513471e-01 -1.80265293e-01 -3.12061787e-01
-5.61503589e-01 -7.25352705e-01 1.15686047e+00 2.52967458e-02
-5.24622858e-01 5.83535075e-01 5.08176506e-01 -6.43186346e-02
-1.30451709e-01 -9.56999481e-01 -8.65630209e-01 -6.57868147e-01
1.94902554e-01 8.04836094e-01 3.42111766e-01 -1.36147916e-01
3.41775179e-01 4.35989618e-01 -1.41873157e+00 5.86242855e-01
3.84077966e-01 9.57872510e-01 5.04528821e-01 8.64621758e-01
-1.77085921e-01 9.50221777e-01 -4.73767132e-01 -4.00125235e-01
2.14677155e-01 -5.13960719e-01 -9.00618792e-01 2.81750351e-01
6.86908960e-01 -4.34603065e-01 -5.95802546e-01 9.93900061e-01
3.39039207e-01 1.14329554e-01 1.16405928e+00 -1.13461673e+00
-1.46265304e+00 6.74456000e-01 -6.44516230e-01 3.08946878e-01
2.88208783e-01 7.63736740e-02 1.02886915e+00 -1.00491548e+00
1.39364445e+00 1.36871433e+00 2.19816521e-01 7.81505883e-01
-1.38100326e+00 -3.71902943e-01 1.98465269e-02 8.18089098e-02
-1.27664101e+00 -3.22345823e-01 5.57691097e-01 -5.73563814e-01
8.70864630e-01 -1.53622434e-01 5.88395238e-01 1.37171078e+00
-7.91650638e-02 9.27679241e-01 1.08674824e+00 -3.96928310e-01
2.50317365e-01 4.67832386e-01 -2.04100087e-01 6.04956031e-01
4.71646450e-02 2.63834000e-02 -6.19758427e-01 1.99871078e-01
8.44353795e-01 -3.61642465e-02 -1.67118164e-03 -4.11511511e-01
-1.52187610e+00 1.03510034e+00 8.11676979e-01 3.36539149e-01
-3.66866738e-01 4.45214063e-01 -1.25025143e-03 1.34230731e-02
5.92040718e-01 4.95121747e-01 -2.38610089e-01 1.18029669e-01
-8.77158999e-01 4.95345145e-01 4.89821732e-01 1.08142912e+00
7.60934174e-01 -8.41571465e-02 -3.75457197e-01 5.96626997e-01
3.43748122e-01 7.63056517e-01 6.05385542e-01 -1.09017861e+00
3.88236314e-01 4.19187486e-01 1.92276716e-01 -1.02255499e+00
5.42322956e-02 -1.61660284e-01 -6.40564203e-01 2.49077275e-01
2.88202256e-01 -7.85930604e-02 -1.21430898e+00 1.44836748e+00
2.23843545e-01 3.30806464e-01 3.92651975e-01 8.11426342e-01
5.65388381e-01 8.99290204e-01 9.99264866e-02 4.22034860e-01
1.29441929e+00 -1.19574940e+00 -3.70394051e-01 -3.66106778e-01
6.27679467e-01 -8.70058656e-01 1.06800807e+00 5.72073720e-02
-8.97820354e-01 -4.68958080e-01 -9.25841987e-01 -1.85305059e-01
-6.95556045e-01 -2.03343723e-02 4.78039831e-01 4.06567156e-01
-1.41724193e+00 4.49593604e-01 -1.97428659e-01 -5.65260470e-01
7.78773844e-01 -4.31734063e-02 -4.74213004e-01 -4.61795956e-01
-1.00145721e+00 6.97186172e-01 4.19866294e-01 -3.81214648e-01
-1.30484271e+00 -9.73497033e-01 -7.80286849e-01 -2.55427390e-01
-9.94563010e-03 -6.80554926e-01 1.03013361e+00 -1.15406740e+00
-1.06659806e+00 1.24907613e+00 -2.03330010e-01 -7.94080138e-01
6.15589738e-01 1.90117806e-02 -1.83179781e-01 3.91187400e-01
5.00769138e-01 1.66740763e+00 1.13553441e+00 -1.63620579e+00
-6.86700821e-01 -3.01223278e-01 1.52971685e-01 2.13496655e-01
-3.37196261e-01 -1.66094512e-01 -5.37080467e-01 -8.34793568e-01
-2.68129557e-01 -1.15722287e+00 -3.64486724e-01 2.51392722e-01
-5.87766886e-01 -1.96693763e-01 1.07359612e+00 -4.29289699e-01
5.33336937e-01 -2.20936036e+00 3.29981297e-01 1.73137173e-01
3.04155886e-01 5.59250005e-02 -5.64904511e-01 3.82526785e-01
-3.47270668e-02 1.77536175e-01 -2.39578649e-01 -2.99547583e-01
-1.73657224e-01 2.72530556e-01 -5.11271119e-01 2.68368632e-01
2.98215479e-01 1.28809106e+00 -1.04946160e+00 -3.76890630e-01
2.96053141e-01 5.05760491e-01 -5.46022058e-01 5.91235049e-02
-6.47954047e-01 2.87143856e-01 -4.98628169e-01 6.78109899e-02
4.69506443e-01 -5.76364756e-01 -4.80953604e-02 -5.89755923e-02
4.02463585e-01 -2.88950384e-01 -7.81934142e-01 1.86581862e+00
-4.95543689e-01 9.87337172e-01 -3.68262768e-01 -7.41617143e-01
6.78612590e-01 9.87367425e-03 3.25972408e-01 -7.97586918e-01
-1.81191146e-01 3.92665341e-02 -2.75198966e-01 -2.89334506e-01
6.69061244e-01 -8.58708769e-02 5.51340133e-02 7.92712867e-01
1.25858903e-01 -2.86447883e-01 2.59838015e-01 7.87246406e-01
1.00646579e+00 -3.30043994e-02 -4.15372759e-01 -3.18260938e-01
3.05246353e-01 2.13933632e-01 -1.84237063e-01 8.51304591e-01
2.33525246e-01 1.01154387e+00 4.00296956e-01 -3.31774265e-01
-1.36032116e+00 -1.33800185e+00 1.21123984e-01 7.99961865e-01
3.51534396e-01 -2.68085420e-01 -9.49327946e-01 -8.83633971e-01
1.85761616e-01 9.77187395e-01 -1.05622172e+00 6.33160174e-02
-1.23479106e-01 -5.23038745e-01 5.08617043e-01 4.19872433e-01
2.60138780e-01 -1.28918552e+00 -2.64916122e-01 8.62693861e-02
4.19637114e-02 -1.39383864e+00 -5.22832394e-01 -3.09787393e-01
-6.62861764e-01 -9.59513783e-01 -1.26356566e+00 -8.73301268e-01
1.14406478e+00 4.78285760e-01 1.12883079e+00 -9.53101218e-02
-5.43832958e-01 6.99750125e-01 -3.73073846e-01 -2.27742195e-01
-5.14799953e-01 4.33476716e-02 -3.80168170e-01 2.80156225e-01
2.45496705e-01 -2.00343937e-01 -8.63536298e-01 1.71122521e-01
-1.31257665e+00 2.19227925e-01 6.06167078e-01 6.09768212e-01
6.51753902e-01 2.13206142e-01 4.25409406e-01 -1.14648139e+00
8.32633674e-01 -6.50904834e-01 -3.10815036e-01 1.83169067e-01
-7.33819485e-01 2.65926480e-01 3.51062953e-01 -4.48276103e-01
-1.12071240e+00 1.83211237e-01 2.05702722e-01 -4.89967138e-01
-1.59682497e-01 2.03879662e-02 3.22362065e-01 4.76258174e-02
8.72294068e-01 4.50181752e-01 -1.63381428e-01 -2.62349546e-01
1.11904073e+00 5.48104227e-01 3.45658243e-01 -5.10158956e-01
7.66643167e-01 1.00902951e+00 -1.98312804e-01 -9.00712132e-01
-9.63189244e-01 -3.37510258e-01 -7.09681571e-01 -1.14254303e-01
1.28526759e+00 -9.90999997e-01 8.42969120e-02 5.65197706e-01
-1.22155142e+00 -5.62602520e-01 -6.83801591e-01 1.48137763e-01
-6.52655602e-01 1.71123728e-01 -4.76740986e-01 -3.97861302e-01
6.69281185e-02 -1.13098538e+00 1.27555239e+00 2.07996711e-01
-2.82927662e-01 -1.32639217e+00 -1.21837780e-01 4.92469579e-01
4.07801747e-01 -2.87747588e-02 9.67749596e-01 -3.58909637e-01
-8.14731061e-01 -1.45863697e-01 -6.69586420e-01 3.47380370e-01
7.14712739e-02 -1.45485014e-01 -9.09591734e-01 -4.76878472e-02
-6.25615180e-01 -4.10969704e-01 1.02835882e+00 4.31028247e-01
1.03003037e+00 -3.91908616e-01 -4.65157628e-01 2.37428680e-01
1.38140500e+00 -2.13669106e-01 8.70702744e-01 2.43528605e-01
6.91935599e-01 8.02100062e-01 3.14931750e-01 2.33124346e-01
5.46081781e-01 5.81016064e-01 2.10882455e-01 -2.62012064e-01
-7.49817073e-01 -7.17748404e-01 3.35505754e-01 3.27771991e-01
3.36477399e-01 -4.92234260e-01 -7.03993142e-01 1.14166319e+00
-1.83412802e+00 -9.89953518e-01 -1.50119290e-01 1.79532170e+00
6.86176240e-01 -9.72164795e-02 -1.08218066e-01 -4.14517760e-01
6.42657995e-01 4.45749789e-01 -5.10580063e-01 -1.64002076e-01
-4.44009125e-01 -4.24318090e-02 7.26472437e-01 6.25374079e-01
-8.08294952e-01 1.40286338e+00 6.88787699e+00 1.00369763e+00
-8.61764252e-01 3.15878242e-01 1.10163236e+00 -9.95263904e-02
-7.50472128e-01 8.56991038e-02 -7.91239262e-01 3.60535353e-01
8.06875467e-01 -2.85696477e-01 6.50946140e-01 7.93309152e-01
1.60202548e-01 -2.21364334e-01 -1.01753342e+00 9.12241161e-01
3.62877995e-01 -1.80279911e+00 5.20684063e-01 2.48304188e-01
1.44987333e+00 2.12318182e-01 5.07461607e-01 -1.65123001e-01
9.13892269e-01 -1.30666018e+00 7.86996365e-01 7.04237640e-01
6.44060850e-01 -3.69572848e-01 3.13047230e-01 2.31642351e-01
-7.13580430e-01 1.48679540e-01 -5.17346799e-01 3.50254089e-01
2.96713859e-01 7.59590983e-01 -9.74753618e-01 1.22309372e-01
6.90377712e-01 9.93557692e-01 -8.42753828e-01 5.70854127e-01
-5.84437013e-01 2.48439193e-01 -1.53149500e-01 -2.49187071e-02
6.51175737e-01 -3.52928676e-02 1.87972650e-01 1.15608823e+00
5.01589715e-01 -2.54422694e-01 -1.27508089e-01 1.13708270e+00
-4.66150373e-01 -1.41347665e-02 -1.00230968e+00 -2.61306226e-01
1.25471517e-01 1.22889340e+00 -8.93893003e-01 -7.62057960e-01
-2.73121685e-01 1.48816633e+00 3.79096299e-01 5.80517352e-01
-6.93848491e-01 -4.99655157e-02 5.45963705e-01 3.05489242e-01
5.82041800e-01 -1.06026165e-01 -1.17508993e-01 -9.58065271e-01
-1.87736586e-01 -5.58086038e-01 1.39878467e-01 -1.46972072e+00
-1.52132058e+00 5.42496204e-01 5.41954227e-02 -6.28191233e-01
-3.17236573e-01 -5.62734365e-01 -4.23695624e-01 8.34510684e-01
-1.59512007e+00 -1.30480218e+00 -2.28029296e-01 6.31449819e-01
7.80734479e-01 -1.66488796e-01 5.87667108e-01 -3.71610001e-02
7.92043656e-02 2.56680958e-02 2.37486854e-01 1.90395176e-01
6.17685497e-01 -1.28046262e+00 6.68089747e-01 6.34424031e-01
8.36087286e-01 3.16297799e-01 5.11801600e-01 -6.89068615e-01
-1.09006619e+00 -1.32090831e+00 8.70800793e-01 -8.55142534e-01
7.69331038e-01 -2.73347437e-01 -6.35517776e-01 6.79257989e-01
5.30060410e-01 9.28404182e-02 2.74625897e-01 -3.55327994e-01
-6.09228969e-01 1.31565422e-01 -1.07123137e+00 8.47698987e-01
1.22812366e+00 -6.26956284e-01 -4.50987130e-01 6.72329366e-01
8.17120612e-01 -4.69282791e-02 -7.10597098e-01 -2.34974131e-01
3.70842516e-01 -9.39010024e-01 1.27317727e+00 -7.95709968e-01
8.99128139e-01 -1.55984133e-01 -1.72027305e-01 -1.52423179e+00
-3.23972940e-01 -2.80404776e-01 1.59364000e-01 1.08553195e+00
7.62308121e-01 -3.72180551e-01 1.01347303e+00 4.14456695e-01
4.61103827e-01 -3.91202480e-01 -6.70253634e-01 -7.38049686e-01
2.81462699e-01 -3.98850471e-01 4.48830575e-01 7.87266195e-01
-4.86514509e-01 4.24218684e-01 -2.22424433e-01 -1.61862765e-02
8.28285694e-01 1.23668108e-02 7.32608378e-01 -8.57939720e-01
-1.78056180e-01 -4.89071816e-01 -5.42975962e-01 -1.29004276e+00
3.72345418e-01 -1.00355995e+00 -6.92172423e-02 -1.89361119e+00
2.80409515e-01 -6.36056960e-01 -2.43450016e-01 3.18229198e-01
-1.81082245e-02 6.00629091e-01 1.76763684e-01 2.92439938e-01
-6.02940023e-01 5.37804484e-01 1.47485292e+00 -5.51693380e-01
2.40322739e-01 -5.41362524e-01 -8.92978013e-01 4.80380148e-01
7.42600083e-01 -7.52976537e-01 -8.91945899e-01 -7.41782665e-01
2.00398773e-01 -2.77180761e-01 6.18104815e-01 -8.31639230e-01
2.22987518e-01 -1.72944427e-01 6.05255902e-01 -2.77913600e-01
3.54378074e-01 -6.77063823e-01 3.65613960e-02 2.88582981e-01
-8.02858174e-01 7.01438859e-02 -6.30137995e-02 1.09742069e+00
-7.79752955e-02 -4.13299710e-01 6.40718758e-01 -3.32649857e-01
-1.03673840e+00 3.95174563e-01 -4.53284204e-01 3.13705415e-01
1.43989444e+00 -3.95686217e-02 -3.96970451e-01 -7.38213360e-01
-6.74379468e-01 5.46745770e-02 7.10618436e-01 8.90239954e-01
8.08084607e-01 -1.41237164e+00 -8.31512272e-01 1.37488157e-01
4.66026753e-01 -7.37274438e-02 3.35597515e-01 3.72454554e-01
-5.42631209e-01 4.39061463e-01 -3.77561748e-02 -5.06101429e-01
-9.96211171e-01 7.00361192e-01 8.32520351e-02 -1.43364176e-01
-7.23005176e-01 9.48428094e-01 6.44078851e-01 -3.85515273e-01
-1.93520963e-01 -1.90512761e-02 5.17723300e-02 -1.69222541e-02
5.06274223e-01 8.41084681e-03 -3.92111093e-01 -9.70752060e-01
-9.34045985e-02 6.33370936e-01 -2.10830688e-01 -6.59088254e-01
1.25767577e+00 -1.13592252e-01 5.73818646e-02 8.17895979e-02
1.40472245e+00 3.25553864e-02 -1.51625097e+00 -3.00754845e-01
4.67452072e-02 -6.96621239e-01 1.49454713e-01 -7.99936295e-01
-1.28921854e+00 1.03264964e+00 5.19590139e-01 1.80736333e-01
5.74137866e-01 4.47692066e-01 7.17441678e-01 1.86813757e-01
1.73738360e-01 -1.15624261e+00 5.36916971e-01 1.56524613e-01
9.75926042e-01 -1.17599595e+00 -1.78647399e-01 -2.96260357e-01
-1.02127528e+00 8.72809827e-01 3.72205436e-01 -4.04793978e-01
7.15849519e-01 -1.09876157e-03 9.99552682e-02 -4.35530812e-01
-8.25364232e-01 -2.11853147e-01 2.03795105e-01 1.10668814e+00
1.55806899e-01 5.67602776e-02 3.08802605e-01 -1.03122927e-01
4.84024249e-02 1.70946214e-02 5.79289079e-01 5.73521435e-01
-4.32599634e-01 -1.02377546e+00 -7.11300075e-02 4.66658801e-01
-7.43605793e-02 -2.61175215e-01 -6.75881982e-01 5.61537802e-01
-6.12018593e-02 8.83778453e-01 1.94784477e-01 -1.17796645e-01
1.14069976e-01 -7.63502866e-02 6.20040238e-01 -7.33786047e-01
-1.14032641e-01 -2.40091473e-01 1.42680295e-02 -4.81676817e-01
-4.48375553e-01 -5.15549123e-01 -1.28151870e+00 -2.12105840e-01
1.73982188e-01 -6.92734718e-02 7.22085476e-01 7.08361804e-01
6.16313577e-01 4.39113677e-01 4.26769108e-01 -7.96892643e-01
-2.52969235e-01 -7.02924609e-01 -7.06820726e-01 9.02445078e-01
2.18404606e-02 -4.44914252e-01 -3.29477370e-01 4.48801041e-01] | [11.115790367126465, -0.01372417714446783] |
f7dc8592-5189-4717-a605-854f78f4fcdd | automated-refugee-case-analysis-an-nlp | 2305.15533 | null | https://arxiv.org/abs/2305.15533v1 | https://arxiv.org/pdf/2305.15533v1.pdf | Automated Refugee Case Analysis: An NLP Pipeline for Supporting Legal Practitioners | In this paper, we introduce an end-to-end pipeline for retrieving, processing, and extracting targeted information from legal cases. We investigate an under-studied legal domain with a case study on refugee law in Canada. Searching case law for past similar cases is a key part of legal work for both lawyers and judges, the potential end-users of our prototype. While traditional named-entity recognition labels such as dates provide meaningful information in legal work, we propose to extend existing models and retrieve a total of 19 useful categories of items from refugee cases. After creating a novel data set of cases, we perform information extraction based on state-of-the-art neural named-entity recognition (NER). We test different architectures including two transformer models, using contextual and non-contextual embeddings, and compare general purpose versus domain-specific pre-training. The results demonstrate that models pre-trained on legal data perform best despite their smaller size, suggesting that domain matching had a larger effect than network architecture. We achieve a F1 score above 90% on five of the targeted categories and over 80% on four further categories. | ['Nehal Bhuta', 'Michael Rovatsos', 'Claire Barale'] | 2023-05-24 | null | null | null | null | ['named-entity-recognition-ner'] | ['natural-language-processing'] | [-8.22561681e-02 1.81383774e-01 -4.11685169e-01 -6.83310986e-01
-1.26934361e+00 -7.85828412e-01 6.77396536e-01 3.38734061e-01
-1.24720645e+00 6.40574694e-01 9.07463253e-01 -5.95698178e-01
-5.61294973e-01 -6.91972494e-01 -3.53079438e-01 1.41083766e-02
-1.96408853e-02 6.64214551e-01 1.04599029e-01 -2.24001318e-01
2.93341219e-01 8.79825413e-01 -8.95343304e-01 4.38963592e-01
9.39065993e-01 4.38124418e-01 -4.41457093e-01 2.60679036e-01
-1.47886217e-01 9.38678741e-01 -9.09247458e-01 -1.29974604e+00
4.24669683e-01 2.05014184e-01 -8.44330907e-01 -6.17876530e-01
8.02124381e-01 -7.20055819e-01 -8.42399657e-01 6.44908905e-01
7.17372000e-01 2.62415588e-01 9.41613555e-01 -7.77404010e-01
-1.24417520e+00 5.10959208e-01 -1.26463726e-01 6.16949916e-01
3.84844780e-01 1.04002774e-01 1.11319733e+00 -6.02236688e-01
1.49275577e+00 9.98646080e-01 9.60807443e-01 7.12385237e-01
-1.02861691e+00 -7.46113360e-01 -1.20981991e-01 3.42123032e-01
-1.11098468e+00 -8.75775158e-01 3.53430063e-01 -5.73568523e-01
1.51482832e+00 -3.75230134e-01 2.44249821e-01 1.27469420e+00
2.86678165e-01 6.43865168e-01 5.89437842e-01 -1.30247220e-01
2.41519794e-01 -1.12195641e-01 4.75211710e-01 1.84397809e-02
6.88557863e-01 1.85732618e-01 -9.53984484e-02 -7.32514322e-01
4.50868189e-01 1.69592410e-01 2.13461995e-01 3.46457250e-02
-7.53143966e-01 7.86289155e-01 3.24840575e-01 6.97252870e-01
-6.91973627e-01 -1.01063199e-01 4.04418200e-01 2.13089108e-01
2.95467913e-01 6.68583870e-01 -5.06738305e-01 -4.62712049e-01
-1.10965347e+00 5.48656106e-01 8.23744357e-01 7.38680899e-01
3.07739884e-01 -3.83925050e-01 -6.09649479e-01 1.05114436e+00
-7.50010461e-02 5.21779716e-01 1.81719154e-01 -7.31143951e-01
9.93222415e-01 8.11692357e-01 2.17053980e-01 -7.08442986e-01
-4.63048905e-01 -2.86256373e-01 -2.96475679e-01 8.56379867e-02
7.85647571e-01 -3.55399609e-01 -1.36171448e+00 1.54296315e+00
1.54088542e-01 -2.60910302e-01 2.21245214e-01 6.73835337e-01
4.84638274e-01 2.00670093e-01 6.79681480e-01 3.06126267e-01
1.39046764e+00 -1.42272621e-01 -7.10274220e-01 -4.53531951e-01
5.68796754e-01 -7.33134925e-01 5.16846657e-01 -1.82407886e-01
-1.00266564e+00 -1.69488505e-01 -5.79531431e-01 -7.14578152e-01
-7.08737195e-01 1.92011669e-01 6.33375943e-01 6.37858748e-01
-6.80928469e-01 5.75300813e-01 -6.83506727e-01 -6.98577523e-01
1.00587666e+00 3.05819027e-02 -6.52606070e-01 -6.05868220e-01
-1.39946342e+00 1.31761444e+00 1.98003963e-01 -1.55085072e-01
-3.93030226e-01 -9.82775807e-01 -1.09693587e+00 1.86494112e-01
1.76502526e-01 -3.16914320e-01 1.11042368e+00 1.75710082e-01
-4.51539218e-01 1.31835973e+00 -2.38335624e-01 -5.14805377e-01
4.36189353e-01 -2.59943277e-01 -9.25838411e-01 2.44997978e-01
7.91206062e-01 4.80200410e-01 2.63462037e-01 -6.17987573e-01
-6.58372998e-01 -4.82982039e-01 1.47761598e-01 -3.02988410e-01
-4.12671059e-01 7.05479264e-01 -2.46050462e-01 -6.26588523e-01
-4.35838372e-01 -5.82877219e-01 -1.64107442e-01 -1.80883586e-01
-6.46079704e-02 -4.23327237e-01 3.25679719e-01 -1.18979156e+00
1.52630436e+00 -2.12647724e+00 -4.73840058e-01 1.86102256e-01
1.21722072e-01 4.94725168e-01 -2.68569618e-01 8.11711490e-01
-1.20995723e-01 3.42632115e-01 -2.27224141e-01 -1.32914409e-01
3.17001671e-01 6.24585450e-02 -2.90961385e-01 3.13470274e-01
6.94263995e-01 1.01725197e+00 -9.12282526e-01 -5.98634422e-01
-1.61460608e-01 4.64281410e-01 -4.81040180e-01 -2.49351323e-01
3.04699630e-01 -1.90445378e-01 -4.53778267e-01 8.65506649e-01
6.23810351e-01 2.02238724e-01 1.74006239e-01 7.13543147e-02
1.23856872e-01 8.12478006e-01 -7.81392872e-01 1.65703762e+00
-4.01033521e-01 7.17527568e-01 2.64671415e-01 -5.71823597e-01
6.47450447e-01 2.66914040e-01 2.57319033e-01 -1.10198271e+00
8.75378549e-02 3.22102189e-01 2.19740063e-01 -8.99387002e-01
6.39381230e-01 -2.18038157e-01 -4.72940296e-01 6.18719280e-01
-1.64068434e-02 5.53217471e-01 6.04374886e-01 4.31924582e-01
1.85803664e+00 -4.56575043e-02 3.29097778e-01 1.34818599e-01
-2.02929646e-01 4.64026302e-01 8.80884767e-01 8.12994123e-01
-5.07852852e-01 4.07595634e-01 5.60914934e-01 -4.20173168e-01
-1.21518695e+00 -1.01061153e+00 -5.10727942e-01 1.03203416e+00
-4.24708873e-01 -1.51963070e-01 -3.90339226e-01 -9.50463831e-01
4.67124611e-01 1.00872600e+00 -6.79931998e-01 -1.37008622e-01
-8.30582559e-01 -4.38707590e-01 1.01878011e+00 8.30234766e-01
2.13939145e-01 -1.32658601e+00 -7.76338398e-01 5.16452610e-01
1.58902094e-01 -1.21824658e+00 -3.72641385e-01 -1.08810112e-01
-5.94241917e-01 -1.42333508e+00 -7.38251388e-01 -6.22941792e-01
3.73284519e-01 -2.74473608e-01 1.06884754e+00 -7.69624189e-02
-4.24764037e-01 2.61547327e-01 -3.21654439e-01 -4.13251042e-01
-3.48595917e-01 3.36395591e-01 -1.16597936e-01 -4.05567884e-01
1.29722047e+00 -4.84069020e-01 -2.27445349e-01 -1.87481374e-01
-1.13325620e+00 -8.14396620e-01 9.09924507e-01 5.98305821e-01
-3.69941518e-02 -5.26950955e-01 7.76039898e-01 -1.06994748e+00
1.33246279e+00 -6.44126296e-01 -1.92470953e-01 5.05636215e-01
-4.74635035e-01 2.41469294e-02 4.60635990e-01 -3.26318026e-01
-1.31611228e+00 -4.09861296e-01 -3.50556493e-01 -1.68512642e-01
-6.17946565e-01 5.88603616e-01 -3.33537199e-02 6.84269786e-01
1.01312673e+00 -3.10165942e-01 -2.76323259e-01 -5.72975099e-01
3.42246920e-01 1.17570686e+00 6.62745595e-01 -5.67855775e-01
7.57546902e-01 3.85440528e-01 -4.29234594e-01 -4.41606492e-01
-6.51014090e-01 -6.96910203e-01 -8.04595590e-01 2.77982384e-01
9.28345263e-01 -6.00116432e-01 -3.89926702e-01 7.24198222e-02
-1.31988156e+00 -1.19468190e-01 -3.34386170e-01 7.20020294e-01
1.38349012e-01 1.35866463e-01 -7.84331143e-01 -7.86397099e-01
-3.70516241e-01 -6.62116945e-01 1.13186395e+00 5.78592382e-02
-4.44942951e-01 -8.16902757e-01 1.41959578e-01 5.66761136e-01
5.43254316e-01 4.32803214e-01 1.21994483e+00 -1.45033467e+00
-2.62614768e-02 -7.44369030e-01 -4.69520330e-01 4.63076830e-02
-6.80944026e-02 -3.48894268e-01 -8.80334854e-01 -5.81006370e-02
-3.44959646e-01 -1.01999961e-01 1.10653770e+00 9.54519883e-02
1.83589146e-01 -7.91599602e-02 -5.12049317e-01 2.07414597e-01
1.20369947e+00 6.55702949e-01 9.02960718e-01 6.13037467e-01
2.77289659e-01 5.50979495e-01 4.50294703e-01 1.33721486e-01
4.22444135e-01 3.75912994e-01 -2.16467395e-01 8.03779066e-02
-9.20721143e-02 -2.43992507e-01 2.93084621e-01 -1.34726360e-01
-1.12918951e-02 -6.24813773e-02 -1.53445733e+00 1.26230621e+00
-1.60929751e+00 -1.38142180e+00 3.09483349e-01 1.85097539e+00
8.32523406e-01 2.31914535e-01 7.47708157e-02 -2.29053617e-01
7.71627426e-01 4.19175029e-02 -7.79419303e-01 -4.52271312e-01
3.68282422e-02 6.39556766e-01 6.73286080e-01 1.28949732e-01
-1.24301505e+00 1.01117396e+00 6.28287888e+00 8.50166261e-01
-5.83680332e-01 2.83354908e-01 3.34790349e-01 -4.37649786e-01
-1.35110900e-01 2.32882742e-02 -9.00162160e-01 3.77015948e-01
1.25993061e+00 -1.45934820e-01 6.40629381e-02 9.91095126e-01
-2.89765513e-03 1.17692798e-01 -1.17525029e+00 7.22433686e-01
1.58311009e-01 -1.38994217e+00 -1.17433824e-01 5.54472148e-01
4.92460549e-01 3.29063922e-01 -3.77848655e-01 7.63202250e-01
4.68266189e-01 -9.44125772e-01 6.19625509e-01 6.61014795e-01
8.81761849e-01 -5.97374201e-01 9.33866799e-01 1.26060441e-01
-6.77502275e-01 -4.98182148e-01 -4.99839813e-01 -8.95272568e-03
5.35366952e-01 5.14322102e-01 -8.19514036e-01 4.70563710e-01
6.16062641e-01 7.07287073e-01 -5.70961237e-01 1.22906697e+00
-3.86829436e-01 5.32530189e-01 -4.16102707e-01 1.25620365e-01
3.21381122e-01 2.07395628e-01 3.63816619e-01 1.50103271e+00
1.36489853e-01 3.67833346e-01 -2.86383957e-01 7.30479121e-01
-5.62804759e-01 -1.54449895e-01 -8.76810312e-01 -1.29212409e-01
8.38413298e-01 1.01715910e+00 -3.58455598e-01 -5.12874126e-01
-3.34975660e-01 5.24786532e-01 6.72627211e-01 4.63913888e-01
-7.01386511e-01 -1.06936014e+00 8.25408697e-01 3.28212678e-01
4.77029711e-01 -1.99169338e-01 -1.73403814e-01 -1.17672455e+00
4.36850399e-01 -7.08860159e-01 7.03842342e-01 -5.97294629e-01
-1.72347236e+00 4.97555852e-01 2.95216829e-01 -9.61758435e-01
-4.67465252e-01 -6.85538828e-01 -5.09081185e-01 7.48195708e-01
-1.53684628e+00 -1.14395344e+00 2.66150266e-01 3.10528755e-01
4.80754152e-02 -3.77946973e-01 6.44586623e-01 7.92013288e-01
-4.58862245e-01 6.49831593e-01 2.05627322e-01 9.92024302e-01
1.03030133e+00 -9.15347397e-01 6.29924953e-01 1.01037252e+00
8.17559585e-02 1.38029170e+00 1.52754504e-02 -1.07585168e+00
-9.76253331e-01 -1.00394380e+00 1.65560198e+00 -7.97102213e-01
7.16990888e-01 -3.02071244e-01 -8.47339809e-01 8.46446872e-01
3.51926923e-01 -4.63970721e-01 9.60497856e-01 5.42445600e-01
-7.10025728e-01 9.04254690e-02 -1.66930401e+00 4.45073724e-01
1.54930866e+00 -8.31191897e-01 -1.46878839e+00 2.50899166e-01
2.70293683e-01 -1.60031050e-01 -1.15255880e+00 -5.25068156e-02
7.24934280e-01 -6.84362888e-01 9.06862199e-01 -1.29536617e+00
4.79177743e-01 1.53046265e-01 1.66360185e-01 -9.83532906e-01
-4.57623482e-01 -1.57136515e-01 3.15161169e-01 1.49759638e+00
8.00784469e-01 -6.87209070e-01 5.33255816e-01 1.41335618e+00
-9.49432477e-02 -4.65114474e-01 -1.28148913e+00 -9.21324492e-01
4.00081158e-01 -7.66529322e-01 7.45740891e-01 1.26692319e+00
1.70133144e-01 3.39205891e-01 6.78697973e-02 3.01156621e-02
4.05845553e-01 -5.21506444e-02 4.04909968e-01 -1.32304704e+00
2.79207677e-01 -3.49686593e-01 -6.08347237e-01 -4.89860088e-01
4.46972698e-01 -1.08803236e+00 -5.31055212e-01 -2.23570848e+00
1.70777410e-01 -3.69913697e-01 -2.20741779e-01 1.04523826e+00
-3.67987528e-02 -1.11379698e-01 3.56665790e-01 2.38700300e-01
-4.66407776e-01 9.13725272e-02 5.86358726e-01 -2.86682338e-01
-2.20508128e-01 -4.29367006e-01 -9.57112014e-01 3.82473379e-01
6.02100492e-01 -9.10697699e-01 7.23836869e-02 -9.54326689e-01
2.17311308e-01 -2.44718015e-01 2.95169234e-01 -9.34652984e-01
4.81260270e-01 -1.19391911e-01 5.60674250e-01 -4.79656696e-01
1.67203218e-01 -8.09413254e-01 -3.39921832e-01 5.27557768e-02
-4.77716416e-01 -1.16008691e-01 3.11744958e-01 3.81974012e-01
-1.59361362e-01 -5.58627665e-01 2.78474003e-01 3.93436588e-02
-5.88485658e-01 -4.88557555e-02 -4.25975144e-01 4.09595788e-01
8.84971857e-01 -3.29353899e-01 -7.70594656e-01 -1.48768753e-01
-8.18224669e-01 3.26694876e-01 3.78696799e-01 4.50411975e-01
5.40408194e-01 -1.19093084e+00 -8.55292201e-01 1.18678212e-02
2.75375485e-01 -3.66712004e-01 1.31646529e-01 5.02814949e-01
-3.80160660e-01 6.71102881e-01 -1.84116364e-01 2.49917969e-01
-8.59749317e-01 4.23162311e-01 -1.27724260e-02 -6.49223566e-01
-4.02672499e-01 3.25096101e-01 -6.36132717e-01 -6.07263565e-01
-6.23380803e-02 -6.08704329e-01 -3.84867907e-01 3.99241030e-01
5.84722102e-01 4.78556812e-01 2.44996116e-01 -6.90126956e-01
-7.35242963e-01 4.45055753e-01 -5.15145123e-01 -1.33542120e-01
1.91666102e+00 6.53502643e-01 1.55846253e-01 -3.57272089e-01
1.18025839e+00 2.84149915e-01 -5.91477513e-01 -1.46233067e-01
4.85426128e-01 -4.61894929e-01 -2.50773877e-01 -1.09810936e+00
-8.50525856e-01 7.62832224e-01 3.45366269e-01 8.39022696e-02
8.44964683e-01 1.14576250e-01 1.00426972e+00 8.67145240e-01
4.13821071e-01 -1.29330039e+00 -8.08852792e-01 7.28000581e-01
7.87043691e-01 -1.00454152e+00 9.42835733e-02 1.54314727e-01
-6.73348904e-01 9.47331071e-01 2.78620631e-01 -1.89009354e-01
5.11119187e-01 8.54706913e-02 2.50790089e-01 -5.09187520e-01
-5.93825161e-01 -4.91806090e-01 1.97811514e-01 7.88955629e-01
3.61357957e-01 -3.44906718e-01 -6.85176075e-01 7.97650874e-01
-1.10887557e-01 2.78878659e-01 3.12834948e-01 1.27664149e+00
-1.37362480e-01 -1.20213711e+00 -3.33543330e-01 9.18038189e-01
-9.71596241e-01 -3.27288836e-01 -6.55492663e-01 1.24978387e+00
2.97171772e-01 1.04158890e+00 1.96978763e-01 -1.38012111e-01
9.60653126e-01 4.04869646e-01 3.92042585e-02 -8.42972159e-01
-1.07151508e+00 -4.57284510e-01 6.90862238e-01 -4.41212475e-01
-3.69842321e-01 -9.12276685e-01 -1.27003384e+00 -1.06586620e-01
9.23197716e-02 1.44230932e-01 5.55969656e-01 9.65147376e-01
9.56154346e-01 8.26003030e-02 -1.76105618e-01 -3.60191047e-01
-5.99050224e-01 -1.30998743e+00 -3.91143680e-01 1.00977683e+00
-5.74857276e-03 -5.87264121e-01 -1.10519789e-01 -3.31355631e-02] | [9.864666938781738, 9.301202774047852] |
38231643-c552-48b8-a531-c5d35f637c1e | sclifd-supervised-contrastive-knowledge | 2302.05929 | null | https://arxiv.org/abs/2302.05929v1 | https://arxiv.org/pdf/2302.05929v1.pdf | SCLIFD:Supervised Contrastive Knowledge Distillation for Incremental Fault Diagnosis under Limited Fault Data | Intelligent fault diagnosis has made extraordinary advancements currently. Nonetheless, few works tackle class-incremental learning for fault diagnosis under limited fault data, i.e., imbalanced and long-tailed fault diagnosis, which brings about various notable challenges. Initially, it is difficult to extract discriminative features from limited fault data. Moreover, a well-trained model must be retrained from scratch to classify the samples from new classes, thus causing a high computational burden and time consumption. Furthermore, the model may suffer from catastrophic forgetting when trained incrementally. Finally, the model decision is biased toward the new classes due to the class imbalance. The problems can consequently lead to performance degradation of fault diagnosis models. Accordingly, we introduce a supervised contrastive knowledge distillation for incremental fault diagnosis under limited fault data (SCLIFD) framework to address these issues, which extends the classical incremental classifier and representation learning (iCaRL) framework from three perspectives. Primarily, we adopt supervised contrastive knowledge distillation (KD) to enhance its representation learning capability under limited fault data. Moreover, we propose a novel prioritized exemplar selection method adaptive herding (AdaHerding) to restrict the increase of the computational burden, which is also combined with KD to alleviate catastrophic forgetting. Additionally, we adopt the cosine classifier to mitigate the adverse impact of class imbalance. We conduct extensive experiments on simulated and real-world industrial processes under different imbalance ratios. Experimental results show that our SCLIFD outperforms the existing methods by a large margin. | ['Weiming Shen', 'Hongwei Wang', 'Gongzhuang Peng', 'Mengxuan Li', 'Hanrong Zhang', 'Peng Peng'] | 2023-02-12 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [ 2.13904440e-01 -5.31751633e-01 -2.44543493e-01 -1.09125383e-01
-1.95803270e-01 5.18164597e-02 1.39114141e-01 2.44949520e-01
3.62760462e-02 8.46175075e-01 -3.99838477e-01 -1.30128577e-01
-6.66717768e-01 -8.16830397e-01 -4.34740692e-01 -9.80687976e-01
8.46184641e-02 4.79281962e-01 3.18769217e-01 -7.25577846e-02
3.67761940e-01 3.93405825e-01 -1.73275375e+00 8.29367414e-02
1.48443246e+00 1.04687405e+00 3.27345431e-01 8.20343345e-02
-2.42551733e-02 8.79410863e-01 -8.41576695e-01 -2.42514908e-02
-4.74057086e-02 -4.37842965e-01 -6.29549801e-01 3.33414763e-01
-3.48588735e-01 -4.35028911e-01 -3.31796229e-01 1.09335864e+00
6.13074780e-01 1.16116196e-01 4.78616744e-01 -1.52821875e+00
-4.19552267e-01 4.52460945e-01 -6.62232399e-01 3.59842747e-01
1.80634350e-01 1.44751877e-01 4.80148971e-01 -1.06762290e+00
2.36079842e-01 1.11812174e+00 7.55273879e-01 3.26208234e-01
-5.67756891e-01 -7.47882903e-01 5.34024775e-01 6.27028227e-01
-1.41232061e+00 -2.53631264e-01 9.88988757e-01 -2.95308143e-01
6.65378928e-01 -3.78929824e-02 6.06305718e-01 7.59930253e-01
2.32574388e-01 8.46171141e-01 9.53427792e-01 -2.81144649e-01
5.57550788e-01 -1.04860395e-01 9.13379863e-02 6.06860578e-01
8.31150353e-01 -9.32538658e-02 -3.41288358e-01 -2.22516373e-01
5.37265956e-01 7.86199450e-01 -4.16617215e-01 -2.75688887e-01
-9.50816810e-01 4.34499592e-01 3.29055727e-01 2.16373861e-01
-5.01236022e-01 -3.70714128e-01 7.55334914e-01 7.73999333e-01
4.91117626e-01 2.07409129e-01 -8.75067592e-01 -7.40705878e-02
-6.24219418e-01 1.40765399e-01 5.30717492e-01 7.50370204e-01
8.18565428e-01 2.98375338e-01 4.20763679e-02 1.01458526e+00
-1.36656528e-02 3.53460014e-01 9.33212578e-01 -3.62874150e-01
5.18894374e-01 9.20405746e-01 5.83725646e-02 -1.05776489e+00
-3.81018668e-01 -9.66294646e-01 -1.14247167e+00 6.32731915e-02
-9.78341848e-02 -1.96925163e-01 -8.19014192e-01 1.26665962e+00
4.45651472e-01 5.41190267e-01 1.81688353e-01 7.04676151e-01
2.66119659e-01 5.46662867e-01 -1.34268209e-01 -8.16443145e-01
9.74437654e-01 -9.43562448e-01 -7.35868216e-01 -1.31425038e-01
7.95688272e-01 -4.62047905e-01 8.12739432e-01 7.04463005e-01
-5.69977224e-01 -6.11852467e-01 -1.26368463e+00 7.65461385e-01
-1.03995755e-01 1.04520194e-01 4.82440919e-01 2.42806211e-01
-2.00936183e-01 7.10120559e-01 -8.16396058e-01 -8.42161179e-02
5.72489917e-01 1.65172517e-01 -1.00304171e-01 -8.08457136e-01
-1.29707766e+00 6.68183744e-01 4.98144060e-01 2.07188204e-01
-8.56538236e-01 -6.86094105e-01 -5.59291363e-01 -1.22808553e-01
7.16912508e-01 -6.47558570e-01 1.20914292e+00 -9.57900524e-01
-1.35261810e+00 -9.95119810e-02 -3.37140784e-02 -3.96916598e-01
6.45495057e-01 -3.99510026e-01 -7.24957228e-01 -2.21640803e-02
1.09813154e-01 -3.09686363e-01 9.38590765e-01 -1.10965133e+00
-9.12538409e-01 -4.29422617e-01 -1.42185390e-01 3.95781457e-01
-6.24025524e-01 -5.88521004e-01 8.02031904e-02 -8.01625967e-01
4.60309297e-01 -6.28750265e-01 -1.51318327e-01 -4.31277603e-01
-2.65775293e-01 -4.23135698e-01 1.22310376e+00 -3.63179237e-01
1.58102119e+00 -2.09813285e+00 -3.53215858e-02 -1.32994160e-01
-2.93272957e-02 4.64567989e-01 1.69146016e-01 2.69521564e-01
-2.17503920e-01 -2.06907615e-01 -4.06460166e-01 1.68098152e-01
-5.05971789e-01 3.81532282e-01 -2.20796227e-01 3.46853971e-01
3.88750017e-01 3.93520832e-01 -1.17394078e+00 -3.06700528e-01
1.31719664e-01 5.53967478e-03 -4.12864298e-01 2.47244537e-01
-1.26740098e-01 2.31925920e-01 -5.72771966e-01 1.03910315e+00
9.92435038e-01 -1.75838068e-01 -4.34754007e-02 -1.70380369e-01
1.05445191e-01 -3.91211808e-02 -1.24943888e+00 1.37643671e+00
-5.22160470e-01 -2.46249035e-01 -2.13943034e-01 -1.53705227e+00
1.02303267e+00 3.34734350e-01 5.59967339e-01 -4.68983829e-01
4.89381477e-02 5.67519367e-01 -4.93928753e-02 -8.11059296e-01
1.74032245e-03 -3.26207221e-01 -9.64956172e-03 2.28102237e-01
-1.32121921e-01 1.32228151e-01 1.31705226e-02 -2.74875790e-01
1.23335934e+00 -1.73388079e-01 2.83153504e-01 -8.23850408e-02
7.66096592e-01 5.21918759e-02 1.26787698e+00 4.60960925e-01
-2.46050522e-01 2.98034936e-01 1.60262749e-01 -6.00046337e-01
-3.91197652e-01 -7.35163510e-01 -1.41103014e-01 7.81892776e-01
4.47140753e-01 -1.24164909e-01 -3.10677081e-01 -1.14608562e+00
3.36425424e-01 5.56487024e-01 -3.03053081e-01 -8.55135024e-01
-5.57034969e-01 -1.26130152e+00 3.41763765e-01 4.86164868e-01
7.34270513e-01 -8.24926913e-01 -3.94874930e-01 4.29632097e-01
5.19213974e-02 -4.64442641e-01 2.39296574e-02 4.02839959e-01
-1.17676055e+00 -1.44941568e+00 -7.78851748e-01 -9.44802999e-01
8.58568192e-01 5.43265343e-01 7.27935374e-01 3.21217865e-01
-3.04907918e-01 -5.47641553e-02 -6.78363562e-01 -3.67430657e-01
-1.50188565e-01 6.96961433e-02 5.13950765e-01 2.26165019e-02
2.81105459e-01 -4.81194794e-01 -6.45028412e-01 4.64171022e-01
-1.01566374e+00 -3.41523468e-01 9.02179420e-01 1.41006958e+00
5.27053297e-01 1.07804918e+00 1.37442708e+00 -9.47616458e-01
5.82331061e-01 -9.01578844e-01 -2.79471457e-01 3.56188923e-01
-1.13872075e+00 -2.39953652e-01 9.09784973e-01 -6.82665646e-01
-1.23806238e+00 -2.06669882e-01 1.42259477e-02 -6.59601271e-01
7.05407262e-02 8.22854936e-01 -3.15363854e-01 1.39770880e-01
3.45437229e-01 4.41717207e-01 6.82896525e-02 -5.78482449e-01
-6.99677914e-02 9.15379226e-01 3.51717740e-01 -4.41148043e-01
8.09429228e-01 1.66350916e-01 -3.93035375e-02 -4.74686772e-01
-7.68884301e-01 -3.15970451e-01 -4.27562296e-01 -5.29757403e-02
-3.92262973e-02 -1.01300001e+00 -3.58955562e-01 9.84542727e-01
-7.37071812e-01 4.66043353e-02 -3.38628173e-01 5.32696605e-01
-2.62297988e-01 4.02541876e-01 -6.34180248e-01 -7.95217276e-01
-4.53335732e-01 -9.19996858e-01 5.48195958e-01 4.11608070e-01
2.09354073e-01 -6.48805022e-01 -1.64881811e-01 9.36254337e-02
3.55349362e-01 2.44823679e-01 1.11615694e+00 -6.85262620e-01
-1.80679768e-01 -3.53422374e-01 1.41273607e-02 5.69191039e-01
6.79678202e-01 -1.90152332e-01 -5.45162678e-01 -6.66013956e-01
3.40019017e-01 -3.88212383e-01 9.09717560e-01 -5.55700213e-02
1.02424133e+00 -2.87001282e-01 -5.76914370e-01 3.37893456e-01
1.18697238e+00 5.46454012e-01 2.51993954e-01 4.30444747e-01
7.67245352e-01 2.51522750e-01 1.31920373e+00 8.33431780e-01
2.00381294e-01 2.31453896e-01 4.34963912e-01 2.41269484e-01
-7.11896643e-02 -1.54105872e-01 2.11504176e-01 1.34951746e+00
2.35871181e-01 -1.74207568e-01 -7.59819269e-01 6.35401964e-01
-1.95193636e+00 -6.00966692e-01 2.06339836e-01 2.12415600e+00
8.52883339e-01 3.78630996e-01 -2.84172207e-01 8.82931054e-01
9.37214196e-01 -1.95082843e-01 -1.18418419e+00 1.22125871e-01
-2.32450608e-02 -6.08112626e-02 2.47462129e-04 -6.52104244e-02
-1.05094445e+00 4.22112465e-01 4.86481476e+00 9.32779372e-01
-1.24451387e+00 5.08464873e-02 5.74917734e-01 6.13464154e-02
3.38286273e-02 -1.54728517e-02 -5.75941920e-01 8.46577883e-01
6.37275219e-01 -3.34821850e-01 1.29500642e-01 1.18795383e+00
-7.14908466e-02 -7.98649788e-02 -8.21613848e-01 9.55361068e-01
1.61133707e-01 -7.97728777e-01 1.46125525e-01 -5.48665583e-01
9.12460029e-01 -3.11557531e-01 -1.06891803e-01 5.76128542e-01
4.56910655e-02 -4.43700552e-01 3.55066925e-01 6.35702968e-01
6.92957103e-01 -1.11022711e+00 1.13889146e+00 5.41066706e-01
-1.12702036e+00 -7.47098386e-01 -6.03368759e-01 -3.19501013e-01
-6.99218065e-02 1.40337336e+00 -7.34122872e-01 1.09227836e+00
6.81375682e-01 1.04773808e+00 -4.98556584e-01 1.31772184e+00
-8.06285366e-02 6.03951275e-01 -3.90957370e-02 2.36713216e-01
-2.80197978e-01 2.77324140e-01 4.73431051e-01 5.08965194e-01
6.36457264e-01 -1.55293956e-01 6.10981405e-01 9.98068154e-02
-3.88935618e-02 -6.13632202e-02 -4.65040863e-01 2.16422185e-01
9.70803082e-01 7.75306940e-01 -6.31363750e-01 -4.31839228e-01
-1.14010930e-01 1.10165644e+00 1.41598195e-01 1.63612664e-01
-4.68209207e-01 -7.61445165e-01 4.73820508e-01 1.10240452e-01
3.15988630e-01 1.53182954e-01 -6.29020855e-02 -1.24415493e+00
1.21136263e-01 -9.23033476e-01 5.49488664e-01 -4.55727011e-01
-1.67534661e+00 4.16672051e-01 -1.87816873e-01 -1.64291608e+00
-7.71984160e-02 -1.78898811e-01 -6.14945650e-01 4.09890711e-01
-1.73327613e+00 -6.14015162e-01 -4.81128603e-01 4.82488334e-01
8.19081306e-01 -1.90902755e-01 5.90501964e-01 6.21930778e-01
-1.01936126e+00 5.92109442e-01 4.59312886e-01 -2.73005158e-01
9.12176609e-01 -8.13428283e-01 -1.39866263e-01 6.15585148e-01
-5.74288785e-01 4.53345209e-01 4.83228803e-01 -9.52682078e-01
-1.53401935e+00 -1.64846957e+00 4.83531237e-01 1.00306109e-01
3.97226214e-01 1.51594549e-01 -1.24915981e+00 3.07576448e-01
-4.33373690e-01 3.07308227e-01 4.67037588e-01 -1.35465682e-01
-1.88792571e-01 -6.46913886e-01 -1.30189764e+00 1.93942875e-01
9.65155900e-01 -2.29852021e-01 -7.05817819e-01 4.94020849e-01
8.68374407e-01 -1.68661848e-01 -9.01222885e-01 1.04691982e+00
2.63346910e-01 -7.41302311e-01 6.72655582e-01 -2.39312112e-01
1.57203555e-01 -6.40264273e-01 2.18222946e-01 -1.57279932e+00
-5.21273494e-01 -1.37066439e-01 -4.58752215e-01 1.32630455e+00
-6.67333379e-02 -1.05819058e+00 4.76447523e-01 -1.09084152e-01
-3.38041723e-01 -1.26821923e+00 -9.17414725e-01 -1.03913546e+00
-1.31869659e-01 -2.76105106e-02 9.40362155e-01 1.17884791e+00
1.09159820e-01 2.72135377e-01 -2.92995274e-01 1.75875992e-01
3.55661362e-01 3.66448522e-01 4.27633196e-01 -1.47897363e+00
-1.67905673e-01 -5.50156645e-02 -4.64573592e-01 -8.22733283e-01
-1.29332721e-01 -6.66850448e-01 2.17527404e-01 -1.29189408e+00
7.55215762e-03 -9.29825127e-01 -9.57190037e-01 5.16955912e-01
-6.09384358e-01 -1.59624383e-01 -3.84662092e-01 6.76305830e-01
-7.62577236e-01 8.04587066e-01 1.27803075e+00 -2.24980250e-01
-1.87209785e-01 2.38060147e-01 -6.24724448e-01 6.60553753e-01
8.56672287e-01 -5.38780868e-01 -8.46978903e-01 -3.03604871e-01
9.05490741e-02 9.37478095e-02 -1.14307806e-01 -1.38145101e+00
3.55166674e-01 -2.27519169e-01 4.66613829e-01 -6.12000823e-01
-2.32863888e-01 -7.24622548e-01 4.12586294e-02 1.08746004e+00
2.47310907e-01 1.06277838e-01 -4.74450663e-02 9.97114718e-01
-3.41747016e-01 -2.95440674e-01 5.62395453e-01 -6.21657595e-02
-6.38194859e-01 3.83795232e-01 -2.31817365e-01 -1.15107343e-01
1.23476529e+00 1.82442088e-02 -4.11749423e-01 2.48746037e-01
-4.15398628e-01 4.01185393e-01 4.27691191e-01 4.77041692e-01
8.08776677e-01 -1.39682007e+00 -6.24166012e-01 4.98872638e-01
2.95160890e-01 3.79802167e-01 6.56031907e-01 9.30696547e-01
-2.62153655e-01 1.84971735e-01 9.75962821e-03 -4.20732588e-01
-8.72514486e-01 8.30440402e-01 7.92468339e-02 -3.21965754e-01
-5.70194244e-01 8.12876940e-01 -1.85918510e-01 -3.06538522e-01
9.97749344e-02 -2.31415212e-01 -1.86401010e-01 2.51508117e-01
5.56332111e-01 7.31725633e-01 5.39418638e-01 -2.06579670e-01
-5.49700975e-01 3.77868772e-01 -5.00140131e-01 7.50425041e-01
1.29912627e+00 -8.34929124e-02 -1.49402574e-01 6.05969310e-01
8.22327077e-01 -3.49822342e-01 -1.34425104e+00 -4.13019598e-01
-1.98170748e-02 -4.15271133e-01 3.30785252e-02 -8.47235799e-01
-1.13488257e+00 6.27664089e-01 6.50204420e-01 2.05576062e-01
1.38136196e+00 -5.09980857e-01 9.74044085e-01 6.95908070e-01
6.93682432e-01 -1.05884874e+00 2.97511816e-01 4.58139807e-01
5.97618878e-01 -1.00469375e+00 2.45284334e-01 -2.60573208e-01
-4.26420271e-01 1.04474437e+00 7.87760735e-01 -2.08762318e-01
8.57721567e-01 1.41914263e-01 -1.53919980e-01 -8.36585928e-03
-8.37076306e-01 1.78655356e-01 -3.32688749e-01 4.88941729e-01
-3.81234474e-02 3.10727730e-02 -3.74808609e-01 9.40861940e-01
2.49668494e-01 2.70631790e-01 3.04902226e-01 1.47850490e+00
-6.69859767e-01 -1.03410852e+00 -3.47967923e-01 8.01315725e-01
-2.74146974e-01 1.90112501e-01 7.36016259e-02 3.75712961e-01
5.84907293e-01 1.06582558e+00 -1.70235515e-01 -7.87493229e-01
4.34248179e-01 1.62461847e-01 2.67267376e-01 -6.38072371e-01
-1.62257999e-01 -2.37451136e-01 -3.26986313e-01 1.31476410e-02
-1.15497492e-01 -5.99454999e-01 -1.28359282e+00 5.52172512e-02
-1.03396380e+00 4.02030349e-01 2.34004006e-01 9.82299507e-01
5.09192884e-01 1.02612948e+00 1.15078688e+00 -4.74925607e-01
-9.01522577e-01 -1.15894771e+00 -6.95688367e-01 2.40932792e-01
2.16690615e-01 -1.19956541e+00 -5.91047823e-01 -1.40402019e-01] | [7.3403639793396, 2.307368755340576] |
865739ac-1abc-418f-9db6-066e9f5fea60 | phase-retrieval-via-non-rigid-image | 2306.15701 | null | https://arxiv.org/abs/2306.15701v1 | https://arxiv.org/pdf/2306.15701v1.pdf | Phase retrieval via non-rigid image registration | Phase retrieval is the numerical procedure of recovering a complex-valued signal from knowledge about its amplitude and some additional information. Here, an indirect registration procedure, based on the large deformation diffeomorphic metric mapping (LDDMM) formalism, is investigated as a phase retrieval method for coherent diffractive imaging. The method attempts to find a deformation which transforms an initial, template image to match an unknown target image by comparing the diffraction pattern to the data. The exterior calculus framework is used to treat different types of deformations in a unified and coordinate-free way. The algorithm performance with respect to measurement noise, image topology, and particular action are explored through numerical examples. | ['Erik Malm'] | 2023-06-26 | null | null | null | null | ['image-registration', 'retrieval'] | ['computer-vision', 'methodology'] | [ 4.98024911e-01 1.19498610e-01 4.24452811e-01 -3.10838014e-01
-5.94277799e-01 -7.08572716e-02 6.49024367e-01 -5.88946879e-01
-6.47624195e-01 5.95835626e-01 -2.62065860e-03 2.79473782e-01
-7.41555989e-01 -8.08300614e-01 -1.39450133e-01 -1.15263319e+00
-3.19144577e-02 7.37467468e-01 2.01329261e-01 -9.36510973e-03
4.31045264e-01 7.03700185e-01 -1.11510563e+00 -2.54507542e-01
5.17294526e-01 5.19576430e-01 3.69069934e-01 2.68075138e-01
2.91611582e-01 4.44649339e-01 -3.17925006e-01 1.23187095e-01
2.71047354e-01 -7.01470912e-01 -1.08458829e+00 2.32937768e-01
6.58664107e-02 1.56480148e-02 -3.91213566e-01 1.30597889e+00
4.72101986e-01 -3.31826583e-02 9.49690819e-01 -5.50681710e-01
-4.99047190e-01 7.76358545e-02 -7.79544935e-02 2.48375922e-01
5.75784147e-01 -1.74981847e-01 2.85721898e-01 -1.07704532e+00
1.36874187e+00 9.16265368e-01 2.57663369e-01 4.63374913e-01
-1.37711513e+00 2.34889295e-02 -8.57829690e-01 4.74027514e-01
-1.58121371e+00 -2.65885890e-01 1.23105299e+00 -5.46260118e-01
5.01848817e-01 3.01230907e-01 4.75131303e-01 4.26536977e-01
7.72528231e-01 -2.14388788e-01 1.63562155e+00 -8.38649690e-01
2.91711271e-01 -2.02556238e-01 2.71782339e-01 7.08362401e-01
2.99345762e-01 2.99210221e-01 -2.40771890e-01 -2.64191031e-01
6.15945578e-01 -4.32295620e-01 -8.07092249e-01 -6.12888813e-01
-1.24446452e+00 4.99806195e-01 2.16657355e-01 9.40336049e-01
-6.32223666e-01 -3.35048467e-01 -3.85792106e-01 4.18200284e-01
5.44835806e-01 7.28399098e-01 1.11712769e-01 1.93569511e-01
-6.85401082e-01 1.41621441e-01 8.21069002e-01 3.82366240e-01
9.84590709e-01 -9.71751846e-03 -5.03791450e-03 1.27811491e-01
5.40389001e-01 8.62482667e-01 4.36560422e-01 -9.28003490e-01
-1.39190748e-01 3.89249831e-01 4.17598374e-02 -9.74146783e-01
-4.33448315e-01 -3.72201577e-02 -6.72871709e-01 6.08879805e-01
4.58095551e-01 1.78936526e-01 -6.62925601e-01 1.54405856e+00
4.14738357e-01 3.18081439e-01 2.02967927e-01 9.36301172e-01
6.36371493e-01 2.34440371e-01 -3.83056164e-01 -7.79219627e-01
1.07055748e+00 1.38016090e-01 -8.77890050e-01 1.64465129e-01
3.17090243e-01 -1.13643157e+00 3.09254199e-01 2.80317664e-01
-1.24781621e+00 -5.98714501e-02 -1.12657547e+00 1.67428732e-01
1.03850223e-01 -1.16830483e-01 -1.22740313e-01 3.78054351e-01
-9.76974487e-01 1.04540503e+00 -9.83861983e-01 -1.36989415e-01
1.15557380e-01 3.76726955e-01 -5.63391030e-01 -1.59122646e-01
-8.51271451e-01 1.22912180e+00 1.20514162e-01 3.45350385e-01
1.97375622e-02 -6.61581159e-01 -5.20393848e-01 -4.09754276e-01
-3.08382511e-01 -7.54408300e-01 6.66114450e-01 -3.96688044e-01
-1.73116577e+00 1.32700694e+00 -1.05442069e-01 -3.07627227e-02
3.19217950e-01 6.39013946e-01 -4.52615470e-01 6.89308822e-01
-1.37078241e-01 -1.10905714e-01 1.02966475e+00 -8.39637935e-01
3.63709062e-01 -7.64979601e-01 -4.90169883e-01 1.35039926e-01
3.79190773e-01 5.10818884e-02 1.22820795e-01 -2.82100141e-01
1.08917165e+00 -9.87885475e-01 -1.50207132e-01 -1.36750326e-01
-1.34373814e-01 3.54788810e-01 8.89462113e-01 -6.98523998e-01
6.76012993e-01 -1.95964146e+00 7.09987521e-01 6.59205317e-01
3.46066833e-01 7.46288002e-02 8.98621008e-02 3.46795589e-01
-2.64058888e-01 -5.06711602e-01 -5.08219957e-01 1.50898010e-01
-3.70649457e-01 4.88403477e-02 2.99559832e-01 1.29029679e+00
7.79781342e-02 9.11434889e-01 -5.62674880e-01 -4.60730702e-01
1.41000539e-01 6.34669185e-01 -5.05893648e-01 2.01707199e-01
2.01026827e-01 1.12358487e+00 -6.42095864e-01 4.24936742e-01
9.96662021e-01 -6.55469298e-02 1.45337150e-01 -5.08549690e-01
-4.89060730e-01 -7.28851631e-02 -1.23481798e+00 1.53917277e+00
7.77010322e-02 6.27007723e-01 3.76359642e-01 -1.32898724e+00
1.05487251e+00 6.13864779e-01 9.16928887e-01 -9.38340187e-01
2.88138539e-01 5.21486461e-01 1.09630764e-01 -9.33754861e-01
1.03081651e-01 -5.15523911e-01 3.49388868e-01 5.85324347e-01
1.47909382e-02 -2.50119179e-01 -7.92817771e-02 -2.71524221e-01
1.22646260e+00 1.08494662e-01 3.09655279e-01 -7.15251029e-01
7.25164592e-01 -1.12216741e-01 1.95947587e-01 1.65120155e-01
1.39494523e-01 6.68721378e-01 9.26800966e-02 -4.60620344e-01
-1.09490669e+00 -1.07495701e+00 -7.15746224e-01 -5.47244489e-01
3.39436829e-01 2.87186325e-01 -9.43623781e-01 2.08642170e-01
-4.03610796e-01 1.40389323e-01 -3.39255959e-01 -2.64849216e-01
-1.04969537e+00 -1.12790537e+00 -1.08971067e-01 -4.53646839e-01
5.14862895e-01 -1.24173856e+00 -7.64847696e-01 3.44128311e-01
-3.50585818e-01 -1.00219262e+00 1.27968471e-02 -2.80909419e-01
-1.00340259e+00 -1.44915998e+00 -6.38493240e-01 -7.47447252e-01
8.32444727e-01 1.45435855e-02 8.35864782e-01 -1.16833441e-01
-6.08899653e-01 8.11674416e-01 -9.97173339e-02 1.37581021e-01
-9.46492493e-01 -5.56125283e-01 6.79507181e-02 1.89504668e-01
3.32071483e-01 -9.52071905e-01 -4.72621173e-01 2.05565587e-01
-9.39762115e-01 -2.70046413e-01 5.76708257e-01 6.60669088e-01
6.90282106e-01 -3.64017151e-02 8.52848962e-02 -6.02685034e-01
5.36580026e-01 -7.86508471e-02 -5.90040386e-01 5.57369813e-02
-8.73739123e-01 2.94273823e-01 -1.91516392e-02 -3.91107857e-01
-1.04308295e+00 -2.73341034e-02 -9.33780968e-02 -3.25634122e-01
-5.80568425e-02 4.38899845e-01 -2.71761447e-01 -9.90886569e-01
8.22830975e-01 4.92471814e-01 5.32284617e-01 -5.60930073e-01
-9.63631421e-02 3.74061674e-01 8.13226879e-01 -5.92039764e-01
1.09909248e+00 6.53822064e-01 7.01843441e-01 -1.04603040e+00
-2.33210534e-01 -4.45483953e-01 -9.99881148e-01 -3.02100718e-01
9.61744785e-01 -2.80881673e-01 -5.40275812e-01 8.55650425e-01
-1.34657955e+00 7.62409046e-02 -4.09666985e-01 1.17105210e+00
-8.79944444e-01 6.45398796e-01 -4.13406253e-01 -3.85217011e-01
-3.30652297e-01 -1.18883419e+00 7.84810483e-01 -3.53912860e-02
5.90178668e-02 -1.03134966e+00 3.56518745e-01 1.36143193e-01
4.39639419e-01 4.44001287e-01 8.17216873e-01 -2.69964576e-01
-1.08123350e+00 -2.01401219e-01 1.60498470e-01 2.76663870e-01
1.33494377e-01 -5.04400074e-01 -7.61274219e-01 -3.41273814e-01
1.12392211e+00 4.50993627e-01 2.89492458e-01 4.40110534e-01
3.35004330e-01 -1.55025452e-01 -2.30654866e-01 6.62283123e-01
1.86789787e+00 3.31663251e-01 1.17936921e+00 2.70001620e-01
4.34187323e-01 3.31475705e-01 5.11158407e-01 1.98750451e-01
-3.16016018e-01 9.76870000e-01 2.09333941e-01 9.99921039e-02
-3.14619511e-01 4.86406624e-01 -2.56039687e-02 9.92875636e-01
-5.03882527e-01 2.22474679e-01 -6.30484521e-01 2.11964734e-02
-1.25685704e+00 -1.32077277e+00 -4.26920980e-01 2.39863825e+00
6.61617398e-01 -1.48961008e-01 -7.27394223e-01 2.78067321e-01
5.95521748e-01 -1.37040883e-01 -1.00741431e-01 1.80472210e-01
-2.30774224e-01 7.91522920e-01 3.06973487e-01 8.47185314e-01
-6.62535489e-01 2.61496335e-01 6.78944063e+00 4.33869421e-01
-1.31721377e+00 1.93885431e-01 -1.37702912e-01 6.31681979e-01
-4.61675882e-01 2.04851165e-01 -3.37573290e-01 5.07281303e-01
8.24917436e-01 -3.17276478e-01 4.10386354e-01 -9.63880122e-03
3.28560919e-01 -3.51840109e-01 -8.54421914e-01 9.99746382e-01
1.41793713e-01 -1.46975434e+00 -1.51444748e-01 4.05235767e-01
3.70935231e-01 -7.25224540e-02 -2.61939198e-01 -5.06646454e-01
-6.42068267e-01 -5.06988168e-01 3.31540197e-01 1.41010082e+00
6.89743161e-01 -3.14699352e-01 5.81546187e-01 2.22091734e-01
-8.75143528e-01 5.28888941e-01 -1.23315044e-01 -4.04107198e-02
4.43640232e-01 6.35480821e-01 -5.30447543e-01 6.83368325e-01
3.07084352e-01 5.43577075e-01 -2.71548718e-01 1.04837978e+00
1.39869407e-01 2.01152146e-01 -3.11988950e-01 3.48168582e-01
-1.40287101e-01 -1.15901232e+00 1.16694641e+00 4.81494427e-01
6.97773397e-01 6.43065870e-01 -1.79500267e-01 1.12461185e+00
4.23274159e-01 1.05945393e-01 -6.98414147e-01 3.57449621e-01
1.12260863e-01 1.15818191e+00 -7.96073914e-01 1.83016956e-01
-1.20806672e-01 6.69766426e-01 -2.85693198e-01 5.58774889e-01
-2.65663207e-01 -4.95879613e-02 2.98362434e-01 6.17527246e-01
1.10695362e-01 -4.21605557e-01 2.27223366e-01 -1.15175438e+00
3.56533617e-01 -3.29813212e-01 -8.50083455e-02 -9.31552231e-01
-8.51744235e-01 5.33254504e-01 4.14105088e-01 -1.51498175e+00
-5.36063671e-01 -6.26633644e-01 -5.98556459e-01 1.13989997e+00
-1.35708714e+00 -8.05633307e-01 -1.86257884e-01 7.35632837e-01
-3.39305073e-01 -2.24507004e-01 9.85450506e-01 2.82694370e-01
1.72240257e-01 -1.30643144e-01 2.93319046e-01 -4.99005951e-02
5.50836146e-01 -1.05854428e+00 -4.18636143e-01 8.72076094e-01
-2.27525443e-01 5.37964463e-01 1.02243876e+00 -3.93606007e-01
-1.50425577e+00 -4.25403118e-01 1.15990329e+00 -2.04504758e-01
5.35547972e-01 1.51424393e-01 -1.08746254e+00 5.41352332e-01
1.18829131e-01 2.12025180e-01 2.61877418e-01 -7.30677307e-01
2.37489522e-01 5.66607155e-02 -1.63021219e+00 1.51633754e-01
7.14275479e-01 -4.54213679e-01 -8.43871057e-01 3.25532526e-01
3.30173552e-01 -5.67059636e-01 -1.49325764e+00 4.19721544e-01
2.42675141e-01 -8.90069783e-01 1.06875110e+00 1.12824803e-02
-1.03218324e-01 -2.65489638e-01 -1.26537874e-01 -1.15770960e+00
-3.60266894e-01 -1.11796021e+00 2.43968710e-01 6.90902233e-01
-2.26539463e-01 -1.07588708e+00 6.08295679e-01 4.78163421e-01
-2.49672398e-01 -4.54428077e-01 -1.41905403e+00 -8.04678917e-01
5.12864552e-02 1.08584724e-02 1.91737399e-01 9.13779199e-01
-2.26138178e-02 1.25426441e-01 -2.86138076e-02 3.64549935e-01
1.18752384e+00 1.55130014e-01 8.34678337e-02 -1.63521481e+00
-1.73393860e-01 -1.91333756e-01 -1.06014764e+00 -4.39843893e-01
2.32635751e-01 -1.08594263e+00 -7.06166103e-02 -1.01588392e+00
-7.02253133e-02 -4.63261932e-01 4.78785345e-03 -2.70115703e-01
4.83806729e-01 2.88483173e-01 -1.59842789e-01 7.28893697e-01
1.08723998e-01 3.94216359e-01 1.67788839e+00 1.26203477e-01
-3.45961899e-02 1.54856458e-01 -9.67868418e-02 6.08774722e-01
4.25364166e-01 -7.05835760e-01 -1.66580513e-01 1.48082733e-01
1.35367945e-01 3.73865187e-01 5.06617188e-01 -1.21795380e+00
3.80214870e-01 -2.56300881e-03 3.72171849e-02 -3.18304121e-01
4.46245879e-01 -8.74976397e-01 8.72882485e-01 6.53703034e-01
1.18032269e-01 -1.12819910e-01 -4.14195895e-01 3.93784314e-01
-3.58206093e-01 -7.84957469e-01 1.08162200e+00 -1.45462647e-01
-5.47302127e-01 3.49220395e-01 -2.44364396e-01 -5.21341059e-03
7.25793064e-01 -3.12979341e-01 -1.09554611e-01 1.19764291e-01
-1.22749698e+00 -7.59426475e-01 6.52468264e-01 -3.10779214e-01
7.31254458e-01 -1.48792005e+00 -7.41141319e-01 6.81239307e-01
-5.74183427e-02 -2.75091678e-01 1.98883995e-01 1.27590966e+00
-7.92014241e-01 2.59269983e-01 -3.25416654e-01 -7.03837872e-01
-1.31090987e+00 3.53656203e-01 7.96378493e-01 -2.94355869e-01
-1.02713871e+00 8.96272212e-02 -2.55771339e-01 -2.95496374e-01
-5.39916277e-01 -1.92863047e-01 -2.45476320e-01 -3.78963053e-01
5.28463244e-01 4.02706563e-01 3.93937439e-01 -1.07780898e+00
-2.72069246e-01 1.23425782e+00 5.69375694e-01 -2.97182769e-01
1.38136292e+00 -1.70079693e-01 -6.25617683e-01 8.77755806e-02
1.58443201e+00 -1.01614617e-01 -8.41993392e-01 -5.63975155e-01
1.53792262e-01 -4.20793355e-01 2.31235534e-01 -2.51182169e-01
-9.59603846e-01 2.98140556e-01 8.35505843e-01 1.08823024e-01
1.14988589e+00 2.67553836e-01 3.05614263e-01 5.07528365e-01
5.05839467e-01 -8.06505084e-01 -8.34761709e-02 1.22999191e-01
1.03089881e+00 -9.21157181e-01 3.05361599e-01 -5.48842371e-01
2.67052799e-01 1.35779774e+00 -2.19374448e-01 -3.94035876e-01
1.29192114e+00 3.17142665e-01 -6.38134554e-02 -5.96725583e-01
-6.87308386e-02 -6.55799638e-03 4.27986115e-01 7.28834569e-01
2.32931390e-01 -8.59751329e-02 -8.69308949e-01 -2.34489381e-01
7.39837885e-02 2.48057917e-01 6.69446945e-01 1.07672489e+00
-4.22041863e-01 -1.34413862e+00 -6.63588345e-01 8.12971517e-02
-3.09899062e-01 3.36380512e-01 4.92521711e-02 9.49068487e-01
-5.31855933e-02 4.96843040e-01 2.59145558e-01 1.30358756e-01
4.59030718e-01 6.78636059e-02 1.21283150e+00 -3.62377852e-01
4.95803505e-02 1.87063724e-01 -1.62087753e-01 -6.00699604e-01
-1.22300923e+00 -9.39995050e-01 -1.41235995e+00 3.06551112e-03
-8.26606601e-02 1.53684214e-01 8.46073151e-01 1.30852437e+00
2.37108935e-02 1.40303180e-01 6.91909790e-01 -9.51513410e-01
-6.53465867e-01 -8.05215418e-01 -1.01389849e+00 4.99044001e-01
1.55598208e-01 -7.84418285e-01 -6.13982677e-01 6.97263330e-02] | [12.907132148742676, -2.7212347984313965] |
b7547f5d-913b-40d9-89ab-c707609b9db9 | spottarget-rethinking-the-effect-of-target | 2306.00899 | null | https://arxiv.org/abs/2306.00899v1 | https://arxiv.org/pdf/2306.00899v1.pdf | SpotTarget: Rethinking the Effect of Target Edges for Link Prediction in Graph Neural Networks | Graph Neural Networks (GNNs) have demonstrated promising outcomes across various tasks, including node classification and link prediction. Despite their remarkable success in various high-impact applications, we have identified three common pitfalls in message passing for link prediction. Particularly, in prevalent GNN frameworks (e.g., DGL and PyTorch-Geometric), the target edges (i.e., the edges being predicted) consistently exist as message passing edges in the graph during training. Consequently, this results in overfitting and distribution shift, both of which adversely impact the generalizability to test the target edges. Additionally, during test time, the failure to exclude the test target edges leads to implicit test leakage caused by neighborhood aggregation. In this paper, we analyze these three pitfalls and investigate the impact of including or excluding target edges on the performance of nodes with varying degrees during training and test phases. Our theoretical and empirical analysis demonstrates that low-degree nodes are more susceptible to these pitfalls. These pitfalls can have detrimental consequences when GNNs are implemented in production systems. To systematically address these pitfalls, we propose SpotTarget, an effective and efficient GNN training framework. During training, SpotTarget leverages our insight regarding low-degree nodes and excludes train target edges connected to at least one low-degree node. During test time, it emulates real-world scenarios of GNN usage in production and excludes all test target edges. Our experiments conducted on diverse real-world datasets, demonstrate that SpotTarget significantly enhances GNNs, achieving up to a 15x increase in accuracy in sparse graphs. Furthermore, SpotTarget consistently and dramatically improves the performance for low-degree nodes in dense graphs. | ['Danai Koutra', 'Xiang Song', 'Wei Ai', 'Shengyi Qian', 'Vassilis N. Ioannidis', 'YuHang Zhou', 'Jing Zhu'] | 2023-06-01 | null | null | null | null | ['link-prediction'] | ['graphs'] | [-4.68945429e-02 1.93064645e-01 -4.91304100e-01 -7.61284083e-02
-8.71475264e-02 -4.96659279e-01 1.95664778e-01 4.41751838e-01
9.36251208e-02 8.24482024e-01 -4.27837372e-01 -9.50237811e-01
-2.85704762e-01 -1.32108676e+00 -8.82595301e-01 -2.58029252e-01
-8.33291233e-01 4.64032024e-01 5.35918832e-01 -2.61185337e-02
-1.12033226e-01 4.95232016e-01 -1.08571792e+00 -7.22892284e-02
7.70526767e-01 4.56072241e-01 -3.71440321e-01 6.43711388e-01
6.19837567e-02 7.75455415e-01 -9.10698235e-01 -5.63679755e-01
4.64876592e-01 -8.11701789e-02 -6.32416666e-01 -1.80377796e-01
5.18776715e-01 -4.71862614e-01 -7.07936883e-01 9.70955074e-01
4.21871513e-01 -2.87066132e-01 4.15793061e-01 -1.95052004e+00
-1.55402407e-01 1.17911053e+00 -8.28666091e-01 2.53772020e-01
7.66108334e-02 3.08486760e-01 1.16247904e+00 -5.04510820e-01
5.71386814e-01 1.03820562e+00 1.21974242e+00 -4.45019640e-02
-1.36186290e+00 -1.02183104e+00 1.65279973e-02 -2.41379723e-01
-1.56802630e+00 -3.74050230e-01 5.12199163e-01 -1.94795206e-01
8.56121004e-01 3.96298230e-01 5.00998557e-01 9.54190493e-01
3.93123686e-01 4.01704758e-01 4.28855479e-01 -3.77812088e-02
1.02270670e-01 -5.19766547e-02 2.71974593e-01 6.74974799e-01
9.71421957e-01 2.28401534e-02 -3.97728622e-01 -3.80638957e-01
7.73820162e-01 -7.39249438e-02 -2.42516577e-01 -2.98767507e-01
-6.40631378e-01 7.98778713e-01 7.96243012e-01 1.04398899e-01
-2.34526351e-01 4.59475130e-01 4.84920293e-01 4.37380701e-01
3.80626321e-01 1.15238272e-01 -4.64832485e-01 4.39911745e-02
-8.89741719e-01 -1.50925696e-01 1.34788620e+00 1.20048308e+00
8.88446152e-01 1.60799265e-01 -2.69090980e-01 6.18089318e-01
3.57377589e-01 3.80636364e-01 -1.78365439e-01 -3.25476825e-01
7.29614735e-01 7.89770067e-01 -6.56581283e-01 -1.48154414e+00
-7.21966147e-01 -1.12744486e+00 -1.09635556e+00 -1.74673826e-01
3.91326308e-01 -5.32204628e-01 -8.84624958e-01 1.71161175e+00
3.43207717e-01 6.47817492e-01 -1.91138074e-01 5.29048860e-01
8.36017430e-01 3.94076884e-01 1.61251619e-01 9.63491946e-02
7.96603739e-01 -8.82241547e-01 -1.93272099e-01 -4.75369453e-01
1.25869155e+00 -5.79777420e-01 6.87095404e-01 5.43176644e-02
-8.29663396e-01 -1.16632447e-01 -1.07702529e+00 5.15222907e-01
-2.58717000e-01 -3.15694451e-01 8.01085174e-01 9.05083895e-01
-1.38129759e+00 8.81987512e-01 -7.57732689e-01 -4.20370698e-01
4.54426110e-01 5.88078141e-01 -3.69576365e-01 -2.63395876e-01
-1.01430988e+00 3.26914489e-01 3.14941078e-01 2.66268924e-02
-8.78477454e-01 -9.41127539e-01 -8.41880202e-01 4.49985534e-01
5.55763662e-01 -4.70556796e-01 9.64942932e-01 -5.36513031e-01
-5.22877097e-01 1.30107373e-01 1.85298279e-01 -5.05551219e-01
4.04727340e-01 2.48648942e-01 -6.32389963e-01 -1.15525074e-01
1.76141374e-02 2.55455375e-01 1.98512867e-01 -1.10602725e+00
-3.84499431e-01 -3.45858000e-02 7.14585111e-02 -2.02439412e-01
-5.40835142e-01 -5.21833539e-01 -7.29217649e-01 -4.49709594e-01
1.02234669e-01 -8.07361245e-01 -2.13457838e-01 -3.18249375e-01
-9.74090934e-01 -1.78507604e-02 9.87568796e-01 -2.17184424e-01
1.60646856e+00 -1.93012321e+00 -6.31537020e-01 1.14730525e+00
8.46934915e-01 2.71525294e-01 -4.17665690e-01 5.03757358e-01
4.74855714e-02 4.97192830e-01 2.95105785e-01 3.91295329e-02
-2.07388952e-01 2.43913338e-01 8.68649557e-02 5.24798214e-01
5.33177666e-02 8.98316979e-01 -9.69345331e-01 -5.00070333e-01
-1.75141260e-01 3.66060287e-01 -7.22371280e-01 -2.56126344e-01
-1.15610629e-01 -5.13231903e-02 -5.05520344e-01 8.22252214e-01
7.81797528e-01 -8.59683216e-01 6.90377772e-01 -6.06613932e-03
4.64916915e-01 5.26671588e-01 -1.09866214e+00 6.03258491e-01
-1.73162699e-01 8.14318776e-01 3.79896313e-02 -7.40407705e-01
9.09977436e-01 -5.75603805e-02 4.10470068e-01 -5.81816256e-01
-4.83127311e-02 2.31968179e-01 3.38782400e-01 -1.18438102e-01
4.18381959e-01 4.17515874e-01 2.18897179e-01 5.57822645e-01
-1.81094393e-01 5.35742998e-01 4.98519391e-01 5.61941326e-01
2.10283875e+00 -7.30922520e-01 -2.58004498e-02 -2.10953012e-01
-7.25093037e-02 -1.16250888e-01 5.29639721e-01 1.16909766e+00
-6.58624619e-02 2.24998474e-01 9.75374997e-01 -1.78700104e-01
-7.66063154e-01 -1.02993345e+00 3.49463448e-02 8.78632963e-01
1.31970420e-01 -7.76602924e-01 -4.27416682e-01 -9.80635583e-01
4.16284293e-01 4.70262825e-01 -2.79774427e-01 -2.95927197e-01
-3.82505089e-01 -8.57976913e-01 8.72670650e-01 4.28352416e-01
2.59042054e-01 -7.78457522e-01 1.34147510e-01 3.78652215e-01
2.71596342e-01 -1.06924701e+00 -2.80211151e-01 1.62388861e-01
-9.82992649e-01 -1.44310153e+00 -1.37643322e-01 -6.88004017e-01
9.71387923e-01 4.93077248e-01 1.55749762e+00 8.66320372e-01
-1.76951811e-01 2.24265039e-01 -2.59517640e-01 -1.91792488e-01
-4.83848453e-01 4.37365144e-01 -1.66448280e-01 -3.74611735e-01
2.85716653e-01 -9.74736214e-01 -5.02756059e-01 5.05586028e-01
-6.70489788e-01 -2.74219960e-01 7.57036030e-01 5.54138303e-01
1.76734135e-01 3.79322052e-01 6.76634967e-01 -1.50855744e+00
6.93167210e-01 -9.29460168e-01 -7.58161604e-01 1.77173272e-01
-8.72111499e-01 -1.39702767e-01 6.69716656e-01 -4.35691267e-01
-1.66563928e-01 -5.30195177e-01 -8.06800202e-02 -1.75944090e-01
3.25890958e-01 8.99671197e-01 1.37171656e-01 -5.03557205e-01
7.65931427e-01 -3.25005710e-01 -1.68573540e-02 -1.05038509e-01
-3.74570012e-01 3.62789840e-01 -5.37986122e-03 -4.06802446e-01
1.07202101e+00 5.74150495e-02 3.22720081e-01 -7.20672607e-01
-3.96726012e-01 -2.09634066e-01 1.19939834e-01 -3.83791804e-01
-1.10242873e-01 -8.84405971e-01 -8.76409888e-01 3.35569859e-01
-8.33946764e-01 -7.87071764e-01 1.26031071e-01 2.79972255e-01
4.29041445e-01 3.12571377e-01 -8.24540019e-01 -5.67534924e-01
-4.60536301e-01 -1.08353364e+00 4.34943497e-01 2.42785946e-01
-3.21491092e-01 -1.30046761e+00 -3.43431085e-01 -3.10773253e-01
6.46499574e-01 2.99713701e-01 1.03515756e+00 -1.10370731e+00
-7.79607773e-01 -2.14282617e-01 -6.00247085e-01 3.94593664e-02
1.33383468e-01 3.08288217e-01 -5.04918396e-01 -7.59284317e-01
-8.83601308e-01 -1.10419132e-01 5.51485598e-01 2.77055770e-01
1.15455735e+00 -3.12174380e-01 -8.88671756e-01 6.85939133e-01
1.53444231e+00 -8.66184756e-02 4.90514219e-01 9.74435434e-02
8.07790160e-01 2.29245141e-01 2.61968285e-01 3.52579892e-01
4.01478380e-01 2.59194344e-01 6.14570081e-01 -3.69600505e-01
-2.63804287e-01 -4.92580414e-01 2.92243063e-01 7.04681218e-01
4.34169680e-01 -8.73135448e-01 -1.21150255e+00 4.72209841e-01
-1.53381944e+00 -5.05849779e-01 -5.79313874e-01 2.32376170e+00
7.18492091e-01 6.26371324e-01 -3.24019976e-02 1.48642749e-01
9.70571935e-01 5.68499789e-03 -6.47553563e-01 -8.40882435e-02
6.58736452e-02 1.73608199e-01 9.44715202e-01 3.93465817e-01
-6.59084558e-01 5.75554192e-01 6.23209381e+00 8.71755302e-01
-1.03245771e+00 -9.24414247e-02 7.69548237e-01 1.16312997e-02
-4.07297522e-01 1.50353685e-01 -9.83406246e-01 4.83963341e-01
1.18525195e+00 -4.25621003e-01 2.79896498e-01 7.66966760e-01
-2.88317315e-02 8.13847110e-02 -1.20689702e+00 5.25085449e-01
-2.72846043e-01 -1.37751770e+00 4.04284038e-02 3.50036293e-01
7.94562399e-01 5.20965099e-01 -1.82779178e-01 6.83481991e-01
7.60547638e-01 -1.06909811e+00 1.92548752e-01 -1.84433907e-01
9.16749835e-01 -9.46391821e-01 9.39034641e-01 2.71328330e-01
-1.20989013e+00 -3.74289677e-02 -2.91261882e-01 -1.06520139e-01
-1.26472592e-01 1.15668488e+00 -1.35080004e+00 6.04742765e-01
3.91933322e-01 4.82279837e-01 -7.68338978e-01 1.39711320e+00
-4.56767576e-03 1.02197874e+00 -7.35657573e-01 -2.69999504e-01
1.64288789e-01 1.80310041e-01 5.84660172e-01 1.08812702e+00
3.76351953e-01 -3.34307164e-01 2.08867297e-01 6.94274247e-01
-6.46997392e-01 -5.50766177e-02 -8.39186609e-01 5.47988601e-02
1.00204945e+00 1.24298191e+00 -9.04982388e-01 -9.00563151e-02
-5.29509544e-01 1.60517648e-01 5.80552697e-01 6.29688680e-01
-9.66564119e-01 -5.26280522e-01 5.80513120e-01 6.72912478e-01
1.17718764e-01 8.76866728e-02 -2.51604229e-01 -6.08295500e-01
9.35972407e-02 -8.51349056e-01 4.47771549e-01 -2.08642960e-01
-1.34776700e+00 4.49975431e-01 -1.68945834e-01 -9.48517323e-01
8.92427117e-02 -3.14630657e-01 -9.40288126e-01 6.08677864e-01
-1.20480943e+00 -8.02818298e-01 -4.07047480e-01 3.29560637e-01
-1.22269548e-01 2.44118320e-03 4.15787101e-01 7.68962622e-01
-7.64278531e-01 1.24951482e+00 -8.74078721e-02 4.55595076e-01
6.58255458e-01 -8.46227765e-01 6.70914710e-01 1.02784181e+00
1.07191227e-01 6.60618007e-01 7.03618824e-01 -1.08704638e+00
-1.32582068e+00 -1.40716422e+00 5.78145325e-01 4.53486443e-02
8.57629657e-01 -4.42947328e-01 -8.80983710e-01 1.03836060e+00
-3.91629249e-01 2.84145236e-01 4.39053804e-01 6.57362163e-01
-2.53471345e-01 -3.30268472e-01 -1.13465428e+00 8.64358962e-01
1.21550310e+00 -1.62785947e-01 4.62750822e-01 5.18330216e-01
6.45271897e-01 -6.11763000e-01 -8.79470706e-01 8.01774979e-01
3.12983930e-01 -9.75838602e-01 7.90575325e-01 -3.94787073e-01
2.91010827e-01 -3.05955112e-03 4.28717077e-01 -1.29734910e+00
-2.45386839e-01 -5.97416043e-01 -3.96946967e-01 1.29762983e+00
9.16465223e-01 -1.03200448e+00 1.25125349e+00 4.96600121e-01
-2.46173237e-02 -9.02116537e-01 -7.19979405e-01 -7.61599898e-01
-2.57022232e-01 -3.95863056e-01 7.38166153e-01 9.54514503e-01
-1.77216426e-01 4.31922436e-01 -1.13579400e-01 5.27697146e-01
6.61611140e-01 -3.02171677e-01 1.13565838e+00 -1.46846473e+00
-3.59440297e-01 -3.59883875e-01 -5.55196226e-01 -1.11896312e+00
1.59589127e-01 -1.11599100e+00 -2.48121783e-01 -1.34670937e+00
1.41990319e-01 -1.17375243e+00 -1.51589438e-01 7.48819232e-01
-1.23077258e-01 3.09663057e-01 -8.04115459e-02 -6.07905909e-02
-5.70190072e-01 -1.02636881e-01 1.02943480e+00 -8.98880959e-02
-3.14262718e-01 2.03600436e-01 -7.18212783e-01 4.92323756e-01
1.03055155e+00 -6.64078057e-01 -5.79124093e-01 -3.44636768e-01
6.52980447e-01 1.47102922e-02 5.54305077e-01 -1.10493684e+00
4.88319635e-01 1.78176314e-01 3.82329017e-01 -7.18178511e-01
-1.10717021e-01 -8.45954597e-01 4.23920214e-01 6.08613253e-01
3.36122550e-02 3.31106693e-01 3.75718117e-01 6.68457210e-01
5.35925590e-02 -9.04807448e-02 2.93390274e-01 4.32005852e-01
-3.46855372e-01 4.77130711e-01 5.76268807e-02 2.02642158e-01
1.06373870e+00 -2.66438365e-01 -9.71772432e-01 -5.14152825e-01
-2.51117349e-01 7.91523218e-01 3.76603752e-01 1.42542049e-01
4.22323316e-01 -1.03109622e+00 -6.62376761e-01 4.30305541e-01
-1.78867392e-02 3.36274169e-02 1.87588215e-01 1.02099073e+00
-6.51445627e-01 9.04287994e-02 1.65916085e-01 -5.81556559e-01
-1.41892457e+00 8.76821727e-02 3.06783199e-01 -6.07884049e-01
-5.61248481e-01 8.45686495e-01 -1.06030226e-01 -4.29708093e-01
4.21463549e-01 -6.50255904e-02 2.04715088e-01 -4.46459651e-01
3.36139984e-02 5.57306588e-01 3.21712136e-01 -1.06721949e-02
-1.97244957e-01 -1.86711803e-01 -3.98683518e-01 5.34458280e-01
1.18594968e+00 1.52437210e-01 -2.98631459e-01 9.68906730e-02
1.10891557e+00 2.39754871e-01 -8.70566010e-01 -1.72257647e-01
5.75177148e-02 -3.71244252e-01 2.44260882e-04 -6.24550760e-01
-1.42624807e+00 2.28077099e-01 -8.07459131e-02 6.99226022e-01
8.82284045e-01 -6.68815449e-02 7.90131509e-01 2.89719194e-01
5.25874674e-01 -6.40554309e-01 -9.98252481e-02 4.98937786e-01
1.09827012e-01 -9.82576013e-01 2.06870064e-01 -9.86413598e-01
1.88822240e-01 9.09705937e-01 1.11763275e+00 1.41265422e-01
7.14987457e-01 5.89305460e-01 -3.46879005e-01 -4.51631159e-01
-1.01825845e+00 3.38539541e-01 -1.96913436e-01 4.20784771e-01
3.16963971e-01 1.95715979e-01 -5.56686223e-02 2.31042922e-01
-1.72927380e-01 -2.80842274e-01 7.49806643e-01 8.13458145e-01
-2.75608003e-01 -1.00398493e+00 -1.81412473e-01 1.17817402e+00
-4.32205141e-01 -4.10517722e-01 -4.16857243e-01 1.21688724e+00
-3.43321145e-01 9.35095906e-01 2.93177158e-01 -6.57715917e-01
1.64630786e-01 -4.37040478e-01 1.02808751e-01 -6.98399007e-01
-7.84584820e-01 -4.12497163e-01 5.49661040e-01 -4.49106425e-01
3.16407293e-01 -4.88398485e-02 -1.29923558e+00 -1.13070118e+00
-5.82217813e-01 2.66427875e-01 4.25876915e-01 3.83700639e-01
6.28963411e-01 7.85754800e-01 6.30577981e-01 -2.31924325e-01
-6.17512524e-01 -9.43232119e-01 -7.43402302e-01 -9.55065712e-03
3.34794909e-01 -5.58338404e-01 -6.99953675e-01 -8.16907942e-01] | [6.958458423614502, 6.074742794036865] |
1f57265e-4505-422a-b5d3-9b80509c0fe5 | domain-translation-with-conditional-gans-from | 1901.08101 | null | http://arxiv.org/abs/1901.08101v1 | http://arxiv.org/pdf/1901.08101v1.pdf | Domain Translation with Conditional GANs: from Depth to RGB Face-to-Face | Can faces acquired by low-cost depth sensors be useful to catch some
characteristic details of the face? Typically the answer is no. However, new
deep architectures can generate RGB images from data acquired in a different
modality, such as depth data. In this paper, we propose a new
\textit{Deterministic Conditional GAN}, trained on annotated RGB-D face
datasets, effective for a face-to-face translation from depth to RGB. Although
the network cannot reconstruct the exact somatic features for unknown
individual faces, it is capable to reconstruct plausible faces; their
appearance is accurate enough to be used in many pattern recognition tasks. In
fact, we test the network capability to hallucinate with some
\textit{Perceptual Probes}, as for instance face aspect classification or
landmark detection. Depth face can be used in spite of the correspondent RGB
images, that often are not available due to difficult luminance conditions.
Experimental results are very promising and are as far as better than
previously proposed approaches: this domain translation can constitute a new
way to exploit depth data in new future applications. | ['Roberto Vezzani', 'Matteo Fabbri', 'Guido Borghi', 'Rita Cucchiara', 'Fabio Lanzi', 'Simone Calderara'] | 2019-01-23 | null | null | null | null | ['face-to-face-translation'] | ['computer-vision'] | [ 3.53445768e-01 5.22769451e-01 2.82361954e-01 -7.64614284e-01
-6.28772557e-01 -4.32207942e-01 6.05036914e-01 -4.98797208e-01
5.38805127e-03 9.24754560e-01 -2.18551293e-01 2.21289262e-01
5.94189577e-02 -9.57961738e-01 -8.94788802e-01 -9.88564372e-01
3.01372737e-01 8.47324550e-01 -1.67737785e-03 -2.10715428e-01
-1.40249869e-02 1.07426929e+00 -2.20431018e+00 4.62688208e-01
1.72005981e-01 1.47052062e+00 2.17968136e-01 2.08867610e-01
-3.64857674e-01 5.87160468e-01 -4.41919148e-01 -5.77831388e-01
3.94221783e-01 -6.87056839e-01 -3.44965726e-01 8.13532621e-02
5.86963654e-01 -6.59521639e-01 -4.97222021e-02 1.02940023e+00
4.73725796e-01 -3.31943095e-01 6.65543795e-01 -1.10510635e+00
-4.97812092e-01 2.42792055e-01 -3.82915765e-01 -2.27149516e-01
7.51920462e-01 2.68145919e-01 4.13853288e-01 -1.01089561e+00
6.63199842e-01 1.48842716e+00 4.91422772e-01 1.29685616e+00
-1.10339880e+00 -6.26179278e-01 -2.52939552e-01 9.58084762e-02
-1.20606637e+00 -9.12642062e-01 1.04168224e+00 -1.48232549e-01
4.14985359e-01 2.41763234e-01 6.42328620e-01 1.65738440e+00
-1.60408318e-01 5.09786069e-01 1.61723757e+00 -2.14378819e-01
3.86942238e-01 2.90467858e-01 -5.67603290e-01 8.37814271e-01
1.17080472e-01 3.53342205e-01 -7.44181633e-01 2.41996706e-01
1.04056001e+00 6.25882074e-02 -6.41588986e-01 -2.44438171e-01
-8.25546384e-01 6.82424724e-01 7.22675622e-01 4.10653114e-01
-5.21343350e-01 1.03257127e-01 -1.32992923e-01 2.08203688e-01
4.70169634e-01 3.26858580e-01 -4.36552703e-01 -3.46498974e-02
-9.26156402e-01 -3.59753668e-02 6.06069326e-01 7.64131367e-01
1.08229828e+00 2.63651162e-01 1.43238500e-01 3.40115726e-01
3.71330142e-01 8.08569968e-01 3.48545551e-01 -1.00581670e+00
1.01462603e-01 4.64520097e-01 3.39415781e-02 -7.59945452e-01
-4.27051753e-01 2.52597891e-02 -9.54804599e-01 6.59946740e-01
6.00480318e-01 -1.83494780e-02 -1.02689910e+00 1.56091785e+00
3.91499341e-01 1.98928759e-01 -1.41713843e-02 1.07907259e+00
1.08939171e+00 4.41479236e-01 -3.74746144e-01 -3.27772260e-01
1.18202186e+00 -2.43868351e-01 -7.09739804e-01 -1.63831979e-01
-1.46641999e-01 -5.91189504e-01 1.00479078e+00 7.07432866e-01
-1.18709791e+00 -6.86006069e-01 -7.62978256e-01 6.74540848e-02
-3.62269551e-01 8.45502168e-02 5.97333491e-01 1.01780343e+00
-1.47887492e+00 6.04204178e-01 -7.38737166e-01 -3.36312145e-01
7.61767387e-01 3.69976819e-01 -8.43548417e-01 -3.30628425e-01
-8.00933897e-01 6.55254364e-01 4.84896675e-02 6.99632704e-01
-1.33080888e+00 -3.07477504e-01 -8.71557772e-01 -7.61757642e-02
1.18178539e-01 -6.80150628e-01 8.51348758e-01 -1.23900092e+00
-1.92759466e+00 1.10772896e+00 -2.59428412e-01 -1.14760451e-01
6.25231922e-01 2.03060135e-01 -1.64884835e-01 4.88786072e-01
-2.63485670e-01 8.43103945e-01 1.38256216e+00 -1.71340573e+00
7.91231357e-03 -1.04609096e+00 -2.87309736e-02 -1.91840112e-01
-2.63849556e-01 -2.33334646e-01 -2.32445613e-01 -2.27217779e-01
3.99444938e-01 -6.67693794e-01 1.74923614e-01 4.17472690e-01
-3.31968576e-01 2.37370357e-01 7.47099459e-01 -6.22673690e-01
8.69927332e-02 -1.91599166e+00 3.05588126e-01 1.60256252e-01
7.62509182e-02 2.05326170e-01 -1.56581029e-01 1.33349612e-01
-7.33043179e-02 3.16822752e-02 -3.99429262e-01 -8.90147805e-01
-2.63873130e-01 3.67391050e-01 -1.87513202e-01 6.83688283e-01
3.96825075e-01 9.71123815e-01 -5.01964331e-01 -1.65945679e-01
3.83847266e-01 9.31572616e-01 -1.99938402e-01 5.12399852e-01
-4.59775269e-01 1.21595263e+00 -3.47130388e-01 8.62608492e-01
9.68114018e-01 3.46992284e-01 -6.25287145e-02 -2.70742953e-01
3.11363220e-01 -1.33941650e-01 -7.98284173e-01 1.86301768e+00
-5.40457666e-01 5.84636927e-01 3.78350943e-01 -1.02306414e+00
1.13920987e+00 3.27217847e-01 2.91882336e-01 -8.61067355e-01
2.43641347e-01 1.86254606e-01 -4.36786175e-01 -6.40971363e-01
-9.44964737e-02 -6.58714414e-01 4.19989347e-01 9.22243819e-02
1.71504319e-01 -5.86429954e-01 -3.96446347e-01 -4.24510539e-01
8.09293389e-01 4.06256378e-01 -1.76398873e-01 2.01266795e-01
5.54900467e-01 -5.10444403e-01 2.18581870e-01 2.85385609e-01
2.11362109e-01 1.12836957e+00 5.16106963e-01 -5.36037326e-01
-8.38368952e-01 -1.07960486e+00 -2.68121094e-01 3.52593154e-01
-5.24767004e-02 1.68722257e-01 -8.25267971e-01 -7.80044854e-01
-1.51966780e-01 4.10431415e-01 -9.00074661e-01 -4.48262952e-02
-3.77394915e-01 -4.75302130e-01 5.03140330e-01 2.57249534e-01
6.04123294e-01 -1.33769512e+00 -5.87785542e-01 1.16444407e-02
6.59764633e-02 -1.37308455e+00 4.05123770e-01 1.17975526e-01
-1.04260194e+00 -1.04739606e+00 -9.73695576e-01 -5.16477764e-01
7.47367263e-01 -1.35278970e-01 1.13455451e+00 2.17011925e-02
-2.85740554e-01 4.54887211e-01 -4.35590029e-01 -2.70748079e-01
-3.31799924e-01 -4.06861842e-01 5.92298955e-02 3.99029404e-01
2.51086384e-01 -8.51922750e-01 -6.27865076e-01 2.65337408e-01
-9.90761459e-01 -2.96120793e-01 5.67710638e-01 6.64847851e-01
5.35744667e-01 6.20085858e-02 5.41075647e-01 -8.22942555e-01
4.99022119e-02 -1.41825587e-01 -5.78955352e-01 -1.93345752e-02
-2.11842179e-01 4.14905548e-02 6.65628552e-01 -1.05871849e-01
-1.23988235e+00 3.77157182e-01 -6.52107179e-01 -6.76653624e-01
-6.90680265e-01 -1.25887692e-02 -7.05418944e-01 -2.45141342e-01
6.97955608e-01 3.73062134e-01 2.48763561e-01 -6.40507638e-01
2.25397363e-01 3.90996009e-01 3.57804269e-01 -4.56837326e-01
8.66954148e-01 7.85140097e-01 3.63017440e-01 -9.27802980e-01
-4.01260734e-01 1.09387755e-01 -6.62170827e-01 -2.80020356e-01
9.02304769e-01 -7.35106289e-01 -9.55937743e-01 4.58726346e-01
-1.34519470e+00 -3.46239954e-01 -4.39375490e-01 1.69851616e-01
-5.69555879e-01 2.31649280e-01 -3.76176387e-01 -1.05803752e+00
1.16815001e-01 -1.09520662e+00 1.58347964e+00 3.33771944e-01
5.32663167e-01 -7.52233863e-01 -3.11281860e-01 4.39957231e-01
2.70234495e-01 5.14382899e-01 6.63053036e-01 2.95056812e-02
-6.53987229e-01 -2.63696849e-01 -1.43625274e-01 5.84900737e-01
2.67220587e-01 -4.27491777e-02 -1.71473312e+00 -2.07591251e-01
4.08011854e-01 -4.93700117e-01 9.04748738e-01 2.90132970e-01
1.49556613e+00 -2.21931815e-01 -2.09502175e-01 7.12347567e-01
1.42412341e+00 4.47974540e-03 1.00119984e+00 -3.04338247e-01
6.83678627e-01 9.78664398e-01 3.06193590e-01 3.72720122e-01
1.06563039e-01 9.17398870e-01 1.08665371e+00 -1.12630054e-01
-4.92814183e-01 -2.82963872e-01 5.12648523e-01 1.40548244e-01
-3.17894131e-01 -1.87274590e-01 -6.98599577e-01 1.37125477e-01
-1.23792613e+00 -7.59862244e-01 -1.51953593e-01 2.23467827e+00
3.97260159e-01 3.72770801e-02 -1.10111600e-02 4.05483574e-01
3.99087608e-01 -2.64245391e-01 -4.63345557e-01 -1.57916799e-01
-4.39842045e-01 7.55658090e-01 -9.74711999e-02 3.40017885e-01
-5.11479497e-01 7.56551504e-01 5.36844015e+00 5.10407567e-01
-1.40115881e+00 2.87321270e-01 7.65127122e-01 1.15440853e-01
-5.15354753e-01 -3.27283293e-01 -6.58025563e-01 3.90163690e-01
8.09556127e-01 6.41354859e-01 5.04937053e-01 6.96105480e-01
-9.37782452e-02 -3.20867151e-01 -1.38906991e+00 1.59246504e+00
4.12941933e-01 -1.09061873e+00 1.56734809e-01 1.95614725e-01
5.92951775e-01 -3.82507831e-01 3.82635772e-01 -4.14699949e-02
-3.55480611e-01 -1.47566354e+00 6.12150192e-01 9.20450687e-01
1.19871998e+00 -7.23560154e-01 7.60341465e-01 4.51027095e-01
-7.64783204e-01 4.27868329e-02 -6.32884026e-01 6.54180348e-02
-2.05682851e-02 8.22124124e-01 -9.77882981e-01 6.99739397e-01
7.33548403e-01 5.65979838e-01 -7.55055010e-01 5.66468418e-01
-4.05363142e-01 1.61443070e-01 -4.37974185e-01 5.21936417e-02
-1.69973493e-01 -1.63203895e-01 2.50523776e-01 4.35335875e-01
8.16571057e-01 1.48595488e-02 -4.56872940e-01 1.05440819e+00
-1.46702975e-01 -1.53109238e-01 -1.04834902e+00 2.34563053e-01
4.58328277e-02 1.26797497e+00 -7.87410855e-01 -3.57820950e-02
-2.68488135e-02 1.44549978e+00 1.68949991e-01 2.99757004e-01
-6.74530625e-01 1.99938983e-01 4.58257914e-01 4.15620238e-01
3.24084520e-01 -8.75497609e-02 -5.48252724e-02 -1.26609111e+00
1.12404585e-01 -6.90203726e-01 -2.11912870e-01 -1.22904587e+00
-1.36159992e+00 1.07591522e+00 -1.55874878e-01 -1.07900882e+00
-5.10959625e-01 -1.10307503e+00 -2.26691455e-01 7.65358865e-01
-1.58501685e+00 -1.44073939e+00 -7.16147065e-01 9.90375757e-01
3.21736813e-01 -2.20711693e-01 1.23174024e+00 2.09319621e-01
-1.88927189e-01 4.88568634e-01 -1.90912366e-01 -1.52865723e-01
5.25882959e-01 -9.78622139e-01 -6.22343905e-02 4.35603231e-01
5.10013282e-01 1.18252747e-01 5.98698795e-01 -2.82416612e-01
-1.69706488e+00 -8.34523618e-01 2.46646509e-01 -7.52365232e-01
-3.02989125e-01 -5.27871609e-01 -6.01475716e-01 5.50736070e-01
-9.00965109e-02 4.68382359e-01 3.74826908e-01 -3.67040098e-01
-2.49723136e-01 -5.66499412e-01 -1.75114429e+00 8.40237364e-02
1.09569287e+00 -5.57034671e-01 -2.86894590e-01 1.86215773e-01
2.31161997e-01 -3.73176426e-01 -6.73626482e-01 4.59620923e-01
4.84741509e-01 -1.66288495e+00 9.26601112e-01 -1.86847314e-01
4.78963226e-01 -1.68047100e-01 -2.59679019e-01 -1.22198963e+00
4.93243933e-01 -2.99032509e-01 -7.84555450e-02 1.41644740e+00
1.40921652e-01 -6.77547991e-01 1.09338725e+00 3.95895809e-01
6.56272545e-02 -5.17894447e-01 -1.44039643e+00 -6.15690887e-01
-5.94648346e-02 -5.58674216e-01 9.23999250e-01 7.27596045e-01
-6.55387998e-01 1.73187420e-01 -3.49151611e-01 2.11265564e-01
8.90366733e-01 7.24082962e-02 8.13402474e-01 -1.41331267e+00
-2.66530812e-01 -2.40536034e-01 -7.21899986e-01 -8.44978869e-01
5.34827173e-01 -6.49053037e-01 -5.04250936e-02 -1.26741803e+00
-3.80093098e-01 -5.16003370e-01 2.68441111e-01 3.79347622e-01
3.94239545e-01 9.35114086e-01 -1.80279929e-02 -1.23196162e-01
1.14561126e-01 8.65703344e-01 1.35409307e+00 -6.61236048e-02
2.00886920e-01 1.21950798e-01 -3.53007764e-01 6.24368191e-01
6.15245223e-01 -3.07789445e-01 -3.04182678e-01 -3.46504450e-01
1.86291426e-01 3.40500116e-01 6.06429279e-01 -1.28369784e+00
-6.04565926e-02 2.11991310e-01 1.02773297e+00 -2.60144651e-01
1.02252924e+00 -1.27493644e+00 4.58993316e-01 2.64517099e-01
2.91035980e-01 -3.26367080e-01 1.28154635e-01 5.09026885e-01
-4.13473517e-01 -1.83739930e-01 8.58750641e-01 -5.18282175e-01
-5.05719364e-01 4.41006958e-01 -6.32616552e-03 -5.01105189e-01
9.56856728e-01 -4.71129298e-01 4.70508300e-02 -5.97632587e-01
-9.52205062e-01 -5.80311358e-01 6.93810046e-01 1.98084563e-01
1.08036304e+00 -1.32126606e+00 -5.02974808e-01 7.05273747e-01
-1.54225377e-03 2.93704331e-01 3.07753056e-01 4.19638515e-01
-5.69442868e-01 2.11364001e-01 -6.34411037e-01 -8.80150497e-01
-9.05408442e-01 7.87098467e-01 5.37508070e-01 3.98295194e-01
-5.29550135e-01 9.83281136e-01 2.80833274e-01 -3.78747523e-01
2.90268380e-02 -2.89057523e-01 -1.33318081e-01 1.46914020e-01
6.02341115e-01 -1.59555286e-01 3.75689954e-01 -7.96778381e-01
-4.11043525e-01 9.93718922e-01 5.61069906e-01 -1.91934839e-01
1.56520116e+00 -1.52428240e-01 -3.40535671e-01 4.63789791e-01
1.05783832e+00 -1.18068784e-01 -1.41678977e+00 1.53348595e-01
-4.42374676e-01 -7.22470284e-01 -6.52701408e-02 -7.38451660e-01
-1.63218594e+00 1.38737476e+00 9.59150851e-01 1.61269620e-01
1.44987512e+00 9.96273607e-02 4.14347321e-01 1.86188519e-01
9.88067508e-01 -4.22427505e-01 2.75652498e-01 -4.81124669e-02
1.20509315e+00 -1.44410026e+00 -3.07685882e-01 -4.59158391e-01
-3.30828577e-01 1.49700272e+00 6.23223662e-01 5.50724268e-02
6.12532973e-01 1.46559328e-01 3.76667758e-03 -3.27133715e-01
-3.07638407e-01 -4.04635340e-01 3.74930501e-02 1.07424748e+00
2.04182684e-01 -1.98463827e-01 3.67427915e-01 2.22861245e-01
-2.27343217e-01 1.80332048e-03 5.34404397e-01 3.65276814e-01
-1.00351341e-01 -1.19571197e+00 -6.80098832e-01 1.55491471e-01
-2.14463234e-01 2.43880838e-01 -5.20190656e-01 7.28935480e-01
3.89656812e-01 8.00104618e-01 6.36246726e-02 -2.53593922e-01
2.77643889e-01 1.66833580e-01 1.26340926e+00 -6.04017973e-01
-2.26197124e-01 -1.61762834e-01 -2.01036006e-01 -6.32425547e-01
-6.62622809e-01 -5.16628504e-01 -9.01711881e-01 -2.68320471e-01
3.08025144e-02 -2.65237093e-01 9.32348371e-01 7.99256206e-01
-2.41214670e-02 2.28611112e-01 8.26241076e-01 -1.46441507e+00
-9.10181031e-02 -9.76601481e-01 -7.02687025e-01 4.11415219e-01
6.39004827e-01 -8.39618087e-01 -3.88683289e-01 1.89766958e-02] | [12.811930656433105, -0.20958364009857178] |
8acb0a96-5df1-422b-aed1-0b6f1313c984 | weakly-supervised-caricature-face-parsing | 1905.05091 | null | https://arxiv.org/abs/1905.05091v1 | https://arxiv.org/pdf/1905.05091v1.pdf | Weakly-supervised Caricature Face Parsing through Domain Adaptation | A caricature is an artistic form of a person's picture in which certain striking characteristics are abstracted or exaggerated in order to create a humor or sarcasm effect. For numerous caricature related applications such as attribute recognition and caricature editing, face parsing is an essential pre-processing step that provides a complete facial structure understanding. However, current state-of-the-art face parsing methods require large amounts of labeled data on the pixel-level and such process for caricature is tedious and labor-intensive. For real photos, there are numerous labeled datasets for face parsing. Thus, we formulate caricature face parsing as a domain adaptation problem, where real photos play the role of the source domain, adapting to the target caricatures. Specifically, we first leverage a spatial transformer based network to enable shape domain shifts. A feed-forward style transfer network is then utilized to capture texture-level domain gaps. With these two steps, we synthesize face caricatures from real photos, and thus we can use parsing ground truths of the original photos to learn the parsing model. Experimental results on the synthetic and real caricatures demonstrate the effectiveness of the proposed domain adaptation algorithm. Code is available at: https://github.com/ZJULearning/CariFaceParsing . | ['Ming-Hsuan Yang', 'Yi-Hsuan Tsai', 'Wei-Chih Hung', 'Wenqing Chu', 'Deng Cai'] | 2019-05-13 | null | null | null | null | ['face-parsing', 'caricature'] | ['computer-vision', 'computer-vision'] | [ 4.91699874e-01 2.55278230e-01 -1.28825933e-01 -7.74722040e-01
-6.29191816e-01 -5.53316653e-01 4.02367979e-01 -6.94238722e-01
1.58506498e-01 5.46391845e-01 5.14419377e-02 1.32751456e-02
3.50572854e-01 -9.52400386e-01 -9.94644046e-01 -6.08610809e-01
5.88671327e-01 3.89234960e-01 -2.19294056e-01 -1.75672516e-01
7.46123791e-02 6.72557175e-01 -1.25194001e+00 4.91352946e-01
7.72174418e-01 1.11957705e+00 1.25370771e-01 3.84451151e-01
-5.84114254e-01 6.00426137e-01 -4.61281776e-01 -8.65608275e-01
1.28823936e-01 -5.84538460e-01 -6.51406288e-01 5.84197044e-01
7.42986619e-01 -4.72333878e-01 -2.57904172e-01 1.34767318e+00
2.89858878e-01 -1.32480711e-01 4.19521809e-01 -1.20843172e+00
-1.11649990e+00 3.20643246e-01 -9.14163411e-01 -2.22958520e-01
3.00246149e-01 8.82450864e-02 4.23722386e-01 -1.20153892e+00
7.20997274e-01 1.72944963e+00 4.33625370e-01 1.04719377e+00
-1.13054097e+00 -1.24780524e+00 1.11026019e-01 1.23897597e-01
-1.13884580e+00 -8.34986925e-01 1.54694474e+00 -2.27432698e-01
2.20379934e-01 -5.54857478e-02 4.90718991e-01 1.11473548e+00
-5.47503382e-02 8.34797561e-01 1.07403290e+00 -2.82521367e-01
4.40789014e-02 1.62421063e-01 -3.39889526e-01 7.79581189e-01
-2.08923563e-01 5.04866131e-02 -3.30187708e-01 1.94671527e-01
1.02091587e+00 1.64892614e-01 -1.15882710e-01 -2.39408419e-01
-6.16309524e-01 7.10573733e-01 4.79130477e-01 -5.84944971e-02
-2.25775197e-01 7.11575523e-02 2.48771831e-01 2.15278476e-01
5.18089890e-01 -7.47143850e-02 -3.12775701e-01 2.50720829e-01
-6.84996545e-01 2.23727718e-01 3.38248551e-01 9.88964558e-01
8.46534967e-01 1.67735502e-01 8.79540518e-02 1.18195677e+00
2.34534770e-01 7.02457309e-01 5.04705198e-02 -1.19172287e+00
4.40281540e-01 6.80813670e-01 -2.30437979e-01 -1.04332805e+00
6.88305274e-02 5.04572205e-02 -9.64969933e-01 2.90731728e-01
5.04538894e-01 -1.48513407e-01 -9.92143929e-01 1.79305708e+00
6.19536579e-01 4.00188535e-01 1.38424560e-01 9.61135507e-01
1.04519475e+00 7.61421263e-01 2.77633607e-01 -1.92806348e-01
1.71216047e+00 -7.36821294e-01 -8.31476748e-01 -5.39197266e-01
8.24809000e-02 -9.55977380e-01 1.44288528e+00 2.93166220e-01
-1.18256116e+00 -5.85119188e-01 -8.33785951e-01 -3.82943600e-01
-1.12042561e-01 3.98456275e-01 5.40919662e-01 5.75453162e-01
-6.75549567e-01 4.28820968e-01 -4.65690017e-01 -1.04875468e-01
1.18059564e+00 2.67022133e-01 -5.94317675e-01 -4.64103460e-01
-9.42822099e-01 3.04125488e-01 -5.00286929e-03 2.68875420e-01
-7.64926970e-01 -8.50488126e-01 -9.75418627e-01 -6.05827384e-02
4.65118587e-01 -4.53648657e-01 1.22460890e+00 -1.58511174e+00
-1.63391960e+00 1.24235308e+00 -3.38010132e-01 3.53864580e-02
3.49182904e-01 -3.88056003e-02 -6.06694460e-01 2.56536663e-01
-2.69859619e-02 7.68892407e-01 1.31804800e+00 -1.34442163e+00
-4.37359661e-01 -5.86346686e-01 -3.53861786e-02 8.72745514e-02
-1.54421896e-01 1.62678361e-01 -5.73846877e-01 -8.86029422e-01
5.73180504e-02 -7.49611080e-01 2.99560249e-01 4.28250045e-01
-3.27988923e-01 -9.15839598e-02 1.19058657e+00 -1.01308429e+00
7.84373879e-01 -2.34749103e+00 5.75409345e-02 -9.12670493e-02
8.86501148e-02 2.15996563e-01 -3.60219479e-01 -6.76610172e-02
-3.56918305e-01 9.75877270e-02 -5.14522076e-01 -3.06863666e-01
-3.05023402e-01 1.57965362e-01 -4.75535363e-01 1.94853038e-01
7.14829743e-01 9.44903970e-01 -6.45818889e-01 -6.63863122e-01
1.44292518e-01 6.64195716e-01 -6.16384566e-01 3.07233155e-01
-3.77104938e-01 6.96886957e-01 -5.40109456e-01 9.24941897e-01
1.17751610e+00 3.71984504e-02 7.58488402e-02 -4.03175533e-01
2.65297204e-01 -2.19824493e-01 -8.53959441e-01 1.76578307e+00
-4.69898254e-01 4.25528586e-01 3.68374467e-01 -9.30213690e-01
1.09143066e+00 1.01488277e-01 2.64184564e-01 -7.29722202e-01
3.23593706e-01 1.30746979e-02 -2.04728886e-01 -5.51572025e-01
-6.41429201e-02 -5.66420853e-01 -8.89150128e-02 3.14641714e-01
-1.28422037e-01 -4.70744103e-01 -2.87725121e-01 6.89174533e-02
4.82224017e-01 3.48339856e-01 7.81916007e-02 3.24210003e-02
9.13535893e-01 2.08914410e-02 8.87727976e-01 -2.51927257e-01
-1.89010784e-01 7.52025723e-01 6.23971641e-01 -6.41096771e-01
-1.08600366e+00 -1.02974463e+00 5.59652643e-03 9.62202311e-01
9.92773920e-02 1.54862866e-01 -1.34317589e+00 -5.90292335e-01
-1.28393993e-01 5.89794755e-01 -7.91535974e-01 -2.51120865e-01
-7.70545006e-01 -3.20530027e-01 3.92926872e-01 4.92134869e-01
9.03818130e-01 -1.47655869e+00 -7.19133839e-02 -5.44565357e-02
-1.68374881e-01 -1.23999274e+00 -8.74373376e-01 -5.68057656e-01
-7.16378510e-01 -1.03367412e+00 -5.72741210e-01 -1.13693392e+00
8.40684772e-01 2.78885841e-01 1.02260292e+00 2.12087974e-01
-2.73594379e-01 1.77012250e-01 -4.03786525e-02 -3.57257515e-01
-4.82713848e-01 -3.57093811e-01 -1.47839367e-01 4.19630885e-01
4.59802061e-01 -7.11611032e-01 -6.28723800e-01 1.14262566e-01
-9.44740355e-01 1.97178498e-01 6.66299164e-01 8.96010935e-01
8.02738607e-01 -9.45453346e-02 6.67101204e-01 -1.33860433e+00
5.30418575e-01 -2.46631652e-01 -5.50279140e-01 3.37990224e-01
-7.98437595e-02 -1.13685578e-01 8.73394608e-01 -7.16732144e-01
-1.65025020e+00 3.61490220e-01 -1.22009970e-01 -8.44773889e-01
-3.11058372e-01 -1.46152586e-01 -8.29604447e-01 -5.93738705e-02
2.92419702e-01 3.55037212e-01 4.09675479e-01 -3.10617149e-01
5.07301033e-01 5.33792675e-01 9.16106284e-01 -8.58567059e-01
9.24059987e-01 6.69178009e-01 -2.28975102e-01 -7.69904733e-01
-7.15461135e-01 3.75966549e-01 -6.21054649e-01 -2.95541137e-01
1.01714635e+00 -8.10291708e-01 -7.04223871e-01 6.78765774e-01
-1.25898111e+00 -3.25162113e-01 -1.38988376e-01 -2.18170658e-01
-4.88774657e-01 2.60912776e-01 -4.30534989e-01 -4.69956726e-01
-3.25764507e-01 -1.01230597e+00 1.23817873e+00 6.58608675e-01
1.53639778e-01 -7.56804526e-01 -3.01209569e-01 7.01031446e-01
-1.39405280e-02 3.63544673e-01 1.17478514e+00 -9.85980704e-02
-5.92243016e-01 -2.04371102e-02 -3.67569864e-01 3.91818315e-01
3.65182877e-01 1.00814015e-01 -1.13190985e+00 5.63719170e-03
-9.25426856e-02 -3.93760353e-01 3.73272806e-01 2.73475975e-01
1.55167389e+00 -3.32776755e-01 -1.73431471e-01 8.19713175e-01
1.10774803e+00 3.24829340e-01 8.68606091e-01 -2.56043673e-01
7.92927980e-01 1.03023386e+00 7.18407393e-01 2.72830695e-01
3.51404876e-01 4.20456290e-01 3.13469112e-01 -3.65598083e-01
-4.85616207e-01 -7.89674997e-01 2.17732698e-01 4.67032969e-01
1.56547740e-01 1.44191593e-01 -7.19734490e-01 4.07074690e-01
-1.29222989e+00 -8.90648544e-01 1.90768331e-01 1.85920453e+00
9.08843577e-01 -1.50473297e-01 -1.81831345e-01 -1.32071495e-01
9.64708567e-01 -9.55587253e-03 -1.02132750e+00 -2.75579661e-01
-5.82405999e-02 3.82059515e-01 4.31051664e-02 4.39854711e-01
-9.13744211e-01 1.50447965e+00 5.01636696e+00 8.43598843e-01
-1.37329865e+00 6.66330978e-02 1.07815802e+00 1.64547283e-02
-3.23432952e-01 -1.31821826e-01 -4.79888350e-01 4.86354619e-01
2.92389095e-01 3.06144953e-02 7.53806472e-01 9.04831529e-01
2.83664972e-01 2.20449135e-01 -9.23864782e-01 1.30802131e+00
1.31454408e-01 -1.20050669e+00 1.36623979e-01 -1.64185986e-01
5.48007607e-01 -7.01011956e-01 3.82097125e-01 2.67588437e-01
1.33736596e-01 -1.17036378e+00 5.01765549e-01 3.83522421e-01
1.31744587e+00 -7.73318172e-01 2.33907998e-01 4.99620587e-02
-1.21811402e+00 -2.04504356e-02 -4.73313779e-01 3.09541404e-01
5.46078384e-02 2.32890770e-01 -5.31090498e-01 -2.72971224e-02
7.23392487e-01 6.13092124e-01 -4.22849208e-01 1.02786794e-01
-3.02962244e-01 4.22798336e-01 -2.03323718e-02 4.19772357e-01
-1.72732472e-01 -4.97295082e-01 1.61414266e-01 6.59512401e-01
2.94561833e-01 5.90720594e-01 -1.54441163e-01 1.00112414e+00
-5.74404776e-01 7.69245252e-02 -5.65600216e-01 -7.03065395e-02
8.02393675e-01 1.41803086e+00 -4.79654849e-01 -3.22589666e-01
-4.97481525e-01 1.16994750e+00 2.95102686e-01 2.91714340e-01
-8.73792410e-01 -1.80652276e-01 9.04975414e-01 2.46864647e-01
1.22722358e-01 3.07133794e-01 -4.08852905e-01 -1.08122146e+00
2.76360922e-02 -1.01393723e+00 2.31307328e-01 -9.75762486e-01
-1.31351483e+00 6.04906976e-01 -2.04654261e-01 -1.11278868e+00
-5.93732297e-02 -5.99028409e-01 -8.68895710e-01 7.58851051e-01
-1.56723893e+00 -1.61893284e+00 -6.57882750e-01 8.26494873e-01
8.87486875e-01 -2.56512582e-01 5.81216037e-01 2.43685707e-01
-7.01194882e-01 7.62826383e-01 -3.43349487e-01 3.85382384e-01
8.77871752e-01 -7.55678236e-01 4.74006653e-01 7.66760111e-01
-5.40592670e-02 4.03817326e-01 4.98602867e-01 -5.91194868e-01
-1.50475895e+00 -1.25628841e+00 4.19076085e-01 -2.87111104e-01
3.11868817e-01 -5.42790413e-01 -1.23964536e+00 6.54458582e-01
3.69386673e-02 3.95400345e-01 4.29953545e-01 -3.89858603e-01
-4.12921250e-01 -4.07934546e-01 -1.41837323e+00 6.10507488e-01
1.25082505e+00 -4.33955699e-01 -4.99965817e-01 1.93782613e-01
5.11540294e-01 -3.88446808e-01 -7.97515988e-01 2.44065776e-01
5.77506304e-01 -8.41381669e-01 1.11107993e+00 -5.19217789e-01
8.07570875e-01 -1.48347884e-01 4.10452075e-02 -1.08193898e+00
-5.86386360e-02 -4.33171630e-01 1.98419467e-01 1.58539402e+00
5.79578057e-02 -5.36311030e-01 1.11525440e+00 7.10786819e-01
6.33795634e-02 -5.51411867e-01 -7.59379089e-01 -2.19926909e-01
4.72105980e-01 -1.66911125e-01 9.88853633e-01 1.08301079e+00
-5.22061765e-01 4.00989711e-01 -3.80644411e-01 1.32220685e-01
7.97840953e-01 2.73012429e-01 1.00047898e+00 -1.13171875e+00
6.43972754e-02 -2.44235069e-01 1.68200694e-02 -9.20373976e-01
6.30958676e-01 -6.70674205e-01 -9.03103426e-02 -1.00346255e+00
3.93087417e-02 -4.85430747e-01 2.31084719e-01 5.67871809e-01
-1.36367321e-01 2.95965880e-01 1.99549705e-01 1.47283345e-01
3.85069810e-02 6.76661670e-01 1.72299969e+00 -2.36072853e-01
-9.52793360e-02 -2.25594401e-01 -8.98073077e-01 9.13190663e-01
7.91393518e-01 -3.30069929e-01 -6.18350506e-01 -6.31125033e-01
-2.51385719e-01 4.19211388e-01 4.38278466e-01 -6.59746349e-01
-3.55691761e-02 -4.51354355e-01 6.87343538e-01 -3.50923985e-01
5.74665129e-01 -8.73684168e-01 1.56299233e-01 2.36576617e-01
-1.27151862e-01 -2.14537293e-01 3.10814917e-01 4.27246720e-01
-3.51719260e-01 4.79485020e-02 1.17196882e+00 -1.13079652e-01
-9.37912762e-01 7.10209489e-01 3.18872631e-01 1.51220500e-01
1.08973503e+00 -3.74179542e-01 -2.33722717e-01 -3.25242847e-01
-6.67209804e-01 -1.64141469e-02 6.13492548e-01 5.82721114e-01
9.20722425e-01 -1.50445986e+00 -7.91340649e-01 6.04766071e-01
6.55430555e-02 2.81750083e-01 6.97026372e-01 3.01888138e-01
-5.38898587e-01 -1.73906647e-02 -5.50150990e-01 -5.04466951e-01
-1.41482997e+00 6.02068961e-01 4.15521711e-01 3.84177238e-01
-5.41896403e-01 8.68580580e-01 7.73031771e-01 -5.68844318e-01
-1.81378156e-01 -1.16824128e-01 -1.92822009e-01 -1.41439423e-01
6.70681715e-01 -1.43318489e-01 -2.83559322e-01 -7.41013944e-01
-3.41736972e-01 1.00405622e+00 -1.04968049e-01 1.19490281e-01
1.25729632e+00 -1.02123491e-01 -9.53360051e-02 -1.96773529e-01
1.16419041e+00 1.05763292e-02 -1.70442331e+00 -3.41548294e-01
-4.74394083e-01 -6.40849769e-01 -1.42317191e-01 -6.29464984e-01
-1.48653507e+00 1.24818921e+00 5.99805653e-01 -5.50182283e-01
1.60855794e+00 1.48228750e-01 9.35566902e-01 6.37615770e-02
2.00282514e-01 -9.53534842e-01 2.53622472e-01 2.48896368e-02
1.20405769e+00 -1.32161570e+00 -2.54700750e-01 -9.53751385e-01
-9.26549852e-01 1.08123159e+00 8.88950646e-01 -1.97425023e-01
5.40342212e-01 2.37117633e-01 1.59163743e-01 -9.34529528e-02
-5.22187829e-01 2.35638589e-01 2.48882398e-02 8.52121532e-01
2.36457154e-01 5.34501970e-02 6.94985464e-02 9.75509644e-01
-3.37512285e-01 5.91345243e-02 3.26164901e-01 4.89206076e-01
-2.67595470e-01 -1.13338768e+00 -6.68653190e-01 1.14853181e-01
-3.16722721e-01 1.75020304e-02 -5.74988782e-01 8.17809165e-01
2.24096969e-01 6.62868559e-01 2.33063966e-01 -2.75540978e-01
3.95058304e-01 1.35317624e-01 6.43418491e-01 -5.53998709e-01
-1.14246480e-01 1.32806301e-02 -2.39829957e-01 -4.35831666e-01
-2.05592409e-01 -6.05298936e-01 -1.46564996e+00 -5.99218130e-01
2.05427572e-01 -2.24231079e-01 5.50483227e-01 6.86195195e-01
3.83891255e-01 3.97295237e-01 7.14914143e-01 -7.36693323e-01
-1.71667784e-01 -8.78318608e-01 -3.54556710e-01 8.01143348e-01
1.04151450e-01 -7.06560671e-01 2.98463590e-02 5.36735475e-01] | [12.611674308776855, -0.13065862655639648] |
ff199b0c-9b37-4ed0-b368-335cdcec4cc3 | language-driven-semantic-segmentation-1 | 2201.03546 | null | https://arxiv.org/abs/2201.03546v2 | https://arxiv.org/pdf/2201.03546v2.pdf | Language-driven Semantic Segmentation | We present LSeg, a novel model for language-driven semantic image segmentation. LSeg uses a text encoder to compute embeddings of descriptive input labels (e.g., "grass" or "building") together with a transformer-based image encoder that computes dense per-pixel embeddings of the input image. The image encoder is trained with a contrastive objective to align pixel embeddings to the text embedding of the corresponding semantic class. The text embeddings provide a flexible label representation in which semantically similar labels map to similar regions in the embedding space (e.g., "cat" and "furry"). This allows LSeg to generalize to previously unseen categories at test time, without retraining or even requiring a single additional training sample. We demonstrate that our approach achieves highly competitive zero-shot performance compared to existing zero- and few-shot semantic segmentation methods, and even matches the accuracy of traditional segmentation algorithms when a fixed label set is provided. Code and demo are available at https://github.com/isl-org/lang-seg. | ['René Ranftl', 'Vladlen Koltun', 'Serge Belongie', 'Kilian Q. Weinberger', 'Boyi Li'] | 2022-01-10 | language-driven-semantic-segmentation | https://openreview.net/forum?id=RriDjddCLN | https://openreview.net/pdf?id=RriDjddCLN | iclr-2022-4 | ['zero-shot-segmentation'] | ['computer-vision'] | [ 2.96246827e-01 2.86037594e-01 -2.21588895e-01 -7.37653136e-01
-1.00191760e+00 -6.01418674e-01 4.01981771e-01 5.21656461e-02
-4.49325085e-01 2.86597133e-01 9.64053646e-02 5.66895492e-02
3.21243018e-01 -9.46152031e-01 -9.44544852e-01 -6.62488461e-01
2.88936794e-01 7.50856876e-01 3.42646718e-01 4.26555909e-02
2.39811257e-01 1.67422757e-01 -1.74975860e+00 6.77428320e-02
6.15647197e-01 9.80651379e-01 5.31279504e-01 6.78417563e-01
-4.23614204e-01 5.46174347e-01 -3.98844391e-01 -1.25164121e-01
1.17329873e-01 -3.35114419e-01 -1.19565511e+00 4.99177337e-01
5.59881747e-01 -4.21280831e-01 -4.08216208e-01 1.17361891e+00
1.55875251e-01 3.18547130e-01 6.99304223e-01 -1.24210954e+00
-1.03945565e+00 4.51478153e-01 -3.29896748e-01 -1.23184741e-01
1.58071175e-01 -1.81964934e-02 1.26918173e+00 -1.16436243e+00
8.56781542e-01 1.10957277e+00 4.89669293e-01 6.22467160e-01
-1.43218648e+00 -3.60006362e-01 2.20401347e-01 -1.68528929e-02
-1.56100523e+00 -2.26114020e-01 5.92213452e-01 -6.34815693e-01
6.28876567e-01 -2.07202151e-01 5.98808408e-01 9.31936800e-01
-1.59009844e-01 8.16651762e-01 7.82713294e-01 -3.79074663e-01
5.10803998e-01 1.01748325e-01 3.04347426e-01 9.27318990e-01
-1.79960042e-01 -3.11431050e-01 -3.95305276e-01 1.47893593e-01
5.98213971e-01 2.09641054e-01 -3.88620123e-02 -6.21115923e-01
-1.07159114e+00 1.05902159e+00 7.97193468e-01 1.35690689e-01
-9.05295610e-02 6.63301647e-01 3.41610521e-01 -1.87668189e-01
7.01625645e-01 1.99613988e-01 -2.87810773e-01 1.39323547e-01
-1.13165772e+00 4.78202999e-02 4.97738183e-01 1.28296459e+00
1.35223496e+00 -1.28484964e-01 -1.53948456e-01 1.05951774e+00
2.33177707e-01 4.41084206e-01 3.31083715e-01 -1.34128368e+00
1.46523878e-01 5.34757137e-01 1.32527173e-01 -4.39225763e-01
-1.78448364e-01 -7.26839975e-02 -3.16918820e-01 3.24989818e-02
1.33980766e-01 7.03385249e-02 -1.66117430e+00 1.71706891e+00
1.70122147e-01 3.27217728e-01 1.09688165e-02 9.56324100e-01
8.54040921e-01 8.82135987e-01 3.02270114e-01 5.73642433e-01
1.50479865e+00 -1.28049374e+00 -3.53696346e-01 -6.81463122e-01
7.67740965e-01 -5.03168821e-01 1.53292692e+00 -9.60978270e-02
-8.25913608e-01 -4.11696494e-01 -9.34070408e-01 -6.10108018e-01
-9.14914370e-01 1.38732597e-01 3.54913205e-01 2.74334759e-01
-1.25296605e+00 5.33254743e-01 -9.61773694e-01 -7.71510601e-01
7.28449225e-01 7.83708543e-02 -2.32737586e-01 -3.10243905e-01
-9.23814476e-01 5.63936055e-01 5.61175168e-01 -3.13561052e-01
-1.17640936e+00 -5.02138734e-01 -1.26862097e+00 8.21261778e-02
1.86812937e-01 -3.32090735e-01 1.42408049e+00 -1.02487743e+00
-1.10295546e+00 1.48713267e+00 -3.12992781e-01 -3.58493775e-01
9.33181867e-02 -2.24812418e-01 -2.41326913e-02 5.67434430e-01
7.85282850e-01 1.57430339e+00 6.50483727e-01 -1.41643608e+00
-4.44154710e-01 -1.97817922e-01 2.55933762e-01 3.37844163e-01
-1.80259570e-01 -8.99650082e-02 -6.93749845e-01 -5.60361505e-01
4.15654629e-01 -8.87817562e-01 -2.34120041e-01 5.44575572e-01
-5.13206005e-01 -1.12731434e-01 9.18455064e-01 -6.02115273e-01
8.18918824e-01 -2.31534266e+00 7.32421456e-03 -5.92105612e-02
-2.02148780e-02 -1.93495929e-01 -2.03546524e-01 4.01069999e-01
-3.25968340e-02 2.47975916e-01 -7.91872144e-01 -4.47199941e-01
8.32308531e-02 6.44330859e-01 -2.03349963e-01 4.97366816e-01
2.77865320e-01 1.06927979e+00 -1.04386616e+00 -6.65379405e-01
4.51358706e-01 3.62193853e-01 -5.26069283e-01 2.08962694e-01
-5.63713074e-01 3.09161723e-01 -3.11154842e-01 7.48040617e-01
4.85506356e-01 -5.75978458e-01 -3.30149323e-01 -1.17039695e-01
-6.63920641e-02 2.18137950e-01 -7.06186473e-01 2.19336390e+00
-5.32032430e-01 9.16132569e-01 -1.73534214e-01 -1.14037085e+00
8.96794975e-01 1.28153443e-01 3.04109335e-01 -4.62496042e-01
2.91295379e-01 1.40285626e-01 -5.81357658e-01 -3.58186424e-01
5.80048084e-01 -8.34808499e-02 -4.86819714e-01 5.32620728e-01
3.60934913e-01 -3.74473929e-01 2.15441316e-01 2.95880884e-01
7.92496026e-01 4.73819762e-01 -1.26718478e-02 -3.96617234e-01
1.14269711e-01 4.48012114e-01 5.50579011e-01 4.58622545e-01
-7.13728443e-02 9.62496579e-01 3.20349813e-01 -2.85452813e-01
-1.24960303e+00 -1.49802804e+00 -2.13628531e-01 1.35425723e+00
6.09715104e-01 -3.43071759e-01 -1.11742437e+00 -4.59331423e-01
-6.79482892e-02 9.37747657e-01 -7.97170877e-01 -4.72025573e-02
-2.95136839e-01 -1.89693555e-01 3.97709727e-01 8.18137288e-01
5.46441853e-01 -9.75470603e-01 -6.89772666e-01 1.79290593e-01
-3.04960221e-01 -1.20270038e+00 -5.19516885e-01 4.51834232e-01
-7.30718970e-01 -8.67021739e-01 -8.63508105e-01 -1.38922393e+00
1.10257196e+00 2.08066493e-01 1.10394168e+00 -1.11519672e-01
-6.19777083e-01 5.04840016e-01 -5.21953166e-01 -8.41379151e-05
-1.00266665e-01 3.62171791e-02 -4.75668788e-01 -1.51244342e-01
4.93848503e-01 -3.16874564e-01 -7.23134458e-01 2.85706073e-01
-9.03824985e-01 3.22546452e-01 1.14247657e-01 7.92786956e-01
9.18975711e-01 -4.02166814e-01 1.79840535e-01 -9.38965261e-01
6.50337338e-02 -4.85125065e-01 -5.60526311e-01 2.11234421e-01
-3.55298966e-01 8.18484053e-02 5.16740203e-01 -2.06432521e-01
-7.69835234e-01 2.32488111e-01 -1.01767682e-01 -6.10017657e-01
-4.33090419e-01 2.40211025e-01 2.98549477e-02 2.39925161e-01
5.07747710e-01 2.60767698e-01 -2.53739417e-01 -5.27322710e-01
8.52820098e-01 9.02221143e-01 8.00274491e-01 -6.83020175e-01
5.62059999e-01 6.02511525e-01 -5.40362716e-01 -7.02794313e-01
-1.15606725e+00 -6.98867917e-01 -8.95015001e-01 -1.03909105e-01
1.42101812e+00 -1.04250956e+00 7.47836381e-02 4.03323650e-01
-1.11154234e+00 -6.58723235e-01 -5.99780023e-01 1.59659490e-01
