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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
942043ba-c25a-4eed-b9ab-3cf1a2dea6e0 | two-independent-teachers-are-better-role | 2306.05745 | null | https://arxiv.org/abs/2306.05745v1 | https://arxiv.org/pdf/2306.05745v1.pdf | Two Independent Teachers are Better Role Model | Recent deep learning models have attracted substantial attention in infant brain analysis. These models have performed state-of-the-art performance, such as semi-supervised techniques (e.g., Temporal Ensembling, mean teacher). However, these models depend on an encoder-decoder structure with stacked local operators to gather long-range information, and the local operators limit the efficiency and effectiveness. Besides, the $MRI$ data contain different tissue properties ($TPs$) such as $T1$ and $T2$. One major limitation of these models is that they use both data as inputs to the segment process, i.e., the models are trained on the dataset once, and it requires much computational and memory requirements during inference. In this work, we address the above limitations by designing a new deep-learning model, called 3D-DenseUNet, which works as adaptable global aggregation blocks in down-sampling to solve the issue of spatial information loss. The self-attention module connects the down-sampling blocks to up-sampling blocks, and integrates the feature maps in three dimensions of spatial and channel, effectively improving the representation potential and discriminating ability of the model. Additionally, we propose a new method called Two Independent Teachers ($2IT$), that summarizes the model weights instead of label predictions. Each teacher model is trained on different types of brain data, $T1$ and $T2$, respectively. Then, a fuse model is added to improve test accuracy and enable training with fewer parameters and labels compared to the Temporal Ensembling method without modifying the network architecture. Empirical results demonstrate the effectiveness of the proposed method. | ['Kun He', 'Ahmed A. Mubarak', 'Afifa Khaled'] | 2023-06-09 | null | null | null | null | ['brain-segmentation'] | ['medical'] | [-5.22401147e-02 1.03951864e-01 -1.37861535e-01 -6.06805623e-01
-5.60600162e-01 1.15719378e-01 1.46007031e-01 -1.21713482e-01
-4.96506959e-01 5.25940716e-01 -6.18727058e-02 -6.27725422e-02
-1.79197475e-01 -7.84123361e-01 -8.63499701e-01 -8.76560688e-01
-1.96940705e-01 1.44717559e-01 6.00819051e-01 2.03294843e-01
1.89278051e-02 1.12733878e-01 -1.31289852e+00 3.65310490e-01
1.19428897e+00 1.44670653e+00 4.88556057e-01 -1.08212139e-02
-2.75953472e-01 8.43048632e-01 -3.67735595e-01 -4.37277138e-01
4.11184086e-03 -4.48742270e-01 -4.63027686e-01 -4.37376618e-01
1.12659976e-01 -7.44575679e-01 -5.64352214e-01 1.09480453e+00
8.83339047e-01 -1.37549005e-02 6.73101723e-01 -8.60657632e-01
-6.84934318e-01 7.17334151e-01 -8.48978162e-01 4.14552331e-01
-2.37103745e-01 2.73838099e-02 4.88966018e-01 -9.31426167e-01
2.58437911e-04 1.11823273e+00 7.42254794e-01 5.91179907e-01
-1.06900358e+00 -1.15948546e+00 3.79942745e-01 4.91231084e-01
-1.53071666e+00 -5.46219587e-01 8.20901692e-01 -5.50052702e-01
8.55755448e-01 -1.90875158e-01 7.37374008e-01 9.20872211e-01
1.14001103e-01 1.10671401e+00 1.13375449e+00 -4.69059460e-02
1.83982626e-01 -6.32537529e-02 9.95899513e-02 9.02970612e-01
-8.48256946e-02 1.39365762e-01 -5.18278718e-01 2.55755275e-01
1.00059736e+00 1.25686854e-01 -3.37467343e-01 -3.12511623e-01
-9.59352970e-01 6.87024772e-01 7.04857767e-01 2.67111838e-01
-3.47576708e-01 2.90857200e-02 5.19049168e-01 -2.36259140e-02
7.02555358e-01 -1.22978680e-01 -6.27855897e-01 -5.62633164e-02
-1.12630498e+00 -2.21761614e-01 -1.17763570e-02 9.64468777e-01
7.71917462e-01 5.72056994e-02 -1.41353071e-01 1.22764051e+00
4.14507657e-01 3.36252183e-01 8.58043909e-01 -5.39224267e-01
6.07497454e-01 5.41535258e-01 -4.85075861e-01 -8.66007864e-01
-5.44059038e-01 -6.63175821e-01 -9.94858444e-01 -9.47183892e-02
4.17570882e-02 -1.73731104e-01 -1.13329363e+00 2.01738310e+00
2.48321012e-01 4.77370709e-01 -2.92055488e-01 8.51953685e-01
1.05701864e+00 6.70843899e-01 1.77581325e-01 -3.79532576e-01
1.14632130e+00 -1.25586951e+00 -6.68451130e-01 -1.04552947e-01
5.53492665e-01 -3.33310992e-01 8.61264348e-01 4.12967771e-01
-1.33014655e+00 -7.49990642e-01 -9.77902889e-01 2.40019942e-03
-2.12343365e-01 3.03626090e-01 5.57499826e-01 3.75006646e-01
-9.95690942e-01 4.84216958e-01 -1.16977060e+00 1.01662606e-01
9.00467694e-01 5.01008272e-01 -1.72644854e-01 -1.26204163e-01
-1.29579508e+00 8.52693021e-01 4.35192645e-01 2.15891734e-01
-1.27344060e+00 -7.46362686e-01 -7.33637094e-01 1.51561081e-01
1.17378563e-01 -2.57884800e-01 1.04737520e+00 -7.41655231e-01
-1.54459214e+00 6.74766541e-01 -1.57253563e-01 -2.63390422e-01
3.78513068e-01 -2.45393261e-01 -3.18125397e-01 2.94967800e-01
6.22412451e-02 8.60115051e-01 8.31651092e-01 -7.61931956e-01
-3.85867536e-01 -5.54166973e-01 -1.37192667e-01 1.25847146e-01
-5.16547441e-01 4.75623012e-02 -5.65876603e-01 -8.11085105e-01
4.74113911e-01 -4.39232081e-01 -4.09720391e-02 5.05236387e-02
-1.88740671e-01 -1.44631386e-01 4.29149240e-01 -8.42259109e-01
1.43585932e+00 -2.25848198e+00 1.10948188e-02 7.06793368e-03
4.41823393e-01 3.47099364e-01 -1.52910441e-01 -1.33743539e-01
-2.54574716e-01 2.72767879e-02 -3.80711079e-01 -1.96208104e-01
-2.19260216e-01 -2.15092627e-03 1.98544152e-02 3.35123450e-01
2.24987343e-01 7.84088969e-01 -7.15342045e-01 -7.93312550e-01
-2.81386580e-02 4.61917162e-01 -6.97777748e-01 2.90253222e-01
5.80464192e-02 4.09067005e-01 -6.57891273e-01 4.18700248e-01
9.72873867e-01 -8.07456151e-02 -2.37762138e-01 -3.41411084e-01
-1.42695516e-01 4.53526139e-01 -6.78531349e-01 1.95947516e+00
-4.90304023e-01 1.15877718e-01 -6.74959645e-02 -1.49248910e+00
8.79126906e-01 3.84872615e-01 5.25184155e-01 -9.77988899e-01
2.65619934e-01 2.70816654e-01 1.79626122e-01 -7.44819939e-01
-3.35573077e-01 -6.25273436e-02 3.96738112e-01 5.06910205e-01
3.37563306e-01 2.41956338e-01 -8.03390443e-02 -1.22647300e-01
1.01047087e+00 3.16125453e-01 -9.11860913e-02 -3.16810459e-01
4.92082566e-01 -5.58986306e-01 9.36606169e-01 2.66870439e-01
-2.39758804e-01 6.25451386e-01 4.51357931e-01 -3.85893762e-01
-6.87271535e-01 -1.09423912e+00 -5.20057499e-01 9.77018535e-01
2.49778479e-01 -2.16873288e-02 -1.07906020e+00 -8.66723239e-01
-3.29289138e-01 4.70784307e-01 -7.43634284e-01 -5.18681884e-01
-6.88030720e-01 -1.05158460e+00 4.92918015e-01 8.22462916e-01
9.05612588e-01 -1.11948764e+00 -7.09722221e-01 3.02787960e-01
-2.08029583e-01 -6.15744233e-01 -4.82946277e-01 2.93993413e-01
-1.09083617e+00 -6.83573782e-01 -9.79541719e-01 -9.09959495e-01
9.11042929e-01 -3.22553106e-02 7.38381207e-01 -6.68477416e-02
-9.45323855e-02 -2.24425450e-01 -3.39996427e-01 -2.31677562e-01
3.39026183e-01 1.56962380e-01 2.67677512e-02 1.14957497e-01
3.25372040e-01 -8.75621855e-01 -9.69344318e-01 4.46573555e-01
-7.24922776e-01 2.32698724e-01 9.08807278e-01 1.22292995e+00
7.25804985e-01 6.99369386e-02 8.65522087e-01 -6.02070272e-01
3.13263535e-01 -5.75791061e-01 -3.14160317e-01 2.93556452e-01
-5.78455389e-01 8.88900906e-02 6.58140898e-01 -6.49417579e-01
-1.17749298e+00 -1.81010351e-01 -4.24145877e-01 -6.54652536e-01
7.15128426e-03 7.00808227e-01 -3.23110104e-01 -7.13041723e-02
2.46952474e-01 4.74349916e-01 1.82114355e-02 -6.91253781e-01
4.13675942e-02 4.97261167e-01 2.33376682e-01 -6.07458949e-01
3.22154820e-01 1.05523758e-01 -3.71946901e-01 -3.16400230e-01
-8.15608799e-01 2.98574697e-02 -6.63995802e-01 -1.01556167e-01
9.20879245e-01 -9.34985518e-01 -4.35157031e-01 8.76030147e-01
-1.12296414e+00 -3.81984681e-01 -7.05434382e-03 9.71432924e-01
-4.50089991e-01 4.06205580e-02 -7.86664844e-01 -5.87468565e-01
-4.24127340e-01 -1.36512995e+00 6.85067713e-01 3.08351398e-01
2.80282944e-01 -6.11827016e-01 -1.43540666e-01 2.03567818e-01
5.04411578e-01 -1.90332130e-01 1.09522879e+00 -5.38864255e-01
-4.49115068e-01 6.19638860e-02 -5.57682216e-01 4.98334467e-01
-7.26029649e-02 -3.05749863e-01 -1.11349463e+00 -3.30053121e-01
2.57272750e-01 -5.52791774e-01 9.92448330e-01 7.02252448e-01
1.74806547e+00 -2.05711976e-01 -4.12957758e-01 8.03406298e-01
9.73837078e-01 6.07126236e-01 5.67216396e-01 -2.49704011e-02
5.78475058e-01 6.38699889e-01 3.34607482e-01 2.65474588e-01
4.59696800e-01 4.58726674e-01 2.87230730e-01 -1.60542890e-01
-1.25812814e-01 -3.77473325e-01 2.04832509e-01 1.38261318e+00
-1.65448505e-02 1.64211467e-01 -8.22124600e-01 6.07473612e-01
-1.65612721e+00 -7.92083263e-01 3.85372341e-01 2.11848855e+00
1.05999339e+00 1.50355846e-01 -1.33240834e-01 4.96917628e-02
6.21851265e-01 1.23577744e-01 -9.15809929e-01 -7.23140091e-02
2.82558203e-01 4.25022453e-01 1.61433339e-01 3.92805487e-02
-9.53591704e-01 6.65561795e-01 5.89642525e+00 1.20749879e+00
-1.28548872e+00 3.94699693e-01 9.97058809e-01 -1.80897221e-01
-1.68110505e-01 -4.12940502e-01 -6.87416017e-01 9.00896609e-01
6.70519471e-01 2.95102715e-01 4.31986749e-01 7.98691928e-01
5.05082421e-02 4.58276682e-02 -1.02009654e+00 1.05090952e+00
1.46538958e-01 -8.20384264e-01 -1.10229537e-01 -3.08726013e-01
5.22526801e-01 2.41203055e-01 1.95296541e-01 4.93124574e-01
-8.37015212e-02 -9.72938836e-01 8.15144181e-01 5.74480772e-01
1.15807712e+00 -6.78026319e-01 6.75137639e-01 5.75341225e-01
-1.26124978e+00 -8.47335681e-02 -5.13430476e-01 1.64297923e-01
3.12851816e-02 6.46342754e-01 -2.35736445e-01 3.26580495e-01
1.06693721e+00 7.05660582e-01 -4.89196539e-01 1.08332896e+00
-2.33571798e-01 6.43675685e-01 -4.33751851e-01 8.93388782e-03
3.74396928e-02 -1.65572807e-01 -9.72957350e-03 9.36787426e-01
5.84452510e-01 3.62620503e-01 -3.78751685e-03 9.98823822e-01
3.02244164e-02 7.31382445e-02 -2.10258320e-01 2.40904093e-01
5.71924269e-01 9.85695362e-01 -4.66039985e-01 -2.63940215e-01
-5.78101099e-01 5.79002202e-01 6.00991845e-01 3.48979175e-01
-1.04648471e+00 -6.07060194e-01 2.22840354e-01 1.24703966e-01
4.04905498e-01 1.18282251e-01 -2.97528595e-01 -1.17727089e+00
4.94629033e-02 -6.22830451e-01 1.73933983e-01 -7.37110019e-01
-1.14462256e+00 8.53584290e-01 1.31011158e-01 -1.16211784e+00
1.33339390e-01 -4.55801249e-01 -6.44893169e-01 8.97562444e-01
-1.40994763e+00 -9.65146184e-01 -1.56956106e-01 5.91863751e-01
3.53750288e-01 -1.09943390e-01 5.95607996e-01 7.34921813e-01
-1.03777385e+00 1.04862142e+00 1.74810421e-02 2.64784157e-01
4.60317045e-01 -7.85176754e-01 5.49374260e-02 7.76141346e-01
-1.76165104e-01 8.42567384e-01 -4.51117828e-02 -4.00790632e-01
-7.66182899e-01 -1.05478656e+00 5.66286027e-01 2.23149359e-01
3.37364674e-01 -3.69125903e-01 -1.01702809e+00 5.65198779e-01
5.30994125e-02 3.45655501e-01 5.91647089e-01 -5.40937968e-02
-2.37813324e-01 -5.39358974e-01 -1.25561488e+00 3.36288273e-01
1.23363304e+00 -4.54401851e-01 -5.77819347e-01 -8.63592997e-02
7.67771065e-01 -4.38358456e-01 -8.45867038e-01 7.12084234e-01
5.86943567e-01 -1.02606297e+00 7.95202434e-01 -2.34439045e-01
5.74437141e-01 -2.98898160e-01 1.82085484e-01 -1.27176046e+00
-4.73450065e-01 -1.76922277e-01 -2.47947916e-01 1.25346208e+00
5.44613302e-01 -7.71559775e-01 6.96090281e-01 5.83137333e-01
-5.32924831e-01 -1.49036825e+00 -1.19838107e+00 -4.87948656e-01
2.17870176e-01 -3.64284515e-01 8.69547665e-01 8.10299993e-01
-2.56886985e-02 2.79782802e-01 -1.86363265e-01 -4.98568080e-02
4.17522371e-01 -4.12771031e-02 6.18863516e-02 -9.31726635e-01
-2.54466534e-01 -4.69629735e-01 -2.74372488e-01 -1.65320849e+00
-3.58784832e-02 -7.90912867e-01 1.17047116e-01 -1.34029090e+00
3.03658485e-01 -9.30439949e-01 -7.35830069e-01 6.67312443e-01
-2.17764676e-01 6.88405521e-03 -3.05444986e-01 1.17911428e-01
-4.98368621e-01 8.75510514e-01 1.49438035e+00 -1.95951298e-01
3.08280028e-02 -7.94614553e-02 -4.05527294e-01 8.73136163e-01
6.76599324e-01 -4.90866035e-01 -7.83928871e-01 -1.08279657e+00
-2.46074662e-01 1.36151731e-01 1.66196033e-01 -1.10553265e+00
2.46891037e-01 1.30085319e-01 7.06902921e-01 -7.14511514e-01
3.21503878e-01 -6.69276178e-01 -2.15143099e-01 4.14379120e-01
-4.30491894e-01 -2.87437718e-02 2.16354392e-02 2.55620271e-01
-3.22560400e-01 -2.92604357e-01 9.97546732e-01 -3.85564123e-03
-4.39943492e-01 9.97238457e-01 9.36083123e-03 8.73086751e-02
1.03750432e+00 -8.49087834e-02 -1.94595978e-01 1.26563251e-01
-5.91435790e-01 3.66035014e-01 5.88721484e-02 1.11415923e-01
8.31900597e-01 -1.44528818e+00 -4.64818507e-01 5.49373806e-01
-1.18296243e-01 5.08667886e-01 7.16438591e-01 1.19005859e+00
-3.80905420e-01 1.38669848e-01 -2.95642376e-01 -5.57187796e-01
-6.25546277e-01 7.23484159e-01 4.00419772e-01 -1.05814092e-01
-5.52080154e-01 1.29916012e+00 7.41141081e-01 -4.23026741e-01
4.06960964e-01 -4.67333853e-01 -4.66651350e-01 5.63117862e-02
6.14990950e-01 2.50344664e-01 2.04582866e-02 -4.50532228e-01
-3.92126977e-01 7.73020387e-01 -3.34203601e-01 6.29628226e-02
1.34813452e+00 -2.50787903e-02 -1.57341301e-01 3.22622955e-01
1.34038365e+00 -3.62182766e-01 -1.41759574e+00 -3.04022253e-01
-4.33165729e-01 -2.26544693e-01 1.95138052e-01 -8.22349787e-01
-1.60997951e+00 1.49743056e+00 7.67747402e-01 -5.50322756e-02
1.42990136e+00 -9.09266919e-02 9.98625159e-01 -2.70665120e-02
4.44008589e-01 -9.36583042e-01 -4.45983671e-02 2.61013806e-01
5.98670244e-01 -1.10648584e+00 -2.51522422e-01 -1.43303990e-01
-5.14780700e-01 7.89582491e-01 1.09046006e+00 -5.08112796e-02
9.29573178e-01 2.65526444e-01 -1.09643741e-02 -8.27507898e-02
-7.01376140e-01 6.99299946e-02 1.87340245e-01 6.28024340e-01
5.27984202e-01 -4.63693663e-02 -4.75081682e-01 1.22542667e+00
4.09644693e-02 1.28212973e-01 -3.06507945e-01 5.39737880e-01
-3.39551538e-01 -7.85151899e-01 1.37403399e-01 8.59297812e-01
-5.22134542e-01 -3.91779393e-01 3.33751351e-01 3.16873938e-01
6.85488522e-01 6.15862429e-01 -7.46495556e-03 -6.47608876e-01
1.08211666e-01 -3.68137583e-02 3.98990691e-01 -4.96953338e-01
-2.89569497e-01 1.77547485e-01 -3.69295895e-01 -4.70844507e-01
-2.85082638e-01 -5.19949198e-01 -1.38201487e+00 -1.05871093e-02
-6.11334443e-01 3.21510762e-01 3.26293409e-01 8.98563743e-01
3.87573063e-01 7.59283006e-01 8.15327823e-01 -6.53277040e-01
-5.80904841e-01 -1.21261621e+00 -6.05435967e-01 7.59284198e-02
1.99941039e-01 -1.04210603e+00 -9.70120803e-02 6.10889588e-03] | [14.508959770202637, -2.5510776042938232] |
b6cc63d6-37ce-434b-bc6f-7bccd7508cba | a-provably-convergent-information-bottleneck | 2102.04729 | null | https://arxiv.org/abs/2102.04729v2 | https://arxiv.org/pdf/2102.04729v2.pdf | A Provably Convergent Information Bottleneck Solution via ADMM | The Information bottleneck (IB) method enables optimizing over the trade-off between compression of data and prediction accuracy of learned representations, and has successfully and robustly been applied to both supervised and unsupervised representation learning problems. However, IB has several limitations. First, the IB problem is hard to optimize. The IB Lagrangian $\mathcal{L}_{IB}:=I(X;Z)-\beta I(Y;Z)$ is non-convex and existing solutions guarantee only local convergence. As a result, the obtained solutions depend on initialization. Second, the evaluation of a solution is also a challenging task. Conventionally, it resorts to characterizing the information plane, that is, plotting $I(Y;Z)$ versus $I(X;Z)$ for all solutions obtained from different initial points. Furthermore, the IB Lagrangian has phase transitions while varying the multiplier $\beta$. At phase transitions, both $I(X;Z)$ and $I(Y;Z)$ increase abruptly and the rate of convergence becomes significantly slow for existing solutions. Recent works with IB adopt variational surrogate bounds to the IB Lagrangian. Although allowing efficient optimization, how close are these surrogates to the IB Lagrangian is not clear. In this work, we solve the IB Lagrangian using augmented Lagrangian methods. With augmented variables, we show that the IB objective can be solved with the alternating direction method of multipliers (ADMM). Different from prior works, we prove that the proposed algorithm is consistently convergent, regardless of the value of $\beta$. Empirically, our gradient-descent-based method results in information plane points that are comparable to those obtained through the conventional Blahut-Arimoto-based solvers and is convergent for a wider range of the penalty coefficient than previous ADMM solvers. | ['Aly El Gamal', 'Teng-Hui Huang'] | 2021-02-09 | null | null | null | null | ['information-plane'] | ['methodology'] | [ 4.62675542e-02 4.48822323e-03 -4.35479909e-01 2.09315196e-02
-9.38868821e-01 -3.28400701e-01 -1.26898999e-03 1.65555134e-01
-4.78403211e-01 8.88475418e-01 -1.10705353e-01 -2.04121724e-01
-5.86619139e-01 -7.36940265e-01 -7.96763301e-01 -9.55826700e-01
-2.93450266e-01 4.99515146e-01 -4.22585785e-01 -3.32176350e-02
2.85838276e-01 2.14012578e-01 -1.32234550e+00 -1.71055228e-01
1.17877102e+00 1.30421817e+00 1.13378093e-01 2.55319536e-01
-1.36256620e-01 6.51362956e-01 -2.48865366e-01 -1.49754688e-01
5.37390709e-01 -3.99192452e-01 -9.86745298e-01 -7.36814439e-02
2.70627856e-01 -1.21320650e-01 -2.28645653e-01 1.35614789e+00
1.09892316e-01 4.91828710e-01 5.97296298e-01 -1.18222642e+00
-2.43719891e-01 2.81350136e-01 -1.00710297e+00 2.61455387e-01
6.15593903e-02 -7.22513944e-02 1.24332011e+00 -9.14298713e-01
6.78615987e-01 9.28307950e-01 7.03535497e-01 2.75954098e-01
-1.51685095e+00 -6.25398219e-01 5.16680002e-01 1.33727685e-01
-1.61237681e+00 -8.84611458e-02 8.46999049e-01 -5.76183081e-01
7.43278205e-01 2.41087005e-01 6.76158786e-01 3.96568656e-01
-5.76361129e-03 7.98613369e-01 9.33703125e-01 -3.87744457e-01
2.74319172e-01 1.70424700e-01 3.61139864e-01 8.66171181e-01
2.72984356e-01 -9.16828886e-02 -6.75734580e-01 -1.64587110e-01
6.27646685e-01 -1.06425405e-01 -4.90405470e-01 -5.20075381e-01
-1.04600859e+00 1.09851336e+00 4.71863031e-01 1.59930065e-01
-3.61045182e-01 2.93501943e-01 1.23289838e-01 1.90012291e-01
5.61249197e-01 5.66124916e-01 -4.32418078e-01 -1.51420385e-01
-1.28224134e+00 4.66858804e-01 6.42702162e-01 8.04730952e-01
9.26438034e-01 1.41025320e-01 7.25947693e-02 8.49216521e-01
4.31599110e-01 3.73959929e-01 2.13974983e-01 -1.32920051e+00
9.06293392e-01 6.67355359e-01 1.69022322e-01 -1.26997721e+00
-3.97584826e-01 -5.55535555e-01 -1.17040420e+00 8.36949348e-02
7.65732169e-01 -1.47168562e-01 -4.91639882e-01 1.85953069e+00
2.73592383e-01 4.34332117e-02 -3.32249478e-02 1.14888179e+00
4.86843795e-01 9.72373486e-01 -2.30623290e-01 -8.15260530e-01
8.71053338e-01 -8.52654397e-01 -7.54152358e-01 -3.33127409e-01
5.49332976e-01 -6.80874109e-01 9.18206453e-01 5.07922053e-01
-1.64791906e+00 -2.15602726e-01 -9.52330530e-01 -3.30667533e-02
-2.47876924e-02 3.60870846e-02 6.49171352e-01 2.23425314e-01
-8.86325657e-01 7.51534104e-01 -8.60868871e-01 2.60762960e-01
3.43971431e-01 6.61924362e-01 -9.97871831e-02 4.85277958e-02
-9.98836756e-01 8.59411895e-01 1.40538886e-01 3.04154724e-01
-5.33563912e-01 -1.07777393e+00 -7.12269545e-01 4.05364372e-02
4.40898865e-01 -4.82289463e-01 7.57517219e-01 -9.90673184e-01
-1.39277792e+00 7.33877838e-01 -3.80814433e-01 -3.96985948e-01
6.19893074e-01 -1.55900970e-01 1.86515152e-01 1.51037604e-01
1.64576262e-01 4.93450195e-01 7.58301139e-01 -1.17186880e+00
-2.32906222e-01 -4.05779958e-01 5.71930483e-02 3.05403411e-01
-4.18107063e-01 -2.51048148e-01 -4.96985435e-01 -5.45090199e-01
5.12306213e-01 -7.84618139e-01 -4.45525557e-01 1.95103943e-01
-2.26921782e-01 1.74082592e-01 5.51310122e-01 -7.37018168e-01
1.49040413e+00 -1.96009588e+00 5.63555539e-01 5.66644609e-01
3.59420776e-01 2.41505563e-01 1.45490706e-01 2.44872883e-01
-5.87895103e-02 1.87083572e-01 -6.32098377e-01 -2.68885106e-01
-1.19939648e-01 1.73709229e-01 -1.51651964e-01 7.16550410e-01
-1.57719180e-01 5.78798771e-01 -8.01596045e-01 -5.34301043e-01
8.39002877e-02 5.87664247e-01 -1.13656330e+00 5.21612680e-03
-7.42211342e-02 6.56814277e-01 -4.76586252e-01 3.13126922e-01
8.85386467e-01 -4.95594263e-01 3.88289779e-01 -1.90017700e-01
-3.66901636e-01 2.11339191e-01 -1.52117908e+00 1.58838248e+00
-3.83467883e-01 5.55284560e-01 5.25043070e-01 -1.73369932e+00
8.67161930e-01 7.74057731e-02 9.97536063e-01 -6.10835850e-01
-2.40948913e-03 2.31053993e-01 -1.95287257e-01 -1.05417944e-01
2.07940012e-01 -1.61022276e-01 1.82146192e-01 4.11218911e-01
-1.11852519e-01 -1.17633797e-01 4.07204986e-01 8.94590616e-02
8.40038002e-01 3.99589315e-02 1.80464029e-01 -5.01291692e-01
6.33260310e-01 3.12304753e-03 8.58517349e-01 6.40672326e-01
5.19005992e-02 5.52854776e-01 7.77120888e-01 -4.54291344e-01
-9.55708444e-01 -9.72541749e-01 -6.17206931e-01 6.13707483e-01
3.87455821e-01 -4.40001935e-01 -7.61444747e-01 -3.66453499e-01
-1.75577048e-02 5.39930940e-01 -4.27461535e-01 -2.74385586e-02
-8.04570854e-01 -1.16790128e+00 1.24464445e-02 2.46388033e-01
6.37429953e-01 -6.09226108e-01 -5.63414991e-01 3.85775596e-01
-5.08016825e-01 -7.84704566e-01 -4.92491424e-01 1.83742076e-01
-1.14880288e+00 -1.00543356e+00 -6.58683240e-01 -4.13214535e-01
9.73623633e-01 -2.55647562e-02 1.10765219e+00 2.16383979e-01
-2.30723262e-01 2.63607472e-01 4.71246056e-02 -1.03376312e-02
3.53076309e-02 -1.07475938e-02 5.07887825e-02 -8.41743276e-02
-2.06033900e-01 -5.74772000e-01 -7.70753086e-01 3.99288863e-01
-8.18201005e-01 2.08298624e-01 1.49491146e-01 1.11446798e+00
1.06420910e+00 1.30956009e-01 2.92027354e-01 -7.69127071e-01
5.01865089e-01 -5.04619837e-01 -1.01169884e+00 1.70974120e-01
-1.01697612e+00 2.74553418e-01 6.64393008e-01 -3.63086253e-01
-8.03537250e-01 -5.72521091e-02 6.95693865e-02 -6.78495407e-01
6.98226035e-01 7.84374356e-01 9.87364575e-02 -8.78465176e-02
3.42870951e-01 2.66735069e-02 4.85107563e-02 -5.39109528e-01
2.79042751e-01 1.34571984e-01 2.14721844e-01 -8.44093800e-01
5.69995403e-01 4.46609706e-01 2.46486872e-01 -6.78035557e-01
-9.20934558e-01 -3.78028125e-01 -5.45716465e-01 -2.34186694e-01
5.26061535e-01 -7.04871774e-01 -1.08580697e+00 1.39391243e-01
-9.05364990e-01 -2.82717019e-01 -5.45102835e-01 6.39358699e-01
-6.69019461e-01 3.94703418e-01 -4.92056310e-01 -8.50477397e-01
-3.43309462e-01 -1.33926821e+00 6.09819293e-01 -6.67565465e-02
-1.68497279e-01 -1.23971558e+00 -8.33139848e-03 5.71612597e-01
3.02500099e-01 1.86055198e-01 9.28398490e-01 1.77322596e-01
-5.42332888e-01 -5.40734306e-02 -1.91448033e-01 3.42339993e-01
-2.30531618e-01 -7.53929242e-02 -4.65107560e-01 -5.94060719e-01
2.52444267e-01 -9.72375944e-02 8.08629036e-01 7.32601106e-01
1.36945319e+00 -9.49759424e-01 -2.69581318e-01 1.06693637e+00
1.50153220e+00 2.08083186e-02 4.80695069e-01 3.68916363e-01
3.37239087e-01 3.74568790e-01 5.98154902e-01 6.71570718e-01
2.24512517e-01 8.77163291e-01 3.37156087e-01 6.16722833e-03
1.41047150e-01 -3.17247324e-02 4.19008195e-01 9.48937595e-01
-2.69434482e-01 9.15606096e-02 -9.13324475e-01 3.87880892e-01
-1.97515368e+00 -8.53985906e-01 -5.80366664e-02 2.53473425e+00
9.96903777e-01 -6.82125539e-02 -7.11069852e-02 1.89089760e-01
5.00889242e-01 1.75485194e-01 -7.96925306e-01 -3.93809050e-01
1.24540739e-02 4.16942209e-01 6.45500779e-01 7.59322941e-01
-8.27828884e-01 5.35532475e-01 6.41728973e+00 7.01139152e-01
-1.20320880e+00 1.38097912e-01 6.73126996e-01 -5.73657751e-01
-3.39631468e-01 1.11534789e-01 -7.57814646e-01 6.88605964e-01
6.09367847e-01 -3.22545439e-01 8.31425488e-01 7.27386951e-01
2.34685853e-01 -1.57521009e-01 -9.13258135e-01 1.15239012e+00
-3.86475809e-02 -1.51293516e+00 -2.97360599e-01 1.71499178e-01
9.39758301e-01 -1.59377441e-01 3.08385283e-01 4.30600643e-02
1.80831417e-01 -1.00518489e+00 6.10335469e-01 3.55806917e-01
8.25384498e-01 -7.81878829e-01 3.75968963e-01 4.15173054e-01
-1.17247331e+00 -1.82650223e-01 -3.80963802e-01 -1.05926350e-01
1.56616703e-01 8.28986108e-01 -2.75256962e-01 3.71669143e-01
7.36662328e-01 7.36954451e-01 -3.70619036e-02 7.34704018e-01
-1.20081510e-02 3.63813818e-01 -5.76928139e-01 3.28625739e-01
3.22443396e-01 -8.48851204e-01 5.78768075e-01 8.25028539e-01
3.69178414e-01 1.60009161e-01 2.34445810e-01 1.15464914e+00
-7.99158365e-02 2.33594745e-01 -4.10554677e-01 3.60698998e-02
1.43866703e-01 8.92205477e-01 -4.61199582e-01 -3.76325212e-02
-3.72776181e-01 4.93004084e-01 5.50016224e-01 6.68617368e-01
-7.69740283e-01 1.31242096e-01 8.79434764e-01 2.88188368e-01
2.01621413e-01 -3.18698466e-01 -5.05641937e-01 -1.39550567e+00
1.79932624e-01 -7.79564083e-01 5.42774260e-01 -3.61328155e-01
-1.01969516e+00 3.04575592e-01 2.50159860e-01 -1.25128853e+00
-2.16332272e-01 -5.36447644e-01 -1.97029322e-01 8.54656339e-01
-1.46834993e+00 -5.04451990e-01 6.36450052e-02 6.31428838e-01
2.39601389e-01 -2.10614707e-02 7.15090215e-01 4.92730618e-01
-9.45472419e-01 7.56147385e-01 5.57749689e-01 -1.01793528e-01
1.10634767e-01 -9.06891227e-01 -3.17717940e-01 7.70027101e-01
-6.38997741e-03 6.57712519e-01 6.65664315e-01 -3.98643255e-01
-1.57715273e+00 -7.79465199e-01 5.92108011e-01 1.11888029e-01
5.88632762e-01 8.66420660e-03 -9.04186487e-01 7.51982808e-01
-3.03359423e-03 1.94211975e-01 5.00657022e-01 2.87224283e-03
-1.36264920e-01 -3.73447329e-01 -1.27734411e+00 3.29351962e-01
8.30079377e-01 -2.14113519e-01 2.18431167e-02 4.42282528e-01
2.19434157e-01 -7.05388010e-01 -1.02058566e+00 4.29837435e-01
4.20224488e-01 -8.26611876e-01 1.33004427e+00 -5.72191596e-01
4.40733552e-01 -1.01235695e-01 -1.46918058e-01 -1.05824113e+00
-9.86548960e-02 -6.56911552e-01 -4.73593622e-01 8.43174934e-01
4.19700742e-01 -7.45062828e-01 8.54679585e-01 8.71078432e-01
9.50041935e-02 -1.14575744e+00 -1.36464882e+00 -6.69897020e-01
4.47865635e-01 -4.02025223e-01 1.94479108e-01 8.96033525e-01
1.12566151e-01 1.74364503e-02 -3.50948930e-01 5.00022843e-02
9.27329540e-01 3.94347489e-01 4.01355147e-01 -1.03921843e+00
-3.26486111e-01 -6.13563359e-01 -5.11615761e-02 -1.29305077e+00
1.15419090e-01 -1.17332780e+00 -2.19356984e-01 -1.35309923e+00
3.06381553e-01 -1.00713730e+00 -2.42149979e-01 3.25853765e-01
-4.89450805e-02 -5.35240136e-02 2.87677318e-01 4.31692958e-01
-2.76190311e-01 5.29253483e-01 1.37825489e+00 -3.88167679e-01
-3.36062878e-01 -1.92756563e-01 -4.90611404e-01 8.80403399e-01
7.16641665e-01 -4.82703298e-01 -3.48677695e-01 -4.93362516e-01
6.18711710e-01 3.59390020e-01 9.55141336e-02 -6.26002133e-01
2.27518320e-01 -3.55048835e-01 9.60530117e-02 -5.09665012e-01
5.76912999e-01 -5.77398300e-01 1.89871296e-01 3.73804539e-01
-3.68355453e-01 -4.41408195e-02 6.54591024e-02 2.37504989e-01
-3.05508196e-01 -5.70360839e-01 1.02242792e+00 -2.50568300e-01
-5.85671701e-02 4.36960161e-01 -2.55319536e-01 4.47286665e-01
8.77804399e-01 -3.68496962e-02 -5.45592010e-02 -4.98478174e-01
-5.88535309e-01 3.86297435e-01 3.73428881e-01 -2.84601986e-01
6.14266396e-01 -1.27689576e+00 -4.90582675e-01 1.88735649e-01
-4.55911994e-01 5.28840899e-01 1.45050734e-01 1.21943021e+00
-6.15941286e-01 2.24200726e-01 1.80473655e-01 -7.29993224e-01
-8.09139252e-01 4.49313730e-01 5.04657865e-01 -7.48323679e-01
-6.18381500e-01 8.49378169e-01 9.99636054e-02 -2.83294320e-01
4.26240981e-01 -2.06080958e-01 -2.75008492e-02 2.91786909e-01
3.39227349e-01 7.01466143e-01 -8.61686096e-02 -5.84207892e-01
-2.94699043e-01 8.44695210e-01 -1.87736228e-02 -9.71275419e-02
1.49784732e+00 -6.30063266e-02 -3.07696402e-01 2.50079960e-01
1.59232724e+00 -2.76283354e-01 -1.40446234e+00 -1.85743079e-01
-2.84616441e-01 -6.59896910e-01 3.12366247e-01 -4.05906677e-01
-1.69075465e+00 9.47798789e-01 4.91649061e-01 5.91088943e-02
1.11902976e+00 -3.38911682e-01 7.46656775e-01 1.73813730e-01
2.77376413e-01 -1.45736778e+00 1.45426989e-01 4.38241571e-01
9.20965254e-01 -1.18562031e+00 4.32212383e-01 -3.56428623e-01
-4.54361498e-01 9.46697116e-01 4.05351996e-01 -6.62816614e-02
7.59138763e-01 2.30606899e-01 -1.32680237e-01 -2.19119668e-01
-5.13585865e-01 1.86651185e-01 4.60116625e-01 4.28251587e-02
4.85009670e-01 -9.15718302e-02 -7.03298926e-01 4.49423373e-01
-2.36692414e-01 -1.43420815e-01 3.00048620e-01 7.63096511e-01
-1.21885622e-02 -1.13864875e+00 -4.62070554e-01 5.10391593e-01
-4.43498015e-01 -7.78212994e-02 1.57667324e-01 6.86456561e-01
4.66153286e-02 7.90905774e-01 5.45633212e-02 2.02909350e-01
1.51786283e-01 -4.24902514e-02 5.46941996e-01 -1.81096524e-01
-2.99237728e-01 9.71377641e-02 -3.31156015e-01 -6.56948626e-01
-5.13277471e-01 -9.13758397e-01 -1.53090143e+00 -5.20480156e-01
-3.85633022e-01 3.53824854e-01 5.81211746e-01 8.23984206e-01
1.25724852e-01 3.25122923e-01 6.70991719e-01 -7.71194458e-01
-4.54922944e-01 -4.69397873e-01 -5.34124553e-01 4.22431499e-01
3.33226502e-01 -9.39003110e-01 -6.22175038e-01 -1.79524481e-01] | [6.905595302581787, 4.337372779846191] |
1649c3ce-a3c3-4bd1-bf06-9193e75e8d4a | osu_chgcg-at-semeval-2016-task-9-chinese | null | null | https://aclanthology.org/S16-1189 | https://aclanthology.org/S16-1189.pdf | OSU\_CHGCG at SemEval-2016 Task 9 : Chinese Semantic Dependency Parsing with Generalized Categorial Grammar | null | ['Lifeng Jin', 'William Schuler', 'Manjuan Duan'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['semantic-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] | [-7.362309455871582, 3.663600206375122] |
69f632db-b407-46e3-b87f-05ba17d88315 | the-effectiveness-of-a-dynamic-loss-function | 2305.10447 | null | https://arxiv.org/abs/2305.10447v1 | https://arxiv.org/pdf/2305.10447v1.pdf | The Effectiveness of a Dynamic Loss Function in Neural Network Based Automated Essay Scoring | Neural networks and in particular the attention mechanism have brought significant advances to the field of Automated Essay Scoring. Many of these systems use a regression-based model which may be prone to underfitting when the model only predicts the mean of the training data. In this paper, we present a dynamic loss function that creates an incentive for the model to predict with the correct distribution, as well as predicting the correct values. Our loss function achieves this goal without sacrificing any performance achieving a Quadratic Weighted Kappa score of 0.752 on the Automated Student Assessment Prize Automated Essay Scoring dataset. | ['Oscar Morris'] | 2023-05-15 | null | null | null | null | ['automated-essay-scoring'] | ['natural-language-processing'] | [-3.24756712e-01 3.42060804e-01 -2.98956960e-01 -6.99937463e-01
-7.54930377e-01 -4.57249820e-01 8.33123177e-02 4.68844503e-01
-6.01127148e-01 9.31923151e-01 -1.80793628e-02 -3.86583716e-01
-2.57813245e-01 -7.92303562e-01 -2.33890250e-01 -3.07763875e-01
4.65421081e-01 5.75596750e-01 6.98532239e-02 -1.43908858e-01
6.83342040e-01 1.45721734e-01 -1.13883054e+00 5.29225580e-02
1.00321281e+00 9.48827267e-01 -2.31633544e-01 8.64042759e-01
-1.69941097e-01 1.17891145e+00 -8.40392053e-01 -1.04563570e+00
1.89565927e-01 -4.07582849e-01 -6.26937747e-01 -5.10335028e-01
8.65840018e-01 -3.93869728e-01 -3.91068399e-01 9.98807430e-01
2.18765721e-01 4.00362998e-01 8.13064992e-01 -1.08610117e+00
-1.04545248e+00 4.87794489e-01 -4.63705719e-01 1.94276944e-01
3.42867345e-01 -4.45210375e-02 1.65260589e+00 -9.28213179e-01
1.21205807e-01 7.34910369e-01 9.48775351e-01 6.37940764e-01
-1.12731481e+00 -6.03260934e-01 -1.21279053e-01 4.32508707e-01
-9.29900765e-01 -1.12558357e-01 7.20654190e-01 -4.10978705e-01
7.26057768e-01 5.81906922e-02 5.93542814e-01 5.37774265e-01
3.64668101e-01 8.78292620e-01 1.07979882e+00 -5.92269480e-01
1.78023517e-01 1.82264671e-01 9.92322087e-01 7.85898805e-01
1.51892275e-01 2.84120440e-02 -6.23242736e-01 -2.12024793e-01
3.89888257e-01 -1.04380518e-01 -1.52729498e-03 -7.57228583e-02
-4.84667271e-01 1.09594738e+00 3.41304958e-01 -1.02822013e-01
-1.98445052e-01 9.02764648e-02 1.35928348e-01 7.03159988e-01
3.90768260e-01 9.10998225e-01 -4.81998444e-01 -3.28697175e-01
-1.23811221e+00 5.48784375e-01 9.28257406e-01 2.92510331e-01
3.24149489e-01 -1.64597984e-02 -3.84146273e-01 9.88196433e-01
8.97581503e-02 2.88499352e-02 6.85498714e-01 -1.10325289e+00
4.09189045e-01 7.97668993e-01 1.53811067e-01 -9.02835369e-01
-5.09366989e-01 -5.62224567e-01 -3.68036717e-01 5.35646856e-01
7.21618772e-01 -2.60109663e-01 -7.32876599e-01 1.60114205e+00
-1.32822528e-01 -3.68253380e-01 -3.20126504e-01 8.33677948e-01
6.89583302e-01 3.98100823e-01 9.00778919e-02 1.54555902e-01
8.16534400e-01 -1.10680318e+00 -7.07279086e-01 -1.05190456e-01
6.30189717e-01 -4.95060086e-01 1.37165225e+00 6.60479903e-01
-1.57718933e+00 -4.25590187e-01 -1.29341972e+00 -1.76770642e-01
-1.55127555e-01 -6.52321130e-02 4.90428239e-01 8.39138806e-01
-9.92937505e-01 7.82849014e-01 -5.76578617e-01 2.21116528e-01
6.04678333e-01 4.93093103e-01 -1.76765114e-01 9.34532136e-02
-1.08404088e+00 1.41253376e+00 6.03610068e-04 -3.12010527e-01
-2.55183488e-01 -1.03427243e+00 -4.75025892e-01 6.97075486e-01
-2.68337280e-01 -4.50240314e-01 1.59462893e+00 -9.09655333e-01
-1.57948256e+00 7.51446187e-01 1.17435820e-01 -5.48724413e-01
7.01258302e-01 -1.45425037e-01 1.31930292e-01 -3.11980158e-01
-1.90639466e-01 3.51348668e-01 1.97905421e-01 -5.03847897e-01
-6.58879638e-01 -5.46744227e-01 -5.67721836e-02 1.23938896e-01
-7.84145296e-01 -1.14993125e-01 3.32142144e-01 -4.73933518e-01
-9.91603732e-02 -7.19328344e-01 -1.14095762e-01 -5.26111610e-02
2.87660351e-03 -7.80625641e-01 4.31464136e-01 -7.53415644e-01
1.41567206e+00 -1.79999387e+00 -2.17720032e-01 2.44045436e-01
4.39814061e-01 3.17235380e-01 3.19895963e-03 -1.05248138e-01
-4.66675358e-03 1.92732364e-02 -5.93920005e-03 -2.10997894e-01
1.42543331e-01 -1.52379051e-01 -1.62934497e-01 2.86269099e-01
4.04107245e-03 7.79137969e-01 -6.60248518e-01 -2.31634244e-01
-2.51305223e-01 1.07320815e-01 -7.56578028e-01 4.20475304e-01
4.59969379e-02 -2.17182189e-01 -1.25897959e-01 3.56616199e-01
3.02839845e-01 -3.41240764e-01 -2.11574361e-02 5.13633490e-01
6.81232810e-02 7.22483218e-01 -6.50977314e-01 1.11910605e+00
-4.17462260e-01 1.07303929e+00 -1.66238993e-01 -9.38199222e-01
1.30147767e+00 1.66287482e-01 2.76482016e-01 -7.00208306e-01
8.13894048e-02 2.35727608e-01 2.95586884e-01 -2.11973056e-01
7.41870761e-01 -2.97189116e-01 1.38161749e-01 8.16229105e-01
1.03352658e-01 -1.23878568e-01 3.03854290e-02 1.21954672e-01
1.22875929e+00 -1.91452309e-01 1.92759901e-01 -1.97934374e-01
4.05825913e-01 6.25821576e-02 4.98930752e-01 9.15997684e-01
-5.17345011e-01 5.99826455e-01 6.83222294e-01 -6.90154791e-01
-1.34166753e+00 -9.62171912e-01 -2.47390568e-01 1.18522286e+00
-5.20490587e-01 -1.78501859e-01 -6.91489339e-01 -9.48617756e-01
3.96667093e-01 1.05782461e+00 -6.93695426e-01 -4.77477819e-01
-3.98195833e-01 -7.09880710e-01 5.23778558e-01 7.47926533e-01
2.65014470e-01 -7.66425967e-01 -4.19948041e-01 2.27725133e-01
-3.33003663e-02 -3.96255046e-01 -3.89581144e-01 4.48290944e-01
-8.71047318e-01 -9.98843610e-01 -6.92229569e-01 -6.62339032e-01
5.56415558e-01 -1.03289522e-01 1.30922592e+00 4.11613047e-01
-1.15331970e-02 2.10668191e-01 1.73445288e-02 -7.28631496e-01
-1.88241392e-01 3.96411359e-01 -1.07172057e-01 -4.61905450e-01
9.66600776e-01 -2.59926021e-01 -2.63002127e-01 9.94220302e-02
-3.83122861e-01 -4.15017486e-01 8.41035992e-02 1.24528420e+00
9.34347417e-03 -4.52444613e-01 1.16456914e+00 -7.90163219e-01
1.23356628e+00 -5.05703866e-01 -5.99404216e-01 3.34506989e-01
-1.46781969e+00 -3.43549363e-02 8.11969161e-01 -3.14254940e-01
-8.36050212e-01 -2.14653179e-01 4.83252145e-02 -9.20059159e-02
2.96510905e-01 7.53535807e-01 4.39620703e-01 -2.04543680e-01
9.75252211e-01 -9.46842209e-02 8.76441672e-02 -4.73165184e-01
-2.94895709e-01 7.48930812e-01 3.59795630e-01 -5.33635676e-01
3.31580192e-01 -2.62345403e-01 1.33872017e-01 -2.40808979e-01
-1.40735936e+00 -3.20069671e-01 -5.96050143e-01 -2.71656454e-01
2.35175416e-01 -3.88679266e-01 -8.49984407e-01 3.49265128e-01
-9.45632398e-01 -3.04693192e-01 -3.44909102e-01 5.03985584e-01
-3.80927086e-01 1.46253422e-01 -7.77016938e-01 -7.61882901e-01
-4.33668524e-01 -7.37118959e-01 2.14640483e-01 6.24363363e-01
-6.16714239e-01 -1.06726527e+00 4.95388031e-01 6.33381426e-01
6.39011264e-01 -3.17497313e-01 1.13628781e+00 -1.15722966e+00
3.52740623e-02 -8.31218600e-01 -1.94375858e-01 4.70033526e-01
-3.41514587e-01 4.23741005e-02 -9.15140212e-01 -1.08466661e-02
-9.31497514e-02 -6.57631636e-01 9.97526109e-01 4.49194402e-01
1.49227560e+00 -1.26455128e-01 3.38025123e-01 3.53272945e-01
1.06978238e+00 1.75866753e-01 4.18725252e-01 5.65306962e-01
2.87513167e-01 5.51353872e-01 3.38751763e-01 3.71252656e-01
4.92134780e-01 5.87570846e-01 2.04894215e-01 2.84746170e-01
1.47620261e-01 -2.13873208e-01 1.68386146e-01 6.69954658e-01
5.22441417e-02 -9.51206032e-03 -1.21101296e+00 5.17955840e-01
-1.72726738e+00 -9.85847592e-01 -3.05162221e-01 2.37365460e+00
1.02455139e+00 2.14033857e-01 1.68953836e-01 3.23232949e-01
4.99743909e-01 -2.30554994e-02 -5.18075883e-01 -1.04870272e+00
7.92872161e-02 5.77801287e-01 3.61329764e-01 8.31329465e-01
-8.62115502e-01 5.18183708e-01 7.73993206e+00 4.29688096e-01
-7.55379260e-01 -1.00677095e-01 7.40324020e-01 -3.93860579e-01
-1.98274776e-01 -2.43478656e-01 -8.67218792e-01 5.56224167e-01
1.21309924e+00 -5.88301241e-01 2.50263631e-01 9.49813366e-01
5.00413291e-02 -5.41065633e-02 -8.77707660e-01 5.06978631e-01
1.26469508e-01 -1.02357781e+00 -1.47643387e-01 4.18821611e-02
8.67518067e-01 -2.85253853e-01 3.08327913e-01 7.58227646e-01
6.33988559e-01 -1.51703858e+00 4.06810701e-01 7.82831013e-01
4.10417497e-01 -9.55073297e-01 1.04454303e+00 5.40988922e-01
-2.71614343e-01 -3.65899056e-01 -5.30375361e-01 -6.69117510e-01
-4.20832306e-01 6.90692484e-01 -1.03022778e+00 -3.80190879e-01
3.16005886e-01 4.03680682e-01 -6.49958491e-01 1.39228821e+00
-3.17295343e-01 1.03117549e+00 -3.56255993e-02 -5.30632079e-01
3.21734071e-01 -7.08419010e-02 1.57576114e-01 1.04035056e+00
2.17869908e-01 5.11507243e-02 1.21258475e-01 8.61089110e-01
-5.15327930e-01 1.87304407e-01 -2.59313345e-01 1.81646019e-01
4.22738433e-01 1.02975607e+00 -1.22912899e-01 -2.78087735e-01
-4.31234658e-01 4.96375173e-01 9.25114572e-01 5.91269396e-02
-6.40317678e-01 -6.77225232e-01 3.70618045e-01 6.15785383e-02
-7.84820244e-02 -1.07753083e-01 -1.22851038e+00 -9.28579569e-01
-1.18450448e-01 -5.80949008e-01 4.93918329e-01 -7.34051585e-01
-1.64398956e+00 1.41531020e-01 -6.29033804e-01 -6.53341770e-01
-1.46110475e-01 -8.30559552e-01 -8.70928645e-01 1.05276573e+00
-1.42288101e+00 -4.26461637e-01 -2.93099344e-01 1.75752074e-01
3.13053727e-01 -5.12923121e-01 7.94869244e-01 3.13109815e-01
-4.41328377e-01 1.27109921e+00 3.22880775e-01 2.89700449e-01
9.93613660e-01 -1.77722907e+00 6.98317438e-02 2.69582570e-01
-7.03430474e-02 4.52047169e-01 6.19027853e-01 -2.17778236e-01
-5.67103922e-01 -6.61533952e-01 1.34072006e+00 -6.58970416e-01
6.31126344e-01 3.29535127e-01 -1.20755339e+00 7.30826616e-01
-6.78565428e-02 -2.83419430e-01 1.07154918e+00 6.15265787e-01
-5.10659337e-01 -5.51445633e-02 -1.26050830e+00 3.62800539e-01
1.84475854e-01 -3.48718524e-01 -7.56430447e-01 2.21631020e-01
-3.90143543e-02 -3.28926682e-01 -9.76110637e-01 6.50405809e-02
1.03192580e+00 -1.03894150e+00 6.23612165e-01 -1.20089519e+00
1.16769505e+00 4.78512406e-01 1.00308426e-01 -1.55130231e+00
-6.13133371e-01 -3.91359664e-02 -4.85002697e-01 7.95938671e-01
7.43953347e-01 -3.01382452e-01 1.44154382e+00 1.35258520e+00
-4.94245738e-02 -1.15270829e+00 -8.22894990e-01 -6.06288016e-01
7.51214325e-01 -1.56680852e-01 3.97100687e-01 1.19502711e+00
5.42211413e-01 1.89890072e-01 -2.26918310e-01 -4.70877469e-01
7.12314129e-01 -2.23320857e-01 4.84475881e-01 -1.65713871e+00
-6.46875426e-02 -1.11969936e+00 -4.79601353e-01 -8.22161853e-01
3.68073434e-01 -9.98515964e-01 -8.53041708e-02 -1.34618425e+00
4.44882423e-01 -3.49751353e-01 -6.81524158e-01 5.31793237e-01
-3.92886162e-01 2.40523145e-01 1.68245181e-01 1.00662820e-01
-5.84448338e-01 3.93642694e-01 1.07104146e+00 -1.06768392e-01
-9.13184583e-02 2.56119460e-01 -8.98591459e-01 7.76849329e-01
1.01680434e+00 -6.97754979e-01 -4.72566150e-02 -4.14808154e-01
4.71867472e-01 4.63236272e-02 2.13087812e-01 -8.42963517e-01
4.91026253e-01 -3.03359449e-01 7.04705656e-01 -2.22839952e-01
1.56066045e-01 -5.75954437e-01 -6.00142896e-01 4.22916323e-01
-1.14033639e+00 2.23507404e-01 -1.22943401e-01 1.80808902e-01
-1.96899205e-01 -8.85777414e-01 1.06338966e+00 -7.10766017e-02
2.91321110e-02 -5.95710948e-02 -2.49901637e-01 1.98040009e-01
6.81340456e-01 -2.29026139e-01 -4.76866543e-01 -6.84678674e-01
-5.93985617e-01 4.34044421e-01 2.99747735e-01 3.03213686e-01
4.93878752e-01 -1.26613963e+00 -1.01802492e+00 8.17143843e-02
-1.96523517e-01 -4.18042362e-01 -8.83206427e-02 6.03761554e-01
-4.53262538e-01 6.29637480e-01 -3.76241744e-01 -6.18922785e-02
-1.31624186e+00 -2.02584729e-01 7.02927828e-01 -6.31250322e-01
-2.31871262e-01 9.09564435e-01 -3.73524785e-01 -7.98642337e-01
3.77108306e-01 1.42852888e-01 -3.71452391e-01 -3.62506695e-02
7.90490508e-01 8.09239924e-01 3.28218758e-01 1.13849314e-02
1.42646164e-01 3.94045338e-02 -4.59325582e-01 -7.83151202e-03
1.35704887e+00 3.59470725e-01 7.64719993e-02 5.08942187e-01
8.30420375e-01 -3.68494764e-02 -1.09875667e+00 -2.03990638e-01
2.55188257e-01 -5.79946280e-01 5.34792721e-01 -1.37838781e+00
-1.01462865e+00 1.01956570e+00 2.19438225e-01 4.09260333e-01
3.98695529e-01 -6.47722602e-01 8.01248491e-01 7.47550845e-01
-1.10540733e-01 -1.40494454e+00 9.18874741e-02 9.47334051e-01
5.68252146e-01 -1.30374777e+00 2.34337956e-01 4.61123228e-01
-6.25828981e-01 1.42201734e+00 9.80323374e-01 -5.28953552e-01
5.59384286e-01 1.48007959e-01 7.15355948e-02 -1.07485555e-01
-1.04614043e+00 5.35597682e-01 5.10256588e-01 3.70300174e-01
9.33964968e-01 1.41422004e-01 -6.65585876e-01 1.09497190e+00
-5.92978895e-01 1.35340229e-01 8.94495666e-01 5.27153134e-01
-8.35283041e-01 -8.64170611e-01 -4.66036111e-01 1.00879526e+00
-8.28382254e-01 -6.34331480e-02 -7.36130178e-01 4.52652365e-01
-5.20617902e-01 8.19919825e-01 1.29860550e-01 -3.04023743e-01
3.34680259e-01 6.71158731e-01 4.55784082e-01 -4.40840900e-01
-9.04497921e-01 -7.54908144e-01 -8.93922001e-02 -3.83999050e-02
2.11037293e-01 -6.44381702e-01 -1.08110738e+00 -7.73477554e-01
-5.21764398e-01 3.73210132e-01 7.38859415e-01 9.15172458e-01
2.15097927e-02 4.80334520e-01 6.63758993e-01 -3.00079465e-01
-1.35548258e+00 -1.08553183e+00 -7.16361523e-01 2.07372099e-01
3.48065674e-01 -5.59924841e-01 -4.98705626e-01 -4.15703177e-01] | [11.329463958740234, 9.37099838256836] |
451c5cea-80df-42d0-a9ff-b96235cb8641 | enhancing-pre-trained-language-model-with | 2012.15070 | null | https://arxiv.org/abs/2012.15070v1 | https://arxiv.org/pdf/2012.15070v1.pdf | Enhancing Pre-trained Language Model with Lexical Simplification | For both human readers and pre-trained language models (PrLMs), lexical diversity may lead to confusion and inaccuracy when understanding the underlying semantic meanings of given sentences. By substituting complex words with simple alternatives, lexical simplification (LS) is a recognized method to reduce such lexical diversity, and therefore to improve the understandability of sentences. In this paper, we leverage LS and propose a novel approach which can effectively improve the performance of PrLMs in text classification. A rule-based simplification process is applied to a given sentence. PrLMs are encouraged to predict the real label of the given sentence with auxiliary inputs from the simplified version. Using strong PrLMs (BERT and ELECTRA) as baselines, our approach can still further improve the performance in various text classification tasks. | ['Hai Zhao', 'Zhuosheng Zhang', 'Jiayi Wang', 'Rongzhou Bao'] | 2020-12-30 | null | null | null | null | ['lexical-simplification'] | ['natural-language-processing'] | [ 6.04853332e-01 3.60410184e-01 -1.13323405e-01 -6.58802509e-01
-5.87213516e-01 -2.78475076e-01 5.58308423e-01 5.06689429e-01
-6.10080421e-01 7.19984412e-01 4.37713683e-01 -3.44779521e-01
4.30942059e-01 -6.79607391e-01 -4.50516343e-01 -2.18200728e-01
7.53636897e-01 2.97972471e-01 -4.55171801e-03 -4.78580773e-01
3.28401327e-01 1.34773031e-01 -1.37146521e+00 7.24204361e-01
1.45883477e+00 5.29303610e-01 6.12344503e-01 5.03087997e-01
-7.87482500e-01 9.79934812e-01 -9.43828464e-01 -6.66694224e-01
-1.24460854e-01 -5.66368639e-01 -9.94728088e-01 -6.35185987e-02
2.47362033e-01 -2.66110841e-02 -9.51649100e-02 1.04661155e+00
3.84080678e-01 2.65125364e-01 8.04877639e-01 -6.90865636e-01
-6.81195378e-01 1.31053197e+00 -4.13436294e-01 5.36915213e-02
4.43931580e-01 -1.63143888e-01 1.04697084e+00 -1.01725197e+00
5.24330318e-01 1.81755650e+00 7.23588884e-01 7.60286868e-01
-1.31942832e+00 -6.87608302e-01 7.56725490e-01 2.66084909e-01
-1.31452751e+00 -5.87290823e-01 6.63070560e-01 -4.09977362e-02
1.26409721e+00 5.06441832e-01 4.68313009e-01 1.12475228e+00
2.09762231e-01 9.99540448e-01 9.82055187e-01 -8.92007113e-01
1.23398285e-02 3.77481401e-01 5.19211292e-01 3.96824777e-01
2.29253814e-01 -6.63347125e-01 -5.08409083e-01 1.60304289e-02
-1.91040993e-01 5.19330427e-02 -2.80326515e-01 2.35652477e-01
-8.84350002e-01 9.87913847e-01 4.53970879e-01 2.94292361e-01
-1.75489917e-01 -3.38066846e-01 4.53547150e-01 4.05615449e-01
6.59424305e-01 9.64760900e-01 -4.86706674e-01 5.09342924e-02
-7.80735612e-01 3.21984977e-01 8.93174469e-01 8.45665157e-01
5.21756232e-01 -2.23365784e-01 -3.54633570e-01 1.18576336e+00
2.24948645e-01 6.17416143e-01 6.86091006e-01 -6.61204696e-01
6.59682631e-01 7.85792172e-01 -1.40038028e-01 -9.58851516e-01
-5.47390640e-01 -5.23349106e-01 -8.59395027e-01 -2.64041454e-01
9.12577957e-02 1.13314971e-01 -8.38826716e-01 1.69814086e+00
1.56813964e-01 -4.48482305e-01 3.04804474e-01 3.85978043e-01
9.45566237e-01 8.45733941e-01 4.68767852e-01 -3.41335148e-01
1.19927883e+00 -1.13892245e+00 -8.54866028e-01 -6.38552904e-01
1.21398854e+00 -9.03410316e-01 1.62660134e+00 2.59712011e-01
-8.14300060e-01 -5.03088653e-01 -1.11768150e+00 -4.43219692e-01
-3.94983500e-01 1.53763788e-02 2.88424760e-01 4.46132869e-01
-7.84707069e-01 6.10052645e-01 -5.38537920e-01 -4.23919588e-01
5.42849541e-01 1.39263496e-01 1.03636058e-02 -3.78384769e-01
-1.42688715e+00 1.32357359e+00 4.70974058e-01 -2.21864581e-01
-1.69482186e-01 -6.49858832e-01 -9.09105718e-01 1.84416682e-01
4.09488350e-01 -7.55135238e-01 1.44948387e+00 -1.09649599e+00
-1.44004703e+00 8.72875631e-01 -5.23870111e-01 -5.31676531e-01
2.89606392e-01 -3.76568794e-01 -2.68007725e-01 -2.12410629e-01
3.26791108e-01 8.04399908e-01 8.74904156e-01 -1.05194986e+00
-5.04542947e-01 -1.29553407e-01 8.20368305e-02 5.32942772e-01
-4.00108546e-01 5.95412515e-02 -4.06265929e-02 -9.28571820e-01
1.49065807e-01 -8.90621662e-01 -1.58262119e-01 -2.69788951e-01
-5.57064474e-01 -6.28478110e-01 5.34715235e-01 -8.62963378e-01
1.53101146e+00 -2.13075066e+00 2.52604306e-01 -9.94269773e-02
3.23062807e-01 6.00364804e-01 -3.88914108e-01 2.49853671e-01
1.64738655e-01 4.08737987e-01 -2.96656787e-01 -7.59403288e-01
-6.89148530e-02 4.28372324e-01 -4.99901026e-01 -2.15422362e-01
2.64710248e-01 1.08620870e+00 -8.33167374e-01 -3.59392941e-01
1.43894687e-01 2.06501305e-01 -4.50688690e-01 1.12673871e-01
-3.40229660e-01 1.62135497e-01 -2.51914501e-01 1.85840294e-01
5.88686287e-01 -1.43718615e-01 1.76678166e-01 2.16877051e-02
2.48126060e-01 1.03908694e+00 -8.24336410e-01 1.35424805e+00
-8.94548297e-01 6.78748012e-01 -4.88992095e-01 -8.02480459e-01
7.95873761e-01 -1.45673707e-01 -3.06781739e-01 -8.41823697e-01
1.81982055e-01 5.83174229e-02 2.19574854e-01 -4.33293968e-01
5.81573367e-01 -2.56528944e-01 -1.63811311e-01 7.20141411e-01
-2.80622870e-01 -2.62364686e-01 2.12723956e-01 3.57244372e-01
8.37854445e-01 -1.51029468e-01 7.54774511e-01 -3.24984819e-01
7.43670166e-01 -6.03590310e-02 4.38373268e-01 1.08854103e+00
1.23488791e-01 2.79496163e-01 4.13440615e-01 -3.81587684e-01
-8.79025161e-01 -7.87384152e-01 -8.37220326e-02 1.47948122e+00
1.56992041e-02 -7.90169835e-01 -7.42231727e-01 -8.42858911e-01
9.25737247e-02 1.60245574e+00 -4.09880579e-01 -4.91147757e-01
-5.45635223e-01 -6.42744422e-01 5.59876859e-01 4.71661389e-01
3.65466684e-01 -1.01718438e+00 -2.79110491e-01 2.87794590e-01
-5.25892198e-01 -1.07053435e+00 -3.79955441e-01 2.04800576e-01
-7.62736619e-01 -7.13849723e-01 -2.77223110e-01 -6.12624347e-01
8.03228736e-01 5.02542853e-01 1.26795602e+00 3.79965961e-01
2.89753318e-01 -4.11004633e-01 -5.92877984e-01 -8.03292871e-01
-1.01600564e+00 5.07985651e-01 -7.12692663e-02 -3.39546412e-01
5.70570230e-01 -6.15925677e-02 -4.31009606e-02 7.35938400e-02
-9.22832012e-01 5.74026048e-01 5.15841603e-01 8.78453374e-01
3.15163851e-01 -9.81874168e-02 6.95836365e-01 -1.45651376e+00
9.82913315e-01 -1.80133373e-01 -4.51089405e-02 4.79148865e-01
-8.11918616e-01 4.02704686e-01 9.81764376e-01 -5.55972517e-01
-1.12880516e+00 -3.40838671e-01 -4.61323798e-01 1.54959261e-01
-5.42971045e-02 7.29614437e-01 -1.90325662e-01 1.43026739e-01
8.04214120e-01 1.24685436e-01 -1.28096685e-01 -7.15131342e-01
3.86765212e-01 9.46931481e-01 6.45977557e-02 -2.98254311e-01
4.14486825e-01 -2.96001676e-02 -3.17640364e-01 -8.27473760e-01
-1.56971157e+00 -2.31220633e-01 -6.15693092e-01 2.01756775e-01
3.64520401e-01 -9.24759507e-01 -1.17779769e-01 4.35391843e-01
-1.51886201e+00 -1.23899341e-01 -1.35757565e-01 2.22654790e-01
6.03768267e-02 4.03915644e-01 -4.72636789e-01 -6.19792342e-01
-4.10955280e-01 -9.84402180e-01 9.76130962e-01 1.78866550e-01
-8.94170105e-01 -1.08133078e+00 -3.94160539e-01 3.52225333e-01
4.80795354e-01 -3.59167755e-01 1.43802011e+00 -9.79354858e-01
4.60724235e-02 -1.87203780e-01 -1.14835396e-01 5.92192054e-01
2.32359722e-01 -1.34014934e-01 -1.03174829e+00 -1.09863598e-02
2.59525657e-01 -3.20543885e-01 1.11539268e+00 1.99484184e-01
1.09243846e+00 -5.71681917e-01 -4.44675922e-01 3.83759052e-01
7.70761728e-01 7.93270096e-02 4.44756985e-01 3.35428983e-01
8.60440493e-01 6.38684154e-01 5.35234571e-01 2.10155383e-01
6.49290323e-01 4.91175860e-01 -1.94649905e-01 -1.15843982e-01
-3.15960586e-01 -3.85640562e-01 4.37507421e-01 1.21375787e+00
5.24702191e-01 -5.23962438e-01 -7.68082440e-01 -9.21738450e-04
-1.68387938e+00 -5.81638098e-01 -1.37286097e-01 1.85500705e+00
1.34909081e+00 4.64709938e-01 -4.19003993e-01 1.77830249e-01
6.84470952e-01 2.29356840e-01 -5.35344243e-01 -6.86854899e-01
-5.40206075e-01 1.12942748e-01 1.44749001e-01 7.96792746e-01
-9.66892362e-01 1.27444375e+00 6.42099571e+00 9.71282959e-01
-1.21193731e+00 2.07510032e-02 6.34769022e-01 -3.04614939e-02
-4.16861206e-01 -1.13446966e-01 -1.10875177e+00 3.70260358e-01
8.57303321e-01 -5.40317535e-01 4.73647773e-01 6.69514775e-01
2.41238490e-01 -1.77579373e-01 -1.26252496e+00 9.41141009e-01
4.86238688e-01 -1.08837712e+00 7.21404254e-01 -5.05903304e-01
7.00937867e-01 -1.82605103e-01 -1.39411017e-01 6.69299781e-01
1.11866631e-01 -1.09761298e+00 6.19006991e-01 4.72327501e-01
4.16850895e-01 -7.18302727e-01 9.51934516e-01 7.43992269e-01
-7.79992819e-01 7.33297691e-02 -5.33569872e-01 -5.74048102e-01
-1.46131441e-02 5.22426605e-01 -9.21070278e-01 2.80971050e-01
2.53337681e-01 8.71271968e-01 -1.12207496e+00 4.55362946e-01
-6.83579385e-01 5.76790273e-01 -2.67282605e-01 -4.84168321e-01
2.82939095e-02 1.74378201e-01 4.86865938e-01 1.39044297e+00
-6.42305166e-02 7.84911811e-02 2.44854212e-01 7.01429665e-01
-3.67875755e-01 5.41048110e-01 -5.78712940e-01 -2.08238706e-01
5.17039597e-01 9.63326395e-01 -6.25210464e-01 -7.23838687e-01
-4.41954434e-01 1.09049046e+00 6.18541896e-01 2.28073418e-01
-4.23696846e-01 -4.96279210e-01 5.19949615e-01 -4.15967517e-02
4.21895310e-02 1.33309707e-01 -8.46691430e-01 -1.38066745e+00
4.78635803e-02 -1.08768964e+00 3.48147720e-01 -7.47525752e-01
-1.37892509e+00 7.14096725e-01 -1.60268471e-01 -8.17543149e-01
-2.41819844e-01 -5.06454647e-01 -4.39967513e-01 1.04134250e+00
-1.58221185e+00 -9.06732440e-01 -9.56246927e-02 2.90404353e-02
1.09829712e+00 -7.46815354e-02 9.48824346e-01 -7.08040819e-02
-4.94217366e-01 8.28909039e-01 5.08975610e-02 3.28492350e-03
7.44141698e-01 -1.13816857e+00 6.83071554e-01 8.42164278e-01
1.44094005e-01 8.11438739e-01 8.75938416e-01 -7.31127501e-01
-7.88970411e-01 -1.25165701e+00 1.54341459e+00 -5.40311992e-01
4.44374532e-01 -5.54528177e-01 -1.42868745e+00 5.37523270e-01
2.04343244e-01 -8.51222515e-01 8.89605820e-01 3.12805116e-01
-3.24989229e-01 9.61664617e-02 -8.50590825e-01 1.00182712e+00
9.93662417e-01 -6.82113647e-01 -1.22206390e+00 4.92667377e-01
1.06017053e+00 -3.88877928e-01 -3.71585131e-01 1.93197668e-01
3.21338147e-01 -5.54568052e-01 6.68722868e-01 -8.14319491e-01
4.55829173e-01 -5.92306703e-02 2.76858769e-02 -1.72799587e+00
-2.86344737e-01 -2.81150222e-01 4.73877192e-02 1.27877128e+00
5.72968721e-01 -6.77222669e-01 1.69731542e-01 7.88763404e-01
-1.62981033e-01 -7.07553267e-01 -6.40879273e-01 -6.70575976e-01
3.48782629e-01 -6.18476570e-01 6.17736995e-01 8.44725490e-01
1.67621329e-01 9.39612925e-01 -3.42655927e-02 -1.91776156e-01
2.18195125e-01 -7.49916583e-02 5.68054616e-01 -1.29857230e+00
5.93016893e-02 -6.92389786e-01 -9.33999345e-02 -1.37900960e+00
8.09670389e-01 -1.22080910e+00 1.18977323e-01 -1.46545327e+00
2.91160673e-01 -3.19229513e-01 -1.75890580e-01 4.93552715e-01
-7.79838979e-01 1.86545849e-02 4.80874568e-01 2.45995104e-01
-6.96171403e-01 7.01463997e-01 1.19640279e+00 -3.46653908e-01
-2.05005184e-01 1.22764930e-01 -1.07074869e+00 9.29484248e-01
7.65054345e-01 -8.38278234e-01 -3.04537356e-01 -6.99529588e-01
3.20050061e-01 -4.53486115e-01 5.51385097e-02 -5.89990556e-01
1.05835997e-01 -5.89893162e-02 3.30396116e-01 -4.18363661e-01
-4.03877348e-02 -4.14719522e-01 -3.22448254e-01 6.41838968e-01
-9.42953050e-01 1.48601070e-01 2.30532348e-01 2.59597361e-01
-1.93996578e-01 -6.98797524e-01 8.12149167e-01 -8.90948251e-02
-5.60682476e-01 -3.57424706e-01 -5.42239487e-01 2.45490342e-01
6.31406665e-01 5.13048247e-02 -3.69117349e-01 -4.38270479e-01
-5.83303213e-01 3.37432265e-01 5.14394104e-01 6.10493183e-01
5.64711273e-01 -1.11089790e+00 -8.10121000e-01 2.88054287e-01
1.85025722e-01 7.26974085e-02 -7.45978719e-03 5.32877624e-01
-3.91160578e-01 5.71130276e-01 1.09880336e-01 -3.64208937e-01
-1.35598278e+00 3.90888959e-01 2.75862724e-01 -3.35935920e-01
-5.74578583e-01 9.46267366e-01 3.71337295e-01 -5.21619141e-01
3.14329743e-01 -6.83046579e-01 -4.45504397e-01 3.70094925e-02
7.28982270e-01 1.42541915e-01 1.70753166e-01 -6.11312509e-01
-2.69011438e-01 2.74236292e-01 -6.32919610e-01 2.46139526e-01
1.03878963e+00 -5.45530915e-01 -1.69808462e-01 7.41150081e-01
1.01507640e+00 1.32242534e-02 -7.67676890e-01 -5.75626075e-01
3.32987040e-01 -3.02013248e-01 3.67756072e-03 -1.01526773e+00
-4.43527609e-01 9.23149884e-01 -3.84229794e-02 1.21216044e-01
1.03427148e+00 6.95892721e-02 8.57228398e-01 9.83037114e-01
-4.84961681e-02 -1.18343711e+00 -1.10032573e-01 9.67671931e-01
9.45863366e-01 -1.34253931e+00 -2.19355058e-02 -7.61361361e-01
-8.97826374e-01 1.19946504e+00 5.18833935e-01 2.76920885e-01
3.45545501e-01 1.16920888e-01 2.99381405e-01 3.18209022e-01
-1.04211283e+00 -5.12177572e-02 3.48057568e-01 4.74215031e-01
7.32731521e-01 9.59778205e-02 -5.37952542e-01 8.64493728e-01
-4.55254912e-01 -3.20038319e-01 5.65874875e-01 7.98191190e-01
-6.27960324e-01 -1.16064262e+00 -1.64240360e-01 8.06375921e-01
-1.90508589e-01 -6.74891233e-01 -6.42335236e-01 5.42641342e-01
8.92629568e-03 1.07221019e+00 -1.19224153e-01 -3.07067811e-01
5.01578391e-01 4.57761437e-01 2.00729623e-01 -1.17225754e+00
-6.18082285e-01 -1.61144376e-01 2.15735972e-01 -2.12570086e-01
-2.33294904e-01 -7.07385838e-01 -1.44491446e+00 -3.41091245e-01
-3.23707104e-01 7.18558654e-02 3.69424343e-01 1.44534433e+00
3.80475909e-01 5.18258274e-01 5.49651504e-01 -4.31329876e-01
-8.60739708e-01 -1.25199282e+00 -3.44184786e-01 7.25008667e-01
1.60853490e-01 -5.91754496e-01 -3.43690902e-01 8.03919435e-02] | [11.143444061279297, 9.123190879821777] |
b2d3600d-941e-4eeb-a035-2c7e31efc355 | an-online-dictionary-for-dialects-of-north | null | null | https://aclanthology.org/2022.eurali-1.15 | https://aclanthology.org/2022.eurali-1.15.pdf | An Online Dictionary for Dialects of North Frisian | Language is an essential part of communication and culture. Documenting, digitizing, and preserving language is a meaningful pursuit. The first author of this work is a speaker of Söl’ring which is a dialect of the North Frisian language spoken on the island of Sylt in the North Frisia region of Germany. Söl’ring is estimated to have only hundreds of native speakers and very limited online language resources making it a prime candidate for language preservation initiatives. To help preserve Söl’ring and provide resources for Söl’ring speakers and learners, we built an online dictionary. Our dictionary, called friisk.org, provides translations for over 28,000 common German words to Söl’ring. In addition, our dictionary supports translations for Söl’ring to German, spell checking for Söl’ring, conjugations for common Söl’ring verbs, and an experimental transcriber from Söl’ring to IPA for pronunciations. Following the release of our online dictionary, we collaborated with neighboring communities to add limited support for additional North Frisian dialects including Fering, Halligen Frisian, Karrharder, Nordergoesharder, Öömrang, and Wiedingharder. | ['Tanno Hüttenrauch', 'Michael Wehar'] | null | null | null | null | eurali-lrec-2022-6 | ['culture'] | ['speech'] | [-6.33466423e-01 6.81903884e-02 -6.03633560e-02 -1.29205823e-01
-7.91228294e-01 -9.63358879e-01 5.89430451e-01 4.07473266e-01
-7.04958916e-01 5.05764663e-01 5.40738583e-01 -5.07761002e-01
-2.03728266e-02 -5.38287282e-01 -2.42262572e-01 -2.24214375e-01
1.79203704e-01 3.62528265e-01 -1.95178971e-01 -7.09353685e-01
1.43458337e-01 5.98694265e-01 -9.84060526e-01 -1.61379352e-01
1.02311611e+00 -1.58629686e-01 6.84415102e-01 3.87694687e-01
2.81503405e-02 3.04216236e-01 -7.93371320e-01 -4.75922525e-01
5.05519193e-03 -5.11081874e-01 -7.75300682e-01 -1.80475950e-01
2.91665941e-01 -3.54351193e-01 -5.29390931e-01 9.45840776e-01
2.83369094e-01 2.88278878e-01 3.86864990e-01 -5.23676634e-01
-4.83270317e-01 1.00914490e+00 1.20919801e-01 5.33225015e-02
4.26184833e-01 -1.00825608e-01 1.00329161e+00 -9.24789846e-01
1.07381487e+00 1.27098143e+00 4.52202857e-01 2.94803739e-01
-7.45036542e-01 -8.26169014e-01 7.33333528e-02 -1.54973552e-01
-1.71852255e+00 -9.20656204e-01 3.67014349e-01 -3.34501237e-01
7.57091761e-01 1.51112869e-01 1.21138871e+00 8.02682042e-01
-1.90709397e-01 4.92773890e-01 1.02198279e+00 -6.89858317e-01
-3.91765721e-02 3.81119937e-01 -1.36766523e-01 6.10517621e-01
3.23638320e-01 -2.92905360e-01 -9.23473239e-01 -5.06347343e-02
5.57399511e-01 -3.60296935e-01 -4.60205376e-01 3.61086279e-01
-1.23789179e+00 7.24391878e-01 -9.60536078e-02 7.76244164e-01
1.91268817e-01 -2.12771758e-01 2.55990088e-01 6.63788021e-01
2.61112511e-01 3.40931356e-01 -4.05430377e-01 -7.13708103e-01
-7.75866091e-01 3.35478306e-01 1.03242207e+00 1.05313146e+00
3.64702821e-01 2.03877211e-01 6.51502311e-01 1.37817168e+00
5.97110689e-01 6.17410958e-01 5.40141940e-01 -6.39809847e-01
3.71227205e-01 2.54402608e-01 6.39600307e-02 -6.37176394e-01
-2.25571051e-01 -2.82321393e-01 -4.19873655e-01 -1.94717035e-01
7.07278788e-01 -2.41953298e-01 -4.72185194e-01 1.55399144e+00
1.46016583e-01 -9.56500351e-01 1.56245917e-01 8.84647548e-01
6.48719430e-01 5.84495485e-01 -1.99177742e-01 -3.02222937e-01
1.12937415e+00 -5.73406637e-01 -6.50217593e-01 -1.03082947e-01
7.84025252e-01 -1.38980615e+00 9.98418570e-01 5.17282009e-01
-1.36235142e+00 -1.69493809e-01 -8.26585412e-01 -1.58870682e-01
-3.94747376e-01 5.62189892e-02 4.62180674e-01 1.00910354e+00
-1.13159406e+00 3.78474534e-01 -7.64636874e-01 -7.15068519e-01
-2.21629679e-01 3.34561706e-01 -8.34100962e-01 -1.32284118e-02
-1.28110838e+00 8.85611653e-01 1.15610503e-01 -2.89459918e-02
-4.97276276e-01 -6.03282154e-01 -1.16942728e+00 -4.37022537e-01
-7.41437972e-02 2.65730709e-01 1.00905383e+00 -7.34905362e-01
-1.32800937e+00 1.28136933e+00 4.06751456e-03 -4.90624495e-02
5.41196465e-01 3.26864421e-02 -1.17391694e+00 -4.30538505e-02
3.07289213e-01 3.89711529e-01 2.31048957e-01 -5.91369689e-01
-5.92906356e-01 -3.24799746e-01 -1.72837734e-01 3.19712460e-01
-2.81761587e-01 4.15431708e-01 -2.24655807e-01 -1.24926805e+00
2.55084068e-01 -6.88574493e-01 4.39988822e-01 -2.69098848e-01
-1.54537916e-01 -1.47746876e-01 4.22750920e-01 -1.36422110e+00
1.32927978e+00 -2.35356092e+00 -3.18785757e-02 4.55769658e-01
6.10271543e-02 6.69311211e-02 1.29852355e-01 1.06704092e+00
2.54172623e-01 1.99037278e-03 -1.39125558e-02 -1.20847650e-01
1.93993598e-01 3.52479994e-01 2.97584925e-02 8.07562768e-01
-2.79697537e-01 4.41182286e-01 -1.02553344e+00 -1.25299469e-01
6.94484711e-02 3.54662031e-01 -2.84032345e-01 -2.94541389e-01
2.61158407e-01 4.26180661e-01 2.38978177e-01 7.58671105e-01
4.28061098e-01 7.12421775e-01 4.97426867e-01 5.12460768e-01
-8.78296494e-01 9.39850926e-01 -1.17788494e+00 1.61482453e+00
-6.23451829e-01 7.92224944e-01 6.29426420e-01 -5.67016244e-01
1.15208244e+00 2.22505152e-01 1.42593384e-02 -5.04376054e-01
7.72282183e-02 7.70283580e-01 4.09729123e-01 -1.86668441e-01
8.92192781e-01 -4.70350832e-02 -2.40298524e-01 4.95695025e-01
-8.23282674e-02 -4.65485692e-01 3.72725576e-01 2.14026630e-01
5.92965901e-01 -8.91449302e-02 4.89483833e-01 -8.01624417e-01
3.58109206e-01 -1.25076426e-02 4.84292775e-01 8.54551867e-02
-2.15708241e-01 1.31711692e-01 2.09962368e-01 2.66357921e-02
-9.94544625e-01 -1.06561172e+00 -5.47148645e-01 1.04959142e+00
-2.83307701e-01 -6.14211977e-01 -8.31287742e-01 -1.30767860e-02
-1.26408279e-01 7.34857440e-01 -7.35449512e-03 3.81439850e-02
-6.60019755e-01 2.37311170e-01 7.26831913e-01 -1.27944008e-01
6.80400312e-01 -8.45221221e-01 -2.05324918e-01 1.56084359e-01
-1.67860106e-01 -6.08367682e-01 -9.20463026e-01 4.92070094e-02
-1.73712343e-01 -7.44897485e-01 -9.98953044e-01 -1.30968595e+00
2.17025727e-01 -1.02853268e-01 6.55636549e-01 -4.30674590e-02
-6.41378164e-02 3.19038659e-01 -2.77328193e-01 -4.41259652e-01
-1.06076419e+00 2.31222242e-01 2.65207201e-01 -5.40978432e-01
3.44657749e-01 -4.31085318e-01 -1.12020984e-01 -1.11207161e-02
-4.83559757e-01 -3.16055834e-01 2.33222008e-01 4.77876067e-01
3.58850241e-01 -4.57366496e-01 5.80118477e-01 -8.06629360e-01
4.59164053e-01 -2.50262499e-01 -5.80093920e-01 -5.59196062e-02
-1.26136154e-01 -3.46876770e-01 6.34134352e-01 -1.49823621e-01
-7.86800086e-01 -4.29899424e-01 -6.51673555e-01 3.61539692e-01
-1.60271630e-01 3.49992663e-01 -4.29194361e-01 -1.26685441e-01
4.54609573e-01 3.22286338e-01 3.47451754e-02 -6.03634655e-01
2.12197676e-01 1.36948895e+00 5.56490779e-01 -2.89245725e-01
5.43252528e-01 7.50719905e-02 -5.28687298e-01 -1.51841593e+00
-2.93868850e-03 -5.70464194e-01 -4.88452137e-01 -1.56782120e-01
5.11943340e-01 -1.07090139e+00 -4.91767347e-01 7.31103778e-01
-8.32803071e-01 -6.13438487e-01 -4.07813966e-01 7.76925206e-01
-7.59794414e-02 2.58566529e-01 -5.18459320e-01 -7.86349416e-01
-3.16270627e-02 -7.96121776e-01 6.31483078e-01 -1.60687268e-01
-7.78133988e-01 -1.15639865e+00 4.29685116e-02 1.45766452e-01
1.33636862e-01 -9.86249670e-02 9.54273462e-01 -6.46848917e-01
-1.30395785e-01 1.69923354e-04 1.74229056e-01 4.12200868e-01
4.00596887e-01 -1.99306831e-01 -4.01836246e-01 -4.27627176e-01
-1.75231054e-01 -6.27892539e-02 3.32699746e-01 -8.68052766e-02
3.47315185e-02 -4.17996168e-01 1.04820453e-01 3.22416604e-01
1.03860271e+00 3.21276158e-01 4.01647627e-01 4.58775699e-01
3.95165622e-01 4.73480344e-01 4.05653477e-01 4.09523219e-01
7.66678393e-01 4.64357823e-01 -3.27105641e-01 3.78265083e-01
-3.60222399e-01 -3.55022073e-01 9.16846752e-01 1.80952191e+00
1.97785601e-01 6.30351678e-02 -1.00256324e+00 1.16281533e+00
-1.06623614e+00 -4.38191831e-01 -2.03250811e-01 2.52691269e+00
9.90116477e-01 -4.69081774e-02 2.58539885e-01 8.98964033e-02
6.94267571e-01 5.69829121e-02 6.00058138e-02 -6.47049546e-01
-3.45704734e-01 4.93594527e-01 5.31971693e-01 9.75727558e-01
-4.64703202e-01 1.23816454e+00 5.92683315e+00 7.12145388e-01
-1.02550042e+00 -3.28698196e-02 7.84505010e-02 -1.57103151e-01
-6.71491504e-01 6.64038360e-02 -9.90418971e-01 2.98530161e-01
1.08280230e+00 -4.83235270e-01 8.11869323e-01 2.44989857e-01
5.25049031e-01 -1.30019411e-01 -8.37118566e-01 1.00350213e+00
3.05300534e-01 -1.08146131e+00 -2.01740682e-01 3.11534375e-01
6.35320723e-01 2.98870295e-01 -2.13387623e-01 1.18065603e-01
3.39107186e-01 -9.30286407e-01 1.27054548e+00 3.08884233e-01
1.07981193e+00 -1.13812876e+00 2.28405014e-01 3.89691293e-01
-9.90330935e-01 3.08072835e-01 -2.70566344e-01 -3.21957827e-01
-9.26101729e-02 3.74518812e-01 -6.08388364e-01 2.04982355e-01
3.41301054e-01 4.86655235e-01 -2.70138234e-01 7.55603492e-01
-5.80216527e-01 9.72604871e-01 -4.20457065e-01 -3.26333612e-01
4.14227307e-01 -6.11021161e-01 7.54344463e-01 1.12831008e+00
3.98734301e-01 -4.70533110e-02 1.57385886e-01 2.43161663e-01
-2.23035321e-01 7.42892623e-01 -5.02032697e-01 -4.80874956e-01
9.46401715e-01 8.74945462e-01 -5.40110469e-01 3.02938782e-02
-3.93371850e-01 9.28275466e-01 2.57802755e-01 2.39699885e-01
-6.04899898e-02 -1.00742793e+00 9.64656234e-01 7.59107888e-01
-3.85374688e-02 -8.39039147e-01 1.50996715e-01 -6.87888265e-01
7.91962445e-02 -1.26410341e+00 2.23122820e-01 -2.01312318e-01
-6.76405787e-01 6.13567054e-01 -1.49534538e-01 -6.38443172e-01
-4.58022147e-01 -5.82934499e-01 -1.27781346e-01 1.18416226e+00
-8.32027256e-01 -1.09549892e+00 3.12259644e-01 6.69745624e-01
4.97902870e-01 -4.52522963e-01 9.54960227e-01 2.09232315e-01
-2.85207957e-01 8.03359151e-01 5.49359798e-01 3.91138911e-01
7.76748121e-01 -1.11595929e+00 3.98408473e-01 6.75948918e-01
3.72969568e-01 8.86616111e-01 4.62102264e-01 -7.35616744e-01
-1.17228663e+00 -9.14401829e-01 1.64239323e+00 1.86991006e-01
1.01860368e+00 -7.14492917e-01 -6.22291803e-01 9.20385659e-01
4.85993803e-01 -8.76501083e-01 9.00799274e-01 6.97994158e-02
-1.15078114e-01 3.04977614e-02 -1.05749881e+00 9.98626411e-01
9.97669637e-01 -6.10967577e-01 -4.89351481e-01 7.32670367e-01
5.46677172e-01 -6.64359450e-01 -1.04424512e+00 -8.32689703e-01
7.18498170e-01 -5.72136343e-01 2.32466131e-01 1.93717644e-01
-5.95312268e-02 -2.29621232e-02 -3.15448791e-02 -1.63363314e+00
1.14634238e-01 -1.46996903e+00 8.20222735e-01 1.35028100e+00
3.48574072e-01 -1.13291335e+00 2.38824680e-01 2.94654518e-01
-5.37449777e-01 5.10024019e-02 -1.29378986e+00 -8.48785877e-01
5.60804427e-01 -5.61279833e-01 3.99483711e-01 1.05752110e+00
4.85920370e-01 1.42839670e-01 -1.49886319e-02 -1.34816602e-01
2.49521583e-01 -4.99138087e-01 4.84763950e-01 -1.01043594e+00
-3.83935124e-01 -5.39929688e-01 -4.78099346e-01 -7.74140656e-01
1.89760253e-01 -1.21807432e+00 -2.37421587e-01 -1.52142859e+00
-6.99297965e-01 -5.56655586e-01 5.77644587e-01 4.20999676e-01
4.56964493e-01 1.10117473e-01 2.28181988e-01 2.51869321e-01
2.99639076e-01 1.69202685e-01 1.27676737e+00 1.92008942e-01
-5.73313415e-01 -8.98537785e-02 -7.62639463e-01 7.44286835e-01
7.88683951e-01 -1.77442715e-01 -2.45120525e-01 -6.91015542e-01
4.31705952e-01 -2.16348886e-01 -4.68702130e-02 -7.60177195e-01
1.95184574e-01 -1.20845139e-01 1.05395578e-01 -3.08054596e-01
2.63220280e-01 -3.12754065e-01 1.67345300e-01 4.08838570e-01
-1.67448130e-02 1.86754495e-01 2.93512642e-01 -2.27159426e-01
-3.43897462e-01 -4.05359626e-01 9.03825641e-01 -3.28076243e-01
-5.57944417e-01 -6.21459484e-02 -1.04865968e+00 2.34040722e-01
8.88867915e-01 -2.93849379e-01 -4.17185351e-02 -6.41221523e-01
-6.39272630e-01 1.16886951e-01 8.18436503e-01 6.46555126e-02
5.14702022e-01 -1.08959484e+00 -1.01133263e+00 6.63712382e-01
3.24308053e-02 -3.43454391e-01 1.31485820e-01 6.84161365e-01
-1.11494303e+00 3.03726256e-01 2.25749929e-02 1.46119565e-01
-9.97348130e-01 -9.90378410e-02 9.78411809e-02 2.53706813e-01
-9.23800111e-01 9.13512290e-01 -3.14160109e-01 -6.17661476e-01
-4.33770102e-03 -1.29422083e-01 8.67170915e-02 1.59165397e-01
3.62058699e-01 3.60118479e-01 3.08309525e-01 -1.16059673e+00
-5.50023437e-01 -7.46674463e-03 -2.37840831e-01 -7.66130090e-01
9.51457202e-01 -2.48745412e-01 -4.38189894e-01 6.72983229e-01
1.08780944e+00 1.00123465e+00 -4.10399467e-01 3.38975281e-01
-2.82903016e-01 -4.38532799e-01 8.80967677e-02 -5.94103515e-01
-4.66904819e-01 5.13285160e-01 1.52518660e-01 -1.19347364e-01
6.88405573e-01 1.59136817e-01 9.04027343e-01 3.51474732e-01
4.19890851e-01 -1.48858285e+00 -6.63708746e-01 8.59221816e-01
9.95413780e-01 -5.63711882e-01 -2.18002811e-01 -2.87009567e-01
-3.68801564e-01 1.00221038e+00 -2.35533327e-01 5.77607378e-02
6.93015218e-01 2.19139323e-01 4.12403822e-01 1.40908092e-01
-4.54881310e-01 -3.17372233e-01 1.12572685e-02 8.15981865e-01
8.25683355e-01 3.16006243e-01 -9.37554896e-01 5.39352119e-01
-1.19676220e+00 -5.00912249e-01 7.67864883e-01 7.13854253e-01
-4.43447739e-01 -1.43961287e+00 -6.65150225e-01 3.37344915e-01
-3.08624059e-01 -5.52684784e-01 -5.41097581e-01 1.38653672e+00
1.54282093e-01 8.78294885e-01 3.15281451e-01 1.29073605e-01
3.42997491e-01 3.59240845e-02 5.04390299e-01 -7.30475128e-01
-9.36402798e-01 7.25972712e-01 5.61716437e-01 2.02873796e-01
1.06016368e-01 -1.25862145e+00 -1.27357650e+00 -8.51825476e-01
4.79224175e-02 5.76217473e-01 8.41518462e-01 8.42161477e-01
-2.46114954e-01 1.10734825e-03 2.50727743e-01 -1.89843938e-01
-2.13463709e-01 -1.05395508e+00 -1.28413796e+00 -2.50887603e-01
-9.63208228e-02 -5.00378311e-02 -3.03412050e-01 -8.16036016e-02] | [10.469358444213867, 10.228668212890625] |
a55b13a1-3828-4db2-9f43-49fb6064a51b | a-comparative-study-of-feature-selection | 1902.06242 | null | http://arxiv.org/abs/1902.06242v1 | http://arxiv.org/pdf/1902.06242v1.pdf | A Comparative Study of Feature Selection Methods for Dialectal Arabic Sentiment Classification Using Support Vector Machine | Unlike other languages, the Arabic language has a morphological complexity
which makes the Arabic sentiment analysis is a challenging task. Moreover, the
presence of the dialects in the Arabic texts have made the sentiment analysis
task is more challenging, due to the absence of specific rules that govern the
writing or speaking system. Generally, one of the problems of sentiment
analysis is the high dimensionality of the feature vector. To resolve this
problem, many feature selection methods have been proposed. In contrast to the
dialectal Arabic language, these selection methods have been investigated
widely for the English language. This work investigated the effect of feature
selection methods and their combinations on dialectal Arabic sentiment
classification. The feature selection methods are Information Gain (IG),
Correlation, Support Vector Machine (SVM), Gini Index (GI), and Chi-Square. A
number of experiments were carried out on dialectical Jordanian reviews with
using an SVM classifier. Furthermore, the effect of different term weighting
schemes, stemmers, stop words removal, and feature models on the performance
were investigated. The experimental results showed that the best performance of
the SVM classifier was obtained after the SVM and correlation feature selection
methods had been combined with the uni-gram model. | ['Omar Al-Harbi'] | 2019-02-17 | null | null | null | null | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-3.40561837e-01 -5.22480190e-01 7.68500492e-02 -4.10719693e-01
2.56864540e-02 -5.43757558e-01 7.05432177e-01 4.55776662e-01
-5.67034841e-01 7.09921300e-01 1.35684222e-01 -1.67041123e-01
-4.37036783e-01 -7.79201686e-01 2.03710154e-01 -9.94854391e-01
1.63945094e-01 1.91779822e-01 5.55333979e-02 -8.89023840e-01
1.00050282e+00 5.13067544e-01 -1.77125180e+00 1.13784716e-01
1.10217488e+00 6.79835796e-01 3.79238687e-02 2.91415304e-01
-4.22985703e-01 5.16888380e-01 -7.68607259e-01 -3.87101442e-01
9.63745732e-03 -6.18050039e-01 -4.85530704e-01 2.81553626e-01
-3.45464200e-01 1.45141453e-01 2.08925247e-01 9.03313935e-01
4.10283506e-01 2.95648009e-01 1.10210848e+00 -1.19515371e+00
-4.83376473e-01 4.34849650e-01 -5.54686725e-01 -7.58121386e-02
4.58286017e-01 -5.04606247e-01 8.01908612e-01 -1.01042831e+00
3.44456494e-01 1.25239062e+00 4.36090440e-01 -1.08538471e-01
-6.17296815e-01 -3.73220384e-01 -6.36645630e-02 5.12021542e-01
-1.19758010e+00 1.11231662e-01 6.90717399e-01 -4.83578533e-01
7.60851681e-01 1.41096428e-01 6.79285228e-01 1.97123170e-01
5.55292368e-01 3.07619333e-01 1.53424871e+00 -9.35714245e-01
1.90785140e-01 7.88619995e-01 6.73059821e-01 5.03757954e-01
3.91254336e-01 -3.94786090e-01 -3.13722581e-01 -1.67335510e-01
-1.26034275e-01 -1.74911529e-01 -1.59565359e-02 -1.17332963e-02
-7.77042329e-01 1.29458821e+00 -2.14268900e-02 7.03792989e-01
-1.12528913e-01 -8.54504585e-01 7.70976722e-01 5.01702309e-01
3.39984715e-01 5.64697266e-01 -7.37318814e-01 -7.36753494e-02
-5.89082062e-01 -3.00596152e-02 1.11634171e+00 3.98280710e-01
4.67832923e-01 9.48993340e-02 3.60713661e-01 1.03852344e+00
4.68225628e-01 6.72427714e-01 9.50888336e-01 8.97535775e-03
1.91655993e-01 1.07576168e+00 -9.99395698e-02 -1.47426403e+00
-5.54263115e-01 1.54144401e-02 -5.20032823e-01 2.33977810e-01
4.95903701e-01 -4.50082421e-01 -7.43279755e-01 1.06673241e+00
4.15681541e-01 -1.03250015e+00 4.30377096e-01 7.54099369e-01
7.62127817e-01 8.37621987e-01 -2.13733539e-01 -4.11099970e-01
1.44613051e+00 -8.53558481e-01 -9.55688059e-01 5.97375110e-02
7.10859656e-01 -1.28769696e+00 1.06757271e+00 7.30416238e-01
-4.52609628e-01 -2.70819962e-01 -1.14671087e+00 4.52047288e-01
-9.51305389e-01 3.70571494e-01 5.81884563e-01 1.20445573e+00
-5.25347352e-01 1.82829037e-01 -6.24674380e-01 -6.41438484e-01
-1.94129318e-01 4.04215634e-01 -5.25757670e-01 1.24342978e-01
-1.24297881e+00 1.51986706e+00 3.17838728e-01 1.98920637e-01
2.85605401e-01 2.46106982e-01 -7.96267271e-01 -2.42244586e-01
-1.01185367e-01 2.83683896e-01 5.01964688e-01 -1.44000256e+00
-1.39665055e+00 4.65028614e-01 9.82922153e-04 1.34893939e-01
8.17343742e-02 2.67791003e-02 -6.79519832e-01 -2.68263426e-02
4.59820181e-02 -1.36653155e-01 4.85655487e-01 -9.49156404e-01
-7.98866510e-01 -5.91726482e-01 -1.62099734e-01 4.42912370e-01
-4.92221117e-01 4.57067907e-01 2.65318211e-02 -4.31661516e-01
3.49871218e-01 -1.08527625e+00 -9.89996269e-02 -7.97053754e-01
-1.13636982e-02 -1.49977356e-01 8.10776234e-01 -9.32646215e-01
1.33484423e+00 -2.24600863e+00 1.65023446e-01 8.00475955e-01
-5.13955712e-01 2.74490416e-01 3.50437224e-01 6.04636014e-01
1.40734598e-01 -3.34178098e-02 -1.80002555e-01 4.58422005e-01
-1.21947475e-01 1.66358382e-01 2.03626364e-01 3.84654641e-01
-2.16433266e-03 -8.40421766e-02 -4.54962432e-01 -5.88414907e-01
1.31492689e-01 5.11861980e-01 -1.92243114e-01 -1.17670231e-01
2.00892314e-01 -3.07061356e-02 -4.13138419e-01 7.41247237e-01
6.14829421e-01 2.39984944e-01 3.35649788e-01 -2.53945112e-01
-2.85076022e-01 4.45750654e-02 -1.44643438e+00 8.00958872e-01
-4.54811275e-01 4.12101746e-01 -1.36682510e-01 -9.58106935e-01
1.27027571e+00 2.54057705e-01 2.20685929e-01 -5.51323950e-01
6.18192315e-01 5.48834622e-01 4.34268832e-01 -6.26812816e-01
7.06895411e-01 -1.12698190e-01 -5.20611405e-02 4.02124494e-01
-1.31985843e-01 -2.32465237e-01 7.40063071e-01 -6.30333796e-02
1.98095992e-01 -1.16503961e-01 5.01427889e-01 -4.89725649e-01
1.00150847e+00 2.86838144e-01 2.26869270e-01 -3.59001644e-02
-2.61684284e-02 2.41786182e-01 7.02766895e-01 -2.36390397e-01
-5.45300543e-01 -3.77648801e-01 -4.44244951e-01 9.00710046e-01
1.36134180e-03 -3.22652161e-01 -7.70285666e-01 -7.39078522e-01
-7.30806738e-02 7.34902799e-01 -3.23081285e-01 -1.59626335e-01
-3.15366089e-01 -1.41715181e+00 1.51083529e-01 -6.67298138e-02
4.43272680e-01 -9.91276622e-01 -5.03441751e-01 4.15837429e-02
-1.09354258e-02 -5.20156682e-01 -3.97611447e-02 4.48450029e-01
-9.05361354e-01 -1.08872640e+00 -4.13771451e-01 -9.84188676e-01
8.79440069e-01 1.12147026e-01 4.65571463e-01 1.46101519e-01
9.95914862e-02 -8.79045390e-03 -9.75079358e-01 -6.44624174e-01
-4.31002647e-01 3.57397974e-01 8.28483924e-02 -4.09295075e-02
8.26137125e-01 5.91873601e-02 -1.64128378e-01 3.43723863e-01
-9.71822679e-01 -4.70474243e-01 5.17762542e-01 9.65080440e-01
1.21897636e-02 5.16234219e-01 8.14593852e-01 -8.97757590e-01
9.70015168e-01 -4.19685096e-01 -4.05177891e-01 3.40120554e-01
-9.17385936e-01 8.54046608e-04 6.21852458e-01 -2.20310614e-01
-1.10428214e+00 -1.87653825e-01 -1.97353363e-01 8.42653632e-01
1.25467787e-02 9.33121681e-01 -2.02199936e-01 -2.10928231e-01
6.88701034e-01 2.28609279e-01 5.02345681e-01 2.16412656e-02
-1.23031564e-01 1.12109065e+00 -5.52101493e-01 -1.66311532e-01
3.98606539e-01 2.41378415e-02 -6.62282109e-02 -1.16675901e+00
-3.90324920e-01 -3.41245383e-01 -6.75410688e-01 -3.84489357e-01
7.52884328e-01 -4.53477681e-01 -4.23337430e-01 8.04463744e-01
-6.92207098e-01 4.36469287e-01 3.80881220e-01 7.90067375e-01
1.94553251e-03 4.35986489e-01 -4.19386566e-01 -8.29638004e-01
-3.68205547e-01 -1.40942132e+00 6.60999715e-02 5.07838845e-01
-4.20245498e-01 -1.03783178e+00 -2.37190098e-01 3.26909870e-01
4.61366057e-01 1.63982108e-01 1.33074474e+00 -9.68206465e-01
4.17454273e-01 -3.94475102e-01 3.85885276e-02 7.50887871e-01
5.84128082e-01 3.92475247e-01 -4.23740268e-01 -2.51255110e-02
2.96560317e-01 -1.24446489e-01 2.70122886e-01 2.43027180e-01
3.00937772e-01 -2.07566321e-01 9.29981917e-02 -1.61599349e-02
1.44057441e+00 7.82973111e-01 4.98322815e-01 8.33444536e-01
4.30090159e-01 8.76440763e-01 1.23690474e+00 5.08296132e-01
1.92058459e-01 2.22568363e-01 9.50281247e-02 1.66454613e-01
4.62348431e-01 4.94481534e-01 5.43591619e-01 1.28942108e+00
-1.21739328e-01 1.76157970e-02 -8.83624971e-01 3.37164909e-01
-1.50944424e+00 -5.97209096e-01 -5.36608219e-01 2.09906030e+00
8.20846617e-01 3.01701371e-02 9.98888463e-02 9.77523685e-01
5.31039476e-01 -1.11739613e-01 1.57579422e-01 -1.13706648e+00
-3.51095855e-01 1.87185690e-01 5.25253594e-01 5.56362867e-01
-1.02960658e+00 7.46655226e-01 5.41512012e+00 7.41414845e-01
-1.48232174e+00 -2.44438633e-01 3.63431275e-01 3.25578898e-01
-1.45006299e-01 -1.22443832e-01 -8.85930359e-01 5.04262865e-01
6.62729919e-01 -7.25159422e-02 1.28119335e-01 6.38275504e-01
1.38428807e-01 -7.55759716e-01 -1.46149710e-01 6.27728820e-01
5.06969333e-01 -4.48069274e-01 6.54233918e-02 -2.02603638e-01
6.45179451e-01 -3.72947633e-01 -4.49742042e-02 9.46587920e-02
-1.41334310e-01 -8.51752937e-01 3.67279768e-01 3.13643306e-01
4.65719327e-02 -1.20708358e+00 1.25707448e+00 1.48126304e-01
-6.15851820e-01 -7.74352774e-02 -2.37214923e-01 -3.26646835e-01
-5.89814074e-02 5.72764158e-01 -6.58529758e-01 4.67248321e-01
5.48969448e-01 4.47893322e-01 -7.34479427e-01 8.50300252e-01
-1.09362900e-01 5.81983864e-01 -3.42125505e-01 -9.35065091e-01
3.55813742e-01 -8.16489041e-01 2.29836479e-01 1.16289556e+00
4.50493038e-01 -8.54792371e-02 -2.00793475e-01 -1.28498539e-01
7.30021000e-01 1.17773914e+00 -3.90180558e-01 -1.19667791e-01
1.69740900e-01 1.22391737e+00 -1.26775920e+00 -5.83005771e-02
-6.74866140e-01 6.72969520e-01 -3.25482488e-01 4.05172221e-02
-5.97332954e-01 -1.03334951e+00 2.99246311e-01 3.45206596e-02
1.14514850e-01 -2.23017991e-01 -5.33243299e-01 -8.77156138e-01
-8.56896341e-02 -1.32816279e+00 4.08994436e-01 -2.86068410e-01
-1.24563706e+00 7.78180242e-01 -1.68378502e-01 -1.08961737e+00
-6.26636744e-02 -9.34047520e-01 -2.96323419e-01 9.97186065e-01
-1.17000520e+00 -8.94238114e-01 8.62310678e-02 5.08187950e-01
4.34005618e-01 -7.33314812e-01 1.00929332e+00 4.80237484e-01
-5.16489446e-01 4.00205344e-01 6.01131856e-01 2.65375003e-02
9.03778970e-01 -1.12696123e+00 -6.31414592e-01 6.43581092e-01
-2.03102455e-01 5.29093742e-01 7.80011773e-01 -6.01903021e-01
-1.18469810e+00 -3.00317883e-01 1.20294654e+00 5.50228823e-03
6.44116521e-01 -6.49949834e-02 -6.90118611e-01 2.27858469e-01
2.60277808e-01 -7.76127934e-01 1.03285873e+00 2.18407184e-01
-3.30971554e-03 -2.80728757e-01 -1.32055807e+00 5.48169255e-01
-3.57547283e-01 -6.70527518e-02 -6.42908454e-01 1.98075503e-01
2.17163377e-02 -2.28014234e-02 -8.47470582e-01 5.69347106e-02
7.42483258e-01 -7.84994960e-01 5.36515772e-01 -4.20375198e-01
4.75688428e-01 -5.63152075e-01 -2.38647580e-01 -1.46459961e+00
-4.81641032e-02 2.15068951e-01 5.68285227e-01 1.32250714e+00
8.69844258e-01 -8.79062951e-01 3.93021733e-01 6.27743900e-01
2.59496421e-01 -6.98624849e-01 -4.89088148e-01 -3.69069219e-01
1.15271002e-01 1.84593707e-01 3.76418829e-01 1.03751194e+00
5.66984773e-01 5.01739800e-01 -1.87946096e-01 -1.84634000e-01
9.96851474e-02 4.85881306e-02 4.59055156e-01 -1.20374787e+00
7.61472955e-02 -4.17979062e-01 -5.84526420e-01 5.03660478e-02
-2.01036837e-02 -7.31015444e-01 -2.02400729e-01 -1.55826783e+00
-3.18081647e-01 -4.48038012e-01 -8.71533602e-02 3.18554610e-01
-3.08929503e-01 -6.00685738e-02 2.69286871e-01 5.53781725e-02
1.60866484e-01 3.49472791e-01 9.85292673e-01 6.19123280e-02
-4.92790401e-01 1.55043274e-01 -7.25103676e-01 7.82322764e-01
1.04006362e+00 -5.59497952e-01 -4.92139339e-01 -4.59467396e-02
6.11238718e-01 -2.33703613e-01 -7.42563605e-01 -7.21246958e-01
3.18772867e-02 -3.25542569e-01 3.79908264e-01 -4.52158839e-01
8.17926452e-02 -9.85908151e-01 -1.25259608e-01 6.34101629e-01
-9.52257495e-03 5.41989684e-01 1.56999648e-01 -9.16352272e-02
-5.99783003e-01 -7.05839932e-01 7.25266397e-01 2.59453547e-03
-6.01117194e-01 -4.39400762e-01 -9.38310981e-01 -2.54144520e-01
1.18688190e+00 -4.39675748e-01 6.51988238e-02 -1.48828596e-01
-4.71880198e-01 9.62868482e-02 3.33691895e-01 2.94533521e-01
4.62641031e-01 -9.24535215e-01 -7.87142158e-01 1.89074382e-01
9.33952853e-02 -5.50445318e-01 -7.22206309e-02 1.04843724e+00
-9.70990121e-01 1.29954249e-01 -6.91162467e-01 9.01119038e-03
-1.67434394e+00 2.65493959e-01 6.58911914e-02 3.66376676e-02
1.37609512e-01 4.22386408e-01 -4.31959838e-01 -5.59837222e-01
-5.25880046e-02 4.90974337e-02 -1.21566200e+00 8.48641038e-01
3.03561032e-01 5.14287233e-01 3.45840424e-01 -1.05750012e+00
-3.09206724e-01 6.88500464e-01 -2.47020036e-01 -3.50553602e-01
1.17059243e+00 -1.03059657e-01 -7.83731878e-01 5.71977437e-01
9.48632061e-01 4.29040968e-01 2.91747209e-02 2.14233175e-01
2.30668336e-01 -4.13384974e-01 -1.33613274e-01 -1.06230843e+00
-8.54351342e-01 5.66167355e-01 4.84016240e-01 4.79423136e-01
1.21981728e+00 -6.43470526e-01 3.65366846e-01 4.42573100e-01
2.04710647e-01 -1.56033635e+00 -4.99403626e-01 1.00260282e+00
5.96987605e-01 -1.34384179e+00 1.47150695e-01 -4.66326267e-01
-1.00861323e+00 1.41558683e+00 4.27924484e-01 -1.69773728e-01
1.20770633e+00 1.80403188e-01 6.00280166e-01 8.91188830e-02
-2.32306585e-01 -9.95650515e-02 1.78607732e-01 3.74858975e-01
1.05914879e+00 6.33956343e-02 -1.60183895e+00 8.75235915e-01
-5.72245121e-01 -1.86398640e-01 8.75553608e-01 1.04519212e+00
-5.34827173e-01 -1.33287501e+00 -7.58276582e-01 7.64607668e-01
-7.15865493e-01 -3.99583057e-02 -3.59264821e-01 8.94774914e-01
1.42099872e-01 1.35469759e+00 -3.90826970e-01 -4.28057373e-01
2.62459189e-01 2.38184676e-01 1.92800239e-01 -3.07414562e-01
-9.73757505e-01 -4.38471101e-02 2.71793067e-01 2.49322742e-01
-4.86012608e-01 -9.07339275e-01 -1.43774235e+00 -3.11650068e-01
-7.55237877e-01 8.08247566e-01 1.35314262e+00 1.11403084e+00
-1.99350566e-01 1.61330014e-01 8.36433887e-01 -2.29175001e-01
-5.64870000e-01 -1.17723680e+00 -9.41854656e-01 3.80118310e-01
-1.64605066e-01 -6.90773845e-01 -6.34523213e-01 -9.58617479e-02] | [11.010150909423828, 6.931622505187988] |
83d5369f-38e9-4997-944d-678db8bba52e | convnext-based-neural-network-for-anti | 2209.06434 | null | https://arxiv.org/abs/2209.06434v5 | https://arxiv.org/pdf/2209.06434v5.pdf | ConvNeXt Based Neural Network for Audio Anti-Spoofing | With the rapid development of speech conversion and speech synthesis algorithms, automatic speaker verification (ASV) systems are vulnerable to spoofing attacks. In recent years, researchers had proposed a number of anti-spoofing methods based on hand-crafted features. However, using hand-crafted features rather than raw waveform will lose implicit information for anti-spoofing. Inspired by the promising performance of ConvNeXt in image classification tasks, we revise the ConvNeXt network architecture and propose a lightweight end-to-end anti-spoofing model. By integrating with the channel attention block and using the focal loss function, the proposed model can focus on the most informative sub-bands of speech representations and the difficult samples that are hard to classify. Experiments show that our proposed system could achieve an equal error rate of 0.64% and min-tDCF of 0.0187 for the ASVSpoof 2019 LA evaluation dataset, which outperforms the state-of-the-art systems. | ['Wing W. Y. Ng', 'Ying Gao', 'Weiheng Liu', 'Yitao Yang', 'Jinghui Zhong', 'Qiaowei Ma'] | 2022-09-14 | null | null | null | null | ['synthetic-speech-detection'] | ['audio'] | [ 2.70065516e-01 -1.91268101e-01 -1.48570403e-01 -2.74846613e-01
-5.05159974e-01 -2.35871196e-01 6.03523910e-01 -1.18624046e-01
-4.24741507e-01 4.02665764e-01 2.97627568e-01 -5.67058444e-01
2.01375321e-01 -4.91618097e-01 -4.86176401e-01 -8.37738037e-01
6.53481036e-02 -3.51032138e-01 1.31643012e-01 -2.45407850e-01
2.00045988e-01 3.78239304e-01 -1.34982431e+00 4.40808535e-01
8.00430596e-01 1.29627252e+00 2.29527026e-01 5.35786033e-01
8.69523641e-03 5.66035867e-01 -8.94573748e-01 -5.22732556e-01
2.60241777e-02 -3.55676830e-01 -3.07805210e-01 -3.18359852e-01
1.89376548e-01 -2.72069424e-01 -7.33184159e-01 1.43563962e+00
9.41471517e-01 -1.75645351e-01 3.47072244e-01 -1.27309632e+00
-6.30705833e-01 6.59134269e-01 -3.89438003e-01 4.55338866e-01
1.28642321e-01 9.84386504e-02 5.80944419e-01 -9.20341253e-01
1.55592889e-01 1.59248841e+00 6.56937182e-01 7.03482866e-01
-8.22117805e-01 -1.41727281e+00 -9.03611630e-02 6.69617891e-01
-1.46326852e+00 -1.11943698e+00 1.07092023e+00 -8.50955844e-02
8.02661359e-01 1.31982654e-01 1.58145592e-01 1.37625718e+00
2.26750284e-01 8.59158218e-01 1.07924938e+00 -3.94026190e-01
-2.66331762e-01 3.09777021e-01 1.10068284e-01 6.26028419e-01
1.75532743e-01 4.49527413e-01 -6.75658286e-01 -4.67558242e-02
2.14538097e-01 -1.18427940e-01 -5.07291317e-01 -3.31856310e-02
-1.10679710e+00 1.01546061e+00 4.61142570e-01 2.95717031e-01
-1.77469537e-01 -8.44644234e-02 7.10431576e-01 2.92070657e-01
2.59524107e-01 1.65198240e-02 -2.83031434e-01 2.00900957e-01
-9.95076835e-01 -3.79042253e-02 3.89718145e-01 6.12510920e-01
3.22791874e-01 3.81999344e-01 -6.69869259e-02 9.40106213e-01
5.76193154e-01 9.16106284e-01 6.44006014e-01 -2.11110204e-01
6.45110190e-01 -1.13889880e-01 -2.12208912e-01 -1.21945596e+00
-3.00349921e-01 -7.60637283e-01 -9.74067032e-01 2.27327533e-02
-5.11692390e-02 -1.27648830e-01 -1.09645379e+00 1.59161997e+00
1.72575787e-01 2.28586778e-01 1.72288388e-01 7.56109655e-01
8.32815051e-01 6.74673975e-01 -2.11733803e-01 -4.16855484e-01
1.33421326e+00 -1.03947854e+00 -1.08118236e+00 -7.00518712e-02
2.29281187e-01 -1.04090524e+00 6.93291843e-01 4.93241161e-01
-4.90023285e-01 -7.67955124e-01 -1.32579124e+00 3.04847300e-01
-2.20863208e-01 -7.08085969e-02 4.30395395e-01 1.39940131e+00
-9.70880091e-01 4.08254385e-01 -5.96159756e-01 6.40049251e-03
7.01950192e-01 2.92195380e-01 -3.39564949e-01 5.72937392e-02
-1.46417284e+00 5.25140405e-01 3.05844933e-01 1.46497324e-01
-1.19703650e+00 -3.52501661e-01 -7.22340405e-01 6.23417720e-02
-7.64382109e-02 -3.09030861e-01 9.64055121e-01 -8.61112714e-01
-1.69518268e+00 4.90020603e-01 -1.44352809e-01 -7.50509918e-01
3.73109132e-01 -1.04270399e-01 -1.15830863e+00 3.53225261e-01
-7.07872510e-02 4.47297961e-01 1.40580344e+00 -9.11458850e-01
-6.20138407e-01 -2.61209548e-01 -5.19968688e-01 -2.58509696e-01
-6.84241354e-01 4.02254730e-01 -4.60774563e-02 -8.34328651e-01
1.93101972e-01 -8.05139482e-01 2.23689258e-01 -2.96938419e-01
-5.14087260e-01 -1.50954142e-01 1.29192507e+00 -1.01603043e+00
1.14529228e+00 -2.52947521e+00 -5.16887307e-01 9.45213065e-02
-1.25422060e-01 1.04026902e+00 -1.75924808e-01 1.86872736e-01
-1.66258439e-01 1.77442014e-01 2.91831400e-02 -3.56817156e-01
-8.80001932e-02 -4.00546581e-01 -5.69135487e-01 8.80694032e-01
2.27915514e-02 2.64133632e-01 -8.38744104e-01 -5.45688093e-01
3.62944156e-01 1.04544377e+00 -5.29616535e-01 1.22862637e-01
3.25138956e-01 2.39782795e-01 -4.18069929e-01 6.21744394e-01
1.23734260e+00 1.44619420e-01 -1.75378516e-01 -9.54376087e-02
1.67262852e-01 7.97812760e-01 -8.16549599e-01 1.43595219e+00
-4.17978913e-01 9.88559783e-01 3.27240288e-01 -9.62381303e-01
1.13483775e+00 5.94284713e-01 1.93193600e-01 -7.09449649e-01
4.13104564e-01 2.76002645e-01 7.64874294e-02 -5.76483667e-01
1.13892436e-01 -1.24716364e-01 2.72040337e-01 6.84366971e-02
2.61712223e-01 2.06238687e-01 -5.35667062e-01 -1.14518022e-02
8.73014748e-01 -4.75239962e-01 -4.07692902e-02 -1.68034345e-01
9.67159629e-01 -6.31163299e-01 6.34892464e-01 6.82383537e-01
-6.83558881e-01 3.48495781e-01 -7.36830384e-02 -2.30165079e-01
-8.31200302e-01 -8.64496469e-01 -5.51307023e-01 7.75714159e-01
1.53784439e-01 -2.99535394e-01 -8.42142940e-01 -7.73796499e-01
-1.18289948e-01 5.23147643e-01 -2.06974000e-01 -2.55589426e-01
-4.62956637e-01 -5.24946034e-01 1.11127222e+00 3.16115282e-02
9.40691531e-01 -1.11124098e+00 -1.01783589e-01 2.92789936e-01
-2.91102231e-01 -1.22973347e+00 -5.13221085e-01 -1.43202871e-01
-5.65602958e-01 -7.02879310e-01 -7.22399235e-01 -8.84224296e-01
3.06777298e-01 6.76957071e-01 4.24460858e-01 1.75374374e-01
6.65609464e-02 -3.62986088e-01 -4.98119920e-01 -5.49015343e-01
-5.40926874e-01 -7.45673701e-02 4.25222337e-01 5.47207594e-01
4.98212993e-01 -4.42591608e-01 -5.76900661e-01 3.72673690e-01
-6.01274908e-01 -1.34810656e-01 3.57497901e-01 1.13099694e+00
-3.11289895e-02 2.38784820e-01 9.02471304e-01 -2.92248666e-01
4.85644966e-01 -1.97947606e-01 -4.07970846e-01 4.64604944e-02
-7.90664256e-01 -2.22811028e-02 5.73080242e-01 -5.79158723e-01
-8.71162653e-01 -1.27310455e-01 -5.58261335e-01 -6.51561975e-01
-1.13011107e-01 4.76316698e-02 -4.50474203e-01 -2.62666047e-01
4.00025815e-01 5.76534152e-01 2.20529616e-01 -5.10172725e-01
4.14899886e-02 1.42314017e+00 5.13094544e-01 5.86400554e-02
7.65420258e-01 4.51986730e-01 -3.41004819e-01 -1.21143448e+00
-3.64603460e-01 -4.05736417e-01 1.49785414e-01 -1.56277474e-02
6.91627443e-01 -9.87326622e-01 -1.01491416e+00 9.64334428e-01
-1.37813115e+00 3.32779527e-01 5.75806677e-01 7.09136426e-01
-1.84097439e-01 7.81140447e-01 -6.02788627e-01 -1.07325995e+00
-7.64498711e-01 -1.43814349e+00 8.02666545e-01 1.44268170e-01
2.24634185e-01 -4.90651190e-01 -1.69516921e-01 4.01851624e-01
8.06379139e-01 -2.84270734e-01 5.17609835e-01 -8.70677590e-01
-2.34210372e-01 -4.65717643e-01 -2.85808504e-01 7.80363500e-01
1.64913014e-01 -3.87905598e-01 -1.67831659e+00 -5.56019008e-01
4.25778955e-01 -6.11925125e-02 1.32490659e+00 3.26973021e-01
1.21568227e+00 -5.33930123e-01 -4.49695170e-01 1.01261258e+00
9.89099503e-01 4.43273813e-01 7.77331173e-01 2.45459765e-01
5.16199768e-01 5.12556255e-01 3.54234129e-01 2.74123937e-01
-2.16853973e-02 7.34236538e-01 4.79600847e-01 2.49990687e-01
-3.18988085e-01 -4.45187986e-01 6.58875823e-01 9.03050959e-01
4.54049855e-01 -4.72253829e-01 -6.25202358e-01 5.15145183e-01
-1.30306387e+00 -1.17301834e+00 1.92604542e-01 2.11693096e+00
5.84064901e-01 3.38354677e-01 -1.20195515e-01 6.57497942e-01
1.15903401e+00 6.58180296e-01 -2.35526726e-01 -3.53466660e-01
-2.12289840e-01 9.40009207e-02 5.52996993e-01 5.48390567e-01
-1.41894853e+00 9.90038097e-01 5.71853685e+00 1.27622962e+00
-1.54434764e+00 3.04086357e-01 5.06109834e-01 1.76836252e-01
-1.09959520e-01 -4.29315776e-01 -1.04473281e+00 7.67479062e-01
1.10232949e+00 1.03618778e-01 5.04099727e-01 7.41612434e-01
8.39348733e-02 6.68038070e-01 -5.07965624e-01 1.30362141e+00
3.74033362e-01 -1.23363447e+00 -1.39596671e-01 9.60090384e-02
3.64824533e-01 2.01633751e-01 3.01486194e-01 4.96449992e-02
-8.48712772e-02 -1.14070737e+00 7.69255280e-01 -1.04496486e-01
8.86231065e-01 -9.48287845e-01 8.74867320e-01 2.40815639e-01
-1.02338135e+00 -3.31952006e-01 -4.46155548e-01 2.45425820e-01
1.14840329e-01 6.89757288e-01 -1.21713805e+00 3.43862593e-01
6.03906214e-01 5.08590281e-01 -1.48733422e-01 9.42654490e-01
-4.28759515e-01 9.23871398e-01 -3.21777135e-01 -1.79770723e-01
3.32819112e-02 4.68999445e-01 6.07353449e-01 1.49771285e+00
4.09059763e-01 -3.44359607e-01 -2.50470549e-01 3.48897099e-01
-1.46640345e-01 -2.20732503e-02 -6.61120355e-01 -1.67312756e-01
7.16171801e-01 8.89861643e-01 -3.10650438e-01 -2.40911767e-01
-8.02761018e-02 8.71468604e-01 -1.50438592e-01 3.00257891e-01
-7.98713446e-01 -9.35788035e-01 8.38556886e-01 -1.59795672e-01
6.07706666e-01 -1.14518560e-01 -1.78759634e-01 -1.02652276e+00
-1.59409925e-01 -1.20545495e+00 1.11481428e-01 -3.59854400e-01
-1.01865375e+00 9.68562961e-01 -6.89016342e-01 -1.19143009e+00
3.21944356e-02 -6.04907513e-01 -4.39114094e-01 8.83355141e-01
-1.73544002e+00 -1.05171180e+00 -6.61238469e-03 8.23512197e-01
5.94085038e-01 -6.42144144e-01 7.19071865e-01 5.85738182e-01
-5.43532670e-01 1.15776217e+00 7.65144974e-02 4.98854399e-01
6.86460972e-01 -3.53644282e-01 6.27733409e-01 1.05597115e+00
1.42032787e-01 5.69351494e-01 7.86107063e-01 -4.94686872e-01
-1.21676219e+00 -8.72351646e-01 1.19020271e+00 1.93458050e-01
3.98942709e-01 -4.07965273e-01 -7.25887954e-01 1.47923946e-01
2.67329305e-01 2.35909730e-01 4.36959386e-01 -2.54651546e-01
-7.37347424e-01 -3.72556776e-01 -1.29279578e+00 1.14935443e-01
8.47018361e-01 -1.02660644e+00 -4.17580605e-01 3.13556790e-01
9.77281272e-01 -6.32006004e-02 -1.75881520e-01 3.44884396e-01
6.65728509e-01 -9.28802609e-01 1.26897871e+00 -5.13584912e-01
-2.86198072e-02 -2.46755764e-01 -2.80197561e-01 -1.23415291e+00
-2.54447669e-01 -9.72562909e-01 -2.91093532e-02 1.39019573e+00
3.42802316e-01 -7.54525721e-01 5.68973780e-01 -3.84353429e-01
-2.46354148e-01 -2.66482115e-01 -1.17581558e+00 -9.20375109e-01
-2.58001894e-01 -6.56164050e-01 8.15929770e-01 9.20009971e-01
2.44132429e-02 1.11798659e-01 -8.48342538e-01 6.20868683e-01
8.27734053e-01 -3.01922023e-01 6.04939997e-01 -8.80482376e-01
-2.15965599e-01 -3.34800571e-01 -6.51908040e-01 -1.18430889e+00
3.19895476e-01 -7.61918426e-01 6.46844953e-02 -8.06914926e-01
-1.96707979e-01 -3.21360916e-01 -7.62907028e-01 1.96360022e-01
-8.72503519e-02 2.47315973e-01 2.40950093e-01 1.08341081e-02
-1.09568909e-01 6.01764262e-01 1.06084907e+00 -5.25179803e-01
7.64255747e-02 9.91806537e-02 -4.31436062e-01 4.92872894e-01
8.00282300e-01 -7.47751355e-01 -1.98445037e-01 -3.73505861e-01
-3.64095777e-01 2.31369108e-01 2.06984654e-01 -1.13209891e+00
2.78274089e-01 1.05197065e-01 2.21281379e-01 -6.67162895e-01
4.41221863e-01 -6.88773751e-01 -1.92924485e-01 8.48627627e-01
-3.27111095e-01 -3.12972844e-01 8.41294706e-04 4.82606620e-01
-3.79771978e-01 -9.26395357e-02 1.01322114e+00 3.48963588e-01
-3.64672124e-01 1.69806674e-01 -2.29727000e-01 -3.43878955e-01
6.47360265e-01 -1.22710653e-02 -5.56169271e-01 -4.49538887e-01
-2.04707772e-01 -2.54504889e-01 7.70600513e-02 5.92777967e-01
9.15102303e-01 -1.21589684e+00 -8.47437441e-01 8.39447260e-01
-1.69786453e-01 -5.53760588e-01 3.90635341e-01 5.10285735e-01
-4.53231186e-01 7.98414230e-01 -2.25672558e-01 -5.45870662e-01
-1.42494392e+00 9.01313841e-01 2.58402735e-01 6.10832237e-02
-5.44746161e-01 1.17510211e+00 1.43477231e-01 -5.23861498e-02
4.61398900e-01 1.17598943e-01 -1.73145398e-01 -1.37562513e-01
1.03557301e+00 2.75193810e-01 2.10429847e-01 -8.48693907e-01
-7.10282087e-01 2.16412961e-01 -1.82894588e-01 -1.13590501e-01
1.13534105e+00 -1.83729932e-01 3.05786669e-01 -2.02639490e-01
1.51451492e+00 2.37371281e-01 -1.00761282e+00 -3.32742065e-01
-3.48736435e-01 -7.24674463e-01 5.08525789e-01 -6.14245772e-01
-1.32916939e+00 1.38738036e+00 1.04441309e+00 2.62132078e-01
1.09047616e+00 -3.74592721e-01 1.02912605e+00 9.59552750e-02
3.19893628e-01 -8.31367552e-01 7.15292022e-02 4.57809299e-01
7.62419105e-01 -1.17290044e+00 -3.25711668e-01 -3.24164987e-01
-3.62587005e-01 9.62297916e-01 5.43067604e-02 6.32300302e-02
7.69530654e-01 1.07505135e-01 2.31091440e-01 2.61596233e-01
-3.50541204e-01 2.85225242e-01 6.12745918e-02 9.58102345e-01
1.93505004e-01 2.28814468e-01 -1.46975711e-01 5.04565775e-01
-4.46700245e-01 -2.02229515e-01 2.28745252e-01 5.41872561e-01
-6.06374741e-01 -9.44716990e-01 -7.50528514e-01 2.21701145e-01
-8.72999787e-01 -2.51538843e-01 -4.32303362e-02 1.52190268e-01
7.13684410e-02 1.55781472e+00 -6.51893169e-02 -1.00944650e+00
-1.32563993e-01 4.23266813e-02 1.05023496e-02 -9.18394711e-04
-3.62504095e-01 1.58229619e-01 1.61693126e-01 -3.43878806e-01
-2.44582906e-01 -4.56848085e-01 -8.26649368e-01 -5.55285394e-01
-8.57875943e-01 1.69921249e-01 1.14596868e+00 7.84224570e-01
3.95937800e-01 4.41195965e-01 1.17187178e+00 -7.53220916e-01
-8.51348042e-01 -1.21876872e+00 -4.69755530e-01 2.82532692e-01
1.18561876e+00 -6.07548296e-01 -5.17200887e-01 -1.76692739e-01] | [14.098994255065918, 5.853119850158691] |
b08ab6f0-b9fd-4202-a547-f74db6ea2bea | consistent-jumpy-predictions-for-videos-and-1 | null | null | https://openreview.net/forum?id=S1gQ5sRcFm | https://openreview.net/pdf?id=S1gQ5sRcFm | Consistent Jumpy Predictions for Videos and Scenes | Stochastic video prediction models take in a sequence of image frames, and generate a sequence of consecutive future image frames. These models typically generate future frames in an autoregressive fashion, which is slow and requires the input and output frames to be consecutive. We introduce a model that overcomes these drawbacks by generating a latent representation from an arbitrary set of frames that can then be used to simultaneously and efficiently sample temporally consistent frames at arbitrary time-points. For example, our model can "jump" and directly sample frames at the end of the video, without sampling intermediate frames. Synthetic video evaluations confirm substantial gains in speed and functionality without loss in fidelity. We also apply our framework to a 3D scene reconstruction dataset. Here, our model is conditioned on camera location and can sample consistent sets of images for what an occluded region of a 3D scene might look like, even if there are multiple possibilities for what that region might contain. Reconstructions and videos are available at https://bit.ly/2O4Pc4R.
| ['Murray Shanahan', 'S. M. Ali Eslami', 'Edward Lockhart', 'Fabio Viola', 'Marta Garnelo', 'Danilo Rezende', 'Ananya Kumar'] | 2019-05-01 | null | null | null | iclr-2019-5 | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 3.38265687e-01 2.10385188e-03 -8.04910213e-02 -1.94180548e-01
-7.60949910e-01 -6.98695064e-01 6.51397407e-01 -4.55538690e-01
1.59947313e-02 5.40130258e-01 3.27868789e-01 -2.04203710e-01
4.02395308e-01 -6.49441004e-01 -1.06503105e+00 -4.90921170e-01
-9.02469233e-02 1.12414867e-01 4.65449959e-01 3.17629367e-01
1.24628715e-01 4.86893654e-01 -1.68599129e+00 6.75962448e-01
1.63319036e-01 7.21796155e-01 6.64631546e-01 1.27581847e+00
2.10241020e-01 1.23004401e+00 -2.55613595e-01 -3.45134437e-02
4.64683473e-01 -5.65004885e-01 -6.93105042e-01 8.08219671e-01
6.06100321e-01 -1.02973902e+00 -6.04764640e-01 8.04155052e-01
1.09124027e-01 1.81789219e-01 3.13596487e-01 -1.01085043e+00
-2.58983433e-01 1.23119548e-01 -3.34742785e-01 1.55916229e-01
9.23011363e-01 5.86267471e-01 7.64094055e-01 -9.29739773e-01
1.00654829e+00 1.23716998e+00 4.07563269e-01 5.79481840e-01
-1.52219987e+00 -3.92412275e-01 3.46586883e-01 1.99961722e-01
-1.22852683e+00 -8.87115419e-01 7.39400387e-01 -4.78323489e-01
7.76361287e-01 2.09897295e-01 9.68766809e-01 1.26318669e+00
1.94297537e-01 8.35893333e-01 6.77825212e-01 -3.15528512e-01
3.08014095e-01 -2.36284569e-01 -3.44550639e-01 6.14588261e-01
-3.20454806e-01 2.11710885e-01 -6.95383608e-01 -1.61489099e-01
1.24747300e+00 3.05949301e-01 -6.31453812e-01 -3.58948767e-01
-1.40308034e+00 4.42404360e-01 -2.27629617e-01 -1.90633491e-01
-6.25121057e-01 4.12470549e-01 -4.37992513e-02 2.98268765e-01
3.72594744e-01 -4.05511037e-02 -2.36854255e-01 -2.91955113e-01
-1.23186100e+00 4.22941327e-01 5.78018486e-01 1.18935168e+00
8.35909545e-01 1.98927239e-01 -1.00617699e-01 4.20556068e-01
1.71875924e-01 5.21775305e-01 1.12249449e-01 -1.91107774e+00
2.63693064e-01 -1.62248462e-01 4.35391247e-01 -7.78529644e-01
1.79213509e-01 3.16249132e-01 -4.39799041e-01 4.22261983e-01
4.79524225e-01 -2.92131692e-01 -7.99445450e-01 1.55743349e+00
3.78648996e-01 8.47684264e-01 -5.44053093e-02 9.10364151e-01
3.09227198e-01 1.13965905e+00 -2.87152201e-01 -3.96297187e-01
8.86049867e-01 -8.27222228e-01 -3.36831510e-01 -1.99742839e-01
2.33804315e-01 -9.23269928e-01 9.21881020e-01 4.05304462e-01
-1.61361516e+00 -7.72892535e-01 -6.64574623e-01 -1.32501557e-01
4.73194540e-01 -4.45081182e-02 3.17894250e-01 1.28837347e-01
-1.33536720e+00 6.24027610e-01 -1.01626289e+00 -1.25177205e-01
8.26271549e-02 7.73421377e-02 -3.27190191e-01 -1.96437076e-01
-6.67590678e-01 3.62271547e-01 1.25557363e-01 -3.50383408e-02
-1.49045491e+00 -5.41857183e-01 -7.91826129e-01 2.58244500e-02
3.00662279e-01 -8.54747057e-01 1.55718124e+00 -1.27662849e+00
-1.59906876e+00 5.81663787e-01 -8.37055326e-01 -5.71577847e-01
6.80992723e-01 -3.08050841e-01 -1.58784196e-01 6.33694768e-01
-4.89015016e-04 9.72880602e-01 1.15923083e+00 -1.23844123e+00
-7.66776025e-01 8.20191279e-02 3.75421852e-01 3.29166532e-01
3.59381825e-01 7.56089166e-02 -8.85582030e-01 -5.66263080e-01
1.68791324e-01 -1.12062931e+00 -3.84791404e-01 3.38861197e-01
-2.87848651e-01 3.17821026e-01 7.84349263e-01 -5.35661340e-01
8.38211954e-01 -2.08199120e+00 1.89579397e-01 -6.43923283e-02
-1.11357905e-01 -1.83088541e-01 -1.08154856e-01 3.28352600e-01
-1.42785057e-01 1.08595595e-01 4.26114500e-02 -5.12278199e-01
-4.45877075e-01 2.26122573e-01 -7.50677586e-01 4.35304880e-01
1.93711922e-01 5.56236327e-01 -1.00099969e+00 -4.28161025e-01
6.40505612e-01 6.56747222e-01 -8.88407886e-01 3.56648117e-01
-6.46673858e-01 8.45903516e-01 -2.87585169e-01 4.07819003e-01
4.31283236e-01 -3.67564887e-01 2.34937459e-01 -9.43098366e-02
-2.31699303e-01 1.98463917e-01 -1.21389508e+00 1.71490002e+00
-3.78729433e-01 1.03201675e+00 -1.39354527e-01 -5.78638494e-01
5.34347773e-01 6.64120138e-01 6.59175634e-01 -2.18399197e-01
-1.42742947e-01 -1.95866421e-01 -3.76914233e-01 -5.76396823e-01
6.77793264e-01 1.07563056e-01 3.95337701e-01 5.17828763e-01
-1.92915171e-01 -3.30377460e-01 2.05097869e-01 2.41800740e-01
1.04690802e+00 6.19845390e-01 -9.02467221e-03 1.10665552e-01
2.57158399e-01 -4.37849686e-02 6.07891202e-01 8.17910135e-01
-1.02421656e-01 1.09919631e+00 5.80470622e-01 -5.96271455e-01
-1.48875475e+00 -1.23843896e+00 1.83754444e-01 5.30367076e-01
3.10526311e-01 -5.40585995e-01 -5.03472686e-01 -2.15704471e-01
-4.33971554e-01 8.17061841e-01 -3.04564774e-01 2.60702521e-01
-7.03449428e-01 1.47429675e-01 -1.10529281e-01 2.00102761e-01
4.85357530e-02 -1.06077385e+00 -9.83675778e-01 3.04391593e-01
-3.92203391e-01 -1.25053823e+00 -5.60948670e-01 -2.65297562e-01
-1.10561621e+00 -1.05002081e+00 -5.92630684e-01 -6.77519321e-01
8.50859284e-01 6.12021387e-01 1.26891232e+00 7.93479457e-02
-5.83403818e-02 6.14082098e-01 -1.72524020e-01 1.12648323e-01
-5.12662470e-01 -5.47841370e-01 4.66728657e-02 1.39271975e-01
-9.39009935e-02 -6.41885996e-01 -8.03299785e-01 2.48850107e-01
-8.78578603e-01 6.37633979e-01 2.04837248e-02 7.10497200e-01
9.56542552e-01 6.89847693e-02 -6.19581826e-02 -7.63430536e-01
-3.33504542e-03 -6.55821741e-01 -7.17467010e-01 -1.73931401e-02
6.65872693e-02 -1.23035692e-01 7.44723439e-01 -4.64213103e-01
-1.11177135e+00 4.11798358e-01 7.62193203e-02 -1.15404356e+00
-3.27128083e-01 1.88617408e-02 2.24743560e-01 4.68595922e-01
3.65689874e-01 5.04207671e-01 -1.37133047e-03 -2.71157265e-01
3.23242128e-01 1.83655038e-01 5.91052234e-01 -5.89313626e-01
5.21514237e-01 8.08791220e-01 -1.92366809e-01 -8.02507281e-01
-4.44852203e-01 -7.61257261e-02 -5.29531956e-01 -5.32681465e-01
6.87243938e-01 -1.29900491e+00 -5.20504713e-01 3.04825783e-01
-1.31250358e+00 -7.64378548e-01 -3.68271023e-01 5.00083208e-01
-9.82319534e-01 3.39172870e-01 -7.31081009e-01 -7.75948167e-01
1.70553520e-01 -1.40674746e+00 1.21282005e+00 9.95370373e-02
-4.02186751e-01 -8.48990142e-01 -2.86042452e-01 -3.47962640e-02
-4.79715317e-03 1.33680567e-01 4.36061591e-01 2.36157298e-01
-1.19262576e+00 -3.30467485e-02 2.16951758e-01 1.25297248e-01
1.53893992e-01 5.62489092e-01 -7.74722695e-01 -3.14949930e-01
8.42594132e-02 -3.60491276e-02 4.43262428e-01 7.00322270e-01
1.33386302e+00 -4.47200209e-01 -1.96813881e-01 6.13724351e-01
1.35825539e+00 3.70921165e-01 7.46955335e-01 -8.46855994e-03
4.30455059e-01 4.63822216e-01 5.35899758e-01 5.88218629e-01
3.58141929e-01 7.58859873e-01 2.83444136e-01 2.07484439e-01
-1.52761713e-01 -5.96835732e-01 6.75638020e-01 5.02946079e-01
-1.18660927e-01 -3.61533523e-01 -6.81184947e-01 7.00752676e-01
-1.76803112e+00 -1.50013566e+00 1.57314669e-02 2.27377462e+00
5.97994268e-01 5.12501262e-02 7.24793077e-02 -1.95188135e-01
7.18056560e-01 3.30184668e-01 -6.65013134e-01 -7.85627216e-02
-4.42775972e-02 -5.93748800e-02 4.29127753e-01 8.46455574e-01
-7.74997175e-01 8.26771319e-01 6.93572569e+00 4.23728943e-01
-1.09377646e+00 -3.13572347e-01 9.56731975e-01 -6.21941805e-01
-5.83342195e-01 4.60311234e-01 -6.53991401e-01 7.35065103e-01
1.07048774e+00 -2.19516933e-01 5.52855670e-01 6.64258182e-01
7.28309691e-01 -4.00351971e-01 -1.30270863e+00 9.86118257e-01
-1.86009035e-01 -1.75614560e+00 1.93744451e-01 -4.40205447e-02
8.40579152e-01 1.95889752e-02 2.21537054e-02 -1.44881949e-01
5.11429250e-01 -7.38313377e-01 1.20868254e+00 8.04078102e-01
7.60775805e-01 -4.67229307e-01 -2.58540045e-02 4.80687410e-01
-1.05215442e+00 -4.07058150e-02 -3.59969795e-01 -2.19231650e-01
6.94873095e-01 3.33190978e-01 -7.51949728e-01 1.40659958e-01
8.30398142e-01 8.55352759e-01 -1.74159229e-01 7.87343442e-01
-3.22691232e-01 7.02421069e-01 -3.93684030e-01 4.39027965e-01
3.07474043e-02 -3.31552982e-01 6.16041541e-01 8.85958910e-01
7.81882703e-01 3.27870697e-01 2.99012661e-01 7.51889884e-01
1.48839459e-01 -4.06962484e-01 -9.91550982e-01 3.44962180e-01
7.27150261e-01 7.34455466e-01 -5.25515676e-01 -6.43753171e-01
-6.69123113e-01 1.25966048e+00 -2.27442682e-02 7.56145597e-01
-9.66009915e-01 2.59489745e-01 7.61245728e-01 2.87274748e-01
4.64047045e-01 -5.62462509e-01 1.13447465e-01 -1.52950215e+00
3.89121436e-02 -8.10310006e-01 3.70785967e-02 -1.35681820e+00
-7.15460896e-01 3.49266350e-01 5.26685733e-03 -1.61014712e+00
-8.16779733e-01 -9.15616304e-02 -5.30060172e-01 6.58582985e-01
-1.24784160e+00 -6.37120247e-01 -2.01019272e-01 7.84401834e-01
1.17764473e+00 1.34727344e-01 6.29135013e-01 1.52738951e-02
-1.87365517e-01 -6.23830482e-02 2.33304482e-02 -1.18433572e-01
4.27476466e-01 -8.63708794e-01 6.54418886e-01 1.12063396e+00
3.34430426e-01 3.53421777e-01 8.21573675e-01 -5.58222592e-01
-1.49643803e+00 -9.66754258e-01 6.74281120e-01 -4.60260004e-01
4.15015519e-01 -1.49096087e-01 -6.33073330e-01 1.10563421e+00
1.40604898e-01 1.70494795e-01 2.99896896e-01 -4.90721226e-01
-4.35573608e-02 1.46619692e-01 -9.49811161e-01 1.03667152e+00
1.10659409e+00 -6.68914676e-01 -1.35819495e-01 3.00344229e-01
7.14674592e-01 -7.30097950e-01 -6.34456277e-01 -1.09046614e-02
6.56140685e-01 -1.59876323e+00 1.11363304e+00 -3.26000184e-01
7.90307581e-01 -5.80194652e-01 -3.90459597e-01 -9.79377687e-01
-1.84947133e-01 -9.57701027e-01 -4.25288856e-01 8.68725657e-01
1.05082966e-01 -2.82700330e-01 9.80200291e-01 9.71566498e-01
4.39142436e-02 -5.49050093e-01 -7.96444356e-01 -5.78410745e-01
-2.47434080e-01 -7.82762766e-01 4.85608965e-01 4.19904917e-01
-2.64337510e-01 3.33473608e-02 -8.15934837e-01 4.98333544e-01
6.03880584e-01 3.64269048e-01 1.06234229e+00 -5.14987230e-01
-5.24602413e-01 -5.89908250e-02 -2.72714674e-01 -1.72038817e+00
5.57166152e-02 -3.33131373e-01 1.36842662e-02 -1.18364918e+00
-1.47584872e-02 -2.89199650e-01 2.30404079e-01 1.95922226e-01
4.24685469e-03 1.49569929e-01 4.91652340e-01 5.92550099e-01
-5.23053348e-01 3.35444123e-01 1.18527126e+00 2.30716079e-01
-1.99139535e-01 -4.97544035e-02 -8.32950473e-02 8.38295698e-01
6.07251525e-01 -2.07662433e-01 -7.22337723e-01 -8.09470177e-01
-1.34293303e-01 9.19414282e-01 6.23094857e-01 -1.02190542e+00
1.21097080e-01 -4.35643584e-01 7.26005316e-01 -5.78194976e-01
8.15006733e-01 -9.01923776e-01 9.62441385e-01 2.37994999e-01
-3.77210259e-01 2.19799712e-01 -4.99939360e-03 6.89487994e-01
-1.21663563e-01 -1.88916698e-01 7.13846803e-01 -5.14802873e-01
-8.68378401e-01 5.09593904e-01 -6.88266873e-01 -3.44920039e-01
1.18197453e+00 -5.70428014e-01 1.05781503e-01 -7.91370749e-01
-9.01860535e-01 1.42547458e-01 1.15819633e+00 3.48583281e-01
8.34157825e-01 -1.22710311e+00 -5.36723673e-01 4.06298637e-01
-2.75301695e-01 2.58967757e-01 4.98956174e-01 3.68652195e-01
-7.60530472e-01 3.97959910e-02 -3.43934894e-02 -1.02230668e+00
-1.21268117e+00 6.19565308e-01 2.59599179e-01 1.17110543e-01
-9.64598775e-01 5.74190140e-01 3.04870695e-01 2.98440039e-01
8.49222485e-03 -3.09617281e-01 1.89347208e-01 -2.92987376e-01
7.44537652e-01 7.77807087e-02 -5.87560534e-01 -7.30397761e-01
9.50499102e-02 5.14937997e-01 6.84984550e-02 -4.02337551e-01
1.26232862e+00 -5.64996421e-01 2.33829528e-01 6.31159902e-01
1.21227658e+00 9.58531573e-02 -2.24905372e+00 -1.16970107e-01
-3.90989214e-01 -9.34179366e-01 -2.04068929e-01 -1.12429157e-01
-9.70610559e-01 6.52436137e-01 3.22093666e-01 1.30560711e-01
1.13453233e+00 -9.78465099e-03 7.92846143e-01 5.46266027e-02
6.67408347e-01 -6.68923378e-01 -2.46235151e-02 3.96407783e-01
6.54299557e-01 -1.06192148e+00 -6.01144731e-02 -2.65550345e-01
-5.83022118e-01 1.36788464e+00 3.03180814e-01 -3.22977334e-01
5.00795364e-01 2.74262875e-01 -4.12939861e-02 5.89799061e-02
-1.35773206e+00 1.27647191e-01 -1.38327658e-01 5.70063412e-01
3.81381869e-01 -1.40771925e-01 2.62006581e-01 -1.53011382e-01
-1.16925657e-01 2.77167916e-01 9.22815442e-01 8.28716218e-01
-2.73622394e-01 -9.64098752e-01 -4.76148218e-01 1.57765627e-01
-3.67104322e-01 -1.71168130e-02 3.17140460e-01 3.66532773e-01
-1.22590557e-01 9.02340949e-01 3.54662359e-01 -1.48392007e-01
-3.40331979e-02 -4.65554260e-02 5.89045525e-01 -5.71718037e-01
9.48497802e-02 3.45168263e-01 -2.38629784e-02 -9.24313366e-01
-6.45072699e-01 -1.10566366e+00 -1.07404172e+00 -3.30793440e-01
9.27160233e-02 -5.86626120e-02 2.37319723e-01 5.10107696e-01
6.09249353e-01 1.40167207e-01 9.26052034e-01 -1.51797438e+00
6.30618185e-02 -4.43751216e-01 -2.43389875e-01 5.04435360e-01
7.74793386e-01 -2.00349748e-01 -2.69806057e-01 9.48194921e-01] | [9.64401626586914, -2.0768940448760986] |
f43889f3-d587-4ada-a5f1-4c2914a1cb42 | small-object-detection-using-deep-learning | 2201.03243 | null | https://arxiv.org/abs/2201.03243v1 | https://arxiv.org/pdf/2201.03243v1.pdf | Small Object Detection using Deep Learning | Now a days, UAVs such as drones are greatly used for various purposes like that of capturing and target detection from ariel imagery etc. Easy access of these small ariel vehicles to public can cause serious security threats. For instance, critical places may be monitored by spies blended in public using drones. Study in hand proposes an improved and efficient Deep Learning based autonomous system which can detect and track very small drones with great precision. The proposed system consists of a custom deep learning model Tiny YOLOv3, one of the flavors of very fast object detection model You Look Only Once (YOLO) is built and used for detection. The object detection algorithm will efficiently the detect the drones. The proposed architecture has shown significantly better performance as compared to the previous YOLO version. The improvement is observed in the terms of resource usage and time complexity. The performance is measured using the metrics of recall and precision that are 93% and 91% respectively. | ['Asif Ullah Khan', 'Tauseef Jamal', 'Ayesha Salar', 'Aleena Ajaz'] | 2022-01-10 | null | null | null | null | ['real-time-object-detection', 'small-object-detection'] | ['computer-vision', 'computer-vision'] | [-5.28944433e-01 -2.40615785e-01 2.39101708e-01 5.59468865e-02
-9.88955945e-02 -6.46651983e-01 5.55376232e-01 -6.86163306e-02
-7.01215863e-01 5.93015373e-01 -6.03021920e-01 -1.23078665e-02
2.56746243e-05 -9.14479733e-01 -5.10106742e-01 -7.13449180e-01
-4.40509260e-01 2.50345260e-01 8.24956357e-01 -4.14322048e-01
-2.63165712e-01 7.59836912e-01 -1.76836693e+00 -2.34758556e-01
1.70421898e-01 1.32273471e+00 2.53440499e-01 1.06868160e+00
6.98630333e-01 6.35857463e-01 -4.58195001e-01 7.21972156e-03
8.31684709e-01 3.12102884e-01 -1.70398340e-01 -2.01452166e-01
5.88500381e-01 -1.03875589e+00 -2.38642886e-01 1.09237266e+00
3.53507727e-01 9.40718725e-02 4.24372435e-01 -1.41049993e+00
-4.45076711e-02 1.13043800e-01 -8.60704958e-01 6.56992614e-01
-2.54299212e-02 2.32427686e-01 4.13222790e-01 -5.98659873e-01
2.87726641e-01 1.14126873e+00 7.69351423e-01 2.04665631e-01
-3.50616664e-01 -1.14785552e+00 -1.83766454e-01 2.81077862e-01
-1.64534187e+00 -1.49086073e-01 7.15056062e-02 -4.45403665e-01
9.19703841e-01 4.81188595e-02 5.45710027e-01 3.45841706e-01
6.46603942e-01 6.03041291e-01 7.92192817e-01 -1.72476396e-01
-2.48316303e-01 4.89872873e-01 3.65396231e-01 1.18073142e+00
9.19692576e-01 5.04237950e-01 1.45899147e-01 2.09859207e-01
3.25805873e-01 2.84602284e-01 -5.35155572e-02 -2.30465740e-01
-9.98858571e-01 8.33684683e-01 6.12951636e-01 6.62192926e-02
-4.43646759e-01 2.71354139e-01 4.32083994e-01 3.45310777e-01
1.14302963e-01 2.39838079e-01 -4.65583324e-01 8.54375958e-02
-8.95726979e-01 4.24645841e-01 3.49599451e-01 1.10901368e+00
6.67389095e-01 5.41200280e-01 3.43802899e-01 1.15896344e-01
5.41717350e-01 1.00243354e+00 2.40365177e-01 -3.57346982e-01
-5.14322659e-03 6.68076336e-01 5.48665941e-01 -1.16866004e+00
-6.19244754e-01 -5.08359790e-01 -5.44555664e-01 7.67305195e-01
-6.06524525e-03 -7.75842428e-01 -1.07851589e+00 1.22219753e+00
5.88920653e-01 7.38524795e-02 2.19678119e-01 9.96318042e-01
1.27280402e+00 1.14151359e+00 1.24453805e-01 1.94880396e-01
1.62197185e+00 -6.85113549e-01 -5.99710047e-01 -1.23380654e-01
4.28171456e-01 -8.59658241e-01 3.74944329e-01 5.79424500e-01
-4.97860640e-01 -7.41574943e-01 -1.55306673e+00 2.66471773e-01
-7.51997113e-01 8.42587352e-01 5.48011899e-01 8.71107280e-01
-9.63771880e-01 -7.83498026e-03 -9.69593525e-01 -8.81213069e-01
3.94478828e-01 6.19910240e-01 -3.93442839e-01 1.64863154e-01
-8.89164209e-01 7.77214706e-01 8.11780930e-01 1.63031258e-02
-1.79597080e+00 -1.73006609e-01 -6.73105121e-01 8.39032978e-02
6.94857776e-01 -2.61329353e-01 9.51026618e-01 -6.60969555e-01
-8.96852911e-01 7.79165268e-01 4.97783095e-01 -8.75625074e-01
4.66288537e-01 -5.57320714e-01 -4.81898993e-01 -3.60714830e-02
-7.86738694e-02 8.12638402e-01 9.39970374e-01 -6.74151480e-01
-1.38156211e+00 -3.76859307e-01 5.51026881e-01 1.62533432e-01
-1.34440958e-01 1.92686930e-01 -5.21128438e-02 -2.11335331e-01
-5.04797459e-01 -1.12595868e+00 1.43122509e-01 1.85313419e-01
-2.17565268e-01 -3.11637461e-01 1.73091042e+00 -3.73266995e-01
8.58490646e-01 -1.93030787e+00 -4.33344513e-01 -1.84481680e-01
3.91802758e-01 9.62315023e-01 2.85544187e-01 3.89174640e-01
9.10313129e-02 -3.24530393e-01 3.16666156e-01 1.93748787e-01
-3.13624710e-01 -2.89623588e-01 -9.21851620e-02 6.11035049e-01
-1.31679133e-01 3.58842045e-01 -6.03374183e-01 -1.79344699e-01
3.91871452e-01 4.24128830e-01 -1.94752693e-01 3.49996328e-01
7.37362541e-03 -1.33534744e-01 -4.06323165e-01 8.60029161e-01
7.99629927e-01 2.89008558e-01 -4.95589674e-01 -3.35014611e-01
-5.74295580e-01 -3.62225354e-01 -1.13164079e+00 7.70148516e-01
-2.39119619e-01 1.14130497e+00 2.54526973e-01 -5.18349826e-01
8.73187006e-01 3.20280075e-01 -1.02959640e-01 -1.75940245e-02
7.03542471e-01 -1.68295145e-01 3.89352217e-02 -5.23633063e-01
5.44640303e-01 2.46322125e-01 7.44180754e-02 1.82255358e-01
4.45557237e-02 1.62303612e-01 2.53933936e-01 1.89896837e-01
9.69270349e-01 -9.47822779e-02 6.55969083e-01 -3.74450266e-01
5.75755596e-01 5.72405875e-01 2.32476965e-01 7.04521120e-01
-3.66186947e-01 -2.74354488e-01 1.47248417e-01 -7.94687271e-01
-8.99682641e-01 -6.61147833e-01 7.86611289e-02 8.97962928e-01
2.53755480e-01 -2.40720704e-01 -7.45628417e-01 -5.49703479e-01
-1.39857501e-01 2.89781898e-01 -4.72315311e-01 -1.27245978e-01
-9.19464827e-02 -7.68258274e-01 5.99307477e-01 3.27637345e-01
1.26678503e+00 -7.94170499e-01 -1.45636868e+00 -8.98562223e-02
7.06801593e-01 -1.21635664e+00 2.53319740e-01 -1.53816804e-01
-5.63345313e-01 -1.24121797e+00 -5.02260745e-01 -7.16551542e-01
3.26366574e-01 8.25512528e-01 4.77783293e-01 2.16346160e-02
-6.36188745e-01 3.99138868e-01 -4.63949025e-01 -1.20336974e+00
1.28764659e-01 -5.13571240e-02 4.07274544e-01 4.00348799e-03
7.00414896e-01 -2.21892800e-02 -5.38165450e-01 -2.63911858e-02
-6.67270839e-01 -4.80720758e-01 8.23935330e-01 3.52224410e-01
1.47240996e-01 5.44316113e-01 6.98456466e-02 -7.16994047e-01
4.64881867e-01 -5.48834443e-01 -1.51672900e+00 -4.32244875e-02
-3.72449279e-01 -4.34323817e-01 5.57768524e-01 -6.63899034e-02
-5.04506171e-01 1.90738052e-01 2.46196240e-01 -4.41949844e-01
-4.56116199e-01 6.80301636e-02 1.05291903e-01 -5.22741854e-01
4.19321060e-01 -1.65170245e-02 -1.33692801e-01 -2.75819153e-01
-1.03530608e-01 7.76575863e-01 1.54818043e-01 3.49255592e-01
1.12540066e+00 5.38215816e-01 3.24717402e-01 -1.27579284e+00
-5.16567707e-01 -5.56618452e-01 -4.20292228e-01 -4.72803950e-01
1.05238557e+00 -1.33121562e+00 -9.99350548e-01 6.41930044e-01
-9.57547009e-01 1.99833095e-01 4.51772839e-01 6.15462422e-01
1.21423386e-01 7.24947527e-02 -2.07507595e-01 -9.66612697e-01
-5.97204506e-01 -1.14555967e+00 9.16675985e-01 5.98944247e-01
4.48548138e-01 -6.41806245e-01 -2.23745834e-02 -8.53312016e-02
4.32087421e-01 4.54382181e-01 2.71131128e-01 -5.80185175e-01
-1.02212477e+00 -1.01201475e+00 -5.40078402e-01 3.36270362e-01
-5.60644595e-03 1.38216496e-01 -8.09684634e-01 -6.91081762e-01
-3.27060223e-01 -3.22298765e-01 8.05829585e-01 4.38418001e-01
4.50068772e-01 -5.95670752e-02 -7.64852226e-01 6.61433578e-01
1.95031989e+00 7.24303246e-01 2.57379144e-01 3.50233495e-01
5.94348550e-01 7.18553066e-02 1.12579882e+00 4.61144865e-01
1.37185499e-01 4.51646000e-01 9.11340952e-01 3.28493118e-03
4.81209792e-02 5.08291796e-02 4.57194716e-01 -6.61391672e-03
-3.88108015e-01 -5.16862869e-01 -9.80673611e-01 5.73062420e-01
-1.57980728e+00 -9.01711583e-01 -1.10227220e-01 2.12247419e+00
-2.91863978e-01 2.50612944e-01 2.57104725e-01 -6.32833019e-02
5.02387166e-01 1.11111879e-01 -1.35464922e-01 -3.36834669e-01
5.23919821e-01 -4.52367738e-02 1.14815664e+00 3.93825263e-01
-1.79613066e+00 1.22133052e+00 5.43571043e+00 6.76638067e-01
-1.38781989e+00 2.90542126e-01 1.54813066e-01 -1.18170463e-01
9.77040291e-01 -2.22918272e-01 -1.46213377e+00 2.27490157e-01
8.34275603e-01 1.13067709e-01 9.12212878e-02 1.26669240e+00
2.77068555e-01 -3.89613956e-01 -4.72383320e-01 9.55109477e-01
1.39633238e-01 -1.03743899e+00 1.55730307e-01 7.33852535e-02
4.01236743e-01 1.96540013e-01 7.14345649e-02 3.03044826e-01
4.02859718e-01 -8.42397511e-01 4.44461048e-01 4.13180828e-01
5.84179282e-01 -1.27311850e+00 1.38594878e+00 4.35683429e-01
-1.44754112e+00 -2.68053770e-01 -9.00010169e-01 -1.31572917e-01
-1.09102353e-01 -8.54113251e-02 -1.21274352e+00 1.52883291e-01
1.13943589e+00 5.18334746e-01 -4.88570303e-01 1.33704245e+00
-5.27847819e-02 5.10129869e-01 -6.35970533e-01 -4.28186268e-01
9.19869006e-01 -6.39085844e-02 9.58456099e-01 9.29747701e-01
5.80018997e-01 2.67632812e-01 3.69906068e-01 4.36905384e-01
1.26685217e-01 -1.00704409e-01 -1.17163861e+00 9.63861682e-03
3.03525597e-01 1.61301494e+00 -6.97116792e-01 -4.02772367e-01
-3.44486207e-01 6.33378625e-01 -2.04855576e-01 -1.19135365e-01
-1.09972084e+00 -8.59782040e-01 5.51071167e-01 4.63410676e-01
4.79281396e-01 -3.86190474e-01 7.25407660e-01 -6.01749599e-01
-4.80349600e-01 -6.24151528e-01 1.50611430e-01 -8.50557923e-01
-3.38544160e-01 1.05137169e+00 1.57093301e-01 -1.45327270e+00
3.13777477e-02 -1.10728061e+00 -5.06438017e-01 2.59630710e-01
-1.34538913e+00 -1.57299387e+00 -7.86743104e-01 4.96842921e-01
7.72093177e-01 -8.75562787e-01 4.73248988e-01 3.46714497e-01
-8.51764441e-01 4.70477521e-01 -1.28583640e-01 3.15884411e-01
5.53645194e-01 -9.91090775e-01 -1.54672265e-01 1.32700646e+00
5.36821112e-02 4.47437078e-01 7.81172752e-01 -6.37147248e-01
-1.39941370e+00 -1.23686802e+00 2.11888373e-01 -4.17137653e-01
4.42822307e-01 -2.10903987e-01 -5.32352850e-02 1.07921731e+00
5.21635950e-01 -5.07408008e-02 2.67040521e-01 -5.36968887e-01
3.17305624e-01 -6.34146273e-01 -1.24838066e+00 1.82095885e-01
1.11141495e-01 1.87914684e-01 -3.55511874e-01 4.62675601e-01
3.89913559e-01 -3.32464665e-01 -3.19634289e-01 3.69176686e-01
6.12008095e-01 -1.11430216e+00 7.85411060e-01 -5.03743052e-01
-2.87530392e-01 -7.99708128e-01 -1.73174888e-01 -9.45222378e-01
-3.19025040e-01 -5.16317427e-01 1.90740123e-01 9.00536776e-01
2.92136610e-01 -4.36295539e-01 7.44672596e-01 6.07850403e-02
-8.86727273e-02 -1.93585396e-01 -5.82059085e-01 -6.57348871e-01
-7.66011298e-01 -2.60772463e-02 3.57744545e-01 5.05605698e-01
-6.86232567e-01 3.17849785e-01 -7.13050902e-01 9.00527775e-01
7.36364186e-01 -2.63675958e-01 1.05699718e+00 -1.38801241e+00
1.47668377e-01 3.89501095e-01 -1.09061503e+00 -7.94388831e-01
-5.67162991e-01 -2.48401746e-01 -1.73490748e-01 -1.15804648e+00
-3.95646803e-02 -2.75658995e-01 -1.36759609e-01 4.54742521e-01
1.94717482e-01 6.36911631e-01 1.01358637e-01 3.55634699e-03
-7.88720250e-01 6.95453286e-02 5.62462091e-01 9.28378478e-03
1.34532340e-02 2.86716729e-01 -6.01838529e-02 8.90304506e-01
7.52724886e-01 -4.40383762e-01 -4.96636450e-01 -2.86464900e-01
1.31017834e-01 -2.16611400e-01 4.57623988e-01 -1.81245995e+00
2.56331235e-01 3.49113882e-01 5.10340750e-01 -9.76972401e-01
5.72792649e-01 -1.22358692e+00 1.47534221e-01 7.90376246e-01
3.96093279e-01 2.75191486e-01 6.33998513e-01 6.01782918e-01
-1.81806669e-01 -2.61705399e-01 1.10135126e+00 -4.51406866e-01
-1.39802611e+00 4.06967521e-01 -8.10470521e-01 -7.44719446e-01
1.81866109e+00 -1.98498979e-01 -2.09565774e-01 -4.16610003e-01
-4.13307458e-01 2.16219738e-01 7.92289376e-02 3.10808510e-01
5.75616777e-01 -8.83753002e-01 -5.84428430e-01 1.72000363e-01
1.89553484e-01 -2.05569342e-01 -5.75234331e-02 7.29193985e-01
-1.37277174e+00 1.07199264e+00 -7.18907356e-01 -3.64422888e-01
-1.78224349e+00 5.73847055e-01 4.74471271e-01 3.04967985e-02
-3.39585960e-01 1.12680364e+00 1.09417304e-01 1.02925099e-01
1.94950283e-01 -3.47210765e-01 -6.69530153e-01 6.45472556e-02
8.33079040e-01 6.26968145e-01 -1.65730581e-01 -6.65313303e-01
-4.39188749e-01 7.45972753e-01 -3.84726733e-01 4.96733874e-01
1.20843971e+00 -1.57650366e-01 -1.96757540e-02 3.01483413e-03
5.06424904e-01 8.10028166e-02 -1.14997506e+00 2.57988691e-01
-3.66812319e-01 -3.14017951e-01 6.52708471e-01 -7.56334901e-01
-1.06882203e+00 8.94495428e-01 1.42902625e+00 2.95573413e-01
8.83536160e-01 -3.41085196e-01 9.04621005e-01 8.30412388e-01
6.69214308e-01 -5.92905462e-01 -3.87578875e-01 3.64890069e-01
4.30976123e-01 -1.68744838e+00 2.60094315e-01 -1.15118049e-01
-4.77805853e-01 1.13074839e+00 8.68576288e-01 -7.31605172e-01
7.73338377e-01 4.19950515e-01 1.52843133e-01 -5.95146775e-01
-6.90822363e-01 -5.93552113e-01 1.28317550e-01 8.58163536e-01
1.73971653e-02 2.30628729e-01 -8.78370404e-02 1.44790187e-01
-1.18619286e-01 -1.02571227e-01 6.10035121e-01 8.40086639e-01
-1.14503181e+00 -3.95209521e-01 -5.07350504e-01 4.20169860e-01
-6.14302516e-01 2.35408191e-02 -3.47452104e-01 1.42465544e+00
5.19168854e-01 9.17589724e-01 -3.23588550e-02 -5.91759562e-01
2.10814606e-02 -4.95838344e-01 5.10251448e-02 -5.45543730e-01
-2.79290825e-01 -2.36403316e-01 8.96007270e-02 -4.11558062e-01
-2.79660046e-01 -1.35972574e-01 -8.24991703e-01 -3.56693804e-01
-4.86820370e-01 9.18942839e-02 7.07820654e-01 8.56424809e-01
9.39487740e-02 2.50548720e-01 2.97015995e-01 -6.94380224e-01
-1.41715243e-01 -9.95926738e-01 -6.31563723e-01 -5.43457806e-01
7.64408469e-01 -1.04325902e+00 -1.25104934e-01 -1.24127947e-01] | [8.519155502319336, -0.9581327438354492] |
bf563f36-0993-4e5f-840f-cd45351cffe9 | candidate-set-re-ranking-for-composed-image | 2305.16304 | null | https://arxiv.org/abs/2305.16304v2 | https://arxiv.org/pdf/2305.16304v2.pdf | Candidate Set Re-ranking for Composed Image Retrieval with Dual Multi-modal Encoder | Composed image retrieval aims to find an image that best matches a given multi-modal user query consisting of a reference image and text pair. Existing methods commonly pre-compute image embeddings over the entire corpus and compare these to a reference image embedding modified by the query text at test time. Such a pipeline is very efficient at test time since fast vector distances can be used to evaluate candidates, but modifying the reference image embedding guided only by a short textual description can be difficult, especially independent of potential candidates. An alternative approach is to allow interactions between the query and every possible candidate, i.e., reference-text-candidate triplets, and pick the best from the entire set. Though this approach is more discriminative, for large-scale datasets the computational cost is prohibitive since pre-computation of candidate embeddings is no longer possible. We propose to combine the merits of both schemes using a two-stage model. Our first stage adopts the conventional vector distancing metric and performs a fast pruning among candidates. Meanwhile, our second stage employs a dual-encoder architecture, which effectively attends to the input triplet of reference-text-candidate and re-ranks the candidates. Both stages utilize a vision-and-language pre-trained network, which has proven beneficial for various downstream tasks. Our method consistently outperforms state-of-the-art approaches on standard benchmarks for the task. | ['Stephen Gould', 'Damien Teney', 'Weixuan Sun', 'Zheyuan Liu'] | 2023-05-25 | null | null | null | null | ['composed-image-retrieval'] | ['computer-vision'] | [ 3.63245308e-01 -5.01562178e-01 -3.38782400e-01 -2.92547673e-01
-1.45041800e+00 -6.41466916e-01 8.16393852e-01 3.80847335e-01
-9.82406437e-01 2.40825027e-01 4.86706384e-02 -1.04166552e-01
-5.92462085e-02 -6.26419365e-01 -6.22367144e-01 -5.66609204e-01
3.79276276e-01 6.48237467e-01 6.46032214e-01 -3.28437760e-02
5.00942647e-01 3.28429878e-01 -1.84332263e+00 3.39617640e-01
5.48282862e-01 1.19643080e+00 6.42943442e-01 5.14575422e-01
-1.01352461e-01 2.29476377e-01 -3.06690961e-01 -6.68207586e-01
4.79664713e-01 -2.05592960e-01 -7.58293152e-01 1.74563468e-01
8.94202530e-01 -4.23392624e-01 -4.12787706e-01 1.05420804e+00
5.74449360e-01 3.92119259e-01 5.04357994e-01 -9.11432683e-01
-8.33369493e-01 1.53871670e-01 -6.54093266e-01 1.40670523e-01
3.72217923e-01 4.38087769e-02 1.44833899e+00 -1.41403234e+00
7.76916087e-01 1.05020952e+00 1.21545315e-01 4.38376039e-01
-1.29556417e+00 -5.24324358e-01 2.60269530e-02 3.44936758e-01
-1.66217554e+00 -4.88867313e-01 6.15647972e-01 -1.86229035e-01
8.25169981e-01 3.27688545e-01 4.79326695e-01 7.33394146e-01
-6.53739274e-02 1.01234102e+00 7.27162838e-01 -4.48221207e-01
-2.00486649e-02 2.20151156e-01 1.37896359e-01 7.85100579e-01
-7.37632811e-03 -4.12569456e-02 -4.71270561e-01 -2.01973617e-01
2.32068717e-01 2.98838019e-01 -4.86537993e-01 -6.56466186e-01
-1.38858926e+00 8.85253012e-01 4.49518204e-01 4.41085488e-01
-4.48679358e-01 2.29893699e-02 3.87647897e-01 3.98542315e-01
1.19256042e-01 5.33388615e-01 -2.33757988e-01 2.42245093e-01
-1.51731455e+00 2.73923695e-01 3.40582728e-01 6.84004486e-01
1.01221931e+00 -5.74958563e-01 -4.12572980e-01 1.02315617e+00
4.25757915e-01 3.62681299e-01 5.99303305e-01 -6.35410666e-01
5.00777721e-01 6.60552680e-01 -2.24663457e-03 -9.93861258e-01
1.55532002e-01 -2.00248629e-01 -5.19519329e-01 -3.33101652e-03
2.31334701e-01 4.56140608e-01 -1.00382423e+00 1.39070809e+00
3.81246835e-01 1.93795823e-02 -3.06519046e-02 1.11649740e+00
6.90118551e-01 6.63526058e-01 -1.92264020e-01 -1.40596285e-01
1.43543804e+00 -1.31185710e+00 -1.20024092e-01 -3.29130381e-01
5.03744841e-01 -1.10017741e+00 1.20665443e+00 8.66088569e-02
-1.02418554e+00 -5.50472558e-01 -1.05865622e+00 -3.22019488e-01
-3.89624029e-01 3.55401903e-01 6.04922101e-02 2.96568453e-01
-1.26711917e+00 3.66458207e-01 -5.84122956e-01 -4.08701062e-01
1.04591921e-01 4.91875172e-01 -5.11768103e-01 -6.31446183e-01
-9.73708153e-01 8.88405561e-01 4.51673508e-01 -3.65831219e-02
-7.85279930e-01 -3.64953965e-01 -9.32319939e-01 7.85555169e-02
4.22163427e-01 -6.54895842e-01 9.98813212e-01 -1.13651288e+00
-1.18253136e+00 1.02348208e+00 -4.96227503e-01 -2.76844025e-01
4.08485562e-01 -4.15121950e-02 -3.51034075e-01 5.25975585e-01
3.19729954e-01 9.90796685e-01 1.10511756e+00 -1.13002753e+00
-7.48914301e-01 -2.45581225e-01 2.34099895e-01 4.03450727e-01
-5.83283365e-01 1.70878202e-01 -1.42414010e+00 -4.76657152e-01
2.49215737e-01 -1.05823016e+00 -2.76511997e-01 2.95145839e-01
-3.01792026e-01 -4.40391243e-01 7.21153378e-01 -2.41809055e-01
1.35370398e+00 -2.13496494e+00 4.15300718e-03 2.89266497e-01
1.17385812e-01 3.57047349e-01 -5.66781402e-01 4.95674312e-01
4.26128916e-02 -1.06144501e-02 5.37040792e-02 -6.09202564e-01
-2.26429999e-02 -6.79695159e-02 -1.94735870e-01 4.41074669e-01
2.94814616e-01 9.72603142e-01 -8.86853278e-01 -8.98293078e-01
4.25099641e-01 3.91455978e-01 -5.56080341e-01 9.56477672e-02
-1.42954111e-01 -1.25624478e-01 -4.40241933e-01 6.56161368e-01
4.49423522e-01 -4.86506045e-01 -3.07377800e-02 -4.91314709e-01
-1.19223863e-01 1.88375488e-01 -1.13572872e+00 1.78153086e+00
-4.33689773e-01 5.51597893e-01 -2.56535292e-01 -1.03989279e+00
7.42611110e-01 1.27545223e-01 4.85431999e-01 -8.45600486e-01
-8.57572556e-02 4.87933546e-01 -1.93533793e-01 -4.28187579e-01
7.81133115e-01 1.48971722e-01 9.31311101e-02 5.39496243e-01
-2.44904328e-02 1.06128961e-01 5.52328944e-01 2.97565371e-01
1.10385025e+00 9.58137959e-02 1.42088935e-01 6.66168407e-02
7.74046004e-01 6.19523376e-02 2.26807371e-01 8.38514507e-01
-3.68841141e-02 8.36897194e-01 1.23027183e-01 -4.95523781e-01
-8.45132768e-01 -1.01797175e+00 -6.41450137e-02 1.17730093e+00
5.45116127e-01 -5.00476539e-01 -1.92958295e-01 -8.23135078e-01
-1.39472932e-01 4.75860715e-01 -5.76119006e-01 -1.97418034e-02
-5.63660681e-01 -4.23573583e-01 1.94385752e-01 1.86539546e-01
2.83202112e-01 -8.52080107e-01 -6.74950480e-01 1.12966791e-01
-1.85027882e-01 -9.98281538e-01 -9.20240939e-01 -2.56141368e-02
-6.08030021e-01 -1.03031182e+00 -8.57932389e-01 -1.12265229e+00
8.21067154e-01 7.25979567e-01 1.08016050e+00 3.86843652e-01
-1.90101996e-01 4.53774571e-01 -3.74729574e-01 2.87697732e-01
-1.00959271e-01 -1.39754608e-01 -2.82591671e-01 3.08143079e-01
4.82349098e-01 -1.75735261e-02 -9.99468744e-01 4.11073536e-01
-1.12939274e+00 -1.17065109e-01 8.83146465e-01 1.09051371e+00
1.04217923e+00 -2.09049210e-01 1.53219461e-01 -3.62135112e-01
5.10944664e-01 -2.18116462e-01 -6.07199788e-01 6.99392378e-01
-7.09519148e-01 2.56521791e-01 4.59751427e-01 -6.04376853e-01
-4.15019095e-01 4.52647358e-01 -1.98893007e-02 -8.92137289e-01
1.56163469e-01 6.59104347e-01 1.10621363e-01 -7.71982372e-02
3.69643152e-01 5.48323452e-01 -1.83850661e-01 -2.26930663e-01
4.81140137e-01 6.28070831e-01 5.15427172e-01 -4.35776770e-01
8.49957228e-01 3.79758030e-01 -2.84733474e-01 -7.46853411e-01
-6.50285065e-01 -9.90935624e-01 -6.56358600e-01 -1.30543247e-01
8.13322365e-01 -8.15866888e-01 -3.56963724e-01 -9.46424305e-02
-1.23363209e+00 1.74049348e-01 3.48771922e-02 6.29925191e-01
-1.37935594e-01 5.88034272e-01 -2.97161669e-01 -5.19638658e-01
-3.94090295e-01 -1.57877481e+00 1.47359872e+00 -2.38452423e-02
1.35368714e-02 -6.97736919e-01 -5.03634289e-02 3.04330379e-01
2.50957251e-01 -1.81663528e-01 8.22890162e-01 -8.50716889e-01
-9.38780189e-01 -6.37124062e-01 -5.07008791e-01 1.55630007e-01
1.17668957e-02 2.41996832e-02 -6.78537250e-01 -5.53401113e-01
-1.82366133e-01 -6.10015392e-01 1.15936613e+00 6.93471804e-02
1.12162209e+00 -1.00753166e-01 -4.89717782e-01 2.47310206e-01
1.70134592e+00 -5.72410077e-02 5.53964198e-01 4.92816627e-01
5.49827576e-01 4.51399893e-01 7.66410828e-01 9.12901610e-02
6.01911664e-01 7.54691660e-01 2.29027122e-01 -1.05232172e-01
1.18350470e-02 -9.67098400e-02 3.48455459e-01 7.50928998e-01
4.23175663e-01 -3.67761374e-01 -8.10326993e-01 7.38862097e-01
-1.71141195e+00 -9.96343851e-01 2.06796944e-01 2.47654748e+00
7.93318450e-01 1.23864457e-01 -1.09858707e-01 1.91051047e-02
6.84603453e-01 2.58681655e-01 -5.03538847e-01 -5.46737760e-02
-2.42553465e-02 3.76880795e-01 3.66977394e-01 3.84882361e-01
-1.22374296e+00 8.67580354e-01 5.37762165e+00 9.53752339e-01
-1.29542208e+00 5.69480248e-02 4.75846171e-01 -4.14103985e-01
-4.11721379e-01 9.61060748e-02 -7.67677248e-01 3.03962559e-01
6.43088222e-01 -1.56089678e-01 3.39273423e-01 7.05805957e-01
-2.85289325e-02 -3.00706089e-01 -1.36055243e+00 1.28111792e+00
5.84757686e-01 -1.38815904e+00 3.03075880e-01 -7.02734068e-02
5.64607024e-01 3.40500861e-01 6.81676194e-02 3.21763784e-01
-1.83426768e-01 -6.93832099e-01 6.53906524e-01 2.16376096e-01
8.83700490e-01 -5.87466121e-01 5.68871975e-01 1.96532890e-01
-1.40438461e+00 6.38772035e-03 -3.58706415e-01 6.44996524e-01
6.86634239e-03 2.80271739e-01 -5.63849688e-01 4.93032962e-01
5.98264754e-01 5.04338562e-01 -9.05874848e-01 1.18504536e+00
9.96642336e-02 7.93789476e-02 -4.22917157e-01 -8.94501060e-02
5.91461480e-01 -6.79282621e-02 3.79816979e-01 1.10178900e+00
4.50872600e-01 -2.67643809e-01 5.38491189e-01 5.66035569e-01
-1.42906144e-01 4.73285705e-01 -5.18316507e-01 -2.65144631e-02
4.56301242e-01 1.47384524e+00 -7.68286645e-01 -4.70820457e-01
-8.41329157e-01 1.41700172e+00 4.41276073e-01 2.65232712e-01
-6.53540492e-01 -5.75675428e-01 3.09918731e-01 -9.84299853e-02
6.76801562e-01 -1.08942397e-01 2.26198763e-01 -1.21996331e+00
4.00844693e-01 -8.56644094e-01 5.76484144e-01 -6.44578636e-01
-1.11313367e+00 7.92982280e-01 -1.98798269e-01 -1.50242043e+00
-2.67616361e-01 -4.10304904e-01 -4.24807727e-01 7.28838325e-01
-1.75336027e+00 -1.32219827e+00 -1.08110651e-01 7.86236346e-01
8.12281251e-01 9.42910686e-02 7.36195266e-01 5.79490960e-01
-4.30120915e-01 8.11538875e-01 2.20813021e-01 1.41142800e-01
8.29121470e-01 -1.01542819e+00 2.04475641e-01 1.02414000e+00
6.25872552e-01 7.54467666e-01 2.51749694e-01 -3.84048074e-01
-1.60882437e+00 -1.15348411e+00 1.11867201e+00 -3.18345606e-01
7.78832972e-01 -2.50909030e-01 -9.95265722e-01 2.32027620e-01
3.40375900e-01 2.65294820e-01 5.23752809e-01 -2.70211190e-01
-5.29513359e-01 -1.43583745e-01 -6.15349889e-01 8.66214037e-01
6.16166830e-01 -8.11105669e-01 -6.11088157e-01 3.28627765e-01
5.90018272e-01 -2.10295737e-01 -5.58447957e-01 2.23989427e-01
7.38683283e-01 -8.64236832e-01 1.00747800e+00 -1.53226942e-01
3.99975926e-01 -5.67828417e-01 -3.54244590e-01 -9.29018319e-01
-1.20177671e-01 -4.41481501e-01 6.78283051e-02 1.05895007e+00
5.90224922e-01 -3.11856270e-01 6.52401149e-01 6.41867876e-01
7.89994095e-03 -1.03689253e+00 -9.15452182e-01 -6.96014941e-01
-3.09639156e-01 -3.49149853e-01 2.85117924e-01 6.43783391e-01
-2.55116045e-01 5.19168019e-01 -1.74879566e-01 1.61808386e-01
4.79622573e-01 5.82938373e-01 6.95210159e-01 -9.93928015e-01
-1.80431247e-01 -5.42653441e-01 -3.56357843e-01 -1.40336716e+00
2.49441624e-01 -1.07899511e+00 2.40721613e-01 -1.58795094e+00
3.25805783e-01 -4.13253665e-01 -5.13174772e-01 3.50709885e-01
-3.83658350e-01 5.11766195e-01 2.05950573e-01 5.24393260e-01
-9.81461406e-01 5.14275610e-01 1.10492158e+00 -4.62733954e-01
-2.37924844e-01 -2.31523678e-01 -4.31541175e-01 4.45929289e-01
4.50455785e-01 -5.47266245e-01 -4.59995210e-01 -7.56528795e-01
-5.52318394e-02 9.84265190e-03 3.49552184e-01 -8.67371857e-01
6.12482369e-01 -9.95524973e-02 3.08512837e-01 -8.50605190e-01
5.03226340e-01 -8.60741913e-01 -2.47485057e-01 2.65738368e-01
-4.64683056e-01 4.42938805e-01 -1.61053002e-01 6.59224212e-01
-4.76665467e-01 -4.66508567e-01 7.66683102e-01 -1.38514703e-02
-8.35329175e-01 6.09145939e-01 -1.49810407e-02 -3.68077047e-02
9.84620035e-01 -4.03007030e-01 -1.37732014e-01 -1.24329425e-01
-2.13219479e-01 3.47242981e-01 6.55560732e-01 5.31899273e-01
1.01583362e+00 -1.42861950e+00 -5.53102136e-01 1.20460786e-01
5.47178745e-01 -2.11936787e-01 -1.40518561e-01 8.14996183e-01
-2.94366419e-01 2.77764708e-01 2.06136152e-01 -7.98098147e-01
-1.49525928e+00 7.73863137e-01 1.96236923e-01 -3.67285758e-01
-5.13138652e-01 8.30282390e-01 2.24543169e-01 -6.42734617e-02
2.51555979e-01 -9.86638889e-02 -1.53237239e-01 2.31597438e-01
6.77865982e-01 -1.91981003e-01 7.66912475e-02 -9.66341674e-01
-4.12988842e-01 7.89563775e-01 -3.90344352e-01 -2.90882081e-01
1.08679998e+00 -3.25753838e-01 -1.71127766e-01 2.30815291e-01
1.71166003e+00 -1.19005196e-01 -7.66501904e-01 -6.80569649e-01
3.54319930e-01 -7.39806592e-01 1.97365671e-01 -3.53558123e-01
-1.12914991e+00 8.72235000e-01 7.17032492e-01 -4.07394134e-02
1.31028068e+00 1.28605708e-01 9.45377290e-01 7.60636687e-01
1.48793921e-01 -9.78207052e-01 4.95476842e-01 3.23179752e-01
8.81874621e-01 -1.47959173e+00 1.59470867e-02 1.49356539e-03
-5.72520137e-01 1.23881817e+00 6.35803580e-01 -9.66746807e-02
4.53942150e-01 -3.71016711e-01 -1.83598977e-02 -2.14550465e-01
-9.16835904e-01 -5.60855031e-01 7.68236041e-01 9.73660424e-02
2.70933181e-01 -3.18901330e-01 -2.70529449e-01 -7.79779330e-02
2.41334602e-01 -2.63443649e-01 9.01140422e-02 8.42179716e-01
-3.54004860e-01 -1.35536385e+00 -2.50826299e-01 7.57208407e-01
-4.75057721e-01 -3.75478923e-01 -1.48853928e-01 5.30430675e-01
4.08391766e-02 9.15457249e-01 1.66768119e-01 -3.76109242e-01
3.50218683e-01 -5.31626418e-02 3.87489557e-01 -6.76184058e-01
-5.24952114e-01 4.27770525e-01 -2.01785132e-01 -7.08163142e-01
-5.80123782e-01 -6.63458049e-01 -9.76135015e-01 1.32137731e-01
-6.99972332e-01 6.40135705e-02 5.94836056e-01 8.88423681e-01
3.94739240e-01 -9.96801555e-02 7.93278873e-01 -9.51640785e-01
-4.43074644e-01 -4.94354695e-01 -4.56683077e-02 5.54268301e-01
3.90801847e-01 -6.13325477e-01 -1.58989638e-01 4.06243205e-02] | [10.678824424743652, 1.0505117177963257] |
6ccfbaf8-0a8d-474d-823c-2103a27551be | from-semantic-categories-to-fixations-a-novel | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Wang_From_Semantic_Categories_to_Fixations_A_Novel_Weakly-Supervised_Visual-Auditory_Saliency_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Wang_From_Semantic_Categories_to_Fixations_A_Novel_Weakly-Supervised_Visual-Auditory_Saliency_CVPR_2021_paper.pdf | From Semantic Categories to Fixations: A Novel Weakly-Supervised Visual-Auditory Saliency Detection Approach | Thanks to the rapid advances in the deep learning techniques and the wide availability of large-scale training sets, the performances of video saliency detection models have been improving steadily and significantly. However, the deep learning based visual-audio fixation prediction is still in its infancy. At present, only a few visual-audio sequences have been furnished with real fixations being recorded in the real visual-audio environment. Hence, it would be neither efficiency nor necessary to re-collect real fixations under the same visual-audio circumstance. To address the problem, this paper advocate a novel approach in a weakly-supervised manner to alleviating the demand of large-scale training sets for visual-audio model training. By using the video category tags only, we propose the selective class activation mapping (SCAM), which follows a coarse-to-fine strategy to select the most discriminative regions in the spatial-temporal-audio circumstance. Moreover, these regions exhibit high consistency with the real human-eye fixations, which could subsequently be employed as the pseudo GTs to train a new spatial-temporal-audio (STA) network. Without resorting to any real fixation, the performance of our STA network is comparable to that of the fully supervised ones. | ['Hong Qin', 'Aimin Hao', 'Deng-Ping Fan', 'Chenglizhao Chen', 'Guotao Wang'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['video-saliency-detection'] | ['computer-vision'] | [ 2.59192914e-01 -1.96239829e-01 -1.68284968e-01 -7.58529454e-02
-5.82692921e-01 -1.24668978e-01 3.64262670e-01 4.25746366e-02
-5.27110219e-01 6.97698474e-01 -1.20338365e-01 5.39474823e-02
-3.92058864e-02 -4.53851640e-01 -6.65337026e-01 -7.91778684e-01
2.35291049e-01 1.91406757e-02 7.33591139e-01 -1.18300386e-01
5.07501066e-01 4.06644270e-02 -2.30399966e+00 3.56109679e-01
7.29309857e-01 1.13895524e+00 7.79946268e-01 2.23109171e-01
9.39982664e-03 6.40343904e-01 -4.87748533e-01 -1.14870161e-01
8.42743833e-03 -5.80353856e-01 -5.10207057e-01 7.28416592e-02
3.10717165e-01 -1.10103479e-02 -2.94808060e-01 1.12345088e+00
5.49613595e-01 2.55937219e-01 2.94324100e-01 -1.43244267e+00
-5.83525777e-01 4.67372179e-01 -6.51395738e-01 6.70860648e-01
4.20253783e-01 1.32164180e-01 8.37185860e-01 -1.09433389e+00
5.03249705e-01 8.68299246e-01 3.40374142e-01 5.13717949e-01
-8.34152281e-01 -5.83263516e-01 3.37667882e-01 7.14107811e-01
-1.60409904e+00 -5.59750080e-01 1.15392554e+00 -3.36274534e-01
6.88676953e-01 1.60963967e-01 1.17404139e+00 1.33178306e+00
-6.10597320e-02 9.87783432e-01 9.64260697e-01 -5.51550865e-01
1.40288830e-01 3.21958899e-01 -2.69423217e-01 6.25713944e-01
-7.79918209e-02 1.46900773e-01 -9.17497933e-01 2.31980383e-01
8.00687015e-01 1.56420961e-01 -4.60622489e-01 -5.22280812e-01
-1.33312666e+00 6.23643219e-01 5.57273567e-01 6.75611377e-01
-3.23946297e-01 -9.12897810e-02 4.46868628e-01 1.12796560e-01
4.99118865e-01 2.68143922e-01 -1.59266740e-01 -8.12542140e-02
-1.28259528e+00 1.33290380e-01 3.15476842e-02 7.90420830e-01
7.28638828e-01 3.67175400e-01 -3.04427624e-01 6.85188413e-01
3.01337481e-01 4.94616717e-01 8.16069365e-01 -6.40301108e-01
3.69304836e-01 4.72356230e-01 1.26280293e-01 -1.36740565e+00
-3.62884760e-01 -7.40903437e-01 -6.44719660e-01 7.53406361e-02
2.87683427e-01 2.85560697e-01 -7.05035508e-01 1.84833324e+00
2.87292361e-01 4.50811625e-01 -2.46466488e-01 1.22659576e+00
5.94068944e-01 4.59175736e-01 1.24257579e-01 -6.04537725e-01
1.14000165e+00 -9.25810575e-01 -8.18913043e-01 -6.80964515e-02
3.04932654e-01 -7.15360940e-01 1.48593867e+00 4.01659042e-01
-1.00064564e+00 -9.34269369e-01 -9.19263721e-01 1.94732711e-01
-3.26247752e-01 9.62966606e-02 4.91645873e-01 3.91901642e-01
-1.10802472e+00 2.23752424e-01 -5.97318232e-01 -3.92840624e-01
5.63178360e-01 1.66027009e-01 -1.71492279e-01 1.20278947e-01
-1.31664789e+00 6.69175684e-01 5.72575092e-01 4.68313187e-01
-1.35034466e+00 -4.50852841e-01 -6.36937976e-01 1.70162112e-01
5.92382908e-01 -4.17792588e-01 9.43165183e-01 -1.50902617e+00
-1.07364559e+00 9.43638027e-01 -2.94928253e-01 -3.69433880e-01
4.17351723e-01 -9.26042199e-02 -5.43369055e-01 2.80642122e-01
2.10174143e-01 1.00622690e+00 1.20099294e+00 -1.29107952e+00
-8.57432365e-01 -1.29668951e-01 9.25432667e-02 2.81564265e-01
-6.20723665e-01 9.76233482e-02 -4.30218220e-01 -7.81013608e-01
6.14951225e-03 -7.70720124e-01 9.64528695e-02 -1.71834350e-01
-1.91601634e-01 -2.75984377e-01 7.26029932e-01 -3.29744875e-01
1.39330816e+00 -2.23901343e+00 2.59378791e-01 -3.19629014e-02
1.98264554e-01 6.09253764e-01 -6.62727058e-02 4.58730273e-02
-2.97279835e-01 -2.58749008e-01 -8.17167610e-02 -3.22777361e-01
-3.59625578e-01 -2.38295123e-01 -2.73482442e-01 3.14646572e-01
1.18112631e-01 7.82767057e-01 -1.32275569e+00 -8.60332489e-01
3.24510634e-01 2.15347350e-01 -5.33194840e-01 3.07612926e-01
-2.17914894e-01 5.34982860e-01 -3.17685664e-01 7.66526043e-01
3.10453862e-01 -1.69952974e-01 -1.78205550e-01 -1.20998167e-01
-3.65268767e-01 -1.20531194e-01 -8.44906986e-01 1.94195104e+00
-3.36885676e-02 7.82870650e-01 -1.91706032e-01 -1.21302617e+00
9.88586187e-01 3.09714198e-01 5.87623358e-01 -9.17507172e-01
1.21537469e-01 4.10210758e-01 9.94544011e-03 -8.22942853e-01
4.00981814e-01 -1.28628820e-01 9.12770405e-02 1.84314534e-01
1.74510375e-01 3.22464645e-01 2.04920337e-01 1.70127809e-04
6.78247333e-01 2.16088191e-01 6.50544390e-02 -2.36210555e-01
7.03047395e-01 2.40223622e-03 4.72298294e-01 4.68112111e-01
-5.25333881e-01 8.62971663e-01 1.82999015e-01 -3.52173835e-01
-8.66166651e-01 -9.14606929e-01 5.73210344e-02 1.22561610e+00
3.45082998e-01 -2.49259651e-01 -8.99994969e-01 -5.56728542e-01
-4.02457803e-01 3.38818848e-01 -6.51531696e-01 -4.48326409e-01
-3.18658084e-01 -4.26697940e-01 3.54250371e-01 3.34951431e-01
5.04666209e-01 -1.65618145e+00 -8.83127451e-01 1.17786877e-01
-4.03110832e-01 -8.12582314e-01 -3.33901763e-01 9.44814160e-02
-6.34537876e-01 -1.01482797e+00 -9.64197218e-01 -1.04302299e+00
4.47449028e-01 7.31995821e-01 7.93615341e-01 1.75103396e-01
-6.52699545e-02 2.51126345e-02 -4.48005915e-01 -4.49818701e-01
7.85675868e-02 1.89527661e-01 1.39509678e-01 4.02722418e-01
6.46741152e-01 -6.32418573e-01 -7.21407592e-01 3.23525608e-01
-8.72250140e-01 2.18635082e-01 4.44950551e-01 6.99937940e-01
6.37005806e-01 -7.03980848e-02 8.05817902e-01 -4.44635242e-01
4.06183362e-01 -4.80440259e-01 -4.48244780e-01 2.04024389e-01
-3.28564227e-01 -3.92220616e-01 4.72663373e-01 -6.46984458e-01
-8.77983928e-01 4.48479364e-03 3.51038128e-02 -8.96315038e-01
-3.32625329e-01 6.66654050e-01 -4.32574973e-02 3.95719260e-02
6.06396496e-01 2.77876168e-01 -1.91718906e-01 -2.26781324e-01
-7.02176914e-02 6.70121372e-01 5.04612148e-01 -1.18083157e-01
7.52413571e-01 2.70121306e-01 -2.12823659e-01 -7.34806359e-01
-1.13694072e+00 -4.66224432e-01 -7.53945172e-01 -6.72386527e-01
9.34442759e-01 -9.82983589e-01 -5.18318295e-01 5.14060795e-01
-1.05724394e+00 -2.14847118e-01 -2.25931779e-01 5.45851529e-01
-6.47694111e-01 4.44353014e-01 -6.56612441e-02 -9.19015229e-01
-1.40219573e-02 -1.16241121e+00 1.02608871e+00 2.36376330e-01
-1.17033832e-01 -5.75932801e-01 2.07604188e-02 2.26272479e-01
3.06599885e-01 1.45399690e-01 6.14673615e-01 -3.83135617e-01
-6.71482146e-01 6.62946478e-02 -2.40763262e-01 1.41423166e-01
4.58515584e-02 -4.08181548e-03 -1.31182277e+00 -2.92431742e-01
2.13294283e-01 -4.72539485e-01 7.90077746e-01 5.37314594e-01
1.54692817e+00 1.60382852e-01 -2.80774862e-01 3.52792352e-01
1.17151845e+00 2.69547015e-01 6.22911036e-01 3.92413259e-01
7.23599017e-01 6.70998573e-01 1.12309718e+00 3.54079485e-01
2.72241652e-01 8.53081644e-01 7.17536986e-01 -2.05434158e-01
-1.07403152e-01 -5.27442813e-01 3.82048011e-01 1.02136159e+00
-2.32490018e-01 -2.91109920e-01 -7.17088699e-01 8.60675097e-01
-2.05361438e+00 -1.26213646e+00 2.43913494e-02 2.28324485e+00
7.03806579e-01 3.88151735e-01 2.36233592e-01 4.63010520e-01
1.04390752e+00 3.30029815e-01 -5.06349802e-01 -4.14585415e-03
-2.02106953e-01 -1.18059225e-01 -1.55796155e-01 1.51976988e-01
-1.19822538e+00 1.01342845e+00 6.08956146e+00 1.02751791e+00
-1.39643335e+00 2.73171395e-01 3.64156038e-01 -3.02662492e-01
-3.60550106e-01 -1.46172822e-01 -5.86699843e-01 7.44824409e-01
9.41065073e-01 1.90839946e-01 4.87555981e-01 7.59691238e-01
4.89526629e-01 -2.19348341e-01 -8.36170793e-01 1.25321221e+00
3.13387066e-01 -1.11109567e+00 2.40208041e-02 -1.46785095e-01
5.03082097e-01 -1.07444979e-01 3.09642762e-01 2.84261465e-01
-4.24581438e-01 -8.73412967e-01 1.06559134e+00 5.70692658e-01
8.64393234e-01 -6.02766514e-01 6.28031492e-01 4.23006862e-01
-1.11101878e+00 -2.83637345e-01 -5.42083323e-01 5.29409535e-02
1.06217206e-01 3.98025155e-01 -5.97883105e-01 5.00816584e-01
1.01465929e+00 8.38403404e-01 -6.93286419e-01 1.33288312e+00
-1.20021768e-01 5.73529243e-01 -2.90763117e-02 -1.23626485e-01
2.31151775e-01 7.39608109e-02 4.99680609e-01 1.00308156e+00
4.29760635e-01 -2.32363299e-01 1.25251755e-01 7.12834597e-01
1.03794038e-01 1.75135285e-01 -5.70584953e-01 -2.07644049e-03
4.57933903e-01 1.13015664e+00 -8.97697508e-01 -2.05498442e-01
-3.97169322e-01 6.12012744e-01 4.48942900e-01 4.47425574e-01
-1.13317704e+00 -3.02534401e-01 1.86623111e-01 2.56619126e-01
4.40614671e-01 8.07168484e-02 -6.63066655e-02 -1.30626893e+00
1.29229546e-01 -8.18777025e-01 1.91989124e-01 -1.21558809e+00
-8.93316269e-01 9.11936462e-01 7.53895119e-02 -1.61744893e+00
-1.70704976e-01 -2.53283560e-01 -4.59107727e-01 5.99547505e-01
-1.62818766e+00 -9.75294769e-01 -4.54594314e-01 9.95823503e-01
7.51266658e-01 -4.28373009e-01 6.82735503e-01 4.95460957e-01
-5.40120780e-01 5.40434062e-01 -1.23865962e-01 -2.10722297e-01
7.44588435e-01 -7.59012878e-01 -2.79054791e-01 9.84722257e-01
4.39107388e-01 4.35703456e-01 8.33780110e-01 -4.18662131e-01
-7.89470136e-01 -8.08838725e-01 9.38014925e-01 -2.67713964e-01
6.73289657e-01 -5.16077161e-01 -1.09465504e+00 3.57529968e-01
2.66901910e-01 1.14792748e-03 4.81641859e-01 3.03687025e-02
4.73867990e-02 -3.57292324e-01 -8.70256662e-01 5.50237775e-01
1.17230594e+00 -7.94940889e-01 -7.95711160e-01 1.84712470e-01
7.56866693e-01 -2.16209158e-01 -2.59806007e-01 4.10080373e-01
3.86950850e-01 -1.00912273e+00 8.40567648e-01 -5.22755146e-01
5.03514886e-01 -5.74008763e-01 -2.35144775e-02 -1.13390136e+00
-3.42234790e-01 -3.66400868e-01 -1.88560456e-01 1.28350735e+00
1.41909182e-01 -1.60107106e-01 6.87349737e-01 -8.59808102e-02
-4.13711876e-01 -6.21024668e-01 -1.02846408e+00 -3.37469280e-01
-6.39190674e-01 -3.37545663e-01 3.79376650e-01 9.80826557e-01
1.68446582e-02 2.05219924e-01 -5.65594554e-01 -8.23803693e-02
3.65865558e-01 3.11848491e-01 4.81399417e-01 -1.38288593e+00
-9.43872407e-02 -3.02517593e-01 -5.41573703e-01 -9.16187108e-01
1.93788812e-01 -6.59995258e-01 2.12834060e-01 -1.00415695e+00
2.63888448e-01 -2.61356920e-01 -9.92908359e-01 3.77093375e-01
-2.23145038e-01 4.89222765e-01 3.90714444e-02 3.93350780e-01
-1.04562783e+00 8.06530833e-01 1.19994414e+00 -1.00811996e-01
-1.89270094e-01 9.83491689e-02 -6.10948086e-01 5.72279751e-01
6.52345777e-01 -5.24566531e-01 -6.91656947e-01 -3.93540293e-01
1.31990120e-01 -2.33616354e-03 3.50814700e-01 -1.20828927e+00
2.29350209e-01 -1.89109474e-01 3.19965124e-01 -7.82780588e-01
3.01888168e-01 -7.98295975e-01 6.44872664e-03 2.54170537e-01
-4.12359536e-01 3.85126681e-03 1.13836065e-01 7.42710114e-01
-6.23458326e-01 -3.24472755e-01 6.72966242e-01 -1.83935717e-01
-9.99579072e-01 3.05559307e-01 -4.30452168e-01 -1.03921369e-01
1.12960267e+00 -5.89292347e-01 -6.34023100e-02 -1.40907556e-01
-7.25804865e-01 9.05171409e-03 4.81516898e-01 6.69628859e-01
6.96591437e-01 -1.44914806e+00 -3.02883238e-01 2.71544814e-01
2.47942597e-01 9.22108535e-03 6.29339814e-01 9.81684208e-01
-1.39204249e-01 5.68895698e-01 -5.83212554e-01 -9.03381705e-01
-1.14602220e+00 1.16021299e+00 4.80724238e-02 1.70756996e-01
-3.07868391e-01 7.25945950e-01 4.17444050e-01 2.72997677e-01
4.52157199e-01 -9.57627818e-02 -6.58540010e-01 3.32079142e-01
6.23040020e-01 1.02918372e-01 8.94856080e-02 -7.82947779e-01
-2.79098094e-01 4.12444890e-01 2.38911912e-01 5.39305732e-02
1.35134387e+00 -4.04542714e-01 1.51308954e-01 8.24688196e-01
9.48179126e-01 -1.92271620e-01 -1.38584721e+00 -2.31246084e-01
-8.51681009e-02 -7.47836411e-01 8.86722356e-02 -2.86188751e-01
-1.32307494e+00 1.31608307e+00 7.48814464e-01 4.77964610e-01
1.30506730e+00 -9.79706943e-02 5.41003644e-01 1.38418436e-01
5.95755279e-01 -1.18836594e+00 4.51073647e-01 2.51042873e-01
7.82462418e-01 -1.29742622e+00 -3.03870410e-01 -1.71829373e-01
-7.68137395e-01 8.65927875e-01 7.27957487e-01 -3.49489711e-02
5.26842952e-01 -2.40421653e-01 -5.34615683e-05 -1.85951814e-01
-6.01209104e-01 -4.19071615e-01 3.03719193e-01 7.68878460e-01
3.54379594e-01 -2.32267916e-01 -1.47730440e-01 3.07095826e-01
6.65480345e-02 7.15362877e-02 3.27718765e-01 7.19808161e-01
-7.85044193e-01 -7.05259204e-01 -3.38531435e-01 2.11378932e-01
-2.90520221e-01 -5.04815765e-02 -3.84632796e-02 6.41107976e-01
2.48495340e-01 9.56603706e-01 7.79462680e-02 -4.52561677e-01
2.41348639e-01 9.10485536e-02 3.27156901e-01 -5.86976290e-01
-4.31109071e-01 3.56929988e-01 -3.50254446e-01 -4.61456656e-01
-1.04197598e+00 -7.94760048e-01 -1.01060915e+00 -6.06558919e-02
-5.00264347e-01 1.95989192e-01 1.46151930e-01 9.46066916e-01
2.43488029e-01 5.87191999e-01 6.66911364e-01 -1.17866027e+00
-1.89489692e-01 -1.00806403e+00 -6.17095530e-01 4.40687567e-01
2.73727238e-01 -1.21066201e+00 -3.06595892e-01 2.72736680e-02] | [9.779098510742188, -0.2537320554256439] |
2001c0ef-f258-4152-a244-84d278251497 | effective-open-intent-classification-with-k | 2304.10220 | null | https://arxiv.org/abs/2304.10220v1 | https://arxiv.org/pdf/2304.10220v1.pdf | Effective Open Intent Classification with K-center Contrastive Learning and Adjustable Decision Boundary | Open intent classification, which aims to correctly classify the known intents into their corresponding classes while identifying the new unknown (open) intents, is an essential but challenging task in dialogue systems. In this paper, we introduce novel K-center contrastive learning and adjustable decision boundary learning (CLAB) to improve the effectiveness of open intent classification. First, we pre-train a feature encoder on the labeled training instances, which transfers knowledge from known intents to unknown intents. Specifically, we devise a K-center contrastive learning algorithm to learn discriminative and balanced intent features, improving the generalization of the model for recognizing open intents. Second, we devise an adjustable decision boundary learning method with expanding and shrinking (ADBES) to determine the suitable decision conditions. Concretely, we learn a decision boundary for each known intent class, which consists of a decision center and the radius of the decision boundary. We then expand the radius of the decision boundary to accommodate more in-class instances if the out-of-class instances are far from the decision boundary; otherwise, we shrink the radius of the decision boundary. Extensive experiments on three benchmark datasets clearly demonstrate the effectiveness of our method for open intent classification. For reproducibility, we submit the code at: https://github.com/lxk00/CLAP | ['Benyou Wang', 'Ruifeng Xu', 'Min Yang', 'Jingjing Mu', 'Jianquan Li', 'Xiaokang Liu'] | 2023-04-20 | null | null | null | null | ['intent-classification'] | ['natural-language-processing'] | [-5.73327346e-03 2.30332315e-01 -3.88152838e-01 -6.54970884e-01
-7.03150153e-01 -8.19006085e-01 3.03675890e-01 4.71004695e-02
-3.05752665e-01 6.11781240e-01 4.37306762e-01 -4.59396720e-01
-2.50207521e-02 -5.40007591e-01 -2.13428199e-01 -6.41275227e-01
-4.59341407e-02 6.23610675e-01 6.05700575e-02 -2.39380270e-01
4.47441459e-01 -3.89652885e-02 -1.25058961e+00 3.49746644e-01
1.21172917e+00 1.05290806e+00 5.42465076e-02 6.48577452e-01
-2.76532829e-01 6.74366176e-01 -6.12509668e-01 -1.02584489e-01
2.90932626e-01 -3.49334836e-01 -1.20959222e+00 6.99177012e-02
1.41302615e-01 -4.60918069e-01 -2.09991679e-01 7.37371266e-01
1.95917130e-01 3.98707718e-01 9.18991148e-01 -1.32512975e+00
-6.84636295e-01 4.39995915e-01 -4.89737034e-01 2.35703245e-01
6.41854703e-01 1.02639176e-01 1.30221665e+00 -8.40857387e-01
3.02497089e-01 8.53449404e-01 5.88754296e-01 7.27946937e-01
-8.94851506e-01 -6.43493295e-01 3.39417785e-01 2.11597711e-01
-1.20805359e+00 -4.30038959e-01 9.52529728e-01 -4.18324381e-01
5.00893891e-01 5.07970154e-01 4.06421959e-01 9.14590061e-01
-1.50865205e-02 9.60875809e-01 9.40232158e-01 -4.64819700e-01
4.39883173e-01 2.12945506e-01 7.89438009e-01 7.49273181e-01
-1.85007602e-01 -1.74886331e-01 -2.19435424e-01 -3.63768250e-01
3.75420690e-01 3.23947817e-01 -5.83430767e-01 -2.18896508e-01
-9.68906522e-01 1.09134376e+00 5.80707073e-01 2.11463898e-01
-3.39769870e-02 -4.64929909e-01 5.71218729e-01 4.62338835e-01
5.84617198e-01 4.86172825e-01 -6.29435956e-01 -2.66784608e-01
-4.79639143e-01 -1.56249385e-02 1.02591944e+00 8.03078413e-01
1.10679424e+00 -7.07441688e-01 -4.77476448e-01 1.18887210e+00
2.47280672e-01 -1.27094500e-02 7.29555368e-01 -7.29347229e-01
5.83566129e-01 1.10675144e+00 -9.80210006e-02 -6.61679089e-01
-4.06765550e-01 -1.69825271e-01 -7.81862020e-01 -3.20539504e-01
4.06982720e-01 -4.33154970e-01 -6.73368692e-01 1.50383687e+00
5.50990462e-01 3.80077846e-02 3.76029700e-01 9.01943803e-01
9.13941264e-01 8.48032594e-01 -2.61022180e-01 -3.21441174e-01
1.36928320e+00 -1.22967827e+00 -5.79629362e-01 -2.66259015e-01
1.09145117e+00 -5.05877197e-01 1.37655890e+00 6.74057603e-02
-3.90685201e-01 -4.38271344e-01 -9.78384912e-01 -1.65191963e-01
-3.40176463e-01 2.10923314e-01 5.86691618e-01 3.80423188e-01
-6.32305682e-01 1.83364883e-01 -4.17673379e-01 -3.44521403e-02
1.40054941e-01 2.38391325e-01 -1.39384478e-01 -1.10180900e-01
-1.37818527e+00 3.73591870e-01 4.95820671e-01 -1.19432315e-01
-5.85731208e-01 -6.65889144e-01 -1.18586349e+00 1.08565204e-01
6.05796993e-01 -3.45605433e-01 1.26424587e+00 -9.06559587e-01
-1.39796638e+00 7.90289819e-01 1.04150372e-02 -2.58625150e-01
4.00366455e-01 4.76123393e-02 -1.21667035e-01 -2.54164971e-02
2.52302051e-01 6.76918507e-01 9.03462231e-01 -9.95564699e-01
-7.21266508e-01 -4.62564826e-01 4.12361234e-01 6.19426072e-01
-7.95241892e-01 -3.19660813e-01 -2.72969157e-01 -4.67702001e-01
2.96535462e-01 -9.64825273e-01 -3.61228846e-02 -1.24682143e-01
-4.99729276e-01 -8.46323609e-01 1.02117169e+00 -5.21778166e-01
1.50793886e+00 -2.42231965e+00 -9.82820690e-02 -6.35605380e-02
4.91117626e-01 2.61061847e-01 4.11739647e-02 2.72273004e-01
-7.34019978e-03 -1.15505747e-01 -3.07502270e-01 -2.44793877e-01
-3.19582522e-02 3.32206130e-01 -3.91849369e-01 2.43682101e-01
-8.50667655e-02 5.95148325e-01 -8.68101537e-01 -3.22118700e-01
1.16586260e-01 -7.91011676e-02 -7.24580288e-01 5.91430545e-01
-3.04477692e-01 3.72063041e-01 -5.48053801e-01 3.41762245e-01
5.59503734e-01 -3.35815370e-01 -1.03897192e-01 1.42395766e-02
-6.33925572e-02 6.81763530e-01 -1.11763060e+00 1.41374874e+00
-6.50097013e-01 5.00321984e-01 -1.19296834e-01 -8.50071728e-01
1.15248430e+00 1.32301927e-01 2.20285580e-01 -2.07247987e-01
1.39497265e-01 1.63178653e-01 -3.87079245e-03 -4.39370900e-01
5.04755795e-01 -9.14641321e-02 -4.05969501e-01 7.36653447e-01
2.34402642e-02 -1.12165384e-04 6.27680048e-02 5.72735704e-02
9.04530466e-01 -3.66609752e-01 4.61532533e-01 -2.40405574e-01
7.96977878e-01 -1.95133850e-01 6.31831408e-01 5.82304716e-01
-5.36872387e-01 5.61216891e-01 7.25802243e-01 -7.22518742e-01
-5.60415924e-01 -6.60592079e-01 -2.79173076e-01 1.50797606e+00
2.61104196e-01 -5.08400798e-01 -7.29510069e-01 -1.39701033e+00
-4.92106229e-02 7.06866682e-01 -6.98540688e-01 -3.88893932e-01
-3.46000403e-01 -4.58342969e-01 2.95471758e-01 4.01488513e-01
7.27207661e-01 -9.31765616e-01 -4.40340608e-01 -2.13421419e-01
-5.09165943e-01 -6.95354819e-01 -1.07816422e+00 5.51865578e-01
-4.44323033e-01 -1.16584575e+00 -6.94910228e-01 -1.08475757e+00
8.73328805e-01 4.17574048e-01 6.52834058e-01 1.54881597e-01
2.72914916e-02 1.44126803e-01 -6.46935940e-01 -4.42900844e-02
-4.78640229e-01 2.81870663e-01 1.25629574e-01 1.25804767e-01
4.93704259e-01 -5.06358296e-02 -4.88784850e-01 5.42369008e-01
-5.19371867e-01 1.65567636e-01 1.78669751e-01 1.14351463e+00
2.31720760e-01 -8.91642570e-02 6.08178437e-01 -7.89331079e-01
8.72331798e-01 -6.18022025e-01 -3.20149660e-01 1.49711579e-01
-4.78440017e-01 2.05624625e-01 7.96156108e-01 -5.05461156e-01
-8.29159856e-01 -1.93686232e-01 -2.83677846e-01 -3.72446060e-01
-3.31283420e-01 5.35849810e-01 -1.60445780e-01 1.40919983e-01
4.57521051e-01 1.79266915e-01 -2.72489283e-02 -1.82688519e-01
2.72172540e-01 1.53235185e+00 1.70705959e-01 -4.77478534e-01
3.54926288e-01 1.76450238e-01 -7.67476678e-01 -7.29147196e-01
-1.13734496e+00 -8.43032658e-01 -6.14994824e-01 -1.50264814e-01
5.52521884e-01 -6.41632438e-01 -4.81695652e-01 4.62674707e-01
-8.30572009e-01 -4.47149843e-01 -2.91258782e-01 3.03213477e-01
-5.04714966e-01 5.36375284e-01 -5.41638672e-01 -6.29384041e-01
-5.59935987e-01 -1.16169536e+00 1.01973367e+00 6.00538850e-01
-4.76675570e-01 -1.14164710e+00 2.72664696e-01 6.82017446e-01
-6.68741763e-02 -5.61395466e-01 9.31493700e-01 -1.50094748e+00
3.84074971e-02 -2.40854517e-01 -1.13833182e-01 3.83375794e-01
4.14691746e-01 -3.50881368e-01 -8.92265379e-01 -3.10308754e-01
2.12521434e-01 -6.82920277e-01 8.53442430e-01 8.54021832e-02
9.74847198e-01 -5.70336938e-01 -4.13825750e-01 6.37081265e-01
7.29055822e-01 2.98869103e-01 3.32295269e-01 2.33753726e-01
5.28376818e-01 5.21850765e-01 8.56241703e-01 7.28776753e-01
5.91705441e-01 5.72242975e-01 1.49990479e-02 4.47642654e-02
1.14722855e-01 -2.50265539e-01 3.38524282e-01 6.44513547e-01
5.01327276e-01 -1.53861910e-01 -9.61678565e-01 3.73255253e-01
-1.63219821e+00 -5.28111398e-01 2.31139883e-01 2.23973107e+00
1.19621205e+00 2.11724177e-01 -1.03045069e-01 1.43459320e-01
9.12435234e-01 2.52500296e-01 -6.95952713e-01 -4.16377515e-01
4.24891174e-01 -4.17505443e-01 -1.13448478e-01 6.54668331e-01
-1.22852504e+00 8.54416072e-01 5.11134243e+00 8.45562220e-01
-1.05083549e+00 -1.01833284e-01 8.98921609e-01 2.79111236e-01
-2.24752083e-01 1.10009924e-01 -1.07829487e+00 4.86830473e-01
4.87796068e-01 8.99659321e-02 2.28739887e-01 9.90060747e-01
-9.38695893e-02 -6.58883378e-02 -1.26970398e+00 8.35567772e-01
3.84679705e-01 -1.01652372e+00 -9.92969051e-02 -2.09881458e-02
6.64102793e-01 -1.65435582e-01 -1.21710911e-01 8.85451615e-01
2.53997535e-01 -4.31166708e-01 1.06537253e-01 8.07416663e-02
6.57145381e-01 -5.46759248e-01 7.73782790e-01 7.27292895e-01
-9.33483779e-01 -3.12493712e-01 -2.54041433e-01 -3.11200678e-01
-2.77355492e-01 4.12401080e-01 -9.50326622e-01 1.69470385e-01
5.19997239e-01 7.54682362e-01 -5.31452358e-01 7.91568756e-01
-3.06220531e-01 4.58652079e-01 -2.49045700e-01 -2.16226473e-01
2.91889012e-01 -3.28169018e-01 4.41818565e-01 8.38837564e-01
-2.42026135e-01 1.54041931e-01 5.97300828e-01 7.01458812e-01
-1.49730951e-01 1.91404387e-01 -5.42356908e-01 1.92652211e-01
5.37753403e-01 1.15688419e+00 -5.07146180e-01 -1.96429878e-01
-5.21058381e-01 9.49801385e-01 6.42669261e-01 1.21331923e-01
-8.45681489e-01 -7.02164412e-01 4.98903900e-01 -2.79386848e-01
1.66250095e-01 2.09971994e-01 -2.36372441e-01 -1.28117204e+00
8.45296830e-02 -6.43025696e-01 8.96965146e-01 -3.77294511e-01
-1.37211215e+00 5.18158972e-01 -9.04264376e-02 -1.39186251e+00
-1.99430630e-01 -4.45740759e-01 -8.97249341e-01 6.18131399e-01
-1.28398359e+00 -9.32831466e-01 -4.30476099e-01 3.68335694e-01
1.01535976e+00 -1.47136137e-01 7.79887855e-01 -1.72573894e-01
-6.74678445e-01 8.12047482e-01 1.98285341e-01 4.54157919e-01
8.83509576e-01 -1.09714460e+00 1.88329220e-02 3.29582036e-01
-8.66352692e-02 6.55504286e-01 3.98101747e-01 -5.30877888e-01
-9.51883912e-01 -9.54864860e-01 8.32393646e-01 -1.96192667e-01
6.39231086e-01 -5.74482501e-01 -1.08794844e+00 7.94095337e-01
4.54177260e-02 7.78736919e-02 1.12786663e+00 4.81981009e-01
-4.07453537e-01 3.91224883e-02 -9.71096814e-01 6.60799682e-01
4.99744922e-01 -3.38037312e-01 -9.72597480e-01 4.68057930e-01
8.92603159e-01 -5.33309698e-01 -7.75779903e-01 4.91422087e-01
4.28818077e-01 -9.14476871e-01 5.84204555e-01 -5.51960289e-01
2.74810761e-01 3.30728628e-02 -6.56419471e-02 -1.37328088e+00
-1.61246121e-01 -5.13554990e-01 -2.55602181e-01 1.28718877e+00
4.70913738e-01 -7.91580021e-01 7.21881449e-01 6.70309007e-01
-4.30532306e-01 -1.28395855e+00 -9.56896782e-01 -4.40870911e-01
9.15328115e-02 -1.49588302e-01 2.90472627e-01 1.11874855e+00
6.91316605e-01 6.95347309e-01 -1.51406541e-01 -7.28006512e-02
1.45241737e-01 5.60923159e-01 7.14979947e-01 -1.10203254e+00
-2.34512717e-01 -2.88937122e-01 -1.33977100e-01 -1.55560076e+00
5.00250876e-01 -7.83881962e-01 7.98864588e-02 -1.28349304e+00
4.55121011e-01 -1.04914987e+00 -6.31322637e-02 8.22566211e-01
-6.11896455e-01 -8.59806463e-02 -1.64771248e-02 5.38541615e-01
-8.72318745e-01 8.58727634e-01 1.14197457e+00 -2.55262196e-01
-6.36082232e-01 4.70491588e-01 -7.52697110e-01 7.95186639e-01
7.98206508e-01 -3.14996809e-01 -4.94943768e-01 3.21906470e-02
-1.72901928e-01 5.64317442e-02 -1.24383131e-02 -7.51200676e-01
1.25730336e-01 -2.57599771e-01 1.66367978e-01 -7.33557522e-01
1.24104545e-01 -5.40210605e-01 -6.86699331e-01 5.17808616e-01
-7.95409501e-01 -5.21610975e-01 2.20616050e-02 3.93573999e-01
-5.16038090e-02 -7.30690956e-01 7.00308800e-01 1.13299720e-01
-8.07949305e-01 1.64149359e-01 -2.19496846e-01 4.03926641e-01
1.25249612e+00 -1.88430101e-01 -4.55831051e-01 -3.13288838e-01
-6.02928638e-01 8.49258780e-01 4.29928094e-01 6.48770273e-01
7.60202408e-01 -9.50874567e-01 -5.13171017e-01 6.40606284e-01
5.51916838e-01 2.35153899e-01 2.85825074e-01 7.47770011e-01
-2.50338733e-01 4.49399710e-01 1.94444641e-01 -4.94859040e-01
-1.26002836e+00 6.54245436e-01 4.32153344e-01 -3.95579994e-01
-6.00602508e-01 8.90410542e-01 4.44617599e-01 -1.10197902e+00
4.33562845e-01 -1.18783243e-01 -5.44248998e-01 1.25859812e-01
6.14991784e-01 1.67064413e-01 -2.12180037e-02 -4.70707625e-01
-2.41258800e-01 2.29817480e-01 -6.37752950e-01 2.81944156e-01
8.18303943e-01 -2.96294987e-01 -1.13101080e-01 6.14189327e-01
1.39510036e+00 -1.92124292e-01 -1.27700281e+00 -1.92973346e-01
-2.61961997e-01 -2.77398646e-01 -7.56544843e-02 -6.59200311e-01
-6.32551789e-01 6.93150699e-01 4.04115140e-01 3.13838691e-01
1.06718981e+00 3.71320605e-01 7.54055619e-01 3.93140614e-01
1.71697319e-01 -9.30220366e-01 2.19288617e-01 9.63378012e-01
7.24295437e-01 -1.37944186e+00 -2.27733284e-01 -4.11127239e-01
-8.58137250e-01 1.11255085e+00 9.69756186e-01 2.08442301e-01
5.30172586e-01 -3.64290439e-02 2.19201803e-01 -4.51235808e-02
-9.11640525e-01 5.08483984e-02 1.13286883e-01 3.88491243e-01
3.17400932e-01 1.34622231e-01 -2.62440830e-01 8.58348668e-01
-1.65668130e-01 -3.30166131e-01 5.68324864e-01 9.82178390e-01
-5.70744932e-01 -7.95404851e-01 -3.53876591e-01 6.97841048e-01
8.30855742e-02 -1.91470712e-01 -4.76582557e-01 2.88619876e-01
2.65690451e-03 1.03147840e+00 2.28208214e-01 -6.00101054e-01
-9.28219780e-03 2.25080878e-01 -2.74633542e-02 -8.97414207e-01
-4.94714051e-01 -2.01584682e-01 4.55743149e-02 1.08233234e-03
1.67007759e-01 -5.74408233e-01 -1.49704027e+00 2.62684235e-03
-9.00386453e-01 6.17712796e-01 2.31774747e-01 1.09388411e+00
3.82973522e-01 2.91396707e-01 1.20735204e+00 -4.20795500e-01
-1.01079786e+00 -1.19780147e+00 -4.34521288e-01 3.03106934e-01
5.08133769e-01 -6.35070920e-01 -6.60250068e-01 -1.60167426e-01] | [12.395990371704102, 7.437747955322266] |
0ef722cc-62d2-496f-93b6-395338d0a7b0 | oriental-language-recognition-olr-2020 | 2107.05365 | null | https://arxiv.org/abs/2107.05365v1 | https://arxiv.org/pdf/2107.05365v1.pdf | Oriental Language Recognition (OLR) 2020: Summary and Analysis | The fifth Oriental Language Recognition (OLR) Challenge focuses on language recognition in a variety of complex environments to promote its development. The OLR 2020 Challenge includes three tasks: (1) cross-channel language identification, (2) dialect identification, and (3) noisy language identification. We choose Cavg as the principle evaluation metric, and the Equal Error Rate (EER) as the secondary metric. There were 58 teams participating in this challenge and one third of the teams submitted valid results. Compared with the best baseline, the Cavg values of Top 1 system for the three tasks were relatively reduced by 82%, 62% and 48%, respectively. This paper describes the three tasks, the database profile, and the final results. We also outline the novel approaches that improve the performance of language recognition systems most significantly, such as the utilization of auxiliary information. | ['Dong Wang', 'Qingyang Hong', 'Lin Li', 'Zheng Li', 'Yiming Zhi', 'Binling Wang', 'Jing Li'] | 2021-07-05 | null | null | null | null | ['dialect-identification'] | ['natural-language-processing'] | [-1.98052764e-01 -6.88906550e-01 -1.29931241e-01 -4.59823966e-01
-1.46913469e+00 -7.85994291e-01 7.41037250e-01 -4.11902815e-01
-8.89100850e-01 5.71355999e-01 1.91027999e-01 -5.88055015e-01
4.96793419e-01 -9.67897773e-02 -2.82815009e-01 -5.70507765e-01
9.32166576e-02 3.29061359e-01 -1.45633385e-01 -4.22897160e-01
1.47684842e-01 7.83657372e-01 -1.27455819e+00 4.92797077e-01
8.18011999e-01 9.66626823e-01 -7.25766935e-04 7.58230984e-01
-1.08101703e-01 6.86975718e-01 -7.47327030e-01 -4.33045715e-01
1.00366831e-01 -5.61840236e-02 -8.04992855e-01 -5.57131916e-02
2.10410595e-01 -2.47304559e-01 -5.44185400e-01 9.83748198e-01
1.12816727e+00 2.20106661e-01 5.19863844e-01 -1.02420175e+00
-6.41636014e-01 8.65331054e-01 -3.58567744e-01 1.24039471e-01
6.28817379e-01 3.17411534e-02 8.20160091e-01 -1.60250378e+00
2.56143212e-01 1.60412049e+00 6.34036779e-01 8.29752743e-01
-1.01605928e+00 -8.92074227e-01 3.32574099e-01 4.07073572e-02
-2.08540082e+00 -1.22255445e+00 4.25132930e-01 -3.83677661e-01
1.33930862e+00 4.59166318e-01 -4.92483564e-03 1.10248792e+00
-4.71824229e-01 1.18389428e+00 1.19425344e+00 -6.99162841e-01
-2.41797775e-01 5.47626615e-01 4.12766367e-01 3.96387845e-01
-1.15499303e-01 -1.67094499e-01 -7.75637329e-01 -1.45122945e-01
1.81283355e-01 -7.51134455e-01 -2.33494326e-01 4.92432922e-01
-1.34836972e+00 6.44818187e-01 -3.07203352e-01 4.96222794e-01
-2.81746015e-02 -2.95829207e-01 5.76396585e-01 4.00743306e-01
5.02103031e-01 4.43185985e-01 -4.70598191e-01 -1.95445076e-01
-7.18812883e-01 -1.41286448e-01 7.85927415e-01 1.15216565e+00
2.89298505e-01 4.58502084e-01 -2.90151626e-01 1.48188066e+00
5.91254592e-01 8.56888831e-01 6.02164984e-01 -3.29848439e-01
7.06467927e-01 3.02832395e-01 -8.09477344e-02 -5.34766972e-01
-1.95489109e-01 -3.39420199e-01 -9.90900278e-01 -5.28364480e-01
4.39036667e-01 -2.71879226e-01 -9.72620428e-01 1.42698717e+00
-3.95376056e-01 -2.61489987e-01 3.69763583e-01 6.90735459e-01
1.20698917e+00 7.34146476e-01 -1.22605063e-01 -3.84967476e-02
1.18044853e+00 -9.78092074e-01 -8.43111873e-01 -1.97194308e-01
8.37519944e-01 -1.27435291e+00 1.06919146e+00 4.74377662e-01
-8.50561082e-01 -5.39084733e-01 -8.74564767e-01 -1.52795315e-01
-5.22607327e-01 6.96477413e-01 4.92314070e-01 1.13690972e+00
-1.28881383e+00 -2.05892220e-01 -2.23596171e-01 -2.41447359e-01
-7.02131838e-02 5.29294670e-01 -5.65481544e-01 -2.43448123e-01
-1.39358103e+00 6.33556128e-01 1.11137003e-01 3.14812094e-01
-6.82495534e-01 -3.90739858e-01 -7.53721952e-01 -3.09240907e-01
-6.68784082e-02 1.98643714e-01 9.11980093e-01 -7.81655371e-01
-1.66907299e+00 1.41687763e+00 -2.83744693e-01 -1.91543669e-01
6.07310414e-01 -2.16927126e-01 -1.05800211e+00 -6.31703496e-01
-8.32137018e-02 4.78286535e-01 3.09360921e-01 -9.74030316e-01
-6.77215636e-01 -3.22543740e-01 -5.33892453e-01 3.20804685e-01
-2.96878099e-01 6.99228346e-01 -1.14009917e+00 -7.45563447e-01
1.29316464e-01 -7.33702481e-01 6.45687282e-02 -9.10566151e-01
-7.27712214e-01 -5.10999858e-01 2.57379025e-01 -1.17921591e+00
1.49067128e+00 -2.61828828e+00 -1.81636676e-01 3.54272038e-01
-7.24441260e-02 3.15765470e-01 -4.49093282e-01 1.91811502e-01
-8.57495591e-02 2.53784567e-01 1.94370791e-01 -5.39973557e-01
1.22851744e-01 -3.23631108e-01 -4.63889956e-01 3.82246464e-01
-2.41428137e-01 6.95293069e-01 -4.88243282e-01 -2.58791000e-02
3.65114212e-02 6.09078050e-01 -2.09920555e-01 2.66924500e-01
5.24512827e-01 1.94946185e-01 -2.43877564e-02 1.02649045e+00
7.90853262e-01 2.58743405e-01 1.12523414e-01 -1.27771884e-01
-2.78236002e-01 6.41446650e-01 -1.45045185e+00 1.09787834e+00
-2.89659560e-01 8.67643297e-01 3.15197438e-01 -5.64511657e-01
1.43249798e+00 4.92768586e-01 4.37333286e-02 -8.62672627e-01
-1.16796084e-01 6.68275297e-01 -3.94832902e-02 -4.16900218e-03
7.38291860e-01 4.86396104e-01 -4.00640607e-01 1.30677477e-01
-5.83948987e-03 2.29681656e-01 -1.26256853e-01 1.06964730e-01
5.97979724e-01 -6.35918856e-01 2.68263549e-01 -4.87453431e-01
1.07565224e+00 -4.76495087e-01 4.75251615e-01 1.11581898e+00
-5.11403024e-01 5.94349623e-01 -1.30847143e-02 -3.23752493e-01
-6.71475768e-01 -1.02310753e+00 -2.04941615e-01 1.26092255e+00
-8.83858129e-02 -2.83039808e-01 -6.90867543e-01 -4.68849897e-01
7.57650658e-02 5.26732564e-01 -6.47428781e-02 1.73332766e-02
-5.86499929e-01 -9.06656265e-01 1.31759095e+00 3.80476952e-01
7.67358243e-01 -8.35646868e-01 7.28068769e-01 4.42872196e-02
-4.56759840e-01 -1.29567218e+00 -1.05496109e+00 2.02750899e-02
-1.37287423e-01 -5.45334458e-01 -9.35510159e-01 -1.09902382e+00
3.65069449e-01 3.52110356e-01 1.01962209e+00 -5.54937944e-02
-1.12904701e-02 4.52398568e-01 -2.58860350e-01 -5.54267466e-02
-5.81583202e-01 4.51718330e-01 5.90764761e-01 3.13371688e-01
7.91783035e-01 2.75836170e-01 -1.39546767e-01 5.70785761e-01
-1.13717861e-01 -2.42601737e-01 4.21254277e-01 1.00580895e+00
5.85073769e-01 -3.11206937e-01 6.60615444e-01 -6.28192604e-01
9.66246545e-01 5.09670861e-02 -6.95250332e-01 7.47063100e-01
-5.86384356e-01 -1.24553747e-01 4.68393058e-01 -6.25021458e-01
-8.95664513e-01 1.92902505e-01 -5.69978595e-01 3.49624492e-02
-1.50845349e-01 2.27965727e-01 -5.50547123e-01 -2.16556102e-01
4.49975342e-01 8.50367188e-01 -3.96779865e-01 -8.77171576e-01
1.42544866e-01 1.62888122e+00 4.76914823e-01 -5.55720747e-01
4.21004385e-01 -1.30571648e-01 -8.28224957e-01 -1.45405567e+00
-2.83978164e-01 -9.41105962e-01 -5.65985143e-01 -1.46562442e-01
4.56573397e-01 -1.55550361e+00 -7.91106820e-01 1.15084147e+00
-1.18065906e+00 -2.23037973e-01 2.39559710e-01 6.81682885e-01
-1.67974061e-03 2.90196419e-01 -6.67048156e-01 -1.23231649e+00
-6.57366097e-01 -1.51367533e+00 1.16156530e+00 -1.58488259e-01
7.01478124e-02 -7.09359765e-01 -2.17734203e-01 6.07455075e-01
5.55405974e-01 -5.93210161e-01 3.93818110e-01 -9.34675395e-01
-3.15431863e-01 -2.05808282e-01 -3.46622437e-01 5.95839679e-01
6.26574010e-02 -1.67913869e-01 -1.25468898e+00 -6.66617274e-01
-5.24797082e-01 -3.12193036e-01 8.50007236e-01 8.14046338e-02
1.01636457e+00 9.42101181e-02 -1.57025933e-01 6.42234087e-01
9.27507699e-01 5.86245596e-01 7.57235050e-01 1.32850662e-01
8.93649220e-01 4.71851140e-01 1.95103735e-01 2.22379327e-01
4.64932889e-01 1.08388782e+00 -5.35055757e-01 -2.28713691e-01
-2.17995465e-01 -2.49035403e-01 9.64945853e-01 1.50840068e+00
1.19720437e-01 -2.83419847e-01 -1.33645952e+00 4.63142484e-01
-1.33591533e+00 -6.82241440e-01 -1.36786506e-01 2.17668796e+00
8.02417517e-01 -1.00891560e-01 2.80384511e-01 1.65999621e-01
8.97312820e-01 -2.38094926e-02 -3.23221058e-01 -4.49593455e-01
-8.49014282e-01 -2.78358966e-01 6.51495755e-01 9.79597628e-01
-1.13164818e+00 1.47037768e+00 7.68159056e+00 1.14668632e+00
-1.27863991e+00 -4.03055288e-02 9.01147544e-01 3.70263785e-01
-6.61411509e-03 -4.28005010e-01 -1.74502146e+00 2.90787250e-01
8.98065269e-01 -2.35582650e-01 8.60606134e-01 8.87210190e-01
-9.03719440e-02 2.64198273e-01 -9.14066672e-01 1.57748175e+00