-9.18003619e-01 1.67381335e-02 -7.40665555e-01 -3.14665973e-01
-1.56019181e-01 -1.21449375e+00 1.46195757e+00 4.31289315e-01
-8.68376642e-02 -1.03597116e+00 -1.74335361e-01 3.40309054e-01
-2.30973940e-02 2.61912376e-01 7.35508978e-01 -4.14318949e-01
-6.73620462e-01 5.41652786e-03 -4.93939221e-01 4.27412003e-01
9.65123698e-02 -2.90852170e-02 -1.19206822e+00 -1.96058527e-01
-6.05534077e-01 -6.22672856e-01 9.48691785e-01 4.40009683e-01
1.34046566e+00 -1.17748335e-01 -4.91957366e-01 8.33663166e-01
1.70243359e+00 3.36353928e-02 5.66163719e-01 2.32108131e-01
7.25231588e-01 4.34284657e-01 7.27897525e-01 2.73623556e-01
4.82636422e-01 3.82869244e-01 3.31782728e-01 -3.63052219e-01
-2.92793036e-01 -5.75092256e-01 4.74600755e-02 4.99895811e-01
5.57002902e-01 -3.81657094e-01 -1.08328986e+00 1.03380406e+00
-1.73723292e+00 -6.15368247e-01 2.42501795e-01 1.96756804e+00
6.70259237e-01 -1.64202541e-01 -2.56657213e-01 -2.64668077e-01
9.89161909e-01 3.74503255e-01 -7.23376215e-01 -5.30067384e-01
3.23097222e-02 3.96083564e-01 8.68752480e-01 5.75994730e-01
-1.37053251e+00 1.61644292e+00 6.16264868e+00 6.68312371e-01
-9.96088684e-01 4.15382594e-01 6.52984321e-01 8.25601891e-02
-3.18571657e-01 3.34061474e-01 -5.91760993e-01 5.37037492e-01
7.84487665e-01 -1.84137851e-01 4.08109695e-01 9.98052001e-01
-1.60372816e-02 -3.24341297e-01 -9.74260390e-01 9.23676252e-01
2.21966520e-01 -1.32030606e+00 3.45423520e-02 -3.70811298e-02
9.46286678e-01 3.04200411e-01 -5.51359355e-02 1.19610667e-01
5.54968417e-01 -1.09812152e+00 9.64158237e-01 3.63197148e-01
1.19285488e+00 -4.38915551e-01 3.19336206e-01 9.53225121e-02
-1.36109388e+00 5.73853925e-02 -5.59391260e-01 2.04932123e-01
3.81744027e-01 1.79657981e-01 -5.93378663e-01 1.80061311e-01
7.16475904e-01 9.39641893e-01 -5.06429076e-01 7.35915124e-01
-5.34291923e-01 5.72160780e-01 -2.14029729e-01 2.18630269e-01
6.77236676e-01 -2.62453228e-01 1.65873170e-01 1.16427529e+00
3.95319700e-01 4.12468724e-02 5.06361544e-01 1.15599847e+00
-3.10145348e-01 5.18645719e-03 -6.40910089e-01 -2.36388639e-01
6.59137249e-01 1.06288683e+00 -1.14538205e+00 -6.70349121e-01
-3.06950182e-01 1.55708051e+00 3.31558704e-01 6.33631110e-01
-9.13216233e-01 -6.76025689e-01 7.00902641e-01 9.63888764e-02
5.21623969e-01 -2.18897730e-01 -3.11175108e-01 -1.01044440e+00
-3.09116185e-01 2.83487979e-02 3.06906998e-01 -1.27228880e+00
-8.73064756e-01 2.99391448e-01 3.26280817e-02 -1.16718197e+00
-1.13091066e-01 -5.85419893e-01 -5.89931130e-01 6.97214007e-01
-1.21588409e+00 -1.39420283e+00 -5.06432235e-01 4.25754517e-01
7.99452782e-01 1.54152885e-01 9.91582215e-01 1.27772942e-01
-3.60800236e-01 3.46105844e-01 2.47917995e-01 5.22773862e-01
5.30825555e-01 -1.30345488e+00 5.18980145e-01 7.25913048e-01
2.89619595e-01 2.30415791e-01 5.60241997e-01 -3.93125772e-01
-6.45686746e-01 -1.45327842e+00 7.59740770e-01 -1.61223233e-01
5.47319055e-01 -7.00014591e-01 -8.34045947e-01 1.07496464e+00
1.00660063e-01 2.82693565e-01 5.88883042e-01 -1.01672523e-01
-4.59433049e-01 1.73728347e-01 -1.12545812e+00 5.36896825e-01
1.12713528e+00 -8.56827199e-01 -6.59624994e-01 5.14860213e-01
9.11637902e-01 -4.00543839e-01 -7.20344722e-01 -1.00989148e-01
2.81019062e-01 -6.32691979e-01 9.84923601e-01 -2.19812110e-01
5.74516058e-01 -3.78893048e-01 -4.22103703e-01 -1.15275240e+00
-2.63665020e-01 1.27821073e-01 4.10593927e-01 1.10837960e+00
3.42812896e-01 -4.96556222e-01 9.36020434e-01 4.22933400e-01
-3.82396758e-01 -6.70084655e-01 -8.44804883e-01 -7.57631361e-01
1.60901025e-01 -3.42104465e-01 4.54455495e-01 8.18709433e-01
-2.63815999e-01 1.42248079e-01 4.24519228e-03 2.99713284e-01
7.35939562e-01 3.02623987e-01 3.82097870e-01 -1.10945952e+00
1.44833386e-01 -2.43921041e-01 -6.30858362e-01 -1.15050602e+00
3.58619899e-01 -1.29986131e+00 4.91363049e-01 -2.11408448e+00
2.19772905e-01 -4.34880495e-01 -2.70272523e-01 8.52768481e-01
1.07025199e-01 6.33143783e-01 -7.33281597e-02 9.66456085e-02
-7.65945435e-01 6.74695790e-01 1.01581669e+00 -3.79960090e-01
1.96671709e-02 -5.78463614e-01 -5.22466838e-01 7.10021079e-01
1.07064426e+00 -5.93538344e-01 -4.84759450e-01 -6.44235492e-01
-3.09238255e-01 -2.83619404e-01 5.74195027e-01 -1.00317883e+00
1.76840961e-01 -9.01351050e-02 1.31281078e-01 -5.12089908e-01
5.22051334e-01 -4.29137558e-01 -1.26312003e-01 2.19856367e-01
-4.70979244e-01 -3.92979056e-01 7.47806728e-02 4.20302778e-01
-3.07707191e-01 -5.02037346e-01 9.95875001e-01 -2.58528143e-01
-1.35566354e+00 3.88380826e-01 -2.94998556e-01 2.91641295e-01
1.03105366e+00 -5.13302743e-01 -1.65887073e-01 -1.74962133e-01
-9.57905650e-01 5.04536271e-01 1.02478313e+00 4.15310144e-01
6.59688830e-01 -1.26684594e+00 -3.32750678e-01 1.08567536e-01
5.77212453e-01 3.51554871e-01 1.40002459e-01 2.55273491e-01
-8.14898193e-01 8.96730199e-02 -1.99070603e-01 -8.26870024e-01
-8.90513361e-01 3.15710306e-01 3.64927679e-01 3.90776962e-01
-1.01303732e+00 1.13062608e+00 5.18939793e-01 -5.98932326e-01
1.41178757e-01 -3.53531450e-01 2.85420507e-01 1.41112609e-02
2.18278468e-01 1.10162059e-02 -4.80136365e-01 -7.99043417e-01
-3.57456356e-01 8.79557848e-01 9.24945995e-02 -2.89743334e-01
1.09236073e+00 -2.18283013e-01 -4.95875953e-03 7.62450993e-01
1.57903945e+00 -6.09411895e-01 -1.51698494e+00 -3.85279566e-01
4.89961207e-02 -4.16954517e-01 1.89206332e-01 -6.16144776e-01
-9.82969105e-01 1.11492920e+00 6.56658649e-01 -3.31376582e-01
9.26849186e-01 5.85034013e-01 8.68717968e-01 2.15957820e-01
4.41898435e-01 -1.48590171e+00 1.54231921e-01 4.67912078e-01
4.60964143e-01 -1.19812012e+00 -2.48191833e-01 -4.51647490e-01
-5.48777282e-01 8.64506781e-01 3.55371296e-01 -3.91320705e-01
5.70525169e-01 1.02525331e-01 2.65614271e-01 -3.82799566e-01
-3.68361235e-01 -6.86092734e-01 7.19225360e-03 5.67199290e-01
3.09571505e-01 2.98973262e-01 3.18831243e-02 1.73812240e-01
-2.52218425e-01 -6.28527477e-02 4.03839946e-01 9.93744373e-01
-8.52684617e-01 -8.07149708e-01 -1.57940581e-01 5.01578987e-01
1.90706495e-02 -1.46182910e-01 -1.45449713e-01 6.42522156e-01
3.08201343e-01 7.65345514e-01 6.52651131e-01 -1.91804126e-01
1.95621531e-02 3.72673064e-01 2.96723425e-01 -1.14886785e+00
5.44313677e-02 2.44514775e-02 -1.06737345e-01 -5.90523183e-01
-2.97924459e-01 -5.22146821e-01 -1.76240039e+00 6.31032959e-02
2.42922734e-02 6.97815837e-03 6.62755668e-01 7.66961932e-01
1.85274675e-01 3.78138721e-01 4.98054504e-01 -8.86033475e-01
4.99895401e-02 -6.98920369e-01 -7.91815042e-01 5.90325356e-01
2.00944334e-01 -7.28277147e-01 -3.01821202e-01 4.91136491e-01] | [9.693638801574707, 0.8477755188941956] |
06167d3d-dfa5-4f70-8dec-8f41bcb98522 | ixa-cogcomp-at-semeval-2023-task-2-context | 2304.10637 | null | https://arxiv.org/abs/2304.10637v3 | https://arxiv.org/pdf/2304.10637v3.pdf | IXA/Cogcomp at SemEval-2023 Task 2: Context-enriched Multilingual Named Entity Recognition using Knowledge Bases | Named Entity Recognition (NER) is a core natural language processing task in which pre-trained language models have shown remarkable performance. However, standard benchmarks like CoNLL 2003 do not address many of the challenges that deployed NER systems face, such as having to classify emerging or complex entities in a fine-grained way. In this paper we present a novel NER cascade approach comprising three steps: first, identifying candidate entities in the input sentence; second, linking the each candidate to an existing knowledge base; third, predicting the fine-grained category for each entity candidate. We empirically demonstrate the significance of external knowledge bases in accurately classifying fine-grained and emerging entities. Our system exhibits robust performance in the MultiCoNER2 shared task, even in the low-resource language setting where we leverage knowledge bases of high-resource languages. | ['Dan Roth', 'Ander Salaberria', 'Oscar Sainz', 'Jon Ander Campos', 'Iker García-Ferrero'] | 2023-04-20 | null | null | null | null | ['named-entity-recognition-ner', 'multilingual-named-entity-recognition'] | ['natural-language-processing', 'natural-language-processing'] | [-4.11918104e-01 -1.59154624e-01 -1.33440465e-01 -4.06122208e-01
-1.02715087e+00 -9.80695069e-01 7.29405463e-01 5.27322412e-01
-1.05785298e+00 9.06223834e-01 6.79979205e-01 -1.72842801e-01
2.03789137e-02 -8.98282170e-01 -5.15828967e-01 1.15873724e-01
-1.52277872e-01 5.49932301e-01 1.40794545e-01 -2.48247460e-01
-9.65944678e-02 6.73749447e-01 -1.03165555e+00 3.26024741e-01
7.27428854e-01 7.69815147e-01 -2.09755585e-01 7.04285502e-01
-4.27198768e-01 9.52124715e-01 -6.57959104e-01 -7.95526206e-01
6.96293116e-02 6.20746724e-02 -1.26968050e+00 -6.12284958e-01
2.27459639e-01 3.06216367e-02 -1.54061243e-01 8.77407432e-01
4.66264784e-01 2.04639941e-01 6.46770954e-01 -7.05561340e-01
-7.75833666e-01 9.41186249e-01 -1.70372814e-01 5.49376845e-01
3.19972247e-01 -2.41443023e-01 1.37643147e+00 -1.32285905e+00
8.66800129e-01 9.35473144e-01 8.12976539e-01 5.03659844e-01
-9.20937479e-01 -6.91838682e-01 6.71941638e-02 -5.05045876e-02
-1.63669550e+00 -4.91481364e-01 9.11266357e-02 -3.96946728e-01
1.43772352e+00 -2.59002335e-02 -1.43135324e-01 8.04295123e-01
-1.02021806e-01 4.44997519e-01 1.00844812e+00 -4.28165525e-01
6.78656399e-02 5.67671508e-02 5.25445819e-01 5.38730323e-01
4.52648520e-01 -2.25723520e-01 -5.63598096e-01 -2.72422045e-01
2.63461530e-01 -2.92749554e-01 9.84215885e-02 4.41202253e-01
-1.23353636e+00 5.80223560e-01 4.39079970e-01 8.51845801e-01
-7.68798470e-01 2.26613265e-02 4.10461694e-01 2.16454685e-01
5.25692105e-01 8.67881238e-01 -1.07741439e+00 -5.72142266e-02
-9.97977138e-01 -2.24899650e-01 1.12337518e+00 1.02177823e+00
8.09483111e-01 -2.03420326e-01 -4.86893594e-01 8.08314979e-01
4.25169282e-02 3.47144127e-01 4.93556619e-01 -4.13594753e-01
7.64586151e-01 7.20715582e-01 4.16873485e-01 -7.00037479e-01
-6.30038559e-01 -3.08014691e-01 -7.66617894e-01 -5.37420928e-01
3.43026340e-01 -6.73619986e-01 -9.93364453e-01 1.66527867e+00
3.48042190e-01 1.24744140e-01 3.96391451e-01 4.07918572e-01
1.08474910e+00 6.09838486e-01 7.77677655e-01 2.16868311e-01
1.67281663e+00 -8.05216134e-01 -4.87522453e-01 -3.95842105e-01
6.40518725e-01 -8.63438368e-01 6.00804806e-01 -2.05863506e-01
-7.83556104e-01 -3.80404323e-01 -5.43430984e-01 -2.09558442e-01
-8.68407905e-01 5.56780636e-01 5.36035895e-01 5.31997383e-01
-1.00014210e+00 4.49667215e-01 -6.94110394e-01 -5.11019170e-01
1.98021665e-01 3.14732820e-01 -7.63126731e-01 3.69783007e-02
-1.54006684e+00 1.15587401e+00 7.27384150e-01 1.28430068e-01
-5.32095492e-01 -8.98715973e-01 -9.59251046e-01 4.23990786e-01
2.36900151e-01 -5.76291203e-01 1.23515868e+00 -5.87817252e-01
-9.26230490e-01 1.00187719e+00 -5.05320072e-01 -5.17405987e-01
7.48167410e-02 -4.49542642e-01 -8.70174587e-01 -7.40138888e-02
5.98975122e-01 4.77402747e-01 1.56289175e-01 -7.18649387e-01
-9.82948303e-01 -1.55745581e-01 1.03700273e-01 1.98565945e-02
-5.15052497e-01 5.44471502e-01 -2.77586311e-01 -6.00157857e-01
-4.08355713e-01 -5.63044727e-01 -3.19690615e-01 -7.19182074e-01
-5.02720773e-01 -7.00656235e-01 1.63330808e-01 -7.26226628e-01
1.47231758e+00 -2.04244661e+00 -3.90908509e-01 5.03534377e-02
2.44205654e-01 4.29192901e-01 -1.30402178e-01 6.49039090e-01
-2.13269234e-01 6.38326526e-01 2.08752409e-01 -1.74484506e-01
1.17367871e-01 -2.13105857e-01 -5.42672992e-01 -1.26155511e-01
7.73011327e-01 1.09047604e+00 -1.09551227e+00 -4.45231199e-01
-3.36094856e-01 4.71450478e-01 -2.48402715e-01 2.17968807e-01
1.08524889e-01 9.51548070e-02 -5.44524848e-01 4.75493014e-01
2.55920351e-01 -4.47637916e-01 1.31338969e-01 -3.25211793e-01
-4.01287407e-01 7.10685253e-01 -1.30153728e+00 1.05569279e+00
-7.00360775e-01 4.38355505e-01 -1.67431101e-01 -4.87331331e-01
8.11001182e-01 5.50853789e-01 5.94827123e-02 -4.36320305e-01
-1.67430058e-01 4.12414908e-01 -2.30459079e-01 -2.12973848e-01
7.88924634e-01 -3.40493292e-01 -6.00221992e-01 3.09892207e-01
5.10796487e-01 5.28037012e-01 4.89347100e-01 2.89061755e-01
1.49944103e+00 -3.35489213e-01 7.14148223e-01 -1.69621810e-01
4.14942235e-01 1.99069232e-01 7.71399915e-01 8.47397745e-01
-1.94865003e-01 3.00110430e-01 1.24267034e-01 -3.96069348e-01
-9.63106453e-01 -9.15304482e-01 -1.83891729e-01 1.43492341e+00
-2.01204777e-01 -4.68614846e-01 -3.74259979e-01 -8.89174879e-01
-1.73679605e-01 7.23200679e-01 -5.16804814e-01 2.37141907e-01
-6.35671735e-01 -7.29825258e-01 1.26723444e+00 7.93002009e-01
5.62173367e-01 -1.21292257e+00 -1.34431422e-01 5.19581497e-01
-3.58805478e-01 -1.47631621e+00 -5.06676257e-01 4.20458108e-01
-3.55874151e-01 -9.30380583e-01 -5.04119396e-01 -8.74815285e-01
5.10365009e-01 -6.61071464e-02 1.50679672e+00 2.83981077e-02
-9.20673013e-02 3.27012926e-01 -4.14368898e-01 -2.53337383e-01
-3.00145209e-01 4.63161379e-01 4.82450202e-02 -6.31472990e-02
8.28006268e-01 -1.24971323e-01 -3.68237913e-01 1.98110580e-01
-6.55416369e-01 -3.94373447e-01 7.42443442e-01 7.95424700e-01
4.11119372e-01 1.69353470e-01 8.94441724e-01 -1.17653763e+00
5.91030955e-01 -7.64017582e-01 -3.22439045e-01 6.83948219e-01
-3.59300137e-01 1.57924891e-01 9.67502773e-01 -7.85525292e-02
-1.35655212e+00 2.88516842e-03 -4.37927485e-01 3.22668433e-01
-6.26841307e-01 8.23355556e-01 -1.34081706e-01 2.91431963e-01
7.12656260e-01 -5.01756892e-02 -1.26995409e+00 -6.27353072e-01
5.13999343e-01 7.70012736e-01 6.24407828e-01 -6.81197464e-01
1.04937923e+00 2.28819758e-01 -2.94897705e-01 -6.84019208e-01
-1.29365885e+00 -9.34266150e-01 -1.02103829e+00 3.54299188e-01
1.11449599e+00 -1.41145301e+00 -4.21354473e-01 2.47813225e-01
-1.33829701e+00 1.00426245e-02 -3.42842996e-01 5.80990732e-01
1.30389914e-01 7.82258622e-03 -1.04918182e+00 -3.90824646e-01
-6.98433220e-01 -3.97685200e-01 9.38824952e-01 5.10724485e-01
-2.76568919e-01 -1.07039368e+00 3.51884902e-01 3.56533587e-01
4.28053528e-01 -7.05398917e-02 7.98625767e-01 -1.41074634e+00
-1.77459359e-01 -3.87309790e-01 -3.13206524e-01 5.65999933e-02
9.15042385e-02 -1.02042787e-01 -8.54298711e-01 -4.37550619e-02
-5.19186080e-01 -4.04753000e-01 8.05554271e-01 -2.21598536e-01
4.21117097e-01 -2.13900313e-01 -3.88302475e-01 3.03706437e-01
1.34469199e+00 -1.32871628e-01 2.29440689e-01 5.69029212e-01
6.27565026e-01 3.98719877e-01 4.85054523e-01 1.23286687e-01
7.57183075e-01 2.36202508e-01 -3.30392480e-01 -2.55001456e-01
-6.90246150e-02 -4.34184343e-01 2.46886328e-01 8.06851685e-01
-7.19944686e-02 -3.66003454e-01 -1.32020664e+00 8.66945148e-01
-1.52860737e+00 -1.20988190e+00 2.85000028e-03 1.77334535e+00
1.06116748e+00 8.58024210e-02 -2.13581041e-01 -5.07283151e-01
9.96233463e-01 -3.70861106e-02 -3.26079696e-01 -3.16775411e-01
-3.72266859e-01 5.92714012e-01 5.18155396e-01 1.45893171e-01
-1.52878928e+00 1.35946882e+00 6.15985775e+00 7.00754881e-01
-9.65350211e-01 2.05387339e-01 4.35129106e-01 4.06534463e-01
-9.89647955e-03 7.97330216e-02 -1.47772622e+00 1.27124876e-01
1.28367019e+00 -4.59805220e-01 1.20536804e-01 7.32833266e-01
-7.26870894e-02 2.96635956e-01 -1.05730379e+00 5.59303761e-01
-1.09346278e-01 -1.37352121e+00 2.89131775e-02 -2.89991558e-01
8.55775535e-01 7.63984680e-01 -5.45474887e-01 7.07009196e-01
9.82520342e-01 -1.06611276e+00 5.63268721e-01 4.27464634e-01
8.28716278e-01 -6.66244984e-01 8.46747458e-01 4.13061500e-01
-1.59339869e+00 9.26791504e-03 -3.43674600e-01 2.39774197e-01
3.42387736e-01 6.80436730e-01 -9.11366403e-01 6.98429406e-01
7.71295011e-01 3.61011297e-01 -7.18445301e-01 1.12930620e+00
-5.21546841e-01 8.72759879e-01 -3.93501014e-01 1.09860264e-01
2.85334438e-01 3.84123743e-01 2.81248808e-01 1.85673118e+00
2.99654514e-01 3.55889171e-01 9.03847069e-02 6.14012182e-01
-8.37586701e-01 3.41559052e-01 -4.09815699e-01 -3.94145340e-01
7.99271345e-01 1.69560790e+00 -7.67136157e-01 -4.21488166e-01
-6.79160595e-01 9.26152706e-01 9.73147452e-01 2.13913143e-01
-3.82975221e-01 -7.32707381e-01 4.94404376e-01 -2.69346237e-01
6.13999844e-01 -3.01482201e-01 5.45052346e-03 -1.68952322e+00
-1.30894437e-01 -5.30117750e-01 8.03944826e-01 -5.60131371e-01
-1.75720215e+00 9.26946342e-01 -4.56199288e-01 -7.53141105e-01
-2.04575673e-01 -5.81332743e-01 -8.25681806e-01 9.57939148e-01
-1.67462230e+00 -1.31005836e+00 1.74244251e-05 5.24437845e-01
3.57459098e-01 4.65008654e-02 1.12079489e+00 7.32339740e-01
-6.74962461e-01 6.40922308e-01 8.79778862e-02 1.03502822e+00
1.00704050e+00 -1.41105556e+00 7.72706866e-01 1.12920868e+00
5.07070720e-01 9.95445251e-01 3.12131166e-01 -7.58008897e-01
-9.38415051e-01 -1.51181293e+00 1.82864308e+00 -8.29889178e-01
1.05596435e+00 -4.37762201e-01 -8.74992728e-01 9.47883785e-01
8.50345194e-02 3.03310305e-01 1.01861894e+00 5.83497286e-01
-7.75183201e-01 1.25121564e-01 -9.10270691e-01 3.57370824e-01
9.04032290e-01 -8.90064418e-01 -9.99391556e-01 1.39562815e-01
6.90829933e-01 -1.58545658e-01 -1.26510084e+00 3.39748055e-01
2.20874995e-01 -1.60517424e-01 8.25147510e-01 -1.25396156e+00
2.73255736e-01 -3.74573618e-01 -6.23494163e-02 -1.42515016e+00
-4.79598939e-01 -3.60437453e-01 1.51390076e-01 1.92041814e+00
9.93479311e-01 -3.92190307e-01 4.25668120e-01 8.61388683e-01
9.72710326e-02 -2.76126176e-01 -7.09146619e-01 -7.64879107e-01
2.13883966e-01 -2.10689098e-01 6.75151825e-01 1.29575658e+00
-7.09392205e-02 8.47285211e-01 -1.57098025e-01 4.45857465e-01
3.48111421e-01 1.09951518e-01 4.29235727e-01 -1.24542117e+00
5.67544177e-02 -1.28020078e-01 -2.75969177e-01 -6.85686409e-01
4.66746509e-01 -1.11647761e+00 1.23180591e-01 -1.58575475e+00
3.79187524e-01 -6.50817394e-01 -6.07937217e-01 7.95620143e-01
-5.57093501e-01 2.71174461e-01 1.23272955e-01 2.79577672e-01
-1.04780138e+00 1.98247597e-01 2.86452711e-01 -2.85380939e-03
-5.63678518e-02 -2.41472155e-01 -9.40365553e-01 7.75795639e-01
6.22883916e-01 -7.41598904e-01 3.46959263e-01 -6.81090176e-01
6.66350663e-01 -3.30299318e-01 2.50394881e-01 -7.37630129e-01
6.51944339e-01 -5.36258966e-02 6.46871924e-01 -3.89287174e-01
-2.56937385e-01 -6.13977909e-01 -2.18561571e-02 7.04311058e-02
-4.03579086e-01 3.21446955e-01 1.94181353e-01 5.52015424e-01
-5.11608899e-01 -2.59716004e-01 6.51794255e-01 -1.60779864e-01
-1.18337357e+00 3.33844274e-01 -3.76588881e-01 6.85200810e-01
8.01376164e-01 3.50239813e-01 -5.88471115e-01 1.27304152e-01
-8.15336764e-01 3.27426136e-01 9.59967673e-02 5.26278794e-01
7.05584586e-02 -1.12381577e+00 -8.93747866e-01 -2.29706943e-01
2.85347313e-01 -3.13190311e-01 1.30139709e-01 4.68258053e-01
-2.94469982e-01 6.17017210e-01 3.77745554e-02 1.46588981e-01
-9.44820762e-01 3.37077826e-01 2.98161358e-01 -9.09271777e-01
-3.22011292e-01 8.65425467e-01 1.02540098e-01 -7.95752943e-01
-9.34431329e-02 1.02101691e-01 -5.77647507e-01 2.37904504e-01
6.23788476e-01 2.20904827e-01 3.09464186e-01 -1.01413357e+00
-6.91203594e-01 3.18580776e-01 -1.13266401e-01 -1.46895451e-02
1.50823081e+00 -1.95828527e-01 -1.89182028e-01 2.06052512e-01
9.82606232e-01 3.53744954e-01 -5.08771122e-01 -5.62926948e-01
8.85563254e-01 1.90526053e-01 -1.15948632e-01 -1.02570999e+00
-6.01350963e-01 6.52484238e-01 9.66299027e-02 7.06433803e-02
9.12817836e-01 2.37405732e-01 9.06715691e-01 8.07678938e-01
4.86784786e-01 -1.15648925e+00 -5.13866723e-01 1.19271469e+00
3.63529503e-01 -1.23464274e+00 -3.59024048e-01 -2.36343205e-01
-5.93805850e-01 1.02241004e+00 6.16214454e-01 -1.62351429e-02
8.03725481e-01 3.46708089e-01 1.09885469e-01 -2.84555048e-01
-9.62872028e-01 -6.14914656e-01 5.45349240e-01 1.76359728e-01
7.70945966e-01 1.43633485e-01 -3.76854002e-01 1.09583533e+00
-2.04591781e-01 -1.24859571e-01 3.36723119e-01 7.67895460e-01
-3.75111789e-01 -1.08011031e+00 -1.17309364e-02 4.45851058e-01
-1.19809330e+00 -6.45005524e-01 -4.66963232e-01 6.35759890e-01
1.99701652e-01 1.03049374e+00 -3.02140750e-02 -1.67301252e-01
7.26074815e-01 2.76431978e-01 -1.44042328e-01 -1.09296978e+00
-1.21410644e+00 -4.07540262e-01 5.43954492e-01 -3.46024930e-01
-3.80640119e-01 -4.00205433e-01 -1.44631064e+00 -1.30156010e-01
-3.77862751e-01 6.35221601e-01 5.37625551e-01 1.15708005e+00
7.46153355e-01 4.72206473e-02 4.27323341e-01 -3.12593281e-01
-4.43949938e-01 -8.32373202e-01 -4.12485540e-01 5.70718765e-01
2.32070703e-02 -2.71691740e-01 -2.41101772e-01 1.45560026e-01] | [9.657301902770996, 9.49191951751709] |
8d19f12d-9bdb-428a-9036-a98d2b660db1 | regularisation-can-mitigate-poisoning-attacks | 2003.00040 | null | https://arxiv.org/abs/2003.00040v2 | https://arxiv.org/pdf/2003.00040v2.pdf | Regularisation Can Mitigate Poisoning Attacks: A Novel Analysis Based on Multiobjective Bilevel Optimisation | Machine Learning (ML) algorithms are vulnerable to poisoning attacks, where a fraction of the training data is manipulated to deliberately degrade the algorithms' performance. Optimal poisoning attacks, which can be formulated as bilevel optimisation problems, help to assess the robustness of learning algorithms in worst-case scenarios. However, current attacks against algorithms with hyperparameters typically assume that these hyperparameters remain constant ignoring the effect the attack has on them. We show that this approach leads to an overly pessimistic view of the robustness of the algorithms. We propose a novel optimal attack formulation that considers the effect of the attack on the hyperparameters by modelling the attack as a multiobjective bilevel optimisation problem. We apply this novel attack formulation to ML classifiers using $L_2$ regularisation and show that, in contrast to results previously reported, $L_2$ regularisation enhances the stability of the learning algorithms and helps to mitigate the attacks. Our empirical evaluation on different datasets confirms the limitations of previous strategies, evidences the benefits of using $L_2$ regularisation to dampen the effect of poisoning attacks and shows how the regularisation hyperparameter increases with the fraction of poisoning points. | ['Phillippa Spencer', 'Luis Muñoz-González', 'Javier Carnerero-Cano', 'Emil C. Lupu'] | 2020-02-28 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [ 2.03846410e-01 2.88104434e-02 -3.96567881e-02 2.59684324e-01
-4.74284798e-01 -8.54087174e-01 9.51363385e-01 2.85653681e-01
-8.05907428e-01 7.36949384e-01 -1.60328284e-01 -4.33281958e-01
-5.96931875e-01 -7.10867882e-01 -8.71246457e-01 -1.26643407e+00
-3.43727976e-01 2.77802825e-01 2.52014786e-01 -1.57329589e-01
6.59242570e-01 7.23226011e-01 -1.36129487e+00 -3.26402625e-03
2.99703807e-01 5.31473577e-01 -7.83872962e-01 9.12664950e-01
2.35629469e-01 6.49738312e-01 -1.31451702e+00 -4.92156595e-01
6.58253789e-01 -2.24181399e-01 -7.18222737e-01 -1.55674443e-01
2.79704899e-01 1.43259659e-01 3.80708873e-02 1.10457075e+00
6.99911475e-01 -1.49444342e-01 6.45323515e-01 -1.57145631e+00
3.30866635e-01 7.05761552e-01 -5.28974771e-01 3.61453086e-01
2.00576708e-01 5.99401712e-01 7.31971622e-01 -2.57390961e-02
2.60083109e-01 1.37590516e+00 8.17687988e-01 4.98320550e-01
-1.55512905e+00 -9.01142776e-01 -4.51770388e-02 -1.37727201e-01
-1.36711955e+00 -4.80138123e-01 6.04992807e-01 -2.67682403e-01
7.64623523e-01 5.42203784e-01 2.21246377e-01 9.95101511e-01
3.24321598e-01 4.88300085e-01 1.09543622e+00 -3.96382749e-01
4.30927873e-01 2.30602130e-01 3.17183398e-02 4.18765008e-01
5.96729875e-01 4.41196769e-01 -3.07206124e-01 -6.98888659e-01
1.68696195e-01 -4.73763764e-01 -3.96069974e-01 -5.39513826e-01
-8.73217225e-01 1.13564682e+00 2.60819793e-01 1.53588369e-01
-1.62587166e-01 4.00148541e-01 6.51879907e-01 5.90136051e-01
1.25727966e-01 1.16545796e+00 -4.15101886e-01 -7.81661645e-02
-6.32535636e-01 2.94070423e-01 1.02300096e+00 2.12228507e-01
2.31817052e-01 1.63476914e-01 1.68213502e-01 4.54459816e-01
3.81971121e-01 3.22054416e-01 2.25180209e-01 -6.47003412e-01
4.87747878e-01 5.33050179e-01 7.08521605e-02 -9.24777329e-01
-3.53972793e-01 -6.15015745e-01 -4.51753527e-01 6.48218036e-01
9.10430908e-01 -3.05204511e-01 -5.63658118e-01 1.90829062e+00
5.91592968e-01 2.62013912e-01 3.09458792e-01 4.47525620e-01
2.55980879e-01 4.22476530e-01 3.12794149e-01 -3.93100142e-01
1.02399731e+00 -2.20201150e-01 -4.89094257e-01 7.73339719e-02
9.88864422e-01 -7.60865092e-01 9.40896809e-01 6.97427213e-01
-1.04417026e+00 3.20065431e-02 -1.49649024e+00 8.42407227e-01
-5.13304532e-01 -7.01679647e-01 2.94671595e-01 1.47426629e+00
-5.19634962e-01 6.93726361e-01 -7.00408041e-01 8.72325525e-03
5.17889619e-01 9.61973250e-01 -1.04580298e-01 3.17865014e-01
-1.27740979e+00 1.04753423e+00 5.02900064e-01 -1.78155646e-01
-9.29614782e-01 -9.07855690e-01 -5.04671216e-01 -2.57818341e-01
5.44516027e-01 -3.75139982e-01 7.11548626e-01 -5.92909932e-01
-1.32195210e+00 5.58317244e-01 5.42474389e-01 -7.42567658e-01
9.76935327e-01 -7.67665356e-02 -3.04787755e-02 -3.48665863e-02
-5.42755723e-01 2.57566065e-01 1.00441611e+00 -1.27481329e+00
-3.09048861e-01 -2.56631166e-01 2.01488450e-01 1.21297814e-01
-3.58350098e-01 8.20047557e-02 1.91582590e-01 -6.11419141e-01
-4.76494849e-01 -1.10797429e+00 -3.93246502e-01 -4.43582505e-01
-3.77249509e-01 1.44714326e-01 1.10529685e+00 -2.16220319e-02
1.29197872e+00 -2.26711297e+00 1.80983290e-01 5.32799125e-01
8.41060877e-02 6.73752546e-01 -2.48713978e-02 5.45699656e-01
-1.51290938e-01 4.08147991e-01 -3.40227038e-01 7.49491341e-03
8.73151794e-02 3.71246666e-01 -4.58488047e-01 1.05201519e+00
-4.67063822e-02 4.00929242e-01 -6.22240067e-01 -1.13586947e-01
2.53690392e-01 5.58563113e-01 -5.34369826e-01 1.08509354e-01
-1.81004554e-01 7.09903166e-02 -3.01861823e-01 9.37251225e-02
4.24091250e-01 2.14313373e-01 1.19871520e-01 -1.70403756e-02
2.04497963e-01 7.90460184e-02 -1.51048338e+00 6.12051785e-01
-1.31262526e-01 3.76373798e-01 1.35406479e-01 -1.03196907e+00
8.25834334e-01 2.89545685e-01 4.13177103e-01 -3.19677293e-01
5.22684097e-01 3.90724139e-03 3.99136335e-01 -1.60700217e-01
-1.74404308e-03 -3.22493672e-01 -2.05874145e-01 7.89355576e-01
-4.46757674e-01 -4.38835084e-01 1.22562043e-01 7.24384934e-02
1.25391257e+00 -4.28665668e-01 3.84798318e-01 -5.18270135e-01
7.30001390e-01 -3.71660143e-02 2.10614815e-01 9.96503770e-01
-7.42413774e-02 1.04717538e-01 9.19647992e-01 -4.39232022e-01
-1.02065516e+00 -7.35328853e-01 -3.52226287e-01 8.67659628e-01
6.36114776e-02 -5.66545248e-01 -1.08293819e+00 -9.85011697e-01
2.04574332e-01 8.74394059e-01 -7.58013606e-01 -6.47785723e-01
-6.51589036e-01 -1.76155460e+00 1.35391712e+00 1.30698755e-01
5.09718597e-01 -8.30521226e-01 -1.15934873e+00 8.60816166e-02
3.93662930e-01 -5.91772974e-01 -1.87252462e-01 4.38063979e-01
-8.08342874e-01 -1.46641147e+00 -1.54987155e-02 -5.00950031e-02
5.20518422e-01 -4.00079161e-01 8.72371435e-01 5.54924548e-01
-4.31023598e-01 3.44966829e-01 -1.98106542e-01 -7.29031324e-01
-8.98628652e-01 3.12349293e-02 1.78561360e-01 -1.41232207e-01
2.74466306e-01 -3.21896046e-01 -2.05029145e-01 4.51986820e-01
-1.28580642e+00 -8.01576793e-01 1.86769351e-01 8.71322870e-01
3.53469588e-02 6.01710916e-01 4.75395560e-01 -1.08842802e+00
9.69723225e-01 -4.02108371e-01 -6.66280568e-01 1.08364463e-01
-8.09310257e-01 5.44346452e-01 6.57190084e-01 -1.07053876e+00
-3.39203656e-01 1.43262502e-02 4.12501255e-03 -3.61795127e-01
-1.18460037e-01 1.15930974e-01 -4.54695344e-01 -5.78174889e-01
1.02098620e+00 -4.30947095e-02 1.18026003e-01 -2.38211751e-01
2.42365271e-01 3.53343815e-01 1.07727900e-01 -6.64604068e-01
1.07829094e+00 2.84536600e-01 5.42368054e-01 -8.83540869e-01
-3.25509846e-01 -3.62186171e-02 -3.05892438e-01 -1.12089671e-01
2.15323895e-01 -3.59751463e-01 -1.02415645e+00 4.21355546e-01
-5.90759933e-01 -4.84533966e-01 -2.75132388e-01 1.04411617e-01
-5.05652308e-01 4.95048195e-01 -1.59507036e-01 -7.54304290e-01
-2.12822318e-01 -1.38357818e+00 3.58977944e-01 -1.56149969e-01
-4.84989971e-01 -1.19562376e+00 1.04723074e-01 3.00166577e-01
3.86474133e-01 6.30681336e-01 1.14267564e+00 -1.13774204e+00
-1.41535193e-01 -2.85813838e-01 3.91197175e-01 2.53876060e-01
-2.12003633e-01 6.10065646e-02 -8.50280702e-01 -7.45474041e-01
3.53091955e-01 -3.51924509e-01 5.72271883e-01 -8.68299976e-03
7.70232320e-01 -8.74982417e-01 -2.20670849e-01 5.96819162e-01
1.36129606e+00 1.45828083e-01 7.33318806e-01 9.20558214e-01
3.60031068e-01 5.45225918e-01 4.41679388e-01 5.02975047e-01
-4.24895465e-01 5.86580455e-01 8.34564984e-01 -6.37500137e-02
3.18989784e-01 1.74216524e-01 4.94007885e-01 -2.61911988e-01
2.49748573e-01 -1.23471290e-01 -9.08931255e-01 1.35542810e-01
-1.44038785e+00 -9.07079816e-01 -1.34934396e-01 2.56080198e+00
1.02914965e+00 7.36305356e-01 5.32808900e-01 7.45556653e-01
4.76933271e-01 1.67358950e-01 -5.62176883e-01 -8.28828335e-01
7.83085600e-02 2.28204608e-01 1.04195237e+00 6.34909391e-01
-1.06309760e+00 6.62910879e-01 7.22427464e+00 1.03122449e+00
-1.18359649e+00 -1.90054089e-01 3.53613466e-01 -4.84443098e-01
1.32290423e-01 -1.01590343e-01 -6.84331834e-01 6.02228284e-01
1.11718011e+00 -1.94688275e-01 4.06352043e-01 4.56278473e-01
-8.98527130e-02 -6.96925372e-02 -1.08400512e+00 5.23484945e-01
2.74935216e-02 -9.67761695e-01 -3.47121246e-02 4.50516135e-01
3.68394852e-01 -2.06305966e-01 3.34616899e-01 1.24140479e-01
8.45310211e-01 -1.24319959e+00 3.62965494e-01 -1.36197992e-02
1.98567167e-01 -1.36305416e+00 6.83991253e-01 5.45104682e-01
-6.26895130e-01 -3.70463520e-01 8.22161790e-03 -1.43212294e-02
-3.81194562e-01 2.45746776e-01 -1.12652969e+00 1.76978126e-01
4.35379356e-01 8.83382000e-03 -9.52230573e-01 1.04135633e+00
-2.63177693e-01 8.38074446e-01 -6.65686965e-01 -6.78287148e-02
3.93026412e-01 6.55712336e-02 9.29274380e-01 1.15681779e+00
-3.40727746e-01 -8.67677629e-02 4.61927280e-02 3.10851276e-01
2.27630064e-01 1.21787138e-01 -6.93276763e-01 -4.28995751e-02
4.92249459e-01 6.78389907e-01 -7.92318285e-01 -2.52022240e-02
1.80741847e-01 2.16958568e-01 7.87634477e-02 8.59855860e-02
-6.89122081e-01 -4.40262824e-01 7.10710049e-01 1.26804635e-01
5.12628667e-02 -3.43066677e-02 -3.68093878e-01 -7.08446383e-01
-3.00380230e-01 -1.72136235e+00 8.67808759e-01 7.10056722e-02
-1.01904273e+00 2.46680081e-01 2.14522213e-01 -8.88582945e-01
-2.23091170e-01 -5.09010553e-01 -5.38915813e-01 5.70570409e-01
-9.65936422e-01 -6.53936684e-01 4.22893494e-01 5.16051531e-01
1.45393565e-01 -2.50610322e-01 6.86598301e-01 -3.30268413e-01
-7.32069552e-01 1.03309333e+00 2.43527025e-01 -1.46616414e-01
5.32981992e-01 -1.16647589e+00 -3.60467136e-02 8.96394253e-01
3.96428220e-02 8.01597953e-01 1.38490045e+00 -4.81435388e-01
-1.25875580e+00 -7.84354150e-01 2.38693401e-01 -7.92923987e-01
6.29667997e-01 -3.42247993e-01 -1.02049923e+00 2.86143273e-01
4.53551523e-02 -2.61242986e-01 7.63303518e-01 5.08333743e-02
-6.18635356e-01 8.89149681e-02 -1.64082694e+00 7.83397019e-01
4.32688862e-01 -1.79250449e-01 -5.71142852e-01 2.28388429e-01
2.28068113e-01 -1.49806514e-01 -1.09651256e+00 7.04529285e-01
3.69564235e-01 -8.21015716e-01 1.33574688e+00 -7.73147285e-01
-2.81416923e-01 -1.91865072e-01 -6.09902069e-02 -1.25305891e+00
1.35264322e-01 -8.45003843e-01 -2.15269029e-01 9.34017599e-01
2.66795516e-01 -8.66417527e-01 8.02866578e-01 5.07164955e-01
6.84336364e-01 -6.32329941e-01 -1.12728965e+00 -7.85154760e-01
4.60240483e-01 -1.93617061e-01 5.72421193e-01 1.09359002e+00
6.48591965e-02 8.61732513e-02 -3.33573937e-01 4.13202941e-01
9.05102551e-01 -6.41090989e-01 1.07053971e+00 -1.08962762e+00
-5.30882120e-01 -7.93544233e-01 -5.98850071e-01 -1.92388117e-01
1.75179586e-01 -5.52776933e-01 -2.22303227e-01 -3.51939499e-01
-1.01036780e-01 -5.49739957e-01 -2.40932420e-01 5.24652123e-01
-3.41879278e-01 3.61917645e-01 3.97559881e-01 1.96339712e-01
-3.82021248e-01 3.34734656e-02 6.53196156e-01 -4.60625179e-02
-3.16840321e-01 1.90068975e-01 -5.60780346e-01 5.64588726e-01
8.54633451e-01 -9.24876750e-01 -4.00260359e-01 9.96199101e-02
4.83074099e-01 -4.53665674e-01 3.05732876e-01 -8.94705653e-01
2.27651566e-01 -2.58442014e-01 2.16436863e-01 -5.01805581e-02
1.79757118e-01 -8.55210721e-01 3.46947968e-01 1.13197851e+00
-5.92175424e-01 -4.08373661e-02 4.51253265e-01 4.42181855e-01
3.43889147e-01 -5.74567258e-01 1.24990976e+00 6.26535097e-04
2.19027579e-01 -1.36031419e-01 -5.85136831e-01 9.06291753e-02
1.39277577e+00 -2.11727649e-01 -2.09485620e-01 -5.12746349e-02
-5.10086834e-01 3.64308625e-01 7.09200263e-01 1.17046081e-01
1.28770158e-01 -7.12778032e-01 -6.70194030e-01 3.45914662e-01
-2.11366698e-01 -2.81299412e-01 -3.52625072e-01 6.21240616e-01
-4.67689931e-01 1.26800478e-01 -1.12440363e-01 -4.81034636e-01
-1.86969304e+00 8.73544753e-01 7.01725006e-01 -3.61162663e-01
-3.51327211e-01 7.12563396e-01 -1.89071953e-01 -2.77901649e-01
6.38915479e-01 4.58231300e-01 -1.90878749e-01 2.32959449e-01
4.88962442e-01 5.56572080e-01 8.45835805e-02 -4.07692820e-01
-4.19523567e-01 4.78035837e-01 -3.32220793e-01 -1.13624245e-01
1.16922951e+00 2.31627539e-01 3.98595631e-02 1.42993048e-01
9.21364129e-01 1.05339132e-01 -1.23313689e+00 -4.03084457e-02
2.51888543e-01 -5.58286965e-01 2.01423973e-01 -7.52649963e-01
-7.78204262e-01 7.31929719e-01 4.39060837e-01 4.77493703e-01
9.61063981e-01 -4.29711729e-01 1.46656513e-01 5.01041830e-01
1.67138681e-01 -8.73569012e-01 1.96406111e-01 1.33870110e-01
5.27116776e-01 -7.01315641e-01 4.30446535e-01 -2.31271520e-01
-3.83608073e-01 9.77993906e-01 3.64224225e-01 -2.34733209e-01
3.27563077e-01 5.52533567e-01 9.95713323e-02 -1.57429114e-01
-7.78317988e-01 2.85457820e-01 -1.36865884e-01 4.53504264e-01
-1.47247940e-01 -2.69447416e-01 -3.72146904e-01 -1.68901067e-02
-3.41909528e-01 -4.68562603e-01 7.22367823e-01 1.11192369e+00
-4.17945385e-01 -1.37417054e+00 -1.12559152e+00 1.18451640e-01
-8.27984214e-01 1.49972022e-01 -6.76953077e-01 1.17852807e+00
1.14788480e-01 1.01018524e+00 -3.46679807e-01 -1.85556516e-01
3.88218790e-01 3.27785760e-01 3.86637062e-01 -2.69840121e-01
-1.19372618e+00 -1.20730713e-01 6.97520077e-02 -3.76422048e-01
-9.81022045e-02 -7.70461738e-01 -8.50625575e-01 -5.68581820e-01
-5.85197389e-01 5.06245136e-01 6.75245762e-01 9.51894939e-01
1.84951872e-02 3.59514415e-01 7.52186596e-01 -5.15134573e-01
-1.32414997e+00 -5.25250077e-01 -4.46469605e-01 5.14896989e-01
4.64443266e-01 -7.20015526e-01 -1.05476737e+00 -2.57182330e-01] | [5.794511318206787, 7.540472507476807] |
a50231f4-010d-4f0d-92e5-d7aaa9246f35 | exploiting-programmatic-behavior-of-llms-dual | 2302.05733 | null | https://arxiv.org/abs/2302.05733v1 | https://arxiv.org/pdf/2302.05733v1.pdf | Exploiting Programmatic Behavior of LLMs: Dual-Use Through Standard Security Attacks | Recent advances in instruction-following large language models (LLMs) have led to dramatic improvements in a range of NLP tasks. Unfortunately, we find that the same improved capabilities amplify the dual-use risks for malicious purposes of these models. Dual-use is difficult to prevent as instruction-following capabilities now enable standard attacks from computer security. The capabilities of these instruction-following LLMs provide strong economic incentives for dual-use by malicious actors. In particular, we show that instruction-following LLMs can produce targeted malicious content, including hate speech and scams, bypassing in-the-wild defenses implemented by LLM API vendors. Our analysis shows that this content can be generated economically and at cost likely lower than with human effort alone. Together, our findings suggest that LLMs will increasingly attract more sophisticated adversaries and attacks, and addressing these attacks may require new approaches to mitigations. | ['Tatsunori Hashimoto', 'Matei Zaharia', 'Carlos Guestrin', 'Ion Stoica', 'Xuechen Li', 'Daniel Kang'] | 2023-02-11 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [-1.79337993e-01 2.35985518e-01 -6.89441800e-01 4.96605262e-02
-1.06237996e+00 -1.16021276e+00 8.04524064e-01 2.00565234e-01
-3.26059699e-01 2.69158244e-01 3.46423328e-01 -1.33747661e+00
4.28372979e-01 -5.41918874e-01 -9.24984634e-01 -2.47013211e-01
-1.92757234e-01 -2.58635521e-01 1.70371741e-01 -3.63893270e-01
5.76973021e-01 6.37133181e-01 -8.03894222e-01 2.39085957e-01
5.71207941e-01 1.82437167e-01 -3.31610829e-01 7.67717600e-01
-5.19717373e-02 1.36210990e+00 -9.67150688e-01 -7.06480324e-01
3.32047462e-01 1.66217297e-01 -4.84378546e-01 -5.59089482e-01
4.02110457e-01 -8.15322459e-01 -5.52730918e-01 1.16259968e+00
1.79020882e-01 -2.70102382e-01 4.00215328e-01 -1.81674874e+00
-6.05501175e-01 1.06369555e+00 -9.86168504e-01 1.76742882e-01
2.66958684e-01 8.71441960e-01 6.93519294e-01 -2.07808465e-01
4.73052233e-01 1.51533449e+00 5.66618800e-01 7.51244724e-01
-1.30015290e+00 -1.19911420e+00 -1.41525283e-01 -6.99987292e-01
-1.04870343e+00 -4.53104496e-01 5.69852531e-01 -5.10726452e-01
1.34150267e+00 3.96579653e-01 -1.36954039e-01 1.66218114e+00
1.06466746e+00 7.94506490e-01 9.83482838e-01 -3.44754696e-01
-1.47077337e-01 2.69540727e-01 4.97262836e-01 7.64967144e-01
8.97265971e-01 3.41426939e-01 -3.52387518e-01 -9.17494655e-01
2.28315100e-01 -1.07999802e-01 -3.26102823e-02 4.56377156e-02
-6.58579230e-01 1.34365928e+00 9.55280438e-02 4.11044419e-01
4.02601324e-02 7.78079689e-01 6.14086330e-01 1.26643673e-01
-1.11549841e-02 1.02464092e+00 -5.59926808e-01 -4.91896093e-01
-6.35317624e-01 -2.67190486e-03 9.42698002e-01 7.27344036e-01
4.31691408e-01 4.39615637e-01 4.79991555e-01 9.43314470e-03
5.99340081e-01 6.64025068e-01 5.24932921e-01 -9.61081326e-01
7.85837889e-01 1.03272587e-01 8.53690598e-03 -1.24700463e+00
-1.15185097e-01 -3.29780072e-01 1.02155238e-01 3.03900182e-01
5.47683239e-01 -3.49776566e-01 -4.27341014e-01 1.71970856e+00
-1.37518331e-01 -1.36632062e-02 -1.82271734e-01 1.18166119e-01
-1.77719340e-01 7.39911914e-01 5.68719089e-01 1.98251471e-01
1.31943655e+00 -6.22096837e-01 -5.21639884e-01 -5.40545166e-01
1.32035458e+00 -8.12811673e-01 1.29309738e+00 4.82300818e-01
-6.81900203e-01 1.87851474e-01 -1.27700436e+00 2.49781206e-01
-6.31010115e-01 -6.88410461e-01 7.90150821e-01 1.72465968e+00
-6.04596972e-01 2.95489669e-01 -9.57554698e-01 2.46484905e-01
4.83820766e-01 2.68766880e-01 -1.66064173e-01 2.56289482e-01
-1.22076952e+00 9.93481398e-01 -1.48358956e-01 -4.02105778e-01
-1.39947855e+00 -9.40050244e-01 -1.15212107e+00 -1.01028748e-01
4.77296710e-01 1.10534720e-01 1.24145305e+00 -5.29322445e-01
-1.05024862e+00 7.33870208e-01 1.82476029e-01 -6.29776478e-01
1.67900726e-01 -5.50747275e-01 -3.95866513e-01 -1.41205147e-01
4.04099040e-02 8.71864334e-02 1.33104432e+00 -1.38449001e+00
-9.79819149e-02 2.02171784e-03 4.00520205e-01 -3.94526839e-01
-9.85796988e-01 8.01253676e-01 3.89925361e-01 -6.30280316e-01
-9.64486063e-01 -9.93301034e-01 -2.16367483e-01 -6.92681551e-01
-6.43180072e-01 1.97023246e-02 1.17958796e+00 -6.19117618e-01
1.55685961e+00 -2.25744867e+00 -5.14560997e-01 1.50598884e-01
6.31012559e-01 7.03715742e-01 -3.75777960e-01 4.80614513e-01
2.33933441e-02 1.13072944e+00 2.29583532e-01 -2.61518896e-01
4.10437167e-01 4.45036311e-03 -9.45299387e-01 6.32319808e-01
1.93109527e-01 1.09880412e+00 -7.81106770e-01 -2.81153560e-01
2.47820951e-02 1.68663040e-01 -6.76828444e-01 5.94782233e-02
-2.39434063e-01 -1.76902354e-01 -5.22956431e-01 7.67049849e-01
3.80351990e-01 9.83298868e-02 1.47806546e-02 4.97039288e-01
-5.04720397e-02 5.58119714e-01 -2.49648750e-01 9.07592833e-01
-8.26910198e-01 8.05079222e-01 5.88784039e-01 -1.48868263e-01
2.31518686e-01 -1.16589256e-02 -1.81482121e-01 -4.31482404e-01
3.83422405e-01 2.00660825e-01 2.86932886e-01 -3.91290128e-01
5.06315410e-01 2.29728401e-01 -6.21825159e-01 1.00709116e+00
-3.44558835e-01 -3.34090173e-01 -3.39531273e-01 6.52926564e-01
1.41226268e+00 -1.26583770e-01 1.38661012e-01 -3.11796308e-01
2.60781348e-01 1.49853647e-01 3.17984819e-01 1.02372766e+00
-5.61597586e-01 -5.47793269e-01 8.88425052e-01 -1.26952752e-01
-7.61019588e-01 -7.87167251e-01 1.89591125e-01 1.33990109e+00
-2.10213795e-01 -7.54621983e-01 -1.01901078e+00 -1.43357873e+00
2.59997934e-01 1.27114677e+00 -1.64932728e-01 -4.63329792e-01
-8.86458576e-01 -4.48570460e-01 1.44331503e+00 3.55436474e-01
-1.84957430e-01 -6.43259168e-01 -6.16623700e-01 4.70084324e-02
2.54090667e-01 -1.23463869e+00 -9.13781345e-01 3.45383346e-01
-5.00415385e-01 -1.00721323e+00 -1.02659695e-01 -2.49133021e-01
4.76978153e-01 2.97707200e-01 1.09840059e+00 4.56348032e-01
-5.98501503e-01 4.06466872e-01 -2.73440462e-02 -5.95849276e-01
-1.25215161e+00 -5.76406866e-02 8.77987817e-02 -5.68267465e-01
5.61021984e-01 -3.54504764e-01 1.36069208e-01 2.00539425e-01
-9.34564114e-01 -8.75157237e-01 5.78484952e-01 3.84511232e-01
-3.22168410e-01 2.05920249e-01 5.42454898e-01 -1.17121625e+00
1.00942504e+00 -7.46046305e-01 -8.99345696e-01 -1.99515715e-01
-5.53920090e-01 7.99501538e-02 1.06395805e+00 -9.16707873e-01
-7.88632393e-01 -2.77390182e-01 -1.56781808e-01 -3.64634901e-01
-3.73858690e-01 1.17904298e-01 -3.62014771e-01 -7.56928802e-01
8.17241132e-01 -1.58840090e-01 -1.74705181e-02 -1.32199883e-01
5.66690683e-01 5.78705728e-01 -5.69647588e-02 -8.22739720e-01
1.43107760e+00 1.34550691e-01 -1.23704650e-01 -1.02633214e+00
-4.11045462e-01 -3.06247696e-02 2.07683131e-01 7.86336064e-02
8.90251279e-01 -6.87150955e-01 -9.45358217e-01 5.46478212e-01
-1.04371762e+00 -5.33612132e-01 3.90932024e-01 -2.85809417e-03
-5.34850471e-02 9.53216612e-01 -1.20269012e+00 -8.71687233e-01
-1.08922809e-01 -1.82413971e+00 8.30388486e-01 -8.33432972e-02
-6.76005721e-01 -9.56483960e-01 -2.06978679e-01 7.36485898e-01
7.37343013e-01 9.91911963e-02 1.42608905e+00 -1.15069413e+00
-6.35226846e-01 -5.76415181e-01 2.57372171e-01 3.39041382e-01
-8.87939706e-02 9.92815793e-02 -8.93407464e-01 -3.30086887e-01
3.55619550e-01 -5.51880121e-01 2.14600146e-01 -1.79162577e-01
8.63208055e-01 -7.24041820e-01 -4.45478708e-01 4.66764778e-01
1.26567173e+00 3.22593421e-01 4.40493554e-01 2.60361433e-01
8.89307141e-01 4.96193796e-01 3.58007133e-01 9.47492868e-02
-9.74285230e-02 2.84818619e-01 4.21969056e-01 2.81705052e-01
1.39173672e-01 -6.71365261e-01 1.05196965e+00 4.82390434e-01
8.22274566e-01 -2.07870573e-01 -1.17501092e+00 2.74989665e-01
-1.12720251e+00 -7.54561305e-01 -2.76707292e-01 1.97728670e+00
7.54214942e-01 6.94376528e-01 -3.74822505e-02 -2.16957048e-01
4.01191503e-01 3.86840433e-01 -3.54711056e-01 -1.12211287e+00
2.89185375e-01 3.45812619e-01 1.13559175e+00 9.36431527e-01
-1.05281377e+00 1.11104965e+00 7.29466438e+00 1.32775056e+00
-9.65568483e-01 2.87902087e-01 7.20885158e-01 2.02382412e-02
-7.24117100e-01 2.05343738e-01 -1.12880898e+00 4.97264594e-01
1.67919183e+00 -3.57662678e-01 5.37317097e-01 1.06917846e+00
1.64018914e-01 -1.70560405e-02 -9.46502209e-01 4.54407752e-01
3.02267939e-01 -1.16612339e+00 -1.37382597e-01 6.05746686e-01
3.89793187e-01 -9.93000492e-02 5.72844863e-01 6.22192979e-01
9.48754787e-01 -1.24055827e+00 4.82666939e-01 -4.62380320e-01
1.91418216e-01 -1.18360031e+00 4.86925542e-01 7.98285484e-01
-7.74406552e-01 -2.90764809e-01 9.47445855e-02 -1.74371704e-01
6.97154924e-02 1.00367039e-01 -7.09082961e-01 -3.11992824e-01
5.79659678e-02 -1.18224651e-01 -7.57868767e-01 5.57863303e-02
-4.83630151e-01 1.02042627e+00 -1.68003306e-01 -2.24407628e-01
5.79438686e-01 2.81198442e-01 6.04127407e-01 1.35377491e+00
-2.44203746e-01 -4.04947251e-01 6.62016049e-02 8.79011452e-01
-2.19838172e-01 -2.86852926e-01 -1.18360496e+00 -8.31195116e-01
6.14284813e-01 1.22450352e+00 -6.56047523e-01 1.06678784e-01
-6.60080731e-01 6.32230222e-01 1.25097483e-01 1.73522979e-01
-1.18523467e+00 -6.38390720e-01 1.12011206e+00 3.13654542e-01
-3.06629211e-01 -6.27274990e-01 -2.01682255e-01 -1.14532244e+00
-3.99333864e-01 -1.67695343e+00 -1.50709469e-02 -3.22937936e-01
-1.23157108e+00 2.18520999e-01 -7.40035390e-03 -4.37596858e-01
-4.11339551e-01 -7.30279326e-01 -6.24051034e-01 6.64896488e-01
-1.22772169e+00 -1.15512681e+00 8.28839898e-01 3.97345960e-01
4.06003952e-01 -2.54554987e-01 6.12179279e-01 1.25152275e-01
-6.28902435e-01 1.05317652e+00 -1.54816717e-01 4.65754777e-01
7.54037142e-01 -8.79840851e-01 8.51263165e-01 1.36874151e+00
9.41005796e-02 1.41235220e+00 6.97378993e-01 -1.06008399e+00
-2.01239657e+00 -7.86467850e-01 7.13045061e-01 -1.27960837e+00
1.62460947e+00 -8.96020114e-01 -5.90949893e-01 1.09156406e+00
4.38936949e-01 -6.45703495e-01 8.71789217e-01 -1.52424082e-01
-9.80588675e-01 7.33891487e-01 -1.07856929e+00 9.96479452e-01
4.83476907e-01 -1.11846220e+00 -5.98071158e-01 3.85491788e-01
1.25898802e+00 8.51614401e-02 -5.87472439e-01 -2.63889402e-01
3.72759581e-01 -6.36318266e-01 9.56725180e-01 -9.90257561e-01
5.40019572e-01 2.21458972e-02 -2.69277036e-01 -9.13319468e-01
8.26051608e-02 -1.36445093e+00 -2.85816967e-01 1.40927672e+00
5.32624722e-01 -9.71001208e-01 6.70417309e-01 1.09268999e+00
3.43225189e-02 -3.64459097e-01 -5.51010728e-01 -9.78966475e-01
8.51723850e-01 -7.23535240e-01 4.71036404e-01 1.20295513e+00
3.52675527e-01 4.35359254e-02 -3.93487513e-01 4.50011402e-01
1.04410160e+00 -5.10019302e-01 7.96699703e-01 -4.56256479e-01
-6.56809151e-01 -6.14640415e-01 -2.04916134e-01 -8.03755224e-01
7.36005962e-01 -7.45820284e-01 -2.38046899e-01 -3.90369385e-01
1.65771827e-01 -1.96180984e-01 1.71960399e-01 6.71863437e-01
-1.65393621e-01 6.48553297e-02 6.78643584e-01 5.57601489e-02
-2.70165354e-01 -5.13330586e-02 5.78357995e-01 -2.03416958e-01
1.58299655e-01 -2.13717312e-01 -9.63526845e-01 1.23572385e+00
7.32232988e-01 -9.73239899e-01 -3.14441174e-01 -2.35342622e-01
3.50086719e-01 -3.87981385e-02 -4.06280421e-02 -6.07846797e-01
-1.48392513e-01 -4.88804638e-01 -3.29349279e-01 -1.31550068e-02
2.78735217e-02 -8.27175021e-01 -3.77841562e-01 9.72747862e-01
-3.66559297e-01 2.24071726e-01 5.90461195e-01 5.77797472e-01
2.92018801e-01 -6.80635154e-01 6.91940546e-01 -1.08860694e-01
-4.51568842e-01 -1.26859918e-01 -1.30549634e+00 3.47439229e-01
1.32893050e+00 1.74873650e-01 -9.46850538e-01 -5.67412972e-01
-8.22142214e-02 9.74999219e-02 8.21310639e-01 6.08693898e-01
2.56576687e-01 -7.97856569e-01 -1.18814498e-01 5.83096482e-02
-2.83505674e-02 -1.12534726e+00 -1.92199245e-01 3.41775507e-01
-5.07735610e-01 4.27105844e-01 1.69225335e-01 7.88900778e-02
-1.45286489e+00 1.06425488e+00 1.62499458e-01 -4.94276524e-01
-1.44717515e-01 8.29830468e-01 4.41044867e-01 -2.34406978e-01
1.53034091e-01 -1.29168510e-01 3.23086113e-01 -3.96274924e-01
7.17241466e-01 3.58117372e-01 -4.88295436e-01 -5.51055014e-01
-5.31182110e-01 -9.90176573e-02 -2.87043035e-01 -4.27886192e-03
9.15748417e-01 3.30322176e-01 -3.23904812e-01 1.42898764e-02
1.30833876e+00 1.21070409e+00 -7.31304288e-01 4.76074427e-01
2.05020860e-01 -6.03118062e-01 -1.03349118e-02 -8.04712415e-01
-7.52804101e-01 8.46126854e-01 -1.22211978e-01 4.12287354e-01
4.07192290e-01 -2.54816890e-01 1.35625029e+00 3.38275880e-01
7.08587825e-01 -6.78727448e-01 3.12903225e-01 4.25311387e-01
3.24234933e-01 -9.73633826e-01 -2.39569530e-01 -4.47070360e-01
-4.36092108e-01 7.19901741e-01 7.31077373e-01 -2.10415870e-01
6.93034947e-01 1.17495191e+00 1.27845451e-01 -7.34942928e-02
-5.54960966e-01 7.09314704e-01 -3.89351606e-01 8.19686472e-01
4.34323996e-01 3.97713035e-01 -3.12782496e-01 6.27068639e-01
5.26832715e-02 -7.24017859e-01 1.12723088e+00 1.18173647e+00
-3.68194669e-01 -1.49203610e+00 -9.70015228e-01 3.23428869e-01
-1.17188323e+00 -5.34866273e-01 -6.46556556e-01 1.08030725e+00
-1.65355071e-01 1.46698928e+00 -4.43952352e-01 -6.88122511e-01
-2.37155721e-01 1.72176197e-01 4.16454673e-02 -6.48662567e-01
-1.14693463e+00 -1.16271161e-01 3.90733093e-01 -9.95701373e-01
6.72005236e-01 -2.88805097e-01 -1.09777832e+00 -7.75676131e-01
-4.10764009e-01 8.49620625e-02 7.27405965e-01 6.13078237e-01
4.75866169e-01 2.70246923e-01 5.58263123e-01 -6.10902786e-01
-1.20382512e+00 -4.45266306e-01 -3.31260443e-01 4.28539097e-01
3.78424644e-01 -3.78753424e-01 -9.81946170e-01 -2.12429896e-01] | [6.109522342681885, 7.848875999450684] |
26d44fa7-ba79-43db-8c8f-b9f0688519b7 | sememnn-a-semantic-matrix-based-memory-neural | 2003.01857 | null | https://arxiv.org/abs/2003.01857v1 | https://arxiv.org/pdf/2003.01857v1.pdf | SeMemNN: A Semantic Matrix-Based Memory Neural Network for Text Classification | Text categorization is the task of assigning labels to documents written in a natural language, and it has numerous real-world applications including sentiment analysis as well as traditional topic assignment tasks. In this paper, we propose 5 different configurations for the semantic matrix-based memory neural network with end-to-end learning manner and evaluate our proposed method on two corpora of news articles (AG news, Sogou news). The best performance of our proposed method outperforms the baseline VDCNN models on the text classification task and gives a faster speed for learning semantics. Moreover, we also evaluate our model on small scale datasets. The results show that our proposed method can still achieve better results in comparison to VDCNN on the small scale dataset. This paper is to appear in the Proceedings of the 2020 IEEE 14th International Conference on Semantic Computing (ICSC 2020), San Diego, California, 2020. | ['Hiroshi Ishiguro', 'Chaoran Liu', 'Changzeng Fu', 'Carlos Toshinori Ishi', 'Yuichiro Yoshikawa'] | 2020-03-04 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [-7.12106079e-02 -4.86886129e-03 -3.02150369e-01 -7.02021360e-01
-2.84745097e-01 -3.36842656e-01 8.85466695e-01 4.62889463e-01
-7.80849397e-01 6.25087440e-01 4.45761919e-01 -1.27291709e-01
-9.88630429e-02 -8.82809579e-01 -3.60787690e-01 -1.37623489e-01
4.15733568e-02 8.18833768e-01 6.65379584e-01 -1.37233332e-01
7.76230216e-01 -7.12874308e-02 -1.76157522e+00 7.80070722e-01
4.75945681e-01 1.20142603e+00 5.36364794e-01 2.70355463e-01
-8.53600919e-01 9.76588488e-01 -5.95097363e-01 -1.60265237e-01
-6.26313016e-02 -1.95292756e-01 -1.38789606e+00 -1.68172672e-01
4.51791495e-01 3.58092010e-01 1.41331673e-01 1.27384973e+00
1.81758001e-01 6.65708661e-01 8.69242728e-01 -1.14663374e+00
-4.59815443e-01 1.06839037e+00 -6.69632316e-01 2.42355391e-01
4.90917638e-02 -1.04216337e+00 1.09716904e+00 -9.09066737e-01
8.12017560e-01 1.52933097e+00 6.74447596e-01 4.65513706e-01
-7.08139718e-01 -6.73586071e-01 4.25935179e-01 5.70339501e-01
-1.11965621e+00 -1.19363271e-01 5.60120344e-01 -3.66384774e-01
9.82836545e-01 2.91230474e-02 1.52876928e-01 6.44211054e-01
9.97806191e-02 8.69648278e-01 8.47993135e-01 -7.08830893e-01
5.27232528e-01 4.69282717e-01 9.96966362e-01 4.50914651e-01
1.15972497e-01 -7.23618150e-01 -5.28170466e-01 -2.74610460e-01
1.11487001e-01 8.89278874e-02 3.15485559e-02 -2.20317543e-01
-1.14591324e+00 1.17494333e+00 5.25895953e-01 6.32443249e-01
-1.94018800e-02 3.17665219e-01 9.05657232e-01 3.69618148e-01
8.48153412e-01 3.90668631e-01 -6.14737213e-01 1.90990239e-01
-7.98828363e-01 4.01735082e-02 9.48164642e-01 9.18254614e-01
5.05672932e-01 -2.16138572e-01 8.29695836e-02 1.33829057e+00
3.06474149e-01 8.12120363e-02 1.25980651e+00 -7.47604132e-01
5.60096085e-01 6.36769414e-01 1.66732445e-01 -1.32433939e+00
-8.78175199e-01 -2.15698406e-01 -6.87039793e-01 -2.74774522e-01
-5.16702421e-02 -5.63257970e-02 -7.98821092e-01 1.30514765e+00
1.36033103e-01 1.17246814e-01 3.24839532e-01 8.98480594e-01
1.28611791e+00 1.14858198e+00 2.68129349e-01 -1.68769792e-01
1.47247350e+00 -1.41068566e+00 -8.45795155e-01 -2.76801378e-01
8.78961682e-01 -8.30697358e-01 1.03915131e+00 4.19244230e-01
-5.25745392e-01 -5.07214069e-01 -9.86523986e-01 2.62284391e-02
-8.53925109e-01 4.66817528e-01 7.08557129e-01 5.09786308e-01
-1.09362054e+00 3.94653708e-01 -7.00752974e-01 -9.92999375e-01
1.40482768e-01 2.55758524e-01 -1.22451268e-01 1.24990314e-01
-1.49178338e+00 9.20886040e-01 1.02722633e+00 -2.44673297e-01
-6.26758099e-01 -1.62881196e-01 -5.92039347e-01 7.56603628e-02
3.85665715e-01 -2.57768750e-01 1.30760276e+00 -1.12238705e+00
-1.25476313e+00 9.95314896e-01 -7.52842203e-02 -5.87646723e-01
3.68927233e-02 -2.03709364e-01 -5.95232725e-01 5.71346842e-02
4.56150949e-01 8.87294292e-01 3.65299851e-01 -1.01927686e+00
-1.02706480e+00 -4.12871897e-01 7.66088590e-02 4.50971544e-01
-1.05470002e+00 2.34113842e-01 -3.61227065e-01 -6.78267360e-01
2.80481815e-01 -7.48704493e-01 -8.55822787e-02 -2.65472502e-01
-2.21386716e-01 -6.84836805e-01 9.92488503e-01 -1.78817406e-01
1.08010173e+00 -2.01027393e+00 -5.36147356e-02 8.27494189e-02
-5.14314882e-02 -2.33258381e-02 1.09394230e-01 2.87417591e-01
4.78678904e-02 -7.33138174e-02 -7.24187568e-02 -3.52020442e-01
-7.78746307e-02 -1.39637262e-01 -4.62380767e-01 1.85762465e-01
-7.49297619e-01 3.59581828e-01 -6.14391208e-01 -6.25350177e-01
-1.03476219e-01 -6.86196163e-02 -1.13847807e-01 -2.89674282e-01
-4.48840708e-01 -3.61171752e-01 -6.39427125e-01 2.71377534e-01
4.31283474e-01 -4.16690707e-01 4.28877205e-01 -5.38102575e-02
1.04089946e-01 2.45134830e-01 -1.24174953e+00 1.84124708e+00
-4.73206937e-01 7.46602476e-01 -1.88513577e-01 -1.38407624e+00
1.10592151e+00 2.89680809e-01 8.38003084e-02 -4.99027967e-01
4.84242111e-01 2.37603351e-01 -2.57933468e-01 -2.30434984e-01
7.78475225e-01 -2.35923678e-01 -1.36234343e-01 8.45360696e-01
2.45643198e-01 1.85640648e-01 4.81458127e-01 4.97473300e-01
6.20381057e-01 -6.24140501e-02 3.15679610e-01 -1.05471826e+00
8.16767097e-01 4.56724703e-01 1.02668241e-01 5.90620458e-01
-6.16891570e-02 3.08204830e-01 3.13126385e-01 -7.72601008e-01
-7.61977136e-01 -3.66624892e-01 -2.64750302e-01 1.64666581e+00
4.64574158e-01 -4.79945749e-01 -8.73835027e-01 -7.57385910e-01
-1.91532448e-02 1.03945136e+00 -6.74951553e-01 -5.20039387e-02
-2.58329183e-01 -9.99193609e-01 4.30758059e-01 4.70558167e-01
7.64552355e-01 -1.25524819e+00 -3.04806381e-01 1.22448981e-01
-3.07371318e-01 -1.05634964e+00 -4.01555151e-01 2.38332391e-01
-1.13019049e+00 -9.02882099e-01 -5.17820895e-01 -1.44262576e+00
4.49617654e-01 4.19540584e-01 8.71531308e-01 -1.02393389e-01
1.01563685e-01 7.12193921e-02 -6.25205219e-01 -5.80418468e-01
-2.15298072e-01 5.46497107e-01 8.38062614e-02 -6.59093782e-02
8.65760028e-01 3.30874100e-02 -3.51912290e-01 3.65098387e-01
-8.33947301e-01 1.27336159e-01 -3.36693004e-02 8.28763247e-01
2.46645391e-01 4.30937141e-01 9.96867895e-01 -1.42100441e+00
8.84536386e-01 -5.46509385e-01 -5.79463422e-01 2.24803507e-01
-1.02447784e+00 7.43822604e-02 8.61951530e-01 -1.68190464e-01
-1.25334299e+00 -2.12805867e-01 1.51697978e-01 1.34587094e-01
-2.15476416e-02 8.31595302e-01 1.47057056e-01 3.36876154e-01
8.47655237e-01 1.23359911e-01 -2.17496976e-01 -6.92948580e-01
2.76960135e-01 1.13420355e+00 2.41946593e-01 -5.93291640e-01
2.52559297e-02 5.90446889e-01 -3.27388585e-01 -5.33192158e-01
-1.31075418e+00 -9.12786722e-01 -4.27781522e-01 -1.20018750e-01
7.82283247e-01 -8.96628261e-01 -4.22455072e-01 4.64641184e-01
-1.15924537e+00 1.18451558e-01 3.78079474e-01 6.20513558e-01
-2.95583576e-01 1.56905338e-01 -7.79190242e-01 -4.68768120e-01
-6.39979124e-01 -7.29077995e-01 8.48270774e-01 6.91136867e-02
-3.16899508e-01 -1.46524143e+00 4.48560193e-02 3.38588417e-01
4.86794651e-01 -4.61698502e-01 1.24940825e+00 -1.36469686e+00
1.83724418e-01 -1.13572299e-01 -4.10308599e-01 3.50483328e-01
-5.09869605e-02 -5.48486292e-01 -9.26053405e-01 -3.14436823e-01
9.19508860e-02 -4.79852021e-01 1.34063876e+00 3.26698750e-01
1.13216388e+00 -8.90530199e-02 -6.70238256e-01 -2.56221765e-03
1.60712981e+00 5.42444110e-01 4.04678166e-01 9.58487868e-01
6.89539671e-01 9.09682155e-01 7.91835546e-01 2.31661662e-01
3.16762805e-01 4.51046020e-01 8.84915739e-02 2.78069019e-01
-5.47674038e-02 1.44447759e-02 2.16098621e-01 1.08389413e+00
3.63101572e-01 -6.67421341e-01 -1.13759553e+00 5.56972802e-01
-2.19791722e+00 -7.31627762e-01 -2.75719851e-01 1.90738285e+00
7.77054191e-01 2.19522178e-01 -1.49535403e-01 -6.46974072e-02
1.18743408e+00 1.55654535e-01 -2.68333822e-01 -4.48838294e-01
-2.25984473e-02 4.46610488e-02 3.51324469e-01 4.39813286e-01
-1.50519431e+00 1.25870192e+00 5.91357565e+00 1.17928255e+00
-1.13040626e+00 5.69315910e-01 7.37449884e-01 1.32704675e-02
-7.25399777e-02 -4.91334908e-02 -1.01063323e+00 4.79834408e-01
1.02664876e+00 -3.11628491e-01 -4.42589857e-02 1.12440670e+00
-2.01778814e-01 -9.11350474e-02 -7.98253000e-01 8.63463283e-01
6.00629866e-01 -1.55003357e+00 2.84307092e-01 -6.13295913e-01
9.73924816e-01 9.22885835e-02 -2.97225982e-01 4.09590632e-01
3.21757525e-01 -8.29764426e-01 6.48453772e-01 1.32898599e-01
4.54959631e-01 -9.55692530e-01 1.32742417e+00 3.82050157e-01
-9.24537718e-01 -1.12160690e-01 -7.50598133e-01 -5.45769483e-02
-1.56639144e-01 5.57394743e-01 -7.62384117e-01 3.34112406e-01
9.00079608e-01 9.27236557e-01 -4.95428532e-01 6.79781497e-01
1.72804072e-01 7.29451299e-01 -4.58532497e-02 -8.01808596e-01
4.41747636e-01 -3.60763408e-02 1.58502355e-01 1.25033867e+00
2.00114742e-01 -1.18520632e-01 3.13808739e-01 3.65471125e-01
-3.26440454e-01 6.03917539e-01 -2.03459188e-01 1.33925259e-01
4.18860465e-01 1.23445451e+00 -1.53958166e+00 -7.79470563e-01
-3.03769916e-01 8.20441604e-01 3.33211601e-01 1.41370267e-01
-5.14796495e-01 -8.56815398e-01 -5.86596318e-02 -1.17215045e-01
4.57401276e-02 3.39558005e-01 -3.33652079e-01 -1.12381721e+00
-1.69997513e-01 -5.54801106e-01 7.13935912e-01 -8.05770397e-01
-1.38384628e+00 8.63676786e-01 -3.62209440e-03 -1.01747715e+00
-1.07943937e-02 -9.09439445e-01 -3.71923178e-01 3.50285023e-01
-1.16761482e+00 -9.32586193e-01 -2.36882687e-01 4.16537404e-01
8.11820567e-01 -7.79230237e-01 9.33908820e-01 3.68707418e-01
-1.81509629e-01 2.55591691e-01 7.76960433e-01 9.90509614e-02
9.70708489e-01 -1.16440892e+00 2.60963261e-01 5.21043062e-01
2.50427425e-01 5.77131391e-01 5.64712167e-01 -5.88182509e-01
-7.01414227e-01 -1.25820744e+00 1.08308005e+00 -3.03841263e-01
7.16749609e-01 -4.47321206e-01 -7.51434505e-01 7.81917512e-01
4.19925988e-01 -4.30028647e-01 6.21877551e-01 2.24116355e-01
-2.88218170e-01 -1.52597979e-01 -1.04386628e+00 2.61666119e-01
6.95973337e-01 -2.83405393e-01 -9.40718949e-01 7.47260988e-01
1.05179417e+00 -1.62484139e-01 -4.85224426e-01 -1.06333904e-02
2.57670462e-01 -4.88745809e-01 6.66664958e-01 -7.15113819e-01
6.17248535e-01 -4.52418476e-01 -3.70171219e-01 -1.27070868e+00
-2.53764123e-01 2.09621385e-01 3.78072083e-01 1.25344515e+00
6.05165005e-01 -7.17614830e-01 8.08584034e-01 4.18682136e-02
-2.19386011e-01 -2.77796149e-01 -7.16434598e-01 -6.78933024e-01
1.78225294e-01 -5.64981163e-01 3.40965867e-01 1.35214019e+00
4.15973544e-01 7.85108984e-01 -2.00666413e-01 -3.29253763e-01
4.58099365e-01 3.39350671e-01 2.05566153e-01 -1.63369656e+00
2.12059226e-02 -4.79933113e-01 -7.37446547e-01 -7.63222992e-01
5.30428052e-01 -1.21768713e+00 -2.16854718e-02 -1.77886140e+00
5.76059878e-01 -3.42399269e-01 -5.40970027e-01 5.12487650e-01
-2.61250772e-02 1.50237411e-01 -2.76489817e-02 4.61370647e-01
-1.06710351e+00 3.85159433e-01 6.99953079e-01 -4.30845499e-01
1.37096569e-01 -1.23605169e-01 -7.59871960e-01 8.60249519e-01
8.33399415e-01 -7.89757550e-01 -7.31452346e-01 -5.76934397e-01
1.39039814e-01 -2.39935040e-01 -1.67959943e-01 -8.70992959e-01
5.46859741e-01 -6.87263086e-02 2.15200275e-01 -5.94127297e-01
1.15694515e-02 -6.82517588e-01 -2.73326337e-01 4.66057509e-01
-8.41143906e-01 -2.44339630e-02 1.84909567e-01 5.20543933e-01
-6.39118552e-01 -7.35879898e-01 7.23755658e-01 -2.73615062e-01
-1.26720524e+00 -1.78328738e-01 -5.00456095e-01 1.07434884e-01
1.02294374e+00 1.75622001e-01 -6.75030887e-01 -2.41636753e-01
-6.81527793e-01 4.14654076e-01 1.48497880e-01 7.98481524e-01
4.49999511e-01 -1.27739477e+00 -4.93111819e-01 -3.30122173e-01
3.28755736e-01 -2.70132035e-01 1.13055639e-01 3.48520458e-01
-8.69405210e-01 9.17104006e-01 -1.47251159e-01 -3.63637120e-01
-1.15200627e+00 6.37907147e-01 1.18897825e-01 -1.01436488e-01
-3.72391582e-01 8.55718255e-01 1.95621178e-01 -9.03090775e-01
4.81174469e-01 1.35205448e-01 -9.00979877e-01 3.00726384e-01
6.47765875e-01 4.29813713e-01 2.14316800e-01 -3.63562733e-01
-5.18910825e-01 4.57122475e-01 -4.19303983e-01 -1.32936597e-01
1.25660610e+00 -1.57977685e-01 -6.79852188e-01 8.40305865e-01
1.43193007e+00 -5.32523394e-01 -6.53368235e-02 -4.08948362e-01
5.46144366e-01 -5.27514927e-02 3.58977437e-01 -7.67490089e-01
-9.52183306e-01 7.83311009e-01 8.02351654e-01 2.18119964e-01
8.50891113e-01 3.74120809e-02 7.42049336e-01 9.54598129e-01
5.50677121e-01 -1.68241394e+00 -2.56243125e-02 6.09231830e-01
5.66310942e-01 -1.18178785e+00 4.62631248e-02 -2.78814167e-01
-7.86632180e-01 1.29830718e+00 7.39994287e-01 -1.44557610e-01
8.37282777e-01 -9.90993381e-02 2.98157632e-01 -3.06242347e-01
-8.99989724e-01 1.64583057e-01 1.28800929e-01 2.47904688e-01
6.85416341e-01 -1.63931586e-02 -7.89932609e-01 5.99192679e-01
-1.45627752e-01 -1.19943716e-01 5.69644868e-01 8.37982655e-01
-8.27011287e-01 -9.47628081e-01 -3.18578809e-01 7.18834937e-01
-5.37126899e-01 -3.49848270e-01 -4.40101653e-01 6.79836333e-01
-1.11416996e-01 9.72439051e-01 3.03752512e-01 -2.66107082e-01
4.50862013e-02 2.47775108e-01 -2.55861223e-01 -8.98925126e-01
-5.40909648e-01 -5.80900200e-02 4.16651428e-01 -2.18189508e-01
-7.03149915e-01 -5.63261032e-01 -1.73871553e+00 -1.77402705e-01
-4.38222140e-01 8.45963418e-01 9.73661840e-01 1.02912402e+00
3.08481395e-01 4.33886588e-01 4.02934313e-01 -3.76128972e-01
-3.45119178e-01 -1.32778168e+00 -8.23296666e-01 4.45960701e-01
-2.46984825e-01 -8.34218860e-01 -4.45236713e-01 2.31081426e-01] | [10.415522575378418, 6.9566731452941895] |
cb1fa575-be3d-4a27-b2b3-bd267b901d98 | hpn-personalized-federated-hyperparameter | 2304.05195 | null | https://arxiv.org/abs/2304.05195v1 | https://arxiv.org/pdf/2304.05195v1.pdf | HPN: Personalized Federated Hyperparameter Optimization | Numerous research studies in the field of federated learning (FL) have attempted to use personalization to address the heterogeneity among clients, one of FL's most crucial and challenging problems. However, existing works predominantly focus on tailoring models. Yet, due to the heterogeneity of clients, they may each require different choices of hyperparameters, which have not been studied so far. We pinpoint two challenges of personalized federated hyperparameter optimization (pFedHPO): handling the exponentially increased search space and characterizing each client without compromising its data privacy. To overcome them, we propose learning a \textsc{H}yper\textsc{P}arameter \textsc{N}etwork (HPN) fed with client encoding to decide personalized hyperparameters. The client encoding is calculated with a random projection-based procedure to protect each client's privacy. Besides, we design a novel mechanism to debias the low-fidelity function evaluation samples for learning HPN. We conduct extensive experiments on FL tasks from various domains, demonstrating the superiority of HPN. | ['Jian Cheng', 'Yaliang Li', 'Zhen Wang', 'Anda Cheng'] | 2023-04-11 | null | null | null | null | ['hyperparameter-optimization'] | ['methodology'] | [ 1.58411320e-02 -5.17560206e-02 -2.94636458e-01 -5.27155638e-01
-1.04570282e+00 -5.57641923e-01 2.01585203e-01 -4.18130845e-01
-3.58273417e-01 8.34062397e-01 2.97232240e-01 -2.03784287e-01
-4.60675597e-01 -6.28192604e-01 -6.21573865e-01 -1.20464098e+00
1.29480332e-01 5.65656304e-01 -2.90326804e-01 1.90943018e-01
8.28565285e-02 3.58198732e-01 -1.39946043e+00 3.18384588e-01
7.60776579e-01 1.17107785e+00 -2.24330008e-01 2.63565600e-01
-8.28672349e-02 3.94188255e-01 -4.85056996e-01 -8.51272523e-01
5.00316858e-01 -1.01036943e-01 -7.55489647e-01 -2.47732773e-02
3.21862787e-01 -5.57332635e-01 -5.22010744e-01 1.26075602e+00
9.13265288e-01 3.30502987e-01 3.73032182e-01 -1.42978299e+00
-5.68635941e-01 9.93459821e-01 -4.32571232e-01 -7.43311495e-02
1.46542549e-01 3.82547051e-01 1.17938387e+00 -5.87728918e-01
3.91948164e-01 1.22995877e+00 6.33699894e-01 6.48343086e-01
-1.38066113e+00 -9.82361317e-01 1.51096925e-01 3.77101421e-01
-1.39566100e+00 -3.87760460e-01 7.14160740e-01 -4.01556343e-02
4.25731421e-01 5.29451132e-01 2.16233775e-01 1.43103051e+00
-3.70096445e-01 9.92952824e-01 9.62002814e-01 -4.90584411e-02
4.91310358e-01 5.03845096e-01 1.57521367e-01 3.25004846e-01
1.90031454e-01 7.79229626e-02 -5.96132636e-01 -8.83728504e-01
2.69168228e-01 2.09210292e-01 -6.87483251e-01 -4.94552672e-01
-6.66315317e-01 9.14877951e-01 -4.14856859e-02 -1.95186839e-01
-2.64937073e-01 -9.94815007e-02 5.42351246e-01 2.28392601e-01
1.50040075e-01 1.63913846e-01 -7.70384669e-01 7.38327391e-04
-5.21202922e-01 3.02491277e-01 9.92684245e-01 1.08775103e+00
9.98609424e-01 -4.30040419e-01 -4.43888724e-01 8.89087737e-01
2.74462432e-01 4.64387894e-01 6.40832782e-01 -1.05750751e+00
8.03544283e-01 4.06504571e-01 2.02627908e-02 -8.32740903e-01
-5.66593707e-02 -2.37800196e-01 -7.72567809e-01 -1.39900044e-01
2.38290116e-01 -5.29681504e-01 -2.48446032e-01 1.82972622e+00
6.24394178e-01 2.93352678e-02 3.82680222e-02 9.42490935e-01
3.26506644e-01 5.74308991e-01 9.19153765e-02 -1.77015767e-01
1.25541627e+00 -9.47649300e-01 -4.32296693e-01 2.51016021e-01
4.80171978e-01 -4.22485769e-01 1.28625643e+00 5.03758609e-01
-9.93084311e-01 1.78274795e-01 -7.29939282e-01 3.99167091e-03
-8.28035399e-02 -1.13183089e-01 6.10874355e-01 9.71028924e-01
-6.77650154e-01 6.37111127e-01 -6.07634008e-01 -1.75489932e-01
8.08997512e-01 5.44992447e-01 -3.16465676e-01 -1.32023349e-01
-1.27548695e+00 1.71665281e-01 3.59191179e-01 -1.87832102e-01
-9.80555177e-01 -8.67298901e-01 -3.36345702e-01 5.32634914e-01
7.02314079e-01 -7.52976537e-01 1.26406610e+00 -5.08395970e-01
-1.71694148e+00 4.80048269e-01 2.59288907e-01 -2.52050430e-01
8.26998234e-01 -4.47460860e-02 -2.92347252e-01 -1.09457746e-01
-2.43532315e-01 -8.45916867e-02 9.27314639e-01 -1.09925365e+00
-8.39520514e-01 -8.47965658e-01 -1.61232129e-01 3.12762767e-01
-1.07813239e+00 -8.07779580e-02 -7.03538537e-01 -5.20572305e-01
-3.14465374e-01 -7.51922488e-01 -1.86596885e-01 -4.08211304e-03
-6.54388666e-01 -2.96548098e-01 9.70183790e-01 -3.08430344e-01
1.39865327e+00 -2.34426618e+00 -2.28610821e-02 4.80489135e-01
3.43342513e-01 4.38612372e-01 -1.56360805e-01 4.30561244e-01
3.30084175e-01 1.53209418e-01 -1.50533214e-01 -6.83626533e-01
5.76513529e-01 1.96248740e-01 -4.61469173e-01 6.79533064e-01
-6.62593722e-01 5.14479935e-01 -4.35442448e-01 -6.15509093e-01
-1.75424799e-01 5.34169436e-01 -7.67413139e-01 4.97180969e-01
-5.53598523e-01 2.94279486e-01 -9.91740704e-01 5.11235058e-01
9.79639709e-01 -4.64250416e-01 4.26036507e-01 -1.37755707e-01
1.25518829e-01 5.05053103e-02 -1.36512899e+00 1.41526604e+00
-4.00960356e-01 -1.60329510e-02 4.33486730e-01 -7.39589691e-01
7.60592997e-01 4.95036453e-01 7.32507646e-01 -3.68787706e-01
3.35113883e-01 4.15217131e-02 -8.02827597e-01 -6.32603407e-01
2.46204436e-01 1.63410380e-01 8.01095590e-02 8.67359638e-01
-3.88578735e-02 5.92609644e-01 -3.82935971e-01 -1.02024212e-01
1.20365739e+00 -1.55897111e-01 -6.17605150e-02 -3.38405371e-02
7.50548840e-01 -4.16216940e-01 8.69210541e-01 7.44247735e-01
-3.97006363e-01 4.83453006e-01 5.17712474e-01 -3.46577734e-01
-8.61983180e-01 -6.03309751e-01 -2.16920108e-01 1.61546695e+00
6.07958324e-02 -5.11141539e-01 -9.08214390e-01 -1.02715838e+00
2.95759529e-01 6.58874512e-01 -5.12602091e-01 -1.40917361e-01
-5.60466826e-01 -9.11423624e-01 7.87227094e-01 3.61951590e-02
5.46830714e-01 -8.85766208e-01 -4.71146554e-01 6.86912015e-02
-3.54405344e-01 -8.04372191e-01 -9.35766697e-01 9.71979052e-02
-6.19636655e-01 -1.08874285e+00 -6.39602184e-01 -3.96866113e-01
4.67245907e-01 1.20653212e-01 7.45006323e-01 -1.93337500e-01
-2.30796948e-01 3.41841161e-01 -2.16395050e-01 2.37300489e-02
-1.39180437e-01 5.09191632e-01 -8.74956027e-02 4.66712832e-01
5.65139651e-01 -7.47057080e-01 -8.37429762e-01 7.28903651e-01
-9.83533680e-01 -5.28775036e-01 4.28380519e-01 8.25478137e-01
6.35182738e-01 1.10323943e-01 2.75015771e-01 -1.38308370e+00
7.18799055e-01 -7.94769824e-01 -6.60292089e-01 5.93940079e-01
-1.12219870e+00 2.06416428e-01 9.84256029e-01 -5.85722148e-01
-1.33911502e+00 -3.37155499e-02 -1.95287764e-02 -6.85348988e-01
-1.00797161e-01 1.80465486e-02 -8.38493824e-01 -1.90533623e-02
6.50241435e-01 8.78962055e-02 -3.94480303e-02 -9.11716998e-01
4.47765172e-01 9.00250494e-01 4.14338499e-01 -9.95220244e-01
5.94103277e-01 3.35079640e-01 -2.26028457e-01 -2.19905242e-01
-6.27228856e-01 -2.49876380e-01 9.09207761e-02 3.06130588e-01
5.36723495e-01 -5.65881550e-01 -1.48483062e+00 2.83103138e-01
-7.54469097e-01 -4.78294753e-02 -3.89395356e-01 3.97780061e-01
-6.77606285e-01 6.05460167e-01 -4.08650905e-01 -6.82506382e-01
-8.46871018e-01 -1.23981726e+00 7.05564320e-01 2.43184417e-01
2.09107146e-01 -7.66701579e-01 3.21994245e-01 6.73492074e-01
6.03717923e-01 -8.72195885e-02 1.10880017e+00 -1.03612053e+00
-6.44740224e-01 -1.89498305e-01 -1.12869263e-01 2.15646490e-01
-8.84462371e-02 -3.85920912e-01 -1.06594396e+00 -5.73142946e-01
1.86879471e-01 -5.20641685e-01 5.06575227e-01 -1.11203104e-01
1.77809191e+00 -1.06874692e+00 -3.02851707e-01 1.42259943e+00
1.34100711e+00 -1.44729406e-01 4.95612592e-01 3.65445793e-01
3.71977568e-01 3.90656501e-01 3.61512601e-01 1.26592445e+00
3.76726002e-01 6.57341719e-01 5.45141101e-01 5.06044865e-01
3.01398367e-01 -4.07925457e-01 3.60551894e-01 8.75157267e-02
2.46105462e-01 -4.55274731e-01 -4.82501835e-01 1.85033813e-01
-1.92722166e+00 -7.33960271e-01 4.31743294e-01 2.31495023e+00
1.03564274e+00 -5.21089435e-01 6.71202093e-02 -2.09742501e-01
9.47934926e-01 1.53020501e-01 -1.16546404e+00 -2.80189272e-02
-1.74855351e-01 -2.51963385e-03 7.85639465e-01 -2.41120812e-02
-9.70846593e-01 6.27423763e-01 5.24838686e+00 9.62227345e-01
-9.54548657e-01 1.27064228e-01 5.78064084e-01 -2.60025471e-01
-5.13998270e-01 -2.41191383e-03 -1.12974954e+00 8.15917611e-01
6.96118414e-01 -4.10891235e-01 9.58475411e-01 1.21241105e+00
-2.74997562e-01 5.90578258e-01 -1.21327162e+00 1.22761834e+00
-4.38571386e-02 -1.06204283e+00 -6.42023981e-02 2.42100239e-01
6.47618532e-01 -4.18738425e-02 4.69138712e-01 4.71577018e-01
6.42078638e-01 -8.53955805e-01 2.75416315e-01 3.52050126e-01
7.77285576e-01 -8.25654387e-01 3.35277796e-01 4.01831210e-01
-6.69263422e-01 -5.13677776e-01 -5.56416929e-01 7.86292553e-01
-9.84275416e-02 3.38980824e-01 -7.40332663e-01 3.96743715e-01
1.06579208e+00 5.68398051e-02 -3.21297139e-01 9.47563648e-01
5.93646094e-02 6.92818224e-01 -5.33359826e-01 3.17267925e-02
-1.23522088e-01 -2.13636875e-01 5.75169027e-01 8.34913254e-01
3.12542826e-01 2.03340307e-01 6.46250397e-02 6.76095545e-01
-5.39417624e-01 4.61014420e-01 -1.05258785e-01 3.72236632e-02
9.06559527e-01 1.16181529e+00 2.78912246e-01 1.67252675e-01
-2.04405367e-01 9.14241016e-01 6.22666478e-01 3.28976929e-01
-8.65030885e-01 -2.75164485e-01 1.02524579e+00 -6.68148696e-02
4.66931671e-01 2.61526972e-01 -6.54110610e-02 -1.27336848e+00
9.91258547e-02 -1.25467253e+00 1.02608776e+00 -1.78888794e-02
-1.82072985e+00 5.11744797e-01 -2.42422819e-01 -9.48512197e-01
-3.79099734e-02 -1.37610883e-01 -4.14495230e-01 7.56615222e-01
-1.38182247e+00 -9.76332307e-01 -1.67996451e-01 1.26174629e+00
4.20726184e-03 -4.47865248e-01 8.79348636e-01 5.82958698e-01
-1.03647196e+00 1.52928233e+00 6.71862006e-01 -2.03247860e-01
9.15298939e-01 -8.28434110e-01 -5.47089539e-02 3.02720815e-01
-3.75245064e-01 6.71640396e-01 4.65028048e-01 -3.42000604e-01
-1.92869234e+00 -1.19821298e+00 5.32174468e-01 -2.04836071e-01
4.25613552e-01 -2.97242463e-01 -9.87576544e-01 7.48952091e-01
-8.07602853e-02 1.89679861e-01 1.05939162e+00 1.36652306e-01
-6.38143659e-01 -3.59714448e-01 -1.58857858e+00 4.38744187e-01
9.19548869e-01 -3.46185505e-01 -3.73826176e-02 4.68182027e-01
8.39809835e-01 -2.36359730e-01 -1.12208068e+00 1.34836763e-01
4.43331122e-01 -9.61320698e-01 8.89787972e-01 -8.07597637e-01
-1.50392622e-01 1.73978850e-01 -4.60850388e-01 -8.57011139e-01
-1.87023357e-01 -1.22656226e+00 -4.49284226e-01 1.58351469e+00
2.08456695e-01 -9.53803957e-01 1.33809853e+00 1.23041964e+00
3.23792845e-01 -7.44555771e-01 -1.04970038e+00 -5.01763344e-01
3.46770487e-03 -1.84737071e-01 1.56186962e+00 9.43767726e-01
-1.54380545e-01 1.13714254e-02 -7.43802428e-01 2.95636088e-01
9.82442856e-01 1.55923322e-01 9.84001040e-01 -1.00955033e+00
-6.78093791e-01 -4.07473087e-01 8.10370818e-02 -9.34786141e-01
2.94862062e-01 -8.96842062e-01 -1.94707602e-01 -7.93435216e-01
2.71045953e-01 -9.22022939e-01 -3.41307849e-01 6.71376467e-01
-2.69902293e-02 -2.77653664e-01 2.21386775e-01 4.30004001e-01
-7.34310925e-01 8.70211720e-01 9.50607181e-01 4.60974537e-02
-1.84489131e-01 4.05708849e-01 -8.92510116e-01 1.97595477e-01
7.28218198e-01 -4.16432768e-01 -5.19001186e-01 -5.40092647e-01
7.01941997e-02 2.12148458e-01 1.34833634e-01 -5.50285041e-01
5.02800286e-01 -4.06099051e-01 2.21894234e-02 -4.37094241e-01
1.64807364e-01 -1.20161152e+00 3.99940938e-01 -7.19016641e-02
-5.85500598e-01 -2.07780868e-01 -3.22919846e-01 7.46945441e-01
2.20678240e-01 -1.60502449e-01 8.52201164e-01 7.46288616e-03
-4.58986759e-02 9.94278014e-01 8.93635005e-02 2.21554518e-01
1.04188979e+00 3.06456864e-01 -5.52161932e-01 -3.05928677e-01
-5.15837312e-01 7.21178472e-01 5.55644274e-01 2.99513154e-02
4.02319551e-01 -1.09755611e+00 -5.02785265e-01 4.09280777e-01
-1.06041636e-02 5.95053658e-02 5.52831292e-01 5.13055205e-01
-1.07091703e-01 2.95768887e-01 1.23096161e-01 -1.70899674e-01
-1.08426249e+00 9.27766442e-01 2.77836293e-01 -3.40076178e-01
-8.52387726e-01 7.94491649e-01 9.72746313e-02 -6.00061357e-01
1.02663755e+00 5.24419069e-01 -1.25246421e-02 -3.65686865e-04
5.41492224e-01 8.62814486e-01 1.19210504e-01 -2.55326658e-01
-1.42492577e-01 1.62640169e-01 -4.05591935e-01 1.01434030e-01
1.50123954e+00 -3.27041745e-01 -5.81528284e-02 -3.04316044e-01
1.64939284e+00 -1.00845829e-01 -1.47527361e+00 -7.46504426e-01
-2.19515368e-01 -9.08581853e-01 -6.02772534e-02 -7.74019659e-01
-1.43053651e+00 5.27655959e-01 7.55336761e-01 -1.05487719e-01
1.25397587e+00 -1.69362485e-01 1.27230740e+00 5.70260763e-01
5.00097692e-01 -1.16640317e+00 -2.74682730e-01 1.98551178e-01
4.99063104e-01 -9.66659904e-01 -3.58814970e-02 -6.25672415e-02
-8.88605416e-01 9.83020008e-01 5.47069311e-01 3.96084070e-01
7.61334002e-01 1.04451835e-01 -1.91079810e-01 2.50314013e-03
-8.80510807e-01 3.48022342e-01 -2.48878166e-01 4.69268173e-01
-2.44635180e-01 -2.18499992e-02 -1.76364526e-01 1.15863419e+00
-1.01649024e-01 5.24554625e-02 6.16997927e-02 8.62510979e-01
-1.21431142e-01 -1.47594440e+00 -5.41324019e-01 3.01930010e-01
-6.71720922e-01 2.05557123e-01 -6.51161075e-02 1.34958416e-01
-6.75130775e-03 7.42204070e-01 -4.04675663e-01 -5.08261979e-01
2.45227873e-01 1.49426058e-01 8.51586685e-02 -1.73567072e-01
-8.49424005e-01 -1.32277578e-01 -2.43654460e-01 -6.63573086e-01
2.80914277e-01 -8.44046652e-01 -1.05690718e+00 -7.59061337e-01
-2.40477949e-01 4.42902178e-01 7.45614529e-01 5.26251912e-01
6.39682353e-01 -2.27031380e-01 1.25247979e+00 -4.21497613e-01
-1.62891245e+00 -3.84468615e-01 -8.32320392e-01 5.79026043e-01
2.43291229e-01 -3.60932648e-01 -6.21327758e-01 -3.72795910e-01] | [5.849506855010986, 6.3966064453125] |
431b1130-c3f0-4fbe-bad3-69c3a999ccf1 | invertible-image-dataset-protection | 2112.14420 | null | https://arxiv.org/abs/2112.14420v1 | https://arxiv.org/pdf/2112.14420v1.pdf | Invertible Image Dataset Protection | Deep learning has achieved enormous success in various industrial applications. Companies do not want their valuable data to be stolen by malicious employees to train pirated models. Nor do they wish the data analyzed by the competitors after using them online. We propose a novel solution for dataset protection in this scenario by robustly and reversibly transform the images into adversarial images. We develop a reversible adversarial example generator (RAEG) that introduces slight changes to the images to fool traditional classification models. Even though malicious attacks train pirated models based on the defensed versions of the protected images, RAEG can significantly weaken the functionality of these models. Meanwhile, the reversibility of RAEG ensures the performance of authorized models. Extensive experiments demonstrate that RAEG can better protect the data with slight distortion against adversarial defense than previous methods. | ['Xinpeng Zhang', 'Zhenxing Qian', 'Sheng Li', 'Qichao Ying', 'Xianhan Zeng', 'Kejiang Chen'] | 2021-12-29 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 5.55816352e-01 4.32672888e-01 5.96326543e-03 -3.76509845e-01
-5.07936835e-01 -1.30361879e+00 4.10739422e-01 -6.89574718e-01
-2.01461762e-01 6.48968995e-01 -4.90669519e-01 -4.84256297e-01
1.58682063e-01 -1.13276613e+00 -1.07747459e+00 -7.60144830e-01
3.61982584e-01 3.43923382e-02 3.44104096e-02 -2.80350029e-01
2.66616583e-01 8.30027521e-01 -1.08349705e+00 6.65620625e-01
8.16075325e-01 6.11046731e-01 -1.72445476e-01 6.91382587e-01
3.93870592e-01 8.74400318e-01 -1.19073999e+00 -1.03373671e+00
1.05924153e+00 -2.13862941e-01 -7.50889719e-01 -5.73550127e-02
5.52129805e-01 -7.67394662e-01 -7.73974955e-01 1.42122006e+00
2.05605358e-01 -2.85816729e-01 8.71497467e-02 -1.63272035e+00
-1.33590758e+00 7.36610651e-01 -4.37386841e-01 -8.90629515e-02
-1.09083131e-01 5.21023929e-01 5.32819986e-01 -2.87918240e-01
4.41973209e-01 1.19284272e+00 2.64568567e-01 9.46193516e-01
-1.06537580e+00 -1.13805819e+00 2.77761035e-02 2.77324626e-03
-1.09170187e+00 -1.32423580e-01 8.63814116e-01 -1.90442756e-01
2.74675459e-01 6.94250584e-01 -6.15149178e-02 1.49062216e+00
4.32432532e-01 3.65481079e-01 1.17456424e+00 -1.43695176e-01
-2.64394462e-01 5.67072451e-01 6.57465234e-02 6.85129762e-01
7.02724218e-01 3.36656511e-01 1.37278317e-02 -2.45954752e-01
7.50439107e-01 3.93575072e-01 -3.05947602e-01 -1.79726079e-01
-8.33918750e-01 7.06364870e-01 6.04859233e-01 7.87709877e-02
-7.64944553e-02 1.79283917e-01 3.04719448e-01 4.85352516e-01
2.90729646e-02 9.02749717e-01 -4.49486822e-01 6.72187388e-01
-3.32590431e-01 8.60990062e-02 5.64462781e-01 9.21260238e-01
4.57346737e-01 4.69356477e-01 8.73610936e-03 2.32602805e-01
1.49731208e-02 6.68293118e-01 4.65763628e-01 -9.17355299e-01
5.21565259e-01 5.78080118e-01 5.76827787e-02 -1.30315638e+00
4.14603949e-01 -4.47551847e-01 -7.26247072e-01 5.02222598e-01
5.21070719e-01 -1.97131991e-01 -9.15938854e-01 1.53979278e+00
3.55997011e-02 -1.52831916e-02 2.02258348e-01 5.09672523e-01
4.39472526e-01 7.58955717e-01 -2.13469043e-01 3.39194119e-01
9.39998209e-01 -8.52676690e-01 -7.31069803e-01 -1.98293358e-01
3.04467946e-01 -5.03341377e-01 1.24675941e+00 5.61123669e-01
-8.75892162e-01 -9.02481616e-01 -1.52440012e+00 1.46720231e-01
-5.94289064e-01 -1.16189018e-01 4.14486915e-01 1.48483253e+00
-4.00345951e-01 5.91895640e-01 -4.09989506e-01 6.04490340e-01
9.29048479e-01 6.93034351e-01 -5.67057967e-01 -1.72463715e-01
-1.47751915e+00 5.12907088e-01 3.05071861e-01 2.22007528e-01
-1.38783801e+00 -4.66808438e-01 -7.96044290e-01 -1.06938165e-02
3.61665785e-01 -2.51953036e-01 1.06663644e+00 -1.09189570e+00
-1.10759115e+00 1.05260372e+00 5.82174420e-01 -8.14319730e-01
9.54698563e-01 -2.12123051e-01 -6.96580708e-01 8.27365890e-02
-2.61048496e-01 3.89639705e-01 1.17500961e+00 -1.55376470e+00
-5.03149688e-01 -3.37689906e-01 5.14481604e-01 -5.05407453e-01
-7.48659015e-01 -1.20567188e-01 9.37673748e-02 -7.93962777e-01
-3.87744576e-01 -1.14834106e+00 -2.94152617e-01 -1.32602096e-01
-8.52628887e-01 4.53069091e-01 1.40674031e+00 -5.44644117e-01
7.23784804e-01 -2.31188583e+00 -6.00952506e-01 3.15341741e-01
1.70422092e-01 9.89124596e-01 -2.54500747e-01 -1.73623636e-02
-3.20291042e-01 7.72365570e-01 -8.43439028e-02 3.19062501e-01
1.81379318e-01 3.17784101e-01 -1.08140004e+00 5.37971199e-01
1.07676432e-01 1.20417571e+00 -6.09963894e-01 -1.01302721e-01
4.15542461e-02 3.98755014e-01 -4.28698748e-01 8.92265961e-02
-3.90026867e-02 2.60905504e-01 -6.22426152e-01 4.81318116e-01
1.10668480e+00 8.53265810e-04 2.35650897e-01 2.38079038e-02
6.02633715e-01 -9.04420167e-02 -8.03611338e-01 6.37112141e-01
-1.83370516e-01 4.97744262e-01 -9.07666534e-02 -8.75138164e-01
1.03046417e+00 1.47825986e-01 -2.00556777e-02 -5.71418881e-01
1.33864582e-01 1.27072975e-01 6.13110699e-02 -3.62031460e-01
3.18138510e-01 -6.36389703e-02 -4.88397837e-01 4.06933963e-01
-4.70147878e-01 9.07931253e-02 -3.78258318e-01 5.28998934e-02
9.80342209e-01 -4.45733815e-02 -3.37281227e-01 2.32332423e-01
8.30967605e-01 -2.42102891e-01 7.37793505e-01 1.09346330e+00
-2.53455281e-01 4.83349562e-01 4.71277654e-01 -7.05940008e-01
-1.22665405e+00 -1.22953928e+00 2.65342772e-01 6.97689950e-01
1.79638088e-01 1.48798987e-01 -1.01598179e+00 -1.63422894e+00
1.82101965e-01 7.37934887e-01 -5.12383044e-01 -7.95780003e-01
-7.23330021e-01 -4.03594792e-01 1.20703495e+00 4.96298969e-01
1.16461730e+00 -1.07743382e+00 6.01351000e-02 -1.64096564e-01
5.77420145e-02 -1.17047679e+00 -5.96154034e-01 -1.39103487e-01
-4.40297931e-01 -1.12381327e+00 -3.54803622e-01 -8.01859438e-01
9.94915068e-01 4.44775641e-01 5.52785933e-01 2.86185980e-01
-3.27590883e-01 4.44066636e-02 -1.28625199e-01 -6.88073754e-01
-1.05605745e+00 -1.11428656e-01 1.20242350e-01 1.55500576e-01
1.56514317e-01 -2.64611989e-01 -2.42345572e-01 6.77695155e-01
-1.41170382e+00 -5.92394948e-01 4.43698555e-01 6.39639735e-01
4.83024001e-01 8.73515606e-01 6.34022236e-01 -1.50934839e+00
4.89923775e-01 2.74661165e-02 -7.92576611e-01 3.76681060e-01
-6.31952882e-01 -1.09252997e-01 1.19142365e+00 -7.92735457e-01
-9.97615814e-01 -1.45849986e-02 2.09713317e-02 -7.11182177e-01
-8.58217031e-02 -2.81687260e-01 -7.77892768e-01 -5.61292470e-01
7.64921486e-01 1.65270656e-01 -9.26126167e-02 -3.02873403e-01
2.99469084e-01 6.05577469e-01 9.32438314e-01 -3.50277454e-01
1.69452596e+00 4.98518527e-01 4.90531996e-02 -2.04008266e-01
-7.73436785e-01 4.93130058e-01 -3.30699205e-01 -2.03728536e-03
5.82537889e-01 -5.84209085e-01 -9.45769012e-01 7.84303963e-01
-9.86965179e-01 -2.81824525e-02 -3.45218986e-01 -6.01603799e-02
-1.80321023e-01 5.81153989e-01 -6.60123825e-01 -5.31062782e-01
-1.50080875e-01 -1.17831361e+00 4.05044824e-01 1.59611985e-01
2.09078580e-01 -5.25465727e-01 -6.02558136e-01 9.88555014e-01
2.42941618e-01 8.20085883e-01 9.64285314e-01 -9.93964553e-01
-9.00117457e-01 -1.10970616e+00 2.05949619e-01 1.10499465e+00
3.11189264e-01 1.17134117e-01 -1.14957821e+00 -4.20934349e-01
3.97786617e-01 -3.06645572e-01 6.68387949e-01 -2.07684413e-01
1.78027177e+00 -9.39011455e-01 -2.68834174e-01 6.31770670e-01
1.10532570e+00 5.87975442e-01 1.12796533e+00 4.98062551e-01
1.04146659e+00 6.36725247e-01 5.23355007e-01 -6.78527728e-02
-4.17037040e-01 2.38527313e-01 8.32602203e-01 -2.55524904e-01
3.47744793e-01 -4.59144920e-01 6.35849714e-01 1.12872735e-01
7.34959543e-02 -5.03982782e-01 -4.15776372e-01 5.38518056e-02
-1.33342743e+00 -1.26249552e+00 -2.79869735e-01 2.20010471e+00
6.00313187e-01 6.44575357e-01 -2.66265690e-01 3.02525312e-01
1.11085713e+00 2.37754822e-01 -6.73990607e-01 -7.18613744e-01
-1.60358950e-01 2.88950622e-01 9.95588839e-01 1.97721928e-01
-1.28932178e+00 9.19041812e-01 6.40833235e+00 7.81140745e-01
-9.06358719e-01 1.19193882e-01 1.02506745e+00 -5.54047190e-02
-4.20326650e-01 -1.33323073e-01 -8.12700391e-01 5.65154374e-01
8.22402835e-01 -3.83262873e-01 3.84280592e-01 1.43460870e+00
-1.53893292e-01 7.89081335e-01 -9.34086025e-01 4.12619233e-01
9.03214980e-03 -1.43972600e+00 3.76563191e-01 4.91654754e-01
8.17062497e-01 -7.69424558e-01 8.93909872e-01 3.63573372e-01
6.55014396e-01 -1.19389868e+00 6.41076386e-01 3.52772683e-01
6.70337975e-01 -1.29774010e+00 7.33317912e-01 4.05229628e-01
-5.13777792e-01 -5.58389366e-01 -6.74040556e-01 3.56735975e-01
-3.34409058e-01 2.79516220e-01 -7.73684859e-01 5.82137287e-01
4.79354769e-01 -1.16722904e-01 -6.60919309e-01 2.14831829e-01
-5.80142617e-01 5.71834385e-01 2.35103816e-01 4.50103760e-01
2.31621101e-01 9.59985144e-03 4.71174687e-01 6.71929657e-01
1.13943085e-01 -2.05615923e-01 7.22365174e-03 1.23417544e+00
-7.18893528e-01 -5.26797652e-01 -1.18086529e+00 -5.06324351e-01
4.76331055e-01 1.20563591e+00 -3.56983542e-01 -9.78895649e-02
-1.23166785e-01 1.01319683e+00 -2.97155797e-01 1.51386589e-01
-1.13862371e+00 -7.75962055e-01 6.32244885e-01 2.90259361e-01
2.07593352e-01 -8.22845176e-02 -1.02751769e-01 -7.88397670e-01
1.78635895e-01 -1.48572540e+00 3.39689165e-01 -7.54620016e-01
-1.46017313e+00 7.51591682e-01 -4.40683663e-01 -1.21090782e+00
-8.55517164e-02 -6.76179051e-01 -5.49849391e-01 7.19425201e-01
-1.20297551e+00 -1.36382055e+00 -1.66826844e-02 8.42963040e-01
2.16119185e-01 -6.26114190e-01 6.62211239e-01 1.23645715e-01
-7.15628326e-01 1.08227634e+00 -6.11272231e-02 6.77692890e-01
5.67395627e-01 -9.74774003e-01 8.27390194e-01 1.26631737e+00
2.06303984e-01 7.99581468e-01 4.49469745e-01 -8.40077460e-01
-1.21036398e+00 -1.45911932e+00 4.44678307e-01 -6.85884476e-01
2.89343596e-01 -6.28010690e-01 -9.22118068e-01 1.05663431e+00
1.76602781e-01 1.84539616e-01 6.78484440e-01 -6.72015369e-01
-6.96562648e-01 -2.74785757e-01 -1.65601385e+00 5.13359666e-01
6.56644821e-01 -7.90096998e-01 -2.63771951e-01 4.97937739e-01
1.19353485e+00 -2.34857708e-01 -4.93169188e-01 3.89593244e-01
3.75752807e-01 -7.31188536e-01 1.40163875e+00 -1.07166576e+00
5.36562085e-01 -2.88111329e-01 -7.99368620e-02 -7.38647938e-01
-2.56215036e-01 -8.02428663e-01 -4.88177538e-02 1.37266409e+00
3.73425633e-01 -9.15713072e-01 9.46584821e-01 8.28210354e-01
1.28502652e-01 -8.27236623e-02 -4.72560316e-01 -1.27548933e+00
3.07131737e-01 -4.27075615e-03 1.09515083e+00 1.40442336e+00
-6.75013304e-01 -4.32719052e-01 -7.40994573e-01 6.92561150e-01
7.07785666e-01 -1.01800762e-01 1.00150156e+00 -8.55956554e-01
-2.60171950e-01 7.59061500e-02 -3.92844200e-01 -4.46361423e-01
3.83145392e-01 -9.57531869e-01 -2.20009446e-01 -7.74729311e-01
-5.77500165e-02 -1.13712788e-01 -5.94172895e-01 8.07909966e-01
-2.28778124e-01 5.63702822e-01 2.41833255e-01 1.06417157e-01
-2.64895894e-03 -5.90678975e-02 1.44265175e+00 -4.41567153e-01
1.54015318e-01 3.22918832e-01 -1.18781900e+00 6.82269216e-01
1.06069088e+00 -9.19240415e-01 -7.11686373e-01 -9.36620533e-02
1.34181961e-01 -4.01310861e-01 7.38114595e-01 -9.47268009e-01
-1.88526660e-01 -3.05364102e-01 7.27736413e-01 -2.62540072e-01
-1.56608954e-01 -1.08400798e+00 3.96153212e-01 7.67125666e-01
-5.14766932e-01 9.07705352e-03 4.39138478e-03 5.79834461e-01
7.37534165e-02 -4.72541630e-01 9.81356859e-01 -3.96200210e-01
-3.27576965e-01 3.44513476e-01 -3.31710249e-01 -3.79589975e-01
1.36465549e+00 -3.13266873e-01 -7.42982507e-01 -2.76758760e-01
-6.82861805e-01 -6.67909682e-02 6.51182234e-01 5.59223533e-01
5.49314320e-01 -1.29637980e+00 -4.61524516e-01 4.71148551e-01
-3.06116700e-01 -2.72990108e-01 2.97522455e-01 -1.76018283e-01
-5.27971148e-01 1.12284288e-01 -4.61656392e-01 1.17189705e-01
-1.52743673e+00 1.33413672e+00 4.81046617e-01 -4.49217737e-01
-4.54710454e-01 7.53994167e-01 4.26219463e-01 -5.52999198e-01
2.77284998e-02 2.03877479e-01 7.84476623e-02 -5.25865316e-01
6.01245582e-01 3.26780707e-01 -9.75767002e-02 -2.69382507e-01
-1.97678372e-01 -5.23838326e-02 -5.90726852e-01 2.46317968e-01
1.25771248e+00 2.31466219e-01 7.92575476e-04 -3.45176518e-01
1.20739543e+00 3.87599558e-01 -1.30096257e+00 1.59588471e-01
-4.03828055e-01 -9.25916672e-01 -1.61315486e-01 -8.65675628e-01
-1.56532407e+00 8.24306548e-01 5.59028387e-01 6.85524881e-01
1.02068162e+00 -5.83881676e-01 1.02255213e+00 7.56680727e-01
3.17880601e-01 -9.23210919e-01 4.44628954e-01 5.22674955e-02
7.29190171e-01 -8.57554078e-01 -5.79443388e-02 -5.96659660e-01
-8.83025646e-01 9.47985411e-01 7.38631785e-01 -3.20499271e-01
2.11624876e-01 3.84652823e-01 6.87100068e-02 9.04259309e-02
-1.78543359e-01 4.94932950e-01 1.17227556e-02 1.03845739e+00
-4.08292383e-01 -1.83548614e-01 1.51681751e-01 7.89400041e-01
-3.09463918e-01 -1.81327060e-01 7.59218931e-01 1.02368033e+00
-2.25671872e-01 -1.56510353e+00 -6.47589445e-01 9.05894488e-03
-9.64256763e-01 1.79196224e-01 -7.89187789e-01 9.49707925e-01
3.92688036e-01 9.55180347e-01 -3.38340312e-01 -7.93493390e-01
4.95258451e-01 -4.58971448e-02 1.94799498e-01 -3.27388257e-01
-9.09285665e-01 -3.85627329e-01 -5.70916049e-02 -4.29111511e-01
1.02783717e-01 -2.11796239e-01 -1.01944566e+00 -6.25994384e-01
-3.09748262e-01 -1.34733394e-02 4.22693223e-01 5.21338999e-01
3.63841474e-01 6.56826317e-01 1.47920871e+00 -2.19377294e-01
-9.63976622e-01 -2.65476495e-01 -6.13492608e-01 6.08080447e-01
2.05639794e-01 -1.71171442e-01 -5.39536417e-01 3.29135925e-01] | [5.612184047698975, 7.984772682189941] |
cd550104-530e-4398-bf3d-4b4d8fcc4f3e | a-simple-method-for-detecting-chaos-in-nature | 1904.00986 | null | http://arxiv.org/abs/1904.00986v3 | http://arxiv.org/pdf/1904.00986v3.pdf | A simple method for detecting chaos in nature | Chaos, or exponential sensitivity to small perturbations, appears everywhere
in nature. Moreover, chaos is predicted to play diverse functional roles in
living systems. A method for detecting chaos from empirical measurements should
therefore be a key component of the biologist's toolkit. But, classic
chaos-detection tools are highly sensitive to measurement noise and break down
for common edge cases, making it difficult to detect chaos in domains, like
biology, where measurements are noisy. However, newer tools promise to overcome
these limitations. Here, we combine several such tools into an automated
processing pipeline, and show that our pipeline can detect the presence (or
absence) of chaos in noisy recordings, even for difficult edge cases. As a
first-pass application of our pipeline, we show that heart rate variability is
not chaotic as some have proposed, and instead reflects a stochastic process in
both health and disease. Our tool is easy-to-use and freely available. | [] | 2020-01-10 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [-1.19798128e-02 -5.35905480e-01 5.46526074e-01 1.00334540e-01
-1.01915061e-01 -9.15295124e-01 5.32659888e-01 5.18213511e-01
-2.31444657e-01 8.65617096e-01 -6.56320900e-02 -3.28178108e-01
4.18020934e-02 -4.92738575e-01 -3.46551806e-01 -9.02316988e-01
-4.63899195e-01 2.88508505e-01 4.39676017e-01 -2.12746978e-01
2.77874440e-01 6.12898886e-01 -1.11307788e+00 -8.23364779e-02
5.09495974e-01 4.06179041e-01 3.33185215e-03 1.18221653e+00
2.47092664e-01 4.43736911e-01 -7.86964834e-01 2.71992743e-01
-5.60573265e-02 -9.15946543e-01 -3.86407167e-01 -5.10015070e-01
-3.17727774e-01 1.17065407e-01 -4.45140041e-02 8.08705389e-01
8.94862950e-01 -3.44316900e-01 3.86750668e-01 -7.89443970e-01
-9.58952308e-02 1.75418645e-01 -2.01149523e-01 6.69980824e-01
3.11653107e-01 7.67226815e-01 4.64673102e-01 -3.68519187e-01
6.38779640e-01 9.53645110e-01 1.15472281e+00 3.57793838e-01
-1.81270432e+00 -4.30700690e-01 -4.62567329e-01 -2.82915473e-01
-1.18804204e+00 -7.02779770e-01 4.73370373e-01 -6.74746990e-01
1.04457319e+00 4.51512247e-01 1.07580698e+00 9.88271415e-01
7.88607359e-01 5.72141521e-02 1.42330492e+00 1.59969345e-01
4.37520057e-01 -1.20822683e-01 -2.48462200e-01 3.48245770e-01
4.47436064e-01 8.43012780e-02 -3.44568670e-01 -4.46858555e-01
8.35564554e-01 1.04694322e-01 -4.25306171e-01 3.00899625e-01
-1.59861326e+00 3.03545237e-01 -8.28854069e-02 5.15009344e-01
-2.29335576e-01 4.38059002e-01 4.06594753e-01 4.68394697e-01
4.49732304e-01 8.26333821e-01 -5.95491350e-01 -8.67375672e-01
-9.87241685e-01 1.46143809e-01 1.05223858e+00 3.30065459e-01
4.40346748e-01 -1.27734318e-01 1.60322681e-01 4.10669476e-01
7.50603229e-02 4.08750951e-01 6.46525085e-01 -1.02155375e+00
-4.20902461e-01 5.12111127e-01 1.69536304e-02 -1.06892061e+00
-1.01745772e+00 -4.46435839e-01 -1.15543699e+00 2.66540796e-02
7.58883417e-01 -2.33833522e-01 -5.27869344e-01 1.59452248e+00
2.42743805e-01 3.69887650e-01 -1.07577398e-01 7.51705229e-01
6.77547097e-01 3.73852223e-01 -1.64146200e-01 -5.40619910e-01
1.38455892e+00 1.96533039e-01 -7.63926327e-01 4.35605273e-02
5.56085348e-01 -8.26147914e-01 8.64462912e-01 2.93219775e-01
-7.61406958e-01 -3.95516939e-02 -8.02570701e-01 2.07087144e-01
-3.99724007e-01 -5.78418434e-01 4.29755747e-01 6.12383544e-01
-1.07196522e+00 1.04683888e+00 -1.31184328e+00 -8.33196402e-01
1.94595218e-01 9.09609050e-02 -1.74078375e-01 5.49087942e-01
-8.40557814e-01 8.65465224e-01 -2.37133443e-01 1.45374462e-01
-5.68102002e-01 -7.58290052e-01 -4.77438986e-01 4.36779298e-03
-2.24365488e-01 -6.77957773e-01 8.18094254e-01 -1.26433089e-01
-1.35122430e+00 8.39595616e-01 -3.06784958e-01 -3.15356225e-01
6.63437903e-01 3.19221288e-01 -6.32148981e-02 1.42019719e-01
-1.56527027e-01 1.07003510e-01 4.63175863e-01 -8.44587743e-01
1.85513422e-02 -3.03616583e-01 -7.26672113e-01 -3.47006530e-01
3.87736806e-03 1.14104059e-02 1.92787617e-01 -4.45584953e-01
1.76544234e-01 -1.03862786e+00 -4.39327240e-01 1.22561239e-01
-2.85363138e-01 2.04498947e-01 8.23069930e-01 -5.14082551e-01
1.13558817e+00 -2.01055217e+00 -3.20842832e-01 1.79127485e-01
5.44535100e-01 1.72572434e-01 3.81801158e-01 7.66942739e-01
3.05511445e-01 4.64839041e-01 -6.05437875e-01 -8.27156082e-02
-1.41151264e-01 3.60420458e-02 -1.60967186e-01 8.51263583e-01
3.98734868e-01 9.30970728e-01 -1.08967805e+00 -2.44563356e-01
1.90346301e-01 6.39744699e-01 -1.85240299e-01 2.22632550e-02
6.33778498e-02 1.07341361e+00 -7.08422624e-03 6.74301624e-01
3.99146378e-01 -5.18067420e-01 1.18743919e-01 4.96046156e-01
-4.51093525e-01 3.20961565e-01 -8.54454458e-01 1.22925270e+00
-3.45895216e-02 1.10968423e+00 1.28928483e-01 -8.32507968e-01
1.04937255e+00 4.13088411e-01 6.70744956e-01 -1.90706149e-01
5.59448637e-02 3.89233798e-01 5.76079071e-01 -6.79379761e-01
-4.63972464e-02 -3.04615349e-01 -9.63409767e-02 5.25854290e-01
-1.97362721e-01 -6.16792560e-01 1.09200262e-01 5.83999883e-03
1.78893054e+00 -9.01471451e-02 4.81930643e-01 -4.53947902e-01
1.60736777e-02 -1.95332915e-02 7.69993126e-01 8.39455068e-01
-8.00459683e-01 8.64178479e-01 9.56152856e-01 -6.44209921e-01
-1.05794144e+00 -6.57753050e-01 -4.29064691e-01 2.38297820e-01
-1.67568717e-02 -6.77980721e-01 -4.69599605e-01 1.08487777e-01
9.48119238e-02 -7.26513416e-02 -6.20432556e-01 -1.28262609e-01
-5.53993821e-01 -1.25596154e+00 8.95792127e-01 -4.44516577e-02
1.50082245e-01 -1.19907701e+00 -1.22998786e+00 5.37091076e-01
-5.30443825e-02 -1.14770913e+00 -5.74544072e-02 4.72165018e-01
-8.01511228e-01 -9.90328074e-01 -6.07688725e-01 -2.62147605e-01
3.19927752e-01 2.37888657e-02 1.06768155e+00 5.79308927e-01
-9.15469110e-01 2.05454797e-01 -9.69208404e-03 -5.68568885e-01
-5.59553266e-01 -2.74748176e-01 3.78715277e-01 -4.51507390e-01
2.38153070e-01 -1.00872111e+00 -9.94724214e-01 2.80680269e-01
-5.94230890e-01 -4.52084571e-01 1.97174162e-01 8.54455590e-01
4.10101563e-01 5.29104285e-03 6.28450751e-01 -4.44897234e-01
8.93152535e-01 -5.88787138e-01 -6.90675318e-01 -2.37640887e-01
-4.99618024e-01 7.04547856e-03 8.82901669e-01 -3.88882667e-01
-2.85054538e-02 2.40325227e-01 -2.12894708e-01 -1.17332384e-01
-4.24154103e-01 5.58995545e-01 4.76423860e-01 -6.78025186e-02
6.79796040e-01 5.03443003e-01 3.12319577e-01 -1.57159165e-01
-2.69709647e-01 6.25715792e-01 2.55781025e-01 -1.89821720e-01
3.68948489e-01 7.49960721e-01 3.37144583e-01 -1.44336402e+00
3.85345034e-02 -3.96324575e-01 -4.83821511e-01 -2.03409284e-01
6.11120224e-01 -5.95043361e-01 -1.21114707e+00 5.27705550e-01
-9.61057901e-01 -6.80332065e-01 -3.14465106e-01 4.41141158e-01
-5.11462331e-01 3.45170826e-01 -6.64702296e-01 -9.44505155e-01
-3.13055843e-01 -1.05010414e+00 9.60908890e-01 3.20051759e-01
-6.22543931e-01 -1.05643725e+00 6.36112273e-01 -4.28388298e-01
7.56482184e-01 6.43196106e-01 4.19728428e-01 -5.73803961e-01
-2.39854872e-01 -1.89628154e-01 2.09568769e-01 -1.40199944e-01
3.07663590e-01 6.98353648e-01 -1.06248438e+00 -7.09762499e-02
2.19021425e-01 -3.31961028e-02 8.85778546e-01 5.94631016e-01
6.42256975e-01 -1.31033659e-01 -3.06760699e-01 7.13737547e-01
1.16082680e+00 -1.21412992e-01 6.60739362e-01 2.04818517e-01
7.64446482e-02 7.92841911e-01 -2.64462903e-02 6.32405162e-01
8.03029984e-02 3.50970536e-01 4.50400859e-02 -8.80641341e-02
6.39314130e-02 1.33581802e-01 2.66107410e-01 8.41163397e-01
-2.67406041e-03 9.93442535e-02 -1.19074500e+00 5.02328753e-01
-1.71958017e+00 -1.31494129e+00 -6.59988880e-01 2.17658210e+00
8.99299383e-01 2.40137443e-01 4.35815752e-01 1.80698901e-01
3.93172443e-01 -3.37466449e-01 -7.09348261e-01 -4.32889998e-01
-1.88032001e-01 5.20776324e-02 3.11726451e-01 3.43076944e-01
-8.76728415e-01 6.91252947e-01 7.86854219e+00 -8.20380636e-03
-1.50235641e+00 -1.57807976e-01 5.13815701e-01 -7.39027485e-02
-3.24581414e-02 1.58705682e-01 -2.65523821e-01 7.99069822e-01
1.23948634e+00 -2.23899469e-01 4.52825904e-01 2.28990331e-01
8.95538270e-01 -3.11726719e-01 -8.15244257e-01 9.48457778e-01
-6.41247272e-01 -1.25207663e+00 -9.82217133e-01 1.83468029e-01
2.77631670e-01 3.34165275e-01 -8.58452693e-02 -3.09966411e-02
2.63610005e-01 -1.11449528e+00 1.54450729e-01 5.47088444e-01
7.43535042e-01 -3.00110817e-01 7.53002644e-01 6.83317840e-01
-9.88888144e-01 2.61933386e-01 -4.92041320e-01 -4.64380980e-01
2.79660374e-01 1.18816781e+00 -1.00930452e+00 -2.52571017e-01
8.84370565e-01 8.25987518e-01 -5.91354787e-01 1.33961320e+00
4.73284684e-02 8.81632507e-01 -7.76978910e-01 -2.91411042e-01
-3.57108206e-01 -1.65800512e-01 7.76586056e-01 1.52102780e+00
4.16356534e-01 1.47661224e-01 -2.36713156e-01 1.04629731e+00
4.17974412e-01 -3.30654383e-01 -8.36013198e-01 -4.73816454e-01
4.99546468e-01 1.45256090e+00 -1.26414621e+00 -2.72601664e-01
6.12124708e-03 7.57271826e-01 -8.05878788e-02 8.19509551e-02
-3.64367783e-01 -6.27810061e-01 9.75384593e-01 9.23151597e-02
-1.49910860e-02 -5.05669236e-01 -5.89553654e-01 -1.22628844e+00
-6.19778596e-02 -7.88393557e-01 7.33635426e-02 -4.63414460e-01
-1.14898634e+00 2.75466114e-01 -4.31712151e-01 -1.29796183e+00
-2.40053564e-01 -2.89194912e-01 -8.91523123e-01 6.25867546e-01
-1.09626245e+00 -5.67138731e-01 -3.00480515e-01 2.42498145e-01
1.78444028e-01 2.83554375e-01 9.80396807e-01 -5.87374792e-02
-6.52381003e-01 1.80217735e-02 4.14747149e-01 -2.15865880e-01
7.37126350e-01 -1.46240366e+00 7.23004758e-01 7.31264174e-01
-1.11344628e-01 9.44125295e-01 1.36844456e+00 -6.78762674e-01
-1.38839936e+00 -6.88849330e-01 9.01623785e-01 -5.86221218e-01
9.22870278e-01 -8.12362492e-01 -1.08668160e+00 2.80177355e-01
-1.20190568e-01 3.91639203e-01 7.29242563e-01 1.31003391e-02
-6.95773289e-02 1.16485238e-01 -1.16192293e+00 4.89113927e-01
7.21487701e-01 -3.55254173e-01 -3.06395799e-01 4.41376835e-01
2.70958364e-01 -1.32764995e-01 -1.00799811e+00 1.83965623e-01
1.05876327e+00 -1.19842660e+00 5.83080053e-01 -1.89078644e-01
4.85727414e-02 -6.28568649e-01 4.61485922e-01 -1.34803653e+00
-2.85493821e-01 -1.33046699e+00 -2.87876949e-02 9.92372572e-01
2.72118777e-01 -9.91272688e-01 3.61598641e-01 2.51878202e-01
4.68167290e-02 -4.30164099e-01 -1.07377720e+00 -9.09160316e-01
1.83722675e-01 -1.72930732e-01 2.31748968e-01 9.96441603e-01
6.81451797e-01 6.42790124e-02 -5.24823479e-02 -1.03708692e-02
3.90705854e-01 -1.69928838e-02 8.89601409e-01 -1.47577131e+00
-3.45548064e-01 -5.94654143e-01 -9.01326001e-01 -4.64718759e-01
-4.69058365e-01 -4.45854634e-01 3.99573624e-01 -1.22167933e+00
1.24184545e-02 -4.49984342e-01 -3.01603734e-01 2.12413833e-01
-2.97101170e-01 6.91213489e-01 -5.44651709e-02 5.97411036e-01
-2.52637118e-01 1.04476579e-01 1.02288103e+00 2.60217518e-01
-5.62557399e-01 -9.81510710e-03 -5.80641448e-01 6.29227340e-01
1.01473856e+00 -7.91483045e-01 1.49541482e-01 2.04354733e-01
4.41326886e-01 -1.27150565e-01 4.42820281e-01 -1.26828957e+00
1.50654331e-01 -1.05617464e-01 4.47561353e-01 -2.52734244e-01
1.45223737e-01 -4.42112327e-01 3.32692891e-01 1.01029170e+00
-1.29200127e-02 3.13407153e-01 3.35411400e-01 4.84050751e-01
1.20208904e-01 2.33089700e-01 8.74675810e-01 -4.52734172e-01
1.81540295e-01 -1.36641279e-01 -9.36734438e-01 9.25238729e-02
7.84165263e-01 -1.39593899e-01 -6.96363389e-01 -5.55946529e-01
-5.75377285e-01 7.01853335e-02 8.40519428e-01 -1.23545617e-01
3.61035675e-01 -5.94317853e-01 -8.55949998e-01 2.44398266e-01
-6.06819801e-02 -2.37413019e-01 -7.35439807e-02 1.54216576e+00
-9.13119078e-01 2.85961390e-01 -9.93246064e-02 -1.04050064e+00
-1.17324817e+00 4.59480286e-01 3.81206304e-01 1.54327247e-02
-8.14750433e-01 5.55347264e-01 -2.38624036e-01 -2.82573462e-01
-2.55792052e-01 -6.25672519e-01 3.93780693e-02 -2.01269135e-01
6.58792913e-01 3.05094451e-01 -3.09120808e-02 -4.15100127e-01
-6.44817889e-01 5.76715112e-01 5.08175015e-01 -1.19709007e-01
1.40833914e+00 -2.89604515e-01 -4.76066917e-01 1.08380651e+00
8.50343049e-01 1.44845679e-01 -1.04426992e+00 3.94395411e-01
1.48242991e-02 -9.69028026e-02 -5.31310380e-01 -5.04259169e-01
-2.99151689e-01 1.00386560e+00 5.08152962e-01 9.30601716e-01
1.01712441e+00 1.02716265e-02 6.78014755e-01 2.37557694e-01
8.80650952e-02 -7.47940481e-01 -3.23616534e-01 4.45521653e-01
5.55194318e-01 -1.07568848e+00 5.91644384e-02 -1.56132402e-02
-1.19896919e-01 1.19191623e+00 2.40472723e-02 -3.15095872e-01
8.06266069e-01 6.06760502e-01 2.81031013e-01 -2.56168008e-01
-1.20302224e+00 -1.74833968e-01 -2.99011260e-01 6.86448872e-01
7.29171395e-01 -8.93181339e-02 -6.67045414e-01 1.39132246e-01
-4.25182134e-01 2.96375975e-02 8.00713062e-01 1.03844619e+00
-5.41989625e-01 -6.88306093e-01 -3.98593366e-01 5.08436978e-01
-7.15297103e-01 -7.09379613e-02 -7.06091344e-01 4.04486448e-01
-7.50473514e-02 1.08907866e+00 1.84444338e-02 -4.88762110e-01
-3.95724829e-03 1.54025063e-01 1.68587983e-01 -3.45444709e-01
-9.10176933e-01 3.15681905e-01 -3.69770676e-01 -5.67329228e-01
-1.65718257e-01 -7.70998061e-01 -1.10335422e+00 -4.21627343e-01
-2.33928248e-01 4.41716313e-02 7.91385472e-01 8.70172322e-01
7.68134594e-01 4.00451005e-01 3.74860197e-01 -6.31296217e-01
-8.82958323e-02 -1.02920270e+00 -6.39739811e-01 3.60449344e-01
8.78913522e-01 -2.24046767e-01 -9.09485221e-01 2.24281475e-01] | [6.317785739898682, 4.313699245452881] |
b02550e0-e2dd-406f-b92c-7b6269350838 | moving-object-detection-in-video-using | 1509.09089 | null | http://arxiv.org/abs/1509.09089v1 | http://arxiv.org/pdf/1509.09089v1.pdf | Moving Object Detection in Video Using Saliency Map and Subspace Learning | Moving object detection is a key to intelligent video analysis. On the one
hand, what moves is not only interesting objects but also noise and cluttered
background. On the other hand, moving objects without rich texture are prone
not to be detected. So there are undesirable false alarms and missed alarms in
many algorithms of moving object detection. To reduce the false alarms and
missed alarms, in this paper, we propose to incorporate a saliency map into an
incremental subspace analysis framework where the saliency map makes estimated
background has less chance than foreground (i.e., moving objects) to contain
salient objects. The proposed objective function systematically takes account
into the properties of sparsity, low-rank, connectivity, and saliency. An
alternative minimization algorithm is proposed to seek the optimal solutions.
Experimental results on the Perception Test Images Sequences demonstrate that
the proposed method is effective in reducing false alarms and missed alarms. | ['Xuelong. Li', 'Li Ye', 'Jing Pan', 'Yanwei Pang'] | 2015-09-30 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 4.56892669e-01 -2.95687884e-01 -1.64915714e-02 2.93421466e-03
-2.66490966e-01 -1.48064807e-01 1.21530317e-01 -6.20328309e-03
-1.02130413e-01 7.64803946e-01 1.82634592e-01 2.12394431e-01
-1.40103638e-01 -5.65862894e-01 -3.92031312e-01 -8.52099001e-01
-1.74995810e-01 -2.03908443e-01 8.75135958e-01 1.22318335e-01
4.62900251e-01 1.97988316e-01 -1.61825061e+00 2.04358518e-01
1.08984840e+00 8.56659293e-01 6.89559996e-01 4.06409711e-01
-1.34297892e-01 9.51637983e-01 -5.30149877e-01 2.04050452e-01
2.12759346e-01 -4.48143363e-01 -3.04682821e-01 5.36350250e-01
2.87040740e-01 -3.54847044e-01 2.08452512e-02 1.64345849e+00
2.44342715e-01 3.78220856e-01 2.56889850e-01 -1.18427551e+00
-3.68959934e-01 3.76673669e-01 -1.11281478e+00 9.85156953e-01
1.44156799e-01 1.49314860e-02 4.51972276e-01 -1.26649940e+00
2.76078671e-01 1.27530301e+00 1.40890107e-01 2.55315602e-01
-7.86515832e-01 -3.55512828e-01 6.09180033e-01 4.91097271e-01
-1.41177142e+00 -4.98121858e-01 9.04728830e-01 -3.49281460e-01
7.80315697e-02 6.70569003e-01 5.79396188e-01 5.49044609e-01
1.67257681e-01 9.88780439e-01 7.38739550e-01 -3.20345640e-01
2.80685723e-01 3.72056007e-01 3.09216052e-01 5.24633467e-01
8.40647280e-01 -2.23900124e-01 -3.72553945e-01 1.31933279e-02
6.77421868e-01 1.73726350e-01 -6.11987472e-01 -4.41677183e-01
-1.43191051e+00 4.54882592e-01 1.13203809e-01 2.99144328e-01
-3.88999343e-01 -2.14819372e-01 9.39345285e-02 -2.71075338e-01
3.19676012e-01 5.11724576e-02 -1.41564652e-01 -2.24343520e-02
-8.62627208e-01 2.36825366e-02 2.69660294e-01 9.08643901e-01
6.23629987e-01 4.74438399e-01 -1.82223707e-01 6.62073433e-01
1.74115717e-01 5.75983465e-01 4.51806039e-01 -6.69800341e-01
5.27907312e-01 6.79231226e-01 3.70035350e-01 -1.77847874e+00
-2.12300584e-01 -7.02759147e-01 -8.35042059e-01 1.03271715e-01
1.86147392e-01 -4.50144075e-02 -6.99841917e-01 1.47270215e+00
5.59474468e-01 5.88816106e-01 -1.19386986e-01 1.28229868e+00
8.13954175e-01 7.65606165e-01 3.07630189e-02 -8.05204928e-01
1.06647635e+00 -8.13226342e-01 -1.09722185e+00 -2.61782438e-01
1.51334912e-01 -9.03996706e-01 8.55825782e-01 2.72598416e-01
-1.01279271e+00 -6.55170023e-01 -9.36789274e-01 5.93436658e-01
1.02210999e-01 3.22955102e-01 4.48186666e-01 3.85665983e-01
-5.54366648e-01 1.10136718e-01 -6.60767019e-01 -3.14775795e-01
2.32511699e-01 1.69391856e-01 7.87493885e-02 1.07582621e-01
-9.83181894e-01 7.32066214e-01 7.69513369e-01 3.82722646e-01
-8.66713762e-01 -1.78661034e-01 -6.05007708e-01 8.48039612e-02
8.60678613e-01 -2.44346082e-01 7.25352049e-01 -1.49276555e+00
-8.07528257e-01 1.45391449e-01 -4.65441257e-01 -2.48286203e-01
4.60927844e-01 -3.50579143e-01 -6.93078220e-01 3.27360451e-01
2.39812419e-01 2.03155175e-01 1.21496344e+00 -1.14239657e+00
-9.81643617e-01 -1.77735388e-01 -3.14688295e-01 4.59354758e-01
-3.78840715e-01 5.16481958e-02 -3.67461473e-01 -7.96542406e-01
6.25295401e-01 -6.26131475e-01 -2.92671472e-01 -2.80623704e-01
-2.96779722e-01 1.12005860e-01 1.23705685e+00 -6.08541787e-01
1.48584509e+00 -2.22608852e+00 5.14014484e-03 2.77514309e-01
1.89216435e-01 2.51321465e-01 -8.06535408e-03 -3.37100834e-01
-7.12681711e-02 -1.43710792e-01 -1.48004070e-01 5.97158253e-01
-6.47504151e-01 -2.11502597e-01 -2.76757509e-01 4.26348418e-01
3.66537184e-01 3.60991448e-01 -9.66791928e-01 -7.09983110e-01
3.89298320e-01 1.48701563e-01 -3.03914338e-01 -1.94540069e-01
1.48211226e-01 4.45831567e-01 -8.61058831e-01 8.98866713e-01
9.44349289e-01 -3.00622970e-01 -1.98282853e-01 -3.41529071e-01
-4.29034114e-01 -1.10551856e-01 -1.77276003e+00 8.69989038e-01
3.76878828e-01 5.79732776e-01 2.02704147e-01 -1.02785897e+00
6.91281140e-01 7.78701827e-02 4.54932570e-01 -4.05122161e-01
9.74467993e-02 1.33258805e-01 1.19854815e-01 -6.02345407e-01
5.79968095e-01 1.24221392e-01 3.86338145e-01 -2.12097034e-01
-4.95472133e-01 5.42750776e-01 2.13254303e-01 3.02267104e-01
7.45967448e-01 -2.20951125e-01 4.25456792e-01 -5.94337642e-01
8.70536804e-01 7.20148683e-02 1.16465294e+00 6.02839470e-01
-3.70459914e-01 3.52842867e-01 -3.10683120e-02 -4.25955385e-01
-5.40164888e-01 -1.06442273e+00 9.75920260e-02 7.77070403e-01
9.23146069e-01 -4.90565188e-02 -7.45657682e-01 -3.27669740e-01
-3.14517260e-01 5.63669801e-01 -3.28765124e-01 -2.41457075e-01
-5.91565728e-01 -8.46805930e-01 -1.95395842e-01 1.93053946e-01
5.98447561e-01 -1.03776288e+00 -8.80315065e-01 2.75972664e-01
-4.74816620e-01 -8.90511930e-01 -5.36141038e-01 -3.98226291e-01
-1.00586593e+00 -1.09583938e+00 -8.43863428e-01 -8.54954720e-01
9.63114917e-01 1.14231575e+00 6.40470445e-01 2.11609468e-01
-2.93993026e-01 1.91472381e-01 -4.88824069e-01 -5.45350492e-01
1.05935335e-02 -4.25384611e-01 4.05868262e-01 5.05997002e-01
2.64141768e-01 -2.04981938e-02 -5.79473257e-01 5.82365394e-01
-8.21614206e-01 8.05518925e-02 6.01575613e-01 5.91171503e-01
6.29537344e-01 7.76188612e-01 5.91946185e-01 -2.71359086e-01
2.79185981e-01 -4.76584256e-01 -7.22942591e-01 2.01249167e-01
-1.08381756e-01 -4.08013850e-01 3.84981066e-01 -6.83658361e-01
-1.16639352e+00 6.47929460e-02 6.26612127e-01 -5.06078720e-01
-8.18265527e-02 3.49587411e-01 -5.03350675e-01 -1.80235967e-01
2.64981061e-01 5.88515997e-01 -3.93502504e-01 -3.07410687e-01
-8.74466822e-02 3.67194772e-01 3.77364427e-01 -3.10229547e-02
9.20127273e-01 6.57653451e-01 3.43309194e-02 -1.28015590e+00
-7.13448405e-01 -6.13199174e-01 -3.37179929e-01 -6.21411145e-01
7.72078753e-01 -9.34422314e-01 -3.31160843e-01 3.45766455e-01
-9.46026564e-01 5.14818430e-01 6.45797700e-02 7.54518747e-01
-9.15996805e-02 6.46974206e-01 -1.74095809e-01 -1.28940380e+00
-1.10748358e-01 -9.84573781e-01 5.48214197e-01 6.95200026e-01
-2.02663075e-02 -4.42808539e-01 -5.09845793e-01 1.06730312e-01
2.92800397e-01 2.32918516e-01 4.92153853e-01 -4.35734928e-01
-1.17266178e+00 -9.23412517e-02 -2.57638097e-01 7.49058798e-02
5.50399184e-01 1.17759496e-01 -5.80597222e-01 -2.43014291e-01
5.40082037e-01 4.09546942e-01 7.57005513e-01 7.56646395e-01
8.79862189e-01 -5.59714913e-01 -5.57796836e-01 2.70551592e-01
1.28363049e+00 7.34481096e-01 5.62616765e-01 3.81890029e-01
8.61131489e-01 6.82496250e-01 1.30117381e+00 4.96700376e-01
-1.14296958e-01 4.83862698e-01 4.19713855e-01 -1.35367170e-01
2.10908845e-01 7.74124861e-02 3.58909726e-01 9.94623244e-01
-5.37772290e-02 -3.43051329e-02 -7.99250603e-01 7.02316523e-01
-1.96059561e+00 -1.27744830e+00 -4.91607845e-01 2.36543822e+00
3.96587193e-01 3.62505406e-01 -7.65968859e-02 2.15697229e-01
1.31601155e+00 -4.84272139e-03 -5.81116438e-01 3.28937858e-01
-4.78595734e-01 -4.62522775e-01 3.86662036e-01 2.68840075e-01
-1.32204521e+00 7.32163668e-01 5.56276894e+00 7.73026824e-01
-9.65973437e-01 -5.85982166e-02 7.00345635e-01 -2.37894312e-01
-8.98209885e-02 1.67413637e-01 -8.42004418e-01 8.38538349e-01
9.91782248e-02 -3.42235684e-01 3.29183102e-01 9.23923135e-01
6.37292087e-01 -6.19104087e-01 -5.26584089e-01 1.01130557e+00
1.95147187e-01 -9.68610048e-01 3.08119595e-01 -2.95099288e-01
6.58346236e-01 -5.87935984e-01 7.67220117e-05 -9.87841412e-02
-1.63628414e-01 -5.06650150e-01 9.62700903e-01 6.42660975e-01
1.47232547e-01 -7.87620723e-01 6.28571689e-01 5.70674837e-01
-1.39403343e+00 -1.57358393e-01 -4.71105874e-01 -1.77817628e-01
1.78807840e-01 8.57036948e-01 -5.00660479e-01 2.93834150e-01
7.50422716e-01 8.40174854e-01 -5.57152927e-01 1.34805012e+00
1.41942918e-01 4.45458859e-01 -1.84522018e-01 -1.19537465e-01
4.69797030e-02 -4.42440629e-01 1.24582517e+00 8.87924790e-01
3.97528201e-01 5.00619590e-01 6.01643562e-01 7.26638019e-01
5.30767798e-01 3.34772080e-01 -5.25949538e-01 1.72431916e-01
4.18803126e-01 1.12719834e+00 -1.18136144e+00 -5.05741060e-01
-4.88758385e-01 8.20876837e-01 -2.84662396e-01 5.01723826e-01
-8.91700923e-01 -3.88801605e-01 3.31410855e-01 2.03793541e-01
3.26801658e-01 -1.60352483e-01 -2.05044836e-01 -1.29498088e+00
1.90043643e-01 -7.87849844e-01 3.46221149e-01 -6.65015221e-01
-8.50985706e-01 1.82887807e-01 -1.71749875e-01 -1.61992371e+00
4.31991875e-01 -1.21056437e-01 -7.51939178e-01 6.55578554e-01
-1.10193896e+00 -4.23829108e-01 -4.44984347e-01 5.29463530e-01
9.99254405e-01 -1.05256669e-01 4.02874239e-02 2.68934101e-01
-8.36365521e-01 4.82217930e-02 7.85197318e-02 -3.93834984e-04
3.19566876e-01 -6.53738379e-01 -2.68699914e-01 1.47799540e+00
-1.11464143e-01 6.10261738e-01 9.81909990e-01 -1.04043365e+00
-1.20353055e+00 -1.12105286e+00 3.36035460e-01 1.28920923e-03
4.90664870e-01 -4.85493317e-02 -1.07809591e+00 2.55070031e-01
-2.03846753e-01 -2.60639787e-01 2.83787996e-01 -3.32516879e-01
3.98921967e-01 -2.11186796e-01 -9.38426256e-01 7.70893812e-01
8.18430305e-01 1.73319668e-01 -6.50340021e-01 1.95028052e-01
5.84539652e-01 -1.73663720e-01 -2.22806912e-03 6.82390153e-01
2.98738599e-01 -8.78797889e-01 9.72367644e-01 -2.89582610e-01
-1.37363579e-02 -1.01188898e+00 -1.16043203e-01 -8.79616976e-01
-7.03738034e-01 -4.50078696e-01 -3.00050467e-01 1.14416456e+00
-1.22348289e-03 -3.00821573e-01 6.44545972e-01 4.65968013e-01
7.10102022e-02 -3.55589360e-01 -8.42100859e-01 -7.83370435e-01
-9.29400086e-01 3.04817092e-02 7.95011893e-02 1.07597613e+00
-7.86249116e-02 2.85467386e-01 -6.22334659e-01 4.88819510e-01
9.18842137e-01 2.15270817e-01 5.66806555e-01 -1.27152205e+00
-4.20528390e-02 -2.43790150e-01 -5.72928488e-01 -8.53949308e-01
-3.57090622e-01 -2.32921839e-01 2.28464901e-01 -1.31237710e+00
5.49322546e-01 -2.20664352e-01 -6.25429928e-01 -1.84241030e-02
-6.93614900e-01 -2.59296354e-02 5.95311038e-02 4.25294429e-01
-9.48239565e-01 5.45529544e-01 1.14753616e+00 -2.14493886e-01
-4.35671538e-01 1.23139732e-01 -5.37007153e-01 1.12856948e+00
8.30901384e-01 -3.34965020e-01 -4.37196285e-01 -2.97390580e-01
-1.13140278e-01 3.01028471e-02 4.09196556e-01 -1.14695251e+00
1.67134047e-01 -5.80460012e-01 6.25875056e-01 -9.10204947e-01
7.32149705e-02 -9.10377920e-01 1.63422048e-01 5.33324480e-01
-1.08898304e-01 -1.03700548e-01 2.49523893e-01 8.96487415e-01
-2.14029372e-01 -1.47953853e-01 8.88976872e-01 -2.13924110e-01
-1.13814318e+00 -3.67779285e-02 -6.25563622e-01 -1.01850390e-01
1.27997255e+00 -6.26113296e-01 -1.80151105e-01 -4.77630258e-01
-6.44394457e-01 3.42318237e-01 2.68330842e-01 5.59753418e-01
9.20411885e-01 -1.29742050e+00 -5.98830044e-01 9.10581648e-02
-1.34778172e-01 -2.50810146e-01 4.11193162e-01 9.98355925e-01
-2.50306815e-01 2.76692390e-01 -1.68891236e-01 -6.31564558e-01
-1.55406749e+00 7.37231970e-01 -1.07957581e-02 2.93026626e-01
-4.05520082e-01 7.96056807e-01 6.24932706e-01 5.42453229e-01
1.61811218e-01 -2.77043074e-01 -5.34743905e-01 1.11505957e-02
8.89435530e-01 8.04485857e-01 -4.35310543e-01 -9.23812866e-01
-5.47327876e-01 5.91149747e-01 -8.47488940e-02 2.66499877e-01
8.40732396e-01 -4.77062404e-01 -2.32336298e-01 5.93322217e-01
5.81769586e-01 2.53223479e-01 -1.05622280e+00 -1.29548937e-01
3.06229115e-01 -9.21644568e-01 7.87779018e-02 -4.16242123e-01
-9.91742074e-01 5.35378933e-01 7.10853994e-01 3.24885845e-01
1.29488075e+00 -2.99668431e-01 4.56651330e-01 2.00865984e-01
6.40484273e-01 -1.18648958e+00 1.88793242e-01 2.86315292e-01
7.06826210e-01 -1.17347872e+00 1.74969897e-01 -8.81237090e-01
-7.14289606e-01 7.46911287e-01 8.92858863e-01 -3.44199538e-02
4.51226205e-01 -7.82862380e-02 -9.01216790e-02 5.67535497e-02
-4.86152619e-01 -2.85564095e-01 3.83095741e-01 3.43486428e-01
2.05367003e-02 -2.13617198e-02 -4.90669161e-01 5.80064595e-01
4.11635578e-01 -1.97519630e-01 6.82449341e-01 9.58570540e-01
-1.11967468e+00 -2.69392967e-01 -9.09502983e-01 4.88873392e-01
-5.90540230e-01 -9.87213925e-02 -1.67763859e-01 3.61807466e-01
1.60387531e-01 1.26165581e+00 3.32517363e-02 -3.06758791e-01
1.43982261e-01 -3.34372669e-01 3.21616381e-02 -5.29600143e-01
1.24861740e-01 4.78393734e-01 -2.05200434e-01 -5.00978351e-01
-5.12295306e-01 -5.93777239e-01 -1.20932341e+00 3.13902736e-01
-8.84916306e-01 2.42246866e-01 1.77358061e-01 6.62303329e-01
1.14627644e-01 5.67265093e-01 6.03910387e-01 -5.78776896e-01
-1.50657639e-01 -7.85965025e-01 -7.28284478e-01 2.99640626e-01
3.02241296e-01 -8.48091722e-01 -4.68397439e-01 2.23800093e-01] | [9.207443237304688, -0.6316173672676086] |
975e4fac-0c52-4103-8ba0-141af2e3b280 | pore-provably-robust-recommender-systems | 2303.14601 | null | https://arxiv.org/abs/2303.14601v1 | https://arxiv.org/pdf/2303.14601v1.pdf | PORE: Provably Robust Recommender Systems against Data Poisoning Attacks | Data poisoning attacks spoof a recommender system to make arbitrary, attacker-desired recommendations via injecting fake users with carefully crafted rating scores into the recommender system. We envision a cat-and-mouse game for such data poisoning attacks and their defenses, i.e., new defenses are designed to defend against existing attacks and new attacks are designed to break them. To prevent such a cat-and-mouse game, we propose PORE, the first framework to build provably robust recommender systems in this work. PORE can transform any existing recommender system to be provably robust against any untargeted data poisoning attacks, which aim to reduce the overall performance of a recommender system. Suppose PORE recommends top-$N$ items to a user when there is no attack. We prove that PORE still recommends at least $r$ of the $N$ items to the user under any data poisoning attack, where $r$ is a function of the number of fake users in the attack. Moreover, we design an efficient algorithm to compute $r$ for each user. We empirically evaluate PORE on popular benchmark datasets. | ['Neil Zhenqiang Gong', 'Yuepeng Hu', 'Yupei Liu', 'Jinyuan Jia'] | 2023-03-26 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [-2.30068058e-01 -1.41567066e-01 -1.25225559e-01 3.93480062e-02
-5.02613485e-01 -1.41667557e+00 2.15408191e-01 -1.71120778e-01
-3.59834254e-01 2.46733040e-01 -9.03524160e-02 -7.60323763e-01
-2.19880804e-01 -1.36021090e+00 -7.96311021e-01 -7.36186564e-01
-1.33214250e-01 4.21709180e-01 4.25711691e-01 -6.26581430e-01
4.73090768e-01 5.94661772e-01 -9.59715724e-01 4.38247621e-01
4.21643287e-01 5.23154676e-01 -5.98967254e-01 7.21545875e-01
4.73010957e-01 6.53929830e-01 -9.64878023e-01 -9.93734777e-01
8.61405849e-01 -4.01387870e-01 -5.95295131e-01 -2.54107594e-01
2.57070094e-01 -6.75432265e-01 -6.99032664e-01 1.31598675e+00
3.17319632e-01 -1.78688928e-01 3.02991569e-01 -1.57698619e+00
-9.67512965e-01 1.21972847e+00 -3.91990960e-01 8.32588673e-02
3.59160364e-01 2.59829283e-01 9.56453681e-01 -4.70119327e-01
2.69126803e-01 1.14628899e+00 3.29522580e-01 9.34216142e-01
-9.81105387e-01 -1.11788666e+00 9.43416283e-02 -2.19298676e-01
-1.30200481e+00 -2.50994772e-01 4.45598453e-01 -1.49928257e-01
1.91535652e-01 8.99686456e-01 3.07813674e-01 1.25591338e+00
-1.03113027e-02 2.92539716e-01 8.58049810e-01 1.32953614e-01
3.48628759e-01 4.08207297e-01 2.45249107e-01 6.04291201e-01
8.96283329e-01 3.62208754e-01 -2.45109782e-01 -1.08077633e+00
6.25678778e-01 4.40743715e-01 -2.05952749e-01 -3.48770440e-01
-7.03150272e-01 1.33595145e+00 6.22313261e-01 2.37165261e-02
-3.01886331e-02 2.15574101e-01 2.54664361e-01 6.94331050e-01
-1.66158732e-02 9.59939897e-01 -6.21853590e-01 4.47094202e-01
-2.79103607e-01 5.33960640e-01 8.69950950e-01 5.04790425e-01
1.38674438e-01 4.31644544e-03 2.04163134e-01 2.61234641e-01
3.25565755e-01 1.01592946e+00 3.59900862e-01 -8.34605753e-01
4.34406757e-01 4.04769123e-01 6.58156693e-01 -1.38672817e+00
4.78023402e-02 -1.92761466e-01 -5.33066213e-01 1.30607367e-01
5.84690869e-01 -2.47958556e-01 -3.24665904e-01 1.78997934e+00
5.33535898e-01 1.77135453e-01 1.80977173e-02 8.96152139e-01
2.68237352e-01 3.72302204e-01 -3.94086301e-01 -1.60809368e-01
1.34507239e+00 -6.74778461e-01 -1.83499947e-01 1.94761425e-01
9.04577851e-01 -3.55974704e-01 1.33058000e+00 8.18786204e-01
-9.74032760e-01 4.29897476e-03 -1.09550965e+00 4.63778794e-01
-6.11019850e-01 -3.72515231e-01 3.33149374e-01 1.47483277e+00
-4.58666414e-01 5.09211302e-01 -5.26127040e-01 7.57012367e-02
5.02069354e-01 5.68935633e-01 -2.24282309e-01 -1.65047422e-02
-1.39242971e+00 4.17486280e-01 -2.03276291e-01 -4.74458277e-01
-1.28536367e+00 -4.63444144e-01 -2.42961437e-01 5.25494926e-02
7.00001717e-01 -5.66456735e-01 9.47961032e-01 -1.92887262e-01
-1.34694672e+00 4.16236371e-01 6.53421402e-01 -8.20729434e-01
4.35931444e-01 -4.98656295e-02 -7.00085700e-01 1.41657069e-01
4.31536213e-02 -2.18793690e-01 1.10036063e+00 -1.21071029e+00
-5.92411935e-01 -5.25924265e-01 7.29636252e-01 -1.58100277e-01
-7.85273492e-01 4.22613412e-01 2.46329948e-01 -6.86485648e-01
-1.65278971e-01 -1.20020294e+00 -4.70602036e-01 -1.46993458e-01
-6.44240856e-01 1.01373270e-01 1.05795228e+00 7.44767115e-03
1.32076943e+00 -1.95341301e+00 -3.44151050e-01 5.42756557e-01
5.77554286e-01 7.72223413e-01 -2.01321334e-01 4.70822781e-01
2.00614497e-01 6.24777794e-01 1.38363123e-01 8.43760967e-02
1.49642467e-01 7.02991411e-02 -1.04395461e+00 7.88591802e-01
-6.78299069e-01 5.70470035e-01 -8.08404624e-01 4.17930812e-01
-5.34764566e-02 1.30919382e-01 -1.12843060e+00 3.30024511e-01
-1.37310565e-01 6.53475448e-02 -7.04087198e-01 3.64403337e-01
1.03855872e+00 -1.43649399e-01 2.56451786e-01 1.30031407e-01
1.91580355e-01 5.11400998e-01 -1.02113926e+00 8.48304927e-01
-6.36307523e-02 -4.32593942e-01 1.23356871e-01 -6.44690037e-01
7.34388292e-01 9.91494134e-02 3.62641871e-01 -3.01080048e-01
4.80010450e-01 2.63086796e-01 1.46606341e-01 -1.24804392e-01
3.71694535e-01 7.48290792e-02 -5.58749080e-01 1.16801798e+00
-6.09753191e-01 4.05614644e-01 -2.33646274e-01 6.93589807e-01
1.63651252e+00 -7.55383492e-01 -1.71577483e-01 -1.65599316e-01
7.19068646e-01 -1.05371810e-01 2.65719682e-01 1.39450705e+00
-1.83704898e-01 5.76971434e-02 4.49902236e-01 -5.31623006e-01
-8.62982035e-01 -1.01951504e+00 3.67638379e-01 1.41558063e+00
4.65996832e-01 -7.43114293e-01 -8.90821040e-01 -1.68282211e+00
2.23118171e-01 5.45390368e-01 -6.84044898e-01 -6.63787127e-01
-3.55260313e-01 -7.39286244e-01 1.13868880e+00 -7.87963197e-02
4.28864211e-01 -7.98115849e-01 -3.64665300e-01 1.30263746e-01
-6.61883503e-02 -8.47026408e-01 -8.24133515e-01 -2.03395918e-01
-5.85154653e-01 -1.56893098e+00 5.22331446e-02 -2.43463859e-01
8.56743217e-01 9.73449826e-01 5.56773603e-01 6.43340945e-01
2.06263408e-01 -7.35744163e-02 -4.94590193e-01 6.56596422e-02
-7.46435881e-01 -6.10042810e-02 6.39584839e-01 -3.16054150e-02
4.18314099e-01 -4.14290845e-01 -7.82556593e-01 1.02720869e+00
-1.07916117e+00 -7.77066410e-01 1.72909841e-01 4.57612365e-01
1.54577971e-01 3.81252617e-01 5.11506081e-01 -1.40053070e+00
9.44279552e-01 -6.56270921e-01 -7.07184196e-01 1.29486978e-01
-5.33340394e-01 -5.15048504e-02 1.40012860e+00 -9.28873897e-01
-1.30746543e-01 -2.87691087e-01 -3.80788594e-01 -4.49758440e-01
-7.50297904e-02 -2.51992829e-02 -5.08083344e-01 -4.20424193e-01
1.25870442e+00 6.12418279e-02 -2.65362382e-01 -6.57935798e-01
9.92509305e-01 7.93640792e-01 5.24549127e-01 -6.62770927e-01
1.49159753e+00 8.37845862e-01 -1.99684978e-01 -1.96160436e-01
-8.99381340e-01 -1.33731291e-01 7.53428340e-02 2.74094492e-01
1.56504795e-01 -3.74330014e-01 -1.31879056e+00 4.08722132e-01
-7.79912710e-01 -1.42719164e-01 -1.04773402e-01 4.96464521e-02
-1.93715110e-01 5.95376909e-01 -7.26356149e-01 -5.84429860e-01
-5.23695946e-01 -1.19163048e+00 1.88377634e-01 -2.23467946e-01
2.55774766e-01 -3.82289827e-01 2.10898742e-01 6.21687949e-01
4.44498867e-01 1.12042029e-03 5.21724641e-01 -1.31737936e+00
-2.47344360e-01 -6.39530659e-01 -7.25828409e-02 2.34223589e-01
8.78201202e-02 -1.33882344e-01 -4.99631226e-01 -7.40771472e-01
3.26481730e-01 -2.35292584e-01 2.72675276e-01 -3.51227373e-01
1.30063665e+00 -1.36309385e+00 -2.10405841e-01 4.57180053e-01
1.11841464e+00 2.89889336e-01 6.42543495e-01 2.14337662e-01
7.51906991e-01 3.29309069e-02 4.93938506e-01 4.95267421e-01
5.93973398e-02 7.60376692e-01 7.08043277e-01 3.81178230e-01
4.58590716e-01 -5.22131741e-01 7.02525795e-01 2.95294285e-01
3.05681735e-01 -4.24293458e-01 -1.79903045e-01 -4.62393463e-02
-1.59388328e+00 -9.28131759e-01 -2.65274197e-01 2.48904252e+00
5.68115532e-01 4.74732190e-01 7.10940957e-01 2.52699256e-01
6.29951537e-01 -1.59915954e-01 -6.21882141e-01 -6.44464552e-01
2.44135946e-01 3.64209026e-01 1.02590692e+00 3.61280262e-01
-9.89457309e-01 1.12577569e+00 5.87726307e+00 7.81672716e-01
-8.30525696e-01 4.49299335e-01 1.84054568e-01 -3.03506464e-01
-3.32157820e-01 8.86420831e-02 -6.89815998e-01 7.42752433e-01
1.25732636e+00 -1.28884584e-01 9.89658415e-01 1.07289565e+00
1.76780909e-01 5.79316318e-01 -6.97670579e-01 4.22595620e-01
5.31424629e-03 -1.30313551e+00 1.86893553e-01 5.74531436e-01
3.86690587e-01 -1.77455544e-01 3.63842875e-01 2.76971459e-01
1.24906862e+00 -6.95486724e-01 4.79055911e-01 -1.87641501e-01
4.19233203e-01 -1.13402700e+00 3.00985873e-01 6.80438876e-01
-6.73412681e-01 -5.73451698e-01 -8.59445393e-01 5.27478196e-02
-2.11005554e-01 2.46661887e-01 -4.41370487e-01 2.17286035e-01
6.49674833e-01 4.60123383e-02 -5.69584370e-01 7.07574248e-01
-4.64099169e-01 7.55145669e-01 -5.27090073e-01 3.06506008e-02
1.92857146e-01 -9.66465399e-02 4.77814823e-01 5.30398250e-01
2.82494575e-01 4.23565060e-01 2.95041502e-01 6.37290001e-01
-6.08484209e-01 4.96503189e-02 -8.50694358e-01 1.87132992e-02
9.36502218e-01 1.14475465e+00 -4.99515355e-01 -3.20555300e-01
8.95286426e-02 8.60695302e-01 2.96292484e-01 -2.26638362e-01
-1.12375867e+00 -3.98048848e-01 1.04339266e+00 4.18180555e-01
3.08153421e-01 -3.61441635e-02 2.63570119e-02 -1.37869370e+00
-5.40463328e-01 -1.50569201e+00 8.17905426e-01 -5.13769090e-01
-1.59439611e+00 4.62273747e-01 -5.01125574e-01 -1.09060240e+00
3.21458727e-01 -3.02625269e-01 -4.92075056e-01 3.03033710e-01
-8.12540472e-01 -8.44302833e-01 2.99910396e-01 1.27135241e+00
-5.33830486e-02 -2.96253443e-01 9.07315850e-01 1.65424883e-01
-6.52320802e-01 1.22627079e+00 8.34900290e-02 3.54869485e-01
7.12497532e-01 -8.41033638e-01 6.04645312e-01 1.23298395e+00
5.73572516e-01 1.16021895e+00 7.33571470e-01 -7.20921457e-01
-1.58658004e+00 -1.08489895e+00 2.49322489e-01 -9.09470856e-01
9.24677789e-01 -5.70901334e-01 -7.52196848e-01 5.32033682e-01
-3.63133162e-01 1.70436829e-01 1.00862682e+00 -2.01470718e-01
-1.27520895e+00 1.91984084e-02 -1.67160904e+00 7.43337214e-01
1.01530695e+00 -5.51373363e-01 -4.01998758e-01 4.95416403e-01
9.40643430e-01 -2.16046199e-01 -6.40288949e-01 7.68619478e-02
6.47321165e-01 -9.62940991e-01 1.22610223e+00 -1.30442429e+00
-5.77123314e-02 -5.60083389e-01 -2.05350101e-01 -9.68762457e-01
-4.33772206e-01 -8.64108682e-01 -5.08208871e-01 7.23653913e-01
4.12169725e-01 -7.60809243e-01 9.55095828e-01 3.63901824e-01
4.59470242e-01 -2.75629669e-01 -6.32682860e-01 -8.87906134e-01
3.93855184e-01 -1.53328508e-01 9.95275080e-01 1.09057617e+00
1.81988478e-01 3.70006301e-02 -1.01625669e+00 8.77653480e-01
8.32751036e-01 2.51832139e-02 1.20432687e+00 -1.07108247e+00
-7.09127665e-01 -8.65749866e-02 3.16772871e-02 -1.06828344e+00
-2.62507468e-01 -6.46339595e-01 -4.84556556e-01 -6.83710396e-01
7.26561323e-02 -6.88958883e-01 -4.96452600e-01 6.21096194e-01
8.10410753e-02 7.42316723e-01 2.43866861e-01 4.86096591e-01
-7.38591015e-01 -2.48171785e-03 9.83411968e-01 8.06879904e-03
-1.73381284e-01 4.74100143e-01 -1.53080034e+00 6.24292612e-01
9.18919683e-01 -1.28251290e+00 -5.11714876e-01 5.74053042e-02
4.59050119e-01 7.86360949e-02 1.97375819e-01 -5.80117404e-01
8.03957433e-02 -2.56993979e-01 -2.36887202e-01 -2.22018808e-01
-1.42169163e-01 -8.86799335e-01 1.04707055e-01 8.87169659e-01
-5.71740985e-01 5.48389666e-02 -5.49800396e-01 7.75466561e-01
7.01579750e-01 -4.19952683e-02 1.11577797e+00 -9.83242691e-02
3.85089636e-01 6.04384661e-01 -4.71684992e-01 -8.19837376e-02
1.02972698e+00 2.80670315e-01 -1.10455537e+00 -3.48285228e-01
-7.75041699e-01 -5.96008217e-03 7.24651873e-01 4.04312551e-01
6.17288530e-01 -1.18120873e+00 -4.60635900e-01 3.14326495e-01
1.17548347e-01 -7.09436893e-01 -4.64248843e-02 4.37158227e-01
-3.22682649e-01 -8.51266310e-02 -1.91542506e-01 2.81401873e-01
-1.50618649e+00 1.40725815e+00 4.34682965e-01 -3.77759933e-01
-4.97759104e-01 7.00798869e-01 -5.73763698e-02 -5.11830807e-01
2.56084412e-01 6.97834566e-02 7.91061483e-03 -5.94193637e-01
9.15127337e-01 3.43742609e-01 -9.16688740e-02 -5.45913637e-01
-2.80896127e-01 -1.42501229e-02 -4.90166277e-01 -2.60400455e-02
8.72833788e-01 -1.28050417e-01 4.70384099e-02 -5.93467951e-01
8.49467397e-01 8.23176205e-01 -6.59778893e-01 -1.65232658e-01
-4.51630056e-01 -9.80287254e-01 -2.82529175e-01 -9.67477322e-01
-1.36898291e+00 4.56691176e-01 3.05407017e-01 1.05093396e+00
7.08789170e-01 -1.99738696e-01 1.13051724e+00 6.26035333e-01
7.72902012e-01 -4.53116745e-01 3.47630233e-01 3.15573886e-02
2.38015264e-01 -4.96556699e-01 -1.94006674e-02 -5.43508708e-01
-6.16577268e-01 6.51071846e-01 5.73588967e-01 -6.31697536e-01
8.09637308e-01 1.09648466e-01 -1.05636179e-01 -3.10959071e-01
-5.96834838e-01 2.96366572e-01 -3.64543080e-01 5.60376346e-01
-4.57341611e-01 1.95523024e-01 -4.80569452e-01 1.18925202e+00
-3.80851537e-01 -2.54625022e-01 1.02184653e+00 7.23543942e-01
-8.14870298e-01 -1.66612959e+00 -5.30500531e-01 3.10825706e-01
-7.97929823e-01 6.01057746e-02 -6.79008663e-01 2.97626585e-01
1.69320911e-01 1.44391465e+00 -6.45043314e-01 -1.05641270e+00
4.20085490e-01 -4.36031640e-01 1.84333771e-02 -6.06323421e-01
-1.10403347e+00 -1.18264116e-01 -1.05276331e-01 -7.14901984e-01
1.92726135e-01 -9.07622799e-02 -9.27345395e-01 -1.10722101e+00
-6.65133715e-01 7.75595188e-01 3.64696026e-01 5.80936372e-01
4.50320333e-01 -2.34077439e-01 1.64935231e+00 1.02079347e-01
-1.07917988e+00 -3.25819254e-01 -9.41109836e-01 5.78870177e-01
-7.36140981e-02 -6.09773993e-01 -9.10733640e-01 -5.16493917e-01] | [5.81056547164917, 7.342576026916504] |
43e879ec-1366-4f94-a9d2-eb1cb2afefe3 | replica-enhanced-feature-pyramid-network-by | 2111.11546 | null | https://arxiv.org/abs/2111.11546v3 | https://arxiv.org/pdf/2111.11546v3.pdf | Lightweight Transformer Backbone for Medical Object Detection | Lesion detection in digital breast tomosynthesis (DBT) is an important and a challenging problem characterized by a low prevalence of images containing tumors. Due to the label scarcity problem, large deep learning models and computationally intensive algorithms are likely to fail when applied to this task. In this paper, we present a practical yet lightweight backbone to improve the accuracy of tumor detection. Specifically, we propose a novel modification of visual transformer (ViT) on image feature patches to connect the feature patches of a tumor with healthy backgrounds of breast images and form a more robust backbone for tumor detection. To the best of our knowledge, our model is the first work of Transformer backbone object detection for medical imaging. Our experiments show that this model can considerably improve the accuracy of lesion detection and reduce the amount of labeled data required in typical ViT. We further show that with additional augmented tumor data, our model significantly outperforms the Faster R-CNN model and state-of-the-art SWIN transformer model. | ['Nicholas Konz', 'Maciej A. Mazurowski', 'Hanxue Gu', 'Haoyu Dong', 'Yifan Zhang'] | 2021-11-22 | null | null | null | null | ['medical-object-detection'] | ['computer-vision'] | [ 4.08387870e-01 4.52205837e-01 -4.22495395e-01 1.12848356e-01
-1.10393083e+00 -1.26208544e-01 2.06190005e-01 1.82990715e-01
-1.14138220e-02 2.41080537e-01 -7.28895664e-02 -5.79009473e-01
3.11884910e-01 -6.51547492e-01 -7.59927332e-01 -7.06787348e-01
2.19796211e-01 3.52555782e-01 6.92077637e-01 -5.64575978e-02
-3.85391325e-01 6.19721174e-01 -1.10095465e+00 4.21438307e-01
5.37141740e-01 1.19021857e+00 5.34161389e-01 4.02665496e-01
7.58782774e-02 1.17521083e+00 -2.07538158e-01 -8.59607011e-02
2.42639661e-01 -4.11606014e-01 -8.22398543e-01 4.39211011e-01
5.62237442e-01 -4.89489317e-01 -5.67222059e-01 1.08542323e+00
6.05528653e-01 -3.99788529e-01 5.96499681e-01 -1.10037529e+00
-3.37744862e-01 5.26593208e-01 -9.12452221e-01 1.71552286e-01
-1.69217601e-01 6.19054288e-02 6.37998939e-01 -1.05409896e+00
7.12328970e-01 8.13730001e-01 9.73210454e-01 5.93664706e-01
-9.55348372e-01 -7.04182088e-01 9.24504474e-02 2.10990071e-01
-1.17892134e+00 -4.50571507e-01 5.78588784e-01 -2.15987071e-01
6.21614873e-01 3.83258253e-01 9.09340322e-01 1.04533136e+00
1.69538289e-01 1.18574023e+00 8.90771866e-01 -5.86582601e-01
-3.77927087e-02 1.67047873e-01 -1.38498127e-01 1.42367935e+00
2.28781834e-01 -1.27268448e-01 -2.00976729e-01 -6.84794933e-02
9.66613948e-01 2.93515116e-01 -2.97990918e-01 -7.03388989e-01
-1.15040970e+00 7.13653266e-01 8.61011624e-01 3.31652999e-01
-2.00459003e-01 3.04466873e-01 4.15383965e-01 -1.28320605e-01
5.02454579e-01 -8.97327289e-02 -9.70124919e-03 5.04142404e-01
-8.89426589e-01 -2.41410255e-01 5.38167894e-01 8.02660167e-01
2.00763762e-01 -8.24439153e-02 -1.75059438e-01 8.51379216e-01
9.84857455e-02 4.72785085e-01 5.66534758e-01 -5.21853268e-01
-8.12051296e-02 7.58486509e-01 -2.49974743e-01 -6.21419549e-01
-5.44223011e-01 -9.69048023e-01 -9.19671059e-01 1.62887692e-01
5.45991242e-01 2.82512635e-01 -1.34107339e+00 1.28592634e+00
6.32976651e-01 1.11939929e-01 -2.03677610e-01 6.90101981e-01
1.07035244e+00 2.48084709e-01 5.09640835e-02 -1.84744939e-01
1.57289374e+00 -1.22132838e+00 -5.63201368e-01 -2.83044040e-01
1.07271481e+00 -7.32206941e-01 8.37122560e-01 1.53148055e-01
-1.07697868e+00 -1.74117088e-01 -9.36289370e-01 -1.24827199e-01
-1.09400116e-01 6.94164157e-01 8.62525582e-01 6.53346717e-01
-9.25251007e-01 2.80910522e-01 -1.24494040e+00 -7.64280856e-01
1.03586245e+00 2.92921364e-01 -4.11734879e-01 -4.40487921e-01
-4.92586643e-01 1.05583930e+00 8.01916048e-02 -1.92231461e-02
-1.52298295e+00 -9.54375446e-01 -7.68438041e-01 -2.11828306e-01
6.43038273e-01 -6.09398842e-01 1.67375576e+00 -1.02729893e+00
-1.04585385e+00 1.19850898e+00 -3.58035080e-02 -5.06754458e-01
8.17727029e-01 1.53681934e-01 1.30733624e-02 3.85157794e-01
1.65957958e-01 7.45784879e-01 7.34381497e-01 -1.13728893e+00
-7.08125710e-01 -2.86057174e-01 -2.39766344e-01 1.20125733e-01
-6.05332017e-01 -5.60874818e-03 -7.62641788e-01 -5.72965324e-01
2.13521078e-01 -9.99020636e-01 -5.67810059e-01 8.15985382e-01
-4.58224952e-01 7.55724087e-02 9.53186214e-01 -7.07961977e-01
8.47477615e-01 -2.16918087e+00 -1.02937281e-01 5.22599630e-02
5.39573789e-01 2.31899589e-01 -9.66218859e-02 -1.07269615e-01
-5.16021177e-02 -1.21769436e-01 -6.68878332e-02 -2.28521749e-01
-3.60914111e-01 3.11099887e-01 1.04213111e-01 8.66747558e-01
-1.69409588e-02 1.12823117e+00 -9.44836915e-01 -9.43355203e-01
2.62145668e-01 3.12886328e-01 -2.94154584e-01 4.93807606e-02
-8.07296857e-03 2.00444564e-01 -3.55168372e-01 1.21219099e+00
5.85678399e-01 -7.26151347e-01 2.68615425e-01 -5.03859639e-01
2.40761146e-01 -8.21181536e-02 -6.11114264e-01 1.47189581e+00
-3.61075073e-01 5.81592441e-01 2.61198819e-01 -1.12439263e+00
4.01711076e-01 3.38089257e-01 8.36895227e-01 -6.68407857e-01
2.45129108e-01 3.82918239e-01 8.27522203e-02 -6.18993819e-01
2.37703957e-02 -2.09826455e-01 3.85653734e-01 2.01097667e-01
-9.35749412e-02 -1.19147981e-02 -5.29552512e-02 1.02731280e-01
1.43639266e+00 -2.17550173e-01 6.66960955e-01 -1.48591757e-01
9.39484015e-02 2.54044354e-01 6.45374835e-01 7.21604407e-01
-3.55265260e-01 5.95913231e-01 4.14510518e-01 -4.93153363e-01
-8.72006655e-01 -9.86147940e-01 -3.34138811e-01 8.08484674e-01
9.56575647e-02 -1.87195405e-01 -6.99740946e-01 -1.02301681e+00
2.15939935e-02 2.50075191e-01 -8.74313593e-01 -2.00824454e-01
-6.60772085e-01 -1.02394795e+00 4.25913155e-01 7.77770281e-01
3.89642149e-01 -7.64079928e-01 -6.38984621e-01 3.78490500e-02
-2.96765149e-01 -1.26858306e+00 -2.80914128e-01 4.92323130e-01
-1.05550444e+00 -1.33358228e+00 -1.02822840e+00 -1.19984365e+00
1.17943001e+00 4.91531283e-01 9.87930357e-01 2.94196725e-01
-1.05284083e+00 1.21249802e-01 -4.04172570e-01 -5.13515651e-01
-7.02264309e-01 -9.06227008e-02 -4.05842900e-01 -9.40939710e-02
3.93690802e-02 -9.06361639e-02 -7.95580268e-01 4.72174168e-01
-8.98801565e-01 3.84907752e-01 9.40266788e-01 1.02653813e+00
7.91010976e-01 8.95573273e-02 1.39766917e-01 -9.78688359e-01
-5.80119863e-02 -4.37842846e-01 -3.23038250e-01 4.16091442e-01
-2.53834903e-01 -3.27036053e-01 1.92558363e-01 -6.23120844e-01
-6.86027825e-01 5.29823601e-01 5.83383292e-02 -6.38373494e-01
1.53068453e-01 2.75128633e-01 2.82755792e-01 -6.85799778e-01
7.58103430e-01 2.26632208e-01 7.69723430e-02 -3.17672640e-01
5.76558784e-02 3.48989695e-01 3.96259278e-01 -1.57258883e-02
6.55288339e-01 8.91148031e-01 3.45924616e-01 -7.01167226e-01
-1.12340105e+00 -5.05618393e-01 -4.05179948e-01 -4.36628699e-01
6.60036981e-01 -9.08894002e-01 -1.21805906e-01 3.51486713e-01
-7.67059028e-01 -4.98333782e-01 -4.27092820e-01 3.11789513e-01
-4.15953964e-01 4.04397249e-01 -8.06749344e-01 -4.77603167e-01
-4.53251481e-01 -1.23405242e+00 1.31848240e+00 -8.30287859e-02
1.82097495e-01 -7.77223825e-01 -3.37510169e-01 4.49149966e-01
4.20534670e-01 4.40837264e-01 8.39156628e-01 -3.55283827e-01
-7.00000107e-01 -4.09952581e-01 -3.65833431e-01 2.44179472e-01
3.46152842e-01 -1.84665859e-01 -9.47403669e-01 -4.88781691e-01
-1.59803599e-01 -3.58030140e-01 1.22081625e+00 7.41646528e-01
1.37544024e+00 1.04787514e-01 -1.16518736e+00 6.46665156e-01
1.17126775e+00 -3.79296653e-02 4.76836979e-01 2.72893220e-01
7.48374224e-01 3.37076306e-01 6.33569360e-01 1.89557984e-01
3.01062707e-02 5.33261776e-01 7.23942339e-01 -1.04403210e+00
-7.62327373e-01 -1.85134456e-01 2.23414391e-01 4.83601421e-01
1.52069002e-01 -2.10993245e-01 -1.09443116e+00 8.12591374e-01
-1.83999860e+00 -4.60449755e-01 -3.49434391e-02 1.87183237e+00
6.41241074e-01 1.00919101e-02 -4.42738123e-02 8.53729323e-02
6.66396737e-01 -2.61763394e-01 -5.02402842e-01 3.22532117e-01
1.61515981e-01 1.38704076e-01 7.94643939e-01 4.47879769e-02
-1.32248747e+00 8.33692074e-01 6.90847206e+00 1.17393327e+00
-1.05757844e+00 4.98658240e-01 8.56081128e-01 -1.83850452e-01
1.68940470e-01 -3.66313249e-01 -7.24465072e-01 2.26504114e-02
3.91693503e-01 9.91731659e-02 -1.27464458e-01 1.00270677e+00
-5.98150492e-02 -1.22934870e-01 -1.24377096e+00 9.91543710e-01
1.55131757e-01 -1.41966879e+00 1.16671883e-02 2.13241249e-01
5.49258113e-01 7.74857625e-02 1.03653610e-01 7.79485554e-02
2.71861553e-01 -1.11231971e+00 6.70656681e-01 1.84732273e-01
9.00991857e-01 -5.00605226e-01 8.20419908e-01 2.27095664e-01
-1.24001622e+00 -1.80447340e-01 -4.09946650e-01 6.26724660e-01
-3.74993421e-02 7.45924056e-01 -1.33705318e+00 2.56443948e-01
6.84763491e-01 8.23529303e-01 -7.48263896e-01 1.25954580e+00
1.19378708e-01 6.04891539e-01 -3.35211307e-01 1.16999023e-01
6.96530491e-02 5.81144750e-01 3.71182799e-01 1.19251132e+00
5.51161587e-01 -1.31127626e-01 6.38916969e-01 5.48731863e-01
-2.38559335e-01 1.65636942e-01 -6.44113243e-01 1.29816249e-01
8.27601478e-02 1.66029394e+00 -1.09361672e+00 -3.50984901e-01
-4.85983759e-01 8.05604279e-01 3.21340650e-01 -3.83141376e-02
-9.39069271e-01 3.49307001e-01 1.04750112e-01 4.88439113e-01
4.10719961e-01 3.11499596e-01 1.57309137e-02 -9.49480057e-01
-4.04528566e-02 -9.21332240e-01 5.26846349e-01 -6.16033554e-01
-1.22401822e+00 4.92394686e-01 -3.78352821e-01 -1.45694387e+00
3.42666596e-01 -7.37510622e-01 -2.89635420e-01 2.15459645e-01
-1.64486587e+00 -1.59651470e+00 -6.18801653e-01 6.91736758e-01
6.05695248e-01 3.12844142e-02 8.86848509e-01 3.18290889e-01
-6.39356434e-01 8.04040432e-01 2.18005646e-02 2.28457257e-01
6.09228492e-01 -1.09366894e+00 -1.09459236e-01 7.14887261e-01
-2.49390796e-01 -3.76913585e-02 3.53113443e-01 -7.12391198e-01
-1.35889137e+00 -1.45015538e+00 3.16113710e-01 -2.31210098e-01
6.19759321e-01 -3.09776992e-01 -5.06017864e-01 8.26270700e-01
2.53474582e-02 7.11264908e-01 6.76402926e-01 -3.50761294e-01
-1.28821686e-01 -2.38014236e-01 -1.23290181e+00 6.89459145e-01
9.47551131e-01 -2.83927560e-01 -1.67336464e-02 9.51860845e-01
4.27369714e-01 -7.43412137e-01 -8.76082599e-01 6.71064973e-01
4.50856388e-01 -6.17916644e-01 1.11922526e+00 -1.86966583e-01
2.58835882e-01 -1.14811547e-01 -6.19256161e-02 -1.00143290e+00
-4.16259915e-01 -3.08558106e-01 -9.76401567e-02 6.51191831e-01
3.77787173e-01 -3.18516612e-01 1.04367983e+00 6.15658164e-02
-3.35070103e-01 -9.36002910e-01 -1.08475900e+00 -6.60045981e-01
-1.69446866e-03 -1.27174422e-01 -2.62934295e-03 7.98823476e-01
-5.56499138e-02 -2.38616094e-01 -2.01501727e-01 2.06918359e-01
8.69697928e-01 1.62986070e-02 3.96780372e-01 -1.03268445e+00
-3.00579935e-01 -4.03826416e-01 -5.58658659e-01 -8.13790560e-01
-2.77518153e-01 -1.17609167e+00 4.21188585e-03 -1.54770768e+00
7.73804784e-01 -4.82817054e-01 -4.20670867e-01 8.79785597e-01
-1.65643096e-02 6.91967547e-01 -1.02152772e-01 4.34939787e-02
-5.55003881e-01 3.01128536e-01 1.63200259e+00 -5.31107843e-01
3.87186348e-01 -1.64202526e-01 -5.16051531e-01 7.66831458e-01
7.39409387e-01 -6.99465692e-01 -1.78346783e-01 -1.69014677e-01
-2.06787884e-01 1.98307827e-01 6.05122507e-01 -9.19292152e-01
2.87206620e-01 -3.47306691e-02 7.20653296e-01 -6.68425918e-01
3.29833835e-01 -8.89564335e-01 3.62848118e-03 1.03692460e+00
-2.07054079e-01 -3.34941894e-01 3.08027983e-01 5.97582459e-01
1.66264828e-02 -1.26007482e-01 1.15576851e+00 -4.97734368e-01
-4.20108587e-01 4.43572611e-01 -4.19360399e-01 -3.66842121e-01
1.49294066e+00 -3.07452649e-01 -3.67756724e-01 1.05548343e-02
-6.64804697e-01 1.88877553e-01 2.94698417e-01 1.76053792e-01
6.81465864e-01 -1.29932141e+00 -7.60292232e-01 7.17765018e-02
3.22221160e-01 8.29818286e-03 1.48283049e-01 1.42367661e+00
-7.15262771e-01 2.68640280e-01 -1.11941293e-01 -8.75708103e-01
-1.72806191e+00 6.10655963e-01 8.52625370e-01 -5.60794711e-01
-1.23165786e+00 9.16652381e-01 7.56110489e-01 -1.56808987e-01
5.49681842e-01 -5.69440126e-01 -4.42880318e-02 -2.35849023e-01
4.91550535e-01 1.30120158e-01 3.94279301e-01 -4.08629417e-01
-4.21973109e-01 3.74481052e-01 -6.05550408e-01 4.32654977e-01
1.23335338e+00 2.87564039e-01 -5.37040047e-02 -4.92334217e-02
1.01072168e+00 -3.28838199e-01 -1.06877363e+00 -4.20343965e-01
-3.02973419e-01 -3.42401654e-01 6.12751603e-01 -8.09034944e-01
-1.58047605e+00 6.55542970e-01 1.09047544e+00 9.64182895e-04
1.26890993e+00 4.35981005e-01 8.28401923e-01 2.24280074e-01
2.03479365e-01 -8.10704827e-01 3.71629149e-01 -1.18428282e-01
7.37193286e-01 -1.31976950e+00 1.03896312e-01 -8.50473762e-01
-3.08803260e-01 1.14875185e+00 6.86110139e-01 -1.35047808e-02
5.76784849e-01 5.81044257e-01 1.70940146e-01 -4.78236616e-01
-6.94909573e-01 -3.21013153e-01 1.38908312e-01 5.88822246e-01
2.54394352e-01 1.31380618e-01 8.01142529e-02 1.26397491e-01
2.97587335e-01 8.40069503e-02 4.24465209e-01 1.09888947e+00
-5.78080595e-01 -9.00471807e-01 -4.39324468e-01 6.62765741e-01
-5.93945682e-01 -2.12756190e-02 -3.57795745e-01 9.42041218e-01
2.39389576e-02 6.39941275e-01 -2.72902966e-01 -1.48834730e-03
2.01786131e-01 -4.05718386e-01 7.17774689e-01 -5.91220498e-01
-5.24673522e-01 4.21749681e-01 8.12727287e-02 -6.50661588e-01
-2.48337448e-01 -5.64083338e-01 -1.08608282e+00 4.22864221e-02
-6.06839180e-01 -4.12387043e-01 5.05923092e-01 6.61957920e-01
-8.03775713e-02 9.05395329e-01 4.39215541e-01 -6.47279620e-01
-6.51320040e-01 -7.69601285e-01 -6.51082695e-01 1.21212266e-01
4.68494952e-01 -7.40283191e-01 -9.50467065e-02 5.17181084e-02] | [15.191291809082031, -2.4718027114868164] |
79ecd76e-c5fe-4e2c-bbb0-6d5a83422212 | accelerating-code-search-with-deep-hashing | 2203.15287 | null | https://arxiv.org/abs/2203.15287v2 | https://arxiv.org/pdf/2203.15287v2.pdf | Accelerating Code Search with Deep Hashing and Code Classification | Code search is to search reusable code snippets from source code corpus based on natural languages queries. Deep learning-based methods of code search have shown promising results. However, previous methods focus on retrieval accuracy but lacked attention to the efficiency of the retrieval process. We propose a novel method CoSHC to accelerate code search with deep hashing and code classification, aiming to perform an efficient code search without sacrificing too much accuracy. To evaluate the effectiveness of CoSHC, we apply our method to five code search models. Extensive experimental results indicate that compared with previous code search baselines, CoSHC can save more than 90% of retrieval time meanwhile preserving at least 99% of retrieval accuracy. | ['Michael R. Lyu', 'Dongmei Zhang', 'Shi Han', 'Hongyu Zhang', 'Lun Du', 'Yanlin Wang', 'Wenchao Gu'] | 2022-03-29 | null | https://aclanthology.org/2022.acl-long.181 | https://aclanthology.org/2022.acl-long.181.pdf | acl-2022-5 | ['code-classification', 'code-search', 'code-search'] | ['computer-code', 'computer-code', 'computer-vision'] | [-5.92060447e-01 -6.41181052e-01 -5.46336949e-01 -2.73644123e-02
-1.15836406e+00 -5.26566327e-01 3.72937083e-01 5.31680942e-01
-4.62160140e-01 4.06605899e-02 1.89488288e-02 -6.31645441e-01
-9.47866887e-02 -7.70601273e-01 -4.75195736e-01 -1.38335973e-01
-1.90258726e-01 -8.09867866e-03 4.03964996e-01 -4.40589078e-02
7.65281439e-01 2.62087397e-02 -1.52466428e+00 3.66743803e-01
9.80501056e-01 8.33285511e-01 4.94692266e-01 4.50080544e-01
-2.15812817e-01 9.95485365e-01 -5.12110054e-01 -5.92512429e-01
1.05734505e-01 4.54562493e-02 -1.02305269e+00 -9.84566867e-01
2.03799039e-01 -5.50158858e-01 -6.04874790e-01 1.28746891e+00
5.76487064e-01 -2.32977167e-01 5.40829778e-01 -9.65029299e-01
-1.10966694e+00 6.29564047e-01 -5.68232596e-01 3.74968112e-01
6.85823679e-01 -1.09562419e-01 1.38501060e+00 -1.14992893e+00
4.79232341e-01 7.58327842e-01 8.47981215e-01 3.47433776e-01
-1.11866987e+00 -7.59322524e-01 -4.60505545e-01 2.21335337e-01
-1.79929078e+00 -3.00949126e-01 4.92316306e-01 -5.83574116e-01
1.46116948e+00 2.81148553e-02 4.19536918e-01 7.25939691e-01
4.65094179e-01 9.84112918e-01 2.93726861e-01 -3.80469531e-01
1.80601418e-01 -7.53570870e-02 3.63111585e-01 1.15047681e+00
3.01598191e-01 -9.73520428e-02 -3.87582928e-01 -7.49811292e-01
1.77387193e-01 2.86237359e-01 -1.70927852e-01 -2.87977457e-01
-9.24943984e-01 1.22317064e+00 6.07884586e-01 4.00588542e-01
-1.50134906e-01 6.52783573e-01 8.62214148e-01 3.97060812e-01
6.23969063e-02 5.51137924e-01 -2.92057812e-01 -5.24546504e-01
-1.08406627e+00 2.82179624e-01 5.96305072e-01 1.33114409e+00
9.75290298e-01 -2.87408054e-01 -2.55458981e-01 9.21231806e-01
6.34583831e-01 6.63278162e-01 9.16446388e-01 -6.81711197e-01
5.21366715e-01 9.67582822e-01 -3.08271557e-01 -1.32120287e+00
-5.75703792e-02 -2.22180098e-01 -3.34224194e-01 -2.03240290e-01
-2.50379562e-01 4.47006851e-01 -3.36165607e-01 1.25647879e+00
-3.88094425e-01 -5.11337638e-01 -1.52630180e-01 4.84556735e-01
8.02334070e-01 6.64948702e-01 -6.34110859e-03 3.20445120e-01
1.32630146e+00 -1.20853806e+00 -2.81455427e-01 -6.21910393e-03
1.15324891e+00 -1.04014969e+00 1.18510687e+00 -2.30507739e-02
-1.01749003e+00 -4.53801632e-01 -9.95536029e-01 -2.98321515e-01
-3.45391542e-01 4.25026745e-01 7.18150616e-01 8.09856892e-01
-1.35277700e+00 2.01212600e-01 -9.36474323e-01 -1.86027408e-01
2.72547424e-01 3.19106489e-01 -2.44924519e-02 -1.82019860e-01
-8.74757051e-01 5.32013834e-01 3.22386801e-01 -5.17004132e-01
-9.86187041e-01 -7.27416873e-01 -8.90512526e-01 6.38086975e-01
-1.30166575e-01 -2.73475498e-01 1.31332731e+00 -4.14437503e-01
-8.27518404e-01 7.89443970e-01 -9.06266943e-02 -4.89312440e-01
4.16897945e-02 -3.30224365e-01 -3.01224589e-01 2.34415993e-01
2.60276496e-01 5.79712093e-01 3.99011433e-01 -6.91973329e-01
-3.67353261e-01 2.52954096e-01 3.33026826e-01 -3.83980840e-01
-9.14548576e-01 2.44124144e-01 -8.62823129e-01 -6.35925889e-01
-3.47238302e-01 -9.30757940e-01 1.54781535e-01 1.08753972e-01
1.39796153e-01 -5.71644247e-01 7.78189123e-01 -3.52672070e-01
1.92411649e+00 -2.46377897e+00 -2.31220812e-01 7.19650015e-02
3.80835295e-01 4.45456117e-01 -4.11222875e-01 8.29393923e-01
4.07928199e-01 1.26673624e-01 -1.33859277e-01 -8.56389701e-02
3.68852705e-01 -2.33106196e-01 -4.84923810e-01 4.67803180e-01
-1.01650223e-01 1.22621799e+00 -8.86246622e-01 -8.36712718e-01
-1.12897120e-01 3.40279222e-01 -1.20196557e+00 1.25693306e-01
4.44898382e-02 -7.06006587e-01 -7.41772354e-01 9.39162791e-01
4.55219984e-01 -6.79861188e-01 -1.50765643e-01 3.45186055e-01
-5.65622859e-02 4.20130849e-01 -3.39175105e-01 1.96088159e+00
-8.32172692e-01 8.00300956e-01 -4.32160378e-01 -6.69895291e-01
9.09134686e-01 1.28377393e-01 3.34331572e-01 -1.18984008e+00
-1.33847147e-01 6.67867780e-01 -2.87549973e-01 -6.54984891e-01
6.30508840e-01 6.11457407e-01 -3.36479098e-01 6.62518084e-01
-1.78639814e-01 8.61059129e-02 1.68154150e-01 3.40183437e-01
1.65340877e+00 -1.36020079e-01 8.15802291e-02 -6.30024672e-01
8.04625630e-01 1.21224307e-01 1.05778642e-01 9.52340066e-01
-2.13502884e-01 3.53794664e-01 1.56132296e-01 -6.37326241e-01
-8.55767131e-01 -6.71967745e-01 -1.94672376e-01 1.35812294e+00
4.06300396e-01 -1.05770802e+00 -7.20052302e-01 -8.06221724e-01
9.09160450e-02 1.23442411e-01 -3.19700658e-01 -5.44384778e-01
-7.38695502e-01 -4.99428958e-01 1.05000567e+00 4.81643617e-01
6.74451053e-01 -9.48996842e-01 -1.04042947e+00 -1.62038170e-02
-2.94143081e-01 -6.88319325e-01 -1.00482178e+00 -3.62899192e-02
-5.65001309e-01 -1.11698079e+00 -9.55609620e-01 -1.24226451e+00
6.37658179e-01 5.92172027e-01 1.37099111e+00 1.06879854e+00
-5.97166419e-01 4.04023111e-01 -7.74359286e-01 2.95709193e-01
-5.17972291e-01 6.72858477e-01 -4.75154102e-01 -8.05404365e-01
6.22550964e-01 -2.94723809e-01 -1.03201270e+00 1.48664281e-01
-9.69538808e-01 -4.68843877e-01 7.87398338e-01 8.84679258e-01
-9.14239436e-02 -1.06411695e-01 5.02368987e-01 -3.46246988e-01
8.62038970e-01 -6.37913883e-01 -8.65337193e-01 4.44536477e-01
-1.18749464e+00 4.76355791e-01 6.33358061e-01 -2.99840569e-01
-6.90472960e-01 2.49587540e-02 -3.37963581e-01 -2.07081676e-01
5.65147460e-01 8.88166368e-01 6.30199075e-01 -5.22193074e-01
8.09870183e-01 6.17232025e-01 -2.42726237e-01 -5.41336715e-01
5.44079207e-03 8.03584814e-01 1.53696194e-01 -7.99947083e-01
6.27333403e-01 1.47430837e-01 -5.27346134e-01 -9.15562734e-02
-7.25556687e-02 -8.02434504e-01 -3.12908143e-01 -4.03037388e-03
7.37811744e-01 -9.74319458e-01 -9.40555274e-01 1.76217202e-02
-1.12365842e+00 -6.57904148e-02 4.07520413e-01 3.91044438e-01
-4.37333554e-01 7.42975652e-01 -9.24936354e-01 -4.69490945e-01
-6.63422465e-01 -1.59709811e+00 1.47116375e+00 -2.00857863e-01
-8.55926201e-02 -7.24871635e-01 4.60362434e-01 2.79736761e-02
8.62568855e-01 -4.79933262e-01 1.42381799e+00 -5.21218479e-01
-9.92823958e-01 -5.25279403e-01 -5.19410610e-01 -2.09266767e-01
-2.02902123e-01 -8.16315264e-02 -7.03282356e-01 -6.64093852e-01
-3.78982991e-01 -6.10978305e-01 9.61013913e-01 -2.38507703e-01
1.26874006e+00 -2.38896802e-01 -4.61270243e-01 5.51645696e-01
1.96829450e+00 5.97992361e-01 6.16075516e-01 6.87141538e-01
2.25932360e-01 -9.76233408e-02 4.07278359e-01 6.29745722e-01
4.84747410e-01 7.60673523e-01 5.55465460e-01 3.79922926e-01
-8.39191005e-02 -2.45665282e-01 4.39352542e-01 1.28503597e+00
3.91645581e-01 9.08284262e-03 -1.45384610e+00 8.15529108e-01
-1.74981487e+00 -6.56328976e-01 1.01250663e-01 2.04817963e+00
9.94408309e-01 -3.13230455e-02 -1.46417439e-01 -9.96639431e-02
4.33655351e-01 9.16799754e-02 -2.86532104e-01 -3.22751760e-01
3.63932550e-01 2.05689490e-01 4.34921920e-01 1.60121053e-01
-9.40357387e-01 7.84137011e-01 7.16664791e+00 1.11063766e+00
-8.66711259e-01 1.88760638e-01 5.01809455e-02 2.03045085e-01
-5.06350100e-01 1.92627549e-01 -6.42769635e-01 6.82970107e-01
8.40931773e-01 -4.56974924e-01 5.05846798e-01 1.51297748e+00
-6.01399958e-01 3.09337955e-02 -1.10883200e+00 1.22837949e+00
1.30654410e-01 -1.41323900e+00 4.41691093e-02 -2.72716340e-02
7.62274146e-01 2.45222062e-01 7.27152154e-02 9.64535713e-01
3.15790206e-01 -8.00399423e-01 7.88132071e-01 2.25237310e-01
5.68905473e-01 -5.67715049e-01 7.45299399e-01 2.05552280e-01
-1.55515981e+00 -4.92628336e-01 -5.00536740e-01 1.02752946e-01
-4.43369418e-01 2.75039971e-01 -4.80663687e-01 1.75114617e-01
8.27243626e-01 6.55796528e-01 -1.05726397e+00 1.44857192e+00
3.17073137e-01 1.58802122e-01 1.04101397e-01 -6.26856387e-01
4.97354805e-01 4.03294414e-01 -1.94302071e-02 1.61564398e+00
6.33230150e-01 -3.54642123e-01 1.31111756e-01 1.16411352e+00
-3.12091053e-01 2.93396831e-01 -7.18975127e-01 -1.75350606e-01
7.48953164e-01 8.31909716e-01 -7.24790275e-01 -3.25066596e-01
-8.70853543e-01 8.88237000e-01 3.88721704e-01 1.68287784e-01
-1.01306450e+00 -9.88767564e-01 3.97560835e-01 -2.72900909e-01
4.67881471e-01 -2.50120968e-01 2.24076793e-01 -1.14675200e+00
3.37566048e-01 -8.69614840e-01 3.76340866e-01 -5.71739912e-01
-7.97650754e-01 6.31455839e-01 -1.61122814e-01 -1.36266196e+00
-9.95051414e-02 -3.95801425e-01 -3.12723637e-01 4.78422195e-01
-1.61158860e+00 -6.92378879e-01 -2.36712500e-01 4.36846465e-01
6.16458952e-01 -5.72630286e-01 8.78758609e-01 1.02291536e+00
-2.23308817e-01 9.63127673e-01 4.55684870e-01 3.62666517e-01
5.22039175e-01 -1.01336753e+00 5.54215848e-01 5.56044877e-01
1.24266595e-01 1.31433237e+00 1.94278404e-01 -3.84768546e-01
-1.89722741e+00 -6.99728131e-01 9.30572510e-01 -3.84275198e-01
8.08939219e-01 -2.91947782e-01 -8.59595060e-01 2.12714136e-01
2.41280824e-01 6.85921907e-02 5.60096145e-01 -8.68834332e-02
-9.18104410e-01 -1.36903644e-01 -7.11401999e-01 4.06859607e-01
7.71135390e-01 -1.21613419e+00 -4.45868611e-01 1.55156180e-01
8.10358703e-01 -1.77022845e-01 -8.71496141e-01 4.32019055e-01
7.49327838e-01 -8.64603460e-01 1.11895454e+00 -1.20901659e-01
6.93721950e-01 -1.32453129e-01 -2.93196857e-01 -4.92580235e-01
-3.21918309e-01 -1.93161368e-01 -1.15597382e-01 9.61221576e-01
3.37914109e-01 -2.76134491e-01 5.28898060e-01 4.43325937e-01
1.12610189e-02 -6.53709531e-01 -6.97191060e-01 -1.06661570e+00
2.33778715e-01 -1.39024094e-01 7.42680609e-01 8.33743870e-01
3.63489836e-01 -1.09278914e-02 -9.82603654e-02 -1.19651340e-01
1.75816461e-01 6.50652289e-01 5.22116244e-01 -1.01496959e+00
-3.09033394e-01 -9.09488618e-01 -4.64993417e-01 -1.20592570e+00
5.00843525e-01 -1.28335476e+00 1.74907774e-01 -1.33763194e+00
9.52243090e-01 -4.72324759e-01 -4.55119520e-01 5.60828924e-01
-9.70273390e-02 2.19631493e-01 2.56002378e-02 8.14433098e-01
-1.10873365e+00 6.41788423e-01 5.47998309e-01 -5.62139213e-01
3.86146545e-01 -3.24919432e-01 -5.16820133e-01 2.27384210e-01
4.01249111e-01 -6.97586834e-01 -4.41769272e-01 -9.67082500e-01
7.94308305e-01 1.78558379e-01 2.09603027e-01 -1.05080342e+00
6.13339841e-01 4.61358339e-01 -2.60469109e-01 -4.12330121e-01
-2.10012406e-01 -8.13072205e-01 -1.67764038e-01 1.04491556e+00
-6.27993166e-01 6.97827876e-01 2.06547648e-01 4.89305258e-01
-4.68890965e-01 -9.96868432e-01 5.38539588e-01 -4.44762930e-02
-8.03192675e-01 1.13195874e-01 -7.19315231e-01 7.23501295e-02
6.14143491e-01 2.36488208e-01 -4.91118103e-01 4.06906977e-02
1.93109550e-03 1.43181831e-01 6.91372097e-01 6.73479319e-01
6.57169878e-01 -1.60396671e+00 -2.88695544e-01 1.74259752e-01
7.04595268e-01 -6.60381019e-01 -2.70537019e-01 4.85026717e-01
-1.09933591e+00 1.01159728e+00 3.05131897e-02 -4.05699670e-01
-1.25802124e+00 9.54092145e-01 1.37360483e-01 -1.57153681e-01
-5.36913216e-01 6.97395504e-01 1.97627898e-02 -4.24811035e-01
4.48738575e-01 -2.49800801e-01 1.10572122e-01 -3.58449459e-01
5.94532728e-01 2.52525896e-01 2.01368526e-01 -1.92250460e-01
-6.43921673e-01 7.56714344e-01 -2.12526157e-01 4.26527828e-01
1.02358150e+00 8.27760398e-02 -3.09697956e-01 -7.40251467e-02
1.89442956e+00 2.14175269e-01 -6.26602054e-01 -2.76527047e-01
6.52694404e-01 -7.85648704e-01 -1.19887292e-02 -5.24400234e-01
-1.21118116e+00 8.37024331e-01 8.36525738e-01 2.83034325e-01
1.11171985e+00 -3.03543129e-05 1.01552999e+00 9.32407975e-01
8.58019650e-01 -8.73991191e-01 4.30209935e-01 6.07024908e-01
4.86298054e-01 -1.33227515e+00 -2.20332500e-02 1.98792025e-01
3.13760340e-02 1.24638772e+00 4.80916560e-01 -1.42203718e-01
7.42502272e-01 4.14626360e-01 -1.13962531e-01 -4.76743519e-01
-7.80087769e-01 3.49427164e-02 2.43484214e-01 5.71837798e-02
6.69817328e-01 -5.16471505e-01 -5.11960685e-01 1.17304318e-01
2.55084544e-01 -1.19097419e-01 2.30471808e-02 1.25912762e+00
-6.01885915e-01 -1.08834887e+00 7.37390071e-02 2.51964122e-01
-6.87158585e-01 -6.87443495e-01 -5.25629334e-02 5.56535244e-01
-4.45166051e-01 6.78120852e-01 -2.33959496e-01 -4.59794074e-01
-1.99928749e-02 -3.84046286e-02 -6.34715036e-02 -4.99037474e-01
-7.94286489e-01 -2.04844952e-01 -5.08342743e-01 -7.25367367e-01
-1.56210244e-01 -3.24085683e-01 -1.23027170e+00 -3.44504535e-01
-6.54015601e-01 4.66065019e-01 4.69048530e-01 3.96923125e-01
7.43472755e-01 2.06570551e-01 6.17067873e-01 -2.57102817e-01
-6.06652379e-01 -6.09655976e-01 -6.04455657e-02 2.62193531e-01
4.69727337e-01 -5.75084329e-01 -3.09640199e-01 -5.04909605e-02] | [7.49930477142334, 8.083356857299805] |
fdd8aad1-0f7e-4eee-a765-1ddf62424856 | learning-to-refine-human-pose-estimation | 1804.07909 | null | http://arxiv.org/abs/1804.07909v1 | http://arxiv.org/pdf/1804.07909v1.pdf | Learning to Refine Human Pose Estimation | Multi-person pose estimation in images and videos is an important yet
challenging task with many applications. Despite the large improvements in
human pose estimation enabled by the development of convolutional neural
networks, there still exist a lot of difficult cases where even the
state-of-the-art models fail to correctly localize all body joints. This
motivates the need for an additional refinement step that addresses these
challenging cases and can be easily applied on top of any existing method. In
this work, we introduce a pose refinement network (PoseRefiner) which takes as
input both the image and a given pose estimate and learns to directly predict a
refined pose by jointly reasoning about the input-output space. In order for
the network to learn to refine incorrect body joint predictions, we employ a
novel data augmentation scheme for training, where we model "hard" human pose
cases. We evaluate our approach on four popular large-scale pose estimation
benchmarks such as MPII Single- and Multi-Person Pose Estimation, PoseTrack
Pose Estimation, and PoseTrack Pose Tracking, and report systematic improvement
over the state of the art. | ['Anna Khoreva', 'Mihai Fieraru', 'Leonid Pishchulin', 'Bernt Schiele'] | 2018-04-21 | null | null | null | null | ['multi-person-pose-estimation-and-tracking'] | ['computer-vision'] | [ 1.24732383e-01 3.31658840e-01 5.83120249e-02 -3.47707868e-01
-7.89665043e-01 -2.73815751e-01 4.68161970e-01 -1.18727818e-01
-5.98555326e-01 7.73973167e-01 3.60114902e-01 5.08811414e-01
1.35114580e-01 -3.33022594e-01 -9.83445287e-01 -3.05177033e-01
-2.62497761e-03 1.04287326e+00 5.26516795e-01 -3.94558072e-01
-2.75951892e-01 1.94033936e-01 -1.40626287e+00 -2.39842813e-02
4.46489066e-01 8.35242689e-01 -2.33952075e-01 6.84741974e-01
4.68623638e-01 2.82070398e-01 -6.05053961e-01 -6.94821596e-01
3.87184113e-01 -2.18767837e-01 -9.12591934e-01 5.82316294e-02
9.42935169e-01 -4.19640154e-01 -3.12488586e-01 7.32023478e-01
7.46527493e-01 1.07794024e-01 3.42505723e-01 -1.30286133e+00
1.02360025e-01 3.58892769e-01 -6.91890359e-01 -3.50035429e-02
7.33686447e-01 9.04940441e-02 6.37958527e-01 -6.06375098e-01
7.87735760e-01 1.49810457e+00 1.16873276e+00 8.11599255e-01
-1.21037900e+00 -5.17641723e-01 3.05646718e-01 -2.92180874e-03
-1.26315141e+00 -2.90362626e-01 5.84026396e-01 -4.22204792e-01
7.85244763e-01 -1.94283612e-02 8.61701190e-01 1.31501186e+00
2.57277220e-01 8.00925732e-01 8.57100189e-01 -2.48057008e-01
-1.56829596e-01 -3.89373332e-01 -1.66437060e-01 8.50280881e-01
4.10947770e-01 1.18058294e-01 -6.23704195e-01 8.05296283e-03
8.46575737e-01 -8.89845267e-02 -2.02019364e-01 -8.21500778e-01
-1.33792710e+00 4.33680028e-01 5.94849050e-01 -1.03792310e-01
-3.62246662e-01 5.33159852e-01 3.28804612e-01 -1.19733915e-01
4.60599542e-01 4.41394061e-01 -6.55756295e-01 -1.59461901e-01
-1.03688204e+00 9.03598189e-01 8.12775612e-01 6.83702290e-01
4.22776371e-01 -2.93405145e-01 -3.17209184e-01 5.04782140e-01
3.74439090e-01 3.98974359e-01 1.76938787e-01 -9.48451519e-01
6.20131910e-01 6.17594302e-01 2.76500583e-01 -8.66341352e-01
-8.75578761e-01 -6.38792217e-01 -6.48725629e-01 9.25453827e-02
7.42094398e-01 -2.44270876e-01 -1.14306784e+00 1.92444849e+00
7.18735456e-01 2.47160289e-02 -3.47061366e-01 1.18883109e+00
7.42842972e-01 1.78152740e-01 1.71975970e-01 2.23915622e-01
1.49924541e+00 -1.18550634e+00 -5.51427364e-01 -5.93485951e-01
3.97999734e-01 -6.16815031e-01 6.23619318e-01 5.00647426e-01
-1.22043622e+00 -7.53704906e-01 -9.92922127e-01 -8.90143290e-02
-1.17638908e-01 3.46951872e-01 5.51766813e-01 5.16697347e-01
-7.63595819e-01 8.16219687e-01 -1.16021121e+00 -5.76274037e-01
3.14922601e-01 7.27354825e-01 -8.28604698e-01 4.99957129e-02
-1.07250834e+00 1.18691933e+00 4.05405343e-01 4.71391678e-01
-8.55977237e-01 -4.65402514e-01 -1.04319537e+00 -2.42125809e-01
7.59243071e-01 -1.12674212e+00 1.40191376e+00 -5.65605283e-01
-1.47883022e+00 8.24096680e-01 1.35827288e-01 -4.28007990e-01
1.03062534e+00 -9.96798515e-01 -5.91836534e-02 4.97156009e-02
5.38896807e-02 9.82433915e-01 8.10470283e-01 -1.03983319e+00
-4.66674149e-01 -5.91513276e-01 8.34332630e-02 3.26247126e-01
-8.12520087e-02 3.90063189e-02 -1.00496280e+00 -5.62618196e-01
2.75752753e-01 -1.37426734e+00 -3.70702803e-01 1.96912378e-01
-5.98342359e-01 -2.09563598e-01 5.38053632e-01 -9.14263785e-01
9.57524240e-01 -1.63899410e+00 7.88100481e-01 9.48020220e-02
2.29834005e-01 8.78675207e-02 -7.63496161e-02 1.35826841e-01
-2.07411908e-02 -3.11400384e-01 -3.79328318e-02 -8.68704140e-01
1.15331344e-01 2.25960016e-01 3.13936114e-01 7.45445311e-01
2.27975398e-01 9.86496508e-01 -7.10144043e-01 -5.87388277e-01
2.36356616e-01 5.93805552e-01 -8.13200712e-01 3.16144913e-01
-2.76907682e-01 9.16437984e-01 -1.70465618e-01 4.93874013e-01
4.52261358e-01 -3.42276067e-01 2.24432260e-01 -5.19359052e-01
1.41445652e-01 8.22829083e-02 -1.52485859e+00 2.20275569e+00
-2.58048344e-02 1.61463693e-01 3.86422849e-03 -5.74736714e-01
5.04819334e-01 4.58305001e-01 7.52453089e-01 -1.81548730e-01
3.05569798e-01 1.02969192e-01 2.78406572e-02 -3.64439070e-01
5.46304941e-01 -1.97055891e-01 -3.72415692e-01 1.72596276e-01
4.29281145e-01 6.96575940e-02 1.58242926e-01 -1.02264680e-01
9.11156774e-01 8.93091977e-01 2.78796524e-01 -2.99506262e-02
6.05580091e-01 -1.56843483e-01 6.50410414e-01 4.86220002e-01
-3.13359588e-01 8.97765160e-01 3.00159127e-01 -7.56375849e-01
-1.00818396e+00 -1.04627132e+00 3.18837345e-01 1.33331549e+00
1.59935370e-01 -6.70844376e-01 -8.79126370e-01 -8.54260147e-01
1.27132282e-01 -2.44619936e-01 -7.47889698e-01 -1.59136459e-01
-9.69625831e-01 -6.71970725e-01 6.77592635e-01 9.23774421e-01
4.83069777e-01 -9.11050916e-01 -6.70551479e-01 1.93609059e-01
-3.69480371e-01 -1.29312956e+00 -4.30393249e-01 -4.17541005e-02
-7.18014300e-01 -1.15893090e+00 -9.43894625e-01 -6.03368461e-01
5.65534174e-01 -4.95215505e-01 1.29209387e+00 7.54548842e-03
-4.15602386e-01 2.71201760e-01 -1.19525336e-01 -3.38288993e-01
-1.40345618e-01 4.81156021e-01 3.73082399e-01 -2.95236558e-01
-2.66040433e-02 -3.11897725e-01 -7.33675659e-01 3.77056032e-01
-3.81182492e-01 2.03494653e-01 6.51495755e-01 7.06902683e-01
5.44116378e-01 -3.13855737e-01 2.37391636e-01 -6.13970637e-01
2.35655397e-01 2.65964121e-02 -4.41631913e-01 1.76903948e-01
-1.75851911e-01 2.98486382e-01 1.54836297e-01 -4.07971501e-01
-8.86172712e-01 7.24993050e-01 -5.76938748e-01 -3.03810090e-01
-3.00211996e-01 3.34441721e-01 -1.40389994e-01 -1.35049611e-01
5.95529258e-01 -2.41638631e-01 1.45163760e-01 -7.94374228e-01
2.45261490e-01 9.00169760e-02 9.90086973e-01 -6.72658026e-01
1.03438103e+00 4.05248284e-01 1.82534829e-01 -3.66320759e-01
-1.10648465e+00 -5.07739484e-01 -1.13462913e+00 -3.46097380e-01
1.12256706e+00 -1.07332361e+00 -9.42175746e-01 5.70552051e-01
-1.20392513e+00 -2.48555988e-01 -7.87192881e-02 4.75183487e-01
-7.35903502e-01 3.89511526e-01 -5.98074317e-01 -5.62344551e-01
-4.75055546e-01 -1.12692177e+00 1.60750365e+00 1.03728168e-01
-5.86411238e-01 -6.73334062e-01 2.92292893e-01 4.46193039e-01
1.84511364e-01 6.32348299e-01 1.41558811e-01 -3.98662925e-01
-3.56788158e-01 -4.16362017e-01 9.09925625e-02 -3.28282975e-02
-1.68987766e-01 -4.22659606e-01 -7.90288448e-01 -5.72994590e-01
-5.11196136e-01 -6.01841211e-01 6.84019029e-01 2.94858217e-01
9.42507386e-01 -7.93752074e-02 -4.76231337e-01 6.91401601e-01
1.01295257e+00 -6.67443812e-01 5.10473967e-01 3.02909881e-01
9.32084680e-01 5.89495301e-01 6.56012952e-01 4.14468288e-01
6.07088327e-01 1.18970525e+00 4.94288206e-01 -2.11328447e-01
-1.90327749e-01 -4.11691964e-01 1.53802380e-01 4.02087927e-01
-6.70024812e-01 4.32107449e-02 -9.93671834e-01 2.22197488e-01
-1.95878422e+00 -6.72822595e-01 7.27883503e-02 2.26088047e+00
8.61500621e-01 2.99815416e-01 5.85843325e-01 1.46640405e-01
4.86064017e-01 7.28624910e-02 -4.72922206e-01 2.45470509e-01
4.14937764e-01 3.10619324e-01 4.69688177e-01 3.25087667e-01
-1.58965790e+00 9.33379352e-01 6.39442682e+00 2.92930931e-01
-8.92068148e-01 6.49762601e-02 2.02265069e-01 -1.22599967e-01
4.29788738e-01 -3.12286794e-01 -1.02284956e+00 1.80232033e-01
6.89453125e-01 4.67301041e-01 9.05670673e-02 7.82682717e-01
-1.16499580e-01 -1.22784689e-01 -1.41577148e+00 1.11108637e+00
1.44557536e-01 -8.62153649e-01 -2.62302458e-01 -1.08673602e-01
5.51083684e-01 -1.11442186e-01 -2.04325676e-01 5.42824805e-01
1.05087928e-01 -1.05716848e+00 8.43130827e-01 6.53364599e-01
6.56288207e-01 -6.52403176e-01 8.92074585e-01 4.91288394e-01
-1.38769639e+00 4.20778431e-02 -3.78316715e-02 -2.13125631e-01
3.69613856e-01 1.43033087e-01 -5.81306159e-01 6.15839720e-01
9.06898677e-01 6.54337227e-01 -9.63776112e-01 1.17697072e+00
-3.59908044e-01 6.71435520e-02 -4.85427856e-01 2.84958839e-01
-2.12043583e-01 3.54614496e-01 3.77627850e-01 1.00649285e+00
1.40098035e-01 -2.51627147e-01 6.64792120e-01 5.90169489e-01
-5.38103171e-02 -1.88249469e-01 -1.31535783e-01 3.10273498e-01
3.58546637e-02 1.23480546e+00 -6.44467771e-01 -2.46414453e-01
-1.84637368e-01 1.34101033e+00 4.61913824e-01 -2.26314832e-02
-9.87971485e-01 1.02245249e-01 6.26599014e-01 3.40512335e-01
1.28717542e-01 -4.00566339e-01 -1.19489767e-02 -1.21869814e+00
1.60442889e-01 -9.33740854e-01 4.00349230e-01 -5.56903064e-01
-1.06192756e+00 3.52978736e-01 2.36763790e-01 -1.10179675e+00
-7.26156831e-01 -6.61559343e-01 -2.55892098e-01 7.08370745e-01
-1.13446057e+00 -1.44723320e+00 -5.64170957e-01 5.33635199e-01
4.30569708e-01 2.93750048e-01 7.76438415e-01 4.37777251e-01
-5.48965156e-01 6.84893191e-01 -6.75823748e-01 3.83112878e-01
9.85449672e-01 -1.33283031e+00 6.75196528e-01 6.45098150e-01
-6.09440394e-02 5.66038311e-01 9.85600591e-01 -8.27615142e-01
-1.33580172e+00 -9.55662251e-01 9.49036598e-01 -9.38527584e-01
2.93356031e-01 -6.30314350e-01 -4.45863634e-01 9.51993465e-01
-1.37035564e-01 3.19316268e-01 2.43400052e-01 3.02514493e-01
-1.64882332e-01 7.15354756e-02 -9.82834935e-01 5.19929945e-01
1.39811683e+00 -1.80876568e-01 -7.57719696e-01 2.79204369e-01
4.53218281e-01 -1.07029819e+00 -9.17944670e-01 7.34748721e-01
9.68848526e-01 -7.62036979e-01 1.38860333e+00 -6.49268210e-01
3.80148649e-01 -2.89313406e-01 2.17816576e-01 -1.11143446e+00
-2.56267220e-01 -4.76294488e-01 -4.99511600e-01 7.55427420e-01
1.69824615e-01 -1.33107349e-01 1.27352738e+00 6.18090153e-01
5.85835278e-02 -7.72027016e-01 -9.72294629e-01 -6.26695037e-01
-4.08076271e-02 -1.43467516e-01 3.73160869e-01 3.76709014e-01
-2.63929486e-01 3.27968657e-01 -9.32077348e-01 1.70125365e-01
7.78387606e-01 -7.31500089e-02 1.28444695e+00 -1.45694208e+00
-6.02340698e-01 -6.66049421e-02 -7.21268356e-01 -1.15179574e+00
2.88536642e-02 -3.37272972e-01 4.12984431e-01 -1.61212409e+00
3.03844392e-01 6.35627732e-02 2.45571434e-02 6.97037101e-01
-3.79454315e-01 7.80493915e-01 4.43101227e-01 2.05924232e-02
-1.05153060e+00 4.90977794e-01 1.23086882e+00 4.33367044e-02
-1.42432433e-02 2.80736476e-01 -2.89455026e-01 9.67508137e-01
4.64727104e-01 -4.28437263e-01 -1.13418475e-01 -2.51046389e-01
4.31261659e-01 8.69114697e-02 8.31102073e-01 -1.59004426e+00
1.80526599e-01 2.49805570e-01 9.34571803e-01 -8.37389588e-01
6.98230982e-01 -8.63885403e-01 2.78817654e-01 7.60050595e-01
-1.65662646e-01 1.59132645e-01 1.31261036e-01 5.21719813e-01
-3.59386504e-02 2.15727285e-01 6.72144413e-01 -3.02898347e-01
-6.50665164e-01 4.90303367e-01 1.63038939e-01 6.74880445e-02
8.64167392e-01 -2.29120761e-01 4.53629578e-03 -3.13097417e-01
-1.15997219e+00 2.67771155e-01 4.42278296e-01 6.85213387e-01
4.00346369e-01 -1.35028803e+00 -7.02286303e-01 1.79226190e-01
1.45027846e-01 1.13338970e-01 1.16398081e-01 1.10092354e+00
-4.93421763e-01 2.64834076e-01 -4.04356480e-01 -7.73948610e-01
-1.33847356e+00 2.34968722e-01 5.80456436e-01 -6.49111688e-01
-6.91795886e-01 8.31644714e-01 -5.38170859e-02 -7.41091251e-01
4.43581849e-01 -3.52558911e-01 -1.88057542e-01 -2.03454107e-01
2.69450396e-01 3.69359165e-01 5.07331118e-02 -9.74334836e-01
-4.79929596e-01 7.04339266e-01 1.23514794e-01 -4.31536362e-02
1.30899286e+00 -4.46722284e-02 1.32994801e-01 1.87943667e-01
9.10264969e-01 -2.02747419e-01 -1.61861134e+00 -1.11377724e-01
1.85440760e-02 -3.61142159e-01 -4.86732870e-01 -9.37928140e-01
-1.00413954e+00 7.56474614e-01 7.16132164e-01 -4.67271835e-01
6.91315114e-01 1.46820664e-01 8.44748795e-01 3.77660125e-01
6.11600697e-01 -1.29576302e+00 3.88043374e-01 6.26842916e-01
1.04279876e+00 -1.34554255e+00 3.08480382e-01 -5.08776724e-01
-2.97626585e-01 9.28580344e-01 9.84670401e-01 -3.12089950e-01
3.98658305e-01 3.09027404e-01 -1.06850766e-01 -2.11227551e-01
-3.15825105e-01 -2.76143789e-01 8.31415117e-01 5.04484296e-01
7.36147106e-01 -1.09861992e-01 -2.84193009e-01 7.08575785e-01
-3.84838909e-01 2.00204685e-01 -6.35505244e-02 9.89189386e-01
-3.43115509e-01 -1.19913244e+00 -7.73013294e-01 2.78303683e-01
-5.84860206e-01 4.33424711e-01 -4.86085504e-01 1.04242015e+00
3.45844001e-01 4.38888401e-01 -1.36447206e-01 -5.02410114e-01
5.44083893e-01 1.75895691e-01 8.57670426e-01 -6.97427750e-01
-7.98743308e-01 1.30351216e-01 1.99302122e-01 -1.03436673e+00
-5.68184733e-01 -6.74220860e-01 -1.17656827e+00 -7.73071051e-02
-2.02879041e-01 -2.50341594e-01 5.51794648e-01 1.18285716e+00
9.62708145e-02 7.37250209e-01 -3.32696587e-01 -1.54994881e+00
-7.10107386e-01 -1.13009942e+00 -2.69967258e-01 7.54615247e-01
2.10275307e-01 -1.14139223e+00 1.99673310e-01 6.14631288e-02] | [7.025851249694824, -0.8275352120399475] |
5fe4bf47-6cd5-4bf0-bdc7-d329f4fa53c6 | xtab-cross-table-pretraining-for-tabular | 2305.06090 | null | https://arxiv.org/abs/2305.06090v1 | https://arxiv.org/pdf/2305.06090v1.pdf | XTab: Cross-table Pretraining for Tabular Transformers | The success of self-supervised learning in computer vision and natural language processing has motivated pretraining methods on tabular data. However, most existing tabular self-supervised learning models fail to leverage information across multiple data tables and cannot generalize to new tables. In this work, we introduce XTab, a framework for cross-table pretraining of tabular transformers on datasets from various domains. We address the challenge of inconsistent column types and quantities among tables by utilizing independent featurizers and using federated learning to pretrain the shared component. Tested on 84 tabular prediction tasks from the OpenML-AutoML Benchmark (AMLB), we show that (1) XTab consistently boosts the generalizability, learning speed, and performance of multiple tabular transformers, (2) by pretraining FT-Transformer via XTab, we achieve superior performance than other state-of-the-art tabular deep learning models on various tasks such as regression, binary, and multiclass classification. | ['Mahsa Shoaran', 'George Karypis', 'Mu Li', 'Nick Erickson', 'Xingjian Shi', 'Bingzhao Zhu'] | 2023-05-10 | null | null | null | null | ['automl'] | ['methodology'] | [-1.07875086e-01 1.94447055e-01 -7.10223556e-01 -8.62524569e-01
-1.10719919e+00 -8.12047601e-01 4.05008048e-01 5.05202353e-01
4.36360657e-01 1.17065573e+00 7.16752782e-02 -6.60009623e-01
-3.71136032e-02 -8.94155324e-01 -1.33126247e+00 -2.61826277e-01
5.99112324e-02 1.24891889e+00 -1.47091672e-01 -1.23976022e-01
-2.72976756e-01 4.79961112e-02 -1.40235043e+00 1.07666469e+00
9.43245709e-01 1.70977128e+00 -7.34739661e-01 1.67838503e-02
-5.26534975e-01 1.48194611e+00 -3.31902236e-01 -9.48863685e-01
3.46061826e-01 -1.36859849e-01 -7.32342601e-01 -5.74794784e-02
1.34477031e+00 -2.58486629e-01 -3.52773279e-01 6.83020830e-01
-1.41368523e-01 -2.27233738e-01 7.63775229e-01 -1.50553775e+00
-9.37418222e-01 1.42239809e+00 -3.37297976e-01 -1.09446563e-01
7.13887289e-02 -5.05836494e-02 1.63164091e+00 -7.09944487e-01
5.35567641e-01 1.52381480e+00 1.16561687e+00 3.15157354e-01
-1.59441340e+00 -1.02355027e+00 3.84432614e-01 2.48374537e-01
-9.11380529e-01 -3.83130729e-01 5.81290960e-01 -5.27317584e-01
1.04546773e+00 2.71613717e-01 3.14223170e-01 1.23556828e+00
7.28870869e-01 1.13097942e+00 1.10388553e+00 5.01308292e-02
1.16883516e-01 3.59268486e-01 2.02272713e-01 9.29862976e-01
8.33771050e-01 -2.68902004e-01 -9.27759409e-01 1.59204811e-01
2.66459465e-01 -1.71246469e-01 3.82161558e-01 -1.20187902e+00
-1.18579280e+00 7.86700130e-01 5.93771338e-01 -2.80876070e-01
1.04750693e-01 3.44154596e-01 8.65021348e-01 5.46188653e-01
4.92996365e-01 8.26849222e-01 -1.13593173e+00 2.45037839e-01
-6.08779550e-01 1.39999226e-01 9.32298064e-01 1.13602972e+00
7.63208151e-01 1.09840848e-01 -2.29224384e-01 9.04423833e-01
5.77749731e-03 8.99144650e-01 3.06758076e-01 -5.40133417e-01
1.07160044e+00 8.74129772e-01 -4.36013937e-01 -6.43939793e-01
-5.59001029e-01 -3.35177928e-01 -9.46132183e-01 -1.60606563e-01
4.28495407e-01 -8.49582721e-03 -1.09675694e+00 1.28321564e+00
9.89878476e-02 -8.02785635e-01 1.44034058e-01 3.53020012e-01
1.17898905e+00 5.82169771e-01 3.01788878e-02 1.55638698e-02
1.01258361e+00 -8.94721746e-01 -8.35260928e-01 -5.44569254e-01
7.91790903e-01 -4.05629843e-01 9.63359296e-01 9.19362307e-01
-9.11314547e-01 -4.76903379e-01 -1.23298705e+00 -4.44620907e-01
-9.54372883e-01 -7.62588009e-02 1.30306184e+00 8.21147978e-01
-6.27753496e-01 5.54396629e-01 -8.21740329e-01 -6.29945770e-02
8.13730776e-01 4.85404134e-01 -6.30553067e-01 -2.28059217e-01
-1.07954741e+00 9.01566088e-01 5.22835970e-01 -1.71885297e-01
-8.18725407e-01 -9.68146443e-01 -1.20803225e+00 2.93516815e-01
7.15409219e-01 -5.81293941e-01 1.27095592e+00 -8.03302526e-01
-1.08121121e+00 7.19416976e-01 3.28824610e-01 -9.45303440e-01
5.46678305e-01 -3.03995848e-01 -2.55192399e-01 -4.62196320e-01
3.54888618e-01 6.74089491e-01 5.91827393e-01 -1.25558698e+00
-6.18921995e-01 -4.99721617e-01 -3.21635067e-01 9.85895917e-02
-4.19233173e-01 -7.72069395e-01 -1.73824310e-01 -3.69854808e-01
2.15820044e-01 -5.78387916e-01 2.10881919e-01 -4.74363230e-02
-7.94662952e-01 -2.71298319e-01 6.82114184e-01 -4.07100916e-01
7.58190453e-01 -1.94905531e+00 -6.05588444e-02 1.83580909e-02
3.87609065e-01 -3.93494040e-01 1.90001369e-01 1.87980950e-01
-2.88745105e-01 -2.88864393e-02 -7.50733763e-02 -4.49699014e-01
3.84383053e-01 3.29764664e-01 -6.14148855e-01 2.85656691e-01
2.25577980e-01 1.10334647e+00 -5.73251903e-01 -6.71438456e-01
1.84257209e-01 -9.04651955e-02 -6.62189186e-01 -4.80608493e-02
-8.04793239e-01 -2.13240132e-01 -3.19470800e-02 1.18177414e+00
5.98139822e-01 -6.58807278e-01 5.85663736e-01 -3.82069021e-01
2.66511261e-01 9.97603297e-01 -9.91058409e-01 1.45275760e+00
-2.87400007e-01 5.59315503e-01 -4.54351813e-01 -1.06957555e+00
1.07896912e+00 -1.90850347e-01 7.35669374e-01 -1.08898616e+00
5.13082370e-02 1.47482693e-01 -2.27105141e-01 7.59565085e-02
3.28891069e-01 -1.18988186e-01 -5.17897308e-01 2.76843578e-01
5.54528713e-01 -2.73303270e-01 4.81095970e-01 1.43157974e-01
8.94062936e-01 2.07362860e-01 2.99821138e-01 -2.87406594e-01
3.70821238e-01 4.55572158e-01 3.44317704e-01 8.38310838e-01
2.80003041e-01 4.04672384e-01 9.25788283e-01 -9.72345889e-01
-9.48753834e-01 -1.31407940e+00 -5.03211439e-01 1.40550995e+00
-1.65157855e-01 -6.71553731e-01 -1.68379441e-01 -1.10331619e+00
9.58599687e-01 6.25698626e-01 -1.04872680e+00 -2.62317270e-01
-2.79229671e-01 -8.85421157e-01 2.20270693e-01 9.90435541e-01
5.27613938e-01 -8.61435413e-01 -4.47715772e-03 1.91129297e-01
-1.45480111e-02 -1.24729776e+00 -1.12988256e-01 1.26472616e+00
-1.12597764e+00 -1.12111998e+00 3.03652793e-01 -5.74330628e-01
2.71729529e-01 -3.15698951e-01 1.68357241e+00 -4.35431570e-01
-3.39427330e-02 1.34732664e-01 6.64236993e-02 -4.98467624e-01
-4.25805897e-01 5.77483356e-01 3.92474793e-02 -1.51133105e-01
2.28518575e-01 -1.13927670e-01 2.01620519e-01 2.09485278e-01
-2.87660778e-01 2.13119432e-01 4.64177370e-01 1.14743733e+00
7.42507935e-01 -4.77021931e-05 3.64441842e-01 -1.56850314e+00
2.29751378e-01 -3.52120519e-01 -1.01575625e+00 5.73303044e-01
-1.14151788e+00 6.52399123e-01 1.01078832e+00 -8.75097439e-02
-7.47380018e-01 7.39614666e-02 3.67282271e-01 -3.33000004e-01
2.46179760e-01 7.16742933e-01 -3.96564841e-01 3.11319441e-01
7.11241603e-01 5.95709421e-02 4.77475598e-02 -3.75955969e-01
4.28663701e-01 3.07168961e-01 6.43969715e-01 -8.58531773e-01
8.00587296e-01 3.80631536e-01 2.51797765e-01 1.19371004e-01
-1.41536570e+00 -3.56954820e-02 -7.75600433e-01 2.81243235e-01
5.87615669e-01 -1.08888102e+00 -8.60011756e-01 1.13935858e-01
-4.60503995e-01 -7.38875449e-01 -1.78525731e-01 -1.20937787e-01
-6.84621572e-01 -2.25552261e-01 -5.29570639e-01 -4.83597577e-01
-5.51545024e-01 -8.64889920e-01 1.08056486e+00 -3.45646977e-01
-1.60171241e-01 -1.08814955e+00 -7.52727091e-02 5.56176305e-01
8.84200633e-02 1.79377407e-01 1.33132792e+00 -8.93503547e-01
-8.35642993e-01 -1.26630545e-01 -3.23888272e-01 1.38434634e-01
5.02760291e-01 -1.37049615e-01 -8.54764640e-01 -1.41738132e-01
-7.05545962e-01 -1.09613812e+00 1.07989597e+00 2.96298295e-01
1.39943242e+00 -6.22829616e-01 -3.29644322e-01 1.07895350e+00
1.23168552e+00 1.37329936e-01 3.73483658e-01 7.19315112e-01
1.02692986e+00 4.06848073e-01 5.63876987e-01 3.20904911e-01
7.68926859e-01 1.51651248e-01 4.78923082e-01 -1.07027665e-01
3.80006135e-02 -5.42639315e-01 1.36478424e-01 3.47302735e-01
7.89488971e-01 -2.56656766e-01 -1.12459493e+00 2.97695816e-01
-1.98081219e+00 -5.97878814e-01 1.02710821e-01 1.94403112e+00
1.13019514e+00 6.82177126e-01 1.55008305e-02 1.64963588e-01
-6.23751655e-02 -6.36627600e-02 -1.01544583e+00 -3.73906523e-01
-3.75828385e-01 1.82495415e-01 9.29298580e-01 1.11973956e-01
-1.65445614e+00 1.04579127e+00 6.59491920e+00 3.43167186e-01
-1.15457988e+00 -2.87744552e-01 1.30405521e+00 -1.41375601e-01
-2.46137202e-01 -4.66695517e-01 -1.07058907e+00 -6.25989884e-02
9.53643560e-01 -1.80347022e-02 4.46889460e-01 1.22166669e+00
-8.15544903e-01 1.29535735e-01 -1.75917327e+00 1.04661787e+00
2.55597383e-01 -1.68137085e+00 5.04034221e-01 -2.62353271e-01
8.40664089e-01 -1.85407382e-02 6.07123315e-01 9.77710247e-01
8.48312616e-01 -1.41787815e+00 8.79134953e-01 1.29835913e-02
1.14049494e+00 -7.30091274e-01 8.13908994e-01 -3.14343631e-01
-1.03018177e+00 -3.26363266e-01 -3.59559298e-01 1.91665530e-01
-7.82191694e-01 2.20979840e-01 -1.10670245e+00 6.11259699e-01
1.22123265e+00 1.29825437e+00 -1.23805916e+00 5.41091681e-01
8.79637748e-02 4.74520862e-01 -1.14709549e-01 -4.30101119e-02
1.91238403e-01 1.97138369e-01 -3.91948313e-01 1.03765476e+00
-2.45313838e-01 -5.11410356e-01 4.69801724e-02 8.43571007e-01
-4.62734908e-01 6.73259050e-02 -5.72188079e-01 -1.20833896e-01
9.77917314e-02 1.09982419e+00 -4.81724024e-01 -5.99151015e-01
-4.66899455e-01 2.70055383e-01 6.13500357e-01 -4.30621430e-02
-8.29566717e-01 1.91443443e-01 5.53447485e-01 1.06425323e-01
6.96780443e-01 2.05739930e-01 -1.38729274e+00 -1.30288041e+00
-8.38778317e-02 -1.49784458e+00 1.09169078e+00 -8.09925795e-01
-1.68333614e+00 5.86423397e-01 1.33305088e-01 -9.70399618e-01
-2.93131799e-01 -1.26313162e+00 2.93322682e-01 4.02907282e-01
-1.19400811e+00 -1.32512021e+00 -2.60675013e-01 6.52248204e-01
3.51255178e-01 -7.12939441e-01 6.91497207e-01 3.32492031e-02
-5.50291061e-01 1.10837901e+00 3.69276136e-01 4.00897145e-01
1.15989542e+00 -1.84255099e+00 6.62262678e-01 4.06031430e-01
2.60306448e-01 6.23748064e-01 6.42353952e-01 -7.82803357e-01
-1.94036472e+00 -1.29966795e+00 6.78461790e-01 -1.06768239e+00
1.06134331e+00 -1.11935711e+00 -8.41760397e-01 1.32504511e+00
1.08390003e-01 1.23619296e-01 4.68454957e-01 8.84486318e-01
-1.20267761e+00 -1.07620001e+00 -1.05879986e+00 3.45975876e-01
7.71155536e-01 -3.98074806e-01 -4.18939829e-01 5.87886631e-01
5.88199377e-01 -7.56165326e-01 -1.04399812e+00 5.75630248e-01
7.99884796e-01 -6.80471778e-01 9.13645625e-01 -9.38038170e-01
9.68428969e-01 1.66481480e-01 -3.76419902e-01 -1.17178571e+00
-3.19404006e-01 -2.73077577e-01 -4.45899129e-01 1.03549159e+00
8.45159471e-01 -5.71934223e-01 1.09912765e+00 5.02175033e-01
-1.59178540e-01 -6.46985531e-01 -7.83241570e-01 -7.22719908e-01
5.51121354e-01 -2.08933830e-01 8.04976702e-01 1.02257812e+00
1.84795946e-01 7.11913288e-01 -2.66862690e-01 -1.87209591e-01
8.17962646e-01 4.91603404e-01 1.05982125e+00 -1.54116011e+00
-1.65574506e-01 -3.82101178e-01 -2.03554884e-01 -7.33214855e-01
2.85109013e-01 -1.09083748e+00 6.21339604e-02 -1.51769435e+00
4.10447091e-01 -6.63596332e-01 -4.24629211e-01 1.14427984e+00
6.95281522e-03 1.38733029e-01 2.43190099e-02 -1.69320911e-01
-1.07907522e+00 4.62058306e-01 1.13677812e+00 -1.01322114e+00
2.08719373e-01 -3.87246251e-01 -9.70731974e-01 8.86107758e-02
5.87806284e-01 -4.68555152e-01 -3.97510231e-01 -3.36378127e-01
5.36585689e-01 1.07029431e-01 -7.51185194e-02 -1.13430643e+00
8.67218599e-02 -1.48035407e-01 1.08014536e+00 -1.14673519e+00
2.12027475e-01 -8.29107106e-01 -1.33669868e-01 2.49023259e-01
-5.22936404e-01 3.36958230e-01 9.07893836e-01 3.86182785e-01
-6.05399795e-02 4.42072749e-01 4.59500432e-01 -1.31902426e-01
-3.96677196e-01 3.02830398e-01 -7.85198063e-02 1.51843280e-01
4.74498361e-01 1.42590642e-01 -7.30066240e-01 -1.14100739e-01
-3.46347034e-01 5.52836597e-01 1.95751756e-01 6.81243360e-01
2.71580428e-01 -1.23296452e+00 -5.08861363e-01 4.91682649e-01
5.00272274e-01 9.71962363e-02 -3.07304919e-01 3.16830605e-01
-5.46277225e-01 1.02486241e+00 -5.37513793e-01 -8.16121817e-01
-1.00660849e+00 9.87819195e-01 6.09511793e-01 -6.67887568e-01
-4.16809916e-01 8.41085196e-01 3.70341033e-01 -9.59239066e-01
6.05990052e-01 -7.60648131e-01 1.53674632e-01 1.09468311e-01
8.22309870e-03 -2.33643070e-01 6.98089004e-01 1.41992748e-01
-5.50783753e-01 -5.62642142e-03 -5.74975431e-01 3.15215737e-01
1.34312057e+00 3.58247638e-01 -1.92244187e-01 9.45746362e-01
8.55597079e-01 -2.96185881e-01 -1.13590789e+00 -1.31420970e-01
2.50888377e-01 1.25080854e-01 -1.60783976e-01 -1.35877395e+00
-1.10652280e+00 5.18216670e-01 1.83204114e-01 4.20290679e-02
8.14292610e-01 -1.42251238e-01 4.78808850e-01 1.03758430e+00
3.63636464e-01 -7.97862828e-01 1.23883359e-01 7.68355131e-01
7.78252244e-01 -1.69018590e+00 4.71024066e-01 -3.37537616e-01
-7.28963494e-01 1.31322551e+00 1.13696480e+00 1.43020496e-01
5.07179618e-01 6.03803277e-01 1.86581254e-01 -1.84977800e-01
-1.51161098e+00 2.82800257e-01 6.15008116e-01 5.04328489e-01
6.80470645e-01 5.03680035e-02 4.91134614e-01 5.88686228e-01
-7.91075349e-01 -2.32726440e-01 3.05029094e-01 6.89578950e-01
4.47072126e-02 -8.40929627e-01 -5.75611591e-01 1.04314542e+00
-1.73991606e-01 -4.20905143e-01 -7.44748831e-01 1.07148802e+00
-8.65635201e-02 6.75881207e-01 3.49189758e-01 -5.45802772e-01
2.03444004e-01 2.39825532e-01 6.82373047e-01 -5.61090708e-01
-7.93982267e-01 -2.31980518e-01 3.64786655e-01 -5.46072900e-01
3.61013710e-01 -9.77573395e-01 -8.30801010e-01 -4.30601567e-01
4.91579138e-02 -8.48236959e-03 3.26778531e-01 7.60975659e-01
-2.50375271e-02 8.38594139e-01 2.15959668e-01 1.60224587e-02
-7.57275462e-01 -9.81994331e-01 -4.76667404e-01 3.50660473e-01
6.16599679e-01 -9.35017526e-01 -1.35331094e-01 4.81001027e-02] | [9.641417503356934, 7.828150749206543] |
aad2a89b-33dc-4383-9501-849a7b8bc022 | g-rank-unsupervised-continuous-learn-to-rank | 2301.12530 | null | https://arxiv.org/abs/2301.12530v1 | https://arxiv.org/pdf/2301.12530v1.pdf | G-Rank: Unsupervised Continuous Learn-to-Rank for Edge Devices in a P2P Network | Ranking algorithms in traditional search engines are powered by enormous training data sets that are meticulously engineered and curated by a centralized entity. Decentralized peer-to-peer (p2p) networks such as torrenting applications and Web3 protocols deliberately eschew centralized databases and computational architectures when designing services and features. As such, robust search-and-rank algorithms designed for such domains must be engineered specifically for decentralized networks, and must be lightweight enough to operate on consumer-grade personal devices such as a smartphone or laptop computer. We introduce G-Rank, an unsupervised ranking algorithm designed exclusively for decentralized networks. We demonstrate that accurate, relevant ranking results can be achieved in fully decentralized networks without any centralized data aggregation, feature engineering, or model training. Furthermore, we show that such results are obtainable with minimal data preprocessing and computational overhead, and can still return highly relevant results even when a user's device is disconnected from the network. G-Rank is highly modular in design, is not limited to categorical data, and can be implemented in a variety of domains with minimal modification. The results herein show that unsupervised ranking models designed for decentralized p2p networks are not only viable, but worthy of further research. | ['Johan Pouwelse', 'Andrew Gold'] | 2023-01-29 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [-3.10869604e-01 6.77335486e-02 -7.97255039e-01 -4.47952718e-01
-6.73244953e-01 -8.13341677e-01 6.21426404e-01 1.24222413e-01
-1.28870070e-01 7.06017911e-01 1.60125718e-02 -4.99357462e-01
-8.77447844e-01 -1.03038979e+00 -2.95112133e-01 -4.64684516e-01
-6.31576300e-01 1.23846292e+00 6.36548698e-01 -2.00625405e-01
4.39643525e-02 6.76652968e-01 -1.83384573e+00 1.69433326e-01
5.28184891e-01 1.25648046e+00 -5.59030548e-02 8.78648520e-01
6.09496646e-02 8.56289804e-01 -5.78336895e-01 -3.74460995e-01
5.35870135e-01 2.80806124e-01 -6.93357766e-01 -4.86366600e-01
3.13850999e-01 -6.28785074e-01 -5.88274479e-01 7.91186869e-01
6.85635448e-01 -2.20869094e-01 4.57754821e-01 -1.73840714e+00
-3.43826592e-01 9.22776937e-01 -2.76708473e-02 -1.12505458e-01
1.27521425e-01 -2.21451998e-01 1.47320902e+00 -5.93838692e-01
8.20279539e-01 8.58356655e-01 5.92687190e-01 4.29521203e-01
-1.13961506e+00 -4.85957831e-01 -1.75067797e-01 3.02890986e-01
-1.23653686e+00 -6.15149021e-01 6.62115514e-01 9.76748578e-03
5.19365489e-01 6.63794756e-01 5.05224586e-01 7.37139106e-01
-2.83235073e-01 6.17278337e-01 5.18645287e-01 -6.08352907e-02
4.11643207e-01 1.42583281e-01 -1.79288968e-01 3.76528263e-01
4.96616513e-01 -9.57247391e-02 -5.94558835e-01 -7.98547208e-01
4.89511251e-01 1.88265592e-01 1.59619257e-01 -9.34972465e-01
-1.28829157e+00 6.09113574e-01 2.11913198e-01 1.43079847e-01
-4.97834921e-01 4.18647408e-01 5.72712958e-01 7.00159132e-01
2.84937143e-01 3.72557044e-01 -7.78993428e-01 -4.08886999e-01
-1.15249348e+00 1.69760823e-01 1.52501965e+00 1.25928235e+00
7.71708012e-01 -3.71407926e-01 7.82002509e-03 8.36481988e-01
1.00647822e-01 5.17818570e-01 1.71575993e-01 -1.64139473e+00
2.94653356e-01 5.90086401e-01 1.50470629e-01 -1.24517787e+00
-2.79279649e-01 -1.39979362e-01 -7.13096440e-01 -2.15399593e-01
1.40246496e-01 -1.43186525e-01 -2.79825270e-01 1.46865213e+00
3.66577387e-01 -4.60940540e-01 5.44636995e-02 7.00289726e-01
8.30553532e-01 3.74993473e-01 -1.31287366e-01 -1.99559733e-01
1.21754479e+00 -7.93775320e-01 -3.06085553e-02 2.64972538e-01
6.87328219e-01 -2.78260350e-01 8.25167954e-01 5.13870120e-01
-1.24414384e+00 2.52741098e-01 -7.00072527e-01 4.48986739e-02
-4.92105037e-01 -2.32038885e-01 9.07169402e-01 6.96368515e-01
-1.51786554e+00 7.41872907e-01 -5.99296331e-01 -8.65201116e-01
3.70312542e-01 8.80146325e-01 -4.15157944e-01 2.32716933e-01
-1.10896599e+00 4.84115034e-01 1.85829535e-01 -3.39893967e-01
-1.08807230e+00 -2.91691363e-01 1.17383115e-01 4.88942862e-01
4.72143680e-01 -8.88692617e-01 1.31722116e+00 -7.38629162e-01
-1.47617078e+00 5.18194616e-01 -2.23488901e-02 -5.38988709e-01
4.81009275e-01 3.04582864e-01 -4.49641317e-01 6.43179357e-01
1.72639593e-01 3.84667158e-01 4.41237420e-01 -1.24556947e+00
-8.31266344e-01 -1.79893166e-01 7.79896975e-02 2.61349112e-01
-1.03244996e+00 3.19374949e-01 -6.44965887e-01 -5.73901944e-02
2.48543650e-01 -7.34899998e-01 -4.04501736e-01 1.79694161e-01
-3.14164817e-01 -4.10711050e-01 1.21630943e+00 -2.24496573e-01
1.12207139e+00 -1.78642380e+00 -4.40546811e-01 1.03493702e+00
6.92827761e-01 7.18306601e-02 -8.54722187e-02 6.67392135e-01
5.35487413e-01 6.34186387e-01 3.66601974e-01 9.07732248e-02
1.76385894e-01 3.87907773e-01 4.15253416e-02 2.48190954e-01
-5.25130153e-01 5.55410385e-01 -1.11882508e+00 -8.53281975e-01
-2.73853064e-01 3.45239230e-02 -8.03065717e-01 -2.17832655e-01
-3.26569587e-01 1.46414787e-01 -8.15898418e-01 1.04934394e+00
2.39372164e-01 -7.38142133e-01 7.76849031e-01 1.21037155e-01
5.64312972e-02 5.22910655e-01 -9.82253611e-01 1.33981287e+00
-3.43486995e-01 6.04627728e-01 7.19566464e-01 -1.19204473e+00
8.30113471e-01 5.30769289e-01 1.23066640e+00 -5.19380689e-01
-1.78170372e-02 8.14196885e-01 -2.65390247e-01 -1.25573173e-01
6.60292268e-01 5.74200749e-01 -7.36452918e-03 8.21410954e-01
-1.39005587e-01 4.39903177e-02 5.68145871e-01 5.69572806e-01
1.82732046e+00 -8.57901990e-01 -2.05166936e-01 -3.52295727e-01
2.26014569e-01 2.34207332e-01 7.59064972e-01 1.10625088e+00
-9.78493169e-02 3.58166605e-01 7.13916242e-01 -4.41215396e-01
-1.29868102e+00 -1.08637452e+00 -8.42914358e-02 1.56380022e+00
2.94696450e-01 -6.73191309e-01 -2.96246439e-01 -7.80524254e-01
2.57052034e-01 -1.57311916e-01 2.33635530e-01 2.94524074e-01
-2.39166915e-01 -2.17755407e-01 7.37828076e-01 1.95687905e-01
5.77138722e-01 -9.43329990e-01 7.44255558e-02 4.02755290e-01
-1.82435974e-01 -1.10930312e+00 -1.16629362e-01 1.61130130e-01
-1.09184110e+00 -1.45264852e+00 -2.82480836e-01 -8.56447518e-01
7.22905874e-01 4.89788532e-01 1.43786871e+00 3.46196651e-01
2.29802981e-01 7.19019115e-01 -3.22192222e-01 3.64017226e-02
-3.00356507e-01 5.54539680e-01 5.37683904e-01 -4.47365135e-01
5.04322767e-01 -1.19850540e+00 -9.83403683e-01 1.02559686e+00
-6.50076747e-01 -6.64225221e-01 6.68577254e-01 5.99152505e-01
4.35723424e-01 4.83540535e-01 7.09364772e-01 -8.99642885e-01
8.25907588e-01 -8.16054165e-01 -5.95825315e-01 2.20833972e-01
-1.17414272e+00 -2.19933182e-01 8.18112433e-01 -2.62840539e-01
-4.02775437e-01 -1.85155049e-01 4.24369544e-01 -1.48579523e-01
1.61795676e-01 4.71430779e-01 -1.64661080e-01 -1.69852227e-01
9.74308789e-01 1.30911572e-02 2.21103579e-01 -5.43050706e-01
5.54056883e-01 1.19229960e+00 4.27201360e-01 -1.07099175e+00
9.78071332e-01 7.79659033e-01 3.91439609e-02 -7.12220967e-01
7.89999738e-02 -6.15914881e-01 -9.27313790e-02 -4.11522314e-02
-1.63313095e-02 -8.75354826e-01 -1.10836756e+00 8.84274915e-02
-7.92992890e-01 -1.97928160e-01 -4.01628256e-01 1.52575225e-01
-5.20455539e-01 3.98412555e-01 -5.38266301e-01 -4.29382354e-01
-6.71974659e-01 -7.06664681e-01 3.99223268e-01 1.49480268e-01
-2.41958454e-01 -8.47418904e-01 -6.45534247e-02 2.75751412e-01
7.82026172e-01 -3.06502610e-01 7.61946976e-01 -1.24667001e+00
-8.93480837e-01 -5.86925805e-01 -5.97956240e-01 4.30597901e-01
-1.40789412e-02 4.38213199e-01 -5.91265619e-01 -2.99816519e-01
-7.03367293e-01 -2.43152753e-01 2.48700261e-01 3.88309285e-02
1.32257187e+00 -9.27219450e-01 -5.11477292e-01 3.29048246e-01
1.02943301e+00 -4.90317121e-02 2.59794921e-01 4.23616737e-01
4.79314834e-01 7.39182234e-01 2.20124558e-01 5.81520498e-01
6.08231902e-01 6.41616285e-01 5.28766692e-01 1.99736997e-01
1.72314003e-01 -3.75909090e-01 1.46899462e-01 8.99229228e-01
-2.59195179e-01 -2.10325196e-01 -9.68877137e-01 9.27204072e-01
-2.03254318e+00 -1.04278016e+00 7.24410787e-02 2.27743959e+00
9.81806695e-01 -7.36807613e-03 3.50570679e-01 2.95179300e-02
7.85589814e-01 -3.89998034e-02 -6.79850459e-01 -3.23702425e-01
-1.54403925e-01 -8.85605719e-03 1.01561427e+00 -9.74524394e-02
-7.09594190e-01 7.19499111e-01 6.98283005e+00 7.79014707e-01
-8.08582008e-01 1.58894882e-01 3.93850923e-01 -3.02932918e-01
-5.14188409e-01 4.80665803e-01 -4.34863180e-01 5.02575278e-01
1.21285820e+00 -5.05900323e-01 6.95316434e-01 1.53392148e+00
3.74799848e-01 3.15067858e-01 -1.25843012e+00 9.16381657e-01
-6.92959666e-01 -1.41787827e+00 -1.68186240e-02 4.46479708e-01
7.02987969e-01 4.94940132e-01 -5.36643080e-02 4.79399487e-02
8.93974900e-01 -7.85829604e-01 3.28555346e-01 1.78798825e-01
8.14569652e-01 -5.11335909e-01 5.39049029e-01 2.38077447e-01
-1.01295733e+00 -4.90691781e-01 -5.56592703e-01 1.02652870e-01
-1.44099668e-01 9.00100529e-01 -6.26709342e-01 2.08599284e-01
1.16561997e+00 6.61934435e-01 -2.88013279e-01 1.32467258e+00
2.61030439e-02 5.34228265e-01 -9.83802676e-01 -4.23207670e-01
-1.22423127e-01 1.81695700e-01 7.05537260e-01 7.96112597e-01
4.98099446e-01 -2.44024575e-01 1.07915223e-01 1.91999644e-01
-7.15910256e-01 1.57736272e-01 -8.06747854e-01 -1.62573233e-01
1.20350063e+00 1.45563555e+00 -5.00761092e-01 -4.74689871e-01
-3.33711088e-01 3.64067674e-01 8.28798562e-02 4.17662829e-01
-2.28779614e-01 -6.84042275e-01 8.81909192e-01 5.94647050e-01
1.50108010e-01 -9.61088464e-02 -1.61966011e-01 -1.02784359e+00
1.84440494e-01 -1.01093602e+00 5.29156089e-01 -4.58086669e-01
-1.54993737e+00 2.70424485e-01 -1.12301230e-01 -1.35502875e+00
-4.28621292e-01 -4.74341989e-01 -5.69476545e-01 2.06519097e-01
-1.31994474e+00 -6.97228789e-01 -3.06947976e-01 9.66843188e-01
-3.14967424e-01 -5.32810330e-01 7.80252874e-01 6.15503728e-01
-1.61887720e-01 5.74703455e-01 6.94481850e-01 1.27974167e-01
6.69981658e-01 -1.09518588e+00 1.43441454e-01 3.35988522e-01
1.27569348e-01 7.60121107e-01 4.68241751e-01 -2.72138357e-01
-1.63618922e+00 -7.85516560e-01 1.08433986e+00 -4.70525712e-01
8.53663146e-01 -2.62661219e-01 -3.05174679e-01 1.83250144e-01
-3.14371198e-01 2.97086596e-01 6.28408968e-01 5.08385360e-01
-6.55747354e-01 -8.35002065e-01 -1.37317073e+00 6.33285463e-01
1.24941242e+00 -7.98528194e-01 1.07734822e-01 7.31010020e-01
4.72622901e-01 1.90902159e-01 -1.15440738e+00 3.63491505e-01
7.56628633e-01 -9.99253750e-01 1.07400572e+00 -6.05126977e-01
6.53719679e-02 -2.12308899e-01 -2.66179383e-01 -7.21461535e-01