4.00311619e-01 -7.79038668e-01 3.70004535e-01 3.58227454e-02
4.22093332e-01 4.31025982e-01 2.28681475e-01 5.32068849e-01
1.33754030e-01 -1.33649445e+00 7.44663179e-01 3.78454328e-01
1.25797415e+00 -8.35768282e-01 7.93818414e-01 1.48224026e-01
-1.31543243e+00 2.78612942e-01 -5.15045747e-02 2.79552251e-01
-9.27493274e-02 3.19278449e-01 -7.51301348e-01 2.85017043e-01
5.99404514e-01 5.18040001e-01 -7.33197749e-01 7.39237010e-01
7.62303919e-02 8.54756534e-01 -4.31969851e-01 -3.52781862e-01
-1.08062938e-01 -7.78970425e-04 6.06771290e-01 1.81831360e+00
1.13138154e-01 -4.04445857e-01 2.01935351e-01 3.16253871e-01
-3.47940683e-01 4.89221811e-01 -6.10633135e-01 -6.79035261e-02
7.15667546e-01 9.68789577e-01 -2.49508113e-01 -2.58786500e-01
-4.59451646e-01 1.04286313e+00 1.14992827e-01 4.93541151e-01
-4.06509757e-01 -5.75571060e-01 8.07622612e-01 -3.41687709e-01
-1.30073994e-01 -2.84932911e-01 -2.38854244e-01 -1.48335075e+00
2.64377911e-02 -1.37818468e+00 4.65434551e-01 -2.58388609e-01
-1.35372102e+00 9.79135573e-01 -5.20501196e-01 -8.20771992e-01
-1.66714579e-01 -9.57554519e-01 -2.00599786e-02 1.39044690e+00
-1.47603190e+00 -9.52121675e-01 1.26208037e-01 6.37557030e-01
5.90167880e-01 -1.12801564e+00 1.12173939e+00 8.89198303e-01
-9.30712521e-01 1.68104231e+00 4.44550902e-01 6.97525978e-01
8.97170246e-01 -1.02967286e+00 6.18158162e-01 9.90267098e-01
1.60506532e-01 6.88290000e-01 6.99971467e-02 -1.32593453e-01
-1.66640878e+00 -1.04876602e+00 1.33639443e+00 -4.79334056e-01
4.76305813e-01 -1.06071758e+00 -7.33628035e-01 5.24486065e-01
-1.49736926e-01 -1.99060932e-01 7.60780573e-01 5.04399002e-01
-7.36683905e-01 -3.21049333e-01 -8.49413991e-01 7.36674607e-01
9.65316772e-01 -1.16079175e+00 -1.37810279e-02 2.31075376e-01
5.54540157e-01 -2.84598082e-01 -8.50545645e-01 1.91784054e-01
7.63358116e-01 -3.45886350e-01 1.04819751e+00 -4.82799947e-01
-5.31099141e-01 -2.12792799e-01 -7.77154148e-01 -1.07887435e+00
-2.67274439e-01 -7.03373492e-01 2.76797891e-01 1.60249317e+00
9.14647758e-01 -7.76747108e-01 2.35641345e-01 3.42811465e-01
1.15105428e-01 -4.04018015e-01 -1.08805072e+00 -9.49770808e-01
1.13293946e-01 -7.73508489e-01 4.11046565e-01 1.06198442e+00
-2.57476956e-01 5.98505259e-01 -5.98914564e-01 1.58547223e-01
3.12496692e-01 -4.61349517e-01 6.85619295e-01 -6.50328159e-01
-1.18463084e-01 -5.60480893e-01 -2.76546091e-01 -1.30587149e+00
1.79939359e-01 -1.09087110e+00 3.84334289e-03 -9.02422249e-01
1.12499170e-01 -3.62627268e-01 -3.95317435e-01 3.14268798e-01
-1.30245551e-01 2.68269360e-01 1.91951960e-01 2.63532281e-01
-8.04331899e-01 3.30634981e-01 6.55214548e-01 -4.38515574e-01
-4.10668880e-01 2.04140786e-02 -8.42182457e-01 3.86330932e-01
7.28403687e-01 1.30818382e-01 1.70457408e-01 -6.20530248e-01
-1.78681880e-01 -3.41010302e-01 -2.67913520e-01 -6.05999351e-01
3.45401168e-01 2.87356138e-01 3.02602053e-01 -6.68495953e-01
3.13579142e-01 -3.30130637e-01 -2.65073210e-01 2.80323833e-01
-3.62728775e-01 3.38696033e-01 4.18319106e-01 -2.21568167e-01
-5.10716975e-01 3.56867105e-01 9.10866380e-01 3.05212229e-01
-9.16441679e-01 -1.48704345e-03 -6.39482200e-01 2.00278595e-01
5.64288676e-01 -5.78480996e-02 -1.91518337e-01 -3.20091784e-01
-6.37104392e-01 3.39299321e-01 -2.13307753e-01 8.20323884e-01
6.61828995e-01 -1.52217150e+00 -1.10832405e+00 5.13176024e-01
3.77342731e-01 -6.01380289e-01 3.97100411e-02 5.71481705e-01
-3.05889636e-01 9.14915442e-01 3.72013867e-01 -5.18724382e-01
-1.60430217e+00 -3.42604443e-02 8.17545414e-01 -3.46749753e-01
-7.95456767e-02 9.58257973e-01 1.33487999e-01 -9.82956469e-01
8.44531298e-01 2.33835712e-01 -3.64415616e-01 1.33225113e-01
1.00975227e+00 3.83552194e-01 3.40257198e-01 -1.08789885e+00
-9.19821918e-01 4.64315563e-01 -5.30414283e-01 -3.41226876e-01
8.94023955e-01 -2.58134395e-01 -1.69648722e-01 6.10545456e-01
1.30698049e+00 2.58426189e-01 -4.11613375e-01 -4.90976274e-01
2.86357701e-01 -2.15350211e-01 9.76733267e-02 -1.15964925e+00
-7.29568124e-01 8.48788321e-01 9.18191314e-01 1.19366206e-01
1.11653864e+00 -1.20668516e-01 4.92858589e-01 5.14530420e-01
4.26168114e-01 -1.36586845e+00 -4.28786635e-01 1.23190856e+00
1.08703887e+00 -1.45564830e+00 -5.54951429e-01 -4.63529736e-01
-5.05086124e-01 8.02190006e-01 4.70139712e-01 4.68441695e-01
7.12743163e-01 5.48991621e-01 5.64278364e-01 4.61829782e-01
-6.81633055e-01 -2.66289145e-01 7.50172496e-01 5.54070294e-01
9.58137751e-01 4.22460049e-01 -2.67486662e-01 1.03955710e+00
-1.12952583e-01 -5.33061445e-01 -1.22015387e-01 3.35423529e-01
-3.02101579e-02 -1.12265801e+00 -4.90009606e-01 1.89561650e-01
-5.55983484e-01 -4.20586228e-01 -8.61720741e-01 6.18273854e-01
-3.04839909e-01 1.10306597e+00 -2.00075164e-01 -8.63217294e-01
7.23835289e-01 2.43524894e-01 9.29961652e-02 -4.14162397e-01
-6.57120109e-01 1.79638237e-01 4.34240520e-01 -3.23882014e-01
1.13395967e-01 -6.32057846e-01 -8.02389264e-01 -5.73458195e-01
-3.59231204e-01 1.16062596e-01 7.17461765e-01 6.65283442e-01
4.51358199e-01 2.75478989e-01 8.90846491e-01 -2.20595077e-01
-7.11716712e-01 -1.09366310e+00 -6.73676133e-01 3.28070492e-01
2.35265523e-01 -1.47312641e-01 -4.84322548e-01 -1.76753566e-01] | [14.200414657592773, 6.656313419342041] |
c318c88c-cef5-4c11-ab50-76e5017a38ec | blcufight-at-semeval-2021-task-10-novel | null | null | https://aclanthology.org/2021.semeval-1.43 | https://aclanthology.org/2021.semeval-1.43.pdf | BLCUFIGHT at SemEval-2021 Task 10: Novel Unsupervised Frameworks For Source-Free Domain Adaptation | Domain adaptation assumes that samples from source and target domains are freely accessible during a training phase. However, such assumption is rarely plausible in the real-world and may causes data-privacy issues, especially when the label of the source domain can be a sensitive attribute as an identifier. SemEval-2021 task 10 focuses on these issues. We participate in the task and propose novel frameworks based on self-training method. In our systems, two different frameworks are designed to solve text classification and sequence labeling. These approaches are tested to be effective which ranks the third among all system in subtask A, and ranks the first among all system in subtask B. | ['Pengyuan Liu', 'Yixiang Liu', 'Yi Wu', 'Weikang Wang'] | 2021-08-01 | null | null | null | semeval-2021 | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 2.61760622e-01 1.23649396e-01 -5.56829691e-01 -7.65919924e-01
-5.87187946e-01 -6.45868421e-01 6.43922746e-01 -5.01090661e-02
-7.47593582e-01 1.41883409e+00 -2.24060535e-01 -1.68557495e-01
1.47495553e-01 -3.92970234e-01 -6.33317649e-01 -5.06522119e-01
3.21153671e-01 8.77350748e-01 2.33234286e-01 -1.23043969e-01
1.53352514e-01 -1.38518633e-02 -1.12795472e+00 5.48815072e-01
1.15085018e+00 1.05387831e+00 -1.42488047e-01 3.29811908e-02
-4.28356856e-01 4.46035475e-01 -8.22816670e-01 -7.28431046e-01
4.72091705e-01 -3.37935388e-01 -1.12907493e+00 -1.66435987e-01
2.65787095e-01 -2.42729694e-01 -1.09517565e-02 1.28550386e+00
4.86542881e-01 5.47951013e-02 8.20502460e-01 -1.54337442e+00
-8.65572274e-01 6.20272160e-01 -3.92725319e-01 -9.96398777e-02
4.26015645e-01 -2.84587860e-01 7.99024820e-01 -6.75616980e-01
7.77948797e-01 8.72640371e-01 4.88407433e-01 1.04223192e+00
-1.02882290e+00 -1.12058771e+00 2.40239218e-01 2.64224350e-01
-1.32705235e+00 -6.72326505e-01 5.90035021e-01 -4.01011765e-01
6.08752310e-01 1.95553109e-01 2.45247763e-02 1.69575131e+00
-2.06354246e-01 9.52874959e-01 1.28874004e+00 -4.00853306e-01
5.79259038e-01 1.00908327e+00 3.69901001e-01 -6.59338161e-02
3.80472600e-01 1.19671714e-03 -5.10576189e-01 -5.47103822e-01
1.28907412e-01 -5.40962964e-02 -3.78418446e-01 -6.64297998e-01
-1.08054483e+00 8.58713508e-01 2.66273343e-03 2.56548256e-01
-1.49358809e-02 -7.84375548e-01 6.85350955e-01 5.58586776e-01
6.02789342e-01 4.66547161e-01 -8.46741259e-01 1.97802871e-01
-8.01368833e-01 3.44365239e-01 1.02625358e+00 1.34689558e+00
4.28602219e-01 -4.67289597e-01 -4.13220376e-02 9.03209388e-01
1.03635103e-01 1.72198251e-01 9.58323896e-01 -4.83968943e-01
7.24016905e-01 5.66806972e-01 4.92778271e-01 -3.33857864e-01
-2.61557341e-01 -1.04258195e-01 -8.51966202e-01 -9.73577201e-02
6.34244263e-01 -4.41816330e-01 -8.34348440e-01 1.68010974e+00
3.72105688e-01 1.79615617e-01 4.10972178e-01 7.39144862e-01
9.79309082e-01 4.51310784e-01 4.51264471e-01 -3.96184564e-01
1.46630657e+00 -7.98788548e-01 -8.71888816e-01 -3.18918884e-01
7.24098325e-01 -4.09780681e-01 1.04353786e+00 2.98750907e-01
-4.75944400e-01 -3.13101292e-01 -1.09528649e+00 3.22804451e-02
-7.07831323e-01 1.25440136e-01 4.53660011e-01 9.76257980e-01
-5.96080720e-01 3.51648569e-01 -2.19437420e-01 -6.97021246e-01
5.16106665e-01 3.63977522e-01 -7.91847587e-01 -7.93855712e-02
-1.84881818e+00 8.31531227e-01 7.18738496e-01 -4.13797438e-01
-5.63402653e-01 -4.31523710e-01 -4.98561919e-01 -1.54286653e-01
3.43916804e-01 -4.56599802e-01 1.28535712e+00 -9.38136697e-01
-1.45661795e+00 1.19141710e+00 -1.12511955e-01 -6.91452146e-01
8.81606460e-01 -1.60238639e-01 -9.09230113e-01 -2.90065795e-01
2.27611512e-01 3.07422072e-01 7.96911418e-01 -1.02030516e+00
-7.75192142e-01 -4.86563623e-01 -6.30055219e-02 3.12402695e-01
-7.42082179e-01 5.86234331e-02 4.60251495e-02 -4.30934697e-01
-2.15177953e-01 -8.63615930e-01 -1.84168547e-01 -1.43721014e-01
-5.22192061e-01 -3.83917749e-01 9.26587939e-01 -4.59091008e-01
1.05599284e+00 -2.28370905e+00 -5.64853013e-01 6.36988729e-02
-1.23986855e-01 6.76936924e-01 2.33778700e-01 2.72582710e-01
-4.04620022e-01 2.10726604e-01 -1.84177935e-01 -3.35320860e-01
9.54938605e-02 1.32214874e-01 -5.22696674e-01 3.78781855e-01
-2.20715508e-01 4.11833704e-01 -6.80895805e-01 -5.28356910e-01
-1.95914224e-01 -9.60115641e-02 -1.41883895e-01 3.89367700e-01
-3.19109172e-01 4.28627133e-01 -8.51126134e-01 5.07034063e-01
1.07540417e+00 -2.31038675e-01 2.75092781e-01 1.88772202e-01
3.00723910e-01 3.34648877e-01 -1.17223287e+00 1.57704508e+00
3.75570506e-02 1.77343741e-01 4.34385613e-02 -1.16992617e+00
1.20087516e+00 7.48890102e-01 3.41585189e-01 -5.77374041e-01
-2.75950637e-02 1.99436098e-01 -2.75654763e-01 -5.66319466e-01
4.90703881e-01 -2.36133024e-01 -1.11353219e-01 4.32942629e-01
2.53480580e-03 2.66969711e-01 -1.53837889e-01 -9.38497670e-03
8.76016557e-01 1.46663755e-01 6.90930843e-01 -2.35706240e-01
7.97929168e-01 -5.63154295e-02 7.85784185e-01 6.11879945e-01
-6.74461842e-01 5.16925931e-01 4.49795276e-01 -4.89211470e-01
-8.52050126e-01 -6.80143714e-01 -2.60183424e-01 1.36673522e+00
5.17211966e-02 -1.30690277e-01 -7.70119607e-01 -1.49159563e+00
1.28058597e-01 9.15515959e-01 -6.94510281e-01 -2.31752262e-01
-5.64980097e-02 -5.74702501e-01 6.33109927e-01 2.67490715e-01
6.89392507e-01 -9.46460009e-01 -2.47200206e-01 1.87793225e-01
-3.42334867e-01 -1.24166429e+00 -4.71354067e-01 4.48778421e-01
-6.35918856e-01 -1.03995275e+00 -5.89137316e-01 -8.55481625e-01
6.33893311e-01 -2.59997156e-02 9.81106877e-01 -4.09446150e-01
2.15601966e-01 -1.41651347e-01 -5.08794308e-01 -7.54035056e-01
-3.27332109e-01 4.53520000e-01 6.91604763e-02 9.83817950e-02
1.13990974e+00 -2.74765134e-01 -2.55035102e-01 5.62863588e-01
-6.79347277e-01 -4.06145930e-01 2.14726999e-01 8.53278339e-01
3.13119739e-01 4.72095199e-02 8.95543396e-01 -1.80946910e+00
8.12124014e-01 -7.65513659e-01 -4.78047162e-01 5.50615668e-01
-7.53967345e-01 -1.21417969e-01 1.12902200e+00 -5.89022815e-01
-1.23307478e+00 4.02214855e-01 -1.05198748e-01 -1.46829441e-01
-6.98564589e-01 1.37894377e-01 -8.13413143e-01 1.21534549e-01
1.02729046e+00 2.30775535e-01 -1.39800027e-01 -6.19073868e-01
3.38147059e-02 1.17354023e+00 1.85442150e-01 -5.62585831e-01
7.32484460e-01 1.61000609e-01 -4.62733984e-01 -5.73292494e-01
-8.39501679e-01 -5.47246695e-01 -6.58274293e-01 4.18102652e-01
5.92239022e-01 -1.07533288e+00 -5.61426997e-01 4.25755799e-01
-1.03236318e+00 7.94350207e-02 -2.73114979e-01 4.42764789e-01
-3.56583834e-01 3.02430272e-01 -1.26471356e-01 -7.30872750e-01
-3.01121771e-01 -9.13605452e-01 5.95547736e-01 1.07506759e-01
-4.15022552e-01 -8.41721177e-01 -8.35886300e-02 4.49038535e-01
3.23816836e-01 -6.28999202e-03 8.56594980e-01 -1.75628662e+00
-7.70937204e-02 -4.05041486e-01 3.02610081e-02 3.35564882e-01
1.38265476e-01 -4.88921195e-01 -1.21431482e+00 -3.10667992e-01
2.86661714e-01 -5.57223082e-01 5.83536029e-01 4.45775464e-02
1.25000215e+00 -4.34220105e-01 -6.09186649e-01 4.91164833e-01
1.09541297e+00 4.63663906e-01 5.25440753e-01 5.03239632e-01
3.42482030e-01 8.26199532e-01 9.44492519e-01 4.79674965e-01
4.37794060e-01 6.87377095e-01 1.08685657e-01 1.49831399e-01
4.87934351e-01 -4.26002294e-01 1.57458767e-01 6.45104349e-02
4.35187191e-01 -4.05723512e-01 -7.53331840e-01 5.21091163e-01
-1.85594070e+00 -9.78946984e-01 -6.83973879e-02 2.41984344e+00
1.10686135e+00 1.67934850e-01 2.82900393e-01 -1.39421314e-01
9.35813248e-01 -8.20264965e-02 -9.04356778e-01 -3.23489815e-01
-1.00688651e-01 -5.12032211e-02 6.44062161e-01 1.20634019e-01
-1.45630848e+00 8.99135590e-01 6.66132307e+00 5.87840497e-01
-8.37365687e-01 1.98767483e-01 7.40794301e-01 3.76624838e-02
-1.97319150e-01 4.32580337e-02 -1.07141590e+00 8.47539961e-01
1.12585890e+00 -4.84946132e-01 -3.22948508e-02 1.18925405e+00
-4.75474954e-01 2.09341377e-01 -1.47787189e+00 7.80653000e-01
-9.92404744e-02 -8.17030609e-01 -2.72850543e-02 8.46414715e-02
5.63623428e-01 -1.68416038e-01 1.44607678e-01 5.34107745e-01
5.64661860e-01 -1.06353807e+00 3.27067167e-01 3.10154036e-02
8.68677974e-01 -6.88114047e-01 9.08154428e-01 8.23020041e-01
-6.05164945e-01 -1.27342820e-01 -7.14567065e-01 7.86974356e-02
-1.24892704e-01 5.91540217e-01 -1.06454265e+00 5.19824266e-01
7.75889516e-01 6.36226475e-01 -3.22823375e-01 9.13647950e-01
-1.18453585e-01 6.64824784e-01 -1.66968927e-01 -1.89894304e-01
-2.00059116e-01 -7.29381442e-02 3.49784970e-01 1.01636815e+00
1.22206643e-01 3.16448286e-02 2.66755193e-01 4.71612841e-01
-4.43288296e-01 4.16165084e-01 -8.46824765e-01 -2.35930420e-02
7.20436096e-01 9.09370422e-01 -2.07734376e-01 -3.92038941e-01
-4.96996284e-01 1.06178737e+00 3.04691881e-01 4.67292637e-01
-5.61216712e-01 -2.91565120e-01 7.24655747e-01 4.98712063e-02
3.36910747e-02 3.08573931e-01 -2.81641424e-01 -1.21004999e+00
8.75775889e-02 -1.06863451e+00 8.23317587e-01 -3.19323450e-01
-1.67466486e+00 6.97555900e-01 -1.07574545e-01 -1.39359677e+00
-3.62601161e-01 -5.25225699e-01 -1.69235066e-01 9.31219876e-01
-1.49273109e+00 -1.00014567e+00 -1.15063585e-01 1.02283061e+00
2.50699252e-01 -7.00883925e-01 1.32304800e+00 4.61489022e-01
-5.39719224e-01 1.23558366e+00 6.36294782e-01 2.93798774e-01
1.39585531e+00 -1.16697562e+00 5.09268641e-01 5.23318172e-01
-1.37646407e-01 7.60032415e-01 6.74858034e-01 -5.90202928e-01
-8.22777450e-01 -1.07971370e+00 1.13531375e+00 -5.26256382e-01
4.03592378e-01 -7.54136026e-01 -1.29903591e+00 9.19254303e-01
1.98417753e-01 1.84155166e-01 1.18229306e+00 2.02801839e-01
-5.02774179e-01 1.06148478e-02 -1.67459428e+00 1.23966388e-01
9.43639040e-01 -5.36133051e-01 -6.68582559e-01 6.20937645e-01
6.76759303e-01 -4.96645570e-01 -8.43310893e-01 2.01772571e-01
2.19640791e-01 -6.16027951e-01 8.45857739e-01 -9.27211165e-01
1.73521802e-01 -1.43214539e-01 -4.81057167e-02 -1.26732421e+00
-1.56142443e-01 -5.08569658e-01 -1.82909745e-04 1.72327816e+00
5.68632782e-01 -1.11162579e+00 1.20920897e+00 1.11251199e+00
5.12001157e-01 -4.15569246e-02 -1.05058515e+00 -1.05341971e+00
3.26457024e-01 1.81184679e-01 1.03449714e+00 1.62372172e+00
1.43280610e-01 5.51517129e-01 -4.76289004e-01 -9.59651321e-02
5.29033363e-01 -2.24172533e-03 7.43365347e-01 -1.57737434e+00
2.27322374e-02 3.31965322e-03 -5.69724068e-02 -9.30543244e-01
5.82498908e-01 -8.41936052e-01 -8.31720456e-02 -1.03580344e+00
8.00910816e-02 -6.54746711e-01 -5.28099835e-01 5.81216633e-01
-2.24223807e-01 -2.87080288e-01 -1.45877898e-01 2.63454556e-01
-5.10365188e-01 4.62353945e-01 7.96600580e-01 -5.11404350e-02
-1.06356241e-01 4.51624453e-01 -1.02828932e+00 5.19819856e-01
9.45905387e-01 -6.49052858e-01 -6.33574903e-01 -1.21978559e-01
-2.72221357e-01 -4.37612124e-02 -1.10380724e-01 -7.50081360e-01
2.85513729e-01 -4.51471031e-01 4.43672150e-01 -4.75553632e-01
1.21508472e-01 -1.44723356e+00 1.38996869e-01 1.20158240e-01
-7.67964602e-01 -4.80688184e-01 -5.83269186e-02 5.40752769e-01
-2.25666583e-01 -5.13906360e-01 7.92683482e-01 -2.37937763e-01
-8.69510233e-01 4.95517105e-01 1.68584317e-01 3.05408120e-01
1.30043149e+00 -3.63268703e-01 -4.56771642e-01 -2.38049194e-01
-4.69129652e-01 2.89996326e-01 5.57451129e-01 4.85162884e-01
1.95954487e-01 -1.23172665e+00 -7.44414806e-01 3.38354588e-01
5.60379267e-01 -1.34379491e-01 1.55839413e-01 2.39615485e-01
1.29270524e-01 4.18069869e-01 -3.93435061e-01 -2.06754118e-01
-1.30212355e+00 7.30196834e-01 1.00926101e-01 -4.13377285e-01
-2.10810915e-01 1.02778196e+00 3.34439844e-01 -8.11756551e-01
4.81764376e-01 2.11383030e-01 -4.99091536e-01 6.60007745e-02
6.67499959e-01 1.32877842e-01 9.80596095e-02 -5.48310399e-01
-5.91194153e-01 -2.48733088e-02 -7.22771943e-01 1.70017987e-01
1.09035254e+00 -1.71549320e-01 7.94393197e-02 2.69469142e-01
1.34379327e+00 -2.96202511e-01 -1.03598607e+00 -6.59215868e-01
2.76958138e-01 -5.49216568e-01 -5.00098884e-01 -1.16712356e+00
-6.32717967e-01 7.69501567e-01 7.57627666e-01 2.16885358e-01
1.00424361e+00 -4.30682451e-01 8.38175654e-01 3.85771573e-01
5.56467950e-01 -1.41864693e+00 -3.39272052e-01 6.03887320e-01
4.91284490e-01 -1.57320893e+00 -8.32866353e-04 -4.02997077e-01
-1.09263730e+00 7.62108505e-01 1.05242181e+00 2.91163206e-01
6.12638891e-01 -2.72785407e-02 1.72259465e-01 3.84714842e-01
-6.77442610e-01 3.04532915e-01 1.61197573e-01 1.04603755e+00
5.67922890e-01 1.67250589e-01 -4.71376687e-01 9.17444646e-01
-8.41339305e-02 1.92888334e-01 2.68861502e-01 9.33952868e-01
-1.79875016e-01 -1.50571275e+00 -1.89022720e-01 3.98086071e-01
-6.15049660e-01 1.54263508e-02 -7.12263107e-01 6.05538070e-01
9.40005332e-02 8.68672729e-01 -2.85467803e-01 -4.25788313e-01
4.02065158e-01 5.25554001e-01 -2.00463459e-01 -6.13667607e-01
-6.58162892e-01 -5.29558897e-01 1.95788547e-01 -3.04907441e-01
-1.94288373e-01 -7.94419169e-01 -9.78615820e-01 -2.20814317e-01
-2.11499050e-01 5.12620270e-01 5.31557202e-01 8.25825036e-01
3.93000752e-01 -1.27192914e-01 7.28963852e-01 1.42757416e-01
-9.61529851e-01 -9.41659272e-01 -8.04958820e-01 7.26067066e-01
3.11708659e-01 -5.93998492e-01 -1.83861122e-01 2.43280809e-02] | [10.374521255493164, 3.1968319416046143] |
fab0c4bf-f980-4d81-ab63-8a409e07e2f4 | bayesian-fusion-for-infrared-and-visible | 2005.05839 | null | https://arxiv.org/abs/2005.05839v1 | https://arxiv.org/pdf/2005.05839v1.pdf | Bayesian Fusion for Infrared and Visible Images | Infrared and visible image fusion has been a hot issue in image fusion. In this task, a fused image containing both the gradient and detailed texture information of visible images as well as the thermal radiation and highlighting targets of infrared images is expected to be obtained. In this paper, a novel Bayesian fusion model is established for infrared and visible images. In our model, the image fusion task is cast into a regression problem. To measure the variable uncertainty, we formulate the model in a hierarchical Bayesian manner. Aiming at making the fused image satisfy human visual system, the model incorporates the total-variation(TV) penalty. Subsequently, the model is efficiently inferred by the expectation-maximization(EM) algorithm. We test our algorithm on TNO and NIR image fusion datasets with several state-of-the-art approaches. Compared with the previous methods, the novel model can generate better fused images with high-light targets and rich texture details, which can improve the reliability of the target automatic detection and recognition system. | ['Chun-Xia Zhang', 'Junmin Liu', 'Jiangshe Zhang', 'Zixiang Zhao', 'Shuang Xu'] | 2020-05-12 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 6.17680490e-01 -4.42726016e-01 2.55780160e-01 -4.16565239e-01
-8.40652525e-01 -1.90596685e-01 4.37656432e-01 -3.39272350e-01
-1.34587049e-01 6.63127303e-01 -9.79707092e-02 -8.52994174e-02
-1.78887025e-01 -6.91443920e-01 -2.52500683e-01 -1.37370670e+00
6.91474438e-01 -3.90418768e-02 5.35887070e-02 -3.79588231e-02
1.41752914e-01 4.01032716e-01 -1.62436199e+00 2.18570232e-01
1.22282398e+00 1.37419343e+00 7.48657882e-01 4.73470539e-01
7.13014528e-02 8.41496468e-01 -2.74364680e-01 -9.99964029e-02
3.15713674e-01 -3.78018051e-01 -2.72799283e-01 4.24217105e-01
1.79663837e-01 -1.90703884e-01 -1.70126319e-01 1.44505990e+00
5.41581511e-01 2.52210498e-01 7.84935772e-01 -8.89531553e-01
-6.57932818e-01 1.82473823e-01 -9.74562705e-01 1.22893415e-02
9.45890024e-02 1.45983040e-01 4.93270338e-01 -8.70737076e-01
1.54557014e-02 1.28647172e+00 1.54977202e-01 -3.14444043e-02
-1.27620018e+00 -5.74604571e-01 3.36808674e-02 2.58481354e-01
-1.29095793e+00 -3.95318002e-01 9.21654224e-01 -4.90507454e-01
1.22858383e-01 5.29593229e-01 4.68973100e-01 7.57194996e-01
6.33404136e-01 6.07393205e-01 1.58575928e+00 -5.63658476e-01
6.10037260e-02 1.03333987e-01 -7.48256892e-02 4.34389263e-01
3.38350147e-01 3.86721790e-01 -4.67163533e-01 -4.33865748e-02
2.90515810e-01 2.82947242e-01 -5.13728797e-01 7.20777959e-02
-8.98318946e-01 4.92911965e-01 3.98713827e-01 1.72577515e-01
-5.99513054e-01 -2.06321180e-01 -3.54726523e-01 -1.56089351e-01
8.79588246e-01 -1.83401287e-01 -1.14877388e-01 6.39143229e-01
-6.80535376e-01 -2.22685754e-01 3.02864611e-01 3.99812222e-01
9.73475695e-01 2.69149765e-02 -5.87436974e-01 9.05955613e-01
9.66962516e-01 1.11233759e+00 -1.38457432e-01 -8.12458336e-01
1.77897632e-01 4.67471331e-01 2.69832879e-01 -1.03779507e+00
-3.56136984e-03 -3.91201764e-01 -1.07207990e+00 7.39215910e-01
1.02968328e-01 -8.99850652e-02 -1.19348717e+00 1.33897138e+00
6.38958216e-01 1.86249614e-01 1.86793610e-01 1.04490840e+00
8.58864903e-01 8.85256052e-01 -1.12684503e-01 -8.00209761e-01
1.45129573e+00 -6.21407270e-01 -9.95528698e-01 -4.23056394e-01
-5.21088317e-02 -1.29642129e+00 2.17272624e-01 4.85622853e-01
-7.27362037e-01 -8.01868260e-01 -9.14902091e-01 2.02219263e-01
8.40690583e-02 4.38469499e-01 3.70786369e-01 6.85021996e-01
-7.51807272e-01 2.37584203e-01 -4.74253327e-01 -1.49453372e-01
1.74943537e-01 8.97958130e-02 -1.84736341e-01 -4.64997411e-01
-9.97429371e-01 9.54651058e-01 5.49346745e-01 7.45476127e-01
-7.25288332e-01 -3.67080986e-01 -7.95280337e-01 -3.09224784e-01
5.28942347e-01 -8.08483303e-01 8.21342289e-01 -7.79541373e-01
-1.65509188e+00 6.88508928e-01 -4.02176857e-01 1.79188862e-01
2.15042308e-01 -1.25108153e-01 -3.76812190e-01 8.99323151e-02
-1.76621988e-01 1.57062307e-01 1.19467485e+00 -1.77894616e+00
-8.36922705e-01 -5.79103231e-01 -1.45423427e-01 5.01817524e-01
1.59714103e-01 1.50477782e-01 -3.31162065e-01 -3.38715434e-01
3.76057565e-01 -7.28552699e-01 -1.98301777e-01 1.54481654e-03
-3.21132004e-01 -8.80390406e-02 8.05428565e-01 -1.02433300e+00
7.74689198e-01 -2.38899469e+00 1.37145251e-01 2.46614769e-01
2.37615675e-01 1.80721357e-01 2.11042866e-01 -5.79424538e-02
8.44651833e-03 -2.46351063e-01 -3.58340740e-01 -2.87424296e-01
-2.35776916e-01 4.32606135e-03 -1.85994282e-01 7.82228827e-01
-2.75456440e-02 6.23162210e-01 -5.61714113e-01 -8.21462274e-01
7.19560564e-01 6.70386195e-01 2.00003102e-01 4.79975700e-01
-1.54705480e-01 8.58833671e-01 -6.07595623e-01 7.91406274e-01
1.23862708e+00 8.07041973e-02 -1.19520955e-01 -6.91435993e-01
-2.75711507e-01 -6.20784998e-01 -1.19138026e+00 1.47570622e+00
-2.46817023e-01 2.46281579e-01 2.84814417e-01 -7.96597779e-01
1.36554623e+00 1.38945729e-01 5.67973793e-01 -8.47222149e-01
2.98809081e-01 -1.79572236e-02 -3.08124751e-01 -5.92483997e-01
3.10124218e-01 -2.73904532e-01 3.00837725e-01 5.59777021e-02
-3.96998584e-01 -3.14965844e-01 -2.62091815e-01 -2.08640933e-01
3.59683871e-01 3.38972926e-01 8.75080824e-02 -8.53519440e-02
6.65832818e-01 -1.34857252e-01 7.08248138e-01 7.79789269e-01
8.39738399e-02 5.00718474e-01 -2.87561417e-01 -8.88506100e-02
-7.32263267e-01 -1.03894436e+00 -2.09323630e-01 7.01155543e-01
5.91311872e-01 2.63484716e-01 -3.19160730e-01 6.15909323e-03
-1.56680897e-01 7.12861776e-01 -3.76124591e-01 -2.27736965e-01
5.54489978e-02 -1.07322228e+00 -1.14643134e-01 7.36913756e-02
8.16005051e-01 -6.96456671e-01 -3.28017890e-01 -1.24951266e-02
-4.49372023e-01 -1.02188051e+00 -1.83797106e-01 -6.44603223e-02
-6.74373627e-01 -9.96236324e-01 -5.81881523e-01 -4.03618574e-01
5.32919824e-01 8.01738501e-01 6.95111036e-01 -1.61694691e-01
-5.19823790e-01 2.40483284e-01 -5.74349105e-01 -5.45965970e-01
-1.96676925e-01 -8.54850829e-01 -3.16578209e-01 7.38875866e-01
2.26881877e-01 -2.87892930e-02 -5.91525495e-01 3.25585514e-01
-9.78234947e-01 4.22414094e-01 8.04495752e-01 8.91988754e-01
7.77235806e-01 4.25900042e-01 -1.75184861e-01 -2.40316287e-01
1.74721479e-01 -6.92196935e-02 -1.06193471e+00 6.92826331e-01
-5.61664283e-01 2.94025876e-02 8.82617906e-02 -1.87148750e-01
-1.98891652e+00 3.87618184e-01 2.49164000e-01 -4.13374364e-01
-1.38817459e-01 6.36167407e-01 -4.12844807e-01 -4.58396763e-01
4.43925738e-01 4.25342917e-01 1.71465978e-01 -5.32569230e-01
2.27869213e-01 9.64016438e-01 5.21595180e-01 -5.53356826e-01
9.24067855e-01 6.51478648e-01 2.44257689e-01 -8.82465720e-01
-1.09349179e+00 -5.95875800e-01 -4.57083315e-01 -6.02991700e-01
1.08977604e+00 -9.92174268e-01 -8.89112234e-01 7.00211644e-01
-1.14235330e+00 8.87880102e-02 1.90862462e-01 9.43090737e-01
-2.62149274e-01 6.93769395e-01 -3.40118498e-01 -1.51294649e+00
-2.74703175e-01 -1.34290659e+00 1.29468751e+00 5.41273773e-01
6.96684897e-01 -7.52332270e-01 7.90173188e-03 7.35917985e-01
2.83102512e-01 3.78909975e-01 4.00542468e-01 2.22124189e-01
-6.72289312e-01 5.41289113e-02 -5.05317867e-01 5.95408857e-01
3.21223557e-01 2.04474866e-01 -1.20702446e+00 -8.45121890e-02
6.41861498e-01 -3.41206342e-02 1.19429064e+00 6.96207762e-01
1.03284872e+00 1.37756273e-01 -2.98959106e-01 7.08183467e-01
1.57295430e+00 4.37794685e-01 9.16702926e-01 1.40475646e-01
7.20081985e-01 7.35805094e-01 1.09122717e+00 5.35645306e-01
1.13889072e-02 6.53075695e-01 6.39879644e-01 -5.15979469e-01
7.19510913e-02 4.11531180e-01 3.22806895e-01 3.66691858e-01
-2.61270940e-01 -3.57696831e-01 -7.43481696e-01 -3.01673048e-04
-2.05891681e+00 -1.07545590e+00 -5.32503247e-01 2.31267095e+00
6.32163227e-01 -4.75984037e-01 -4.98881638e-01 -2.31835246e-02
1.14244390e+00 1.47901937e-01 -3.60207707e-01 3.09405267e-01
-3.55532318e-01 -3.74040268e-02 5.57753086e-01 6.78021491e-01
-1.03502989e+00 5.75356126e-01 6.34426880e+00 1.03154254e+00
-1.04013944e+00 -1.55540034e-02 7.61352539e-01 3.43815446e-01
-2.35725701e-01 2.39465132e-01 -6.33709669e-01 5.31309843e-01
3.49357933e-01 4.64627035e-02 4.56535906e-01 1.00426406e-01
4.15339619e-01 -7.86091685e-01 -5.62083483e-01 1.35218525e+00
2.66028494e-01 -6.65316403e-01 -2.57507712e-01 1.80091038e-01
6.91600919e-01 -2.04951659e-01 2.10359111e-01 -4.21697497e-01
4.63798821e-01 -1.01237631e+00 6.14601254e-01 1.22725499e+00
6.44566298e-01 -5.60555160e-01 9.43190873e-01 3.34123939e-01
-1.23188686e+00 -1.57075927e-01 -4.59875315e-01 1.34259522e-01
2.56241769e-01 1.20149446e+00 -1.38938591e-01 1.38666809e+00
7.73557365e-01 7.25689948e-01 -5.46567678e-01 1.03514051e+00
-3.10059398e-01 2.83170760e-01 -3.81402224e-01 3.65566373e-01
-2.30822593e-01 -8.80152941e-01 3.91095996e-01 7.43602872e-01
4.50249672e-01 4.10492867e-01 5.03904641e-01 9.46802974e-01
4.94110197e-01 -9.89962891e-02 -4.72944111e-01 1.58874929e-01
1.53966457e-01 1.56729782e+00 -5.18555582e-01 -3.06256086e-01
-2.94566035e-01 7.98720241e-01 -1.33624420e-01 5.98399341e-01
-8.24320495e-01 5.42054847e-02 2.31235936e-01 -5.51150501e-01
5.51695861e-02 -9.23308879e-02 -3.93312573e-01 -1.13767326e+00
9.10697654e-02 -6.00381196e-01 3.67342532e-01 -1.30833113e+00
-1.34423888e+00 3.60807806e-01 1.28078923e-01 -1.32758057e+00
3.02393734e-01 -6.30826950e-01 -4.27203000e-01 1.35783756e+00
-1.63842928e+00 -1.34563100e+00 -8.38429630e-01 6.52261376e-01
1.87851861e-01 7.22955093e-02 3.88211221e-01 2.82746926e-02
-6.69531822e-01 -1.45112619e-01 4.66288477e-01 -3.72237921e-01
8.18919718e-01 -6.31965578e-01 -5.20677626e-01 1.18248439e+00
-1.71702594e-01 6.29622415e-02 8.16396832e-01 -8.10511947e-01
-1.32708311e+00 -9.47740853e-01 2.32132167e-01 -1.38876855e-01
3.20359379e-01 1.06489576e-01 -6.98368073e-01 2.76833922e-01
2.61156648e-01 8.09154706e-04 4.85944003e-01 -2.30401322e-01
-1.60088956e-01 -4.93475169e-01 -1.09405816e+00 1.71785191e-01
5.97395420e-01 -4.26715136e-01 -4.96858120e-01 4.55995142e-01
4.70510840e-01 -3.71406734e-01 -7.85545111e-01 8.52238774e-01
4.62148398e-01 -9.73217249e-01 9.72975552e-01 1.64767176e-01
1.32142827e-01 -8.81677747e-01 -3.17251474e-01 -1.22873068e+00
-5.87721765e-01 -2.08876431e-01 1.42169029e-01 1.34859002e+00
1.01422258e-01 -6.83552384e-01 9.49595422e-02 5.54785788e-01
-1.03578180e-01 -2.26795033e-01 -7.55065978e-01 -4.90300864e-01
-7.16473043e-01 -1.17287405e-01 1.46229222e-01 6.18170619e-01
-5.07406533e-01 1.28800869e-01 -7.12630391e-01 6.94852233e-01
1.29472029e+00 2.44763240e-01 4.83315676e-01 -1.27849710e+00
-3.14656377e-01 -1.24114811e-01 -2.37119988e-01 -6.56708956e-01
4.66751121e-02 -6.23485982e-01 5.12382686e-01 -1.77703571e+00
6.60418153e-01 -1.07413799e-01 -3.58060658e-01 3.42989177e-01
-5.91026425e-01 2.08093122e-01 7.84519836e-02 3.14591736e-01
-3.60263437e-01 8.60605419e-01 1.28810549e+00 -5.78499734e-01
3.97652350e-02 -4.78796996e-02 -7.47410655e-01 3.88353914e-01
7.18663931e-01 -4.77331549e-01 -2.15160877e-01 -4.76714820e-01
1.15822181e-02 2.42668226e-01 6.61621869e-01 -7.92433858e-01
5.11383653e-01 -6.49312317e-01 6.48269355e-01 -8.53489339e-01
5.74262500e-01 -1.05085254e+00 6.03478789e-01 3.76733512e-01
1.76617444e-01 -6.46029472e-01 -4.20419499e-02 7.37537622e-01
-2.02841133e-01 -1.26975939e-01 9.94742751e-01 -6.38556108e-03
-5.88179946e-01 2.79937863e-01 -1.46314666e-01 -8.05171907e-01
1.17355156e+00 -3.10530931e-01 -6.01126492e-01 -1.69167861e-01
-5.05659044e-01 2.26190791e-01 3.58561814e-01 1.52802601e-01
7.89328814e-01 -1.19302654e+00 -1.04658127e+00 -7.11971149e-02
2.59173334e-01 7.13025257e-02 7.14145184e-01 1.07807004e+00
-2.33934388e-01 -1.10127717e-01 2.03598049e-02 -9.48568761e-01
-1.57447410e+00 6.44330919e-01 3.81353587e-01 5.04098088e-02
-3.11730683e-01 5.02862751e-01 5.22663653e-01 -6.16274290e-02
-1.38211772e-01 2.60202307e-02 -4.68205452e-01 -2.49032393e-01
7.14578331e-01 2.98236251e-01 -1.10861897e-01 -8.36500645e-01
-1.99560016e-01 8.42836440e-01 1.63365796e-01 -3.28672752e-02
1.08951223e+00 -5.11122048e-01 -5.51521420e-01 4.12875950e-01
7.60729909e-01 -1.42606795e-01 -1.30327260e+00 -3.98651540e-01
-4.75256443e-01 -8.20159554e-01 6.46218121e-01 -7.12709010e-01
-1.08811772e+00 6.26867175e-01 1.04477215e+00 -1.18594266e-01
1.42146015e+00 -7.82416668e-03 1.98238522e-01 9.66920778e-02
3.41470897e-01 -9.74582672e-01 -6.44936040e-02 7.82660469e-02
7.41989672e-01 -1.46732163e+00 3.36373806e-01 -7.38696218e-01
-4.89009559e-01 1.08940673e+00 4.84273702e-01 3.72713774e-01
4.52261865e-01 2.10155979e-01 1.65710792e-01 -6.37114700e-03
-3.99201035e-01 -6.25983536e-01 5.47091067e-01 6.67031646e-01
2.01911211e-01 3.81315872e-02 -1.33798897e-01 1.70617625e-01
5.47446847e-01 2.03666016e-02 1.30459517e-01 8.43035161e-01
-8.08630943e-01 -7.35078037e-01 -1.14556336e+00 2.96712518e-01
-3.05364996e-01 -9.11560282e-02 -2.24842161e-01 -6.76807314e-02
2.16793846e-02 1.71875095e+00 -2.76478082e-01 -5.54533839e-01
-1.00819662e-01 -2.16921180e-01 4.39906865e-01 -3.81229609e-01
3.54658179e-02 5.88881314e-01 -1.63197666e-01 -4.37878281e-01
-9.32460487e-01 -3.68960708e-01 -6.89111114e-01 -1.80664554e-01
-9.80558157e-01 5.61148785e-02 8.77020895e-01 1.17535079e+00
6.89309835e-02 6.85196519e-01 9.05069709e-01 -1.03728080e+00
-4.10135001e-01 -1.04351425e+00 -9.42667246e-01 2.47052476e-01
7.54858404e-02 -7.42027462e-01 -5.88965654e-01 1.38185918e-01] | [10.563213348388672, -1.9487202167510986] |
ae23e0f9-a088-49e4-8c74-230e71fb64e8 | laneaf-robust-multi-lane-detection-with-1 | 2212.11533 | null | https://arxiv.org/abs/2212.11533v5 | https://arxiv.org/pdf/2212.11533v5.pdf | Multi Lane Detection | Lane detection is a long-standing task and a basic module in autonomous driving. The task is to detect the lane of the current driving road, and provide relevant information such as the ID, direction, curvature, width, length, with visualization. Our work is based on CNN backbone DLA-34, along with Affinity Fields, aims to achieve robust detection of various lanes without assuming the number of lanes. Besides, we investigate novel decoding methods to achieve more efficient lane detection algorithm. | ['Fei Wu', 'Luoyu Chen'] | 2022-12-22 | null | null | null | null | ['lane-detection'] | ['computer-vision'] | [-1.35801062e-01 -1.64818943e-01 -3.18577826e-01 -6.72939062e-01
-1.30356610e-01 -5.25540650e-01 3.13937008e-01 -1.33278713e-01
-5.28603733e-01 4.94985789e-01 2.64111459e-01 -9.32962894e-01
1.55517057e-01 -9.14097667e-01 -6.39612436e-01 -6.47534966e-01
-3.31700593e-02 3.47697474e-02 7.79324055e-01 -5.73468268e-01
7.49264061e-01 8.21146905e-01 -1.61520827e+00 -1.83402270e-01
8.36856365e-01 8.60214055e-01 2.23769173e-01 6.20203674e-01
-2.86587059e-01 3.52023542e-01 -9.05889422e-02 -8.92207175e-02
-1.23864390e-01 1.57297730e-01 -2.59987652e-01 -1.28546268e-01
5.14945388e-01 -5.36998332e-01 -6.78798139e-01 1.13433981e+00
3.64602834e-01 -3.42964344e-02 8.62676263e-01 -1.22924995e+00
1.62736207e-01 4.10814583e-01 -5.48161328e-01 4.17317599e-01
-9.94240642e-02 3.42233449e-01 6.72024786e-01 -7.57754087e-01
5.84111452e-01 1.29422963e+00 2.75591522e-01 2.87205130e-01
-3.58576804e-01 -7.36965477e-01 1.40840784e-01 5.63456833e-01
-1.37107301e+00 -4.95436609e-01 8.75603139e-01 -4.02844191e-01
5.74616551e-01 1.90273225e-02 2.58449614e-01 5.44343948e-01
7.19542027e-01 7.78473496e-01 7.64041901e-01 -6.21012561e-02
5.40943779e-02 -2.21188679e-01 5.81648588e-01 1.13334513e+00
1.80832058e-01 -8.62948522e-02 -2.56172776e-01 3.40211660e-01
4.65646356e-01 -2.91214019e-01 7.35685453e-02 -4.41377521e-01
-1.29926360e+00 8.50306749e-01 4.98743415e-01 -1.34513974e-01
-5.02196960e-02 2.36397475e-01 5.75404227e-01 -6.49356917e-02
-1.38419837e-01 7.45552927e-02 -1.62430376e-01 -3.72527063e-01
-4.91727740e-01 5.85821271e-01 4.92303431e-01 1.19611716e+00
1.29663336e+00 -1.31929249e-01 -1.73099145e-01 4.70743418e-01
5.03350616e-01 7.78247774e-01 3.33662815e-02 -1.20834899e+00
8.21296692e-01 5.24491370e-01 -6.28346056e-02 -1.25778449e+00
-7.72863448e-01 -2.26145819e-01 -7.84623742e-01 2.60024279e-01
2.66381413e-01 -2.87003458e-01 -9.41938043e-01 1.40174961e+00
3.64443399e-02 2.95904279e-01 -3.01837660e-02 1.09555089e+00
8.98385346e-01 6.19428098e-01 -1.03736743e-01 2.43493453e-01
1.44008899e+00 -1.18611228e+00 -7.24383533e-01 -8.91744733e-01
8.74679387e-01 -7.66298950e-01 5.83239615e-01 2.89750248e-01
-6.63946569e-01 -7.02979565e-01 -1.30875540e+00 -5.70101738e-01
-7.78744161e-01 5.63701868e-01 7.59660423e-01 5.24341404e-01
-8.92957330e-01 -4.96980958e-02 -1.96145430e-01 -1.84024274e-02
2.03190356e-01 1.62999541e-01 -3.45799148e-01 -1.94103569e-01
-1.49270213e+00 9.02184606e-01 5.19876003e-01 6.25022829e-01
-5.63655734e-01 -7.52161592e-02 -1.25617981e+00 -2.24186584e-01
3.73400003e-01 -2.95350492e-01 9.20137882e-01 -6.47843033e-02
-1.29332960e+00 6.89843953e-01 -6.64176762e-01 -5.34236073e-01
6.70525074e-01 -5.20310774e-02 -6.99259162e-01 -3.07736367e-01
1.93996236e-01 1.13215995e+00 5.69557786e-01 -8.84528399e-01
-1.28081071e+00 -3.03773344e-01 -1.49119183e-01 2.61323214e-01
3.42420429e-01 -4.21575516e-01 -8.47287655e-01 -3.01688295e-02
2.84859598e-01 -8.87619436e-01 -3.96849662e-01 -1.03925958e-01
-7.30862260e-01 -3.17340136e-01 1.26583660e+00 -5.59752941e-01
1.39698219e+00 -2.55779743e+00 -3.44903678e-01 4.74630147e-01
4.37245280e-01 1.49223447e-01 7.54866675e-02 -2.53589619e-02
5.00038743e-01 -2.09610417e-01 -1.65721834e-01 1.34998307e-01
1.86215401e-01 2.64041305e-01 -5.36121607e-01 5.50448418e-01
2.35636905e-01 8.93955469e-01 -6.85314238e-01 -5.57067692e-01
4.99427557e-01 2.17981577e-01 -1.53394595e-01 7.16513172e-02
7.81129226e-02 8.94459039e-02 -3.67750645e-01 3.55903387e-01
1.14825380e+00 4.18666691e-01 -2.79712319e-01 -3.13303411e-01
-8.12034607e-01 3.92212331e-01 -1.10634744e+00 1.25585842e+00
-1.94707125e-01 1.29760790e+00 1.75633445e-01 -5.99515617e-01
1.34359288e+00 -5.64356327e-01 -7.79745802e-02 -1.06833732e+00
3.04564297e-01 1.81382358e-01 3.21950614e-01 -6.17876530e-01
9.27220702e-01 7.88647294e-01 -5.49905181e-01 -8.38539526e-02
-7.52772510e-01 1.18851461e-01 3.80402744e-01 1.39252469e-01
8.72517943e-01 -2.94294715e-01 -2.08486438e-01 -3.34199250e-01
6.64866984e-01 4.13125083e-02 6.12640560e-01 5.35006762e-01
-6.00405812e-01 2.15522304e-01 7.54538298e-01 -5.40394604e-01
-1.02256095e+00 -9.35342133e-01 -1.90116435e-01 8.11878681e-01
7.92352796e-01 -1.47553623e-01 -6.29499793e-01 -5.14525533e-01
4.14374381e-01 7.00445473e-01 -5.50621212e-01 -1.96563452e-01
-9.25711989e-01 -2.34912559e-01 4.36543286e-01 6.76654756e-01
6.80436432e-01 -8.18314850e-01 -7.41221547e-01 1.95753917e-01
-5.06024249e-02 -1.19060075e+00 -7.14134693e-01 3.39376807e-01
-2.90378213e-01 -9.95570600e-01 -1.57388672e-01 -1.15927386e+00
5.40146053e-01 5.81847548e-01 6.55321777e-01 2.40022559e-02
-1.12074077e-01 -5.95271111e-01 3.00387293e-01 -4.85600233e-01
-4.35043275e-02 3.71605426e-01 -3.08706611e-01 -1.91013634e-01
5.40058911e-01 4.67067733e-02 -7.72436619e-01 8.12149584e-01
-1.57977134e-01 3.68776947e-01 7.85907984e-01 3.38914365e-01
4.62642044e-01 3.53896879e-02 4.57086951e-01 -8.47552717e-01
6.96861863e-01 -3.64141256e-01 -7.14757264e-01 -1.72396258e-01
-4.51607466e-01 1.16106547e-01 4.97924775e-01 1.11002393e-01
-7.37802446e-01 1.30770311e-01 -5.88574290e-01 7.66880065e-02
-5.09206176e-01 1.69330046e-01 -6.20197892e-01 -1.12486236e-01
1.64135024e-01 3.18585247e-01 -5.02149947e-02 -2.08070893e-02
5.66649437e-01 9.36024129e-01 9.85697985e-01 2.36693816e-03
3.95220995e-01 5.12322843e-01 2.09385306e-01 -8.07877362e-01
-4.02082920e-01 -7.91848242e-01 -1.19227588e+00 -4.48453635e-01
6.78290904e-01 -8.74067962e-01 -1.03592873e+00 4.96794939e-01
-1.16525209e+00 -3.31560880e-01 6.15706861e-01 2.50491202e-01
-4.14476603e-01 3.59748304e-01 -3.31566095e-01 -6.61056221e-01
-1.97173178e-01 -1.42411458e+00 1.07820654e+00 4.36795771e-01
2.85178542e-01 -7.89707899e-01 -4.75627780e-01 1.98694542e-01
4.13454711e-01 4.20545228e-02 9.17939246e-01 -1.41215697e-01
-7.26321995e-01 -2.33622372e-01 -7.90040553e-01 -1.22182399e-01
-1.39238358e-01 2.22632363e-01 -6.86405599e-01 2.00765990e-02
-9.61946011e-01 2.51403272e-01 1.33835542e+00 3.37901682e-01
1.33636236e+00 1.68540075e-01 -6.58910811e-01 8.25666726e-01
9.56485033e-01 3.90091419e-01 8.65272939e-01 3.97574514e-01
9.99272227e-01 7.05579698e-01 1.05663514e+00 4.45608124e-02
8.72836471e-01 4.54397559e-01 7.26462245e-01 -1.75906166e-01
-9.14372355e-02 -3.33317816e-01 3.89811486e-01 7.30728626e-01
5.45379877e-01 -3.11298043e-01 -9.13624346e-01 5.14312029e-01
-1.99274433e+00 -7.28682518e-01 -8.59614313e-01 1.87345982e+00
5.36446087e-02 9.97148454e-01 1.37665004e-01 2.74486709e-02
8.21419001e-01 3.21006298e-01 -8.39637578e-01 -8.79318833e-01
-1.67737126e-01 -6.94973826e-01 1.34140515e+00 5.96980095e-01
-1.31908369e+00 1.20476520e+00 7.27540922e+00 9.02083874e-01
-1.51111686e+00 -4.69977617e-01 6.26110673e-01 6.75365210e-01
-2.72581697e-01 -2.23371267e-01 -1.50030529e+00 7.22380817e-01
8.45132172e-01 1.04765967e-01 1.93705440e-01 7.70490766e-01
5.62955678e-01 -4.43967402e-01 -5.21927416e-01 8.04224968e-01
5.83131202e-02 -1.31820416e+00 -5.67775741e-02 2.09133476e-01
3.42314124e-01 2.34634295e-01 -3.24417278e-02 4.77341801e-01
-1.37729454e-03 -9.26641524e-01 7.09411621e-01 6.53807163e-01
8.09117377e-01 -1.21289706e+00 1.00323999e+00 4.95213836e-01
-1.47658038e+00 -1.42418995e-01 -5.75009465e-01 -4.78755422e-02
3.45034063e-01 4.77746308e-01 -1.00938129e+00 9.60610434e-02
2.60796905e-01 9.54306483e-01 -7.45912492e-01 1.15175724e+00
-4.16609615e-01 6.10373914e-01 -1.21740542e-01 -3.02490205e-01
6.21019721e-01 -2.65842825e-01 3.74073178e-01 1.62142193e+00
1.07939728e-01 -3.99602115e-01 2.92842567e-01 5.55066347e-01
1.32983133e-01 6.10324033e-02 -8.42323899e-01 5.59755445e-01
6.91614449e-01 1.25106514e+00 -6.47628486e-01 -1.78379446e-01
-3.75504017e-01 4.64343548e-01 3.18979114e-01 8.67463499e-02
-8.34586024e-01 -9.80998755e-01 9.18795824e-01 3.07181060e-01
2.79039741e-01 -6.89191222e-01 -3.73882532e-01 -3.88580978e-01
-1.47321612e-01 -1.95292518e-01 -1.26145855e-02 -8.58276665e-01
-6.19271636e-01 1.17734112e-01 -3.82090211e-01 -1.00594318e+00
-9.20837969e-02 -8.46353650e-01 -9.44132328e-01 9.01963532e-01
-1.99840832e+00 -9.49541271e-01 -6.05878949e-01 3.60319346e-01
7.13839650e-01 -9.04562101e-02 5.85474186e-02 5.67271888e-01
-7.43790269e-01 6.14577055e-01 2.00926095e-01 3.71360064e-01
6.65654242e-01 -1.26811576e+00 1.00635898e+00 7.19795048e-01
-6.66494846e-01 2.89413184e-01 8.12871099e-01 -6.19937658e-01
-1.63989139e+00 -1.41134453e+00 6.41516805e-01 -5.39051369e-02
6.09894693e-01 -4.85669345e-01 -8.31719160e-01 3.71320993e-01
-1.64535698e-02 -1.99778855e-01 -5.78531437e-02 -1.77616641e-01
-1.24765681e-02 -2.91981280e-01 -6.51318550e-01 6.59986377e-01
1.16066325e+00 -2.56132871e-01 -1.39946163e-01 -8.52730721e-02
4.98274893e-01 -6.69863641e-01 -1.60453379e-01 3.19344193e-01
5.07521927e-01 -8.29963207e-01 7.55758464e-01 7.60808541e-03
1.91239342e-01 -7.97143877e-01 2.31826544e-01 -1.05975425e+00
-6.05360508e-01 -3.03464681e-01 2.06744894e-01 8.34157884e-01
5.04481077e-01 -7.42951691e-01 9.05347645e-01 2.30791092e-01
-6.91411674e-01 -7.25760341e-01 -9.03633058e-01 -3.09515208e-01
-1.67192847e-01 -6.01426721e-01 5.70330620e-01 4.08768952e-01
-3.02400500e-01 3.59273434e-01 -3.47210437e-01 5.42642653e-01
4.72052425e-01 1.18782409e-02 1.07872176e+00 -1.16505039e+00
8.11707497e-01 -7.75172472e-01 -8.05750072e-01 -1.66431510e+00
3.85652453e-01 -6.93183303e-01 5.92763484e-01 -1.61661971e+00
-1.48092762e-01 -3.68038803e-01 -9.41854119e-02 3.52462649e-01
7.25560114e-02 9.21772718e-02 -3.73365015e-01 -1.11766182e-01
-9.14821684e-01 1.72924563e-01 1.36754310e+00 -3.87946218e-01
2.26676017e-02 3.89879197e-02 -4.26553696e-01 6.05539382e-01
9.08621490e-01 2.55638003e-01 -2.65557766e-01 -3.29478949e-01
2.10645020e-01 -1.34240746e-01 2.14425847e-01 -1.08546340e+00
7.00498283e-01 -2.83365786e-01 2.10847989e-01 -1.49458599e+00
1.72763824e-01 -4.89640802e-01 -4.39375907e-01 4.65545177e-01
-2.72519320e-01 4.43958074e-01 2.62689710e-01 6.15989089e-01
-1.52575001e-01 -9.56925005e-02 4.69874620e-01 4.36626971e-01
-1.72457087e+00 1.39378041e-01 -7.06723690e-01 -2.33243302e-01
1.12577665e+00 -3.78314883e-01 -7.06018925e-01 -1.56342581e-01
-3.17849070e-01 8.93914819e-01 4.04632598e-01 6.95099175e-01
8.04208636e-01 -1.20771027e+00 -6.92672253e-01 5.55046678e-01
4.25769567e-01 9.82120261e-02 1.86588779e-01 5.92387199e-01
-8.51218879e-01 7.12474763e-01 -4.51671064e-01 -6.12382233e-01
-1.16311455e+00 5.06661475e-01 3.43830079e-01 4.23111081e-01
-6.87270343e-01 3.81611347e-01 1.68055922e-01 -2.24857256e-01
3.09207678e-01 -3.04932177e-01 -6.07667744e-01 -9.06007513e-02
6.11545086e-01 4.58584368e-01 1.76496193e-01 -1.01629186e+00
-5.18638313e-01 8.41076732e-01 -1.61076114e-01 2.44485736e-01
6.40408874e-01 -5.41436374e-01 3.44056003e-02 3.66400272e-01
1.15482581e+00 -4.89209592e-02 -1.41212273e+00 2.27692991e-01
3.15241925e-02 -3.81509513e-01 4.21109557e-01 -5.77379823e-01
-9.80333269e-01 1.19243097e+00 7.72853255e-01 -9.86918807e-02
3.94387066e-01 -4.32805359e-01 1.06495368e+00 4.65653092e-01
3.81968558e-01 -1.19941270e+00 -6.67617738e-01 1.03776395e+00
5.87418973e-01 -1.45288336e+00 -2.62413383e-01 -6.25184417e-01
-3.90716016e-01 1.39401555e+00 1.00867486e+00 1.19605914e-01
7.08713949e-01 3.75731558e-01 2.72640973e-01 -3.28137368e-01
-7.31494963e-01 -5.65603733e-01 3.94125909e-01 5.41960716e-01
2.61960894e-01 3.29842627e-01 -3.31221223e-01 -7.53790513e-02
-3.36971045e-01 -4.68603343e-01 5.30156791e-01 7.34281063e-01
-1.16550696e+00 -5.36946774e-01 -1.32375449e-01 6.47771180e-01
1.43329412e-01 1.52803436e-01 -1.51627749e-01 6.20300889e-01
4.13195789e-01 8.07315648e-01 5.01075029e-01 -5.20441115e-01
5.65707445e-01 -3.42767835e-01 -3.37514073e-01 -1.03824608e-01
2.12787613e-01 -2.97005802e-01 2.37413481e-01 -4.22390521e-01
3.51725936e-01 -5.43364346e-01 -1.64907062e+00 -5.99775493e-01
-2.38054127e-01 3.30804139e-02 1.11300659e+00 9.32789087e-01
4.93906021e-01 4.15352583e-01 7.97558129e-01 -7.02122748e-01
1.70941249e-01 -6.69186056e-01 -5.23174167e-01 2.35307172e-01
6.01023853e-01 -8.18248332e-01 -2.09095582e-01 -5.06683946e-01] | [8.042351722717285, -1.4236348867416382] |
b5d33892-2bcb-4266-a9e6-72f17fb9eb54 | detecting-oriented-text-in-natural-images-by | 1703.06520 | null | http://arxiv.org/abs/1703.06520v3 | http://arxiv.org/pdf/1703.06520v3.pdf | Detecting Oriented Text in Natural Images by Linking Segments | Most state-of-the-art text detection methods are specific to horizontal Latin
text and are not fast enough for real-time applications. We introduce Segment
Linking (SegLink), an oriented text detection method. The main idea is to
decompose text into two locally detectable elements, namely segments and links.
A segment is an oriented box covering a part of a word or text line; A link
connects two adjacent segments, indicating that they belong to the same word or
text line. Both elements are detected densely at multiple scales by an
end-to-end trained, fully-convolutional neural network. Final detections are
produced by combining segments connected by links. Compared with previous
methods, SegLink improves along the dimensions of accuracy, speed, and ease of
training. It achieves an f-measure of 75.0% on the standard ICDAR 2015
Incidental (Challenge 4) benchmark, outperforming the previous best by a large
margin. It runs at over 20 FPS on 512x512 images. Moreover, without
modification, SegLink is able to detect long lines of non-Latin text, such as
Chinese. | ['Serge Belongie', 'Baoguang Shi', 'Xiang Bai'] | 2017-03-19 | detecting-oriented-text-in-natural-images-by-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Shi_Detecting_Oriented_Text_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Shi_Detecting_Oriented_Text_CVPR_2017_paper.pdf | cvpr-2017-7 | ['curved-text-detection'] | ['computer-vision'] | [ 2.67239183e-01 1.16272934e-01 -2.35209346e-01 1.58060029e-01
-7.93023348e-01 -8.05557609e-01 7.31693983e-01 6.56559348e-01
-4.19037819e-01 1.46635726e-01 3.60373929e-02 -1.90375775e-01
6.07428432e-01 -5.04793406e-01 -8.14044416e-01 -1.89050063e-01
1.08156763e-01 5.29025376e-01 9.32207167e-01 2.55140215e-02
2.16270104e-01 2.86350697e-01 -1.01961339e+00 3.63612324e-01
6.63203239e-01 9.55293596e-01 3.43840336e-03 7.62483537e-01
-3.26675028e-01 4.34715092e-01 -5.90449154e-01 -6.37590945e-01
1.12420149e-01 -2.60553688e-01 -5.51648319e-01 4.16967154e-01
8.66118014e-01 -3.10559422e-01 -4.13424850e-01 9.32137966e-01
3.33087146e-01 -6.61421537e-01 5.91256082e-01 -7.73873389e-01
-4.49095815e-01 8.85432959e-01 -1.13456857e+00 7.51085803e-02
2.89613962e-01 -3.68478864e-01 1.07949722e+00 -1.23452306e+00
8.52761030e-01 1.11556304e+00 1.01238203e+00 7.26184919e-02
-1.09539628e+00 -3.37387085e-01 2.47356772e-01 -2.67576069e-01
-1.35042882e+00 -5.14557771e-02 2.50309855e-01 -6.15544736e-01
7.70371377e-01 1.11333430e-01 4.55386311e-01 8.28369081e-01
4.18078661e-01 1.26598322e+00 5.89062214e-01 -6.76577151e-01
-1.13224275e-01 -1.38428118e-02 1.68106243e-01 1.11890149e+00
5.32005787e-01 -6.18001819e-01 -6.75155461e-01 3.09266523e-02
6.04221106e-01 -2.03277171e-01 -2.92698294e-01 -4.75478709e-01
-1.54482305e+00 7.73487329e-01 6.11476362e-01 3.60847741e-01
-7.13921189e-02 1.30314842e-01 6.14846408e-01 -9.33216959e-02
6.11057997e-01 7.31021240e-02 -1.73817992e-01 1.46172807e-01
-1.40592945e+00 3.44899446e-02 6.23664618e-01 9.81914163e-01
3.32821667e-01 -2.50562429e-01 -2.50481606e-01 7.63572574e-01
3.15794259e-01 7.70641029e-01 3.87513906e-01 -2.85793930e-01
8.92761886e-01 9.01688039e-01 -1.95837930e-01 -8.90109897e-01
-7.29073703e-01 -7.02717364e-01 -8.57269704e-01 3.11960783e-02
4.86727804e-01 -2.23315999e-01 -1.01459634e+00 9.81267571e-01
2.25702569e-01 -3.48996669e-01 -3.42019945e-01 5.81050575e-01
7.39698589e-01 6.67897999e-01 -1.52489170e-01 3.33147421e-02
1.83610523e+00 -1.28817821e+00 -6.59622550e-01 -6.56345248e-01
7.73180604e-01 -1.29580939e+00 8.20728719e-01 5.29352665e-01
-1.16349483e+00 -5.11829555e-01 -1.24926043e+00 -2.82997727e-01
-4.74664450e-01 6.73049450e-01 -5.20042554e-02 5.30801415e-01
-7.82519221e-01 2.78506786e-01 -9.34042871e-01 -8.50079894e-01
3.62786800e-01 1.85794711e-01 -2.88555324e-01 2.06541598e-01
-7.35694110e-01 3.07719797e-01 3.47708583e-01 -1.01602100e-01
-2.39317894e-01 -3.07798177e-01 -8.01246524e-01 8.24807659e-02
2.91179806e-01 -4.38307196e-01 1.06848025e+00 -9.52125609e-01
-9.60301995e-01 1.16327977e+00 -1.22838572e-01 -7.46798933e-01
1.23167336e+00 -6.17484748e-01 -4.96518731e-01 5.37295282e-01
4.20372754e-01 9.33266938e-01 1.08211279e+00 -9.83052492e-01
-9.33501244e-01 -2.86666393e-01 -6.05414033e-01 7.58030117e-02
-4.47401524e-01 1.79127574e-01 -1.34107006e+00 -1.08014905e+00
3.66681099e-01 -8.44189763e-01 3.67198050e-01 4.14537072e-01
-1.10394156e+00 -2.87395746e-01 1.30884695e+00 -7.94092298e-01
1.29691589e+00 -2.08652234e+00 -1.49656907e-01 3.13644484e-02
4.24358159e-01 5.06359220e-01 -7.84476753e-03 5.18154562e-01
1.53599232e-01 2.80765563e-01 -1.90042943e-01 -6.20936513e-01
-2.45346595e-02 -4.83527064e-01 -3.21280509e-01 7.32021451e-01
7.04752877e-02 9.39889967e-01 -5.67810357e-01 -7.43671894e-01
1.69269562e-01 4.96966273e-01 -4.90194857e-02 -2.84417868e-01
-2.96122670e-01 -2.00497344e-01 -2.11015537e-01 6.00126147e-01
8.53006542e-01 -3.91618401e-01 5.35926744e-02 -1.64050117e-01
-2.69488782e-01 8.81185308e-02 -1.18773866e+00 1.51274455e+00
1.39398053e-01 1.45146775e+00 -1.18295290e-01 -4.85493302e-01
9.23244894e-01 1.37083307e-01 1.06707774e-01 -6.37377203e-01
1.66243583e-01 3.71156186e-01 -4.82953936e-01 -4.23367918e-01
8.00812960e-01 5.75857878e-01 -1.15274727e-01 2.95640558e-01
-2.01851472e-01 6.57888204e-02 5.89777231e-01 4.09364104e-01
1.10635674e+00 7.78401196e-02 2.03918010e-01 -1.98640227e-01
4.37137455e-01 -1.09706417e-01 1.92535326e-01 8.51132751e-01
-4.38196100e-02 1.00315106e+00 8.15062523e-01 -2.63787270e-01
-1.22193778e+00 -1.17344701e+00 -3.36860746e-01 1.01395631e+00
1.33262873e-01 -7.21005440e-01 -1.00712967e+00 -9.66146767e-01
5.83929271e-02 1.71577543e-01 -6.91008329e-01 3.44583839e-01
-7.66244888e-01 -4.01018947e-01 7.68732607e-01 6.43890500e-01
8.78692329e-01 -8.57820272e-01 -5.50627947e-01 1.97729558e-01
-2.08351478e-01 -1.51788783e+00 -8.58874202e-01 1.56464130e-01
-6.72160029e-01 -1.01941872e+00 -1.18706918e+00 -1.12030387e+00
7.15848207e-01 4.14336830e-01 1.13724756e+00 -2.52911658e-03
-4.93958384e-01 3.18840817e-02 -3.46597999e-01 -2.73663193e-01
-3.10045689e-01 4.11474794e-01 -3.54251295e-01 6.53418228e-02
2.44351104e-01 3.86797071e-01 -6.58641338e-01 5.53557515e-01
-6.98488116e-01 1.11274175e-01 6.91069305e-01 5.97431183e-01
5.80485523e-01 -1.46126807e-01 7.71312192e-02 -8.34456563e-01
3.53390038e-01 9.27617867e-03 -7.57948041e-01 2.60997415e-01
-3.45260292e-01 -2.49499947e-01 4.62945640e-01 -2.81380385e-01
-5.83825707e-01 3.38984489e-01 -7.20447488e-03 -8.63903537e-02
-1.51827484e-01 4.04557228e-01 3.37578744e-01 1.98731020e-01
7.26142585e-01 2.21385375e-01 -3.52601677e-01 -4.12311971e-01
2.92164385e-01 7.33225405e-01 6.76871657e-01 1.86648760e-02
7.78116047e-01 7.74755657e-01 -1.28937989e-01 -1.07059169e+00
-7.83550084e-01 -7.12314010e-01 -8.25644851e-01 -2.30105966e-02
1.02102685e+00 -9.80422020e-01 -2.41518661e-01 7.38897681e-01
-1.08558738e+00 -3.63529027e-01 8.30957368e-02 2.03342527e-01
-1.69485152e-01 7.44130731e-01 -7.48676956e-01 -3.80368352e-01
-5.72208166e-01 -8.01016331e-01 1.72853136e+00 2.31057346e-01
-1.49799228e-01 -8.41858745e-01 -1.02367759e-01 2.76756465e-01
-4.89516146e-02 1.76516682e-01 5.79437792e-01 -5.69014370e-01
-2.99610674e-01 -7.69330084e-01 -6.08311594e-01 4.90282513e-02
-2.79653370e-01 3.91497701e-01 -7.29177833e-01 -4.48197305e-01
-7.24603951e-01 -2.05287322e-01 1.25374961e+00 2.91878849e-01
8.16953540e-01 -1.62356738e-02 -7.75179207e-01 2.99316078e-01
1.28287351e+00 -1.57749265e-01 3.81514758e-01 5.53797245e-01
8.20288420e-01 3.89903992e-01 4.69586462e-01 2.06714630e-01
1.66665971e-01 7.71278262e-01 3.12482893e-01 -5.85958898e-01
-4.24055368e-01 -1.85620308e-01 3.13405335e-01 3.94231200e-01
5.69092631e-01 -7.83019722e-01 -1.23345923e+00 4.64196235e-01
-1.94202650e+00 -5.49328625e-01 -8.27455401e-01 2.07316136e+00
4.94927645e-01 8.62110198e-01 3.75332743e-01 4.22042698e-01
1.08130002e+00 8.25073868e-02 -5.09622216e-01 -2.14817211e-01
-4.77168143e-01 -1.22848324e-01 7.11985350e-01 2.15564728e-01
-1.65359592e+00 1.17639315e+00 5.69533873e+00 1.01182711e+00
-1.15874255e+00 -8.98519382e-02 6.96414769e-01 2.07298160e-01
3.65038693e-01 -3.89516175e-01 -1.13778198e+00 2.68593431e-01
4.13466036e-01 3.02927822e-01 -2.62086630e-01 7.01406419e-01
1.74933359e-01 -3.84366840e-01 -8.39037836e-01 6.65280104e-01
3.28381211e-01 -1.36387515e+00 -5.45721911e-02 3.68886031e-02
9.74865258e-01 3.65085453e-01 6.08912185e-02 9.02753416e-03
-6.28235936e-02 -8.28999221e-01 1.12888932e+00 1.46033898e-01
8.54035974e-01 -5.16131043e-01 6.97001338e-01 2.67264187e-01
-1.56084764e+00 2.62417942e-01 -2.19516054e-01 5.66470921e-01
7.80901238e-02 6.85351789e-01 -7.93337286e-01 3.52380365e-01
6.36815071e-01 7.42341459e-01 -9.74789739e-01 1.38620055e+00
-5.53942978e-01 8.09522033e-01 -3.96072686e-01 -5.28384298e-02
5.28501868e-01 1.70658499e-01 4.46634620e-01 1.80759859e+00
2.46437207e-01 -7.47890890e-01 2.33355910e-01 6.08295679e-01
-5.36670983e-01 2.77460694e-01 -3.36678654e-01 7.68932104e-02
3.34644973e-01 1.41094232e+00 -1.43956065e+00 -3.87653321e-01
-5.73677659e-01 1.33871400e+00 1.93469539e-01 1.82763964e-01
-9.03877854e-01 -8.68168294e-01 -1.06324062e-01 2.38533437e-01
8.32181871e-01 -2.76117325e-01 -4.46515262e-01 -9.65015829e-01
3.59322846e-01 -7.83576071e-01 2.56756216e-01 -7.50928283e-01
-7.87661552e-01 6.88006282e-01 -7.70098090e-01 -1.30707860e+00
1.89265639e-01 -7.38107324e-01 -4.95354414e-01 5.13042390e-01
-1.23713112e+00 -1.46186399e+00 -6.08428001e-01 3.07133853e-01
8.51469159e-01 4.95049218e-03 5.32473266e-01 2.44954918e-02
-8.90351534e-01 6.78096116e-01 5.82556069e-01 6.75917387e-01
8.95140529e-01 -1.33862519e+00 9.85570908e-01 1.17237604e+00
3.88750523e-01 1.69536501e-01 5.14451087e-01 -8.68742824e-01
-9.12385702e-01 -1.30190539e+00 1.14592957e+00 -3.23582321e-01
6.93293691e-01 -8.39511693e-01 -8.49959433e-01 6.32562935e-01
2.95455039e-01 7.33221769e-02 1.98489130e-01 -2.13498518e-01
-6.89747810e-01 2.71086186e-01 -7.49100804e-01 6.39695048e-01
6.72970533e-01 -2.75906831e-01 -3.97312671e-01 6.45866096e-01
4.04779553e-01 -5.13991416e-01 -4.67413455e-01 2.11295187e-01
4.79849517e-01 -1.16255200e+00 8.31400514e-01 1.04759760e-01
5.78509867e-01 -3.43120873e-01 2.65630364e-01 -7.34315991e-01
-1.34063497e-01 -6.50821567e-01 -1.31247476e-01 1.29975545e+00
5.38673103e-01 -3.12739253e-01 8.99689674e-01 -5.57084262e-01
-1.53898165e-01 -6.33761704e-01 -8.97228360e-01 -8.64163339e-01
1.80096164e-01 -3.36477786e-01 1.13904707e-01 6.71111584e-01
1.20922633e-01 4.83040452e-01 -9.02668685e-02 9.85167101e-02
5.22368312e-01 8.11695158e-02 6.82087779e-01 -1.37932348e+00
1.39749080e-01 -9.68947470e-01 -4.24307585e-01 -1.51034021e+00
-2.58337349e-01 -6.71756685e-01 4.70022708e-02 -1.86112452e+00
1.67825326e-01 -6.30526617e-02 2.80677080e-01 3.94969314e-01
-2.44370744e-01 8.91380131e-01 2.40567222e-01 4.10611063e-01
-8.57733071e-01 1.52521268e-01 8.37929308e-01 -2.77486145e-01
6.57183584e-03 -8.26659724e-02 -1.58412769e-01 1.02268863e+00
7.62851477e-01 -3.08579415e-01 2.48509020e-01 -4.95320886e-01
4.08070594e-01 -2.87648708e-01 1.22736230e-01 -1.26023114e+00
2.90808618e-01 7.27940142e-01 6.19337618e-01 -1.23913753e+00
1.42170815e-02 -5.72806001e-01 -5.92930138e-01 6.51578128e-01
-2.65636772e-01 5.92822535e-03 4.27604705e-01 5.77111363e-01
-1.23922713e-01 -2.75516808e-01 9.73423421e-01 3.74545008e-01
-3.71174127e-01 -8.04276839e-02 -5.89923918e-01 7.24289715e-02
1.15126240e+00 -3.01369011e-01 -7.04310954e-01 -6.05000071e-02
-3.63728613e-01 2.18995750e-01 4.81640995e-01 4.89069849e-01
3.40335995e-01 -8.75304222e-01 -8.00470710e-01 1.85687602e-01
3.76281261e-01 -5.03993556e-02 -1.49060130e-01 1.00640666e+00
-1.14874971e+00 9.47409451e-01 3.60902280e-01 -9.76100504e-01
-1.60511148e+00 4.31998223e-01 3.41141492e-01 -4.75551516e-01
-1.11442924e+00 8.95359576e-01 2.16981173e-01 -5.08191846e-02
6.48005486e-01 -4.08310831e-01 -1.20992966e-01 3.62210184e-01
5.66415608e-01 3.41437042e-01 3.10704768e-01 -6.49783552e-01
-3.39576542e-01 1.18607211e+00 -2.91147500e-01 -1.46055240e-02
7.89866328e-01 -1.33634001e-01 1.55552000e-01 4.38459605e-01
9.73836601e-01 3.87871563e-01 -1.20920086e+00 -4.28351969e-01
1.17852487e-01 -1.21640168e-01 8.89388099e-02 -8.72607887e-01
-1.00136745e+00 9.00367200e-01 7.84630120e-01 6.83536410e-01
6.94175959e-01 2.25695401e-01 8.34230244e-01 2.54096985e-01
-1.40356824e-01 -1.18799901e+00 4.28110033e-01 5.39103687e-01
7.55419135e-01 -1.24543524e+00 8.68065730e-02 -5.94166517e-01
-4.09198225e-01 1.63659644e+00 2.86094517e-01 7.81315118e-02
2.53914148e-01 4.21918631e-01 1.67794511e-01 -2.44008616e-01
-2.64887720e-01 -2.35975981e-01 6.26640558e-01 1.00912154e-01
6.50755584e-01 -1.78273365e-01 -2.96768785e-01 -2.24741474e-01
-3.81850265e-02 -3.17764431e-01 4.46034670e-01 8.73254478e-01
-7.96412885e-01 -6.34254217e-01 -6.67158604e-01 4.27771598e-01
-6.93822086e-01 -1.41441047e-01 -6.76621735e-01 1.05849147e+00
5.13648801e-02 1.00440633e+00 5.37280262e-01 1.05423294e-03
3.11877191e-01 -1.24383889e-01 2.84365918e-02 -4.30394799e-01
-5.80939412e-01 7.83276618e-01 -2.68545449e-02 -2.68565744e-01
-5.05500771e-02 -8.57311189e-01 -1.39312005e+00 -3.60284150e-01
-6.16078317e-01 -2.62657940e-01 7.92039275e-01 7.33435094e-01
2.89036155e-01 6.45353079e-01 3.26160550e-01 -5.83898425e-01
3.06528118e-02 -1.08624160e+00 -5.67186654e-01 1.36208698e-01
3.06604356e-01 -1.74569920e-01 -2.45303541e-01 2.90734887e-01] | [12.006707191467285, 2.354170799255371] |
e9cff06e-3893-4621-9780-bb3b1852114a | dreamdecompiler-improved-bayesian-program | 2306.07856 | null | https://arxiv.org/abs/2306.07856v1 | https://arxiv.org/pdf/2306.07856v1.pdf | DreamDecompiler: Improved Bayesian Program Learning by Decompiling Amortised Knowledge | Solving program induction problems requires searching through an enormous space of possibilities. DreamCoder is an inductive program synthesis system that, whilst solving problems, learns to simplify search in an iterative wake-sleep procedure. The cost of search is amortised by training a neural search policy, reducing search breadth and effectively "compiling" useful information to compose program solutions across tasks. Additionally, a library of program components is learnt to express discovered solutions in fewer components, reducing search depth. In DreamCoder, the neural search policy has only an indirect effect on the library learnt through the program solutions it helps discover. We present an approach for library learning that directly leverages the neural search policy, effectively "decompiling" its amortised knowledge to extract relevant program components. This provides stronger amortised inference: the amortised knowledge learnt to reduce search breadth is now also used to reduce search depth. We integrate our approach with DreamCoder and demonstrate faster domain proficiency with improved generalisation on a range of domains, particularly when fewer example solutions are available. | ['N. Siddharth', 'Christopher G. Lucas', 'Alessandro B. Palmarini'] | 2023-06-13 | null | null | null | null | ['program-synthesis', 'program-induction'] | ['computer-code', 'computer-code'] | [ 2.61247844e-01 2.86578149e-01 -7.61947393e-01 -3.00861150e-01
-7.69136965e-01 -8.89801919e-01 3.10011119e-01 1.15855835e-01
-5.63892424e-01 7.81800032e-01 2.21408278e-01 -8.68277669e-01
-7.94081017e-02 -1.04488885e+00 -1.12606609e+00 -1.46409258e-01
-7.51386359e-02 3.94571513e-01 -8.87030959e-02 -8.38517100e-02
3.77806067e-01 2.46449888e-01 -1.86241603e+00 4.95304048e-01
9.54101205e-01 4.40146327e-01 4.16412026e-01 9.23168182e-01
-3.57577860e-01 1.31638157e+00 -6.79972351e-01 -1.48857683e-01
1.48779303e-01 -5.06137647e-02 -1.37311113e+00 -5.62464297e-01
2.79038727e-01 -2.00591534e-01 -3.20092402e-02 9.13759708e-01
1.22759111e-01 2.36106887e-01 2.01212000e-02 -9.97350633e-01
-5.43594122e-01 1.26083660e+00 -3.25189769e-01 3.82595837e-01
6.55445814e-01 5.74916720e-01 1.37522554e+00 -4.33242291e-01
6.22129381e-01 1.19621789e+00 6.36391759e-01 7.28857934e-01
-1.72304165e+00 -8.77831042e-01 8.35367292e-02 3.83140184e-02
-1.06379378e+00 -3.93578023e-01 5.33635557e-01 -3.83966386e-01
1.70330274e+00 5.02601385e-01 7.85107493e-01 6.09422743e-01
-7.31368139e-02 9.86706614e-01 7.77571917e-01 -7.33141005e-01
2.54783511e-01 3.86126846e-01 2.32065767e-01 1.01490617e+00
-3.20483036e-02 2.67358482e-01 -4.91387993e-01 -2.29166567e-01
3.10313821e-01 -9.22863856e-02 -2.23977819e-01 -2.55512565e-01
-8.25763941e-01 8.55377495e-01 5.13348758e-01 1.83695540e-01
8.34051594e-02 5.01585066e-01 5.58308244e-01 5.61934292e-01
-8.30307156e-02 1.51523745e+00 -9.37522471e-01 -4.53012347e-01
-1.04340410e+00 4.75146323e-01 1.40931261e+00 1.06166971e+00
1.33912742e+00 5.31322770e-02 -1.53358072e-01 8.15817237e-01
-4.02701385e-02 3.32308382e-01 7.12186813e-01 -1.26849198e+00
5.82202077e-01 9.77032125e-01 -3.67823392e-01 -4.35672760e-01
-1.99586511e-01 -3.51023704e-01 -8.53582546e-02 3.77643377e-01
2.22738802e-01 -3.05512577e-01 -6.39186144e-01 1.94493532e+00
-1.30844727e-01 -2.06593439e-01 1.66746259e-01 4.58475798e-01
6.38128221e-01 7.45762289e-01 1.67140171e-01 1.34624407e-01
1.36415648e+00 -9.67236578e-01 -9.07316580e-02 -7.38268554e-01
1.21701038e+00 -3.30267102e-01 1.50261021e+00 6.48774028e-01
-1.36895299e+00 -2.56451994e-01 -1.07262671e+00 -3.51416200e-01
-3.60419095e-01 -1.19204707e-01 1.27102041e+00 4.75383580e-01
-1.20730841e+00 7.40317643e-01 -3.81894886e-01 1.73346385e-01
6.97561026e-01 7.91251600e-01 -1.04971744e-01 -2.85294771e-01
-8.28417063e-01 8.84642065e-01 6.73357666e-01 -5.82820356e-01
-1.02431464e+00 -1.48516715e+00 -1.03947616e+00 3.87337267e-01
5.50277650e-01 -8.38165998e-01 1.67652941e+00 -9.06095862e-01
-1.38242185e+00 7.84923851e-01 -2.70825654e-01 -4.83909875e-01
-1.56184569e-01 -2.03707535e-02 2.53451318e-01 -3.15061122e-01
1.29903853e-01 7.28970468e-01 3.76757801e-01 -8.81632090e-01
-6.36822999e-01 -1.67417169e-01 6.86649263e-01 9.58677009e-02
-1.71774864e-01 -4.77853455e-02 -2.94919729e-01 -3.95440936e-01
-4.37068820e-01 -9.42081749e-01 -2.07604513e-01 -4.61329639e-01
5.51969558e-02 -5.94049752e-01 4.61642683e-01 -5.88227570e-01
1.44892216e+00 -1.97948432e+00 3.46796542e-01 1.99723318e-01
4.03986245e-01 3.21840085e-02 -2.51158506e-01 -1.72244273e-02
-2.15519637e-01 2.87084192e-01 -2.70215452e-01 1.25374228e-01
2.23261341e-01 4.51636642e-01 -4.15071040e-01 -1.13187775e-01
2.55452305e-01 1.21647561e+00 -1.03919506e+00 -2.50098884e-01
-3.03167664e-02 -2.31483713e-01 -1.40692282e+00 2.14838311e-01
-9.75020111e-01 -1.31593898e-01 -3.15923303e-01 6.25261664e-01
3.65426838e-01 -2.67361611e-01 8.69948938e-02 2.41704091e-01
-1.85518548e-01 7.64260054e-01 -8.55338335e-01 2.01170611e+00
-1.31333590e+00 9.14225101e-01 1.04592368e-02 -1.17970514e+00
7.15823114e-01 -1.13154076e-01 -5.44376820e-02 -8.89253259e-01
-3.11174214e-01 1.50207102e-01 2.77888551e-02 -7.18337357e-01
3.50387633e-01 -1.09872229e-01 -3.71796578e-01 8.32160115e-01
2.30290279e-01 -5.14377952e-01 2.72397041e-01 2.15374276e-01
1.60611248e+00 4.15595770e-01 1.99077353e-01 -3.36238146e-01
5.22847116e-01 5.56463182e-01 4.23499227e-01 9.49285746e-01
4.66158181e-01 -2.44008750e-01 8.27850461e-01 -5.78686416e-01
-9.29871023e-01 -6.93603933e-01 2.05044836e-01 1.75220335e+00
-5.36058545e-01 -6.95565820e-01 -6.56466007e-01 -8.25434387e-01
2.09193248e-02 1.12300241e+00 -4.41528171e-01 -4.25762415e-01
-9.93346810e-01 -4.65819031e-01 7.77601480e-01 5.30090451e-01
2.24897563e-01 -1.44792080e+00 -9.94919121e-01 1.09605208e-01
3.24227847e-02 -3.71814758e-01 -3.03560019e-01 8.98359179e-01
-1.08190715e+00 -9.57335889e-01 -1.16034858e-01 -8.94955039e-01
8.46169233e-01 -1.54918358e-01 1.63076127e+00 4.75416034e-01
-7.13983178e-01 2.14021936e-01 4.43297140e-02 -4.53447968e-01
-5.64221263e-01 4.05622482e-01 -4.56773996e-01 -1.12593627e+00
6.71783328e-01 -7.84364939e-01 -9.43619609e-02 -3.02695781e-01
-6.35755181e-01 -9.32462420e-03 8.88639569e-01 8.57625186e-01
9.31768939e-02 4.28331941e-01 2.71723777e-01 -1.28068221e+00
8.37266684e-01 -6.56924844e-01 -1.00015008e+00 2.32195079e-01
-7.59620368e-01 1.07781816e+00 9.53227520e-01 -5.55843890e-01
-1.00659657e+00 1.71666741e-01 3.53865139e-02 -3.86827618e-01
-1.59500033e-01 5.51114380e-01 -2.47086510e-02 -9.71380547e-02
1.29055953e+00 3.25005710e-01 -6.51606694e-02 -4.01052833e-01
5.22421837e-01 1.64743289e-01 6.07653737e-01 -1.47082996e+00
7.24288940e-01 -2.50938684e-01 -2.14258254e-01 -3.42517912e-01
-5.87793112e-01 -3.42067868e-01 -6.46214634e-02 3.82837802e-01
4.22671497e-01 -8.12673509e-01 -1.12539256e+00 -1.81089088e-01
-1.15945363e+00 -1.12665689e+00 -6.69484198e-01 3.72354239e-02
-5.22963822e-01 -1.82014778e-01 -3.12818706e-01 -4.94579464e-01
-4.36240405e-01 -1.47960186e+00 7.68395722e-01 1.85421124e-01
-8.43449175e-01 -9.59687114e-01 1.43916115e-01 8.46920684e-02
3.81804198e-01 -2.93179363e-01 1.61010122e+00 -6.45236909e-01
-8.88226271e-01 -2.04316899e-02 -1.82999417e-01 2.17677400e-01
-3.15029144e-01 -4.97449249e-01 -1.00127947e+00 4.18702327e-02
1.84810609e-02 -4.53751206e-01 7.67241418e-01 1.70959339e-01
1.60668230e+00 -8.79688382e-01 -4.16303217e-01 9.65819597e-01
1.36536646e+00 1.06352009e-01 4.16571051e-01 6.04231536e-01
6.15279138e-01 5.29055059e-01 1.79797158e-01 2.11184219e-01
2.77718574e-01 3.22084814e-01 7.90790543e-02 2.08797559e-01
-4.61480655e-02 -3.84179711e-01 3.09605271e-01 2.24385366e-01
7.49680698e-02 3.89613837e-01 -1.19683993e+00 5.89059472e-01
-1.62248361e+00 -7.72767484e-01 5.06273925e-01 1.92426121e+00
1.62707412e+00 1.46941498e-01 1.46000981e-01 -3.28763761e-02
1.29686117e-01 -1.24054007e-01 -6.76858246e-01 -1.03249550e+00
5.31350315e-01 1.04859769e+00 6.85681462e-01 7.66663611e-01
-5.22033989e-01 1.04417062e+00 6.68278790e+00 7.46620238e-01
-1.04287589e+00 -2.13362679e-01 2.00223327e-01 -3.18541288e-01
-6.82457387e-01 3.67184907e-01 -9.34864759e-01 1.80719063e-01
1.23798120e+00 -4.42195684e-01 1.24363637e+00 1.59159195e+00
-4.78731871e-01 -3.05621415e-01 -1.75045455e+00 6.15426123e-01
-2.93017179e-01 -1.71880722e+00 -1.39429599e-01 4.94190957e-03
5.26087642e-01 -1.17267273e-01 -3.62468772e-02 1.09821188e+00
9.79299784e-01 -1.34114218e+00 5.86398602e-01 3.09357911e-01
6.13360345e-01 -9.28927720e-01 3.70440513e-01 6.55434906e-01
-8.74443412e-01 -6.51283801e-01 -1.53889447e-01 -6.23800218e-01
-6.40488803e-01 2.51298785e-01 -1.31579792e+00 -2.12035388e-01
6.62620306e-01 5.02226174e-01 -7.53466845e-01 7.62743533e-01
-6.18272126e-01 3.72431725e-01 -2.24201083e-01 -3.49784017e-01
3.92138325e-02 2.37773880e-01 4.12739396e-01 1.42244565e+00
7.00926185e-02 1.63260430e-01 8.09835345e-02 1.60278916e+00
-1.85303807e-01 -3.48257422e-01 -7.02639401e-01 -1.25150740e-01
6.44142687e-01 1.13055253e+00 -2.85889357e-01 -4.54685986e-01
-2.67877579e-01 3.70117247e-01 5.10002673e-01 2.01864988e-01
-4.23277080e-01 -6.26598656e-01 4.84691828e-01 1.23059452e-01
2.92569190e-01 2.67496947e-02 -6.58506036e-01 -9.40362692e-01
-1.90804064e-01 -1.50507534e+00 4.42853659e-01 -6.86964571e-01
-7.05042362e-01 1.27886027e-01 4.78477269e-01 -3.62809718e-01
-5.30341268e-01 -5.01788259e-01 -8.20359826e-01 1.34239256e+00
-1.40572941e+00 -6.61085367e-01 -5.06488932e-03 2.85387188e-01
7.02402651e-01 -3.33614141e-01 1.00128508e+00 -1.12155624e-01
-4.47263032e-01 8.86749387e-01 -6.30842328e-01 -2.06696823e-01
2.56890118e-01 -1.62754989e+00 4.14850473e-01 6.23068333e-01
-2.31609166e-01 1.35203838e+00 6.11571491e-01 -4.28796142e-01
-2.06491733e+00 -7.96007037e-01 6.56763792e-01 -5.85720420e-01
7.63332844e-01 -3.35697591e-01 -8.78982544e-01 7.32611060e-01
5.03848456e-02 -3.45567912e-01 6.26617968e-01 6.85084760e-01
-7.77001441e-01 -2.22801566e-01 -9.47044730e-01 5.32589972e-01
8.48435283e-01 -7.91842043e-01 -9.24688339e-01 1.71965227e-01
1.14286447e+00 -4.74069595e-01 -7.79027104e-01 1.66470021e-01
4.83759493e-01 -5.76180875e-01 1.05054080e+00 -5.21553338e-01
8.72636795e-01 -5.12865707e-02 4.15478945e-02 -1.25593507e+00
-1.80674434e-01 -7.56349325e-01 -5.09015501e-01 8.32870603e-01
6.97023869e-01 -5.26116014e-01 9.95567262e-01 1.03620303e+00
-2.52202749e-01 -8.34879160e-01 -2.81619370e-01 -3.86941075e-01
2.81864494e-01 -6.59520268e-01 7.97146142e-01 7.99893141e-01
4.45792198e-01 6.71779931e-01 2.54618675e-01 -2.14080319e-01
4.63921785e-01 2.41563439e-01 8.91275823e-01 -1.06745124e+00
-1.11469364e+00 -7.02683330e-01 3.55585039e-01 -1.01451778e+00
6.15446329e-01 -1.56741560e+00 2.25963563e-01 -9.13901985e-01
3.74477118e-01 -7.30400383e-01 4.60192226e-02 1.01277351e+00
2.13380586e-02 -4.87861782e-01 1.10788524e-01 1.04675882e-01
-4.78901923e-01 -3.79158229e-01 1.00472212e+00 -2.93295711e-01
-5.48741996e-01 -1.81322843e-02 -1.16845071e+00 7.16346681e-01
5.92322826e-01 -5.83656311e-01 -5.93666852e-01 -4.88287330e-01
8.36556554e-01 1.94534570e-01 3.31239104e-01 -1.00297987e+00
5.25037050e-01 -2.00101569e-01 3.38266581e-01 -1.29934728e-01
-2.05820754e-01 -5.62144160e-01 -6.02003979e-03 6.36333346e-01
-9.23002481e-01 4.26735543e-02 8.62708926e-01 1.98684901e-01
2.21946016e-01 -7.99729824e-01 5.79656363e-01 -7.78729200e-01
-9.30705488e-01 -2.10100144e-01 -3.14478546e-01 2.85480559e-01
5.83828747e-01 -1.67233780e-01 -1.37907043e-01 1.14773527e-01
-5.57596743e-01 3.94178987e-01 4.63109046e-01 2.40821943e-01
2.29241163e-01 -8.21081579e-01 -1.16042823e-01 6.63264155e-01
-4.85997424e-02 4.21658546e-01 -2.03573927e-01 4.30918694e-01
-4.79614317e-01 6.04225993e-01 -1.11096807e-01 -4.75675225e-01
-1.19789708e+00 8.11110616e-01 1.18897341e-01 -5.69787323e-01
-4.55720633e-01 1.19868565e+00 2.07779527e-01 -7.32695282e-01
4.08304036e-01 -6.48974717e-01 1.73259348e-01 -1.16327733e-01
6.24812663e-01 1.46911189e-01 -5.16281910e-02 5.10131955e-01
-2.56614923e-01 3.12072963e-01 -3.85671884e-01 5.60586713e-02
1.69421327e+00 5.30479670e-01 -5.66381395e-01 4.08374295e-02
1.34052420e+00 -1.09857814e-02 -1.03717220e+00 -3.37415487e-01
2.94850379e-01 -1.49879172e-01 2.65237361e-01 -1.07605803e+00
-5.48423469e-01 7.39552438e-01 2.60035321e-02 1.14376433e-02
1.19240737e+00 1.31177872e-01 5.91450751e-01 1.02596390e+00
4.04044390e-01 -9.84326541e-01 1.29363433e-01 8.09956849e-01
5.94036520e-01 -1.11823452e+00 1.45557344e-01 1.61575503e-03
-1.93456098e-01 1.33114111e+00 8.09499383e-01 -1.87326610e-01
1.49165973e-01 8.93816531e-01 -7.64461517e-01 -2.63498873e-01
-9.42861855e-01 7.94814527e-02 1.66069880e-01 5.96352994e-01
5.08827329e-01 -1.60682783e-01 1.69466719e-01 8.13110352e-01
-6.17996693e-01 4.02944833e-01 3.05655211e-01 1.04295337e+00
-5.09799838e-01 -1.28791940e+00 -3.93692970e-01 5.64142644e-01
-3.33249897e-01 -5.00702620e-01 -1.25165850e-01 8.23513746e-01
2.90354520e-01 2.74518877e-01 -5.79687394e-02 -3.39527637e-01
2.81942636e-01 3.21920633e-01 7.08149195e-01 -1.18167734e+00
-7.05266118e-01 -4.37796563e-01 2.77307153e-01 -7.36211419e-01
1.55272231e-01 -3.28870624e-01 -1.37416136e+00 -4.92685556e-01
-1.53069139e-01 3.10563117e-01 6.23494625e-01 8.59838426e-01
3.23411316e-01 8.13902318e-01 2.23446757e-01 -7.17809677e-01
-6.86307549e-01 -4.66071278e-01 1.96891055e-01 -1.47876456e-01
5.31509101e-01 -2.04722404e-01 -3.21216643e-01 1.41890928e-01] | [8.726836204528809, 7.239196300506592] |
12455bff-b18f-4fe8-ba15-fea948347e93 | progressive-purification-for-instance | 2206.00830 | null | https://arxiv.org/abs/2206.00830v2 | https://arxiv.org/pdf/2206.00830v2.pdf | Progressive Purification for Instance-Dependent Partial Label Learning | Partial label learning (PLL) aims to train multiclass classifiers from the examples each annotated with a set of candidate labels where a fixed but unknown candidate label is correct. In the last few years, the instance-independent generation process of candidate labels has been extensively studied, on the basis of which many theoretical advances have been made in PLL. Nevertheless, the candidate labels are always instance-dependent in practice and there is no theoretical guarantee that the model trained on the instance-dependent PLL examples can converge to an ideal one. In this paper, a theoretically grounded and practically effective approach named POP, i.e. PrOgressive Purification for instance-dependent partial label learning, is proposed. Specifically, POP updates the learning model and purifies each candidate label set progressively in every epoch. Theoretically, we prove that POP enlarges the region appropriately fast where the model is reliable, and eventually approximates the Bayes optimal classifier with mild assumptions. Technically, POP is flexible with arbitrary PLL losses and could improve the performance of the previous PLL losses in the instance-dependent case. Experiments on the benchmark datasets and the real-world datasets validate the effectiveness of the proposed method. | ['Xin Geng', 'Congyu Qiao', 'Biao Liu', 'Jiaqi Lv', 'Ning Xu'] | 2022-06-02 | null | null | null | null | ['partial-label-learning'] | ['methodology'] | [ 4.61197674e-01 3.36285740e-01 -3.56477559e-01 -5.17013013e-01
-1.04812372e+00 -3.24315399e-01 2.88106203e-01 4.38149542e-01
-5.25077403e-01 1.10505879e+00 -6.41007483e-01 -1.11263909e-03
-3.90681118e-01 -7.68356919e-01 -7.40919769e-01 -1.12742984e+00
1.47014856e-01 8.20575774e-01 2.52498567e-01 3.82825315e-01
1.52927250e-01 4.02357191e-01 -1.85602582e+00 1.05678581e-01
1.06079662e+00 1.25627756e+00 7.54559636e-02 4.26759541e-01
-1.29128113e-01 7.73275673e-01 -5.87544918e-01 -5.83988786e-01
2.63703138e-01 -4.84392405e-01 -6.93925619e-01 1.55317187e-01
4.06679899e-01 2.37428352e-01 1.61442325e-01 1.14963114e+00
3.66712958e-01 -2.29341146e-02 6.69492126e-01 -1.41036117e+00
-2.87841976e-01 6.75696909e-01 -6.12969935e-01 -3.03526700e-01
-2.59714425e-01 -2.13935286e-01 1.19498253e+00 -8.47192705e-01
2.41826475e-01 9.19803381e-01 7.85887301e-01 6.49959087e-01
-1.26843989e+00 -8.15941930e-01 2.28643522e-01 5.51608726e-02
-1.44148529e+00 5.81102306e-03 8.11496437e-01 -2.89960682e-01
2.25470930e-01 1.95667580e-01 4.03521985e-01 6.92128956e-01
-1.02020882e-01 1.03124893e+00 1.29800856e+00 -6.33448958e-01
3.75204027e-01 7.22317815e-01 5.12693226e-01 6.40745759e-01
3.68099302e-01 1.30095840e-01 -3.44269425e-01 -2.93449700e-01
5.62243238e-02 -4.97913361e-02 -2.24107817e-01 -5.88739932e-01
-7.92600632e-01 7.39852011e-01 1.38802961e-01 2.75084138e-01
-1.95720851e-01 2.74683274e-02 3.51992697e-01 3.05276364e-01
7.87427604e-01 3.21555018e-01 -6.03464067e-01 2.70059764e-01
-1.02923119e+00 7.90075958e-02 8.50547731e-01 9.45037186e-01
9.49988604e-01 -2.34029040e-01 -1.58958867e-01 1.08584690e+00
8.96773189e-02 4.90497172e-01 2.98179835e-01 -7.13719368e-01
2.29673341e-01 4.82260168e-01 2.34188586e-01 -5.53147554e-01
-3.74563307e-01 -1.03969240e+00 -6.54222071e-01 6.37183487e-02
8.03473532e-01 -1.17195003e-01 -5.38539290e-01 1.98733974e+00
4.88750994e-01 4.28084701e-01 -2.14057248e-02 5.48021674e-01
2.27185771e-01 8.00342560e-01 3.66433442e-01 -5.80258965e-01
8.87543917e-01 -8.22321057e-01 -5.20986915e-01 -8.15910697e-02
6.84074104e-01 -4.71138388e-01 1.05637264e+00 8.88390958e-01
-8.10511708e-01 -6.80089891e-01 -1.03558838e+00 4.74866569e-01
3.27777937e-02 4.47054863e-01 5.75893462e-01 8.09996307e-01
-6.59452319e-01 7.37764597e-01 -4.86899406e-01 -7.88454115e-02
4.30380762e-01 4.07720983e-01 -1.63739800e-01 -2.02595219e-01
-1.10965955e+00 5.64518452e-01 7.72622228e-01 1.69823825e-01
-7.96397269e-01 -6.50255203e-01 -4.64273900e-01 1.26237288e-01
5.15827179e-01 -3.36380631e-01 1.24137211e+00 -1.37666047e+00
-1.33755922e+00 7.94308305e-01 -5.20165153e-02 -5.85705638e-01
7.78446138e-01 7.17942044e-02 -1.48318768e-01 9.05167162e-02
-6.52927719e-03 4.90888834e-01 9.63845015e-01 -1.42806697e+00
-1.32266593e+00 -3.03895652e-01 4.10751477e-02 4.07251939e-02
-5.19033730e-01 -3.27439219e-01 -4.34298106e-02 -2.56100982e-01
1.36635020e-01 -1.05729496e+00 -5.97219095e-02 -9.53019038e-02
-2.01265559e-01 -7.24410355e-01 3.88077915e-01 -1.94642141e-01
1.05004847e+00 -2.19334626e+00 -1.34700030e-01 3.50354463e-01
-6.57703876e-02 2.90297538e-01 -5.39448531e-03 2.19170719e-01
-9.11750644e-02 3.80126126e-02 -5.22677898e-01 -5.54587364e-01
5.48089854e-02 2.33435452e-01 -4.01421756e-01 6.42722070e-01
1.10958949e-01 3.67231220e-01 -1.08329284e+00 -5.58864117e-01
2.78383903e-02 3.13981354e-01 -4.44221944e-01 1.82535857e-01
-3.63730758e-01 3.73294353e-01 -5.52819788e-01 4.71090853e-01
6.99908674e-01 -3.44502270e-01 1.91821501e-01 1.78899601e-01
6.38074726e-02 7.61052081e-03 -1.28326237e+00 1.10283494e+00
-6.33898199e-01 2.52291679e-01 -1.23720475e-01 -1.34661710e+00
1.05921113e+00 3.83881152e-01 5.45981646e-01 -3.18382323e-01
1.94448314e-03 5.31181633e-01 -3.19976628e-01 -1.45179585e-01
1.54179230e-01 -7.24325120e-01 1.98890362e-02 3.13449502e-01
1.50966734e-01 8.51766691e-02 2.06681162e-01 -1.66529179e-01
7.10606277e-01 9.72622856e-02 2.14831874e-01 -4.41507846e-02
8.67019057e-01 -2.03790516e-01 7.89410472e-01 9.32772815e-01
-2.63348430e-01 4.42920595e-01 4.92803037e-01 -2.59480059e-01
-7.71552861e-01 -1.00123692e+00 -6.89131081e-01 9.67798471e-01
2.26236135e-01 -9.50510055e-02 -6.82038069e-01 -1.07690620e+00
-5.19220792e-02 9.25868928e-01 -4.73302484e-01 -1.78137779e-01
-4.54894155e-01 -1.06603801e+00 9.65781063e-02 1.37864292e-01
3.53627592e-01 -1.16653109e+00 -1.92686871e-01 4.53974217e-01
-1.15459234e-01 -6.91173315e-01 -9.59052797e-03 4.67711717e-01
-8.69713426e-01 -1.17594171e+00 -6.17138088e-01 -9.18523490e-01
9.06707883e-01 -3.19489487e-03 1.00766170e+00 1.22853406e-01
-1.06502781e-02 1.59381881e-01 -3.84509176e-01 -3.63254875e-01
-5.21338046e-01 1.89603299e-01 -1.20680053e-02 5.36684334e-01
3.21975589e-01 -4.38386679e-01 -3.05247396e-01 3.30453873e-01
-7.58630872e-01 -2.01010630e-01 6.56232715e-01 1.16905463e+00
9.90486622e-01 5.18882394e-01 1.16265690e+00 -1.31405044e+00
2.55895138e-01 -5.18295765e-01 -7.78631091e-01 5.27658045e-01
-8.80745351e-01 9.75081623e-02 9.93881822e-01 -4.41478789e-01
-9.86738980e-01 2.54871905e-01 -1.11324362e-01 -1.93570331e-01
-1.13908477e-01 4.35903698e-01 -2.29386270e-01 1.71837397e-02
6.00735903e-01 1.99447870e-01 -2.51843035e-01 -4.82727408e-01
1.62045807e-01 6.82930112e-01 3.17224026e-01 -8.71997178e-01
7.03132212e-01 3.68462443e-01 1.16453610e-01 -5.58610380e-01
-1.36242604e+00 -4.78881747e-01 -5.84345281e-01 -3.16868514e-01
2.48913795e-01 -5.87651968e-01 -6.52950525e-01 4.85702783e-01
-9.19113100e-01 -3.44036311e-01 -5.22631109e-01 5.07365048e-01
-6.00085139e-01 3.58606368e-01 -4.14756268e-01 -1.27937472e+00
-1.68255016e-01 -9.59254980e-01 7.94634700e-01 3.35702211e-01
7.33533874e-02 -9.06215727e-01 -4.29236069e-02 4.51675989e-02
-1.98122337e-01 9.53999981e-02 9.43033159e-01 -9.20824349e-01
-5.49364209e-01 -5.72233737e-01 1.26929665e-02 8.04766774e-01
-2.36028526e-02 -8.24926123e-02 -1.12967193e+00 -5.21632254e-01
1.10762633e-01 -5.16484678e-01 9.78342235e-01 3.86197031e-01
1.26622307e+00 -1.07799508e-01 -4.32626396e-01 1.39664322e-01
1.56766903e+00 2.62164831e-01 2.06843168e-01 8.71246532e-02
2.83925563e-01 7.40024865e-01 1.06310308e+00 4.40897346e-01
-3.72717157e-02 4.80559945e-01 3.40646178e-01 1.41011268e-01
1.56643301e-01 -2.09691182e-01 1.19881831e-01 6.62323654e-01
3.02540362e-01 -4.95874077e-01 -6.21691525e-01 6.05585754e-01
-1.76342857e+00 -7.64945865e-01 1.77238032e-03 2.65332675e+00
9.93405759e-01 3.86657059e-01 -9.24596377e-03 5.47917366e-01
1.00969923e+00 -3.05872440e-01 -5.76420009e-01 -1.16098978e-01
-1.00839004e-01 3.37006003e-01 3.55306923e-01 5.52726269e-01
-1.26368237e+00 5.83820581e-01 5.25535297e+00 1.13792694e+00
-1.00177038e+00 2.37835154e-01 8.69861960e-01 5.38419932e-04
-7.83476681e-02 3.57256904e-02 -1.21642387e+00 5.88517725e-01
1.00477409e+00 -2.15114579e-01 9.42051560e-02 1.17202830e+00
-1.41810909e-01 -1.33052856e-01 -1.17922986e+00 7.58302093e-01
-2.85663679e-02 -8.24426353e-01 -1.49533615e-01 -6.14230074e-02
7.33292282e-01 -4.17975157e-01 2.17943013e-01 5.67573249e-01
1.68707415e-01 -5.21189511e-01 8.19704235e-01 4.30806965e-01
7.29919970e-01 -1.30252564e+00 9.16371226e-01 8.11164558e-01
-8.69547963e-01 -4.61169511e-01 -5.32182574e-01 2.68563032e-01
-3.80761776e-05 1.00900531e+00 -8.06483269e-01 5.70712864e-01
3.02148312e-01 6.82209074e-01 -4.57633734e-01 1.29576683e+00
-3.96698833e-01 1.11659098e+00 -5.37854552e-01 -1.29578531e-01
2.67709017e-01 -2.82746851e-01 2.82357752e-01 1.00364673e+00
3.38518590e-01 -6.86798245e-02 3.12397599e-01 4.77217823e-01
-3.79228666e-02 4.32818085e-01 -2.07415417e-01 3.05464536e-01
5.18035650e-01 1.24153328e+00 -8.02026391e-01 -2.77885914e-01
-2.29200184e-01 5.05479634e-01 6.10535800e-01 1.25445113e-01
-8.95508051e-01 -3.13629895e-01 3.06121688e-02 -1.22548034e-02
3.37335676e-01 4.70553935e-01 -1.63696632e-01 -7.58971632e-01
1.83750629e-01 -5.37553489e-01 6.83562994e-01 -2.53829032e-01
-1.65705144e+00 6.09342098e-01 -1.52361169e-01 -1.36250246e+00
-5.94520457e-02 -4.59786087e-01 -2.54387945e-01 7.44806409e-01
-1.66427946e+00 -7.10771859e-01 -1.16494000e-02 1.39406487e-01
3.56343836e-01 1.19581647e-01 6.75451458e-01 4.48458046e-01
-5.61346650e-01 8.30969155e-01 4.16918546e-01 -2.41433948e-01
7.60632634e-01 -1.36837816e+00 -3.81134152e-01 6.17671132e-01
2.49364898e-01 2.61437953e-01 7.61257827e-01 -3.65787148e-01
-5.90797365e-01 -1.30186176e+00 1.09259164e+00 -1.59885362e-01
4.95549142e-01 -2.94324011e-01 -1.08752179e+00 3.99828792e-01
-3.37552994e-01 2.13948756e-01 6.40738189e-01 8.41699988e-02
-1.45532712e-01 -6.39711678e-01 -1.25747871e+00 3.06860983e-01
6.67842031e-01 -2.43396312e-01 -3.27211529e-01 6.54603362e-01
3.87088478e-01 -5.33842146e-02 -7.20562577e-01 5.82138777e-01
4.16236788e-01 -7.94156492e-01 7.72234678e-01 -5.02823353e-01
6.59326315e-02 -3.96154970e-01 9.37353596e-02 -1.19465256e+00
-7.55909383e-02 -3.95096362e-01 -1.04829259e-02 1.42434549e+00
6.37460232e-01 -8.17490518e-01 8.40901077e-01 4.96516287e-01
-1.06709845e-01 -1.01064920e+00 -1.10995245e+00 -9.02312994e-01
1.91106811e-01 -4.84664381e-01 4.06151742e-01 7.30760217e-01
-9.64071080e-02 9.18044895e-02 -4.98310745e-01 2.49500677e-01
9.22481000e-01 3.77130300e-01 3.93495977e-01 -1.59488058e+00
-5.14227808e-01 -3.46947759e-01 -1.68789238e-01 -1.16958714e+00
5.52530825e-01 -1.11192369e+00 3.07501256e-01 -1.25497139e+00
2.87751436e-01 -1.23673499e+00 -8.46920609e-01 4.86174077e-01
-5.02489507e-01 1.52400643e-01 1.44675642e-01 2.45007083e-01
-7.06005931e-01 5.33263445e-01 1.09354579e+00 -1.66841835e-01
-6.75549880e-02 6.27667487e-01 -5.86397350e-01 5.64622641e-01
8.29811871e-01 -1.04211390e+00 -6.17048800e-01 2.64931470e-01
2.32282817e-01 -4.38410491e-02 5.91060594e-02 -1.07285225e+00
-5.57497367e-02 1.69426445e-02 5.51120825e-02 -5.78999996e-01
2.16984510e-01 -9.66873884e-01 5.50751612e-02 5.04726052e-01
-8.69371235e-01 -7.81115651e-01 -2.38339663e-01 8.95327151e-01
-1.25711009e-01 -8.35062325e-01 1.15920293e+00 1.48629770e-01
-3.10901016e-01 3.91770214e-01 9.65169743e-02 1.78311124e-01
1.09748638e+00 -8.65567662e-03 9.06504840e-02 -1.14488997e-01
-7.67532468e-01 1.65692315e-01 1.95845857e-01 -1.01510277e-02
2.05735981e-01 -1.14916158e+00 -8.51539731e-01 -9.10402387e-02
9.10622329e-02 1.77031644e-02 1.61605775e-01 7.51995504e-01
7.80821033e-03 4.43285584e-01 3.00401539e-01 -7.11533010e-01
-1.02037930e+00 8.54173958e-01 3.03788006e-01 -5.25323749e-01
-4.88145173e-01 8.89221847e-01 3.56444359e-01 -4.60954100e-01
4.35634255e-01 -1.40874460e-01 -2.69099742e-01 5.00303768e-02
5.23157179e-01 2.93336987e-01 1.37704894e-01 -3.26428413e-01
-5.92117906e-02 4.92432773e-01 -8.49135220e-02 1.22204609e-01
1.17805851e+00 -7.09167048e-02 -2.60602217e-02 7.84540296e-01
1.21811616e+00 -1.30731925e-01 -1.52276957e+00 -4.08237129e-01
2.77340889e-01 -3.10789406e-01 -1.26777068e-01 -7.67713130e-01
-1.16217899e+00 8.89124393e-01 4.68459755e-01 4.12223488e-01
1.21250641e+00 -1.45235613e-01 5.45107782e-01 3.65186721e-01
7.61446238e-01 -1.07092500e+00 -1.32794499e-01 1.50765643e-01
4.66016591e-01 -9.82433975e-01 -1.69372588e-01 -5.09180725e-01
-5.07925630e-01 1.01364148e+00 5.96485138e-01 -4.28724922e-02
5.38604438e-01 -6.06169961e-02 -3.01511407e-01 3.83559644e-01
-8.31816971e-01 2.56303884e-02 1.46629006e-01 3.69525701e-01
2.56762236e-01 1.42469302e-01 -6.20393097e-01 7.27492332e-01
5.56608737e-02 1.09284837e-02 2.98702121e-01 5.41461468e-01
-5.60658097e-01 -1.29663980e+00 -2.92892158e-01 5.01383126e-01
-4.03439045e-01 9.17766690e-02 3.79004888e-02 8.11250508e-01
4.16247159e-01 9.40102041e-01 -1.47613019e-01 6.01835968e-03
2.01247066e-01 4.05683041e-01 6.24890447e-01 -6.37440205e-01
-4.13464040e-01 -6.68278933e-02 3.39302756e-02 -1.48214430e-01
-4.89436239e-01 -7.93351948e-01 -1.36475325e+00 1.46130651e-01
-6.99957430e-01 6.87373400e-01 4.82712895e-01 9.54653502e-01
-7.70888180e-02 1.66996986e-01 1.05336320e+00 -3.46249789e-01
-8.93928885e-01 -8.46590638e-01 -9.27190483e-01 2.15751901e-01
1.57745168e-01 -8.31628501e-01 -8.61537993e-01 -6.37738928e-02] | [9.259793281555176, 4.068922519683838] |
3f333000-5331-4926-9730-b63e38314940 | simultaneous-face-detection-and-360-degree | 2111.11604 | null | https://arxiv.org/abs/2111.11604v1 | https://arxiv.org/pdf/2111.11604v1.pdf | Simultaneous face detection and 360 degree headpose estimation | With many practical applications in human life, including manufacturing surveillance cameras, analyzing and processing customer behavior, many researchers are noticing face detection and head pose estimation on digital images. A large number of proposed deep learning models have state-of-the-art accuracy such as YOLO, SSD, MTCNN, solving the problem of face detection or HopeNet, FSA-Net, RankPose model used for head pose estimation problem. According to many state-of-the-art methods, the pipeline of this task consists of two parts, from face detection to head pose estimation. These two steps are completely independent and do not share information. This makes the model clear in setup but does not leverage most of the featured resources extracted in each model. In this paper, we proposed the Multitask-Net model with the motivation to leverage the features extracted from the face detection model, sharing them with the head pose estimation branch to improve accuracy. Also, with the variety of data, the Euler angle domain representing the face is large, our model can predict with results in the 360 Euler angle domain. Applying the multitask learning method, the Multitask-Net model can simultaneously predict the position and direction of the human head. To increase the ability to predict the head direction of the model, we change there presentation of the human face from the Euler angle to vectors of the Rotation matrix. | ['Long Tran Quoc', 'Duc Tran Minh', 'Tuan Nguyen Dinh', 'Linh Nguyen Viet', 'Hoang Nguyen Viet'] | 2021-11-23 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-4.18348372e-01 5.99229597e-02 -4.57989424e-03 -3.94128621e-01
-3.60542208e-01 -3.81505370e-01 2.35281169e-01 -6.46058857e-01
-2.82735199e-01 2.30381295e-01 -2.10256688e-02 6.35487884e-02
8.04559048e-03 -3.57288033e-01 -6.14418864e-01 -7.56879926e-01
-1.89274014e-03 6.54650927e-01 1.28672749e-01 -9.75375529e-03
1.33450016e-01 8.35913658e-01 -1.84230483e+00 2.49703005e-01
5.59258349e-02 1.21884632e+00 1.17536128e-01 5.74661672e-01
2.95218736e-01 5.01631975e-01 -5.99361420e-01 -3.57571423e-01
3.92657906e-01 2.07337946e-01 -4.72488016e-01 1.92477182e-01
7.82934248e-01 -8.50683391e-01 -3.30643535e-01 7.47291505e-01
1.08447492e+00 -1.85657680e-01 4.40568924e-01 -1.69524181e+00
-3.84022370e-02 1.40536115e-01 -9.49493527e-01 -1.09107755e-01
4.13889855e-01 1.54798746e-01 6.09096766e-01 -1.19696772e+00
4.84424978e-01 1.56275833e+00 7.96768248e-01 4.55027193e-01
-3.70420784e-01 -9.94922936e-01 2.00571597e-01 4.03390765e-01
-1.49254322e+00 -6.77507043e-01 6.93591416e-01 -4.44695830e-01
7.85341442e-01 -7.76447430e-02 4.80587542e-01 1.05270696e+00
1.32326066e-01 7.81466722e-01 9.53467011e-01 -1.82468235e-01
-1.57025710e-01 6.85138255e-02 1.01242416e-01 9.20268416e-01
1.54030025e-01 -2.27168668e-02 -8.90460789e-01 9.29884389e-02
8.24635804e-01 3.01909178e-01 -2.99003460e-02 -4.27349478e-01
-7.33005762e-01 8.67046893e-01 1.06244206e-01 -8.91053975e-02
-2.45687008e-01 -9.82588977e-02 4.03354675e-01 -4.55335602e-02
3.75551879e-01 -2.26763234e-01 -8.02490294e-01 1.18491963e-01
-1.01722550e+00 3.59340459e-01 1.02107048e+00 1.05194414e+00
6.10434055e-01 -4.40894961e-02 -1.92330062e-01 5.51631749e-01
6.60672724e-01 6.54014647e-01 4.68509078e-01 -7.52656221e-01
3.18596542e-01 5.05788684e-01 -9.75212157e-02 -8.93280923e-01
-1.09786880e+00 -3.89207751e-01 -5.79158008e-01 2.58226961e-01
5.41067898e-01 -5.33759117e-01 -9.58368242e-01 1.39420271e+00
7.94361055e-01 1.42965689e-01 -3.50429565e-01 1.03869092e+00
1.03558540e+00 3.98292959e-01 -3.28551084e-01 -8.01225230e-02
1.67524314e+00 -1.13829815e+00 -6.84291244e-01 -5.68137825e-01
5.68494380e-01 -1.01127625e+00 5.11310518e-01 5.88859856e-01
-8.26484382e-01 -6.53971493e-01 -9.61604595e-01 -2.05216169e-01
-3.31002235e-01 6.13269567e-01 6.15197420e-01 6.81432128e-01
-1.09509122e+00 2.68252492e-01 -6.27102673e-01 -4.24396634e-01
3.20467114e-01 7.71303713e-01 -5.05851507e-01 -2.72094220e-01
-8.70563388e-01 1.08478928e+00 1.57443225e-01 3.64052981e-01
-1.00589979e+00 -4.24781650e-01 -8.68839502e-01 7.77495876e-02
6.14923537e-01 -6.06099904e-01 1.40547776e+00 -6.24220371e-01
-1.54299843e+00 8.78422439e-01 -2.75228322e-01 -1.46264210e-01
5.91680884e-01 -5.34806132e-01 -3.16064090e-01 -6.54323846e-02
-5.77070192e-02 7.83134341e-01 1.32817209e+00 -9.65535045e-01
-6.85818136e-01 -1.01850295e+00 -1.87053353e-01 1.28652185e-01
-2.95008957e-01 3.45169514e-01 -7.52850294e-01 -8.34914073e-02
3.78212705e-02 -1.14196301e+00 1.83583140e-01 2.20408916e-01
-4.35837895e-01 -4.74448323e-01 1.40161526e+00 -8.74723852e-01
1.06038916e+00 -2.31256223e+00 -2.19785333e-01 2.77276873e-03
3.15254956e-01 3.07527095e-01 2.19114944e-02 1.41062662e-01
-1.52196914e-01 -4.18586999e-01 4.79001939e-01 -6.98795199e-01
8.05078447e-02 4.96391430e-02 1.18319742e-01 8.89964044e-01
2.73897145e-02 7.31899440e-01 -3.48365098e-01 -5.04587948e-01
2.09253833e-01 6.44006789e-01 -5.16239822e-01 3.44161540e-01
1.99651882e-01 1.66000545e-01 -3.43014926e-01 8.21687758e-01
1.03119373e+00 -1.00145802e-01 2.75197685e-01 -5.49820006e-01
-1.15655221e-01 -5.09173758e-02 -1.56394887e+00 1.37733746e+00
-3.78532946e-01 4.56245810e-01 5.13621092e-01 -6.42030001e-01
7.61908293e-01 4.44945961e-01 5.40182889e-01 -4.46173817e-01
3.65087241e-01 -2.37573963e-02 1.28823087e-01 -7.35795200e-01
1.94690347e-01 1.02318317e-01 2.95422524e-01 2.75376916e-01
3.14500570e-01 1.58986583e-01 -1.80355776e-02 -2.74184555e-01
7.00115681e-01 1.32607952e-01 2.54671216e-01 9.37309954e-03
3.84393275e-01 -5.08526266e-01 6.10342264e-01 3.32654387e-01
-3.29532623e-01 7.19143450e-01 4.51424152e-01 -5.78984439e-01
-7.78568268e-01 -6.41327918e-01 -1.46353133e-02 1.56392324e+00
-3.64966542e-01 -2.61395156e-01 -1.09132075e+00 -7.45685756e-01
2.66932547e-01 2.10668668e-01 -6.65790319e-01 9.35105458e-02
-8.53469670e-01 -6.02265596e-01 4.02148515e-01 6.42464161e-01
4.16488379e-01 -1.04221332e+00 -7.02722013e-01 -1.89510435e-01
7.60359764e-02 -1.38919330e+00 -5.40068507e-01 1.39530271e-01
-5.34122288e-01 -1.32305717e+00 -7.47993350e-01 -7.83786893e-01
5.58255017e-01 1.87042817e-01 7.94795275e-01 -7.98969865e-02
-4.31146175e-01 2.36964971e-01 -1.57140102e-02 -8.06637764e-01
3.11448306e-01 3.36031169e-02 3.77449512e-01 1.19890191e-01
4.27967489e-01 -1.92079499e-01 -6.68380141e-01 4.07695979e-01
-3.93658936e-01 -2.32899770e-01 6.39336944e-01 4.27865326e-01
1.58339053e-01 -1.21120080e-01 4.07977968e-01 -6.77955449e-01
1.54898286e-01 -3.38722050e-01 -5.66725671e-01 8.87817517e-02
-4.19573218e-01 -2.66966254e-01 4.14286852e-01 -5.01228333e-01
-8.16537261e-01 4.49530929e-01 -1.34277537e-01 -7.55901515e-01
-2.71206498e-01 2.08844125e-01 -3.85901123e-01 1.00873958e-03
2.49795184e-01 -1.03725173e-01 3.01092833e-01 -5.41295111e-01
1.63745865e-01 6.83746397e-01 3.29926729e-01 3.87139991e-03
5.66381633e-01 4.47186530e-01 3.36309016e-01 -8.13309193e-01
-9.94150937e-01 -5.91417372e-01 -9.28612292e-01 -3.85086477e-01
8.69642437e-01 -9.96955872e-01 -1.32924354e+00 9.49249506e-01
-1.44179475e+00 7.67260343e-02 3.49276900e-01 4.33536559e-01
-2.92605609e-01 1.58265412e-01 -5.56052387e-01 -9.39658225e-01
-5.00421882e-01 -1.33614349e+00 1.56358755e+00 2.20196992e-01
6.63919896e-02 -6.73403263e-01 -4.47073936e-01 4.27997380e-01
3.15472305e-01 -9.37819555e-02 5.07892430e-01 -7.90395021e-01
-3.90983820e-01 -3.95260721e-01 -2.04889521e-01 6.59141019e-02
-1.34307534e-01 3.81407663e-02 -1.41637540e+00 -6.32003725e-01
2.33198330e-01 -2.38110453e-01 5.55315793e-01 7.62430847e-01
1.19890547e+00 -1.89349875e-01 -4.78134364e-01 6.60734296e-01
7.80177653e-01 -2.36530434e-02 3.10096651e-01 1.70765698e-01
9.99399722e-01 8.80625188e-01 5.19418180e-01 6.60068512e-01
6.60924315e-01 9.38529074e-01 7.57217467e-01 -9.59091187e-02
-1.31224990e-01 3.92423663e-03 6.24464869e-01 5.40519476e-01
2.46527255e-01 -9.23588797e-02 -8.47594798e-01 9.28755924e-02
-1.78515553e+00 -5.76190352e-01 -1.93425156e-02 2.10433435e+00
2.60171205e-01 -1.15263099e-02 4.96053934e-01 8.61507207e-02
8.10043573e-01 2.38922387e-02 -6.36305511e-01 -8.48826990e-02
2.71059722e-01 -1.48190722e-01 6.13396347e-01 2.22916991e-01
-1.10226870e+00 8.31143379e-01 6.51938915e+00 5.38872659e-01
-1.52211630e+00 1.97433084e-01 4.27962333e-01 -4.01166081e-01
6.12133205e-01 -5.09801865e-01 -1.60904610e+00 2.84626156e-01
8.10959458e-01 3.78278404e-01 3.44496906e-01 1.13132703e+00
1.68738961e-01 -7.58081377e-02 -1.40399837e+00 1.31579018e+00
6.01927996e-01 -6.55804932e-01 -2.84531802e-01 2.15423569e-01
1.98804140e-01 3.95446718e-02 3.07600439e-01 4.82047349e-01
-6.57427311e-02 -1.13303101e+00 7.21506178e-01 3.64005804e-01
7.29065597e-01 -7.18725622e-01 8.31957757e-01 5.76708555e-01
-1.09742606e+00 -5.42757452e-01 -1.84604108e-01 -2.05268972e-02
2.92378038e-01 5.56464970e-01 -1.06168163e+00 2.85509199e-01
8.28575194e-01 4.11681175e-01 -5.98064601e-01 8.57685566e-01
-2.04178020e-01 2.41945848e-01 -4.24389243e-01 2.90760189e-01
-1.46087289e-01 1.38932779e-01 2.28863180e-01 9.47678626e-01
4.89364892e-01 -1.19897947e-01 4.82532561e-01 2.67331094e-01
-4.44687679e-02 -1.30901918e-01 -5.91544211e-01 4.37729090e-01
3.27689707e-01 1.61482561e+00 -4.73354042e-01 3.04362364e-02
-6.53768480e-01 6.87585533e-01 1.43930420e-01 9.31945890e-02
-8.52647841e-01 6.95082247e-02 5.73206961e-01 4.05577421e-01
5.02958536e-01 -1.51514471e-01 -1.66092515e-01 -6.44721329e-01
1.56251088e-01 -9.58182931e-01 2.69969404e-01 -6.94455802e-01
-9.75922942e-01 4.77880239e-01 7.07601383e-02 -8.70969653e-01
-4.59356934e-01 -1.20782447e+00 -6.08212054e-01 7.61068225e-01
-1.42929447e+00 -1.65532506e+00 -4.39981848e-01 6.54347479e-01
6.92495465e-01 -2.81081766e-01 5.25996149e-01 5.58053136e-01
-9.12490964e-01 8.15585256e-01 -4.31807131e-01 3.40967983e-01
1.05650783e+00 -8.54644358e-01 2.96666473e-01 4.90184009e-01
-2.81104803e-01 6.42596602e-01 5.41453362e-01 -4.13986862e-01
-1.68619478e+00 -9.33057129e-01 6.88927889e-01 -4.55303490e-01
3.10332745e-01 -5.88904321e-01 -4.54269677e-01 8.42338860e-01
-3.54513079e-02 3.71699065e-01 4.35999721e-01 2.69869268e-01
-1.53231248e-01 -3.63382488e-01 -1.24208450e+00 1.58241495e-01
8.33740592e-01 -3.87731075e-01 -1.55476198e-01 4.05996025e-01
3.13042134e-01 -6.49417698e-01 -5.36823452e-01 4.21194911e-01
8.98583949e-01 -1.02060437e+00 9.63445961e-01 -3.45373601e-01
1.52200028e-01 -2.38418337e-02 8.46621813e-04 -1.10759187e+00
-2.91721255e-01 -2.88840383e-01 -5.91446757e-01 1.10425878e+00
1.47803545e-01 -5.15877008e-01 9.90459263e-01 3.82608563e-01
-1.01336502e-01 -9.02529836e-01 -1.06574118e+00 -4.59941089e-01
-2.87306905e-01 -1.85201630e-01 6.91327989e-01 5.41359067e-01
-4.50834543e-01 5.86524546e-01 -4.96063232e-01 4.24361169e-01
4.56695884e-01 -1.19792268e-01 9.50683594e-01 -1.57385767e+00
-1.75842673e-01 -2.08741114e-01 -2.03917712e-01 -1.09638178e+00
2.55846024e-01 -3.85812849e-01 -1.01825766e-01 -1.25484240e+00
2.93039769e-01 1.47871226e-01 7.74969831e-02 6.87043011e-01
7.92267099e-02 4.39756997e-02 4.25605297e-01 2.05112025e-02
-4.48487818e-01 1.88750476e-01 1.17335773e+00 1.48163557e-01
3.46648432e-02 2.50106812e-01 -4.06336159e-01 1.15660369e+00
4.11216527e-01 -3.68002445e-01 -2.03087360e-01 -5.27880371e-01
-2.69279419e-03 1.71062931e-01 4.70685452e-01 -9.71317172e-01
5.45863032e-01 2.44242653e-01 8.67856562e-01 -9.21728015e-01
7.23797739e-01 -9.73908663e-01 -1.11230791e-01 4.40773487e-01
2.73118824e-01 3.59204739e-01 2.22247541e-01 -2.24017464e-02
-4.23345044e-02 -2.26675928e-01 8.18727672e-01 -9.32465419e-02
-6.15474999e-01 5.38845897e-01 -1.31473243e-01 -3.47806841e-01
1.05104518e+00 -3.06810975e-01 -2.64272749e-01 -4.40016329e-01
-6.88276291e-01 2.89227962e-01 8.13344494e-02 7.03364193e-01
4.04208988e-01 -1.03622484e+00 -6.38348341e-01 4.49810445e-01
-1.63238466e-01 1.39849335e-01 3.18586975e-01 9.93145585e-01
-1.97464913e-01 6.02022767e-01 -2.31003240e-01 -7.35569537e-01
-1.59477890e+00 5.73032677e-01 2.90224254e-01 -1.67209566e-01
-1.72727659e-01 8.62346709e-01 3.73346239e-01 -6.62333369e-01
4.36099172e-01 -4.73404750e-02 -4.96064216e-01 5.67718089e-01
6.12176359e-01 6.83785141e-01 4.29721832e-01 -8.51664245e-01
-5.57027102e-01 6.89648271e-01 -3.55817527e-01 1.89450979e-01
1.37938607e+00 -7.85944685e-02 -7.80437142e-02 4.94093746e-02
1.25416028e+00 -1.69544563e-01 -1.32298231e+00 1.44649148e-01
-2.94898748e-01 -2.32816145e-01 2.39698350e-01 -7.25919545e-01
-1.43152511e+00 1.10838294e+00 1.12176681e+00 -1.63268551e-01
1.00281453e+00 -1.19811259e-01 5.88638783e-01 2.63444632e-01
4.27819848e-01 -1.24920523e+00 2.44353652e-01 7.32055783e-01
8.42460752e-01 -1.45982087e+00 1.54143572e-01 -5.83309650e-01
-5.70469797e-01 1.16023612e+00 1.04356873e+00 2.87103444e-01
7.63876677e-01 5.70428908e-01 1.37446553e-01 -4.02779788e-01
-4.91840273e-01 -3.62852551e-02 3.27098846e-01 6.34741247e-01
5.62618017e-01 -7.65196830e-02 2.35463843e-01 6.36260092e-01
-2.03867897e-01 -2.14361846e-02 1.26832202e-01 7.49197483e-01
-4.67522740e-01 -8.62737298e-01 -7.38658190e-01 4.00400072e-01
-5.73461473e-01 1.48503959e-01 -1.78438380e-01 8.49934220e-01
3.16155553e-01 8.50042760e-01 1.08683214e-01 -2.75090307e-01
4.59082931e-01 1.72319442e-01 4.25894231e-01 -6.90608621e-01
-4.91607130e-01 2.48496547e-01 -1.07297748e-01 -6.46679401e-01
1.64130837e-01 -7.77484596e-01 -1.10132897e+00 -3.48364025e-01
-5.10692537e-01 -3.92650664e-01 8.25366199e-01 1.18162680e+00
3.57572109e-01 3.84259820e-01 7.10757196e-01 -1.53619862e+00
-7.90870965e-01 -1.29723465e+00 -7.30403900e-01 2.62208451e-02
3.98438483e-01 -1.02688932e+00 -2.41845414e-01 -1.79554567e-01] | [13.63985538482666, 0.3038301467895508] |
e6445bba-9317-4f5c-a1fc-edb346d8f9a4 | learning-rigidity-in-dynamic-scenes-with-a | 1804.04259 | null | http://arxiv.org/abs/1804.04259v2 | http://arxiv.org/pdf/1804.04259v2.pdf | Learning Rigidity in Dynamic Scenes with a Moving Camera for 3D Motion Field Estimation | Estimation of 3D motion in a dynamic scene from a temporal pair of images is
a core task in many scene understanding problems. In real world applications, a
dynamic scene is commonly captured by a moving camera (i.e., panning, tilting
or hand-held), increasing the task complexity because the scene is observed
from different view points. The main challenge is the disambiguation of the
camera motion from scene motion, which becomes more difficult as the amount of
rigidity observed decreases, even with successful estimation of 2D image
correspondences. Compared to other state-of-the-art 3D scene flow estimation
methods, in this paper we propose to \emph{learn} the rigidity of a scene in a
supervised manner from a large collection of dynamic scene data, and directly
infer a rigidity mask from two sequential images with depths. With the learned
network, we show how we can effectively estimate camera motion and projected
scene flow using computed 2D optical flow and the inferred rigidity mask. For
training and testing the rigidity network, we also provide a new semi-synthetic
dynamic scene dataset (synthetic foreground objects with a real background) and
an evaluation split that accounts for the percentage of observed non-rigid
pixels. Through our evaluation we show the proposed framework outperforms
current state-of-the-art scene flow estimation methods in challenging dynamic
scenes. | ['Zhaoyang Lv', 'Deqing Sun', 'Jan Kautz', 'James M. Rehg', 'Kihwan Kim', 'Alejandro Troccoli'] | 2018-04-12 | learning-rigidity-in-dynamic-scenes-with-a-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Zhaoyang_Lv_Learning_Rigidity_in_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Zhaoyang_Lv_Learning_Rigidity_in_ECCV_2018_paper.pdf | eccv-2018-9 | ['scene-flow-estimation'] | ['computer-vision'] | [ 5.23617566e-01 -2.17618212e-01 1.00189127e-01 -4.73964997e-02
-1.93867654e-01 -8.26814473e-01 4.71617579e-01 -4.90403414e-01
-4.19336706e-01 2.94224292e-01 1.39639452e-01 -4.38004695e-02
5.54718226e-02 -5.49589038e-01 -7.82603264e-01 -6.57469034e-01
5.36719598e-02 5.12469351e-01 6.25001490e-01 6.21845163e-02
3.91593397e-01 5.91978967e-01 -1.42294300e+00 4.60423566e-02
4.42612469e-01 7.46822655e-01 6.36412680e-01 1.08692050e+00
1.90072611e-01 9.77737725e-01 -2.01492906e-01 -1.27419069e-01
7.48705924e-01 -3.05967152e-01 -9.67214942e-01 5.84738135e-01
9.09727395e-01 -8.05074513e-01 -4.76968199e-01 1.00316298e+00
9.37338620e-02 4.94471282e-01 4.03576106e-01 -7.96321571e-01
1.94756523e-01 1.45667598e-01 -7.87234128e-01 5.29960871e-01
6.14268243e-01 2.53668964e-01 6.35062277e-01 -6.62294090e-01
1.04492545e+00 1.16522050e+00 2.76822418e-01 4.40271318e-01
-1.29837096e+00 -3.26637626e-01 3.40569943e-01 2.23544970e-01
-1.12213683e+00 -5.54293036e-01 1.19592154e+00 -7.95283198e-01
7.32382715e-01 1.56594187e-01 7.58291960e-01 1.08033347e+00
-1.03661872e-01 4.99542683e-01 8.09676528e-01 -2.44236723e-01
2.19150305e-01 -1.67095512e-01 -1.18737854e-01 6.77115798e-01
1.66620806e-01 1.55443624e-01 -5.12596548e-01 2.28726804e-01
1.20365143e+00 -1.66569576e-02 -6.19609475e-01 -7.28909791e-01
-1.50143540e+00 2.80872136e-01 9.10521448e-02 1.07715860e-01
-1.40867159e-01 1.50554731e-01 1.98186800e-01 -3.29335639e-03
3.51443291e-01 1.95096999e-01 -3.60530198e-01 -4.54057515e-01
-8.70808184e-01 2.70397305e-01 8.22592735e-01 6.60789847e-01
7.58314371e-01 9.75514799e-02 4.31147069e-01 4.24195647e-01
8.68627951e-02 5.27797639e-01 3.61114055e-01 -1.49839139e+00
6.14132226e-01 4.35627103e-01 3.55220020e-01 -1.27165461e+00
-3.29645902e-01 1.96959022e-02 -7.79598653e-01 2.93896049e-01
1.03278899e+00 7.73619860e-02 -7.18946159e-01 1.67259943e+00
5.99418759e-01 7.48210490e-01 4.78817709e-02 1.25245953e+00
3.70196015e-01 5.12561798e-01 -6.06060147e-01 -4.43204165e-01
9.65611815e-01 -1.06624234e+00 -4.30684477e-01 -4.61648315e-01
3.38561267e-01 -8.27153563e-01 8.75494063e-01 5.24799883e-01
-1.07418942e+00 -6.24444008e-01 -6.62943304e-01 -1.44881159e-01
1.10787556e-01 -1.69449165e-01 3.60669822e-01 4.05703694e-01
-7.51782358e-01 6.11573756e-01 -1.17739284e+00 -3.89688075e-01
6.23382777e-02 1.29916295e-01 -6.29279017e-01 -2.31717482e-01
-5.88609755e-01 7.54183590e-01 4.59267646e-01 3.03047746e-01
-1.02207541e+00 -6.47860587e-01 -9.65764046e-01 -1.86993062e-01
5.97322941e-01 -8.84370685e-01 1.05625403e+00 -9.84630287e-01
-1.82355046e+00 9.80965853e-01 -1.94301128e-01 -1.49697259e-01
9.88436222e-01 -5.43258965e-01 7.05965906e-02 5.40819466e-01
-5.98339513e-02 3.66573006e-01 9.18823719e-01 -1.40029263e+00
-4.36531484e-01 -2.35136777e-01 2.55700439e-01 4.64414805e-01
1.16919383e-01 -1.57143697e-01 -5.46061575e-01 -2.73632467e-01
2.34408811e-01 -1.12656033e+00 -2.52194941e-01 1.58381298e-01
-4.79957998e-01 4.74295199e-01 1.13534665e+00 -5.10560215e-01
8.10381949e-01 -1.88589227e+00 5.70874512e-01 -7.66531974e-02
1.98340148e-01 1.39736101e-01 5.47421016e-02 -9.23947990e-02
-6.14234172e-02 -4.10047323e-01 -4.03784722e-01 -4.38553005e-01
-5.75404406e-01 2.18835562e-01 -3.30283791e-01 7.88993835e-01
-2.40935966e-01 4.33380336e-01 -1.08027434e+00 -3.67739439e-01
8.99423957e-01 3.87176722e-01 -6.83048725e-01 4.71110880e-01
-6.86216578e-02 1.10207832e+00 -3.00266296e-01 2.51676708e-01
7.64919758e-01 -4.20755088e-01 2.62914270e-01 -2.39784837e-01
-1.90821677e-01 9.17025879e-02 -1.69323957e+00 2.08594751e+00
-5.14911771e-01 8.74051154e-01 4.94095199e-02 -1.11390090e+00
5.13002634e-01 1.13879420e-01 6.70178056e-01 -3.12588811e-01
1.60380706e-01 1.21060917e-02 -8.41158442e-03 -9.31171834e-01
5.40307641e-01 1.18554585e-01 2.81496346e-01 3.04414183e-01
-1.96195275e-01 -4.08422619e-01 2.57237494e-01 -6.23072498e-02
1.17062962e+00 3.44113141e-01 2.32207641e-01 -9.86328870e-02
7.69828498e-01 -2.15375051e-01 5.76955616e-01 4.71598953e-01
-2.52132505e-01 8.94187748e-01 3.94110978e-01 -7.15154946e-01
-1.21982300e+00 -1.07562947e+00 2.37161126e-02 4.01216477e-01
6.65911078e-01 -1.83681041e-01 -6.29777193e-01 -4.19573814e-01
-2.48631015e-01 2.02741235e-01 -5.44988930e-01 2.64656544e-01
-1.02472210e+00 -4.98650193e-01 9.51432735e-02 2.18253344e-01
7.05142677e-01 -8.52046669e-01 -1.29909253e+00 -2.68630392e-04
-6.80311680e-01 -1.95811427e+00 -5.18453419e-01 -2.86840886e-01
-8.48895133e-01 -1.41260159e+00 -4.54014748e-01 -4.61872101e-01
6.19422078e-01 7.14684367e-01 1.12083149e+00 -2.52957016e-01
-2.29709789e-01 4.01621640e-01 -7.07656816e-02 3.88616055e-01
-3.24576050e-01 -3.01157802e-01 1.54921487e-01 3.22270215e-01
-3.88430566e-01 -8.11861396e-01 -9.89342928e-01 5.93650222e-01
-9.68659043e-01 6.63530052e-01 -1.81346778e-02 5.28085828e-01
4.85191256e-01 1.43501595e-01 -4.03305054e-01 -7.06158459e-01
-3.07591975e-01 -6.01574220e-02 -9.25394416e-01 -1.57603279e-01
1.18989572e-01 -9.67778489e-02 7.26529002e-01 -6.61006272e-01
-1.20315194e+00 4.52640533e-01 1.52187318e-01 -9.02824402e-01
-2.65894026e-01 -6.85725734e-02 -3.77514884e-02 -5.17023504e-02
4.67388839e-01 1.22482046e-01 -1.64938852e-01 -1.47712559e-01
4.15852159e-01 5.80133274e-02 7.58200407e-01 -4.28940445e-01
8.85452628e-01 1.12960446e+00 3.51300925e-01 -1.13439310e+00
-9.34739828e-01 -6.23542786e-01 -1.05133235e+00 -5.04052877e-01
1.18348169e+00 -8.45479906e-01 -9.23054814e-01 7.20269501e-01
-1.32871664e+00 -7.53070831e-01 -1.91060141e-01 7.79194534e-01
-8.36755037e-01 6.65001452e-01 -5.46089709e-01 -5.74553788e-01
3.69788148e-02 -1.41637981e+00 1.18832743e+00 3.76899466e-02
-1.20201930e-01 -1.15956855e+00 2.87451804e-01 6.13881588e-01
-3.29759195e-02 6.65024519e-01 4.17446315e-01 2.27738664e-01
-9.94303167e-01 1.59170985e-01 -7.95021206e-02 3.44427496e-01
2.58805037e-01 2.07866818e-01 -1.09480512e+00 -6.29473329e-02
2.87317425e-01 -7.87572861e-02 5.58840811e-01 7.31663525e-01
1.07065737e+00 -2.01552570e-01 -7.29616433e-02 1.00675297e+00
1.51508236e+00 1.38082718e-02 5.29557228e-01 2.21080050e-01
1.22508419e+00 7.93087661e-01 3.88098389e-01 4.01605427e-01
1.72910720e-01 9.35907960e-01 6.21352494e-01 1.55847117e-01
-1.37132257e-01 -8.37522969e-02 3.56009692e-01 8.06160390e-01
-3.81681025e-01 -2.73897469e-01 -9.42809463e-01 3.37846488e-01
-1.90245247e+00 -1.16935718e+00 -1.30238846e-01 2.40776396e+00
5.10018706e-01 1.03183441e-01 5.69883222e-03 1.68775097e-01
6.56412780e-01 3.93489331e-01 -6.89383566e-01 9.38828513e-02
1.91808399e-02 -2.65509635e-01 4.50539976e-01 1.04383564e+00
-9.11259532e-01 9.75215614e-01 5.59453392e+00 2.60625154e-01
-1.46855533e+00 -1.61419600e-01 6.02415979e-01 -2.62223870e-01
1.71381626e-02 2.38240272e-01 -5.48870981e-01 4.23820525e-01
4.67888802e-01 6.38302788e-02 5.78085005e-01 6.31629944e-01
5.66566706e-01 -4.73340422e-01 -1.33253860e+00 1.26902127e+00
2.75140256e-01 -1.24057353e+00 -1.20394059e-01 3.18984538e-02
7.65311778e-01 9.23879668e-02 -2.22939357e-01 -4.96536940e-01
1.06779702e-01 -6.57969177e-01 7.77979076e-01 4.98639464e-01
6.84051871e-01 -2.40522310e-01 3.16829592e-01 6.30329430e-01
-1.22700024e+00 -5.08197956e-02 -1.72022641e-01 -2.95335680e-01
6.76340282e-01 7.20566988e-01 -4.84522015e-01 5.11271238e-01
8.15325439e-01 1.12856197e+00 -4.39933985e-01 8.11605453e-01
-8.13855305e-02 2.58912921e-01 -4.81013089e-01 6.09351516e-01
-1.63288862e-02 -6.65403545e-01 8.04858029e-01 9.49767590e-01
2.71950930e-01 2.09653556e-01 3.15593332e-01 8.67952645e-01
8.21323544e-02 -3.98187250e-01 -8.36743414e-01 4.86943752e-01
-2.13158373e-02 1.34999669e+00 -1.07471764e+00 -3.80913377e-01
-3.64410073e-01 1.21939516e+00 2.04973236e-01 4.57752943e-01
-7.07135499e-01 2.42351741e-01 5.98266542e-01 3.14992011e-01
1.72026128e-01 -6.48830235e-01 9.20891762e-02 -1.72161460e+00
1.43785849e-01 -4.21915531e-01 6.99889883e-02 -1.07911789e+00
-7.33169854e-01 4.34909135e-01 3.47339511e-01 -1.41898382e+00
-3.88984799e-01 -6.32247567e-01 -5.66165745e-01 5.70123494e-01
-1.41621685e+00 -6.59444511e-01 -8.45243156e-01 6.81258202e-01
8.87052298e-01 3.71580839e-01 2.82312959e-01 1.81962281e-01
-5.55887640e-01 -1.69081107e-01 -1.14326723e-01 2.23655865e-01
5.63822210e-01 -1.07394469e+00 2.36121058e-01 1.23066604e+00
7.08681196e-02 2.96117574e-01 7.57773042e-01 -3.04158390e-01
-1.51662934e+00 -8.66036296e-01 2.80147076e-01 -8.00717950e-01
6.40589058e-01 -4.91539836e-01 -7.80477285e-01 5.48310339e-01
-2.63418239e-02 5.18660069e-01 -4.52596648e-03 -5.57232499e-01
-1.10762581e-01 -1.33507684e-01 -7.99222350e-01 5.40323377e-01
1.37991047e+00 -4.31507438e-01 -2.70965248e-01 1.84141070e-01
6.38795435e-01 -9.61386323e-01 -4.99118686e-01 4.01687831e-01
7.15960562e-01 -1.36715031e+00 1.09978068e+00 -1.20020173e-01
7.44423091e-01 -6.05832696e-01 -1.41725972e-01 -1.03429186e+00
1.29753694e-01 -8.45479608e-01 -3.44111800e-01 7.01121449e-01
-2.35635579e-01 -2.26101488e-01 1.03654754e+00 5.52704334e-01
-3.08425222e-02 -3.93205374e-01 -9.26763177e-01 -6.09717369e-01
-3.31866950e-01 -5.19387305e-01 -1.05550542e-01 1.04773331e+00
-4.93590444e-01 3.74451190e-01 -4.62547004e-01 3.26204330e-01
7.27847159e-01 2.09537774e-01 1.17770970e+00 -9.55598652e-01
-5.66577911e-01 -3.29529107e-01 -6.55662358e-01 -1.54650152e+00
2.64228910e-01 -4.60051060e-01 6.25352710e-02 -1.05419338e+00
1.93994537e-01 -1.29738227e-01 3.67702931e-01 -1.65680781e-01
-1.20895416e-01 1.16134144e-01 1.99006885e-01 3.65746647e-01
-4.23408866e-01 3.07279050e-01 1.56811678e+00 1.79864362e-01
-4.65288788e-01 -5.29642925e-02 8.68087932e-02 1.08573043e+00
4.30961847e-01 -2.23868132e-01 -6.65222287e-01 -6.18273675e-01
2.95694232e-01 5.24516046e-01 4.88983035e-01 -9.38288867e-01
1.40553728e-01 -4.96607035e-01 2.74601430e-01 -4.70275909e-01
4.77265865e-01 -9.81686056e-01 2.75571167e-01 4.10173208e-01
-2.20675737e-01 2.75025487e-01 7.99167156e-02 6.85849428e-01
-3.40373465e-03 -6.63940236e-02 9.38512504e-01 -3.43243599e-01
-6.86608613e-01 3.93348783e-01 -8.73231366e-02 3.15728784e-01
9.63675678e-01 -5.16527116e-01 -3.69217128e-01 -3.69139940e-01
-5.73432803e-01 -1.08903214e-01 7.65225947e-01 3.48772883e-01
7.84699798e-01 -8.44137967e-01 -4.36035037e-01 3.18532318e-01
-8.08105916e-02 7.26273298e-01 4.50964272e-01 9.05339897e-01
-1.09008873e+00 8.16375483e-03 -1.89271674e-01 -1.16537094e+00
-1.26796007e+00 4.94128823e-01 6.31820977e-01 -2.61622548e-01
-1.04412723e+00 4.06050861e-01 7.22706199e-01 -2.37273768e-01
3.86782140e-02 -5.39780080e-01 -8.74548629e-02 -3.92645001e-01
4.54669148e-01 5.66619754e-01 -1.89547256e-01 -7.90271819e-01
-1.53862104e-01 1.20060802e+00 1.95084751e-01 -3.22065324e-01
1.07074559e+00 -3.63862425e-01 -1.33775063e-02 5.36116719e-01
1.25451934e+00 6.20726980e-02 -1.81394053e+00 -1.16333976e-01
-4.59212780e-01 -9.04428899e-01 -1.22302495e-01 -2.96975579e-02
-1.21175563e+00 8.64953339e-01 5.09674430e-01 -4.59874533e-02
1.05820715e+00 -1.38816327e-01 6.15240753e-01 4.33497161e-01
1.68503642e-01 -8.50037158e-01 3.92803520e-01 4.77495551e-01
5.22987425e-01 -1.32498670e+00 6.54562637e-02 -7.46555865e-01
-4.95320708e-01 1.31485856e+00 7.14136362e-01 -2.45826885e-01
5.61123490e-01 2.10590720e-01 -9.07144882e-03 3.51773761e-02
-5.55346072e-01 8.33064392e-02 6.66295215e-02 3.94601971e-01
3.65056470e-02 -3.41777325e-01 3.59980196e-01 -4.70609933e-01
-2.61604130e-01 -2.25938931e-01 1.09642017e+00 7.15385139e-01
-1.40778750e-01 -7.30225682e-01 -3.00804317e-01 -3.18225436e-02
-3.38731796e-01 2.68658549e-01 -6.24779910e-02 7.30475187e-01
2.81288084e-02 7.37588227e-01 2.70541906e-01 -1.24654792e-01
3.38459730e-01 -3.93200964e-01 8.65470171e-01 -5.01701057e-01
-3.31038833e-02 1.38903677e-01 -7.73168802e-02 -8.52749527e-01
-1.04883814e+00 -7.22921968e-01 -1.05489898e+00 -2.58572251e-01
-1.74569339e-01 -3.56692582e-01 4.97930944e-01 9.37657714e-01
-1.36433477e-02 3.36051106e-01 7.08257973e-01 -1.30548275e+00
4.82095964e-02 -6.35320485e-01 -4.41539794e-01 8.78298700e-01
7.67937660e-01 -6.94449663e-01 -6.93828821e-01 6.70559108e-01] | [8.590841293334961, -2.0555076599121094] |
1c0601a9-9b25-4043-bded-dcf0b6df0d96 | multigbs-a-multi-layer-graph-approach-to | 2008.11908 | null | https://arxiv.org/abs/2008.11908v2 | https://arxiv.org/pdf/2008.11908v2.pdf | MultiGBS: A multi-layer graph approach to biomedical summarization | Automatic text summarization methods generate a shorter version of the input text to assist the reader in gaining a quick yet informative gist. Existing text summarization methods generally focus on a single aspect of text when selecting sentences, causing the potential loss of essential information. In this study, we propose a domain-specific method that models a document as a multi-layer graph to enable multiple features of the text to be processed at the same time. The features we used in this paper are word similarity, semantic similarity, and co-reference similarity, which are modelled as three different layers. The unsupervised method selects sentences from the multi-layer graph based on the MultiRank algorithm and the number of concepts. The proposed MultiGBS algorithm employs UMLS and extracts the concepts and relationships using different tools such as SemRep, MetaMap, and OGER. Extensive evaluation by ROUGE and BERTScore shows increased F-measure values. | ['Nasser Ghadiri', 'Ensieh Davoodijam', 'Fabio Rinaldi', 'Maryam Lotfi Shahreza'] | 2020-08-27 | null | null | null | null | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 4.66057032e-01 6.73031926e-01 -7.77313933e-02 -1.42454384e-02
-5.06142735e-01 -4.92384285e-01 6.41261637e-01 1.21210182e+00
-2.11663052e-01 6.93188131e-01 9.94631231e-01 9.47751030e-02
-6.05728090e-01 -8.16788137e-01 4.29817215e-02 -1.68555781e-01
1.02638841e-01 5.79524934e-01 3.66464406e-01 -5.63081563e-01
1.01037204e+00 1.49308564e-02 -1.73247039e+00 5.82579136e-01
1.44423997e+00 2.74760276e-01 3.57824802e-01 8.43168378e-01
-8.82399678e-01 6.10124767e-01 -8.12612116e-01 -1.78950250e-01
-1.65920407e-01 -7.52981007e-01 -1.09040880e+00 -1.12942979e-02
2.52937466e-01 2.81107336e-01 2.55888909e-01 9.35563445e-01
3.85775954e-01 1.74161345e-01 9.02499378e-01 -8.73914540e-01
8.84177759e-02 9.49035645e-01 -4.12667483e-01 7.43829608e-02
1.01446176e+00 -5.63097119e-01 1.11027122e+00 -5.46175301e-01
8.48915875e-01 1.39422941e+00 4.09055263e-01 3.04082669e-02
-8.40516686e-01 -7.04641938e-02 -2.65378170e-02 1.27482742e-01
-9.96740580e-01 -1.64880902e-01 8.29692781e-01 -3.53816420e-01
1.08460462e+00 7.71603644e-01 7.28918433e-01 2.87463039e-01
3.44420999e-01 6.85513079e-01 7.79040694e-01 -7.52351999e-01
2.79449165e-01 7.01829791e-02 8.07579219e-01 4.77108181e-01
7.85125792e-01 -1.03891218e+00 -7.45167911e-01 -3.29805911e-01
-1.50707290e-01 -1.33091494e-01 -1.46613359e-01 -1.03198946e-01
-1.02575004e+00 6.64779663e-01 -1.31895971e-02 8.00111413e-01
-5.05415857e-01 -5.00941038e-01 6.39818788e-01 2.89747894e-01
5.81819057e-01 8.32699835e-01 -9.36759263e-02 -4.28051390e-02
-1.34388340e+00 2.39081755e-01 1.08311045e+00 9.73628283e-01
7.24386692e-01 -4.92917746e-01 -3.32499772e-01 5.21370053e-01
3.22712570e-01 5.77477133e-03 8.15598786e-01 -3.21864694e-01
6.70481682e-01 1.42054927e+00 -1.58842087e-01 -1.35799432e+00
-4.02997077e-01 -4.18269932e-01 -6.43483102e-01 -1.71914652e-01
-3.35618496e-01 -6.37854189e-02 -6.75883353e-01 1.01435709e+00
2.84266979e-01 -4.29985344e-01 3.46501529e-01 2.62948781e-01
1.34202099e+00 6.92392409e-01 1.72153860e-02 -5.10028422e-01
1.12270963e+00 -8.17253292e-01 -9.39075947e-01 5.41400388e-02
7.70244777e-01 -1.00766253e+00 7.08392739e-01 3.35347801e-01
-1.02886438e+00 -4.52891797e-01 -1.32922864e+00 -4.93350513e-02
-7.36330450e-01 8.61792713e-02 2.37218827e-01 4.17305291e-01
-1.01005864e+00 7.77810037e-01 -1.76314771e-01 -9.82832313e-01
-1.51007146e-01 8.53833929e-02 -1.91251606e-01 7.91479349e-02
-1.08303332e+00 8.37270677e-01 1.00687516e+00 -4.08073515e-01
1.44655928e-01 -3.77560586e-01 -6.22457445e-01 3.22089642e-01
4.11245883e-01 -1.01462924e+00 6.26642406e-01 -7.63424873e-01
-1.14349222e+00 4.68919039e-01 -1.83953732e-01 -3.60745639e-01
4.49224442e-01 -2.65759259e-01 -2.10882038e-01 6.37698829e-01
3.13693494e-01 2.37687841e-01 5.44988453e-01 -1.19597614e+00
-8.01094830e-01 -3.84243727e-01 -1.51864246e-01 7.30738938e-01
-4.13774818e-01 1.21589974e-01 -2.84319341e-01 -5.09920299e-01
2.37122238e-01 -2.18998864e-01 -2.29891576e-02 -1.06628931e+00
-6.75798297e-01 -3.91446918e-01 7.39830375e-01 -9.51374948e-01
1.85007656e+00 -1.56654179e+00 3.33012640e-01 4.39976722e-01
5.38483560e-01 1.16924822e-01 2.58434657e-02 1.25113142e+00
1.50590569e-01 5.48048615e-01 -2.61910915e-01 1.58659499e-02
-1.49472132e-01 -1.77520692e-01 1.10422224e-01 -1.77784279e-01
-2.75795490e-01 3.56108218e-01 -1.10467577e+00 -1.07128298e+00
1.05858058e-01 -1.41879516e-02 -2.73981363e-01 1.23232352e-02
-1.56716183e-01 -1.16015792e-01 -5.99293232e-01 2.89686590e-01
2.99497068e-01 2.36346707e-01 3.15944135e-01 -2.72226691e-01
-2.57865787e-01 3.52286845e-01 -1.24530625e+00 1.50066316e+00
-2.52622068e-01 5.23061574e-01 -4.43706691e-01 -8.17470312e-01
9.65192556e-01 1.62152365e-01 5.64884186e-01 -3.47319841e-01
1.33431479e-01 7.93536566e-03 -1.76462039e-01 -6.67129695e-01
1.17676222e+00 4.07746792e-01 -1.92829356e-01 5.27001321e-01
9.83756706e-02 -2.87115932e-01 1.06324267e+00 1.07282448e+00
1.03472328e+00 -1.34308953e-02 8.92387748e-01 -4.30399001e-01
6.33548737e-01 4.40158010e-01 4.38074628e-03 8.13123822e-01
6.80874050e-01 4.07832295e-01 7.67119110e-01 8.97348151e-02
-1.01452219e+00 -5.91722548e-01 3.65512937e-01 7.46531725e-01
1.08522072e-01 -1.18800545e+00 -9.00993705e-01 -6.94658577e-01
-1.72093824e-01 1.18597162e+00 -4.68471378e-01 -2.26934716e-01
-2.99493581e-01 -4.41176623e-01 2.40312278e-01 -1.82689145e-01
3.72102797e-01 -7.76479185e-01 -6.60397410e-01 3.09859365e-01
-3.95313561e-01 -5.33739090e-01 -2.10250735e-01 -2.97130078e-01
-9.89216328e-01 -1.06354547e+00 -2.55177140e-01 -5.60061276e-01
7.84812987e-01 3.03123832e-01 9.96028364e-01 1.87492371e-01
-7.94101506e-02 3.13881099e-01 -9.63466585e-01 -5.08467495e-01
-7.86346018e-01 6.54690742e-01 -3.19660962e-01 -2.40967989e-01
1.73662320e-01 -4.30409461e-01 -2.72954524e-01 -3.62896264e-01
-1.16289711e+00 3.38791758e-01 5.92725039e-01 4.37451303e-01
2.39239499e-01 2.43320689e-01 8.41521084e-01 -1.06574893e+00
1.60856998e+00 -5.66646218e-01 2.02929631e-01 7.34798431e-01
-9.49042261e-01 3.08815986e-01 5.72149396e-01 1.25852838e-01
-1.10175765e+00 -3.56568456e-01 2.59059876e-01 6.17667198e-01
-2.85310950e-02 1.15988421e+00 -1.30238131e-01 3.67773563e-01
5.27740896e-01 3.30942601e-01 -1.29019842e-01 -5.34643531e-01
4.82306540e-01 9.06666517e-01 1.11666657e-01 -1.18684962e-01
4.11863178e-01 3.10881599e-03 4.27643768e-02 -1.24178958e+00
-6.60323381e-01 -1.01301014e+00 -8.85779381e-01 -6.40476525e-01
5.38553238e-01 -4.61819857e-01 -9.77022201e-02 -2.84622163e-02
-1.04962373e+00 4.86420304e-01 -3.88885140e-01 2.36706063e-01
-1.53507769e-01 9.31033254e-01 1.48873553e-01 -6.10008717e-01
-9.71854031e-01 -4.30159956e-01 7.60973454e-01 4.25161988e-01
-7.83170700e-01 -8.14949393e-01 1.79213196e-01 2.38303930e-01
7.32983053e-02 4.28050280e-01 1.23235273e+00 -1.23862696e+00
-8.63490552e-02 -4.15783942e-01 8.42822343e-02 1.54401913e-01
3.71980995e-01 2.49293685e-01 -2.63630033e-01 -1.15564995e-01
-1.24144323e-01 1.45535007e-01 9.78786469e-01 2.06544206e-01
5.45419455e-01 -6.39776886e-01 -4.94058609e-01 -9.16539356e-02
1.47393727e+00 8.24920163e-02 5.79325855e-01 4.43136930e-01
6.49771869e-01 8.08454692e-01 7.44871736e-01 4.86155123e-01
3.67647707e-01 6.22284003e-02 1.93034977e-01 2.49149859e-01
-1.60413116e-01 -2.15542600e-01 1.10739216e-01 1.46478176e+00
1.59284785e-01 -5.16448140e-01 -8.89049888e-01 6.07227325e-01
-2.01415324e+00 -1.03557706e+00 -4.49130982e-01 1.97542787e+00
7.38332987e-01 2.98808545e-01 2.27179363e-01 3.95900995e-01
7.62648225e-01 1.42850190e-01 7.57450461e-02 -7.11084127e-01
-1.28195897e-01 -3.25883413e-03 3.25222731e-01 6.59209847e-01
-5.09221971e-01 8.96715164e-01 5.45836258e+00 7.60936975e-01
-5.57329953e-01 -3.89987439e-01 9.47007909e-02 -6.94470704e-02
-6.92180574e-01 3.68795693e-01 -6.17847204e-01 3.19840431e-01
7.45830774e-01 -1.01631796e+00 -1.52044281e-01 4.82910782e-01
4.23159987e-01 -7.28581250e-01 -6.69126987e-01 5.60074568e-01
6.83383822e-01 -1.29511988e+00 6.71214938e-01 -2.73706049e-01
6.76190615e-01 -3.63869071e-01 -6.33338630e-01 -1.46734327e-01
3.01472515e-01 -5.31438231e-01 5.32237589e-01 9.69749570e-01
2.76797622e-01 -9.13652718e-01 9.67043102e-01 3.82144839e-01
-1.02333260e+00 2.62415856e-01 -1.95344940e-01 2.99368259e-02
-1.09980274e-02 5.40793836e-01 -1.10108411e+00 1.42049527e+00
2.23422348e-01 7.08857298e-01 -1.11523449e+00 1.17736268e+00
-9.07328874e-02 4.63902652e-01 -2.03325927e-01 -6.70903563e-01
2.12331206e-01 -4.27208066e-01 1.02991235e+00 1.50117910e+00
4.23415840e-01 -6.11988418e-02 2.03495294e-01 2.97403514e-01
1.69225797e-01 9.15855229e-01 -4.70440477e-01 -1.27362520e-01
4.95914519e-01 1.25398374e+00 -1.02458429e+00 -7.22296953e-01
2.70936731e-02 7.11932063e-01 -9.43906605e-02 3.60149220e-02
-2.11472646e-03 -1.08267689e+00 -1.12971231e-01 2.25571647e-01
-1.29616156e-01 -1.50339708e-01 -5.42961478e-01 -9.90342736e-01
-7.17337877e-02 -8.43362153e-01 6.00826621e-01 -7.61860132e-01
-7.57439673e-01 5.30121863e-01 3.86865497e-01 -1.04700255e+00
-4.43064839e-01 6.00942858e-02 -8.15138936e-01 7.42181003e-01
-8.75434577e-01 -8.59550834e-01 -4.17787790e-01 1.40671834e-01
7.70266652e-01 -1.76294148e-01 6.41488791e-01 -2.81625688e-01
-4.40334588e-01 -3.67962569e-02 3.30774933e-01 -1.92369655e-01
5.83819449e-01 -1.57417703e+00 2.06874102e-01 1.10360372e+00
-9.08527896e-02 6.17072403e-01 1.07516193e+00 -1.26338041e+00
-7.98397005e-01 -6.85132086e-01 1.42439711e+00 3.61118168e-02
4.78160709e-01 6.58674305e-03 -7.60157406e-01 1.14633881e-01
7.51893580e-01 -1.39712167e+00 7.37033844e-01 -9.98225063e-02
2.05618948e-01 -1.69962943e-02 -1.08151340e+00 6.40349448e-01
6.60510004e-01 -7.06903487e-02 -1.27364361e+00 2.59187460e-01
6.65754974e-01 1.02518864e-01 -7.31458664e-01 1.27568021e-02
3.51691455e-01 -9.15932715e-01 4.39069748e-01 -5.54034770e-01
4.28234339e-01 -3.08817685e-01 2.49939874e-01 -1.55004072e+00
-1.46734595e-01 -6.92370772e-01 4.51599471e-02 1.53457725e+00
5.13307095e-01 -3.94697905e-01 4.42774743e-01 1.20025739e-01
-1.12425692e-01 -3.91984522e-01 -4.32615548e-01 -3.68184030e-01
-3.01202238e-01 8.40687081e-02 5.62940717e-01 7.88499236e-01
7.25277781e-01 7.97460496e-01 4.14496996e-02 -2.38817021e-01
6.72790289e-01 1.41606063e-01 7.41627932e-01 -1.67284656e+00
4.22600150e-01 -5.03648341e-01 -3.54808420e-01 -3.23967427e-01
-1.96329042e-01 -1.05868697e+00 -2.17102170e-01 -2.67728806e+00
3.85718256e-01 2.66660094e-01 2.76223999e-02 6.01091757e-02
-2.29483962e-01 -6.19848430e-01 1.53462380e-01 2.96079367e-01
-7.79518843e-01 3.03276479e-01 8.44693720e-01 -1.16631582e-01
-5.32447636e-01 -1.08162306e-01 -8.54082286e-01 7.01922297e-01
1.07665515e+00 -6.47503614e-01 -7.02771187e-01 5.52049689e-02
4.85202760e-01 -1.19199105e-01 -2.87436008e-01 -1.14413273e+00
4.89135176e-01 -1.10713936e-01 1.49111226e-01 -9.52545822e-01
-2.88557857e-01 -4.10085201e-01 7.02307075e-02 5.03657818e-01
-5.78418314e-01 4.14669812e-01 -9.56273973e-02 3.89061272e-01
-4.03149903e-01 -7.74241865e-01 2.33234942e-01 -3.25285971e-01
-5.31922698e-01 -3.39206189e-01 -6.51314318e-01 1.50208518e-01
6.37066305e-01 -5.70514143e-01 -3.34679335e-01 -4.97679174e-01
-5.21768570e-01 3.53533059e-01 3.83813918e-01 3.87089878e-01
6.17239475e-01 -8.83178234e-01 -9.54757273e-01 -3.87137651e-01
1.63732216e-01 -9.39083695e-02 1.22523397e-01 3.93171638e-01
-7.95054913e-01 4.10244763e-01 -4.04014587e-01 -9.87093821e-02
-1.75198531e+00 1.43960565e-01 -1.99148104e-01 -6.45773649e-01
-5.25610030e-01 1.59239009e-01 -5.14383733e-01 2.81083677e-02
-1.52605951e-01 -2.91557789e-01 -1.11300802e+00 8.44077945e-01
4.73315686e-01 7.52751768e-01 2.37884641e-01 -6.74888134e-01
-2.41225287e-01 4.72586811e-01 -2.26108238e-01 -2.42728040e-01
1.42742395e+00 -4.51606780e-01 -6.07196987e-01 3.60585481e-01
8.58869255e-01 3.24586064e-01 -3.31444949e-01 -3.21941301e-02
6.65255845e-01 -5.64935543e-02 6.02317825e-02 -7.83359826e-01
-1.48389503e-01 2.49088287e-01 3.58444043e-02 8.05193543e-01
1.06298602e+00 -3.13013270e-02 3.91719967e-01 4.57179636e-01
-2.19975442e-01 -1.62135541e+00 7.52770379e-02 4.41695660e-01
1.00666404e+00 -6.04852438e-01 6.03418648e-01 -5.50226748e-01
-7.69239128e-01 1.36254811e+00 2.85544842e-01 1.15396038e-01
2.91302443e-01 -1.44439444e-01 -1.93229392e-01 -3.58460128e-01
-6.63862050e-01 -4.03595626e-01 5.72412908e-01 4.42683280e-01
5.93619764e-01 -1.01883493e-01 -1.29006457e+00 4.34149623e-01
-5.04362881e-01 -2.37698883e-01 1.01704609e+00 1.01010764e+00
-9.87950444e-01 -1.02630913e+00 -2.27945849e-01 8.66070390e-01
-3.43689471e-01 -5.91793619e-02 -1.16904378e+00 6.59023404e-01
-1.31905526e-01 1.25672269e+00 -2.08369613e-01 -3.90700340e-01
4.89840448e-01 2.12968111e-01 3.13504249e-01 -9.21706438e-01
-9.22572851e-01 2.46106740e-03 5.67068338e-01 9.74015705e-03
-5.60333312e-01 -8.51559997e-01 -1.28308010e+00 -1.00466236e-01
-3.03948551e-01 6.85824335e-01 9.11787391e-01 9.55521643e-01
4.66841280e-01 7.26753831e-01 3.90658230e-01 -4.35697407e-01
-2.67173141e-01 -1.37067235e+00 -3.06932330e-01 4.41867977e-01
-4.46169227e-02 -8.85290429e-02 -3.02633256e-01 -1.62490066e-02] | [12.42379093170166, 9.558388710021973] |
25eb6aab-db73-4cd3-870c-2fdd63fc9c01 | seq-nms-for-video-object-detection | 1602.08465 | null | http://arxiv.org/abs/1602.08465v3 | http://arxiv.org/pdf/1602.08465v3.pdf | Seq-NMS for Video Object Detection | Video object detection is challenging because objects that are easily
detected in one frame may be difficult to detect in another frame within the
same clip. Recently, there have been major advances for doing object detection
in a single image. These methods typically contain three phases: (i) object
proposal generation (ii) object classification and (iii) post-processing. We
propose a modification of the post-processing phase that uses high-scoring
object detections from nearby frames to boost scores of weaker detections
within the same clip. We show that our method obtains superior results to
state-of-the-art single image object detection techniques. Our method placed
3rd in the video object detection (VID) task of the ImageNet Large Scale Visual
Recognition Challenge 2015 (ILSVRC2015). | ['Honghui Shi', 'Prajit Ramachandran', 'Wei Han', 'Tom Le Paine', 'Thomas S. Huang', 'Pooya Khorrami', 'Shuicheng Yan', 'Mohammad Babaeizadeh', 'Jianan Li'] | 2016-02-26 | null | null | null | null | ['object-proposal-generation'] | ['computer-vision'] | [ 4.40239489e-01 -5.33834636e-01 6.65587513e-03 -2.91577607e-01
-8.27047110e-01 -5.50338447e-01 5.32305658e-01 2.44250402e-01
-8.06122482e-01 3.75200272e-01 -9.18722451e-02 9.95498672e-02
4.46199745e-01 -3.95677567e-01 -1.05099893e+00 -4.65969473e-01
-2.20841408e-01 4.37990353e-02 1.39963984e+00 2.14236453e-01
3.02776784e-01 5.53429961e-01 -1.68051445e+00 6.88866436e-01
6.53522089e-02 1.23517060e+00 4.45858896e-01 1.14053404e+00
1.60375074e-01 9.24960613e-01 -6.12266898e-01 -2.75786966e-02
3.50013942e-01 -3.63025367e-01 -6.51041448e-01 4.06794220e-01
9.32565749e-01 -6.89239919e-01 -2.42786512e-01 1.25181901e+00
3.07453334e-01 1.22320272e-01 3.70229572e-01 -1.43863702e+00
-4.43566628e-02 2.31755331e-01 -8.90218854e-01 1.00326896e+00
3.01241547e-01 2.09527120e-01 7.31167912e-01 -1.44815588e+00
4.68637705e-01 1.28080773e+00 5.32462478e-01 3.18336189e-01
-1.07031548e+00 -7.34084189e-01 3.00089985e-01 6.13342226e-01
-1.60419059e+00 -5.09456217e-01 3.30140203e-01 -6.13714337e-01
8.85900080e-01 1.15792766e-01 6.32254899e-01 5.29404819e-01
-1.42647862e-01 1.09503686e+00 6.36007786e-01 -3.03307980e-01
2.36508369e-01 -5.51425852e-02 2.49152899e-01 6.02114975e-01
4.11412567e-01 1.84375003e-01 -5.04776120e-01 -1.27359271e-01
8.41323078e-01 6.37866557e-02 -1.77045703e-01 -3.71935368e-01
-1.30997562e+00 7.33490646e-01 5.87381065e-01 1.69104800e-01
-5.79690039e-01 3.23391467e-01 4.29910988e-01 8.04784000e-02
1.10979117e-01 2.00825483e-02 -3.38843703e-01 -4.00021207e-03
-1.12905598e+00 4.10264760e-01 4.28695202e-01 1.01591730e+00
5.19321144e-01 -2.32040331e-01 -4.70326364e-01 6.82346284e-01
3.40963989e-01 1.47268549e-01 9.74556059e-02 -7.92446792e-01
4.87782210e-01 3.27667594e-01 3.20576996e-01 -8.76922429e-01
-2.62229830e-01 -2.16920614e-01 -3.52724016e-01 5.23608983e-01
4.72553849e-01 6.75316229e-02 -1.03861046e+00 1.35592449e+00
5.85527897e-01 4.57950592e-01 -3.02132070e-01 1.16836357e+00
1.04602325e+00 8.03668737e-01 2.44295314e-01 -1.63522825e-01
1.56689823e+00 -1.21897173e+00 -3.98398429e-01 -2.65972614e-01
2.30921715e-01 -1.02163124e+00 4.16175365e-01 2.56829381e-01
-1.18113136e+00 -9.83108997e-01 -1.06122184e+00 1.40829429e-01
-2.23863974e-01 3.14302146e-01 3.98779660e-01 6.06908083e-01
-9.50644493e-01 1.76035285e-01 -8.01968038e-01 -4.78384465e-01
7.39000261e-01 2.90075332e-01 -3.50982994e-01 -1.13674618e-01
-6.42564893e-01 7.09545374e-01 6.36416852e-01 -1.97575521e-02
-1.34350121e+00 -5.51238060e-01 -6.61455572e-01 -1.33407399e-01
7.59024024e-01 -4.42667633e-01 1.25572896e+00 -1.23584235e+00
-7.90769935e-01 1.01171470e+00 -3.15657496e-01 -6.06675386e-01
5.37198424e-01 -3.47873241e-01 -3.58048916e-01 3.72630805e-01
2.35657960e-01 1.05880988e+00 1.17666078e+00 -1.07746768e+00
-1.40183902e+00 -1.35756791e-01 6.66728169e-02 5.14918119e-02
1.12028614e-01 8.54544222e-01 -1.06514597e+00 -7.69984305e-01
5.68969455e-03 -7.54445970e-01 -5.36977649e-02 4.39931422e-01
-1.52142435e-01 -6.26298726e-01 1.04940891e+00 -4.11244899e-01
1.05806696e+00 -1.99250972e+00 -1.39916867e-01 -3.80906947e-02
1.35843560e-01 6.52345538e-01 -3.73350352e-01 -2.34265402e-01
-1.41347036e-01 -3.40470485e-02 2.31181711e-01 -1.56060904e-01
-4.25528020e-01 -2.77072251e-01 -1.94673643e-01 5.09768069e-01
4.61936474e-01 8.32114577e-01 -9.04940665e-01 -6.83635473e-01
1.62303120e-01 1.43142655e-01 -6.02960527e-01 1.96919367e-01
-1.21587496e-02 -6.08574860e-02 -1.43400177e-01 1.03784370e+00
5.97374558e-01 -2.16772273e-01 -2.20344171e-01 -6.18028879e-01
-2.32805908e-01 1.72736645e-02 -1.63293028e+00 1.33114922e+00
4.41853255e-01 8.75136077e-01 1.35557070e-01 -8.91916156e-01
3.62691432e-01 2.55411595e-01 4.00127083e-01 -3.95095825e-01
-8.92585702e-03 -4.86014262e-02 2.81343818e-01 -6.24164701e-01
5.56457937e-01 1.96228608e-01 4.02250439e-01 1.25234798e-01
1.33111998e-01 1.76693261e-01 7.58140683e-01 2.83231348e-01
1.31351876e+00 9.89421234e-02 4.85302359e-01 -6.96571218e-03
5.11747479e-01 -1.06681600e-01 6.82373226e-01 1.30148804e+00
-7.69324601e-01 9.03611422e-01 2.15606794e-01 -5.57243347e-01
-1.06731868e+00 -1.18109322e+00 -3.37390490e-02 1.41659915e+00
3.66299301e-01 -5.07627070e-01 -5.89584887e-01 -8.59887958e-01
1.62546352e-01 -1.18038595e-01 -7.11197555e-01 1.82854906e-01
-7.19773471e-01 -6.32716894e-01 5.18558502e-01 8.75209928e-01
3.69118273e-01 -1.20370018e+00 -8.83734286e-01 2.42612690e-01
-1.38891593e-01 -1.33398676e+00 -7.24834561e-01 -8.98076445e-02
-7.82736182e-01 -1.21701336e+00 -7.59832740e-01 -8.46462071e-01
7.00016439e-01 7.85991669e-01 1.05104256e+00 3.28942925e-01
-7.76980400e-01 2.60237604e-01 -4.23687458e-01 -6.34591162e-01
-1.64259240e-01 -3.39395612e-01 2.07295343e-01 2.10084811e-01
4.24154580e-01 2.30043575e-01 -6.54551923e-01 6.74028039e-01
-1.00199831e+00 -7.88228661e-02 6.98494911e-01 4.41657871e-01
6.36870623e-01 9.53096431e-03 4.58350390e-01 -3.11976016e-01
-4.45044646e-03 -3.05017531e-01 -8.76441836e-01 2.94293135e-01
-1.19214632e-01 -5.41159391e-01 -3.57996002e-02 -6.13514900e-01
-6.90745473e-01 4.11342323e-01 8.82494375e-02 -5.63444018e-01
-2.39974469e-01 3.41476426e-02 1.32368207e-01 -2.86219120e-01
5.74150085e-01 1.06527820e-01 -2.49130204e-01 -3.75381947e-01
-1.25186276e-02 4.54755098e-01 6.12921119e-01 -5.44547439e-02
9.02404010e-01 5.60162365e-01 -2.06550062e-01 -8.53287995e-01
-9.73964691e-01 -9.77781057e-01 -7.16977060e-01 -5.81198871e-01
1.02786660e+00 -1.21104026e+00 -5.55915713e-01 4.98750627e-01
-1.30356407e+00 -1.46325409e-01 -7.17169791e-02 6.11355126e-01
-1.60261050e-01 3.61208946e-01 -4.62915361e-01 -8.48482847e-01
-2.17404336e-01 -1.07207787e+00 1.13913882e+00 5.35409868e-01
-4.71488386e-02 -1.87817067e-01 -2.27364421e-01 3.40055525e-01
3.91181648e-01 -7.93704838e-02 1.98178783e-01 -5.02501965e-01
-9.51197326e-01 -4.60432887e-01 -6.50753498e-01 4.67891604e-01
-2.38134209e-02 1.95398018e-01 -8.12177420e-01 -5.12845159e-01
-4.18507367e-01 -3.74869019e-01 1.21875298e+00 5.21429002e-01
1.03644967e+00 3.09305005e-02 -4.66143668e-01 4.06382114e-01
1.35721767e+00 3.67240280e-01 7.90536940e-01 4.12010819e-01
5.40390015e-01 2.29221448e-01 8.80082667e-01 2.10068196e-01
8.23898092e-02 1.04366720e+00 3.12099785e-01 -2.31682017e-01
-4.43282187e-01 1.78392172e-01 5.36483586e-01 -6.59458563e-02
-3.25281531e-01 -4.53372858e-02 -7.53357768e-01 7.09407449e-01
-2.00271916e+00 -1.28182244e+00 -2.28930309e-01 2.02989435e+00
5.61805308e-01 4.54877466e-01 6.44940197e-01 -7.04864636e-02
1.02467215e+00 -1.39068859e-02 -4.43893373e-01 3.60710979e-01
-2.57179528e-01 -1.24171395e-02 4.01542962e-01 1.00643918e-01
-1.63453853e+00 9.25355375e-01 6.82023525e+00 6.26789093e-01
-7.22091675e-01 2.42277250e-01 5.28586149e-01 -3.74885499e-01
8.03917050e-01 -5.96433543e-02 -1.25252724e+00 4.89167273e-01
3.93069327e-01 1.55187115e-01 -9.30313915e-02 1.04873848e+00
2.34937772e-01 -6.56127870e-01 -1.26990771e+00 1.11240315e+00
4.14027095e-01 -1.53892648e+00 2.65532359e-02 -2.66376257e-01
7.32476175e-01 2.89916217e-01 -3.44764888e-01 2.69080192e-01
-1.72190860e-01 -7.19446242e-01 1.05306184e+00 3.23272824e-01
3.99879664e-01 -4.69778448e-01 6.89445078e-01 7.33622387e-02
-1.54916239e+00 -2.46197596e-01 -4.60124969e-01 1.59249201e-01
2.91995525e-01 3.49000543e-01 -7.77261257e-01 -1.08424582e-01
1.17401052e+00 5.57503939e-01 -8.19066882e-01 1.90755987e+00
-1.27498239e-01 7.29344547e-01 -3.15662593e-01 -3.07971705e-02
1.09279685e-01 2.15200663e-01 8.94147158e-01 1.46570647e+00
-4.76545468e-03 1.71445340e-01 5.26551127e-01 7.20954299e-01
-1.90564737e-01 -1.90352470e-01 -1.64263576e-01 2.18540236e-01
2.75736868e-01 1.46555305e+00 -1.13121378e+00 -7.52010226e-01
-6.34514511e-01 9.71737862e-01 3.59969884e-02 2.04060137e-01
-9.75986063e-01 -4.05676126e-01 5.90904415e-01 1.27593338e-01
9.49528396e-01 -1.15046881e-01 2.84942389e-01 -1.04410470e+00
2.14820713e-01 -7.66339004e-01 5.84450245e-01 -8.20626497e-01
-1.00622582e+00 3.92160505e-01 -7.72849098e-02 -1.32887220e+00
5.08624241e-02 -6.07124805e-01 -6.87967658e-01 3.98747385e-01
-1.36501861e+00 -8.36577773e-01 -5.12056887e-01 5.82580030e-01
9.33341205e-01 -7.00417161e-02 2.58955747e-01 6.22867763e-01
-5.68752766e-01 4.27234948e-01 -1.99948609e-01 5.97092927e-01
6.82946026e-01 -1.03329957e+00 3.52810204e-01 1.34953010e+00
2.78473347e-01 2.79123008e-01 5.50440848e-01 -6.73477173e-01
-1.22187150e+00 -1.27326775e+00 5.19294679e-01 -5.09606659e-01
5.12858093e-01 -4.01235521e-01 -8.98779452e-01 5.22460938e-01
-1.77413136e-01 5.91428161e-01 2.33765379e-01 -3.21106434e-01
-3.07166755e-01 -1.11705191e-01 -8.66672218e-01 3.02108526e-01
1.05920386e+00 -2.93224305e-01 -5.69177449e-01 3.68567377e-01
3.80159467e-01 -3.20831507e-01 -3.94529194e-01 5.34222901e-01
6.67268395e-01 -7.49738038e-01 1.31288552e+00 -7.62931645e-01
1.01438709e-01 -1.00859618e+00 -1.68781728e-01 -5.57707846e-01
-4.70919877e-01 -5.50098956e-01 -3.49471033e-01 1.05272889e+00
2.01293916e-01 1.50051370e-01 7.15424359e-01 3.71813387e-01
2.08249465e-01 -3.08893591e-01 -8.61104310e-01 -1.01127112e+00
-6.57123566e-01 -4.54524100e-01 -8.70682746e-02 3.89367282e-01
-3.76047850e-01 2.04800725e-01 -2.28768930e-01 3.01817358e-01
8.86150360e-01 5.88867776e-02 9.15368319e-01 -1.10166228e+00
-3.21762383e-01 -5.27250230e-01 -9.25801992e-01 -1.07708347e+00
-5.14084339e-01 -4.45951998e-01 3.90771627e-01 -1.37273407e+00
8.08753848e-01 -1.82203770e-01 -5.19313812e-01 5.37046313e-01
-5.77022076e-01 9.77466881e-01 5.73577225e-01 3.28697264e-01
-1.39145553e+00 1.73152555e-02 6.68476760e-01 -1.03628866e-01
1.01926336e-02 7.45067373e-02 -3.08065116e-01 9.20647264e-01
3.19604129e-01 -7.31606603e-01 2.13797346e-01 -1.21081851e-01
-1.71864420e-01 -2.40186349e-01 7.74945557e-01 -1.39588463e+00
3.94654661e-01 3.08317551e-03 9.13545072e-01 -1.01706278e+00
2.88288951e-01 -6.75878644e-01 -2.00779274e-01 6.78995013e-01
-1.56517103e-01 8.80865082e-02 3.11967313e-01 6.79198742e-01
-1.96157068e-01 -3.50490928e-01 1.11974156e+00 -2.59013567e-02
-1.24341989e+00 2.49336049e-01 -5.62368631e-01 -7.01079220e-02
1.39884448e+00 -2.50421882e-01 -2.49858320e-01 -9.36442465e-02
-6.22068822e-01 1.38267532e-01 1.77004322e-01 7.97408164e-01
7.69999087e-01 -1.19733024e+00 -1.02816308e+00 -1.00517325e-01
2.20758930e-01 -3.22692931e-01 7.65981823e-02 9.89918172e-01
-4.33466882e-01 2.90331960e-01 -1.28092736e-01 -9.74611700e-01
-1.90852749e+00 6.31075025e-01 2.45666876e-01 5.92908496e-03
-6.23331964e-01 1.18793726e+00 3.83877754e-01 4.07123625e-01
5.92042267e-01 -2.72098482e-01 -1.60282403e-01 1.77351147e-01
1.23161066e+00 4.31191266e-01 4.18997742e-02 -7.16870248e-01
-7.53660858e-01 5.71783543e-01 -3.98599803e-01 -5.82499765e-02
1.20851517e+00 1.27274618e-02 1.55129224e-01 2.37689093e-01
1.06811047e+00 -4.27331537e-01 -1.48095179e+00 -5.06263614e-01
8.72217864e-02 -8.01037252e-01 1.24410748e-01 -6.09337628e-01
-9.93154883e-01 3.81649703e-01 9.74197507e-01 2.30312049e-01
1.01270604e+00 3.43675733e-01 4.86412644e-01 4.84325647e-01
2.73274183e-01 -1.30582070e+00 3.73312056e-01 3.35453689e-01
8.52829456e-01 -1.46575701e+00 1.93287551e-01 -3.74034941e-01
-4.35850531e-01 1.15268719e+00 7.17625320e-01 -1.26722455e-01
5.08408725e-01 3.39480072e-01 -1.32297993e-01 -7.00081959e-02
-7.04709351e-01 -5.55095911e-01 6.65193975e-01 4.16862965e-01
2.73209512e-01 -3.00952017e-01 -1.79954637e-02 2.92805284e-01
7.11062431e-01 1.51733905e-01 4.87455755e-01 1.21434355e+00
-9.09895897e-01 -6.16427422e-01 -7.24326551e-01 5.45145690e-01
-6.97516263e-01 -1.01286799e-01 -2.53889233e-01 5.73027313e-01
2.37446681e-01 1.14129102e+00 2.54709512e-01 -1.04244694e-01
2.67520845e-01 -1.89641744e-01 5.54690599e-01 -6.85041487e-01
-4.69440669e-01 3.15685421e-01 -1.04953229e-01 -9.44114387e-01
-7.34297216e-01 -9.47138846e-01 -1.23081446e+00 3.42967898e-01
-5.63042879e-01 -1.57243416e-01 7.55875528e-01 8.50056648e-01
3.21945280e-01 5.48046887e-01 2.63289899e-01 -1.28240323e+00
-1.82322726e-01 -8.39235067e-01 -4.73874718e-01 4.96755838e-01
5.02501547e-01 -7.25564003e-01 -1.90928616e-02 5.35618007e-01] | [8.663561820983887, -0.21679627895355225] |
dc755bc0-eaca-4730-913a-0e3225ace153 | fine-grained-image-to-image-transformation | 2001.03856 | null | https://arxiv.org/abs/2001.03856v2 | https://arxiv.org/pdf/2001.03856v2.pdf | Fine-grained Image-to-Image Transformation towards Visual Recognition | Existing image-to-image transformation approaches primarily focus on synthesizing visually pleasing data. Generating images with correct identity labels is challenging yet much less explored. It is even more challenging to deal with image transformation tasks with large deformation in poses, viewpoints, or scales while preserving the identity, such as face rotation and object viewpoint morphing. In this paper, we aim at transforming an image with a fine-grained category to synthesize new images that preserve the identity of the input image, which can thereby benefit the subsequent fine-grained image recognition and few-shot learning tasks. The generated images, transformed with large geometric deformation, do not necessarily need to be of high visual quality but are required to maintain as much identity information as possible. To this end, we adopt a model based on generative adversarial networks to disentangle the identity related and unrelated factors of an image. In order to preserve the fine-grained contextual details of the input image during the deformable transformation, a constrained nonalignment connection method is proposed to construct learnable highways between intermediate convolution blocks in the generator. Moreover, an adaptive identity modulation mechanism is proposed to transfer the identity information into the output image effectively. Extensive experiments on the CompCars and Multi-PIE datasets demonstrate that our model preserves the identity of the generated images much better than the state-of-the-art image-to-image transformation models, and as a result significantly boosts the visual recognition performance in fine-grained few-shot learning. | ['Lin Ma', 'Wenhan Luo', 'Jiebo Luo', 'Yixuan Zhang', 'Yutong He', 'Wei Xiong'] | 2020-01-12 | fine-grained-image-to-image-transformation-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Xiong_Fine-Grained_Image-to-Image_Transformation_Towards_Visual_Recognition_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Xiong_Fine-Grained_Image-to-Image_Transformation_Towards_Visual_Recognition_CVPR_2020_paper.pdf | cvpr-2020-6 | ['fine-grained-image-recognition'] | ['computer-vision'] | [ 5.27391195e-01 6.37267977e-02 -1.40831739e-01 -3.36444736e-01
-5.06732702e-01 -6.64059699e-01 6.63496792e-01 -7.34014452e-01
-5.80753312e-02 6.26408696e-01 4.74266708e-02 1.50539935e-01
9.86202583e-02 -1.03741074e+00 -1.02836490e+00 -9.72164273e-01
7.07122326e-01 2.04489067e-01 -1.33315161e-01 -2.60038078e-01
1.47254169e-01 6.91535711e-01 -1.65192223e+00 1.90865546e-01
9.11474586e-01 8.78826976e-01 3.71040702e-02 4.12884295e-01
1.72310006e-02 6.54069066e-01 -4.25494581e-01 -6.59073770e-01
5.42949855e-01 -7.63809204e-01 -6.10451937e-01 3.98142844e-01
8.36026490e-01 -4.01794195e-01 -4.36934710e-01 1.33099544e+00
4.39615726e-01 1.88663214e-01 8.52342129e-01 -1.49871111e+00
-1.28952980e+00 1.48080319e-01 -6.26605570e-01 -2.07211539e-01
1.77706808e-01 4.05982286e-01 6.62706137e-01 -9.02296722e-01
8.23276520e-01 1.44734764e+00 2.15447947e-01 7.70565629e-01
-1.45322323e+00 -9.50831771e-01 -9.71699879e-02 2.56521821e-01
-1.51773715e+00 -5.14367282e-01 1.02020383e+00 -4.34518874e-01
2.64040530e-01 2.94846147e-01 5.39570808e-01 1.28852379e+00
2.00968683e-01 1.49082914e-01 1.15044045e+00 -3.75968963e-01
7.03318641e-02 2.51385048e-02 -6.80029392e-01 8.33905756e-01
2.96269298e-01 2.84232855e-01 -4.08256561e-01 3.14657480e-01
1.17733979e+00 2.15765938e-01 -3.20322514e-01 -6.53936565e-01
-1.26889646e+00 8.09776843e-01 6.00194216e-01 2.21083716e-01
-3.24165881e-01 1.80176906e-02 -1.78017884e-01 2.96797544e-01
3.73639911e-01 5.38318157e-01 9.40616205e-02 3.94682586e-01
-7.63713241e-01 1.59302980e-01 4.98705268e-01 9.99989867e-01
1.12284553e+00 4.28541392e-01 -4.78566855e-01 7.48505533e-01
8.35935678e-03 6.43675983e-01 3.32613021e-01 -1.00178564e+00
3.96665126e-01 6.08073175e-01 2.22944990e-02 -1.14757466e+00
1.90963969e-01 -2.12562591e-01 -1.25510395e+00 6.27146542e-01
3.31893861e-01 4.60403487e-02 -1.17202902e+00 1.91501939e+00
4.46246326e-01 3.11755091e-01 1.04576387e-01 8.38247120e-01
7.48323321e-01 6.23912394e-01 1.10834390e-01 -5.91583140e-02
1.29924703e+00 -9.25814152e-01 -7.08779335e-01 -3.64582300e-01
-6.36857823e-02 -1.00743687e+00 1.14060020e+00 -2.81980932e-01
-1.10297275e+00 -9.98646319e-01 -1.16658843e+00 -2.26113737e-01
-2.59566098e-01 -1.59276813e-01 3.26253325e-01 6.11567497e-01
-7.96954334e-01 4.60818022e-01 -6.07690811e-01 -9.07656252e-02
5.96974611e-01 2.48113066e-01 -7.10957944e-01 -2.97507763e-01
-1.18605483e+00 8.58163178e-01 3.71970415e-01 -1.63445771e-01
-9.13228095e-01 -8.48595679e-01 -1.20115411e+00 1.14697456e-01
2.10014418e-01 -1.00002027e+00 6.80892944e-01 -1.20997787e+00
-1.53386831e+00 1.10923314e+00 3.22967954e-02 3.76150422e-02
5.44947386e-01 3.59199494e-01 -3.54307652e-01 5.54421358e-02
2.02759057e-01 8.66055846e-01 1.39385581e+00 -1.37935472e+00
-4.11745369e-01 -3.85957658e-01 -1.60507336e-02 2.71536291e-01
-3.81717861e-01 -8.47213641e-02 -4.62735236e-01 -1.01032615e+00
-9.81199518e-02 -1.01897514e+00 -3.15295905e-02 2.89074689e-01
-3.25594366e-01 3.05653065e-01 8.27784598e-01 -6.84766829e-01
5.34700513e-01 -2.23640108e+00 3.68431270e-01 7.16471439e-03
6.89347088e-02 2.63554811e-01 -5.68847954e-01 1.89951956e-01
-2.67752886e-01 5.00826500e-02 -2.10945085e-01 -1.72945201e-01
-1.14682704e-01 2.65430093e-01 -3.67005765e-01 5.08376598e-01
4.76600409e-01 1.33252490e+00 -7.27776825e-01 -4.49658394e-01
3.70672226e-01 6.41226470e-01 -5.85805953e-01 6.08598471e-01
6.15836047e-02 9.05319989e-01 -3.89576942e-01 5.17445326e-01
7.70758510e-01 -1.55308887e-01 -2.09040523e-01 -5.24772882e-01
2.54403323e-01 -4.32062715e-01 -1.00965273e+00 1.67303109e+00
-5.58908343e-01 3.27760786e-01 -1.10941499e-01 -9.61415350e-01
1.03140664e+00 2.22698018e-01 4.32644367e-01 -9.25367892e-01
2.22643211e-01 -5.20751625e-02 8.11728761e-02 -3.95893008e-01
1.45571217e-01 -4.67236519e-01 -1.46502286e-01 3.70411158e-01
1.67798489e-01 -4.94708151e-01 -4.88145687e-02 -3.99411879e-02
5.40239394e-01 2.11274445e-01 2.26914272e-01 2.49633053e-03
5.72625935e-01 -3.52806658e-01 6.57546103e-01 3.59019101e-01
5.54164313e-02 9.96904016e-01 1.48364201e-01 -3.90303820e-01
-1.46190703e+00 -1.17358398e+00 7.47886002e-02 7.95956254e-01
5.05484283e-01 2.78432101e-01 -1.00756121e+00 -4.41041440e-01
-1.11527316e-01 6.46627665e-01 -7.80015111e-01 -6.80611372e-01
-5.44965208e-01 -5.15983045e-01 5.03596365e-01 4.09611672e-01
8.07193935e-01 -1.27221596e+00 -1.23160988e-01 2.25956254e-02
-3.59270900e-01 -1.34151423e+00 -1.02818763e+00 -3.22102040e-01
-4.04411733e-01 -1.00295615e+00 -8.51189494e-01 -1.08852863e+00
1.15695119e+00 2.85499275e-01 7.01447785e-01 -2.51405656e-01
-4.58796859e-01 -2.53055543e-02 -1.61881432e-01 -1.92722157e-02
-6.85370445e-01 -2.42101401e-01 1.70069113e-02 7.18243957e-01
-7.49783590e-02 -6.71201587e-01 -6.33213103e-01 5.06794989e-01
-1.18874335e+00 4.28863317e-01 6.74135327e-01 1.10616696e+00
6.06065989e-01 1.13539658e-01 5.72094023e-01 -8.25762689e-01
2.89831012e-01 -9.61769447e-02 -4.49140519e-01 4.38759118e-01
-5.20436764e-01 2.01364964e-01 9.14098501e-01 -8.06486309e-01
-1.34967697e+00 1.22632138e-01 -8.13141763e-02 -7.61271000e-01
-1.33411303e-01 -1.68905184e-01 -7.33816445e-01 -5.41789293e-01
7.05549121e-01 4.36351091e-01 2.06207469e-01 -7.03853071e-02
8.07668746e-01 3.76441777e-01 8.54414940e-01 -5.13336778e-01
1.48040080e+00 6.00665212e-01 6.38786852e-02 -5.18905997e-01
-6.44649506e-01 6.60596564e-02 -7.21896708e-01 -2.32641324e-02
1.11182940e+00 -8.95312488e-01 -5.08026421e-01 7.03935146e-01
-1.04078674e+00 -1.47631615e-01 -5.15297115e-01 2.17484742e-01
-7.63156950e-01 2.20676959e-01 -3.98337454e-01 -1.39733881e-01
-3.77497524e-01 -1.18913615e+00 1.03699613e+00 5.42763352e-01
4.93730307e-02 -7.10464656e-01 -1.53491691e-01 5.22481322e-01
3.95549029e-01 5.68755865e-01 9.93456125e-01 -4.82359044e-02
-8.26231837e-01 -8.18765461e-02 -3.63600880e-01 4.84684080e-01
5.81276715e-01 -1.64045990e-01 -7.52795815e-01 -4.47090417e-01
2.79298183e-02 -3.30013424e-01 5.59715867e-01 -3.47743854e-02
1.13092768e+00 -6.30201578e-01 -1.40558481e-02 1.11697900e+00
1.25734198e+00 2.98645556e-01 1.01190114e+00 2.49124435e-03
1.15454733e+00 5.48315167e-01 5.37764788e-01 7.74661303e-02
1.37767017e-01 7.64293611e-01 3.76752049e-01 -4.24976379e-01
-6.54237390e-01 -5.86800218e-01 1.87879607e-01 4.43700522e-01
-1.66416109e-01 -1.16362922e-01 -3.42839539e-01 4.05638903e-01
-1.58111429e+00 -1.21615314e+00 4.21660274e-01 2.06799912e+00
9.99587774e-01 -3.08526725e-01 -3.94524634e-01 -1.89022750e-01
1.06695175e+00 3.01962376e-01 -7.95110583e-01 -1.18938394e-01
-2.10232720e-01 4.56908077e-01 3.60163361e-01 3.44820321e-01
-9.37838078e-01 1.23263359e+00 5.08583641e+00 9.63571191e-01
-1.31304133e+00 1.10341392e-01 7.87116706e-01 1.59633681e-01
-3.15496922e-01 -8.48927125e-02 -5.50056279e-01 5.07692039e-01
3.04080307e-01 -2.91040748e-01 7.48378694e-01 8.08542013e-01
1.76104140e-02 5.19951880e-01 -9.05693054e-01 1.13373470e+00
3.53876472e-01 -1.39528131e+00 4.70594853e-01 -6.69504376e-03
1.23230946e+00 -7.27783203e-01 3.09095711e-01 1.04174495e-01
2.08878487e-01 -1.20613766e+00 7.12128937e-01 5.97802043e-01
1.42937136e+00 -8.44601989e-01 3.48307699e-01 1.42826661e-01
-1.08107114e+00 1.11907981e-01 -4.78051305e-01 2.25845292e-01
1.73745304e-01 6.04351237e-02 -5.27306139e-01 4.85003859e-01
4.37151402e-01 5.38388908e-01 -5.49151957e-01 6.07795119e-01
-5.28862596e-01 2.08905805e-02 2.82997668e-01 4.11092877e-01
-6.14092173e-03 -3.53245556e-01 2.87101895e-01 5.62675834e-01
5.23759961e-01 1.81833237e-01 -9.81349614e-04 1.24711382e+00
-4.25264239e-01 3.45907807e-02 -8.24626684e-01 4.02383022e-02
4.02473181e-01 1.44578707e+00 -4.37663287e-01 -2.29100838e-01
-3.23239923e-01 1.39176726e+00 2.53851056e-01 5.49315095e-01
-9.93711948e-01 -4.06855702e-01 9.20890450e-01 2.09051207e-01
4.07435596e-01 4.20220494e-02 -1.81759045e-01 -1.32716417e+00
2.69908458e-02 -1.10401750e+00 -4.82301451e-02 -8.05008411e-01
-1.29377460e+00 7.40272522e-01 -2.64732033e-01 -1.30043197e+00
-4.21205014e-01 -3.43713820e-01 -7.35028386e-01 1.04544103e+00
-1.37360215e+00 -1.81872761e+00 -7.67192781e-01 9.50114667e-01
3.94643664e-01 -2.07090154e-01 8.72391880e-01 3.18912894e-01
-3.66337687e-01 8.87067735e-01 -1.34074882e-01 3.78262520e-01
9.47278261e-01 -7.94725478e-01 6.73560143e-01 1.03583086e+00
1.86303392e-01 5.10321140e-01 4.00686532e-01 -5.71702063e-01
-1.34973300e+00 -1.51182759e+00 4.90982592e-01 -2.47380391e-01
3.01579207e-01 -3.81801814e-01 -1.00117803e+00 5.09756505e-01
2.68660367e-01 4.53714043e-01 3.89353096e-01 -5.70175946e-01
-6.19078159e-01 -2.85373390e-01 -1.27358556e+00 7.72897899e-01
1.22321427e+00 -7.18501389e-01 -5.44209778e-01 9.51433554e-02
7.16937423e-01 -3.58991265e-01 -8.82830083e-01 5.98112822e-01
6.32257581e-01 -7.67431915e-01 1.38439560e+00 -7.24100769e-01
5.18860519e-01 -4.20283973e-01 -6.76744059e-02 -1.49489450e+00
-5.73182344e-01 -6.13812268e-01 2.39700824e-01 1.42972469e+00
5.94353303e-02 -5.70016444e-01 7.07421362e-01 6.24049962e-01
1.34553835e-01 -3.92178416e-01 -7.84859478e-01 -8.10645819e-01
1.67224616e-01 3.03088933e-01 8.74037027e-01 1.10652077e+00
-6.16031349e-01 3.71567130e-01 -6.83322370e-01 7.21304193e-02
8.18066120e-01 3.91803235e-01 1.04444921e+00 -8.67140472e-01
-1.58531338e-01 -3.38588566e-01 -5.81558883e-01 -5.75314999e-01
4.63730395e-01 -8.56717288e-01 -1.85835268e-02 -1.23626435e+00
2.46052697e-01 -2.78889388e-01 -1.05244033e-01 5.99587739e-01
-3.81539404e-01 9.30382311e-01 2.31755823e-01 1.97502390e-01
-9.61285681e-02 9.18742418e-01 1.94413948e+00 -4.17036027e-01
1.49171799e-01 -1.01624563e-01 -9.01120365e-01 4.61124659e-01
6.83059871e-01 -3.55294794e-01 -5.87180912e-01 -3.16785604e-01
-2.53763914e-01 4.33475748e-02 5.68669200e-01 -8.02688122e-01
1.05666913e-01 -5.64689100e-01 6.23658419e-01 6.20464012e-02
4.08599824e-01 -7.34894931e-01 6.20653331e-01 3.77764642e-01
-1.87880114e-01 -2.23290071e-01 -9.12249088e-02 3.36478889e-01
-3.25457275e-01 1.48392254e-02 1.39885247e+00 -1.17034860e-01
-7.41758049e-01 8.39797437e-01 2.65827090e-01 1.59294501e-01
1.36802793e+00 -2.65131325e-01 -3.56317997e-01 -3.66348445e-01
-4.94707763e-01 -2.62731820e-01 9.42229807e-01 9.25228417e-01
6.78249300e-01 -1.80679655e+00 -8.28012884e-01 7.53377020e-01
4.27874357e-01 -1.51484683e-01 4.91662860e-01 3.36771935e-01
-4.27015156e-01 4.27878164e-02 -9.31498945e-01 -2.59453952e-01
-1.16702545e+00 8.50721419e-01 2.68111825e-01 4.10362035e-02
-4.66504365e-01 6.03578448e-01 7.75243163e-01 -4.02922928e-01
-1.33765683e-01 9.34986323e-02 -4.02750447e-02 -6.11082502e-02
5.03732741e-01 5.04451469e-02 -1.89221621e-01 -9.29717600e-01
1.45330504e-02 1.01484966e+00 -4.33853827e-03 5.48431128e-02
1.15111959e+00 -1.20977834e-01 -9.32355970e-02 -1.56286091e-01
1.35380244e+00 -1.35559095e-02 -1.64543808e+00 -3.72797370e-01
-7.52251685e-01 -8.70136201e-01 -2.03152910e-01 -5.34627557e-01
-1.38003170e+00 9.12703514e-01 5.63835859e-01 -1.56824097e-01
1.24292839e+00 -1.39705896e-01 7.21018136e-01 -5.89874201e-02
4.18788970e-01 -6.98146522e-01 4.79259253e-01 7.86489099e-02
1.07552230e+00 -1.38596094e+00 -2.59115160e-01 -5.15997410e-01
-6.89351380e-01 7.83779919e-01 8.63819361e-01 -1.88794285e-01
2.67826259e-01 4.60990667e-02 9.94259045e-02 8.34784433e-02
-1.52450442e-01 -1.45348907e-01 5.84874809e-01 8.83451939e-01
-4.10266370e-02 2.12376602e-02 2.10317355e-02 3.05734515e-01
-2.55803734e-01 -2.11055174e-01 2.59763032e-01 4.20320302e-01
-1.24445453e-01 -1.31675065e+00 -4.51268345e-01 7.09894970e-02
-2.87013531e-01 -8.85578394e-02 -2.55098760e-01 7.10022628e-01
3.59671861e-01 6.78941369e-01 -2.27053892e-02 -3.85827124e-01
2.99730897e-01 -1.49324104e-01 7.45807767e-01 -6.57399654e-01
-1.72886714e-01 -1.47117868e-01 -4.16214436e-01 -5.36061943e-01
-2.08290026e-01 -3.68631721e-01 -1.00915730e+00 -4.63434935e-01
-1.48043837e-02 -2.62255400e-01 4.35918629e-01 7.55669594e-01
3.92745525e-01 6.15802765e-01 9.60256934e-01 -1.00270140e+00
-4.89836633e-01 -7.24592090e-01 -5.47362804e-01 9.92869079e-01
1.16957992e-01 -7.38726079e-01 -2.25027487e-01 4.73672539e-01] | [12.003471374511719, -0.316432923078537] |
4639b897-6aff-4376-8c63-3964f381fc9a | qascore-an-unsupervised-unreferenced-metric | 2210.04320 | null | https://arxiv.org/abs/2210.04320v1 | https://arxiv.org/pdf/2210.04320v1.pdf | QAScore -- An Unsupervised Unreferenced Metric for the Question Generation Evaluation | Question Generation (QG) aims to automate the task of composing questions for a passage with a set of chosen answers found within the passage. In recent years, the introduction of neural generation models has resulted in substantial improvements of automatically generated questions in terms of quality, especially compared to traditional approaches that employ manually crafted heuristics. However, the metrics commonly applied in QG evaluations have been criticized for their low agreement with human judgement. We therefore propose a new reference-free evaluation metric that has the potential to provide a better mechanism for evaluating QG systems, called QAScore. Instead of fine-tuning a language model to maximize its correlation with human judgements, QAScore evaluates a question by computing the cross entropy according to the probability that the language model can correctly generate the masked words in the answer to that question. Furthermore, we conduct a new crowd-sourcing human evaluation experiment for the QG evaluation to investigate how QAScore and other metrics can correlate with human judgements. Experiments show that QAScore obtains a stronger correlation with the results of our proposed human evaluation method compared to existing traditional word-overlap-based metrics such as BLEU and ROUGE, as well as the existing pretrained-model-based metric BERTScore. | ['Yvette Graham', 'Liting Zhou', 'Gareth Jones', 'Chenyang Lyu', 'Tianbo Ji'] | 2022-10-09 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [-1.85411185e-01 7.44978338e-02 4.98371482e-01 -2.55008221e-01
-1.34707785e+00 -6.43028140e-01 6.61875427e-01 4.44687635e-01
-6.81646585e-01 8.03334951e-01 5.31795144e-01 -2.35144392e-01
-6.94114566e-02 -9.96226788e-01 -2.18686745e-01 -3.37974399e-01
4.67573285e-01 4.76598710e-01 2.83347964e-01 -6.40719593e-01
5.56223512e-01 -2.66560793e-01 -1.41634572e+00 4.42464143e-01
1.34455097e+00 8.09385002e-01 2.21962005e-01 7.55978405e-01
-4.53551650e-01 8.74760807e-01 -1.12799215e+00 -9.08605695e-01
1.43812727e-02 -1.14384997e+00 -1.13201344e+00 -3.10077101e-01
2.01708272e-01 7.82034621e-02 1.57583039e-02 9.56875741e-01
6.47877097e-01 4.83639032e-01 6.06512249e-01 -8.36347640e-01
-9.71437216e-01 6.47002339e-01 1.17119737e-01 1.83403358e-01
8.91236544e-01 2.39862412e-01 1.37500072e+00 -8.34343791e-01
4.90434319e-01 1.23753905e+00 4.10182863e-01 4.78698671e-01
-7.64730275e-01 -2.18803108e-01 -3.18967968e-01 4.14531827e-01
-1.11133492e+00 -7.87639767e-02 5.06305456e-01 -2.97309428e-01
8.99222672e-01 3.52362752e-01 2.90329158e-01 6.98437989e-01
1.04294434e-01 6.17736697e-01 1.13862240e+00 -8.13042402e-01
3.41248393e-01 2.75822908e-01 4.17488851e-02 5.00912011e-01
-1.18800290e-01 -1.88865408e-01 -5.04253328e-01 -2.18525082e-01
2.08098307e-01 -6.14480853e-01 -3.45220417e-01 1.78337544e-01
-1.22621894e+00 1.04641616e+00 4.29099947e-01 4.64638859e-01
-3.95296901e-01 -2.40703359e-01 2.42905870e-01 2.84872741e-01
3.12783480e-01 1.24894357e+00 -2.67483532e-01 -2.91667223e-01
-8.58135581e-01 6.74642026e-01 9.42791224e-01 5.79729915e-01
5.56095064e-01 -1.96185604e-01 -1.10082078e+00 8.75503004e-01
2.19193980e-01 3.04078192e-01 9.36062157e-01 -9.78456914e-01
5.12582481e-01 8.23152184e-01 4.88492429e-01 -9.94663775e-01
-1.17382213e-01 -4.02195424e-01 -4.24086899e-01 -6.81967735e-02
6.34857476e-01 -2.00270250e-01 -4.50990766e-01 1.61290073e+00
2.38527238e-01 -4.68203127e-01 1.28316998e-01 8.96826982e-01
1.08108377e+00 7.59894490e-01 6.98127002e-02 -1.74048059e-02
1.38525319e+00 -1.08332360e+00 -7.61745512e-01 7.81622976e-02
5.05403757e-01 -9.41358149e-01 1.65233123e+00 2.40864947e-01
-1.02169693e+00 -7.80871928e-01 -9.45481062e-01 -9.65634361e-03
-2.91127384e-01 1.13940313e-02 3.86600345e-02 7.35601366e-01
-1.04101706e+00 3.66198242e-01 -1.33484788e-03 -3.19555700e-01
-1.43252537e-01 -1.69489294e-01 9.20918137e-02 1.45764887e-01
-1.67450249e+00 1.12949002e+00 2.84946561e-01 -1.39150560e-01
-7.39544809e-01 -2.10136652e-01 -4.94360298e-01 1.67861581e-01
2.82137185e-01 -8.31632078e-01 1.57915199e+00 -7.66472101e-01
-1.61509526e+00 7.02149332e-01 -2.00625226e-01 -4.16672945e-01
4.75754440e-01 -1.67649865e-01 -4.84989941e-01 1.24193840e-01
3.06086034e-01 4.69093978e-01 4.38792586e-01 -9.73940909e-01
-6.16183043e-01 8.37889612e-02 5.00326514e-01 3.12695533e-01
-2.64623761e-01 1.51944846e-01 -2.33589545e-01 -5.29813886e-01
-2.25638986e-01 -5.19482017e-01 -2.45768741e-01 -4.90457922e-01
-2.48736426e-01 -7.42975473e-01 -7.13057667e-02 -8.15051615e-01
1.50795007e+00 -1.60949385e+00 -2.89898694e-01 1.40918851e-01
8.51617381e-02 4.79400456e-01 -3.20630342e-01 7.30266154e-01
5.84654689e-01 1.94212660e-01 -1.01303957e-01 -3.58831696e-03
1.77499980e-01 -2.06389889e-01 -6.40033409e-02 -3.33836585e-01
1.22224100e-01 9.72810388e-01 -1.23445857e+00 -6.20499969e-01
-3.22484970e-01 9.29772779e-02 -5.01662314e-01 7.58639216e-01
-5.21713912e-01 2.24628106e-01 -3.25487286e-01 1.03050590e-01
1.67911470e-01 -3.58705699e-01 -2.34877802e-02 1.95448801e-01
1.86876714e-01 6.61981881e-01 -8.69508803e-01 1.41408014e+00
-6.11200631e-01 2.75258690e-01 -5.92009425e-01 -4.31267589e-01
1.28192043e+00 4.84418422e-01 -1.13960363e-01 -9.10831451e-01
1.60948649e-01 1.65403754e-01 1.07162625e-01 -6.63168550e-01
9.03792202e-01 -1.00477368e-01 -2.13344947e-01 6.83766663e-01
1.75241426e-01 -3.09905618e-01 6.32513881e-01 3.50047708e-01
1.10661364e+00 2.10143596e-01 4.01032776e-01 -1.48765892e-02
8.68610799e-01 2.08816752e-02 2.44647488e-01 9.18303490e-01
-2.73026615e-01 7.11365044e-01 3.17551732e-01 -8.35663080e-02
-8.33714247e-01 -9.23990428e-01 3.13161701e-01 1.14366841e+00
-2.26795170e-02 -4.99563098e-01 -1.18443501e+00 -8.81753445e-01
-4.05651242e-01 1.11047924e+00 -4.73856062e-01 -1.68668225e-01
-4.12404507e-01 -6.92473352e-01 6.61007226e-01 2.60021031e-01
7.27008879e-01 -1.41483319e+00 -5.70654929e-01 4.86140609e-01
-9.38276231e-01 -9.33732867e-01 -6.23732686e-01 -3.03017437e-01
-5.25252879e-01 -9.16828573e-01 -8.84931028e-01 -6.88705146e-01
4.32290673e-01 9.50198174e-02 1.57594657e+00 3.15302134e-01
3.00977945e-01 1.67506307e-01 -9.52545524e-01 -3.05429786e-01
-6.86792433e-01 2.68545479e-01 -4.28401053e-01 7.79150613e-03
6.40298784e-01 -4.59030569e-02 -8.42640340e-01 3.47876966e-01
-1.00253892e+00 -2.24872574e-01 4.29639190e-01 6.60981536e-01
2.57042825e-01 -2.19875693e-01 1.10397947e+00 -6.39176250e-01
1.58134854e+00 -4.67408866e-01 -2.00571448e-01 6.70933664e-01
-8.59171450e-01 2.97872812e-01 7.25333810e-01 -1.74456060e-01
-1.19689810e+00 -6.45575762e-01 -3.26001763e-01 4.05811131e-01
3.60522000e-03 6.54344380e-01 -1.17711581e-01 1.67056754e-01
1.07014465e+00 1.99175924e-01 -2.31284395e-01 -3.00154060e-01
4.55427259e-01 6.39917910e-01 4.26918238e-01 -5.42639613e-01
6.88409984e-01 -2.62098402e-01 -4.61020887e-01 -3.81983846e-01
-1.14272702e+00 -5.32775402e-01 -2.42355704e-01 -4.97174323e-01
9.64650095e-01 -5.42844713e-01 -6.26811385e-01 1.76065728e-01
-1.33555603e+00 2.49202959e-02 -3.04297805e-01 3.48103225e-01
-5.16759932e-01 4.88980442e-01 -4.52739388e-01 -8.74426246e-01
-6.95641041e-01 -7.95872867e-01 7.14798987e-01 4.58356589e-01
-7.01325417e-01 -9.13800538e-01 4.61215228e-01 7.87513196e-01
6.35272920e-01 1.24861784e-01 1.13153946e+00 -7.95505941e-01
-2.65539587e-01 -3.24923873e-01 -9.45084617e-02 5.54052591e-01
-4.97216880e-02 -2.28722200e-01 -8.15206885e-01 5.43131195e-02
1.18229672e-01 -4.35783923e-01 5.54204285e-01 -1.13782503e-01
7.81443417e-01 -4.55436170e-01 3.72086138e-01 -1.60211518e-01
1.36578882e+00 2.01841414e-01 6.73697948e-01 5.50192595e-01
1.98996112e-01 7.22749591e-01 7.54921854e-01 2.37206414e-01
6.25076771e-01 6.83805943e-01 1.26714155e-01 2.51032293e-01
-1.09734714e-01 -5.72414100e-01 2.97350287e-01 1.07747495e+00
-4.59353067e-02 -6.66700184e-01 -8.43810439e-01 6.74276114e-01
-1.72901440e+00 -8.77291262e-01 -1.97206005e-01 2.31927371e+00
1.02547550e+00 1.26899332e-01 1.97207555e-01 2.12545648e-01
6.01783276e-01 6.34932145e-03 -3.85224484e-02 -5.55747628e-01
-1.70770615e-01 5.45460403e-01 -2.31908530e-01 6.55131996e-01
-5.97485840e-01 8.18051696e-01 6.38472271e+00 8.14198434e-01
-4.83191609e-01 1.63174555e-01 5.09692013e-01 1.07774675e-01
-6.45525753e-01 -5.95227927e-02 -6.24751687e-01 5.06586909e-01
1.18715370e+00 -6.77311778e-01 2.55022883e-01 5.45157373e-01
2.63647795e-01 -1.21080399e-01 -8.92045259e-01 6.18302822e-01
4.39524829e-01 -9.43587422e-01 3.63660306e-01 -3.38840723e-01
9.35442626e-01 -4.49231297e-01 -2.29297519e-01 6.18539393e-01
4.80250180e-01 -9.86118734e-01 7.41063833e-01 6.74741447e-01
2.54281282e-01 -7.24122524e-01 1.11210942e+00 5.01353204e-01
-7.19229221e-01 1.11267991e-01 -3.62110853e-01 -1.36428058e-01
2.65926540e-01 5.08749962e-01 -1.14922869e+00 5.06514370e-01
4.55760062e-01 -2.82548547e-01 -8.89055789e-01 1.28680527e+00
-6.64233208e-01 7.56769717e-01 2.31425151e-01 -8.18273783e-01
3.96724343e-01 -8.10620487e-02 3.19335341e-01 1.10574901e+00
4.87421960e-01 -9.43903718e-03 -1.48540050e-01 9.63446915e-01
-1.36922047e-01 7.47539937e-01 -1.12209544e-01 8.90994295e-02
4.31785107e-01 1.18281019e+00 -4.27699149e-01 -4.40261275e-01
-2.24923670e-01 9.66979861e-01 3.30290556e-01 3.08957845e-01
-6.53723359e-01 -6.51606321e-01 4.72053885e-02 2.16976658e-01
9.69063714e-02 -4.33976799e-02 -2.30172843e-01 -9.31464136e-01
2.10935012e-01 -1.21199322e+00 3.26027721e-01 -9.73611355e-01
-1.30660331e+00 1.13998747e+00 -1.99060768e-01 -1.21459329e+00
-6.64684653e-01 -2.11476341e-01 -7.78394341e-01 1.23248756e+00
-1.29749632e+00 -6.32226229e-01 -4.80826914e-01 2.72532642e-01
4.25569206e-01 -1.21655330e-01 9.94775057e-01 1.37621447e-01
-1.17314056e-01 6.73236787e-01 -5.55158108e-02 1.56051189e-01
8.27169299e-01 -1.48196244e+00 5.90717494e-01 9.06041920e-01
3.71925950e-01 5.49211144e-01 8.14119875e-01 -3.69928241e-01
-5.00654519e-01 -8.52313936e-01 1.57030690e+00 -6.81959748e-01
2.95547634e-01 7.27902129e-02 -1.00377822e+00 -7.04674646e-02
4.92190033e-01 -7.13248789e-01 9.10095632e-01 -2.84916647e-02
-2.33223721e-01 -3.10245510e-02 -1.11787927e+00 4.71879452e-01
5.98458171e-01 -6.36502922e-01 -9.34033453e-01 3.27833384e-01
8.01768422e-01 -6.35503558e-03 -6.20358765e-01 1.77986234e-01
3.46052110e-01 -9.78652954e-01 5.88675797e-01 -6.63209617e-01
5.73287487e-01 -4.78237689e-01 -5.18682711e-02 -1.57859337e+00
-4.58086282e-01 -4.93042737e-01 3.77370380e-02 1.42508805e+00
7.18771100e-01 -3.24475914e-01 5.56829929e-01 5.69188416e-01
1.19615182e-01 -7.29365051e-01 -7.23644912e-01 -6.96589947e-01
2.21603692e-01 -6.49822652e-02 9.77133274e-01 5.13409078e-01
2.21345797e-02 6.00518227e-01 -1.90650269e-01 -2.45150670e-01
2.53041089e-01 -3.98712084e-02 7.12351263e-01 -1.02107823e+00
-3.76378208e-01 -5.30627310e-01 -8.51618722e-02 -8.80786121e-01
-1.93300575e-01 -8.35738897e-01 5.16636133e-01 -1.90549946e+00
2.13605538e-02 -7.60911480e-02 -3.71552169e-01 -6.09824397e-02
-9.52248693e-01 2.72232413e-01 2.16259778e-01 3.66388224e-02
-8.65412235e-01 6.90759540e-01 1.48856163e+00 6.74642995e-02
-1.08908407e-01 -2.59783734e-02 -1.01006556e+00 4.20660406e-01
8.38951766e-01 -4.52469915e-01 -3.73282611e-01 -4.72353071e-01
6.27884567e-01 1.70630291e-01 1.14804611e-01 -1.20183980e+00
1.58372954e-01 -6.07295856e-02 1.04825810e-01 -2.01394856e-01
-1.67596728e-01 -2.57516235e-01 -3.48701894e-01 1.55392423e-01
-7.47147918e-01 4.40598249e-01 -2.79241115e-01 3.09120357e-01
-6.15692675e-01 -6.80971444e-01 4.65725929e-01 -4.36385185e-01
-4.87053245e-01 -1.11842975e-01 -4.39894199e-01 4.85975116e-01
6.44848347e-01 -6.84603900e-02 -2.67320007e-01 -9.12643552e-01
-2.42964849e-01 2.28878200e-01 7.88158625e-02 6.41334891e-01
5.63411951e-01 -1.37441480e+00 -1.19919813e+00 -3.18147838e-01
4.90975320e-01 -4.31191564e-01 1.06549129e-01 2.77710021e-01
-5.08158743e-01 5.41100085e-01 -1.63805000e-02 -8.06839541e-02
-9.55439508e-01 2.27888599e-01 4.22476739e-01 -8.77701521e-01
1.25844330e-01 8.39839101e-01 -1.76444367e-01 -5.62156320e-01
1.28559321e-01 -2.77012706e-01 -5.91869533e-01 1.40212318e-02
6.93358421e-01 5.77112019e-01 3.54927182e-01 -4.50447023e-01
-1.93955861e-02 1.89708143e-01 1.37107581e-01 -4.77240682e-01
7.52520978e-01 -1.46779776e-01 -7.89809898e-02 2.65234917e-01
8.00158441e-01 8.91894400e-02 -5.20768762e-01 -3.80899161e-01
2.25534156e-01 -2.72317141e-01 -2.72858739e-01 -1.28304422e+00
-4.68767405e-01 7.19000459e-01 3.72437090e-01 4.61477131e-01
9.78884339e-01 -2.54672945e-01 1.09100032e+00 5.48364818e-01
4.30372596e-01 -1.24855208e+00 5.23757994e-01 7.29781985e-01
1.04097497e+00 -1.15378153e+00 -4.33032334e-01 4.22105417e-02
-8.35848987e-01 8.47833574e-01 8.35016727e-01 6.28982782e-02
1.84390441e-01 -4.82417405e-01 4.84222978e-01 -6.59322925e-03
-7.58069634e-01 -3.91676426e-01 6.93761051e-01 4.27848101e-01
8.18101048e-01 1.62574351e-01 -8.13465476e-01 7.18760133e-01
-8.02763522e-01 8.11907426e-02 4.52682436e-01 4.78461474e-01
-7.22657979e-01 -1.24387085e+00 -3.35199922e-01 4.52390075e-01
-5.39895296e-01 -1.71609357e-01 -6.33164763e-01 2.43128821e-01
8.40171650e-02 1.53719985e+00 -3.94337624e-01 -4.35603440e-01
4.98809397e-01 3.42952043e-01 2.51729697e-01 -8.10370922e-01
-1.17616236e+00 -5.18275559e-01 3.43545437e-01 -2.07430676e-01
-3.64763767e-01 -2.55713314e-01 -9.26622808e-01 -9.10042226e-02
-4.61321354e-01 8.50707114e-01 3.51070106e-01 1.07344449e+00
1.75817445e-01 4.17870969e-01 6.29943967e-01 2.64548715e-02
-8.71528983e-01 -1.44053829e+00 -1.72918260e-01 7.85271287e-01
-6.04798757e-02 -1.39416084e-01 -4.05897677e-01 -2.43128374e-01] | [11.548017501831055, 8.253650665283203] |
78ff3ebb-34af-4c38-b77b-9f40d0b82118 | image2stylegan-how-to-embed-images-into-the | 1904.03189 | null | https://arxiv.org/abs/1904.03189v2 | https://arxiv.org/pdf/1904.03189v2.pdf | Image2StyleGAN: How to Embed Images Into the StyleGAN Latent Space? | We propose an efficient algorithm to embed a given image into the latent space of StyleGAN. This embedding enables semantic image editing operations that can be applied to existing photographs. Taking the StyleGAN trained on the FFHQ dataset as an example, we show results for image morphing, style transfer, and expression transfer. Studying the results of the embedding algorithm provides valuable insights into the structure of the StyleGAN latent space. We propose a set of experiments to test what class of images can be embedded, how they are embedded, what latent space is suitable for embedding, and if the embedding is semantically meaningful. | ['Peter Wonka', 'Rameen Abdal', 'Yipeng Qin'] | 2019-04-05 | image2stylegan-how-to-embed-images-into-the-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Abdal_Image2StyleGAN_How_to_Embed_Images_Into_the_StyleGAN_Latent_Space_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Abdal_Image2StyleGAN_How_to_Embed_Images_Into_the_StyleGAN_Latent_Space_ICCV_2019_paper.pdf | iccv-2019-10 | ['image-morphing'] | ['computer-vision'] | [ 6.75599575e-01 5.37678480e-01 -1.12639554e-01 -4.90102917e-01
-1.33129239e-01 -9.53810453e-01 8.52794945e-01 -4.26572382e-01
-1.63577870e-01 4.44739699e-01 4.08929378e-01 -1.42949879e-01
2.98934758e-01 -9.32110131e-01 -9.35840428e-01 -6.19330406e-01
2.34915361e-01 1.23268694e-01 -1.20155461e-01 -5.41129634e-02
-1.28602199e-02 4.81214166e-01 -1.19047451e+00 2.07860485e-01
4.45401311e-01 6.13626122e-01 -2.77433619e-02 6.12138927e-01
-1.48045316e-01 4.48806673e-01 -6.42705619e-01 -6.24659240e-01
5.03118277e-01 -8.40926290e-01 -1.08421969e+00 4.74350303e-01
4.89899933e-01 -5.83827794e-01 -4.02837694e-01 1.13515794e+00
-4.42478396e-02 -8.95873178e-03 6.72633588e-01 -1.36675167e+00
-1.18477476e+00 4.03769463e-01 9.10969377e-02 -2.02070668e-01
3.29587072e-01 6.45958483e-02 8.60361397e-01 -4.72997755e-01
1.34463239e+00 1.39585638e+00 1.95696935e-01 8.21874559e-01
-1.43633604e+00 -3.85410309e-01 -1.95026711e-01 -1.65710911e-01
-1.01913476e+00 -4.72272843e-01 9.39090908e-01 -3.57194126e-01
3.89061749e-01 3.71855080e-01 8.33063662e-01 1.08735514e+00
2.80958027e-01 6.01886988e-01 1.17977095e+00 -5.86148143e-01
1.64655581e-01 5.45614362e-01 -3.21404070e-01 7.22313702e-01
-3.64372367e-03 1.79043069e-01 -3.88689667e-01 -2.44467594e-02
1.01308596e+00 -4.79457416e-02 -1.62941709e-01 -6.74293578e-01
-1.13502240e+00 1.12702382e+00 5.50539494e-01 2.17293173e-01
-5.08545637e-02 4.20268118e-01 2.45619372e-01 6.35126054e-01
5.39058149e-01 7.16591895e-01 -8.17025974e-02 5.49239889e-02
-6.88174784e-01 -5.34885637e-02 5.04704475e-01 1.15267336e+00
9.24552619e-01 4.76035103e-02 -1.14447460e-01 5.27370989e-01
9.31710750e-02 3.77130926e-01 3.12290549e-01 -1.44659019e+00
9.32697803e-02 6.00900292e-01 -1.26113161e-01 -8.14215899e-01
5.17017543e-01 2.04833359e-01 -3.91819239e-01 2.59591192e-01
9.05151144e-02 -6.87343627e-02 -9.25134838e-01 1.74068809e+00
5.01552522e-02 -5.77585958e-02 9.85486358e-02 5.19028187e-01
3.88841033e-01 7.58342147e-01 4.66601066e-02 6.86201751e-02
1.16158473e+00 -1.01848423e+00 -7.82252908e-01 -2.31086642e-01
8.42549920e-01 -8.29623640e-01 1.22349989e+00 -3.21138889e-01
-1.05321026e+00 -5.25005221e-01 -1.05454504e+00 -4.77269948e-01
-4.77560878e-01 -2.79212184e-03 6.28854513e-01 7.52910376e-01
-1.24288976e+00 7.50811994e-01 -5.26483953e-01 -6.40885353e-01
2.34327570e-01 1.26987755e-01 -7.59282887e-01 -4.01090309e-02
-1.04448044e+00 7.65981376e-01 4.25977647e-01 -2.68746883e-01
-8.46918404e-01 -6.06918275e-01 -9.55734432e-01 -6.59464719e-03
-5.29546849e-02 -8.77422094e-01 6.28623903e-01 -1.58919501e+00
-1.72126758e+00 1.34117842e+00 -3.14374119e-02 -2.77745903e-01
5.54062843e-01 6.08765706e-02 -2.06663176e-01 4.56918091e-01
8.88968557e-02 1.21657431e+00 1.26175439e+00 -1.12367094e+00
-2.60317981e-01 -8.42277780e-02 4.52025265e-01 9.94633585e-02
-4.37545151e-01 1.83802679e-01 -3.32756281e-01 -9.02443528e-01
-2.14965671e-01 -1.25185275e+00 3.87151055e-02 4.78372484e-01
-2.86097437e-01 2.53848523e-01 1.07965159e+00 -8.12096834e-01
6.62225306e-01 -2.57523060e+00 5.90829730e-01 2.08818451e-01
1.53331876e-01 -1.60737529e-01 -3.52790892e-01 5.34660995e-01
-3.12348008e-01 5.28873801e-01 -3.22795510e-01 -3.47158194e-01
1.61806688e-01 4.69723791e-01 -5.27412295e-01 3.19257110e-01
1.86230436e-01 1.15796566e+00 -7.00201094e-01 -3.10320646e-01
1.17660202e-01 5.28468132e-01 -8.07731986e-01 4.66893077e-01
-1.67300135e-01 5.35433948e-01 -2.99215078e-01 1.92059800e-01
3.64714056e-01 -2.05809046e-02 1.35186031e-01 -3.04782242e-01
1.65676206e-01 -5.57799684e-03 -7.24306107e-01 1.67189145e+00
-3.32288176e-01 1.15268350e+00 -2.15390414e-01 -5.76403260e-01
8.66010666e-01 1.45435050e-01 1.43778384e-01 -3.79927129e-01
-3.93630974e-02 -4.17896658e-02 -4.40126330e-01 -2.57057309e-01
6.77251935e-01 -5.09745240e-01 -1.38991058e-01 7.23807514e-01
2.92227507e-01 -2.29440868e-01 -1.14537720e-02 2.74137139e-01
8.36232722e-01 1.06883891e-01 1.15192719e-02 -2.71682411e-01
1.35357782e-01 -2.36237705e-01 1.59097120e-01 3.11197072e-01
1.60958633e-01 5.52106082e-01 6.27860606e-01 -5.50600946e-01
-1.48061621e+00 -1.16560936e+00 1.12794980e-01 8.42392743e-01
-2.01889984e-02 -4.69892114e-01 -1.01855230e+00 -8.36165965e-01
-1.01014040e-01 7.58911967e-01 -8.87058198e-01 -5.53767025e-01
-3.10497314e-01 4.83172089e-02 5.56426883e-01 4.60460365e-01
5.77019513e-01 -1.17471588e+00 -5.20503640e-01 -2.63830066e-01
-1.18726671e-01 -1.24681509e+00 -8.25576007e-01 -2.29850158e-01
-9.72182095e-01 -7.20248461e-01 -6.54467642e-01 -8.80198956e-01
1.17582786e+00 7.59699643e-02 9.70332086e-01 1.41241103e-01
-8.20242092e-02 7.68702865e-01 -3.35161120e-01 -1.35441408e-01
-8.27979982e-01 -4.16382551e-02 -1.31163552e-01 1.64335266e-01
1.51198599e-02 -5.21164536e-01 -3.46189439e-01 2.43640080e-01
-1.33683264e+00 2.60775894e-01 3.72985959e-01 7.20060349e-01
3.16239566e-01 -4.57882956e-02 -1.55395329e-01 -1.15166104e+00
4.33555633e-01 -2.30200700e-02 -5.36023855e-01 4.34454620e-01
-4.83607918e-01 3.75053287e-01 2.83367664e-01 -3.46144438e-01
-9.52562332e-01 4.23741080e-02 -5.66941947e-02 -3.86467844e-01
-1.00659899e-01 1.35723889e-01 -4.32387978e-01 -3.16987276e-01
3.37604225e-01 7.08650947e-02 1.56032577e-01 -3.80547553e-01
1.02453864e+00 3.46423954e-01 3.14597070e-01 -4.39997941e-01
1.09409606e+00 7.48885751e-01 -1.09941997e-01 -8.78682315e-01
-5.29067218e-01 5.99129051e-02 -8.78304720e-01 1.54091641e-01
1.26233196e+00 -6.21121407e-01 -6.04521632e-02 2.53895193e-01
-1.05876899e+00 -4.67356354e-01 -9.18877482e-01 7.92628527e-02
-8.35118711e-01 3.29739302e-01 -5.25845647e-01 -1.19376399e-01
4.98081837e-03 -1.19201303e+00 9.58829522e-01 -1.51356980e-01
-5.31316757e-01 -1.43728507e+00 -2.12752242e-02 2.26855919e-01
2.86798924e-01 3.76832366e-01 1.27491093e+00 -1.90081876e-02
-7.14527488e-01 -2.82534342e-02 -1.36691198e-01 5.51990867e-01
3.35858941e-01 1.19358078e-01 -8.76260161e-01 -4.81832653e-01
2.63886936e-02 -2.11028103e-02 6.91564679e-01 1.62394688e-01
9.37369704e-01 -5.77244401e-01 -3.32048029e-01 1.09275675e+00
1.41494954e+00 1.03270680e-01 1.02929962e+00 3.30902785e-01
7.24147856e-01 6.21070921e-01 2.06936583e-01 -1.32325009e-01
-2.35167332e-02 7.65203714e-01 7.25165457e-02 -3.26187491e-01
-1.32749602e-01 -7.55208731e-01 7.11418867e-01 8.06259573e-01
2.33898461e-01 -1.14073522e-01 -3.15810859e-01 4.93941814e-01
-1.35411274e+00 -7.55562484e-01 5.05564570e-01 1.62051797e+00
8.27381313e-01 -8.80509317e-02 -1.44189194e-01 -1.26331404e-01
6.65155232e-01 3.07895005e-01 -1.46666124e-01 -8.14967334e-01
-1.99246883e-01 3.60703200e-01 6.57925427e-01 5.92869759e-01
-8.66420686e-01 1.12672842e+00 7.79455328e+00 4.92550224e-01
-1.24538028e+00 1.93986565e-01 4.84827995e-01 1.01277001e-01
-8.75475407e-01 3.96298289e-01 -3.57123435e-01 3.51690918e-01
7.59369552e-01 -2.73366451e-01 6.70093238e-01 6.97928071e-01
-1.10646643e-01 2.29270741e-01 -1.41087437e+00 7.06169963e-01
7.26431385e-02 -1.55999303e+00 7.64674306e-01 3.27879548e-01
7.54264772e-01 -6.29257262e-01 4.43332762e-01 -1.46320760e-01
1.02876708e-01 -9.86717582e-01 7.96684206e-01 4.45728958e-01
1.17400289e+00 -5.32748580e-01 4.83164698e-01 -1.42881900e-01
-6.81706607e-01 1.38630360e-01 -4.27526176e-01 1.21405154e-01
8.67828205e-02 6.11808486e-02 -5.61779618e-01 2.63873816e-01
4.32337642e-01 7.72436559e-01 -7.89056361e-01 2.98381448e-01
-7.03000307e-01 5.30929804e-01 -7.86996782e-02 4.60535914e-01
1.91069037e-01 -4.41410422e-01 6.55710399e-01 9.11383808e-01
5.47282457e-01 -1.40901916e-02 -3.69274735e-01 1.31971931e+00
-3.75037640e-01 -9.78673697e-02 -1.07930624e+00 -6.37427270e-01
8.93922821e-02 1.11487067e+00 -8.14932585e-01 -3.94646883e-01
-7.40307048e-02 1.73275030e+00 -6.74284771e-02 5.83124399e-01
-6.79398298e-01 -4.76394624e-01 7.60169566e-01 1.58953071e-01
2.88743019e-01 -2.43747264e-01 -7.36874044e-02 -1.18795145e+00
-2.03615516e-01 -6.43794656e-01 7.60961249e-02 -1.15660763e+00
-8.92382920e-01 4.72226113e-01 1.49729505e-01 -9.00391638e-01
-1.65067673e-01 -5.61377406e-01 -7.06682980e-01 7.13947535e-01
-1.15836906e+00 -1.44474602e+00 -2.66063869e-01 4.83067602e-01
3.72029573e-01 -2.63925582e-01 1.06342936e+00 2.93560661e-02
-8.35393667e-02 7.39468932e-01 1.03451557e-01 3.71374846e-01
6.83177710e-01 -1.11885703e+00 6.50345922e-01 8.98294866e-01
2.57837564e-01 7.14661837e-01 7.39557028e-01 -5.58912635e-01
-1.25496328e+00 -1.13856649e+00 7.73455024e-01 -6.17453635e-01
4.09200341e-01 -4.70633298e-01 -6.20621741e-01 1.31110668e+00
5.58171868e-01 -1.30130872e-01 7.90136099e-01 -2.50782043e-01
-3.54946733e-01 1.15802892e-01 -1.20696497e+00 6.79162025e-01
1.12000322e+00 -1.00956023e+00 -5.16488254e-01 1.24156028e-01
1.06512618e+00 -1.59264207e-01 -9.50057805e-01 -1.18150316e-01
6.36869550e-01 -7.07263827e-01 7.99442112e-01 -7.32331395e-01
7.36344039e-01 -2.23196849e-01 -1.58562750e-01 -1.52010095e+00
-4.36954528e-01 -7.64138758e-01 2.85416394e-01 1.31572497e+00
1.51673421e-01 -6.26337469e-01 6.80688918e-01 6.74362183e-01
1.85429245e-01 -1.24639891e-01 -6.46977007e-01 -7.39273787e-01
7.42873698e-02 1.53579712e-01 8.62127781e-01 1.13099897e+00
-3.68488401e-01 2.07465649e-01 -5.00199497e-01 -2.01240271e-01
5.72341025e-01 2.22861961e-01 9.23549533e-01 -7.21381843e-01
-2.77928531e-01 -8.26271176e-02 -7.15204418e-01 -8.50744247e-01
3.54164511e-01 -1.04903185e+00 -4.99503493e-01 -1.25334024e+00
1.44730151e-01 -1.43617719e-01 -7.53856450e-02 4.16071296e-01
1.21602543e-01 5.87776065e-01 6.92676783e-01 2.36979142e-01
-3.03661048e-01 5.89430094e-01 1.39276874e+00 -2.76408792e-01
7.75744841e-02 -5.57264566e-01 -5.92392862e-01 4.99147594e-01
7.78910279e-01 -4.62429404e-01 -4.62688804e-01 -6.12587452e-01
3.67575288e-01 -2.40036011e-01 4.22912031e-01 -4.89399314e-01
-3.28072935e-01 -2.05835655e-01 3.26723397e-01 7.76229054e-02
3.16671103e-01 -8.90276015e-01 6.28842533e-01 3.88810098e-01
-6.21156275e-01 1.28137276e-01 1.41952679e-01 3.53048265e-01
-2.94327527e-01 -2.13215455e-01 8.36377919e-01 -2.03178167e-01
-7.94461966e-01 1.53324544e-01 -2.94934601e-01 -5.95217161e-02
1.01961374e+00 -4.43084627e-01 -3.63673605e-02 -5.53274155e-01
-9.98080194e-01 -2.84742534e-01 1.18890381e+00 5.69475591e-01
6.08589768e-01 -1.92523348e+00 -3.66428912e-01 5.86015165e-01
1.72830820e-01 -5.48908114e-01 -5.46155274e-02 3.32880497e-01
-7.37637281e-01 1.40639186e-01 -9.12887692e-01 -2.90690392e-01
-1.33299649e+00 6.13576591e-01 2.37117052e-01 -1.11710690e-02
-6.07373238e-01 6.76494002e-01 5.30273914e-01 -2.57974386e-01
-2.56350398e-01 -3.06682080e-01 2.31029421e-01 -1.47534713e-01
6.85358644e-02 1.24695420e-01 -5.73465824e-01 -7.10254133e-01
-7.94584211e-03 7.33774483e-01 2.44357258e-01 -2.78501064e-01
1.21933115e+00 -2.98288405e-01 -4.61654991e-01 4.44289088e-01
1.63726807e+00 6.67390674e-02 -1.31576383e+00 3.08810230e-02
-2.77429312e-01 -7.08967268e-01 -9.77155268e-02 -3.54726553e-01
-1.07011247e+00 1.04436255e+00 6.07830465e-01 2.42687181e-01
9.96476710e-01 -2.84320563e-02 8.23241532e-01 1.62879042e-02
2.66403943e-01 -1.13238072e+00 1.80750623e-01 3.00197363e-01
1.09964442e+00 -7.77203381e-01 1.59228463e-02 -4.28621501e-01
-4.78070915e-01 1.10764992e+00 7.20881820e-02 -3.57401818e-01
5.63954771e-01 -4.87496555e-02 1.60356406e-02 -2.24807620e-01
-3.75856876e-01 2.47945994e-01 3.19016039e-01 4.62535203e-01
8.79538730e-02 2.30535150e-01 -2.44353712e-02 -9.59064364e-02
-4.62861806e-01 -1.71696059e-02 5.73488355e-01 8.86466801e-01
-6.06500804e-02 -1.37215602e+00 -4.64451946e-02 5.94287142e-02
-2.04932347e-01 1.38120145e-01 -7.23906755e-01 5.94316483e-01
1.93056569e-01 3.26289594e-01 1.28208265e-01 -2.62691081e-01
1.23233303e-01 2.50699669e-01 8.77204061e-01 -6.82062209e-01
-3.31114441e-01 -2.08680585e-01 -1.00987725e-01 -5.64822733e-01
-4.81052101e-01 -3.99839669e-01 -8.95876348e-01 -4.10543680e-01
-7.78341219e-02 9.14031640e-02 5.45153618e-01 8.31127226e-01
3.34386945e-01 3.51744622e-01 6.00907862e-01 -5.63617527e-01
-1.68810427e-01 -6.07708097e-01 -6.09376729e-01 9.13563669e-01
2.57523686e-01 -3.72490436e-01 -3.49585295e-01 7.50171363e-01] | [11.781255722045898, -0.29270240664482117] |
a91c1f46-872e-4f52-8607-b9080c4c9f01 | learning-homographic-disambiguation | 2304.05860 | null | https://arxiv.org/abs/2304.05860v2 | https://arxiv.org/pdf/2304.05860v2.pdf | Learning Homographic Disambiguation Representation for Neural Machine Translation | Homographs, words with the same spelling but different meanings, remain challenging in Neural Machine Translation (NMT). While recent works leverage various word embedding approaches to differentiate word sense in NMT, they do not focus on the pivotal components in resolving ambiguities of homographs in NMT: the hidden states of an encoder. In this paper, we propose a novel approach to tackle homographic issues of NMT in the latent space. We first train an encoder (aka "HDR-encoder") to learn universal sentence representations in a natural language inference (NLI) task. We further fine-tune the encoder using homograph-based synset sentences from WordNet, enabling it to learn word-level homographic disambiguation representations (HDR). The pre-trained HDR-encoder is subsequently integrated with a transformer-based NMT in various schemes to improve translation accuracy. Experiments on four translation directions demonstrate the effectiveness of the proposed method in enhancing the performance of NMT systems in the BLEU scores (up to +2.3 compared to a solid baseline). The effects can be verified by other metrics (F1, precision, and recall) of translation accuracy in an additional disambiguation task. Visualization methods like heatmaps, T-SNE and translation examples are also utilized to demonstrate the effects of the proposed method. | ['Qun Liu', 'Wei Peng', 'Weixuan Wang'] | 2023-04-12 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 4.90464151e-01 1.35260403e-01 -3.57343316e-01 -1.73164502e-01
-8.27987850e-01 -7.39579141e-01 8.08420360e-01 -6.90734759e-02
-2.84867138e-01 8.29572260e-01 5.52998304e-01 -5.39014578e-01
-9.55114290e-02 -8.29346776e-01 -7.60739088e-01 -6.41988993e-01
3.58922392e-01 5.45776784e-01 -3.10984671e-01 -7.09813297e-01
1.31194085e-01 7.88272247e-02 -8.15780401e-01 1.24390200e-01
1.25658095e+00 4.02037472e-01 3.71349722e-01 2.24623650e-01
-4.22944695e-01 6.70463443e-02 -7.29964912e-01 -8.57265770e-01
4.02540356e-01 -5.38336694e-01 -7.42466033e-01 -1.89063281e-01
4.89580631e-01 5.60710244e-02 -2.55445361e-01 1.37943554e+00
5.48230648e-01 -8.15083552e-03 7.34721839e-01 -8.33852112e-01
-1.37082279e+00 8.33406746e-01 -4.38297451e-01 2.90723115e-01
4.51188773e-01 -1.09345138e-01 1.45051575e+00 -9.57484603e-01
6.67140245e-01 1.52997231e+00 4.62927550e-01 4.88527596e-01
-1.21567321e+00 -7.09575295e-01 -8.01766738e-02 1.19847439e-01
-1.35987902e+00 -1.84088826e-01 5.34368336e-01 -2.06181243e-01
1.34052873e+00 2.16667056e-01 3.66351187e-01 1.28936923e+00
6.81323946e-01 4.52676654e-01 1.24819386e+00 -6.43514633e-01
-1.13728575e-01 7.77056217e-02 4.84925769e-02 6.41528368e-01
4.98530775e-01 8.00682988e-04 -6.75249457e-01 1.80308074e-01
9.01844859e-01 -1.69459879e-01 -1.79291546e-01 -2.42830906e-02
-1.48027897e+00 9.18083012e-01 4.60891902e-01 5.90086579e-01
-3.71999204e-01 -5.37787564e-03 3.95729810e-01 5.13964832e-01
4.71559495e-01 8.19456458e-01 -2.46627361e-01 1.05101839e-01
-7.78507590e-01 -4.42124233e-02 5.15484512e-01 1.14940870e+00
8.00167441e-01 1.32533148e-01 -5.18825173e-01 8.65167916e-01
1.53606191e-01 6.52091265e-01 8.84814620e-01 -2.75671780e-01
7.13949025e-01 6.46437943e-01 -1.00875080e-01 -9.93498445e-01
1.11181878e-01 -8.16355824e-01 -8.21102202e-01 -3.94280374e-01
-2.14000255e-01 -7.32009634e-02 -9.55233574e-01 1.79639578e+00
-2.77192350e-02 8.91156718e-02 4.94382322e-01 9.81376112e-01
5.22907555e-01 8.95131230e-01 -2.85657085e-02 -1.40761971e-01
1.54228079e+00 -1.05379999e+00 -1.16831255e+00 -3.85193944e-01
5.73399305e-01 -1.10784030e+00 1.22483742e+00 -1.62733167e-01
-9.59749997e-01 -6.23002350e-01 -1.33006394e+00 -2.36116350e-01
-5.76415658e-01 1.98232934e-01 4.93262678e-01 3.57566953e-01
-1.03288007e+00 6.24440968e-01 -6.08260095e-01 -6.04452491e-01
3.96137610e-02 4.14591074e-01 -3.06123555e-01 9.08901542e-02
-1.90223932e+00 1.40090966e+00 4.00308579e-01 4.56625670e-02
-5.39236307e-01 -5.27956009e-01 -9.10855949e-01 6.69768900e-02
-8.03728029e-03 -1.06660545e+00 8.81695271e-01 -8.40849996e-01
-1.52916873e+00 8.06730092e-01 -1.49234161e-01 -6.25308037e-01
1.33447260e-01 -3.36213142e-01 -6.03371561e-01 -1.02497496e-01
5.20014703e-01 6.53485060e-01 6.41034305e-01 -8.70371521e-01
-3.77371907e-01 -2.68177986e-01 2.46144950e-01 5.70846796e-01
-4.20852214e-01 -2.77725216e-02 -1.62167117e-01 -8.54593158e-01
1.00433283e-01 -9.37421501e-01 -8.16130191e-02 -6.16511822e-01
-4.81339723e-01 -3.31926078e-01 5.94210505e-01 -7.81386256e-01
1.25881433e+00 -1.80275285e+00 5.78082204e-01 -1.48688465e-01
-1.01881251e-01 3.67962837e-01 -3.13541085e-01 7.28342593e-01
-3.54923680e-03 2.42356703e-01 -2.91664749e-01 -2.59655088e-01
2.18891546e-01 5.76629639e-01 -5.05097330e-01 1.09214090e-01
5.77490211e-01 1.37022626e+00 -9.12590027e-01 -4.46048498e-01
2.00280398e-01 5.45497060e-01 -2.52131969e-01 5.69414683e-02
-1.52061164e-01 3.33722502e-01 -2.59159178e-01 3.81134838e-01
4.15812075e-01 -8.07494298e-02 4.32718337e-01 -3.27832431e-01
-3.49165648e-02 7.70495653e-01 -7.72212267e-01 2.02735114e+00
-8.77750576e-01 5.71967423e-01 -5.94968021e-01 -7.85127819e-01
1.19042492e+00 4.98187900e-01 -6.96376041e-02 -7.36582160e-01
6.96373209e-02 3.17991585e-01 7.04899356e-02 -3.30542862e-01
8.18038404e-01 -3.77228230e-01 -1.65622354e-01 4.20700848e-01
4.48316544e-01 -7.00789969e-03 9.48795080e-02 -9.26936511e-03
8.59860420e-01 2.94861227e-01 3.60214651e-01 -4.37568218e-01
3.80507469e-01 5.75389788e-02 4.10383165e-01 3.83117855e-01
2.82964259e-01 3.93020242e-01 2.46810138e-01 -2.60332018e-01
-1.20480621e+00 -1.03766358e+00 -1.21565834e-01 9.65082824e-01
2.69253820e-01 -5.82164705e-01 -7.53961980e-01 -5.28010249e-01
-1.94863752e-01 9.56546307e-01 -4.92030382e-01 -2.82878757e-01
-7.10820079e-01 -8.78314018e-01 6.94465280e-01 4.90079969e-01
5.12831867e-01 -7.67646015e-01 -2.06361845e-01 2.13067934e-01
-4.47861403e-01 -1.24878764e+00 -6.38900101e-01 2.00872555e-01
-9.93757844e-01 -4.99730706e-01 -6.41435385e-01 -1.01202118e+00
6.05604649e-01 2.16813982e-01 1.05365121e+00 -1.96879268e-01
1.07016191e-01 -1.98830992e-01 -3.27856302e-01 -1.89959239e-02
-5.20578444e-01 4.29921627e-01 2.93275416e-01 -6.03093877e-02
5.54594636e-01 -6.91757083e-01 -4.22578394e-01 1.56205267e-01
-9.13476825e-01 1.31877646e-01 8.74210298e-01 9.70790923e-01
6.06702030e-01 -3.20516348e-01 5.40772378e-01 -8.92263114e-01
1.10161543e+00 -3.32330883e-01 -5.12030125e-01 5.28590262e-01
-9.17688251e-01 5.34666896e-01 5.23107827e-01 -4.46604043e-01
-8.54157746e-01 -2.94877827e-01 -8.62540584e-03 -4.16169077e-01
3.45509052e-01 5.88276863e-01 -2.36280993e-01 2.13286504e-01
6.02537036e-01 4.19032037e-01 -2.37446219e-01 -5.23598135e-01
6.74004972e-01 6.54803038e-01 5.01988232e-01 -5.43094993e-01
1.02349925e+00 -2.62100875e-01 -2.48268709e-01 -2.98661828e-01
-8.00854087e-01 -2.08297685e-01 -8.17592740e-01 2.88444519e-01
9.52523947e-01 -9.31163490e-01 -1.15109533e-01 -2.73708075e-01
-1.42469239e+00 2.15900511e-01 -7.32914731e-02 5.12555003e-01
-3.07605803e-01 2.67493546e-01 -6.70684278e-01 -3.60290021e-01
-7.67769516e-01 -1.22628272e+00 1.31698704e+00 1.58045679e-01
-4.26118046e-01 -1.20439148e+00 1.81695223e-01 4.23094809e-01
4.89655197e-01 -3.21621187e-02 1.42185402e+00 -6.94685698e-01
-6.05209172e-01 9.56954509e-02 -1.63397655e-01 3.02187294e-01
3.61115724e-01 -3.53067130e-01 -8.30124855e-01 -1.81327462e-01
-1.27507806e-01 6.69437274e-02 7.55536497e-01 8.22359547e-02
3.28344256e-01 -6.03432119e-01 -2.16034979e-01 5.40082574e-01
1.66537774e+00 1.34132147e-01 7.43021786e-01 3.94553781e-01
7.63457119e-01 3.36168021e-01 4.92915511e-01 -1.55470788e-01
2.22224191e-01 8.70619535e-01 1.39080122e-01 -1.43308893e-01
-4.49178338e-01 -6.12691820e-01 6.54951811e-01 1.57650459e+00
-8.78705978e-02 -1.98874950e-01 -7.44709969e-01 5.58062375e-01
-1.71413350e+00 -6.37284994e-01 4.85955104e-02 1.89912772e+00
1.14287615e+00 1.19435355e-01 -4.75517482e-01 -5.39278798e-02
9.73182559e-01 2.70043731e-01 -2.51653761e-01 -8.35382760e-01
-3.29988927e-01 5.81554055e-01 5.50964713e-01 8.11353803e-01
-7.74846733e-01 1.59515202e+00 5.75539637e+00 8.72295856e-01
-1.14452922e+00 3.65196258e-01 1.28675327e-01 4.28820074e-01
-8.03204060e-01 1.92592874e-01 -7.87120342e-01 2.08759785e-01
8.05426359e-01 -2.29428038e-01 5.24714708e-01 4.62293178e-01
-2.06431579e-02 4.59655762e-01 -1.05897498e+00 9.13637459e-01
1.62142411e-01 -1.32295382e+00 5.34540653e-01 -6.57694321e-03
8.96323085e-01 -1.23534977e-01 1.51403844e-01 4.37140495e-01
2.46261597e-01 -1.04595160e+00 4.18360084e-01 2.98891485e-01
1.05234754e+00 -6.85407698e-01 9.16678548e-01 1.27138849e-02
-1.02298582e+00 3.77528429e-01 -6.50352418e-01 -9.53491703e-02
8.10687691e-02 4.51803833e-01 -1.18446088e+00 9.83428955e-01
1.18138932e-01 7.47981846e-01 -4.74060953e-01 3.41497421e-01
-7.44711637e-01 4.70205456e-01 1.33664146e-01 -2.04547837e-01
3.83251429e-01 -4.49827999e-01 6.13659143e-01 1.38908768e+00
4.88084048e-01 -2.06666455e-01 -1.85885519e-01 1.12620831e+00
-2.22278461e-01 2.53812969e-01 -7.31943548e-01 -3.83595556e-01
5.39387882e-01 1.23226440e+00 -4.31608021e-01 -3.55958045e-01
-2.83400565e-01 1.43525827e+00 2.34591395e-01 3.76039833e-01
-8.25809121e-01 -3.82102072e-01 8.97885621e-01 -1.42788693e-01
1.25337079e-01 -3.63775015e-01 -3.02387565e-01 -1.34043264e+00
-7.07884654e-02 -9.36959445e-01 -2.95022298e-02 -7.67536461e-01
-1.34645820e+00 9.61054265e-01 -9.01943594e-02 -1.25357783e+00
-3.67587537e-01 -6.36775374e-01 -4.15268660e-01 1.24353158e+00
-1.49714243e+00 -1.19230580e+00 2.06977680e-01 3.31177622e-01
6.69855595e-01 -3.17945302e-01 1.17070830e+00 3.54309529e-01
-6.40692472e-01 8.73779595e-01 7.65780881e-02 2.39157274e-01
7.43749380e-01 -1.31944811e+00 8.57604146e-01 1.04893935e+00
5.68384051e-01 1.18997467e+00 6.60089850e-01 -8.15883279e-01
-1.53061378e+00 -1.20256972e+00 1.44576585e+00 -4.53992754e-01
7.46371150e-01 -5.12684762e-01 -7.42103517e-01 7.79254556e-01
6.98326111e-01 -4.92973119e-01 6.06528938e-01 3.36868405e-01
-4.65615749e-01 5.86681589e-02 -6.66932821e-01 8.04490805e-01
1.13583100e+00 -8.20697188e-01 -1.12705207e+00 3.47055674e-01
1.28668237e+00 -3.89285266e-01 -9.68985081e-01 3.77870351e-01
4.63493705e-01 -3.69161576e-01 9.11815643e-01 -7.68554151e-01
6.33906066e-01 -2.50238270e-01 -3.94172937e-01 -1.60252655e+00
-3.73839974e-01 -6.67942524e-01 1.99124426e-01 1.17971468e+00
6.76092505e-01 -6.59278810e-01 2.20291808e-01 -1.11198969e-01
-2.35212162e-01 -8.19859743e-01 -9.58373368e-01 -7.95269608e-01
2.63685644e-01 -1.68695655e-02 8.20817053e-01 1.15107965e+00
9.44244191e-02 1.10999191e+00 -4.69684213e-01 1.97525740e-01
2.15326592e-01 3.46892864e-01 3.72863173e-01 -8.58156383e-01
-3.25330526e-01 -3.03297818e-01 -4.60457712e-01 -1.07223558e+00
2.88544804e-01 -1.45254993e+00 -1.93201452e-01 -1.69273508e+00
2.42719397e-01 -1.68390930e-01 -5.01461267e-01 4.83560324e-01
-4.06084359e-01 1.81487203e-01 9.15851444e-02 3.63984793e-01
-2.95218647e-01 6.33168340e-01 1.38582063e+00 -2.58434743e-01
1.13013603e-01 -3.59122515e-01 -7.17506766e-01 2.01683521e-01
8.45302284e-01 -6.03621662e-01 -4.35247481e-01 -9.83897805e-01
3.09955359e-01 4.94960546e-02 1.67132914e-01 -6.92374289e-01
3.01675182e-02 -9.06499382e-03 1.60581648e-01 -2.68860191e-01
2.07807645e-01 -5.89020908e-01 1.08168386e-01 3.97656947e-01
-5.27224898e-01 7.28143871e-01 9.92985517e-02 3.27258229e-01
-2.86114007e-01 -2.37770408e-01 3.67709696e-01 -6.90019280e-02
-4.28153396e-01 7.76533335e-02 -8.75017941e-02 5.28559685e-02
4.53732967e-01 -9.13501531e-02 -2.96453565e-01 -1.69175133e-01
-3.31241548e-01 -1.00295134e-01 3.79797339e-01 7.71210492e-01
4.49204773e-01 -1.67463708e+00 -7.23942816e-01 2.15674683e-01
2.32212394e-01 -4.12092656e-01 -2.99798906e-01 7.74205983e-01
-3.13641578e-01 6.72775626e-01 -5.16675115e-01 -3.72071058e-01
-8.93754542e-01 4.27218497e-01 3.05770725e-01 -4.10542876e-01
-4.81340170e-01 7.41585076e-01 1.61234558e-01 -4.56405967e-01
4.12672497e-02 -4.26872969e-01 -8.39546174e-02 -1.94906108e-02
1.60196692e-01 6.33905903e-02 3.17518800e-01 -7.23917782e-01
-3.53523850e-01 6.01060390e-01 -1.15409516e-01 -3.36039454e-01
1.06309867e+00 -1.95495933e-01 -2.03315780e-01 4.07225013e-01
1.20361400e+00 1.44508868e-01 -6.31463110e-01 -3.96617979e-01
1.16785794e-01 -2.38289326e-01 -2.09736913e-01 -8.99504185e-01
-7.57546484e-01 1.08312452e+00 5.87756157e-01 -2.23625690e-01
9.10061240e-01 -1.76659286e-01 1.17704093e+00 2.22997412e-01
3.13994080e-01 -8.55182052e-01 -3.67779471e-02 6.43729806e-01
8.65274370e-01 -1.06269062e+00 -3.60756695e-01 -3.04096282e-01
-6.81393325e-01 1.15799928e+00 3.41927767e-01 -1.45570204e-01
5.02217524e-02 1.77163035e-01 2.84256518e-01 -8.24759975e-02
-6.89716399e-01 -2.29858711e-01 4.90928382e-01 3.60575676e-01
6.65168107e-01 2.67700881e-01 -8.37414801e-01 5.94330847e-01
-6.21205211e-01 -3.61061484e-01 2.21473292e-01 4.78703260e-01
-3.22212905e-01 -1.49566293e+00 -1.28828347e-01 8.46070237e-04
-3.75717759e-01 -6.50930464e-01 -5.73913574e-01 5.81112325e-01
2.25283384e-01 7.66089320e-01 -1.41634822e-01 -7.61487901e-01
1.62968248e-01 3.02148193e-01 5.67996383e-01 -8.04977953e-01
-5.76304138e-01 6.23155348e-02 1.73247486e-01 -2.64019579e-01
-3.15402776e-01 -3.04447293e-01 -1.07735431e+00 -1.45596519e-01
-4.58029807e-01 3.13344389e-01 7.25804865e-01 9.95528162e-01
3.29506189e-01 6.81117535e-01 3.95319939e-01 -3.49042684e-01
-6.58472717e-01 -1.27857244e+00 -1.16349369e-01 4.71581995e-01
2.07993872e-02 -5.94141424e-01 -1.11167684e-01 3.56275542e-03] | [11.539843559265137, 10.09378433227539] |
0d40ee99-2659-41b6-bd91-db76b7c65172 | compressive-mr-fingerprinting-reconstruction | 2006.15271 | null | https://arxiv.org/abs/2006.15271v3 | https://arxiv.org/pdf/2006.15271v3.pdf | Compressive MR Fingerprinting reconstruction with Neural Proximal Gradient iterations | Consistency of the predictions with respect to the physical forward model is pivotal for reliably solving inverse problems. This consistency is mostly un-controlled in the current end-to-end deep learning methodologies proposed for the Magnetic Resonance Fingerprinting (MRF) problem. To address this, we propose ProxNet, a learned proximal gradient descent framework that directly incorporates the forward acquisition and Bloch dynamic models within a recurrent learning mechanism. The ProxNet adopts a compact neural proximal model for de-aliasing and quantitative inference, that can be flexibly trained on scarce MRF training datasets. Our numerical experiments show that the ProxNet can achieve a superior quantitative inference accuracy, much smaller storage requirement, and a comparable runtime to the recent deep learning MRF baselines, while being much faster than the dictionary matching schemes. Code has been released at https://github.com/edongdongchen/PGD-Net. | ['Mohammad Golbabaee', 'Mike E. Davies', 'Dongdong Chen'] | 2020-06-27 | null | null | null | null | ['de-aliasing', 'magnetic-resonance-fingerprinting'] | ['computer-vision', 'medical'] | [-1.74925238e-01 1.17601901e-01 -3.36521119e-01 -5.72548211e-01
-1.02701104e+00 -1.27184018e-02 4.03854460e-01 -3.71398836e-01
-2.31992483e-01 7.69357204e-01 4.65554476e-01 -3.60266268e-01
-4.25835609e-01 -5.31604111e-01 -9.31205153e-01 -7.11216033e-01
-2.47939095e-01 4.08335865e-01 -5.60527854e-02 4.92659174e-02
1.45272657e-01 4.09615666e-01 -8.67514551e-01 9.62047130e-02
6.75988913e-01 1.18640816e+00 5.69423020e-01 4.29448009e-01
3.92166793e-01 1.12919092e+00 2.66746074e-01 -1.67686954e-01
2.88269162e-01 -1.86098292e-01 -9.17300522e-01 -8.30163598e-01
5.60102582e-01 -8.30751479e-01 -9.25302565e-01 9.58337486e-01
1.01792383e+00 2.27113098e-01 3.42608929e-01 -8.01170826e-01
-7.57384777e-01 7.62126863e-01 -3.54466528e-01 3.43540281e-01
2.37853955e-02 -9.61621478e-02 9.49712992e-01 -1.18772233e+00
6.53986573e-01 8.74470651e-01 1.16983128e+00 5.88493109e-01
-1.21008015e+00 -7.65120387e-01 -9.01763067e-02 2.76624411e-01
-1.41413426e+00 -6.31346524e-01 6.98583007e-01 -2.45381802e-01
6.69313729e-01 1.35271624e-01 4.06362414e-01 1.15049756e+00
3.47946137e-01 7.57492661e-01 1.07449126e+00 -6.56245127e-02
8.60501751e-02 -3.06130439e-01 6.00245185e-02 9.15121377e-01
-6.98934942e-02 5.35838425e-01 -6.88075721e-01 -3.23277891e-01
1.11552954e+00 9.15425364e-03 -4.95681792e-01 -4.86873239e-01
-1.38136137e+00 7.01158166e-01 5.87950706e-01 1.62162483e-01
-4.76163924e-01 7.70709395e-01 5.23512006e-01 2.37485737e-01
4.92519468e-01 1.13509640e-01 -4.76071328e-01 -1.70003474e-01
-1.09507728e+00 4.90968585e-01 5.68129241e-01 7.72007763e-01
4.39782023e-01 2.68044192e-02 -2.35457733e-01 1.01280522e+00
4.55749273e-01 7.18065560e-01 3.91301662e-01 -1.28950953e+00
2.74905324e-01 -2.31262058e-01 3.48383039e-02 -1.16947758e+00
-6.63647056e-01 -5.99563539e-01 -9.71989155e-01 -2.00194195e-02
3.10963184e-01 -9.42404345e-02 -6.10181451e-01 1.98554754e+00
5.00155151e-01 7.06758499e-01 -5.57986975e-01 1.33032620e+00
8.33508551e-01 3.84025156e-01 -7.94228390e-02 9.93961468e-02
1.14510369e+00 -9.65773821e-01 -6.04518950e-01 1.06463823e-02
7.27290213e-01 -6.08493924e-01 1.12073863e+00 3.33951920e-01
-1.12266076e+00 -4.78131503e-01 -1.02020872e+00 -4.87399280e-01
1.20257787e-01 2.48516098e-01 7.64475822e-01 3.45089018e-01
-1.09695959e+00 1.04845500e+00 -1.30289125e+00 2.47206196e-01
3.37383717e-01 3.49079043e-01 -2.58857280e-01 -1.36912212e-01
-1.47276402e+00 8.24117780e-01 5.75927943e-02 5.40727973e-01
-1.05348837e+00 -1.32823896e+00 -7.22263157e-01 -2.19161004e-01
1.36452913e-01 -8.50716233e-01 1.24067259e+00 -3.07043195e-01
-1.75603902e+00 6.78342283e-01 5.29897735e-02 -5.86612284e-01
7.43588448e-01 -4.89528745e-01 -2.56979048e-01 1.28699541e-01
7.50685036e-02 5.40780902e-01 6.86085165e-01 -7.74987638e-01
-1.67955756e-02 -2.45237976e-01 -1.35326207e-01 -5.73170520e-02
1.07874483e-01 -2.47495189e-01 -3.68670821e-01 -9.21519339e-01
2.86863327e-01 -1.03801656e+00 -3.79694432e-01 5.62363148e-01
-2.41800919e-01 1.19952478e-01 3.31372470e-01 -9.76830423e-01
1.04982829e+00 -1.91530466e+00 7.87924305e-02 2.82619834e-01
3.82068187e-01 1.08103342e-01 -1.56709790e-01 1.70154959e-01
-1.33021176e-01 -4.53741789e-01 -1.62986502e-01 -2.92569160e-01
1.47794336e-01 -3.23165022e-02 -4.83497918e-01 9.08292115e-01
-3.38106602e-01 1.11372042e+00 -9.86383557e-01 -3.32134336e-01
1.26246288e-01 7.73124516e-01 -6.22715235e-01 1.88132867e-01
8.21049437e-02 5.69319546e-01 -6.25682175e-01 5.52371740e-01
7.27328897e-01 -4.26791877e-01 3.69950771e-01 -6.57677293e-01
7.65718073e-02 6.31162405e-01 -1.08472204e+00 2.31164312e+00
-6.50378108e-01 3.73888493e-01 2.37419039e-01 -1.44091725e+00
8.93660247e-01 3.12267065e-01 7.98760414e-01 -1.19349515e+00
5.49461506e-02 6.28900290e-01 -2.88957059e-01 -3.78650725e-01
2.65890002e-01 -2.26764411e-01 1.71052352e-01 5.92903912e-01
1.78127393e-01 3.69082034e-01 -2.45621800e-01 1.75560579e-01
9.89989102e-01 6.03165388e-01 -1.14254348e-01 -6.66293025e-01
3.70146394e-01 -3.64828408e-01 7.29621172e-01 1.04024482e+00
-1.76224038e-01 8.02293301e-01 4.87213507e-02 -5.28360069e-01
-1.16260231e+00 -1.15575755e+00 -6.27629578e-01 1.02130842e+00
2.94225514e-01 -2.66543955e-01 -5.61788261e-01 -1.77433878e-01
1.83661431e-01 1.85848594e-01 -5.12174308e-01 -2.69786745e-01
-9.19277072e-01 -6.38103366e-01 6.35968328e-01 6.53654873e-01
5.27139425e-01 -7.56981254e-01 -3.20213497e-01 4.89724815e-01
-3.72535467e-01 -1.02033210e+00 -4.96455520e-01 -4.84890081e-02
-9.68087256e-01 -8.29024673e-01 -9.33559775e-01 -5.84976673e-01
3.60203266e-01 -1.01039030e-01 1.17494476e+00 1.66634351e-01
-3.05597603e-01 1.20722353e-01 -1.03647821e-01 2.57637829e-01
-1.32188484e-01 1.84679762e-01 1.31156445e-01 -2.35044196e-01
2.37025078e-02 -9.24514830e-01 -1.15912390e+00 2.24005416e-01
-4.79173124e-01 2.75280714e-01 5.14212489e-01 1.22564161e+00
8.75877976e-01 -7.03166962e-01 6.45661473e-01 -9.51650560e-01
3.59912813e-01 -6.45291626e-01 -4.74290878e-01 1.51281819e-01
-8.64585578e-01 2.84140438e-01 6.03733957e-01 -3.32316637e-01
-1.00734222e+00 -1.37895509e-03 -5.74557304e-01 -4.47817087e-01
2.97478795e-01 7.57002354e-01 2.73656726e-01 -4.32602584e-01
5.95389366e-01 2.15258151e-01 1.65166274e-01 -7.93677807e-01
5.13231218e-01 3.64380062e-01 8.72248232e-01 -1.07066166e+00
4.10464436e-01 5.73684871e-01 -4.03858125e-02 -4.16370451e-01
-9.32737291e-01 -2.58993477e-01 -4.34638143e-01 -1.62039623e-01
5.08536816e-01 -1.09320486e+00 -6.87151194e-01 3.56166691e-01
-8.87290597e-01 -8.15832496e-01 -3.40042152e-02 8.67096126e-01
-8.94255102e-01 5.56391060e-01 -1.06931603e+00 -2.48798519e-01
-8.07564080e-01 -1.05468190e+00 1.10662842e+00 -2.61435896e-01
-2.07266942e-01 -1.08068383e+00 3.19983721e-01 3.31432968e-01
6.55103564e-01 2.95746803e-01 7.11771607e-01 -2.02970549e-01
-5.90042293e-01 1.59287065e-01 -3.35158199e-01 1.33607313e-01
-1.56668097e-01 -5.66691399e-01 -8.43958318e-01 -5.40038884e-01
1.86716318e-01 -4.06687796e-01 8.64528000e-01 5.68516672e-01
1.50161171e+00 -3.05234700e-01 -2.97870487e-01 1.19323707e+00
1.29160035e+00 -3.23661149e-01 7.28434324e-01 2.87639916e-01
7.02136040e-01 1.13416374e-01 4.42364037e-01 4.54443455e-01
3.86846185e-01 7.30110347e-01 -1.16601377e-03 -2.05585569e-01
-4.56674635e-01 -3.25822771e-01 1.90016344e-01 1.08831263e+00
2.57288832e-02 3.47738147e-01 -9.43933129e-01 4.41657364e-01
-1.98522699e+00 -9.48983014e-01 1.31913302e-02 1.85332525e+00
1.09119654e+00 -1.35054082e-01 -1.33608848e-01 -2.56765574e-01
5.34445524e-01 2.45128319e-01 -7.79107988e-01 1.90980211e-02
2.58312732e-01 4.10141319e-01 6.47759378e-01 6.18393660e-01
-7.93405890e-01 4.55286711e-01 6.38492537e+00 9.79259372e-01
-1.46466708e+00 5.54983079e-01 5.91342688e-01 -1.53370440e-01
-4.35500056e-01 -7.13907406e-02 -4.56888735e-01 6.13646448e-01
1.00852776e+00 1.05654225e-01 7.65939593e-01 7.45954216e-01
4.66851473e-01 2.38783926e-01 -1.01600587e+00 1.21681333e+00
-4.24626440e-01 -1.74218416e+00 -2.99525201e-01 -1.23157173e-01
4.80001420e-01 6.42403603e-01 1.05984308e-01 1.28005564e-01
3.07702813e-02 -1.09265769e+00 9.05128539e-01 9.23648000e-01
9.85057950e-01 -4.74187642e-01 4.59081084e-01 2.09014505e-01
-1.03263950e+00 1.09159343e-01 -4.96394932e-01 1.46194980e-01
3.68194163e-01 1.02117276e+00 -3.61723959e-01 4.60034758e-01
8.65382612e-01 7.66967177e-01 7.37113357e-02 7.96853065e-01
-6.32798821e-02 7.00552464e-01 -4.30851549e-01 3.80550295e-01
8.95823017e-02 -2.31868848e-01 3.60077888e-01 1.14947236e+00
2.50625819e-01 1.36350557e-01 -4.79919184e-03 1.28143597e+00
-1.68775812e-01 -2.14740664e-01 -1.60616711e-01 8.70411769e-02
5.80679595e-01 1.19318032e+00 -3.14283997e-01 -2.22802818e-01
-3.78890812e-01 8.31110954e-01 5.45715213e-01 3.95663738e-01
-1.12297237e+00 -4.58952457e-01 5.92581987e-01 1.68685749e-01
2.22064957e-01 -3.79622757e-01 -3.70323747e-01 -1.25042927e+00
3.39646488e-01 -7.37632811e-01 1.53468326e-01 -7.96741366e-01
-1.43497705e+00 3.87563050e-01 -9.53330472e-02 -8.84126723e-01
-2.63033807e-01 -6.61491454e-01 -3.85913461e-01 9.18149650e-01
-1.54735756e+00 -1.09361207e+00 -1.90693870e-01 7.07777858e-01
-1.27994776e-01 9.31518748e-02 7.56600857e-01 8.58054161e-01
-2.66857713e-01 7.47656763e-01 4.88739014e-01 1.93993211e-01
6.31227374e-01 -1.02167857e+00 4.66981381e-01 6.33638203e-01
-8.16441253e-02 9.60099578e-01 5.95814705e-01 -5.02244473e-01
-1.68815219e+00 -9.49380577e-01 6.65149629e-01 -1.61207765e-01
8.26594114e-01 -4.07745689e-01 -1.13904405e+00 5.39216399e-01
-2.47269750e-01 6.45552576e-01 3.10344458e-01 9.02341977e-02
-3.63315433e-01 -3.13466251e-01 -1.06351233e+00 3.09417754e-01
1.30137491e+00 -1.16307771e+00 -3.21196795e-01 4.38230038e-01
5.07068753e-01 -9.27581072e-01 -1.22960138e+00 5.61549246e-01
8.64897370e-01 -6.79330707e-01 1.29977536e+00 -3.56110364e-01
4.20719385e-01 -1.95575699e-01 -7.06526190e-02 -9.18912292e-01
-4.67079580e-01 -6.05234623e-01 -4.98610705e-01 7.25209773e-01
2.74974495e-01 -6.59991860e-01 8.34191322e-01 7.68095255e-01
-2.33331263e-01 -1.14524662e+00 -1.05524099e+00 -7.79986560e-01
3.72873336e-01 -4.63247359e-01 6.20124340e-01 9.56240952e-01
-6.23322465e-02 -1.82311016e-03 -8.22796822e-01 9.47450623e-02
9.05452728e-01 2.75740594e-01 3.11956346e-01 -7.78287172e-01
-6.38125837e-01 -1.87537014e-01 -1.70873687e-01 -1.49157405e+00
2.10351437e-01 -1.29070365e+00 4.64646406e-02 -1.34003496e+00
2.24380851e-01 -8.77489150e-01 -5.97255468e-01 3.77188027e-01
7.31133819e-02 2.99607486e-01 -6.65397570e-02 3.44650388e-01
-2.95316875e-01 7.05307245e-01 1.38475013e+00 2.58113984e-02
2.26935759e-01 -2.95600235e-01 -4.76511449e-01 4.33358431e-01
7.38390684e-01 -6.07096255e-01 -3.61353368e-01 -8.05890799e-01
3.10823232e-01 3.65526885e-01 6.66048408e-01 -8.42223942e-01
4.13910776e-01 1.08188316e-01 3.17593932e-01 -3.55063528e-01
2.29606420e-01 -4.82688129e-01 3.99663240e-01 5.96741140e-01
-6.64849520e-01 -2.95963809e-02 -4.27564271e-02 4.52256113e-01
1.49740372e-02 1.71008166e-02 8.49800885e-01 5.16802678e-03
-6.19895577e-01 7.75518119e-01 2.86645908e-02 2.31856763e-01
2.14621633e-01 1.94696337e-01 -3.50238346e-02 -1.99301258e-01
-8.66315007e-01 8.85339603e-02 -5.14691975e-03 2.21383587e-01
5.82814872e-01 -1.64126468e+00 -6.14887297e-01 1.35851474e-02
-3.76456767e-01 -2.21274301e-01 4.30953324e-01 1.36806393e+00
-6.82192445e-01 3.40579897e-01 -1.06875531e-01 -5.89007318e-01
-4.70928550e-01 2.64651299e-01 8.60940874e-01 -1.93806812e-01
-1.13364840e+00 7.29789555e-01 2.78385524e-02 -8.81800115e-01
2.62591749e-01 -3.07216257e-01 4.99158084e-01 -5.95626354e-01
5.84096551e-01 5.86109936e-01 2.25817844e-01 -2.52349317e-01
-4.56007332e-01 3.87721539e-01 -2.85869446e-02 -8.99240151e-02
1.65849292e+00 -1.44432321e-01 -9.61915329e-02 4.46998589e-02
1.64149010e+00 -2.03490138e-01 -1.32825637e+00 -5.63042879e-01
2.85597611e-03 -3.29515606e-01 5.77903628e-01 -8.49166453e-01
-1.42028213e+00 8.06461334e-01 8.84086967e-01 -4.98438030e-01
7.25407958e-01 -1.34179890e-01 1.42416775e+00 5.15256286e-01
5.45330346e-01 -1.06736624e+00 -2.10645765e-01 3.50596875e-01
9.53065753e-01 -1.09654117e+00 1.08900763e-01 -8.87488276e-02
-1.24584302e-01 1.17196882e+00 2.53658593e-01 -4.29151535e-01
1.07618403e+00 5.05733848e-01 -5.56016713e-02 -3.93257678e-01
-4.72082764e-01 4.51936394e-01 3.02653015e-01 2.70957589e-01
6.14583790e-01 2.56412998e-02 -4.46520656e-01 5.60666800e-01
-1.06205195e-01 4.37813699e-01 -1.15677387e-01 6.83142483e-01
-1.16556413e-01 -1.02456594e+00 1.48541033e-01 4.15040165e-01
-6.08760536e-01 -3.11323941e-01 2.83088297e-01 4.75125611e-01
-1.24841988e-01 2.34419316e-01 -2.20094293e-01 -2.12087929e-01
1.81929752e-01 -4.31982905e-01 6.37007117e-01 -4.67171036e-02
-2.61692762e-01 -1.20543607e-01 6.98622223e-03 -1.20805669e+00
-4.45402086e-01 -5.44464767e-01 -1.55571556e+00 -5.90232968e-01
-2.57281423e-01 2.57330295e-02 4.89861637e-01 8.92112792e-01
6.49834037e-01 1.94313005e-01 5.16569495e-01 -8.60084474e-01
-9.40375745e-01 -6.94935739e-01 -6.17686450e-01 3.42680335e-01
4.63675678e-01 -8.40488911e-01 -1.53792292e-01 -2.65147388e-01] | [13.526860237121582, -2.395923137664795] |
61fc1389-7a5a-415d-b7ba-52dba9c0f214 | scalable-weight-reparametrization-for | 2302.13435 | null | https://arxiv.org/abs/2302.13435v1 | https://arxiv.org/pdf/2302.13435v1.pdf | Scalable Weight Reparametrization for Efficient Transfer Learning | This paper proposes a novel, efficient transfer learning method, called Scalable Weight Reparametrization (SWR) that is efficient and effective for multiple downstream tasks. Efficient transfer learning involves utilizing a pre-trained model trained on a larger dataset and repurposing it for downstream tasks with the aim of maximizing the reuse of the pre-trained model. However, previous works have led to an increase in updated parameters and task-specific modules, resulting in more computations, especially for tiny models. Additionally, there has been no practical consideration for controlling the number of updated parameters. To address these issues, we suggest learning a policy network that can decide where to reparametrize the pre-trained model, while adhering to a given constraint for the number of updated parameters. The policy network is only used during the transfer learning process and not afterward. As a result, our approach attains state-of-the-art performance in a proposed multi-lingual keyword spotting and a standard benchmark, ImageNet-to-Sketch, while requiring zero additional computations and significantly fewer additional parameters. | ['Simyung Chang', 'Seunghan Yang', 'Jun-Tae Lee', 'Byeonggeun Kim'] | 2023-02-26 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 1.97672799e-01 -9.69293863e-02 -3.55881929e-01 -2.58371532e-01
-1.08201742e+00 -2.32840002e-01 5.24845839e-01 -9.13056657e-02
-8.96671653e-01 6.11095548e-01 1.04146570e-01 -4.67176616e-01
1.81043241e-02 -5.54828167e-01 -7.45019674e-01 -7.00071156e-01
2.24918216e-01 5.67216814e-01 1.56977296e-01 -1.24297790e-01
5.30589700e-01 4.07204300e-01 -1.28605986e+00 9.67202932e-02
7.37776518e-01 9.39419925e-01 7.10713863e-01 4.20532495e-01
-7.71120042e-02 4.01002735e-01 -3.64712954e-01 -4.81404603e-01
3.55952442e-01 -2.16910347e-01 -7.45911479e-01 4.42664418e-03
3.16086262e-01 -4.67191458e-01 -2.17211157e-01 8.11878622e-01
5.54401994e-01 3.48334879e-01 8.74744892e-01 -1.06267643e+00
-7.64037192e-01 6.64710402e-01 -6.63265705e-01 1.47460476e-01
-2.62631685e-01 -1.03490375e-01 1.11608052e+00 -1.16986525e+00
3.00378770e-01 9.54685986e-01 4.51633304e-01 5.04880548e-01
-1.17833745e+00 -9.35438395e-01 7.22349733e-02 3.02855670e-01
-1.72609591e+00 -5.48621237e-01 6.18290126e-01 -1.73559368e-01
1.07584393e+00 3.53794992e-02 4.32607412e-01 8.02773058e-01
6.79578409e-02 5.85298359e-01 7.33078301e-01 -7.84766674e-01
1.49619550e-01 4.14264262e-01 -2.41521716e-01 6.50574982e-01
1.96028307e-01 -3.83082628e-01 -5.62724769e-01 -2.12203875e-01
8.29706371e-01 -6.03567213e-02 -2.13592708e-01 -5.28921366e-01
-1.26171350e+00 9.10981357e-01 1.30806997e-01 3.87203068e-01
-2.55822420e-01 1.59969404e-01 4.76945281e-01 4.99681830e-01
7.74894536e-01 3.77777338e-01 -6.70297682e-01 -1.11437656e-01
-1.08959687e+00 -7.69662112e-02 5.82200527e-01 9.80683208e-01
1.17142296e+00 -1.94592178e-01 -4.31729145e-02 1.17393768e+00
1.75402522e-01 1.55403093e-01 8.39461327e-01 -6.40242636e-01
8.45336080e-01 4.60671604e-01 1.37058005e-03 -6.25310421e-01
5.43323159e-02 -1.44502118e-01 -6.37792945e-01 -6.52026013e-02
3.24801624e-01 -1.23904906e-01 -9.74686801e-01 1.72912240e+00
2.20230699e-01 3.42596203e-01 3.54469679e-02 7.44346082e-01
2.01221287e-01 8.79942954e-01 2.17597306e-01 1.23876102e-01
1.20485413e+00 -1.19347978e+00 -3.07761878e-01 -3.37437510e-01
1.00643778e+00 -1.00762343e+00 1.28796482e+00 2.75447667e-01
-8.70700240e-01 -4.91291821e-01 -1.12135196e+00 -3.60768229e-01
-4.73000377e-01 6.93652153e-01 3.92092466e-01 4.21638519e-01
-1.10480654e+00 6.56875610e-01 -6.56798422e-01 -5.14730513e-01
1.82673916e-01 5.72842538e-01 -4.90028143e-01 -1.29503325e-01
-1.02627575e+00 1.20959628e+00 5.51766455e-01 -1.19239375e-01
-5.89632392e-01 -8.45541358e-01 -6.79063439e-01 3.22551668e-01
3.56030613e-01 -5.24565041e-01 1.20126629e+00 -8.89059961e-01
-1.82062292e+00 7.11831570e-01 -6.63885549e-02 -3.69591951e-01
3.98863912e-01 -1.77798197e-01 -8.97944197e-02 -4.00520004e-02
-6.08793683e-02 9.24829125e-01 1.18002379e+00 -8.37461174e-01
-6.57265961e-01 -1.44622594e-01 6.43387958e-02 4.58898574e-01
-1.03959370e+00 8.97836536e-02 -9.23054159e-01 -8.70206535e-01
-1.65065050e-01 -1.11895120e+00 -1.50477231e-01 1.62300676e-01
6.73917979e-02 -5.63846588e-01 7.67701149e-01 -6.94097638e-01
1.20304871e+00 -2.19182062e+00 4.49045092e-01 1.27022654e-01
-7.79499561e-02 4.61027592e-01 -4.58263129e-01 4.31255072e-01
8.73475522e-02 6.58489242e-02 -1.97095647e-01 -6.15681708e-01
-2.22822681e-01 2.07468346e-01 -2.49232620e-01 3.69226515e-01
6.25295490e-02 6.70121193e-01 -6.08910859e-01 -6.32927775e-01
2.46238604e-01 4.15487438e-01 -6.69613361e-01 3.22931677e-01
3.92410122e-02 1.64413869e-01 -5.04211307e-01 2.79958576e-01
5.92256427e-01 -2.54337370e-01 4.66955127e-03 -3.42819601e-01
-2.07163706e-01 2.83846915e-01 -1.05431569e+00 1.85010540e+00
-9.29613829e-01 5.33460796e-01 -1.26774788e-01 -1.22522736e+00
8.89451504e-01 3.25275630e-01 4.69975621e-01 -5.08240283e-01
3.33006047e-02 4.77548867e-01 -1.43923298e-01 -3.78257155e-01
5.08449376e-01 1.34635061e-01 1.21012419e-01 7.39635289e-01
2.55830616e-01 -1.18799796e-02 5.63396327e-02 1.03178009e-01
7.51321971e-01 2.10085541e-01 3.22869539e-01 -3.11719090e-01
6.18082225e-01 -1.40077159e-01 2.45693818e-01 4.57119912e-01
1.18896641e-01 3.97406399e-01 2.75768284e-02 -2.92148650e-01
-1.34281993e+00 -5.19640028e-01 6.81294799e-02 1.64024329e+00
3.72419432e-02 -3.44301790e-01 -7.21673369e-01 -4.95298982e-01
5.63706048e-02 8.05194378e-01 -6.07859790e-01 -3.12276751e-01
-7.38180220e-01 -5.24339497e-01 5.50012529e-01 3.75495762e-01
4.12101179e-01 -9.10124779e-01 -3.48378927e-01 1.39565811e-01
-1.51240170e-01 -9.51539397e-01 -1.04120111e+00 5.02583347e-02
-9.47955489e-01 -6.48287535e-01 -1.20590353e+00 -9.12547946e-01
9.95641530e-01 5.94700873e-01 6.75359845e-01 4.42303009e-02
-1.21570416e-01 1.33722961e-01 -2.64979661e-01 -2.82782793e-01
-2.56272703e-01 6.76664889e-01 -1.42232701e-01 1.37261719e-01
3.62349987e-01 -3.10529530e-01 -5.78019083e-01 3.33909988e-01
-8.75505328e-01 2.85905480e-01 8.88318837e-01 7.65610218e-01
3.90550643e-01 -4.95266736e-01 5.99146664e-01 -7.16135740e-01
8.41858387e-01 -3.84652704e-01 -5.44037461e-01 7.11275160e-01
-9.43274319e-01 4.57548857e-01 6.93009555e-01 -7.66079724e-01
-1.17353296e+00 -1.14576280e-01 1.68252185e-01 -6.03057086e-01
3.32321793e-01 4.84228998e-01 1.59926221e-01 -2.76459843e-01
4.61430281e-01 3.00465316e-01 7.58885741e-02 -6.29829943e-01
5.87993145e-01 8.54226947e-01 7.34567866e-02 -6.89527869e-01
7.53852010e-01 1.82180583e-01 -3.38953584e-01 -7.63808489e-01
-6.07512176e-01 -6.78982258e-01 -7.81169415e-01 3.51678617e-02
5.08343160e-01 -9.49198902e-01 -3.73516500e-01 3.60242575e-01
-1.09286320e+00 -3.67076665e-01 1.31126910e-01 6.16062343e-01
-4.64147717e-01 3.44419688e-01 -4.51822937e-01 -4.00599420e-01
-6.78070188e-01 -1.18011963e+00 1.01305676e+00 -1.01625018e-01
-1.57837942e-01 -1.01740229e+00 7.25276917e-02 2.77375638e-01
8.59254241e-01 -5.94784319e-01 1.02876687e+00 -6.53982282e-01
-5.00518262e-01 -1.47370189e-01 -4.08375680e-01 4.04782861e-01
1.99348047e-01 -2.33920828e-01 -7.01988935e-01 -5.60850203e-01
-2.62354016e-01 -4.95453566e-01 9.66407359e-01 1.34524088e-02
1.36839461e+00 -3.40811461e-01 -4.25333887e-01 5.94956279e-01
1.30132329e+00 4.63431180e-02 4.35842723e-01 4.21289146e-01
6.06398404e-01 4.78550106e-01 5.65019906e-01 3.13146144e-01
4.20747757e-01 7.63326406e-01 -1.00728840e-01 -6.13261573e-02
-1.08080618e-01 -3.17440450e-01 2.34566152e-01 1.04902709e+00
-1.77063599e-01 -2.53373869e-02 -6.92946553e-01 6.56977713e-01
-1.93720257e+00 -5.25621593e-01 6.71680272e-01 2.39211583e+00
1.15277135e+00 -1.58817187e-01 -1.65317982e-01 -3.00805360e-01
6.90642476e-01 1.13997407e-01 -6.22186422e-01 -4.07535106e-01
3.94861519e-01 3.83565515e-01 8.25977623e-01 4.21148002e-01
-7.67033398e-01 1.36915314e+00 5.86228704e+00 1.11829007e+00
-1.48240161e+00 2.06466570e-01 4.09018397e-01 -6.85550719e-02
-2.32820094e-01 1.08190753e-01 -8.59028339e-01 2.73370922e-01
1.03177214e+00 -5.43162107e-01 6.87140048e-01 9.68391955e-01
4.40841243e-02 -3.98615412e-02 -1.11445892e+00 8.43433201e-01
2.79462665e-01 -1.26302242e+00 3.53839844e-01 2.27741897e-02
5.37579596e-01 2.73732871e-01 2.73234155e-02 6.37217879e-01
-1.34727033e-02 -5.65432608e-01 5.33888578e-01 1.31670088e-01
8.25461328e-01 -8.46046388e-01 4.27509159e-01 4.08519328e-01
-1.01347756e+00 1.02045476e-01 -7.09264278e-01 1.90651223e-01
-1.50460705e-01 2.05469757e-01 -1.01466405e+00 2.50877887e-01
5.37951589e-01 3.69610280e-01 -3.89336824e-01 9.21876669e-01
-9.65944231e-02 4.72091049e-01 -2.90325522e-01 -8.40481222e-02
3.23061407e-01 -2.88410246e-01 4.23564091e-02 1.19986379e+00
6.67245924e-01 -1.17140815e-01 1.02421246e-01 5.19254506e-01
-3.58477980e-01 5.02487063e-01 -4.41643804e-01 -4.49212082e-02
6.17573202e-01 1.37589860e+00 -4.96651590e-01 -4.39065903e-01
-6.08852744e-01 1.12093663e+00 7.99210548e-01 4.26861852e-01
-8.09609711e-01 -4.77498502e-01 4.94473904e-01 7.11116716e-02
4.83190030e-01 -4.15352225e-01 -1.73918884e-02 -1.29190898e+00
-8.22517946e-02 -6.29330456e-01 3.19020927e-01 -5.05460799e-01
-1.06590199e+00 3.97698551e-01 2.46406034e-01 -1.15080535e+00
-3.15887868e-01 -5.27760029e-01 -4.82877761e-01 1.01752782e+00
-1.85218382e+00 -1.24899793e+00 8.98564458e-02 6.86507344e-01
7.88330436e-01 -2.60734528e-01 9.32474256e-01 4.27165508e-01
-5.46967685e-01 9.10605371e-01 2.14082301e-01 -6.80791363e-02
1.18822515e+00 -8.86588514e-01 2.70613223e-01 6.03020847e-01
1.13888085e-01 6.85227275e-01 3.79664540e-01 -2.91965902e-01
-1.35155487e+00 -1.13161552e+00 1.04030859e+00 2.59373579e-02
8.64382386e-01 -4.03322935e-01 -9.68824208e-01 8.03516984e-01
2.55753636e-01 -5.75588048e-02 6.56917691e-01 9.18833539e-02
-4.03893590e-01 -2.35812590e-01 -9.89022851e-01 8.51101875e-01
7.22328067e-01 -5.44012547e-01 -5.30282915e-01 3.42978448e-01
7.22347438e-01 -2.98844934e-01 -8.49595308e-01 2.25732565e-01
5.86167097e-01 -2.14739159e-01 9.50427890e-01 -4.14499521e-01
2.89025366e-01 3.71934613e-04 1.03052728e-01 -1.46970570e+00
-3.72595340e-01 -5.41379452e-01 -1.33507416e-01 1.06227088e+00
6.75439596e-01 -7.77869284e-01 6.96981132e-01 8.30377460e-01
-3.09791118e-02 -8.01990569e-01 -1.11085987e+00 -6.12592816e-01
1.16901189e-01 -2.02423304e-01 4.06210214e-01 9.11160231e-01
-1.20216869e-01 3.59863758e-01 -6.73463464e-01 1.33996196e-02
5.05590737e-01 3.06115840e-02 7.90202081e-01 -9.09393191e-01
-1.26678243e-01 -2.64434785e-01 1.53158471e-01 -1.22400737e+00
2.81361014e-01 -9.16046560e-01 -3.69507819e-02 -1.26499152e+00
2.24851951e-01 -6.36887550e-01 -5.20470679e-01 8.98185015e-01
-1.86037078e-01 2.08681867e-01 2.11581588e-01 6.38859153e-01
-4.54909891e-01 5.47519505e-01 1.26390553e+00 -8.91026929e-02
-1.92864716e-01 2.13648728e-03 -6.63034737e-01 4.66768354e-01
1.02341175e+00 -5.37210703e-01 -6.20173037e-01 -7.36077130e-01
9.35004372e-03 -3.02040726e-01 -7.49625489e-02 -6.06490731e-01
5.46477079e-01 -2.39741072e-01 1.46226421e-01 -3.84106249e-01
5.11113405e-01 -7.90630698e-01 -2.60977596e-01 2.77764827e-01
-4.97811347e-01 1.10209063e-01 1.72098815e-01 4.94734645e-01
-1.13159619e-01 -5.49620271e-01 8.03601027e-01 -3.95215116e-02
-8.39827716e-01 3.40618044e-01 -2.63366103e-01 -2.63024509e-01
1.11424267e+00 -6.44974485e-02 -3.10091618e-02 -2.45061427e-01
-3.73632729e-01 1.66727841e-01 2.97678888e-01 7.05818474e-01
4.96914744e-01 -1.25108838e+00 -6.10999703e-01 2.82115459e-01
3.05034906e-01 -2.45921642e-01 -1.18171563e-02 5.29864669e-01
-3.60647202e-01 6.36022508e-01 -1.63935542e-01 -3.25341940e-01
-1.17938495e+00 4.88702327e-01 -1.25523079e-02 -4.05815482e-01
-5.64241230e-01 8.10059309e-01 -4.52823974e-02 -4.40272719e-01
2.49460742e-01 -3.79105389e-01 -1.17233641e-01 1.33494973e-01
3.46743613e-01 3.23381901e-01 2.68731862e-01 -4.14342314e-01
-1.25013456e-01 7.22505331e-01 -6.53658926e-01 -1.71355739e-01
1.30644071e+00 -1.49952844e-01 -4.57916260e-02 1.90929458e-01
1.47669876e+00 -2.14611471e-01 -1.20618248e+00 -6.58052266e-01
7.84881562e-02 -4.10084814e-01 3.33742648e-01 -3.47436607e-01
-9.19125795e-01 7.72371590e-01 3.63337696e-01 -3.39368016e-01
9.20050263e-01 -1.14927463e-01 8.76188636e-01 7.52010882e-01
5.78594148e-01 -1.42366397e+00 2.06746072e-01 4.56993043e-01
8.66001248e-01 -1.19532847e+00 1.14055671e-01 -1.14965163e-01
-6.24272287e-01 1.15308189e+00 5.72864652e-01 -1.82187542e-01
6.26934469e-01 1.13369904e-01 -1.29746050e-01 2.16293022e-01
-7.53090501e-01 1.53532058e-01 4.11585778e-01 2.02570677e-01
3.60557735e-01 -5.47859892e-02 -4.28175002e-01 2.56073792e-02
4.95223887e-02 1.60100490e-01 1.03355356e-01 8.91654193e-01
-5.65407813e-01 -1.47199821e+00 -1.64658248e-01 4.42158014e-01
-3.40796590e-01 -3.21956366e-01 -1.24533111e-02 7.65415668e-01
-1.19648457e-01 5.00967324e-01 2.32339967e-02 -1.50030971e-01
1.53269231e-01 1.92571953e-01 5.39811850e-01 -8.23307812e-01
-2.53292590e-01 2.50373706e-02 -9.88504216e-02 -1.96072325e-01
-4.34133142e-01 -2.85265476e-01 -8.33047092e-01 -2.32397512e-01
-6.68150187e-01 1.49837196e-01 9.32574749e-01 1.00042570e+00
5.68471909e-01 7.00388402e-02 5.84754884e-01 -1.14366233e+00
-1.03821218e+00 -1.23273492e+00 -2.53308445e-01 2.34324649e-01
-7.24167526e-02 -5.88666439e-01 -3.02247554e-01 9.12948400e-02] | [10.127636909484863, 2.4715464115142822] |
e08e27bd-0f25-4393-aedf-143a31883b6d | aerial-imagery-pile-burn-detection-using-deep | 2012.14036 | null | https://arxiv.org/abs/2012.14036v1 | https://arxiv.org/pdf/2012.14036v1.pdf | Aerial Imagery Pile burn detection using Deep Learning: the FLAME dataset | Wildfires are one of the costliest and deadliest natural disasters in the US, causing damage to millions of hectares of forest resources and threatening the lives of people and animals. Of particular importance are risks to firefighters and operational forces, which highlights the need for leveraging technology to minimize danger to people and property. FLAME (Fire Luminosity Airborne-based Machine learning Evaluation) offers a dataset of aerial images of fires along with methods for fire detection and segmentation which can help firefighters and researchers to develop optimal fire management strategies. This paper provides a fire image dataset collected by drones during a prescribed burning piled detritus in an Arizona pine forest. The dataset includes video recordings and thermal heatmaps captured by infrared cameras. The captured videos and images are annotated and labeled frame-wise to help researchers easily apply their fire detection and modeling algorithms. The paper also highlights solutions to two machine learning problems: (1) Binary classification of video frames based on the presence [and absence] of fire flames. An Artificial Neural Network (ANN) method is developed that achieved a 76% classification accuracy. (2) Fire detection using segmentation methods to precisely determine fire borders. A deep learning method is designed based on the U-Net up-sampling and down-sampling approach to extract a fire mask from the video frames. Our FLAME method approached a precision of 92% and a recall of 84%. Future research will expand the technique for free burning broadcast fire using thermal images. | ['Erik Blasch', 'Peter Z Fulé', 'Liming Zheng', 'Abolfazl Razi', 'Fatemeh Afghah', 'Alireza Shamsoshoara'] | 2020-12-28 | null | null | null | null | ['fire-detection'] | ['time-series'] | [ 7.45491326e-01 -7.62719572e-01 -2.05054343e-01 -3.05385655e-03
-2.14917004e-01 -6.83700740e-01 3.94961327e-01 -2.14488178e-01
-8.03881407e-01 7.33768642e-01 -9.81812272e-03 -3.97070438e-01
-2.07275853e-01 -1.28372645e+00 -1.89689934e-01 -8.69454324e-01
-4.08219844e-01 4.21711169e-02 1.14192560e-01 -1.54785961e-01
1.67801768e-01 9.68265653e-01 -1.51651311e+00 3.83670300e-01
9.54529867e-02 1.06262243e+00 2.94379860e-01 1.06658244e+00
5.17442763e-01 6.31219625e-01 -5.85579395e-01 3.93104851e-01
1.02565992e+00 -2.17635021e-01 -6.77912414e-01 -5.74741475e-02
3.47545624e-01 -9.65110600e-01 -4.02434319e-01 7.06152022e-01
3.51581067e-01 5.91709256e-01 7.84862339e-01 -1.37052906e+00
9.75621566e-02 2.62693495e-01 -7.87109375e-01 7.53394127e-01
-2.58986093e-02 5.17003953e-01 3.86463076e-01 -4.20065761e-01
2.92641819e-01 8.44934225e-01 8.48077118e-01 2.56122082e-01
-1.02614152e+00 -9.99152422e-01 -2.38596842e-01 1.45727798e-01
-1.66074598e+00 -3.27078372e-01 4.40749317e-01 -7.50894785e-01
1.32969975e+00 5.63709140e-01 1.08311868e+00 8.09842408e-01
1.69555098e-01 7.20024109e-02 1.08962095e+00 -3.87350053e-01
3.13524306e-01 -7.47244775e-01 -3.35656404e-01 7.96910286e-01
1.39700592e-01 9.78997231e-01 -1.06641144e-01 -6.85665235e-02
1.11491168e+00 3.80753338e-01 -3.62292975e-01 5.38148880e-01
-1.07988179e+00 8.47180545e-01 8.96568120e-01 6.75420836e-02
-6.58531189e-01 1.91745862e-01 3.54083151e-01 -1.04429781e-01
5.79579473e-01 1.41239613e-01 -2.89541245e-01 2.77250670e-02
-1.52068281e+00 2.53455490e-01 3.67505461e-01 3.51477474e-01
6.83398306e-01 2.80477703e-01 -3.37942247e-03 6.29203439e-01
2.60574251e-01 9.56153214e-01 -1.07186012e-01 -1.13172364e+00
8.48328974e-03 4.03126866e-01 1.43503264e-01 -1.30105782e+00
-4.46856022e-01 -1.94431040e-02 -1.10177922e+00 9.51424539e-01
2.05550686e-01 -4.64753717e-01 -1.15960300e+00 9.83194828e-01
1.06078953e-01 2.27116451e-01 -4.76606250e-01 1.15304673e+00
3.14937115e-01 9.65654314e-01 3.04178268e-01 -2.57527798e-01
1.21272027e+00 -5.67784727e-01 -3.45365852e-01 -3.57781082e-01
-1.44817114e-01 -2.99537033e-01 4.92114514e-01 1.64215222e-01
-4.05433059e-01 -3.73273343e-01 -1.03493476e+00 4.47609693e-01
-5.85415125e-01 4.58071493e-02 3.82555038e-01 6.80458426e-01
-7.18234122e-01 6.48372531e-01 -9.14311528e-01 -6.73036635e-01
8.82983446e-01 3.88365060e-01 -1.03807636e-01 9.36053991e-02
-8.89901102e-01 8.77046227e-01 5.95664144e-01 4.81017351e-01
-1.35937274e+00 -5.41217208e-01 -5.77705681e-01 -1.61198094e-01
2.29610682e-01 -5.98167479e-01 1.00541890e+00 -9.54402685e-01
-7.79549837e-01 7.07164466e-01 1.93782032e-01 -5.65710247e-01
5.16446829e-01 -1.74546868e-01 -3.39587986e-01 3.68935019e-01
2.76958913e-01 9.90548491e-01 9.26742077e-01 -1.17835319e+00
-1.20590329e+00 -4.58879739e-01 2.92191654e-02 1.11763708e-01
-9.35791731e-02 1.93602264e-01 8.05252254e-01 -8.33431840e-01
-2.51285136e-01 -7.28326976e-01 -2.77373314e-01 1.88144237e-01
3.69588472e-02 5.23613036e-01 1.30633807e+00 -9.20400441e-01
1.24371481e+00 -1.67604017e+00 -3.21830928e-01 1.70741677e-01
-9.55500677e-02 4.86160725e-01 7.17702135e-02 2.91609377e-01
3.64825949e-02 4.74864721e-01 -9.54385221e-01 7.69499898e-01
-5.43352723e-01 2.93955386e-01 -2.39129156e-01 3.51783931e-01
1.06451675e-01 2.21163273e-01 -1.00843740e+00 -1.17156118e-01
5.90296626e-01 5.31791031e-01 -1.08547784e-01 6.63211644e-02
-5.55703565e-02 1.44043118e-01 2.24282108e-02 9.36509013e-01
6.42477870e-01 8.24584663e-01 -2.65252948e-01 -2.74315774e-01
-5.91176987e-01 -3.06139618e-01 -1.10028005e+00 9.31577563e-01
-2.05932766e-01 6.88845634e-01 3.43059331e-01 -6.99737966e-01
9.63477671e-01 4.32904422e-01 8.40864480e-01 -1.38475254e-01
2.27644607e-01 -1.27421871e-01 -4.29345667e-01 -8.00133765e-01
4.35879022e-01 -6.22894287e-01 1.68112293e-01 6.68723166e-01
-2.17913881e-01 -1.91162348e-01 2.48557910e-01 -3.97178173e-01
1.26234841e+00 1.50529772e-01 2.31399581e-01 -3.10647994e-01
1.86751969e-02 6.91315532e-01 4.33357120e-01 6.14340365e-01
-6.36103153e-01 4.73739028e-01 -3.92427474e-01 -1.16574943e+00
-9.77488875e-01 -9.33708608e-01 3.81692150e-03 9.88340557e-01
-1.35093912e-01 -1.07225992e-01 -7.40244806e-01 -4.96018261e-01
-1.49468556e-01 6.57728255e-01 -3.06464970e-01 -1.36795685e-01
-4.13312227e-01 -9.64827418e-01 8.65472257e-01 7.61265159e-01
1.17559612e+00 -1.36784029e+00 -1.56717801e+00 1.06699884e-01
-4.90442425e-01 -7.90690720e-01 -1.17028221e-01 3.85841459e-01
-7.74357140e-01 -1.36129045e+00 -5.34509063e-01 -5.92408895e-01
4.97798413e-01 6.88569725e-01 5.57244003e-01 1.27544682e-02
-9.69129801e-01 3.78187269e-01 -3.59093815e-01 -4.97281194e-01
-9.14256200e-02 -6.13190159e-02 1.38781285e-02 -3.97427350e-01
6.02982342e-01 -6.16304100e-01 -7.00079858e-01 2.72184849e-01
-1.03267550e+00 -1.44287404e-02 2.91451275e-01 4.60500836e-01
2.21921772e-01 9.29221809e-01 -7.28698671e-02 -1.71205312e-01
4.85603809e-01 -3.33982944e-01 -5.40896416e-01 7.02529475e-02
-1.78416640e-01 -5.39230645e-01 3.60676318e-01 3.90834734e-03
-9.56939399e-01 5.11333704e-01 3.68485391e-01 -2.68562406e-01
-8.77239525e-01 3.66388112e-01 3.86597425e-01 4.42638323e-02
9.61311996e-01 -2.52280802e-01 -2.67191619e-01 -5.08127036e-03
-7.42163658e-02 1.00794089e+00 9.71744597e-01 -2.66065717e-01
8.86551440e-01 7.88570464e-01 9.05391648e-02 -1.56390774e+00
-5.37986875e-01 -6.04575515e-01 -1.08528316e+00 -1.14644825e+00
1.03686953e+00 -6.27749741e-01 -5.38453758e-01 6.00648522e-01
-9.90054011e-01 -4.79365289e-01 -2.18779966e-01 5.47624469e-01
-3.85457844e-01 2.43726242e-02 -3.36913407e-01 -1.36649740e+00
-5.95552266e-01 -6.14239037e-01 7.63571739e-01 2.91390806e-01
-1.66136399e-01 -4.19756502e-01 1.37284651e-01 5.27872741e-01
3.74398172e-01 7.70184159e-01 5.24518430e-01 3.98830265e-01
-3.09121519e-01 -1.63921826e-02 -2.52855808e-01 5.04086912e-01
4.49901402e-01 5.18282235e-01 -1.05785346e+00 -4.41419594e-02
-4.52467725e-02 1.01048037e-01 1.27529180e+00 8.37318778e-01
6.28364086e-01 -4.07044351e-01 -3.46647769e-01 6.34516835e-01
1.63938820e+00 8.09773207e-01 6.83984160e-01 6.54464364e-01
4.56758767e-01 3.65762144e-01 4.11639333e-01 7.49706864e-01
-3.34443212e-01 1.22538924e-01 8.06891739e-01 -1.55701116e-01
5.80548197e-02 2.03359928e-02 5.11738420e-01 -3.75449240e-01
-1.06962788e+00 -1.01373047e-01 -1.19432271e+00 4.12459522e-01
-1.56716335e+00 -1.85518956e+00 -1.25755176e-01 2.21682191e+00
2.55615890e-01 -2.29949221e-01 4.95994598e-01 5.00275373e-01
1.06843734e+00 3.51697385e-01 -2.01099902e-01 -1.86249822e-01
1.53986290e-01 4.78197277e-01 1.05370784e+00 4.98320878e-01
-1.93913066e+00 8.44039977e-01 6.35068846e+00 3.36324602e-01
-1.10132241e+00 -1.79677248e-01 4.67229843e-01 -2.51744837e-01
4.76385474e-01 3.52598280e-02 -4.47849274e-01 1.29041523e-01
7.02740312e-01 1.21317521e-01 1.03677046e+00 3.34226519e-01
7.96647310e-01 -6.05595052e-01 -1.75910324e-01 6.01089656e-01
-1.83505744e-01 -1.01634324e+00 -1.65994391e-01 -1.15663409e-01
4.74922359e-01 6.00299388e-02 -4.71555650e-01 -1.30768806e-01
4.11657989e-01 -9.69166994e-01 6.14178538e-01 6.04861856e-01
8.24322045e-01 -8.74203682e-01 5.63333094e-01 4.64613497e-01
-1.47308242e+00 -5.61659575e-01 -3.55558932e-01 -8.16082835e-01
3.54039639e-01 4.15445536e-01 -9.16036725e-01 2.83163667e-01
1.09752440e+00 5.22388220e-01 -1.29864872e-01 1.13163304e+00
-2.04625636e-01 7.31192350e-01 -7.62137830e-01 2.78301150e-01
5.10202706e-01 -3.96344960e-01 6.19314313e-01 1.31295300e+00
4.03087288e-01 5.88267803e-01 4.98831511e-01 8.43497872e-01
3.90547335e-01 -5.97137153e-01 -9.72652793e-01 -1.47851437e-01
3.58530551e-01 1.31108165e+00 -1.16896784e+00 3.45150977e-02
9.42463502e-02 6.98642075e-01 -3.92686605e-01 3.36644828e-01
-8.23540151e-01 -2.58789569e-01 6.89286351e-01 4.83629107e-01
-1.00973152e-01 -6.72926605e-01 -2.24040687e-01 -5.57110965e-01
-3.19628984e-01 -2.85723001e-01 5.79240978e-01 -8.17075074e-01
-9.95951116e-01 4.32009667e-01 4.19993520e-01 -1.01275265e+00
6.88109174e-02 -4.80563134e-01 -5.87067187e-01 5.19686162e-01
-1.11016619e+00 -1.24653149e+00 -7.99388289e-01 4.31914777e-01
7.04151630e-01 -9.26589295e-02 1.02045870e+00 9.72119793e-02
-5.79381883e-01 -5.36668122e-01 -1.95651650e-01 4.47624326e-01
1.72727928e-01 -6.75037146e-01 6.60941303e-02 1.08269370e+00
-7.40443030e-03 -1.96114480e-01 1.68289945e-01 -1.02236497e+00
-7.63302088e-01 -1.48020291e+00 5.06508470e-01 2.50062823e-01
1.80843219e-01 4.92327251e-02 -1.60233483e-01 6.19647503e-01
5.96962720e-02 -2.75843680e-01 5.15257955e-01 -4.57673639e-01
2.09257141e-01 -1.54064447e-01 -1.53398871e+00 4.70667690e-01
9.22855198e-01 -3.87223572e-01 -3.22735518e-01 3.69492352e-01
-2.09255606e-01 1.91558331e-01 -4.29528475e-01 5.63051939e-01
1.03252399e+00 -1.00668848e+00 9.59744751e-01 -3.96482527e-01
4.61688280e-01 -5.14248669e-01 -4.43095654e-01 -1.16396654e+00
-7.83427119e-01 -1.57019004e-01 3.70352417e-01 9.90942240e-01
5.00466116e-02 1.20725660e-02 5.67800820e-01 6.69964910e-01
1.98718518e-01 -2.12448575e-02 -8.56985927e-01 -7.91410685e-01
-3.80045444e-01 -6.11020148e-01 9.97118205e-02 7.39426434e-01
-3.46812189e-01 -3.77129950e-02 -1.96042642e-01 5.22572160e-01
5.09162247e-01 -3.59177858e-01 2.93398231e-01 -1.32059753e+00
4.96625632e-01 -4.66290355e-01 -4.34382141e-01 -9.51617062e-02
-2.59941891e-02 -7.11840034e-01 1.28953427e-01 -1.62221813e+00
1.32693186e-01 -2.31800497e-01 5.46669066e-02 1.02627587e+00
2.89802730e-01 5.52527845e-01 1.98893070e-01 1.57453284e-01
5.12649775e-01 2.51146913e-01 7.13366508e-01 -6.14639819e-01
-1.59695983e-01 1.47069991e-01 -4.16828692e-03 8.92839074e-01
1.10321558e+00 -7.32310176e-01 -1.18608810e-01 -5.07710814e-01
-2.67174393e-01 -6.45552948e-02 1.10330474e+00 -1.59920967e+00
1.91757068e-01 -7.05512226e-01 6.63722873e-01 -5.39621413e-01
1.51820108e-01 -1.45897520e+00 5.31034350e-01 8.12408209e-01
-5.25752157e-02 1.00840181e-01 4.32496846e-01 4.03065085e-01
1.74782604e-01 -3.93433809e-01 9.13023055e-01 -5.91018796e-01
-9.63790119e-01 2.58098722e-01 -1.22870195e+00 -4.09080356e-01
1.38742042e+00 -3.80769938e-01 -3.82449239e-01 -2.18399048e-01
-5.51809430e-01 -2.02511668e-01 2.00674474e-01 2.22145721e-01
6.45883918e-01 -1.10385001e+00 -8.03024113e-01 5.81327796e-01
-5.10360181e-01 -8.53408203e-02 1.35230109e-01 3.28629792e-01
-1.24294209e+00 1.16776504e-01 -8.57026160e-01 -4.04537618e-01
-1.29303849e+00 3.74541581e-01 5.46332896e-01 2.79077679e-01
-5.62637627e-01 7.35228539e-01 -4.84842896e-01 -1.69357046e-01
-3.30192149e-02 -3.34764183e-01 2.05135942e-02 2.41820022e-01
7.72439003e-01 1.01383150e+00 2.32120790e-02 -8.32419336e-01
-4.72735226e-01 3.46011966e-01 5.31531155e-01 -4.81430620e-01
1.54479742e+00 2.67842710e-01 -4.43413742e-02 2.66974062e-01
6.25708342e-01 -8.69742751e-01 -1.41557741e+00 5.64054072e-01
-2.80926943e-01 -6.79700911e-01 8.61414731e-01 -1.13321388e+00
-1.38474607e+00 7.64747798e-01 1.02329576e+00 2.26720259e-01
1.50829029e+00 -7.63551474e-01 7.92649329e-01 3.95276845e-01
3.94758910e-01 -1.22385383e+00 -6.01672828e-01 3.31827015e-01
8.64120007e-01 -1.00982654e+00 1.52018890e-01 -1.26398483e-03
-2.83108324e-01 1.50913870e+00 3.82168472e-01 -1.70412004e-01
7.96623468e-01 8.59923720e-01 1.52082399e-01 -1.40051678e-01
-3.03806305e-01 -4.74897712e-01 -3.93774271e-01 1.06196284e+00
1.45432547e-01 5.01013458e-01 5.16684679e-03 1.10303000e-01
7.75339305e-02 3.96806628e-01 1.42538577e-01 1.41111028e+00
-1.00932503e+00 -5.28362870e-01 -9.64558542e-01 7.92280316e-01
-4.09189641e-04 -1.29087031e-01 -7.47653782e-01 6.64943576e-01
4.49382842e-01 1.33615005e+00 7.06884041e-02 -6.63110971e-01
2.06748068e-01 4.91758250e-02 3.59882444e-01 -2.05797926e-01
-9.16273236e-01 -2.54282989e-02 -6.90851435e-02 -9.31841657e-02
-8.61283123e-01 -5.86280286e-01 -9.90503967e-01 -6.03730381e-01
-2.60238320e-01 -3.27138036e-01 7.17525959e-01 7.88701236e-01
-1.84824586e-01 3.51995617e-01 6.28961682e-01 -1.41284621e+00
-3.98150422e-02 -9.58133101e-01 -9.54993367e-01 -1.12586051e-01
1.36297435e-01 -7.02054679e-01 -4.83039439e-01 5.17573357e-01] | [9.181795120239258, -1.2553658485412598] |
c8e687cb-82a0-45fc-800a-7ad300a82a92 | nc-dre-leveraging-non-entity-clue-information | 2204.00255 | null | https://arxiv.org/abs/2204.00255v1 | https://arxiv.org/pdf/2204.00255v1.pdf | NC-DRE: Leveraging Non-entity Clue Information for Document-level Relation Extraction | Document-level relation extraction (RE), which requires reasoning on multiple entities in different sentences to identify complex inter-sentence relations, is more challenging than sentence-level RE. To extract the complex inter-sentence relations, previous studies usually employ graph neural networks (GNN) to perform inference upon heterogeneous document-graphs. Despite their great successes, these graph-based methods, which normally only consider the words within the mentions in the process of building graphs and reasoning, tend to ignore the non-entity clue words that are not in the mentions but provide important clue information for relation reasoning. To alleviate this problem, we treat graph-based document-level RE models as an encoder-decoder framework, which typically uses a pre-trained language model as the encoder and a GNN model as the decoder, and propose a novel graph-based model NC-DRE that introduces decoder-to-encoder attention mechanism to leverage Non-entity Clue information for Document-level Relation Extraction. | ['Yidong Cheng', 'Liang Zhang'] | 2022-04-01 | null | null | null | null | ['document-level-relation-extraction'] | ['natural-language-processing'] | [ 1.17680803e-01 6.60845995e-01 -3.98392558e-01 -3.98979366e-01
-6.58769786e-01 -4.94615346e-01 5.34541070e-01 6.51456654e-01
-3.09355527e-01 7.14154065e-01 5.43080747e-01 -6.93821847e-01
3.66675295e-02 -1.29809904e+00 -7.50851452e-01 1.55109644e-03
6.94551840e-02 5.97225904e-01 2.20448464e-01 -3.72742504e-01
-3.03945076e-02 1.46042764e-01 -6.02314830e-01 3.30127716e-01
1.07367587e+00 5.86172283e-01 -1.20682992e-01 7.49617219e-01
-5.73275208e-01 1.52493858e+00 -8.26892674e-01 -1.05826640e+00
-2.21026406e-01 -6.53981328e-01 -1.04142761e+00 -1.81894436e-01
1.11588255e-01 -2.33476624e-01 -8.47713530e-01 1.29044366e+00
3.61532658e-01 -3.17750797e-02 7.71379173e-01 -9.03100252e-01
-1.18078268e+00 1.35383260e+00 -6.09027505e-01 4.16066647e-01
4.48401123e-01 -4.25722718e-01 1.62197757e+00 -7.56034195e-01
7.78027654e-01 1.30701828e+00 5.06228507e-01 2.53187776e-01
-8.57969463e-01 -5.38317263e-01 3.89051318e-01 4.08314705e-01
-1.54352617e+00 -3.40899944e-01 9.26445603e-01 -1.24954939e-01
1.78032219e+00 1.80615798e-01 4.20935869e-01 1.01743853e+00
3.83332610e-01 7.98232973e-01 2.78685391e-01 -3.80763888e-01
-2.35629737e-01 -2.90562958e-01 4.29266512e-01 6.53261840e-01
5.78244925e-01 -5.27869642e-01 -3.06994140e-01 2.55355388e-01
5.23444951e-01 -1.74576789e-01 -4.05915767e-01 2.29344130e-01
-9.78921890e-01 8.94297481e-01 8.49754930e-01 4.64079112e-01
-4.82463956e-01 6.87230453e-02 6.82561576e-01 2.51168847e-01
6.91509962e-01 5.32120228e-01 -5.21521211e-01 2.56027907e-01
-6.07149661e-01 -6.42824769e-02 1.17900527e+00 1.37473714e+00
5.40299416e-01 -1.78763852e-01 -5.80455542e-01 6.50783122e-01
3.14045370e-01 1.49803892e-01 2.17848793e-01 1.59032792e-02
1.30779397e+00 1.16436541e+00 -5.07441282e-01 -1.34086561e+00
-2.84275442e-01 -8.37543905e-01 -1.23184395e+00 -7.53982902e-01
-3.31704706e-01 -4.04554009e-01 -8.58061969e-01 1.44794810e+00
1.95413440e-01 -1.01298377e-01 4.45099920e-01 5.56142628e-01
1.65717041e+00 6.06979072e-01 1.42854810e-01 -1.85440436e-01
1.31855202e+00 -1.34168959e+00 -1.06466722e+00 -5.03973603e-01
9.89661813e-01 -3.56757581e-01 5.90547681e-01 -3.46548289e-01
-9.21881199e-01 -3.35868418e-01 -1.20440590e+00 -5.89396238e-01
-6.43951237e-01 4.78277393e-02 6.26098812e-01 -7.82106966e-02
-8.07133734e-01 4.63547170e-01 -5.28802633e-01 -1.61383525e-01
4.53869134e-01 2.05463842e-01 -4.16060686e-01 -3.12564708e-02
-1.85415471e+00 1.19351995e+00 8.27633262e-01 6.11143053e-01
-3.34205866e-01 -3.35643470e-01 -1.41912520e+00 5.34595668e-01
9.57960010e-01 -9.28253770e-01 8.89924586e-01 -3.69619817e-01
-1.11483014e+00 6.32768750e-01 -3.53023261e-01 -3.60166013e-01
2.43468173e-02 -1.54224396e-01 -7.06614137e-01 6.11033738e-02
2.08892480e-01 5.26732132e-02 2.66627312e-01 -9.49536324e-01
-3.11025232e-01 -1.42938122e-01 2.59609252e-01 3.75682652e-01
-5.42814359e-02 2.38555863e-01 -7.96018422e-01 -6.52816415e-01
1.92350090e-01 -4.61831421e-01 -1.39345288e-01 -9.66386616e-01
-1.30723739e+00 -6.68717504e-01 4.39922065e-01 -9.56356227e-01
1.77261817e+00 -1.64388371e+00 2.45690733e-01 1.00097418e-01
8.37717950e-01 2.71570086e-01 -2.94294178e-01 8.12313557e-01
-2.55000323e-01 3.89919400e-01 -9.84349996e-02 -2.09356636e-01
7.59792402e-02 1.09797880e-01 -1.48848876e-01 -1.19017437e-01
6.57702327e-01 1.65105116e+00 -1.11081803e+00 -7.15051651e-01
-2.43852824e-01 3.47281307e-01 -1.62975654e-01 3.32035750e-01
-3.24374884e-01 1.03501320e-01 -5.86451173e-01 4.04516846e-01
4.72401768e-01 -6.65555835e-01 4.97435629e-01 -5.86202979e-01
5.29119611e-01 1.00875592e+00 -7.22918808e-01 1.25155604e+00
-4.70397413e-01 5.88865817e-01 -2.48108268e-01 -9.90078151e-01
8.33763897e-01 3.11359823e-01 1.48384392e-01 -6.97995126e-01
6.96671754e-02 -2.27959454e-02 2.83103287e-01 -7.01877773e-01
5.42568624e-01 -5.51881175e-03 -1.34249717e-01 2.16439918e-01
2.49642178e-01 -5.44053577e-02 5.00327528e-01 7.91075468e-01
1.58449924e+00 -1.17371775e-01 6.07738137e-01 2.12685764e-01
6.65915132e-01 -2.27286085e-01 5.87413251e-01 3.92624348e-01
3.79914343e-01 4.13797200e-01 1.10034895e+00 -1.82633605e-02
-7.76215255e-01 -7.27109730e-01 3.97094280e-01 5.51025212e-01
1.24073878e-01 -1.07238388e+00 -4.32657719e-01 -1.17822051e+00
-2.05510765e-01 7.35149324e-01 -5.56452692e-01 -3.71888578e-01
-7.70906210e-01 -5.44668734e-01 5.05595684e-01 6.02542162e-01
5.61113238e-01 -1.08828092e+00 4.14263546e-01 4.58333850e-01
-5.03503919e-01 -1.70921969e+00 -5.31792283e-01 3.17302108e-01
-4.44966733e-01 -1.12537730e+00 -4.31940198e-01 -8.13763559e-01
7.95173883e-01 -1.12231791e-01 1.63208365e+00 2.72420406e-01
2.65942484e-01 -2.60088772e-01 -6.98501408e-01 -1.70712277e-01
-3.76002103e-01 7.37060487e-01 -5.92075169e-01 -3.17261487e-01
7.87792087e-01 -5.62433243e-01 -2.12075144e-01 -2.17058629e-01
-6.75200522e-01 2.73499340e-01 7.82763183e-01 6.90555990e-01
4.59353387e-01 3.67123127e-01 6.64480805e-01 -1.50374281e+00
1.20707011e+00 -5.91242254e-01 -1.86647654e-01 7.70381570e-01
-5.40706456e-01 3.29012543e-01 7.69334674e-01 -1.40787646e-01
-7.35462904e-01 -5.84027290e-01 -3.21199238e-01 -7.72295296e-02
2.29704842e-01 1.29727352e+00 -5.26568115e-01 4.26504731e-01
3.20928484e-01 1.36865735e-01 -6.31743193e-01 -1.92407891e-01
4.60744262e-01 6.17771745e-01 4.47456688e-01 -4.12163526e-01
7.58771241e-01 -3.55182379e-01 1.02933034e-01 -3.47093493e-01
-1.33498371e+00 -3.52383643e-01 -7.97772765e-01 1.22150570e-01
1.14880264e+00 -1.03728878e+00 -6.38929904e-01 7.38725290e-02
-1.66576803e+00 -1.97910726e-01 -5.63100390e-02 2.95464426e-01
1.39349818e-01 3.53311896e-01 -1.07057023e+00 -5.78734219e-01
-7.21369803e-01 -9.65544283e-01 1.07179189e+00 2.34631941e-01
-3.36691976e-01 -1.21399438e+00 -1.62214935e-01 2.89952129e-01
5.93784153e-02 2.69823849e-01 1.36031616e+00 -9.49158013e-01
-6.83057606e-01 -3.92395556e-01 -7.90295422e-01 1.38546616e-01
3.91050100e-01 -1.52296931e-01 -5.23187816e-01 1.01609364e-01
-4.01459873e-01 -1.43742383e-01 8.39450061e-01 -7.97019228e-02
8.83885443e-01 -6.27825439e-01 -5.19424081e-01 4.19302523e-01
1.27375960e+00 -1.58492345e-02 6.60338402e-01 -1.25216730e-02
1.38180435e+00 4.88025099e-01 1.53499663e-01 -1.69652253e-01
1.16377795e+00 4.63225424e-01 2.20130846e-01 -1.70756504e-01
-4.05148447e-01 -6.52916133e-01 3.40093166e-01 1.59545016e+00
-3.52558345e-02 -8.23245525e-01 -8.61535430e-01 4.54354227e-01
-1.85266268e+00 -8.07626486e-01 -4.19801772e-01 1.53234577e+00
1.17946815e+00 4.65939373e-01 -3.05291086e-01 -2.15963367e-02
7.30455399e-01 4.26997960e-01 -3.54138225e-01 -4.41094875e-01
-1.19530104e-01 1.54910937e-01 3.93431753e-01 7.06288576e-01
-9.56902385e-01 1.37666321e+00 4.96998024e+00 8.23681295e-01
-7.36610353e-01 -4.31726761e-02 2.59663641e-01 3.77119154e-01
-7.22041965e-01 2.37069786e-01 -9.19010103e-01 2.64704049e-01
1.06067157e+00 -1.76650852e-01 2.00601637e-01 5.14266670e-01
-2.57835567e-01 8.31935182e-02 -1.34292185e+00 7.30533123e-01
2.47567073e-01 -1.31181335e+00 3.03228259e-01 -5.79894669e-02
3.92908990e-01 -2.55861282e-02 -6.17019892e-01 6.69248879e-01
6.66053116e-01 -1.13421929e+00 2.50616550e-01 4.60817039e-01
8.01546752e-01 -7.17606962e-01 1.09862864e+00 1.62190825e-01
-1.61566556e+00 4.30186003e-01 -3.98651034e-01 3.34505960e-02
3.68226647e-01 9.64728951e-01 -8.64603043e-01 1.27706695e+00
1.21876158e-01 1.07576871e+00 -7.65653253e-01 5.36913514e-01
-9.85103488e-01 4.87012774e-01 3.26559432e-02 -3.84235859e-01
1.71646386e-01 -1.51937932e-01 3.66843730e-01 1.47851264e+00
1.67238638e-02 1.52079567e-01 1.35186911e-01 8.45348597e-01
-6.52465880e-01 8.98679867e-02 -6.73764348e-01 -4.65811461e-01
3.60592425e-01 1.31934428e+00 -7.68573880e-01 -4.62447286e-01
-6.76241815e-01 1.15679991e+00 1.04847026e+00 5.42814136e-01
-6.57420516e-01 -1.03103530e+00 1.86353177e-01 -2.59960771e-01
2.78496623e-01 -5.28971493e-01 -2.22479284e-01 -1.46479630e+00
3.12450588e-01 -7.35807538e-01 5.06359041e-01 -7.85971820e-01
-1.44332492e+00 1.00199628e+00 -1.06646620e-01 -9.15831983e-01
-3.31925541e-01 -3.16634417e-01 -5.16750991e-01 1.07016706e+00
-1.56713402e+00 -1.27699864e+00 1.08124658e-01 2.89474130e-01
3.18953604e-01 9.62087959e-02 7.01559544e-01 5.23891389e-01
-8.96639526e-01 7.09507108e-01 -5.60291886e-01 9.90513206e-01
2.57362455e-01 -1.33035004e+00 7.89663017e-01 1.05736864e+00
4.46293533e-01 8.37184429e-01 2.93721467e-01 -1.10511565e+00
-1.26835704e+00 -1.26152837e+00 1.80164325e+00 -4.89245117e-01
9.25172508e-01 -6.98078573e-01 -9.75434721e-01 1.04504275e+00
5.23329735e-01 -2.38164365e-02 5.67537904e-01 5.75839579e-01
-4.28756893e-01 1.10075518e-01 -4.86125916e-01 9.08239841e-01
1.32476020e+00 -1.00234067e+00 -7.38243043e-01 3.85734856e-01
1.18372095e+00 -6.90075397e-01 -1.13077176e+00 6.06244385e-01
-1.49948880e-01 -2.86500394e-01 8.64402950e-01 -7.48192787e-01
9.12712932e-01 -1.70385659e-01 1.55798018e-01 -1.42684698e+00
-3.24059635e-01 -4.52199280e-01 -7.66297579e-01 1.64118207e+00
1.05709898e+00 -4.75256979e-01 5.15283346e-01 5.13993502e-01
-1.32128537e-01 -8.89854372e-01 -4.98926044e-01 -5.58688700e-01
-7.85901695e-02 -3.15994948e-01 6.67328477e-01 9.87644792e-01
3.51475298e-01 1.61047101e+00 -1.89386651e-01 2.83176392e-01
1.90764964e-01 2.44250730e-01 5.14831662e-01 -1.10033095e+00
-1.56354427e-01 -4.06098247e-01 -2.56797314e-01 -1.06093442e+00
7.43183911e-01 -1.33183396e+00 1.19037740e-01 -2.39708376e+00
3.04861754e-01 -1.03697203e-01 -1.06139354e-01 4.69322354e-01
-6.66129827e-01 -3.04274619e-01 -1.44600645e-01 -3.68420221e-02
-7.72918224e-01 7.72569716e-01 1.62409627e+00 -4.39874560e-01
1.33845970e-01 -1.81563154e-01 -9.86232817e-01 4.48698282e-01
3.70232314e-01 -5.97549379e-01 -6.24718070e-01 -6.22515142e-01
9.06912088e-01 2.22408742e-01 2.04129834e-02 -5.82924008e-01
6.21262968e-01 -8.04026574e-02 2.33811215e-01 -6.01159275e-01
-4.17297184e-02 -7.02929974e-01 -3.18991452e-01 2.86713410e-02
-6.00529671e-01 2.40225613e-01 -2.18092769e-01 5.04356325e-01
-6.71287179e-01 -1.70535132e-01 -5.45261316e-02 -1.27841324e-01
-3.90374273e-01 6.47185743e-01 -8.61058012e-02 4.45666701e-01
5.80792964e-01 1.77746609e-01 -7.01338947e-01 -6.25245631e-01
-6.88406587e-01 4.04856980e-01 -2.18731701e-01 4.55022216e-01
6.14053726e-01 -1.32881236e+00 -9.98391509e-01 -1.49392396e-01
1.97967052e-01 5.95083773e-01 1.19290061e-01 6.52121961e-01
-3.43861669e-01 4.14657176e-01 5.85936368e-01 1.11227624e-01
-1.01453054e+00 8.15605700e-01 2.56102204e-01 -9.96996224e-01
-8.07985067e-01 1.01655650e+00 -3.35629247e-02 -4.10818219e-01
-4.96171080e-02 -5.21872878e-01 -6.24912143e-01 3.01674940e-02
2.00888976e-01 -2.81091779e-01 2.31969923e-01 -6.39605403e-01
-4.05045658e-01 3.15419704e-01 -2.99422145e-01 2.09638894e-01
1.26137447e+00 -1.92317396e-01 -6.69883072e-01 2.81235874e-01
1.34494197e+00 1.39825836e-01 -6.28826439e-01 -5.56317151e-01
5.02762318e-01 1.71873331e-01 3.63727026e-02 -6.46688402e-01
-1.14666104e+00 9.26631212e-01 -7.22830117e-01 4.37918723e-01
8.21452975e-01 2.99619019e-01 1.13375163e+00 5.87171674e-01
8.72319639e-02 -9.28871810e-01 -1.42683849e-01 1.09427440e+00
1.05256331e+00 -1.24395680e+00 1.58285052e-01 -1.04388487e+00
-6.60769820e-01 1.20536983e+00 8.54830325e-01 6.66958913e-02
6.74607992e-01 4.44481879e-01 -1.95849195e-01 -4.89365280e-01
-7.91313410e-01 -5.48570514e-01 6.59297466e-01 4.78759021e-01
8.10151339e-01 -4.24595171e-04 -4.06561255e-01 9.33501065e-01
-2.71190941e-01 -1.40401796e-01 2.78180063e-01 6.24881983e-01
1.30235121e-01 -1.12564504e+00 5.72982550e-01 6.34194911e-01
-4.59943563e-01 -7.61610508e-01 -6.99271083e-01 5.41824579e-01
-1.90990523e-01 1.17977834e+00 -7.20010027e-02 -7.04856515e-01
3.95026922e-01 -3.42925414e-02 3.73387277e-01 -1.07522249e+00
-7.89827704e-01 -3.49377692e-01 8.43585789e-01 -2.37839401e-01
-3.71700615e-01 -1.02030687e-01 -1.49604976e+00 -2.29454845e-01
-5.57052672e-01 4.03337687e-01 1.21952914e-01 1.34291124e+00
4.46799010e-01 1.30636108e+00 3.38650018e-01 -4.20208648e-02
-4.31779064e-02 -1.21868384e+00 -4.77764994e-01 2.97510892e-01
2.75575757e-01 -3.16347092e-01 -1.87388062e-02 -3.13243389e-01] | [9.20391845703125, 8.587726593017578] |
7dc18a35-29a0-41c2-87d1-d69309d9a706 | detector-in-detector-multi-level-analysis-for | 1902.07017 | null | http://arxiv.org/abs/1902.07017v1 | http://arxiv.org/pdf/1902.07017v1.pdf | Detector-in-Detector: Multi-Level Analysis for Human-Parts | Vision-based person, hand or face detection approaches have achieved
incredible success in recent years with the development of deep convolutional
neural network (CNN). In this paper, we take the inherent correlation between
the body and body parts into account and propose a new framework to boost up
the detection performance of the multi-level objects. In particular, we adopt a
region-based object detection structure with two carefully designed detectors
to separately pay attention to the human body and body parts in a
coarse-to-fine manner, which we call Detector-in-Detector network (DID-Net).
The first detector is designed to detect human body, hand, and face. The second
detector, based on the body detection results of the first detector, mainly
focus on the detection of small hand and face inside each body. The framework
is trained in an end-to-end way by optimizing a multi-task loss. Due to the
lack of human body, face and hand detection dataset, we have collected and
labeled a new large dataset named Human-Parts with 14,962 images and 106,879
annotations. Experiments show that our method can achieve excellent performance
on Human-Parts. | ['Qing Song', 'Xiaojie Li', 'Lu Yang', 'Fuqiang Zhou'] | 2019-02-19 | null | null | null | null | ['body-detection', 'hand-detection'] | ['computer-vision', 'computer-vision'] | [-1.90324351e-01 5.51032349e-02 3.67316268e-02 -1.45525694e-01
-2.82151669e-01 -8.39532465e-02 3.80618483e-01 -3.89632910e-01
-3.86579752e-01 1.67817861e-01 2.00662166e-01 4.91796941e-01
4.65626121e-01 -7.21156061e-01 -5.11054516e-01 -6.66739225e-01
2.11866781e-01 4.47280943e-01 6.22769654e-01 -7.63326045e-03
-2.90789813e-01 3.42258066e-01 -1.37714696e+00 5.57994135e-02
4.80795503e-01 1.27664852e+00 -9.75554213e-02 2.35881418e-01
2.98299253e-01 5.79909742e-01 -3.63035560e-01 -6.77796721e-01
4.70355541e-01 -5.10506749e-01 -3.41772467e-01 5.02228200e-01
5.42181969e-01 -6.60198867e-01 -3.82683992e-01 1.01727223e+00
9.06761765e-01 -2.07241386e-01 5.83746016e-01 -1.21499217e+00
-2.72949070e-01 3.77640873e-01 -1.32494581e+00 8.65183249e-02
2.05039889e-01 3.96496505e-01 7.29806185e-01 -1.23447859e+00
2.79951006e-01 1.73104477e+00 8.10482502e-01 7.34401047e-01
-7.85113037e-01 -8.70025814e-01 1.94207773e-01 4.99231368e-02
-1.80725503e+00 -4.92855310e-01 6.50137126e-01 -6.36818647e-01
4.72153872e-01 -3.10593814e-01 7.68860221e-01 8.09599817e-01
-1.29896834e-01 9.72193480e-01 7.53302097e-01 -4.31208700e-01
-3.03272516e-01 1.06950793e-02 6.80170283e-02 1.29173911e+00
3.82603824e-01 1.36634916e-01 -1.50602385e-01 5.51607683e-02
9.43118334e-01 2.58305788e-01 -6.41298741e-02 -4.46809590e-01
-9.21082675e-01 6.89726353e-01 7.04943836e-01 9.84679759e-02
-4.69357401e-01 1.86728872e-02 4.89632189e-01 -3.66479993e-01
2.99246341e-01 -5.44912040e-01 -2.45195910e-01 6.17674768e-01
-9.47707057e-01 1.96043894e-01 4.42238361e-01 8.22305322e-01
4.14531857e-01 -2.17576638e-01 -8.37170064e-01 9.10043597e-01
6.11613154e-01 6.30192876e-01 2.14336008e-01 -3.56083661e-01
5.38667679e-01 8.63166332e-01 -3.14015932e-02 -8.78171980e-01
-6.63646936e-01 -4.80281770e-01 -1.04911864e+00 1.34708241e-01
5.36948204e-01 -3.66928279e-01 -1.14333463e+00 1.50296426e+00
9.59526181e-01 -1.92718834e-01 -3.58606517e-01 1.21592021e+00
1.12394452e+00 3.45809221e-01 2.54340887e-01 -1.30874394e-02
1.90692925e+00 -1.25444186e+00 -4.64606404e-01 -2.79268801e-01
1.45264968e-01 -5.44477761e-01 6.60072803e-01 1.06450632e-01
-1.05006170e+00 -8.55296969e-01 -7.58362830e-01 -6.39655218e-02
-6.66277409e-02 7.70677030e-01 3.25093627e-01 8.55826616e-01
-7.06134200e-01 1.49618417e-01 -7.08416104e-01 -5.08216143e-01
7.81596363e-01 3.15387458e-01 -2.11613730e-01 -8.67533535e-02
-8.13856304e-01 7.55205929e-01 4.60101336e-01 3.26887041e-01
-1.04041207e+00 -1.30503818e-01 -8.55649590e-01 6.94823712e-02
5.98241091e-01 -9.52605307e-01 1.07313442e+00 -8.69603276e-01
-1.37971306e+00 1.13545859e+00 1.69254355e-02 -1.56242758e-01
8.06574047e-01 -2.01623261e-01 -3.26590240e-01 2.06468880e-01
1.94520265e-01 7.56260812e-01 1.10146463e+00 -9.07791257e-01
-8.90138924e-01 -7.21758485e-01 -3.78384233e-01 3.97818722e-02
-3.47976267e-01 6.38075411e-01 -1.09681118e+00 -5.97334445e-01
2.79055852e-02 -8.81391943e-01 1.27835408e-01 3.78065795e-01
-6.48486078e-01 -6.54231668e-01 7.22951949e-01 -8.24387610e-01
1.11706877e+00 -2.00098348e+00 9.97482613e-02 -2.52834614e-02
4.39969420e-01 5.27877271e-01 -4.90885116e-02 -3.32308486e-02
1.56372577e-01 -3.88079196e-01 -1.33516952e-01 -6.03106499e-01
-8.67589116e-02 -1.79737031e-01 3.67187560e-01 7.13289022e-01
3.16464365e-01 8.46592665e-01 -5.68201125e-01 -1.10839176e+00
2.06133708e-01 6.19231462e-01 -4.23111409e-01 2.13082388e-01
1.43230379e-01 3.01663488e-01 -5.46103358e-01 1.16278923e+00
8.79699469e-01 -2.24048480e-01 -1.32688954e-02 -2.95076311e-01
6.61006570e-02 -2.38343343e-01 -1.25799775e+00 1.28823030e+00
-9.17245075e-02 1.13831513e-01 5.50279617e-01 -4.26598758e-01
6.77141964e-01 4.22991365e-01 4.36781943e-01 -5.38612008e-01
4.17832583e-01 -1.93938725e-02 2.59369314e-01 -3.90546978e-01
-2.05950797e-01 -2.60052085e-01 1.40399009e-01 9.18137580e-02
1.58162192e-01 5.57696223e-01 2.85083354e-01 -1.45638809e-02
7.53559589e-01 8.97736475e-02 3.85549396e-01 -2.72468030e-02
8.48626971e-01 -4.02141243e-01 8.22658360e-01 3.65056455e-01
-5.61527491e-01 7.11994588e-01 2.54262954e-01 -4.29544389e-01
-7.33199954e-01 -7.53845334e-01 -1.35924906e-01 1.35590410e+00
4.14341241e-02 -3.40656221e-01 -9.10858691e-01 -9.45956409e-01
2.81252205e-01 -2.45261937e-01 -8.16893518e-01 -4.91273627e-02
-7.26810753e-01 -6.76064610e-01 5.32605469e-01 8.70466232e-01
9.97883379e-01 -1.38446748e+00 -6.30866945e-01 1.67364135e-01
-6.01905771e-02 -1.11561143e+00 -8.55039477e-01 -1.76317915e-01
-5.02546906e-01 -1.25338209e+00 -1.35147631e+00 -8.98339808e-01
5.06716192e-01 1.79038778e-01 9.78674471e-01 1.35007218e-01
-6.85876429e-01 2.45468155e-01 -1.12096496e-01 -5.33957541e-01
1.54535631e-02 -1.63884953e-01 9.75022465e-02 4.29198861e-01
2.18289614e-01 -1.63591579e-01 -8.66507709e-01 5.38244665e-01
-2.83875525e-01 -2.58274645e-01 9.83216643e-01 6.68313980e-01
2.99013168e-01 -1.05143143e-02 1.67084545e-01 -4.04471755e-01
1.43771231e-01 -1.32800117e-01 -6.07713163e-01 4.32328492e-01
-3.03323448e-01 -3.50262940e-01 2.40463823e-01 -4.18190777e-01
-9.87969637e-01 5.64971030e-01 -2.65237480e-01 -6.17990196e-01
-1.63236663e-01 -7.54266158e-02 -6.28178596e-01 -2.00769663e-01
3.78674954e-01 2.75237978e-01 5.75088635e-02 -7.29881883e-01
2.43712202e-01 6.47171497e-01 6.89733505e-01 -4.27509725e-01
8.39933515e-01 4.54545856e-01 2.04276796e-02 -7.52164423e-01
-1.01544011e+00 -8.07867885e-01 -8.88161004e-01 -4.35579181e-01
1.01752484e+00 -1.18636596e+00 -8.84174287e-01 8.19746494e-01
-1.22815049e+00 1.50978249e-02 1.54426079e-02 2.15312406e-01
-5.25208786e-02 5.03609538e-01 -6.96095049e-01 -1.11109042e+00
-9.24229622e-01 -8.61155152e-01 1.50069678e+00 4.17862356e-01
2.03282893e-01 -3.85442704e-01 -1.47791132e-01 2.83470511e-01
7.43660182e-02 3.31261426e-01 2.73715109e-01 -1.96527243e-01
-3.51801634e-01 -5.01054466e-01 -6.35240734e-01 4.18991357e-01
-7.02119991e-02 -6.55070171e-02 -9.57518220e-01 -4.12763417e-01
-2.88396388e-01 -4.06561702e-01 1.14031875e+00 5.10268271e-01
8.80246580e-01 -2.00723529e-01 -6.52199984e-01 5.11355758e-01
1.22319865e+00 -1.71820238e-01 4.04009730e-01 -1.02763675e-01
1.04483187e+00 6.75806999e-01 4.87182438e-01 6.89657450e-01
3.64979923e-01 8.61398757e-01 3.78542602e-01 -3.00991476e-01
-5.66405714e-01 -2.85158664e-01 4.00754392e-01 1.33377269e-01
-3.46661299e-01 3.88819352e-02 -7.86430061e-01 4.66761678e-01
-1.69385612e+00 -8.92548501e-01 -1.54233292e-01 2.08113432e+00
6.25854969e-01 1.53250039e-01 9.35376644e-01 -4.87061478e-02
1.10144019e+00 -3.05827036e-02 -4.91532892e-01 4.82934624e-01
8.03189948e-02 3.67636629e-03 3.55923891e-01 -1.23347722e-01
-1.54437971e+00 8.99755001e-01 5.63242865e+00 7.45672524e-01
-8.46611798e-01 2.57714212e-01 3.06178182e-01 -2.37781629e-02
8.35558355e-01 -4.67151910e-01 -1.51552558e+00 6.48237228e-01
1.42363608e-01 3.09785217e-01 -3.76548283e-02 9.89039421e-01
5.83272986e-02 1.89072236e-01 -9.80256796e-01 1.30905437e+00
1.06753275e-01 -6.39530241e-01 -1.56122237e-01 1.22530283e-02
4.56381947e-01 -2.70195752e-01 -1.25590652e-01 4.67247695e-01
-6.05339594e-02 -8.35526824e-01 9.21850741e-01 3.51764709e-01
7.13734865e-01 -6.83226407e-01 7.36644685e-01 5.22597909e-01
-1.70386553e+00 -3.39679658e-01 -4.82968509e-01 6.93817297e-03
2.47615669e-02 6.11900330e-01 -6.08742714e-01 2.25922406e-01
8.25650632e-01 5.84959924e-01 -5.98152339e-01 1.35264468e+00
-3.63592952e-01 3.58285785e-01 -3.84008318e-01 1.07085124e-01
-1.05227001e-01 1.62559673e-01 2.95815647e-01 1.22325790e+00
-6.08220659e-02 3.19785625e-01 7.18805969e-01 9.18943524e-01
-2.59867191e-01 1.42194645e-03 -1.44814014e-01 3.45364153e-01
2.29854614e-01 1.69938862e+00 -8.12983394e-01 -4.54245716e-01
-5.60913801e-01 1.01959813e+00 2.24613443e-01 -6.23701066e-02
-9.07582402e-01 -1.77290693e-01 3.64534050e-01 3.91858846e-01
5.69809914e-01 1.77098453e-01 1.40076935e-01 -1.08022952e+00
3.15077424e-01 -7.04346061e-01 6.91655219e-01 -3.73688638e-01
-1.45476806e+00 3.59109700e-01 -2.24338919e-01 -9.99379754e-01
1.14342310e-01 -6.62227094e-01 -7.24701464e-01 9.94095147e-01
-1.20728028e+00 -1.59849608e+00 -7.84578800e-01 8.22716713e-01
3.63709599e-01 -3.68145593e-02 3.75778019e-01 5.77728570e-01
-1.07517827e+00 8.60741258e-01 -4.23903406e-01 9.05584157e-01
7.32920051e-01 -9.34589028e-01 2.80851901e-01 8.37435126e-01
-2.26425022e-01 4.43221509e-01 2.48650044e-01 -6.89046681e-01
-1.00448334e+00 -1.24575496e+00 7.20675766e-01 -4.53443378e-01
-4.93716225e-02 -4.10373449e-01 -5.62114418e-01 4.84405845e-01
-3.33123729e-02 4.66463655e-01 2.21692652e-01 1.40561551e-01
-4.46130902e-01 -2.20503539e-01 -1.20348442e+00 2.03996226e-01
1.11298192e+00 -1.67367607e-01 -5.46563327e-01 4.65331018e-01
3.04616779e-01 -4.05258745e-01 -5.02054274e-01 6.60571337e-01
7.64551640e-01 -9.16407645e-01 1.26074886e+00 -3.19274724e-01
3.20125937e-01 -4.40622091e-01 2.71394283e-01 -5.89885950e-01
-6.47794724e-01 -1.05540447e-01 -4.39573497e-01 1.38587785e+00
-1.13854937e-01 -2.67636716e-01 8.88378739e-01 3.62481505e-01
1.23599097e-01 -7.14693606e-01 -6.77529752e-01 -9.56310570e-01
-1.05730668e-01 2.16449678e-01 3.19741666e-01 4.36139673e-01
-3.88889700e-01 5.80416381e-01 -6.17436051e-01 1.72393993e-01
9.44826007e-01 3.04474592e-01 9.57930803e-01 -1.40483212e+00
-3.57164472e-01 -5.21267414e-01 -4.47746456e-01 -1.12708008e+00
-2.70017326e-01 -5.07753015e-01 2.55188704e-01 -1.40049469e+00
9.10495937e-01 3.08346935e-02 -2.94321209e-01 7.91620612e-01
-5.51042199e-01 5.22926033e-01 3.13129663e-01 2.89396167e-01
-8.79978836e-01 6.67218864e-01 1.35568106e+00 -2.00053304e-01
-2.91810948e-02 2.08674759e-01 -4.67054129e-01 9.28349316e-01
3.13534826e-01 -3.54170233e-01 3.24016422e-01 -1.22772129e-02
-5.27578652e-01 -1.09086052e-01 8.46227825e-01 -1.31161988e+00
2.85250753e-01 2.85754234e-01 9.26950514e-01 -8.66732955e-01
3.46588582e-01 -6.28624976e-01 -2.65529484e-01 8.59580100e-01
1.26349255e-01 -4.20426995e-01 8.74866843e-02 5.37638009e-01
-6.54101595e-02 1.05562985e-01 1.22250009e+00 -3.42833638e-01
-7.73807108e-01 6.35180533e-01 1.94890574e-01 -7.39143640e-02
1.14671659e+00 -8.12352747e-02 1.64905220e-01 1.22893229e-01
-7.63508379e-01 4.53479171e-01 2.02357516e-01 4.00969118e-01
5.20294487e-01 -1.24473226e+00 -1.01170671e+00 1.44111380e-01
2.20891029e-01 -4.65928316e-02 3.33315492e-01 8.63778770e-01
-1.94114551e-01 3.59760523e-01 -2.80533671e-01 -6.44906580e-01
-1.73406434e+00 8.41225743e-01 5.74075341e-01 -2.15328664e-01
-7.91177452e-01 1.13683033e+00 5.74207842e-01 -2.32042342e-01
6.27908289e-01 -8.73401240e-02 -2.70600200e-01 1.90520540e-01
6.63694680e-01 3.70933652e-01 -2.05361173e-01 -9.38443124e-01
-7.40059078e-01 9.06627119e-01 -2.17253789e-02 2.72508889e-01
9.10259843e-01 -2.15632166e-03 -1.94510203e-02 -2.00109988e-01
1.08591688e+00 -7.03110546e-02 -1.33470166e+00 -4.21472698e-01
-3.61257821e-01 -3.40984225e-01 4.79656979e-02 -8.18409503e-01
-1.47197127e+00 1.08374858e+00 1.01644158e+00 -1.57636821e-01
1.05849409e+00 1.96189493e-01 9.18377042e-01 -5.26819937e-02
4.29934531e-01 -1.03371704e+00 2.70753086e-01 3.48738909e-01
8.26489449e-01 -1.49911320e+00 1.01755671e-01 -7.04300046e-01
-4.27260131e-01 9.26329136e-01 9.36693013e-01 7.86196291e-02
6.24449253e-01 -7.31780902e-02 -1.75244138e-01 -3.20289493e-01
-1.21122561e-01 -9.01689827e-01 5.87171197e-01 5.66649377e-01
3.65115643e-01 1.25874355e-01 -2.39619508e-01 8.52699041e-01
3.22858930e-01 2.17219800e-01 -3.39415997e-01 7.35686362e-01
-6.02420270e-01 -8.95313799e-01 -8.25280070e-01 3.13747108e-01
-5.39305985e-01 9.36818793e-02 -5.84711075e-01 8.51591229e-01
7.02168226e-01 8.78892422e-01 -1.42552987e-01 -2.44755611e-01
4.84938115e-01 6.95933998e-02 5.90350628e-01 -7.53930986e-01
-5.74683845e-01 2.71509260e-01 -2.12459177e-01 -5.45409620e-01
-2.87938416e-01 -6.37092471e-01 -1.11726487e+00 -1.18160345e-01
-4.18176174e-01 -3.33371133e-01 1.87049389e-01 8.63472044e-01
1.78622693e-01 4.40066308e-01 3.93743068e-01 -1.10572422e+00
-7.50010014e-01 -1.28128994e+00 -8.23617041e-01 3.49588156e-01
2.11259618e-01 -9.85095382e-01 1.21729575e-01 -1.07960500e-01] | [7.950474739074707, -0.5548710823059082] |
e7789942-85fd-4ee9-a2dd-a48782acd153 | mrgan360-multi-stage-recurrent-generative | 2303.08525 | null | https://arxiv.org/abs/2303.08525v1 | https://arxiv.org/pdf/2303.08525v1.pdf | MRGAN360: Multi-stage Recurrent Generative Adversarial Network for 360 Degree Image Saliency Prediction | Thanks to the ability of providing an immersive and interactive experience, the uptake of 360 degree image content has been rapidly growing in consumer and industrial applications. Compared to planar 2D images, saliency prediction for 360 degree images is more challenging due to their high resolutions and spherical viewing ranges. Currently, most high-performance saliency prediction models for omnidirectional images (ODIs) rely on deeper or broader convolutional neural networks (CNNs), which benefit from CNNs' superior feature representation capabilities while suffering from their high computational costs. In this paper, inspired by the human visual cognitive process, i.e., human being's perception of a visual scene is always accomplished by multiple stages of analysis, we propose a novel multi-stage recurrent generative adversarial networks for ODIs dubbed MRGAN360, to predict the saliency maps stage by stage. At each stage, the prediction model takes as input the original image and the output of the previous stage and outputs a more accurate saliency map. We employ a recurrent neural network among adjacent prediction stages to model their correlations, and exploit a discriminator at the end of each stage to supervise the output saliency map. In addition, we share the weights among all the stages to obtain a lightweight architecture that is computationally cheap. Extensive experiments are conducted to demonstrate that our proposed model outperforms the state-of-the-art model in terms of both prediction accuracy and model size. | ['Wei Xiang', 'Rong Quan', 'Xinlang Chen', 'Pan Gao'] | 2023-03-15 | null | null | null | null | ['saliency-prediction'] | ['computer-vision'] | [ 2.97067434e-01 1.54140577e-01 1.47288514e-03 -2.57158995e-01
-2.15741210e-02 3.09778452e-02 3.39805573e-01 -2.30776280e-01
-9.28544328e-02 2.49391034e-01 2.02319145e-01 -1.90674469e-01
1.72604188e-01 -6.75868511e-01 -6.24706149e-01 -5.04369497e-01
2.08116531e-01 -3.03862631e-01 6.27716184e-01 -2.81600207e-01
3.25516164e-01 1.45421773e-01 -1.60183668e+00 3.71130891e-02
1.09306240e+00 1.27419114e+00 8.10899198e-01 4.25199121e-01
2.39702567e-01 1.09224129e+00 -2.58698225e-01 -3.51025313e-01
4.95406501e-02 -4.05878931e-01 -4.02496189e-01 2.90653370e-02
5.59383631e-02 -2.28593498e-01 -6.21893823e-01 1.19155097e+00
4.88314062e-01 1.46644130e-01 2.87968755e-01 -1.17135632e+00
-9.95440125e-01 4.89867210e-01 -6.21369362e-01 2.36756876e-01
2.34464571e-01 -4.03170660e-03 8.97233307e-01 -1.05728149e+00
4.84129936e-01 9.92471695e-01 2.91597992e-01 5.72100043e-01
-8.33303809e-01 -6.79453552e-01 3.98286611e-01 4.71883625e-01
-1.33471215e+00 -2.26505831e-01 1.30938637e+00 -1.01166725e-01
5.60348153e-01 1.03822663e-01 8.17674041e-01 8.97530556e-01
5.53009331e-01 1.18853641e+00 9.10502434e-01 1.25831604e-01
2.22423583e-01 1.49427637e-01 -4.29792911e-01 6.78311169e-01
-1.14012159e-01 4.02950868e-02 -4.22721505e-01 4.88938183e-01
1.17544818e+00 4.03691560e-01 -2.03475550e-01 -6.79919004e-01
-1.18820119e+00 6.54616714e-01 1.25497282e+00 1.96450818e-02
-6.58959210e-01 -6.17525615e-02 1.61839247e-01 -6.48568571e-02
4.30541992e-01 4.08911347e-01 -5.62854148e-02 1.82008848e-01
-8.78884375e-01 1.67411342e-01 1.81161970e-01 8.45464528e-01
5.82147777e-01 3.92631143e-01 -1.19177205e-02 7.75473058e-01
3.69747162e-01 4.29291487e-01 6.76925480e-01 -4.68592167e-01
1.82238266e-01 7.16076553e-01 1.01491295e-01 -1.42430902e+00
-4.17225450e-01 -9.12201226e-01 -1.12284231e+00 2.81853825e-01
-2.32574597e-01 7.94640481e-02 -9.42281067e-01 1.59206462e+00
1.59937054e-01 4.25603390e-01 3.50088984e-01 1.42864907e+00
9.65766668e-01 7.29783475e-01 1.41374066e-01 2.16361120e-01
1.04157555e+00 -1.20660770e+00 -3.93771619e-01 -5.33840239e-01
-6.75101802e-02 -7.16793895e-01 9.88504410e-01 2.08042368e-01
-1.08495665e+00 -9.43084180e-01 -1.30341530e+00 -2.60459930e-01
-1.56316087e-01 1.92396626e-01 6.07363760e-01 4.59674709e-02
-1.08258986e+00 2.72163779e-01 -6.50088847e-01 1.12024192e-02
3.92896950e-01 2.24151194e-01 1.65297493e-01 1.83096990e-01
-1.29317772e+00 8.41924250e-01 1.37012392e-01 2.79735148e-01
-1.05557728e+00 -4.86472577e-01 -1.10664260e+00 3.00041318e-01
2.63359696e-01 -5.56823492e-01 1.11294329e+00 -1.15259624e+00
-1.52202475e+00 5.20089507e-01 -1.77572235e-01 -4.79187340e-01
4.23315525e-01 -2.36433551e-01 -5.05716383e-01 -1.46720046e-02
1.35248318e-01 9.03743088e-01 1.04306173e+00 -1.14988804e+00
-7.13750958e-01 -2.89808124e-01 1.97850481e-01 5.19965112e-01
-8.43307078e-02 -1.12654030e-01 -3.88595045e-01 -6.48340642e-01
5.35244882e-01 -9.63074565e-01 -5.20397425e-01 -1.36389241e-01
-5.66882253e-01 6.31368831e-02 1.05520320e+00 -7.30752647e-01
9.55355585e-01 -2.24301767e+00 3.05872619e-01 -1.51358783e-01
3.49672914e-01 3.02389145e-01 -2.12551579e-02 -2.31428280e-01
-4.19347268e-03 -2.21573442e-01 1.53724067e-02 -4.06916618e-01
-2.31765151e-01 -1.68659925e-01 -4.64134037e-01 1.87722191e-01
3.01734269e-01 1.21353352e+00 -1.12175071e+00 -3.41308683e-01
6.98698521e-01 5.78146756e-01 -4.74478364e-01 3.23013514e-01
-1.43616632e-01 5.49321413e-01 -5.75942278e-01 6.47190213e-01
6.87569499e-01 -5.73019266e-01 -6.69921860e-02 -1.92766696e-01
-2.89803445e-01 3.21114361e-01 -9.11160648e-01 1.77671456e+00
-6.77859843e-01 6.35296404e-01 -3.62673938e-01 -8.74292493e-01
1.23217690e+00 1.28217906e-01 1.32263273e-01 -1.01108420e+00
1.25904262e-01 2.20422447e-01 -2.44651530e-02 -1.85392171e-01
8.21366906e-01 1.62431091e-01 -2.44721100e-01 1.43692121e-02
-1.46764457e-01 -1.83963835e-01 -4.39854532e-01 9.03514028e-02
4.87032920e-01 1.75573304e-01 1.95170864e-01 -4.83390912e-02
6.02725506e-01 -2.99508661e-01 6.53808653e-01 3.63009065e-01
-1.94382399e-01 9.72620726e-01 1.95024312e-01 -6.77712679e-01
-1.22175562e+00 -1.18362284e+00 1.37367994e-01 8.53069484e-01
9.97955859e-01 1.57302305e-01 -6.22138321e-01 -3.74152511e-01
-3.45124513e-01 5.89825571e-01 -6.80514574e-01 -3.59064847e-01
-3.78035694e-01 -3.16440672e-01 -9.64791849e-02 5.78111053e-01
1.06521618e+00 -1.41650021e+00 -1.10591173e+00 2.69713521e-01
-2.14125484e-01 -1.05500019e+00 -3.79521310e-01 -1.29075378e-01
-6.81349933e-01 -7.24911749e-01 -1.06914735e+00 -1.05470133e+00
5.19150734e-01 7.13370979e-01 9.09611881e-01 -2.41576314e-01
9.89246741e-02 -3.42894584e-01 -2.90994525e-01 -6.02519691e-01
-2.94033680e-02 3.33902836e-01 -6.96715191e-02 1.07155867e-01
1.28703579e-01 -7.09938586e-01 -1.02554083e+00 4.75894868e-01
-8.93800676e-01 7.93895543e-01 8.81976426e-01 6.05371237e-01
5.99068880e-01 -1.49274364e-01 6.87455773e-01 -4.34632361e-01
3.27515990e-01 -7.08599091e-01 -4.82964605e-01 3.44446935e-02
-3.57736796e-01 -6.33905232e-02 7.53136516e-01 -2.91994154e-01
-1.08677411e+00 1.76119842e-02 -9.84103158e-02 -6.79486811e-01
1.18661806e-01 5.42818010e-01 -1.07717656e-01 5.59175620e-03
4.68765378e-01 5.82717001e-01 -4.85411063e-02 -1.19233064e-01
1.25608504e-01 4.79696065e-01 7.31901467e-01 4.07012522e-01
8.16598773e-01 3.00141901e-01 -2.15647265e-01 -5.59152126e-01
-1.10169613e+00 -1.27416536e-01 -4.02266711e-01 -4.21411514e-01
9.40385044e-01 -1.23089814e+00 -6.26251876e-01 6.45798743e-01
-1.13918245e+00 -1.68275923e-01 -3.26772928e-02 3.81512076e-01
-4.69777733e-01 -1.27292762e-03 -4.02637035e-01 -6.08463228e-01
-4.30182517e-01 -1.24089170e+00 9.78680313e-01 9.67187285e-01
1.55890852e-01 -6.84045315e-01 -3.43020231e-01 1.70140564e-01
6.27246141e-01 1.78929508e-01 5.39106607e-01 -2.85145223e-01
-8.34545374e-01 6.63008587e-03 -5.25671899e-01 2.87132293e-01
2.63590336e-01 -3.00702065e-01 -8.15160275e-01 -2.74593346e-02
1.05940752e-01 -1.75169751e-01 7.07347214e-01 5.65256655e-01
1.28641224e+00 -2.74704844e-01 -2.03856230e-01 5.75452924e-01
1.29240417e+00 2.96500653e-01 8.95194352e-01 4.73995715e-01
1.00991428e+00 4.03551966e-01 7.51888573e-01 3.64945263e-01
7.86278844e-01 5.77864885e-01 9.40341353e-01 -4.93758082e-01
5.07361852e-02 -6.10322416e-01 2.63101399e-01 8.28802526e-01
-5.14546111e-02 -3.08320839e-02 -6.10516489e-01 6.64041519e-01
-1.86344039e+00 -9.86617982e-01 2.93116003e-01 1.98936987e+00
4.07144725e-01 2.90231347e-01 -4.91801500e-02 8.82880017e-02
7.86431968e-01 5.47901750e-01 -9.33105707e-01 -3.89706045e-01
-1.40289411e-01 -9.18201283e-02 2.96899050e-01 1.44805029e-01
-1.03175175e+00 1.17592156e+00 5.37979984e+00 6.21633768e-01
-1.69758654e+00 -7.34258369e-02 9.25352275e-01 -5.34250140e-02
-3.40151340e-01 -3.16647977e-01 -4.82670724e-01 6.44783199e-01
4.74077523e-01 1.09925242e-02 4.95681077e-01 1.19261301e+00
1.70203418e-01 -2.72438824e-02 -5.80028117e-01 8.29032540e-01
2.08115697e-01 -1.36559391e+00 8.96956772e-02 -1.02467790e-01
1.12553859e+00 1.52471125e-01 7.03470945e-01 2.56562442e-01
1.52260363e-01 -1.10473740e+00 9.71400619e-01 5.82161844e-01
4.84678984e-01 -1.11811149e+00 8.27707052e-01 3.67478162e-01
-1.19552863e+00 -3.06331545e-01 -5.85733593e-01 -1.76856264e-01
1.48214474e-01 5.49682081e-01 -7.19334960e-01 4.99458224e-01
7.47570932e-01 9.53388631e-01 -5.95868647e-01 9.64529395e-01
-5.75118482e-01 1.29501477e-01 3.96280251e-02 -3.20559412e-01
3.72691959e-01 -1.21230975e-01 5.61878324e-01 6.49503052e-01
4.59147871e-01 2.47990340e-02 4.55365106e-02 1.05895925e+00
2.23256070e-02 -3.73486988e-02 -5.77858329e-01 4.66253996e-01
3.36660773e-01 1.28814745e+00 -4.28679824e-01 -3.36779237e-01
-4.52798307e-01 1.18772292e+00 2.28838846e-01 3.07705790e-01
-1.04073381e+00 -4.74060714e-01 6.85209870e-01 6.33985400e-02
5.94662309e-01 -5.31070344e-02 -5.45929611e-01 -1.22411263e+00
4.05043364e-02 -5.78265548e-01 -1.02898806e-01 -1.24758315e+00
-6.35323822e-01 9.41017687e-01 -5.51532328e-01 -1.45464647e+00
-3.48026931e-01 -1.86157480e-01 -7.60921240e-01 1.01188469e+00
-1.83419704e+00 -1.37484443e+00 -6.30811930e-01 5.86389244e-01
7.78870344e-01 -2.07345486e-02 4.48334098e-01 -4.29675356e-03
-3.89852732e-01 3.59946877e-01 -1.19963057e-01 1.12039603e-01
1.54054374e-01 -9.85243499e-01 7.35152900e-01 1.04083872e+00
1.51520025e-03 3.41507822e-01 7.32868075e-01 -4.63561624e-01
-1.04209375e+00 -1.25847995e+00 8.63045275e-01 1.87679902e-02
4.01291609e-01 -3.46617579e-01 -8.23982775e-01 5.97455025e-01
-5.75672835e-03 5.74341752e-02 1.86577782e-01 -1.52758524e-01
-2.12414395e-02 -1.33780047e-01 -9.59357321e-01 1.01351142e+00
1.05014157e+00 -5.90736568e-01 -4.56746429e-01 -2.38259748e-01
8.90755236e-01 -6.39489174e-01 -4.16548252e-01 6.32951140e-01
5.55840015e-01 -1.23413360e+00 1.07688296e+00 -5.44914417e-02
9.75962341e-01 -4.42265332e-01 2.48362347e-02 -1.34151149e+00
-5.06399572e-01 -2.76717693e-01 5.31360917e-02 8.15414071e-01
2.39094034e-01 -4.89973754e-01 8.09835672e-01 3.37785512e-01
-4.27864671e-01 -9.49370384e-01 -7.27234662e-01 -1.65596455e-01
-4.80518430e-01 -2.02207446e-01 6.18970931e-01 6.53486371e-01
-9.26845148e-02 4.22183603e-01 -6.89138055e-01 3.46837938e-01
4.22373682e-01 4.35605258e-01 6.31695688e-01 -1.11141348e+00
-1.93970948e-02 -3.09960186e-01 -7.40487099e-01 -1.41467535e+00
-6.71143383e-02 -6.46608293e-01 1.66405454e-01 -1.36364579e+00
1.98517323e-01 -3.80040973e-01 -6.37367666e-01 2.44128585e-01
-3.99177492e-01 6.01840138e-01 2.85353214e-01 2.78819829e-01
-7.52574623e-01 9.26601708e-01 1.66583788e+00 -2.16583654e-01
-3.00965399e-01 1.26621500e-01 -8.73818278e-01 7.95521855e-01
8.12472939e-01 3.40177980e-03 -6.56907439e-01 -5.39961338e-01
1.51934355e-01 2.67799824e-01 7.44648159e-01 -1.44143021e+00
3.54150057e-01 1.09056337e-02 6.88861132e-01 -7.32260942e-01
4.87401426e-01 -5.52771986e-01 8.98194760e-02 4.76808578e-01
-3.50678235e-01 4.87994626e-02 3.64900902e-02 3.84531826e-01
-4.13591295e-01 8.99634361e-02 8.82157624e-01 -4.49483842e-02
-1.07443309e+00 5.25547922e-01 -1.52795404e-01 -3.88798624e-01
9.66074288e-01 -3.47085983e-01 -3.96263376e-02 -5.39575815e-01
-6.47230327e-01 1.58453986e-01 6.84933007e-01 7.97466040e-01
1.01697791e+00 -1.37483823e+00 -5.39491773e-01 2.88823187e-01
-1.50977457e-02 3.00719321e-01 6.25279605e-01 5.80236256e-01
-3.16111147e-01 3.88060868e-01 -6.27356589e-01 -7.00027466e-01
-8.50937605e-01 7.38243639e-01 1.52002782e-01 -1.97635263e-01
-4.77025241e-01 7.25689054e-01 6.49683058e-01 -2.20782474e-01
-1.08974077e-01 -2.66715705e-01 -6.39877081e-01 -3.81308824e-01
5.20202935e-01 7.14384951e-03 -2.70952076e-01 -7.63332188e-01
-2.14446560e-01 4.55846101e-01 -5.96440099e-02 2.94877440e-02
1.30567825e+00 -3.50436777e-01 2.49519348e-01 2.70463765e-01
1.00163507e+00 -2.50362843e-01 -1.63557923e+00 -4.18681413e-01
-4.59076196e-01 -4.63151693e-01 1.71558395e-01 -6.31347716e-01
-1.34874129e+00 8.40315521e-01 6.04242086e-01 2.57461984e-02
1.20856822e+00 4.55430932e-02 9.50802445e-01 -1.23428449e-01
3.72230560e-01 -8.73294234e-01 3.08129668e-01 4.43330467e-01
8.72726202e-01 -1.32980824e+00 -2.16116711e-01 -4.72062856e-01
-1.06064880e+00 6.96389616e-01 7.90637374e-01 -4.92642254e-01
4.23800647e-01 -3.15952480e-01 1.30609006e-01 -5.38508296e-02
-4.57468450e-01 -4.88546193e-02 3.58907819e-01 5.77626646e-01
6.61690012e-02 1.25623524e-01 1.49059251e-01 6.62440360e-01
-4.75629091e-01 -5.10426983e-03 3.76950681e-01 4.98829186e-01
-4.05373275e-01 -3.97811532e-01 6.62589222e-02 2.24446863e-01
-3.29423696e-01 -1.72473013e-01 6.86284676e-02 4.21419561e-01
5.95015660e-02 8.52253675e-01 2.42339522e-01 -7.89880931e-01
2.06487134e-01 -5.83174825e-01 -1.17337862e-02 -4.08810884e-01
-2.93874055e-01 -1.96131140e-01 -3.88009369e-01 -5.26992619e-01
-4.35436130e-01 -6.24561489e-01 -1.23781526e+00 -1.92603752e-01
-1.14777550e-01 4.64396998e-02 7.78840065e-01 8.25234294e-01
5.47112405e-01 8.29632998e-01 1.10386443e+00 -1.21985590e+00
-2.76915222e-01 -9.27857399e-01 -3.84139597e-01 2.15606198e-01
2.82469898e-01 -7.89222002e-01 4.96285444e-04 -2.28238985e-01] | [9.78117847442627, -0.39375096559524536] |
50eb6a86-c533-4f02-baf2-f2bbcca3c875 | crab-learning-certifiably-fair-predictive | 2212.10839 | null | https://arxiv.org/abs/2212.10839v2 | https://arxiv.org/pdf/2212.10839v2.pdf | Consistent Range Approximation for Fair Predictive Modeling | This paper proposes a novel framework for certifying the fairness of predictive models trained on biased data. It draws from query answering for incomplete and inconsistent databases to formulate the problem of consistent range approximation (CRA) of fairness queries for a predictive model on a target population. The framework employs background knowledge of the data collection process and biased data, working with or without limited statistics about the target population, to compute a range of answers for fairness queries. Using CRA, the framework builds predictive models that are certifiably fair on the target population, regardless of the availability of external data during training. The framework's efficacy is demonstrated through evaluations on real data, showing substantial improvement over existing state-of-the-art methods. | ['Babak Salimi', 'Sainyam Galhotra', 'Nazanin Sabri', 'Jiongli Zhu'] | 2022-12-21 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [-1.94321200e-01 5.21291614e-01 -8.79635155e-01 -1.03452039e+00
-1.24004102e+00 -3.41165334e-01 4.77967143e-01 3.68296385e-01
-3.21721226e-01 1.08425987e+00 -5.29769324e-02 -4.53957498e-01
-4.21709657e-01 -9.72087026e-01 -7.25109339e-01 -1.45698950e-01
9.28310826e-02 1.14985573e+00 -1.15960337e-01 -7.57475644e-02
4.21814799e-01 2.04917610e-01 -1.62041438e+00 2.75044143e-01
1.05826330e+00 1.42279255e+00 -9.47992861e-01 4.33519214e-01
7.90598243e-02 1.35844862e+00 -7.01795042e-01 -1.36949766e+00
4.02724415e-01 -5.97522631e-02 -8.75454009e-01 -6.62328005e-01
7.89117396e-01 -7.48326600e-01 -8.97050947e-02 9.86511409e-01
2.47304067e-01 -5.38444854e-02 7.29917765e-01 -1.68088281e+00
-8.03790033e-01 5.49638391e-01 6.81399554e-02 8.41767117e-02
2.97291309e-01 -4.80447449e-02 1.30108035e+00 -7.16362894e-01
3.61944139e-01 1.33017504e+00 6.25815749e-01 6.86892867e-01
-1.40086007e+00 -1.01659203e+00 -4.41203266e-02 -3.44386920e-02
-1.62917376e+00 -7.02710509e-01 2.95248121e-01 -2.24912450e-01
6.26522481e-01 7.81943798e-01 3.33623081e-01 7.96172976e-01
4.50339131e-02 5.35834491e-01 1.10672367e+00 -1.22713476e-01
7.80678988e-01 5.95357180e-01 3.33728433e-01 1.41630307e-01
7.56095707e-01 6.58427179e-01 -8.54487181e-01 -1.16364753e+00
5.02254330e-02 -6.24175966e-02 1.90817460e-01 -4.59026366e-01
-3.92777622e-02 1.22427523e+00 3.23373437e-01 -3.45094919e-01
-5.82648635e-01 5.19004054e-02 2.96805203e-01 3.06295544e-01
9.71073687e-01 3.82689118e-01 -4.91997331e-01 1.06617659e-01
-1.31005609e+00 9.75672305e-01 1.17497706e+00 1.25669026e+00
4.03332353e-01 -3.53034914e-01 -5.17960191e-01 5.01812935e-01
4.07130808e-01 9.14320529e-01 -1.75858900e-01 -1.49902642e+00
4.47047740e-01 6.38012409e-01 6.29415274e-01 -9.75021899e-01
1.64718971e-01 -3.00901204e-01 -5.22027791e-01 -1.24113888e-01
5.01910329e-01 3.17542046e-01 -3.25878859e-01 1.89474297e+00
4.18888986e-01 -1.71665683e-01 1.18560597e-01 7.78337359e-01
5.70489764e-01 2.12507814e-01 4.52853411e-01 -5.19042611e-01
8.88792932e-01 -9.27410647e-02 -5.49634755e-01 8.71835798e-02
2.90672541e-01 -2.59223163e-01 8.15661252e-01 5.16835451e-01
-1.34386170e+00 -2.31734708e-01 -5.10462761e-01 -1.33548722e-01
-1.87907159e-01 -4.75977838e-01 6.21560931e-01 1.15986025e+00
-9.03043687e-01 2.99449950e-01 -3.02536458e-01 3.28022838e-02
9.57657397e-01 4.24406230e-01 -2.01788694e-02 -3.55717599e-01
-1.27963197e+00 8.39645684e-01 9.76180434e-02 -1.90439358e-01
-1.00085700e+00 -1.24814963e+00 -5.04883409e-01 1.20077305e-01
4.85009760e-01 -1.02254379e+00 1.39537013e+00 -8.58554423e-01
-7.76105642e-01 1.16561484e+00 -3.30440998e-01 -7.98153222e-01
1.10822594e+00 -3.31203133e-01 -3.43003869e-01 -1.10586256e-01
2.28137404e-01 2.89614528e-01 6.57529414e-01 -1.34286785e+00
-8.77729118e-01 -7.47159779e-01 -1.27594098e-01 -1.38967007e-01
-3.68271768e-02 3.10035050e-02 -7.80923963e-02 -2.47102827e-01
-2.69305021e-01 -4.46764678e-01 -2.53389806e-01 9.39846635e-02
-4.65807855e-01 -4.88232970e-01 3.91141355e-01 -6.25281632e-01
1.50447524e+00 -1.67620969e+00 -6.09833837e-01 8.30386698e-01
3.51588249e-01 -2.34172903e-02 2.49667570e-01 1.68980151e-01
3.18362206e-01 5.21757364e-01 -8.51441696e-02 -1.98597521e-01
2.28941783e-01 3.94612581e-01 -9.10979211e-01 4.26901728e-01
-2.15670210e-03 6.37143612e-01 -6.30631447e-01 -8.32363188e-01
-2.79606044e-01 -2.02929631e-01 -8.11949313e-01 6.16879106e-01
-3.22179317e-01 -1.21229328e-02 -4.92151856e-01 7.14895666e-01
9.21003044e-01 9.38346535e-02 2.60466635e-01 1.40081763e-01
2.69770235e-01 2.60424852e-01 -1.00475919e+00 7.57294536e-01
6.51571378e-02 -3.23792607e-01 -5.50079085e-02 -1.03358662e+00
1.11836207e+00 5.99918626e-02 3.00935298e-01 -7.73886740e-01
1.07149817e-02 4.37198818e-01 -5.43671072e-01 -4.49525267e-01
6.93533361e-01 -5.47900975e-01 -4.62013453e-01 5.27021050e-01
-1.62689462e-01 -2.58500665e-01 -1.04435846e-01 3.64955008e-01
6.67389333e-01 -2.81124562e-01 3.66870731e-01 -4.24459815e-01
4.50021356e-01 2.51281977e-01 5.99614143e-01 1.35420716e+00
-4.02392030e-01 1.59704149e-01 6.10127270e-01 -7.52124369e-01
-1.12947285e+00 -1.30459583e+00 -3.62322271e-01 1.29807281e+00
8.03465843e-02 -2.14222088e-01 -5.08728743e-01 -8.11335027e-01
6.48338795e-01 1.16108763e+00 -9.34975743e-01 -3.13275039e-01
1.20502133e-02 -4.19757336e-01 7.44208395e-01 3.92542273e-01
2.74522990e-01 -3.29096228e-01 -6.37771189e-01 -1.09232053e-01
-2.02689052e-01 -7.55077004e-01 -2.27130473e-01 -3.83745104e-01
-1.00416768e+00 -1.52782786e+00 -1.41024098e-01 9.70083056e-04
5.33021212e-01 -3.57468843e-01 1.84133101e+00 3.34544301e-01
1.46534741e-01 5.58077872e-01 3.40542555e-01 -8.57613504e-01
-6.81822717e-01 -3.87627453e-01 1.08817473e-01 -2.54389912e-01
8.90753508e-01 -1.36120498e-01 -5.57142913e-01 4.12603617e-01
-7.50093758e-01 -5.91319263e-01 3.27573717e-02 8.32808912e-01
8.36863101e-01 -1.67743966e-01 7.16974437e-01 -1.57223761e+00
9.15721774e-01 -6.25373602e-01 -9.14990366e-01 6.87674165e-01
-1.32301319e+00 -1.75972432e-02 4.77147102e-01 -7.91433454e-02
-1.13894367e+00 -5.02214313e-01 4.86636877e-01 -4.06479299e-01
1.27867805e-02 2.38400877e-01 -1.36523530e-01 2.85284609e-01
9.65941370e-01 -1.98082775e-01 2.38254905e-01 -3.21695179e-01
4.71085280e-01 7.72774756e-01 7.14740276e-01 -1.11979187e+00
3.79603356e-01 3.64409268e-01 1.29607335e-01 1.18870221e-01
-1.10452580e+00 -2.26530239e-01 -1.29387537e-02 -2.72802878e-02
6.57262355e-02 -8.07057679e-01 -1.20351708e+00 1.61692721e-03
-6.56932533e-01 1.88842639e-02 -5.51324368e-01 1.03806749e-01
-5.80537856e-01 -2.74242431e-01 -1.50961742e-01 -1.53422558e+00
-6.79533124e-01 -3.90331924e-01 8.29558730e-01 -2.30674352e-02
-5.00455737e-01 -8.84525836e-01 2.20121607e-01 7.78727829e-01
6.57934666e-01 6.30326793e-02 1.13307595e+00 -1.36394250e+00
-5.20177960e-01 -8.72449398e-01 -1.81048080e-01 3.02026242e-01
-4.51936066e-01 1.86031520e-01 -1.13974059e+00 -1.36909932e-01
-3.20503227e-02 -7.23820925e-01 4.60061520e-01 5.02626300e-01
1.42309368e+00 -1.07006562e+00 -2.68505484e-01 6.46198029e-03
1.32508504e+00 -3.10873687e-01 5.28572321e-01 -3.76321077e-02
-1.27780259e-01 8.65925670e-01 9.48685408e-01 8.60898972e-01
8.00216973e-01 3.50545466e-01 4.88405079e-01 3.66802692e-01
6.35649979e-01 -7.61045396e-01 -2.22057551e-01 -2.77206957e-01
-5.63453548e-02 1.46751359e-01 -7.74348021e-01 5.36510587e-01
-1.76433992e+00 -1.21299219e+00 1.65190786e-01 2.81795049e+00
9.93470907e-01 -1.51472597e-03 4.57867056e-01 -1.83364108e-01
5.91104507e-01 -3.42325211e-01 -1.05290043e+00 -8.30037117e-01
-6.27019703e-02 3.91635805e-01 6.41045690e-01 5.91972113e-01
-7.79630363e-01 4.98589963e-01 8.06688976e+00 8.40589046e-01
-5.73709011e-01 1.22420900e-01 1.44003248e+00 -3.37841928e-01
-9.07600224e-01 1.79214299e-01 -5.99774599e-01 3.64584893e-01
1.17774677e+00 -6.71455681e-01 3.87822419e-01 1.19860387e+00
4.41958569e-02 -2.64584064e-01 -1.45029199e+00 6.81643248e-01
-8.69367570e-02 -1.28353548e+00 1.31425902e-01 3.10505722e-02
8.19501519e-01 -4.27396357e-01 2.72390872e-01 6.53513372e-01
8.06163549e-01 -1.37874806e+00 9.08276081e-01 1.09916842e+00
7.50825763e-01 -9.29135501e-01 1.14549458e+00 6.87093735e-01
-2.87106156e-01 -3.85235786e-01 -5.30777156e-01 -1.29424669e-02
-4.27773684e-01 8.44851732e-01 -8.23754668e-01 5.74235260e-01
8.27448487e-01 1.88033268e-01 -6.44130111e-01 6.76640391e-01
2.41313338e-01 8.73955965e-01 -4.39258486e-01 9.39825326e-02
-4.57667142e-01 1.37551859e-01 -1.90919396e-02 6.51385188e-01
2.00322032e-01 1.89378083e-01 -2.20709424e-02 1.31775987e+00
-4.73850340e-01 2.26799816e-01 -6.81097209e-01 2.80615658e-01
9.93185639e-01 6.77777052e-01 5.34122050e-01 -6.42002940e-01
1.21159829e-01 -7.61082694e-02 3.94966274e-01 3.81321937e-01
-5.29894531e-01 4.49804604e-01 4.07841116e-01 1.97750017e-01
-2.49012172e-01 7.61181474e-01 -7.51354218e-01 -6.46171749e-01
7.99191818e-02 -1.15087533e+00 1.30708277e+00 -5.15930831e-01
-1.93050170e+00 1.44051434e-02 3.80032599e-01 -7.63188362e-01
-6.04678094e-01 -1.03662454e-01 -3.15571278e-01 1.18462765e+00
-1.20538580e+00 -1.24513853e+00 1.02513535e-02 7.01011002e-01
-3.04928511e-01 -3.42691630e-01 9.61047947e-01 -9.88240018e-02
-2.53096253e-01 1.03848505e+00 -1.99598670e-01 -3.32021445e-01
7.34164476e-01 -1.04906166e+00 -2.39534631e-01 4.07606989e-01
-3.16297770e-01 9.23542798e-01 9.90984142e-01 -6.83255076e-01
-1.07996869e+00 -9.72544968e-01 1.22309983e+00 -9.92676377e-01
1.43717289e-01 4.97762337e-02 -7.83524156e-01 5.84124923e-01
-3.86640817e-01 3.37251484e-01 1.06966436e+00 4.24047738e-01
-8.85091245e-01 -7.50905156e-01 -1.92426705e+00 -1.88729521e-02
7.42149830e-01 -5.81112802e-01 -6.14031851e-01 9.40961763e-02
3.54757309e-01 -4.50674355e-01 -9.22238827e-01 4.71664429e-01
7.83118963e-01 -1.14928365e+00 8.70927751e-01 -1.48215783e+00
4.89913940e-01 1.64147601e-01 -5.75837195e-01 -5.79580784e-01
1.18282847e-01 -4.95080501e-01 -2.80690372e-01 1.00084662e+00
4.93324339e-01 -8.57210875e-01 8.76047254e-01 1.93523216e+00
6.31770253e-01 -6.62321210e-01 -1.35615635e+00 -5.19425750e-01
4.58182514e-01 -6.36205256e-01 1.08912539e+00 7.01461673e-01
-2.03340203e-01 -2.22511634e-01 -5.47666192e-01 1.83284909e-01
1.19604456e+00 4.22438055e-01 1.02826774e+00 -1.49352407e+00
5.75695932e-02 -2.90735722e-01 -2.22212553e-01 -3.93568963e-01
2.88124770e-01 -4.82774168e-01 -1.64752096e-01 -8.88610661e-01
6.87045872e-01 -7.16063440e-01 -1.72280893e-01 4.46779221e-01
-3.74827802e-01 -1.04293309e-01 -3.33549716e-02 3.08348686e-01
-8.25992405e-01 3.97060275e-01 6.64738476e-01 -1.00730732e-01
5.70714921e-02 5.26827991e-01 -1.38593376e+00 2.59081423e-01
6.95282876e-01 -8.06380868e-01 -4.41755444e-01 2.00440541e-01
4.44484800e-01 8.85384455e-02 8.02801788e-01 -4.02947038e-01
2.99055785e-01 -4.97996420e-01 2.63183385e-01 -4.86517966e-01
1.19615523e-02 -9.66070771e-01 3.22672933e-01 5.45265377e-01
-1.13198042e+00 -1.28311738e-01 -7.29317963e-02 6.74137533e-01
-1.55655667e-01 -2.13637799e-02 5.50261915e-01 -4.79492266e-03
-5.65690361e-02 3.68724912e-01 4.35584128e-01 7.33585119e-01
8.49351943e-01 1.88543618e-01 -6.90876484e-01 -7.69449949e-01
-4.92749006e-01 5.89116335e-01 4.62285608e-01 -3.09358593e-02
6.46248221e-01 -1.30443573e+00 -9.08498764e-01 2.55117118e-01
4.69885886e-01 -1.48229629e-01 1.60003901e-01 5.03784537e-01
-7.14845881e-02 5.34577012e-01 1.55976295e-01 -4.17154849e-01
-9.84740794e-01 7.79716194e-01 5.43119133e-01 -2.63901055e-01
2.32084394e-01 6.82848930e-01 -1.98174566e-01 -6.65929914e-01
3.27627063e-01 -1.90370418e-02 2.99809456e-01 -2.90255845e-01
5.29270172e-01 5.87495983e-01 -7.69995749e-02 -3.87760639e-01
-3.95973921e-01 -6.26010522e-02 1.01693816e-01 -2.88895681e-03
1.02034819e+00 -1.49094284e-01 -3.00453275e-01 3.14313322e-01
7.66434133e-01 3.01258508e-02 -8.97465944e-01 -4.32648748e-01
7.53059760e-02 -1.07888961e+00 -1.92214236e-01 -1.08682418e+00
-8.72681618e-01 5.95017910e-01 5.90926111e-01 2.43919402e-01
9.30696845e-01 9.30006206e-02 1.51882112e-01 2.62843490e-01
5.24212956e-01 -1.21180439e+00 -4.97422814e-01 -1.89428315e-01
9.25264299e-01 -1.35942566e+00 3.31272155e-01 -8.27985331e-02
-8.46314311e-01 6.97012424e-01 5.96196711e-01 4.71576452e-02
8.23886096e-01 -1.12872951e-01 7.02594742e-02 -1.99716806e-01
-1.16572666e+00 3.55040997e-01 4.71341044e-01 6.35173917e-01
2.90738679e-02 4.14942682e-01 -3.11214238e-01 1.23650444e+00
-2.83956140e-01 5.48215151e-01 1.21585429e-01 5.39296210e-01
-2.42709696e-01 -8.13527763e-01 -3.94467920e-01 9.49885368e-01
-6.88676775e-01 -5.11768758e-02 -5.07545114e-01 7.32826829e-01
1.40032917e-01 1.18950784e+00 6.27101287e-02 -9.73608121e-02
5.84332764e-01 1.45977095e-01 1.24227874e-01 -2.17340901e-01
-5.82491279e-01 -5.07989287e-01 3.02191347e-01 -9.07174468e-01
-5.10618865e-01 -6.63672209e-01 -6.32410705e-01 -8.86671185e-01
3.13346870e-02 5.44354141e-01 1.74031109e-01 8.08376014e-01
4.23305064e-01 -3.87878597e-01 7.20913947e-01 1.18075691e-01
-1.55966294e+00 -5.99858105e-01 -6.09644890e-01 8.36653411e-01
1.85507894e-01 -3.55809003e-01 -3.11573535e-01 -3.31567138e-01] | [8.871914863586426, 5.302847862243652] |
b4c02f68-9a04-47f9-8832-e933f7cf7b50 | a-residual-solver-and-its-unfolding-neural | 2009.03477 | null | https://arxiv.org/abs/2009.03477v1 | https://arxiv.org/pdf/2009.03477v1.pdf | A Residual Solver and Its Unfolding Neural Network for Total Variation Regularized Models | This paper proposes to solve the Total Variation regularized models by finding the residual between the input and the unknown optimal solution. After analyzing a previous method, we developed a new iterative algorithm, named as Residual Solver, which implicitly solves the model in gradient domain. We theoretically prove the uniqueness of the gradient field in our algorithm. We further numerically confirm that the residual solver can reach the same global optimal solutions as the classical method on 500 natural images. Moreover, we unfold our iterative algorithm into a convolution neural network (named as Residual Solver Network). This network is unsupervised and can be considered as an "enhanced version" of our iterative algorithm. Finally, both the proposed algorithm and neural network are successfully applied on several problems to demonstrate their effectiveness and efficiency, including image smoothing, denoising, and biomedical image reconstruction. The proposed network is general and can be applied to solve other total variation regularized models. | ['Yuanhao Gong'] | 2020-09-08 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 2.87356496e-01 9.55217704e-02 2.59833872e-01 -1.25198111e-01
-3.65018815e-01 1.14467843e-02 -5.03783561e-02 -4.96061444e-01
-3.93124521e-01 9.04991806e-01 -1.70320123e-01 2.01221276e-02
-2.66150266e-01 -5.13062418e-01 -7.95235753e-01 -1.00974560e+00
1.10311508e-01 -9.64129791e-02 -9.18439105e-02 -3.26764137e-02
2.84204572e-01 2.68661678e-01 -9.74891245e-01 -2.47912169e-01
1.29030180e+00 8.69132817e-01 3.91140252e-01 4.29540098e-01
-1.39093399e-01 6.75713181e-01 -2.11600289e-01 1.78327471e-01
3.08601528e-01 -6.61082745e-01 -8.98355782e-01 3.64763141e-01
2.83785313e-01 -2.34227225e-01 -3.05685937e-01 1.35266149e+00
5.91029167e-01 4.48305458e-01 3.83618772e-01 -7.13720381e-01
-8.45308542e-01 2.86950350e-01 -7.99293697e-01 9.54605415e-02
-7.98810199e-02 -2.23985627e-01 3.27564269e-01 -1.03598940e+00
6.71591341e-01 9.83147025e-01 1.05557036e+00 5.89771211e-01
-1.17610466e+00 -3.66375446e-01 2.22066343e-01 -2.93819644e-02
-1.35378957e+00 -1.60764426e-01 9.86501992e-01 -4.65769440e-01
4.26358044e-01 1.67894498e-01 4.73491460e-01 5.50151169e-01
2.45752484e-01 7.99451888e-01 1.18698001e+00 -4.20761138e-01
1.77744739e-02 8.91126245e-02 4.11459088e-01 8.43369067e-01
1.24934673e-01 1.11296482e-01 1.29318982e-02 -9.14486125e-02
9.81640875e-01 -1.22603349e-01 -8.10426176e-01 -4.84361611e-02
-1.06094968e+00 7.76569724e-01 5.09387016e-01 4.54243034e-01
-4.00380164e-01 6.93652257e-02 1.66109189e-01 9.84136388e-02
7.87532747e-01 2.54617393e-01 -2.72208095e-01 4.00944352e-01
-1.07907975e+00 1.50411040e-01 7.63334334e-01 8.11876237e-01
8.86018515e-01 4.17385876e-01 -2.16988504e-01 1.06713879e+00
2.74407834e-01 5.21674454e-01 6.49020612e-01 -1.13248014e+00
-3.45375426e-02 4.31583673e-01 1.65252432e-01 -1.30396080e+00
-5.79022348e-01 -7.07820237e-01 -1.34179568e+00 2.52675060e-02
3.64380151e-01 -5.68081439e-01 -7.80955493e-01 1.72203231e+00
5.08923173e-01 8.09773088e-01 -4.72974665e-02 1.11140227e+00
9.92865860e-01 7.64740348e-01 -2.96297044e-01 -6.20679259e-01
8.50900769e-01 -1.09841990e+00 -1.10787237e+00 -1.18931286e-01
4.82325494e-01 -7.27095306e-01 6.75276697e-01 4.58151996e-01
-1.16132081e+00 -5.44478416e-01 -8.51386487e-01 -5.09559885e-02
9.47247222e-02 5.79791009e-01 3.80150467e-01 3.91126126e-01
-1.21277523e+00 9.65715766e-01 -7.89305449e-01 -1.40756160e-01
2.25973040e-01 2.37936184e-01 -1.69667020e-01 2.52095580e-01
-1.02515161e+00 7.23171175e-01 1.68593943e-01 8.46932650e-01
-5.65355539e-01 -7.60111153e-01 -7.14215875e-01 -1.89509556e-01
3.89831036e-01 -6.68698132e-01 8.61015141e-01 -1.22571051e+00
-1.68675101e+00 7.34127820e-01 -4.57608789e-01 -8.93455595e-02
4.45354760e-01 -2.64423490e-01 -2.06020605e-02 -7.18568936e-02
-1.46075012e-02 1.35236681e-01 9.11248088e-01 -1.30616808e+00
-1.89359576e-01 -1.09144792e-01 -3.40033054e-01 -7.47698247e-02
-2.87169665e-01 -2.13376194e-01 -6.21352136e-01 -8.36147487e-01
3.64148974e-01 -8.69952917e-01 -7.08516836e-01 -9.42176208e-02
-4.74131525e-01 1.73859224e-01 4.78466928e-01 -1.14850807e+00
1.24410379e+00 -2.11636543e+00 3.78335863e-01 4.48621631e-01
2.99557567e-01 2.13003069e-01 -1.65432975e-01 -5.68991229e-02
-3.73809725e-01 -3.75105254e-02 -9.48621869e-01 -2.52582163e-01
-4.24935222e-01 1.12033360e-01 -1.09731548e-01 8.16660047e-01
-1.88924447e-02 9.42626119e-01 -6.74284101e-01 -4.46257174e-01
1.46834448e-01 7.99213648e-01 -5.00761926e-01 1.40597150e-01
1.28234357e-01 9.46628630e-01 -5.47908068e-01 1.15819059e-01
1.35743701e+00 -4.27218974e-01 1.93214551e-01 -4.18298095e-01
-2.71370947e-01 -4.58289206e-01 -1.60962808e+00 1.56291652e+00
-4.41124469e-01 6.22918010e-01 5.88347554e-01 -1.59050155e+00
1.06546330e+00 9.09605324e-02 6.92344666e-01 -5.07749259e-01
2.65625268e-01 2.56800979e-01 -4.62660640e-01 -6.95673645e-01
4.30843234e-01 -3.03768963e-01 5.74706137e-01 2.19063565e-01
-3.32817249e-02 1.03530876e-01 1.22946337e-01 8.35862756e-02
6.36872768e-01 2.11307272e-01 2.69310981e-01 -5.34456968e-01
9.92346168e-01 -1.68128029e-01 7.18821645e-01 7.52757668e-01
-5.64390384e-02 6.52764618e-01 4.22022849e-01 -4.19151485e-01
-7.28875995e-01 -7.87210763e-01 -4.58899319e-01 3.69545013e-01
2.96605140e-01 -3.53054218e-02 -9.97570992e-01 -3.46787900e-01
-2.54182428e-01 1.90688223e-01 -6.15410924e-01 6.01125509e-02
-7.00322986e-01 -1.09747422e+00 3.56152087e-01 4.38633233e-01
9.23127234e-01 -9.41395462e-01 -1.81964189e-01 2.78803706e-01
-3.44717562e-01 -1.04886878e+00 -6.33827031e-01 -7.82088935e-02
-1.03730619e+00 -1.26116669e+00 -9.84954238e-01 -8.51983488e-01
9.65618730e-01 7.26951212e-02 8.98482919e-01 4.98297960e-01
-2.60744184e-01 5.14975429e-01 2.73677241e-02 1.02402613e-01
-1.97378874e-01 -2.09106609e-01 -1.44620210e-01 3.86844248e-01
-2.54217505e-01 -5.07348657e-01 -6.78302526e-01 3.17775071e-01
-8.63835931e-01 -9.54758450e-02 3.27215582e-01 1.06063199e+00
9.29150701e-01 1.56066582e-01 5.32451868e-01 -1.08824432e+00
8.56785476e-01 -3.62112314e-01 -8.63835037e-01 2.71259576e-01
-6.80921853e-01 1.77498013e-01 6.94585562e-01 -3.71407092e-01
-1.38979971e+00 2.53696382e-01 -1.51788130e-01 -4.43950564e-01
9.66301635e-02 6.72908068e-01 1.57110646e-01 -4.82199937e-01
4.50845331e-01 2.82585114e-01 3.14999223e-01 -6.57250941e-01
3.38849932e-01 2.38674775e-01 7.20103145e-01 -5.28942168e-01
7.73564816e-01 6.52008057e-01 2.84160525e-01 -9.97763395e-01
-8.04572701e-01 -4.23020810e-01 -4.63518471e-01 -2.22832784e-01
9.06406939e-01 -5.94247341e-01 -9.30593371e-01 8.32683623e-01
-1.22631383e+00 -4.87708449e-01 -1.40726075e-01 4.66593266e-01
-4.96122867e-01 8.59352648e-01 -7.97292829e-01 -7.69697249e-01
-4.08724368e-01 -1.12491417e+00 6.13394439e-01 3.71850252e-01
2.60714650e-01 -1.59368527e+00 3.39151204e-01 -9.27819163e-02
5.76922297e-01 3.78689080e-01 4.68197972e-01 -8.97371843e-02
-2.48371512e-01 7.69288242e-02 -3.80507886e-01 6.81800663e-01
1.31452680e-02 -3.41391489e-02 -7.56412745e-01 -2.81306297e-01
7.01715648e-01 7.26131499e-02 1.03135514e+00 1.00850844e+00
1.20952797e+00 -2.92319328e-01 -3.64523619e-01 1.16797256e+00
1.67764962e+00 -1.54350586e-02 7.02027559e-01 1.27036914e-01
6.53491497e-01 5.69578350e-01 2.30705306e-01 3.64478141e-01
-1.11033060e-01 4.98909771e-01 1.57821119e-01 -4.90709037e-01
4.27935906e-02 2.82567948e-01 8.96846950e-02 1.17550981e+00
-2.67296463e-01 1.33796960e-01 -6.01378441e-01 4.47880059e-01
-2.03068757e+00 -9.02889490e-01 -6.45971358e-01 1.95225763e+00
7.81243324e-01 -5.14530063e-01 -3.34595442e-01 -6.21913262e-02
1.00822699e+00 -1.48653733e-02 -6.50330961e-01 -4.59781945e-01
-1.87493637e-01 4.62756276e-01 5.07462382e-01 7.92504191e-01
-1.24475729e+00 7.29950905e-01 7.29500484e+00 9.26760018e-01
-1.28816533e+00 8.86774361e-02 6.96096480e-01 1.47962019e-01
-1.59461081e-01 -1.54926702e-01 -5.50522685e-01 3.40238482e-01
5.53733349e-01 -2.22116336e-01 7.56983042e-01 7.77065933e-01
5.26520431e-01 -6.15893304e-02 -5.97886086e-01 9.82135773e-01
1.81710780e-01 -1.35395253e+00 -3.53318989e-01 -1.94664061e-01
1.03913164e+00 -3.32992136e-01 4.67908792e-02 -2.57067513e-02
3.53676714e-02 -1.11170435e+00 4.59670499e-02 8.47367346e-01
5.87233126e-01 -5.63099623e-01 7.68159807e-01 3.15908074e-01
-1.16672122e+00 1.24933839e-01 -5.32681167e-01 -1.07223630e-01
2.93967634e-01 9.60755646e-01 -1.17554843e-01 8.06388438e-01
4.84443247e-01 1.05351365e+00 -2.35889435e-01 1.13542533e+00
-9.33013633e-02 3.74176651e-01 -1.96099773e-01 2.84413308e-01
2.30461225e-01 -8.37255359e-01 7.21714079e-01 9.84265029e-01
4.06619251e-01 2.65869945e-01 1.20303661e-01 1.21594608e+00
6.90876320e-02 3.89562815e-01 -3.59142989e-01 3.66890877e-01
-1.06106885e-01 1.36567140e+00 -7.33621359e-01 -1.70475617e-01
-3.69557023e-01 9.28277612e-01 2.45131299e-01 8.63738000e-01
-7.66853392e-01 -4.79848057e-01 3.16267639e-01 -2.68653691e-01
2.49512285e-01 -4.65219282e-02 -4.64375377e-01 -1.27266562e+00
1.15543799e-02 -6.87419295e-01 1.26189768e-01 -7.68500745e-01
-1.26953936e+00 6.41808927e-01 -5.92056923e-02 -1.06326556e+00
9.05913338e-02 -8.15842032e-01 -7.07389653e-01 1.03135788e+00
-1.44089520e+00 -5.35753071e-01 -4.98719066e-01 7.77741134e-01
2.09258571e-01 1.12481199e-01 4.17794853e-01 5.38456857e-01
-8.98116708e-01 3.92612189e-01 4.59498972e-01 4.77056438e-03
5.25389075e-01 -1.06715775e+00 -1.97301403e-01 1.01714051e+00
-4.10138577e-01 7.92313576e-01 5.57786405e-01 -7.01703072e-01
-1.13512421e+00 -1.10113931e+00 5.48217654e-01 1.29025206e-01
5.41725159e-01 2.01008499e-01 -1.15185714e+00 7.75965035e-01
3.19325000e-01 -3.66223510e-03 1.59365043e-01 -2.14215547e-01
4.30811644e-01 -7.75238350e-02 -1.17728984e+00 4.27389622e-01
9.34146285e-01 -1.37025312e-01 -2.86275327e-01 3.51025701e-01
6.37983561e-01 -8.21759820e-01 -6.70972288e-01 5.33029795e-01
1.93155333e-01 -7.57070005e-01 1.00199819e+00 -3.59356016e-01
5.10077834e-01 -2.15751514e-01 1.46728486e-01 -1.34613478e+00
-3.01933110e-01 -7.94489741e-01 1.43972477e-02 9.90976274e-01
1.25214726e-01 -8.81129682e-01 4.50087816e-01 6.40976667e-01
-2.73989886e-01 -9.22580659e-01 -9.86614168e-01 -6.72931015e-01
2.37726763e-01 -4.13082451e-01 2.21802071e-01 1.04906309e+00
-2.83421367e-01 -1.33973181e-01 -7.26747453e-01 1.55983031e-01
8.85717928e-01 -2.67004669e-02 5.10782182e-01 -9.38727796e-01
-2.47506499e-01 -4.60528314e-01 -4.00375910e-02 -1.39061630e+00
4.62694436e-01 -1.00648749e+00 2.55440772e-01 -1.52868044e+00
2.08721116e-01 -2.24050432e-01 -3.03169578e-01 1.10702492e-01
-2.97907472e-01 1.98660791e-01 -9.24980715e-02 7.02752247e-02
-1.13052847e-02 5.38566530e-01 1.63532889e+00 -8.68276134e-02
-5.98084271e-01 1.62349734e-02 -6.32781029e-01 8.54216337e-01
5.98564684e-01 -2.13014260e-01 -2.51164883e-01 -3.65633816e-01
1.22485541e-01 1.62339117e-02 4.20354456e-01 -7.49377131e-01
2.12294415e-01 -1.04989521e-01 1.57407060e-01 -2.52498299e-01
-3.57195698e-02 -7.30746686e-01 2.28102550e-01 3.85628074e-01
-2.88427949e-01 -3.36818963e-01 2.82026768e-01 4.23324764e-01
-2.31950045e-01 -6.19651556e-01 1.00519192e+00 -1.95139617e-01
-5.36074758e-01 3.12735319e-01 -2.25465640e-01 1.35229886e-01
7.45615661e-01 -1.07710987e-01 -1.32664129e-01 -3.37800860e-01
-1.23621178e+00 2.07857385e-01 1.32389694e-01 -3.17059547e-01
7.64066041e-01 -1.30451524e+00 -6.45792544e-01 2.87429273e-01
-6.87512994e-01 -8.65535885e-02 5.68002164e-01 1.32540178e+00
-6.81704283e-01 1.54452235e-01 2.18820661e-01 -6.09455168e-01
-9.20507669e-01 6.25426471e-01 7.64730513e-01 -4.14325297e-01
-9.15976703e-01 6.29561543e-01 3.74714822e-01 -5.89086831e-01
1.66311577e-01 -2.50190943e-01 -2.28137374e-01 -3.78816098e-01
4.66047108e-01 7.51889825e-01 -3.17366064e-01 -5.91558933e-01
-3.49441946e-01 1.15881491e+00 4.32670861e-01 1.13382287e-01
1.42244506e+00 -3.05918545e-01 -6.78793073e-01 1.59335867e-01
1.45301819e+00 1.38409799e-02 -1.18677127e+00 -1.92107648e-01
-2.45215610e-01 -2.19789848e-01 3.04889947e-01 -3.05810601e-01
-1.68368495e+00 6.98818386e-01 6.65143907e-01 7.06528425e-02
1.34560204e+00 -4.69353110e-01 8.33933115e-01 3.04789424e-01
-2.20399708e-01 -1.13547909e+00 -3.16191167e-02 5.22562444e-01
1.03148091e+00 -1.06815565e+00 1.43861458e-01 -5.73084533e-01
-3.34901899e-01 1.18688309e+00 5.32291591e-01 -5.45021415e-01
9.85917330e-01 2.95648903e-01 -7.06679281e-03 -6.12995103e-02
-2.08043233e-01 -1.16301496e-02 3.92154664e-01 3.33752096e-01
5.38827479e-01 -3.36290389e-01 -9.56438124e-01 5.79709351e-01
2.38418624e-01 4.90031272e-01 5.51745117e-01 4.24140096e-01
-1.86125949e-01 -7.47947156e-01 -5.19757509e-01 2.21328527e-01
-3.96430373e-01 -1.42128497e-01 3.67942192e-02 6.77736640e-01
-9.82863829e-02 8.61751974e-01 -7.64117688e-02 -9.02605057e-02
2.70771652e-01 3.86307910e-02 5.39551616e-01 -9.36567336e-02
-4.87178981e-01 2.87332207e-01 -3.52992475e-01 -6.70261502e-01
-7.58341908e-01 -2.16822773e-01 -1.37005615e+00 -1.64186999e-01
-3.44124317e-01 1.67317688e-01 4.21195477e-01 1.04167187e+00
1.37789533e-01 7.20568836e-01 6.32437706e-01 -8.19679022e-01
-3.17402512e-01 -9.08955812e-01 -6.54280722e-01 3.95950615e-01
3.93178254e-01 -5.88732541e-01 -8.15009654e-01 2.20887080e-01] | [11.651649475097656, -2.524151086807251] |
7affb89d-23db-4989-9009-2e5d3fea5044 | noise-aware-physics-informed-machine-learning | 2206.12901 | null | https://arxiv.org/abs/2206.12901v5 | https://arxiv.org/pdf/2206.12901v5.pdf | Noise-aware Physics-informed Machine Learning for Robust PDE Discovery | This work is concerned with discovering the governing partial differential equation (PDE) of a physical system. Existing methods have demonstrated the PDE identification from finite observations but failed to maintain satisfying results against noisy data, partly owing to suboptimal estimated derivatives and found PDE coefficients. We address the issues by introducing a noise-aware physics-informed machine learning (nPIML) framework to discover the governing PDE from data following arbitrary distributions. We propose training a couple of neural networks, namely solver and preselector, in a multi-task learning paradigm, which yields important scores of basis candidates that constitute the hidden physical constraint. After they are jointly trained, the solver network estimates potential candidates, e.g., partial derivatives, for the sparse regression algorithm to initially unveil the most likely parsimonious PDE, decided according to the information criterion. We also propose the denoising physics-informed neural networks (dPINNs), based on Discrete Fourier Transform (DFT), to deliver a set of the optimal finetuned PDE coefficients respecting the noise-reduced variables. The denoising PINNs are structured into forefront projection networks and a PINN, by which the formerly learned solver initializes. Our extensive experiments on five canonical PDEs affirm that the proposed framework presents a robust and interpretable approach for PDE discovery, applicable to a wide range of systems, possibly complicated by noise. | ['Takashi Morita', 'Ken-ichi Fukui', 'Masayuki Numao', 'Pongpisit Thanasutives'] | 2022-06-26 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 2.38410041e-01 -2.32231587e-01 1.93170354e-01 2.60770936e-02
-9.39979672e-01 -4.95403796e-01 3.82425725e-01 -3.77477616e-01
6.66891634e-02 1.12033927e+00 -5.32509643e-04 -5.76372668e-02
-8.82747829e-01 -4.96322423e-01 -7.42300808e-01 -1.24416721e+00
-1.49094984e-01 7.48920500e-01 -7.09004328e-02 -8.80503282e-02
4.62630510e-01 8.06648910e-01 -1.33731055e+00 9.81611684e-02
1.10031903e+00 1.12440693e+00 3.41961235e-01 4.79161084e-01
8.60298350e-02 8.02023947e-01 -3.06003422e-01 5.77364936e-02
4.11387801e-01 -5.72000325e-01 -6.83954239e-01 9.91358534e-02
2.17227295e-01 -2.05642611e-01 -3.32465172e-01 1.07755899e+00
3.43218297e-01 5.48748612e-01 8.14598203e-01 -1.09463179e+00
-6.12476707e-01 2.84910440e-01 -3.13269526e-01 2.84345657e-01
1.18107878e-01 2.49094695e-01 1.07604325e+00 -9.69807148e-01
6.07500732e-01 1.24049270e+00 8.82117867e-01 2.90949225e-01
-1.40852892e+00 -3.85016799e-01 -2.12588627e-02 2.29001135e-01
-1.55250883e+00 -3.38827878e-01 1.28013813e+00 -5.05768418e-01
7.33214676e-01 3.04408848e-01 3.38873923e-01 1.16735172e+00
5.82844436e-01 3.32550049e-01 1.19750905e+00 -1.92389693e-02
4.73353177e-01 9.36494321e-02 3.95852178e-02 6.34848118e-01
1.24341525e-01 2.96469212e-01 -5.37613094e-01 -7.28446364e-01
1.14741075e+00 -4.09574062e-01 -5.55665195e-01 -3.55933249e-01
-8.60381424e-01 9.05479789e-01 1.99359030e-01 2.09890455e-01
-8.04264009e-01 -2.05246165e-01 1.02142483e-01 1.50895789e-01
3.67570072e-01 9.02999938e-01 -7.67159104e-01 2.18126047e-02
-1.02057087e+00 3.58932108e-01 1.08737254e+00 5.65610707e-01
8.77233505e-01 3.16524923e-01 -1.10596351e-01 7.12106705e-01
2.15235606e-01 2.91900456e-01 1.46097094e-01 -1.31744230e+00
1.37955502e-01 4.54460651e-01 1.76864535e-01 -1.31575763e+00
-3.00076097e-01 -5.98339021e-01 -1.36456823e+00 3.06467354e-01
1.16082899e-01 -3.85678560e-01 -7.61685312e-01 1.46575391e+00
4.22073394e-01 5.84001720e-01 4.27040681e-02 1.27428210e+00
6.98212028e-01 9.01287913e-01 -1.62771672e-01 -4.44716424e-01
8.91324282e-01 -6.37326717e-01 -5.55524290e-01 8.01610500e-02
3.18586111e-01 -5.37373543e-01 6.06906176e-01 8.79677355e-01
-1.05859900e+00 -7.29436398e-01 -6.77120030e-01 2.37128049e-01
7.05082491e-02 2.87855893e-01 5.25229633e-01 1.92662794e-02
-1.05817735e+00 1.34289551e+00 -8.94437909e-01 2.21332684e-01
1.95680112e-01 4.55798656e-01 -1.90701783e-01 2.72275567e-01
-1.06492543e+00 8.84678543e-01 4.39108759e-01 5.91382802e-01
-1.11896741e+00 -9.78908837e-01 -5.94631553e-01 7.15778917e-02
5.16915262e-01 -9.75904822e-01 6.06556296e-01 -8.68529141e-01
-1.76455045e+00 3.17134142e-01 -1.52082801e-01 -3.68437648e-01
3.63909900e-01 -7.57918730e-02 -3.16362381e-01 2.52718121e-01
5.23225404e-02 3.20451826e-01 1.17897546e+00 -1.59346163e+00
-2.86372125e-01 1.54744446e-01 -3.97749469e-02 2.81043649e-02
-9.43750963e-02 -2.68763185e-01 -2.99444288e-01 -6.27500296e-01
4.49661583e-01 -8.69537532e-01 -6.24046445e-01 -1.05010450e-01
-7.09265769e-01 -1.42155051e-01 7.61021495e-01 -9.33805048e-01
1.13683999e+00 -1.81370842e+00 9.00104165e-01 4.50482398e-01
3.28232318e-01 5.31104617e-02 -9.80418697e-02 4.31689262e-01
-2.76523978e-01 -1.51546389e-01 -5.78959346e-01 -3.70099425e-01
-1.99317336e-01 4.72741783e-01 -4.43016261e-01 6.24713719e-01
6.23434842e-01 4.31914568e-01 -6.06680870e-01 -4.05856907e-01
2.62267858e-01 5.15903652e-01 -5.56790352e-01 4.89131212e-01
-3.15195709e-01 1.20934772e+00 -9.22733247e-01 4.37989861e-01
9.13253427e-01 -4.41768378e-01 -1.31067947e-01 -5.71325243e-01
-2.11417601e-01 -7.87566826e-02 -1.76879370e+00 1.45091200e+00
-2.61586577e-01 3.73473614e-02 6.28881872e-01 -1.45476758e+00
1.17746174e+00 2.68247753e-01 8.61883700e-01 -2.15681374e-01
1.94204435e-01 3.60855401e-01 -4.55233343e-02 -8.08822274e-01
2.68234432e-01 -2.56275862e-01 9.91664380e-02 3.53137963e-02
1.66879892e-01 -2.29687184e-01 1.46798655e-01 -6.98003918e-02
1.26314163e+00 3.53821009e-01 2.49391973e-01 -6.27451718e-01
1.12043095e+00 1.49377197e-01 7.89027989e-01 7.64114022e-01
1.33193716e-01 7.50186980e-01 5.01341164e-01 -5.98652065e-01
-1.02796388e+00 -7.85388470e-01 -2.25810140e-01 5.14237046e-01
6.25636950e-02 -4.67870682e-02 -3.40511769e-01 -2.22222015e-01
8.62277076e-02 5.27968585e-01 -4.29445893e-01 -1.07706882e-01
-9.69547212e-01 -7.28833258e-01 1.83210820e-01 9.18682143e-02
2.29959980e-01 -1.14285719e+00 -3.27925801e-01 4.28107172e-01
9.90307853e-02 -9.96398032e-01 3.09387539e-02 4.05622661e-01
-9.22967970e-01 -1.05094731e+00 -6.77076817e-01 -5.89366615e-01
5.96486568e-01 -2.94323146e-01 8.36820304e-01 6.04482666e-02
-1.96181715e-01 2.55787134e-01 2.13180669e-02 4.60120052e-01
-5.92812717e-01 -4.52028587e-02 2.16050819e-01 2.58306146e-01
-3.80688369e-01 -1.07218385e+00 -2.97282010e-01 1.30024338e-02
-6.75351739e-01 -1.32987395e-01 7.69580722e-01 1.02334952e+00
9.85533774e-01 4.01849657e-01 3.38748157e-01 -6.73439920e-01
6.76793873e-01 -8.37092638e-01 -7.78520525e-01 9.47781876e-02
-2.82142311e-01 3.63220781e-01 1.33822680e+00 -6.92859828e-01
-1.22823036e+00 2.72450000e-01 -1.71076432e-01 -1.14731610e+00
-3.90931100e-01 7.14172781e-01 -7.76864635e-03 -5.14022470e-01
5.78198075e-01 3.77974182e-01 -2.49994799e-01 -8.11626673e-01
1.21763974e-01 1.28968051e-02 6.51431978e-01 -1.00434768e+00
9.80726957e-01 3.28393459e-01 5.19376457e-01 -8.66872847e-01
-8.64226878e-01 -4.54165131e-01 -7.82429099e-01 -3.39623205e-02
5.92943251e-01 -7.00769901e-01 -9.30931628e-01 3.60833079e-01
-1.35967255e+00 5.69146499e-02 -1.54724345e-01 4.69649673e-01
-4.96443719e-01 5.12304246e-01 -6.29864097e-01 -9.84350801e-01
-2.74228841e-01 -1.13306844e+00 1.04776454e+00 1.23902775e-01
-3.00005376e-01 -1.19001150e+00 2.89934516e-01 1.24508679e-01
3.26827675e-01 5.68752587e-01 1.05682540e+00 -4.52477455e-01
-7.56034911e-01 1.97708100e-01 -4.42852229e-02 3.94855380e-01
4.15405408e-02 1.65804476e-01 -8.79321337e-01 -3.55827987e-01
6.24017715e-01 -2.66831160e-01 9.12272036e-01 8.52378130e-01
1.27669299e+00 -4.65627760e-01 -3.41545880e-01 1.06712103e+00
1.68830490e+00 2.06257403e-01 2.27164343e-01 -7.66538829e-02
8.67828071e-01 3.22999924e-01 3.82606834e-02 7.26076543e-01
-1.35543466e-01 2.85745203e-01 4.81863499e-01 -4.15709568e-03
7.52466694e-02 1.29304215e-01 2.20024586e-01 9.68876719e-01
-2.43525192e-01 -3.02137673e-01 -6.46411479e-01 3.05826247e-01
-1.74887288e+00 -6.04824901e-01 -2.65351206e-01 1.73221588e+00
7.17623293e-01 7.93240294e-02 -1.99607015e-01 -9.07369703e-02
5.91633379e-01 1.10656716e-01 -9.34018135e-01 -3.02382976e-01
-2.34090194e-01 4.60872084e-01 2.70461470e-01 6.49682820e-01
-1.12269342e+00 5.43171525e-01 6.11131144e+00 9.51791942e-01
-1.22437179e+00 -7.12375492e-02 4.51238632e-01 1.30848125e-01
-3.71714503e-01 1.07472241e-01 -8.36814642e-01 4.15055305e-01
8.69348049e-01 -8.59699547e-02 7.35857189e-01 7.38380313e-01
6.72157705e-01 1.04388028e-01 -9.52015996e-01 8.93546820e-01
-1.34790331e-01 -1.52247858e+00 2.35316351e-01 5.41045703e-03
8.84316325e-01 -1.26595154e-01 -1.32849976e-01 4.50377390e-02
1.93685248e-01 -9.76521134e-01 6.15104556e-01 1.08819556e+00
3.44403297e-01 -4.89362091e-01 5.30810595e-01 5.60636342e-01
-1.01431537e+00 -2.01373681e-01 -5.16524911e-01 -2.34600350e-01
3.32010001e-01 8.27388167e-01 -2.92893559e-01 8.21032226e-01
5.39854825e-01 1.08819997e+00 -3.52249034e-02 9.72703993e-01
3.99085619e-02 6.21598542e-01 -4.47571784e-01 1.54820219e-01
3.74159992e-01 -7.81147182e-01 1.02935469e+00 7.06633985e-01
6.38412476e-01 5.45082748e-01 4.52696353e-01 1.57708752e+00
1.90805718e-01 -8.30204785e-02 -3.58258456e-01 1.34258404e-01
2.64221817e-01 1.47019470e+00 -7.90738404e-01 -1.07475907e-01
-4.21314500e-02 8.33461046e-01 2.57391691e-01 6.38584197e-01
-7.99872160e-01 2.45053530e-01 6.45067751e-01 -2.90131181e-01
5.84692776e-01 -2.20131084e-01 -2.94447124e-01 -1.12184477e+00
1.54965192e-01 -9.24686134e-01 3.64737719e-01 -6.15282834e-01
-1.75934494e+00 6.68077528e-01 5.14931418e-02 -1.24039984e+00
-1.01292253e-01 -8.14467430e-01 -9.07378674e-01 9.71269011e-01
-1.48889232e+00 -7.58205891e-01 -1.19949065e-01 6.23943329e-01
3.29435498e-01 5.72084822e-02 7.55658031e-01 2.01333642e-01
-7.58475840e-01 -6.70776591e-02 3.57672632e-01 -3.02518994e-01
1.89694986e-01 -1.10164762e+00 -1.94888428e-01 6.65637970e-01
-1.58874422e-01 4.93839830e-01 1.03472507e+00 -8.72362912e-01
-1.86319149e+00 -8.67538095e-01 4.09478903e-01 -2.25571454e-01
7.83340871e-01 3.85617018e-02 -1.33590114e+00 3.20181072e-01
-6.38776347e-02 -1.09318318e-02 2.17708405e-02 -1.65888742e-01
2.79346645e-01 -2.83496734e-02 -1.12538350e+00 1.84876651e-01
9.03072417e-01 -2.72217959e-01 -5.35421729e-01 5.44751823e-01
4.58137661e-01 -6.53931797e-01 -1.06731093e+00 4.06384677e-01
1.19775660e-01 -7.47405827e-01 1.19268513e+00 -6.69980288e-01
6.81311965e-01 -2.46747524e-01 9.14977938e-02 -1.29702151e+00
-5.89499414e-01 -1.01981831e+00 -4.00346309e-01 1.07416046e+00
2.84707159e-01 -5.74690044e-01 7.30091095e-01 5.81006408e-01
-4.68134671e-01 -1.03623748e+00 -1.22726011e+00 -7.63375401e-01
9.18870494e-02 -2.61774123e-01 2.41495326e-01 1.16292620e+00
-6.65400982e-01 2.20897317e-01 -6.53623044e-01 6.43411636e-01
7.77347982e-01 3.04062784e-01 4.41651106e-01 -1.42596865e+00
-6.50015593e-01 -3.52413952e-01 9.98796225e-02 -1.01803899e+00
3.75438362e-01 -8.06868494e-01 1.97364196e-01 -1.18394196e+00
1.13097262e-02 -4.39735323e-01 -3.34315687e-01 1.78305075e-01
6.43823892e-02 -2.39736840e-01 -6.97360933e-02 4.36024040e-01
-2.61883736e-01 8.05730879e-01 1.55744255e+00 2.53876150e-02
-3.74404579e-01 1.91004500e-01 -4.88533229e-01 8.98190916e-01
4.70826626e-01 -6.40761137e-01 -2.32960790e-01 -1.86784223e-01
2.88667101e-02 5.46468079e-01 6.92228019e-01 -1.11202061e+00
3.99297327e-01 -3.63025784e-01 3.32642436e-01 -5.90224564e-01
5.91545224e-01 -8.12604129e-01 4.47505087e-01 3.20503801e-01
-2.19880834e-01 -6.87829182e-02 2.76657283e-01 3.74042064e-01
-1.08618401e-01 -4.25965011e-01 6.18810654e-01 -2.29307592e-01
-6.87920332e-01 4.34689522e-01 -3.62468392e-01 1.20252455e-02
5.43431461e-01 -1.14663623e-01 2.76936740e-01 -1.73793346e-01
-9.66095507e-01 1.08359233e-01 -4.55026217e-02 -7.82177895e-02
4.79737997e-01 -1.10160625e+00 -7.78551102e-01 3.61175269e-01
-5.95944703e-01 5.87892048e-02 4.28378165e-01 7.90812850e-01
-3.13404649e-01 1.99982747e-01 -2.02600807e-01 -7.02265739e-01
-6.59045339e-01 4.87757862e-01 5.09664237e-01 -5.01986325e-01
-8.45698655e-01 9.79812741e-01 -6.47275522e-02 -3.85387748e-01
2.51523107e-02 -6.51332378e-01 -1.53144553e-01 -1.09115593e-01
-3.60736027e-02 5.96105456e-01 -9.48954076e-02 -6.69808984e-01
-2.82465398e-01 8.31946194e-01 4.31135923e-01 3.17884773e-01
1.70056593e+00 3.17555852e-02 -4.47504222e-01 1.08335838e-01
1.31352484e+00 -2.14791641e-01 -1.54127324e+00 -2.88068116e-01
-6.24819994e-02 -9.31371599e-02 2.54503429e-01 -5.88093579e-01
-1.19808578e+00 5.63141108e-01 1.28302857e-01 3.27758610e-01
1.19653606e+00 -1.02329947e-01 8.65390182e-01 6.00025654e-01
5.87791763e-02 -8.45563471e-01 -1.10642388e-01 5.64043343e-01
1.06022346e+00 -1.04704797e+00 9.34594199e-02 -4.29924488e-01
-2.23184779e-01 1.46832895e+00 5.70548594e-01 -6.17329121e-01
8.06328177e-01 3.46281737e-01 -2.70819992e-01 -4.36427742e-01
-6.58432543e-01 1.27202854e-01 6.47022069e-01 2.72973537e-01
-8.57907459e-02 -3.15494746e-01 -6.98489100e-02 7.63795435e-01
-1.43051669e-01 -6.34938404e-02 2.66524822e-01 5.10983646e-01
-3.34115475e-01 -8.45914662e-01 -5.67068815e-01 4.56563056e-01
-1.43223293e-02 8.58902931e-04 -2.11501539e-01 7.34624982e-01
4.34906691e-01 5.69322765e-01 -1.70715228e-01 -8.31597447e-02
2.75407940e-01 -1.31629348e-01 1.69659525e-01 -3.67843300e-01
-4.82737660e-01 2.35123456e-01 -1.86658166e-02 -5.26311815e-01
-3.70835930e-01 -7.12077141e-01 -1.11617208e+00 -2.19837546e-01
-2.31421515e-01 2.84702808e-01 1.89383682e-02 1.40436459e+00
4.81106907e-01 6.14872515e-01 7.13805735e-01 -1.28772020e+00
-7.61018395e-01 -7.49708176e-01 -6.12871826e-01 3.40602875e-01
5.30720115e-01 -9.07941937e-01 -7.23139048e-01 3.77067998e-02] | [6.513397216796875, 3.510263442993164] |
5f36755a-5af0-4857-b69a-a57ac3962354 | iterative-signal-processing-for-integrated | 2306.05235 | null | https://arxiv.org/abs/2306.05235v1 | https://arxiv.org/pdf/2306.05235v1.pdf | Iterative Signal Processing for Integrated Sensing and Communication Systems | Integrated sensing and communication (ISAC), with sensing and communication sharing the same wireless resources and hardware, has the advantages of high spectrum efficiency and low hardware cost, which is regarded as one of the key technologies of the fifth generation advanced (5G-A) and sixth generation (6G) mobile communication systems. ISAC has the potential to be applied in the intelligent applications requiring both communication and high accurate sensing capabilities. The fundamental challenges of ISAC system are the ISAC signal design and ISAC signal processing. However, the existing ISAC signal has low anti-noise capability. And the existing ISAC signal processing algorithms have the disadvantages of quantization errors and high complexity, resulting in large energy consumption. In this paper, phase coding is applied in ISAC signal design to improve the anti-noise performance of ISAC signal. Then, the effect of phase coding method on improving the sensing accuracy is analyzed. In order to improve the sensing accuracy with low-complexity algorithm, the iterative ISAC signal processing methods are proposed. The proposed methods improve the sensing accuracy with low computational complexity, realizing energy efficient ISAC signal processing. Taking the scenarios of short distance and long distance sensing into account, the iterative two-dimensional (2D) fast Fourier transform (FFT) and iterative cyclic cross-correlation (CC) methods are proposed, respectively, realizing high sensing accuracy and low computational complexity. Finally, the feasibility of the proposed ISAC signal processing methods are verified by simulation results. | ['Zhiyong Feng', 'Huici Wu', 'Kaifeng Han', 'Wangjun Jiang', 'Hanyang Qu', 'Zhiqing Wei'] | 2023-06-08 | null | null | null | null | ['quantization'] | ['methodology'] | [ 7.02018857e-01 -3.60637039e-01 -3.97010833e-01 1.81889653e-01
-4.16137993e-01 -1.98318213e-01 1.61454007e-01 -5.17981173e-03
-2.59387374e-01 4.35085833e-01 1.28801502e-02 -4.76183116e-01
-6.36151195e-01 -7.72162735e-01 1.38403550e-01 -1.00165391e+00
-4.57549691e-01 -4.22344476e-01 4.87940386e-02 -3.93421873e-02
6.23511747e-02 2.90497392e-01 -8.78335655e-01 -4.03631836e-01
1.05848336e+00 1.59050143e+00 4.94981915e-01 5.09122014e-01
2.08048284e-01 3.34065616e-01 -7.54514039e-01 1.02792576e-01
2.34772131e-01 -7.81799138e-01 5.97942472e-02 -1.74027205e-01
-7.15345025e-01 -5.98346032e-02 -1.16798297e-01 1.20537782e+00
7.47363508e-01 -1.98336795e-01 5.01318932e-01 -1.00332677e+00
-2.99763739e-01 7.36474633e-01 -8.59533727e-01 4.87057306e-02
1.73631072e-01 -1.83057666e-01 4.44705933e-01 -7.76670218e-01
-1.72834277e-01 9.64146733e-01 7.97005951e-01 1.59059003e-01
-6.61920309e-01 -1.11884248e+00 -4.95566696e-01 4.45931137e-01
-1.76495111e+00 -4.18236107e-01 1.02032208e+00 -1.97789092e-02
4.03020471e-01 3.64575237e-01 9.62758541e-01 7.82608837e-02
1.16707504e-01 2.69940883e-01 1.01529765e+00 -8.62659037e-01
3.68771434e-01 -3.14218014e-01 -2.08829895e-01 6.88278735e-01
7.63039470e-01 4.34334368e-01 -1.53233692e-01 3.49826887e-02
5.20588219e-01 8.77455994e-02 -5.17954946e-01 1.63373366e-01
-1.30067968e+00 3.83793443e-01 5.11236489e-01 8.27482879e-01
-2.90004820e-01 4.77007717e-01 -8.95018969e-03 -1.04349824e-02
-1.73340872e-01 1.04694545e-01 9.61951688e-02 -1.57147557e-01
-1.00027168e+00 -4.61367816e-01 3.52269322e-01 9.79071498e-01
4.69764054e-01 5.09330750e-01 5.81489205e-02 6.21651173e-01
8.34236205e-01 1.47607207e+00 6.36211753e-01 -7.90744245e-01
3.93139333e-01 2.47683376e-01 3.71166095e-02 -1.47071314e+00
-7.63911784e-01 -1.00745583e+00 -1.25763345e+00 -2.63025105e-01
-5.34957200e-02 -4.54567403e-01 -4.65808839e-01 1.45293868e+00
-8.54676366e-02 2.96796262e-01 5.44123292e-01 7.32424617e-01
5.63050210e-01 7.80370116e-01 -3.03034931e-01 -8.76301110e-01
1.54399979e+00 -2.54878670e-01 -1.25657582e+00 -3.01770955e-01
3.04532319e-01 -9.19739783e-01 3.42790782e-01 2.38141611e-01
-8.39630365e-01 -7.81609535e-01 -1.81735742e+00 5.01274884e-01
8.77982154e-02 6.58988953e-01 4.37745422e-01 1.39038002e+00
-5.24231672e-01 -1.62206233e-01 -6.40754223e-01 -1.27338141e-01
4.25448805e-01 2.71813869e-01 6.11735702e-01 -7.82902241e-02
-1.32290113e+00 4.40859705e-01 3.89394820e-01 1.02488190e-01
1.08289272e-01 -6.95007026e-01 -5.07753491e-01 2.74305463e-01
2.98119575e-01 -1.42560378e-01 1.05320263e+00 -5.47154188e-01
-1.59308565e+00 -2.57967468e-02 -2.25101057e-02 -6.20421529e-01
-1.47359744e-01 2.67228901e-01 -1.34881306e+00 2.82986313e-01
-1.34000376e-01 1.81126501e-02 5.45780241e-01 -7.34057903e-01
-7.54761636e-01 -4.95865881e-01 -5.04373252e-01 8.65784958e-02
-6.26324043e-02 -4.21613246e-01 -1.98065206e-01 -7.79540300e-01
7.16477692e-01 -7.69391596e-01 -3.57493013e-01 -2.14531660e-01
-5.08244596e-02 3.86330694e-01 1.24868596e+00 -4.25625533e-01
1.51610672e+00 -2.44581437e+00 -2.98855633e-01 6.74698055e-01
-2.02834874e-01 5.89149952e-01 1.30444780e-01 3.13582927e-01
4.26102936e-01 8.53173584e-02 -1.84491888e-01 2.19517902e-01
-3.24999720e-01 -1.40564710e-01 2.12723270e-01 5.04478633e-01
-2.33456612e-01 6.65384293e-01 -7.09623516e-01 -4.39732075e-01
3.16139549e-01 3.04602295e-01 -2.90556669e-01 -5.72119772e-01
2.76000410e-01 4.82840300e-01 -6.57218874e-01 7.60265172e-01
8.97086740e-01 -2.22666800e-01 2.89082050e-01 -6.76419675e-01
-2.24757001e-01 1.48646813e-02 -1.43517745e+00 1.49420655e+00
-6.38514161e-01 3.33057702e-01 4.15323615e-01 -1.07446492e+00
1.24287033e+00 5.11389017e-01 8.08414638e-01 -1.24207282e+00
2.72410750e-01 8.03563356e-01 4.28237110e-01 -2.69305378e-01
1.29139066e-01 -1.55424520e-01 -2.62986362e-01 3.90099615e-01
-3.73883098e-01 -5.66963740e-02 9.37741324e-02 -1.74893700e-02
6.85847938e-01 -4.07321393e-01 8.29838157e-01 -5.60938179e-01
1.15574074e+00 -1.08208880e-01 1.01383817e+00 2.11169273e-01
-2.06157357e-01 -1.42529890e-01 -5.36595464e-01 1.91569507e-01
-4.32705373e-01 -5.88376880e-01 5.41965701e-02 -1.02024421e-01
1.00507164e+00 -3.91135663e-01 -2.72314578e-01 1.84527561e-01
-7.07043484e-02 4.33880746e-01 3.39073777e-01 -3.36829811e-01
-5.71521342e-01 -7.23108113e-01 5.91850579e-01 3.04975867e-01
1.36219621e+00 -2.20298514e-01 -1.06661665e+00 4.51192051e-01
-2.51800597e-01 -1.14408755e+00 -5.07075906e-01 -2.16215178e-01
-4.90810305e-01 -9.55927312e-01 -5.47533572e-01 -5.84809363e-01
2.25064754e-01 7.45883584e-01 3.49196076e-01 1.63731933e-01
-3.05513561e-01 3.71849000e-01 -4.38431203e-01 -5.92169940e-01
7.70548061e-02 -4.14893657e-01 2.74872124e-01 1.80163547e-01
2.44360059e-01 -5.51865876e-01 -9.55749512e-01 3.64666313e-01
-3.99951965e-01 -3.19583192e-02 8.71894896e-01 4.14250940e-01
3.93572360e-01 8.70539188e-01 8.22067380e-01 -2.92979479e-01
3.54762673e-01 -1.01130418e-01 -9.69905794e-01 1.14766330e-01
-8.28778744e-01 -4.27911997e-01 8.21567357e-01 -1.13067754e-01
-1.10477602e+00 7.92812183e-02 -2.43409708e-01 1.39002591e-01
7.35438526e-01 6.98022127e-01 -4.87006128e-01 -4.26266104e-01
3.97962034e-01 4.74949211e-01 1.24976777e-01 -2.48240426e-01
5.27382530e-02 7.77265489e-01 5.67154408e-01 -1.98034629e-01
1.13640308e+00 3.81493479e-01 4.66464818e-01 -8.93289685e-01
-3.46856356e-01 -3.92533243e-01 4.56178859e-02 -3.06158215e-01
6.89693511e-01 -1.18041742e+00 -8.34223986e-01 2.99023688e-01
-8.49236786e-01 8.10417831e-02 6.54751842e-04 1.22649002e+00
-1.26169622e-01 4.20797765e-01 -1.49781197e-01 -1.06085694e+00
-6.12647533e-01 -1.16376448e+00 4.92952704e-01 5.71227133e-01
-1.63292900e-01 -6.31109893e-01 -5.00495434e-01 1.22732051e-01
8.00522566e-01 1.16315648e-01 7.84743369e-01 9.00811180e-02
-7.73022056e-01 2.15654690e-02 -1.57780886e-01 -2.54475892e-01
3.79995197e-01 -5.55174530e-01 -3.97259295e-01 -3.26170683e-01
-8.48051906e-02 5.59462905e-01 3.43716472e-01 5.99974811e-01
9.82579470e-01 -2.91608900e-01 -7.82599986e-01 7.04788506e-01
1.83999741e+00 1.07862723e+00 8.18578780e-01 -3.10773224e-01
7.58225620e-02 -9.12737399e-02 1.03693306e+00 5.28736711e-01
-1.13365494e-01 6.68298781e-01 2.58180499e-01 1.35328636e-01
-8.67098048e-02 -1.80275857e-01 1.88196063e-01 1.06907630e+00
-1.08832121e-02 -2.78333724e-01 -4.85632151e-01 1.64231747e-01
-1.46890736e+00 -1.26334310e+00 -5.44755340e-01 2.07030249e+00
4.80073959e-01 -3.07027604e-02 -2.86184549e-01 8.71679604e-01
5.62948465e-01 1.36326611e-01 -4.55800414e-01 -1.30736763e-02
-9.28653032e-02 3.74057859e-01 1.14540005e+00 4.81356412e-01
-8.03780079e-01 -8.33854079e-02 5.13201189e+00 1.21903467e+00
-1.23750365e+00 2.51582772e-01 1.60094693e-01 3.70814413e-01
-2.74114937e-01 4.12336625e-02 -6.20309949e-01 8.71454418e-01
6.61844432e-01 -2.86256701e-01 2.01551273e-01 4.60575372e-01
4.96736825e-01 -3.23765337e-01 -3.88913929e-01 1.53579235e+00
-2.80010939e-01 -1.43008399e+00 -5.01377165e-01 1.63764447e-01
5.88381290e-01 -5.52801311e-01 1.75794009e-02 1.05541490e-01
-2.89039046e-01 -5.82781792e-01 5.73079169e-01 5.09421110e-01
1.25475109e+00 -7.75738537e-01 7.11049914e-01 5.29014230e-01
-1.93295348e+00 -4.74475116e-01 -6.68479949e-02 -2.02584356e-01
3.16091508e-01 1.35615075e+00 -1.62156254e-01 8.83857667e-01
4.91264611e-01 6.54393673e-01 -8.59655365e-02 8.52253199e-01
2.06243526e-02 6.80489957e-01 -5.90546548e-01 -4.15125072e-01
-1.38465747e-01 -4.96880800e-01 5.61385870e-01 8.66847932e-01
1.15070462e+00 7.56819367e-01 2.82989204e-01 5.02192974e-01
1.57490820e-01 1.75009891e-01 -2.56209701e-01 1.23647854e-01
1.36076975e+00 9.13667738e-01 -5.73020458e-01 -2.71530747e-01
-6.32026911e-01 2.68260330e-01 -1.08151555e+00 2.46134683e-01
-6.76304221e-01 -8.66409481e-01 1.22158892e-01 7.89994746e-02
1.49740994e-01 -5.26290298e-01 -6.29975259e-01 -4.54722404e-01
-4.81747799e-02 -6.33077741e-01 7.13294148e-02 -3.06991130e-01
-5.37733674e-01 -1.90899223e-01 -5.36552250e-01 -1.82915342e+00
-2.99936850e-02 -1.39806330e-01 -7.38732696e-01 8.69485974e-01
-1.44788682e+00 -6.31887853e-01 -5.98800480e-01 6.36854291e-01
2.10780665e-01 -3.72057587e-01 6.84896469e-01 7.66480505e-01
-1.94127247e-01 5.91834724e-01 4.02471930e-01 -8.39667767e-02
-4.92322855e-02 -5.58374643e-01 -3.49462420e-01 1.25425303e+00
-3.58240068e-01 5.28902590e-01 2.94563949e-01 -5.90227962e-01
-1.62136149e+00 -1.05075622e+00 5.88576674e-01 9.94132817e-01
1.26864001e-01 -6.73554987e-02 -2.17132613e-01 -8.91565755e-02
6.07865937e-02 1.04065485e-01 9.09524381e-01 -4.77208585e-01
2.52960622e-01 -7.86968470e-01 -1.36473095e+00 5.83488703e-01
8.25132072e-01 -2.83128977e-01 1.59067705e-01 -9.23289582e-02
5.13425350e-01 -6.73990399e-02 -8.73545527e-01 7.51808703e-01
7.26953924e-01 -6.46765471e-01 1.00607753e+00 1.00207126e+00
-4.15279597e-01 -1.04921830e+00 -5.95584452e-01 -9.68584001e-01
-5.11550486e-01 -9.30409670e-01 -1.87006369e-01 1.20191979e+00
2.62885183e-01 -8.92966747e-01 6.91113353e-01 -3.47361058e-01
6.09663688e-02 -5.30556440e-01 -9.27373886e-01 -1.07919824e+00
-6.93715036e-01 -3.85128528e-01 6.74182475e-01 5.17050982e-01
3.80706161e-01 5.03762484e-01 -3.52354318e-01 3.71745586e-01
6.95879340e-01 3.76338176e-02 3.41571182e-01 -1.42684615e+00
-5.42207539e-01 -3.57998699e-01 -3.89584720e-01 -1.52548051e+00
-7.23525584e-01 -5.84823787e-01 -1.39052510e-01 -1.50356078e+00
-3.20930362e-01 -8.40521872e-01 -1.68513909e-01 -1.01004444e-01
8.61388966e-02 4.47833180e-01 1.50004193e-01 2.80002832e-01
-2.46940836e-01 4.38061088e-01 1.18753648e+00 -3.02865922e-01
-8.84483531e-02 4.93813127e-01 -4.63189751e-01 4.52871948e-01
9.21840072e-01 6.19431064e-02 -7.92680264e-01 -5.62374704e-02
6.55926242e-02 5.57592690e-01 3.59898582e-02 -1.59496248e+00
5.57390273e-01 4.31948714e-02 4.48701888e-01 -9.47424054e-01
6.01314902e-01 -1.50980234e+00 4.89130199e-01 1.32361293e+00
3.34558666e-01 -3.48297924e-01 -1.34302855e-01 7.30825841e-01
-2.15652108e-01 1.17662903e-02 9.93547499e-01 2.47947842e-01
-8.01179171e-01 1.37837321e-01 -6.88273549e-01 -3.08911890e-01
1.34382093e+00 -4.66212034e-01 -1.74353778e-01 -5.66902936e-01
-1.78473607e-01 2.76445746e-01 -9.68891233e-02 -1.24104366e-01
7.40069628e-01 -1.64901197e+00 -3.84195268e-01 3.36568028e-01
-8.30509514e-02 -3.54924917e-01 3.81568104e-01 1.17022491e+00
-4.71701205e-01 7.74602175e-01 4.90836017e-02 -7.71179914e-01
-1.30465448e+00 3.14869612e-01 3.45073253e-01 -1.09699860e-01
-1.85278162e-01 2.24012509e-01 -5.86508155e-01 1.58607438e-01
-8.82274508e-02 -1.19067803e-01 -1.47326276e-01 -3.92788738e-01
5.99589825e-01 5.24015486e-01 -1.33272588e-01 -5.18606186e-01
-5.17666638e-01 1.22545862e+00 6.33787036e-01 -8.82836655e-02
5.01894295e-01 -5.05030870e-01 1.48803487e-01 -1.78656504e-01
1.18125880e+00 7.35809445e-01 -6.48145854e-01 -3.36127788e-01
1.47629296e-02 -3.98182929e-01 1.34353891e-01 -7.81056583e-01
-1.23365045e+00 6.45271480e-01 1.05904114e+00 1.34267047e-01
1.80416930e+00 -5.34801900e-01 1.07064497e+00 9.98669416e-02
1.10958815e+00 -1.03139496e+00 3.71304341e-02 -7.22968206e-02
2.62574673e-01 -7.44418263e-01 6.01612747e-01 -7.49055564e-01
1.02983408e-01 1.08648276e+00 2.77512893e-02 1.92439392e-01
1.03037524e+00 4.30306137e-01 -3.45160626e-02 1.62840430e-02
-1.41962737e-01 -3.12529147e-01 4.38785776e-02 8.49983573e-01
2.32735753e-01 3.43574584e-01 -1.05006230e+00 8.14469397e-01
-1.29965976e-01 -7.50280172e-02 3.92731488e-01 7.27119625e-01
-7.37018466e-01 -1.14448309e+00 -6.67002022e-01 4.76316214e-01
-5.82123101e-01 2.34115571e-01 2.21652910e-01 6.01581752e-01
5.64959407e-01 1.57635939e+00 -1.37773350e-01 -6.44988954e-01
1.68226376e-01 -5.84248304e-01 2.30911821e-01 -7.11751506e-02
2.32503206e-01 3.99015188e-01 3.77398357e-02 -4.00941670e-01
-6.96364582e-01 -4.67084229e-01 -1.77197659e+00 -1.62837803e-01
-5.13839900e-01 4.52334762e-01 7.69674301e-01 8.85988295e-01
4.68779713e-01 7.29116082e-01 9.91624653e-01 -1.30398571e-01
-2.48922408e-01 -5.64138412e-01 -7.79784620e-01 -1.40802607e-01
2.07718581e-01 -5.12232244e-01 2.79771090e-02 -1.86511084e-01] | [6.347332954406738, 1.2148429155349731] |
753b08b3-bdee-4c39-8c1c-01228e0ca500 | video-enhancement-with-task-oriented-flow | 1711.09078 | null | https://arxiv.org/abs/1711.09078v3 | https://arxiv.org/pdf/1711.09078v3.pdf | Video Enhancement with Task-Oriented Flow | Many video enhancement algorithms rely on optical flow to register frames in a video sequence. Precise flow estimation is however intractable; and optical flow itself is often a sub-optimal representation for particular video processing tasks. In this paper, we propose task-oriented flow (TOFlow), a motion representation learned in a self-supervised, task-specific manner. We design a neural network with a trainable motion estimation component and a video processing component, and train them jointly to learn the task-oriented flow. For evaluation, we build Vimeo-90K, a large-scale, high-quality video dataset for low-level video processing. TOFlow outperforms traditional optical flow on standard benchmarks as well as our Vimeo-90K dataset in three video processing tasks: frame interpolation, video denoising/deblocking, and video super-resolution. | ['Baian Chen', 'Jiajun Wu', 'Donglai Wei', 'William T. Freeman', 'Tianfan Xue'] | 2017-11-24 | null | null | null | null | ['video-denoising', 'video-enhancement'] | ['computer-vision', 'computer-vision'] | [ 2.70924568e-01 -7.42396533e-01 -4.65695500e-01 -3.31875473e-01
-5.36860824e-01 -3.15096378e-01 3.24930012e-01 -5.15404403e-01
-5.00306606e-01 7.61642814e-01 5.21088481e-01 -5.99751174e-02
1.39607102e-01 -4.55567390e-01 -8.33572984e-01 -5.42038560e-01
-4.73899871e-01 -1.36771634e-01 2.48449355e-01 8.44021067e-02
3.50775570e-01 2.45004207e-01 -1.30377936e+00 6.62748992e-01
5.93964219e-01 1.14676154e+00 3.06843191e-01 1.32268047e+00
1.17502116e-01 1.54283333e+00 -2.90554255e-01 -1.27379030e-01
4.25180703e-01 -3.86701018e-01 -1.08970904e+00 2.97066718e-01
1.23436487e+00 -1.06328273e+00 -7.78641164e-01 9.62590098e-01
3.43051553e-01 3.73418897e-01 4.23762769e-01 -1.20924103e+00
-6.35203660e-01 3.53345692e-01 -5.85290909e-01 9.14468050e-01
4.77867097e-01 5.74064851e-01 8.63461077e-01 -8.54083836e-01
8.79564047e-01 1.34738839e+00 6.85080051e-01 7.50763357e-01
-1.30300117e+00 -5.75902045e-01 -8.42665285e-02 4.27520543e-01
-1.00850844e+00 -9.81889307e-01 7.08344817e-01 -5.79290748e-01
9.20226157e-01 -1.28180787e-01 5.06992340e-01 1.09416890e+00
4.95274454e-01 7.73621500e-01 7.09130764e-01 1.03831537e-01
4.80975062e-02 -7.23717213e-01 -4.60164621e-02 7.99369872e-01
2.41841935e-02 3.77216727e-01 -9.17601049e-01 2.81719744e-01
1.33149552e+00 -1.53268665e-01 -7.70150840e-01 -2.84894817e-02
-1.37741244e+00 3.61881703e-01 3.52312177e-01 -3.62092555e-02
-4.41007674e-01 6.95863903e-01 6.73186183e-01 6.22917652e-01
5.19075453e-01 2.16119155e-01 -4.24951881e-01 -4.29400682e-01
-1.37163043e+00 4.01606202e-01 5.30027688e-01 8.82220745e-01
1.00232875e+00 5.32845140e-01 -4.23367441e-01 6.52735591e-01
2.45205343e-01 2.09393755e-01 5.99694669e-01 -1.89193904e+00
6.41450405e-01 -1.26650110e-01 3.13765854e-01 -1.01976442e+00
-1.01905912e-01 1.54284105e-01 -1.05393457e+00 4.04410779e-01
5.16986132e-01 -2.55489022e-01 -8.92034471e-01 1.62545455e+00
-1.57180130e-02 9.25742269e-01 6.13161214e-02 1.23888648e+00
9.17264879e-01 9.93193686e-01 1.31142482e-01 -2.83015132e-01
9.28366423e-01 -1.50276601e+00 -8.33058178e-01 -2.28731945e-01
5.37478268e-01 -7.14930773e-01 6.81817710e-01 5.50607562e-01
-1.56651425e+00 -8.61800730e-01 -9.76165414e-01 -5.14758408e-01
3.55184734e-01 -1.95691824e-01 6.86361432e-01 3.14586163e-01
-1.40112317e+00 9.33163762e-01 -9.23224747e-01 1.61613539e-01
7.07376003e-01 2.41705894e-01 -6.15058482e-01 -3.22791100e-01
-1.00505042e+00 3.32901806e-01 5.04424214e-01 1.42991543e-01
-1.32347357e+00 -1.09370995e+00 -1.12599576e+00 2.99077854e-02
1.47493929e-01 -1.06564689e+00 1.07126689e+00 -1.36226976e+00
-1.50405765e+00 6.80527329e-01 -3.89951319e-01 -6.70930028e-01
5.84858000e-01 -3.79139751e-01 -4.52049047e-01 6.13946557e-01
1.02791131e-01 1.13196051e+00 1.58486211e+00 -9.32494342e-01
-9.06164467e-01 1.64528400e-01 1.09571151e-01 -8.89190100e-03
-2.16652900e-02 1.73317075e-01 -5.59912264e-01 -1.07791460e+00
-3.53264064e-01 -4.96468693e-01 -2.24451721e-01 3.97578686e-01
6.95178732e-02 -7.00185774e-03 1.01219535e+00 -9.82225299e-01
1.27484822e+00 -2.14943123e+00 2.08232835e-01 -4.97308552e-01
5.20282567e-01 5.21722019e-01 -4.75832373e-01 -2.61197448e-01
-2.14918122e-01 -8.39219019e-02 -3.04886490e-01 -5.23340046e-01
-6.12091839e-01 2.14050859e-01 -7.82073215e-02 3.47710192e-01
5.95724106e-01 9.81452823e-01 -1.27022922e+00 -6.52435958e-01
4.67437029e-01 6.42488837e-01 -1.03595674e+00 3.68738472e-01
-1.59204975e-01 4.83164966e-01 -9.38621685e-02 5.46056509e-01
7.29656219e-01 -2.79089063e-01 -5.31523377e-02 -5.39391279e-01
5.50178885e-02 4.99661602e-02 -1.14725053e+00 2.06651974e+00
-5.22173107e-01 1.25986505e+00 3.32438201e-01 -7.55978107e-01
4.74542081e-01 3.85656029e-01 8.98686588e-01 -5.62740386e-01
5.32225147e-02 -3.39621231e-02 -3.08712244e-01 -7.34527171e-01
9.08197820e-01 2.64159769e-01 5.96801579e-01 3.81048709e-01
5.17372251e-01 2.64256269e-01 6.44975901e-01 2.50307024e-01
1.33383942e+00 3.57336640e-01 6.48492277e-02 -3.24909657e-01
7.83350587e-01 -3.82894188e-01 9.34818864e-01 6.60214305e-01
-6.93047941e-01 1.00270021e+00 3.71874064e-01 -8.07526886e-01
-1.11658013e+00 -1.05560219e+00 1.21678822e-01 1.01596308e+00
2.34754533e-01 -5.75103819e-01 -7.26107538e-01 -6.78832233e-01
-1.17507473e-01 3.52058783e-02 -4.30540353e-01 -3.92068960e-02
-1.08519256e+00 -2.97608048e-01 3.07772964e-01 4.86511141e-01
6.23586774e-01 -1.19833064e+00 -5.31901300e-01 5.73343635e-01
-4.21120346e-01 -1.71147251e+00 -1.15756381e+00 -2.60546267e-01
-1.05283284e+00 -1.23970079e+00 -8.71000230e-01 -9.39672112e-01
4.12912458e-01 6.41947329e-01 1.53455138e+00 1.76077589e-01
-1.12046294e-01 3.25896204e-01 -1.73794627e-01 4.51289266e-01
-6.00275636e-01 -9.23343077e-02 -1.27185911e-01 2.89704442e-01
-1.41969966e-02 -6.58118248e-01 -9.32872891e-01 1.52346686e-01
-1.17618477e+00 3.10720764e-02 1.17232308e-01 9.10792768e-01
3.65273327e-01 -2.17345536e-01 2.83815593e-01 -6.64531946e-01
2.26521820e-01 -2.57922173e-01 -5.65487564e-01 -2.64487509e-02
-1.58989921e-01 7.34134167e-02 6.82004213e-01 -2.68522650e-01
-1.10381567e+00 5.79769090e-02 -6.84174746e-02 -1.03625882e+00
-4.33598869e-02 1.37052247e-02 3.72366644e-02 -2.97060430e-01
5.46574056e-01 2.42269367e-01 -3.77875799e-03 -2.44811922e-01
1.59746394e-01 3.61381531e-01 1.02328491e+00 -4.42453474e-01
7.78501272e-01 6.25248790e-01 3.51375304e-02 -8.60271275e-01
-7.51984715e-01 -4.66074407e-01 -6.15563214e-01 -3.75451595e-01
1.04030573e+00 -1.18217385e+00 -7.41046548e-01 7.99855471e-01
-1.39516437e+00 -7.69980967e-01 -2.47967049e-01 5.95995188e-01
-7.75248289e-01 6.06154680e-01 -9.99676406e-01 -1.19851716e-01
-5.46260059e-01 -1.43684995e+00 1.07282507e+00 1.18632160e-01
7.45451823e-02 -1.23222649e+00 6.04013428e-02 4.36911702e-01
4.10638630e-01 1.20261967e-01 2.33593181e-01 4.92359936e-01
-9.55180109e-01 4.40691739e-01 -5.49586713e-01 8.24158430e-01
3.75440270e-01 2.40786120e-01 -8.80703270e-01 -5.20806909e-01
-1.96004838e-01 -5.57349741e-01 1.35895002e+00 7.28950977e-01
1.44386160e+00 -2.61541516e-01 2.05419809e-01 1.41050172e+00
1.38225520e+00 -1.18932776e-01 9.46491897e-01 1.60483718e-01
1.03843009e+00 3.39162230e-01 4.01259184e-01 3.65891069e-01
3.94208968e-01 5.99004686e-01 3.53338838e-01 2.24875316e-01
-6.77295327e-01 -1.90890376e-02 7.68568456e-01 8.66757512e-01
-4.35034335e-01 -1.53665483e-01 -5.11828542e-01 5.00339568e-01
-1.70619226e+00 -1.34175575e+00 -9.50712990e-03 1.80639291e+00
9.70515251e-01 4.54436708e-03 -2.32568365e-02 -3.43836583e-02
6.03327572e-01 7.90915787e-01 -3.54131073e-01 -2.27554604e-01
-2.16520190e-01 1.39415637e-01 5.26027679e-01 9.01979506e-01
-1.45044827e+00 1.09115469e+00 6.57065964e+00 6.50617242e-01
-1.29017353e+00 1.52439162e-01 7.41716862e-01 -2.10187748e-01
-2.36103460e-02 -2.27116749e-01 -5.91020823e-01 6.74790025e-01
8.51303995e-01 -2.16867551e-02 6.07142031e-01 4.20479000e-01
6.13524795e-01 -5.19625004e-03 -1.22091126e+00 1.44895399e+00
2.53423620e-02 -1.90663946e+00 1.49318039e-01 -1.85681432e-01
9.62122023e-01 1.22129194e-01 -1.25262424e-01 -5.34772389e-02
1.03543706e-01 -1.02066970e+00 6.15361691e-01 4.29825544e-01
9.29107845e-01 -3.30606639e-01 3.74694198e-01 -2.22047001e-01
-1.41395998e+00 -4.07864451e-02 -3.13191950e-01 1.23999584e-02
6.11963034e-01 4.45530266e-01 9.33885016e-03 4.68905002e-01
1.00799549e+00 1.65565133e+00 -4.45019811e-01 1.01307762e+00
-2.71844165e-03 6.95036232e-01 1.18024461e-01 9.36828434e-01
3.92072588e-01 -1.39621839e-01 5.32866657e-01 1.32468891e+00
8.45749974e-02 -8.67049471e-02 1.32595062e-01 6.01038873e-01
-4.85326022e-01 -4.00281310e-01 -2.53023982e-01 1.44988850e-01
7.59405047e-02 1.24350297e+00 -4.64596570e-01 -5.35871923e-01
-6.34573102e-01 1.20320952e+00 1.64197311e-01 6.24896824e-01
-7.38781691e-01 -7.16172010e-02 1.25071192e+00 -2.24859975e-02
2.90219814e-01 -3.61477792e-01 2.87666202e-01 -1.64944780e+00
-1.94453087e-03 -8.83976936e-01 2.73425728e-01 -8.51038158e-01
-1.08741224e+00 4.44622070e-01 -3.56326401e-01 -1.33941329e+00
-5.04933715e-01 -6.92693532e-01 -5.32584906e-01 5.81529915e-01
-1.95601022e+00 -7.03170717e-01 -6.09383404e-01 9.09195840e-01
9.57803726e-01 -1.02582812e-01 1.97815731e-01 8.40391278e-01
-5.82076967e-01 4.37528938e-01 1.20068807e-02 5.49505651e-01
1.10239565e+00 -1.09179318e+00 6.05477571e-01 1.09488940e+00
1.10326007e-01 1.03444979e-01 4.98325855e-01 -3.98887247e-01
-1.59443033e+00 -1.46752036e+00 5.47352195e-01 -2.05724925e-01
5.45136571e-01 -2.43955757e-02 -1.11067402e+00 8.21696639e-01
3.91913295e-01 9.77221549e-01 5.66220954e-02 -6.81165457e-01
-3.51263046e-01 -3.02052438e-01 -8.62901092e-01 4.22160298e-01
1.29927897e+00 -5.70503116e-01 -2.73630232e-01 1.30045399e-01
7.85470068e-01 -4.53890890e-01 -1.06921268e+00 2.00933158e-01
5.11079073e-01 -1.16154027e+00 1.30503976e+00 -6.55672073e-01
1.07704997e+00 -3.75907689e-01 -2.34375112e-02 -1.18128085e+00
-3.81002486e-01 -1.10165250e+00 -6.36400342e-01 1.05318224e+00
-2.59707093e-01 -9.23877954e-02 1.10548973e+00 2.08831966e-01
-1.39162883e-01 -2.81797677e-01 -7.09174454e-01 -6.34247601e-01
-1.58885121e-01 -5.10867417e-01 2.51205415e-01 1.11374927e+00
-6.18430853e-01 1.18998913e-02 -9.15444195e-01 -6.32771105e-02
9.07181382e-01 -2.81779468e-01 7.91382909e-01 -7.14501441e-01
-4.10649478e-01 -4.47497696e-01 -6.40950918e-01 -1.60380173e+00
5.37476003e-01 -5.37246287e-01 8.00108463e-02 -1.12910628e+00
-1.00595228e-01 3.40413712e-02 -2.87508518e-01 6.04652502e-02
-3.60035241e-01 4.60034400e-01 3.13665718e-01 2.43887976e-01
-7.80367196e-01 3.87468457e-01 1.58128786e+00 -3.64350736e-01
-2.91777492e-01 -2.94035852e-01 -1.33856207e-01 7.50991464e-01
4.75313574e-01 -1.31401762e-01 -3.72206062e-01 -9.35784340e-01
-1.02571286e-01 5.70423782e-01 4.75151598e-01 -1.08209085e+00
2.17338547e-01 -2.65243620e-01 4.91474092e-01 -2.99081802e-01
2.57581383e-01 -5.79446137e-01 -1.56858921e-01 3.87798190e-01
-4.71162081e-01 2.06343368e-01 -1.13304406e-01 4.58028048e-01
-5.85580170e-01 -1.28051490e-01 9.59851384e-01 -1.45294547e-01
-1.42397344e+00 8.17299545e-01 -4.78194386e-01 5.08674920e-01
5.82459986e-01 -3.11446756e-01 -3.28450054e-01 -4.44117188e-01
-4.86416161e-01 1.85272709e-01 3.19523662e-01 5.87213814e-01
9.51883197e-01 -1.32251632e+00 -8.89631748e-01 2.72839367e-01
-2.40995988e-01 2.52901781e-02 4.09009516e-01 6.05908096e-01
-1.10616899e+00 1.49965897e-01 -7.08279312e-01 -6.61982715e-01
-1.06099319e+00 5.17170966e-01 4.52467054e-01 -9.70475748e-02
-7.68403530e-01 7.77718067e-01 2.38571361e-01 1.58240676e-01
1.66363582e-01 -4.65144068e-01 -1.81991413e-01 -1.68778330e-01
1.06125951e+00 6.65165186e-01 -2.47006431e-01 -9.10348773e-01
-1.56006560e-01 6.30440474e-01 3.75264995e-02 1.55115888e-01
1.06144726e+00 -4.28062141e-01 -1.10207886e-01 -1.28951957e-02
1.53463423e+00 -5.37714601e-01 -2.09914851e+00 -2.76194036e-01
-4.89590652e-02 -9.05229867e-01 3.97070378e-01 -4.82083410e-02
-1.75050414e+00 9.53402042e-01 4.45472836e-01 -2.06184283e-01
1.36654067e+00 -5.50597727e-01 1.13921571e+00 3.12849015e-01
2.45289803e-01 -8.70125890e-01 2.53977090e-01 5.91653287e-01
6.57681882e-01 -1.53255796e+00 5.71846999e-02 -3.92426938e-01
-3.43852788e-01 1.28339553e+00 6.31074786e-01 -3.72532099e-01
6.43582106e-01 3.26158226e-01 7.45114163e-02 4.28105623e-01
-1.10818088e+00 5.28395101e-02 3.61939579e-01 6.21804893e-01
5.84833741e-01 -5.04364491e-01 2.15158403e-01 -1.65664688e-01
2.14821786e-01 6.71018362e-01 7.12221980e-01 7.61579931e-01
-3.02464604e-01 -8.55485499e-01 -2.57046431e-01 2.94799924e-01
-6.66290045e-01 -1.89870477e-01 2.91424394e-01 3.96426618e-01
-8.53500366e-02 8.48625422e-01 3.33796769e-01 -2.33010992e-01
-1.39795214e-01 -4.58802342e-01 6.58015192e-01 -1.11560613e-01
-5.93733788e-01 -1.27897292e-01 -4.53126105e-03 -1.30679882e+00
-8.88791680e-01 -4.42596167e-01 -9.62762892e-01 -6.86459422e-01
4.39104348e-01 -2.02889025e-01 1.09617099e-01 7.56815493e-01
2.90605545e-01 6.33298993e-01 6.15441322e-01 -1.30081868e+00
-1.42187506e-01 -5.17387092e-01 -3.60588521e-01 8.13928306e-01
9.77235734e-01 -3.66459906e-01 -2.90439010e-01 7.32951581e-01] | [10.719393730163574, -1.4765853881835938] |