-1.04970537e-01 -1.21660459e+00 -6.33754611e-01 9.83349323e-01
6.21397555e-01 -8.73958707e-01 1.50428748e+00 1.00257421e+00
3.29565048e-01 -3.78349036e-01 -1.00918055e+00 -9.90958393e-01
-1.89934686e-01 -3.07159543e-01 9.30034935e-01 7.39430070e-01
4.81127828e-01 -6.64225128e-03 -1.12313829e-01 1.53704032e-01
9.09495234e-01 -1.05845749e-01 1.13985276e+00 -1.87571025e+00
-1.92205116e-01 -6.27277315e-01 -4.65092033e-01 -1.26485085e+00
-1.89394221e-01 -8.23850930e-01 -1.66827649e-01 -1.65854311e+00
1.92334145e-01 -1.26328874e+00 -5.28004467e-01 7.24179029e-01
7.12862015e-01 1.47871628e-01 -2.07691878e-01 1.03345537e+00
-1.07569683e+00 1.40575990e-01 8.43801677e-01 -2.04518586e-01
-2.49465331e-01 3.26949954e-01 -1.05238485e+00 4.57456708e-01
9.52351332e-01 -6.32573247e-01 -6.63029850e-01 -2.56542414e-01
7.30492473e-01 2.20464885e-01 2.51322776e-01 -8.43921244e-01
9.39783573e-01 -8.03599432e-02 -2.60081530e-01 -3.81284773e-01
-1.23619493e-02 -1.24322104e+00 2.23311484e-02 2.94768848e-02
-9.34127644e-02 -4.13805544e-02 -5.97840309e-01 5.69425404e-01
-3.12558800e-01 6.39093593e-02 1.88497201e-01 -1.49422124e-01
-4.60343778e-01 6.03670180e-01 -2.42162108e-01 -4.45530973e-02
8.44281733e-01 -3.89452994e-01 -7.60562897e-01 -8.66166830e-01
-5.82699776e-01 6.69501781e-01 7.09921300e-01 1.78516582e-01
3.37497354e-01 -1.21889758e+00 -4.88166094e-01 -1.84845716e-01
9.52180922e-02 7.35879689e-02 -2.90356547e-01 6.59552455e-01
-6.83989108e-01 4.06370759e-01 1.39392659e-01 -5.00307262e-01
-1.13323307e+00 1.66804999e-01 3.86412032e-02 -3.72928649e-01
-4.29157764e-01 2.00859681e-01 -5.71733177e-01 -8.44684780e-01
4.78541166e-01 2.07767233e-01 1.75003275e-01 1.92719743e-01
2.15972394e-01 6.65008187e-01 1.23284146e-01 -1.00143634e-01
-4.70195025e-01 1.62225455e-01 -1.97926182e-02 -1.30736291e-01
1.40879345e+00 -1.30985647e-01 -5.37538707e-01 -3.21105450e-01
1.17727399e+00 1.40581634e-02 -8.78051758e-01 -3.20693612e-01
1.74752027e-01 -2.40908489e-01 -6.82897121e-02 -5.60482979e-01
-1.33723342e+00 1.85185760e-01 4.57962677e-02 7.92448819e-01
8.89990330e-01 6.57064095e-02 7.67343640e-01 1.15153611e+00
9.75406051e-01 -1.43566024e+00 -6.43575191e-02 3.92540693e-01
1.29785329e-01 -9.36486781e-01 2.49302849e-01 -5.82541153e-02
-2.46014401e-01 9.54632699e-01 1.49926364e-01 1.72962129e-01
9.82751846e-01 -6.21360913e-02 -1.79316923e-02 -4.70191360e-01
-1.13892293e+00 -5.82724158e-03 -1.80240452e-01 8.40939641e-01
-5.02650551e-02 1.90217514e-02 -3.64094734e-01 4.59680378e-01
-2.32326999e-01 -1.37061691e-02 3.70226473e-01 1.09015119e+00
-5.28089106e-01 -1.49212050e+00 4.28015068e-02 8.62564683e-01
-5.62720835e-01 -2.14743674e-01 -5.26280940e-01 5.12230217e-01
-4.38481003e-01 1.23177397e+00 -2.97722947e-02 -5.21604359e-01
2.03395188e-01 -4.45279218e-02 -3.93757313e-01 -4.95817721e-01
-5.34604788e-01 -5.08470953e-01 5.68158329e-01 -8.43404293e-01
-2.72813916e-01 -3.98233801e-01 -1.01122808e+00 -1.00971925e+00
-2.30841562e-01 8.80360901e-01 9.69232142e-01 2.26917595e-01
1.04283583e+00 -3.08919042e-01 1.12156987e+00 -6.28468633e-01
-6.11593127e-01 -5.18451571e-01 -8.00416291e-01 -7.16443509e-02
-4.77977917e-02 -2.52252847e-01 -6.96881950e-01 -3.52899611e-01] | [6.042801380157471, 6.202227592468262] |
ea8ab3b3-3816-4fe0-8b75-954324663368 | analysis-of-the-influence-of-final-resolution | 2307.00388 | null | https://arxiv.org/abs/2307.00388v1 | https://arxiv.org/pdf/2307.00388v1.pdf | Analysis of the influence of final resolution on ADC accuracy | This work is devoted to the study of the influence of quantization noise on the spectral characteristics of a digital signal and the assessment of spectrum measurement errors that arise due to the quantization noise of an analog-to-digital converter. To achieve more accurate and reliable measurements of the spectrum, an error assessment was carried out, which allows taking into account the impact of quantization noise on the spectral data. This is important for obtaining more accurate results and ensuring high-quality measurements of the spectral components of the vibration signal. In addition, further research is aimed at developing methods for estimating spectrum measurement errors taking into account other possible sources of errors and contributing to the development of compensation algorithms to reduce the impact of quantization noise. | ['Anzhelika Stakhova'] | 2023-07-01 | null | null | null | null | ['quantization'] | ['methodology'] | [ 5.85587144e-01 -5.62998474e-01 1.49179429e-01 -9.82672814e-03
-5.21902621e-01 -1.10467657e-01 1.27450258e-01 6.76625371e-01
-3.02724242e-01 3.74400944e-01 1.07495993e-01 -1.79066375e-01
-2.37246841e-01 -9.93228555e-01 -2.81275958e-01 -4.90740836e-01
8.14496428e-02 7.63312355e-02 3.31576109e-01 1.18648447e-01
2.39327341e-01 6.18973851e-01 -1.62862968e+00 -4.59721237e-01
6.00647867e-01 8.86044145e-01 1.73527077e-01 5.79459429e-01
1.10860839e-01 3.21610361e-01 -1.25888836e+00 1.50850192e-01
-1.13566756e-01 -7.06209183e-01 -2.57140398e-01 4.52061445e-01
-6.24609888e-02 -2.70893812e-01 2.79492170e-01 1.39488161e+00
4.12911177e-01 2.90079236e-01 8.26348424e-01 -6.47241771e-01
-7.29918778e-02 4.53136772e-01 1.85686395e-01 9.08829346e-02
5.57458997e-01 -7.49393925e-02 3.45508754e-01 -3.45092237e-01
3.64654392e-01 8.19789171e-01 8.33518356e-02 -1.75099447e-01
-1.40173495e+00 -5.06004035e-01 -6.33093536e-01 5.93170285e-01
-1.50189209e+00 -1.28955737e-01 1.02958167e+00 -6.03367686e-01
4.20661926e-01 1.27979726e-01 5.34420609e-01 5.09179831e-01
4.11355078e-01 -2.03530103e-01 8.80410373e-01 -9.01683271e-01
4.02575433e-01 1.90787315e-01 2.40768328e-01 8.80348012e-02
4.23775911e-01 4.60611172e-02 -1.43365279e-01 1.96129140e-02
8.01304579e-01 -3.95639241e-01 -6.06495440e-01 -2.40488529e-01
-6.47821486e-01 6.17743969e-01 2.76939720e-01 9.04140651e-01
-5.85009634e-01 -1.35853626e-02 4.09912974e-01 7.55530968e-02
-5.58764413e-02 4.00361866e-01 1.22341506e-01 -4.87844706e-01
-7.13270366e-01 -2.52371170e-02 5.52385032e-01 2.89802670e-01
4.71591800e-01 5.94390869e-01 3.00739974e-01 3.97103965e-01
1.26115710e-01 7.26804078e-01 5.03428519e-01 -1.04000771e+00
2.89508671e-01 5.91526866e-01 2.67171383e-01 -1.24771333e+00
-3.33068252e-01 -4.33684349e-01 -5.04907370e-01 4.55160052e-01
4.66771275e-01 -2.60628946e-02 -3.88872445e-01 1.15614414e+00
1.36729300e-01 -2.35386595e-01 -7.98653904e-03 8.63445818e-01
2.37563010e-02 5.62389374e-01 -1.96479455e-01 -4.86029267e-01
1.22171867e+00 1.11344807e-01 -1.25542104e+00 2.03250557e-01
2.26814643e-01 -9.86942053e-01 1.11637127e+00 4.78506744e-01
-4.98520017e-01 -7.28472471e-01 -1.17085934e+00 3.13656598e-01
-1.53033912e-01 3.36103141e-01 -2.53540486e-01 8.64895463e-01
-2.55625397e-01 7.10903287e-01 -7.44911253e-01 1.20001391e-01
-6.00196004e-01 1.43766209e-01 -1.29935205e-01 2.96696454e-01
-1.16441643e+00 1.15112197e+00 2.44167164e-01 3.06519866e-01
-1.15683511e-01 -3.75062793e-01 -6.35304093e-01 2.64754176e-01
3.31694901e-01 1.13415875e-01 1.18840039e+00 -9.11779404e-01
-1.31146157e+00 -1.10586565e-02 -5.09524792e-02 -4.26212817e-01
2.68035144e-01 -4.56279628e-02 -7.70663798e-01 2.47174263e-01
-1.78743541e-01 -6.12306118e-01 7.44332075e-01 -5.55443883e-01
-1.48257501e-02 -5.01271784e-01 -6.24645770e-01 -2.32898742e-01
-1.82121545e-01 -2.27235511e-01 1.51387393e-01 -4.87202317e-01
1.95561677e-01 -6.73643053e-01 9.51118693e-02 -4.02739227e-01
7.69616142e-02 2.18002558e-01 6.66970313e-01 -8.26851726e-01
1.30859900e+00 -2.11891413e+00 4.37402911e-02 5.70329666e-01
-5.42324305e-01 6.19786680e-01 2.68030316e-01 6.74993336e-01
7.74077699e-02 -3.21203589e-01 -6.32704273e-02 2.61916041e-01
-2.97537148e-01 -1.80040643e-01 6.75562844e-02 3.13157380e-01
1.96046323e-01 -2.58986682e-01 -4.61468786e-01 8.29099715e-02
7.48169839e-01 6.02312624e-01 -2.64303740e-02 7.49043375e-02
-3.25208679e-02 2.21873552e-01 -2.73598701e-01 1.32067218e-01
1.49444595e-01 2.99549848e-01 2.19877452e-01 -5.04113972e-01
-2.53610730e-01 4.49171185e-01 -1.60229945e+00 6.40165329e-01
-5.21854043e-01 8.59490931e-01 1.79765835e-01 -7.53679097e-01
1.33724129e+00 5.24730802e-01 3.17101866e-01 -9.58832085e-01
6.61388338e-01 5.24170041e-01 2.90167015e-02 -4.15876299e-01
5.54348111e-01 -2.20915854e-01 2.52859414e-01 -3.68241929e-02
-1.84905738e-01 -4.20259356e-01 3.33645672e-01 -2.67725468e-01
1.77823901e-01 -2.15721518e-01 3.38031739e-01 -7.79812858e-02
7.07119703e-01 -3.50781262e-01 3.15013796e-01 -1.18397303e-01
-1.32555768e-01 5.20960167e-02 4.01983827e-01 3.26648951e-02
-9.05814648e-01 -3.45678300e-01 -2.38548502e-01 6.11916818e-02
1.27453610e-01 -2.15258405e-01 -8.15837324e-01 3.56346995e-01
9.00048465e-02 8.98359239e-01 -1.11738637e-01 -4.79660362e-01
-2.90938437e-01 -4.50979173e-01 2.29173098e-02 4.06358987e-01
2.05607563e-01 -5.74008048e-01 -6.80795729e-01 8.05227101e-01
-4.15066123e-01 -9.22441542e-01 -7.36442357e-02 1.02647752e-01
-1.05764365e+00 -1.37113380e+00 -4.19789642e-01 -2.71512151e-01
3.92770886e-01 -1.13784246e-01 3.39554280e-01 5.85977323e-02
-3.13772529e-01 4.56194818e-01 -3.66226047e-01 -5.77678800e-01
-9.24055278e-01 -2.15982661e-01 -8.02706704e-02 8.17814283e-03
2.60905445e-01 -2.91396350e-01 -1.78811014e-01 4.39420998e-01
-8.44934583e-01 -6.21299028e-01 3.31015199e-01 2.02316001e-01
4.15916353e-01 6.58219278e-01 5.04984915e-01 -2.64582098e-01
9.80485320e-01 7.95916393e-02 -1.30677307e+00 -2.80056506e-01
-7.57620394e-01 1.93849578e-01 9.49689806e-01 -2.50407994e-01
-7.09569633e-01 -6.47816658e-02 -3.29135269e-01 -1.45154595e-01
-1.16698407e-01 3.90194029e-01 -4.58328873e-02 -1.57122165e-01
8.38345587e-01 1.96907014e-01 5.04149556e-01 -5.29656053e-01
-3.34440291e-01 9.66581106e-01 6.45723760e-01 3.44330532e-04
3.91436726e-01 -3.03705573e-01 4.19283360e-01 -1.49065566e+00
-3.95792238e-02 -6.53058231e-01 -1.87249750e-01 -3.93731177e-01
6.07883751e-01 -4.95926887e-01 -5.23549736e-01 7.20014334e-01
-8.77290130e-01 -4.96843755e-02 1.14557318e-01 8.63039255e-01
-3.67517740e-01 6.45785689e-01 -5.09443402e-01 -1.15151501e+00
-2.70878285e-01 -1.34650993e+00 3.31841767e-01 2.06262961e-01
-4.02630121e-01 -8.28727186e-01 -1.29687577e-01 3.26605022e-01
4.17854965e-01 2.37455055e-01 7.46531367e-01 6.57583177e-02
-3.88162255e-01 -3.68732810e-01 3.35094601e-01 7.07671881e-01
1.44158363e-01 3.11464727e-01 -4.66257125e-01 -1.08134277e-01
4.84192967e-01 1.76246017e-01 2.13009343e-01 2.52790511e-01
3.76320124e-01 1.08912274e-01 1.25347495e-01 -2.52679825e-01
1.76469302e+00 4.26149428e-01 7.87039816e-01 1.87622681e-01
3.85193125e-04 3.80846441e-01 8.44234765e-01 2.62039781e-01
-2.00167388e-01 1.09568477e+00 4.40529108e-01 1.84605032e-01
-1.46643639e-01 5.85918548e-03 7.58232474e-02 5.74649453e-01
-1.06822670e-01 -1.32264242e-01 -6.63395107e-01 4.19673353e-01
-1.01795793e+00 -7.33512521e-01 -7.89024711e-01 2.45679760e+00
3.85751247e-01 1.44549817e-01 2.49716744e-01 1.36982393e+00
8.91069412e-01 -1.86630324e-01 -1.76143143e-02 -6.29678667e-01
3.60091507e-01 5.28388061e-02 5.52415192e-01 8.20335567e-01
-6.83681488e-01 -9.66706127e-02 6.52587938e+00 4.49208707e-01
-1.57473719e+00 -6.64357305e-01 -1.07954063e-01 3.86582822e-01
-1.17920812e-04 -4.58181947e-01 -5.46983242e-01 7.22648203e-01
1.13775051e+00 -1.94638401e-01 5.68285108e-01 6.19260669e-01
7.06241190e-01 -6.25075221e-01 -1.63249522e-01 4.94218707e-01
-6.75040558e-02 -8.22363555e-01 -2.40135461e-01 7.06532747e-02
3.39690417e-01 -5.89437485e-01 -2.33824417e-01 -3.36251020e-01
-6.17464304e-01 -2.43547127e-01 7.65653789e-01 5.36871254e-01
4.17296469e-01 -9.34585333e-01 1.17226481e+00 3.04034889e-01
-1.06761467e+00 -6.84485435e-02 -2.67086208e-01 -4.16244417e-01
2.31706932e-01 1.08861697e+00 -1.00170159e+00 5.42151034e-01
3.74338329e-02 -1.57601759e-01 -1.37980863e-01 1.02876747e+00
-4.21070248e-01 7.72053659e-01 -2.04696119e-01 -5.14276624e-01
-2.96724349e-01 -3.56160998e-01 4.23934221e-01 6.59140527e-01
4.15834725e-01 5.14173321e-02 -1.51657030e-01 8.88990700e-01
3.58379871e-01 2.82532275e-01 -1.63573604e-02 -3.75268042e-01
7.16835320e-01 6.92686439e-01 -5.15805364e-01 -4.52664867e-02
-1.87590957e-01 3.88089925e-01 -2.55344123e-01 1.85975373e-01
-4.88890022e-01 -6.70391023e-01 4.27389950e-01 6.20337725e-01
2.54698873e-01 -4.38577950e-01 2.67956760e-02 -3.93155038e-01
1.50460586e-01 -5.63198686e-01 -3.42256278e-02 -4.62432593e-01
-5.69232702e-01 1.75353572e-01 -1.88462764e-01 -1.18033373e+00
-8.27744186e-01 -5.74867308e-01 -4.90335077e-01 1.09323335e+00
-9.68482673e-01 -1.77784622e-01 -1.15988545e-01 3.13889772e-01
4.07080725e-02 2.55069762e-01 1.01239645e+00 3.76661837e-01
-1.12162881e-01 -3.21288072e-02 2.29145944e-01 1.41643612e-02
4.50368822e-01 -8.59407067e-01 -1.84941009e-01 9.16190743e-01
-2.58060843e-01 7.27526993e-02 1.04323053e+00 -5.11729360e-01
-1.14031494e+00 -6.08992934e-01 9.84853983e-01 2.56436527e-01
5.01168370e-01 7.83779845e-02 -1.03613985e+00 -1.92508940e-02
-4.05060679e-01 -5.06428599e-01 4.98481035e-01 -2.35722154e-01
3.74800086e-01 -3.90225232e-01 -1.05838704e+00 9.41702947e-02
-1.32610425e-01 -7.76304007e-01 -7.54830539e-01 -4.36221927e-01
3.31472963e-01 -2.56889880e-01 -1.12958658e+00 7.78901502e-02
3.36083382e-01 -1.05881143e+00 6.78934515e-01 6.76712394e-01
-8.13264176e-02 -5.25507569e-01 2.18631908e-01 -1.14200509e+00
-1.24457337e-01 -1.38677642e-01 4.64063555e-01 9.88631666e-01
5.95470183e-02 -6.15434170e-01 3.79451603e-01 3.92957687e-01
2.47798115e-01 -1.46015868e-01 -6.99741542e-01 -7.68277407e-01
-4.46936339e-01 -5.59914827e-01 2.39617646e-01 3.92330408e-01
2.60706544e-01 1.30968735e-01 -2.31706232e-01 3.58844399e-01
4.91268933e-01 -1.33918747e-01 3.74668568e-01 -1.10795522e+00
-1.77909344e-01 -1.81487218e-01 -7.14748502e-01 -7.76352882e-02
-5.18397510e-01 1.25270754e-01 1.20719068e-01 -1.32936954e+00
-6.59226775e-01 2.06430018e-01 -1.51133062e-02 -2.36005604e-01
-2.66453028e-01 4.45535958e-01 4.08497542e-01 -8.40011388e-02
2.98437655e-01 4.21126455e-01 1.00416601e+00 1.56337544e-01
-1.27768114e-01 6.04247630e-01 2.13524606e-02 6.15787625e-01
6.73745513e-01 -4.62885916e-01 -4.25568700e-01 -1.40966726e-02
-1.52772844e-01 4.27099794e-01 2.93109745e-01 -1.49477088e+00
2.12984700e-02 9.25376266e-02 2.29745582e-01 -5.97761035e-01
2.34288856e-01 -1.20338237e+00 6.81766212e-01 8.53052318e-01
-8.83656889e-02 -2.80055463e-01 1.13283440e-01 2.89071321e-01
-6.13745809e-01 -7.54672229e-01 8.20141196e-01 2.10459471e-01
-4.35628533e-01 -6.61269188e-01 -9.76065338e-01 -4.94446069e-01
7.35615492e-01 -3.12933028e-01 9.60274972e-03 -4.47788656e-01
-8.15845668e-01 -3.01110893e-01 5.48877180e-01 -6.95444122e-02
1.97209492e-01 -9.72357452e-01 -2.03457326e-01 4.84372944e-01
-2.01377138e-01 -4.06920075e-01 2.59121239e-01 4.79244381e-01
-6.58463836e-01 4.78576332e-01 -2.85259843e-01 -2.01664150e-01
-1.66860580e+00 3.43013525e-01 5.28043389e-01 -4.85915244e-02
7.06942454e-02 3.99250798e-02 -1.16109240e+00 3.18895906e-01
4.09962714e-01 -5.76007366e-01 -4.13987517e-01 1.42320082e-01
6.69164479e-01 1.02891183e+00 4.75048214e-01 -4.93311793e-01
-2.54387915e-01 8.85194004e-01 8.86438489e-01 -1.95075065e-01
5.60425997e-01 -8.88583288e-02 4.26886342e-02 6.28806770e-01
9.47782040e-01 2.78478324e-01 -8.67294848e-01 7.56951347e-02
7.39939436e-02 -4.17220175e-01 3.11803937e-01 -6.60783708e-01
-4.13934410e-01 8.41698170e-01 6.21880233e-01 6.11048281e-01
1.38930047e+00 -7.47922003e-01 3.80285114e-01 1.40349055e-02
4.68460530e-01 -1.25694501e+00 -5.32976463e-02 1.06937468e-01
6.59362912e-01 -5.60935676e-01 2.54465699e-01 -6.41549647e-01
-1.41563043e-01 1.55317390e+00 -9.89720225e-02 -2.25694358e-01
4.09187585e-01 1.72209665e-01 3.42869908e-01 4.30519581e-01
5.84208854e-02 -1.82266504e-01 3.65968317e-01 6.77914441e-01
4.45468634e-01 2.24714294e-01 -9.69778717e-01 1.36153713e-01
3.69569170e-03 4.49654371e-01 8.42806101e-01 7.29875863e-01
-7.40948081e-01 -1.39514339e+00 -1.07158363e+00 1.28128186e-01
-4.05527115e-01 4.80851561e-01 -3.39453399e-01 6.99384987e-01
1.03402585e-01 1.07570064e+00 -3.90870646e-02 -2.62090504e-01
8.64255846e-01 2.62996376e-01 4.42947268e-01 -1.52217776e-01
-4.27348077e-01 3.73898834e-01 2.19504923e-01 -4.25591767e-01
-3.05168003e-01 -4.85070407e-01 -1.28298187e+00 -2.05224380e-01
-7.13405013e-01 4.84926701e-01 1.32556975e+00 7.09782660e-01
9.42579061e-02 8.90534520e-01 5.24340928e-01 -5.42873681e-01
-9.23656583e-01 -9.95541632e-01 -9.17724311e-01 3.68566126e-01
1.35428742e-01 -4.71055597e-01 -6.56995058e-01 -1.98769748e-01] | [15.05114459991455, 5.64572811126709] |
2208ab8c-282e-42d0-b467-a18221121ac5 | multi-agent-embodied-question-answering-in | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2013_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123580647.pdf | Multi-Agent Embodied Question Answering in Interactive Environments | We investigate a new AI task --- Multi-Agent Interactive Question Answering --- where several agents explore the scene jointly in interactive environments to answer a question. To cooperate efficiently and answer accurately, agents must be well-organized to have balanced work division and share knowledge about the objects involved. We address this new problem in two stages: Multi-Agent 3D Reconstruction in Interactive Environments and Question Answering. Our proposed framework features multi-layer structural and semantic memories shared by all agents, as well as a question answering model built upon a 3D-CNN network to encode the scene memories. During the reconstruction, agents simultaneously explore and scan the scene with a clear division of work, organized by next viewpoints planning. We evaluate our framework on the IQuADv1 dataset and outperform the IQA baseline in a single-agent scenario. In multi-agent scenarios, our framework shows favorable speedups while remaining high accuracy. | ['Huaping Liu', 'Di Guo', 'Weilai Xiang', 'Sinan Tan', 'Fuchun Sun'] | null | null | null | null | eccv-2020-8 | ['embodied-question-answering'] | ['computer-vision'] | [-3.11628759e-01 3.95111620e-01 3.46750289e-01 -5.32482028e-01
-8.45475793e-01 -7.39721000e-01 5.49160779e-01 4.34242368e-01
-4.55301642e-01 5.40723741e-01 2.83720791e-01 7.51992166e-02
-2.06766173e-01 -1.06847739e+00 -9.65026259e-01 -1.85309246e-01
-2.13011086e-01 1.67590964e+00 7.86357403e-01 -3.99596751e-01
2.66368717e-01 3.24232966e-01 -1.43961680e+00 5.76720953e-01
6.22782886e-01 6.84257269e-01 6.32720888e-01 1.14987099e+00
-2.56572157e-01 1.88661468e+00 -7.87421703e-01 -2.86790252e-01
4.04390454e-01 -2.84871608e-01 -1.85132003e+00 5.36572099e-01
2.62681961e-01 -9.96340275e-01 -4.40185368e-01 6.38129056e-01
1.41642481e-01 4.72035497e-01 1.94395691e-01 -1.25842118e+00
-6.41315103e-01 4.96067405e-01 -3.15487683e-01 1.65591240e-01
9.42209423e-01 4.99088854e-01 9.83616829e-01 -4.61025208e-01
1.00572348e+00 1.62771034e+00 -5.15089277e-03 2.17958570e-01
-8.41174006e-01 -4.08750623e-02 5.74491262e-01 6.87125027e-01
-9.77539659e-01 -2.60657787e-01 5.63128173e-01 -9.94198397e-02
1.57740974e+00 9.48110446e-02 6.91513717e-01 5.26875556e-01
2.42693394e-01 1.10582829e+00 8.42649579e-01 1.59948189e-02
3.94025981e-01 -1.18501529e-01 7.83262029e-02 1.29776347e+00
-1.82151541e-01 -2.83982098e-01 -7.06736445e-01 -1.02273062e-01
9.88462687e-01 -2.75447057e-03 2.85349220e-01 -5.10047376e-01
-1.57810390e+00 6.00125849e-01 6.56772971e-01 3.70587893e-02
-7.49272585e-01 4.27207172e-01 3.19303781e-01 7.08411872e-01
1.95159331e-01 8.19790065e-01 -4.00843114e-01 1.00142108e-02
-1.44920155e-01 9.35448170e-01 1.27082527e+00 1.21252966e+00
9.98029411e-01 -3.56486917e-01 7.30369985e-02 3.18020552e-01
3.28152597e-01 4.32553440e-01 -3.97040874e-01 -1.89741576e+00
6.77061200e-01 1.05145895e+00 4.57545757e-01 -9.21122432e-01
-7.10952580e-01 2.88315080e-02 -4.44283724e-01 3.96014363e-01
4.49484736e-01 7.39377066e-02 -6.49267197e-01 1.37354434e+00
8.56986523e-01 1.05544828e-01 3.63344580e-01 1.13151240e+00
1.12317801e+00 7.18904555e-01 4.11808118e-03 1.61005825e-01
1.60219479e+00 -1.98678076e+00 -5.05961955e-01 -3.70311469e-01
5.56220472e-01 -1.39333531e-01 8.04925859e-01 2.28104070e-01
-1.69671845e+00 -5.60463965e-01 -6.99275017e-01 -5.29955447e-01
-1.26204684e-01 -4.90733117e-01 5.69749057e-01 -2.84194231e-01
-1.41286552e+00 8.35542846e-03 -9.34843659e-01 -4.10756409e-01
3.67194682e-01 4.53086555e-01 -4.11086947e-01 -2.57284373e-01
-5.59660435e-01 1.09891081e+00 2.19589889e-01 -1.97245255e-01
-1.87440419e+00 -3.14240932e-01 -8.10945749e-01 9.03452784e-02
9.29293573e-01 -1.32945573e+00 1.80094469e+00 -5.12275159e-01
-1.63908756e+00 7.64596999e-01 -2.19336096e-02 -4.16163683e-01
3.55651140e-01 -3.91775101e-01 2.59115875e-01 5.66713512e-01
3.52564633e-01 8.65725636e-01 2.22029820e-01 -1.59707355e+00
-8.55562270e-01 -7.09608078e-01 1.20516086e+00 8.77060950e-01
5.38727582e-01 -1.07738942e-01 -6.04559660e-01 1.77916199e-01
1.37522787e-01 -9.46761370e-01 -6.78036392e-01 -1.06615342e-01
-2.29268029e-01 -4.24246132e-01 5.55260479e-01 -4.79319155e-01
1.49075672e-01 -1.64633620e+00 5.93469918e-01 -1.28776222e-01
5.94877303e-01 -3.36979836e-01 -5.70799410e-01 7.10249543e-01
6.53527677e-01 -3.63738835e-01 1.00655593e-01 -6.72433496e-01
2.42649391e-01 4.50612158e-01 -2.46171542e-02 3.33620697e-01
4.35421243e-02 1.30692160e+00 -1.01762414e+00 -6.29095972e-01
1.66126713e-01 -3.84414718e-02 -8.08004677e-01 8.40277076e-01
-1.19719398e+00 6.41009748e-01 -1.00112534e+00 6.90998852e-01
5.22074759e-01 -8.10681462e-01 2.83818454e-01 4.04671013e-01
3.34067792e-02 3.24542493e-01 -1.06179380e+00 2.57048678e+00
-3.76437306e-01 4.48986948e-01 6.41174555e-01 -7.50467837e-01
6.75115407e-01 1.58177987e-01 3.06336075e-01 -9.87635076e-01
-5.70410974e-02 -1.10252671e-01 -2.48639137e-01 -8.23551893e-01
7.07087398e-01 4.59643722e-01 -2.17587233e-01 8.48894179e-01
4.88258293e-03 -5.62418461e-01 1.84461236e-01 5.67108333e-01
1.43503308e+00 1.86766516e-02 8.18743631e-02 -1.81803107e-01
1.93214953e-01 8.01488280e-01 3.38415965e-03 1.05963922e+00
-1.92064837e-01 1.22671440e-01 3.59930098e-01 -1.05085731e+00
-7.73682117e-01 -9.77869749e-01 8.46621454e-01 1.48479033e+00
1.04970181e+00 -1.22894906e-01 -7.65320659e-01 -8.54140818e-01
-2.82467932e-01 5.86565495e-01 -5.31450987e-01 4.07223105e-01
-1.03823566e+00 -1.03062965e-01 1.84498161e-01 3.93499047e-01
1.00295794e+00 -1.53957045e+00 -1.50636971e+00 2.71174669e-01
-5.34980834e-01 -1.28360498e+00 -1.89211249e-01 -2.12686986e-01
-4.64703053e-01 -1.43828154e+00 -3.26739848e-01 -1.07524860e+00
5.93503714e-01 6.54376984e-01 1.76858103e+00 4.59488481e-01
8.67340639e-02 1.01296794e+00 -5.08149028e-01 -3.33487779e-01
-3.95492226e-01 1.95640400e-01 -4.19122159e-01 -3.84007215e-01
-1.45756990e-01 -4.92345154e-01 -7.56852269e-01 4.05763149e-01
-6.67119563e-01 4.30057615e-01 4.77526844e-01 1.84479117e-01
4.14543182e-01 -1.40424699e-01 4.26514000e-01 -6.96961999e-01
6.12195969e-01 -5.12484372e-01 -6.96162224e-01 5.88582218e-01
3.31280869e-03 -2.16235429e-01 3.44388157e-01 -9.27172303e-02
-1.26604497e+00 -1.42854154e-01 5.90487421e-02 -1.00178912e-01
-5.79932153e-01 2.64464796e-01 -1.39510125e-01 4.65323888e-02
5.41784763e-01 1.70767948e-01 -1.85974911e-01 -9.80249867e-02
5.47886193e-01 -5.01473360e-02 5.54190457e-01 -8.19869936e-01
3.83355588e-01 6.85236096e-01 -1.18427545e-01 -3.97026956e-01
-6.85896873e-01 -3.65049005e-01 -6.17481470e-01 -3.23362619e-01
1.14702618e+00 -1.05704749e+00 -1.55765080e+00 4.02871281e-01
-1.82882440e+00 -9.70363855e-01 -4.18019563e-01 1.29921719e-01
-9.57562029e-01 5.12273423e-02 -7.23929882e-01 -7.33649850e-01
-7.16273040e-02 -1.42737079e+00 1.24937129e+00 3.89580429e-01
-2.67946441e-02 -8.91985178e-01 4.15036410e-01 9.57285464e-01
4.48396385e-01 1.74543560e-01 7.70304918e-01 -7.94777691e-01
-1.64889777e+00 3.42115670e-01 -4.01666582e-01 -7.31770039e-01
-3.91233504e-01 -8.37146103e-01 -5.56368172e-01 -2.15884268e-01
-1.64578885e-01 -9.49511290e-01 5.78932166e-01 2.04145029e-01
8.92150521e-01 -3.17384303e-01 -3.83611202e-01 8.38579014e-02
1.09492457e+00 3.29856038e-01 3.35123539e-01 4.70545352e-01
3.43854487e-01 7.62729645e-01 6.10190511e-01 5.33106744e-01
1.71142972e+00 5.01817405e-01 1.13152218e+00 -1.91238359e-01
-1.65341720e-02 4.67812968e-03 1.34763755e-02 4.37932938e-01
-7.29417577e-02 -7.48219728e-01 -1.03159285e+00 7.43748128e-01
-2.19111562e+00 -8.67006361e-01 -1.77721679e-01 1.24282193e+00
2.89707422e-01 -2.42794663e-01 2.41716847e-01 -5.45238972e-01
2.76731282e-01 5.31035662e-01 -9.37499166e-01 -1.84930727e-01
3.50140333e-02 -2.08289847e-01 -1.51655123e-01 8.09408128e-01
-8.41163695e-01 1.10409367e+00 6.33554411e+00 -3.08617596e-02
-7.43526891e-02 4.63144094e-01 5.28451622e-01 -1.50293708e-01
-2.10494906e-01 -1.06189251e-01 -3.17257017e-01 -3.99092197e-01
6.28108859e-01 1.87475830e-01 6.81741893e-01 6.88883960e-01
-7.41204768e-02 -6.91658914e-01 -1.41222548e+00 7.76562989e-01
2.04728410e-01 -1.81093228e+00 5.05866334e-02 -9.25736800e-02
7.68948913e-01 2.86977381e-01 -2.44229928e-01 4.66618896e-01
9.52837646e-01 -9.79708791e-01 8.16709638e-01 6.01597369e-01
1.11141182e-01 -4.73258078e-01 6.76369965e-01 8.17183137e-01
-1.23752034e+00 -1.45827129e-01 -1.21508524e-01 -4.13652688e-01
5.42883277e-01 -3.65372986e-01 -9.45626557e-01 8.04120362e-01
8.29866290e-01 3.33515644e-01 -2.29585141e-01 6.86475277e-01
1.43561393e-01 -2.04516575e-01 -1.20781556e-01 -1.03551731e-01
5.12853980e-01 -1.31644070e-01 4.89170790e-01 7.25179136e-01
-1.92820609e-01 8.26590955e-01 9.23015594e-01 9.29495335e-01
-6.62357956e-02 -3.08739543e-01 -3.11369210e-01 2.30676964e-01
3.48801404e-01 9.55920100e-01 -7.83243239e-01 -5.99036694e-01
-4.58331496e-01 1.22572482e+00 8.18991244e-01 5.15614092e-01
-8.90114605e-01 1.67458981e-01 6.34257317e-01 5.98883769e-03
2.47330219e-02 -6.62759542e-01 -3.69153768e-02 -9.95410681e-01
-2.23716889e-02 -1.08363521e+00 5.23525536e-01 -1.02422941e+00
-8.07654321e-01 9.70221460e-01 -6.21463172e-02 -4.10705686e-01
-2.96539545e-01 -2.26384237e-01 -4.51219290e-01 3.71728837e-01
-1.74639106e+00 -1.63267624e+00 -8.47643077e-01 8.29492331e-01
1.22814906e+00 -2.26352334e-01 1.04452205e+00 -7.94522092e-02
-2.91940600e-01 -2.64621645e-01 -5.50851285e-01 -3.91874425e-02
5.90945780e-02 -1.27217865e+00 6.09620154e-01 2.41564244e-01
3.06289256e-01 3.76745015e-02 4.49868113e-01 -6.65421128e-01
-1.83664119e+00 -8.75327468e-01 3.52000713e-01 -7.94837773e-01
3.63460809e-01 -2.87885189e-01 -8.22139919e-01 1.11153221e+00
8.46257806e-01 -9.99473035e-02 3.34944159e-01 -2.71759345e-03
-2.24527717e-01 3.22767258e-01 -1.09423316e+00 4.47518826e-01
1.38532960e+00 -4.61248040e-01 -8.80357385e-01 7.04326868e-01
1.18515062e+00 -7.94717669e-01 -6.23537540e-01 3.22872177e-02
2.34822392e-01 -1.39417100e+00 1.06490481e+00 -8.48337531e-01
5.41203618e-01 -3.11913043e-01 -3.71323496e-01 -1.15726662e+00
-2.03608453e-01 -5.27918696e-01 -1.44793510e-01 5.29005706e-01
3.34421515e-01 -3.44431996e-01 9.70053852e-01 8.02534938e-01
-2.36604840e-01 -5.21259248e-01 -6.69422328e-01 -2.26160824e-01
-4.46014911e-01 -2.23387763e-01 9.09289539e-01 5.66904247e-01
-3.79443206e-02 5.01510203e-01 -5.11641167e-02 7.63915181e-01
5.82057357e-01 6.22213364e-01 1.34611499e+00 -9.76859033e-01
-3.92052770e-01 -1.07363664e-01 1.06282108e-01 -1.76541483e+00
4.64191496e-01 -5.61935067e-01 2.19861954e-01 -2.33146548e+00
4.58951682e-01 -4.50645983e-01 3.24280560e-01 2.15407804e-01
2.04321206e-01 -2.55923271e-01 4.78487492e-01 3.21385771e-01
-1.98063409e+00 5.44651985e-01 1.79629099e+00 -2.77417868e-01
-3.12457025e-01 -2.94899344e-01 -3.71135473e-01 7.29697049e-01
5.16085327e-01 -1.14986584e-01 -6.54726386e-01 -1.47445130e+00
3.47777545e-01 9.27110493e-01 7.05953538e-01 -9.49606180e-01
9.08896685e-01 -4.03563887e-01 3.99574749e-02 -7.64960885e-01
9.76450801e-01 -8.34130108e-01 1.01626754e-01 2.90798575e-01
-5.55495441e-01 5.81223488e-01 -4.58947346e-02 6.28231287e-01
-1.27253473e-01 1.14690885e-01 1.42079324e-01 -8.66915166e-01
-9.65123773e-01 4.67839628e-01 -4.81045395e-01 2.20288873e-01
1.38016129e+00 7.07882270e-02 -6.92564666e-01 -7.59444296e-01
-6.43310726e-01 1.02138519e+00 5.15195608e-01 2.10826889e-01
7.75986969e-01 -8.30320776e-01 -9.45869923e-01 -1.13946974e-01
-2.78377589e-02 8.41347754e-01 6.19104505e-01 3.45908940e-01
-8.87435198e-01 3.38694304e-01 -2.45958939e-01 -4.31661129e-01
-9.00358975e-01 5.16751647e-01 5.20888984e-01 -6.15668714e-01
-6.19063973e-01 1.03750086e+00 6.11659288e-01 -8.01250815e-01
3.95145565e-01 -2.71872908e-01 -3.06250274e-01 -2.61306882e-01
6.89336479e-01 5.10872841e-01 -2.23809198e-01 -4.00542706e-01
-3.28057051e-01 3.09040159e-01 -1.10386722e-01 -2.44437754e-01
1.54647517e+00 -5.86324930e-01 -3.44003737e-01 1.39990479e-01
5.93182921e-01 -5.28528094e-01 -1.35911119e+00 -4.79325742e-01
-2.14754771e-02 -4.57780987e-01 -2.74483025e-01 -9.90796089e-01
-7.36904383e-01 4.79093164e-01 -1.54987484e-01 6.69518888e-01
7.25182056e-01 7.07741737e-01 7.89063513e-01 1.03121614e+00
8.70255828e-01 -8.46036315e-01 9.90309775e-01 7.30397165e-01
1.38599288e+00 -1.46067846e+00 -1.24018982e-01 -2.47666761e-01
-9.26387787e-01 6.97921693e-01 1.21047533e+00 -2.76406407e-01
2.01776907e-01 1.96742088e-01 2.80199319e-01 -8.61592531e-01
-1.29001415e+00 -3.02390754e-01 -1.99862033e-01 6.45652950e-01
-4.72023964e-01 -3.17428708e-02 8.72908294e-01 3.55588108e-01
-4.24798913e-02 -3.12877953e-01 5.32496691e-01 1.02241600e+00
-5.63431799e-01 -4.68343914e-01 -2.37027273e-01 9.09409299e-02
1.71582714e-01 5.64914286e-01 -7.04029977e-01 6.63083613e-01
-9.22743902e-02 1.39072132e+00 3.83154958e-01 -2.07232788e-01
6.76330030e-01 -2.56302059e-01 5.68092525e-01 -6.31030262e-01
-1.03106463e+00 -3.49239856e-01 4.01503682e-01 -1.09125054e+00
-7.66008556e-01 -4.58610356e-01 -1.78018200e+00 -5.09129882e-01
-1.83537100e-02 1.53430685e-01 4.05956030e-01 1.23611403e+00
5.56611538e-01 6.88052297e-01 2.79426545e-01 -1.04686677e+00
-2.49385118e-01 -6.51428580e-01 -2.16419742e-01 1.80836976e-01
4.82852638e-01 -3.66421103e-01 3.32496077e-01 -2.02449754e-01] | [4.3974151611328125, 0.5948408842086792] |
5c9602d0-8020-40aa-b27d-2a57033159e2 | on-high-dimensional-poisson-models-with | 2301.00139 | null | https://arxiv.org/abs/2301.00139v1 | https://arxiv.org/pdf/2301.00139v1.pdf | On High dimensional Poisson models with measurement error: hypothesis testing for nonlinear nonconvex optimization | We study estimation and testing in the Poisson regression model with noisy high dimensional covariates, which has wide applications in analyzing noisy big data. Correcting for the estimation bias due to the covariate noise leads to a non-convex target function to minimize. Treating the high dimensional issue further leads us to augment an amenable penalty term to the target function. We propose to estimate the regression parameter through minimizing the penalized target function. We derive the L1 and L2 convergence rates of the estimator and prove the variable selection consistency. We further establish the asymptotic normality of any subset of the parameters, where the subset can have infinitely many components as long as its cardinality grows sufficiently slow. We develop Wald and score tests based on the asymptotic normality of the estimator, which permits testing of linear functions of the members if the subset. We examine the finite sample performance of the proposed tests by extensive simulation. Finally, the proposed method is successfully applied to the Alzheimer's Disease Neuroimaging Initiative study, which motivated this work initially. | ['Yanyuan Ma', 'Jianxuan Liu', 'Yeqing Zhou', 'Fei Jiang'] | 2022-12-31 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 7.62296394e-02 8.84072036e-02 -2.49551814e-02 -5.98396897e-01
-1.04017651e+00 -7.83883035e-02 -5.14889583e-02 7.35785440e-02
-6.40213609e-01 1.18223989e+00 6.09535053e-02 -9.49112102e-02
-3.83833140e-01 -6.55968964e-01 -6.84272349e-01 -9.14718568e-01
-2.42044687e-01 4.93393898e-01 -2.04923004e-01 4.00182217e-01
5.75613556e-03 1.16351377e-02 -1.22904408e+00 -5.79499602e-01
1.18254209e+00 1.04044187e+00 1.80061191e-01 5.09650707e-01
2.82556862e-01 1.21774711e-01 -4.36445981e-01 -2.14014217e-01
2.38903224e-01 -2.32860476e-01 -3.03448796e-01 1.13962352e-01
1.26599684e-01 -2.70010650e-01 3.10555607e-01 1.45924878e+00
7.22176194e-01 -5.75704873e-02 9.07111526e-01 -1.24421787e+00
-3.18211764e-01 4.46662635e-01 -6.81495547e-01 1.04561485e-01
-9.57229808e-02 -2.94339329e-01 1.01453388e+00 -1.03705955e+00
4.20535624e-01 1.25823319e+00 7.91328609e-01 3.87249410e-01
-1.32225239e+00 -7.39763558e-01 -4.19013388e-02 -1.81395561e-01
-1.59404016e+00 -4.28511947e-01 1.37865692e-01 -5.53081810e-01
2.35762730e-01 5.60173355e-02 3.43759388e-01 1.11102152e+00
2.60316253e-01 3.66696984e-01 1.08174193e+00 -2.11557016e-01
7.22019494e-01 1.47750825e-01 5.98620296e-01 4.83863860e-01
7.92359769e-01 -2.35091094e-02 -1.92675982e-02 -7.92784095e-01
8.85800004e-01 1.47238444e-03 -2.21715689e-01 -1.56390950e-01
-9.25615191e-01 1.13587987e+00 -1.17975630e-01 -1.32322505e-01
-5.78701138e-01 1.46535665e-01 3.10118467e-01 2.39862680e-01
8.56545389e-01 -1.17241412e-01 -4.01152998e-01 3.25695537e-02
-7.89370060e-01 3.56566817e-01 8.02460015e-01 1.14021611e+00
1.75151691e-01 -9.52006802e-02 -3.47331256e-01 1.01487362e+00
3.38567615e-01 8.86153340e-01 6.58876225e-02 -9.54169750e-01
6.16496205e-01 1.77090093e-01 3.64621878e-01 -5.00920296e-01
-5.60953677e-01 -4.96903569e-01 -1.43989170e+00 -1.79036409e-01
4.35343832e-01 -5.66481471e-01 -5.16959727e-01 2.07343602e+00
5.27910531e-01 2.02294245e-01 -2.59751171e-01 7.97759235e-01
2.16424093e-01 1.34962395e-01 2.30070144e-01 -9.00009990e-01
1.30392468e+00 2.87038484e-03 -9.06998396e-01 2.73851771e-02
6.79306626e-01 -2.95410782e-01 9.40421402e-01 4.49031204e-01
-1.15152073e+00 -1.18314251e-01 -6.69968486e-01 2.43616581e-01
3.56989592e-01 1.87446047e-02 2.96890557e-01 7.92957723e-01
-8.33645284e-01 1.54997885e-01 -7.86971688e-01 -1.48298576e-01
7.72099495e-01 5.24325967e-01 -2.70953774e-01 -1.99631423e-01
-9.74574506e-01 3.44252378e-01 -9.30128619e-02 3.23299855e-01
-4.50685978e-01 -9.80407417e-01 -6.94967449e-01 1.65181771e-01
2.36163095e-01 -9.64425027e-01 9.48075414e-01 -5.70374787e-01
-8.32642674e-01 5.72102726e-01 -5.49628615e-01 -3.02794188e-01
9.93050754e-01 -1.44601330e-01 7.91573077e-02 -1.36617407e-01
5.94255686e-01 3.97617966e-02 8.09805691e-01 -7.22611904e-01
-5.29352009e-01 -7.61252165e-01 -5.11094332e-01 -7.87611082e-02
-2.27347746e-01 1.00943036e-01 -7.41105452e-02 -8.46352339e-01
1.67087346e-01 -7.59685099e-01 -6.75159752e-01 -6.60682172e-02
-5.45605600e-01 -1.28804207e-01 1.51691154e-01 -4.57597673e-01
1.21175623e+00 -2.23187280e+00 -1.12698242e-01 5.29000998e-01
4.94417816e-01 -5.63691556e-01 -4.28145602e-02 -1.49041280e-01
6.10368326e-04 3.10435683e-01 -5.81263244e-01 -2.70957559e-01
-1.04223356e-01 -1.32285701e-02 6.15295507e-02 1.03789794e+00
1.31194681e-01 4.71886098e-01 -5.85406721e-01 -3.21882844e-01
-2.91818738e-01 2.30302930e-01 -8.64663899e-01 1.48649901e-01
7.36915171e-02 3.94973844e-01 -8.26978207e-01 5.07089198e-01
1.04551446e+00 -4.11003619e-01 6.97154254e-02 8.89332816e-02
1.39702186e-01 -2.92070448e-01 -1.53166914e+00 6.61016345e-01
-1.57196909e-01 -6.17965730e-03 3.98567319e-01 -1.28024089e+00
8.50208819e-01 2.62381673e-01 7.00012028e-01 -2.93556452e-01
2.48066083e-01 3.86756599e-01 -4.11785729e-02 -6.21515691e-01
-1.07641533e-01 -5.63023448e-01 -3.49487126e-01 1.20946638e-01
-2.17424676e-01 3.86318505e-01 -7.37127885e-02 -8.88369828e-02
1.19631696e+00 -5.68516970e-01 7.71877408e-01 -5.29810131e-01
3.36171895e-01 -5.03096938e-01 7.92944133e-01 1.16948175e+00
-1.47389889e-01 5.78300893e-01 9.07837093e-01 3.66908982e-02
-1.13912117e+00 -1.03110611e+00 -9.09879506e-01 5.64302504e-01
-2.17591986e-01 3.96242701e-02 -5.83160043e-01 -5.27668893e-01
2.81272382e-01 4.49541420e-01 -7.54218459e-01 9.93623883e-02
-2.73681134e-01 -1.55283833e+00 3.23051959e-01 4.41187412e-01
2.87342608e-01 -4.72025067e-01 -1.43068865e-01 1.86566815e-01
3.21006328e-02 -9.25532639e-01 -3.98086101e-01 2.67581850e-01
-1.06913853e+00 -1.21405566e+00 -8.84163857e-01 -5.66282451e-01
7.54315376e-01 -6.37376234e-02 7.06899524e-01 -1.71319216e-01
-5.94293252e-02 3.83440882e-01 2.31048353e-02 -3.06198567e-01
-1.09723814e-01 -1.92292660e-01 4.27447677e-01 -3.57000306e-02
3.52857262e-01 -4.63205695e-01 -6.12062752e-01 5.41696548e-01
-7.08245993e-01 -5.63490272e-01 3.02716762e-01 1.04613292e+00
9.03246164e-01 -4.16968837e-02 1.22890806e+00 -1.14538288e+00
7.07368553e-01 -1.07597601e+00 -8.19413364e-01 5.95587306e-02
-4.85042781e-01 1.21465489e-01 3.43757659e-01 -6.55670941e-01
-6.85169935e-01 -6.87551275e-02 4.99771237e-02 -1.08932033e-01
1.08896904e-01 6.09669566e-01 -3.97884697e-01 1.17776409e-01
3.28045875e-01 -3.06737185e-01 8.18237290e-02 -5.14670610e-01
1.27141222e-01 7.11574256e-01 2.68663019e-01 -5.39062798e-01
5.35065055e-01 3.51989210e-01 3.73642862e-01 -9.65093970e-01
-7.33666599e-01 -4.60800022e-01 -3.28584611e-01 1.10449441e-01
6.50689185e-01 -9.38040316e-01 -1.06394863e+00 2.05605045e-01
-9.20254946e-01 -1.04469627e-01 -3.44252080e-01 8.28647792e-01
-6.39056206e-01 3.49456489e-01 -3.94437641e-01 -1.27056873e+00
-2.46401414e-01 -1.01482987e+00 9.66545463e-01 -1.89110726e-01
-1.82730779e-01 -9.97850120e-01 7.19657913e-02 -1.21266246e-02
1.19830243e-01 2.19949737e-01 1.03923631e+00 -9.50465500e-01
-2.10935384e-01 -4.08975124e-01 -3.86199504e-01 2.66593963e-01
-1.26037851e-01 -4.51784492e-01 -5.76945961e-01 -2.24283636e-01
6.43940389e-01 -4.08658013e-02 7.03742087e-01 1.26174557e+00
1.33049321e+00 -4.22042191e-01 -2.93373495e-01 5.40485501e-01
1.36660755e+00 -2.51009256e-01 4.37924117e-01 1.59358025e-01
4.37177151e-01 4.65073586e-01 6.28541291e-01 9.80312467e-01
2.44954824e-01 3.35813254e-01 1.31026328e-01 1.69178873e-01
5.65207899e-01 3.04156572e-01 2.45060921e-01 6.93288505e-01
1.47699177e-01 -1.48883402e-01 -6.19991243e-01 3.25139254e-01
-1.94880712e+00 -7.56326973e-01 -6.09830737e-01 2.63526750e+00
9.15748775e-01 -1.16141133e-01 2.30518192e-01 -1.45098731e-01
1.01520395e+00 -4.52260613e-01 -8.04180562e-01 2.59399433e-02
-1.83395311e-01 2.98798770e-01 8.52090061e-01 5.54424167e-01
-9.24922764e-01 2.73153782e-01 7.31267929e+00 8.33594263e-01
-4.89410520e-01 2.97350258e-01 9.27220643e-01 -2.14732304e-01
-1.93997234e-01 -2.80603528e-01 -9.78079617e-01 6.32647932e-01
1.06113565e+00 -3.11173588e-01 6.86472729e-02 8.94145131e-01
8.32232296e-01 -4.02120590e-01 -1.11686742e+00 8.22628379e-01
-2.58816391e-01 -7.01081395e-01 -4.31286722e-01 5.28098464e-01
5.82902968e-01 -9.37689319e-02 8.94783810e-03 1.66987896e-01
1.19958907e-01 -1.08818281e+00 3.63937080e-01 6.28493667e-01
9.44221377e-01 -7.04997182e-01 1.04309046e+00 3.02057028e-01
-7.00498044e-01 -2.48848692e-01 -7.66308069e-01 4.16887850e-02
2.30591208e-01 1.17618346e+00 -6.99601829e-01 -4.26441804e-02
4.10134614e-01 4.45775509e-01 -3.97988856e-01 1.45086277e+00
2.18886584e-01 7.55144715e-01 -5.97905517e-01 3.59214284e-02
-3.52188826e-01 -6.40220821e-01 6.72060311e-01 8.50565076e-01
7.22490549e-01 1.88499495e-01 8.68046805e-02 9.25716460e-01
-2.76357174e-01 3.72012734e-01 -4.78353947e-01 3.89996558e-01
6.15856528e-01 9.14468944e-01 -4.68934566e-01 -1.11986607e-01
-5.33174157e-01 5.15607178e-01 2.23380193e-01 5.69755673e-01
-6.27792716e-01 4.42016013e-02 5.84675610e-01 3.00677001e-01
3.08294415e-01 -6.06723540e-02 -7.80514181e-01 -9.93830144e-01
4.64678019e-01 -5.84276557e-01 4.82895315e-01 -2.23846585e-01
-1.74328613e+00 1.17582008e-01 1.91780210e-01 -1.05356002e+00
-7.43649527e-02 -3.42262149e-01 -5.09470522e-01 1.03860784e+00
-1.09443879e+00 -4.50806350e-01 -6.96500540e-02 5.14518857e-01
1.48676321e-01 -1.26356825e-01 6.17693126e-01 3.74193937e-01
-7.44543016e-01 7.18374193e-01 5.07113874e-01 -1.75962433e-01
6.66523218e-01 -1.07111168e+00 -1.80921480e-02 4.96839106e-01
-7.35868156e-01 7.51790285e-01 7.60304213e-01 -1.11606729e+00
-1.04005075e+00 -1.08188236e+00 6.86461091e-01 -1.73902392e-01
8.74145210e-01 -5.39772928e-01 -9.51618135e-01 4.96822596e-01
-6.60782099e-01 2.97092617e-01 6.82296216e-01 3.99050802e-01
2.57979389e-02 -1.25934798e-02 -1.53706968e+00 2.57787734e-01
8.76438797e-01 -1.90734081e-02 -1.09340265e-01 5.58085382e-01
5.28305829e-01 8.73842090e-02 -1.11394882e+00 4.59459901e-01
5.32811284e-01 -5.62816679e-01 8.66944432e-01 -8.20794344e-01
2.26443425e-01 1.26876637e-01 -3.38450968e-01 -9.68069792e-01
-2.07054287e-01 -4.81738985e-01 1.00894965e-01 1.15000212e+00
3.46123010e-01 -6.73545182e-01 5.78553021e-01 8.88898671e-01
3.91070813e-01 -6.67330861e-01 -1.52355134e+00 -7.53198802e-01
2.22046211e-01 -5.85243940e-01 5.70915490e-02 5.70269942e-01
-1.13753438e-01 2.34105453e-01 -4.57043737e-01 3.06973338e-01
1.18199873e+00 -5.11646628e-01 3.85308892e-01 -1.61144376e+00
-2.09637135e-01 2.90600327e-03 -5.12511730e-01 -6.71287060e-01
2.78348804e-01 -6.06693029e-01 2.80226141e-01 -9.04402375e-01
7.31242716e-01 -6.80713356e-01 -1.25701278e-01 -4.04829420e-02
-5.61955810e-01 -1.47999391e-01 -3.24610502e-01 2.32739657e-01
-4.87831920e-01 6.36996567e-01 1.03465033e+00 9.37247574e-02
-3.99871260e-01 7.27975130e-01 -8.32555890e-01 7.47716546e-01
6.70760274e-01 -8.41784239e-01 -3.09976816e-01 1.23963371e-01
2.88005203e-01 2.77671933e-01 5.39692581e-01 -5.47171652e-01
-1.32900193e-01 -8.73018578e-02 4.63697225e-01 -3.22182238e-01
4.66656052e-02 -7.64128089e-01 -4.42505926e-02 1.57588467e-01
-4.75720674e-01 -1.02002621e-01 -2.70051897e-01 1.01136160e+00
2.34475717e-01 -5.18833280e-01 1.06033051e+00 1.37275636e-01
3.00390661e-01 4.37142402e-01 -3.68428946e-01 5.09694874e-01
8.54692936e-01 1.86754778e-01 -9.54573527e-02 -4.08415765e-01
-9.62730706e-01 4.24728811e-01 -3.86512130e-02 -1.91781700e-01
6.66162550e-01 -1.34337890e+00 -9.66566324e-01 2.78389275e-01
-1.31091684e-01 -1.33944064e-01 1.27431557e-01 1.37294519e+00
1.21049717e-01 -5.19379675e-02 2.74514169e-01 -6.88621163e-01
-1.14429212e+00 4.03744757e-01 1.99713901e-01 2.79423557e-02
-6.52163863e-01 4.64657277e-01 4.12830532e-01 -6.08357489e-02
3.77453834e-01 -4.58358347e-01 -1.63061619e-02 9.82530266e-02
6.65542662e-01 6.88319683e-01 -9.17210728e-02 -2.05974251e-01
-3.68966669e-01 3.27396601e-01 2.12651417e-02 -3.59084666e-01
1.60939801e+00 -4.55876559e-01 -3.53079408e-01 4.55068529e-01
1.34902239e+00 3.98468934e-02 -9.38895822e-01 -2.51819044e-01
2.76088536e-01 -3.11725348e-01 -1.48986308e-02 -2.11897731e-01
-7.57293522e-01 4.68420446e-01 8.13623309e-01 2.45054588e-01
8.40517104e-01 8.52381438e-02 3.37818563e-01 2.71305978e-01
3.19792837e-01 -9.70787942e-01 -4.61496949e-01 4.23765391e-01
7.38159657e-01 -1.36041045e+00 1.19786084e-01 -5.21184921e-01
-4.34752852e-01 8.17221880e-01 1.73516393e-01 -3.75145346e-01
1.03172755e+00 4.38587517e-01 -4.83381569e-01 -8.42210662e-04
-6.17524326e-01 -3.20382342e-02 9.94521976e-02 6.40577197e-01
3.20009768e-01 2.17459157e-01 -9.02799547e-01 1.17657888e+00
-8.66131783e-02 1.57236189e-01 6.67821288e-01 4.11766231e-01
-5.55071473e-01 -8.39348793e-01 -5.51878750e-01 1.09533334e+00
-6.54879332e-01 -2.77948231e-01 1.96852043e-01 6.27134979e-01
-1.70692429e-01 8.79342139e-01 1.03642568e-02 3.36947262e-01
2.78887331e-01 1.50182381e-01 2.50403136e-01 -5.17419338e-01
-7.82837495e-02 5.60882092e-01 -4.41411808e-02 -3.25746775e-01
-2.81704515e-01 -9.66259122e-01 -9.07235086e-01 -1.87352747e-01
-4.15871382e-01 6.52473494e-02 4.08146083e-01 9.63962734e-01
2.58137211e-02 1.77018166e-01 7.91592300e-01 -3.78585368e-01
-1.22569239e+00 -1.09188068e+00 -1.05816829e+00 1.48035407e-01
5.99152625e-01 -7.38950670e-01 -6.04269385e-01 -1.45072088e-01] | [7.591695308685303, 4.6826491355896] |
246bf292-f92c-4b27-bbfe-8ade9fb0a523 | fast-online-video-super-resolution-with | 2202.01731 | null | https://arxiv.org/abs/2202.01731v2 | https://arxiv.org/pdf/2202.01731v2.pdf | Fast Online Video Super-Resolution with Deformable Attention Pyramid | Video super-resolution (VSR) has many applications that pose strict causal, real-time, and latency constraints, including video streaming and TV. We address the VSR problem under these settings, which poses additional important challenges since information from future frames is unavailable. Importantly, designing efficient, yet effective frame alignment and fusion modules remain central problems. In this work, we propose a recurrent VSR architecture based on a deformable attention pyramid (DAP). Our DAP aligns and integrates information from the recurrent state into the current frame prediction. To circumvent the computational cost of traditional attention-based methods, we only attend to a limited number of spatial locations, which are dynamically predicted by the DAP. Comprehensive experiments and analysis of the proposed key innovations show the effectiveness of our approach. We significantly reduce processing time and computational complexity in comparison to state-of-the-art methods, while maintaining a high performance. We surpass state-of-the-art method EDVR-M on two standard benchmarks with a speed-up of over $3\times$. | ['Luc van Gool', 'Radu Timofte', 'Martin Danelljan', 'Dario Fuoli'] | 2022-02-03 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 2.94149965e-01 -3.20014656e-01 -1.94591835e-01 -1.10533334e-01
-8.08015883e-01 -1.87920436e-01 2.29198769e-01 -2.63621420e-01
-3.30385149e-01 5.93346238e-01 3.17387253e-01 -3.07212770e-02
2.53467690e-02 -5.99856913e-01 -5.78522742e-01 -4.13315684e-01
1.42634306e-02 -6.54950365e-02 8.57567310e-01 -4.31896538e-01
2.11231887e-01 3.46437812e-01 -1.66977465e+00 3.81166548e-01
7.86075115e-01 1.13908887e+00 7.69927084e-01 6.13468349e-01
1.47288427e-01 1.10304570e+00 -2.68036395e-01 -1.76495135e-01
1.11162521e-01 -1.61413059e-01 -6.50081277e-01 8.40838477e-02
4.10463542e-01 -6.68436527e-01 -9.29747701e-01 9.61422503e-01
4.97875005e-01 3.57477576e-01 -1.53746322e-01 -8.05359602e-01
-7.75607347e-01 4.94456947e-01 -8.86285126e-01 1.01214743e+00
4.94886041e-01 5.40236011e-02 9.37401772e-01 -1.05945218e+00
6.76418483e-01 1.23511732e+00 3.20618272e-01 5.47809362e-01
-1.06595993e+00 -6.74666584e-01 7.56583273e-01 6.09106541e-01
-1.57566190e+00 -8.28244448e-01 5.87642133e-01 -1.92530617e-01
9.02497828e-01 3.20209593e-01 4.17922556e-01 8.38630259e-01
4.55947593e-02 6.89684093e-01 6.40203118e-01 -4.76823412e-02
9.48826298e-02 -4.38780367e-01 9.88550112e-02 4.94481504e-01
-9.52711031e-02 -1.76376626e-01 -9.73528504e-01 1.09332107e-01
1.25967133e+00 7.08993003e-02 -6.53607488e-01 1.04893401e-01
-1.24918151e+00 4.89461988e-01 2.41167784e-01 2.63790995e-01
-5.46712101e-01 1.72974721e-01 1.75235391e-01 -5.52634411e-02
5.33932328e-01 -7.78803825e-02 -2.88824290e-01 -3.29269826e-01
-1.08996463e+00 7.97230601e-02 1.09175272e-01 1.04977965e+00
4.61886883e-01 2.41582990e-01 -5.04008055e-01 7.21740305e-01
1.19320233e-03 1.83126375e-01 4.10427213e-01 -1.23748243e+00
6.07215047e-01 2.80663550e-01 3.36656928e-01 -1.18692935e+00
-1.91400081e-01 -3.32806498e-01 -9.74281907e-01 -2.23613158e-01
-1.03900306e-01 1.04399502e-01 -7.58382440e-01 1.71894705e+00
4.12579715e-01 7.71952212e-01 -5.71617819e-02 1.25989223e+00
7.34141767e-01 9.36654329e-01 -2.83630401e-01 -7.94718623e-01
1.20005703e+00 -1.18060529e+00 -1.01292312e+00 -2.32407749e-01
2.34137420e-02 -6.71033263e-01 9.27059710e-01 2.73136050e-01
-1.52138901e+00 -7.94278741e-01 -8.99183095e-01 -3.10081154e-01
4.38709140e-01 9.26848967e-03 4.71366495e-01 1.05888739e-01
-1.18200338e+00 6.76132143e-01 -1.10780632e+00 -1.26382932e-01
3.82373422e-01 5.30385971e-01 -1.53181285e-01 -5.54308966e-02
-1.02664018e+00 4.82200980e-01 -2.02580411e-02 1.92478165e-01
-6.92103684e-01 -8.85222673e-01 -7.01570988e-01 1.52755573e-01
5.87148547e-01 -5.04912138e-01 1.39035320e+00 -6.79121494e-01
-1.62417364e+00 4.35014248e-01 -7.86561787e-01 -5.91283023e-01
4.89394784e-01 -6.13068044e-01 -5.03131568e-01 3.90364081e-01
-2.36092322e-02 3.38693380e-01 8.05508435e-01 -1.04625595e+00
-9.94411647e-01 -2.23141566e-01 4.05354679e-01 4.19487178e-01
-3.62232417e-01 3.15894634e-01 -1.11113298e+00 -7.86892891e-01
2.70511478e-01 -7.11866856e-01 -3.81356657e-01 -1.06806390e-01
6.98579922e-02 1.31225144e-03 9.83239770e-01 -7.79438257e-01
1.79222238e+00 -2.22621655e+00 3.80454302e-01 -4.03881460e-01
5.21063864e-01 4.10434991e-01 7.67057315e-02 -5.52031510e-02
5.26136085e-02 -4.25239578e-02 1.84497103e-01 -3.39987040e-01
-3.55528086e-01 3.37822512e-02 -5.77820420e-01 3.65888208e-01
1.08909555e-01 6.60183191e-01 -8.27263713e-01 -5.19364953e-01
3.02692860e-01 6.63153946e-01 -7.54301548e-01 3.41661155e-01
-1.24630757e-01 5.49914777e-01 -4.69209671e-01 5.31064510e-01
5.31637549e-01 -5.86430430e-01 2.30907962e-01 -5.03508806e-01
-3.76286447e-01 3.29350889e-01 -1.17807841e+00 1.78396928e+00
-4.02399182e-01 6.21276975e-01 1.29466699e-02 -7.39398479e-01
7.69686580e-01 3.02589446e-01 4.97385979e-01 -1.00084627e+00
1.84958894e-02 -7.25733340e-02 -2.12857634e-01 -3.55008364e-01
9.63351965e-01 4.08871323e-01 2.46059820e-01 1.47173539e-01
-2.07564458e-01 7.77760923e-01 2.27583781e-01 3.03073257e-01
1.14936554e+00 2.29442105e-01 4.14635658e-01 -1.50767356e-01
5.98858356e-01 -3.58567238e-01 1.11762929e+00 5.49672842e-01
-2.75570691e-01 8.45953941e-01 3.56313169e-01 -5.63396990e-01
-9.05045927e-01 -8.57685030e-01 2.18590006e-01 1.24603128e+00
6.81821823e-01 -5.71422219e-01 -6.48767650e-01 -2.71660656e-01
-5.14089227e-01 3.89848888e-01 -5.05516231e-01 1.68105617e-01
-1.11695683e+00 -5.89150131e-01 -4.97255847e-02 7.02177525e-01
5.26433468e-01 -7.14166105e-01 -9.52721298e-01 5.48539042e-01
-5.28836906e-01 -1.63851047e+00 -8.06625128e-01 -6.27174973e-01
-7.86697626e-01 -7.28062749e-01 -6.92533135e-01 -5.62563539e-01
3.56861711e-01 8.63254011e-01 1.13677251e+00 3.80704440e-02
-6.42733201e-02 6.05070964e-02 -4.91507769e-01 2.79368013e-01
5.63070066e-02 -2.52942964e-02 1.49393350e-01 3.02411675e-01
3.31086665e-02 -7.07021594e-01 -9.70468879e-01 6.27284825e-01
-7.61382222e-01 4.96694952e-01 3.52881938e-01 6.43686414e-01
1.05854940e+00 9.89269465e-02 6.18999958e-01 -7.31944144e-01
2.55566090e-02 -3.68216515e-01 -8.03874195e-01 2.27206916e-01
-1.40689552e-01 -1.62397340e-01 5.85115254e-01 -4.52087402e-01
-1.29341936e+00 3.16277482e-02 7.90549442e-04 -8.70109677e-01
2.84879923e-01 1.84063226e-01 -1.49159625e-01 1.31238112e-02
1.76277369e-01 3.61536205e-01 -3.95568728e-01 -4.46493149e-01
9.05726701e-02 4.96770978e-01 7.94263005e-01 -3.95941526e-01
6.15712583e-01 6.86658144e-01 -1.95153281e-01 -7.94819295e-01
-8.70771527e-01 -3.38984996e-01 -4.40775365e-01 -3.19605947e-01
6.70644760e-01 -1.27526140e+00 -7.75621414e-01 2.81113088e-01
-1.26570618e+00 -1.72866717e-01 -8.21592286e-02 2.46811777e-01
-4.91624206e-01 3.92395020e-01 -7.71280348e-01 -7.30985761e-01
-4.60117996e-01 -1.34563017e+00 9.84239459e-01 3.14752042e-01
7.32458159e-02 -4.55163121e-01 -2.29546458e-01 4.32271451e-01
5.98136425e-01 8.62392336e-02 3.47595632e-01 -4.49015610e-02
-1.18815827e+00 2.71853447e-01 -5.25488436e-01 -2.08460957e-01
1.24229185e-01 -7.33148009e-02 -7.36922860e-01 -4.74322647e-01
-5.58702871e-02 2.05732778e-01 8.47053826e-01 4.46467936e-01
1.47962904e+00 -2.40879714e-01 -2.39431947e-01 7.11870074e-01
1.41201282e+00 5.25097489e-01 6.85590327e-01 1.19049959e-01
8.53826106e-01 1.94425687e-01 9.54691172e-01 7.67364085e-01
6.55232549e-01 1.12908709e+00 3.98017257e-01 4.27033082e-02
-2.15793267e-01 -1.28655583e-01 4.19441521e-01 1.00651062e+00
-4.81792778e-01 -3.51203740e-01 -6.20917022e-01 5.26665866e-01
-2.30397701e+00 -1.15107810e+00 1.02536224e-01 2.18586946e+00
6.76236629e-01 1.23559877e-01 8.19422491e-03 4.78709638e-02
1.00005913e+00 5.50626040e-01 -6.61257267e-01 2.14927495e-02
2.11339388e-02 2.02439591e-01 3.02319556e-01 5.04670084e-01
-1.10343122e+00 1.10317922e+00 6.27167606e+00 7.92055368e-01
-1.11961865e+00 3.10298324e-01 6.10946417e-01 -5.13278902e-01
-6.37763292e-02 -1.51985481e-01 -9.48557377e-01 5.87948263e-01
9.32506502e-01 -3.49392474e-01 7.32004821e-01 6.06981575e-01
5.62403262e-01 3.99822071e-02 -1.08659840e+00 1.32064080e+00
4.69480753e-02 -1.75763953e+00 6.64390102e-02 -4.86109182e-02
6.61154926e-01 -7.29213804e-02 1.37350604e-01 6.69169948e-02
1.53322620e-02 -7.71497309e-01 8.58121097e-01 4.88727450e-01
9.71486390e-01 -8.62218261e-01 4.43430990e-01 -3.16050090e-02
-1.87989354e+00 -1.97572783e-01 -3.95458162e-01 -1.16433091e-01
6.48954153e-01 4.84912664e-01 6.93717748e-02 7.02237487e-01
1.02004743e+00 9.99388218e-01 -3.44200701e-01 7.23933995e-01
1.53079823e-01 3.39726508e-01 -1.75217018e-01 5.82218409e-01
-5.72525784e-02 6.78888559e-02 6.93648815e-01 1.04248273e+00
5.19574344e-01 7.50828087e-01 2.03018099e-01 5.60020149e-01
-1.49338350e-01 -1.44994572e-01 -2.34761044e-01 2.13466480e-01
7.64837086e-01 1.03875351e+00 -5.91086686e-01 -4.33199495e-01
-7.06261635e-01 1.10770047e+00 3.37227672e-01 3.14679652e-01
-1.21149111e+00 -5.30714318e-02 9.00417686e-01 3.24558198e-01
6.44602299e-01 -2.99443811e-01 -1.97337996e-02 -1.46393645e+00
2.99917281e-01 -8.56656849e-01 3.17582875e-01 -7.52149940e-01
-7.72189677e-01 9.06719565e-01 -2.28304803e-01 -1.33095050e+00
-3.10533196e-02 7.30877295e-02 -2.81628191e-01 4.97052759e-01
-1.74465847e+00 -8.37146699e-01 -5.44630229e-01 6.88467026e-01
1.06652367e+00 2.52176654e-02 3.49563539e-01 6.85657024e-01
-9.56733286e-01 5.61232150e-01 -1.35741085e-01 -8.46374482e-02
4.96407121e-01 -7.39042640e-01 6.18428528e-01 1.32641828e+00
-9.74500552e-02 4.18993324e-01 6.02778971e-01 -4.21933472e-01
-1.59576333e+00 -1.22015464e+00 5.41662872e-01 -7.42240921e-02
6.16040707e-01 -1.80530280e-01 -1.10716200e+00 7.40014434e-01
6.69481978e-02 4.27470565e-01 3.03639650e-01 -2.62684584e-01
-1.86351702e-01 -2.95137912e-01 -7.92661190e-01 7.36539066e-01
1.54381526e+00 -3.69404346e-01 -3.90305698e-01 -1.97040159e-02
1.26119530e+00 -8.48936498e-01 -9.21286643e-01 4.80085641e-01
5.48761189e-01 -1.24301815e+00 1.11279321e+00 -2.23023653e-01
5.71735620e-01 -5.28401196e-01 -4.66743797e-01 -7.29071677e-01
-6.41628206e-01 -1.02041304e+00 -7.04244554e-01 1.28025949e+00
-4.68966030e-02 -3.60643268e-01 6.37571275e-01 7.57301867e-01
-1.68248221e-01 -7.79603601e-01 -1.07103121e+00 -4.64820534e-01
-6.94561958e-01 -3.07703644e-01 5.02401829e-01 7.63699472e-01
-2.73463249e-01 3.50231677e-01 -6.88522339e-01 5.94136000e-01
6.54804885e-01 2.81410068e-01 4.88717079e-01 -8.73886943e-01
-4.04255033e-01 -1.11000121e-01 -4.80171919e-01 -1.51160800e+00
-5.93952164e-02 -1.82705641e-01 -1.68335393e-01 -1.48003650e+00
2.44755730e-01 -2.72141933e-01 -4.91562247e-01 2.82131702e-01
-2.74410188e-01 2.65738249e-01 5.36881685e-01 3.02025467e-01
-1.06938326e+00 6.54465139e-01 1.13158703e+00 3.34306479e-01
-3.70155215e-01 -3.29691559e-01 -6.79591417e-01 9.16205108e-01
5.63142002e-01 -1.37734622e-01 -5.16251087e-01 -9.53193188e-01
-6.11924641e-02 6.26675248e-01 1.59174785e-01 -1.13369012e+00
4.41230744e-01 -3.10158461e-01 1.16161630e-01 -7.71717668e-01
6.15544736e-01 -7.06460595e-01 2.89908618e-01 1.16124302e-01
-2.08412230e-01 4.22734439e-01 1.29733518e-01 7.25731850e-01
-2.44635597e-01 4.13085938e-01 1.04305720e+00 9.59314182e-02
-1.08546758e+00 6.78213418e-01 -1.28048286e-01 1.16078742e-01
1.15243709e+00 -1.27020657e-01 -4.32237357e-01 -2.77382433e-01
-5.39034486e-01 3.16509873e-01 5.07336915e-01 5.91724217e-01
8.96254420e-01 -1.25283074e+00 -6.33508205e-01 1.67136639e-02
-2.46079773e-01 1.71234697e-01 8.32086325e-01 7.88134992e-01
-4.86483365e-01 3.70987892e-01 -1.48054942e-01 -5.93035161e-01
-1.40563619e+00 7.85358608e-01 2.41676029e-02 -2.52296478e-01
-1.09360921e+00 6.21493459e-01 5.45841515e-01 5.76583028e-01
2.13978484e-01 -1.25973895e-01 -4.69675273e-01 -6.56100363e-02
1.09860539e+00 5.71399093e-01 -1.33789361e-01 -8.79608154e-01
-3.57917547e-01 7.78571248e-01 -3.94353688e-01 8.32001865e-02
1.44640863e+00 -6.84200287e-01 9.21930522e-02 1.69687465e-01
7.27202952e-01 -1.38394594e-01 -1.67206907e+00 -6.07108593e-01
-3.61864388e-01 -9.43774819e-01 3.17697436e-01 -2.78414577e-01
-1.44676888e+00 5.90400398e-01 5.96820772e-01 2.34513208e-02
1.47672570e+00 -1.29632428e-01 1.16577303e+00 -6.49143457e-02
6.04774356e-01 -8.93430114e-01 -1.64551083e-02 3.58081609e-01
7.97678113e-01 -1.06453156e+00 4.62916084e-02 -6.27837121e-01
-6.02510929e-01 7.73284197e-01 8.85716140e-01 -6.23161942e-02
4.17655408e-01 3.75823528e-01 -2.54534036e-01 2.32747450e-01
-1.24559295e+00 -1.76783100e-01 1.14363082e-01 4.18737024e-01
3.39462578e-01 -2.86351323e-01 -2.46746182e-01 6.96892321e-01
2.08277345e-01 1.95152313e-01 6.44094467e-01 9.11295056e-01
-3.99092346e-01 -7.83698559e-01 -2.91710585e-01 1.85321495e-01
-8.75841558e-01 -1.29953682e-01 4.18067157e-01 4.75847423e-01
-2.28369385e-01 9.03365195e-01 3.28606457e-01 -5.34834445e-01
3.46590310e-01 -5.46133280e-01 2.67249912e-01 -3.09804171e-01
-1.83231235e-01 2.82715887e-01 -1.35005256e-02 -1.14993942e+00
-6.33916199e-01 -5.14771283e-01 -1.19003761e+00 -4.82686251e-01
-1.96031824e-01 -2.63487190e-01 2.25428827e-02 8.28164458e-01
8.27038646e-01 9.76665854e-01 7.64531493e-01 -1.28104830e+00
-1.70978844e-01 -5.22735417e-01 -1.65958330e-01 1.60800293e-01
5.59810400e-01 -7.22060144e-01 1.07497327e-01 3.08662802e-01] | [11.082282066345215, -1.8759123086929321] |
23415d94-c8a3-44b2-8e6e-423781e5e2da | shrimp-sparser-random-feature-models-via | 2112.04002 | null | https://arxiv.org/abs/2112.04002v1 | https://arxiv.org/pdf/2112.04002v1.pdf | SHRIMP: Sparser Random Feature Models via Iterative Magnitude Pruning | Sparse shrunk additive models and sparse random feature models have been developed separately as methods to learn low-order functions, where there are few interactions between variables, but neither offers computational efficiency. On the other hand, $\ell_2$-based shrunk additive models are efficient but do not offer feature selection as the resulting coefficient vectors are dense. Inspired by the success of the iterative magnitude pruning technique in finding lottery tickets of neural networks, we propose a new method -- Sparser Random Feature Models via IMP (ShRIMP) -- to efficiently fit high-dimensional data with inherent low-dimensional structure in the form of sparse variable dependencies. Our method can be viewed as a combined process to construct and find sparse lottery tickets for two-layer dense networks. We explain the observed benefit of SHRIMP through a refined analysis on the generalization error for thresholded Basis Pursuit and resulting bounds on eigenvalues. From function approximation experiments on both synthetic data and real-world benchmark datasets, we show that SHRIMP obtains better than or competitive test accuracy compared to state-of-art sparse feature and additive methods such as SRFE-S, SSAM, and SALSA. Meanwhile, SHRIMP performs feature selection with low computational complexity and is robust to the pruning rate, indicating a robustness in the structure of the obtained subnetworks. We gain insight into the lottery ticket hypothesis through SHRIMP by noting a correspondence between our model and weight/neuron subnetworks. | ['Rachel Ward', 'Hayden Schaeffer', 'Bobby Shi', 'Yuege Xie'] | 2021-12-07 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 2.02307656e-01 9.57551673e-02 -1.49874553e-01 -2.12770551e-01
-3.74005973e-01 -1.30402416e-01 8.59703869e-02 -2.89418370e-01
-9.65079442e-02 8.45625281e-01 -9.28631425e-02 1.36164635e-01
-9.09581244e-01 -7.42114961e-01 -9.66057718e-01 -9.91154850e-01
-6.05818450e-01 3.24773997e-01 -8.06666911e-02 -1.45852745e-01
9.38299671e-02 5.30237377e-01 -1.61795533e+00 1.13232642e-01
8.20671797e-01 1.39012003e+00 1.03777528e-01 1.09600946e-01
1.65428847e-01 6.61774814e-01 -5.43080345e-02 -3.21982771e-01
6.07995987e-01 -2.08272621e-01 -3.68097901e-01 1.60951354e-02
4.94261742e-01 5.76283373e-02 -5.32088578e-01 1.06286776e+00
3.06072801e-01 1.95399553e-01 7.60488629e-01 -1.38377261e+00
-7.59497344e-01 8.77893507e-01 -6.53418481e-01 7.29890838e-02
-2.80691573e-04 -5.79999276e-02 1.18095458e+00 -1.29875255e+00
4.76360738e-01 1.18935680e+00 1.02828324e+00 1.89965665e-01
-1.53570271e+00 -8.56079340e-01 2.93441623e-01 9.38872248e-02
-1.61251009e+00 -6.37226164e-01 8.62311125e-01 -3.31077904e-01
1.02736771e+00 3.21452349e-01 6.46146297e-01 9.58334982e-01
6.57258853e-02 7.34809458e-01 7.78302789e-01 -3.03942800e-01
3.62758189e-01 8.45845640e-02 4.12268102e-01 1.28432512e+00
6.91911519e-01 2.68131703e-01 -9.89566267e-01 -5.77641249e-01
1.00437880e+00 2.23956808e-01 -5.01574457e-01 -7.56185532e-01
-9.79238510e-01 1.13928092e+00 4.20796305e-01 1.73269525e-01
-5.60211897e-01 1.80854827e-01 1.49292633e-01 6.06105745e-01
2.57751316e-01 5.63178957e-01 -5.89777827e-01 1.54418781e-01
-1.19779611e+00 2.17453137e-01 9.98705685e-01 9.18672860e-01
9.62632835e-01 5.75801075e-01 4.58045043e-02 7.99734712e-01
2.43874148e-01 3.83612782e-01 3.80330235e-01 -9.61599648e-01
3.76099557e-01 6.70814514e-01 -3.38356525e-01 -1.31917644e+00
-6.43606782e-01 -9.83568668e-01 -1.38622129e+00 1.35369733e-01
3.90892982e-01 -8.75293538e-02 -7.36549377e-01 1.92509580e+00
8.42251629e-02 3.75848651e-01 -3.30052376e-02 7.86981225e-01
5.87999225e-01 3.00164700e-01 -4.15639877e-01 -2.30013683e-01
8.85016084e-01 -5.69382787e-01 -2.00698629e-01 1.83646455e-02
4.64038759e-01 -1.87263504e-01 7.13763595e-01 5.84089518e-01
-1.24600530e+00 -2.42585376e-01 -9.52396750e-01 2.53818989e-01
-3.66047001e-03 7.10522681e-02 1.18935871e+00 6.28274977e-01
-1.07409525e+00 9.69089866e-01 -8.53426278e-01 1.59708276e-01
7.24900842e-01 8.96413088e-01 -5.02775550e-01 -1.61114886e-01
-9.85389948e-01 4.97111946e-01 -6.65653720e-02 2.36075073e-01
-7.08536923e-01 -9.19834197e-01 -9.42351580e-01 4.98096466e-01
3.36753964e-01 -9.99450684e-01 5.46880960e-01 -9.87090766e-01
-1.13104796e+00 2.97939271e-01 -3.37123215e-01 -9.17506874e-01
1.00282937e-01 4.21096049e-02 6.67711198e-02 3.29624712e-01
-2.50193514e-02 5.58679581e-01 1.20446229e+00 -9.29995418e-01
-2.88414925e-01 -5.23754358e-01 -3.10276330e-01 -2.58257687e-01
-6.27298653e-01 -4.18599933e-01 4.85819653e-02 -6.81510448e-01
4.92195994e-01 -7.15017557e-01 -5.35284758e-01 1.77625269e-01
-3.26536745e-01 5.04219271e-02 6.74123347e-01 -9.73734856e-02
1.23114610e+00 -2.15443897e+00 3.63034666e-01 9.36265647e-01
6.73197210e-01 -2.27750167e-01 -2.51536161e-01 7.76725560e-02
-3.55083257e-01 -1.54180070e-02 -4.20824558e-01 -1.73252508e-01
4.68239151e-02 5.95810488e-02 -3.14311922e-01 8.19254816e-01
2.15645805e-01 6.75659776e-01 -5.28151333e-01 -1.86123431e-01
-7.80636668e-02 6.14557981e-01 -9.28733826e-01 -2.72475302e-01
1.88730568e-01 -1.76907673e-01 -4.78421509e-01 9.01937783e-01
5.10260642e-01 -5.94978750e-01 -4.05608118e-02 -3.56852829e-01
6.83169737e-02 -9.01023597e-02 -1.45035899e+00 1.28077221e+00
-3.76470506e-01 5.85921645e-01 3.10796112e-01 -1.24592364e+00
9.65656519e-01 1.81515485e-01 7.41257191e-01 -4.50472653e-01
1.53320476e-01 3.81235480e-01 1.17218435e-01 -1.76495612e-01
2.42997572e-01 -4.80486862e-02 1.25365853e-01 -1.94036833e-03
3.81255686e-01 2.79501796e-01 3.17038029e-01 2.17167720e-01
1.37029552e+00 -1.89790070e-01 3.03398788e-01 -6.65476918e-01
2.36150205e-01 -2.05436215e-01 6.86360955e-01 8.63349915e-01
1.74237460e-01 5.16644657e-01 6.48966968e-01 -4.31472003e-01
-6.97543859e-01 -8.61510575e-01 -3.24849039e-01 1.00493658e+00
-1.39160484e-01 -2.68325061e-01 -3.64140838e-01 -2.75204539e-01
3.46305579e-01 3.25917423e-01 -7.38644600e-01 -2.66165346e-01
-4.23332810e-01 -8.86816263e-01 4.28134143e-01 4.08941120e-01
2.31240287e-01 -7.76584744e-01 -5.19374907e-01 8.99805352e-02
4.14172262e-01 -8.46038580e-01 -2.71865278e-01 9.16216135e-01
-9.84497070e-01 -1.07614291e+00 -7.09738851e-01 -8.79846394e-01
8.25594544e-01 1.41137451e-01 1.13131583e+00 6.07105345e-02
-3.31215858e-01 3.10516357e-01 -1.33898318e-01 1.50248762e-02
2.36744568e-01 -3.28193679e-02 4.37584698e-01 1.98212057e-01
1.56614482e-01 -1.11444104e+00 -4.31272537e-01 3.15915883e-01
-7.28230000e-01 -2.20989794e-01 7.65373111e-01 1.33369434e+00
6.22571588e-01 2.18034238e-01 7.34278440e-01 -8.31152439e-01
4.49263632e-01 -7.59588301e-01 -6.66370392e-01 -3.54671478e-02
-6.44882858e-01 3.86258453e-01 9.19636726e-01 -6.55391097e-01
-4.90002245e-01 4.55722570e-01 8.03511590e-04 -8.10059130e-01
2.85464168e-01 6.96646333e-01 9.49111655e-02 -7.95262277e-01
6.89593673e-01 4.31040615e-01 7.86367133e-02 -4.61691797e-01
-1.05333412e-02 9.85322669e-02 1.92141160e-01 -6.12353086e-01
9.05977726e-01 5.17554522e-01 4.63742405e-01 -1.17903519e+00
-6.79779291e-01 -2.49384195e-01 -2.84910053e-01 3.06364089e-01
2.12780431e-01 -8.72658551e-01 -9.79849279e-01 1.78907588e-01
-7.51493454e-01 -1.13397472e-01 -4.94846046e-01 4.87138689e-01
-6.33095682e-01 3.57550591e-01 -5.16485214e-01 -8.47505569e-01
-3.53745818e-01 -1.01784277e+00 7.59278476e-01 7.69423246e-02
-2.18689188e-01 -6.91299260e-01 -3.47218156e-01 -1.02476142e-01
5.73581874e-01 1.16798967e-01 1.15741646e+00 -7.71299362e-01
-5.65946281e-01 -9.82254520e-02 -1.61160827e-01 1.17931068e-01
-2.88192749e-01 -1.52610853e-01 -6.77567601e-01 -2.91561991e-01
1.40211210e-01 -3.38658720e-01 1.25277531e+00 8.76412392e-01
1.15780652e+00 -4.87430215e-01 -2.59303778e-01 1.04119420e+00
1.40560985e+00 -1.83275566e-01 4.17854548e-01 1.02717228e-01
5.76124489e-01 4.00219262e-01 4.95825596e-02 7.38521099e-01
-6.81489408e-02 3.59046698e-01 5.37177742e-01 -7.75695518e-02
5.10722436e-02 -5.93095645e-02 3.24531674e-01 5.58498323e-01
-4.99602333e-02 1.68572471e-01 -4.41137791e-01 5.51798940e-01
-1.80151665e+00 -1.10371029e+00 -1.79746032e-01 2.20832825e+00
6.73686445e-01 2.11496979e-01 1.66316599e-01 3.45952868e-01
4.28789735e-01 -9.24402568e-03 -6.86098278e-01 -6.16375096e-02
-4.92940575e-01 4.57877874e-01 7.06887543e-01 3.14827234e-01
-8.78729045e-01 6.18386149e-01 6.14049244e+00 9.29284573e-01
-1.05370283e+00 -1.00903831e-01 5.96195281e-01 -5.97582400e-01
-4.27290708e-01 -2.36861005e-01 -1.11124444e+00 2.99399763e-01
7.81535327e-01 7.14639621e-03 5.46787739e-01 1.12888753e+00
7.46726990e-02 1.88095570e-01 -1.08085525e+00 1.11587048e+00
-3.36973593e-02 -1.68445909e+00 -2.73122713e-02 7.64431581e-02
7.61746645e-01 1.88429102e-01 2.11418986e-01 4.04223770e-01
1.61451995e-01 -1.13286698e+00 6.03643239e-01 4.11465496e-01
6.59548938e-01 -7.87286520e-01 3.30091536e-01 1.90718085e-01
-1.19998121e+00 -5.52800775e-01 -6.49806678e-01 -2.00001702e-01
-1.69666424e-01 9.29224789e-01 -4.11706895e-01 1.84728682e-01
5.60236096e-01 6.81791008e-01 -3.89082074e-01 1.24383175e+00
3.06431562e-01 6.92417026e-01 -7.64046073e-01 -1.20280340e-01
2.22300127e-01 -2.02515244e-01 7.21612573e-01 1.00144494e+00
4.77405667e-01 -1.09569356e-01 8.95462930e-02 1.09859502e+00
-8.31022579e-03 3.48094180e-02 -7.65600860e-01 5.56530245e-02
4.21397507e-01 1.11712730e+00 -6.07040644e-01 -2.58539729e-02
-5.05201459e-01 5.81189334e-01 3.33388090e-01 5.04481554e-01
-5.79658389e-01 -3.09561670e-01 7.30321586e-01 2.20967114e-01
7.66373575e-01 -2.80683637e-01 -6.20739281e-01 -1.18876290e+00
6.55576438e-02 -1.04261494e+00 2.28698462e-01 -2.23504916e-01
-1.40043378e+00 7.02094316e-01 -3.15669477e-01 -1.14995015e+00
-1.99426308e-01 -6.16638899e-01 -3.13979566e-01 6.77229166e-01
-1.37336075e+00 -8.72987807e-01 1.01379998e-01 1.02438319e+00
4.83803660e-01 -5.45552731e-01 8.47844720e-01 2.62782067e-01
-6.42038643e-01 8.49067688e-01 2.54464477e-01 -1.76736608e-01
7.26132467e-02 -7.85816252e-01 3.82430367e-02 7.26851940e-01
3.08365196e-01 9.81733263e-01 6.85217142e-01 -4.95309830e-01
-1.90347552e+00 -8.69698524e-01 5.05258739e-01 1.64640948e-01
8.73102546e-01 -4.46839631e-01 -8.71029675e-01 3.45620632e-01
-1.94560513e-01 2.51828492e-01 6.66026652e-01 4.93805259e-01
-2.70028710e-01 -1.21620297e-01 -1.23766792e+00 3.24038416e-01
1.18105769e+00 -6.04168139e-02 -2.83911675e-01 3.36899698e-01
5.95435858e-01 -6.96601570e-02 -8.17023337e-01 3.70431364e-01
7.02991843e-01 -9.52082276e-01 1.22047567e+00 -5.94721556e-01
3.67934495e-01 7.95980841e-02 -5.28129637e-01 -1.11226654e+00
-6.44661367e-01 -7.84630716e-01 -5.92875123e-01 7.10053384e-01
6.41827226e-01 -7.45795071e-01 1.07683933e+00 6.38075590e-01
-1.78273380e-01 -1.28641486e+00 -1.16938555e+00 -1.00801253e+00
-2.47438625e-01 -3.67469400e-01 2.81207472e-01 8.98636043e-01
-3.17454971e-02 3.15282196e-01 -4.61777419e-01 1.08402357e-01
8.44240606e-01 2.74792701e-01 2.97999650e-01 -1.52032578e+00
-6.66087627e-01 -6.21046424e-01 -5.14598250e-01 -9.93832946e-01
3.36286157e-01 -9.53737259e-01 -2.68568903e-01 -8.62286925e-01
1.50766790e-01 -7.09152400e-01 -4.17877108e-01 6.73656464e-01
3.34070653e-01 3.36742967e-01 9.97587219e-02 1.88989356e-01
-5.05964041e-01 7.03443468e-01 8.82355809e-01 4.70370576e-02
-4.81268555e-01 2.73287416e-01 -1.02209580e+00 8.92821133e-01
4.61131722e-01 -6.11450255e-01 -1.52425095e-01 -1.28492951e-01
5.19472599e-01 1.91577122e-01 4.97125059e-01 -1.13384271e+00
3.01374286e-01 9.44704656e-03 5.83211660e-01 -3.79007339e-01
5.58663905e-01 -9.84703064e-01 -3.83809917e-02 4.53710496e-01
-2.28472516e-01 3.65403704e-02 3.04673575e-02 6.04416966e-01
-8.72441605e-02 -3.36854011e-01 7.05064058e-01 -1.28221929e-01
-4.22013164e-01 4.63293076e-01 -3.19585472e-01 -1.27564415e-01
5.83944082e-01 -4.74801600e-01 -1.96701303e-01 -3.10199529e-01
-8.44817638e-01 1.33376241e-01 8.46735388e-02 -1.61756083e-01
8.30432594e-01 -1.19024575e+00 -6.03417575e-01 6.53474450e-01
-5.04156470e-01 -1.15887806e-01 2.50201941e-01 1.13228047e+00
4.63919155e-02 3.73856068e-01 -2.32934222e-01 -5.64811468e-01
-1.01223409e+00 4.73619640e-01 1.56962395e-01 -2.34073654e-01
-6.33975387e-01 1.15777016e+00 2.43542254e-01 -6.56177774e-02
4.59755272e-01 -3.46519113e-01 7.73193538e-02 7.44010359e-02
4.06956166e-01 4.25289452e-01 4.05959375e-02 -3.12103331e-01
-4.17002201e-01 4.63698983e-01 9.30000171e-02 3.31257522e-01
1.92580891e+00 1.06864527e-01 -4.51718830e-02 1.16294339e-01
1.40377676e+00 2.22115088e-02 -1.24327946e+00 -2.28865162e-01
-1.49320021e-01 -2.61648893e-01 2.28966832e-01 -4.28496778e-01
-1.47968531e+00 5.69508731e-01 3.56400371e-01 3.13668221e-01
1.20513332e+00 -1.02230951e-01 5.24498880e-01 7.94355631e-01
4.73984182e-01 -7.71563947e-01 -1.65321738e-01 6.24633551e-01
9.06386554e-01 -9.52441752e-01 8.38506669e-02 -4.85885471e-01
-3.32516819e-01 1.13074732e+00 3.59519243e-01 -6.96305335e-01
8.05678666e-01 5.80085754e-01 -7.56463349e-01 -3.55200917e-01
-8.69658530e-01 -1.36452436e-01 3.00221801e-01 5.14047027e-01
1.66289866e-01 -1.98819920e-01 -1.61679372e-01 1.02420747e+00
-1.92133859e-01 5.54195344e-02 3.27581048e-01 5.68274140e-01
-6.09659612e-01 -7.51341939e-01 -3.75107974e-01 1.02528965e+00
-3.66335660e-01 -4.47344184e-01 -1.32576808e-01 8.06090355e-01
-2.92467266e-01 5.13018548e-01 -1.50969982e-01 -4.29312110e-01
1.44877791e-01 5.55777401e-02 4.77514148e-01 -3.89982492e-01
-6.97516680e-01 1.34616002e-01 -8.31844956e-02 -6.95077658e-01
-1.12115763e-01 -7.01877475e-01 -1.12652457e+00 -1.27179816e-01
-4.30518717e-01 1.29591450e-01 4.76542145e-01 8.79142761e-01
5.87118149e-01 2.96005607e-01 5.95936775e-01 -1.06327379e+00
-8.52421522e-01 -6.87125862e-01 -9.68417585e-01 7.78526440e-02
3.29226613e-01 -9.78187978e-01 -6.61386311e-01 -3.34620148e-01] | [8.264572143554688, 3.739588737487793] |
58597401-6466-47ce-a031-015406caab72 | cross-domain-few-shot-segmentation-with | 2211.14745 | null | https://arxiv.org/abs/2211.14745v1 | https://arxiv.org/pdf/2211.14745v1.pdf | Cross-domain Few-shot Segmentation with Transductive Fine-tuning | Few-shot segmentation (FSS) expects models trained on base classes to work on novel classes with the help of a few support images. However, when there exists a domain gap between the base and novel classes, the state-of-the-art FSS methods may even fail to segment simple objects. To improve their performance on unseen domains, we propose to transductively fine-tune the base model on a set of query images under the few-shot setting, where the core idea is to implicitly guide the segmentation of query images using support labels. Although different images are not directly comparable, their class-wise prototypes are desired to be aligned in the feature space. By aligning query and support prototypes with an uncertainty-aware contrastive loss, and using a supervised cross-entropy loss and an unsupervised boundary loss as regularizations, our method could generalize the base model to the target domain without additional labels. We conduct extensive experiments under various cross-domain settings of natural, remote sensing, and medical images. The results show that our method could consistently and significantly improve the performance of prototypical FSS models in all cross-domain tasks. | ['Song Wang', 'Zhenyao Wu', 'Xinyi Wu', 'Yuhang Lu'] | 2022-11-27 | null | null | null | null | ['cross-domain-few-shot'] | ['computer-vision'] | [ 5.90269089e-01 3.96097898e-01 -3.57480705e-01 -8.67683113e-01
-9.93521094e-01 -5.58828056e-01 4.61996436e-01 -2.51411181e-02
-2.92783082e-01 6.13766789e-01 -3.06417197e-01 1.76707417e-01
-1.08513877e-01 -8.19716752e-01 -7.95648515e-01 -6.62407100e-01
3.55778635e-01 6.85188711e-01 6.87325895e-01 -2.23698840e-02
-4.35279496e-02 1.34776235e-01 -1.33540952e+00 2.46583670e-01
1.35775042e+00 9.04753089e-01 2.11826533e-01 1.60729602e-01
-1.48017064e-01 1.94250762e-01 -2.85989076e-01 -4.02317941e-01
5.76368451e-01 -4.93965119e-01 -1.05793273e+00 6.56461239e-01
4.62065846e-01 -1.74795568e-01 -6.51882663e-02 1.29442692e+00
2.64246285e-01 5.19161582e-01 1.04209459e+00 -1.09580612e+00
-9.22195494e-01 4.35164660e-01 -6.45456254e-01 -5.69721833e-02
-1.20692011e-02 2.24938735e-01 1.02570009e+00 -8.09082687e-01
8.37713301e-01 1.18913352e+00 6.73828602e-01 6.95119202e-01
-1.59574533e+00 -5.60534358e-01 5.29329300e-01 7.94227645e-02
-1.48014891e+00 -3.33326161e-01 7.77345479e-01 -5.32737017e-01
3.93062025e-01 2.49963365e-02 3.46109688e-01 9.86320257e-01
-2.95535624e-01 9.98175800e-01 1.02755582e+00 -3.09643000e-01
5.57509243e-01 4.86932904e-01 3.56648207e-01 5.73192179e-01
6.24054931e-02 -9.59448144e-02 -1.58399805e-01 -8.88187513e-02
4.13928688e-01 1.60447031e-01 -2.14462936e-01 -8.71558070e-01
-1.01378167e+00 1.10818672e+00 6.88127100e-01 2.70596802e-01
-1.14637159e-01 -3.44639212e-01 2.44201496e-01 7.97831193e-02
8.10827315e-01 5.34644604e-01 -3.88253391e-01 5.00961602e-01
-1.11842263e+00 6.77418336e-02 5.67268312e-01 1.18756843e+00
1.07438684e+00 -3.68828714e-01 -4.90931064e-01 1.06527340e+00
1.20991327e-01 3.21894825e-01 4.02755678e-01 -9.85632122e-01
2.63270915e-01 6.03761077e-01 7.59905204e-02 -6.88863158e-01
-1.80881903e-01 -4.12014425e-01 -6.94626272e-01 -2.34600112e-01
2.39529029e-01 -9.16557834e-02 -1.54380023e+00 1.77638769e+00
5.81382930e-01 3.34062845e-01 1.18156709e-01 8.48540664e-01
6.42990768e-01 6.53042555e-01 1.24048576e-01 -2.31617376e-01
9.55871701e-01 -1.05676317e+00 -4.37699288e-01 -5.75242221e-01
6.30917609e-01 -3.40826720e-01 1.44674253e+00 -1.72915965e-01
-7.61501849e-01 -5.80133677e-01 -1.14241099e+00 6.10644445e-02
-4.10515785e-01 -2.31465712e-01 3.44994932e-01 4.65505272e-01
-6.25261843e-01 6.33087277e-01 -8.28620017e-01 -5.94919026e-01
8.94435108e-01 -5.64252958e-02 -6.91474527e-02 -4.71693277e-01
-1.24551046e+00 6.29983783e-01 6.85883403e-01 -2.58506656e-01
-9.47837293e-01 -7.97605515e-01 -9.82108474e-01 -4.62290645e-02
6.73159003e-01 -7.08094835e-01 1.16590273e+00 -1.17976165e+00
-1.18137050e+00 1.17126560e+00 1.26388595e-01 -3.37777972e-01
5.25993705e-01 2.58652885e-02 -2.04752609e-01 3.29821110e-01
6.49545312e-01 1.12484062e+00 9.90488589e-01 -1.50178802e+00
-7.08328307e-01 -3.69500011e-01 3.78633328e-02 3.55408072e-01
-2.93445200e-01 -4.78277713e-01 -5.20620525e-01 -6.00425482e-01
3.20611924e-01 -9.53850150e-01 -3.48897398e-01 2.67581880e-01
-5.07951081e-01 -2.09458217e-01 7.65785158e-01 -2.69892126e-01
7.92542815e-01 -2.34648561e+00 -1.53385112e-02 4.37039975e-03
-1.22694932e-01 2.03578055e-01 -2.61903256e-01 -1.36332110e-01
2.84854412e-01 2.39730310e-02 -9.85363364e-01 -3.15586440e-02
-1.10453151e-01 5.37259161e-01 -4.58867252e-01 3.65677863e-01
5.38791895e-01 8.29292178e-01 -1.19696915e+00 -7.49243677e-01
-4.22693649e-03 3.24335955e-02 -4.63362366e-01 1.97778270e-01
-5.62484980e-01 4.87159401e-01 -6.25902295e-01 6.15336299e-01
7.24274218e-01 -5.17111897e-01 -7.56285042e-02 1.78322308e-02
3.72393548e-01 -8.28000680e-02 -1.00733089e+00 1.95342124e+00
-1.34793743e-01 3.39917541e-01 -1.32301718e-01 -1.50276673e+00
9.07908320e-01 -9.70520452e-02 4.12711293e-01 -4.08363938e-01
-1.37541741e-01 2.03885481e-01 -1.17741093e-01 -4.95374233e-01
1.13981299e-01 -5.47726452e-01 -1.41554594e-01 2.61085391e-01
3.20883751e-01 -4.54343319e-01 1.29275218e-01 1.55309245e-01
7.29718983e-01 1.18344359e-01 1.87470615e-01 -2.42181018e-01
1.23793639e-01 2.52094835e-01 8.72891009e-01 1.02147758e+00
-5.04918158e-01 9.97069359e-01 2.16981158e-01 -1.73941806e-01
-8.49745214e-01 -1.35597599e+00 -4.94663417e-01 1.23529077e+00
5.79998016e-01 2.67265707e-01 -7.71038473e-01 -1.22598398e+00
-2.43554376e-02 8.91651571e-01 -7.61907101e-01 -4.05451030e-01
-1.46650404e-01 -8.26175690e-01 3.63379538e-01 4.69563723e-01
7.26766050e-01 -8.26842606e-01 -5.01911283e-01 3.00673068e-01
-3.12526464e-01 -1.03456831e+00 -6.41709149e-01 2.50455201e-01
-9.48374748e-01 -9.79224741e-01 -1.11155927e+00 -1.08420312e+00
8.45483720e-01 3.69254947e-01 9.28800106e-01 -3.95027488e-01
-3.39625061e-01 3.27900618e-01 -4.32296604e-01 -1.90687895e-01
-1.97773650e-01 1.31254494e-01 -2.92933255e-01 1.59615681e-01
4.67799425e-01 -3.40067327e-01 -5.80540955e-01 4.28862959e-01
-1.04039621e+00 4.94522974e-02 2.74381697e-01 1.09977043e+00
8.97904277e-01 -2.02653348e-01 7.34771192e-01 -1.17419565e+00
1.36583105e-01 -6.14743948e-01 -3.39968085e-01 6.10032439e-01
-6.68648243e-01 5.21877930e-02 3.43706846e-01 -6.54454470e-01
-1.19706488e+00 1.81735918e-01 2.89921641e-01 -5.31491578e-01
-2.05931827e-01 4.06924605e-01 -1.72906369e-01 2.87368655e-01
1.00056469e+00 1.96483493e-01 -2.04550594e-01 -3.74868363e-01
5.78681052e-01 7.38786876e-01 4.07929838e-01 -6.56933725e-01
5.57123661e-01 6.65345192e-01 -4.56537217e-01 -7.62654424e-01
-1.47312725e+00 -6.83289051e-01 -7.49905407e-01 4.38031591e-02
9.42027986e-01 -8.88469815e-01 3.91534179e-01 3.76248837e-01
-9.39543605e-01 -2.91706175e-01 -6.84699953e-01 2.78729945e-01
-6.62650824e-01 3.56827736e-01 -2.74793684e-01 -5.31762421e-01
-9.79766622e-02 -1.05047607e+00 1.12190032e+00 4.25030887e-01
4.88524772e-02 -9.08341646e-01 -6.76611289e-02 3.81496251e-01
5.91860227e-02 2.33296663e-01 9.59457099e-01 -1.02331483e+00
-4.06557620e-01 -1.26148790e-01 -4.03034627e-01 5.98029137e-01
3.09403241e-01 -2.53045589e-01 -1.00792277e+00 -2.16716617e-01
1.37619674e-01 -6.93441391e-01 1.32011461e+00 5.06260455e-01
1.22228968e+00 -1.11152507e-01 -5.90548933e-01 6.67704582e-01
1.26253319e+00 3.17643851e-01 4.69037563e-01 6.04314692e-02
4.86802667e-01 7.72128642e-01 9.58062947e-01 2.63434350e-01
2.55508035e-01 4.33706641e-01 1.86277270e-01 -1.70864481e-02
2.68173963e-02 -2.47880116e-01 7.52034187e-02 4.15754914e-01
4.62183535e-01 -1.89601704e-01 -9.52414870e-01 9.28178310e-01
-1.88310766e+00 -7.95973301e-01 5.22009969e-01 2.14203095e+00
1.13941133e+00 2.05288887e-01 -7.58144483e-02 -4.36690181e-01
1.07056057e+00 2.08588958e-01 -1.21113706e+00 1.84086394e-02
-4.26528491e-02 -3.60075161e-02 3.36525589e-01 2.80467272e-01
-1.51476920e+00 1.13356483e+00 6.00962830e+00 1.02235329e+00
-1.05717635e+00 1.67932257e-01 9.78820324e-01 2.18631346e-02
-3.06246102e-01 1.21507831e-01 -7.93127716e-01 3.81771445e-01
5.76752663e-01 -2.42030874e-01 1.49134204e-01 1.03434503e+00
-2.94989020e-01 -6.98859394e-02 -1.31269073e+00 7.23706961e-01
5.99537715e-02 -1.22028899e+00 4.12385687e-02 -2.68076122e-01
1.06709766e+00 2.06648052e-01 1.48805931e-01 4.81712759e-01
3.23783249e-01 -7.57138312e-01 4.45657283e-01 3.36036801e-01
8.87607574e-01 -3.73523802e-01 4.15941417e-01 4.63512897e-01
-8.17557275e-01 4.40860493e-03 -5.93980491e-01 2.92540222e-01
7.30647072e-02 5.71205914e-01 -8.74113560e-01 4.29648012e-01
6.77083313e-01 9.61325824e-01 -4.46365446e-01 9.73105788e-01
1.98243354e-02 5.22559941e-01 -3.79293978e-01 2.21159369e-01
5.28257608e-01 -3.51985127e-01 4.67944294e-01 1.09090579e+00
1.56369969e-01 2.89009392e-01 5.27219772e-01 1.13894773e+00
-1.93602532e-01 -5.55840842e-02 -5.58765590e-01 -5.52595258e-02
4.47848856e-01 9.13656116e-01 -9.36838508e-01 -6.41955137e-01
-3.63318384e-01 1.02929056e+00 2.20742330e-01 4.66738224e-01
-7.64241874e-01 -6.00314617e-01 5.27684569e-01 9.65214148e-02
4.28339690e-01 3.17463964e-01 -4.83990252e-01 -1.27367556e+00
3.03283613e-02 -4.73454148e-01 5.81415534e-01 -5.66878974e-01
-1.85099244e+00 3.78044069e-01 2.07973495e-01 -1.42335546e+00
-1.92152653e-02 -2.84586400e-01 -4.22818542e-01 5.39280117e-01
-1.54900682e+00 -1.21102285e+00 2.20526960e-02 5.40339231e-01
7.68567383e-01 2.86749117e-02 6.41213536e-01 1.15215413e-01
-3.16959083e-01 4.95486349e-01 5.47801435e-01 1.88826591e-01
8.94590795e-01 -1.12412310e+00 1.41723663e-01 7.38977909e-01
5.23038432e-02 3.54073912e-01 5.97637773e-01 -6.33987904e-01
-4.02924895e-01 -1.54397488e+00 4.82374787e-01 -2.06200540e-01
4.11161065e-01 -2.53553689e-01 -1.31465101e+00 7.14514613e-01
-6.62921518e-02 4.94025826e-01 7.01653540e-01 3.00251832e-03
-5.55395663e-01 -3.73357348e-02 -1.45123899e+00 3.76312643e-01
1.11646140e+00 -4.24114704e-01 -1.03864670e+00 6.52648926e-01
9.97887611e-01 -2.45388597e-01 -6.26510024e-01 6.12230539e-01
2.43342757e-01 -6.49567187e-01 9.03963566e-01 -8.83025467e-01
3.85316283e-01 -2.41982594e-01 -3.11518818e-01 -1.39216840e+00
-2.68070638e-01 -8.50583538e-02 2.74126410e-01 1.20378613e+00
5.05230069e-01 -4.91592705e-01 7.11963832e-01 7.38629043e-01
-3.30759734e-01 -5.80621421e-01 -9.32176530e-01 -1.21407831e+00
2.94974595e-01 -2.24672496e-01 4.21448678e-01 1.11791027e+00
-2.44270191e-01 4.63418722e-01 -8.14271942e-02 1.50500059e-01
7.63993025e-01 5.15354395e-01 3.63573402e-01 -1.38456786e+00
-2.45712385e-01 -2.52431214e-01 -2.34271765e-01 -1.12753844e+00
2.66882211e-01 -9.77490485e-01 5.63117623e-01 -1.57819688e+00
4.78557140e-01 -7.23059177e-01 -3.91623259e-01 5.59832811e-01
-3.03639174e-01 1.12680614e-01 6.77026212e-02 2.14742035e-01
-8.91165793e-01 8.96285057e-01 1.38090968e+00 -5.76315224e-01
-4.38922971e-01 1.48253307e-01 -7.60410011e-01 8.11135590e-01
6.23926640e-01 -7.45155036e-01 -6.35739148e-01 -3.74773830e-01
-2.27222502e-01 -2.60654032e-01 1.46725431e-01 -8.79773200e-01
3.26826945e-02 -5.89195967e-01 2.17237711e-01 -3.57200146e-01
2.73525327e-01 -6.85087204e-01 -1.55446067e-01 3.22949708e-01
-5.68075657e-01 -9.73046303e-01 -9.57146361e-02 9.28872585e-01
-1.50149822e-01 -3.22447628e-01 1.33284211e+00 -1.70431793e-01
-8.45601916e-01 5.67070961e-01 9.02598277e-02 6.11145914e-01
1.29183948e+00 -4.14658576e-01 -1.95691422e-01 3.71004716e-02
-9.48981822e-01 4.49038595e-01 5.93637109e-01 4.56894636e-01
5.80729783e-01 -1.14748478e+00 -6.63496614e-01 1.28147975e-01
6.22758687e-01 5.95540702e-01 4.18146551e-01 4.73677069e-01
-1.40363097e-01 1.11310378e-01 -7.32361674e-02 -1.02047157e+00
-7.53140867e-01 6.86286628e-01 4.72919196e-01 -6.17319420e-02
-5.51804900e-01 1.07371008e+00 8.39408755e-01 -6.29148901e-01
2.24382982e-01 -2.33016342e-01 2.73563564e-02 2.24668771e-01
2.96017379e-01 1.65379688e-01 -2.01924220e-01 -4.57578927e-01
-3.27928722e-01 5.33535838e-01 -3.63293946e-01 2.77708042e-02
1.31268787e+00 -3.03663254e-01 2.93818951e-01 6.74896777e-01
1.24181736e+00 -7.09898353e-01 -1.72714114e+00 -6.81643069e-01
1.62871137e-01 -4.79331404e-01 -1.56587735e-01 -7.74607241e-01
-9.44875777e-01 9.28796291e-01 7.39982486e-01 1.39100505e-02
1.11184049e+00 4.53998178e-01 7.94116080e-01 4.42668676e-01
2.18015611e-01 -1.47924256e+00 2.25376368e-01 4.80828911e-01
5.77405035e-01 -1.69891965e+00 -3.19129258e-01 -5.06611168e-01
-8.94003153e-01 7.16543555e-01 6.79989398e-01 -4.73652482e-02
8.07626188e-01 -3.25669199e-01 -1.81247100e-01 -9.50098187e-02
-4.14034396e-01 -3.72052848e-01 3.57562751e-01 9.00107503e-01
-4.84405681e-02 1.28268078e-01 -1.39002934e-01 8.00328195e-01
4.14027333e-01 1.13019086e-01 3.99467826e-01 9.22911704e-01
-7.59259522e-01 -6.31838322e-01 -1.72126114e-01 7.51840949e-01
-9.48015600e-02 3.11928336e-02 -2.20097348e-01 5.65440834e-01
2.54883677e-01 8.49773467e-01 1.31758317e-01 -2.33268663e-01
2.78936177e-01 2.10064113e-01 3.57000858e-01 -1.17146850e+00
-6.22566184e-03 -7.10257888e-02 -3.15002024e-01 -1.49390489e-01
-5.84669769e-01 -7.32190847e-01 -1.33032441e+00 2.56927788e-01
-5.29900551e-01 -9.82903782e-03 2.34035119e-01 1.11948705e+00
3.68295968e-01 2.11128101e-01 7.36378312e-01 -5.37261546e-01
-8.37657869e-01 -7.96733081e-01 -8.54240656e-01 7.30217099e-01
2.44992808e-01 -7.87082255e-01 -3.50523859e-01 3.94780219e-01] | [9.638983726501465, 1.4004461765289307] |
8e70cfb8-37d2-45d3-b558-286efe48302a | through-wall-human-pose-estimation-using | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Zhao_Through-Wall_Human_Pose_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Zhao_Through-Wall_Human_Pose_CVPR_2018_paper.pdf | Through-Wall Human Pose Estimation Using Radio Signals | This paper demonstrates accurate human pose estimation through walls and occlusions. We leverage the fact that wireless signals in the WiFi frequencies traverse walls and reflect off the human body. We introduce a deep neural network approach that parses such radio signals to estimate 2D poses. Since humans cannot annotate radio signals, we use state-of-the-art vision model to provide cross-modal supervision. Specifically, during training the system uses synchronized wireless and visual inputs, extracts pose information from the visual stream, and uses it to guide the training process. Once trained, the network uses only the wireless signal for pose estimation. We show that, when tested on visible scenes, the radio-based system is almost as accurate as the vision-based system used to train it. Yet, unlike vision-based pose estimation, the radio-based system can estimate 2D poses through walls despite never trained on such scenarios. Demo videos are available at our website (http://rfpose.csail.mit.edu). | ['Ming-Min Zhao', 'Tianhong Li', 'Hang Zhao', 'Antonio Torralba', 'Dina Katabi', 'Yonglong Tian', 'Mohammad Abu Alsheikh'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['rf-based-pose-estimation'] | ['computer-vision'] | [ 1.58222064e-01 2.91386694e-01 -7.31001571e-02 -3.07611138e-01
-7.93807387e-01 -6.53574944e-01 7.04394802e-02 -3.31988245e-01
-5.31474829e-01 4.84765381e-01 1.86346292e-01 -1.24745257e-01
3.07650894e-01 -6.48998022e-01 -1.18188500e+00 -1.33810148e-01
-2.77141899e-01 4.95906353e-01 -1.28682390e-01 -5.34407310e-02
-4.68243808e-01 1.63704351e-01 -1.06733191e+00 -5.06699570e-02
-3.33388038e-02 1.09495401e+00 -7.77720660e-02 1.27747071e+00
4.70210403e-01 5.21912277e-01 -6.71322107e-01 -1.01302490e-01
4.02851731e-01 1.04656562e-01 -4.37512964e-01 -1.16082929e-01
8.22155237e-01 -8.10701847e-01 -9.89293098e-01 5.01133084e-01
8.61022830e-01 -1.65469989e-01 1.99025944e-01 -1.20778275e+00
-2.89595011e-03 4.32773024e-01 -5.64699948e-01 -2.56473988e-01
1.06127036e+00 1.52265638e-01 6.48180485e-01 -7.79079854e-01
5.65449715e-01 9.54664350e-01 1.17656386e+00 5.93101442e-01
-9.92832661e-01 -6.59771979e-01 1.21573947e-01 -4.30789530e-01
-1.36557794e+00 -6.07892811e-01 6.79958224e-01 -2.90466934e-01
5.98896265e-01 1.77367851e-01 1.02888191e+00 1.83752441e+00
4.32510190e-02 7.44364738e-01 5.48764288e-01 -4.23518777e-01
3.70541848e-02 -2.29096621e-01 -7.78131485e-02 1.02513301e+00
2.88119912e-01 1.27840877e-01 -1.06651688e+00 -1.37453303e-01
1.10078025e+00 4.73432951e-02 -5.88158369e-01 -5.66283524e-01
-1.32065749e+00 2.02049106e-01 6.94742322e-01 1.96685456e-03
-3.36198539e-01 1.13388431e+00 -5.14297560e-02 2.05087550e-02
5.71845807e-02 2.06159979e-01 -4.04466659e-01 -2.86300838e-01
-1.15127254e+00 -6.13380373e-02 8.27455819e-01 9.61117864e-01
3.99126947e-01 2.82853972e-02 1.79892421e-01 5.60198784e-01
8.71724904e-01 1.16704655e+00 1.30116520e-02 -1.22663128e+00
7.65495360e-01 -2.92003959e-01 3.98461938e-01 -1.13872278e+00
-8.36406589e-01 -6.97399080e-01 -3.36359054e-01 1.15655744e-02
5.28262973e-01 -8.00773561e-01 -1.12805045e+00 1.90463436e+00
4.27990288e-01 2.41551280e-01 -2.53865957e-01 1.20234275e+00
8.08989167e-01 1.50068372e-01 -1.15307145e-01 3.53110701e-01
1.21612418e+00 -7.01295972e-01 -4.45948482e-01 -8.23482931e-01
3.28106254e-01 -6.31618559e-01 6.42473459e-01 5.15961826e-01
-1.03662360e+00 -7.56399572e-01 -1.49899518e+00 2.69994706e-01
1.13058090e-01 3.84531498e-01 7.40631819e-01 7.88378298e-01
-1.00812900e+00 3.24518532e-01 -1.31284654e+00 -7.69159973e-01
2.10828885e-01 5.38477421e-01 -3.43339205e-01 -1.18318178e-01
-1.33772564e+00 6.25991523e-01 -2.20323503e-01 3.78009975e-01
-1.23616016e+00 -2.91197449e-01 -1.00066638e+00 -3.23396623e-01
1.82078168e-01 -1.12622380e+00 1.62848616e+00 -8.12902629e-01
-1.31399786e+00 5.75595915e-01 -2.35575978e-02 -4.44484532e-01
7.56618440e-01 -7.98244119e-01 -3.50655854e-01 5.18144369e-01
1.09724581e-01 7.59550452e-01 9.29775476e-01 -1.54390061e+00
-3.93772006e-01 -4.31376785e-01 4.55285966e-01 1.82156146e-01
-3.63403820e-02 -6.89901114e-01 -1.02022326e+00 -4.32456672e-01
4.72143233e-01 -1.27583003e+00 -1.00021787e-01 4.46891934e-01
-6.01421714e-01 7.03180552e-01 5.86292803e-01 -7.72600651e-01
6.97226524e-01 -2.06270409e+00 -2.91105062e-01 7.14038730e-01
3.47132832e-01 -2.54992634e-01 7.59586468e-02 1.96819231e-01
2.99841076e-01 -2.10258082e-01 1.77010730e-01 -7.57156372e-01
5.41819222e-02 -1.24172373e-02 4.83741499e-02 8.59799743e-01
-4.42430854e-01 6.50507867e-01 -8.05700898e-01 -4.08832192e-01
3.05495054e-01 8.50744128e-01 -4.34491426e-01 6.21539764e-02
1.93183199e-01 5.63824952e-01 -5.25434077e-01 7.31028676e-01
5.06272674e-01 -2.23538086e-01 6.04221702e-01 -5.10652125e-01
3.80967230e-01 6.29148930e-02 -1.19279754e+00 2.23147869e+00
-8.95353913e-01 8.32595229e-01 5.24988770e-01 -7.16252863e-01
5.92652857e-01 4.79973853e-01 5.95094860e-01 -5.62522173e-01
3.07043552e-01 -1.58795670e-01 -7.06325650e-01 -5.79673648e-01
2.65046626e-01 3.02661240e-01 -5.73974729e-01 1.90897495e-01
1.04990639e-01 2.83016413e-02 -3.58004630e-01 2.39247605e-01
1.55823934e+00 7.51945555e-01 -1.19530782e-01 6.17874920e-01
7.56922504e-03 -1.53360730e-02 2.74235576e-01 1.05504382e+00
-9.71577689e-02 7.62262523e-01 -2.28931144e-01 -4.04427767e-01
-7.60464430e-01 -1.62146747e+00 1.43041521e-01 1.13402057e+00
4.65605140e-01 -4.92497295e-01 -6.86911702e-01 -7.11674392e-01
2.11008981e-01 1.25567108e-01 -5.62486827e-01 7.70364478e-02
-5.93821108e-01 -7.21892118e-02 9.84449863e-01 8.13623130e-01
6.27763391e-01 -3.63043100e-01 -8.41109931e-01 2.07454085e-01
-5.92376232e-01 -1.28894830e+00 -2.80317843e-01 2.03847274e-01
-4.10590798e-01 -1.05914712e+00 -8.31425786e-01 -4.91508991e-01
5.92169106e-01 3.27669412e-01 1.06728137e+00 1.94786042e-01
-3.26719195e-01 1.35048699e+00 4.28250106e-03 -3.53106678e-01
1.72995239e-01 -1.95584707e-02 2.23509267e-01 -2.69385785e-01
3.28768715e-02 -7.20990181e-01 -6.92407548e-01 2.46859372e-01
1.23851269e-01 -1.75319150e-01 5.52586734e-01 4.80455577e-01
3.78816932e-01 -4.81294811e-01 -6.67411163e-02 -5.65082967e-01
3.81390899e-02 -3.87011707e-01 -4.58771646e-01 1.35995392e-02
9.41019580e-02 -1.96769565e-01 3.43314499e-01 -2.37997994e-01
-8.62481296e-01 6.89489543e-01 -1.39711294e-02 -5.45507610e-01
-3.94488186e-01 3.73123229e-01 4.04727049e-02 -1.33052319e-01
8.17126274e-01 -1.70276567e-01 -4.19721276e-01 -3.07878494e-01
2.58586466e-01 6.89762890e-01 9.64834690e-01 -4.62496608e-01
1.03697896e+00 8.80407691e-01 -3.65841538e-02 -9.62934315e-01
-9.89378870e-01 -3.98221135e-01 -5.35546482e-01 -7.14328885e-01
1.01489723e+00 -1.37882090e+00 -9.98937011e-01 1.26793489e-01
-1.17955029e+00 -3.46526086e-01 2.72490114e-01 8.76507640e-01
-6.74627185e-01 1.19033746e-01 -6.08501375e-01 -1.03209889e+00
-2.12104395e-01 -7.00701535e-01 1.43020248e+00 1.61802799e-01
-6.80407345e-01 -6.84588671e-01 1.64054170e-01 7.98325717e-01
4.99092937e-02 4.87940639e-01 -6.81321248e-02 6.48183823e-02
-5.46917200e-01 -4.78220731e-01 2.38384694e-01 -3.63978177e-01
-2.55685031e-01 -3.29976469e-01 -1.40692055e+00 -4.24605370e-01
-5.33075988e-01 -5.09139955e-01 5.86266220e-01 6.55530751e-01
7.87928462e-01 9.97475684e-02 -8.06402922e-01 8.96823764e-01
1.16026568e+00 -2.81170040e-01 2.81752586e-01 3.76811236e-01
9.37004030e-01 2.00609148e-01 5.82690597e-01 5.32131374e-01
4.71102506e-01 8.31475139e-01 6.01968944e-01 -3.75925124e-01
-1.49101257e-01 -5.39598882e-01 4.32547480e-01 1.28178880e-01
-2.03452215e-01 -1.91840008e-01 -9.53630626e-01 1.36847168e-01
-1.79277885e+00 -8.27001631e-01 -3.03278640e-02 2.08138275e+00
3.49193245e-01 4.87117141e-01 1.25551656e-01 -4.82217781e-02
6.81431472e-01 1.52363807e-01 -4.03512836e-01 -7.58544654e-02
5.30535638e-01 1.75623640e-01 1.10395539e+00 6.15054071e-01
-1.46611249e+00 6.02776229e-01 6.42546844e+00 2.36420128e-02
-1.11541450e+00 1.29238084e-01 3.17233987e-02 -4.79740411e-01
1.69112850e-02 -3.27684134e-01 -5.36534846e-01 1.65043920e-01
8.74325514e-01 7.98201561e-01 3.98963451e-01 1.08124375e+00
2.90385127e-01 -2.03722551e-01 -1.18857241e+00 1.24619377e+00
2.21342575e-02 -9.96853411e-01 -7.74371922e-01 1.65781230e-01
2.40918398e-01 2.84334749e-01 4.90076141e-03 2.39264697e-01
4.35292631e-01 -9.22293961e-01 1.17424762e+00 6.17505610e-01
9.40944493e-01 -6.97771072e-01 5.75652122e-01 1.62166968e-01
-1.52394331e+00 1.59812778e-01 4.78030443e-02 -7.05864727e-02
3.63899499e-01 6.04030967e-01 -9.20259476e-01 4.23777819e-01
9.82839704e-01 4.82309073e-01 -4.50023055e-01 8.82494450e-01
-6.86041236e-01 6.55082762e-01 -8.22172344e-01 1.83656439e-01
-8.68561938e-02 5.41888237e-01 3.98650318e-01 1.03239119e+00
3.13860118e-01 -2.36998603e-01 7.17274427e-01 3.63422722e-01
-1.37181863e-01 -6.47382438e-01 -9.04647708e-01 3.04385304e-01
7.49173164e-01 1.14011252e+00 -5.90250194e-01 -4.80653197e-02
-3.27081442e-01 1.15112996e+00 3.97742912e-02 5.70338547e-01
-1.27391243e+00 -3.79811257e-01 4.79893506e-01 1.69974536e-01
2.95381337e-01 -7.24925041e-01 -9.75596085e-02 -9.84842479e-01
5.58196791e-02 -5.72035432e-01 5.13112433e-02 -1.10534406e+00
-6.69451714e-01 2.46272147e-01 -2.95893192e-01 -1.32358932e+00
-3.49662632e-01 -3.82267863e-01 -2.72346854e-01 3.96005690e-01
-8.77344728e-01 -1.47298217e+00 -8.47917438e-01 5.66631854e-01
1.99249670e-01 2.71071255e-01 8.63643706e-01 4.60933119e-01
-2.65511751e-01 7.51519501e-01 -2.48556450e-01 7.00460196e-01
7.59393692e-01 -1.03189647e+00 5.94359517e-01 6.96519554e-01
4.36170638e-01 7.80560076e-01 8.69703591e-01 -7.56154358e-01
-1.66940999e+00 -8.06140780e-01 2.79921860e-01 -5.66341877e-01
2.97421217e-01 -5.99738836e-01 1.11776583e-01 1.02945101e+00
-1.16154179e-01 2.11799487e-01 6.26369417e-01 3.12374234e-01
-2.58513212e-01 -2.24563569e-01 -9.51888084e-01 5.72192729e-01
1.30518079e+00 -6.07432306e-01 -4.90027905e-01 2.33559966e-01
4.54685777e-01 -7.78284192e-01 -5.91178775e-01 3.62513475e-02
1.27992177e+00 -6.52291536e-01 1.38126910e+00 -6.15621172e-02
3.54871829e-03 -3.38436633e-01 -3.74179363e-01 -1.09916651e+00
-1.37510896e-01 -4.85007793e-01 -3.49611938e-01 6.97613776e-01
4.71827984e-01 -2.75573999e-01 1.17561364e+00 3.64940912e-01
2.37085477e-01 -2.67752409e-01 -8.67010713e-01 -5.42514920e-01
-8.73568356e-01 -8.90825152e-01 1.61580503e-01 5.97372293e-01
-1.44970655e-01 3.53886366e-01 -7.44103611e-01 7.04760194e-01
9.57864463e-01 -2.52030164e-01 1.21126688e+00 -9.81108606e-01
-8.20188820e-01 4.43132937e-01 -4.43413138e-01 -1.50479293e+00
-2.55170137e-01 -3.81137699e-01 4.18378919e-01 -1.69396496e+00
-4.80748385e-01 -2.98487306e-01 8.21776539e-02 5.72192788e-01
4.33079481e-01 8.20667982e-01 2.70727694e-01 -7.11429194e-02
-7.55882859e-01 1.66418850e-01 7.65988708e-01 -3.40972900e-01
7.22584920e-03 2.35273570e-01 -3.38264763e-01 1.13696516e+00
8.66452515e-01 -4.35668677e-01 -2.16225758e-01 -8.13602924e-01
3.46808046e-01 2.70988196e-01 7.65136719e-01 -1.80883002e+00
3.37226927e-01 3.11678201e-01 1.37721825e+00 -5.01846910e-01
1.00309050e+00 -1.15206599e+00 1.68937922e-01 7.61636972e-01
-1.74462423e-01 -1.57032296e-01 8.27592015e-02 7.62492120e-01
4.14082944e-01 1.09685138e-01 3.24862868e-01 -2.63222605e-01
-4.52492684e-01 9.88832861e-02 -5.59385657e-01 -1.18220292e-01
5.88065505e-01 -2.77900636e-01 -1.49878502e-01 -1.04187310e+00
-1.08832586e+00 2.43607014e-01 3.76358926e-01 2.07435548e-01
6.40378475e-01 -1.41458976e+00 -2.37796139e-02 8.13156590e-02
-7.07092956e-02 -6.94506466e-02 8.14800635e-02 6.68052554e-01
-7.14410126e-01 1.73028946e-01 2.15098951e-02 -9.70292032e-01
-1.28562021e+00 -7.85574038e-03 5.38448095e-01 9.09759104e-02
-4.51861560e-01 9.39728677e-01 -1.41937330e-01 -4.10470009e-01
6.93489492e-01 -1.34131134e-01 1.66384250e-01 -2.69784123e-01
3.16476732e-01 2.67884523e-01 -3.82880047e-02 -4.63773310e-01
-7.39007354e-01 9.54483926e-01 4.95187342e-01 -8.35783899e-01
1.17587936e+00 -3.30499411e-01 8.21010470e-01 3.20280582e-01
1.13578963e+00 3.10790092e-01 -1.50476587e+00 -3.08626592e-02
-4.63511199e-01 -3.27813685e-01 2.18041599e-01 -8.19177330e-01
-9.64892924e-01 7.41798162e-01 8.52399826e-01 -2.48558726e-02
7.54988730e-01 -6.47877902e-02 8.98918688e-01 7.53765345e-01
8.67861569e-01 -1.36156607e+00 1.26753405e-01 3.87357533e-01
5.88780940e-01 -1.03352404e+00 1.45761460e-01 -3.32436979e-01
-2.08754897e-01 9.89570975e-01 6.90604806e-01 -1.39044046e-01
5.91417909e-01 7.32623816e-01 4.65768993e-01 -1.91033497e-01
-1.16489381e-01 -2.62590230e-01 -2.58546341e-02 1.09971082e+00
3.43536019e-01 -4.66284752e-02 5.24692118e-01 3.66658002e-01
-5.48318982e-01 7.16124028e-02 2.34423667e-01 1.32142770e+00
-4.95229214e-01 -5.63647687e-01 -9.47249472e-01 1.41071796e-01
-3.92868012e-01 2.42249459e-01 -3.89945328e-01 7.83612311e-01
2.62665421e-01 1.04714429e+00 -2.14146972e-01 -7.59490848e-01
3.82300526e-01 -1.49458617e-01 9.04696941e-01 -4.43020493e-01
-4.06287491e-01 1.82312950e-01 6.17109478e-01 -1.18469429e+00
-3.94994825e-01 -3.02774698e-01 -1.39442587e+00 -1.65267795e-01
-1.39234975e-01 -1.18307449e-01 1.01600075e+00 8.58928800e-01
1.36647955e-01 8.20552707e-01 5.48921764e-01 -1.44181275e+00
-1.91473559e-01 -5.79485297e-01 -3.39998186e-01 1.55242249e-01
5.68936884e-01 -7.48087406e-01 -1.13784291e-01 -4.39846376e-03] | [6.839484214782715, 0.26518136262893677] |
5fee27fc-1276-4b20-8cc2-ed416173a080 | interpretable-reinforcement-learning-via-1 | 2303.10382 | null | https://arxiv.org/abs/2303.10382v2 | https://arxiv.org/pdf/2303.10382v2.pdf | Interpretable Reinforcement Learning via Neural Additive Models for Inventory Management | The COVID-19 pandemic has highlighted the importance of supply chains and the role of digital management to react to dynamic changes in the environment. In this work, we focus on developing dynamic inventory ordering policies for a multi-echelon, i.e. multi-stage, supply chain. Traditional inventory optimization methods aim to determine a static reordering policy. Thus, these policies are not able to adjust to dynamic changes such as those observed during the COVID-19 crisis. On the other hand, conventional strategies offer the advantage of being interpretable, which is a crucial feature for supply chain managers in order to communicate decisions to their stakeholders. To address this limitation, we propose an interpretable reinforcement learning approach that aims to be as interpretable as the traditional static policies while being as flexible and environment-agnostic as other deep learning-based reinforcement learning solutions. We propose to use Neural Additive Models as an interpretable dynamic policy of a reinforcement learning agent, showing that this approach is competitive with a standard full connected policy. Finally, we use the interpretability property to gain insights into a complex ordering strategy for a simple, linear three-echelon inventory supply chain. | ['Johannes S. Otterbach', 'Sebastian Schulze', 'Maximilian Schambach', 'Julien Siems'] | 2023-03-18 | null | null | null | null | ['additive-models'] | ['methodology'] | [-1.43270582e-01 1.28799781e-01 -1.31177127e-01 -1.23016220e-02
-8.32914375e-03 -9.51839805e-01 3.88379186e-01 3.81454647e-01
-3.09920937e-01 8.28520179e-01 2.84244239e-01 -4.97237921e-01
-6.11150146e-01 -8.36799204e-01 -8.94774079e-01 -5.92304409e-01
-3.27360779e-01 1.18978500e+00 -1.71606302e-01 -6.44287765e-01
2.08684534e-01 7.42440403e-01 -1.07846427e+00 3.63953263e-02
5.13904989e-01 1.02384508e+00 3.45512688e-01 6.41941905e-01
6.16191849e-02 7.30606019e-01 -5.56398034e-01 -4.22753617e-02
5.06635904e-01 -1.56978462e-02 -4.80232209e-01 3.23625773e-01
-4.24336523e-01 -7.09107101e-01 -1.02676272e-01 8.82214427e-01
1.94718242e-01 6.93473220e-02 5.70164502e-01 -1.34092355e+00
-7.09774315e-01 9.74196792e-01 -4.76980321e-02 1.95370987e-01
-7.47182146e-02 4.19524372e-01 9.91240680e-01 -2.82841567e-02
4.72978562e-01 1.27497935e+00 3.52255464e-01 1.71179026e-01
-1.43546045e+00 -3.26163828e-01 6.78060651e-01 2.87848324e-01
-6.73295081e-01 1.14674374e-01 6.72711074e-01 -4.97937679e-01
8.58427048e-01 1.76036820e-01 1.04851055e+00 1.24881196e+00
7.76811779e-01 6.05233669e-01 1.32936549e+00 -1.71355814e-01
4.93972093e-01 2.71599777e-02 -3.14328313e-01 8.57919157e-02
5.08842230e-01 2.83140540e-01 1.49517432e-01 2.67957151e-01
7.82823563e-01 2.80628890e-01 -7.84661025e-02 -4.90641922e-01
-1.05996251e+00 1.03168356e+00 4.32185501e-01 2.37090170e-01
-9.02217567e-01 4.58628654e-01 4.01079386e-01 8.16872180e-01
2.37773970e-01 1.08287644e+00 -7.08292007e-01 -5.28597794e-02
-4.30328935e-01 4.21117276e-01 1.08882427e+00 7.94070423e-01
2.81572759e-01 3.63835454e-01 3.83489281e-02 -6.10219091e-02
3.23639423e-01 5.28646886e-01 -2.16920804e-02 -1.23247015e+00
3.19923878e-01 2.67807752e-01 6.46463156e-01 -8.94910336e-01
-9.05597031e-01 -5.73827803e-01 -8.16546977e-01 3.61854166e-01
3.09027910e-01 -3.68242800e-01 -7.45280504e-01 1.56102383e+00
-3.62696336e-03 -2.90928394e-01 6.30560070e-02 8.80049586e-01
-3.02397311e-01 7.55236089e-01 -1.49329349e-01 -4.27258670e-01
1.28212154e+00 -6.06926799e-01 -1.04641056e+00 -6.93492666e-02
1.08742207e-01 -3.62011999e-01 8.61205935e-01 7.72191048e-01
-1.02127564e+00 -1.29890263e-01 -1.06780267e+00 6.36039913e-01
-6.84441745e-01 -7.33938098e-01 4.70868587e-01 1.67076945e-01
-9.32696879e-01 6.79043412e-01 -8.36843908e-01 -2.44602069e-01
-1.20683648e-01 5.68543375e-01 9.32966694e-02 2.61761189e-01
-1.52484596e+00 1.31795633e+00 6.32307112e-01 2.34565869e-01
-9.57946599e-01 -5.08553982e-01 -4.89926010e-01 4.54392552e-01
9.67980027e-01 -4.93221313e-01 1.57089293e+00 -1.11970699e+00
-1.57548177e+00 -8.19594711e-02 6.46146774e-01 -6.51710093e-01
8.46359909e-01 -6.39177561e-02 -4.06733811e-01 5.49379794e-04
-3.02773446e-01 2.71079630e-01 5.82698584e-01 -1.57874167e+00
-7.25119412e-01 -2.40545839e-01 4.52371150e-01 1.02649786e-01
-3.90409715e-02 -3.09040219e-01 2.12749958e-01 -5.68514645e-01
-2.55317003e-01 -1.04487813e+00 -5.57293713e-01 -5.47948778e-01
-3.24510336e-01 1.36718929e-01 7.19714701e-01 -7.18936443e-01
9.49385047e-01 -1.61507273e+00 3.25875223e-01 3.20884228e-01
-3.70410159e-02 -1.69521496e-02 -1.34062245e-01 8.83083045e-01
1.17719546e-01 1.47160128e-01 2.27845907e-02 1.23672402e-02
4.84790266e-01 6.82360530e-01 -2.53401905e-01 2.19544038e-01
4.40925568e-01 8.76497090e-01 -1.03680551e+00 1.55276194e-01
1.43664852e-01 7.19765574e-02 -4.26278859e-01 2.64441878e-01
-9.60718155e-01 8.68752897e-01 -4.29218799e-01 4.99937207e-01
4.84058172e-01 -8.72532576e-02 8.12539339e-01 4.55594668e-03
-2.94724613e-01 -1.57966748e-01 -1.25049210e+00 9.77169991e-01
-7.28297055e-01 3.29070479e-01 3.74143958e-01 -1.06845737e+00
8.12307894e-01 2.80097038e-01 6.58463478e-01 -1.12033951e+00
2.18294516e-01 3.09648275e-01 3.08976680e-01 -5.20691752e-01
5.83372474e-01 -4.77580458e-01 -3.57215524e-01 7.00091422e-01
-4.38521236e-01 -3.47297713e-02 2.01572016e-01 -2.77982354e-01
9.64670002e-01 1.87794883e-02 2.98514962e-01 -3.60485673e-01
5.12950346e-02 -1.71540663e-01 5.22595108e-01 7.01365948e-01
-2.41956692e-02 7.38180652e-02 9.42585349e-01 -6.45465076e-01
-1.09783685e+00 -9.79889810e-01 2.65340865e-01 8.78666699e-01
1.84139326e-01 3.61106038e-01 -6.50648057e-01 -5.38096189e-01
3.21636558e-01 8.98454249e-01 -6.25123382e-01 -1.99680123e-02
-8.36088777e-01 -4.29636538e-01 -1.57297418e-01 6.74631238e-01
2.24723995e-01 -1.28443503e+00 -1.07153440e+00 8.53674710e-01
3.49126086e-02 -9.47601974e-01 -3.32240075e-01 7.29799271e-01
-5.08429468e-01 -1.09778035e+00 -3.63037527e-01 -4.00647193e-01
4.48815972e-01 -2.13756993e-01 1.08289039e+00 -2.54236013e-01
3.30358773e-01 5.02097130e-01 -1.67977095e-01 -4.98621345e-01
-7.86302388e-01 2.92530298e-01 3.27725619e-01 4.72248532e-02
-7.99921677e-02 -4.75620329e-01 -7.02608824e-01 2.51470059e-01
-1.29164219e+00 -1.49898604e-01 6.62041783e-01 8.23574662e-01
5.33981979e-01 2.61111289e-01 1.07702017e+00 -8.65284979e-01
1.00639868e+00 -6.39415503e-01 -1.12119877e+00 3.53400111e-01
-1.01816618e+00 3.75703961e-01 1.17341626e+00 -4.92830098e-01
-8.73085916e-01 -9.62140560e-02 5.23918569e-02 -2.26126954e-01
7.78015563e-03 6.65282190e-01 -2.73980796e-01 4.18805331e-01
-1.14938155e-01 2.03150272e-01 1.40959606e-01 -5.25012493e-01
3.56651545e-01 5.74816227e-01 2.62055159e-01 -4.24919695e-01
6.97549820e-01 2.30123162e-01 2.21577808e-01 -2.30424806e-01
-3.29906821e-01 1.64028257e-02 -8.25954914e-01 -2.69351900e-01
7.56035686e-01 -4.00087535e-01 -1.26832891e+00 2.08408847e-01
-1.12969220e+00 -5.83677351e-01 -5.63945889e-01 2.09522396e-01
-1.07876337e+00 -5.09988554e-02 -6.47027373e-01 -8.43713343e-01
-1.26177773e-01 -1.25170922e+00 5.57521343e-01 3.16212848e-02
-1.53188914e-01 -1.31700754e+00 2.09595144e-01 3.75905633e-02
8.34632277e-01 5.63685834e-01 1.26116192e+00 -9.51944172e-01
-7.13831365e-01 3.74205448e-02 9.89130884e-02 3.33645135e-01
3.21828097e-01 -2.22858414e-01 -3.26631159e-01 -6.36724055e-01
9.59780738e-02 2.21281103e-03 3.07713807e-01 3.72664154e-01
8.88712466e-01 -1.01381445e+00 5.97637743e-02 2.91430295e-01
1.48870397e+00 8.84110093e-01 2.67324567e-01 8.20541620e-01
3.27171355e-01 8.62856328e-01 8.34175527e-01 7.24002957e-01
5.76076806e-01 6.11955106e-01 1.06018913e+00 5.39688170e-02
4.73591089e-01 -2.38813415e-01 3.01669151e-01 6.50843501e-01
6.39840513e-02 -5.91298819e-01 -7.13252485e-01 3.13509345e-01
-2.02827358e+00 -9.60900009e-01 3.04182708e-01 1.89139557e+00
5.55432796e-01 4.84786481e-01 3.83996993e-01 -9.25378595e-03
5.29087603e-01 1.36746302e-01 -9.19646204e-01 -1.00707507e+00
-7.76593806e-03 -2.37782896e-01 7.89013505e-01 4.36925441e-01
-8.81823361e-01 5.95251918e-01 6.20770645e+00 7.50495642e-02
-1.24865079e+00 -1.06614046e-01 5.67973197e-01 -6.83608055e-02
-5.35562456e-01 -1.19636625e-01 -3.47085357e-01 7.07168519e-01
1.21162498e+00 1.75684150e-02 9.56012845e-01 4.54974979e-01
8.36207628e-01 7.81478584e-02 -1.21691692e+00 3.68768930e-01
-4.43522185e-01 -1.52190340e+00 -3.33721861e-02 2.67038256e-01
5.19260466e-01 -1.68363869e-01 1.22221902e-01 1.40453860e-01
5.95214605e-01 -1.09170794e+00 1.06459081e+00 6.41248226e-01
2.71585971e-01 -1.11217606e+00 9.82732534e-01 3.53280753e-01
-9.39292192e-01 -6.58195674e-01 6.35064989e-02 -1.81664571e-01
4.92232203e-01 -1.42472582e-02 -8.97993386e-01 5.58808684e-01
3.70647579e-01 2.80538559e-01 -5.29989824e-02 7.58142054e-01
2.84004491e-02 3.18787783e-01 -2.35973373e-01 -1.47025123e-01
6.20892525e-01 -3.32210243e-01 4.05491173e-01 1.08764899e+00
2.14285970e-01 -5.02173975e-02 3.97210628e-01 7.05450475e-01
8.20001736e-02 -3.96170259e-01 -6.37471378e-01 -1.82435691e-01
2.71603674e-01 8.20779860e-01 -8.83564115e-01 7.91957080e-02
-2.01865017e-01 6.68548822e-01 -1.10206837e-02 3.95699739e-01
-6.78660929e-01 -7.85880908e-02 6.15909517e-01 2.90874571e-01
6.08290136e-01 -3.65290910e-01 -4.26803939e-02 -7.42911220e-01
-1.46538049e-01 -1.19792902e+00 1.46376893e-01 -3.85252595e-01
-1.30299032e+00 3.12187821e-01 6.67717829e-02 -1.10978580e+00
-4.51421142e-01 -7.53900170e-01 -3.73925447e-01 4.12498713e-01
-1.85691214e+00 -1.02772462e+00 2.74682343e-01 2.31003866e-01
5.47696829e-01 -2.35924758e-02 4.72564608e-01 -1.93006378e-02
-4.11000043e-01 5.12464866e-02 7.02426553e-01 -1.76111564e-01
1.68800190e-01 -1.72501469e+00 1.60085618e-01 6.74810112e-01
-4.74707723e-01 2.70904183e-01 1.03316987e+00 -7.29959488e-01
-1.97395778e+00 -1.06177390e+00 3.22554201e-01 -1.08496577e-01
1.03111184e+00 -3.65806580e-01 -8.03223372e-01 8.64241064e-01
5.16523063e-01 -5.01237214e-01 2.58258402e-01 -2.00107574e-01
7.06352815e-02 -4.33310926e-01 -1.18169391e+00 5.92630386e-01
6.12122357e-01 2.19327342e-02 -4.01032329e-01 2.20471188e-01
9.82334197e-01 -8.87390301e-02 -1.09441805e+00 2.52362072e-01
5.98004878e-01 -5.53094029e-01 6.17574751e-01 -7.84822941e-01
1.24624148e-01 -1.25929013e-01 -3.05813760e-01 -1.77471232e+00
-5.77072382e-01 -8.47732842e-01 -3.76297683e-01 1.07544959e+00
2.71236241e-01 -1.02692556e+00 4.39627558e-01 3.55328500e-01
-7.44691044e-02 -7.88633287e-01 -9.66660857e-01 -1.03902817e+00
2.32771158e-01 -4.48112637e-02 1.07102668e+00 7.94913888e-01
-1.05358317e-01 1.50093026e-02 -6.08946860e-01 4.02200937e-01
4.05116498e-01 3.93694431e-01 3.58982652e-01 -1.07936001e+00
-6.18264914e-01 -5.32025218e-01 5.59681877e-02 -8.22572172e-01
-1.73761398e-02 -4.74760801e-01 1.19630493e-01 -1.66180551e+00
-1.72172308e-01 -6.46958351e-01 -6.55047953e-01 2.89207548e-01
5.49763620e-01 -3.27761441e-01 6.88941002e-01 -4.18292284e-02
-5.00366092e-01 4.91488039e-01 1.47994125e+00 -4.48655546e-01
-1.74735233e-01 3.35835397e-01 -8.67600739e-01 3.96415710e-01
1.03011441e+00 -3.99551213e-01 -4.23398018e-01 -3.74449372e-01
6.20763719e-01 4.79360133e-01 2.03843461e-03 -5.40661454e-01
-6.79239184e-02 -7.63508201e-01 2.17536092e-01 -5.00129938e-01
1.49982631e-01 -1.38341439e+00 6.40223384e-01 1.09598231e+00
-2.89889723e-01 8.08961749e-01 1.23949476e-01 7.27900624e-01
-4.29554582e-02 -1.88752592e-01 4.73510295e-01 -3.47910553e-01
-5.86440742e-01 1.89317018e-01 -8.56496751e-01 -2.58482844e-01
1.10428607e+00 -4.19621430e-02 -3.21366519e-01 -6.23864889e-01
-7.17142642e-01 7.78564811e-01 4.54595447e-01 5.82025349e-01
3.26072335e-01 -8.70932937e-01 -6.26578271e-01 -1.27784582e-02
-2.36496583e-01 -1.07408479e-01 -2.14227177e-02 6.65520668e-01
-7.74066150e-01 5.80086648e-01 -6.14514589e-01 -1.80341989e-01
-4.21222180e-01 9.99749660e-01 5.33642650e-01 -8.12736213e-01
-4.23727602e-01 -7.57070929e-02 -1.26883350e-02 -4.20091003e-01
2.29955062e-01 -7.83160090e-01 -1.36280134e-01 4.53202635e-01
1.52186319e-01 2.99543858e-01 -3.14891413e-02 -7.06320927e-02
-1.45002574e-01 2.49413833e-01 8.74204487e-02 -1.21774830e-01
1.70591557e+00 -2.95114309e-01 8.68147090e-02 5.12163699e-01
6.85897946e-01 -3.30306709e-01 -1.68673015e+00 1.14370864e-02
4.68197078e-01 -2.23810479e-01 -2.80171961e-01 -1.23813772e+00
-1.12504685e+00 6.96112931e-01 6.28082573e-01 9.27506089e-01
1.18451989e+00 -1.84699938e-01 6.97502971e-01 5.39021671e-01
3.43327165e-01 -1.33319616e+00 2.05266699e-01 6.34682059e-01
1.07138360e+00 -1.02397943e+00 -2.56518036e-01 2.82396168e-01
-6.44565940e-01 1.32243299e+00 2.91092306e-01 -1.73941050e-02
4.59184021e-01 5.27778387e-01 5.83659746e-02 -5.68259507e-02
-8.05331945e-01 -1.33573100e-01 -3.49301457e-01 6.67310894e-01
-1.99394867e-01 4.58340347e-01 -5.45078143e-02 4.81275678e-01
-2.43831589e-03 -1.35581478e-01 8.89678895e-01 9.97850657e-01
-3.60491633e-01 -1.11329663e+00 -4.07451898e-01 2.53374487e-01
-2.81153738e-01 3.52128595e-01 -9.00396630e-02 1.11696398e+00
1.94650062e-03 8.60302329e-01 4.10025239e-01 -1.41677082e-01
5.98021030e-01 -1.83769971e-01 3.74258637e-01 -3.13172042e-01
-8.34316254e-01 8.70692879e-02 -4.48263250e-02 -3.48881125e-01
8.58239271e-03 -7.75522172e-01 -1.22202849e+00 -4.37262684e-01
-8.16520527e-02 9.47447028e-03 6.66500032e-01 9.12245452e-01
2.66124517e-01 8.67704391e-01 9.77101207e-01 -8.20938170e-01
-1.22398353e+00 -6.32814825e-01 -7.34937608e-01 2.56203055e-01
6.88101053e-01 -6.81714714e-01 -2.29844488e-02 -3.56293559e-01] | [4.3895697593688965, 2.429342031478882] |
5e8a3482-bd9f-4da4-97bf-1ce3460cf0da | proofver-natural-logic-theorem-proving-for | 2108.11357 | null | https://arxiv.org/abs/2108.11357v2 | https://arxiv.org/pdf/2108.11357v2.pdf | ProoFVer: Natural Logic Theorem Proving for Fact Verification | Fact verification systems typically rely on neural network classifiers for veracity prediction which lack explainability. This paper proposes ProoFVer, which uses a seq2seq model to generate natural logic-based inferences as proofs. These proofs consist of lexical mutations between spans in the claim and the evidence retrieved, each marked with a natural logic operator. Claim veracity is determined solely based on the sequence of these operators. Hence, these proofs are faithful explanations, and this makes ProoFVer faithful by construction. Currently, ProoFVer has the highest label accuracy and the second-best Score in the FEVER leaderboard. Furthermore, it improves by 13.21% points over the next best model on a dataset with counterfactual instances, demonstrating its robustness. As explanations, the proofs show better overlap with human rationales than attention-based highlights and the proofs help humans predict model decisions correctly more often than using the evidence directly. | ['Andreas Vlachos', 'Sebastian Riedel', 'Amrith Krishna'] | 2021-08-25 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 6.54959157e-02 9.05026615e-01 -8.52834404e-01 -2.22336166e-02
-8.12643826e-01 -6.00580692e-01 9.23755348e-01 1.98605210e-01
2.35393152e-01 1.25273538e+00 5.37146986e-01 -9.54119503e-01
-2.21732557e-01 -7.00925708e-01 -1.11256695e+00 -2.74171941e-02
1.14001654e-01 5.13212323e-01 -1.27798855e-01 -1.39795765e-01
6.19960308e-01 -6.36431873e-02 -1.11919796e+00 8.60820472e-01
1.02558708e+00 8.04774702e-01 -4.35399383e-01 9.33871269e-01
-1.06630020e-01 1.99190354e+00 -6.42931879e-01 -1.21585715e+00
-1.40533775e-01 -6.59923851e-01 -7.98634768e-01 -8.21276546e-01
8.97275150e-01 -5.10111213e-01 -4.05515611e-01 8.39200377e-01
1.06971420e-01 -4.50019836e-01 7.91785419e-01 -1.80978906e+00
-1.32054996e+00 1.44061017e+00 -2.11134523e-01 2.47357979e-01
7.52951205e-01 3.01401317e-01 1.56409657e+00 -7.96856999e-01
9.67098773e-01 1.07920933e+00 1.02377117e+00 7.42768764e-01
-1.24757564e+00 -6.31503940e-01 -2.90235251e-01 8.90224099e-01
-6.72533691e-01 -4.35965478e-01 8.81558597e-01 -5.40239990e-01
1.22305143e+00 3.87814373e-01 5.44026017e-01 1.55275035e+00
5.43950975e-01 8.66922021e-01 1.05880249e+00 -3.84474695e-01
2.21203625e-01 -2.21629795e-02 2.29410529e-01 1.10348523e+00
6.70747221e-01 3.59619737e-01 -8.46513033e-01 -3.06285113e-01
2.50712126e-01 -1.63189709e-01 -4.98133212e-01 -1.46662772e-01
-1.49503303e+00 1.13664329e+00 5.21013081e-01 -8.32426641e-03
-6.67937338e-01 6.59515321e-01 4.28632051e-01 1.59868672e-01
4.50986437e-02 9.47077692e-01 -6.13731205e-01 -2.43542120e-01
-9.71081316e-01 6.15249574e-01 1.11229062e+00 6.15440547e-01
7.31419921e-02 4.66505624e-02 -6.86588407e-01 2.81589240e-01
5.87362750e-03 7.76763558e-01 1.62655130e-01 -1.41162622e+00
5.02720594e-01 6.14679635e-01 3.36842358e-01 -1.27181339e+00
-3.24003905e-01 -3.11457932e-01 -7.00956583e-01 4.11762446e-02
5.42139113e-01 1.08609840e-01 -4.36982244e-01 1.83171892e+00
-1.80261999e-01 7.17367977e-02 3.63975734e-01 6.86793089e-01
8.46107960e-01 3.12699795e-01 1.32152304e-01 -1.72629759e-01
1.21110106e+00 -8.37674022e-01 -8.05074215e-01 -7.30500221e-02
6.59378350e-01 -2.29178756e-01 1.11133635e+00 6.54588282e-01
-9.92061734e-01 -1.28699392e-01 -1.22674191e+00 -1.14327937e-01
-2.41744056e-01 -4.91460375e-02 1.18125379e+00 2.44202763e-01
-7.30410397e-01 7.03243673e-01 -3.28812413e-02 1.96798563e-01
6.49560213e-01 -2.02233776e-01 -1.92928717e-01 -1.49894461e-01
-1.73045278e+00 1.12224090e+00 4.22073543e-01 -1.06513172e-01
-9.23178256e-01 -8.67610097e-01 -8.33441913e-01 3.47344697e-01
5.78124166e-01 -8.59975338e-01 1.31714082e+00 -7.12172031e-01
-9.99688387e-01 6.74785316e-01 -6.28641844e-02 -1.09142447e+00
8.76423001e-01 -2.03393400e-01 -5.16120136e-01 1.38148636e-01
5.29350698e-01 5.91831744e-01 6.71515703e-01 -1.13086319e+00
-3.70444447e-01 1.93870500e-01 5.04076123e-01 -5.72786331e-01
2.76594818e-01 -3.00948769e-01 4.16693479e-01 -4.04927582e-01
-2.31083438e-01 -7.17315018e-01 4.57238883e-01 -1.38061896e-01
-9.11980212e-01 -6.81548178e-01 2.31247455e-01 -8.32873881e-01
1.22043526e+00 -1.59219813e+00 -4.46946830e-01 -3.01969498e-02
6.05171919e-01 -1.97891258e-02 -8.58525261e-02 3.29707205e-01
-3.15282673e-01 5.62501490e-01 1.13077968e-01 3.86663884e-01
5.91416180e-01 -1.19969897e-01 -9.33591604e-01 8.19777176e-02
2.51838148e-01 1.35418034e+00 -1.05178297e+00 -6.28080368e-01
-3.40763807e-01 -1.57704845e-01 -6.84817970e-01 -9.31263268e-02
-8.63231719e-01 -4.77000087e-01 -2.35040903e-01 4.63032842e-01
4.01454002e-01 -5.96497893e-01 3.08775723e-01 1.06357627e-01
1.60714924e-01 9.76944387e-01 -3.36276114e-01 1.31832957e+00
-2.31431305e-01 9.08756375e-01 -4.82220203e-01 -4.88712579e-01
6.81903183e-01 5.02682447e-01 -7.12709576e-02 -7.18512535e-01
-5.52104115e-02 3.29218328e-01 2.48403460e-01 -7.97335923e-01
4.47087884e-01 -4.38826382e-01 -9.67336595e-02 8.46737206e-01
-4.31306511e-01 1.64220519e-02 2.25940645e-01 7.71115661e-01
1.39082778e+00 1.68556243e-01 4.04168218e-01 1.22740678e-01
2.10574359e-01 5.66763520e-01 6.37509108e-01 1.23672831e+00
-1.59886003e-01 1.38315395e-01 1.23498857e+00 -7.02401102e-01
-9.74457383e-01 -1.15514934e+00 2.51681000e-01 6.34208739e-01
-2.01007694e-01 -4.43024516e-01 -6.18230760e-01 -1.22650027e+00
6.38318777e-01 1.67956090e+00 -9.36616004e-01 -3.87698859e-01
-1.31747440e-01 2.44628608e-01 1.03459990e+00 6.50217831e-01
4.21290576e-01 -1.53739274e+00 -5.14080524e-01 8.71791393e-02
-6.82149768e-01 -7.77688920e-01 -1.51676118e-01 -2.09150631e-02
-4.57590818e-01 -1.73895466e+00 -4.08728905e-02 -9.47412401e-02
1.82039395e-01 4.98036072e-02 1.32138610e+00 6.38225675e-01
1.43278882e-01 -3.66746075e-02 -8.95301700e-02 -6.03332043e-01
-8.78290534e-01 -2.15346456e-01 1.67824998e-01 -5.95981300e-01
4.96298373e-01 -3.25950652e-01 -1.26178220e-01 -3.11406106e-01
-3.70242834e-01 3.58146161e-01 6.47770286e-01 1.11195028e+00
1.08462162e-01 -2.44773537e-01 8.53869855e-01 -1.00765562e+00
9.53886211e-01 -6.92120492e-01 -3.20742548e-01 4.70723122e-01
-1.10392952e+00 3.48292589e-01 8.66173685e-01 -3.58778149e-01
-8.98892760e-01 -7.40054429e-01 5.14167190e-01 -4.59578812e-01
4.15632129e-02 6.78410351e-01 1.83856905e-01 7.82111526e-01
1.10172176e+00 -1.10575119e-02 -1.09274119e-01 1.73159450e-01
6.47686422e-01 3.83801281e-01 7.14486301e-01 -6.54235303e-01
6.75401032e-01 4.17163111e-02 1.33607527e-02 2.03895941e-01
-1.37976813e+00 3.58360827e-01 9.17105004e-03 -1.03163026e-01
6.80861592e-01 -3.46580207e-01 -1.46866298e+00 -1.47152141e-01
-1.82904983e+00 -4.24742490e-01 -2.25804672e-01 5.99442184e-01
-7.33499885e-01 8.87261033e-02 -7.28819489e-01 -1.04816151e+00
-3.71452779e-01 -4.62866157e-01 4.88896757e-01 -1.97097942e-01
-7.91743875e-01 -7.43822694e-01 -5.06710783e-02 6.30627692e-01
4.97506969e-02 2.79423058e-01 1.57097626e+00 -8.79522800e-01
-6.82323575e-01 -1.56921640e-01 -4.41320032e-01 -2.44421363e-02
-3.04187387e-01 2.02336013e-02 -9.30191636e-01 5.18265188e-01
-2.24618301e-01 -6.25191391e-01 1.02954865e+00 3.60336393e-01
9.31808472e-01 -1.04761326e+00 -2.89766788e-01 3.85943912e-02
1.15227103e+00 -1.49247512e-01 6.54697239e-01 4.51683760e-01
3.45195591e-01 4.74032789e-01 3.73126507e-01 1.42784074e-01
5.83642423e-01 1.17358387e-01 5.55506408e-01 2.61325926e-01
-2.35539421e-01 -9.71251190e-01 5.84805787e-01 8.70391279e-02
1.43579487e-02 -2.93847382e-01 -1.00340056e+00 4.94349360e-01
-1.90425014e+00 -1.92483568e+00 -6.36173189e-01 1.58556700e+00
9.42209899e-01 4.93928164e-01 -2.18255594e-01 3.08339089e-01
5.44023275e-01 4.35671508e-02 -6.44318163e-01 -7.75588095e-01
-3.98871750e-01 -1.33675262e-01 2.06905097e-01 6.93509817e-01
-5.59848368e-01 7.78672099e-01 7.07013178e+00 5.74994922e-01
-7.39621699e-01 2.37576663e-02 2.78761387e-01 -2.33663097e-01
-1.07310331e+00 1.25889093e-01 -1.03810750e-01 4.94541645e-01
7.46043622e-01 -1.11533619e-01 5.61582029e-01 9.09385502e-01
2.86418378e-01 -9.36829373e-02 -1.41131973e+00 6.59176826e-01
2.20242158e-01 -2.05945015e+00 2.56655365e-01 -1.20126903e-01
8.48026156e-01 -2.60757118e-01 -3.72596532e-02 5.15851736e-01
7.59761810e-01 -1.20774746e+00 1.30846095e+00 8.81200850e-01
7.45217919e-01 -5.17231286e-01 1.00957537e+00 3.97829831e-01
-7.39890784e-02 -4.67904091e-01 -1.13145381e-01 -4.04917985e-01
-5.91426902e-02 7.62278497e-01 -1.44394302e+00 2.51983196e-01
1.22627199e-01 5.50754249e-01 -4.20463324e-01 5.69324911e-01
-1.25519466e+00 9.93819118e-01 2.44757563e-01 -5.53651929e-01
5.58729135e-02 5.35689831e-01 4.55346018e-01 1.06265795e+00
1.34355247e-01 1.78461462e-01 -1.87766910e-01 1.70125854e+00
-5.90896487e-01 -4.44417804e-01 -8.04653645e-01 -3.39876324e-01
5.64397573e-01 8.40920508e-01 -1.15090571e-01 -7.10588813e-01
-9.32452921e-03 6.82053089e-01 4.04441625e-01 2.86993235e-01
-1.38427246e+00 -1.37512267e-01 3.04964840e-01 1.24515846e-01
1.90614372e-01 4.19030696e-01 -7.02206314e-01 -9.08633113e-01
1.33129537e-01 -1.05092180e+00 4.26457047e-01 -1.69369423e+00
-1.44098234e+00 2.97283649e-01 -3.10244411e-01 -6.76007807e-01
-6.43420398e-01 -6.86825633e-01 -8.01050067e-01 6.72554493e-01
-1.63353622e+00 -1.22015619e+00 -5.75293563e-02 7.20135644e-02
1.61245331e-01 -2.78271586e-01 9.55962956e-01 -5.03216267e-01
-2.17347220e-01 5.35463333e-01 -4.72367704e-01 2.75378257e-01
6.25155687e-01 -1.42392254e+00 3.05080324e-01 8.14010680e-01
2.05519229e-01 9.02197599e-01 9.81792569e-01 -1.02467251e+00
-1.11169338e+00 -5.95311582e-01 1.60663414e+00 -8.34015489e-01
9.06094074e-01 5.96681200e-02 -7.01938748e-01 7.23034561e-01
3.59275699e-01 -3.95284861e-01 4.81340706e-01 5.87071598e-01
-1.27384579e+00 2.38079906e-01 -1.04165745e+00 6.91575825e-01
1.07599676e+00 -8.79868448e-01 -1.23836350e+00 3.74330789e-01
7.03909278e-01 2.04263814e-02 -3.01495463e-01 -8.58400539e-02
7.63377726e-01 -1.19200456e+00 4.68373746e-01 -1.27804732e+00
1.45724976e+00 -2.69561797e-01 -9.75956544e-02 -1.17049432e+00
-5.48321605e-01 -7.03714192e-01 -6.74819708e-01 9.37396288e-01
9.30621684e-01 -5.55265188e-01 7.15763807e-01 6.43230259e-01
1.29083797e-01 -7.94252634e-01 -7.20068932e-01 -7.25091219e-01
3.49405669e-02 -7.61966288e-01 9.39517438e-01 1.38890982e+00
6.71712518e-01 6.52040780e-01 -1.69112459e-01 -2.14589089e-01
6.47750199e-01 5.21441281e-01 6.31638408e-01 -1.32735074e+00
-2.99536765e-01 -7.53194332e-01 3.55483256e-02 -5.69824636e-01
6.34875357e-01 -1.36535990e+00 7.88083747e-02 -1.84718585e+00
6.37342572e-01 -3.06339771e-01 -1.12669230e-01 1.03376806e+00
-3.90952408e-01 -1.25792161e-01 1.49094775e-01 2.41972450e-02
-5.56479990e-01 2.09457695e-01 1.09617186e+00 -5.00212789e-01
3.60267848e-01 -3.87963474e-01 -1.14840293e+00 7.91939497e-01
8.15879285e-01 -6.55766070e-01 -3.46860498e-01 -3.52189571e-01
9.41141248e-01 4.61257726e-01 1.00153911e+00 -5.33958793e-01
1.50130078e-01 -5.20248175e-01 4.30777669e-01 -3.99642110e-01
1.70828681e-02 -2.06189647e-01 -4.05619964e-02 6.85229957e-01
-9.64961410e-01 2.08104998e-01 -2.14140359e-02 6.68148518e-01
2.78310418e-01 -1.66579798e-01 2.26734146e-01 3.53048928e-03
-4.49197769e-01 -1.96299106e-01 -7.28480220e-02 2.69628555e-01
5.38625777e-01 -9.31857303e-02 -1.04450679e+00 -7.70820677e-01
-3.61555934e-01 -3.78665254e-02 2.87553668e-01 3.35449487e-01
8.25677931e-01 -1.36871254e+00 -1.10529113e+00 -2.70417213e-01
1.80402890e-01 -6.21094644e-01 3.01445089e-02 1.11747468e+00
-1.48749068e-01 6.91422939e-01 -6.21923022e-02 9.25085545e-02
-7.91860223e-01 8.75961244e-01 2.33263269e-01 -4.54680264e-01
-4.43764269e-01 7.58139372e-01 -2.88340718e-01 -1.89633995e-01
-1.58819869e-01 -4.09106553e-01 -2.31258310e-02 -2.59985238e-01
6.11707330e-01 5.37995219e-01 -3.88877898e-01 -3.51225287e-02
-3.17278326e-01 -3.01470995e-01 2.00909272e-01 -1.21642642e-01
1.17519486e+00 7.58968890e-01 -3.20111185e-01 3.13780069e-01
5.91493547e-01 5.18115163e-01 -9.35942531e-01 2.20119879e-01
2.33896166e-01 -3.45381528e-01 -2.17965677e-01 -1.74995697e+00
-5.94270468e-01 7.74358869e-01 -4.59849119e-01 5.38042128e-01
2.04145342e-01 1.29783943e-01 6.82538271e-01 5.68363369e-01
2.57798463e-01 -8.42336833e-01 8.46701637e-02 4.42659736e-01
1.29381454e+00 -1.01278305e+00 2.67393701e-02 -4.14195471e-02
-8.03531170e-01 9.34716046e-01 4.63001400e-01 1.89373549e-02
-2.42437065e-01 4.83168662e-02 -6.06723651e-02 -4.79033411e-01
-1.39184570e+00 3.47494155e-01 1.99068055e-01 6.35097623e-01
4.01447386e-01 3.69508535e-01 -3.77739966e-01 1.04457104e+00
-6.36129975e-01 3.59467715e-01 7.71987319e-01 2.67373044e-02
-2.54735351e-01 -5.14811099e-01 -4.04717863e-01 7.03746259e-01
-3.91286045e-01 -5.17589450e-01 -9.02400613e-01 8.95766556e-01
1.48654491e-01 1.17532647e+00 -3.52616280e-01 -4.94873285e-01
1.21300243e-01 3.71968180e-01 5.06769419e-01 -2.24376619e-01
-5.37584007e-01 -1.00544453e+00 7.56149769e-01 -5.95932782e-01
-2.45886054e-02 -4.57341582e-01 -1.59936452e+00 -1.03719926e+00
-3.58748972e-01 5.04519761e-01 3.82697433e-01 9.68836904e-01
4.62494224e-01 4.31377888e-01 1.70854583e-01 -2.29062270e-02
-1.02992892e+00 -1.03416502e+00 -2.05903575e-01 5.95255673e-01
3.84549528e-01 -7.13796556e-01 -4.50701267e-01 2.16195756e-03] | [9.852505683898926, 7.818010330200195] |
4c12b986-3bf7-4235-928c-ab07aee9840d | a-study-on-the-invariance-in-security | 2302.11527 | null | https://arxiv.org/abs/2302.11527v1 | https://arxiv.org/pdf/2302.11527v1.pdf | A study on the invariance in security whatever the dimension of images for the steganalysis by deep-learning | In this paper, we study the performance invariance of convolutional neural networks when confronted with variable image sizes in the context of a more "wild steganalysis". First, we propose two algorithms and definitions for a fine experimental protocol with datasets owning "similar difficulty" and "similar security". The "smart crop 2" algorithm allows the introduction of the Nearly Nested Image Datasets (NNID) that ensure "a similar difficulty" between various datasets, and a dichotomous research algorithm allows a "similar security". Second, we show that invariance does not exist in state-of-the-art architectures. We also exhibit a difference in behavior depending on whether we test on images larger or smaller than the training images. Finally, based on the experiments, we propose to use the dilated convolution which leads to an improvement of a state-of-the-art architecture. | ['Frédéric Comby', 'Marc Chaumont', 'Kévin Planolles'] | 2023-02-22 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 4.58962083e-01 -6.25283942e-02 2.90645808e-01 -1.09378599e-01
1.79095700e-01 -5.79779088e-01 7.52259433e-01 -1.71445414e-01
-8.04343641e-01 5.20947397e-01 -2.99849778e-01 -5.94206631e-01
-7.04497322e-02 -9.96298552e-01 -8.84690523e-01 -8.42060626e-01
-2.64653653e-01 -1.95436433e-01 3.69350612e-01 -6.52065039e-01
2.82915205e-01 4.95528966e-01 -1.62259948e+00 2.95404315e-01
3.54464799e-01 9.84377086e-01 1.24259926e-01 8.11510444e-01
1.90957785e-01 3.80078405e-01 -6.26366913e-01 -5.21446466e-01
7.40547895e-01 -3.27341557e-01 -8.92885387e-01 -3.05555373e-01
4.95186687e-01 -3.21197063e-01 -2.40745544e-01 1.36723053e+00
3.53208303e-01 -3.78908932e-01 4.57360953e-01 -1.11885142e+00
-6.72916949e-01 8.00630212e-01 -3.26118618e-01 1.69305637e-01
1.05300076e-01 3.96478385e-01 4.53436404e-01 -4.08845730e-02
6.86898708e-01 1.17097592e+00 7.24953711e-01 7.22711146e-01
-1.29380202e+00 -9.00818408e-01 -9.91399214e-02 6.19595088e-02
-1.28632617e+00 -2.08463192e-01 5.44875145e-01 -3.03184241e-01
5.15399635e-01 4.53230411e-01 5.38127959e-01 1.30808449e+00
3.03228587e-01 -4.43214513e-02 1.30407679e+00 -6.67593062e-01
2.66561322e-02 4.22699422e-01 2.93605253e-02 5.84961951e-01
7.56857097e-01 2.38658726e-01 2.85393357e-01 1.55315846e-01
6.67254269e-01 -1.44873902e-01 -4.35386389e-01 -3.79990488e-01
-1.18493462e+00 9.42384124e-01 3.73505890e-01 1.06381798e+00
-9.07026008e-02 2.41141066e-01 5.85679114e-01 8.88066888e-01
8.14626589e-02 6.66115165e-01 -3.39440554e-01 3.12332511e-01
-7.35913217e-01 -1.68474503e-02 8.78048480e-01 6.10955596e-01
7.65456557e-01 1.57981321e-01 2.23335758e-01 1.27653256e-01
-2.42715348e-02 3.50386053e-01 7.24234819e-01 -4.18472469e-01
4.69591796e-01 4.87418860e-01 3.88071351e-02 -1.26297939e+00
-5.14637291e-01 -4.82942134e-01 -1.32097566e+00 8.13511431e-01
6.56790316e-01 -2.24411592e-01 -9.43417311e-01 2.05771017e+00
-2.20059782e-01 1.50722161e-01 5.79211831e-01 4.80793953e-01
5.96521378e-01 3.72055739e-01 -6.52690604e-02 -9.85156000e-02
1.63859940e+00 -4.13130403e-01 -5.22036493e-01 3.23735446e-01
8.24554861e-01 -6.87432051e-01 1.07193255e+00 5.76310396e-01
-6.89424038e-01 -7.64712572e-01 -1.39479768e+00 4.20665234e-01
-1.07899773e+00 -3.51805873e-02 2.09000409e-01 1.52599680e+00
-1.18601906e+00 8.37237775e-01 -3.05797398e-01 -5.72093427e-01
-1.27741881e-02 3.70686173e-01 -6.32327437e-01 3.71201903e-01
-1.36384392e+00 8.88061345e-01 8.84859502e-01 -2.98956167e-02
-1.05108237e+00 -4.02191132e-01 -5.35675108e-01 3.27471852e-01
-1.19400593e-02 -3.88818651e-01 5.54820359e-01 -1.22340381e+00
-1.35204113e+00 1.22241580e+00 4.54195291e-01 -8.72634292e-01
8.49982321e-01 -2.28115860e-02 -6.56450212e-01 4.96254787e-02
-4.39126015e-01 6.09015405e-01 9.51016366e-01 -1.34660864e+00
-4.39943135e-01 -9.67974588e-02 4.55028057e-01 -4.56559002e-01
-6.48487329e-01 1.66573718e-01 1.24428652e-01 -5.91803789e-01
-1.35849074e-01 -1.05675173e+00 -1.39494881e-01 -3.51524442e-01
-3.27008456e-01 3.95517617e-01 1.24293542e+00 -3.25060248e-01
1.12251651e+00 -2.31600285e+00 -1.67119931e-02 4.62191850e-01
1.69763550e-01 5.55217981e-01 -3.97124022e-01 2.64989644e-01
-6.28298283e-01 6.73304498e-01 -1.07358821e-01 -6.24186546e-02
-1.70983449e-02 1.15551353e-01 -4.00994122e-02 5.99284232e-01
5.86052574e-02 4.36200172e-01 -3.54433656e-01 -2.43095100e-01
1.21084098e-02 2.57849962e-01 -4.10992354e-01 1.60445452e-01
5.58935776e-02 4.58722621e-01 -2.44201571e-01 2.18138471e-01
1.22348285e+00 -2.39405155e-01 4.19055581e-01 -4.12945956e-01
-2.75003940e-01 -5.11371255e-01 -1.39376915e+00 1.30680454e+00
-3.09781462e-01 7.27004051e-01 9.11025926e-02 -1.19301951e+00
1.01466632e+00 4.60115701e-01 4.38601226e-02 -6.00700974e-01
6.09115779e-01 3.10135275e-01 3.22340935e-01 -5.02813578e-01
4.38470155e-01 1.79001153e-01 9.82135236e-02 4.18303400e-01
7.73884682e-03 1.01325601e-01 3.52956861e-01 -1.23902276e-01
1.10650063e+00 -2.10522041e-01 3.58773708e-01 -7.15531588e-01
1.01962543e+00 -4.67738032e-01 5.67220673e-02 1.01722562e+00
-8.79613012e-02 3.54808599e-01 8.13489735e-01 -6.89036608e-01
-1.27506256e+00 -7.06473947e-01 -1.81247428e-01 7.99476445e-01
2.40922689e-01 -4.53499593e-02 -1.13288391e+00 -7.56239653e-01
-3.04718941e-01 2.98424423e-01 -1.06669140e+00 -2.04085298e-02
-7.17694283e-01 -8.59884262e-01 1.14167106e+00 -5.46827652e-02
1.24202716e+00 -1.10333729e+00 -9.39535141e-01 -3.27627778e-01
1.11466266e-01 -1.21161890e+00 2.37623587e-01 2.32319698e-01
-5.90697110e-01 -1.13764882e+00 -6.29173875e-01 -8.31858873e-01
5.84517419e-01 4.97241020e-02 1.07028735e+00 5.46695769e-01
3.64083685e-02 -1.67399105e-02 -7.23192155e-01 -3.02015036e-01
-1.00972474e+00 3.69842976e-01 4.11149003e-02 1.42538324e-01
1.50625231e-02 -6.65242732e-01 -6.67500913e-01 2.36207262e-01
-1.45999825e+00 -2.60669887e-01 7.21234381e-01 6.13714337e-01
-1.28890514e-01 2.53153116e-01 1.55793086e-01 -1.01553082e+00
7.08877921e-01 -4.12944764e-01 -6.60458386e-01 3.86867493e-01
-8.20479810e-01 4.31581497e-01 1.17411876e+00 -6.36687338e-01
-5.91587663e-01 -2.13000149e-01 -1.24133140e-01 -3.11096579e-01
-6.08161747e-01 -6.96777850e-02 -1.65969923e-01 -6.69820428e-01
7.57279813e-01 2.79515535e-01 8.97610858e-02 -4.04429257e-01
1.93540752e-01 4.32383925e-01 3.64874244e-01 -4.18972790e-01
1.14457476e+00 4.81102675e-01 2.67219752e-01 -6.49119973e-01
7.18932897e-02 6.51315525e-02 -4.26594853e-01 2.05732480e-01
1.05906427e+00 -4.34798628e-01 -9.30924237e-01 8.76469553e-01
-1.20497847e+00 -2.14463234e-01 -1.86843827e-01 4.87829953e-01
-4.28257585e-01 7.61368036e-01 -5.67548037e-01 -4.11064267e-01
-3.54852885e-01 -1.39201236e+00 3.65522742e-01 2.05381066e-01
2.75344074e-01 -9.57761049e-01 1.17822148e-01 -2.37875223e-01
5.66072702e-01 3.62298101e-01 8.54798079e-01 -9.62157190e-01
-3.37812543e-01 -4.08749916e-02 -4.81933981e-01 6.48731411e-01
3.40728909e-02 4.22609299e-02 -9.90555823e-01 -7.45044351e-01
8.94178152e-02 -3.01562380e-02 8.78336191e-01 -1.33108357e-02
9.85760152e-01 -4.88228291e-01 -4.72396277e-02 9.48682249e-01
1.84523666e+00 2.47164026e-01 1.08849251e+00 7.70382941e-01
2.49033600e-01 5.85778058e-01 1.86111704e-01 2.72740573e-01
-2.45039314e-01 4.16731328e-01 8.19353044e-01 -3.75868887e-01
9.33188125e-02 1.61189884e-01 2.09343180e-01 3.28367740e-01
-2.44389728e-01 -5.97797334e-01 -6.18483663e-01 3.65087152e-01
-1.35513318e+00 -1.12648559e+00 -1.67661265e-01 2.09691787e+00
4.54520911e-01 2.28480086e-01 -2.66752224e-02 3.32001299e-01
8.98843944e-01 4.79658127e-01 1.59617707e-01 -8.99945498e-01
-2.46562958e-01 5.47227442e-01 1.01752174e+00 2.10120633e-01
-1.31237400e+00 7.65075147e-01 6.70162058e+00 9.17586267e-01
-1.58015645e+00 1.66601717e-01 4.54007745e-01 6.17023826e-01
-1.45710647e-01 4.49537188e-02 -5.29495239e-01 7.58346319e-01
9.51808155e-01 3.17272961e-01 1.43179983e-01 4.80251521e-01
-3.78867052e-02 1.15016267e-01 -7.64636993e-01 7.09048212e-01
3.77964936e-02 -1.27445281e+00 3.15316558e-01 2.47090057e-01
5.78636646e-01 -2.07839504e-01 3.09916943e-01 -2.12801117e-02
2.55898446e-01 -1.07547748e+00 6.18744612e-01 2.11713403e-01
9.81247485e-01 -4.96156514e-01 1.00616217e+00 2.06338122e-01
-1.16870677e+00 -3.17905635e-01 -4.78139520e-01 6.12550750e-02
-3.36572230e-01 1.07217424e-01 -4.64210898e-01 7.18432009e-01
7.39362001e-01 -6.84479699e-02 -7.87078857e-01 7.01893389e-01
9.25284177e-02 2.68451422e-01 -1.83608621e-01 -1.62591696e-01
5.18305719e-01 -6.91616237e-02 3.77041996e-01 1.53663647e+00
5.88124931e-01 -1.28539026e-01 -3.52451324e-01 7.04086959e-01
-8.07133242e-02 2.08143756e-01 -9.13220644e-01 1.91145942e-01
1.57062143e-01 1.12249517e+00 -8.23927164e-01 -3.58756781e-01
-3.88168544e-01 8.47517073e-01 -2.14279637e-01 1.98823839e-01
-7.64096856e-01 -5.32888770e-01 4.20110315e-01 3.28343846e-02
5.72199821e-01 -1.37068987e-01 -6.70212228e-03 -1.20263803e+00
-2.62710124e-01 -1.10625064e+00 2.07249254e-01 -4.49192554e-01
-8.37137938e-01 7.78500557e-01 2.33799759e-02 -1.47610366e+00
-9.30288732e-02 -1.02922237e+00 -6.35911047e-01 6.74008131e-01
-1.46483541e+00 -1.14329219e+00 -4.39919740e-01 8.78488183e-01
2.35128682e-03 -5.36391973e-01 1.03778911e+00 2.77487069e-01
-3.58354121e-01 8.96955311e-01 3.07519972e-01 4.49423343e-01
4.66333747e-01 -8.54661047e-01 4.34198737e-01 1.12565088e+00
-1.00223415e-01 4.46409106e-01 9.74646151e-01 -3.42893213e-01
-9.53287840e-01 -6.70813143e-01 5.16492426e-01 -2.24504266e-02
5.06513059e-01 -1.91323623e-01 -7.47378945e-01 5.70001245e-01
5.73924243e-01 -1.77166790e-01 2.77134389e-01 -3.06694329e-01
-8.40869844e-01 -2.18703702e-01 -1.59774578e+00 5.34929991e-01
7.75599062e-01 -3.23975682e-01 -3.24615091e-01 -1.44920066e-01
7.67967582e-01 9.88463778e-03 -7.54551649e-01 4.37978148e-01
8.05819750e-01 -1.47779357e+00 7.23227262e-01 -5.90323865e-01
1.88876405e-01 -1.81995541e-01 -2.66768515e-01 -1.05509925e+00
-3.44984680e-01 -5.82226038e-01 3.65676224e-01 9.05246377e-01
3.90502006e-01 -9.38979745e-01 9.31349516e-01 -2.85497252e-02
3.60696524e-01 -1.15085773e-01 -8.97081435e-01 -9.17818248e-01
2.96721637e-01 -8.35638791e-02 7.68534482e-01 1.20186937e+00
-3.22363585e-01 -3.82114232e-01 -5.06109297e-01 3.15354347e-01
5.15108883e-01 -2.13420570e-01 8.80898654e-01 -1.27549565e+00
-3.41467112e-01 -5.34754932e-01 -8.93604457e-01 -5.92617989e-01
2.47759447e-02 -3.60204875e-01 -3.88322264e-01 -4.24930394e-01
1.41492993e-01 -3.50677311e-01 -4.70469743e-01 3.17689687e-01
3.68081003e-01 5.77003896e-01 4.21765774e-01 7.60039613e-02
-9.46004614e-02 -2.05187142e-01 9.40484583e-01 -1.59685403e-01
1.19630724e-01 -1.33774579e-01 -5.50590456e-01 6.52092755e-01
1.06604910e+00 -4.89512354e-01 -2.92337626e-01 -3.16679597e-01
3.46639544e-01 -3.27067614e-01 5.46074569e-01 -1.48869050e+00
-1.90545604e-01 7.84667507e-02 1.33880556e-01 9.66714323e-02
-1.31795192e-02 -1.22633934e+00 3.24491978e-01 1.25145996e+00
-1.98987111e-01 3.46680611e-01 1.04763143e-01 2.25699812e-01
8.15187395e-02 -6.82218015e-01 1.17956400e+00 -3.77892941e-01
-5.89594066e-01 -7.70405158e-02 -2.08027229e-01 -3.18183810e-01
1.09047592e+00 -5.49444675e-01 -5.56330264e-01 -2.67193109e-01
-5.61224341e-01 -4.26242918e-01 6.51585579e-01 2.17998981e-01
2.39654317e-01 -9.64388251e-01 -7.76422322e-01 5.56418180e-01
-6.29873574e-02 -7.51039565e-01 1.59085289e-01 6.38520777e-01
-1.12648892e+00 3.56998086e-01 -8.35015535e-01 -2.56793976e-01
-1.53449428e+00 8.52819979e-01 4.60235685e-01 -6.53661311e-01
-2.06633240e-01 4.43349272e-01 2.73887068e-01 -2.22527117e-01
1.34510472e-01 -4.14205462e-01 -4.58147317e-01 -6.22303374e-02
3.64289939e-01 2.71312177e-01 3.25676464e-02 -5.19496560e-01
-7.20852390e-02 6.91128552e-01 -4.52876389e-02 9.27687585e-02
1.01903343e+00 -7.24501833e-02 -7.06431344e-02 -1.56526327e-01
1.15361166e+00 8.30368251e-02 -7.21474409e-01 1.21409439e-01
-3.09368104e-01 -3.14634711e-01 -4.83855844e-01 -4.52272236e-01
-1.12661374e+00 9.05623257e-01 1.32222605e+00 1.02138901e+00
1.33265865e+00 -6.26908362e-01 4.29147989e-01 5.40851653e-01
2.86662787e-01 -6.77102149e-01 -1.60077531e-02 4.73723859e-01
5.81182361e-01 -1.02311563e+00 -1.19448423e-01 -1.82622299e-01
-1.81493402e-01 1.32499158e+00 4.07387733e-01 -3.91498685e-01
6.09963655e-01 4.61257815e-01 4.69151400e-02 -7.12497309e-02
-2.35586688e-01 -1.93954036e-01 -1.80767417e-01 8.00156593e-01
1.02474503e-01 -7.50247203e-03 -8.41406584e-01 9.97915789e-02
-2.62112916e-01 -6.44894764e-02 7.12826669e-01 6.78611219e-01
-4.17399377e-01 -1.34101868e+00 -6.76693141e-01 -9.90478247e-02
-8.24005485e-01 -1.89499721e-01 -2.24885464e-01 1.45531225e+00
7.12245226e-01 7.70541191e-01 1.20506175e-02 -8.12788069e-01
9.77695659e-02 -9.08288807e-02 3.34003657e-01 -3.06074340e-02
-1.03020513e+00 -5.13943672e-01 -4.15705256e-02 -3.50278258e-01
-5.94167352e-01 -6.42348975e-02 -6.24479234e-01 -8.41186702e-01
-2.74045676e-01 1.48789898e-01 7.21196711e-01 8.03934813e-01
1.12324648e-01 2.56377399e-01 7.91237772e-01 -6.97328508e-01
-7.11003840e-01 -1.01513350e+00 -6.81520402e-01 7.22368181e-01
3.67302537e-01 -2.04371542e-01 -5.88368177e-01 -3.21990810e-02] | [4.400973320007324, 8.0281400680542] |
17d111f5-c35e-4545-8795-a9563b852e2f | naiverole-author-contribution-extraction-and | 1912.10170 | null | https://arxiv.org/abs/1912.10170v1 | https://arxiv.org/pdf/1912.10170v1.pdf | NaïveRole: Author-Contribution Extraction and Parsing from Biomedical Manuscripts | Information about the contributions of individual authors to scientific publications is important for assessing authors' achievements. Some biomedical publications have a short section that describes authors' roles and contributions. It is usually written in natural language and hence author contributions cannot be trivially extracted in a machine-readable format. In this paper, we present 1) A statistical analysis of roles in author contributions sections, and 2) Na\"iveRole, a novel approach to extract structured authors' roles from author contribution sections. For the first part, we used co-clustering techniques, as well as Open Information Extraction, to semi-automatically discover the popular roles within a corpus of 2,000 contributions sections from PubMed Central. The discovered roles were used to automatically build a training set for Na\"iveRole, our role extractor approach, based on Na\"ive Bayes. Na\"iveRole extracts roles with a micro-averaged precision of 0.68, recall of 0.48 and F1 of 0.57. It is, to the best of our knowledge, the first attempt to automatically extract author roles from research papers. This paper is an extended version of a previous poster published at JCDL 2018. | ['Dominika Tkaczyk', 'Joeran Beel', 'Andrew Collins'] | 2019-12-15 | null | null | null | null | ['open-information-extraction'] | ['natural-language-processing'] | [ 2.52792567e-01 4.73829329e-01 -3.94787997e-01 5.38042709e-02
-8.49145651e-01 -7.80909359e-01 5.56474984e-01 8.85879815e-01
-5.18676758e-01 1.33127594e+00 5.97946942e-01 -3.35861564e-01
-4.10136223e-01 -5.15385330e-01 -6.19780600e-01 -5.84951282e-01
1.58822790e-01 7.48371780e-01 1.52697876e-01 1.83536336e-01
6.87472761e-01 3.82349402e-01 -1.50078773e+00 2.93617666e-01
9.78837729e-01 2.46927992e-01 3.07020575e-01 7.97234952e-01
-4.57164407e-01 7.95435727e-01 -8.84353876e-01 -3.10599536e-01
-3.37411702e-01 -5.88774025e-01 -9.60768282e-01 -4.40331906e-01
-1.98829100e-02 5.74135222e-02 1.63080215e-01 7.54139066e-01
3.21248829e-01 -6.68096989e-02 8.77135336e-01 -8.63623440e-01
-3.62487525e-01 1.20333254e+00 -4.48376715e-01 5.35095870e-01
6.22705340e-01 -5.67298770e-01 1.37028301e+00 -6.26555145e-01
1.27204919e+00 1.01886714e+00 2.34867036e-01 5.91786206e-01
-1.13507056e+00 -8.58768821e-01 2.77251881e-02 -2.46065587e-01
-1.04293513e+00 -3.25502843e-01 6.70661688e-01 -8.92923594e-01
6.26747012e-01 3.05885434e-01 7.18717039e-01 1.06180191e+00
2.37441599e-01 6.19526565e-01 1.02220488e+00 -5.95037282e-01
2.43223354e-01 1.37856066e-01 4.78669077e-01 2.69623429e-01
6.76132441e-01 -7.43803740e-01 -5.36459327e-01 -7.41516292e-01
4.18395102e-01 -2.65858680e-01 -5.05545214e-02 3.77492338e-01
-1.50263166e+00 4.62971330e-01 -2.36944526e-01 8.28701913e-01
-1.99880064e-01 -6.36161491e-02 4.54673678e-01 1.66882053e-02
5.55456936e-01 8.80899012e-01 -3.75030220e-01 -2.63310730e-01
-9.12961304e-01 4.08393681e-01 8.89541924e-01 7.93183148e-01
1.92609966e-01 -5.76129675e-01 -3.37052643e-01 8.35901141e-01
2.87823945e-01 -3.85135487e-02 5.15903354e-01 -7.39634454e-01
5.15750535e-02 8.80005956e-01 1.73404887e-01 -6.82819724e-01
-3.15289617e-01 -4.46480066e-01 -5.77387214e-01 -4.38774496e-01
6.37102365e-01 1.11345323e-02 -2.59987533e-01 1.38346207e+00
2.12244079e-01 -2.66459376e-01 -3.73351052e-02 4.01283383e-01
1.36854279e+00 3.41518492e-01 2.30134025e-01 -5.95307291e-01
1.68314004e+00 -3.78812790e-01 -1.09852016e+00 6.62776113e-01
2.90650427e-01 -1.01723552e+00 4.92759317e-01 6.62767410e-01
-9.98596728e-01 -2.36829847e-01 -8.51608038e-01 -1.55359939e-01
-5.52895963e-01 3.04139256e-01 6.24379396e-01 4.05259222e-01
-1.03633749e+00 5.43989420e-01 -3.88404340e-01 -2.09945887e-01
7.35427737e-01 2.68966407e-01 -2.82154411e-01 2.28276283e-01
-1.16300547e+00 5.22055864e-01 1.28100470e-01 -5.43530643e-01
-6.52639985e-01 -9.66553688e-01 -3.10829163e-01 -2.72881120e-01
5.24814665e-01 -6.55571163e-01 1.01031578e+00 -3.86186004e-01
-9.30154979e-01 1.26830292e+00 -6.44282639e-01 -2.08310828e-01
3.51459920e-01 -2.44551569e-01 -2.84742445e-01 4.03192937e-01
6.71248317e-01 3.16940755e-01 2.28642285e-01 -1.11110449e+00
-6.82421505e-01 -5.51288247e-01 -1.64589226e-01 -2.14553580e-01
-3.85103345e-01 6.05900347e-01 -3.43601137e-01 -9.60526466e-01
-1.40838418e-02 -5.27607739e-01 -2.37395659e-01 -3.29570651e-01
-3.91997814e-01 -9.00009215e-01 1.21817686e-01 -6.72709644e-01
1.40624750e+00 -1.64548516e+00 1.75379530e-01 1.83082268e-01
9.38887537e-01 -1.19145676e-01 4.65707392e-01 5.05643427e-01
-1.63186237e-01 6.52043641e-01 -2.16203615e-01 -7.57509395e-02
-2.44408190e-01 -4.34278473e-02 -1.67616785e-01 4.51027244e-01
-1.67488441e-01 6.67945981e-01 -1.26680541e+00 -6.12013102e-01
-4.07092124e-01 2.09734663e-01 -2.47425839e-01 1.49861023e-01
-1.84095964e-01 3.98659140e-01 -5.04413128e-01 5.67672729e-01
5.18371403e-01 -1.56607077e-01 3.30639392e-01 1.82893664e-01
-4.61378783e-01 5.26855826e-01 -8.18651438e-01 1.33879280e+00
3.62790450e-02 6.61957026e-01 4.79410887e-02 -8.48619282e-01
9.30711448e-01 6.05152965e-01 1.01949751e+00 -5.24899550e-02
2.53820747e-01 3.52798402e-01 1.15831964e-01 -2.79623508e-01
4.84080315e-01 3.03625758e-03 -1.54273137e-01 6.93605185e-01
3.33456472e-02 3.49257052e-01 8.84224176e-01 5.59383929e-01
1.15888238e+00 2.45422255e-02 7.35434115e-01 -5.78804731e-01
7.88049757e-01 -5.11626154e-02 4.08709079e-01 7.38369107e-01
1.24237604e-01 4.75011706e-01 1.10530591e+00 -2.16221228e-01
-7.39768565e-01 -7.20228016e-01 -2.79398233e-01 8.27311099e-01
-6.02252126e-01 -8.37017298e-01 -7.86153197e-01 -5.77422917e-01
-6.95983544e-02 3.72631252e-01 -7.25289047e-01 3.96846592e-01
-4.90526021e-01 -7.74809241e-01 8.96483958e-01 1.84716970e-01
-1.08712830e-01 -9.28024888e-01 -4.20764148e-01 2.23522440e-01
-2.23539889e-01 -1.08201027e+00 -3.90280783e-01 3.58951390e-01
-8.63578260e-01 -1.50299883e+00 -9.13610697e-01 -6.31342173e-01
9.69959617e-01 -1.50300130e-01 1.03728437e+00 4.17755060e-02
-4.19154495e-01 -6.53436035e-02 -4.48831171e-01 -8.68393242e-01
-5.81124187e-01 -7.98591319e-03 1.04914375e-01 -3.91938716e-01
5.61832666e-01 -5.41235745e-01 -2.20107719e-01 -4.59475033e-02
-7.55992115e-01 -2.88241059e-01 7.67779768e-01 7.89493263e-01
8.06690454e-01 4.59306091e-02 6.06227219e-01 -1.32817757e+00
8.66856277e-01 -5.49908817e-01 -3.93900603e-01 -6.63356930e-02
-8.57987046e-01 1.06358893e-01 6.49502695e-01 -2.42332518e-01
-8.42798054e-01 -1.56800359e-01 -2.16580361e-01 8.51549059e-02
-3.98298085e-01 4.82084632e-01 -4.16849911e-01 5.15705407e-01
5.26667416e-01 1.56564265e-01 -5.13825268e-02 -7.53518939e-01
4.89585400e-02 8.94463181e-01 3.32917005e-01 -6.78199828e-01
4.14041042e-01 3.79625380e-01 1.20937116e-01 -9.97889817e-01
-7.09829926e-01 -9.84402597e-01 -7.24712074e-01 -5.16469683e-03
6.58470929e-01 -8.14599872e-01 -8.75318766e-01 2.26712734e-01
-1.36544061e+00 1.75185263e-01 -1.21021554e-01 5.30093253e-01
-1.15772092e-03 5.10207236e-01 -4.02956992e-01 -7.40328789e-01
-5.36006510e-01 -6.37402654e-01 8.07326734e-01 1.37349650e-01
-8.34398210e-01 -8.16207588e-01 2.85162479e-01 7.69847631e-01
-2.51144856e-01 4.17600960e-01 8.23679209e-01 -9.89316583e-01
-4.08415049e-02 -3.17275226e-02 -6.45386875e-02 -8.12746510e-02
2.11731225e-01 9.86951515e-02 -8.28238904e-01 1.56437308e-01
-3.76867861e-01 2.13621527e-01 9.14628386e-01 3.80242169e-01
1.34631419e+00 -5.62806189e-01 -6.80416644e-01 -1.50783956e-01
7.94549286e-01 4.62534785e-01 6.42853022e-01 2.63720751e-01
6.80762112e-01 7.22813904e-01 5.27553678e-01 5.43770969e-01
2.59289473e-01 4.51201379e-01 -1.00938067e-01 2.79805392e-01
1.15652449e-01 -1.24608353e-01 1.01749964e-01 6.41069293e-01
-5.31093001e-01 -9.90107730e-02 -9.68988478e-01 6.92187130e-01
-1.50071251e+00 -9.37827706e-01 -8.00582528e-01 1.97735763e+00
1.20447457e+00 3.65483850e-01 3.57013762e-01 2.88683057e-01
6.25022352e-01 -2.25772575e-01 -1.14475511e-01 -3.55067253e-01
-3.57525200e-02 4.62637067e-01 2.97739089e-01 2.46708706e-01
-8.89571905e-01 5.59052765e-01 6.63582563e+00 7.53545403e-01
-5.10016739e-01 2.31028542e-01 4.47145849e-01 -7.77814090e-02
-4.42418277e-01 1.49147481e-01 -1.00059175e+00 6.96810901e-01
9.71818149e-01 -3.25739831e-01 -2.52410978e-01 6.96548641e-01
3.50734383e-01 -2.38668054e-01 -1.06511807e+00 7.87373841e-01
1.68026388e-01 -1.53123546e+00 1.45939425e-01 5.29249370e-01
5.14548481e-01 -5.98220885e-01 -5.56960940e-01 -1.03522286e-01
2.62187421e-01 -1.09757411e+00 6.03136718e-01 6.28607094e-01
7.08295286e-01 -8.57579231e-01 9.28025126e-01 3.45533401e-01
-8.63586426e-01 1.44415319e-01 -2.45094150e-01 -3.24524820e-01
-3.67403299e-01 1.28826356e+00 -1.08928430e+00 7.04956591e-01
6.16371155e-01 9.72600043e-01 -3.68833393e-01 7.68369496e-01
-5.01170039e-01 7.91092694e-01 3.50899637e-01 -5.91696024e-01
-2.43554413e-01 7.40163177e-02 7.61425793e-01 1.27625060e+00
5.74392229e-02 3.47304940e-01 -9.27715003e-02 9.02008653e-01
-4.38343227e-01 3.12855661e-01 -3.63820016e-01 -7.84008622e-01
4.99687105e-01 1.30518568e+00 -1.22651279e+00 -3.56240958e-01
1.91993006e-02 5.17808378e-01 -6.81833103e-02 -1.57545403e-01
-3.50582242e-01 -7.54334390e-01 4.87916410e-01 6.53407097e-01
-1.76555470e-01 1.31321236e-01 -4.24197078e-01 -9.36112642e-01
-9.63680819e-02 -7.15805411e-01 7.54408896e-01 -4.47079927e-01
-1.22096252e+00 3.32579166e-01 7.62256905e-02 -1.07873511e+00
9.24006999e-02 -6.07338965e-01 -2.48756930e-01 9.95560229e-01
-7.92147577e-01 -8.39250684e-01 1.11132860e-01 -6.32051006e-02
5.02978683e-01 -3.49741668e-01 8.18296075e-01 1.72659457e-01
-4.91535813e-01 4.54211175e-01 2.68663794e-01 2.69955754e-01
8.46607625e-01 -1.52453041e+00 -1.02714062e-01 6.18294835e-01
-1.59887269e-01 1.28824890e+00 7.21554101e-01 -9.63769913e-01
-1.12327158e+00 -8.40060651e-01 1.62802875e+00 -8.64089549e-01
6.87117338e-01 -4.70193863e-01 -6.04651868e-01 3.53138834e-01
2.34543711e-01 -5.03263772e-01 1.52636015e+00 1.77873358e-01
-6.28607068e-03 7.39555061e-02 -1.03704715e+00 4.40237910e-01
1.04565680e+00 -1.37573913e-01 -9.03203309e-01 1.74923807e-01
5.95559657e-01 -3.45891118e-01 -1.26142371e+00 -2.12419316e-01
7.23034143e-01 -2.95719057e-01 1.01917422e+00 -5.00223339e-01
8.35310876e-01 -3.35951746e-01 3.04995030e-01 -7.48885214e-01
-3.50528717e-01 -7.35113680e-01 -1.25688016e-01 1.38701761e+00
4.76648927e-01 -4.22526985e-01 4.77576226e-01 1.57464787e-01
-5.97872697e-02 -7.91344047e-01 -9.08958077e-01 -2.58148044e-01
9.55592170e-02 -6.73055425e-02 3.84175032e-01 1.02491844e+00
3.29380721e-01 5.73642969e-01 1.20670237e-02 -4.00480330e-01
6.46785378e-01 6.80911094e-02 4.95809197e-01 -1.81526637e+00
-2.47597218e-01 -6.41899526e-01 -1.06505327e-01 -3.80479485e-01
2.43283853e-01 -9.97423530e-01 -2.35564575e-01 -1.83447087e+00
7.05188394e-01 -2.72358984e-01 -2.19795033e-01 5.42189002e-01
-2.83540100e-01 -8.05775747e-02 -5.60876846e-01 5.70993543e-01
-6.35564446e-01 -1.75044909e-01 1.18611395e+00 -1.21450409e-01
-4.09805417e-01 5.34177897e-03 -1.57751572e+00 6.47343278e-01
7.19556212e-01 -7.54073560e-01 -1.96530536e-01 3.83683950e-01
5.62693536e-01 -2.02586263e-01 9.30958912e-02 -5.04044175e-01
6.08877949e-02 -3.48380536e-01 6.36720002e-01 -8.46143007e-01
-3.20578784e-01 -5.22913113e-02 3.14602852e-01 4.99091089e-01
-6.03101909e-01 -2.81049043e-01 1.14725970e-01 4.20415074e-01
5.81663148e-03 -4.99777913e-01 1.90831318e-01 -6.59067929e-01
-1.54250771e-01 -2.86386430e-01 -9.34097350e-01 -3.35254450e-03
7.95747995e-01 -4.41099294e-02 -2.58258849e-01 1.41389936e-01
-6.59085155e-01 1.03555806e-01 3.87204826e-01 3.82193565e-01
6.20573401e-01 -8.57832313e-01 -8.58629584e-01 -4.87407118e-01
1.42431319e-01 -3.83521356e-02 7.82676265e-02 1.00979805e+00
-4.35280353e-01 7.03941643e-01 -5.25308736e-02 -1.61695734e-01
-1.77448165e+00 4.61934566e-01 -4.75685805e-01 -6.80976450e-01
-3.06107014e-01 8.02924693e-01 1.66088436e-02 -2.88816333e-01
3.09724845e-02 -1.87997848e-01 -1.03909385e+00 6.95346713e-01
8.51469755e-01 5.45314193e-01 9.75876302e-02 -6.20466173e-01
-7.32515693e-01 2.24846855e-01 -3.35574150e-01 -2.42728770e-01
1.44562542e+00 2.62085885e-01 -9.16689634e-01 7.43250251e-01
8.31770062e-01 7.36109853e-01 -1.01413988e-01 2.27157757e-01
1.67842790e-01 -2.46485010e-01 -2.43679270e-01 -8.59832823e-01
-4.38052446e-01 4.27601188e-01 -7.60746300e-02 -1.56164870e-01
6.75329626e-01 3.97985339e-01 3.04992110e-01 1.03934435e-02
1.18149139e-01 -1.14040112e+00 -3.92535143e-03 3.13091397e-01
8.08096290e-01 -7.17691541e-01 5.01991630e-01 -6.23939216e-01
-1.49685040e-01 1.19346821e+00 3.02041978e-01 4.35688496e-01
5.84881842e-01 3.04195732e-01 -2.33950451e-01 -3.95335317e-01
-6.91000283e-01 3.41503359e-02 3.41200650e-01 4.16992962e-01
1.07708919e+00 1.50819734e-01 -1.36951780e+00 1.32108867e+00
-1.79716185e-01 1.96446568e-01 8.96648347e-01 9.54552948e-01
-3.52740973e-01 -1.51061523e+00 -7.28402674e-01 7.50989139e-01
-1.23403323e+00 -5.27598970e-02 -1.11702871e+00 3.61546099e-01
5.28320432e-01 1.00091207e+00 -1.87007204e-01 1.84925590e-02
-2.67061368e-02 5.41642196e-02 3.43683451e-01 -7.75614917e-01
-6.91431582e-01 3.02693039e-01 2.55086094e-01 3.84113453e-02
-7.82686234e-01 -9.55491245e-01 -1.51363862e+00 -1.03088312e-01
3.95567082e-02 8.06888580e-01 4.21575606e-01 7.67679155e-01
3.53625417e-01 9.60215628e-01 1.36967927e-01 -1.84995592e-01
1.08181410e-01 -9.76430774e-01 -5.04680753e-01 2.17806339e-01
-9.94965360e-02 -5.81772804e-01 -2.27993459e-01 2.54633605e-01] | [8.804450035095215, 8.619308471679688] |
f60b2aa6-a50e-4a80-a569-806be714a675 | a-large-dataset-for-improving-patch-matching | 1801.01466 | null | http://arxiv.org/abs/1801.01466v3 | http://arxiv.org/pdf/1801.01466v3.pdf | A Large Dataset for Improving Patch Matching | We propose a new dataset for learning local image descriptors which can be
used for significantly improved patch matching. Our proposed dataset consists
of an order of magnitude more number of scenes, images, and positive and
negative correspondences compared to the currently available Multi-View Stereo
(MVS) dataset from Brown et al. The new dataset also has better coverage of the
overall viewpoint, scale, and lighting changes in comparison to the MVS
dataset. Our dataset also provides supplementary information like RGB patches
with scale and rotations values, and intrinsic and extrinsic camera parameters
which as shown later can be used to customize training data as per application.
We train an existing state-of-the-art model on our dataset and evaluate on
publicly available benchmarks such as HPatches dataset and Strecha et
al.\cite{strecha} to quantify the image descriptor performance. Experimental
evaluations show that the descriptors trained using our proposed dataset
outperform the current state-of-the-art descriptors trained on MVS by 8%, 4%
and 10% on matching, verification and retrieval tasks respectively on the
HPatches dataset. Similarly on the Strecha dataset, we see an improvement of
3-5% for the matching task in non-planar scenes. | ['Sharat Chandran', 'Rahul Mitra', 'Arjun Jain', 'Utkarsh Gautam', 'Shuaib Ahmed', 'Sanath Narayan', 'Nehal Doiphode'] | 2018-01-04 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 1.07658960e-01 -5.81799805e-01 -1.03630058e-01 -5.25499642e-01
-1.09147811e+00 -5.90351403e-01 6.97379529e-01 8.54404643e-02
-2.10561112e-01 2.46002063e-01 1.05281048e-01 3.86673152e-01
-9.58804488e-02 -7.50067949e-01 -9.10960019e-01 -7.73380756e-01
1.27908438e-01 3.48314643e-01 5.30266941e-01 -4.71376121e-01
6.08948469e-01 6.67727053e-01 -2.09224677e+00 5.34963727e-01
1.71617463e-01 1.40582418e+00 1.15313895e-01 4.10588533e-01
4.28138673e-01 4.92477477e-01 -9.88713130e-02 -3.31227362e-01
8.28809619e-01 1.86537385e-01 -5.58753848e-01 -1.69915501e-02
1.50006771e+00 -2.98548311e-01 -5.59660375e-01 8.36219370e-01
7.90334165e-01 9.68436822e-02 5.29202104e-01 -1.06592083e+00
-7.73876250e-01 1.90576501e-02 -8.11559498e-01 1.36490837e-02
8.58741701e-01 1.28290266e-01 1.00515246e+00 -1.23447871e+00
8.96049559e-01 1.12540805e+00 8.21525514e-01 5.54420128e-02
-1.04812765e+00 -7.62088299e-01 -2.53063917e-01 4.66715932e-01
-1.72714436e+00 -4.45267707e-01 6.70824885e-01 -3.19009542e-01
1.24005353e+00 3.92113805e-01 5.73357284e-01 8.94501626e-01
5.15591428e-02 6.31105006e-01 1.28637362e+00 -4.65465307e-01
-1.17375337e-01 9.10004824e-02 -4.76922058e-02 5.33646941e-01
-9.50892642e-02 3.85973394e-01 -6.76866472e-01 -1.78535804e-01
7.48000741e-01 2.55775638e-03 -1.93082258e-01 -9.70109999e-01
-1.35582221e+00 7.01958537e-01 5.73929191e-01 -9.45932195e-02
-1.89162657e-01 6.23750798e-02 5.51587164e-01 4.20444578e-01
2.71895140e-01 2.23734066e-01 -3.92982513e-01 2.33873818e-02
-9.03807998e-01 3.03551167e-01 5.18309116e-01 1.38357401e+00
1.16703749e+00 -2.49310970e-01 2.83738784e-02 1.25815666e+00
8.94376114e-02 1.00396824e+00 1.65222511e-01 -1.03779435e+00
6.34844005e-01 6.73943281e-01 6.07096516e-02 -1.25499630e+00
-2.83768654e-01 -2.04754025e-01 -6.95788383e-01 -6.41972059e-04
-5.12072183e-02 6.78710282e-01 -9.71724391e-01 1.24211264e+00
2.27277547e-01 1.33916780e-01 -5.98121509e-02 9.11218822e-01
1.16381562e+00 3.73026639e-01 -5.03385544e-01 3.74443114e-01
1.20055413e+00 -1.04517019e+00 -1.56576827e-01 -1.02045454e-01
3.99080485e-01 -1.41710603e+00 9.28769827e-01 4.78582352e-01
-9.81584609e-01 -7.98991501e-01 -1.17993486e+00 -1.49176553e-01
-5.72678268e-01 2.70727217e-01 6.57074332e-01 4.97156709e-01
-1.26679146e+00 5.49953163e-01 -4.77939636e-01 -7.34048009e-01
2.32316896e-01 5.95517039e-01 -9.75412428e-01 -5.32000482e-01
-8.90539825e-01 8.67823899e-01 9.11911875e-02 -1.59953237e-01
-7.34141588e-01 -8.71365607e-01 -1.06771958e+00 -2.70130247e-01
-7.94092268e-02 -5.80667377e-01 5.07160544e-01 -4.28010404e-01
-1.12222540e+00 1.43781722e+00 3.68773825e-02 -6.07118271e-02
3.09549659e-01 8.04364402e-03 -5.00797749e-01 3.10533166e-01
2.87814826e-01 1.01793480e+00 6.21000767e-01 -1.29214990e+00
-5.99054813e-01 -4.56743687e-01 9.52335298e-02 1.69710025e-01
-7.45896026e-02 -3.51981670e-02 -8.81902933e-01 -5.45569301e-01
2.21851006e-01 -1.19098020e+00 7.58026540e-02 2.38360956e-01
-7.00773820e-02 1.44400522e-01 8.85155201e-01 -5.02048314e-01
7.99283564e-01 -2.26859689e+00 1.12155870e-01 4.14118141e-01
-3.34020585e-01 1.05689816e-01 -3.98832887e-01 7.44031847e-01
-2.68514544e-01 -4.50706273e-01 6.34588525e-02 -2.79860646e-01
-7.59461150e-02 2.91810632e-01 -1.49600700e-01 8.89122605e-01
-9.97069255e-02 4.51436520e-01 -5.96613348e-01 -4.67668414e-01
7.66641140e-01 6.21988833e-01 -6.01860344e-01 1.55438572e-01
3.74667048e-01 3.39797616e-01 8.47619995e-02 9.86708045e-01
1.16621232e+00 -9.05353799e-02 -3.53482485e-01 -6.63911641e-01
-4.65197973e-02 -1.21746823e-01 -1.62872934e+00 2.04959130e+00
-5.04533887e-01 6.22842491e-01 -3.19361895e-01 -7.93813705e-01
1.13939440e+00 -7.23054931e-02 5.59213459e-01 -9.99164283e-01
-1.95149511e-01 3.49157214e-01 -5.91453135e-01 -2.88323313e-01
6.19753480e-01 3.78568053e-01 2.78497692e-02 -1.08367473e-01
3.11257273e-01 -1.17688723e-01 1.60377011e-01 3.25740315e-02
8.68194699e-01 1.78014830e-01 2.49504372e-01 -4.50163424e-01
1.00443137e+00 -1.09949648e-01 3.17793906e-01 6.23921096e-01
3.93818542e-02 9.95848715e-01 -1.61598504e-01 -6.42237842e-01
-1.19218016e+00 -1.05082989e+00 -5.92305124e-01 8.34390163e-01
4.48670596e-01 -5.54832757e-01 -2.76726812e-01 -2.02131435e-01
3.36909890e-01 -5.92972971e-02 -6.56779587e-01 1.50904775e-01
-5.21147847e-01 -3.25126290e-01 4.99556124e-01 6.46106839e-01
9.38667059e-01 -7.36724555e-01 -4.05752540e-01 -3.72994512e-01
1.01303697e-01 -1.45776951e+00 -6.21430874e-01 -2.60969758e-01
-6.10332370e-01 -1.37105799e+00 -6.70151889e-01 -9.66957748e-01
7.05861628e-01 5.47270715e-01 1.26341486e+00 -3.83005887e-02
-5.58446884e-01 8.04191053e-01 -4.18218851e-01 1.47814322e-02
7.18237227e-03 -1.83409661e-01 -1.00144476e-01 -1.07600139e-02
3.20878744e-01 -5.76789916e-01 -1.18068981e+00 7.76019692e-01
-8.71898711e-01 -4.08212617e-02 5.61023057e-01 1.06428719e+00
8.83602977e-01 -5.25825739e-01 -1.37495980e-01 -7.69121289e-01
-1.73844621e-01 -1.69132113e-01 -8.13605428e-01 3.11217993e-01
-7.02798724e-01 -5.16815633e-02 2.87167370e-01 -2.53721207e-01
-6.98703110e-01 4.03946996e-01 -1.19297989e-01 -6.43239081e-01
-1.16657346e-01 7.73659125e-02 2.41198521e-02 -1.09542072e+00
6.40444398e-01 2.98143417e-01 -1.84173405e-01 -4.60341305e-01
4.18970793e-01 7.42757797e-01 6.69259667e-01 -4.96129841e-01
8.97341430e-01 7.85561323e-01 3.38630974e-01 -6.57218754e-01
-3.99323195e-01 -1.18010998e+00 -9.58161235e-01 -7.26552978e-02
4.24516767e-01 -1.37609756e+00 -6.24731421e-01 6.16231203e-01
-9.69764709e-01 2.39028156e-01 -4.44116518e-02 2.65376866e-01
-8.12291324e-01 6.08645916e-01 -3.85061473e-01 -2.68123895e-01
-3.67395371e-01 -1.40005195e+00 1.83217037e+00 -4.73479778e-02
2.40620479e-01 -1.02013421e+00 2.90056378e-01 6.26812339e-01
6.66600764e-01 4.86428142e-01 4.96522635e-01 -1.72086015e-01
-7.92681992e-01 -4.10349548e-01 -3.10714364e-01 3.90526563e-01
6.61434792e-03 -6.69334307e-02 -1.01954055e+00 -6.93064928e-01
-4.58292902e-01 -3.25953156e-01 1.02268314e+00 -3.86941358e-02
9.09718871e-01 3.27372663e-02 -3.23959231e-01 1.13960004e+00
1.92979133e+00 -1.97775140e-02 9.45547104e-01 6.53408766e-01
6.86161876e-01 4.02829885e-01 9.87354875e-01 5.30100703e-01
5.21833360e-01 1.32689738e+00 5.17995894e-01 -2.11097628e-01
-2.82028496e-01 -1.54704705e-01 2.88225859e-01 7.97113717e-01
-3.37122492e-02 2.67128080e-01 -8.35041523e-01 6.43961847e-01
-1.77948642e+00 -1.04699111e+00 -1.71407521e-01 2.42636156e+00
4.53906536e-01 -4.46822017e-01 -4.66578193e-02 1.31529063e-01
6.18566632e-01 3.08040291e-01 -2.98567414e-01 -1.81077346e-01
-4.87681925e-01 5.41115046e-01 1.05693805e+00 3.23066950e-01
-1.41360056e+00 8.98287654e-01 6.25914001e+00 8.28930736e-01
-1.40829086e+00 -3.83086167e-02 2.00273857e-01 1.93149045e-01
1.95713621e-02 1.78396985e-01 -9.39121902e-01 1.38470724e-01
5.80694199e-01 2.43193448e-01 4.77198958e-01 8.75207901e-01
-3.52294832e-01 -8.21738690e-02 -1.18107963e+00 1.65049517e+00
6.03959024e-01 -1.37067854e+00 2.45676056e-01 -6.06242977e-02
1.06313670e+00 4.41005379e-01 1.59978241e-01 1.00614116e-01
-2.32624829e-01 -8.23914707e-01 6.24396622e-01 3.55474681e-01
9.81039882e-01 -5.56292057e-01 8.94115567e-01 -2.63465583e-01
-1.48674417e+00 1.77347347e-01 -7.85680890e-01 2.87815273e-01
-3.07437330e-01 4.51216102e-02 -4.30773467e-01 9.14295077e-01
1.05135560e+00 1.25568318e+00 -1.25033915e+00 1.25562775e+00
2.14433685e-01 3.69616561e-02 -4.58205372e-01 4.96610731e-01
6.08373918e-02 -1.11619823e-01 2.97478646e-01 1.23332477e+00
2.69950509e-01 -2.27011651e-01 3.84952910e-02 4.30923611e-01
1.17523447e-01 4.02769536e-01 -8.69873464e-01 6.03524327e-01
4.12360460e-01 1.37906945e+00 -2.54880548e-01 -1.98705226e-01
-7.01611340e-01 1.02140439e+00 8.89957696e-02 1.72754422e-01
-6.47425711e-01 -3.48487109e-01 5.26593506e-01 1.85557663e-01
7.22959042e-01 -2.52480228e-02 1.83338195e-01 -1.21878660e+00
1.74318075e-01 -9.10069048e-01 5.57372570e-01 -8.99266541e-01
-1.33758104e+00 5.23456931e-01 1.45907149e-01 -1.70909703e+00
-2.04733685e-01 -8.84074032e-01 -1.73242047e-01 7.48285294e-01
-1.75614321e+00 -1.64097333e+00 -9.06475484e-01 1.01138997e+00
2.86344886e-01 -5.36715150e-01 1.02423131e+00 7.36494482e-01
-6.38523772e-02 7.92623639e-01 3.55682880e-01 1.68073580e-01
1.24628687e+00 -9.72517729e-01 3.64515483e-01 4.84935284e-01
2.95680851e-01 5.26706219e-01 4.10595983e-01 -1.14192046e-01
-1.93132079e+00 -1.02478588e+00 3.24477911e-01 -6.22773111e-01
3.80268663e-01 -3.81889790e-01 -5.41752517e-01 6.15550697e-01
3.45392317e-01 7.05475867e-01 6.72622681e-01 -1.50552705e-01
-9.46108758e-01 -6.07621849e-01 -1.24300802e+00 4.64821868e-02
1.12803090e+00 -7.57629156e-01 -1.76106736e-01 4.31643277e-01
1.37939334e-01 -9.60741520e-01 -1.40301263e+00 8.00782859e-01
8.08541417e-01 -1.39154315e+00 1.32658386e+00 -1.26530990e-01
1.71023145e-01 -4.19506609e-01 -8.04739892e-01 -9.90338027e-01
-3.27682823e-01 -2.94994354e-01 4.77662385e-01 1.25364995e+00
7.44740218e-02 -5.35427928e-01 6.20584786e-01 1.11115471e-01
-2.78397352e-02 -6.58893764e-01 -1.09212089e+00 -8.13282669e-01
-1.95408523e-01 -2.06346214e-01 5.26715815e-01 8.70710552e-01
-4.09727186e-01 -1.16202580e-02 -4.27645832e-01 3.29837769e-01
8.32728982e-01 4.02066350e-01 1.19305444e+00 -9.70507860e-01
-2.56586164e-01 -1.27777860e-01 -1.26464617e+00 -8.56117904e-01
-2.38703676e-02 -8.81942809e-01 -3.95620495e-01 -1.29211104e+00
5.14297783e-01 -4.97031987e-01 -3.98502260e-01 3.41706842e-01
2.33744025e-01 7.52295911e-01 2.73181021e-01 4.03686464e-01
-7.65543878e-01 3.10938179e-01 1.06930447e+00 -3.18059117e-01
2.66972840e-01 -4.44314510e-01 -2.20391616e-01 2.73158103e-01
2.96517313e-01 -1.45054579e-01 -1.10730827e-01 -4.17073935e-01
1.53384268e-01 -8.94579664e-02 5.23280084e-01 -1.17137897e+00
3.53298396e-01 9.66835842e-02 4.62090939e-01 -8.16806972e-01
7.49811769e-01 -9.72859859e-01 5.69402516e-01 1.63818657e-01
3.59135792e-02 3.76733929e-01 3.09320271e-01 5.00461459e-01
-6.40711248e-01 2.74721533e-01 8.87870073e-01 -7.87474141e-02
-9.89834905e-01 3.13120276e-01 5.59331119e-01 -1.41624376e-01
1.10390437e+00 -4.69253778e-01 -5.51556587e-01 -4.03459758e-01
-3.17089230e-01 1.93270952e-01 9.10821199e-01 7.40180850e-01
8.11488628e-01 -1.76238406e+00 -5.91609657e-01 6.29171908e-01
7.87550151e-01 -2.24227980e-01 2.71926999e-01 9.06667054e-01
-9.05429542e-01 6.87900841e-01 -6.70102358e-01 -1.02634895e+00
-1.55671287e+00 5.51759601e-01 2.40937397e-01 9.20312479e-02
-3.95388842e-01 5.13563275e-01 1.41491279e-01 -4.84046251e-01
3.23628187e-01 -2.55081594e-01 -7.48375058e-02 -1.03071034e-01
2.29098171e-01 4.03371900e-01 3.00096482e-01 -1.15630388e+00
-6.60408854e-01 1.39650810e+00 -3.33893336e-02 2.57977664e-01
1.56429279e+00 -7.69348443e-02 -2.56291509e-01 8.15405324e-02
1.66082251e+00 1.23557992e-01 -8.12862754e-01 -2.61185676e-01
-1.79300174e-01 -9.66233194e-01 -6.82513565e-02 -5.13160229e-01
-1.33640015e+00 7.72350132e-01 1.20476210e+00 -4.30343926e-01
1.21888745e+00 -1.81115214e-02 7.51810968e-01 2.95515239e-01
8.57244968e-01 -9.33306456e-01 3.48538123e-02 3.11439961e-01
1.17812145e+00 -1.54568267e+00 2.70678252e-01 -5.21214008e-01
-4.96206194e-01 1.14643919e+00 4.88725483e-01 -4.86955255e-01
7.77969241e-01 7.08274767e-02 8.86289850e-02 -3.31632465e-01
-4.46598351e-01 -1.57029614e-01 5.60895324e-01 7.13030815e-01
2.84104645e-01 -1.31692529e-01 2.85605956e-02 -2.19210908e-01
-1.12278484e-01 -3.33375454e-01 6.21968359e-02 8.52514148e-01
-1.02214232e-01 -1.47444344e+00 -4.95339245e-01 2.04127267e-01
-2.40566701e-01 -8.88043195e-02 -3.63503754e-01 9.33232725e-01
1.58421487e-01 6.58049047e-01 -1.04561582e-01 -6.16827011e-01
7.37997174e-01 -4.02065814e-01 8.98678482e-01 -4.02746230e-01
-6.25763834e-01 -6.94871321e-02 2.66881776e-03 -1.12542009e+00
-9.25850213e-01 -8.01466763e-01 -7.26998329e-01 -3.18379939e-01
-3.67835999e-01 -2.56299645e-01 5.82604766e-01 4.30132538e-01
5.01102746e-01 -2.26365879e-01 7.78474331e-01 -1.07466090e+00
-4.12283242e-01 -7.21505523e-01 -3.96119863e-01 9.38779593e-01
2.01058477e-01 -9.30530131e-01 -3.16177815e-01 -1.39583662e-01] | [8.13115406036377, -2.0152735710144043] |
480f05a0-fdd6-4ab9-8ed0-255ceb86559d | head-pose-estimation-of-occluded-faces-using | 1602.00997 | null | http://arxiv.org/abs/1602.00997v1 | http://arxiv.org/pdf/1602.00997v1.pdf | Head Pose Estimation of Occluded Faces using Regularized Regression | This paper presents regression methods for estimation of head pose from
occluded 2-D face images. The process primarily involves reconstructing a face
from its occluded image, followed by classification. Typical methods for
reconstruction assume that the pixel errors of the occluded regions are
independent. However, such an assumption is not true in the case of occlusion,
because of its inherent contiguous nature. Hence, we use nuclear norm as a
metric that can describe well the structure of the error. We also use LASSO
Regression based l1 - regularization to improve reconstruction. Next, we
implement Nuclear Norm Regularized Regression (NR), and also our proposed
method, for reconstruction and subsequent classification. Finally, we compare
the performance of the methods in terms of accuracy of head pose estimation of
occluded faces. | ['Rishabh Bindal', 'Michael Rotkowitz', 'Soumya Indela', 'Amit Kumar'] | 2016-02-02 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-4.92982604e-02 1.61487967e-01 -1.51342511e-01 -6.69849098e-01
-8.12753081e-01 -4.27981187e-03 4.71274592e-02 -3.52041900e-01
-2.39274800e-01 9.26542222e-01 4.16034162e-01 1.69163689e-01
3.92375961e-02 -4.33718503e-01 -7.71735966e-01 -9.27745759e-01
2.72590593e-02 3.15075785e-01 -5.56652606e-01 2.65182555e-01
-4.42317463e-02 8.01233947e-01 -1.56490302e+00 -2.43784130e-01
6.11057997e-01 8.84828269e-01 4.80375905e-03 1.09693252e-01
6.50782436e-02 5.97737491e-01 -4.48531896e-01 -1.98656619e-01
1.96741134e-01 -2.86708385e-01 -3.81803632e-01 5.39849222e-01
5.48102140e-01 -4.31842357e-01 -3.43756229e-01 9.81984138e-01
5.21002531e-01 1.83936000e-01 9.70936358e-01 -8.88427317e-01
-1.80643231e-01 1.48051515e-01 -9.27322268e-01 -1.62962928e-01
7.11374164e-01 -4.26185817e-01 5.15592039e-01 -1.13995934e+00
6.47728682e-01 1.48483872e+00 6.32862687e-01 6.57961190e-01
-1.29242945e+00 -7.07234979e-01 2.25404665e-01 -1.64673906e-02
-1.62121153e+00 -1.00173867e+00 8.55106115e-01 -4.95004147e-01
2.90328145e-01 1.77786663e-01 4.76532876e-01 1.04138243e+00
-4.69165966e-02 7.99181759e-01 1.28076720e+00 -4.73632574e-01
1.97296724e-01 7.67299011e-02 1.62387133e-01 6.78767383e-01
4.16679174e-01 -2.22921576e-02 -4.65738922e-01 -2.32763261e-01
6.40880823e-01 -1.83908209e-01 -3.03015798e-01 -3.01219314e-01
-5.45255959e-01 8.52204382e-01 1.01162449e-01 8.42366889e-02
-5.32669783e-01 -8.52553844e-02 1.52053192e-01 -1.17612816e-01
8.92379642e-01 -3.27857018e-01 3.42194773e-02 4.32711512e-01
-1.20890880e+00 1.42924100e-01 6.28668725e-01 7.84533799e-01
5.18159449e-01 2.78827459e-01 -3.27341929e-02 9.96705592e-01
7.60274589e-01 5.36733329e-01 1.80059791e-01 -1.03357911e+00
1.61671311e-01 1.34118930e-01 1.64000392e-02 -9.79477942e-01
-4.37734544e-01 -3.23169351e-01 -8.83569121e-01 4.24452543e-01
4.83076423e-01 -8.99446160e-02 -8.74290764e-01 1.75468540e+00
6.90121412e-01 2.27935076e-01 -1.24790289e-01 9.35519338e-01
8.99450302e-01 3.81126761e-01 -7.63545185e-02 -8.89386952e-01
1.07915688e+00 -5.72975993e-01 -1.29314935e+00 -7.51104578e-02
2.63114780e-01 -5.88811278e-01 4.80272770e-01 3.22562635e-01
-1.17859292e+00 -3.57974499e-01 -8.22870851e-01 -6.01787679e-02
9.81951579e-02 3.60335141e-01 3.77151281e-01 6.28549457e-01
-7.53378808e-01 2.94735938e-01 -7.60438621e-01 -1.62381694e-01
4.63787615e-01 4.79434222e-01 -7.35432386e-01 -1.58138528e-01
-6.14188612e-01 9.18049097e-01 -2.46774942e-01 5.34742236e-01
-6.74240112e-01 -3.09526026e-01 -1.03755820e+00 -2.18849182e-01
2.80559938e-02 -3.23709905e-01 8.62207294e-01 -6.60573661e-01
-1.46911836e+00 1.05017936e+00 -7.53215373e-01 5.61014842e-03
5.99848688e-01 -2.08942965e-01 -7.96322748e-02 -2.56953090e-02
5.07240556e-02 4.53969002e-01 1.23637629e+00 -1.23882902e+00
-4.11394313e-02 -9.75818217e-01 -3.30883682e-01 1.59050286e-01
8.63315091e-02 1.17925100e-01 -5.03120303e-01 -5.60514033e-01
8.36793363e-01 -8.72534692e-01 -7.30960295e-02 2.85658240e-01
-4.60022867e-01 -1.73354551e-01 7.33620405e-01 -1.04991114e+00
1.02627683e+00 -2.49735522e+00 1.83044195e-01 3.65614325e-01
1.13663837e-01 -1.81487843e-01 5.70632033e-02 -1.71484619e-01
-2.42766634e-01 -3.09126765e-01 -3.39557290e-01 -9.52106893e-01
-1.09294869e-01 1.97085261e-01 -1.29895166e-01 1.21447301e+00
-1.66900963e-01 4.86330003e-01 -3.50478917e-01 -5.99832714e-01
8.31827447e-02 8.39540184e-01 -5.40859938e-01 2.93602228e-01
1.07043922e-01 9.20512199e-01 -3.31821352e-01 6.81514502e-01
1.14371157e+00 2.81854570e-01 2.68878281e-01 -3.25837910e-01
-1.29139617e-01 8.65321308e-02 -1.46847498e+00 1.35913396e+00
-3.86798799e-01 5.43917060e-01 6.96106374e-01 -1.03558528e+00
9.96408641e-01 6.00531161e-01 7.40991652e-01 -2.74923921e-01
1.34052411e-01 1.49576634e-01 -4.76336777e-01 -5.81469238e-01
-6.00700006e-02 -1.62194774e-01 5.10601759e-01 3.73447597e-01
-2.66810000e-01 4.27067988e-02 -1.11908019e-01 -3.22061181e-01
5.68699598e-01 3.41961145e-01 3.07553351e-01 -2.66046554e-01
6.19845092e-01 -5.21754563e-01 7.79612541e-01 1.76563397e-01
-1.58028349e-01 6.95614517e-01 4.03630137e-01 -3.40900630e-01
-7.63507485e-01 -8.42469454e-01 -7.76374459e-01 7.35598505e-01
-1.09084107e-01 6.28480688e-02 -9.11019385e-01 -6.04166031e-01
1.88788757e-01 4.21020061e-01 -6.71516955e-01 2.54524738e-01
-7.27250755e-01 -8.34243298e-01 2.17577621e-01 3.69261444e-01
3.72243911e-01 -7.43995070e-01 -1.50425076e-01 -9.72607825e-03
-3.47783178e-01 -1.09813309e+00 -3.82430285e-01 -2.63892370e-03
-9.85374570e-01 -1.07391644e+00 -7.10753202e-01 -8.49771261e-01
1.17400444e+00 2.30473846e-01 6.92923188e-01 -3.97601314e-02
-2.28855029e-01 3.73589128e-01 3.88489328e-02 -1.31473869e-01
-1.46826059e-02 -5.48982918e-01 5.13036728e-01 2.74019063e-01
2.12329373e-01 -5.20255983e-01 -4.68601614e-01 4.50500995e-01
-5.17079949e-01 -4.00243074e-01 2.92427361e-01 7.53878057e-01
9.18127894e-01 -7.49829859e-02 3.68769318e-01 -8.62911284e-01
2.48725846e-01 -3.81566077e-01 -6.63430750e-01 2.03651801e-01
-2.13976428e-01 -6.93516359e-02 1.14389516e-01 -3.85160625e-01
-1.26366067e+00 4.18468118e-01 -2.83195436e-01 -5.86615145e-01
-2.59331703e-01 2.90135771e-01 -3.55083674e-01 -1.69480905e-01
4.57774073e-01 2.37030126e-02 2.33465165e-01 -8.17304671e-01
8.77043009e-02 7.68510997e-01 3.34775507e-01 -3.64215642e-01
5.50531566e-01 7.74972618e-01 2.36374393e-01 -1.28852808e+00
-8.96591365e-01 -3.96064699e-01 -7.94946373e-01 -1.45775512e-01
9.85354364e-01 -1.04067147e+00 -7.27656603e-01 2.17722878e-01
-1.18465114e+00 1.22975158e-02 -1.99431330e-01 8.88357162e-01
-5.88360965e-01 4.47425187e-01 -4.51986343e-01 -1.27055752e+00
-1.77675784e-01 -1.22883022e+00 1.05285132e+00 -1.02517955e-01
-2.65406165e-02 -8.36088300e-01 3.40965344e-03 6.28285468e-01
2.72351671e-02 3.46803367e-01 5.67413509e-01 -2.15804189e-01
-2.99171329e-01 -3.74777764e-01 -1.66969076e-01 3.59456211e-01
7.89659992e-02 -9.96974781e-02 -1.34351206e+00 -5.22907972e-01
4.92208898e-01 -1.49822563e-01 5.46412110e-01 9.40355957e-01
1.19151390e+00 -4.21790093e-01 -3.12256813e-01 7.41492391e-01
1.14751840e+00 -1.02249607e-01 7.12850332e-01 -1.11661710e-01
4.86603469e-01 1.02192950e+00 6.28032386e-01 5.14244795e-01
2.10515276e-01 9.71008897e-01 2.97160894e-01 -8.18328112e-02
-2.61160403e-01 -3.67579795e-02 4.75189000e-01 6.93348110e-01
-1.21270418e-01 1.36266679e-01 -3.74145001e-01 1.15649424e-01
-1.67681921e+00 -7.68036187e-01 -2.09946647e-01 2.38403964e+00
4.10021931e-01 -4.21801388e-01 1.58998266e-01 2.37876788e-01
8.63720000e-01 3.80421691e-02 -4.09163922e-01 -1.10566624e-01
-1.62789240e-01 -5.88501021e-02 2.76870638e-01 7.74476767e-01
-1.07489622e+00 6.03641748e-01 7.49340010e+00 6.51480615e-01
-9.51265216e-01 1.91185743e-01 5.77492297e-01 -1.65410805e-02
-7.74906203e-02 -3.15402091e-01 -1.02419531e+00 3.28623056e-01
4.38050658e-01 3.35063607e-01 5.08836746e-01 8.28853071e-01
4.45893317e-01 -2.43582815e-01 -9.85761881e-01 1.17424572e+00
3.73160362e-01 -5.34912288e-01 -4.09427613e-01 2.71219045e-01
5.94364583e-01 -4.94964242e-01 1.36307508e-01 6.42471835e-02
-2.67051607e-01 -1.13073778e+00 6.41982675e-01 5.14498711e-01
8.72808278e-01 -5.02422750e-01 5.71179748e-01 3.02010387e-01
-1.11422694e+00 8.12385306e-02 -5.34110785e-01 7.08824843e-02
1.55766696e-01 8.26387584e-01 -4.82388049e-01 1.01802878e-01
5.59257209e-01 5.31130314e-01 -9.99487936e-02 9.42016006e-01
-4.22356457e-01 2.91004062e-01 -4.36151147e-01 5.44159651e-01
-2.74791211e-01 -5.60730100e-01 6.53808892e-01 8.98124397e-01
2.31862396e-01 2.92565405e-01 1.83450580e-01 6.19449377e-01
-9.66910794e-02 4.20762002e-01 -8.25759053e-01 5.53672373e-01
3.10084492e-01 1.00167668e+00 -4.23694551e-01 -8.39202702e-02
-6.32235944e-01 7.53667533e-01 2.91957915e-01 4.71930712e-01
-6.87668562e-01 2.13472411e-01 6.21582329e-01 3.06079805e-01
1.71335697e-01 -2.73764461e-01 -2.70103246e-01 -1.12149167e+00
2.57174551e-01 -6.84770584e-01 2.62573451e-01 -4.83858973e-01
-1.09543633e+00 5.39396286e-01 1.17682248e-01 -9.90390897e-01
-1.19553022e-01 -4.96535450e-01 -3.69149357e-01 7.19246984e-01
-1.50192595e+00 -1.11386442e+00 -3.42544168e-01 6.78323448e-01
5.79070032e-01 6.14809431e-02 7.52693892e-01 4.19790834e-01
-9.82171595e-01 6.30386293e-01 2.08466023e-01 1.82483923e-02
6.31601334e-01 -8.77158821e-01 -2.41036057e-01 5.79436481e-01
-7.99766630e-02 6.12499714e-01 7.72753716e-01 -7.51436472e-01
-1.34938085e+00 -9.71114695e-01 7.57478714e-01 -1.20636925e-01
2.60629896e-02 -2.39416763e-01 -9.18613672e-01 8.19000721e-01
-3.31218570e-01 3.99720132e-01 6.46776974e-01 2.07408294e-01
-2.65410721e-01 -1.39018163e-01 -1.51124430e+00 2.11722240e-01
8.51364613e-01 -4.93066221e-01 -3.78612757e-01 6.26269460e-01
9.31444615e-02 -2.99535543e-01 -8.90144169e-01 6.14459455e-01
7.73014128e-01 -7.70313442e-01 1.04681230e+00 -2.84995794e-01
-1.05864845e-01 -2.35272810e-01 -5.31599641e-01 -8.53359938e-01
-3.53098541e-01 -2.54590511e-01 -1.64192379e-01 1.07557189e+00
2.18789190e-01 -5.02373457e-01 9.90170479e-01 6.07968330e-01
1.37826666e-01 -5.99213302e-01 -1.26648068e+00 -7.58355319e-01
-1.92906737e-01 -3.34275335e-01 3.26913208e-01 7.30024576e-01
-1.52201667e-01 8.75822231e-02 -5.48808694e-01 3.50131780e-01
1.14429092e+00 1.27123939e-02 6.72998130e-01 -1.18470323e+00
3.25733386e-02 -3.55828032e-02 -3.96951199e-01 -8.93680334e-01
7.05508351e-01 -6.09350443e-01 2.12866247e-01 -1.25362539e+00
1.71528801e-01 -4.16040480e-01 1.39168620e-01 2.70006508e-01
5.49896806e-03 5.63085973e-01 -9.42058191e-02 4.41245943e-01
-1.70993447e-01 6.38274729e-01 1.16596663e+00 -2.41461210e-02
-3.38626921e-01 2.84204215e-01 -1.68579280e-01 1.03880692e+00
4.96568173e-01 -4.92154121e-01 7.65672280e-03 -3.40271384e-01
-3.25385958e-01 2.53072858e-01 7.97741786e-02 -7.90045321e-01
-1.32468333e-02 -1.56293325e-02 6.47589266e-01 -5.54822683e-01
8.31905901e-01 -1.02691495e+00 3.43601018e-01 2.96536386e-01
-1.27778813e-01 -3.01304281e-01 -2.70997975e-02 3.75955224e-01
-1.36771664e-01 -4.43568140e-01 1.18090963e+00 2.26273257e-02
-7.68091455e-02 3.37735176e-01 -2.26481453e-01 -4.38658416e-01
9.18549001e-01 -3.93244207e-01 2.37893462e-01 -5.40221751e-01
-1.18496037e+00 -1.68922350e-01 4.39252198e-01 -6.75719678e-02
8.07027817e-01 -1.24635935e+00 -9.05678093e-01 5.36066115e-01
-2.73325205e-01 -1.72060564e-01 9.78336856e-02 1.15187919e+00
-3.99536282e-01 2.64953524e-01 3.29661444e-02 -6.31019235e-01
-1.53798115e+00 6.06962919e-01 2.59399623e-01 1.11266248e-01
-5.89351535e-01 6.24880612e-01 4.94317591e-01 -3.32483411e-01
6.31539881e-01 1.59273013e-01 -5.40851355e-01 2.59726077e-01
6.00113332e-01 6.18026018e-01 3.41555551e-02 -1.14534712e+00
-4.48703110e-01 9.56209064e-01 2.49103740e-01 -1.09244660e-01
1.41227043e+00 -3.82595271e-01 -4.73741412e-01 3.27345461e-01
1.32638490e+00 2.95606494e-01 -1.04116130e+00 -4.26324904e-02
-1.90116405e-01 -6.05227172e-01 2.60617942e-01 -6.48002774e-02
-1.25111115e+00 7.21711755e-01 7.46805370e-01 -3.30590844e-01
1.08554745e+00 -1.05227172e-01 2.91713566e-01 1.28471181e-01
2.90486634e-01 -8.93042803e-01 -2.67749995e-01 1.74052700e-01
1.10252118e+00 -1.27129817e+00 4.19801772e-01 -8.86775851e-01
-2.55302101e-01 1.02978957e+00 4.17631954e-01 -9.36759710e-02
8.90982389e-01 1.47254750e-01 5.74249551e-02 -1.13643855e-01
-2.64156193e-01 -1.77832514e-01 3.37975800e-01 4.92274433e-01
6.16980910e-01 -8.01927224e-02 -5.29490650e-01 1.57162949e-01
-2.22547427e-02 -2.38012061e-01 1.81186378e-01 7.84943700e-01
-2.99916297e-01 -8.42849553e-01 -1.01342261e+00 2.94734061e-01
-4.80249733e-01 3.79222929e-01 4.50426899e-02 4.60040390e-01
1.11664347e-01 9.83507633e-01 5.53840538e-03 5.47200255e-02
3.25011313e-01 1.07806968e-02 7.94980228e-01 -5.30662954e-01
3.62202264e-02 5.77646673e-01 -1.25292391e-01 -5.00107884e-01
-4.90741819e-01 -8.55147719e-01 -9.08249497e-01 -2.10728884e-01
-6.43680990e-01 2.33140424e-01 1.06893158e+00 7.92305291e-01
-9.75138322e-02 9.62022468e-02 1.02222300e+00 -9.99060869e-01
-4.74378526e-01 -1.02115309e+00 -9.68925178e-01 2.16617838e-01
5.15838742e-01 -9.85079885e-01 -6.61107659e-01 -8.70094970e-02] | [13.219351768493652, 0.3282988965511322] |
6397298f-aefd-4f9c-bb42-1a36f5146016 | benchmark-for-license-plate-character | 1607.02937 | null | http://arxiv.org/abs/1607.02937v2 | http://arxiv.org/pdf/1607.02937v2.pdf | Benchmark for License Plate Character Segmentation | Automatic License Plate Recognition (ALPR) has been the focus of many
researches in the past years. In general, ALPR is divided into the following
problems: detection of on-track vehicles, license plates detection, segmention
of license plate characters and optical character recognition (OCR). Even
though commercial solutions are available for controlled acquisition
conditions, e.g., the entrance of a parking lot, ALPR is still an open problem
when dealing with data acquired from uncontrolled environments, such as roads
and highways when relying only on imaging sensors. Due to the multiple
orientations and scales of the license plates captured by the camera, a very
challenging task of the ALPR is the License Plate Character Segmentation (LPCS)
step, which effectiveness is required to be (near) optimal to achieve a high
recognition rate by the OCR. To tackle the LPCS problem, this work proposes a
novel benchmark composed of a dataset designed to focus specifically on the
character segmentation step of the ALPR within an evaluation protocol.