dd37a252-3bc1-4103-9eb1-a3f4e43a9934 | action-unit-memory-network-for-weakly | 2104.14135 | null | https://arxiv.org/abs/2104.14135v1 | https://arxiv.org/pdf/2104.14135v1.pdf | Action Unit Memory Network for Weakly Supervised Temporal Action Localization | Weakly supervised temporal action localization aims to detect and localize actions in untrimmed videos with only video-level labels during training. However, without frame-level annotations, it is challenging to achieve localization completeness and relieve background interference. In this paper, we present an Action Unit Memory Network (AUMN) for weakly supervised temporal action localization, which can mitigate the above two challenges by learning an action unit memory bank. In the proposed AUMN, two attention modules are designed to update the memory bank adaptively and learn action units specific classifiers. Furthermore, three effective mechanisms (diversity, homogeneity and sparsity) are designed to guide the updating of the memory network. To the best of our knowledge, this is the first work to explicitly model the action units with a memory network. Extensive experimental results on two standard benchmarks (THUMOS14 and ActivityNet) demonstrate that our AUMN performs favorably against state-of-the-art methods. Specifically, the average mAP of IoU thresholds from 0.1 to 0.5 on the THUMOS14 dataset is significantly improved from 47.0% to 52.1%. | ['Yongdong Zhang', 'Feng Wu', 'Tao Mei', 'Jingen Liu', 'Wenfei Yang', 'Tianzhu Zhang', 'Wang Luo'] | 2021-04-29 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Luo_Action_Unit_Memory_Network_for_Weakly_Supervised_Temporal_Action_Localization_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Luo_Action_Unit_Memory_Network_for_Weakly_Supervised_Temporal_Action_Localization_CVPR_2021_paper.pdf | cvpr-2021-1 | ['weakly-supervised-action-localization', 'weakly-supervised-temporal-action'] | ['computer-vision', 'computer-vision'] | [ 2.69967288e-01 -4.71236520e-02 -8.79039049e-01 -1.42467201e-01
-5.05551040e-01 -2.21640036e-01 3.84098262e-01 -3.61475945e-01
-5.71683347e-01 7.36238718e-01 2.95736670e-01 1.38847262e-01
2.80750781e-01 -2.94767886e-01 -9.02234316e-01 -7.94905126e-01
-3.59720886e-01 -1.66940108e-01 7.98420787e-01 2.78734237e-01
1.19733557e-01 1.57054454e-01 -1.32406306e+00 7.14797795e-01
5.58605134e-01 1.26698637e+00 2.90481627e-01 4.49993372e-01
3.43929619e-01 1.78016388e+00 -4.35888022e-01 2.88654655e-01
1.02390073e-01 -5.83060086e-01 -9.06842232e-01 2.12226704e-01
3.96961689e-01 -7.68825471e-01 -7.94329286e-01 8.65933597e-01
3.91269058e-01 4.36840296e-01 2.82913953e-01 -1.11667931e+00
-6.46604955e-01 7.41977692e-01 -5.47385871e-01 5.40082335e-01
2.04441369e-01 3.52719873e-01 9.68656182e-01 -8.73934865e-01
5.79970121e-01 1.03718805e+00 3.39147359e-01 7.15327203e-01
-8.98830175e-01 -6.55252039e-01 6.95597887e-01 6.28719628e-01
-1.43924451e+00 -5.56029916e-01 4.69148517e-01 -3.08366656e-01
1.11459136e+00 -7.38543719e-02 4.69892859e-01 1.30244076e+00
9.85977575e-02 1.26894081e+00 7.89744854e-01 -1.34943753e-01
2.97742397e-01 -3.82557124e-01 2.06149407e-02 9.23551321e-01
-9.59873796e-02 -1.56491950e-01 -1.03060937e+00 2.76772559e-01
1.13501716e+00 2.49053076e-01 -3.53522688e-01 -3.09811532e-01
-1.46784949e+00 4.17304426e-01 5.40749371e-01 4.77662176e-01
-3.31586897e-01 7.61256993e-01 5.61243057e-01 8.43525678e-03
4.17419374e-01 1.07990369e-01 -5.05264461e-01 -4.92298603e-01
-7.35831797e-01 -4.61240858e-01 4.11904454e-01 9.94268060e-01
5.45370579e-01 1.59151271e-01 -4.52362746e-01 5.40376961e-01
-4.29369211e-02 2.50739872e-01 6.53279781e-01 -1.26813829e+00
3.68110836e-01 6.01050138e-01 1.66223824e-01 -7.80242860e-01
-3.13574702e-01 -3.51507187e-01 -6.86953783e-01 -2.04714835e-01
1.73634738e-01 -1.74546912e-02 -9.05274510e-01 1.85701847e+00
8.42949152e-02 9.80383515e-01 -1.81750357e-01 1.04051030e+00
8.90577555e-01 5.61678588e-01 3.09219509e-01 -2.93539852e-01
8.29730153e-01 -1.59262311e+00 -8.55722845e-01 -4.47485447e-01
9.37810957e-01 -2.78745383e-01 1.09399128e+00 1.13578461e-01
-1.06663013e+00 -8.27992558e-01 -1.01871002e+00 7.26492703e-02
2.24605780e-02 5.40980875e-01 6.91815734e-01 6.75641671e-02
-8.39333177e-01 5.07716537e-01 -1.43584776e+00 -3.42804641e-01
7.13904977e-01 3.47578347e-01 -3.92916799e-01 6.69544889e-03
-1.08066308e+00 5.69078088e-01 5.61283350e-01 9.52308401e-02
-1.36901474e+00 -5.22529185e-01 -8.44774783e-01 6.85948953e-02
7.54718542e-01 -2.30576456e-01 1.41208756e+00 -1.26829553e+00
-1.53764272e+00 4.92892712e-01 -2.01079294e-01 -8.15684557e-01
4.12075967e-01 -4.38014805e-01 -2.62993991e-01 4.87890273e-01
1.14619024e-01 8.92850697e-01 7.57331669e-01 -7.62237728e-01
-8.22795570e-01 3.76042910e-02 3.70973706e-01 1.03115886e-01
-6.97040737e-01 -4.25195396e-02 -8.54489446e-01 -8.20241332e-01
7.08577931e-02 -8.54096174e-01 -2.26993576e-01 1.02394454e-01
5.21869138e-02 -1.71584085e-01 9.24082160e-01 -4.90648299e-01
1.53763938e+00 -2.28843760e+00 1.42947257e-01 -2.44701669e-01
1.74443260e-01 1.86657995e-01 -2.84286588e-01 -1.09896630e-01
4.75608855e-02 -1.26933277e-01 -3.02679669e-02 -3.55197549e-01
-6.36356473e-02 4.55273271e-01 -9.79953259e-02 6.92091942e-01
2.84074489e-02 1.07867420e+00 -1.08182430e+00 -5.74013889e-01
2.10238203e-01 3.49699020e-01 -6.49220467e-01 1.23409264e-01
-2.32727870e-01 5.42118371e-01 -3.28831553e-01 1.03049648e+00
1.29505873e-01 -5.40826142e-01 2.62161463e-01 -3.54409486e-01
-2.63437536e-02 3.27804476e-01 -1.03039873e+00 2.05705810e+00
-2.08607808e-01 6.62746072e-01 -2.10397691e-01 -1.05252874e+00
3.82808805e-01 3.16967517e-01 9.83164132e-01 -8.95621657e-01
2.31346756e-01 -5.20456582e-02 4.95282821e-02 -3.65061015e-01
2.88632452e-01 3.80499482e-01 -2.13763025e-02 4.47813958e-01
4.77762878e-01 8.65511835e-01 4.04364318e-01 1.23286784e-01
1.48926377e+00 2.94414639e-01 2.22027183e-01 -4.79146242e-02
6.02898359e-01 -3.31897706e-01 9.85628784e-01 8.01017880e-01
-7.43992090e-01 4.27095860e-01 4.82290536e-01 -6.13508582e-01
-4.28841144e-01 -9.01511788e-01 1.87660351e-01 1.52086842e+00
3.62459004e-01 -6.44237876e-01 -7.66811848e-01 -1.01652861e+00
-3.92189771e-01 2.56860346e-01 -8.26699018e-01 -5.17522693e-01
-8.31329346e-01 -4.97426838e-01 5.95738709e-01 1.17868769e+00
1.06083977e+00 -1.21069360e+00 -7.28819311e-01 1.55701995e-01
-4.04870510e-01 -1.51217747e+00 -8.07500839e-01 1.04501687e-01
-9.03275549e-01 -1.11466384e+00 -5.33779860e-01 -9.12367225e-01
8.26129079e-01 4.37491208e-01 9.00011241e-01 -4.12795879e-02
-3.11780125e-02 4.65010732e-01 -5.83010793e-01 1.41924754e-01
8.79471079e-02 1.53755769e-01 1.95820406e-01 2.92075396e-01
3.34004730e-01 -5.44677615e-01 -7.67118752e-01 7.30336905e-01
-8.06866884e-01 7.34139010e-02 8.30195129e-01 7.45208144e-01
8.78047466e-01 -3.66926789e-02 3.72751504e-01 -5.61033487e-01
-1.10629432e-01 -3.83355618e-01 -4.24085915e-01 2.02639371e-01
-2.49735117e-01 -1.78175375e-01 3.14689100e-01 -9.38951552e-01
-9.35474396e-01 4.60503906e-01 3.20236146e-01 -7.61997640e-01
-2.51844134e-02 3.27392370e-01 -1.23971477e-01 -2.34507680e-01
4.55941409e-01 4.33597744e-01 -3.70897382e-01 -3.09181571e-01
3.31736326e-01 2.09228545e-01 6.80820167e-01 -3.43714505e-01
3.01997721e-01 6.84630215e-01 -2.09172457e-01 -4.58084047e-01
-1.18017375e+00 -4.84278262e-01 -8.62946153e-01 -4.08571780e-01
1.00334072e+00 -1.13681805e+00 -6.60764873e-01 5.05743921e-01
-8.93456697e-01 -9.03252006e-01 -2.73985952e-01 5.58309257e-01
-7.29052663e-01 3.46198469e-01 -9.11918044e-01 -4.29379225e-01
-1.02966435e-01 -1.02205074e+00 9.71636474e-01 5.84434457e-02
-1.42010570e-01 -8.32744837e-01 -1.70497417e-01 4.44425404e-01
1.71422705e-01 1.42760888e-01 1.65396824e-01 -2.60338515e-01
-8.18017483e-01 -3.54013219e-02 -1.64799616e-01 4.77341712e-01
2.94347644e-01 -3.61387610e-01 -7.62965143e-01 -3.44086379e-01
-1.26851737e-01 -4.66696262e-01 1.22803545e+00 4.62181330e-01
1.51724422e+00 -3.68116796e-01 -3.25467616e-01 5.56867719e-01
8.95176828e-01 2.77099669e-01 6.42283678e-01 2.52397776e-01
8.21995080e-01 -2.87771337e-02 8.94448042e-01 5.98700762e-01
2.45212704e-01 6.68890297e-01 5.83842814e-01 4.20380421e-02
-2.23801970e-01 -2.70344585e-01 8.83859932e-01 5.85405529e-01
-2.31939197e-01 -1.31825849e-01 -6.55437171e-01 4.60713178e-01
-2.31879163e+00 -1.19219887e+00 2.18384460e-01 1.89743376e+00
8.09287608e-01 4.45950150e-01 1.41271353e-01 -1.10469349e-01
5.27117133e-01 5.62128305e-01 -6.70000613e-01 3.31331909e-01
-1.33349612e-01 -4.74978201e-02 6.25781417e-01 3.71266007e-01
-1.66917396e+00 1.23570693e+00 6.03579235e+00 7.78117776e-01
-9.66914773e-01 4.48466033e-01 5.55194914e-01 -5.34021616e-01
6.15002751e-01 -6.10851236e-02 -7.39712119e-01 6.55850172e-01
8.28461111e-01 1.69284031e-01 3.14127117e-01 9.67650950e-01
4.02429402e-01 -2.19464034e-01 -1.23101676e+00 1.10678422e+00
2.84034461e-01 -1.30698347e+00 -7.56035447e-02 -1.69235721e-01
9.43514943e-01 7.37241358e-02 -8.57493933e-03 5.04447341e-01
4.06127274e-02 -8.76367331e-01 8.12551081e-01 5.67817330e-01
5.89405656e-01 -6.29794002e-01 6.93156660e-01 2.45740771e-01
-1.57039070e+00 -3.60751420e-01 -2.82174468e-01 -2.22344011e-01
2.16321737e-01 1.49013072e-01 -3.52074623e-01 -4.32055295e-02
9.44284976e-01 1.41823304e+00 -5.87357044e-01 7.66235888e-01
-4.64643776e-01 8.35164726e-01 -5.99951521e-02 1.41193539e-01
5.59820950e-01 3.70308489e-01 1.97060660e-01 1.19102407e+00
4.14230600e-02 1.97834730e-01 6.64057136e-01 5.11440277e-01
-3.52061719e-01 -2.14850873e-01 -2.69046992e-01 -1.88697681e-01
1.96066767e-01 1.02521563e+00 -6.11033976e-01 -3.93699378e-01
-5.72007895e-01 1.30072224e+00 3.62198055e-01 4.47409689e-01
-1.38671422e+00 5.70076928e-02 6.74521685e-01 1.53025582e-01
4.33398783e-01 -4.02402461e-01 3.03802043e-02 -1.36967862e+00
3.98284383e-02 -7.12560296e-01 6.10548913e-01 -6.97357297e-01
-7.21293926e-01 2.89544880e-01 -2.46819600e-01 -1.39553416e+00
1.53823551e-02 -4.19753283e-01 -3.55306745e-01 4.62434664e-02
-1.17200577e+00 -1.12953162e+00 -4.59125310e-01 6.73386633e-01
8.74202371e-01 -1.18081897e-01 5.02956867e-01 5.87502241e-01
-9.09055889e-01 5.95434904e-01 -2.89140046e-01 5.08743346e-01
7.54363060e-01 -9.49840486e-01 -5.15150242e-02 1.06956863e+00
2.76946574e-01 5.11819243e-01 1.92246586e-01 -5.40071607e-01
-1.28994370e+00 -1.46036959e+00 4.69211012e-01 -3.75634760e-01
8.99594843e-01 -4.33079928e-01 -8.63586843e-01 1.04831862e+00
9.68545824e-02 6.25416756e-01 5.94898582e-01 -2.94020087e-01
-2.23367721e-01 -1.83363363e-01 -6.15307868e-01 5.71380496e-01
1.53289056e+00 -5.35832226e-01 -2.85770178e-01 5.33430398e-01
7.63786197e-01 -5.81805646e-01 -7.97056735e-01 5.05785227e-01
4.76716936e-01 -8.36554229e-01 8.80842209e-01 -7.23482490e-01
2.44439408e-01 -5.73736787e-01 -2.78754532e-01 -7.39050388e-01
-5.22553742e-01 -5.34595788e-01 -9.53481674e-01 1.05796456e+00
1.88682973e-01 -2.25776702e-01 9.37518477e-01 2.88807869e-01
-2.14118823e-01 -8.50128174e-01 -1.04807138e+00 -9.75734890e-01
-5.04012227e-01 -3.97825152e-01 1.25782723e-02 7.74863720e-01
3.34670581e-02 2.35112548e-01 -8.18131447e-01 1.81732476e-01
3.10738117e-01 3.21528362e-03 5.94568193e-01 -5.46557784e-01
-4.35659945e-01 -2.77394950e-01 -5.89886248e-01 -1.59912407e+00
4.67425227e-01 -5.72281957e-01 1.45381853e-01 -1.33087373e+00
3.76720876e-01 7.22470731e-02 -8.76961887e-01 9.17279899e-01
-5.72132356e-02 3.09740782e-01 4.83284853e-02 1.55779526e-01
-1.54862320e+00 6.92786932e-01 1.09403586e+00 -3.09953541e-01
-1.77503541e-01 -2.01920360e-01 -3.90728444e-01 9.86189187e-01
9.37400579e-01 -3.95569175e-01 -6.14709914e-01 -6.67942643e-01
-3.09109241e-01 -2.77532399e-01 4.08717215e-01 -1.29486740e+00
5.39667368e-01 -3.75136852e-01 3.54667306e-01 -5.57220578e-01
3.03430289e-01 -6.51930094e-01 -2.82856375e-01 6.30054653e-01
-4.38448340e-01 -2.09759399e-01 1.05556436e-01 7.73612380e-01
-3.27745557e-01 2.22587481e-01 8.59093726e-01 -8.58425274e-02
-1.48157978e+00 5.73543072e-01 -4.71880198e-01 2.31677089e-02
1.23769271e+00 -4.99503911e-02 -2.88400263e-01 -2.99852312e-01
-7.94971526e-01 4.19075578e-01 1.88960940e-01 5.06215990e-01
6.80354178e-01 -1.62159300e+00 -3.28216553e-01 -5.97375110e-02
1.62454754e-01 -2.15774164e-01 3.97809237e-01 1.23313391e+00
-2.20006168e-01 4.87095535e-01 -1.50169104e-01 -6.68614805e-01
-1.18623364e+00 5.78333497e-01 5.34731448e-01 -2.37482280e-01
-5.11052728e-01 8.97666574e-01 3.94320339e-01 9.73088592e-02
7.86035538e-01 -5.68990409e-01 -8.62687901e-02 -5.15994690e-02
8.57367516e-01 5.12454510e-01 -2.45228782e-01 -6.70952082e-01
-5.25877893e-01 2.98207313e-01 1.53207150e-03 1.91564217e-01
1.25465596e+00 1.33023234e-02 2.89547984e-02 3.51433903e-01
1.01777518e+00 -4.90112096e-01 -1.85555530e+00 -4.02446866e-01
-2.68095247e-02 -4.61327344e-01 8.92942473e-02 -6.19337261e-01
-1.42640972e+00 6.89534128e-01 5.92921138e-01 -3.45784187e-01
1.38325262e+00 9.77475643e-02 8.76968861e-01 6.59516335e-01
3.86642039e-01 -1.39055324e+00 7.90146530e-01 6.60388589e-01
8.04457366e-01 -1.27577400e+00 -1.70750707e-01 -2.30322242e-01
-5.52372158e-01 7.48741388e-01 1.25517941e+00 -9.59566981e-02
5.24513543e-01 2.64652044e-01 -2.73392558e-01 5.95688596e-02
-9.50833201e-01 -3.57303292e-01 3.26243192e-01 2.44359553e-01
3.08785766e-01 -1.92266688e-01 -2.86231041e-01 7.33231843e-01
7.07724273e-01 2.29873225e-01 2.44525149e-01 1.24779963e+00
-6.75138652e-01 -7.59541512e-01 1.96396876e-02 2.75232017e-01
-4.64058697e-01 -1.78365912e-02 -4.45225745e-01 6.25511944e-01
2.37628981e-01 8.55699360e-01 -4.37075226e-03 -5.26886463e-01
2.76276737e-01 -2.23616913e-01 4.29139048e-01 -5.43937802e-01
-2.44891912e-01 2.42093056e-01 -3.66761573e-02 -1.25074124e+00
-8.86709988e-01 -5.55743814e-01 -1.56298530e+00 -1.00245334e-01
-3.21786031e-02 -1.72688603e-01 -1.25532910e-01 8.48811507e-01
4.59070832e-01 7.64050782e-01 3.61338556e-01 -8.48829508e-01
-2.64341980e-01 -9.67782438e-01 -5.20140886e-01 2.55769879e-01
1.75533250e-01 -9.44219530e-01 -2.46710241e-01 3.27011526e-01] | [8.411396980285645, 0.536492645740509] |
012c3997-5381-42ab-ba5f-47b00df5a381 | generalizable-re-identification-from-videos | 2211.03663 | null | https://arxiv.org/abs/2211.03663v2 | https://arxiv.org/pdf/2211.03663v2.pdf | Generalizable Re-Identification from Videos with Cycle Association | In this paper, we are interested in learning a generalizable person re-identification (re-ID) representation from unlabeled videos. Compared with 1) the popular unsupervised re-ID setting where the training and test sets are typically under the same domain, and 2) the popular domain generalization (DG) re-ID setting where the training samples are labeled, our novel scenario combines their key challenges: the training samples are unlabeled, and collected form various domains which do no align with the test domain. In other words, we aim to learn a representation in an unsupervised manner and directly use the learned representation for re-ID in novel domains. To fulfill this goal, we make two main contributions: First, we propose Cycle Association (CycAs), a scalable self-supervised learning method for re-ID with low training complexity; and second, we construct a large-scale unlabeled re-ID dataset named LMP-video, tailored for the proposed method. Specifically, CycAs learns re-ID features by enforcing cycle consistency of instance association between temporally successive video frame pairs, and the training cost is merely linear to the data size, making large-scale training possible. On the other hand, the LMP-video dataset is extremely large, containing 50 million unlabeled person images cropped from over 10K Youtube videos, therefore is sufficient to serve as fertile soil for self-supervised learning. Trained on LMP-video, we show that CycAs learns good generalization towards novel domains. The achieved results sometimes even outperform supervised domain generalizable models. Remarkably, CycAs achieves 82.2% Rank-1 on Market-1501 and 49.0% Rank-1 on MSMT17 with zero human annotation, surpassing state-of-the-art supervised DG re-ID methods. Moreover, we also demonstrate the superiority of CycAs under the canonical unsupervised re-ID and the pretrain-and-finetune scenarios. | ['Liang Zheng', 'Shengjin Wang', 'YaLi Li', 'Yifan Sun', 'Jingwei Zhang', 'Zhaopeng Dou', 'Zhongdao Wang'] | 2022-11-07 | null | null | null | null | ['generalizable-person-re-identification'] | ['computer-vision'] | [ 2.62433216e-02 -2.59063214e-01 -4.70813781e-01 -2.89061755e-01
-5.33892870e-01 -6.37614131e-01 6.86891079e-01 -2.63787150e-01
-5.40906310e-01 9.22158062e-01 2.12088645e-01 2.68995345e-01
-1.40513610e-02 -4.23301935e-01 -7.17447758e-01 -4.53738660e-01
-5.37877567e-02 7.88367629e-01 3.51092629e-02 -3.72979492e-02
-2.82471895e-01 1.26215294e-01 -1.70559978e+00 3.80973145e-02
8.60828340e-01 8.79449248e-01 1.85126048e-02 3.94567370e-01
4.24074680e-01 6.00589812e-01 -5.12099743e-01 -4.46921825e-01
3.46264809e-01 -5.72133243e-01 -8.30222249e-01 5.00904381e-01
6.79827869e-01 -4.44453597e-01 -7.05883861e-01 9.81783330e-01
4.52698290e-01 4.48058188e-01 6.79661810e-01 -1.37079072e+00
-9.55249071e-01 2.36271292e-01 -4.68394369e-01 3.77500057e-01
5.37276566e-01 1.59372881e-01 8.48737121e-01 -7.97621727e-01
8.49751472e-01 1.04503536e+00 5.97500622e-01 1.16831958e+00
-1.15370572e+00 -9.63511884e-01 3.79080683e-01 4.89799976e-01
-1.50064075e+00 -3.53234947e-01 6.04810834e-01 -5.44038117e-01
7.49916673e-01 -7.83200040e-02 5.27503252e-01 1.55617416e+00
-6.39165342e-01 8.53142262e-01 1.00381327e+00 -1.72262505e-01
1.66628331e-01 1.15304172e-01 1.85724124e-01 2.13007942e-01
3.12205166e-01 1.70462042e-01 -4.62165117e-01 1.61013350e-01
7.64970124e-01 3.69173169e-01 -3.47301334e-01 -3.15849274e-01
-1.43093145e+00 3.63197833e-01 1.11054264e-01 2.25606531e-01
5.24498187e-02 -1.71844125e-01 7.10670292e-01 5.59969783e-01
3.46970707e-01 2.45414734e-01 -4.08138812e-01 -2.06087947e-01
-7.72971272e-01 2.94043094e-01 5.30844212e-01 1.20833015e+00
7.72580683e-01 -1.11168064e-02 -7.39291832e-02 9.75575984e-01
-1.96013585e-01 6.68156266e-01 9.07415330e-01 -7.39896953e-01
5.78482747e-01 5.10314226e-01 2.45789737e-01 -7.54068792e-01
-2.40338758e-01 -3.10293406e-01 -1.05713224e+00 -2.79660910e-01
5.72703242e-01 -1.84257105e-01 -9.16713774e-01 1.89695776e+00
6.62082955e-02 6.97304308e-01 3.07661682e-01 9.05511737e-01
8.56658161e-01 3.94243956e-01 1.47144556e-01 -2.71568775e-01
1.26360965e+00 -1.04905069e+00 -3.61577690e-01 -1.89185441e-01
5.21397650e-01 -7.34002069e-02 9.98854995e-01 3.78589422e-01
-6.96798503e-01 -1.04759347e+00 -1.03946209e+00 2.43173406e-01
-1.97997838e-01 3.90910715e-01 2.93954432e-01 5.00260234e-01
-8.19416583e-01 3.51369083e-01 -4.32517022e-01 -6.14905000e-01
3.71514291e-01 4.38693613e-01 -8.28143060e-01 -3.74153048e-01
-1.09437013e+00 3.93501282e-01 6.64495647e-01 -2.62059033e-01
-1.17147255e+00 -6.25862539e-01 -7.47765779e-01 -1.93784907e-02
4.83637363e-01 -4.10160571e-01 1.00461948e+00 -1.33821607e+00
-1.32513726e+00 1.18905890e+00 -2.09845141e-01 -5.34539700e-01
6.37724400e-01 -3.39203596e-01 -1.04028571e+00 2.60987490e-01
3.01251620e-01 5.66275537e-01 1.01848161e+00 -1.15298975e+00
-8.35969090e-01 -3.55812073e-01 -1.43476166e-02 1.31120607e-01
-8.35372448e-01 -2.33127967e-01 -8.16854954e-01 -9.55569386e-01
-3.48401606e-01 -1.16485238e+00 9.66826230e-02 -4.83743310e-01
-2.56926000e-01 -4.10606772e-01 1.01495838e+00 -5.68281353e-01
1.09496975e+00 -2.36836004e+00 1.31405130e-01 9.23016146e-02
3.30574125e-01 5.36165655e-01 -2.53810704e-01 9.20000076e-02
-4.18918252e-01 -7.03390408e-03 8.65248404e-03 -4.20806140e-01
-1.10865854e-01 1.93325192e-01 -2.66821295e-01 4.03259516e-01
6.93241542e-04 8.72935534e-01 -1.02028310e+00 -3.01028579e-01
8.63897353e-02 8.72366205e-02 -5.97242177e-01 4.14276272e-01
-4.78755916e-03 8.47697973e-01 -2.67670304e-01 5.58138549e-01
5.95282078e-01 -5.12215436e-01 4.47127193e-01 -3.12820584e-01
2.44538292e-01 -6.04370721e-02 -1.20731390e+00 1.68593705e+00
-1.64777249e-01 5.13046205e-01 -4.42051739e-01 -1.51669836e+00
8.75777900e-01 3.47165406e-01 7.03445494e-01 -7.73486674e-01
-2.36105233e-01 1.42005354e-01 -3.79259527e-01 -3.74518454e-01
3.10373276e-01 7.23910108e-02 -2.49700591e-01 4.66222793e-01
5.27682304e-01 5.95031142e-01 4.86438841e-01 1.28095984e-01
8.16623151e-01 1.62607267e-01 3.20214242e-01 -1.34552553e-01
7.85176039e-01 -5.62368110e-02 9.34614658e-01 6.96787357e-01
-4.03831452e-01 5.78920424e-01 2.60085046e-01 -4.95541126e-01
-1.08581400e+00 -1.14234781e+00 2.23553311e-02 1.08246732e+00
5.78454554e-01 -4.26005006e-01 -6.67706251e-01 -1.13446391e+00
7.61106461e-02 1.75007924e-01 -5.48348486e-01 2.70847473e-02
-7.48338997e-01 -5.60009599e-01 5.56475163e-01 7.23556757e-01
1.07929802e+00 -7.17231035e-01 8.05010423e-02 8.47355276e-02
-3.88361096e-01 -1.49041629e+00 -8.04605901e-01 -2.87980825e-01
-6.95591509e-01 -1.35486460e+00 -1.21129811e+00 -1.11646819e+00
8.10114086e-01 5.24354994e-01 1.10058188e+00 -1.60102367e-01
-8.94883797e-02 9.03065920e-01 -5.73030174e-01 1.38942137e-01
-1.33061215e-01 1.55257909e-02 7.39972174e-01 1.69230357e-01
8.15573215e-01 -6.58920527e-01 -6.45373583e-01 8.97878826e-01
-6.70474946e-01 -1.06578566e-01 2.86276281e-01 1.18410206e+00
7.36826897e-01 1.36353716e-01 9.44669425e-01 -7.74994373e-01
1.78274617e-01 -5.81462264e-01 -3.76888126e-01 4.23714936e-01
-5.70052862e-01 -2.34269544e-01 8.14097166e-01 -9.26374316e-01
-1.05584431e+00 9.66893882e-02 1.06202856e-01 -6.21736228e-01
-3.59948248e-01 9.24418345e-02 -3.43691528e-01 1.30910888e-01
7.18044341e-01 4.63590890e-01 -1.70035407e-01 -5.21200120e-01
1.40411302e-01 7.67984748e-01 1.02994633e+00 -7.75283992e-01
1.05506277e+00 5.27727187e-01 -4.67874646e-01 -7.53318250e-01
-6.42724335e-01 -7.29623079e-01 -7.23710656e-01 -1.68760628e-01
7.78736651e-01 -1.45461690e+00 -7.81271696e-01 6.77973509e-01
-6.75113618e-01 -4.00885344e-01 -4.45914835e-01 5.30097783e-01
-5.36303401e-01 6.60799921e-01 -4.30811405e-01 -4.46064085e-01
-1.20143235e-01 -8.87360692e-01 7.40352094e-01 3.92900169e-01
-8.57984200e-02 -9.14133430e-01 1.37358054e-03 5.07512391e-01
-2.22057179e-02 7.39558563e-02 3.54918122e-01 -1.04324687e+00
-2.54859030e-01 -1.33546799e-01 -3.37800860e-01 6.43009841e-01
4.04336691e-01 -6.62180841e-01 -1.00136673e+00 -7.24379539e-01
-3.33410859e-01 -5.61625659e-01 7.39682257e-01 -8.53578895e-02
1.29236364e+00 -2.94225603e-01 -4.47145313e-01 7.69343019e-01
1.13488305e+00 1.02396205e-01 5.71190059e-01 5.46314657e-01
8.29473615e-01 4.75901008e-01 7.46600389e-01 5.62311769e-01
4.55116183e-01 9.67236340e-01 -6.25497149e-03 3.58170718e-02
-2.11825117e-01 -5.75431526e-01 5.13467669e-01 6.36129022e-01
-5.07683516e-01 -1.73478901e-01 -7.30307281e-01 6.92922115e-01
-1.95949614e+00 -1.20689678e+00 3.61829400e-01 2.31513643e+00
6.65519893e-01 -1.02133371e-01 7.62017012e-01 5.78651316e-02
1.03084767e+00 -2.15837378e-02 -9.17818725e-01 4.52433288e-01
-2.54948258e-01 -6.43933564e-02 5.42990685e-01 -4.85409610e-02
-1.43115032e+00 8.73402417e-01 5.68431234e+00 8.31797719e-01
-1.00987899e+00 1.99125528e-01 5.13087928e-01 -4.13882136e-02
1.42115563e-01 -3.61072898e-01 -1.11031508e+00 8.03801537e-01
7.56438136e-01 -3.99508268e-01 5.19317925e-01 1.05565965e+00
-6.18143678e-02 3.76321435e-01 -1.33254838e+00 1.63773656e+00
2.92952001e-01 -1.12606204e+00 1.51696384e-01 -1.27441250e-02
9.98499334e-01 -1.32002145e-01 1.52388945e-01 5.30552506e-01
2.94687510e-01 -9.18941736e-01 4.86680448e-01 1.62592337e-01
1.44730616e+00 -7.01247871e-01 6.31840944e-01 1.88894570e-01
-1.38006401e+00 -3.79127204e-01 -4.41881835e-01 9.94108170e-02
2.58874409e-02 2.74171948e-01 -5.24051487e-01 6.74288690e-01
1.11007118e+00 1.42418122e+00 -5.84949493e-01 7.83240557e-01
-1.32281959e-01 6.48734629e-01 -6.38458654e-02 5.50802112e-01
-1.39374942e-01 4.54400405e-02 4.41921264e-01 1.26443040e+00
3.90066385e-01 9.81716588e-02 3.86799365e-01 3.52894455e-01
-3.69814515e-01 -1.46377832e-01 -4.66126651e-01 -5.89421540e-02
4.61247802e-01 7.72664428e-01 -2.79312283e-01 -5.88438630e-01
-6.89312398e-01 1.44350350e+00 2.92556405e-01 6.65658772e-01
-8.23086619e-01 -5.18316403e-02 9.01145160e-01 1.97690114e-01
4.58723098e-01 -1.70652077e-01 3.44206005e-01 -1.77872813e+00
1.42739683e-01 -1.18141735e+00 7.25300848e-01 -2.82254398e-01
-1.78628504e+00 5.87535679e-01 2.10906595e-01 -1.83265746e+00
-5.22247970e-01 -7.22728193e-01 -2.91634053e-01 4.04355109e-01
-1.57557368e+00 -1.25507522e+00 -6.39589012e-01 1.20853412e+00
5.96192896e-01 -7.88456023e-01 7.51857698e-01 5.75390041e-01
-7.32971251e-01 1.19248283e+00 4.09769297e-01 5.86357534e-01
1.18578684e+00 -1.03854585e+00 3.25095266e-01 9.74040806e-01
9.17487442e-02 5.81622005e-01 2.96092540e-01 -6.63043201e-01
-1.23691857e+00 -1.37987638e+00 5.71226299e-01 -4.12449270e-01
4.86078382e-01 -3.61090124e-01 -8.15072358e-01 9.36269939e-01
-2.95359552e-01 3.50876123e-01 8.07306588e-01 6.40885979e-02
-6.65347338e-01 -3.80461901e-01 -1.12946761e+00 4.45546389e-01
1.66178167e+00 -6.15342081e-01 -6.68740869e-01 4.08888757e-01
5.35000443e-01 -2.31856078e-01 -1.11611366e+00 4.01757598e-01
5.19129217e-01 -7.27503896e-01 1.26190639e+00 -7.39723504e-01
1.03358328e-01 -3.55820745e-01 -1.28693134e-01 -1.16388941e+00
-4.74113166e-01 -5.82964957e-01 -2.78380364e-01 1.54451144e+00
-5.09426445e-02 -7.59678304e-01 1.02306557e+00 4.54258561e-01
2.02374339e-01 -7.94464648e-02 -9.17809486e-01 -1.36077690e+00
-1.19172581e-01 -1.82501078e-01 7.39447296e-01 1.19269502e+00
1.88350491e-02 1.24548055e-01 -8.47882628e-01 5.68133295e-02
7.91977704e-01 -6.47609606e-02 1.08086634e+00 -1.32974148e+00
-5.56088269e-01 1.69967674e-02 -6.20846748e-01 -1.60843611e+00
4.56925929e-01 -7.85695076e-01 -2.31896639e-01 -8.83008838e-01
4.84708875e-01 -6.77116215e-01 -4.15504992e-01 5.29629290e-01
-2.56573915e-01 4.43685800e-01 2.34425172e-01 8.01565230e-01
-9.39092934e-01 4.04592156e-01 1.06049860e+00 -4.67445582e-01
-2.62489557e-01 -3.64190042e-02 -7.00228572e-01 5.42399168e-01
6.20819807e-01 -8.61229599e-02 -5.74782848e-01 -3.29299152e-01
-3.17215472e-01 -2.63367236e-01 5.72626829e-01 -1.20161355e+00
2.95869857e-01 -1.00759506e-01 5.41425824e-01 -2.93054163e-01
1.13693342e-01 -7.19287157e-01 9.88666564e-02 1.34906933e-01
-2.34794468e-01 -8.70436728e-02 3.27863954e-02 8.75236750e-01
-3.72686923e-01 -4.53128777e-02 6.89843535e-01 -9.39567015e-02
-1.53858089e+00 6.01426899e-01 7.99659938e-02 3.44097912e-01
1.19031084e+00 -4.13393527e-01 -4.36362684e-01 -4.06841755e-01
-8.41423571e-01 3.13274056e-01 7.16687441e-01 5.89618385e-01
5.61339080e-01 -1.49089086e+00 -7.43037224e-01 3.47536683e-01
5.62800229e-01 -9.96502265e-02 5.48590124e-01 4.37787503e-01
-2.12370213e-02 3.79408419e-01 -3.68763953e-01 -7.73843408e-01
-1.19299436e+00 8.48350227e-01 1.58586800e-01 -3.62918258e-01
-8.24130595e-01 5.62727451e-01 5.37714422e-01 -4.00092691e-01
2.35444874e-01 3.24346632e-01 -3.78240079e-01 -6.34632111e-02
8.70774090e-01 4.53344136e-01 -2.35724598e-01 -8.61059070e-01
-4.36902702e-01 6.00992918e-01 -2.94902146e-01 1.34169057e-01
1.18064392e+00 -2.88642913e-01 3.41894031e-01 1.15240723e-01
1.20969689e+00 -2.38224417e-01 -1.61529601e+00 -5.83109856e-01
-2.60579139e-02 -4.19465125e-01 -6.07578337e-01 -5.96987247e-01
-1.01033187e+00 4.36104923e-01 8.55281413e-01 -2.79428482e-01
1.29736447e+00 1.54681608e-01 9.86145198e-01 4.84420061e-01
5.18804669e-01 -1.40348363e+00 5.21853745e-01 3.75969410e-01
4.75851566e-01 -1.46417046e+00 -2.38190349e-02 -2.44092137e-01
-8.79773378e-01 8.39627802e-01 9.10753846e-01 -1.49115413e-01
3.99088621e-01 -2.14615300e-01 -2.33312756e-01 2.69943595e-01
-3.41943473e-01 -3.36204737e-01 3.06607932e-01 1.20369208e+00
8.57164785e-02 6.99567720e-02 5.76891750e-02 9.55180287e-01
2.86919605e-02 2.89079309e-01 1.98888525e-01 5.03215849e-01
1.94821656e-02 -1.19835615e+00 -3.07289362e-01 3.00215691e-01
-3.09238583e-01 2.97773510e-01 -2.42433287e-02 8.87267113e-01
3.41985494e-01 9.84509826e-01 1.05075702e-01 -7.05485821e-01
3.52831692e-01 -2.10209787e-01 3.80589515e-01 -6.12821519e-01
-2.33246982e-01 -2.77144223e-01 6.93207234e-03 -4.34291542e-01
-6.79518700e-01 -6.88671410e-01 -9.08873677e-01 -2.51570553e-01
8.77384245e-02 1.78463250e-01 -6.98603466e-02 9.84324098e-01
3.93045187e-01 1.84589773e-01 6.70512736e-01 -8.48208368e-01
-4.65975106e-01 -7.81380951e-01 -6.44135773e-01 1.04501605e+00
3.02479088e-01 -9.08889472e-01 -3.45303118e-01 5.41032732e-01] | [14.741125106811523, 1.0589454174041748] |
c6c93fbc-b13d-49e6-a1e9-67c69ff5a978 | quantum-algorithms-applied-to-satellite | 2302.07181 | null | https://arxiv.org/abs/2302.07181v1 | https://arxiv.org/pdf/2302.07181v1.pdf | Quantum algorithms applied to satellite mission planning for Earth observation | Earth imaging satellites are a crucial part of our everyday lives that enable global tracking of industrial activities. Use cases span many applications, from weather forecasting to digital maps, carbon footprint tracking, and vegetation monitoring. However, there are also limitations; satellites are difficult to manufacture, expensive to maintain, and tricky to launch into orbit. Therefore, it is critical that satellites are employed efficiently. This poses a challenge known as the satellite mission planning problem, which could be computationally prohibitive to solve on large scales. However, close-to-optimal algorithms can often provide satisfactory resolutions, such as greedy reinforcement learning, and optimization algorithms. This paper introduces a set of quantum algorithms to solve the mission planning problem and demonstrate an advantage over the classical algorithms implemented thus far. The problem is formulated as maximizing the number of high-priority tasks completed on real datasets containing thousands of tasks and multiple satellites. This work demonstrates that through solution-chaining and clustering, optimization and machine learning algorithms offer the greatest potential for optimal solutions. Most notably, this paper illustrates that a hybridized quantum-enhanced reinforcement learning agent can achieve a completion percentage of 98.5% over high-priority tasks, which is a significant improvement over the baseline greedy methods with a completion rate of 63.6%. The results presented in this work pave the way to quantum-enabled solutions in the space industry and, more generally, future mission planning problems across industries. | ['Alexey Melnikov', 'Markus Pflitsch', 'Mohammad Kordzanganeh', 'Egor Barashov', 'Miras Koblan', 'Daria Lemtiuzhnikova', 'Karan Pinto', 'Saaketh Rayaprolu', 'Sergei Iudin', 'Igor Tokarev', 'Serge Rainjonneau'] | 2023-02-14 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [ 4.40496683e-01 -3.39626335e-02 -2.47637004e-01 1.67312939e-02
-8.01287115e-01 -5.21696806e-01 3.28221738e-01 -1.75058126e-01
-4.67444718e-01 1.16037714e+00 -3.93009394e-01 -3.48529756e-01
-6.81296468e-01 -8.09485555e-01 -7.29652166e-01 -1.09046185e+00
-4.93477196e-01 8.19937289e-01 -2.51729786e-01 -5.31319976e-01
5.01779556e-01 3.54511380e-01 -1.68725312e+00 -3.75315547e-01
1.16061747e+00 9.41758037e-01 7.82767534e-01 4.10928220e-01
3.46021026e-01 2.63496518e-01 -4.86457437e-01 1.94866788e-02
2.95061350e-01 -1.76049277e-01 -8.33596528e-01 2.90491700e-01
-1.80300355e-01 6.07322827e-02 -6.63619637e-02 1.25308931e+00
4.39185560e-01 2.09382191e-01 2.66641110e-01 -1.40139735e+00
-2.26838797e-01 2.56207198e-01 -5.34991205e-01 2.73911744e-01
1.34872556e-01 6.38694540e-02 1.15400565e+00 -1.49498045e-01
5.09114563e-01 8.08233500e-01 2.17585713e-01 1.93924636e-01
-1.06156349e+00 -4.96015787e-01 -4.06393200e-01 4.34325308e-01
-1.25136495e+00 -3.26015592e-01 1.55859470e-01 -1.68250695e-01
1.36141145e+00 4.83387470e-01 6.94049001e-01 4.90157276e-01
7.04608560e-01 2.49612600e-01 1.34564650e+00 -3.33813995e-01
4.91258204e-01 -3.44540000e-01 -3.42331946e-01 5.89469075e-01
4.28686857e-01 4.15510029e-01 -6.87782586e-01 -1.49942875e-01
5.72416902e-01 -1.93820953e-01 -1.16474837e-01 -1.10887386e-01
-1.55465651e+00 8.96747112e-01 5.65487266e-01 3.18100959e-01
-7.68114090e-01 4.55696255e-01 -9.46905613e-02 3.72448653e-01
2.78399736e-01 9.33830380e-01 -1.74740463e-01 -2.76254058e-01
-9.49009240e-01 4.71392781e-01 6.44463956e-01 8.08611989e-01
9.06289756e-01 -2.09682360e-02 1.41254842e-01 3.68183963e-02
1.70090556e-01 8.95633221e-01 8.57454762e-02 -1.54500675e+00
2.45548651e-01 3.30445878e-02 4.88431782e-01 -6.73330724e-01
-8.65199327e-01 -9.06198978e-01 -8.65419805e-01 1.59167692e-01
1.35689080e-01 -1.72916561e-01 -5.01683056e-01 1.40679741e+00
5.30011654e-01 1.95856467e-02 1.70192137e-01 1.26573348e+00
2.02062931e-02 7.55374193e-01 -2.59307265e-01 -5.58063686e-01
1.63879204e+00 -8.98517966e-01 -6.92625523e-01 -4.96921003e-01
7.24638820e-01 -8.70749593e-01 7.20663130e-01 6.59571767e-01
-8.83123517e-01 -3.60763492e-03 -1.04767501e+00 4.80992913e-01
-1.51963219e-01 -1.33563608e-01 1.16864574e+00 9.25549567e-01
-1.10104501e+00 1.01340091e+00 -8.09775651e-01 -4.97187555e-01
1.00312531e-01 7.46726036e-01 -4.04122844e-03 -1.97095394e-01
-1.16572201e+00 8.90131652e-01 6.45073235e-01 -2.06011891e-01
-7.53383994e-01 -1.15496404e-01 -1.95657358e-01 1.04325965e-01
6.86101317e-01 -9.02492881e-01 1.23487663e+00 -5.67159891e-01
-1.46459460e+00 4.75891769e-01 1.42624810e-01 -6.37192607e-01
-1.33824468e-01 2.66788810e-01 -3.36745679e-01 1.04937688e-01
3.75333607e-01 5.67353368e-01 8.81691575e-01 -5.04472017e-01
-8.95942569e-01 -4.16519076e-01 1.94389775e-01 6.35869980e-01
-4.52856690e-01 1.25253469e-01 -8.18242654e-02 -1.22703418e-01
2.53498763e-01 -1.57171381e+00 -6.09417200e-01 -7.00067043e-01
-1.43382221e-01 -2.99149245e-01 2.40936816e-01 -5.27867794e-01
8.91342878e-01 -1.90840948e+00 4.66777205e-01 -2.08173897e-02
-3.88008654e-02 -4.37699020e-01 -3.17923445e-03 5.18775642e-01
2.80359477e-01 1.79799069e-02 -2.49867991e-01 1.33374199e-01
2.89709717e-01 4.66344595e-01 -8.01168606e-02 5.60983896e-01
-1.25055447e-01 7.02996969e-01 -1.01862943e+00 -1.73574865e-01
-3.45148370e-02 -2.61715472e-01 -3.86468977e-01 -3.88113320e-01
-4.17641103e-01 6.14789724e-01 -6.06440365e-01 8.82900536e-01
4.29453909e-01 -6.62150085e-01 5.13816416e-01 1.07835449e-01
-4.02639836e-01 3.81636411e-01 -1.20465493e+00 1.79341924e+00
-2.18173295e-01 1.81648865e-01 3.78328830e-01 -1.28792620e+00
4.00495768e-01 -1.20553719e-02 7.69208550e-01 -1.08315635e+00
-7.57634863e-02 4.48985815e-01 3.94745804e-02 -2.89222181e-01
1.05234969e+00 -4.34141457e-01 -4.36442792e-01 5.83121657e-01
-2.04671934e-01 -5.40197551e-01 4.74266738e-01 1.99895784e-01
1.11563492e+00 -2.02601422e-02 2.09813267e-01 -4.93248522e-01
-3.65201943e-02 5.38762808e-01 5.61979830e-01 6.77027047e-01
-8.43464509e-02 -4.92197946e-02 1.85608387e-01 -2.68673480e-01
-1.06271875e+00 -6.79674566e-01 -2.01903269e-01 1.18484306e+00
4.51705337e-01 -4.04680461e-01 -6.98435009e-01 3.72799747e-02
-5.37008829e-02 7.56618142e-01 5.16240783e-02 1.84807740e-02
-1.63791656e-01 -1.30871177e+00 1.11958876e-01 -1.04102299e-01
5.58323503e-01 -6.10605896e-01 -7.70634532e-01 3.69488597e-01
-5.01316011e-01 -1.33380687e+00 -3.30049545e-02 2.97034264e-01
-1.09717333e+00 -7.88599849e-01 -6.07829332e-01 -2.19072640e-01
3.60496372e-01 6.82973087e-01 1.02459168e+00 -2.96190530e-02
-2.62964100e-01 1.73360363e-01 -2.25602537e-01 -2.79274255e-01
-7.21417665e-02 2.81279385e-01 7.14777231e-01 -4.07524377e-01
-1.00496277e-01 -3.97954136e-01 -5.07850766e-01 4.29656923e-01
-6.31770730e-01 -1.48107320e-01 8.77141058e-01 7.94450104e-01
7.69774020e-01 7.81866133e-01 3.92848104e-01 -2.31252611e-01
4.20645058e-01 -4.88984227e-01 -9.12451863e-01 2.38218099e-01
-8.49276125e-01 2.33550653e-01 3.14547598e-01 2.01403499e-01
-6.59791291e-01 2.44988129e-01 2.75146902e-01 2.88605690e-02
1.69579014e-01 6.81529164e-01 4.04320091e-01 -5.32213748e-01
7.23977864e-01 3.47105712e-01 2.68755388e-02 -9.79716145e-03
2.07735568e-01 7.53545642e-01 1.23070799e-01 -5.35322249e-01
6.92131937e-01 5.32089651e-01 5.41867197e-01 -1.14727354e+00
-6.35466754e-01 -4.15235281e-01 -1.00785255e-01 -3.59832376e-01
9.16088462e-01 -1.05376899e+00 -8.90203416e-01 7.70755783e-02
-8.59596491e-01 -1.88466564e-01 -1.06857918e-01 8.03784549e-01
-8.20097387e-01 4.75310475e-01 -1.11569621e-01 -8.60858619e-01
-2.09614798e-01 -1.24112320e+00 1.20409107e+00 3.96031469e-01
1.57822385e-01 -5.85878491e-01 -1.90561399e-01 8.33832443e-01
4.27273482e-01 1.25872731e-01 6.05857015e-01 1.03812635e-01
-1.27986467e+00 2.21198827e-01 8.92388914e-03 -2.52206236e-01
-1.11544922e-01 -4.60608900e-01 -5.94517589e-01 -7.28238225e-01
1.84690841e-02 -5.25418878e-01 6.32442415e-01 4.42532390e-01
6.54144824e-01 4.56612669e-02 -4.57075953e-01 4.55702543e-01
1.19834733e+00 -3.42047252e-02 6.02323592e-01 7.42515922e-01
1.15976967e-01 4.77449656e-01 1.20092380e+00 6.56320453e-01
3.22854906e-01 8.21855783e-01 9.66648400e-01 4.45550084e-01
4.05886948e-01 5.02228796e-01 4.35072511e-01 1.03275156e+00
-5.39245069e-01 2.71803904e-02 -1.06230474e+00 2.87998378e-01
-1.75881827e+00 -9.30343747e-01 -2.56744802e-01 2.38465810e+00
3.56044322e-01 5.97644262e-02 7.12832361e-02 -9.07354578e-02
6.94046080e-01 2.21571159e-02 -7.40886331e-01 -2.91632116e-01
-8.93428251e-02 3.58729176e-02 1.06733131e+00 7.08332881e-02
-1.06546259e+00 9.08236206e-01 6.23855209e+00 9.90051925e-01
-9.82881546e-01 5.59009910e-01 1.24522988e-02 -3.69111180e-01
-3.00365299e-01 3.13635558e-01 -7.12391198e-01 2.91289836e-01
1.37143028e+00 -4.38750178e-01 1.22261024e+00 6.85973763e-01
1.70690939e-01 -4.03377712e-01 -5.69209218e-01 1.15957153e+00
-2.48392940e-01 -1.31499422e+00 -4.22164321e-01 7.05330849e-01
8.94317925e-01 6.35589480e-01 7.53372386e-02 1.66987136e-01
3.26676041e-01 -8.95055234e-01 6.23646080e-01 3.53615344e-01
7.17866898e-01 -9.84818339e-01 3.44178796e-01 8.32575619e-01
-8.77061248e-01 -4.22109187e-01 -7.37427652e-01 -5.63131869e-01
4.07034010e-01 8.94741654e-01 -7.21460640e-01 1.06708145e+00
8.39561343e-01 2.94243813e-01 7.51245096e-02 1.11491454e+00
-4.56411280e-02 2.02689871e-01 -6.62072420e-01 -5.02865732e-01
5.56368887e-01 -7.01146483e-01 6.61300302e-01 2.89283365e-01
9.62534308e-01 1.49276376e-01 3.50092560e-01 5.75108349e-01
-1.20904576e-03 -1.47931099e-01 -7.44922221e-01 -7.48186171e-01
3.52825969e-01 1.38664699e+00 -1.00564623e+00 2.37852931e-02
-1.99559748e-01 9.66277957e-01 1.13486692e-01 5.06574325e-02
-9.58231509e-01 -1.70298070e-01 7.10373163e-01 -2.59143144e-01
2.82651275e-01 -6.28999472e-01 -1.07843384e-01 -9.98824000e-01
-1.82689980e-01 -1.03856838e+00 1.77571282e-01 -7.32473195e-01
-7.87922263e-01 2.12762490e-01 -3.33586991e-01 -1.29658401e+00
-2.89170861e-01 -6.39654160e-01 -6.22853227e-02 6.46509111e-01
-1.63154125e+00 -5.41571915e-01 -3.22835259e-02 3.40851873e-01
1.88086987e-01 -1.57679766e-01 8.70721936e-01 2.58007646e-01
-4.95079905e-01 -1.44425139e-01 8.36229861e-01 -9.65236187e-01
3.70483369e-01 -1.06547534e+00 2.63384134e-01 6.60262048e-01
2.36800551e-01 1.78511798e-01 8.36205900e-01 -6.46628618e-01
-2.11309481e+00 -8.34243119e-01 7.67675757e-01 2.37075537e-02
8.81484032e-01 -4.27871756e-02 -3.95674929e-02 4.03953403e-01
3.02302837e-01 -4.87045944e-01 5.72576582e-01 2.28849083e-01
4.87006873e-01 -6.29819110e-02 -9.10117328e-01 3.94928962e-01
1.00961792e+00 -4.76818681e-01 -1.44851610e-01 1.38848722e+00
6.51551664e-01 -5.11731982e-01 -9.70079541e-01 1.08801745e-01
2.05253094e-01 -8.49812567e-01 9.76119161e-01 -3.32191527e-01
7.15668797e-02 -2.97146559e-01 -4.33373630e-01 -1.34948170e+00
-6.41230166e-01 -1.00708890e+00 1.52739614e-01 5.15195489e-01
3.31285417e-01 -8.00090075e-01 7.27810025e-01 4.34327751e-01
-2.60349035e-01 -3.72335732e-01 -1.46206105e+00 -1.27249491e+00
-9.83865783e-02 -3.94630879e-01 5.90637982e-01 9.66552079e-01
-8.59700590e-02 3.84525031e-01 -4.62664753e-01 4.95993912e-01
1.17134917e+00 5.82314432e-01 3.95906389e-01 -1.24874580e+00
-5.71907163e-01 -1.67461142e-01 -1.91613570e-01 -8.19277704e-01
-4.60691378e-02 -1.04481232e+00 -1.11931570e-01 -1.49239910e+00
1.02600165e-01 -7.05144882e-01 -1.54102340e-01 2.32619107e-01
2.85003245e-01 2.04454839e-01 2.79480010e-01 5.68789244e-01
-8.30140650e-01 6.11213267e-01 1.38438523e+00 -2.21212819e-01
1.18043214e-01 2.55914629e-02 -6.16580367e-01 4.44135725e-01
1.07989550e+00 -8.12794209e-01 -1.28866673e-01 -5.08331716e-01
9.18636918e-01 4.14057493e-01 3.20241779e-01 -1.21951926e+00
4.86148447e-01 -3.72211099e-01 -2.45010182e-01 -4.22414929e-01
5.13872385e-01 -5.02895594e-01 3.76864791e-01 9.60456073e-01
5.94224870e-01 -1.27907973e-02 -2.12574840e-01 7.44840503e-01
7.55908266e-02 -5.06106019e-01 3.85830969e-01 -3.42669278e-01
-8.88033450e-01 -1.46065531e-02 -4.52893108e-01 -2.09211245e-01
1.31111848e+00 6.66486546e-02 -4.57461983e-01 -1.56894892e-01
-7.30044842e-01 5.19276857e-01 4.48786736e-01 5.19887581e-02
1.79895326e-01 -1.23967159e+00 -6.21679068e-01 -3.85893524e-01
-1.53414279e-01 -3.33987921e-01 3.33997190e-01 1.27993929e+00
-5.29009759e-01 8.48523498e-01 -3.90773654e-01 -8.47476661e-01
-9.92054582e-01 5.51052928e-01 1.49526447e-01 -4.61086810e-01
-2.25790888e-01 5.91767311e-01 -3.53231609e-01 -5.51198184e-01
-2.35732734e-01 -6.06331155e-02 1.97353289e-01 1.03965878e-01
2.85139054e-01 6.30094051e-01 2.21815884e-01 -3.69486898e-01
-4.75806534e-01 5.15438795e-01 3.15910012e-01 -2.12314740e-01
1.32230866e+00 -1.90070719e-01 -4.42818999e-01 6.14326894e-02
4.22813535e-01 -5.06187975e-01 -8.33031476e-01 -6.32767752e-02
8.53936672e-02 -1.87175155e-01 6.05338216e-01 -5.15745163e-01
-6.48555279e-01 7.06850708e-01 5.95701516e-01 6.58735037e-01
1.14973819e+00 -4.65041399e-02 7.98497736e-01 1.15858865e+00
1.16720831e+00 -1.55993092e+00 -1.85733587e-02 4.75296348e-01
5.29744864e-01 -1.17350161e+00 3.29034090e-01 -3.95333409e-01
-4.47190732e-01 1.00798976e+00 -2.63022119e-03 1.42065421e-01
1.12358660e-01 1.26515105e-01 -5.66353798e-01 -3.90333474e-01
-6.64746225e-01 -7.24598289e-01 -2.39264578e-01 2.94762582e-01
3.30997519e-02 5.00252783e-01 -5.87211907e-01 2.25182742e-01
-2.50957668e-01 6.65622205e-02 6.15338922e-01 1.16629148e+00
-1.02233160e+00 -1.25413311e+00 -8.68408561e-01 5.29493868e-01
-3.03592026e-01 -1.61551043e-01 -3.16481777e-02 5.09059012e-01
-1.34424902e-02 1.08320928e+00 -7.52474554e-03 -3.84939432e-01
-2.41263695e-02 -9.62383226e-02 8.43654752e-01 -3.97127658e-01
-2.70099759e-01 -3.58916819e-02 4.64686871e-01 -6.90860987e-01
-7.08846986e-01 -7.84231186e-01 -1.43313086e+00 -5.16433358e-01
-5.46701014e-01 3.26870263e-01 1.11650276e+00 1.10417914e+00
5.69678426e-01 6.47462368e-01 7.32707500e-01 -1.19745660e+00
-1.17977977e+00 -5.37798584e-01 -1.06325221e+00 -2.32901990e-01
-1.83059230e-01 -8.67322981e-01 -4.94393483e-02 -3.92690897e-01] | [5.601892471313477, 4.766911029815674] |
c8758889-6d08-48db-83ff-89efbb6dab2e | tailoring-to-the-tails-risk-measures-for-fine | 2208.03066 | null | https://arxiv.org/abs/2208.03066v2 | https://arxiv.org/pdf/2208.03066v2.pdf | Tailoring to the Tails: Risk Measures for Fine-Grained Tail Sensitivity | Expected risk minimization (ERM) is at the core of many machine learning systems. This means that the risk inherent in a loss distribution is summarized using a single number - its average. In this paper, we propose a general approach to construct risk measures which exhibit a desired tail sensitivity and may replace the expectation operator in ERM. Our method relies on the specification of a reference distribution with a desired tail behaviour, which is in a one-to-one correspondence to a coherent upper probability. Any risk measure, which is compatible with this upper probability, displays a tail sensitivity which is finely tuned to the reference distribution. As a concrete example, we focus on divergence risk measures based on f-divergence ambiguity sets, which are a widespread tool used to foster distributional robustness of machine learning systems. For instance, we show how ambiguity sets based on the Kullback-Leibler divergence are intricately tied to the class of subexponential random variables. We elaborate the connection of divergence risk measures and rearrangement invariant Banach norms. | ['Robert C. Williamson', 'Christian Fröhlich'] | 2022-08-05 | null | null | null | null | ['machine-learning', 'machine-learning'] | ['methodology', 'miscellaneous'] | [ 1.01559833e-01 3.70898128e-01 1.31831452e-01 -4.01894331e-01
-8.20396781e-01 -5.52048624e-01 6.00996614e-01 3.90188873e-01
-6.33976519e-01 8.55913877e-01 -9.14439932e-02 -1.32550195e-01
-4.86953020e-01 -9.38573003e-01 -5.37984550e-01 -9.99830782e-01
-3.04827243e-01 2.00771570e-01 -9.16955471e-02 -3.50041270e-01
2.61589140e-01 7.65544891e-01 -1.56499517e+00 -3.26066822e-01
8.64507794e-01 1.16082549e+00 -3.37056965e-01 5.78435004e-01
-1.21503912e-01 3.33059460e-01 -6.33861423e-01 -5.45955420e-01
1.52128667e-01 -4.29934561e-01 -7.24033654e-01 -3.31039131e-01
6.35217584e-04 2.96663523e-01 4.17458981e-01 1.49732757e+00
3.24011803e-01 1.31298304e-01 1.48526406e+00 -1.37072122e+00
-3.35637957e-01 5.24114013e-01 -1.71931937e-01 -1.40892863e-02
1.41926691e-01 -3.16455036e-01 1.09600425e+00 -7.37415552e-01
2.09462091e-01 8.90776396e-01 7.27084816e-01 4.94091094e-01
-1.46207762e+00 -6.89672492e-03 -2.25671247e-01 -3.41836303e-01
-1.32364535e+00 -5.90879209e-02 6.32759571e-01 -8.08527350e-01
3.77059914e-02 5.58225513e-01 2.70086586e-01 8.26806605e-01
7.22374499e-01 2.28520095e-01 9.30958271e-01 -4.57795829e-01
6.18342519e-01 3.52583021e-01 6.59120008e-02 4.92345721e-01
4.63243604e-01 3.49406190e-02 3.21103446e-02 -1.82938263e-01
4.31514263e-01 -1.53459206e-01 -4.15119201e-01 -8.46512794e-01
-8.46416771e-01 9.38613415e-01 2.52734646e-02 6.46653891e-01
-1.47471219e-01 1.27810791e-01 5.89929819e-01 4.88297194e-01
4.94722426e-01 4.13835496e-01 -1.85008720e-01 1.14903606e-01
-5.36474764e-01 1.45480484e-01 1.09496427e+00 7.35961258e-01
4.52438831e-01 -1.82073966e-01 -3.44737113e-01 4.19898719e-01
4.30361897e-01 5.81640959e-01 1.20863438e-01 -8.20411146e-01
8.24836567e-02 2.75629818e-01 3.28355074e-01 -6.68716192e-01
-3.08035702e-01 -6.15216851e-01 -1.03175223e+00 5.60184956e-01
6.12915397e-01 1.40548304e-01 2.12386362e-02 2.25561547e+00
2.23271325e-01 -4.22091514e-01 1.14392772e-01 5.53062797e-01
-2.59199768e-01 2.96055615e-01 1.05190851e-01 -5.71490586e-01
1.00292778e+00 -1.18892297e-01 -6.72846615e-01 6.03814244e-01
7.12179422e-01 -4.12340879e-01 1.35087574e+00 6.02858841e-01
-1.12229609e+00 -3.38661000e-02 -1.13581240e+00 3.42625588e-01
-4.65898424e-01 -3.11565906e-01 2.21220381e-03 1.04053116e+00
-1.01411307e+00 1.12040329e+00 -2.22694963e-01 -1.93251923e-01
7.07539022e-02 6.30141869e-02 -4.63601112e-01 6.25647962e-01
-1.22069871e+00 1.08897233e+00 3.17743748e-01 5.95497638e-02
-5.66806078e-01 -8.58371854e-01 -8.07463527e-01 1.84232667e-01
-7.92403743e-02 -5.08059859e-01 1.14933157e+00 -8.52656782e-01
-1.54766202e+00 1.08253825e+00 5.81486642e-01 -4.76035506e-01
9.99971569e-01 -2.05277856e-02 -2.33933151e-01 -3.85677069e-02
-7.27919564e-02 -1.62971199e-01 1.04531789e+00 -8.88669968e-01
-9.01183784e-02 -3.07985395e-01 -1.54093668e-01 -3.02839160e-01
-3.56899053e-01 1.84358180e-01 6.14125311e-01 -7.54569292e-01
-2.02181920e-01 -4.83140141e-01 -1.03168137e-01 1.86536223e-01
-2.93046147e-01 -3.90170604e-01 1.80190414e-01 -9.83264372e-02
1.29256094e+00 -2.25183392e+00 3.26827586e-01 5.26063323e-01
-3.96674536e-02 8.86544120e-03 2.24210545e-01 4.56672400e-01
-4.14566725e-01 3.32233459e-01 -6.06027663e-01 -1.43699825e-01
6.27321005e-01 -2.27503657e-01 -3.91282022e-01 1.14377010e+00
2.40200281e-01 2.49544337e-01 -7.16344833e-01 -3.29751074e-01
5.48039041e-02 3.87266427e-01 -4.45828825e-01 2.25077242e-01
-3.30710113e-01 1.40210077e-01 -4.16050553e-01 -2.94247698e-02
7.23668456e-01 8.26800391e-02 -1.27815455e-01 -7.37031400e-02
-2.21028805e-01 -2.16988713e-01 -1.26411247e+00 1.21901524e+00
-3.48829836e-01 1.04575105e-01 2.93578327e-01 -1.25324428e+00
1.22623825e+00 2.00308546e-01 5.55691183e-01 -2.11307686e-02
5.25819004e-01 4.01517600e-01 -3.85682404e-01 -2.19827279e-01
1.64204493e-01 -7.24738896e-01 -3.34380955e-01 5.46956718e-01
3.29488255e-02 -2.90925890e-01 3.31098586e-02 -2.17730552e-01
7.37649262e-01 -1.64453402e-01 4.38873291e-01 -1.17896056e+00
9.07156050e-01 -7.48359323e-01 3.09199840e-01 5.20529628e-01
-3.57152253e-01 4.76129174e-01 9.96742308e-01 -6.87907338e-02
-1.03420830e+00 -1.49504805e+00 -6.55798912e-01 7.12924123e-01
-2.45973527e-01 -1.77594215e-01 -9.91730273e-01 -6.69099092e-01
-6.97861537e-02 8.40073526e-01 -7.96548486e-01 -5.33294618e-01
-5.98106869e-02 -5.13176203e-01 7.12200284e-01 4.91964072e-02
2.14788020e-01 -7.42894650e-01 -5.24320424e-01 9.64809954e-03
1.64959580e-01 -3.97517681e-01 -6.67894483e-01 4.66928303e-01
-7.99918354e-01 -1.02914214e+00 -9.92323101e-01 -3.92437816e-01
4.18103307e-01 -6.30847037e-01 1.19555616e+00 -3.85756254e-01
-2.36245215e-01 7.21269727e-01 -6.94280416e-02 -6.76195443e-01
-7.77964413e-01 -3.51090461e-01 4.04468447e-01 2.29420930e-01
9.40341651e-02 -5.88725567e-01 -3.53722811e-01 2.92687714e-01
-1.29733348e+00 -8.33772719e-01 1.24665588e-01 6.54226840e-01
6.31193101e-01 2.28901491e-01 8.12506735e-01 -4.08369392e-01
8.37929845e-01 -4.55461770e-01 -8.16811562e-01 4.40199822e-01
-6.13247395e-01 5.27058125e-01 6.30857706e-01 -4.24599648e-01
-7.02363789e-01 -1.44677311e-01 -2.21392494e-02 -1.28306061e-01
7.79328570e-02 3.70443970e-01 -5.25632381e-01 -2.51744706e-02
7.65380800e-01 -9.78654772e-02 2.58185685e-01 -3.29046577e-01
4.12921578e-01 6.23762012e-01 4.41175789e-01 -9.26116467e-01
4.97093320e-01 2.99570292e-01 5.75064480e-01 -8.65565240e-01
-8.41752470e-01 -2.07401067e-01 -4.86416399e-01 -2.61420786e-01
8.44370723e-01 -1.51756793e-01 -9.23446119e-01 6.07281506e-01
-9.90970314e-01 -8.01658407e-02 -1.01834261e+00 6.01631880e-01
-1.22721887e+00 3.19430560e-01 -3.95580828e-01 -1.14013803e+00
-3.22570264e-01 -7.07522511e-01 8.30539107e-01 -7.99760520e-02
-7.66129196e-02 -1.28418791e+00 3.71126324e-01 -4.31041688e-01
5.20893991e-01 3.98199588e-01 9.27528918e-01 -8.63828063e-01
9.57186669e-02 -2.99708962e-01 2.46849172e-02 9.35193419e-01
3.29702534e-02 4.60940301e-02 -7.60319889e-01 -2.08363488e-01
7.20007062e-01 2.06628144e-02 7.43176699e-01 3.96857440e-01
1.09290373e+00 -3.43976676e-01 2.81003207e-01 4.40537483e-01
1.56873286e+00 -2.24768445e-01 5.05929887e-01 1.81869313e-01
-1.78605709e-02 1.03054464e+00 5.77684522e-01 4.91577685e-01
-1.31971702e-01 4.33618486e-01 4.40911740e-01 4.70166504e-01
6.51692569e-01 -2.18642186e-02 4.09697741e-01 7.02803910e-01
1.69987977e-01 1.56236961e-01 -6.07725561e-01 2.58969873e-01
-1.52275443e+00 -1.16389012e+00 -4.12598588e-02 2.77923250e+00
1.08318627e+00 2.84277260e-01 3.35581064e-01 3.53451729e-01
9.75472152e-01 -1.53999448e-01 -2.14327887e-01 -7.08226979e-01
-1.17045492e-01 9.08149928e-02 4.86655623e-01 6.88488305e-01
-1.05460978e+00 2.31850684e-01 6.64717388e+00 1.03236485e+00
-9.05276656e-01 4.13514767e-03 4.37050134e-01 3.63470942e-01
-5.76762378e-01 -3.09981197e-01 -5.44126153e-01 6.52653754e-01
9.91350591e-01 -5.14284730e-01 -4.25029136e-02 9.42596734e-01
1.27290234e-01 7.04274550e-02 -1.30454600e+00 8.33686352e-01
-2.03812003e-01 -9.09003496e-01 -1.80210173e-01 2.97181994e-01
3.42140973e-01 -5.10329962e-01 2.38423929e-01 -3.08420863e-02
8.92282128e-02 -1.26680005e+00 7.52735138e-01 1.02408457e+00
8.19584489e-01 -1.02241206e+00 8.19104433e-01 3.93671632e-01
-8.90084088e-01 2.12526515e-01 -4.00718808e-01 1.94340408e-01
1.75405130e-01 1.21682966e+00 -1.73146725e-01 5.02967298e-01
3.46907884e-01 2.70597637e-01 -2.66504567e-02 1.07052314e+00
1.83735237e-01 1.64855719e-01 -4.69600916e-01 -2.39099458e-01
-1.34971276e-01 -6.59770191e-01 7.85922170e-01 1.32286453e+00
6.23473227e-01 -2.81279683e-01 -3.14162105e-01 1.11778092e+00
-2.27770805e-02 4.21037674e-01 -9.86725450e-01 3.78549993e-02
2.30184689e-01 1.00545049e+00 -5.45022726e-01 1.13423757e-01
-1.55146807e-01 7.06077397e-01 9.84207019e-02 -8.07270333e-02
-8.30767393e-01 -7.83883810e-01 8.36213648e-01 -5.61044440e-02
1.10236801e-01 -2.18225345e-02 -2.15223715e-01 -1.02763700e+00
2.82377243e-01 -4.34391648e-01 4.07469958e-01 -2.92924970e-01
-1.81475091e+00 4.79101598e-01 1.14945367e-01 -1.24439406e+00
-7.20950030e-03 -9.03255343e-01 -7.14696646e-01 1.03634620e+00
-1.12327647e+00 -5.17808557e-01 2.51831144e-01 7.62895584e-01
-3.31989825e-01 -1.02945911e-02 9.41193104e-01 3.34455706e-02
-3.41524392e-01 5.74023306e-01 2.52408415e-01 -1.89222708e-01
5.53957880e-01 -1.67782712e+00 -2.27350459e-01 5.22636533e-01
-1.88884750e-01 5.96445322e-01 1.15434313e+00 -3.22019845e-01
-9.34489608e-01 -9.83618796e-01 9.96336102e-01 -3.95458192e-01
1.01259589e+00 -1.23268671e-01 -8.64924252e-01 4.05066967e-01
-1.35185584e-01 4.21604775e-02 6.96316898e-01 -8.27432647e-02
-4.33664560e-01 -3.24212313e-01 -1.52404261e+00 3.97442609e-01
6.49766743e-01 -5.26542008e-01 -6.99208200e-01 2.41203874e-01
4.70221579e-01 2.86408484e-01 -1.19192970e+00 5.59432149e-01
3.95472258e-01 -1.23261130e+00 8.21119010e-01 -6.44729137e-01
-1.18955903e-01 -2.78096169e-01 -4.40906703e-01 -1.24804342e+00
-1.52291339e-02 -9.29613531e-01 2.10094929e-01 1.27568960e+00
1.80311263e-01 -8.91330361e-01 2.72601694e-01 5.46475291e-01
2.45190054e-01 -7.87150741e-01 -1.24995017e+00 -1.25358343e+00
7.99115121e-01 -4.41380769e-01 4.22781497e-01 6.08718932e-01
4.06158686e-01 -1.12935409e-01 -4.42850478e-02 -1.09617352e-01
9.43530917e-01 -4.40646917e-01 1.63760796e-01 -1.56794381e+00
-3.11382085e-01 -9.08712626e-01 -7.27571845e-01 -4.32029396e-01
2.38054767e-01 -9.37871754e-01 1.86905399e-01 -8.21990371e-01
5.25918370e-03 -3.34705681e-01 -4.85371143e-01 -3.20813149e-01
3.83548409e-01 -2.29960397e-01 -8.00695717e-02 -1.13172494e-01
-4.79210526e-01 8.52149725e-01 8.00931573e-01 1.87430173e-01
-4.05312441e-02 5.44264317e-01 -6.68951035e-01 9.90438104e-01
1.06721389e+00 -4.76175219e-01 -3.60204726e-01 3.74259919e-01
6.65236592e-01 -1.22420266e-01 3.15056622e-01 -9.62958753e-01
1.20191082e-01 -1.51386142e-01 -2.87584007e-01 -5.95591143e-02
-1.80123523e-02 -9.02395010e-01 7.47786313e-02 4.11347032e-01
-6.71455324e-01 -7.84367919e-02 -3.28811884e-01 6.14531994e-01
-9.07735750e-02 -8.54122221e-01 1.18800163e+00 6.50347993e-02
2.37873435e-01 1.95491061e-01 -4.12498206e-01 3.81705672e-01
1.18152654e+00 6.12820350e-02 -1.97942913e-01 -3.38187128e-01
-7.75596142e-01 -1.09482966e-01 4.31068152e-01 -9.19788554e-02
4.08289731e-01 -1.30064809e+00 -7.09978223e-01 1.23164438e-01
1.53442204e-01 -3.39920431e-01 -8.10433775e-02 1.00085521e+00
-3.13512385e-01 1.65063754e-01 -1.74748421e-01 -3.71512413e-01
-1.02200127e+00 8.23444247e-01 7.49497414e-01 -1.81095630e-01
-3.67693305e-01 6.26646340e-01 6.18826076e-02 -3.85894477e-01
4.72271174e-01 -4.24945086e-01 -9.64264721e-02 1.00415684e-01
6.03136301e-01 5.40440798e-01 -4.72885040e-05 -4.50661749e-01
-3.81982654e-01 6.97022736e-01 4.66421485e-01 -4.31915998e-01
9.94613647e-01 -1.49551239e-02 -4.90002751e-01 7.85346031e-01
1.49654329e+00 2.69489706e-01 -1.16794288e+00 1.60256967e-01
4.93351072e-01 -1.83571391e-02 -2.21438080e-01 -3.92086923e-01
-5.59474468e-01 7.84702718e-01 4.78180498e-01 9.01880562e-01
9.75926340e-01 1.23521253e-01 1.28316537e-01 3.32490832e-01
4.41318601e-01 -1.17501473e+00 -5.45768589e-02 3.96151274e-01
1.21524811e+00 -8.54302526e-01 -3.67867619e-01 -2.25512654e-01
-3.89499724e-01 1.29719722e+00 -5.05851880e-02 -3.40982854e-01
1.23591030e+00 5.37385702e-01 -3.47606778e-01 1.17055789e-01
-3.01990777e-01 -2.13068202e-01 3.19824964e-01 7.13856936e-01
3.64645272e-01 1.72519878e-01 -7.02997684e-01 9.26353633e-01
-2.11571917e-01 -3.52653116e-02 4.35990125e-01 5.50938785e-01
-7.36654103e-01 -1.16381788e+00 -3.45137417e-01 1.66001946e-01
-7.10977197e-01 1.15581490e-01 -1.00882053e-01 5.25450826e-01
-1.94862947e-01 7.75280118e-01 -5.92622980e-02 -1.44019216e-01
3.33233118e-01 2.26596892e-01 5.44401705e-01 -3.63526553e-01
-1.61838889e-01 -5.00009716e-01 -1.92773417e-01 -4.66112286e-01
-2.96351790e-01 -6.12298727e-01 -9.57638621e-01 -4.03853476e-01
-1.97346732e-01 4.88072902e-01 8.24426353e-01 8.37378442e-01
-2.82088965e-02 1.50824592e-01 7.03119159e-01 -5.41465104e-01
-1.53214121e+00 -7.23359823e-01 -1.30804956e+00 2.30701610e-01
5.39407670e-01 -6.14200592e-01 -1.01217401e+00 -2.46928826e-01] | [7.223840236663818, 4.048754692077637] |
49f1ca4e-33e4-4bc2-8234-d0d5adc29648 | an-asr-based-tutor-for-learning-to-read-how | 2306.04190 | null | https://arxiv.org/abs/2306.04190v1 | https://arxiv.org/pdf/2306.04190v1.pdf | An ASR-Based Tutor for Learning to Read: How to Optimize Feedback to First Graders | The interest in employing automatic speech recognition (ASR) in applications for reading practice has been growing in recent years. In a previous study, we presented an ASR-based Dutch reading tutor application that was developed to provide instantaneous feedback to first-graders learning to read. We saw that ASR has potential at this stage of the reading process, as the results suggested that pupils made progress in reading accuracy and fluency by using the software. In the current study, we used children's speech from an existing corpus (JASMIN) to develop two new ASR systems, and compared the results to those of the previous study. We analyze correct/incorrect classification of the ASR systems using human transcripts at word level, by means of evaluation measures such as Cohen's Kappa, Matthews Correlation Coefficient (MCC), precision, recall and F-measures. We observe improvements for the newly developed ASR systems regarding the agreement with human-based judgment and correct rejection (CR). The accuracy of the ASR systems varies for different reading tasks and word types. Our results suggest that, in the current configuration, it is difficult to classify isolated words. We discuss these results, possible ways to improve our systems and avenues for future research. | ['Helmer Strik', 'Catia Cucchiarini', 'Ferdy Hubers', 'Cristian Tejedor-Garcia', 'Yu Bai'] | 2023-06-07 | null | null | null | null | ['automatic-speech-recognition'] | ['speech'] | [ 2.90278465e-01 4.19960529e-01 7.34471753e-02 -4.08977121e-01
-9.80399549e-01 -7.16740966e-01 6.70145512e-01 7.72695243e-01
-7.53312111e-01 6.78019226e-01 5.38531601e-01 -8.15063477e-01
-2.46611387e-01 -3.62789899e-01 -3.37549657e-01 -2.25033965e-02
6.54487669e-01 3.20363373e-01 7.24246860e-01 -7.20928252e-01
8.34141672e-01 4.44202662e-01 -2.11828518e+00 4.52749789e-01
1.33143115e+00 1.80671543e-01 4.87447411e-01 1.26966155e+00
-3.20452571e-01 1.09960699e+00 -1.35580957e+00 -3.63596886e-01
-2.87973136e-01 -5.04017472e-01 -1.23052371e+00 -9.46836099e-02
7.13380158e-01 -2.25937113e-01 -1.48810044e-01 7.32173860e-01
6.95287943e-01 5.19810140e-01 4.86058950e-01 -3.91879886e-01
-1.04474342e+00 7.20812321e-01 1.25268295e-01 4.70684409e-01
1.25120080e+00 1.43987015e-01 5.67922413e-01 -5.54118216e-01
1.84062839e-01 1.17010176e+00 3.49851608e-01 8.90016735e-01
-9.33660865e-01 -3.63735318e-01 6.08642101e-02 2.99624979e-01
-1.05286205e+00 -7.14215398e-01 5.95064685e-02 -4.98233676e-01
1.51059437e+00 4.34894502e-01 8.52407098e-01 9.70996976e-01
-1.36277899e-01 8.45651150e-01 1.65915334e+00 -1.20069504e+00
1.98966727e-01 5.70712388e-01 6.73524559e-01 1.67197391e-01
1.72898639e-02 -7.26562589e-02 -7.88933575e-01 4.50561017e-01
2.79174984e-01 -7.81831622e-01 -5.76687872e-01 4.84868944e-01
-9.59591329e-01 8.02343726e-01 -2.67780125e-01 7.53018200e-01
-5.24714179e-02 -5.55009544e-01 2.15286344e-01 8.29864144e-01
6.52691960e-01 7.48730719e-01 -5.98984480e-01 -1.08743620e+00
-9.26463962e-01 -5.08663654e-02 1.10038197e+00 8.67628098e-01
-2.01908618e-01 -1.06466331e-01 -4.58206236e-01 1.44230402e+00
4.85462368e-01 4.68751997e-01 1.07101107e+00 -6.60855591e-01
4.70131218e-01 6.40536606e-01 -9.72707048e-02 -5.04389763e-01
-1.20197043e-01 -5.61075844e-02 2.59141438e-03 1.94891304e-01
6.17730677e-01 1.62971556e-01 -1.07959187e+00 1.17846668e+00
-1.46686640e-02 -4.28056806e-01 3.30763310e-01 6.16360128e-01
1.34955728e+00 5.65599680e-01 9.74002630e-02 -3.79780561e-01
1.07900190e+00 -9.12126660e-01 -1.08186197e+00 2.68546753e-02
1.05072451e+00 -1.26046360e+00 1.29777825e+00 7.49950647e-01
-1.40382195e+00 -6.64261460e-01 -1.03695500e+00 5.84283471e-03
-4.15759861e-01 1.64517581e-01 6.20540380e-02 1.22856271e+00
-1.54769671e+00 4.27868575e-01 -6.09701216e-01 -6.68659210e-01
-3.20474893e-01 3.59052509e-01 -1.95665076e-01 1.54242560e-01
-1.12210965e+00 1.36401415e+00 -4.95563112e-02 -1.14617027e-01
-2.60947973e-01 -3.38430047e-01 -8.50908756e-01 -1.98691830e-01
-3.44439931e-02 1.15677565e-01 2.00191474e+00 -8.76269042e-01
-2.31684494e+00 1.15358222e+00 -9.26725715e-02 -1.49239466e-01
3.74066114e-01 -3.05345595e-01 -5.42251289e-01 7.07411841e-02
2.65279952e-02 1.92984089e-01 2.53692223e-03 -6.12394631e-01
-8.84927571e-01 -3.71633202e-01 7.62689114e-02 4.41067219e-01
-2.38650423e-02 4.80226129e-01 1.76921919e-01 -5.82761765e-01
8.45588073e-02 -6.22099400e-01 3.24297279e-01 -7.30495512e-01
1.63946256e-01 -9.30968285e-01 7.65430331e-02 -1.14828527e+00
1.55520654e+00 -1.88565123e+00 -1.58643752e-01 1.03748851e-01
-7.51189888e-03 1.09188879e+00 -3.43949199e-02 7.00174272e-01
-1.17037341e-01 2.86793619e-01 2.98949271e-01 1.22694187e-01
-2.11748570e-01 2.29702611e-02 2.64491111e-01 1.38997898e-01
-5.44328503e-02 5.17173231e-01 -1.08333945e+00 -2.37064838e-01
5.02391517e-01 3.05133283e-01 1.44727275e-01 5.76372683e-01
1.99383497e-01 -3.35366987e-02 1.74299516e-02 3.20478022e-01
1.00984275e-01 3.49963099e-01 -3.14224422e-01 7.71452010e-01
-5.83468795e-01 1.30191839e+00 -9.81662810e-01 1.39709985e+00
-6.64626002e-01 1.06743145e+00 -1.91396087e-01 -9.10676003e-01
1.22258747e+00 8.02496314e-01 -5.36628246e-01 -1.00031531e+00
-6.19511455e-02 5.22940576e-01 4.91121024e-01 -8.23059797e-01
7.37018764e-01 3.64755809e-01 6.39575958e-01 3.94444227e-01
1.35698542e-01 -3.74438226e-01 2.73385197e-01 9.68580171e-02
1.17033637e+00 -1.16620235e-01 4.42300558e-01 -2.26141140e-01
8.46916318e-01 1.64721906e-01 -4.00256276e-01 9.23943222e-01
-1.54494062e-01 5.16322374e-01 4.53388952e-02 2.31701493e-01
-5.74930847e-01 -7.47135580e-01 -7.93943033e-02 1.07528734e+00
-6.77504897e-01 -2.45631531e-01 -1.12133420e+00 -6.29012048e-01
-5.09321690e-01 1.41644156e+00 -1.58766091e-01 -1.05125979e-01
-3.98144186e-01 9.99818146e-02 4.82876301e-01 4.30229664e-01
2.55002137e-02 -1.23807275e+00 -4.48929578e-01 4.59449828e-01
1.71772435e-01 -7.56550491e-01 -2.83415020e-01 2.19452102e-02
-6.46642923e-01 -8.10410678e-01 -9.27377164e-01 -1.02166533e+00
4.04005796e-01 4.13570732e-01 1.15551865e+00 4.99619067e-01
1.05377309e-01 9.27202165e-01 -1.12853098e+00 -4.16551322e-01
-1.08490932e+00 2.15217263e-01 -1.79632142e-01 -9.04995203e-01
8.42143059e-01 -9.99760032e-02 -2.33015031e-01 2.17823297e-01
-5.15170157e-01 -9.22036767e-02 5.27038336e-01 6.80566669e-01
-1.69591412e-01 -3.59264910e-01 7.69691408e-01 -7.44749904e-01
1.31120813e+00 -5.69451749e-02 -3.37570906e-01 4.46987569e-01
-8.39323759e-01 -2.24344924e-01 2.15539306e-01 -6.16546035e-01
-9.44319248e-01 -5.16269088e-01 -7.48132765e-01 2.92731732e-01
-7.15326846e-01 5.02351344e-01 1.97064221e-01 -2.00450003e-01
7.47106552e-01 3.02588284e-01 3.51574719e-01 -4.48024094e-01
6.01576117e-04 1.74767828e+00 2.85039216e-01 -5.23793101e-01
1.88280329e-01 -9.93438184e-01 -8.45502257e-01 -1.25494552e+00
-7.78074741e-01 -7.34145105e-01 -6.50934339e-01 -6.26314402e-01
6.53841555e-01 -9.06188309e-01 -7.42426872e-01 5.83299100e-01
-1.26577771e+00 -5.94120324e-01 -3.70316982e-01 8.85998368e-01
-3.11993062e-01 1.43135428e-01 -5.00618398e-01 -1.27051342e+00
-3.13214242e-01 -1.34508741e+00 8.04433405e-01 6.49158001e-01
-8.87052834e-01 -1.11476922e+00 2.32796550e-01 1.00975776e+00
4.17014927e-01 -6.46273077e-01 7.59787440e-01 -1.37878966e+00
1.20016940e-01 -4.26287912e-02 2.03232139e-01 8.37951779e-01
6.82441965e-02 2.24459052e-01 -8.10757816e-01 -5.94988540e-02
9.69968513e-02 -3.82489383e-01 2.24887908e-01 1.42042443e-01
7.30745852e-01 -2.84964502e-01 2.84357488e-01 -4.22498375e-01
8.43502462e-01 7.13148892e-01 6.59092367e-01 5.35088658e-01
4.89160001e-01 1.01693261e+00 8.44520092e-01 -3.38955462e-01
3.90572220e-01 7.06667840e-01 -4.67672169e-01 4.38235343e-01
-4.44615036e-01 -3.15767914e-01 8.89953494e-01 1.69399571e+00
-1.77974805e-01 -3.28043044e-01 -1.05943704e+00 9.44733977e-01
-1.49689865e+00 -7.27295041e-01 -6.10974669e-01 2.46158814e+00
1.11915970e+00 1.47805110e-01 2.34731629e-01 4.85730857e-01
7.33834684e-01 -1.70170605e-01 2.45925859e-01 -1.32680511e+00
2.38484025e-01 9.43474710e-01 2.59915888e-01 9.35787022e-01
-1.68572217e-01 9.53909695e-01 6.71189260e+00 8.96831274e-01
-9.17095363e-01 4.08309847e-02 3.07799309e-01 2.39759624e-01
-2.26369396e-01 -3.48445237e-01 -8.11989903e-01 4.04902577e-01
1.55012619e+00 -2.50809848e-01 4.42946404e-01 5.17243624e-01
3.82064819e-01 -5.25646448e-01 -1.04420829e+00 4.74201232e-01
3.84245604e-01 -7.86659300e-01 -2.36374423e-01 -3.70763958e-01
3.70510787e-01 -1.27279490e-01 -3.31795782e-01 5.22849441e-01
3.08846980e-01 -1.17022145e+00 5.73721230e-01 3.81794274e-01
6.85343146e-01 -6.17698073e-01 8.96990180e-01 4.13632631e-01
-5.28560579e-01 4.29207414e-01 9.23366472e-03 -5.57667136e-01
-4.36030000e-01 -7.14569539e-02 -1.27972925e+00 8.57573301e-02
4.63646889e-01 3.28656524e-01 -9.87380385e-01 9.56198215e-01
-6.24588549e-01 1.17744863e+00 -1.51400775e-01 -9.31444049e-01
4.31044139e-02 -1.93081975e-01 4.08371717e-01 1.34372354e+00
3.78438890e-01 4.97020900e-01 -2.65813679e-01 2.20414206e-01
3.96837115e-01 7.03438818e-01 -4.36731756e-01 -7.65168220e-02
7.10136056e-01 7.78935730e-01 -7.23031938e-01 -2.39840060e-01
-6.18324339e-01 7.17932224e-01 2.40719214e-01 7.36287907e-02
-1.11123651e-01 -5.81673861e-01 2.55706698e-01 3.51225585e-01
-2.85816133e-01 -6.43729046e-02 -4.45289135e-01 -7.19788551e-01
9.42703635e-02 -1.53207958e+00 -5.11168428e-02 -9.44442570e-01
-8.96788597e-01 3.77975374e-01 6.32775426e-02 -8.90828252e-01
-2.51569241e-01 -7.24273622e-01 -5.86733997e-01 1.16280341e+00
-9.43800449e-01 -3.50969851e-01 -1.46838173e-01 1.24867164e-01
1.20230973e+00 -4.93206114e-01 8.65392327e-01 2.29890291e-02
-2.90676594e-01 8.91588211e-01 -1.01386353e-01 -6.10241145e-02
6.97771430e-01 -1.70913637e+00 3.11593652e-01 7.18271434e-01
2.68430561e-01 5.88521898e-01 8.29113126e-01 -6.26924396e-01
-9.83227611e-01 -2.65980870e-01 1.27967358e+00 -4.27175671e-01
7.46635795e-01 8.06457549e-02 -1.13481069e+00 3.76648366e-01
7.09119201e-01 -1.00209212e+00 8.64652574e-01 6.49613217e-02
1.82497770e-01 2.49601051e-01 -1.09791517e+00 5.46508729e-01
8.02437067e-01 -6.35578871e-01 -1.32110000e+00 3.61567706e-01
8.49307597e-01 -6.78111434e-01 -1.22032452e+00 7.61336461e-02
3.91379684e-01 -8.31581175e-01 3.13477099e-01 -2.88810730e-01
4.40960079e-01 -2.52967235e-02 1.24050155e-01 -1.87703824e+00
1.14616565e-01 -6.13606811e-01 2.67118186e-01 1.56714761e+00
7.72477865e-01 -7.43808329e-01 3.52673888e-01 8.55849802e-01
-4.74266618e-01 -6.98216379e-01 -8.18369448e-01 -5.07580817e-01
2.18847841e-01 -4.88349468e-01 2.89361537e-01 6.74472630e-01
4.67351675e-01 5.21428287e-01 3.44567597e-01 -7.45312357e-03
-1.83593750e-01 -7.54370213e-01 5.85619032e-01 -1.20873344e+00
-1.20136417e-01 -5.80419719e-01 -5.66983521e-01 -9.92158353e-01
2.16780435e-02 -5.94133317e-01 1.87723234e-01 -1.71777511e+00
-3.56944770e-01 -3.26005965e-02 1.86041221e-01 3.85497063e-01
-3.77116174e-01 -3.80535066e-01 2.78095186e-01 -2.86190420e-01
-3.26749712e-01 1.60656050e-01 1.16464853e+00 1.17868066e-01
-3.81338924e-01 4.93675441e-01 -6.96071804e-01 4.17031884e-01
7.98755050e-01 -3.21124226e-01 -4.86833066e-01 -3.29658180e-01
9.24341902e-02 2.31009766e-01 -2.33488157e-01 -9.20196116e-01
3.20586711e-01 -5.97976223e-02 1.57412857e-01 -4.71799016e-01
-2.48905167e-01 -5.32570183e-01 -3.42197686e-01 1.68568954e-01
-8.66202712e-01 3.56056482e-01 3.10148835e-01 -2.88041532e-01
-3.46862823e-01 -1.03345191e+00 5.98451555e-01 -5.85313067e-02
-3.92814249e-01 -7.95984626e-01 -1.06076598e+00 1.96993187e-01
9.31759655e-01 -4.29518849e-01 -5.57371080e-01 -6.29243910e-01
-6.45102441e-01 3.23805720e-01 3.79049510e-01 7.22239852e-01
6.54008627e-01 -7.03660786e-01 -6.93887174e-01 9.70493332e-02
1.78111285e-01 -1.12610562e-02 -1.34704858e-01 7.64461935e-01
-8.51065874e-01 6.52807057e-01 4.99345139e-02 -4.20223415e-01
-1.93768632e+00 -6.94948062e-02 2.16977090e-01 4.08600830e-02
-3.50717872e-01 7.57601798e-01 -7.01833427e-01 -4.75177020e-01
5.86120129e-01 -4.32987362e-01 -1.01449144e+00 1.95993915e-01
7.78994977e-01 7.51193702e-01 5.20992398e-01 -6.34797215e-01
1.88297167e-01 4.21091825e-01 -4.68128234e-01 -4.85772371e-01
1.06716537e+00 -1.68518007e-01 6.21311069e-02 1.06372499e+00
6.69583142e-01 4.21723753e-01 -2.22790301e-01 1.88576072e-01
2.29661316e-01 -3.36809874e-01 9.36176851e-02 -1.32131863e+00
-4.29701328e-01 7.57057607e-01 7.13420630e-01 7.44710863e-01
8.10367644e-01 -2.74921507e-02 3.17511111e-01 3.80710930e-01
9.51636881e-02 -1.52142894e+00 -2.92323589e-01 7.34899759e-01
8.01194429e-01 -1.21226466e+00 -1.99268371e-01 -3.86940271e-01
-6.49301469e-01 1.21725368e+00 8.74673784e-01 3.22960347e-01
1.68831617e-01 -9.10796374e-02 4.51016545e-01 -3.78672034e-02
-8.14473689e-01 -4.38620389e-01 4.99271035e-01 7.88214684e-01
1.32340658e+00 2.38043606e-01 -1.23009503e+00 3.58090758e-01
-7.98131645e-01 -7.04664886e-02 1.08554626e+00 9.91284847e-01
-7.49635696e-01 -1.22579837e+00 -5.20442903e-01 6.14924133e-01
-6.16813421e-01 -1.25784382e-01 -8.65557313e-01 8.66399407e-01
-3.85721028e-01 1.63341951e+00 -1.13447376e-01 -4.12286222e-01
8.26763928e-01 3.27638656e-01 5.21981597e-01 -1.39617109e+00
-1.07911634e+00 -3.71295184e-01 5.47096550e-01 -1.38730109e-01
-5.11299372e-01 -7.99500167e-01 -8.21161807e-01 -5.75244576e-02
-7.55188882e-01 5.19326508e-01 8.91700327e-01 1.21450365e+00
-3.10794979e-01 7.05013692e-01 3.28748614e-01 -1.12964481e-01
-7.56479919e-01 -1.53633463e+00 -2.04156160e-01 -4.97314781e-02
1.93918735e-01 -2.10980907e-01 -6.06969774e-01 -2.70460099e-01] | [11.907395362854004, 8.56702995300293] |
7d91a860-5244-4328-a2ce-56df911c155b | imputing-missing-observations-with-time | 2201.05634 | null | https://arxiv.org/abs/2201.05634v1 | https://arxiv.org/pdf/2201.05634v1.pdf | Imputing Missing Observations with Time Sliced Synthetic Minority Oversampling Technique | We present a simple yet novel time series imputation technique with the goal of constructing an irregular time series that is uniform across every sample in a data set. Specifically, we fix a grid defined by the midpoints of non-overlapping bins (dubbed "slices") of observation times and ensure that each sample has values for all of the features at that given time. This allows one to both impute fully missing observations to allow uniform time series classification across the entire data and, in special cases, to impute individually missing features. To do so, we slightly generalize the well-known class imbalance algorithm SMOTE \cite{smote} to allow component wise nearest neighbor interpolation that preserves correlations when there are no missing features. We visualize the method in the simplified setting of 2-dimensional uncoupled harmonic oscillators. Next, we use tSMOTE to train an Encoder/Decoder long-short term memory (LSTM) model with Logistic Regression for predicting and classifying distinct trajectories of different 2D oscillators. After illustrating the the utility of tSMOTE in this context, we use the same architecture to train a clinical model for COVID-19 disease severity on an imputed data set. Our experiments show an improvement over standard mean and median imputation techniques by allowing a wider class of patient trajectories to be recognized by the model, as well as improvement over aggregated classification models. | ['Jennifer Hadlock', 'Qi Wei', 'Sevda Molani', 'Andrew Baumgartner'] | 2022-01-14 | null | null | null | null | ['irregular-time-series'] | ['time-series'] | [ 1.60237581e-01 1.86094232e-02 -3.11533719e-01 -4.33686197e-01
-1.12920320e+00 -4.48141277e-01 2.30826005e-01 1.82628274e-01
-2.00325102e-01 9.56381083e-01 3.87440026e-01 -4.45749238e-02
-4.44490463e-01 -7.64640749e-01 -8.07729423e-01 -8.65712285e-01
-3.71292442e-01 6.48776054e-01 -5.80624700e-01 1.81174930e-02
-3.43274102e-02 2.16702506e-01 -1.16689622e+00 4.66254592e-01
7.72142708e-01 6.34413064e-01 -5.20311236e-01 4.56792027e-01
3.21407169e-01 5.87238133e-01 -7.75478005e-01 -1.76229209e-01
7.58468136e-02 -4.05641049e-01 -5.72647333e-01 -4.58732367e-01
2.31521651e-01 -1.07075997e-01 -3.06866676e-01 3.54363859e-01
5.60995281e-01 5.05456142e-02 7.10652053e-01 -1.44576848e+00
-6.58237040e-01 6.56806052e-01 -5.26200354e-01 -1.06279731e-01
2.11324111e-01 6.61664596e-03 4.99319822e-01 -3.24972719e-01
6.69140995e-01 6.46477163e-01 1.30253792e+00 4.44323242e-01
-2.02039289e+00 -6.69438899e-01 -2.97073841e-01 -8.55064839e-02
-1.40449238e+00 -3.58983248e-01 5.36115110e-01 -6.09869599e-01
1.11835861e+00 4.07407463e-01 5.92443109e-01 1.45459378e+00
8.77016187e-01 1.42930999e-01 8.29378605e-01 3.99911925e-02
3.33961010e-01 -4.19003636e-01 3.39245468e-01 1.15062771e-02
-4.34179641e-02 1.66608512e-01 -5.74240506e-01 -6.38749361e-01
4.41446126e-01 4.93511587e-01 -1.45966962e-01 -2.57366121e-01
-1.67710197e+00 7.69477129e-01 3.17763895e-01 9.30305645e-02
-4.44453746e-01 3.20260584e-01 4.59315419e-01 4.72745270e-01
5.99242389e-01 3.28385472e-01 -5.30470490e-01 -8.97842497e-02
-1.31192720e+00 3.31734627e-01 6.94229364e-01 5.34530699e-01
4.78025854e-01 -3.85418199e-02 -4.59773958e-01 4.64504629e-01
-3.35441381e-01 4.86499190e-01 5.33651948e-01 -1.25641155e+00
3.51036906e-01 2.33780295e-01 2.77308553e-01 -6.76363707e-01
-7.39876151e-01 -5.98875940e-01 -1.40892410e+00 4.05649953e-02
6.48829579e-01 -3.21447283e-01 -1.02886856e+00 2.37425137e+00
1.24409497e-02 6.88536406e-01 9.40106437e-02 6.06003582e-01
3.53207976e-01 5.22909284e-01 1.12591451e-03 -4.40573603e-01
1.08684576e+00 -3.01245093e-01 -6.84580505e-01 2.26007983e-01
7.34750628e-01 -3.76622587e-01 5.73536634e-01 3.61823618e-01
-9.54913497e-01 -2.91517675e-01 -7.76796281e-01 -1.72962278e-01
-4.08770978e-01 -2.72111744e-01 3.88002515e-01 3.61240208e-01
-8.97352695e-01 9.84115362e-01 -1.37468326e+00 2.35190094e-02
3.22034359e-01 7.38021195e-01 -4.50959772e-01 2.87023038e-01
-1.29208529e+00 7.30721712e-01 -1.64102077e-01 1.44655094e-01
-5.88957965e-01 -1.17412579e+00 -7.63979316e-01 -7.16424920e-03
-4.73552763e-01 -1.12480187e+00 8.28044593e-01 -8.54026675e-01
-8.23719263e-01 5.80393493e-01 -5.21236479e-01 -6.44598782e-01
4.23260778e-01 2.22696975e-01 -5.85772276e-01 -5.08744061e-01
7.97035545e-02 4.54779983e-01 5.43820798e-01 -6.17551386e-01
7.98364133e-02 -5.33632278e-01 -5.88446319e-01 -2.61223048e-01
-1.62233505e-02 -3.31848919e-01 2.68978894e-01 -8.85321438e-01
1.71975628e-01 -9.93360102e-01 -2.43783370e-01 -2.63079137e-01
-7.03043938e-01 1.75186023e-01 3.88854563e-01 -9.09888566e-01
1.28928018e+00 -1.97900307e+00 3.11119616e-01 2.93340147e-01
3.76320541e-01 -5.95732987e-01 -2.06455395e-01 6.91896558e-01
-6.44907892e-01 -2.70376168e-02 -7.33597696e-01 -7.51873434e-01
-1.06313966e-01 4.04034942e-01 -4.89953816e-01 6.35274470e-01
9.91013795e-02 8.33936691e-01 -6.34255469e-01 -6.90925196e-02
3.49088721e-02 9.07318830e-01 -6.32560730e-01 -1.09143518e-01
5.26664034e-02 1.00199401e+00 -7.57446960e-02 3.13919872e-01
8.02035809e-01 -2.93106675e-01 -6.33627027e-02 8.99946541e-02
-4.54859026e-02 3.80445510e-01 -1.02165902e+00 1.81707716e+00
-5.24330497e-01 5.85152030e-01 -4.45807040e-01 -9.43518579e-01
7.24102557e-01 6.03399396e-01 1.02779400e+00 -4.50150639e-01
-1.90802887e-01 2.25775629e-01 -2.34995197e-04 -3.74650925e-01
1.53364092e-01 -2.91483879e-01 -4.14530009e-01 6.05069041e-01
-4.61644642e-02 4.22990561e-01 -2.63674200e-01 -2.33362064e-01
1.31665516e+00 8.52591172e-02 -4.68791798e-02 -4.64529581e-02
-4.89852801e-02 1.94411166e-02 8.11328173e-01 7.64829278e-01
9.52680856e-02 1.05432701e+00 6.96882725e-01 -5.63645065e-01
-1.30274165e+00 -1.35000277e+00 -7.57810950e-01 5.61803639e-01
-4.31977868e-01 -1.12671949e-01 -6.90338433e-01 -2.95653552e-01
2.73986042e-01 7.02396512e-01 -1.08315969e+00 -9.21663791e-02
-6.16358519e-01 -1.36383283e+00 7.20865667e-01 4.46924746e-01
-2.18364775e-01 -1.02345288e+00 -4.99425024e-01 3.81061882e-01
-6.03752553e-01 -3.49574357e-01 -4.56250191e-01 6.13829553e-01
-1.06172132e+00 -7.70495534e-01 -6.34081781e-01 -5.88546395e-01
5.42955399e-01 -4.37606871e-01 1.14792335e+00 -2.86583126e-01
-2.25183681e-01 -2.13289320e-01 9.07414183e-02 -1.14546597e-01
-2.92534858e-01 3.72089773e-01 1.60145923e-01 -2.24874597e-02
6.29158318e-01 -1.00004554e+00 -7.75835276e-01 1.02638602e-01
-7.93788314e-01 1.62409004e-02 3.36692482e-02 1.11531472e+00
7.43794858e-01 -3.90767485e-01 9.63610888e-01 -6.12986028e-01
3.83710891e-01 -1.01742601e+00 -3.32300276e-01 4.55188975e-02
-4.43096995e-01 2.43013024e-01 8.44897866e-01 -4.51454043e-01
-4.05916810e-01 1.16072431e-01 -2.01121837e-01 -4.68302131e-01
-8.78694504e-02 5.37614107e-01 2.03758717e-01 3.47901940e-01
5.71417570e-01 2.69347697e-01 2.90355802e-01 -4.98885453e-01
1.67651847e-01 8.60221267e-01 7.07778096e-01 -3.11964154e-01
3.16179186e-01 4.84728277e-01 1.61778510e-01 -3.24532390e-01
-5.39414585e-01 -3.37106586e-02 -7.01297283e-01 2.64015883e-01
8.83377194e-01 -9.70605850e-01 -1.08612001e+00 4.63331789e-01
-1.03536522e+00 -5.18054128e-01 -5.15861392e-01 6.88983083e-01
-8.49071980e-01 -9.63109136e-02 -6.90053880e-01 -5.72500229e-01
-3.60092670e-01 -9.98027205e-01 1.14963222e+00 -2.32195586e-01
-6.78792000e-01 -1.13460100e+00 6.19084954e-01 -1.86368767e-02
4.16336805e-01 9.20093358e-01 1.00318611e+00 -7.77410388e-01
-3.18361968e-01 -3.39060754e-01 3.73931229e-01 -6.15046546e-02
2.26540521e-01 -1.46557137e-01 -1.01785707e+00 -2.25056529e-01
1.86999798e-01 2.24143416e-02 1.02896380e+00 8.77754986e-01
1.37546468e+00 -3.59603554e-01 -5.09697318e-01 9.82051671e-01
1.27446973e+00 -5.80823310e-02 7.37563848e-01 1.84935872e-02
4.77316767e-01 4.80039835e-01 -1.85730103e-02 4.21680599e-01
5.38781762e-01 8.11147392e-01 1.46993726e-01 -2.16769561e-01
2.21229583e-01 -2.09004834e-01 2.93669641e-01 8.32134724e-01
9.80167538e-02 -1.52159154e-01 -8.35847199e-01 8.03219020e-01
-1.84289098e+00 -1.34518683e+00 -3.43080759e-01 2.77977252e+00
1.02391648e+00 -2.06871048e-01 1.88541114e-01 1.29520699e-01
5.17001629e-01 -6.68373927e-02 -7.70334661e-01 -5.41165590e-01
-1.43730745e-01 5.29334903e-01 5.93913555e-01 5.40391028e-01
-1.11668062e+00 7.90294036e-02 6.70446777e+00 5.27808011e-01
-1.09811449e+00 1.67652607e-01 8.50469291e-01 -4.28677261e-01
-2.39651501e-01 -1.67412713e-01 -2.62407482e-01 8.57128203e-01
1.52965164e+00 -1.65343001e-01 6.12298906e-01 1.00649789e-01
5.83890021e-01 1.86484650e-01 -1.39351046e+00 6.82037711e-01
-2.63851345e-01 -1.25647771e+00 -3.98746699e-01 1.23610787e-01
6.95380926e-01 3.51234883e-01 1.96983203e-01 1.52568296e-01
3.50963295e-01 -1.45356417e+00 2.74958998e-01 9.33792531e-01
8.93763900e-01 -7.36039579e-01 6.60479844e-01 4.45367664e-01
-7.22922921e-01 -3.85743938e-02 -8.24727565e-02 -1.24731466e-01
3.56293678e-01 7.87743270e-01 -4.19137776e-01 5.85853457e-01
4.97696579e-01 8.26580048e-01 -1.44485891e-01 1.12604821e+00
4.15892661e-01 5.77916324e-01 -6.05664253e-01 5.74492276e-01
-3.18636119e-01 -3.13902199e-01 3.91239077e-01 8.42739165e-01
7.24352598e-01 -8.79937336e-02 -2.12455064e-01 9.84157860e-01
5.86125962e-02 -4.52235460e-01 -5.96969962e-01 6.66725099e-01
5.93389750e-01 7.80172944e-01 -2.28654400e-01 -9.03234631e-02
-2.94296980e-01 8.44249785e-01 1.01675026e-01 4.67069209e-01
-1.00302601e+00 -3.97166580e-01 8.94370019e-01 2.16441214e-01
-1.95160247e-02 -2.09454507e-01 -7.57873654e-01 -1.23839927e+00
5.84317856e-02 -6.10217988e-01 6.32704794e-01 -8.92263472e-01
-1.57058144e+00 6.33135200e-01 -1.88314885e-01 -1.50085998e+00
-7.56959319e-01 -5.51545843e-02 -5.76000690e-01 1.32414222e+00
-1.12223613e+00 -1.08989656e+00 -4.07857411e-02 5.99089384e-01
-8.20110291e-02 4.10588771e-01 1.34368014e+00 3.55744541e-01
-6.74718440e-01 6.94935203e-01 7.25826740e-01 2.68792748e-01
1.02376235e+00 -1.17118454e+00 5.48251867e-01 4.42588240e-01
-3.41704190e-02 8.30365360e-01 7.71842599e-01 -7.71050155e-01
-1.09287477e+00 -1.44896710e+00 1.23944533e+00 -8.85985255e-01
4.21272069e-01 -5.54016232e-01 -1.10971558e+00 1.09700274e+00
8.93002823e-02 -1.34747904e-02 9.39787626e-01 1.89955443e-01
-2.77569264e-01 -2.76045859e-01 -1.32256949e+00 4.57456589e-01
6.96297050e-01 -4.77920860e-01 -6.44346535e-01 3.33297521e-01
5.96325755e-01 -5.08693278e-01 -1.48035502e+00 2.97116876e-01
8.83727670e-01 -8.28053594e-01 1.07701480e+00 -7.09853768e-01
4.37004149e-01 -3.10899764e-01 -7.38887861e-03 -1.46514714e+00
-1.39334023e-01 -6.18512809e-01 -8.61318931e-02 8.63111973e-01
6.16180539e-01 -9.05265689e-01 7.33105004e-01 8.05499673e-01
-1.01717234e-01 -7.47838557e-01 -1.36827934e+00 -5.92215836e-01
6.03268802e-01 -4.89139766e-01 9.29548264e-01 1.12123156e+00
3.73667449e-01 -2.62623340e-01 -6.40625179e-01 1.55781731e-01
7.85725892e-01 3.18819493e-01 3.09751034e-01 -1.23886180e+00
-3.67788374e-01 -8.69822204e-02 -3.03283066e-01 -4.98675704e-01
1.16889611e-01 -1.12869680e+00 -2.75370985e-01 -1.21075869e+00
3.77330661e-01 -6.45011246e-01 -5.50392091e-01 7.94833124e-01
1.94891065e-01 6.39889240e-01 -2.36544892e-01 4.21728820e-01
-6.39962181e-02 3.62123847e-01 9.40499485e-01 6.38963655e-02
-4.91912961e-01 2.64152527e-01 -5.08926690e-01 2.14120850e-01
6.75489247e-01 -8.62803817e-01 -7.53146634e-02 -4.91067529e-01
1.72459945e-01 7.73171127e-01 7.62171507e-01 -8.30894887e-01
2.15487853e-01 3.32609974e-02 8.02734792e-01 -5.97198427e-01
3.49407256e-01 -7.30846465e-01 1.03447020e+00 4.60574389e-01
-5.22967041e-01 2.63165474e-01 1.77759096e-01 3.76968265e-01
-1.52660176e-01 4.00652736e-01 4.72531527e-01 4.13303465e-01
3.44719231e-01 3.29936206e-01 -2.96040237e-01 -4.35348526e-02
7.44899869e-01 -2.14719340e-01 -3.97796810e-01 -3.50952297e-01
-1.33930314e+00 3.31355780e-01 5.58983862e-01 3.03477407e-01
2.27730617e-01 -1.47731769e+00 -9.44599330e-01 5.92967749e-01
9.61231515e-02 -9.54571441e-02 5.23681998e-01 1.32876384e+00
-2.42346874e-03 2.78187752e-01 -2.21600905e-01 -8.13152254e-01
-7.12745428e-01 6.08738005e-01 6.42691731e-01 -2.67952085e-01
-7.10345149e-01 6.21858872e-02 -1.24403246e-01 -6.71144307e-01
-4.88209017e-02 -6.26691043e-01 1.40403837e-01 1.02496795e-01
3.06040853e-01 3.13973427e-01 4.14793551e-01 -4.69482690e-01
-3.58848304e-01 6.19767427e-01 3.23605686e-01 -6.42290041e-02
1.57692790e+00 -3.76488119e-02 -2.90127248e-01 6.97684944e-01
1.29596794e+00 -7.57434219e-02 -1.35917330e+00 2.61423200e-01
-1.70172021e-01 4.62026820e-02 -3.45170557e-01 -1.09985304e+00
-8.91405165e-01 7.67125249e-01 7.05543518e-01 2.42064878e-01
1.26450956e+00 -3.39410096e-01 7.99961627e-01 -1.39447898e-01
2.50218242e-01 -4.41239446e-01 -7.90325284e-01 2.33080685e-01
6.19072020e-01 -7.40023553e-01 -4.59441513e-01 1.78304315e-01
-2.95752406e-01 1.02009130e+00 -1.04626775e-01 -4.09401804e-01
6.74057245e-01 6.34595454e-01 4.92360517e-02 6.06095269e-02
-9.07118618e-01 3.57043207e-01 7.50012249e-02 6.41660511e-01
6.47826970e-01 8.60135406e-02 -2.54098922e-01 9.64601278e-01
-4.15489167e-01 4.98230487e-01 6.77800298e-01 4.40402836e-01
3.60001385e-01 -1.23935938e+00 -4.26413327e-01 7.70700037e-01
-5.88395000e-01 -1.09932184e-01 2.09850028e-01 6.38522148e-01
1.75880030e-01 7.36301422e-01 7.12731302e-01 -2.22320199e-01
2.42638856e-01 4.08585757e-01 2.27555409e-01 -1.76391929e-01
-7.64043391e-01 -1.57407112e-02 -1.17634498e-01 -4.88704950e-01
-2.14062557e-01 -9.78933275e-01 -1.26644766e+00 -5.48482895e-01
3.82315405e-02 7.66129419e-02 4.42698777e-01 8.14862669e-01
7.52889812e-01 7.05095947e-01 6.73780382e-01 -6.82410836e-01
-5.37282526e-01 -8.76914799e-01 -5.41440964e-01 5.43735325e-01
1.03517771e+00 -3.43581945e-01 -4.64239419e-01 1.24557331e-01] | [7.105462074279785, 3.341601848602295] |
7ad7db10-85a9-441b-ac21-fcf56b084915 | position-agnostic-multi-microphone-speech | 2010.11875 | null | https://arxiv.org/abs/2010.11875v2 | https://arxiv.org/pdf/2010.11875v2.pdf | Scene-Agnostic Multi-Microphone Speech Dereverberation | Neural networks (NNs) have been widely applied in speech processing tasks, and, in particular, those employing microphone arrays. Nevertheless, most existing NN architectures can only deal with fixed and position-specific microphone arrays. In this paper, we present an NN architecture that can cope with microphone arrays whose number and positions of the microphones are unknown, and demonstrate its applicability in the speech dereverberation task. To this end, our approach harnesses recent advances in deep learning on set-structured data to design an architecture that enhances the reverberant log-spectrum. We use noisy and noiseless versions of a simulated reverberant dataset to test the proposed architecture. Our experiments on the noisy data show that the proposed scene-agnostic setup outperforms a powerful scene-aware framework, sometimes even with fewer microphones. With the noiseless dataset we show that, in most cases, our method outperforms the position-aware network as well as the state-of-the-art weighted linear prediction error (WPE) algorithm. | ['Sharon Gannot', 'Haggai Maron', 'Ethan Fetaya', 'Yochai Yemini'] | 2020-10-22 | null | null | null | null | ['speech-dereverberation'] | ['speech'] | [ 3.81874591e-01 -3.38428795e-01 7.52691925e-01 -2.89073408e-01
-9.01012778e-01 -4.87599373e-01 4.35951561e-01 -2.62020856e-01
-5.75484991e-01 3.82963985e-01 4.69379485e-01 -6.10589266e-01
-4.15596306e-01 -4.15604770e-01 -7.01890349e-01 -9.01137471e-01
-1.39821082e-01 8.42195451e-02 6.31953180e-02 -3.84859830e-01
-2.05185339e-01 5.14959216e-01 -1.71298540e+00 1.42963737e-01
4.99576926e-01 1.16604209e+00 4.20527101e-01 1.09610808e+00
2.81068802e-01 5.01081824e-01 -9.05214429e-01 3.35765891e-02
5.00322700e-01 -1.59944832e-01 3.27170268e-02 -2.18156159e-01
7.64446557e-01 -3.55155133e-02 -3.96401912e-01 8.82939100e-01
1.28552449e+00 6.00750089e-01 2.97498107e-01 -5.05645633e-01
-3.61541778e-01 7.06423402e-01 -3.85633446e-02 5.22111773e-01
5.77374510e-02 1.58298109e-02 7.55061030e-01 -1.20948672e+00
-9.60671455e-02 1.11292291e+00 1.01913905e+00 4.35774505e-01
-1.04351306e+00 -5.47868192e-01 1.50958508e-01 4.04831767e-01
-1.23883522e+00 -1.03132319e+00 1.02012742e+00 -1.50895566e-01
1.14945340e+00 3.99818569e-01 3.39544237e-01 1.52527666e+00
-3.35194170e-01 5.90584993e-01 1.04224801e+00 -7.79578865e-01
5.11230409e-01 -2.08900556e-01 -1.50663763e-01 2.90782023e-02
-1.41640872e-01 5.86681604e-01 -6.63758934e-01 7.03502446e-02
3.61769378e-01 -3.75856370e-01 -6.03599668e-01 -2.56741405e-01
-1.21139801e+00 5.37127078e-01 4.31417346e-01 7.55149603e-01
-4.25236434e-01 1.48008287e-01 3.23812932e-01 3.71178478e-01
7.66001046e-01 5.65619826e-01 -7.35846579e-01 -1.70330375e-01
-1.07044709e+00 1.31970391e-01 1.00956047e+00 5.02073586e-01
1.91376060e-01 7.06595182e-01 -4.03811522e-02 1.41121614e+00
2.17306435e-01 6.52459085e-01 5.42936862e-01 -7.22561479e-01
5.15662730e-01 -4.16784972e-01 -5.35062365e-02 -9.30858493e-01
-7.87356496e-01 -1.32195854e+00 -1.17172134e+00 3.02394539e-01
3.44707459e-01 -5.58475137e-01 -7.39759803e-01 1.81350195e+00
2.80176580e-01 6.74002290e-01 8.46927911e-02 9.40931320e-01
5.73104322e-01 5.69824159e-01 -4.34999436e-01 -1.99077204e-01
7.57607222e-01 -1.23797679e+00 -8.22946310e-01 -3.88193667e-01
1.41567886e-01 -9.34276164e-01 8.24050605e-01 8.38514626e-01
-9.84977603e-01 -9.29841995e-01 -1.14919400e+00 3.34921688e-01
-4.61134970e-01 -8.65265206e-02 1.10850446e-01 1.04082286e+00
-1.39314198e+00 6.45456731e-01 -7.24177599e-01 -4.57802452e-02
1.28745800e-02 2.00969830e-01 -1.07181564e-01 -8.19737390e-02
-1.17100716e+00 8.02006125e-01 1.08822063e-01 6.53952360e-01
-9.30589378e-01 -8.55623364e-01 -7.87279129e-01 2.08941564e-01
3.98669422e-01 -6.60677075e-01 1.57433057e+00 -8.64011467e-01
-1.82628798e+00 1.38961121e-01 -1.94980010e-01 -7.84881294e-01
4.27916169e-01 -4.34289932e-01 -8.23075831e-01 -1.36975527e-01
-3.88692409e-01 1.77249834e-01 9.28380072e-01 -1.31280899e+00
-5.19741893e-01 -2.33962238e-01 -1.37488708e-01 2.01456755e-01
-5.65868974e-01 -1.00568265e-01 1.00592062e-01 -8.64367723e-01
1.21921889e-01 -5.61850429e-01 -4.32191193e-01 -1.14705272e-01
-4.82961476e-01 8.78467113e-02 5.06603181e-01 -7.08220601e-01
1.04963863e+00 -2.36352134e+00 9.76003706e-02 2.42615849e-01
-2.13484894e-02 7.09986746e-01 -4.01839375e-01 4.11413014e-01
-4.24836904e-01 -2.03160986e-01 -1.91044331e-01 -9.09166873e-01
2.00043708e-01 1.13513298e-01 -1.63704321e-01 3.83697182e-01
-2.07436740e-01 3.14926684e-01 -7.94368863e-01 6.51273504e-02
6.80654824e-01 9.90993917e-01 -4.59192485e-01 3.15019608e-01
1.29508689e-01 5.66675246e-01 2.71375299e-01 2.12863743e-01
7.36753345e-01 3.59039366e-01 -1.65710986e-01 -1.41712695e-01
-2.57577300e-01 3.74749690e-01 -1.35484493e+00 1.68906724e+00
-1.12872124e+00 1.04407310e+00 6.57411039e-01 -1.10838461e+00
9.37423408e-01 6.06369376e-01 1.64775759e-01 -7.05833733e-01
-1.07233889e-01 6.68513954e-01 3.60929281e-01 -5.71123183e-01
2.22150907e-01 -1.38827816e-01 5.53706467e-01 3.46980407e-03
9.02903602e-02 -1.05778269e-01 -2.34886721e-01 -4.80664402e-01
1.38822341e+00 -4.06758428e-01 1.91522270e-01 -5.57550881e-03
7.24962771e-01 -7.43666887e-01 2.01964825e-01 1.20740950e+00
-1.01855449e-01 9.22975242e-01 -2.65339404e-01 -2.38999844e-01
-1.05814052e+00 -1.00588584e+00 -2.13559449e-01 1.19972050e+00
-4.01483804e-01 6.43752739e-02 -8.56717348e-01 -1.30267054e-01
-4.05630112e-01 9.07006562e-01 -2.75760025e-01 1.26130983e-01
-7.96588361e-01 -4.80936378e-01 4.98127282e-01 3.78294200e-01
3.40732872e-01 -8.99680555e-01 -2.90080428e-01 4.54048008e-01
3.19695324e-02 -1.14653766e+00 -2.05257595e-01 5.31062484e-01
-4.10679489e-01 -6.72582328e-01 -8.94582093e-01 -8.33350420e-01
1.15815274e-01 4.17908639e-01 9.67080474e-01 -3.12118173e-01
1.28066197e-01 4.61777955e-01 -4.11321580e-01 -6.35733664e-01
-4.88470107e-01 3.29495631e-02 3.58241379e-01 2.38087550e-01
-6.17830008e-02 -1.21357453e+00 -7.87749648e-01 2.53951907e-01
-7.86774397e-01 -3.08589160e-01 5.45090675e-01 9.03514743e-01
2.73139387e-01 4.34166551e-01 7.27083504e-01 -4.24552411e-01
7.78470755e-01 -3.27990085e-01 -5.30968010e-01 -8.38192701e-02
-2.84230769e-01 -2.20408976e-01 1.12644064e+00 -5.86284876e-01
-1.06146443e+00 -1.26594469e-01 -8.54036272e-01 -4.52350110e-01
-4.96437281e-01 4.73327965e-01 -3.58996540e-01 -9.87339839e-02
9.95965660e-01 1.59663707e-01 -5.34374654e-01 -9.72957909e-01
2.36124262e-01 1.02943885e+00 7.52340913e-01 -1.22029595e-01
7.97826648e-01 3.09684247e-01 9.77616981e-02 -1.10722530e+00
-6.86740458e-01 -7.41465509e-01 -3.91955823e-01 -1.45620048e-01
3.50979209e-01 -7.93138802e-01 -3.44858050e-01 6.42625690e-01
-1.31117547e+00 -4.34350044e-01 -3.14145237e-01 7.26601243e-01
-4.78528947e-01 1.36612162e-01 -3.27301532e-01 -1.30037189e+00
-1.92869395e-01 -9.76059198e-01 1.03012478e+00 -1.55712023e-01
1.44393265e-01 -1.19599867e+00 3.08851123e-01 2.13747755e-01
1.00036740e+00 -3.56151015e-02 5.44240713e-01 -9.45568085e-01
-7.92449638e-02 -1.92093372e-01 1.64576620e-01 9.30729330e-01
-1.20722153e-03 -5.32935262e-01 -1.59488678e+00 -2.65565902e-01
5.22183716e-01 1.08833745e-01 8.87894392e-01 6.15136027e-01
1.27024817e+00 -2.55550742e-01 2.15709388e-01 6.56492352e-01
1.27983093e+00 3.09276909e-01 3.60861450e-01 2.36760527e-01
6.47085607e-01 5.76036096e-01 -4.94367257e-02 2.48832241e-01
-1.36711612e-01 8.86823237e-01 6.25517786e-01 -3.02498400e-01
-5.22140145e-01 -8.63714889e-02 3.00599545e-01 1.14070404e+00
1.69082820e-01 -7.01005220e-01 -7.60486901e-01 7.66821623e-01
-1.58144569e+00 -8.16577196e-01 -1.32276919e-02 2.12063241e+00
4.78276938e-01 -9.14577097e-02 -2.25598782e-01 6.67577446e-01
6.08723760e-01 4.24496949e-01 -4.88524258e-01 -3.71869594e-01
-3.96054476e-01 5.89043438e-01 2.43651867e-01 7.20842779e-01
-1.16440976e+00 3.33872646e-01 6.18310595e+00 9.54606593e-01
-1.03371751e+00 4.25901890e-01 3.65312040e-01 -2.93173790e-01
-4.72446270e-02 -6.81334615e-01 -3.63577515e-01 2.63470203e-01
1.34393048e+00 4.09784466e-01 8.19728136e-01 7.73051560e-01
5.99183083e-01 3.07593584e-01 -1.29888928e+00 1.24649608e+00
2.55250782e-01 -1.02977872e+00 -5.42280674e-01 -2.20400896e-02
7.24925339e-01 3.92191678e-01 4.70980227e-01 2.74393588e-01
-7.16975033e-02 -1.12598944e+00 7.46695280e-01 4.60200965e-01
3.45842838e-01 -6.11675978e-01 1.01108634e+00 4.93579865e-01
-8.02776337e-01 -3.97811830e-01 -3.76559317e-01 -2.83992201e-01
3.00271571e-01 9.80471909e-01 -9.22252655e-01 4.74941999e-01
8.16402078e-01 4.14823234e-01 -2.44622946e-01 1.51451433e+00
-2.30192885e-01 9.04244840e-01 -5.71142673e-01 3.08254994e-02
2.91228592e-01 1.98966399e-01 9.85087454e-01 1.40082085e+00
7.29722977e-01 -1.79034323e-01 -1.66473761e-01 4.32100564e-01
-1.10337384e-01 5.13192415e-02 -6.72006845e-01 5.27072191e-01
4.59267348e-01 1.07164299e+00 -7.48528391e-02 6.85664415e-02
-2.78703809e-01 5.84330618e-01 1.40192658e-01 6.69898689e-01
-4.94043857e-01 -4.27268326e-01 7.38094091e-01 -5.44673651e-02
7.87002206e-01 -3.43903840e-01 -2.57516056e-01 -6.83755159e-01
2.42973328e-01 -1.02102900e+00 -2.61227071e-01 -9.99002159e-01
-1.43997133e+00 9.68859434e-01 -2.05827817e-01 -1.20714200e+00
-4.81510907e-01 -1.02205610e+00 -5.86395502e-01 9.56921518e-01
-1.62948203e+00 -7.95984983e-01 -1.00274188e-02 4.35568005e-01
7.38765121e-01 -2.79831856e-01 9.04360712e-01 7.43070304e-01
-4.70477074e-01 5.91558754e-01 5.71951032e-01 4.87223342e-02
5.14411807e-01 -1.21950912e+00 4.93622720e-01 1.05348539e+00
6.16441309e-01 4.86206651e-01 1.06903362e+00 1.42063767e-01
-1.04162395e+00 -1.12021542e+00 7.60022521e-01 -2.66180009e-01
6.56883299e-01 -7.82950997e-01 -7.98391998e-01 2.51934469e-01
3.16879302e-01 3.00555527e-01 6.24809265e-01 3.21670026e-01
-3.14231336e-01 -4.29738998e-01 -9.01221514e-01 6.06843531e-01
1.26604354e+00 -4.80805904e-01 -4.87560928e-01 6.14211619e-01
8.90935302e-01 -5.76817572e-01 -5.07043004e-01 3.57094556e-01
4.47674155e-01 -1.35017276e+00 1.18510413e+00 -2.49414131e-01
5.88819645e-02 -2.89538056e-01 -6.42210305e-01 -1.86549938e+00
-1.33097872e-01 -6.93221986e-01 -2.86387026e-01 1.17398357e+00
6.15286827e-01 -8.24024618e-01 4.97678339e-01 -4.14530188e-02
-6.66465044e-01 -6.41197622e-01 -1.53177714e+00 -9.89358902e-01
8.53260085e-02 -1.24900138e+00 6.69297457e-01 4.32560772e-01
-5.56854427e-01 3.09089512e-01 -5.84507883e-01 5.44415414e-01
4.66594249e-01 -5.19577265e-01 4.99630332e-01 -9.87385392e-01
-8.18343520e-01 -5.14633179e-01 -4.48366672e-01 -1.33573556e+00
2.14876473e-01 -3.50105345e-01 4.70124304e-01 -1.38428867e+00
-6.98297441e-01 -3.97139281e-01 -4.78035927e-01 6.40258491e-02
3.76987010e-02 1.58727378e-01 1.63483292e-01 -3.65013182e-01
-4.67076749e-01 5.88679910e-01 8.05146456e-01 -2.14534208e-01
-1.39507860e-01 3.56439561e-01 -5.12481570e-01 8.90065253e-01
7.73195863e-01 -3.47451687e-01 -3.68079692e-01 -8.33074450e-01
2.55442888e-01 -1.57444000e-01 4.78289604e-01 -1.61345983e+00
5.58107138e-01 5.59556425e-01 3.06751877e-01 -4.33710992e-01
7.50647664e-01 -1.12890828e+00 1.13381676e-01 1.57669201e-01
-4.94600594e-01 -2.49924988e-01 2.63004422e-01 7.19655037e-01
-2.84542114e-01 -2.12802663e-01 5.30559301e-01 1.41709864e-01
-2.89603382e-01 -5.21665215e-02 -4.39761281e-01 -1.92483023e-01
2.80393988e-01 -6.89839497e-02 -9.08412561e-02 -7.54694998e-01
-8.01859736e-01 -2.37948596e-01 -2.24888951e-01 3.76583636e-01
5.43247640e-01 -1.14323521e+00 -7.53588438e-01 3.38908017e-01
-1.99674129e-01 1.43826514e-01 5.21703660e-01 6.62050545e-01
-1.05407648e-01 5.25427580e-01 3.69556308e-01 -6.95847809e-01
-1.15627789e+00 5.83547652e-01 8.57997894e-01 -6.59100264e-02
-2.69166201e-01 1.14103270e+00 1.61640540e-01 -8.67197752e-01
6.83597863e-01 -5.05490363e-01 -1.93102092e-01 -4.09128517e-01
7.26964533e-01 6.22906327e-01 5.85199475e-01 -6.35217071e-01
-1.51463568e-01 4.98474866e-01 4.10587996e-01 -4.55733627e-01
1.49190879e+00 -2.24881783e-01 1.54430836e-01 7.04124868e-01
1.36121583e+00 3.82786065e-01 -1.19313955e+00 -4.20167297e-01
-2.63714969e-01 -3.30472201e-01 3.67747247e-01 -1.04441094e+00
-9.70990121e-01 1.24357009e+00 9.95401382e-01 5.25305510e-01
1.43224573e+00 -3.35910290e-01 6.23211026e-01 6.40902877e-01
3.74212563e-01 -7.51135826e-01 -1.95785865e-01 6.53903067e-01
9.72030520e-01 -1.14726377e+00 -5.79092622e-01 -9.53295231e-02
1.05323806e-01 1.05985618e+00 3.15645814e-01 4.46424857e-02
8.98894370e-01 3.57570916e-01 3.42185229e-01 2.25033984e-01
-6.04327440e-01 -2.31430903e-01 2.05548108e-01 1.01902032e+00
2.98416644e-01 2.11731791e-02 3.82263988e-01 4.88428056e-01
-6.90868378e-01 -4.28802669e-01 1.28779724e-01 4.46796924e-01
-3.14669877e-01 -1.01064956e+00 -8.39589477e-01 1.21467605e-01
-4.64497060e-01 -4.63006914e-01 -3.24796349e-01 4.56264287e-01
3.18102330e-01 1.63968718e+00 -2.96557974e-02 -4.82133925e-01
8.13682139e-01 -5.99652976e-02 2.50370651e-01 -4.92228955e-01
-8.97995055e-01 2.87993193e-01 2.42198125e-01 -3.63629341e-01
-4.15080339e-01 -5.06060004e-01 -5.19063473e-01 -1.73895732e-02
-4.31236923e-01 3.20502557e-02 1.06934714e+00 9.16779399e-01
2.67866939e-01 1.06757808e+00 7.52062559e-01 -1.32117903e+00
-7.57916331e-01 -1.31315672e+00 -5.60952723e-01 -2.90255137e-02
1.08977211e+00 -3.05067539e-01 -8.62331688e-01 -3.86060864e-01] | [15.090444564819336, 5.823320388793945] |
07a4d152-f88a-43a0-a8a1-5823c5bf6407 | evaluating-the-utility-of-document-embedding | 1907.08184 | null | https://arxiv.org/abs/1907.08184v1 | https://arxiv.org/pdf/1907.08184v1.pdf | Evaluating the Utility of Document Embedding Vector Difference for Relation Learning | Recent work has demonstrated that vector offsets obtained by subtracting pretrained word embedding vectors can be used to predict lexical relations with surprising accuracy. Inspired by this finding, in this paper, we extend the idea to the document level, in generating document-level embeddings, calculating the distance between them, and using a linear classifier to classify the relation between the documents. In the context of duplicate detection and dialogue act tagging tasks, we show that document-level difference vectors have utility in assessing document-level similarity, but perform less well in multi-relational classification. | ['Timothy Baldwin', 'Jingyuan Zhang'] | 2019-07-18 | null | null | null | null | ['document-embedding'] | ['methodology'] | [ 1.82990789e-01 5.09882212e-01 -2.91168034e-01 -3.91906768e-01
-3.55457276e-01 -6.21590257e-01 9.28284407e-01 8.61218154e-01
-7.93608189e-01 4.66111869e-01 7.68354237e-01 -5.17857909e-01
-4.00343418e-01 -8.50975096e-01 -2.93197408e-02 -3.09765786e-01
-2.15894952e-01 5.55089056e-01 2.41624430e-01 -6.09646082e-01
4.97484982e-01 4.47461337e-01 -1.28763378e+00 6.11217916e-01
4.96746302e-01 6.14701986e-01 -2.93305784e-01 6.04772329e-01
-4.65769589e-01 7.56689906e-01 -7.87450016e-01 -6.64367318e-01
-1.37535170e-01 -5.07873893e-01 -1.28378332e+00 -2.87919611e-01
1.14717245e-01 3.84732969e-02 -4.29718703e-01 7.11728215e-01
3.97460580e-01 2.42835209e-01 9.56347644e-01 -9.47058856e-01
-8.02138329e-01 8.25556517e-01 1.53466314e-02 4.16811079e-01
9.03926969e-01 -8.15718234e-01 1.53838682e+00 -8.12022209e-01
6.85084701e-01 1.23886406e+00 8.45975280e-01 3.25297058e-01
-1.38327563e+00 -2.44972005e-01 -2.18511328e-01 3.18803042e-01
-1.26950014e+00 -3.59585494e-01 6.08630180e-01 -3.99239421e-01
1.58091807e+00 2.69678682e-01 4.11005586e-01 1.15710783e+00
2.48415217e-01 2.59747893e-01 9.96032000e-01 -1.05459070e+00
-1.78714603e-01 4.70568091e-01 5.19992888e-01 5.56878567e-01
2.21791923e-01 -1.46789402e-01 -4.75478470e-01 -3.61547291e-01
2.01948315e-01 -1.58160910e-01 -3.25687021e-01 -2.73720950e-01
-1.00836039e+00 1.15203667e+00 2.25636005e-01 1.15985525e+00
-1.95581898e-01 -3.04230839e-01 6.99084163e-01 7.05623865e-01
7.82804906e-01 1.01664209e+00 -3.71909946e-01 -2.03317150e-01
-5.75132906e-01 1.59926713e-01 1.08837926e+00 4.68252510e-01
5.44340253e-01 -5.22469044e-01 -2.34988123e-01 1.11881983e+00
8.14460441e-02 -3.66659492e-01 1.18960428e+00 -5.25598645e-01
3.06152701e-01 6.73630118e-01 -2.13063791e-01 -1.14798629e+00
-5.62742174e-01 1.36218131e-01 -4.13514555e-01 -1.76936314e-01
2.03297555e-01 1.21611491e-01 -1.13351978e-01 1.38702011e+00
1.39533028e-01 -1.56339452e-01 6.06292188e-01 4.82167810e-01
7.60118902e-01 3.55407655e-01 -6.70145303e-02 -4.43883330e-01
1.35329592e+00 -6.65298998e-01 -8.81227314e-01 4.41873781e-02
1.31168878e+00 -8.11406016e-01 8.28051507e-01 -1.18167400e-02
-9.10485506e-01 -5.20717680e-01 -1.13977706e+00 -2.01338738e-01
-9.56076801e-01 -2.25809515e-01 7.28622913e-01 5.95783353e-01
-1.01324975e+00 9.72141564e-01 -2.41000444e-01 -7.11910784e-01
-1.68395728e-01 1.62722230e-01 -5.97490132e-01 5.49802221e-02
-1.57073200e+00 1.54063296e+00 5.40608644e-01 -2.03765437e-01
3.00953180e-01 -4.07760412e-01 -9.62087333e-01 3.47098783e-02
3.70460972e-02 -1.55383706e-01 7.06543863e-01 -2.70459563e-01
-1.24331415e+00 1.00954771e+00 -1.12711444e-01 -4.82946277e-01
-2.78545022e-02 5.33173122e-02 -7.09998310e-01 -3.50144110e-03
-4.93279137e-02 3.35913301e-01 3.72047842e-01 -8.62226009e-01
-3.76304984e-01 -3.75392735e-01 2.00926319e-01 2.67286390e-01
-9.05280054e-01 2.06096396e-01 1.06384858e-01 -4.42857563e-01
2.05892414e-01 -7.11996615e-01 2.85501759e-02 -5.36837041e-01
-2.80512869e-02 -1.06543946e+00 6.10882759e-01 -4.84524488e-01
1.42048299e+00 -2.16795206e+00 1.65682048e-01 2.00441048e-01
3.20864737e-01 5.06758153e-01 -2.68210530e-01 9.79730666e-01
-2.75600195e-01 3.54144961e-01 1.21803150e-01 -1.48193046e-01
-5.30354632e-03 5.18650889e-01 -3.36388141e-01 1.68837413e-01
3.33064854e-01 8.47098589e-01 -9.33853030e-01 -6.17938876e-01
3.96963507e-02 5.17014742e-01 -3.48794192e-01 2.09306702e-01
3.47207189e-01 -3.90913814e-01 -1.31967634e-01 -2.56442223e-02
1.79271162e-01 2.21232370e-01 7.28897512e-01 -2.01436356e-01
4.93341163e-02 7.78366148e-01 -8.11497569e-01 1.33318007e+00
-7.23691225e-01 1.02838421e+00 -5.42458713e-01 -1.33025050e+00
1.13256907e+00 4.59156692e-01 3.90703499e-01 -6.32979155e-01
5.55029996e-02 -2.69465167e-02 4.10577893e-01 -6.10288680e-01
9.27150369e-01 -2.35083297e-01 -2.61565119e-01 5.63805223e-01
3.09288740e-01 -5.15509583e-02 2.74460465e-01 2.11179942e-01
1.33702624e+00 -1.32499471e-01 6.20461822e-01 -5.79329468e-02
6.59585774e-01 -6.23067468e-02 -1.04027450e-01 3.41815084e-01
-2.10328922e-02 3.64943415e-01 7.44546056e-01 -3.10369283e-01
-8.62827063e-01 -1.03587723e+00 -5.79173803e-01 1.19804728e+00
2.35055350e-02 -9.48875070e-01 -2.76523083e-01 -8.93774509e-01
2.64790088e-01 7.21502900e-01 -8.53497207e-01 -4.35802847e-01
-6.45750105e-01 -6.46731257e-01 8.96335959e-01 5.18440783e-01
-2.15254650e-01 -7.94866621e-01 -3.42050433e-01 4.34811592e-01
-1.19451247e-01 -1.14039934e+00 -1.45610705e-01 6.20249987e-01
-8.03704202e-01 -1.05467391e+00 -2.72673577e-01 -9.43392754e-01
3.20638776e-01 1.09064676e-01 1.18148613e+00 1.82488635e-01
-2.68141955e-01 3.55815023e-01 -8.57076406e-01 -5.87623455e-02
-7.31462538e-01 2.48846754e-01 1.85448617e-01 -4.12061661e-01
6.09092653e-01 -5.17587543e-01 1.15889661e-01 1.24925837e-01
-9.22294021e-01 -7.30785787e-01 3.80667806e-01 1.00313962e+00
-5.73872682e-03 2.25419682e-02 5.71918786e-01 -9.32292223e-01
1.48066533e+00 -4.54586923e-01 3.24168742e-01 4.20840234e-01
-1.01912796e+00 4.72321481e-01 6.55392826e-01 -4.42243904e-01
-6.83865249e-01 -4.72945482e-01 -3.05318475e-01 6.21700585e-02
-1.26216456e-01 5.29299855e-01 3.17901343e-01 2.09563095e-02
7.61232615e-01 4.41119410e-02 -2.67398804e-02 -3.58446270e-01
7.34735429e-01 1.00434756e+00 1.03449300e-01 -1.59290239e-01
3.81483674e-01 -5.35029545e-02 1.88366529e-02 -9.63307142e-01
-7.23431826e-01 -8.04758489e-01 -1.06021297e+00 8.27636942e-02
9.90479231e-01 -3.36544067e-01 -3.60839695e-01 -4.41188902e-01
-1.48316884e+00 2.57296562e-01 -3.82323235e-01 5.89374185e-01
-3.04084748e-01 5.22700071e-01 -4.98766780e-01 -6.60029709e-01
4.98564169e-02 -5.21904469e-01 6.35600984e-01 -1.90597087e-01
-9.04906154e-01 -1.47082686e+00 5.46742439e-01 2.19133735e-01
2.05065578e-01 -1.57823205e-01 1.28616095e+00 -1.37749803e+00
3.84915739e-01 -5.47165096e-01 -5.59120476e-02 3.81154180e-01
2.28584021e-01 -2.30081007e-01 -8.71461809e-01 1.17986396e-01
-6.54623434e-02 -3.60760927e-01 8.47913861e-01 -2.90637016e-01
6.39433920e-01 -1.20849505e-01 -5.62664747e-01 8.52580145e-02
1.10981894e+00 2.11533785e-01 5.50510466e-01 3.44462723e-01
4.79713500e-01 9.84448671e-01 4.58497405e-01 1.37066916e-01
2.66444445e-01 1.01108348e+00 -2.02765882e-01 3.63576293e-01
-1.98618293e-01 -2.17604116e-01 1.89730942e-01 1.07799339e+00
-1.11281360e-02 -3.02474022e-01 -7.11387813e-01 5.02705395e-01
-1.68578029e+00 -1.07815993e+00 -2.32109845e-01 1.77117825e+00
9.06347573e-01 1.78591192e-01 8.85353312e-02 4.91965711e-01
4.34019208e-01 4.27692115e-01 2.60438263e-01 -1.17494464e+00
-4.03302871e-02 3.66370529e-01 3.43124121e-01 5.96683383e-01
-6.88791275e-01 9.80420470e-01 7.10101175e+00 5.03042459e-01
-7.47896016e-01 1.48876414e-01 2.09319741e-02 2.38671556e-01
-4.52802241e-01 -9.09692943e-02 -5.05759895e-01 1.56105012e-01
1.21625996e+00 -1.91249058e-01 5.19668460e-02 4.19384450e-01
-3.18322718e-01 1.22273806e-02 -1.34968328e+00 6.10822499e-01
5.29140055e-01 -1.10221791e+00 -7.59939896e-03 1.83015287e-01
2.10669547e-01 -3.61926794e-01 -2.15936929e-01 4.49555755e-01
1.35326877e-01 -1.04782665e+00 -8.82332623e-02 3.52614313e-01
2.58113682e-01 -6.54496074e-01 1.13661587e+00 6.38316125e-02
-1.01144874e+00 1.49728283e-01 -6.57757163e-01 -4.07654345e-01
-2.65393443e-02 3.36583704e-01 -1.27878809e+00 6.81418836e-01
3.58314216e-01 5.56697309e-01 -6.35623276e-01 4.19192225e-01
-9.71401930e-02 3.35159212e-01 7.96379372e-02 -4.41577971e-01
2.28393033e-01 -3.28179777e-01 1.61621466e-01 1.51266980e+00
8.50583613e-02 2.49298513e-02 -3.05167943e-01 4.37099278e-01
5.39427549e-02 4.27777052e-01 -9.95571315e-01 -3.50309312e-01
4.99333769e-01 1.27833629e+00 -5.61425924e-01 -2.88930744e-01
-5.05192339e-01 1.33131242e+00 7.66675770e-01 -2.33496115e-01
-5.37930846e-01 -7.54245222e-01 7.63524532e-01 -2.33913109e-01
3.26756179e-01 -3.29233378e-01 -1.11787476e-01 -9.35648024e-01
-5.76071814e-02 -3.40017498e-01 3.18000942e-01 -4.69550461e-01
-1.49518871e+00 9.26227391e-01 4.23279889e-02 -9.00723279e-01
-6.11057103e-01 -8.31594646e-01 -5.73535085e-01 8.97007227e-01
-1.14023030e+00 -8.80170882e-01 2.17722088e-01 1.71213552e-01
2.17738301e-01 -1.81557477e-01 1.39271033e+00 1.34235367e-01
-2.92566448e-01 7.56934702e-01 1.59384519e-01 4.27309752e-01
8.22136641e-01 -1.43017364e+00 1.90311566e-01 1.49163872e-01
8.89716506e-01 7.58137763e-01 7.14481711e-01 -3.00202966e-01
-8.65835786e-01 -6.12827539e-01 1.87345350e+00 -7.41685927e-01
1.07453585e+00 -2.85205305e-01 -1.05671036e+00 3.62015039e-01
3.95447046e-01 -8.36278871e-02 1.20340335e+00 7.36667275e-01
-7.28715062e-01 1.76788986e-01 -1.04910016e+00 5.00559032e-01
1.28013813e+00 -1.02324724e+00 -1.34456706e+00 4.08891708e-01
7.82318056e-01 2.88794097e-02 -1.38261271e+00 1.04602553e-01
6.05737925e-01 -1.02438021e+00 1.04696572e+00 -1.00293529e+00
5.25790632e-01 3.04524809e-01 -2.70868331e-01 -1.65266860e+00
-3.57214361e-01 -1.08182199e-01 2.47720658e-04 1.42650270e+00
5.59364676e-01 -7.09993422e-01 6.16442442e-01 2.08431169e-01
-9.48066562e-02 -7.90191829e-01 -1.06508470e+00 -1.06049681e+00
3.97963107e-01 -1.78640217e-01 2.97620952e-01 1.15330577e+00
8.98177803e-01 7.92030394e-01 -1.41895667e-01 -1.24777220e-01
-6.73060343e-02 3.23197655e-02 1.72444955e-01 -1.42430496e+00
-2.56221890e-01 -6.17725551e-01 -9.19264555e-01 -7.66505539e-01
7.73393452e-01 -1.19744611e+00 -8.90263915e-02 -1.46706367e+00
-1.79974467e-01 -1.79089636e-01 -2.83108741e-01 1.31039575e-01
-1.01705059e-01 2.86985129e-01 4.68785912e-02 2.25511685e-01
-4.16596740e-01 4.12073642e-01 5.56535423e-01 -1.19763620e-01
-8.17237981e-03 -2.70603150e-01 -5.86347818e-01 4.71875042e-01
7.27859199e-01 -6.00886524e-01 -2.81482011e-01 -1.24441378e-01
1.92174807e-01 -1.59459263e-01 -3.90077457e-02 -7.90376425e-01
5.25632240e-02 2.56444216e-01 2.11058021e-01 -1.07477143e-01
3.91607255e-01 -8.24926853e-01 -4.60771203e-01 4.57131535e-01
-1.00150144e+00 3.80145133e-01 -1.75686195e-01 4.68409538e-01
-5.37099063e-01 -8.66657436e-01 2.48605937e-01 5.51062450e-02
-5.61212003e-01 -3.45299035e-01 -6.29849195e-01 5.17491288e-02
8.33741486e-01 -1.82240099e-01 -2.48007908e-01 -1.81856051e-01
-7.43737161e-01 -2.25494042e-01 5.29536866e-02 7.77682483e-01
6.02407753e-01 -1.48712790e+00 -3.99441123e-01 1.85815454e-01
5.13815522e-01 -8.01764131e-01 -2.80487239e-01 6.22794330e-01
-2.71452777e-02 7.54192233e-01 -8.68469626e-02 -1.30781472e-01
-1.67074645e+00 5.96727192e-01 1.31020010e-01 -3.76512945e-01
-3.03238064e-01 8.20812464e-01 -5.74022055e-01 -5.55144727e-01
2.77327560e-02 -3.06154877e-01 -6.05002344e-01 7.60181069e-01
4.18257773e-01 2.85439372e-01 3.49687278e-01 -7.32291162e-01
-4.98027295e-01 4.97648835e-01 -1.94906771e-01 -2.63654500e-01
1.31497908e+00 4.35268283e-02 -2.62773305e-01 7.21396208e-01
1.69337833e+00 1.72368914e-01 -1.60465449e-01 -3.17958742e-01
6.69017494e-01 -3.63471717e-01 -1.64869297e-02 -3.46436024e-01
-2.63911694e-01 8.06682229e-01 2.88429677e-01 1.11505401e+00
4.64682907e-01 3.77372056e-01 5.64095795e-01 5.79905629e-01
2.47696303e-02 -1.06546175e+00 1.68420225e-01 7.50184178e-01
6.69681966e-01 -1.18849063e+00 3.83346155e-02 -3.60296786e-01
-5.62025726e-01 1.54348588e+00 2.83773869e-01 -6.93216920e-02
6.58744812e-01 1.36369407e-01 2.10448261e-02 -2.98878461e-01
-8.18823338e-01 -4.06766623e-01 4.45107669e-01 9.95230794e-01
1.10001779e+00 -1.52081504e-01 -8.56153131e-01 3.27008992e-01
-2.77140856e-01 -4.62078363e-01 7.27114499e-01 7.62834787e-01
-3.64023864e-01 -1.65858066e+00 6.52804673e-02 6.01542175e-01
-3.28635901e-01 -4.23018724e-01 -8.71387362e-01 7.23196626e-01
1.07152916e-01 1.07371855e+00 4.46926415e-01 -7.69782126e-01
5.25094867e-01 5.80217600e-01 5.74012399e-01 -9.37118828e-01
-7.86591053e-01 -6.06955528e-01 7.06499100e-01 -2.63424784e-01
-5.23769081e-01 -6.07892334e-01 -9.38474476e-01 -1.19771637e-01
-4.82042640e-01 5.76483250e-01 5.25202692e-01 1.08955395e+00
1.35073230e-01 5.77992320e-01 6.32846415e-01 -4.08940285e-01
-7.78751075e-01 -1.29618680e+00 -4.71795291e-01 6.53048456e-01
-2.16089655e-02 -5.53700089e-01 -5.15595019e-01 -4.38567728e-01] | [10.533355712890625, 8.768109321594238] |
d6cd2540-8562-4eba-ba1f-97e2a25b512c | assisting-clinical-decisions-for-scarcely | 2307.03315 | null | https://arxiv.org/abs/2307.03315v1 | https://arxiv.org/pdf/2307.03315v1.pdf | Assisting Clinical Decisions for Scarcely Available Treatment via Disentangled Latent Representation | Extracorporeal membrane oxygenation (ECMO) is an essential life-supporting modality for COVID-19 patients who are refractory to conventional therapies. However, the proper treatment decision has been the subject of significant debate and it remains controversial about who benefits from this scarcely available and technically complex treatment option. To support clinical decisions, it is a critical need to predict the treatment need and the potential treatment and no-treatment responses. Targeting this clinical challenge, we propose Treatment Variational AutoEncoder (TVAE), a novel approach for individualized treatment analysis. TVAE is specifically designed to address the modeling challenges like ECMO with strong treatment selection bias and scarce treatment cases. TVAE conceptualizes the treatment decision as a multi-scale problem. We model a patient's potential treatment assignment and the factual and counterfactual outcomes as part of their intrinsic characteristics that can be represented by a deep latent variable model. The factual and counterfactual prediction errors are alleviated via a reconstruction regularization scheme together with semi-supervision, and the selection bias and the scarcity of treatment cases are mitigated by the disentangled and distribution-matched latent space and the label-balancing generative strategy. We evaluate TVAE on two real-world COVID-19 datasets: an international dataset collected from 1651 hospitals across 63 countries, and a institutional dataset collected from 15 hospitals. The results show that TVAE outperforms state-of-the-art treatment effect models in predicting both the propensity scores and factual outcomes on heterogeneous COVID-19 datasets. Additional experiments also show TVAE outperforms the best existing models in individual treatment effect estimation on the synthesized IHDP benchmark dataset. | ['Chenyang Lu', 'Philip Payne', 'Hanqing Yang', 'Neel Shah', 'Hanyang Liu', 'Ziqi Xu', 'Ahmed Sameh Said', 'Bing Xue'] | 2023-07-06 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [ 4.29746732e-02 7.37980232e-02 -8.90585899e-01 -3.70328486e-01
-5.47401547e-01 -2.20773295e-01 2.96430200e-01 -2.05924101e-02
-2.16809571e-01 1.15474975e+00 8.14483643e-01 -3.04402769e-01
-4.89222050e-01 -6.88645720e-01 -4.29727703e-01 -1.17822707e+00
2.54382759e-01 1.09211838e+00 -8.42581391e-01 3.38241756e-01
-3.15292597e-01 2.12490931e-01 -9.89216506e-01 1.98826700e-01
1.19269514e+00 7.29251027e-01 -1.48180574e-02 6.42902106e-02
1.76656902e-01 8.54495287e-01 -1.13724522e-01 -3.94399941e-01
1.95192665e-01 -4.61492717e-01 -4.38814908e-01 7.91753232e-02
1.33344959e-02 -5.90828478e-01 -5.63382268e-01 5.98300576e-01
1.03072155e+00 -1.70103654e-01 1.35423398e+00 -1.48160672e+00
-4.58284587e-01 7.21009254e-01 -3.54368001e-01 -1.39942080e-01
-1.31060973e-01 3.89339328e-01 9.25600708e-01 -5.53687155e-01
5.69468856e-01 1.07821751e+00 5.24915040e-01 8.25485587e-01
-1.40088654e+00 -6.75658822e-01 -2.26548612e-01 -8.16299841e-02
-9.89384830e-01 -2.07230344e-01 6.96933985e-01 -1.09342349e+00
5.01708448e-01 5.84566891e-02 6.56960309e-01 1.67238772e+00
6.57166243e-01 7.67900229e-01 1.10957873e+00 8.43356326e-02
3.28292340e-01 1.23565242e-01 -1.04552552e-01 3.70864540e-01
1.56222641e-01 5.09068549e-01 -2.29732260e-01 -5.65878034e-01
9.05801713e-01 7.25949109e-01 -6.04518414e-01 -7.44683444e-01
-1.32034516e+00 1.13758385e+00 2.45970085e-01 -1.34760618e-01
-1.05515075e+00 -7.44569749e-02 4.45837736e-01 -5.49138337e-02
4.16185468e-01 1.93916917e-01 -8.52323413e-01 1.35324463e-01
-9.10434902e-01 2.11272731e-01 6.05356395e-01 7.35922873e-01
1.53612882e-01 1.90887243e-01 -7.08133876e-01 8.84417534e-01
4.56912369e-01 8.10498357e-01 5.85236132e-01 -1.11478889e+00
5.34997523e-01 5.19969404e-01 2.17825651e-01 -6.38956666e-01
-5.14297903e-01 -5.65707028e-01 -1.26845670e+00 -1.98307678e-01
2.28661910e-01 -6.47183359e-01 -7.71832108e-01 2.03863811e+00
2.91253507e-01 2.55401880e-01 1.42727107e-01 7.59418607e-01
7.84425080e-01 3.23096275e-01 4.60189819e-01 -7.81802058e-01
1.27525926e+00 -6.64107859e-01 -1.13135302e+00 -7.72262067e-02
9.56972301e-01 -2.13671878e-01 4.77273077e-01 -1.03279473e-02
-9.88919914e-01 1.97249815e-01 -4.25026983e-01 3.09904724e-01
2.04590306e-01 1.66064158e-01 7.87391365e-01 3.93892944e-01
-3.64928901e-01 6.61031127e-01 -7.66313732e-01 -8.57314393e-02
9.42296267e-01 3.95605773e-01 -4.21284944e-01 -1.63964272e-01
-1.30874479e+00 7.53522992e-01 7.23404437e-02 8.96190405e-02
-1.21749568e+00 -1.29844213e+00 -8.75784159e-01 4.07763004e-01
4.49150532e-01 -1.51485324e+00 8.64668965e-01 -4.61833894e-01
-1.39473963e+00 7.20075428e-01 1.26581620e-02 -3.37058783e-01
8.42042267e-01 1.89313918e-01 -2.27877066e-01 -1.01518720e-01
5.02932034e-02 2.90921926e-01 7.10911751e-01 -1.07252586e+00
-3.05107653e-01 -5.61131358e-01 -6.13025129e-01 4.28513765e-01
-1.94775909e-01 -2.69841164e-01 1.79765165e-01 -8.40661347e-01
-1.10993765e-01 -1.13033223e+00 -4.29887980e-01 -2.07972243e-01
-3.55300725e-01 -1.59123942e-01 2.51468241e-01 -6.30514920e-01
1.14323616e+00 -2.05891347e+00 5.59850931e-01 -3.09268564e-01
4.97061253e-01 -1.10003032e-01 1.01326585e-01 4.51569051e-01
-3.69475067e-01 2.30476797e-01 -5.23221314e-01 -3.87623310e-01
2.85573512e-01 3.14325839e-01 -3.61859977e-01 7.19530046e-01
-3.32234919e-01 9.82031465e-01 -8.47153068e-01 -6.39109254e-01
4.36468810e-01 4.67422932e-01 -8.81217301e-01 5.61494708e-01
7.68492594e-02 7.28930652e-01 -5.69544613e-01 6.11151040e-01
6.77496850e-01 -5.70190012e-01 4.25121754e-01 -1.01320721e-01
4.43868428e-01 -7.86633492e-02 -8.53821695e-01 1.45598269e+00
-3.48750532e-01 2.18609199e-01 9.12782829e-03 -9.59254444e-01
3.71026874e-01 8.45251739e-01 1.15814638e+00 -2.91607916e-01
6.50432050e-01 1.56060562e-01 6.66006655e-02 -7.78320611e-01
-2.82341361e-01 -1.04896140e+00 -3.23359221e-02 3.60934049e-01
2.58905757e-02 -6.47469908e-02 -4.57156569e-01 5.34013063e-02
8.69670093e-01 -3.83728333e-02 3.29712749e-01 -2.86359221e-01
-1.19309686e-02 -3.61206561e-01 1.41882229e+00 5.75293660e-01
-4.59031075e-01 8.44023943e-01 6.37557149e-01 -3.09242576e-01
-7.67840385e-01 -1.15546811e+00 -7.38775611e-01 3.14239860e-01
-3.06746930e-01 1.54133007e-01 -3.85565847e-01 -8.45338464e-01
4.78801340e-01 9.78322446e-01 -9.54604805e-01 -2.94421166e-01
-1.76996261e-01 -1.24481356e+00 1.06828384e-01 6.70538247e-01
7.96757191e-02 -1.13258159e+00 -1.83762953e-01 4.21101272e-01
-4.09611970e-01 -7.58344173e-01 -3.99679035e-01 3.26303877e-02
-9.70600605e-01 -1.11147642e+00 -1.19901085e+00 -3.39852840e-01
4.25445735e-01 -2.61055738e-01 1.08870196e+00 -3.89785141e-01
-3.56994458e-02 1.78200305e-01 1.04933575e-01 -4.69090343e-01
-3.44909579e-01 -1.89981356e-01 5.36501467e-01 2.69967943e-01
2.87768304e-01 -6.43622458e-01 -1.10519397e+00 3.16030890e-01
-6.97423935e-01 1.19727366e-01 4.06300306e-01 1.35528970e+00
5.54854989e-01 -3.48863125e-01 9.75863695e-01 -1.21557784e+00
3.21562558e-01 -1.03238869e+00 -3.70894521e-01 2.58524060e-01
-9.67646718e-01 -2.35333592e-01 3.26333374e-01 -4.89173144e-01
-1.27884221e+00 -1.72660664e-01 2.56208777e-01 -8.47299159e-01
-4.18870673e-02 5.54502368e-01 -2.14068606e-01 7.23313868e-01
2.75300086e-01 -8.15854445e-02 8.36788118e-03 -3.94738764e-01
2.47219559e-02 9.47968781e-01 2.15162951e-02 -5.43214619e-01
2.56087333e-01 5.18287539e-01 6.85945973e-02 -1.10285230e-01
-8.00707698e-01 -1.09910339e-01 -2.69276887e-01 7.08894581e-02
1.06681025e+00 -1.23641312e+00 -1.13078499e+00 3.66635591e-01
-8.35239410e-01 -5.77574432e-01 -3.41281593e-01 1.08323765e+00
-8.70480657e-01 1.42909005e-01 -6.20097637e-01 -5.17342031e-01
-4.41945255e-01 -1.41281617e+00 9.12569880e-01 1.32246306e-02
-7.52426758e-02 -1.37449861e+00 4.06338036e-01 7.02825010e-01
2.56038129e-01 4.84745860e-01 1.25165629e+00 -5.46468735e-01
-3.58493358e-01 -1.69437483e-01 -4.59257793e-03 3.06684792e-01
4.24756348e-01 -2.18980506e-01 -6.94925249e-01 -4.97095257e-01
2.56668448e-01 -1.65366665e-01 6.54382408e-01 1.17095399e+00
1.28159034e+00 -2.27195621e-01 -4.50054973e-01 9.07046199e-01
1.36638796e+00 1.36818394e-01 4.01063800e-01 -2.10072860e-01
8.01509559e-01 7.34913290e-01 2.63558567e-01 8.20853889e-01
6.64434910e-01 5.43582857e-01 4.94970530e-01 1.03003168e-02
2.56218016e-01 -3.27519357e-01 2.42907763e-03 7.53659010e-01
7.29876906e-02 -6.55745506e-01 -1.07336402e+00 6.93468571e-01
-2.01605558e+00 -9.19351697e-01 -2.97841340e-01 2.21132612e+00
7.98926651e-01 -4.22161728e-01 -2.40996763e-01 -4.05875534e-01
6.47702813e-01 7.03056082e-02 -6.87570989e-01 -2.13996515e-01
-2.44662598e-01 -3.28677207e-01 5.22634447e-01 2.34516114e-01
-9.13051009e-01 3.95456016e-01 6.20860863e+00 6.98392153e-01
-1.02163947e+00 2.07074836e-01 9.01303530e-01 -3.43922585e-01
-3.64424974e-01 -7.66028315e-02 -3.11319441e-01 8.15462053e-01
8.59417677e-01 -2.69532710e-01 2.44318426e-01 5.54953456e-01
5.47523141e-01 2.18561634e-01 -1.38810611e+00 1.00024199e+00
-4.88997139e-02 -1.33662748e+00 -1.11491948e-01 3.71633500e-01
1.05275965e+00 6.36386052e-02 2.86287069e-01 4.74938661e-01
3.87176514e-01 -1.14358032e+00 2.47624978e-01 7.64238477e-01
1.01367760e+00 -3.86854559e-01 1.18820834e+00 2.83725590e-01
-4.45059001e-01 -4.75495934e-01 -1.39495537e-01 2.68619329e-01
5.50881624e-01 5.83670199e-01 -5.83539605e-01 6.82560444e-01
4.69155908e-01 8.89354944e-01 2.72623926e-01 8.96086991e-01
-1.69572949e-01 7.93968856e-01 -1.57698486e-02 5.85842848e-01
-1.70540124e-01 -3.63735467e-01 5.36325216e-01 5.07776201e-01
3.82733226e-01 5.27918220e-01 -3.11313284e-04 8.74183774e-01
-4.12929982e-01 1.85111597e-01 -6.64451897e-01 -1.41953006e-01
3.48957390e-01 9.26079929e-01 -1.22795245e-02 -4.20258194e-01
-3.06089133e-01 6.46570563e-01 6.68190122e-02 6.23357475e-01
-8.73184204e-01 4.37406540e-01 6.19160593e-01 7.95856863e-02
-4.77022753e-04 6.05195761e-01 -5.65522254e-01 -1.69664264e+00
-3.53195429e-01 -9.30226862e-01 8.20070028e-01 -7.63200462e-01
-1.66188383e+00 2.36695975e-01 -1.09441198e-01 -1.42233658e+00
-2.72312909e-01 -3.45994681e-01 -3.56256902e-01 1.13182044e+00
-1.44304776e+00 -9.91693616e-01 -4.94877808e-02 4.23510015e-01
2.82783717e-01 -2.19978303e-01 9.02752876e-01 4.58567828e-01
-1.06918323e+00 3.96386713e-01 4.11353648e-01 -5.54078966e-02
9.38741744e-01 -1.18567073e+00 -7.22750127e-01 7.28640631e-02
-5.99921525e-01 5.75364709e-01 6.20833814e-01 -8.48788619e-01
-1.08467352e+00 -1.06298697e+00 8.51455212e-01 -7.17776299e-01
4.12702918e-01 -1.56363472e-02 -9.51165557e-01 9.30293858e-01
1.26507103e-01 7.69865513e-02 1.23114932e+00 1.02675343e-02
1.56851232e-01 1.43211082e-01 -1.27992189e+00 4.33083624e-01
8.65304828e-01 -4.14258629e-01 -8.14823151e-01 5.40186167e-01
5.96687436e-01 -3.22528780e-01 -1.26756227e+00 7.73110569e-01
4.73254353e-01 -5.46209991e-01 7.96352386e-01 -1.27792025e+00
8.96735668e-01 2.70970792e-01 -2.87408173e-01 -1.46319580e+00
-5.33195376e-01 -5.29072702e-01 -1.97495744e-01 1.10155404e+00
1.73744872e-01 -6.77783251e-01 8.58840048e-01 1.14092064e+00
4.67619933e-02 -9.13576603e-01 -1.04513907e+00 -3.35936576e-01
4.97521043e-01 -5.17161489e-02 8.58577251e-01 1.48549688e+00
-1.63088903e-01 1.98368639e-01 -8.65028083e-01 1.25959545e-01
8.65843236e-01 3.60940903e-01 4.53748494e-01 -1.31035852e+00
-2.44824708e-01 -3.61428529e-01 6.54547513e-02 -3.85172576e-01
5.12320518e-01 -9.28585231e-01 -3.34078431e-01 -1.84705210e+00
8.09357584e-01 -4.41661358e-01 -6.97818339e-01 4.30330247e-01
-3.65974188e-01 -4.72574025e-01 6.05139472e-02 1.88983545e-01
1.28692556e-02 1.30675697e+00 1.41305673e+00 -1.97735593e-01
-2.28931308e-01 1.30121922e-02 -5.40545344e-01 7.08155811e-01
4.74684298e-01 -1.00947237e+00 -4.86768246e-01 -2.75288105e-01
6.58463314e-02 1.02562439e+00 2.18065962e-01 -3.31863910e-01
-7.71567449e-02 -8.19775403e-01 1.83922410e-01 -4.87469256e-01
1.01698563e-01 -1.03483868e+00 5.31610548e-01 6.01585329e-01
-4.28330302e-01 -2.14843974e-01 -1.55097842e-01 7.94531643e-01
5.76303788e-02 1.47821397e-01 6.00936055e-01 2.23668292e-01
1.27859354e-01 9.49644148e-01 -4.00996238e-01 2.21982628e-01
9.98103678e-01 2.24298984e-01 -2.20510587e-01 -3.76630276e-01
-9.68433678e-01 6.72629893e-01 3.99836838e-01 3.70699883e-01
4.71785069e-01 -1.53801978e+00 -1.08510005e+00 8.46616924e-02
1.24689430e-01 2.81751491e-02 1.09031689e+00 1.13026488e+00
-2.02273563e-01 3.81663710e-01 -1.43259808e-01 -4.71605510e-01
-8.60111535e-01 9.03877676e-01 6.17418945e-01 -4.76520568e-01
-5.90803564e-01 4.09604549e-01 8.69790554e-01 -6.31487906e-01
4.25845906e-02 1.17973067e-01 -3.27239543e-01 2.55553186e-01
4.19644266e-02 4.57581103e-01 -3.59973460e-01 -5.93513191e-01
-3.37557882e-01 2.63080448e-01 1.84642121e-01 2.01563284e-01
1.71775293e+00 -2.65678853e-01 -1.46231741e-01 3.17886978e-01
1.19846869e+00 -3.54873002e-01 -1.36458564e+00 -2.80545533e-01
-6.38378084e-01 -5.17849684e-01 3.77013057e-01 -9.96144295e-01
-1.33762991e+00 1.03576136e+00 6.84167564e-01 -3.30392718e-01
9.34227109e-01 -2.97818899e-01 5.82090855e-01 -2.70418257e-01
1.71365082e-01 -8.72088671e-01 -2.59325415e-01 1.88393742e-02
9.93146777e-01 -1.55846286e+00 -7.08454475e-03 -2.47591823e-01
-9.36434090e-01 4.73005205e-01 4.22160625e-01 2.36722022e-01
8.76747191e-01 -7.88419023e-02 1.78783566e-01 -2.53331244e-01
-9.70650494e-01 4.55356896e-01 3.07677925e-01 1.94254532e-01
3.12601835e-01 6.09189034e-01 -4.21232432e-01 1.24657285e+00
1.44718856e-01 2.93756366e-01 4.76527601e-01 5.49571276e-01
5.58094800e-01 -9.30139244e-01 -9.48162153e-02 9.23661947e-01
-6.13933146e-01 -2.02865712e-02 2.41377443e-01 4.67610478e-01
3.40233780e-02 6.36626184e-01 -1.99152872e-01 -1.30821876e-02
2.32750580e-01 2.12741435e-01 8.98611248e-02 -4.68410522e-01
-4.70203847e-01 1.23114698e-01 -1.78339988e-01 -3.85791302e-01
-2.93321013e-01 -8.53472292e-01 -1.07331359e+00 -2.59005010e-01
-2.80263364e-01 4.53328900e-02 3.39824349e-01 8.79196882e-01
6.31165206e-01 9.32493091e-01 8.78600001e-01 -5.86629927e-01
-9.81819749e-01 -8.30580831e-01 -9.20514405e-01 7.17609286e-01
4.93079364e-01 -9.41854894e-01 -6.82165027e-01 -2.47535869e-01] | [7.980721950531006, 5.5434794425964355] |
b96bf4d2-ef86-4f3a-a2ab-6b85887fd3ea | illinois-math-solver-math-reasoning-on-the | null | null | https://aclanthology.org/N16-3011 | https://aclanthology.org/N16-3011.pdf | Illinois Math Solver: Math Reasoning on the Web | null | ['Subhro Roy', 'Dan Roth'] | 2016-06-01 | null | null | null | naacl-2016-6 | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [-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] | [-7.2480998039245605, 3.8105437755584717] |
81ae52cf-6ce5-43d1-a046-bc3abd2d85f5 | on-crowdsourcing-design-with-comparison | 2306.01538 | null | https://arxiv.org/abs/2306.01538v1 | https://arxiv.org/pdf/2306.01538v1.pdf | On Crowdsourcing-design with Comparison Category Rating for Evaluating Speech Enhancement Algorithms | Speech enhancement techniques improve the quality or the intelligibility of an audio signal by removing unwanted noise. It is used as preprocessing in numerous applications such as speech recognition, hearing aids, broadcasting and telephony. The evaluation of such algorithms often relies on reference-based objective metrics that are shown to correlate poorly with human perception. In order to evaluate audio quality as perceived by human observers it is thus fundamental to resort to subjective quality assessment. In this paper, a user evaluation based on crowdsourcing (subjective) and the Comparison Category Rating (CCR) method is compared against the DNSMOS, ViSQOL and 3QUEST (objective) metrics. The overall quality scores of three speech enhancement algorithms from real time communications (RTC) are used in the comparison using the P.808 toolkit. Results indicate that while the CCR scale allows participants to identify differences between processed and unprocessed audio samples, two groups of preferences emerge: some users rate positively by focusing on noise suppression processing, while others rate negatively by focusing mainly on speech quality. We further present results on the parameters, size considerations and speaker variations that are critical and should be considered when designing the CCR-based crowdsourcing evaluation. | ['Sneha Das', 'Line H. Clemmensen', 'Clément Laroche', 'Angélica S. Z. Suárez'] | 2023-06-02 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 2.22424977e-03 -1.14475951e-01 5.14828026e-01 -4.04274821e-01
-9.92489219e-01 -6.02844596e-01 3.16741675e-01 6.40641749e-01
-8.89487684e-01 6.38787329e-01 6.13002598e-01 -2.78625160e-01
-2.55722135e-01 -3.54093283e-01 7.86144957e-02 -5.33505738e-01
1.85075268e-01 -4.46449742e-02 2.54288495e-01 -6.04487002e-01
3.08816820e-01 3.53286713e-01 -2.03017569e+00 3.38562399e-01
8.89150560e-01 9.97500241e-01 1.68123931e-01 9.84795630e-01
-3.37274857e-02 3.32541943e-01 -1.31288385e+00 -3.14875841e-01
-8.75240006e-03 -2.27583513e-01 -4.72199112e-01 6.06041662e-02
4.75036576e-02 -2.61942863e-01 4.43651468e-01 1.03206730e+00
1.33886826e+00 4.39159185e-01 1.65960446e-01 -1.12786186e+00
-4.04146135e-01 3.42829585e-01 2.76542723e-01 3.69958341e-01
1.02986586e+00 2.56777853e-01 7.48785675e-01 -7.92949080e-01
2.59438038e-01 1.13759589e+00 5.83941340e-01 3.43682289e-01
-1.14685369e+00 -4.86821175e-01 -2.30474368e-01 2.53795475e-01
-1.31651461e+00 -9.16793287e-01 5.92790663e-01 -4.53862548e-01
9.53806520e-01 7.69889534e-01 5.61066747e-01 7.56925106e-01
-2.23553360e-01 1.15851864e-01 1.53848338e+00 -5.97029030e-01
8.02024782e-01 7.67618120e-01 -1.00631073e-01 -1.05330385e-01
-1.36155486e-01 3.43710810e-01 -8.30106854e-01 -4.07854974e-01
2.55648613e-01 -7.58104444e-01 -3.81608903e-01 1.62189052e-01
-8.01863909e-01 5.90216458e-01 -6.42702729e-02 7.13550329e-01
-5.47484696e-01 -4.18237418e-01 5.52988768e-01 7.13445425e-01
6.68993175e-01 5.00828207e-01 -3.08575749e-01 -7.33918846e-01
-8.68944645e-01 2.19803914e-01 8.62885475e-01 4.87306952e-01
2.75384814e-01 2.05841959e-01 -2.88221836e-01 1.32734179e+00
4.18702096e-01 5.44148386e-01 5.92761457e-01 -1.01940191e+00
8.40853006e-02 2.33309343e-01 5.43233752e-01 -9.91987348e-01
-3.71151030e-01 -1.92244202e-01 -1.45429924e-01 7.67674685e-01
3.62562478e-01 -2.53552228e-01 -3.92296255e-01 1.43215764e+00
2.64113635e-01 -7.45009422e-01 -5.39947823e-02 1.16136444e+00
9.71169591e-01 4.27477181e-01 3.17932546e-01 -4.71242011e-01
1.34215391e+00 -3.73074859e-01 -1.15779901e+00 8.46115351e-02
3.61063689e-01 -1.31310356e+00 1.43623650e+00 7.25856245e-01
-1.39227545e+00 -8.35193574e-01 -1.00746465e+00 1.82147086e-01
-4.05695438e-01 -1.77237377e-01 -5.78016192e-02 1.60319602e+00
-1.41201544e+00 4.53842938e-01 -2.25831494e-01 -4.45656687e-01
-2.05207020e-01 4.17834491e-01 -3.69585484e-01 3.99076641e-01
-1.14321327e+00 9.08546805e-01 -7.94745311e-02 -9.12934914e-02
-4.57457244e-01 -3.71728033e-01 -6.69644713e-01 -3.10628046e-03
-1.07912295e-01 -1.43270632e-02 1.52886605e+00 -1.16788137e+00
-1.98011684e+00 8.77892911e-01 -2.14855805e-01 -1.51328817e-01
5.62586069e-01 1.17648998e-02 -1.07013929e+00 3.10141772e-01
5.30258156e-02 3.16116065e-01 8.31024468e-01 -1.27355778e+00
-8.62441719e-01 -2.13773817e-01 1.77839398e-01 2.73582309e-01
-1.68261066e-01 7.76594758e-01 6.68965206e-02 -6.20702267e-01
-2.15065405e-02 -5.10094821e-01 7.55628198e-02 -1.59672387e-02
9.95659083e-02 -2.11044505e-01 5.10303974e-01 -1.03314877e+00
1.35193121e+00 -2.26384926e+00 -4.44372207e-01 3.88772309e-01
2.32594740e-02 5.07579923e-01 -3.36179249e-02 4.34369057e-01
-1.46166207e-02 2.28944495e-01 1.59055829e-01 -3.38979095e-01
2.39974886e-01 -1.80549711e-01 1.51644632e-01 4.75381196e-01
-1.39410168e-01 1.11312613e-01 -9.78513956e-01 -4.76958156e-01
3.33825201e-01 6.63207710e-01 -5.15078008e-01 2.48124957e-01
4.95494515e-01 5.32413781e-01 1.60106480e-01 5.29532969e-01
6.74640536e-01 5.36879420e-01 -2.09533915e-01 -1.22118130e-01
-4.48202848e-01 4.19795275e-01 -1.35896921e+00 1.26670146e+00
-6.33016050e-01 7.51013339e-01 5.90016484e-01 -3.96909565e-01
9.92509127e-01 1.03478718e+00 1.73859537e-01 -9.58240092e-01
2.15856582e-01 4.86562222e-01 1.34288833e-01 -8.12235415e-01
7.27190495e-01 -3.56005162e-01 3.99533421e-01 5.02457976e-01
-1.23249032e-01 -3.31939995e-01 1.01381727e-01 -1.47210881e-01
7.69564569e-01 -5.77668130e-01 3.97150576e-01 -3.23532224e-01
7.10370719e-01 -4.23119575e-01 2.13201165e-01 4.18743372e-01
-8.08116317e-01 5.18597007e-01 1.34929135e-01 3.05778295e-01
-8.42061818e-01 -1.01675808e+00 -8.24486837e-02 1.28008270e+00
-1.40346080e-01 -2.06461087e-01 -1.03193593e+00 -3.37742418e-02
-4.51555997e-01 7.87661731e-01 -2.05960706e-01 7.59113729e-02
8.35702047e-02 -1.78455114e-01 5.47015071e-01 1.84741288e-01
2.51731992e-01 -1.13353109e+00 -6.43606603e-01 4.28680956e-01
-4.21903908e-01 -8.64108026e-01 -4.50374961e-01 -1.94255441e-01
-6.32978261e-01 -4.09071088e-01 -9.44768429e-01 -5.77346563e-01
1.39577791e-01 2.92511642e-01 8.37791264e-01 2.51534164e-01
1.74752884e-02 8.68265390e-01 -8.35860789e-01 -5.87452412e-01
-7.40738750e-01 -2.91901499e-01 3.61649960e-01 -7.95306731e-03
4.87266034e-01 -5.77595592e-01 -6.51351094e-01 7.48138547e-01
-8.72644126e-01 -6.08067453e-01 1.42215133e-01 3.55531067e-01
4.05947082e-02 1.81140989e-01 8.53982270e-01 -1.43028721e-01
1.44149077e+00 -8.58080089e-02 -1.88873187e-01 -9.48398411e-02
-5.33972800e-01 -6.88575923e-01 2.15608284e-01 -4.81106579e-01
-1.13395405e+00 -4.67272669e-01 -5.97302079e-01 2.28441700e-01
-4.72753882e-01 3.15646112e-01 -3.34464878e-01 -2.44675279e-01
1.05773199e+00 -2.26402327e-01 1.21068187e-01 -4.36173320e-01
-1.30231947e-01 1.57300115e+00 3.22692156e-01 -2.82159269e-01
4.47971880e-01 2.96830028e-01 -7.86625445e-01 -1.46473324e+00
-9.96117741e-02 -8.98029327e-01 -1.84984893e-01 -6.32607937e-01
8.19234669e-01 -8.02465856e-01 -9.18227017e-01 3.51158619e-01
-1.17719197e+00 -1.37525469e-01 -3.77866596e-01 7.27357447e-01
-4.03757960e-01 4.01723802e-01 -3.45209926e-01 -1.70052135e+00
-2.41760790e-01 -1.19053209e+00 1.01189327e+00 2.42218912e-01
-7.03146338e-01 -8.47195685e-01 1.17268879e-02 7.12931097e-01
7.56612360e-01 -2.16657728e-01 4.61441815e-01 -5.49304783e-01
3.47414523e-01 -3.60732019e-01 1.25300989e-01 6.79066956e-01
3.53951544e-01 -1.38296187e-01 -1.50082719e+00 -2.40106180e-01
3.52861017e-01 -8.21307525e-02 -3.87919210e-02 3.27538341e-01
6.25989437e-01 -6.05351515e-02 5.16704381e-01 -9.39253494e-02
1.06961620e+00 6.62671983e-01 8.85673702e-01 2.66685367e-01
-1.68288291e-01 1.18597174e+00 5.90776742e-01 5.55952668e-01
-9.63474065e-02 8.08412969e-01 1.33046985e-01 -2.04063594e-01
8.16903859e-02 -2.72979494e-03 6.31865621e-01 9.70959723e-01
-2.48573765e-01 -3.13534200e-01 -5.65586746e-01 4.83749121e-01
-1.17125058e+00 -8.92179072e-01 -1.03632152e-01 2.55888104e+00
6.63666606e-01 1.05527133e-01 5.85957170e-01 9.63374376e-01
1.00267446e+00 -1.49395689e-01 9.93474051e-02 -9.23543632e-01
-1.97165921e-01 4.09823269e-01 1.64931521e-01 9.22201514e-01
-6.65717363e-01 3.96385610e-01 6.63495493e+00 7.96918213e-01
-1.04854786e+00 3.82022232e-01 3.91946405e-01 -4.32200171e-02
-1.73131734e-01 -3.34987611e-01 -1.85074165e-01 4.46303755e-01
1.09200549e+00 -2.05105633e-01 5.72038651e-01 3.54640126e-01
1.17486036e+00 -4.85802203e-01 -6.73821330e-01 1.16138434e+00
-1.05041526e-01 -5.97589672e-01 -3.66758525e-01 4.50819777e-03
2.88672805e-01 -4.55155671e-01 6.83373138e-02 1.58547729e-01
-3.12089294e-01 -8.95490944e-01 1.03250992e+00 3.11519444e-01
6.50810421e-01 -4.59220797e-01 8.91094446e-01 -8.03200006e-02
-9.70323741e-01 1.71112886e-03 -2.46050268e-01 -3.50784600e-01
4.25757706e-01 6.10271633e-01 -8.90285850e-01 1.55134976e-01
8.05387676e-01 -3.83358866e-01 -3.67329240e-01 1.38624549e+00
-1.36245787e-01 7.51845121e-01 -2.19225675e-01 -2.25437373e-01
-1.70226365e-01 -5.72857782e-02 8.16470206e-01 1.24268246e+00
4.71672714e-01 2.04395443e-01 -4.25223440e-01 6.30343258e-01
3.76784533e-01 7.47694910e-01 -3.09680432e-01 -4.24679890e-02
5.29444754e-01 1.10926008e+00 -7.09245980e-01 -1.74201727e-01
-2.94584781e-01 8.16995680e-01 -5.82663834e-01 4.40749049e-01
-5.27592957e-01 -5.55194914e-01 5.53761601e-01 2.57049918e-01
-2.04429552e-01 -1.09788282e-02 -3.30684006e-01 -3.13879877e-01
1.58775792e-01 -1.22318017e+00 -9.43792462e-02 -1.01071322e+00
-1.03876829e+00 7.94416487e-01 -1.55997366e-01 -1.47585559e+00
3.38007277e-03 -5.26436210e-01 -5.62426627e-01 1.32090712e+00
-9.64386702e-01 -3.66657466e-01 -3.67205709e-01 4.50727522e-01
4.69795555e-01 -9.08408910e-02 7.33881295e-01 6.80000067e-01
-8.21178332e-02 6.00333989e-01 -1.88716486e-01 -3.25901985e-01
8.66580427e-01 -1.18256664e+00 -8.59097242e-02 3.69097352e-01
-3.11792850e-01 5.35466492e-01 1.26826847e+00 -4.29819584e-01
-5.58053076e-01 -3.77409935e-01 1.13559937e+00 -7.17210993e-02
4.52225715e-01 -6.62610531e-02 -8.27700138e-01 -3.95461440e-01
3.74094635e-01 -5.46326041e-01 1.17327631e+00 -2.60520745e-02
3.99409421e-02 -1.85684621e-01 -1.58089530e+00 4.90373820e-01
7.11330533e-01 -9.83159006e-01 -5.48974633e-01 -3.73011548e-03
5.25678098e-01 6.04450814e-02 -1.01708198e+00 -1.55652300e-01
6.74662471e-01 -1.37584400e+00 6.60864949e-01 -1.78055868e-01
-7.02620298e-02 -3.22842956e-01 -3.23967338e-01 -1.52484536e+00
2.09831689e-02 -9.32345867e-01 7.33804941e-01 1.52898860e+00
6.61943376e-01 -8.58474553e-01 4.03841168e-01 8.29135776e-01
-5.49350791e-02 -1.72128491e-02 -1.15598059e+00 -7.87771761e-01
-2.58322328e-01 -8.09732735e-01 5.66566527e-01 6.33156955e-01
4.82485861e-01 9.61400941e-02 -1.16864324e-01 1.83218762e-01
1.77610695e-01 -9.22929049e-01 6.01494670e-01 -1.13201749e+00
-2.77853876e-01 -5.12713790e-01 -6.83636427e-01 -4.01456386e-01
-5.40138721e-01 -1.71676278e-01 1.67653456e-01 -1.37976050e+00
-4.59366441e-01 -1.22922346e-01 -2.20642149e-01 -1.64627150e-01
-2.60752626e-02 2.11778671e-01 2.62904853e-01 -1.50699541e-01
-2.76518375e-01 4.50638652e-01 9.75002408e-01 7.18304813e-02
-5.77312291e-01 2.87579417e-01 -7.18876302e-01 5.71811497e-01
9.46179509e-01 -3.26484144e-01 -3.38369191e-01 -2.49612749e-01
3.74174237e-01 1.29862696e-01 4.01871622e-01 -1.12955785e+00
2.64435202e-01 1.36980593e-01 -9.12019983e-02 -6.44113198e-02
5.14993310e-01 -9.26388621e-01 5.02943322e-02 2.92501032e-01
-3.76273483e-01 1.86210856e-01 2.36460835e-01 2.57948786e-01
-4.58883226e-01 -5.67903399e-01 7.74218678e-01 9.78855789e-02
-4.41528887e-01 -4.34879571e-01 -9.57412899e-01 -1.93698615e-01
5.01931965e-01 -6.90058649e-01 -1.06108464e-01 -1.09266567e+00
-1.06372356e+00 -2.33449131e-01 3.47650945e-01 2.63194442e-01
5.42437971e-01 -1.21054852e+00 -6.48206055e-01 1.24243787e-02
2.47596651e-01 -8.29925776e-01 4.97377396e-01 1.04168415e+00
-6.48035705e-01 2.05127254e-01 -2.32885286e-01 -3.36175859e-01
-1.68111944e+00 1.77288517e-01 3.45579952e-01 2.84160674e-01
9.23743472e-02 6.96031272e-01 -2.92975932e-01 -2.50866115e-01
5.07397175e-01 -2.38067999e-01 -6.41250551e-01 2.95541823e-01
8.83170307e-01 1.10898149e+00 5.57769418e-01 -9.48856235e-01
-2.83248872e-01 2.11754456e-01 4.77911353e-01 -1.07180643e+00
8.38891089e-01 -4.32484210e-01 6.13024458e-02 5.32515347e-01
1.05685031e+00 6.04549050e-01 -5.68880618e-01 1.11734763e-01
4.77514379e-02 -6.81874812e-01 2.71355182e-01 -1.03577185e+00
-6.34003341e-01 9.59522069e-01 1.49255884e+00 7.92896628e-01
1.28088987e+00 -4.42881942e-01 4.18970019e-01 1.45472065e-01
3.73682112e-01 -1.81922746e+00 -4.96942773e-02 -1.18092485e-02
1.11142015e+00 -1.05163002e+00 -4.32528883e-01 -4.60301101e-01
-7.16858327e-01 8.45884264e-01 4.86736558e-02 4.40162301e-01
7.46838093e-01 7.83752352e-02 6.23331010e-01 2.69969776e-02
-2.51702994e-01 -5.38862050e-01 2.45695263e-01 1.17836416e+00
8.96707475e-01 1.95225805e-01 -8.63874555e-01 6.89509630e-01
-4.64768887e-01 -6.44356459e-02 4.67513025e-01 9.11951542e-01
-7.88841009e-01 -1.07647717e+00 -1.02488327e+00 1.56686798e-01
-6.34465158e-01 4.55765314e-02 -6.08949840e-01 3.86568338e-01
1.89363539e-01 2.14983535e+00 -2.13989690e-01 -5.46235561e-01
8.21647048e-01 -2.54670139e-02 7.82043710e-02 -4.46479172e-01
-1.22354531e+00 2.83679664e-01 5.04995704e-01 -2.80551225e-01
-6.87450588e-01 -4.78109121e-01 -7.29409933e-01 -3.63875270e-01
-4.40349936e-01 5.38121581e-01 1.11277974e+00 8.13127458e-01
6.41864389e-02 4.51079488e-01 5.56809127e-01 -8.72804523e-01
-4.15491581e-01 -1.25388229e+00 -9.29478109e-01 4.34411854e-01
3.33662540e-01 -4.55248326e-01 -7.62666583e-01 -9.06551927e-02] | [15.113118171691895, 5.701748847961426] |
e87c2e41-5361-4ffb-b583-af831a55c781 | binary-domain-generalization-for-sparsifying | 2306.13515 | null | https://arxiv.org/abs/2306.13515v1 | https://arxiv.org/pdf/2306.13515v1.pdf | Binary domain generalization for sparsifying binary neural networks | Binary neural networks (BNNs) are an attractive solution for developing and deploying deep neural network (DNN)-based applications in resource constrained devices. Despite their success, BNNs still suffer from a fixed and limited compression factor that may be explained by the fact that existing pruning methods for full-precision DNNs cannot be directly applied to BNNs. In fact, weight pruning of BNNs leads to performance degradation, which suggests that the standard binarization domain of BNNs is not well adapted for the task. This work proposes a novel more general binary domain that extends the standard binary one that is more robust to pruning techniques, thus guaranteeing improved compression and avoiding severe performance losses. We demonstrate a closed-form solution for quantizing the weights of a full-precision network into the proposed binary domain. Finally, we show the flexibility of our method, which can be combined with other pruning strategies. Experiments over CIFAR-10 and CIFAR-100 demonstrate that the novel approach is able to generate efficient sparse networks with reduced memory usage and run-time latency, while maintaining performance. | ['Maria A. Zuluaga', 'Francesco Galati', 'Riccardo Schiavone'] | 2023-06-23 | null | null | null | null | ['domain-generalization'] | ['methodology'] | [ 3.23872060e-01 1.57181285e-02 -2.90172428e-01 -6.40945315e-01
-6.48940355e-02 7.03957006e-02 1.35526806e-01 8.97947233e-04
-7.12341845e-01 8.45643759e-01 -3.11896831e-01 -4.49107409e-01
-5.08035541e-01 -1.04434204e+00 -5.73408544e-01 -6.06423378e-01
-3.09539717e-02 3.76954168e-01 4.65808094e-01 -2.49201581e-02
-6.10688105e-02 1.03694773e+00 -1.89268851e+00 2.65070647e-01
5.11559129e-01 1.48310423e+00 4.68681484e-01 3.05490345e-01
-1.98661700e-01 7.17245698e-01 -1.05397141e+00 -5.34295261e-01
5.94210804e-01 -1.41455799e-01 -3.93916726e-01 -2.62758881e-01
9.44955528e-01 -5.16219795e-01 -5.98217845e-01 1.30055428e+00
2.85432011e-01 7.28223920e-02 4.06037182e-01 -1.11688244e+00
-1.92084327e-01 9.97509480e-01 -2.33443886e-01 4.26555902e-01
-5.70449114e-01 -4.15532202e-01 8.91016304e-01 -5.58866441e-01
3.79439682e-01 1.20075130e+00 7.78515577e-01 5.80214620e-01
-1.23008394e+00 -8.09583008e-01 1.35344848e-01 3.47991884e-01
-1.54180181e+00 -6.82732224e-01 4.88374352e-01 1.17577620e-01
1.31779754e+00 1.31714359e-01 8.88419390e-01 8.38226318e-01
-8.07226524e-02 4.39210892e-01 5.40745497e-01 -4.62452352e-01
5.66949666e-01 -7.14332983e-02 3.46843988e-01 7.15123594e-01
1.07660103e+00 -3.92688401e-02 -5.25055408e-01 -4.88992557e-02
8.43764961e-01 1.83271706e-01 -2.08187312e-01 -2.97266454e-01
-6.65667057e-01 9.81714547e-01 6.82676971e-01 5.11609554e-01
-3.83420467e-01 4.09462214e-01 4.68313992e-01 3.49019825e-01
3.72162834e-02 3.77568230e-02 -3.40855628e-01 -1.27312377e-01
-1.54816771e+00 4.34417278e-01 8.91283095e-01 9.93296385e-01
5.54147363e-01 6.40524447e-01 -2.85781510e-02 1.02715313e+00
1.69728458e-01 4.62960124e-01 7.28842974e-01 -8.71621013e-01
5.17986655e-01 3.70498627e-01 -5.46763301e-01 -1.01087987e+00
-4.20254439e-01 -7.62166739e-01 -1.26144385e+00 2.97492176e-01
2.83428907e-01 -4.76993807e-02 -1.00517046e+00 1.63231492e+00
-9.86061990e-02 2.08715424e-02 2.07033604e-01 7.00279355e-01
9.35848415e-01 5.63601911e-01 -1.13040037e-01 -6.88518882e-02
1.31825483e+00 -8.96562219e-01 -6.23357177e-01 -4.41264719e-01
5.10774791e-01 -3.81963342e-01 6.33117259e-01 8.20613980e-01
-1.13452661e+00 -4.06170249e-01 -1.55309618e+00 -1.74722150e-01
-3.50743949e-01 2.67537087e-01 8.91846538e-01 1.12023175e+00
-1.24065220e+00 1.02196777e+00 -8.85119379e-01 -8.59235078e-02
8.70616138e-01 7.89151549e-01 -1.44627139e-01 -1.67442575e-01
-8.82983208e-01 7.93716431e-01 8.29946876e-01 1.13204353e-01
-4.80963141e-01 -6.61471844e-01 -7.80538917e-01 7.11761415e-01
2.19159424e-02 -3.92755955e-01 1.23020184e+00 -8.40147555e-01
-1.43697238e+00 3.87041390e-01 -4.66882028e-02 -1.39163482e+00
1.32848069e-01 -2.77940463e-03 -4.10533905e-01 4.50174242e-01
-4.30367142e-01 9.73083794e-01 8.65575492e-01 -5.34100354e-01
-7.53007352e-01 -2.80324548e-01 -5.32895289e-02 -1.78824678e-01
-9.12978590e-01 -2.60860592e-01 -2.17421517e-01 -8.89875174e-01
3.68166625e-01 -5.88107705e-01 -3.43253314e-01 1.73350558e-01
-1.11115225e-01 -1.73881605e-01 8.52789938e-01 -9.08651426e-02
1.36313725e+00 -2.01407599e+00 -1.28428608e-01 4.91933167e-01
3.36219132e-01 8.72043014e-01 -3.52706075e-01 -3.00142318e-01
-5.60171232e-02 -7.57873878e-02 -3.08794398e-02 -4.43457663e-01
-1.44292578e-01 5.86644828e-01 -3.43269616e-01 2.78092861e-01
1.13304332e-01 5.93759716e-01 -4.29821849e-01 -3.98231089e-01
1.28234133e-01 7.43710756e-01 -6.92120910e-01 -3.70837748e-01
-1.72277316e-01 -6.44880474e-01 4.27102949e-03 5.22506535e-01
8.83614540e-01 -7.19961803e-03 2.28805259e-01 -3.55223149e-01
2.99113989e-01 4.17053610e-01 -1.20635819e+00 1.39601994e+00
-3.62779319e-01 8.34637463e-01 4.41523567e-02 -1.39163315e+00
1.28620040e+00 2.22728495e-02 2.43278742e-01 -8.55816245e-01
3.75050664e-01 5.05184293e-01 2.58570611e-01 1.45341545e-01
5.77614784e-01 -1.35582626e-01 3.47657919e-01 8.59669000e-02
5.38426340e-01 3.50145549e-02 6.65865719e-01 -6.64980188e-02
1.07381260e+00 -7.47423768e-01 1.65209547e-01 -2.21711591e-01
3.48653138e-01 -2.88823932e-01 4.34420228e-01 7.87529945e-01
-1.18685134e-01 4.03315037e-01 5.13783991e-01 -5.00324130e-01
-1.11381495e+00 -8.22889268e-01 -4.67472464e-01 7.45054007e-01
-2.45410085e-01 -4.72888350e-01 -8.50156188e-01 -2.71745831e-01
-9.01750624e-02 4.69027013e-01 -1.74877554e-01 -2.04039648e-01
-6.00390971e-01 -8.24272633e-01 9.19339597e-01 5.54065704e-01
7.49963522e-01 -8.40408027e-01 -1.05535650e+00 2.64997333e-01
1.33450121e-01 -1.47284031e+00 1.93946138e-01 7.04918683e-01
-1.66425753e+00 -6.57398641e-01 -9.14624572e-01 -8.94367337e-01
4.42248106e-01 3.67081553e-01 1.03117180e+00 2.54662931e-01
-9.15107280e-02 -3.38262081e-01 -2.24039212e-01 -2.30065644e-01
-2.04274163e-01 4.27832246e-01 1.53539672e-01 -2.93341458e-01
5.39717436e-01 -8.34902585e-01 -4.59217966e-01 -9.08940583e-02
-1.01397049e+00 -1.55881569e-01 8.96016419e-01 8.74597430e-01
6.91207051e-01 4.40392107e-01 6.57363176e-01 -7.87723720e-01
5.86038768e-01 -1.22417465e-01 -8.80200624e-01 7.04630688e-02
-7.54125714e-01 2.51744270e-01 9.13683057e-01 -6.09095097e-01
-6.41485333e-01 1.99854523e-01 -4.91433889e-01 -5.98204434e-01
-8.81636739e-02 4.70040679e-01 -1.02255337e-01 -4.80198473e-01
7.50817060e-01 8.12191367e-02 -8.66922140e-02 -5.95462382e-01
1.24322310e-01 5.63619018e-01 5.60534894e-01 -2.66873866e-01
5.40009737e-01 3.53590667e-01 4.38519776e-01 -1.03226912e+00
-5.99817872e-01 -1.44767582e-01 -3.13271761e-01 3.10596049e-01
2.47991279e-01 -8.65835071e-01 -4.50902909e-01 2.07455039e-01
-1.21662700e+00 -3.03214312e-01 -4.70409870e-01 5.38042963e-01
-3.01226735e-01 4.21424031e-01 -7.24412680e-01 -5.93305409e-01
-5.71871579e-01 -1.06209540e+00 7.12171912e-01 2.35623658e-01
-3.32855955e-02 -5.51036775e-01 -3.93402457e-01 -4.91245613e-02
7.22001255e-01 -3.05119783e-01 1.04039598e+00 -7.12349474e-01
-5.34056723e-01 -2.51154512e-01 -6.86314464e-01 7.47671187e-01
-4.03571308e-01 -1.17928222e-01 -8.30497742e-01 -2.67552108e-01
1.73621297e-01 -3.12862247e-01 1.20003712e+00 6.12548411e-01
1.36096430e+00 -5.71305931e-01 -1.08517945e-01 9.07995164e-01
1.69398344e+00 1.95313975e-01 8.34984303e-01 2.94855744e-01
2.29755953e-01 1.33857712e-01 1.06535561e-01 5.31127334e-01
-2.14427441e-01 6.51092410e-01 5.66022038e-01 5.23045808e-02
-3.09679002e-01 8.39450806e-02 9.11005437e-02 6.46508932e-01
-1.17459573e-01 -2.44947210e-01 -6.29305124e-01 4.29862201e-01
-1.50816786e+00 -9.98147190e-01 1.26479670e-01 2.14213061e+00
7.99201846e-01 5.71201921e-01 -4.93050702e-02 8.29506636e-01
3.66513968e-01 1.58804044e-01 -3.59361440e-01 -7.35277534e-01
-1.91599756e-01 8.95310819e-01 8.18741381e-01 1.74310654e-01
-1.01940131e+00 7.04527497e-01 6.58287430e+00 1.27906561e+00
-1.30623531e+00 9.11248177e-02 6.41845882e-01 -5.10102332e-01
5.85485063e-02 -5.26449919e-01 -1.30930948e+00 2.31254682e-01
1.40768349e+00 -2.25694459e-02 4.46089715e-01 1.34589446e+00
-2.40952179e-01 1.02625623e-01 -1.02093220e+00 1.32290149e+00
5.87271899e-02 -1.55472159e+00 2.92700589e-01 -5.92184393e-03
5.89670539e-01 2.18666315e-01 3.36786844e-02 3.84995490e-01
-2.29180247e-01 -1.15590739e+00 9.06674385e-01 -8.23967680e-02
9.14097428e-01 -1.00081420e+00 9.37684298e-01 2.52001792e-01
-9.90230143e-01 -2.76880503e-01 -1.14837813e+00 -1.26792938e-01
-7.13020004e-03 1.12944829e+00 -6.82422161e-01 -4.50060144e-03
8.22802305e-01 4.36058402e-01 -5.75514793e-01 1.27503526e+00
1.26710564e-01 4.55992341e-01 -6.92491412e-01 -2.59961784e-01
2.21347764e-01 -3.67267653e-02 2.58062810e-01 1.40035319e+00
6.96341336e-01 -6.00372367e-02 -5.09692013e-01 8.61652911e-01
-2.86224842e-01 -4.18692492e-02 -6.05204761e-01 -1.16409525e-01
6.20465696e-01 1.03863394e+00 -1.05305111e+00 -4.09425437e-01
-2.48917937e-01 6.67064488e-01 4.38396037e-01 2.39043385e-01
-7.79029667e-01 -7.72413850e-01 6.34002149e-01 3.58288363e-02
8.91832948e-01 -2.49180183e-01 -5.97219229e-01 -1.00604892e+00
1.12622574e-01 -8.26840639e-01 1.01182714e-01 -2.96814173e-01
-7.58173883e-01 1.14108920e+00 -6.32918766e-03 -9.85626638e-01
-5.48157878e-02 -9.74603117e-01 4.04481627e-02 4.32948142e-01
-1.64744425e+00 -5.46221495e-01 -1.98544741e-01 5.64539135e-01
3.86844277e-01 -4.15498555e-01 8.41283023e-01 8.28056633e-01
-4.90729213e-01 1.04266596e+00 4.19028327e-02 -1.00423954e-01
7.68179968e-02 -7.28658617e-01 1.59721613e-01 9.11855936e-01
5.31459689e-01 6.96794629e-01 5.53264737e-01 -2.25910708e-01
-1.02198446e+00 -1.22005820e+00 9.40700889e-01 5.80443740e-01
3.45649123e-01 -3.75725687e-01 -9.01598454e-01 3.98331702e-01
-2.45635599e-01 1.10634007e-01 4.82682258e-01 3.31089236e-02
-5.35556436e-01 -6.41826034e-01 -1.37400377e+00 6.48805082e-01
1.08330882e+00 -6.57344311e-02 -4.40820932e-01 2.96288729e-01
7.71047771e-01 -4.50869322e-01 -6.41170919e-01 6.21028960e-01
5.24283171e-01 -1.21242344e+00 1.20095289e+00 -1.48683950e-01
4.57423478e-01 9.14786234e-02 -3.87427390e-01 -6.81526303e-01
-1.55207381e-01 -2.54838496e-01 -7.30463982e-01 8.35131466e-01
2.21003726e-01 -6.55954421e-01 1.31820738e+00 4.20397282e-01
-9.27997604e-02 -9.71655548e-01 -1.40272355e+00 -1.15196776e+00
-6.02358878e-02 -7.10242808e-01 7.78838813e-01 3.06828946e-01
-3.10905248e-01 2.03268856e-01 -1.77176282e-01 -1.06555939e-01
6.35286927e-01 -2.06411868e-01 2.03453794e-01 -1.46653724e+00
-2.75121033e-01 -1.02785981e+00 -8.93478751e-01 -1.44353831e+00
5.15884608e-02 -7.20452309e-01 -9.82740223e-02 -1.18605781e+00
-1.24088041e-01 -8.20750296e-01 -2.67379612e-01 5.66300571e-01
6.76851034e-01 7.40131021e-01 2.38055110e-01 3.84595692e-02
-2.92051136e-01 4.54212964e-01 7.02047467e-01 -1.58498690e-01
-5.10980450e-02 -3.52453906e-03 -4.95527774e-01 8.18239391e-01
8.32443953e-01 -6.93866789e-01 -6.61729455e-01 -5.54757535e-01
1.07096359e-01 -2.84108788e-01 2.54853308e-01 -1.72252059e+00
4.29690421e-01 2.28132740e-01 2.66966701e-01 -7.89158344e-01
5.16848207e-01 -1.17708158e+00 1.14861056e-02 7.29576826e-01
-1.57455683e-01 -9.06040445e-02 3.47988486e-01 3.64597976e-01
-3.45422655e-01 -9.22671497e-01 1.02206421e+00 1.22645058e-01
-4.62028325e-01 3.06148440e-01 -3.64163071e-01 -3.02809179e-01
7.05467224e-01 -5.37822843e-01 -3.84701461e-01 -5.09695671e-02
-4.73061293e-01 -3.48316610e-01 1.47391055e-02 -4.46290970e-02
8.48307788e-01 -1.30335069e+00 -3.28371763e-01 5.13385177e-01
-4.72029120e-01 2.09420070e-01 1.60832062e-01 5.51189303e-01
-1.09267819e+00 9.95258272e-01 -4.20270115e-01 -4.22731161e-01
-1.36896455e+00 3.51479262e-01 3.81584674e-01 -3.63213211e-01
-8.04199219e-01 1.13154340e+00 -3.13502133e-01 9.93705243e-02
8.91553104e-01 -8.49329889e-01 -1.46344587e-01 -8.86765569e-02
7.68892229e-01 4.03892577e-01 7.66294658e-01 -4.18008000e-01
-2.25593820e-01 2.62284517e-01 -8.11181515e-02 8.23218748e-02
1.47650170e+00 2.75513262e-01 -4.24681306e-02 -1.64442241e-01
1.17996323e+00 -6.08191729e-01 -7.90294647e-01 -2.92969108e-01
-2.93922294e-02 -1.99226066e-01 3.66083443e-01 -3.34577650e-01
-1.87827396e+00 9.85397756e-01 7.45072126e-01 1.04307331e-01
1.49682593e+00 -3.70456129e-01 1.14443552e+00 9.95969296e-01
4.83431697e-01 -9.64783728e-01 -1.89368039e-01 6.59883380e-01
5.72701097e-01 -7.76882231e-01 2.92028159e-01 -3.55615526e-01
-3.44414897e-02 1.45460582e+00 4.69426811e-01 -2.45411977e-01
8.06872427e-01 6.09898567e-01 -3.52634400e-01 -8.67696255e-02
-7.22909272e-01 5.38770892e-02 8.98610577e-02 6.97896719e-01
1.56126440e-01 -1.00997545e-01 -4.49473768e-01 7.64203429e-01
-5.66094398e-01 2.27527961e-01 3.64882410e-01 6.90649152e-01
-6.87121868e-01 -1.10828757e+00 -2.23503128e-01 8.07774365e-01
-5.61935484e-01 -3.15375924e-01 1.49766728e-01 5.49029708e-01
3.46581697e-01 7.37989128e-01 2.70987093e-01 -3.52106810e-01
1.99354798e-01 -1.96017414e-01 5.78149855e-01 -5.84967971e-01
-4.69115168e-01 -1.67192012e-01 2.02230969e-03 -5.50667226e-01
-3.28020930e-01 -1.07124224e-01 -1.15552604e+00 -5.45044243e-01
-3.81881684e-01 -3.49368677e-02 8.20518911e-01 6.59613550e-01
3.09054643e-01 5.06285191e-01 -6.02884181e-02 -9.62376893e-01
-6.78287148e-01 -8.11255515e-01 -7.53948808e-01 3.26117175e-03
2.72948205e-01 -6.20879292e-01 -2.70233363e-01 -1.83814973e-01] | [8.53123950958252, 3.0515730381011963] |
18bd4fdf-e1f9-45ba-9e61-27c9d5b5f48a | on-the-impact-of-voice-anonymization-on | 2304.02181 | null | https://arxiv.org/abs/2304.02181v1 | https://arxiv.org/pdf/2304.02181v1.pdf | On the Impact of Voice Anonymization on Speech-Based COVID-19 Detection | With advances seen in deep learning, voice-based applications are burgeoning, ranging from personal assistants, affective computing, to remote disease diagnostics. As the voice contains both linguistic and paralinguistic information (e.g., vocal pitch, intonation, speech rate, loudness), there is growing interest in voice anonymization to preserve speaker privacy and identity. Voice privacy challenges have emerged over the last few years and focus has been placed on removing speaker identity while keeping linguistic content intact. For affective computing and disease monitoring applications, however, the paralinguistic content may be more critical. Unfortunately, the effects that anonymization may have on these systems are still largely unknown. In this paper, we fill this gap and focus on one particular health monitoring application: speech-based COVID-19 diagnosis. We test two popular anonymization methods and their impact on five different state-of-the-art COVID-19 diagnostic systems using three public datasets. We validate the effectiveness of the anonymization methods, compare their computational complexity, and quantify the impact across different testing scenarios for both within- and across-dataset conditions. Lastly, we show the benefits of anonymization as a data augmentation tool to help recover some of the COVID-19 diagnostic accuracy loss seen with anonymized data. | ['Tiago H. Falk', 'Carolyn Côté-Lussier', 'Mohamed Imoussaïne-Aïkous', 'Yi Zhu'] | 2023-04-05 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [-4.39700484e-02 1.14285082e-01 8.42508897e-02 -5.22849381e-01
-8.29074502e-01 -7.38549709e-01 3.02419960e-01 2.73044050e-01
-3.81734073e-01 5.70158899e-01 7.66618192e-01 -8.19968656e-02
-7.12247714e-02 -3.35866213e-01 -2.16803581e-01 -4.81425017e-01
-1.54919103e-01 2.45075509e-01 -6.30440235e-01 8.67597237e-02
-4.69684035e-01 5.89191794e-01 -1.26453125e+00 1.15979716e-01
6.37774587e-01 7.35297799e-01 -7.61837602e-01 6.86897576e-01
5.99706396e-02 3.82626086e-01 -9.67678726e-01 -6.44181788e-01
2.91329861e-01 -2.95974463e-01 -6.79907024e-01 -1.94903031e-01
2.43367687e-01 -2.84418315e-01 -4.04252440e-01 9.03991401e-01
1.16013634e+00 -1.17477216e-01 1.35522291e-01 -1.37181509e+00
-3.37858349e-01 6.62539124e-01 -2.54873186e-01 7.65528530e-02
4.23024088e-01 2.07015157e-01 8.83925021e-01 -4.21126515e-01
5.73979914e-01 1.22525036e+00 1.17457211e+00 6.25281036e-01
-1.46782339e+00 -7.63235390e-01 -2.82509387e-01 -1.56255752e-01
-1.63362622e+00 -9.09816265e-01 6.61868215e-01 -4.61847365e-01
8.26193988e-01 5.34267426e-01 5.83578765e-01 1.46996582e+00
6.30655419e-03 5.24383485e-01 9.25787270e-01 -8.44063610e-02
2.29268640e-01 4.54068393e-01 5.42656444e-02 3.27461869e-01
9.78872851e-02 -1.08230794e-02 -6.85367942e-01 -6.72789037e-01
3.07498127e-02 -1.21756181e-01 -4.15005445e-01 -4.57069185e-03
-8.89029503e-01 7.64359474e-01 -2.13848054e-01 1.24999247e-01
-2.70772457e-01 -7.31991976e-02 8.62555444e-01 2.98153788e-01
6.59017742e-01 6.78198516e-01 -3.53298932e-01 -5.08365750e-01
-9.39190149e-01 2.89683789e-01 1.13464487e+00 6.59040272e-01
3.45583826e-01 1.74530610e-01 -3.18390787e-01 1.02731133e+00
-6.88214451e-02 5.82453012e-01 6.83150768e-01 -7.38572121e-01
3.00056696e-01 1.77547425e-01 -2.22512975e-01 -1.09532881e+00
-7.40278542e-01 -3.64769161e-01 -9.54116762e-01 -3.70130301e-01
2.07705840e-01 -5.74431837e-01 -6.24623656e-01 2.06263804e+00
3.23971152e-01 1.22763000e-01 1.48184508e-01 7.72355020e-01
8.30720961e-01 2.67165333e-01 1.28581047e-01 -2.42574140e-01
1.26631975e+00 -3.62718880e-01 -1.01302218e+00 -1.24905080e-01
4.44628805e-01 -8.16105723e-01 9.49690759e-01 2.37455130e-01
-7.73244560e-01 -6.31873682e-02 -7.33452082e-01 3.95374782e-02
-3.05619985e-01 -1.32346958e-01 5.05762637e-01 1.26219618e+00
-9.95461226e-01 5.88474631e-01 -8.85501862e-01 -4.09259379e-01
4.08845901e-01 5.66980004e-01 -6.59571588e-01 1.61961362e-01
-1.45752656e+00 5.17089963e-01 -2.57170141e-01 -4.14450020e-02
-3.83201540e-01 -1.12048531e+00 -7.65867651e-01 4.53293622e-02
2.45965943e-02 -6.69077635e-01 1.12248039e+00 -7.81485081e-01
-1.40961778e+00 1.00076807e+00 -5.85222207e-02 -5.87477803e-01
7.69387186e-01 -1.07211865e-01 -9.71553683e-01 -8.39286968e-02
8.34993925e-03 4.91547853e-01 7.51692057e-01 -9.02558565e-01
-5.21265157e-02 -4.43513751e-01 -4.72858399e-01 -1.90530326e-02
-5.76315880e-01 1.37512982e-01 -1.42570227e-01 -7.59792566e-01
-1.82123557e-01 -1.18701935e+00 8.26111808e-02 -1.36941403e-01
-6.90285981e-01 3.63320440e-01 8.40957403e-01 -1.06968033e+00
1.09407103e+00 -2.50460124e+00 -1.45935103e-01 1.52181610e-01
3.72381717e-01 3.34113568e-01 -5.61073143e-03 3.82878631e-01
-2.58247614e-01 4.30156380e-01 -3.39977443e-01 -6.76679850e-01
9.65781277e-04 1.59229189e-01 -2.26181448e-01 7.08730996e-01
1.88659295e-01 6.21347487e-01 -5.08083940e-01 -3.10406774e-01
1.07592717e-01 7.37083793e-01 -7.85831451e-01 1.10406518e-01
1.88422114e-01 5.40807188e-01 -6.92561045e-02 7.07518399e-01
7.03655124e-01 2.32597098e-01 1.21722579e-01 -9.97529179e-02
2.29190663e-01 3.46289873e-01 -8.03595781e-01 1.41826677e+00
-5.95192790e-01 7.40631580e-01 5.37826538e-01 -6.21543288e-01
7.68145025e-01 6.02971435e-01 9.92889285e-01 -2.52937585e-01
2.37214506e-01 -8.78660157e-02 3.31268609e-01 -6.53264463e-01
5.81399679e-01 -3.55909824e-01 -2.34409854e-01 4.62197185e-01
-2.25966811e-01 -7.79355317e-03 -3.69460434e-01 2.49217497e-03
1.15390456e+00 -7.16733515e-01 2.57711977e-01 -1.07571907e-01
1.50180504e-01 -2.62455493e-01 7.91499078e-01 4.72429752e-01
-7.16063738e-01 9.74701405e-01 4.74554539e-01 -1.17027918e-02
-8.55393350e-01 -1.06350374e+00 -2.98024952e-01 8.38155568e-01
-4.45194304e-01 -5.28352082e-01 -9.11759555e-01 -3.83231819e-01
3.84177536e-01 7.31993794e-01 -5.26180148e-01 -5.74613154e-01
-2.88627684e-01 -8.84628713e-01 1.45194888e+00 2.64488041e-01
1.93236202e-01 -9.51374292e-01 -3.07068348e-01 1.03277631e-01
-3.48043323e-01 -1.30295372e+00 -7.02092946e-01 2.98735797e-02
-4.78043050e-01 -6.44200861e-01 -4.30433840e-01 -1.78345308e-01
8.19282457e-02 -2.52775177e-02 1.07207525e+00 -3.92213702e-01
-5.62282801e-01 7.25810111e-01 -1.68611839e-01 -6.34456515e-01
-6.95102990e-01 3.50402981e-01 4.78745788e-01 2.68540829e-01
3.66098940e-01 -7.11721063e-01 -4.21678782e-01 3.51140387e-02
-7.46264040e-01 -4.03675616e-01 1.54901162e-01 6.60117149e-01
3.14669162e-01 -2.07402438e-01 8.39407504e-01 -1.12421000e+00
1.10610008e+00 -8.28594208e-01 -4.71051186e-02 -1.48652181e-01
-5.41130483e-01 -2.73598403e-01 7.42247760e-01 -3.74420464e-01
-7.87986398e-01 1.60845537e-02 -3.66942346e-01 -6.56654775e-01
-1.76377445e-01 4.27215099e-01 -4.07425165e-01 2.35394642e-01
6.49468422e-01 -4.50507887e-02 3.04184377e-01 -4.20367599e-01
2.70955294e-01 1.25620520e+00 6.78288877e-01 -3.17895949e-01
3.45992208e-01 5.13902724e-01 -3.34442586e-01 -1.27292943e+00
-4.89821702e-01 -4.99791801e-01 -2.99407035e-01 1.08935989e-01
8.05190802e-01 -9.43450153e-01 -7.96448469e-01 5.62474430e-01
-1.05537403e+00 1.54101372e-01 -5.28508246e-01 4.50266838e-01
-3.05055887e-01 2.91771054e-01 -5.24652421e-01 -8.02414834e-01
-8.05007398e-01 -1.14989197e+00 1.03129387e+00 -2.13208094e-01
-9.55560267e-01 -7.81420648e-01 1.49093315e-01 5.71814358e-01
5.04392982e-01 2.82609642e-01 8.70216310e-01 -9.72792327e-01
1.36789903e-01 -4.54991490e-01 7.31997117e-02 3.45282584e-01
6.23822212e-01 -2.41246447e-01 -1.57539093e+00 -3.24072242e-01
4.74398434e-02 -1.61267430e-01 5.31589866e-01 2.32576594e-01
1.09658909e+00 -6.05615914e-01 -1.35361820e-01 8.58622789e-01
7.24261463e-01 -4.53304797e-02 3.67898524e-01 -2.32519582e-01
9.04748738e-01 8.22465599e-01 1.99043099e-02 6.27740562e-01
1.62765697e-01 6.25420630e-01 8.60097483e-02 -6.02399670e-02
-1.23572119e-01 -2.02943519e-01 3.44039410e-01 9.46809947e-01
5.09952784e-01 -3.20388585e-01 -1.00339484e+00 6.90512598e-01
-1.26123393e+00 -9.09494698e-01 2.86503822e-01 2.24296379e+00
1.02422535e+00 -3.99139583e-01 3.98694634e-01 1.98868796e-01
7.29116857e-01 1.54522866e-01 -9.30111527e-01 -6.35056555e-01
-1.48820490e-01 2.12767944e-01 5.14440358e-01 3.32280517e-01
-1.02631497e+00 6.74093664e-01 6.32741308e+00 3.03640604e-01
-1.60491490e+00 3.03750392e-02 7.09596395e-01 -5.40474296e-01
-2.91155308e-01 -5.49216747e-01 -3.68646264e-01 4.39042240e-01
1.18720233e+00 -4.12937373e-01 5.26624799e-01 7.06033051e-01
6.28273249e-01 1.93334341e-01 -1.16166782e+00 1.28312111e+00
1.23203136e-01 -9.26349998e-01 -1.55824035e-01 2.47381732e-01
4.61727589e-01 1.96428791e-01 2.06432298e-01 4.32540588e-02
1.11481063e-01 -1.22994769e+00 4.10597116e-01 2.37612277e-01
1.19968390e+00 -9.37140822e-01 7.95971632e-01 -1.09413840e-01
-6.92377031e-01 5.29872328e-02 8.37401599e-02 2.88349003e-01
7.09199458e-02 8.38387132e-01 -1.22832286e+00 3.22311252e-01
6.73673034e-01 4.52305317e-01 -2.92873114e-01 7.45253563e-01
1.79975525e-01 8.49748731e-01 -4.27113980e-01 2.08119377e-01
-2.95780003e-01 5.61244190e-02 8.02719176e-01 1.28242946e+00
2.97088146e-01 8.19373429e-02 -1.54150069e-01 7.57440329e-01
-4.10136700e-01 1.51170909e-01 -7.53298104e-01 -4.71922666e-01
7.35384345e-01 1.07388043e+00 -2.59320706e-01 7.18017668e-02
-1.86689988e-01 1.02273536e+00 4.39911196e-03 1.80184379e-01
-6.35732770e-01 -3.24534208e-01 1.56213975e+00 2.41840214e-01
-1.61316276e-01 1.16462626e-01 -3.87894779e-01 -9.67621088e-01
4.75601666e-02 -1.35834825e+00 3.77936900e-01 -3.02944422e-01
-1.32199812e+00 4.75815028e-01 -3.17548692e-01 -9.37527955e-01
-5.38197815e-01 1.72728933e-02 -3.67067099e-01 9.58993852e-01
-8.23405981e-01 -8.71559143e-01 -1.71865538e-01 4.72969532e-01
1.02330744e-01 -1.78613111e-01 1.14143538e+00 6.46318734e-01
-6.68482661e-01 1.23330986e+00 1.95655569e-01 3.47289562e-01
1.03092062e+00 -9.14797485e-01 5.52683413e-01 5.43103218e-01
7.27139264e-02 7.57968426e-01 7.95901656e-01 -4.46703851e-01
-1.50151896e+00 -1.33485436e+00 7.48595655e-01 -4.09007847e-01
5.28690159e-01 -5.79790294e-01 -8.77317131e-01 6.26698494e-01
-1.13510452e-01 3.34842987e-02 1.08538866e+00 2.92469233e-01
-4.98965979e-01 -3.32573265e-01 -1.60437119e+00 5.86568952e-01
7.91772425e-01 -1.02043879e+00 -1.25636622e-01 9.48370323e-02
9.04360890e-01 -3.57383877e-01 -9.74088490e-01 2.03201413e-01
7.43167102e-01 -9.44028616e-01 8.87499869e-01 -6.14728212e-01
8.45568553e-02 5.24341613e-02 -1.22972518e-01 -1.59526682e+00
-4.87451367e-02 -9.53741193e-01 3.01620454e-01 1.69787681e+00
4.09179628e-01 -8.77925754e-01 8.90805960e-01 8.93053353e-01
7.66237751e-02 -4.13281888e-01 -1.04580879e+00 -6.18837833e-01
-1.39136333e-02 -6.63066030e-01 8.18453193e-01 1.40497386e+00
4.63402867e-02 3.58426988e-01 -7.20997393e-01 2.41232902e-01
2.45837703e-01 -2.19612956e-01 7.69006729e-01 -1.13060951e+00
-3.13019127e-01 -1.80894375e-01 -6.04453027e-01 -1.03142872e-01
3.15398037e-01 -9.36064601e-01 -2.05379039e-01 -8.36228669e-01
-1.11317597e-02 -3.39631170e-01 -3.42011228e-02 6.10019386e-01
-4.17578220e-02 7.69074336e-02 2.50972301e-01 -5.49982525e-02
-3.61005776e-02 6.07726693e-01 5.73287129e-01 -2.11087346e-01
-5.45051217e-01 1.93429872e-01 -8.52276087e-01 5.44817209e-01
9.98588085e-01 -5.97886264e-01 -4.66629028e-01 -3.15127522e-01
-1.65449783e-01 1.60980448e-01 2.20657319e-01 -9.83512878e-01
4.61729895e-03 2.85048336e-01 6.92741871e-02 -1.20012209e-01
5.39043665e-01 -7.67405450e-01 3.90302122e-01 3.82816821e-01
-3.47608149e-01 1.27543807e-01 5.45121253e-01 4.80239213e-01
-1.34646192e-01 2.31749147e-01 7.61939049e-01 2.48743027e-01
1.49344960e-02 2.59309977e-01 -4.35874492e-01 4.67665941e-01
6.14906192e-01 -3.13594677e-02 -1.47259101e-01 -7.23025501e-01
-7.92762756e-01 6.63794354e-02 3.23414952e-01 7.33686209e-01
3.47740203e-01 -9.92912173e-01 -8.26559901e-01 4.59631383e-01
2.77511299e-01 -4.09214824e-01 3.83982390e-01 8.41186106e-01
-1.90792143e-01 2.79573053e-01 -3.78448213e-03 -5.44905126e-01
-1.61140370e+00 3.73801440e-01 5.05437553e-01 1.82063162e-01
-4.99878794e-01 6.64260447e-01 -6.55340776e-02 -7.66371727e-01
3.73131394e-01 -3.11126918e-01 2.27587700e-01 3.12236845e-01
2.92347908e-01 5.41898906e-01 3.94827276e-01 -6.71127915e-01
-6.42830491e-01 7.59997964e-02 -3.68597656e-02 -2.09347889e-01
1.21112919e+00 -1.41114876e-01 -7.95033798e-02 5.00411034e-01
1.54647946e+00 4.89377558e-01 -6.73639476e-01 -4.63290475e-02
-1.84398100e-01 -2.51233160e-01 -3.84053886e-02 -8.37036014e-01
-1.21601391e+00 7.58045375e-01 8.06010246e-01 2.38830104e-01
1.05789268e+00 -1.72396839e-01 9.30473387e-01 9.67763662e-02
-2.11694110e-02 -9.77017283e-01 -5.55303276e-01 2.28185862e-01
7.14558601e-01 -1.03308606e+00 -9.04900134e-02 -4.04523611e-01
-6.77851737e-01 4.48082834e-01 1.49805561e-01 4.99505997e-01
8.54445577e-01 5.38635612e-01 3.99796039e-01 6.06322894e-03
-4.94581342e-01 2.51521319e-01 -7.27606565e-02 6.78070009e-01
6.48119509e-01 4.09575015e-01 -1.72766745e-02 8.62360656e-01
-9.84904528e-01 -2.04081550e-01 5.00826895e-01 5.27197659e-01
3.29606831e-01 -8.76796842e-01 -4.68914807e-01 6.86381578e-01
-9.24403191e-01 -2.33207747e-01 -9.00937438e-01 6.81196332e-01
4.82420847e-02 1.14420152e+00 1.51305810e-01 -5.49514651e-01
4.62973893e-01 2.18352303e-01 2.03463268e-02 -5.19706011e-01
-8.49679053e-01 -1.72578573e-01 4.29388285e-01 -5.49417436e-01
-2.63245068e-02 -1.06776261e+00 -1.20555592e+00 -7.35254228e-01
5.71230389e-02 1.15234442e-01 7.41723061e-01 7.09762931e-01
7.86996245e-01 5.41731596e-01 5.72722852e-01 -3.07276964e-01
-7.24215448e-01 -8.55965376e-01 -6.72889531e-01 4.58436102e-01
6.19340420e-01 -2.33856305e-01 -4.37989950e-01 -1.30761668e-01] | [14.016050338745117, 5.878878116607666] |
93022461-c0cf-41f4-a38d-3d3ccd5216cf | hm-net-a-regression-network-for-object-center | 2110.09881 | null | https://arxiv.org/abs/2110.09881v2 | https://arxiv.org/pdf/2110.09881v2.pdf | HM-Net: A Regression Network for Object Center Detection and Tracking on Wide Area Motion Imagery | Wide Area Motion Imagery (WAMI) yields high-resolution images with a large number of extremely small objects. Target objects have large spatial displacements throughout consecutive frames. This nature of WAMI images makes object tracking and detection challenging. In this paper, we present our deep neural network-based combined object detection and tracking model, namely, Heat Map Network (HM-Net). HM-Net is significantly faster than state-of-the-art frame differencing and background subtraction-based methods, without compromising detection and tracking performances. HM-Net follows the object center-based joint detection and tracking paradigm. Simple heat map-based predictions support an unlimited number of simultaneous detections. The proposed method uses two consecutive frames and the object detection heat map obtained from the previous frame as input, which helps HM-Net monitor spatio-temporal changes between frames and keeps track of previously predicted objects. Although reuse of prior object detection heat map acts as a vital feedback-based memory element, it can lead to an unintended surge of false-positive detections. To increase the robustness of the method against false positives and to eliminate low confidence detections, HM-Net employs novel feedback filters and advanced data augmentations. HM-Net outperforms state-of-the-art WAMI moving object detection and tracking methods on the WPAFB dataset with its 96.2% F1 and 94.4% mAP detection scores while achieving a 61.8% mAP tracking score on the same dataset. This performance corresponds to an improvement of 2.1% for F1, 6.1% for mAP scores on detection, and 9.5% for mAP score on tracking over the state-of-the-art. | ['Bahadir Gunturk', 'H. Fatih Ugurdag', 'Hasan F. Ates', 'Hakki Motorcu'] | 2021-10-19 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 3.09149534e-01 -3.36816669e-01 1.01706833e-01 2.46191360e-02
-4.31187540e-01 -3.89532566e-01 5.42473078e-01 -9.17831808e-02
-7.62666464e-01 5.82958281e-01 -3.98670226e-01 3.94590711e-03
3.81186426e-01 -9.04102504e-01 -9.04074132e-01 -8.68408382e-01
-4.34686929e-01 8.45852196e-02 1.23728132e+00 -1.94125175e-02
-1.01733804e-01 4.85295534e-01 -1.75846887e+00 4.00767535e-01
4.72982407e-01 1.30114746e+00 5.29463768e-01 9.59292591e-01
1.24725208e-01 8.49978983e-01 -8.34150076e-01 3.18402201e-02
4.48207557e-01 -7.79749751e-02 -1.75063968e-01 -6.50746450e-02
1.03065562e+00 -7.07183242e-01 -5.79627991e-01 9.23236310e-01
5.00193477e-01 2.47521251e-01 2.39505619e-01 -1.39653289e+00
-3.51502001e-01 2.07741365e-01 -9.22979772e-01 9.62585688e-01
-1.30983189e-01 5.07756889e-01 3.47375959e-01 -9.79071975e-01
4.29750949e-01 1.00727534e+00 9.80361700e-01 4.38972265e-01
-1.18045449e+00 -1.19770908e+00 1.96693063e-01 9.24274176e-02
-1.72484779e+00 -2.54174769e-01 1.55247375e-01 -6.95596993e-01
1.10195148e+00 3.49467635e-01 7.32554138e-01 5.39926648e-01
3.01047385e-01 8.26987624e-01 4.89824355e-01 -1.18063830e-01
3.45254713e-03 -4.98141125e-02 9.18307230e-02 6.45780265e-01
5.69324851e-01 5.15471280e-01 -4.36140716e-01 8.50470662e-02
1.09726918e+00 1.55539036e-01 -3.02529544e-01 -1.36144087e-01
-1.39427495e+00 4.93735969e-01 7.29217172e-01 2.20929667e-01
-4.81730103e-01 3.01016212e-01 1.25859439e-01 -1.51883289e-01
3.23147953e-01 -1.74920529e-01 -4.40444767e-01 1.31402448e-01
-1.30837679e+00 3.41180861e-01 7.42578804e-02 7.94430375e-01
5.10778427e-01 5.59482515e-01 -5.96383512e-01 3.53787035e-01
1.46385878e-01 1.05088949e+00 2.00863734e-01 -6.13995910e-01
4.02237058e-01 5.62395990e-01 4.89836305e-01 -1.32064271e+00
-6.31312072e-01 -7.77916968e-01 -8.93252552e-01 7.53431976e-01
5.15871644e-01 -2.06604093e-01 -1.21617103e+00 1.69438183e+00
4.75079894e-01 5.75072885e-01 -1.42405644e-01 1.01083481e+00
7.27950633e-01 9.25621629e-01 3.01200598e-01 -1.96864977e-01
1.12334871e+00 -6.88598394e-01 -6.61496818e-01 -4.09011811e-01
2.74209321e-01 -5.10041416e-01 4.85777110e-01 1.13594703e-01
-9.57168579e-01 -1.07697320e+00 -1.00882542e+00 5.29384673e-01
-2.69793063e-01 4.75725949e-01 2.19536901e-01 5.59031963e-01
-9.97881234e-01 2.89562702e-01 -1.08326268e+00 -2.83241540e-01
6.12361908e-01 6.81886792e-01 -2.50625342e-01 1.17477819e-01
-8.86482358e-01 9.77088094e-01 8.39266658e-01 1.81434900e-01
-9.76904511e-01 -8.76363635e-01 -6.32487059e-01 2.05322597e-02
5.06415009e-01 -3.35567415e-01 1.00537908e+00 -8.10654044e-01
-8.84360135e-01 5.18029928e-01 -8.55196863e-02 -8.48320663e-01
7.92635202e-01 -4.55572873e-01 -6.59297943e-01 2.96458974e-02
1.97764039e-01 1.19005334e+00 9.28487599e-01 -9.17825460e-01
-1.44645894e+00 -7.75924176e-02 -2.86911488e-01 -1.35261938e-01
-3.05083655e-02 4.55007218e-02 -5.60497820e-01 -6.88033342e-01
-7.06271604e-02 -9.32628453e-01 6.65352270e-02 3.22346061e-01
2.32787542e-02 1.06942371e-01 1.37134087e+00 -6.32337153e-01
1.35613036e+00 -1.93327963e+00 -4.99121785e-01 -2.86155343e-01
2.54953712e-01 9.84540522e-01 -1.64208636e-01 -1.64072856e-01
5.56083433e-02 -2.61154413e-01 -1.23155817e-01 7.50489160e-02
-4.65377897e-01 -1.73431456e-01 -2.11184412e-01 5.81690669e-01
5.22270441e-01 9.16164935e-01 -8.42167795e-01 -1.99375138e-01
6.92278862e-01 6.44282401e-01 -2.80109495e-01 -1.51313227e-02
-3.50609124e-02 3.60520959e-01 5.80213964e-02 7.75123179e-01
8.98589134e-01 -2.73326933e-01 -2.54900455e-01 -3.87078792e-01
-5.23069799e-01 -3.73370945e-01 -1.44173396e+00 1.15234244e+00
1.49435624e-01 1.17429876e+00 -1.18203826e-01 -4.52499837e-01
7.25647211e-01 1.82482183e-01 4.96145517e-01 -7.55011916e-01
1.48265511e-01 -1.02229044e-01 1.88654989e-01 -1.85078397e-01
7.61716604e-01 2.45399371e-01 3.21737111e-01 -5.27051277e-02
-9.67127904e-02 7.69199312e-01 2.69554496e-01 6.57618642e-02
1.15179884e+00 7.83962905e-02 1.96541995e-01 -1.14974953e-01
5.05986392e-01 3.18218201e-01 8.18656981e-01 1.17677319e+00
-4.45125282e-01 5.41762054e-01 -4.78250772e-01 -7.74914682e-01
-9.13192153e-01 -1.11450672e+00 -1.71929911e-01 1.16096270e+00
3.63192052e-01 -3.52163166e-02 -4.05054271e-01 -3.80823553e-01
5.80684543e-02 4.70298648e-01 -7.80286491e-01 -1.98486701e-01
-8.19551170e-01 -1.03957450e+00 5.62545240e-01 8.58118773e-01
9.30669427e-01 -9.40647244e-01 -1.35064840e+00 5.03413618e-01
-3.10067236e-01 -1.09183109e+00 -2.38716796e-01 -6.10353127e-02
-6.94287419e-01 -1.06653404e+00 -8.45277131e-01 -4.95023489e-01
3.93650025e-01 7.98386216e-01 9.29353118e-01 -7.47606624e-03
-5.11441886e-01 -7.36698415e-03 -1.96526155e-01 -6.04278386e-01
-1.96410730e-01 -3.86992574e-01 2.23975554e-01 7.37822801e-02
3.31914634e-01 -1.52876088e-02 -7.91046202e-01 6.78708017e-01
-8.71374309e-01 1.05718754e-01 5.91270149e-01 6.88177466e-01
5.98567784e-01 4.72244956e-02 2.79440105e-01 -1.88092500e-01
-2.33732745e-01 -2.06457376e-01 -1.26495969e+00 2.15713903e-01
-2.66598254e-01 -4.15486753e-01 1.69155393e-02 -8.68271828e-01
-8.50789666e-01 4.50056791e-01 2.86013991e-01 -6.32169843e-01
-4.73016761e-02 -1.70202255e-01 1.63489372e-01 -3.17318112e-01
7.82862365e-01 3.34195137e-01 -1.34860352e-01 -1.76001728e-01
1.15736835e-01 3.75792742e-01 1.11430550e+00 3.55091572e-01
9.12369728e-01 6.81906164e-01 -1.15130700e-01 -7.78917253e-01
-6.19781375e-01 -7.17230976e-01 -6.34377658e-01 -6.66141629e-01
1.05529785e+00 -1.31965482e+00 -6.81491613e-01 6.41341031e-01
-1.05805230e+00 -4.01701629e-01 7.87114538e-03 4.21065032e-01
-8.65621865e-02 6.69196695e-02 -3.29111338e-01 -1.06739807e+00
-5.42350888e-01 -7.72782326e-01 9.58867550e-01 4.94730443e-01
2.91329343e-02 -4.67173666e-01 -2.47374848e-01 -4.44355272e-02
7.55279720e-01 4.58740264e-01 -1.11133099e-01 -3.43243152e-01
-9.80782032e-01 -6.50780678e-01 -4.91954327e-01 7.25770195e-04
-1.09659657e-02 6.86198696e-02 -9.88297760e-01 -4.22095150e-01
-3.73153120e-01 2.21996903e-01 1.07690859e+00 9.22365963e-01
6.11745715e-01 -3.05165406e-02 -7.58845389e-01 4.40118343e-01
1.52053881e+00 5.20193815e-01 6.41571820e-01 5.98533750e-01
6.80391014e-01 -1.10312207e-02 1.01529860e+00 4.99947608e-01
-5.76268043e-03 8.67734671e-01 6.53389573e-01 -2.63788044e-01
-3.13860923e-01 2.16621399e-01 4.36541051e-01 -8.73518959e-02
-5.80389835e-02 -2.30658546e-01 -1.13737857e+00 6.03630006e-01
-2.16442275e+00 -1.44487941e+00 -4.20638055e-01 2.41767859e+00
2.67779738e-01 3.69710118e-01 4.10336494e-01 -1.14034787e-01
1.13895845e+00 -1.08555734e-01 -5.55285513e-01 6.39875174e-01
-2.80469894e-01 -2.11188465e-01 1.10219860e+00 2.16886058e-01
-1.75080967e+00 9.80957925e-01 5.55241537e+00 6.13508403e-01
-1.28442121e+00 1.99760899e-01 3.66552085e-01 -3.70089114e-01
8.51155043e-01 -5.00935793e-01 -1.30555344e+00 5.68147838e-01
9.47363794e-01 1.28116652e-01 -4.66087833e-02 7.84053326e-01
3.57075661e-01 -5.09974241e-01 -6.99633539e-01 1.07552314e+00
-1.39834225e-01 -1.70738709e+00 -2.49648452e-01 -3.06783523e-02
6.80487812e-01 5.25600016e-01 -7.21958503e-02 4.06160176e-01
1.69911474e-01 -6.85050905e-01 1.00493133e+00 4.51847821e-01
6.36342704e-01 -7.17899740e-01 9.95090067e-01 5.24319768e-01
-1.64808929e+00 -2.21143842e-01 -5.20777762e-01 -1.58658013e-01
3.20249110e-01 3.02465796e-01 -6.14786625e-01 2.74342865e-01
1.10343754e+00 4.74570155e-01 -7.28815794e-01 1.61234128e+00
2.45789602e-01 6.42973006e-01 -6.04198992e-01 3.01489592e-01
3.24308783e-01 2.86661655e-01 7.35683143e-01 1.58731520e+00
2.96262532e-01 2.16861963e-01 5.64101696e-01 7.35076308e-01
2.81494349e-01 -5.46881974e-01 -4.42722023e-01 3.88667315e-01
5.60653925e-01 1.25475633e+00 -8.85385871e-01 -7.29457080e-01
-4.20153528e-01 8.12871277e-01 -1.10099867e-01 1.84773445e-01
-1.28235912e+00 -2.57698685e-01 7.59501696e-01 2.77960300e-01
8.44792783e-01 -2.16350004e-01 6.94303364e-02 -8.67445230e-01
-1.52426064e-01 -3.54087532e-01 3.79120797e-01 -6.58012986e-01
-7.88478494e-01 7.14642644e-01 3.28736342e-02 -1.62486577e+00
-1.07112013e-01 -5.34979582e-01 -5.61840117e-01 6.60231471e-01
-1.33417714e+00 -1.21811032e+00 -7.18508899e-01 3.83936822e-01
5.91802835e-01 -6.37146682e-02 3.54775459e-01 6.02044463e-01
-7.48672545e-01 4.62474495e-01 1.13226503e-01 6.52630031e-01
5.37502348e-01 -8.12041700e-01 6.36763513e-01 1.45823395e+00
4.16694954e-02 1.15356490e-01 6.30360544e-01 -9.65566874e-01
-1.17318189e+00 -1.80312204e+00 3.31989676e-01 -5.75425982e-01
4.69743341e-01 -1.85811624e-01 -1.14944577e+00 6.09632015e-01
-1.71410531e-01 5.35097539e-01 1.36030868e-01 -6.49499953e-01
9.19060186e-02 -1.22064792e-01 -1.02492940e+00 3.36208999e-01
8.81473720e-01 1.19441606e-01 -3.86759967e-01 1.34607121e-01
5.72621405e-01 -5.31623900e-01 -4.27757651e-01 6.76324010e-01
7.17847288e-01 -8.94212306e-01 1.35289645e+00 -2.06037894e-01
-1.93877771e-01 -1.05745196e+00 -1.59018874e-01 -7.59299099e-01
-9.22292590e-01 -1.43910691e-01 -3.88062179e-01 1.08500195e+00
1.63641483e-01 -2.93278754e-01 8.45860362e-01 5.44576526e-01
-2.46950388e-02 -2.31894359e-01 -9.62460101e-01 -1.16656673e+00
-5.03459215e-01 -5.42049944e-01 1.00264318e-01 6.62048340e-01
-6.95504367e-01 -3.54636163e-02 -5.49950778e-01 7.68582523e-01
8.23083162e-01 -7.08861798e-02 8.69334817e-01 -1.19385874e+00
-3.97641212e-03 -3.63631666e-01 -8.26708615e-01 -7.25947678e-01
-5.89944601e-01 -4.33539629e-01 2.50112385e-01 -1.39303851e+00
1.03753656e-01 -1.11579917e-01 -4.33597624e-01 6.92627966e-01
-4.14164394e-01 8.44263971e-01 5.61386347e-01 3.22832525e-01
-8.40682924e-01 2.12838307e-01 7.05491006e-01 -2.43161097e-01
-4.43624884e-01 2.06619389e-02 8.94426368e-03 6.61665976e-01
8.34018111e-01 -6.66942596e-01 -1.50092710e-02 -4.07791048e-01
-4.09883827e-01 -1.75479308e-01 7.76428580e-01 -1.80493450e+00
4.72997248e-01 -5.51593304e-02 1.11562383e+00 -1.23038054e+00
4.01761770e-01 -8.29233885e-01 5.78908861e-01 8.89714897e-01
1.44422501e-01 1.03181317e-01 8.87292147e-01 8.00146341e-01
-9.25390199e-02 2.91604519e-01 1.00424993e+00 -7.32231587e-02
-1.27858305e+00 2.50771463e-01 -6.78617597e-01 -1.96807578e-01
1.22669101e+00 -6.36662841e-01 -3.78906786e-01 -3.92403305e-02
-5.68985403e-01 2.13204682e-01 1.83920249e-01 5.53221285e-01
5.54531157e-01 -1.18235040e+00 -7.25351155e-01 9.75145176e-02
2.03522351e-02 -9.00075771e-03 2.96838015e-01 1.04270446e+00
-5.03705204e-01 4.97120947e-01 -5.21900594e-01 -1.12317216e+00
-1.52412999e+00 3.68096590e-01 3.50394070e-01 -1.10395614e-03
-8.32120299e-01 9.17326391e-01 4.43642676e-01 2.40462214e-01
3.20264101e-01 -4.60118353e-01 3.08781560e-03 -7.28772283e-02
1.28096497e+00 5.66884041e-01 -3.82019393e-02 -7.71920443e-01
-5.93195498e-01 4.10964161e-01 -1.22625642e-01 6.90415576e-02
1.10963988e+00 -9.59863290e-02 4.62389678e-01 2.30406448e-01
5.63970804e-01 -4.15220827e-01 -1.83651888e+00 -3.50373477e-01
-9.54418406e-02 -7.61737645e-01 4.59685296e-01 -1.04673088e+00
-1.18140626e+00 6.60351932e-01 1.43468475e+00 -4.30432744e-02
1.01599038e+00 -2.07663864e-01 6.41788125e-01 2.44498983e-01
2.77187735e-01 -7.62365699e-01 -9.13115125e-03 4.87697661e-01
5.94585955e-01 -1.54636836e+00 2.57772766e-02 -1.61416769e-01
-5.63277423e-01 8.87132943e-01 1.08150542e+00 3.31530371e-03
2.63710856e-01 4.79155511e-01 7.42466599e-02 3.53305899e-02
-5.11695206e-01 -6.25376284e-01 5.43672264e-01 7.39863455e-01
2.30217651e-02 -4.50646356e-02 2.53394336e-01 1.05593294e-01
3.70006651e-01 9.37485471e-02 2.43896395e-01 1.12305343e+00
-1.00626409e+00 -2.59409308e-01 -7.94713795e-01 4.31766272e-01
-3.93853694e-01 1.39811793e-02 -3.89257669e-02 1.14627457e+00
3.70274067e-01 9.02069807e-01 5.00937998e-01 -4.80350107e-01
1.26738608e-01 -1.70350060e-01 1.97960839e-01 -1.76851958e-01
-5.95500350e-01 1.57930478e-01 -2.13575572e-01 -5.32439232e-01
-5.04398108e-01 -5.52302539e-01 -1.37407255e+00 -3.03214133e-01
-7.91184187e-01 -4.42748129e-01 5.59292376e-01 7.16182232e-01
4.77987647e-01 7.24577248e-01 -1.28209382e-01 -1.31790745e+00
-3.20236571e-02 -1.06166518e+00 -2.48776123e-01 1.44375011e-01
6.85783982e-01 -8.80716443e-01 1.70750931e-01 3.13279659e-01] | [6.4963788986206055, -2.062049627304077] |
27c63198-1daf-45ea-a7a9-4cb28fded4f5 | transfer-to-a-low-resource-language-via-close | 2304.08823 | null | https://arxiv.org/abs/2304.08823v1 | https://arxiv.org/pdf/2304.08823v1.pdf | Transfer to a Low-Resource Language via Close Relatives: The Case Study on Faroese | Multilingual language models have pushed state-of-the-art in cross-lingual NLP transfer. The majority of zero-shot cross-lingual transfer, however, use one and the same massively multilingual transformer (e.g., mBERT or XLM-R) to transfer to all target languages, irrespective of their typological, etymological, and phylogenetic relations to other languages. In particular, readily available data and models of resource-rich sibling languages are often ignored. In this work, we empirically show, in a case study for Faroese -- a low-resource language from a high-resource language family -- that by leveraging the phylogenetic information and departing from the 'one-size-fits-all' paradigm, one can improve cross-lingual transfer to low-resource languages. In particular, we leverage abundant resources of other Scandinavian languages (i.e., Danish, Norwegian, Swedish, and Icelandic) for the benefit of Faroese. Our evaluation results show that we can substantially improve the transfer performance to Faroese by exploiting data and models of closely-related high-resource languages. Further, we release a new web corpus of Faroese and Faroese datasets for named entity recognition (NER), semantic text similarity (STS), and new language models trained on all Scandinavian languages. | ['Ivan Vulić', 'Goran Glavaš', 'Annika Simonsen', 'Vésteinn Snæbjarnarson'] | 2023-04-18 | null | null | null | null | ['zero-shot-cross-lingual-transfer', 'named-entity-recognition-ner', 'cross-lingual-transfer', 'xlm-r'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-2.52533644e-01 -2.60436416e-01 -4.73949552e-01 -2.95254618e-01
-1.07675099e+00 -1.03943455e+00 5.03441811e-01 2.40961224e-01
-1.01629162e+00 1.10003603e+00 4.26404119e-01 -5.88540614e-01
2.19913363e-01 -4.58645910e-01 -7.39342630e-01 -1.44427627e-01
2.49358192e-01 7.18012214e-01 2.08417565e-01 -5.89223683e-01
-7.48395920e-02 1.56687155e-01 -9.59977090e-01 3.07483405e-01
1.06182528e+00 2.43557151e-02 3.55407685e-01 7.53974989e-02
-3.16150010e-01 2.48854145e-01 -1.87846720e-01 -8.94480169e-01
2.02218577e-01 -4.11519140e-01 -9.20301795e-01 -7.65997708e-01
4.48281378e-01 2.09086522e-01 -1.19478673e-01 7.77372897e-01
5.74127316e-01 -8.38245228e-02 7.03718245e-01 -8.89643729e-01
-8.39068413e-01 1.11161721e+00 -4.26160902e-01 2.50457764e-01
1.57338768e-01 2.48750392e-02 1.27967358e+00 -8.96647751e-01
1.28235054e+00 1.25206339e+00 8.09293330e-01 6.33402824e-01
-1.12465644e+00 -9.18291569e-01 1.22173846e-01 4.00891975e-02
-1.51103175e+00 -7.38882422e-01 2.72096753e-01 -2.97019780e-01
1.36400127e+00 -3.73109132e-02 2.68206716e-01 1.15223801e+00
-2.00252101e-01 7.16850340e-01 1.09117842e+00 -6.45009696e-01
-2.87585646e-01 3.16322953e-01 -1.26191616e-01 6.17299378e-01
2.47174710e-01 2.45539293e-01 -7.48975694e-01 -1.02673769e-01
3.01752895e-01 -5.36220372e-01 -2.54435062e-01 -1.01038866e-01
-1.41953719e+00 7.22182930e-01 2.51768082e-01 8.47071230e-01
-4.06275690e-02 -1.85082480e-01 7.96684504e-01 4.87084210e-01
6.80132031e-01 4.30316478e-01 -1.08078980e+00 -2.16212869e-01
-7.86657691e-01 -1.84179574e-01 9.77210104e-01 1.15544879e+00
9.20691907e-01 -5.31239882e-02 3.52582514e-01 1.21123922e+00
9.49105993e-02 5.86192966e-01 7.58471191e-01 -3.02574843e-01
9.22318518e-01 2.02092856e-01 -2.18258262e-01 -3.39789629e-01
-1.60601184e-01 -1.24177471e-01 -4.90471870e-01 -2.99466521e-01
4.09262449e-01 -2.89358586e-01 -6.66696370e-01 2.13466167e+00
4.17669415e-01 -4.34413850e-02 5.61528981e-01 5.61578035e-01
6.79112792e-01 4.90210921e-01 4.03099716e-01 -1.45698031e-02
1.41885686e+00 -7.55639732e-01 -2.22822756e-01 -4.75509912e-01
1.22184753e+00 -8.88773263e-01 1.30840659e+00 -1.22849502e-01
-9.29280400e-01 -4.31119949e-01 -7.04911470e-01 -3.48396093e-01
-7.90549219e-01 4.05588597e-02 6.10994279e-01 4.87756252e-01
-9.24421787e-01 5.94720781e-01 -7.29121029e-01 -8.93661976e-01
7.03456402e-02 -6.66895183e-03 -8.81777287e-01 -3.33761066e-01
-1.62425673e+00 1.30034566e+00 7.73964763e-01 -4.21403915e-01
-5.90386331e-01 -1.17134714e+00 -8.79687905e-01 -2.59097457e-01
1.34632319e-01 -5.25387228e-01 9.56457734e-01 -9.20007229e-01
-1.27135861e+00 1.47470284e+00 2.53096260e-02 -3.33606213e-01
2.80434310e-01 -3.40387225e-01 -6.30850911e-01 -2.16601476e-01
4.52380538e-01 7.35017598e-01 2.25578137e-02 -8.44877243e-01
-7.99928069e-01 -4.01458532e-01 -2.41332099e-01 2.95664728e-01
-5.43579817e-01 7.74275422e-01 -3.54870170e-01 -5.84995210e-01
-4.61798638e-01 -1.03517580e+00 1.08632110e-01 -3.30096096e-01
-1.03706665e-01 -3.09502691e-01 2.19467267e-01 -1.06841040e+00
1.27635658e+00 -2.16473842e+00 3.09444964e-01 -1.03858285e-01
-4.45015579e-01 4.07433271e-01 -3.53558928e-01 7.90602505e-01
5.48218517e-03 5.15269578e-01 -2.48923585e-01 -2.69585341e-01
-2.93620825e-02 4.66885149e-01 -6.76055700e-02 3.86313975e-01
2.15195939e-01 1.01904082e+00 -1.01097488e+00 -6.03469253e-01
4.76396084e-03 4.52507377e-01 -4.88824397e-01 -8.65225121e-02
3.17103788e-02 4.87265378e-01 -2.21744180e-02 4.19139355e-01
1.92061901e-01 2.05861226e-01 4.57847834e-01 -4.46513966e-02
-3.69958609e-01 9.51329052e-01 -5.58968663e-01 2.03137517e+00
-1.11998832e+00 4.13202494e-01 9.19414088e-02 -5.64316809e-01
6.68212652e-01 3.68347287e-01 5.09642400e-02 -5.95035374e-01
-1.33844614e-01 1.01687276e+00 3.90382946e-01 -1.25906467e-01
4.34356630e-01 -4.62954015e-01 -3.73177946e-01 3.24219137e-01
3.90513837e-01 -8.44604522e-02 2.63658583e-01 4.16787751e-02
8.56424749e-01 4.85009938e-01 7.20516264e-01 -5.54422021e-01
4.35377359e-01 1.28497466e-01 7.59044170e-01 2.45570287e-01
-4.33325768e-02 1.58170566e-01 -7.71359131e-02 -1.56766828e-02
-1.22993827e+00 -1.08416402e+00 -3.79170477e-01 1.58133674e+00
-3.82258445e-01 -4.83164251e-01 -5.53677022e-01 -8.17877591e-01
1.38807043e-01 1.01013911e+00 -3.11913252e-01 -3.07673141e-02
-1.01083505e+00 -5.08579731e-01 1.17421508e+00 2.78876364e-01
9.72792283e-02 -1.12867749e+00 -2.16469951e-02 3.95196795e-01
-1.89241111e-01 -1.31318724e+00 -6.86045170e-01 1.39143810e-01
-4.91119713e-01 -7.03590035e-01 -8.11336398e-01 -9.35549021e-01
1.30126685e-01 -4.41354588e-02 1.39200568e+00 -2.65579611e-01
4.84237932e-02 7.43603632e-02 -4.07547206e-01 -3.25218797e-01
-7.66586185e-01 7.27225542e-01 3.48107278e-01 -4.13540542e-01
5.78760684e-01 -6.18360758e-01 1.42885655e-01 2.70756811e-01
-5.37446439e-01 -1.04884543e-01 3.53803396e-01 6.58189178e-01
4.77214038e-01 -5.53419888e-01 7.90733457e-01 -1.03328443e+00
2.28418231e-01 -6.97597563e-01 -5.36096394e-01 5.95490336e-01
-1.61910370e-01 1.55211568e-01 9.31075692e-01 -4.52240407e-01
-1.16331208e+00 -3.51217896e-01 -1.94517002e-01 -8.78522098e-02
-1.91427633e-01 6.80069864e-01 -3.71782899e-01 1.35586321e-01
7.27781892e-01 1.17374487e-01 -5.34812093e-01 -9.25582945e-01
7.53236175e-01 9.50200856e-01 5.39429545e-01 -9.08873856e-01
6.28417671e-01 1.63043454e-01 -3.90108705e-01 -8.77638102e-01
-8.78756106e-01 -5.18831015e-01 -1.01503360e+00 3.56936842e-01
9.55163777e-01 -1.15134776e+00 -3.01538199e-01 3.37510377e-01
-1.28200138e+00 -4.00394887e-01 -2.07057372e-01 7.74079561e-01
-1.44582912e-01 1.15100019e-01 -7.68058836e-01 -3.20108861e-01
-4.24926102e-01 -6.60579205e-01 8.99155259e-01 -1.33813962e-01
-2.96946168e-01 -1.36570096e+00 4.66471136e-01 1.90838486e-01
3.51466656e-01 -1.38841122e-01 1.26698673e+00 -1.23206925e+00
-8.31252411e-02 2.57773966e-01 -5.22740856e-02 3.99502486e-01
2.93304294e-01 -2.32619837e-01 -5.64123094e-01 -4.81420040e-01
-4.87810433e-01 -4.80317980e-01 5.75316429e-01 -2.31777102e-01
8.60000253e-02 -2.05797717e-01 -4.01575238e-01 7.53241062e-01
1.38915241e+00 -1.67056933e-01 1.70395017e-01 3.77491176e-01
8.53390753e-01 8.59907269e-01 4.97592419e-01 -1.49382845e-01
7.50786185e-01 6.48896039e-01 -3.74000102e-01 -1.42247714e-02
-3.55608284e-01 -5.59419990e-01 7.47165620e-01 1.74770761e+00
1.58445075e-01 -1.95302993e-01 -1.41331911e+00 1.04569173e+00
-1.55335653e+00 -4.95466679e-01 6.64292499e-02 2.25971460e+00
1.47457242e+00 -1.98442996e-01 -1.06549278e-01 -7.25856185e-01
8.38744998e-01 -1.10095680e-01 -3.28117609e-01 -4.35540825e-01
-4.44347441e-01 4.83397692e-01 5.77401519e-01 6.12823248e-01
-7.82758594e-01 1.85201740e+00 5.27153111e+00 1.17481148e+00
-1.13068211e+00 4.29694831e-01 1.63853511e-01 5.61334798e-03
-5.00605881e-01 3.52312088e-01 -1.39233506e+00 4.48325187e-01
1.36621976e+00 -5.29224992e-01 5.16426086e-01 5.95626771e-01
-3.27587992e-01 3.07877809e-01 -1.37251163e+00 6.81558430e-01
7.62144625e-02 -1.04246330e+00 -3.75034139e-02 -1.16617559e-02
6.14741623e-01 8.73735130e-01 -3.26243699e-01 5.64692140e-01
8.92109632e-01 -8.06695044e-01 6.17283225e-01 3.17269787e-02
1.18263602e+00 -7.94132590e-01 6.12430274e-01 4.37919647e-01
-1.13640881e+00 4.49190706e-01 -4.64540064e-01 3.67158979e-01
4.07082975e-01 3.27609032e-01 -6.79005384e-01 8.99576485e-01
5.47611594e-01 7.54613221e-01 -5.00814140e-01 5.22896349e-01
-5.04618943e-01 7.43974745e-01 -5.93223929e-01 4.77136374e-02
4.14741635e-01 -3.61517258e-02 5.69593489e-01 1.61689854e+00
4.60803717e-01 -2.34927535e-01 2.20862359e-01 4.48085487e-01
-3.68334919e-01 8.94070029e-01 -7.99740076e-01 -1.47300914e-01
6.55106723e-01 1.11629224e+00 -1.91894129e-01 -3.29826623e-01
-7.12132752e-01 1.06672227e+00 9.14571762e-01 5.03640249e-02
-5.00517964e-01 -3.83644760e-01 7.26099432e-01 5.92780523e-02
3.80006343e-01 -3.18453789e-01 1.68409884e-01 -1.34810507e+00
-9.71964449e-02 -9.83083069e-01 5.46491861e-01 -4.56062615e-01
-2.00796294e+00 7.76772141e-01 3.14556360e-02 -8.04188073e-01
-4.94018346e-01 -6.29401684e-01 -2.02098563e-01 1.29534709e+00
-1.78177834e+00 -1.72970462e+00 5.67871928e-01 6.70926750e-01
2.03592479e-01 -4.88747835e-01 9.61784899e-01 6.10875368e-01
-2.39130065e-01 7.55002856e-01 3.39386463e-01 3.47538531e-01
1.23124921e+00 -8.81962836e-01 8.41898680e-01 6.49542749e-01
5.24731398e-01 7.89268553e-01 2.10886508e-01 -7.93212056e-01
-1.05637169e+00 -1.04794395e+00 1.59359026e+00 -3.83767903e-01
1.30253804e+00 -7.30207026e-01 -1.16381299e+00 9.31240976e-01
3.91132176e-01 -1.28897454e-03 9.93485272e-01 6.56725228e-01
-7.56978810e-01 1.21564388e-01 -9.02662396e-01 7.98633039e-01
1.43616998e+00 -9.84984040e-01 -8.77294183e-01 2.45277166e-01
9.72440183e-01 1.44561613e-02 -1.08882344e+00 2.10859120e-01
5.70281565e-01 -4.70216781e-01 6.77301884e-01 -1.19001138e+00
2.05246270e-01 -5.39528094e-02 -4.22785282e-01 -1.69847238e+00
-7.05636516e-02 -6.60479605e-01 6.00704074e-01 1.90576649e+00
7.08193839e-01 -8.12459886e-01 5.39162457e-02 -8.32401402e-03
-3.26795787e-01 -1.88453555e-01 -1.30685639e+00 -1.19388580e+00
8.58271241e-01 -3.08818191e-01 4.78263766e-01 1.52695489e+00
6.51014000e-02 8.23912919e-01 -2.28508636e-01 -7.88486674e-02
2.88663357e-01 -1.20410815e-01 6.26774251e-01 -1.20444739e+00
-4.42511737e-01 -1.27801895e-01 -2.55106479e-01 -5.93312621e-01
8.75467181e-01 -1.66020060e+00 5.65271489e-02 -1.10261142e+00
4.44126688e-02 -7.81043708e-01 -4.25957114e-01 7.61121333e-01
-2.45402023e-01 1.35453463e-01 3.53249967e-01 2.77405530e-01
-2.50843406e-01 4.81121510e-01 7.49465227e-01 4.12576914e-01
-1.55737758e-01 -6.36361420e-01 -6.34832144e-01 8.07437003e-01
7.32423067e-01 -8.32661808e-01 1.26862794e-01 -8.32195520e-01
3.46724510e-01 -8.51718038e-02 -1.64196745e-01 -5.79534471e-01
8.44664425e-02 -1.04528673e-01 -2.06572667e-01 -1.49961233e-01
-3.05135804e-03 -5.71474433e-01 1.75388642e-02 2.97989845e-01
-2.22200587e-01 1.56283513e-01 4.65013564e-01 2.66342368e-02
-2.37566441e-01 -3.35306138e-01 7.63580799e-01 -3.68089378e-01
-7.25032687e-01 2.19404981e-01 -1.28055379e-01 8.19404900e-01
6.35886252e-01 2.13847935e-01 -4.72622961e-01 1.14463836e-01
-3.59927416e-01 2.08953358e-02 6.73711479e-01 7.34466851e-01
-1.72013581e-01 -1.33763409e+00 -1.33644462e+00 9.81862620e-02
5.43281078e-01 -5.65816224e-01 -2.23008133e-02 6.33930564e-01
-7.37293586e-02 4.42292005e-01 -3.43894839e-01 -2.37973481e-01
-1.12266743e+00 4.22925115e-01 1.15831397e-01 -6.01371229e-01
-2.52210349e-01 7.63458133e-01 3.55791241e-01 -1.30279255e+00
-4.30660516e-01 1.04077563e-01 1.51197091e-01 1.49271697e-01
1.26683777e-02 1.26852438e-01 7.70088881e-02 -1.08879411e+00
-8.04306924e-01 5.62939465e-01 -2.99998701e-01 -4.34331626e-01
1.40011203e+00 -1.61474094e-01 -3.12804163e-01 6.19003475e-01
1.41204488e+00 5.93927324e-01 -4.81915921e-01 -5.48042834e-01
4.54568684e-01 -1.99055836e-01 -2.22224191e-01 -9.11629260e-01
-6.66302025e-01 9.49065626e-01 3.20302211e-02 -5.78318894e-01
6.48744464e-01 1.93856224e-01 9.29038703e-01 4.06250328e-01
6.68725669e-01 -1.20159411e+00 -5.55950761e-01 1.06893528e+00
6.54462814e-01 -1.18272257e+00 -3.59391272e-01 -1.97373286e-01
-7.37643898e-01 5.97099781e-01 4.73301947e-01 2.17550576e-01
5.23340404e-01 2.53513306e-01 1.85070321e-01 1.13278963e-01
-7.84806252e-01 -3.77355069e-01 3.05201560e-01 6.25003278e-01
9.33589876e-01 1.51925832e-01 -4.78327602e-01 5.89731812e-01
-5.33437014e-01 -2.63723910e-01 1.98391639e-02 7.33340383e-01
-1.07167631e-01 -1.48961830e+00 6.20643757e-02 2.66420152e-02
-8.08696091e-01 -1.01952970e+00 -3.50165159e-01 1.23683929e+00
3.49125713e-01 5.95298767e-01 9.29307714e-02 6.29371777e-02
3.02136838e-01 4.35965300e-01 5.16005337e-01 -9.44599330e-01
-1.07872546e+00 -3.33163179e-02 5.90557873e-01 -1.58607677e-01
-2.97397763e-01 -7.06337929e-01 -1.10702968e+00 -3.58276188e-01
-2.28442520e-01 2.58212715e-01 7.65879035e-01 8.32684338e-01
4.29453045e-01 -1.94089022e-02 2.47077614e-01 -4.17960227e-01
-2.70526022e-01 -9.98705864e-01 -5.37280083e-01 3.79024416e-01
-1.54846400e-01 -3.33757728e-01 -3.65026951e-01 -1.29465535e-01] | [10.778732299804688, 9.899868965148926] |
43ee1dc5-6a89-4bc4-9186-d2ec8f340ccd | good-news-vs-bad-news-what-are-they-talking | null | null | https://aclanthology.org/R17-1044 | https://aclanthology.org/R17-1044.pdf | Good News vs. Bad News: What are they talking about? | Today{'}s massive news streams demand the automate analysis which is provided by various online news explorers. However, most of them do not provide sentiment analysis. The main problem of sentiment analysis of news is the differences between the writers and readers attitudes to the news text. News can be good or bad but have to be delivered in neutral words as pure facts. Although there are applications for sentiment analysis of news, the task of news analysis is still a very actual problem because the latest news impacts people{'}s lives daily. In this paper, we explored the problem of sentiment analysis for Ukrainian and Russian news, developed a corpus of Ukrainian and Russian news and annotated each text using one of three categories: positive, negative and neutral. Each text was marked by at least three independent annotators via the web interface, the inter-annotator agreement was analyzed and the final label for each text was computed. These texts were used in the machine learning experiments. Further, we investigated what kinds of named entities such as Locations, Organizations, Persons are perceived as good or bad by the readers and which of them were the cause for text annotation ambiguity. | ['Olga Kanishcheva', 'Victoria Bobicev'] | 2017-09-01 | null | null | null | ranlp-2017-9 | ['text-annotation'] | ['natural-language-processing'] | [-3.85416150e-01 1.65834770e-01 -2.71307975e-01 -6.56280518e-01
-1.62247583e-01 -1.01806748e+00 7.11978316e-01 6.30981147e-01
-5.96023321e-01 8.95421565e-01 7.65294909e-01 -1.84865013e-01
4.38465089e-01 -7.11996555e-01 -2.08834603e-01 -4.79059637e-01
6.55353606e-01 4.35116023e-01 1.80081129e-01 -7.29142308e-01
5.94565332e-01 -1.58694535e-01 -1.42055631e+00 6.15384102e-01
7.20914900e-01 9.00367737e-01 -4.73448709e-02 3.37774724e-01
-9.98657227e-01 1.09669507e+00 -8.23398292e-01 -8.79613340e-01
-1.36429269e-03 -3.98090750e-01 -9.35086668e-01 1.99961830e-02
-1.61022812e-01 3.41952294e-01 5.34319162e-01 1.63838041e+00
3.21316212e-01 -1.61447972e-01 4.54695612e-01 -1.13201666e+00
-8.75855267e-01 1.01887226e+00 -5.21950603e-01 3.64342183e-01
4.90681350e-01 -6.38059914e-01 1.02790642e+00 -8.54095042e-01
9.05869901e-01 1.05008757e+00 5.68309367e-01 3.34117472e-01
-2.16820344e-01 -4.46729243e-01 3.69585484e-01 -1.05159637e-02
-8.78705382e-01 -2.13322565e-01 5.66561878e-01 -8.64725173e-01
6.74283803e-01 3.73822272e-01 7.83682346e-01 8.77229989e-01
5.02602994e-01 3.80970746e-01 1.41603279e+00 -3.97238314e-01
3.43748868e-01 9.85565066e-01 6.28255904e-01 8.14487040e-02
1.78356364e-01 -8.71610105e-01 -3.58157963e-01 -2.88254887e-01
-1.42617539e-01 8.53234455e-02 -2.84005046e-01 3.46900105e-01
-1.16770995e+00 9.95655119e-01 8.88759345e-02 6.09998822e-01
-4.85985070e-01 -6.43368185e-01 8.91738892e-01 5.74774265e-01
1.07806766e+00 6.04674816e-01 -8.36528897e-01 -3.37884039e-01
-3.30474526e-01 1.54659435e-01 1.14048398e+00 7.01865971e-01
4.46763694e-01 -3.31361353e-01 2.34690234e-01 8.05441380e-01
4.26039159e-01 6.64290011e-01 7.21267462e-01 -7.72093460e-02
4.20604795e-01 9.05110121e-01 5.13242781e-01 -1.59048975e+00
-5.90567350e-01 -5.14323711e-01 -5.02424479e-01 -4.29207012e-02
2.10707739e-01 -6.96071446e-01 -6.67576015e-01 1.27417612e+00
4.34142232e-01 -6.22918069e-01 4.22028303e-01 9.69198704e-01
1.24534059e+00 1.01301146e+00 1.78958729e-01 -6.23453379e-01
1.92145133e+00 -8.24832976e-01 -1.34287429e+00 -3.55280459e-01
7.61842668e-01 -1.52351344e+00 9.28117096e-01 4.16430712e-01
-7.57887900e-01 -1.66338772e-01 -7.15186000e-01 1.16680160e-01
-9.75425005e-01 1.08004473e-01 4.19974774e-01 6.04220629e-01
-6.32192850e-01 1.19039521e-01 -3.35439086e-01 -6.36519790e-01
-3.53797339e-02 -4.00503054e-02 -3.32747012e-01 6.26004875e-01
-1.61236370e+00 1.21936941e+00 3.99764657e-01 4.23610397e-02
1.72979385e-01 -7.38601573e-03 -6.50212526e-01 -3.16939056e-01
2.53005743e-01 -2.42679909e-01 1.28241920e+00 -1.70935798e+00
-1.22753906e+00 1.20629239e+00 -8.67102817e-02 -2.89044101e-02
3.18232834e-01 -1.58552095e-01 -1.23217773e+00 -2.83844501e-01
6.94200695e-01 -2.12280452e-01 3.46912116e-01 -1.06750500e+00
-1.16781890e+00 -4.40851957e-01 1.48826823e-01 4.39369231e-01
-4.36540127e-01 7.66727507e-01 -1.38212353e-01 -6.70102179e-01
2.58508563e-01 -7.86402822e-01 -1.51313812e-01 -5.52162170e-01
-3.13613564e-01 -3.36747766e-01 6.78103149e-01 -7.48199642e-01
1.25316095e+00 -2.06931186e+00 -3.18066537e-01 1.82592958e-01
2.84215137e-02 -2.26682182e-02 5.14046788e-01 5.85218906e-01
2.31479127e-02 3.56496066e-01 2.23593980e-01 1.08993560e-01
-1.25540763e-01 1.52780831e-01 -5.99696100e-01 4.53786701e-01
-3.61050487e-01 2.59391308e-01 -1.01546073e+00 -4.97141391e-01
-2.27834508e-01 2.64605045e-01 -3.43132913e-02 -1.11589156e-01
-1.85650259e-01 3.80235702e-01 -8.68278623e-01 5.49680352e-01
4.33402419e-01 -2.14018166e-01 7.65011832e-02 -2.87781805e-01
-7.14261591e-01 3.81504923e-01 -1.05747676e+00 9.31559145e-01
-2.90729940e-01 5.49595058e-01 9.60185751e-02 -7.14669287e-01
1.21586478e+00 5.93624949e-01 2.14167595e-01 -6.45213366e-01
5.53122282e-01 3.91517997e-01 -2.49885455e-01 -8.59294951e-01
1.07516098e+00 -4.05935973e-01 -5.27664721e-01 5.55520415e-01
-4.20926541e-01 -3.11039295e-02 3.17560732e-01 1.98127389e-01
4.78168339e-01 -3.35145026e-01 7.07332850e-01 -5.15960634e-01
5.34177542e-01 5.39250195e-01 7.23631978e-01 2.38111168e-01
-1.48568287e-01 2.42237031e-01 8.50114584e-01 -8.16928387e-01
-1.00813377e+00 -2.54557967e-01 -2.80385762e-01 1.42555368e+00
4.55945820e-01 -6.20504558e-01 -6.23095870e-01 -7.35834002e-01
-5.80949545e-01 7.86022723e-01 -6.49237156e-01 3.90947461e-01
-1.81390084e-02 -9.88680542e-01 8.25101733e-02 3.29706855e-02
2.96922982e-01 -9.40486073e-01 -4.79096621e-01 1.25939384e-01
-5.10438383e-01 -1.02971911e+00 -1.02446310e-01 8.37396681e-02
-2.17287198e-01 -8.05027783e-01 -5.46053350e-01 -8.89096141e-01
8.58184934e-01 4.50199582e-02 1.06872904e+00 -1.74209982e-01
5.36370218e-01 -2.12083325e-01 -1.07460093e+00 -1.06976891e+00
-4.88683790e-01 1.01147592e-01 2.83509284e-01 2.22523604e-02
6.15523279e-01 1.02230899e-01 -3.38743240e-01 4.37249601e-01
-8.21789503e-01 1.13264956e-01 1.51906863e-01 3.67595434e-01
2.38367051e-01 3.86854678e-01 6.86076045e-01 -1.58841407e+00
9.67688322e-01 -8.29352140e-01 -1.37116432e-01 1.59789667e-01
-5.26687562e-01 -2.40046233e-01 9.00092065e-01 -5.08332551e-02
-1.34232056e+00 -3.95656258e-01 -5.16255140e-01 8.15027654e-01
-8.74163657e-02 1.01897466e+00 1.20736726e-01 4.16463733e-01
7.50845671e-01 -2.55502760e-01 -4.13895845e-01 -1.85595661e-01
1.74797162e-01 1.23839974e+00 8.36789086e-02 -1.19357333e-01
3.16563457e-01 5.24298608e-01 -8.60074162e-01 -6.44033372e-01
-1.47905898e+00 -7.55379438e-01 -1.75612152e-01 -5.61706603e-01
1.08499062e+00 -9.43749011e-01 -3.45886737e-01 4.17786866e-01
-1.34662569e+00 4.84593242e-01 -9.58517343e-02 5.82069993e-01
1.53908014e-01 1.35565862e-01 -5.49741924e-01 -8.25236619e-01
-4.84019876e-01 -9.18539524e-01 3.65014285e-01 5.05480945e-01
-6.69729829e-01 -1.05838597e+00 2.71398038e-01 4.55270857e-01
3.86045605e-01 1.76638916e-01 7.64638841e-01 -1.30442369e+00
6.94460511e-01 -4.94749665e-01 -1.50197838e-02 2.40932584e-01
2.61530906e-01 2.45831147e-01 -8.07647467e-01 1.80423379e-01
4.44936007e-01 -2.69621998e-01 2.34379724e-01 3.90060395e-02
4.46045637e-01 -6.33744955e-01 -2.28532568e-01 -1.93794906e-01
1.18065965e+00 4.00333464e-01 4.57281530e-01 7.68725395e-01
4.47364092e-01 9.00051713e-01 9.45714414e-01 6.00496113e-01
4.69982147e-01 2.83130974e-01 2.29803607e-01 3.51872258e-02
7.42848396e-01 -1.25837531e-02 5.59704900e-01 1.29081213e+00
-2.45148063e-01 -5.08549929e-01 -1.00778997e+00 5.11234760e-01
-1.93795097e+00 -1.00514400e+00 -6.48985565e-01 1.59530485e+00
8.49805534e-01 4.59839165e-01 -1.93571687e-01 1.00414261e-01
1.07698238e+00 2.34604076e-01 6.45043626e-02 -7.15692699e-01
-3.85391563e-01 -4.41049844e-01 4.34251368e-01 2.63078094e-01
-1.04684436e+00 8.13346326e-01 5.78061056e+00 5.22949696e-01
-1.29656136e+00 1.34748951e-01 7.63423085e-01 3.67064327e-01
-5.72755337e-01 1.01774417e-01 -8.99709523e-01 5.62382638e-01
6.61417842e-01 -3.82172793e-01 -3.35537016e-01 1.25865853e+00
3.25705469e-01 -1.26204804e-01 -3.04136515e-01 6.49332702e-01
2.69989699e-01 -1.19115186e+00 -1.84746817e-01 -4.74130183e-01
9.83188868e-01 3.93180549e-02 -1.83170021e-01 2.04931840e-01
2.80027092e-01 -5.58254242e-01 9.40230429e-01 3.57982486e-01
1.57031178e-01 -7.75345266e-01 1.41410851e+00 4.87677306e-01
-7.88689554e-01 2.27496654e-01 -3.53238761e-01 -5.44692218e-01
5.22819996e-01 8.99129331e-01 -5.81273735e-01 3.91310334e-01
9.52811241e-01 7.21445918e-01 -3.96005422e-01 3.88025135e-01
-2.64176667e-01 5.06719768e-01 -1.19206555e-01 -7.76750326e-01
2.92856634e-01 -1.57618180e-01 4.09256697e-01 1.33915663e+00
3.76958758e-01 2.14889333e-01 3.31269354e-01 7.71521404e-02
6.99579529e-03 8.44321191e-01 -3.64842176e-01 -2.28603005e-01
1.85590550e-01 1.62069392e+00 -1.40454423e+00 -5.34947395e-01
-6.72976196e-01 6.90154254e-01 -3.36446725e-02 6.39202371e-02
-5.88385999e-01 -5.53043067e-01 1.93725094e-01 2.22834021e-01
-2.20155910e-01 4.26696122e-01 -3.73514265e-01 -1.29494178e+00
-3.98352519e-02 -8.93473864e-01 4.00461495e-01 -9.91890371e-01
-1.61090732e+00 1.13461256e+00 -3.65601867e-01 -1.14967084e+00
1.51015550e-01 -5.97996414e-01 -5.73005438e-01 5.36530018e-01
-1.25948381e+00 -6.65422082e-01 -2.15974092e-01 3.70011270e-01
4.54401344e-01 -1.91506401e-01 8.39954853e-01 3.51574987e-01
-3.91140729e-01 -1.79699063e-01 2.00130925e-01 2.93407649e-01
9.78311360e-01 -1.13132596e+00 1.98900521e-01 6.19615674e-01
-2.31113896e-01 4.92074609e-01 1.44812167e+00 -9.36576724e-01
-6.74689710e-01 -6.62632823e-01 1.63212454e+00 -4.37913179e-01
1.09384239e+00 -1.77118387e-02 -6.15027368e-01 7.22846448e-01
5.98457873e-01 -2.96282619e-01 1.15888333e+00 3.23829383e-01
-1.13437958e-01 2.79723704e-01 -1.01592553e+00 5.82977414e-01
3.10521841e-01 -2.04809636e-01 -1.00391483e+00 7.03579962e-01
6.01066172e-01 -5.14725804e-01 -6.84017003e-01 -1.80878676e-02
2.76026189e-01 -8.10420752e-01 2.97987849e-01 -7.67705262e-01
7.45555282e-01 -6.50855362e-01 -3.98350619e-02 -1.54999399e+00
-8.62440020e-02 -1.19468562e-01 8.08939040e-01 1.48555481e+00
8.01658452e-01 -7.52403378e-01 3.03025395e-01 6.73726082e-01
-2.37489492e-01 -3.64287317e-01 -3.77960235e-01 2.57340842e-03
-1.35733262e-01 -5.23066401e-01 4.05826151e-01 1.54469717e+00
7.62855470e-01 7.04617620e-01 -5.10692596e-01 -3.53431664e-02
-1.88858375e-01 1.61619961e-01 5.04474759e-01 -1.31252444e+00
2.38351986e-01 -2.54959941e-01 -4.44296807e-01 -4.80008721e-01
-3.54261920e-02 -5.16664088e-01 2.85002906e-02 -1.84156811e+00
8.75318944e-02 -2.72895187e-01 2.98515372e-02 2.87912428e-01
-1.03981949e-01 1.26853108e-01 -1.63109183e-01 3.42962652e-01
-7.17328191e-01 4.92218137e-02 1.11265981e+00 1.15734003e-01
-1.08148135e-01 1.53003946e-01 -1.22534692e+00 1.31312478e+00
9.03530717e-01 -6.28477812e-01 -1.68430775e-01 -2.46897772e-01
1.66005886e+00 -2.50581205e-01 -1.80181086e-01 -4.58018363e-01
3.78282875e-01 -5.07520437e-01 1.50137454e-01 -5.20036578e-01
-1.19827136e-01 -8.79271209e-01 1.94058031e-01 1.64537281e-01
-4.15603578e-01 4.08374548e-01 -2.62342811e-01 2.47936845e-01
-6.60883665e-01 -5.24575889e-01 6.88413501e-01 -3.73688757e-01
-7.88433373e-01 -1.08494349e-01 -7.57518828e-01 2.94722319e-01
1.20902467e+00 -1.08765410e-02 -7.13600934e-01 -5.27190089e-01
-7.94497132e-01 2.51827717e-01 5.22968769e-01 4.03275192e-01
1.45240650e-01 -1.15605795e+00 -7.25952923e-01 -3.58362824e-01
4.19401556e-01 -2.28156060e-01 3.09846669e-01 8.16749275e-01
-7.24545956e-01 5.22995740e-02 -2.15977326e-01 -5.02237491e-02
-1.29133582e+00 4.37479556e-01 -2.50668209e-02 -4.98917587e-02
-2.66924143e-01 7.01666772e-01 -6.07200824e-02 -5.50593853e-01
-1.02826543e-01 -1.14552289e-01 -1.26790679e+00 7.71527231e-01
8.11696827e-01 7.19140396e-02 -5.66339083e-02 -1.34786713e+00
-3.83088201e-01 6.13138258e-01 -3.35005522e-01 -1.41503707e-01
1.31436872e+00 -5.19282639e-01 -6.01390362e-01 9.74337459e-01
9.07045007e-01 4.80886906e-01 -1.55686677e-01 -1.14810593e-01
9.94136184e-02 -3.50063652e-01 -5.98870739e-02 -8.22409391e-01
-8.56597126e-01 2.31236890e-01 1.08175553e-01 1.07111740e+00
8.08418870e-01 1.70059308e-01 6.41280949e-01 3.38743418e-01
1.28867045e-01 -1.80149364e+00 -4.89861041e-01 8.87462854e-01
6.30820334e-01 -1.49035656e+00 -4.82957577e-03 -5.29900312e-01
-1.43173015e+00 1.22256219e+00 3.58039558e-01 1.14202559e-01
9.47591305e-01 3.81012559e-01 8.84528339e-01 -5.63898742e-01
-5.51323771e-01 1.71900675e-01 1.18162863e-01 1.04013123e-01
1.02015281e+00 -2.45355293e-02 -1.07682991e+00 1.00394797e+00
-7.66602993e-01 -1.80829123e-01 9.42459941e-01 1.01519454e+00
-7.46671140e-01 -6.68419003e-01 -4.34138119e-01 4.40450907e-01
-1.22210324e+00 1.03256896e-01 -5.60961843e-01 3.40220243e-01
3.98878396e-01 1.21572506e+00 1.12859324e-01 -3.36301416e-01
4.09307152e-01 -2.00816482e-01 -5.44380128e-01 -7.23003387e-01
-9.36384678e-01 -3.42499427e-02 5.76734960e-01 8.08249637e-02
-9.01455760e-01 -5.49938083e-01 -1.50140750e+00 -3.21261972e-01
-5.30613840e-01 9.35511291e-01 1.08141708e+00 1.30207431e+00
-5.61757386e-02 2.23466828e-01 6.78768933e-01 -3.88152897e-02
-9.43874940e-02 -8.78078520e-01 -6.11568987e-01 6.65924311e-01
1.30652979e-01 -3.35609347e-01 -6.25024557e-01 2.40988806e-01] | [11.083269119262695, 6.9499125480651855] |
a8ac150b-f4ec-4c27-8c4a-8a22f32f7f95 | pager-progressive-attribute-guided-extendable | 2206.00162 | null | https://arxiv.org/abs/2206.00162v2 | https://arxiv.org/pdf/2206.00162v2.pdf | PAGER: Progressive Attribute-Guided Extendable Robust Image Generation | This work presents a generative modeling approach based on successive subspace learning (SSL). Unlike most generative models in the literature, our method does not utilize neural networks to analyze the underlying source distribution and synthesize images. The resulting method, called the progressive attribute-guided extendable robust image generative (PAGER) model, has advantages in mathematical transparency, progressive content generation, lower training time, robust performance with fewer training samples, and extendibility to conditional image generation. PAGER consists of three modules: core generator, resolution enhancer, and quality booster. The core generator learns the distribution of low-resolution images and performs unconditional image generation. The resolution enhancer increases image resolution via conditional generation. Finally, the quality booster adds finer details to generated images. Extensive experiments on MNIST, Fashion-MNIST, and CelebA datasets are conducted to demonstrate generative performance of PAGER. | ['C. -C. Jay Kuo', 'Zohreh Azizi'] | 2022-06-01 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 6.05626225e-01 3.03722918e-02 -7.38338828e-02 -2.27228194e-01
-9.81073260e-01 -4.94200706e-01 8.45479071e-01 -7.77018547e-01
4.44243886e-02 1.01639807e+00 3.26756269e-01 3.29250954e-02
7.34058395e-02 -9.22961414e-01 -7.84932375e-01 -8.85476172e-01
3.51758093e-01 3.01969618e-01 -2.65757918e-01 1.13294609e-02
-3.07974387e-02 2.99878657e-01 -1.26957834e+00 3.53668183e-01
1.14804780e+00 7.19784915e-01 4.64333713e-01 6.52645469e-01
-2.23660059e-02 9.33405638e-01 -5.06867945e-01 -3.42571855e-01
1.94937497e-01 -8.47083330e-01 -4.34651911e-01 5.24185300e-01
2.32771769e-01 -6.48346782e-01 -4.95823413e-01 1.10000098e+00
6.36762619e-01 1.11913905e-01 8.06183100e-01 -1.29242051e+00
-1.60501289e+00 8.27585161e-01 -7.54631102e-01 -3.23605947e-02
2.72974521e-01 3.33588153e-01 6.76696420e-01 -1.20427859e+00
8.81616890e-01 1.47614729e+00 3.52072358e-01 7.56473839e-01
-1.72536039e+00 -7.26633430e-01 -8.81549269e-02 -1.19568497e-01
-1.40437138e+00 -3.83626938e-01 9.39061761e-01 -2.72693098e-01
3.99114132e-01 2.41856307e-01 5.97248077e-01 1.52239716e+00
4.60817441e-02 9.29791808e-01 1.54709637e+00 -3.40568662e-01
3.21284801e-01 1.68471709e-01 -3.81515086e-01 5.55702269e-01
2.14838445e-01 3.19573104e-01 -6.40471697e-01 -2.59760134e-02
1.44058168e+00 -7.18139410e-02 -3.23729962e-01 -3.50327730e-01
-1.12359822e+00 9.19309437e-01 4.13236469e-01 -2.00675488e-01
-4.24394339e-01 2.10696176e-01 -2.46421441e-01 2.77308989e-02
5.34908056e-01 2.27065265e-01 1.52699202e-01 3.99933070e-01
-1.13375747e+00 2.37130895e-01 4.77566361e-01 1.23432946e+00
7.47695148e-01 8.14718485e-01 -5.39766371e-01 9.06435966e-01
3.84129316e-01 8.43576491e-01 5.90694606e-01 -1.29941249e+00
1.96107373e-01 3.28855693e-01 -9.75770652e-02 -7.31706202e-01
1.86603263e-01 -6.68627739e-01 -1.21331429e+00 3.37002695e-01
-1.78819835e-01 -2.67723471e-01 -1.33198583e+00 1.77788496e+00
2.96547078e-02 9.86146275e-03 3.15601110e-01 8.44636321e-01
1.14470541e+00 9.35856104e-01 6.58668876e-02 -1.51603132e-01
9.08355713e-01 -9.07891393e-01 -8.38004351e-01 -5.44083156e-02
-2.43179724e-01 -8.69113326e-01 1.08695650e+00 2.04226524e-01
-1.36269474e+00 -8.85134161e-01 -1.06469870e+00 -1.33014724e-01
-1.09357856e-01 4.93319422e-01 7.38916755e-01 5.68771899e-01
-1.14958274e+00 3.15552413e-01 -4.01099592e-01 -4.67512272e-02
6.90421760e-01 2.00486425e-02 -1.57230973e-01 -1.26948103e-01
-1.03872359e+00 2.76187748e-01 4.27616537e-01 -1.54153183e-01
-1.39228857e+00 -6.98020041e-01 -1.02929103e+00 -4.27088775e-02
-2.52428278e-02 -1.15438533e+00 9.47186291e-01 -1.08702660e+00
-1.66814077e+00 7.42766321e-01 3.46604851e-03 -4.68969047e-01
6.64710581e-01 -1.31372139e-01 -1.64945632e-01 1.34584561e-01
5.47774099e-02 1.25451243e+00 1.35315943e+00 -1.77612686e+00
-2.80609339e-01 -3.36510763e-02 -3.10926497e-01 3.25721174e-01
-1.08329274e-01 -3.71933758e-01 -4.04650897e-01 -1.03707337e+00
1.45483583e-01 -9.30247188e-01 -3.04272205e-01 -1.86149687e-01
-6.63324833e-01 2.16663122e-01 7.61145115e-01 -7.39812672e-01
8.01295221e-01 -2.24042368e+00 2.23286778e-01 1.31463632e-01
2.54965454e-01 -2.28621215e-02 -5.52411020e-01 2.19974190e-01
-2.15276495e-01 5.18678874e-02 -1.37050956e-01 -1.85687438e-01
1.93646476e-02 4.75012399e-02 -8.21320117e-01 7.78161064e-02
2.08405703e-01 1.27800071e+00 -6.46311581e-01 -4.47756410e-01
1.11033551e-01 8.93439054e-01 -5.35838902e-01 1.64127588e-01
-2.02332273e-01 4.09059942e-01 -2.09289312e-01 7.01075375e-01
7.90377319e-01 -3.99866223e-01 7.32031977e-03 -4.75109905e-01
8.32046494e-02 -2.65284449e-01 -1.11870599e+00 1.65107155e+00
-2.12736621e-01 4.57680821e-01 -1.41605183e-01 -3.70000303e-01
1.19099760e+00 1.18700102e-01 1.58633113e-01 -7.16255069e-01
-1.32605582e-01 -8.99800658e-02 -3.12790126e-01 -2.61300594e-01
7.78615475e-01 -2.97415890e-02 2.83240806e-03 4.61782515e-01
2.94045210e-01 -1.43877402e-01 2.39810318e-01 5.70207417e-01
5.80469608e-01 6.66453421e-01 1.88901331e-02 -1.05046220e-01
3.13763410e-01 -2.71681935e-01 4.76601124e-01 6.13409281e-01
1.60235643e-01 9.22806919e-01 2.27257147e-01 -2.16125503e-01
-1.47880042e+00 -1.81204522e+00 1.46929517e-01 9.50868845e-01
2.38423660e-01 -1.82903439e-01 -8.85449946e-01 -2.72035688e-01
-3.71363163e-01 9.44980264e-01 -5.69848716e-01 -2.37757817e-01
-3.84273261e-01 -9.97209132e-01 3.64887595e-01 6.40985787e-01
1.02952170e+00 -1.27106762e+00 -1.55756131e-01 -6.17672876e-02
-6.73828781e-01 -1.00495851e+00 -6.96264029e-01 -3.05106223e-01
-9.94158030e-01 -5.21480739e-01 -9.85968232e-01 -8.28563273e-01
1.04129326e+00 2.37126604e-01 1.12266719e+00 -4.17599171e-01
-5.09284198e-01 4.37358201e-01 -1.26594216e-01 -2.56093442e-01
-7.22211480e-01 -8.24447498e-02 -1.67918444e-01 -5.16177155e-02
-2.09725976e-01 -6.34825945e-01 -5.95330477e-01 1.19380705e-01
-1.06953597e+00 5.41129470e-01 1.06562507e+00 9.84539330e-01
1.17252946e+00 2.38833189e-01 6.42594576e-01 -1.02053761e+00
7.98270285e-01 -1.53343588e-01 -4.76517320e-01 1.64040014e-01
-7.20464945e-01 1.92407683e-01 4.64149177e-01 -6.42276168e-01
-1.69872117e+00 2.23490521e-01 4.44460250e-02 -5.00973523e-01
-1.67250052e-01 2.17705797e-02 -2.86418080e-01 1.74749240e-01
8.12446535e-01 9.86150801e-01 3.64763588e-02 -3.68986279e-01
7.79451013e-01 3.37743253e-01 8.99231970e-01 -5.19532740e-01
1.33941031e+00 6.56915784e-01 -1.97524548e-01 -6.53665304e-01
-6.32242501e-01 3.02252352e-01 -4.53239053e-01 -1.84528217e-01
8.24979424e-01 -1.28606570e+00 -9.17566568e-02 5.02298415e-01
-8.06734741e-01 -4.10118461e-01 -6.07064009e-01 2.46325105e-01
-9.60193038e-01 8.54448974e-02 -7.31306791e-01 -7.54218996e-01
-6.60063922e-01 -9.38984752e-01 9.33834970e-01 4.99407530e-01
-5.84574565e-02 -6.50831640e-01 -1.96083754e-01 4.65503842e-01
6.38861954e-01 3.79997820e-01 8.10388863e-01 6.47578984e-02
-1.16673303e+00 1.32050693e-01 -3.24856699e-01 3.99056137e-01
2.15545475e-01 -6.84663057e-02 -8.39602351e-01 -3.91411990e-01
-2.85902709e-01 -3.77805173e-01 1.08415532e+00 5.80268919e-01
1.18480480e+00 -6.23384356e-01 -8.97787586e-02 9.76929545e-01
1.45657051e+00 1.35140240e-01 1.20363855e+00 1.38579637e-01
7.51430333e-01 3.14349771e-01 2.98984408e-01 2.38489956e-01
2.94924408e-01 2.47140676e-01 1.47659153e-01 -5.90503871e-01
-7.73894250e-01 -6.01257741e-01 5.72458744e-01 5.75918019e-01
-3.59681696e-01 -1.58017516e-01 -2.44770765e-01 3.27874541e-01
-1.59629762e+00 -1.34439039e+00 2.60665685e-01 1.86234260e+00
1.00030291e+00 -1.45109728e-01 1.21352248e-01 -2.09424078e-01
8.05017829e-01 1.37534235e-02 -7.15525448e-01 1.13868853e-03
-5.53950250e-01 -1.24500003e-02 1.94457710e-01 2.82294720e-01
-9.79968071e-01 1.00950038e+00 6.89051104e+00 1.07549930e+00
-7.96666980e-01 -2.77760196e-02 1.10967815e+00 -8.41493085e-02
-5.91914058e-01 -2.21313536e-01 -7.88197160e-01 3.59569699e-01
4.71594602e-01 -1.88286006e-01 7.79263794e-01 9.21982229e-01
1.53065860e-01 1.55258730e-01 -7.67404258e-01 1.07041514e+00
3.19363058e-01 -1.45570517e+00 6.43256307e-01 9.27070826e-02
1.19033957e+00 -3.41764063e-01 8.45008135e-01 1.23824596e-01
7.35612094e-01 -1.17558384e+00 8.29908192e-01 7.81593084e-01
1.13590467e+00 -1.04328370e+00 2.68052042e-01 9.27737802e-02
-9.78508890e-01 -4.34488170e-02 -5.78173578e-01 4.91536885e-01
1.68553397e-01 5.15278757e-01 -5.21091998e-01 4.21079814e-01
6.54624403e-01 4.22280818e-01 -7.14852035e-01 6.45933867e-01
-5.15929759e-01 5.80797791e-01 2.17927381e-01 4.17994261e-01
-1.40989274e-01 -3.97167742e-01 6.65200770e-01 1.04480648e+00
4.82847273e-01 -5.99654876e-02 9.76874158e-02 1.54668331e+00
-3.32156777e-01 -1.57721400e-01 -5.67816556e-01 -2.31286287e-02
4.91150856e-01 1.53132033e+00 -8.17371786e-01 -3.42832714e-01
-3.31829973e-02 1.35466647e+00 -3.01088095e-02 8.33438098e-01
-8.84441495e-01 -1.83073208e-01 2.20345467e-01 1.08913444e-01
4.11192238e-01 -1.88269362e-01 -2.95703620e-01 -1.03025687e+00
-3.66264343e-01 -1.14841831e+00 1.48651004e-01 -1.18898463e+00
-1.33273196e+00 6.89725220e-01 4.89685461e-02 -1.17573416e+00
-4.28880692e-01 -1.66798785e-01 -4.47127998e-01 9.80143249e-01
-1.38640594e+00 -1.67890644e+00 -4.89772141e-01 7.09228396e-01
8.12989831e-01 -4.69247252e-01 7.03323126e-01 8.13934430e-02
-5.04243970e-01 5.24842620e-01 9.18357149e-02 1.74070224e-01
5.48421025e-01 -1.27002990e+00 3.42023224e-01 9.76614237e-01
2.85882205e-01 7.11872220e-01 6.62174344e-01 -7.65381813e-01
-1.26072741e+00 -1.51022804e+00 2.86688805e-01 -2.90049732e-01
1.36136875e-01 -3.70592445e-01 -5.67451239e-01 6.82628930e-01
5.62894821e-01 -3.36151242e-01 7.46559024e-01 -2.86702216e-01
-5.39092541e-01 -1.45999625e-01 -1.07112324e+00 8.00316393e-01
9.64574516e-01 -2.89988071e-01 -2.12247193e-01 8.89267698e-02
7.29136348e-01 -3.92012745e-01 -7.69541323e-01 2.59885103e-01
4.47303742e-01 -8.58619630e-01 1.34529865e+00 -8.96636769e-02
8.86530757e-01 -4.28774178e-01 -2.37132788e-01 -1.38820672e+00
-8.53578925e-01 -6.75224841e-01 -3.65554392e-01 1.59700346e+00
4.52755243e-01 -3.29480827e-01 5.95911324e-01 3.50456685e-01
1.22395165e-01 -5.23662269e-01 -3.93932283e-01 -8.22677255e-01
-4.80460376e-02 9.65534896e-02 7.70206094e-01 6.98418856e-01
-6.37524426e-01 5.37888646e-01 -7.83419430e-01 -4.60993638e-03
1.29255760e+00 5.74931562e-01 7.92718947e-01 -9.01427388e-01
-5.68477929e-01 -2.77096987e-01 9.70252156e-02 -9.41392303e-01
-6.12527877e-02 -8.79611611e-01 5.94688281e-02 -1.78390145e+00
7.74514318e-01 -1.06831484e-01 -9.82729066e-03 4.04378593e-01
-2.95346677e-01 7.36373544e-01 2.79162556e-01 3.93407524e-01
-5.80194950e-01 7.52079308e-01 1.69605803e+00 -1.93153173e-01
-2.70429760e-01 -2.36939862e-01 -1.06121123e+00 6.42847598e-01
7.87026286e-01 -1.06050305e-01 -8.13070536e-01 -3.14258844e-01
-3.48152071e-02 -1.48953721e-01 6.79723203e-01 -8.38989794e-01
-1.45179346e-01 -1.50765911e-01 1.27910483e+00 -7.76457906e-01
3.68322730e-01 -2.65496731e-01 6.79803133e-01 3.47650588e-01
-5.48516393e-01 -1.00992016e-01 -2.95257401e-02 5.23553669e-01
6.66891644e-03 1.49623573e-01 1.06574917e+00 -3.29248041e-01
-7.23271430e-01 4.12472486e-01 -3.20781916e-01 -8.52303132e-02
8.29706490e-01 -3.80268842e-01 -4.19515610e-01 -6.29881978e-01
-7.88730264e-01 -1.76027626e-01 4.44973260e-01 4.47171152e-01
1.11209464e+00 -1.77988207e+00 -1.07941377e+00 3.79558414e-01
-1.50479183e-01 -9.63452905e-02 5.34976840e-01 2.76036441e-01
-2.32500300e-01 1.28248155e-01 -3.01550746e-01 -4.97776628e-01
-1.03662682e+00 6.22704685e-01 -2.71342732e-02 -1.78388283e-01
-6.70991480e-01 8.82041454e-01 6.52213931e-01 -2.37916455e-01
-1.12216763e-01 3.46220970e-01 -7.92465657e-02 -1.57090694e-01
5.89808941e-01 5.05679548e-01 -5.68834007e-01 -7.86898673e-01
2.16497883e-01 2.69531995e-01 -1.75236568e-01 -3.65831286e-01
1.29944050e+00 -3.67163658e-01 -7.28804991e-02 1.61873341e-01
7.62697875e-01 -2.01654941e-01 -1.61930335e+00 -2.56239414e-01
-7.20464706e-01 -4.77527857e-01 1.29799858e-01 -9.86699700e-01
-1.22930348e+00 5.28691351e-01 7.20660985e-01 -1.68175921e-02
1.51108873e+00 -3.12915482e-02 7.35679686e-01 1.09613221e-02
2.34342679e-01 -1.06527209e+00 6.08426869e-01 1.86994806e-01
1.14633107e+00 -8.67685974e-01 1.12158552e-01 -4.37803149e-01
-8.68726671e-01 8.11014354e-01 8.55675936e-01 -1.47192255e-01
2.26827845e-01 4.72046316e-01 4.08733301e-02 1.27469733e-01
-6.88662410e-01 -1.12581037e-01 4.30270106e-01 1.09466481e+00
2.86438972e-01 4.41914946e-02 -1.22217648e-02 4.82421964e-01
-3.97472262e-01 4.16614115e-02 5.58396876e-01 4.98649210e-01
-3.90068799e-01 -9.14106309e-01 -5.18477798e-01 3.64096612e-01
-3.07257235e-01 -3.35046589e-01 -2.27267772e-01 5.46450317e-01
1.16501406e-01 6.78663254e-01 -5.91055825e-02 -2.36714333e-01
-9.84526724e-02 -1.36191517e-01 7.17311561e-01 -3.13423574e-01
1.19551076e-02 5.55275202e-01 -3.70784014e-01 -4.16295886e-01
-3.90658438e-01 -5.28678656e-01 -1.01040494e+00 -2.08265796e-01
-2.31132075e-01 -1.69426426e-01 4.83262271e-01 4.91910070e-01
4.82429475e-01 7.90036678e-01 6.78815782e-01 -8.57062221e-01
-3.78130972e-01 -8.56535137e-01 -6.29328609e-01 4.68559921e-01
5.56349196e-03 -2.24962354e-01 -2.30217446e-02 7.30065286e-01] | [11.58603572845459, -0.5429757237434387] |
4f1dd253-45f0-475d-8d8f-bea445ceddae | cl4ac-a-contrastive-loss-for-audio-captioning | 2107.09990 | null | https://arxiv.org/abs/2107.09990v3 | https://arxiv.org/pdf/2107.09990v3.pdf | CL4AC: A Contrastive Loss for Audio Captioning | Automated Audio captioning (AAC) is a cross-modal translation task that aims to use natural language to describe the content of an audio clip. As shown in the submissions received for Task 6 of the DCASE 2021 Challenges, this problem has received increasing interest in the community. The existing AAC systems are usually based on an encoder-decoder architecture, where the audio signal is encoded into a latent representation, and aligned with its corresponding text descriptions, then a decoder is used to generate the captions. However, training of an AAC system often encounters the problem of data scarcity, which may lead to inaccurate representation and audio-text alignment. To address this problem, we propose a novel encoder-decoder framework called Contrastive Loss for Audio Captioning (CL4AC). In CL4AC, the self-supervision signals derived from the original audio-text paired data are used to exploit the correspondences between audio and texts by contrasting samples, which can improve the quality of latent representation and the alignment between audio and texts, while trained with limited data. Experiments are performed on the Clotho dataset to show the effectiveness of our proposed approach. | ['Wenwu Wang', 'Mark D. Plumbley', 'H Lilian Tang', 'Tom Ko', 'Xinhao Mei', 'Qiushi Huang', 'Xubo Liu'] | 2021-07-21 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 6.96200430e-01 1.48901224e-01 1.19282585e-02 -2.63115764e-01
-1.39653146e+00 -4.26844269e-01 4.60631609e-01 -4.22427356e-02
1.19682692e-01 5.99748492e-01 8.75521421e-01 2.88215399e-01
2.76665926e-01 -9.30995345e-02 -8.26887429e-01 -5.36455095e-01
2.75034398e-01 2.52989411e-01 -3.08386594e-01 9.21852291e-02
-1.29622906e-01 -4.11217242e-01 -1.44868827e+00 8.09611082e-01
9.04753268e-01 1.12676632e+00 3.98890257e-01 4.39166665e-01
-2.92096198e-01 7.93342233e-01 -4.91904974e-01 -3.55887800e-01
1.10890225e-01 -8.70846689e-01 -7.44572639e-01 1.03469089e-01
2.70962775e-01 -2.18709856e-01 -3.47349733e-01 1.06155849e+00
6.50667846e-01 -1.21501088e-01 4.38637137e-01 -1.41647983e+00
-4.48894560e-01 9.08194005e-01 -3.23675543e-01 -1.80027321e-01
7.20682263e-01 -2.06130192e-01 1.25236952e+00 -1.02279222e+00
3.68945330e-01 1.25266016e+00 5.15556633e-01 5.84948659e-01
-1.16385138e+00 -8.03401113e-01 4.29860055e-02 4.03653294e-01
-1.40150428e+00 -9.16491807e-01 1.00866830e+00 -4.26003724e-01
3.33529651e-01 2.79556185e-01 5.74176610e-01 1.36028206e+00
-1.54773891e-01 1.11805642e+00 7.03922868e-01 -5.12190759e-01
-7.55694732e-02 1.44362539e-01 -4.10586804e-01 1.03240825e-01
-5.29476881e-01 -1.89605877e-01 -9.78537381e-01 -1.31954178e-01
2.72791982e-01 -3.74946266e-01 -6.02982402e-01 -1.65850490e-01
-1.40750611e+00 5.04849136e-01 2.00883299e-01 8.74169096e-02
-4.53927517e-01 -7.49970824e-02 6.96160257e-01 3.36110026e-01
4.63797301e-01 3.90228868e-01 -3.50506157e-02 -5.26528358e-01
-1.09148943e+00 2.17375219e-01 5.09937406e-01 1.00985205e+00
2.74686813e-01 -2.93727666e-02 -3.83153260e-01 1.05271494e+00
5.15303016e-01 4.37237620e-01 8.05245697e-01 -7.03214228e-01
1.18596590e+00 1.84356645e-01 1.88114285e-01 -9.93525803e-01
2.42822677e-01 -3.54875416e-01 -8.75947893e-01 -4.56269234e-01
-4.48339107e-03 -1.65993989e-01 -4.96473730e-01 1.90315056e+00
8.87739435e-02 4.75552261e-01 3.68395299e-01 1.14915168e+00
6.14293754e-01 1.26845574e+00 -4.51343618e-02 -4.28921789e-01
1.13097215e+00 -1.17293918e+00 -1.17150879e+00 -2.75757253e-01
3.54754567e-01 -1.25232327e+00 1.12834489e+00 2.60988384e-01
-1.18840492e+00 -6.98026955e-01 -1.05622721e+00 -4.41401228e-02
2.37400368e-01 2.69471377e-01 -1.90766469e-01 4.09648605e-02
-5.62254429e-01 2.57347196e-01 -5.21963894e-01 6.88072070e-02
8.73638690e-02 4.03271392e-02 -3.42324227e-01 -1.06224142e-01
-1.49765587e+00 3.33948106e-01 4.74910706e-01 4.69360292e-01
-1.00310647e+00 -6.99243188e-01 -8.55317593e-01 1.75877526e-01
2.12890729e-01 -3.88608724e-01 1.39206696e+00 -1.52128506e+00
-1.63805151e+00 3.88258159e-01 -2.24966466e-01 -5.68548024e-01
5.50493538e-01 -5.24693549e-01 -5.82262814e-01 2.10425466e-01
3.41224432e-01 7.69136429e-01 1.09015429e+00 -1.10426247e+00
-5.83592176e-01 2.23473106e-02 -3.72271508e-01 6.00094736e-01
-4.29783285e-01 7.38417879e-02 -4.98536885e-01 -8.90087903e-01
2.99616884e-02 -9.56049263e-01 3.22018623e-01 -1.81525484e-01
-4.73420233e-01 -3.07395905e-02 9.64182973e-01 -1.13915169e+00
1.22709405e+00 -2.72174025e+00 5.16663730e-01 -1.14863522e-01
-1.70968607e-01 1.51903331e-01 -2.79755890e-01 6.40206337e-01
-2.68443227e-01 -2.47992426e-01 -4.31617230e-01 -6.51286423e-01
1.66935429e-01 -1.16913117e-01 -9.52390313e-01 2.03335181e-01
2.73364097e-01 4.82945800e-01 -8.78030121e-01 -6.75138652e-01
-1.33091703e-01 5.82010031e-01 -3.74572337e-01 8.26052606e-01
-3.78933519e-01 6.64155900e-01 -3.00603092e-01 2.84267396e-01
3.48198950e-01 4.57556024e-02 1.54151376e-02 -3.25958252e-01
-4.73817214e-02 6.34873450e-01 -9.21453416e-01 1.96346772e+00
-6.40197515e-01 8.98538411e-01 1.46199703e-01 -8.33378553e-01
9.77880836e-01 1.06602836e+00 5.68330169e-01 -5.90691984e-01
-9.85284299e-02 4.06962872e-01 -2.25286275e-01 -6.40489459e-01
2.68738538e-01 -2.01303616e-01 -1.55022487e-01 2.99510807e-01
-1.44419774e-01 -2.84408361e-01 -4.03623730e-02 9.09999683e-02
6.84185803e-01 2.43530244e-01 -1.02324896e-01 3.12171638e-01
7.71674871e-01 -1.35547578e-01 4.58698481e-01 7.49322623e-02
2.94234604e-02 1.05482101e+00 4.10148978e-01 -1.44122643e-02
-1.30908179e+00 -8.24909627e-01 1.42390400e-01 7.69788444e-01
-2.57724337e-02 -6.84700787e-01 -8.85591209e-01 -4.39144284e-01
-3.66894007e-01 6.03821933e-01 -3.27881664e-01 -2.98688859e-01
-5.10997832e-01 -2.76987981e-02 6.84223175e-01 3.00484389e-01
5.06666780e-01 -1.05070019e+00 -5.91105744e-02 4.50762868e-01
-1.09071076e+00 -1.39996469e+00 -1.00218952e+00 -2.68856794e-01
-5.02200067e-01 -6.61590338e-01 -1.07856250e+00 -1.07648730e+00
5.10226011e-01 1.21301822e-01 6.37714922e-01 -4.21576858e-01
1.73925847e-01 1.82377920e-01 -6.27597570e-01 -2.90495247e-01
-6.76959455e-01 2.27334455e-01 1.05379999e-01 7.70826459e-01
8.13114420e-02 -6.63986564e-01 -3.71547073e-01 2.86517143e-01
-8.59439135e-01 7.33414412e-01 7.20693827e-01 8.79199564e-01
6.00158274e-01 -3.33231419e-01 7.92072296e-01 -4.44175214e-01
8.30295444e-01 -5.37380517e-01 -3.39637369e-01 1.79749027e-01
-2.04534695e-01 4.05541398e-02 9.77551341e-01 -5.80807388e-01
-9.96693015e-01 3.25021982e-01 -9.80458632e-02 -7.46512473e-01
1.19345032e-01 7.35421181e-01 -5.05403578e-01 5.93090773e-01
2.46515676e-01 4.98219281e-01 9.98659283e-02 -6.80912971e-01
2.05564752e-01 1.40477324e+00 8.38758171e-01 -5.04965186e-01
7.63411522e-01 -3.11559942e-02 -3.98151070e-01 -6.91958189e-01
-9.25936222e-01 -4.52251673e-01 -3.35486859e-01 -3.24347705e-01
7.69394577e-01 -1.27065408e+00 -7.95470998e-02 2.92623550e-01
-1.46167600e+00 -1.86849069e-02 -2.50826448e-01 7.17936754e-01
-8.37765038e-01 3.90767992e-01 -3.39518160e-01 -7.73540258e-01
-5.20291030e-01 -1.36235464e+00 1.27616847e+00 -1.19132467e-01
-4.28851098e-01 -5.11627316e-01 1.73232630e-01 6.06420815e-01
1.83838561e-01 8.73010159e-02 9.27335024e-01 -5.56984127e-01
-3.52409124e-01 -3.52958620e-01 -1.16987363e-01 5.09055138e-01
3.61663342e-01 -3.71738583e-01 -1.07213569e+00 -3.24669093e-01
-5.76428361e-02 -5.80532968e-01 3.80667090e-01 -5.57736345e-02
1.15411186e+00 -6.59713030e-01 -2.49044206e-02 2.87519127e-01
8.90321791e-01 3.77862677e-02 5.53210914e-01 4.70654517e-02
6.62527084e-01 7.63825119e-01 7.66748726e-01 4.00471151e-01
2.64906526e-01 1.17903543e+00 3.06892425e-01 9.55333859e-02
-3.84319067e-01 -8.02085638e-01 8.18969488e-01 1.61319268e+00
4.21018004e-01 -1.69878215e-01 -7.94647455e-01 7.42109478e-01
-1.93485689e+00 -7.30286062e-01 -2.34013144e-02 2.05004621e+00
1.22607076e+00 -4.61041220e-02 4.14760560e-02 3.79876673e-01
9.65789139e-01 1.78098813e-01 -3.57588083e-01 1.63778907e-03
1.12868898e-01 -2.97939479e-01 -1.76648229e-01 3.68130296e-01
-8.27209711e-01 5.37891746e-01 5.41055822e+00 9.34932888e-01
-1.19928133e+00 1.96338326e-01 2.82724738e-01 -1.13483481e-01
-2.86804706e-01 4.11036983e-02 -3.55897486e-01 8.35531533e-01
1.14346063e+00 -4.51653928e-01 4.69517112e-01 5.82899630e-01
5.10623574e-01 5.04304826e-01 -1.44958270e+00 1.29682755e+00
1.80646002e-01 -9.65773821e-01 2.98857659e-01 -2.17876017e-01
6.42700315e-01 -3.56325358e-01 2.87978590e-01 2.38158897e-01
-4.11482960e-01 -9.15630519e-01 1.10307777e+00 3.57888997e-01
1.00786757e+00 -5.42001724e-01 8.07188988e-01 2.18608558e-01
-1.31415033e+00 -1.07634412e-02 -1.42298162e-01 1.45484686e-01
3.83631706e-01 3.29416752e-01 -9.84305382e-01 7.63738155e-01
4.73463774e-01 7.88134396e-01 -1.62773088e-01 1.13958907e+00
-1.21449694e-01 8.78311336e-01 -1.23020202e-01 1.91979930e-01
2.58344471e-01 -1.77527349e-02 7.76979625e-01 1.08748496e+00
4.87050533e-01 -3.16435874e-01 5.58298118e-02 7.48455942e-01
-3.29674125e-01 5.28982937e-01 -4.23513532e-01 -4.36864823e-01
4.55740660e-01 8.25053513e-01 5.57405539e-02 -2.06552520e-01
-2.77003109e-01 1.09995198e+00 5.70428232e-03 2.69214600e-01
-8.62647057e-01 -5.40039420e-01 3.52228492e-01 6.98692277e-02
3.38770561e-02 -1.44253420e-02 1.15309529e-01 -1.18050742e+00
4.65254158e-01 -1.24828947e+00 1.39340714e-01 -1.12387550e+00
-1.18938649e+00 7.67304540e-01 -2.99170776e-03 -1.93794048e+00
-4.42203879e-01 1.01381257e-01 -4.34699774e-01 8.76061141e-01
-1.38285553e+00 -1.14704454e+00 -2.18804017e-01 4.99648511e-01
9.63120759e-01 -3.28382522e-01 7.69920051e-01 7.19801724e-01
-2.92326957e-01 7.07673073e-01 2.57181436e-01 1.74809158e-01
9.88032460e-01 -9.27222431e-01 2.66515285e-01 6.28301740e-01
3.33369583e-01 1.90699816e-01 9.06834722e-01 -4.97780353e-01
-1.20961142e+00 -1.32072997e+00 1.01877820e+00 -3.15590849e-04
5.67191362e-01 -5.03313959e-01 -1.01837516e+00 5.68510294e-01
4.59912747e-01 -2.00266257e-01 7.80406713e-01 -2.88539439e-01
-2.61188298e-01 -3.49047780e-01 -5.39282918e-01 5.81356287e-01
5.84132493e-01 -9.78806734e-01 -7.52897739e-01 3.20944905e-01
1.03083444e+00 -5.18698037e-01 -6.49001479e-01 1.82908684e-01
6.31319702e-01 -2.66930491e-01 5.64236522e-01 -4.41308498e-01
7.06137657e-01 -3.51973534e-01 -2.48217657e-01 -1.38206935e+00
1.60985127e-01 -1.01289463e+00 1.27050681e-02 1.63370955e+00
4.94745135e-01 -3.86746414e-03 3.78908396e-01 1.26285508e-01
-4.12988275e-01 -3.93580854e-01 -1.21674037e+00 -5.71975708e-01
-3.31247240e-01 -3.81909400e-01 6.91076815e-01 9.10096169e-01
2.47456864e-01 6.80477917e-01 -9.04946148e-01 2.41833702e-01
3.82776052e-01 4.61176001e-02 8.00120950e-01 -9.30316687e-01
-2.40128830e-01 -5.23293465e-02 -2.49663368e-01 -1.26707447e+00
3.12685072e-01 -9.21714723e-01 5.35773754e-01 -1.32789624e+00
-2.65004039e-02 -1.78505078e-01 -1.13719761e-01 3.06579709e-01
2.46078409e-02 2.16797829e-01 3.77229214e-01 5.60259521e-01
-5.80592036e-01 1.16815925e+00 9.88234460e-01 -4.43893671e-01
-2.50244051e-01 7.15449378e-02 -3.40239525e-01 3.96242291e-01
5.41052282e-01 -5.87003767e-01 -4.28079039e-01 -5.62901437e-01
1.88641295e-01 5.05796134e-01 1.37507170e-01 -1.23119175e+00
2.64146745e-01 1.32294059e-01 -5.90601824e-02 -5.11431098e-01
6.41759813e-01 -1.11148512e+00 3.21904778e-01 1.55357674e-01
-8.78795922e-01 -4.55660224e-02 9.12703648e-02 6.18297458e-01
-9.43541944e-01 -2.22143650e-01 4.90491211e-01 2.86817670e-01
-3.74107361e-02 1.57976463e-01 -3.10460418e-01 7.27248192e-02
7.19805539e-01 1.39982313e-01 9.76455361e-02 -8.02568138e-01
-6.65885210e-01 3.44204396e-01 1.53703406e-03 7.01068640e-01
6.32898808e-01 -1.92842114e+00 -1.11358190e+00 1.89909503e-01
3.87174010e-01 5.92624694e-02 2.45307073e-01 7.59314835e-01
-1.69544011e-01 4.77699190e-01 -1.21024095e-01 -5.70634723e-01
-1.22586191e+00 2.98375368e-01 2.32981503e-01 -1.55851737e-01
-6.55477345e-01 4.91308898e-01 1.70441598e-01 -3.76497917e-02
5.26246071e-01 -6.97055683e-02 -1.62909344e-01 8.15203115e-02
5.73245883e-01 4.57335934e-02 -5.06714918e-02 -9.01520729e-01
-6.37596399e-02 3.69921803e-01 -8.63132551e-02 -6.13437116e-01
1.20761108e+00 -4.22801107e-01 2.88659725e-02 7.48669624e-01
1.34536910e+00 6.44561499e-02 -1.20857632e+00 -4.28991050e-01
-1.31475002e-01 -4.10293072e-01 -8.93361717e-02 -5.16253173e-01
-7.07784534e-01 1.15970743e+00 5.81835747e-01 2.25986943e-01
1.11467373e+00 -2.01265827e-01 1.24629283e+00 1.40146226e-01
-3.31573784e-02 -1.01620305e+00 2.86241949e-01 3.34538132e-01
1.35122538e+00 -9.34729576e-01 -4.33917671e-01 -2.68027902e-01
-8.12862277e-01 1.11483622e+00 5.01218736e-01 3.17037374e-01
1.25341028e-01 -9.22516510e-02 1.81633577e-01 3.72091144e-01
-7.62943149e-01 2.60670096e-01 4.04572248e-01 4.20566499e-01
4.29974705e-01 -1.01546444e-01 -3.13712545e-02 9.21691358e-01
-4.85965222e-01 -3.54951131e-03 4.09241706e-01 6.01937294e-01
-8.78828168e-02 -1.31455517e+00 -4.93121415e-01 5.13344258e-02
-4.36458468e-01 -1.40610531e-01 -4.99282777e-01 4.48664203e-02
-1.21641204e-01 9.63043034e-01 -1.38426609e-02 -4.33633238e-01
3.93663257e-01 4.28303838e-01 4.66461666e-02 -5.47172070e-01
-3.51733267e-01 3.40539873e-01 -2.79520079e-03 -2.44840786e-01
-4.43411618e-01 -6.72225535e-01 -9.82778966e-01 2.96192378e-01
-2.84807652e-01 8.69424820e-01 6.31247818e-01 9.69399214e-01
4.76943552e-01 5.79744697e-01 9.80862319e-01 -7.14856029e-01
-5.99604249e-01 -1.16854560e+00 -3.90022069e-01 5.18939257e-01
7.17822969e-01 -3.56403470e-01 -4.09678787e-01 6.16764367e-01] | [15.258832931518555, 4.912491798400879] |
f5461dcc-6579-416d-8954-b0d04911be5d | domain-adaptation-for-semantic-segmentation | 1810.07911 | null | http://arxiv.org/abs/1810.07911v2 | http://arxiv.org/pdf/1810.07911v2.pdf | Domain Adaptation for Semantic Segmentation via Class-Balanced Self-Training | Recent deep networks achieved state of the art performance on a variety of
semantic segmentation tasks. Despite such progress, these models often face
challenges in real world `wild tasks' where large difference between labeled
training/source data and unseen test/target data exists. In particular, such
difference is often referred to as `domain gap', and could cause significantly
decreased performance which cannot be easily remedied by further increasing the
representation power. Unsupervised domain adaptation (UDA) seeks to overcome
such problem without target domain labels. In this paper, we propose a novel
UDA framework based on an iterative self-training procedure, where the problem
is formulated as latent variable loss minimization, and can be solved by
alternatively generating pseudo labels on target data and re-training the model
with these labels. On top of self-training, we also propose a novel
class-balanced self-training framework to avoid the gradual dominance of large
classes on pseudo-label generation, and introduce spatial priors to refine
generated labels. Comprehensive experiments show that the proposed methods
achieve state of the art semantic segmentation performance under multiple major
UDA settings. | ['Yang Zou', 'Zhiding Yu', 'Jinsong Wang', 'B. V. K. Vijaya Kumar'] | 2018-10-18 | null | null | null | null | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 6.56756222e-01 3.01732749e-01 -2.60967523e-01 -5.97421348e-01
-8.99663746e-01 -5.11608958e-01 5.61260879e-01 -3.09260469e-02
-6.19961679e-01 8.29793155e-01 -1.00218348e-01 2.28959303e-02
-6.59581199e-02 -6.50997937e-01 -6.65153861e-01 -9.07675445e-01
5.75258136e-01 7.09908307e-01 4.40089583e-01 1.51843518e-01
1.06357023e-01 4.36284430e-02 -1.40464556e+00 -1.11262593e-02
1.32006466e+00 1.03655028e+00 4.54754174e-01 -4.15389724e-02
-5.62905967e-01 4.74743962e-01 -7.89242625e-01 -2.67197728e-01
1.46951810e-01 -5.61944723e-01 -1.10307932e+00 5.07736146e-01
2.34965771e-01 -1.11813441e-01 9.09675285e-02 1.32890213e+00
4.96766806e-01 3.49716336e-01 7.51457512e-01 -1.14869571e+00
-6.81663632e-01 4.24971700e-01 -9.33488131e-01 1.96859129e-02
-2.17248708e-01 -1.56077668e-01 8.28607559e-01 -6.49008095e-01
5.94180703e-01 1.21165335e+00 6.12993598e-01 8.41624260e-01
-1.45014966e+00 -7.22004890e-01 3.93889427e-01 3.28555256e-02
-1.30341017e+00 -2.29510143e-01 1.09168994e+00 -2.99542248e-01
3.64492923e-01 -6.84956908e-02 2.48513266e-01 1.24493504e+00
-3.78281057e-01 1.05447471e+00 1.28149211e+00 -4.79863793e-01
5.68069756e-01 2.18514591e-01 2.75246769e-01 4.12730634e-01
1.46538362e-01 -2.26927862e-01 -2.77873516e-01 5.82938045e-02
7.36779392e-01 -2.69516230e-01 -1.17779307e-01 -6.79037988e-01
-1.04178929e+00 8.71275067e-01 4.12957460e-01 1.96431175e-01
-2.59184003e-01 -2.37598822e-01 3.76699001e-01 -1.40022123e-02
8.35201323e-01 2.80102611e-01 -5.51422060e-01 1.74900696e-01
-9.26450849e-01 1.11786157e-01 3.15385133e-01 9.81562197e-01
9.05827820e-01 4.79818434e-02 -1.68563932e-01 1.48093057e+00
4.45690870e-01 2.45107338e-01 7.41789758e-01 -8.61149013e-01
4.32429880e-01 5.91920435e-01 1.47385538e-01 -6.97014451e-01
-3.16628426e-01 -7.14910567e-01 -7.88830996e-01 8.82560611e-02
6.75688326e-01 -2.27903858e-01 -1.39589500e+00 2.00892353e+00
5.91437697e-01 4.46631312e-01 1.72636449e-01 9.83093321e-01
6.76737905e-01 6.25414550e-01 3.56568068e-01 -2.66879737e-01
1.12548614e+00 -1.24675179e+00 -7.99680412e-01 -6.17466450e-01
6.12113416e-01 -5.65401196e-01 1.03755438e+00 2.31288895e-01
-8.38604689e-01 -7.86775351e-01 -1.03014612e+00 -1.13903575e-01
-2.69942671e-01 6.19000122e-02 4.32907045e-01 5.41322351e-01
-9.17029738e-01 4.83712971e-01 -7.60632813e-01 -3.03323865e-01
7.65904605e-01 3.01235139e-01 -1.93856850e-01 -1.84562206e-01
-1.07663631e+00 7.06323385e-01 6.32302105e-01 2.00996712e-01
-9.68033254e-01 -6.39002621e-01 -8.67660046e-01 -2.92657912e-01
6.79580688e-01 -4.33037102e-01 1.36464834e+00 -1.21934104e+00
-1.43887436e+00 1.26726520e+00 -1.83007166e-01 -3.66084933e-01
5.63294590e-01 -3.71671110e-01 -3.25899154e-01 -1.05095215e-01
5.15632927e-01 1.00285947e+00 9.09364045e-01 -1.67495847e+00
-6.87909067e-01 -5.19473851e-01 -4.71475758e-02 4.27584469e-01
-2.91600913e-01 -1.48971736e-01 -5.44018686e-01 -8.36305797e-01
4.97951210e-01 -8.89796555e-01 -3.77411157e-01 -1.17703624e-01
-4.29497421e-01 -4.63396758e-01 9.00675237e-01 -4.54479933e-01
9.87221301e-01 -2.17884231e+00 1.31666645e-01 -2.60618292e-02
1.58058122e-01 5.37350774e-01 -2.23245025e-01 -1.07617699e-01
-1.38427779e-01 -3.22878063e-02 -8.32385719e-01 -3.77868950e-01
2.32784953e-02 4.12607193e-01 -3.76206130e-01 3.65576178e-01
2.97622114e-01 6.47212446e-01 -1.09371340e+00 -7.04434097e-01
4.08282466e-02 4.10855025e-01 -2.93052912e-01 2.77005285e-01
-5.88233471e-01 8.36006343e-01 -6.80404842e-01 5.13545811e-01
9.52627718e-01 -3.03168327e-01 1.37431487e-01 2.49093343e-02
1.60503536e-01 1.85024291e-01 -1.18486941e+00 1.95329738e+00
-3.10685426e-01 3.56818199e-01 1.91356361e-01 -1.48189008e+00
1.08377838e+00 1.42279610e-01 3.74924272e-01 -7.50833869e-01
2.81567276e-01 2.46729478e-01 -2.25855976e-01 -3.63588125e-01
2.25608677e-01 -2.88610488e-01 5.49614569e-03 3.42915475e-01
6.94651976e-02 2.60066483e-02 2.96921097e-02 -1.11811772e-01
7.06519246e-01 5.56023300e-01 -3.04895379e-02 -2.66473114e-01
5.81287801e-01 2.42318928e-01 1.01685894e+00 5.93283415e-01
-4.61661279e-01 6.21770680e-01 4.33742255e-01 -1.40314639e-01
-9.27557290e-01 -1.03781962e+00 -2.65983731e-01 1.28865480e+00
4.93223846e-01 1.98303401e-01 -1.11461461e+00 -1.04924572e+00
-2.54930407e-01 6.69170678e-01 -5.81328630e-01 -1.42299950e-01
-5.60381770e-01 -1.07238257e+00 4.33886200e-01 6.07838392e-01
9.44049478e-01 -1.18561983e+00 -3.24539185e-01 3.66978019e-01
-2.46732324e-01 -1.08841014e+00 -2.80603021e-01 3.11667442e-01
-9.65062916e-01 -7.85577834e-01 -1.08668113e+00 -1.25017059e+00
8.91390979e-01 3.33363622e-01 1.00305200e+00 -2.94649303e-01
-5.00597619e-02 -1.00460120e-01 -4.76117194e-01 -2.25493923e-01
-3.05099547e-01 9.19246897e-02 -1.00880191e-01 2.31297597e-01
3.30232710e-01 -5.12890637e-01 -6.77103162e-01 5.09594440e-01
-1.08089411e+00 1.05218440e-01 4.73240256e-01 1.03013957e+00
8.80012572e-01 2.52763093e-01 8.86255741e-01 -1.23048151e+00
4.74574149e-01 -5.92172205e-01 -4.20263648e-01 2.64615148e-01
-6.82615280e-01 1.68715328e-01 5.19168854e-01 -4.50784266e-01
-1.46443915e+00 2.47081611e-02 -1.34692594e-01 -4.61046994e-01
-4.95712280e-01 3.02054256e-01 -5.31718194e-01 1.27874300e-01
5.95481753e-01 9.85296443e-02 -1.71473935e-01 -7.10140049e-01
3.56142998e-01 7.82598436e-01 5.90895295e-01 -9.03035283e-01
6.56737328e-01 5.36985397e-01 -1.97047725e-01 -4.68103409e-01
-1.31439710e+00 -5.71553111e-01 -7.71419346e-01 3.85809056e-02
1.07000434e+00 -8.81899655e-01 6.24601096e-02 8.31468940e-01
-9.18711245e-01 -5.62703669e-01 -2.77737230e-01 2.43359700e-01
-4.79065895e-01 3.91126812e-01 -2.58447468e-01 -5.88564217e-01
-3.92728522e-02 -1.22642004e+00 1.04164767e+00 4.70627725e-01
-1.56603716e-02 -1.24988377e+00 -5.79397194e-03 6.16354942e-01
2.02075735e-01 2.51878172e-01 8.40776563e-01 -7.98300683e-01
-2.62478501e-01 1.13765270e-01 -5.80641091e-01 6.33480012e-01
3.58371437e-01 -6.00708961e-01 -1.11706638e+00 -3.74661744e-01
1.37322485e-01 -5.10747313e-01 8.82831693e-01 4.14093852e-01
1.26606715e+00 4.02916595e-02 -4.96915907e-01 6.37712836e-01
1.30896831e+00 3.30052882e-01 2.71813691e-01 4.36907738e-01
8.74601603e-01 8.15892279e-01 8.87032807e-01 1.71177566e-01
3.41272414e-01 6.38184786e-01 2.34501630e-01 -2.69711077e-01
-2.93851078e-01 -2.00727835e-01 -1.57890424e-01 6.83232903e-01
2.12587655e-01 -2.26499602e-01 -1.01048422e+00 8.11961472e-01
-1.88110495e+00 -4.70424265e-01 3.38779017e-02 2.09206319e+00
9.73224461e-01 2.22269118e-01 9.47656035e-02 4.45329733e-02
1.07188642e+00 2.10000217e-01 -9.23397243e-01 -2.62259617e-02
-1.57688811e-01 2.47077912e-01 4.97283220e-01 4.05962914e-01
-1.44269383e+00 1.21955121e+00 5.78694868e+00 1.09944177e+00
-1.11155152e+00 3.88105184e-01 8.62466455e-01 3.43956411e-01
-2.82433033e-01 -8.43892768e-02 -7.14659750e-01 5.81273973e-01
6.98157310e-01 -3.85017172e-02 1.25196338e-01 9.47547317e-01
4.47558500e-02 -8.53834897e-02 -8.79149973e-01 8.29330266e-01
5.77807240e-02 -8.63103449e-01 -9.43279192e-02 -7.37087578e-02
9.68903840e-01 -2.14366511e-01 1.00276165e-01 3.76240730e-01
5.35927176e-01 -6.01766586e-01 5.89611113e-01 -6.93803728e-02
8.37374449e-01 -6.85867906e-01 5.53661466e-01 4.70601708e-01
-9.10198390e-01 -1.46887847e-03 -4.08161342e-01 2.23562285e-01
1.15599945e-01 5.64573526e-01 -7.04320490e-01 4.72770393e-01
6.15476906e-01 7.13595033e-01 -4.52752173e-01 1.01097250e+00
-3.65048885e-01 7.34858215e-01 -2.19420388e-01 3.77519965e-01
4.80892330e-01 -3.79137546e-01 3.73451978e-01 9.28359151e-01
1.15720786e-01 -3.59222032e-02 2.63616323e-01 8.69305372e-01
-1.15580000e-01 1.34955034e-01 -4.01868403e-01 1.47533134e-01
5.12165487e-01 9.25554216e-01 -1.20857775e+00 -3.53911251e-01
-3.32049757e-01 1.22459340e+00 4.11864161e-01 6.06375039e-01
-9.74518776e-01 -9.46222916e-02 3.99594545e-01 4.15483210e-03
1.76477686e-01 8.73641968e-02 -4.11372453e-01 -1.04376078e+00
3.55215259e-02 -5.15240133e-01 4.01092768e-01 -6.23463333e-01
-1.43989706e+00 4.79870975e-01 1.66343570e-01 -1.19865203e+00
-7.26924017e-02 -4.79683518e-01 -4.37772512e-01 7.67547429e-01
-1.68727577e+00 -1.07382417e+00 -1.99185044e-01 5.37840366e-01
7.83939838e-01 1.85695142e-02 5.57493925e-01 5.32284498e-01
-7.96262205e-01 7.05602467e-01 3.96573007e-01 1.16754882e-01
8.58627677e-01 -1.37172711e+00 2.67915279e-01 7.99989998e-01
-9.32435021e-02 3.43113214e-01 5.36907554e-01 -7.08363950e-01
-6.21697009e-01 -1.31532788e+00 5.57797372e-01 -1.08411901e-01
4.05847013e-01 -4.12936121e-01 -1.13929844e+00 7.16272712e-01
5.40947318e-02 1.62245408e-01 5.59060812e-01 1.40018284e-01
-2.51644492e-01 -1.00520670e-01 -1.31149721e+00 3.62179428e-01
1.16868162e+00 -2.56836414e-01 -7.70762086e-01 4.01259750e-01
8.41319501e-01 -4.12906289e-01 -5.56006014e-01 6.72235906e-01
3.52760777e-02 -7.36315310e-01 9.50971365e-01 -4.17742848e-01
3.22460860e-01 -3.29031110e-01 7.47602656e-02 -1.35799384e+00
-1.59438729e-01 -3.82467419e-01 1.81668416e-01 1.56292307e+00
2.56441355e-01 -6.57344103e-01 1.09153008e+00 5.69205284e-01
-4.02545631e-01 -6.46572292e-01 -9.88948107e-01 -7.88294375e-01
3.22980821e-01 -2.43458495e-01 5.21413267e-01 1.25701261e+00
-4.51687187e-01 3.63807619e-01 -8.90694782e-02 1.44453943e-01
8.28908682e-01 7.43711293e-02 3.90955299e-01 -1.27237582e+00
-1.49310008e-01 -3.72364759e-01 -1.19910851e-01 -1.22626591e+00
3.34825605e-01 -8.11019957e-01 4.94590908e-01 -1.59987557e+00
1.03312425e-01 -8.68051231e-01 -5.28419554e-01 5.38862050e-01
-3.43177140e-01 2.92416036e-01 -5.23191355e-02 4.12463963e-01
-7.22099960e-01 6.44491136e-01 1.43094683e+00 -1.42582506e-01
-2.57997066e-01 3.61938815e-04 -7.04966128e-01 8.64051640e-01
7.43861020e-01 -5.12970269e-01 -8.28769684e-01 -7.47758150e-01
-2.58464873e-01 -1.10647812e-01 2.65682042e-01 -1.01552320e+00
-2.48705670e-02 -1.88498974e-01 1.79832116e-01 -4.37733591e-01
1.90931842e-01 -6.45039022e-01 -9.26463753e-02 1.01133129e-02
-3.71450514e-01 -4.48773414e-01 1.92784116e-01 6.88562155e-01
-3.83563399e-01 -4.12787825e-01 1.12749743e+00 -1.12020567e-01
-1.01774681e+00 1.23328730e-01 9.59466677e-03 2.57170409e-01
1.18146777e+00 -3.79841894e-01 -1.88243613e-01 1.15372673e-01
-9.71082091e-01 3.98088783e-01 4.12676424e-01 3.77134353e-01
2.52114028e-01 -1.25478494e+00 -4.68276978e-01 1.82431147e-01
1.16716363e-02 7.18363881e-01 4.96724665e-01 5.00855386e-01
-3.00504506e-01 2.47715622e-01 -9.32055861e-02 -6.77671015e-01
-8.23141396e-01 4.82203364e-01 1.85826764e-01 -3.61192465e-01
-5.79835773e-01 1.20284700e+00 6.94311082e-01 -6.75159693e-01
3.95075738e-01 1.22611020e-02 -2.65208364e-01 1.95868656e-01
2.40098223e-01 1.88527316e-01 -7.85128549e-02 -8.34605515e-01
-2.28643611e-01 6.02780998e-01 -3.69132698e-01 -7.59550259e-02
1.21373272e+00 -3.54961485e-01 1.69617385e-01 3.16920519e-01
1.11965132e+00 -3.68901521e-01 -1.71798003e+00 -5.00473559e-01
4.64314111e-02 -3.73788834e-01 6.85541928e-02 -1.00155532e+00
-1.20098019e+00 9.79146421e-01 6.57609820e-01 4.67461459e-02
1.27421069e+00 7.09375367e-03 1.02472520e+00 3.81194316e-02
2.60291010e-01 -1.42038369e+00 1.67286873e-01 3.08718711e-01
4.48060632e-01 -1.42668009e+00 -3.57638627e-01 -6.40787661e-01
-6.85896933e-01 5.17039001e-01 1.00325024e+00 -1.89805627e-01
4.80355263e-01 -1.63666457e-01 3.48349392e-01 5.63361961e-03
-1.83856457e-01 -3.27319741e-01 5.40671982e-02 7.00739145e-01
1.68615699e-01 -3.31026465e-02 -3.65562290e-01 7.62348652e-01
1.03509158e-01 -9.66755599e-02 1.02536440e-01 1.01301467e+00
-4.35711533e-01 -1.31373286e+00 -4.52095538e-01 2.23802119e-01
-4.17901099e-01 1.15096897e-01 -9.72198918e-02 6.64104223e-01
4.61990356e-01 8.09793174e-01 1.13631159e-01 -6.70069456e-02
2.08822891e-01 2.36722127e-01 2.66945899e-01 -7.57407606e-01
-5.43045625e-02 4.05102372e-01 -1.49799600e-01 -2.55247027e-01
-6.15387917e-01 -7.66693115e-01 -1.46063554e+00 2.31014132e-01
-3.79026830e-01 2.61043936e-01 5.56939185e-01 1.12428117e+00
2.95566410e-01 6.47171617e-01 5.63308418e-01 -6.37000322e-01
-5.03085077e-01 -1.03103983e+00 -5.89353621e-01 7.88743496e-01
1.74031302e-01 -1.05532050e+00 -2.72866696e-01 2.59016722e-01] | [9.636284828186035, 1.3823785781860352] |
ca4fd891-346d-4239-a967-e4f1f79c67c7 | criminal-investigation-tracker-with-suspect | 2302.10423 | null | https://arxiv.org/abs/2302.10423v1 | https://arxiv.org/pdf/2302.10423v1.pdf | Criminal Investigation Tracker with Suspect Prediction using Machine Learning | An automated approach to identifying offenders in Sri Lanka would be better than the current system. Obtaining information from eyewitnesses is one of the less reliable approaches and procedures still in use today. Automated criminal identification has the ability to save lives, notwithstanding Sri Lankan culture's lack of awareness of the issue. Using cutting-edge technology like biometrics to finish this task would be the most accurate strategy. The most notable outcomes will be obtained by applying fingerprint and face recognition as biometric techniques. The main responsibilities will be image optimization and criminality. CCTV footage may be used to identify a person's fingerprint, identify a person's face, and identify crimes involving weapons. Additionally, we unveil a notification system and condense the police report to Additionally, to make it simpler for police officers to understand the essential points of the crime, we develop a notification system and condense the police report. Additionally, if an incident involving a weapon is detected, an automated notice of the crime with all the relevant facts is sent to the closest police station. The summarization of the police report is what makes this the most original. In order to improve the efficacy of the overall image, the system will quickly and precisely identify the full crime scene, identify, and recognize the suspects using their faces and fingerprints, and detect firearms. This study provides a novel approach for crime prediction based on real-world data, and criminality incorporation. A crime or occurrence should be reported to the appropriate agencies, and the suggested web application should be improved further to offer a workable channel of communication. | ['Pradeepa Bandara', 'D. D. G. Delgasdeniya', 'S. P. S. Mandula', 'Erandika Lakmali', 'R. A. T. M. Rajapaksha', 'S. J. Dilmini'] | 2023-02-21 | null | null | null | null | ['crime-prediction'] | ['miscellaneous'] | [ 4.22333598e-01 -4.30302352e-01 -2.16738567e-01 -3.77394885e-01
-3.34734857e-01 -7.56190300e-01 1.63937107e-01 2.08628982e-01
-5.74866951e-01 5.59295356e-01 -3.47842835e-02 -6.44203365e-01
-4.08906490e-01 -9.37148809e-01 -2.97885891e-02 -5.82293987e-01
1.72797069e-01 3.91726464e-01 -3.16169634e-02 5.04615083e-02
8.63807023e-01 1.05519772e+00 -1.06097829e+00 2.04218611e-01
5.92807591e-01 6.08503997e-01 2.02121511e-01 4.46168602e-01
-7.33302021e-03 4.57513988e-01 -4.30910259e-01 -6.47551179e-01
3.44384402e-01 6.65976182e-02 -7.56735861e-01 -1.77616328e-02
2.17404693e-01 -1.15405631e+00 -8.35902020e-02 8.33687484e-01
4.37580675e-01 -1.26490355e-01 6.82271659e-01 -7.41004527e-01
-4.28143501e-01 1.25416264e-01 -8.34796250e-01 2.63702393e-01
8.25336516e-01 1.15786754e-01 2.12019131e-01 -5.84899485e-01
3.76172811e-01 9.93044794e-01 6.44095004e-01 6.34263933e-01
-9.08353925e-01 -9.92259443e-01 -4.55889463e-01 2.62341529e-01
-1.39219749e+00 -6.50595903e-01 5.68461657e-01 -4.89657789e-01
7.99299896e-01 2.17756182e-01 4.94341284e-01 7.25565970e-01
6.50823340e-02 1.98249131e-01 1.06107569e+00 -6.84161007e-01
-2.21704811e-01 3.92520905e-01 3.27079147e-01 8.36266696e-01
5.68010926e-01 7.19920844e-02 -2.55863458e-01 -3.55224013e-01
5.61071515e-01 4.99173582e-01 6.35480583e-02 4.18677926e-01
-4.64106083e-01 6.23282135e-01 -2.55233824e-01 4.32857096e-01
-4.80013996e-01 -4.23818082e-01 3.48786891e-01 -1.48038372e-01
2.42847592e-01 2.58427233e-01 1.31856412e-01 -5.33206165e-01
-9.30225551e-01 -7.57330358e-02 3.68212432e-01 3.49352509e-01
5.60841441e-01 -2.03658000e-01 1.83647022e-01 8.49133492e-01
2.37480700e-01 6.98230743e-01 -2.17591807e-01 -7.11118639e-01
7.19822943e-01 8.10927749e-01 5.19080181e-03 -1.34079146e+00
-1.59006149e-01 2.38399819e-01 -5.11700153e-01 2.84971446e-01
5.39501607e-01 -2.48563990e-01 -6.37965858e-01 8.53128731e-01
1.47720754e-01 1.40154719e-01 -2.41194710e-01 5.20558774e-01
5.25739849e-01 5.32510638e-01 1.38669550e-01 -1.36209980e-01
1.53172112e+00 1.11277513e-01 -6.52769208e-01 2.89105084e-02
3.64504725e-01 -9.27191615e-01 3.24764013e-01 2.70713508e-01
-6.60128832e-01 -2.00200647e-01 -6.77296042e-01 4.40904737e-01
-3.84776682e-01 2.81916261e-01 6.46639049e-01 1.38705587e+00
-5.32033622e-01 4.19839233e-01 -6.64095700e-01 -7.03074217e-01
5.20464361e-01 6.91397309e-01 -6.08797371e-01 -2.96377718e-01
-7.36534059e-01 1.11887193e+00 1.97033897e-01 3.66252124e-01
-1.77447170e-01 -2.34853566e-01 -8.40473533e-01 -2.29894266e-01
6.30604103e-02 -1.36137130e-02 5.23804247e-01 -3.63490075e-01
-9.67487037e-01 8.81741643e-01 -4.75833654e-01 1.65076796e-02
2.39615709e-01 3.40447843e-01 -6.11039340e-01 5.51492572e-01
4.50446635e-01 1.06000990e-01 6.76366985e-01 -8.88332129e-01
-9.59581971e-01 -7.94580162e-01 -4.50822264e-02 -1.34588420e-01
-4.27485824e-01 8.09889019e-01 -1.78905070e-01 -1.66555405e-01
-3.70205566e-02 -6.35474920e-01 2.82815039e-01 -6.25352681e-01
-2.80094147e-01 -1.16895869e-01 1.15124726e+00 -1.20030677e+00
1.29094601e+00 -2.17561126e+00 -9.62482989e-01 5.77312052e-01
-9.47468802e-02 7.13490367e-01 9.56532732e-02 5.90793967e-01
9.15332288e-02 2.29918569e-01 3.58761661e-02 -4.33718897e-02
-3.06631058e-01 -1.08966090e-01 -8.78487900e-02 5.23947179e-01
2.52794981e-01 4.19250309e-01 -6.83693469e-01 -5.38855672e-01
4.81343716e-01 4.24070776e-01 2.63957456e-02 -9.61980671e-02
9.13728416e-01 3.44341129e-01 -7.34299421e-01 1.24893081e+00
8.83020222e-01 4.66342390e-01 6.96473047e-02 1.64513215e-01
-2.75585473e-01 4.80955839e-02 -1.19974518e+00 5.98449051e-01
-1.48276836e-01 5.95786512e-01 2.36789837e-01 -8.71051848e-01
1.20233023e+00 6.37637556e-01 4.61735338e-01 -7.27901816e-01
-7.61683518e-03 2.68631220e-01 -3.17950487e-01 -1.12069690e+00
5.50647736e-01 1.52366525e-02 2.05383956e-01 8.21743548e-01
-4.55824256e-01 3.67914259e-01 2.31732339e-01 -4.26129401e-02
9.39093590e-01 -9.11187604e-02 1.48279309e-01 2.23430455e-01
4.42423344e-01 -6.38383031e-02 3.74594450e-01 7.25133836e-01
-3.10982585e-01 1.50068268e-01 1.49196580e-01 -6.97115660e-01
-8.69618654e-01 -5.79719126e-01 -4.13992077e-01 6.08495653e-01
-1.20890453e-01 1.69625670e-01 -8.02797377e-01 -5.69512546e-01
-1.27310291e-01 5.43094218e-01 -2.33637646e-01 2.25493982e-01
-3.71707439e-01 -8.20939779e-01 7.63894856e-01 1.82568595e-01
6.57748759e-01 -9.89630699e-01 -7.32308030e-01 3.64614308e-01
-1.44994900e-01 -9.15311754e-01 -2.61048615e-01 -4.21441287e-01
-7.10050821e-01 -1.32080936e+00 -6.67081177e-01 -3.33091676e-01
9.64525521e-01 3.87324661e-01 -2.23549418e-02 3.95677298e-01
-4.32499349e-01 4.30573344e-01 -2.85562992e-01 -5.77973485e-01
-2.80324697e-01 -1.53474480e-01 1.39097482e-01 2.35516965e-01
9.16625679e-01 -1.93291008e-01 -4.51610446e-01 1.13558192e-02
-6.42666876e-01 -4.17218387e-01 2.81462371e-01 3.40203851e-01
-5.79257980e-02 2.24299297e-01 5.93527615e-01 -8.11040223e-01
9.28648949e-01 -2.92513698e-01 -5.18976212e-01 4.71091688e-01
-5.47080934e-01 -4.91413236e-01 2.94248372e-01 2.13345904e-02
-1.20832634e+00 2.43873999e-01 -9.28665772e-02 2.25555986e-01
-7.29279101e-01 1.82503119e-01 8.51608664e-02 -2.51985848e-01
2.78159887e-01 2.67807007e-01 -7.91641790e-03 -4.49743122e-01
-5.44311881e-01 1.33609700e+00 5.79539120e-01 -4.90235209e-01
5.37138581e-01 4.85319912e-01 6.46689758e-02 -9.26510692e-01
-1.23835377e-01 -8.86867285e-01 -7.75577366e-01 -5.65035701e-01
8.92029762e-01 -2.27668241e-01 -1.38351429e+00 6.30328178e-01
-1.59214377e+00 6.10793293e-01 5.49688756e-01 8.02865088e-01
5.07337898e-02 6.36055410e-01 -2.50371456e-01 -1.65599966e+00
-3.61953229e-01 -1.00853777e+00 8.21321905e-01 4.83155787e-01
-2.82562822e-01 -8.04840744e-01 -4.91013452e-02 8.24869573e-01
1.72232687e-01 2.84682989e-01 7.75353551e-01 -5.11107087e-01
-3.24200988e-01 -7.62614250e-01 -6.06298208e-01 1.36608154e-01
5.72930515e-01 4.00477111e-01 -1.05254114e+00 3.08737792e-02
-1.52382970e-01 3.21162194e-01 5.79570174e-01 2.45795473e-01
7.10018218e-01 -2.11353883e-01 -4.73555028e-01 3.45970750e-01
1.47647429e+00 1.18063438e+00 7.73471057e-01 5.74396610e-01
4.12743777e-01 9.58284259e-01 5.21122098e-01 3.59957784e-01
5.00388026e-01 5.94664037e-01 1.75518587e-01 -9.94618312e-02
2.97641695e-01 -1.06879540e-01 3.01996142e-01 1.73702791e-01
-9.58995104e-01 1.81711301e-01 -1.18174648e+00 3.90657485e-01
-1.45278621e+00 -1.64136159e+00 -3.10956150e-01 2.33108997e+00
3.00042897e-01 -3.19645941e-01 3.39414358e-01 3.47004622e-01
1.13030767e+00 -4.27900344e-01 7.06982240e-02 -8.14413190e-01
3.90796959e-01 3.24843049e-01 6.38261020e-01 5.42262077e-01
-9.61027920e-01 6.59212768e-01 6.35728121e+00 5.04906833e-01
-1.16692734e+00 -1.11317277e-01 7.62304544e-01 2.42852226e-01
1.04163609e-01 1.80842262e-02 -1.05045915e+00 7.75250971e-01
7.29473948e-01 2.13459209e-01 4.90740180e-01 3.85194629e-01
7.06126034e-01 -6.63972437e-01 -6.67397738e-01 9.26944017e-01
8.31924975e-02 -1.28057992e+00 -2.00876042e-01 4.54544872e-01
1.92480341e-01 -7.23390281e-01 -1.19791903e-01 -1.52364671e-01
-2.36329034e-01 -1.00376010e+00 2.75726438e-01 9.91407931e-01
8.95543039e-01 -1.09229493e+00 1.03103220e+00 4.35470194e-01
-1.11849451e+00 -3.04368496e-01 -1.90430373e-01 -1.99627399e-01
2.13777065e-01 6.61071613e-02 -1.30967009e+00 3.47398460e-01
5.20925343e-01 1.59420818e-01 -3.82637054e-01 1.02493596e+00
-5.09665720e-03 4.86076176e-01 -2.61004061e-01 -1.34642452e-01
1.91307738e-01 -3.32520634e-01 2.73289502e-01 1.26069725e+00
4.18397963e-01 4.46245581e-01 -2.46758237e-01 6.19408786e-01
4.17211175e-01 1.33914560e-01 -9.75697517e-01 -4.01738985e-03
6.80578709e-01 1.07527649e+00 -8.50920856e-01 -8.17314312e-02
-5.30902982e-01 6.83649957e-01 -4.40620184e-02 2.18953431e-01
-3.09829921e-01 -7.60224164e-01 3.38889450e-01 4.52646077e-01
-2.15583399e-01 -4.87902835e-02 -3.79337102e-01 -5.72897971e-01
9.54968631e-02 -7.61141717e-01 2.73583621e-01 -3.66286993e-01
-8.27430248e-01 2.96031237e-01 1.88420549e-01 -8.94605517e-01
-3.67639124e-01 -5.02671838e-01 -8.82407784e-01 1.22683895e+00
-1.14523423e+00 -1.18232799e+00 1.74303308e-01 5.79889596e-01
2.66544700e-01 -5.02540410e-01 9.65316653e-01 4.42119509e-01
-6.59027934e-01 4.92469102e-01 2.26643551e-02 6.09346211e-01
3.51287633e-01 -5.60594201e-01 2.68770717e-02 8.80351007e-01
-2.02078253e-01 8.98860812e-01 2.39133671e-01 -1.05652761e+00
-1.17042017e+00 -6.18348956e-01 1.36127031e+00 -5.31304061e-01
4.36976939e-01 -1.44884616e-01 -4.72615629e-01 3.55908662e-01
-1.42232761e-01 -6.84921145e-01 7.97456801e-01 6.16125911e-02
1.74573109e-01 -2.68435776e-01 -1.45568228e+00 3.06801379e-01
4.57914233e-01 -8.20853353e-01 -5.27370393e-01 2.33474031e-01
-3.84236395e-01 4.54177335e-02 -4.57489282e-01 7.36625027e-03
9.60641325e-01 -8.97394657e-01 6.77422225e-01 -3.28731984e-01
1.46552011e-01 1.52488975e-02 2.25909889e-01 -6.15506291e-01
-1.50076017e-01 -3.97888839e-01 7.62742460e-01 1.59866309e+00
3.75353187e-01 -9.10085380e-01 8.53191674e-01 1.29746926e+00
5.07054269e-01 -4.75809127e-01 -9.65068340e-01 -5.40181935e-01
-6.99744463e-01 -7.20536292e-01 7.48458266e-01 8.61954987e-01
3.34174067e-01 -2.32699588e-01 -5.63787341e-01 4.15907741e-01
7.43242323e-01 -3.58564645e-01 5.39575994e-01 -1.00432765e+00
8.88007283e-02 -4.05133255e-02 -6.18240654e-01 -3.39590088e-02
-8.88053030e-02 -4.56268638e-01 -5.83531678e-01 -1.41940546e+00
5.41133761e-01 -4.80763227e-01 1.23685159e-01 6.61875188e-01
1.46911681e-01 3.78649652e-01 3.01652402e-01 2.60184199e-01
1.51028976e-01 -4.55006093e-01 9.42561388e-01 -1.23078570e-01
-2.28304952e-01 4.55695003e-01 -7.06544757e-01 7.89539874e-01
8.48789454e-01 -6.43747449e-01 -6.57539144e-02 -3.97984713e-01
-2.25080959e-02 3.85721534e-01 4.53180671e-01 -5.77821016e-01
5.46581388e-01 -3.88695389e-01 5.69958806e-01 -6.80941939e-01
4.28988606e-01 -1.08873188e+00 2.49393433e-01 3.23314488e-01
2.06086636e-01 1.05265416e-02 1.72418520e-01 5.89902215e-02
3.73119898e-02 -8.98156822e-01 4.40820545e-01 -9.61106792e-02
-5.01685500e-01 2.37955317e-01 -6.75327241e-01 -9.19253111e-01
1.25243616e+00 -1.03212249e+00 -4.93314028e-01 -3.76687497e-01
-5.05406797e-01 -1.90709457e-01 2.29070529e-01 -1.77991260e-02
8.02466929e-01 -9.83780622e-01 -6.27037466e-01 3.35012287e-01
-2.18553230e-01 -7.94723928e-01 3.79439801e-01 6.35528862e-01
-7.91522324e-01 6.83447063e-01 -6.09188795e-01 -1.33589074e-01
-1.74864185e+00 2.10359782e-01 -4.34161127e-02 -1.96704697e-02
-2.06489697e-01 3.59859288e-01 -3.73032659e-01 -1.74115032e-01
-1.57827288e-02 3.49561006e-01 -7.63928056e-01 1.01006798e-01
1.04562175e+00 7.64502823e-01 2.90692840e-02 -1.01166487e+00
-6.14924431e-01 6.80532992e-01 -2.70111352e-01 -7.02410331e-03
1.50631928e+00 -2.04191372e-01 -3.07964027e-01 -3.01915228e-01
8.22508991e-01 3.27966660e-01 -7.52343833e-01 3.06627154e-01
-4.41009514e-02 -9.97604609e-01 -1.95717722e-01 -6.59108043e-01
-8.43536019e-01 8.79414439e-01 6.24521971e-01 3.37304980e-01
1.16013658e+00 -3.46438259e-01 5.02494037e-01 3.31762373e-01
5.45328677e-01 -1.21356034e+00 -6.73639596e-01 2.84174621e-01
4.39626455e-01 -1.29293060e+00 7.26013258e-02 -2.25981772e-01
-4.67716724e-01 1.58893895e+00 1.47345871e-01 2.43958741e-01
3.10063601e-01 3.22605789e-01 -3.96851189e-02 -2.19477817e-01
-2.09129937e-02 -6.33925712e-03 1.00691132e-01 1.08021510e+00
3.26332033e-01 1.75169230e-01 -5.35160601e-01 5.18767655e-01
9.21080932e-02 7.15258196e-02 4.74088699e-01 8.08513284e-01
-5.09808242e-01 -1.37130010e+00 -1.06652486e+00 6.67981267e-01
-9.78085577e-01 -4.25544046e-02 -6.22939944e-01 6.51875377e-01
4.38816905e-01 1.35090506e+00 -1.34367600e-01 -3.61240566e-01
3.66545320e-01 -6.38470575e-02 3.06967437e-01 -5.63140631e-01
-6.02545142e-01 -2.45142981e-01 2.49604538e-01 -7.27428570e-02
-4.64130670e-01 -8.55470657e-01 -9.75530922e-01 -9.04211879e-01
-3.45881164e-01 1.96506232e-02 1.06134117e+00 1.17803991e+00
-2.67306659e-02 -3.80133331e-01 7.53069401e-01 -6.00437760e-01
-7.64370933e-02 -6.03331327e-01 -7.91414678e-01 -6.91177743e-03
2.19199192e-02 -3.64013284e-01 1.50477007e-01 -1.72152591e-03] | [13.033305168151855, 0.976281464099884] |
e0b33119-331d-41ac-a66e-b93e55e9c5e9 | residual-generation-using-physically-based | 2008.04644 | null | https://arxiv.org/abs/2008.04644v1 | https://arxiv.org/pdf/2008.04644v1.pdf | Residual Generation Using Physically-Based Grey-Box Recurrent Neural Networks For Engine Fault Diagnosis | Data-driven fault diagnosis is complicated by unknown fault classes and limited training data from different fault realizations. In these situations, conventional multi-class classification approaches are not suitable for fault diagnosis. One solution is the use of anomaly classifiers that are trained using only nominal data. Anomaly classifiers can be used to detect when a fault occurs but give little information about its root cause. Hybrid fault diagnosis methods combining physically-based models and available training data have shown promising results to improve fault classification performance and identify unknown fault classes. Residual generation using grey-box recurrent neural networks can be used for anomaly classification where physical insights about the monitored system are incorporated into the design of the machine learning algorithm. In this work, an automated residual design is developed using a bipartite graph representation of the system model to design grey-box recurrent neural networks and evaluated using a real industrial case study. Data from an internal combustion engine test bench is used to illustrate the potentials of combining machine learning and model-based fault diagnosis techniques. | ['Daniel Jung'] | 2020-08-11 | null | null | null | null | ['anomaly-classification'] | ['computer-vision'] | [ 2.34974056e-01 9.87614617e-02 2.50693798e-01 -9.56699103e-02
-1.44351244e-01 -5.49785793e-03 2.26058826e-01 1.91486642e-01
6.47576690e-01 4.96002734e-01 -4.36178833e-01 -6.74978614e-01
-7.28367686e-01 -7.46117175e-01 -2.81238407e-01 -8.51033390e-01
-1.39592513e-01 7.48682201e-01 1.59167469e-01 -4.90932554e-01
2.17783064e-01 8.66318464e-01 -1.72986948e+00 3.56487900e-01
7.60907888e-01 9.77944911e-01 -6.96394444e-02 6.88259184e-01
4.66093980e-02 1.06756938e+00 -9.89008248e-01 5.52055836e-01
1.66534379e-01 -9.70849693e-01 -6.58644795e-01 3.42168093e-01
-3.14441442e-01 1.62110142e-02 -3.89241368e-01 7.64685988e-01
3.31195831e-01 2.72757560e-01 7.45704293e-01 -1.33741808e+00
-1.34937331e-01 2.95642436e-01 1.94862057e-02 1.95201859e-01
2.37033635e-01 4.04736353e-03 6.00563228e-01 -8.11751842e-01
1.50969341e-01 1.00298762e+00 7.90488243e-01 3.35696220e-01
-1.08155262e+00 -9.96671095e-02 -2.21347198e-01 5.90693772e-01
-1.04542470e+00 -2.42291421e-01 1.37497437e+00 -5.96535683e-01
1.32173347e+00 3.28360677e-01 7.72306323e-01 5.80989182e-01
7.91174889e-01 2.66023636e-01 1.06632805e+00 -6.12284780e-01
6.37931287e-01 -3.59729499e-01 3.62986177e-01 7.30022371e-01
6.07428491e-01 5.16461253e-01 -1.55791566e-01 -3.33856761e-01
6.26714706e-01 3.03973794e-01 -3.89664173e-01 -3.99569243e-01
-5.48556328e-01 7.38708496e-01 4.37677443e-01 6.25031173e-01
-6.16353214e-01 1.53717212e-02 5.80838084e-01 9.38239813e-01
5.55613995e-01 8.08631897e-01 -4.63954955e-01 9.64601487e-02
-8.85277629e-01 7.82672316e-02 9.06166613e-01 1.20460108e-01
6.57108486e-01 1.00915718e+00 3.07654679e-01 7.46004820e-01
5.36496282e-01 4.93744373e-01 7.35744298e-01 -7.12223232e-01
-2.76169986e-01 1.07768285e+00 -2.15389788e-01 -8.53871584e-01
-6.69193923e-01 -4.94022340e-01 -7.94422209e-01 1.00883794e+00
6.29872680e-02 -2.96246082e-01 -1.31372166e+00 7.60441661e-01
5.66903427e-02 1.33081838e-01 4.58427131e-01 8.40254188e-01
1.69802040e-01 5.36074340e-01 -7.17475355e-01 -1.79219618e-01
8.08144391e-01 -4.96924132e-01 -7.09256113e-01 -2.57191271e-01
9.94893730e-01 -3.23211908e-01 4.23116982e-01 6.83197021e-01
-6.63408995e-01 -4.26572561e-01 -1.36524296e+00 9.19100285e-01
-5.11855125e-01 1.67972609e-01 1.39979288e-01 6.56052947e-01
-6.25430882e-01 1.06561852e+00 -1.11231434e+00 -1.40710548e-01
-1.04076177e-01 4.71864194e-01 -8.69445503e-02 -2.95035124e-01
-1.35916471e+00 1.17572105e+00 4.73569930e-01 6.93530858e-01
-8.45199227e-01 -2.56811500e-01 -7.45577097e-01 -8.44259635e-02
4.20772195e-01 -4.92918313e-01 1.05743313e+00 -9.69962835e-01
-1.52420616e+00 -1.16868407e-01 2.18397722e-01 -6.72139704e-01
1.74927682e-01 -2.07122546e-02 -8.22466791e-01 3.01024050e-01
-2.46066317e-01 -5.72537482e-01 1.16087139e+00 -1.14819610e+00
-2.19387501e-01 -1.77958727e-01 -3.97697926e-01 -3.03487122e-01
2.27420956e-01 -3.98120522e-01 7.17648447e-01 -4.65608269e-01
5.66997230e-01 -8.87721956e-01 -2.68870682e-01 -5.69981217e-01
-3.39446217e-01 5.84607311e-02 1.60809326e+00 -6.72976196e-01
9.83233929e-01 -1.85326254e+00 -2.55453512e-02 8.11839044e-01
-1.11660093e-01 3.90466541e-01 3.60541523e-01 7.94349730e-01
-5.69610417e-01 -3.48924071e-01 -6.14727974e-01 4.70821112e-01
-3.32419723e-01 6.56041563e-01 -7.24394619e-02 4.67989266e-01
5.94326377e-01 5.11447310e-01 -5.81999660e-01 2.39479929e-01
6.97216451e-01 2.43631572e-01 -1.73892558e-01 2.65901476e-01
-1.41704261e-01 3.34583402e-01 -4.02188241e-01 7.29148030e-01
-2.31620087e-03 -9.23108011e-02 -5.19865705e-03 -1.23590534e-03
3.57363582e-01 -3.35633159e-02 -1.26645637e+00 1.06403887e+00
-6.25657022e-01 5.00114262e-01 -3.63145694e-02 -1.69679928e+00
1.32959497e+00 7.99773097e-01 5.57032585e-01 -4.98591870e-01
8.91504213e-02 6.20455027e-01 5.50667584e-01 -7.27327645e-01
-1.97157562e-01 -3.03021431e-01 1.54162571e-01 2.87804455e-01
1.52279630e-01 -2.91077286e-01 5.88295683e-02 -3.67573738e-01
1.56485057e+00 -1.33669255e-02 6.20463453e-02 -3.83416176e-01
7.27881074e-01 3.53752613e-01 5.60367227e-01 4.48262721e-01
3.50092143e-01 5.18120110e-01 5.66558063e-01 -5.90702236e-01
-8.93710971e-01 -4.71403420e-01 -7.90661108e-03 1.29719973e-01
-1.90770730e-01 -5.28647155e-02 -5.34454167e-01 -7.87638605e-01
1.78842947e-01 7.83210754e-01 -4.19054121e-01 -9.88772988e-01
-2.79632092e-01 -8.41710925e-01 4.23243135e-01 3.91968995e-01
2.26379991e-01 -1.14720011e+00 -8.49727273e-01 5.15986800e-01
4.99881625e-01 -2.64963180e-01 6.89713776e-01 7.93549061e-01
-1.17039919e+00 -1.56239367e+00 -4.30621296e-01 -4.63318497e-01
9.70315933e-01 -1.38038129e-01 8.53796363e-01 4.53883857e-01
-6.90416753e-01 4.38225478e-01 -4.69738841e-01 -3.55201721e-01
-1.02677262e+00 -4.07228708e-01 1.80472061e-01 -1.08333491e-01
-3.60639133e-02 -4.71581429e-01 -1.70631602e-01 4.96294588e-01
-8.50820959e-01 -6.23944044e-01 7.84290969e-01 1.48292398e+00
1.19284637e-01 9.16606843e-01 1.01125979e+00 -7.96630085e-01
7.25527346e-01 -6.30081773e-01 -5.33577740e-01 2.09446415e-01
-1.26225221e+00 2.97877163e-01 6.50291204e-01 -2.17284113e-01
-1.04410982e+00 6.01712801e-02 -1.74043283e-01 -7.79399693e-01
-4.42655534e-01 1.15838504e+00 -3.88940498e-02 -1.90315917e-01
8.14346910e-01 -1.87312275e-01 5.36716163e-01 -4.61387098e-01
-2.65085846e-01 6.48279607e-01 3.69477868e-01 -3.43860805e-01
4.68418539e-01 -1.70008838e-01 5.14472365e-01 -8.41648161e-01
-3.14889282e-01 -4.97436821e-01 -6.38301969e-01 -6.72892392e-01
1.39035538e-01 -5.32254934e-01 -3.85718137e-01 8.18066299e-01
-6.42902493e-01 -3.26142639e-01 -7.09041119e-01 5.60016990e-01
-4.79251474e-01 9.32722390e-02 -6.15905762e-01 -1.10484731e+00
-3.02461743e-01 -1.07070172e+00 6.79909706e-01 -1.81362897e-01
-2.06656292e-01 -1.30313551e+00 1.30274579e-01 -8.36735666e-02
5.28567314e-01 3.96897852e-01 1.08479166e+00 -1.15436268e+00
4.33526672e-02 -8.54723513e-01 4.28436100e-01 7.39189863e-01
5.67107320e-01 1.69030741e-01 -7.95952618e-01 -3.83098751e-01
5.63293755e-01 1.35187972e-02 6.69323385e-01 2.57743835e-01
4.66984212e-01 -1.32520810e-01 -3.82145315e-01 -1.93846866e-01
1.48870742e+00 5.78273654e-01 4.86498713e-01 2.30803996e-01
7.89515972e-01 6.75628603e-01 5.89347124e-01 2.04960942e-01
-5.40156662e-01 3.50210100e-01 4.91462946e-01 6.60806745e-02
1.21810997e-03 3.43166649e-01 5.63291132e-01 7.57236540e-01
6.78695366e-02 -1.35646865e-01 -1.36711478e+00 3.08584034e-01
-1.89881790e+00 -1.02901030e+00 -5.49693108e-01 1.94703054e+00
3.42870853e-03 5.55000544e-01 -1.39017746e-01 1.25252867e+00
8.08354080e-01 -4.25006390e-01 -6.43410563e-01 -5.45325518e-01
-1.23872841e-03 1.23998366e-01 2.57341951e-01 2.22141251e-01
-7.82010078e-01 2.92048696e-02 6.14842653e+00 3.47051471e-01
-1.20615005e+00 -2.36908004e-01 3.03817898e-01 2.68468916e-01
1.80231899e-01 2.03041032e-01 9.30177420e-03 2.57016450e-01
1.61587489e+00 8.68117660e-02 6.27732053e-02 8.05377543e-01
3.50353092e-01 -2.55602688e-01 -8.97410393e-01 5.57495177e-01
1.31953359e-01 -9.95251656e-01 -2.24847123e-01 4.90287729e-02
5.82590759e-01 -8.27434808e-02 -4.36694741e-01 1.78851597e-02
6.46003261e-02 -6.51067197e-01 3.08465242e-01 1.00330901e+00
1.10550590e-01 -8.28672171e-01 1.28898740e+00 1.02521695e-01
-7.76725411e-01 -6.09812975e-01 -6.87700361e-02 -2.95636773e-01
1.46175817e-01 1.07245493e+00 -9.83041346e-01 9.96534109e-01
4.61235791e-01 8.47095907e-01 -6.09408319e-01 1.26534498e+00
-1.09122572e-02 9.27800238e-01 -3.68779510e-01 3.03842574e-01
1.07018901e-02 -2.22770184e-01 7.17203677e-01 4.57112044e-01
6.38061404e-01 -2.90606797e-01 1.45205900e-01 8.12005043e-01
8.24553609e-01 -3.48769993e-01 -1.04030883e+00 -1.54366881e-01
-1.06607424e-02 1.02062213e+00 -1.00699627e+00 -1.03684515e-01
-3.11474144e-01 7.78378129e-01 -4.25719291e-01 3.11127871e-01
-2.60462552e-01 -6.39517426e-01 4.59821373e-01 2.51059592e-01
9.45674926e-02 -3.79055105e-02 -3.61525528e-02 -6.59103990e-01
-1.40107378e-01 -8.05012882e-01 2.96266049e-01 -1.03406918e+00
-1.34913254e+00 6.06976271e-01 1.00545496e-01 -1.38811052e+00
-1.09087586e+00 -9.98010337e-01 -1.02458489e+00 8.46182942e-01
-8.00093591e-01 -8.46271396e-01 -6.71691075e-02 3.65572691e-01
3.52778643e-01 -6.02164090e-01 8.74485493e-01 5.43816984e-02
-8.75452816e-01 -2.05690473e-01 5.80259025e-01 2.78534251e-03
2.33813509e-01 -1.41109383e+00 1.86112806e-01 1.07017958e+00
5.98632358e-02 -7.31240511e-02 9.83919680e-01 -1.14273906e+00
-1.44579983e+00 -9.65699315e-01 1.87688753e-01 4.98375669e-02
7.69447565e-01 2.29383215e-01 -1.47141707e+00 2.99568146e-01
-1.44324943e-01 3.93887103e-01 2.59727061e-01 -2.06792697e-01
3.68136704e-01 3.67706781e-03 -1.21100569e+00 8.89836773e-02
2.63375610e-01 -5.47093809e-01 -8.12004447e-01 1.91318974e-01
7.88515061e-02 -2.13050112e-01 -1.04117548e+00 5.85569680e-01
7.12675899e-02 -8.39108348e-01 5.66519439e-01 -3.96469891e-01
-2.57955007e-02 -5.08679390e-01 2.44713366e-01 -1.77107430e+00
-1.27069697e-01 -3.47884506e-01 -3.94361138e-01 9.62943673e-01
4.36236531e-01 -1.05661285e+00 7.54726827e-01 3.92803818e-01
-5.62871397e-01 -8.65943611e-01 -9.40884531e-01 -8.76819193e-01
-2.87384719e-01 -3.17011118e-01 3.52074713e-01 7.01721489e-01
1.93390399e-01 2.88446993e-01 1.64062660e-02 2.77129710e-01
1.46793216e-01 2.91719865e-02 3.95746231e-02 -1.77012432e+00
-2.27769062e-01 -2.63237357e-01 -1.03453231e+00 3.05262834e-01
4.02770519e-01 -6.05769515e-01 2.89493382e-01 -1.48958647e+00
-7.98663914e-01 -4.56063122e-01 -4.32328075e-01 7.10595846e-01
1.21893466e-01 -1.27196806e-02 -4.71538365e-01 2.31722146e-01
3.01821142e-01 4.49659258e-01 5.22292912e-01 -2.30201706e-01
-9.74317044e-02 3.47600490e-01 2.23475292e-01 6.29552901e-01
1.06557238e+00 -5.08298755e-01 -7.37503827e-01 3.48353773e-01
2.31880739e-01 4.53604817e-01 5.89401007e-01 -1.66001213e+00
1.63534939e-01 2.52601117e-01 4.09328312e-01 -4.78403568e-01
5.50511926e-02 -1.36844456e+00 6.52832031e-01 9.55283284e-01
6.03430569e-02 2.63524294e-01 1.89971611e-01 7.94959664e-01
-6.18918538e-01 -6.64943755e-01 3.28303993e-01 5.38243614e-02
-7.15018392e-01 -3.07681561e-01 -8.70825589e-01 -6.13244236e-01
1.04213119e+00 -3.32431018e-01 -1.02992661e-01 -1.64706260e-01
-1.11447477e+00 -1.60956364e-02 3.70912045e-01 2.68323541e-01
9.32414532e-01 -1.21099579e+00 -4.56112266e-01 6.49575591e-01
1.24316849e-01 -1.31880596e-01 7.43272677e-02 7.17192233e-01
-6.59132540e-01 2.92917818e-01 -3.85636836e-01 -6.58432186e-01
-1.13338542e+00 5.79028487e-01 1.03821266e+00 3.36567760e-02
-7.37556219e-01 1.68258905e-01 -7.78982580e-01 -3.21654260e-01
-5.04242897e-01 -3.00728232e-01 -3.09683140e-02 2.16978253e-03
9.85702500e-03 6.35213375e-01 9.30200219e-01 -6.07659876e-01
-3.20253670e-01 2.87283003e-01 4.50450540e-01 9.74955224e-03
1.26773059e+00 2.10654840e-01 -2.06408456e-01 1.09475756e+00
8.92650247e-01 -6.79416060e-01 -9.30008709e-01 1.44444093e-01
4.49858278e-01 4.73929606e-02 4.50006664e-01 -9.22457457e-01
-1.45152318e+00 7.64959753e-01 8.67656350e-01 9.09350395e-01
1.30207610e+00 -2.88040340e-01 1.26276582e-01 7.28998899e-01
2.08509102e-01 -1.02875876e+00 -1.48689643e-01 2.96329260e-01
9.35980439e-01 -9.24107730e-01 -5.22529520e-02 -1.54882580e-01
-3.14899325e-01 1.75914729e+00 4.13196981e-01 -5.04740477e-01
9.85526502e-01 3.56904924e-01 1.18221380e-01 -6.95149422e-01
-7.83020020e-01 -4.40027453e-02 4.57852095e-01 4.33765531e-01
2.56702393e-01 -2.91081548e-01 7.93391615e-02 8.23230147e-02
2.01601699e-01 -1.32038400e-01 5.96254826e-01 1.52941835e+00
-4.57317919e-01 -1.22126126e+00 -7.94241846e-01 1.14123893e+00
-3.74498546e-01 4.57264602e-01 -4.31172609e-01 7.30690062e-01
-2.16674224e-01 1.25234365e+00 1.37365237e-01 -4.63055611e-01
4.65806067e-01 6.47863925e-01 1.67940095e-01 -5.97734571e-01
-4.67555106e-01 -1.65767491e-01 1.91852421e-01 -3.44804525e-01
-3.70989293e-01 -6.31939530e-01 -1.41367161e+00 8.67554471e-02
-8.92006636e-01 7.08017588e-01 6.50631666e-01 1.20709717e+00
3.89335334e-01 1.35823524e+00 7.18677104e-01 -9.08933640e-01
-4.44703698e-01 -1.06043363e+00 -9.34267521e-01 7.63410777e-02
5.21254957e-01 -1.06689572e+00 -6.76973104e-01 -1.26197770e-01] | [6.681512355804443, 2.4642674922943115] |
97b3b318-cf50-4290-9a7b-0fe2792b7933 | sequential-quantiles-via-hermite-series | 1507.05073 | null | http://arxiv.org/abs/1507.05073v2 | http://arxiv.org/pdf/1507.05073v2.pdf | Sequential Quantiles via Hermite Series Density Estimation | Sequential quantile estimation refers to incorporating observations into
quantile estimates in an incremental fashion thus furnishing an online estimate
of one or more quantiles at any given point in time. Sequential quantile
estimation is also known as online quantile estimation. This area is relevant
to the analysis of data streams and to the one-pass analysis of massive data
sets. Applications include network traffic and latency analysis, real time
fraud detection and high frequency trading. We introduce new techniques for
online quantile estimation based on Hermite series estimators in the settings
of static quantile estimation and dynamic quantile estimation. In the static
quantile estimation setting we apply the existing Gauss-Hermite expansion in a
novel manner. In particular, we exploit the fact that Gauss-Hermite
coefficients can be updated in a sequential manner. To treat dynamic quantile
estimation we introduce a novel expansion with an exponentially weighted
estimator for the Gauss-Hermite coefficients which we term the Exponentially
Weighted Gauss-Hermite (EWGH) expansion. These algorithms go beyond existing
sequential quantile estimation algorithms in that they allow arbitrary
quantiles (as opposed to pre-specified quantiles) to be estimated at any point
in time. In doing so we provide a solution to online distribution function and
online quantile function estimation on data streams. In particular we derive an
analytical expression for the CDF and prove consistency results for the CDF
under certain conditions. In addition we analyse the associated quantile
estimator. Simulation studies and tests on real data reveal the Gauss-Hermite
based algorithms to be competitive with a leading existing algorithm. | ['Melvin Varughese', 'Iain Macdonald', 'Michael Stephanou'] | 2015-07-17 | null | null | null | null | ['sequential-distribution-function-estimation', 'sequential-quantile-estimation', 'data-summarization'] | ['miscellaneous', 'miscellaneous', 'miscellaneous'] | [-2.48842046e-01 -2.86916941e-01 3.95361967e-02 -4.57746893e-01
-9.80018020e-01 -8.02656233e-01 5.37227243e-02 9.24374163e-01
-3.38171303e-01 7.24327743e-01 -1.55579820e-01 -3.84805113e-01
-5.67924500e-01 -1.02086866e+00 -6.60203338e-01 -4.96352077e-01
-5.48258185e-01 6.24635100e-01 3.09763730e-01 -1.42567560e-01
3.31205100e-01 7.50459611e-01 -1.12105811e+00 -2.16972858e-01
7.33525753e-01 1.45130563e+00 -6.26040459e-01 9.78967786e-01
-7.60839060e-02 8.57302547e-02 -6.21530771e-01 -4.98890340e-01
3.43662739e-01 -3.05526406e-01 -2.71790177e-01 -2.17786968e-01
-1.17257861e-02 -7.40378261e-01 4.19203758e-01 6.78538680e-01
4.37697947e-01 2.18250930e-01 6.76896811e-01 -1.59217930e+00
-6.82480773e-03 5.34274936e-01 -1.04899979e+00 3.97753119e-01
4.32030261e-01 -7.30557382e-01 8.53595674e-01 -5.92000484e-01
8.07241350e-02 9.51959133e-01 9.37576056e-01 -1.83741316e-01
-1.36290193e+00 -5.74411035e-01 -1.31990954e-01 -1.49068311e-01
-1.13664234e+00 -1.36166409e-01 5.23060679e-01 -2.98748374e-01
4.22742873e-01 1.32207781e-01 6.00956202e-01 2.99881250e-01
3.47244799e-01 4.83144492e-01 8.15180421e-01 -2.73375332e-01
5.08689761e-01 -5.61623201e-02 2.43011221e-01 1.69288799e-01
3.04167181e-01 1.18263438e-01 -3.87201726e-01 -8.25464249e-01
7.59181201e-01 1.38419658e-01 1.35195509e-01 -2.20174074e-01
-7.73888707e-01 1.02496684e+00 -8.06608573e-02 -1.83487654e-01
-6.08219624e-01 2.79168367e-01 8.45882535e-01 5.54847717e-01
9.87367272e-01 -2.75339663e-01 -4.10152793e-01 -4.20575500e-01
-1.09480286e+00 5.63383758e-01 1.26111209e+00 9.44829524e-01
5.77605426e-01 4.98774601e-03 -2.43035823e-01 5.43247998e-01
3.36460993e-02 5.29136300e-01 4.05717380e-02 -8.58700037e-01
5.44888496e-01 -9.42195125e-04 3.98911089e-01 -5.68139553e-01
-3.44543815e-01 -4.52323645e-01 -6.88429534e-01 -1.10285230e-01
8.57055962e-01 -4.25833136e-01 -1.20003968e-01 1.60977650e+00
7.25091457e-01 3.11443567e-01 -2.38327235e-01 3.75967830e-01
-3.37003440e-01 7.01913178e-01 -5.90635054e-02 -5.04980564e-01
1.10715604e+00 7.73540214e-02 -4.58377391e-01 7.52243936e-01
2.25868613e-01 -7.45542705e-01 4.55531746e-01 6.85142577e-01
-1.05240369e+00 -2.08455577e-01 -6.54363632e-01 4.22446311e-01
-2.31109262e-01 -2.96455622e-01 4.24558163e-01 1.10266507e+00
-6.80866957e-01 8.09744060e-01 -6.73818529e-01 2.64790654e-02
1.26750469e-01 2.57503271e-01 4.17334437e-02 5.15010059e-01
-1.04530764e+00 1.33332074e-01 1.33147955e-01 3.89702953e-02
-2.47955218e-01 -1.08948147e+00 -5.72331131e-01 3.06114256e-01
2.18864396e-01 -4.82935816e-01 1.49672139e+00 -7.09172904e-01
-1.41761923e+00 1.75315499e-01 -2.84361452e-01 -6.76997662e-01
9.10254002e-01 -1.79189712e-01 -1.67558938e-01 2.52638429e-01
1.63633779e-01 -4.16490555e-01 8.13920021e-01 -5.75525224e-01
-8.12979102e-01 -4.94608760e-01 -1.76116586e-01 -2.70447791e-01
-1.60717651e-01 -3.47475745e-02 4.10964698e-01 -5.28046906e-01
1.28945246e-01 -4.29128021e-01 -1.93090588e-01 -1.56433046e-01
-1.20862052e-01 -3.68524730e-01 5.45557261e-01 -4.28433061e-01
1.29028678e+00 -1.88042176e+00 -5.97305655e-01 7.66856909e-01
-3.33118550e-02 -4.14181888e-01 5.12587309e-01 1.13391292e+00
-3.23784072e-03 -1.35847807e-01 -9.77097601e-02 -2.42425904e-01
3.04272771e-01 -1.60660341e-01 -7.62050629e-01 9.34748292e-01
1.31577542e-02 4.06830251e-01 -9.23346162e-01 -3.54016662e-01
9.82898846e-02 1.66865140e-01 -6.81438923e-01 2.57238775e-01
6.02188483e-02 2.93869954e-02 -4.30876881e-01 5.85745871e-01
9.22324598e-01 1.18266560e-01 -3.67485695e-02 5.14215603e-02
-4.81134892e-01 -2.94272214e-01 -1.50381744e+00 6.79093301e-01
-7.05193639e-01 3.90618533e-01 -4.07472514e-02 -1.34086847e+00
1.04279315e+00 5.31888127e-01 7.82119095e-01 -3.58483106e-01
7.94695616e-02 6.94629610e-01 -7.23727047e-01 -1.40733495e-01
6.28962398e-01 -6.38405502e-01 -4.89093482e-01 4.86010760e-01
-2.45458111e-01 2.19404981e-01 5.55038273e-01 2.06088856e-01
7.00659513e-01 -5.69974333e-02 5.51401615e-01 -3.73371691e-01
3.89449626e-01 -3.03512096e-01 2.94211239e-01 8.60685825e-01
-1.82463706e-01 2.51832724e-01 1.17115283e+00 -5.53658344e-02
-1.15138507e+00 -1.43491244e+00 -4.70378608e-01 1.08988309e+00
-2.11718574e-01 -5.42406589e-02 -5.19975483e-01 -1.36940390e-01
4.94965017e-01 6.00921392e-01 -6.11930430e-01 1.10743314e-01
-2.88803339e-01 -5.28603375e-01 4.74146038e-01 6.48619294e-01
2.99390614e-01 -3.46017331e-01 -8.29340518e-01 6.30948484e-01
2.45875403e-01 -8.62429082e-01 -6.00142121e-01 2.93816060e-01
-1.00457716e+00 -9.43345726e-01 -8.70659769e-01 -9.14262757e-02
1.99652731e-01 -1.38599202e-01 8.88130188e-01 -5.25980294e-01
1.16253778e-01 7.00394154e-01 -3.40479165e-01 -7.06343234e-01
-1.14048250e-01 -1.46966785e-01 -8.54292810e-02 3.98726851e-01
1.33048654e-01 -8.28306973e-01 -6.98256373e-01 3.05855244e-01
-1.06585658e+00 -7.27234066e-01 7.80954352e-03 8.35698009e-01
5.59963584e-01 -1.09252244e-01 1.37949824e+00 -1.21119344e+00
9.93987322e-01 -1.01558769e+00 -1.28827202e+00 -4.66377148e-03
-6.55031860e-01 -1.11251742e-01 1.11614990e+00 -4.11440700e-01
-9.25768793e-01 -9.98017490e-02 -1.67457923e-01 -3.70933209e-03
2.07624108e-01 6.72620475e-01 1.62070155e-01 1.28063887e-01
2.93638915e-01 6.86635152e-02 -2.27427304e-01 -4.89970237e-01
2.09318206e-01 7.18711495e-01 6.13631248e-01 -9.95798707e-01
5.09232342e-01 6.47527397e-01 5.23354590e-01 -7.05642879e-01
-6.57271862e-01 -1.00543880e+00 -3.50816786e-01 -3.72483999e-01
3.58915552e-02 -4.49569255e-01 -1.52655160e+00 3.77659798e-01
-9.90381837e-01 4.04919451e-03 -4.81425822e-01 4.18741465e-01
-9.59008753e-01 3.95974845e-01 -6.67427599e-01 -1.58557701e+00
-2.59048045e-01 -3.29897165e-01 1.14837301e+00 4.38850731e-01
-8.97340029e-02 -1.28433609e+00 2.68842041e-01 -4.01725382e-01
4.91243094e-01 8.67944002e-01 6.88879907e-01 -7.84022391e-01
-1.69936866e-02 -8.28174174e-01 -2.94299841e-01 1.42861918e-01
-2.43339464e-01 1.45844549e-01 -4.86921370e-01 -2.59504467e-01
-6.38921335e-02 1.77562371e-01 3.57386023e-01 6.33184552e-01
1.09965515e+00 -2.69195080e-01 -1.28911719e-01 5.63711524e-01
1.60911989e+00 -7.12695271e-02 2.92214930e-01 2.75053918e-01
-2.04467639e-01 4.66189831e-01 7.62889445e-01 1.36581779e+00
4.53066230e-02 3.89126301e-01 3.31775993e-01 4.49638039e-01
9.40452576e-01 -2.57111460e-01 2.61715293e-01 2.03011900e-01
5.98853901e-02 -2.26016402e-01 -5.51908076e-01 7.91716635e-01
-1.75315297e+00 -1.00665557e+00 -5.54875314e-01 2.85820198e+00
6.86955452e-01 1.79299459e-01 8.28438461e-01 4.39400882e-01
8.42764676e-01 -3.80020350e-01 -5.60705483e-01 -8.92037094e-01
2.36330450e-01 7.18310654e-01 8.68211627e-01 3.42950612e-01
-7.95237541e-01 -4.02235016e-02 5.66949749e+00 7.71182120e-01
-7.55777001e-01 9.75657031e-02 4.87146169e-01 4.28193882e-02
-1.74253196e-01 2.64194697e-01 -8.22601616e-01 7.14822292e-01
1.45432198e+00 -8.67645264e-01 9.34558734e-03 9.59671080e-01
4.99026597e-01 -5.21385849e-01 -9.39984679e-01 6.43745422e-01
-7.90627182e-01 -1.05421102e+00 -4.82147306e-01 3.33394945e-01
5.25177896e-01 -4.91651654e-01 -8.05366933e-02 1.98542535e-01
-9.74444374e-02 -5.05157411e-01 6.08907461e-01 8.34215760e-01
6.76355720e-01 -1.47524428e+00 8.37072968e-01 3.24609905e-01
-1.40374696e+00 -1.36307329e-01 -3.84131342e-01 -4.26634960e-02
5.81614614e-01 1.36357093e+00 -6.73644125e-01 7.94006586e-01
2.21585512e-01 2.20471278e-01 -5.02165742e-02 1.49921918e+00
4.92972165e-01 8.95793378e-01 -9.29860175e-01 -1.96222171e-01
1.27924591e-01 -3.41687620e-01 4.15122658e-01 1.30217826e+00
8.31836402e-01 -6.40425161e-02 -6.86784238e-02 5.62736690e-01
1.36844382e-01 3.89844239e-01 -1.16838910e-01 3.51788729e-01
5.40820837e-01 9.18527961e-01 -7.67104030e-01 -3.70711714e-01
-3.15297097e-01 6.85966909e-01 -1.05126649e-01 4.36937362e-01
-8.43379617e-01 -9.94727612e-01 2.75196820e-01 4.82551485e-01
3.29402477e-01 -3.00488591e-01 -2.85152227e-01 -6.50867760e-01
2.83472300e-01 -9.87093151e-02 7.36593306e-01 -9.95335430e-02
-1.44666159e+00 1.83271933e-02 6.48850560e-01 -1.15426373e+00
-8.12230945e-01 -3.37831289e-01 -7.86402285e-01 8.84627759e-01
-1.60650921e+00 -5.90440094e-01 -5.67358658e-02 5.40840685e-01
4.62070368e-02 2.37521112e-01 4.93813604e-01 3.39635223e-01
-1.44194260e-01 6.89686060e-01 8.32410216e-01 -1.14618100e-01
5.91418624e-01 -1.40468085e+00 1.20520309e-01 5.30248761e-01
-3.63851637e-01 4.72199827e-01 1.08427060e+00 -6.36361122e-01
-1.09438884e+00 -7.68173516e-01 8.50936830e-01 1.43969104e-01
1.00316405e+00 -1.83875889e-01 -8.67025495e-01 3.74636859e-01
-2.75548011e-01 1.52811870e-01 7.84937441e-01 6.20553549e-03
-4.11543772e-02 -4.98211980e-01 -1.28295267e+00 3.44238157e-04
3.41869652e-01 -2.65352279e-01 4.23129648e-02 2.80802965e-01
1.95962965e-01 -2.14070827e-01 -1.27436662e+00 1.78785861e-01
9.59047258e-01 -1.39204979e+00 6.43116355e-01 -5.87040663e-01
1.01701804e-01 -1.20575145e-01 -1.84900448e-01 -1.00038230e+00
2.74120569e-01 -1.08818948e+00 -3.62412512e-01 1.32711935e+00
5.12766577e-02 -1.30920708e+00 5.73563874e-01 6.08481109e-01
2.97971398e-01 -9.02699113e-01 -1.26314318e+00 -1.14881504e+00
1.20894201e-01 -6.27862751e-01 7.16103971e-01 4.05387461e-01
2.16044977e-01 -3.52616906e-01 -2.46424183e-01 1.50895149e-01
1.11674702e+00 3.88861150e-01 7.33431399e-01 -1.26999438e+00
-5.66386163e-01 -3.71432304e-01 -5.12889564e-01 -1.13485062e+00
-1.40949279e-01 -3.49960446e-01 -1.30226970e-01 -8.99565399e-01
-2.82037482e-02 -3.79307121e-01 -2.47335792e-01 -2.55565435e-01
6.37060776e-02 3.38189816e-03 -2.82980591e-01 -1.80590913e-01
-2.97200680e-01 4.58273381e-01 4.50278908e-01 6.25187278e-01
-2.28219002e-01 8.29298556e-01 -2.73354322e-01 2.82094985e-01
9.14531469e-01 -6.31522238e-01 -3.10867369e-01 4.74165708e-01
5.57280481e-01 7.44767487e-01 4.85871226e-01 -6.59458280e-01
3.80518466e-01 -1.28073752e-01 3.26894045e-01 -9.14437711e-01
-7.71086439e-02 -5.86162567e-01 6.95617199e-02 2.13484883e-01
-1.85845718e-01 2.04772100e-01 7.31710196e-02 9.66232657e-01
-2.20074549e-01 -4.83077139e-01 6.01970792e-01 3.23814631e-01
-1.08486339e-01 2.72798181e-01 -4.84440833e-01 1.78761482e-01
1.05794084e+00 -3.16200167e-01 3.01922446e-05 -8.52495611e-01
-7.81646132e-01 3.79817247e-01 -2.80558765e-02 -3.02665323e-01
5.01176417e-01 -1.21060073e+00 -7.75723755e-01 1.52452171e-01
-1.72427297e-01 -2.54572779e-01 3.26037705e-01 1.16653752e+00
-4.44951534e-01 2.27947637e-01 9.21940804e-02 -4.35659438e-01
-7.71150112e-01 1.50879905e-01 3.30606818e-01 -3.41979116e-01
-1.84538171e-01 3.83857161e-01 -9.47124511e-02 6.80436939e-02
4.74801511e-02 -3.39840651e-01 1.77068591e-01 3.85087907e-01
4.93495882e-01 8.81638348e-01 1.33583039e-01 -1.94925647e-02
-1.51892945e-01 3.58314842e-01 2.32228667e-01 -3.63570452e-01
1.37460291e+00 -3.08679521e-01 -7.14962631e-02 8.61893356e-01
1.42167091e+00 3.66145879e-01 -1.22307634e+00 -5.14407270e-02
3.54033917e-01 -4.06195670e-01 -3.97453576e-01 -1.79805025e-01
-7.29393959e-01 6.77748263e-01 3.01346093e-01 1.01828218e+00
1.31001818e+00 -2.66743690e-01 9.99719858e-01 6.60979329e-03
3.09478194e-01 -1.32461286e+00 -3.68230492e-01 2.84491450e-01
4.54963923e-01 -7.72680819e-01 2.21586432e-02 -2.67952651e-01
-7.15947226e-02 1.86507273e+00 -1.50445223e-01 -5.54393053e-01
1.00656605e+00 3.59346867e-01 -3.91522318e-01 2.86303312e-01
-5.62567592e-01 -1.07200332e-01 -2.37652920e-02 2.16472760e-01
6.10562325e-01 1.77424759e-01 -7.00026691e-01 7.23580360e-01
-2.45928988e-01 4.32554781e-02 8.57690334e-01 7.59294391e-01
-3.79438013e-01 -1.16711807e+00 -4.85751748e-01 7.19334960e-01
-7.91628540e-01 2.14767501e-01 3.42139989e-01 7.43453443e-01
-5.38317859e-01 9.54516470e-01 2.13824600e-01 2.95266211e-01
5.50312579e-01 3.04464102e-01 2.53441036e-01 -1.78907052e-01
-4.20115381e-01 2.85166949e-01 -7.67581910e-02 -2.91273385e-01
4.86625135e-02 -1.07275522e+00 -1.18940902e+00 -6.76663876e-01
-4.02627647e-01 5.73588312e-01 9.07335579e-01 7.95518458e-01
-6.06773384e-02 2.12118760e-01 1.09168029e+00 -6.06379688e-01
-1.14787614e+00 -5.39823890e-01 -1.08915007e+00 2.21985765e-03
5.23750246e-01 -3.11948299e-01 -4.12569582e-01 -2.43808344e-01] | [7.048947811126709, 3.9317073822021484] |
fd3c1399-eafd-4cab-ba92-93706d70b985 | speed-experimental-design-for-policy | 2301.12357 | null | https://arxiv.org/abs/2301.12357v2 | https://arxiv.org/pdf/2301.12357v2.pdf | SPEED: Experimental Design for Policy Evaluation in Linear Heteroscedastic Bandits | In this paper, we study the problem of optimal data collection for policy evaluation in linear bandits. In policy evaluation, we are given a target policy and asked to estimate the expected reward it will obtain when executed in a multi-armed bandit environment. Our work is the first work that focuses on such optimal data collection strategy for policy evaluation involving heteroscedastic reward noise in the linear bandit setting. We first formulate an optimal design for weighted least squares estimates in the heteroscedastic linear bandit setting that reduces the MSE of the value of the target policy. We then use this formulation to derive the optimal allocation of samples per action during data collection. We then introduce a novel algorithm SPEED (Structured Policy Evaluation Experimental Design) that tracks the optimal design and derive its regret with respect to the optimal design. Finally, we empirically validate that SPEED leads to policy evaluation with mean squared error comparable to the oracle strategy and significantly lower than simply running the target policy. | ['Robert Nowak', 'Josiah Hanna', 'Qiaomin Xie', 'Subhojyoti Mukherjee'] | 2023-01-29 | null | null | null | null | ['experimental-design'] | ['methodology'] | [ 4.59485725e-02 7.67709017e-02 -1.06781304e+00 -3.16494077e-01
-1.42525220e+00 -8.78278375e-01 3.14897835e-01 3.86932045e-02
-8.88637722e-01 1.27091956e+00 3.22055966e-01 -1.11851430e+00
-7.00366795e-01 -4.38160807e-01 -1.15559494e+00 -8.14899266e-01
-1.49826556e-02 8.68104160e-01 -3.93350512e-01 6.12064362e-01
2.66111553e-01 4.78356540e-01 -9.14176345e-01 -6.25448748e-02
7.09876657e-01 1.48792553e+00 5.36511987e-02 9.62155223e-01
3.93367589e-01 8.70492339e-01 -5.40793598e-01 -3.37436885e-01
6.36135161e-01 -4.75170910e-01 -6.92088842e-01 1.39515579e-01
2.87070334e-01 -1.03186095e+00 -4.87397835e-02 1.01942873e+00
4.97371346e-01 4.16292548e-01 5.46210587e-01 -8.73724997e-01
-1.69091493e-01 8.08425248e-01 -4.17083174e-01 4.86396104e-01
-2.13931091e-02 8.76573473e-02 1.27526426e+00 3.63722704e-02
1.68364376e-01 1.38453102e+00 1.80153757e-01 4.00587499e-01
-1.65834641e+00 -2.93459237e-01 2.94188648e-01 -2.83571845e-03
-6.15862012e-01 -6.34750664e-01 1.94105953e-01 -2.76518106e-01
5.02263546e-01 4.73740757e-01 8.05853486e-01 1.07272100e+00
-4.21645731e-01 1.18902850e+00 1.42998219e+00 -5.37660956e-01
7.68812418e-01 3.90494555e-01 2.51283079e-01 1.92646995e-01
5.47266722e-01 6.95482016e-01 -6.79135323e-02 -4.31026429e-01
6.79879010e-01 -7.13583604e-02 -1.53991356e-01 -4.26720023e-01
-8.42922568e-01 9.64485943e-01 9.27995741e-02 -2.57895082e-01
-1.00008130e+00 6.70111299e-01 1.51984304e-01 4.22312051e-01
6.29394472e-01 5.02164781e-01 -4.85903800e-01 -5.45800567e-01
-9.08819318e-01 7.18331456e-01 9.22275603e-01 9.31670129e-01
1.95924118e-01 5.44102527e-02 -9.56574857e-01 7.31147707e-01
3.21162865e-02 1.01637816e+00 -3.24361399e-02 -1.42567372e+00
8.66798222e-01 -2.96492010e-01 1.15696454e+00 -2.58088827e-01
-3.41885090e-02 -8.11477602e-01 -6.42112195e-02 7.23988190e-02
6.91858113e-01 -8.05658698e-01 -5.59486270e-01 1.66526675e+00
4.64366704e-01 -8.60098526e-02 1.76885411e-01 8.16581309e-01
-3.71329278e-01 6.59637690e-01 -2.39276037e-01 -7.90090144e-01
9.10956204e-01 -8.95799577e-01 -6.20682001e-01 -2.37726435e-01
4.88195062e-01 -4.69662607e-01 8.72625172e-01 5.11437953e-01
-1.38145089e+00 2.45397702e-01 -5.50886512e-01 5.91266513e-01
3.62068743e-01 -4.22979966e-02 5.27472615e-01 1.03313804e+00
-5.23038805e-01 5.70408404e-01 -8.55392277e-01 8.49281698e-02
6.72549784e-01 2.64274150e-01 3.18215281e-01 6.46245033e-02
-5.06052375e-01 9.56021786e-01 4.98956263e-01 -1.08201489e-01
-1.28418827e+00 -9.28849459e-01 -2.46774197e-01 2.71766543e-01
1.13907373e+00 -5.19199908e-01 2.19548965e+00 -9.98977721e-01
-1.75847960e+00 1.42331213e-01 -3.33544433e-01 -1.04487014e+00
8.12442601e-01 -3.38965893e-01 2.75617927e-01 -1.98679313e-01
-3.55655104e-02 -1.16644368e-01 6.95380747e-01 -9.45229888e-01
-1.03941429e+00 -5.50862134e-01 2.75284439e-01 2.68898338e-01
-2.30958626e-01 5.27012832e-02 4.15414851e-03 -5.12764990e-01
-3.56353551e-01 -1.08773732e+00 -3.37242961e-01 -5.85314631e-01
-4.40047711e-01 5.62495254e-02 2.02343315e-02 -5.06780386e-01
1.29966462e+00 -1.73276782e+00 -1.67447641e-01 5.33169508e-01
-4.40793097e-01 6.45955047e-03 -6.90389285e-03 2.54324526e-01
2.25069910e-01 4.13897857e-02 2.48498783e-01 -1.12482741e-01
4.08156723e-01 2.03178331e-01 -7.17376053e-01 7.39089012e-01
-5.46539068e-01 8.84161890e-01 -7.30063558e-01 -9.74349454e-02
-3.58004458e-02 -5.85053504e-01 -7.89498746e-01 2.88713068e-01
-6.15775526e-01 2.60513127e-01 -8.59748244e-01 3.83461386e-01
4.07151610e-01 -1.12121604e-01 1.50377229e-01 3.48075420e-01
-1.81656018e-01 1.75352857e-01 -1.14404809e+00 9.68714952e-01
-4.83992577e-01 1.65086344e-01 2.36744300e-01 -1.43363357e+00
4.07873034e-01 2.60446697e-01 6.28948569e-01 -5.78604579e-01
2.72317678e-01 2.97359496e-01 -1.78334996e-01 -4.94669318e-01
2.28987843e-01 -3.50002706e-01 1.57500595e-01 7.67286777e-01
-3.42721760e-01 2.09454447e-01 7.70033225e-02 -2.66361088e-01
9.25678492e-01 -1.40522599e-01 3.11381727e-01 -2.53008574e-01
-2.85701156e-02 3.42456414e-03 3.21526647e-01 1.60345566e+00
-1.91127099e-02 -1.79338977e-01 9.75460052e-01 -3.37457597e-01
-1.20686662e+00 -7.57787704e-01 -4.55159731e-02 1.33823550e+00
-3.40057701e-01 4.54341233e-01 -5.50442576e-01 -6.09964848e-01
6.30231023e-01 1.23925006e+00 -7.23879814e-01 2.79185683e-01
-5.89299574e-02 -7.80821085e-01 9.64143574e-02 2.95283765e-01
2.41791084e-01 -6.97348654e-01 -7.79932022e-01 4.18942750e-01
1.36407375e-01 -9.25794244e-01 -7.16878176e-01 1.42868996e-01
-9.01138961e-01 -9.40754652e-01 -8.48929584e-01 1.53050423e-01
4.71672446e-01 1.09029729e-02 7.43926466e-01 -8.04257274e-01
4.36294198e-01 6.79833591e-01 -1.63843796e-01 -9.54661310e-01
-1.54941037e-01 -8.02815706e-02 -1.11950286e-01 2.98780531e-01
1.41221017e-01 -1.07341938e-01 -8.18831027e-01 2.05970764e-01
-6.26869321e-01 -2.99323052e-01 4.84556258e-01 8.61961722e-01
5.74481189e-01 -2.89057016e-01 4.41949397e-01 -9.65163469e-01
1.16308057e+00 -5.00672758e-01 -1.48789895e+00 5.54216623e-01
-7.90884197e-01 4.14627165e-01 3.75247329e-01 -5.75118601e-01
-1.10164380e+00 -9.18195471e-02 4.14378047e-01 -2.87921339e-01
2.14809760e-01 4.87787247e-01 4.02124345e-01 2.11217925e-01
6.93443835e-01 1.36735871e-01 7.15383589e-02 -5.47969580e-01
2.99926460e-01 7.68874407e-01 1.84389159e-01 -1.30391860e+00
2.66347360e-02 2.49505743e-01 1.49854407e-01 -3.07367563e-01
-1.30487227e+00 -3.49946737e-01 3.14493030e-01 -5.36773577e-02
2.86520481e-01 -5.81904173e-01 -1.37277019e+00 -1.29599541e-01
-8.29660296e-01 -8.06088328e-01 -6.91304088e-01 1.00357211e+00
-1.53184235e+00 -1.88735843e-01 -5.98110594e-02 -1.81914020e+00
-2.91789323e-01 -9.27679062e-01 7.10937500e-01 -8.31515640e-02
1.64808020e-01 -8.60619903e-01 2.85599023e-01 4.26197261e-01
3.47293109e-01 1.53901223e-02 6.47960186e-01 -7.29498029e-01
-5.74349999e-01 -2.09580794e-01 -1.88978210e-01 4.19676304e-01
-2.19363242e-01 -4.72267658e-01 -5.14666498e-01 -6.33095920e-01
-3.56177948e-02 -3.70602459e-01 6.64974928e-01 1.29500842e+00
1.41292596e+00 -1.22171998e+00 -5.19446135e-02 5.26332855e-01
1.27847135e+00 6.13566160e-01 1.36922747e-01 6.74854398e-01
-2.05979005e-01 3.17719519e-01 9.22772348e-01 9.87441003e-01
-1.37704641e-01 6.58856273e-01 3.23055208e-01 2.42395788e-01
8.68079126e-01 -3.51147443e-01 3.45138073e-01 -4.34016913e-01
-9.99397188e-02 -1.98375240e-01 -5.46955049e-01 6.14042461e-01
-2.32851934e+00 -1.28626931e+00 3.79070371e-01 2.95277977e+00
9.28840339e-01 2.56093573e-02 8.02017629e-01 -3.79424334e-01
5.45570970e-01 -2.36290842e-01 -1.17388856e+00 -6.59204125e-01
2.87157625e-01 1.42746449e-01 1.41170096e+00 7.85856307e-01
-9.88870740e-01 5.78635275e-01 6.93435144e+00 8.95590365e-01
-6.37848437e-01 2.06061721e-01 9.91578758e-01 -7.18267858e-01
-1.43726751e-01 -4.14082780e-02 -9.06675577e-01 5.49561977e-01
1.39437759e+00 -5.48571646e-01 1.09422243e+00 1.04323483e+00
8.03755820e-01 -3.25347394e-01 -1.15580964e+00 8.00917864e-01
-6.92141354e-01 -1.37299323e+00 -5.22480130e-01 6.30572855e-01
1.03111327e+00 8.75987113e-02 1.93803027e-01 2.00560778e-01
1.02559042e+00 -7.52054632e-01 9.05535817e-01 6.15775228e-01
4.99888480e-01 -8.81697416e-01 5.47356665e-01 7.32321560e-01
-1.58570915e-01 -7.77026534e-01 -3.78110409e-01 -4.46164273e-02
-8.02688971e-02 4.78082120e-01 -1.19509780e+00 1.13848768e-01
2.39822745e-01 1.38572738e-01 3.53132576e-01 1.47095203e+00
3.94493574e-03 9.90705132e-01 -6.15871251e-01 -5.28180957e-01
7.23764598e-01 -5.22242546e-01 5.69323182e-01 1.04206109e+00
5.54376304e-01 1.93959773e-02 2.69444436e-01 7.09599853e-01
7.59087084e-03 2.00128689e-01 -3.87845099e-01 -2.06725910e-01
5.57940245e-01 5.90298593e-01 -1.06076077e-01 -5.35607636e-01
-9.09138694e-02 3.28323394e-01 3.58929724e-01 7.34254360e-01
-6.94590092e-01 3.31605166e-01 6.58398569e-01 -1.64263219e-01
5.95517397e-01 1.96918070e-01 -2.10251465e-01 -6.60423219e-01
2.19153315e-01 -8.81215155e-01 5.38680375e-01 -3.01581681e-01
-1.06395280e+00 -1.58743441e-01 4.26975787e-01 -7.51705110e-01
-6.04926884e-01 -3.66384953e-01 -1.26631215e-01 8.73090506e-01
-1.33274138e+00 -2.40152583e-01 6.66203558e-01 3.80342722e-01
3.41356188e-01 -9.38813612e-02 5.48681021e-01 -1.84279114e-01
-6.87874258e-01 5.88545024e-01 1.26693928e+00 -2.63663143e-01
1.40937552e-01 -1.49691367e+00 -3.33179772e-01 5.80688298e-01
-2.08128199e-01 4.24337238e-01 8.71993780e-01 -4.62168902e-01
-1.46876049e+00 -9.70470846e-01 -9.59752966e-03 -6.29748926e-02
9.41441357e-01 7.73304626e-02 -1.40405670e-01 9.83006954e-01
-1.81586593e-01 -1.30530998e-01 4.55084085e-01 3.32119733e-01
-1.50873344e-02 -5.72737813e-01 -1.27600873e+00 4.92034048e-01
6.06849074e-01 -4.69172047e-03 -1.36701927e-01 7.23220885e-01
5.81416011e-01 -4.68063831e-01 -8.74393463e-01 1.63214862e-01
9.49619293e-01 -6.13660455e-01 7.16239154e-01 -1.45881617e+00
-4.16759141e-02 3.62379164e-01 -2.31974393e-01 -1.46289659e+00
-1.26213893e-01 -1.24213791e+00 -3.41394305e-01 6.49812400e-01
7.54208565e-01 -8.25763285e-01 1.03498495e+00 9.15283740e-01
3.97646457e-01 -7.66238928e-01 -1.19908762e+00 -1.16842079e+00
1.90631971e-01 -5.71011186e-01 5.32052815e-01 2.04916328e-01
-1.68337762e-01 -8.77851173e-02 -7.48695254e-01 2.56768260e-02
9.13649619e-01 5.06860852e-01 8.73614192e-01 -6.19534016e-01
-8.71114016e-01 -6.12661541e-01 3.26255709e-01 -1.67504823e+00
7.81701207e-02 -2.74272949e-01 1.43461496e-01 -1.20447779e+00
4.54128921e-01 -5.07643640e-01 -5.06932795e-01 2.29541838e-01
-1.38846993e-01 -7.55274475e-01 2.89675355e-01 -1.66387156e-01
-7.22595394e-01 3.04617435e-01 1.23573530e+00 -1.03437155e-01
-3.20500433e-01 9.54209745e-01 -9.43431795e-01 1.02854460e-01
8.79789054e-01 -7.60361135e-01 -4.72873420e-01 -2.19301328e-01
1.14589572e-01 7.90936232e-01 2.23733068e-01 -3.71534348e-01
-1.98436938e-02 -9.29446518e-01 6.27803430e-02 -4.89823371e-01
1.14694528e-01 -1.03639448e+00 1.26555145e-01 5.43228507e-01
-9.76047337e-01 -2.34772742e-01 3.04494910e-02 9.42777574e-01
3.76134336e-01 -6.78138793e-01 5.77241898e-01 -2.21388906e-01
2.06999704e-01 4.14948434e-01 -3.31206888e-01 1.50375843e-01
1.14955318e+00 2.64698684e-01 -1.97872654e-01 -1.03198409e+00
-7.11346567e-01 5.05340338e-01 -9.60393026e-02 -2.56399125e-01
1.51091248e-01 -1.18880761e+00 -8.18464339e-01 -2.30360627e-01
-2.86883801e-01 -4.88222837e-01 -2.08878040e-01 8.81386101e-01
5.20353466e-02 9.77790236e-01 3.65533143e-01 -2.67134875e-01
-1.11540270e+00 6.91680849e-01 5.33127248e-01 -5.59248686e-01
7.05512762e-02 5.11765003e-01 -2.57719457e-01 6.61050342e-03
6.66426897e-01 -4.20685440e-01 3.66621315e-01 -8.47388059e-02
7.39333868e-01 7.46198177e-01 -4.57026213e-02 2.42864296e-01
3.03994775e-01 -5.62253371e-02 -4.17209938e-02 -8.15550804e-01
1.35804486e+00 2.27428991e-02 2.38032520e-01 3.44247103e-01
1.03976405e+00 -1.90120131e-01 -1.65249884e+00 -5.87490320e-01
2.25559726e-01 -9.23807144e-01 5.14940143e-01 -9.50452447e-01
-7.98629582e-01 2.83228904e-01 5.48026621e-01 7.26572871e-01
7.42237687e-01 -2.04623699e-01 2.33396962e-01 7.53876507e-01
2.95196354e-01 -1.27519250e+00 -5.32496214e-01 1.76253647e-01
7.67638743e-01 -1.06815457e+00 1.50794610e-01 4.47704494e-01
-6.39133036e-01 7.46173680e-01 -2.95137987e-02 -1.09072119e-01
4.26071018e-01 -3.57644930e-02 -4.18079525e-01 1.89456716e-01
-8.84584725e-01 -4.86202329e-01 1.73106298e-01 2.54795730e-01
-1.81016579e-01 6.15802348e-01 -4.62501913e-01 1.80458471e-01
-1.47654057e-01 2.29195446e-01 2.51040131e-01 8.83722186e-01
-6.72930241e-01 -9.44615602e-01 -7.63969302e-01 8.99769843e-01
-8.72352421e-01 9.27811861e-02 -8.85161459e-02 4.04559940e-01
-7.31721997e-01 1.12748158e+00 8.52624774e-02 4.77177113e-01
6.17455542e-01 -1.03528611e-02 6.77329838e-01 -2.32501686e-01
-1.11100748e-01 4.34336066e-01 3.61943722e-01 -5.91170251e-01
-3.50368828e-01 -7.43042409e-01 -3.76483500e-01 -5.10277092e-01
-3.45321685e-01 6.32547796e-01 7.94555664e-01 9.45804954e-01
1.49349809e-01 2.28031904e-01 9.76072550e-01 -4.84723896e-01
-1.73945200e+00 -9.86125350e-01 -6.08285010e-01 1.88905686e-01
8.88833106e-01 -4.91405666e-01 -2.82320321e-01 -5.72786629e-01] | [4.487074375152588, 3.2356362342834473] |
b885bb73-3a93-4b8c-9c25-07824f7a1716 | vidit-virtual-image-dataset-for-illumination | 2005.05460 | null | https://arxiv.org/abs/2005.05460v2 | https://arxiv.org/pdf/2005.05460v2.pdf | VIDIT: Virtual Image Dataset for Illumination Transfer | Deep image relighting is gaining more interest lately, as it allows photo enhancement through illumination-specific retouching without human effort. Aside from aesthetic enhancement and photo montage, image relighting is valuable for domain adaptation, whether to augment datasets for training or to normalize input test data. Accurate relighting is, however, very challenging for various reasons, such as the difficulty in removing and recasting shadows and the modeling of different surfaces. We present a novel dataset, the Virtual Image Dataset for Illumination Transfer (VIDIT), in an effort to create a reference evaluation benchmark and to push forward the development of illumination manipulation methods. Virtual datasets are not only an important step towards achieving real-image performance but have also proven capable of improving training even when real datasets are possible to acquire and available. VIDIT contains 300 virtual scenes used for training, where every scene is captured 40 times in total: from 8 equally-spaced azimuthal angles, each lit with 5 different illuminants. | ['Sabine Süsstrunk', 'Majed El Helou', 'Ruofan Zhou', 'Johan Barthas'] | 2020-05-11 | null | null | null | null | ['image-relighting'] | ['computer-vision'] | [ 7.03656137e-01 -1.91438407e-01 3.00764948e-01 -3.58028620e-01
-5.28420746e-01 -7.19006717e-01 8.42271328e-01 6.67008460e-02
-3.90519798e-01 9.03090239e-01 1.63243726e-01 -3.16816159e-02
2.51731634e-01 -7.18446195e-01 -8.22339952e-01 -7.33360946e-01
2.08507538e-01 3.58155042e-01 5.56964763e-02 -2.75687069e-01
1.54262424e-01 7.25796282e-01 -1.90833247e+00 1.28825650e-01
8.08598638e-01 8.27577055e-01 2.69699097e-01 6.42689943e-01
-2.25290917e-02 4.78134245e-01 -6.66454732e-01 -5.94370663e-01
3.54666531e-01 -5.05837262e-01 -4.82697129e-01 1.77755892e-01
8.45756054e-01 -2.37600058e-01 -7.32090622e-02 7.18215406e-01
6.46947384e-01 4.32996929e-01 5.79861343e-01 -1.36109483e+00
-6.03493512e-01 -1.51946232e-01 -5.69366634e-01 -1.92194402e-01
3.64792407e-01 2.95834482e-01 6.59307301e-01 -7.88490772e-01
9.11309063e-01 9.56810653e-01 5.42422533e-01 4.15954500e-01
-1.49370527e+00 -4.41773355e-01 -3.70204657e-01 3.78847361e-01
-1.22025049e+00 -4.99655634e-01 8.64285827e-01 -2.79891223e-01
8.14047575e-01 5.36076725e-01 7.79560268e-01 1.29676163e+00
-2.02693250e-02 7.02824056e-01 1.50357556e+00 -5.65514803e-01
2.52050102e-01 5.56155205e-01 -4.43077505e-01 1.87374070e-01
2.29956269e-01 2.78085113e-01 -4.20301229e-01 2.68876880e-01
7.60203838e-01 -2.49420330e-01 -6.70599997e-01 -5.51220894e-01
-1.08306897e+00 5.42509198e-01 5.59536278e-01 5.77462930e-03
-3.69174063e-01 9.65179224e-03 2.57991731e-01 3.61815214e-01
4.30317014e-01 6.68021381e-01 -4.49852943e-01 -2.00875074e-01
-8.01427901e-01 1.51630759e-01 4.87203807e-01 6.48293614e-01
8.72536778e-01 2.96936870e-01 -1.08675152e-01 9.51717556e-01
-2.45058328e-01 7.05200791e-01 1.81595266e-01 -1.04264772e+00
2.78001070e-01 4.88129854e-01 3.11924279e-01 -9.00076926e-01
-1.96080327e-01 -3.03042054e-01 -9.08108592e-01 6.19693041e-01
4.11220491e-01 -5.97099066e-02 -9.56689239e-01 1.51974106e+00
3.05215031e-01 7.90328979e-02 9.31699425e-02 1.02359855e+00
7.74404645e-01 6.35384500e-01 2.84017175e-02 -3.81700397e-02
1.25369906e+00 -7.56729484e-01 -5.22563338e-01 -3.26272160e-01
3.81643564e-01 -1.20871627e+00 1.43301773e+00 8.01956713e-01
-1.06016994e+00 -5.23847222e-01 -8.68095636e-01 -1.85144871e-01
-5.91469765e-01 -3.76946144e-02 5.96006095e-01 8.33755672e-01
-1.00123322e+00 4.72220808e-01 -2.98048347e-01 -4.08317327e-01
4.31743741e-01 -6.00275509e-02 -6.85535610e-01 -4.46724504e-01
-1.00828588e+00 1.02914417e+00 1.65989354e-01 -9.57561955e-02
-6.98957324e-01 -6.82871103e-01 -9.13335025e-01 -7.68349618e-02
2.47826561e-01 -4.41672981e-01 8.38894606e-01 -1.30880117e+00
-1.38727498e+00 1.10806668e+00 1.42970225e-02 -3.60399455e-01
7.64198661e-01 -1.51041701e-01 -5.15105367e-01 -6.24818280e-02
-2.99527764e-01 8.19381416e-01 1.08753169e+00 -1.68958962e+00
-5.68428226e-02 -1.10142894e-01 -1.49533823e-02 3.73413622e-01
-4.19636548e-01 -1.65600643e-01 -5.24444342e-01 -7.23470926e-01
-3.22371483e-01 -8.85310292e-01 3.29789929e-02 1.88269585e-01
-3.78579378e-01 6.14424229e-01 8.33297491e-01 -7.67797709e-01
6.19897366e-01 -2.13822603e+00 -7.77463689e-02 2.66757272e-02
-2.58173764e-01 5.96042931e-01 -3.62199277e-01 5.21231532e-01
-2.44638264e-01 -1.74660292e-02 -2.85536706e-01 -4.87779051e-01
-1.25345498e-01 4.15972769e-01 -2.92218149e-01 4.13591057e-01
2.33653888e-01 9.81464088e-01 -6.84705436e-01 -4.52660471e-01
7.99169838e-01 9.70611751e-01 -3.18276703e-01 1.25612095e-01
-2.28110820e-01 6.89542711e-01 2.77049243e-01 4.38239157e-01
9.52970862e-01 1.62432626e-01 -1.61934882e-01 -3.55964541e-01
-1.51146308e-01 -2.70139128e-01 -1.24868000e+00 1.51498950e+00
-7.71317363e-01 1.17815256e+00 -1.49319038e-01 -6.34400010e-01
9.14353192e-01 9.71541852e-02 3.83771300e-01 -1.34621847e+00
1.89295068e-01 -9.43747461e-02 -5.63806057e-01 -4.43090945e-01
9.78819072e-01 -9.82784331e-02 1.91152751e-01 2.95738071e-01
-3.21168989e-01 -7.03232646e-01 2.02679753e-01 1.79164615e-02
6.27754033e-01 2.80648291e-01 3.13552842e-02 5.33669852e-02
3.77732724e-01 1.58121124e-01 2.42709115e-01 3.04868042e-01
-1.57436896e-02 1.10081184e+00 1.83394179e-01 -4.92266238e-01
-1.26435947e+00 -1.27351534e+00 -3.10734630e-01 7.91177213e-01
2.71934897e-01 8.35467204e-02 -5.51191151e-01 -3.13121855e-01
-2.38381103e-01 1.15002525e+00 -6.69536591e-01 -8.96513313e-02
-2.34872684e-01 -6.11125529e-01 2.29509175e-01 2.02220917e-01
7.32220411e-01 -1.14717603e+00 -7.05453217e-01 -7.94173181e-02
-3.94872427e-01 -1.17757070e+00 -2.31204182e-01 2.32729346e-01
-5.50659716e-01 -1.05717039e+00 -9.30575192e-01 -4.55660850e-01
7.55976439e-01 6.60450280e-01 1.44223022e+00 -7.84568042e-02
-6.39238119e-01 5.71761727e-01 -4.10056949e-01 -5.75145721e-01
-4.02393222e-01 -3.02132785e-01 -3.45016032e-01 -3.38494405e-02
-1.02708101e-01 -5.04497111e-01 -8.49083960e-01 6.67055607e-01
-1.38899100e+00 2.67841965e-01 4.65559602e-01 7.64278173e-01
5.38854122e-01 -9.40926932e-03 1.23060726e-01 -7.45832264e-01
2.54923820e-01 -6.01233318e-02 -6.63118660e-01 2.00327665e-01
-4.10065174e-01 -2.62467414e-01 5.31338573e-01 -3.04713875e-01
-1.50641906e+00 -1.94526106e-01 -6.59216419e-02 -3.22380394e-01
-3.61550540e-01 1.83776822e-02 -1.41321629e-01 -5.06402433e-01
9.26803052e-01 1.44056961e-01 -9.22714770e-02 -2.96066523e-01
5.76854169e-01 4.02800828e-01 7.10094273e-01 -3.36818010e-01
9.29595053e-01 6.29232764e-01 1.62003919e-01 -1.14584470e+00
-4.53830361e-01 -2.42014259e-01 -5.40915668e-01 -4.46941823e-01
6.57029748e-01 -7.04029024e-01 -4.11637098e-01 4.39470023e-01
-8.57177436e-01 -8.21547627e-01 -5.63317299e-01 1.22675888e-01
-3.35505217e-01 2.20033661e-01 -1.72450662e-01 -5.46327651e-01
-8.16520210e-03 -1.05776989e+00 1.09072578e+00 4.25326407e-01
-4.13963981e-02 -1.01341975e+00 3.96467410e-02 5.95211446e-01
4.60880905e-01 5.47441840e-01 7.79682338e-01 2.33154073e-01
-4.73211735e-01 -2.25277260e-01 -4.88561928e-01 6.93779230e-01
1.56234592e-01 2.05796644e-01 -1.36565232e+00 -3.20117474e-01
-4.14092928e-01 -5.57428360e-01 6.73622668e-01 1.41293660e-01
1.14994645e+00 2.67863390e-03 1.09842293e-01 7.78317630e-01
1.68531477e+00 -3.10520474e-02 1.20331836e+00 5.02967298e-01
5.87776959e-01 6.34112477e-01 7.26284087e-01 3.68757516e-01
1.31763473e-01 9.79938209e-01 7.49978423e-01 -8.07440162e-01
-6.14450514e-01 3.95036815e-03 4.28260192e-02 1.29499659e-01
-2.30458558e-01 -5.17151117e-01 -8.31778169e-01 3.61936837e-01
-1.18442118e+00 -7.82479405e-01 -3.78086329e-01 2.71352792e+00
6.59071743e-01 -2.15680882e-01 -1.57839909e-01 2.42992267e-01
3.86874497e-01 3.05721909e-02 -4.89084780e-01 -5.07007539e-01
-5.70492744e-01 3.10355186e-01 4.57946658e-01 4.41148341e-01
-8.38336527e-01 9.00290072e-01 5.48392868e+00 6.03243232e-01
-1.44420660e+00 -1.52639553e-01 7.31463373e-01 -1.26400366e-02
-4.10574555e-01 -2.43386775e-01 -2.94349343e-01 2.21118450e-01
4.84831631e-01 2.46982425e-01 6.77453458e-01 7.03649342e-01
2.84964949e-01 -5.52137792e-01 -7.16398180e-01 1.17009091e+00
3.24049264e-01 -1.09910369e+00 -7.17387795e-02 4.12807949e-02
1.02757716e+00 -1.14954459e-02 3.19502532e-01 8.58370811e-02
1.12676032e-01 -9.48697567e-01 5.61542392e-01 3.29921037e-01
1.12119043e+00 -6.95540607e-01 5.84088385e-01 4.34405804e-02
-8.64495099e-01 2.53678679e-01 -4.25284863e-01 5.08576855e-02
1.58891633e-01 3.93589258e-01 -1.07960606e+00 6.51598990e-01
6.03554487e-01 3.51165593e-01 -8.69003952e-01 1.20841384e+00
-3.20251167e-01 3.53415489e-01 -2.12793887e-01 2.03544021e-01
-1.19112032e-02 -2.97200084e-01 4.43744063e-01 1.11590195e+00
3.63282353e-01 -2.58659840e-01 -2.69032478e-01 4.57754493e-01
-8.08429718e-02 1.90090567e-01 -7.68455148e-01 1.85899630e-01
2.15425298e-01 1.39211118e+00 -8.05774987e-01 -5.07757105e-02
-3.21712673e-01 1.51187944e+00 -6.00814167e-03 6.92415297e-01
-9.32068169e-01 -3.44088018e-01 5.78727901e-01 1.35527000e-01
1.58900693e-02 -3.04185450e-01 -2.86542207e-01 -1.04127431e+00
-1.25678629e-01 -8.05789888e-01 8.57537761e-02 -1.47852945e+00
-9.35734332e-01 6.70155346e-01 6.67232201e-02 -1.29035461e+00
-4.70257699e-02 -5.49239874e-01 -4.91752893e-01 8.37779224e-01
-1.80089605e+00 -1.31497324e+00 -1.14739120e+00 6.53242588e-01
4.37384963e-01 2.39934951e-01 8.83228600e-01 3.56171191e-01
-3.23126346e-01 4.60574001e-01 3.57112050e-01 -3.27970922e-01
1.18271506e+00 -1.19366324e+00 3.85601401e-01 7.67392814e-01
3.57013732e-01 9.52560902e-02 1.04402685e+00 -1.00036696e-01
-1.39407456e+00 -1.18595278e+00 5.73017716e-01 -5.21008074e-01
1.19949415e-01 -4.28554356e-01 -9.80127633e-01 3.48875642e-01
3.61415535e-01 -2.81815231e-03 6.27086520e-01 -1.49988547e-01
-3.11659664e-01 -3.67212832e-01 -1.24941671e+00 8.81823778e-01
9.11243320e-01 -3.33032846e-01 -1.68632776e-01 2.55378753e-01
2.93310761e-01 -4.51381594e-01 -6.00167096e-01 3.58576655e-01
5.47572255e-01 -1.46830595e+00 1.27864873e+00 3.09398421e-03
5.65755069e-01 -2.75028914e-01 -1.71108231e-01 -1.60856640e+00
2.16511227e-02 -5.23638844e-01 4.89440829e-01 1.44121790e+00
2.28103250e-01 -7.85691559e-01 8.54323506e-01 6.40876889e-01
-1.64794475e-01 -2.53639847e-01 -6.72103822e-01 -7.82148778e-01
-2.47579813e-01 -6.80030286e-01 5.85960686e-01 1.04409313e+00
-7.11020410e-01 2.00797558e-01 -5.49747169e-01 -7.23451301e-02
5.96231222e-01 1.02397807e-01 1.36879385e+00 -9.83205795e-01
-3.08698148e-01 -3.58519405e-01 -3.30781013e-01 -5.83787203e-01
-2.65956104e-01 -5.72627664e-01 -4.64359336e-02 -1.46468139e+00
-3.76463518e-03 -4.61468846e-01 6.79748878e-02 4.08259571e-01
-6.02799691e-02 1.12113309e+00 2.05760583e-01 -7.06385598e-02
-2.52718687e-01 6.51661694e-01 1.54154313e+00 -5.11217825e-02
-1.62453562e-01 -2.58312047e-01 -5.09439647e-01 5.36529779e-01
8.16003203e-01 -1.01368234e-01 -7.32282639e-01 -4.77623999e-01
3.35750639e-01 -1.86022222e-01 6.93717837e-01 -1.14041662e+00
-3.65701079e-01 -2.07932010e-01 6.48973286e-01 -5.37399590e-01
7.48682320e-01 -9.63691890e-01 5.26694417e-01 1.91319957e-01
-2.66720187e-02 -2.44654581e-01 6.06906056e-01 3.62734288e-01
-2.05508232e-01 -4.60398532e-02 1.04512525e+00 3.92580144e-02
-1.06374621e+00 8.85927156e-02 1.08623505e-01 7.28993937e-02
1.03824401e+00 -6.41448677e-01 -3.95624757e-01 -4.81907517e-01
-2.74148941e-01 -1.09190211e-01 1.14660156e+00 3.78730208e-01
5.06987751e-01 -1.13447809e+00 -5.47524571e-01 3.89889508e-01
3.89497340e-01 -1.40826628e-01 5.82312584e-01 5.37903905e-01
-7.71303833e-01 -8.40237923e-03 -5.34564257e-01 -6.01509869e-01
-1.62681651e+00 5.29591382e-01 3.65998484e-02 2.64872368e-02
-6.74952149e-01 8.63419116e-01 4.55892920e-01 -1.14111707e-01
2.24440947e-01 -8.04450177e-03 -1.54494392e-02 3.84258591e-02
5.50478041e-01 2.96446741e-01 4.00179595e-01 -4.32128996e-01
-7.51770139e-02 5.57065129e-01 2.11964369e-01 -2.07394809e-01
1.27451158e+00 -2.49932721e-01 1.47816837e-01 2.46826336e-01
1.07059860e+00 -2.08120458e-02 -1.38493872e+00 -1.09315991e-01
-4.72507805e-01 -1.04573834e+00 2.41179183e-01 -1.16157496e+00
-9.86752987e-01 1.03810072e+00 8.52852225e-01 5.33386618e-02
1.42458522e+00 -4.20821905e-01 6.31692350e-01 2.36168727e-01
4.07793552e-01 -1.00588799e+00 4.02797461e-01 2.35939741e-01
1.14592874e+00 -1.45808363e+00 4.00584452e-02 -2.70661950e-01
-8.96772802e-01 9.13590312e-01 4.16631401e-01 2.24938080e-01
3.18962187e-02 2.56297022e-01 3.22501183e-01 6.47865608e-02
-3.00760090e-01 -2.51799941e-01 3.32327098e-01 1.08821547e+00
4.52043116e-01 -9.92158204e-02 2.03944683e-01 -2.14408875e-01
-1.55813679e-01 -1.41491950e-01 5.26775420e-01 6.04991972e-01
-1.65314287e-01 -1.15453935e+00 -7.00380802e-01 2.44549021e-01
-1.13431653e-02 -1.17288582e-01 -4.53668892e-01 1.02554977e+00
1.57111481e-01 7.62824893e-01 1.20094260e-02 -1.33960649e-01
4.13919896e-01 -2.53703028e-01 8.04550111e-01 -2.10326359e-01
-2.94996947e-01 -1.14907287e-01 2.47725859e-01 -4.81605053e-01
-1.79883391e-01 -6.51488602e-01 -6.67666376e-01 -4.25388396e-01
-1.03794813e-01 -1.07864656e-01 1.10246634e+00 5.29831350e-01
7.61329457e-02 6.09460711e-01 6.92998409e-01 -1.34859943e+00
2.46874075e-02 -7.39325941e-01 -5.45221806e-01 7.46946275e-01
1.48440972e-01 -7.60305941e-01 -2.75082678e-01 3.22689414e-01] | [10.47054386138916, -2.375222682952881] |
bec1bd31-b930-4ae8-b943-8d30ad5928a7 | transferable-adversarial-robustness-for | 2306.04064 | null | https://arxiv.org/abs/2306.04064v1 | https://arxiv.org/pdf/2306.04064v1.pdf | Transferable Adversarial Robustness for Categorical Data via Universal Robust Embeddings | Research on adversarial robustness is primarily focused on image and text data. Yet, many scenarios in which lack of robustness can result in serious risks, such as fraud detection, medical diagnosis, or recommender systems often do not rely on images or text but instead on tabular data. Adversarial robustness in tabular data poses two serious challenges. First, tabular datasets often contain categorical features, and therefore cannot be tackled directly with existing optimization procedures. Second, in the tabular domain, algorithms that are not based on deep networks are widely used and offer great performance, but algorithms to enhance robustness are tailored to neural networks (e.g. adversarial training). In this paper, we tackle both challenges. We present a method that allows us to train adversarially robust deep networks for tabular data and to transfer this robustness to other classifiers via universal robust embeddings tailored to categorical data. These embeddings, created using a bilevel alternating minimization framework, can be transferred to boosted trees or random forests making them robust without the need for adversarial training while preserving their high accuracy on tabular data. We show that our methods outperform existing techniques within a practical threat model suitable for tabular data. | ['Nicolas Flammarion', 'Carmela Troncoso', 'Maksym Andriushchenko', 'Klim Kireev'] | 2023-06-06 | null | null | null | null | ['adversarial-robustness', 'medical-diagnosis', 'fraud-detection'] | ['adversarial', 'medical', 'miscellaneous'] | [ 2.33377054e-01 1.31826416e-01 -1.64070111e-02 -1.66875809e-01
-8.65775704e-01 -1.01827800e+00 6.59918070e-01 2.06808001e-01
-1.82718202e-01 7.38353908e-01 1.09231390e-01 -3.61390591e-01
-1.61727399e-01 -1.18384016e+00 -9.84552145e-01 -8.10944498e-01
3.52156796e-02 4.16454285e-01 -8.98588821e-02 -4.48956072e-01
-1.66014396e-02 6.71816766e-01 -1.06571996e+00 1.89208910e-01
8.18374336e-01 1.17202485e+00 -6.99047804e-01 5.64640462e-01
1.04871757e-01 8.66007566e-01 -7.55382001e-01 -1.14397287e+00
7.06981122e-01 6.48716092e-02 -4.59379554e-01 -2.53763914e-01
8.84652138e-01 -2.88942844e-01 -7.49558210e-01 1.23863435e+00
5.98630309e-01 8.94906446e-02 8.35073113e-01 -1.51874530e+00
-1.10344982e+00 6.33878529e-01 -2.09669277e-01 -3.73956442e-01
6.74212277e-02 1.23101860e-01 7.85288930e-01 -6.05930805e-01
5.96999109e-01 1.58106232e+00 8.83821189e-01 8.14365268e-01
-1.22805190e+00 -7.34515429e-01 1.02888383e-01 1.69262230e-01
-8.65533769e-01 -4.07516718e-01 7.65215516e-01 -4.53890026e-01
2.75567085e-01 7.28793442e-01 7.91940019e-02 1.55813432e+00
5.16432643e-01 6.32349670e-01 9.73603249e-01 8.10796469e-02
1.70643717e-01 3.93027067e-01 -3.77006441e-01 5.60346782e-01
3.09486598e-01 2.77626604e-01 -1.57729283e-01 -1.72692135e-01
3.01151752e-01 3.40136200e-01 -3.55052859e-01 -7.26889312e-01
-1.27306032e+00 1.23110902e+00 8.46798837e-01 -2.02115085e-02
3.79653126e-02 1.40136540e-01 8.66889775e-01 5.75626075e-01
4.23066139e-01 8.02848458e-01 -4.15834218e-01 4.22801793e-01
-7.33651817e-01 3.17286581e-01 5.52833915e-01 5.86653173e-01
2.91879416e-01 1.26389116e-01 -2.67914802e-01 7.73259699e-01
2.35577729e-02 7.51542091e-01 3.51355135e-01 -7.61577785e-01
6.32324815e-01 4.30239022e-01 -1.87299207e-01 -1.52404261e+00
-4.46740568e-01 -1.38156056e-01 -1.50088203e+00 4.17156607e-01
4.61159080e-01 -1.88341662e-02 -1.09125996e+00 1.78173351e+00
3.36887211e-01 -5.85975349e-01 2.08075017e-01 7.76785910e-01
7.33642757e-01 4.86950755e-01 -2.09087938e-01 1.13886356e-01
1.10415745e+00 -9.30279493e-01 -7.15646446e-01 -1.69925392e-01
2.93699771e-01 -6.82203591e-01 9.19083595e-01 5.37225366e-01
-8.74425352e-01 -2.02663392e-01 -1.07124519e+00 -1.86250716e-01
-1.02035260e+00 -5.11817873e-01 5.06214619e-01 9.83681023e-01
-8.16567838e-01 7.51331508e-01 -4.01527196e-01 -1.97264522e-01
6.12507164e-01 2.66370535e-01 -8.85489941e-01 -2.29444072e-01
-1.69917846e+00 1.13250363e+00 2.37786576e-01 1.90939590e-01
-9.16502953e-01 -5.21904528e-01 -1.11964083e+00 -8.88209417e-02
3.94391030e-01 -7.97002017e-01 6.42673790e-01 -9.99481022e-01
-1.24252915e+00 6.39454067e-01 5.53096473e-01 -8.22307527e-01
1.00359941e+00 -1.64200962e-01 -3.11081350e-01 1.63562506e-01
-9.81916953e-03 4.97009695e-01 1.53470671e+00 -1.12984622e+00
-1.33103535e-01 -4.38375175e-01 4.00077969e-01 -6.32973015e-02
-7.94884205e-01 -1.02079809e-01 -2.34492317e-01 -1.21750343e+00
-1.30643547e-01 -1.05008686e+00 -5.18318892e-01 4.98297691e-01
-6.58056855e-01 1.80042759e-01 9.04099941e-01 -7.18059957e-01
9.43980217e-01 -2.01565313e+00 4.20407414e-01 2.07993805e-01
2.74347514e-01 3.00951034e-01 -7.05526993e-02 3.05685699e-01
-3.25460941e-01 5.20241380e-01 -5.91674030e-01 -9.36544761e-02
2.62326032e-01 2.24861324e-01 -5.34057260e-01 7.24620461e-01
2.62405157e-01 9.38279510e-01 -6.29413426e-01 -3.01818967e-01
3.93433005e-01 3.86373430e-01 -5.63302934e-01 1.89951986e-01
-1.54140532e-01 2.16196090e-01 -1.30743444e-01 7.49348998e-01
9.30713713e-01 1.93775371e-01 -6.64895251e-02 -2.42826283e-01
4.29090708e-01 -1.95834443e-01 -9.48324978e-01 1.26571286e+00
-5.54216206e-01 4.71047252e-01 1.70800969e-01 -1.22430503e+00
6.30023777e-01 1.61560684e-01 2.78800577e-01 -5.24728060e-01
6.35283627e-03 5.66298328e-02 -4.13816661e-01 -2.38161176e-01
3.99543345e-01 -3.28914285e-01 -6.28398716e-01 2.99074113e-01
-1.49976552e-01 -3.66261512e-01 -1.11012354e-01 3.74577194e-01
1.08728242e+00 -1.89157858e-01 -4.88918796e-02 6.17138296e-02
4.87622172e-01 -5.30517623e-02 4.17664111e-01 8.31681430e-01
-1.38884053e-01 8.54625463e-01 6.92457676e-01 -5.81318319e-01
-1.10204923e+00 -1.15973890e+00 -3.33341599e-01 8.07375729e-01
-1.74152814e-02 -2.08929420e-01 -6.64527714e-01 -1.33964777e+00
3.33927125e-01 4.47080821e-01 -1.07677054e+00 -7.02080548e-01
-3.82571518e-01 -8.75097215e-01 7.54039407e-01 4.70116884e-01
4.91295159e-01 -9.07685816e-01 -7.77411163e-02 3.53261493e-02
-1.16062857e-01 -8.99511993e-01 -4.75272477e-01 3.96794617e-01
-7.53296554e-01 -1.10379839e+00 -6.76278651e-01 -3.56620699e-01
6.82926416e-01 -2.44240724e-02 1.04605460e+00 -9.47887003e-02
-1.96323901e-01 2.89812773e-01 -3.74902695e-01 -2.92623103e-01
-6.25299454e-01 8.30065738e-03 3.13115120e-01 7.74415508e-02
-3.00645590e-01 -3.87285352e-01 -4.04640913e-01 4.43357408e-01
-1.27262986e+00 -4.26764309e-01 4.84231949e-01 1.34150803e+00
3.01954210e-01 2.27429435e-01 6.10783517e-01 -1.11676788e+00
3.17473680e-01 -6.48041248e-01 -4.95449007e-01 2.11020544e-01
-5.26710987e-01 1.28483936e-01 1.19010127e+00 -5.47482550e-01
-5.10920107e-01 -1.41681328e-01 -3.18274736e-01 -8.79816115e-01
3.59354056e-02 3.59076411e-01 -4.67559427e-01 -1.77602738e-01
9.61643100e-01 -6.60252571e-02 6.89753965e-02 -2.25219250e-01
6.61725521e-01 5.10360777e-01 5.69800675e-01 -3.98365676e-01
1.54968166e+00 6.19578063e-01 1.64193556e-01 -2.13027686e-01
-7.08208203e-01 1.82726368e-01 -4.99877989e-01 3.71984020e-02
7.55643785e-01 -6.32470071e-01 -4.97924358e-01 5.17456830e-01
-7.65052080e-01 -1.29008502e-01 -2.81421483e-01 -4.61288914e-03
-5.64676881e-01 5.62613964e-01 -4.49718028e-01 -5.36726713e-01
-4.76427436e-01 -1.00535476e+00 8.86000514e-01 -2.30788678e-01
-3.02096712e-03 -1.15167367e+00 -2.46985123e-01 5.27580798e-01
4.85241801e-01 8.05121660e-01 1.00376391e+00 -6.43705130e-01
-4.55565721e-01 -7.70930827e-01 -1.93925798e-01 6.17563546e-01
2.49509528e-01 -1.97548568e-02 -9.76388991e-01 -6.06995404e-01
3.54419425e-02 -6.78235710e-01 1.08834195e+00 4.90537770e-02
1.61164629e+00 -9.16570246e-01 -7.64169917e-02 1.03817177e+00
1.15932381e+00 -1.56346142e-01 7.69119084e-01 6.20553553e-01
1.01856840e+00 8.31856191e-01 5.99646747e-01 9.42113176e-02
1.83032915e-01 6.11187160e-01 1.06810534e+00 -2.13701934e-01
2.00002685e-01 -3.29164416e-01 5.76127827e-01 3.74805301e-01
4.61673707e-01 -3.97547036e-01 -7.90567219e-01 5.85312724e-01
-1.87094700e+00 -1.07061231e+00 6.44682497e-02 2.31189322e+00
7.95343399e-01 2.42663026e-01 1.00774253e-02 3.37382227e-01
6.56298697e-01 2.04793975e-01 -6.71260476e-01 -7.22334683e-01
-4.41688895e-01 -1.31748393e-02 7.88296640e-01 1.81443498e-01
-1.67384279e+00 8.10809076e-01 6.30055857e+00 9.99190331e-01
-1.22251618e+00 1.22754127e-01 8.89451683e-01 -2.16694906e-01
-3.53184432e-01 -4.36450630e-01 -1.74669325e-01 4.91496474e-01
9.11403537e-01 1.80996060e-02 4.52647001e-01 9.73342180e-01
-2.21026927e-01 5.05356133e-01 -1.07499707e+00 8.92042935e-01
1.40385836e-01 -1.27476406e+00 4.13891882e-01 -4.23402414e-02
6.93473399e-01 -4.58071887e-01 8.31789494e-01 3.86218548e-01
5.81504166e-01 -1.59721386e+00 7.40535378e-01 2.37841815e-01
9.40566897e-01 -1.01908052e+00 7.77292728e-01 7.95866624e-02
-5.60915053e-01 -2.39386171e-01 -6.02928340e-01 3.70282263e-01
-2.44897231e-01 7.22457707e-01 -4.97723162e-01 7.92562127e-01
1.05558503e+00 7.71741331e-01 -5.81871748e-01 5.08902788e-01
-1.97530046e-01 3.25136662e-01 -2.43045434e-01 2.83972651e-01
2.35591725e-01 4.32049893e-02 6.58882737e-01 9.58446681e-01
1.61386371e-01 -3.73068035e-01 -1.35498554e-01 6.38284385e-01
-4.52169389e-01 8.20342898e-02 -1.07573152e+00 2.24734899e-02
3.48045491e-02 1.29816091e+00 -3.67071390e-01 -1.17258556e-01
-3.07054460e-01 1.08565927e+00 2.60881484e-01 6.42684177e-02
-9.58858848e-01 -3.44187617e-01 8.05505514e-01 8.39373320e-02
2.19815508e-01 1.29752070e-01 -4.65728998e-01 -1.41013169e+00
1.06600173e-01 -1.47487354e+00 6.10518277e-01 -4.60345834e-01
-1.84429264e+00 6.58788919e-01 -2.71366537e-01 -1.26692557e+00
-1.56691581e-01 -8.26033711e-01 -4.18488950e-01 6.47895038e-01
-1.31653595e+00 -1.39761829e+00 -6.57200441e-03 1.05367386e+00
1.80848122e-01 -3.06586415e-01 9.04939413e-01 2.80945569e-01
-7.51806021e-01 1.17386937e+00 5.18014729e-01 2.58645475e-01
1.06204450e+00 -1.50772214e+00 5.59532046e-01 9.75556552e-01
-5.29886931e-02 5.49266815e-01 6.93186164e-01 -4.78304118e-01
-1.51190281e+00 -1.55528283e+00 3.89797986e-01 -9.15230691e-01
8.18082929e-01 -7.49471426e-01 -9.18173969e-01 6.06246531e-01
-3.95310782e-02 2.38868937e-01 4.60172862e-01 -2.39219256e-02
-8.71307731e-01 -2.79464662e-01 -1.62048793e+00 7.70082951e-01
7.59082317e-01 -6.02001786e-01 -4.03177738e-01 5.96691251e-01
7.94201255e-01 -3.72325480e-01 -1.00564075e+00 5.01486719e-01
4.94858623e-01 -6.96852565e-01 1.21944118e+00 -7.69298196e-01
6.67432964e-01 -2.17498973e-01 -3.73615354e-01 -1.48608303e+00
-7.36748129e-02 -6.76195383e-01 -1.88706055e-01 1.10987926e+00
3.71952206e-01 -8.73976529e-01 6.21499002e-01 5.63930631e-01
1.35984242e-01 -4.80594099e-01 -1.17325222e+00 -8.34518611e-01
7.59572506e-01 -3.54956925e-01 6.06337309e-01 1.23882556e+00
-2.99272120e-01 -3.09543490e-01 -8.22721004e-01 1.68899536e-01
7.25351870e-01 -3.86660099e-02 8.73335361e-01 -1.05129158e+00
4.09982018e-02 -3.05122584e-01 -6.95018411e-01 -2.32750177e-01
3.00437987e-01 -9.36869144e-01 4.41857502e-02 -1.28686309e+00
-3.81632447e-02 -6.37329698e-01 -3.41751546e-01 7.75626302e-01
-3.43505979e-01 7.36370981e-01 2.29510844e-01 -1.93872694e-02
-2.05470949e-01 6.30338788e-01 1.03903651e+00 -7.35792160e-01
4.25458342e-01 -4.70781215e-02 -9.53382432e-01 6.57602668e-01
8.93658876e-01 -6.32315874e-01 -1.02501377e-01 -2.42052257e-01
4.28728282e-01 1.83738787e-02 4.72958863e-01 -9.56392348e-01
7.75538087e-02 -1.82735845e-01 5.32509327e-01 -1.66838393e-01
1.96807131e-01 -1.13410258e+00 6.52965084e-02 3.83062154e-01
-2.45958194e-01 3.89786810e-02 4.11231011e-01 6.81316972e-01
-1.60903320e-01 -9.89601165e-02 1.10377872e+00 4.26588655e-02
-6.93132803e-02 5.45690119e-01 -2.26595327e-01 3.78398784e-03
1.10284913e+00 9.96760130e-02 -4.55908090e-01 -5.52975774e-01
-7.14159846e-01 2.96485454e-01 7.46432781e-01 7.24134743e-01
6.73470795e-01 -1.62637627e+00 -8.62498343e-01 1.58774123e-01
1.48257270e-01 -2.29390375e-02 3.05742055e-01 4.83037531e-01
-5.44317484e-01 2.12733164e-01 -3.85882288e-01 -2.58785278e-01
-1.12636352e+00 1.43055487e+00 5.88953018e-01 -4.56294119e-01
-3.80806059e-01 8.23590219e-01 4.06265318e-01 -8.98541212e-01
2.51582205e-01 1.04127198e-01 -2.59761494e-02 5.35363913e-01
2.51591712e-01 1.22755833e-01 2.76048303e-01 -5.73532820e-01
-3.87560964e-01 5.40780485e-01 -1.47007316e-01 8.97479057e-02
1.18844581e+00 1.82019264e-01 -2.36620635e-01 1.46773070e-01
1.21067822e+00 2.60255396e-01 -1.10624373e+00 9.01236385e-03
-4.08078909e-01 -5.82870245e-01 6.47340063e-03 -8.63449931e-01
-1.52882683e+00 1.00698519e+00 5.38910031e-01 3.24303478e-01
1.22378218e+00 -4.08061504e-01 7.78229713e-01 3.45417470e-01
1.92913324e-01 -8.41032028e-01 1.80765614e-01 3.00523251e-01
1.20273805e+00 -1.34339797e+00 -2.71646790e-02 -2.37156972e-01
-5.79961360e-01 1.11382723e+00 4.82336521e-01 -1.37611240e-01
4.52356160e-01 1.75979272e-01 2.07845822e-01 1.88830733e-01
-6.34775221e-01 6.56896532e-02 3.24977130e-01 8.64374042e-01
-1.08625188e-01 -6.52366206e-02 6.77661523e-02 4.15084124e-01
-3.18966568e-01 -7.21538723e-01 5.71430147e-01 5.57549596e-01
8.10744017e-02 -1.14817512e+00 -9.38021719e-01 5.85470200e-01
-8.84258330e-01 -2.02284440e-01 -5.02713203e-01 5.88107169e-01
8.37919638e-02 9.81078327e-01 -2.05628544e-01 -6.02764726e-01
4.38715190e-01 -1.66865081e-01 1.92803353e-01 -2.60550916e-01
-9.04350102e-01 -6.09835148e-01 -1.38371691e-01 -5.14960766e-01
9.60174054e-02 -6.14455998e-01 -4.99666870e-01 -9.04657185e-01
-1.55869707e-01 8.83900840e-03 5.45503676e-01 4.62972224e-01
-1.26848876e-01 5.24884760e-01 9.44490969e-01 -7.93260634e-01
-7.38674462e-01 -6.83488667e-01 -3.53912026e-01 7.44221747e-01
5.74508727e-01 -5.53377092e-01 -7.41223574e-01 -2.48809189e-01] | [5.610775947570801, 7.906092643737793] |
b9c4000f-3fa3-4a7a-8364-41edac576a68 | hardware-software-co-design-with-adc-less-in | 2211.02167 | null | https://arxiv.org/abs/2211.02167v2 | https://arxiv.org/pdf/2211.02167v2.pdf | Hardware/Software co-design with ADC-Less In-memory Computing Hardware for Spiking Neural Networks | Spiking Neural Networks (SNNs) are bio-plausible models that hold great potential for realizing energy-efficient implementations of sequential tasks on resource-constrained edge devices. However, commercial edge platforms based on standard GPUs are not optimized to deploy SNNs, resulting in high energy and latency. While analog In-Memory Computing (IMC) platforms can serve as energy-efficient inference engines, they are accursed by the immense energy, latency, and area requirements of high-precision ADCs (HP-ADC), overshadowing the benefits of in-memory computations. We propose a hardware/software co-design methodology to deploy SNNs into an ADC-Less IMC architecture using sense-amplifiers as 1-bit ADCs replacing conventional HP-ADCs and alleviating the above issues. Our proposed framework incurs minimal accuracy degradation by performing hardware-aware training and is able to scale beyond simple image classification tasks to more complex sequential regression tasks. Experiments on complex tasks of optical flow estimation and gesture recognition show that progressively increasing the hardware awareness during SNN training allows the model to adapt and learn the errors due to the non-idealities associated with ADC-Less IMC. Also, the proposed ADC-Less IMC offers significant energy and latency improvements, $2-7\times$ and $8.9-24.6\times$, respectively, depending on the SNN model and the workload, compared to HP-ADC IMC. | ['Kaushik Roy', 'Utkarsh Saxena', 'Adarsh Kumar Kosta', 'Marco Paul E. Apolinario'] | 2022-11-03 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 3.19334745e-01 -3.76550466e-01 -6.77581578e-02 1.15182422e-01
5.88996485e-02 -3.36792290e-01 2.98966706e-01 2.54339844e-01
-9.86464024e-01 5.46975970e-01 -4.32856858e-01 -3.61187071e-01
1.54958963e-01 -7.83841491e-01 -7.45813727e-01 -6.26178920e-01
-6.16458356e-02 -1.46949872e-01 5.48035324e-01 1.15671247e-01
2.51127541e-01 5.64186215e-01 -1.63755691e+00 -1.06811546e-01
8.10776353e-01 1.40090215e+00 2.28745922e-01 7.80083656e-01
1.42938606e-02 6.90284848e-01 -5.34521163e-01 -1.91153750e-01
1.84313521e-01 -3.55363309e-01 5.61237857e-02 -7.60552585e-01
1.90447867e-01 -4.33149546e-01 -5.78199446e-01 8.80073845e-01
5.95003247e-01 1.33120790e-02 4.63298142e-01 -1.26519358e+00
-9.61419120e-02 2.44689614e-01 -3.82746518e-01 5.27028799e-01
-5.98404668e-02 3.77177894e-01 2.14842275e-01 -5.12400627e-01
3.05973679e-01 6.69006348e-01 8.79496574e-01 5.52199006e-01
-9.52921510e-01 -8.95672560e-01 -2.05383599e-01 5.50332785e-01
-1.40076995e+00 -5.94191730e-01 5.54598927e-01 -5.81973419e-02
1.79769468e+00 1.41038924e-01 1.07830024e+00 9.07138944e-01
6.59779191e-01 4.99276608e-01 1.03441405e+00 -1.18613988e-01
8.90809774e-01 -1.45309865e-01 2.17192490e-02 6.76274359e-01
7.85934269e-01 -1.01898168e-03 -8.01721096e-01 3.65216732e-02
1.00033069e+00 -4.48171273e-02 -3.32317650e-01 2.47634038e-01
-8.86499584e-01 2.63843536e-01 7.32661784e-01 4.41095345e-02
-5.89243114e-01 9.22365785e-01 6.84066057e-01 -1.18763268e-01
-2.34486550e-01 4.02091444e-01 -2.23400116e-01 -6.12322032e-01
-1.13275027e+00 -1.79381683e-01 6.63850129e-01 9.15975630e-01
3.72467399e-01 6.03831410e-01 1.95650846e-01 -3.23522501e-02
1.77340060e-01 5.94309926e-01 7.81512082e-01 -8.75716329e-01
1.35079101e-01 6.85924828e-01 1.23644002e-01 -7.26504326e-01
-7.82445431e-01 -5.39939106e-01 -1.10452664e+00 3.49990934e-01
3.58701140e-01 -6.91096485e-02 -9.29645121e-01 1.51648462e+00
2.99451649e-02 4.37896937e-01 7.36797601e-02 9.28647041e-01
5.27118683e-01 7.02038705e-01 2.53790051e-01 -2.62352616e-01
1.50750959e+00 -7.29380012e-01 -8.02873671e-01 -4.85498488e-01
2.14711517e-01 -2.79742599e-01 9.32811737e-01 1.74774587e-01
-1.02557492e+00 -5.02260447e-01 -1.68457401e+00 -2.65664846e-01
-2.32605889e-01 3.71298864e-02 7.74238288e-01 9.92340207e-01
-1.13335681e+00 6.50009453e-01 -1.59813261e+00 -1.25840962e-01
5.32399535e-01 7.77733028e-01 2.60979801e-01 4.13701534e-01
-7.38758445e-01 7.00293124e-01 3.10176760e-01 2.02267691e-01
-5.57673037e-01 -8.25371802e-01 -3.60965073e-01 3.59887868e-01
2.76245233e-02 -7.56561875e-01 8.56324196e-01 -9.64818776e-01
-1.90668964e+00 1.12583481e-01 -1.66879982e-01 -9.40641761e-01
1.92091584e-01 7.40867555e-02 -5.50331295e-01 3.23185980e-01
-6.55502319e-01 7.64927328e-01 7.37123311e-01 -3.48749518e-01
-2.81441242e-01 -3.11650813e-01 -1.02585748e-01 -8.10856093e-03
-6.24398112e-01 -4.35882777e-01 -1.40503347e-01 -3.69940221e-01
1.55362889e-01 -1.08148217e+00 -3.01354110e-01 5.12092173e-01
1.63518444e-01 2.73506045e-01 1.03376961e+00 -1.64552167e-01
1.24770558e+00 -1.96900439e+00 -4.63907182e-01 -3.46584409e-03
9.28031728e-02 6.92534089e-01 2.90410221e-01 -2.16221735e-01
5.79186797e-01 -3.27757776e-01 1.26704797e-02 -1.56176733e-02
-3.67891371e-01 1.11272655e-01 1.06407851e-01 4.74696279e-01
1.89434469e-01 9.10082817e-01 -5.19518018e-01 -1.33525178e-01
2.22625762e-01 8.01021159e-01 -5.17909050e-01 -2.91513026e-01
9.29138958e-02 2.87230313e-01 -1.90652415e-01 5.65128446e-01
4.30581748e-01 -4.32424545e-01 3.09666961e-01 -5.03706574e-01
-2.32463732e-01 2.94850409e-01 -1.11432612e+00 1.61370099e+00
-5.97503066e-01 1.01407158e+00 2.71824181e-01 -7.21907258e-01
8.41138184e-01 2.45941579e-01 5.13501503e-02 -1.14548802e+00
5.80727756e-01 6.27696991e-01 2.11116791e-01 -1.19378388e-01
5.55769145e-01 1.30609021e-01 1.41017392e-01 1.48351982e-01
-1.69937015e-02 2.41167694e-01 -1.54345647e-01 -1.38989121e-01
1.17095160e+00 1.61152288e-01 2.11512700e-01 -5.83129168e-01
3.28701019e-01 5.31976623e-03 6.20927632e-01 4.40018207e-01
-4.82877433e-01 -5.12759611e-02 6.16176948e-02 -5.32097399e-01
-9.58149076e-01 -9.23316360e-01 -1.92723274e-01 5.07673085e-01
5.04912078e-01 -3.45762342e-01 -8.80493760e-01 2.40293458e-01
-3.38314116e-01 4.66821522e-01 1.85913616e-03 -3.20612282e-01
-6.62644744e-01 -8.32335174e-01 9.06216502e-01 9.38472629e-01
9.78333354e-01 -4.95343536e-01 -2.15955591e+00 5.33327162e-01
3.80538642e-01 -1.30967343e+00 -3.37233782e-01 6.09172046e-01
-1.32677627e+00 -5.13158023e-01 -2.88460225e-01 -5.32318830e-01
6.97652996e-01 5.59454598e-02 7.11440206e-01 -6.65531307e-02
-5.66505611e-01 -3.92808057e-02 1.76807135e-01 -4.30868536e-01
4.71185781e-02 -1.64774567e-01 2.51651615e-01 -2.41276503e-01
4.34529632e-01 -7.75781631e-01 -1.30139077e+00 -5.54220611e-03
-6.29643142e-01 3.32497388e-01 5.95833421e-01 7.05860496e-01
7.00784922e-01 -1.75059184e-01 5.55793881e-01 -4.40121830e-01
1.71645343e-01 -1.49196386e-02 -9.53419864e-01 -2.58637369e-02
-1.03041387e+00 3.68442833e-01 1.16274035e+00 -7.12544620e-01
-7.73210704e-01 4.97713655e-01 -1.25210211e-02 -3.36390197e-01
3.78880054e-01 1.03431533e-03 1.45941183e-01 -5.56659997e-01
7.07240224e-01 2.46885821e-01 -1.06997102e-01 1.11891076e-01
-1.79038242e-01 4.95855391e-01 7.92369485e-01 -1.86006591e-01
5.12832776e-02 4.95507717e-01 5.09377837e-01 -7.25964665e-01
3.19873333e-01 8.35898742e-02 3.65036838e-02 -3.59575987e-01
7.51554132e-01 -1.17656684e+00 -1.19852233e+00 5.52634597e-01
-8.99963260e-01 -3.29115748e-01 6.69303611e-02 6.42752528e-01
-2.41185263e-01 -1.62229892e-02 -8.25527847e-01 -8.86187255e-01
-1.07012367e+00 -1.41527045e+00 4.89974976e-01 1.09407461e+00
-3.12482148e-01 -6.16043627e-01 -5.59145451e-01 1.55542351e-04
1.03577864e+00 2.27094904e-01 8.99552226e-01 -6.57873005e-02
-8.79054844e-01 -1.75709188e-01 -3.01413774e-01 -7.59910718e-02
-3.59442800e-01 -1.71540782e-01 -1.17427003e+00 -4.46964324e-01
-5.42668477e-02 -1.08419314e-01 6.37796462e-01 6.01233602e-01
1.04189336e+00 -2.19593257e-01 -4.90789264e-01 8.84050190e-01
1.91453183e+00 6.10514939e-01 6.23313367e-01 1.53206393e-01
4.29879159e-01 -3.20841819e-01 -7.79210627e-02 5.39693773e-01
3.17435384e-01 5.60917616e-01 5.41972518e-01 1.37597129e-01
-9.98840034e-02 -1.48282379e-01 3.62189591e-01 8.60933065e-01
-1.97918460e-01 -2.95649230e-01 -7.87395477e-01 3.33236784e-01
-1.45969200e+00 -4.25619423e-01 -3.24545890e-01 2.34363294e+00
6.56410038e-01 4.55250412e-01 -1.85815006e-01 2.58237541e-01
4.26300555e-01 -2.09840879e-01 -1.05300975e+00 -7.57748961e-01
2.14184951e-02 7.99185693e-01 1.14988923e+00 6.59071654e-02
-5.78428626e-01 4.63992864e-01 5.20935440e+00 6.80732846e-01
-1.66384697e+00 2.35680014e-01 3.87414068e-01 -5.22452116e-01
6.60296157e-02 -3.94051187e-02 -7.23446906e-01 8.18215013e-01
1.58825886e+00 -1.00744493e-01 4.82963979e-01 9.41355705e-01
3.59011777e-02 -5.90086579e-01 -1.14073861e+00 1.33711267e+00
-2.90872723e-01 -1.52723384e+00 -2.32597142e-01 -1.82542771e-01
5.26102006e-01 -5.06784543e-02 -1.87861957e-02 6.15113880e-03
-4.38300163e-01 -8.60661328e-01 8.08114111e-01 1.80378705e-01
1.09366417e+00 -7.32414007e-01 7.05361962e-01 2.96286106e-01
-1.32639956e+00 -9.34713259e-02 -3.92317295e-01 -4.86203671e-01
2.25299690e-02 4.79363769e-01 -5.99125445e-01 -7.42430389e-02
8.94963861e-01 1.47989094e-01 -3.67637694e-01 8.14966559e-01
2.70527422e-01 4.10548747e-01 -7.71304548e-01 -8.04149985e-01
1.26999423e-01 -6.33688718e-02 2.21438736e-01 1.24709582e+00
5.39205730e-01 3.43111724e-01 -4.80857134e-01 7.16114044e-01
-1.58418104e-01 -3.36034000e-01 -8.69265348e-02 -7.10758418e-02
9.20323551e-01 1.25573301e+00 -1.19496715e+00 -3.68832916e-01
-4.03625995e-01 1.11273634e+00 -5.95263354e-02 -3.43900435e-02
-1.02008092e+00 -8.43972743e-01 6.08022749e-01 2.63519853e-01
8.31078812e-02 -4.29032505e-01 -7.51773298e-01 -8.09468806e-01
1.63153127e-01 -3.26179981e-01 -2.00977296e-01 -5.43486714e-01
-3.00127685e-01 7.27588475e-01 -6.19533360e-01 -1.06905937e+00
-3.08296811e-02 -7.88428664e-01 -4.40836310e-01 5.73908865e-01
-1.44048047e+00 -5.32051921e-01 -7.43235648e-01 4.55786526e-01
3.11524063e-01 8.88347551e-02 9.01295304e-01 4.59306240e-01
-6.70747936e-01 8.52919996e-01 1.04977250e-01 -1.80539265e-01
2.04152152e-01 -6.31960392e-01 5.13124049e-01 1.03628528e+00
-2.11490571e-01 5.82666039e-01 5.95419884e-01 -3.92416149e-01
-2.20821071e+00 -9.36522007e-01 3.20310235e-01 3.38660538e-01
3.55311871e-01 -4.17036861e-01 -8.00734639e-01 1.60857528e-01
4.13959920e-02 4.01863575e-01 5.12896359e-01 -7.20877230e-01
-9.16081741e-02 -4.41955090e-01 -1.33287668e+00 7.18859673e-01
1.07450581e+00 -4.37827229e-01 1.94782540e-01 -1.98295623e-01
2.21691549e-01 -5.60390949e-01 -5.89486599e-01 2.35702053e-01
8.77839386e-01 -7.84723103e-01 7.47521758e-01 2.41677850e-01
9.22471285e-02 -4.52657759e-01 -7.94309378e-02 -7.46996760e-01
-5.53033948e-02 -6.43174887e-01 -4.62191403e-01 6.13134623e-01
2.34834805e-01 -7.26791918e-01 1.10041940e+00 7.08742917e-01
-1.68923810e-01 -8.02035749e-01 -1.36928296e+00 -7.63380647e-01
-4.10559565e-01 -3.86148304e-01 2.98718154e-01 2.64827818e-01
1.47439882e-01 1.55510589e-01 6.44981638e-02 2.07763031e-01
5.43857932e-01 -2.82844216e-01 1.50685847e-01 -8.85125339e-01
-4.69869196e-01 -5.50472081e-01 -8.18578959e-01 -1.08966148e+00
-4.11684662e-01 -5.22977114e-01 1.27584249e-01 -1.02752650e+00
-2.02280477e-01 -5.24928272e-01 -2.01133445e-01 2.51929343e-01
-3.74827185e-03 4.25419897e-01 1.58011943e-01 1.19589910e-01
-4.97996032e-01 1.97703913e-01 6.23022258e-01 1.30550459e-01
-3.28574061e-01 -3.37418079e-01 -9.05220509e-02 6.11571789e-01
5.91535151e-01 -3.79850745e-01 -5.70109427e-01 -6.77943230e-01
2.43002430e-01 1.77779168e-01 3.76598358e-01 -1.80835533e+00
9.50121701e-01 1.80736586e-01 7.47417152e-01 -2.55534470e-01
4.99570638e-01 -9.48067963e-01 5.99712372e-01 1.14597654e+00
1.67290464e-01 3.26172233e-01 6.50186121e-01 5.22767842e-01
1.42682329e-01 -8.57514814e-02 9.68990207e-01 2.07722872e-01
-1.01233566e+00 -4.84853722e-02 -7.62133300e-01 -2.36757919e-01
1.04307294e+00 -6.61722958e-01 -5.55740476e-01 -2.60077626e-03
-1.48936167e-01 -1.50385946e-01 4.56310362e-01 -1.17598481e-01
6.13204658e-01 -9.91185486e-01 -2.46904194e-02 5.13839722e-01
-3.34067672e-01 -1.44013241e-01 5.08171797e-01 8.49967539e-01
-1.21383345e+00 5.69242775e-01 -7.49827862e-01 -6.31157637e-01
-8.69388580e-01 3.01636606e-01 2.52490014e-01 -7.63650751e-03
-6.02525055e-01 8.61221135e-01 -4.82989341e-01 5.61544836e-01
2.76142359e-01 -5.32937825e-01 3.69832188e-01 -3.66908371e-01
5.33548772e-01 7.35873044e-01 3.13177854e-01 8.62216726e-02
-7.25362420e-01 5.30063272e-01 2.94374943e-01 9.83434319e-02
1.18899310e+00 8.84480327e-02 1.27305225e-01 1.01599798e-01
9.75115418e-01 -5.18258572e-01 -1.59564579e+00 1.41528875e-01
5.18552400e-02 9.60848257e-02 5.31198978e-01 -7.78223634e-01
-1.20566905e+00 8.31746399e-01 1.22253191e+00 -5.09270430e-01
1.74496317e+00 -6.55334353e-01 1.15450096e+00 3.22232872e-01
6.55459523e-01 -1.25425553e+00 -9.80926603e-02 1.86450154e-01
-4.19116160e-03 -7.85601318e-01 7.24902824e-02 -1.48853973e-01
-1.53521359e-01 1.34725058e+00 7.67659247e-01 -3.88441980e-01
5.21541893e-01 1.06036532e+00 -4.31015104e-01 2.48310804e-01
-7.31843293e-01 2.83497572e-01 -1.97924763e-01 4.55469698e-01
1.80748969e-01 1.14782862e-01 -4.38024044e-01 7.46976256e-01
4.19829264e-02 5.63673019e-01 7.24382222e-01 1.12567306e+00
-2.48750433e-01 -5.07147431e-01 2.39200108e-02 6.04263246e-01
-4.04037476e-01 -8.35032910e-02 2.58675665e-01 4.51158166e-01
2.45841265e-01 7.24884152e-01 5.91168225e-01 -5.95904827e-01
1.50073633e-01 9.48898681e-03 5.97938240e-01 1.59032091e-01
-8.25497389e-01 -7.41477758e-02 -1.92126989e-01 -6.31922007e-01
-1.88047349e-01 -2.23605976e-01 -1.74719346e+00 -6.15256310e-01
-2.69913197e-01 -3.41433376e-01 1.30643702e+00 6.77579403e-01
8.66421580e-01 7.77361155e-01 1.02670930e-01 -7.60136068e-01
-1.88367769e-01 -5.53795040e-01 -2.12067321e-01 -2.96607494e-01
3.17562371e-01 -5.35349250e-01 -1.92140147e-01 -1.37220487e-01] | [8.29780101776123, 2.485051155090332] |
a259a3dc-70ae-45b4-9b53-6e21e757889b | expanding-the-text-classification-toolbox | 1903.09878 | null | http://arxiv.org/abs/1903.09878v2 | http://arxiv.org/pdf/1903.09878v2.pdf | Expanding the Text Classification Toolbox with Cross-Lingual Embeddings | Most work in text classification and Natural Language Processing (NLP)
focuses on English or a handful of other languages that have text corpora of
hundreds of millions of words. This is creating a new version of the digital
divide: the artificial intelligence (AI) divide. Transfer-based approaches,
such as Cross-Lingual Text Classification (CLTC) - the task of categorizing
texts written in different languages into a common taxonomy, are a promising
solution to the emerging AI divide. Recent work on CLTC has focused on
demonstrating the benefits of using bilingual word embeddings as features,
relegating the CLTC problem to a mere benchmark based on a simple averaged
perceptron.