Furthermore, we propose the Jaccard-Centroid coefficient, a new evaluation
measure more suitable than the Jaccard coefficient regarding the location of
the bounding box within the ground-truth annotation. The dataset is composed of
2,000 Brazilian license plates consisting of 14,000 alphanumeric symbols and
their corresponding bounding box annotations. We also present a new
straightforward approach to perform LPCS efficiently. Finally, we provide an
experimental evaluation for the dataset based on four LPCS approaches and
demonstrate the importance of character segmentation for achieving an accurate
OCR. | ['William Robson Schwartz', 'Gabriel Resende Gonçalves', 'David Menotti', 'Sirlene Pio Gomes da Silva'] | 2016-07-11 | null | null | null | null | ['license-plate-recognition', 'license-plate-detection'] | ['computer-vision', 'computer-vision'] | [ 1.12612717e-01 -5.18354177e-01 1.16712764e-01 -1.11193389e-01
-8.03096950e-01 -8.44686031e-01 4.20410901e-01 1.21624790e-01
-7.10375428e-01 5.66258609e-01 -5.83738744e-01 -2.19232008e-01
-6.25233948e-02 -7.84744620e-01 -7.09421694e-01 -5.35874426e-01
3.57581258e-01 7.42075980e-01 5.85231364e-01 -8.24989229e-02
6.87347233e-01 1.07739365e+00 -1.50726724e+00 -2.21051216e-01
9.60429311e-01 9.09085751e-01 3.24176729e-01 6.04109347e-01
-1.77578196e-01 2.73983330e-01 -7.34103978e-01 -5.87552965e-01
3.78930569e-01 -6.28132373e-02 -2.29454830e-01 4.77429509e-01
2.92616934e-01 -1.60788998e-01 -1.46068737e-01 1.21139383e+00
9.92967039e-02 2.17651710e-01 9.44580853e-01 -1.00969732e+00
-1.49351090e-01 1.89191848e-02 -3.65979731e-01 6.41639903e-02
1.29077837e-01 2.24240378e-01 7.98546195e-01 -8.47024620e-01
6.09593451e-01 5.38458169e-01 7.51659870e-01 2.96250135e-01
-9.36455131e-01 -4.12229836e-01 -2.81279713e-01 4.72527832e-01
-1.90037107e+00 -2.08509579e-01 5.81278324e-01 -7.62534678e-01
4.14615691e-01 3.80845308e-01 4.16946799e-01 4.61245835e-01
-2.58134246e-01 5.93550026e-01 1.12493932e+00 -3.72301072e-01
4.38765198e-01 4.89337951e-01 2.42497310e-01 2.99860150e-01
4.92994577e-01 -4.85045642e-01 1.92579702e-01 2.37268507e-01
6.80224121e-01 -7.74561465e-02 -7.36005977e-02 -7.25196749e-02
-9.07878935e-01 6.13851190e-01 -4.27764691e-02 5.91968238e-01
-1.15106210e-01 -1.78854167e-01 1.77466378e-01 -2.97037631e-01
6.06141612e-02 4.14402038e-01 8.25446174e-02 -2.86721408e-01
-1.33974671e+00 2.03984961e-01 6.71655297e-01 9.95985270e-01
9.41652775e-01 -3.36347036e-02 1.26641365e-02 9.73551512e-01
2.83791184e-01 8.47613931e-01 2.41634101e-01 -4.20861781e-01
7.55928457e-01 7.16757119e-01 2.45918795e-01 -1.18520164e+00
-2.13243097e-01 -1.70136690e-01 -5.09899378e-01 2.89190263e-01
7.25671411e-01 -5.83918169e-02 -6.60179973e-01 8.23825717e-01
-7.21017495e-02 1.34243086e-01 3.43236253e-02 9.77323294e-01
3.68313998e-01 9.35196996e-01 -1.84178919e-01 1.25385812e-02
1.50575697e+00 -7.69725084e-01 -5.67669392e-01 -2.94003695e-01
4.48320061e-01 -8.96075249e-01 7.70656765e-01 2.36160621e-01
-6.86220706e-01 -3.47215950e-01 -1.02538908e+00 1.78007886e-01
-6.57181501e-01 8.32051933e-01 -1.20472692e-01 9.24884617e-01
-6.61908388e-01 2.72952080e-01 -3.67245525e-01 -2.29877502e-01
4.64452691e-02 2.97283828e-01 -4.91624296e-01 -2.05241755e-01
-8.33696246e-01 9.67463911e-01 3.01053047e-01 3.77878845e-01
-3.16154540e-01 -2.02320352e-01 -5.32422841e-01 9.31859389e-02
4.29029882e-01 2.37849310e-01 5.85225821e-01 -8.45672429e-01
-1.49259806e+00 9.76150692e-01 6.44186363e-02 -2.96385139e-01
8.62014055e-01 1.88243300e-01 -6.79726899e-01 2.52361864e-01
5.54510988e-02 1.17555432e-01 7.11152136e-01 -1.19315004e+00
-7.24018633e-01 -4.01659906e-01 -2.01401293e-01 1.65048853e-01
-1.85255259e-01 1.28874451e-01 -9.93819296e-01 -4.19411749e-01
-1.05180301e-01 -1.11220324e+00 -3.65230367e-02 -4.51041251e-01
-5.90857804e-01 -1.82227254e-01 7.81595230e-01 -9.16692913e-01
1.15369785e+00 -2.44301987e+00 -3.76698464e-01 6.23470128e-01
-7.49607831e-02 6.48956001e-01 1.63624972e-01 2.29694024e-01
9.64803174e-02 1.06716797e-01 -6.35346949e-01 -3.73898029e-01
1.48045897e-01 -1.16165787e-01 -3.25280726e-02 8.18733335e-01
1.48543328e-01 4.64517117e-01 -2.03452453e-01 -6.81378663e-01
4.07037258e-01 2.98482180e-01 -5.14102951e-02 5.25636747e-02
4.90512475e-02 9.73814577e-02 -3.55914801e-01 7.67048836e-01
1.10333431e+00 2.86246628e-01 -1.17095210e-01 -8.04385245e-02
-6.41091228e-01 -5.09425282e-01 -1.43947315e+00 7.98484385e-01
-1.47521093e-01 1.08911228e+00 3.57935950e-02 -8.48407984e-01
1.52907908e+00 1.82889029e-02 3.58386964e-01 -6.32040679e-01
1.08275115e-01 7.40035176e-01 -2.10661530e-01 -4.32203621e-01
9.04748559e-01 1.27040297e-01 -1.09126918e-01 -1.16444960e-01
-4.04848725e-01 -8.21070746e-02 7.24861503e-01 -3.16467911e-01
7.88809478e-01 -2.49920428e-01 1.68631766e-02 -1.39879048e-01
1.12251115e+00 3.45241249e-01 3.69829327e-01 5.29254854e-01
-2.21501991e-01 7.37140894e-01 4.40651000e-01 -1.98605865e-01
-1.35751951e+00 -7.25378215e-01 -2.78718024e-01 3.56695175e-01
3.83580834e-01 4.37937398e-03 -9.79875326e-01 -3.81239921e-01
-2.63714399e-02 6.10288858e-01 -2.04768881e-01 4.34590936e-01
-8.06871593e-01 -7.21413434e-01 7.67332017e-01 2.18005449e-01
7.51659453e-01 -7.19968557e-01 -3.67280781e-01 8.74140486e-02
-1.46975920e-01 -1.68622065e+00 -4.10976052e-01 -2.63497919e-01
-6.34556234e-01 -1.07786131e+00 -1.06237197e+00 -6.99415386e-01
6.27156019e-01 1.24021232e-01 6.73133910e-01 -1.41141623e-01
-7.86736384e-02 2.02652395e-01 -3.88342023e-01 4.26698215e-02
-6.50493324e-01 1.17800400e-01 -2.15698019e-01 6.24361813e-01
4.18755919e-01 -5.75024448e-02 -3.27922463e-01 9.31726098e-01
-9.46436524e-01 -3.35468948e-01 8.41218293e-01 4.76601645e-02
6.00981891e-01 1.82347581e-01 9.76038128e-02 -5.34441471e-01
3.99264872e-01 -2.24855661e-01 -1.21456861e+00 3.41237903e-01
-1.74611509e-01 -4.05712038e-01 8.09624255e-01 -5.53940050e-02
-6.92514360e-01 3.05905640e-01 -4.27586108e-01 -3.21881264e-01
-6.34677649e-01 1.17234826e-01 -2.24585235e-01 -3.53585511e-01
2.50087947e-01 6.86295450e-01 -4.70260978e-02 -4.22162116e-01
-9.05127451e-02 1.07059360e+00 6.18520021e-01 -3.08481514e-01
1.05831599e+00 3.91261131e-01 1.17723443e-01 -1.50031698e+00
1.14046849e-01 -9.59341645e-01 -6.11848176e-01 -6.85403824e-01
1.04409063e+00 -5.56973934e-01 -9.68575656e-01 7.06390858e-01
-1.07209432e+00 1.34668872e-02 -4.14445773e-02 5.96948147e-01
-3.91613364e-01 7.57816315e-01 -2.58995682e-01 -9.78255510e-01
3.43951397e-02 -1.39161766e+00 9.22180653e-01 4.37935948e-01
3.30397367e-01 -7.09719181e-01 1.27709866e-01 6.59011722e-01
2.84737825e-01 1.37559578e-01 7.00816572e-01 -9.24440444e-01
-8.86611223e-01 -7.54092336e-01 -3.99614096e-01 5.65016925e-01
-2.94175386e-01 3.80188018e-01 -6.67109609e-01 1.73689246e-01
-3.90974373e-01 2.19005853e-01 7.00810015e-01 2.44004801e-01
7.75695324e-01 5.48454113e-02 -1.36135638e-01 3.59880000e-01
1.76206684e+00 4.33649719e-01 1.11274767e+00 6.53317392e-01
6.41404450e-01 5.60562551e-01 8.25228930e-01 4.40808147e-01
2.46545464e-01 1.05959153e+00 3.11281651e-01 1.84267998e-01
1.65676296e-01 8.89422745e-02 3.97296339e-01 6.60089731e-01
-4.49887484e-01 -1.47476166e-01 -1.12185121e+00 3.10933352e-01
-1.43156421e+00 -8.43413711e-01 -5.89067400e-01 2.53651094e+00
2.49402791e-01 1.02434248e-01 2.79597372e-01 3.73422831e-01
1.26598072e+00 -8.87456238e-02 -9.84250382e-02 -4.67344612e-01
-3.63418937e-01 -3.78106117e-01 9.74043310e-01 4.15402323e-01
-1.20273590e+00 8.17677557e-01 5.40006208e+00 1.17272079e+00
-1.21495450e+00 -2.71549284e-01 4.16320175e-01 5.93807638e-01
1.08518146e-01 -2.26809278e-01 -1.41147363e+00 8.41268420e-01
7.38386214e-01 1.81951672e-01 2.28518113e-01 8.40795875e-01
2.18561530e-01 -3.43448162e-01 -5.67649543e-01 1.13226175e+00
3.99283081e-01 -1.18336236e+00 -1.41151324e-01 2.15308890e-01
5.25279462e-01 -8.94396603e-02 1.98662952e-02 7.62790591e-02
-5.86036146e-01 -7.75841296e-01 7.96200156e-01 5.62570214e-01
7.60074914e-01 -8.16154599e-01 7.73832083e-01 3.35976273e-01
-1.22573435e+00 -5.19037805e-02 -4.95184809e-01 4.83582616e-01
1.03905953e-01 5.07447064e-01 -9.94609594e-01 2.76427865e-01
1.62736312e-01 4.28448051e-01 -7.81914890e-01 1.63827693e+00
-3.42138042e-03 3.92259032e-01 -3.17022562e-01 -4.13480192e-01
5.15743911e-01 -8.27503502e-01 8.36850464e-01 1.59018636e+00
5.42762041e-01 -3.19154710e-01 -2.19787791e-01 8.86604726e-01
-2.60537304e-02 4.74248737e-01 -3.52023840e-01 -2.93779001e-02
2.82732576e-01 1.29831541e+00 -1.16234076e+00 -1.85697347e-01
-3.20131272e-01 8.22012901e-01 -3.20133418e-01 2.24914432e-01
-1.07615280e+00 -6.89381659e-01 4.49949533e-01 3.71428221e-01
4.95586455e-01 -4.88804191e-01 -8.60008374e-02 -1.03459561e+00
2.42987767e-01 -6.06489897e-01 -5.66872209e-02 -5.40282965e-01
-8.37570786e-01 5.19089937e-01 -2.10859090e-01 -1.65974474e+00
-2.39628106e-02 -1.14405572e+00 -3.77463907e-01 6.96792901e-01
-1.61465979e+00 -7.67286479e-01 -2.91800380e-01 6.26598358e-01
4.98593926e-01 -2.58631408e-01 4.78656799e-01 7.79498219e-01
-8.49360883e-01 4.02881503e-01 7.23273516e-01 5.65189183e-01
4.39142048e-01 -1.17462862e+00 1.36965886e-01 1.03860855e+00
-8.20390508e-02 8.86103362e-02 6.15861714e-01 -6.07047498e-01
-1.22469485e+00 -1.02395976e+00 7.33545780e-01 -1.28278196e-01
5.36461592e-01 -4.07138348e-01 -8.02487910e-01 3.72994244e-01
-4.16388094e-01 -1.80975422e-02 5.43077052e-01 -5.29558659e-01
-5.95161431e-02 -3.52313936e-01 -1.03644502e+00 4.50281441e-01
1.59002051e-01 -3.54021698e-01 -2.70504713e-01 2.72522748e-01
-1.66018710e-01 -2.43212253e-01 -7.41819739e-01 6.48357347e-02
4.61025566e-01 -8.99340808e-01 8.23922455e-01 1.39791891e-01
2.73443937e-01 -7.13585854e-01 -1.02436401e-01 -8.05807471e-01
8.87428895e-02 -8.14950615e-02 5.21591783e-01 1.39751291e+00
5.08466482e-01 -4.78718221e-01 9.52832818e-01 6.45645857e-01
-1.46599770e-01 -2.68159866e-01 -1.11349285e+00 -9.47110295e-01
-3.06363940e-01 -5.52040637e-01 2.75767356e-01 5.27579248e-01
-5.06878674e-01 -1.95279047e-01 -3.26871455e-01 4.06067729e-01
7.48923302e-01 9.69359651e-02 8.15322578e-01 -1.20822120e+00
-1.45293802e-01 -7.25235105e-01 -1.10795581e+00 -7.97064900e-01
2.72900555e-02 -5.51366687e-01 2.39681855e-01 -1.41521060e+00
-1.43665880e-01 -4.84065801e-01 2.49028504e-01 -1.69501781e-01
2.39478812e-01 6.65434301e-01 5.45330942e-01 4.77757692e-01
-6.96655810e-01 6.13858625e-02 9.12827790e-01 -3.35482240e-01
-8.92861187e-02 3.91198575e-01 1.50051087e-01 7.39849806e-01
6.47312701e-01 -2.96606481e-01 2.79611051e-01 -8.94319937e-02
2.48375878e-01 -6.16585091e-02 1.81508243e-01 -1.27843070e+00
5.77521443e-01 1.08966976e-01 2.07605332e-01 -8.19097996e-01
4.52590734e-01 -1.09016955e+00 7.88663402e-02 2.60154963e-01
1.65768623e-01 -7.35735223e-02 1.21657290e-01 3.73633832e-01
-4.76513863e-01 -9.03240263e-01 1.01877880e+00 2.39108726e-02
-1.12337041e+00 -3.46894190e-02 -9.54877317e-01 -2.24935338e-02
1.38330460e+00 -7.23664761e-01 -1.83690950e-01 2.59209480e-02
-5.41838348e-01 -1.70979276e-01 4.70972270e-01 1.91686545e-02
5.52911520e-01 -9.85216498e-01 -7.61471808e-01 -2.91905887e-02
3.08776557e-01 -3.27499002e-01 4.13035423e-01 7.78849840e-01
-1.33637142e+00 6.46619141e-01 -3.45471323e-01 -5.68340838e-01
-1.36735451e+00 2.24941477e-01 4.66257840e-01 -1.95858955e-01
-2.90372401e-01 1.35155290e-01 -3.17644775e-01 -1.25914469e-01
-4.74172533e-02 -1.28413826e-01 -6.98353648e-01 2.93780714e-01
5.40531158e-01 7.76193202e-01 3.54965568e-01 -1.16093528e+00
-4.44236547e-01 1.17729402e+00 3.00819159e-01 -9.70182568e-02
1.01799262e+00 -3.17580514e-02 -1.16828918e-01 2.53043860e-01
1.13246679e+00 4.04930711e-01 -1.12433469e+00 1.58201694e-01
1.63876250e-01 -5.65196693e-01 -2.60701418e-01 -4.40223068e-01
-8.83883357e-01 7.49409020e-01 5.57578146e-01 2.95282006e-01
7.90151954e-01 -3.45925361e-01 4.98080254e-01 5.70424914e-01
2.30834588e-01 -1.50233817e+00 -3.69762480e-01 5.46542585e-01
5.34138680e-01 -1.04468977e+00 -2.47174621e-01 -4.75010544e-01
-6.25290871e-01 1.52352035e+00 2.14040369e-01 -1.18071228e-01
2.93561131e-01 2.80057341e-01 -1.96903303e-01 2.95630425e-01
3.45342040e-01 -4.90612805e-01 2.24531084e-01 2.69561261e-01
-9.75295156e-02 2.51149654e-01 -4.20019329e-01 4.77999389e-01
1.36835262e-01 -2.29795545e-01 8.29878509e-01 5.42659760e-01
-5.53483605e-01 -8.97958100e-01 -9.94779050e-01 2.12315857e-01
-4.11185145e-01 1.58143803e-01 -3.12401861e-01 1.05915499e+00
2.56568789e-01 7.68494546e-01 1.69950664e-01 -2.02365041e-01
5.90121269e-01 2.57243589e-02 7.76081011e-02 -1.99524447e-01
-2.89012641e-01 -4.62969802e-02 -8.15335885e-02 -9.00294334e-02
-2.93876261e-01 -8.14852059e-01 -1.06555021e+00 -3.00588787e-01
-2.93573439e-01 2.26632491e-01 1.27563846e+00 9.89302337e-01
-1.35842904e-01 1.40297502e-01 7.74943471e-01 -6.37906492e-01
-2.93191940e-01 -6.59100890e-01 -1.06579733e+00 4.11564201e-01
-8.72663930e-02 -3.63312393e-01 -4.01629448e-01 5.06938696e-02] | [9.81925106048584, -4.9673614501953125] |
b9cbfb5c-ccdd-4dea-b3a5-13f19b32ddaa | view-consistency-aware-holistic-triangulation | 2302.11301 | null | https://arxiv.org/abs/2302.11301v2 | https://arxiv.org/pdf/2302.11301v2.pdf | View Consistency Aware Holistic Triangulation for 3D Human Pose Estimation | The rapid development of multi-view 3D human pose estimation (HPE) is attributed to the maturation of monocular 2D HPE and the geometry of 3D reconstruction. However, 2D detection outliers in occluded views due to neglect of view consistency, and 3D implausible poses due to lack of pose coherence, remain challenges. To solve this, we introduce a Multi-View Fusion module to refine 2D results by establishing view correlations. Then, Holistic Triangulation is proposed to infer the whole pose as an entirety, and anatomy prior is injected to maintain the pose coherence and improve the plausibility. Anatomy prior is extracted by PCA whose input is skeletal structure features, which can factor out global context and joint-by-joint relationship from abstract to concrete. Benefiting from the closed-form solution, the whole framework is trained end-to-end. Our method outperforms the state of the art in both precision and plausibility which is assessed by a new metric. | ['Xu Zhao', 'Zhuo Chen', 'Xiaoyue Wan'] | 2023-02-22 | null | null | null | null | ['3d-human-pose-estimation', 'anatomy'] | ['computer-vision', 'miscellaneous'] | [-1.81414187e-01 2.37571850e-01 -3.35130394e-02 -2.30038762e-01
-7.14093864e-01 -4.76337314e-01 3.21756303e-01 -4.13343340e-01
-4.62390818e-02 4.51718658e-01 5.76351643e-01 3.68369907e-01
-1.90014765e-01 -4.36840534e-01 -6.04291141e-01 -2.99442440e-01
1.46141142e-01 4.11327988e-01 2.61708975e-01 -1.57437995e-01
-7.24206911e-03 2.94791788e-01 -1.30951142e+00 4.60922010e-02
6.42227352e-01 8.14026356e-01 8.86774138e-02 2.92860866e-01
4.83875930e-01 2.35352814e-01 -2.28487611e-01 -4.53338325e-01
4.10129547e-01 -1.15618929e-01 -3.40345830e-01 4.10399199e-01
7.11159885e-01 -6.77030802e-01 -3.24975610e-01 9.03762639e-01
8.10457945e-01 -3.18466872e-01 3.16928774e-01 -1.07160282e+00
-1.28425807e-01 -6.56052902e-02 -9.94380474e-01 -1.68260992e-01
8.88951302e-01 8.54069516e-02 7.66128302e-01 -1.40962470e+00
9.65684474e-01 1.31848192e+00 6.69737697e-01 2.64402747e-01
-1.02319694e+00 -4.57832724e-01 4.16969270e-01 -2.99677178e-02
-1.31328702e+00 -2.55079746e-01 1.11879814e+00 -4.29661453e-01
5.10252416e-01 3.40407938e-01 1.02914822e+00 1.29061711e+00
5.04754663e-01 7.00465977e-01 1.13266981e+00 -2.11028665e-01
-9.94789302e-02 -2.53132164e-01 -3.58396709e-01 8.84232163e-01
4.13425088e-01 2.17743620e-01 -7.64600039e-01 -1.04276650e-01
1.02489352e+00 3.29450130e-01 -3.42036456e-01 -9.79638934e-01
-1.33759320e+00 2.61551380e-01 4.84799922e-01 -1.98032796e-01
-4.55107808e-01 -1.16120331e-01 1.88243315e-01 -2.31231645e-01
3.27569276e-01 1.18422404e-01 -5.92466235e-01 -4.06683572e-02
-7.51662433e-01 4.13972497e-01 3.99381548e-01 1.13331091e+00
4.67342943e-01 -1.70956120e-01 1.37596056e-02 5.37434876e-01
6.76817298e-01 5.21768451e-01 1.53248876e-01 -1.12789643e+00
6.34200215e-01 9.59882796e-01 -3.94674875e-02 -1.09313953e+00
-7.43206322e-01 -7.65303791e-01 -7.51783729e-01 1.33990049e-01
3.34150642e-01 1.06866315e-01 -9.17942226e-01 1.64264405e+00
8.17425787e-01 -2.28201628e-01 -4.76997763e-01 1.21129215e+00
8.03105354e-01 -5.19392900e-02 -4.69868511e-01 -3.24147433e-01
1.59661710e+00 -8.24218154e-01 -5.81046104e-01 -3.86736482e-01
9.57922861e-02 -7.99768627e-01 7.96842933e-01 5.91421485e-01
-1.21522987e+00 -6.22014999e-01 -1.04431772e+00 -3.11598420e-01
9.76474658e-02 1.99759364e-01 2.53880411e-01 4.25844818e-01
-3.69767547e-01 3.03222179e-01 -1.12509584e+00 -2.90441871e-01
2.41366223e-01 2.83737212e-01 -8.08943331e-01 -3.32223952e-01
-8.39719236e-01 9.81363952e-01 2.56742984e-01 5.44641912e-01
-6.12803519e-01 -7.15173304e-01 -9.72197711e-01 -2.54293382e-01
7.11840212e-01 -1.33003128e+00 7.49728143e-01 -9.67861488e-02
-1.38761270e+00 9.26329076e-01 3.54973180e-03 2.34345868e-01
1.06993949e+00 -7.03857243e-01 -7.22697973e-02 4.78075445e-01
1.54638663e-01 3.42522293e-01 6.37973845e-01 -1.50147283e+00
-2.57998496e-01 -1.06203759e+00 -6.13311455e-02 6.82975233e-01
7.76707307e-02 -5.23797095e-01 -9.13779855e-01 -5.15209377e-01
1.04012990e+00 -1.12267244e+00 -2.47906059e-01 3.67649496e-01
-5.00766754e-01 2.46038184e-01 4.86652106e-01 -1.17838907e+00
1.19999945e+00 -2.09410381e+00 6.12383783e-01 3.40285778e-01
5.22658408e-01 -3.19562942e-01 3.88896674e-01 2.47410327e-01
-3.08087841e-02 -2.56071448e-01 3.32676955e-02 -5.19129217e-01
-1.09907806e-01 2.47186318e-01 1.94586977e-01 8.64038050e-01
2.84783840e-02 7.82898426e-01 -8.00943375e-01 -7.59688616e-01
4.56083506e-01 4.34057742e-01 -6.75036430e-01 3.75284940e-01
1.88575208e-01 9.13386941e-01 -6.59689844e-01 8.75859141e-01
1.02305079e+00 -2.26280868e-01 1.61446273e-01 -6.55324459e-01
9.17510390e-02 1.01216286e-01 -1.36550498e+00 2.58386159e+00
-2.33493105e-01 -3.68372768e-01 1.30350888e-01 -3.52514774e-01
5.30334651e-01 3.52336496e-01 6.82485580e-01 -3.43042463e-01
1.17637835e-01 3.16088349e-01 -1.45791188e-01 -5.15929222e-01
4.36890349e-02 -4.50266451e-02 -2.12597381e-02 -1.21475232e-03
1.56677272e-02 -1.41857013e-01 -2.90497094e-01 1.41847461e-01
6.76829875e-01 8.76493692e-01 6.33907378e-01 1.46275954e-02
4.05142814e-01 -3.30583096e-01 6.17538095e-01 3.84738818e-02
-9.79412571e-02 1.16410518e+00 3.90437186e-01 -2.78093964e-01
-8.74813557e-01 -1.53612137e+00 -1.18917309e-01 2.42093712e-01
2.66433090e-01 -5.65382242e-01 -5.13831377e-01 -7.27009833e-01
1.34113533e-02 1.50085598e-01 -6.17186129e-01 -1.93972602e-01
-6.63217008e-01 -4.19906527e-01 3.70306894e-02 5.24884105e-01
3.70733470e-01 -2.77465582e-01 -8.37712586e-01 3.34131904e-02
-4.12940085e-01 -1.35781634e+00 -4.04404223e-01 -7.84061328e-02
-1.18092072e+00 -1.27564585e+00 -8.55825305e-01 -1.97956279e-01
7.43572295e-01 2.28371173e-01 9.64177370e-01 -6.69852197e-02
-2.20468178e-01 3.40104818e-01 -2.29600027e-01 3.34867500e-02
1.48542989e-02 -1.39754191e-01 2.51709491e-01 -2.21422911e-01
-2.02199727e-01 -1.01237321e+00 -1.06996381e+00 5.48143446e-01
-4.88866091e-01 2.30458409e-01 8.01057398e-01 7.13550687e-01
6.99517012e-01 -3.99259478e-01 -8.84374976e-03 -5.44878900e-01
-1.27566814e-01 -2.21276253e-01 -2.15731770e-01 2.52928197e-01
-4.11324471e-01 -2.29647011e-01 1.34202257e-01 -1.10607304e-01
-1.18337214e+00 4.15443331e-01 -1.25369474e-01 -7.95498013e-01
-2.27976948e-01 3.38314235e-01 -5.88480234e-01 1.44934699e-01
4.74954426e-01 1.11412644e-01 2.39202097e-01 -6.52135074e-01
3.79174769e-01 9.89753827e-02 6.31964803e-01 -5.93319595e-01
8.78552973e-01 7.19687462e-01 2.82534242e-01 -4.65677351e-01
-9.62319911e-01 -4.24573570e-01 -1.17042220e+00 -4.14704114e-01
8.92276466e-01 -1.29854643e+00 -5.36102355e-01 1.68709919e-01
-1.31206214e+00 6.92885399e-01 1.61772773e-01 8.43045771e-01
-6.25086725e-01 8.00391793e-01 -4.87696797e-01 -6.74623072e-01
-1.45586997e-01 -1.28441274e+00 1.58439422e+00 -2.02322736e-01
-3.54583502e-01 -5.01091838e-01 -4.97024320e-02 6.01116359e-01
-3.60373497e-01 8.08453202e-01 5.92433155e-01 -1.27886981e-01
-7.48122692e-01 -3.41766745e-01 1.81971882e-02 2.00305790e-01
2.15549418e-03 -1.07572995e-01 -9.87149537e-01 -3.37789208e-01
3.51833433e-01 -3.44224423e-02 3.80038410e-01 5.47233343e-01
7.11461842e-01 -1.79126002e-02 -3.24670076e-01 8.04367304e-01
1.21321833e+00 -3.94308269e-01 5.77027142e-01 2.07747579e-01
1.07123268e+00 8.24607909e-01 7.67120540e-01 5.65484345e-01
5.96734703e-01 8.14598858e-01 7.76117027e-01 -3.22042368e-02
-2.05625400e-01 -5.65261364e-01 1.96791485e-01 1.09362578e+00
-4.92149770e-01 1.99952677e-01 -7.54768133e-01 1.44585259e-02
-1.79356873e+00 -5.99349201e-01 -2.21492723e-01 2.37236261e+00
4.53065693e-01 3.76462311e-01 9.76518393e-02 6.93689883e-02
4.13950205e-01 2.25009285e-02 -5.24850428e-01 4.20450658e-01
1.07546374e-01 -3.07547361e-01 2.20963389e-01 2.63929725e-01
-6.94822729e-01 5.55254042e-01 5.52791071e+00 5.53835690e-01
-8.74461472e-01 8.61665383e-02 2.66048282e-01 -2.62542754e-01
-2.67010540e-01 1.02332875e-01 -5.05466163e-01 2.15912640e-01
1.73338279e-02 3.98314804e-01 -1.39843106e-01 7.48365760e-01
9.91459787e-02 -3.27704102e-01 -1.23698330e+00 1.27395380e+00
2.71698356e-01 -7.75438249e-01 -4.55550663e-02 3.95401537e-01
6.03178561e-01 -3.48961800e-01 -1.23393677e-01 3.34263369e-02
-4.72316861e-01 -6.51294768e-01 9.69278336e-01 7.28056788e-01
7.69398332e-01 -6.92366540e-01 5.94867826e-01 6.47626936e-01
-1.29184985e+00 1.23114958e-01 -1.42553240e-01 -1.67748272e-01
4.90246385e-01 6.86405301e-01 -4.57675010e-01 1.39577568e+00
5.89989841e-01 7.60444283e-01 -6.48343921e-01 7.98501372e-01
-3.75234962e-01 -1.17955185e-01 -4.12851632e-01 7.18978107e-01
-2.50430852e-01 -1.43888816e-01 7.65290141e-01 5.79195201e-01
2.79972494e-01 1.03876173e-01 4.00143564e-01 6.95978701e-01
3.78673792e-01 -1.65017381e-01 -4.90257353e-01 7.02715099e-01
2.30851680e-01 1.38758206e+00 -6.78983212e-01 -4.03239764e-03
-4.01000351e-01 1.16773868e+00 1.50055259e-01 3.10606033e-01
-9.99411047e-01 4.37357813e-01 1.99865147e-01 4.34277624e-01
1.59793526e-01 -3.27235430e-01 -4.91230756e-01 -1.54044378e+00
7.42974341e-01 -7.99937904e-01 3.65584224e-01 -9.13019836e-01
-1.09323478e+00 4.80102926e-01 2.61498839e-01 -1.86392021e+00
-3.52789909e-01 -4.33478534e-01 -1.62259668e-01 7.07624435e-01
-1.14492869e+00 -1.46248496e+00 -4.64331836e-01 4.10898000e-01
4.42446709e-01 2.94207841e-01 6.51667714e-01 2.57266581e-01
-3.42722744e-01 4.76969838e-01 -6.53757572e-01 -1.67855635e-01
9.77671087e-01 -9.99249697e-01 3.44328843e-02 8.42752159e-01
3.34157310e-02 7.94306755e-01 6.83809817e-01 -9.07336771e-01
-1.78208935e+00 -6.18399262e-01 5.67131758e-01 -1.12326574e+00
1.77386463e-01 -3.12613636e-01 -4.48493123e-01 6.66868448e-01
-2.98165590e-01 2.35707715e-01 4.59883422e-01 1.72226414e-01
-4.27290559e-01 4.26234715e-02 -1.01035130e+00 5.17714560e-01
1.42071676e+00 -3.68730396e-01 -7.33295918e-01 1.30339731e-02
5.94595909e-01 -1.13864243e+00 -1.23576152e+00 6.32291913e-01
1.08649802e+00 -1.27740145e+00 1.29444504e+00 -3.18579286e-01
5.93184352e-01 -5.35805225e-01 -3.25861394e-01 -8.79048645e-01
-1.85382515e-01 -4.87280726e-01 -4.11393404e-01 9.97637808e-01
5.43458760e-02 -2.48256937e-01 7.89829135e-01 4.54255909e-01
-2.83430308e-01 -9.80486035e-01 -1.23093534e+00 -6.58035576e-01
-3.22459549e-01 -3.66893858e-01 2.19570816e-01 7.09760189e-01
-1.19614087e-01 4.92408395e-01 -6.59247041e-01 6.05896115e-01
9.05296862e-01 2.53734440e-01 9.73603904e-01 -1.25650346e+00
-4.60790008e-01 -4.13964540e-02 -5.91546118e-01 -1.26612604e+00
-4.72053379e-01 -5.27721763e-01 -1.66435823e-01 -1.38653100e+00
2.96275735e-01 1.94052294e-01 4.00620997e-02 -6.08014092e-02
-3.54617953e-01 1.65956095e-01 2.67679602e-01 3.11845303e-01
-4.50233310e-01 7.49626517e-01 1.78228641e+00 5.13959885e-01
8.44438076e-02 -1.88285224e-02 -4.48935062e-01 1.16225994e+00
1.40419468e-01 -3.98833543e-01 -4.73072439e-01 -4.57104087e-01
4.19788629e-01 4.66108739e-01 6.59938335e-01 -9.24993813e-01
7.87092298e-02 8.85223746e-02 9.50088680e-01 -1.02807391e+00
7.39705563e-01 -1.09063756e+00 3.51982594e-01 5.02496362e-01
1.84808403e-01 2.40599707e-01 -2.98581123e-01 8.17932665e-01
-5.78242876e-02 2.97923654e-01 6.25337899e-01 -4.31680113e-01
-2.16987893e-01 6.24685109e-01 4.77850854e-01 1.30065471e-01
7.57526994e-01 -6.69840515e-01 1.56633720e-01 -2.02335969e-01
-1.05048144e+00 2.42324293e-01 7.44090319e-01 3.52663547e-01
8.18741143e-01 -1.36332488e+00 -5.81921220e-01 3.11900616e-01
2.86995888e-01 5.51686585e-01 5.70411682e-01 1.41018152e+00
-4.84071225e-01 2.15260237e-02 -2.18673885e-01 -1.07485819e+00
-1.12193203e+00 5.06426513e-01 1.56454101e-01 -3.61383229e-01
-7.64035702e-01 5.87747157e-01 4.95247364e-01 -4.93150383e-01
1.78961277e-01 -3.84970725e-01 1.39021918e-01 -8.62733498e-02
2.55626351e-01 4.20987934e-01 8.62277672e-02 -7.01259494e-01
-6.80442452e-01 1.13302755e+00 1.07319176e-01 -7.91700259e-02
1.16447723e+00 -5.24950385e-01 8.54573101e-02 3.82868946e-01
1.15125906e+00 3.74815971e-01 -1.48084247e+00 3.32997292e-02
-4.32538629e-01 -7.54986405e-01 -3.37123364e-01 -7.90300906e-01
-8.42307985e-01 9.80215490e-01 6.28064632e-01 -5.52602112e-01
9.15482163e-01 2.39077001e-03 6.75267100e-01 -4.78560999e-02
6.31540656e-01 -1.00009036e+00 1.63170859e-01 1.62938520e-01
1.30789590e+00 -1.26748204e+00 7.56850421e-01 -9.98972833e-01
-4.98251289e-01 1.14322627e+00 8.51835012e-01 -1.87854856e-01
6.92019224e-01 -2.26387009e-02 -1.32336333e-01 -4.53660399e-01
-2.87806243e-01 1.26179978e-01 7.21477449e-01 5.53938448e-01
3.29980791e-01 -2.09574625e-01 -4.11625564e-01 6.64553463e-01
-2.99205065e-01 -1.47192463e-01 7.20508546e-02 9.64416265e-01
-4.87519652e-02 -8.12723339e-01 -4.92063671e-01 1.71436295e-02
-3.92029822e-01 2.54441768e-01 -1.17700405e-01 1.04918408e+00
2.61030436e-01 6.15020812e-01 -3.56639355e-01 -4.16271478e-01
7.82607853e-01 -1.83132425e-01 9.22505081e-01 -6.95868134e-01
-3.90929490e-01 7.91687608e-01 1.16281673e-01 -1.06113303e+00
-3.89366925e-01 -5.19348621e-01 -1.06187952e+00 6.86725378e-02
-4.17765230e-01 -3.48341942e-01 4.55424994e-01 1.05253029e+00
4.26049054e-01 4.28297073e-01 4.95814621e-01 -1.05397832e+00
-8.09921861e-01 -8.31522167e-01 -5.24798036e-01 5.64259827e-01
2.40913674e-01 -1.06454849e+00 -2.08539888e-01 -1.38773158e-01] | [6.9997878074646, -0.9658544063568115] |
62e3a2e8-c69e-40d1-b55e-3a2f7c551c2e | parting-with-misconceptions-about-learning | 2306.07962 | null | https://arxiv.org/abs/2306.07962v1 | https://arxiv.org/pdf/2306.07962v1.pdf | Parting with Misconceptions about Learning-based Vehicle Motion Planning | The release of nuPlan marks a new era in vehicle motion planning research, offering the first large-scale real-world dataset and evaluation schemes requiring both precise short-term planning and long-horizon ego-forecasting. Existing systems struggle to simultaneously meet both requirements. Indeed, we find that these tasks are fundamentally misaligned and should be addressed independently. We further assess the current state of closed-loop planning in the field, revealing the limitations of learning-based methods in complex real-world scenarios and the value of simple rule-based priors such as centerline selection through lane graph search algorithms. More surprisingly, for the open-loop sub-task, we observe that the best results are achieved when using only this centerline as scene context (\ie, ignoring all information regarding the map and other agents). Combining these insights, we propose an extremely simple and efficient planner which outperforms an extensive set of competitors, winning the nuPlan planning challenge 2023. | ['Kashyap Chitta', 'Andreas Geiger', 'Marcel Hallgarten', 'Daniel Dauner'] | 2023-06-13 | null | null | null | null | ['misconceptions', 'motion-planning'] | ['miscellaneous', 'robots'] | [-4.25773486e-02 4.60632682e-01 -4.92788136e-01 -2.50029266e-01
-7.53974140e-01 -7.80780911e-01 1.16024327e+00 5.41358665e-02
-4.83151615e-01 8.80468786e-01 6.01025462e-01 -6.02855504e-01
-1.44862220e-01 -7.28839517e-01 -6.58462346e-01 -4.87811774e-01
-5.43954968e-01 9.99148846e-01 6.72023773e-01 -5.88859200e-01
4.17720348e-01 6.12866879e-01 -1.39401650e+00 -3.89263928e-01
5.73305428e-01 4.13810104e-01 3.92615706e-01 7.25333333e-01
3.23877275e-01 7.35750675e-01 8.13407227e-02 -6.57928661e-02
5.23881912e-01 6.17721640e-02 -8.70281816e-01 5.13524972e-02
4.40331578e-01 -5.85177958e-01 -8.85706604e-01 6.60274148e-01
4.07771856e-01 5.60172975e-01 3.07578355e-01 -1.61931980e+00
2.89267898e-01 2.19790488e-01 -2.55206108e-01 1.88445866e-01
5.31510532e-01 7.78979540e-01 1.00138581e+00 -4.00171906e-01
9.33773220e-01 1.17361557e+00 7.76770413e-01 2.47971371e-01
-8.80255997e-01 -1.66382879e-01 5.19962907e-01 5.93115270e-01
-1.10268188e+00 -6.85547292e-01 6.31320834e-01 -4.33871567e-01
1.32670474e+00 9.12871957e-02 5.55956841e-01 9.08615112e-01
5.16422510e-01 8.07967424e-01 6.55296087e-01 5.55156171e-02
2.67640471e-01 -4.71911490e-01 -1.37451515e-01 9.01609719e-01
1.68880582e-01 6.27758026e-01 -3.00654918e-01 -4.04605679e-02
4.82024908e-01 -2.89899558e-01 -8.51893649e-02 -7.52698302e-01
-1.58964407e+00 7.78653622e-01 4.12079215e-01 -6.22144714e-02
-5.01255870e-01 4.19930696e-01 3.59729588e-01 -4.85639870e-02
-4.00304757e-02 4.68402594e-01 -3.47887278e-01 -3.25035900e-01
-8.75254095e-01 7.22165525e-01 8.79032195e-01 1.38997698e+00
1.08582497e+00 -1.55064553e-01 -1.96308553e-01 1.69340849e-01
3.81446481e-01 3.18048954e-01 -1.73807591e-01 -1.35013270e+00
6.81260169e-01 1.57074854e-01 5.88740051e-01 -1.04610860e+00
-1.13371277e+00 -1.63684592e-01 -2.37036198e-01 2.18809158e-01
6.55463040e-01 -4.46040809e-01 -1.00261176e+00 1.75504422e+00
4.68777865e-01 4.61809337e-01 4.17636111e-02 8.70164394e-01
3.60593915e-01 6.23651385e-01 1.23779289e-02 -7.78702721e-02
9.47152495e-01 -1.50045431e+00 -5.42027771e-01 -8.91212523e-01
8.29631925e-01 -5.92430532e-01 7.16671169e-01 9.74672064e-02
-9.00253594e-01 -2.97955096e-01 -1.27052450e+00 -4.63984441e-03
-4.14645851e-01 -2.60006845e-01 1.04596841e+00 3.70538175e-01
-1.25766826e+00 4.05019492e-01 -9.86122906e-01 -6.71843708e-01
2.56233394e-01 1.52816549e-01 -3.87495220e-01 -3.15626472e-01
-9.62421000e-01 1.44981933e+00 4.24268901e-01 8.09116587e-02
-1.29572654e+00 -6.18093014e-01 -1.01291323e+00 -2.21947759e-01
8.12672496e-01 -6.15153491e-01 1.54663050e+00 -8.52745846e-02
-1.50370622e+00 3.70711029e-01 -1.65358812e-01 -6.98152244e-01
7.62912989e-01 -2.32144728e-01 -2.28040665e-01 -2.66706526e-01
3.07253480e-01 9.38795626e-01 1.93144277e-01 -1.31441391e+00
-1.15787613e+00 -1.24458291e-01 3.51414174e-01 6.04839563e-01
6.11618578e-01 -4.98729736e-01 -7.36880422e-01 6.61944151e-02
1.92029685e-01 -1.29727745e+00 -9.41390932e-01 -3.25068682e-01
-3.60828221e-01 -3.72652709e-01 7.55547106e-01 -4.66813803e-01
9.38019276e-01 -1.68875015e+00 -2.44429614e-02 1.58337131e-01
8.16974342e-02 -1.60765164e-02 -1.86704546e-01 8.49913895e-01
5.37765443e-01 -1.88446581e-01 3.67179140e-02 -2.70016700e-01
4.24497366e-01 5.03970146e-01 -3.10724258e-01 7.13524401e-01
-1.57985330e-01 1.15282440e+00 -1.38708150e+00 -5.05123675e-01
6.68856978e-01 -3.13852206e-02 -4.11286145e-01 -9.64355692e-02
-6.25054419e-01 4.35022384e-01 -6.68042839e-01 3.30921739e-01
4.92336065e-01 -3.61231081e-02 3.13472062e-01 1.04309909e-01
-5.83602190e-01 5.01575947e-01 -1.10162699e+00 2.03505135e+00
-2.50555992e-01 9.56020832e-01 6.36790693e-02 -4.58025128e-01
4.97008532e-01 -7.40217231e-03 6.41444087e-01 -8.91433001e-01
-1.75009340e-01 1.03424154e-01 -1.85981944e-01 -4.29885238e-01
8.64534378e-01 2.79368281e-01 -2.23989204e-01 1.27603456e-01
-1.58072352e-01 -4.06556875e-01 2.97076225e-01 1.75234780e-01
1.40573740e+00 4.07685548e-01 3.66467386e-01 -3.50578398e-01
1.88655987e-01 7.87199080e-01 6.28452837e-01 8.90989840e-01
-7.14300573e-01 3.49148929e-01 3.56995851e-01 -7.34903276e-01
-1.02671874e+00 -7.63087034e-01 2.27892265e-01 1.06882346e+00
6.60927057e-01 -4.41100717e-01 -5.13538182e-01 -6.97286367e-01
1.72276478e-02 1.03323030e+00 -4.09427196e-01 3.10005814e-01
-1.04827249e+00 -5.57951450e-01 5.09703875e-01 4.04729426e-01
3.97030085e-01 -8.50227237e-01 -1.07692528e+00 4.82855201e-01
-2.62033254e-01 -1.40448666e+00 -4.64484811e-01 -9.04121995e-02
-5.03776729e-01 -1.20911860e+00 -1.99484944e-01 -5.56358397e-01
1.95052743e-01 5.21007180e-01 1.23453009e+00 9.67647787e-03
2.90657669e-01 4.33493018e-01 -4.79514077e-02 -1.97049230e-01
-1.67530432e-01 2.25139081e-01 -5.46510033e-02 -5.55216312e-01
4.81309630e-02 -5.10195851e-01 -5.04472315e-01 4.80538517e-01
-5.50483465e-02 4.43458050e-01 6.21920824e-01 4.17583793e-01
3.94849241e-01 1.57561973e-02 5.39868355e-01 -6.45355642e-01
2.93893397e-01 -6.38034821e-01 -9.42424297e-01 4.95895781e-02
-7.83066988e-01 6.89222664e-03 1.83334827e-01 2.43694410e-01
-1.15015507e+00 2.33817995e-01 -2.11539656e-01 2.51088310e-02
-3.33838433e-01 5.59370756e-01 -1.60798490e-01 -2.46761322e-01
5.11170208e-01 5.60391918e-02 -1.87607765e-01 -2.84037236e-02
5.30077815e-01 -6.62646741e-02 8.44075203e-01 -2.93078452e-01
8.07966769e-01 8.21443260e-01 3.29901576e-01 -7.35391200e-01
-5.90519369e-01 -5.85077643e-01 -6.70517743e-01 -4.79682922e-01
8.18080366e-01 -9.29814637e-01 -7.03616381e-01 1.71414748e-01
-1.25504410e+00 -8.71514142e-01 -1.05183728e-01 5.28680980e-01
-1.18282068e+00 3.21384937e-01 -1.38763189e-01 -5.45043647e-01
2.02966794e-01 -1.39817607e+00 9.95063722e-01 1.09337211e-01
-9.51300375e-03 -1.14168584e+00 5.61567843e-01 1.89884648e-01
3.91444832e-01 4.16211963e-01 6.16811275e-01 -5.80467641e-01
-1.25473809e+00 -1.36793062e-01 -1.45363435e-01 -6.43972576e-01
-1.84853256e-01 -3.33571017e-01 -7.05554605e-01 -2.04783559e-01
-5.86791813e-01 -1.07553490e-02 8.35743487e-01 4.70013708e-01
3.51562411e-01 -2.11284146e-01 -8.82042050e-01 5.80042243e-01
1.37085211e+00 2.23097444e-01 6.81472540e-01 7.11656868e-01
7.98693419e-01 7.00511813e-01 1.06150746e+00 4.22322512e-01
1.28853071e+00 9.40259635e-01 7.90024698e-01 1.68775633e-01
-2.06746355e-01 -7.02253342e-01 9.67495218e-02 2.27405712e-01
-7.39645660e-02 -4.40163642e-01 -1.45007324e+00 8.79404902e-01
-2.42249632e+00 -1.08474076e+00 -1.07051633e-01 2.05098534e+00
2.50491321e-01 3.38364273e-01 4.10965644e-02 -4.89509344e-01
3.17546159e-01 7.08221436e-01 -6.07058525e-01 4.39517200e-02
2.61872243e-02 -6.48599684e-01 1.16381168e+00 1.06521904e+00
-1.39763248e+00 1.46825635e+00 7.30561781e+00 6.14757299e-01
-7.28127122e-01 -1.48450281e-03 1.93745926e-01 6.45355657e-02
-4.09455776e-01 5.07302344e-01 -9.59708631e-01 1.17546655e-01
1.04175782e+00 -1.98062360e-01 6.42290592e-01 7.39821672e-01
5.21168113e-01 -5.28453946e-01 -1.23911595e+00 6.47399604e-01
-8.66323188e-02 -1.75459087e+00 -4.73114818e-01 2.47191668e-01
9.21679378e-01 8.14703524e-01 -3.12593818e-01 4.50520009e-01
7.72339344e-01 -1.08435023e+00 7.50482142e-01 4.65386152e-01
1.31163388e-01 -6.87078476e-01 5.75650871e-01 6.86212897e-01
-1.27904415e+00 -5.84805198e-02 -7.44592324e-02 -2.41259560e-01
7.94693947e-01 3.39787342e-02 -1.24064314e+00 7.64469922e-01
1.89948633e-01 8.14522088e-01 -4.02429163e-01 1.36309910e+00
-1.72755584e-01 2.28095204e-01 -5.38358450e-01 1.29982129e-01
9.49893475e-01 -6.78412989e-02 8.47639978e-01 1.25402045e+00
-8.12692642e-02 2.99537517e-02 7.88734317e-01 3.83714259e-01
4.05933112e-01 -5.61945915e-01 -1.05272686e+00 3.43573332e-01
5.06169140e-01 1.07746851e+00 -6.53484344e-01 -1.07825905e-01
-4.29270208e-01 3.60729069e-01 2.37504125e-01 4.96848494e-01
-9.19202566e-01 1.63099274e-01 7.69169390e-01 -6.36612326e-02
2.50525415e-01 -8.80577147e-01 -3.58580321e-01 -5.70132136e-01
-1.39837533e-01 -4.20789540e-01 1.80536360e-01 -5.63912988e-01
-5.67392111e-01 2.91529953e-01 3.50252390e-01 -1.17143381e+00
-6.47935092e-01 -3.35775495e-01 -5.56074440e-01 4.49991405e-01
-2.11270022e+00 -1.32866740e+00 -2.60676712e-01 3.83202583e-01
8.67614925e-01 -5.53330630e-02 4.19334620e-01 1.61894470e-01
-1.90582514e-01 9.55540687e-02 -1.17446110e-01 -2.59156406e-01
3.84706408e-01 -1.02194583e+00 1.00640297e+00 1.04038644e+00
-6.80211037e-02 -3.10861524e-02 1.10243201e+00 -7.64902592e-01
-1.81305027e+00 -1.10914552e+00 1.00986814e+00 -7.51444995e-01
5.95402241e-01 -1.19096644e-01 -3.12717915e-01 1.20744288e+00
3.18830937e-01 -2.27239564e-01 -6.98918477e-02 9.66597423e-02
1.11473113e-01 3.11796099e-01 -7.62002289e-01 1.04241645e+00
1.34923530e+00 -1.44180492e-01 -3.85489106e-01 6.58213854e-01
6.95178747e-01 -8.59204888e-01 -4.36569124e-01 6.65067732e-01
5.15261292e-01 -1.05628538e+00 9.41639721e-01 -5.09409070e-01
-1.51989087e-01 -4.66230571e-01 -2.98997045e-01 -1.21590745e+00
-4.64192212e-01 -1.13879120e+00 -6.53905421e-02 5.85898101e-01
6.06293380e-01 -5.51948607e-01 1.13499105e+00 7.42180884e-01
-7.28488445e-01 -6.76562846e-01 -1.04820454e+00 -6.62382245e-01
-2.31859520e-01 -7.50148475e-01 7.01626480e-01 6.17146790e-01
5.35777286e-02 2.31968552e-01 -5.59579253e-01 5.35749793e-01
6.07500613e-01 -7.10921595e-03 1.26526642e+00 -8.75111163e-01
1.49187952e-01 -3.96042138e-01 -4.81062174e-01 -1.51640332e+00
3.62457663e-01 -4.86088425e-01 5.82638383e-01 -2.04644132e+00
-2.65492976e-01 -6.91511929e-01 2.55731642e-01 3.31266522e-01
1.61977813e-01 -4.24137801e-01 2.99775720e-01 1.32972032e-01
-9.96803641e-01 6.31540000e-01 1.36062920e+00 -2.34519262e-02
-3.55510026e-01 1.37481004e-01 -3.27231854e-01 9.15935934e-01
7.61352539e-01 1.17780948e-02 -6.95687711e-01 -7.25494206e-01
1.96332470e-01 3.97559792e-01 3.92592996e-01 -1.25326967e+00
8.78979146e-01 -8.60448837e-01 -2.84974575e-01 -1.09938931e+00
4.81192201e-01 -8.10504317e-01 2.25414172e-01 2.43428558e-01
8.19619372e-03 3.08430791e-01 3.56500208e-01 8.59437287e-01
1.28610075e-01 1.54844403e-01 3.10569614e-01 -1.83868363e-01
-1.75334239e+00 5.82865119e-01 -4.51523840e-01 1.15953445e-01
1.26411021e+00 -4.19010609e-01 -6.40779555e-01 -5.05357206e-01
-3.98442566e-01 9.74662006e-01 5.55779040e-01 5.67711949e-01
4.02371824e-01 -1.10866940e+00 -6.68140471e-01 -6.29506558e-02
1.61213905e-01 2.11251214e-01 2.96956241e-01 1.01676464e+00
-6.37538552e-01 8.72527122e-01 -2.54694134e-01 -5.79525828e-01
-6.58313394e-01 4.44506407e-01 3.54598522e-01 -5.41242063e-01
-9.49554861e-01 5.07281959e-01 9.63748526e-03 -6.25195503e-01
1.63590148e-01 -1.27136663e-01 -2.73560375e-01 -1.49207994e-01
3.81781459e-01 6.84032202e-01 7.15500638e-02 -7.76083112e-01
-4.79994625e-01 2.66214907e-01 1.92517534e-01 -3.65308702e-01
1.21089113e+00 -5.06006181e-01 4.03402269e-01 1.89695090e-01
7.11005092e-01 -1.16912216e-01 -2.04498172e+00 -7.47651309e-02
3.45423251e-01 -4.09109801e-01 2.38971367e-01 -6.73854351e-01
-6.04991674e-01 4.01833892e-01 2.11240575e-01 -3.57730500e-02
4.89657283e-01 -2.95425821e-02 7.42095232e-01 7.05195785e-01
1.01637876e+00 -1.27640080e+00 -3.42788249e-01 1.06728172e+00
8.70871425e-01 -1.41772902e+00 1.33540675e-01 -2.60165244e-01
-7.58113027e-01 7.16819167e-01 5.96017957e-01 -2.37450361e-01
4.91589487e-01 1.15821391e-01 -6.97819963e-02 -3.00709814e-01
-1.06594574e+00 -4.85912353e-01 1.20098785e-01 7.94254780e-01
-3.13829005e-01 2.72078604e-01 2.72939005e-03 -1.48133915e-02
-3.68303329e-01 -2.05706224e-01 5.57819605e-01 1.07428575e+00
-7.31775165e-01 -8.21934044e-01 -8.06498080e-02 2.51011968e-01
4.04067665e-01 2.42750809e-01 6.48339987e-02 1.18602610e+00
-2.33696103e-01 1.04274642e+00 1.91735532e-02 -4.28097665e-01
3.95496577e-01 -1.66589841e-01 3.90954196e-01 -4.14051116e-01
-8.82611647e-02 -1.73578128e-01 7.41455555e-01 -1.22522151e+00
-2.64935404e-01 -8.99032831e-01 -1.32510364e+00 -6.87465191e-01
7.31005520e-02 -1.02584720e-01 6.15545809e-01 1.25485873e+00
5.30796289e-01 3.26409221e-01 3.58869165e-01 -1.63334918e+00
-4.00566071e-01 -4.53383595e-01 -4.86672446e-02 -2.29414001e-01
6.90723002e-01 -8.99500072e-01 -4.03433740e-02 -2.68119693e-01] | [5.638774871826172, 0.8445616364479065] |
368b51bb-63cd-44b9-a388-b42cb481b7c9 | chemformer-a-pre-trained-transformer-for | null | null | https://iopscience.iop.org/article/10.1088/2632-2153/ac3ffb | https://iopscience.iop.org/article/10.1088/2632-2153/ac3ffb/pdf | Chemformer: a pre-trained transformer for computational chemistry | Transformer models coupled with a simplified molecular line entry system (SMILES) have recently proven to be a powerful combination for solving challenges in cheminformatics. These models, however, are often developed specifically for a single application and can be very resource-intensive to train. In this work we present the Chemformer model—a Transformer-based model which can be quickly applied to both sequence-to-sequence and discriminative cheminformatics tasks. Additionally, we show that self-supervised pre-training can improve performance and significantly speed up convergence on downstream tasks. On direct synthesis and retrosynthesis prediction benchmark datasets we publish state-of-the-art results for top-1 accuracy. We also improve on existing approaches for a molecular optimisation task and show that Chemformer can optimise on multiple discriminative tasks simultaneously. Models, datasets and code will be made available after publication. | ['Esben Jannik Bjerrum', 'Jiazhen He', 'Spyridon Dimitriadis', 'Ross Irwin'] | 2022-01-31 | null | null | null | machine-learning-science-and-technology-2022 | ['retrosynthesis'] | ['medical'] | [ 6.76618814e-01 -2.81309575e-01 -3.28656495e-01 -3.94241780e-01
-1.09636056e+00 -1.11762059e+00 6.93242371e-01 6.05043709e-01
-2.96805888e-01 1.19076991e+00 -2.19842270e-01 -6.73142433e-01
-1.60804853e-01 -1.43340498e-01 -9.46021795e-01 -1.01115453e+00
-1.03673518e-01 8.29896748e-01 1.52144179e-01 -4.72265661e-01
6.97505355e-01 5.50732911e-01 -1.13095081e+00 5.57654142e-01
9.98271346e-01 6.15416586e-01 6.86139882e-01 6.36985421e-01
4.59161624e-02 5.65834224e-01 -2.98154294e-01 -4.36050504e-01
1.40844449e-01 -4.32660431e-01 -7.62677491e-01 -8.11558425e-01
5.36308110e-01 3.92262071e-01 9.50395539e-02 5.96362174e-01
1.11447728e+00 2.42134094e-01 7.64764547e-01 -6.18176579e-01
7.23293051e-02 4.54347670e-01 -9.20275971e-02 1.57177001e-01
4.88656759e-01 4.40863848e-01 1.02857673e+00 -1.04830897e+00
8.58601391e-01 1.08141506e+00 7.76709974e-01 3.99488509e-01
-1.70542717e+00 -8.75521779e-01 1.14713326e-01 5.41231334e-01
-1.04202342e+00 -7.57535815e-01 2.61843294e-01 -5.13227165e-01
1.87980020e+00 2.94435531e-01 3.42958689e-01 1.13217485e+00
5.27614474e-01 5.30384123e-01 9.29683745e-01 -2.00586587e-01
3.42896342e-01 -1.26320243e-01 -3.36040288e-01 5.60934305e-01
-2.11351588e-01 3.56087774e-01 -7.81109154e-01 -2.33301491e-01
-7.57293426e-04 -1.93057090e-01 -9.50458869e-02 -6.27252996e-01
-1.19446087e+00 1.05334651e+00 5.96694171e-01 7.61405081e-02
-2.61564046e-01 2.66272258e-02 5.55197477e-01 3.43390405e-01
3.66173416e-01 1.31719565e+00 -9.97116268e-01 -2.48543397e-01
-1.06898952e+00 5.30786097e-01 9.32459176e-01 9.27706480e-01
5.17827630e-01 -2.20666409e-01 -9.75685045e-02 5.66146612e-01
1.57621339e-01 -1.42345652e-01 2.04473689e-01 -3.94838452e-01
4.46483403e-01 1.06288932e-01 -1.66294593e-02 -1.25844106e-01
-6.89102650e-01 -4.79962915e-01 -5.25579631e-01 1.50077268e-01
2.65799940e-01 1.32246260e-02 -9.75873411e-01 1.44338989e+00
5.38411140e-01 -1.86000779e-01 1.62371576e-01 5.49778461e-01
8.22787404e-01 9.62047994e-01 7.22255290e-01 -4.98137593e-01
9.73270655e-01 -1.19083285e+00 -4.48944867e-01 1.26966745e-01
1.16419601e+00 -1.26141286e+00 4.27013963e-01 8.82904589e-01
-1.10047829e+00 -3.02691281e-01 -1.16997170e+00 -2.24440262e-01
-1.00965178e+00 4.77532409e-02 9.46661055e-01 8.12493682e-01
-7.65606940e-01 1.28577673e+00 -5.86041152e-01 -9.80429575e-02
3.11820865e-01 9.60629165e-01 -4.30365384e-01 1.28052071e-01
-9.43212688e-01 1.42331791e+00 9.37372029e-01 -1.83316544e-01
-1.34364545e+00 -1.38554585e+00 -5.11374533e-01 -9.49521363e-02
5.62702954e-01 -8.07921290e-01 1.49634075e+00 -5.29785514e-01
-1.90287066e+00 4.30413067e-01 -1.84933186e-01 -5.10505617e-01
3.57694626e-01 -3.45927365e-02 -2.13315830e-01 -2.00915933e-01
-3.98151837e-02 7.46605814e-01 4.61192787e-01 -7.87511706e-01
-4.61392075e-01 -2.65114635e-01 -3.60438734e-01 2.68092245e-01
1.91916570e-01 2.93214232e-01 7.01046549e-04 -5.23106277e-01
-4.23927486e-01 -9.28954840e-01 -6.83606803e-01 -5.38377523e-01
-5.17433226e-01 -1.60905570e-01 5.23786724e-01 -4.58610803e-01
1.14662182e+00 -1.30867052e+00 6.71117723e-01 3.24284554e-01
-4.11630645e-02 5.95948040e-01 -3.16456825e-01 8.48678589e-01
-8.93255055e-01 7.99777135e-02 -3.85728627e-01 -4.85804565e-02
-8.67151842e-02 -1.60197243e-01 2.01431364e-02 3.77866417e-01
2.10504949e-01 7.86602080e-01 -1.08777833e+00 6.89886808e-02
3.60171705e-01 3.95654351e-01 -7.78254032e-01 1.54308587e-01
-7.71310568e-01 7.10388720e-01 -3.14919770e-01 6.45947337e-01
5.78079522e-01 -2.30043218e-01 5.53596258e-01 -3.32835585e-01
-5.57258666e-01 6.64065838e-01 -7.57476389e-01 2.03558803e+00
-4.44391102e-01 4.52358983e-02 -1.91973463e-01 -1.19265306e+00
6.61038458e-01 4.00347561e-01 5.21045685e-01 -7.44852781e-01
-1.54572755e-01 4.65033263e-01 3.88471574e-01 -1.88038886e-01
2.69060552e-01 -5.17454863e-01 3.57196689e-01 -1.48012266e-02
3.99464399e-01 -4.70064133e-01 6.63972378e-01 1.83035165e-01
8.69463801e-01 7.68800378e-01 5.57578683e-01 -4.43567783e-01
8.81442726e-01 3.54666770e-01 2.19666749e-01 4.67035145e-01
4.93816257e-01 2.71330029e-01 3.17262918e-01 -5.63004613e-01
-1.20362926e+00 -6.05998158e-01 -3.56199950e-01 1.55605590e+00
-4.53310996e-01 -1.22261894e+00 -4.58884269e-01 -6.65250301e-01
-1.02085307e-01 8.65623653e-01 -2.28245392e-01 -6.83963597e-02
-3.84897619e-01 -1.22095704e+00 2.26460099e-01 1.41113147e-01
-3.97815228e-01 -6.77395165e-01 2.65000109e-02 7.89626360e-01
4.38483834e-01 -7.04284191e-01 -3.75310361e-01 9.72599506e-01
-7.95286775e-01 -1.23606765e+00 -5.79740763e-01 -4.68743086e-01
2.22742081e-01 2.39381306e-02 1.14820147e+00 -3.30020517e-01
-4.94915724e-01 -2.85673350e-01 -4.76764813e-02 -7.52942324e-01
-6.66363358e-01 5.38162053e-01 -5.30567728e-02 -5.76483190e-01
3.57400887e-02 -5.93896866e-01 -7.35349357e-01 1.93644300e-01
-6.35483682e-01 1.79541200e-01 5.32183349e-01 8.79530430e-01
7.17019618e-01 -4.75153506e-01 5.54701626e-01 -1.09096301e+00
5.20016909e-01 -3.86862993e-01 -8.78930628e-01 3.68415207e-01
-1.01317441e+00 6.19675100e-01 9.16507483e-01 -6.33416176e-02
-8.41537118e-01 7.35932827e-01 -4.88065422e-01 -3.89829166e-02
1.57426335e-02 7.65831351e-01 -2.44897157e-01 -5.63050628e-01
6.89850807e-01 8.85280445e-02 -1.01156890e-01 -7.62865067e-01
5.17760336e-01 1.74163058e-01 1.12925738e-01 -8.54159892e-01
5.14043272e-01 -1.64175972e-01 6.71370268e-01 -6.77235842e-01
-5.81072211e-01 -7.55337536e-01 -6.21482432e-01 3.45156670e-01
7.14988112e-01 -1.02592242e+00 -1.05614424e+00 -1.18430369e-01
-1.15345073e+00 -4.00676250e-01 3.42564255e-01 3.52849334e-01
-6.95809186e-01 4.90021229e-01 -1.99893951e-01 -4.39720988e-01
-6.22498155e-01 -1.59714830e+00 1.22772324e+00 -1.47170603e-01
-3.45961004e-01 -1.14809859e+00 3.78615111e-01 2.76558936e-01
1.68563470e-01 2.47848943e-01 1.02642536e+00 -9.24127221e-01
-6.40653670e-01 1.50783032e-01 2.45353192e-01 -1.11232206e-01
-4.26324308e-02 5.43519575e-03 -8.83744419e-01 -3.49113256e-01
-6.44782245e-01 -4.87168312e-01 1.19332600e+00 2.42433563e-01
1.34055948e+00 -3.86714190e-01 -7.88828611e-01 8.24128270e-01
1.31716919e+00 4.32530195e-01 5.15276313e-01 4.64826643e-01
8.14715445e-01 4.81171966e-01 7.52610564e-01 3.83437663e-01
-1.49749547e-01 9.61804211e-01 3.01958054e-01 -1.94458023e-01
2.87802905e-01 -3.31780583e-01 4.02303666e-01 2.63660550e-01
-4.42548633e-01 -3.22204202e-01 -8.95646572e-01 -5.67507371e-02
-1.92466617e+00 -1.00654066e+00 -4.80549157e-01 2.23620749e+00
1.23690879e+00 -2.21118823e-01 4.65824693e-01 4.40328242e-03
1.92284361e-01 -1.81888387e-01 -6.22185826e-01 -6.94229960e-01
-7.68394326e-04 1.02323306e+00 7.69902527e-01 6.31663263e-01
-1.12305057e+00 1.33944738e+00 7.17621756e+00 1.34319723e+00
-1.10676396e+00 4.60619144e-02 6.35494709e-01 -2.97642767e-01
-2.35953227e-01 3.72167170e-01 -1.02811575e+00 4.87895012e-01
1.33156502e+00 -1.29798830e-01 3.67935538e-01 8.69587719e-01
4.18161243e-01 1.54233366e-01 -1.72602499e+00 1.02824450e+00
-4.36683923e-01 -2.14196801e+00 1.26842439e-01 1.25275794e-02
7.40393281e-01 1.47722960e-01 7.87723139e-02 1.69725358e-01
2.68642247e-01 -1.56067002e+00 4.93498236e-01 1.14601329e-01
7.54375398e-01 -9.74810362e-01 2.75667876e-01 1.76161617e-01
-1.01534998e+00 2.23519251e-01 -2.92594045e-01 2.27617308e-01
-1.00670205e-02 3.42617214e-01 -1.15359640e+00 9.98380840e-01
3.07840377e-01 1.08162093e+00 -3.86900634e-01 1.11543608e+00
-7.14409351e-02 2.59140790e-01 -3.07464927e-01 -2.81926185e-01
5.49421012e-01 -4.49466139e-01 2.87325799e-01 1.62456274e+00
2.28997767e-01 5.02233952e-02 2.44279131e-01 7.20836282e-01
1.24026410e-01 2.31914610e-01 -4.62981552e-01 -2.90923357e-01
2.81093922e-02 1.28746116e+00 -3.98492187e-01 -3.23581606e-01
2.50528213e-02 7.70964682e-01 1.51880741e-01 2.13006958e-01
-7.28341103e-01 -3.19227070e-01 8.30895364e-01 -1.44219086e-01
5.88533580e-01 -1.87387645e-01 -1.51141416e-02 -8.53802800e-01
-6.85943604e-01 -1.27020001e+00 4.58241522e-01 -5.29453993e-01
-1.02501702e+00 3.15806009e-02 -1.57973245e-01 -7.64858365e-01
-1.05884373e-01 -1.26709831e+00 -4.85458285e-01 9.87542331e-01
-1.66050279e+00 -7.83218443e-01 4.15547788e-01 2.90073276e-01
4.84538287e-01 -2.53805697e-01 1.11602414e+00 2.30641857e-01
-6.02965236e-01 3.80895227e-01 6.94513738e-01 -1.05756164e+00
1.05188870e+00 -1.42581725e+00 4.41634774e-01 3.43403220e-01
-1.49434894e-01 1.02299190e+00 9.70319450e-01 -5.67435622e-01
-1.84325504e+00 -9.97395992e-01 7.73857176e-01 -6.52199328e-01
7.29805350e-01 -5.41643977e-01 -6.24266326e-01 2.18363613e-01
2.22941279e-01 -6.77244663e-01 1.16023552e+00 2.95402765e-01
-3.07610482e-01 3.72262448e-02 -9.12102103e-01 3.58635128e-01
1.22520518e+00 -2.47403458e-01 -1.26015678e-01 1.19493461e+00
4.33944255e-01 -7.15420187e-01 -1.28543913e+00 2.99688548e-01
3.71992648e-01 -8.41148615e-01 1.46545756e+00 -1.00075996e+00
4.63397592e-01 -3.91301811e-01 2.37009332e-01 -1.39507389e+00
-3.55305344e-01 -1.46953571e+00 -2.67665796e-02 6.07875228e-01
9.98910487e-01 -5.94262958e-01 5.52626312e-01 2.00607255e-01
-5.09002864e-01 -5.89726448e-01 -9.60012913e-01 -8.68738472e-01
2.78187245e-01 -3.68402563e-02 4.06090736e-01 7.21683204e-01
4.10631955e-01 1.09362435e+00 -3.01335782e-01 -1.82136670e-01
3.21045071e-01 4.36299369e-02 6.38560474e-01 -1.03409481e+00
-5.02815008e-01 -6.52758598e-01 -4.56012562e-02 -9.05238330e-01
3.53621878e-02 -1.34084105e+00 -6.73088655e-02 -1.21419764e+00
2.06146106e-01 -2.80424565e-01 -3.49487454e-01 5.90411067e-01
-2.70339996e-01 -2.25470319e-01 -1.63490903e-02 -5.78015409e-02
-5.99987030e-01 6.80587649e-01 1.11395478e+00 -4.00552154e-01
-2.17317209e-01 -1.79238379e-01 -6.83187366e-01 -1.42145948e-02
7.45297730e-01 -6.00538731e-01 -3.81908208e-01 2.38222137e-01
5.02717435e-01 -1.84470952e-01 -1.25962123e-01 -6.78137481e-01
1.38845235e-01 -1.99325562e-01 5.33364773e-01 -7.06005216e-01
4.20987964e-01 -5.45084774e-01 4.03941721e-01 6.44487560e-01
-3.33279759e-01 5.14664687e-02 6.12462699e-01 5.46406746e-01
1.17683090e-01 -2.48220250e-01 7.09824502e-01 -3.64946574e-01
-3.68059367e-01 5.79683661e-01 -3.94431859e-01 -4.66024339e-01
1.00190890e+00 -1.57167643e-01 -2.56778061e-01 1.00131802e-01
-5.25091827e-01 3.20798725e-01 3.98207247e-01 4.05697487e-02
2.85299450e-01 -6.93751991e-01 -4.69962984e-01 -2.39401057e-01
2.02082977e-01 -1.78472504e-01 -1.47234768e-01 1.04479229e+00
-5.96689582e-01 1.17076182e+00 6.12041280e-02 -4.71018791e-01
-1.69457412e+00 9.30027246e-01 5.14083803e-01 -4.04565930e-01
1.50334254e-01 8.43701422e-01 2.97796279e-01 -5.22463143e-01
3.39175016e-02 -2.12293565e-01 -9.40028299e-03 1.63886055e-01
4.22566563e-01 1.68196902e-01 6.11023009e-01 -3.28755856e-01
-6.54322982e-01 5.44921935e-01 -5.37621975e-01 3.21216792e-01
1.74487233e+00 7.22670734e-01 -3.14647779e-02 -2.69916624e-01
1.17567563e+00 -4.86375004e-01 -1.08604848e+00 1.75352946e-01
2.25567892e-01 -2.04566605e-02 1.23284481e-01 -1.23108590e+00
-2.89589882e-01 1.03011274e+00 5.19828856e-01 -4.70819771e-01
7.25735426e-01 -1.66859508e-01 4.45253581e-01 9.03950989e-01
-1.20123461e-01 -1.20038843e+00 -4.39107344e-02 5.83383083e-01
7.97296345e-01 -1.14356041e+00 5.52759290e-01 -4.22738969e-01
-3.84500355e-01 1.46079636e+00 1.61612898e-01 5.54189801e-01
1.63369402e-01 3.96920145e-02 -5.23431718e-01 -2.56738544e-01
-9.75580931e-01 -2.14297146e-01 3.68653387e-01 7.30191708e-01
1.11146963e+00 -9.81863365e-02 -3.77189815e-01 1.46402255e-01
-1.86371416e-01 -1.87383711e-01 -6.42375974e-03 1.07651067e+00
-3.71130228e-01 -2.06406856e+00 -1.96365416e-01 -1.13134431e-02
-5.42296410e-01 -7.38449812e-01 -1.00268197e+00 3.85841459e-01
1.21367443e-02 9.11986947e-01 -6.93528354e-01 -1.45882756e-01
2.52809554e-01 2.62572467e-01 1.14794779e+00 -6.36459351e-01
-9.81506526e-01 1.15714841e-01 6.31349564e-01 -6.05643988e-01
-4.43073928e-01 -4.11477745e-01 -1.05118382e+00 -4.64130521e-01
-3.23954105e-01 5.15060008e-01 9.13684070e-01 9.08469021e-01
6.89829171e-01 4.86583978e-01 4.83056515e-01 -1.16332150e+00
-6.13991201e-01 -7.12473273e-01 1.07679576e-01 -2.31533751e-01
1.47858232e-01 -5.23880363e-01 4.16110642e-02 1.01667993e-01] | [4.694423198699951, 5.921141624450684] |
c6c10c0a-b070-45d4-b09c-63149ec51ca8 | sparse-depth-completion-with-semantic-mesh | 2112.05498 | null | https://arxiv.org/abs/2112.05498v1 | https://arxiv.org/pdf/2112.05498v1.pdf | Sparse Depth Completion with Semantic Mesh Deformation Optimization | Sparse depth measurements are widely available in many applications such as augmented reality, visual inertial odometry and robots equipped with low cost depth sensors. Although such sparse depth samples work well for certain applications like motion tracking, a complete depth map is usually preferred for broader applications, such as 3D object recognition, 3D reconstruction and autonomous driving. Despite the recent advancements in depth prediction from single RGB images with deeper neural networks, the existing approaches do not yield reliable results for practical use. In this work, we propose a neural network with post-optimization, which takes an RGB image and sparse depth samples as input and predicts the complete depth map. We make three major contributions to advance the state-of-the-art: an improved backbone network architecture named EDNet, a semantic edge-weighted loss function and a semantic mesh deformation optimization method. Our evaluation results outperform the existing work consistently on both indoor and outdoor datasets, and it significantly reduces the mean average error by up to 19.5% under the same settings of 200 sparse samples on NYU-Depth-V2 dataset. | ['Sinem Guven', 'Matias Aiskovich', 'Bing Zhou'] | 2021-12-10 | null | null | null | null | ['3d-object-recognition', 'depth-completion'] | ['computer-vision', 'computer-vision'] | [ 2.41611466e-01 4.44397964e-02 -4.83699054e-01 -6.03383124e-01
-5.40979564e-01 1.17849052e-01 2.51751333e-01 -1.17253512e-01
-4.58948404e-01 5.59441507e-01 1.28153875e-01 -2.07315966e-01
8.92401487e-02 -1.05271316e+00 -1.04747510e+00 -2.62943149e-01
-1.32700419e-02 5.82302034e-01 2.37428948e-01 -2.10918307e-01
3.09523106e-01 4.51961011e-01 -1.93438160e+00 -1.61632225e-01
8.10564339e-01 1.63470912e+00 2.48039037e-01 4.24120635e-01
-1.47275433e-01 7.58124411e-01 1.01186953e-01 1.56672612e-01
6.40781343e-01 -1.81057621e-02 -4.82267171e-01 1.61939964e-01
6.88625634e-01 -7.18577027e-01 -8.04699183e-01 9.88605678e-01
7.25870848e-01 2.32643455e-01 1.67617202e-01 -1.14820182e+00
-4.16475147e-01 1.47252262e-01 -6.12539172e-01 -2.01526389e-01
4.15141523e-01 7.13132769e-02 7.02802598e-01 -9.30361390e-01
7.09187567e-01 1.29590094e+00 9.96795654e-01 6.93620265e-01
-7.67978251e-01 -6.30448639e-01 1.87388435e-01 4.01503533e-01
-1.20317817e+00 -3.96620512e-01 9.80057597e-01 -1.21595144e-01
1.23765373e+00 -2.74921954e-01 7.81432211e-01 9.98927891e-01
1.93888411e-01 8.16019773e-01 1.00060844e+00 6.42890036e-02
3.22870344e-01 -2.66990900e-01 -2.58112252e-01 9.13209975e-01
4.02375758e-01 2.28046224e-01 -8.32468510e-01 3.44739646e-01
8.68396878e-01 2.24298224e-01 -2.53309608e-01 -8.00469160e-01
-1.18809497e+00 7.38333285e-01 8.44119847e-01 -2.79556900e-01
-4.78014410e-01 4.74326044e-01 1.93488374e-01 2.76702374e-01
7.09395707e-01 -5.42710125e-02 -6.28792524e-01 -4.91871417e-01
-5.82826734e-01 4.89547759e-01 6.06002569e-01 1.07356799e+00
1.16507459e+00 8.85872841e-02 4.96196687e-01 6.19370282e-01
5.19090652e-01 7.52706528e-01 4.11534399e-01 -1.26279724e+00
7.26051629e-01 8.33559036e-01 1.62587851e-01 -9.76678789e-01
-7.45780349e-01 -1.07779719e-01 -8.45347404e-01 4.04108763e-01
1.37240618e-01 -2.43516304e-02 -1.01607251e+00 1.42895520e+00
4.91712898e-01 5.91768026e-01 -2.29398869e-02 1.17722058e+00
1.07234478e+00 2.63132364e-01 -5.65781653e-01 3.67725074e-01
8.46367776e-01 -1.08853805e+00 -5.78130126e-01 -7.87784755e-01
6.10874891e-01 -2.73120731e-01 1.05666065e+00 4.35781300e-01
-9.05821085e-01 -4.45033938e-01 -1.45829749e+00 -6.81997120e-01
-1.96891889e-01 -1.91578209e-01 9.39431727e-01 4.51161742e-01
-1.09709394e+00 7.59386480e-01 -1.25683570e+00 -3.89145553e-01
5.99728703e-01 4.17870969e-01 -4.16899651e-01 -5.97724020e-01
-8.59066665e-01 8.03249121e-01 7.69468546e-02 2.52356261e-01
-8.29120755e-01 -7.08426535e-01 -1.40391910e+00 -4.00166601e-01
2.81980425e-01 -1.01768470e+00 1.16327858e+00 -2.37559646e-01
-1.78752422e+00 9.00019765e-01 -1.92981020e-01 -5.88987529e-01
6.03243947e-01 -5.52885711e-01 -1.33990059e-02 -1.36630639e-01
1.09292403e-01 1.02284741e+00 4.09330070e-01 -1.11857855e+00
-7.38403678e-01 -8.18634868e-01 1.58095017e-01 3.61743629e-01
-3.15900929e-02 -8.04743052e-01 -6.07844353e-01 -1.52503073e-01
8.94859374e-01 -1.05994546e+00 -4.66221750e-01 4.11606342e-01
-2.81568438e-01 2.07742125e-01 9.14766431e-01 -3.36430490e-01
5.54780960e-01 -1.85341299e+00 3.25737536e-01 -9.57157612e-02
2.59106040e-01 -3.16893548e-01 1.62072465e-01 -1.06420018e-01
4.57358927e-01 -3.87939960e-01 -4.29388642e-01 -1.11138260e+00
1.35660455e-01 5.81673682e-01 3.73910666e-02 7.70602465e-01
-5.49277030e-02 9.17738855e-01 -9.20622051e-01 -4.57925498e-02
5.98706126e-01 8.03989232e-01 -7.95773387e-01 8.69744718e-02
-2.20212802e-01 5.15422642e-01 -3.04685682e-01 9.03618753e-01
7.04205275e-01 -2.90000319e-01 -3.44639540e-01 -1.46448284e-01
2.13527642e-02 6.48851156e-01 -1.29057431e+00 2.68420911e+00
-5.95427811e-01 7.74809778e-01 1.85233474e-01 -8.28862607e-01
1.00976646e+00 -1.17105477e-01 6.81934714e-01 -9.94032502e-01
2.69964993e-01 4.33650881e-01 -5.35037398e-01 -3.90078574e-01
9.36333776e-01 2.03353073e-02 -3.43370847e-02 3.69792357e-02
-2.11067572e-01 -5.85170507e-01 -4.31080818e-01 -1.61529064e-01
1.14537537e+00 5.31959772e-01 -3.22063342e-02 1.63315371e-01
1.92110866e-01 1.19148940e-01 7.59203851e-01 3.74819636e-01
-2.37342730e-01 8.14982355e-01 4.01290320e-02 -6.44904613e-01
-1.01054990e+00 -7.74073839e-01 -1.88590020e-01 4.08706874e-01
8.48300934e-01 -1.41354769e-01 -3.53250563e-01 -3.18753690e-01
4.25399780e-01 2.47000024e-01 -5.35889149e-01 -1.08971342e-01
-4.64175165e-01 -4.31473613e-01 2.98267543e-01 8.43979716e-01
9.68072474e-01 -7.28488386e-01 -8.82242322e-01 2.18282104e-01
-5.82868122e-02 -1.54200995e+00 8.27863291e-02 2.68077999e-01
-1.34834874e+00 -9.16688144e-01 -4.78485852e-01 -6.23359144e-01
4.97169644e-01 4.70083028e-01 1.11495268e+00 -1.48883522e-01
-1.33859724e-01 1.81597590e-01 -1.60562053e-01 -3.60450119e-01
2.07993045e-01 1.32965893e-01 2.28157297e-01 -3.99262905e-01
4.28749025e-01 -7.38267303e-01 -8.23443651e-01 1.98147684e-01
-5.20204067e-01 1.59331471e-01 3.75125378e-01 5.90048909e-01
9.71157253e-01 -1.85736597e-01 9.18492600e-02 -6.98508263e-01
-8.97598639e-03 -3.82861227e-01 -6.87220275e-01 -5.56193471e-01
-8.20923328e-01 -2.64879391e-02 1.67213440e-01 -6.53857784e-03
-6.17619395e-01 3.39052856e-01 -5.56570828e-01 -7.19225705e-01
-5.43397814e-02 4.51358467e-01 -1.05905227e-01 -5.39107263e-01
6.06697619e-01 -9.34644938e-02 1.01508148e-01 -4.67730761e-01
1.55278936e-01 4.55247402e-01 5.74273884e-01 -2.46227235e-01
6.08605444e-01 9.01191592e-01 4.07273352e-01 -7.10087299e-01
-8.33037317e-01 -5.57048321e-01 -4.10502851e-01 4.58807545e-03
7.05502629e-01 -1.41563559e+00 -7.96901286e-01 7.39601374e-01
-1.03989279e+00 -6.61077261e-01 -2.02364370e-01 6.46270633e-01
-7.73421764e-01 4.36632782e-01 -5.51062524e-01 -5.68698704e-01
-1.52862862e-01 -1.36857271e+00 1.49477887e+00 1.69286698e-01
2.60223389e-01 -8.56643736e-01 -6.08939603e-02 4.22120452e-01
3.27449679e-01 6.02687657e-01 3.41040820e-01 2.95053571e-01
-9.70500588e-01 -2.23640755e-01 -2.53347099e-01 1.88793167e-01
1.51736021e-01 -5.25868595e-01 -1.08121753e+00 -1.40849277e-01
-1.50404736e-01 -6.08811915e-01 1.03409290e+00 5.02490044e-01
1.20152807e+00 2.20271900e-01 -2.48800784e-01 1.40222812e+00
1.49712229e+00 -1.16854042e-01 7.74816513e-01 6.34001851e-01
1.16893721e+00 3.98290604e-01 6.07400656e-01 4.75349873e-01
1.04748666e+00 6.37065947e-01 1.03612232e+00 5.29854670e-02
-2.26160854e-01 -3.64117175e-01 1.25434488e-01 7.55146146e-01
3.60694267e-02 1.86192486e-02 -9.26003039e-01 5.69257200e-01
-1.90249562e+00 -4.42325711e-01 1.51689153e-03 2.05434608e+00
4.35147345e-01 3.79552871e-01 -2.69307733e-01 3.02457690e-01
2.17722297e-01 3.15195590e-01 -1.14067101e+00 3.67779136e-02
-2.30695412e-01 1.81688279e-01 9.97819483e-01 6.16074383e-01
-9.80373919e-01 8.56503248e-01 5.48908043e+00 2.03278199e-01
-1.35966206e+00 1.11494511e-01 3.74086797e-01 -3.29067916e-01
-4.03722197e-01 -2.46476457e-01 -1.00400698e+00 2.81431973e-01
6.50242567e-01 2.16433123e-01 4.36751366e-01 1.21746016e+00
1.40914395e-01 -3.64540249e-01 -1.02063358e+00 1.58401358e+00
7.41912499e-02 -1.43872130e+00 -2.99034297e-01 2.41892174e-01
1.03056657e+00 8.35813880e-01 -2.11546998e-02 1.05651200e-01
1.08514383e-01 -9.33996379e-01 7.70159841e-01 2.65340835e-01
8.93288791e-01 -7.10913002e-01 7.74221957e-01 4.11436856e-01
-1.22114301e+00 4.30176817e-02 -7.00376570e-01 -5.88439643e-01
3.92255872e-01 8.90942812e-01 -5.00538588e-01 4.94286239e-01
9.17946517e-01 1.33037794e+00 -1.37657374e-01 9.84650433e-01
-3.17565918e-01 -1.01431571e-02 -6.10405982e-01 8.14402327e-02
2.62229532e-01 -9.47958604e-02 3.67061973e-01 4.22459960e-01
5.45094073e-01 -2.32202038e-02 2.07140207e-01 6.14092827e-01
-1.80468574e-01 -2.91364610e-01 -8.82671893e-01 2.72659063e-01
4.58458185e-01 8.83710742e-01 -5.79707503e-01 -1.16667978e-01
-5.10533929e-01 1.01152813e+00 2.64933407e-01 1.06325440e-01
-6.37134790e-01 -4.02688712e-01 1.15664971e+00 2.21348956e-01
3.00452232e-01 -7.22906590e-01 -8.79658580e-01 -1.21589208e+00
2.18559757e-01 -4.08147991e-01 -2.63673067e-01 -7.85467625e-01
-7.83193707e-01 3.98962677e-01 -4.09374326e-01 -1.33160853e+00
-2.90666103e-01 -8.92830431e-01 -4.29211184e-02 6.94632947e-01
-2.06892633e+00 -8.89297724e-01 -1.16839659e+00 4.72897828e-01
5.41719496e-01 1.67734742e-01 6.20759726e-01 6.81043625e-01
-4.20174778e-01 2.29287669e-01 -8.99605080e-02 -1.28077060e-01
4.82959092e-01 -1.12722790e+00 8.90083492e-01 6.56153262e-01
-1.36326194e-01 1.97881907e-01 5.61133385e-01 -6.43871665e-01
-2.06428051e+00 -1.09059095e+00 6.53526962e-01 -3.74927342e-01
3.03950638e-01 -3.23376149e-01 -4.78593081e-01 6.09223485e-01
-4.17213112e-01 6.55027032e-01 -5.18849380e-02 -8.25233534e-02
6.10543834e-03 -2.27169707e-01 -1.24840856e+00 3.17762285e-01
1.73943615e+00 -5.17206550e-01 -1.21700279e-01 2.15048015e-01
9.77315187e-01 -1.31301081e+00 -8.20545673e-01 8.27543199e-01
6.93601966e-01 -1.28434372e+00 1.18447471e+00 -4.31123823e-02
6.38470352e-01 -3.41767311e-01 -6.57176137e-01 -1.14980578e+00
-1.75979722e-03 -3.23272914e-01 -5.43016434e-01 4.27076638e-01
1.34900391e-01 -6.38871789e-01 1.53252494e+00 5.17835677e-01
-6.28736913e-01 -1.04288018e+00 -1.12590933e+00 -5.64061821e-01
-3.88088584e-01 -9.46632028e-01 7.24027991e-01 6.75152063e-01
-4.93438751e-01 5.01159206e-02 -3.89801651e-01 2.05225557e-01
8.02293837e-01 -3.47195342e-02 1.17909372e+00 -1.17643726e+00
-4.16648015e-02 -2.60527343e-01 -1.02865422e+00 -1.90201354e+00
2.26453006e-01 -4.44447815e-01 3.61782014e-01 -1.93703830e+00
-5.73730648e-01 -7.09664643e-01 -1.54645257e-02 2.71149218e-01
9.38251242e-02 5.49554944e-01 -3.05387855e-01 -8.10557157e-02
-4.79316056e-01 9.76954579e-01 1.24844456e+00 -2.44278729e-01
-2.91262895e-01 -9.63416621e-02 -4.00064826e-01 8.26029480e-01
4.66841608e-01 -2.65847981e-01 -5.52670181e-01 -9.73808110e-01
5.29319167e-01 1.03260390e-01 3.50495577e-01 -1.44236267e+00
2.44407445e-01 -6.57146797e-02 3.24932367e-01 -9.23286974e-01
7.14535356e-01 -8.79891455e-01 4.63872366e-02 4.94850188e-01
2.73045629e-01 1.26787797e-01 1.56875089e-01 6.52980208e-01
-2.16529340e-01 2.47900650e-01 6.44194365e-01 -1.68206915e-01
-1.39491093e+00 8.24979246e-01 4.32189018e-01 -7.20640197e-02
8.19938838e-01 -7.67719328e-01 -2.19733983e-01 -4.43485498e-01
-3.27672571e-01 3.21001530e-01 1.01434338e+00 5.60552835e-01
9.44699943e-01 -1.50907290e+00 -3.88817668e-01 4.28618252e-01
2.83152342e-01 1.02258325e+00 3.09825033e-01 8.15313637e-01
-8.78507674e-01 3.04547280e-01 -3.21712941e-01 -1.00807941e+00
-5.91995180e-01 9.67542902e-02 3.42752278e-01 2.31057838e-01
-1.05392349e+00 1.13990915e+00 -8.19244012e-02 -6.29039764e-01
5.47142029e-01 -9.14094448e-01 1.16749115e-01 -4.39265847e-01
3.27163547e-01 5.56988955e-01 4.38277066e-01 -6.26499891e-01
-4.37954098e-01 9.39173818e-01 4.41906214e-01 5.68832941e-02
1.49990749e+00 -4.16063368e-01 1.88493878e-01 6.58355176e-01
1.37894702e+00 -3.27919781e-01 -1.83936572e+00 -3.53726566e-01
-3.31936538e-01 -6.56208098e-01 4.76130605e-01 -2.11220309e-01
-1.26847768e+00 8.66569400e-01 6.48309946e-01 -2.67533392e-01
9.68714118e-01 -1.82209134e-01 1.33748591e+00 7.61857390e-01
9.64042068e-01 -9.08372998e-01 2.01002043e-02 8.48411322e-01
5.17190635e-01 -1.66714215e+00 1.66453734e-01 -3.22461426e-01
-3.06708068e-01 7.18414307e-01 7.05025196e-01 -3.61978441e-01
7.10296154e-01 2.41054162e-01 7.30336159e-02 -2.59232134e-01
-4.08105582e-01 -2.28727743e-01 1.43125877e-01 6.84683025e-01
1.29493192e-01 -1.33908749e-01 1.85540646e-01 1.24645166e-01
-3.63362551e-01 1.81961447e-01 3.79274189e-01 1.07370758e+00
-5.80047429e-01 -8.27609181e-01 -1.08209081e-01 4.69549894e-01
-8.80773365e-02 1.77262485e-01 -2.89479904e-02 6.10464752e-01
1.28168672e-01 6.36241496e-01 2.38348067e-01 -7.67803013e-01
5.80598176e-01 -2.58013457e-01 7.22760022e-01 -5.84161460e-01
2.03169342e-02 -4.64848638e-01 8.28803927e-02 -1.17505026e+00
-6.48566961e-01 -6.10588133e-01 -1.45402431e+00 -6.05815113e-01
6.58185259e-02 -5.88479519e-01 1.21697688e+00 9.48428333e-01
5.21690905e-01 4.46731210e-01 5.03216565e-01 -1.48715532e+00
-2.15392187e-01 -7.82489538e-01 -5.97753942e-01 2.24008158e-01
6.29210174e-01 -1.13124573e+00 -2.61692792e-01 -2.78982192e-01] | [8.520731925964355, -2.520075559616089] |
e38244e2-0b76-46f6-a746-bda43c0c5853 | deduction-under-perturbed-evidence-probing | 2305.14507 | null | https://arxiv.org/abs/2305.14507v1 | https://arxiv.org/pdf/2305.14507v1.pdf | Deduction under Perturbed Evidence: Probing Student Simulation Capabilities of Large Language Models | We explore whether Large Language Models (LLMs) are capable of logical reasoning with distorted facts, which we call Deduction under Perturbed Evidence (DUPE). DUPE presents a unique challenge to LLMs since they typically rely on their parameters, which encode mostly accurate information, to reason and make inferences. However, in DUPE, LLMs must reason over manipulated or falsified evidence present in their prompts, which can result in false conclusions that are valid only under the manipulated evidence. Our goal with DUPE is to determine whether LLMs can arrive at these false conclusions and identify whether the dominant factor influencing the deduction process is the encoded data in the parameters or the manipulated evidence in the prompts. To evaluate the DUPE capabilities of LLMs, we create a DUPEd version of the StrategyQA dataset, where facts are manipulated to reverse the answer to the question. Our findings show that even the most advanced GPT models struggle to reason on manipulated facts - showcasing poor DUPE skills - with accuracy dropping by 45% compared to the original dataset. We also investigate prompt settings inspired from student simulation models, which mitigate the accuracy drop to some extent. Our findings have practical implications for understanding the performance of LLMs in real-world applications such as student simulation models that involve reasoning over inaccurate information. | ['Richard G. Baraniuk', 'Shashank Sonkar'] | 2023-05-23 | null | null | null | null | ['logical-reasoning', 'strategyqa'] | ['reasoning', 'reasoning'] | [-1.74343754e-02 6.16609871e-01 4.04378958e-02 -9.88329053e-02
-9.18081343e-01 -9.59543049e-01 4.14791882e-01 4.61202592e-01
-2.43375093e-01 9.23086286e-01 9.64360684e-02 -1.04087698e+00
-3.87080818e-01 -1.03979659e+00 -1.13183105e+00 -4.12306227e-02
3.08112621e-01 5.46344995e-01 4.33544189e-01 -4.67765093e-01
6.27165020e-01 2.64182955e-01 -1.70955300e+00 6.45873010e-01
1.33518541e+00 4.14653659e-01 -1.02688335e-01 9.23011959e-01
-4.88657147e-01 1.61575150e+00 -1.12378049e+00 -8.20836604e-01
5.19878790e-02 -1.02746166e-01 -9.54441369e-01 -7.89405882e-01
6.95023894e-01 -4.93849844e-01 -1.87177993e-02 1.18818402e+00
4.82425272e-01 -7.75519088e-02 3.54676187e-01 -1.43737721e+00
-5.43119788e-01 1.21483541e+00 -3.80298123e-02 6.12618387e-01
1.17089009e+00 5.40043712e-01 8.68463159e-01 -5.74258626e-01
5.31714439e-01 1.58416176e+00 8.49511147e-01 3.25768620e-01
-1.19381034e+00 -5.19550920e-01 3.36934298e-01 5.25441170e-01
-1.14469576e+00 -5.54489553e-01 2.93510526e-01 -2.00228140e-01
1.18574524e+00 6.34388328e-01 5.21265149e-01 1.09363604e+00
3.82558614e-01 6.90667987e-01 1.49722886e+00 -6.05921686e-01
3.76193941e-01 3.17923248e-01 4.01943982e-01 8.13585818e-01
6.58487320e-01 1.75693363e-01 -1.05049729e+00 -2.86160231e-01
4.95360911e-01 -8.61873507e-01 -3.41869026e-01 9.90093350e-02
-9.99219775e-01 6.90838456e-01 -1.15549639e-01 8.95018354e-02
-4.80527312e-01 -2.49605924e-02 7.11754560e-02 8.18823099e-01
-1.56652853e-01 1.16259027e+00 -6.71616077e-01 -7.63458669e-01
-6.29336715e-01 8.85664582e-01 1.33201385e+00 8.41412246e-01
3.83554436e-02 -6.30107522e-02 -5.05000770e-01 3.08986753e-01
1.28360048e-01 5.99009633e-01 4.18292820e-01 -1.31983840e+00
7.85658479e-01 6.14827871e-01 6.68005943e-01 -1.03979158e+00
-3.53363603e-01 -3.20113719e-01 3.00284505e-01 8.31056014e-02
1.12339354e+00 -1.98048145e-01 -3.32785487e-01 2.05183935e+00
1.88114896e-01 1.34540573e-01 3.61936003e-01 6.88425183e-01
7.93572307e-01 5.04569471e-01 1.92704961e-01 8.78527835e-02
1.27223516e+00 -3.41617167e-01 -8.14374447e-01 -4.37732220e-01
1.11474788e+00 -5.00691175e-01 1.70971847e+00 6.99037731e-01
-1.75917542e+00 -2.16258839e-02 -9.59402740e-01 -1.99415088e-01
-2.94551879e-01 -5.62820852e-01 5.98923385e-01 6.94090068e-01
-1.06044817e+00 5.44014096e-01 -4.74629402e-01 1.08764730e-02
1.54619217e-01 -4.69405018e-02 1.04765549e-01 -4.59939033e-01
-1.77035356e+00 1.36046171e+00 3.17209482e-01 6.20137493e-04
-6.79199040e-01 -1.27876484e+00 -9.74598110e-01 5.86923361e-01
8.01012814e-01 -5.50668657e-01 1.82559705e+00 -4.58906710e-01
-1.35113513e+00 4.34014499e-01 -1.67022765e-01 -5.56782722e-01
8.94547641e-01 -1.30270809e-01 -2.43146703e-01 1.26557752e-01
1.85958281e-01 2.21176341e-01 3.19342166e-01 -1.12783420e+00
-4.51505870e-01 -4.46519181e-02 5.04851460e-01 1.27871349e-01
3.79991472e-01 -4.53418307e-02 3.16265732e-01 -2.41471857e-01
4.88830358e-02 -4.39760417e-01 2.49087617e-01 -4.12767589e-01
-2.74265856e-01 -2.56379664e-01 3.74191701e-01 -7.09053218e-01
1.31512976e+00 -1.63766038e+00 -2.82702804e-01 2.96138048e-01
2.17800334e-01 4.17267114e-01 -2.85080969e-01 4.72679883e-01
8.79268274e-02 4.39211726e-01 2.00972542e-01 2.69388288e-01
4.09040481e-01 3.17503870e-01 -8.50498199e-01 -1.92537129e-01
2.78820008e-01 1.11902428e+00 -1.07123184e+00 -6.45015895e-01
-8.53084773e-02 -1.03265762e-01 -6.82955742e-01 1.26326039e-01
-6.76311016e-01 -2.95013133e-02 -3.36022347e-01 4.64808613e-01
6.70079529e-01 -2.01206535e-01 1.45830736e-01 2.70441741e-01
-1.05172902e-01 9.36795056e-01 -1.29253590e+00 1.14328384e+00
-3.56757730e-01 3.85129064e-01 -1.69702191e-02 -6.69419527e-01
4.53064024e-01 9.02999043e-02 -4.18654978e-01 -1.04380441e+00
-5.51453717e-02 2.69419968e-01 3.88069481e-01 -8.39667022e-01
5.79988420e-01 -3.03459376e-01 -6.86235353e-02 5.40294468e-01
-2.80629903e-01 -6.69119298e-01 5.06879032e-01 6.51088417e-01
1.30999017e+00 -8.32317621e-02 9.89381596e-02 -2.78374404e-01
4.78539705e-01 3.65151733e-01 6.36886358e-01 1.52908123e+00
-2.34818056e-01 9.37752128e-02 1.22785258e+00 -2.10581303e-01
-6.16633594e-01 -1.04769611e+00 -3.09175812e-02 9.76834714e-01
-4.52090912e-02 -4.05462116e-01 -7.76874006e-01 -7.19421446e-01
2.32129976e-01 1.78084767e+00 -2.39696816e-01 -5.23844898e-01
-3.32535535e-01 -2.58273989e-01 1.08898246e+00 5.35623789e-01
5.25409460e-01 -9.49677289e-01 -9.17457461e-01 2.25289047e-01
-5.14578760e-01 -1.07639349e+00 9.09511298e-02 -4.46286192e-03
-7.00158834e-01 -1.29108500e+00 1.54620886e-01 -6.43104687e-02
4.02764022e-01 -3.34895961e-02 1.44817007e+00 5.76518297e-01
3.43913108e-01 6.21343315e-01 -1.75487816e-01 -8.15674543e-01
-7.81740487e-01 -2.06061512e-01 -2.08795413e-01 -9.33904469e-01
5.27163804e-01 -2.80096948e-01 1.81552805e-02 2.66858369e-01
-9.25213277e-01 7.00578764e-02 3.73275280e-01 6.15942717e-01
-1.62961986e-02 2.43300065e-01 6.31743729e-01 -1.13275695e+00
1.19149041e+00 -5.68322182e-01 -7.03755617e-01 6.05156720e-01
-5.99591196e-01 4.46276039e-01 7.81215727e-01 -4.95043188e-01
-1.26422286e+00 -1.16444457e+00 5.70851453e-02 -1.05354182e-01
2.37514451e-02 9.83444989e-01 -5.84967919e-02 6.71541493e-04
8.97147775e-01 1.58241257e-01 -9.65756550e-02 4.09749858e-02
1.35813564e-01 1.70375958e-01 5.95592022e-01 -1.58158386e+00
5.79380691e-01 -2.15084180e-01 2.10694764e-02 -2.72088051e-01
-1.15473843e+00 3.27839762e-01 1.70751080e-01 -1.11687645e-01
2.67274231e-01 -7.16067672e-01 -1.24884796e+00 3.21682513e-01
-9.52851593e-01 -7.97201455e-01 -3.83190602e-01 3.11218768e-01
-6.05120540e-01 3.00982535e-01 -9.07260597e-01 -8.75485063e-01
7.77753517e-02 -1.20496798e+00 6.14889979e-01 3.20568085e-01
-6.23606086e-01 -9.10172522e-01 -3.48534524e-01 5.62780976e-01
4.68975067e-01 -1.76926628e-01 1.59815335e+00 -1.09167540e+00
-5.73904276e-01 -4.79084179e-02 -4.17657569e-02 1.47495465e-02
-5.04572749e-01 1.13536879e-01 -8.20887148e-01 1.00389592e-01
2.59003907e-01 -5.33651233e-01 2.18567938e-01 5.34869768e-02
1.07010984e+00 -5.67361951e-01 1.43567964e-01 1.11423172e-01
9.50961113e-01 3.88404578e-02 5.50207734e-01 3.71965945e-01
6.76877722e-02 6.54783309e-01 7.44074762e-01 1.54201731e-01
7.61451781e-01 9.28355530e-02 1.38965517e-01 8.20441365e-01
3.98583598e-02 -7.39118457e-01 4.07208234e-01 5.74767232e-01
2.96109438e-01 -2.99873948e-01 -1.19505155e+00 4.43933636e-01
-1.69912040e+00 -1.02961564e+00 -2.32146472e-01 2.16435647e+00
1.36367965e+00 6.27668321e-01 -4.27969515e-01 4.77743328e-01
2.42981240e-01 -1.20522045e-01 -5.58941245e-01 -9.12402570e-01
-1.33178964e-01 4.22125816e-01 9.87112671e-02 9.67422664e-01
-1.48366513e-02 8.55151534e-01 6.73823071e+00 5.93491495e-01
-7.54963696e-01 -1.97544321e-01 2.82257110e-01 -1.89813092e-01
-8.20354760e-01 4.02650945e-02 -5.24477124e-01 5.45316875e-01
1.44848752e+00 -4.20862049e-01 5.20228684e-01 5.58282375e-01
1.34652317e-01 -6.76976144e-01 -1.31879926e+00 3.61079127e-01
-1.78762585e-01 -1.14532888e+00 1.52317092e-01 -2.98264563e-01
6.07866228e-01 -7.50787675e-01 1.96379632e-01 9.15639043e-01
9.03484523e-01 -1.09521544e+00 1.15329695e+00 6.86996460e-01
2.73114443e-01 -7.26279795e-01 9.30251360e-01 9.23290372e-01
-3.64159554e-01 -2.23112032e-01 -2.24137157e-01 -7.34543443e-01
-2.05755994e-01 4.87884074e-01 -1.17317176e+00 3.06145221e-01
3.94293427e-01 -8.41869861e-02 -7.07892597e-01 7.41745293e-01
-6.42620742e-01 9.08607244e-01 -3.78394663e-01 -1.20489433e-01
-3.86693217e-02 1.95047379e-01 5.31734645e-01 7.49607623e-01
2.72854745e-01 4.88010496e-01 -1.63139105e-01 1.27336061e+00
1.36093140e-01 -5.07939875e-01 -3.48573804e-01 -2.08055377e-01
1.03111756e+00 6.31495953e-01 -2.93543875e-01 -6.05603039e-01
-1.31180063e-01 1.90319523e-01 5.30762076e-01 5.13556838e-01
-7.87208080e-01 -5.63224331e-02 5.73062301e-01 1.08206391e-01
-1.29850626e-01 1.51820956e-02 -6.22373819e-01 -1.16496575e+00
6.59849867e-02 -1.52139986e+00 4.39620763e-01 -1.33931351e+00
-1.10370445e+00 -1.34729609e-01 4.10925210e-01 -2.79356003e-01
-4.65158701e-01 -6.58921123e-01 -6.86136365e-01 1.10366309e+00
-1.56622565e+00 -4.41474319e-01 -2.31619298e-01 3.57792318e-01
2.97382444e-01 4.17052716e-01 7.18325675e-01 -1.20043993e-01
-5.41632414e-01 6.75246954e-01 -5.89176178e-01 -2.80861616e-01
6.61560118e-01 -1.49628234e+00 2.86257505e-01 9.75691795e-01
-2.97818512e-01 1.09659815e+00 1.25205982e+00 -9.11226749e-01
-1.62810576e+00 -4.59735304e-01 9.93059397e-01 -9.32072997e-01
7.46539056e-01 -5.51564284e-02 -1.28593242e+00 1.00278068e+00
-1.55651882e-01 -2.10928947e-01 5.53516805e-01 2.11021323e-02
-5.71850002e-01 3.99933726e-01 -1.62209225e+00 9.00052309e-01
1.03260303e+00 -5.41501999e-01 -1.39014804e+00 7.39079043e-02
8.63986611e-01 -1.07528472e+00 -6.86967492e-01 8.33830312e-02
3.82977515e-01 -1.07010686e+00 7.63393044e-01 -1.09377933e+00
6.04203343e-01 -1.79799139e-01 -1.02744102e-01 -1.51194394e+00
-5.12191933e-03 -4.51972991e-01 -3.82091314e-01 1.04779124e+00
5.39329350e-01 -1.06785083e+00 4.20813739e-01 1.44627416e+00
-5.71236538e-04 -6.46889746e-01 -7.31140971e-01 -5.12693286e-01
3.81571889e-01 -8.85318756e-01 1.13276350e+00 8.78786862e-01
1.87044591e-01 -1.81802750e-01 4.51355487e-01 4.74478632e-01
4.33055609e-01 -9.38958954e-04 4.92426217e-01 -1.24899244e+00
-1.52523637e-01 -3.73858958e-01 1.47328615e-01 -7.05635726e-01
4.17830318e-01 -6.98191404e-01 4.16581929e-02 -1.44508469e+00
-2.17177883e-01 -6.02195501e-01 1.27156109e-01 5.38933754e-01
-5.36181152e-01 -6.25172734e-01 3.74780327e-01 -2.86270291e-01
-4.12258387e-01 1.52169004e-01 1.26489556e+00 1.79538891e-01
6.16466440e-02 -2.58645743e-01 -1.01995802e+00 8.29667807e-01
5.73836982e-01 -3.03273916e-01 -5.59364438e-01 -4.70645815e-01
1.02980423e+00 3.36218208e-01 6.81922793e-01 -7.93972015e-01
4.66972113e-01 -4.96800959e-01 1.80063203e-01 -2.77651906e-01
-1.17155891e-02 -6.23532712e-01 -2.25935310e-01 4.17882472e-01
-7.73212254e-01 4.37837481e-01 6.89610064e-01 8.86254236e-02
4.48415019e-02 -5.10085881e-01 1.79822356e-01 -2.28242651e-01
-5.55891514e-01 -6.80713773e-01 -6.07216299e-01 6.00516260e-01
7.61436999e-01 2.81018987e-02 -1.13766694e+00 -5.06483853e-01
-4.05776560e-01 5.26667655e-01 2.85816282e-01 -4.76299599e-02
5.49495280e-01 -8.31636727e-01 -5.72895110e-01 1.61587834e-01
-5.07155545e-02 2.66944051e-01 2.35625163e-01 8.92492771e-01
-5.54174006e-01 6.00273252e-01 -1.36831691e-02 -2.21066222e-01
-9.02147949e-01 5.10520101e-01 4.67144877e-01 -3.92219990e-01
-2.48368636e-01 8.83004844e-01 -3.08855265e-01 -6.44061327e-01
1.06437668e-01 -1.04826844e+00 9.55474004e-02 -8.17787200e-02
5.93933761e-01 4.48742628e-01 3.65059882e-01 1.77895635e-01
-1.72662377e-01 -9.49516892e-02 -4.19763811e-02 -2.64680266e-01
9.31127548e-01 2.02731211e-02 4.99348268e-02 4.95555371e-01
2.01726571e-01 4.60550547e-01 -8.66484463e-01 -1.82979599e-01
5.66911604e-03 -6.74824119e-01 2.80616619e-02 -1.49236727e+00
-4.39842343e-01 5.34790039e-01 -2.60556251e-01 4.28737938e-01
7.03742981e-01 -3.72747451e-01 4.54449922e-01 6.73524261e-01
5.93711436e-01 -8.69005799e-01 -2.18485460e-01 6.40689135e-01
9.04875755e-01 -9.67624784e-01 -1.32162467e-01 -4.33049619e-01
-6.15006685e-01 8.59238207e-01 8.95783305e-01 3.37341070e-01
-2.12670997e-01 6.85751915e-01 6.93791211e-02 -2.59309888e-01
-1.41952348e+00 2.24740922e-01 -2.67988443e-01 5.38235068e-01
2.05015481e-01 -5.67732230e-02 -3.19282323e-01 1.03607845e+00
-9.59235311e-01 2.89653629e-01 1.00501668e+00 1.15298867e+00
-3.27446252e-01 -6.83438599e-01 -9.13676322e-01 5.36000848e-01
-4.64001268e-01 -3.20943683e-01 -4.79241937e-01 8.24158370e-01
-3.16911250e-01 1.28012991e+00 -1.46948650e-01 -1.03002198e-01
6.44169152e-01 6.44390047e-01 7.18872905e-01 -5.27355611e-01
-7.69015431e-01 -1.01870096e+00 4.41816151e-01 -7.66117632e-01
2.87162215e-01 -6.23384297e-01 -1.43105197e+00 -8.86743188e-01
-2.25717761e-02 4.71201122e-01 3.06679875e-01 1.12771451e+00
3.39775175e-01 5.89077652e-01 -1.97687447e-01 7.03880750e-03
-1.29590297e+00 -9.06067371e-01 -1.67631209e-01 2.72771984e-01
2.71582872e-01 -7.94993103e-01 -6.38605773e-01 -5.71325421e-01] | [9.501642227172852, 7.31986665725708] |
029822de-e264-4809-bb3b-167defc880e7 | posecnn-a-convolutional-neural-network-for-6d | 1711.00199 | null | http://arxiv.org/abs/1711.00199v3 | http://arxiv.org/pdf/1711.00199v3.pdf | PoseCNN: A Convolutional Neural Network for 6D Object Pose Estimation in Cluttered Scenes | Estimating the 6D pose of known objects is important for robots to interact
with the real world. The problem is challenging due to the variety of objects
as well as the complexity of a scene caused by clutter and occlusions between
objects. In this work, we introduce PoseCNN, a new Convolutional Neural Network
for 6D object pose estimation. PoseCNN estimates the 3D translation of an
object by localizing its center in the image and predicting its distance from
the camera. The 3D rotation of the object is estimated by regressing to a
quaternion representation. We also introduce a novel loss function that enables
PoseCNN to handle symmetric objects. In addition, we contribute a large scale
video dataset for 6D object pose estimation named the YCB-Video dataset. Our
dataset provides accurate 6D poses of 21 objects from the YCB dataset observed
in 92 videos with 133,827 frames. We conduct extensive experiments on our
YCB-Video dataset and the OccludedLINEMOD dataset to show that PoseCNN is
highly robust to occlusions, can handle symmetric objects, and provide accurate
pose estimation using only color images as input. When using depth data to
further refine the poses, our approach achieves state-of-the-art results on the
challenging OccludedLINEMOD dataset. Our code and dataset are available at
https://rse-lab.cs.washington.edu/projects/posecnn/. | ['Yu Xiang', 'Venkatraman Narayanan', 'Dieter Fox', 'Tanner Schmidt'] | 2017-11-01 | null | null | null | null | ['6d-pose-estimation-using-rgbd'] | ['computer-vision'] | [-2.04823360e-01 -4.89761084e-01 -8.98576453e-02 -3.82364213e-01
-5.82840800e-01 -6.99521005e-01 2.22253859e-01 -3.21733654e-01
-5.02132595e-01 2.52275020e-01 -9.34173465e-02 1.30159527e-01
2.48320714e-01 -3.22840154e-01 -1.30881572e+00 -4.62846488e-01
-4.04502004e-01 8.76847029e-01 3.43192488e-01 1.21129053e-02
1.59832343e-01 1.09650505e+00 -1.35111713e+00 1.13334842e-01
1.49243534e-01 1.33636427e+00 4.24951971e-01 8.29542816e-01
4.13066119e-01 5.26617706e-01 -7.06067324e-01 -3.08513716e-02
7.74193645e-01 2.74860203e-01 -5.96353233e-01 3.57282460e-01
1.17400098e+00 -9.60001349e-01 -6.51612580e-01 8.58711064e-01
3.35061431e-01 1.09046057e-01 4.53566343e-01 -1.64241230e+00
-2.03737244e-01 7.86601678e-02 -7.12855458e-01 3.88892852e-02
4.32297528e-01 1.67133018e-01 7.18355715e-01 -1.13308322e+00
7.64559567e-01 1.63931715e+00 6.21397734e-01 3.26873064e-01
-9.63264942e-01 -8.60338986e-01 2.13860616e-01 3.41228902e-01
-1.46109235e+00 -3.27692538e-01 6.15216911e-01 -4.16318119e-01
1.07068443e+00 -8.84532705e-02 8.10520828e-01 9.22656953e-01
1.28553538e-02 9.18015540e-01 4.03341353e-01 -1.27958104e-01
-4.44030538e-02 -5.93292236e-01 -3.80433977e-01 7.57283390e-01
5.35105288e-01 -9.37825367e-02 -5.73449790e-01 1.76076114e-01
1.09733653e+00 1.61886156e-01 -2.03398481e-01 -1.11336541e+00
-1.53865480e+00 5.38333893e-01 7.84289479e-01 -4.90101844e-01
-6.83769658e-02 7.54685462e-01 1.80628404e-01 3.66559662e-02
1.90247506e-01 4.01086003e-01 -7.17777669e-01 -1.89405039e-01
-1.82519123e-01 5.54824650e-01 7.82767713e-01 1.42200601e+00
6.83286846e-01 -1.51011527e-01 3.98362160e-01 5.85290253e-01
4.37201142e-01 1.00720787e+00 -4.52120528e-02 -1.63356698e+00
7.60830104e-01 4.39519465e-01 3.52414191e-01 -1.16890728e+00
-6.67169392e-01 -2.71614671e-01 -3.33532035e-01 3.43660384e-01
5.13121605e-01 -3.43603455e-02 -8.73599887e-01 1.61199903e+00
5.46013713e-01 4.97983880e-02 -2.72626489e-01 1.16989672e+00
8.63102078e-01 5.89258552e-01 -5.43932498e-01 1.89277709e-01
1.14226866e+00 -1.18090284e+00 -3.47678781e-01 -5.03469884e-01
4.30828452e-01 -8.39838147e-01 6.33819938e-01 5.36818504e-01
-8.54127467e-01 -4.83770490e-01 -1.15370989e+00 -3.97898138e-01
-1.08301282e-01 3.70550811e-01 6.41170919e-01 8.57037604e-02
-8.28314722e-01 1.63852036e-01 -1.17881584e+00 -3.92746359e-01
4.00621742e-01 5.07676482e-01 -6.68049455e-01 -3.26918036e-01
-6.03304744e-01 1.06476235e+00 4.65691477e-01 2.13447437e-01
-1.13277233e+00 -4.42057163e-01 -1.09872484e+00 -1.85931697e-01
6.49689078e-01 -4.41381395e-01 1.49373555e+00 -4.38371062e-01
-1.10850132e+00 7.12084889e-01 1.22352429e-01 -3.82116228e-01
6.56057298e-01 -7.11848557e-01 1.09908067e-01 4.82298523e-01
9.19308960e-02 1.20270967e+00 8.16381454e-01 -1.28610516e+00
-6.29989326e-01 -6.36417210e-01 2.02596128e-01 5.09222329e-01
3.90902534e-02 -2.00929672e-01 -1.06152666e+00 -4.25025553e-01
6.30207300e-01 -1.49500573e+00 -6.01023026e-02 7.87534058e-01
-2.80815303e-01 -1.20128781e-01 1.28718448e+00 -2.85643041e-01
2.76812911e-01 -2.13931799e+00 2.13654861e-01 -2.02784613e-01
1.54930890e-01 8.66826400e-02 -4.49774027e-01 2.33362794e-01
-6.99217096e-02 -3.65177423e-01 2.91421026e-01 -5.23616493e-01
2.53000185e-02 3.40004683e-01 -8.29777718e-02 9.12664831e-01
1.96010202e-01 7.45701194e-01 -7.02913404e-01 -3.03359598e-01
3.08856219e-01 6.96249962e-01 -6.31836236e-01 2.82046437e-01
-3.05946499e-01 2.11299464e-01 -2.38984033e-01 8.50434601e-01
9.37537014e-01 -2.37620384e-01 -5.31076901e-02 -6.62344158e-01
-2.51365956e-02 1.37510344e-01 -1.36406791e+00 1.96390820e+00
-1.22206032e-01 9.06339765e-01 1.12293243e-01 -6.69460893e-01
8.37546289e-01 -1.14556000e-01 7.48635352e-01 -1.83533669e-01
2.27216631e-01 1.32232353e-01 -1.36854038e-01 -4.60150570e-01
5.60886204e-01 5.05962849e-01 -6.99112490e-02 2.25598052e-01
6.73991516e-02 -6.14823520e-01 4.01745260e-01 1.31826475e-01
9.42206919e-01 4.56528395e-01 1.31684944e-01 8.82683098e-02
2.45961584e-02 9.03606713e-02 4.96056944e-01 4.54033285e-01
-2.82647491e-01 7.88376391e-01 3.17279249e-01 -6.90569699e-01
-1.15844071e+00 -1.15541017e+00 -1.73404604e-01 6.33935571e-01
6.05173171e-01 -3.36071640e-01 -3.24615747e-01 -4.45830941e-01
4.56166446e-01 1.22987248e-01 -3.96986574e-01 -1.04083374e-01
-7.94171929e-01 -2.82830656e-01 2.21871912e-01 7.72736311e-01
6.26385629e-01 -5.61574578e-01 -8.95780742e-01 -1.25143733e-02
-3.99936110e-01 -1.59383225e+00 -5.75032651e-01 3.17402154e-01
-8.52636933e-01 -1.34930563e+00 -4.87141103e-01 -9.00791585e-01
6.46871984e-01 7.56630003e-01 9.94836032e-01 -1.95201933e-01
-4.06228840e-01 3.79156649e-01 -3.79550934e-01 -4.65966612e-01
3.25977020e-02 2.43442645e-03 3.52423757e-01 -4.53486115e-01
2.72134811e-01 -1.88768610e-01 -6.90105736e-01 6.70611978e-01
-5.80606282e-01 -5.43920621e-02 3.92357290e-01 4.28953111e-01
5.77506483e-01 -1.37715608e-01 -1.13736451e-01 -7.47023225e-02
-3.08401793e-01 -2.85256444e-03 -9.98699784e-01 -1.32107958e-01
9.33335498e-02 -1.94482371e-01 1.64518654e-01 -6.50380135e-01
-5.78755796e-01 6.52425289e-01 1.97169363e-01 -8.45361829e-01
-1.28619626e-01 1.05380200e-01 -8.88488218e-02 -2.70003110e-01
4.89707917e-01 -3.84029359e-01 1.36975959e-01 -4.76752400e-01
2.05735445e-01 4.54572737e-01 7.83146620e-01 -4.78043288e-01
1.00503635e+00 6.69676244e-01 3.04246187e-01 -7.71307886e-01
-9.45178509e-01 -5.59724092e-01 -8.94590139e-01 -3.79815191e-01
7.99966514e-01 -1.35931313e+00 -1.13812184e+00 7.01553941e-01
-1.41726589e+00 -3.07259917e-01 1.28858089e-01 8.74541700e-01
-6.83568776e-01 2.34887734e-01 -6.27139509e-01 -4.55328971e-01
-4.20526639e-02 -1.42471445e+00 1.55925405e+00 -9.08756182e-02
2.79451255e-02 -3.61241609e-01 -2.83091962e-01 3.84027302e-01
-4.25600670e-02 4.11986321e-01 3.62395436e-01 -1.89835310e-01
-1.05930197e+00 -2.67951936e-01 -2.88129121e-01 2.13105619e-01
1.01783469e-01 1.20728187e-01 -5.72720110e-01 -6.31191373e-01
-1.73189998e-01 -7.29638398e-01 7.73992777e-01 3.65599304e-01
1.08361876e+00 -9.02002901e-02 -2.22190186e-01 1.05701411e+00
1.36027908e+00 1.07690021e-01 2.75701046e-01 5.91015100e-01
9.84273970e-01 3.37771595e-01 9.40466344e-01 5.81869960e-01
4.46762294e-01 8.72104228e-01 1.15819526e+00 1.72807172e-01
-8.97845253e-02 -1.32214889e-01 3.22671711e-01 4.75074142e-01
-1.40433656e-02 -2.65033484e-01 -9.42631185e-01 1.23375230e-01
-1.67532933e+00 -5.44371068e-01 -5.76885641e-02 1.96945262e+00
5.65619767e-01 1.98307499e-01 7.39861699e-03 -1.46191508e-01
6.73199356e-01 1.43224344e-01 -8.15700412e-01 2.80993193e-01
-3.06754503e-02 -3.76068383e-01 7.61593342e-01 3.89272988e-01
-1.39192891e+00 9.43916559e-01 5.84300566e+00 4.01484430e-01
-1.15604436e+00 -4.14376974e-01 1.99226052e-01 -4.60524768e-01
3.79689544e-01 -3.01398426e-01 -1.07500565e+00 3.53236124e-02
1.87025443e-01 2.43485749e-01 4.50628996e-01 1.09060395e+00
-1.26190469e-01 -2.54419386e-01 -1.38071525e+00 1.31102836e+00
4.58304584e-01 -1.02163613e+00 -1.57613143e-01 -4.89999466e-02
7.67482162e-01 6.03329182e-01 1.38587028e-01 -2.04824489e-02
5.82957901e-02 -8.12651098e-01 1.17270756e+00 4.37047556e-02
7.88840175e-01 -6.66831613e-01 5.96524894e-01 2.31869072e-01
-1.28688061e+00 -9.63144675e-02 -6.86276197e-01 -1.62862048e-01
3.55477892e-02 2.28418142e-01 -1.21720886e+00 3.90844978e-02
1.19663548e+00 1.04420054e+00 -7.13480532e-01 1.23389781e+00
-2.67766386e-01 -7.00795650e-02 -6.97800159e-01 -4.67054499e-03
9.12087262e-02 2.61308048e-02 6.13892794e-01 7.49992132e-01
4.79032606e-01 -9.70182533e-04 4.29247200e-01 5.82582235e-01
-1.76553845e-01 -4.08368438e-01 -5.54238021e-01 1.07370958e-01
5.86445391e-01 1.19698811e+00 -8.38339567e-01 -1.22959763e-01
-1.86474636e-01 8.32813740e-01 3.52410525e-01 3.32355410e-01
-9.61551905e-01 -1.49049371e-01 8.76948297e-01 -2.44878471e-01
7.09890366e-01 -9.44435060e-01 1.46560490e-01 -1.26571202e+00
1.61308423e-01 -8.84708047e-01 1.41508933e-02 -1.29107094e+00
-9.77707088e-01 4.58428741e-01 2.53742218e-01 -1.57983959e+00
-5.45506775e-02 -1.17837179e+00 9.80497599e-02 3.69278282e-01
-1.21872568e+00 -1.06397569e+00 -7.40258992e-01 4.36297804e-01
7.02920377e-01 7.19467402e-02 5.60903728e-01 1.62597671e-01
-1.41932547e-01 2.59111047e-01 -7.03972625e-03 3.02411109e-01
1.00010157e+00 -1.07061815e+00 5.33127964e-01 5.07392585e-01
1.58083498e-01 4.34703410e-01 6.51475489e-01 -5.77483177e-01
-1.98653674e+00 -1.02467489e+00 4.55310047e-01 -7.49655306e-01
5.30937850e-01 -6.72882318e-01 -3.54995459e-01 1.00932157e+00
-2.36275703e-01 3.55947256e-01 1.54174492e-01 -3.21853280e-01
-5.09910047e-01 -3.08265835e-01 -8.00714433e-01 5.55681169e-01
1.26014483e+00 -2.47206151e-01 -3.85158896e-01 4.83079642e-01
1.12937140e+00 -1.19264734e+00 -7.69955635e-01 5.45858443e-01
8.69345903e-01 -7.64154851e-01 1.23154724e+00 -2.80421883e-01
2.79786319e-01 -6.46616459e-01 -4.40354139e-01 -1.12672770e+00
-1.54482931e-01 -1.65363774e-01 -2.14739233e-01 5.80590546e-01
9.69096869e-02 -2.39511743e-01 9.38768148e-01 3.31401646e-01
-9.82703120e-02 -5.78102767e-01 -8.22898865e-01 -8.79385889e-01
-3.59302580e-01 -5.41086257e-01 5.07844746e-01 5.15538752e-01
-5.89213312e-01 1.79330885e-01 -3.73213619e-01 5.16609490e-01
7.10237086e-01 3.39944571e-01 1.37332308e+00 -1.04274631e+00
-2.83243507e-02 -2.64261160e-02 -8.04903448e-01 -1.67355740e+00
1.63910836e-01 -4.58443999e-01 2.72599906e-01 -1.34699523e+00
1.13522224e-01 -2.83301532e-01 3.07285190e-01 4.67093557e-01
2.51444548e-01 6.93491995e-01 4.67508346e-01 2.02155784e-01
-8.37915003e-01 5.33395886e-01 1.49996245e+00 -3.34626853e-01
6.40731230e-02 -7.70833939e-02 1.14780553e-02 9.06932354e-01
7.41960645e-01 -3.65234107e-01 -2.04198048e-01 -9.77002919e-01
2.62148112e-01 2.89685111e-02 5.73980331e-01 -1.16460264e+00
1.54404253e-01 -7.90470019e-02 9.41779912e-01 -1.34815753e+00
9.17968869e-01 -1.20461130e+00 6.35874793e-02 6.59138978e-01
-1.29956707e-01 3.31956804e-01 1.79200336e-01 4.42928880e-01
1.28708547e-02 -6.07016571e-02 6.90603971e-01 -1.71856403e-01
-9.22800839e-01 6.75727367e-01 -2.71340683e-02 -7.21833762e-03
9.56955969e-01 -1.71981603e-01 -4.23916370e-01 -4.92017329e-01
-2.71871001e-01 4.47309852e-01 7.51053333e-01 6.48716211e-01
7.06956565e-01 -1.51398301e+00 -5.30111432e-01 3.08938116e-01
2.48654112e-01 6.80600166e-01 -1.41611472e-01 6.70131624e-01
-1.31478524e+00 3.48266482e-01 -3.36314023e-01 -1.19317031e+00
-1.34825742e+00 4.98078555e-01 2.18215644e-01 4.74496067e-01
-4.24238503e-01 1.01113689e+00 2.38172367e-01 -5.98282874e-01
7.90373027e-01 -6.89725995e-01 1.17783867e-01 -1.12425603e-01
4.03712690e-01 4.85717028e-01 -7.14001507e-02 -8.55763614e-01
-5.66202521e-01 8.40397716e-01 3.27080302e-02 5.17781377e-02
1.45905328e+00 -1.89962223e-01 -2.27201566e-01 2.42936343e-01
1.65282261e+00 -3.23649615e-01 -1.88020778e+00 -1.96838200e-01
-2.96811610e-01 -8.08440447e-01 -1.94761649e-01 -3.77087951e-01
-1.21911490e+00 7.84175634e-01 5.86037040e-01 -4.04890120e-01
8.02428007e-01 1.97317541e-01 7.09447682e-01 1.13070130e+00
6.07073128e-01 -9.81909156e-01 6.97131991e-01 9.68930781e-01
1.17777193e+00 -1.45060813e+00 2.68523842e-01 -4.88662869e-01
-3.31273884e-01 1.35736704e+00 1.06107020e+00 -2.51033455e-01
4.33301866e-01 4.13504511e-01 2.82248825e-01 -5.56273735e-04
-4.93352592e-01 8.80161747e-02 2.56348491e-01 5.19827366e-01
-2.79255938e-02 -1.28876537e-01 4.29329634e-01 -1.39460564e-01
-3.94411147e-01 -3.54548633e-01 5.39103746e-01 1.00293589e+00
-5.64539492e-01 -6.33408308e-01 -7.13124573e-01 -2.95564067e-02
-1.80150524e-01 3.01195115e-01 -3.71725112e-01 9.36376154e-01
1.33146256e-01 7.26107836e-01 2.88762391e-01 -3.40142280e-01
3.41419667e-01 -4.98529047e-01 8.74435902e-01 -5.26630461e-01
1.22171544e-01 4.08146501e-01 8.08268934e-02 -9.61616516e-01
-6.30316913e-01 -7.50895977e-01 -1.35053587e+00 -1.02349892e-01
-3.62694085e-01 -3.71346235e-01 8.91428947e-01 5.80702126e-01
9.28297043e-02 1.91274151e-01 5.55669665e-01 -1.60980642e+00
-6.10888004e-01 -8.80932450e-01 -5.33431232e-01 3.91280025e-01
6.24974132e-01 -9.46306288e-01 -3.60757947e-01 5.44562712e-02] | [7.314334869384766, -2.472534656524658] |
7bc52841-e5ab-409b-bdf2-914dc66439d9 | batgpt-a-bidirectional-autoregessive-talker | 2307.00360 | null | https://arxiv.org/abs/2307.00360v1 | https://arxiv.org/pdf/2307.00360v1.pdf | BatGPT: A Bidirectional Autoregessive Talker from Generative Pre-trained Transformer | BatGPT is a large-scale language model designed and trained jointly by Wuhan University and Shanghai Jiao Tong University. It is capable of generating highly natural and fluent text in response to various types of input, including text prompts, images, and audio. In the modeling level, we employ a bidirectional autoregressive architecture that allows the model to efficiently capture the complex dependencies of natural language, making it highly effective in tasks such as language generation, dialog systems, and question answering. Moreover, the bidirectional autoregressive modeling not only operates from left to right but also from right to left, effectively reducing fixed memory effects and alleviating model hallucinations. In the training aspect, we propose a novel parameter expansion method for leveraging the pre-training of smaller models and employ reinforcement learning from both AI and human feedback, aimed at improving the model's alignment performance. Overall, these approaches significantly improve the effectiveness of BatGPT, and the model can be utilized for a wide range of natural language applications. | ['Dongjie Yang', 'Yifei Yang', 'Hai Zhao', 'Shitou Zhang', 'Zuchao Li'] | 2023-07-01 | null | null | null | null | ['text-generation', 'question-answering'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.54217882e-02 -3.80133055e-02 5.86788077e-03 -2.54168481e-01
-5.48860252e-01 -2.45317072e-01 6.39976740e-01 -1.23476535e-01
-1.65992349e-01 3.84775579e-01 5.96143603e-01 -4.11415011e-01
4.21307296e-01 -8.89027119e-01 -5.59086919e-01 -3.42435271e-01
4.11514759e-01 5.47920763e-01 3.56089175e-02 -5.37932396e-01
1.70017302e-01 2.17218965e-01 -9.14619446e-01 4.14362848e-01
1.03247654e+00 8.87260258e-01 7.20701337e-01 7.43471920e-01
-3.36501002e-01 1.41536129e+00 -5.49716175e-01 -1.11752376e-01
-2.33127952e-01 -4.19519842e-01 -6.14600837e-01 1.24411955e-01
-1.17096692e-01 -6.65870965e-01 -6.59589589e-01 4.86946493e-01
7.02630341e-01 1.56356901e-01 4.08598393e-01 -1.00237429e+00
-1.09191573e+00 6.53368831e-01 -4.43647057e-01 -1.59025177e-01
4.76941049e-01 4.63414550e-01 9.25144136e-01 -1.04794216e+00
1.97194993e-01 1.81990337e+00 1.71782359e-01 7.15724289e-01
-8.73206496e-01 -8.64599645e-01 5.97062707e-01 1.87294737e-01
-1.03056943e+00 -4.35426176e-01 7.14533329e-01 -4.68121767e-01
1.03244209e+00 2.17275396e-01 5.05526066e-01 1.26178372e+00
1.80983528e-01 1.11278272e+00 8.10810626e-01 -4.34973389e-01
-4.08019908e-02 1.27260268e-01 -2.72672951e-01 5.54853916e-01
-6.09524667e-01 -4.28129286e-02 -6.30968451e-01 -1.25906751e-01
9.29376960e-01 4.04783785e-02 -2.05952734e-01 2.88544805e-03
-1.30071795e+00 8.18925023e-01 5.81715703e-01 2.15718359e-01
-5.80960155e-01 1.22220434e-01 2.13717327e-01 -1.81903988e-02
3.85497749e-01 5.73785245e-01 -2.91394472e-01 -2.98915617e-02
-4.16821122e-01 2.16912434e-01 3.02527457e-01 1.10984361e+00
4.83269781e-01 4.46422815e-01 -6.50309861e-01 1.13562870e+00
4.87944424e-01 7.48734713e-01 6.82577550e-01 -8.01679969e-01
7.73670137e-01 7.78563559e-01 1.30403355e-01 -9.50592339e-01
-3.78624707e-01 -2.67537236e-01 -1.20841348e+00 -3.75518560e-01
-1.27955556e-01 -2.87069440e-01 -9.50713813e-01 1.91371989e+00
1.47630706e-01 9.87012684e-02 1.81643069e-01 6.41665280e-01
9.20184255e-01 1.28663874e+00 3.17256093e-01 -1.47033125e-01
1.31850040e+00 -1.33553970e+00 -9.75589454e-01 -6.83252513e-01
3.89818668e-01 -7.98569918e-01 1.49751246e+00 2.92609870e-01
-1.05956173e+00 -7.59546757e-01 -6.36232793e-01 -3.53680193e-01
-6.21892437e-02 4.13634807e-01 4.09567147e-01 -4.81304713e-03
-8.97675216e-01 8.34016502e-03 -6.29021347e-01 2.41393666e-03
5.20148128e-02 2.11669385e-01 2.60479823e-02 -3.03133577e-01
-1.46888399e+00 8.73376131e-01 2.55320072e-01 5.73787689e-01
-7.07270980e-01 -3.86470556e-01 -9.29654598e-01 3.04058641e-01
3.01249266e-01 -8.06762695e-01 1.38091147e+00 -6.30026281e-01
-1.89528430e+00 2.82752395e-01 -5.28431058e-01 -3.57577801e-01
4.53891665e-01 -4.74404424e-01 -3.95140976e-01 -6.01742305e-02
-1.87424660e-01 9.63760555e-01 8.52990091e-01 -9.80476856e-01
-2.26107717e-01 -2.37991348e-01 -3.78009165e-04 4.23548222e-01
-5.77991068e-01 5.18686287e-02 -6.85193062e-01 -8.26861203e-01
-1.65545896e-01 -9.33221877e-01 -4.67456192e-01 -3.53643984e-01
-3.71380627e-01 -5.08145750e-01 9.13733184e-01 -8.47200155e-01
1.36080456e+00 -2.07398891e+00 1.21930383e-01 -1.42545238e-01
1.04044430e-01 3.46367627e-01 -5.53085983e-01 6.69866025e-01
7.97718838e-02 1.77913830e-01 8.67745094e-03 -3.43633771e-01
-1.18172757e-01 9.77765620e-02 -8.55559707e-01 -3.16515505e-01
4.10119712e-01 1.17305279e+00 -8.14085424e-01 -3.96708369e-01
2.33279735e-01 3.32955092e-01 -6.93799198e-01 7.80963838e-01
-5.67122936e-01 6.33020461e-01 -6.29184663e-01 1.65617794e-01
3.71404529e-01 -4.31237608e-01 -5.89807816e-02 1.21438399e-01
7.13383313e-03 2.96137422e-01 -6.91232026e-01 1.51794970e+00
-9.29476321e-01 4.96665180e-01 -8.22286904e-02 -5.81509233e-01
1.38970590e+00 4.59288508e-01 7.79959857e-02 -1.02435029e+00
1.16805332e-02 -5.79556562e-02 6.64079487e-02 -7.19177783e-01
6.27813160e-01 -3.36293839e-02 -1.42055273e-01 5.93267739e-01
-9.37667564e-02 -4.01270032e-01 1.42430529e-01 3.97484303e-01
6.38901234e-01 -1.72047615e-01 3.95497419e-02 1.79476082e-01
7.40901172e-01 -4.02563572e-01 4.73311633e-01 7.49165118e-01
2.02549696e-01 2.72839308e-01 4.38478708e-01 -2.43030101e-01
-8.93073261e-01 -7.23401248e-01 4.75082964e-01 1.52904618e+00
1.60427332e-01 -4.74570692e-01 -5.59514046e-01 -2.21842110e-01
-3.12052637e-01 1.02471399e+00 -4.49561477e-01 -6.63268864e-01
-6.65592551e-01 -2.68597126e-01 4.03090835e-01 7.96116412e-01
6.24203861e-01 -1.58785093e+00 -1.08442865e-01 2.85128444e-01
-5.00873387e-01 -1.30563593e+00 -8.70914817e-01 -4.86170083e-01
-7.89033651e-01 -4.68016893e-01 -8.75520229e-01 -9.51782048e-01
5.74937880e-01 3.41516316e-01 1.04359901e+00 2.18911752e-01
1.92137584e-01 2.68463552e-01 -2.99466014e-01 -5.62073886e-01
-6.17777228e-01 9.23399851e-02 -9.92577374e-02 6.17708825e-02
3.45170014e-02 -2.27866605e-01 -3.35583866e-01 4.68736172e-01
-1.02106762e+00 5.52673042e-01 8.00025165e-01 1.07432330e+00
1.99107245e-01 -4.48377371e-01 8.33413541e-01 -8.02858174e-01
1.20729184e+00 -3.00151289e-01 -4.45705235e-01 5.51684201e-01
-4.47784126e-01 4.59214225e-02 1.00527298e+00 -7.07550406e-01
-1.26027691e+00 -2.35277250e-01 -2.63312876e-01 -3.41689050e-01
3.58017460e-02 5.82617342e-01 -2.53454179e-01 2.98139393e-01
4.24787045e-01 5.38047194e-01 -1.70486078e-01 -2.80306727e-01
3.68031383e-01 9.33487296e-01 5.49887002e-01 -4.43035603e-01
5.89655876e-01 -2.95673441e-02 -4.48889464e-01 -8.03801119e-01
-6.14168108e-01 -2.81327665e-01 -2.23054811e-01 -7.13189691e-02
7.42283225e-01 -1.10361838e+00 -8.33201468e-01 7.48112679e-01
-1.55682337e+00 -4.29578006e-01 2.25122303e-01 4.35183495e-01
-4.17874068e-01 1.61469162e-01 -8.03211510e-01 -8.55584860e-01
-5.97939610e-01 -1.22381949e+00 9.87352431e-01 5.12745321e-01
-2.29994610e-01 -8.86135578e-01 -8.86046737e-02 4.15860981e-01
5.83764017e-01 -4.94532645e-01 1.24946380e+00 -6.56668782e-01
-6.56598985e-01 -8.95458236e-02 -1.41323805e-01 3.80744845e-01
-3.00949048e-02 -7.47468695e-02 -7.21737266e-01 -1.88124508e-01
-2.52692755e-02 -5.83780050e-01 5.70383012e-01 3.85394581e-02
1.26315343e+00 -5.62619865e-01 -1.41606871e-02 2.31059805e-01
7.28141010e-01 4.34890598e-01 6.97360814e-01 -3.45660858e-02
7.31836736e-01 6.65477514e-01 6.06413484e-01 4.92918760e-01
6.66121066e-01 6.43747389e-01 2.24879041e-01 -3.66344333e-01
1.17470011e-01 -8.89023423e-01 3.22704703e-01 1.09093428e+00
4.95654047e-01 -5.23029089e-01 -9.59403992e-01 4.95393634e-01
-1.88175249e+00 -7.22811222e-01 1.12426639e-01 1.81810951e+00
8.67907047e-01 1.12020999e-01 -3.28581482e-01 -4.69618380e-01
7.23281562e-01 2.40255296e-01 -7.03220606e-01 -4.47426528e-01
-1.26037840e-02 -1.34294480e-01 -1.25123337e-01 5.94683349e-01
-5.77911913e-01 1.33947039e+00 6.48356247e+00 7.18452692e-01
-1.37301910e+00 -2.88939238e-01 7.96895027e-01 1.80368289e-01
-4.22002465e-01 -7.49302506e-02 -5.87574184e-01 4.48153764e-01
7.88261116e-01 -4.27032799e-01 4.96884525e-01 7.74323344e-01
6.47607446e-01 1.01293609e-01 -8.24766815e-01 8.90756726e-01
2.52781305e-02 -1.31987214e+00 4.67966706e-01 -2.63676316e-01
6.88069344e-01 -2.22371489e-01 2.26311594e-01 8.21858287e-01
3.76034498e-01 -1.19247782e+00 4.87489164e-01 5.13035893e-01
5.30221999e-01 -6.18368328e-01 5.68902731e-01 9.27698672e-01
-1.04069853e+00 -1.91176146e-01 -2.21802101e-01 -5.33580445e-02
3.83660823e-01 7.50577524e-02 -1.13780856e+00 2.60765344e-01
2.29514956e-01 4.77409124e-01 -4.80611533e-01 6.53577447e-01
-5.90366125e-01 6.73837185e-01 -4.95267585e-02 -2.82352358e-01
1.71938956e-01 -1.32976547e-01 2.28218883e-01 1.05235410e+00
2.49813959e-01 3.29260349e-01 4.86888200e-01 8.12683284e-01
-3.13918024e-01 3.08510303e-01 -4.19658273e-01 -3.32186937e-01
6.92856014e-01 1.13414156e+00 -2.82393750e-02 -5.54927647e-01
-9.87156853e-02 7.19498813e-01 3.44454885e-01 6.87683105e-01
-6.86960280e-01 -2.33053580e-01 9.82621312e-02 -1.76292479e-01
-5.41905984e-02 -4.22884166e-01 -3.41600984e-01 -1.08979332e+00
-5.06470166e-02 -1.20143414e+00 1.20354556e-01 -1.37328243e+00
-1.01858103e+00 8.96055758e-01 -2.16860428e-01 -9.81319368e-01
-8.40740383e-01 -3.44615102e-01 -8.11037719e-01 1.13731837e+00
-1.35583973e+00 -1.27795124e+00 -3.31401378e-01 5.97561896e-01
8.58237147e-01 -2.45244488e-01 8.26025724e-01 -5.45993932e-02
-6.90021157e-01 4.62426662e-01 -2.12816536e-01 1.59666523e-01
9.00440276e-01 -9.22579348e-01 5.93791008e-01 8.66619051e-01
4.76736501e-02 9.13647652e-01 5.75008035e-01 -5.42558789e-01
-1.24746418e+00 -1.09376335e+00 9.48922157e-01 3.69766774e-03
5.66513479e-01 -5.32846332e-01 -1.09661853e+00 5.64339459e-01
3.93202752e-01 -4.61110294e-01 4.97707218e-01 4.28380072e-02
-2.96652377e-01 -7.57863447e-02 -5.05016744e-01 8.90924633e-01
4.69522744e-01 -7.01278090e-01 -3.09216589e-01 5.50611496e-01
1.08956027e+00 -7.23297179e-01 -4.94650900e-01 3.93575817e-01
3.16808432e-01 -5.95563352e-01 8.47473741e-01 -7.56815016e-01
8.35096836e-01 2.27836948e-02 2.65956849e-01 -1.32942414e+00
-4.43548232e-01 -8.78599703e-01 -2.48518884e-01 1.33746648e+00
3.02518874e-01 -5.55830359e-01 3.99121910e-01 7.94132292e-01
-1.90050066e-01 -9.58476484e-01 -5.01680315e-01 -2.92030871e-01
3.41408812e-02 -2.56891698e-01 6.64498568e-01 4.47019547e-01
-1.67362556e-01 9.51963484e-01 -9.70170438e-01 2.63855215e-02
-2.35326305e-01 1.55028284e-01 1.04549885e+00 -7.43421495e-01
-3.77902240e-01 -3.28674406e-01 3.34667802e-01 -2.00812602e+00
1.86972782e-01 -4.47034001e-01 4.93716151e-01 -1.60229218e+00
-5.64165860e-02 -3.51415157e-01 4.18854877e-02 4.06348765e-01
-5.75469196e-01 -3.76051217e-01 4.70635206e-01 3.64642799e-01
-3.75638545e-01 1.00727332e+00 1.71865487e+00 -1.96717128e-01
-1.66338265e-01 1.46463007e-01 -9.06223059e-01 6.83492303e-01
5.61285377e-01 -1.25778407e-01 -5.30655205e-01 -8.57597172e-01
-3.18941958e-02 2.97610193e-01 2.87008971e-01 -5.44179738e-01
5.03190756e-01 -4.23712671e-01 3.23261529e-01 -5.82368553e-01
5.34164906e-01 -5.38170457e-01 -3.39125901e-01 3.38238746e-01
-8.79959941e-01 3.47109109e-01 2.83072263e-01 5.62947869e-01
-4.30300891e-01 -7.52000837e-03 6.12912059e-01 -1.66263089e-01
-5.82404196e-01 4.50868279e-01 -4.07134145e-01 -2.47359518e-02
7.36317813e-01 3.72760862e-01 -3.00945401e-01 -1.04780543e+00
-3.28666151e-01 8.84060264e-01 9.32621881e-02 8.71702552e-01
8.01218271e-01 -1.31962919e+00 -8.36433709e-01 2.56739348e-01
-2.00542565e-02 2.18170017e-01 6.78719699e-01 5.78063786e-01
-2.27461532e-01 7.15293407e-01 1.51346728e-01 -4.95195240e-01
-9.91651177e-01 5.61937213e-01 8.72875601e-02 -7.00690567e-01
-2.64231592e-01 8.40498388e-01 6.77736402e-01 -5.80798566e-01
2.54812151e-01 -5.71385100e-02 -4.25783485e-01 -3.14852178e-01
7.40312815e-01 -5.47274761e-03 -2.08492324e-01 -4.07449901e-01
3.55670415e-02 1.71239331e-01 -4.35433447e-01 -2.39685729e-01
1.05857205e+00 -2.11710200e-01 -5.25573380e-02 4.27325606e-01
8.51428986e-01 -3.72490436e-02 -1.15823591e+00 -4.13983762e-01
-3.60482752e-01 -2.29393229e-01 -1.77973539e-01 -9.01376486e-01
-6.96883380e-01 1.44203699e+00 1.22834302e-01 8.13654512e-02
1.08668685e+00 -1.78596050e-01 1.13319767e+00 6.64734066e-01
-7.01687019e-03 -1.05617046e+00 6.50921941e-01 9.34869587e-01
1.38069582e+00 -9.07244802e-01 -3.40155840e-01 -1.44210383e-01
-1.12689841e+00 1.11053872e+00 9.90906835e-01 7.22642541e-02
-6.66595250e-02 2.54782643e-02 2.96532214e-01 2.83158183e-01
-1.33978391e+00 1.80114537e-01 2.99104869e-01 4.90686178e-01
5.45138180e-01 -1.65997893e-01 -2.59134043e-02 6.30438268e-01
-2.20411494e-01 4.81099673e-02 4.17701811e-01 5.59865713e-01
-4.43135291e-01 -1.11129701e+00 -3.09845626e-01 1.83746710e-01
7.60812163e-02 -4.04807657e-01 -4.87498194e-01 3.61319363e-01
-2.44419143e-01 1.19586492e+00 -1.18386649e-01 -3.99716765e-01
3.59243631e-01 -1.36161387e-01 -5.64863831e-02 -6.52284980e-01
-5.41880071e-01 2.31659979e-01 -1.19672984e-01 -2.22931877e-01
-6.98910840e-03 -1.38776869e-01 -1.23961568e+00 3.70700397e-02
-4.64142412e-01 2.24406257e-01 4.27015483e-01 1.15406227e+00
6.80953979e-01 6.44812644e-01 9.81295884e-01 -5.37572443e-01
-8.13263476e-01 -1.31849134e+00 -9.04512480e-02 2.59673059e-01
1.24663413e-01 -3.80952567e-01 9.80076343e-02 5.70596009e-02] | [12.402861595153809, 8.556004524230957] |