In this paper, we explore more extensively and systematically two flavors of
the CLTC problem: news topic classification and textual churn intent detection
(TCID) in social media. In particular, we test the hypothesis that embeddings
with context are more effective, by multi-tasking the learning of multilingual
word embeddings and text classification; we explore neural architectures for
CLTC; and we move from bi- to multi-lingual word embeddings. For all
architectures, types of word embeddings and datasets, we notice a consistent
gain trend in favor of multilingual joint training, especially for
low-resourced languages. | ["Meryem M'hamdi", 'Michael Baeriswyl', 'Robert West', 'Claudiu Musat', 'Andreea Hossmann'] | 2019-03-23 | null | null | null | null | ['multilingual-word-embeddings'] | ['methodology'] | [-1.97706386e-01 -2.03259513e-01 -6.76989555e-01 -2.53828406e-01
-7.42778480e-01 -5.58291674e-01 1.03545129e+00 6.37187898e-01
-8.26982737e-01 3.47542882e-01 8.64957392e-01 -7.02487648e-01
8.81546065e-02 -6.37336552e-01 -1.96563259e-01 -4.51596528e-01
1.03966407e-01 9.09245610e-01 -3.13102394e-01 -3.89846444e-01
3.85313392e-01 -6.12383932e-02 -1.45955789e+00 4.81438279e-01
9.66587067e-01 8.37092042e-01 -1.18002310e-01 6.18621290e-01
-5.75123787e-01 7.01148927e-01 -4.40612793e-01 -3.99930179e-01
-2.71925814e-02 -2.93070108e-01 -1.06669831e+00 -2.15268195e-01
4.19420719e-01 1.26609117e-01 -2.71530412e-02 8.10866535e-01
4.84857708e-01 -2.17177898e-01 8.11586261e-01 -1.08834350e+00
-1.03843844e+00 9.59055722e-01 -5.92716396e-01 4.19309795e-01
3.33522588e-01 -3.01500440e-01 1.83037412e+00 -8.06692660e-01
8.42685580e-01 1.35066640e+00 7.14859664e-01 9.05590504e-02
-1.26788628e+00 -4.54623550e-01 -3.36046377e-03 2.58087009e-01
-1.15759325e+00 -8.69038031e-02 5.35437703e-01 -9.42006290e-01
1.35124862e+00 5.30773252e-02 3.74609083e-01 1.38857937e+00
4.40857708e-01 8.68691504e-01 1.16186714e+00 -8.56883466e-01
-4.31936979e-02 6.18868053e-01 6.49061799e-01 3.50631416e-01
4.87638772e-01 -7.03301430e-02 -5.37681162e-01 -3.10743302e-01
-9.19017941e-02 -1.57823578e-01 3.39752936e-04 -2.03508556e-01
-1.21144962e+00 1.60991848e+00 1.04520291e-01 9.58423615e-01
-1.49867898e-02 -2.26004317e-01 1.09210134e+00 6.99858427e-01
1.11455011e+00 9.18627381e-01 -6.40869796e-01 -1.31050855e-01
-7.52738118e-01 1.34401307e-01 1.10585892e+00 7.07010448e-01
6.46184981e-01 -2.48578086e-01 6.12541027e-02 1.08386540e+00
2.67409563e-01 1.76877931e-01 1.32505894e+00 -1.11363128e-01
6.53292000e-01 7.18293011e-01 -3.78415436e-01 -1.04014981e+00
-5.19825161e-01 -3.19955915e-01 -4.45216149e-01 -3.23675722e-01
3.07190061e-01 -2.14483663e-01 -5.34298897e-01 1.27166009e+00
-5.36196604e-02 -3.40906233e-01 2.65464753e-01 6.19086385e-01
5.14794767e-01 9.15438771e-01 1.13445416e-01 -1.16102681e-01
1.61292744e+00 -1.07937813e+00 -6.38561249e-01 -5.29651999e-01
1.13959014e+00 -8.18490326e-01 1.18004680e+00 2.45181620e-01
-3.30498457e-01 -3.68733197e-01 -9.62704420e-01 -5.19723594e-01
-1.18453312e+00 -3.40874456e-02 4.71881807e-01 7.13813245e-01
-7.77813435e-01 3.93734902e-01 -3.44859868e-01 -7.67654240e-01
2.14183182e-01 -9.22332779e-02 -2.74805754e-01 -2.99423337e-01
-1.36255825e+00 1.33482492e+00 5.19911647e-01 -6.87348485e-01
-2.61790484e-01 -6.85989439e-01 -8.90128195e-01 -1.45036653e-01
-6.25141189e-02 -1.64276958e-02 8.97484243e-01 -1.08471608e+00
-9.86150801e-01 1.28941596e+00 -1.26982052e-02 -4.71693426e-01
1.77263603e-01 -3.49740744e-01 -6.29689515e-01 -3.33727688e-01
4.07420903e-01 4.76581067e-01 5.76048315e-01 -1.03639638e+00
-7.92690277e-01 -4.90990341e-01 -6.37375355e-01 1.59047350e-01
-9.16234076e-01 4.38806832e-01 1.98932528e-01 -6.86198056e-01
-4.33316022e-01 -8.00985336e-01 2.22650662e-01 -5.77982545e-01
-7.50021711e-02 -9.51093435e-01 9.09421802e-01 -8.11629415e-01
1.44342780e+00 -2.12392735e+00 2.18096197e-01 -1.31465867e-01
2.73989975e-01 2.15007797e-01 -1.92804888e-01 8.12664211e-01
1.07034653e-01 4.38503832e-01 6.48609474e-02 -3.33813697e-01
2.12767467e-01 3.82637978e-01 -3.82973760e-01 5.45580268e-01
2.19106093e-01 1.00241566e+00 -9.18114781e-01 -4.65892136e-01
-2.08317354e-01 2.50292540e-01 -6.00340188e-01 -8.61929059e-02
-2.00664461e-01 -1.06606476e-01 -2.86387116e-01 4.89331305e-01
2.28117898e-01 -2.50594243e-02 4.92875248e-01 1.91276118e-01
-5.78237951e-01 6.45896375e-01 -4.22464699e-01 1.49565399e+00
-6.88969910e-01 1.28330243e+00 -3.65428597e-01 -1.16214085e+00
9.69383121e-01 4.09139812e-01 4.48440969e-01 -5.78041852e-01
5.48147202e-01 5.69679260e-01 2.32615456e-01 -6.28336966e-01
7.76510537e-01 -2.48088747e-01 -5.06697178e-01 8.01388979e-01
4.92059529e-01 -2.74149757e-02 2.41047606e-01 3.96055356e-02
9.00631845e-01 -2.18941316e-01 5.68138421e-01 -7.17236042e-01
6.41097054e-02 4.89755809e-01 3.04655340e-02 6.15712583e-01
-2.41732970e-01 1.76002011e-01 5.40905416e-01 -4.34410870e-01
-1.41570449e+00 -6.07179880e-01 -6.17830038e-01 1.80308306e+00
-3.11332047e-01 -5.00864744e-01 -2.70972431e-01 -7.07284153e-01
4.07256365e-01 8.51456523e-01 -6.63408518e-01 3.56459394e-02
-4.79414910e-01 -9.43651795e-01 5.77984154e-01 3.19626987e-01
-1.07937112e-01 -9.91911292e-01 -5.21997511e-01 1.87330991e-01
-9.19803828e-02 -1.15694892e+00 -4.82213736e-01 7.02207386e-01
-6.68139577e-01 -7.65253484e-01 -5.14657557e-01 -1.26993036e+00
8.59038830e-02 9.85440835e-02 1.27089691e+00 -2.76912361e-01
-4.37787324e-01 2.62027830e-01 -9.70110416e-01 -6.30620301e-01
-6.85469151e-01 5.32721341e-01 3.00266922e-01 -1.19554952e-01
1.09129345e+00 -3.12187850e-01 1.75215658e-02 -1.57003880e-01
-9.53633130e-01 -2.72713572e-01 6.04798138e-01 1.03153741e+00
-1.90566152e-01 -1.30583882e-01 6.58043444e-01 -1.22304237e+00
1.13259065e+00 -1.01970875e+00 -3.93748470e-02 1.42719537e-01
-9.33461487e-01 1.59882858e-01 7.10196793e-01 -5.33767164e-01
-6.40550494e-01 -5.40257156e-01 -1.01592720e-01 -4.78001609e-02
-1.46894529e-01 1.01290667e+00 5.24622858e-01 4.25332367e-01
9.58862543e-01 7.02954009e-02 -4.77988087e-02 -5.41777015e-01
6.83704317e-01 1.37241983e+00 -2.67223418e-02 -5.15266359e-01
5.74625075e-01 1.58272758e-01 -9.25419152e-01 -1.16011214e+00
-1.01703262e+00 -1.00808001e+00 -8.74126792e-01 5.34496754e-02
1.14834976e+00 -8.40271831e-01 -9.71314311e-02 3.20258620e-03
-1.19391406e+00 -7.52805546e-02 -2.15162113e-01 6.43293262e-01
-1.74418062e-01 9.56490189e-02 -7.96553493e-01 -5.83138347e-01
-2.49615282e-01 -8.60663772e-01 9.71567452e-01 -2.77241200e-01
-5.67659795e-01 -1.44190133e+00 6.83562517e-01 3.05299431e-01
4.88676369e-01 6.39438257e-03 1.40137362e+00 -1.39018929e+00
4.44059670e-01 -4.46463734e-01 -2.74184048e-01 5.07187068e-01
2.66594216e-02 -1.58694014e-01 -1.08097863e+00 -4.13663238e-01
-9.99413431e-02 -7.98069179e-01 1.02291489e+00 1.21723488e-01
5.07115722e-01 -4.04069573e-01 -4.38364297e-01 1.71019156e-02
1.62860703e+00 1.57239556e-01 1.94233447e-01 5.06441355e-01
6.37971401e-01 9.00962830e-01 2.49179125e-01 2.52305329e-01
4.40819293e-01 5.07778168e-01 -1.51057571e-01 1.71370134e-01
1.18648507e-01 -2.25679651e-01 5.14130533e-01 1.68075299e+00
2.42944777e-01 -3.14550847e-01 -1.32376564e+00 7.49392629e-01
-1.65942132e+00 -7.36771286e-01 -6.85150996e-02 1.74013829e+00
1.04207981e+00 -2.94169970e-02 1.07173182e-01 9.34454575e-02
4.91297066e-01 3.17721337e-01 -8.63617733e-02 -9.77112293e-01
-1.40730850e-02 2.26445451e-01 4.32595134e-01 4.36111510e-01
-1.19105899e+00 1.01108658e+00 6.18307877e+00 8.65093410e-01
-1.16833389e+00 5.42179585e-01 5.12122691e-01 3.38512361e-01
-4.08769310e-01 -1.66394144e-01 -9.12649214e-01 4.34507757e-01
1.21520567e+00 -3.56209964e-01 2.30687857e-01 9.96683776e-01
-3.31751317e-01 -5.02272956e-02 -1.30668318e+00 7.92992353e-01
6.88917220e-01 -1.25274539e+00 1.29440218e-01 1.37851208e-01
8.25264037e-01 6.44680262e-01 4.82092537e-02 7.77181447e-01
4.81027782e-01 -1.03706503e+00 5.36227822e-01 -2.59524524e-01
6.62447691e-01 -4.66029376e-01 1.06430256e+00 2.58699626e-01
-8.41696084e-01 -2.79614449e-01 -5.14001191e-01 -7.45520890e-02
-7.02254176e-02 4.36424792e-01 -7.57558882e-01 4.76697773e-01
7.31223524e-01 1.04609954e+00 -6.77282512e-01 5.58332562e-01
2.73485273e-01 6.50517166e-01 -3.57107446e-02 -5.23462355e-01
5.86117327e-01 -2.30422154e-01 3.62259805e-01 1.75964856e+00
1.29307479e-01 -4.52044070e-01 2.01777086e-01 5.46498239e-01
-2.23867103e-01 6.27653003e-01 -9.57174838e-01 -6.35055184e-01
2.94589967e-01 1.01709199e+00 -5.41267693e-01 -2.32693315e-01
-7.66347110e-01 8.13958168e-01 6.18777275e-01 -8.08769166e-02
-5.94921708e-01 -5.68751395e-01 6.08794332e-01 -1.91263452e-01
1.67979762e-01 -3.26901883e-01 -3.92595142e-01 -1.28967571e+00
-2.47535497e-01 -9.22817230e-01 4.32969689e-01 -1.28002226e-01
-1.93516934e+00 8.45462680e-01 -1.17452689e-01 -1.06029415e+00
-2.59392142e-01 -9.65486407e-01 -4.82990116e-01 4.52058882e-01
-1.62470806e+00 -1.08656049e+00 1.88909277e-01 2.78962791e-01
8.69282842e-01 -4.60362762e-01 9.71609294e-01 5.71146905e-01
-3.76049012e-01 5.08805096e-01 6.31832838e-01 3.49499017e-01
9.14955854e-01 -1.25950801e+00 1.86336368e-01 2.11810589e-01
4.54882056e-01 2.60957479e-01 6.03553772e-01 -3.88201565e-01
-1.13791811e+00 -9.78252769e-01 1.64458776e+00 -8.76780689e-01
1.34591889e+00 -7.90440142e-01 -7.93428659e-01 7.67023146e-01
7.10814834e-01 -4.22591418e-01 1.06660390e+00 6.61439419e-01
-7.30609477e-01 2.37358645e-01 -5.77082217e-01 3.70595485e-01
4.50768113e-01 -7.87948966e-01 -9.37974393e-01 6.26325667e-01
1.09088254e+00 3.24223369e-01 -9.37993109e-01 -2.22892925e-01
5.47710240e-01 -4.74931985e-01 4.68765378e-01 -9.67883348e-01
8.56216609e-01 3.25881183e-01 -1.72836542e-01 -1.57000661e+00
-2.57453471e-01 -7.06629455e-02 4.47713017e-01 1.19606507e+00
6.77597642e-01 -8.47019315e-01 2.54268974e-01 -5.23356274e-02
-2.43173882e-01 -5.93840241e-01 -1.10143971e+00 -7.31393278e-01
8.53185415e-01 -5.38422227e-01 -9.36656073e-02 1.71504056e+00
7.31626153e-01 9.87479150e-01 -2.35104665e-01 -4.15360898e-01
9.09342766e-02 6.13977723e-02 4.16897207e-01 -1.45184600e+00
9.53268446e-03 -7.54201591e-01 -4.42601472e-01 -6.57828808e-01
4.08621371e-01 -1.40376627e+00 1.41422659e-01 -1.37290859e+00
2.51195103e-01 -4.41638827e-01 -6.92206845e-02 2.02189505e-01
1.23420276e-01 1.58089980e-01 6.29838835e-03 3.34130377e-01
-4.73086923e-01 3.60037267e-01 5.91800332e-01 -3.53221923e-01
-8.09985548e-02 -5.08973598e-01 -6.61924422e-01 4.01572317e-01
7.99281299e-01 -7.08687663e-01 6.56188503e-02 -5.34595370e-01
4.49998379e-01 -4.87874359e-01 -1.93638891e-01 -6.89231455e-01
6.77413046e-02 1.10760152e-01 -6.66991323e-02 -1.91014335e-01
-3.52984406e-02 -7.04720199e-01 -7.19283342e-01 3.88734818e-01
-7.86277831e-01 1.72515541e-01 1.42807603e-01 6.40593469e-01
-4.69120800e-01 -4.18733925e-01 6.17734015e-01 4.39184532e-02
-3.71087193e-01 -1.29245415e-01 -9.53328907e-01 4.65396553e-01
8.80500853e-01 3.66821438e-02 -5.43993711e-01 -8.41898844e-02
-5.46474338e-01 2.51188517e-01 8.64485875e-02 9.56029832e-01
6.39888421e-02 -1.20492256e+00 -9.92121518e-01 1.45408750e-01
5.63412070e-01 -6.20594323e-01 -3.54538232e-01 8.28910530e-01
-4.32101905e-01 1.03045976e+00 -8.34401324e-02 -3.54662657e-01
-9.75192010e-01 6.59850657e-01 7.59319682e-03 -4.97868568e-01
-3.47629219e-01 6.67314470e-01 -1.76071852e-01 -7.98923135e-01
1.09336525e-01 -3.89245391e-01 -5.04011989e-01 7.79576838e-01
2.00587168e-01 2.38113105e-01 1.79442525e-01 -7.38433719e-01
-2.30089515e-01 3.04873794e-01 -2.63077408e-01 -4.31907997e-02
1.45883632e+00 -3.15771401e-02 -3.86455625e-01 1.07449889e+00
1.88417649e+00 -3.61564406e-03 -2.17463642e-01 -3.97325099e-01
6.60290003e-01 -1.35170519e-01 1.81988358e-01 -5.59233963e-01
-4.78559464e-01 1.01140976e+00 5.48113823e-01 7.51935065e-01
3.52358133e-01 3.43000829e-01 8.38260889e-01 3.74089777e-01
4.18235436e-02 -1.56062925e+00 1.80041209e-01 9.03229296e-01
6.67448103e-01 -1.42493546e+00 3.84219326e-02 2.62207329e-01
-9.47447777e-01 1.18801260e+00 3.07863861e-01 -2.68692195e-01
9.46192563e-01 5.55176474e-02 1.68314949e-01 -3.13114077e-01
-7.97820687e-01 -2.74255514e-01 3.29072773e-01 3.66384119e-01
9.78824079e-01 1.59813270e-01 -6.80651307e-01 4.76831079e-01
-3.39343995e-01 -5.24500906e-01 4.27081734e-01 9.55790818e-01
-6.57509029e-01 -1.13609064e+00 -1.72402665e-01 8.74816895e-01
-5.11505842e-01 -3.90819669e-01 -5.15903652e-01 8.75483215e-01
1.82329774e-01 9.10714686e-01 5.05121291e-01 -4.77507174e-01
-1.21946804e-01 6.86042547e-01 -9.74351466e-02 -1.01536870e+00
-7.93222725e-01 -5.36157750e-02 2.84511149e-01 -1.12788249e-02
-3.75790119e-01 -7.77247608e-01 -5.78985155e-01 -3.02343637e-01
-4.61218178e-01 4.33415562e-01 7.56337285e-01 1.07646203e+00
3.05883378e-01 2.20150143e-01 5.07416308e-01 -7.83291519e-01
-5.18427253e-01 -1.34027779e+00 -6.81820035e-01 3.45356256e-01
2.38318995e-01 -3.99489850e-01 -5.81976235e-01 -7.56243840e-02] | [10.54782485961914, 9.639232635498047] |
957da8b1-7fbc-4dc5-92d3-e320b6216a3b | multi-instance-learning-for-end-to-end | 1903.02652 | null | http://arxiv.org/abs/1903.02652v1 | http://arxiv.org/pdf/1903.02652v1.pdf | Multi-Instance Learning for End-to-End Knowledge Base Question Answering | End-to-end training has been a popular approach for knowledge base question
answering (KBQA). However, real world applications often contain answers of
varied quality for users' questions. It is not appropriate to treat all
available answers of a user question equally.
This paper proposes a novel approach based on multiple instance learning to
address the problem of noisy answers by exploring consensus among answers to
the same question in training end-to-end KBQA models. In particular, the QA
pairs are organized into bags with dynamic instance selection and different
options of instance weighting. Curriculum learning is utilized to select
instance bags during training. On the public CQA dataset, the new method
significantly improves both entity accuracy and the Rouge-L score over a
state-of-the-art end-to-end KBQA baseline. | ['Qiong Zhang', 'Yifan He', 'Luo Si', 'Mengxi Wei'] | 2019-03-06 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-3.93977553e-01 8.31749290e-02 1.09189734e-01 -6.37520850e-01
-1.81673133e+00 -7.36296415e-01 9.71157625e-02 5.38376570e-01
-7.58110166e-01 1.09444189e+00 3.40596139e-01 -3.04614931e-01
-1.52381599e-01 -1.03020537e+00 -7.09722877e-01 -3.27262312e-01
5.43883443e-01 1.10671151e+00 8.00073028e-01 -7.23528445e-01
7.40527287e-02 -1.22933924e-01 -1.23562157e+00 1.00786126e+00
1.47687066e+00 9.71390128e-01 1.22071497e-01 7.13010669e-01
-7.26504624e-01 9.63445723e-01 -7.53128469e-01 -1.17073786e+00
-9.59459096e-02 -3.16284865e-01 -1.29746711e+00 -3.68458271e-01
8.62800062e-01 -1.41524926e-01 -3.72403301e-02 6.80569649e-01
9.19329107e-01 3.50199431e-01 4.11702126e-01 -8.92073870e-01
-7.84957349e-01 3.94116044e-01 -1.10884041e-01 3.55159193e-01
5.48311114e-01 -2.95874849e-02 1.48828661e+00 -8.42178345e-01
4.49887484e-01 1.19715738e+00 4.49513227e-01 6.97729051e-01
-1.14527154e+00 -4.41027105e-01 -6.36570081e-02 7.21532524e-01
-9.47577119e-01 -2.00863123e-01 5.23090720e-01 -1.76570058e-01
9.86891568e-01 3.46246362e-01 2.55530000e-01 4.44882452e-01
-3.75241786e-01 1.13976073e+00 9.33547616e-01 -5.85249484e-01
1.84709936e-01 3.67303252e-01 6.60394967e-01 5.93285918e-01
-1.17963396e-01 -6.92429900e-01 -4.13520902e-01 -5.40053904e-01
-7.86169097e-02 -3.90025169e-01 -2.06111208e-01 -2.93444604e-01
-8.52472246e-01 1.02501798e+00 4.38466221e-01 -8.80869851e-03
-3.46849233e-01 -9.51574221e-02 4.35039252e-01 3.77354473e-01
3.55904430e-01 8.46297026e-01 -9.66474414e-01 -1.70435786e-01
-5.51373661e-01 9.85494494e-01 1.02446556e+00 7.88092732e-01
8.12972307e-01 -6.54406011e-01 -7.22173393e-01 1.09869456e+00
1.85624674e-01 4.94930446e-01 2.39942253e-01 -8.77517343e-01
1.07059717e+00 7.67491519e-01 4.20837343e-01 -7.71760881e-01
-1.62956536e-01 -2.67690599e-01 -4.29585069e-01 -5.89106500e-01
7.40612328e-01 -1.03253275e-01 -9.13822889e-01 1.57754707e+00
7.63101995e-01 1.49758965e-01 3.15056592e-01 8.00450027e-01
1.59421647e+00 8.22698057e-01 4.88464534e-01 -1.08131549e-04
1.56881118e+00 -1.16138637e+00 -9.19665456e-01 -1.28432184e-01
7.81959713e-01 -6.75483406e-01 1.26810050e+00 2.02323616e-01
-1.15910709e+00 -6.75617933e-01 -4.56142515e-01 -5.05983651e-01
-4.40603495e-01 -6.24664240e-02 1.82286963e-01 6.68544233e-01
-7.75812447e-01 1.61107667e-02 -1.08948797e-01 9.18416083e-02
3.34871590e-01 3.00725460e-01 -2.03011543e-01 -5.25480032e-01
-1.64403772e+00 9.67776120e-01 3.67253900e-01 -1.35522142e-01
-5.81395388e-01 -8.88858140e-01 -6.31741107e-01 1.82434604e-01
4.54759538e-01 -9.97870922e-01 1.69112337e+00 -4.26485479e-01
-1.33888221e+00 6.78091824e-01 -3.28637898e-01 -3.85289907e-01
1.63162112e-01 -4.83280450e-01 -3.37269634e-01 8.61029699e-02
2.41565824e-01 6.19369328e-01 4.66405958e-01 -1.36801159e+00
-7.93521583e-01 -4.74424124e-01 4.93906021e-01 4.58016008e-01
-1.51088476e-01 -5.38573861e-02 -5.83551288e-01 -6.76378533e-02
2.70493925e-02 -5.11557281e-01 -1.30452022e-01 -6.53572440e-01
-4.46579931e-03 -8.03079665e-01 3.51494849e-01 -1.17666662e+00
1.38871634e+00 -1.70621383e+00 7.66820014e-02 3.95309813e-02
4.60542254e-02 5.09689152e-01 -3.09498012e-01 3.56206715e-01
3.66137147e-01 1.20054828e-02 -9.17559341e-02 -2.86041260e-01
6.32476658e-02 2.75719881e-01 -5.77338338e-01 -2.22756833e-01
2.51212388e-01 8.59409332e-01 -1.34889233e+00 -8.61621559e-01
-3.45772356e-01 7.95081034e-02 -7.17133760e-01 6.74355507e-01
-6.95538759e-01 4.55507785e-01 -5.78639328e-01 5.76756477e-01
6.02214932e-01 -2.28215247e-01 -4.87532616e-02 -1.21155016e-01
6.21691525e-01 6.59458935e-01 -1.06261098e+00 1.73610413e+00
-5.99071324e-01 -1.85443717e-03 -2.25857198e-01 -6.57118142e-01
7.91777074e-01 5.72604179e-01 6.17945492e-02 -1.14409196e+00
-3.04758906e-01 4.10720170e-01 -2.21723542e-02 -8.69176745e-01
8.73076081e-01 -1.93566725e-01 -1.60739765e-01 4.78635579e-02
3.32159787e-01 -4.67963248e-01 3.50931853e-01 4.73313868e-01
1.04883969e+00 -5.53205721e-02 1.35213286e-01 1.59246936e-01
7.15981483e-01 3.07352185e-01 4.50402230e-01 8.53910863e-01
-2.66172916e-01 5.29382408e-01 2.12984219e-01 -2.48877555e-01
-7.03746498e-01 -1.17985463e+00 -4.41067703e-02 1.42726612e+00
5.06680422e-02 -3.08323532e-01 -6.55828476e-01 -1.18198991e+00
2.93468893e-01 9.04185057e-01 -4.03412044e-01 1.38287777e-02
-6.33136094e-01 -5.14099658e-01 4.21467066e-01 3.86601031e-01
5.33388495e-01 -1.12726390e+00 -1.95446104e-01 4.65675265e-01
-9.47007358e-01 -9.80793476e-01 -2.46309742e-01 -6.63293079e-02
-9.39133644e-01 -1.00829220e+00 -6.58164322e-01 -6.70447111e-01
3.12882662e-01 3.15376341e-01 2.05503917e+00 1.07600302e-01
3.36574838e-02 6.26314878e-01 -6.97506249e-01 -5.19387007e-01
-2.74735779e-01 1.90593019e-01 -4.27743018e-01 -2.32636750e-01
7.01811492e-01 -5.87668084e-02 -8.17010820e-01 2.64971852e-01
-8.29901457e-01 -3.68445694e-01 1.17902480e-01 1.04983282e+00
6.69252694e-01 -2.07353175e-01 1.27041376e+00 -1.07968283e+00
1.03721833e+00 -7.54219055e-01 -3.42658162e-01 8.23165178e-01
-3.92818689e-01 -8.16493668e-03 5.77295184e-01 -1.24001458e-01
-1.32174921e+00 -3.10641438e-01 -4.14146006e-01 1.22709438e-01
-9.04803276e-02 6.04672670e-01 -3.04395288e-01 2.01419547e-01
7.91866660e-01 -1.15849450e-01 -4.30854380e-01 -3.49552721e-01
8.01119566e-01 8.10962796e-01 3.05518419e-01 -8.96006346e-01
3.01372796e-01 7.13929236e-02 -5.80182910e-01 -3.82395267e-01
-1.44816256e+00 -1.24549127e+00 -5.07676899e-01 -1.99981034e-01
8.34690630e-01 -1.03175414e+00 -7.35529482e-01 1.75454214e-01
-1.35444832e+00 -1.57582104e-01 -2.08980560e-01 9.08632502e-02
-4.15428996e-01 2.65548319e-01 -4.83412206e-01 -7.89483607e-01
-6.00725293e-01 -9.52287793e-01 9.41788197e-01 5.04188895e-01
-8.97519737e-02 -9.16573286e-01 5.50750375e-01 1.41307533e+00
2.97259092e-01 -2.50245333e-01 1.16064918e+00 -9.24575269e-01
-5.01415372e-01 -1.95895687e-01 -1.87935472e-01 4.02139992e-01
-9.73598808e-02 -4.83107418e-01 -1.01743937e+00 -1.46471933e-01
-2.39542931e-01 -9.37530518e-01 9.70177650e-01 -4.48261760e-02
1.06362975e+00 -2.43191645e-01 9.50361714e-02 -2.41812408e-01
1.43971241e+00 5.96119910e-02 8.54993284e-01 3.90124261e-01
6.28019154e-01 8.18021894e-01 9.94546473e-01 7.21787140e-02
8.93259823e-01 6.53112829e-01 3.85868639e-01 2.86946595e-01
9.68767107e-02 -8.68223086e-02 2.84138601e-02 8.34913850e-01
2.64604151e-01 -4.60162640e-01 -1.04450655e+00 1.12500262e+00
-1.92210186e+00 -9.75359082e-01 -2.69220382e-01 2.20313072e+00
1.34622812e+00 -1.62437066e-01 1.44009784e-01 -1.07818834e-01
3.42652023e-01 -3.57689150e-02 -2.91486651e-01 -2.39548981e-01
5.50993197e-02 6.09878421e-01 8.14044103e-02 5.22619128e-01
-9.94373202e-01 9.34211195e-01 5.28245354e+00 1.01010537e+00
-3.37870479e-01 3.99573267e-01 6.09171629e-01 2.35273361e-01
-5.61970413e-01 -9.69291106e-03 -1.13994217e+00 3.74062806e-01
1.06126797e+00 1.87963769e-01 -7.33675659e-02 6.54629111e-01
-2.04909816e-01 -3.42476726e-01 -7.58301258e-01 6.39122844e-01
9.33949202e-02 -1.25165796e+00 2.32183337e-01 -6.21726394e-01
8.37971568e-01 -1.82199031e-01 1.71684437e-02 9.46705341e-01
4.57679927e-01 -8.49684179e-01 2.90637910e-01 5.51439226e-01
2.20847473e-01 -8.22908938e-01 1.09275329e+00 4.32562560e-01
-7.58546770e-01 -2.02281445e-01 -5.59693396e-01 2.88985193e-01
3.04731637e-01 4.58386689e-01 -1.27902639e+00 6.94174647e-01
8.16820741e-01 -1.28255099e-01 -7.09490776e-01 1.31018734e+00
-3.25503260e-01 1.00759029e+00 -1.95801482e-02 -3.26861113e-01
4.12403286e-01 -1.11910589e-01 1.33798763e-01 1.21195722e+00
8.35228115e-02 5.59515595e-01 3.15133249e-03 4.95080888e-01
-3.94819111e-01 5.14940441e-01 -1.75095648e-01 3.01561952e-01
5.13891459e-01 1.13774347e+00 -7.84744546e-02 -4.29236352e-01
-4.95869160e-01 8.70693862e-01 7.53592908e-01 2.41571710e-01
-6.51165128e-01 -3.30484331e-01 3.28658164e-01 4.45714407e-02
4.04464364e-01 1.45723209e-01 -6.06585704e-02 -9.05146599e-01
2.51117140e-01 -1.21525204e+00 1.03133833e+00 -7.23037958e-01
-1.67891860e+00 3.11222166e-01 -1.09742247e-01 -9.20776546e-01
-1.89513057e-01 -3.48799437e-01 -4.40799356e-01 1.07339013e+00
-1.83675480e+00 -1.13710451e+00 -3.52699757e-01 6.42484546e-01
4.39349890e-01 -2.01651882e-02 9.92686391e-01 6.30683899e-01
-1.17506959e-01 6.27261579e-01 3.57861102e-01 2.37102956e-01
1.15754950e+00 -1.75019443e+00 1.30770817e-01 4.02561516e-01
3.26104254e-01 5.98283052e-01 6.98970675e-01 -6.27456367e-01
-1.10029471e+00 -1.01251280e+00 1.22479928e+00 -9.00631785e-01
3.18223953e-01 1.19875120e-02 -1.36547470e+00 3.76817465e-01
4.94508803e-01 -4.73104231e-02 1.01447666e+00 6.66938424e-01
-5.04748881e-01 -4.74051178e-01 -1.36840081e+00 3.78851950e-01
2.36884668e-01 -6.92846358e-01 -1.01074350e+00 5.30085027e-01
1.05383492e+00 -6.75432920e-01 -8.46697271e-01 4.62976962e-01
1.78552091e-01 -7.51438737e-01 1.09036982e+00 -1.22345948e+00
2.68495739e-01 -4.05598253e-01 -1.82417631e-01 -1.44762385e+00
-1.50267363e-01 -4.30041701e-02 -4.01021153e-01 1.37461400e+00
7.48753190e-01 -1.36548787e-01 8.79351199e-01 8.55034590e-01
-7.57143870e-02 -9.69694376e-01 -1.17228365e+00 -3.57966036e-01
3.63283068e-01 -1.41987711e-01 6.74441755e-01 8.86980951e-01
-4.49442379e-02 6.31255269e-01 6.00555306e-03 3.64071399e-01
3.31367344e-01 2.01136798e-01 9.37200427e-01 -1.06321752e+00
-3.92990053e-01 1.58851631e-02 -9.61991847e-02 -1.17221296e+00
8.04482996e-02 -7.55271614e-01 2.64790922e-01 -2.00215411e+00
2.51560837e-01 -6.42903388e-01 -4.18039918e-01 1.64800942e-01
-9.55128431e-01 -6.31624833e-02 8.18485841e-02 -2.09725335e-01
-1.26204050e+00 5.83869696e-01 1.32881236e+00 -2.60041416e-01
-2.25973785e-01 1.26684070e-01 -5.73249698e-01 1.15163743e-01
7.93440938e-01 -5.69203734e-01 -6.18281484e-01 -6.48233235e-01
7.21197486e-01 3.83534491e-01 1.86383516e-01 -6.89740837e-01
3.10503274e-01 1.92446169e-02 -2.29866728e-02 -8.35074842e-01
4.21138674e-01 -7.30681717e-01 -4.89358306e-01 -9.27553773e-02
-4.37557131e-01 4.33533043e-02 3.42067152e-01 7.87876964e-01
-6.34813070e-01 -5.99283993e-01 5.34441411e-01 -4.41340655e-01
-6.26158357e-01 1.48261204e-01 1.72402337e-01 8.01117659e-01
6.29437447e-01 2.72706211e-01 -5.43975651e-01 -6.64786637e-01
-5.99540353e-01 7.79069185e-01 -2.86330968e-01 2.41834775e-01
6.85498774e-01 -1.27327323e+00 -1.06186950e+00 -4.94109631e-01
5.18627346e-01 5.00362039e-01 7.17947781e-01 4.86207902e-01
-3.96935105e-01 6.39944911e-01 1.49662271e-01 -4.84572530e-01
-1.46771502e+00 2.24038869e-01 4.51059401e-01 -9.00463820e-01
1.62103534e-01 1.37073410e+00 -2.72810906e-01 -1.04727757e+00
2.65672266e-01 -1.00873284e-01 -6.01558805e-01 3.38340431e-01
6.10914946e-01 3.73086125e-01 4.91359144e-01 -3.14874381e-01
7.84756988e-02 8.47555399e-02 -3.71408731e-01 -1.30138576e-01
1.14228582e+00 -1.03424221e-01 -2.50115246e-01 4.20824647e-01
9.20900583e-01 5.99112399e-02 -6.02457106e-01 -5.35039604e-01
1.64914340e-01 -5.60777843e-01 -3.41836989e-01 -1.25139391e+00
-6.76456451e-01 1.02344537e+00 5.77454984e-01 1.85608774e-01
9.52721119e-01 8.54063407e-02 1.18719852e+00 8.18396866e-01
3.61848831e-01 -1.17362070e+00 3.79726648e-01 7.94970453e-01
8.83774757e-01 -1.62000966e+00 -3.26462775e-01 -2.32338414e-01
-7.15229750e-01 8.34301472e-01 1.04213572e+00 3.23493242e-01
2.97723264e-01 -5.96760511e-01 3.97449762e-01 -2.11425945e-01
-6.55821264e-01 -3.70936960e-01 5.82502902e-01 3.64199758e-01
5.66415548e-01 -5.58434892e-03 -4.83373970e-01 7.74712861e-01
2.73850840e-02 -2.30954200e-01 1.21895812e-01 8.11147928e-01
-5.64514756e-01 -1.25538290e+00 -3.04516107e-01 5.46590269e-01
-6.05501592e-01 -3.71471792e-01 -2.32501909e-01 4.64396417e-01
1.33696899e-01 1.34845245e+00 -2.33126983e-01 -1.64926037e-01
6.69861138e-01 6.18226111e-01 4.94727165e-01 -8.17908227e-01
-1.11344051e+00 -6.45717442e-01 4.61660445e-01 -1.41990706e-01
-4.69448000e-01 -4.11138684e-01 -1.28480995e+00 5.83753698e-02
-6.55651987e-01 6.84003413e-01 3.53793085e-01 9.78539884e-01
4.03795838e-01 3.70488465e-01 3.45580339e-01 3.28691334e-01
-8.05988848e-01 -1.14772272e+00 -2.97661245e-01 7.72001863e-01
3.50982338e-01 -3.97796869e-01 -1.08545370e-01 -5.92412017e-02] | [11.262449264526367, 8.036029815673828] |
edd4e4fb-4944-434b-a51d-92734180cf55 | masked-conditional-video-diffusion-for | 2205.09853 | null | https://arxiv.org/abs/2205.09853v4 | https://arxiv.org/pdf/2205.09853v4.pdf | MCVD: Masked Conditional Video Diffusion for Prediction, Generation, and Interpolation | Video prediction is a challenging task. The quality of video frames from current state-of-the-art (SOTA) generative models tends to be poor and generalization beyond the training data is difficult. Furthermore, existing prediction frameworks are typically not capable of simultaneously handling other video-related tasks such as unconditional generation or interpolation. In this work, we devise a general-purpose framework called Masked Conditional Video Diffusion (MCVD) for all of these video synthesis tasks using a probabilistic conditional score-based denoising diffusion model, conditioned on past and/or future frames. We train the model in a manner where we randomly and independently mask all the past frames or all the future frames. This novel but straightforward setup allows us to train a single model that is capable of executing a broad range of video tasks, specifically: future/past prediction -- when only future/past frames are masked; unconditional generation -- when both past and future frames are masked; and interpolation -- when neither past nor future frames are masked. Our experiments show that this approach can generate high-quality frames for diverse types of videos. Our MCVD models are built from simple non-recurrent 2D-convolutional architectures, conditioning on blocks of frames and generating blocks of frames. We generate videos of arbitrary lengths autoregressively in a block-wise manner. Our approach yields SOTA results across standard video prediction and interpolation benchmarks, with computation times for training models measured in 1-12 days using $\le$ 4 GPUs. Project page: https://mask-cond-video-diffusion.github.io ; Code : https://github.com/voletiv/mcvd-pytorch | ['Christopher Pal', 'Alexia Jolicoeur-Martineau', 'Vikram Voleti'] | 2022-05-19 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [ 1.34598106e-01 -3.28046918e-01 3.75236059e-03 -2.57178664e-01
-8.60144138e-01 -3.00781012e-01 7.80670404e-01 -5.86318374e-01
-8.84390026e-02 7.34010994e-01 2.69478589e-01 -4.10504133e-01
4.95635539e-01 -8.58834386e-01 -1.22875595e+00 -7.20547259e-01
-2.70111829e-01 1.26302361e-01 3.72948438e-01 -8.59773904e-02
-2.15486258e-01 3.14701527e-01 -1.63041699e+00 6.48048222e-01
5.03221989e-01 8.85399997e-01 3.21700037e-01 1.35213387e+00
3.49651515e-01 1.14106691e+00 -3.65409613e-01 -3.86223704e-01
4.91773009e-01 -8.03834319e-01 -4.37314540e-01 1.43988401e-01
3.91119391e-01 -8.82105708e-01 -6.61594510e-01 6.53812170e-01
4.84905094e-01 9.77559686e-02 5.47849715e-01 -1.28069174e+00
-6.57857120e-01 5.08805454e-01 -4.51974988e-01 3.15582454e-01
3.43217045e-01 6.54291093e-01 6.48183644e-01 -1.03416133e+00
7.72192836e-01 1.18696082e+00 6.40881419e-01 9.33463693e-01
-1.43870890e+00 -7.91467667e-01 1.11971006e-01 8.74919370e-02
-1.30143392e+00 -6.23078346e-01 5.30495405e-01 -6.76489055e-01
9.89570975e-01 2.41044521e-01 6.47446811e-01 1.51605177e+00
4.91648376e-01 8.27287436e-01 8.44335973e-01 -7.68700391e-02
2.17925653e-01 -4.21065360e-01 -3.45434576e-01 5.24430692e-01
-3.23767006e-01 2.67544419e-01 -5.88055909e-01 2.81694923e-02
1.14242959e+00 5.86465783e-02 -3.58689189e-01 9.19869766e-02
-1.49584317e+00 7.43484378e-01 -3.74953598e-02 2.93954536e-02
-5.95580280e-01 6.87031925e-01 2.76979834e-01 2.66870588e-01
6.19477749e-01 -1.94879308e-01 -3.39784116e-01 -3.63086641e-01
-1.52019501e+00 6.32253587e-01 6.70811236e-01 1.14898252e+00
8.24304879e-01 4.67649937e-01 -5.57454348e-01 5.14433920e-01
7.43722618e-02 4.21699733e-01 2.56900609e-01 -1.51396656e+00
3.51570040e-01 -2.68956840e-01 2.31067732e-01 -6.82445824e-01
6.51672184e-02 -1.58468306e-01 -9.84913468e-01 4.80487883e-01
2.51584828e-01 -5.72563589e-01 -1.25286138e+00 1.81630480e+00
2.04609618e-01 7.67404258e-01 -5.13571538e-02 9.02252197e-01
7.47831225e-01 1.17272449e+00 1.15398310e-01 -3.74694526e-01
8.11365783e-01 -1.12806606e+00 -3.92065823e-01 -7.82658532e-03
3.63254994e-01 -1.01172984e+00 8.69481683e-01 4.97648358e-01
-1.52104449e+00 -8.73113811e-01 -8.61822307e-01 -1.57404020e-01
2.25299180e-01 1.44978285e-01 4.31999773e-01 3.74869019e-01
-1.50812984e+00 8.47216964e-01 -1.24502051e+00 1.09274372e-01
4.30321485e-01 2.46926978e-01 -2.20108673e-01 -2.52209663e-01
-8.27862024e-01 4.17940736e-01 1.17133893e-01 -1.15548987e-02
-1.55017662e+00 -8.49542439e-01 -8.37560296e-01 -8.42939131e-03
-9.56122472e-04 -9.47782934e-01 1.26419389e+00 -1.29041898e+00
-1.53410971e+00 4.51076806e-01 -5.10860026e-01 -8.38137805e-01
9.14456666e-01 -3.96006763e-01 -3.60058963e-01 9.89406928e-02
7.47749135e-02 1.12653661e+00 1.28627801e+00 -1.07750022e+00
-4.67870504e-01 2.59785831e-01 -4.48616184e-02 4.11458574e-02
1.41217753e-01 1.43440098e-01 -7.45832860e-01 -1.14632034e+00
-3.40089798e-01 -1.02844453e+00 -1.90464392e-01 9.04678702e-02
-3.06030542e-01 1.35851741e-01 9.84040558e-01 -8.66682768e-01
1.18502402e+00 -2.11332154e+00 2.98985779e-01 -3.48038703e-01
2.73575317e-02 1.60903290e-01 -1.55863956e-01 2.29762614e-01
-3.00209552e-01 1.88878477e-01 -2.88258761e-01 -7.86514580e-01
-1.08303927e-01 3.56375784e-01 -5.72036207e-01 1.53575301e-01
4.32294309e-01 6.92691803e-01 -8.44877303e-01 -3.34296942e-01
3.65757614e-01 9.71119165e-01 -9.68039453e-01 4.51554269e-01
-5.90311110e-01 8.25479507e-01 3.58447526e-03 4.44503516e-01
6.45816803e-01 -2.91807950e-01 -3.66958999e-03 -3.81292552e-02
-2.16546524e-02 2.59989351e-01 -1.16342831e+00 1.81447518e+00
-3.55426580e-01 8.19303215e-01 -1.82573088e-02 -5.58551252e-01
5.24168313e-01 5.20462573e-01 5.30253351e-01 -2.43227988e-01
-1.06441833e-01 6.57117590e-02 -2.35204831e-01 -2.68826485e-01
4.82122928e-01 -4.08605561e-02 3.05729240e-01 3.98120403e-01
2.90036678e-01 -4.20216061e-02 5.45760155e-01 2.60178626e-01
1.28411150e+00 7.37492800e-01 -3.63025665e-01 8.89782161e-02
1.73251644e-01 -1.51578128e-01 7.24967778e-01 6.09105885e-01
8.88869092e-02 1.14811969e+00 6.01898193e-01 -3.73231590e-01
-1.42483532e+00 -1.24323380e+00 1.63955167e-01 1.11649370e+00
-1.50801644e-01 -6.55851007e-01 -7.55205691e-01 -4.35927868e-01
-2.31481075e-01 8.95117402e-01 -4.85003024e-01 1.03314230e-02
-7.84906387e-01 -6.93267047e-01 4.04554397e-01 5.80502272e-01
3.58792067e-01 -1.02162898e+00 -4.50158536e-01 4.27633226e-01
1.66366808e-02 -1.03850353e+00 -5.19520104e-01 -9.30795074e-02
-8.17494035e-01 -6.09368145e-01 -1.12789571e+00 -4.84403431e-01
3.66786927e-01 2.60937698e-02 1.45405018e+00 1.01702020e-01
3.07265799e-02 3.04783493e-01 -1.32266119e-01 -1.02176694e-02
-8.15543234e-01 -3.66607100e-01 -1.29584804e-01 8.35187733e-02
-8.99341423e-03 -7.40623295e-01 -9.86460865e-01 3.06781232e-01
-9.86031473e-01 5.69561481e-01 2.43830562e-01 9.91232634e-01
7.91489303e-01 -1.68788970e-01 2.48047367e-01 -6.47702217e-01
2.70502150e-01 -7.35665321e-01 -5.83189249e-01 -1.24381348e-01
-8.51493850e-02 -1.94660485e-01 5.95190227e-01 -6.06989741e-01
-1.05625212e+00 1.38860628e-01 -5.41706562e-01 -1.03765619e+00
-2.76014090e-01 2.00588122e-01 9.33302641e-02 3.24566990e-01
6.01810634e-01 5.00698745e-01 -8.97729248e-02 -3.01907957e-01
2.38167003e-01 1.27030268e-01 5.98043263e-01 -6.32683218e-01
5.04679263e-01 5.19269764e-01 -1.01911262e-01 -6.39488637e-01
-3.92081469e-01 5.37406169e-02 -2.25018352e-01 -4.55115318e-01
1.10858047e+00 -1.43467033e+00 -3.35111141e-01 6.76072538e-01
-1.22881925e+00 -1.02542698e+00 -2.57109880e-01 4.75949734e-01
-7.22736120e-01 3.75520080e-01 -1.01355135e+00 -5.69688797e-01
-3.08554828e-01 -1.60090303e+00 1.15592086e+00 2.27760170e-02
-2.70047426e-01 -8.29912364e-01 -7.17195421e-02 3.90866883e-02
4.47927922e-01 3.46554548e-01 4.09788936e-01 1.29300132e-02
-9.66316760e-01 1.30319923e-01 3.83350924e-02 7.30906069e-01
-1.59255415e-01 5.17510772e-01 -8.37670088e-01 -3.37279528e-01
-1.54854089e-01 -1.16189353e-01 1.02186370e+00 7.11873233e-01
1.28481936e+00 -3.17846000e-01 -1.11873373e-01 9.35576260e-01
1.28158236e+00 5.15449569e-02 8.69627476e-01 -1.48233488e-01
7.17810154e-01 -2.08502654e-02 3.56513709e-01 6.42638564e-01
3.13430220e-01 6.22872293e-01 4.15040702e-01 9.88639425e-03
-5.58852017e-01 -2.28797674e-01 8.02859843e-01 5.97791016e-01
-3.53140265e-01 -6.07137501e-01 -7.18187392e-01 6.56128287e-01
-1.78826618e+00 -1.24033952e+00 -2.94558376e-01 2.12307692e+00
9.61044252e-01 1.34540930e-01 1.53298944e-01 -6.63679615e-02
6.62554264e-01 2.38111183e-01 -3.83046508e-01 -1.41225502e-01
-1.03779748e-01 4.21207339e-01 3.92941177e-01 5.83011389e-01
-1.20761228e+00 8.72135520e-01 5.74474001e+00 9.19520617e-01
-1.32514918e+00 2.71653563e-01 1.16083050e+00 -4.46472257e-01
-3.20665866e-01 2.79658232e-02 -8.61639023e-01 8.58555555e-01
1.21299136e+00 8.56019631e-02 4.89443332e-01 8.02682400e-01
4.31916803e-01 -1.31097525e-01 -1.19673145e+00 9.42891717e-01
-1.40707642e-01 -1.75617361e+00 2.44986087e-01 -1.41057357e-01
1.08352995e+00 2.85420507e-01 1.86091140e-01 3.03935051e-01
4.97042507e-01 -1.03465414e+00 1.13795507e+00 7.12942481e-01
1.00119865e+00 -5.03693223e-01 3.14948022e-01 3.11732531e-01
-1.08371937e+00 9.02416930e-02 -1.68779999e-01 -6.05015196e-02
6.10787511e-01 7.29245484e-01 -5.31141460e-01 3.67673069e-01
9.06998098e-01 9.89203513e-01 -2.18986362e-01 8.52539659e-01
-1.95205837e-01 9.96304631e-01 -4.35035378e-01 5.43344557e-01
1.20987631e-01 -1.18482895e-01 5.48933446e-01 1.38454247e+00
9.04373288e-01 1.17960528e-01 -5.01399823e-02 9.78535891e-01
-1.16213225e-01 -4.10801739e-01 -4.76998419e-01 3.12568307e-01
1.66154772e-01 9.08655345e-01 -5.12917161e-01 -6.38812542e-01
-4.80252087e-01 1.23030341e+00 6.86930073e-03 6.16645098e-01
-1.40429389e+00 1.99498653e-01 8.32989991e-01 2.57643014e-01
7.16690004e-01 -5.32467246e-01 -1.78081051e-01 -1.41756725e+00
5.92046678e-02 -8.81404161e-01 2.71539390e-01 -1.08640599e+00
-9.52306390e-01 6.56735063e-01 1.38261691e-02 -1.26674426e+00
-7.19477415e-01 -3.05446923e-01 -7.22124219e-01 9.60623801e-01
-1.24950492e+00 -1.13889349e+00 -3.32363993e-01 7.68689215e-01
8.85619402e-01 4.77445358e-03 6.37578666e-01 5.14364719e-01
-4.71300930e-01 3.00001502e-01 1.96393635e-02 1.09421112e-01
6.87440276e-01 -9.80870903e-01 7.13546276e-01 1.23617077e+00
7.89576955e-03 2.47402668e-01 9.15960133e-01 -7.18584597e-01
-1.31036663e+00 -1.32843161e+00 6.85490727e-01 -2.69639343e-01
4.61082876e-01 -3.74862790e-01 -8.24501932e-01 9.43950236e-01
4.35565114e-01 3.44184250e-01 2.61038959e-01 -5.58729351e-01
-1.78655714e-01 -7.77340904e-02 -7.68247902e-01 7.96712637e-01
1.07157326e+00 -3.47229034e-01 3.33501063e-02 3.93702924e-01
7.89089561e-01 -7.28049874e-01 -8.93570900e-01 3.06114107e-01
4.22608614e-01 -1.39055502e+00 1.06664598e+00 -2.32086733e-01
1.01814485e+00 -4.35841918e-01 -2.09017113e-01 -9.87540483e-01
-2.02683747e-01 -1.03554022e+00 -6.26233935e-01 1.09406018e+00
3.50406736e-01 -2.77122021e-01 7.83599496e-01 5.85522413e-01
-3.75822097e-01 -6.79187477e-01 -8.42684567e-01 -6.66633010e-01
1.27163321e-01 -8.02523315e-01 4.63919371e-01 6.08461320e-01
-6.44901395e-01 -9.75023548e-04 -7.92640507e-01 1.33993462e-01
4.83288467e-01 -7.59576485e-02 9.35234785e-01 -3.87945592e-01
-7.62274265e-01 -4.04286742e-01 -2.13179320e-01 -1.40004909e+00
-2.17466727e-02 -5.27622342e-01 -1.77102406e-02 -1.39725649e+00
-6.41519725e-02 -2.63361812e-01 4.73561063e-02 4.46830541e-01
-9.30916891e-02 4.18501347e-01 2.60960937e-01 2.34105900e-01
-3.41735512e-01 4.88172621e-01 1.19084358e+00 -5.04636904e-04
-4.52634394e-02 -7.56648853e-02 -5.43332146e-03 5.80968440e-01
6.36239171e-01 -3.03420573e-01 -4.39276427e-01 -7.52472997e-01
1.94613449e-02 5.48353910e-01 8.54531765e-01 -1.36388803e+00
-2.43167137e-03 -1.71355069e-01 6.45762980e-01 -5.90345085e-01
6.98131144e-01 -3.91359508e-01 8.37834656e-01 2.72731870e-01
-1.20320253e-01 2.38031328e-01 1.99652180e-01 4.87402201e-01
-1.65481254e-01 1.73414618e-01 7.77633071e-01 -1.84909657e-01
-8.20390105e-01 5.31842053e-01 -6.00513697e-01 -9.82528105e-02
1.08852506e+00 -5.03152795e-02 -1.26795888e-01 -6.46149457e-01
-1.04032612e+00 -8.36751983e-02 6.86702371e-01 3.11104476e-01
6.70604110e-01 -1.30582058e+00 -9.51444864e-01 2.36661330e-01
-4.53435212e-01 1.25823468e-01 5.64693689e-01 7.83075690e-01
-8.36908519e-01 -5.86663224e-02 -1.59630194e-01 -8.23268175e-01
-1.18169713e+00 6.88666582e-01 2.29369715e-01 -2.56230921e-01
-6.50784254e-01 1.01034570e+00 3.87390524e-01 1.77409559e-01
-1.67352706e-02 -4.19288039e-01 3.36483777e-01 -3.38512182e-01
5.09601533e-01 1.74429595e-01 -1.82633266e-01 -7.99965203e-01
1.23784490e-01 1.73823237e-01 1.45262092e-01 -2.24730283e-01
1.38185227e+00 1.90250296e-02 9.01011825e-02 1.80398151e-01
1.15753031e+00 -2.54742056e-01 -2.05708313e+00 9.62702557e-02
-6.56289995e-01 -4.64085072e-01 5.01767471e-02 -5.27401209e-01
-1.29157460e+00 9.61474717e-01 4.73083615e-01 -5.10937534e-02
1.31274986e+00 -2.57356107e-01 9.36380625e-01 -3.13071966e-01
3.31678569e-01 -6.52167320e-01 -4.84364890e-02 4.49652582e-01
9.72110868e-01 -1.00704312e+00 -4.45837490e-02 -2.95662493e-01
-6.17501915e-01 1.00424910e+00 4.30152863e-01 -4.10879672e-01
7.79735267e-01 4.51105624e-01 -1.75291762e-01 2.60646641e-01
-1.21826386e+00 4.73904051e-02 2.21470863e-01 4.61737752e-01
6.35711908e-01 -2.63436325e-03 2.49370560e-03 3.45963240e-01
-7.72634819e-02 3.46906215e-01 5.53176641e-01 9.95044410e-01
8.82833749e-02 -1.03446114e+00 -3.31542999e-01 3.85790944e-01
-6.98469162e-01 -3.80193740e-01 4.11590964e-01 4.76962686e-01
1.78356826e-01 7.79880464e-01 2.20384076e-01 -4.51066703e-01
-2.55532771e-01 3.76879424e-02 3.94932061e-01 -3.68099660e-01
-4.00624454e-01 4.71424252e-01 1.30583897e-01 -7.93276846e-01
-4.99119967e-01 -8.91838491e-01 -1.05664456e+00 -7.00250328e-01
1.44372314e-01 -2.08877549e-01 3.79906029e-01 6.69762731e-01
6.34643495e-01 5.68855643e-01 3.70473534e-01 -1.50354123e+00
-1.28201514e-01 -7.92907357e-01 -1.80656195e-01 3.84011507e-01
4.57368761e-01 -4.14139450e-01 -1.65451676e-01 8.07979584e-01] | [10.733820915222168, -0.7174512147903442] |
45fb54a9-98d8-42d9-92d6-47518adffcb8 | joint-3d-scene-reconstruction-and-class | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Hane_Joint_3D_Scene_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Hane_Joint_3D_Scene_2013_CVPR_paper.pdf | Joint 3D Scene Reconstruction and Class Segmentation | Both image segmentation and dense 3D modeling from images represent an intrinsically ill-posed problem. Strong regularizers are therefore required to constrain the solutions from being 'too noisy'. Unfortunately, these priors generally yield overly smooth reconstructions and/or segmentations in certain regions whereas they fail in other areas to constrain the solution sufficiently. In this paper we argue that image segmentation and dense 3D reconstruction contribute valuable information to each other's task. As a consequence, we propose a rigorous mathematical framework to formulate and solve a joint segmentation and dense reconstruction problem. Image segmentations provide geometric cues about which surface orientations are more likely to appear at a certain location in space whereas a dense 3D reconstruction yields a suitable regularization for the segmentation problem by lifting the labeling from 2D images to 3D space. We show how appearance-based cues and 3D surface orientation priors can be learned from training data and subsequently used for class-specific regularization. Experimental results on several real data sets highlight the advantages of our joint formulation. | ['Andrea Cohen', 'Roland Angst', 'Marc Pollefeys', 'Christopher Zach', 'Christian Hane'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 5.53648353e-01 2.64605135e-01 -1.26762152e-01 -4.61423069e-01
-7.31803834e-01 -3.20294976e-01 4.61254179e-01 -1.37253165e-01
-2.35166863e-01 3.81544054e-01 -3.21559757e-02 3.97589915e-02
-7.84324780e-02 -5.47294140e-01 -6.59812450e-01 -8.76180053e-01
2.93647438e-01 5.41669846e-01 2.60108471e-01 1.34853810e-01
3.25679034e-01 7.91078866e-01 -1.39739776e+00 -1.83741733e-01
8.06804061e-01 8.62911642e-01 4.76959318e-01 3.70899439e-01
-3.96689624e-01 2.86748976e-01 -8.62396508e-02 1.08202666e-01
4.70567197e-01 -2.22304985e-01 -7.99189925e-01 1.08137333e+00
5.47010124e-01 -5.28188720e-02 1.78831771e-01 1.26734900e+00
6.47392869e-02 1.09167166e-01 9.55486059e-01 -7.93025553e-01
-3.67122114e-01 -1.93639711e-01 -9.53851700e-01 -1.23571768e-01
2.89743721e-01 -1.11601345e-01 7.60785699e-01 -1.04131484e+00
7.16505051e-01 1.13876569e+00 6.32520676e-01 4.46899325e-01
-1.47435713e+00 1.41903922e-01 5.25017381e-01 -4.78950739e-01
-1.23716831e+00 -4.41361457e-01 1.30212820e+00 -7.25062311e-01
4.36587840e-01 4.73417848e-01 4.38238293e-01 7.10461259e-01
-1.62003949e-01 9.83035386e-01 1.18782783e+00 -5.45585394e-01
2.40952447e-01 2.73555011e-01 2.57002413e-01 6.62823856e-01
3.31574768e-01 -1.55619774e-02 -1.60957262e-01 -7.90563319e-03
1.39567745e+00 3.40329260e-02 -3.69509399e-01 -7.29189813e-01
-9.33775067e-01 7.59079516e-01 3.20818633e-01 2.49998569e-01
-4.93072897e-01 8.95701256e-03 -2.14207724e-01 1.03439972e-01
8.77879798e-01 3.65050226e-01 -3.15494150e-01 5.04020810e-01
-9.78395939e-01 3.64590943e-01 5.10438144e-01 9.15972769e-01
1.21779311e+00 -4.66797017e-02 3.99614632e-01 9.35367227e-01
6.39489114e-01 3.35038334e-01 -2.44729221e-01 -1.26489377e+00
7.38935396e-02 5.59991837e-01 3.41207504e-01 -1.04297674e+00
-2.96256691e-01 -3.73245418e-01 -6.42245412e-01 3.71211916e-01
7.71833003e-01 1.69801533e-01 -1.27622914e+00 1.50980759e+00
4.69995141e-01 1.07032880e-01 -2.79290944e-01 1.17896211e+00
4.10642654e-01 5.92806816e-01 -1.19142845e-01 -3.28564644e-01
1.08716643e+00 -4.57810163e-01 -6.90945923e-01 -5.71173251e-01
2.98669875e-01 -1.02098703e+00 1.03488111e+00 2.47396857e-01
-1.42199242e+00 -4.19179827e-01 -7.65994847e-01 -8.79102796e-02
7.81283230e-02 -1.07335441e-01 6.37442231e-01 4.26259756e-01
-8.71608853e-01 4.26464587e-01 -9.96101439e-01 -2.63675094e-01
5.18200934e-01 2.89880604e-01 -1.62266403e-01 -1.97363809e-01
-6.01538420e-01 9.00100648e-01 -1.28373340e-01 2.45123029e-01
-5.36583900e-01 -4.61265743e-01 -8.97745669e-01 -4.65843320e-01
4.02644336e-01 -6.52525485e-01 9.95925426e-01 -8.86878192e-01
-1.18707371e+00 1.46174550e+00 -3.11380923e-01 -7.12945238e-02
4.63871777e-01 -1.45956233e-01 2.07834020e-01 4.12490457e-01
3.08592588e-01 6.17588937e-01 1.12264109e+00 -1.87743807e+00
-3.89472246e-01 -5.68328679e-01 2.91995108e-02 2.70806700e-01
2.01130122e-01 -2.14103386e-01 -8.21722925e-01 -6.95398033e-01
8.24981689e-01 -8.25927913e-01 -8.34282160e-01 2.39486709e-01
-6.04379237e-01 1.06102109e-01 8.07136953e-01 -4.46187973e-01
6.50804341e-01 -2.04011059e+00 4.83215213e-01 5.24696827e-01
2.69099563e-01 -4.88349944e-02 1.67895835e-02 4.95139230e-03
1.41946644e-01 2.38538861e-01 -7.28383780e-01 -5.91721356e-01
-1.22320578e-01 6.53768480e-01 8.03599358e-02 9.81949389e-01
4.27950859e-01 6.06101334e-01 -6.74449325e-01 -5.35102427e-01
5.58684826e-01 7.20322430e-01 -6.66223109e-01 3.60696346e-01
-3.44052702e-01 9.50233042e-01 -7.74681151e-01 5.13567328e-01
8.47681046e-01 -3.94992560e-01 -1.49804363e-02 -2.09970742e-01
-2.03544915e-01 4.49032225e-02 -1.25322962e+00 1.77430987e+00
-4.19492811e-01 2.61592656e-01 8.53977740e-01 -1.32569242e+00
9.17902589e-01 1.73986822e-01 7.67640650e-01 -3.44317049e-01
2.31260478e-01 1.56627670e-01 -5.18521965e-01 -6.37085736e-01
2.51875073e-01 -6.59863532e-01 5.75539805e-02 2.57927597e-01
-2.40811318e-01 -6.88583612e-01 -6.56050667e-02 -4.24690545e-02
4.97559130e-01 3.18681777e-01 -7.33001307e-02 -4.86033082e-01
4.88168299e-01 -2.28944886e-02 8.26678932e-01 5.91693640e-01
2.74244137e-02 9.52082992e-01 2.94962376e-01 -3.87539089e-01
-1.00273430e+00 -1.03298938e+00 -5.01807928e-01 3.68855268e-01
4.35037762e-01 2.07036454e-02 -7.54395425e-01 -5.32424629e-01
-1.45312563e-01 5.05992770e-01 -5.18406391e-01 2.41922319e-01
-6.14726305e-01 -7.01414824e-01 -3.13865870e-01 1.46747783e-01
2.85629392e-01 -6.75061584e-01 -6.72033429e-01 1.65714249e-01
-2.21972406e-01 -1.26405525e+00 -5.31525791e-01 3.45641643e-01
-9.93776798e-01 -1.16769302e+00 -9.88478661e-01 -9.57842886e-01
1.31167436e+00 4.34277028e-01 1.12499809e+00 2.33902320e-01
-3.05938333e-01 6.55777633e-01 -1.38465434e-01 -1.02369394e-02
-3.94735098e-01 -3.73402625e-01 -6.64607286e-02 2.88420945e-01
-5.40204458e-02 -7.76095033e-01 -4.49099898e-01 4.35533106e-01
-1.00533330e+00 2.76947320e-01 4.57730889e-01 6.32862985e-01
9.21290934e-01 7.20690787e-02 2.08177537e-01 -1.30392325e+00
1.44536510e-01 -2.17140466e-01 -6.76663458e-01 -5.95595315e-03
-4.16361511e-01 8.90380964e-02 7.07907826e-02 -2.26666600e-01
-1.30388355e+00 6.27182662e-01 -2.91909575e-01 -5.55639923e-01
-5.99749684e-01 4.20156032e-01 -3.76530766e-01 -1.60993308e-01
4.48661476e-01 6.92835078e-02 2.20604941e-01 -8.94209862e-01
3.67367923e-01 2.73623973e-01 3.57619554e-01 -7.46465504e-01
9.60491598e-01 8.13756645e-01 1.21024288e-01 -1.18289709e+00
-1.07699120e+00 -6.08334124e-01 -9.16168749e-01 -2.12855071e-01
1.04973865e+00 -7.79338658e-01 -6.50680512e-02 3.33053172e-01
-1.03886771e+00 -4.13068861e-01 -4.75648820e-01 2.95395225e-01
-7.61638224e-01 7.11978734e-01 -3.92086357e-01 -9.24960494e-01
2.72381961e-01 -1.29004061e+00 1.32647538e+00 2.18249261e-02
-1.77346274e-01 -1.34727228e+00 -2.05375448e-01 4.74399179e-01
1.02599338e-01 4.82884794e-01 7.79476881e-01 1.33695260e-01
-7.97992706e-01 -1.94100872e-01 -2.48758629e-01 4.36318606e-01
1.97392210e-01 -1.07073985e-01 -9.66526568e-01 5.23753576e-02
4.86205369e-01 -1.08945379e-02 9.20223117e-01 9.01787460e-01
9.33349907e-01 -1.56513914e-01 -3.58632445e-01 8.64427865e-01
1.48595262e+00 -1.37616172e-01 5.43971896e-01 -3.53310034e-02
9.54200566e-01 9.20429945e-01 5.71718216e-01 1.73917204e-01
1.13803826e-01 7.10689247e-01 5.06389201e-01 -4.45539117e-01
-1.53819144e-01 -9.09761861e-02 -7.94525351e-03 6.92782104e-01
-1.70321986e-01 1.15038097e-01 -7.35076010e-01 4.49373305e-01
-1.70803332e+00 -4.25509751e-01 -5.59948087e-01 2.09250760e+00
8.21842015e-01 2.39900216e-01 3.30335647e-02 1.58671692e-01
6.62490368e-01 1.40497327e-01 -4.32374239e-01 4.18732502e-03
-2.19786212e-01 6.51145652e-02 5.30825615e-01 1.10850072e+00
-1.14264071e+00 9.49579239e-01 6.40157652e+00 7.26247072e-01
-9.02582645e-01 -1.76494166e-01 7.64070153e-01 2.24205896e-01
-8.18471909e-01 2.09083796e-01 -7.19583392e-01 1.90329358e-01
3.22116643e-01 4.24593419e-01 1.25339389e-01 4.74992871e-01
3.33154023e-01 -3.53229076e-01 -8.83186817e-01 1.06060922e+00
-2.67135084e-01 -1.31292737e+00 -1.76617963e-04 2.47139215e-01
7.97333360e-01 -2.58686990e-01 8.32178339e-04 -4.86712337e-01
2.75687158e-01 -8.88343871e-01 7.66386211e-01 6.01594210e-01
6.55733407e-01 -4.20859873e-01 1.81000277e-01 6.25888467e-01
-8.89324546e-01 5.19162714e-01 -3.88527781e-01 -1.34565294e-01
5.01537621e-01 1.07634795e+00 -4.41479206e-01 2.58387774e-01
4.10581380e-01 9.04354274e-01 -2.57456213e-01 8.39330614e-01
-3.55957925e-01 2.87413955e-01 -4.78262991e-01 4.37761277e-01
3.71413648e-01 -6.84134066e-01 8.65467250e-01 9.84586656e-01
-3.96508127e-02 3.16213816e-01 4.79984224e-01 1.19267488e+00
2.83713669e-01 -1.21869072e-01 -7.64393449e-01 2.38539353e-01
9.19113904e-02 1.21155429e+00 -1.17491889e+00 -1.18004018e-02
-4.75613028e-01 9.99229372e-01 1.56708270e-01 7.83339322e-01
-3.79117906e-01 2.62874544e-01 5.61489999e-01 4.80745465e-01
1.63631678e-01 -5.11331618e-01 -6.27379537e-01 -1.39565730e+00
-7.33224973e-02 -3.84736836e-01 3.69035196e-03 -5.53787649e-01
-1.46220124e+00 2.69974530e-01 7.13128969e-02 -1.11361992e+00
1.45856710e-02 -6.63760304e-01 -4.77166951e-01 9.34205174e-01
-1.66729701e+00 -9.68245864e-01 -3.10501251e-02 5.03731012e-01
5.75828373e-01 4.69678104e-01 6.22304022e-01 1.81026906e-01
-5.00789821e-01 -1.10014290e-01 -1.70922786e-01 -4.95677590e-02
1.02454789e-01 -1.38533115e+00 1.47186920e-01 9.83779848e-01
1.64170176e-01 5.88265836e-01 8.16857100e-01 -6.52425706e-01
-1.39174104e+00 -8.84553671e-01 7.11315691e-01 -4.44042951e-01
2.24238470e-01 -4.08247411e-01 -9.97570693e-01 6.12936914e-01
-8.85438621e-02 1.40765801e-01 4.77238923e-01 -1.42716775e-02
-3.24305929e-02 4.29633826e-01 -1.11393309e+00 5.24233043e-01
9.70946312e-01 -4.91467714e-01 -5.04222989e-01 3.89331579e-01
2.86998987e-01 -2.97118574e-01 -8.08245420e-01 2.61520296e-01
2.41192691e-02 -1.01504230e+00 1.29413772e+00 -3.03581536e-01
1.95730954e-01 -3.23304921e-01 -3.54103565e-01 -1.01218581e+00
-2.10221320e-01 -7.62394071e-01 2.14713618e-01 9.69041288e-01
2.65091032e-01 -2.89295971e-01 1.01145399e+00 8.19191515e-01
-3.93750459e-01 -7.28033721e-01 -7.21535683e-01 -6.12625957e-01
1.85634688e-01 -7.26125360e-01 -8.80726650e-02 9.82334375e-01
-3.80101889e-01 3.60113323e-01 -2.13410810e-01 3.74480754e-01
1.20945787e+00 3.51078302e-01 5.89244545e-01 -1.36504197e+00
4.35460592e-03 -4.52438205e-01 -1.86351195e-01 -1.64734066e+00
2.05852523e-01 -8.80631268e-01 3.13392729e-01 -1.61068773e+00
-4.04954627e-02 -7.50133038e-01 2.39179432e-01 1.30281791e-01
-1.20396979e-01 3.81746143e-01 -1.29628465e-01 3.94887537e-01
-2.98575729e-01 4.99901831e-01 1.61290944e+00 -2.16402467e-02
-2.29941100e-01 2.36650705e-01 -7.53126323e-01 1.19809413e+00
5.58339298e-01 -1.44747302e-01 -4.28267539e-01 -5.03554940e-01
1.29783109e-01 2.07595617e-01 5.02373457e-01 -4.48206931e-01
-1.28830910e-01 -3.60252798e-01 4.11962211e-01 -6.30770147e-01
5.66583216e-01 -1.01131928e+00 3.95669378e-02 3.07161026e-02
-2.84059972e-01 -7.69397080e-01 4.16631363e-02 6.49733067e-01
-5.97440414e-02 -4.95388985e-01 1.17329443e+00 -5.01779020e-01
-5.96479774e-01 4.65596169e-01 -4.70542997e-01 1.61343127e-01
7.54348516e-01 -5.87061703e-01 5.96969843e-01 -3.43118936e-01
-1.31870604e+00 7.27366582e-02 8.79505932e-01 -6.66650385e-02
8.90995145e-01 -1.16913748e+00 -5.52726269e-01 4.76766318e-01
-2.56785750e-01 5.34223199e-01 9.04287249e-02 9.27383184e-01
-4.75106150e-01 1.70363247e-01 -8.93948078e-02 -9.57458973e-01
-1.07269573e+00 3.98824096e-01 4.35863853e-01 -7.28490427e-02
-9.41170990e-01 1.05061591e+00 6.06336951e-01 -4.80540454e-01
2.63624638e-01 -4.91472840e-01 -5.15441746e-02 -8.66782814e-02
1.18907556e-01 7.39423186e-02 -2.03380600e-01 -8.52596104e-01
-2.97248304e-01 1.07921147e+00 -6.26932606e-02 -2.33781368e-01
1.57358706e+00 -6.62853241e-01 -3.59188691e-02 4.40632671e-01
1.20434165e+00 -7.43247792e-02 -1.86733997e+00 -4.21024978e-01
1.48885921e-01 -7.33795524e-01 3.95308763e-01 -2.99159914e-01
-1.34191787e+00 9.76400375e-01 8.15288201e-02 4.52757061e-01
1.06190562e+00 3.28634560e-01 6.79772437e-01 -1.62219882e-01
1.98601261e-01 -1.05832708e+00 7.62694003e-03 3.05961996e-01
9.05966341e-01 -1.28592110e+00 1.81779206e-01 -1.03035307e+00
-4.17030096e-01 9.61612642e-01 3.03844005e-01 -4.60128695e-01
8.36981177e-01 2.17459485e-01 8.76333490e-02 -5.81634045e-01
-2.08113208e-01 -4.40936387e-01 4.97176111e-01 6.89005196e-01
3.47012877e-01 -1.45359322e-01 -3.30295712e-02 2.87891924e-01
2.61779040e-01 -3.43836039e-01 4.49933201e-01 9.60460246e-01
-5.21319926e-01 -1.16443086e+00 -5.82036018e-01 2.25222439e-01
-4.15948510e-01 3.14798564e-01 -3.49951148e-01 7.26309955e-01
7.12208599e-02 7.18921304e-01 7.98721313e-02 4.43739593e-02
1.56558171e-01 -1.67629629e-01 8.68602157e-01 -9.20273185e-01
2.08462309e-02 1.03905249e+00 -5.05362079e-02 -5.23115098e-01
-7.70245910e-01 -9.17624056e-01 -1.38787329e+00 2.10552007e-01
-3.76807690e-01 2.16732360e-02 5.32628298e-01 9.96004999e-01
-3.59821618e-02 2.88132608e-01 6.01634145e-01 -1.30575311e+00
-2.74687141e-01 -5.42756677e-01 -9.57655966e-01 5.27071178e-01
4.13562179e-01 -7.77992725e-01 -6.11879230e-01 4.14290488e-01] | [9.064848899841309, -3.051060199737549] |
4cc84a1c-b451-4877-b672-f81ddf803373 | unsupervised-meta-learning-for-few-shot-image | 1811.11819 | null | https://arxiv.org/abs/1811.11819v2 | https://arxiv.org/pdf/1811.11819v2.pdf | Unsupervised Meta-Learning For Few-Shot Image Classification | Few-shot or one-shot learning of classifiers requires a significant inductive bias towards the type of task to be learned. One way to acquire this is by meta-learning on tasks similar to the target task. In this paper, we propose UMTRA, an algorithm that performs unsupervised, model-agnostic meta-learning for classification tasks. The meta-learning step of UMTRA is performed on a flat collection of unlabeled images. While we assume that these images can be grouped into a diverse set of classes and are relevant to the target task, no explicit information about the classes or any labels are needed. UMTRA uses random sampling and augmentation to create synthetic training tasks for meta-learning phase. Labels are only needed at the final target task learning step, and they can be as little as one sample per class. On the Omniglot and Mini-Imagenet few-shot learning benchmarks, UMTRA outperforms every tested approach based on unsupervised learning of representations, while alternating for the best performance with the recent CACTUs algorithm. Compared to supervised model-agnostic meta-learning approaches, UMTRA trades off some classification accuracy for a reduction in the required labels of several orders of magnitude. | ['Ladislau Bölöni', 'Siavash Khodadadeh', 'Mubarak Shah'] | 2018-11-28 | unsupervised-meta-learning-for-few-shot-image-1 | http://papers.nips.cc/paper/9203-unsupervised-meta-learning-for-few-shot-image-classification | http://papers.nips.cc/paper/9203-unsupervised-meta-learning-for-few-shot-image-classification.pdf | neurips-2019-12 | ['unsupervised-few-shot-learning', 'unsupervised-few-shot-image-classification'] | ['computer-vision', 'computer-vision'] | [ 6.04853988e-01 2.96863765e-01 -4.71550465e-01 -4.73407596e-01
-8.34621549e-01 -2.34707907e-01 1.02853024e+00 3.84205341e-01
-5.98237514e-01 7.10154057e-01 -1.28749967e-01 2.49402644e-03
-1.04782293e-02 -8.38124394e-01 -7.23872721e-01 -7.12620020e-01
1.82086423e-01 8.07896495e-01 4.01493609e-01 -1.49258584e-01
1.85053214e-01 7.24494904e-02 -2.11842346e+00 7.06333876e-01
6.18517876e-01 9.94585872e-01 2.10602850e-01 8.68003726e-01
-4.08417940e-01 1.12051511e+00 -5.93375385e-01 -1.45316452e-01
1.69559836e-01 -5.77482343e-01 -1.02952862e+00 3.52616489e-01
5.15694380e-01 -2.16021240e-02 2.42813155e-02 8.89030874e-01
4.12978888e-01 3.71906966e-01 1.07868469e+00 -1.28360379e+00
-3.92739549e-02 7.06885576e-01 -3.35771978e-01 3.22208196e-01
-2.33111501e-01 1.11979723e-01 9.31568980e-01 -9.67500806e-01
7.08518803e-01 9.43489850e-01 4.51257974e-01 7.89786518e-01
-1.36565137e+00 -5.50090253e-01 7.50472024e-02 3.83507192e-01
-1.02135742e+00 -5.88217080e-01 5.81518888e-01 -5.56171238e-01
7.35344470e-01 1.04354374e-01 3.69423211e-01 1.23413205e+00
-1.24356851e-01 7.19127178e-01 1.31416166e+00 -8.39700520e-01
8.46168160e-01 4.85965192e-01 6.60252571e-01 4.03335512e-01
1.18536450e-01 4.75913547e-02 -3.93726528e-01 -2.00777829e-01
2.67036468e-01 3.15023780e-01 2.37381563e-01 -6.71098650e-01
-1.05572844e+00 9.41767037e-01 2.92470694e-01 3.62557262e-01
-3.02045494e-01 1.95177436e-01 7.85007894e-01 4.81506914e-01
7.05156803e-01 5.10266066e-01 -6.01281524e-01 4.25598361e-02
-9.71194327e-01 2.15364597e-03 6.86630070e-01 1.00280857e+00
1.24574590e+00 1.03952944e-01 -3.93750042e-01 9.89143312e-01
-1.15230016e-01 6.66873762e-03 9.95657563e-01 -9.12146449e-01
1.80886865e-01 6.17292047e-01 1.46596268e-01 -2.13617250e-01
-1.29472926e-01 -5.28796613e-01 -4.95975047e-01 3.66090953e-01
2.99654931e-01 -2.43002266e-01 -1.44789135e+00 1.43202531e+00
3.21415246e-01 5.43567717e-01 1.68380797e-01 3.77270550e-01
7.10322618e-01 6.17955506e-01 3.69928867e-01 -3.09673935e-01
1.23788893e+00 -1.29256248e+00 -4.17473495e-01 -4.28133816e-01
9.41653073e-01 -3.87769073e-01 1.05060065e+00 2.65013248e-01
-7.21607327e-01 -7.92250335e-01 -1.13341296e+00 3.32798600e-01
-6.64132416e-01 -1.06973939e-01 6.01676702e-01 7.67605722e-01
-6.37194633e-01 7.85842299e-01 -5.72342515e-01 -3.83950800e-01
5.49913108e-01 8.15496147e-02 -1.74369395e-01 -2.65704274e-01
-9.29005325e-01 8.37972581e-01 8.19993973e-01 -6.60788536e-01
-1.44977045e+00 -8.71197820e-01 -9.21570957e-01 9.29532945e-02
5.24392009e-01 -3.79073769e-01 1.54720938e+00 -1.37803316e+00
-1.17902577e+00 1.00124168e+00 -1.62027478e-02 -7.43897796e-01
4.52407390e-01 8.84850621e-02 -2.63096750e-01 6.65485933e-02
4.57421206e-02 8.42155397e-01 1.20893478e+00 -1.34747326e+00
-8.31943631e-01 -1.74679041e-01 2.19654381e-01 6.16652705e-02
-3.49584430e-01 -2.21724108e-01 -1.64672256e-01 -4.94508892e-01
-3.24614763e-01 -9.39197004e-01 -3.93256426e-01 -3.19697917e-01
-7.94953108e-02 -1.52613744e-01 8.62317681e-01 9.91099626e-02
8.34042966e-01 -2.03843713e+00 -9.38150957e-02 -1.66233554e-01
7.34294727e-02 3.79708439e-01 -3.70509923e-01 3.93241465e-01
-3.54514778e-01 -1.05990164e-01 -3.60218912e-01 -4.38364297e-01
-2.90430486e-01 3.53667378e-01 -2.48289675e-01 3.21478128e-01
1.09557666e-01 7.32924998e-01 -1.18196285e+00 -4.92191941e-01
4.49759841e-01 5.16029447e-02 -2.47532129e-01 2.46317372e-01
-4.91640121e-01 2.67128468e-01 -1.64676487e-01 3.67697299e-01
3.70465547e-01 -2.94794858e-01 1.34886339e-01 6.90868720e-02
1.79683283e-01 1.62936509e-01 -1.05492675e+00 1.69428265e+00
-8.83778155e-01 5.77622354e-01 -5.59104383e-01 -1.37936640e+00
7.87194848e-01 4.29438084e-01 4.36577886e-01 -4.44634289e-01
2.43035540e-01 1.68538153e-01 9.99482498e-02 -3.05289745e-01
1.45724088e-01 -3.71730357e-01 8.83054361e-03 7.26158679e-01
7.30482936e-01 6.45985305e-02 5.84654391e-01 9.28006917e-02
1.14204073e+00 1.08123906e-01 4.31598723e-01 -1.85609266e-01
2.38975585e-01 3.11063766e-01 3.92238647e-01 9.14725661e-01
-9.07915458e-02 5.97530782e-01 2.12901205e-01 -5.59010983e-01
-1.15189552e+00 -7.77928531e-01 -1.90626681e-01 1.64167809e+00
-1.92676768e-01 -3.68534446e-01 -8.12418640e-01 -1.00635242e+00
-3.97991329e-01 1.16353071e+00 -1.16849828e+00 -4.43378210e-01
-8.38502422e-02 -8.58658493e-01 9.88545343e-02 3.77553910e-01
3.23716283e-01 -1.23479903e+00 -9.73743260e-01 2.45001256e-01
2.13803172e-01 -9.25668001e-01 -1.15101874e-01 6.77988112e-01
-1.13158619e+00 -1.26508737e+00 -8.26270401e-01 -7.99235404e-01
9.10636663e-01 3.71751457e-01 1.11834490e+00 -2.64187217e-01
-4.82670218e-01 5.15319526e-01 -6.83206975e-01 -7.47767508e-01
-6.13905609e-01 1.87022328e-01 -1.70517012e-01 3.65567029e-01
4.64957565e-01 -5.99710047e-01 -3.78119975e-01 1.57485470e-01
-8.26112568e-01 5.80877550e-02 6.01702273e-01 1.06260455e+00
4.43549752e-01 -6.43508285e-02 7.27875054e-01 -1.57701242e+00
2.11572960e-01 -7.14787602e-01 -3.26168776e-01 3.35310936e-01
-6.04881883e-01 1.91256419e-01 6.87612712e-01 -7.84118891e-01
-1.04659629e+00 3.09778064e-01 3.51114273e-01 -6.95929110e-01
-4.60898370e-01 3.32687825e-01 8.48937258e-02 5.54798096e-02
1.19236100e+00 2.03969315e-01 -1.50368229e-01 -5.05292296e-01
5.32091022e-01 6.04237795e-01 1.26983166e-01 -3.36362600e-01
5.62993526e-01 4.47572947e-01 -1.46366775e-01 -9.26072598e-01
-1.30156481e+00 -7.67929316e-01 -8.83911610e-01 -1.88502043e-01
5.62712193e-01 -8.80520642e-01 -7.55917886e-03 3.47123384e-01
-8.05127680e-01 -6.48736238e-01 -8.72995079e-01 3.03014189e-01
-7.93453753e-01 -4.58950661e-02 -2.07021371e-01 -7.83258438e-01
-2.10154146e-01 -8.51905644e-01 7.23186195e-01 3.74186481e-03
-2.72942454e-01 -1.12941957e+00 2.12515950e-01 1.03426747e-01
3.05786878e-01 6.79041967e-02 8.92164111e-01 -1.19327378e+00
-1.18045621e-01 -4.57420260e-01 1.19004026e-01 4.96553421e-01
3.28986943e-01 -2.72857815e-01 -1.47678030e+00 -3.10667187e-01
6.79859221e-02 -8.72485638e-01 1.06514406e+00 2.68147111e-01
1.13909030e+00 -1.41219273e-01 -2.22631693e-01 2.27732673e-01
1.43203604e+00 2.66035646e-01 4.91866350e-01 3.99847031e-01
4.38893795e-01 6.94599450e-01 8.92809093e-01 4.86396760e-01
-4.43304591e-02 5.66915035e-01 4.06273246e-01 6.05624430e-02
-2.63664275e-01 -1.06503434e-01 2.13199943e-01 4.84516978e-01
1.00994222e-01 1.81756958e-01 -9.68000352e-01 7.63117909e-01
-1.91579199e+00 -1.08677554e+00 2.21966505e-01 2.51019621e+00
7.48769939e-01 3.44965965e-01 2.60918438e-01 3.60826731e-01
7.66102433e-01 2.37124786e-01 -6.51476443e-01 -3.64379197e-01
3.15618753e-01 3.66250694e-01 3.78625840e-01 2.25529656e-01
-1.25150383e+00 9.50801611e-01 6.53508615e+00 9.50821340e-01
-1.12084877e+00 5.21667182e-01 7.75230050e-01 -1.47214025e-01
1.06002308e-01 1.70824319e-01 -8.10363114e-01 2.61016667e-01
1.38127840e+00 -3.58277112e-01 1.23511508e-01 1.22816837e+00
-2.18062535e-01 -2.13110864e-01 -1.37386966e+00 7.75434494e-01
1.96968630e-01 -1.39410174e+00 9.51299742e-02 -1.38508394e-01
1.05000257e+00 2.01046422e-01 -8.13072994e-02 8.19972515e-01
5.04326820e-01 -7.95877755e-01 5.58588207e-01 3.71620119e-01
7.72813082e-01 -8.73650908e-01 6.50644124e-01 7.81733334e-01
-8.82826447e-01 -3.59662473e-01 -6.50550544e-01 -5.08323126e-02
-3.00466061e-01 4.39363450e-01 -9.52296674e-01 3.27424914e-01
4.46588725e-01 6.43406153e-01 -7.36102700e-01 1.12227571e+00
-6.26774654e-02 7.99199760e-01 2.99246684e-02 2.77857874e-02
3.64637941e-01 1.65384710e-01 2.47225240e-01 1.22300828e+00
5.54963108e-03 -9.15691350e-03 4.06501323e-01 4.39954817e-01
-1.29013836e-01 1.15544058e-01 -7.18062043e-01 -2.49579102e-02
4.16448891e-01 1.38543320e+00 -7.29117751e-01 -9.55305517e-01
-3.43106270e-01 8.74904037e-01 3.16961050e-01 2.19859108e-01
-5.42379439e-01 -4.69974399e-01 1.44273818e-01 2.74138957e-01
3.45981449e-01 2.32415825e-01 -1.30331993e-01 -1.01129413e+00
-4.75628197e-01 -7.48003840e-01 4.46283102e-01 -5.28735459e-01
-1.20458019e+00 5.53592920e-01 2.81861484e-01 -1.52803946e+00
-6.53496444e-01 -5.49086809e-01 -9.30303812e-01 6.05164528e-01
-1.42999411e+00 -1.19565630e+00 -4.14706945e-01 4.80872393e-01
1.13989055e+00 -4.59156364e-01 1.08804035e+00 -7.71984085e-02
-3.58984321e-01 3.22000921e-01 1.28205582e-01 1.17865220e-01
7.56024241e-01 -1.21627057e+00 2.99420476e-01 3.99115473e-01
3.53847444e-01 2.55193979e-01 7.55301178e-01 -4.19946134e-01
-8.60970795e-01 -1.50623775e+00 5.68291843e-01 -4.09264535e-01
6.04931295e-01 -3.37913603e-01 -8.80911827e-01 7.42666662e-01
1.84491634e-01 4.43147004e-01 7.85807490e-01 1.68104038e-01
-5.35900474e-01 -1.37823403e-01 -1.04330349e+00 3.93267095e-01
7.63441741e-01 -3.87437254e-01 -8.35005999e-01 5.66210806e-01
5.75775445e-01 4.93624657e-02 -6.02475524e-01 2.60410637e-01
2.23960623e-01 -8.81543815e-01 7.00822115e-01 -9.80141759e-01
5.19752264e-01 -7.91364256e-03 -6.88595325e-02 -1.55842507e+00
-1.91671446e-01 -6.77922517e-02 -2.84933388e-01 1.04739475e+00
5.18562675e-01 -3.52379948e-01 9.76429164e-01 2.33466730e-01
-4.89019118e-02 -5.99806309e-01 -7.49948978e-01 -1.07881033e+00
-4.82998742e-03 -4.68371063e-01 1.74488813e-01 1.16042316e+00
-4.00529243e-02 6.87467575e-01 -4.25700963e-01 -3.94198179e-01
1.00517738e+00 4.85939421e-02 8.89740646e-01 -1.45085514e+00
-4.47899491e-01 -9.30005461e-02 -3.08683187e-01 -3.06018859e-01
2.21484393e-01 -1.10445845e+00 -4.26202640e-02 -1.41737032e+00
4.92397457e-01 -4.52488273e-01 -5.47948241e-01 7.17732430e-01
-2.27981675e-02 1.40603378e-01 8.38209465e-02 2.03803316e-01
-7.31331646e-01 4.02545035e-01 8.72348487e-01 -3.85173738e-01
-2.38740981e-01 2.55929708e-01 -4.04416025e-01 8.31273615e-01
8.94358158e-01 -7.24444151e-01 -7.97034740e-01 -1.25075197e-02
-2.03092486e-01 -1.11282334e-01 2.78279573e-01 -1.21362364e+00
1.47200242e-01 -1.59511715e-01 3.08598489e-01 -3.17631900e-01
5.62034667e-01 -7.45301425e-01 1.87691897e-02 5.54168463e-01
-6.42351985e-01 -4.85268384e-01 2.21036207e-02 6.18761957e-01
-1.22351587e-01 -8.69884849e-01 1.26804161e+00 -5.81770539e-01
-9.86860871e-01 3.55012387e-01 -3.17068338e-01 1.34349391e-01
1.28610802e+00 -2.41110295e-01 -1.48269415e-01 -1.84794068e-01
-1.07214642e+00 -3.32691856e-02 4.28745568e-01 4.21101868e-01
4.45322633e-01 -1.14312160e+00 -6.01737916e-01 2.84141246e-02
6.58547819e-01 -1.76235914e-01 2.57746339e-01 5.96222997e-01
-4.01008800e-02 1.67557135e-01 -4.06337947e-01 -6.04110539e-01
-1.20478189e+00 8.48415315e-01 3.14390689e-01 -3.26825857e-01
-5.17830074e-01 5.36544383e-01 1.67136326e-01 -4.03400570e-01
2.58018017e-01 1.80679888e-01 -4.11004603e-01 4.97668147e-01
7.83087015e-01 4.96386826e-01 1.60245970e-01 -3.92779201e-01
9.54175070e-02 1.52339309e-01 -5.05197346e-01 -1.49871483e-01
1.38291073e+00 2.13321418e-01 3.19175869e-01 1.23186922e+00
1.18755805e+00 -4.94853616e-01 -1.34981620e+00 -4.66199189e-01
3.18664491e-01 -3.65012199e-01 2.70705130e-02 -7.58909464e-01
-8.58374238e-01 1.16401196e+00 7.65030861e-01 2.79241428e-02
8.13272059e-01 -1.00266978e-01 2.58686423e-01 5.78095317e-01
6.61130846e-01 -1.28340292e+00 4.31727290e-01 5.17089367e-01
5.06708741e-01 -1.52234507e+00 -6.92843944e-02 -1.26306981e-01
-8.00998032e-01 9.74391818e-01 7.09040344e-01 -2.81286240e-01
7.05519795e-01 -3.09953373e-02 -3.06697842e-02 -6.71346113e-02
-1.16802180e+00 -4.26285535e-01 2.35631526e-01 6.93062782e-01
3.24191719e-01 -7.96223134e-02 -2.31585093e-02 3.15464854e-01
3.46817523e-01 2.92202588e-02 4.56628829e-01 1.22121084e+00
-9.00939226e-01 -1.15422380e+00 -2.02764854e-01 7.32150197e-01
-1.80305511e-01 -5.03483377e-02 -2.50109613e-01 7.00714588e-01
9.99767259e-02 7.88817108e-01 1.28329739e-01 -2.82016933e-01
1.99614108e-01 5.05497217e-01 4.61092234e-01 -1.39122534e+00
-5.76500475e-01 -1.17874928e-01 1.20442830e-01 -1.75368279e-01
-5.48886120e-01 -4.28925574e-01 -9.30229068e-01 1.19292870e-01
-3.34065050e-01 2.44670600e-01 6.19172156e-01 1.11968160e+00
1.75796561e-02 5.43429375e-01 8.99618506e-01 -1.12430823e+00
-5.37251532e-01 -1.20681596e+00 -6.00978374e-01 4.60950792e-01
6.81184530e-02 -8.25918972e-01 -3.84766579e-01 2.22425804e-01] | [9.942405700683594, 3.0413849353790283] |
ea881f7d-d844-4fb9-95c0-3bf397f389cc | adversarial-semi-supervised-audio-source | 1711.00048 | null | http://arxiv.org/abs/1711.00048v2 | http://arxiv.org/pdf/1711.00048v2.pdf | Adversarial Semi-Supervised Audio Source Separation applied to Singing Voice Extraction | The state of the art in music source separation employs neural networks
trained in a supervised fashion on multi-track databases to estimate the
sources from a given mixture. With only few datasets available, often extensive
data augmentation is used to combat overfitting. Mixing random tracks, however,
can even reduce separation performance as instruments in real music are
strongly correlated. The key concept in our approach is that source estimates
of an optimal separator should be indistinguishable from real source signals.
Based on this idea, we drive the separator towards outputs deemed as realistic
by discriminator networks that are trained to tell apart real from separator
samples. This way, we can also use unpaired source and mixture recordings
without the drawbacks of creating unrealistic music mixtures. Our framework is
widely applicable as it does not assume a specific network architecture or
number of sources. To our knowledge, this is the first adoption of adversarial
training for music source separation. In a prototype experiment for singing
voice separation, separation performance increases with our approach compared
to purely supervised training. | ['Sebastian Ewert', 'Simon Dixon', 'Daniel Stoller'] | 2017-10-31 | null | null | null | null | ['audio-source-separation', 'music-source-separation'] | ['audio', 'music'] | [ 4.89963889e-01 8.33586082e-02 7.17204586e-02 1.60194308e-01
-1.09938884e+00 -9.27046537e-01 3.17214340e-01 -1.67595088e-01
-1.29292160e-01 8.17163408e-01 1.77670032e-01 1.08374227e-02
-3.45785588e-01 -3.88161868e-01 -7.68708885e-01 -8.27740192e-01
3.10135838e-02 4.65879261e-01 -1.39909431e-01 -2.70387501e-01
-2.21703723e-01 4.53714579e-01 -1.45820725e+00 2.28769436e-01
8.63110960e-01 8.50535870e-01 -1.51390567e-01 9.35825109e-01
1.97865982e-02 6.99272215e-01 -1.21713006e+00 -2.94488311e-01
5.43737411e-01 -1.02953815e+00 -4.32626933e-01 -4.86072488e-02
5.40432870e-01 -1.57219782e-01 -2.65364736e-01 9.99612808e-01
7.42347956e-01 -1.06096780e-02 6.70812607e-01 -1.25977504e+00
-2.28437886e-01 1.31335366e+00 -2.12910399e-01 -1.85010247e-02
2.03380600e-01 3.38072889e-02 8.49868834e-01 -5.68540454e-01
1.66898459e-01 7.06002951e-01 1.11899602e+00 4.63578075e-01
-1.64060175e+00 -9.04213905e-01 -3.58318985e-01 -3.04024577e-01
-1.33376074e+00 -1.00053644e+00 1.07791960e+00 -2.39004314e-01
3.72183412e-01 5.84216297e-01 5.38408458e-01 1.29770505e+00
-6.19838715e-01 8.96800220e-01 1.06524587e+00 -6.24734521e-01
1.85448363e-01 3.61048698e-01 -2.51976818e-01 -7.66502544e-02
1.80884644e-01 1.88130558e-01 -5.92891991e-01 -2.48658001e-01
7.50581920e-01 -4.58253324e-01 -6.15663648e-01 -3.55634212e-01
-1.28864741e+00 4.98620331e-01 1.44172043e-01 6.06005549e-01
-2.43433505e-01 1.66372806e-01 3.69321287e-01 6.68124914e-01
2.34290376e-01 8.98733974e-01 -1.33762375e-01 -1.12743005e-01
-1.69730377e+00 3.46245021e-01 1.06616187e+00 6.24445260e-01
1.14180990e-01 7.23265171e-01 -2.45943200e-02 8.37317169e-01
2.85399463e-02 5.98412395e-01 8.56287837e-01 -1.19533551e+00
2.01638386e-01 -6.64935187e-02 1.60038009e-01 -8.11489046e-01
-5.82343861e-02 -9.61397350e-01 -8.60681474e-01 5.87871671e-01
1.06892848e+00 -4.52975452e-01 -8.15569162e-01 1.87901139e+00
-3.53368483e-02 5.79992056e-01 4.38499153e-01 8.57226431e-01
5.86780250e-01 1.90706283e-01 -6.15291178e-01 -2.07373843e-01
8.22829902e-01 -7.87593663e-01 -8.89250994e-01 -3.12434047e-01
7.89019689e-02 -1.10075033e+00 7.15759218e-01 7.71243513e-01
-1.23419893e+00 -6.38319671e-01 -1.35159433e+00 4.30448472e-01
-7.81918615e-02 2.55049497e-01 4.06620145e-01 1.05354917e+00
-6.64029896e-01 1.02194595e+00 -5.48316896e-01 3.05504631e-02
2.72467911e-01 4.62870836e-01 -2.84396559e-01 5.02946973e-01
-1.16645896e+00 6.43607438e-01 4.47020084e-01 8.12691078e-02
-9.75925744e-01 -6.26309812e-01 -5.99265039e-01 -8.72408226e-03
2.23772228e-01 -4.50348675e-01 1.34973431e+00 -1.76950192e+00
-1.73624599e+00 4.83981699e-01 1.20389089e-01 -7.90760756e-01
6.52340949e-01 -3.91792148e-01 -6.57286882e-01 7.53079727e-02
-1.29216462e-01 2.24988088e-01 1.41604125e+00 -1.54458904e+00
-1.75955087e-01 5.94402924e-02 -3.06897163e-01 6.23086356e-02
-2.01916426e-01 -1.74687862e-01 6.55176044e-02 -1.08839798e+00
1.15830779e-01 -9.53383446e-01 -6.05953820e-02 -3.20696205e-01
-6.75296843e-01 5.36179781e-01 4.77332294e-01 -6.72347188e-01
9.08743083e-01 -2.33968735e+00 2.41537437e-01 3.77255648e-01
8.76088962e-02 4.73346651e-01 -2.13765740e-01 4.14706916e-01
-3.75593245e-01 -8.69662091e-02 -3.38642985e-01 -5.83819866e-01
7.83208609e-02 -5.25530577e-02 -7.49072850e-01 4.91620034e-01
1.00937791e-01 6.21827900e-01 -7.75205135e-01 -3.30398120e-02
2.13661660e-02 5.23782134e-01 -3.98221463e-01 2.49104649e-01
-9.34279189e-02 7.85282791e-01 1.92718953e-01 1.90044910e-01
5.66552818e-01 2.24376589e-01 2.48581260e-01 -8.12068507e-02
2.14494959e-01 4.49303567e-01 -1.67916620e+00 1.95522821e+00
-3.54875654e-01 7.93820262e-01 3.98584664e-01 -1.04945886e+00
9.46253479e-01 8.38268757e-01 4.22563046e-01 4.59222794e-02
2.59051472e-01 6.31875813e-01 4.67584193e-01 -4.84130457e-02
3.20231438e-01 -5.56617081e-01 1.04805753e-01 5.49749613e-01
3.11761051e-01 -3.11910391e-01 -6.79262504e-02 -1.57429099e-01
8.67622018e-01 1.92987055e-01 5.53979836e-02 2.61027962e-02
2.42033914e-01 -1.86374754e-01 5.73052585e-01 8.97386730e-01
1.48427054e-01 1.11254096e+00 2.67333061e-01 2.16562167e-01
-8.85901093e-01 -1.27950728e+00 5.52004501e-02 5.76959252e-01
-2.47128949e-01 -2.18934476e-01 -8.44474256e-01 -3.12924236e-01
-9.89610627e-02 7.95807421e-01 -3.37670863e-01 -2.26431832e-01
-5.61482251e-01 -4.26450521e-01 1.42143035e+00 3.57420385e-01
3.75437848e-02 -1.04655325e+00 -3.17234695e-01 3.33017200e-01
-1.54459953e-01 -8.95318508e-01 -4.85082477e-01 5.52440047e-01
-8.39824855e-01 -8.83739412e-01 -1.06284952e+00 -6.02841079e-01
2.58287340e-01 7.54219666e-02 1.03343248e+00 -2.34788388e-01
1.33257076e-01 9.34215337e-02 -1.24470837e-01 -7.22888052e-01
-9.16233540e-01 5.18289246e-02 4.09992337e-01 2.74972588e-01
6.88142926e-02 -1.18801677e+00 -3.21390808e-01 2.56157637e-01
-9.67965186e-01 -2.22316980e-01 4.88352805e-01 6.67852879e-01
2.15985373e-01 4.64368820e-01 9.15832639e-01 -6.71236515e-01
7.96371043e-01 -4.19575691e-01 -2.02139109e-01 -7.38633722e-02
-3.46221417e-01 2.76538748e-02 1.05357492e+00 -9.07441199e-01
-7.62735188e-01 -7.25731440e-03 -8.22894275e-02 -8.68428767e-01
-5.59674382e-01 2.34794319e-01 -2.80152380e-01 1.32519081e-01
9.91352201e-01 2.56449908e-01 6.70470148e-02 -6.78141356e-01
4.29076105e-01 6.60405457e-01 9.12732542e-01 -4.03658450e-01
1.04994464e+00 4.62991059e-01 -2.20694900e-01 -7.34518528e-01
-7.35278070e-01 -3.22503150e-01 -7.06898451e-01 4.53929529e-02
4.08831477e-01 -8.14149559e-01 -5.48318446e-01 6.52268589e-01
-9.24898565e-01 -3.75396162e-01 -7.97670901e-01 8.25875044e-01
-6.41964495e-01 2.10131839e-01 -4.43229467e-01 -1.12638569e+00
-1.59679472e-01 -8.18335712e-01 6.97141945e-01 1.19353689e-01
-5.39108217e-01 -8.90170932e-01 2.76783198e-01 2.43972778e-01
4.16378796e-01 2.57827491e-01 3.67922902e-01 -1.10370874e+00
-1.78568542e-01 -3.96795362e-01 5.51378369e-01 7.70214975e-01
4.61090118e-01 -8.17594752e-02 -1.55779982e+00 -2.32187748e-01
3.72927308e-01 -3.78640115e-01 7.47838914e-01 3.59442145e-01
5.58743775e-01 -3.54196995e-01 -4.09481525e-02 5.88319778e-01
1.04139721e+00 6.62702024e-02 6.10679686e-01 9.32077542e-02
6.80709898e-01 4.54621762e-01 6.91047460e-02 7.17369094e-02
-3.96004051e-01 6.09753907e-01 1.98826864e-01 -2.11296856e-01
-3.76282483e-01 -3.49631488e-01 5.97015381e-01 8.32656920e-01
-2.14309990e-02 -3.29789907e-01 -5.41932106e-01 6.35062039e-01
-1.42587757e+00 -1.26754701e+00 -8.32513943e-02 2.54644156e+00
1.02846503e+00 2.31804863e-01 4.86106843e-01 9.21095550e-01
6.50075555e-01 1.05008185e-01 -4.98734921e-01 -1.32503822e-01
-3.45539123e-01 5.35851657e-01 3.15473795e-01 3.21343780e-01
-1.10066307e+00 4.64438796e-01 6.47160816e+00 9.25146461e-01
-1.33035827e+00 4.88423705e-02 3.47976424e-02 -5.80671787e-01
-2.01592967e-01 -5.28555997e-02 -2.36958027e-01 3.69380832e-01
1.03869998e+00 1.40982404e-01 7.62402773e-01 4.56020594e-01
1.80657171e-02 2.09211066e-01 -1.23149776e+00 9.64534998e-01
1.93289220e-01 -9.79287565e-01 -3.91933173e-01 -4.45119850e-02
5.45221090e-01 -1.41496152e-01 2.07686529e-01 1.01207688e-01
2.38915414e-01 -1.38030279e+00 1.11074769e+00 4.88791794e-01
5.86964846e-01 -8.53031635e-01 5.27614832e-01 7.30198324e-01
-7.81082571e-01 1.54429913e-01 -1.49885584e-02 4.35238658e-03
1.58600628e-01 6.31852090e-01 -8.89158070e-01 6.98717356e-01
1.45260483e-01 3.92297000e-01 -2.19196156e-01 1.11784410e+00
-2.59709269e-01 1.05537665e+00 -4.64123338e-01 5.90051293e-01
-1.34244531e-01 -2.08533987e-01 9.68393087e-01 1.03221643e+00
4.63707536e-01 -5.14585137e-01 -7.71865696e-02 9.11411345e-01
-3.29871401e-02 -1.36055768e-01 -7.50012159e-01 -2.46023446e-01
4.29644704e-01 1.01018131e+00 -6.91666961e-01 -1.92278609e-01
-5.79431327e-03 1.05753613e+00 3.89435235e-03 3.72929424e-01
-6.86361074e-01 -4.12094474e-01 5.66402137e-01 1.17498659e-01
4.17363226e-01 -9.34449062e-02 -3.17050308e-01 -1.14106238e+00
-5.88193387e-02 -1.47968125e+00 -6.36229478e-03 -5.93813658e-01
-1.29825139e+00 7.82792330e-01 -2.89252728e-01 -1.58173943e+00
-4.94252056e-01 -2.36965463e-01 -8.07970166e-01 1.05057824e+00
-1.07457113e+00 -9.69063818e-01 1.08760998e-01 4.97735858e-01
2.91346222e-01 -3.29721451e-01 1.10562599e+00 2.77386278e-01
-2.35951975e-01 6.77366376e-01 4.74436939e-01 2.81944633e-01
9.74049509e-01 -1.47322309e+00 2.24405617e-01 8.95712972e-01
9.55994904e-01 6.16782188e-01 9.89752710e-01 -3.86748075e-01
-8.98627281e-01 -6.84136093e-01 3.93888682e-01 -4.63536620e-01
6.07666612e-01 -4.48020428e-01 -1.08193445e+00 4.64373440e-01
4.28497970e-01 -1.43769309e-01 1.10491550e+00 8.88686329e-02
-3.66568804e-01 -9.05958042e-02 -9.10167873e-01 5.13748407e-01
7.30043411e-01 -6.38757348e-01 -8.02273333e-01 3.14014778e-02
2.49075323e-01 -4.45059210e-01 -7.73114324e-01 2.04339191e-01
7.03250766e-01 -1.19096029e+00 1.16419637e+00 -5.00472665e-01
2.37683535e-01 -4.70918804e-01 -2.99630556e-02 -1.64940763e+00
8.71667359e-03 -1.01002753e+00 -2.98395604e-01 1.66113830e+00
4.61501449e-01 -4.55328822e-01 6.94160879e-01 1.25134483e-01
6.77268207e-02 -5.06927669e-02 -7.19805479e-01 -1.14911771e+00
2.90987998e-01 -5.45747995e-01 6.19141161e-01 1.08215475e+00
-1.77498713e-01 3.34618628e-01 -7.71685719e-01 3.30442846e-01
7.44672239e-01 2.73249805e-01 9.92520988e-01 -1.22418106e+00
-1.03206897e+00 -5.19870639e-01 -1.35751575e-01 -7.57168293e-01
2.30828539e-01 -1.04259753e+00 1.56824812e-01 -1.05563974e+00
-4.99858499e-01 -6.22817934e-01 -4.67677057e-01 2.28120819e-01
-6.55744448e-02 5.71325660e-01 4.53222960e-01 3.90858263e-01
1.51626363e-01 3.84396285e-01 9.53458488e-01 -1.69799581e-01
-4.24164236e-01 5.30977905e-01 -8.64261627e-01 1.00648630e+00
9.65525270e-01 -8.16584885e-01 -6.39674425e-01 -2.54235983e-01
-7.67729804e-02 1.73540816e-01 4.49926615e-01 -1.38824868e+00
1.21763796e-01 3.16226929e-01 4.41493303e-01 -1.14043169e-01
6.87296271e-01 -8.80992830e-01 6.50169134e-01 1.75855845e-01
-4.38510060e-01 -6.56946838e-01 4.47778851e-01 4.59639847e-01
-3.51756424e-01 -5.65023959e-01 6.79787993e-01 -1.38411941e-02
1.52918145e-01 -2.63851881e-01 -3.22768420e-01 1.58098772e-01
5.92820287e-01 -2.98708797e-01 2.29528010e-01 -7.70499647e-01
-8.63129258e-01 -4.22419548e-01 3.43262821e-01 1.70427531e-01
1.88831940e-01 -1.39363837e+00 -8.05224180e-01 3.74879122e-01
-3.33995432e-01 -5.00393147e-03 -6.88204765e-02 9.22287822e-01
-8.23607817e-02 5.86946160e-02 -7.25644156e-02 -3.65218163e-01
-1.28101265e+00 5.36458552e-01 5.19251049e-01 -9.48781744e-02
-5.19955456e-01 6.94947004e-01 8.83589685e-03 -4.01553482e-01
2.43545502e-01 -1.54256970e-01 -4.14101779e-02 2.13090211e-01
5.42989314e-01 2.75783062e-01 3.49427462e-02 -6.01447105e-01
-1.07656173e-01 2.72867560e-01 3.15462023e-01 -5.86067319e-01
1.14879036e+00 1.62657619e-01 8.17025527e-02 8.97550166e-01
9.13822651e-01 9.35510278e-01 -1.11744022e+00 -1.04515389e-01
-1.24978468e-01 -3.92918915e-01 -1.00743994e-01 -7.78713942e-01
-1.04622459e+00 8.17926347e-01 4.69959795e-01 6.12600923e-01
1.23799217e+00 -2.54040837e-01 6.55511618e-01 1.37880221e-01
2.17285473e-02 -7.10272133e-01 -1.70101926e-01 1.72162905e-01
9.30453718e-01 -9.90072250e-01 -3.21306258e-01 -9.70066786e-02
-5.08183241e-01 1.08597958e+00 1.52265236e-01 -4.77405965e-01
4.06967342e-01 6.53448761e-01 4.87791449e-01 1.51094615e-01
-1.74677774e-01 -2.57107794e-01 4.80592757e-01 6.63874030e-01
4.79603976e-01 -1.38812035e-01 2.80702084e-01 7.38158286e-01
-6.20975912e-01 2.14471053e-02 5.37934482e-01 5.33541083e-01
-1.20691851e-01 -1.41227257e+00 -9.85688746e-01 2.78011739e-01
-8.60274851e-01 -2.82896698e-01 -6.09045684e-01 5.92231095e-01
1.89780146e-01 1.18920410e+00 -3.52491625e-02 -2.57962167e-01
2.01207548e-01 3.34249645e-01 5.70531428e-01 -5.07536471e-01
-8.89538407e-01 5.40802360e-01 6.50659874e-02 6.98593259e-02
-5.75852871e-01 -5.21851182e-01 -1.02788103e+00 2.14429852e-03
-5.79746425e-01 4.35612887e-01 5.21548331e-01 7.75263250e-01
7.01122135e-02 8.50723863e-01 4.94256288e-01 -1.01947153e+00
-7.06210554e-01 -1.02633834e+00 -8.93733323e-01 4.35174674e-01
7.20640242e-01 -2.93121606e-01 -5.58542013e-01 2.53906548e-01] | [15.537339210510254, 5.5534563064575195] |
7f94d40f-ec45-4154-a09a-48ba505d8c51 | coordinating-cross-modal-distillation-for | 2211.16712 | null | https://arxiv.org/abs/2211.16712v1 | https://arxiv.org/pdf/2211.16712v1.pdf | Coordinating Cross-modal Distillation for Molecular Property Prediction | In recent years, molecular graph representation learning (GRL) has drawn much more attention in molecular property prediction (MPP) problems. The existing graph methods have demonstrated that 3D geometric information is significant for better performance in MPP. However, accurate 3D structures are often costly and time-consuming to obtain, limiting the large-scale application of GRL. It is an intuitive solution to train with 3D to 2D knowledge distillation and predict with only 2D inputs. But some challenging problems remain open for 3D to 2D distillation. One is that the 3D view is quite distinct from the 2D view, and the other is that the gradient magnitudes of atoms in distillation are discrepant and unstable due to the variable molecular size. To address these challenging problems, we exclusively propose a distillation framework that contains global molecular distillation and local atom distillation. We also provide a theoretical insight to justify how to coordinate atom and molecular information, which tackles the drawback of variable molecular size for atom information distillation. Experimental results on two popular molecular datasets demonstrate that our proposed model achieves superior performance over other methods. Specifically, on the largest MPP dataset PCQM4Mv2 served as an "ImageNet Large Scale Visual Recognition Challenge" in the field of graph ML, the proposed method achieved a 6.9% improvement compared with the best works. And we obtained fourth place with the MAE of 0.0734 on the test-challenge set for OGB-LSC 2022 Graph Regression Task. We will release the code soon. | ['Meng Liu', 'Yingyi Zhang', 'Lei Shen', 'Ruixin Zhang', 'Nan Zhang', 'Hao Zhang'] | 2022-11-30 | null | null | null | null | ['graph-regression', 'molecular-property-prediction'] | ['graphs', 'miscellaneous'] | [ 1.45662636e-01 -1.21573612e-01 -4.44404423e-01 -5.21551780e-02
-4.39335495e-01 -3.50860655e-01 3.32894087e-01 4.68907863e-01
-2.70482302e-01 1.06145811e+00 -1.62998334e-01 -6.32850409e-01
-7.54458308e-02 -8.34594607e-01 -9.34101462e-01 -9.55546975e-01
-1.82159007e-01 4.34645474e-01 9.03878361e-02 -2.08864003e-01
3.96011263e-01 7.57415235e-01 -1.00804472e+00 1.76908359e-01
1.10453546e+00 9.01856601e-01 2.46860370e-01 3.82835776e-01
-2.45066702e-01 6.79885983e-01 -2.43070930e-01 -3.73464763e-01
2.55766362e-01 -4.27482426e-01 -7.30788410e-01 -3.82520944e-01
6.92233264e-01 3.60285752e-02 -3.11841905e-01 1.04373777e+00
6.93768084e-01 2.50913918e-01 8.45988870e-01 -1.04931438e+00
-7.96081901e-01 4.13462937e-01 -7.91039169e-01 2.71739345e-02
4.71424818e-01 3.77765566e-01 1.07470679e+00 -7.92539060e-01
8.42864811e-01 1.22608256e+00 5.24039924e-01 3.55522811e-01
-1.37726963e+00 -7.40858495e-01 2.61374354e-01 5.76603711e-01
-1.59547329e+00 4.75911684e-02 9.03193891e-01 -4.82120246e-01
1.35503089e+00 1.61842510e-01 4.78161812e-01 1.07592547e+00
4.40019965e-01 3.52014810e-01 1.09522855e+00 -1.46335512e-01
2.46718466e-01 -9.16107073e-02 1.91382602e-01 1.03228152e+00
4.42360491e-01 7.80206248e-02 -4.42295820e-01 -1.39403611e-01
5.73137879e-01 5.98725230e-02 -5.17044187e-01 -4.62119102e-01
-9.03851330e-01 9.33969021e-01 1.06264985e+00 -5.02269492e-02
-2.59589374e-01 1.14163384e-01 2.29972884e-01 2.74401933e-01
3.94770533e-01 7.63392031e-01 -2.75828093e-01 9.29212421e-02
-4.91034150e-01 1.90834001e-01 7.90419400e-01 8.20161402e-01
6.99307382e-01 4.85075265e-02 7.15503842e-02 7.28624940e-01
3.01110536e-01 2.92274088e-01 1.64750069e-01 -2.74748027e-01
8.10668170e-01 7.29398847e-01 -3.05082202e-01 -1.33314753e+00
-6.09068036e-01 -3.72590065e-01 -1.17267096e+00 1.99140683e-01
3.18563998e-01 1.88879922e-01 -9.85263586e-01 1.43217480e+00
3.52140963e-01 8.11054483e-02 -6.89811036e-02 9.21901643e-01
1.26850462e+00 8.33457947e-01 3.62742782e-01 -2.78749913e-01
1.04140711e+00 -9.78074253e-01 -3.08692455e-01 5.93187846e-02
9.97401476e-01 -5.47978103e-01 6.80828154e-01 5.21566629e-01
-6.99018419e-01 -5.60592115e-01 -1.26087260e+00 -2.52280504e-01
-6.90270245e-01 -4.83526848e-02 1.00567698e+00 4.97930020e-01
-7.13053644e-01 1.07677209e+00 -5.75840533e-01 -1.07430153e-01
5.39776266e-01 5.48191190e-01 -6.97207212e-01 -1.54835433e-01
-1.14956772e+00 9.09027278e-01 7.80398250e-01 1.39985472e-01
-7.02084124e-01 -8.93123269e-01 -7.24079013e-01 -1.31521206e-02
6.04893744e-01 -6.72774017e-01 3.39599103e-01 -3.56175870e-01
-1.32717693e+00 5.35254180e-01 2.81499140e-02 -5.67975283e-01
4.90076512e-01 1.11828506e-01 -2.71539390e-01 1.90313548e-01
-1.96306780e-01 5.67670643e-01 5.74219525e-01 -1.01157308e+00
3.59186530e-02 -5.12249887e-01 1.56548060e-02 3.23503137e-01
-2.20839214e-02 -6.30215883e-01 -4.58319008e-01 -5.18361151e-01
1.58083782e-01 -8.70606601e-01 -4.04491603e-01 -9.73430052e-02
-6.35899603e-01 -1.39444023e-01 5.34142971e-01 -5.23987412e-01
1.14167094e+00 -1.77503490e+00 5.26234567e-01 2.73245722e-01
5.98705351e-01 3.62664104e-01 -2.78923035e-01 6.05927825e-01
-5.69119751e-01 3.37125063e-01 -1.60633191e-01 -5.09583056e-02
-2.80964047e-01 -5.01134954e-02 -1.52029350e-01 4.45128351e-01
1.09114856e-01 1.05508447e+00 -8.39903474e-01 -4.18482929e-01
2.36731067e-01 6.77703738e-01 -6.82192087e-01 3.36449146e-02
-3.47766459e-01 4.61554170e-01 -5.96897006e-01 8.05588543e-01
1.05170691e+00 -6.20024800e-01 3.28923911e-01 -5.90584457e-01
-4.01244797e-02 2.84170449e-01 -9.76919949e-01 1.72273505e+00
4.81505841e-02 1.51050881e-01 -4.01970148e-01 -1.09608305e+00
9.66505647e-01 -6.35190904e-02 4.99271661e-01 -8.39973092e-01
-1.09133825e-01 1.80356041e-01 1.96081430e-01 -2.32990235e-01
3.51360470e-01 -2.19469279e-01 3.92945975e-01 -8.90942812e-02
-7.43380561e-02 -3.19076777e-01 1.36352569e-01 3.30522925e-01
8.57323647e-01 2.20879614e-01 4.98140067e-01 -1.75175473e-01
6.34587348e-01 -1.18339315e-01 3.64414543e-01 5.14564395e-01
-5.47911925e-03 4.59958583e-01 8.12984884e-01 -6.41683578e-01
-7.70215273e-01 -7.86656141e-01 -9.96000022e-02 5.93197107e-01
2.91955560e-01 -9.18865323e-01 -3.79560918e-01 -9.12330925e-01
1.89063832e-01 3.98012042e-01 -4.57598090e-01 -2.15920866e-01
-4.85427290e-01 -1.03264892e+00 1.85653090e-01 2.89656967e-01
4.19766366e-01 -8.27610791e-01 1.42340884e-01 3.22778165e-01
2.71390736e-01 -1.04494298e+00 -3.66633296e-01 1.62548155e-01
-8.43298435e-01 -1.31950676e+00 -5.92902780e-01 -5.32314122e-01
6.33015096e-01 4.38349217e-01 9.04710472e-01 2.33664364e-01
-5.94912589e-01 -2.53489047e-01 -2.99875617e-01 -1.34618878e-01
-1.05014190e-01 3.03256929e-01 -1.50775556e-02 -2.06479773e-01
8.68604407e-02 -6.30439997e-01 -8.18896472e-01 2.36866280e-01
-5.86884022e-01 3.91842514e-01 5.53848624e-01 7.66138315e-01
8.95454884e-01 -4.48465645e-02 5.33224046e-01 -1.07774436e+00
5.12019753e-01 -4.04465437e-01 -7.41614401e-01 2.14028224e-01
-8.68559122e-01 2.61767626e-01 1.00127125e+00 -2.16028020e-01
-5.46464443e-01 3.04309726e-02 -2.97183067e-01 -4.32744771e-01
1.54904783e-01 7.01509595e-01 -1.38380736e-01 -5.72713315e-01
5.66775799e-01 2.97212720e-01 -7.63255805e-02 -5.52852869e-01
3.99674177e-01 1.68229192e-01 -6.00754172e-02 -7.65609205e-01
6.46101773e-01 9.30092186e-02 7.36914456e-01 -9.42998827e-01
-4.70012605e-01 -2.01006249e-01 -5.58298111e-01 1.94413230e-01
8.61448109e-01 -8.70222032e-01 -1.22156549e+00 2.60166675e-01
-1.03184998e+00 -1.72364905e-01 1.66249111e-01 4.73268479e-01
-2.66242623e-01 9.99536872e-01 -4.94115651e-01 -4.36222702e-01
-6.07745588e-01 -1.55894065e+00 8.68555129e-01 5.35498431e-04
1.30382001e-01 -9.68581378e-01 5.79401478e-03 5.59295058e-01
2.29450747e-01 6.47439003e-01 1.33809376e+00 -5.70731819e-01
-8.25195014e-01 1.43982926e-02 -5.36124229e-01 8.19470882e-02
1.84406877e-01 -1.00672252e-01 -7.28921890e-01 -5.05862355e-01
-4.72945333e-01 -4.07657951e-01 1.07739866e+00 1.81801930e-01
1.42124844e+00 -1.97360948e-01 -5.05723357e-01 8.66426826e-01
1.51097620e+00 2.30402470e-01 5.45273781e-01 -8.10598731e-02
1.37874198e+00 2.99122244e-01 3.88869613e-01 2.62721300e-01
2.50945985e-01 8.76614988e-01 5.91817379e-01 -1.95676044e-01
-2.07597196e-01 -5.90610087e-01 2.02377647e-01 8.53680074e-01
-5.25332868e-01 -3.04553002e-01 -7.88024545e-01 -1.64873824e-01
-1.80533588e+00 -8.20499361e-01 -3.43228042e-01 2.24724436e+00
8.33740413e-01 1.42663315e-01 -4.42757197e-02 -2.04923600e-01
3.79073828e-01 4.30386066e-01 -7.70721972e-01 -2.81431139e-01
-1.22803099e-01 3.87669921e-01 7.16568112e-01 6.78396523e-01
-8.17857146e-01 1.14843869e+00 5.58848953e+00 1.16195118e+00
-1.24839091e+00 -3.79509300e-01 7.15984762e-01 7.53180981e-02
-2.17229366e-01 1.51435239e-02 -9.99079585e-01 6.09257102e-01
7.78843939e-01 4.37254384e-02 4.10098225e-01 8.24352384e-01
1.07265517e-01 7.11406767e-02 -1.20933270e+00 1.42120743e+00
-1.77447293e-02 -1.72237468e+00 7.11660445e-01 4.12285417e-01
5.10909915e-01 1.66221619e-01 -6.89288378e-02 3.59987378e-01
-1.37716621e-01 -1.51836050e+00 7.16223642e-02 3.68781596e-01
8.64787340e-01 -7.98768580e-01 4.16383922e-01 1.48625314e-01
-1.44704449e+00 4.78160530e-01 -7.20462203e-01 -5.51756993e-02
-9.95625630e-02 4.53376114e-01 -9.97124195e-01 1.07312262e+00
1.59750134e-01 1.13373339e+00 -6.73811972e-01 9.76428390e-01
-1.79011434e-01 2.76927948e-01 -1.95990220e-01 -1.07739188e-01
2.26692840e-01 -7.79601932e-01 3.41715246e-01 9.31866884e-01
1.49611607e-01 1.39090642e-01 2.78856128e-01 8.54085088e-01
-4.39605206e-01 3.47359657e-01 -7.77675688e-01 -3.31496954e-01
1.78100392e-01 1.15330255e+00 -5.54430366e-01 -1.95297912e-01
-3.14467311e-01 8.77150834e-01 6.27839804e-01 4.15604681e-01
-9.87342477e-01 -3.87109876e-01 4.83821630e-01 8.25996771e-02
2.94335544e-01 -4.26585197e-01 1.85468987e-01 -1.31328595e+00
-1.11311845e-01 -9.46706891e-01 3.94727916e-01 -6.36587799e-01
-1.32562733e+00 5.42700946e-01 -1.35407418e-01 -8.61927569e-01
2.82876521e-01 -1.22431612e+00 -3.62955213e-01 9.96383846e-01
-1.81027222e+00 -9.33107615e-01 -2.01171264e-01 4.09531355e-01
2.69076705e-01 -3.87538448e-02 8.27302456e-01 3.87356609e-01
-7.99041927e-01 6.97249115e-01 8.13076571e-02 -1.93452865e-01
6.65053129e-01 -1.20864677e+00 3.94987375e-01 2.97269881e-01
1.32485211e-01 6.56293631e-01 4.49338078e-01 -8.21300387e-01
-1.98804557e+00 -9.79932070e-01 5.35359621e-01 -3.87994051e-01
6.16666794e-01 -3.72472584e-01 -1.14757204e+00 2.82529294e-01
-1.72344849e-01 7.01072961e-02 7.10235119e-01 -4.79876902e-03
-6.42054439e-01 -3.39880437e-02 -1.02132785e+00 5.15331388e-01
1.25009441e+00 -4.12912905e-01 -1.43339440e-01 7.20322132e-01
7.83566833e-01 -4.28281188e-01 -1.09364355e+00 5.63791811e-01
2.92553097e-01 -7.40411043e-01 1.27880776e+00 -8.58963668e-01
2.93367743e-01 -5.10753036e-01 -1.04817212e-01 -1.08126044e+00
-2.82110840e-01 -6.92584932e-01 -2.57082015e-01 6.50672197e-01
4.86600995e-01 -8.77261579e-01 8.45795631e-01 3.82383704e-01
-4.81631272e-02 -1.16622818e+00 -9.27557647e-01 -7.92649269e-01
1.34154499e-01 -7.97270536e-02 4.37204987e-01 9.67388570e-01
-7.90122226e-02 8.21118414e-01 -4.77638870e-01 3.83024924e-02
4.84615803e-01 4.14943933e-01 7.50833333e-01 -1.24263835e+00
-5.34718871e-01 -4.75567669e-01 -5.84192574e-01 -1.34867144e+00
5.48748784e-02 -1.30828428e+00 -6.39493763e-01 -1.47866893e+00
4.11604524e-01 -2.99657702e-01 -3.89384955e-01 4.90463346e-01
-1.58791482e-01 -4.04964853e-03 2.89738327e-01 1.59532368e-01
-4.80216742e-01 7.15718508e-01 1.68369317e+00 -5.08318543e-01
-2.52903223e-01 -3.37915063e-01 -6.23583019e-01 2.50413507e-01
6.45083368e-01 -2.40433827e-01 -5.64724088e-01 -7.51153156e-02
3.99025351e-01 5.24478545e-03 1.91005304e-01 -7.96947837e-01
1.61060452e-01 -3.43349725e-01 4.11912978e-01 -7.48196006e-01
5.55157602e-01 -7.13810503e-01 2.37600073e-01 5.67752957e-01
4.05042917e-02 -2.00396731e-01 2.32315257e-01 8.10220063e-01
5.35958372e-02 -9.38450545e-03 6.97209358e-01 -2.09200665e-01
-4.92842317e-01 8.92155051e-01 2.97806293e-01 -1.79397121e-01
9.68863130e-01 -2.30517402e-01 -5.12934029e-01 -1.48243727e-02
-5.21047056e-01 1.93950355e-01 5.08440077e-01 1.94911033e-01
7.54442215e-01 -1.12933123e+00 -4.92799133e-01 2.10152715e-01
9.86758694e-02 3.40625569e-02 2.56484121e-01 7.54292786e-01
-8.05277884e-01 6.56737745e-01 -1.73960060e-01 -4.22850728e-01
-1.26876485e+00 8.51001680e-01 3.19537669e-01 -4.59418595e-01
-5.69071651e-01 6.90445900e-01 3.53606552e-01 -1.01506554e-01
1.41022087e-03 -3.24788004e-01 -1.39838636e-01 7.42833689e-02
3.10489327e-01 2.38140717e-01 2.37628892e-01 -5.26742280e-01
-5.60536981e-01 9.58094954e-01 -5.35708547e-01 7.29277134e-01
1.36353064e+00 3.45383435e-01 -7.03800023e-02 1.34915244e-02
1.43864989e+00 -1.11870296e-01 -1.05373764e+00 -8.28737542e-02
-2.53745526e-01 -3.33957404e-01 -1.55915627e-02 -6.90744281e-01
-1.04859865e+00 1.07365799e+00 4.90898311e-01 5.40301763e-02
7.26675034e-01 -2.06364334e-01 6.88960910e-01 7.32885182e-01
4.09891635e-01 -7.37888098e-01 2.55783405e-02 4.83428627e-01
9.64901567e-01 -1.38385201e+00 6.67691886e-01 -7.02440143e-01
-4.88343805e-01 1.23718286e+00 6.57912493e-01 8.57630670e-02
4.03383136e-01 -4.19317186e-01 -5.20329237e-01 -6.21612489e-01
-5.18528044e-01 2.09040586e-02 6.27529442e-01 4.02913213e-01
5.82733989e-01 1.88992098e-01 -3.56117725e-01 1.36935040e-01
-7.50496089e-02 -2.63835341e-01 1.17123172e-01 7.23327518e-01
-2.82120317e-01 -1.34613466e+00 1.20824955e-01 2.37789035e-01
-2.09209025e-01 -4.30131882e-01 -6.29595399e-01 8.85539770e-01
-6.19763844e-02 6.16308451e-01 -4.86184150e-01 -3.44972521e-01
1.78015694e-01 -2.11954508e-02 6.78610027e-01 -5.76781809e-01
-2.86019325e-01 -1.44774824e-01 1.87269934e-02 -6.37361705e-01
-2.59849608e-01 1.70356870e-01 -1.40934145e+00 -5.09274840e-01
-3.66536468e-01 3.48290741e-01 5.84062815e-01 5.28366506e-01
7.64239490e-01 4.62320298e-01 5.25982440e-01 -6.98042572e-01
-3.90250742e-01 -7.30241537e-01 -5.15852869e-01 3.67796570e-01
2.04852119e-01 -7.54145682e-01 -2.47550637e-01 -3.24620575e-01] | [5.116727828979492, 5.826181888580322] |
0881f095-ffae-480d-b0f1-31865d3c1d9f | network-modeling-and-pathway-inference-from | 1810.00839 | null | http://arxiv.org/abs/1810.00839v1 | http://arxiv.org/pdf/1810.00839v1.pdf | Network Modeling and Pathway Inference from Incomplete Data ("PathInf") | In this work, we developed a network inference method from incomplete data
("PathInf") , as massive and non-uniformly distributed missing values is a
common challenge in practical problems. PathInf is a two-stages inference
model. In the first stage, it applies a data summarization model based on
maximum likelihood to deal with the massive distributed missing values by
transforming the observation-wise items in the data into state matrix. In the
second stage, transition pattern (i.e. pathway) among variables is inferred as
a graph inference problem solved by greedy algorithm with constraints. The
proposed method was validated and compared with the state-of-art Bayesian
network method on the simulation data, and shown consistently superior
performance. By applying the PathInf on the lymph vascular metastasis data, we
obtained the holistic pathways of the lymph node metastasis with novel
discoveries on the jumping metastasis among nodes that are physically apart.