8ffdd5a0-726a-4729-b7fe-cfe960ab92c4 | data-needs-for-integrated-economic | 2209.01487 | null | https://arxiv.org/abs/2209.01487v1 | https://arxiv.org/pdf/2209.01487v1.pdf | Data needs for integrated economic-epidemiological models of pandemic mitigation policies | The COVID-19 pandemic and the mitigation policies implemented in response to it have resulted in economic losses worldwide. Attempts to understand the relationship between economics and epidemiology has lead to a new generation of integrated mathematical models. The data needs for these models transcend those of the individual fields, especially where human interaction patterns are closely linked with economic activity. In this article, we reflect upon modelling efforts to date, discussing the data needs that they have identified, both for understanding the consequences of the pandemic and policy responses to it through analysis of historic data and for the further development of this new and exciting interdisciplinary field. | ['Katharina D. Hauck', 'Peter C. Smith', 'Patrick Doohan', 'Robert Johnson', 'Giovanni Forchini', 'Christian Morgenstern', 'David J. Haw'] | 2022-09-03 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 1.61562100e-01 -5.30797653e-02 -1.42438054e-01 1.94136858e-01
1.92161471e-01 -3.54732484e-01 8.55763912e-01 7.38580585e-01
-8.53287220e-01 6.69969201e-01 5.59823811e-01 -8.27289462e-01
-7.41471469e-01 -8.52587342e-01 -2.17200086e-01 -4.46000487e-01
-6.38862193e-01 6.16447151e-01 -7.96248913e-02 -4.69561279e-01
1.41668558e-01 6.00131810e-01 -1.10807633e+00 -3.41080159e-01
6.88888133e-01 3.15722376e-01 1.91025749e-01 6.27394617e-01
-1.11641839e-01 6.96763873e-01 -6.24064624e-01 -1.28201753e-01
-1.32416934e-01 -4.56744105e-01 -5.56439757e-01 -4.68860835e-01
-8.43612790e-01 -2.01181993e-01 -1.50153309e-01 4.98975515e-01
4.90400046e-01 1.34432495e-01 7.87692189e-01 -1.20652461e+00
-4.84188527e-01 1.29751965e-01 -4.64136690e-01 6.67090893e-01
4.59111691e-01 5.45455813e-02 1.31934777e-01 -3.63301903e-01
7.87641406e-01 1.33301198e+00 9.86369550e-01 6.97753802e-02
-1.06036544e+00 -5.14104486e-01 -9.93646607e-02 -4.34027165e-02
-1.21375465e+00 -2.82204777e-01 1.73874214e-01 -7.42971897e-01
1.18228865e+00 2.02608615e-01 9.28987324e-01 6.72542393e-01
3.95367414e-01 -1.78425461e-01 9.45932865e-01 -5.39927900e-01
-5.64992800e-03 2.02886820e-01 -9.66869965e-02 1.07347846e-01
6.52033508e-01 6.15656078e-01 -7.71595836e-02 -4.99094814e-01
5.89515090e-01 4.92501169e-01 -1.24524422e-01 2.46688217e-01
-8.24815929e-01 8.54125619e-01 1.89551860e-01 6.13796473e-01
-6.65422499e-01 1.32788410e-02 1.65582016e-01 9.13070589e-02
6.47148669e-01 2.42341951e-01 -4.04274434e-01 -1.02648653e-01
-7.29753911e-01 3.82055134e-01 6.86186016e-01 4.80863489e-02
2.79738069e-01 -1.47944227e-01 4.32769984e-01 4.36263978e-01
5.35403371e-01 8.66784096e-01 -2.21831426e-01 -7.95044422e-01
1.40144750e-01 4.14821655e-01 3.46358687e-01 -1.52563679e+00
-7.41057515e-01 -2.37438887e-01 -7.89532423e-01 -1.94830388e-01
5.23204207e-01 -4.30608481e-01 -5.72405934e-01 1.72475612e+00
3.56206685e-01 1.22527041e-01 -6.35992289e-02 2.77145624e-01
1.57104135e-01 9.25004482e-01 5.32344162e-01 -4.73113209e-01
1.02313590e+00 8.45315233e-02 -8.07053626e-01 1.04693651e-01
5.40697157e-01 -6.41188443e-01 2.04870865e-01 7.60196103e-03
-9.45654869e-01 4.74770330e-02 -4.05350298e-01 5.68729460e-01
-9.56703126e-01 -1.07532740e+00 3.81033599e-01 4.60532904e-01
-1.10644007e+00 2.88595408e-01 -8.55174661e-01 -1.05914211e+00
1.90678462e-01 4.34740223e-02 1.86883777e-01 1.45737320e-01
-1.50966895e+00 1.42857349e+00 6.42234534e-02 -1.03021776e-02
-1.86788335e-01 -9.24897254e-01 -4.71379697e-01 1.64366066e-02
1.42325938e-01 -6.19109094e-01 6.12933397e-01 -3.60892653e-01
-4.41727042e-01 5.04782021e-01 -2.71541085e-02 -4.47015733e-01
3.83833289e-01 2.68593699e-01 -7.43855774e-01 -7.92268838e-04
9.75908190e-02 1.04826823e-01 -2.89661497e-01 -1.12528205e+00
-7.83690929e-01 -5.43196380e-01 -3.29238147e-01 1.68497562e-01
-1.65320143e-01 1.00719869e+00 4.22846496e-01 -8.10001791e-01
-5.81191540e-01 -5.83870471e-01 -5.99645376e-01 -5.72020292e-01
3.28366190e-01 -2.46212929e-02 7.30642736e-01 -7.92776108e-01
1.25430942e+00 -1.74257636e+00 -1.14767939e-01 2.23545507e-01
1.39520407e-01 3.07700813e-01 2.20445871e-01 1.32732785e+00
3.63298357e-01 4.54877645e-01 -3.16145062e-01 1.97148547e-02
-2.15908706e-01 2.45704755e-01 -5.53504765e-01 5.76091588e-01
-4.80059721e-02 5.51786721e-01 -1.18527400e+00 -1.30084246e-01
3.67262363e-01 8.91504824e-01 -4.05673206e-01 -4.19359915e-02
1.81823179e-01 5.88700533e-01 -4.09327805e-01 2.20957235e-01
6.59469068e-01 -4.10959199e-02 4.85448271e-01 7.92606652e-01
-6.84109449e-01 7.94083532e-03 -4.53096688e-01 5.78418791e-01
-2.50105262e-01 3.40139121e-01 2.67176837e-01 -9.46355283e-01
5.48960865e-01 5.99701047e-01 8.30805659e-01 -8.01336467e-01
2.01947346e-01 1.18507653e-01 1.28705978e-01 -5.85584521e-01
2.36511856e-01 -3.97877067e-01 7.32659027e-02 8.88350189e-01
-3.81546855e-01 5.71295619e-02 1.62769854e-01 -5.75037077e-02
6.49741113e-01 -1.23628303e-01 4.92143720e-01 -6.30867064e-01
1.62712321e-01 3.30425113e-01 5.62747121e-01 5.79076886e-01
-3.99185717e-01 -1.37627959e-01 2.20064774e-01 -6.73213899e-01
-8.21796238e-01 -9.34941232e-01 -4.13434565e-01 6.83727980e-01
-3.81008014e-02 -1.34181321e-01 -4.63493884e-01 7.24862814e-02
8.71090144e-02 6.98463023e-01 -6.70026183e-01 9.62912887e-02
-6.42202735e-01 -1.44617534e+00 5.06582677e-01 1.82152718e-01
2.01318964e-01 -1.04574907e+00 -1.19979370e+00 3.94843578e-01
-1.36762232e-01 -8.42929244e-01 1.00391679e-01 -1.28440395e-01
-7.77713597e-01 -1.38280368e+00 -7.88634062e-01 -3.85061532e-01
6.07830465e-01 3.39340419e-02 9.02551293e-01 2.37586424e-01
-4.48772281e-01 7.26187408e-01 -1.58113778e-01 -9.63224828e-01
-5.67415237e-01 -2.82186061e-01 3.38191718e-01 -4.47115839e-01
5.07796168e-01 -4.49264705e-01 -6.36976957e-01 1.33341506e-01
-9.89666104e-01 -3.40272188e-01 1.81211308e-01 1.74732506e-01
-2.33332794e-02 1.95075765e-01 1.10772896e+00 -6.68980002e-01
8.14573944e-01 -1.26165891e+00 -4.85551506e-01 3.40590179e-01
-9.20273066e-01 -5.84302366e-01 2.55884647e-01 -9.35076401e-02
-1.14150786e+00 -6.56428814e-01 6.76445812e-02 5.01262903e-01
-4.32179183e-01 1.01481676e+00 5.48432291e-01 3.93004149e-01
3.53542328e-01 -1.26615763e-01 2.44969785e-01 -5.47029734e-01
-7.28620740e-04 7.16734946e-01 3.42615277e-01 -2.78386176e-01
7.26455510e-01 5.79727829e-01 -2.59298068e-02 -1.37315106e+00
-6.64868578e-02 -2.40277424e-01 -1.84055164e-01 -5.67148209e-01
7.61724055e-01 -5.13215482e-01 -4.75112796e-01 7.31467545e-01
-1.06900692e+00 -5.21625638e-01 -1.07763559e-01 7.47882187e-01
-2.13494420e-01 -6.75474927e-02 -3.56686652e-01 -1.31712592e+00
2.10681725e-02 -7.51533747e-01 5.50723337e-02 1.98793039e-01
-5.19769251e-01 -1.77406585e+00 9.58488524e-01 8.53962973e-02
1.03611779e+00 7.44336605e-01 9.26962674e-01 -4.07330602e-01
-5.35195656e-02 -2.89542936e-02 -9.16117504e-02 -3.82630408e-01
4.10389811e-01 2.13808298e-01 -5.32331288e-01 -1.83064416e-01
2.80374795e-01 8.51146877e-02 4.46677327e-01 6.15377784e-01
2.25279704e-01 -6.66184723e-01 -6.06346965e-01 3.10676217e-01
1.47059202e+00 7.28288054e-01 2.49799043e-01 4.87255067e-01
-9.70367193e-02 1.28206837e+00 2.96622247e-01 4.46819574e-01
7.65776217e-01 4.96479154e-01 5.60126677e-02 -3.31668645e-01
2.86686480e-01 -1.39441833e-01 1.32685229e-02 5.87689221e-01
-4.94756818e-01 -2.69870996e-01 -1.58062029e+00 7.06807673e-01
-1.69813824e+00 -1.28555894e+00 -1.78027347e-01 2.17546558e+00
5.44412374e-01 -1.77253798e-01 5.08653641e-01 -2.05056265e-01
5.94153464e-01 1.88256323e-01 -1.00606658e-01 -6.78415120e-01
-4.05610353e-02 2.74011749e-03 6.36617780e-01 7.32641578e-01
-5.48337758e-01 3.74062777e-01 8.57691479e+00 2.95057267e-01
-8.57240319e-01 -2.64370367e-02 6.89717770e-01 9.90432352e-02
-4.33628738e-01 -7.48491660e-03 -3.98667395e-01 4.25399423e-01
1.44507730e+00 -5.73948920e-01 5.91597021e-01 -7.44995326e-02
1.08110404e+00 -4.86081362e-01 -1.33061484e-01 2.02847198e-02
-1.82369500e-01 -1.21263158e+00 -1.99671388e-01 5.19045353e-01
5.56128800e-01 2.52324343e-01 1.64999142e-01 -7.54403025e-02
5.90934396e-01 -1.09482872e+00 2.94895291e-01 7.27498114e-01
5.20769775e-01 -8.95983994e-01 7.03421116e-01 5.19512534e-01
-9.99823272e-01 -7.35361874e-02 2.87434962e-02 -6.78496182e-01
6.86208665e-01 5.05969703e-01 -6.11618519e-01 3.01771820e-01
6.99326158e-01 4.47967261e-01 8.90410915e-02 8.74482512e-01
3.40971917e-01 5.59091985e-01 -4.92251009e-01 1.50693312e-01
1.35200679e-01 -3.59361172e-01 6.24151766e-01 1.24163294e+00
3.35089594e-01 3.81964862e-01 -1.16133943e-01 6.47517979e-01
4.86998022e-01 5.01037575e-02 -1.04763079e+00 -5.26778519e-01
5.58677375e-01 5.83420634e-01 -6.97202384e-01 -1.84732694e-02
-4.02863234e-01 1.25882685e-01 -9.00632963e-02 5.62186360e-01
-6.83188319e-01 -3.03917527e-01 9.67304289e-01 4.62352335e-01
-2.41853490e-01 -4.24832672e-01 -8.04889798e-02 -5.63882470e-01
-6.53610587e-01 -5.52924156e-01 4.51986939e-01 -1.47344366e-01
-8.33309650e-01 1.65585726e-01 6.75501525e-01 -1.40310079e-01
-4.76512402e-01 -1.82526693e-01 -8.09491158e-01 1.02720487e+00
-1.53103542e+00 -9.03677583e-01 3.66013050e-01 1.27035394e-01
-9.15634334e-02 3.21755022e-01 8.96635592e-01 3.58754277e-01
-5.83148241e-01 -4.25155759e-02 6.05384171e-01 -2.09517404e-01
2.77255416e-01 -6.66784406e-01 4.00789291e-01 6.11090183e-01
-6.41509473e-01 7.97047317e-01 7.54678905e-01 -1.04997814e+00
-8.43516707e-01 -8.05309057e-01 1.36739755e+00 -4.26165223e-01
7.78847694e-01 -1.18196823e-01 -5.45863330e-01 6.93753242e-01
3.85980040e-01 -9.93953049e-01 8.37936640e-01 -1.02738231e-01
8.40470567e-02 2.29224116e-01 -1.46877253e+00 6.33228481e-01
7.75370538e-01 -2.88825065e-01 -6.55812144e-01 1.53507769e-01
4.41821784e-01 3.96927208e-01 -9.52853680e-01 4.33282793e-01
7.16167092e-01 -7.04557717e-01 1.14566076e+00 -6.79215014e-01
-1.30957263e-02 5.98359704e-02 3.79807986e-02 -1.23775971e+00
-3.40780854e-01 -6.19601250e-01 2.15978175e-01 1.11529291e+00
2.51283467e-01 -1.27172780e+00 1.74502283e-01 6.49733961e-01
4.87133592e-01 -7.01510072e-01 -6.94795668e-01 -6.63481534e-01
4.71753389e-01 -2.53945827e-01 8.42067480e-01 1.10937858e+00
3.43486458e-01 -3.78095001e-01 -3.00236940e-01 8.22554678e-02
5.99814832e-01 -4.47053611e-01 5.04516661e-01 -1.20821357e+00
1.23011850e-01 -6.35157883e-01 -3.04275781e-01 -2.91894041e-02
-4.70951915e-01 -2.68965393e-01 -3.87780160e-01 -1.61271024e+00
3.00493568e-01 -6.07716739e-01 -2.54220814e-01 1.33270651e-01
6.04341403e-02 1.41248286e-01 2.18205616e-01 2.00102985e-01
2.98890769e-01 2.40869284e-01 6.65041924e-01 3.00306290e-01
-4.76830840e-01 -1.65242329e-01 -7.26213038e-01 8.12925041e-01
1.16912067e+00 -5.51882982e-01 -4.61536437e-01 -1.99623376e-01
6.49089634e-01 1.39426723e-01 3.99865061e-01 -7.67995298e-01
-2.59980536e-03 -9.93641257e-01 3.32008339e-02 -5.27696192e-01
5.37656769e-02 -8.74920845e-01 7.80424893e-01 1.06485116e+00
4.71243486e-02 5.78593075e-01 4.02226686e-01 4.46502656e-01
1.72000647e-01 8.67653489e-02 5.90297043e-01 3.17425206e-02
-3.88150327e-02 9.59772095e-02 -1.03458357e+00 5.33282936e-01
1.12815475e+00 -3.65428403e-02 -6.59313083e-01 -5.01782417e-01
-6.90147936e-01 4.30675656e-01 7.86448479e-01 1.35195389e-01
3.33968163e-01 -7.22829878e-01 -7.79753923e-01 -1.10419884e-01
-4.98269379e-01 -4.46043491e-01 2.97280103e-01 1.04569304e+00
-7.64571249e-01 8.43776047e-01 -4.11768556e-01 1.77909926e-01
-7.67016292e-01 8.55509400e-01 5.96140623e-01 -3.46466362e-01
-4.83441561e-01 1.44115791e-01 1.69450283e-01 -4.95666921e-01
1.20868105e-02 5.45723699e-02 -4.47467297e-01 2.49802068e-01
6.93021774e-01 1.07255793e+00 -5.97736657e-01 -9.25098062e-01
-6.69086099e-01 4.75878149e-01 4.59789515e-01 -3.74991953e-01
1.69689441e+00 -4.54453021e-01 -3.58253360e-01 4.18210357e-01
8.36583555e-01 5.32264933e-02 -7.37837851e-01 1.26954749e-01
1.58368662e-01 -1.79650784e-01 -1.83243051e-01 -9.75155473e-01
-7.03023434e-01 5.41650057e-01 3.40790480e-01 6.51533902e-01
9.57357287e-01 -2.22518206e-01 7.01040149e-01 -1.45026058e-01
1.24351561e-01 -8.93444538e-01 -5.59994936e-01 2.21007109e-01
9.07095432e-01 -5.47478139e-01 9.89023317e-03 3.16143073e-02
-7.21469596e-02 5.55311859e-01 -1.46713078e-01 1.14424840e-01
1.40366697e+00 3.09055477e-01 8.61668140e-02 -3.26860934e-01
-4.92043167e-01 -1.06255040e-01 -2.39424258e-01 1.11366785e+00
2.01848820e-01 1.92203864e-01 -8.19349706e-01 1.53159007e-01
1.93075284e-01 3.46676707e-01 4.69926983e-01 8.66633713e-01
-4.09459978e-01 -1.08860779e+00 -8.30177546e-01 3.76473516e-01
-5.65708280e-01 5.30147925e-03 -4.59246695e-01 1.08982813e+00
2.43442386e-01 1.19684935e+00 2.58560628e-01 -9.17670578e-02
-5.25785536e-02 2.29507256e-02 1.03328237e-02 -1.35024980e-01
-6.83440268e-01 -1.30933672e-02 2.00208891e-02 -9.53430086e-02
-5.62615395e-01 -1.06918061e+00 -1.21074426e+00 -1.11723757e+00
-1.55150797e-03 3.07348996e-01 5.77170134e-01 9.27426338e-01
3.99988115e-01 3.65262836e-01 4.57177579e-01 -6.94549918e-01
-2.64430970e-01 -8.04855704e-01 -6.52701139e-01 -9.37292203e-02
3.47769171e-01 -5.67183256e-01 -5.60954750e-01 -3.45664889e-01] | [5.960150241851807, 4.3710713386535645] |
c591906a-0916-4b7d-bd63-e72ddeff6598 | modeep-a-deep-learning-framework-using-motion | 1409.7963 | null | http://arxiv.org/abs/1409.7963v1 | http://arxiv.org/pdf/1409.7963v1.pdf | MoDeep: A Deep Learning Framework Using Motion Features for Human Pose Estimation | In this work, we propose a novel and efficient method for articulated human
pose estimation in videos using a convolutional network architecture, which
incorporates both color and motion features. We propose a new human body pose
dataset, FLIC-motion, that extends the FLIC dataset with additional motion
features. We apply our architecture to this dataset and report significantly
better performance than current state-of-the-art pose detection systems. | ['Yann Lecun', 'Jonathan Tompson', 'Arjun Jain', 'Christoph Bregler'] | 2014-09-28 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [-1.27163798e-01 -3.48051816e-01 -7.41912723e-02 -2.36183301e-01
-4.01905060e-01 -2.87958324e-01 3.29511493e-01 -7.15736866e-01
-9.12533462e-01 5.69922864e-01 5.45240700e-01 5.77229917e-01
4.85673308e-01 -2.04555273e-01 -5.40272593e-01 -1.12481326e-01
-4.16409463e-01 6.12154543e-01 4.83190179e-01 -4.19832468e-01
-3.60541731e-01 4.13969040e-01 -1.44880795e+00 1.39521763e-01
6.53498843e-02 8.56163383e-01 -4.00936484e-01 9.98929560e-01
8.18374753e-01 5.61748385e-01 -5.64967811e-01 -4.39196050e-01
3.59416127e-01 -5.25802314e-01 -9.55688477e-01 1.22209758e-01
1.06869507e+00 -8.26228797e-01 -8.79260004e-01 4.17415440e-01
9.76941943e-01 2.41199672e-01 3.61876339e-01 -1.04158723e+00
-1.31215736e-01 2.57683545e-01 -3.12014252e-01 2.84028649e-01
9.81002629e-01 2.93874502e-01 6.34397626e-01 -1.01197124e+00
9.01709199e-01 1.57628012e+00 1.16805458e+00 8.86191964e-01
-8.30824614e-01 -3.87807786e-01 -1.14575155e-01 2.06623942e-01
-1.38361752e+00 -4.43641007e-01 8.25905561e-01 -4.22478974e-01
8.95918667e-01 -3.87603156e-02 1.39044499e+00 1.41585982e+00
3.31254900e-01 1.27186310e+00 4.50213283e-01 -2.68103868e-01
-1.77018449e-01 -1.11762512e+00 -2.88140714e-01 1.18035316e+00
3.00839424e-01 1.86815366e-01 -7.83393502e-01 -5.51348180e-03
9.07609344e-01 -9.70853493e-02 -2.48207599e-01 -7.14510679e-01
-1.47459507e+00 6.16913378e-01 5.78912795e-01 5.37762940e-02
-3.30208451e-01 8.49990547e-01 6.31170690e-01 -1.63407251e-01
2.02948287e-01 2.16585204e-01 -4.06013995e-01 -3.65722150e-01
-1.20851076e+00 6.82903826e-01 8.42741966e-01 6.66882753e-01
4.09423597e-02 -1.09777525e-02 -2.81546235e-01 4.45178390e-01
5.33356071e-01 5.11273801e-01 3.60437304e-01 -1.23076141e+00
2.04667836e-01 2.97741681e-01 1.60225525e-01 -9.62856591e-01
-1.00128508e+00 -3.12960982e-01 -5.27804732e-01 -8.64683613e-02
5.03910363e-01 -2.47203663e-01 -1.06036472e+00 1.60916901e+00
5.41135430e-01 8.31967453e-04 -4.78165746e-01 1.44547081e+00
1.23384130e+00 -9.72426683e-03 3.44468802e-02 2.32688725e-01
1.33282495e+00 -1.33803594e+00 -6.89891756e-01 -2.91889518e-01
4.18373793e-01 -5.32036543e-01 6.53945446e-01 5.25105655e-01
-1.18389559e+00 -7.65471935e-01 -1.16772270e+00 -2.51590192e-01
1.03592850e-01 6.27801776e-01 9.19781923e-01 6.50847495e-01
-1.05453598e+00 7.41390526e-01 -1.20558298e+00 -5.65323055e-01
3.39417398e-01 5.88933349e-01 -6.29919410e-01 1.86872885e-01
-1.25265098e+00 7.69395709e-01 3.63067687e-01 5.32952964e-01
-9.07330930e-01 5.85968271e-02 -1.28400695e+00 -4.16394353e-01
3.99245888e-01 -1.23784018e+00 1.52737749e+00 -5.28849304e-01
-1.81895590e+00 8.82546484e-01 1.95471738e-02 -4.66847688e-01
9.57530975e-01 -1.10390711e+00 -4.14067060e-01 5.53409040e-01
-1.87719882e-01 9.80584979e-01 9.39862251e-01 -6.68657899e-01
-1.59299508e-01 -2.89309353e-01 -1.54398263e-01 1.01673022e-01
-1.19315848e-01 2.08217919e-01 -1.02202499e+00 -8.62588167e-01
4.21226323e-02 -1.41333115e+00 -2.13963762e-01 4.25338000e-01
-2.57476121e-01 -2.41779730e-01 7.93204725e-01 -7.67745912e-01
1.26423025e+00 -1.55269551e+00 6.90208733e-01 -3.00139505e-02
3.86261970e-01 3.81843358e-01 -3.57184745e-02 1.92353576e-01
2.58747786e-01 -4.65083152e-01 -4.99901995e-02 -4.83907968e-01
7.58746341e-02 1.89962849e-01 4.67222154e-01 9.56073940e-01
7.48682693e-02 1.22379994e+00 -8.02452803e-01 -8.08691084e-01
3.77463669e-01 6.15788877e-01 -9.22840416e-01 8.26015249e-02
4.39041704e-02 6.03490949e-01 -4.44776565e-03 9.49095011e-01
3.30811322e-01 -1.45215109e-01 3.23542118e-01 -4.98758048e-01
2.69059151e-01 -1.16398133e-01 -1.21157146e+00 2.36590958e+00
1.50213212e-01 5.41739166e-01 2.53369641e-02 -3.89110625e-01
4.49923366e-01 3.73917371e-01 9.06267762e-01 -1.12906583e-02
5.93085289e-01 -1.17036952e-02 7.12919980e-02 -3.59256893e-01
6.46143556e-01 6.07952625e-02 -3.06793690e-01 8.55368897e-02
5.34035087e-01 1.37377664e-01 3.08262378e-01 -9.31339115e-02
1.13626170e+00 8.30157280e-01 4.76799667e-01 -7.29832724e-02
5.88568330e-01 -2.98587561e-01 4.07969773e-01 4.47710842e-01
-9.58186626e-01 7.54566073e-01 -6.41509518e-02 -9.62591052e-01
-7.60848284e-01 -1.23621237e+00 3.15708846e-01 1.30112433e+00
6.50340095e-02 -7.44783282e-01 -6.72021091e-01 -7.75662363e-01
1.95804834e-01 -6.56854928e-01 -7.81131089e-01 8.22606459e-02
-1.12888825e+00 -1.14762060e-01 8.55312049e-01 1.24166894e+00
6.00968719e-01 -1.13108933e+00 -1.06361449e+00 2.29343832e-01
-3.28655332e-01 -1.18863225e+00 -8.65957379e-01 -3.83329034e-01
-7.15117633e-01 -1.28000879e+00 -9.67922866e-01 -7.44112313e-01
1.41425684e-01 -3.65768313e-01 1.28295732e+00 5.36988303e-02
-6.69527590e-01 6.84829473e-01 -2.56737888e-01 -1.50715902e-01
1.83405578e-01 1.92471400e-01 3.95939618e-01 -4.88937557e-01
2.78157502e-01 -7.56343603e-02 -1.04588580e+00 3.58100802e-01
-5.06767154e-01 -1.53102040e-01 4.25265253e-01 8.64534974e-01
3.34225684e-01 -7.92666435e-01 -6.60700724e-02 -1.58331066e-01
1.67824581e-01 1.94126263e-01 -2.15012759e-01 -1.23686358e-01
2.18238086e-01 3.09997499e-02 1.38356378e-02 -4.83666867e-01
-6.75003707e-01 8.76019478e-01 -5.51147699e-01 -5.56565166e-01
-1.56791881e-01 3.55167240e-01 -6.69707060e-02 -4.52786416e-01
4.78937387e-01 -2.02649817e-01 1.37318850e-01 -5.24698555e-01
5.91041028e-01 1.83024243e-01 1.27983451e+00 -4.31041956e-01
7.34481156e-01 6.99349046e-01 2.08456501e-01 -7.00615466e-01
-6.78645730e-01 -8.70628178e-01 -1.29737496e+00 -6.98305249e-01
1.13343143e+00 -1.29037881e+00 -9.63791549e-01 7.15717494e-01
-1.13110840e+00 -1.28169924e-01 1.86110526e-01 8.62095714e-01
-9.52756166e-01 6.42118394e-01 -1.07543099e+00 -7.14383304e-01
-4.98093665e-01 -8.57161164e-01 1.60018122e+00 -9.52397659e-02
-8.59544873e-01 -6.59031391e-01 4.71212894e-01 2.99052715e-01
2.88914964e-02 8.11835766e-01 -3.13156486e-01 -2.60630459e-01
-2.11641476e-01 -3.47635537e-01 2.57057101e-01 2.86996067e-02
-1.33306503e-01 -1.81622058e-01 -5.30006886e-01 -7.54508674e-01
-5.84131658e-01 -6.88228071e-01 1.24766982e+00 5.62646806e-01
9.79035914e-01 3.84975411e-02 -3.46781045e-01 8.89461040e-01
8.89520764e-01 -3.57847154e-01 6.91850305e-01 4.77178335e-01
1.06320143e+00 1.29506439e-01 5.34522653e-01 6.03600323e-01
4.16982174e-01 1.01138496e+00 2.45516643e-01 -2.18302473e-01
-2.52005339e-01 -2.52688706e-01 3.72153759e-01 8.24044883e-01
-7.87939966e-01 4.31959443e-02 -9.01471019e-01 4.99068111e-01
-1.91294789e+00 -1.09759831e+00 8.44582617e-02 1.57144237e+00
6.89440548e-01 -1.12033114e-01 8.48477900e-01 1.54675052e-01
5.61590314e-01 2.75928766e-01 -3.01653743e-01 3.11545637e-02
7.24708363e-02 5.44776082e-01 3.12408596e-01 -7.84040953e-04
-2.11661315e+00 1.03167868e+00 8.22823620e+00 1.27811551e-01
-8.01066637e-01 -1.30280405e-01 -2.29704157e-01 -4.86330181e-01
5.87935209e-01 -7.04132199e-01 -6.65724218e-01 2.23123878e-02
5.82010806e-01 4.03195351e-01 3.07106096e-02 1.04737651e+00
-1.93091661e-01 1.43900752e-01 -1.21265876e+00 1.30476785e+00
6.96594536e-01 -9.16810453e-01 -2.82614261e-01 -1.26040811e-02
5.22839785e-01 -5.82439974e-02 -1.60571709e-01 2.10825890e-01
1.74858019e-01 -1.02273428e+00 9.17108297e-01 6.13463163e-01
8.34499896e-01 -7.93563843e-01 7.66903460e-01 -2.59386182e-01
-1.78030193e+00 1.60633679e-02 7.31125707e-03 -2.72230864e-01
3.80674809e-01 -4.62733246e-02 -3.73660803e-01 4.26110804e-01
9.64014351e-01 1.03526998e+00 -8.61394763e-01 1.22590446e+00
-3.47283810e-01 2.98471719e-01 -2.83085197e-01 1.62422553e-01
-6.13152571e-02 5.30844092e-01 5.05206704e-01 1.29795253e+00
1.46069944e-01 -2.17829365e-02 7.81861961e-01 3.06726992e-01
-2.19980583e-01 -1.22477718e-01 -3.30552787e-01 -6.26935735e-02
3.06153204e-02 1.20475197e+00 -6.24956012e-01 -4.48250085e-01
-4.38537002e-01 1.34816968e+00 1.73593640e-01 -1.67386103e-02
-1.00612879e+00 -2.32162192e-01 6.44393146e-01 -1.21747740e-01
5.24013996e-01 -5.84293485e-01 3.93174320e-01 -1.42630875e+00
1.71758491e-03 -8.51048172e-01 6.91425681e-01 -6.16031051e-01
-1.27554607e+00 2.05767378e-01 2.73790713e-02 -1.34972870e+00
-6.83285773e-01 -9.68128324e-01 -1.78929955e-01 3.17597926e-01
-7.84938514e-01 -1.55844247e+00 -3.99849415e-01 7.53493130e-01
3.95855963e-01 2.15674583e-02 6.89798176e-01 2.87986338e-01
-2.69964278e-01 6.75924242e-01 -5.78157067e-01 8.07536066e-01
1.06634665e+00 -1.16887426e+00 8.54021966e-01 7.47493863e-01
1.41187593e-01 6.21054471e-01 5.98545969e-01 -8.09886634e-01
-1.69280505e+00 -8.90000999e-01 5.25676012e-01 -8.31629872e-01
3.42839271e-01 -3.54371130e-01 -2.26419777e-01 8.61812711e-01
7.20482022e-02 2.75115967e-01 5.96491933e-01 1.21444222e-02
-1.82157859e-01 4.71321374e-01 -9.62273777e-01 5.55838645e-01
1.67741323e+00 -3.68575722e-01 -1.03463840e+00 -1.30004629e-01
5.26594877e-01 -8.68910730e-01 -9.64461148e-01 8.77362788e-01
1.34621251e+00 -6.42604470e-01 1.34960806e+00 -7.06475496e-01
3.88530493e-01 -3.70177388e-01 6.70789927e-03 -1.08677936e+00
-6.50080442e-01 -4.81709123e-01 -7.75106966e-01 2.49663189e-01
-2.51959234e-01 1.63046286e-01 1.12730217e+00 1.09012879e-01
-1.14549071e-01 -5.40072858e-01 -9.66709018e-01 -8.26387525e-01
-3.20362262e-02 -1.92509145e-01 2.70704627e-01 4.66829151e-01
-3.52718383e-02 1.00356720e-01 -1.00387704e+00 -1.95757791e-01
6.77942574e-01 7.45462999e-02 1.12945485e+00 -1.31217694e+00
-2.07625791e-01 -1.78381413e-01 -1.00844026e+00 -1.05110395e+00
1.90439895e-01 -3.19884181e-01 3.39806944e-01 -1.46399462e+00
3.19710851e-01 9.12583351e-01 -1.82278976e-01 4.54925030e-01
-2.40956306e-01 9.73813891e-01 4.61647600e-01 5.08799329e-02
-1.27228343e+00 6.14396632e-01 1.27105415e+00 -1.98514849e-01
4.85959351e-02 -2.28538141e-01 1.31018639e-01 1.03658962e+00
3.86254162e-01 4.03557420e-02 1.99953616e-01 -1.87399954e-01
-3.85044217e-02 -2.19567884e-02 7.25105107e-01 -1.74027753e+00
1.83446914e-01 3.23608667e-01 1.37090421e+00 -1.20529830e+00
6.30185544e-01 -4.63700265e-01 3.92561555e-02 1.01555097e+00
-1.43718854e-01 2.81583905e-01 1.59616604e-01 5.47427833e-01
-9.36897770e-02 6.01509690e-01 6.30924225e-01 -4.29896384e-01
-1.20983922e+00 4.97860789e-01 -4.07761723e-01 2.24976406e-01
7.68827319e-01 -9.30954665e-02 -9.56871957e-02 -4.81098443e-01
-1.06226027e+00 1.60154849e-01 4.93822694e-01 7.26690531e-01
7.90775180e-01 -1.85936832e+00 -6.01092041e-01 2.24429891e-02
2.70905465e-01 -4.93311435e-01 4.21105772e-02 7.99740136e-01
-9.83915031e-01 4.87911135e-01 -5.66728771e-01 -8.03414404e-01
-1.54955006e+00 2.91055292e-01 3.84011090e-01 -5.05701043e-02
-7.04452395e-01 8.27227175e-01 -3.92959535e-01 -4.88963425e-01
2.81635433e-01 -3.35854590e-01 -1.22596622e-01 -1.99068323e-01
4.83831465e-01 6.02714598e-01 -2.31002674e-01 -1.20987511e+00
-8.94256234e-01 6.77561998e-01 4.65902567e-01 -2.08125979e-01
1.05703831e+00 1.77755393e-02 1.16622463e-01 2.99177229e-01
1.05785704e+00 -5.96974641e-02 -1.23711026e+00 1.68358609e-02
-7.54966680e-03 -5.26264071e-01 -3.38133842e-01 -7.22628593e-01
-1.03474569e+00 6.81562960e-01 8.99179757e-01 -6.53234720e-01
9.39060688e-01 1.78283140e-01 1.13547981e+00 4.92135584e-01
5.52800179e-01 -1.39348197e+00 6.13370895e-01 7.75859654e-01
9.57462847e-01 -1.20704758e+00 4.04201508e-01 -3.28024596e-01
-4.09576237e-01 1.15625155e+00 8.66285563e-01 -3.76673967e-01
5.40632904e-01 1.56707272e-01 1.90462321e-01 -4.14644301e-01
-4.45274413e-01 -6.39313102e-01 1.01882792e+00 5.97181618e-01
9.09720004e-01 1.50567427e-01 -6.48959935e-01 5.77437699e-01
-3.60126436e-01 4.49295461e-01 5.43385595e-02 1.17436385e+00
-3.87493610e-01 -8.52847517e-01 -6.11810207e-01 1.65086407e-02
-6.23825133e-01 3.45453113e-01 -9.03459489e-01 1.07608700e+00
3.15351725e-01 5.96849859e-01 -1.07043386e-01 -6.75081730e-01
4.07987416e-01 1.57766417e-01 1.08380556e+00 -4.39090639e-01
-5.23734570e-01 3.90088350e-01 4.40549344e-01 -1.27541590e+00
-9.80136693e-01 -5.76548338e-01 -1.10714447e+00 -1.79357558e-01
6.52462943e-04 -5.26407599e-01 2.21538872e-01 8.98651600e-01
5.32592684e-02 5.80906451e-01 -2.26368561e-01 -1.67033350e+00
-3.60247672e-01 -1.07964003e+00 -5.04853904e-01 6.56291306e-01
2.29237095e-01 -1.18468201e+00 1.70985579e-01 8.97486880e-02] | [7.028767108917236, -0.8089232444763184] |
e57154f8-1381-416f-b116-69febf1bdf07 | progressive-and-aligned-pose-attention | 2103.11622 | null | https://arxiv.org/abs/2103.11622v1 | https://arxiv.org/pdf/2103.11622v1.pdf | Progressive and Aligned Pose Attention Transfer for Person Image Generation | This paper proposes a new generative adversarial network for pose transfer, i.e., transferring the pose of a given person to a target pose. We design a progressive generator which comprises a sequence of transfer blocks. Each block performs an intermediate transfer step by modeling the relationship between the condition and the target poses with attention mechanism. Two types of blocks are introduced, namely Pose-Attentional Transfer Block (PATB) and Aligned Pose-Attentional Transfer Bloc ~(APATB). Compared with previous works, our model generates more photorealistic person images that retain better appearance consistency and shape consistency compared with input images. We verify the efficacy of the model on the Market-1501 and DeepFashion datasets, using quantitative and qualitative measures. Furthermore, we show that our method can be used for data augmentation for the person re-identification task, alleviating the issue of data insufficiency. Code and pretrained models are available at https://github.com/tengteng95/Pose-Transfer.git. | ['Xiang Bai', 'Wenqing Cheng', 'Baoguang Shi', 'Mengde Xu', 'Tengteng Huang', 'Zhen Zhu'] | 2021-03-22 | null | null | null | null | ['pose-transfer'] | ['computer-vision'] | [-5.03930449e-02 -4.21579108e-02 2.85958320e-01 -4.57671970e-01
-4.21936214e-01 -5.84564507e-01 6.13407433e-01 -5.14145792e-01
-2.73910642e-01 7.39685476e-01 2.77861953e-01 2.57208377e-01
3.49872291e-01 -8.06519866e-01 -8.14098358e-01 -5.71650803e-01
2.46373862e-01 5.14552534e-01 4.50440273e-02 -2.51621604e-01
-6.74634799e-02 4.16295916e-01 -1.22773528e+00 1.66579440e-01
6.56080782e-01 6.94206953e-01 -1.23844475e-01 6.00991130e-01
3.95075828e-01 2.52692848e-01 -8.35676014e-01 -9.62248743e-01
5.84148705e-01 -5.86238682e-01 -7.68500507e-01 1.74544305e-01
8.30284655e-01 -5.44124246e-01 -6.94795310e-01 1.12190640e+00
9.34394896e-01 1.59446865e-01 6.49726331e-01 -1.62237120e+00
-1.05423450e+00 1.16808631e-01 -8.47077847e-01 1.57805488e-01
6.55361056e-01 3.13783705e-01 4.51861590e-01 -9.18538034e-01
5.16002595e-01 1.57129693e+00 8.18371236e-01 9.48958576e-01
-1.24794877e+00 -1.04824507e+00 1.24021560e-01 2.53994405e-01
-1.47193396e+00 -3.99060607e-01 8.74723911e-01 -4.77525890e-01
3.45007509e-01 3.90561491e-01 7.43887544e-01 1.46455920e+00
7.78862387e-02 6.93841815e-01 1.07766068e+00 -2.91951090e-01
-3.30730230e-01 6.78782761e-02 -1.69968188e-01 5.90252817e-01
2.58296728e-01 3.09687227e-01 -4.02172625e-01 1.66914985e-01
9.98821974e-01 7.08839148e-02 -3.91289204e-01 -3.60192180e-01
-9.94918764e-01 5.02112627e-01 7.91701496e-01 -6.74815774e-02
-2.91442335e-01 1.20364673e-01 2.12078616e-01 2.12359563e-01
4.67035741e-01 1.69135645e-01 -1.79362539e-02 3.38358045e-01
-4.94249284e-01 4.28814590e-01 2.69007176e-01 1.25874043e+00
6.00083768e-01 2.09552757e-02 -5.38807034e-01 8.19699526e-01
2.85263896e-01 7.80329406e-01 3.47803861e-01 -6.49270535e-01
5.02360225e-01 5.37642837e-01 1.81363136e-01 -1.06705248e+00
-3.99418734e-02 -3.35112512e-01 -1.01725185e+00 1.45323426e-01
2.96358973e-01 -3.90916884e-01 -1.01130295e+00 1.92753744e+00
4.27595496e-01 1.72351107e-01 -2.33538255e-01 1.05759430e+00
1.03629398e+00 5.78946471e-01 2.08024740e-01 2.64073074e-01
1.44326448e+00 -1.05284595e+00 -6.50992155e-01 -2.96933353e-01
-4.93770093e-02 -7.19611287e-01 1.00561929e+00 3.05717066e-02
-1.13178921e+00 -1.08182442e+00 -8.08128059e-01 -1.49721816e-01
-3.45928639e-01 4.54493135e-01 3.81180763e-01 6.95763826e-01
-1.20512199e+00 3.61417860e-01 -5.01702428e-01 -5.44875503e-01
4.69122350e-01 4.76426661e-01 -5.26613414e-01 1.66869745e-01
-1.22967589e+00 6.76965356e-01 2.82896012e-01 2.81026185e-01
-9.06079173e-01 -6.26755416e-01 -8.18200767e-01 -1.83613256e-01
-6.08916096e-02 -8.68488133e-01 9.95416701e-01 -1.21081257e+00
-1.30025744e+00 1.11628222e+00 2.86572129e-02 -2.00541377e-01
8.54233563e-01 -4.84189183e-01 -3.41223836e-01 -1.33420164e-02
1.03175625e-01 1.24148214e+00 1.10391164e+00 -1.46505356e+00
-3.95218045e-01 -4.02333796e-01 1.35934755e-01 3.92212719e-01
-4.34464455e-01 8.34934115e-02 -9.00639236e-01 -1.00983906e+00
-4.08328593e-01 -1.14815116e+00 7.45191053e-02 5.05543128e-02
-6.52434170e-01 -3.85518447e-02 7.73895144e-01 -1.11477947e+00
9.77730572e-01 -2.10688376e+00 4.05850589e-01 1.11824527e-01
1.13202527e-01 4.83678818e-01 -3.05227339e-01 4.21247870e-01
-5.41122019e-01 7.15082418e-03 8.13669711e-03 -7.49515176e-01
-3.49633805e-02 -2.48512879e-01 -2.47645259e-01 4.46455538e-01
1.87022299e-01 1.32758117e+00 -5.73124588e-01 -3.94530445e-01
3.12494040e-01 6.56278372e-01 -4.51235145e-01 4.69540358e-01
1.93258241e-01 7.45447457e-01 -2.25323796e-01 4.66381997e-01
9.03459609e-01 2.13201847e-02 -2.33972490e-01 -4.39918160e-01
2.43934155e-01 -1.33105457e-01 -1.09608805e+00 1.47336280e+00
-5.01533337e-02 4.16857094e-01 -1.02021679e-01 -4.28354740e-01
9.62517500e-01 2.85555959e-01 3.11278403e-01 -6.11308098e-01
3.54443163e-01 -3.39347810e-01 -1.24905676e-01 -1.66795611e-01
4.42228079e-01 2.69237943e-02 -7.04373345e-02 3.17024350e-01
1.01318814e-01 2.18753725e-01 1.72342584e-02 7.30036050e-02
5.85823953e-01 2.98367262e-01 3.05204485e-02 -2.02658564e-01
6.92959309e-01 -3.09479058e-01 4.56735283e-01 3.00976694e-01
-4.31386024e-01 8.74328613e-01 1.13409594e-01 -5.47644079e-01
-1.28463876e+00 -1.13815284e+00 1.69103667e-01 1.04751778e+00
3.55373085e-01 -2.76856780e-01 -1.11256540e+00 -6.02913737e-01
1.00425772e-01 3.13507706e-01 -1.01513934e+00 -3.49239826e-01
-6.08780980e-01 -6.17992818e-01 5.78814030e-01 7.72219300e-01
1.00338793e+00 -1.27854228e+00 -1.23197980e-01 -3.62347990e-01
-2.99151808e-01 -8.49752188e-01 -1.13649642e+00 -7.02852130e-01
-4.53089923e-01 -9.68766928e-01 -1.13255966e+00 -1.00226486e+00
1.16570997e+00 1.91140011e-01 1.01136053e+00 9.55994576e-02
-2.21923381e-01 5.75566888e-01 -3.10649842e-01 -4.30937886e-01
-2.49703243e-01 9.70775262e-02 2.06591994e-01 3.20789725e-01
2.83860832e-01 -3.86691630e-01 -1.09648919e+00 6.06322289e-01
-5.89237213e-01 4.09691244e-01 5.40355742e-01 7.23008513e-01
4.62984353e-01 -2.41740987e-01 4.37744349e-01 -7.15340614e-01
7.33787060e-01 -2.37113647e-02 -3.45020771e-01 2.59542853e-01
-2.94326603e-01 -3.42609197e-01 4.21508849e-01 -5.82525909e-01
-1.16774035e+00 2.36530289e-01 -7.67284110e-02 -6.15807414e-01
-2.37028986e-01 -2.61247575e-01 -4.84995335e-01 -1.32759139e-01
6.47806227e-01 2.79274136e-01 1.73891142e-01 -4.23862994e-01
2.85039455e-01 6.24947071e-01 8.37849557e-01 -6.42793894e-01
1.38424897e+00 4.03673589e-01 -3.34894329e-01 -3.77077013e-01
-5.20628989e-01 -2.09269777e-01 -8.38123858e-01 -4.75059479e-01
1.01603079e+00 -9.68085289e-01 -7.39794612e-01 9.21649218e-01
-1.08873510e+00 -3.12605143e-01 -2.63189673e-01 5.20948023e-02
-3.73404562e-01 2.65663862e-01 -7.64644444e-01 -4.90828097e-01
-6.28913343e-01 -9.02685106e-01 1.20086265e+00 4.63503033e-01
-1.54080287e-01 -8.14308345e-01 2.14437947e-01 5.06260574e-01
1.92986295e-01 3.93015176e-01 2.82159239e-01 -2.82037467e-01
-4.52639371e-01 -4.77261543e-01 -2.64838904e-01 3.06427449e-01
2.76413947e-01 -2.47303322e-01 -9.42368627e-01 -9.23026145e-01
-3.43441218e-01 -3.35720688e-01 5.52396894e-01 1.49870485e-01
1.14358079e+00 -3.40932280e-01 -4.06489223e-01 7.46669114e-01
1.00738108e+00 2.11973652e-01 1.01151645e+00 2.36436054e-01
1.00643706e+00 5.20365000e-01 4.04997647e-01 3.27739492e-02
3.26988518e-01 9.43460286e-01 1.42807379e-01 -5.52152634e-01
-5.06423473e-01 -5.80604732e-01 2.76867062e-01 3.90544415e-01
-5.05472243e-01 -2.51951993e-01 -7.70892441e-01 3.15252513e-01
-1.63325524e+00 -1.15480328e+00 1.38222471e-01 2.24722338e+00
7.65707314e-01 -1.41422659e-01 5.05544066e-01 -5.69493882e-02
1.18042541e+00 -1.42180651e-01 -5.33534467e-01 9.29297432e-02
8.22222531e-02 9.62975807e-03 3.31829339e-01 5.37609220e-01
-1.17923999e+00 9.84826922e-01 5.91170454e+00 6.02713227e-01
-9.19416964e-01 2.34347563e-02 8.14269125e-01 5.99438511e-02
-6.62872419e-02 -4.94607210e-01 -8.85164857e-01 6.99247360e-01
4.04534101e-01 -1.20660149e-01 3.64411384e-01 5.04126489e-01
4.48774500e-03 3.30557883e-01 -1.06885719e+00 1.13386846e+00
3.77689302e-01 -8.00836146e-01 3.69617254e-01 1.39235467e-01
6.72026575e-01 -5.19277573e-01 4.19568568e-01 3.80683690e-01
1.67923480e-01 -1.05805147e+00 7.33648956e-01 5.85222125e-01
1.13387060e+00 -8.30333531e-01 7.21245527e-01 -9.20084119e-02
-1.30627966e+00 8.51791129e-02 -2.01987982e-01 2.41933346e-01
7.35922083e-02 -1.67961061e-01 -6.20062351e-01 7.43198097e-01
8.76533806e-01 6.73018336e-01 -1.00146782e+00 1.00107265e+00
-2.98720628e-01 1.34868175e-01 4.82540689e-02 3.34404200e-01
-3.62206966e-01 -2.41058573e-01 5.35478115e-01 1.01538479e+00
2.31474668e-01 1.04816832e-01 6.90943468e-03 9.18094456e-01
-1.93546802e-01 -1.06785484e-01 -4.27529395e-01 3.33939552e-01
4.33642864e-01 1.18275774e+00 -4.11352187e-01 -1.82767540e-01
-2.17498630e-01 1.48632586e+00 3.08948159e-01 4.08884048e-01
-1.11422229e+00 -2.33711436e-01 5.38417459e-01 3.49839807e-01
1.36033401e-01 3.68257910e-02 -1.01004876e-01 -1.13385499e+00
7.22984225e-02 -1.06561029e+00 3.59608293e-01 -9.97401834e-01
-1.52530909e+00 8.16341281e-01 2.04902321e-01 -1.36927378e+00
-1.01828789e-02 -5.66847086e-01 -6.57720625e-01 1.17436659e+00
-9.23522532e-01 -1.65268755e+00 -7.68091023e-01 7.94372976e-01
4.73554552e-01 -3.15749973e-01 7.24948406e-01 5.15273452e-01
-7.99322128e-01 1.05155993e+00 -4.19603080e-01 6.24630153e-01
9.94460404e-01 -1.25679719e+00 8.22989643e-01 1.07170308e+00
-1.40095636e-01 7.54200995e-01 6.22029066e-01 -7.99110174e-01
-9.83375251e-01 -1.42194271e+00 4.65966135e-01 -7.04536498e-01
9.35158730e-02 -6.44273877e-01 -6.19021475e-01 1.00773418e+00
4.85067755e-01 -1.66523173e-01 5.26947677e-01 -6.98061287e-02
-2.90631890e-01 -2.90118128e-01 -1.21286333e+00 8.18319917e-01
1.32843220e+00 -3.02566469e-01 -5.09610534e-01 2.21617192e-01
5.10763645e-01 -7.95511663e-01 -6.85518384e-01 3.27493817e-01
5.52418232e-01 -8.12897086e-01 1.34821856e+00 -5.03918409e-01
3.60383451e-01 -4.04500514e-01 2.61247426e-01 -1.43387663e+00
-6.63429976e-01 -6.22096062e-01 -1.32493032e-02 1.47358739e+00
1.27610460e-01 -6.18817449e-01 8.12165320e-01 5.75646520e-01
5.07774949e-02 -4.71691996e-01 -6.30105257e-01 -7.58033574e-01
1.49820536e-01 2.43600532e-01 8.14293504e-01 7.56143272e-01
-4.42587763e-01 4.94360387e-01 -7.87862122e-01 2.31617242e-01
7.08301127e-01 -6.21220544e-02 1.13315773e+00 -9.06050563e-01
-2.98574328e-01 -2.16937453e-01 -4.97150123e-01 -9.49280083e-01
8.97499733e-03 -7.54330456e-01 -1.50934532e-01 -1.41321743e+00
4.69047725e-01 -2.94127405e-01 -1.56922176e-01 6.50421143e-01
-4.15667146e-01 7.50896871e-01 4.55629677e-01 3.33453268e-01
-3.64577293e-01 7.07801282e-01 1.62866879e+00 -2.35327989e-01
-1.03733912e-01 7.48569146e-02 -6.56627297e-01 4.32617754e-01
9.46468830e-01 -3.97762917e-02 -5.73961020e-01 -4.57345754e-01
-1.29097775e-01 -3.17748725e-01 7.91070163e-01 -1.19569051e+00
4.53272909e-02 1.33962914e-01 1.00639391e+00 -5.11259913e-01
5.63216388e-01 -7.13358939e-01 5.61872780e-01 6.21278584e-01
-2.51600295e-01 3.80184084e-01 4.47098225e-01 4.33109075e-01
8.03223401e-02 1.58763781e-01 8.40968132e-01 9.49646458e-02
-4.90439892e-01 7.03657925e-01 2.74239808e-01 -8.44578370e-02
1.25360119e+00 -2.65567243e-01 -4.72287446e-01 -6.08695447e-01
-7.49759793e-01 2.63487101e-01 5.87660611e-01 7.79513597e-01
6.47970200e-01 -1.88383877e+00 -8.98308694e-01 3.64382625e-01
1.34757653e-01 -1.09032400e-01 4.09250438e-01 4.34379011e-01
-5.18777907e-01 2.27241829e-01 -7.20928907e-01 -3.79953504e-01
-1.46859181e+00 7.06086516e-01 6.78147137e-01 2.28336547e-02
-5.94190657e-01 8.98991525e-01 6.41888320e-01 -4.58141595e-01
2.51375377e-01 1.93530947e-01 -1.73974946e-01 -2.26646096e-01
5.71108699e-01 3.16489190e-01 -3.27064395e-01 -1.02108777e+00
-3.59786779e-01 5.91323376e-01 -1.94128424e-01 -1.06239885e-01
1.07031643e+00 -1.13301471e-01 5.40047735e-02 -7.72679970e-02
8.61580849e-01 -1.01656072e-01 -1.53542626e+00 -2.79391967e-02
-7.55463898e-01 -6.99390292e-01 -5.03682852e-01 -8.69387567e-01
-1.31479156e+00 6.32746637e-01 9.58573759e-01 -8.24239850e-02
1.12930620e+00 -6.86998963e-02 7.93402195e-01 -2.04983607e-01
3.83377939e-01 -7.54738033e-01 4.05108273e-01 7.47624040e-02
1.45260143e+00 -1.02921283e+00 -1.43669710e-01 -6.12494290e-01
-6.33924961e-01 5.42432785e-01 1.14631760e+00 -1.89065322e-01
3.58386964e-01 -7.67010301e-02 7.96571448e-02 -7.11251944e-02
-1.72662750e-01 -1.28463715e-01 6.47605419e-01 1.00073540e+00
2.77030498e-01 1.56572416e-01 -1.65113471e-02 4.26749468e-01
-6.00171626e-01 -2.52213567e-01 -5.54805920e-02 4.66646671e-01
9.69021842e-02 -1.20670497e+00 -6.65427923e-01 9.96092632e-02
-3.25008333e-02 4.14586440e-02 -7.48886883e-01 8.55685830e-01
4.34391111e-01 6.63116872e-01 2.42940351e-01 -7.23766625e-01
5.71719587e-01 -1.13532200e-01 6.48829162e-01 -5.53572297e-01
-6.15286052e-01 -5.85783832e-02 -1.76481888e-01 -3.64989996e-01
-2.53704309e-01 -7.06893444e-01 -6.45856380e-01 -5.85147679e-01
6.48373691e-03 1.45062879e-02 3.78509648e-02 4.14625347e-01
4.73512709e-01 6.02524519e-01 6.06975973e-01 -1.06947887e+00
-2.35449404e-01 -1.17852366e+00 -1.83190063e-01 9.90534365e-01
5.98110631e-02 -7.00840831e-01 2.59551331e-02 4.50967222e-01] | [12.036809921264648, -0.800792932510376] |
89c8f10c-93fc-4d16-99b4-51e3210562fe | physics-constrained-backdoor-attacks-on-power | 2211.04445 | null | https://arxiv.org/abs/2211.04445v1 | https://arxiv.org/pdf/2211.04445v1.pdf | Physics-Constrained Backdoor Attacks on Power System Fault Localization | The advances in deep learning (DL) techniques have the potential to deliver transformative technological breakthroughs to numerous complex tasks in modern power systems that suffer from increasing uncertainty and nonlinearity. However, the vulnerability of DL has yet to be thoroughly explored in power system tasks under various physical constraints. This work, for the first time, proposes a novel physics-constrained backdoor poisoning attack, which embeds the undetectable attack signal into the learned model and only performs the attack when it encounters the corresponding signal. The paper illustrates the proposed attack on the real-time fault line localization application. Furthermore, the simulation results on the 68-bus power system demonstrate that DL-based fault line localization methods are not robust to our proposed attack, indicating that backdoor poisoning attacks pose real threats to DL implementations in power systems. The proposed attack pipeline can be easily generalized to other power system tasks. | ['Zuyi Li', 'Ren Wang', 'Jianing Bai'] | 2022-11-07 | null | null | null | null | ['fault-localization'] | ['computer-code'] | [-1.59115165e-01 -2.11351782e-01 -2.05939129e-01 1.70557454e-01
-5.42676032e-01 -9.23798978e-01 5.32516479e-01 1.11665703e-01
4.28169996e-01 7.78757870e-01 -8.01084518e-01 -7.35870421e-01
-4.17723864e-01 -7.55578518e-01 -8.12958837e-01 -1.08462870e+00
-8.65204632e-01 5.80390096e-02 2.07741871e-01 -1.70492753e-01
3.26353669e-01 9.86274779e-01 -3.77211481e-01 -5.46500459e-02
5.18103838e-01 8.02689910e-01 -4.96557713e-01 5.10106981e-01
6.94557488e-01 5.64779401e-01 -1.19769704e+00 2.05543324e-01
4.28151280e-01 -9.80157033e-02 -5.01468241e-01 -4.24153864e-01
2.66763180e-01 -6.17366850e-01 -9.69760716e-01 1.53201210e+00
5.16592145e-01 -3.83516312e-01 3.41919482e-01 -1.90095985e+00
-2.35890374e-01 8.19841385e-01 -7.97116876e-01 6.88186824e-01
2.89350539e-01 3.50161552e-01 6.78323269e-01 -5.60036302e-01
-1.88537166e-01 1.09370768e+00 6.05129242e-01 2.19728500e-01
-9.90945280e-01 -1.28984869e+00 -2.76892364e-01 2.17277378e-01
-1.47314644e+00 1.52928606e-01 1.09232259e+00 -1.03175640e-01
1.13902044e+00 7.25154132e-02 3.90294880e-01 1.16070187e+00
1.41504443e+00 5.36031783e-01 1.03434467e+00 -1.88237354e-01
4.81198370e-01 -1.67717323e-01 2.72176564e-01 8.24109256e-01
6.82372570e-01 5.63126683e-01 -4.35007066e-01 -7.12628841e-01
5.82411170e-01 -4.18775380e-02 -5.34917891e-01 -1.26777232e-01
-6.51348591e-01 7.13373959e-01 8.20809126e-01 4.49533701e-01
2.31067985e-01 6.48820877e-01 4.43480700e-01 4.67812687e-01
-1.20664239e-01 8.07558954e-01 -6.35736287e-01 2.41419747e-01
-9.46196318e-01 1.87288463e-01 8.32366943e-01 7.53410459e-01
3.00909191e-01 9.63692784e-01 1.67542219e-01 -2.54513651e-01
3.39735925e-01 6.38156712e-01 2.74599101e-02 -6.57574460e-02
4.88326579e-01 5.20219430e-02 -7.24512190e-02 -1.06098044e+00
-1.03435552e+00 -9.10282969e-01 -7.84493089e-01 4.17805135e-01
2.01058835e-01 -7.13564038e-01 -4.70276713e-01 1.32051110e+00
5.78833045e-03 7.23484874e-01 1.43444046e-01 6.34621978e-01
2.18661979e-01 8.04327369e-01 -3.08272779e-01 3.39484960e-02
1.13917661e+00 -5.29685080e-01 -8.05776894e-01 -1.50400158e-02
5.92500687e-01 -3.31543416e-01 5.91704667e-01 8.28277171e-01
-5.74622452e-01 -3.34730059e-01 -2.23294401e+00 7.89278030e-01
-3.07011068e-01 -5.74590228e-02 8.12493742e-01 1.29587662e+00
-5.45660079e-01 6.36951208e-01 -1.06630552e+00 4.03064825e-02
4.10369784e-01 6.41206026e-01 8.66290461e-03 1.74603313e-01
-1.65156472e+00 7.95151293e-01 3.22859168e-01 4.02407497e-01
-1.77911139e+00 -8.24268997e-01 -6.15047395e-01 3.80592227e-01
6.30056918e-01 -6.92117736e-02 8.74160349e-01 -5.92493005e-02
-1.52716148e+00 -1.66991383e-01 7.03824580e-01 -8.53459239e-01
6.41675368e-02 -3.11653823e-01 -8.97594154e-01 1.45410091e-01
-3.34840447e-01 -3.49136949e-01 1.04569459e+00 -1.17599356e+00
-4.25590038e-01 -2.07033083e-01 2.53602087e-01 -4.11137611e-01
-5.60682952e-01 -2.55377501e-01 6.23914480e-01 -4.34287935e-01
-1.28620327e-01 -8.74902666e-01 -1.20581970e-01 -3.76429617e-01
-8.58805656e-01 -1.29739583e-01 1.42457569e+00 -4.77935784e-02
1.04568064e+00 -2.07820749e+00 -3.24167222e-01 5.59702814e-01
9.66513604e-02 3.64103228e-01 2.71713316e-01 9.87317741e-01
-3.13347787e-01 -6.51907325e-02 -6.70115575e-02 3.08378547e-01
2.83836156e-01 -1.84281711e-02 -1.12460864e+00 1.24805892e+00
2.77107120e-01 8.03157270e-01 -7.93039262e-01 2.99770713e-01
2.57041872e-01 6.29179627e-02 -2.82515883e-01 -6.05378300e-02
-7.62769878e-02 2.91875273e-01 -7.18445122e-01 5.46908021e-01
8.52889419e-01 2.22842887e-01 1.55474827e-01 -3.98677528e-01
1.72169641e-01 -7.76431859e-02 -8.77980232e-01 1.13064253e+00
-2.27525368e-01 7.76000977e-01 -3.02273240e-02 -1.10770702e+00
9.53924537e-01 3.77186179e-01 4.36020583e-01 -4.02100652e-01
3.67321163e-01 7.61277303e-02 5.27655244e-01 -6.74847513e-02
3.78683768e-03 -6.82897270e-02 -7.76239693e-01 5.16135693e-01
3.68920296e-01 -3.41459423e-01 -3.61949056e-01 2.38509104e-01
1.58949780e+00 -3.49808604e-01 4.44721356e-02 -8.92731130e-01
7.09413290e-01 -4.80818935e-02 6.99608564e-01 9.56106484e-01
-9.89614949e-02 -3.44339460e-01 7.39994705e-01 -3.03584844e-01
-5.50866008e-01 -1.51810563e+00 -4.43741620e-01 -9.46803614e-02
3.35723519e-01 -3.93497884e-01 -5.37006378e-01 -1.13177407e+00
2.67922610e-01 8.54737461e-01 -7.25542232e-02 -9.11188424e-01
-4.08673197e-01 -9.79254782e-01 1.39875424e+00 5.84913671e-01
5.95997691e-01 -4.67681527e-01 -4.56785440e-01 2.01632977e-01
9.68173504e-01 -1.22447193e+00 -5.64541072e-02 7.11305737e-01
-5.53417265e-01 -1.26806402e+00 -2.34974828e-02 -6.21682286e-01
6.39196575e-01 -5.72927967e-02 5.92783630e-01 2.45638341e-01
-6.27755463e-01 1.56924605e-01 2.46134941e-02 -2.07916632e-01
-3.85875940e-01 -1.80435181e-01 6.70130432e-01 -3.36260587e-01
2.17623010e-01 -4.29900169e-01 -1.52607873e-01 2.07864925e-01
-7.79327929e-01 -9.12995160e-01 2.32126176e-01 9.76804197e-01
-2.59440511e-01 1.63801491e+00 8.49242985e-01 -6.47248209e-01
8.05840731e-01 -5.58450341e-01 -1.27086210e+00 -8.84141997e-02
-4.37089175e-01 -3.25096585e-02 1.25342691e+00 -3.81653100e-01
-6.88737810e-01 -2.13495657e-01 -1.72203481e-01 -4.16640490e-01
6.16587847e-02 5.14322937e-01 -6.46989465e-01 -6.76526487e-01
3.83696556e-01 -2.39042062e-02 -7.36200988e-01 -1.77787170e-01
9.09705162e-02 1.88256145e-01 7.89659321e-01 -6.99136198e-01
1.73279858e+00 3.90252262e-01 7.62918591e-01 -9.56691623e-01
-4.32483345e-01 3.75705659e-01 -3.96394998e-01 -6.09067902e-02
2.96970516e-01 -7.86704779e-01 -1.09794974e+00 6.96374536e-01
-9.11001861e-01 -1.86691985e-01 2.80553848e-01 3.53289396e-01
-2.02568799e-01 6.03439867e-01 -9.44516778e-01 -6.08022213e-01
-2.06554398e-01 -1.37543535e+00 5.72238088e-01 3.03051054e-01
1.23650447e-01 -1.02022064e+00 -1.67909727e-01 -2.81034946e-01
1.39562950e-01 3.62347096e-01 1.44009387e+00 -9.50013757e-01
-7.35721946e-01 -6.54617429e-01 3.78131926e-01 3.72899860e-01
1.86924443e-01 8.11761171e-02 -1.00913346e+00 -1.08583510e+00
6.27824306e-01 -3.09340268e-01 1.99769795e-01 7.06999898e-02
9.70556915e-01 -1.33340240e-01 -5.91314375e-01 5.64546227e-01
1.48309708e+00 6.40596032e-01 3.31822217e-01 1.53300494e-01
6.87149286e-01 -1.87273905e-01 4.23891962e-01 5.17845213e-01
-4.99823511e-01 4.28817481e-01 6.86846733e-01 -1.61690652e-01
4.17303205e-01 -1.70538098e-01 6.23753607e-01 4.47195053e-01
1.00521064e+00 -5.22642910e-01 -8.88745308e-01 -1.37058590e-02
-1.36809742e+00 -6.29712760e-01 -1.71628892e-01 1.95893681e+00
3.58565658e-01 9.65058565e-01 -4.67075109e-01 7.35764146e-01
4.21071738e-01 3.11990902e-02 -7.61093855e-01 -3.02930325e-01
6.12404458e-02 3.70175272e-01 7.81413555e-01 5.11204422e-01
-1.16738963e+00 8.08853209e-01 6.40860081e+00 1.12818241e+00
-1.34753978e+00 -2.01625060e-02 2.26417765e-01 2.67123133e-01
1.40314698e-01 2.35738963e-01 -8.79753172e-01 2.70603776e-01
9.39776540e-01 -4.80899125e-01 1.92081854e-02 6.23739719e-01
1.09750889e-01 -1.32866204e-01 -1.49503732e+00 5.74199021e-01
-8.93362425e-03 -1.00618076e+00 -2.89434701e-01 9.62867960e-02
6.98456466e-01 -2.79219419e-01 3.76010299e-01 2.94151843e-01
3.60152185e-01 -1.05057657e+00 2.93049157e-01 -1.08512528e-01
2.24143147e-01 -1.41343546e+00 8.95390809e-01 3.66082489e-01
-9.61121321e-01 -3.77654582e-01 -3.35220039e-01 -1.38958409e-01
4.41398136e-02 7.40361869e-01 -1.07160604e+00 9.25338686e-01
5.88698566e-01 4.96570796e-01 -4.00932610e-01 9.48604226e-01
-6.98984921e-01 1.05934989e+00 -4.68971640e-01 9.80580375e-02
4.33501244e-01 3.75312060e-01 7.16202557e-01 7.65189409e-01
1.22473076e-01 -4.90999579e-01 5.57431757e-01 1.00077486e+00
1.35156408e-01 -7.63187945e-01 -9.45355415e-01 -4.13193777e-02
7.43826091e-01 1.31937277e+00 -9.59325194e-01 2.63850182e-01
-3.24196011e-01 5.44079483e-01 -1.16643526e-01 2.55036682e-01
-1.16316247e+00 -4.71354634e-01 5.02082527e-01 -3.22730720e-01
-5.58634028e-02 -6.94073141e-01 -1.64746374e-01 -8.13573420e-01
-3.11722100e-01 -8.64654541e-01 2.88473278e-01 -4.30198014e-01
-1.43006623e+00 3.81111890e-01 2.36430049e-01 -1.13145852e+00
-2.24888674e-03 -7.94075668e-01 -9.22286570e-01 5.33806860e-01
-1.20892632e+00 -7.33061850e-01 4.53251421e-01 7.08729565e-01
3.00112575e-01 -6.57973707e-01 6.65951073e-01 1.81464300e-01
-9.39169943e-01 7.61681855e-01 1.94417283e-01 2.21415028e-01
2.90508419e-01 -1.08060086e+00 4.68119204e-01 1.14921200e+00
5.60788065e-02 4.80660588e-01 8.98320675e-01 -8.24564993e-01
-2.09499836e+00 -7.80542672e-01 -1.97838530e-01 -1.63084224e-01
1.34935319e+00 -8.70597124e-01 -7.71475673e-01 6.56431615e-01
5.42501152e-01 1.29686505e-01 4.29840475e-01 -3.66358697e-01
-1.52253851e-01 -1.92196697e-01 -1.55509329e+00 5.96746504e-01
1.08946040e-01 -6.83475614e-01 -5.91126561e-01 5.60480595e-01
5.88262081e-01 -3.35862219e-01 -9.90525007e-01 5.14278471e-01
-5.05264513e-02 -3.57069820e-01 9.34215784e-01 -2.61864930e-01
-2.32653499e-01 -7.13588536e-01 2.61971615e-02 -1.44664443e+00
-2.67152190e-01 -9.37791049e-01 -4.31100816e-01 1.21702456e+00
1.97003648e-01 -8.43472183e-01 5.74419379e-01 -5.74217215e-02
-1.60543874e-01 -5.84387898e-01 -1.39861226e+00 -1.09285235e+00
5.30257821e-01 -3.81432533e-01 7.35354900e-01 1.08566475e+00
4.17407215e-01 2.41197944e-01 -4.10632163e-01 1.47352803e+00
9.02419269e-01 -2.69321799e-01 3.71871561e-01 -8.16523254e-01
-5.36100388e-01 -1.91004902e-01 -5.34626007e-01 -5.70028007e-01
4.01777595e-01 -7.62511969e-01 -8.23182315e-02 -6.42128885e-01
-2.19749153e-01 -3.93197715e-01 -5.93786359e-01 5.16583383e-01
4.86547425e-02 6.15111478e-02 5.87104596e-02 -1.71539530e-01
-7.97982588e-02 6.58837914e-01 6.96272910e-01 -6.48378074e-01
4.18883443e-01 -6.53862581e-02 -8.29735585e-03 6.71469748e-01
1.02753079e+00 -8.05270195e-01 -5.80765367e-01 1.98841915e-02
1.33725718e-01 3.31531674e-01 2.50471532e-01 -1.58713710e+00
5.15548646e-01 1.46655023e-01 4.79897290e-01 -7.30468392e-01
1.42886892e-01 -1.39057350e+00 -1.48631170e-01 1.21207476e+00
2.59742439e-01 3.30534369e-01 9.09866512e-01 5.18882453e-01
1.23504192e-01 -3.67941886e-01 6.65155351e-01 4.10222769e-01
-4.99991030e-01 2.20544085e-01 -9.14722741e-01 -3.83014321e-01
1.30786407e+00 3.71480614e-01 -6.14002883e-01 -1.61990091e-01
-3.26006353e-01 5.21781981e-01 8.76973011e-03 6.27082169e-01
5.89906812e-01 -1.25750148e+00 -3.66247743e-01 6.44369423e-01
-4.49330956e-01 -4.63618785e-01 1.05127856e-01 4.23338085e-01
-3.51685941e-01 7.47299194e-01 -2.02889040e-01 -6.41907275e-01
-9.23825860e-01 8.45194101e-01 6.44171298e-01 -2.80284435e-01
-9.07808661e-01 4.39958662e-01 6.98110163e-02 -1.41368344e-01
2.26000622e-01 -2.06143826e-01 2.35617355e-01 -7.12461054e-01
2.34134093e-01 3.63982856e-01 1.68198749e-01 -3.56158130e-02
-6.09386146e-01 4.69274282e-01 -1.10075325e-01 2.36021578e-01
9.51376140e-01 3.18033546e-01 -1.51192665e-01 2.89772958e-01
8.53555739e-01 -7.24099576e-02 -1.13940430e+00 9.80334282e-02
2.63083071e-01 -2.93005556e-01 5.74595749e-01 -9.11337554e-01
-1.42653596e+00 8.82955253e-01 7.37593114e-01 4.70339179e-01
8.74118805e-01 -4.90107328e-01 6.84284568e-01 7.91908860e-01
9.98131633e-01 -5.83344400e-01 3.32757026e-01 4.03828382e-01
4.28845555e-01 -4.38825160e-01 3.05781156e-01 -2.16662556e-01
1.91131935e-01 1.34509623e+00 8.12519908e-01 -7.20185518e-01
1.04403377e+00 1.32286417e+00 -3.26322764e-01 -3.71348828e-01
-6.87921464e-01 8.50867927e-01 -4.32370938e-02 5.23934722e-01
-5.65729961e-02 -1.40757576e-01 2.99958372e-03 5.64746797e-01
-2.44570356e-02 -3.30625594e-01 9.67828214e-01 1.29297674e+00
-3.44213158e-01 -1.03055286e+00 -8.29605281e-01 1.96406528e-01
-5.83062410e-01 1.38236985e-01 -1.84913367e-01 9.56432939e-01
-9.06434283e-02 1.05607748e+00 -2.39227578e-01 -4.55365986e-01
2.50448465e-01 -2.85170704e-01 4.67435390e-01 -3.75092417e-01
-5.54544270e-01 -4.26733308e-02 -2.70377398e-01 -5.69577098e-01
5.57772756e-01 -3.09162885e-01 -1.59541607e+00 -2.12782159e-01
-6.78916633e-01 3.61326694e-01 3.69689494e-01 9.96189415e-01
-1.30697608e-01 1.07821202e+00 1.04136884e+00 -8.68025601e-01
-1.01771736e+00 -5.66053212e-01 -1.11778927e+00 -4.25168365e-01
6.29719555e-01 -1.14624262e+00 -7.70159364e-01 -1.00900650e+00] | [5.434837818145752, 7.427256107330322] |
fc39cece-c6c4-4cb4-86fd-7ca002e0d00b | human-gender-prediction-based-on-deep | 2205.09850 | null | https://arxiv.org/abs/2205.09850v3 | https://arxiv.org/pdf/2205.09850v3.pdf | Human Gender Prediction Based on Deep Transfer Learning from Panoramic Radiograph Images | Panoramic Dental Radiography (PDR) image processing is one of the most extensively used manual methods for gender determination in forensic medicine. With the assistance of the PDR images, a person's biological gender determination can be performed through analyzing skeletal structures expressing sexual dimorphism. Manual approaches require a wide range of mandibular parameter measurements in metric units. Besides being time-consuming, these methods also necessitate the employment of experienced professionals. In this context, deep learning models are widely utilized in the auto-analysis of radiological images nowadays, owing to their high processing speed, accuracy, and stability. In our study, a data set consisting of 24,000 dental panoramic images was prepared for binary classification, and the transfer learning method was used to accelerate the training and increase the performance of our proposed DenseNet121 deep learning model. With the transfer learning method, instead of starting the learning process from scratch, the existing patterns learned beforehand were used. Extensive comparisons were made using deep transfer learning (DTL) models VGG16, ResNet50, and EfficientNetB6 to assess the classification performance of the proposed model in PDR images. According to the findings of the comparative analysis, the proposed model outperformed the other approaches by achieving a success rate of 97.25% in gender classification. | ['I. Atas'] | 2022-05-19 | null | null | null | null | ['gender-prediction'] | ['computer-vision'] | [-9.47395631e-04 1.07666127e-01 4.76858318e-02 -4.47216660e-01
-3.42107832e-01 1.98469654e-01 3.44550163e-01 2.03900516e-01
-8.94422770e-01 5.42099774e-01 -1.92969754e-01 -9.06630978e-02
-1.15310125e-01 -1.15077448e+00 -3.71409029e-01 -8.83406103e-01
1.50186926e-01 7.96510518e-01 3.48705910e-02 -2.03100398e-01
4.05534655e-01 6.37673020e-01 -1.62948930e+00 -1.71112776e-01
7.38053441e-01 9.40899968e-01 2.61271149e-01 2.61196136e-01
-1.89167231e-01 5.16840279e-01 -5.12520313e-01 -8.32123518e-01
1.09007932e-01 -9.74390507e-02 -5.61838925e-01 -3.11183394e-03
1.73737928e-01 -5.07730007e-01 -3.46732318e-01 9.86281037e-01
8.56613815e-01 1.00909159e-01 6.53483033e-01 -7.11374521e-01
-5.63019097e-01 4.23475415e-01 -7.95747221e-01 3.30528975e-01
9.01846855e-04 3.58633883e-02 6.69859111e-01 -7.81882763e-01
5.31466484e-01 1.29069936e+00 5.23282826e-01 6.62634313e-01
-7.78272569e-01 -1.06701779e+00 -4.96204048e-01 7.20405817e-01
-1.25946057e+00 -1.70745984e-01 7.08545387e-01 -3.51099849e-01
4.21618372e-01 -1.59128025e-01 9.08966780e-01 8.05944502e-01
5.42509913e-01 4.61339265e-01 1.27240539e+00 -2.84227669e-01
1.46845043e-01 -2.13960484e-01 3.02136806e-03 6.70312345e-01
3.53165388e-01 1.29807589e-03 -1.96182132e-01 1.51673868e-01
6.77081943e-01 1.78194996e-02 1.18591957e-01 -3.43036950e-02
-4.65963572e-01 1.04648209e+00 5.33016205e-01 5.23535073e-01
-3.50613892e-01 -1.54871404e-01 5.52675068e-01 -8.48556161e-02
3.44385743e-01 5.00930361e-02 1.05182573e-01 -4.08733413e-02
-8.80582392e-01 -3.86109576e-02 1.80993736e-01 1.78742453e-01
5.30751884e-01 1.55013427e-01 2.29832277e-01 1.28271139e+00
4.39501733e-01 6.02036357e-01 9.60670829e-01 -6.69684291e-01
1.48839533e-01 6.91138387e-01 -7.11328030e-01 -1.14966190e+00
-3.44434202e-01 -1.91947669e-01 -1.03325772e+00 2.35560805e-01
5.36581039e-01 2.81412393e-01 -1.32244694e+00 1.34293461e+00
4.08147931e-01 -1.73833102e-01 -1.45800605e-01 7.03939140e-01
9.59159374e-01 4.86675233e-01 3.93908024e-01 -2.87644733e-02
1.38406634e+00 -3.73802006e-01 -5.15371799e-01 -3.68545204e-02
2.34911084e-01 -7.19856620e-01 8.35781455e-01 5.29929757e-01
-7.46028483e-01 -5.21411002e-01 -1.15776742e+00 -2.13751912e-01
-2.13455990e-01 2.28039354e-01 6.58165932e-01 8.15893471e-01
-7.82524049e-01 6.57377243e-01 -1.03596878e+00 -3.15873086e-01
7.55370319e-01 7.15263247e-01 -5.53095043e-01 -2.10309461e-01
-9.23233509e-01 9.42349374e-01 3.02104175e-01 3.61230582e-01
-6.93255961e-01 -4.85043794e-01 -7.38107383e-01 -5.21090440e-03
-1.99929714e-01 -4.06745076e-01 9.17010725e-01 -4.07010555e-01
-1.50306201e+00 1.30781984e+00 3.22676986e-01 -4.48908448e-01
6.58772767e-01 -5.71551025e-02 -2.47567818e-01 4.38630551e-01
1.46391526e-01 7.68465638e-01 6.03724360e-01 -9.57341731e-01
-4.23372000e-01 -1.00364065e+00 -1.78258598e-01 9.31011513e-03
-4.58250254e-01 -1.09482504e-01 -4.06896323e-01 -3.82112175e-01
3.75529408e-01 -9.08913076e-01 -1.35598049e-01 -1.99423432e-02
-1.76037341e-01 -3.75104934e-01 6.45166874e-01 -9.93961692e-01
7.69432187e-01 -2.32421231e+00 -1.11558780e-01 5.20635307e-01
3.05912256e-01 2.89714217e-01 1.62493780e-01 -3.85082364e-02
-1.16273137e-02 -2.73567438e-01 -4.16964501e-01 -3.05435538e-01
-3.93764585e-01 3.16116899e-01 2.52426535e-01 6.50257945e-01
-3.52584183e-01 5.00221968e-01 -5.90304315e-01 -1.13361478e+00
4.17274863e-01 6.16054058e-01 -4.30629313e-01 1.47182703e-01
2.30694175e-01 6.10633492e-01 -3.29371303e-01 7.96067178e-01
7.90094435e-01 4.05771956e-02 2.48410538e-01 -2.19431221e-01
3.31186444e-01 -3.08196157e-01 -8.11262965e-01 1.46567881e+00
-7.13671148e-01 4.21476483e-01 -3.18426825e-02 -1.06630921e+00
1.32949507e+00 2.08049640e-01 7.25722075e-01 -1.04444659e+00
5.70205986e-01 6.13529801e-01 4.19251174e-01 -5.89529991e-01
2.92541534e-01 -3.99143308e-01 1.50913015e-01 2.82891721e-01
2.22620770e-01 -2.12204069e-01 3.60176414e-01 -2.32838094e-01
5.23672760e-01 -2.58713633e-01 2.40723670e-01 -6.59720525e-02
8.42505693e-01 -3.94152492e-01 5.88148177e-01 9.19177011e-02
-6.29112646e-02 5.30039966e-01 1.99198708e-01 -6.65929317e-01
-9.57535028e-01 -1.00416112e+00 -6.27652049e-01 7.34223545e-01
-1.03670089e-02 1.01779498e-01 -6.62811339e-01 -2.46076405e-01
-9.39956084e-02 4.19610083e-01 -8.58219445e-01 -1.66016191e-01
-9.18965995e-01 -1.07520163e+00 6.23623550e-01 5.28926551e-01
8.82328272e-01 -1.32071888e+00 -5.83090901e-01 9.50067192e-02
-8.83583794e-04 -8.63798022e-01 6.89246729e-02 -1.07093938e-01
-1.21597970e+00 -1.36085558e+00 -1.11608124e+00 -8.59712541e-01
7.91654050e-01 -4.04380828e-01 5.18736601e-01 2.90728807e-01
-5.15772939e-01 7.50806630e-02 2.73251701e-02 -2.36039370e-01
-4.95878339e-01 3.20437878e-01 -1.23439834e-01 -1.93805154e-02
5.52311420e-01 -7.93226123e-01 -8.63264620e-01 1.23602152e-01
-9.15961146e-01 -2.40193054e-01 8.39460611e-01 8.48206401e-01
5.06211996e-01 -1.66004255e-01 5.96185863e-01 -8.44562769e-01
6.57408834e-01 -3.09002489e-01 -4.39904094e-01 2.85528544e-02
-7.91993260e-01 -8.58648643e-02 5.26692927e-01 -2.90592104e-01
-1.05205858e+00 -2.24689230e-01 -6.61246121e-01 -2.31695056e-01
3.93268019e-02 3.29552412e-01 -6.00933656e-02 -1.03623830e-01
4.44014192e-01 3.50721866e-01 5.03656566e-01 -4.57211167e-01
-4.75579165e-02 7.03626871e-01 7.69005239e-01 -6.39912128e-01
4.65491861e-01 4.76239443e-01 2.04648495e-01 -8.34584951e-01
-3.57545942e-01 -1.46698967e-01 -5.45384049e-01 -3.72555405e-01
1.20157421e+00 -6.20055914e-01 -8.32260668e-01 8.27261984e-01
-6.86784983e-01 1.01294644e-01 1.60752144e-02 6.65524662e-01
-3.14496398e-01 5.69107831e-01 -8.55171680e-01 -4.50166434e-01
-7.06124544e-01 -1.44307327e+00 7.88339555e-01 3.93526852e-01
-1.27282925e-02 -8.96786094e-01 1.03861012e-01 5.80769360e-01
3.44504476e-01 1.77119792e-01 1.23120439e+00 -6.59957588e-01
6.23469055e-02 -3.99209261e-01 -1.61705106e-01 4.70840305e-01
1.89483717e-01 -2.08108071e-02 -8.74109328e-01 -1.17691174e-01
1.47928506e-01 -3.27207327e-01 7.93950617e-01 3.66024673e-01
1.58473325e+00 8.99246559e-02 -1.18481539e-01 5.50618112e-01
1.29774964e+00 6.72118008e-01 9.23445463e-01 5.94367683e-01
7.43344963e-01 5.94966888e-01 4.25181568e-01 3.28930020e-01
5.72145343e-01 4.20484364e-01 4.81277555e-01 -6.54854029e-02
-1.99952304e-01 -1.36712313e-01 -1.85581595e-01 8.53389204e-01
-7.30058730e-01 3.17186177e-01 -1.04758966e+00 3.85008663e-01
-1.23703992e+00 -7.51556098e-01 1.80221736e-01 2.09784651e+00
7.47268677e-01 8.31756294e-02 4.01730835e-02 6.19055748e-01
8.96410584e-01 -1.89516786e-02 -5.44883251e-01 -3.81966591e-01
2.21387446e-01 7.37438798e-01 3.69332582e-01 8.17695707e-02
-8.70434403e-01 6.67298257e-01 5.48010397e+00 9.83503461e-01
-1.56444287e+00 1.83908604e-02 8.87814283e-01 6.51529953e-02
7.39257485e-02 -6.12683177e-01 -6.31492496e-01 5.33398509e-01
7.76792467e-01 7.45095685e-02 1.01456374e-01 9.32661116e-01
-4.12889048e-02 -2.66422123e-01 -8.20135653e-01 1.21095657e+00
5.70486374e-02 -1.01545215e+00 6.09777635e-03 2.51969188e-01
2.66440302e-01 -3.04980338e-01 3.02169919e-01 3.84147376e-01
-7.84194022e-02 -1.16806948e+00 2.03861177e-01 1.22054346e-01
7.37277627e-01 -9.67540324e-01 1.00634181e+00 -4.86584082e-02
-8.37759256e-01 -1.94993451e-01 -5.10543108e-01 2.45182812e-01
1.01914443e-01 5.23096144e-01 -9.00582194e-01 3.11807632e-01
8.93422425e-01 4.68002319e-01 -5.12510598e-01 8.88033807e-01
-3.48766893e-01 4.66152966e-01 -3.62570137e-01 3.73302624e-02
1.07302167e-01 -3.20065886e-01 4.16776650e-02 8.20175290e-01
4.15391922e-01 2.67546494e-02 -3.50553185e-01 5.32639980e-01
-1.73025072e-01 5.98517716e-01 -2.70179898e-01 1.97171703e-01
2.72941917e-01 1.31760418e+00 -1.04131961e+00 -2.79885352e-01
-9.46388543e-02 2.47865677e-01 -5.15408292e-02 -2.80127913e-01
-6.35703683e-01 -1.69370040e-01 1.71474561e-01 3.50048840e-01
1.80420443e-01 -4.07336652e-02 -4.59189862e-01 -8.21092486e-01
-1.73078001e-01 -8.65132153e-01 5.55755258e-01 -4.34530854e-01
-1.13142359e+00 5.40889800e-01 1.48927167e-01 -1.00744414e+00
-2.70679563e-01 -6.40107453e-01 -6.33626044e-01 4.36025590e-01
-1.05865967e+00 -1.11555099e+00 -3.96162897e-01 6.23162389e-01
3.70318711e-01 -2.99711615e-01 7.38131464e-01 7.73235857e-01
-5.95730484e-01 6.48223281e-01 -8.28874707e-02 3.23690027e-01
4.87912834e-01 -1.05543411e+00 -1.93628818e-01 4.58518445e-01
-1.59193471e-01 6.33275092e-01 4.07596380e-01 -5.44743538e-01
-8.35741162e-01 -5.93009949e-01 5.77642858e-01 3.88824850e-01
4.38596070e-01 -3.97376083e-02 -7.76453733e-01 3.71233284e-01
-1.18984088e-01 -8.77947286e-02 8.07906330e-01 -2.68985406e-02
-1.26818612e-01 -2.92883575e-01 -1.48145878e+00 3.14163923e-01
5.08314371e-01 -2.31263489e-01 -7.19367087e-01 5.05975559e-02
-2.18620412e-02 -5.20372510e-01 -1.21897745e+00 5.12862206e-01
8.69701385e-01 -8.49786162e-01 1.05075741e+00 4.75004129e-02
8.79006863e-01 9.96196419e-02 -2.73310076e-02 -8.63208950e-01
1.07670791e-01 4.91629362e-01 3.10477495e-01 1.05453205e+00
2.34142784e-02 -6.19541228e-01 9.39605296e-01 4.16550636e-01
-1.34572074e-01 -9.83067513e-01 -1.12967730e+00 -3.86901796e-01
4.08349901e-01 -1.79346815e-01 6.23154402e-01 6.10399783e-01
-4.25101399e-01 -2.41077524e-02 -1.19992800e-01 -1.80272013e-01
7.61547565e-01 1.25160098e-01 4.60460663e-01 -1.25485015e+00
-2.39136681e-01 -3.40354055e-01 -8.80979240e-01 -2.90742993e-01
3.57687138e-02 -9.64494705e-01 -2.34486535e-01 -1.47134018e+00
2.12135747e-01 -3.56223226e-01 -2.90260851e-01 4.34378475e-01
1.10792123e-01 7.48635530e-01 1.61101013e-01 3.95087034e-01
8.18618014e-02 5.26952326e-01 1.52792203e+00 -2.74538398e-01
-1.18058650e-02 2.36992091e-01 -3.96117836e-01 9.38302219e-01
8.83629441e-01 -5.55663109e-01 -3.41773391e-01 -2.92342275e-01
-5.90073951e-02 -1.83783844e-02 1.06280349e-01 -1.07082176e+00
1.08871579e-01 2.97005862e-01 6.44261360e-01 -6.47803605e-01
5.28519094e-01 -5.52634597e-01 3.32563408e-02 8.99593771e-01
3.76316980e-02 -7.72372484e-02 8.68779495e-02 9.04872194e-02
-4.07157362e-01 -3.50786626e-01 1.04019737e+00 -2.62922853e-01
-6.53267145e-01 4.06450272e-01 -3.14311534e-01 -2.94352829e-01
9.34609175e-01 -4.62277412e-01 -1.01228617e-01 -7.25530013e-02
-1.07313693e+00 -2.15363860e-01 2.40595713e-01 2.05174774e-01
6.74934447e-01 -1.02680004e+00 -4.93741632e-01 3.42169441e-02
-2.55827210e-03 1.10480338e-01 5.82655907e-01 8.39897037e-01
-9.75197434e-01 6.48024678e-02 -9.14271235e-01 -7.88377225e-01
-1.19107997e+00 1.56375766e-01 2.25417450e-01 -3.33254278e-01
-7.49871910e-01 5.31895995e-01 1.07952483e-01 -4.88423675e-01
1.52661027e-02 -9.82006490e-02 -7.43419111e-01 1.93133652e-01
1.78722277e-01 5.98768473e-01 4.07763898e-01 -7.19831824e-01
-3.62060964e-01 9.42131698e-01 -2.85466015e-01 3.96582633e-02
1.66534770e+00 2.24540442e-01 -2.04803586e-01 3.19832191e-02
1.20123541e+00 -1.02785036e-01 -6.87893689e-01 6.42245635e-02
-1.45941958e-01 -3.82127136e-01 7.61235086e-03 -3.54869485e-01
-1.65855002e+00 1.13621497e+00 1.09498715e+00 -3.99594456e-01
1.09892547e+00 -1.11576736e-01 8.99172246e-01 2.22052112e-01
5.24828076e-01 -1.11849320e+00 7.66705871e-02 -2.10408792e-02
5.64271629e-01 -1.20847845e+00 1.44654036e-01 -1.79259241e-01
-3.64739597e-01 1.38526523e+00 7.03844726e-01 -3.07409883e-01
7.13918686e-01 9.23337322e-03 2.14518264e-01 -3.89066219e-01
1.69144079e-01 2.57786363e-01 6.65597692e-02 5.45539081e-01
4.89148289e-01 -1.20637797e-01 -9.99265730e-01 4.73687112e-01
-7.35674620e-01 1.21011391e-01 3.36984426e-01 7.00310051e-01
-1.89566955e-01 -1.25533855e+00 -4.19548631e-01 5.75414062e-01
-7.99707949e-01 3.84459585e-01 2.15184927e-01 1.03706384e+00
2.72160947e-01 6.57478333e-01 5.58640957e-02 -2.96362191e-01
9.96412858e-02 -2.46801808e-01 6.33998632e-01 -4.80680525e-01
-4.02505517e-01 -1.33817673e-01 -2.34336674e-01 -2.29293093e-01
-4.84136671e-01 -5.43762088e-01 -1.32446826e+00 -4.14715499e-01
-1.01194188e-01 -1.05982712e-02 9.65987504e-01 1.02734017e+00
-2.11884215e-01 4.84374434e-01 4.94211823e-01 -6.90404892e-01
-5.67107856e-01 -1.16783071e+00 -6.57662749e-01 4.26989943e-01
-4.25124526e-01 -9.40700769e-01 -2.44826198e-01 -4.69517708e-02] | [14.344772338867188, -2.0359063148498535] |
306a75cd-5f50-4cd5-86ab-47828610bab6 | zeroeggs-zero-shot-example-based-gesture | 2209.07556 | null | https://arxiv.org/abs/2209.07556v2 | https://arxiv.org/pdf/2209.07556v2.pdf | ZeroEGGS: Zero-shot Example-based Gesture Generation from Speech | We present ZeroEGGS, a neural network framework for speech-driven gesture generation with zero-shot style control by example. This means style can be controlled via only a short example motion clip, even for motion styles unseen during training. Our model uses a Variational framework to learn a style embedding, making it easy to modify style through latent space manipulation or blending and scaling of style embeddings. The probabilistic nature of our framework further enables the generation of a variety of outputs given the same input, addressing the stochastic nature of gesture motion. In a series of experiments, we first demonstrate the flexibility and generalizability of our model to new speakers and styles. In a user study, we then show that our model outperforms previous state-of-the-art techniques in naturalness of motion, appropriateness for speech, and style portrayal. Finally, we release a high-quality dataset of full-body gesture motion including fingers, with speech, spanning across 19 different styles. | ['Marc-André Carbonneau', 'Nikolaus F. Troje', 'Daniel Holden', 'Ylva Ferstl', 'Saeed Ghorbani'] | 2022-09-15 | null | null | null | null | ['gesture-generation'] | ['robots'] | [ 3.45231712e-01 1.41438887e-01 -1.81358293e-01 -5.10225832e-01
-5.42480290e-01 -8.75468552e-01 1.00188255e+00 -9.15961027e-01
-3.03273708e-01 4.47041512e-01 8.21898580e-01 1.52183741e-01
1.84570327e-01 -6.02118194e-01 -5.83956182e-01 -5.66382766e-01
8.16509351e-02 4.41082537e-01 1.43010691e-01 -4.08911198e-01
5.44080138e-03 5.38617969e-01 -1.34361053e+00 3.98938835e-01
9.13795680e-02 4.63269114e-01 1.62099227e-01 1.25350237e+00
-1.25657767e-01 4.40186888e-01 -7.41113186e-01 -1.68862835e-01
1.51761323e-01 -4.75134909e-01 -6.87665522e-01 1.21311933e-01
5.87118089e-01 -8.77396166e-01 -6.21327400e-01 4.57285553e-01
7.64897048e-01 4.80979890e-01 7.98600078e-01 -1.09871328e+00
-9.90764976e-01 5.14752448e-01 -3.17130499e-02 -3.50949645e-01
6.25247419e-01 6.66370571e-01 1.07145274e+00 -8.13148081e-01
1.20559430e+00 1.82558000e+00 4.10814226e-01 1.55081940e+00
-1.51836896e+00 -4.09536511e-01 2.28326410e-01 -5.26481569e-01
-9.78869021e-01 -6.06883109e-01 7.20776498e-01 -3.11540931e-01
8.19751680e-01 4.11135793e-01 9.73784447e-01 2.11945844e+00
-2.26051331e-01 1.23015058e+00 7.61180401e-01 -4.87360507e-01
4.38185602e-01 -4.05103683e-01 -3.46791834e-01 4.65045154e-01
-4.07858044e-01 3.64924043e-01 -9.01901186e-01 -2.83439189e-01
1.37273824e+00 -2.66533226e-01 -7.85786882e-02 -4.90835756e-01
-1.47846663e+00 6.89950943e-01 1.43874511e-01 4.72711176e-02
-8.16287659e-03 7.19002724e-01 1.30160645e-01 1.27841562e-01
2.05020264e-01 4.29154217e-01 -3.16373020e-01 -7.73426235e-01
-1.05476010e+00 9.10202742e-01 1.00464761e+00 1.32982147e+00
5.22426702e-02 3.56099963e-01 -7.68937469e-01 7.44718492e-01
3.30262393e-01 8.66470337e-01 3.72387260e-01 -1.55621779e+00
2.59744465e-01 -1.09427525e-02 2.79866576e-01 -6.30448043e-01
-1.83448717e-01 2.74983436e-01 -5.55770159e-01 5.67951918e-01
7.08338499e-01 -4.00471300e-01 -1.29681957e+00 2.12243366e+00
2.91241612e-02 -1.52986750e-01 -2.03133017e-01 9.65378523e-01
6.53364182e-01 5.64521134e-01 3.00642580e-01 2.20512137e-01
1.20912385e+00 -8.67004097e-01 -7.47072518e-01 -1.37686551e-01
1.40412703e-01 -6.24689400e-01 1.68318617e+00 3.14391822e-01
-1.20306945e+00 -5.24822235e-01 -7.54996955e-01 -2.21924946e-01
3.58842760e-02 -5.17085828e-02 7.17666090e-01 6.62202835e-01
-1.21586633e+00 8.53418529e-01 -1.16008818e+00 -5.95974565e-01
3.52460027e-01 3.07135403e-01 -2.24141523e-01 3.59898895e-01
-9.54815030e-01 5.15390694e-01 -7.86875635e-02 -1.73282295e-01
-7.30404139e-01 -6.00845635e-01 -8.93183947e-01 -2.48176500e-01
8.34267437e-02 -1.09630919e+00 1.66953278e+00 -8.47854376e-01
-2.31236815e+00 6.96439087e-01 -3.61548096e-01 1.17629454e-01
8.81978035e-01 -5.18301487e-01 -1.14104122e-01 1.77915618e-01
-1.06028795e-01 1.35821986e+00 9.69927251e-01 -1.15665972e+00
-2.45184109e-01 2.87266690e-02 1.13347314e-01 3.25130880e-01
-2.53122002e-01 -5.66416197e-02 -6.04599953e-01 -1.26404774e+00
-1.48668289e-01 -1.28351319e+00 -1.65492803e-01 5.21885812e-01
-4.68573689e-01 -5.28462380e-02 9.36210275e-01 -3.28534633e-01
1.13080716e+00 -2.07559967e+00 7.48053312e-01 4.78302203e-02
-5.89554310e-02 -1.32193506e-01 -3.60964417e-01 3.84208888e-01
3.27047050e-01 4.62044477e-01 -2.81358570e-01 -7.32979774e-01
3.93001705e-01 5.33162892e-01 -3.85519147e-01 -7.11742043e-02
3.24047685e-01 1.22684276e+00 -9.88977194e-01 -4.49001610e-01
1.72377467e-01 4.55190271e-01 -9.58330214e-01 4.86882150e-01
-5.86033344e-01 9.07459199e-01 -4.37180817e-01 6.62629247e-01
1.03701107e-01 -5.14785275e-02 1.42011493e-01 -1.00187780e-02
3.85556929e-02 7.44158551e-02 -1.14684069e+00 2.47209024e+00
-3.93953532e-01 7.47269869e-01 -1.29835180e-03 1.63022950e-01
5.71379602e-01 3.94579262e-01 2.92817324e-01 4.16408991e-03
-2.06010323e-02 -7.94811267e-03 -2.93836355e-01 -6.82697177e-01
7.09816515e-01 -2.31703639e-01 -4.11306053e-01 8.32366765e-01
1.01848207e-01 -6.27139270e-01 -8.34993869e-02 1.20208450e-01
8.41383100e-01 7.87116587e-01 -1.51332691e-01 -7.44731426e-02
-2.94437885e-01 -1.30803972e-01 2.10053191e-01 9.99100268e-01
-1.64543852e-01 1.14627087e+00 3.96567583e-01 -3.66237551e-01
-1.32821512e+00 -1.40529835e+00 2.47308016e-01 1.38937318e+00
-9.86737534e-02 -8.49871710e-02 -8.40955198e-01 -4.56410736e-01
-1.54131958e-02 7.40509331e-01 -7.38797843e-01 8.68045241e-02
-1.04237592e+00 -2.91856945e-01 9.36448634e-01 7.60446429e-01
1.49873719e-01 -1.52917576e+00 -7.76457429e-01 1.01781681e-01
2.03305334e-02 -9.10841942e-01 -9.35541391e-01 -2.91040182e-01
-8.72867048e-01 -4.43009168e-01 -1.24543464e+00 -5.99501550e-01
2.86644816e-01 -3.26417506e-01 1.21164536e+00 -1.27944261e-01
-1.68191940e-01 5.87466955e-01 -2.47808486e-01 -3.37894596e-02
-6.44144118e-01 2.72348523e-01 1.46232933e-01 -2.76482254e-01
4.97642122e-02 -8.08061361e-01 -7.49016941e-01 8.76968447e-03
-1.03125787e+00 1.26214311e-01 2.76725173e-01 9.92503703e-01
1.92472920e-01 -1.05948770e+00 1.18686013e-01 -5.98361731e-01
9.13737774e-01 -1.00164942e-01 -2.22539846e-02 -5.91400592e-03
-2.60211620e-02 4.43855256e-01 5.19273616e-02 -9.43381310e-01
-1.14577866e+00 1.73331097e-01 -1.66543975e-01 -4.91682082e-01
-3.76298130e-01 -2.48237878e-01 -2.05194503e-01 2.06539840e-01
7.25087166e-01 1.23004034e-01 1.63514793e-01 -4.59592998e-01
1.26536500e+00 5.25905609e-01 7.40342975e-01 -8.76180589e-01
7.91291356e-01 6.61626339e-01 -1.51025012e-01 -9.31640804e-01
-1.91652715e-01 2.87577122e-01 -8.20456803e-01 -1.53558120e-01
1.14000130e+00 -5.89246988e-01 -7.85400867e-01 5.98752320e-01
-1.18184841e+00 -9.60752487e-01 -5.08291125e-01 1.01923853e-01
-9.85603929e-01 1.80316448e-01 -8.98510873e-01 -8.52129281e-01
-1.12091094e-01 -1.30700469e+00 1.60838962e+00 1.18752331e-01
-1.24390471e+00 -9.55638885e-01 5.53349815e-02 -2.31323600e-01
4.50176388e-01 4.97575939e-01 8.09832752e-01 -8.71776864e-02
-5.56292236e-01 2.22635433e-01 3.53255063e-01 -6.38465062e-02
3.68489891e-01 2.15208545e-01 -1.04420078e+00 -1.72597960e-01
-6.04793727e-01 -4.32405949e-01 7.00296402e-01 5.98395646e-01
9.16821063e-01 -2.71250308e-01 -1.92136437e-01 9.10204828e-01
7.63760448e-01 1.11038357e-01 5.27591467e-01 1.36174157e-01
1.03439116e+00 5.23978233e-01 2.54850984e-01 4.83428091e-01
1.23963110e-01 8.87106061e-01 -6.29774928e-02 8.66837129e-02
-3.10687751e-01 -5.46304703e-01 3.19505692e-01 3.49738449e-01
-3.26942831e-01 -2.96012968e-01 -4.20196772e-01 4.52803582e-01
-1.58675408e+00 -1.14866996e+00 3.80625248e-01 1.78099620e+00
1.04629517e+00 1.59043465e-02 5.98397374e-01 -2.57950008e-01
2.97035187e-01 5.44612527e-01 -6.50353312e-01 -6.51362360e-01
-1.22906022e-01 4.14368212e-01 2.60823131e-01 6.69623375e-01
-1.07959938e+00 1.32209349e+00 8.02105331e+00 5.34003735e-01
-1.08906102e+00 -1.65304095e-01 4.29300368e-01 -8.28902781e-01
-8.26928020e-01 -3.07389349e-01 -4.86280620e-01 3.39661419e-01
4.50185835e-01 3.08108777e-01 6.93138719e-01 6.57260835e-01
5.06388485e-01 3.94593775e-01 -1.40083790e+00 8.89539838e-01
-1.35057926e-01 -1.44885087e+00 2.81333089e-01 -1.22399352e-01
6.00089669e-01 -3.38188291e-01 3.24850023e-01 2.11639836e-01
5.81374407e-01 -1.11358941e+00 1.14912367e+00 6.00928545e-01
1.45301783e+00 -3.20248395e-01 -2.77029604e-01 5.55639006e-02
-8.88474345e-01 1.01753943e-01 2.17517987e-01 -6.81108907e-02
5.90269566e-01 -2.38291517e-01 -3.40341032e-01 -1.51688412e-01
4.87776548e-01 3.25124621e-01 -2.32886747e-01 1.53676242e-01
-3.58588964e-01 5.31869531e-01 -3.40232372e-01 -3.23006719e-01
1.28653228e-01 1.07676439e-01 9.96390700e-01 1.43344426e+00
3.49802196e-01 2.74532914e-01 4.41733338e-02 1.25160336e+00
1.60761282e-01 -3.66654932e-01 -7.43096530e-01 -2.00125009e-01
6.87858522e-01 7.90767491e-01 -4.45737094e-01 -4.27200854e-01
-1.59758441e-02 1.55695713e+00 1.14199929e-02 8.26817453e-01
-4.79735434e-01 -1.85536757e-01 1.18175411e+00 -1.79716393e-01
2.69286722e-01 -7.19736874e-01 -6.00764155e-01 -1.24053597e+00
-5.90358488e-02 -8.62750411e-01 -1.39797226e-01 -8.11641753e-01
-1.12777805e+00 5.26782751e-01 2.78373361e-01 -1.10901892e+00
-9.79100287e-01 -6.91430986e-01 -6.51264966e-01 9.86982226e-01
-6.06091440e-01 -1.35337973e+00 -1.45546004e-01 2.88896024e-01
9.57955062e-01 -2.14488462e-01 1.09390640e+00 -1.85476765e-01
-1.10009857e-01 7.87545383e-01 -2.79078722e-01 1.58715963e-01
7.94890046e-01 -1.43377590e+00 1.22228086e+00 4.89733577e-01
1.55076191e-01 8.36905122e-01 9.66028869e-01 -6.65450573e-01
-1.45406461e+00 -6.67590797e-01 5.02292991e-01 -8.93835008e-01
4.98044282e-01 -6.94003522e-01 -5.15988946e-01 6.55735493e-01
1.78603023e-01 -2.54982084e-01 4.54387933e-01 -1.78802274e-02
-2.30898634e-01 5.90293646e-01 -1.01387095e+00 1.37452328e+00
1.68190324e+00 -5.61457992e-01 -6.58073068e-01 -1.43924817e-01
9.92131472e-01 -7.65824556e-01 -7.23153293e-01 1.24978811e-01
1.31267548e+00 -7.66695559e-01 1.06039488e+00 -8.92164409e-01
7.08843231e-01 2.17051748e-02 -2.01512486e-01 -1.33485830e+00
-4.32588011e-01 -1.14789200e+00 -3.43561769e-01 1.09481347e+00
3.06901425e-01 -2.85476837e-02 1.04099929e+00 1.06102049e+00
1.55517757e-01 -6.24202132e-01 -6.25753999e-01 -6.10852122e-01
3.56637299e-01 -5.49022734e-01 7.75199354e-01 6.15674436e-01
-1.40586540e-01 -3.68716754e-02 -6.40512943e-01 -1.93331748e-01
5.10713458e-01 -1.83214284e-02 1.03292239e+00 -7.53714323e-01
-8.65474999e-01 -6.46896422e-01 -7.08060041e-02 -1.48021626e+00
8.28348622e-02 -5.75782061e-01 1.10345744e-01 -1.34175277e+00
-9.03680176e-02 -1.95555761e-01 2.71267772e-01 3.97584021e-01
-2.30590865e-01 2.66657799e-01 6.32985234e-01 3.33729506e-01
-5.05175665e-02 6.52102411e-01 1.72979593e+00 -1.17747791e-01
-6.70842230e-01 -5.61754853e-02 -4.83528793e-01 7.04535544e-01
4.98595864e-01 -2.65426099e-01 -5.24462819e-01 -7.77570963e-01
-1.45870596e-01 2.11808354e-01 4.12066877e-01 -6.56820118e-01
-2.66199976e-01 -5.44522345e-01 4.58710700e-01 -1.77693024e-01
6.93011343e-01 -2.23340228e-01 4.65946048e-02 2.81719774e-01
-8.13388288e-01 5.65604419e-02 1.24995284e-01 5.57627201e-01
3.35103124e-01 2.01982230e-01 4.15985525e-01 -1.97719574e-01
-7.60588169e-01 3.18877697e-01 -6.27633691e-01 1.40409738e-01
3.61101627e-01 -3.12063515e-01 1.21119767e-01 -8.05424094e-01
-1.19498670e+00 -7.71608725e-02 6.31093740e-01 8.96392524e-01
5.84219754e-01 -1.84457397e+00 -5.59207022e-01 3.14440012e-01
7.51256421e-02 2.38123164e-02 -3.95283811e-02 6.05913065e-02
-6.15622222e-01 -1.44337282e-01 -2.17923462e-01 -8.09696436e-01
-1.11665237e+00 7.07870498e-02 1.30511627e-01 2.51642585e-01
-8.48175704e-01 8.96266580e-01 1.13716260e-01 -4.56073701e-01
1.82322279e-01 -6.31386638e-01 2.39369065e-01 -3.75031382e-01
5.32699049e-01 1.73056602e-01 -7.34210610e-01 -4.57397997e-01
-1.36063993e-01 5.92323303e-01 4.18050826e-01 -1.18463552e+00
1.08216369e+00 -3.21585983e-02 3.26060176e-01 8.29854012e-01
9.11020458e-01 3.22979204e-02 -1.97416651e+00 2.03487873e-01
-1.64201647e-01 -4.33768392e-01 -5.28697968e-01 -8.17273319e-01
-5.87056816e-01 8.32341611e-01 5.40580988e-01 -2.15007290e-01
7.71799862e-01 -9.65020582e-02 9.60862994e-01 4.20784414e-01
2.95549303e-01 -1.04685187e+00 3.59234810e-01 6.23219013e-01
1.13766515e+00 -9.32265282e-01 -5.36885917e-01 -2.83754002e-02
-9.46185172e-01 1.18612075e+00 2.77800113e-01 -3.48262519e-01
4.87835020e-01 5.10478795e-01 4.60155696e-01 7.18684541e-03
-5.76209724e-01 1.59936979e-01 4.14722890e-01 8.04196417e-01
6.03890657e-01 3.31785858e-01 -2.31423005e-02 3.51001471e-01
-6.93284273e-01 1.88635021e-01 2.60107845e-01 1.02913439e+00
-1.06295764e-01 -1.41381896e+00 -3.11863422e-01 -2.36239582e-02
-1.82186574e-01 -2.76936404e-02 -5.76147318e-01 9.33549941e-01
-2.25303024e-01 7.93816507e-01 3.78331617e-02 -4.56283063e-01
3.96285355e-01 3.96323293e-01 8.81263852e-01 -5.98825634e-01
-2.83883095e-01 2.57368356e-01 1.90519124e-01 -8.09667408e-01
-2.55495042e-01 -7.41305709e-01 -1.14945507e+00 -1.83052734e-01
3.68247956e-01 -6.02596045e-01 3.50185603e-01 8.91225338e-01
3.15432787e-01 4.86159503e-01 2.99581349e-01 -1.65171456e+00
-6.64471924e-01 -1.09566975e+00 -4.61806029e-01 9.25602555e-01
4.26376760e-01 -6.19559288e-01 -2.08821774e-01 4.87839758e-01] | [5.653951644897461, -0.11354607343673706] |
d6bbaccd-f0cf-4557-9d27-e01b77b1c052 | sequence-to-point-learning-based-on | 2006.00250 | null | https://arxiv.org/abs/2006.00250v1 | https://arxiv.org/pdf/2006.00250v1.pdf | Sequence to Point Learning Based on Bidirectional Dilated Residual Network for Non Intrusive Load Monitoring | Non Intrusive Load Monitoring (NILM) or Energy Disaggregation (ED), seeks to save energy by decomposing corresponding appliances power reading from an aggregate power reading of the whole house. It is a single channel blind source separation problem (SCBSS) and difficult prediction problem because it is unidentifiable. Recent research shows that deep learning has become a growing popularity for NILM problem. The ability of neural networks to extract load features is closely related to its depth. However, deep neural network is difficult to train because of exploding gradient, vanishing gradient and network degradation. To solve these problems, we propose a sequence to point learning framework based on bidirectional (non-casual) dilated convolution for NILM. To be more convincing, we compare our method with the state of art method, Seq2point (Zhang) directly and compare with existing algorithms indirectly via two same datasets and metrics. Experiments based on REDD and UK-DALE data sets show that our proposed approach is far superior to existing approaches in all appliances. | ['Zhenrong Zhang', 'Ziyue Jia', 'Linfeng Yang', 'Hui Liu', 'Fannie Kong'] | 2020-05-30 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 8.47708061e-02 -2.97909707e-01 -3.28082502e-01 -2.42987871e-01
-7.37668693e-01 -6.28940701e-01 4.30948257e-01 -4.44178253e-01
2.20367327e-01 7.75061548e-01 4.38555211e-01 -3.30305636e-01
-2.58094907e-01 -5.71355283e-01 -6.15121961e-01 -9.18847442e-01
-6.27551377e-02 1.51593253e-01 -4.21052039e-01 -1.13102600e-01
-1.42631382e-01 4.26453084e-01 -1.24396217e+00 2.45987903e-02
9.07475114e-01 1.52170897e+00 2.66070038e-01 5.58883548e-01
1.72948036e-02 1.30172086e+00 -6.35826707e-01 -1.90391138e-01
3.76057118e-01 -2.83824176e-01 -9.78285372e-01 -2.09119186e-01
5.60389124e-02 -7.64592707e-01 -3.10163379e-01 1.23723483e+00
8.58406186e-01 -1.20711654e-01 6.82625711e-01 -1.94538987e+00
-9.65755880e-01 8.64763200e-01 -8.30761015e-01 3.39389443e-01
2.99682975e-01 -5.29086664e-02 9.43132520e-01 -2.72052020e-01
-3.30785900e-01 9.75135684e-01 8.54684114e-01 2.63453841e-01
-1.14607573e+00 -8.12682569e-01 5.73857091e-02 5.99183500e-01
-1.12885070e+00 -5.54761350e-01 1.05454206e+00 -1.02521136e-01
1.42048454e+00 6.21024966e-01 2.03808948e-01 1.29916954e+00
-4.57194783e-02 1.20985293e+00 1.04475009e+00 -2.92936921e-01
4.30103451e-01 -9.43473168e-03 -2.56460086e-02 4.04538751e-01
1.68134242e-01 4.97153923e-02 -3.17809403e-01 -7.02073500e-02
1.30987749e-01 1.31332502e-01 -4.60000843e-01 -2.03149170e-01
-9.26811934e-01 6.49610162e-01 2.99805731e-01 3.39195251e-01
-2.82496244e-01 1.37881890e-01 5.22327602e-01 2.65879422e-01
4.54523355e-01 -2.33343747e-02 -8.09628725e-01 -4.48774636e-01
-1.18698514e+00 -2.71175027e-01 1.09446597e+00 8.00293684e-01
3.97638828e-01 3.78280997e-01 -6.61046505e-02 7.07541227e-01
3.31334740e-01 7.00274348e-01 9.02515292e-01 -8.73153687e-01
5.28452933e-01 7.05278814e-01 1.36170357e-01 -5.55309057e-01
-7.74545670e-01 -1.76222980e-01 -1.58846879e+00 2.22191840e-01
8.91055688e-02 -3.46631557e-01 -4.75896686e-01 1.68583417e+00
-2.22605802e-02 5.26568472e-01 5.50488457e-02 8.01642299e-01
7.72781610e-01 9.21447217e-01 1.06835775e-02 -4.03467923e-01
1.04837739e+00 -1.09553695e+00 -9.89593387e-01 -6.39812127e-02
2.68000096e-01 -5.64462304e-01 8.02801490e-01 5.78053057e-01
-8.53594482e-01 -3.68539512e-01 -1.37207234e+00 -1.09864801e-01
-8.38515878e-01 2.75344819e-01 5.60359240e-01 6.73222542e-01
-9.19115603e-01 7.41875410e-01 -6.95661128e-01 -5.69877386e-01
6.84759557e-01 4.51552063e-01 5.45111559e-02 3.01868349e-01
-9.58134592e-01 9.31087911e-01 3.47902149e-01 1.35671780e-01
-8.41857195e-01 -6.66857481e-01 -7.57137001e-01 2.66583323e-01
1.39106199e-01 -3.76421720e-01 1.28948510e+00 -9.10934687e-01
-1.75796390e+00 3.82004559e-01 5.92206381e-02 -5.85666120e-01
4.90330100e-01 -4.03462738e-01 -7.82322049e-01 -2.72445977e-01
-2.19634436e-02 2.58723497e-01 1.04524720e+00 -1.08364487e+00
-8.91281724e-01 -4.88043487e-01 -9.53016356e-02 -1.45504490e-01
-5.16837418e-01 -3.35216731e-01 3.52591306e-01 -5.99831104e-01
-3.05260450e-01 -5.16477048e-01 2.13399440e-01 -5.24342239e-01
-7.47728109e-01 -5.87981999e-01 1.36512005e+00 -1.15660763e+00
1.22715878e+00 -2.03238845e+00 -1.35032743e-01 -9.02644470e-02
7.42123201e-02 3.23135972e-01 -1.46715775e-01 2.33504742e-01
-5.62374532e-01 -1.65344954e-01 -1.77703559e-01 -4.08674300e-01
7.00149477e-01 1.78736314e-01 -3.10099483e-01 6.85566545e-01
-1.37785703e-01 1.40808618e+00 -8.85851443e-01 -1.12501428e-01
6.76185727e-01 4.42627460e-01 3.15306962e-01 2.37348005e-01
4.24948670e-02 4.05096449e-02 1.20698489e-01 9.34055626e-01
8.37315202e-01 -4.60942328e-01 5.30789420e-02 -7.01319754e-01
5.42872101e-02 2.64689982e-01 -1.23147142e+00 1.73219740e+00
-6.98966146e-01 8.53977501e-01 1.01454251e-01 -1.55297410e+00
5.74892640e-01 4.97377455e-01 9.10548270e-01 -9.56612468e-01
3.10676426e-01 1.07078264e-02 -4.58828866e-01 -5.39571822e-01
3.24548371e-02 2.06485823e-01 1.26881879e-02 3.86526406e-01
1.72909677e-01 3.87601078e-01 4.75978255e-02 -2.15895504e-01
1.29865658e+00 -2.26908419e-02 4.18891072e-01 -4.41212922e-01
3.93802524e-01 -5.79903066e-01 5.36205709e-01 5.36318362e-01
-5.05947411e-01 1.23093642e-01 2.82544792e-01 -3.79859120e-01
-8.21730494e-01 -1.22213852e+00 3.73992883e-02 9.38809633e-01
6.09554797e-02 -8.10778886e-02 -7.26938367e-01 -8.43537629e-01
7.28398412e-02 1.02754712e+00 -2.59146839e-01 -2.26617366e-01
-3.55140328e-01 -1.15738225e+00 3.89423579e-01 8.46682966e-01
9.46915507e-01 -9.38684344e-01 -5.98096788e-01 3.43318582e-01
-3.80105376e-01 -1.09247816e+00 -3.96813750e-01 7.39787519e-01
-4.90790367e-01 -1.13188100e+00 -6.47092700e-01 -7.86464691e-01
1.56610668e-01 2.65225291e-01 1.52952719e+00 -6.78252280e-01
-3.63386661e-01 2.46508852e-01 -1.58500850e-01 -6.48769736e-01
-1.49167612e-01 3.28273237e-01 1.70368463e-01 -9.04103443e-02
1.07738853e+00 -1.14404917e+00 -6.23499870e-01 3.34085017e-01
-6.85150445e-01 -2.02667683e-01 5.56399465e-01 4.06246930e-01
1.71709612e-01 8.65744472e-01 8.10698271e-01 -2.73467928e-01
8.15188944e-01 -7.58714378e-01 -9.04552460e-01 2.30387464e-01
-1.12567592e+00 -1.10687152e-01 7.68314600e-01 -2.71954328e-01
-9.12699759e-01 1.37037873e-01 -6.55337423e-02 -4.84498829e-01
-3.48576128e-01 -3.59156579e-01 -5.48477411e-01 3.38204741e-01
1.50191948e-01 2.88194478e-01 -3.79688263e-01 -7.46668577e-01
4.50720519e-01 8.60129893e-01 7.51551867e-01 -2.50979692e-01
7.95472682e-01 4.35030997e-01 -1.25235617e-01 -6.41172528e-01
-6.60430849e-01 -5.16411185e-01 -4.56317902e-01 2.84505263e-02
9.98817325e-01 -1.00104105e+00 -1.27382576e+00 7.83647537e-01
-1.08555102e+00 -3.34470242e-01 -4.92676944e-01 9.90814418e-02
-4.24469590e-01 2.57546067e-01 -5.00506580e-01 -1.07675421e+00
-7.35772192e-01 -1.06414807e+00 1.32941186e+00 2.48558521e-01
-2.87439734e-01 -9.40719187e-01 -1.97440516e-02 3.15468758e-01
5.38074374e-01 4.44154680e-01 9.56101716e-01 -5.98776102e-01
-2.80496716e-01 -1.33438841e-01 -3.23872209e-01 8.53710473e-01
6.63211405e-01 -4.87063199e-01 -1.27725661e+00 -4.34512615e-01
3.70185792e-01 -6.90986216e-02 7.08013833e-01 5.32231033e-01
1.53007185e+00 -6.08653128e-01 -3.82369995e-01 7.02999115e-01
1.79639733e+00 2.64431387e-01 6.50129735e-01 1.06561288e-01
7.29062140e-01 2.99086757e-02 -3.80218923e-01 5.19298792e-01
5.78945220e-01 1.91672549e-01 5.37875891e-01 -3.37601602e-01
-5.38558513e-02 -2.26702183e-01 5.72508276e-01 8.64603460e-01
2.46581286e-01 -4.96508151e-01 -2.82921672e-01 5.59809268e-01
-2.13209033e+00 -9.94309366e-01 9.82622728e-02 1.69393277e+00
4.44177419e-01 2.80559883e-02 4.11402166e-01 7.59699166e-01
5.04562080e-01 3.01345944e-01 -1.09009969e+00 -2.95088798e-01
-2.55340397e-01 -4.43327352e-02 8.42991590e-01 1.37832716e-01
-1.20305848e+00 -4.21977788e-02 6.41985798e+00 1.06093919e+00
-7.96961248e-01 4.72968638e-01 7.08103299e-01 -3.44725102e-01
2.60536939e-01 -7.88637757e-01 -4.41661775e-01 1.04680145e+00
1.04934037e+00 1.16469875e-01 1.03948236e+00 9.64856684e-01
2.29990080e-01 -2.35474169e-01 -1.58241093e+00 1.70710039e+00
5.03956228e-02 -9.36175764e-01 -5.16366899e-01 -3.53699662e-02
8.95654082e-01 3.10129583e-01 -1.43817782e-01 2.65723556e-01
6.82197213e-01 -1.01626849e+00 3.86350751e-01 5.63858509e-01
4.15486783e-01 -7.55290270e-01 9.74622428e-01 3.58693391e-01
-1.45429206e+00 -4.56076682e-01 1.98063138e-03 -1.58903778e-01
9.58267972e-02 8.89138162e-01 -3.07832301e-01 5.04452586e-01
1.19928944e+00 7.11384833e-01 -3.08347553e-01 4.92001891e-01
-9.99955684e-02 6.53464675e-01 -7.07939804e-01 4.41696122e-02
4.21237620e-03 -1.69474199e-01 2.11760372e-01 1.01186121e+00
2.35739708e-01 -1.99845731e-01 4.80190385e-03 9.91447330e-01
-2.97539979e-01 -3.42141688e-01 -5.95876276e-01 7.66567513e-02
2.73483902e-01 1.50370288e+00 -5.15851617e-01 -2.96733499e-01
-7.45025218e-01 1.54423821e+00 -3.03395204e-02 4.85587239e-01
-9.09128845e-01 -4.30290997e-01 9.34675097e-01 -3.11499804e-01
3.02059114e-01 2.23454952e-01 -3.51148546e-01 -1.17198753e+00
1.92820415e-01 -6.33913636e-01 2.95253247e-01 -8.53243470e-01
-1.93606973e+00 5.34145646e-02 -1.45182282e-01 -9.30210710e-01
-2.15095520e-01 -4.99380141e-01 -8.04810107e-01 6.86017454e-01
-1.90652251e+00 -1.11645246e+00 -3.16096753e-01 8.17732871e-01
7.42908478e-01 -2.31361717e-01 7.70913184e-01 6.93661749e-01
-8.98931742e-01 5.50022125e-01 6.55693769e-01 2.73822814e-01
2.09217951e-01 -1.70627069e+00 3.19113225e-01 6.61298156e-01
-6.33673370e-02 -2.89224982e-01 5.20268798e-01 -1.37875482e-01
-1.47295642e+00 -1.27225065e+00 4.78343427e-01 -6.25939667e-01
5.50444603e-01 -4.35997307e-01 -4.39276397e-01 6.88164055e-01
7.29793429e-01 -3.53547454e-01 7.51906633e-01 -1.94289505e-01
1.47436718e-02 -7.04069257e-01 -1.43720496e+00 -1.63607877e-02
9.26420748e-01 -5.24896622e-01 -5.20874262e-01 4.37826186e-01
4.49412942e-01 1.10947847e-01 -8.31702292e-01 4.28180844e-01
2.80533373e-01 -1.09314394e+00 9.12295282e-01 -1.43648952e-01
1.37453392e-01 -3.54740590e-01 -6.59174383e-01 -1.45463347e+00
-4.31138396e-01 -9.87956703e-01 -1.06118584e+00 1.86808574e+00
5.60239293e-02 -7.39579737e-01 5.62010705e-01 4.76080686e-01
2.65825003e-01 -5.84118307e-01 -1.02247751e+00 -9.23871279e-01
-4.64820229e-02 -5.97923458e-01 1.38973176e+00 8.81383002e-01
5.29291667e-02 6.53556228e-01 -5.18301487e-01 5.40364563e-01
7.97971785e-01 1.68435350e-01 2.38992289e-01 -1.25531662e+00
-6.75987676e-02 -6.89709961e-01 -1.91055313e-01 -1.03155887e+00
3.39915097e-01 -7.63118744e-01 -7.91723803e-02 -1.58803856e+00
1.56943530e-01 1.47513092e-01 -7.88312614e-01 5.99963903e-01
1.77242935e-01 1.51428282e-01 3.75751182e-02 -1.41077135e-02
-6.64907038e-01 5.12510836e-01 4.41554278e-01 -6.86072886e-01
-1.94423106e-02 -1.00119606e-01 -6.96271479e-01 8.15871358e-01
1.17252135e+00 -1.18860956e-02 -5.86623132e-01 -4.16836739e-01
1.66678682e-01 -2.82293200e-01 4.40691024e-01 -1.09904456e+00
4.21051793e-02 3.22203100e-01 8.01306725e-01 -9.16868329e-01
8.56884792e-02 -1.29573750e+00 1.64070427e-01 3.45062762e-01
1.00569025e-01 1.72122613e-01 8.58188048e-02 4.02177811e-01
1.81904823e-01 -8.09367094e-03 5.28574407e-01 -7.79577419e-02
-9.65570748e-01 1.56044215e-01 -1.00399561e-01 -2.53368497e-01
9.30043161e-01 5.71441650e-02 -2.66020596e-01 -3.95214111e-01
-4.11044717e-01 4.87467170e-01 -1.27709419e-01 7.70504296e-01
1.85684115e-02 -1.88101578e+00 -4.24983025e-01 2.98935235e-01
-3.03395838e-01 -1.32439405e-01 -6.91438019e-02 6.70440793e-01
1.29378140e-01 6.19538367e-01 2.72636771e-01 -5.55601120e-01
-7.11329579e-01 9.92095828e-01 5.98793268e-01 -2.83682197e-01
-4.41506892e-01 4.15813088e-01 -1.34952478e-02 -2.30482876e-01
6.66786075e-01 -7.94918597e-01 6.64917976e-02 2.47393191e-01
6.41090274e-01 1.16155791e+00 3.55991662e-01 -4.02225226e-01
-3.34102780e-01 4.32422996e-01 4.39199597e-01 6.61842585e-01
1.51682520e+00 -4.62058514e-01 -2.82002687e-01 5.36141634e-01
1.66662431e+00 -6.08820617e-01 -1.07786036e+00 2.27059405e-02
-1.49065871e-02 1.54150337e-01 4.30327475e-01 -1.23891282e+00
-1.50728989e+00 7.03656197e-01 1.45002627e+00 9.12527084e-01
1.68542016e+00 -2.88912430e-02 1.40453720e+00 2.02375010e-01
2.09971666e-01 -1.38315117e+00 -3.24984610e-01 -1.82766184e-01
5.93660712e-01 -1.58637738e+00 -3.59048247e-01 1.25618413e-01
1.28138632e-01 9.72265542e-01 4.86400425e-01 1.07148059e-01
1.08314466e+00 6.99879408e-01 -1.54711246e-01 -1.14897033e-02
-3.27251881e-01 -1.10689566e-01 -1.16096977e-02 7.63135970e-01
-9.06445310e-02 2.33230650e-01 2.74227440e-01 8.52197945e-01
-2.36978903e-01 3.67155015e-01 -4.68379818e-02 7.48669386e-01
1.97275840e-02 -5.67803204e-01 -5.37238717e-01 7.54508734e-01
-4.44534421e-01 -1.14065092e-02 7.95145929e-02 3.36790830e-01
3.74012649e-01 1.36143899e+00 1.06780857e-01 -3.05433154e-01
4.53498542e-01 2.13738516e-01 2.37584978e-01 2.97277987e-01
5.63269816e-02 -1.09848574e-01 -3.22404802e-01 -8.05100381e-01
-6.32043719e-01 -3.23031604e-01 -7.66164839e-01 -7.11866617e-01
-3.32768977e-01 -2.71017253e-01 6.18864775e-01 1.10040104e+00
2.77213782e-01 9.35916185e-01 9.88777101e-01 -9.66643870e-01
-7.02222526e-01 -9.99804854e-01 -9.62747574e-01 5.00729799e-01
7.75058031e-01 -4.47245806e-01 -3.96024108e-01 1.23424120e-01] | [16.065776824951172, 7.579491138458252] |
114739a5-a9f4-4cf7-a86a-63823d087d02 | a-physically-informed-deep-learning-approach | 2208.04938 | null | https://arxiv.org/abs/2208.04938v1 | https://arxiv.org/pdf/2208.04938v1.pdf | A physically-informed Deep-Learning approach for locating sources in a waveguide | Inverse source problems are central to many applications in acoustics, geophysics, non-destructive testing, and more. Traditional imaging methods suffer from the resolution limit, preventing distinction of sources separated by less than the emitted wavelength. In this work we propose a method based on physically-informed neural-networks for solving the source refocusing problem, constructing a novel loss term which promotes super-resolving capabilities of the network and is based on the physics of wave propagation. We demonstrate the approach in the setup of imaging an a-priori unknown number of point sources in a two-dimensional rectangular waveguide from measurements of wavefield recordings along a vertical cross-section. The results show the ability of the method to approximate the locations of sources with high accuracy, even when placed close to each other. | ['Dmitry Batenkov', 'Eli Turkel', 'Symeon Papadimitropoulos', 'Adar Kahana'] | 2022-08-07 | null | null | null | null | ['geophysics'] | ['miscellaneous'] | [ 4.78206098e-01 -4.60841507e-02 8.13578725e-01 -2.91271538e-01
-5.14833987e-01 -2.25514069e-01 2.16480255e-01 -2.21274704e-01
-7.23312199e-01 7.28968203e-01 6.65558800e-02 -8.68762508e-02
-8.62484872e-01 -7.45307267e-01 -4.08908695e-01 -1.03031826e+00
-3.65320683e-01 6.01880491e-01 8.91641304e-02 -5.12688085e-02
2.02043489e-01 9.46701050e-01 -1.45789385e+00 -1.31255180e-01
4.61274773e-01 9.46190357e-01 1.15069181e-01 4.27799374e-01
9.20171961e-02 3.11250240e-01 -6.18009031e-01 2.34976351e-01
9.87788364e-02 -3.55922312e-01 -3.26857418e-01 -3.46543729e-01
1.66681647e-01 -5.11064470e-01 -7.77010694e-02 1.07870507e+00
5.37620664e-01 2.80076236e-01 9.41366851e-01 -6.25072300e-01
-2.75397420e-01 6.78741395e-01 -4.29197758e-01 2.66673595e-01
-5.88377081e-02 -2.82897860e-01 5.17809451e-01 -6.68516159e-01
3.65479946e-01 7.10717857e-01 8.73428524e-01 5.20193398e-01
-1.06720841e+00 -3.19880396e-01 -7.47776210e-01 1.67408153e-01
-1.27992070e+00 -5.29378355e-01 1.12694216e+00 -6.47011757e-01
5.14854908e-01 9.66483876e-02 2.21211597e-01 7.74252713e-01
4.42952057e-03 -2.89695531e-01 1.07322180e+00 -7.96622396e-01
4.06440735e-01 1.74036339e-01 3.98422897e-01 4.71701324e-01
3.97311926e-01 5.93806565e-01 -5.49162388e-01 -1.69630751e-01
1.01025212e+00 -2.84452945e-01 -7.82855392e-01 -2.78207749e-01
-9.08533692e-01 7.60427952e-01 2.36541778e-01 9.11041737e-01
-3.57912928e-01 -7.91023746e-02 -2.51906127e-01 -1.10575035e-01
6.07485056e-01 8.08581233e-01 -1.21333143e-02 1.65846914e-01
-1.15078115e+00 8.31115320e-02 8.48407626e-01 2.92857111e-01
5.52447379e-01 2.06404045e-01 2.65292317e-01 5.71765721e-01
3.93963516e-01 8.34100485e-01 3.59830111e-01 -9.91684735e-01
-3.27415168e-02 -1.22330360e-01 2.32387006e-01 -1.06416094e+00
-6.64536655e-01 -7.98615456e-01 -9.49098885e-01 8.05825710e-01
6.61125362e-01 -4.90243405e-01 -6.66472137e-01 1.67705178e+00
3.87429088e-01 5.02135634e-01 2.52608955e-01 1.28052592e+00
7.13492513e-01 7.36164927e-01 -5.19882381e-01 -2.85744727e-01
1.00773656e+00 -4.02449369e-01 -6.43185735e-01 -7.75225610e-02
5.12047783e-02 -4.99274343e-01 2.65512466e-01 6.63755000e-01
-1.24566996e+00 -3.98461133e-01 -9.17380810e-01 3.03772122e-01
-1.19650252e-01 -1.75659731e-01 7.56112486e-02 4.27485794e-01
-9.45823073e-01 8.68942261e-01 -6.01625443e-01 3.58830750e-01
-8.37485865e-02 1.23084597e-01 -2.97634840e-01 3.32726508e-01
-9.62406576e-01 5.82590342e-01 1.48017379e-02 3.86146635e-01
-5.13148904e-01 -8.84281218e-01 -5.60371816e-01 3.15313101e-01
-9.51122493e-02 -4.02708411e-01 1.04799318e+00 -5.93671978e-01
-1.47241843e+00 3.92998368e-01 -2.00512614e-02 -3.15443009e-01
3.06945503e-01 7.53095187e-03 -4.76636916e-01 6.64582074e-01
1.58004135e-01 -2.15024129e-01 8.91434491e-01 -1.53808236e+00
-1.79955065e-01 -4.07577038e-01 -2.71254539e-01 -2.25872129e-01
-1.18228480e-01 -2.61587828e-01 4.34964180e-01 -1.68641180e-01
6.79865718e-01 -4.42449301e-01 -2.43764415e-01 -1.81029681e-02
-4.55865830e-01 3.35707217e-01 6.98794127e-01 -5.77567935e-01
3.67536783e-01 -2.00911832e+00 6.99897259e-02 4.13988441e-01
3.28726918e-01 3.04444224e-01 -1.27923414e-01 6.66254580e-01
-2.36121595e-01 -2.81444639e-01 -5.67368925e-01 -2.17767358e-01
-4.18906599e-01 -1.59873188e-01 -2.32546598e-01 7.47239053e-01
-1.84464782e-01 2.06577644e-01 -6.79606676e-01 -1.45725593e-01
1.33548945e-01 7.11921692e-01 -3.04954082e-01 3.39190960e-01
1.05279751e-01 8.28700662e-01 -5.01434982e-01 7.79223666e-02
1.03842974e+00 -9.98681970e-03 -3.37758452e-01 -1.36554062e-01
-6.64494574e-01 -1.37799680e-01 -1.40223467e+00 1.31687200e+00
-6.32196784e-01 7.69913316e-01 7.48060703e-01 -1.20094442e+00
8.96031618e-01 5.51503897e-01 5.34371018e-01 -7.18578219e-01
1.64590657e-01 4.60531414e-01 2.14929935e-02 -9.85125482e-01
6.39044493e-02 -6.95864499e-01 5.74160397e-01 5.46411872e-01
7.78554752e-02 -1.47961318e-01 -3.99934687e-02 -3.42490524e-01
1.02739477e+00 -2.74479300e-01 -3.17880884e-02 -3.38623822e-01
4.89280790e-01 -1.44503564e-01 9.37778205e-02 8.89554679e-01
2.76588738e-01 7.82718122e-01 8.40870664e-02 -4.25756425e-01
-8.13370824e-01 -9.97469366e-01 -4.91170913e-01 2.39402920e-01
1.41574964e-01 5.73278904e-01 -7.82846987e-01 1.10563688e-01
-1.68312982e-01 7.59564757e-01 -3.90875757e-01 4.32701819e-02
-7.82443464e-01 -6.88664675e-01 4.60193545e-01 -3.18348855e-02
4.68221486e-01 -1.05971360e+00 -1.06867504e+00 1.78757370e-01
-2.77231246e-01 -1.30092084e+00 5.25044382e-01 2.99807340e-01
-6.98866665e-01 -9.73080099e-01 -9.76673543e-01 -4.59205240e-01
5.76869428e-01 1.91265076e-01 8.11668873e-01 -2.36885503e-01
-2.41725966e-01 4.17708486e-01 -1.39493108e-01 -5.54684639e-01
-6.12631381e-01 -3.50254357e-01 7.06362724e-02 2.09738895e-01
-8.18051919e-02 -1.08903790e+00 -5.61918259e-01 1.05780907e-01
-7.87944436e-01 -2.50764966e-01 4.55356061e-01 5.32310188e-01
2.41645854e-02 3.00324261e-01 4.53556210e-01 -3.65865678e-01
5.45291007e-01 -4.02398229e-01 -9.91140604e-01 -2.07489446e-01
-3.07596266e-01 8.17450508e-02 6.67807400e-01 -3.60700876e-01
-1.23085582e+00 -3.64123225e-01 -4.75912660e-01 -7.44626969e-02
-7.13384390e-01 5.25675595e-01 2.19704419e-01 -6.06293917e-01
7.27857471e-01 2.74221063e-01 2.41690641e-03 -6.66647673e-01
-1.78725109e-01 4.73428845e-01 7.02617228e-01 -2.80509233e-01
9.88489449e-01 6.78984702e-01 6.62380874e-01 -1.46973753e+00
-5.46625912e-01 -5.31375587e-01 -3.05296898e-01 -3.04417878e-01
6.76117957e-01 -3.75025004e-01 -8.68994534e-01 5.71010232e-01
-1.50847542e+00 -1.15773313e-01 -3.97499233e-01 1.10775936e+00
-2.04195097e-01 4.24358845e-01 -4.08565849e-01 -1.11201811e+00
-1.18815728e-01 -7.46301115e-01 7.41580844e-01 2.95292348e-01
3.48228633e-01 -1.16182554e+00 5.45953989e-01 1.37713566e-01
7.18248665e-01 2.47248784e-01 8.00617099e-01 -3.69588673e-01
-4.59226400e-01 -4.24394980e-02 -6.74719736e-02 3.39189053e-01
-2.41657212e-01 -3.08674812e-01 -1.16429901e+00 1.46881249e-02
8.60921860e-01 4.71866615e-02 8.62490892e-01 8.90974998e-01
6.93921506e-01 -8.45168456e-02 -3.36779445e-01 7.34181225e-01
1.91654563e+00 3.17052156e-01 4.18319613e-01 -2.62138750e-02
2.17796609e-01 9.73637640e-01 -1.87800363e-01 4.17067021e-01
-3.47072929e-01 4.63914037e-01 6.43312454e-01 -2.11447179e-01
-4.31574732e-02 3.93947929e-01 -2.51516283e-01 3.43652546e-01
-3.47419500e-01 -5.70062757e-01 -7.81556726e-01 5.77869475e-01
-1.17659593e+00 -9.59945202e-01 -6.35929108e-01 2.19635463e+00
3.94871205e-01 -1.01102754e-01 -4.95515764e-01 3.37608963e-01
4.79915857e-01 -9.90486592e-02 -3.14374603e-02 -8.94854814e-02
-7.37960637e-02 5.31318247e-01 4.58740383e-01 9.83543396e-01
-6.06750309e-01 -6.40649050e-02 6.49993277e+00 4.60233569e-01
-1.49104762e+00 2.10913286e-01 -1.24636209e-02 3.98814864e-02
-3.41094017e-01 -3.12509567e-01 -5.64426064e-01 3.20033461e-01
7.72733271e-01 3.07479560e-01 2.45841518e-01 1.32240728e-01
3.59330684e-01 -4.85139102e-01 -7.01265752e-01 6.65183425e-01
-7.07116798e-02 -1.25783360e+00 -1.81639329e-01 -1.51556805e-01
4.37779576e-01 -2.90755071e-02 -9.11633074e-02 -5.76253355e-01
-4.26633149e-01 -8.01779032e-01 6.40265942e-01 6.95489466e-01
6.09148264e-01 -4.20548826e-01 6.07080400e-01 8.69816840e-01
-3.03153217e-01 -9.14658885e-03 -2.66430825e-01 -3.17939043e-01
7.19763160e-01 1.26301742e+00 -6.25692606e-01 5.02909005e-01
6.14784956e-01 4.33165021e-02 2.54743934e-01 1.17458880e+00
-1.04438193e-01 8.65824342e-01 -6.43301725e-01 -2.96606892e-03
3.05060804e-01 -5.13923705e-01 8.92559946e-01 1.01246858e+00
8.60191166e-01 4.80046421e-01 -4.97611523e-01 1.35887706e+00
3.05911332e-01 -3.44456017e-01 -6.01670265e-01 4.20207828e-01
6.62151799e-02 1.14748561e+00 -7.63527930e-01 2.36257896e-01
-2.20091999e-01 2.96517640e-01 -2.12306693e-01 7.93642759e-01
-4.50561851e-01 -6.20790482e-01 1.82730481e-01 4.48556542e-01
2.83364445e-01 -4.45796728e-01 -7.29129612e-02 -5.95519662e-01
1.81592017e-01 -7.93055631e-03 -3.00011963e-01 -8.51186633e-01
-1.07104445e+00 7.90986538e-01 4.15285617e-01 -9.51378644e-01
-2.42779315e-01 -8.35414648e-01 -6.80759311e-01 1.19020712e+00
-1.92247188e+00 -8.55125904e-01 -3.13252330e-01 3.45909119e-01
-2.83877075e-01 5.70111871e-02 8.16764355e-01 2.71627873e-01
4.88154031e-03 -2.94807851e-01 5.80607772e-01 2.47076541e-01
2.62311250e-01 -9.84067619e-01 -1.54844135e-01 8.90054941e-01
4.68492433e-02 4.84673083e-01 1.23119843e+00 -2.56777465e-01
-8.00582409e-01 -4.00295317e-01 8.61596823e-01 3.68979961e-01
4.04821157e-01 -2.58880913e-01 -9.71014261e-01 1.82594508e-01
3.49032849e-01 1.89093500e-01 5.32006443e-01 -2.40642205e-01
2.14804575e-04 -2.24200428e-01 -1.29588377e+00 -1.00446135e-01
5.64728677e-01 -1.81880087e-01 -6.90730393e-01 3.51705760e-01
1.69341028e-01 -2.02444136e-01 -4.18859899e-01 3.53298396e-01
4.06022608e-01 -1.54091108e+00 1.10010850e+00 -2.76561212e-02
5.79642296e-01 4.53542806e-02 2.49693692e-01 -1.43924284e+00
-3.65514517e-01 -4.46523011e-01 3.13216895e-01 9.46218491e-01
2.11643949e-01 -1.12000132e+00 7.84123600e-01 5.58000710e-03
-1.26785755e-01 -3.91495496e-01 -1.40513229e+00 -6.87217593e-01
6.56130388e-02 -3.88745040e-01 1.30980000e-01 6.36916161e-01
-1.46550775e-01 1.44101515e-01 -1.25806332e-01 1.03800535e+00
1.25627291e+00 1.39834851e-01 -4.96487841e-02 -1.64156485e+00
-6.42849267e-01 -3.06856275e-01 -1.92693159e-01 -1.00594151e+00
1.83057114e-01 -3.67234111e-01 4.16864842e-01 -1.39668500e+00
-2.97530413e-01 -7.11464107e-01 -6.99131982e-03 -1.20144144e-01
6.16416633e-01 3.50506335e-01 -3.98288578e-01 -2.20148228e-02
2.66622931e-01 2.38957033e-01 1.09792399e+00 9.49432552e-02
-3.75952646e-02 3.71205747e-01 -5.70560545e-02 1.03560841e+00
6.15390599e-01 -8.09916019e-01 -3.18911284e-01 -8.43240201e-01
2.61208832e-01 4.31681365e-01 7.58986652e-01 -1.51453269e+00
5.30340910e-01 4.38149944e-02 1.95880577e-01 -2.07721233e-01
5.85268021e-01 -1.17067122e+00 4.23835278e-01 4.54349369e-01
-3.21750402e-01 -6.28358781e-01 1.82520896e-02 3.62767220e-01
-3.42204630e-01 -1.02652013e+00 9.27690148e-01 -2.97510892e-01
-3.24579567e-01 -2.78137922e-01 -4.95772064e-01 -9.38658137e-03
4.77558464e-01 -3.81467789e-02 -2.48107016e-01 -4.57587332e-01
-7.25405335e-01 -5.82451522e-01 -1.46971673e-01 -2.94548184e-01
6.99939489e-01 -6.72674954e-01 -8.94187629e-01 5.34373879e-01
-3.17035675e-01 -8.18612948e-02 4.42536891e-01 7.86509573e-01
-7.68687546e-01 3.29161763e-01 -1.47205114e-01 -5.15401125e-01
-1.13935888e+00 2.52171814e-01 9.42427218e-01 -1.71466339e-02
-5.09121776e-01 1.00377762e+00 1.54915288e-01 -1.52512833e-01
4.76332707e-03 -2.00217098e-01 -4.65035647e-01 -3.50168645e-01
6.48367882e-01 5.48782051e-01 1.35241106e-01 -6.18949294e-01
8.09718575e-03 1.09765589e+00 6.91323459e-01 -5.47757745e-01
1.54027891e+00 7.29018673e-02 -3.53261977e-01 4.66019154e-01
1.17856240e+00 4.03954238e-01 -9.04523849e-01 -1.79188669e-01
-3.82472813e-01 -1.98642612e-01 3.55278581e-01 -8.30309451e-01
-9.61566389e-01 1.19583559e+00 6.34130836e-01 6.49695218e-01
1.00573635e+00 2.62845345e-02 4.96431708e-01 4.04801130e-01
3.47900152e-01 -6.38331354e-01 -2.44732335e-01 2.79421091e-01
6.44021809e-01 -8.75330508e-01 -2.93322682e-01 -4.82422352e-01
3.74682754e-01 1.44732690e+00 -5.35363480e-02 -2.38938004e-01
1.02749789e+00 7.39086032e-01 3.11504751e-01 -2.02595606e-01
3.11223697e-02 9.08729061e-02 2.19542444e-01 8.18226516e-01
1.60608739e-01 -2.28163421e-01 -1.68730825e-01 2.93241143e-01
2.04649583e-01 4.27239090e-02 7.59181976e-01 6.89139664e-01
-7.37089157e-01 -7.65346587e-01 -7.50857890e-01 -5.54243289e-02
-5.14031172e-01 2.46976409e-02 1.43467605e-01 5.28442740e-01
4.30140048e-01 1.02896333e+00 1.37200341e-01 2.83082068e-01
3.74798000e-01 -7.33907893e-02 4.53670263e-01 -2.34611362e-01
-5.83047792e-02 2.83611596e-01 -1.48465499e-01 -9.36191827e-02
-7.79027045e-01 -4.95507300e-01 -1.31028867e+00 2.35684171e-01
-4.03659880e-01 7.88566470e-01 1.00241578e+00 9.52854037e-01
1.26901671e-01 3.40917528e-01 5.18712580e-01 -1.13815689e+00
-6.84865057e-01 -9.96897161e-01 -1.25698936e+00 4.01249621e-03
1.06811285e+00 -5.31319320e-01 -9.74680364e-01 -1.14747256e-01] | [12.46377182006836, -2.5769975185394287] |
4d2244c1-067b-4f81-bd36-246fadb0c452 | depth-estimation-using-modified-cost-function | 1703.00919 | null | http://arxiv.org/abs/1703.00919v2 | http://arxiv.org/pdf/1703.00919v2.pdf | Depth Estimation using Modified Cost Function for Occlusion Handling | The paper presents a novel approach to occlusion handling problem in depth
estimation using three views. A solution based on modification of similarity
cost function is proposed. During the depth estimation via optimization
algorithms like Graph Cut similarity metric is constantly updated so that only
non-occluded fragments in side views are considered. At each iteration of the
algorithm non-occluded fragments are detected based on side view virtual depth
maps synthesized from the best currently estimated depth map of the center
view. Then similarity metric is updated for correspondence search only in
non-occluded regions of the side views. The experimental results, conducted on
well-known 3D video test sequences, have proved that the depth maps estimated
with the proposed approach provide about 1.25 dB virtual view quality
improvement in comparison to the virtual view synthesized based on depth maps