The discovery indicates the possible presence of sentinel node groups in the
lung lymph nodes which have been previously speculated yet never found. The
pathway map can also improve the current dissection examination protocol for
better individualized treatment planning, for higher diagnostic accuracy and
reducing the patients trauma. | ['Quanzheng Li', 'Xing Wang', 'Nan Wu', 'Qitian Chen', 'Xiang Li', 'Ning Guo'] | 2018-10-01 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 3.96934956e-01 5.37859797e-01 -6.17909133e-01 -1.92106694e-01
-3.39207649e-01 -1.68927118e-01 2.69832760e-01 3.30284059e-01
-4.37171496e-02 1.24581385e+00 5.41612208e-01 -6.27709806e-01
-1.06433082e+00 -9.73471165e-01 -4.69241470e-01 -1.04198372e+00
-2.05970332e-01 9.32326019e-01 5.71424142e-02 1.50101244e-01
1.06620781e-01 4.53056246e-01 -9.49883938e-01 1.58615425e-01
7.99593627e-01 4.77946520e-01 2.39152282e-01 5.08861542e-01
-1.02357224e-01 7.62961566e-01 -1.30177930e-01 -2.06538588e-01
-8.31377953e-02 -4.54528600e-01 -5.76027095e-01 -3.48098665e-01
-2.53707588e-01 -3.07716787e-01 -3.03705841e-01 7.89428890e-01
3.69148910e-01 -1.57732084e-01 8.89007688e-01 -1.28365529e+00
2.06310555e-01 1.03211343e+00 -6.67379498e-01 -1.33986115e-01
1.74433872e-01 -2.26589933e-01 7.68886983e-01 -7.07404315e-01
9.61513221e-01 1.23359096e+00 6.80918217e-01 3.03469777e-01
-1.32264304e+00 -6.80276155e-01 -7.23062009e-02 3.95045847e-01
-1.52909422e+00 -1.81915592e-02 6.43831074e-01 -5.20993948e-01
5.01752079e-01 4.14892554e-01 7.41949916e-01 1.17666662e+00
7.86744893e-01 6.59815148e-02 8.46501052e-01 -1.09991409e-01
3.66005391e-01 -2.04606414e-01 2.49877617e-01 7.45046616e-01
7.33792782e-01 2.08233565e-01 -7.28147447e-01 -5.52462339e-01
4.70603853e-01 4.17549103e-01 3.41615938e-02 -5.82040586e-02
-1.31608379e+00 6.97988987e-01 3.41377795e-01 -1.91390701e-02
-7.06270814e-01 1.57953292e-01 1.41526967e-01 -2.78633028e-01
1.66854694e-01 -2.71470785e-01 -3.82058561e-01 3.00906718e-01
-9.08815980e-01 7.50188250e-03 1.04510522e+00 5.74184060e-01
6.87750638e-01 -5.76842844e-01 -4.27971691e-01 2.25161254e-01
7.87672698e-01 3.99894029e-01 -1.01795860e-01 -8.60878050e-01
3.17718387e-01 9.95942235e-01 -1.07422642e-01 -1.20685816e+00
-9.11077678e-01 -4.26995575e-01 -1.47221959e+00 -1.64786175e-01
5.14988363e-01 -3.95229042e-01 -6.68943226e-01 1.75736797e+00
6.29881740e-01 6.02994680e-01 4.70403954e-02 6.01068795e-01
9.69475806e-01 4.73067790e-01 1.46017775e-01 -7.33653784e-01
1.53334141e+00 -3.77285957e-01 -1.11966336e+00 2.70105958e-01
6.33137822e-01 -3.80503327e-01 -5.94676137e-02 6.17598057e-01
-6.43484235e-01 -1.83667690e-01 -6.76992655e-01 4.31613237e-01
-1.44892618e-01 -1.95327085e-02 8.82114947e-01 5.62052786e-01
-7.79428124e-01 4.88395214e-01 -9.73082721e-01 -3.84702235e-01
3.03143024e-01 4.22821492e-01 -3.69298279e-01 -3.40586245e-01
-1.28179717e+00 6.26100659e-01 5.35104871e-01 6.69480443e-01
-1.01552045e+00 -8.51757526e-01 -5.45502305e-01 1.13045052e-01
8.47411573e-01 -1.28321409e+00 5.12544036e-01 5.36417067e-02
-1.11069238e+00 1.48219675e-01 -6.65132821e-01 -1.20769829e-01
3.38374287e-01 4.94027615e-01 -2.27895617e-01 7.99935460e-02
-1.71667963e-01 3.06874514e-01 2.91200787e-01 -1.09122550e+00
-3.78871143e-01 -5.54816604e-01 -6.33595407e-01 -5.75343939e-03
3.69783901e-02 -4.94718552e-01 -2.59567171e-01 -2.93264240e-01
4.76114154e-01 -8.32193375e-01 -6.88421011e-01 -3.13711651e-02
-8.76324832e-01 -1.81460142e-01 5.74402928e-01 -7.78373897e-01
1.44995606e+00 -1.74150097e+00 5.42367160e-01 7.56885052e-01
4.54913497e-01 -4.08144891e-01 2.72222310e-01 8.53697121e-01
6.02405220e-02 2.15718314e-01 -3.64261538e-01 -1.79740772e-01
-4.47736561e-01 5.79084694e-01 1.75660044e-01 4.99370664e-01
-2.01987550e-01 6.99943066e-01 -8.53504777e-01 -1.04234600e+00
9.35733095e-02 1.73297644e-01 -5.52575886e-01 -1.76939685e-02
-2.04140723e-01 7.93199539e-01 -6.40276253e-01 8.34671855e-01
9.29021716e-01 -4.06587064e-01 7.25604117e-01 -4.11311209e-01
3.84700149e-02 -3.70481312e-01 -1.21169281e+00 1.77808630e+00
1.33762702e-01 -4.08085249e-02 2.73690313e-01 -7.61260867e-01
8.11815619e-01 4.19526130e-01 8.92599344e-01 -5.16060777e-02
6.37522191e-02 -3.99461091e-02 3.08702528e-01 -7.53245711e-01
1.55568659e-01 -2.40179762e-01 -9.41245258e-02 1.92772761e-01
-1.81609169e-01 1.75749555e-01 2.83166883e-03 6.60424292e-01
1.51918364e+00 -1.69545561e-01 5.61075032e-01 -3.42497528e-01
4.28294748e-01 2.35794589e-01 9.44421887e-01 1.00184393e+00
1.99262902e-01 2.34700441e-01 1.04744196e+00 6.61387667e-02
-4.25694376e-01 -1.20710814e+00 -3.89448762e-01 2.50732571e-01
1.50136918e-01 -4.50103700e-01 -3.04379970e-01 -5.56807041e-01
3.33379865e-01 4.91672754e-01 -8.21578264e-01 -1.55037940e-01
-8.45748037e-02 -1.28017497e+00 4.08121914e-01 1.22456983e-01
8.23892802e-02 -4.68785167e-01 6.56819716e-02 3.21595490e-01
-4.70208019e-01 -7.69688845e-01 4.67171043e-01 2.41242081e-01
-1.15234792e+00 -1.37063873e+00 -1.39952794e-01 -2.36053705e-01
1.05664957e+00 -2.06823260e-01 6.03871882e-01 2.56153405e-01
-4.81971174e-01 2.97613181e-02 -2.44393483e-01 -2.82861263e-01
-3.72223556e-01 -1.59779891e-01 -1.21788066e-02 8.41432363e-02
-7.97895491e-02 -5.03739238e-01 -7.59343505e-01 4.62889045e-01
-7.53210127e-01 2.90587276e-01 1.00421989e+00 1.01729798e+00
8.12258720e-01 5.63696921e-01 6.36756659e-01 -1.20731282e+00
3.47028643e-01 -1.25101244e+00 -4.43556488e-01 2.38391742e-01
-6.93673611e-01 8.06648210e-02 2.95408934e-01 -1.07456788e-01
-1.25113404e+00 1.88886255e-01 -9.62299202e-03 6.40286058e-02
-2.80143116e-02 1.19608259e+00 -2.47897461e-01 2.35288396e-01
2.90317714e-01 -1.89101566e-02 4.82169300e-01 -3.99651200e-01
3.56247611e-02 3.94315422e-01 2.19602779e-01 -3.71961623e-01
5.64417183e-01 8.03459227e-01 1.15207875e+00 -5.98902225e-01
-5.75698197e-01 -3.67418230e-01 -7.98737109e-01 -4.31938022e-01
8.00109982e-01 -6.95352197e-01 -1.19560981e+00 2.74284720e-01
-9.14366066e-01 -1.73779503e-01 1.20785005e-01 8.80603790e-01
-9.40340906e-02 5.27224898e-01 -5.24742663e-01 -7.07050681e-01
-8.66997018e-02 -9.76706207e-01 7.26801157e-01 -2.34623756e-02
-2.68008918e-01 -1.20705938e+00 3.59757781e-01 2.88917124e-01
2.49537423e-01 5.60749114e-01 1.35388160e+00 -5.90341032e-01
-8.97223473e-01 -1.05575897e-01 -1.68870062e-01 -4.18646663e-01
1.37527972e-01 3.01473022e-01 -3.27873558e-01 -1.51111975e-01
-3.36772680e-01 3.52889746e-01 7.96747088e-01 7.08292544e-01
1.31618547e+00 -2.12990403e-01 -1.14700222e+00 4.64203835e-01
1.39657843e+00 -1.24248929e-01 5.82217634e-01 -3.70630801e-01
5.66084743e-01 8.55424225e-01 5.88236094e-01 6.72062874e-01
4.80088174e-01 5.66916645e-01 8.80157351e-01 -9.44988579e-02
-6.90816790e-02 -3.11656177e-01 -6.88693374e-02 6.82125330e-01
-3.72902937e-02 -5.45668006e-01 -9.09429133e-01 2.72883564e-01
-2.24510574e+00 -9.80609119e-01 -8.91658008e-01 1.95582628e+00
5.02865076e-01 1.75240726e-04 -2.45668724e-01 -1.43523902e-01
6.98688924e-01 -3.61253470e-01 -5.40167332e-01 1.06086046e-01
1.50588185e-01 -1.25258595e-01 6.05350196e-01 6.78375781e-01
-2.23472565e-01 3.27797413e-01 6.61428738e+00 7.24981666e-01
-4.97491598e-01 7.67526180e-02 3.98835272e-01 8.15456733e-02
-5.47132909e-01 5.35181940e-01 -1.01356494e+00 4.68258113e-01
9.02432561e-01 -5.32475188e-02 7.60901794e-02 1.13958739e-01
7.99078643e-01 -7.19149113e-01 -9.24700618e-01 5.01425326e-01
-2.00526372e-01 -1.42267358e+00 3.25462520e-02 5.69080353e-01
5.45018137e-01 -1.71737626e-01 -5.66690624e-01 -1.12448903e-02
4.26879019e-01 -8.89605820e-01 1.47811389e-02 1.25069308e+00
5.31759322e-01 -6.75285399e-01 1.10364783e+00 6.50013983e-01
-9.52086508e-01 -1.26684800e-01 -3.71991992e-01 -1.12185001e-01
6.42393947e-01 1.32653713e+00 -1.32192194e+00 1.24365366e+00
2.85793394e-01 8.07026863e-01 -4.17909801e-01 1.07630491e+00
-2.40368098e-01 8.65915060e-01 -4.98079002e-01 8.38024076e-03
-2.38631338e-01 -2.13257208e-01 7.25575507e-01 7.41467178e-01
5.90849519e-01 1.53198704e-01 -9.94406827e-03 8.89544964e-01
1.55172974e-01 -2.87781984e-01 -5.25621235e-01 3.93658161e-01
6.93599582e-01 1.48363841e+00 -7.55416751e-01 -1.17505774e-01
-1.75458521e-01 3.23577106e-01 1.96314920e-02 4.27383214e-01
-7.36097038e-01 8.74075741e-02 1.42030194e-01 1.43111527e-01
-1.42415771e-02 -4.57992963e-02 -3.30011129e-01 -7.28998423e-01
-4.10328120e-01 -2.66613007e-01 9.58837390e-01 -5.24138570e-01
-1.26241779e+00 -3.74781974e-02 5.02953529e-01 -9.05383050e-01
-2.27992967e-01 -4.64596212e-01 -6.51053190e-01 8.03618610e-01
-1.15223587e+00 -1.05117524e+00 -4.92716521e-01 7.48675048e-01
-1.20937780e-01 1.20432891e-01 8.05135608e-01 7.26071447e-02
-9.85491276e-01 1.87731534e-01 2.37666532e-01 -2.70381421e-01
4.51577455e-01 -9.70244050e-01 -5.18245637e-01 6.70160115e-01
-5.94402432e-01 6.16026640e-01 6.29645944e-01 -1.47558200e+00
-1.70907295e+00 -9.51832056e-01 7.77020395e-01 -2.89751351e-01
6.43224418e-01 -3.82438451e-01 -6.77801430e-01 5.68814933e-01
-1.26202017e-01 -2.14118659e-01 9.76581097e-01 2.98594981e-01
3.97146344e-01 -2.45541126e-01 -1.30460346e+00 5.56242824e-01
9.15783823e-01 1.94332883e-01 -2.73847282e-01 4.67823952e-01
5.01445591e-01 -1.51907429e-01 -1.25013638e+00 6.99291050e-01
5.27546108e-01 -6.46619618e-01 9.40432787e-01 -3.92692715e-01
1.20290995e-01 -5.34032524e-01 -1.31795658e-02 -9.97828841e-01
-5.82682014e-01 -5.55845022e-01 -2.52713323e-01 1.10489976e+00
5.48001051e-01 -5.70519447e-01 1.04450560e+00 4.65013117e-01
-8.65388066e-02 -7.29228556e-01 -1.37022412e+00 -4.25194472e-01
-3.40021819e-01 -2.82745898e-01 3.92378449e-01 6.82249367e-01
1.86700448e-01 1.31020039e-01 -5.38778186e-01 4.92908120e-01
1.12294102e+00 -1.89404532e-01 6.48590863e-01 -1.69667208e+00
-3.94599617e-01 7.02603161e-03 -3.14929426e-01 -5.53935587e-01
6.85499832e-02 -1.17797172e+00 -2.59937674e-01 -1.88229239e+00
8.54299307e-01 -5.14586210e-01 -1.77599877e-01 5.39903343e-01
-2.13434678e-02 -4.42891926e-01 -4.37421322e-01 9.35998186e-02
-3.56054157e-01 3.76541346e-01 1.36932325e+00 4.40971404e-02
-5.53787462e-02 2.29231134e-01 -6.46729708e-01 5.83047211e-01
7.09884703e-01 -9.68650997e-01 -3.58899623e-01 1.81731686e-01
7.29379654e-01 1.03108513e+00 6.48002505e-01 -6.63235486e-01
6.73745632e-01 -4.42272782e-01 3.06101292e-01 -1.52672994e+00
3.16703767e-01 -1.31098747e+00 1.02312362e+00 9.60911512e-01
-3.07542980e-02 -2.66792029e-01 2.46347617e-02 1.15492058e+00
8.73718634e-02 -3.63398671e-01 6.63401708e-02 7.98528865e-02
-2.58377790e-01 2.23792806e-01 -6.39261663e-01 -5.98924339e-01
1.29712486e+00 -8.83064568e-02 -4.44532901e-01 -2.42125064e-01
-1.15536857e+00 5.88377357e-01 -1.48656920e-01 -2.61848241e-01
7.01261044e-01 -1.09269881e+00 -1.01992881e+00 1.71570182e-01
-1.43802717e-01 3.17659765e-01 8.50728571e-01 1.49303150e+00
-2.96466231e-01 2.27900013e-01 -3.93785862e-03 -8.38250101e-01
-1.25175202e+00 4.44223374e-01 1.37797937e-01 -6.94341302e-01
-4.25882876e-01 4.19902503e-01 -1.19351164e-01 -4.76967663e-01
1.66999221e-01 -1.44223556e-01 -1.87748015e-01 2.88612455e-01
1.05128147e-01 8.32392871e-01 -1.01147994e-01 -8.56461599e-02
-5.03618896e-01 3.59603912e-01 -6.09633550e-02 4.32472788e-02
1.28158462e+00 -4.26130503e-01 -7.90278375e-01 3.50085527e-01
7.42147565e-01 2.19592974e-02 -9.43335414e-01 -1.04102023e-01
-1.89346284e-01 -1.95127621e-01 5.89725794e-03 -1.08139265e+00
-8.85092795e-01 3.59965324e-01 2.01718405e-01 -1.43999025e-01
9.36534464e-01 2.33025089e-01 1.84216246e-01 3.74348402e-01
3.06360126e-01 -6.06239676e-01 -2.89997876e-01 2.46299878e-01
5.41740835e-01 -9.93388712e-01 4.97712851e-01 -7.33498454e-01
-2.00078323e-01 1.11952567e+00 3.15635383e-01 3.59952390e-01
1.15119064e+00 3.57239157e-01 -4.52334672e-01 -5.48930883e-01
-1.09225178e+00 2.95586586e-01 9.74126086e-02 3.51329565e-01
2.30442770e-02 2.49973223e-01 -6.67281628e-01 7.19021499e-01
9.61781219e-02 1.82346061e-01 6.24690235e-01 5.24627090e-01
-2.15752766e-01 -1.28691053e+00 -5.69606245e-01 7.83305049e-01
-2.68019676e-01 -5.56482002e-03 -4.25213307e-01 8.97386909e-01
1.48459300e-01 1.06495726e+00 -1.61768645e-01 -2.80093938e-01
1.65075600e-01 -1.06957868e-01 2.18256056e-01 -3.49927515e-01
-3.53016734e-01 3.69348899e-02 2.28976071e-01 -4.05495703e-01
-3.63636851e-01 -7.95817673e-01 -1.33448410e+00 -3.75192046e-01
-4.73577440e-01 2.82477587e-01 7.88288653e-01 8.78374219e-01
3.47100854e-01 9.54947531e-01 4.72576499e-01 -4.45153534e-01
-3.80033970e-01 -8.06008875e-01 -7.45995641e-01 -4.61897701e-02
1.29469067e-01 -7.68543899e-01 -5.34074068e-01 -3.97384554e-01] | [7.7691192626953125, 5.319459438323975] |
0c616e5f-591f-4ef9-9f91-5f3049f8fca6 | ampose-alternatively-mixed-global-local | 2210.04216 | null | https://arxiv.org/abs/2210.04216v4 | https://arxiv.org/pdf/2210.04216v4.pdf | AMPose: Alternatively Mixed Global-Local Attention Model for 3D Human Pose Estimation | The graph convolutional networks (GCNs) have been applied to model the physically connected and non-local relations among human joints for 3D human pose estimation (HPE). In addition, the purely Transformer-based models recently show promising results in video-based 3D HPE. However, the single-frame method still needs to model the physically connected relations among joints because the feature representations transformed only by global relations via the Transformer neglect information on the human skeleton. To deal with this problem, we propose a novel method in which the Transformer encoder and GCN blocks are alternately stacked, namely AMPose, to combine the global and physically connected relations among joints towards HPE. In the AMPose, the Transformer encoder is applied to connect each joint with all the other joints, while GCNs are applied to capture information on physically connected relations. The effectiveness of our proposed method is evaluated on the Human3.6M dataset. Our model also shows better generalization ability by testing on the MPI-INF-3DHP dataset. Code can be retrieved at https://github.com/erikervalid/AMPose. | ['PeiYuan Wu', 'Yunwei Chiu', 'Hongxin Lin'] | 2022-10-09 | null | null | null | null | ['3d-human-pose-estimation'] | ['computer-vision'] | [-4.85582203e-01 1.74772710e-01 7.35225007e-02 -7.40686432e-02
-1.19700752e-01 -3.06528658e-02 3.53276938e-01 -3.87043089e-01
-1.67898938e-01 3.78637880e-01 2.12801501e-01 1.92828581e-01
-2.09021166e-01 -8.26392412e-01 -8.46976638e-01 -5.11357903e-01
-3.15418005e-01 5.77088356e-01 5.50697505e-01 -3.30295742e-01
-3.67189735e-01 5.11794031e-01 -1.15552139e+00 -5.99275455e-02
4.77162331e-01 8.77543628e-01 -2.78933975e-03 5.68753660e-01
2.51155287e-01 8.46203208e-01 -3.04869682e-01 -3.02934468e-01
2.62469083e-01 -4.25407410e-01 -6.95414603e-01 -9.15806834e-03
3.02648127e-01 -5.83939016e-01 -9.72340286e-01 6.66302979e-01
6.88731134e-01 2.55515665e-01 3.99258018e-01 -1.41584766e+00
-2.92715997e-01 2.10712001e-01 -6.54183030e-01 -6.59713075e-02
5.59987962e-01 5.68541475e-02 7.73280144e-01 -7.93276846e-01
7.24548757e-01 1.48351812e+00 6.41615391e-01 4.89098668e-01
-6.04217708e-01 -5.20495355e-01 2.54922528e-02 5.37115633e-01
-1.39970624e+00 -1.39166951e-01 9.24486458e-01 -2.42061451e-01
9.31719601e-01 7.54740909e-02 1.13076282e+00 1.00796723e+00
5.47792077e-01 8.06034207e-01 6.15145504e-01 -1.29146725e-01
-1.72044113e-01 -6.24580264e-01 1.00947887e-01 1.09786093e+00
4.16050628e-02 9.25647914e-02 -6.90155447e-01 -3.54300579e-03
1.17842829e+00 -3.89761105e-02 -3.16989064e-01 -5.93335748e-01
-1.26577437e+00 4.96056557e-01 8.94150198e-01 -3.16873305e-02
-4.35181260e-01 5.95635056e-01 3.62599611e-01 -1.29074812e-01
3.54848385e-01 -1.17225632e-01 -3.73160899e-01 -8.77840891e-02
-5.41301787e-01 4.02706236e-01 7.43174732e-01 1.03654814e+00
5.68202138e-01 -2.30181560e-01 -5.08633479e-02 8.49407613e-01
5.54558933e-01 4.38376427e-01 1.05105959e-01 -1.11545432e+00
5.48600852e-01 6.29516065e-01 -1.25990227e-01 -1.22957420e+00
-7.38437951e-01 -4.84996200e-01 -1.06583798e+00 4.79222164e-02
2.68463314e-01 -8.25894922e-02 -9.70805883e-01 1.60533977e+00
6.30313933e-01 2.11971924e-01 -3.22396040e-01 1.29771042e+00
9.94222343e-01 6.80295348e-01 -1.97961360e-01 3.27025145e-01
1.45042968e+00 -1.15887868e+00 -6.49941206e-01 -5.95901236e-02
5.17096519e-01 -6.28197551e-01 6.27428412e-01 1.86563239e-01
-1.17868614e+00 -6.75121665e-01 -9.48653340e-01 -3.98286670e-01
-1.80003569e-02 2.26389199e-01 4.23909932e-01 2.31868457e-02
-9.22167718e-01 4.91904080e-01 -1.32357287e+00 -4.12328184e-01
2.17640787e-01 4.33775634e-01 -6.15951955e-01 -1.51440293e-01
-1.47274065e+00 9.50939536e-01 2.36043513e-01 6.51437402e-01
-7.93048620e-01 -2.17841998e-01 -9.51841295e-01 -2.85412427e-02
5.61017096e-01 -1.06766498e+00 9.56522346e-01 -2.99895376e-01
-1.57100534e+00 3.71865183e-01 1.42961308e-01 -3.19113284e-01
6.77087665e-01 -5.78429759e-01 -7.48367310e-02 4.85549986e-01
-1.69193819e-01 8.67785156e-01 4.34023529e-01 -9.90704238e-01
-2.55918294e-01 -3.91610533e-01 4.31865513e-01 5.26583374e-01
3.73819359e-02 -3.02313030e-01 -1.04808545e+00 -6.05238020e-01
3.17820400e-01 -1.23027432e+00 -1.09157331e-01 3.38418156e-01
-5.81481457e-01 -2.82084167e-01 9.40091491e-01 -9.14430678e-01
1.01254237e+00 -1.98957789e+00 7.72419691e-01 1.95401996e-01
3.69336724e-01 1.86417222e-01 -2.44650105e-03 4.72085297e-01
1.42748922e-01 -3.07353169e-01 -7.91087672e-02 -3.31816912e-01
1.14843054e-02 5.04249632e-01 4.61464792e-01 5.07180154e-01
1.12174928e-01 9.62017715e-01 -7.19988346e-01 -8.03924799e-01
4.88876224e-01 9.57091510e-01 -5.43726921e-01 1.83107093e-01
-3.31431441e-02 4.88906145e-01 -6.23866081e-01 5.84120631e-01
7.40395308e-01 -2.01829582e-01 1.27789930e-01 -5.05101681e-01
2.65373826e-01 2.13817120e-01 -1.15854752e+00 1.96175480e+00
-2.06961840e-01 1.69684738e-01 9.68758762e-02 -7.63327003e-01
7.34934509e-01 4.24148470e-01 7.24300802e-01 -6.26554251e-01
2.39150643e-01 -2.39231978e-02 -4.46141921e-02 -4.63091165e-01
1.22511223e-01 1.75740451e-01 -5.24774082e-02 -3.31840813e-02
1.44958735e-01 1.61685973e-01 3.27959508e-02 2.88461715e-01
1.08394539e+00 7.37796664e-01 1.79907873e-01 -1.18554376e-01
5.81439793e-01 -1.00592263e-01 6.62141621e-01 1.21775314e-01
-1.79065824e-01 6.31151438e-01 5.54585457e-01 -4.39782172e-01
-8.75210822e-01 -1.31153941e+00 2.86143780e-01 5.24362266e-01
4.77971554e-01 -7.00864255e-01 -7.97776818e-01 -5.92476070e-01
-5.03355265e-02 6.57949522e-02 -4.23626184e-01 -3.03375483e-01
-8.29489052e-01 -3.59680057e-01 5.08403897e-01 6.76410437e-01
9.02181029e-01 -8.02109718e-01 -8.85772824e-01 1.83160976e-01
-5.10847032e-01 -1.22857785e+00 -3.59557748e-01 -1.07004166e-01
-7.23412752e-01 -1.28232121e+00 -8.74886394e-01 -6.93762124e-01
4.67627585e-01 2.33491976e-02 8.02074969e-01 2.97934026e-01
-1.48481578e-01 3.43875378e-01 -3.81056875e-01 1.61700070e-01
1.22323841e-01 1.47847131e-01 -3.99583299e-03 -2.47582912e-01
-1.05263814e-01 -6.27816379e-01 -8.26187491e-01 7.40012646e-01
-6.66558981e-01 3.92351627e-01 4.80732322e-01 8.34924340e-01
5.71502566e-01 1.04450518e-02 1.01821236e-01 -4.44281429e-01
1.53220356e-01 -3.06659430e-01 -2.02718124e-01 1.15990832e-01
1.85010478e-01 -4.58154269e-02 3.99573296e-01 -3.03862333e-01
-9.81317282e-01 3.30741018e-01 -4.63444382e-01 -6.64055467e-01
-1.39502725e-02 4.87746626e-01 -4.63804632e-01 2.06893552e-02
2.61424124e-01 -7.82688335e-02 7.42502287e-02 -5.76740384e-01
2.10058346e-01 2.37480223e-01 5.06979704e-01 -6.99324846e-01
6.45864964e-01 3.77626330e-01 4.47174877e-01 -7.30324507e-01
-4.24961597e-01 -2.41045162e-01 -8.28383088e-01 -4.85065699e-01
9.77762878e-01 -9.60057139e-01 -7.14482725e-01 7.88701057e-01
-1.26333332e+00 -4.06963438e-01 5.79180615e-03 7.51960695e-01
-6.51007593e-01 5.70285976e-01 -1.01285577e+00 -5.17877579e-01
-2.65129328e-01 -1.27064395e+00 1.24149466e+00 3.37405950e-02
-1.73249066e-01 -7.88961172e-01 -9.05602500e-02 3.42891127e-01
-1.82862245e-02 5.70792556e-01 7.39142954e-01 -6.17137179e-02
-5.44438183e-01 -2.25060567e-01 -1.24659218e-01 1.81099176e-01
4.29035611e-02 -3.66478786e-02 -4.37688768e-01 -1.92518115e-01
-2.03953385e-01 -1.18687265e-01 6.23105049e-01 3.15515071e-01
8.00975263e-01 -5.94389699e-02 -3.68079275e-01 5.66312730e-01
1.01108277e+00 8.52370635e-03 8.62045109e-01 2.08884522e-01
1.21178329e+00 6.41608357e-01 5.42251706e-01 3.86240810e-01
5.96995354e-01 9.69271302e-01 6.07231200e-01 -9.28271040e-02
-4.17618662e-01 -3.38727862e-01 3.81472141e-01 1.06893504e+00
-5.86726189e-01 -4.40976858e-01 -9.19730186e-01 2.13743031e-01
-1.98502266e+00 -5.23339152e-01 -4.43860352e-01 1.87645125e+00
4.06464607e-01 1.98988348e-01 1.94392696e-01 6.40896857e-02
7.28738010e-01 1.50861382e-01 -3.86746675e-01 6.67524487e-02
1.32980436e-01 1.36455849e-01 2.82135397e-01 4.69025880e-01
-9.26505566e-01 7.48165369e-01 5.03223133e+00 7.51045167e-01
-7.14248896e-01 1.65264994e-01 1.96369037e-01 -1.44802257e-01
2.23564923e-01 3.27678658e-02 -6.44982636e-01 2.50515044e-01
5.40806949e-01 3.17104429e-01 1.15930490e-01 5.20172119e-01
2.05738500e-01 -2.39527598e-01 -9.57560956e-01 8.41869175e-01
-2.00867772e-01 -7.75212288e-01 1.61469936e-01 8.51713791e-02
2.71503925e-01 -1.72981128e-01 -2.49834210e-01 3.72283347e-03
-3.68960574e-02 -6.94456100e-01 8.41859460e-01 8.26832771e-01
5.71067393e-01 -7.96678245e-01 8.38542879e-01 4.36606109e-01
-1.62468934e+00 2.24090174e-01 -2.63055205e-01 -1.60495147e-01
4.01023537e-01 5.68106651e-01 -3.74156535e-01 1.06348240e+00
8.42328191e-01 8.12474787e-01 -4.68390614e-01 1.00044155e+00
-4.68592972e-01 2.17307046e-01 -5.81845760e-01 2.58314282e-01
1.04176894e-01 -1.07078232e-01 6.27356708e-01 7.83461571e-01
3.64582896e-01 1.43579140e-01 1.06997520e-01 5.95332086e-01
1.55803055e-01 -1.32723212e-01 -3.78773957e-01 2.48309091e-01
5.69620766e-02 1.15947330e+00 -6.74176037e-01 -2.41531983e-01
-3.94879609e-01 1.10000849e+00 2.75743514e-01 3.38993460e-01
-1.17811918e+00 -2.20092222e-01 4.70856428e-01 3.74509752e-01
1.47990853e-01 -7.55388141e-01 2.60954082e-01 -1.12130201e+00
2.79024035e-01 -6.49127424e-01 3.96358728e-01 -1.02424300e+00
-1.06819391e+00 4.92246389e-01 3.86945784e-01 -1.27027535e+00
-5.25909476e-02 -5.32519519e-01 -2.41032168e-01 6.76319838e-01
-1.06473875e+00 -1.37103188e+00 -5.09915769e-01 8.31136942e-01
3.17223817e-01 3.93324405e-01 4.68972534e-01 4.32204634e-01
-6.09532237e-01 3.41597825e-01 -4.93750930e-01 3.92355412e-01
4.94319558e-01 -9.22048569e-01 4.98332292e-01 6.07377350e-01
-8.79443809e-02 5.29946804e-01 3.50472420e-01 -9.43310261e-01
-1.42081416e+00 -9.33592439e-01 6.91385448e-01 -1.00311726e-01
3.55758339e-01 -4.67868388e-01 -8.20835412e-01 7.94060111e-01
1.92207955e-02 2.72645473e-01 1.30911723e-01 -1.70905352e-01
-1.83624074e-01 -6.56963214e-02 -8.30050409e-01 4.85453695e-01
1.46021962e+00 -1.86160177e-01 -6.07816339e-01 2.92295069e-01
6.77326024e-01 -9.59901631e-01 -1.15697181e+00 5.71151972e-01
8.41065168e-01 -8.50274682e-01 1.18144083e+00 -1.92422763e-01
6.36169612e-01 -5.90262413e-01 -4.93273623e-02 -1.16247487e+00
-2.57307857e-01 -2.02976719e-01 -4.11591172e-01 8.48887026e-01
-9.18478146e-02 -5.24016619e-01 7.97146976e-01 2.77005583e-01
-2.56883502e-01 -9.66005802e-01 -1.17544186e+00 -8.40476036e-01
-1.95653081e-01 -1.21694215e-01 4.52676803e-01 6.18412077e-01
-5.66438735e-02 3.28032255e-01 -6.95289016e-01 2.54971951e-01
6.30068481e-01 -2.94301510e-01 8.87462795e-01 -1.05989671e+00
-4.93119419e-01 -6.68367296e-02 -7.91286707e-01 -1.46618414e+00
-6.27257973e-02 -6.95080221e-01 1.43350353e-02 -1.79636633e+00
3.34119536e-02 -2.38087252e-01 -1.29899263e-01 4.62881148e-01
4.73105125e-02 1.48265406e-01 4.15408522e-01 8.40810388e-02
-5.86982548e-01 7.51553059e-01 1.70574343e+00 -1.04936967e-02
8.47608130e-03 -1.26229376e-01 1.46190152e-01 8.26682210e-01
7.48774350e-01 -2.82877654e-01 -4.73642975e-01 -5.81134319e-01
1.25204206e-01 4.57807362e-01 9.30036724e-01 -1.44629169e+00
4.38642949e-01 1.76586509e-01 6.26838744e-01 -8.71042907e-01
7.20657587e-01 -8.94716620e-01 7.34455585e-01 7.31150329e-01
-3.03418785e-02 1.34172797e-01 1.84220150e-02 5.35127521e-01
-1.97272718e-01 2.22274542e-01 4.70710754e-01 -2.40408778e-01
-6.36717379e-01 6.19397283e-01 -1.70979932e-01 -1.67507663e-01
9.00708318e-01 -5.40631376e-02 -1.87151834e-01 -3.58754307e-01
-1.07927537e+00 4.20070887e-01 3.77347618e-01 3.46483886e-01
7.91277289e-01 -1.60869217e+00 -4.73802894e-01 1.44217368e-02
-7.86184743e-02 4.68024731e-01 6.07032180e-01 1.00939429e+00
-9.36131597e-01 2.33616114e-01 -4.79667604e-01 -7.47903466e-01
-1.21503580e+00 3.82498533e-01 6.24954760e-01 -3.48479569e-01
-9.06103492e-01 8.43178272e-01 2.77743965e-01 -4.49392766e-01
1.53447434e-01 -3.05188268e-01 -2.29914747e-02 -2.94524908e-01
-4.92618792e-02 7.19501317e-01 -7.34923705e-02 -9.11126196e-01
-5.75068057e-01 9.53922808e-01 2.67532885e-01 -7.33877793e-02
1.17924142e+00 -1.65770710e-01 -1.48024395e-01 2.31694981e-01
1.33659458e+00 -3.07065487e-01 -1.40946710e+00 -1.25265509e-01
-3.75940353e-01 -3.42678040e-01 -1.57690331e-01 -4.96860236e-01
-1.44281292e+00 1.09588504e+00 4.62974280e-01 -1.66977525e-01
1.15352333e+00 5.39799891e-02 1.16171825e+00 2.22663924e-01
6.77336216e-01 -9.23406184e-01 1.65782094e-01 4.83641148e-01
1.00545359e+00 -7.07604289e-01 2.54430830e-01 -8.73205185e-01
-4.39087480e-01 1.13727152e+00 8.60197008e-01 -4.08045769e-01
8.22938204e-01 1.00616522e-01 -3.15023601e-01 -4.42790270e-01
-4.84571546e-01 -4.33711261e-02 4.12635267e-01 5.06062508e-01
4.03220862e-01 9.90809426e-02 -4.12951529e-01 3.96482259e-01
-1.70362920e-01 1.76333919e-01 1.61483198e-01 1.06655383e+00
-1.42320842e-01 -1.21284711e+00 -3.59715790e-01 1.95679396e-01
-1.69780791e-01 2.39681080e-01 -3.94169331e-01 1.09850788e+00
3.28235507e-01 7.15560377e-01 -2.85977237e-02 -6.23189270e-01
6.75346315e-01 -7.77508095e-02 7.31022596e-01 -3.73933226e-01
-3.63667309e-01 4.09060210e-01 2.75956094e-01 -9.48810697e-01
-6.52865350e-01 -5.23527384e-01 -1.58986366e+00 -3.92331868e-01
-1.46766961e-01 1.46832494e-02 3.02705914e-01 8.03055346e-01
3.72110903e-01 7.44174063e-01 1.71892211e-01 -1.01782203e+00
-1.42877623e-01 -8.35066199e-01 -5.71247816e-01 4.11693156e-01
1.28003746e-01 -1.24533427e+00 -9.60187465e-02 -2.15193570e-01] | [7.153324127197266, -0.5972241759300232] |
122dffed-124e-486c-a67c-fcbfbc594b27 | dialogue-discourse-aware-graph-convolutional | 2012.03502 | null | https://arxiv.org/abs/2012.03502v2 | https://arxiv.org/pdf/2012.03502v2.pdf | Dialogue Discourse-Aware Graph Model and Data Augmentation for Meeting Summarization | Meeting summarization is a challenging task due to its dynamic interaction nature among multiple speakers and lack of sufficient training data. Existing methods view the meeting as a linear sequence of utterances while ignoring the diverse relations between each utterance. Besides, the limited labeled data further hinders the ability of data-hungry neural models. In this paper, we try to mitigate the above challenges by introducing dialogue-discourse relations. First, we present a Dialogue Discourse-Dware Meeting Summarizer (DDAMS) to explicitly model the interaction between utterances in a meeting by modeling different discourse relations. The core module is a relational graph encoder, where the utterances and discourse relations are modeled in a graph interaction manner. Moreover, we devise a Dialogue Discourse-Aware Data Augmentation (DDADA) strategy to construct a pseudo-summarization corpus from existing input meetings, which is 20 times larger than the original dataset and can be used to pretrain DDAMS. Experimental results on AMI and ICSI meeting datasets show that our full system can achieve SOTA performance. Our codes will be available at: https://github.com/xcfcode/DDAMS. | ['Xinwei Geng', 'Bing Qin', 'Xiaocheng Feng', 'Xiachong Feng'] | 2020-12-07 | null | null | null | null | ['meeting-summarization'] | ['natural-language-processing'] | [ 3.10959518e-01 8.17085445e-01 -1.25576958e-01 -6.61417484e-01
-7.34833241e-01 -3.63580555e-01 7.58783579e-01 2.02949658e-01
-8.58229399e-03 5.71587503e-01 9.53471661e-01 -2.67636508e-01
3.72520179e-01 -5.73325098e-01 -5.82936406e-01 -7.86665231e-02
1.11200869e-01 6.94842160e-01 8.20430070e-02 -5.50673485e-01
-6.22223765e-02 -2.94175833e-01 -1.15157378e+00 6.32811844e-01
1.12186313e+00 4.16035891e-01 2.70650357e-01 8.82505596e-01
-3.30669641e-01 1.25756383e+00 -1.05190802e+00 -3.35470259e-01
-1.83856696e-01 -1.01798296e+00 -1.10600686e+00 4.23684567e-01
1.74464390e-01 -5.21229684e-01 -6.52449071e-01 6.37559116e-01
4.79278952e-01 3.87222409e-01 5.29519022e-01 -1.01373911e+00
-7.06768572e-01 1.34626174e+00 -2.90860832e-01 4.63688821e-02
6.72398508e-01 -2.35066667e-01 1.14422619e+00 -6.69711649e-01
5.84202766e-01 1.49800825e+00 1.83607474e-01 7.39241838e-01
-9.10031557e-01 -3.88193637e-01 5.18060327e-01 3.19249272e-01
-1.07788074e+00 -7.82434821e-01 1.05493701e+00 -1.59838289e-01
1.13656151e+00 6.18946314e-01 5.23987651e-01 1.14445090e+00
-2.03373328e-01 1.13199043e+00 2.63937593e-01 -5.27382612e-01
-5.40073030e-02 -1.18280038e-01 5.53288937e-01 5.68900347e-01
-1.76084742e-01 -7.88293183e-01 -6.38941646e-01 3.35688852e-02
4.49322790e-01 -1.78411931e-01 -4.78688627e-01 1.16424179e-02
-1.05981851e+00 7.55202293e-01 3.72103274e-01 2.89574623e-01
-1.69052258e-01 -2.06321403e-01 6.48234248e-01 4.82639700e-01
6.10629439e-01 5.03718019e-01 -1.18086085e-01 -1.64321005e-01
-5.43344855e-01 2.82332063e-01 1.16459358e+00 1.16712356e+00
4.03297096e-01 4.59463373e-02 -3.27277005e-01 1.06287134e+00
2.29548097e-01 1.77400991e-01 4.71246094e-01 -8.32557619e-01
9.13047493e-01 8.05194259e-01 -2.71798611e-01 -9.36677456e-01
-4.68258917e-01 4.15216945e-02 -9.62839544e-01 -7.48654842e-01
1.98458172e-02 -3.80156934e-01 -6.03677750e-01 1.69169700e+00
3.23412299e-01 6.15869239e-02 4.15432870e-01 6.78479195e-01
1.64402103e+00 1.00769031e+00 -1.93680182e-01 -5.39631009e-01
1.20298553e+00 -1.38101661e+00 -1.26638758e+00 -2.97397226e-01
8.78953815e-01 -4.55007434e-01 1.05690086e+00 5.01919910e-02
-1.31760502e+00 -5.06721377e-01 -9.92798924e-01 -4.23643976e-01
1.03084318e-01 2.46713515e-02 3.42128545e-01 3.21219563e-02
-9.74906385e-01 2.04116255e-01 -8.96851659e-01 -2.81209648e-01
1.15673214e-01 2.71184027e-01 1.92295685e-02 6.68900236e-02
-1.34273612e+00 8.49293292e-01 5.27703524e-01 2.17717066e-01
-5.11511683e-01 -3.39373738e-01 -1.37858546e+00 1.55020088e-01
7.18829751e-01 -5.29441118e-01 1.69851875e+00 -7.42679238e-01
-1.81773472e+00 4.19130296e-01 -4.09144670e-01 -4.94193524e-01
2.93875635e-01 -2.04962552e-01 -2.14008614e-01 -1.77840255e-02
-1.65271118e-01 3.99538875e-01 2.94428110e-01 -1.34802938e+00
-3.68259162e-01 -2.22545385e-01 3.37106138e-01 7.88292825e-01
-3.02689314e-01 -5.35817482e-02 -6.67449534e-01 -3.95135313e-01
1.18810050e-02 -8.84002149e-01 -1.20881245e-01 -1.04022694e+00
-1.00642359e+00 -6.23231351e-01 7.80847728e-01 -8.48693132e-01
1.62090933e+00 -2.00289583e+00 6.26201153e-01 -2.95214802e-01
3.51584285e-01 4.24125701e-01 -2.19591707e-01 9.08250570e-01
8.46410915e-02 2.55304519e-02 -2.82427430e-01 -7.25520015e-01
-3.02623939e-02 2.27678478e-01 -2.09964469e-01 -9.75029636e-03
2.53999352e-01 1.01551485e+00 -7.65875697e-01 -5.36978066e-01
7.07517490e-02 2.63751268e-01 -4.04559344e-01 6.88283324e-01
-5.56373477e-01 6.48370385e-01 -5.87576210e-01 2.31400385e-01
4.01687890e-01 -3.17361385e-01 6.19086802e-01 -1.73907563e-01
9.98639092e-02 8.92454863e-01 -8.70718002e-01 1.83705640e+00
-4.05413747e-01 6.60169303e-01 4.16800044e-02 -1.06908953e+00
1.05396342e+00 4.89385247e-01 2.60950923e-01 -5.25884271e-01
3.13455403e-01 -2.20127046e-01 1.73045769e-01 -6.77593529e-01
8.46827447e-01 1.62415534e-01 -3.57291639e-01 6.38940692e-01
1.32332951e-01 -3.29380870e-01 4.53588992e-01 5.63231468e-01
1.10465503e+00 -2.52214670e-01 4.43889886e-01 1.71261802e-01
6.62906408e-01 6.14035921e-03 5.54592490e-01 5.11398852e-01
-3.17415372e-02 5.84975839e-01 7.09767222e-01 -9.24623981e-02
-7.52793193e-01 -6.40806317e-01 2.66553938e-01 1.23579800e+00
2.22410157e-01 -8.18128884e-01 -8.10954690e-01 -6.61651671e-01
-3.49959284e-01 1.01726043e+00 -4.03620541e-01 -1.54714152e-01
-1.07477367e+00 -3.77080321e-01 4.55384910e-01 4.15479928e-01
4.61556315e-01 -1.13133013e+00 -1.50329128e-01 1.58718735e-01
-7.48907208e-01 -1.28485799e+00 -7.23108888e-01 -1.02075115e-01
-5.27350426e-01 -9.41478252e-01 -2.99794793e-01 -9.47503686e-01
5.48402488e-01 3.89445782e-01 1.15598285e+00 1.06436282e-01
1.93387464e-01 2.83541024e-01 -6.21173859e-01 -5.50538063e-01
-9.59794223e-01 4.34751719e-01 -1.64262444e-01 -2.21693695e-01
3.31930220e-01 -4.14320022e-01 -2.69505739e-01 -9.44660157e-02
-8.71206105e-01 6.34465456e-01 2.23048314e-01 8.18251729e-01
2.16212109e-01 -1.72830492e-01 1.01705158e+00 -1.30915654e+00
1.08751631e+00 -4.22069639e-01 -2.52430830e-02 3.74793261e-01
-5.17150946e-03 -1.35507420e-01 6.27802491e-01 -3.65285516e-01
-1.35098410e+00 -1.04829684e-01 -1.02846466e-01 -1.43006928e-02
-7.32813478e-02 8.75337541e-01 -4.27240491e-01 7.89095581e-01
4.31992918e-01 3.10250856e-02 1.44356698e-01 -2.60230750e-01
5.79730034e-01 1.03589058e+00 5.17112970e-01 -3.99182320e-01
4.36213106e-01 -1.04364716e-01 -6.65347278e-01 -1.13897610e+00
-1.10446870e+00 -5.18697381e-01 -7.04151690e-01 -2.68619359e-01
8.88422012e-01 -1.01113486e+00 -6.34051442e-01 1.93457961e-01
-1.40159047e+00 -5.82422554e-01 -1.95252955e-01 3.23284984e-01
-4.69683796e-01 4.89556372e-01 -8.78526390e-01 -8.57097030e-01
-5.14961720e-01 -9.48528349e-01 8.41600478e-01 3.06554556e-01
-6.14502490e-01 -9.82892215e-01 7.35813454e-02 8.22744489e-01
9.97798070e-02 2.35386044e-01 8.35555553e-01 -1.10500240e+00
-2.80276746e-01 2.62873117e-02 6.83837682e-02 2.82784641e-01
4.84134436e-01 2.71243211e-02 -7.32848465e-01 -2.10372135e-01
1.01357818e-01 -5.58669925e-01 7.10552275e-01 3.25654209e-01
9.25074220e-01 -6.01623535e-01 -2.19490454e-01 9.99078602e-02
6.68712020e-01 2.64457315e-01 3.21202070e-01 -1.33775137e-02
1.12085187e+00 8.86939287e-01 4.70345408e-01 3.66269320e-01
1.12772000e+00 5.77655196e-01 1.69330224e-01 -2.62167994e-02
-2.76240289e-01 -2.36540705e-01 3.80878478e-01 1.84766316e+00
1.07961901e-01 -6.01168215e-01 -1.01840425e+00 5.62069774e-01
-2.18900061e+00 -8.64628375e-01 -2.60490805e-01 1.75253582e+00
1.22913468e+00 -1.05667878e-02 9.27205086e-02 -1.43783927e-01
7.03029513e-01 4.28869069e-01 -5.33024073e-01 -5.61321497e-01
4.71548773e-02 -3.17848176e-01 -1.40461013e-01 8.01791608e-01
-1.01622534e+00 9.87047970e-01 5.41715336e+00 3.94602716e-01
-7.14525342e-01 -1.15632549e-01 4.93655056e-01 -1.99800625e-01
-4.07814592e-01 -3.21729332e-01 -7.09758997e-01 1.81859866e-01
1.16788161e+00 -6.50434613e-01 2.27824122e-01 4.71258134e-01
2.77688026e-01 1.69346511e-01 -1.37817299e+00 7.20675766e-01
4.85250831e-01 -1.29981816e+00 1.60692960e-01 -2.13008821e-01
5.74616432e-01 -1.05674930e-01 -4.00288373e-01 7.64653206e-01
5.19748867e-01 -9.35210228e-01 3.28953534e-01 2.23420322e-01
5.58155298e-01 -7.43362010e-01 7.21722901e-01 6.64996445e-01
-1.00908089e+00 1.38025001e-01 -9.99795347e-02 -2.08008379e-01
4.03811991e-01 1.29088461e-01 -1.35134184e+00 9.99056876e-01
1.97781712e-01 7.81888306e-01 -3.94803852e-01 3.12120616e-01
-2.67110348e-01 6.34933650e-01 -1.24523252e-01 -1.14312060e-01
-4.82899770e-02 -1.68355525e-01 6.05435610e-01 1.28863728e+00
-9.31140110e-02 5.66691101e-01 5.59558272e-01 5.31717777e-01
-3.70373398e-01 1.27285853e-01 -6.88943326e-01 -1.41372651e-01
7.10780978e-01 1.08148885e+00 -2.26135388e-01 -5.06491303e-01
-5.26558757e-01 9.00192976e-01 5.23017704e-01 4.14037585e-01
-7.53269076e-01 -1.92171961e-01 2.19495669e-01 -1.43265828e-01
-5.93432523e-02 -2.07608119e-01 2.73340736e-02 -1.37180257e+00
1.09409876e-01 -1.28201222e+00 3.95548046e-01 -3.50835562e-01
-1.08654392e+00 9.15699124e-01 1.49811506e-01 -9.69313323e-01
-6.40388727e-01 6.42592385e-02 -8.69605839e-01 5.95003068e-01
-1.18468738e+00 -1.19583941e+00 -3.30190152e-01 3.06493700e-01
1.21450365e+00 -8.02123994e-02 9.00211394e-01 1.42244667e-01
-1.03207493e+00 5.32448769e-01 -3.03153247e-01 2.87472457e-01
6.93891644e-01 -1.30830777e+00 5.01760900e-01 6.52573764e-01
1.57495234e-02 4.59070683e-01 8.30778003e-01 -6.68947816e-01
-1.40325820e+00 -1.12452352e+00 1.02915192e+00 -3.84087414e-01
5.76790154e-01 -5.23352921e-01 -1.40604842e+00 1.08641064e+00
7.98500597e-01 -7.89624393e-01 1.03048193e+00 4.25062835e-01
1.27975062e-01 2.13528469e-01 -5.16764045e-01 8.79209995e-01
9.48766410e-01 -4.12684828e-01 -1.14724374e+00 4.25688565e-01
1.27737200e+00 -8.41182172e-01 -7.86756098e-01 4.98961359e-01
-3.85777019e-02 -5.52135766e-01 5.53916216e-01 -5.68539441e-01
6.90582633e-01 -1.79789285e-03 7.35471258e-03 -1.55175495e+00
6.95126429e-02 -7.78975904e-01 -5.22016048e-01 1.71403110e+00
4.43349868e-01 -2.85149038e-01 5.46527743e-01 7.59653986e-01
-5.46095550e-01 -7.16235280e-01 -6.65623307e-01 -4.22041982e-01
-1.22206800e-01 -1.45907234e-02 5.77806175e-01 1.09062183e+00
7.01649368e-01 1.29143310e+00 -4.25757796e-01 9.73021090e-02
2.30313316e-01 1.55325219e-01 1.10759747e+00 -9.61033762e-01
-3.11680496e-01 -3.65962297e-01 1.79378748e-01 -1.34933567e+00
4.71235782e-01 -9.61371601e-01 2.67734647e-01 -2.08006883e+00
1.68152675e-01 -8.45617503e-02 2.04760179e-01 2.85550058e-01
-4.64053333e-01 -5.17770350e-01 2.56388575e-01 1.43809855e-01
-8.97749901e-01 1.07452559e+00 1.37650239e+00 -3.56959879e-01
-6.84708118e-01 2.28861347e-02 -7.76106119e-01 7.03657389e-01
1.04108763e+00 -1.35836601e-01 -8.27043116e-01 -6.65871382e-01
-2.66571790e-01 5.77239037e-01 -2.46600524e-01 -5.94087481e-01
4.84234631e-01 -7.40232542e-02 -1.65660843e-01 -9.14166093e-01
5.27411580e-01 -2.44222566e-01 -1.30825296e-01 6.56503066e-02
-8.77071083e-01 -9.35026556e-02 1.64357170e-01 4.55594420e-01
-5.94219267e-01 -1.36260182e-01 1.77221626e-01 -3.69150303e-02
-3.75585198e-01 8.08912665e-02 -2.97627866e-01 1.67756453e-01
8.70604157e-01 1.34638816e-01 -5.11686862e-01 -8.62665236e-01
-6.79403126e-01 8.06425214e-01 1.23767719e-01 5.90450168e-01
5.24457991e-01 -1.16595328e+00 -1.16860795e+00 -1.20618641e-01
6.51583895e-02 6.93517506e-01 4.17598188e-01 4.55307662e-01
-3.00588369e-01 4.10689443e-01 1.90569744e-01 -4.19228911e-01
-1.60222363e+00 2.81772584e-01 1.00308843e-01 -3.90750498e-01
-8.58884096e-01 8.85373056e-01 3.83883119e-01 -6.50405824e-01
4.66643631e-01 -3.66134703e-01 -6.87521875e-01 2.00791895e-01
6.22191012e-01 1.11418612e-01 -8.04973170e-02 -6.39395416e-01
-1.09859511e-01 -1.04261868e-01 -4.68896508e-01 3.87534201e-02
1.48525727e+00 -5.11938274e-01 -2.14633465e-01 9.07116234e-01
9.51400399e-01 -3.46103273e-02 -1.05467010e+00 -5.54257989e-01
1.32531747e-01 -2.28212727e-03 -3.22485179e-01 -4.43612993e-01
-5.88268995e-01 6.23515189e-01 -3.85536283e-01 7.85606861e-01
9.68977213e-01 2.26487234e-01 9.06051815e-01 5.65594554e-01
-2.63261765e-01 -1.01410246e+00 2.90916920e-01 7.94196606e-01
1.38396764e+00 -1.22521830e+00 -1.19901737e-02 -7.08846688e-01
-1.20156574e+00 9.72501934e-01 8.76784801e-01 6.17100112e-02
1.70230985e-01 1.88402861e-01 2.38374799e-01 -2.94715613e-01
-1.19568634e+00 -6.29670843e-02 2.81130970e-01 4.08409357e-01
8.96016598e-01 1.95639029e-01 -2.91249722e-01 8.25854897e-01
-5.08979440e-01 -3.38448197e-01 8.99356723e-01 7.61246502e-01
-3.35475653e-01 -1.13644159e+00 -1.50854308e-02 4.02498156e-01
-2.37380713e-01 -4.94820476e-02 -7.50169575e-01 6.29573941e-01
-5.32141030e-01 1.36798048e+00 2.25393623e-01 -4.29134220e-01
7.19464839e-01 1.56266242e-01 1.34385973e-01 -1.07784331e+00
-5.46810031e-01 1.50330380e-01 8.02795351e-01 -1.84085429e-01
-3.98000360e-01 -5.75056732e-01 -1.56316161e+00 -3.09919149e-01
-3.90386790e-01 4.11417812e-01 2.48362884e-01 8.94004703e-01
3.28259587e-01 1.21820915e+00 6.79530621e-01 -6.47045434e-01
-5.51274776e-01 -1.42351604e+00 -1.85065821e-01 3.79218072e-01
3.91225249e-01 -1.91521481e-01 -4.98225680e-03 2.08529040e-01] | [12.512981414794922, 8.178709983825684] |
9449d857-cad6-4650-bf51-bac08f8cd6cd | corefi-a-crowd-sourcing-suite-for-coreference | 2010.02588 | null | https://arxiv.org/abs/2010.02588v1 | https://arxiv.org/pdf/2010.02588v1.pdf | CoRefi: A Crowd Sourcing Suite for Coreference Annotation | Coreference annotation is an important, yet expensive and time consuming, task, which often involved expert annotators trained on complex decision guidelines. To enable cheaper and more efficient annotation, we present CoRefi, a web-based coreference annotation suite, oriented for crowdsourcing. Beyond the core coreference annotation tool, CoRefi provides guided onboarding for the task as well as a novel algorithm for a reviewing phase. CoRefi is open source and directly embeds into any website, including popular crowdsourcing platforms. CoRefi Demo: aka.ms/corefi Video Tour: aka.ms/corefivideo Github Repo: https://github.com/aribornstein/corefi | ['Ido Dagan', 'Arie Cattan', 'Aaron Bornstein'] | 2020-10-06 | null | https://aclanthology.org/2020.emnlp-demos.27 | https://aclanthology.org/2020.emnlp-demos.27.pdf | emnlp-2020-11 | ['cross-document-coreference-resolution'] | ['natural-language-processing'] | [-2.46753469e-01 6.45897150e-01 -2.35246539e-01 -3.10901310e-02
-1.30011153e+00 -1.32485926e+00 4.36977684e-01 1.43453524e-01
-3.36000204e-01 8.28285813e-01 5.98901331e-01 -6.46500811e-02
1.94716044e-02 -1.59351349e-01 -3.96933287e-01 -2.67500013e-01
3.31622928e-01 1.38130713e+00 5.51242232e-01 -4.82242495e-01
8.39402676e-02 2.81668544e-01 -1.28321815e+00 4.07012641e-01
6.22531295e-01 4.85455573e-01 2.86581099e-01 9.69308972e-01
4.98980694e-02 6.62969589e-01 -5.50251901e-01 -8.40836883e-01
1.52799681e-01 -2.58114755e-01 -1.48620951e+00 -6.17210984e-01
4.78216678e-01 6.33901954e-02 -3.61045688e-01 1.07955837e+00
8.09013069e-01 4.07198042e-01 -1.10974766e-01 -1.38363302e+00
-4.27016020e-01 8.29681695e-01 5.45005500e-02 2.65335530e-01
1.24198651e+00 -3.40544917e-02 1.29146600e+00 -9.34953094e-01
1.11238015e+00 1.18267059e+00 7.69371927e-01 6.86391950e-01
-8.46829474e-01 -6.75555706e-01 -2.65046716e-01 4.15793151e-01
-1.44815540e+00 -6.02653027e-01 4.57804441e-01 -4.80364442e-01
9.14812207e-01 5.55307388e-01 4.17285174e-01 1.12647402e+00
-4.96108562e-01 7.70491123e-01 8.23859274e-01 -4.45515484e-01
3.12499324e-04 2.00042762e-02 2.29418322e-01 6.84684157e-01
1.42147660e-01 -2.31117770e-01 -7.96076000e-01 -5.45648754e-01
4.23251897e-01 -4.72724617e-01 -6.82114780e-01 -1.53096974e-01
-9.66736972e-01 6.09748363e-01 -3.52497138e-02 4.42694813e-01
-1.22726932e-01 2.99060475e-02 7.32072055e-01 2.44314894e-01
4.78644669e-02 5.09278834e-01 -2.96199054e-01 -5.68827391e-01
-7.19437838e-01 4.31120127e-01 1.28037882e+00 1.50345325e+00
7.15167582e-01 -6.91664517e-01 -3.23626518e-01 7.13778138e-01
2.90870577e-01 2.93620974e-01 4.09168810e-01 -1.70730269e+00
3.69941801e-01 8.08165789e-01 7.51245022e-01 -9.16722536e-01
-5.87746263e-01 2.95666605e-01 -1.69452623e-01 -8.61474723e-02
4.82260853e-01 -1.94148481e-01 -1.87489450e-01 1.15573466e+00
6.11837387e-01 -3.46111536e-01 -1.58234417e-01 1.27935505e+00
1.45539212e+00 -1.59804016e-01 3.45017731e-01 -2.26340577e-01
1.80521929e+00 -1.17942286e+00 -9.84243214e-01 -1.14054330e-01
9.05112028e-01 -1.24286354e+00 8.23938727e-01 3.42615992e-01
-1.38219488e+00 -2.22259566e-01 -6.95447445e-01 -5.55709660e-01
-4.88048851e-01 2.89782077e-01 5.28704047e-01 3.30244660e-01
-1.16780162e+00 5.36359489e-01 -5.46539903e-01 -9.03473079e-01
3.71041149e-02 3.92623872e-01 -9.04734373e-01 3.35430796e-03
-1.45857346e+00 1.00413716e+00 5.34467518e-01 -1.02357842e-01
-7.02598155e-01 -3.92920077e-01 -8.74631703e-01 -3.80343199e-01
7.44838119e-01 -3.33932638e-01 1.85279810e+00 -8.42496574e-01
-1.41415679e+00 1.49263620e+00 -4.77416694e-01 -2.71684438e-01
5.62053204e-01 -5.49972117e-01 -5.64524829e-01 5.61201990e-01
6.72582805e-01 5.73582172e-01 8.98693278e-02 -9.85299647e-01
-7.82531977e-01 1.97401997e-02 3.63437057e-01 2.11565614e-01
1.53362796e-01 5.32237113e-01 -7.74905205e-01 -3.41503888e-01
-3.16327095e-01 -1.14880323e+00 1.83674246e-01 -1.34124175e-01
-2.47499093e-01 -6.75051332e-01 5.52242279e-01 -7.68026114e-01
1.42215276e+00 -1.94608629e+00 -5.13069145e-02 -2.54242182e-01
3.82009596e-01 7.58873463e-01 1.15710460e-01 9.81800079e-01
-8.13261718e-02 1.13620244e-01 1.26759494e-02 -2.10837469e-01
3.33494842e-01 1.25600532e-01 8.84502232e-02 3.96538019e-01
-2.55344331e-01 7.61970460e-01 -1.56161690e+00 -7.58468032e-01
1.00271605e-01 1.49000019e-01 -2.13805601e-01 1.81776345e-01
-1.72470510e-01 5.74743390e-01 -3.35630774e-01 6.56499088e-01
4.31518942e-01 -4.07714486e-01 5.77480137e-01 -1.85290366e-01
-2.21428514e-01 1.68199569e-01 -1.32034886e+00 2.01022220e+00
3.59411468e-03 7.27128983e-01 5.40760279e-01 -3.87766540e-01
8.40648711e-01 1.00771058e+00 3.63665193e-01 -1.11526452e-01
3.29845637e-01 4.90299284e-01 -3.90377402e-01 -7.72670329e-01
7.90657461e-01 4.94273245e-01 -2.03924030e-01 6.47758305e-01
3.30360621e-01 7.25183636e-02 5.76276779e-01 5.32573521e-01
1.20811534e+00 4.50119168e-01 5.38335502e-01 -2.50175387e-01
8.37138712e-01 7.93113410e-01 8.39093387e-01 6.41382873e-01
-7.16338217e-01 6.89367712e-01 3.31238240e-01 -5.11165500e-01
-8.67114723e-01 -3.79102200e-01 5.32003455e-02 1.40251160e+00
-8.59091160e-05 -9.30840135e-01 -9.32681024e-01 -6.43537521e-01
-3.52774352e-01 4.28161114e-01 -5.06161511e-01 3.88274133e-01
-4.72407907e-01 2.11566776e-01 9.01553929e-01 4.28010136e-01
4.83931601e-01 -9.81707096e-01 -4.80107874e-01 1.94579884e-01
-8.16047072e-01 -1.29769444e+00 -9.04550254e-01 -7.50359073e-02
-3.39850858e-02 -1.80421221e+00 -5.95900357e-01 -9.02705133e-01
3.92836869e-01 2.44540259e-01 1.40480483e+00 2.50698060e-01
-2.51264364e-01 8.22954953e-01 -8.41471910e-01 -4.78404403e-01
-2.50263780e-01 2.40283370e-01 -4.78366874e-02 -5.38983405e-01
1.12485099e+00 -3.51470977e-01 -6.91657305e-01 6.75627232e-01
-4.65108961e-01 -2.58918852e-01 -2.49942049e-01 4.58712190e-01
4.79130387e-01 -9.41312015e-01 2.46514842e-01 -8.39176357e-01
8.24969113e-01 -3.50496590e-01 -6.53098464e-01 3.07930946e-01
-3.01742494e-01 -6.18185937e-01 4.32722941e-02 6.58508539e-02
-9.74531889e-01 2.38749504e-01 -2.14612082e-01 -3.84711534e-01
-3.87733757e-01 1.46495968e-01 -5.57103753e-03 -1.55936152e-01
9.32523310e-01 -4.03963089e-01 -1.15958355e-01 -4.32233602e-01
3.71427149e-01 8.65682900e-01 8.88194561e-01 -5.39330840e-01
5.40326893e-01 5.22805333e-01 -5.09608269e-01 -3.36378455e-01
-8.99476707e-01 -9.98495817e-01 -9.71858203e-01 -4.06917185e-01
9.67249393e-01 -1.04343569e+00 -1.15203607e+00 -2.43194148e-01
-1.42876530e+00 -6.48486674e-01 -2.91934758e-01 2.88101315e-01
-5.15888572e-01 6.44184291e-01 -4.54888910e-01 -5.68369806e-01
-7.00768948e-01 -8.92716229e-01 8.82909298e-01 5.26531100e-01
-1.09366131e+00 -1.12027061e+00 6.27837837e-01 8.43224168e-01
-1.40680894e-02 1.63657889e-01 -8.86314809e-02 -1.27769566e+00
-1.01061292e-01 -6.51017547e-01 -4.79194745e-02 -2.24336997e-01
-3.13745767e-01 8.90814066e-02 -1.00293446e+00 -5.33619057e-03
-6.64003372e-01 -3.93321902e-01 4.29202951e-02 -1.80478126e-01
3.40983331e-01 -2.59178191e-01 -7.82128155e-01 1.72307611e-01
9.77853477e-01 8.39283839e-02 4.13594007e-01 9.25463736e-01
3.17392856e-01 6.54782057e-01 1.31844616e+00 5.48570752e-01
6.92206681e-01 9.24457371e-01 1.20702803e-01 3.91779095e-01
1.01468218e-02 -1.53917342e-01 1.58683345e-01 6.95408463e-01
-4.94054079e-01 -6.92183450e-02 -1.39187312e+00 7.53953338e-01
-2.37109232e+00 -1.01747680e+00 -6.27107441e-01 1.90386665e+00
7.46958137e-01 -4.17520434e-01 3.03786367e-01 -2.23939702e-01
1.01740420e+00 -9.26795825e-02 1.04670547e-01 -4.20816272e-01
-1.77964255e-01 5.02302349e-02 2.95315981e-01 8.53741467e-01
-1.10442209e+00 1.37280369e+00 6.36193466e+00 7.16784954e-01
-4.69007939e-01 8.05914342e-01 -2.80262619e-01 -1.53186068e-01
-1.30088210e-01 2.19094127e-01 -8.28159928e-01 3.20819020e-01
8.13836336e-01 -5.27681351e-01 3.36573005e-01 9.31113601e-01
3.26071173e-01 -2.44688615e-01 -9.38919842e-01 9.03855562e-01
6.60535470e-02 -1.55488884e+00 -5.34503043e-01 -2.88142562e-01
3.73366445e-01 3.61457109e-01 -9.09032404e-01 2.25917667e-01
9.03314710e-01 -5.84503293e-01 6.46184742e-01 2.69297987e-01
8.56698096e-01 -3.20941687e-01 1.13588083e+00 -5.52351363e-02
-1.25636733e+00 1.58840984e-01 -2.15932876e-01 2.69365907e-02
4.86480325e-01 1.75101578e-01 -5.54361284e-01 6.53081954e-01
1.00087488e+00 2.95086354e-01 -4.05326277e-01 1.19424963e+00
-8.53226423e-01 8.09131190e-02 -3.58711891e-02 4.95943725e-02
-1.27531981e-04 3.01631372e-02 9.00530756e-01 1.56394506e+00
-8.94079581e-02 7.62876391e-01 3.56412351e-01 3.05161268e-01
-1.78278744e-01 9.66225341e-02 -5.61486602e-01 -1.77403342e-03
1.13081396e+00 1.79892576e+00 -5.80970228e-01 -3.25083286e-01
-2.62333781e-01 1.14656341e+00 4.01929945e-01 1.19174644e-01
-6.58615291e-01 -5.69522917e-01 5.33232868e-01 6.87780231e-02
2.43268514e-04 -2.47607566e-03 3.40605348e-01 -8.29955757e-01
-1.44652471e-01 -1.21388328e+00 9.61056173e-01 -1.09166515e+00
-6.51121438e-01 8.42524052e-01 -7.51348287e-02 -1.33675349e+00
-2.34306633e-01 -5.92364788e-01 -7.44145334e-01 7.39509940e-01
-9.66634274e-01 -1.20748150e+00 -7.54634559e-01 9.22721684e-01
4.24255371e-01 -1.00460216e-01 1.16645491e+00 3.50802481e-01
-6.41007662e-01 5.50965548e-01 -3.25140983e-01 5.24445176e-01
1.20182717e+00 -1.21999395e+00 1.26219615e-01 4.66957331e-01
-2.93665826e-01 6.61415696e-01 9.17687058e-01 -7.69736290e-01
-1.20554078e+00 -7.61966586e-01 1.22378957e+00 -1.04086983e+00
1.03557932e+00 -8.18761140e-02 -7.44960070e-01 1.00799298e+00
6.31795645e-01 -6.66352063e-02 1.15590847e+00 2.61140615e-02
-2.74647623e-01 4.23710227e-01 -1.07181644e+00 4.31639522e-01
1.18124807e+00 -8.65650594e-01 -9.83398616e-01 9.48057055e-01
5.21090984e-01 -8.18557262e-01 -1.36985934e+00 -1.01368129e-01
4.19905603e-01 -7.72350490e-01 3.72976959e-01 -3.08089435e-01
1.94725916e-01 -4.67598617e-01 3.93310077e-02 -8.35700154e-01
-4.27858442e-01 -1.36480355e+00 -6.48483261e-02 1.61601603e+00
4.35294002e-01 -5.40877879e-01 6.03779078e-01 1.12957323e+00
-3.73119265e-01 -1.10784478e-01 -1.02042270e+00 -6.49904549e-01
-3.39289486e-01 -4.39439714e-01 3.63385141e-01 1.07245219e+00
8.75912488e-01 2.84288973e-01 -1.28597289e-01 2.99068689e-01
4.28715557e-01 -2.13640615e-01 8.25470865e-01 -1.33334172e+00
-5.98273147e-03 -2.53878206e-01 -9.24869850e-02 -4.03749168e-01
1.79836884e-01 -9.14558232e-01 2.65915766e-02 -1.66995454e+00
-1.24443017e-01 -1.80593491e-01 3.66302490e-01 7.77132332e-01
-2.95904055e-02 4.95411962e-01 2.78340131e-01 7.68136203e-01
-1.30809855e+00 -1.50634885e-01 8.50469291e-01 1.70863196e-01
-7.49612227e-02 -3.24494123e-01 -7.78633773e-01 8.92601192e-01
7.48212218e-01 -6.64477825e-01 1.47337735e-01 -5.09682834e-01
5.43293595e-01 -1.63809285e-01 2.05978453e-01 -9.65678275e-01
9.30208206e-01 1.88884616e-01 -1.51312381e-01 -3.83391172e-01
1.76953986e-01 -6.41474128e-01 3.60208154e-01 5.90650029e-02
-1.46422178e-01 9.27899778e-02 3.73253137e-01 2.96259493e-01
-4.13029581e-01 -7.87789047e-01 2.04914391e-01 -6.62620664e-01
-8.90436172e-01 1.18428268e-01 -3.67857516e-01 4.61608976e-01
1.07333028e+00 -2.50252187e-01 -7.07877815e-01 -4.43937302e-01
-1.25343096e+00 8.39284718e-01 6.16846323e-01 3.91090214e-01
-5.16155213e-02 -1.21522725e+00 -5.38859427e-01 -5.81877351e-01
4.06653404e-01 -7.80439004e-02 3.34538072e-01 1.24381137e+00
-7.21103072e-01 7.14334548e-01 -5.84312752e-02 -2.11457551e-01
-1.57509327e+00 4.99100178e-01 2.73810178e-01 -4.31491770e-02
-5.04117966e-01 7.64247179e-01 -4.10214663e-01 -6.44571781e-01
3.17684650e-01 4.54413861e-01 -4.74911094e-01 4.95550722e-01
9.28790867e-01 7.28992462e-01 9.12471190e-02 -9.18296993e-01
-6.60809457e-01 3.94154161e-01 1.09700933e-01 -2.64637262e-01
9.04971659e-01 -5.09872615e-01 -2.83283144e-01 2.38974467e-01
5.49822628e-01 1.50950655e-01 -6.10214770e-01 -1.26303956e-01
4.58977908e-01 -3.54534358e-01 -3.82193714e-01 -8.91931713e-01
-4.38883543e-01 3.49850506e-01 4.30619597e-01 2.98773050e-01
6.14864349e-01 3.50023657e-01 5.85421205e-01 4.65317667e-01
6.59754932e-01 -1.50049937e+00 -3.98021638e-01 6.24730289e-01
1.28876913e+00 -1.44755530e+00 -1.37097584e-02 -4.19423223e-01
-8.23511600e-01 1.07476926e+00 6.32636487e-01 2.84807123e-02
4.26203579e-01 3.42754312e-02 4.89784449e-01 -5.78725338e-01
-5.91337502e-01 -4.88696903e-01 3.93552147e-03 8.72039378e-01
7.35944271e-01 3.23649449e-03 -7.41006970e-01 9.91953194e-01
-1.91704080e-01 3.13846081e-01 6.40522301e-01 9.19004738e-01
-2.58415103e-01 -1.25527883e+00 -6.30865872e-01 -5.22091016e-02
-6.29477382e-01 -2.85329316e-02 -8.34139466e-01 8.85623455e-01
6.05505593e-02 1.23715246e+00 -2.36575063e-02 -2.12071240e-01
6.37513518e-01 8.67406428e-02 -4.78776805e-02 -7.86980748e-01
-1.17658043e+00 1.03152301e-02 8.85613799e-01 -8.70375574e-01
-7.12585628e-01 -5.76100290e-01 -1.29070568e+00 -4.92237300e-01
-2.71439105e-01 9.09417093e-01 1.95484683e-01 8.46981645e-01
4.84469354e-01 -7.01651350e-02 -1.71913803e-01 -1.16293705e+00
2.57748455e-01 -9.98197436e-01 -2.66520947e-01 4.99910742e-01
-1.88167691e-01 -6.17478192e-01 -2.84046143e-01 -3.32995169e-02] | [9.36915111541748, 9.2487154006958] |
0e9f56de-fca0-4136-bca1-bd70f433733f | loop-closure-detection-based-on-object-level | 2304.05146 | null | https://arxiv.org/abs/2304.05146v2 | https://arxiv.org/pdf/2304.05146v2.pdf | Loop Closure Detection Based on Object-level Spatial Layout and Semantic Consistency | Visual simultaneous localization and mapping (SLAM) systems face challenges in detecting loop closure under the circumstance of large viewpoint changes. In this paper, we present an object-based loop closure detection method based on the spatial layout and semanic consistency of the 3D scene graph. Firstly, we propose an object-level data association approach based on the semantic information from semantic labels, intersection over union (IoU), object color, and object embedding. Subsequently, multi-view bundle adjustment with the associated objects is utilized to jointly optimize the poses of objects and cameras. We represent the refined objects as a 3D spatial graph with semantics and topology. Then, we propose a graph matching approach to select correspondence objects based on the structure layout and semantic property similarity of vertices' neighbors. Finally, we jointly optimize camera trajectories and object poses in an object-level pose graph optimization, which results in a globally consistent map. Experimental results demonstrate that our proposed data association approach can construct more accurate 3D semantic maps, and our loop closure method is more robust than point-based and object-based methods in circumstances with large viewpoint changes. | ['Fei Wen', 'Rendong Ying', 'Xiang Chen', 'Haochen Niu', 'Peilin Liu', 'Xingwu Ji'] | 2023-04-11 | null | null | null | null | ['simultaneous-localization-and-mapping', 'loop-closure-detection', 'graph-matching'] | ['computer-vision', 'computer-vision', 'graphs'] | [-1.95768133e-01 -3.89575899e-01 -7.28648379e-02 -3.95454556e-01
-3.05286139e-01 -7.20946074e-01 4.35153931e-01 4.72224206e-01
-9.43446532e-02 6.14223890e-02 -5.99372983e-02 8.27593058e-02
-3.13234985e-01 -6.35276973e-01 -7.76380837e-01 -1.76314160e-01
1.04758605e-01 6.74032748e-01 6.73695922e-01 3.29907089e-02
5.24939775e-01 8.17763686e-01 -1.35961950e+00 -3.61461073e-01
6.68963015e-01 8.79469752e-01 6.07791543e-01 3.24996024e-01
-1.02078952e-01 3.43312681e-01 -2.36051232e-02 9.82401893e-02
3.85548621e-01 -5.70304766e-02 -5.34070373e-01 6.69682622e-01
6.48319185e-01 -1.76516578e-01 -3.51501673e-01 1.23477638e+00
1.91385508e-01 2.20156208e-01 1.79697394e-01 -1.60272229e+00
-3.57675254e-01 -2.28157327e-01 -8.75123560e-01 -2.32866436e-01
7.59145141e-01 -1.30422473e-01 9.40464377e-01 -1.21304429e+00
8.49194407e-01 1.37486124e+00 5.19064248e-01 -6.72732666e-02
-1.12148035e+00 -3.83167624e-01 6.05744064e-01 2.74441838e-01
-1.88675654e+00 -3.43298316e-01 1.11322355e+00 -4.75509971e-01
7.78849900e-01 3.00671190e-01 7.97601223e-01 3.08358938e-01
3.74713391e-02 3.65636110e-01 8.71323287e-01 -3.98564219e-01
1.82915747e-01 2.81758428e-01 -2.72491556e-02 1.10847688e+00
4.82107997e-01 -1.59624457e-01 -6.46440804e-01 -3.06475490e-01
8.83209884e-01 4.25638109e-01 -1.88539892e-01 -1.40576148e+00
-1.60053313e+00 5.68198800e-01 6.57792509e-01 4.80745509e-02
-1.46232471e-01 1.17071487e-01 -9.53506380e-02 -1.08224615e-01
2.78721362e-01 2.14039177e-01 -8.32012892e-02 3.70275408e-01
-5.09902298e-01 1.01712339e-01 4.64916974e-01 1.72964466e+00
1.31124794e+00 -3.94156873e-01 3.11109871e-01 4.92978454e-01
7.50186563e-01 7.13045061e-01 -3.04973125e-01 -1.05680311e+00
5.16363204e-01 1.07657254e+00 2.89394557e-01 -1.78084433e+00
-5.24375558e-01 -1.18820846e-01 -5.09839714e-01 6.00508712e-02
-1.15729779e-01 4.91514117e-01 -6.89803421e-01 1.34069288e+00
7.42898703e-01 2.46206686e-01 -1.65724412e-01 1.07626569e+00
5.70402563e-01 3.44985604e-01 -3.61576080e-01 -1.18321240e-01
1.23797238e+00 -7.57549047e-01 -5.78259885e-01 -3.94122094e-01
4.92565900e-01 -8.04480493e-01 6.44791186e-01 -1.58447772e-01
-8.20765018e-01 -4.72085506e-01 -1.15486324e+00 -1.35961771e-01
-1.43576249e-01 1.09405071e-02 2.81523675e-01 3.74284610e-02
-8.87553096e-01 2.76635867e-02 -6.63617849e-01 -7.05050707e-01
1.55138103e-02 2.14320764e-01 -6.34433687e-01 -1.55873761e-01
-4.88058060e-01 9.07212436e-01 6.43827856e-01 2.40974545e-01
-5.79775929e-01 -3.41325402e-01 -1.19953227e+00 -2.84929603e-01
5.88405788e-01 -7.57585645e-01 6.91498637e-01 -3.25999290e-01
-8.79835308e-01 9.67601657e-01 -4.43413377e-01 1.24472797e-01
3.77515882e-01 3.01447250e-02 -2.49417499e-01 1.47059798e-01
3.64366472e-01 5.38567185e-01 4.26276356e-01 -1.77835572e+00
-9.69259322e-01 -7.71254957e-01 -2.29630172e-02 7.34440744e-01
2.22539172e-01 -2.06957981e-01 -9.78327572e-01 1.06102601e-01
1.33895588e+00 -9.59847808e-01 -2.75496453e-01 3.25370371e-01
-3.78823847e-01 -7.71575496e-02 1.14391696e+00 -4.80110705e-01
9.33956027e-01 -2.22230506e+00 3.12769830e-01 7.10376024e-01
4.31387365e-01 -4.99186099e-01 6.81727100e-03 3.77415001e-01
4.25623864e-01 -2.00079486e-01 8.24030302e-03 -4.35583860e-01
-1.56407416e-01 2.66328007e-01 6.29333928e-02 8.86435032e-01
-2.50018865e-01 6.19846821e-01 -1.15153873e+00 -5.90061009e-01
6.13440037e-01 1.49983168e-01 -4.55631196e-01 2.01417834e-01
-1.89162135e-01 4.04695719e-01 -6.88358784e-01 7.96858132e-01
8.76995802e-01 -2.29447842e-01 2.87907701e-02 -6.11490488e-01
-4.28913027e-01 -2.54991818e-02 -1.57464218e+00 2.22020125e+00
-2.36209005e-01 4.81844753e-01 4.09105755e-02 -4.70385969e-01
1.26142597e+00 -1.55914441e-01 5.40632248e-01 -4.69863564e-01
-4.57019499e-03 2.16957778e-01 -5.63344896e-01 -2.20211431e-01
6.13573074e-01 4.72717345e-01 -1.24328971e-01 1.18244365e-01
-2.30606660e-01 -3.82578343e-01 -1.70399725e-01 3.89397681e-01
7.34318554e-01 2.83451885e-01 3.91534716e-01 -1.69469818e-01
5.55303097e-01 1.72308877e-01 6.36810482e-01 4.63716388e-01
-7.92638287e-02 5.26031911e-01 4.27347384e-02 -4.43185180e-01
-1.03188062e+00 -1.20612812e+00 1.40248656e-01 4.09836024e-01
1.34579945e+00 -5.05787551e-01 -2.67994672e-01 -5.33760488e-01
3.23762238e-01 3.56613904e-01 -1.92493334e-01 -1.79564595e-01
-4.88663256e-01 -1.39957219e-01 -3.08966428e-01 2.11661220e-01
5.56506455e-01 -3.18074942e-01 -5.84060013e-01 2.46329047e-02
-2.22514942e-01 -1.28318262e+00 -8.08254838e-01 -3.42151970e-01
-8.22546065e-01 -1.15072846e+00 6.76001459e-02 -8.74630928e-01
1.17262638e+00 1.01263106e+00 5.36237717e-01 1.55240804e-01
-1.16440378e-01 7.55051672e-01 -2.57761091e-01 -3.87684144e-02
-6.16939850e-02 -3.69764686e-01 4.24870908e-01 2.34485865e-01
2.63793916e-01 -5.13040721e-01 -5.62681437e-01 6.93938911e-01
-3.29475880e-01 4.48020488e-01 3.67240041e-01 2.11181313e-01
1.02268934e+00 -5.23319542e-02 -2.78501719e-01 -4.09172326e-01
5.63810579e-02 -1.83953047e-01 -1.02824414e+00 5.24978757e-01
-6.89091921e-01 -1.68635741e-01 -8.37493539e-02 -6.27593100e-02
-7.96483755e-01 5.99299729e-01 6.93274319e-01 -8.47082555e-01
-5.95505759e-02 3.07225883e-01 -5.35750091e-01 -4.51336265e-01
3.93849403e-01 2.19676808e-01 5.01081757e-02 -5.77086449e-01
4.93815303e-01 4.13442850e-01 4.80314344e-01 -3.01650167e-01
9.91033554e-01 8.84189725e-01 2.80266732e-01 -5.85348368e-01
-6.59940898e-01 -9.90991116e-01 -1.05995357e+00 -4.71336961e-01
8.54366064e-01 -1.14569151e+00 -6.58078372e-01 2.05958724e-01
-1.35315919e+00 3.24672908e-01 1.86509117e-01 8.09328258e-01
-6.90037549e-01 5.80716252e-01 -1.00208133e-01 -6.65573835e-01
1.25222757e-01 -1.12896621e+00 1.32482851e+00 2.95028533e-03
8.14866945e-02 -9.42648649e-01 -4.05416042e-02 1.74645513e-01
-3.52753401e-01 4.92353827e-01 5.66581786e-01 -1.97380245e-01
-1.32980990e+00 -1.74857453e-01 -4.55902606e-01 -3.17648262e-01
2.80510306e-01 -4.63083312e-02 -4.59481835e-01 -4.56471145e-01
-3.49124074e-02 5.63734233e-01 2.75876790e-01 1.16952136e-01
6.64765477e-01 -4.01163213e-02 -8.55198562e-01 9.29135323e-01
1.71629214e+00 1.29855826e-01 7.16201216e-02 4.26603884e-01
1.27548158e+00 7.15527236e-01 1.01124859e+00 4.85801637e-01
7.62917995e-01 8.60074878e-01 7.98673093e-01 -1.00597534e-02
3.36288065e-02 -7.51961887e-01 2.53505930e-02 9.22866642e-01
2.69600391e-01 1.01720378e-01 -1.05391204e+00 4.21724468e-01
-2.19162703e+00 -4.99797463e-01 -4.36395973e-01 2.36326456e+00
6.33693337e-02 -4.65147384e-02 -6.78445846e-02 -4.53904212e-01
1.12487042e+00 3.18955034e-02 -5.47411323e-01 1.82019651e-01
-8.22123736e-02 -9.69934940e-01 8.62914205e-01 7.77123511e-01
-8.72188747e-01 9.68784988e-01 5.23094416e+00 2.33212665e-01
-6.24122322e-01 1.09127350e-01 -2.33747482e-01 1.89538494e-01
-5.02686620e-01 5.76044142e-01 -9.31420684e-01 2.30457574e-01
-3.74739841e-02 -1.11215189e-01 4.51714367e-01 7.40913272e-01
2.33833030e-01 -2.64373869e-01 -1.15475869e+00 1.34195375e+00
3.75100732e-01 -1.33112681e+00 1.88443512e-01 2.81716347e-01
6.39618754e-01 1.89423449e-02 -3.56180251e-01 -3.97324294e-01
6.90061003e-02 -2.36377001e-01 1.11756432e+00 5.82339108e-01
5.09440124e-01 -6.15738034e-01 2.89298862e-01 4.95571613e-01
-1.67448092e+00 -1.32106887e-02 -4.59864259e-01 9.00162235e-02
3.04202229e-01 4.30376142e-01 -9.03523803e-01 8.42854977e-01
7.31886983e-01 1.00629890e+00 -5.37905693e-01 1.20363057e+00
-4.85279635e-02 -3.62374216e-01 -4.49237317e-01 1.24426730e-01
5.32146059e-02 -6.56436980e-01 9.47214365e-01 6.56085789e-01
3.82543236e-01 1.50574401e-01 7.13873386e-01 9.86907303e-01
3.29692751e-01 1.49117574e-01 -8.40549529e-01 5.49352467e-01
9.50486422e-01 1.22123289e+00 -1.03583086e+00 -1.28578842e-01
-5.81373036e-01 9.97955024e-01 3.81379992e-01 3.70419204e-01
-7.40830481e-01 -8.46188143e-02 5.73778689e-01 1.48554593e-01
1.23014562e-02 -7.27471113e-01 -1.17181830e-01 -1.18073094e+00
2.98350602e-01 -1.97025582e-01 1.89164117e-01 -1.11682642e+00
-7.97300637e-01 2.40174219e-01 1.50438964e-01 -1.54035211e+00
2.97158152e-01 -3.15398097e-01 -2.70988166e-01 6.88728929e-01
-1.25778985e+00 -1.33726263e+00 -7.72938967e-01 6.76750004e-01
2.83234596e-01 1.58084229e-01 3.75543773e-01 6.48995265e-02
-1.74237147e-01 -7.32542798e-02 6.49313182e-02 -6.05989397e-02
4.63648677e-01 -1.05565655e+00 2.66256660e-01 9.53217506e-01
3.86354685e-01 5.99070847e-01 4.59847659e-01 -9.39581156e-01
-1.84376132e+00 -1.20568120e+00 7.87280083e-01 -6.80138350e-01
3.93392742e-01 -7.34486580e-01 -5.87049067e-01 9.31720853e-01
-3.88612390e-01 4.90070134e-03 7.56415650e-02 -4.55084778e-02
-3.68851990e-01 -2.75793791e-01 -9.69195724e-01 5.06543815e-01
1.51335680e+00 -6.56371117e-01 -4.46925402e-01 4.46652114e-01
9.31010008e-01 -8.40908945e-01 -7.11443484e-01 5.40433407e-01
4.46672618e-01 -7.56519794e-01 1.13907897e+00 -7.61074647e-02
-5.92054367e-01 -1.05249131e+00 -4.15292948e-01 -1.03458118e+00
-5.08568883e-01 -3.21152598e-01 1.38710856e-01 1.12403476e+00
-5.42080728e-03 -6.62124634e-01 5.58898747e-01 4.41254139e-01
-2.36997932e-01 -2.33702973e-01 -8.76343727e-01 -8.46436977e-01
-1.00976288e+00 -1.68232679e-01 6.70255065e-01 9.32107151e-01
-2.30660349e-01 2.51542181e-01 -2.60423362e-01 9.00960326e-01
9.43693578e-01 6.27141774e-01 1.12019134e+00 -1.34523511e+00
1.65297657e-01 -2.15824470e-01 -8.74735355e-01 -1.09581995e+00
4.33482528e-02 -9.32716191e-01 2.88769037e-01 -1.68811429e+00
2.96145290e-01 -6.75992012e-01 -9.55384150e-02 6.31038621e-02
-1.45907234e-02 -4.56822244e-03 2.63105422e-01 6.06252134e-01
-1.05842531e+00 4.85165149e-01 1.05063486e+00 4.69550267e-02
-3.74042720e-01 -2.85420716e-01 -1.30921930e-01 7.31805861e-01
4.00217444e-01 -2.72586942e-01 -3.90512675e-01 -7.24821270e-01
1.77682474e-01 5.29136099e-02 6.42283440e-01 -8.36057246e-01
5.40072799e-01 -3.65744323e-01 1.90634683e-01 -1.00034726e+00
5.29637694e-01 -1.31667161e+00 6.30667627e-01 3.70921910e-01
5.97454607e-02 2.46956229e-01 -6.99458122e-02 9.22861576e-01
-3.74629647e-02 8.33678097e-02 4.81202751e-01 -2.52522290e-01
-1.28378642e+00 7.18291104e-01 2.26520777e-01 -3.20222408e-01
1.28597295e+00 -5.78397274e-01 -5.78989051e-02 -1.98315397e-01
-6.11463249e-01 5.48533976e-01 1.17260599e+00 7.71878123e-01
1.01988792e+00 -1.66128409e+00 -4.74158913e-01 5.40740252e-01
7.11459696e-01 2.92470127e-01 8.27678069e-02 9.24804807e-01
-7.02582657e-01 3.00916612e-01 9.12552886e-03 -1.31188571e+00
-1.44539773e+00 7.00062454e-01 2.05841720e-01 6.08298719e-01
-6.46506906e-01 5.97857714e-01 5.02637446e-01 -6.13535106e-01
1.81643218e-01 -2.08049566e-01 1.72313899e-01 -2.60143846e-01
1.32978335e-01 5.11309624e-01 -8.53365809e-02 -1.06555676e+00
-7.79445529e-01 1.19568598e+00 2.57102489e-01 -1.69291627e-02
9.02098417e-01 -8.80505204e-01 -5.10933697e-01 5.08724511e-01
1.24353135e+00 8.14393088e-02 -1.20983052e+00 -4.59946364e-01
5.86382002e-02 -1.01816893e+00 6.30534021e-03 -9.38374102e-02
-7.18909383e-01 5.74110091e-01 6.45739257e-01 -1.17081977e-01
7.85620511e-01 3.02908063e-01 3.81041914e-01 3.24301630e-01
9.93937016e-01 -9.30808842e-01 -7.26521760e-02 3.10593635e-01
7.87356138e-01 -1.17764997e+00 2.49805287e-01 -8.08419168e-01
-2.79594898e-01 8.47698987e-01 8.65750730e-01 -5.91753982e-02
5.63400328e-01 -2.51140356e-01 -1.71749353e-01 -4.85823154e-01
-2.78283358e-01 -2.70222366e-01 3.94307196e-01 5.75344920e-01
-3.57914120e-01 3.71839069e-02 6.42067045e-02 -2.63208061e-01
1.74838603e-01 -6.24171376e-01 2.09796354e-01 9.16283369e-01
-8.72481823e-01 -7.32400298e-01 -6.14549160e-01 7.78221758e-03
5.46391904e-01 2.85702050e-01 -3.48881274e-01 7.55345762e-01
2.13326409e-01 8.12324464e-01 3.33070070e-01 -5.61548650e-01
6.08665049e-01 -3.12879920e-01 7.15232432e-01 -7.90806115e-01
2.74614692e-01 2.04363197e-01 -2.05785722e-01 -9.21128929e-01
-5.22093773e-01 -6.97329104e-01 -1.34610045e+00 -1.22780792e-01
-6.06980562e-01 2.69496925e-02 9.76050854e-01 7.69188941e-01
6.27159417e-01 -1.73952937e-01 9.02855933e-01 -7.86174893e-01
-7.93765411e-02 -2.42413715e-01 -7.54515767e-01 5.29335082e-01
3.68289173e-01 -9.41146314e-01 -3.32615972e-01 -1.32429540e-01] | [7.309749603271484, -2.303487539291382] |
fd0f8a0e-4f9a-425a-a217-5113c47013b6 | unsupervised-machine-learning-framework-for | 2208.01439 | null | https://arxiv.org/abs/2208.01439v3 | https://arxiv.org/pdf/2208.01439v3.pdf | Unsupervised machine learning framework for discriminating major variants of concern during COVID-19 | Due to the high mutation rate of the virus, the COVID-19 pandemic evolved rapidly. Certain variants of the virus, such as Delta and Omicron, emerged with altered viral properties leading to severe transmission and death rates. These variants burdened the medical systems worldwide with a major impact to travel, productivity, and the world economy. Unsupervised machine learning methods have the ability to compress, characterize, and visualize unlabelled data. This paper presents a framework that utilizes unsupervised machine learning methods to discriminate and visualize the associations between major COVID-19 variants based on their genome sequences. These methods comprise a combination of selected dimensionality reduction and clustering techniques. The framework processes the RNA sequences by performing a k-mer analysis on the data and further visualises and compares the results using selected dimensionality reduction methods that include principal component analysis (PCA), t-distributed stochastic neighbour embedding (t-SNE), and uniform manifold approximation projection (UMAP). Our framework also employs agglomerative hierarchical clustering to visualize the mutational differences among major variants of concern and country-wise mutational differences for selected variants (Delta and Omicron) using dendrograms. We also provide country-wise mutational differences for selected variants via dendrograms. We find that the proposed framework can effectively distinguish between the major variants and has the potential to identify emerging variants in the future. | ['Seshadri Vasan', 'Laurence O. W. Wilson', 'Pranjal Singh', 'Vinti Agarwal', 'Tom Blau', 'Mingyue Kang', 'Chaarvi Bansal', 'Rohitash Chandra'] | 2022-08-01 | null | null | null | null | ['machine-learning', 'machine-learning'] | ['methodology', 'miscellaneous'] | [ 2.22023860e-01 -5.37714362e-01 2.57631063e-01 4.21511047e-02
-1.12699404e-01 -8.22027028e-01 7.38956630e-01 4.47467089e-01
-3.71043503e-01 6.52669370e-01 3.38078916e-01 -4.74234670e-01
-5.19584179e-01 -5.14287353e-01 1.36241645e-01 -1.14132130e+00
-5.03342271e-01 1.00116420e+00 -3.80881906e-01 -1.83631748e-01
2.52249897e-01 8.13271940e-01 -1.40125930e+00 1.55190915e-01
1.01435995e+00 9.62939560e-02 2.51300156e-01 8.83138716e-01
-3.25942993e-01 -9.66628268e-02 -7.41409957e-01 3.66845205e-02
2.45130032e-01 -4.78076428e-01 -2.82966197e-01 -1.48697406e-01
-2.90996075e-01 2.18700588e-01 4.56336856e-01 8.23077679e-01
4.23307180e-01 -4.83368151e-02 1.22844696e+00 -1.34059465e+00
-4.40141857e-01 -1.03414185e-01 -6.90726519e-01 3.21649879e-01
3.74328226e-01 3.66731063e-02 5.11077642e-01 -9.03974891e-01
1.16113329e+00 1.35829484e+00 8.01786482e-01 2.82268971e-01
-1.55239928e+00 -2.22024068e-01 -5.16701698e-01 2.04133034e-01
-1.47462666e+00 6.85260072e-02 6.71940982e-01 -1.26903582e+00
1.15695620e+00 6.61101639e-01 8.39757919e-01 8.22215319e-01
5.25242448e-01 1.63354665e-01 1.11653531e+00 -7.13796616e-02
3.84270310e-01 -5.86904921e-02 3.00260726e-02 4.63921368e-01
5.96213341e-01 -1.97072625e-02 3.03597432e-02 -9.38572526e-01
1.78151101e-01 6.72978878e-01 -2.28016183e-01 -6.78198576e-01
-1.23112392e+00 1.00718284e+00 -1.19905688e-01 2.78342515e-01
-6.33283794e-01 -5.29329777e-01 4.16191012e-01 1.80826277e-01
3.17744792e-01 4.36417311e-01 -5.72286904e-01 -2.99829572e-01
-9.18464184e-01 -9.46673229e-02 2.63524979e-01 2.81879693e-01
8.41013968e-01 -1.52270601e-03 3.14187199e-01 6.78664327e-01
5.51466346e-01 5.56944311e-01 5.77371418e-01 -3.61430973e-01
-5.36394790e-02 9.17717576e-01 6.90705404e-02 -1.25549698e+00
-7.76803732e-01 1.86145768e-01 -9.45531189e-01 2.51760602e-01
2.49467837e-03 -1.87945008e-01 -7.94057965e-01 1.37656212e+00
4.52183068e-01 3.11036687e-02 1.81491733e-01 5.31768262e-01
5.06916165e-01 8.68891895e-01 1.11994550e-01 -5.82167566e-01
1.21051228e+00 -2.48206779e-01 -6.01955354e-01 7.65973508e-01
6.82102203e-01 -5.85123658e-01 7.03328848e-01 6.01841845e-02
-4.90458012e-01 -2.63316929e-01 -8.55416298e-01 5.84924161e-01
-6.83646560e-01 -2.88355798e-01 3.05084847e-02 9.92526233e-01
-1.30874288e+00 3.14948976e-01 -1.06688190e+00 -9.69357073e-01
9.16277170e-02 3.35816503e-01 -4.71655428e-01 2.23004699e-01
-8.61036599e-01 7.09339023e-01 4.20476228e-01 8.20653420e-03
-5.96017540e-01 -5.52657425e-01 -7.59393275e-01 -2.68029988e-01
-4.46813017e-01 -8.06905270e-01 -2.95220204e-02 -3.90609771e-01
-1.16853726e+00 7.11579919e-01 -4.01577592e-01 -2.03060284e-01
2.64986098e-01 1.38895690e-01 -6.38294935e-01 2.72745728e-01
-9.47962701e-02 3.65139961e-01 7.37699389e-01 -1.19580555e+00
-4.34265703e-01 -6.97785199e-01 -7.03284562e-01 9.68684554e-02
-1.16250679e-01 2.81925559e-01 1.44137472e-01 -6.70314431e-01
-8.44942257e-02 -1.26235068e+00 -3.26022446e-01 -5.61106682e-01
-2.70808339e-01 -2.42762014e-01 1.23929930e+00 -6.54179394e-01
9.43954885e-01 -2.16669321e+00 5.53434789e-01 4.57811713e-01
5.30687511e-01 3.50875407e-01 -1.31173879e-02 1.04716325e+00
-1.29232854e-01 2.58270055e-01 -6.25720978e-01 1.39612451e-01
-3.04535180e-01 3.35186511e-01 7.16842040e-02 8.91171515e-01
1.49825513e-01 5.50399005e-01 -9.82929051e-01 -2.06249177e-01
2.73070753e-01 9.65659857e-01 -4.63408887e-01 9.39352065e-02
7.03311563e-02 5.33444107e-01 -4.56843786e-02 4.18793797e-01
8.51172626e-01 -1.34100057e-02 6.69081390e-01 1.17296800e-01
-1.82948753e-01 -3.30676913e-01 -6.76227272e-01 1.09192705e+00
2.82980114e-01 6.55210912e-01 -7.13622943e-02 -7.89881051e-01
9.72476721e-01 1.40082806e-01 5.23798168e-01 7.63299242e-02
-8.07059631e-02 -7.75793716e-02 1.35540277e-01 -6.53583169e-01
3.72711748e-01 -1.16239585e-01 3.21568847e-01 4.63956982e-01
-3.04449767e-01 3.98773968e-01 2.41388276e-01 3.19614917e-01
9.14513767e-01 -2.66735286e-01 4.45269138e-01 -5.72926521e-01
6.27127647e-01 2.73600459e-01 4.89205688e-01 1.02093637e-01
-2.39103615e-01 2.87827730e-01 5.44132471e-01 -4.38462496e-01
-1.09838617e+00 -1.41053617e+00 -1.53471977e-01 8.84702206e-01
-8.51415470e-02 -4.61352915e-01 -5.47041178e-01 -3.48267287e-01
4.17580828e-03 4.79549497e-01 -6.70795858e-01 -4.22411319e-03
-3.14929694e-01 -1.29403722e+00 5.59479892e-01 7.07531273e-02
-1.83018163e-01 -8.24380219e-01 -8.41146767e-01 -2.20753536e-01
1.02556661e-01 -2.43768394e-01 -1.27190441e-01 7.98332617e-02
-1.05984390e+00 -1.16925788e+00 -6.12952411e-01 -6.78319812e-01
8.83534312e-01 4.28444557e-02 5.84910035e-01 -1.40201956e-01
-6.63635492e-01 4.08415914e-01 -2.96119869e-01 -2.49281600e-01
-6.50442183e-01 -3.85050416e-01 5.54552257e-01 -1.11331925e-01
6.54664099e-01 -5.04261851e-01 -6.29225969e-01 2.12749764e-01
-8.65667224e-01 -3.92695695e-01 3.86412561e-01 6.00230873e-01
5.42474031e-01 3.01252007e-01 1.89781114e-01 -8.36691439e-01
9.33246195e-01 -8.46962929e-01 -3.45978022e-01 1.25160456e-01
-7.38101840e-01 4.37627807e-02 4.51060295e-01 -1.46842808e-01
-8.72609198e-01 -1.41747743e-01 2.57215768e-01 -2.11834982e-01
-4.18400079e-01 5.08172274e-01 -4.78380471e-02 4.21153843e-01
6.42983377e-01 4.83111054e-01 4.02010828e-01 -5.79248846e-01
4.88126576e-01 8.79367590e-01 2.10835278e-01 7.05500841e-02
7.44197488e-01 8.92905354e-01 6.47940394e-03 -1.64987004e+00
6.38482451e-01 -7.88383782e-01 -8.81677926e-01 -1.89809963e-01
1.28581607e+00 -5.30132234e-01 -6.10108674e-01 3.51847529e-01
-8.14563513e-01 1.68571606e-01 1.80236101e-01 5.45927823e-01
-2.67357916e-01 7.12707698e-01 -6.14369333e-01 -7.70521343e-01
-3.19567442e-01 -1.02383685e+00 7.38365173e-01 -2.30288059e-01
-5.80583990e-01 -1.21179473e+00 9.90757346e-01 1.43124178e-01
4.55868423e-01 7.02367365e-01 1.53641903e+00 -8.23180199e-01
1.30314767e-01 1.37609884e-01 -1.30143855e-02 -7.96287805e-02
5.52473545e-01 4.79094684e-01 -3.76277089e-01 -6.16146445e-01
-1.13582704e-02 4.01627600e-01 6.73687816e-01 4.41636473e-01
1.75476894e-01 -2.05747813e-01 -6.48156703e-01 6.15423977e-01
1.27040529e+00 8.49315226e-01 4.00398344e-01 2.88632423e-01
7.41165578e-01 8.59685600e-01 2.93646723e-01 4.20652777e-01
6.71410933e-02 3.94651741e-01 3.06136012e-01 2.91395374e-03
3.94009084e-01 1.94787517e-01 2.95444489e-01 1.05891514e+00
-4.60100353e-01 5.53201810e-02 -1.18944597e+00 4.59454477e-01
-1.42330110e+00 -1.11727798e+00 -3.42617810e-01 2.06559730e+00
3.14750254e-01 -4.49279010e-01 4.65990573e-01 5.66593604e-03
8.94546568e-01 8.23859498e-02 -3.73355865e-01 -8.41397166e-01
-2.13107005e-01 -1.77952588e-01 1.38960898e-01 3.03350389e-01
-8.77562582e-01 3.00911099e-01 6.86376190e+00 3.68286371e-01
-1.06778252e+00 -9.20410752e-02 3.06259036e-01 1.28881969e-02
-3.66972804e-01 -2.30695993e-01 -1.84172109e-01 5.77761710e-01
1.12150633e+00 -1.11696675e-01 2.80307770e-01 5.42576134e-01
5.23979247e-01 1.86013296e-01 -6.77075922e-01 1.11745489e+00
1.79020062e-01 -1.28798664e+00 2.97582209e-01 4.29411560e-01
7.11077273e-01 2.01136693e-01 4.98048365e-02 -2.78062940e-01
2.18900889e-01 -9.03378367e-01 -1.97790504e-01 4.81886417e-01
8.92858863e-01 -1.13656986e+00 7.36767113e-01 5.09899817e-02
-1.20615089e+00 8.77635777e-02 -4.84444946e-01 4.21002321e-03
3.27639937e-01 4.63567585e-01 -1.26830304e+00 4.53535736e-01
5.85778296e-01 7.01251864e-01 -5.16187012e-01 8.50831807e-01
1.62506104e-01 3.54049832e-01 -1.50675416e-01 -2.09377497e-01
1.13726016e-02 -1.01942849e+00 8.83200169e-01 1.47539973e+00
3.49103987e-01 1.15533569e-03 -1.36687383e-01 5.38230121e-01
6.14753664e-01 2.26702243e-01 -6.64640009e-01 -4.86663699e-01
3.91241103e-01 1.09135258e+00 -1.07642722e+00 -3.94715875e-01
-1.37616828e-01 1.04434168e+00 -1.29396349e-01 4.84476358e-01
-3.10648292e-01 -4.51128185e-01 1.34271634e+00 1.28378347e-01
3.68983090e-01 -4.59828168e-01 2.86266714e-01 -7.32646048e-01
-4.74809587e-01 -8.24646115e-01 7.03636169e-01 -2.44706884e-01
-1.32407665e+00 8.81399333e-01 6.39898255e-02 -1.17519093e+00
-4.24316406e-01 -6.30382538e-01 -5.09073734e-01 7.92531669e-01
-7.87507951e-01 -5.17453730e-01 1.36772707e-01 6.19282305e-01
2.39258289e-01 -5.33296704e-01 1.05272114e+00 -7.43325278e-02
-6.14905536e-01 -3.64876837e-02 1.07961953e+00 -1.41418934e-01
3.35351825e-01 -1.29381990e+00 3.29078019e-01 5.13715208e-01
-7.62244612e-02 1.18795741e+00 7.39671707e-01 -1.00272548e+00
-1.33172190e+00 -1.11041915e+00 7.84516573e-01 -4.19690043e-01
6.38857782e-01 -3.98711473e-01 -5.31166017e-01 3.29188555e-01
2.54560113e-01 -6.48192704e-01 1.47636127e+00 -1.67415574e-01
-3.46509635e-01 5.14970958e-01 -1.38642001e+00 6.13803208e-01
5.82638800e-01 -5.67814350e-01 -8.40386689e-01 2.76198506e-01
4.76311266e-01 5.04637122e-01 -8.28648269e-01 1.18868895e-01
6.45324409e-01 -1.22160518e+00 9.43580985e-01 -9.21093702e-01
-1.23519771e-01 -7.12677360e-01 -3.65793258e-02 -1.47296047e+00
-5.14484286e-01 -5.92031002e-01 2.43161619e-01 8.63272667e-01
3.30922365e-01 -9.34967518e-01 6.07505381e-01 9.33741853e-02
3.18680644e-01 -4.62214440e-01 -8.23310971e-01 -7.30671823e-01
-1.77814111e-01 8.39037001e-02 5.00618279e-01 1.17121327e+00
2.07936451e-01 1.24487087e-01 -2.19126433e-01 1.92684278e-01
6.19354248e-01 2.36185312e-01 5.83335221e-01 -1.51505518e+00
-2.21778881e-02 -4.31314796e-01 -1.04119325e+00 -1.37830749e-01
-2.57330596e-01 -9.64378238e-01 -3.67835850e-01 -1.54866016e+00
3.66549343e-01 -1.49490565e-01 -2.87458301e-01 -9.97326523e-02
-3.87693085e-02 1.82272524e-01 1.46072656e-01 6.24965072e-01
-1.54356867e-01 1.95185781e-01 3.85756433e-01 4.21633422e-02
-6.73935771e-01 -2.46891707e-01 -1.70296252e-01 6.25946105e-01
9.47521448e-01 -3.62260759e-01 -4.79708165e-01 3.61927748e-02
2.28687927e-01 -3.06817502e-01 -2.46818885e-01 -5.35389483e-01
-3.23483169e-01 -9.57777426e-02 4.55174536e-01 -8.67579281e-01
1.70594439e-01 -8.20753515e-01 8.74144971e-01 1.18318605e+00
1.29455626e-01 7.99363732e-01 4.54666726e-02 8.12684596e-01
1.26171947e-01 2.65792251e-01 5.99393189e-01 1.01212710e-01
-4.57872033e-01 -1.22575656e-01 -1.13917506e+00 -2.96594411e-01
1.29469466e+00 -5.72471619e-01 -3.74054849e-01 -9.97058675e-02
-7.40080774e-01 4.28452715e-03 8.66166711e-01 3.63738745e-01
7.78665483e-01 -1.07214236e+00 -8.97981763e-01 3.95867825e-01
1.06234282e-01 -6.85251355e-01 3.59262347e-01 1.10072589e+00
-1.26366174e+00 4.70621198e-01 -5.63551366e-01 -9.68299985e-01
-1.76918805e+00 1.17933559e+00 -1.73800424e-01 2.37877667e-02
-5.42130411e-01 5.55205107e-01 2.02250779e-01 -6.16461933e-01
-1.91159591e-01 1.70356870e-01 -7.22595811e-01 4.98512983e-01
6.64002836e-01 9.55951095e-01 -1.66075855e-01 -1.27379858e+00
-8.19544256e-01 7.17792273e-01 8.97647217e-02 2.64351517e-01
1.26213980e+00 -2.32490927e-01 -7.32610762e-01 5.35199761e-01
1.42865014e+00 1.93932205e-01 -6.46110058e-01 3.92144203e-01
2.23153308e-01 -2.81442553e-01 -6.01304173e-01 -5.14002442e-01
-5.85283935e-01 6.58498406e-01 1.02334285e+00 2.69231409e-01
1.05823636e+00 -3.09931599e-02 4.60088879e-01 2.26152673e-01
5.25826588e-02 -6.10761344e-01 -4.75539207e-01 1.97986335e-01
3.64749402e-01 -6.53793931e-01 -1.64652362e-01 -2.04882309e-01
-5.76862633e-01 9.07102406e-01 -2.42905393e-01 9.77660269e-02
5.51703036e-01 2.13464186e-01 5.02471387e-01 -6.08854830e-01
-5.56063592e-01 1.29029781e-01 1.54645458e-01 1.11786652e+00
3.89316529e-01 4.57725227e-01 -7.29918420e-01 -1.68629095e-01
-1.86738268e-01 -7.14658380e-01 2.75845796e-01 8.99870753e-01
-3.59103829e-01 -1.07972682e+00 -7.60663927e-01 6.56203687e-01
-2.86395159e-02 -2.23088712e-01 -8.42071652e-01 5.81685066e-01
1.82886362e-01 8.76744092e-01 2.99966335e-01 -4.51041341e-01
-1.30999103e-01 4.11300808e-01 2.05949564e-02 -3.25299919e-01
-3.68531525e-01 2.67724961e-01 -2.07414895e-01 -2.75195748e-01
-6.16233349e-01 -7.75501788e-01 -1.26223183e+00 -2.67215997e-01
1.42689452e-01 4.97085512e-01 6.86258256e-01 6.36537194e-01
7.43615627e-01 1.11793377e-01 4.99442756e-01 -8.50568712e-01
6.31509349e-02 -7.10629642e-01 -8.59195650e-01 5.94501615e-01
2.66636163e-01 -4.81302291e-01 -4.02357072e-01 1.46211430e-01] | [5.0355987548828125, 5.229383945465088] |
60c7cbd5-b137-4769-8aa6-fb610439c8bf | covariance-regression-with-random-forests | 2209.08173 | null | https://arxiv.org/abs/2209.08173v3 | https://arxiv.org/pdf/2209.08173v3.pdf | Covariance regression with random forests | Capturing the conditional covariances or correlations among the elements of a multivariate response vector based on covariates is important to various fields including neuroscience, epidemiology and biomedicine. We propose a new method called Covariance Regression with Random Forests (CovRegRF) to estimate the covariance matrix of a multivariate response given a set of covariates, using a random forest framework. Random forest trees are built with a splitting rule specially designed to maximize the difference between the sample covariance matrix estimates of the child nodes. We also propose a significance test for the partial effect of a subset of covariates. We evaluate the performance of the proposed method and significance test through a simulation study which shows that the proposed method provides accurate covariance matrix estimates and that the Type-1 error is well controlled. An application of the proposed method to thyroid disease data is also presented. CovRegRF is implemented in a freely available R package on CRAN. | ['Aurelie Labbe', 'Denis Larocque', 'Cansu Alakus'] | 2022-09-16 | null | null | null | null | ['epidemiology'] | ['medical'] | [ 0.5142719 -0.12711503 -0.24231252 -0.7449912 -0.45839885 -0.01769425
0.46986362 0.0791059 -0.31017244 0.9519614 0.39447033 -0.42116126
-0.545911 -0.8506434 -0.56825066 -0.894358 -0.20333152 0.4112067
-0.0903176 0.32232797 0.20941842 0.3396842 -1.1185379 -0.37258145
0.9801147 0.55692637 0.31006762 0.60006994 0.23888065 0.21911784
-0.31355885 -0.14801684 0.02903957 -0.2825456 -0.36120853 -0.04291353
0.01323573 0.09073698 0.39822435 0.8701096 0.44741207 0.1287749
1.2124016 -1.2278643 -0.15806612 0.58189803 -0.88403785 -0.13858435
-0.10069534 -0.23310831 0.82745826 -0.6296517 0.43166 1.2510502
0.6317237 0.04126607 -1.7168689 -0.93928975 -0.04315846 0.05424258
-1.5186024 -0.15209371 0.43030712 -0.70031106 0.42564702 0.25769284
0.32467222 0.6857471 0.7849082 0.25527304 1.6265501 -0.23188314
0.5281671 0.09549396 0.42845318 0.50899196 0.44728586 0.2130433
-0.34504247 -0.53835845 0.6724435 0.15684529 0.13015765 -0.44314027
-1.0397615 0.9920855 0.27321538 0.06992736 -0.652376 0.12568992
-0.00942171 -0.1095083 0.5668259 -0.36938044 -0.38562667 0.3540308
-0.92360085 0.21776636 0.6115336 0.58355886 0.6891431 -0.15409929
-0.243028 0.84501123 0.49708042 1.0296718 0.03064452 -0.6159262
0.14223559 0.37979427 0.00990504 -1.0018038 -0.65532607 -0.55605954
-1.367891 0.12521839 0.4000124 -0.32172334 -0.91911036 1.8768244
0.6761234 0.22553934 -0.29481986 0.49779853 0.49663904 0.24604768
0.40789798 -0.6635118 1.2491603 -0.3357993 -0.84814864 -0.08191559
0.28603172 -0.34545612 0.6082764 0.32546905 -0.602802 -0.21700957
-0.47718617 0.34252042 0.01616783 0.12324001 0.67191356 0.79028785
-0.690411 0.3406912 -0.9661149 -0.1858722 0.18492335 0.5770536
-0.43733975 -0.0855219 -0.8393349 0.6967819 -0.40888485 0.35825065
-0.6613844 -0.70920783 -0.6100965 0.04556843 0.10951938 -0.83240604
0.68704385 -0.61406946 -1.0734025 0.50951254 -0.6984083 -0.16227897
0.5052351 0.02216353 -0.08508419 -0.52809435 0.3131987 0.22896844
0.5420711 -0.63980234 -0.28875983 -0.8828018 -0.77364767 -0.14931297
0.2466915 0.1506844 0.03182634 -0.6034514 0.33953437 -0.88063496
-0.86285454 -0.24732217 -0.5237708 -0.05751548 0.05994352 -0.66340816
1.2106892 -1.7474366 0.15149164 0.7269633 0.01922265 -0.5323535
0.04017098 0.4590776 -0.08709797 -0.08551615 -0.6741104 0.21746919
-0.4588506 0.01724075 0.18667208 0.728462 0.11360793 0.39444187
-0.5252187 -0.30835068 0.01875164 0.56167513 -0.5321015 0.19653593
0.27051297 0.5092781 -0.4974157 0.38191938 0.9884047 -0.16045102
0.61526173 0.09038145 -0.22907257 0.00958349 -1.4190123 0.86555225
-0.2806724 0.24925536 0.10387679 -0.88495284 1.274312 0.07578109
0.45037067 -0.03675453 0.21662585 -0.01708038 0.28603902 -0.26242542
0.03259618 -0.50511545 -0.05762589 0.37066615 -0.25471377 0.1810331
0.01308066 0.04516315 1.2246302 -0.07679523 0.9477449 -0.56257254
0.58368486 -0.2967084 0.7392363 0.8243416 0.07232963 0.21144716
0.6243192 -0.02290907 -0.73995894 -1.0974427 -0.7510785 0.76545185
-0.49234322 -0.05865512 -0.31572214 -0.33802927 0.30060452 0.61685497
-0.92268825 0.28241032 -0.12408849 -1.3445255 0.04021511 0.50776416
0.19539736 -0.7566957 -0.30822453 0.08809735 -0.02502379 -0.6723781
-0.12527154 0.4683698 -1.2318333 -1.0478864 -0.5830939 -0.4805748
0.8148284 0.05718653 0.6163307 -0.32456195 -0.06175264 -0.10652541
0.01389722 -0.30876666 -0.24319732 -0.07376132 0.0098492 -0.02568317
0.26650897 -0.752256 -0.54954815 0.5291671 -0.5832695 0.14137533
0.5474408 0.97176766 0.65369 0.0440258 0.7774013 -1.2193756
0.4244299 -0.79150885 -0.84101325 0.18566473 -0.87665033 0.21449083
0.11198186 -0.39599463 -0.9740517 0.32757908 0.06937603 0.16160712
-0.06897897 0.76965433 -0.3045386 0.32053924 0.34212443 -0.01814225
0.10034701 -0.6945841 0.16336589 0.8191319 0.3934236 -0.4113035
0.47441894 0.29699805 0.7519439 -0.69879586 -0.29034492 -0.44998324
-0.65809983 -0.19761986 0.6620031 -0.7715399 -0.88550436 0.4249871
-0.48177457 -0.23295115 0.29999954 0.8548766 -0.584692 0.07786264
-0.0396571 -1.0770394 -0.3338187 -1.038172 0.73696846 0.0355502
-0.23582116 -0.9750284 0.3918604 0.23918773 0.07040544 0.26940677
1.1716977 -0.60631984 -0.24966846 -0.22178808 -0.20266977 -0.04007343
0.25422207 0.23482637 -0.62475425 -0.07783835 -0.09155655 0.30155146
0.668975 1.1011375 0.9127878 -0.22271518 -0.5230143 0.37164834
1.441026 0.13695596 0.46812573 0.03298857 0.36216283 0.8397037
0.6909324 0.6993615 0.38718963 0.63518745 0.00985259 0.00971117
0.2131358 -0.21685237 0.157292 0.6550621 0.03737011 -0.07294356
-0.90139633 0.5090169 -1.7651664 -0.7190166 -0.88235354 2.5247664
0.78107 -0.31813496 0.21777725 0.06087749 0.80449545 -0.35359043
-0.4510651 -0.56012064 0.07847439 0.25616056 0.7736549 0.63251495
-0.75852823 0.538319 7.4821463 0.46002176 -0.77113146 -0.15857047
0.60681736 0.24806139 -0.3497398 0.20832117 -0.8460731 0.2664631
1.0337352 -0.15459405 0.15332593 0.5634513 0.8045815 -0.7068922
-0.7561853 0.14137225 -0.42031175 -0.72929037 -0.57346797 0.34761357
0.669529 -0.11866523 -0.04449522 -0.02261691 0.81801385 -0.9642498
0.11401186 0.673252 0.7137809 -0.6295753 0.7133505 0.44990772
-0.7266206 0.17608562 -0.39974177 -0.19673058 0.03269348 1.1920766
-1.3363246 0.2738466 0.5120167 0.56248766 -0.5534007 1.1721241
-0.37929222 0.95868313 -0.26903108 -0.14910194 -0.38056475 -0.4327733
0.52100503 1.0167001 0.32714653 -0.01586104 -0.28539434 0.6775607
0.41598636 0.55160445 -0.4005519 0.34579867 0.39410374 1.0647337
-0.7562319 0.02496265 -0.3482665 0.15635139 0.1633588 0.35315683
-0.4755936 0.10810652 0.5164019 0.09501979 0.41645503 0.05241185
-0.63307726 -0.9132481 -0.23586038 -0.6462351 0.48604485 -0.71045107
-1.2165397 0.11220227 0.414792 -0.8387811 -0.35177633 -0.38098225
-0.4420208 1.1126372 -0.9571093 -0.79708093 -0.04195718 0.41051447
0.05409402 0.11879634 0.76858073 -0.12293145 -0.7658942 0.30867657
0.5790504 -0.2582194 0.6058949 -1.0319533 0.03741299 0.67471594
-0.3288231 0.8576269 0.9819 -1.2085778 -0.93403274 -0.7963953
1.1644112 0.0216706 0.65589136 -0.38247645 -0.6454369 0.6700171
-0.2494967 -0.15242171 0.93896717 0.59694743 -0.26479152 -0.2926885
-1.3636928 0.33144847 0.5027833 0.03417704 -0.14036833 0.08973479
0.20691784 0.18156879 -1.0903124 0.72088814 1.0294967 -0.80273145
0.6413727 -0.6211118 0.24796835 -0.42125523 -0.31142843 -1.2151369
-0.490769 -0.06771234 0.5667116 1.1924392 0.42340714 -0.5759762
0.7152707 0.69730645 0.5503822 -0.53064305 -0.9807501 -0.5657453
0.138463 -0.47845554 0.42484653 0.53707486 -0.25978988 0.46157682
-0.56986433 0.15974857 0.88730586 0.08207797 0.82459575 -1.4446727
-0.31235358 0.22109438 -0.4552092 -0.2975875 0.03805716 -0.66430074
0.14414442 -1.3809276 0.8693601 -0.53570616 -0.29110867 0.37550423
-0.45569775 -0.20211531 -0.22308452 -0.07311841 0.22299692 0.38320315
1.1063584 0.12080593 -0.26216212 0.7693759 -0.6568162 0.52857214
0.58548945 -1.1227734 -0.33046886 0.24198653 0.09147702 0.5995943
0.25562745 -0.57713634 -0.0665677 -0.49305844 0.4167767 -0.7939627
-0.01113503 -0.8357616 0.8436939 0.54850626 -0.5214165 0.15154973
-0.12742199 0.7785533 0.13671431 -0.2598746 0.68985945 0.35272256
0.11583629 -0.12281364 -0.68290854 -0.25367206 0.68889153 0.12750843
-0.03185141 -0.519646 -0.8129441 0.22473928 0.01217955 -0.07094463
0.3308267 -0.9924086 -0.9110176 0.2822777 -0.13318963 -0.48136866
0.31033167 1.1633977 0.02084265 0.32962063 -0.25120193 -0.5894921
-1.7314292 0.2791572 -0.02284689 -0.46123737 -0.16022646 0.3870247
0.46918607 -0.5628337 0.02525532 -0.22575536 -0.33638567 -0.09181365
0.39487568 0.75167334 -0.08793928 -0.6089639 -0.62610114 0.5663904
0.24546275 -0.40621996 1.5060353 -0.16779563 -0.38717967 0.49047634
0.9875488 0.23587981 -0.87688196 -0.2379083 0.04777582 -0.40732753
0.15495992 -0.9430646 -0.9101302 0.67669326 0.7267098 -0.34230962
1.2951534 -0.2245612 -0.26528463 0.02236527 0.15496816 -0.43976936
-0.6871172 0.14149171 0.63526237 -1.0264173 0.46293744 -0.5171561
-0.66489625 0.7187854 0.22895817 -0.15467077 1.0638846 0.3886431
-0.0491585 0.11646327 -1.0233555 -0.21417716 0.23427771 0.7684138
0.6838494 0.5148853 -1.0002161 0.6893256 -0.2161737 0.23231085
0.5984788 0.41973388 -0.26680872 -1.2709702 -0.54420996 0.6794082
-0.5570378 -0.13788107 -0.22801608 0.65496415 -0.18487427 1.0283347
0.02565982 -0.23447932 0.00793226 0.09630878 0.19430801 -0.29697034
-0.5374038 0.5318664 0.03568164 -0.22479247 -0.4920622 -1.1121967
-0.854913 -0.15376836 -0.59584284 0.21391961 0.97968423 0.759453
0.03400757 0.28251025 0.86839217 -0.01774903 -0.49654087 -1.061309
-0.826811 -0.22862822 0.06698623 -0.70337343 -0.27391824 -0.16467638] | [7.727998733520508, 4.947638034820557] |
1954a9be-004b-4d2a-9c3e-ebbf636232c7 | diffusion-policies-as-an-expressive-policy | 2208.06193 | null | https://arxiv.org/abs/2208.06193v2 | https://arxiv.org/pdf/2208.06193v2.pdf | Diffusion Policies as an Expressive Policy Class for Offline Reinforcement Learning | Offline reinforcement learning (RL), which aims to learn an optimal policy using a previously collected static dataset, is an important paradigm of RL. Standard RL methods often perform poorly in this regime due to the function approximation errors on out-of-distribution actions. While a variety of regularization methods have been proposed to mitigate this issue, they are often constrained by policy classes with limited expressiveness that can lead to highly suboptimal solutions. In this paper, we propose representing the policy as a diffusion model, a recent class of highly-expressive deep generative models. We introduce Diffusion Q-learning (Diffusion-QL) that utilizes a conditional diffusion model to represent the policy. In our approach, we learn an action-value function and we add a term maximizing action-values into the training loss of the conditional diffusion model, which results in a loss that seeks optimal actions that are near the behavior policy. We show the expressiveness of the diffusion model-based policy, and the coupling of the behavior cloning and policy improvement under the diffusion model both contribute to the outstanding performance of Diffusion-QL. We illustrate the superiority of our method compared to prior works in a simple 2D bandit example with a multimodal behavior policy. We then show that our method can achieve state-of-the-art performance on the majority of the D4RL benchmark tasks. | ['Mingyuan Zhou', 'Jonathan J Hunt', 'Zhendong Wang'] | 2022-08-12 | null | null | null | null | ['d4rl'] | ['robots'] | [-2.59866774e-01 1.14908054e-01 -7.82791436e-01 -5.49769513e-02
-1.00029230e+00 -4.52777952e-01 8.07447135e-01 -1.13628186e-01
-5.44002056e-01 1.08489835e+00 5.41359961e-01 -3.40196460e-01
-3.30688685e-01 -7.25495577e-01 -8.62876892e-01 -1.01272500e+00
1.16647534e-01 5.74587882e-01 -2.79287323e-02 -2.64626652e-01
4.71588001e-02 2.80150741e-01 -1.15754306e+00 -2.85246992e-03
1.07809460e+00 1.12029052e+00 1.02449089e-01 4.91851121e-01
-6.68242350e-02 1.27211559e+00 -6.82746828e-01 -3.23578119e-01
4.51327115e-01 -5.37501812e-01 -4.73093390e-01 -1.30694099e-02
-2.35272720e-02 -8.37115765e-01 -5.51380754e-01 1.01806808e+00
6.30666316e-01 4.16930586e-01 8.18018198e-01 -1.12550116e+00
-7.79068112e-01 6.58327103e-01 -6.34310722e-01 3.14368121e-02
-1.56886373e-02 4.62339431e-01 1.24673355e+00 -2.86025494e-01
4.76414323e-01 1.50027251e+00 3.99992853e-01 7.10267901e-01
-1.35558629e+00 -5.48779488e-01 4.45025355e-01 -1.12524137e-01
-8.27998996e-01 -1.71458170e-01 8.35828602e-01 -2.70771503e-01
8.38881552e-01 -1.89969435e-01 7.86979139e-01 1.39920855e+00
1.06839433e-01 1.43660736e+00 1.33113134e+00 -2.32124254e-01
6.11048579e-01 -7.90041536e-02 -3.70096505e-01 6.03589296e-01
-1.05977803e-01 5.54273069e-01 -3.46705884e-01 -3.01951021e-01
9.97726023e-01 1.21115997e-01 4.62395558e-03 -7.48281956e-01
-6.44118607e-01 1.26432955e+00 4.86397505e-01 1.00807343e-02
-7.18883634e-01 7.11121261e-01 3.77166122e-01 3.24206948e-01
7.58307219e-01 2.68701494e-01 -3.26803058e-01 -4.80759740e-01
-6.77983940e-01 6.99998796e-01 6.35077238e-01 5.78526020e-01
6.33725166e-01 2.59614468e-01 -7.30063617e-01 9.53286886e-01
2.82426000e-01 5.61333477e-01 4.16924804e-01 -1.25774670e+00
4.36248153e-01 2.36608505e-01 5.09640813e-01 -4.63680089e-01
-1.79193288e-01 -6.34538829e-01 -5.00006199e-01 3.67034316e-01
6.24917746e-01 -3.61141026e-01 -8.22529435e-01 2.04302907e+00
3.76024872e-01 -9.15110558e-02 1.03264578e-01 7.58179188e-01
2.79213078e-02 6.35703862e-01 2.43802890e-01 -4.14721638e-01
5.56610584e-01 -1.11257124e+00 -7.70948172e-01 -1.34293064e-01
5.85823596e-01 -2.87178069e-01 1.34635162e+00 3.69196117e-01
-1.21323121e+00 -2.25591555e-01 -6.78844690e-01 3.42845529e-01
9.01270881e-02 -1.07557833e-01 7.31721461e-01 4.63403046e-01
-9.73739028e-01 8.53871047e-01 -8.86031389e-01 -6.53297529e-02
7.16907144e-01 2.69640207e-01 2.91898936e-01 -3.76781709e-02
-1.08642900e+00 8.49019587e-01 1.42620653e-01 -4.03236687e-01
-1.31933618e+00 -7.90526927e-01 -4.10844594e-01 9.57541466e-02
9.03079569e-01 -6.74898028e-01 1.76435292e+00 -1.19695961e+00
-2.14506412e+00 2.16387570e-01 2.73837507e-01 -7.91234910e-01
9.04973507e-01 -4.38288480e-01 2.88981423e-02 1.56569377e-01
-1.18052103e-02 5.38715243e-01 1.18265104e+00 -1.14877534e+00
-7.31076717e-01 -1.97378427e-01 2.91090250e-01 4.47661459e-01
-3.53322774e-01 -3.66539478e-01 -3.79250944e-01 -8.32766473e-01
-7.09307373e-01 -9.39328194e-01 -3.17858130e-01 -2.03211278e-01
-8.46130848e-02 -4.79569793e-01 5.72743475e-01 -3.64037037e-01
1.39709032e+00 -2.00318003e+00 2.86295295e-01 1.18009873e-01
-3.69127057e-02 2.68136710e-01 -9.32628140e-02 4.87303436e-01
4.80007857e-01 -1.70536861e-02 -2.36462384e-01 -4.15417433e-01
4.82380569e-01 5.59986949e-01 -6.32145226e-01 5.23413599e-01
-6.92256773e-03 1.16789460e+00 -1.04912519e+00 -1.14968255e-01
6.82034716e-02 2.31252089e-01 -9.03517663e-01 4.32951987e-01
-8.57136548e-01 4.96371031e-01 -7.67734826e-01 4.10958201e-01
2.21500546e-01 -1.20514356e-01 1.78652883e-01 3.22581291e-01
5.38397171e-02 9.61313248e-02 -8.22731853e-01 1.69341373e+00
-3.58450741e-01 -2.34521623e-03 9.20911580e-02 -1.12167346e+00
6.77596509e-01 1.50626197e-01 8.77675116e-01 -8.82816732e-01
-4.11032438e-02 3.03130537e-01 -4.10083905e-02 -3.37274611e-01
2.46652767e-01 -4.05415803e-01 5.59434481e-02 7.89593697e-01
5.17975017e-02 -1.91942975e-01 1.41688719e-01 2.52360441e-02
1.01659679e+00 6.11360729e-01 3.14639807e-01 -2.33066142e-01
-3.92416529e-02 -2.62359113e-01 5.78350961e-01 1.20463955e+00
-3.49694073e-01 2.19225325e-02 8.80869329e-01 -1.99118361e-01
-1.03588188e+00 -8.61729026e-01 1.72253042e-01 1.35149491e+00
-2.31346652e-01 -1.76571205e-01 -5.45737147e-01 -1.16760778e+00
5.71463525e-01 8.37962210e-01 -8.21900964e-01 -4.18772787e-01
-5.72604239e-01 -7.66780913e-01 4.79345322e-01 5.69059253e-01
5.17412484e-01 -1.02940536e+00 -5.37001669e-01 4.43492502e-01
2.19716340e-01 -6.95612013e-01 -5.32663107e-01 1.80337563e-01
-8.62823427e-01 -7.65962601e-01 -1.12009907e+00 -1.66413605e-01
2.89023072e-01 -2.43677676e-01 1.09622478e+00 -1.87238142e-01
2.83864230e-01 7.39726424e-01 -2.55127132e-01 -4.60260540e-01
-4.91298348e-01 -4.36219014e-02 1.08953498e-01 7.78085068e-02
2.07614470e-02 -4.85345215e-01 -9.15514410e-01 5.54390289e-02
-9.73936856e-01 -3.00190777e-01 5.81796587e-01 1.10888386e+00
5.56488752e-01 -2.12319806e-01 7.92808592e-01 -8.57792854e-01
1.04972184e+00 -6.39515102e-01 -9.18498278e-01 1.14137560e-01
-9.16336775e-01 4.90579426e-01 7.13200808e-01 -7.49356091e-01
-1.15848565e+00 -2.66715109e-01 -1.99864805e-01 -7.54206300e-01
2.42813170e-01 3.74443918e-01 1.85977638e-01 1.03979662e-01
4.75835830e-01 1.67105243e-01 2.35421181e-01 -4.96538520e-01
6.47982657e-01 3.29643250e-01 1.23229183e-01 -1.18936598e+00
4.31061298e-01 4.92431253e-01 -1.62688177e-02 -3.53374451e-01
-1.17446589e+00 -6.44399002e-02 1.31229013e-01 -2.40636289e-01
5.62330365e-01 -7.53083766e-01 -8.52334380e-01 2.74338543e-01
-5.83444834e-01 -1.08773255e+00 -8.73384356e-01 5.08656085e-01
-1.16808128e+00 9.98201147e-02 -6.22469127e-01 -1.30184066e+00
-2.22459197e-01 -1.14515913e+00 9.11643207e-01 1.21472500e-01
1.76169485e-01 -1.26447749e+00 4.57998782e-01 -5.82323633e-02
6.63811564e-01 1.39688432e-01 1.02543104e+00 -4.57545251e-01
-3.42142224e-01 2.27382362e-01 1.46903366e-01 5.51096618e-01
4.36140504e-03 -2.24832103e-01 -7.36216009e-01 -5.02296984e-01
6.25582784e-02 -9.02200937e-01 1.15327573e+00 8.01745474e-01
1.22397184e+00 -5.21373868e-01 -1.67434365e-02 6.23202205e-01
1.33419299e+00 3.99735063e-01 3.88715446e-01 3.77127886e-01
3.93176913e-01 2.81544864e-01 6.95771098e-01 8.33951473e-01
2.11145669e-01 6.57203853e-01 5.48006654e-01 3.55718695e-02
1.60800800e-01 -7.64901221e-01 6.64017320e-01 3.65957439e-01
-7.73945749e-02 -3.38371366e-01 -6.27291739e-01 2.44096726e-01
-2.26909661e+00 -9.99475658e-01 7.27803826e-01 2.08106613e+00
1.02193654e+00 2.29132354e-01 6.26143873e-01 -2.90260494e-01
1.76847070e-01 3.31163496e-01 -1.08412719e+00 -2.89023519e-01
-5.73035590e-02 1.88917816e-01 5.76853216e-01 5.38279951e-01
-1.00604892e+00 1.12157977e+00 6.86151075e+00 1.24307764e+00
-1.10562491e+00 2.63737053e-01 7.12994158e-01 -4.20327485e-01
-4.03706551e-01 -1.81103200e-01 -7.93493450e-01 5.01240492e-01
6.64957702e-01 -9.37590897e-02 7.21257031e-01 9.46843266e-01
3.77242714e-01 -1.43717378e-01 -9.17184591e-01 8.16328764e-01
-3.37011755e-01 -1.27809632e+00 -1.66136157e-02 3.95531863e-01
1.03537607e+00 8.15133452e-02 5.04533589e-01 8.62500668e-01
9.99611080e-01 -9.66937125e-01 9.25195575e-01 5.50051570e-01
6.48963630e-01 -1.01218283e+00 2.23308221e-01 6.02724791e-01
-5.93860269e-01 -6.50760710e-01 -5.09282112e-01 -2.67219860e-02
-4.95680235e-02 3.27813536e-01 -4.54433501e-01 2.89271802e-01
4.16680574e-01 7.89745092e-01 5.93432449e-02 7.71513641e-01
-4.17441607e-01 8.72903049e-01 -2.75978237e-01 -3.07493061e-01
8.41422617e-01 -4.84100580e-01 5.89352548e-01 7.96102166e-01
1.42873406e-01 -9.55477580e-02 4.01386529e-01 9.60670471e-01
-2.97017485e-01 8.91602933e-02 -5.74104249e-01 -3.10559422e-01
1.65730238e-01 7.46091604e-01 -2.75855929e-01 -2.50904918e-01
-3.62863213e-01 7.19273448e-01 8.58432591e-01 7.35185802e-01
-8.80806625e-01 2.93995500e-01 7.83246994e-01 -1.27948701e-01
5.89336574e-01 -2.35344380e-01 1.93927035e-01 -1.00148737e+00
-1.46722332e-01 -1.07535958e+00 4.77359861e-01 -2.73492247e-01
-1.43733251e+00 3.45488638e-02 1.15651175e-01 -9.88186300e-01
-5.13413489e-01 -3.81811708e-01 -3.12669516e-01 4.56695259e-01
-1.72585464e+00 -8.84374201e-01 4.60720778e-01 7.26850092e-01
5.22878766e-01 -2.91577488e-01 4.75996166e-01 1.51492164e-01
-5.42652786e-01 5.03460407e-01 7.63430655e-01 -2.48731390e-01
6.98779941e-01 -1.54081583e+00 -1.25555590e-01 3.03157806e-01
-1.72630306e-02 3.10335487e-01 6.18121266e-01 -6.06081665e-01
-1.39524674e+00 -9.67713535e-01 -3.08580827e-02 -2.03688815e-01
8.62141490e-01 -1.98609367e-01 -6.23784542e-01 7.04982817e-01
1.06685922e-01 3.46644372e-02 3.58311534e-01 3.12184505e-02
-1.25880763e-01 -2.05357060e-01 -1.00690174e+00 6.70822740e-01
9.12412584e-01 -3.47716957e-01 -3.27094734e-01 4.03314918e-01
6.78768337e-01 -3.28347057e-01 -7.74624288e-01 1.17022805e-01
4.42524910e-01 -9.73837912e-01 8.67765546e-01 -1.00230014e+00
4.07830060e-01 2.05495328e-01 -5.10058291e-02 -1.72291696e+00
-2.21769392e-01 -1.11406565e+00 -7.00821996e-01 9.48764861e-01
1.03682630e-01 -6.38237476e-01 7.35272944e-01 4.42624897e-01
1.08107939e-01 -1.21885002e+00 -9.49497342e-01 -9.85573709e-01
6.80322826e-01 -3.88151944e-01 6.64048076e-01 5.09540975e-01
-1.91189513e-01 9.93093625e-02 -7.57313371e-01 -4.77691650e-01
6.62306428e-01 5.58726639e-02 7.44871676e-01 -6.85326338e-01
-8.15092385e-01 -8.43064904e-01 3.77934724e-01 -1.58326387e+00
4.40595835e-01 -6.29875124e-01 -3.76019813e-02 -1.36007965e+00
9.08662155e-02 -5.76259017e-01 -4.37009990e-01 3.73552054e-01
-2.39513293e-01 -3.64369273e-01 2.99733490e-01 3.53300303e-01
-6.80290222e-01 1.27182674e+00 1.59669590e+00 -1.02804013e-01
-5.45065165e-01 2.71632314e-01 -6.37346148e-01 6.92370176e-01
7.70655751e-01 -6.25910878e-01 -6.89222276e-01 -2.72786856e-01
2.73056924e-01 2.20354229e-01 1.13587148e-01 -5.43404818e-01
-2.14932840e-02 -5.26404858e-01 1.51856974e-01 -1.60994008e-01
4.09706771e-01 -5.78691244e-01 -3.31294417e-01 6.47373259e-01
-6.56294346e-01 -1.50570631e-01 -1.39537957e-02 9.80025828e-01
5.54189552e-03 -7.22095091e-03 8.53350163e-01 -1.22462705e-01
-3.43856782e-01 5.75931370e-01 -2.29144946e-01 4.53421593e-01
9.70315814e-01 3.49546283e-01 -2.30419040e-01 -7.60928988e-01
-4.22707319e-01 3.49847943e-01 2.92748898e-01 1.74431115e-01
3.77287388e-01 -1.48050284e+00 -6.62425280e-01 -1.42222002e-01
-2.51151681e-01 -2.23721936e-01 -1.76180992e-02 9.34481025e-01
-9.21686590e-02 2.88666964e-01 3.28992791e-02 -3.18356037e-01
-4.43536878e-01 6.56493306e-01 5.49245834e-01 -9.44955230e-01
-7.34731436e-01 6.51858389e-01 2.69875854e-01 -2.19742358e-01
4.83877629e-01 -2.60024577e-01 -2.29074974e-02 -6.67049214e-02
2.92963028e-01 3.53791028e-01 -3.95948887e-01 -1.29398495e-01
1.64081648e-01 2.27336362e-01 2.45827846e-02 -5.32920301e-01
1.32461357e+00 1.67417139e-01 3.89264464e-01 3.56848478e-01
9.22597766e-01 -6.94609284e-02 -2.12953687e+00 -4.35603321e-01
-2.84675937e-02 -4.35176849e-01 1.96678996e-01 -8.94343138e-01
-1.05799401e+00 6.76194608e-01 4.65515345e-01 2.47578621e-01
9.14682686e-01 -3.80093679e-02 7.57017553e-01 2.91274220e-01
2.40992546e-01 -1.55952275e+00 4.13443834e-01 5.67687154e-01
8.70235741e-01 -1.26998615e+00 -2.03965735e-02 4.23746854e-01
-9.98714626e-01 7.49957621e-01 3.90448034e-01 -3.73769313e-01
4.69714791e-01 1.88002825e-01 -1.36307701e-01 1.08457424e-01
-9.42937374e-01 -4.66311455e-01 1.34376571e-01 4.76127595e-01
1.37323625e-02 7.94282928e-02 -3.61275315e-01 5.69874048e-01
1.73577189e-01 1.53327793e-01 -9.39120948e-02 1.11002147e+00
-4.53756005e-01 -1.17302859e+00 -1.57465398e-01 4.67188030e-01
-6.42017663e-01 1.80092826e-01 -2.63679326e-01 9.12897289e-01
-3.53338808e-01 7.47289717e-01 -3.57277244e-02 -3.57198603e-02
2.73249269e-01 6.25681505e-02 7.15957761e-01 -1.91126972e-01
-6.44721866e-01 5.03485680e-01 -1.87188208e-01 -8.29691827e-01
-4.10660028e-01 -6.06853247e-01 -1.15998268e+00 -2.35083267e-01
1.33285612e-01 1.62196115e-01 4.38859165e-01 9.67817545e-01
2.70121872e-01 4.38996375e-01 6.56182170e-01 -5.56843996e-01
-1.56533921e+00 -8.47202837e-01 -8.90531301e-01 5.83204389e-01
5.29723287e-01 -9.52564538e-01 -3.14081937e-01 -5.74010015e-01] | [4.0703043937683105, 2.3204474449157715] |
125c0945-151c-4a29-8b73-e7ea8994656b | video-to-video-synthesis | 1808.06601 | null | http://arxiv.org/abs/1808.06601v2 | http://arxiv.org/pdf/1808.06601v2.pdf | Video-to-Video Synthesis | We study the problem of video-to-video synthesis, whose goal is to learn a
mapping function from an input source video (e.g., a sequence of semantic
segmentation masks) to an output photorealistic video that precisely depicts
the content of the source video. While its image counterpart, the
image-to-image synthesis problem, is a popular topic, the video-to-video
synthesis problem is less explored in the literature. Without understanding
temporal dynamics, directly applying existing image synthesis approaches to an
input video often results in temporally incoherent videos of low visual
quality. In this paper, we propose a novel video-to-video synthesis approach
under the generative adversarial learning framework. Through carefully-designed
generator and discriminator architectures, coupled with a spatio-temporal
adversarial objective, we achieve high-resolution, photorealistic, temporally
coherent video results on a diverse set of input formats including segmentation
masks, sketches, and poses. Experiments on multiple benchmarks show the
advantage of our method compared to strong baselines. In particular, our model
is capable of synthesizing 2K resolution videos of street scenes up to 30
seconds long, which significantly advances the state-of-the-art of video
synthesis. Finally, we apply our approach to future video prediction,
outperforming several state-of-the-art competing systems. | ['Jun-Yan Zhu', 'Ming-Yu Liu', 'Ting-Chun Wang', 'Jan Kautz', 'Andrew Tao', 'Guilin Liu', 'Bryan Catanzaro'] | 2018-08-20 | video-to-video-synthesis-1 | http://papers.nips.cc/paper/7391-video-to-video-synthesis | http://papers.nips.cc/paper/7391-video-to-video-synthesis.pdf | neurips-2018-12 | ['video-to-video-synthesis'] | ['computer-vision'] | [ 7.12799847e-01 -7.70029724e-02 -5.54601997e-02 3.56910154e-02
-9.37162757e-01 -8.63465726e-01 8.41823280e-01 -7.58032560e-01
1.11778274e-01 6.61019921e-01 2.19264135e-01 -9.40125883e-02
3.75771791e-01 -6.84453011e-01 -1.42835486e+00 -6.01578236e-01
3.33243728e-01 9.68806222e-02 2.48843491e-01 -1.89397007e-01
5.18785529e-02 3.95662069e-01 -1.46304238e+00 5.61236918e-01
5.76991022e-01 1.09739256e+00 1.46471739e-01 1.01768708e+00
2.28950620e-01 1.09363246e+00 -5.57028830e-01 -5.50386965e-01
6.07713521e-01 -8.61403644e-01 -6.28017068e-01 4.85335618e-01
9.45055902e-01 -6.88868523e-01 -8.85724843e-01 9.27954078e-01
1.72270775e-01 1.86537772e-01 5.49662054e-01 -1.37458348e+00
-9.78383839e-01 3.47899258e-01 -4.40852463e-01 1.86582059e-02
5.85552335e-01 7.07620740e-01 8.36384594e-01 -8.73107135e-01
1.12084699e+00 1.40346551e+00 2.60218918e-01 7.82983303e-01
-1.60094273e+00 -7.57169247e-01 1.75385505e-01 -8.21969658e-02
-1.27446651e+00 -6.42629981e-01 7.30146766e-01 -5.37602484e-01
7.39409924e-01 1.50041312e-01 6.79939568e-01 1.62467539e+00
2.29873523e-01 7.97746122e-01 9.44301248e-01 -1.29935443e-01
1.07418597e-01 -3.61262560e-01 -7.82636881e-01 6.52490556e-01
-1.85217589e-01 3.77808511e-01 -5.84905505e-01 1.93495005e-01
1.22292328e+00 1.10982403e-01 -3.98012906e-01 -3.24823380e-01
-1.38573527e+00 6.16652310e-01 2.37607121e-01 1.27604026e-02
-3.50907117e-01 5.66745996e-01 1.75463736e-01 4.60790634e-01
3.81489217e-01 4.70813274e-01 3.97292711e-02 -2.99766865e-02
-1.32067811e+00 6.98942363e-01 5.98615170e-01 1.13209307e+00
4.08584237e-01 5.66254497e-01 -4.39899445e-01 4.88752007e-01
-1.33625790e-01 7.88985670e-01 1.34060308e-01 -1.45695806e+00
5.58761716e-01 1.26261234e-01 1.42732307e-01 -1.01445472e+00
3.54268759e-01 3.30600403e-02 -8.45422685e-01 2.44813249e-01
4.01007205e-01 -1.59162030e-01 -1.05846477e+00 1.67438269e+00
6.80081993e-02 6.92093372e-01 8.91154855e-02 9.90459442e-01
7.31464565e-01 1.08987701e+00 -1.21124096e-01 -2.40978688e-01
9.20758963e-01 -1.30360353e+00 -7.21001923e-01 -3.17042440e-01
-1.49565116e-01 -7.78905153e-01 9.25449252e-01 3.63933772e-01
-1.46181011e+00 -7.32881427e-01 -8.60050797e-01 -1.82016090e-01
3.20506841e-02 -2.13014185e-01 2.06457511e-01 2.13166460e-01
-1.18893301e+00 7.21022844e-01 -6.33225262e-01 -1.36628017e-01
3.77073407e-01 1.04080029e-01 -5.23717046e-01 -2.82712877e-01
-1.04707420e+00 4.03413236e-01 3.38182241e-01 -1.77298203e-01
-1.38340652e+00 -9.33082640e-01 -9.15475190e-01 -7.77123347e-02
5.89872777e-01 -9.42554533e-01 1.33991587e+00 -1.44409728e+00
-1.69932151e+00 6.72939837e-01 -4.54465933e-02 -5.64199388e-01
8.90711367e-01 -2.48437390e-01 -4.05296981e-01 4.92810220e-01
5.03079519e-02 1.09269297e+00 1.40950179e+00 -1.39895332e+00
-5.85949957e-01 1.96364626e-01 1.70369774e-01 6.91171736e-02
6.93120062e-02 -6.56128228e-02 -9.53420997e-01 -1.18523991e+00
-4.00880337e-01 -1.01482022e+00 -2.23656878e-01 2.71357149e-01
-3.56109560e-01 3.71184647e-01 1.02115202e+00 -8.10525596e-01
1.02369881e+00 -2.17642927e+00 7.16863096e-01 -2.77414888e-01
6.84732422e-02 1.06587209e-01 -3.33530158e-01 4.98226583e-01
-1.61681205e-01 1.67197898e-01 -2.75032818e-01 -3.82303655e-01
-1.62171349e-01 1.49873272e-01 -8.55364203e-01 3.27658534e-01
4.91975456e-01 1.29391658e+00 -1.05469882e+00 -4.33534175e-01
3.75052541e-01 4.35543150e-01 -8.69865954e-01 5.66796064e-01
-6.30581975e-01 7.38981068e-01 -1.92270473e-01 7.12376118e-01
2.85581857e-01 -3.45900685e-01 5.93508258e-02 -1.89402655e-01
1.98019430e-01 -1.49097025e-01 -8.50735486e-01 1.93039095e+00
-3.99071574e-01 8.33212376e-01 -1.06223993e-01 -7.57804990e-01
6.94557488e-01 3.39237869e-01 5.33580959e-01 -6.95341408e-01
-5.07128611e-02 1.91708490e-01 -3.84561986e-01 -4.08571094e-01
5.78802466e-01 -9.81479883e-02 -1.65695220e-01 2.36366481e-01
6.90315366e-02 -5.55581331e-01 2.90252298e-01 2.86091834e-01
1.10991871e+00 5.96552849e-01 5.43633364e-02 1.67165369e-01
2.59122491e-01 -1.54685557e-01 3.61446112e-01 5.95074773e-01
1.07804909e-01 1.13553262e+00 5.76298535e-01 -3.58451217e-01
-1.65322471e+00 -1.26940525e+00 3.11479807e-01 6.08554780e-01
2.57922590e-01 -4.11849022e-01 -9.15942073e-01 -5.39031506e-01
1.57324225e-02 5.09977400e-01 -5.53723216e-01 -1.28781438e-01
-8.91486943e-01 1.02705114e-01 5.78905702e-01 4.52863693e-01
4.57291216e-01 -1.26299596e+00 -2.94597268e-01 2.67785281e-01
-2.87355304e-01 -1.58241963e+00 -9.56183970e-01 -5.36614537e-01
-7.42813587e-01 -8.82264793e-01 -1.11559498e+00 -7.36305118e-01
3.92902166e-01 3.62747997e-01 1.35648715e+00 -1.02845140e-01
-2.47117892e-01 4.95444238e-01 -2.12483674e-01 2.06977651e-01
-7.95069396e-01 -2.56305546e-01 -8.88623670e-02 3.64265501e-01
-4.38849181e-01 -6.85935020e-01 -8.00367236e-01 2.70652801e-01
-1.41994047e+00 4.68813688e-01 4.80098456e-01 8.91418934e-01
7.35006690e-01 -1.29161417e-01 2.23399803e-01 -8.78389359e-01
2.43819058e-01 -4.94286269e-01 -5.69093943e-01 2.06662863e-01
-1.37086689e-01 5.29287606e-02 1.03839982e+00 -7.48745978e-01
-8.83418381e-01 1.50860488e-01 1.47768065e-01 -1.31736565e+00
-4.67422418e-02 -6.48631305e-02 -1.37199670e-01 -3.93794812e-02
5.19034505e-01 5.19640505e-01 2.65436564e-02 -1.71772942e-01
6.79190397e-01 2.00611934e-01 1.08490467e+00 -6.69998109e-01
1.14785194e+00 5.56340754e-01 -1.27908085e-02 -7.00287044e-01
-4.75439399e-01 1.25323888e-02 -3.53230506e-01 -2.86095291e-01
9.88777459e-01 -1.19996321e+00 -4.46336895e-01 5.26430964e-01
-1.22206938e+00 -7.34686732e-01 -3.63935202e-01 3.21597494e-02
-1.08720469e+00 2.69514084e-01 -7.14386761e-01 -3.74720633e-01
-1.43356919e-01 -1.46405649e+00 1.56869781e+00 7.57683516e-02
-1.54581055e-01 -7.77010798e-01 -1.10668495e-01 4.45474684e-01
2.25373104e-01 6.83012426e-01 5.42598963e-01 4.82742600e-02
-1.31717241e+00 1.20724641e-01 -1.46492347e-01 4.83493268e-01
-1.04547832e-02 3.50392342e-01 -6.05066001e-01 -3.15011919e-01
-3.50884020e-01 -4.96239096e-01 9.13135529e-01 3.04668903e-01
1.25798368e+00 -6.04693174e-01 -2.20193923e-01 9.70727742e-01
1.43941391e+00 3.05611312e-01 9.68773842e-01 -1.99279282e-02
8.53691518e-01 3.08104515e-01 5.67727864e-01 2.54714042e-01
5.15422337e-02 1.02730763e+00 4.14635807e-01 -6.11397065e-02
-5.60190558e-01 -7.83170581e-01 6.57191396e-01 4.28908736e-01
-6.98962137e-02 -6.39688134e-01 -5.17647624e-01 4.70697194e-01
-1.83159804e+00 -1.36413670e+00 3.80691767e-01 1.85177171e+00
8.01316440e-01 1.95515598e-03 1.93361834e-01 -4.12167236e-02
6.37444496e-01 5.93908191e-01 -6.63426042e-01 -1.19252086e-01
-1.84320524e-01 3.05965692e-01 4.72347319e-01 4.57368910e-01
-1.17072487e+00 1.23285234e+00 6.28256655e+00 1.04793835e+00
-1.29900169e+00 -8.45249593e-02 7.41894186e-01 -2.46205032e-01
-5.36854565e-01 -1.29469126e-01 -5.30718327e-01 6.99514627e-01
9.41064835e-01 -1.90429434e-01 9.55830872e-01 7.14720070e-01
2.35100776e-01 2.26990640e-01 -1.26891279e+00 1.15288639e+00
2.21745372e-01 -1.89797115e+00 4.65321898e-01 -2.39793235e-03
1.28905141e+00 -3.45431060e-01 3.47135633e-01 1.38692766e-01
3.42249334e-01 -1.31088603e+00 1.25739503e+00 6.00835860e-01
1.45519638e+00 -6.17887735e-01 4.82432246e-02 1.82758435e-04
-1.20381308e+00 -6.45862520e-02 2.92894449e-02 1.28159091e-01
4.34015274e-01 7.48030916e-02 -3.31008017e-01 5.53400874e-01
5.67903042e-01 1.06796920e+00 -3.63338888e-01 6.24517143e-01
-1.50449663e-01 4.66284990e-01 4.51365449e-02 3.86864275e-01
3.96858603e-01 -1.70093343e-01 5.95124185e-01 1.06324732e+00
4.57613498e-01 3.12344611e-01 3.44119579e-01 1.07558823e+00
-4.69984323e-01 -1.50067031e-01 -1.08712292e+00 -3.99821460e-01
3.28719348e-01 9.03535843e-01 -6.80203497e-01 -4.12191689e-01
-3.57122421e-01 1.44053805e+00 3.67545821e-02 5.76041043e-01
-1.06688035e+00 1.36594381e-02 8.20535839e-01 2.70748109e-01
5.80658197e-01 -1.87392652e-01 -2.91949846e-02 -1.29699135e+00
1.39953613e-01 -1.44483662e+00 -8.58328417e-02 -1.01139045e+00
-9.41381335e-01 6.24428391e-01 -7.07624629e-02 -1.40415192e+00
-6.05789304e-01 -2.91386247e-01 -5.08683801e-01 6.88894033e-01
-1.15333533e+00 -1.17421937e+00 -3.02897871e-01 6.50945306e-01
1.13621056e+00 -3.77803296e-01 4.69893038e-01 3.23817223e-01
-2.04001412e-01 4.86985624e-01 7.99540654e-02 1.24338858e-01
6.95013225e-01 -8.60181630e-01 9.08004284e-01 1.23411453e+00
3.31081420e-01 1.32753447e-01 8.04263413e-01 -7.14444101e-01
-1.69882429e+00 -1.42020833e+00 3.41796756e-01 -5.13860881e-01
6.18068218e-01 -3.76609743e-01 -6.51037931e-01 7.24806130e-01
4.11210239e-01 3.30711126e-01 3.51902731e-02 -9.32398260e-01
-4.57796067e-01 -3.17437351e-02 -8.76908362e-01 1.05019414e+00
1.36833489e+00 -5.97802818e-01 -2.02918887e-01 3.15935135e-01
1.19942081e+00 -8.41801763e-01 -7.91522086e-01 2.83647746e-01
6.38183057e-01 -9.47879672e-01 1.33990228e+00 -7.66758204e-01
1.39622498e+00 -2.85689920e-01 -3.44805866e-01 -1.17757952e+00
-2.10455701e-01 -1.16214859e+00 -2.12990180e-01 1.02214503e+00
9.51947421e-02 -1.02140410e-02 7.51731813e-01 3.62907976e-01
-9.82463136e-02 -7.18620360e-01 -5.43037117e-01 -8.13570142e-01
-8.76786709e-02 -2.55783349e-01 4.68173891e-01 6.52197063e-01
-6.11178339e-01 2.24555090e-01 -1.01669431e+00 2.84671281e-02
5.55468738e-01 3.09218615e-01 1.13670313e+00 -4.68700558e-01
-6.30095422e-01 -4.53565180e-01 -3.53099018e-01 -1.24070418e+00
4.46636707e-01 -7.10926712e-01 6.89270124e-02 -1.25119567e+00
-7.91252591e-03 5.20515479e-02 2.75710553e-01 1.71019882e-01
-1.86460853e-01 6.28530145e-01 6.75110996e-01 1.77743703e-01
-5.14269948e-01 5.48488677e-01 1.74352205e+00 -2.82410860e-01
-4.25604396e-02 -2.65831679e-01 -3.98604780e-01 3.77762288e-01
3.40989292e-01 -2.83363581e-01 -5.30025303e-01 -4.67311472e-01
-1.26411319e-01 7.31879711e-01 6.49494886e-01 -9.74233449e-01
2.15147678e-02 -5.03963888e-01 3.91361505e-01 -2.00085372e-01
6.35070622e-01 -6.73751831e-01 7.26255119e-01 4.26843822e-01
-4.40908313e-01 2.16276899e-01 1.06078058e-01 8.56415272e-01
-3.56568575e-01 2.76188046e-01 8.70204628e-01 -2.22721413e-01
-8.04722905e-01 6.70949757e-01 1.48496330e-02 3.95454913e-01
1.13476026e+00 -1.82924703e-01 -2.21275970e-01 -6.60681248e-01
-5.21142423e-01 -4.47773598e-02 9.06852663e-01 8.25296760e-01
7.82908976e-01 -1.54273355e+00 -7.54382551e-01 2.82129258e-01
-1.26982853e-01 -5.46565577e-02 3.43805790e-01 2.52159536e-01
-8.34793329e-01 4.39146399e-01 -4.19379354e-01 -6.99178219e-01
-1.05456197e+00 8.22046101e-01 1.56789288e-01 -1.55980110e-01
-8.64991248e-01 7.17471898e-01 5.67137778e-01 5.52336946e-02
1.80098653e-01 -2.19523743e-01 2.75293350e-01 -2.78446555e-01
4.57124263e-01 1.46170527e-01 -3.27137411e-01 -6.97893262e-01
1.03934221e-01 5.67957401e-01 1.43346652e-01 -3.54054540e-01
1.14042473e+00 8.07062909e-02 2.62567580e-01 2.69268841e-01
1.29206419e+00 -7.20280558e-02 -1.98702395e+00 -1.92012653e-01
-4.71975327e-01 -7.61898875e-01 -3.08362424e-01 -4.66665030e-01
-1.42289817e+00 7.07007825e-01 1.21406019e-01 -1.32733613e-01
1.13581169e+00 -1.76198915e-01 1.09872115e+00 2.45969817e-02
3.72467428e-01 -7.10593462e-01 5.61698198e-01 1.69475973e-01
1.13628709e+00 -1.18430853e+00 -2.65110135e-01 -4.16681558e-01
-8.53789508e-01 1.05378783e+00 4.70359087e-01 -4.79316890e-01
2.53623635e-01 3.86498064e-01 -1.05370589e-01 2.09131241e-01
-1.02273834e+00 1.08963907e-01 4.65682030e-01 4.61987406e-01
9.56714824e-02 -3.99061292e-03 1.43332615e-01 2.28897676e-01
-2.03578070e-01 2.29553416e-01 5.55538714e-01 5.77962995e-01
6.43236116e-02 -1.04692650e+00 -3.51842523e-01 1.35288432e-01
-5.66616356e-01 -1.70488253e-01 -2.70391256e-01 8.03851008e-01
-5.91583960e-02 7.20405161e-01 1.40726134e-01 -4.05655205e-01
2.49772161e-01 -7.70046040e-02 6.49914861e-01 -5.01857281e-01
-4.53545988e-01 2.74526119e-01 -7.65131563e-02 -1.04353213e+00
-4.11801159e-01 -6.36360765e-01 -8.04899216e-01 -4.24885601e-01
3.70322645e-01 -3.03840280e-01 2.47757405e-01 7.04552233e-01
3.45044374e-01 6.50615096e-01 6.32314265e-01 -1.28347313e+00
-3.57633621e-01 -4.54855055e-01 -2.81783104e-01 7.79956639e-01
4.89818990e-01 -4.68719393e-01 -1.73956946e-01 7.00181246e-01] | [10.856088638305664, -0.614187479019165] |
dc0c02b5-c77f-49ba-adb4-cf23f1159668 | classification-of-audio-segments-in-call | 2106.02422 | null | https://arxiv.org/abs/2106.02422v1 | https://arxiv.org/pdf/2106.02422v1.pdf | Classification of Audio Segments in Call Center Recordings using Convolutional Recurrent Neural Networks | Detailed statistical analysis of call center recordings is critical in the customer relationship management point of view. With the recent advances in artificial intelligence, many tasks regarding the calculation of call statistics are now performed automatically. This work proposes a neural network framework where the aim is to correctly identify audio segments and classify them as either customer or agent sections. Accurately identifying these sections gives a fair metric for evaluating agents' performances. We inherited the convolutional recurrent neural network (CRNN) architecture commonly used for such problems as music genre classification. We also tested the same architecture's performance, where the previous class information and the gender information of speakers are also added to the training data labels. We saw that CRNN could generalize the training data and perform well on validation data for this problem with and without the gender information. Moreover, even the training was performed using Turkish speech samples; the trained network was proven to achieve high accuracy for call center recordings in other languages like German and English. | ['Şükrü Ozan'] | 2021-06-04 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [-1.21680081e-01 4.04674672e-02 2.17509434e-01 -5.60091496e-01
-7.51071274e-01 -3.66513640e-01 4.70643997e-01 2.75283188e-01
-6.55733407e-01 4.91344661e-01 8.90070423e-02 -1.60601154e-01
-1.78582683e-01 -6.58448040e-01 -4.13381070e-01 -6.19316041e-01
-1.27654135e-01 9.22013283e-01 -3.31869453e-01 -3.50864410e-01
1.49752766e-01 5.21091700e-01 -1.90920579e+00 5.48874199e-01
3.39917988e-01 1.33473051e+00 -2.12377921e-01 9.14232254e-01
-1.36301160e-01 1.01263821e+00 -1.21272147e+00 -4.72087622e-01
-4.44724970e-02 -2.91672111e-01 -9.08156395e-01 -1.39352798e-01
2.35756159e-01 -5.64381182e-02 1.64719865e-01 6.58318937e-01
7.95251846e-01 2.72832185e-01 5.98675489e-01 -1.10404813e+00
-1.13845140e-01 1.32935321e+00 -9.14697871e-02 1.34344697e-01
2.92989463e-01 -5.81813097e-01 1.25450647e+00 -5.50674021e-01
2.56169051e-01 1.04584777e+00 8.27135801e-01 4.20935661e-01
-9.51440275e-01 -8.15170288e-01 -2.70036627e-02 4.98838335e-01
-1.33811104e+00 -4.82255459e-01 8.53312910e-01 -5.99938333e-01
9.10420358e-01 2.99768299e-01 4.16690558e-01 1.35459960e+00
-3.91924292e-01 6.76667452e-01 7.47855365e-01 -5.50141394e-01
6.86323792e-02 4.03363466e-01 2.89357394e-01 1.65471807e-01
-3.17590058e-01 -1.28042087e-01 -3.25254232e-01 -3.88338566e-02
3.84991884e-01 -1.99919179e-01 -4.29969765e-02 2.07265034e-01
-1.00735641e+00 9.98719394e-01 2.04974696e-01 8.85959446e-01
-4.74702924e-01 3.06515004e-02 9.50361431e-01 4.29611921e-01
4.18653280e-01 6.39245689e-01 -5.96449196e-01 -3.93950939e-01
-8.92769814e-01 7.29226694e-02 1.06447554e+00 5.36553204e-01
2.40247473e-01 4.32028502e-01 -1.01204686e-01 1.22770560e+00
-3.03377897e-01 2.10013941e-01 9.32152212e-01 -8.70582104e-01
3.79829437e-01 5.42349815e-01 -7.61181638e-02 -9.77923393e-01
-8.54480267e-01 -1.05201244e+00 -9.20245230e-01 -4.25490402e-02
7.56722748e-01 -3.05815428e-01 -5.76625526e-01 1.73942244e+00
-1.53179690e-01 7.34377280e-02 2.15391204e-01 7.76904404e-01
1.13162172e+00 5.54673672e-01 -2.76049048e-01 -2.65682608e-01
1.28499782e+00 -6.23403251e-01 -8.40313256e-01 1.85544327e-01
5.98034620e-01 -8.26062679e-01 1.01115572e+00 8.44718933e-01
-8.59737217e-01 -6.64261162e-01 -9.22234297e-01 2.78532147e-01
-5.02076328e-01 6.76325023e-01 5.75474441e-01 9.18894231e-01
-8.10251415e-01 4.75283295e-01 -4.99888182e-01 -2.54455477e-01
8.03502873e-02 6.47179306e-01 -2.34970376e-01 6.13255799e-01
-1.05779111e+00 6.58163309e-01 3.06953728e-01 3.36393356e-01
-6.81582451e-01 -2.67496914e-01 -5.28632641e-01 3.49564224e-01
6.08865768e-02 -2.12230131e-01 1.60826290e+00 -1.36166573e+00
-1.68778419e+00 9.20444965e-01 1.95979163e-01 -8.48175764e-01
4.30941910e-01 -7.26421773e-02 -6.52150929e-01 -1.26923129e-01
-2.69505709e-01 2.11884737e-01 7.32374489e-01 -1.02159989e+00
-7.06455052e-01 -4.05438542e-01 2.88842563e-02 -1.43471971e-01
-5.15384555e-01 1.04029894e-01 -1.11788966e-01 -6.76428616e-01
1.57800034e-01 -9.91437972e-01 2.31469870e-01 -1.14616227e+00
-5.26546419e-01 -4.48325843e-01 3.65863234e-01 -8.34636033e-01
1.21195996e+00 -2.16391563e+00 -6.98218420e-02 5.50689757e-01
-8.44055787e-02 2.06819788e-01 7.36892000e-02 4.11966205e-01
-2.51882941e-01 -2.48505130e-01 2.52636164e-01 -4.81004387e-01
1.81609794e-01 -7.52988160e-02 -3.33469003e-01 1.52063116e-01
-1.10960297e-01 5.04220188e-01 -2.88695067e-01 -2.93156892e-01
1.77269112e-02 4.44035769e-01 -6.51305258e-01 2.38971964e-01
-4.51806039e-02 4.54980046e-01 -1.10617049e-01 4.51313019e-01
1.99786425e-01 2.76289463e-01 1.79488137e-01 -3.09015512e-01
1.15334429e-01 2.98108369e-01 -1.07451093e+00 1.31781876e+00
-7.72552550e-01 8.86587858e-01 2.99217314e-01 -1.34407520e+00
1.34324729e+00 6.31472111e-01 5.58653176e-01 -6.38728797e-01
6.67905509e-01 3.83995205e-01 4.71605569e-01 -4.72997069e-01
6.16237223e-01 -8.59792754e-02 -3.35057765e-01 4.30723310e-01
2.11352587e-01 5.95072210e-02 2.14512542e-01 -3.62628371e-01
7.08663166e-01 -4.01163816e-01 -1.04906000e-01 1.16026007e-01
7.51518071e-01 -2.09931925e-01 5.81533492e-01 5.54574490e-01
-4.61051315e-02 6.58991992e-01 5.29435873e-01 -5.22150934e-01
-7.63328254e-01 -5.97455561e-01 -1.38716683e-01 1.61748683e+00
-6.04993343e-01 -3.40193510e-01 -8.86956990e-01 -3.93037409e-01
-3.21967870e-01 7.82312095e-01 -5.63594043e-01 -1.60090290e-02
-6.60088122e-01 -6.07168972e-01 9.16132510e-01 5.74781716e-01
2.13066146e-01 -1.42995548e+00 -5.13995945e-01 3.53590965e-01
-3.83001179e-01 -1.10419786e+00 6.95061311e-03 5.03661215e-01
-4.20010954e-01 -1.04515111e+00 -5.97640157e-01 -1.06799686e+00
1.45561099e-01 -5.74253023e-01 1.19037354e+00 -1.27476603e-01
-2.08415404e-01 1.76845804e-01 -5.15942991e-01 -7.92437077e-01
-5.46898544e-01 6.32289827e-01 1.72006786e-01 3.93660307e-01
4.08377200e-01 -7.12037742e-01 -2.76433319e-01 4.15151685e-01
-5.53281009e-01 -4.19664055e-01 3.64465445e-01 8.13438416e-01
7.33818635e-02 3.02046031e-01 1.07556033e+00 -9.86096382e-01
8.63449097e-01 -3.97045165e-02 -1.93241715e-01 -3.25601697e-02
-2.38168254e-01 -2.01454118e-01 7.46804655e-01 -4.54064578e-01
-6.67343438e-01 -1.47580400e-01 -7.63115644e-01 -2.82510102e-01
-3.21358681e-01 6.55787468e-01 -6.09621219e-02 4.41891342e-01
5.24193287e-01 -1.12041645e-01 -7.65062049e-02 -7.52097428e-01
-2.94677198e-01 1.20468807e+00 5.85032761e-01 -4.11393642e-01
2.01889232e-01 6.84892759e-02 -3.45926166e-01 -8.40657055e-01
-9.61254120e-01 -5.23805678e-01 -5.74389279e-01 -2.88130850e-01
7.24341691e-01 -5.72081923e-01 -1.48103893e+00 5.34984648e-01
-1.12477577e+00 -1.47536442e-01 -1.90759227e-01 9.09753025e-01
-5.68840384e-01 -2.60260046e-01 -7.08534062e-01 -1.12090814e+00
-4.48836416e-01 -1.05594647e+00 7.57514536e-01 -1.12623312e-02
-4.29613441e-01 -8.18385363e-01 -1.57135576e-02 5.13747454e-01
3.46574157e-01 1.22809172e-01 1.17710948e+00 -1.45340216e+00
2.30110705e-01 -5.64338744e-01 2.16430590e-01 6.10619009e-01
-5.73169030e-02 6.36334419e-02 -1.48559511e+00 -1.02404714e-01
-6.65195053e-03 -3.23160261e-01 8.89769971e-01 5.11779010e-01
1.17546856e+00 -8.33334401e-02 2.22443476e-01 3.09001774e-01
7.22741306e-01 4.39898252e-01 4.85380888e-01 2.53928334e-01
5.66995621e-01 9.43247616e-01 4.75999981e-01 4.90899950e-01
1.17194809e-01 8.63702774e-01 4.41009909e-01 -9.54281911e-02
2.86748767e-01 1.94352996e-02 3.06369573e-01 1.18245161e+00
-3.59325141e-01 -1.63599744e-01 -9.55670595e-01 4.02830273e-01
-1.80573082e+00 -1.21186244e+00 3.41821164e-02 2.14794326e+00
6.07681215e-01 1.29067525e-01 4.69756037e-01 8.94302845e-01
8.36642981e-01 -2.28082940e-01 -1.85640693e-01 -8.57438743e-01
-1.08179204e-01 2.40298018e-01 1.80213884e-01 1.95355400e-01
-1.26920557e+00 6.25512540e-01 6.14770126e+00 9.82921064e-01
-1.34809065e+00 1.92104094e-02 6.98695421e-01 -1.23155795e-01
2.26245090e-01 -7.46928751e-01 -7.22062171e-01 1.24966107e-01
1.13379049e+00 3.59526604e-01 5.24552047e-01 7.01373577e-01
2.34418347e-01 1.60889983e-01 -1.15288842e+00 1.26097476e+00
2.80108750e-01 -9.90098774e-01 -9.33022276e-02 5.79296760e-02
3.49443525e-01 -1.23338185e-01 1.94750145e-01 8.49894047e-01
4.12996449e-02 -1.30411327e+00 7.72037327e-01 3.62129331e-01
5.92652798e-01 -1.31724966e+00 1.38565981e+00 5.19250631e-01
-7.69227326e-01 -4.40712690e-01 -1.18554227e-01 -9.14450213e-02
-1.14210524e-01 3.76435995e-01 -1.46442413e+00 4.02791172e-01
7.93604910e-01 5.31157792e-01 -6.01379335e-01 9.00792062e-01
3.06763291e-01 7.82673776e-01 -2.05415562e-01 4.12744321e-02
1.53346658e-01 -3.16015869e-01 3.97053570e-01 1.10383034e+00
4.95820910e-01 -5.21653116e-01 6.03627525e-02 3.80045563e-01
1.33562908e-02 5.71009874e-01 -1.78355575e-01 -2.74993151e-01
7.76442736e-02 1.27834606e+00 -9.35841382e-01 -1.10036336e-01
1.29616961e-01 5.07383108e-01 1.65299967e-01 2.29124084e-01
-8.08099449e-01 -5.60935736e-01 4.21359390e-01 1.22817978e-02
3.30155700e-01 5.64279109e-02 -1.85985044e-01 -7.08641589e-01
-1.68652177e-01 -8.88313293e-01 4.60384399e-01 -6.24760330e-01
-1.08848131e+00 1.11359656e+00 -3.92208487e-01 -1.36438370e+00
-8.32419932e-01 -7.21727371e-01 -5.40674508e-01 5.04660010e-01
-1.01768768e+00 -1.18189549e+00 -2.66038924e-01 4.83307034e-01
5.34663916e-01 -8.69916737e-01 1.31730545e+00 6.95421755e-01
-6.23035908e-01 8.73314559e-01 1.59045100e-01 7.26417542e-01
7.83111572e-01 -1.22735393e+00 -3.12496156e-01 9.02323052e-02
5.68708181e-01 4.09413189e-01 7.68065751e-01 -3.56526300e-02
-7.48247921e-01 -8.68353665e-01 9.58243370e-01 -1.65555254e-01
5.16998351e-01 -5.16727328e-01 -5.52598178e-01 6.12888157e-01
1.58287048e-01 -5.64130247e-01 1.00259924e+00 6.85659230e-01
-2.58395225e-01 -4.05042082e-01 -8.44277322e-01 1.41783774e-01
4.93606716e-01 -6.75553083e-01 -4.27850753e-01 1.03305906e-01
4.55747753e-01 -3.80959839e-01 -7.69547760e-01 3.81192923e-01
7.43429005e-01 -1.18183243e+00 7.61076152e-01 -5.88711917e-01
2.14748800e-01 -6.58167303e-02 -3.46273661e-01 -1.34005070e+00
1.82311729e-01 -3.56346041e-01 3.60319853e-01 1.65046394e+00
6.11938417e-01 -5.01976967e-01 8.56641054e-01 1.75578501e-02
-1.97458938e-01 -4.78647470e-01 -8.40360582e-01 -5.36020458e-01
-1.13108225e-01 -8.11160743e-01 7.71557391e-01 8.13875735e-01
-1.27650425e-01 6.17208421e-01 -2.87739247e-01 -2.83460140e-01
-6.49609417e-02 3.29773217e-01 7.96957970e-01 -1.73546827e+00
-3.34971637e-01 -6.63937747e-01 -6.67190611e-01 -2.78215647e-01
6.36059403e-01 -9.14998531e-01 -1.81194380e-01 -1.20083511e+00
-2.28733778e-01 -4.62790728e-01 -5.15748978e-01 1.69618011e-01
4.96551514e-01 5.76700747e-01 2.83080861e-02 -3.53076458e-02
-3.75268131e-01 4.12439495e-01 7.56511986e-01 -2.61175871e-01
-4.36354429e-01 9.29126680e-01 -4.32257533e-01 7.36406505e-01
8.97863209e-01 -2.74987966e-01 -2.23019674e-01 -7.90124014e-02
3.59968215e-01 -6.98454306e-03 1.73762992e-01 -1.15500581e+00
5.66906035e-02 4.54459697e-01 2.85326481e-01 -8.97881687e-01
7.60687053e-01 -1.06521499e+00 1.12991251e-01 2.24021792e-01
-7.69756973e-01 1.01362914e-02 -2.42807865e-02 9.09867361e-02
-7.31944382e-01 -3.12752873e-01 4.81114149e-01 2.01752022e-01
-2.25347653e-01 -1.88699245e-01 -3.62954706e-01 -1.33198753e-01
6.48410976e-01 -2.34974563e-01 4.99275737e-02 -8.71855915e-01
-1.23137331e+00 -1.43998519e-01 -8.66006240e-02 4.08397764e-01
1.79245472e-01 -1.12719190e+00 -7.97423840e-01 2.61040151e-01
1.13973163e-01 -3.68925810e-01 1.87003851e-01 1.00243938e+00
-5.70926607e-01 5.49551129e-01 -1.54839084e-01 -7.15910852e-01
-1.47049367e+00 2.84163713e-01 4.85791951e-01 -2.11115196e-01
-1.52398691e-01 7.28212535e-01 -1.05668150e-01 -7.38372982e-01
8.46169472e-01 -5.32931328e-01 -1.04688919e+00 8.80587578e-01
3.48384738e-01 4.63070869e-01 6.80665672e-01 -8.58324468e-01
-2.70329863e-01 4.84965712e-01 5.80718601e-03 -2.47648001e-01
1.58425987e+00 1.26141950e-01 2.20510755e-02 9.96717393e-01
1.35366023e+00 2.50678778e-01 -2.86027282e-01 -3.03464383e-02
2.07872763e-01 2.07126066e-01 8.77561346e-02 -7.46414483e-01
-1.21336114e+00 9.98012841e-01 5.14220655e-01 7.14353740e-01
1.00904357e+00 -1.82883322e-01 5.86649060e-01 5.10910451e-01
2.76889414e-01 -1.44764745e+00 -2.40327716e-01 8.30369174e-01
8.50196540e-01 -1.11806309e+00 -4.41554219e-01 -8.08820203e-02
-7.77173817e-01 1.25562119e+00 2.81684905e-01 -4.50335182e-02
6.45650566e-01 9.99662057e-02 4.06530708e-01 -1.06474698e-01
-6.02566421e-01 -3.34257841e-01 2.94112593e-01 5.26120842e-01
7.98640549e-01 2.03165531e-01 4.22688061e-03 1.25408745e+00
-1.05247915e+00 -3.62316757e-01 4.38767523e-01 3.09357733e-01
-1.25968471e-01 -9.12531197e-01 -4.89254713e-01 5.21600544e-01
-9.29969549e-01 1.66186109e-01 -5.89045286e-01 6.94225490e-01
1.72687575e-01 1.17623889e+00 4.36163515e-01 -6.10326707e-01
4.94919062e-01 3.13717306e-01 2.12837294e-01 -5.48658848e-01
-1.35628474e+00 2.61094421e-01 4.20729488e-01 -1.29145443e-01
-5.55034459e-01 -5.60779512e-01 -1.21886575e+00 -1.73213407e-01
-3.27293903e-01 7.74551213e-01 1.03961551e+00 7.48305500e-01
7.89394137e-03 1.02720273e+00 9.16826069e-01 -9.79798496e-01
-3.73684734e-01 -1.43969452e+00 -9.12754834e-01 4.50114578e-01
3.67529541e-01 -5.29201448e-01 -3.92296165e-01 7.48939440e-02] | [15.784043312072754, 5.242916584014893] |
17e2c512-5261-4d25-97e2-95c2deba8bb7 | re-ranking-via-metric-fusion-for-object | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Bai_Re-Ranking_via_Metric_Fusion_for_Object_Retrieval_and_Person_Re-Identification_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Bai_Re-Ranking_via_Metric_Fusion_for_Object_Retrieval_and_Person_Re-Identification_CVPR_2019_paper.pdf | Re-Ranking via Metric Fusion for Object Retrieval and Person Re-Identification | This work studies the unsupervised re-ranking procedure for object retrieval and person re-identification with a specific concentration on an ensemble of multiple metrics (or similarities). While the re-ranking step is involved by running a diffusion process on the underlying data manifolds, the fusion step can leverage the complementarity of multiple metrics. We give a comprehensive summary of existing fusion with diffusion strategies, and systematically analyze their pros and cons. Based on the analysis, we propose a unified yet robust algorithm which inherits their advantages and discards their disadvantages. Hence, we call it Unified Ensemble Diffusion (UED). More interestingly, we derive that the inherited properties indeed stem from a theoretical framework, where the relevant works can be elegantly summarized as special cases of UED by imposing additional constraints on the objective function and varying the solver of similarity propagation. Extensive experiments with 3D shape retrieval, image retrieval and person re-identification demonstrate that the proposed framework outperforms the state of the arts, and at the same time suggest that re-ranking via metric fusion is a promising tool to further improve the retrieval performance of existing algorithms.
| [' Longin Jan Latecki', ' Philip H.S. Torr', ' Peng Tang', 'Song Bai'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['3d-shape-retrieval'] | ['computer-vision'] | [ 1.61938474e-01 -3.65256041e-01 -3.16577516e-02 -4.49165069e-02
-6.65913999e-01 -7.28587270e-01 9.88163710e-01 2.71689653e-01
-4.16961581e-01 2.40656957e-01 1.07617296e-01 2.08479792e-01
-8.67080212e-01 -5.53596377e-01 -4.32724357e-02 -1.01645529e+00
-4.15867493e-02 4.93030876e-01 9.38612502e-03 -2.84601003e-01
3.85742545e-01 8.01756561e-01 -1.77725065e+00 -3.92441720e-01
1.07215810e+00 9.45501626e-01 -2.04232931e-01 3.98628831e-01
1.42477170e-01 3.05548221e-01 -4.15242910e-01 -7.96331167e-01
4.16582078e-01 -2.81893849e-01 -7.10545480e-01 2.17605799e-01
3.89977008e-01 -1.45811543e-01 -4.67896670e-01 1.20907581e+00
7.94380963e-01 2.04524457e-01 9.37142432e-01 -1.17080772e+00
-9.55832899e-01 1.28896803e-01 -4.18817282e-01 1.60770848e-01
5.34763575e-01 -4.07079846e-01 1.19180202e+00 -1.09348595e+00
5.39692342e-01 1.06262815e+00 5.97088218e-01 3.40443850e-01
-1.06088459e+00 -2.79525846e-01 1.58177540e-01 2.31188104e-01
-1.64136267e+00 -3.94182324e-01 1.00659680e+00 -6.57710671e-01
3.48795027e-01 4.98371243e-01 5.75561941e-01 7.24453032e-01
-2.77759790e-01 8.62780929e-01 1.00705564e+00 -4.34015989e-01
6.98314011e-02 6.50673732e-02 5.44555843e-01 7.76431799e-01
5.63978910e-01 1.70949638e-01 -5.00406444e-01 -4.00014877e-01
4.23290193e-01 5.23486286e-02 -2.98125625e-01 -9.26952600e-01
-1.36353290e+00 6.12617433e-01 2.75678992e-01 4.16313261e-01
-1.90704718e-01 -2.55206674e-01 1.51477128e-01 4.47818846e-01
6.86475933e-01 2.59431630e-01 2.48202384e-01 2.57160097e-01
-1.00240850e+00 3.77924234e-01 6.04183197e-01 6.84519231e-01
7.54428804e-01 -2.41892800e-01 -3.31924289e-01 9.40049589e-01
3.59009206e-01 6.78375959e-01 3.71454418e-01 -7.50433326e-01
1.26812607e-01 6.92813516e-01 2.65230745e-01 -1.35664117e+00
-2.73581833e-01 -5.70559740e-01 -8.45921516e-01 5.74592091e-02
4.27244395e-01 1.88115492e-01 -4.95842606e-01 1.61644208e+00
3.85990530e-01 3.31773072e-01 1.06582753e-02 9.72690344e-01
6.86475754e-01 9.48809311e-02 -1.80313185e-01 -2.41972536e-01
1.11628449e+00 -8.32889736e-01 -6.75501049e-01 5.07241666e-01
4.23063099e-01 -7.45886743e-01 3.95858616e-01 2.45342806e-01
-1.19762576e+00 -5.41095257e-01 -1.19717038e+00 6.47096559e-02
-4.91573423e-01 2.00509682e-01 6.48914874e-01 9.86135960e-01
-1.32844007e+00 8.61385524e-01 -5.79292417e-01 -5.41157603e-01
-2.12104502e-03 4.62328464e-01 -2.80533701e-01 4.96050790e-02
-9.54349995e-01 8.67829859e-01 -8.65289345e-02 3.23737741e-01
-4.05893445e-01 -5.04505098e-01 -4.83337641e-01 -2.19278932e-01
2.70416200e-01 -1.00541139e+00 8.20898950e-01 -5.47060430e-01
-1.55578446e+00 1.02890599e+00 -6.05289303e-02 -2.66144097e-01
8.17487180e-01 -3.03580135e-01 -3.21136445e-01 3.82773072e-01
-8.52797702e-02 1.99087456e-01 9.21977043e-01 -1.25907338e+00
-5.04707634e-01 -6.52743220e-01 2.54291177e-01 3.65465939e-01
-8.07847857e-01 8.65912959e-02 -4.86701310e-01 -7.91801512e-01
1.59252360e-01 -1.09261894e+00 9.08742771e-02 -1.32487506e-01
-2.32227325e-01 -5.24870932e-01 5.27018487e-01 -3.96851271e-01
1.44634056e+00 -1.91745758e+00 9.49231923e-01 5.70592463e-01
5.00133693e-01 1.21483654e-01 -1.11410722e-01 5.10507822e-01
3.46981697e-02 4.22401316e-02 -2.25936607e-01 -6.74849689e-01
1.92224056e-01 -1.47372440e-01 -1.44105971e-01 8.84377003e-01
-4.92967404e-02 9.49205637e-01 -9.55933750e-01 -4.68806744e-01
3.06896508e-01 7.63978422e-01 -2.18216613e-01 -6.88880309e-02
4.98977810e-01 5.10938287e-01 -6.41235173e-01 7.39926934e-01
6.69063389e-01 -2.00344995e-01 2.99993623e-02 -4.73540902e-01
-2.13184819e-01 -1.43957183e-01 -1.52997661e+00 1.76773083e+00
1.45941172e-02 3.08353454e-01 1.58629324e-02 -1.23160803e+00
8.67708385e-01 1.04283109e-01 1.06626225e+00 -3.41378480e-01
-5.35450242e-02 5.05933523e-01 -3.47312748e-01 -2.97255784e-01
7.18710721e-01 1.18015654e-01 -6.16534567e-03 5.56198895e-01
9.05106142e-02 3.31016243e-01 3.29732209e-01 2.39940390e-01
7.26682365e-01 7.91960210e-02 9.40356478e-02 -4.38117623e-01
9.91720617e-01 -3.64428848e-01 1.00357011e-01 8.55189562e-01
-1.72887325e-01 5.61384976e-01 4.65560257e-02 -5.18195182e-02
-9.25686717e-01 -1.11122537e+00 -2.76201665e-01 8.02396894e-01
6.08555675e-01 -5.20641267e-01 -6.39678955e-01 -6.83348417e-01
2.24554524e-01 1.25407711e-01 -6.38598979e-01 -8.88931677e-02
-3.64400417e-01 -1.09087420e+00 4.80206817e-01 2.89530456e-01
5.52532434e-01 -3.52338016e-01 -1.80985019e-01 -1.66788563e-01
-1.01641193e-01 -8.62058222e-01 -5.80579817e-01 -5.64164996e-01
-7.83070266e-01 -1.14709592e+00 -1.47339559e+00 -6.42720520e-01
4.20419127e-01 7.62394488e-01 8.27651203e-01 2.97837317e-01
-3.35963406e-02 1.38501227e+00 -4.50454772e-01 4.50063087e-02
-1.95395481e-02 1.21744506e-01 4.25338894e-01 6.15943134e-01
3.13840210e-01 -6.67273045e-01 -7.78403103e-01 4.79238808e-01
-8.16584945e-01 -3.28918844e-01 3.46855193e-01 6.92851901e-01
4.97535616e-01 1.25818960e-02 6.52415276e-01 -3.90921831e-01
7.45809317e-01 -3.39935303e-01 -5.13105690e-01 6.28476501e-01
-9.04668868e-01 1.46204263e-01 -1.13034762e-01 -3.41111064e-01
-9.04784024e-01 -1.88649103e-01 3.04121435e-01 -2.48097554e-01
2.79336840e-01 3.24200809e-01 -4.25080173e-02 -4.99164224e-01
1.72420248e-01 2.76216775e-01 -6.51379824e-02 -7.92256355e-01
5.34572661e-01 6.17624521e-01 2.76251197e-01 -6.50560737e-01
1.07512069e+00 9.55465615e-01 1.92818895e-01 -8.51636767e-01
-7.43612707e-01 -6.39641941e-01 -9.50757742e-01 -4.99720842e-01
6.83383048e-01 -6.17193937e-01 -8.39721084e-01 6.16583765e-01
-9.56278801e-01 2.59604722e-01 -2.94259787e-01 4.47248578e-01
-4.29423481e-01 8.53821635e-01 -4.68638629e-01 -9.86600161e-01
-2.03237563e-01 -1.02796733e+00 1.17544639e+00 1.66815922e-01
9.09592137e-02 -1.23484445e+00 2.14011788e-01 3.10763121e-01
3.29501837e-01 1.09965071e-01 6.70643330e-01 -7.37767339e-01
-6.00396574e-01 -3.00752223e-01 -2.56650746e-01 3.04514945e-01
2.44237185e-01 -1.59843042e-01 -7.85618544e-01 -4.46414173e-01
8.81730171e-04 2.38664448e-01 1.01393282e+00 1.07886165e-01
7.06212521e-01 1.71544582e-01 -4.22642022e-01 4.39847380e-01
1.30435503e+00 -1.39861912e-01 3.22087407e-01 3.70235324e-01
6.33510947e-01 8.31196070e-01 3.74893695e-01 3.21117908e-01
5.49013078e-01 1.03528202e+00 9.51542780e-02 1.01554111e-01
-3.28290373e-01 1.10551357e-01 2.76342213e-01 1.14252639e+00
-7.19879508e-01 -6.86050355e-02 -6.18154645e-01 3.85343045e-01
-2.01814508e+00 -1.03734982e+00 -2.91275866e-02 2.59666419e+00
2.47644156e-01 -3.84232968e-01 4.63523775e-01 4.27271903e-01
9.33223784e-01 2.07786411e-01 -3.65130424e-01 2.60882109e-01
-5.26323974e-01 -2.41861697e-02 3.10542166e-01 7.44471073e-01
-1.18639684e+00 4.89290655e-01 6.26494122e+00 7.53118515e-01
-7.70694911e-01 2.07619682e-01 2.47812010e-02 -5.47543429e-02
-3.96796077e-01 -1.94614768e-01 -7.72119999e-01 1.95681304e-01
3.96940529e-01 -5.11671066e-01 4.56932724e-01 2.84658730e-01
-2.18650177e-01 1.88684836e-01 -1.02623057e+00 1.25006270e+00
5.15547037e-01 -9.82371449e-01 1.94419831e-01 2.52709895e-01
8.00099969e-01 -2.58010268e-01 4.20953542e-01 -1.15054801e-01
-1.20067120e-01 -5.21149874e-01 6.87667966e-01 1.09000921e+00
4.18527573e-01 -5.85852385e-01 6.49912000e-01 -4.40038778e-02
-1.29563606e+00 3.57804075e-02 -1.12650149e-01 2.53995568e-01
1.51602820e-01 5.96486211e-01 1.04489915e-01 1.26728761e+00
3.83730948e-01 1.12455416e+00 -8.45562339e-01 1.31315672e+00
2.24722371e-01 -2.09434703e-02 -2.73503453e-01 -7.92725664e-03
-5.31207882e-02 -6.98077023e-01 1.01576257e+00 1.21448696e+00
6.58454061e-01 -1.15363754e-01 3.70893977e-03 5.39282084e-01
1.70208752e-01 3.62295717e-01 -4.80525404e-01 1.11742906e-01
2.41129562e-01 1.24472046e+00 -6.20133042e-01 -3.42163652e-01
-4.32871193e-01 1.18161273e+00 1.78858161e-01 5.40033221e-01
-6.47337615e-01 -2.35020980e-01 6.41881108e-01 -1.68947250e-01
2.24217787e-01 -4.14887846e-01 -1.62141949e-01 -1.37696958e+00
2.74839520e-01 -4.64559376e-01 5.03531456e-01 -4.89342391e-01
-1.52554488e+00 5.23832738e-01 3.42989177e-01 -1.47958410e+00
-7.53429085e-02 -7.13317156e-01 -1.13731921e-01 5.10917723e-01
-1.79338217e+00 -1.07191098e+00 -3.08838248e-01 6.58971846e-01
2.10790291e-01 -2.66055763e-01 7.36952245e-01 5.68317056e-01
-7.07006395e-01 6.16490304e-01 4.75878417e-01 -2.40276635e-01
7.13049352e-01 -1.34293532e+00 -2.77946204e-01 7.02906847e-01
1.97374433e-01 7.76883841e-01 5.34460962e-01 -5.23865223e-01
-1.45053744e+00 -6.44028425e-01 1.02302754e+00 -6.76706195e-01
8.30302656e-01 -1.17468551e-01 -7.23684728e-01 3.05176049e-01
1.25345901e-01 -3.40542644e-01 8.01213622e-01 1.93377227e-01
-3.38554144e-01 -2.55673409e-01 -8.79366100e-01 5.72568297e-01
1.43361974e+00 -5.83030343e-01 -6.59115195e-01 4.21527505e-01
2.26605833e-01 -9.66609339e-04 -1.15402985e+00 5.87806225e-01
7.20076621e-01 -1.20729256e+00 1.41996825e+00 -4.17071879e-01
-1.96648061e-01 -4.29770768e-01 -3.41757119e-01 -9.60614860e-01
-4.06172395e-01 -8.08946729e-01 -4.71954793e-01 1.46902180e+00
2.26750528e-03 -7.95137703e-01 4.57873017e-01 5.83110809e-01
1.27868891e-01 -6.49735451e-01 -9.21947658e-01 -1.01332104e+00
1.52235806e-01 -2.47464940e-01 5.74040711e-01 7.98629463e-01
3.35227661e-02 1.41215324e-01 -5.13848126e-01 2.12106690e-01
1.07726061e+00 1.70019835e-01 5.18341720e-01 -1.69396949e+00
-1.69350833e-01 -7.68420875e-01 -6.37715042e-01 -1.31697595e+00
1.72529638e-01 -1.12734079e+00 -5.45212746e-01 -1.26954639e+00
3.33330065e-01 -5.01330614e-01 -7.16605484e-01 -7.72729144e-02
-1.68192282e-01 4.84230340e-01 4.60462272e-01 8.35018218e-01
-8.68072927e-01 7.38941669e-01 1.06519568e+00 -2.54784375e-01
-2.19209582e-01 2.31905282e-01 -6.58908010e-01 6.49106264e-01
4.62142110e-01 -9.78804678e-02 -2.18408719e-01 -4.43780750e-01
3.77435267e-01 -3.75027001e-01 6.81350231e-01 -9.23418820e-01
5.45121491e-01 3.11705619e-01 -1.80279985e-02 -5.68688273e-01
5.21008849e-01 -7.27109909e-01 -5.52487932e-03 2.48085096e-01
-2.96489596e-01 9.35050473e-02 -2.87375271e-01 8.84878516e-01
-2.00862005e-01 -3.36676002e-01 4.25584674e-01 2.35477671e-01
-4.95232433e-01 3.33655030e-01 -9.80801582e-02 -4.36662227e-01
8.87231052e-01 -3.60439360e-01 -9.08548757e-02 -3.18303823e-01
-9.62580621e-01 9.82385278e-02 4.24170941e-01 5.85395992e-01
5.46826303e-01 -1.71587336e+00 -7.69548118e-01 2.78508924e-02
1.55123040e-01 -6.74713552e-01 3.10482502e-01 1.29926145e+00
9.41285491e-02 5.31821430e-01 1.67973951e-01 -6.80142701e-01
-1.28615320e+00 6.44052386e-01 3.35124552e-01 -3.49293649e-01
-5.52854955e-01 5.70101500e-01 6.38566837e-02 -1.77884236e-01
3.08410257e-01 1.90578356e-01 -5.18684268e-01 3.84072691e-01
4.48343903e-01 8.69041383e-01 2.56099701e-02 -1.03022003e+00
-5.00003934e-01 1.23166680e+00 4.70350422e-02 -2.18506634e-01
1.20115101e+00 -6.10521555e-01 -1.63797259e-01 4.36965585e-01
1.14276361e+00 3.10314327e-01 -8.95452976e-01 -4.72761571e-01
1.90269053e-01 -2.83181936e-01 -1.37825506e-02 -3.71635258e-01
-1.04302955e+00 5.78691781e-01 7.71747351e-01 5.15726566e-01
1.25234175e+00 6.36740178e-02 5.33823967e-01 3.75783861e-01
4.14776325e-01 -9.79945242e-01 1.48927540e-01 1.60909668e-01
8.99576664e-01 -9.14271653e-01 1.76595211e-01 -5.45782864e-01
-2.69113630e-01 1.06936646e+00 -2.76361462e-02 -4.94700581e-01
8.43168676e-01 -3.81544411e-01 -3.55903268e-01 -2.21893191e-01
-9.03901607e-02 -6.06696725e-01 7.80030787e-01 5.53387821e-01
2.90763140e-01 -1.57811448e-01 -7.27132440e-01 4.43679750e-01
1.01106994e-01 -1.56471401e-01 5.31870546e-03 5.51464796e-01
-1.73053563e-01 -1.41316164e+00 -4.44555581e-01 1.44944236e-01
-2.28224710e-01 6.39880374e-02 -5.66006184e-01 7.61631429e-01
1.51363865e-03 1.06797433e+00 -2.65564561e-01 -4.78300989e-01
3.95980209e-01 4.94301459e-03 7.77519763e-01 -1.68842867e-01
-4.22898829e-01 7.10433647e-02 -1.48533776e-01 -4.19907510e-01
-1.14014518e+00 -9.89663661e-01 -5.11612117e-01 -1.61428392e-01
-5.29961586e-01 1.99437931e-01 4.72314835e-01 9.55776870e-01
4.29114997e-01 1.80895120e-01 7.06489086e-01 -9.17771816e-01
-7.15171695e-01 -7.58773327e-01 -6.61980510e-01 5.81696153e-01
3.50825250e-01 -1.15202987e+00 -5.46441913e-01 -1.31116912e-01] | [11.019030570983887, 0.939283549785614] |
460075e8-d0a7-4780-85b6-d66623b3653b | multi-person-3d-pose-estimation-in-crowded | 2007.10986 | null | https://arxiv.org/abs/2007.10986v1 | https://arxiv.org/pdf/2007.10986v1.pdf | Multi-person 3D Pose Estimation in Crowded Scenes Based on Multi-View Geometry | Epipolar constraints are at the core of feature matching and depth estimation in current multi-person multi-camera 3D human pose estimation methods. Despite the satisfactory performance of this formulation in sparser crowd scenes, its effectiveness is frequently challenged under denser crowd circumstances mainly due to two sources of ambiguity. The first is the mismatch of human joints resulting from the simple cues provided by the Euclidean distances between joints and epipolar lines. The second is the lack of robustness from the naive formulation of the problem as a least squares minimization. In this paper, we depart from the multi-person 3D pose estimation formulation, and instead reformulate it as crowd pose estimation. Our method consists of two key components: a graph model for fast cross-view matching, and a maximum a posteriori (MAP) estimator for the reconstruction of the 3D human poses. We demonstrate the effectiveness and superiority of our proposed method on four benchmark datasets. | ['He Chen', 'Gim Hee Lee', 'Pengfei Li', 'Pengfei Guo', 'Gregory Chirikjian'] | 2020-07-21 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/672_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123480545.pdf | eccv-2020-8 | ['3d-multi-person-pose-estimation'] | ['computer-vision'] | [ 9.40634757e-02 -1.63697109e-01 3.49639654e-01 -1.13538444e-01
-5.43262899e-01 -2.08822712e-01 6.46146119e-01 1.60265751e-02
-5.60576856e-01 5.46732724e-01 4.99181956e-01 4.75727588e-01
-7.57426098e-02 -4.19872224e-01 -5.36256194e-01 -4.80880916e-01
3.52018833e-01 6.11179531e-01 3.07168424e-01 -3.14480811e-01
8.60467926e-02 5.53702950e-01 -1.46099257e+00 -3.59415263e-01
5.69033921e-01 6.53618872e-01 -1.43793181e-01 5.61028063e-01
4.36705023e-01 2.64631212e-01 -5.30469179e-01 -5.63697994e-01
5.24726391e-01 -1.91090256e-01 -3.11405122e-01 6.71625853e-01
7.00973451e-01 -4.02085841e-01 -5.88596106e-01 8.63576055e-01
8.42640698e-01 3.62297803e-01 4.97504115e-01 -1.41563654e+00
2.13747755e-01 -3.71242881e-01 -8.36476088e-01 -8.61919075e-02
1.13782346e+00 2.04035547e-02 8.90897989e-01 -1.14668858e+00
7.38979459e-01 1.41486669e+00 7.35355675e-01 2.95508832e-01
-8.77815723e-01 -1.82487309e-01 -3.24851647e-02 -8.69210053e-04
-1.61210787e+00 -6.43620670e-01 7.05562770e-01 -7.89270103e-01
7.41014659e-01 7.30532706e-02 9.24402833e-01 9.10371542e-01
2.31340751e-01 4.94807065e-01 8.77567828e-01 -4.19175714e-01
-1.29199093e-02 -1.85288060e-02 -1.88046426e-01 7.94408917e-01
5.49801946e-01 -2.92487741e-02 -8.28358591e-01 -4.98456210e-01
9.93983567e-01 2.29669869e-01 -3.44877124e-01 -9.37497497e-01
-1.24374402e+00 6.66818380e-01 2.93355286e-01 -1.51967585e-01
-4.29232925e-01 -7.25974375e-03 1.36926070e-01 -2.13711053e-01
1.71776786e-01 -2.83370074e-03 1.08127788e-01 1.47871198e-02
-7.51306951e-01 8.05338502e-01 5.68532526e-01 1.13465941e+00
6.60172522e-01 -2.47961625e-01 8.75136107e-02 5.68170488e-01
6.09073102e-01 6.80649400e-01 4.14501801e-02 -8.87631774e-01
6.57945931e-01 6.04689181e-01 3.85710508e-01 -1.45960963e+00
-4.22288805e-01 -3.47995609e-01 -7.62207091e-01 -8.18550214e-02
7.71821856e-01 -3.10055375e-01 -2.99635589e-01 1.50232399e+00
7.03754187e-01 -1.58061296e-01 -3.73263150e-01 1.27184951e+00
6.05432808e-01 1.13668844e-01 -4.00420904e-01 -1.68517791e-02
1.23598921e+00 -9.06920373e-01 -5.65792620e-01 -6.64865434e-01
2.14921758e-01 -8.21994007e-01 3.15757722e-01 7.53383040e-02
-1.11207569e+00 -4.75968838e-01 -9.66376483e-01 -9.81039107e-02
4.68320698e-02 -5.11540659e-02 2.65179068e-01 5.09999335e-01
-6.42417192e-01 2.70530105e-01 -5.57399988e-01 -6.13626003e-01
-5.53675592e-02 3.65783453e-01 -6.52434766e-01 -3.22512060e-01
-8.47070813e-01 9.71025229e-01 -8.92016515e-02 3.91855806e-01
-2.77510107e-01 -3.51845562e-01 -1.11335433e+00 -3.17850113e-01
6.22355521e-01 -1.16712296e+00 9.29672658e-01 -3.58817190e-01
-1.35442090e+00 9.75675583e-01 -3.91164005e-01 -5.31787910e-02
1.07143962e+00 -6.30181372e-01 1.30816549e-01 2.05124453e-01
1.77490354e-01 2.82948524e-01 6.90888047e-01 -1.26718426e+00
-4.52980757e-01 -8.95401895e-01 -4.40681353e-02 6.14999115e-01
1.59615815e-01 -1.07121110e-01 -9.38769221e-01 -5.26866138e-01
3.89184654e-01 -1.25133276e+00 -3.87670934e-01 2.60084957e-01
-3.95964950e-01 7.23913684e-02 3.38920653e-01 -7.32835770e-01
8.88654292e-01 -1.85440969e+00 5.83495021e-01 2.91528344e-01
3.75738531e-01 -7.43093863e-02 2.54128814e-01 5.31076610e-01
3.30064565e-01 -5.93715847e-01 6.34195805e-02 -6.68837011e-01
5.12768663e-02 5.77361928e-03 3.39641064e-01 1.09620225e+00
7.45584117e-03 7.16328740e-01 -7.83098221e-01 -6.18262172e-01
2.83740997e-01 7.19564378e-01 -4.42756683e-01 3.13171059e-01
3.39517862e-01 5.45082510e-01 -4.62678969e-01 5.50487757e-01
6.24340057e-01 1.22238239e-02 9.67859104e-02 -3.50055516e-01
6.14459580e-03 -1.84078112e-01 -1.80895138e+00 1.93428528e+00
1.97388723e-01 1.88418582e-01 2.15253159e-01 -6.30877912e-01
7.54636168e-01 3.89915168e-01 6.76871061e-01 -2.80129552e-01
2.35773310e-01 1.65161178e-01 -3.14926118e-01 -2.70560354e-01
5.58906794e-01 -1.24996662e-01 -1.25071928e-01 1.75014883e-01
-2.23060161e-01 -9.83905047e-02 5.36012352e-02 8.09605643e-02
8.33842814e-01 3.20778668e-01 5.92218459e-01 -8.28547552e-02
5.88076770e-01 -2.68389493e-01 7.72483230e-01 3.90849024e-01
-3.29278946e-01 8.32872272e-01 2.94073224e-01 -4.58959699e-01
-1.08212852e+00 -1.03321171e+00 1.04349844e-01 3.62283707e-01
2.46374786e-01 -5.24605513e-01 -5.14752746e-01 -3.77427638e-01
2.48161227e-01 -2.17771187e-01 -3.38385403e-01 1.81431949e-01
-8.06770027e-01 -4.84276325e-01 2.40528286e-01 3.46440822e-01
4.13081408e-01 -1.74584880e-01 -8.92970979e-01 1.23513322e-02
-4.95292455e-01 -1.64276433e+00 -6.97229147e-01 -4.68356192e-01
-6.69470906e-01 -1.24640059e+00 -1.12177336e+00 -4.78054971e-01
7.98345625e-01 5.88558733e-01 8.97773206e-01 1.12374857e-01
-3.97309929e-01 7.12879658e-01 -1.40047804e-01 -1.94028944e-01
1.86491042e-01 -3.48246753e-01 4.42127287e-01 2.23823339e-01
2.88923860e-01 -3.55652571e-01 -8.08871686e-01 6.25579596e-01
-4.14480269e-01 4.22539376e-02 3.74275923e-01 6.35240614e-01
5.21082044e-01 -9.96361300e-02 3.43741989e-03 -3.85992706e-01
2.76443720e-01 -1.66215062e-01 -4.54600513e-01 9.92021188e-02
-1.57950744e-01 -2.58818805e-01 1.37540298e-02 -1.22491054e-01
-8.26984704e-01 6.51402593e-01 1.26685813e-01 -2.74817109e-01
-1.32390559e-01 1.18955046e-01 -3.87376487e-01 -2.90164649e-01
3.60204011e-01 5.63075766e-02 1.26204103e-01 -3.67963046e-01
8.47454146e-02 3.99914622e-01 5.87569714e-01 -6.00814760e-01
1.00805533e+00 7.23395169e-01 3.95691365e-01 -1.24597311e+00
-5.49942732e-01 -1.02634954e+00 -1.04612136e+00 -5.57327211e-01
9.88609195e-01 -1.23403895e+00 -8.95216763e-01 6.30981863e-01
-1.25440359e+00 3.01690638e-01 3.28017622e-02 6.61276162e-01
-5.88622868e-01 1.00903976e+00 -4.61603850e-01 -9.99456704e-01
-8.44563991e-02 -1.19970322e+00 1.25854015e+00 1.26249753e-02
-3.34653795e-01 -8.58680189e-01 1.79014623e-01 6.74835801e-01
-1.22367479e-01 7.85598338e-01 3.14104676e-01 -3.64972912e-02
-5.28564274e-01 -6.81568563e-01 5.06013706e-02 -1.45005792e-01
-1.77739784e-01 -2.45931849e-01 -7.60017574e-01 -5.42591691e-01
-7.37116858e-02 -7.63161555e-02 3.84761542e-01 3.78652960e-01
5.76697104e-02 -1.53806945e-02 -3.88076216e-01 5.39692402e-01
1.42968249e+00 -3.77015471e-01 3.87880325e-01 3.37446958e-01
8.48019600e-01 9.20551300e-01 6.06455266e-01 9.00776803e-01
7.17134535e-01 1.22155964e+00 3.11222643e-01 1.07493795e-01
1.33979525e-02 -3.33743274e-01 3.13430130e-01 7.97921836e-01
-4.03104037e-01 -6.07893467e-02 -1.05099690e+00 2.92565227e-01
-2.09636712e+00 -7.72251010e-01 -2.28560448e-01 2.50099564e+00
1.66447937e-01 -4.29063328e-02 5.48427343e-01 1.94567338e-01
7.52309084e-01 2.20865719e-02 -2.88393408e-01 4.22840923e-01
-2.03448191e-01 -3.59967917e-01 2.57250845e-01 5.84014416e-01
-9.38128471e-01 4.82805252e-01 5.76908445e+00 2.03221709e-01
-3.60719711e-01 -1.13556691e-01 -1.13660470e-01 -1.22044288e-01
1.12051874e-01 -1.18497643e-03 -1.03468335e+00 1.85267121e-01
6.34613782e-02 1.50075287e-01 7.48871863e-02 4.20289576e-01
1.24311611e-01 -4.37142193e-01 -1.12427509e+00 1.32138884e+00
4.64600503e-01 -7.45162308e-01 -9.18656811e-02 3.76827598e-01
6.47815764e-01 -3.99784267e-01 -2.60688543e-01 -1.95361733e-01
-2.90572673e-01 -7.13188708e-01 9.29705501e-01 5.57618856e-01
3.77199560e-01 -6.26100957e-01 6.98818207e-01 5.11356950e-01
-1.38608849e+00 8.22339058e-02 -3.60950440e-01 -3.69203180e-01
3.92786056e-01 5.71207166e-01 -4.66168970e-01 6.59340799e-01
3.80889714e-01 5.97866595e-01 -4.34348285e-01 1.27295363e+00
-1.20917097e-01 -4.33213919e-01 -4.12853748e-01 1.57629430e-01
-2.14779884e-01 -3.56852174e-01 8.75084639e-01 9.95535016e-01
2.03317881e-01 7.35538974e-02 3.87913793e-01 4.10240293e-01
3.45102340e-01 1.32983640e-01 -5.73961854e-01 4.77370888e-01
3.03794086e-01 1.10983610e+00 -5.10638297e-01 -6.15736982e-03
-6.92869365e-01 1.20581663e+00 3.53914469e-01 3.78408641e-01
-6.55181229e-01 -4.08138856e-02 6.36269987e-01 4.53092903e-01
2.07454830e-01 -6.68712854e-01 7.22753257e-02 -1.55718136e+00
5.68176091e-01 -1.05245423e+00 2.87643820e-01 -5.85163534e-01
-1.04631662e+00 3.22490126e-01 1.41901776e-01 -1.30405903e+00
-3.64040732e-01 -5.00140369e-01 -2.59506285e-01 8.11307013e-01
-1.24867988e+00 -1.21422470e+00 -5.53272963e-01 8.09433639e-01
3.34477901e-01 8.07206482e-02 6.36305690e-01 4.10049468e-01
-5.33934116e-01 5.04606009e-01 -2.06594646e-01 1.82597488e-01
8.07475328e-01 -1.02502239e+00 3.55110794e-01 9.76646364e-01
-1.93347648e-01 5.77750027e-01 8.75135243e-01 -7.51996696e-01
-1.78489721e+00 -5.49173534e-01 1.09545910e+00 -6.26369894e-01
2.97334254e-01 -4.74945694e-01 -3.08654904e-01 5.79946101e-01
-3.16461891e-01 2.10667439e-02 5.06771922e-01 -6.24056235e-02
-1.81269363e-01 1.16665833e-01 -1.00655615e+00 4.73669261e-01
1.06964099e+00 -2.80682385e-01 -5.27955174e-01 1.87834457e-01
1.98794395e-01 -6.94969356e-01 -9.42814589e-01 2.86545724e-01
8.47258687e-01 -1.09077096e+00 1.31773317e+00 -1.66980267e-01
5.41424453e-02 -4.01521355e-01 -4.05667484e-01 -9.76822734e-01
-2.77978957e-01 -6.58558965e-01 -5.27536944e-02 8.76877904e-01
-1.39949381e-01 -2.87530839e-01 9.83050525e-01 7.83222914e-01
3.43889922e-01 -5.05386174e-01 -1.02423370e+00 -5.21417379e-01
-4.52980608e-01 -5.23271561e-02 2.26847142e-01 5.16354918e-01
-1.80285629e-02 5.36837220e-01 -7.00831413e-01 3.08164865e-01
1.25319803e+00 2.13559642e-02 1.33723891e+00 -1.38028717e+00
-5.08990169e-01 -7.45073855e-02 -7.75797069e-01 -1.43278432e+00
-1.55125648e-01 -2.28970990e-01 1.99438930e-01 -1.39278972e+00
3.94025654e-01 -1.72772463e-02 4.64222193e-01 -2.78715461e-01
-3.33338052e-01 9.61661488e-02 4.51056272e-01 4.35313672e-01
-6.13848448e-01 5.82913101e-01 1.19427729e+00 2.62925029e-01
-5.25918119e-02 2.13813663e-01 -1.74671173e-01 9.85440075e-01
2.83689648e-01 -3.20424736e-01 -1.37968659e-01 -5.44123232e-01
3.19429725e-01 5.74542999e-01 6.90895677e-01 -1.13665712e+00
4.84990716e-01 2.48941090e-02 5.62766731e-01 -5.83257198e-01
8.45093787e-01 -8.44474077e-01 3.58485490e-01 5.38942754e-01
2.25701168e-01 3.85868788e-01 -2.80757129e-01 7.77407169e-01
-2.66364533e-02 -2.41313167e-02 7.64399529e-01 -3.77995104e-01
-4.92427111e-01 3.68969887e-01 -2.50377841e-02 1.87137544e-01
8.69674742e-01 -6.19466305e-01 1.94851886e-02 -5.23720503e-01
-6.92317307e-01 1.33312076e-01 7.62504280e-01 3.16542745e-01
7.87192166e-01 -1.33601773e+00 -7.58548200e-01 2.45471582e-01
2.16487363e-01 1.43798977e-01 2.44976029e-01 1.12252092e+00
-4.53969270e-01 4.87783402e-01 -1.70640290e-01 -6.99609280e-01
-1.48612750e+00 2.18824804e-01 3.94362330e-01 -1.59803897e-01
-5.64125359e-01 4.97765541e-01 9.78214368e-02 -3.57414693e-01
4.68004704e-01 2.57038295e-01 1.11613479e-02 -1.42172128e-01
4.18369204e-01 8.78920913e-01 -1.48341149e-01 -1.40097356e+00
-5.44405639e-01 1.19205225e+00 3.52621794e-01 -2.21566364e-01
1.18144000e+00 -4.95236427e-01 4.55598570e-02 2.78448194e-01
1.00839400e+00 1.94110870e-01 -1.21210325e+00 -5.40759981e-01
-2.37708867e-01 -8.18120003e-01 -3.03529769e-01 -1.16108067e-01
-6.24289751e-01 7.86872149e-01 3.90936404e-01 -3.27577204e-01
6.52542472e-01 -2.15864852e-01 8.94878566e-01 4.32113647e-01
5.35722792e-01 -1.07656729e+00 1.99034393e-01 3.29479933e-01
8.20352614e-01 -1.30133665e+00 5.72359502e-01 -6.89335585e-01
-4.48801070e-01 9.77075517e-01 5.51104426e-01 -1.88858658e-01
5.35802662e-01 -3.30695361e-02 -8.62764195e-03 -2.78387666e-01
-1.55428961e-01 -2.71935195e-01 4.87035543e-01 6.07791305e-01
3.78269166e-01 -1.57678366e-01 -2.14831010e-01 2.86481291e-01
-6.79334924e-02 -1.67774901e-01 3.48694354e-01 1.06182182e+00
-4.19303775e-01 -8.72697651e-01 -7.42468417e-01 -1.74706221e-01
-3.48268211e-01 4.67315108e-01 -3.81545305e-01 7.72474766e-01
7.12785497e-02 1.09495020e+00 -1.46613553e-01 -2.67668247e-01
8.38635266e-01 -1.16735652e-01 8.07425976e-01 -4.56626385e-01
-3.86970222e-01 3.96567106e-01 -1.32771619e-02 -7.65211701e-01
-6.21101737e-01 -1.05426538e+00 -7.54783332e-01 -3.68379623e-01
-3.37689370e-01 -9.93110240e-02 5.54378390e-01 9.64711130e-01
1.72350556e-01 -5.28812408e-02 3.16129088e-01 -1.19034672e+00
-7.55261898e-01 -6.03107274e-01 -6.66219413e-01 7.22211182e-01
3.06440115e-01 -9.52021897e-01 -3.17697346e-01 -1.32232293e-01] | [7.071255683898926, -0.9983372688293457] |
1d3a8d53-05c4-45ba-860f-68a7707a7b06 | adaptive-unimodal-cost-volume-filtering-for | 1909.03751 | null | https://arxiv.org/abs/1909.03751v2 | https://arxiv.org/pdf/1909.03751v2.pdf | Adaptive Unimodal Cost Volume Filtering for Deep Stereo Matching | State-of-the-art deep learning based stereo matching approaches treat disparity estimation as a regression problem, where loss function is directly defined on true disparities and their estimated ones. However, disparity is just a byproduct of a matching process modeled by cost volume, while indirectly learning cost volume driven by disparity regression is prone to overfitting since the cost volume is under constrained. In this paper, we propose to directly add constraints to the cost volume by filtering cost volume with unimodal distribution peaked at true disparities. In addition, variances of the unimodal distributions for each pixel are estimated to explicitly model matching uncertainty under different contexts. The proposed architecture achieves state-of-the-art performance on Scene Flow and two KITTI stereo benchmarks. In particular, our method ranked the $1^{st}$ place of KITTI 2012 evaluation and the $4^{th}$ place of KITTI 2015 evaluation (recorded on 2019.8.20). The codes of AcfNet are available at: https://github.com/DeepMotionAIResearch/DenseMatchingBenchmark. | ['Zhiwei Li', 'Kun Yu', 'Youmin Zhang', 'Yimin Chen', 'Suihanjin Yu', 'Xiao Bai', 'Kuiyuan Yang'] | 2019-09-09 | null | null | null | null | ['stereo-matching'] | ['computer-vision'] | [-1.73143908e-01 -1.93600550e-01 -1.48295030e-01 -5.49066663e-01
-1.04787242e+00 -2.30735675e-01 5.22900403e-01 -1.30783051e-01
-6.50451183e-01 7.85141647e-01 2.25620300e-01 -1.24527469e-01
1.69811696e-01 -8.84254992e-01 -1.03073967e+00 -5.35513222e-01
1.63819604e-02 3.60590547e-01 2.19520256e-01 1.40996231e-02
5.13269246e-01 1.54336900e-01 -1.82043457e+00 2.87529051e-01
1.07906032e+00 1.28124237e+00 -2.97560799e-03 6.55501425e-01
-1.35357633e-01 8.62467766e-01 -3.13148350e-01 -6.01434052e-01
7.08628953e-01 -2.89492667e-01 -4.42245960e-01 -2.62594879e-01
1.29168415e+00 -6.97271049e-01 -7.63661683e-01 1.29942691e+00
5.88379741e-01 1.57867625e-01 6.22410893e-01 -1.30574226e+00
-2.80437618e-01 2.12473869e-01 -8.93636227e-01 3.66076410e-01
1.55440539e-01 5.80806196e-01 9.15633321e-01 -1.03810990e+00
4.36511040e-01 1.27633691e+00 5.43255866e-01 2.82847881e-01
-1.18269563e+00 -9.06630218e-01 1.88910246e-01 3.47316504e-01
-1.32711923e+00 -6.69817328e-01 5.94437897e-01 -7.87171781e-01
8.16007316e-01 5.81229553e-02 6.48074985e-01 7.48296797e-01
2.27610931e-01 7.73477674e-01 1.03353560e+00 -2.86750663e-02
1.22276567e-01 -2.62118816e-01 -3.66882756e-02 6.65563047e-01
2.71568209e-01 7.33111322e-01 -6.03670239e-01 2.01711044e-01
8.56421947e-01 -2.51721680e-01 -4.33930486e-01 -3.81645292e-01
-1.00805771e+00 7.61325479e-01 7.17019916e-01 -3.59105110e-01
-1.19987689e-01 6.56081855e-01 3.08319122e-01 5.38327768e-02
5.35498261e-01 -1.81883164e-02 -2.44192347e-01 -3.68664891e-01
-9.85736847e-01 4.31780756e-01 3.64412934e-01 1.04098856e+00
1.23838151e+00 5.15304022e-02 -4.05464709e-01 7.79111981e-01
2.73299277e-01 7.11461723e-01 1.03715353e-01 -1.45399880e+00
9.07581449e-01 2.34410763e-01 8.16869885e-02 -8.20579171e-01
-1.86107635e-01 -2.90333539e-01 -7.16153145e-01 6.60424232e-01
8.19846570e-01 -1.74427494e-01 -1.03401732e+00 1.74628901e+00
1.59380198e-01 5.43751359e-01 -3.45077038e-01 1.19105422e+00
1.03426254e+00 5.68893969e-01 -1.41106397e-01 1.92746133e-01
8.89927745e-01 -1.08873069e+00 -2.42670849e-01 -4.29268211e-01
3.17897052e-01 -8.82094502e-01 9.72097516e-01 2.73880243e-01
-1.24598861e+00 -7.53834188e-01 -9.21068966e-01 -2.98888534e-01
2.13683601e-02 -2.66653955e-01 5.51057696e-01 4.37113583e-01
-1.14644599e+00 6.52097166e-01 -6.84733987e-01 3.90263610e-02
6.23487592e-01 2.85648882e-01 -1.02880813e-01 -2.96717405e-01
-1.06096995e+00 6.94893181e-01 -2.25775782e-02 1.03069864e-01
-8.26397479e-01 -1.04794919e+00 -1.06592691e+00 -6.83396682e-02
3.40963006e-02 -8.97924721e-01 1.03317940e+00 -7.62189746e-01
-1.35364270e+00 1.21178734e+00 -3.21998745e-01 -4.51469839e-01
9.84845161e-01 -3.26230943e-01 -3.07036135e-02 -9.50739253e-03
2.89369553e-01 1.09888208e+00 8.23263705e-01 -1.14738441e+00
-9.79449689e-01 -2.36528903e-01 1.73936501e-01 1.48729354e-01
1.76732376e-01 -5.82616687e-01 -5.85632920e-01 -4.53120023e-01
9.30766389e-02 -6.56092525e-01 -7.99408481e-02 4.05845195e-01
-3.56393129e-01 2.03242540e-01 2.55470842e-01 -4.92133826e-01
1.11709845e+00 -2.07405901e+00 -1.78569421e-01 9.92724672e-02
3.07805359e-01 -1.18366197e-01 -8.18999857e-02 -7.32069463e-02
1.16794869e-01 -2.06191704e-01 -3.46521080e-01 -4.08262104e-01
8.52000341e-02 -2.45307997e-01 -2.74264872e-01 7.00467467e-01
9.98158194e-03 7.79580534e-01 -8.57698441e-01 -5.14837325e-01
6.34077072e-01 5.30000567e-01 -8.01526546e-01 2.02789098e-01
4.74267900e-02 5.28917909e-01 -1.07709788e-01 6.41455531e-01
1.21840763e+00 -1.04869239e-01 -3.47561419e-01 -3.42975140e-01
-3.11180174e-01 2.02369496e-01 -1.20975173e+00 1.84182930e+00
-4.09724444e-01 1.01373494e+00 -1.71108931e-01 -6.36298954e-01
8.99724066e-01 -2.89430112e-01 5.81829727e-01 -9.25676227e-01
6.27710368e-04 2.54306823e-01 3.28720734e-02 -2.95758873e-01
5.33829153e-01 3.31054702e-02 1.52094379e-01 1.13565922e-02
-1.70918569e-01 -5.30732334e-01 2.16661930e-01 -1.48524687e-01
8.50460708e-01 3.38993222e-01 -8.43490139e-02 -3.49578857e-01
5.55644691e-01 -6.80679157e-02 8.42552841e-01 6.68454170e-01
-5.63660026e-01 1.12003314e+00 4.25220817e-01 -5.10202110e-01
-1.07901371e+00 -1.40927923e+00 -4.69750166e-01 5.42479873e-01
5.53782105e-01 7.42213503e-02 -6.04772568e-01 -2.93752879e-01
3.28030914e-01 4.88084078e-01 -6.91061437e-01 4.01331894e-02
-6.91267312e-01 -3.44838887e-01 4.89318520e-01 5.85399508e-01
8.77924383e-01 -7.41119385e-01 -6.86653793e-01 6.90747723e-02
-3.05186689e-01 -1.05911505e+00 -8.21641386e-01 -2.54314572e-01
-8.28776240e-01 -1.24355423e+00 -8.60067248e-01 -4.90581959e-01
5.10656476e-01 1.86728597e-01 1.53213620e+00 -1.04506031e-01
-4.52634066e-01 7.02051744e-02 1.32487372e-01 -1.95876688e-01
1.87619731e-01 -1.45682245e-01 -3.51465523e-01 -6.68816715e-02
4.73098755e-01 -4.48663861e-01 -1.26018310e+00 3.18058491e-01
-6.27316952e-01 1.18589640e-01 1.65065780e-01 8.61925244e-01
6.64356232e-01 -4.86505479e-01 6.64908662e-02 -5.12134075e-01
7.09173232e-02 -2.71640778e-01 -1.16173613e+00 -1.17627263e-01
-5.45003295e-01 3.33918445e-02 2.62134820e-01 -1.47144943e-01
-1.26319671e+00 -1.19195119e-01 -1.10769115e-01 -6.95683479e-01
-7.03963786e-02 -1.71608422e-02 8.81257281e-02 3.69834304e-02
5.32199919e-01 2.41602696e-02 -6.78989142e-02 -1.37848169e-01
1.70874700e-01 2.88541704e-01 6.36269033e-01 -6.95223033e-01
6.80968344e-01 9.73315537e-01 2.15733215e-01 -4.50486511e-01
-7.04950213e-01 -4.00793284e-01 -3.05215806e-01 -5.65114558e-01
8.35679233e-01 -1.23130882e+00 -8.98961723e-01 7.55674005e-01
-9.68972266e-01 -6.19545281e-01 -3.66401374e-01 7.19976842e-01
-6.83663785e-01 3.02097559e-01 -5.90978265e-01 -5.45386851e-01
-3.15878719e-01 -1.37418056e+00 1.25239074e+00 4.62846428e-01
-8.44124407e-02 -9.92615283e-01 2.77762771e-01 5.09290636e-01
5.33412099e-01 2.32171610e-01 4.54319626e-01 2.97657013e-01
-1.04120123e+00 -2.69608945e-03 -6.33480191e-01 4.48951334e-01
-2.26115286e-02 1.62040502e-01 -1.41669202e+00 -2.50492603e-01
-1.87910825e-01 -2.58427680e-01 1.26155972e+00 1.15189302e+00
1.20382905e+00 2.37978414e-01 -2.49428786e-02 1.29696739e+00
1.74825454e+00 1.90239981e-01 1.06429994e+00 5.35863936e-01
7.03512967e-01 7.03760207e-01 6.42755687e-01 5.84825516e-01
6.01143360e-01 7.12162018e-01 5.80862343e-01 -7.27488995e-02
-4.94580805e-01 -3.27998579e-01 1.25706866e-01 1.55914620e-01
-9.71591845e-02 -7.28791505e-02 -1.02826250e+00 4.63204384e-01
-1.80343759e+00 -1.03522980e+00 -3.54958534e-01 2.52939773e+00
6.16790771e-01 3.70993316e-01 -6.75864369e-02 -1.77861139e-01
8.30798447e-01 3.79387170e-01 -7.49293029e-01 -2.35000148e-01
-3.69012654e-01 2.02292845e-01 7.74474263e-01 1.06098211e+00
-1.05593026e+00 9.25819278e-01 5.39859247e+00 8.44467819e-01
-1.19507062e+00 -6.57844096e-02 1.03890395e+00 -3.54334056e-01
-3.53277057e-01 1.07652240e-01 -7.22524762e-01 8.24115336e-01
3.63452107e-01 -2.40098864e-01 2.32562140e-01 6.85045421e-01
3.43302637e-01 -6.11140847e-01 -1.27012336e+00 1.36377990e+00
-1.97202563e-01 -1.55280507e+00 -2.16594011e-01 3.25572491e-01
9.78765070e-01 3.27575982e-01 4.23840731e-01 2.61741400e-01
1.65513586e-04 -1.11940181e+00 8.46753955e-01 6.36088073e-01
9.93516266e-01 -6.00342393e-01 6.89497650e-01 -1.47742718e-01
-1.26267278e+00 1.34071574e-01 -4.66974795e-01 -2.58276284e-01
2.68656343e-01 9.12023246e-01 -3.67733724e-02 2.61075377e-01
8.90144348e-01 9.39980686e-01 -2.84916937e-01 1.42515242e+00
-1.28644720e-01 1.05020881e-01 -2.21596003e-01 4.19355690e-01
8.59964415e-02 -4.46455389e-01 5.25486946e-01 1.26050675e+00
3.92359436e-01 -6.48749396e-02 -1.16073944e-01 9.67603147e-01
-2.55050212e-01 -1.78694040e-01 -5.02296805e-01 7.51667559e-01
4.15087372e-01 7.92027473e-01 -2.96804219e-01 -3.89510125e-01
-4.78173584e-01 6.98582709e-01 2.80968666e-01 4.23906505e-01
-1.00762010e+00 -2.83666193e-01 1.23210418e+00 1.95494816e-01
2.43793145e-01 1.88772410e-01 -7.83983648e-01 -1.45598090e+00
1.98921278e-01 -5.59428155e-01 2.56373823e-01 -7.56451607e-01
-1.36780226e+00 5.81856310e-01 2.05465895e-03 -1.69738090e+00
-1.92659333e-01 -5.33821881e-01 -7.28205085e-01 1.07044280e+00
-1.81970978e+00 -5.76118350e-01 -6.84179962e-01 6.14392281e-01
5.69704711e-01 3.08361351e-02 2.14692503e-01 6.65054739e-01
-4.51362342e-01 6.66895509e-01 4.53855135e-02 1.67519912e-01
8.87902558e-01 -1.08097756e+00 5.36462843e-01 9.12632525e-01
-2.73270935e-01 -4.76792715e-02 7.27819562e-01 -3.52410167e-01
-9.25866246e-01 -9.38093483e-01 7.32036650e-01 -2.86767483e-01
5.18118739e-01 -1.03274032e-01 -7.94099689e-01 4.41757411e-01
1.11406401e-01 3.71079773e-01 7.15311021e-02 -2.66258270e-01
-4.28632557e-01 -5.38226247e-01 -1.30854917e+00 5.35628915e-01
1.30435014e+00 -4.53097433e-01 -3.84881757e-02 -1.90467238e-02
4.10251677e-01 -9.17626917e-01 -6.51434481e-01 5.57878196e-01
6.41399682e-01 -1.63139856e+00 1.15230775e+00 -9.50714946e-02
9.29602504e-01 -3.09882432e-01 -3.89979035e-01 -1.04477572e+00
-9.07715932e-02 -3.67616594e-01 -1.43206045e-01 9.73148465e-01
4.13195074e-01 -8.02703559e-01 1.13273907e+00 8.37868273e-01
-2.13028073e-01 -7.81734586e-01 -1.26342225e+00 -7.02539444e-01
2.42920145e-01 -6.10472023e-01 5.30171275e-01 6.99684143e-01
-2.94840991e-01 -1.15448155e-01 -3.12583297e-01 -8.77178386e-02
1.26251280e+00 9.71566513e-02 8.92807484e-01 -8.45205843e-01
-4.34900552e-01 -8.91155422e-01 -6.31949961e-01 -1.37991405e+00
1.11130960e-01 -5.03730953e-01 1.26355797e-01 -1.40413201e+00
1.69051006e-01 -5.27737439e-01 -8.13920870e-02 -9.79457870e-02
-2.27679446e-01 2.25653276e-01 3.02946895e-01 1.57178149e-01
-1.37773633e-01 6.44300878e-01 1.20649552e+00 -3.11864316e-01
-8.43941942e-02 3.19145508e-02 -3.12073708e-01 7.08753169e-01
8.54105055e-01 -2.12411806e-01 -4.25622255e-01 -9.23795164e-01
1.35917366e-02 1.50701627e-01 4.70509470e-01 -1.13102245e+00
2.36269027e-01 -2.35105172e-01 3.30631226e-01 -7.11332977e-01
6.46153152e-01 -4.82602507e-01 4.62132916e-02 4.28825170e-01
-2.45144650e-01 1.19074928e-02 3.71878594e-01 2.47410670e-01
-4.91306931e-01 1.11551248e-01 1.00184226e+00 3.87633443e-02
-8.32773507e-01 6.31266057e-01 1.41775474e-01 6.76118970e-01
7.20356107e-01 -5.39152026e-01 -5.83128154e-01 -5.44667363e-01
-2.02215061e-01 3.83212358e-01 5.26764989e-01 2.33202934e-01
7.89792955e-01 -1.32218087e+00 -9.21935856e-01 1.12189665e-01
1.04755558e-01 2.55873561e-01 5.12296140e-01 6.25704885e-01
-8.56991351e-01 3.23482454e-01 -3.05916458e-01 -1.00017798e+00
-7.95464396e-01 1.20509028e-01 6.74692452e-01 -1.22442730e-01
-5.04604936e-01 1.19749582e+00 5.72675765e-01 -2.71984816e-01
5.57439923e-01 -3.87045056e-01 3.00268114e-01 -7.27372691e-02
3.86916012e-01 6.73745692e-01 -7.88261741e-02 -5.63598573e-01
-4.25287932e-01 9.53780115e-01 2.75182694e-01 -2.95275480e-01
9.78099942e-01 -2.53779799e-01 1.72279969e-01 1.33982494e-01
1.54629958e+00 -3.11772406e-01 -2.01441336e+00 -1.85875148e-01
-3.30251455e-01 -1.13008046e+00 1.41195908e-01 -5.28319538e-01
-1.49994433e+00 1.12237108e+00 9.30160165e-01 -5.23150444e-01
1.02408016e+00 -1.72042951e-01 8.64017427e-01 -1.29980266e-01
3.28790873e-01 -1.16733849e+00 -5.41132465e-02 5.04772425e-01
7.46317685e-01 -1.83976495e+00 -7.74281994e-02 -2.76752144e-01
-4.58795577e-01 9.82414365e-01 1.05414975e+00 -3.00729096e-01
8.55372012e-01 3.87594789e-01 2.78663367e-01 4.58723381e-02
-5.05889416e-01 -2.76964933e-01 4.03056294e-01 5.75382054e-01
6.02731586e-01 5.87081052e-02 -1.02186732e-01 -2.70368814e-01
-1.63628504e-01 2.86109578e-02 3.81572872e-01 7.15317547e-01
-2.77392238e-01 -7.99169600e-01 -2.92619646e-01 3.90914917e-01
-2.48181462e-01 -4.31182057e-01 3.56622279e-01 7.24868476e-01
1.18961707e-01 8.20144773e-01 5.99306941e-01 -3.47236484e-01
5.42350113e-01 -5.20422757e-01 4.50020701e-01 -2.91559607e-01
-2.95548737e-01 -9.56040770e-02 3.01766675e-02 -9.45023596e-01
-2.95237333e-01 -6.67461872e-01 -1.09376991e+00 -7.03048110e-01
2.05330685e-01 -3.70809555e-01 5.10190010e-01 5.51143467e-01
1.86016768e-01 1.91199422e-01 6.76782668e-01 -1.04992986e+00
-3.33833545e-01 -6.88849926e-01 -3.97463739e-01 5.35959661e-01
5.93750894e-01 -5.92904568e-01 -6.20600998e-01 -5.24873473e-02] | [8.791584014892578, -2.2831223011016846] |
efcc6e66-c896-4eba-a4e7-8165340663bd | interpretable-neural-embeddings-with-sparse | 2306.14135 | null | https://arxiv.org/abs/2306.14135v1 | https://arxiv.org/pdf/2306.14135v1.pdf | Interpretable Neural Embeddings with Sparse Self-Representation | Interpretability benefits the theoretical understanding of representations. Existing word embeddings are generally dense representations. Hence, the meaning of latent dimensions is difficult to interpret. This makes word embeddings like a black-box and prevents them from being human-readable and further manipulation. Many methods employ sparse representation to learn interpretable word embeddings for better interpretability. However, they also suffer from the unstable issue of grouped selection in $\ell1$ and online dictionary learning. Therefore, they tend to yield different results each time. To alleviate this challenge, we propose a novel method to associate data self-representation with a shallow neural network to learn expressive, interpretable word embeddings. In experiments, we report that the resulting word embeddings achieve comparable and even slightly better interpretability than baseline embeddings. Besides, we also evaluate that our approach performs competitively well on all downstream tasks and outperforms benchmark embeddings on a majority of them. | ['Hao Zhu', 'Minxue Xia'] | 2023-06-25 | null | null | null | null | ['dictionary-learning', 'word-embeddings'] | ['methodology', 'methodology'] | [-6.45515621e-02 2.97682911e-01 -6.42017782e-01 -5.02026558e-01
-4.12975580e-01 -5.10688424e-01 5.61048985e-01 4.30164695e-01
-4.60586220e-01 3.87963533e-01 7.59464383e-01 -5.19545019e-01
2.21571140e-02 -6.37206793e-01 -3.15980852e-01 -3.96860063e-01
1.19342431e-01 4.81119215e-01 -3.94701362e-01 -2.39963174e-01
1.55124277e-01 1.85419172e-01 -1.40447211e+00 2.14001417e-01
9.63058472e-01 9.10696864e-01 2.95687504e-02 1.33245155e-01
-3.86175334e-01 5.46683729e-01 -4.82355684e-01 -5.63873649e-01
1.38253182e-01 -2.77169365e-02 -8.52552533e-01 -5.73896877e-02
1.62250161e-01 -4.73141760e-01 -4.23614591e-01 9.92252052e-01
1.18780538e-01 2.48571157e-01 8.11126530e-01 -1.15031517e+00
-1.79534268e+00 7.55313337e-01 -3.85764062e-01 1.76733792e-01
1.58241376e-01 -2.68075671e-02 1.79033291e+00 -1.28045774e+00
3.25655639e-01 1.44923389e+00 5.12674570e-01 4.81485516e-01
-1.37168932e+00 -4.94935334e-01 5.05812705e-01 1.57576099e-01
-1.33243585e+00 -2.64543861e-01 5.83084464e-01 -1.84843183e-01
1.21907067e+00 3.97970557e-01 5.35840511e-01 1.30227315e+00
-1.56038597e-01 6.18344009e-01 6.23327553e-01 -4.16232377e-01
2.54704565e-01 1.08993404e-01 5.56348205e-01 6.43374979e-01
8.44434679e-01 -1.05940834e-01 -3.05717289e-01 -6.96931779e-02
6.85940206e-01 4.82069582e-01 -1.04138240e-01 -1.74589694e-01
-1.07380939e+00 1.32022309e+00 6.52126193e-01 3.25308293e-01
-1.46505833e-01 5.97971439e-01 5.07859409e-01 2.99463779e-01
7.52798498e-01 8.13287318e-01 -3.27591419e-01 -1.09356388e-01
-2.45057613e-01 1.36258587e-01 4.39156085e-01 7.51713395e-01
7.83681989e-01 2.64268011e-01 -1.31919578e-01 9.61622059e-01
6.00140989e-01 2.70347625e-01 7.52630055e-01 -6.30375206e-01
4.53619570e-01 8.18900704e-01 -7.53671303e-03 -1.28598642e+00
-2.60898560e-01 -3.18686545e-01 -9.10776615e-01 -1.07757486e-01
1.16636202e-01 1.88847199e-01 -1.01462400e+00 1.63008821e+00
-1.41738877e-01 -5.28939292e-02 1.03567764e-01 1.03684652e+00
7.79512465e-01 8.61143172e-01 4.19041336e-01 1.90028951e-01
1.72844958e+00 -1.12532234e+00 -1.06783235e+00 -7.49056935e-01
7.91852236e-01 -5.76242387e-01 1.64404333e+00 1.49002582e-01
-8.11914742e-01 -5.47502816e-01 -1.23496377e+00 -6.87693954e-01
-6.69467926e-01 6.44010901e-02 1.22424150e+00 5.78454614e-01
-7.42495179e-01 3.88721198e-01 -8.68858278e-01 -1.21779785e-01
5.43177128e-01 2.43303373e-01 -4.47736382e-01 -1.40294537e-01
-1.31750190e+00 1.08217478e+00 4.99903560e-01 2.28339564e-02
-4.23885077e-01 -6.40813410e-01 -1.26269889e+00 2.21305937e-01
-7.77504966e-02 -6.15201473e-01 9.56607699e-01 -9.13613021e-01
-1.13256347e+00 7.41785347e-01 -4.35846061e-01 -5.33388078e-01
-1.16636358e-01 -6.63189948e-01 -4.02931005e-01 -2.17081010e-01
8.76085386e-02 6.86141372e-01 8.29007328e-01 -1.05826104e+00
-2.33064711e-01 -2.35149235e-01 5.55953145e-01 6.28014505e-02
-1.08926177e+00 -2.95060813e-01 -3.02977562e-01 -1.06378675e+00
1.28569929e-02 -6.46614015e-01 -4.56625253e-01 2.76492923e-01
-2.05432907e-01 -6.01833940e-01 6.80140615e-01 -3.37720305e-01
1.66546428e+00 -2.32126045e+00 2.65535206e-01 -1.11766458e-01
7.16023326e-01 3.95729095e-01 -4.43032444e-01 3.14773947e-01
-3.43097419e-01 7.33153939e-01 4.58250716e-02 -6.87360406e-01
2.98001379e-01 7.91999519e-01 -6.15114689e-01 3.27837378e-01
6.12658322e-01 1.13960540e+00 -1.04845774e+00 -9.74746421e-02
1.89831138e-01 5.29170930e-01 -8.95336270e-01 2.91333735e-01
-9.14348066e-02 -9.57581773e-02 -4.89015609e-01 4.19520497e-01
5.38826048e-01 -4.34202999e-01 2.04779401e-01 -2.52348632e-01
1.72530636e-01 5.14977634e-01 -1.07384598e+00 1.77258503e+00
-8.62801552e-01 8.99116814e-01 -5.54430306e-01 -1.26662040e+00
9.49377239e-01 1.87380478e-01 -1.11618340e-01 -5.11472166e-01
2.50141285e-02 1.87961787e-01 -2.84862649e-02 -4.20424074e-01
8.98278177e-01 -3.50256324e-01 -2.47496173e-01 6.57612026e-01
-9.22618806e-03 6.16804734e-02 4.79041934e-02 1.33161917e-01
8.71891916e-01 -3.00426126e-01 5.15159607e-01 -3.55254024e-01
1.67461827e-01 -4.04739767e-01 5.44694006e-01 4.39652503e-01
9.18011144e-02 7.91024923e-01 5.49227357e-01 -9.20098126e-01
-1.06120050e+00 -1.25402474e+00 -1.77735135e-01 1.25503397e+00
3.37907851e-01 -8.34410965e-01 -4.00716424e-01 -4.90346581e-01
1.73935741e-01 1.05485141e+00 -8.36858094e-01 -6.03612065e-01
-3.98315102e-01 -6.24266684e-01 4.18040037e-01 1.06039071e+00
-2.12675869e-01 -7.79836357e-01 -3.04394752e-01 1.00210771e-01
-1.12640865e-01 -1.01086819e+00 -4.43300992e-01 3.90031517e-01
-8.79698336e-01 -5.76117575e-01 -2.91640878e-01 -7.44632840e-01
1.06940627e+00 5.21729171e-01 1.41365314e+00 3.70278448e-01
-1.65941462e-01 -7.91923851e-02 -6.05301976e-01 -4.11661595e-01
-1.40319347e-01 2.84701109e-01 3.81199211e-01 -2.21785992e-01
8.47302914e-01 -4.53761399e-01 -6.44016087e-01 6.28755316e-02
-1.26474822e+00 -1.72819078e-01 4.59653705e-01 1.13533688e+00
3.52821946e-01 -3.26478720e-01 6.15149975e-01 -7.84816861e-01
1.18029213e+00 -6.25226498e-01 -1.40150174e-01 5.43274693e-02
-9.13623273e-01 5.52796602e-01 9.03566658e-01 -4.77314681e-01
-5.39390087e-01 -3.46906573e-01 -1.38073817e-01 -4.08529073e-01
1.28340423e-01 6.25857949e-01 -6.38029799e-02 4.83495176e-01
7.64986634e-01 -1.01920426e-01 -2.08729804e-02 -5.28038919e-01
9.55819845e-01 8.44487846e-01 2.48071868e-02 -6.04287624e-01
1.04273653e+00 3.69922251e-01 -5.67619860e-01 -6.29065752e-01
-1.20725965e+00 -3.24150831e-01 -4.17733967e-01 4.55730587e-01
8.41732144e-01 -1.06762505e+00 -1.76359698e-01 -2.53755152e-01
-1.52233160e+00 2.46735528e-01 -5.79634607e-01 3.65333587e-01
-7.25698397e-02 4.39795911e-01 -4.48738605e-01 -5.85636735e-01
-3.03956181e-01 -1.27685773e+00 1.12186491e+00 8.57471898e-02
-9.26432908e-01 -1.32176888e+00 -5.30532040e-02 3.81255507e-01
5.23675978e-01 2.56939456e-02 1.12526572e+00 -7.25955427e-01
-2.95603603e-01 -3.88798237e-01 -4.40158993e-01 5.00109076e-01
3.83230060e-01 -2.05549970e-01 -1.09688735e+00 -2.07859695e-01
-3.37293386e-01 -4.96493816e-01 1.00626147e+00 -8.66184533e-02
1.62216139e+00 -6.10920668e-01 -1.37132630e-01 6.31558180e-01
1.21911979e+00 -2.79888839e-01 7.36042500e-01 1.94537446e-01
7.78199255e-01 3.44095409e-01 4.61327195e-01 4.54876691e-01
4.04736340e-01 4.22606915e-01 4.97424513e-01 -2.58241802e-01
9.61391404e-02 -3.56784046e-01 5.16316116e-01 1.09697056e+00
-2.54778326e-01 -3.50534320e-01 -7.61744857e-01 8.20406437e-01
-1.74715304e+00 -6.62311316e-01 5.36099635e-02 1.77093840e+00
8.52877378e-01 1.35149330e-01 -2.59317636e-01 1.60276368e-01
3.80179137e-01 5.99521279e-01 -4.09872919e-01 -9.81474459e-01
7.26065636e-02 6.09647691e-01 2.89397448e-01 5.29214621e-01
-8.61275434e-01 1.05625415e+00 6.28325415e+00 7.39133894e-01
-8.89355063e-01 2.88464397e-01 6.69571221e-01 -2.35613640e-02
-1.17483783e+00 -6.05602562e-02 -5.73111773e-01 3.75493139e-01
8.69455755e-01 -2.08030447e-01 3.71252120e-01 9.26084161e-01
4.17459123e-02 4.01466399e-01 -1.37887275e+00 1.13089728e+00
-2.67039873e-02 -1.47139120e+00 5.72885990e-01 9.09850094e-03
4.63810951e-01 -1.93296820e-01 4.33107674e-01 4.00998622e-01
2.95122147e-01 -1.72249031e+00 6.50022626e-01 3.64586487e-02
8.73296320e-01 -7.12776959e-01 7.25703716e-01 -5.92580298e-03
-1.20988834e+00 -1.32505566e-01 -8.95323455e-01 -5.86460114e-01
2.88696159e-02 5.22841513e-01 -4.03999150e-01 2.23805785e-01
3.48747432e-01 8.13674152e-01 -5.33487022e-01 2.73266912e-01
-6.68667674e-01 4.68376577e-01 -4.19791304e-02 -3.26823086e-01
5.20541728e-01 -9.38940868e-02 1.86550334e-01 1.14097607e+00
2.15553626e-01 -3.91755067e-02 -4.22847457e-03 1.10546458e+00
-3.62584203e-01 2.10860167e-02 -9.63445365e-01 -5.77795982e-01
5.34440100e-01 1.21154952e+00 -4.36036646e-01 -2.59538293e-01
-6.36359513e-01 1.06949723e+00 7.52083898e-01 3.25030506e-01
-8.94076705e-01 -3.71978551e-01 1.47849238e+00 4.69744243e-02
2.41631120e-01 -5.33772528e-01 -5.29230714e-01 -1.49549711e+00
7.48298839e-02 -6.71770155e-01 3.06985825e-01 -5.59613466e-01
-1.55657685e+00 6.70214295e-01 -1.20427273e-01 -1.14999056e+00
-2.63418764e-01 -1.01343834e+00 -5.30155003e-01 6.05246127e-01
-1.59510660e+00 -9.27806377e-01 -2.23866418e-01 -3.64422873e-02
7.34884381e-01 -4.95645292e-02 1.34251225e+00 2.31036082e-01
-6.58100545e-01 9.40455377e-01 1.87167719e-01 2.33133480e-01
4.14252102e-01 -1.37107754e+00 7.68938422e-01 5.81929266e-01
6.19980514e-01 1.12371671e+00 6.40681088e-01 -1.19530633e-01
-1.65174365e+00 -1.18941355e+00 8.81247163e-01 -7.09746480e-01
9.62835550e-01 -6.52596891e-01 -1.00406706e+00 8.36474657e-01
4.28926080e-01 2.65703559e-01 1.14379549e+00 6.84430897e-01
-8.30591679e-01 -1.13311363e-02 -7.00617313e-01 8.32272589e-01
1.15564585e+00 -7.46782362e-01 -1.10188484e+00 2.60501832e-01
1.12065530e+00 -2.04989538e-02 -8.63632441e-01 -7.59746879e-02
4.84531224e-01 -6.66876435e-01 1.24432075e+00 -1.08567262e+00
7.85680294e-01 -1.66163892e-01 -1.81805819e-01 -1.56434524e+00
-6.39357328e-01 -4.11176145e-01 -3.66978735e-01 1.21576571e+00
5.87510765e-01 -8.47552121e-01 5.13473749e-01 7.78137207e-01
-1.22459447e-02 -8.55935931e-01 -7.95151353e-01 -8.45762491e-01
3.72127920e-01 -5.56114554e-01 7.89052546e-01 1.12467659e+00
2.74897218e-01 7.13223696e-01 -2.21454173e-01 1.09937228e-02
4.07250494e-01 1.57736167e-01 3.75013232e-01 -1.12192249e+00
-5.48094027e-02 -4.30991381e-01 -6.20931566e-01 -1.31647062e+00
5.28056204e-01 -1.22737908e+00 -1.53864324e-01 -1.48491347e+00
2.89025959e-02 -4.64888275e-01 -5.87472200e-01 5.32783270e-01
-5.17441273e-01 3.46886575e-01 -2.93988921e-02 -2.71475641e-03
-5.81518710e-01 7.14672804e-01 9.38672960e-01 -2.30010524e-01
1.35491222e-01 -7.25344181e-01 -1.30416274e+00 7.31368601e-01
1.00141025e+00 -3.64704400e-01 -6.73826456e-01 -1.28705823e+00
3.66788119e-01 -8.14106822e-01 2.19268799e-01 -4.59870011e-01
-2.41282091e-01 -1.92588180e-01 1.82953283e-01 -6.89617097e-02
2.44835421e-01 -7.29779303e-01 -2.81219870e-01 3.37318063e-01
-4.47030216e-01 3.58721077e-01 7.71403983e-02 7.03498960e-01
-3.63853842e-01 -3.40950847e-01 5.09346366e-01 2.80737933e-02
-8.18245232e-01 3.23974729e-01 -4.15755302e-01 3.76342535e-01
7.74872005e-01 -2.37436995e-01 -1.23521067e-01 -3.03804189e-01
-1.81661233e-01 1.22476660e-01 4.21684980e-01 8.21391404e-01
7.95511067e-01 -1.78628123e+00 -4.70847785e-01 3.21480989e-01
4.67841506e-01 2.21501410e-01 -1.99259341e-01 2.82672524e-01
-4.81519878e-01 4.82750565e-01 4.56001945e-02 -2.71091610e-01
-7.07719147e-01 5.57781994e-01 -1.26959518e-01 -2.38316610e-01
-6.19718671e-01 9.19844151e-01 4.44200814e-01 -3.30966681e-01
1.42003894e-01 -8.65550518e-01 -2.28266135e-01 2.08986819e-01
7.80297399e-01 2.48387009e-02 -2.82821149e-01 -4.81119275e-01
-3.11291993e-01 4.85998631e-01 -2.25429580e-01 3.64904344e-01
1.58227503e+00 -5.29186092e-02 -1.02313362e-01 4.56201375e-01
1.57606924e+00 -1.66059434e-01 -7.97434509e-01 -2.09960729e-01
1.23557307e-01 -7.07478046e-01 2.20469795e-02 -1.15445778e-01
-1.01404560e+00 1.18375206e+00 2.99427211e-01 4.78763461e-01
7.93442726e-01 7.77511522e-02 9.50505018e-01 5.32393038e-01
-1.29754648e-01 -1.00130987e+00 2.72203594e-01 6.11489415e-01
9.74416137e-01 -1.32502317e+00 1.64359167e-01 -3.52689356e-01
-5.87324858e-01 1.31466866e+00 5.94823539e-01 -3.02675188e-01
5.93858480e-01 8.55556279e-02 5.94374202e-02 -2.12314546e-01
-8.30123961e-01 -7.78745338e-02 3.93375039e-01 6.69257522e-01
6.99541152e-01 1.50675192e-01 -3.39474469e-01 8.97112787e-01
-1.86891675e-01 -2.84421980e-01 2.48904407e-01 4.93358135e-01
-3.05396914e-01 -1.29013050e+00 6.98782317e-03 4.08276558e-01
-3.10160518e-01 -4.50276196e-01 -1.91569328e-01 6.95893943e-01
-5.38516007e-02 7.45767593e-01 3.38310838e-01 -3.45628768e-01
2.62114257e-01 -6.13553487e-02 5.90104386e-02 -8.29499662e-01
-2.33317927e-01 -5.33500016e-01 1.49937481e-01 -4.99439359e-01
-1.10658571e-01 -2.09021166e-01 -1.32317519e+00 -3.95614952e-01
-3.02975953e-01 1.94242761e-01 2.50675231e-01 9.47989166e-01
5.13401449e-01 4.02088225e-01 5.36067545e-01 -4.93495792e-01
-8.74467790e-01 -8.81964087e-01 -4.35347378e-01 7.41822243e-01
2.78201908e-01 -8.72776449e-01 -5.35335243e-01 -5.94980009e-02] | [10.509124755859375, 8.646236419677734] |
980a18c5-8131-43eb-aa1e-a5b2b4e94a73 | learning-object-permanence-from-video | 2003.10469 | null | https://arxiv.org/abs/2003.10469v4 | https://arxiv.org/pdf/2003.10469v4.pdf | Learning Object Permanence from Video | Object Permanence allows people to reason about the location of non-visible objects, by understanding that they continue to exist even when not perceived directly. Object Permanence is critical for building a model of the world, since objects in natural visual scenes dynamically occlude and contain each-other. Intensive studies in developmental psychology suggest that object permanence is a challenging task that is learned through extensive experience. Here we introduce the setup of learning Object Permanence from data. We explain why this learning problem should be dissected into four components, where objects are (1) visible, (2) occluded, (3) contained by another object and (4) carried by a containing object. The fourth subtask, where a target object is carried by a containing object, is particularly challenging because it requires a system to reason about a moving location of an invisible object. We then present a unified deep architecture that learns to predict object location under these four scenarios. We evaluate the architecture and system on a new dataset based on CATER, and find that it outperforms previous localization methods and various baselines. | ['Gal Chechik', 'Amir Globerson', 'Aviv Shamsian', 'Ofri Kleinfeld'] | 2020-03-23 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2481_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123610035.pdf | eccv-2020-8 | ['video-object-tracking'] | ['computer-vision'] | [ 1.80352420e-01 -2.91299373e-01 6.04502819e-02 -2.48657644e-01
-9.13784578e-02 -7.41488159e-01 6.77709222e-01 1.33598745e-01
-5.25949538e-01 4.92009968e-01 4.95557114e-02 -1.25407204e-01
6.99209124e-02 -3.69716763e-01 -1.29757798e+00 -7.38220751e-01
-1.98630348e-01 2.26801708e-01 5.92392623e-01 2.32306466e-01
3.22269380e-01 4.88825858e-01 -1.93425214e+00 3.98440510e-01
2.57112294e-01 7.45371580e-01 8.30338597e-01 5.86670816e-01
2.00986043e-01 8.01485837e-01 -6.24892473e-01 -8.57165307e-02