generated by the state-of-the-art MPEG Depth Estimation Reference Software. | ['Olgierd Stankiewicz', 'Marek Domanski', 'Krzysztof Wegner'] | 2017-03-02 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 3.03943127e-01 2.97610432e-01 1.73462063e-01 -3.52098703e-01
-3.79275233e-01 -3.39494437e-01 2.01792464e-01 2.52844006e-01
-3.02856982e-01 7.54557550e-01 2.28592753e-01 1.67365864e-01
-1.48144722e-01 -1.06565595e+00 -2.92527556e-01 -4.76131290e-01
-5.78019954e-02 5.50369501e-01 7.49494016e-01 1.13683632e-02
8.80104661e-01 6.82184577e-01 -1.85114229e+00 2.87337184e-01
6.13299668e-01 9.66980875e-01 3.27440947e-01 8.85718822e-01
-1.04353234e-01 2.57938921e-01 -4.71856266e-01 -4.65691127e-02
6.48464382e-01 -4.37490523e-01 -4.75939721e-01 5.84097266e-01
7.41918147e-01 -7.88207650e-01 -2.87283212e-01 9.71913397e-01
4.64651108e-01 -8.57651886e-03 4.78702873e-01 -9.87230003e-01
5.23955584e-01 -4.53913771e-02 -8.55989158e-01 3.69882822e-01
9.46040750e-01 -2.67006993e-01 5.55444360e-01 -9.44224775e-01
1.21488261e+00 1.05444157e+00 2.12064594e-01 2.02148288e-01
-1.00308824e+00 -2.30384126e-01 2.19545420e-02 4.49012935e-01
-1.31493151e+00 -2.64257491e-01 9.23539519e-01 -3.85712892e-01
1.18006098e+00 2.08381712e-01 7.77743459e-01 6.13977835e-02
5.32028973e-01 1.98095530e-01 8.96907210e-01 -5.25945127e-01
3.98709565e-01 2.69141585e-01 -1.70773670e-01 9.39992011e-01
3.18846583e-01 1.80314302e-01 -5.32853007e-01 -5.77408001e-02
9.08988655e-01 -5.42765558e-01 -6.16472483e-01 -9.43554759e-01
-1.16959023e+00 6.21987641e-01 -4.82342988e-02 2.48774707e-01
-4.62595552e-01 -2.17386365e-01 4.61688399e-01 2.14785337e-01
5.30915141e-01 -2.53388405e-01 -3.77181381e-01 -1.74371839e-01
-9.31335986e-01 2.22977176e-01 7.38349497e-01 1.18660462e+00
8.74243855e-01 -4.78175059e-02 4.98240381e-01 4.43536073e-01
5.09672940e-01 1.96187451e-01 5.42835928e-02 -1.16310775e+00
7.70018756e-01 7.46644497e-01 8.29738081e-02 -1.00916326e+00
-1.95068121e-01 -2.64339328e-01 -2.99239814e-01 9.89051878e-01
4.24706161e-01 1.16706133e-01 -7.47731626e-01 9.71890926e-01
8.05475116e-01 -2.27269322e-01 2.51592815e-01 8.59909236e-01
8.65400255e-01 6.14563942e-01 -6.96878433e-01 -6.51110232e-01
9.90327239e-01 -7.88305581e-01 -7.50366032e-01 1.32294148e-01
3.51569146e-01 -1.16807425e+00 3.50787550e-01 9.74703670e-01
-1.51515734e+00 -7.21390665e-01 -1.59998322e+00 8.28146487e-02
6.82113096e-02 -3.06958735e-01 2.57468577e-02 7.72949457e-01
-1.01023865e+00 5.33692837e-01 -5.27449369e-01 -3.66059095e-01
-3.27334143e-02 3.14927042e-01 -5.20039558e-01 -2.54297286e-01
-5.79655468e-01 6.30993426e-01 3.85199875e-01 -5.13691269e-02
-8.48939598e-01 -4.99148279e-01 -7.01262176e-01 -2.30802149e-01
2.04145119e-01 -6.70720816e-01 6.25016510e-01 -8.53641391e-01
-1.38843727e+00 1.21581447e+00 -3.87584537e-01 2.16250750e-03
8.17470312e-01 -2.93787196e-02 -7.86062479e-02 8.11559319e-01
2.28257682e-02 4.12834644e-01 5.79225242e-01 -1.47251737e+00
-9.50386345e-01 -5.91866791e-01 1.49019614e-01 7.41790056e-01
4.67188090e-01 -3.31820190e-01 -6.13668382e-01 -1.02716744e-01
1.10469663e+00 -3.91820401e-01 -3.90001200e-02 3.60980868e-01
-7.92560540e-03 3.36221933e-01 1.04462838e+00 -7.21958280e-01
1.04013145e+00 -1.85509634e+00 3.77410352e-01 4.58898991e-01
2.98121065e-01 -1.16065800e-01 3.78547877e-01 5.36794782e-01
-1.21147968e-02 -3.04957449e-01 -6.38213605e-02 -8.17454979e-02
-5.36856294e-01 -1.09133972e-02 4.39352393e-01 6.48631990e-01
-6.67978525e-01 -3.04742008e-01 -7.84986377e-01 -6.84304833e-01
7.25896955e-01 5.16277552e-01 -5.14671266e-01 3.65831524e-01
-2.33608652e-02 4.14183170e-01 -2.30048016e-01 6.69402063e-01
1.26244724e+00 4.74359632e-01 7.37367943e-02 -3.56891662e-01
-3.29340547e-01 -3.51603865e-03 -1.60782564e+00 1.90526319e+00
-3.44011694e-01 6.90922737e-01 1.64584368e-01 -6.64906681e-01
1.04645252e+00 5.96627474e-01 7.10898697e-01 -5.71320474e-01
9.86344516e-02 2.22619310e-01 -1.88962907e-01 -6.28368199e-01
3.90985668e-01 8.10892209e-02 7.68488288e-01 3.58115345e-01
-1.90235049e-01 -3.85271370e-01 3.14434469e-01 1.44051507e-01
8.35823119e-01 4.40768391e-01 4.89318013e-01 -2.07607165e-01
1.25695777e+00 -5.33706956e-02 4.97173846e-01 1.14014000e-01
-3.59449863e-01 8.41826856e-01 4.61529315e-01 -3.59566271e-01
-1.21274030e+00 -1.29733360e+00 -4.61188555e-01 7.00507835e-02
3.73794556e-01 -1.89372063e-01 -8.83827329e-01 -4.99064714e-01
-3.07001293e-01 3.11034143e-01 -3.34789604e-01 5.68958163e-01
-6.65977716e-01 -1.88980505e-01 -3.10602129e-01 -5.15962951e-02
7.35951841e-01 -7.24201202e-01 -1.05535126e+00 4.94125545e-01
-6.53735548e-02 -1.08796287e+00 -2.99161300e-02 -3.11592668e-01
-1.60463595e+00 -1.34609556e+00 -8.18934619e-01 -7.85909235e-01
8.17498207e-01 4.45519507e-01 1.18990064e+00 1.46685541e-01
-3.13164681e-01 5.28321624e-01 -2.63867825e-01 3.62452902e-02
-4.04071480e-01 -5.77405453e-01 -4.85068560e-01 -1.45881698e-01
2.97100484e-01 -8.17975700e-01 -1.09190500e+00 5.52084148e-01
-6.16342962e-01 1.28077254e-01 2.46305801e-02 2.88385421e-01
6.68088257e-01 2.38105997e-01 2.67445028e-01 -7.35646784e-01
-9.15440693e-02 -1.09036840e-01 -1.12818575e+00 -8.21596533e-02
-6.22886896e-01 -1.99189618e-01 1.64544046e-01 1.13672927e-01
-1.20533097e+00 -6.97824582e-02 5.67128183e-04 -1.39438435e-01
-7.28663355e-02 -5.57210408e-02 -3.65880460e-01 -3.69057506e-01
3.51633847e-01 2.90499404e-02 -2.22390547e-01 -1.40904933e-01
-4.55562621e-02 4.75086689e-01 4.48840559e-01 -1.89950958e-01
5.11179388e-01 7.95653522e-01 3.81248593e-01 -8.21170151e-01
1.05080776e-01 -7.19842732e-01 -8.13480794e-01 -8.46672118e-01
1.01176918e+00 -7.98515022e-01 -5.25078893e-01 4.32295680e-01
-1.26924002e+00 3.79780799e-01 1.79465801e-01 7.83998191e-01
-8.97677004e-01 7.42898583e-01 -3.49550635e-01 -9.57571387e-01
-2.15372801e-01 -1.24438393e+00 9.68508005e-01 8.58421400e-02
-1.97362602e-02 -1.22951496e+00 7.23926201e-02 5.85210443e-01
-3.30086760e-02 5.64346075e-01 9.45471525e-01 2.51103848e-01
-1.25088549e+00 -2.20359609e-01 1.89818162e-02 1.79913163e-01
1.79304495e-01 1.73246354e-01 -1.03923070e+00 -1.11093201e-01
4.05846119e-01 2.96938509e-01 3.29705507e-01 6.45387352e-01
4.01157051e-01 3.78548294e-01 -3.87067229e-01 7.16265798e-01
2.26991534e+00 7.41421044e-01 9.18344498e-01 4.81661677e-01
4.93716031e-01 6.87194765e-01 1.07139397e+00 6.10170066e-01
1.22137181e-01 7.38997817e-01 8.29291999e-01 1.91849127e-01
-1.23426318e-01 -1.47720408e-02 -1.68212935e-01 5.65625429e-01
-2.29402527e-01 -4.84062850e-01 -5.04759908e-01 4.34758723e-01
-1.37423623e+00 -8.65162075e-01 -5.46190977e-01 2.70879841e+00
1.31785154e-01 6.30097032e-01 -1.04485020e-01 7.85659730e-01
6.30865872e-01 1.87733695e-01 -1.88508973e-01 -7.01342881e-01
5.46460934e-02 1.87976882e-01 4.51283842e-01 1.23805737e+00
-5.52711546e-01 4.50162619e-01 5.68652344e+00 1.83473632e-01
-6.49121761e-01 -4.14994657e-02 1.52890548e-01 -2.16258809e-01
-3.77960473e-01 1.99323252e-01 -6.72672868e-01 2.12637186e-01
4.07956243e-01 -2.17311665e-01 -6.32909760e-02 6.58262312e-01
5.06668031e-01 -1.30721712e+00 -1.14175284e+00 1.22453082e+00
3.66740942e-01 -1.03755105e+00 -1.14672139e-01 2.68211842e-01
1.02508295e+00 -4.64812875e-01 -3.42870593e-01 -6.18031085e-01
-6.33053005e-01 -3.44937950e-01 3.96964490e-01 3.20742428e-01
5.27719855e-01 -1.07431829e+00 8.66557240e-01 9.60986167e-02
-1.39096808e+00 3.94630700e-01 -4.34519857e-01 4.06930372e-02
5.64603090e-01 5.85981429e-01 -8.76211643e-01 7.42227912e-01
3.46083283e-01 4.84182328e-01 -1.39107138e-01 1.34564054e+00
-1.65882912e-02 -9.71121788e-02 -1.58497602e-01 4.67195004e-01
1.45287439e-01 -6.24801397e-01 8.62537384e-01 6.06103599e-01
4.17338282e-01 2.12274909e-01 -2.95912206e-01 4.03291285e-01
5.89393377e-01 3.25065374e-01 -9.68993425e-01 8.35291505e-01
1.81935653e-01 7.46255577e-01 -9.08775926e-01 -5.48762083e-01
-6.44276083e-01 1.18124902e+00 -3.57902497e-01 3.48107010e-01
-7.01340318e-01 -4.23245341e-01 2.48181134e-01 5.96322656e-01
3.01216781e-01 -1.74180418e-01 -2.29489326e-01 -8.30124140e-01
1.73957765e-01 -5.85067809e-01 3.19861352e-01 -9.68377352e-01
-4.24041390e-01 7.31331587e-01 3.22651625e-01 -1.57059526e+00
-5.05804300e-01 -3.29812318e-01 -4.75061834e-01 1.01651573e+00
-1.15639806e+00 -3.91132891e-01 -6.76286042e-01 4.84213680e-01
1.09104383e+00 -3.71127501e-02 5.84732652e-01 3.41183305e-01
6.45897165e-02 1.01610705e-01 2.86087189e-02 -6.09840035e-01
2.93553263e-01 -9.08656240e-01 3.15708108e-02 8.77414346e-01
-2.47936800e-01 -1.03522271e-01 9.50273097e-01 -8.00657749e-01
-9.93886769e-01 -7.86463767e-02 9.94108796e-01 -7.56297484e-02
-7.34906718e-02 -1.62122339e-01 -8.69331121e-01 3.66741598e-01
3.72199088e-01 -2.13318646e-01 2.10786566e-01 -7.14773595e-01
1.99934497e-01 -1.38435289e-01 -1.65982640e+00 1.20624520e-01
1.10351670e+00 -2.98246086e-01 -3.31520170e-01 2.16174841e-01
5.92717864e-02 -6.48103595e-01 -8.92284572e-01 4.46951270e-01
6.89955592e-01 -2.01047397e+00 1.00339866e+00 2.88311154e-01
2.57461280e-01 -4.70769018e-01 -3.49907815e-01 -7.35588789e-01
2.18704358e-01 -4.36429232e-01 2.44463235e-01 8.46618295e-01
1.08145811e-01 -4.95009840e-01 1.37717700e+00 2.89677560e-01
-2.08193853e-01 -6.58106267e-01 -1.11950028e+00 -4.08062518e-01
-7.04991162e-01 -9.12386701e-02 -5.11731058e-02 4.52077001e-01
-7.53761828e-02 1.23469725e-01 1.88220382e-01 3.82126957e-01
1.29896557e+00 1.15139000e-02 8.35965753e-01 -1.18405950e+00
-3.34350586e-01 -5.67358127e-03 -1.13248050e+00 -9.07877505e-01
-3.13237697e-01 -3.24787259e-01 -3.58852983e-01 -1.91963041e+00
-4.48932610e-02 -1.94804370e-01 2.49480188e-01 -6.75068855e-01
2.38917977e-01 1.59646362e-01 -2.39102244e-01 -2.94662952e-01
1.23514749e-01 4.03330214e-02 1.39744246e+00 3.92471015e-01
-2.97404170e-01 1.96328953e-01 3.15510541e-01 9.21908855e-01
5.51019728e-01 -3.99351209e-01 -1.05541503e+00 -2.69860446e-01
6.82388768e-02 8.11285913e-01 -1.52219802e-01 -1.39244676e+00
-5.62125184e-02 6.62345588e-02 4.66217339e-01 -1.52944648e+00
7.36268342e-01 -1.11689603e+00 4.12335396e-01 6.06946886e-01
2.00492531e-01 3.78137231e-01 -4.77236770e-02 4.70182657e-01
-2.91215360e-01 -7.50219107e-01 8.39459181e-01 -3.61411691e-01
-7.81305552e-01 7.75620118e-02 -2.63262182e-01 -1.48776948e-01
1.33487499e+00 -1.29509223e+00 1.13515534e-01 -6.00800991e-01
-9.43784237e-01 -3.05169582e-01 7.91760206e-01 -1.46610156e-01
1.04758453e+00 -1.07393897e+00 -5.72636604e-01 4.48169321e-01
5.72825484e-02 -9.57130939e-02 4.07756180e-01 4.47178662e-01
-1.43938899e+00 4.17193949e-01 -4.92245018e-01 -8.17283154e-01
-1.85717821e+00 5.75717807e-01 4.62803841e-01 -3.88740748e-02
-8.19205940e-01 6.24100626e-01 4.11572933e-01 4.36055452e-01
3.48763555e-01 -2.94973969e-01 -2.36429840e-01 -1.39104605e-01
4.50199097e-01 8.41590106e-01 2.03815550e-01 -5.68987370e-01
-2.86835611e-01 1.27430987e+00 2.19516069e-01 -6.84664786e-01
1.11768615e+00 -6.81697428e-01 9.43966210e-02 3.16370666e-01
1.31167352e+00 4.77968425e-01 -1.33790457e+00 1.17167167e-01
-3.35985541e-01 -1.24717700e+00 2.32117206e-01 -5.38371086e-01
-1.17260993e+00 8.22352529e-01 1.09772146e+00 -1.44593716e-01
1.31759334e+00 -4.54969764e-01 4.23884302e-01 -6.32500872e-02
7.86658168e-01 -1.04073858e+00 2.63855662e-02 -5.31123839e-02
8.22594643e-01 -1.04170418e+00 5.22369742e-01 -9.77003515e-01
-2.81543702e-01 1.46162915e+00 7.40730226e-01 -2.41594240e-01
7.04268217e-01 2.30321258e-01 5.56961559e-02 -2.56461471e-01
-6.13644898e-01 1.07196145e-01 -1.36129549e-02 8.62764895e-01
4.58937466e-01 -4.98984069e-01 -6.53665960e-01 -8.24894011e-01
1.97792232e-01 -1.58618093e-01 9.68117416e-01 1.03969574e+00
-8.47046971e-01 -9.62185025e-01 -7.08707094e-01 -2.07450753e-03
-3.75766218e-01 8.87274519e-02 1.43151209e-01 1.19461274e+00
2.07996875e-01 1.01263034e+00 2.68379569e-01 8.04995149e-02
5.13453901e-01 -2.14493707e-01 1.15557945e+00 -4.60712224e-01
-4.46830004e-01 4.59509432e-01 3.67072791e-01 -8.76559556e-01
-5.11725426e-01 -6.10096991e-01 -1.33982563e+00 -2.67686307e-01
-9.48663279e-02 6.53703511e-02 1.03480923e+00 2.70170808e-01
-1.59473121e-01 2.95042634e-01 8.52912247e-01 -8.91587794e-01
1.46118000e-01 -6.20494604e-01 -7.14099526e-01 2.36153975e-01
1.64686352e-01 -7.40942359e-01 -7.66005933e-01 -9.16450378e-03] | [9.302372932434082, -2.5382800102233887] |
8d27698f-263b-4093-8786-d0f7edada3af | unsupervised-behavior-change-detection-in | 1908.05103 | null | https://arxiv.org/abs/1908.05103v1 | https://arxiv.org/pdf/1908.05103v1.pdf | Unsupervised Behavior Change Detection in Multidimensional Data Streams for Maritime Traffic Monitoring | The worldwide growth of maritime traffic and the development of the Automatic Identification System (AIS) has led to advances in monitoring systems for preventing vessel accidents and detecting illegal activities. In this work, we describe research gaps and challenges in machine learning for vessel behavior change and event detection, considering several constraints imposed by real-time data streams and the maritime monitoring domain. As a starting point, we investigate how unsupervised and semi-supervised change detection methods may be employed for identifying shifts in vessel behavior, aiming to detect and label unusual events. | ['Stan Matwin', 'Vania Bogorny', 'Lucas May Petry', 'Amilcar Soares'] | 2019-08-14 | null | null | null | null | ['semi-supervised-change-detection'] | ['computer-vision'] | [ 3.15737396e-01 -1.58797041e-01 -2.09898099e-01 -5.03589571e-01
-3.49134594e-01 -7.05156207e-01 6.17675900e-01 8.51059258e-01
-5.66053152e-01 6.53790712e-01 2.20460147e-01 -1.62432745e-01
-5.04032850e-01 -8.60350311e-01 -7.18311742e-02 -5.99095404e-01
-5.43993473e-01 1.73108771e-01 3.26443881e-01 -1.66156471e-01
2.69019306e-01 1.16925383e+00 -1.54853046e+00 3.08517665e-01
2.35238045e-01 8.06805789e-01 -4.88493800e-01 8.06249619e-01
-8.00121650e-02 9.70557928e-01 -6.33670509e-01 -3.30434412e-01
3.46429020e-01 -4.33549941e-01 -7.00540900e-01 3.98719639e-01
-4.45548519e-02 2.00918004e-01 -7.37304807e-01 8.79089653e-01
3.17752630e-01 4.77464855e-01 7.16413379e-01 -1.32372272e+00
1.49928972e-01 6.56098962e-01 -1.85337052e-01 1.19807029e+00
3.69689524e-01 6.45830482e-02 5.86632311e-01 -4.31113094e-01
5.87073147e-01 7.94560313e-01 7.17501462e-01 3.47504050e-01
-8.14040840e-01 -3.35221171e-01 2.86820829e-01 5.98852158e-01
-1.28167522e+00 -4.23861086e-01 8.85935187e-01 -5.50526261e-01
8.90131533e-01 4.70778316e-01 6.84913099e-01 9.72605407e-01
1.42527804e-01 8.90584409e-01 6.08480811e-01 -4.77665663e-01
4.07938510e-01 -9.53842513e-03 1.86868459e-01 2.96896607e-01
2.60650605e-01 3.80301595e-01 -5.50086200e-01 -1.67682856e-01
4.91066128e-01 2.68204629e-01 3.92717689e-01 1.79266050e-01
-8.55832398e-01 7.32652605e-01 -5.78920655e-02 6.82925463e-01
-7.85924017e-01 -3.49423259e-01 1.03015161e+00 6.65714264e-01
5.26627898e-01 5.18909752e-01 -6.07912183e-01 -6.98823273e-01
-7.81000793e-01 2.32556928e-02 9.36462402e-01 8.42143297e-01
4.14534807e-01 2.73391545e-01 -1.09366067e-01 6.19632363e-01
-9.95402858e-02 1.72854424e-01 4.16965455e-01 -6.50127411e-01
1.88134834e-01 6.80546045e-01 -1.41220197e-01 -1.05455077e+00
-8.94523084e-01 -4.83220488e-01 -6.23711824e-01 -5.80038317e-02
4.34181720e-01 -1.56908810e-01 -6.02873981e-01 1.02217007e+00
2.98253238e-01 5.11989951e-01 1.03941530e-01 2.78463691e-01
4.41053391e-01 3.58860642e-01 3.07893932e-01 -7.06762075e-01
1.04628849e+00 -5.78727245e-01 -1.02212417e+00 4.90384549e-02
8.62547815e-01 -3.97270918e-01 2.52692908e-01 3.59536946e-01
-7.73382962e-01 -5.72303653e-01 -4.59029019e-01 6.55081570e-01
-7.99768925e-01 -3.08035135e-01 5.19511282e-01 8.11971724e-01
-2.06010848e-01 4.06956851e-01 -1.30226600e+00 -4.79818135e-01
4.05336171e-01 7.41026849e-02 -2.67753661e-01 3.02180320e-01
-1.16743708e+00 9.86315429e-01 5.47806978e-01 3.91582221e-01
-8.81290019e-01 -5.13403773e-01 -9.96588111e-01 -2.84339249e-01
3.63224268e-01 1.81465253e-01 1.26731420e+00 -6.72484457e-01
-1.18872404e+00 5.74833810e-01 -1.47305848e-02 -8.93061757e-01
6.04880452e-01 1.39229238e-01 -1.28222859e+00 3.50112724e-03
1.11179717e-01 -2.01276809e-01 4.31242913e-01 -6.86666548e-01
-1.49580169e+00 -3.06402415e-01 -3.70962024e-01 -1.64875522e-01
-2.82915920e-01 4.81000215e-01 5.66262975e-02 -8.13382864e-01
3.34117375e-03 -6.95011258e-01 -4.31565136e-01 -4.99671429e-01
-1.42015442e-01 -4.19896364e-01 9.28024173e-01 -7.15104818e-01
1.60382533e+00 -2.22733927e+00 -3.48230094e-01 6.89330339e-01
-2.08659559e-01 3.32044095e-01 1.92200199e-01 6.22317135e-01
-2.25946158e-01 -2.10875317e-01 -2.69007891e-01 -1.76526994e-01
-3.14507544e-01 7.04126060e-01 -1.27225056e-01 5.48973441e-01
2.45696560e-01 5.37968040e-01 -1.29066992e+00 -2.09981889e-01
6.71060622e-01 -1.36433661e-01 9.81701910e-02 -2.05306895e-02
4.18599129e-01 8.58810067e-01 -4.16343421e-01 9.43834543e-01
2.47805148e-01 6.26674354e-01 -2.29505792e-01 1.24602608e-01
-6.01765931e-01 2.72779435e-01 -1.29243839e+00 1.21261370e+00
-1.26225859e-01 9.07938480e-01 -2.92980075e-01 -1.50217104e+00
1.10244215e+00 4.53901023e-01 1.07344794e+00 -8.56573462e-01
9.99821052e-02 -1.75210834e-02 -1.71009809e-01 -1.25844097e+00
2.78004676e-01 -1.67117059e-01 -4.31417704e-01 1.49319530e-01
-1.49278268e-01 4.81643111e-01 7.48332322e-01 -3.16790521e-01
1.52557731e+00 -4.96954918e-01 5.64429402e-01 1.69939578e-01
7.13517070e-01 4.16047990e-01 7.80086279e-01 8.70868564e-01
-5.98148763e-01 -1.51796281e-01 1.59314811e-01 -1.00954449e+00
-6.21769965e-01 -8.99111032e-01 -2.28479922e-01 1.22683275e+00
-6.84451610e-02 -9.79693457e-02 -2.37291709e-01 -7.85128117e-01
1.14745989e-01 8.26600134e-01 -5.59192419e-01 -5.43329120e-01
-9.07777548e-01 -7.30752289e-01 9.01552439e-01 7.92462587e-01
2.83085197e-01 -1.27480376e+00 -6.42155409e-01 8.23766232e-01
4.06867824e-02 -1.01134849e+00 -7.15807006e-02 2.56312102e-01
-8.74817908e-01 -1.25695598e+00 -1.21027559e-01 -7.52880096e-01
5.59288919e-01 -6.72581568e-02 8.34451556e-01 -2.86584169e-01
-7.00499773e-01 4.93584245e-01 -6.88853800e-01 -7.03242421e-01
-7.40116358e-01 2.25578234e-01 2.43497655e-01 3.04015547e-01
8.47504735e-01 -4.09162730e-01 -2.55096018e-01 4.45335597e-01
-1.11402643e+00 -8.34930658e-01 4.87946272e-01 2.67622501e-01
2.20598429e-01 7.62703478e-01 8.73261511e-01 -9.15218353e-01
6.28342688e-01 -7.33046114e-01 -3.64073426e-01 2.57319152e-01
-6.39574111e-01 -3.73102546e-01 2.20519200e-01 -4.16910678e-01
-1.09506249e+00 3.19961369e-01 -2.97591984e-01 -4.03316356e-02
-1.06427836e+00 7.20357358e-01 1.46504492e-01 1.22050852e-01
7.74933279e-01 2.33614385e-01 -8.08304474e-02 -7.12801754e-01
-9.02744606e-02 9.49155867e-01 6.98986411e-01 -1.23135485e-02
7.46606827e-01 8.13812912e-01 -1.00070387e-01 -9.86150384e-01
-8.30783784e-01 -1.10256672e+00 -1.19245315e+00 -7.92498112e-01
5.96572459e-01 -6.13977790e-01 -6.08786702e-01 4.90893334e-01
-8.58146548e-01 8.61850083e-02 -7.70093262e-01 7.17613280e-01
-3.48342687e-01 2.70093888e-01 -5.03132522e-01 -1.21464705e+00
1.21353589e-01 -3.97176564e-01 6.42738819e-01 1.98433101e-01
-4.54127282e-01 -1.22879124e+00 3.95016879e-01 1.34637922e-01
3.22615504e-01 6.74678385e-01 2.16676131e-01 -9.86463010e-01
1.06955104e-01 -8.63406479e-01 3.39778572e-01 3.15669626e-01
5.91560066e-01 -1.17347822e-01 -6.39926255e-01 -1.81697100e-01
-1.28723651e-01 6.00955427e-01 6.21976376e-01 4.56031173e-01
7.34424293e-01 -1.42463654e-01 -5.08343458e-01 1.32721841e-01
8.47021878e-01 8.95629466e-01 3.44856232e-01 7.23679841e-01
2.86203444e-01 8.05593431e-01 8.47613215e-01 8.00460339e-01
2.35451028e-01 3.02185565e-01 2.21087664e-01 -2.69782394e-01
2.01163739e-01 -3.33037041e-02 1.70420259e-01 5.76835513e-01
-4.23333436e-01 8.07930604e-02 -1.03651953e+00 8.74989569e-01
-2.08613181e+00 -1.51425660e+00 -4.33644354e-01 2.04736066e+00
5.00640750e-01 3.90115529e-01 6.05827272e-01 5.33840716e-01
8.03010166e-01 -1.58400998e-01 -6.24082267e-01 -3.32969904e-01
2.07120758e-02 -2.60921806e-01 8.35824728e-01 5.87331690e-02
-1.49370492e+00 6.41431272e-01 7.68713140e+00 4.12220508e-01
-8.85690689e-01 7.65189976e-02 2.52938032e-01 -8.58230591e-02
3.53422910e-01 -3.65181655e-01 -7.36616910e-01 5.18421471e-01
1.16596472e+00 -2.15920553e-01 4.82225828e-02 6.28206491e-01
8.15816820e-01 1.59235835e-01 -9.46810424e-01 7.69123435e-01
1.17467493e-01 -1.15370929e+00 -2.16297284e-01 -1.29770845e-01
5.48353553e-01 6.01045378e-02 -4.47602808e-01 3.93260956e-01
1.18680798e-01 -2.91082978e-01 6.10470116e-01 8.03151071e-01
-3.82170863e-02 -9.55766201e-01 8.95164847e-01 3.19559604e-01
-1.31987965e+00 -6.12814009e-01 9.84028727e-03 -4.08052176e-01
7.35431552e-01 6.04651213e-01 -8.65329325e-01 5.31067729e-01
8.49064171e-01 1.17545795e+00 -3.94087613e-01 1.48432112e+00
7.86355808e-02 9.72918034e-01 -3.39826852e-01 -1.94005945e-04
3.08141649e-01 4.20865752e-02 9.36422825e-01 1.60656106e+00
-5.22481613e-02 -6.43048882e-02 4.11719352e-01 1.38568535e-01
4.85976249e-01 -1.14926785e-01 -7.43940055e-01 1.01837069e-01
3.14076751e-01 9.14502740e-01 -8.82580936e-01 -2.24287570e-01
-5.38244367e-01 6.31100357e-01 -1.03067912e-01 1.92098826e-01
-8.70132923e-01 -5.74503958e-01 8.27202857e-01 2.02725664e-01
1.51024938e-01 -5.02278924e-01 -2.14124963e-01 -9.21418130e-01
-1.26793087e-02 -3.37208152e-01 1.23743594e+00 -8.98041483e-03
-1.49665725e+00 3.72685969e-01 2.96862245e-01 -1.68931639e+00
-2.09999472e-01 -3.74365002e-01 -7.98959255e-01 -4.54371935e-03
-1.76879776e+00 -7.44630754e-01 -3.68593372e-02 7.30094016e-01
9.21827614e-01 -4.78297800e-01 5.18732369e-01 6.55468404e-01
-8.15551400e-01 3.75656754e-01 1.52254045e-01 5.97814679e-01
3.49378735e-01 -9.67648208e-01 2.58207053e-01 9.73344445e-01
2.02152640e-01 3.74433920e-02 8.79704416e-01 -7.08348453e-01
-9.07450795e-01 -1.35950696e+00 9.34779286e-01 -2.88532645e-01
1.07933605e+00 -1.77736282e-02 -8.13036263e-01 8.11243951e-01
-1.78775564e-01 -3.18114042e-01 1.10811245e+00 8.48370269e-02
5.58883213e-02 -2.54211009e-01 -1.15181553e+00 2.30973914e-01
1.13977194e+00 -3.48476261e-01 -7.33151495e-01 6.21185422e-01
4.51422147e-02 -2.06214890e-01 -1.17491496e+00 3.71512234e-01
1.06410295e-01 -2.77214468e-01 8.06682646e-01 -1.16929018e+00
-3.06714445e-01 -2.90780872e-01 4.55906749e-01 -1.19333625e+00
-5.89097381e-01 -7.78788984e-01 -4.47093807e-02 1.21665180e+00
4.85206962e-01 -6.06746376e-01 7.01721072e-01 9.56788242e-01
-3.52155894e-01 1.21005580e-01 -1.34748352e+00 -9.14446652e-01
-4.40230310e-01 -8.68792415e-01 3.74970049e-01 1.13733923e+00
3.80594522e-01 -7.89828375e-02 -4.56765294e-01 4.37373698e-01
3.85414273e-01 -2.41623431e-01 3.74061018e-01 -1.54005229e+00
1.86339393e-01 -4.25557584e-01 -1.17407537e+00 -5.46592712e-01
1.74535085e-02 -5.73998868e-01 -1.29769361e-02 -1.05887377e+00
-3.97393405e-01 -1.35407329e-01 -5.97420752e-01 6.61930799e-01
2.80197948e-01 6.58628419e-02 -3.81349117e-01 2.11806640e-01
-7.63556659e-01 1.25817388e-01 5.12438118e-01 -1.56065971e-01
-5.58649600e-01 7.22794652e-01 -2.50560701e-01 7.12381065e-01
1.03869820e+00 -7.57725120e-01 1.80099875e-01 -1.28123149e-01
3.60062599e-01 -1.02193020e-01 1.99684739e-01 -9.89778876e-01
7.28090286e-01 -2.69221187e-01 -7.08579049e-02 -7.20292270e-01
-3.28329295e-01 -1.19459140e+00 4.51348275e-02 7.58255661e-01
-5.93710482e-01 1.36505470e-01 1.99500978e-01 1.01580369e+00
-6.61767662e-01 -4.46571887e-01 5.71987152e-01 -1.27348593e-02
-1.16513956e+00 1.21527486e-01 -1.43965459e+00 -3.19685698e-01
1.63655257e+00 -3.75449628e-01 -2.48061381e-02 -9.19479802e-02
-1.37011623e+00 3.87490273e-01 -4.65970635e-01 8.32866311e-01
3.41103435e-01 -1.18403959e+00 -9.73277509e-01 4.65654582e-01
3.72525632e-01 -3.90383393e-01 3.02947164e-01 1.00357568e+00
-4.91189927e-01 2.19395071e-01 -1.62042320e-01 -3.74004394e-01
-1.32580924e+00 5.62160075e-01 3.66246372e-01 -1.25558779e-01
-7.35291779e-01 4.18966472e-01 -9.24261332e-01 -2.87129849e-01
3.75962943e-01 -3.29172730e-01 -6.79461539e-01 3.42375785e-01
8.46536100e-01 1.07083035e+00 4.02089924e-01 -7.22908020e-01
-3.47153842e-01 8.79515037e-02 2.78050592e-03 2.79117227e-01
1.50238967e+00 -2.41894901e-01 1.14785850e-01 6.23430192e-01
9.23073947e-01 -4.35005188e-01 -9.43999052e-01 -3.80201280e-01
7.80067921e-01 -3.49251509e-01 6.21725023e-02 -4.99948710e-01
-1.00187421e+00 2.30737895e-01 7.61055350e-01 1.19881999e+00
1.11997712e+00 2.43910760e-01 7.54970968e-01 7.49058604e-01
2.96489656e-01 -1.49179029e+00 -5.86649001e-01 5.25211275e-01
4.57126468e-01 -1.32387686e+00 -4.41317946e-01 -3.98799658e-01
-7.56172895e-01 1.24771857e+00 1.47241473e-01 -7.30832620e-03
1.10466897e+00 4.71670270e-01 2.89944887e-01 -4.11283880e-01
-5.27163029e-01 -6.26516998e-01 2.27041468e-02 7.35194743e-01
6.73540542e-03 5.22643998e-02 -3.79766226e-01 1.79624438e-01
1.45343706e-01 -1.14124484e-01 2.87368059e-01 1.40388882e+00
-5.39285362e-01 -1.05680740e+00 -2.64545888e-01 6.54271066e-01
-4.24117655e-01 5.46031713e-01 -3.54139745e-01 5.97469866e-01
2.85255313e-01 1.43186367e+00 4.50389087e-01 -2.28320003e-01
1.14985359e+00 1.23457253e-01 -2.66921341e-01 -4.69627380e-01
-9.02862191e-01 -1.14561968e-01 4.62513149e-01 -3.72535139e-01
-9.62731779e-01 -1.45860600e+00 -1.12325287e+00 4.61550280e-02
-3.08272213e-01 2.97938794e-01 5.90687513e-01 1.12939596e+00
2.94109315e-01 7.66128242e-01 9.61012244e-01 -3.84724736e-01
-1.72347918e-01 -1.04402947e+00 -8.16272497e-01 7.06829548e-01
1.71313763e-01 -6.10583901e-01 -5.79156697e-01 5.62503278e-01] | [7.187201976776123, 2.7496418952941895] |
4a576248-f03f-4e46-b3f3-8c9ca1e012bc | blending-target-domain-adaptation-by-1 | 1907.03389 | null | https://arxiv.org/abs/1907.03389v1 | https://arxiv.org/pdf/1907.03389v1.pdf | Blending-target Domain Adaptation by Adversarial Meta-Adaptation Networks | (Unsupervised) Domain Adaptation (DA) seeks for classifying target instances when solely provided with source labeled and target unlabeled examples for training. Learning domain-invariant features helps to achieve this goal, whereas it underpins unlabeled samples drawn from a single or multiple explicit target domains (Multi-target DA). In this paper, we consider a more realistic transfer scenario: our target domain is comprised of multiple sub-targets implicitly blended with each other, so that learners could not identify which sub-target each unlabeled sample belongs to. This Blending-target Domain Adaptation (BTDA) scenario commonly appears in practice and threatens the validities of most existing DA algorithms, due to the presence of domain gaps and categorical misalignments among these hidden sub-targets. To reap the transfer performance gains in this new scenario, we propose Adversarial Meta-Adaptation Network (AMEAN). AMEAN entails two adversarial transfer learning processes. The first is a conventional adversarial transfer to bridge our source and mixed target domains. To circumvent the intra-target category misalignment, the second process presents as ``learning to adapt'': It deploys an unsupervised meta-learner receiving target data and their ongoing feature-learning feedbacks, to discover target clusters as our ``meta-sub-target'' domains. These meta-sub-targets auto-design our meta-sub-target DA loss, which empirically eliminates the implicit category mismatching in our mixed target. We evaluate AMEAN and a variety of DA algorithms in three benchmarks under the BTDA setup. Empirical results show that BTDA is a quite challenging transfer setup for most existing DA algorithms, yet AMEAN significantly outperforms these state-of-the-art baselines and effectively restrains the negative transfer effects in BTDA. | ['Liang Lin', 'Xiaodan Liang', 'Ziliang Chen', 'Jingyu Zhuang'] | 2019-07-08 | blending-target-domain-adaptation-by | http://openaccess.thecvf.com/content_CVPR_2019/html/Chen_Blending-Target_Domain_Adaptation_by_Adversarial_Meta-Adaptation_Networks_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Chen_Blending-Target_Domain_Adaptation_by_Adversarial_Meta-Adaptation_Networks_CVPR_2019_paper.pdf | cvpr-2019-6 | ['multi-target-domain-adaptation'] | ['computer-vision'] | [ 3.75948578e-01 3.17610912e-02 -4.07648832e-01 -3.90386671e-01
-9.51561034e-01 -9.22357798e-01 7.18944430e-01 -1.34142131e-01
-2.44430497e-01 7.99580276e-01 1.09042712e-02 -1.99689120e-01
2.89122045e-01 -8.33592534e-01 -8.73251200e-01 -8.14278603e-01
1.89380795e-01 7.30590343e-01 2.62822151e-01 -3.34524184e-01
-2.80549139e-01 2.94389009e-01 -1.16965580e+00 4.84767139e-01
1.07916915e+00 1.01945543e+00 -1.19250081e-01 2.26021230e-01
-2.83439815e-01 6.31240368e-01 -7.70016313e-01 -5.84168017e-01
4.92537051e-01 -4.85000491e-01 -7.44024932e-01 1.73179895e-01
5.10174215e-01 -2.81990826e-01 -2.01117158e-01 1.00758731e+00
3.25006038e-01 2.67180782e-02 1.10667169e+00 -1.67106378e+00
-9.64818954e-01 4.87380445e-01 -6.75367236e-01 1.89098641e-02
-1.34088606e-01 4.31528151e-01 8.37194800e-01 -1.06801140e+00
3.87651384e-01 1.27645326e+00 7.37863183e-01 8.07645679e-01
-1.45384467e+00 -9.71122265e-01 3.42765301e-01 -7.76396915e-02
-1.06635451e+00 -3.74919415e-01 1.03328371e+00 -5.54870129e-01
4.11490411e-01 4.31048451e-03 1.04596756e-01 1.72903681e+00
-1.56562492e-01 9.64937806e-01 1.07450771e+00 -4.07063276e-01
3.47322524e-01 4.84948337e-01 2.05937713e-01 5.01049906e-02
4.81930822e-02 4.76410955e-01 -1.77516475e-01 -1.44667462e-01
7.44174540e-01 6.69328198e-02 -1.11012071e-01 -8.50314498e-01
-1.14628935e+00 9.24334466e-01 6.51472926e-01 1.07143305e-01
-2.69214123e-01 -3.29359651e-01 5.58739662e-01 7.41137445e-01
4.85620141e-01 3.75019401e-01 -6.53619766e-01 2.56665468e-01
-5.00482619e-01 2.69289732e-01 5.74779451e-01 1.23078322e+00
1.00944078e+00 2.61973858e-01 -3.42249990e-01 1.08028233e+00
9.21975449e-03 5.49595356e-01 7.17777252e-01 -6.09129369e-01
7.42379129e-01 8.18600476e-01 1.38845861e-01 -4.24640417e-01
8.82857386e-03 -6.30425036e-01 -9.18520153e-01 4.07477319e-01
7.56290853e-01 -3.61956716e-01 -1.22746170e+00 2.07017303e+00
5.44938862e-01 1.60901889e-01 3.20629865e-01 7.59154260e-01
6.03765190e-01 4.96553272e-01 2.49178052e-01 2.82820426e-02
1.08177400e+00 -1.14213085e+00 -3.44763815e-01 -5.54890573e-01
4.79515404e-01 -5.05813658e-01 1.32247102e+00 1.78239241e-01
-7.67520070e-01 -7.73153901e-01 -1.06207228e+00 2.55723894e-01
-5.41919529e-01 -1.26071036e-01 2.14221045e-01 5.40145636e-01
-6.76299393e-01 3.15750450e-01 -4.80495811e-01 -1.47093207e-01
6.92605555e-01 3.47761154e-01 -3.68046343e-01 -3.32577199e-01
-1.27915573e+00 9.34293628e-01 4.25784677e-01 -3.60970289e-01
-1.20736182e+00 -1.10797691e+00 -7.96718299e-01 -6.57295138e-02
3.71663630e-01 -6.29565060e-01 1.43831778e+00 -1.83493769e+00
-1.34813488e+00 1.11918330e+00 8.19389597e-02 -4.00681049e-01
6.82997704e-01 -1.66495085e-01 -7.25266278e-01 -1.95474207e-01
1.81517348e-01 7.64599800e-01 1.21500456e+00 -1.55670261e+00
-6.19298637e-01 -3.61285985e-01 -9.88109410e-02 2.76062340e-01
-5.42247057e-01 -2.86970139e-01 -5.89131042e-02 -1.01711464e+00
-3.19427103e-01 -8.79458070e-01 -3.21403407e-02 -8.48200023e-02
-3.14225227e-01 -1.65866584e-01 1.11954105e+00 -2.37654105e-01
9.73585427e-01 -2.21718526e+00 1.93974048e-01 9.18726772e-02
2.29655206e-01 6.30167425e-01 -4.63334978e-01 3.16568702e-01
-3.71182382e-01 -1.81964859e-01 -5.47148585e-01 -2.15795949e-01
-2.58957036e-02 2.47014716e-01 -5.87284207e-01 1.78072929e-01
6.39190316e-01 8.65311921e-01 -9.59403872e-01 -1.43079340e-01
-7.43577331e-02 1.84638306e-01 -4.38578874e-01 5.56705594e-01
-4.24954087e-01 7.14162827e-01 -3.66900206e-01 8.24562013e-01
7.96587527e-01 -1.69273689e-01 3.91659141e-02 5.23695908e-03
3.62165511e-01 3.13044451e-02 -9.64320421e-01 1.47088754e+00
-2.10716620e-01 2.30573833e-01 3.17762867e-02 -1.28490126e+00
1.10164869e+00 1.40341237e-01 4.17918265e-01 -5.80501378e-01
1.60645675e-02 3.68754685e-01 1.20633587e-01 6.66335374e-02
1.37229040e-01 -4.68346894e-01 -2.45145679e-01 2.87825793e-01
3.18262637e-01 1.79923892e-01 -2.72965312e-01 6.10233806e-02
1.10957766e+00 2.63553590e-01 3.70756865e-01 -1.64032698e-01
5.50830245e-01 3.13368261e-01 8.24868619e-01 6.14710510e-01
-6.15507841e-01 5.91651022e-01 2.92644471e-01 -4.18991059e-01
-1.15872478e+00 -1.50168335e+00 1.41859632e-02 1.44753873e+00
1.39505208e-01 2.86571562e-01 -6.44344330e-01 -1.35760045e+00
3.65878999e-01 7.81873167e-01 -9.50754464e-01 -5.63809872e-01
-4.32468474e-01 -5.11535466e-01 6.36884332e-01 6.83841467e-01
4.99318272e-01 -1.11846709e+00 -1.14091672e-01 3.21214020e-01
-6.16251901e-02 -7.72087991e-01 -6.93935096e-01 4.71865863e-01
-7.37101436e-01 -1.08624458e+00 -9.13443267e-01 -8.73133361e-01
6.65113509e-01 2.00003564e-01 1.42896104e+00 -4.67873722e-01
2.24376619e-01 2.76434213e-01 -4.51519907e-01 -5.79315424e-01
-8.26487362e-01 9.22213793e-02 1.75826728e-01 2.70224810e-01
8.22070897e-01 -7.99965024e-01 -4.30120468e-01 6.56211495e-01
-9.86223578e-01 -1.74481079e-01 7.89365590e-01 1.20314491e+00
5.66090047e-01 -2.83926964e-01 1.10923123e+00 -1.21977937e+00
3.66113156e-01 -1.03763151e+00 -4.34001148e-01 2.50935048e-01
-5.66889465e-01 -6.58110604e-02 9.52228665e-01 -1.13043511e+00
-1.02922285e+00 -1.81614831e-02 1.94917455e-01 -1.04396844e+00
-4.46084410e-01 2.89565563e-01 -6.69971108e-01 2.25724787e-01
1.07608604e+00 1.42209724e-01 2.50992626e-01 -3.75708282e-01
3.23932886e-01 6.78104818e-01 7.60960877e-01 -9.02411699e-01
1.31803966e+00 1.80395037e-01 -4.62083310e-01 -2.34694004e-01
-9.73431230e-01 -4.34100240e-01 -8.76612902e-01 1.27336323e-01
5.01697302e-01 -1.19511592e+00 -1.51525484e-03 6.54812157e-01
-8.73661101e-01 -7.23361671e-01 -6.15942240e-01 1.36760905e-01
-5.24263084e-01 -2.89644860e-02 -2.41100550e-01 -2.92518944e-01
-2.42433727e-01 -1.05068183e+00 6.29296124e-01 7.01171681e-02
-1.93894997e-01 -1.17417431e+00 2.04793662e-01 9.26927105e-02
4.76196319e-01 4.65057969e-01 9.47895765e-01 -1.23333657e+00
-4.72211726e-02 3.27039813e-03 -7.36176968e-02 7.11864650e-01
5.44918776e-01 -4.05979544e-01 -1.24147773e+00 -6.34516537e-01
-9.84511077e-02 -7.20809758e-01 6.70255363e-01 3.52707170e-02
1.03558004e+00 -4.13134426e-01 -3.03244710e-01 5.18419385e-01
1.18677127e+00 3.00529003e-01 5.11519194e-01 4.63378876e-01
7.68379986e-01 5.85325837e-01 8.65663826e-01 5.44101335e-02
3.05295616e-01 6.20234013e-01 4.71078008e-01 -2.78729379e-01
-2.55814344e-01 -3.79052222e-01 6.60885096e-01 5.08807778e-01
4.57362652e-01 -1.91036165e-01 -1.01293993e+00 7.68521667e-01
-1.69851720e+00 -7.69350171e-01 3.13371181e-01 2.33746743e+00
1.27904975e+00 1.42822355e-01 5.90788424e-01 -7.70102963e-02
9.87457097e-01 -1.69554248e-01 -1.20134306e+00 -2.46854752e-01
-1.15472637e-01 2.41652057e-01 2.84103900e-01 1.68157041e-01
-1.43837404e+00 9.60964799e-01 5.25339556e+00 8.57957006e-01
-1.30939579e+00 1.49355829e-01 6.63261354e-01 1.83193475e-01
-2.70660907e-01 -3.23460817e-01 -7.35572755e-01 5.79848588e-01
9.58026171e-01 -1.98272720e-01 3.19714993e-01 1.08881426e+00
-3.82736802e-01 5.25526285e-01 -1.55322540e+00 6.44504189e-01
-1.93860769e-01 -9.27812636e-01 2.72400528e-01 1.39617637e-01
8.64959955e-01 9.43605378e-02 4.14109766e-01 9.51159716e-01
7.86262989e-01 -9.28971589e-01 6.91057384e-01 -1.53216302e-01
1.15947139e+00 -8.50377858e-01 4.79948193e-01 4.13632691e-01
-9.33069050e-01 -3.48141581e-01 -2.39835307e-01 2.52448559e-01
-2.77516246e-01 3.37703675e-01 -9.26333249e-01 5.92665672e-01
5.24388909e-01 6.89873934e-01 -5.22427917e-01 7.11630702e-01
-2.67095566e-02 7.72453368e-01 -4.69907932e-02 6.09023869e-01
2.68865973e-01 -1.33952901e-01 6.10376835e-01 1.02683353e+00
1.78159297e-01 -1.65144533e-01 3.05808902e-01 1.00007343e+00
-4.11227643e-01 -1.58703312e-01 -7.23892570e-01 4.33220789e-02
8.32045913e-01 9.34177697e-01 -1.58606470e-01 -4.04113859e-01
-4.78239447e-01 1.06031239e+00 5.18096387e-01 5.05745709e-01
-9.13493454e-01 -2.19782159e-01 9.55136895e-01 1.98166952e-01
3.54093850e-01 3.14981610e-01 -3.86478305e-01 -1.29493701e+00
-6.06021732e-02 -1.38301110e+00 6.89335108e-01 -2.35687450e-01
-2.11034322e+00 5.56599379e-01 -1.58477537e-02 -1.80708873e+00
-3.50423753e-01 -5.82006574e-01 -7.14973330e-01 1.02912605e+00
-1.68352330e+00 -1.59356964e+00 -2.50119716e-01 1.06523287e+00
6.64170146e-01 -4.29942101e-01 9.60142195e-01 3.15952331e-01
-5.08800864e-01 1.23378611e+00 3.49719882e-01 3.93520385e-01
1.27026844e+00 -1.42829096e+00 4.00187910e-01 8.33111346e-01
-1.31604478e-01 4.72522497e-01 3.54108304e-01 -5.67160308e-01
-1.03318083e+00 -1.69280636e+00 2.71033078e-01 -7.44893253e-01
6.48394227e-01 -3.46406311e-01 -1.38961291e+00 1.08232033e+00
-3.58407199e-03 2.77572155e-01 8.59364450e-01 4.25540060e-02
-8.81969631e-01 -2.80877829e-01 -1.43717897e+00 4.27300811e-01
7.97020018e-01 -3.67939472e-01 -8.75375330e-01 1.14581734e-01
9.15762305e-01 -4.12013859e-01 -9.83913541e-01 6.34416640e-01
1.76739022e-01 -7.41439402e-01 1.09182262e+00 -9.00239050e-01
5.60608268e-01 -1.95000827e-01 -8.40960070e-02 -1.79488766e+00
-3.22508574e-01 -3.52450937e-01 -2.31647640e-01 1.60967135e+00
2.75892526e-01 -8.06744218e-01 7.10744619e-01 4.55591530e-01
-3.74279648e-01 -4.24347758e-01 -7.98988044e-01 -1.08232272e+00
9.25924897e-01 -2.73695365e-02 8.28697741e-01 1.62969089e+00
-3.27484816e-01 4.12153900e-01 -1.60606623e-01 1.92714080e-01
7.08301187e-01 6.37369752e-02 1.03890371e+00 -1.26060236e+00
-2.83335000e-01 -4.03486460e-01 6.75831968e-03 -9.36114430e-01
2.50616342e-01 -9.97802377e-01 -1.06924335e-02 -7.21069574e-01
3.45248319e-02 -8.71078908e-01 -6.52547240e-01 7.94867933e-01
-4.20075566e-01 9.84599814e-02 6.75916374e-02 3.83225322e-01
-2.84915030e-01 7.05643654e-01 1.22706389e+00 -3.96156907e-01
-4.06129688e-01 1.05292872e-01 -9.67164099e-01 5.02343476e-01
6.87166989e-01 -4.70698416e-01 -6.10017061e-01 -3.29698712e-01
-6.30309403e-01 -1.91853032e-01 3.40972573e-01 -9.17731524e-01
-6.30673766e-02 -3.82753491e-01 6.07199967e-01 -3.53989869e-01
1.35473162e-01 -1.01724172e+00 3.74260731e-02 2.03454792e-01
-2.15398371e-01 -2.46322930e-01 3.82878929e-01 6.19616866e-01
-3.96396011e-01 1.20472431e-01 1.23492336e+00 -7.74648180e-03
-7.68374801e-01 3.60342622e-01 9.53985378e-02 4.19900715e-01
1.26420081e+00 -3.29657465e-01 -5.06068885e-01 -1.22172013e-01
-6.41251206e-01 4.08527136e-01 6.65012002e-01 5.92391551e-01
3.58945519e-01 -1.62066352e+00 -7.73135722e-01 3.02663952e-01
4.43493783e-01 4.28120047e-01 1.05500437e-01 5.31221271e-01
1.29657686e-01 -8.31955001e-02 -5.27199745e-01 -7.69017875e-01
-8.04998100e-01 8.72571051e-01 4.39317912e-01 -4.70303863e-01
-3.53190094e-01 7.71688640e-01 8.11955333e-01 -9.92378354e-01
2.33583525e-01 3.90607044e-02 -1.19909845e-01 5.18346615e-02
4.07141656e-01 1.44998819e-01 8.67504068e-03 -5.51687121e-01
-1.33919403e-01 1.01757847e-01 -4.09765542e-01 2.26805165e-01
1.20065010e+00 1.20509155e-02 3.87109995e-01 4.11938697e-01
1.14150298e+00 -1.29834592e-01 -1.69504082e+00 -6.97093129e-01
7.76668489e-02 -2.68906862e-01 -4.90244329e-01 -1.27221489e+00
-9.13904369e-01 8.97549510e-01 5.98617315e-01 -9.47647542e-02
1.30511463e+00 -1.48201004e-01 7.56557465e-01 1.00102134e-01
2.00521290e-01 -8.27102423e-01 4.82538223e-01 5.56294978e-01
9.16251183e-01 -1.48196697e+00 -4.88843977e-01 -3.27552736e-01
-9.83025908e-01 7.93886125e-01 1.29090452e+00 -1.07793078e-01
3.28045756e-01 -2.90363492e-03 3.54569674e-01 2.10724726e-01
-5.70894182e-01 -8.14739242e-02 2.96771735e-01 1.07573605e+00
-1.97489485e-02 8.93518254e-02 3.40500981e-01 9.42225993e-01
7.13165700e-02 -2.09563106e-01 1.91078439e-01 8.44125509e-01
-1.54657677e-01 -1.49140465e+00 -5.34730256e-01 2.91487038e-01
-2.37798110e-01 2.66727433e-02 -5.51614821e-01 1.02443326e+00
2.01194465e-01 6.92450106e-01 1.13317333e-01 -4.75378692e-01
5.20480692e-01 2.89654255e-01 2.00545505e-01 -7.10458159e-01
-8.05240512e-01 -8.11456069e-02 -2.49235421e-01 -2.01270387e-01
-2.12037906e-01 -4.75831777e-01 -1.07188523e+00 -1.22338668e-01
-6.29522353e-02 1.80801407e-01 1.38375074e-01 7.34292865e-01
4.12264675e-01 4.67263401e-01 9.83070254e-01 -7.15509415e-01
-1.07681167e+00 -1.13323486e+00 -4.09665227e-01 8.30122113e-01
5.29296100e-01 -1.00832427e+00 -3.93192023e-01 2.27318659e-01] | [10.335018157958984, 3.0913710594177246] |
11c912f4-689e-499e-940f-a950a82559fb | push-pull-characterizing-the-adversarial | 2210.00753 | null | https://arxiv.org/abs/2210.00753v1 | https://arxiv.org/pdf/2210.00753v1.pdf | Push-Pull: Characterizing the Adversarial Robustness for Audio-Visual Active Speaker Detection | Audio-visual active speaker detection (AVASD) is well-developed, and now is an indispensable front-end for several multi-modal applications. However, to the best of our knowledge, the adversarial robustness of AVASD models hasn't been investigated, not to mention the effective defense against such attacks. In this paper, we are the first to reveal the vulnerability of AVASD models under audio-only, visual-only, and audio-visual adversarial attacks through extensive experiments. What's more, we also propose a novel audio-visual interaction loss (AVIL) for making attackers difficult to find feasible adversarial examples under an allocated attack budget. The loss aims at pushing the inter-class embeddings to be dispersed, namely non-speech and speech clusters, sufficiently disentangled, and pulling the intra-class embeddings as close as possible to keep them compact. Experimental results show the AVIL outperforms the adversarial training by 33.14 mAP (%) under multi-modal attacks. | ['Jyh-Shing Roger Jang', 'Hung-Yi Lee', 'Helen Meng', 'Haibin Wu', 'Xuanjun Chen'] | 2022-10-03 | null | null | null | null | ['audio-visual-active-speaker-detection'] | ['computer-vision'] | [-8.92238542e-02 3.13125789e-01 1.42578259e-01 -2.44254693e-02
-1.27318418e+00 -1.04762495e+00 5.92504621e-01 3.93672995e-02
-2.47150198e-01 1.58644661e-01 2.32400328e-01 -5.36647260e-01
3.57887931e-02 -2.73001343e-01 -5.53199410e-01 -8.57892871e-01
-4.57896173e-01 1.52544007e-01 4.05690104e-01 -9.22425389e-02
-8.64842981e-02 7.00218737e-01 -1.09231091e+00 1.31270632e-01
4.62103635e-01 7.76367843e-01 -3.93912256e-01 7.46174395e-01
2.76961297e-01 5.48882842e-01 -9.57795501e-01 -8.20204079e-01
3.18636358e-01 -7.73993880e-02 -3.39104593e-01 -3.27475339e-01
4.25769269e-01 -5.59900522e-01 -8.32042634e-01 1.23255181e+00
9.33017254e-01 4.95869406e-02 5.23920119e-01 -1.76753104e+00
-4.42635715e-01 7.41329908e-01 -7.53475964e-01 3.46366197e-01
3.61970633e-01 4.58698571e-01 1.00299466e+00 -7.47098804e-01
6.64440319e-02 1.57798421e+00 4.07314748e-01 7.72083819e-01
-1.12210238e+00 -1.14695096e+00 4.63320911e-01 1.03974834e-01
-1.39325631e+00 -7.96040118e-01 9.34809208e-01 -4.12852198e-01
6.29600883e-01 6.06796145e-01 2.24303395e-01 1.53046358e+00
-2.71824509e-01 7.97030210e-01 5.71089625e-01 -3.60022873e-01
1.99786335e-01 4.03872341e-01 -2.11241916e-01 3.96791875e-01
6.61267489e-02 3.11045110e-01 -7.44592488e-01 -7.07661092e-01
1.97640195e-01 -2.30306759e-01 -3.33722979e-01 -4.79687095e-01
-7.72522569e-01 1.04076934e+00 2.09014729e-01 -3.78995389e-03
3.89742032e-02 5.52441329e-02 5.30030072e-01 2.47805893e-01
4.86284375e-01 3.19119394e-01 4.14372087e-02 -3.98675464e-02
-7.62210071e-01 1.05544463e-01 4.88377452e-01 6.89740181e-01
1.35590598e-01 4.48783755e-01 -3.47844362e-02 5.59640288e-01
5.76862156e-01 7.33810127e-01 1.13720531e-02 -6.13916874e-01
9.14399862e-01 1.61760196e-01 -1.48358107e-01 -1.12487566e+00
7.02347085e-02 -1.13746487e-01 -7.47093499e-01 6.41765833e-01
5.08710682e-01 -3.83208632e-01 -4.37513560e-01 1.82502699e+00
4.35169697e-01 4.00198340e-01 -6.10922314e-02 8.61064076e-01
5.53265154e-01 7.32703269e-01 4.25204411e-02 -1.66601166e-01
1.09459066e+00 -5.02146900e-01 -6.90072179e-01 -4.41005826e-01
3.10707420e-01 -9.18736637e-01 9.05857503e-01 2.84918934e-01
-1.01891124e+00 -1.48858428e-01 -1.34480107e+00 3.66547346e-01
-2.42147028e-01 -4.65700388e-01 3.64291817e-01 1.23694313e+00
-6.16116047e-01 -1.54637769e-01 -8.61607075e-01 2.49999329e-01
5.07105291e-01 3.69905263e-01 -4.71728384e-01 7.67172426e-02
-1.61785710e+00 5.98051429e-01 -1.14771150e-01 1.86096095e-02
-1.60538101e+00 -8.04942846e-01 -7.75607765e-01 -1.06807455e-01
4.26722586e-01 -2.52503425e-01 1.08471930e+00 -7.14139640e-01
-1.46094763e+00 6.20997846e-01 2.30096266e-01 -6.45729721e-01
7.79375196e-01 -5.98338135e-02 -6.49554312e-01 3.60157788e-01
-2.43399262e-01 2.05597326e-01 1.23529255e+00 -1.51509166e+00
-1.55678719e-01 -5.62173545e-01 3.47020894e-01 1.78125471e-01
-9.37137008e-01 3.30373496e-01 -3.60084087e-01 -8.49416554e-01
-2.25316986e-01 -1.06192970e+00 -2.62671933e-02 3.32491934e-01
-8.14950287e-01 7.90254176e-02 1.24567533e+00 -5.85552275e-01
1.24736047e+00 -2.92129874e+00 -2.08309721e-02 3.04573148e-01
3.99289727e-01 7.27008641e-01 -1.04064740e-01 4.25108820e-01
-2.15684697e-01 3.90315115e-01 -3.15966122e-02 -7.70727277e-01
4.36468691e-01 -1.28844798e-01 -8.66931379e-01 7.44320691e-01
4.21068147e-02 4.25504297e-01 -5.35616398e-01 -3.63232434e-01
3.22202116e-01 6.71564281e-01 -6.01707578e-01 3.19626153e-01
1.92586064e-01 1.68005764e-01 -3.38800490e-01 5.33035338e-01
9.39914227e-01 4.30392444e-01 1.08895218e-02 2.85432097e-02
2.00860158e-01 -5.57583161e-02 -1.21676779e+00 1.06546175e+00
-2.36560747e-01 7.46584952e-01 4.75030273e-01 -8.52679670e-01
6.67869270e-01 5.38258433e-01 2.32174322e-01 -5.08753538e-01
-4.24625091e-02 -9.32383724e-03 -2.21535354e-03 -2.08837897e-01
1.25037596e-01 -5.78758456e-02 -3.72862011e-01 3.56728971e-01
-1.42455906e-01 2.20158055e-01 -5.04291356e-01 7.21660435e-01
9.04184878e-01 -8.47598255e-01 -1.42093629e-01 2.37863347e-01
4.97744977e-01 -6.70954823e-01 4.23387080e-01 8.29008400e-01
-6.30032063e-01 4.86762047e-01 7.33817637e-01 2.19987050e-01
-8.06288362e-01 -1.45190811e+00 2.02679500e-01 1.00288153e+00
1.65146619e-01 -3.00198138e-01 -8.03342640e-01 -1.02533567e+00
2.22302541e-01 6.33213520e-01 -4.85765219e-01 -5.21265984e-01
-4.68986303e-01 -4.18011039e-01 1.31868374e+00 2.66076326e-01
2.45312974e-01 -4.82571185e-01 1.28847480e-01 -1.90648973e-01
7.52118826e-02 -1.19877100e+00 -8.96260142e-01 -2.12515984e-03
-2.02402353e-01 -8.73530090e-01 -6.97702587e-01 -5.67454696e-01
3.73420775e-01 3.14376026e-01 5.58957100e-01 -2.46832460e-01
-2.74782985e-01 4.00440067e-01 -2.59837002e-01 -5.96194446e-01
-6.27969921e-01 -2.35814601e-01 3.43850344e-01 2.40965992e-01
3.93355414e-02 -5.64129770e-01 -4.43280965e-01 5.47895253e-01
-7.60017514e-01 -5.74085176e-01 2.31037289e-01 5.49498081e-01
1.90281674e-01 2.41713449e-01 5.04985631e-01 -4.95638162e-01
4.81917799e-01 -4.56143051e-01 -4.44069922e-01 -1.40833175e-02
-1.10518523e-01 -4.76434827e-01 7.14103043e-01 -9.80285227e-01
-8.23476255e-01 -6.22777306e-02 -3.53518635e-01 -8.28611553e-01
-8.79021641e-03 -7.46151209e-02 -8.00524473e-01 -2.66340554e-01
7.17252433e-01 2.19854135e-02 6.84923201e-04 -3.38952601e-01
4.84087616e-01 8.49045813e-01 3.28254402e-01 -2.55496591e-01
1.40036654e+00 5.61268449e-01 -2.67757118e-01 -1.07656944e+00
-4.70340282e-01 -3.15422654e-01 -1.46762460e-01 -3.28604817e-01
7.38470733e-01 -9.85851467e-01 -8.24670434e-01 6.29450798e-01
-9.75508869e-01 -1.52273834e-01 2.09237053e-03 4.85623598e-01
-1.69624522e-01 8.38109612e-01 -4.02279586e-01 -1.17458534e+00
-2.46090457e-01 -1.30480421e+00 7.60710299e-01 -1.22286148e-01
-5.97355030e-02 -8.09669256e-01 -1.06485203e-01 6.15741968e-01
2.33757630e-01 3.99226964e-01 6.95737302e-01 -9.26690459e-01
-4.77960438e-01 -5.14189482e-01 1.41999006e-01 4.84862924e-01
-1.55716836e-01 1.70041025e-01 -1.36469054e+00 -5.97180903e-01
7.03578666e-02 -2.43875727e-01 6.53805733e-01 1.85348853e-01
1.08820772e+00 -5.95743537e-01 -2.91577011e-01 4.97561097e-01
8.49398792e-01 3.35428923e-01 7.19311476e-01 -4.57138615e-03
5.89887142e-01 4.27684575e-01 4.67850566e-01 4.98755395e-01
-6.23574406e-02 9.80112374e-01 9.75154757e-01 -9.74402800e-02
-9.27300304e-02 -3.66708308e-01 8.24037671e-01 5.27606070e-01
5.66900313e-01 -6.29879892e-01 -7.66987741e-01 4.01050538e-01
-1.33307600e+00 -1.18024814e+00 5.58293462e-02 2.38112926e+00
6.56697035e-01 6.63032591e-01 4.24807489e-01 5.31536520e-01
9.55350339e-01 5.84089100e-01 -5.19343495e-01 -3.67474467e-01
-1.49308190e-01 -1.06147096e-01 3.75227243e-01 7.45965123e-01
-1.42702234e+00 6.60599768e-01 5.56460524e+00 1.15521348e+00
-8.72061372e-01 2.60985136e-01 3.60976338e-01 -4.27527159e-01
-2.40902245e-01 -7.56287277e-02 -7.88295686e-01 5.77114999e-01
8.39116991e-01 -1.49764523e-01 2.32155785e-01 8.52764189e-01
1.43153965e-01 4.96761858e-01 -8.37865114e-01 8.47007215e-01
2.17070878e-01 -9.20282900e-01 2.03926340e-02 3.49739522e-01
9.24751833e-02 -3.24091226e-01 6.96512699e-01 2.20828906e-01
2.89730251e-01 -9.01404440e-01 9.68905687e-01 1.49270240e-03
6.84810102e-01 -1.18625987e+00 3.59818339e-01 3.73542458e-01
-1.13202095e+00 -2.73477525e-01 -4.22096886e-02 5.42106748e-01
2.63323545e-01 4.95680809e-01 -5.68946719e-01 2.89563924e-01
5.85569024e-01 9.95605290e-02 -3.88636202e-01 7.91436017e-01
-1.53009325e-01 8.62637341e-01 -2.65830785e-01 7.29433596e-02
2.44306982e-01 4.74262625e-01 1.16676736e+00 1.01501393e+00
-7.81950483e-04 -2.27684274e-01 1.90619156e-01 5.02136528e-01
-9.95535180e-02 -1.25860065e-01 -7.56584525e-01 -3.36182863e-01
8.10470700e-01 9.05063868e-01 -9.69146937e-02 2.39282757e-01
-1.61217406e-01 9.71738040e-01 1.18806930e-02 3.27600241e-01
-1.16612387e+00 -7.08469450e-01 1.10368466e+00 1.57379836e-01
6.91458955e-02 -3.05209726e-01 3.83186387e-03 -8.90275955e-01
-8.30060840e-02 -1.08850205e+00 4.54390496e-01 -4.02720362e-01
-1.47048509e+00 4.94608730e-01 -9.76615250e-02 -1.30897105e+00
3.37816663e-02 -2.67817497e-01 -6.35461152e-01 8.54677975e-01
-9.88551497e-01 -1.24913383e+00 1.62300169e-01 8.63906503e-01
1.93315744e-01 -5.49483657e-01 8.72849524e-01 6.00011706e-01
-7.93202698e-01 1.55917418e+00 5.78215495e-02 5.90294600e-01
8.19099844e-01 -8.84648323e-01 4.81143326e-01 1.07548130e+00
1.73754662e-01 6.98250532e-01 8.82964492e-01 -1.79012150e-01
-1.39208937e+00 -8.66780460e-01 4.20231998e-01 -2.95942456e-01
9.40223455e-01 -9.11715984e-01 -8.42092752e-01 4.80715662e-01
1.69165477e-01 2.07940508e-02 9.47493434e-01 -1.17091633e-01
-9.80601728e-01 -1.99345633e-01 -1.23768163e+00 6.95158005e-01
7.67271936e-01 -9.91517484e-01 -2.82010466e-01 3.63928080e-01
1.07051528e+00 -2.65629739e-01 -6.72033668e-01 4.36715074e-02
3.16118836e-01 -9.41760242e-01 1.53732324e+00 -5.18905878e-01
-6.46793544e-02 -2.81570375e-01 -4.26379263e-01 -1.05752730e+00
-9.04092006e-03 -1.07017493e+00 -3.93840760e-01 1.75607443e+00
3.95917803e-01 -7.55533993e-01 7.52879858e-01 2.19820797e-01
-2.81175077e-01 -3.13529968e-01 -1.46516430e+00 -9.10323977e-01
9.70661715e-02 -5.95796764e-01 4.54017013e-01 1.03788459e+00
-6.17737547e-02 -1.04973251e-02 -6.08647704e-01 9.33379173e-01
8.25333953e-01 -4.99275744e-01 8.60508204e-01 -7.88899362e-01
-5.44803262e-01 -4.58830506e-01 -6.39556944e-01 -7.93084145e-01
3.14419895e-01 -6.44191623e-01 -2.91981876e-01 -9.13051069e-01
-6.00430071e-02 -4.59523886e-01 -2.14908421e-01 4.15364802e-01
-7.00433180e-02 3.70485365e-01 3.60662162e-01 -1.77969821e-02
-3.52587372e-01 4.92341548e-01 8.42553854e-01 -4.92561400e-01
3.88376154e-02 3.19609135e-01 -8.74641478e-01 6.09242678e-01
6.98671639e-01 -6.89393044e-01 -6.72793806e-01 -1.22004882e-01
-9.38831940e-02 5.03680781e-02 5.69691837e-01 -7.05389619e-01
1.84674338e-01 -7.62315392e-02 -2.92302631e-02 -4.78081048e-01
7.83301055e-01 -9.14437413e-01 2.62994245e-02 4.35847938e-01
-4.13993180e-01 -3.71691436e-01 3.96148235e-01 8.89978886e-01
-2.22291440e-01 6.51894137e-02 1.11905885e+00 3.82915318e-01
-2.10676610e-01 3.31600249e-01 -3.72474790e-01 3.83743495e-01
1.39649415e+00 -8.93327966e-02 -5.52976966e-01 -7.32613981e-01
-7.14198530e-01 1.29019886e-01 6.21910281e-02 5.14169514e-01
7.20739245e-01 -1.31174159e+00 -7.93465018e-01 3.09205621e-01
1.50598109e-01 -3.11181009e-01 8.84520590e-01 4.93716419e-01
-3.22522074e-01 1.24951079e-01 2.77039677e-01 -4.92240787e-01
-1.88870883e+00 8.62331271e-01 4.11835432e-01 -4.24520625e-03
-4.43377733e-01 1.08021235e+00 3.60070229e-01 -2.83355594e-01
8.56831551e-01 6.17753386e-01 5.67854941e-02 2.27752939e-01
7.19549298e-01 3.51458460e-01 -6.01629354e-02 -8.39898884e-01
-5.98512292e-01 2.12031484e-01 -2.06913888e-01 -3.68361741e-01
8.88304293e-01 -3.10202260e-02 2.66493976e-01 1.27979264e-01
1.44181633e+00 5.52875757e-01 -1.18315232e+00 -1.55936867e-01
-4.55266237e-01 -7.20782638e-01 -1.18587427e-02 -4.72448558e-01
-1.13757765e+00 1.36010587e+00 7.95028150e-01 6.80377662e-01
9.49498653e-01 8.25706199e-02 6.34342909e-01 -1.01984680e-01
1.58632975e-02 -6.93530917e-01 6.01127863e-01 1.54736981e-01
9.57715869e-01 -6.87235355e-01 -1.58783361e-01 -4.83628124e-01
-9.61466551e-01 5.90397716e-01 5.05484283e-01 9.23906937e-02
7.76210368e-01 4.21265125e-01 1.21888570e-01 1.94398314e-01
-5.09347081e-01 2.47513980e-01 2.30771884e-01 8.41879368e-01
-2.01657370e-01 2.25718506e-02 6.16747320e-01 5.42022109e-01
-5.47054745e-02 -1.00124133e+00 3.61012042e-01 6.27353609e-01
-2.97853857e-01 -9.02006745e-01 -7.80217826e-01 -4.82265390e-02
-6.71320498e-01 -5.61959259e-02 -5.83035052e-01 7.59701669e-01
-7.63250440e-02 1.38285613e+00 -1.68257162e-01 -7.28912950e-01
5.07049263e-01 -8.17855448e-02 5.19171804e-02 -3.45192850e-01
-5.26813090e-01 1.38694853e-01 7.35507533e-02 -1.27508759e-01
9.28106606e-02 -6.22274876e-01 -8.57197762e-01 -7.32553244e-01
-4.15613651e-01 1.46148846e-01 5.14707506e-01 4.74326879e-01
2.43599027e-01 3.94727468e-01 1.23580515e+00 -7.34254360e-01
-9.33297992e-01 -6.60154641e-01 -5.86942136e-01 2.55410671e-01
5.82004130e-01 -6.12936676e-01 -1.21757388e+00 -5.22855997e-01] | [14.011585235595703, 5.8440656661987305] |
8b1142a0-3612-49c8-a49e-25e8e07238d7 | cnn-based-patch-matching-for-optical-flow | 1607.08064 | null | https://arxiv.org/abs/1607.08064v4 | https://arxiv.org/pdf/1607.08064v4.pdf | CNN-based Patch Matching for Optical Flow with Thresholded Hinge Embedding Loss | Learning based approaches have not yet achieved their full potential in optical flow estimation, where their performance still trails heuristic approaches. In this paper, we present a CNN based patch matching approach for optical flow estimation. An important contribution of our approach is a novel thresholded loss for Siamese networks. We demonstrate that our loss performs clearly better than existing losses. It also allows to speed up training by a factor of 2 in our tests. Furthermore, we present a novel way for calculating CNN based features for different image scales, which performs better than existing methods. We also discuss new ways of evaluating the robustness of trained features for the application of patch matching for optical flow. An interesting discovery in our paper is that low-pass filtering of feature maps can increase the robustness of features created by CNNs. We proved the competitive performance of our approach by submitting it to the KITTI 2012, KITTI 2015 and MPI-Sintel evaluation portals where we obtained state-of-the-art results on all three datasets. | ['Kiran varanasi', 'Christian Bailer', 'Didier Stricker'] | 2016-07-27 | cnn-based-patch-matching-for-optical-flow-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Bailer_CNN-Based_Patch_Matching_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Bailer_CNN-Based_Patch_Matching_CVPR_2017_paper.pdf | cvpr-2017-7 | ['patch-matching'] | ['computer-vision'] | [-3.54176790e-01 -4.39800054e-01 4.22327556e-02 -1.53318802e-02
-5.34676909e-01 -4.98041809e-01 6.45623803e-01 -6.53284788e-02
-6.62351727e-01 9.18716133e-01 1.86822549e-01 -3.38307954e-02
-2.19381437e-01 -7.52239168e-01 -6.41839325e-01 -4.65320379e-01
-5.16935766e-01 1.56092778e-01 5.70291817e-01 -2.60276377e-01
5.69574893e-01 8.42203498e-01 -1.51417685e+00 2.03640446e-01
5.27978837e-01 9.35107708e-01 -1.75906509e-01 1.07558501e+00
2.70906061e-01 9.54477608e-01 -4.90531981e-01 -3.58715475e-01
7.57897198e-01 -4.70191896e-01 -1.10760880e+00 -2.76677042e-01
1.33917212e+00 -4.72397000e-01 -6.60062373e-01 6.20677888e-01
5.97344637e-01 5.61539650e-01 5.01187623e-01 -1.22858512e+00
-1.57222301e-01 1.00083113e-01 -1.94497973e-01 6.84524596e-01
1.96732447e-01 3.38245481e-01 9.99258518e-01 -8.31174076e-01
1.02255869e+00 1.19437635e+00 7.68129587e-01 4.11951870e-01
-1.22212255e+00 -4.25009459e-01 -2.08897993e-01 5.65695286e-01
-1.02898133e+00 -5.84798872e-01 4.97790456e-01 -5.62949717e-01
1.20008481e+00 7.02182874e-02 6.87976956e-01 5.71088195e-01
9.50561240e-02 7.65287817e-01 1.10456288e+00 -3.30387682e-01
-1.43630892e-01 -1.78342164e-02 -2.47248396e-01 1.02104998e+00
-7.02803209e-03 5.01355886e-01 -5.69209278e-01 6.96026608e-02
1.07759082e+00 -4.61388290e-01 -3.72385532e-01 -3.86687189e-01
-1.43536878e+00 9.14569020e-01 8.30393553e-01 4.70323354e-01
-1.83115453e-02 5.57361722e-01 5.77347517e-01 4.15376574e-01
5.98991573e-01 6.25579417e-01 -3.90239447e-01 -3.75111789e-01
-1.23562241e+00 5.85789740e-01 7.79684663e-01 5.16167879e-01
9.78465021e-01 -1.19303428e-01 -4.18891639e-01 5.37824333e-01
-6.59128502e-02 1.50275245e-01 1.57539189e-01 -1.68644679e+00
3.75658959e-01 -1.15042869e-02 1.26969412e-01 -1.16668475e+00
-4.79143173e-01 -4.58687097e-01 -6.55532300e-01 6.27455711e-01
9.94742930e-01 -1.41961142e-01 -5.26694179e-01 1.56339645e+00
1.80186685e-02 7.14852631e-01 -1.15888610e-01 1.00209844e+00
5.24969578e-01 3.61298352e-01 -3.50414902e-01 9.44280103e-02
6.91236973e-01 -1.23970056e+00 -2.99020529e-01 3.32599461e-01
7.02847064e-01 -9.97877777e-01 8.18265915e-01 4.52688783e-01
-1.34116459e+00 -6.22807443e-01 -9.19351578e-01 -2.23619819e-01
-1.93835035e-01 1.41028361e-02 8.67461205e-01 5.61318994e-01
-1.48836851e+00 1.41173756e+00 -7.77169645e-01 -4.73214656e-01
7.03786135e-01 4.80919540e-01 -5.02712846e-01 1.85769126e-01
-8.79544139e-01 9.78134036e-01 -1.49634093e-01 2.82953493e-02
-6.89134419e-01 -1.04451835e+00 -7.28409410e-01 1.18191138e-01
-1.18894000e-02 -7.70184338e-01 1.04482412e+00 -7.53077209e-01
-1.65246499e+00 8.32441330e-01 -1.75404325e-01 -8.45155895e-01
9.50517893e-01 -3.25054377e-02 1.09336451e-01 4.76059854e-01
-1.43516287e-02 1.01362586e+00 7.86791563e-01 -7.45458007e-01
-6.52070403e-01 -1.03411004e-02 2.60569900e-01 -1.21503875e-01
-1.58847377e-01 -1.11909829e-01 -3.26778963e-02 -5.54226339e-01
-3.17513227e-01 -8.28398764e-01 -2.49457836e-01 3.99146765e-01
-8.44837502e-02 -1.25314921e-01 5.41303396e-01 -3.51446539e-01
8.26274574e-01 -1.82532930e+00 2.95881275e-02 8.71977061e-02
3.63692313e-01 5.52350283e-01 -3.51856321e-01 4.26715195e-01
6.53344989e-02 6.90174699e-02 -1.65019631e-01 -5.80532312e-01
-1.08998463e-01 9.71053243e-02 -2.90026188e-01 7.54460156e-01
4.15514886e-01 7.98619926e-01 -9.13038969e-01 -5.55957019e-01
5.78258753e-01 7.97250092e-01 -8.43549311e-01 1.39094121e-03
2.35164508e-01 6.46961510e-01 1.51231617e-01 2.36671507e-01
6.73056781e-01 -8.96875486e-02 -4.09933925e-01 -2.36728311e-01
-3.59632552e-01 4.41676438e-01 -1.18262768e+00 1.67645085e+00
-4.81574446e-01 1.37330496e+00 -2.35580221e-01 -9.79684591e-01
5.75792789e-01 1.68812126e-01 6.82906866e-01 -6.93931043e-01
-5.51668629e-02 3.15693021e-01 5.16540632e-02 -3.21933240e-01
3.50772023e-01 -4.60284539e-02 7.71215677e-01 2.57825047e-01
4.23235774e-01 -2.08461985e-01 7.00528622e-01 1.51570499e-01
1.26318705e+00 2.19735011e-01 -1.08337469e-01 -4.20495003e-01
7.96797991e-01 -7.98309892e-02 3.02988499e-01 8.30834150e-01
-6.51929915e-01 8.20351958e-01 5.94969869e-01 -8.16242039e-01
-9.57486093e-01 -9.34616387e-01 -3.17743361e-01 6.24076903e-01
-9.58634689e-02 -5.08575022e-01 -5.32583714e-01 -8.78072143e-01
8.78025070e-02 -1.30687281e-01 -8.85148346e-01 2.53189772e-01
-8.49682331e-01 -4.91674662e-01 7.02388406e-01 4.75237876e-01
6.59233749e-01 -1.07187355e+00 -6.80218220e-01 2.22857967e-01
-1.60036579e-01 -1.33287501e+00 -3.80730629e-01 -2.48896375e-01
-1.12286294e+00 -1.30771232e+00 -8.18817437e-01 -5.59220195e-01
3.46489727e-01 6.43627569e-02 1.20282304e+00 4.03141737e-01
-5.09181857e-01 3.22443604e-01 -2.03738704e-01 -6.09740056e-02
-1.56426921e-01 3.31396013e-01 -2.23960772e-01 -1.21071398e-01
-1.21321388e-01 -6.51522994e-01 -9.82772589e-01 2.43010506e-01
-6.92284882e-01 -4.61067408e-01 2.12490737e-01 7.30846882e-01
7.69525766e-02 -3.54916126e-01 2.35906690e-01 -6.38179123e-01
4.18483198e-01 3.06002777e-02 -7.06030071e-01 -3.78827393e-01
-6.17555857e-01 1.82533592e-01 6.62505865e-01 -6.81701452e-02
-5.25950193e-01 1.77096725e-02 -2.94433177e-01 -4.98875022e-01
7.42540142e-05 4.35938388e-02 7.34378994e-01 -9.13465679e-01
7.50238657e-01 -2.12656811e-01 2.99138427e-01 -2.74149179e-01
1.90134585e-01 7.17098964e-03 4.14980829e-01 -3.75960857e-01
7.86119699e-01 8.78270805e-01 6.34760737e-01 -8.17496777e-01
-5.47854781e-01 -6.54894412e-01 -6.96768880e-01 -2.78726369e-01
5.27706563e-01 -6.02898657e-01 -1.09616852e+00 5.23472071e-01
-1.26219392e+00 -5.12909353e-01 -3.13397586e-01 6.59540236e-01
-8.62748802e-01 4.96762753e-01 -9.28039551e-01 -4.32731807e-01
-9.15090591e-02 -1.27563655e+00 8.17889869e-01 1.70007467e-01
-2.58733127e-02 -1.51905251e+00 5.09472489e-01 1.91806629e-01
8.94819021e-01 2.50280976e-01 3.43328416e-01 -1.66885048e-01
-7.50519395e-01 1.24942241e-02 -4.37552780e-01 5.08759856e-01
-2.45516896e-01 4.40482169e-01 -1.16641462e+00 -3.85195225e-01
-4.42383170e-01 -2.15561166e-01 1.37764621e+00 5.14620245e-01
8.72943103e-01 2.78364830e-02 -2.01735664e-02 9.36894834e-01
1.61253250e+00 -3.49091917e-01 1.06029999e+00 4.98421580e-01
4.88757104e-01 6.28799796e-01 3.21550459e-01 2.17776671e-01
1.49239033e-01 8.68870854e-01 5.04352868e-01 -2.74126977e-01
-5.72504163e-01 1.90931678e-01 2.42102310e-01 4.46731925e-01
-4.61241007e-01 -8.68445635e-02 -5.72476029e-01 6.92121387e-01
-1.86674774e+00 -1.17301190e+00 -5.09786367e-01 2.25517988e+00
4.79566634e-01 7.53963068e-02 3.56526881e-01 1.84739083e-01
3.76318485e-01 9.35075432e-02 2.24369653e-02 -5.49405634e-01
-4.97095697e-02 7.59255230e-01 6.14837229e-01 1.00268590e+00
-1.27364731e+00 1.02079523e+00 6.97402716e+00 4.60572213e-01
-1.09935260e+00 1.05229907e-01 4.19982702e-01 -1.66517422e-01
3.50219578e-01 1.19771093e-01 -6.10038638e-01 3.62339586e-01
6.72264576e-01 8.30787271e-02 3.24657470e-01 3.44673783e-01
3.22725594e-01 -3.01832467e-01 -9.92555201e-01 9.41378415e-01
2.24494096e-02 -1.86967623e+00 -2.21244991e-01 2.55718827e-01
8.67826164e-01 4.35255170e-01 -3.24929343e-03 2.99062692e-02
-2.30777580e-02 -1.04568970e+00 2.16915742e-01 5.51164865e-01
3.81464750e-01 -6.25739038e-01 7.91848063e-01 -2.73250520e-01
-1.03216660e+00 1.74157679e-01 -4.82872218e-01 -3.32121015e-01
-5.78261390e-02 6.78379536e-01 -6.26434326e-01 5.36375403e-01
5.80566883e-01 1.20977986e+00 -7.87567735e-01 1.85208952e+00
5.16260415e-03 5.46168208e-01 -3.76498163e-01 7.13312179e-02
5.15535593e-01 -4.73432578e-02 7.04741895e-01 1.36872196e+00
6.60908520e-02 -5.73026061e-01 -1.04584850e-01 8.36168110e-01
-7.50524178e-02 2.56884724e-01 -5.73807836e-01 4.82824326e-01
-5.22467270e-02 1.32561100e+00 -6.70240760e-01 -2.32347324e-01
-3.39290261e-01 8.99925888e-01 5.03905714e-01 2.68018186e-01
-4.56324518e-01 -6.99070752e-01 8.55753064e-01 1.87257141e-01
4.64710355e-01 -3.48047018e-01 -1.30508561e-02 -1.29854333e+00
4.63414900e-02 -3.25864911e-01 1.65046245e-01 -5.53259015e-01
-1.17837238e+00 4.73000616e-01 -1.43202901e-01 -1.23721683e+00
-3.33793163e-01 -9.06374872e-01 -7.50622869e-01 8.43051016e-01
-1.88040709e+00 -6.67505205e-01 -3.84446502e-01 6.61445260e-01
2.09232584e-01 -4.60205823e-02 5.75116634e-01 5.84939063e-01
-2.19908088e-01 3.72522801e-01 -3.03285979e-02 4.00284290e-01
1.06812429e+00 -1.31055713e+00 5.42973995e-01 1.03159225e+00
3.93398017e-01 3.41735542e-01 7.32497036e-01 -7.51222968e-02
-8.91746879e-01 -7.04101205e-01 1.06343818e+00 -6.18812859e-01
7.06115186e-01 -1.10647343e-01 -7.73497224e-01 6.08777404e-01
3.33538443e-01 4.71123070e-01 2.54316121e-01 -9.47510302e-02
-3.01083952e-01 -1.99612945e-01 -1.06775153e+00 2.66552240e-01
9.69378114e-01 -3.24039102e-01 -1.27556160e-01 3.96837562e-01
2.72706985e-01 -4.86803889e-01 -6.69789970e-01 3.63532841e-01
6.30172789e-01 -1.64691782e+00 1.05015194e+00 -6.30442262e-01
6.61478400e-01 -1.11950293e-01 1.81680053e-01 -1.33095407e+00
-1.47951961e-01 -8.62185597e-01 -2.81525292e-02 7.07560778e-01
2.36927241e-01 -8.25012743e-01 9.89580929e-01 1.12855494e-01
-4.22018953e-02 -5.21592736e-01 -1.03027534e+00 -8.35263073e-01
4.77826864e-01 -4.98129696e-01 1.44564673e-01 7.41346419e-01
-1.44225076e-01 2.18178355e-03 -1.96192294e-01 -3.27442378e-01
7.04653561e-01 -7.59363472e-02 8.88568699e-01 -1.16388607e+00
-2.42315531e-01 -8.89349878e-01 -1.17221057e+00 -8.91127408e-01
2.24793747e-01 -8.79307985e-01 -3.29698443e-01 -1.44364822e+00
-1.62633568e-01 -3.32632363e-01 -2.59120971e-01 2.52274662e-01
7.18102902e-02 8.52931499e-01 5.53722680e-01 1.30207151e-01
-4.57815766e-01 1.72691941e-01 1.63162601e+00 1.13324635e-01
-4.70675305e-02 6.13592751e-02 -3.07374131e-02 3.01554739e-01
8.65898192e-01 -2.88871139e-01 -1.91847771e-01 -4.44524467e-01
1.37007415e-01 -1.73187777e-01 9.12651718e-01 -1.44192696e+00
2.49184027e-01 2.23522916e-01 2.31472090e-01 -2.28394210e-01
2.54230261e-01 -4.18404073e-01 -5.46932220e-01 5.56274951e-01
-3.51245105e-01 4.52314243e-02 2.67664462e-01 7.94053674e-02
-3.37568849e-01 -1.71673208e-01 9.75457013e-01 -1.36100590e-01
-5.64858794e-01 4.68387842e-01 -2.84431756e-01 5.06035328e-01
5.87322772e-01 -1.05674744e-01 -4.04515326e-01 -5.23363948e-01
-6.27281547e-01 -5.33999391e-02 3.36755514e-01 2.46250138e-01
4.37974066e-01 -1.11288226e+00 -9.05734777e-01 3.64939690e-01
-1.80697918e-01 -4.63092417e-01 -4.83826622e-02 1.24134624e+00
-1.24310279e+00 5.39547980e-01 -5.67161024e-01 -6.90119147e-01
-1.15341485e+00 2.07352236e-01 8.26345026e-01 -2.80711830e-01
-6.36216521e-01 8.09726894e-01 -1.55363128e-01 -2.56554842e-01
2.85922647e-01 -3.68139416e-01 -1.32501023e-02 -1.31928712e-01
6.51377380e-01 7.44127095e-01 3.79424125e-01 -5.91731429e-01
-3.94647390e-01 8.27508092e-01 1.58088833e-01 -2.83651263e-01
1.22932994e+00 9.51295644e-02 -1.91734970e-01 1.54404685e-01
1.66539192e+00 7.60077452e-03 -1.49236643e+00 7.41099417e-02
-2.60908812e-01 -9.45949078e-01 2.08604127e-01 -5.75750649e-01
-1.38452947e+00 1.25528800e+00 6.58335745e-01 1.69657305e-01
8.32924724e-01 -2.93759853e-01 8.45128596e-01 3.62537444e-01
1.29283816e-01 -8.58164132e-01 5.96920960e-02 6.60953701e-01
5.94347537e-01 -1.45731676e+00 -6.36996329e-02 -3.81330580e-01
-7.76128750e-03 1.47691369e+00 4.13701862e-01 -7.16087162e-01
6.04978740e-01 1.93922445e-01 2.00135082e-01 -5.15934676e-02
-6.69771671e-01 -6.52379930e-01 4.04521376e-01 7.32446313e-01
5.01737118e-01 -3.72912705e-01 -2.92826891e-01 -6.71349525e-01
-2.22775400e-01 1.68663800e-01 7.01883137e-01 7.44837403e-01
-2.68294871e-01 -1.28899717e+00 -2.29037162e-02 2.47019917e-01
-6.84689164e-01 -2.03936771e-01 -3.00431818e-01 1.10875142e+00
-1.54951185e-01 7.90658951e-01 2.12958038e-01 -1.94185510e-01
3.54875594e-01 -1.94866195e-01 9.87312198e-01 -2.54419565e-01
-8.84973347e-01 -3.04474205e-01 -1.12746097e-01 -1.06839943e+00
-7.81873465e-01 -6.20828986e-01 -9.88239408e-01 -6.49064839e-01
1.92678664e-02 3.82600836e-02 6.78621769e-01 8.19689214e-01
2.60307580e-01 2.35881716e-01 6.10367060e-01 -1.19022977e+00
-1.47319511e-01 -6.56394839e-01 -2.95377433e-01 4.50711727e-01
6.93745077e-01 -7.27157891e-01 -7.31707692e-01 -4.71740179e-02] | [8.809144020080566, -1.8217999935150146] |
e9c5b3be-8da8-4850-ba01-816e2764aa90 | one-time-of-interaction-may-not-be-enough-go | null | null | https://aclanthology.org/P19-1001 | https://aclanthology.org/P19-1001.pdf | One Time of Interaction May Not Be Enough: Go Deep with an Interaction-over-Interaction Network for Response Selection in Dialogues | Currently, researchers have paid great attention to retrieval-based dialogues in open-domain. In particular, people study the problem by investigating context-response matching for multi-turn response selection based on publicly recognized benchmark data sets. State-of-the-art methods require a response to interact with each utterance in a context from the beginning, but the interaction is performed in a shallow way. In this work, we let utterance-response interaction go deep by proposing an interaction-over-interaction network (IoI). The model performs matching by stacking multiple interaction blocks in which residual information from one time of interaction initiates the interaction process again. Thus, matching information within an utterance-response pair is extracted from the interaction of the pair in an iterative fashion, and the information flows along the chain of the blocks via representations. Evaluation results on three benchmark data sets indicate that IoI can significantly outperform state-of-the-art methods in terms of various matching metrics. Through further analysis, we also unveil how the depth of interaction affects the performance of IoI. | ['Chongyang Tao', 'Wenpeng Hu', 'Wei Wu', 'Rui Yan', 'Dongyan Zhao', 'Can Xu'] | 2019-07-01 | null | null | null | acl-2019-7 | ['conversational-response-selection'] | ['natural-language-processing'] | [ 5.16775489e-01 2.33216375e-01 -8.13521668e-02 -7.39480615e-01
-1.08742702e+00 -4.53746796e-01 6.89137518e-01 8.94531906e-02
-4.32780057e-01 3.25209022e-01 7.97412455e-01 -2.71398425e-01
-3.77028994e-02 -4.43354666e-01 -1.40134931e-01 -3.73070121e-01
3.11477661e-01 6.17677450e-01 1.27766997e-01 -6.86044335e-01
5.98003924e-01 -7.33528808e-02 -1.05316722e+00 8.06908906e-01
7.79319465e-01 7.55901039e-01 6.62083998e-02 4.93189096e-01
-4.46863949e-01 8.37264895e-01 -6.45486116e-01 -3.41063499e-01
6.74147019e-03 -7.93775618e-01 -1.56370795e+00 -1.38545454e-01
7.17296228e-02 -3.63604933e-01 -2.85979420e-01 6.04485691e-01
6.68992937e-01 6.92912936e-01 4.75777656e-01 -7.85044193e-01
-5.04289508e-01 8.16251993e-01 -1.93384677e-01 7.26407720e-03
1.06016266e+00 1.45754650e-01 1.44974220e+00 -1.07311094e+00
5.90199351e-01 1.45502245e+00 8.72474387e-02 7.51111269e-01
-1.09321451e+00 -4.33957666e-01 4.10359502e-01 2.31372237e-01
-9.16794658e-01 -5.61558485e-01 8.63374412e-01 -2.77044028e-01
1.02260268e+00 5.57596564e-01 1.05323806e-01 7.90241539e-01
-2.50098348e-01 1.19105959e+00 6.42333925e-01 -5.46466053e-01
-1.17035776e-01 -9.44997817e-02 8.24888527e-01 2.50443608e-01
-8.00439894e-01 -2.38713905e-01 -6.06310368e-01 -1.97761759e-01
2.89008737e-01 -1.53874993e-01 -3.24633777e-01 2.95128405e-01
-1.01159227e+00 9.69284594e-01 5.09144366e-01 3.95659477e-01
-2.83042192e-01 -3.79065484e-01 6.77641034e-01 7.73154378e-01
5.97188473e-01 5.12950540e-01 -2.59264350e-01 -1.63242146e-01
-5.57044864e-01 4.31985289e-01 1.15782452e+00 5.93483090e-01
6.89780772e-01 -6.04452610e-01 -7.29800880e-01 1.36890805e+00
2.70707071e-01 -1.73019335e-01 5.02779245e-01 -7.19781399e-01
8.61631930e-01 9.27647352e-01 7.08595365e-02 -9.63190019e-01
-5.85443377e-01 7.60101900e-02 -7.70886183e-01 -5.34687340e-01
6.64318621e-01 -1.43204674e-01 -3.13396394e-01 1.67962515e+00
5.79991996e-01 -3.38417262e-01 2.77880162e-01 1.29476619e+00
1.37340999e+00 7.62620032e-01 -1.45275429e-01 -1.89124644e-01
1.41271186e+00 -1.37772882e+00 -7.94940472e-01 -2.05444336e-01
7.62521386e-01 -8.42362881e-01 1.20427942e+00 6.64690360e-02
-1.08270824e+00 -5.70285916e-01 -6.10097408e-01 -2.54277408e-01
7.69007057e-02 1.02234706e-01 3.34403157e-01 -4.69566993e-02
-9.10568535e-01 3.23206156e-01 -4.95348014e-02 -2.79209971e-01
-3.91867846e-01 2.55021691e-01 -3.10321242e-01 -8.66097510e-02
-1.81285203e+00 9.40342367e-01 -9.22413543e-02 3.65893871e-01
-5.96659899e-01 -3.48792255e-01 -7.59414554e-01 8.34296867e-02
5.69204986e-01 -4.53018337e-01 1.74899662e+00 -9.30952847e-01
-2.01728034e+00 7.31181562e-01 -3.55698168e-01 -2.45855808e-01
3.30279320e-01 -3.80417466e-01 -2.04242796e-01 2.51404643e-01
-1.37367070e-01 4.33465838e-01 5.42788386e-01 -1.02901661e+00
-4.79963273e-01 -4.16462839e-01 6.20914161e-01 6.21753395e-01
-1.54228816e-02 4.67716992e-01 -6.85034156e-01 -1.34607673e-01
2.86093980e-01 -1.17991817e+00 -1.89910904e-01 -6.50866687e-01
-4.48368788e-01 -1.03316987e+00 4.45297599e-01 -6.27236605e-01
1.67582166e+00 -2.21630573e+00 4.55132514e-01 -1.33461684e-01
2.74282575e-01 2.68339515e-01 -4.37975854e-01 9.47479248e-01
-6.54342622e-02 8.07210058e-02 4.47421074e-02 -3.62160265e-01
5.43292053e-02 -1.63042530e-01 -2.53390372e-01 2.56107599e-01
9.81739908e-03 7.34113753e-01 -1.05808568e+00 -3.44017863e-01
-9.63619202e-02 1.04267402e-02 -7.00701892e-01 8.91477168e-01
-3.21926326e-01 7.78516471e-01 -6.95358157e-01 1.81929022e-01
4.35212314e-01 -2.44247019e-01 3.25551420e-01 -9.77132022e-02
-9.86561403e-02 9.78075922e-01 -7.06766427e-01 1.81106567e+00
-7.88058579e-01 5.16371310e-01 1.17432468e-01 -8.63527060e-01
1.03369987e+00 4.00460929e-01 2.59817064e-01 -8.30440521e-01
-8.67143050e-02 1.24856293e-01 3.23230803e-01 -5.40527284e-01
8.17519844e-01 2.92078108e-01 -5.00395656e-01 8.41920495e-01
-2.08189756e-01 7.83285573e-02 9.59697440e-02 5.38294315e-01
1.07273662e+00 -1.60838783e-01 1.45466477e-01 4.90498058e-02
9.14180338e-01 -3.18729401e-01 2.52578199e-01 8.15179348e-01
-1.18403189e-01 4.70684767e-01 4.89732534e-01 -4.05378491e-01
-4.21590686e-01 -5.17015398e-01 2.16103733e-01 1.72579122e+00
1.92042172e-01 -4.21875238e-01 -8.36610436e-01 -9.00386274e-01
-2.92629451e-01 5.33104599e-01 -5.55796921e-01 -1.71282098e-01
-8.01047921e-01 -1.85168922e-01 4.10516739e-01 1.59132197e-01
6.72497928e-01 -1.37742400e+00 -3.41585547e-01 4.05943513e-01
-8.14399719e-01 -9.40308690e-01 -8.65141928e-01 -4.83018868e-02
-3.91028017e-01 -8.21258783e-01 -5.79941988e-01 -6.76948905e-01
5.24544835e-01 5.33097088e-01 1.16041636e+00 3.70331645e-01
1.65412873e-01 1.71350643e-01 -9.55125451e-01 2.29372904e-01
-3.80300164e-01 4.57139820e-01 -2.29585350e-01 3.18864852e-01
4.16949630e-01 -1.62809253e-01 -6.53344572e-01 7.29158461e-01
-7.19108820e-01 1.22630447e-01 3.30292284e-01 1.05282831e+00
-1.99238062e-02 -5.52836955e-01 6.62746310e-01 -9.44384277e-01
1.09552193e+00 -6.24116480e-01 -1.20338917e-01 3.27952206e-01
-1.73231661e-01 6.48524761e-02 5.33047557e-01 -2.50338435e-01
-1.36124778e+00 -2.84853876e-01 -2.31827870e-01 6.64374977e-02
1.03787640e-02 7.86896884e-01 -2.67072469e-01 3.64167392e-01
4.96683270e-01 2.90004592e-02 -7.52752200e-02 -4.98386711e-01
3.75179261e-01 1.11808264e+00 1.80186957e-01 -6.36798501e-01
1.35257080e-01 -3.33926231e-02 -6.44348204e-01 -6.52260959e-01
-9.77990746e-01 -8.65006924e-01 -5.46861291e-01 -4.70853746e-01
7.09845483e-01 -5.26091158e-01 -9.35894310e-01 4.93575543e-01
-1.47894442e+00 -3.38090897e-01 2.75748789e-01 2.61409044e-01
-2.06470743e-01 1.91480070e-01 -7.37776577e-01 -9.07801151e-01
-4.06621218e-01 -1.38940597e+00 8.97204459e-01 3.36281240e-01
-6.04377210e-01 -6.96661532e-01 1.26299217e-01 6.91658914e-01
2.44242385e-01 -4.35973048e-01 8.33750546e-01 -1.33525026e+00
-3.55613291e-01 -2.38364533e-01 -1.82956234e-01 9.57996994e-02
1.16228737e-01 -2.17881009e-01 -9.21814263e-01 -4.59259972e-02
3.35033126e-02 -5.33919394e-01 8.45740855e-01 2.20039487e-02
9.15629804e-01 -4.45769757e-01 -1.53599709e-01 -1.14872437e-02
7.48261213e-01 3.52640182e-01 5.89296639e-01 1.87267289e-01
5.94107032e-01 1.33165526e+00 8.63345563e-01 2.56885409e-01
6.69210911e-01 9.98988450e-01 1.58325478e-01 -4.02535126e-02
1.90227330e-01 -1.49750814e-01 2.57764608e-01 9.99504030e-01
4.48531955e-01 -4.79741544e-01 -7.98259795e-01 4.17919904e-01
-2.13224006e+00 -9.03688908e-01 -1.60524070e-01 2.26301432e+00
1.06740999e+00 -1.52362222e-02 9.12812352e-02 -1.89693440e-02
7.58659720e-01 3.82421643e-01 -4.25412297e-01 -7.69192159e-01
3.58765483e-01 -2.07060352e-01 -2.72902429e-01 8.53931665e-01
-7.17856944e-01 1.07983923e+00 5.58668423e+00 6.53011203e-01
-8.36196065e-01 -4.47620638e-02 6.39260769e-01 1.22462094e-01
-3.94513369e-01 2.60035008e-01 -9.17042673e-01 2.53217101e-01
9.17210102e-01 5.85297728e-03 4.88826126e-01 4.46262568e-01
2.57480949e-01 -3.29660177e-01 -1.37541664e+00 5.73178470e-01
7.43559673e-02 -1.02534294e+00 -9.65582728e-02 -2.56740034e-01
2.04675302e-01 -2.46817023e-01 -1.72782555e-01 6.45248652e-01
3.94632339e-01 -7.93098092e-01 1.95367396e-01 5.55550992e-01
4.66988444e-01 -5.80298543e-01 7.07591236e-01 5.58278143e-01
-1.09715903e+00 3.99293825e-02 -2.89693139e-02 -2.19953761e-01
1.48797214e-01 3.50605994e-01 -9.75540280e-01 8.44663262e-01
3.70768577e-01 3.34181547e-01 -1.96384043e-01 5.43483615e-01
-3.41173440e-01 4.98336524e-01 -5.81469722e-02 -2.80574411e-01
4.34800327e-01 -3.45911443e-01 4.42752779e-01 1.32400489e+00
-3.13905984e-01 5.30105948e-01 5.70367515e-01 5.20554066e-01
-1.60674259e-01 3.64340097e-01 -4.02463555e-01 -1.49264354e-02
6.86852574e-01 1.17574215e+00 -1.84290886e-01 -3.12062949e-01
-3.55411470e-01 1.02391875e+00 5.80855310e-01 3.56135309e-01
-5.19958615e-01 -5.09160459e-01 5.62840819e-01 -2.57723957e-01
-1.64398611e-01 1.44050628e-01 3.15245613e-02 -9.42468584e-01
1.02499060e-01 -1.10357392e+00 4.51714069e-01 -4.76136655e-01
-1.26509941e+00 8.01863492e-01 -2.31349207e-02 -1.16070414e+00
-6.57887042e-01 7.31645450e-02 -6.97722971e-01 1.12731659e+00
-1.17355943e+00 -8.65592301e-01 -8.05546939e-02 3.37107539e-01
1.13500333e+00 1.05059378e-01 9.58903968e-01 1.63524881e-01
-6.93761051e-01 8.13670099e-01 -2.05558851e-01 4.41509008e-01
8.49485040e-01 -8.54543209e-01 4.71972913e-01 4.37847406e-01
-1.51882060e-02 9.54349279e-01 4.99009192e-01 -3.63100559e-01
-1.26598215e+00 -4.72430617e-01 1.30639565e+00 -2.50932515e-01
4.92570877e-01 -2.76221693e-01 -1.13224506e+00 3.02026957e-01
5.85085034e-01 -3.54343414e-01 8.88895392e-01 4.58550692e-01
-3.05425584e-01 9.78151858e-02 -8.09018731e-01 6.15761340e-01
7.96401203e-01 -8.98598373e-01 -8.42716634e-01 2.72075772e-01
1.02921855e+00 -6.87086582e-01 -6.49105847e-01 3.15355510e-01
6.38223588e-01 -8.41534197e-01 7.42780209e-01 -7.67535627e-01
6.54651761e-01 1.53320059e-02 -9.15034935e-02 -1.40781116e+00
-1.91046998e-01 -9.62529063e-01 2.39171699e-01 1.46046686e+00
5.14567673e-01 -4.09560561e-01 2.05347732e-01 8.91763389e-01
-2.46082887e-01 -9.44032013e-01 -6.50382519e-01 -2.22344786e-01
-1.91408722e-03 -1.05531834e-01 6.51307344e-01 7.09598303e-01
6.36259317e-01 9.65495944e-01 -4.06805366e-01 -1.71940252e-01
-5.79630621e-02 3.85816842e-01 1.06804204e+00 -1.05759442e+00
-2.84221590e-01 -3.82695645e-01 2.56105572e-01 -1.69883978e+00
4.92778152e-01 -7.54961908e-01 2.93055594e-01 -1.51621902e+00
2.69095212e-01 -3.66525382e-01 -2.91521028e-02 3.55140835e-01
-6.12149894e-01 -3.58447254e-01 3.96025538e-01 5.40858686e-01
-8.16607714e-01 4.99253392e-01 1.23232126e+00 -1.80593461e-01
-4.29830700e-01 2.13178128e-01 -5.88631570e-01 5.94004035e-01
6.49616718e-01 -3.48494202e-01 -4.65785772e-01 -4.99380350e-01
1.19536854e-02 8.14791620e-01 -1.24804676e-01 -3.39994371e-01
4.86100376e-01 -1.42643914e-01 -3.34074438e-01 -6.37092233e-01
3.32059503e-01 -4.31039035e-01 -3.46359462e-01 -2.48467829e-02
-1.26542234e+00 -3.77365825e-04 -2.45063886e-01 4.37682122e-01
-4.73227441e-01 -4.56431508e-01 4.85840648e-01 -2.44017035e-01
-4.45786357e-01 5.29869981e-02 -4.60698813e-01 3.19214761e-01
4.46075648e-01 1.30940378e-01 -2.06528768e-01 -7.50331044e-01
-8.51088107e-01 4.88671899e-01 2.83351690e-02 6.19913399e-01
6.51618719e-01 -1.10329497e+00 -8.35866928e-01 6.41672909e-02
3.10417026e-01 3.58056761e-02 3.81965846e-01 8.04386854e-01
1.62515655e-01 5.23276925e-01 3.74501765e-01 -5.77281237e-01
-1.49327266e+00 1.59779727e-01 2.22912997e-01 -8.73613000e-01
-3.35169375e-01 1.14602721e+00 4.17052090e-01 -8.49463344e-01
4.30818975e-01 -4.63304594e-02 -6.78064346e-01 4.23927158e-01
6.54769242e-01 4.66380902e-02 1.22144558e-01 -6.87120378e-01
-3.42711568e-01 1.34377047e-01 -6.73370481e-01 -4.09488827e-01
9.02182996e-01 -4.36969250e-01 -2.40456983e-01 5.06268263e-01
1.49141526e+00 -2.90474445e-01 -7.79345214e-01 -7.04062879e-01
1.74140036e-01 -6.42027557e-01 -2.72952706e-01 -9.24835563e-01
-8.33390117e-01 7.87574351e-01 3.82738896e-02 3.89772475e-01
9.97620821e-01 -4.01273333e-02 9.18493390e-01 5.91527879e-01
2.52803475e-01 -1.14088714e+00 4.08234596e-01 1.06698692e+00
1.19025099e+00 -1.22571564e+00 -2.59212643e-01 -3.00436944e-01
-1.01256073e+00 9.89582777e-01 8.49900961e-01 6.75214604e-02
3.13895822e-01 -2.14592710e-01 1.89800933e-01 -1.97050855e-01
-1.13085711e+00 -1.51953325e-01 3.23408186e-01 6.59942105e-02
9.91577148e-01 -6.90359995e-02 -7.66003788e-01 6.55037642e-01
1.08874358e-01 -3.40108424e-01 2.97294766e-01 7.69271970e-01
-2.87194192e-01 -1.38327956e+00 -1.93836421e-01 2.92418838e-01
-3.69123191e-01 -3.72965455e-01 -1.04755318e+00 3.57686281e-01
-7.34086335e-01 1.54982841e+00 -1.04055181e-01 -6.60285950e-01
5.20972788e-01 1.87720433e-01 -7.24978149e-02 -8.26776266e-01
-1.33543730e+00 7.61389136e-02 6.27384186e-01 -6.63961053e-01
-2.84471571e-01 -5.79085052e-01 -1.44063747e+00 -8.83675292e-02
-6.27483487e-01 5.96120715e-01 5.04464805e-01 1.11764634e+00
4.11487579e-01 5.20035625e-01 1.30098510e+00 -8.35619867e-01
-9.25332367e-01 -1.34173310e+00 7.88602456e-02 5.88799238e-01
3.90368640e-01 -4.62422550e-01 -5.12785494e-01 -4.90699530e-01] | [12.471622467041016, 7.839714527130127] |
dfaa8915-dea8-4d5a-9803-4519b5a12a87 | multicqa-zero-shot-transfer-of-self | 2010.00980 | null | https://arxiv.org/abs/2010.00980v1 | https://arxiv.org/pdf/2010.00980v1.pdf | MultiCQA: Zero-Shot Transfer of Self-Supervised Text Matching Models on a Massive Scale | We study the zero-shot transfer capabilities of text matching models on a massive scale, by self-supervised training on 140 source domains from community question answering forums in English. We investigate the model performances on nine benchmarks of answer selection and question similarity tasks, and show that all 140 models transfer surprisingly well, where the large majority of models substantially outperforms common IR baselines. We also demonstrate that considering a broad selection of source domains is crucial for obtaining the best zero-shot transfer performances, which contrasts the standard procedure that merely relies on the largest and most similar domains. In addition, we extensively study how to best combine multiple source domains. We propose to incorporate self-supervised with supervised multi-task learning on all available source domains. Our best zero-shot transfer model considerably outperforms in-domain BERT and the previous state of the art on six benchmarks. Fine-tuning of our model with in-domain data results in additional large gains and achieves the new state of the art on all nine benchmarks. | ['Iryna Gurevych', 'Jonas Pfeiffer', 'Andreas Rücklé'] | 2020-10-02 | null | https://aclanthology.org/2020.emnlp-main.194 | https://aclanthology.org/2020.emnlp-main.194.pdf | emnlp-2020-11 | ['question-similarity', 'answer-selection'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.10251166e-01 1.45198433e-02 -2.66204149e-01 -2.80195981e-01
-1.64297760e+00 -6.33283496e-01 9.21579778e-01 2.30598912e-01
-6.13290787e-01 7.48579621e-01 5.96898317e-01 -1.89473227e-01
-3.10936570e-01 -7.28686750e-01 -7.43673921e-01 9.16626900e-02
4.31338251e-01 1.04777098e+00 8.44836652e-01 -8.92966866e-01
2.19384015e-01 -4.95319128e-01 -1.57116866e+00 8.73248875e-01
1.09828472e+00 8.75127852e-01 3.88213643e-03 6.72303319e-01
-5.21382034e-01 1.01104295e+00 -4.48024452e-01 -7.32371032e-01
1.04159854e-01 -4.80816931e-01 -1.41047537e+00 -5.21081507e-01
9.42153633e-01 -4.60032761e-01 -4.39858645e-01 6.64124548e-01
7.00789392e-01 2.87033170e-01 8.28144312e-01 -9.45758700e-01
-1.17776811e+00 3.51501077e-01 -6.97335005e-02 4.33107823e-01
7.23040462e-01 -9.94332433e-02 1.59417212e+00 -1.04430830e+00
8.99668396e-01 1.25795102e+00 6.92231297e-01 4.64842200e-01
-1.25208533e+00 -5.79882503e-01 -3.13311189e-01 5.30268550e-01
-8.86482477e-01 -5.93215764e-01 4.18264747e-01 -4.08953160e-01
1.30173397e+00 5.14377048e-03 -3.46506864e-01 1.11233211e+00
-1.50650859e-01 6.84722006e-01 9.14294064e-01 -7.49134421e-01
7.86071122e-02 2.71419853e-01 7.20041156e-01 4.89463001e-01
-1.83950737e-02 -2.34456033e-01 -4.69593167e-01 -7.93638647e-01
-5.92313483e-02 -2.36930519e-01 -9.43714157e-02 -4.68519539e-01
-1.29094279e+00 1.02398682e+00 3.49135935e-01 3.92907292e-01
7.58074448e-02 -8.66747051e-02 4.77322102e-01 1.17463398e+00
7.46795058e-01 7.22517908e-01 -6.02751255e-01 -6.44942559e-03
-7.37587869e-01 5.06498516e-01 1.24916482e+00 1.14714718e+00
1.01172090e+00 -6.80414915e-01 -5.70925951e-01 1.44419491e+00
-1.80592880e-01 5.57652354e-01 5.81240535e-01 -1.31726348e+00
8.97912860e-01 7.62381792e-01 2.89864928e-01 -4.99122202e-01
-1.08022071e-01 -2.46273249e-01 -3.65732342e-01 -3.27855885e-01
7.54780650e-01 -2.43587345e-01 -5.33481836e-01 1.66018295e+00
3.03092986e-01 -3.95862460e-02 2.47046888e-01 5.95391333e-01
1.25056696e+00 4.94354218e-01 9.77350324e-02 3.84215713e-01
1.57943833e+00 -1.30678999e+00 -4.16183323e-01 -5.29315174e-01
1.04164088e+00 -7.62913048e-01 1.24256098e+00 -2.84769922e-01
-9.01610017e-01 -7.13305771e-01 -8.93700004e-01 -4.24294621e-01
-4.91100103e-01 -2.04392985e-01 1.31321043e-01 3.32605630e-01
-1.02643728e+00 6.71633422e-01 5.57432584e-02 -8.90382051e-01
1.99272767e-01 -3.32591645e-02 -4.38812405e-01 -4.16163892e-01
-1.80223870e+00 1.22641313e+00 8.90887603e-02 -9.99402583e-01
-5.75919032e-01 -1.09489977e+00 -6.30022109e-01 6.69912770e-02
2.00312138e-01 -9.80929673e-01 1.77828717e+00 -6.98616147e-01
-1.12404323e+00 1.39360666e+00 -2.41022095e-01 -5.23777306e-01
4.23673093e-01 -3.16162646e-01 -3.59568805e-01 3.64598453e-01
4.11437839e-01 7.25833178e-01 5.80666780e-01 -6.78729594e-01
-4.98961717e-01 -9.67447087e-02 2.19881773e-01 2.14777634e-01
-8.71979058e-01 3.68914098e-01 -2.65745848e-01 -2.56466717e-01
-5.85177600e-01 -6.87817633e-01 1.25418812e-01 1.22517198e-01
2.92892873e-01 -6.87632859e-01 4.56381589e-01 -6.63263738e-01
9.68916655e-01 -1.71225107e+00 -1.32188067e-01 -4.17036533e-01
2.14235112e-01 3.83048981e-01 -6.10606611e-01 8.19210887e-01
3.95520963e-02 -2.81130105e-01 -1.64733902e-01 -2.62946576e-01
2.30455309e-01 -1.39842974e-02 -5.85720837e-01 1.18337043e-01
8.51888731e-02 1.33202302e+00 -1.12206376e+00 -6.56968653e-01
-1.31838262e-01 -8.11438188e-02 -6.68758273e-01 5.12905359e-01
-4.05836672e-01 1.22694420e-02 -5.12026489e-01 2.92756796e-01
2.71620274e-01 -8.89975727e-01 -2.15820465e-02 2.14849383e-01
5.04029334e-01 7.28885949e-01 -6.44883931e-01 2.04397225e+00
-4.74704385e-01 4.94279772e-01 -4.75309864e-02 -1.10085285e+00
1.08219957e+00 4.80438530e-01 4.11391526e-01 -1.06158733e+00
-1.98062390e-01 5.43290079e-01 -9.72150266e-02 -4.55201566e-01
4.55193102e-01 1.25214281e-02 -2.32371509e-01 8.75598371e-01
5.19910693e-01 -7.84274638e-02 1.19643241e-01 5.31471193e-01
1.46844327e+00 -5.07307099e-03 3.87092024e-01 -4.91805196e-01
5.01540244e-01 3.89347970e-01 -4.88001704e-02 1.04394209e+00
-2.59251654e-01 5.80579519e-01 3.39812279e-01 -3.39260213e-02
-1.20019972e+00 -9.70299482e-01 -1.57835186e-01 1.98138976e+00
1.35509744e-01 -1.02634259e-01 -6.82394922e-01 -8.75576854e-01
4.08255726e-01 7.84604490e-01 -7.65859127e-01 -3.27373624e-01
-3.75928342e-01 -4.10221606e-01 6.02170289e-01 4.09141272e-01
5.42459846e-01 -9.11630988e-01 -1.58530280e-01 3.91960651e-01
-5.61865211e-01 -1.18952858e+00 -5.66136062e-01 -5.63395880e-02
-7.60803223e-01 -1.11018717e+00 -1.27552617e+00 -9.34030175e-01
-7.66010731e-02 6.81344390e-01 1.82098961e+00 7.54028708e-02
-1.45347759e-01 5.75666189e-01 -6.15157545e-01 -7.73733929e-02
-5.30513465e-01 5.90750635e-01 -3.21402013e-01 -4.04591978e-01
8.69675457e-01 -5.42900980e-01 -5.67044675e-01 5.23819745e-01
-5.02404213e-01 -4.22976226e-01 1.80736676e-01 1.03770626e+00
-1.06829010e-01 -1.09936738e+00 1.29229903e+00 -1.16299510e+00
9.52109098e-01 -1.05529916e+00 -7.85590187e-02 8.48831534e-01
-5.91155708e-01 1.81467235e-01 5.08784175e-01 -1.98333099e-01
-1.28589165e+00 -5.51353276e-01 -1.77450016e-01 -1.63781196e-01
-1.42449901e-01 6.11784682e-02 2.85884976e-01 -5.79115227e-02
1.53715098e+00 -1.35463148e-01 -7.45818615e-02 -7.07839131e-01
8.15305769e-01 1.05508959e+00 5.15424669e-01 -8.69366825e-01
6.48049474e-01 3.84861290e-01 -7.69213319e-01 -5.07982373e-01
-1.20533931e+00 -1.18785119e+00 -6.22618973e-01 1.67819008e-01
6.67089641e-01 -1.07228196e+00 -2.10869744e-01 1.68096393e-01
-1.37901151e+00 -3.83269697e-01 -2.27114573e-01 -1.63347647e-02
-5.17632067e-01 5.55538952e-01 -8.70974481e-01 -2.56316841e-01
-7.14462221e-01 -5.19531369e-01 1.06479645e+00 -7.84855857e-02
-5.01725674e-01 -1.20797586e+00 7.72442579e-01 7.38062739e-01
7.00784862e-01 -3.30789447e-01 9.56023991e-01 -1.48080432e+00
-1.50748357e-01 -1.22556034e-02 -5.58731675e-01 1.31953195e-01
-1.37160169e-02 -7.99117506e-01 -1.13874006e+00 -2.01641142e-01
-1.40576795e-01 -1.09900844e+00 1.23924494e+00 -1.95003763e-01
5.36493957e-01 1.09441824e-01 -4.58320707e-01 1.46141887e-01
1.24670959e+00 -2.96231925e-01 4.00755465e-01 6.14607096e-01
2.50996351e-01 1.06660748e+00 6.32136166e-01 1.40365690e-01
6.15459323e-01 6.64155126e-01 -3.87151949e-02 8.91915858e-02
-3.47114027e-01 -3.10199380e-01 1.96957529e-01 8.82053375e-01
3.88878822e-01 -1.28530607e-01 -1.01074314e+00 9.00336742e-01
-1.99500406e+00 -1.06322002e+00 -1.04776099e-01 2.06368160e+00
1.05462360e+00 -1.07864216e-01 2.09294185e-01 -5.11633277e-01
8.22574317e-01 1.51008219e-01 -6.00812078e-01 -2.17677847e-01
-2.02556685e-01 6.54760420e-01 2.78718680e-01 5.98286211e-01
-1.06956565e+00 8.48131597e-01 7.14828873e+00 8.88815045e-01
-1.60840347e-01 7.23745525e-01 2.93639332e-01 5.40664643e-02
-6.10614061e-01 4.93442528e-02 -9.35487330e-01 2.98110455e-01
1.20522428e+00 -6.69310927e-01 1.08758360e-01 9.39185023e-01
-6.94640756e-01 2.96737701e-01 -1.35887027e+00 5.55289805e-01
2.81188667e-01 -1.43648005e+00 -7.53726661e-02 -2.61319548e-01
1.03088379e+00 6.59324050e-01 -2.89895356e-01 9.43400204e-01
8.57200384e-01 -6.87965810e-01 1.48444369e-01 2.74599969e-01
7.90382743e-01 -1.48360372e-01 5.72590292e-01 5.12940824e-01
-8.28377485e-01 -1.50464162e-01 -6.73835993e-01 -1.11870930e-01
9.58671197e-02 2.67263263e-01 -6.01400018e-01 6.69379830e-01
6.07811391e-01 6.56695366e-01 -6.38827562e-01 8.94233227e-01
3.47501449e-02 5.44884741e-01 -1.48768470e-01 -2.57386744e-01
2.04923257e-01 2.40747839e-01 2.28594750e-01 1.42092180e+00
8.99243429e-02 -3.14449263e-03 7.70937949e-02 7.95985103e-01
-4.91320699e-01 3.64681482e-01 -6.56053960e-01 2.82895356e-01
6.84172928e-01 1.05730402e+00 1.64626464e-01 -8.39561045e-01
-8.61438751e-01 1.05200744e+00 8.81340683e-01 3.36757898e-01
-5.89775443e-01 -6.74026847e-01 5.48971832e-01 8.13469663e-02
2.50557810e-01 1.87603846e-01 -7.82284439e-02 -1.44998264e+00
2.95925289e-02 -1.16899097e+00 9.87293482e-01 -5.68732440e-01
-2.09633756e+00 5.34079075e-01 -2.71970942e-03 -1.17580009e+00
-6.28352761e-01 -5.24254203e-01 -7.86474884e-01 1.00028539e+00
-1.99314559e+00 -1.03402591e+00 -3.15762073e-01 8.26773822e-01
6.18421912e-01 -3.73869568e-01 1.11698067e+00 6.06458008e-01
5.80637343e-02 9.44907546e-01 7.01844573e-01 2.38005072e-01
1.61701107e+00 -1.11177480e+00 8.40367436e-01 3.09200794e-01
1.55494332e-01 5.22369683e-01 3.77911389e-01 -4.56252813e-01
-9.84381676e-01 -7.68310487e-01 1.29652655e+00 -9.66500044e-01
1.14660418e+00 -3.02742451e-01 -1.38353670e+00 6.31601393e-01
6.29220963e-01 -4.67223451e-02 7.25444674e-01 4.26017344e-01
-9.83867943e-01 2.60219723e-02 -9.37039733e-01 1.88123137e-01
1.03704345e+00 -9.40631688e-01 -1.54315066e+00 7.24013507e-01
1.03560793e+00 1.14611881e-02 -1.06963003e+00 3.03019196e-01
3.87018681e-01 -8.53488803e-01 1.26647294e+00 -1.08664346e+00
8.50032568e-01 3.01711529e-01 -1.82414159e-01 -1.14558148e+00
-5.14597595e-01 -3.54758084e-01 -1.31267354e-01 1.27853167e+00
5.84117591e-01 -7.45140254e-01 6.83877528e-01 6.88865483e-01
7.60087892e-02 -2.77720630e-01 -1.00498426e+00 -9.19171095e-01
8.68711710e-01 1.39981450e-03 3.28976691e-01 1.15962481e+00
4.09517407e-01 8.82704377e-01 -2.33771592e-01 -5.73360622e-01
7.13306963e-01 3.65235895e-01 8.68022799e-01 -1.36550224e+00
-4.38893110e-01 -3.74205619e-01 1.77995905e-01 -1.35811996e+00
5.26487947e-01 -1.18100834e+00 -1.06688559e-01 -1.69036138e+00
5.91490567e-01 -3.25960755e-01 -4.62437272e-01 2.73467422e-01
-7.16942072e-01 1.71944842e-01 -8.18245932e-02 5.83550572e-01
-1.18405998e+00 6.49452329e-01 9.06323493e-01 -2.72213638e-01
1.76253498e-01 -2.78405368e-01 -8.25142622e-01 2.28590593e-01
6.38938189e-01 -4.79269445e-01 -4.65891212e-01 -7.51755893e-01
-1.33110983e-02 -3.18817683e-02 1.86571836e-01 -7.69675314e-01
6.25266910e-01 3.02202761e-01 -8.08824375e-02 -1.87775001e-01
2.78781921e-01 -2.66508549e-01 -5.57082057e-01 3.61874253e-01
-9.17553782e-01 -1.85066640e-01 2.21034475e-02 6.39262915e-01
-3.12283814e-01 -3.54279697e-01 8.33018422e-01 -2.67474800e-01
-8.74601126e-01 3.82559448e-02 9.01972875e-03 1.12155080e+00
5.80631912e-01 1.57078907e-01 -8.99166107e-01 -5.84319234e-01
-2.34278753e-01 6.12884820e-01 3.12182009e-01 8.61616969e-01
1.44057527e-01 -1.41709220e+00 -1.21938741e+00 -4.17884082e-01
8.71756613e-01 -6.64912462e-01 3.48273516e-01 4.90984440e-01
-1.02473795e-01 7.06990123e-01 -1.11155428e-01 -4.17441815e-01
-9.95806813e-01 5.97187281e-01 3.41338158e-01 -7.17680216e-01
-3.84879798e-01 8.10341895e-01 2.10600823e-01 -1.04026675e+00
6.64392188e-02 2.01025099e-01 -2.76884139e-01 1.88933715e-01
6.64935231e-01 5.00809610e-01 2.51843244e-01 -1.11912817e-01
-2.38713428e-01 5.10013521e-01 -3.57921004e-01 -1.74568325e-01
1.12872517e+00 -1.33534998e-01 -2.15828791e-02 2.20534056e-01
1.58330023e+00 -4.38760698e-01 -8.44452381e-01 -1.09915626e+00
4.10687029e-01 -2.32806861e-01 -4.36792284e-01 -7.35697269e-01
-1.89298898e-01 1.06051290e+00 1.47451252e-01 -7.99773857e-02
5.82618237e-01 3.11768204e-01 1.15428483e+00 9.79621708e-01
3.55105340e-01 -1.31382787e+00 4.14646953e-01 9.14460421e-01
6.73262894e-01 -1.56750715e+00 -2.89807379e-01 -1.00914948e-01
-6.37264252e-01 8.32173586e-01 7.60209858e-01 -3.29630464e-01
6.04060590e-01 -1.91300660e-01 1.94493551e-02 -1.91153288e-01
-1.18000996e+00 -4.56328899e-01 5.25976896e-01 5.42364717e-01
6.63493156e-01 -4.13771719e-01 -2.44159266e-01 5.46787143e-01
8.79310146e-02 1.20647505e-01 -6.27276450e-02 7.18202829e-01
-8.89132380e-01 -1.32749534e+00 -1.43387303e-01 5.60772479e-01
-5.26744485e-01 -4.68864948e-01 -6.35156453e-01 6.55490696e-01
-4.54052687e-01 1.19577742e+00 -1.02622258e-02 -2.97039628e-01
4.79282498e-01 5.79225242e-01 3.38253498e-01 -7.93302953e-01
-9.58979845e-01 -6.13023698e-01 5.52045286e-01 -4.30395186e-01
-3.20575535e-01 -3.88030350e-01 -6.30178094e-01 -3.02518904e-01
-2.51152992e-01 4.24608529e-01 4.72355410e-02 9.77070093e-01
6.52465224e-01 3.48192751e-02 5.09532392e-01 -8.81860778e-02
-1.25481558e+00 -1.32185876e+00 -1.57319233e-01 8.90426993e-01
1.69541433e-01 -6.14505172e-01 -4.74448204e-01 -1.09376445e-01] | [11.378213882446289, 7.947145938873291] |
e47b30e0-89a9-46a8-b181-71f47b233d80 | offline-reinforcement-learning-with-7 | 2307.02752 | null | https://arxiv.org/abs/2307.02752v1 | https://arxiv.org/pdf/2307.02752v1.pdf | Offline Reinforcement Learning with Imbalanced Datasets | The prevalent use of benchmarks in current offline reinforcement learning (RL) research has led to a neglect of the imbalance of real-world dataset distributions in the development of models. The real-world offline RL dataset is often imbalanced over the state space due to the challenge of exploration or safety considerations. In this paper, we specify properties of imbalanced datasets in offline RL, where the state coverage follows a power law distribution characterized by skewed policies. Theoretically and empirically, we show that typically offline RL methods based on distributional constraints, such as conservative Q-learning (CQL), are ineffective in extracting policies under the imbalanced dataset. Inspired by natural intelligence, we propose a novel offline RL method that utilizes the augmentation of CQL with a retrieval process to recall past related experiences, effectively alleviating the challenges posed by imbalanced datasets. We evaluate our method on several tasks in the context of imbalanced datasets with varying levels of imbalance, utilizing the variant of D4RL. Empirical results demonstrate the superiority of our method over other baselines. | ['Zhao Ding', 'Wai Kin Chan', 'Haoran Xu', 'JieLin Qiu', 'Sijie Chen', 'Li Jiang'] | 2023-07-06 | null | null | null | null | ['q-learning', 'reinforcement-learning-1', 'retrieval', 'offline-rl', 'd4rl'] | ['methodology', 'methodology', 'methodology', 'playing-games', 'robots'] | [-9.72018987e-02 5.43829938e-03 -6.06803536e-01 -1.89193726e-01
-8.04715216e-01 -6.15697563e-01 3.25223029e-01 4.46121275e-01
-5.65721989e-01 1.01689279e+00 3.08777809e-01 -4.38621491e-01
-3.44047159e-01 -6.51585162e-01 -7.76092768e-01 -4.04181600e-01
-2.42843136e-01 6.61470354e-01 -2.37997353e-01 -3.04795802e-01
3.87299240e-01 2.32665583e-01 -1.71572542e+00 1.31824061e-01
1.01891208e+00 9.80530322e-01 -2.27228686e-01 3.28111917e-01
-4.00072448e-02 1.18706048e+00 -1.06230044e+00 -1.26788139e-01
4.90955114e-01 -3.70792687e-01 -6.00949287e-01 -1.34200454e-01
2.95825712e-02 -5.61454713e-01 -3.40990454e-01 8.41442466e-01
6.74249828e-01 3.77831280e-01 2.89887071e-01 -1.78254485e+00
-1.96518421e-01 6.06238544e-01 -5.48010230e-01 2.90087909e-01
2.84836024e-01 4.78989959e-01 9.63810921e-01 -1.15541890e-01
6.16737902e-01 1.28658485e+00 4.09081191e-01 6.00098550e-01
-1.35636294e+00 -7.75451601e-01 5.74484289e-01 2.36074015e-01
-7.85361886e-01 -2.65289664e-01 7.48499751e-01 -1.70616940e-01
1.10680711e+00 3.64324711e-02 9.15697575e-01 1.29518700e+00
1.30690336e-01 1.37807655e+00 1.36590362e+00 -3.41692626e-01
8.55405211e-01 1.16654471e-01 -2.72437576e-02 6.89773634e-02
4.76627052e-01 5.47216058e-01 -8.61065567e-01 -3.98008138e-01
2.36806586e-01 -7.01644644e-02 1.89697996e-01 -7.49166667e-01
-7.61347950e-01 9.56584871e-01 -2.65014805e-02 -2.19308570e-01
-3.93057853e-01 8.23450312e-02 7.48867691e-01 6.05956733e-01
6.96854591e-01 9.10709500e-01 -6.38025761e-01 -5.70968568e-01
-7.52558231e-01 7.90835559e-01 1.04395974e+00 7.91008353e-01
4.57689792e-01 -1.01376258e-01 -4.57797766e-01 5.85934758e-01
-1.36141442e-02 4.42440987e-01 7.18504190e-01 -1.00051165e+00
6.31718814e-01 5.53379416e-01 5.14743030e-01 -6.79866612e-01
-4.60161060e-01 -5.38453043e-01 -3.72404099e-01 -5.72626805e-03
2.86785066e-01 -2.19051376e-01 -6.90008640e-01 1.99778128e+00
3.56970459e-01 -2.50694335e-01 2.67293870e-01 6.91045284e-01
-5.93213961e-02 2.87728310e-01 2.57138848e-01 -5.15664756e-01
5.46293914e-01 -7.75273204e-01 -1.00077915e+00 -2.31847852e-01
7.34800220e-01 -3.99770103e-02 1.33441770e+00 8.44661415e-01
-1.02399158e+00 -1.76724598e-01 -1.06331670e+00 4.26519006e-01
-2.09877640e-01 -3.44838411e-01 8.06372583e-01 5.65879464e-01
-7.54007280e-01 6.95503056e-01 -9.26956534e-01 -1.06143720e-01
7.43237376e-01 2.77201056e-01 1.74437851e-01 -4.57337312e-02
-1.37089717e+00 1.04448497e+00 4.48400289e-01 -1.16667822e-01
-1.09304047e+00 -9.01250303e-01 -6.59568369e-01 -1.01047546e-01
8.08817208e-01 -3.29521656e-01 1.52770758e+00 -1.11798143e+00
-1.33420599e+00 3.72810543e-01 3.92777562e-01 -8.72759640e-01
8.89223695e-01 -6.38977647e-01 -3.10048848e-01 -1.13648390e-02
-9.93705988e-02 3.71073782e-01 7.11703658e-01 -1.08980000e+00
-5.46196818e-01 -5.55807590e-01 2.48534277e-01 5.88841498e-01
-3.91887099e-01 -4.09257382e-01 8.61068070e-02 -3.04520011e-01
-4.80866194e-01 -7.55550086e-01 -4.71709907e-01 -5.51327229e-01
-2.72474915e-01 -2.98099816e-01 5.95403492e-01 -3.13458562e-01
1.49313200e+00 -1.92748761e+00 -2.30892420e-01 2.49777615e-01
6.92968220e-02 1.81299597e-01 -1.98384702e-01 8.21829736e-01
1.61031425e-01 1.44024134e-01 8.09276998e-02 -1.88124806e-01
2.60794580e-01 3.22053015e-01 -7.34438002e-01 5.95134020e-01
1.38085678e-01 7.37793863e-01 -1.23904169e+00 -3.62331837e-01
-1.40280962e-01 -2.86924124e-01 -7.89888322e-01 5.25412261e-01
-7.79147327e-01 4.15023357e-01 -5.59678435e-01 6.28455400e-01
4.81618375e-01 -8.99771750e-02 4.74037230e-01 2.22430691e-01
-1.28120054e-02 3.51730615e-01 -1.01883638e+00 1.68741059e+00
-5.03881216e-01 2.38623992e-01 -3.66226912e-01 -1.05223441e+00
8.36972892e-01 -2.80188974e-02 6.66725099e-01 -1.33477283e+00
-1.34938657e-01 -3.75164934e-02 5.40925153e-02 -6.50665402e-01
4.83025253e-01 -1.77438766e-01 -2.31082499e-01 7.44916201e-01
-4.46027555e-02 -2.88866132e-01 3.57876718e-01 1.70473218e-01
1.37942755e+00 4.28908139e-01 2.46449217e-01 -1.39494821e-01
-2.36145616e-01 1.60549626e-01 8.83715153e-01 1.02877378e+00
-6.30318522e-01 1.31732095e-02 9.18628216e-01 -4.49816912e-01
-9.73811567e-01 -8.46805215e-01 6.35536760e-02 1.04324853e+00
2.44236633e-01 -3.38254064e-01 -5.46609640e-01 -8.53426337e-01
4.34268653e-01 9.44942474e-01 -5.26772499e-01 -5.90503037e-01
-3.75392050e-01 -9.08965051e-01 5.59717596e-01 4.59944159e-01
2.85566181e-01 -1.27783763e+00 -1.31114674e+00 3.38971376e-01
-1.04154490e-01 -9.06999290e-01 -1.07122888e-03 4.33082521e-01
-9.65452909e-01 -1.13773906e+00 -9.41932797e-02 3.11963111e-02
2.94224739e-01 -2.71824718e-01 1.53746438e+00 -1.61168069e-01
-3.18159759e-01 5.37720919e-01 -2.83919990e-01 -9.28907871e-01
-2.48600438e-01 1.01312570e-01 3.30004066e-01 -3.61383110e-01
5.77357531e-01 -3.87999713e-01 -8.18936706e-01 -9.48899388e-02
-1.06342328e+00 -2.64932156e-01 4.77441877e-01 9.62839067e-01
4.68517035e-01 1.86053142e-01 1.25519788e+00 -1.16518712e+00
1.00975072e+00 -7.42550194e-01 -6.30340219e-01 2.48218343e-01
-9.84447181e-01 1.81615919e-01 7.01138735e-01 -6.74519300e-01
-8.96780849e-01 -2.40963876e-01 3.13850939e-01 -5.96596658e-01
-4.35559303e-02 3.63625228e-01 -1.86892927e-01 4.66683447e-01
6.90029502e-01 8.67731199e-02 2.56039649e-01 -2.23700017e-01
2.79974401e-01 5.35354793e-01 4.67671789e-02 -9.32304561e-01
2.98361868e-01 3.37331682e-01 -2.67468154e-01 -3.12783480e-01
-1.29917729e+00 -2.90216684e-01 -4.89860103e-02 -1.82830676e-01
1.63978919e-01 -9.63709533e-01 -9.27979410e-01 3.21376920e-01
-4.01369512e-01 -8.00145030e-01 -7.57176518e-01 3.22961509e-01
-1.29398501e+00 -6.16772473e-02 -3.71097147e-01 -1.42284322e+00
-2.21483245e-01 -9.23531830e-01 8.43161404e-01 2.24053711e-01
-2.01873451e-01 -7.36894608e-01 5.74197412e-01 2.47883901e-01
3.94330829e-01 3.42802793e-01 1.06092703e+00 -9.57120299e-01
-2.28896290e-01 1.89773947e-01 2.83887863e-01 2.73478538e-01
2.07728054e-03 -2.45979086e-01 -1.00273275e+00 -6.19939268e-01
1.84711087e-02 -1.15339589e+00 6.57176137e-01 1.60911947e-01
1.66846406e+00 -4.73345011e-01 -2.34347191e-02 2.12854281e-01
1.55952477e+00 3.67223054e-01 3.11308563e-01 5.54095864e-01
5.49660809e-02 7.68369853e-01 1.42423844e+00 1.05688906e+00
4.48220730e-01 2.57946432e-01 7.41855741e-01 4.94629890e-02
4.79145318e-01 -8.28964353e-01 4.30733055e-01 3.74781162e-01
6.05021298e-01 -3.14963758e-01 -9.53769088e-01 8.00778091e-01
-2.06194043e+00 -8.09155345e-01 9.02522743e-01 2.36907291e+00
1.27112246e+00 3.28235537e-01 4.35308069e-01 1.46089554e-01
2.73353785e-01 1.33958906e-01 -1.41741920e+00 -6.38705015e-01
3.63906696e-02 2.40296111e-01 5.70683539e-01 8.46727714e-02
-9.93281007e-01 7.74331510e-01 6.50148535e+00 6.90471351e-01
-9.68853116e-01 -2.22089782e-01 9.20979619e-01 -6.18354380e-01
-4.19147700e-01 -1.27950370e-01 -6.79050505e-01 4.46840495e-01
1.14078975e+00 -3.14327598e-01 5.62402427e-01 7.31034994e-01
1.97303519e-01 -3.23930740e-01 -1.39411950e+00 8.51781666e-01
-8.92160013e-02 -7.85203874e-01 -1.96303442e-01 -5.62650673e-02
9.45920944e-01 -2.74964273e-02 3.24918807e-01 9.29917157e-01
7.48893857e-01 -1.20963824e+00 7.67031431e-01 4.61312532e-01
6.47934020e-01 -1.21187866e+00 4.71364111e-01 7.70348310e-01
-2.95760661e-01 -7.31040418e-01 -2.37147748e-01 -1.78372055e-01
-4.15213704e-01 5.57115972e-01 -8.27350736e-01 2.77391076e-01
7.27889776e-01 7.09399045e-01 -2.69811422e-01 8.75067651e-01
1.31131679e-01 6.96946502e-01 -1.74481079e-01 -1.39423251e-01
1.84636205e-01 -7.67059699e-02 2.48840630e-01 7.78620839e-01
-1.43338636e-01 -2.03465894e-01 4.38908219e-01 6.57210648e-01
-5.30158319e-02 5.08011989e-02 -9.27915335e-01 -2.61149019e-01
6.56503975e-01 8.42255592e-01 -2.47274369e-01 -1.54065207e-01
-3.84499766e-02 2.41750658e-01 6.69842422e-01 3.96536827e-01
-6.45472050e-01 -7.97379240e-02 8.13938320e-01 -1.16466701e-01
1.13187946e-01 7.24972133e-03 -2.33052611e-01 -9.97663140e-01
1.31252885e-01 -1.60630596e+00 5.99467158e-01 -2.26328880e-01
-1.44734442e+00 2.17529207e-01 1.27898529e-01 -1.22531247e+00
-4.81382966e-01 -1.28901839e-01 -8.26079100e-02 4.56457555e-01
-1.91808498e+00 -5.25697529e-01 1.69489384e-01 4.32924002e-01
5.19038141e-01 3.50792296e-02 6.65323079e-01 -3.07288766e-02
-5.50834060e-01 4.93693620e-01 2.50986993e-01 -3.81079197e-01
8.87732089e-01 -1.42997015e+00 5.91088347e-02 3.23497653e-01
-1.18912935e-01 4.16063666e-01 9.25517678e-01 -7.02804208e-01
-1.71354890e+00 -1.12030852e+00 2.48812884e-01 -4.58182603e-01
5.79350948e-01 -4.65734661e-01 -6.60654604e-01 5.03199100e-01
8.48661512e-02 2.98891868e-02 7.26365328e-01 1.70583442e-01
-3.34576070e-01 -4.51483727e-01 -1.37986207e+00 4.97623503e-01
1.04995584e+00 -4.40092832e-01 -4.43172693e-01 3.76550257e-01
8.07718396e-01 -5.02907872e-01 -8.52528274e-01 5.40949166e-01
4.88488495e-01 -9.37391222e-01 6.48352504e-01 -1.32710171e+00
4.77892965e-01 1.54236272e-01 -3.87408808e-02 -1.64441562e+00
3.38499159e-01 -6.70014799e-01 -6.47817492e-01 8.89238775e-01
8.28448609e-02 -4.43074495e-01 7.06216097e-01 7.35909879e-01
4.32588786e-01 -1.13739932e+00 -7.00204372e-01 -9.27098453e-01
2.50897884e-01 -2.77716011e-01 7.76550055e-01 6.28252685e-01
1.55574471e-01 -2.08923563e-01 -2.57078826e-01 -2.37080842e-01
6.95205271e-01 4.61548716e-01 7.44936466e-01 -1.02588856e+00
-4.35080856e-01 -1.39752254e-01 -1.80992242e-02 -7.46422887e-01
4.85520422e-01 -3.68428469e-01 2.31730521e-01 -1.04232442e+00
1.52530983e-01 -9.08523798e-01 -7.56581068e-01 3.84210169e-01
-1.07468508e-01 -2.56445646e-01 2.22359627e-01 -1.33596882e-02
-1.14595544e+00 9.42458332e-01 1.15898144e+00 -5.68309203e-02
-4.21392858e-01 4.28235484e-03 -8.20116460e-01 4.68538344e-01
9.88824844e-01 -4.86174196e-01 -9.15785134e-01 -1.21407434e-01
7.38226473e-01 3.10015202e-01 -1.11492522e-01 -7.82588780e-01
-4.44554240e-02 -4.38145965e-01 2.20969647e-01 -5.80609739e-01
1.28447250e-01 -7.19865859e-01 -3.05391401e-01 5.86357713e-01
-9.04035211e-01 2.90858686e-01 4.07988101e-01 7.12789893e-01
-2.47426301e-01 2.13613078e-01 5.03385544e-01 -1.88993067e-01
-3.84363830e-01 3.26183707e-01 -5.07587641e-02 7.83728957e-01
1.12623227e+00 2.35574588e-01 -4.63974983e-01 -5.65426648e-01
-3.54477942e-01 8.29551816e-01 3.46546859e-01 5.02218425e-01
4.64271039e-01 -1.00478458e+00 -4.53049302e-01 2.32091770e-01
2.18337834e-01 7.15684518e-02 1.99199483e-01 7.23681211e-01
-5.09619564e-02 2.58842021e-01 -4.33809698e-01 -4.04163748e-01
-6.91414654e-01 8.12107503e-01 3.46401423e-01 -8.64466012e-01
-4.52459306e-01 2.90621340e-01 8.89656041e-03 -5.17296255e-01
6.58459663e-01 -2.89942026e-01 -2.26892918e-01 2.88994521e-01
3.61156553e-01 4.14112300e-01 3.26706588e-01 2.83501357e-01
-1.87022850e-01 -2.41029352e-01 -1.85945824e-01 -2.35844657e-01
1.41510677e+00 1.59745738e-02 3.23100477e-01 8.33439648e-01
7.32645810e-01 -2.88693488e-01 -1.51918900e+00 -9.88567173e-02
3.61524910e-01 -5.05799890e-01 -2.13747561e-01 -1.27018011e+00
-8.16562295e-01 6.34949684e-01 6.34129703e-01 2.64246225e-01
1.28862238e+00 -4.45897192e-01 5.57083428e-01 3.85605484e-01
7.35057950e-01 -1.61810482e+00 2.15313420e-01 4.64641631e-01
6.59297466e-01 -1.31620049e+00 -7.47149764e-03 4.93734181e-01
-9.47813630e-01 5.86481512e-01 1.02676952e+00 -2.60647744e-01
4.96001601e-01 4.33491498e-01 2.29841381e-01 6.68867826e-02
-1.59949982e+00 -1.56986043e-01 -4.25467372e-01 5.28536499e-01
-2.56818570e-02 1.35896534e-01 -3.47848803e-01 5.89712977e-01
-2.12408289e-01 2.11187780e-01 5.44765413e-01 1.25599754e+00
2.03441251e-02 -1.18854606e+00 -3.81471775e-02 6.38160646e-01
-5.94969392e-01 1.72227308e-01 -2.01768503e-01 7.71606684e-01
-1.18051320e-01 9.66364205e-01 2.92408735e-01 -1.19590469e-01
3.54638398e-01 -1.70610901e-02 5.62391341e-01 -4.27171230e-01
-8.96007717e-01 -2.03820005e-01 5.14029637e-02 -1.16610324e+00
-3.22267085e-01 -6.99317992e-01 -1.13113832e+00 1.54803157e-01
-5.41449599e-02 1.56510055e-01 7.30566442e-01 8.66315126e-01
5.62766135e-01 4.24850166e-01 9.80680346e-01 -2.12095380e-01
-1.60255158e+00 -8.09183538e-01 -6.82102621e-01 6.32272005e-01
6.19858325e-01 -8.09289515e-01 -3.75196844e-01 -5.76990724e-01] | [4.081406593322754, 2.278317928314209] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.