2.96715111e-01 -3.09739947e-01 -9.01687562e-01 3.01659796e-02
1.10646880e+00 -6.60845697e-01 -2.35022083e-01 9.06049073e-01
1.38237953e-01 2.63716131e-01 4.73680496e-01 -1.22989702e+00
-1.45660484e+00 6.56441212e-01 -5.83521008e-01 6.85293853e-01
4.56411988e-01 1.41168132e-01 8.63338888e-01 -9.22427714e-01
6.03405297e-01 1.25672555e+00 3.25385690e-01 6.45075738e-01
-1.28862536e+00 -1.50580332e-01 8.24483931e-01 3.21935058e-01
-1.35537589e+00 -6.00008488e-01 5.76401055e-01 -6.28467500e-01
1.00444400e+00 3.61980319e-01 9.37090576e-01 1.33477700e+00
3.52848321e-01 7.46119201e-01 9.72813427e-01 -5.24095118e-01
2.90510803e-01 2.84118522e-02 2.44014159e-01 4.46132809e-01
3.20415854e-01 3.74917351e-02 -5.95205188e-01 3.89291137e-01
7.00422525e-01 2.87019342e-01 -3.45081300e-01 -5.20334840e-01
-1.29503238e+00 4.13106717e-02 7.63433814e-01 5.04702151e-01
-1.86979651e-01 4.49717641e-01 -2.98054487e-01 1.36011615e-01
1.38377324e-01 3.21523100e-01 -4.55242991e-01 2.14650873e-02
-7.53765643e-01 2.53164858e-01 2.70031095e-01 1.07620728e+00
5.84086120e-01 -2.47623265e-01 -1.57767996e-01 3.61008406e-01
3.23945731e-01 4.13704515e-01 4.60538656e-01 -6.56673193e-01
2.88153440e-01 4.41123813e-01 5.81657290e-01 -7.98207998e-01
-4.01493102e-01 -4.85549301e-01 -1.81100786e-01 4.39294964e-01
4.83200848e-01 3.20572346e-01 -1.02551126e+00 2.23630214e+00
4.47335482e-01 4.54137564e-01 -2.08454981e-01 1.14539647e+00
9.29820776e-01 5.23505151e-01 1.78849921e-01 -6.41345233e-02
1.53623366e+00 -1.06564343e+00 -5.78968465e-01 -7.13511169e-01
1.94479123e-01 -3.40407491e-01 1.23628628e+00 2.58850217e-01
-1.16617835e+00 -7.78411865e-01 -1.09676886e+00 -5.06960690e-01
-7.04168558e-01 2.49290109e-01 9.27471817e-01 2.52319694e-01
-1.46024632e+00 3.77984077e-01 -1.01831377e+00 -5.41722894e-01
5.52324593e-01 1.99493974e-01 -6.20232761e-01 1.36847898e-01
-6.82506144e-01 1.06709862e+00 4.43341225e-01 4.01067674e-01
-1.47665405e+00 -6.32972181e-01 -9.74893689e-01 2.76311457e-01
2.23523900e-01 -8.02057922e-01 1.25763643e+00 -1.04168880e+00
-7.82448351e-01 1.28369820e+00 -4.41429079e-01 -1.23397961e-01
3.03373277e-01 -6.14871919e-01 -3.95484716e-01 2.57360116e-02
3.34961563e-01 7.32823074e-01 1.01283014e+00 -1.57170689e+00
-6.48346782e-01 -6.87443197e-01 2.74434149e-01 2.99896449e-01
-1.56605281e-02 7.15767890e-02 -4.23866838e-01 -3.47145915e-01
6.42602921e-01 -8.33114922e-01 4.01570082e-01 4.33903515e-01
-3.67679328e-01 -4.22453851e-01 1.02723563e+00 -4.92550910e-01
8.04282486e-01 -2.14931679e+00 -4.66274917e-02 -6.68546200e-01
3.67980033e-01 9.07055885e-02 -1.47367612e-01 4.14698631e-01
-4.46701974e-01 1.34721743e-02 3.13306302e-01 -5.42687356e-01
-1.49002641e-01 -6.18728958e-02 -6.81960762e-01 7.16565311e-01
2.81360056e-02 1.10684860e+00 -1.07526040e+00 -1.05583683e-01
5.11484332e-02 5.42128623e-01 -2.62966454e-01 4.86587286e-02
-1.94215000e-01 5.79075336e-01 -5.44136614e-02 8.15540016e-01
7.69769013e-01 -3.75686139e-01 -2.23135009e-01 9.24985260e-02
-4.21789795e-01 2.97574490e-01 -1.01593697e+00 1.85094786e+00
-6.83100075e-02 9.33487773e-01 -1.50377095e-01 -5.98868787e-01
2.98661560e-01 1.27071247e-01 -5.65822385e-02 -7.58064628e-01
-1.90215372e-02 -8.67848992e-02 1.44335255e-01 -7.08500326e-01
3.45596164e-01 2.09973201e-01 3.64865631e-01 5.88110447e-01
5.39710186e-02 3.64304394e-01 7.22794756e-02 1.10733211e-01
1.08933926e+00 3.72621089e-01 1.52767107e-01 -1.61440685e-01
4.60308716e-02 -6.01388872e-01 5.12824595e-01 1.09610212e+00
-4.36898857e-01 6.53302491e-01 1.91847339e-01 -7.68803895e-01
-5.80697477e-01 -1.52924895e+00 -1.22907534e-01 1.48531175e+00
7.68610477e-01 1.27019897e-01 -4.52279270e-01 -3.93416941e-01
2.23153792e-02 7.22376227e-01 -1.02849150e+00 -2.36002430e-01
-5.19049168e-01 -1.56390458e-01 -2.89350629e-01 7.39841998e-01
3.38970542e-01 -1.45500720e+00 -1.33132946e+00 -1.57466307e-01
-7.56598935e-02 -9.38969374e-01 -4.24542099e-01 1.16414294e-01
-5.00786960e-01 -9.18031335e-01 -5.09923637e-01 -1.09016943e+00
9.93873954e-01 8.51648390e-01 1.10274744e+00 4.80863422e-01
-4.84131247e-01 6.54899597e-01 -1.25175834e-01 -6.16261780e-01
2.32296303e-01 -3.23646188e-01 4.25888032e-01 6.42187968e-02
3.06427330e-01 -6.33481562e-01 -8.54359210e-01 1.94905221e-01
-7.04167604e-01 1.44903585e-01 5.30616999e-01 3.15913379e-01
5.46875238e-01 2.81647027e-01 1.56332180e-01 -4.27210689e-01
7.68418834e-02 -3.50010037e-01 -5.88155746e-01 5.06690562e-01
1.02931447e-01 -2.40336984e-01 1.19105019e-01 -8.37469995e-01
-1.15905392e+00 1.65817916e-01 4.39619899e-01 -2.86387146e-01
-3.51626933e-01 1.64356187e-03 -3.82244051e-01 4.37280238e-02
6.34521484e-01 2.67728299e-01 -6.08370304e-01 -4.59256977e-01
4.49365854e-01 1.42669216e-01 7.51933515e-01 -4.40097600e-01
7.54534960e-01 9.02705610e-01 -3.22667092e-01 -7.62680173e-01
-1.14753997e+00 -4.15945768e-01 -1.05027175e+00 -3.28949183e-01
9.01383460e-01 -1.02027404e+00 -9.92803812e-01 6.37691140e-01
-1.43076408e+00 -4.76927608e-01 -3.03019546e-02 5.20307481e-01
-3.50860804e-01 -1.79364309e-01 -1.83485225e-01 -7.34579504e-01
3.53776127e-01 -8.26316893e-01 1.23630750e+00 4.93232369e-01
-1.78563427e-02 -8.60298157e-01 5.62039204e-02 6.97110221e-02
2.44652495e-01 1.45617828e-01 7.73105860e-01 -4.25746769e-01
-9.91353333e-01 -1.23828821e-01 -1.86687469e-01 -1.57078192e-01
3.06755692e-01 2.48100180e-02 -1.22812986e+00 -5.96120536e-01
2.60438383e-01 -2.14113325e-01 1.05674720e+00 3.98176461e-01
1.37425160e+00 -2.91899264e-01 -8.52631748e-01 5.22919595e-01
1.30998552e+00 1.29964992e-01 3.78589809e-01 2.25351408e-01
5.69125473e-01 5.87511599e-01 3.65705222e-01 -2.06967704e-02
3.76406461e-01 5.76731563e-01 8.57372403e-01 -7.95914978e-02
-3.33872765e-01 -3.93233448e-01 9.62603018e-02 3.76183800e-02
-1.73468947e-01 -4.02587950e-01 -9.08604860e-01 8.55140567e-01
-1.74886692e+00 -1.22568464e+00 -8.24434161e-02 2.20835590e+00
6.58349335e-01 1.83151841e-01 -8.52503479e-02 -2.31512845e-01
7.78699100e-01 1.15409501e-01 -9.05894816e-01 1.31147504e-02
-2.27385700e-01 -3.39447379e-01 -1.20759159e-01 2.85166681e-01
-1.23279667e+00 1.17589557e+00 6.76261282e+00 -1.22158989e-01
-1.11355197e+00 1.94046274e-01 1.61816403e-01 -2.38609865e-01
-9.30375978e-02 2.72154272e-01 -8.73405457e-01 5.50150394e-01
3.07466775e-01 1.65785089e-01 4.92512882e-01 9.42092776e-01
-1.61510602e-01 -5.41969180e-01 -1.87691331e+00 8.84568930e-01
3.04336220e-01 -7.35833883e-01 -1.17872365e-01 -1.43759996e-01
5.87988138e-01 -2.41126344e-01 6.87832296e-01 3.06305856e-01
3.24721038e-02 -1.02261508e+00 1.25944805e+00 5.70360184e-01
4.30328727e-01 1.06528938e-01 -9.49889235e-03 5.17666101e-01
-1.03226984e+00 -2.87851781e-01 -6.29853010e-01 -5.36325812e-01
4.64142151e-02 2.11121053e-01 -6.76006317e-01 -2.01162115e-01
1.06207132e+00 5.98998606e-01 -9.00803685e-01 1.39048767e+00
-6.20639026e-01 1.75755367e-01 1.91326849e-02 1.21095642e-01
-4.47850898e-02 1.53217778e-01 5.91796696e-01 6.56248093e-01
1.49526536e-01 2.47009844e-02 -2.36247078e-01 1.36565769e+00
-6.04606457e-02 -3.45377743e-01 -5.87504983e-01 2.37627894e-01
5.32999456e-01 1.13134706e+00 -9.85024512e-01 -7.63002783e-02
-4.53354597e-01 1.31701601e+00 7.72203386e-01 6.67622864e-01
-9.88266289e-01 -1.23372726e-01 5.27277231e-01 2.63142914e-01
5.44652700e-01 -3.66294086e-01 1.00081593e-01 -1.07501733e+00
5.06335795e-01 -2.92051584e-01 1.18852101e-01 -1.41471505e+00
-1.08100152e+00 4.67487782e-01 -7.95771927e-02 -1.01641071e+00
2.71345586e-01 -8.51038694e-01 -6.14623904e-01 6.76012576e-01
-1.34007204e+00 -1.42257154e+00 -5.58607578e-01 3.97872299e-01
5.85066259e-01 2.01688364e-01 5.71262836e-01 -5.31047164e-03
-4.09168571e-01 3.16585481e-01 -2.21327275e-01 -1.03980526e-01
5.16789377e-01 -1.27976370e+00 4.37836766e-01 1.14814615e+00
7.92795658e-01 1.20307922e+00 6.39977157e-01 -7.91295648e-01
-1.28928888e+00 -7.35374987e-01 9.06924069e-01 -1.34587431e+00
4.58556116e-01 -1.09641469e+00 -1.09030962e+00 1.30811989e+00
3.66755910e-02 4.28410113e-01 3.57281506e-01 4.45668399e-01
-6.36256456e-01 -1.45290673e-01 -8.22509646e-01 7.25014150e-01
1.40297139e+00 -5.93362272e-01 -1.03120482e+00 4.14829284e-01
8.67606878e-01 -5.89230478e-01 2.33619604e-02 3.57735381e-02
8.05944741e-01 -1.20748043e+00 1.00008106e+00 -9.87266481e-01
1.83170423e-01 -5.16734540e-01 1.40594780e-01 -9.97908533e-01
-4.96982634e-01 -4.65326101e-01 -4.21325058e-01 1.00834930e+00
2.17148110e-01 -5.42364478e-01 3.33368361e-01 7.13769138e-01
-8.63703862e-02 -3.98885012e-01 -1.10499907e+00 -8.41188192e-01
-1.36275664e-01 -1.39134750e-01 6.70869589e-01 8.38108480e-01
-2.17563316e-01 2.70545840e-01 -2.91507006e-01 7.09706545e-01
3.29507262e-01 4.30072725e-01 4.65076625e-01 -1.28375006e+00
-2.35109162e-02 -2.76439786e-01 -5.70137560e-01 -1.21814632e+00
1.71401620e-01 -6.64446115e-01 2.43276179e-01 -1.65703189e+00
5.57885766e-01 -2.08054647e-01 -4.93894845e-01 8.16186905e-01
-2.36492798e-01 9.14771184e-02 7.13471398e-02 2.85600185e-01
-1.08190465e+00 2.80798078e-01 1.21814358e+00 9.26807430e-03
-1.77800477e-01 6.59369826e-02 -7.76983202e-01 8.58285010e-01
4.31669652e-01 -2.75545597e-01 -5.86374819e-01 -1.13304758e+00
5.17450094e-01 -3.06486398e-01 9.33470905e-01 -1.21702552e+00
3.65557283e-01 -1.15964979e-01 9.68191206e-01 -7.83879876e-01
5.62996745e-01 -1.04433692e+00 3.14452052e-02 3.90559971e-01
-3.45053911e-01 7.55942911e-02 2.65147924e-01 6.43710554e-01
3.50532383e-01 -1.68483064e-01 5.29605865e-01 -1.61964148e-01
-1.13051021e+00 3.42606664e-01 3.05446591e-02 -8.12639520e-02
1.38914776e+00 -2.56863654e-01 -7.81615078e-01 -8.47583339e-02
-1.05515039e+00 1.13018192e-01 4.82716709e-01 8.71762037e-01
8.37407410e-01 -1.22741723e+00 -4.17334676e-01 2.22187355e-01
1.52260661e-01 -9.76339653e-02 3.92384857e-01 5.42459607e-01
-2.53414661e-01 3.89354289e-01 -2.33837187e-01 -8.14528167e-01
-1.10875916e+00 1.09061503e+00 2.86295742e-01 7.49672472e-01
-7.35720873e-01 1.39636683e+00 7.73724020e-01 1.16661496e-01
4.33301926e-01 -5.52922666e-01 -2.35164240e-01 -1.37554571e-01
8.69034111e-01 2.32171658e-02 -1.68291286e-01 -6.80619836e-01
-5.37957549e-01 5.09346008e-01 -2.90968865e-01 5.86759523e-02
1.35623193e+00 -3.39718193e-01 -3.28447610e-01 9.99495268e-01
8.29474688e-01 -2.66895257e-02 -1.61743617e+00 -3.30439746e-01
-1.19729765e-01 -8.89186442e-01 -1.90581039e-01 -9.59895670e-01
-6.89157605e-01 9.49643195e-01 6.13600016e-01 1.85030058e-01
9.26874876e-01 6.70483351e-01 2.19680578e-01 3.50687087e-01
5.45742691e-01 -7.16302514e-01 5.99497974e-01 2.62422591e-01
1.15620100e+00 -1.27348304e+00 5.71736656e-02 -1.10518567e-01
-4.81020749e-01 6.61252081e-01 1.11223924e+00 1.76828802e-01
4.14618939e-01 -1.80537298e-01 -1.83217868e-01 -3.90326768e-01
-7.79184580e-01 -2.87860006e-01 5.26559949e-01 7.11201131e-01
1.65656537e-01 -1.29837826e-01 3.52685869e-01 5.87803006e-01
-2.38487795e-01 -5.62434793e-01 2.66094357e-01 1.07730508e+00
-5.47049403e-01 -4.23631877e-01 -1.40584812e-01 9.52231959e-02
-2.52788931e-01 -1.15848452e-01 -4.91477817e-01 9.03637409e-01
5.52041531e-01 6.47829592e-01 4.42150056e-01 2.00967535e-01
5.99898808e-02 -5.15401810e-02 8.70523810e-01 -1.03038549e+00
1.07552884e-02 -2.79718786e-01 -7.40299284e-01 -5.70551336e-01
-4.49699342e-01 -7.15923667e-01 -1.11954439e+00 3.40139568e-01
-3.66205812e-01 -2.76695758e-01 7.61253655e-01 8.49783838e-01
1.99831903e-01 5.55679142e-01 2.86228150e-01 -7.51631260e-01
-1.37932763e-01 -6.65391684e-01 -6.85208619e-01 2.40546629e-01
9.85701501e-01 -1.05212867e+00 -5.08600414e-01 2.41691843e-01] | [9.95349407196045, 1.1168246269226074] |
e3bf2203-381c-457d-b5a2-9af80a74524c | a-fast-geometric-regularizer-to-mitigate | 2212.07350 | null | https://arxiv.org/abs/2212.07350v1 | https://arxiv.org/pdf/2212.07350v1.pdf | A Fast Geometric Regularizer to Mitigate Event Collapse in the Contrast Maximization Framework | Event cameras are emerging vision sensors and their advantages are suitable for various applications such as autonomous robots. Contrast maximization (CMax), which provides state-of-the-art accuracy on motion estimation using events, may suffer from an overfitting problem called event collapse. Prior works are computationally expensive or cannot alleviate the overfitting, which undermines the benefits of the CMax framework. We propose a novel, computationally efficient regularizer based on geometric principles to mitigate event collapse. The experiments show that the proposed regularizer achieves state-of-the-art accuracy results, while its reduced computational complexity makes it two to four times faster than previous approaches. To the best of our knowledge, our regularizer is the only effective solution for event collapse without trading off runtime. We hope our work opens the door for future applications that unlocks the advantages of event cameras. | ['Guillermo Gallego', 'Yoshimitsu Aoki', 'Shintaro Shiba'] | 2022-12-14 | null | null | null | null | ['event-based-vision', 'event-based-motion-estimation'] | ['computer-vision', 'computer-vision'] | [ 1.74056828e-01 -1.78035423e-01 -1.39009342e-01 -1.00124285e-01
-5.46845615e-01 -2.42256522e-01 6.33481562e-01 1.12402383e-02
-7.45593846e-01 4.15460587e-01 -1.79273576e-01 -1.37649268e-01
5.60164414e-02 -7.06350029e-01 -7.10813224e-01 -7.61512637e-01
-9.45025310e-02 -3.87221910e-02 8.48102987e-01 1.15514621e-01
3.45304847e-01 5.04430056e-01 -1.81820571e+00 -2.80267864e-01
7.69755781e-01 1.05088174e+00 5.35351634e-01 4.27853495e-01
1.56117857e-01 7.45712340e-01 -3.42308462e-01 -3.20958585e-01
3.68652970e-01 -1.93799496e-01 -3.13009351e-01 1.50340810e-01
3.20914656e-01 -5.40758252e-01 -6.17687523e-01 1.06708562e+00
3.00503463e-01 1.11948192e-01 2.58254379e-01 -1.39210820e+00
-3.44328165e-01 1.24336518e-01 -9.79270518e-01 2.16883361e-01
2.04410225e-01 2.06704199e-01 8.44720125e-01 -9.37304378e-01
5.76077104e-01 1.08000135e+00 7.12379873e-01 3.23479921e-01
-7.81803966e-01 -4.66459870e-01 3.93213600e-01 4.31340575e-01
-1.37894094e+00 -3.61185253e-01 6.83082640e-01 -2.20857516e-01
9.41392422e-01 -2.77180970e-02 6.42012119e-01 9.58490074e-01
2.44445100e-01 1.07497799e+00 8.42766583e-01 -1.71571746e-01
3.20277274e-01 -3.00040990e-01 -6.94690645e-02 7.32938230e-01
6.36025488e-01 1.97232857e-01 -5.44356346e-01 -3.14391218e-03
9.50687289e-01 2.13601574e-01 -3.17676008e-01 -5.30599236e-01
-1.37927425e+00 8.33786368e-01 3.73572558e-01 -1.28961504e-01
-3.94616753e-01 3.22006583e-01 2.71723449e-01 -1.12360798e-01
2.26101160e-01 1.42819971e-01 -2.97499150e-01 -3.57031405e-01
-7.90523708e-01 2.34333858e-01 5.51128149e-01 9.37868595e-01
6.32639170e-01 1.20304219e-01 2.42483839e-01 5.05567014e-01
4.18606669e-01 6.26825571e-01 3.52456599e-01 -1.12676072e+00
2.25925773e-01 5.61507344e-01 3.13770682e-01 -8.86884928e-01
-4.57152694e-01 -2.59614199e-01 -6.81520998e-01 4.67752844e-01
5.31685710e-01 -6.28141977e-04 -8.01001430e-01 1.65294969e+00
4.20981437e-01 4.34256881e-01 2.05011312e-02 1.12345803e+00
6.09696329e-01 6.94370687e-01 -8.93793032e-02 -4.84787256e-01
1.29226327e+00 -1.08567083e+00 -6.47627115e-01 -4.13056523e-01
1.06859691e-01 -7.65632033e-01 9.77857888e-01 4.41731453e-01
-1.08232141e+00 -2.13050276e-01 -1.38426554e+00 7.25747719e-02
3.09793651e-02 3.70043181e-02 1.07229733e+00 3.11600566e-01
-6.51364982e-01 4.35844988e-01 -1.38025022e+00 -6.68316782e-01
3.73358607e-01 2.90420830e-01 -1.90450728e-01 -4.19819988e-02
-6.23614967e-01 9.82321739e-01 1.61502615e-01 1.28749590e-02
-8.06066632e-01 -4.48890597e-01 -1.12241101e+00 -2.36125857e-01
7.42743850e-01 -5.41371584e-01 1.44969082e+00 -5.07943571e-01
-1.64513087e+00 6.48768246e-01 -3.67030859e-01 -7.25463927e-01
5.87637246e-01 -4.12531167e-01 -9.20415521e-02 2.01433212e-01
5.55778854e-02 6.47965670e-01 7.94141710e-01 -9.48700666e-01
-7.49485075e-01 -2.71570086e-01 -1.42838946e-02 1.36900783e-01
-4.03141886e-01 -1.86799057e-02 -7.86515594e-01 -5.54333925e-01
2.94340730e-01 -9.98016953e-01 -5.76595306e-01 1.74927309e-01
-3.14444080e-02 -8.35345909e-02 9.46476519e-01 -1.43816233e-01
1.14180720e+00 -2.04149842e+00 -1.49543896e-01 -3.17767292e-01
4.36062552e-02 2.12343708e-01 1.30661875e-01 2.42462426e-01
3.71712476e-01 -4.38501179e-01 -3.07829916e-01 -5.16103387e-01
-2.66835719e-01 1.84089422e-01 -2.46625513e-01 8.45669329e-01
2.44570166e-01 6.45797968e-01 -1.06500459e+00 -3.99867862e-01
8.35310757e-01 5.42698681e-01 -2.52854973e-01 4.83309254e-02
-2.37166911e-01 3.56669098e-01 -3.47215116e-01 5.19960284e-01
7.69283414e-01 -4.72346634e-01 -7.46997297e-02 -2.49374703e-01
-4.62314129e-01 1.62516057e-01 -1.49783313e+00 1.71553612e+00
5.96282538e-03 8.00417483e-01 -5.05287983e-02 -8.41335058e-01
7.59909213e-01 -4.84758951e-02 6.65895998e-01 -5.23760378e-01
1.96402624e-01 5.39415777e-02 -1.93730235e-01 -4.76448059e-01
7.34674215e-01 1.14784967e-02 -1.30179296e-05 -3.52197737e-02
-1.45936340e-01 1.22490071e-01 2.23110512e-01 6.00612164e-02
1.09378505e+00 4.72307682e-01 7.94132113e-01 -9.96840671e-02
2.21537426e-01 1.34360194e-01 8.15562785e-01 7.60928214e-01
-3.79710555e-01 8.37720752e-01 2.72299126e-02 -5.50445020e-01
-9.14091766e-01 -1.16160834e+00 -9.25061852e-02 6.26086831e-01
6.83120489e-01 -3.36199760e-01 -4.48903471e-01 -3.83107007e-01
-1.74952760e-01 2.47013703e-01 -1.69991955e-01 8.84137154e-02
-7.07788706e-01 -9.52890098e-01 4.29873794e-01 8.22788656e-01
7.15889812e-01 -7.03345299e-01 -1.25891697e+00 3.17049861e-01
-2.60096371e-01 -1.75004375e+00 -3.20869714e-01 -9.77452658e-03
-1.05950773e+00 -1.09631228e+00 -7.16808200e-01 -5.44108212e-01
6.95447683e-01 8.18963885e-01 7.51269937e-01 -1.14011332e-01
-2.85533756e-01 3.05644423e-01 -3.06726038e-01 -4.77766931e-01
4.94296327e-02 -2.49716118e-01 2.11498499e-01 -6.33810684e-02
2.83464164e-01 -5.37463784e-01 -8.79831672e-01 3.84097070e-01
-9.42487597e-01 6.20849468e-02 6.94722891e-01 4.44211930e-01
8.12058985e-01 1.73461407e-01 3.09343755e-01 -5.19359767e-01
-7.47695193e-02 -4.08771485e-01 -1.07569194e+00 -1.43130302e-01
-6.33566082e-01 1.22977249e-01 4.81597066e-01 -6.25429392e-01
-1.22028613e+00 5.06025076e-01 2.15422884e-01 -6.09733522e-01
1.98100898e-02 3.75534326e-01 4.12113778e-02 -2.62073129e-01
4.99823421e-01 6.87113106e-02 -1.61400840e-01 -2.42546409e-01
3.26611489e-01 2.96863168e-01 9.26770151e-01 -2.45617419e-01
5.16038597e-01 1.16998816e+00 3.06690097e-01 -9.50571477e-01
-7.80582130e-01 -7.93698549e-01 -9.23104584e-02 -2.11962089e-01
8.15871477e-01 -1.08479106e+00 -8.05268824e-01 6.88660324e-01
-1.20734215e+00 2.42736563e-03 -1.55078381e-01 7.71008074e-01
-5.29013515e-01 8.03191066e-01 -5.59409678e-01 -1.02919734e+00
-1.71517551e-01 -1.12018251e+00 1.04118013e+00 5.79450190e-01
1.18114889e-01 -6.09711945e-01 -7.16391653e-02 1.94998592e-01
1.80528894e-01 2.96301126e-01 -2.18936559e-02 -1.10429987e-01
-9.31379914e-01 -1.49144605e-01 -2.89165616e-01 2.05881298e-02
-4.52130288e-02 1.04511835e-01 -9.56305802e-01 -3.39736611e-01
2.57659614e-01 8.90646502e-02 1.07696676e+00 6.87882066e-01
9.57984686e-01 1.16713196e-01 -3.79391193e-01 5.59480190e-01
1.51631689e+00 2.75355220e-01 9.08316016e-01 5.93485057e-01
6.06593728e-01 2.80625850e-01 9.25353110e-01 8.52213681e-01
7.46416330e-01 8.59856308e-01 8.42481315e-01 4.70590144e-02
-9.43231732e-02 -1.12990297e-01 6.60530865e-01 5.05478978e-01
-1.03248596e-01 -4.12743270e-01 -7.08814383e-01 9.03343320e-01
-2.30846667e+00 -8.97163272e-01 -3.57556164e-01 2.36412311e+00
3.92904907e-01 2.38332719e-01 4.41162065e-02 -2.58302409e-02
7.31274188e-01 3.13156217e-01 -6.45080328e-01 5.69689870e-02
-1.27366394e-01 -1.48536667e-01 7.80836344e-01 2.98074841e-01
-1.28893161e+00 8.82178783e-01 6.37803555e+00 6.40897870e-01
-1.06368101e+00 -1.49888843e-02 9.01557580e-02 -1.75454244e-01
1.15488030e-01 2.53850788e-01 -9.86503124e-01 5.27482629e-01
5.78992724e-01 -8.96081887e-03 2.18299955e-01 9.78572011e-01
3.71491492e-01 -4.93137151e-01 -1.01412773e+00 1.22016943e+00
2.89873719e-01 -1.24488711e+00 -2.26928249e-01 1.25791028e-01
7.55247414e-01 3.62033814e-01 -1.29929736e-01 -1.10817499e-01
2.20519334e-01 -6.22997940e-01 8.30099285e-01 3.21416974e-01
3.32614779e-01 -6.57669485e-01 7.45993137e-01 2.90027589e-01
-1.46794128e+00 1.08763091e-01 -4.33425188e-01 -3.86875570e-01
6.82105720e-01 9.04415786e-01 -6.01248205e-01 4.62608218e-01
7.91427851e-01 7.56928384e-01 -4.38394547e-01 1.32088351e+00
-1.94491282e-01 3.33057910e-01 -6.86764359e-01 1.12555996e-02
1.42086297e-01 -2.90292706e-02 9.33784425e-01 1.06679201e+00
4.95880127e-01 6.06927909e-02 3.38068753e-01 6.41851008e-01
3.73463482e-02 -3.62544984e-01 -7.48097479e-01 6.75948337e-02
5.88347793e-01 1.13114309e+00 -9.64133561e-01 -1.94428325e-01
-8.12999725e-01 1.02368307e+00 2.14172497e-01 1.60569981e-01
-9.36387420e-01 -2.94248074e-01 5.37544847e-01 -6.28236681e-02
6.87815428e-01 -5.88450134e-01 -5.56424737e-01 -1.39425206e+00
4.44780111e-01 -4.07961309e-01 4.13735241e-01 -5.73001862e-01
-1.18411922e+00 2.11121976e-01 8.78026187e-02 -1.58590579e+00
-2.59466350e-01 -5.61964869e-01 -5.52230299e-01 -5.22718579e-02
-1.66307640e+00 -1.02445328e+00 -6.22899055e-01 4.57395345e-01
7.73707807e-01 1.17527738e-01 3.83599073e-01 2.28586361e-01
-5.07015765e-01 2.80138344e-01 -4.19481210e-02 -1.85014933e-01
6.20579004e-01 -9.84088719e-01 3.05242866e-01 1.35731316e+00
5.05699143e-02 3.62365633e-01 8.69727612e-01 -5.62517643e-01
-1.57176852e+00 -9.58805144e-01 5.64029038e-01 -3.08572948e-01
6.99078858e-01 4.32616174e-02 -6.59113169e-01 6.28492832e-01
9.23724622e-02 1.34746328e-01 2.71805078e-01 -2.61420667e-01
-1.49226829e-01 -1.25183299e-01 -9.92503941e-01 7.25302696e-01
1.04820645e+00 -1.06374763e-01 -5.69161296e-01 1.80908013e-02
6.31775200e-01 -4.73455787e-01 -6.25558913e-01 6.73790634e-01
5.98136008e-01 -1.01607037e+00 9.91384506e-01 2.76229233e-01
2.05505237e-01 -7.00817108e-01 -1.76932156e-01 -8.00522089e-01
4.83770593e-04 -7.36599147e-01 -5.86821616e-01 1.11355138e+00
-1.31155699e-02 -6.28768623e-01 7.09232152e-01 3.82186979e-01
-1.14075698e-01 -6.27002776e-01 -1.05461502e+00 -1.13534439e+00
-4.39294815e-01 -7.86189437e-01 1.74955800e-01 7.36610770e-01
-1.08158156e-01 2.58940428e-01 -5.08322775e-01 4.53039587e-01
1.03199172e+00 8.83908123e-02 7.87005723e-01 -1.06421733e+00
-2.69890368e-01 -3.06649148e-01 -6.51796997e-01 -1.40777767e+00
-7.55500346e-02 -1.57864586e-01 2.57286072e-01 -1.46680558e+00
4.17899609e-01 -1.10558219e-01 1.49340350e-02 2.01084152e-01
-2.17632979e-01 3.17569345e-01 2.38112628e-01 2.56416202e-01
-9.70371485e-01 6.64132059e-01 9.39139366e-01 1.12593107e-01
-1.97423235e-01 -1.62076250e-01 -6.08301342e-01 1.29068148e+00
6.88987792e-01 -3.89378369e-01 -2.94775605e-01 -5.92877388e-01
2.59929411e-02 -9.18942243e-02 5.71983099e-01 -1.07792985e+00
6.42802119e-01 -3.28377664e-01 1.12138204e-01 -7.84236491e-01
4.47368771e-01 -8.13892007e-01 2.06588566e-01 3.10472995e-01
2.39057273e-01 6.02093861e-02 1.85503438e-01 8.01301479e-01
-2.25642100e-01 -2.13549227e-01 1.00916767e+00 -1.18173622e-01
-1.01842165e+00 2.17909813e-01 -5.36942780e-01 -1.04433775e-01
1.25608063e+00 -3.85455102e-01 -3.62692505e-01 -4.60103989e-01
-9.28069353e-02 3.47582370e-01 7.86542177e-01 3.97126377e-01
7.80838192e-01 -1.29335678e+00 -5.09659410e-01 -8.49293098e-02
1.44942760e-01 3.12224567e-01 8.56019631e-02 9.75548863e-01
-5.17368734e-01 2.70134985e-01 1.87349856e-01 -9.33685720e-01
-1.19352126e+00 5.61264277e-01 -2.77165398e-02 -2.98995852e-01
-9.97515559e-01 5.66417992e-01 4.19766188e-01 2.34345142e-02
3.09899092e-01 -3.65681171e-01 9.99630094e-02 -2.30106801e-01
6.17691696e-01 8.48192155e-01 -1.27715454e-01 -4.52395856e-01
-6.37599766e-01 7.55441129e-01 -1.78946108e-01 -1.96721181e-01
1.13563514e+00 -3.98348987e-01 2.19289258e-01 3.06984216e-01
8.55548143e-01 -8.16225186e-02 -1.68333995e+00 -7.75199980e-02
-1.47819802e-01 -5.56295395e-01 1.28281131e-01 -2.00473502e-01
-1.06758249e+00 6.33314013e-01 5.31061411e-01 -3.90712470e-02
1.27772367e+00 6.35220259e-02 1.14774323e+00 4.79193717e-01
7.29612112e-01 -1.00735176e+00 9.40893963e-02 4.97786760e-01
4.98983741e-01 -1.51842868e+00 2.88462341e-01 -5.94889402e-01
-5.84048808e-01 8.93832326e-01 6.78223372e-01 -3.84303242e-01
3.39684486e-01 4.76297647e-01 -1.60075799e-01 -1.16071381e-01
-6.16427302e-01 -5.07363498e-01 -3.89966853e-02 5.19467056e-01
-3.33246514e-02 -5.40222377e-02 -5.60423553e-01 4.21767563e-01
1.64137706e-01 1.06119186e-01 6.20544672e-01 1.09956121e+00
-5.15029013e-01 -8.62266958e-01 -4.08120394e-01 1.30709589e-01
-5.26005805e-01 6.98030367e-02 -4.52804714e-02 9.22831476e-01
9.61147714e-03 1.15734673e+00 9.80343372e-02 -2.53688276e-01
4.19840693e-01 -4.44316089e-01 6.24527216e-01 -1.38211414e-01
-3.82259414e-02 1.71079621e-01 -7.48356581e-02 -8.90032828e-01
-8.02663088e-01 -7.62996256e-01 -1.65655732e+00 -3.54070663e-01
-5.18688500e-01 -4.04646367e-01 4.76248324e-01 9.39338267e-01
4.94690835e-01 2.16852173e-01 4.64059561e-01 -9.44968700e-01
-3.88726413e-01 -6.86735928e-01 -1.50640979e-01 3.30357939e-01
3.74275506e-01 -9.91411507e-01 -3.47006351e-01 1.14390023e-01] | [8.450549125671387, -1.3897593021392822] |
d5599aea-f129-4662-8324-9efca252ea37 | eeg-aided-boosting-of-single-lead-ecg-based | 2211.13125 | null | https://arxiv.org/abs/2211.13125v1 | https://arxiv.org/pdf/2211.13125v1.pdf | EEG aided boosting of single-lead ECG based sleep staging with Deep Knowledge Distillation | An electroencephalogram (EEG) signal is currently accepted as a standard for automatic sleep staging. Lately, Near-human accuracy in automated sleep staging has been achievable by Deep Learning (DL) based approaches, enabling multi-fold progress in this area. However, An extensive and expensive clinical setup is required for EEG based sleep staging. Additionally, the EEG setup being obtrusive in nature and requiring an expert for setup adds to the inconvenience of the subject under study, making it adverse in the point of care setting. An unobtrusive and more suitable alternative to EEG is Electrocardiogram (ECG). Unsurprisingly, compared to EEG in sleep staging, its performance remains sub-par. In order to take advantage of both the modalities, transferring knowledge from EEG to ECG is a reasonable approach, ultimately boosting the performance of ECG based sleep staging. Knowledge Distillation (KD) is a promising notion in DL that shares knowledge from a superior performing but usually more complex teacher model to an inferior but compact student model. Building upon this concept, a cross-modality KD framework assisting features learned through models trained on EEG to improve ECG-based sleep staging performance is proposed. Additionally, to better understand the distillation approach, extensive experimentation on the independent modules of the proposed model was conducted. Montreal Archive of Sleep Studies (MASS) dataset consisting of 200 subjects was utilized for this study. The results from the proposed model for weighted-F1-score in 3-class and 4-class sleep staging showed a 13.40 \% and 14.30 \% improvement, respectively. This study demonstrates the feasibility of KD for single-channel ECG based sleep staging's performance enhancement in 3-class (W-R-N) and 4-class (W-R-L-D) classification. | ['Mohanasankar Sivaprakasam', 'Preejith SP', 'Sricharan V', 'Vaibhav Joshi'] | 2022-11-18 | null | null | null | null | ['sleep-staging', 'eeg-based-sleep-staging', 'ecg-based-sleep-staging'] | ['medical', 'time-series', 'time-series'] | [ 6.92365617e-02 6.20853491e-02 -9.49248746e-02 -4.70820218e-01
-5.95150471e-01 -1.49068728e-01 7.24711791e-02 2.28391156e-01
-6.95630252e-01 1.05060828e+00 2.74886768e-02 -2.14696720e-01
-3.60712141e-01 -4.00181770e-01 -1.78772807e-01 -8.24831486e-01
-1.13582335e-01 3.05058211e-01 -6.15189783e-02 -8.39123949e-02
2.13088080e-01 2.78596759e-01 -1.37739658e+00 3.97174135e-02
1.15259135e+00 1.15390003e+00 3.27844709e-01 5.15771091e-01
7.35311806e-02 2.36004695e-01 -8.11842442e-01 -1.93415120e-01
1.30392089e-01 -6.57637954e-01 -8.80651474e-01 -1.63670078e-01
4.53521200e-02 1.18277222e-01 -1.73928235e-02 6.99771762e-01
8.40786636e-01 1.64585337e-01 5.43122649e-01 -1.14051831e+00
-2.01290682e-01 2.07028434e-01 -4.16299492e-01 7.46046305e-01
5.08630812e-01 4.94990638e-03 6.12213135e-01 -4.59958076e-01
-1.87159628e-01 2.08444998e-01 9.66631651e-01 6.45075738e-01
-1.09740841e+00 -9.19252813e-01 -3.85160357e-01 4.85898405e-01
-1.63953316e+00 -2.52008766e-01 6.11272037e-01 -2.04769850e-01
1.02504456e+00 4.10753459e-01 1.42581618e+00 8.68157446e-01
6.23055577e-01 4.15431827e-01 1.70594335e+00 -2.79509455e-01
3.80277723e-01 6.05731785e-01 2.36922517e-01 5.24905264e-01
2.58879423e-01 -1.17563345e-01 -9.19493914e-01 1.18619695e-01
3.26889277e-01 2.87899345e-01 -3.36485028e-01 1.65307388e-01
-7.13879585e-01 4.68646169e-01 5.09961367e-01 5.62265813e-01
-3.41199666e-01 -2.41547376e-01 3.33569527e-01 3.58996481e-01
4.45045918e-01 6.24926031e-01 -5.34310222e-01 -6.20293319e-01
-1.71121967e+00 -3.04996341e-01 7.23097086e-01 5.18918395e-01
5.20839095e-01 -4.87238765e-02 -9.16296337e-03 6.36300206e-01
2.39974067e-01 2.98998863e-01 1.02643704e+00 -4.59348112e-01
1.16147935e-01 8.18597555e-01 -2.32732907e-01 -6.04966044e-01
-9.55842495e-01 -8.33794415e-01 -1.05718052e+00 -8.48448090e-03
6.78563341e-02 3.54806036e-02 -8.00704420e-01 1.43961990e+00
2.04066914e-02 4.17060763e-01 -3.83180603e-02 8.33591938e-01
8.97024751e-01 2.75581390e-01 2.51931787e-01 -4.49869454e-01
1.64648211e+00 -5.50255120e-01 -5.08112073e-01 -1.36958718e-01
2.61076719e-01 -5.15002906e-01 1.06192780e+00 7.91103363e-01
-1.00129843e+00 -7.35653758e-01 -1.23694599e+00 9.30730924e-02
-3.43959838e-01 2.55310476e-01 5.42252183e-01 1.03824782e+00
-1.25220668e+00 6.28069639e-01 -1.10408998e+00 -7.82461107e-01
6.61806822e-01 1.00379515e+00 -4.38469380e-01 3.36849153e-01
-1.06589186e+00 1.11436737e+00 2.80708909e-01 1.04452357e-01
-6.84793711e-01 -8.16614330e-01 -5.96079588e-01 8.48399103e-02
-9.18510631e-02 -7.84563124e-01 8.54675710e-01 -7.29051888e-01
-1.57810104e+00 7.98033357e-01 -2.37209737e-01 -5.22347391e-01
2.69572556e-01 -1.48728922e-01 -6.01209402e-01 3.35609257e-01
5.54350577e-02 5.27992070e-01 8.48656654e-01 -7.38904357e-01
-5.52769721e-01 -5.19571841e-01 6.30523786e-02 3.54028255e-01
-4.31977153e-01 -1.82625085e-01 9.83142629e-02 -3.81139666e-01
-3.62272039e-02 -8.92390728e-01 9.84554514e-02 -2.94023722e-01
-1.41410306e-01 -3.01913530e-01 4.90447640e-01 -7.56930709e-01
1.39904749e+00 -2.17146373e+00 -7.25770965e-02 7.59685040e-02
5.10208070e-01 2.48864964e-01 5.32146931e-01 2.11456463e-01
-2.01773852e-01 -1.11955144e-01 -2.26632282e-01 -4.24233794e-01
-1.68042839e-01 2.64660418e-01 3.95104378e-01 6.14573240e-01
2.79409904e-02 6.47959292e-01 -6.25782609e-01 -4.38538969e-01
2.25609794e-01 4.57707435e-01 -4.08843577e-01 1.92908436e-01
8.33275199e-01 7.87366152e-01 -8.10242444e-02 5.37235975e-01
3.80008161e-01 -2.13422239e-01 -1.90869704e-01 -2.50605524e-01
-4.23347354e-02 1.93703502e-01 -8.07483852e-01 1.90600884e+00
-6.29272878e-01 6.11710310e-01 -2.71751851e-01 -1.13426876e+00
7.51383483e-01 5.45009375e-01 3.61353010e-01 -7.50746667e-01
2.50045776e-01 1.82789505e-01 2.78067231e-01 -5.58098435e-01
2.87004381e-01 -7.55942881e-01 2.23625302e-02 2.80221969e-01
3.89223248e-01 -1.22017220e-01 -2.39779018e-02 -4.75178137e-02
1.06919467e+00 -5.28477393e-02 6.60795629e-01 -6.13634646e-01
6.13939941e-01 -2.61143386e-01 5.86193442e-01 5.17615199e-01
-3.60504806e-01 4.65060681e-01 3.49344015e-02 -3.54095131e-01
-4.20356065e-01 -9.98699248e-01 -5.19898832e-01 6.05620027e-01
1.39188483e-01 -5.46751022e-01 -6.77794755e-01 -6.10829592e-01
-3.96803081e-01 6.37271345e-01 -6.34319961e-01 -4.95875090e-01
-1.66384459e-01 -1.11727905e+00 7.20830739e-01 7.34708607e-01
6.89628839e-01 -8.95588040e-01 -1.05047822e+00 1.47358580e-02
-6.42897561e-02 -6.25364065e-01 -1.47911564e-01 7.92415559e-01
-1.06339514e+00 -9.74652886e-01 -8.68186414e-01 -5.14390707e-01
6.47410989e-01 -3.69294696e-02 1.02131736e+00 8.80820528e-02
-4.90200192e-01 4.45181161e-01 -3.23227704e-01 -7.27210581e-01
5.90961203e-02 2.04540104e-01 4.87667292e-01 -1.35063857e-01
5.69246948e-01 -9.32343960e-01 -1.03675008e+00 1.63328528e-01
-5.93351960e-01 -1.02968335e-01 8.49534452e-01 7.90017664e-01
3.85915041e-01 8.24507400e-02 7.58156300e-01 -4.59571809e-01
6.76715374e-01 -5.47026575e-01 -3.61700952e-02 4.73425537e-02
-1.03973567e+00 -2.75527298e-01 7.36630917e-01 -2.40795434e-01
-8.08263600e-01 -2.12710142e-01 -3.08580846e-01 -7.21030310e-02
-3.88134390e-01 3.23847115e-01 9.76996198e-02 -6.65969253e-02
5.49084902e-01 4.74721074e-01 -4.57166508e-02 -3.61887366e-01
-3.51547599e-01 8.88180554e-01 4.40993339e-01 -2.48019874e-01
5.74730575e-01 2.75550127e-01 4.08531278e-02 -1.09613180e+00
-7.66342521e-01 -6.37812018e-01 -7.79083908e-01 -8.98466036e-02
1.15009534e+00 -8.81021202e-01 -6.91573322e-01 3.24995756e-01
-4.40641612e-01 -2.57102042e-01 -3.58355969e-01 7.04200149e-01
-3.13871711e-01 2.54610419e-01 -4.39978130e-02 -7.44839966e-01
-7.57223427e-01 -1.15780020e+00 8.35401773e-01 7.14155257e-01
-6.92921042e-01 -1.05659437e+00 6.94624558e-02 5.61915219e-01
5.44735372e-01 -1.10287450e-01 9.01449502e-01 -8.63671422e-01
3.96055840e-02 -2.79456466e-01 3.55560482e-02 4.84110057e-01
2.57482797e-01 -7.21220672e-01 -1.21311569e+00 -2.97259122e-01
4.66980010e-01 -2.47943461e-01 3.22782487e-01 3.07636827e-01
9.63252187e-01 2.50825703e-01 -1.36583939e-01 6.91974580e-01
1.31524217e+00 3.91993195e-01 5.83464861e-01 4.54515457e-01
2.81466633e-01 3.73040587e-02 2.51864105e-01 4.35365081e-01
6.82757080e-01 5.08068204e-01 2.85883807e-02 -2.66872197e-01
-9.69389454e-02 2.50209033e-01 3.60288382e-01 9.98171091e-01
-3.16432089e-01 2.21541360e-01 -7.74241626e-01 2.71675050e-01
-1.21246529e+00 -7.52410889e-01 3.09650600e-02 2.10578132e+00
7.94966280e-01 1.87885568e-01 3.04929852e-01 6.90864444e-01
1.19048856e-01 -3.61815006e-01 -4.40293729e-01 -6.46423578e-01
3.09744209e-01 9.13139939e-01 1.54689461e-01 -1.99513242e-01
-7.04091132e-01 2.48197809e-01 5.65102863e+00 6.47634923e-01
-1.29369402e+00 4.68609661e-01 4.33478087e-01 -3.38775545e-01
3.66780847e-01 -2.72869974e-01 -6.79006100e-01 8.64168763e-01
1.41593289e+00 -1.52506456e-01 3.58632356e-01 6.10666990e-01
4.89192635e-01 -7.40700841e-01 -1.18710220e+00 1.35553288e+00
2.68964171e-01 -9.55062389e-01 -4.71759051e-01 -6.04632311e-02
3.18170547e-01 -5.04868962e-02 -8.89770463e-02 5.95195770e-01
-7.55991638e-01 -1.11855233e+00 4.62594092e-01 6.20633066e-01
1.04910100e+00 -7.15383947e-01 1.27841437e+00 3.53728592e-01
-1.20468938e+00 -4.47926261e-02 -1.25128299e-01 -2.80760974e-01
-1.16302483e-01 2.26983339e-01 -8.09773326e-01 7.36251056e-01
1.11217022e+00 6.82765543e-01 -8.51920128e-01 1.32099319e+00
-1.42706826e-01 8.88567984e-01 -2.65745044e-01 -3.70297953e-02
-2.73856390e-02 -1.04823932e-01 1.06391542e-01 1.08064485e+00
4.26393390e-01 3.00024629e-01 -1.24353983e-01 5.87041020e-01
6.55017197e-02 -9.78875384e-02 -3.37905973e-01 3.10813874e-01
2.11807285e-02 1.46199512e+00 -1.05711925e+00 -1.98704153e-01
-4.07089353e-01 1.15641224e+00 1.14650624e-02 -5.80357537e-02
-7.71266460e-01 -5.50640643e-01 3.10557991e-01 2.44923666e-01
-9.00556669e-02 1.46740839e-01 -5.06521940e-01 -9.63307619e-01
-1.20589837e-01 -7.42279291e-01 3.55880946e-01 -6.51669502e-01
-1.20360243e+00 7.16052055e-01 1.13411650e-01 -1.29765821e+00
1.95429340e-01 -2.65439153e-01 -8.26710463e-01 1.00440168e+00
-1.46020985e+00 -8.65365624e-01 -6.38166964e-01 7.18745351e-01
6.28586948e-01 -1.90866962e-01 1.06850612e+00 5.68457901e-01
-6.35447145e-01 8.13356340e-01 -2.03557745e-01 -2.87754059e-01
6.14394665e-01 -1.54963660e+00 -3.35442722e-01 5.87784767e-01
5.75543987e-03 8.52618098e-01 4.13142353e-01 -2.70226032e-01
-1.21395004e+00 -8.10319126e-01 7.91472912e-01 -4.85491991e-01
4.00279790e-01 -1.29728094e-01 -6.58083439e-01 1.42298415e-01
3.87866974e-01 -2.97388464e-01 1.42071784e+00 1.09111838e-01
5.18144369e-01 -5.55514514e-01 -1.33811283e+00 2.68119782e-01
6.38813078e-01 -5.89940310e-01 -1.01022315e+00 -5.76188713e-02
-1.08469628e-01 -2.04759017e-01 -1.07654989e+00 2.19077334e-01
6.71845913e-01 -1.13509655e+00 5.25311053e-01 -4.74012643e-02
7.71167278e-02 -2.43236199e-01 2.47862026e-01 -1.35248137e+00
-1.53308332e-01 -6.17015779e-01 3.12865637e-02 1.03916764e+00
3.04010451e-01 -5.38726449e-01 8.35115075e-01 5.90167284e-01
-4.84130651e-01 -1.10625494e+00 -1.21177089e+00 -6.17041886e-01
-2.43217647e-01 -4.82705384e-01 3.25659811e-01 6.17734790e-01
3.00533026e-01 5.77755094e-01 -1.93521142e-01 1.36896715e-01
2.30485991e-01 -2.07125843e-01 4.87925053e-01 -1.51392758e+00
-2.86557674e-01 -1.64467350e-01 -7.94319272e-01 -5.31667233e-01
-8.43842998e-02 -1.15426612e+00 -1.52053148e-01 -1.70398808e+00
4.29429203e-01 -3.59079719e-01 -8.92041445e-01 3.79501551e-01
-1.81026593e-01 7.30814755e-01 4.04364371e-04 7.77465105e-02
-5.75931549e-01 4.07611340e-01 8.76127183e-01 1.48367494e-01
-4.01059449e-01 3.49031210e-01 -8.45707357e-01 7.43426561e-01
1.01930368e+00 -6.49071395e-01 -7.97326744e-01 -3.09298368e-04
7.65556470e-02 8.07188377e-02 5.39902672e-02 -1.40806413e+00
3.77394676e-01 5.37683308e-01 8.18152964e-01 -4.22775477e-01
6.01468444e-01 -8.90667737e-01 1.82182804e-01 5.47624767e-01
1.37272656e-01 3.47759187e-01 3.31467003e-01 4.34302777e-01
-1.57837331e-01 -2.23392814e-01 8.43601823e-01 -1.15203805e-01
-4.83668953e-01 -1.80489346e-02 -5.50113738e-01 -8.78097562e-05
9.77838457e-01 -8.83777797e-01 1.05354220e-01 -1.59595147e-01
-9.41989064e-01 1.53534859e-02 1.07694112e-01 3.67860980e-02
5.59189200e-01 -7.83842862e-01 -1.56065762e-01 5.22700846e-01
4.69141379e-02 -2.08609432e-01 3.20662469e-01 1.79926383e+00
-3.79537940e-01 3.93534005e-01 -5.57135582e-01 -6.57631934e-01
-1.27374411e+00 3.25825153e-04 3.11353534e-01 -2.21078932e-01
-7.07134724e-01 1.02052474e+00 -2.33007088e-01 4.56900857e-02
2.54881322e-01 -5.50016463e-01 -3.51307482e-01 2.49076322e-01
4.70533073e-01 6.53945982e-01 5.35839379e-01 -3.06630939e-01
-5.81703305e-01 4.57849592e-01 -2.58063916e-02 1.84146091e-01
1.41229129e+00 -1.98767602e-01 -1.08142592e-01 6.46942854e-01
8.66953135e-01 -1.59655333e-01 -6.96044922e-01 4.52155232e-01
1.69146992e-02 2.95742135e-02 1.86660647e-01 -1.13461041e+00
-8.10546517e-01 9.67381835e-01 1.16190231e+00 9.13609266e-02
1.49476635e+00 -1.98909357e-01 7.02202499e-01 3.34930569e-01
5.90879977e-01 -9.98650610e-01 -1.74241677e-01 -1.17044792e-01
3.36685717e-01 -1.15684330e+00 1.77869886e-01 3.20730120e-01
-7.00333357e-01 1.15023947e+00 3.96941394e-01 1.43274162e-02
8.35614860e-01 1.02488726e-01 7.91462883e-02 -3.84577483e-01
-2.79788673e-01 -1.71757676e-02 4.07245249e-01 5.31353116e-01
4.82249737e-01 -1.90689228e-02 -4.76533383e-01 1.05773365e+00
-4.50031489e-01 2.84598202e-01 3.89911920e-01 8.71951401e-01
-1.71759576e-01 -9.95068252e-01 -1.48841873e-01 9.87336755e-01
-7.38816977e-01 -1.98546454e-01 -3.84382941e-02 8.82885098e-01
5.74756563e-01 1.17133796e+00 -2.01948956e-01 -4.46554154e-01
2.88087636e-01 3.97760004e-01 5.22736847e-01 -7.78850138e-01
-9.85408604e-01 -5.83244264e-02 -1.86993897e-01 -4.06267047e-01
-6.96202219e-01 -4.62052107e-01 -1.24219012e+00 -4.19937298e-02
-4.04660195e-01 5.63409090e-01 5.90222657e-01 1.22076750e+00
3.25106353e-01 7.62679100e-01 3.04889083e-01 -8.15752089e-01
-1.85611755e-01 -1.08868098e+00 -9.56406772e-01 -8.30560625e-02
2.53211528e-01 -8.37160289e-01 -3.26213241e-01 -3.64784710e-02] | [13.53005313873291, 3.4997591972351074] |
62e4baa3-c63d-4c6c-8bde-582949b9627e | on-the-computational-modeling-of-meaning | 2307.04518 | null | https://arxiv.org/abs/2307.04518v1 | https://arxiv.org/pdf/2307.04518v1.pdf | On the Computational Modeling of Meaning: Embodied Cognition Intertwined with Emotion | This document chronicles this author's attempt to explore how words come to mean what they do, with a particular focus on child language acquisition and what that means for models of language understanding.\footnote{I say \emph{historical} because I synthesize the ideas based on when I discovered them and how those ideas influenced my later thinking.} I explain the setting for child language learning, how embodiment -- being able to perceive and enact in the world, including knowledge of concrete and abstract concepts -- is crucial, and how emotion and cognition relate to each other and the language learning process. I end with what I think are some of the requirements for a language-learning agent that learns language in a setting similar to that of children. This paper can act as a potential guide for ongoing and future work in modeling language. | ['Casey Kennington'] | 2023-07-10 | null | null | null | null | ['language-acquisition'] | ['natural-language-processing'] | [ 2.91522872e-02 4.00094420e-01 -2.39839152e-01 -6.08470082e-01
1.89374313e-01 -6.51170135e-01 5.91189504e-01 5.43923080e-01
-4.35603589e-01 3.96291792e-01 5.86306155e-01 -5.98330438e-01
-6.13588654e-02 -8.05914998e-01 -4.44355786e-01 -3.42826903e-01
6.16699718e-02 1.65702984e-01 -3.04682434e-01 -4.08117503e-01
3.80083740e-01 4.50731248e-01 -1.50492835e+00 1.21090688e-01
6.47337854e-01 -8.21854174e-02 3.82019132e-01 5.65755010e-01
-5.84096372e-01 1.59812403e+00 -4.60187137e-01 -2.24736378e-01
-5.04805744e-01 -6.51017249e-01 -1.21354711e+00 2.67707109e-01
-8.99872407e-02 -5.50000370e-01 1.16282880e-01 8.70351553e-01
-1.17774792e-02 3.00813138e-01 6.54809058e-01 -8.51569831e-01
-9.61947978e-01 1.03794158e+00 -9.14799273e-02 3.10140133e-01
4.85317558e-01 -1.46612078e-01 5.36218345e-01 -5.35151482e-01
5.29022694e-01 1.36383367e+00 3.48610818e-01 1.23406339e+00
-1.05408728e+00 -6.86586082e-01 7.92449057e-01 -1.03405014e-01
-1.08121371e+00 -5.96243083e-01 5.67185104e-01 -7.63060093e-01
1.22793090e+00 -9.34908092e-02 1.18364680e+00 8.49549890e-01
1.36238992e-01 9.28043783e-01 1.00991416e+00 -9.55174863e-01
1.02867186e-01 6.36629105e-01 4.46980387e-01 5.51609457e-01
1.05129637e-01 1.30393773e-01 -7.56427884e-01 2.46916607e-01
9.43545878e-01 -3.35616022e-01 -9.49003100e-02 -1.81766432e-02
-8.82616580e-01 8.68005812e-01 -1.35130018e-01 1.05787671e+00
-1.80924580e-01 4.54897523e-01 1.91522852e-01 5.38557172e-01
2.38725230e-01 4.84656155e-01 -4.29230273e-01 -5.72983623e-01
-5.92819512e-01 2.50748713e-02 8.23356986e-01 6.60204113e-01
3.62690598e-01 3.72341663e-01 7.81544924e-01 7.98468709e-01
8.75450611e-01 4.01198357e-01 3.85001361e-01 -1.02099216e+00
-1.67778760e-01 4.11724418e-01 -2.07589716e-01 -5.30055165e-01
-4.36978251e-01 -3.60809952e-01 2.47713089e-01 4.63914454e-01
4.53891426e-01 -4.08672094e-01 -5.53443789e-01 2.28044915e+00
1.18056238e-01 -7.17553049e-02 4.99467731e-01 4.60702330e-01
1.07694376e+00 7.96231449e-01 7.75627494e-01 -5.36246836e-01
1.09767354e+00 -4.40316737e-01 -5.78925371e-01 -5.53570747e-01
1.15708005e+00 -6.27693057e-01 9.60250556e-01 4.17067766e-01
-1.55409241e+00 -2.15164706e-01 -1.01131070e+00 -1.97947428e-01
-2.66165316e-01 -3.04413736e-01 1.30153525e+00 8.89719069e-01
-1.33156860e+00 2.74355978e-01 -1.01629579e+00 -9.27946091e-01
-4.05863672e-02 1.14121236e-01 -8.50972384e-02 3.09765995e-01
-7.87965834e-01 1.09680891e+00 5.24855852e-01 -7.66115710e-02
-6.63855076e-01 -4.64316100e-01 -8.85979950e-01 -1.95454180e-01
-4.21315916e-02 -5.95561147e-01 1.54429829e+00 -1.45265019e+00
-1.51750970e+00 1.45903254e+00 -1.22948974e-01 -4.47623543e-02
-3.00320089e-01 -3.39243531e-01 -5.25614560e-01 2.24728435e-01
-2.89405771e-02 7.05063701e-01 3.12221080e-01 -1.40204358e+00
-7.08396852e-01 -3.92479748e-01 4.97127384e-01 4.09332365e-01
-2.72401184e-01 6.76744521e-01 9.23000798e-02 -7.54869878e-01
5.15906274e-01 -4.78060275e-01 7.15973452e-02 -1.52930068e-02
4.02967155e-01 -5.92914879e-01 7.31289759e-02 -2.93894410e-01
1.29121530e+00 -2.22129321e+00 8.28200504e-02 1.30391583e-01
3.83920550e-01 3.85186337e-02 -8.25416148e-02 1.18496656e+00
-3.20704639e-01 3.16478580e-01 4.32461113e-01 5.65952286e-02
1.66362435e-01 3.42108369e-01 -4.97032434e-01 2.47426987e-01
-2.94007268e-03 8.77646863e-01 -1.00434887e+00 -2.91402578e-01
1.11663170e-01 8.05512190e-01 -3.84338289e-01 9.96819884e-02
-9.60438028e-02 4.81152266e-01 -6.62219942e-01 2.57937849e-01
7.25649223e-02 -1.14503324e-01 3.00599486e-01 7.67327368e-01
-4.27508622e-01 7.53487647e-01 -9.20740306e-01 1.57265413e+00
-5.57272494e-01 5.75212479e-01 4.08874512e-01 -1.01174688e+00
8.13763320e-01 7.83560693e-01 -2.11432040e-01 -5.30190706e-01
2.36616746e-01 2.03557000e-01 5.78419626e-01 -8.97461832e-01
-2.69589752e-01 -7.03690171e-01 2.10415423e-01 1.13693285e+00
-1.63761616e-01 -6.05599463e-01 5.11376709e-02 4.96804893e-01
4.85334605e-01 2.33095437e-01 4.34453279e-01 -7.09253132e-01
5.83782375e-01 -3.59916568e-01 3.43407452e-01 5.14189422e-01
5.79710901e-02 -3.50428671e-01 2.04940081e-01 -4.48868603e-01
-4.23163116e-01 -1.16782176e+00 -1.25332633e-02 1.56379342e+00
-1.04283072e-01 -2.17117608e-01 -8.08925152e-01 2.52307057e-01
-6.45154357e-01 1.50797498e+00 -4.39385891e-01 -3.35618742e-02
-4.76661772e-01 -4.71489280e-02 2.63716280e-01 7.19520867e-01
-5.06651253e-02 -1.50139868e+00 -1.01032734e+00 2.98889488e-01
4.01184618e-01 -5.38334966e-01 2.53286093e-01 2.42797017e-01
-1.04447722e+00 -6.19982541e-01 -9.78683680e-02 -1.28712714e+00
7.11308658e-01 1.68280937e-02 1.10752630e+00 1.01663363e+00
2.01531217e-01 1.17430019e+00 -5.28458476e-01 -9.14624393e-01
-7.24795461e-01 -4.94152069e-01 -1.08840741e-01 -8.08130622e-01
8.00915778e-01 -1.06031561e+00 5.35490736e-03 -6.41510069e-01
-1.02949238e+00 5.35960734e-01 6.79710060e-02 3.60583633e-01
-1.89158991e-01 5.77513538e-02 4.69322354e-01 -8.78368616e-01
7.12742865e-01 -3.34863007e-01 -5.07464930e-02 4.29435223e-01
-3.37480485e-01 -3.24766450e-02 -3.87434411e-04 -4.74025577e-01
-1.23914051e+00 -6.07957721e-01 -1.43351302e-01 6.55743659e-01
-7.06250727e-01 5.75146258e-01 -8.80755298e-03 2.87538469e-01
3.03867579e-01 3.17204803e-01 -5.15015200e-02 -3.02639365e-01
3.66781622e-01 4.06042397e-01 4.58141863e-01 -1.36549401e+00
6.60133183e-01 1.57827690e-01 -2.38169521e-01 -1.21600640e+00
-7.79737234e-01 -1.47400841e-01 -6.91978097e-01 -1.33810341e-01
9.46071923e-01 -1.17116249e+00 -6.18333101e-01 4.14625466e-01
-1.14252126e+00 -8.15194547e-01 -3.80397797e-01 1.04572999e+00
-6.81484103e-01 -2.43107844e-02 -8.43821168e-01 -1.33718944e+00
-1.04875848e-01 -8.38694394e-01 3.99438262e-01 6.31688833e-01
-8.36665630e-01 -1.81975436e+00 -1.29135698e-01 3.73043448e-01
2.64779806e-01 7.48369843e-02 1.54864776e+00 -7.42976844e-01
-2.21662045e-01 5.97201347e-01 4.24533129e-01 4.77847643e-02
1.17339596e-01 3.02618980e-01 -8.39875996e-01 -2.00973824e-01
4.20980930e-01 -7.82193422e-01 1.73573196e-01 1.45048708e-01
5.12126327e-01 -1.07644260e-01 -8.84171203e-02 2.79153705e-01
1.56306982e+00 7.44391918e-01 6.56547993e-02 2.63282448e-01
1.84163943e-01 1.08146238e+00 2.21328527e-01 1.28993586e-01
6.93097889e-01 8.45552012e-02 -3.33071142e-01 8.41866434e-02
4.52277586e-02 -2.92040199e-01 4.15167511e-01 1.10143292e+00
-7.99886510e-02 -1.07631482e-01 -1.22523332e+00 8.11995804e-01
-1.58684480e+00 -9.67151165e-01 1.15237519e-01 1.64023399e+00
1.31481016e+00 1.18439257e-01 5.50703928e-02 7.67760873e-02
2.07594886e-01 5.53411208e-02 -1.52227536e-01 -1.23147452e+00
1.74436137e-01 7.22141743e-01 -6.26471400e-01 1.08257961e+00
-2.70270348e-01 1.28446496e+00 7.24921608e+00 3.39229926e-02
-1.20475829e+00 3.18205170e-02 5.50539374e-01 1.20634012e-01
-8.53001237e-01 -5.84521738e-04 -4.36189711e-01 -2.46206760e-01
7.41424382e-01 -5.00751972e-01 5.91334820e-01 4.69935298e-01
2.86373675e-01 -3.37007821e-01 -1.70178866e+00 5.50164461e-01
1.11497343e-01 -8.44235420e-01 -6.22884855e-02 -2.76826620e-01
3.07949662e-01 -3.28610867e-01 1.06228944e-02 1.87784940e-01
4.78786618e-01 -1.39267838e+00 1.02251995e+00 3.72764230e-01
3.09829891e-01 -5.33445656e-01 1.16801307e-01 8.73068511e-01
-1.00122595e+00 1.11705892e-01 2.13543564e-01 -1.03136945e+00
-4.88241650e-02 -2.58096009e-01 -6.08593106e-01 -4.31959927e-01
3.96130592e-01 3.90417665e-01 -2.25643083e-01 3.70791644e-01
-6.79515719e-01 7.48137414e-01 -2.98716754e-01 -4.86594826e-01
1.98397428e-01 -7.95151517e-02 3.51874799e-01 1.16075194e+00
-1.46354456e-02 1.08964264e+00 2.01786071e-01 9.08397973e-01
6.59268260e-01 4.79594827e-01 -5.29574931e-01 -5.31606257e-01
6.88867211e-01 6.93969429e-01 -1.16362000e+00 -3.98175955e-01
-7.84852684e-01 3.02971035e-01 1.17904015e-01 4.52773154e-01
-2.45941162e-01 1.87430938e-03 8.65768075e-01 9.68432575e-02
-9.24492925e-02 -4.84844565e-01 -1.43318251e-01 -1.09812462e+00
-4.58341986e-01 -9.32091057e-01 -8.89546350e-02 -1.03095973e+00
-8.48085999e-01 1.42522067e-01 4.40953434e-01 -1.21833913e-01
-5.27518153e-01 -6.05537713e-01 -9.48232830e-01 7.28492439e-01
-8.77160311e-01 -1.20311153e+00 3.44350129e-01 2.89682984e-01
6.92005277e-01 9.65107307e-02 1.23997438e+00 -3.83899182e-01
-1.84558168e-01 2.20477432e-01 -6.55379474e-01 1.94896966e-01
-1.60278857e-01 -1.08014405e+00 2.64601529e-01 7.86563814e-01
5.13667762e-01 1.34555006e+00 8.91948462e-01 -4.36963737e-01
-1.17015254e+00 -9.69377980e-02 1.28970480e+00 -4.94635195e-01
9.16334331e-01 -2.25115925e-01 -1.02959883e+00 1.10175955e+00
6.66636229e-01 -7.38459587e-01 1.20415986e+00 1.08789913e-01
-9.24022645e-02 3.07634950e-01 -1.02314341e+00 7.39834845e-01
1.19551873e+00 -7.43073881e-01 -1.30026591e+00 1.97214648e-01
9.48604643e-01 -1.73484728e-01 -7.97284186e-01 -9.19168741e-02
6.40019834e-01 -1.08585584e+00 6.72100842e-01 -7.65303016e-01
1.71745226e-01 -8.23241547e-02 -8.25470537e-02 -9.47579503e-01
-4.12880838e-01 -6.89376593e-01 3.17415982e-01 1.58369172e+00
4.81545299e-01 -8.75883281e-01 4.34438020e-01 1.09485149e+00
1.59600332e-01 -6.69929206e-01 -5.20046115e-01 -1.37271598e-01
9.14416790e-01 -8.67704034e-01 5.84606111e-01 1.06382895e+00
7.81832874e-01 3.28935832e-01 4.99057502e-01 1.07123256e-02
2.82686532e-01 -7.18383417e-02 4.35003936e-01 -1.26139271e+00
-3.22489798e-01 -4.50478405e-01 -1.45499334e-01 -8.92056108e-01
2.02496246e-01 -7.98098743e-01 -4.03445512e-01 -1.63979864e+00
4.54817377e-02 -3.56122762e-01 -1.82741031e-01 5.72725117e-01
2.78097063e-01 -3.66184384e-01 4.92151499e-01 -1.99066341e-01
-2.28747696e-01 5.38238809e-02 9.96263027e-01 4.54169780e-01
-3.47102851e-01 -3.97593111e-01 -1.22112131e+00 1.25013673e+00
8.89598727e-01 -1.79595381e-01 -1.00310385e+00 -7.27345467e-01
5.91573179e-01 2.54855547e-02 5.87964244e-02 -8.62916052e-01
2.14384288e-01 -9.24388111e-01 3.43204886e-01 -2.53282964e-01
2.71753699e-01 -7.71543801e-01 8.25689211e-02 6.59949720e-01
-6.37042522e-01 2.68704027e-01 4.72703457e-01 -2.25302801e-01
7.67030120e-02 -5.77557802e-01 6.26008272e-01 -4.35217649e-01
-1.06199455e+00 -3.86377662e-01 -1.05368471e+00 1.59524694e-01
1.02400482e+00 -3.22401732e-01 -1.09824352e-01 -5.38850784e-01
-1.09468424e+00 3.26697320e-01 6.73249960e-01 5.29819369e-01
5.67624986e-01 -9.03897345e-01 -4.26475257e-01 2.56827027e-01
-3.27082202e-02 -2.17941105e-01 -2.51454890e-01 3.92077804e-01
-6.64021850e-01 4.42074448e-01 -5.59763052e-02 -1.20509595e-01
-1.44874263e+00 2.93630987e-01 3.85125369e-01 3.84190857e-01
-6.70777917e-01 1.47038388e+00 6.27559662e-01 -1.25790164e-01
4.38917994e-01 -5.83819687e-01 -6.15071177e-01 -1.63376287e-01
8.03655386e-01 -1.94648579e-01 -8.38190258e-01 -7.55152464e-01
-4.17879432e-01 8.40170026e-01 1.75022800e-02 -4.56649363e-01
1.60936749e+00 -4.64171141e-01 -7.50274956e-01 1.14579213e+00
6.77475750e-01 2.51921505e-01 -6.63114429e-01 -2.70387560e-01
7.11040944e-02 -2.56680837e-03 -2.17184275e-01 -9.74123240e-01
-5.97902834e-01 1.13773549e+00 2.53833711e-01 1.39729112e-01
1.08960760e+00 4.45205063e-01 1.45273209e-01 4.30209965e-01
5.15704691e-01 -1.17830861e+00 4.19337340e-02 4.02359962e-01
8.54430795e-01 -6.40548825e-01 2.57801801e-01 -3.58371675e-01
-4.20596570e-01 1.46241236e+00 7.14232326e-01 1.55590892e-01
8.42561901e-01 4.44934756e-01 3.90886098e-01 -4.71698165e-01
-1.29897010e+00 2.10728478e-02 -2.49163166e-01 8.43756974e-01
1.23045814e+00 1.64514199e-01 -7.11285233e-01 4.48395103e-01
-6.59946144e-01 1.26051098e-01 5.46283424e-01 1.10135233e+00
-9.41384733e-01 -1.31800532e+00 -2.92121798e-01 -1.98735178e-01
-5.75203896e-01 -1.63626909e-01 -6.00560963e-01 8.86172652e-01
7.10989714e-01 1.33487833e+00 1.03350945e-01 7.94272572e-02
-2.96750247e-01 4.14105147e-01 8.89802277e-01 -1.25786877e+00
-7.21568882e-01 1.72145233e-01 1.90832838e-01 -1.31151125e-01
-8.00421774e-01 -8.53514552e-01 -2.03093886e+00 -3.90638500e-01
-8.67265910e-02 3.78704876e-01 4.93767917e-01 1.31625032e+00
-4.86251593e-01 1.04994908e-01 1.90182235e-02 -1.78292781e-01
-2.35669930e-02 -6.75425649e-01 -5.57153106e-01 -1.55977622e-01
3.33677620e-01 -1.04483232e-01 -3.73970777e-01 1.72275022e-01] | [10.363302230834961, 8.578219413757324] |
96cd6e04-3f6c-4343-8638-d920a9f16a19 | heads-up-unsupervised-constituency-parsing | 2010.09517 | null | https://arxiv.org/abs/2010.09517v1 | https://arxiv.org/pdf/2010.09517v1.pdf | Heads-up! Unsupervised Constituency Parsing via Self-Attention Heads | Transformer-based pre-trained language models (PLMs) have dramatically improved the state of the art in NLP across many tasks. This has led to substantial interest in analyzing the syntactic knowledge PLMs learn. Previous approaches to this question have been limited, mostly using test suites or probes. Here, we propose a novel fully unsupervised parsing approach that extracts constituency trees from PLM attention heads. We rank transformer attention heads based on their inherent properties, and create an ensemble of high-ranking heads to produce the final tree. Our method is adaptable to low-resource languages, as it does not rely on development sets, which can be expensive to annotate. Our experiments show that the proposed method often outperform existing approaches if there is no development set present. Our unsupervised parser can also be used as a tool to analyze the grammars PLMs learn implicitly. For this, we use the parse trees induced by our method to train a neural PCFG and compare it to a grammar derived from a human-annotated treebank. | ['Frank Keller', 'Reinald Kim Amplayo', 'Taeuk Kim', 'Bowen Li'] | 2020-10-19 | null | https://aclanthology.org/2020.aacl-main.43 | https://aclanthology.org/2020.aacl-main.43.pdf | asian-chapter-of-the-association-for | ['constituency-parsing'] | ['natural-language-processing'] | [ 2.58984596e-01 5.83319247e-01 -1.91335976e-01 -5.62415540e-01
-1.10876727e+00 -9.07119393e-01 6.10987961e-01 2.69652516e-01
-3.77026469e-01 7.28495598e-01 2.63910085e-01 -6.82017803e-01
2.81386346e-01 -8.67534697e-01 -8.92638564e-01 -4.36859280e-01
1.88412279e-01 7.58803010e-01 5.01816392e-01 -1.31813973e-01
7.80845881e-02 1.89665973e-01 -1.37645864e+00 4.01757807e-01
1.18277252e+00 5.40949464e-01 7.38089323e-01 3.59109402e-01
-5.63414037e-01 8.19201648e-01 -4.11669016e-01 -6.07273579e-01
-1.11708671e-01 -5.84658504e-01 -1.14486408e+00 -1.16866127e-01
1.77655026e-01 8.36845040e-02 3.59239697e-01 1.03584898e+00
7.70470053e-02 -4.29620780e-02 4.40757453e-01 -6.67337894e-01
-4.61939037e-01 1.21861815e+00 -2.78563946e-01 2.20679283e-01
2.07607880e-01 -2.65395463e-01 1.58539200e+00 -8.24783862e-01
8.14869046e-01 1.44660258e+00 3.18283498e-01 6.89593315e-01
-1.40723443e+00 -5.18253207e-01 3.93163592e-01 1.07747301e-01
-8.09069157e-01 -3.69543165e-01 8.29138219e-01 -2.66951740e-01
1.19328082e+00 -9.76880789e-02 3.32510799e-01 9.86680329e-01
-7.81482682e-02 1.00485420e+00 1.25441253e+00 -1.03885472e+00
1.30508915e-01 2.06188112e-01 4.18204993e-01 8.86163712e-01
2.00097233e-01 -2.32457712e-01 -3.88853729e-01 1.15939662e-01
4.95673150e-01 -6.94039583e-01 -9.12322197e-03 -1.98826551e-01
-9.15821016e-01 1.08716738e+00 6.53551100e-03 7.05876887e-01
-2.90576607e-01 -1.91631392e-01 2.63060629e-01 2.32675523e-01
4.70984042e-01 5.60920894e-01 -6.98920071e-01 3.27004641e-02
-7.52437472e-01 -4.46762480e-02 1.00397003e+00 8.49913180e-01
1.05197513e+00 -2.09000751e-01 2.97922660e-02 1.08390665e+00
3.47935230e-01 2.68692374e-01 5.02021730e-01 -7.30035186e-01
6.93787038e-01 6.44513488e-01 -1.90063372e-01 -4.02913153e-01
-2.41619185e-01 -1.91307619e-01 -2.90006578e-01 -6.95826560e-02
5.07066965e-01 -8.47875401e-02 -9.04412448e-01 2.03478909e+00
3.01727563e-01 -3.39177519e-01 1.77568376e-01 4.94004965e-01
6.09108210e-01 7.04214633e-01 3.54279160e-01 -6.97472394e-02
1.44221818e+00 -8.19870830e-01 -5.71933448e-01 -5.07167220e-01
9.63577032e-01 -5.90086520e-01 1.31178522e+00 3.85940343e-01
-1.04759955e+00 -4.27898824e-01 -7.81753123e-01 -1.96338281e-01
-3.71855319e-01 2.52849042e-01 5.75618565e-01 7.19580889e-01
-1.09742725e+00 7.34515369e-01 -9.62879300e-01 -4.76967692e-01
2.26672217e-01 4.86373693e-01 -4.38299716e-01 -4.47485819e-02
-1.23287082e+00 1.08700645e+00 6.75954938e-01 7.11428421e-03
-7.54567444e-01 -2.09378704e-01 -1.10550511e+00 1.58834737e-02
4.09654588e-01 -3.46485674e-01 1.49756026e+00 -1.24820817e+00
-1.71687651e+00 1.10371745e+00 -3.02423805e-01 -4.42365050e-01
-5.70139848e-02 -3.58356059e-01 7.69817783e-03 -1.19180091e-01
4.19049531e-01 6.56059086e-01 3.99342179e-01 -1.08318996e+00
-8.63709807e-01 -3.12409341e-01 1.35712624e-01 -4.81125750e-02
-2.68337190e-01 5.13893902e-01 -5.13404787e-01 -4.05190498e-01
5.87205403e-04 -8.12536240e-01 -3.18689108e-01 -8.67947757e-01
-4.86524254e-01 -6.28638804e-01 3.72550339e-01 -6.94737554e-01
1.23589265e+00 -1.71083772e+00 2.96704769e-01 7.77040869e-02
-2.42224067e-01 3.38440597e-01 -2.34775364e-01 4.27204370e-01
9.07668322e-02 4.02588785e-01 -5.02839029e-01 -4.60531294e-01
8.74082968e-02 5.97179294e-01 -2.93668717e-01 -1.44235296e-02
7.55980551e-01 7.49699593e-01 -9.69167173e-01 -7.67164946e-01
-3.35579552e-02 1.24133982e-01 -5.69589078e-01 4.66010571e-01
-5.30763626e-01 6.69157445e-01 -5.77276826e-01 5.82572103e-01
2.78592050e-01 -4.92360853e-02 6.87929809e-01 2.68203825e-01
-3.07249010e-01 1.06807125e+00 -7.37650871e-01 1.57783186e+00
-7.55302310e-01 4.12119031e-01 -5.78891598e-02 -1.30938303e+00
9.33866143e-01 2.17878073e-01 -1.81921840e-01 -5.75873196e-01
8.80655646e-02 5.49963892e-01 3.80982667e-01 -3.58373076e-01
1.52500287e-01 -3.24782103e-01 -4.62457836e-01 3.52259696e-01
6.44840837e-01 5.83239757e-02 5.97927570e-01 -7.69863501e-02
1.26819479e+00 4.92802501e-01 5.52529156e-01 -5.14432192e-01
7.65306771e-01 8.55485648e-02 9.21029329e-01 6.33211434e-01
2.68217802e-01 2.62758255e-01 7.11453915e-01 -2.96635896e-01
-8.36898744e-01 -1.01633787e+00 -1.07148126e-01 1.52916300e+00
-3.52896839e-01 -5.09982228e-01 -1.05700302e+00 -1.05251312e+00
-4.97193724e-01 8.73082459e-01 -4.61425573e-01 3.64852786e-01
-1.24060440e+00 -5.71438849e-01 4.51943427e-01 6.49233520e-01
4.56986725e-02 -1.69483423e+00 -4.53522474e-01 6.75401926e-01
-2.50785738e-01 -1.31662035e+00 7.36708194e-02 5.74027002e-01
-1.02479696e+00 -1.02491617e+00 -3.53453815e-01 -1.24160624e+00
5.65506995e-01 -5.84590912e-01 1.55069268e+00 -7.02821314e-02
2.96756506e-01 -2.34936625e-02 -4.71922070e-01 -5.83549440e-01
-8.57972205e-01 6.41711295e-01 -1.29772499e-01 -2.00112481e-02
5.36222219e-01 -6.60694361e-01 4.13029604e-02 -5.74017763e-02
-6.32317901e-01 8.34326744e-02 6.69885397e-01 7.65095711e-01
7.32055426e-01 -2.20669359e-01 7.20517576e-01 -1.57528496e+00
4.39946592e-01 -3.10551077e-01 -9.53212142e-01 2.59390593e-01
-4.32751745e-01 6.96648717e-01 9.31058586e-01 -2.20591158e-01
-1.37947941e+00 2.70185798e-01 -4.64909583e-01 -1.48869744e-02
-3.30735505e-01 6.83735192e-01 -7.18205273e-01 3.17770988e-01
2.38804355e-01 -1.86943300e-02 -4.95651037e-01 -9.04394448e-01
3.63003939e-01 4.30873513e-01 4.95689124e-01 -9.80748892e-01
7.25241244e-01 -9.78793055e-02 -1.35448113e-01 -6.44425809e-01
-1.12572789e+00 -2.40802824e-01 -9.29317892e-01 2.30507120e-01
8.99468303e-01 -5.27818084e-01 -2.25302696e-01 4.15666075e-03
-1.47474825e+00 -5.79226196e-01 -1.54679250e-02 4.72721189e-01
-4.67858255e-01 2.73994505e-01 -7.61611104e-01 -8.59936416e-01
-3.36621135e-01 -1.07966030e+00 1.02419472e+00 1.76733490e-02
-3.35379362e-01 -1.04076600e+00 3.86842161e-01 1.10097326e-01
7.83371553e-02 5.56404106e-02 1.48852730e+00 -9.56520975e-01
-6.86160266e-01 2.07236469e-01 6.01039156e-02 3.18974525e-01
2.62763351e-02 -1.03235550e-01 -1.06207752e+00 -7.70968711e-03
-2.26361722e-01 -3.94295514e-01 8.41418266e-01 2.36255631e-01
9.52387214e-01 -2.37535566e-01 -4.81193960e-01 2.68502265e-01
1.33020341e+00 8.40833560e-02 4.61796135e-01 5.03411233e-01
6.69031143e-01 1.13459265e+00 6.48079813e-01 -1.80953562e-01
4.24460679e-01 3.33331108e-01 2.41761491e-01 3.26637179e-02
1.42941251e-01 -4.94031072e-01 7.30064213e-01 1.27906454e+00
-3.21291275e-02 -4.13640559e-01 -1.11335909e+00 8.64199877e-01
-1.66874456e+00 -4.17785019e-01 -4.64630164e-02 2.14428687e+00
9.96590436e-01 3.30137491e-01 -9.55972597e-02 4.19387557e-02
7.30305493e-01 3.66808325e-02 1.00937165e-01 -7.37624824e-01
1.03360415e-01 9.83405054e-01 1.41399324e-01 6.83910847e-01
-1.24430680e+00 1.54817426e+00 5.45150900e+00 5.67560256e-01
-9.42948103e-01 2.56078392e-01 2.68030167e-01 4.53797907e-01
-4.08148199e-01 4.20597464e-01 -1.17691803e+00 1.57099113e-01
1.26478231e+00 8.85251611e-02 5.87759400e-03 8.32743227e-01
-3.22778262e-02 4.99913320e-02 -1.53984308e+00 3.67490828e-01
-1.70490652e-01 -9.69517529e-01 1.09587885e-01 6.95369160e-03
2.66758561e-01 2.07346484e-01 -3.02350849e-01 6.36475563e-01
7.83166885e-01 -9.32398677e-01 5.61950028e-01 -5.00262342e-02
6.61335170e-01 -6.81086421e-01 8.17204416e-01 6.67705297e-01
-1.31594741e+00 8.33283141e-02 -3.89390588e-01 3.85546102e-03
2.86625862e-01 3.43242526e-01 -1.01913464e+00 4.27853137e-01
5.77970922e-01 3.06845874e-01 -5.49006701e-01 6.40454233e-01
-8.61662269e-01 1.10149026e+00 -3.97974730e-01 -2.78937101e-01
4.23685759e-01 -1.08767651e-01 3.88931930e-01 1.50608277e+00
2.05907881e-01 -1.06025510e-01 3.85575742e-01 9.19437230e-01
-1.65844887e-01 6.06684744e-01 -6.74040437e-01 -1.72508731e-01
3.11802655e-01 1.35132217e+00 -8.71784091e-01 -3.32103699e-01
-6.31700516e-01 8.07397604e-01 9.71660793e-01 -5.87092638e-02
-4.59899336e-01 -3.56492192e-01 2.12991878e-01 8.75024870e-02
5.06006777e-01 -5.93151636e-02 1.49001986e-01 -1.18060637e+00
2.02547424e-02 -9.14831936e-01 5.45264781e-01 -4.00677234e-01
-1.27356791e+00 1.02569652e+00 1.57353237e-01 -7.67718494e-01
-7.18385160e-01 -7.63067722e-01 -6.79298759e-01 9.10692453e-01
-1.72581029e+00 -1.19551241e+00 2.28667408e-01 2.25844085e-01
7.15098500e-01 -2.24939868e-04 1.16044271e+00 1.10444777e-01
-5.47625780e-01 5.39721966e-01 -3.66140395e-01 5.06628215e-01
5.07822394e-01 -1.54907668e+00 6.43847764e-01 1.20172369e+00
5.99105597e-01 6.37019217e-01 5.81896544e-01 -5.36091447e-01
-1.05449486e+00 -1.14212203e+00 1.46901178e+00 -5.74562609e-01
7.00982571e-01 -7.60718763e-01 -1.23831964e+00 1.00345337e+00
3.85795474e-01 -5.80009334e-02 7.24194229e-01 5.95246196e-01
-2.92853087e-01 8.26501697e-02 -7.17559218e-01 3.29546720e-01
9.60825443e-01 -3.22498262e-01 -1.10744691e+00 2.03450974e-02
6.59437180e-01 -2.13157550e-01 -6.78370595e-01 4.26479757e-01
2.83156455e-01 -6.37307823e-01 2.45185927e-01 -7.21301734e-01
5.29074430e-01 -3.77292484e-02 -2.43724994e-02 -1.32216573e+00
-3.90099794e-01 -4.96960193e-01 1.14378147e-01 1.65172017e+00
9.07077968e-01 -6.63786530e-01 7.68325686e-01 3.03985864e-01
-3.09275091e-01 -6.23750985e-01 -7.26098120e-01 -7.08422363e-01
3.59267533e-01 -3.76525819e-01 4.50048387e-01 7.70052612e-01
7.50441775e-02 9.07172620e-01 9.09832865e-02 2.57137626e-01
5.09862363e-01 3.38127702e-01 7.03294039e-01 -1.64233065e+00
-5.84592283e-01 -2.79626518e-01 6.46702887e-04 -9.59494412e-01
7.78928518e-01 -1.18421173e+00 4.28075612e-01 -1.49601567e+00
1.35198176e-01 -7.06394553e-01 -4.00150806e-01 7.99627304e-01
-1.79456741e-01 -9.56059322e-02 5.82068451e-02 -1.79499364e-03
-4.59080547e-01 2.87611932e-01 8.15827549e-01 1.93179935e-01
-2.47587487e-01 -1.76288322e-01 -6.34727240e-01 1.13417888e+00
7.93503463e-01 -8.86591017e-01 -1.15355752e-01 -5.68856061e-01
1.40883982e-01 -2.94470671e-03 -1.97128519e-01 -6.82933450e-01
-3.19843553e-02 1.53094728e-03 -3.40097435e-02 -4.28986043e-01
-1.62868485e-01 -5.04412830e-01 -2.71001905e-01 1.92427263e-01
-2.32777357e-01 2.72227637e-02 2.32187137e-01 1.82454959e-01
-3.75323236e-01 -7.28198826e-01 6.38257086e-01 -5.06818116e-01
-6.80265427e-01 1.21541396e-01 -5.58386624e-01 3.47622931e-01
4.26838815e-01 1.33678257e-01 -3.10041010e-03 3.94049734e-02
-6.60428405e-01 7.64089078e-02 3.23347569e-01 3.48177850e-01
8.34482536e-02 -9.66702223e-01 -6.71125054e-01 1.72909856e-01
1.55761540e-01 2.65005440e-01 -4.21146214e-01 4.35213536e-01
-3.37351859e-01 5.45011342e-01 -7.90186524e-02 -4.71707046e-01
-1.12492526e+00 5.53303659e-01 5.90437837e-02 -9.82367456e-01
-5.81370652e-01 7.32000649e-01 5.94270289e-01 -6.87821090e-01
-8.99807215e-02 -6.12461329e-01 -4.66598928e-01 -2.88172625e-02
2.22782046e-01 -3.13548505e-01 1.75792709e-01 -6.81511462e-01
-4.94898856e-01 5.31699002e-01 -2.00424373e-01 -2.91241109e-01
1.42569900e+00 1.40425205e-01 -1.86897099e-01 5.09310484e-01
9.88771021e-01 4.63228703e-01 -7.77946413e-01 -2.34083459e-01
7.85318196e-01 1.13597028e-01 -1.98352560e-01 -5.17612994e-01
-6.93027973e-01 1.04414964e+00 -3.56088318e-02 1.82069272e-01
9.79891658e-01 4.67003852e-01 7.88134754e-01 5.55960476e-01
5.77458918e-01 -9.81705070e-01 -2.37800270e-01 9.52126086e-01
5.80941916e-01 -1.17008078e+00 -5.99009216e-01 -6.54674888e-01
-3.37098420e-01 1.02840650e+00 6.07417762e-01 -2.19508484e-01
3.53153318e-01 4.36840415e-01 1.37423947e-02 6.90043718e-03
-9.11315918e-01 -6.72575295e-01 1.50594413e-01 6.20118082e-01
8.84667814e-01 -4.37150300e-02 -5.52903175e-01 9.65720236e-01
-4.12482709e-01 -2.62741685e-01 2.44476989e-01 9.34234500e-01
-5.69410920e-01 -1.92017055e+00 -1.40141919e-01 1.72123417e-01
-1.02643573e+00 -4.13982600e-01 -5.57025909e-01 8.99281979e-01
1.12587176e-01 7.87489295e-01 -5.92281334e-02 -5.76091511e-03
2.75881708e-01 4.91121680e-01 6.83890939e-01 -1.29510117e+00
-5.42471528e-01 1.38885587e-01 5.29850602e-01 -3.64257455e-01
-5.35137057e-01 -6.68518662e-01 -1.32020426e+00 5.14116466e-01
-5.56190312e-01 4.29848135e-01 4.02283192e-01 1.12420392e+00
3.27368155e-02 2.82326549e-01 4.54036683e-01 -6.67627335e-01
-2.63387114e-01 -1.10847414e+00 -1.90450460e-01 2.28444412e-01
-8.41655582e-02 -6.66041434e-01 -1.80964589e-01 1.78314388e-01] | [10.406314849853516, 9.550872802734375] |
0297bebd-52e7-41db-8b93-75393937813c | large-scale-traffic-data-imputation-with | 2301.11691 | null | https://arxiv.org/abs/2301.11691v1 | https://arxiv.org/pdf/2301.11691v1.pdf | Large-Scale Traffic Data Imputation with Spatiotemporal Semantic Understanding | Large-scale data missing is a challenging problem in Intelligent Transportation Systems (ITS). Many studies have been carried out to impute large-scale traffic data by considering their spatiotemporal correlations at a network level. In existing traffic data imputations, however, rich semantic information of a road network has been largely ignored when capturing network-wide spatiotemporal correlations. This study proposes a Graph Transformer for Traffic Data Imputation (GT-TDI) model to impute large-scale traffic data with spatiotemporal semantic understanding of a road network. Specifically, the proposed model introduces semantic descriptions consisting of network-wide spatial and temporal information of traffic data to help the GT-TDI model capture spatiotemporal correlations at a network level. The proposed model takes incomplete data, the social connectivity of sensors, and semantic descriptions as input to perform imputation tasks with the help of Graph Neural Networks (GNN) and Transformer. On the PeMS freeway dataset, extensive experiments are conducted to compare the proposed GT-TDI model with conventional methods, tensor factorization methods, and deep learning-based methods. The results show that the proposed GT-TDI outperforms existing methods in complex missing patterns and diverse missing rates. The code of the GT-TDI model will be available at https://github.com/KP-Zhang/GT-TDI. | ['Zhengbing He', 'Na Xie', 'Liang Zheng', 'Lan Wu', 'Kunpeng Zhang'] | 2023-01-27 | null | null | null | null | ['traffic-data-imputation'] | ['time-series'] | [-2.92489752e-02 -1.79545596e-01 -4.91629511e-01 -5.27805448e-01
-4.09322053e-01 4.82316390e-02 3.32484394e-01 -3.76231730e-01
1.70925543e-01 1.02429676e+00 7.05217957e-01 -5.58340251e-01
-5.90861559e-01 -1.22078454e+00 -1.02338445e+00 -4.13140446e-01
1.47981033e-01 6.51377797e-01 -9.83043313e-02 -3.51343602e-01
-1.58137027e-02 2.80819714e-01 -1.36324763e+00 3.37805659e-01
1.15907705e+00 9.21993494e-01 6.39030412e-02 9.86042693e-02
-2.98676163e-01 1.01962996e+00 -1.69828787e-01 -5.20507693e-01
2.03516543e-01 5.33099584e-02 -5.23118258e-01 -2.27601498e-01
2.04577625e-01 -3.08715582e-01 -1.09830225e+00 8.04638743e-01
1.94460794e-01 3.02582294e-01 4.11630183e-01 -2.14478183e+00
-9.33249235e-01 6.82705641e-01 -5.60402751e-01 -6.99492497e-03
2.95498557e-02 6.77441433e-02 4.23457891e-01 -6.54679060e-01
4.57282215e-01 1.48908019e+00 1.01802552e+00 1.53872430e-01
-1.12435997e+00 -1.04844940e+00 1.97351024e-01 5.89998841e-01
-1.63726556e+00 -4.14370298e-01 9.84896183e-01 -5.47184289e-01
8.08696866e-01 8.18188563e-02 4.41730410e-01 1.25309062e+00
2.25587234e-01 7.28780210e-01 8.35685611e-01 3.14307690e-01
7.18407854e-02 -2.73089647e-01 3.22334319e-01 3.61391127e-01
3.86292130e-01 1.46169916e-01 -5.42562425e-01 -8.80986601e-02
6.56542718e-01 5.39203167e-01 4.05316532e-01 1.02386609e-01
-1.25110924e+00 6.31079257e-01 6.54555976e-01 -1.27126381e-01
-7.29341030e-01 3.11088830e-01 3.93506229e-01 1.39667273e-01
7.92219639e-01 -5.39265633e-01 -3.42701823e-01 -2.44011804e-01
-6.93062425e-01 2.84737378e-01 5.47726870e-01 1.33012366e+00
1.05870366e+00 2.80460507e-01 -1.21782236e-01 8.45260680e-01
1.59013599e-01 7.30903089e-01 -1.55224055e-01 -1.16180491e+00
1.15778291e+00 7.69922614e-01 1.08006462e-01 -1.49959946e+00
-4.36204165e-01 -2.70329833e-01 -1.56159902e+00 -3.52201581e-01
3.58221352e-01 -5.53799510e-01 -8.04050505e-01 1.84049773e+00
3.88616860e-01 8.16256642e-01 -4.89758141e-02 9.78640199e-01
8.19375753e-01 6.43972337e-01 1.49282873e-01 2.33808905e-01
9.22302663e-01 -7.28916228e-01 -1.05733621e+00 3.33151594e-02
6.61001265e-01 -4.05921102e-01 5.61958909e-01 -4.51558381e-02
-6.30277932e-01 -7.18778372e-01 -4.04278487e-01 -1.47178575e-01
-7.46568501e-01 1.31743029e-01 7.32164621e-01 3.71839613e-01
-8.44099462e-01 1.01814136e-01 -6.87708497e-01 -4.87819374e-01
7.00876892e-01 3.11697423e-01 -4.90195394e-01 -5.63635409e-01
-1.51111042e+00 7.09206581e-01 -2.91491505e-02 7.70283520e-01
-7.80316710e-01 -9.49976087e-01 -7.81097949e-01 -6.55512884e-02
3.70296478e-01 -9.98907566e-01 5.25232077e-01 -3.21240157e-01
-7.67278910e-01 3.45531106e-02 -8.15711737e-01 -2.93572068e-01
4.22055572e-01 1.45928562e-01 -8.25079322e-01 -4.91420567e-01
4.56573784e-01 4.25724387e-01 3.71806473e-01 -1.13738167e+00
-4.17914271e-01 -6.37701750e-01 2.67650206e-02 -7.63894245e-02
-7.32972811e-04 -3.72053504e-01 -2.04947427e-01 -5.72030067e-01
1.29099533e-01 -9.39605474e-01 -3.20730716e-01 -2.62197793e-01
-8.28584909e-01 -2.97168136e-01 1.05893588e+00 -8.84377599e-01
1.17384410e+00 -1.72988319e+00 -3.10276467e-02 3.44847322e-01
3.79376411e-01 1.38645053e-01 -4.69282329e-01 6.93548739e-01
-5.49714975e-02 6.62479177e-02 -1.71440452e-01 -5.73486805e-01
8.75467882e-02 6.95401609e-01 -1.82843834e-01 2.05615610e-01
-3.01000830e-02 1.09324181e+00 -6.49616599e-01 -2.55466521e-01
5.92228353e-01 6.77986026e-01 -3.03903788e-01 -3.94857153e-02
2.16792319e-02 7.80304432e-01 -6.29730642e-01 6.73909962e-01
1.32181203e+00 1.04431458e-01 -4.60454971e-02 -5.44518232e-01
-1.67346448e-01 2.90851351e-02 -1.08662140e+00 1.63960385e+00
-4.41303939e-01 4.16764349e-01 -1.51551086e-02 -1.30204701e+00
1.09430921e+00 1.57703623e-01 8.07029486e-01 -1.12566209e+00
-1.62304237e-01 -1.28498122e-01 -2.76966274e-01 -6.54226840e-01
5.75902462e-01 1.35431886e-01 -5.35535440e-02 3.58376265e-01
-3.63618761e-01 6.60425127e-01 7.28861988e-02 4.27293867e-01
1.11913919e+00 2.19410062e-02 -7.10628211e-01 1.18677102e-01
4.35246706e-01 2.65763372e-01 1.01064181e+00 7.24601746e-01
-7.99499676e-02 4.78465706e-01 4.15562570e-01 -8.14452350e-01
-1.03508568e+00 -1.03911805e+00 7.52700027e-03 6.86870694e-01
1.30056992e-01 -1.73143327e-01 -4.98952866e-01 -4.43813026e-01
4.61243629e-01 6.36770546e-01 -7.01416790e-01 -5.77504523e-02
-4.31545466e-01 -8.57896924e-01 6.88890636e-01 4.97167647e-01
7.28453815e-01 -6.85772300e-01 7.04344571e-01 1.34951457e-01
-1.12210464e+00 -1.40065145e+00 -2.10687742e-01 -6.92789614e-01
-8.89725685e-01 -1.11648548e+00 -9.28567275e-02 -3.34568083e-01
7.54637122e-01 7.04006910e-01 7.94629037e-01 8.74906704e-02
5.42962626e-02 7.53810033e-02 -2.39552930e-01 -1.46593645e-01
1.19517155e-01 2.21639618e-01 6.12688735e-02 3.59197319e-01
8.48067105e-01 -9.63323593e-01 -4.63567376e-01 7.38414764e-01
-7.32310236e-01 3.89780045e-01 3.71190995e-01 4.73260403e-01
4.33645517e-01 3.18744540e-01 8.66115272e-01 -7.79661059e-01
5.32191575e-01 -1.38074374e+00 -3.31802249e-01 1.14767700e-01
-3.82188916e-01 -2.54291117e-01 4.90816623e-01 -1.13043010e-01
-1.12155616e+00 -1.66060522e-01 -1.52818829e-01 -5.76078176e-01
-4.42700684e-01 8.29679966e-01 -6.31617188e-01 2.35256955e-01
-4.07597562e-03 1.25482485e-01 1.56665280e-01 -5.75053215e-01
2.98875839e-01 8.75490248e-01 3.69207114e-01 -6.91581607e-01
7.83644080e-01 5.87216794e-01 2.39181057e-01 -7.49160707e-01
-8.19692671e-01 -3.12403828e-01 -9.22970653e-01 -4.64453131e-01
5.24011970e-01 -1.28267276e+00 -8.37635517e-01 6.68473005e-01
-1.18837059e+00 -2.66915470e-01 -8.22766945e-02 6.34406269e-01
-4.39544499e-01 1.10113263e-01 -3.84070396e-01 -7.65170932e-01
-1.40390859e-03 -9.78977144e-01 8.42218518e-01 -1.82029963e-01
8.61898735e-02 -1.15134358e+00 -8.08085948e-02 1.17626297e+00
7.15068579e-01 3.98694485e-01 8.77259076e-01 -1.80357233e-01
-8.84970427e-01 -1.96211055e-01 -7.20968306e-01 1.02888875e-01
3.54866415e-01 -1.35802642e-01 -6.47603154e-01 1.74449190e-01
-5.34404039e-01 2.26310462e-01 8.28903258e-01 7.25955963e-01
1.43972373e+00 -5.68022311e-01 -3.67728263e-01 6.77375495e-01
1.27411520e+00 -5.12850843e-02 1.05542219e+00 9.77686569e-02
1.32570493e+00 7.15044200e-01 5.15952349e-01 4.85145926e-01
1.37621641e+00 5.55848181e-01 5.46047986e-01 -2.37600207e-01
-2.49854326e-01 -7.26431608e-01 1.05917156e-01 8.24338853e-01
-2.26278007e-01 -4.08356458e-01 -9.01279032e-01 9.41674054e-01
-2.50687432e+00 -1.29805136e+00 -9.29047167e-01 1.90149724e+00
2.03409977e-02 -2.81884134e-01 3.10314387e-01 2.06542648e-02
8.88964951e-01 1.16681144e-01 -6.95335031e-01 -2.19035342e-01
-2.92264432e-01 -2.87605435e-01 6.97367251e-01 5.37346542e-01
-7.37722516e-01 8.41096342e-01 5.03414488e+00 8.07366788e-01
-5.45084834e-01 3.57852042e-01 4.10737187e-01 3.84923182e-02
-5.48226297e-01 2.32227907e-01 -6.44143224e-01 8.44615340e-01
1.27879763e+00 7.28504211e-02 7.13237286e-01 2.23757952e-01
1.15658474e+00 1.31691948e-01 -5.98712265e-01 8.03714871e-01
-3.64170432e-01 -1.61855638e+00 2.41188586e-01 3.14709723e-01
7.97211826e-01 4.85275835e-01 3.66973155e-03 2.31571108e-01
6.27725661e-01 -9.87628400e-01 1.79490983e-01 1.14713514e+00
5.87432146e-01 -8.48485708e-01 8.18414330e-01 3.36186796e-01
-1.41594672e+00 -1.80722073e-01 -5.66198587e-01 -3.29046518e-01
5.62919199e-01 9.67013240e-01 -3.43014896e-01 1.07753611e+00
6.40286803e-01 1.47914445e+00 -3.57112765e-01 8.82991314e-01
1.20921776e-01 9.25907314e-01 -3.67454559e-01 5.28830647e-01
2.07093745e-01 -8.05886388e-01 3.93434614e-01 8.04219604e-01
6.15667880e-01 1.09882236e-01 1.50883853e-01 1.02348483e+00
-2.11746737e-01 -4.36102778e-01 -9.98294473e-01 2.61015236e-01
7.14803219e-01 1.04305363e+00 1.30902231e-01 -2.82802254e-01
-5.52357197e-01 3.03682715e-01 2.37824485e-01 8.86641860e-01
-1.05151713e+00 -1.01710185e-01 1.10099173e+00 4.98325527e-01
5.82444854e-02 -4.71402973e-01 -5.88991106e-01 -1.18738341e+00
2.58313883e-02 -3.12602878e-01 3.05400848e-01 -1.27249014e+00
-1.56852961e+00 2.25961611e-01 1.03703499e-01 -1.25954485e+00
1.85672745e-01 -1.03207760e-01 -6.40236855e-01 1.01619327e+00
-1.59575891e+00 -1.91539490e+00 -6.19609237e-01 1.23971164e+00
1.65609270e-01 -2.65535623e-01 3.78780365e-01 8.73983383e-01
-1.03254497e+00 2.51705289e-01 4.31875408e-01 3.65219980e-01
2.31112197e-01 -4.74862039e-01 5.82698166e-01 8.67319524e-01
-3.77749741e-01 6.01236939e-01 3.79987746e-01 -1.14269626e+00
-1.71678257e+00 -1.78256953e+00 1.17178059e+00 -4.37732697e-01
7.75215089e-01 -2.78165966e-01 -7.86251426e-01 1.09163392e+00
-4.73733395e-02 2.12049797e-01 5.37754834e-01 4.38871086e-01
-3.82911503e-01 -7.44274139e-01 -1.28095305e+00 2.25001499e-01
1.40515172e+00 -5.28661728e-01 -1.32382080e-01 5.20961106e-01
7.24674702e-01 8.87990966e-02 -1.13402224e+00 3.91683757e-01
5.47490835e-01 -5.89740098e-01 1.02291465e+00 -7.15878606e-01
3.40053529e-01 -6.30963981e-01 -3.77938449e-01 -1.28544378e+00
-4.70580697e-01 -1.52066693e-01 2.19836924e-03 1.34595323e+00
3.10709983e-01 -8.37196171e-01 8.68073523e-01 1.02548027e+00
-2.34367549e-01 -1.20335750e-01 -1.26344824e+00 -6.01660550e-01
4.20708545e-02 -8.91447425e-01 1.34094059e+00 1.07604110e+00
-3.27579677e-01 1.45436302e-01 -9.78199720e-01 4.94060338e-01
1.07095432e+00 -6.04411773e-02 1.13643646e+00 -1.33125567e+00
6.80206299e-01 1.93319365e-01 -4.93068963e-01 -7.70874143e-01
4.76470411e-01 -8.68166149e-01 -5.81790864e-01 -1.91273022e+00
1.79494768e-01 -8.24012220e-01 -3.63402873e-01 7.59456456e-01
3.89817774e-01 9.80524197e-02 -6.95936754e-02 1.63474977e-01
-5.41740596e-01 9.28415775e-01 1.16045916e+00 -4.72398192e-01
6.26348257e-02 9.75258723e-02 -6.93385541e-01 2.03721076e-01
1.02024055e+00 -5.96285880e-01 -5.92931271e-01 -9.63941693e-01
1.87725440e-01 4.24887419e-01 8.35292220e-01 -9.25695062e-01
6.19079649e-01 -4.75059599e-01 2.36790478e-01 -1.19735432e+00
4.51122046e-01 -1.16858411e+00 8.98395658e-01 -2.75616813e-02
-9.60684121e-02 3.12289953e-01 1.23822331e-01 7.91379809e-01
-2.90019453e-01 6.57203436e-01 -1.54680580e-01 1.25745937e-01
-6.44229949e-01 8.29096854e-01 -4.98709470e-01 -1.08131178e-01
7.29921579e-01 -4.82611507e-01 -6.98497713e-01 -4.54016060e-01
-8.55986416e-01 6.86592221e-01 8.01095665e-02 8.33474159e-01
8.92065525e-01 -1.69104826e+00 -9.28074062e-01 4.17786688e-01
8.73571932e-02 -9.72053334e-02 1.04228103e+00 1.20494783e+00
5.94522804e-02 4.74740386e-01 -3.30488920e-01 -4.50995386e-01
-6.94388688e-01 5.22124529e-01 1.00893766e-01 1.64668858e-01
-4.79181945e-01 2.81786155e-02 -1.69278666e-01 -1.16199315e+00
-1.50181115e-01 -2.30311438e-01 -3.70984748e-02 -8.17191303e-02
1.00467004e-01 9.16454732e-01 1.08010642e-01 -1.15390038e+00
-2.91619450e-01 4.99219447e-01 4.23220098e-01 3.16134751e-01
1.56243646e+00 -8.13002527e-01 -3.50916266e-01 2.58804172e-01
1.10127103e+00 -5.77469468e-01 -1.00236738e+00 -5.12130797e-01
-4.01685685e-01 -8.18677127e-01 -3.50317284e-02 -8.52541625e-01
-1.55754316e+00 9.49108660e-01 2.45911673e-01 -2.32559323e-01
9.21837986e-01 -4.43209589e-01 1.22238696e+00 1.45235062e-01
6.44755125e-01 -8.83574903e-01 -6.58545077e-01 5.84806561e-01
6.77012920e-01 -1.33263969e+00 -3.41996282e-01 -3.88348997e-01
-7.40999460e-01 6.73537135e-01 7.58355677e-01 1.51054831e-02
1.07301688e+00 5.14421202e-02 -3.07684273e-01 -1.53706342e-01
-8.43985617e-01 -3.05749565e-01 2.52562240e-02 1.00578058e+00
-2.13965755e-02 4.35635567e-01 1.45068407e-01 5.99712253e-01
-1.05033882e-01 5.68248630e-01 5.35372019e-01 3.16782326e-01
3.19910161e-02 -1.04102385e+00 -3.34912866e-01 7.32541680e-01
2.14079231e-01 9.57992896e-02 1.24366004e-02 5.68506896e-01
4.31409121e-01 1.53214407e+00 1.17658600e-01 -7.89570451e-01
4.68044043e-01 -2.52887905e-01 -2.03845426e-01 1.13876909e-01
-1.48616463e-01 -5.09458840e-01 3.46933961e-01 -7.73727953e-01
-5.67818642e-01 -6.15662515e-01 -9.07085955e-01 -1.33624411e+00
1.40968755e-01 3.65942389e-01 7.16915250e-01 9.99717236e-01
8.57277215e-01 4.66824979e-01 5.81644356e-01 -5.92114687e-01
3.50140244e-01 -8.75001729e-01 -5.56174695e-01 5.48202932e-01
3.75462383e-01 -8.64190817e-01 -1.01691842e-01 -3.03838193e-01] | [6.547829627990723, 2.0693421363830566] |
53b42db9-bb59-4376-a084-ab1b4e0961db | event-driven-two-stage-solution-to-non | 2107.12582 | null | https://arxiv.org/abs/2107.12582v1 | https://arxiv.org/pdf/2107.12582v1.pdf | Event-driven Two-stage Solution to Non-intrusive Load Monitoring | Existing methods of non-intrusive load monitoring (NILM) in literatures generally suffer from high computational complexity and/or low accuracy in identifying working household appliances. This paper proposes an event-driven Factorial Hidden Markov model (eFHMM) for multiple appliances with multiple states in a household, aiming for low computational complexity and high load disaggregation accuracy. The proposed eFHMM decreases the computational complexity to be linear to the event number, which ensures online load disaggregation. Furthermore, the eFHMM is solved in two stages, where the first stage identifies state-changing appliance using transient signatures and the second stage confirms the inferred states using steady-state signatures. The combination of transient and steady-state signatures, which are extracted from transient and steady periods segmented by detected events, enhances the uniqueness of each state transition and associated appliances, which ensures accurate load disaggregation. The event-driven two-stage NILM solution, termed as eFHMM-TS, is naturally fit into an edge-cloud framework, which makes possible the real-world application of NILM. The proposed eFHMM-TS method is validated on the LIFTED and synD datasets. Results demonstrate that the eFHMM-TS method outperforms other methods and can be applied in practice. | ['Zuyi Li', 'Jiayu Han', 'Wei Tian', 'Lei Yan'] | 2021-07-27 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 1.28679797e-01 -2.97643691e-01 -2.78155565e-01 -2.17345148e-01
-4.05879200e-01 -1.10100985e-01 3.26265514e-01 1.73759967e-01
3.13984573e-01 6.06795669e-01 -1.85363635e-01 -6.04880787e-02
-2.72834808e-01 -8.32012177e-01 -2.21953183e-01 -9.09490526e-01
-1.56694576e-01 4.17663336e-01 2.94486344e-01 4.31948364e-01
-3.56205106e-01 1.37692332e-01 -1.80202746e+00 -6.70472085e-02
8.90931189e-01 1.21770775e+00 4.46463794e-01 3.75850290e-01
8.89461935e-02 7.82315075e-01 -4.67414737e-01 2.25462615e-02
3.20496112e-02 -4.21121120e-01 -4.90649879e-01 4.18689668e-01
-6.75301909e-01 -5.03811479e-01 1.04269505e-01 9.18671787e-01
4.34439480e-01 1.11036584e-01 3.48542064e-01 -1.92584825e+00
-1.05643021e-02 4.40330863e-01 -8.04661810e-01 1.81711271e-01
6.33986592e-01 2.15570301e-01 7.59194732e-01 -3.04704487e-01
1.53262675e-01 9.82439637e-01 7.09442139e-01 -6.43663630e-02
-1.47166657e+00 -7.39989042e-01 2.21095771e-01 5.75466633e-01
-1.69870591e+00 -2.66959250e-01 7.33302355e-01 -3.53450865e-01
1.29923522e+00 5.91073453e-01 6.88884497e-01 6.28544271e-01
1.16022162e-01 1.03269863e+00 1.22168756e+00 -4.19361293e-01
5.88289022e-01 6.76833512e-03 2.93889999e-01 1.05212018e-01
4.86735582e-01 -1.27593771e-01 -2.31594160e-01 -4.47213322e-01
2.88847983e-01 5.66319227e-01 -6.00838363e-02 1.44410044e-01
-7.57801056e-01 5.84157944e-01 -3.03655207e-01 4.41825479e-01
-9.47527051e-01 -3.71304929e-01 6.19608521e-01 -1.91864341e-01
4.97695357e-01 -3.30205381e-01 -4.09843028e-01 -1.63467139e-01
-1.30564797e+00 -1.74879149e-01 8.39231133e-01 1.15010214e+00
8.04769039e-01 2.55366750e-02 -2.29020968e-01 3.22574675e-01
4.59495544e-01 4.61369097e-01 3.05018425e-01 -7.24991143e-01
1.40845627e-01 9.53194678e-01 2.10411578e-01 -7.80186415e-01
-6.79054022e-01 -5.08834608e-04 -1.08813715e+00 -2.60060579e-01
-2.55359113e-01 -2.22687989e-01 -4.25896823e-01 1.63414669e+00
8.14301789e-01 5.93147814e-01 -4.18136381e-02 4.64275926e-01
2.96915054e-01 1.00629961e+00 3.32278550e-01 -1.03271961e+00
1.53832197e+00 -3.57264578e-01 -1.28306973e+00 3.69391829e-01
3.70634019e-01 -5.95870852e-01 5.79339981e-01 2.68460095e-01
-9.32394028e-01 -2.71873236e-01 -6.86376035e-01 6.07743800e-01
-2.90700316e-01 1.06458232e-01 4.21651602e-01 4.67503518e-01
-5.94772935e-01 3.29401076e-01 -1.05918133e+00 -6.83919370e-01
1.64187580e-01 4.72740293e-01 2.13571295e-01 1.92525804e-01
-1.10770833e+00 6.46372676e-01 8.09694827e-01 2.72703141e-01
-6.51324153e-01 -5.65307856e-01 -7.01573491e-01 2.43980363e-01
3.36611688e-01 -3.49720955e-01 1.33479619e+00 -3.63348156e-01
-1.43831968e+00 3.02817166e-01 -3.88669431e-01 -3.65902066e-01
3.60889673e-01 2.63783813e-01 -1.02610159e+00 5.36294021e-02
2.00736701e-01 -2.06938222e-01 6.57554388e-01 -1.15403008e+00
-1.00078607e+00 -4.77810323e-01 -3.80058765e-01 -5.25729880e-02
-2.60323763e-01 6.21135794e-02 -2.04075336e-01 -2.70073533e-01
-2.40141153e-02 -8.76302838e-01 1.92626454e-02 -8.10076237e-01
-6.99913979e-01 -4.72734332e-01 1.34152865e+00 -9.24432993e-01
1.74675369e+00 -2.12703371e+00 -4.33971524e-01 4.07448262e-01
-2.58066446e-01 2.60758638e-01 5.07599711e-01 8.85949969e-01
1.76172704e-03 -4.77329403e-01 -1.38839006e-01 -5.74886084e-01
4.23076630e-01 4.52833414e-01 -1.04528412e-01 5.74807405e-01
-6.82146475e-02 7.80851305e-01 -9.57452416e-01 -7.25232363e-01
9.67295110e-01 5.02317727e-01 8.51020142e-02 3.00802916e-01
-2.38153130e-01 4.45745707e-01 -2.60915935e-01 7.69932508e-01
8.55804086e-01 -3.99474204e-01 5.83445549e-01 -1.70879632e-01
-3.40268433e-01 1.06060773e-01 -1.72766280e+00 9.79774475e-01
-5.62055349e-01 1.38798326e-01 6.76340461e-02 -8.48088384e-01
5.97758532e-01 8.44266951e-01 9.82931733e-01 -5.73709488e-01
1.80582404e-01 5.99783510e-02 -5.66293478e-01 -5.05838156e-01
2.00999856e-01 8.43421742e-02 -8.62927139e-02 5.56253076e-01
-1.16299190e-01 4.27252501e-01 3.24787796e-01 -9.43402126e-02
9.87382054e-01 -3.69804241e-02 6.67418182e-01 -3.26552749e-01
4.00268018e-01 -3.38853806e-01 1.00788379e+00 1.92241430e-01
-3.10603082e-01 -3.20566058e-01 1.13606215e-01 -2.22121581e-01
-5.20717502e-01 -7.87324905e-01 -1.01237513e-01 7.73468435e-01
1.90572485e-01 -4.42915827e-01 -7.23229289e-01 -3.03940356e-01
-1.19935740e-02 1.16365075e+00 -3.03836763e-01 -7.59916157e-02
-1.85085624e-01 -1.16486239e+00 -1.37310931e-02 3.61450613e-01
7.85587668e-01 -1.13321722e+00 -9.91116166e-01 7.50668764e-01
-3.03621620e-01 -1.06351757e+00 -2.95627475e-01 3.58539432e-01
-6.21555984e-01 -9.74517286e-01 -1.70567870e-01 -4.80599284e-01
5.08713126e-01 2.47005343e-01 9.32789505e-01 -3.19866627e-01
-3.01921606e-01 2.88838387e-01 -2.09842503e-01 -4.60443854e-01
-2.98573554e-01 7.76751190e-02 2.45077088e-01 5.76113164e-01
9.69175160e-01 -9.24017787e-01 -6.67871058e-01 3.68301600e-01
-9.56964731e-01 1.69011746e-02 3.98027986e-01 3.51084232e-01
7.34764218e-01 1.00661969e+00 7.90380716e-01 -6.48171484e-01
1.99271798e-01 -7.89617181e-01 -1.00690675e+00 2.98962802e-01
-1.15060091e+00 -4.14962232e-01 6.29828334e-01 -4.42222834e-01
-1.25826657e+00 3.85723382e-01 -1.92880314e-02 -4.55911219e-01
-2.93958306e-01 9.79605168e-02 -4.98223424e-01 6.85270071e-01
-3.26740533e-01 5.06894946e-01 -4.31246996e-01 -6.77052498e-01
-6.27529547e-02 7.48208821e-01 5.65249860e-01 -1.61018193e-01
4.72478569e-01 4.08251524e-01 6.77746087e-02 -6.87225819e-01
-3.87508035e-01 -8.48478079e-01 -4.42896247e-01 -4.19705987e-01
8.66584420e-01 -9.46069300e-01 -1.46549582e+00 7.71089911e-01
-8.89235973e-01 -1.70893013e-01 -4.70178992e-01 3.30035359e-01
-4.58377302e-01 3.19470614e-01 -9.00598645e-01 -1.57251668e+00
-4.30721402e-01 -9.57371473e-01 1.22927833e+00 3.87019664e-01
-5.22605538e-01 -9.66707468e-01 -6.34603016e-03 2.34658159e-02
1.92493558e-01 5.88469565e-01 8.72089744e-01 -4.56578583e-01
-5.04924178e-01 -4.05915141e-01 2.72465739e-02 -1.17067970e-01
5.17013252e-01 4.12179530e-02 -9.96404409e-01 -5.62023282e-01
2.41278484e-01 3.59521329e-01 -7.89845549e-03 5.38790107e-01
9.06619549e-01 -5.90484083e-01 -5.37849486e-01 6.62122965e-02
1.71021199e+00 5.86257458e-01 5.49610734e-01 1.59449667e-01
3.41820776e-01 3.08296412e-01 6.64192200e-01 1.12147474e+00
7.25687802e-01 6.34509385e-01 4.96202916e-01 -2.13870972e-01
3.28823060e-01 -4.16061729e-01 4.35047925e-01 1.16935110e+00
2.07343698e-01 -2.18722284e-01 -5.15971780e-01 7.36551642e-01
-2.20413733e+00 -1.10265160e+00 -3.64191443e-01 2.11907101e+00
6.23429894e-01 4.26382609e-02 4.76600140e-01 7.88566828e-01
1.07220316e+00 -1.49139717e-01 -8.30848992e-01 -2.77871728e-01
1.38160110e-01 -8.01000744e-02 3.81306648e-01 1.34048671e-01
-7.74922252e-01 1.00003503e-01 5.79728603e+00 7.95288563e-01
-6.00235105e-01 5.28701544e-01 3.94363731e-01 1.06297880e-01
2.99669296e-01 -3.44407521e-02 -9.64894354e-01 1.15526402e+00
1.41512227e+00 -2.79412895e-01 4.83783394e-01 8.45499337e-01
8.47397923e-01 -5.99615574e-01 -1.12374067e+00 1.10544801e+00
-5.54281831e-01 -7.03672647e-01 -5.13087511e-01 2.81889200e-01
8.15487027e-01 -2.90676266e-01 -4.16960299e-01 1.50081158e-01
3.98307025e-01 -2.08289802e-01 5.02874613e-01 4.82625067e-01
5.21001220e-01 -9.39406455e-01 7.42767572e-01 6.05163991e-01
-1.93243623e+00 -3.45929414e-01 3.23551059e-01 4.76139709e-02
5.56926489e-01 1.06469917e+00 -6.76518798e-01 8.91180873e-01
8.45777988e-01 4.73920852e-01 -6.25364259e-02 6.77931428e-01
1.54408172e-01 8.81121099e-01 -7.75854409e-01 2.03801498e-01
-2.35282183e-01 -3.72395605e-01 3.06029081e-01 1.13225138e+00
2.46680632e-01 2.19891801e-01 5.56381881e-01 7.87695587e-01
4.42428380e-01 -1.38298854e-01 -2.31894106e-01 2.00615823e-01
7.86218107e-01 1.32834208e+00 -1.01660848e+00 -6.33872509e-01
-5.07825017e-01 1.04118788e+00 -4.33488697e-01 2.05211326e-01
-9.76273537e-01 -2.62452185e-01 5.13560534e-01 1.19776763e-01
2.63804078e-01 1.43404245e-01 -6.55716211e-02 -1.00352645e+00
7.77711067e-03 -5.71473241e-01 6.47440672e-01 -5.40480137e-01
-1.32749176e+00 1.94487378e-01 2.45726585e-01 -1.26227450e+00
-4.89324450e-01 2.73441374e-01 -8.85798335e-01 7.28093386e-01
-1.35943818e+00 -1.18212855e+00 -2.97121942e-01 8.55604649e-01
5.11998713e-01 5.38240075e-01 8.89583886e-01 7.57570624e-01
-1.28088593e+00 2.58014202e-01 2.93388873e-01 -3.68262619e-01
-6.62661344e-02 -1.15717435e+00 2.04033688e-01 1.01196373e+00
-5.20488262e-01 6.59016669e-02 8.28620315e-01 -8.84467721e-01
-1.40026712e+00 -1.50303411e+00 9.98088896e-01 -2.43963413e-02
4.67636734e-01 -3.49733412e-01 -8.30317974e-01 8.29541683e-01
1.80115908e-01 -2.64844775e-01 7.07541466e-01 -2.51870215e-01
3.29940408e-01 -3.22783798e-01 -1.46905339e+00 5.00602685e-02
5.20453334e-01 -5.42857885e-01 -2.07465947e-01 4.94525373e-01
3.97653401e-01 1.80834889e-01 -1.28476489e+00 3.59558344e-01
1.64417893e-01 -8.11670661e-01 4.44151342e-01 8.33118707e-02
-5.69373369e-01 -6.09851480e-01 -3.40097100e-01 -9.72211421e-01
-5.37962675e-01 -1.10087919e+00 -9.87174869e-01 1.99872386e+00
-2.11496413e-01 -8.87314677e-01 2.04134986e-01 7.75033116e-01
1.51559606e-01 -5.32698691e-01 -1.11727798e+00 -1.03426361e+00
-9.78994250e-01 -4.35634226e-01 1.29964757e+00 9.12922382e-01
8.38943347e-02 1.23010606e-01 -3.99702549e-01 5.19635677e-01
8.29400778e-01 4.92982566e-01 4.00261164e-01 -1.25434732e+00
-1.74414754e-01 8.90769344e-03 -2.84139514e-01 -5.48136353e-01
-6.73372597e-02 -4.05600250e-01 1.32669717e-01 -1.52349639e+00
4.71957862e-01 -1.20793149e-01 -5.00897586e-01 5.67642212e-01
-1.11759767e-01 1.99438885e-01 1.26408458e-01 3.11047971e-01
-7.08304107e-01 5.42391062e-01 1.52783662e-01 8.42668265e-02
-5.47560692e-01 3.20593834e-01 -9.14717838e-03 7.18432546e-01
8.61939728e-01 -3.46843421e-01 -3.61007601e-01 3.60919595e-01
-1.75713301e-01 1.04868859e-01 1.90475062e-01 -9.48651791e-01
1.24747567e-01 -2.61205286e-01 3.48421991e-01 -1.27771592e+00
2.67945647e-01 -1.34835720e+00 1.19047773e+00 8.27727079e-01
4.23956454e-01 3.03508461e-01 5.57058044e-02 5.84407926e-01
1.17034391e-01 1.08807117e-01 6.75272763e-01 6.45061880e-02
-7.25906134e-01 3.60219538e-01 -5.07863343e-01 -6.56731367e-01
1.55619848e+00 -1.71642110e-01 1.60963148e-01 -2.43402451e-01
-7.09915996e-01 4.70036238e-01 2.15629533e-01 1.58579394e-01
-1.15145087e-01 -1.56463575e+00 -2.50173301e-01 1.59163892e-01
-1.05818316e-01 -2.26901636e-01 4.86088872e-01 1.19726682e+00
2.05014944e-01 4.82567102e-01 3.20026815e-01 -8.98941696e-01
-1.09615064e+00 1.10100007e+00 -4.85077873e-02 -5.78973055e-01
-7.09139049e-01 -2.43114695e-01 -1.80182129e-01 1.63413495e-01
1.23487145e-01 -3.57648760e-01 8.70368779e-02 4.06475931e-01
5.11234999e-01 1.06754589e+00 2.35022664e-01 -8.91215801e-01
-4.31778610e-01 3.38349015e-01 4.06359613e-01 2.70319611e-01
1.28299975e+00 -7.76001453e-01 -2.45566741e-01 7.92224646e-01
1.10033035e+00 -6.03018761e-01 -1.08177733e+00 -1.75754368e-01
2.32144088e-01 -1.05135411e-01 1.65467590e-01 -5.21672606e-01
-1.03358233e+00 3.08696151e-01 9.27086949e-01 8.64861369e-01
1.79106092e+00 -1.89222768e-01 1.44137526e+00 -3.60099554e-01
6.59183860e-01 -1.22225463e+00 -7.83600092e-01 -2.02086538e-01
8.30654353e-02 -9.06848490e-01 -4.84671742e-02 -3.30688149e-01
-2.61449695e-01 5.92605233e-01 4.17350948e-01 3.48672509e-01
8.52606416e-01 4.71471786e-01 -5.36855876e-01 -5.57770841e-02
-7.39754021e-01 -3.61225575e-01 -3.06315958e-01 3.78108352e-01
-2.56955445e-01 6.57653153e-01 -3.38037848e-01 7.33533919e-01
1.79046601e-01 4.12025154e-01 -1.01180203e-01 1.27188170e+00
-3.28541994e-02 -7.69745588e-01 -5.60417354e-01 5.71014941e-01
-3.57081175e-01 4.55836058e-01 2.60090351e-01 5.28333008e-01
4.74606633e-01 1.42220616e+00 1.52073622e-01 -2.58356392e-01
4.45718646e-01 4.60834116e-01 4.27498445e-02 -2.46762395e-01
-5.13534069e-01 4.23801780e-01 -2.97454000e-01 -6.05712414e-01
-4.50413734e-01 -9.67682958e-01 -1.22691000e+00 -6.90283716e-01
-7.87224889e-01 1.87924325e-01 5.17196894e-01 9.41372097e-01
6.34007812e-01 5.32926083e-01 1.21041834e+00 -9.56679821e-01
-4.15896177e-01 -9.99634266e-01 -9.38057661e-01 5.04351497e-01
1.71362549e-01 -7.09054172e-01 -3.06243479e-01 1.96368903e-01] | [5.998030662536621, 2.602268695831299] |
bfbffeb6-4621-4154-9e7d-7f2f9a106b48 | large-scale-real-time-personalized-similar | 2004.05716 | null | https://arxiv.org/abs/2004.05716v1 | https://arxiv.org/pdf/2004.05716v1.pdf | Large-scale Real-time Personalized Similar Product Recommendations | Similar product recommendation is one of the most common scenes in e-commerce. Many recommendation algorithms such as item-to-item Collaborative Filtering are working on measuring item similarities. In this paper, we introduce our real-time personalized algorithm to model product similarity and real-time user interests. We also introduce several other baseline algorithms including an image-similarity-based method, item-to-item collaborative filtering, and item2vec, and compare them on our large-scale real-world e-commerce dataset. The algorithms which achieve good offline results are also tested on the online e-commerce website. Our personalized method achieves a 10% improvement on the add-cart number in the real-world e-commerce scenario. | ['Li Chen', 'Jing Gao', 'Yan Huang', 'Dong Li', 'Zhi Liu'] | 2020-04-12 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [-4.31970984e-01 -8.62113714e-01 -5.04917502e-01 -6.64178848e-01
-4.60047573e-01 -6.35131419e-01 2.27389291e-01 7.71791935e-02
-3.84083271e-01 -1.43025130e-01 5.37702918e-01 -2.49143884e-01
-3.33996207e-01 -1.07799149e+00 -2.12682620e-01 -1.47303119e-01
-1.50253743e-01 5.71647406e-01 2.62590796e-01 -6.87697887e-01
6.17846310e-01 1.41920656e-01 -1.36834264e+00 7.38355994e-01
7.61136293e-01 1.07602000e+00 3.43149841e-01 4.68068153e-01
-3.35779846e-01 -2.50683818e-02 -2.68913172e-02 -1.09420061e+00
7.63232112e-01 -4.07407850e-01 -5.74403346e-01 2.45894805e-01
5.91925085e-01 -4.12110984e-01 -1.21544808e-01 1.13377631e+00
6.40379369e-01 6.87021673e-01 3.59885693e-01 -9.96134698e-01
-1.43200386e+00 5.69645345e-01 -5.17983675e-01 3.61392707e-01
5.74172139e-01 -2.99493343e-01 1.39290297e+00 -1.22242594e+00
7.29283869e-01 1.11049390e+00 7.56150067e-01 -1.90628633e-01
-7.39015639e-01 -4.01040941e-01 4.31863308e-01 6.10668659e-01
-1.26542521e+00 2.14157626e-01 5.37261546e-01 -4.18701209e-02
1.03564668e+00 1.94323838e-01 8.00746560e-01 6.42198324e-01
1.58682346e-01 8.65738034e-01 5.85979819e-01 -1.90840781e-01
1.55679202e-02 5.04876375e-01 4.28794414e-01 2.92063683e-01
2.34548748e-01 1.20162100e-01 -2.48934805e-01 -2.19126955e-01
1.18680561e+00 9.98800933e-01 2.38955513e-01 -1.07563876e-01
-9.87044215e-01 1.25028276e+00 5.84952772e-01 4.54608083e-01
-4.46490049e-01 -4.34172690e-01 3.46418977e-01 7.78460443e-01
5.16737103e-01 5.32392740e-01 -6.15297735e-01 2.73854196e-01
-5.31575978e-01 2.62614191e-01 7.96261489e-01 1.24370813e+00
4.62955475e-01 -3.85907263e-01 -2.16260552e-02 1.05563903e+00
5.32620311e-01 2.63139546e-01 8.20486426e-01 -6.38917267e-01
9.93032232e-02 2.13954911e-01 2.79281288e-01 -1.34313738e+00
-7.39666671e-02 -8.96398783e-01 -5.73529541e-01 -3.99269432e-01
2.10102916e-01 -3.72559242e-02 -2.26371557e-01 8.42717707e-01
3.76427859e-01 4.12234455e-01 -2.72325337e-01 1.26603842e+00
7.95338929e-01 7.53781140e-01 -1.21305116e-01 -3.86550367e-01
1.19884872e+00 -1.72023642e+00 -5.60909212e-01 3.20413828e-01
6.30434096e-01 -1.34971213e+00 9.68168914e-01 6.55767739e-01
-1.10432899e+00 -1.03764558e+00 -5.92018545e-01 4.17292491e-02
-4.22372788e-01 -3.89690906e-01 1.35976315e+00 8.19945157e-01
-6.17033243e-01 8.37570310e-01 -7.77437240e-02 -6.79172397e-01
5.30540310e-02 2.40115106e-01 -1.36432543e-01 -3.26513261e-01
-1.06436574e+00 6.63054347e-01 3.07678524e-02 -5.80770195e-01
-2.08358005e-01 -7.53083766e-01 -3.36210817e-01 2.19361454e-01
2.61345327e-01 -8.70174587e-01 1.10618341e+00 -1.23213804e+00
-1.44861615e+00 4.54330981e-01 1.33510664e-01 -1.82660013e-01
5.03656529e-02 -3.89746815e-01 -1.14515138e+00 -4.44960803e-01
-2.66293734e-01 1.63993686e-01 4.11511689e-01 -1.07643378e+00
-1.14625800e+00 -5.90501428e-01 2.08086699e-01 3.78487796e-01
-7.63490617e-01 3.25008869e-01 -9.02272284e-01 -7.93096066e-01
3.49981666e-01 -7.44992197e-01 -7.00440705e-01 -8.84261169e-03
4.78704840e-01 -2.24087134e-01 3.13625455e-01 -4.03496474e-01
1.18116510e+00 -2.16875887e+00 -4.71039146e-01 3.88968199e-01
-2.84566015e-01 1.18148513e-01 -7.41929889e-01 5.52258611e-01
1.18785806e-01 -1.47256285e-01 9.28844273e-01 6.90547079e-02
1.02318367e-02 9.56408083e-02 -1.58610214e-02 2.94348627e-01
-7.56040573e-01 1.12669849e+00 -8.63389552e-01 -2.39287958e-01
7.59613216e-02 2.32860669e-01 -9.24286187e-01 1.08106367e-01
1.29358470e-01 4.06217650e-02 -4.09040332e-01 4.85687256e-01
8.61703098e-01 -3.38102669e-01 3.69777948e-01 -3.70296568e-01
1.45147637e-01 1.50898352e-01 -1.44225597e+00 1.89848626e+00
-8.14898193e-01 -2.25844637e-01 -2.39172786e-01 -6.02544188e-01
9.95884717e-01 -1.44992322e-01 7.31534660e-01 -1.17197835e+00
2.55317450e-01 1.25456229e-01 -2.97391154e-02 -2.16270715e-01
1.04675877e+00 1.98607653e-01 2.91091442e-01 6.92258000e-01
1.31361037e-01 3.05084288e-01 1.06813155e-01 3.51089150e-01
6.65553689e-01 -2.33777091e-01 3.74129385e-01 -3.71704251e-01
2.08844647e-01 -4.99500856e-02 2.46844202e-01 7.11375713e-01
-9.42868665e-02 5.51446438e-01 -6.37152791e-01 -7.91916132e-01
-1.12563396e+00 -9.68040884e-01 -6.41646683e-02 1.64768922e+00
6.50930464e-01 -6.00646019e-01 -4.06672284e-02 -8.74555349e-01
3.87679219e-01 6.13767028e-01 -2.67619669e-01 9.26337689e-02
-1.46284103e-02 -4.50984329e-01 -6.85450852e-01 6.20777428e-01
3.49311501e-01 -7.11855829e-01 5.37835598e-01 5.75976253e-01
1.30258992e-01 -6.75777614e-01 -1.35891497e+00 -4.49903578e-01
-1.11229169e+00 -5.99971771e-01 -7.67886281e-01 -1.18228328e+00
6.13485575e-01 1.19786739e+00 1.43243635e+00 1.60136312e-01
8.21480006e-02 3.89753878e-01 -9.93369102e-01 -2.24284813e-01
2.90662408e-01 -4.34198886e-01 2.57884443e-01 1.88492343e-01
8.44080806e-01 -4.41537529e-01 -1.14607012e+00 1.14046276e+00
-6.13822520e-01 -5.85421324e-01 2.67940074e-01 7.77996182e-01
8.79065514e-01 3.62351358e-01 6.17197752e-01 -1.08284187e+00
7.92748570e-01 -7.43421376e-01 -3.48979890e-01 2.24145755e-01
-1.24348271e+00 -5.50438166e-01 7.40958989e-01 -8.81005883e-01
-1.06486380e+00 -2.42577851e-01 -5.43869674e-01 -9.48629230e-02
-3.34845856e-02 6.93407774e-01 -1.06195971e-01 -3.86049077e-02
4.86943066e-01 5.10821491e-02 -5.65169096e-01 -8.93663585e-01
9.38915014e-01 8.92682016e-01 3.32911313e-01 3.52226868e-02
5.72729647e-01 2.85359591e-01 -6.33239448e-01 -1.79505050e-01
-8.64816368e-01 -1.35784984e+00 -5.29088140e-01 2.14473963e-01
1.15287885e-01 -8.77585411e-01 -6.14337504e-01 1.15179792e-01
-8.50368440e-01 2.78045952e-01 -4.59605187e-01 1.07446599e+00
-2.29311615e-01 5.46328902e-01 -9.26493824e-01 -3.57525140e-01
-6.53950632e-01 -6.64369226e-01 4.06230956e-01 1.25444889e-01
8.32564570e-03 -1.11419129e+00 2.58794278e-01 5.01789570e-01
7.30291903e-01 -9.01337087e-01 4.46837813e-01 -1.24069440e+00
-2.89054722e-01 -3.26579124e-01 -3.27582985e-01 1.97123826e-01
1.89134076e-01 -3.81856441e-01 -1.23200014e-01 -3.78352493e-01
-5.55183589e-02 2.47865081e-01 5.94558418e-01 3.00510257e-01
8.90312433e-01 -2.46784940e-01 -5.58144413e-02 5.99131048e-01
1.73442543e+00 4.73544180e-01 8.34172368e-01 1.40012994e-01
4.29290235e-01 3.68836820e-01 1.18540895e+00 7.86150336e-01
4.21113223e-01 9.92243707e-01 1.49554342e-01 -6.85147420e-02
1.34120375e-01 -4.86853391e-01 1.47489056e-01 1.27059329e+00
-1.38186395e-01 -5.21360517e-01 7.16325939e-02 5.12707114e-01
-2.05534887e+00 -1.17826378e+00 -4.95097131e-01 2.31739688e+00
2.64872789e-01 -1.76115721e-01 5.27815640e-01 -3.80193740e-01
6.92484796e-01 -3.09961051e-01 -3.73793125e-01 -6.58660769e-01
-2.92883869e-02 2.10183620e-01 4.37311023e-01 5.38948290e-02
-1.14210021e+00 8.44201207e-01 6.66401148e+00 1.05917466e+00
-7.42343724e-01 5.37400067e-01 4.78338778e-01 -1.66306794e-01
-7.06508160e-01 -4.33613837e-01 -8.23783994e-01 4.76767242e-01
8.97120416e-01 -4.90942657e-01 4.99647886e-01 1.27804089e+00
1.63365528e-01 3.16032350e-01 -9.03400660e-01 1.35789323e+00
3.68866861e-01 -1.36618960e+00 1.03721648e-01 1.99840680e-01
1.25064921e+00 -7.65126050e-02 2.15296268e-01 3.38698089e-01
5.61321497e-01 -4.15201217e-01 1.30081594e-01 2.17423856e-01
3.27059388e-01 -9.25418258e-01 1.12658668e+00 6.79381415e-02
-1.30060780e+00 -1.20129079e-01 -9.03249443e-01 -1.30447552e-01
5.14520943e-01 7.36506522e-01 -1.53000221e-01 6.35519981e-01
8.71306479e-01 1.08353806e+00 -1.76544264e-01 1.48843253e+00
3.68137032e-01 4.10915494e-01 -1.78333789e-01 -3.27565402e-01
1.41525790e-01 -7.63305426e-01 -2.88762525e-02 1.08881307e+00
7.84920573e-01 3.44837606e-01 3.07283342e-01 2.34314695e-01
-1.18449241e-01 1.12049758e+00 -3.71429980e-01 4.48099077e-02
1.02576077e-01 1.56019163e+00 -6.64921820e-01 -4.27039266e-01
-1.13100338e+00 1.51121068e+00 -1.37225136e-01 -1.33715332e-01
-8.81149709e-01 -1.70320958e-01 7.58166969e-01 6.16130121e-02
1.01444983e+00 -1.66471839e-01 -1.78685591e-01 -1.07821774e+00
-4.02339041e-01 -7.96701312e-01 5.22998989e-01 -5.36577702e-01
-2.21559882e+00 3.58632743e-01 -6.87887788e-01 -1.49971843e+00
1.96482465e-01 -5.29596627e-01 -7.59888291e-01 6.18573129e-01
-1.00181222e+00 -8.42966497e-01 2.88556367e-02 1.03254402e+00
8.79392862e-01 -4.88310784e-01 8.74870002e-01 8.66432488e-01
-9.35564861e-02 1.07228041e+00 8.88326764e-01 -3.19129527e-01
8.70888054e-01 -8.03121150e-01 8.68051469e-01 4.63130116e-01
7.92685330e-01 1.00818014e+00 4.08790141e-01 -6.82699263e-01
-1.78084075e+00 -1.12880743e+00 1.00748515e+00 -3.43411028e-01
8.05879653e-01 1.11521054e-02 -5.87333083e-01 5.94550431e-01
3.53653103e-01 -7.17922002e-02 1.35439670e+00 7.76228786e-01
-6.30872250e-01 -1.94955990e-01 -1.47092712e+00 4.19700176e-01
1.55208039e+00 -1.57896936e-01 -4.76784945e-01 7.81603456e-01
7.17617512e-01 1.83114335e-01 -1.37598991e+00 -1.29722685e-01
9.82080638e-01 -1.00431716e+00 1.30364156e+00 -9.86140430e-01
3.52659762e-01 3.47374007e-02 -4.57818121e-01 -1.47935224e+00
-1.43322229e+00 -3.20137352e-01 2.41334185e-01 1.02174652e+00
5.30426145e-01 -6.39079630e-01 8.22249293e-01 3.59392107e-01
8.73081386e-02 -7.00706720e-01 -2.46796943e-03 -1.09070659e+00
-2.56006449e-01 -3.79187226e-01 9.66398656e-01 1.09278810e+00
1.34762004e-01 4.13844317e-01 -5.16148269e-01 -1.09149896e-01
4.03499395e-01 7.78069258e-01 6.86878681e-01 -1.14584577e+00
-9.47147191e-01 -2.63320595e-01 -4.06607836e-01 -1.62986851e+00
-5.81088781e-01 -9.49872136e-01 -4.68374491e-01 -1.36933887e+00
6.35594606e-01 -2.53955990e-01 -8.70561361e-01 -2.45254681e-01
-2.82282308e-02 6.44499481e-01 2.65239775e-01 2.64559150e-01
-1.24608159e+00 1.34162173e-01 1.54621565e+00 -1.34043515e-01
-2.74036139e-01 2.40439221e-01 -9.61289287e-01 3.80341142e-01
6.15606427e-01 -1.82153702e-01 -4.24739718e-01 -3.99751186e-01
3.89625669e-01 -1.80126250e-01 -5.28443813e-01 -3.88249367e-01
4.16802198e-01 -3.93745720e-01 4.09117967e-01 -6.87476158e-01
-9.16640908e-02 -1.09290218e+00 3.62816960e-01 2.63410330e-01
-3.84963691e-01 2.62586027e-01 -5.16335189e-01 4.88034070e-01
-2.14352757e-01 -4.21735346e-01 4.84852821e-01 -4.02998209e-01
-1.13217163e+00 7.22574830e-01 5.64222224e-03 -5.30913234e-01
9.99574721e-01 -1.98270664e-01 -2.40214407e-01 -7.12347925e-01
-7.79738426e-01 4.50919829e-02 3.31272244e-01 6.70705140e-01
6.45228863e-01 -1.60477757e+00 -7.97421694e-01 1.66117504e-01
3.61448944e-01 -1.18086338e+00 6.04800999e-01 4.79436696e-01
-9.69067365e-02 3.33870232e-01 -3.87682378e-01 -4.88508567e-02
-1.53447938e+00 1.34249532e+00 -2.48743862e-01 -4.08007056e-01
-1.99238822e-01 1.19824278e+00 7.39102811e-02 -4.36736494e-01
6.21830486e-02 3.51665616e-01 -4.87233728e-01 5.76747470e-02
9.71041143e-01 5.11811852e-01 1.87667608e-01 -6.07767999e-01
-3.61775085e-02 6.34881556e-01 -5.43844759e-01 3.52492303e-01
1.32200098e+00 -5.51454604e-01 2.30894461e-01 -1.11030661e-01
1.47496974e+00 2.03959852e-01 -3.63146007e-01 -6.36193693e-01
-4.41455692e-01 -1.21651924e+00 3.83981466e-01 -7.87694514e-01
-1.62185323e+00 3.01146626e-01 1.04496491e+00 2.97944725e-01
1.12512922e+00 -8.44314173e-02 1.28065705e+00 4.51272517e-01
1.21890128e+00 -1.21189070e+00 6.95457980e-02 4.43110585e-01
7.49828339e-01 -1.42146361e+00 1.53014660e-01 -5.90003848e-01
-9.50862646e-01 7.68933773e-01 4.10610139e-01 -4.92749870e-01
1.14283383e+00 -8.28588754e-02 2.20786318e-01 2.39912532e-02
-6.96014643e-01 -4.70470279e-01 3.40849668e-01 5.23227215e-01
6.77667916e-01 1.73587590e-01 -9.52016115e-01 1.04964793e+00
7.54347295e-02 2.31680438e-01 -2.34061666e-02 5.36752522e-01
-6.01208389e-01 -1.57897806e+00 -7.58153722e-02 7.27322161e-01
-3.97162139e-01 -3.38699758e-01 -4.64182124e-02 6.60471767e-02
1.50927663e-01 1.25152636e+00 3.25939924e-01 -8.88451040e-01
6.46981776e-01 -4.08315569e-01 6.17116570e-01 -6.37426555e-01
-1.18160713e+00 4.89024162e-01 -1.88350677e-04 -7.60533869e-01
-2.50604749e-01 -4.61359620e-01 -9.71625328e-01 -7.95621037e-01
-7.99079716e-01 3.27818543e-01 7.42273331e-01 4.57213193e-01
7.17942774e-01 -1.07038677e-01 1.25668883e+00 -7.49432266e-01
-5.05511403e-01 -8.50159645e-01 -1.35715795e+00 1.27937865e+00
-7.79294789e-01 -2.21276164e-01 -2.85528712e-02 -1.51450112e-01] | [10.052803993225098, 5.740522384643555] |
9d0645dc-5312-4e6f-af2b-f881ff7e4044 | sentiment-analysis-for-reinforcement-learning | 2010.02316 | null | https://arxiv.org/abs/2010.02316v1 | https://arxiv.org/pdf/2010.02316v1.pdf | Sentiment Analysis for Reinforcement Learning | While reinforcement learning (RL) has been successful in natural language processing (NLP) domains such as dialogue generation and text-based games, it typically faces the problem of sparse rewards that leads to slow or no convergence. Traditional methods that use text descriptions to extract only a state representation ignore the feedback inherently present in them. In text-based games, for example, descriptions like "Good Job! You ate the food}" indicate progress, and descriptions like "You entered a new room" indicate exploration. Positive and negative cues like these can be converted to rewards through sentiment analysis. This technique converts the sparse reward problem into a dense one, which is easier to solve. Furthermore, this can enable reinforcement learning without rewards, in which the agent learns entirely from these intrinsic sentiment rewards. This framework is similar to intrinsic motivation, where the environment does not necessarily provide the rewards, but the agent analyzes and realizes them by itself. We find that providing dense rewards in text-based games using sentiment analysis improves performance under some conditions. | ['Eve Fleisig', 'Ameet Deshpande'] | 2020-10-05 | null | null | null | null | ['text-based-games'] | ['playing-games'] | [ 5.11037670e-02 3.84208739e-01 -4.64973718e-01 -4.03973937e-01
-5.00703216e-01 -6.50431097e-01 7.09382236e-01 4.25591409e-01
-6.90064728e-01 1.16841388e+00 3.98906916e-01 -2.42364630e-01
1.32387325e-01 -9.23997223e-01 -4.18239146e-01 -5.82067907e-01
-3.24793234e-02 4.26053941e-01 -2.61738390e-01 -8.20276320e-01
3.74676466e-01 -1.08077928e-01 -1.48573482e+00 7.74136707e-02
5.10138273e-01 6.88184738e-01 3.64478290e-01 7.43431866e-01
-4.00575608e-01 1.52075982e+00 -7.54223943e-01 -6.11844771e-02
1.45136043e-01 -6.20101631e-01 -8.52585495e-01 3.87356095e-02
-5.04366755e-01 -5.93270898e-01 -8.93306583e-02 1.10297954e+00
1.34413868e-01 4.58297223e-01 3.03248674e-01 -1.30367374e+00
-5.89677513e-01 8.02778304e-01 -4.57519174e-01 -1.58518374e-01
8.06664348e-01 4.75090176e-01 1.25455403e+00 -4.82846677e-01
6.80206001e-01 1.16736102e+00 8.07579458e-02 1.04437649e+00
-1.28469479e+00 -4.81217891e-01 3.97828668e-01 -2.63402760e-01
-5.35789609e-01 -2.95668185e-01 7.21942186e-01 -3.01822692e-01
1.16964984e+00 1.30681664e-01 9.08622086e-01 1.20429075e+00
1.44428387e-01 1.19761229e+00 1.25803912e+00 -3.26105028e-01
6.69808626e-01 4.47459340e-01 -1.85005859e-01 5.53635538e-01
-1.92180693e-01 5.93000829e-01 -7.94860840e-01 -2.84205135e-02
8.09137881e-01 1.49874955e-01 1.44844651e-01 -3.91503811e-01
-1.20814192e+00 1.24006855e+00 1.89531818e-01 1.39249623e-01
-5.97048342e-01 4.28595394e-01 4.05272752e-01 7.20825911e-01
7.32474700e-02 1.22174680e+00 -5.16808033e-01 -7.31086612e-01
-6.43088281e-01 4.77696985e-01 9.25141692e-01 6.83471918e-01
1.02412796e+00 4.42420214e-01 -1.19598582e-01 6.60460413e-01
3.26416165e-01 4.21659559e-01 8.33807111e-01 -1.23952460e+00
2.79192060e-01 4.21734124e-01 5.02705634e-01 -6.37117028e-01
-5.59177518e-01 -1.02848016e-01 -3.86378646e-01 5.96541107e-01
4.26085532e-01 -9.57752109e-01 -7.20870674e-01 2.13480163e+00
1.51217848e-01 -2.80889899e-01 5.75704455e-01 1.07361972e+00
5.94383299e-01 7.17556119e-01 1.26588419e-01 -2.98215419e-01
9.80715811e-01 -7.66978443e-01 -8.99891734e-01 -7.68053889e-01
8.49395871e-01 -3.61700594e-01 1.39754939e+00 4.42545116e-01
-1.16019034e+00 -7.75802061e-02 -9.18288350e-01 1.89402536e-01
-3.53350222e-01 -2.10303634e-01 1.14121449e+00 4.47594941e-01
-1.22877884e+00 6.37300789e-01 -8.32956851e-01 -2.29260087e-01
-2.11623777e-03 6.48420393e-01 -1.75045028e-01 3.39279920e-01
-1.34780192e+00 8.96466196e-01 1.40508279e-01 -1.21319093e-01
-1.10302353e+00 -1.98329583e-01 -1.27870095e+00 1.46899045e-01
7.50532687e-01 -4.16981399e-01 1.82524133e+00 -1.40536010e+00
-2.12589192e+00 4.90999013e-01 -1.02550350e-02 -3.53514880e-01
2.59542763e-01 -1.86250746e-01 1.19109355e-01 -1.60265416e-01
2.55553603e-01 1.06820440e+00 8.97850156e-01 -1.00168622e+00
-5.40815890e-01 -2.11013064e-01 6.62091255e-01 6.91007733e-01
-3.57081383e-01 1.36403278e-01 2.17437834e-01 -2.91686088e-01
-2.44945243e-01 -8.09407413e-01 -8.33852768e-01 -2.87273914e-01
-2.98243523e-01 -2.60511935e-01 4.37093168e-01 3.08611915e-02
7.41238415e-01 -1.83329451e+00 4.08593528e-02 2.81269103e-01
2.17992216e-01 -3.06204945e-01 -3.10851127e-01 6.26916111e-01
-3.17138210e-02 1.02339149e-01 2.22568452e-01 -2.09047526e-01
9.79721621e-02 3.96820396e-01 -2.52458215e-01 -4.15060706e-02
5.74834108e-01 8.73782098e-01 -1.34454882e+00 -1.73997983e-01
1.18596859e-01 -1.41972587e-01 -9.08578753e-01 3.06824863e-01
-5.99589348e-01 3.79913658e-01 -8.27722251e-01 2.40683481e-01
-1.09949142e-01 -3.51362258e-01 3.74159157e-01 8.50990057e-01
-1.49247035e-01 6.76689267e-01 -1.10749519e+00 1.65710890e+00
-4.03220862e-01 3.41450781e-01 2.73098022e-01 -9.23951745e-01
1.02646077e+00 3.99837464e-01 5.55180192e-01 -8.71989369e-01
1.98014006e-01 -1.71510875e-01 1.52616903e-01 -4.14083004e-01
1.00479698e+00 -4.68235701e-01 -4.60321844e-01 1.03967285e+00
-5.55389225e-02 -6.99453771e-01 4.59938943e-01 6.28078878e-01
1.34010243e+00 2.94464260e-01 2.74132371e-01 5.14440127e-02
3.67907956e-02 4.96292263e-01 5.62699378e-01 9.13983941e-01
-1.22093678e-01 1.31932199e-01 9.81630921e-01 -2.69021302e-01
-9.08247054e-01 -7.32592285e-01 6.05541348e-01 1.36087418e+00
6.83540304e-04 -6.87589347e-01 -3.63422483e-01 -4.08572286e-01
-2.74287641e-01 7.34916210e-01 -6.18348598e-01 -3.12262803e-01
-1.98235676e-01 -4.83557224e-01 1.17057517e-01 5.20086169e-01
1.41909152e-01 -1.67080617e+00 -8.08824480e-01 4.67296958e-01
-1.11287897e-02 -5.90730727e-01 -2.95613557e-01 7.94804215e-01
-8.37944567e-01 -6.47786915e-01 -5.52489221e-01 -5.47935665e-01
6.83362722e-01 8.46568868e-02 1.12384689e+00 6.37112977e-03
2.40873545e-02 5.31597018e-01 -3.97481799e-01 -5.68318367e-01
-5.00808656e-01 -1.92241147e-01 2.51197487e-01 -4.49226350e-01
4.40905809e-01 -3.91482085e-01 -2.85750091e-01 -1.34969801e-01
-7.00899959e-01 5.20720556e-02 5.01772404e-01 1.16869187e+00
2.72543103e-01 -1.96370706e-01 1.05095577e+00 -1.06402791e+00
1.39792562e+00 -5.94094932e-01 -4.64224994e-01 -2.88849115e-01
-6.80493653e-01 4.42387909e-01 7.91941762e-01 -5.66021621e-01
-1.01453400e+00 3.46838087e-01 -3.23738828e-02 -1.27651185e-01
-1.88661009e-01 8.30193579e-01 3.74779373e-01 4.04885113e-01
9.20740783e-01 1.96472511e-01 3.04401994e-01 1.70775816e-01
2.77971447e-01 3.20663601e-01 1.20405808e-01 -8.71850431e-01
5.17177761e-01 -2.69491822e-02 -3.85763109e-01 -5.10614276e-01
-6.98574960e-01 -3.19978327e-01 5.87076098e-02 -3.34249675e-01
7.48429537e-01 -8.45265508e-01 -1.08066845e+00 -7.41311759e-02
-8.48446488e-01 -8.94562364e-01 -9.10072803e-01 6.21119678e-01
-1.02923810e+00 2.15284694e-02 -8.59838486e-01 -1.14504528e+00
9.43175778e-02 -1.16753924e+00 7.67465115e-01 6.95444465e-01
-6.08247995e-01 -8.33709717e-01 1.08069196e-01 -2.25326065e-02
4.80020136e-01 -1.69152647e-01 7.38981664e-01 -5.78100741e-01
-3.27735007e-01 -1.96366608e-02 2.87059546e-01 5.03806286e-02
2.23107487e-01 -1.76812217e-01 -6.80456758e-01 -1.24343522e-01
5.07159978e-02 -1.03541648e+00 4.14896131e-01 3.51502806e-01
7.41108894e-01 -4.16069627e-01 1.66248456e-01 -1.14233382e-01
1.08584368e+00 4.22526807e-01 3.33940506e-01 4.05695051e-01
1.10798009e-01 8.09755802e-01 8.65596414e-01 8.42668891e-01
4.38005835e-01 2.34385952e-01 5.12798488e-01 -2.46237680e-01
5.94482064e-01 -3.93929899e-01 9.92833376e-01 4.52021271e-01
7.44262785e-02 1.20237768e-01 -5.57893813e-01 3.52846056e-01
-2.13389683e+00 -1.13259482e+00 4.23750937e-01 2.01229835e+00
1.26252770e+00 3.18261653e-01 2.20019951e-01 -1.66005716e-02
2.65681326e-01 2.66037971e-01 -8.95860612e-01 -9.71364975e-01
1.05084762e-01 4.22953814e-01 2.40764335e-01 6.12132728e-01
-7.01936901e-01 1.32473922e+00 6.86047077e+00 2.89634049e-01
-1.21472430e+00 -1.37321189e-01 6.42211556e-01 -3.43022048e-01
-4.86345530e-01 8.73956010e-02 -4.84193236e-01 -2.14917194e-02
9.11985457e-01 -3.34768832e-01 7.83334553e-01 1.13374591e+00
5.06390929e-01 -5.32122076e-01 -1.32897151e+00 7.10012972e-01
-2.64528036e-01 -1.20236027e+00 -3.31351578e-01 6.42308518e-02
6.55598640e-01 -3.15408269e-03 1.78815544e-01 7.38910079e-01
1.28147614e+00 -1.10925198e+00 7.13456869e-01 1.30581811e-01
6.58385813e-01 -7.77099907e-01 5.29251337e-01 7.40393519e-01
-7.60452688e-01 -3.10124427e-01 -3.52797896e-01 -7.63186932e-01
-9.13532004e-02 1.04576580e-01 -1.21434033e+00 -8.66289157e-03
3.13622355e-01 6.05206788e-01 -9.39883515e-02 4.18740451e-01
-6.42243624e-01 3.78773779e-01 -2.51165777e-01 -7.44635642e-01
6.34494483e-01 -3.84381980e-01 4.34700817e-01 7.43096828e-01
9.17473584e-02 1.46687955e-01 5.27190387e-01 1.11669099e+00
8.99455696e-02 1.92866355e-01 -9.65973854e-01 -4.26133782e-01
5.89221306e-02 1.24118435e+00 -6.66531682e-01 -4.34480906e-01
-3.55033070e-01 6.84364796e-01 3.36512804e-01 4.72741455e-01
-2.99526691e-01 -2.63509482e-01 7.58522093e-01 -1.13254674e-01
-7.80650675e-02 -1.46489739e-01 -1.02055989e-01 -1.25695038e+00
-3.31952661e-01 -1.17271912e+00 1.03008933e-01 -8.84985864e-01
-9.20623720e-01 3.37209612e-01 -2.49917507e-01 -9.38625634e-01
-1.06580544e+00 -4.33401883e-01 -6.40233457e-01 6.40237510e-01
-1.46777487e+00 -3.35165083e-01 2.95281112e-01 8.60876799e-01
7.02727437e-01 -1.60238564e-01 1.10137451e+00 -2.15543494e-01
-3.83169383e-01 2.29297802e-01 -1.80002302e-01 1.97538659e-01
7.02659547e-01 -1.64061463e+00 4.22578417e-02 4.64244872e-01
7.57824183e-02 6.07101440e-01 8.82645309e-01 -6.12693906e-01
-1.69779646e+00 -5.01716495e-01 6.42854333e-01 -2.10548222e-01
8.38947654e-01 -2.51072764e-01 -4.70928639e-01 6.44248307e-01
2.41758376e-01 -3.10693711e-01 5.64675510e-01 3.99896592e-01
-5.87010831e-02 2.53907531e-01 -1.01035094e+00 9.05005395e-01
4.28070843e-01 -5.21491170e-01 -5.90254366e-01 5.57083249e-01
6.78163171e-01 -4.75800246e-01 -4.12449390e-01 -3.48845601e-01
2.17453927e-01 -7.55171776e-01 5.49808145e-01 -1.07283938e+00
7.66045034e-01 -1.18051305e-01 7.24493638e-02 -1.77467120e+00
-4.23946977e-02 -9.67501998e-01 1.53027132e-01 6.78362191e-01
7.74534464e-01 -5.83240211e-01 1.21734452e+00 1.07900417e+00
-2.65618544e-02 -6.22781992e-01 -4.51123953e-01 -4.96306926e-01
3.17273773e-02 -4.52528000e-01 5.28926551e-01 9.65695560e-01
9.35833752e-01 6.01413429e-01 -5.17554402e-01 -4.23145115e-01
1.93213508e-01 7.36709237e-02 6.39911175e-01 -1.01200390e+00
-5.11917651e-01 -3.85281146e-01 9.75882336e-02 -1.24687421e+00
1.54414013e-01 -7.37152219e-01 5.77046633e-01 -1.38695669e+00
1.05990537e-01 -7.94899881e-01 -2.19765082e-01 8.38847458e-01
3.04707885e-02 -2.52681404e-01 3.11424226e-01 -9.37474743e-02
-7.82991409e-01 6.14861548e-01 1.51062155e+00 -1.56527936e-01
-6.87717974e-01 4.52136621e-02 -1.15275121e+00 6.31052196e-01
1.07086098e+00 -5.97629964e-01 -5.77071905e-01 -2.70954072e-02
8.64436030e-01 6.89367592e-01 -4.41899411e-02 -5.40985405e-01
3.11789781e-01 -7.44033098e-01 2.77055830e-01 8.37459665e-05
6.65642798e-01 -5.19177258e-01 -4.13551748e-01 5.51637352e-01
-8.38368118e-01 2.50499129e-01 4.47913185e-02 4.46688920e-01
-3.30430537e-01 -4.94394422e-01 3.31428498e-01 -5.80456972e-01
-5.98569810e-01 8.40855315e-02 -1.05054533e+00 7.91392401e-02
9.38360214e-01 -1.43312067e-01 -2.48192534e-01 -1.29008138e+00
-8.30862939e-01 6.51485920e-01 2.98346490e-01 3.60888779e-01
7.07140326e-01 -1.12259734e+00 -4.22640681e-01 1.71273351e-02
-1.69502974e-01 9.80611593e-02 -1.38031736e-01 5.75586379e-01
3.37356515e-02 9.02624056e-02 -5.17773747e-01 -3.16238135e-01
-9.31773782e-01 4.56527054e-01 1.45837963e-01 -5.17730355e-01
-4.07088190e-01 7.76593924e-01 2.26618305e-01 -5.77308774e-01
2.68254578e-01 -3.85081261e-01 -4.95254725e-01 1.07080966e-01
4.77772772e-01 -3.40005100e-01 -3.08911502e-01 3.89655605e-02
3.16052102e-02 -1.11111730e-01 -2.49637276e-01 -7.02707469e-01
1.46501327e+00 8.32571387e-02 1.31976753e-01 5.88766396e-01
4.46038395e-01 -7.17011616e-02 -1.37519491e+00 -1.51604563e-01
7.87906162e-03 -2.94174612e-01 -2.75895316e-02 -7.90707588e-01
-8.61029446e-01 8.01749527e-01 2.29786560e-02 5.00662744e-01
7.81282902e-01 -1.63003445e-01 4.93567079e-01 1.02711916e+00
5.24600208e-01 -1.51121318e+00 6.88575864e-01 8.62177610e-01
6.44683480e-01 -1.46359384e+00 -2.26137802e-01 3.36825103e-01
-1.30527532e+00 1.04732907e+00 8.95356059e-01 -1.80685490e-01
2.74826109e-01 5.68385363e-01 8.00963938e-02 -2.03323990e-01
-1.34269297e+00 -2.87467211e-01 -6.30839050e-01 6.85273230e-01
3.98356080e-01 2.50336647e-01 -1.59454960e-02 5.74751794e-01
-5.54812312e-01 -1.52977007e-02 1.16083097e+00 1.25244391e+00
-6.45713449e-01 -1.43375802e+00 -2.57644922e-01 7.47208953e-01
-4.32288200e-01 -1.53817222e-01 -5.37135780e-01 4.44987118e-01
-4.86024559e-01 1.15206039e+00 2.94802543e-02 -4.40497577e-01
6.14267737e-02 -1.60770789e-02 2.31823802e-01 -1.20762408e+00
-1.10196149e+00 9.84908417e-02 3.10804814e-01 -8.53672326e-01
-3.66555721e-01 -7.15129197e-01 -1.80066335e+00 -2.33590782e-01
-2.29840279e-01 5.72663724e-01 7.38525093e-01 8.61443996e-01
7.43149668e-02 5.69308758e-01 7.28129148e-01 -7.23356366e-01
-8.48768055e-01 -7.49658704e-01 -6.79174364e-01 1.59089595e-01
4.46560919e-01 -5.74231327e-01 -3.05821896e-01 -1.70153081e-01] | [4.058199405670166, 1.5664467811584473] |
68b309cf-be53-417e-9e73-2594bf85c29d | experience-feedback-using-representation | 2109.13027 | null | https://arxiv.org/abs/2109.13027v1 | https://arxiv.org/pdf/2109.13027v1.pdf | Experience feedback using Representation Learning for Few-Shot Object Detection on Aerial Images | This paper proposes a few-shot method based on Faster R-CNN and representation learning for object detection in aerial images. The two classification branches of Faster R-CNN are replaced by prototypical networks for online adaptation to new classes. These networks produce embeddings vectors for each generated box, which are then compared with class prototypes. The distance between an embedding and a prototype determines the corresponding classification score. The resulting networks are trained in an episodic manner. A new detection task is randomly sampled at each epoch, consisting in detecting only a subset of the classes annotated in the dataset. This training strategy encourages the network to adapt to new classes as it would at test time. In addition, several ideas are explored to improve the proposed method such as a hard negative examples mining strategy and self-supervised clustering for background objects. The performance of our method is assessed on DOTA, a large-scale remote sensing images dataset. The experiments conducted provide a broader understanding of the capabilities of representation learning. It highlights in particular some intrinsic weaknesses for the few-shot object detection task. Finally, some suggestions and perspectives are formulated according to these insights. | ['Hanene Azzag', 'Anissa Mokraoui', 'Mustapha Lebbah', 'Pierre Le Jeune'] | 2021-09-27 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 4.90372181e-01 5.84760346e-02 -1.64510816e-01 -4.55135524e-01
-1.76941276e-01 -2.99763441e-01 4.39789861e-01 2.49909714e-01
-5.30924797e-01 4.60296035e-01 -2.00357452e-01 1.01722710e-01
-3.33655298e-01 -1.19249439e+00 -3.22802991e-01 -8.61657083e-01
-4.71319407e-01 2.62737870e-01 4.16794121e-01 -2.82414079e-01
3.05726882e-02 9.50974166e-01 -2.10007429e+00 4.38138783e-01
6.06208861e-01 1.00573814e+00 4.88839000e-01 8.52275491e-01
-2.11397163e-03 6.83467090e-01 -5.93616903e-01 -6.21668203e-03
5.54264009e-01 -2.54531205e-01 -5.90346873e-01 4.43660825e-01
2.87169337e-01 -5.23845553e-01 -2.28080571e-01 1.03340530e+00
5.48894405e-01 6.60806119e-01 6.69334888e-01 -8.80436718e-01
-7.80145407e-01 5.67510307e-01 -3.76721770e-01 5.84102690e-01
-1.14785329e-01 2.89447695e-01 9.35103476e-01 -1.25197256e+00
5.54767728e-01 9.03539538e-01 7.35619724e-01 6.52393043e-01
-8.78446817e-01 -3.63108426e-01 2.17970818e-01 2.75651574e-01
-1.37876487e+00 -3.68754625e-01 5.53835869e-01 -4.97159153e-01
8.60661209e-01 2.04065070e-01 8.87802541e-01 7.12650776e-01
-2.25307480e-01 6.15983903e-01 7.42214441e-01 -4.65873867e-01
7.36963451e-01 6.21615529e-01 3.40535253e-01 6.02496445e-01
2.00876996e-01 1.96374372e-01 -6.56765327e-02 4.66423072e-02
4.89985257e-01 4.05327946e-01 -7.45273978e-02 -4.14936006e-01
-7.28970766e-01 1.02427125e+00 9.17658031e-01 4.07749265e-01
-5.83702981e-01 -2.87475437e-01 5.28750181e-01 1.55874103e-01
7.12839901e-01 5.60487032e-01 -4.73990887e-01 4.92883265e-01
-9.08618629e-01 -3.37234028e-02 5.34167886e-01 6.74931467e-01
9.56643879e-01 2.56910086e-01 -3.13522279e-01 9.02865291e-01
3.63467298e-02 2.14192480e-01 6.14031315e-01 -4.26795214e-01
7.21104145e-02 8.59045863e-01 4.62476024e-03 -1.13549232e+00
-4.40408230e-01 -5.70639133e-01 -7.44449914e-01 4.76911187e-01
1.10447906e-01 -3.04010749e-01 -9.96625304e-01 1.21147883e+00
5.52260756e-01 1.50882751e-01 2.15585425e-01 7.54143178e-01
9.00497317e-01 1.04829574e+00 -5.08428663e-02 -1.58904567e-01
1.20612919e+00 -8.58663499e-01 -4.32281077e-01 -3.05021614e-01
4.72197145e-01 -2.31309399e-01 8.49431157e-01 1.26567870e-01
-5.85643053e-01 -9.91323769e-01 -9.83369529e-01 5.92313588e-01
-9.30006087e-01 4.81248915e-01 6.30990744e-01 5.53622723e-01
-7.43648887e-01 8.11030149e-01 -5.87034941e-01 -8.72014821e-01
7.88360178e-01 1.07997410e-01 1.25413612e-02 4.73686904e-02
-9.60936308e-01 7.12705195e-01 7.28102982e-01 4.08694983e-01
-1.11785007e+00 -4.30607766e-01 -8.30958545e-01 6.66050985e-02
3.82406533e-01 5.88679947e-02 1.00809777e+00 -1.28620720e+00
-1.28308046e+00 9.25623536e-01 2.80398905e-01 -7.66338527e-01
2.98262566e-01 1.47782452e-02 -5.92933714e-01 2.48771086e-01
1.32480100e-01 7.53042281e-01 1.01750565e+00 -1.12853360e+00
-9.35164273e-01 -4.46250945e-01 1.22481532e-01 1.40231803e-01
-6.08238280e-01 3.95195410e-02 1.32324114e-01 -4.92613941e-01
1.55088110e-02 -5.96409082e-01 -5.01241684e-01 4.65165049e-01
-6.60918728e-02 -2.10175782e-01 9.18063939e-01 -1.77829653e-01
1.11154616e+00 -2.20104146e+00 -4.47421484e-02 1.42914057e-01
-1.31506138e-02 7.64670491e-01 -3.34506720e-01 4.09974754e-01
-3.95972490e-01 -8.29889551e-02 -1.50577918e-01 2.82589167e-01
-4.40457493e-01 1.46086797e-01 -6.59375548e-01 4.33689892e-01
6.46160066e-01 5.57056665e-01 -1.13166308e+00 -1.09690852e-01
2.59395391e-01 1.98646843e-01 -2.00556770e-01 3.28810781e-01
-2.41481826e-01 -1.37280732e-01 -2.55524874e-01 8.34610283e-01
5.54610014e-01 -2.30833441e-01 1.95656084e-02 -1.85099706e-01
-4.41686839e-01 -3.76873195e-01 -1.23022330e+00 1.17205405e+00
-1.54437110e-01 6.33372068e-01 -4.44508165e-01 -1.22497725e+00
1.42942071e+00 8.03595185e-02 1.87499970e-01 -1.21581227e-01
1.35631233e-01 -4.46837768e-02 8.61826614e-02 -7.50015199e-01
6.12927079e-01 -1.38449594e-02 3.45285743e-01 2.04435572e-01
2.29042023e-01 1.59434870e-01 4.25808996e-01 -5.96695393e-02
8.85111690e-01 1.68001264e-01 7.29391515e-01 -1.77139975e-02
4.02677625e-01 3.78065705e-01 5.08363187e-01 1.10607183e+00
-5.29559314e-01 3.31351638e-01 2.68767159e-02 -1.20259488e+00
-9.44582105e-01 -7.39808500e-01 -2.78989166e-01 1.39164758e+00
-4.84440029e-02 -1.14997245e-01 -2.91503578e-01 -7.67011821e-01
-7.51744136e-02 5.85307837e-01 -1.15274072e+00 -3.40960711e-01
-8.51927251e-02 -1.03193855e+00 1.48586109e-01 6.52140796e-01
4.39062595e-01 -1.43406343e+00 -1.29974198e+00 3.92107487e-01
3.99426430e-01 -6.19154990e-01 2.64008790e-01 6.25735462e-01
-1.20109534e+00 -1.15447593e+00 -6.48273289e-01 -9.07889187e-01
7.85974562e-01 7.60254562e-01 6.00874364e-01 -5.69506362e-03
-8.30964804e-01 2.28205144e-01 -7.67409384e-01 -4.88606691e-01
-3.32983226e-01 -1.28927156e-01 8.14058110e-02 2.83452630e-01
7.35715091e-01 -2.58417428e-01 -6.29635274e-01 2.20187560e-01
-9.30593491e-01 -4.78872418e-01 7.73825049e-01 9.82381344e-01
4.40709323e-01 1.44448817e-01 6.08212113e-01 -8.31471384e-01
3.32010806e-01 -6.42567754e-01 -6.04660153e-01 3.21595401e-01
-3.99656117e-01 -3.84643376e-01 7.33105838e-01 -8.68303537e-01
-8.80043864e-01 3.65931243e-01 3.05544227e-01 -4.30690587e-01
-3.24944019e-01 3.98253620e-01 4.13898915e-01 2.54087187e-02
1.26415265e+00 3.35374415e-01 -5.19770719e-02 -1.35327160e-01
4.31670815e-01 7.94558764e-01 3.01544130e-01 8.31460506e-02
9.09540772e-01 6.59343123e-01 -4.86040294e-01 -1.29880893e+00
-1.09305096e+00 -6.65743411e-01 -9.74382460e-01 -4.78887975e-01
7.46031225e-01 -9.46537971e-01 -1.71904743e-01 2.47131899e-01
-9.23460245e-01 -2.79370785e-01 -1.07271564e+00 3.11093897e-01
-2.71697700e-01 -6.41771108e-02 -1.32442564e-01 -1.25000000e+00
-3.20871741e-01 -6.08098030e-01 7.71611392e-01 6.97199166e-01
1.66726843e-01 -7.12041378e-01 2.93560237e-01 -2.64810264e-01
3.75227332e-01 4.50404696e-02 5.23940444e-01 -1.00719750e+00
-2.26968348e-01 -5.61935484e-01 -2.65402019e-01 5.43699622e-01
1.78517267e-01 4.39798474e-01 -1.29280376e+00 -4.58197534e-01
-1.69246703e-01 -5.18105567e-01 9.77271020e-01 3.86860490e-01
1.24996436e+00 -1.78053021e-01 -3.98232877e-01 6.43698514e-01
1.59074342e+00 5.54155290e-01 6.14427507e-01 3.10397923e-01
1.99490204e-01 6.02692723e-01 8.57853711e-01 9.08172190e-01
-2.54601270e-01 2.99549699e-01 5.95128775e-01 -2.00093493e-01
8.24954808e-02 -2.91024502e-02 3.03703517e-01 4.32341069e-01
-1.36191517e-01 -2.69147065e-02 -7.72645712e-01 6.31798208e-01
-1.69265234e+00 -1.25273728e+00 2.43572474e-01 2.12769651e+00
3.97684544e-01 3.42296138e-02 1.85782358e-01 8.58887509e-02
9.39129710e-01 3.73009920e-01 -7.02867568e-01 -6.31376430e-02
-1.88667893e-01 3.99419695e-01 3.52325588e-01 -1.71687249e-02
-1.49748802e+00 1.17198384e+00 6.23612165e+00 4.31089461e-01
-1.03688014e+00 -5.62381484e-02 6.16239786e-01 6.52875677e-02
4.84712839e-01 -1.17897745e-02 -8.79459620e-01 8.27560276e-02
6.62766993e-01 -1.25271365e-01 2.46248290e-01 1.31504893e+00
6.52666911e-02 -9.10784453e-02 -8.36229742e-01 7.48335898e-01
3.96262258e-01 -1.54778838e+00 2.31330127e-01 -3.06848228e-01
7.10404932e-01 1.81032062e-01 -4.53676432e-02 6.41176462e-01
1.76758111e-01 -6.02489114e-01 3.84277016e-01 5.19015551e-01
6.30775094e-01 -7.14382529e-01 7.97928989e-01 2.34716058e-01
-1.37069023e+00 -7.70591557e-01 -1.24010479e+00 -3.30187231e-01
-3.40314329e-01 3.12231213e-01 -1.05864656e+00 3.16732347e-01
8.18432987e-01 9.59705830e-01 -9.69817042e-01 1.35767186e+00
-2.37471014e-01 4.54442501e-01 5.95367327e-02 -3.29464197e-01
4.28552687e-01 -2.42666066e-01 6.06463611e-01 1.11151719e+00
3.55113924e-01 1.83735371e-01 2.82751620e-01 9.43280995e-01
2.43005395e-01 1.74496442e-01 -1.18430698e+00 -1.32700562e-01
5.25137544e-01 1.72521436e+00 -1.05088985e+00 -6.64699733e-01
-1.55709967e-01 8.74877751e-01 3.81906360e-01 1.93418011e-01
-5.56016624e-01 -7.40733206e-01 4.29092258e-01 -8.64509940e-02
6.53091729e-01 2.61565357e-01 2.84506261e-01 -9.79848921e-01
-2.98786730e-01 -6.01428926e-01 6.24879062e-01 -8.21133733e-01
-1.35422969e+00 6.24735057e-01 -1.38520390e-01 -1.54648781e+00
1.58428401e-01 -7.02731133e-01 -9.57789183e-01 3.58427733e-01
-1.42308521e+00 -7.93804884e-01 -6.42858744e-01 2.28073448e-01
8.54709864e-01 -5.95437527e-01 1.09342611e+00 2.63937702e-03
-7.78486490e-01 1.21268444e-01 2.28057191e-01 2.73421794e-01
3.21370244e-01 -1.03498864e+00 1.10331744e-01 9.94505465e-01
1.53402999e-01 2.91670263e-01 5.48648298e-01 -5.05698025e-01
-1.05019712e+00 -1.67013025e+00 4.48988378e-01 -1.15693927e-01
7.34550893e-01 -2.04843238e-01 -1.15213954e+00 4.69078600e-01
-2.67530859e-01 4.31491137e-01 8.65627229e-01 -1.80314764e-01
-1.12228796e-01 -4.29132640e-01 -1.20948827e+00 3.51041198e-01
8.11084151e-01 -2.59891659e-01 -7.27175951e-01 7.52782643e-01
4.95799810e-01 1.73680276e-01 -4.24266726e-01 1.40547812e-01
3.66358995e-01 -9.36832666e-01 8.00758660e-01 -1.00851047e+00
3.88705909e-01 -3.94986719e-01 -2.57954985e-01 -1.19079447e+00
-6.95825100e-01 -3.80517513e-01 -6.04729876e-02 8.16258371e-01
1.79946229e-01 -4.66384828e-01 6.59663081e-01 -1.11749567e-01
-5.17327338e-02 -6.38756216e-01 -6.17336035e-01 -9.95991409e-01
-4.03295219e-01 -1.54942170e-01 3.26481968e-01 8.98943365e-01
-2.16123477e-01 3.19272280e-01 -2.67522454e-01 4.63972628e-01
5.35224974e-01 1.72795236e-01 7.97172189e-01 -1.53450620e+00
-1.43121123e-01 -1.78965554e-01 -6.53282821e-01 -6.26371682e-01
-2.34845936e-01 -9.76191401e-01 2.35313624e-01 -1.20440280e+00
2.20907360e-01 -2.45248511e-01 -4.05330390e-01 6.14793181e-01
-9.06059146e-03 2.44716138e-01 1.39193565e-01 2.50748575e-01
-6.36448085e-01 5.89396834e-01 6.88729525e-01 -4.22457337e-01
-7.47875094e-01 3.36045742e-01 -4.77676809e-01 7.64240205e-01
9.55151439e-01 -5.69379985e-01 -2.97503889e-01 -1.49776727e-01
6.24036118e-02 -2.51452953e-01 4.56183821e-01 -1.27040064e+00
1.71162665e-01 -1.17381558e-01 7.93702841e-01 -8.06422412e-01
1.15637064e-01 -8.05646241e-01 -2.12970704e-01 8.39149892e-01
-3.95416856e-01 -1.01942942e-01 2.15646148e-01 9.03328300e-01
1.11198463e-01 -7.02206612e-01 1.29837263e+00 -3.92137855e-01
-1.19935429e+00 3.84374142e-01 -5.44545829e-01 -2.94451505e-01
1.48012042e+00 -6.31582856e-01 -1.16257198e-01 6.77987784e-02
-1.04495072e+00 3.91218476e-02 7.03634843e-02 3.62234324e-01
9.32855904e-01 -1.09991348e+00 -6.31596386e-01 4.45579678e-01
5.92548192e-01 -7.84304067e-02 2.18593523e-01 4.73139256e-01
-3.88613373e-01 -6.45082518e-02 -5.47039866e-01 -6.56251729e-01
-1.08807850e+00 7.06608713e-01 4.97139633e-01 6.41955286e-02
-7.59288192e-01 9.13012087e-01 -7.04059377e-02 -3.47026259e-01
3.23066443e-01 -2.11187884e-01 -7.03034341e-01 5.25290430e-01
1.03302419e+00 4.88641560e-01 -1.40432060e-01 -3.34222108e-01
-1.94949254e-01 2.37363636e-01 -2.75875688e-01 2.77709395e-01
1.95919561e+00 9.74856317e-02 -5.08962385e-02 8.22968245e-01
6.75067484e-01 -6.59660757e-01 -1.35180652e+00 -4.96410549e-01
4.23164293e-03 -5.60839653e-01 -8.62550810e-02 -7.15136349e-01
-9.89141643e-01 8.87660503e-01 1.29341435e+00 3.89773965e-01
1.00380063e+00 -1.22160666e-01 8.33335817e-02 9.65489328e-01
7.58495778e-02 -1.44020903e+00 4.94354486e-01 5.76571167e-01
7.38548338e-01 -1.36700022e+00 1.46292955e-01 2.72414863e-01
-5.81296504e-01 1.63556099e+00 6.95247769e-01 -3.87023658e-01
5.70018351e-01 -2.99608827e-01 -9.21203122e-02 -4.91790324e-01
-6.43152773e-01 -5.60941517e-01 1.04111552e-01 9.75513816e-01
-1.76013056e-02 -2.25833873e-03 3.23618740e-01 1.12412333e-01
2.91356683e-01 -6.57622749e-03 6.33512914e-01 1.02737033e+00
-1.09996760e+00 -3.62793654e-01 -2.65472293e-01 6.22975528e-01
1.64315417e-01 5.70100695e-02 -5.14212072e-01 8.23977768e-01
2.98601687e-01 7.37745583e-01 2.94962466e-01 -3.77878159e-01
3.78103375e-01 7.96464235e-02 4.02247235e-02 -1.24639189e+00
-5.51411688e-01 -3.64718467e-01 -4.51873034e-01 -2.24779412e-01
-5.21887183e-01 -5.52883804e-01 -8.17521453e-01 3.68431151e-01
-6.79521978e-01 -1.04801610e-01 5.40071726e-01 5.79720259e-01
1.38934150e-01 5.75159848e-01 1.02641118e+00 -1.04748166e+00
-9.70097661e-01 -9.73980546e-01 -6.83515131e-01 1.85841113e-01
1.82699755e-01 -4.66974437e-01 -5.63688695e-01 3.43796760e-02] | [9.84738540649414, 2.3821194171905518] |
b0683fe9-86d8-4773-b797-bd3d6ec533da | nested-named-entity-recognition-via-second | 1909.02250 | null | https://arxiv.org/abs/1909.02250v3 | https://arxiv.org/pdf/1909.02250v3.pdf | Nested Named Entity Recognition via Second-best Sequence Learning and Decoding | When an entity name contains other names within it, the identification of all combinations of names can become difficult and expensive. We propose a new method to recognize not only outermost named entities but also inner nested ones. We design an objective function for training a neural model that treats the tag sequence for nested entities as the second best path within the span of their parent entity. In addition, we provide the decoding method for inference that extracts entities iteratively from outermost ones to inner ones in an outside-to-inside way. Our method has no additional hyperparameters to the conditional random field based model widely used for flat named entity recognition tasks. Experiments demonstrate that our method performs better than or at least as well as existing methods capable of handling nested entities, achieving the F1-scores of 85.82%, 84.34%, and 77.36% on ACE-2004, ACE-2005, and GENIA datasets, respectively. | ['Takashi Shibuya', 'Eduard Hovy'] | 2019-09-05 | null | null | null | null | ['nested-named-entity-recognition', 'nested-mention-recognition'] | ['natural-language-processing', 'natural-language-processing'] | [-1.61035925e-01 2.38017187e-01 -1.67925730e-01 -7.70699680e-01
-7.99614429e-01 -8.76406252e-01 1.87694147e-01 1.68135434e-01
-5.82830310e-01 1.14366364e+00 -9.98490900e-02 -5.18778384e-01
1.16719101e-02 -9.75187719e-01 -8.63947988e-01 -4.17522907e-01
-3.54246795e-01 5.37869871e-01 1.45153850e-01 3.00539285e-01
-5.40987067e-02 4.88665789e-01 -1.18359005e+00 1.61737144e-01
1.15272486e+00 6.64440751e-01 1.24378153e-03 2.63862401e-01
-4.77143884e-01 6.73473299e-01 -6.64464116e-01 -9.88265336e-01
8.12862739e-02 5.17000295e-02 -9.63108480e-01 -4.15256023e-01
5.56976795e-01 -2.16986954e-01 -1.88988239e-01 1.07927108e+00
2.79432148e-01 1.13962151e-01 5.12159288e-01 -8.99391413e-01
-7.72456288e-01 1.35499799e+00 -4.86891478e-01 7.18800947e-02
1.66078344e-01 -4.99459982e-01 1.34020984e+00 -9.14133430e-01
7.47919858e-01 8.41618657e-01 7.32858479e-01 5.09837151e-01
-1.10450935e+00 -8.40089917e-01 1.74562857e-01 -4.75207195e-02
-1.92990279e+00 -4.00874108e-01 -7.18021616e-02 -4.14390981e-01
1.21094382e+00 2.47166291e-01 -7.47866333e-02 5.88543892e-01
-1.91714242e-01 7.77381897e-01 5.04477739e-01 -3.87766987e-01
1.07585318e-01 1.51001979e-02 3.41829330e-01 8.52320313e-01
6.15905046e-01 -2.82172203e-01 -2.66952485e-01 -4.67132449e-01
4.83317107e-01 -4.79240060e-01 -9.51723978e-02 -4.27560285e-02
-1.07584584e+00 6.73591912e-01 2.95232296e-01 4.86569822e-01
-4.98237848e-01 1.14255182e-01 -5.85848130e-02 -1.69698313e-01
2.38546103e-01 6.36519194e-01 -1.16799116e+00 -1.84618682e-01
-9.81678724e-01 -2.81559564e-02 1.10640013e+00 1.48311126e+00
8.37541938e-01 -2.12114215e-01 -1.28258884e-01 7.75249243e-01
4.05244052e-01 4.53914732e-01 -3.12810764e-03 -6.21794999e-01
7.13782847e-01 6.71762824e-01 3.77571106e-01 -6.88272417e-01
-4.94018793e-01 -6.43591046e-01 -7.74993956e-01 -6.26480043e-01
4.20385152e-01 -5.41622639e-01 -1.09357750e+00 2.09215641e+00
4.25579071e-01 3.23394299e-01 5.14551438e-02 3.74026328e-01
9.37912405e-01 7.80095935e-01 4.44923967e-01 -8.33911225e-02
1.48752964e+00 -7.10047960e-01 -6.87622547e-01 -2.89733946e-01
8.44917238e-01 -3.74265194e-01 2.03097284e-01 -7.40046427e-02
-9.98005152e-01 -3.34062606e-01 -6.95894182e-01 -5.47559634e-02
-4.68137145e-01 7.13822365e-01 8.55940700e-01 8.21200669e-01
-8.12431574e-01 5.27709424e-01 -8.81245852e-01 -8.35561380e-02
1.21143177e-01 4.67239052e-01 -5.74493587e-01 2.84343064e-01
-1.53966892e+00 7.63340056e-01 9.94160593e-01 2.99101144e-01
-2.94442773e-01 -7.26021945e-01 -1.08394313e+00 3.92296523e-01
4.15102363e-01 -5.41989684e-01 1.28734624e+00 -5.86524725e-01
-9.03080463e-01 9.03933406e-01 -5.57074249e-01 -4.76772398e-01
1.21567801e-01 -3.52851272e-01 -8.29042792e-01 -2.64426261e-01
3.91684026e-01 6.71453834e-01 6.40411442e-03 -1.01078355e+00
-9.12338555e-01 -2.73309052e-01 6.10767230e-02 -7.92777911e-02
-2.68093556e-01 1.95687518e-01 -7.30151415e-01 -5.66288233e-01
1.89861074e-01 -7.81897664e-01 -2.20368400e-01 -4.50184733e-01
-9.54870522e-01 -4.33730662e-01 -1.50216846e-02 -8.39487433e-01
1.83106351e+00 -2.00400066e+00 -2.22591966e-01 3.92136067e-01
1.96228608e-01 3.85405809e-01 1.98587865e-01 4.66244400e-01
-2.76514858e-01 4.97704655e-01 -2.52962470e-01 -1.11628003e-01
2.13978305e-01 6.78032488e-02 -2.97560036e-01 1.41794398e-01
3.40423942e-01 7.01657534e-01 -7.97420263e-01 -3.85638505e-01
-4.37108248e-01 3.00131083e-01 -4.71791536e-01 8.85314867e-02
-3.99982631e-02 2.47896854e-02 -5.58117747e-01 5.65478504e-01
8.44556987e-01 -5.64001799e-01 6.06827080e-01 -5.05912537e-03
-2.20217809e-01 7.27176905e-01 -1.50400317e+00 1.18217909e+00
-2.56775200e-01 1.73345610e-01 -5.56346104e-02 -6.69574022e-01
9.51494575e-01 4.56462979e-01 1.59342036e-01 -3.67309719e-01
-2.96982825e-01 4.40807581e-01 -1.10980362e-01 -4.93293345e-01
6.07667208e-01 5.37423901e-02 -3.50828826e-01 2.24698358e-03
1.90817982e-01 8.79686415e-01 5.56009591e-01 2.27000087e-01
1.11594892e+00 3.52533646e-02 6.13481343e-01 -2.23181341e-05
6.55317366e-01 -3.36463422e-01 1.18732917e+00 9.57077861e-01
1.39732435e-02 2.90294141e-01 5.17422318e-01 -2.98145592e-01
-8.64814103e-01 -1.19414139e+00 -3.70827049e-01 1.20768917e+00
2.23287567e-02 -5.11366010e-01 -6.41209364e-01 -9.11877871e-01
1.86842367e-01 1.03981388e+00 -4.29464549e-01 1.64298654e-01
-8.37254047e-01 -6.32100523e-01 1.07174373e+00 7.90839493e-01
4.78440434e-01 -9.52054620e-01 9.71675217e-02 3.73043865e-01
-3.71843368e-01 -1.33733165e+00 -4.96573627e-01 4.94045347e-01
-5.96366167e-01 -8.14379752e-01 -6.35298193e-01 -1.03696120e+00
6.76633716e-01 -4.54626530e-01 1.28580296e+00 4.42300458e-03
2.71539509e-01 -3.53803486e-01 -1.01485439e-01 -6.33929148e-02
-1.79008916e-01 5.51469386e-01 -1.31876990e-01 1.52349677e-02
6.04316115e-01 -2.65047073e-01 -7.73284808e-02 3.95597905e-01
-5.73931754e-01 -7.87804499e-02 6.94369853e-01 7.76660025e-01
4.40124661e-01 -5.41577227e-02 7.81368315e-01 -1.26649725e+00
6.50924146e-02 -8.46159935e-01 -7.05392420e-01 5.96508384e-01
-5.09253144e-01 3.23426306e-01 7.29485214e-01 -1.91008851e-01
-1.27187121e+00 3.53040576e-01 -3.97073984e-01 1.49893865e-01
-4.72611129e-01 8.55175078e-01 -5.51674783e-01 5.12310326e-01
3.68233144e-01 1.51850849e-01 -9.61194873e-01 -7.54624903e-01
3.68822962e-01 7.63614178e-01 5.55812895e-01 -6.61940873e-01
8.17220807e-01 5.86088784e-02 -4.18158680e-01 -5.85971773e-01
-8.96832824e-01 -5.29909134e-01 -7.24959671e-01 3.34565252e-01
7.68839240e-01 -9.92754340e-01 -7.00881064e-01 4.60307032e-01
-1.35157955e+00 1.47429034e-01 1.39729097e-01 5.20593762e-01
1.24525741e-01 5.15406489e-01 -8.14532042e-01 -7.29932189e-01
-3.67276043e-01 -6.92877233e-01 8.69832397e-01 5.97577214e-01
-2.03329042e-01 -7.57800996e-01 -5.31798340e-02 1.93155110e-01
1.05867788e-01 -7.39720762e-02 9.93198156e-01 -1.34486651e+00
-6.41559601e-01 -2.43582875e-01 -3.29449058e-01 -2.37340350e-02
-4.96234335e-02 1.60897419e-01 -6.55754566e-01 -1.01105496e-01
-7.42453396e-01 5.14500886e-02 8.40127349e-01 1.35191187e-01
8.43968511e-01 -4.00937945e-01 -7.94613540e-01 6.67668998e-01
1.35641873e+00 6.31525636e-01 6.46030605e-01 4.12883908e-01
6.57181859e-01 3.21450770e-01 3.67636889e-01 3.35140973e-01
6.81995988e-01 4.80659068e-01 5.26734963e-02 1.24769762e-01
5.33905625e-01 -3.97868663e-01 5.42421639e-02 6.44197464e-01
1.66187569e-01 -6.15206063e-01 -1.10463369e+00 7.82952309e-01
-1.60720825e+00 -1.05105984e+00 -3.55523050e-01 2.26043272e+00
1.21153176e+00 5.24860993e-02 -2.51295775e-01 -4.56595600e-01
1.28829730e+00 -1.46259489e-02 -3.54060441e-01 -2.25193292e-01
-1.50289834e-01 4.04335707e-01 7.37632990e-01 2.82383054e-01
-1.61085510e+00 1.05946469e+00 6.37220907e+00 6.96692050e-01
-5.57839990e-01 -2.15009302e-01 4.43193227e-01 4.41770196e-01
-3.31713885e-01 1.44875258e-01 -1.52098131e+00 5.78076005e-01
1.10568607e+00 -5.88925295e-02 2.30875358e-01 9.43040967e-01
-2.72858530e-01 4.10784297e-02 -1.00381815e+00 6.04724109e-01
-3.34292352e-01 -9.98868108e-01 -1.80667579e-01 -5.89283369e-02
8.76738846e-01 8.47232118e-02 -4.20068055e-01 4.80430394e-01
6.03319883e-01 -9.18297052e-01 6.65934563e-01 4.33551788e-01
7.13645101e-01 -7.92245269e-01 8.90293360e-01 4.89320606e-01
-1.42512548e+00 6.52502328e-02 -2.09457695e-01 2.82955289e-01
3.37871850e-01 7.63185203e-01 -9.01876867e-01 7.82058239e-01
5.92014492e-01 2.77505398e-01 -3.77206445e-01 1.44439840e+00
-6.07870877e-01 8.94937456e-01 -5.38646400e-01 -1.75424188e-01
1.24205165e-01 8.17872137e-02 3.56410533e-01 1.58135581e+00
3.61476570e-01 -8.31022672e-03 1.95551645e-02 8.93950582e-01
-5.67393601e-01 1.38696894e-01 -3.01630616e-01 -3.02961439e-01
1.08623147e+00 1.11179674e+00 -5.35100996e-01 -5.34691930e-01
-4.35155779e-01 7.17827082e-01 6.96509838e-01 4.35817063e-01
-1.12174034e+00 -9.97901976e-01 5.77921271e-01 -3.30874354e-01
9.28080916e-01 -1.54860005e-01 -2.61682868e-01 -1.27656710e+00
2.79444665e-01 -5.77474296e-01 6.09938562e-01 -5.77985585e-01
-1.33787775e+00 6.35274231e-01 -1.02677226e-01 -7.90644526e-01
-2.17964575e-01 -4.89540428e-01 -3.35578650e-01 1.03036404e+00
-1.24304307e+00 -8.56973231e-01 1.22886963e-01 2.79192310e-02
-2.13661715e-01 2.98608541e-02 1.02447855e+00 7.75793195e-01
-8.07508826e-01 1.05840707e+00 2.99609929e-01 9.76622939e-01
6.05686843e-01 -1.34204400e+00 9.34598088e-01 1.15560031e+00
2.70911157e-01 1.27889466e+00 2.32604042e-01 -7.81298876e-01
-9.57237005e-01 -1.09803247e+00 1.74525642e+00 -1.59835607e-01
4.88127500e-01 -3.27881753e-01 -1.00055909e+00 1.15069842e+00
1.35080948e-01 -2.76120663e-01 8.52254987e-01 4.83656824e-01
-6.27005577e-01 1.62798449e-01 -1.06128550e+00 4.28530067e-01
1.05697060e+00 -3.78887057e-01 -7.83696115e-01 -2.12320648e-02
7.03250825e-01 -5.18391371e-01 -1.11500132e+00 5.38010955e-01
5.52072465e-01 -5.62624395e-01 8.95432234e-01 -8.18726659e-01
2.05834374e-01 -4.68646348e-01 -1.15609534e-01 -9.21002805e-01
-4.17432189e-01 -4.88086671e-01 -3.41883004e-01 1.81105101e+00
1.11454284e+00 -7.91820765e-01 7.41661251e-01 8.20257127e-01
4.15521190e-02 -6.10013783e-01 -9.41046715e-01 -7.52966464e-01
-8.19674507e-02 -2.31854424e-01 8.04723680e-01 1.27139115e+00
-2.31324345e-01 3.32176626e-01 -2.25841582e-01 7.49334455e-01
4.55732644e-01 2.48202726e-01 4.17012006e-01 -1.26849174e+00
-1.99810177e-01 -1.31724313e-01 -3.00875962e-01 -1.37302709e+00
3.17545146e-01 -9.77181315e-01 1.33257627e-01 -1.57596934e+00
2.01456293e-01 -7.86792159e-01 -5.30949831e-01 9.74058270e-01
-4.50263649e-01 -2.11593106e-01 -1.74561560e-01 -1.03342220e-01
-5.81603169e-01 3.39890085e-02 6.40339315e-01 -7.06118718e-03
-7.43879452e-02 8.11051726e-02 -7.60355711e-01 6.19248033e-01
6.94693565e-01 -7.99578011e-01 1.75406069e-01 -5.18794358e-01
4.80872095e-01 1.28748804e-01 -9.27294269e-02 -7.36041784e-01
3.54278535e-01 -1.14819057e-01 3.52336079e-01 -8.97178471e-01
-1.34982076e-02 -5.18158376e-01 3.86626542e-01 1.04048669e-01
-4.14469212e-01 3.78799178e-02 8.91419873e-02 3.59435380e-01
-2.54588425e-01 -5.39491594e-01 4.63614523e-01 -5.96611351e-02
-1.02782094e+00 1.13406032e-01 -3.33853811e-01 3.08885098e-01
8.14138353e-01 6.99425265e-02 -3.60571146e-01 -7.86888525e-02
-9.06733751e-01 5.42945027e-01 6.14152364e-02 4.86020148e-01
2.07386956e-01 -1.16667926e+00 -6.74518764e-01 4.35223207e-02
1.13700829e-01 -5.83889112e-02 1.54224858e-01 5.10198474e-01
-5.75185299e-01 5.96254170e-01 2.32621685e-01 -1.74795195e-01
-1.18211818e+00 4.11796331e-01 3.67218703e-01 -4.21342164e-01
-2.60364771e-01 1.14524484e+00 1.85226485e-01 -1.01154721e+00
3.42697382e-01 -2.30615273e-01 -2.70652413e-01 -3.19646783e-02
4.86741871e-01 2.43341371e-01 9.06198695e-02 -7.22544551e-01
-6.60075426e-01 2.07049295e-01 -2.98510134e-01 2.19155744e-01
1.20430779e+00 2.83717755e-02 -3.50419462e-01 1.88549846e-01
1.16549647e+00 3.02652180e-01 -5.18770516e-01 -3.79805177e-01
5.96597731e-01 -1.03603549e-01 -3.42865229e-01 -1.02939773e+00
-9.76446748e-01 4.16946262e-01 1.54604346e-01 2.58265764e-01
8.16953599e-01 7.52785150e-03 8.35507035e-01 7.08574176e-01
5.77442527e-01 -9.30982709e-01 -9.56145108e-01 8.97232115e-01
1.13431871e-01 -8.79510760e-01 -2.05598056e-01 -6.74088359e-01
-3.92765492e-01 9.37774479e-01 6.44581258e-01 2.52334058e-01
5.18437624e-01 4.76535141e-01 -7.00395554e-02 2.29776762e-02
-8.07461083e-01 -2.53063947e-01 4.09993023e-01 1.52235940e-01
7.23806679e-01 1.72962934e-01 -5.40245473e-01 7.57472515e-01
1.07325846e-02 -4.65772897e-02 2.44433075e-01 9.30305481e-01
-4.82155710e-01 -1.08777511e+00 -1.35780901e-01 4.47775811e-01
-1.08791363e+00 -5.11294484e-01 -2.51717389e-01 7.97607839e-01
3.56909335e-01 8.92157018e-01 1.82527483e-01 -2.63516933e-01
2.92883009e-01 5.61415076e-01 5.03682904e-02 -6.34992957e-01
-6.35407865e-01 -2.17228964e-01 6.64565146e-01 -7.46419579e-02
-2.12521940e-01 -4.18641359e-01 -1.68040586e+00 -2.98397183e-01
-8.82716298e-01 5.94833553e-01 4.94866878e-01 9.33482707e-01
8.29396486e-01 2.44095132e-01 4.71692890e-01 2.77345721e-02
-4.85348552e-01 -9.26536143e-01 -6.38299346e-01 2.46024460e-01
3.08159878e-03 -6.56584799e-01 -9.46463048e-02 2.38303430e-02] | [9.636587142944336, 9.481165885925293] |
317ab7e3-20b8-47c6-9be1-82e7da9cbe8b | class-incremental-mixture-of-gaussians-for | 2307.04094 | null | https://arxiv.org/abs/2307.04094v1 | https://arxiv.org/pdf/2307.04094v1.pdf | Class-Incremental Mixture of Gaussians for Deep Continual Learning | Continual learning models for stationary data focus on learning and retaining concepts coming to them in a sequential manner. In the most generic class-incremental environment, we have to be ready to deal with classes coming one by one, without any higher-level grouping. This requirement invalidates many previously proposed methods and forces researchers to look for more flexible alternative approaches. In this work, we follow the idea of centroid-driven methods and propose end-to-end incorporation of the mixture of Gaussians model into the continual learning framework. By employing the gradient-based approach and designing losses capable of learning discriminative features while avoiding degenerate solutions, we successfully combine the mixture model with a deep feature extractor allowing for joint optimization and adjustments in the latent space. Additionally, we show that our model can effectively learn in memory-free scenarios with fixed extractors. In the conducted experiments, we empirically demonstrate the effectiveness of the proposed solutions and exhibit the competitiveness of our model when compared with state-of-the-art continual learning baselines evaluated in the context of image classification problems. | ['Bartosz Krawczyk', 'Lukasz Korycki'] | 2023-07-09 | null | null | null | null | ['image-classification', 'continual-learning'] | ['computer-vision', 'methodology'] | [ 2.14304864e-01 -2.17882693e-01 -4.39263992e-02 -4.28582966e-01
-9.49795663e-01 -4.37861532e-01 7.87979066e-01 3.63754243e-01
-9.80323553e-01 6.11172497e-01 -1.71494097e-01 -2.63921112e-01
-4.97431338e-01 -4.93275225e-01 -8.05887878e-01 -9.06533897e-01
-8.89352486e-02 4.21434939e-01 2.46814653e-01 1.07983537e-01
3.33867908e-01 6.01632893e-01 -1.71173215e+00 9.80819166e-02
8.03367674e-01 8.70732248e-01 3.39334577e-01 5.24340093e-01
-1.61106318e-01 7.58308887e-01 -2.28194013e-01 -4.28306550e-01
4.02516186e-01 -1.23713344e-01 -5.99565685e-01 6.84449136e-01
8.24455798e-01 -1.18842430e-01 -7.32521713e-02 7.98217535e-01
4.59852964e-01 5.03950894e-01 8.43458176e-01 -1.07442582e+00
-3.15671414e-01 2.91781098e-01 -8.70856702e-01 1.21163011e-01
-1.95868954e-01 -6.45566210e-02 8.82533669e-01 -1.13329232e+00
4.47794408e-01 1.01700687e+00 7.05907702e-01 3.68140131e-01
-1.45132577e+00 -3.73262346e-01 6.74001575e-01 3.23705196e-01
-1.46046603e+00 -6.44943833e-01 9.04417038e-01 -2.56111741e-01
8.64005029e-01 2.51470298e-01 4.52244490e-01 8.43303144e-01
-6.74381778e-02 1.18646348e+00 9.47934151e-01 -8.93072665e-01
3.94118577e-01 4.54447627e-01 3.56085539e-01 6.19657040e-01
7.89454356e-02 -3.79026890e-01 -4.83606488e-01 -1.76153749e-01
3.61960232e-01 2.66499162e-01 -6.01755269e-02 -1.04949975e+00
-7.13171482e-01 8.19255114e-01 3.89753908e-01 2.35700294e-01
-4.14618939e-01 7.30029568e-02 2.44033784e-01 1.71169400e-01
7.29288459e-01 -6.12047911e-02 -2.99172342e-01 9.69110355e-02
-1.51289034e+00 3.62630486e-01 5.08251667e-01 8.80480111e-01
7.11651742e-01 -1.97459280e-01 -1.32561207e-01 1.04581714e+00
2.86427498e-01 -1.16097983e-02 5.46902537e-01 -8.11765313e-01
2.97526479e-01 2.57681191e-01 1.01255365e-01 -6.92445278e-01
-2.68477887e-01 -1.10196972e+00 -6.40033662e-01 2.49398947e-01
2.62214094e-01 1.10215716e-01 -8.88103306e-01 1.80554867e+00
5.61991394e-01 4.80539531e-01 6.07256852e-02 3.67432564e-01
1.58466410e-03 4.61088181e-01 8.69529098e-02 -5.52514136e-01
9.42688823e-01 -1.32498741e+00 -4.70525205e-01 -3.71354930e-02
4.58866894e-01 -8.89990211e-01 1.09359288e+00 7.23027945e-01
-1.13009381e+00 -7.37888753e-01 -9.22038853e-01 -5.80542013e-02
-2.31315330e-01 3.89237136e-01 5.19093871e-01 7.26790190e-01
-1.17125797e+00 7.72104502e-01 -1.28646481e+00 -3.31057250e-01
4.93177623e-01 4.62579548e-01 -2.28994250e-01 -1.95052072e-01
-4.03982729e-01 6.38930380e-01 3.68109405e-01 1.90571457e-01
-8.03587019e-01 -4.94565964e-01 -5.24447978e-01 5.32651693e-02
5.17749071e-01 -7.52786100e-01 1.12885618e+00 -7.90551603e-01
-1.24195874e+00 4.73652899e-01 -4.32636201e-01 -6.81935787e-01
8.33873570e-01 -7.23390222e-01 2.29746595e-01 1.08478911e-01
-1.29976630e-01 7.29263961e-01 1.16148543e+00 -1.42114997e+00
-8.69554937e-01 -3.39603901e-01 1.18654735e-01 4.47235674e-01
-1.01700950e+00 -3.71780962e-01 -5.75302839e-01 -5.92911601e-01
2.96183139e-01 -9.29936171e-01 -3.22543770e-01 -3.58940959e-02
-1.12946488e-01 -3.10698748e-01 9.07739043e-01 -4.54831779e-01
1.23779833e+00 -2.24334168e+00 2.51031488e-01 6.96846247e-02
1.16956629e-01 2.29391053e-01 -1.18808849e-02 3.87834996e-01
-8.57222825e-02 -1.96401343e-01 -5.04118741e-01 -1.26501524e+00
-5.58029525e-02 4.40658927e-02 -2.87587553e-01 6.61987185e-01
5.82687594e-02 5.64353704e-01 -7.54307687e-01 -6.18836045e-01
2.78888077e-01 6.43677413e-01 -8.68990302e-01 9.69169289e-02
-7.41792247e-02 2.23130018e-01 -2.48182639e-01 3.25103521e-01
8.36545587e-01 -7.12031946e-02 -3.63279157e-03 -2.73142476e-02
-2.16083363e-01 5.86902164e-02 -1.38189256e+00 1.93768859e+00
-6.06144726e-01 4.52374548e-01 -2.95707304e-02 -1.28306127e+00
5.93697846e-01 9.36961621e-02 4.78089005e-01 -2.02519849e-01
-2.84332812e-01 3.63748312e-01 -1.81521893e-01 -1.80686206e-01
4.59183455e-01 -1.87946215e-01 2.62024879e-01 1.70380771e-01
4.32694584e-01 1.73953921e-01 2.63035208e-01 1.83578566e-01
7.87186742e-01 4.68563050e-01 1.73154145e-01 -2.39722982e-01
6.31076396e-01 -2.57984012e-01 2.91441172e-01 9.58146513e-01
-5.21743372e-02 5.47128499e-01 5.00150770e-02 -1.48473546e-01
-8.31683695e-01 -1.11056781e+00 -3.51922452e-01 1.14468384e+00
-7.63924271e-02 -2.73981631e-01 -7.82225966e-01 -8.42967987e-01
-1.58107072e-01 9.21120286e-01 -4.29048747e-01 -1.24518082e-01
-6.56574786e-01 -1.11468160e+00 4.28799354e-02 2.78084964e-01
3.44515920e-01 -6.44279659e-01 -5.94320953e-01 3.54860246e-01
3.31641696e-02 -7.72972763e-01 -2.50385821e-01 6.09130859e-01
-1.11786032e+00 -6.42226696e-01 -9.05575156e-01 -9.02532101e-01
7.56032705e-01 4.44522440e-01 8.29007149e-01 -1.92674369e-01
-3.58257443e-01 6.35253608e-01 -3.07333052e-01 -2.32139334e-01
7.65199587e-02 3.98318022e-01 -3.23348232e-02 4.99879122e-01
2.01934814e-01 -7.96203732e-01 -7.76408494e-01 -5.83436117e-02
-1.01346374e+00 -6.21262789e-02 6.26963437e-01 9.24604952e-01
6.09559894e-01 2.09830657e-01 6.42968595e-01 -7.70565748e-01
4.20824200e-01 -4.55673993e-01 -4.78561044e-01 3.79050404e-01
-7.83025205e-01 6.61391094e-02 6.78083956e-01 -4.98962581e-01
-1.08414865e+00 3.69579077e-01 -1.81214795e-01 -4.92094755e-01
-1.81876093e-01 1.70931622e-01 -1.46519884e-01 -3.09557300e-02
4.47405040e-01 4.83342230e-01 -2.51028866e-01 -6.82251811e-01
6.12577200e-01 6.66017354e-01 4.72638190e-01 -6.08719528e-01
6.91133559e-01 6.30423129e-01 -7.22028017e-02 -8.65038574e-01
-8.59021842e-01 -9.05040801e-01 -9.30926919e-01 -1.50003076e-01
4.64614362e-01 -9.06096518e-01 -3.81601423e-01 6.76466703e-01
-9.48158562e-01 -1.46144256e-01 -4.24075663e-01 5.74250638e-01
-7.12135792e-01 6.72866344e-01 -4.19430405e-01 -1.09345949e+00
-1.16487205e-01 -1.08409691e+00 1.05447698e+00 1.08303085e-01
1.04314096e-01 -1.13727760e+00 -3.15291211e-02 2.35601351e-01
4.31046903e-01 -9.76577699e-02 7.18450010e-01 -7.58403242e-01
-5.47448695e-01 -2.58384407e-01 1.38326854e-01 5.70060313e-01
2.40470827e-01 -3.11518818e-01 -1.18625641e+00 -7.05016494e-01
3.62256646e-01 -5.46039283e-01 1.30145800e+00 4.30571318e-01
1.23228145e+00 -3.07421088e-01 -3.90908778e-01 7.53564894e-01
1.56137025e+00 -3.20559032e-02 3.43875587e-01 5.11835873e-01
6.58066452e-01 6.62178695e-01 5.41631103e-01 4.36578333e-01
3.14695269e-01 6.56388521e-01 3.55332136e-01 -4.77214269e-02
-4.85987961e-02 -1.17511805e-02 6.09397106e-02 6.18029952e-01
1.09572724e-01 -4.02860701e-01 -6.34041011e-01 7.87847459e-01
-1.93068278e+00 -8.58245909e-01 1.07451558e-01 2.39597964e+00
8.75570059e-01 2.83078223e-01 1.05689734e-01 2.92466044e-01
4.93228704e-01 3.49486433e-02 -4.62979585e-01 -5.80132492e-02
9.28252488e-02 2.06447586e-01 4.04890627e-01 6.72977984e-01
-1.45769763e+00 9.95778084e-01 5.37709332e+00 1.33478808e+00
-1.13096833e+00 2.25694805e-01 6.71253264e-01 -4.14711177e-01
-1.63380533e-01 1.37871638e-01 -9.60737944e-01 1.80058271e-01
8.22875738e-01 1.40619442e-01 2.86938578e-01 8.88190269e-01
1.25642255e-01 -1.82058856e-01 -1.10874474e+00 8.01729143e-01
1.55664176e-01 -9.58585143e-01 -1.43219624e-02 -2.41651083e-03
8.23078752e-01 -1.18846737e-01 3.48529339e-01 3.03240150e-01
-5.91797493e-02 -7.02433467e-01 8.52519333e-01 5.84379315e-01
2.28918284e-01 -8.01395595e-01 4.48072791e-01 6.66247845e-01
-9.47151601e-01 -2.49500930e-01 -4.23210770e-01 1.63084134e-01
1.38505414e-01 6.20173097e-01 -7.63072193e-01 7.09312856e-01
3.20443302e-01 4.89041716e-01 -6.70189500e-01 1.35610151e+00
1.42482787e-01 8.66166055e-01 -4.39609319e-01 2.14933440e-01
5.47841847e-01 -1.69147059e-01 4.51555699e-01 1.37915838e+00
3.90892565e-01 -4.54701692e-01 2.76598454e-01 4.69813436e-01
-9.93332081e-03 3.85742128e-01 -1.98534980e-01 2.92015314e-01
2.95850188e-01 1.19323611e+00 -1.13968229e+00 -3.62123698e-01
-4.99048561e-01 1.21689165e+00 6.28164589e-01 3.69769573e-01
-6.97879255e-01 -2.46857747e-01 1.48371041e-01 1.88947350e-01
6.74317241e-01 -3.52458805e-01 -1.74605668e-01 -1.07106626e+00
3.87384385e-01 -7.72274733e-01 3.74915898e-01 -1.96875677e-01
-1.21347034e+00 6.59738779e-01 1.77578285e-01 -1.16985726e+00
-2.80709118e-01 -3.80980521e-01 -4.25681949e-01 7.42706895e-01
-1.86566186e+00 -1.33631647e+00 -5.95146343e-02 6.66240454e-01
9.42993164e-01 -1.13698877e-01 6.22887075e-01 4.65826243e-01
-3.70876372e-01 7.60091603e-01 4.76816744e-01 -3.86235982e-01
7.75478482e-01 -1.32445455e+00 1.78302214e-01 9.77644503e-01
4.97186810e-01 9.92210925e-01 6.80018127e-01 -3.61228347e-01
-8.71544302e-01 -1.09464896e+00 7.80192077e-01 -1.89427093e-01
5.25789917e-01 -5.89578152e-01 -9.93430495e-01 5.32862365e-01
1.00515731e-01 1.58178896e-01 6.28121853e-01 1.71712562e-01
-9.50215980e-02 -2.96049744e-01 -9.89702284e-01 3.72897565e-01
1.00325203e+00 -3.16538543e-01 -4.59929556e-01 4.00086731e-01
4.32312429e-01 -6.03662618e-02 -2.86208928e-01 3.92440915e-01
5.30972660e-01 -9.36642587e-01 9.99333918e-01 -5.61014652e-01
4.78042066e-02 -3.15679163e-01 -1.53774783e-01 -1.13937068e+00
-2.63702810e-01 -5.81137955e-01 -2.60745168e-01 1.38469136e+00
2.21324295e-01 -4.10598248e-01 8.01677048e-01 2.81985044e-01
-1.63192391e-01 -8.29198182e-01 -1.07285333e+00 -8.45957220e-01
1.84187546e-01 -5.16452551e-01 3.63216549e-02 5.54508388e-01
-4.54280406e-01 1.80275217e-01 -5.57060659e-01 2.25870803e-01
8.29045713e-01 1.07761525e-01 6.56581283e-01 -1.02086902e+00
-8.31940413e-01 -4.44542974e-01 -2.90511191e-01 -1.30293632e+00
6.60268441e-02 -7.88817644e-01 4.22346480e-02 -1.36154270e+00
4.98980999e-01 -6.09499812e-01 -6.54727340e-01 3.82860899e-01
-4.36579019e-01 2.08557591e-01 3.86947781e-01 4.09872860e-01
-7.94825792e-01 7.42101312e-01 7.21723378e-01 -4.58211042e-02
-3.27166289e-01 2.73960233e-01 -6.34355187e-01 7.21647322e-01
7.20261812e-01 -6.07534766e-01 -8.49131823e-01 -3.90540749e-01
-1.49514839e-01 -4.61800188e-01 3.37403476e-01 -1.18805611e+00
4.75600421e-01 2.86502857e-02 2.73695081e-01 -5.64152181e-01
5.20714104e-01 -9.13014531e-01 -3.26402962e-01 1.63337871e-01
-4.66239214e-01 -3.07341248e-01 1.69690937e-01 7.80385375e-01
-1.74861073e-01 -5.67904234e-01 7.40443468e-01 -1.56493634e-02
-4.55937266e-01 1.60932958e-01 -1.83513299e-01 -2.30043486e-01
1.09284091e+00 -1.73818931e-01 1.64469630e-01 -2.29014263e-01
-9.56685364e-01 1.97123334e-01 3.49617124e-01 2.53355801e-01
4.53582615e-01 -9.55827773e-01 -6.46431983e-01 3.17827053e-02
-6.64599314e-02 8.74484032e-02 3.82992208e-01 9.06440854e-01
-1.65065736e-01 3.00451815e-01 5.06673381e-02 -7.42444992e-01
-1.24354291e+00 6.97922409e-01 2.67869115e-01 -4.72652942e-01
-4.84743804e-01 1.06166387e+00 3.48377615e-01 -3.57633293e-01
5.81568062e-01 2.38408819e-02 -2.59652257e-01 2.96469688e-01
3.95934701e-01 3.51442337e-01 3.68496001e-01 -2.78724641e-01
-2.37358674e-01 4.73298788e-01 -4.17765409e-01 -3.95380348e-01
1.62986934e+00 -3.63594651e-01 1.06033131e-01 6.97983563e-01
1.28124201e+00 2.37606168e-01 -1.60008574e+00 -2.44250655e-01
1.45376772e-01 -3.86972964e-01 2.33528778e-01 -5.64062417e-01
-7.74935007e-01 9.96613920e-01 9.14817750e-01 -1.83876511e-02
1.26011145e+00 -1.97952732e-01 5.25030434e-01 4.30135071e-01
2.29171157e-01 -1.12294722e+00 1.33635625e-01 1.61527321e-01
5.90932727e-01 -1.12883079e+00 2.25283932e-02 -3.08944732e-01
-1.82492867e-01 1.13265824e+00 3.04311067e-01 -1.64662257e-01
6.80881917e-01 1.59321502e-01 -1.42124727e-01 1.95137024e-01
-7.46432304e-01 -3.02784413e-01 2.81452209e-01 3.79205465e-01
4.75426137e-01 -3.25657964e-01 -5.23580909e-01 1.94228441e-01
7.61867017e-02 -8.18043575e-02 2.68998712e-01 1.19824445e+00
-5.47190368e-01 -1.40665317e+00 -2.73633957e-01 1.91409022e-01
-4.79720652e-01 -2.21000597e-01 4.47526053e-02 7.28020668e-01
2.76401341e-01 9.76399720e-01 1.03579406e-02 1.01712734e-01
1.77651346e-01 4.70486164e-01 6.05553448e-01 -5.40226400e-01
-2.10966468e-01 5.51714778e-01 -3.19034874e-01 -1.23532049e-01
-5.79442501e-01 -1.18276775e+00 -9.30846572e-01 2.90309787e-01
-5.32676935e-01 1.71780467e-01 8.09059203e-01 9.56131041e-01
2.43973315e-01 4.98088926e-01 7.75237024e-01 -8.75642419e-01
-1.00653350e+00 -9.21436489e-01 -3.36655438e-01 2.61165082e-01
5.43688595e-01 -6.67506576e-01 -3.44903201e-01 2.72705287e-01] | [9.443745613098145, 2.3882265090942383] |
b07fab74-9de0-40bf-a6a2-9be73f62b984 | monocular-3d-human-pose-estimation-for-sports | 2304.04437 | null | https://arxiv.org/abs/2304.04437v1 | https://arxiv.org/pdf/2304.04437v1.pdf | Monocular 3D Human Pose Estimation for Sports Broadcasts using Partial Sports Field Registration | The filming of sporting events projects and flattens the movement of athletes in the world onto a 2D broadcast image. The pixel locations of joints in these images can be detected with high validity. Recovering the actual 3D movement of the limbs (kinematics) of the athletes requires lifting these 2D pixel locations back into a third dimension, implying a certain scene geometry. The well-known line markings of sports fields allow for the calibration of the camera and for determining the actual geometry of the scene. Close-up shots of athletes are required to extract detailed kinematics, which in turn obfuscates the pertinent field markers for camera calibration. We suggest partial sports field registration, which determines a set of scene-consistent camera calibrations up to a single degree of freedom. Through joint optimization of 3D pose estimation and camera calibration, we demonstrate the successful extraction of 3D running kinematics on a 400m track. In this work, we combine advances in 2D human pose estimation and camera calibration via partial sports field registration to demonstrate an avenue for collecting valid large-scale kinematic datasets. We generate a synthetic dataset of more than 10k images in Unreal Engine 5 with different viewpoints, running styles, and body types, to show the limitations of existing monocular 3D HPE methods. Synthetic data and code are available at https://github.com/tobibaum/PartialSportsFieldReg_3DHPE. | ['Stefanie Klatt', 'Tobias Baumgartner'] | 2023-04-10 | null | null | null | null | ['camera-calibration', '3d-pose-estimation', '3d-human-pose-estimation', 'monocular-3d-human-pose-estimation', '2d-human-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-2.24800944e-01 -2.78983742e-01 -1.69989333e-01 -1.78856522e-01
-7.17310548e-01 -8.63422513e-01 2.58678794e-01 -2.84752667e-01
-4.54048753e-01 3.06406289e-01 1.37245536e-01 1.51466191e-01
7.73503855e-02 -4.03052568e-01 -1.02263474e+00 -1.71597406e-01
-9.80660915e-02 7.22733974e-01 3.77912730e-01 -3.26578349e-01
2.45195910e-01 6.94632173e-01 -1.27159369e+00 -3.87358628e-02
1.00049496e-01 4.70325291e-01 -1.72667317e-02 1.12281382e+00
4.99343961e-01 3.35858345e-01 -6.54098511e-01 -7.09187865e-01
7.52127290e-01 -3.85281593e-01 -4.01125133e-01 5.15033484e-01
9.22176957e-01 -4.44883645e-01 -7.72556961e-01 6.88731611e-01
3.78250211e-01 1.25804335e-01 1.37831315e-01 -1.19934046e+00
6.34317920e-02 1.34235295e-02 -8.94218564e-01 5.94293252e-02
8.73390257e-01 1.72478870e-01 6.40179157e-01 -7.70466685e-01
1.03644168e+00 9.09232497e-01 1.04190981e+00 1.84825078e-01
-8.21623683e-01 -6.46365047e-01 -3.32369238e-01 1.02579802e-01
-1.73081541e+00 -1.90487072e-01 9.11104262e-01 -4.65181381e-01
3.83594036e-01 3.77619565e-01 1.34612286e+00 1.03633320e+00
3.78764719e-01 3.23625118e-01 9.09444213e-01 -5.56637824e-01
-3.57921049e-02 -2.29042545e-01 3.67189273e-02 7.75418580e-01
3.00286710e-01 1.37817204e-01 -1.14839649e+00 1.39835715e-01
1.37968671e+00 -4.09784429e-02 -9.12615284e-02 -6.71350241e-01
-1.37531650e+00 4.43943739e-01 1.83131352e-01 -3.94756317e-01
-9.62589681e-02 4.14636344e-01 1.07013635e-01 -1.22708552e-01
9.09439549e-02 4.07982826e-01 -1.32626876e-01 -5.39392591e-01
-7.55227864e-01 5.97460508e-01 5.46718240e-01 1.37279296e+00
7.72336364e-01 -6.29996136e-03 3.91461790e-01 3.11747402e-01
-3.81918810e-02 7.77959466e-01 -1.99757237e-02 -1.61548889e+00
5.97500145e-01 4.99568909e-01 7.31394216e-02 -1.03514636e+00
-6.67943835e-01 -3.68923396e-01 -1.44602954e-01 2.29433522e-01
9.28515136e-01 -2.95031726e-01 -5.34433544e-01 1.23972261e+00
6.83947682e-01 1.09074764e-01 -2.53789485e-01 1.47149146e+00
3.94036472e-01 3.07427585e-01 -5.41398048e-01 4.34968799e-01
1.70941973e+00 -5.78103960e-01 -3.20333302e-01 -5.81068218e-01
3.94098282e-01 -9.83697891e-01 9.28522289e-01 4.31655169e-01
-1.08945918e+00 -4.82050657e-01 -1.11456668e+00 -1.09454989e-01
1.93146527e-01 3.90286386e-01 6.00103199e-01 5.73466659e-01
-4.16981786e-01 2.95307308e-01 -9.48818386e-01 -3.16030443e-01
-1.28481761e-01 2.42362112e-01 -8.31358194e-01 -5.26495650e-02
-1.02528811e+00 9.45297420e-01 1.59045339e-01 4.11495537e-01
-6.83157146e-01 -6.84553027e-01 -9.10500169e-01 -4.32437509e-01
8.04232657e-01 -7.29423106e-01 1.33919501e+00 -5.94746232e-01
-1.41114569e+00 1.07586443e+00 3.43668401e-01 -1.30763516e-01
7.59284556e-01 -6.13665342e-01 -1.87505379e-01 5.01183033e-01
1.11713931e-01 4.55550164e-01 5.30281901e-01 -1.11392224e+00
-5.22411168e-01 -5.18398583e-01 2.26978451e-01 7.08412647e-01
1.40907958e-01 -1.01741232e-01 -1.24977362e+00 -3.47373515e-01
5.43904901e-01 -1.39515483e+00 -2.62227267e-01 1.35286957e-01
-5.87991536e-01 7.70151615e-01 3.66595715e-01 -9.54768658e-01
8.02622318e-01 -1.99068582e+00 3.80221903e-01 2.26556540e-01
9.67893600e-02 -2.63988674e-01 3.03275108e-01 3.48388255e-01
1.01526655e-01 -4.98969585e-01 9.79517475e-02 -1.92540824e-01
-2.68220156e-01 2.50833452e-01 1.74642831e-01 1.06754172e+00
-2.69426376e-01 6.64629757e-01 -6.00594401e-01 -6.54316008e-01
2.21039280e-01 4.30449724e-01 -5.85493207e-01 9.91664454e-02
2.14938223e-01 5.02622485e-01 -3.57536137e-01 6.98198318e-01
5.97964704e-01 3.03324550e-01 2.14214101e-01 -3.07454735e-01
-2.59035647e-01 -8.07259306e-02 -1.74151218e+00 2.22962809e+00
-2.02236418e-02 6.32876933e-01 1.55466229e-01 -4.75104213e-01
7.31042325e-01 9.99972448e-02 7.36745596e-01 -4.40096557e-01
1.89919472e-01 -5.17089069e-02 -2.97438323e-01 -3.30492288e-01
9.58585024e-01 -6.89399019e-02 -6.17890656e-01 8.66169110e-02
-1.35814603e-02 -4.88872916e-01 1.47785902e-01 1.50479704e-01
7.51552343e-01 5.85070789e-01 -9.61165875e-03 -6.57549128e-02
-7.95574412e-02 6.13469899e-01 5.98695993e-01 3.21112156e-01
4.96512540e-02 1.09237289e+00 4.32968229e-01 -4.39189613e-01
-1.50363362e+00 -1.23517001e+00 1.18380301e-01 3.12077433e-01
5.11927366e-01 -7.18383312e-01 -1.03460240e+00 1.21110454e-02
3.99212576e-02 1.25060692e-01 -4.33244944e-01 -1.98355690e-01
-8.81080389e-01 -3.46709073e-01 7.33695984e-01 6.48393750e-01
3.55307668e-01 -1.45874128e-01 -1.14284682e+00 -9.56275389e-02
-4.50516224e-01 -1.53898013e+00 -7.50083804e-01 -9.85859782e-02
-7.18049645e-01 -1.50813866e+00 -6.56668723e-01 -4.55067664e-01
6.19815648e-01 3.67761940e-01 9.14260089e-01 -1.83061436e-01
-5.39043367e-01 6.37140512e-01 -1.63594633e-01 -1.72982961e-01
-5.86518049e-02 -1.26983598e-01 3.61410260e-01 -2.22518712e-01
-1.25913888e-01 -3.27703178e-01 -6.23213410e-01 7.08601415e-01
-3.75051200e-01 2.91151434e-01 2.12418020e-01 4.32580948e-01
8.28775764e-01 -1.86296791e-01 -5.87052226e-01 -4.66510236e-01
-6.94455430e-02 1.31542549e-01 -8.53844881e-01 6.49081320e-02
2.43566468e-01 -3.26739728e-01 9.72350612e-02 -4.75209892e-01
-9.30746138e-01 6.37466788e-01 1.69398978e-01 -7.72360504e-01
-4.13460135e-02 1.61358073e-01 5.42138331e-02 -9.37933177e-02
9.82436538e-01 -4.77500968e-02 6.10713586e-02 -4.18980628e-01
4.49978471e-01 2.99673140e-01 1.22743738e+00 -8.05931926e-01
9.91728306e-01 7.73478270e-01 3.07152152e-01 -9.54808652e-01
-7.46199667e-01 -7.22475648e-01 -1.32225204e+00 -7.14968503e-01
1.08178723e+00 -1.14641714e+00 -9.43438530e-01 4.00247037e-01
-1.00268745e+00 -2.54401505e-01 -4.55148578e-01 1.10567498e+00
-7.48297572e-01 5.32742083e-01 -6.24815524e-01 -5.36039710e-01
1.03666432e-01 -1.14786696e+00 1.30289972e+00 1.29362717e-01
-3.90908450e-01 -5.59044361e-01 2.61893392e-01 7.24113405e-01
-5.91496289e-01 4.95238900e-01 -5.40161058e-02 3.58034462e-01
-6.66554749e-01 -4.66506898e-01 5.36230564e-01 -3.32889929e-02
-4.04920101e-01 2.43725568e-01 -6.84465945e-01 -1.12184234e-01
-1.76369578e-01 1.60760432e-02 1.46272346e-01 4.68846411e-01
5.13326764e-01 2.79243827e-01 -2.92949844e-02 1.12093794e+00
1.15453565e+00 -2.67811805e-01 7.15412021e-01 6.29942894e-01
9.62321937e-01 6.81101680e-01 9.46571529e-01 5.77205181e-01
4.92072105e-01 1.24558377e+00 3.31304610e-01 -1.35561094e-01
-3.30550641e-01 -6.92595720e-01 3.69300067e-01 7.29372740e-01
-6.02390647e-01 3.62068295e-01 -1.02318311e+00 1.50805950e-01
-1.49373579e+00 -8.03876519e-01 -6.42716050e-01 2.41684532e+00
5.51743567e-01 1.60028890e-01 3.78731489e-01 -3.22598778e-02
8.21594954e-01 -2.48005480e-01 -4.00091112e-01 -1.40783444e-01
-2.22291481e-02 -1.11441240e-01 1.21585238e+00 5.90805948e-01
-8.48861933e-01 8.49875331e-01 6.22359705e+00 4.82171059e-01
-8.36831152e-01 -7.89041892e-02 -5.38131185e-02 -3.66917163e-01
1.07107490e-01 4.44699168e-01 -1.03339934e+00 1.16934270e-01
6.84408426e-01 1.83160044e-02 1.82331979e-01 7.88765252e-01
8.14973116e-02 -5.05430877e-01 -8.66259813e-01 1.23463106e+00
2.08665550e-01 -1.34469640e+00 -5.71606815e-01 2.48539537e-01
5.53149343e-01 -2.33993560e-01 -2.51040995e-01 -2.75240809e-01
5.59990481e-02 -4.46006656e-01 1.45941353e+00 6.17717624e-01
8.21531355e-01 -6.52592897e-01 3.19109499e-01 5.05316079e-01
-1.29309869e+00 1.19495727e-01 -3.90107393e-01 -3.33474219e-01
4.07231808e-01 2.23078698e-01 -7.97098398e-01 7.34515548e-01
7.50815511e-01 5.95230162e-01 -6.37905598e-01 9.69909668e-01
-3.55896413e-01 3.13241214e-01 -6.04626834e-01 4.39027756e-01
-1.13182694e-01 -5.21345139e-01 6.08646095e-01 7.93724895e-01
3.79572481e-01 2.96707124e-01 1.55054986e-01 4.72016931e-01
1.03675976e-01 -1.24266952e-01 -3.37670177e-01 4.32306349e-01
2.33644709e-01 1.28561747e+00 -9.91924107e-01 -4.74344343e-02
-1.69759601e-01 9.39279199e-01 -6.56432360e-02 2.96375155e-02
-1.10445249e+00 -2.13136580e-02 7.67030060e-01 6.80444419e-01
-2.80572679e-02 -9.40209270e-01 -3.10643017e-01 -1.38661051e+00
4.13816720e-01 -8.27284098e-01 2.94801146e-01 -1.08753693e+00
-5.65392256e-01 1.33920789e-01 4.17427182e-01 -1.59224200e+00
-2.17460632e-01 -5.11848390e-01 -1.86236173e-01 6.95632875e-01
-7.44869471e-01 -1.19394445e+00 -5.25240242e-01 7.03997672e-01
4.68074441e-01 3.07460785e-01 4.60465521e-01 2.92986423e-01
-3.23833823e-01 3.80699068e-01 -1.77566633e-01 3.45227182e-01
8.13599110e-01 -1.01796222e+00 5.97731948e-01 9.76879418e-01
4.43061084e-01 3.18642795e-01 8.16268981e-01 -8.16000223e-01
-2.00498748e+00 -4.28148806e-01 4.06412989e-01 -1.04183316e+00
4.77238566e-01 -7.32196927e-01 -3.27562809e-01 9.46969748e-01
-4.69791353e-01 -1.38052404e-01 3.19611073e-01 -5.54211773e-02
-4.72982228e-02 -2.82784671e-01 -5.07937014e-01 6.21836960e-01
1.11055017e+00 -3.62820089e-01 -6.03649199e-01 4.40400928e-01
1.47987872e-01 -1.37544262e+00 -1.00303376e+00 -1.31207600e-01
9.66362953e-01 -9.16014194e-01 1.39572740e+00 -2.59412915e-01
2.70525753e-01 -5.52006721e-01 -8.85286257e-02 -9.49055314e-01
1.78482309e-01 -6.61496580e-01 4.52726334e-01 8.30138206e-01
6.66105561e-03 7.02545187e-03 1.31295288e+00 7.19802797e-01
-2.16685057e-01 -2.94790626e-01 -1.04742408e+00 -7.43019164e-01
-1.59000933e-01 -7.51591086e-01 1.56513631e-01 7.39816427e-01
6.46417439e-02 2.67473664e-02 -7.76415527e-01 5.24296880e-01
8.41951132e-01 1.50103107e-01 1.48041391e+00 -8.68697107e-01
-5.44180930e-01 1.72148168e-01 -9.76958215e-01 -1.12378633e+00
-2.79115915e-01 -4.47524965e-01 -1.27298012e-01 -1.07386422e+00
7.56767578e-03 -2.21575797e-01 7.34920382e-01 1.76967368e-01
1.49892315e-01 5.85766912e-01 3.63259643e-01 3.42431337e-01
-3.17996204e-01 1.46332923e-02 1.36084938e+00 4.02990460e-01
5.69195747e-02 1.39288623e-02 -1.86982006e-01 9.83176351e-01
5.64029574e-01 -4.78675663e-01 -2.02428535e-01 -6.61616147e-01
2.81310737e-01 5.00301242e-01 7.26414502e-01 -1.32349229e+00
3.28828871e-01 -1.93436474e-01 6.64420903e-01 -6.35605037e-01
8.67794633e-01 -7.71158159e-01 9.04988348e-01 4.90757078e-01
-1.06567867e-01 4.05272335e-01 2.70147193e-02 3.18285018e-01
-2.65566278e-02 -1.94320008e-01 4.26935673e-01 -5.13020158e-01
-9.41396892e-01 1.50161192e-01 -4.36206721e-02 1.95608556e-01
1.14632452e+00 -8.40362370e-01 -4.59518284e-02 -4.70801562e-01
-7.95510411e-01 1.47115335e-01 9.40773904e-01 2.85853267e-01
6.58217788e-01 -1.17652798e+00 -6.25553668e-01 3.44905853e-01
4.45574820e-02 2.28217870e-01 5.76130569e-01 8.79132867e-01
-1.38623667e+00 7.04842061e-02 -5.67411661e-01 -1.03798532e+00
-1.46789670e+00 1.81421980e-01 4.88930285e-01 3.02410841e-01
-1.00596929e+00 5.35468817e-01 -1.01385981e-01 -3.41746539e-01
-1.82874426e-01 -1.77324295e-01 3.00061285e-01 -1.84785917e-01
2.99565017e-01 6.90857053e-01 1.33557573e-01 -1.03036952e+00
-3.82751822e-01 1.22605813e+00 4.59826469e-01 -4.91507679e-01
1.12995720e+00 -1.75626710e-01 4.80245918e-01 1.87429249e-01
9.79646206e-01 4.60484743e-01 -1.76211917e+00 1.00085899e-01
-4.22637135e-01 -8.32634985e-01 -2.71054924e-01 -2.78839380e-01
-1.03664494e+00 9.22829092e-01 3.41405898e-01 -4.83796090e-01
7.85798192e-01 1.00668035e-01 5.71567953e-01 1.11858852e-01
6.39207780e-01 -1.33661091e+00 -1.13560624e-01 1.67338163e-01
6.16593301e-01 -8.03050876e-01 4.85058576e-01 -7.54195690e-01
-8.83979678e-01 1.32562494e+00 5.66471338e-01 -3.15403849e-01
2.43745402e-01 6.10377252e-01 4.00060236e-01 -4.00986880e-01
-2.63859313e-02 1.17531136e-01 3.17231834e-01 5.75297177e-01
4.02669283e-03 9.56745371e-02 -1.43924356e-01 2.86078781e-01
-7.18466699e-01 -2.68081754e-01 9.05719757e-01 1.19582164e+00
-1.06616370e-01 -1.02623379e+00 -1.03006995e+00 -3.39337826e-01
-1.98491633e-01 4.65525717e-01 -3.73963028e-01 1.17310643e+00
1.21641181e-01 5.72800040e-01 9.45496708e-02 -5.37222981e-01
9.60791886e-01 -2.44634554e-01 8.89736116e-01 -4.89955455e-01
-3.75099480e-01 2.26915151e-01 4.04484570e-01 -7.65331268e-01
-3.01232934e-01 -9.06894088e-01 -1.25150073e+00 -4.22743618e-01
-2.24275246e-01 7.04290867e-02 9.76694584e-01 5.52083790e-01
-2.80563403e-02 2.15753973e-01 3.05993527e-01 -1.21850300e+00
-4.83383864e-01 -3.78613472e-01 -8.40610623e-01 5.39590359e-01
-1.07229397e-01 -8.24345946e-01 -1.30989328e-02 4.77309346e-01] | [7.329680442810059, -1.0851832628250122] |
db03938e-26e8-4d33-a160-735a38f91eec | systematic-visual-reasoning-through-object | 2306.02500 | null | https://arxiv.org/abs/2306.02500v1 | https://arxiv.org/pdf/2306.02500v1.pdf | Systematic Visual Reasoning through Object-Centric Relational Abstraction | Human visual reasoning is characterized by an ability to identify abstract patterns from only a small number of examples, and to systematically generalize those patterns to novel inputs. This capacity depends in large part on our ability to represent complex visual inputs in terms of both objects and relations. Recent work in computer vision has introduced models with the capacity to extract object-centric representations, leading to the ability to process multi-object visual inputs, but falling short of the systematic generalization displayed by human reasoning. Other recent models have employed inductive biases for relational abstraction to achieve systematic generalization of learned abstract rules, but have generally assumed the presence of object-focused inputs. Here, we combine these two approaches, introducing Object-Centric Relational Abstraction (OCRA), a model that extracts explicit representations of both objects and abstract relations, and achieves strong systematic generalization in tasks involving complex visual displays. | ['Jonathan D. Cohen', 'Shanka Subhra Mondal', 'Taylor W. Webb'] | 2023-06-04 | null | null | null | null | ['visual-reasoning', 'systematic-generalization', 'visual-reasoning'] | ['computer-vision', 'reasoning', 'reasoning'] | [ 3.22310030e-01 3.32593590e-01 1.36676565e-01 -4.90159810e-01
-6.11430109e-02 -7.41467953e-01 1.04126239e+00 3.86181831e-01
-3.16266268e-01 3.73119026e-01 3.08919787e-01 -3.03493261e-01
-4.55224097e-01 -7.68784881e-01 -5.64905286e-01 -3.04870218e-01
-1.49638802e-02 6.36431932e-01 4.17026579e-01 -1.86242789e-01
3.66614819e-01 1.04610753e+00 -1.86462831e+00 8.95218134e-01
6.02488458e-01 7.31924415e-01 2.58509248e-01 7.63971031e-01
-2.10842147e-01 1.13125801e+00 -4.92627174e-01 -2.47158155e-01
7.40173236e-02 -3.91196907e-01 -8.98158669e-01 2.78010927e-02
7.65233278e-01 -1.55606940e-01 -1.85238615e-01 7.67018020e-01
-4.63637384e-03 1.47683457e-01 7.26941824e-01 -1.09393513e+00
-1.42233026e+00 6.67635441e-01 -1.51507840e-01 5.02834797e-01
5.80030203e-01 5.46677172e-01 1.27069497e+00 -1.04097450e+00
6.32295966e-01 1.73531044e+00 4.02249634e-01 6.23707771e-01
-1.91310179e+00 -4.71797138e-01 6.37224376e-01 5.57643324e-02
-1.20913982e+00 -2.22165674e-01 5.82519293e-01 -6.40988946e-01
1.41958046e+00 4.26138073e-01 1.06714761e+00 8.36916268e-01
-3.15099843e-02 6.75182223e-01 1.71674955e+00 -4.73568767e-01
2.16749400e-01 4.59795743e-01 2.94724733e-01 4.88543838e-01
6.52633607e-01 3.22672486e-01 -6.57827377e-01 7.64442608e-02
9.10854101e-01 2.79108912e-01 -3.16168666e-01 -2.94730365e-01
-1.22698343e+00 4.49396908e-01 9.92029428e-01 3.35461557e-01
-4.61166859e-01 -7.43341865e-03 1.14202976e-01 4.76710111e-01
-1.44348636e-01 1.05374134e+00 -2.44340554e-01 4.75728661e-01
-5.06836832e-01 3.83055657e-01 5.64866006e-01 1.12072325e+00
9.07593131e-01 1.92778766e-01 -3.90249521e-01 6.36114180e-01
2.13932306e-01 6.05589509e-01 3.35115910e-01 -1.02928948e+00
2.74998456e-01 1.09421182e+00 -1.36666283e-01 -9.54560220e-01
-3.81672144e-01 -4.79976118e-01 -5.18658996e-01 7.02809036e-01
2.96035141e-01 4.16295677e-01 -1.11631179e+00 1.98592067e+00
-1.12473823e-01 -8.48314583e-01 2.07249224e-01 8.41576219e-01
8.36645067e-01 4.57845777e-01 5.35010040e-01 2.89777517e-02
1.32011569e+00 -4.24510151e-01 -3.25471640e-01 -5.48447728e-01
4.74859864e-01 -3.83942351e-02 1.45977080e+00 4.04635131e-01
-1.30041623e+00 -8.59439373e-01 -1.11055923e+00 -4.73886669e-01
-9.05095637e-01 -1.99764654e-01 8.56705308e-01 3.98998886e-01
-1.26759732e+00 4.42254215e-01 -2.10815802e-01 -5.42588472e-01
8.87109637e-01 3.44525129e-01 -3.71978372e-01 -1.87425375e-01
-8.56000364e-01 1.22656560e+00 6.85094237e-01 2.61242203e-02
-8.11724544e-01 -6.54764593e-01 -9.87543344e-01 4.01210845e-01
2.47626245e-01 -1.03706789e+00 1.09690261e+00 -1.24816310e+00
-8.02291274e-01 1.06438315e+00 -1.18854150e-01 -1.87700838e-01
1.03323525e-02 -1.45667940e-01 -2.41793588e-01 3.23200524e-01
-1.54232323e-01 9.53173161e-01 8.29114437e-01 -1.80109704e+00
-3.84974390e-01 -5.36965013e-01 4.29927677e-01 3.07879239e-01
-4.76572625e-02 -2.20084324e-01 1.99955225e-01 -5.48239291e-01
3.47925007e-01 -8.18761587e-01 8.64953250e-02 1.83874756e-01
1.85104981e-02 -3.43022734e-01 5.88970304e-01 -2.65114129e-01
9.63455021e-01 -2.21829724e+00 4.50774521e-01 2.22395092e-01
4.83701319e-01 1.68823302e-01 -4.04020958e-02 4.81525362e-01
-2.60991603e-01 1.99106827e-01 -1.05357304e-01 -9.43890959e-02
1.77439362e-01 4.70654309e-01 -7.68502295e-01 -1.65048540e-01
5.90380609e-01 1.28831911e+00 -1.05153084e+00 -3.12640429e-01
-8.85221213e-02 4.90107341e-03 -5.25154054e-01 4.82370239e-03
-2.75075793e-01 -5.58298640e-02 -1.93688720e-02 4.85935092e-01
4.04178917e-01 -3.70499134e-01 3.36641967e-01 -1.72973890e-02
-1.30638015e-02 4.11249191e-01 -9.47887778e-01 1.17292070e+00
-3.90088528e-01 8.28252017e-01 -3.71598303e-01 -5.60904503e-01
8.55613291e-01 1.83205485e-01 -3.56672168e-01 -8.72557700e-01
-1.12169869e-01 7.22724050e-02 6.01854026e-01 -4.13514405e-01
1.73103184e-01 -6.26828253e-01 1.96711287e-01 6.31229699e-01
1.45967498e-01 -6.46627903e-01 8.65124762e-02 5.86493611e-01
7.46551156e-01 2.20094338e-01 4.22873527e-01 -1.68813497e-01
3.67038846e-01 4.83424665e-05 2.28401631e-01 1.14999807e+00
2.17749491e-01 4.00243342e-01 4.75089669e-01 -6.82663918e-01
-9.27454412e-01 -1.50845420e+00 -1.50461057e-02 1.20862305e+00
-3.01146526e-02 -3.55649859e-01 -1.08730242e-01 -5.38411498e-01
2.56760716e-01 1.16716027e+00 -1.00435758e+00 -4.16740030e-01
-3.94297332e-01 -2.35238314e-01 1.93023354e-01 9.68579054e-01
9.55914631e-02 -1.75728476e+00 -1.00499237e+00 -1.95499003e-01
4.24061060e-01 -8.76081169e-01 1.32720307e-01 2.91386545e-01
-9.95165527e-01 -8.75089288e-01 -2.79910833e-01 -5.67100704e-01
9.34346914e-01 2.94562250e-01 1.26100445e+00 4.61998343e-01
-4.61989909e-01 7.83416808e-01 -1.33905977e-01 -8.80878270e-01
-4.12211210e-01 -3.27648997e-01 -9.84924212e-02 -1.62852034e-01
5.17091632e-01 -6.05744362e-01 -2.09562749e-01 -2.14029793e-02
-1.11964250e+00 2.69101262e-01 9.95791018e-01 6.39300168e-01
2.06207931e-01 -3.12041491e-01 6.57110214e-01 -8.95885289e-01
7.24829018e-01 -2.33146295e-01 -2.93173105e-01 2.55060643e-01
-4.32084888e-01 3.67270350e-01 5.32839537e-01 -7.54641593e-01
-1.28539205e+00 -2.36876428e-01 5.08060932e-01 -3.33627284e-01
-3.48536015e-01 4.86676961e-01 2.88079288e-02 6.19000383e-02
1.08587086e+00 4.47957873e-01 -1.70518473e-01 -1.25562862e-01
6.01019382e-01 2.04994932e-01 2.51805156e-01 -1.01475966e+00
8.64666343e-01 5.89935482e-01 3.05071205e-01 -8.32244158e-01
-1.02640522e+00 -4.82798964e-02 -9.55886066e-01 -7.84092918e-02
7.01648951e-01 -7.55065382e-01 -8.50735962e-01 3.98383476e-02
-1.09385622e+00 -4.33000058e-01 -8.53857696e-01 4.52613860e-01
-6.97130680e-01 1.58381891e-02 -5.00699937e-01 -8.70337844e-01
2.18544275e-01 -8.17675114e-01 7.27226198e-01 1.28352419e-01
-6.66025996e-01 -7.86167145e-01 -3.05005997e-01 -1.62688956e-01
2.73748964e-01 5.68891056e-02 1.51500022e+00 -6.87287509e-01
-1.06920636e+00 7.23090488e-03 -5.48524916e-01 2.32788920e-01
7.67307207e-02 -1.43640369e-01 -1.24789786e+00 -1.29236817e-01
-7.20727220e-02 -6.57913804e-01 1.16637218e+00 -1.24155983e-01
1.02898812e+00 -2.82898396e-01 -3.24402750e-01 4.21871990e-01
1.30964243e+00 3.09496135e-01 6.68438613e-01 -9.32654890e-04
5.83550334e-01 8.51010025e-01 6.21703938e-02 -1.14366740e-01
3.28041524e-01 3.56402755e-01 1.63746495e-02 2.65190415e-02
-3.67288440e-01 -2.90665627e-01 -1.65983513e-01 2.53055364e-01
-4.12690789e-01 3.11405033e-01 -1.03166294e+00 7.24353850e-01
-1.61628008e+00 -1.32844150e+00 1.34339839e-01 2.04090643e+00
9.72460270e-01 5.41855931e-01 1.24584483e-02 1.87581271e-01
1.67816818e-01 -1.22154184e-01 -5.02783298e-01 -9.13446128e-01
-2.68739820e-01 3.01508546e-01 -1.82122886e-01 4.72757161e-01
-3.17346424e-01 9.62436378e-01 7.50357199e+00 2.48404220e-02
-7.52814293e-01 -3.33328575e-01 1.19867995e-01 -3.45187783e-02
-6.63103044e-01 2.69596815e-01 -5.42921007e-01 -1.64330080e-01
5.91070533e-01 2.29189266e-02 5.70329845e-01 5.59511960e-01
-5.68209112e-01 -3.21330756e-01 -1.75072408e+00 7.91609228e-01
1.88135967e-01 -1.21220529e+00 1.05140042e+00 -7.93223605e-02
6.11385524e-01 -6.41385913e-01 3.17396939e-01 5.70758164e-01
5.21455169e-01 -1.55368805e+00 7.72963166e-01 9.15128827e-01
5.31293452e-01 -2.34524012e-01 1.76399231e-01 3.58548909e-01
-9.15385187e-01 -5.88586509e-01 -3.33010584e-01 -6.63224757e-01
-3.61904740e-01 -1.81976959e-01 -6.90913618e-01 1.20407447e-01
8.56155574e-01 4.87581074e-01 -9.78691757e-01 7.24437714e-01
-4.80946988e-01 1.85396045e-01 -3.64924520e-02 -1.88837215e-01
1.50486641e-03 1.89702913e-01 4.51518685e-01 1.22307038e+00
-1.16753288e-01 5.18035531e-01 -8.03452730e-02 1.30987835e+00
2.09108114e-01 -2.86159933e-01 -8.47744286e-01 -4.45375107e-02
5.23802280e-01 9.24503922e-01 -5.97026050e-01 -6.23390496e-01
-5.18237770e-01 6.05055153e-01 8.09822559e-01 8.08660150e-01
-2.03125134e-01 -3.01050693e-01 4.43519711e-01 2.42533073e-01
4.20517683e-01 -4.46967244e-01 -6.29682302e-01 -7.65054405e-01
-5.22682779e-02 -8.79382730e-01 5.19145966e-01 -1.36678731e+00
-1.20623875e+00 5.09025872e-01 3.50611150e-01 -6.55266464e-01
2.24162750e-02 -1.04526389e+00 -3.27204168e-01 1.31297457e+00
-1.13049626e+00 -1.36379623e+00 -3.72062236e-01 5.47939301e-01
3.91416520e-01 -8.39931443e-02 8.49032342e-01 -5.39904356e-01
1.48579687e-01 2.60287941e-01 -7.05405354e-01 7.61779025e-02
4.40281630e-01 -1.46558464e+00 2.81913668e-01 4.05086577e-01
4.08972502e-01 1.39529431e+00 6.53065264e-01 -5.19315064e-01
-1.13664865e+00 -5.27989447e-01 8.10720444e-01 -1.20988977e+00
3.46735507e-01 -6.18462443e-01 -1.44285798e+00 1.26038337e+00
2.47459501e-01 3.42226364e-02 5.98945200e-01 4.66050118e-01
-1.12684548e+00 -2.13498861e-01 -1.00531149e+00 1.00562894e+00
1.29420912e+00 -9.00288284e-01 -1.48900449e+00 -1.02422282e-01
7.07826316e-01 7.68061206e-02 -5.93694687e-01 2.47241616e-01
5.60668886e-01 -1.09196198e+00 1.13076735e+00 -1.03293371e+00
2.59755313e-01 -5.12275755e-01 -2.21350476e-01 -1.32261646e+00
-1.09435058e+00 -4.31139022e-02 -1.84183508e-01 8.58274519e-01
4.43258554e-01 -8.51045728e-01 -6.43624039e-03 6.15428090e-01
3.35335359e-02 -5.31904697e-01 -3.68941367e-01 -7.51039445e-01
2.05101836e-02 -1.94914773e-01 5.29289365e-01 7.80359149e-01
2.45173126e-01 7.06749916e-01 3.74651045e-01 1.66291207e-01
4.45058078e-01 4.31729794e-01 5.50182104e-01 -1.64246881e+00
-2.09441036e-01 -7.80391276e-01 -6.20560408e-01 -6.86266363e-01
1.34181991e-01 -1.07349801e+00 -2.15881795e-01 -1.72245026e+00
3.08677673e-01 -3.10716927e-01 -4.43415374e-01 5.12041569e-01
-2.28220806e-01 1.16260327e-01 5.56469798e-01 3.41123939e-01
-3.49642485e-01 2.78517604e-01 1.38057005e+00 -1.25844792e-01
-2.39390418e-01 -3.35588336e-01 -1.19623256e+00 8.45373809e-01
5.76490164e-01 -2.94996172e-01 -6.61526203e-01 -4.23842072e-01
7.09947944e-01 -4.04750317e-01 9.29620802e-01 -1.14430761e+00
2.74101924e-02 -4.26794410e-01 1.16677272e+00 -2.98223048e-01
5.62406659e-01 -1.02581704e+00 -3.02145910e-02 4.75172192e-01
-7.96683073e-01 2.81144202e-01 6.29033685e-01 5.71687460e-01
2.63564778e-03 -9.84400809e-02 8.04769337e-01 -6.42164826e-01
-9.34547007e-01 -1.96338296e-01 -3.58708411e-01 5.12059927e-02
8.26585650e-01 -5.98824263e-01 -5.75061142e-01 -8.02755952e-02
-1.05763233e+00 -6.80019036e-02 4.43710327e-01 4.30029631e-01
8.34848821e-01 -1.36058497e+00 -4.58292335e-01 2.43927613e-01
4.95228797e-01 -2.40955949e-01 4.49132547e-02 4.60513949e-01
-2.62792110e-01 4.75841492e-01 -5.06511927e-01 -6.12176239e-01
-9.88461912e-01 1.36263144e+00 2.29904130e-01 3.74239087e-01
-7.21066058e-01 7.90936053e-01 6.93394959e-01 -2.88258135e-01
-8.48540813e-02 -7.73029566e-01 -1.47334680e-01 -2.52007227e-02
5.94725788e-01 9.61745717e-03 -4.09149170e-01 -4.77422804e-01
-3.45514417e-01 6.61044240e-01 -2.00519979e-01 -1.78248942e-01
9.21568751e-01 5.76269403e-02 -1.73856974e-01 9.77976203e-01
5.80728948e-01 -5.18402681e-02 -1.19707334e+00 -4.68380481e-01
1.58211309e-02 -5.39519489e-01 -5.27447104e-01 -1.21097255e+00
-1.96577609e-01 1.24543846e+00 1.89320073e-01 2.99803287e-01
9.75450814e-01 3.87244701e-01 -2.03753427e-01 7.48941064e-01
2.87293285e-01 -8.19097519e-01 5.16766369e-01 4.56462353e-01
1.37860298e+00 -1.07813013e+00 2.96273679e-01 -3.41999382e-01
-7.37793028e-01 1.15871191e+00 9.61443961e-01 -1.00080192e-01
2.81954139e-01 -1.18975900e-02 -6.09676205e-02 -4.95859057e-01
-1.11246467e+00 -4.72485095e-01 5.29423773e-01 8.89980137e-01
3.25739413e-01 2.45150328e-02 3.18164021e-01 1.45439282e-01
-9.85881612e-02 -1.30355790e-01 3.49199623e-01 1.00245714e+00
-6.46254718e-01 -3.22947919e-01 -3.23178619e-01 3.98889482e-01
1.24379590e-01 -3.47431123e-01 -7.55713224e-01 1.01455855e+00
1.77468374e-01 5.18683970e-01 2.98303872e-01 1.32924557e-01
5.58202386e-01 5.05730271e-01 8.95896971e-01 -9.64346051e-01
-6.75632894e-01 -5.55904210e-01 -2.16180041e-01 -3.69794846e-01
-3.43122900e-01 -6.15389705e-01 -1.29743016e+00 1.90157279e-01
2.63055682e-01 -1.54722407e-01 1.30828962e-01 8.56270313e-01
2.13986531e-01 5.88633060e-01 -9.06079262e-02 -6.65042520e-01
-4.50492382e-01 -8.81712973e-01 -4.93649244e-01 9.84873891e-01
6.78445160e-01 -6.93811834e-01 -3.29282671e-01 5.41071035e-03] | [10.581801414489746, 2.252000093460083] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.