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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3907e8b0-fa04-4a1b-b87e-5e30d00e09a2 | climax-a-foundation-model-for-weather-and | 2301.10343 | null | https://arxiv.org/abs/2301.10343v3 | https://arxiv.org/pdf/2301.10343v3.pdf | ClimaX: A foundation model for weather and climate | Most state-of-the-art approaches for weather and climate modeling are based on physics-informed numerical models of the atmosphere. These approaches aim to model the non-linear dynamics and complex interactions between multiple variables, which are challenging to approximate. Additionally, many such numerical models are computationally intensive, especially when modeling the atmospheric phenomenon at a fine-grained spatial and temporal resolution. Recent data-driven approaches based on machine learning instead aim to directly solve a downstream forecasting or projection task by learning a data-driven functional mapping using deep neural networks. However, these networks are trained using curated and homogeneous climate datasets for specific spatiotemporal tasks, and thus lack the generality of numerical models. We develop and demonstrate ClimaX, a flexible and generalizable deep learning model for weather and climate science that can be trained using heterogeneous datasets spanning different variables, spatio-temporal coverage, and physical groundings. ClimaX extends the Transformer architecture with novel encoding and aggregation blocks that allow effective use of available compute while maintaining general utility. ClimaX is pre-trained with a self-supervised learning objective on climate datasets derived from CMIP6. The pre-trained ClimaX can then be fine-tuned to address a breadth of climate and weather tasks, including those that involve atmospheric variables and spatio-temporal scales unseen during pretraining. Compared to existing data-driven baselines, we show that this generality in ClimaX results in superior performance on benchmarks for weather forecasting and climate projections, even when pretrained at lower resolutions and compute budgets. The source code is available at https://github.com/microsoft/ClimaX. | ['Aditya Grover', 'Jayesh K. Gupta', 'Ashish Kapoor', 'Johannes Brandstetter', 'Tung Nguyen'] | 2023-01-24 | null | null | null | null | ['weather-forecasting'] | ['miscellaneous'] | [-3.84365141e-01 -4.15030539e-01 -1.12005752e-02 -4.24279988e-01
-4.71127808e-01 -8.29728961e-01 8.46894979e-01 1.30175591e-01
-6.15095310e-02 9.33138609e-01 1.86341062e-01 -8.97067189e-01
-5.82816601e-02 -1.27174389e+00 -7.32628822e-01 -7.13636160e-01
-4.42219466e-01 5.86526513e-01 -1.94743909e-02 -5.08432090e-01
-2.79932678e-01 7.42794931e-01 -1.40519464e+00 3.77964862e-02
8.88650894e-01 9.48390305e-01 1.43726215e-01 9.34296429e-01
-1.30787864e-01 5.25414050e-01 -2.41960958e-01 3.18638533e-01
4.48878378e-01 -2.66773999e-01 -6.25856459e-01 -3.87555033e-01
5.37058473e-01 -3.62670779e-01 -3.59244496e-01 5.45338154e-01
4.76350129e-01 3.53725702e-01 5.47832906e-01 -1.03512573e+00
-6.13425195e-01 2.37685919e-01 -2.22907737e-01 4.29759532e-01
-4.14849162e-01 4.86924142e-01 8.93825948e-01 -8.93408000e-01
1.01778895e-01 1.11877239e+00 1.14164805e+00 2.29771271e-01
-1.48161757e+00 -4.84156758e-01 3.94860685e-01 -2.01179460e-01
-1.37697101e+00 -2.63961643e-01 3.26546103e-01 -7.85077691e-01
1.34906483e+00 4.04351324e-01 5.48435330e-01 6.76433027e-01
3.33642423e-01 -1.12839550e-01 1.30631912e+00 8.04081783e-02
3.42383683e-01 -8.61524418e-02 -6.04955480e-02 4.64855731e-01
-5.01211435e-02 6.89567864e-01 -3.96917999e-01 -1.83518738e-01
6.27694786e-01 -2.07875911e-02 -2.77615964e-01 1.49685800e-01
-1.07430375e+00 8.98223102e-01 8.01332355e-01 8.38609934e-02
-5.22328734e-01 3.09359580e-01 2.35907987e-01 2.65831232e-01
1.07002473e+00 6.82493925e-01 -1.12729430e+00 2.74180248e-02
-1.52894998e+00 7.28696585e-01 7.31672585e-01 5.16546726e-01
1.00976956e+00 3.89720649e-01 1.49318010e-01 3.93664837e-01
1.28388852e-01 1.15630889e+00 6.61065951e-02 -1.03203917e+00
3.00204843e-01 3.73565376e-01 3.66421908e-01 -8.63854170e-01
-7.73544073e-01 -5.95402718e-01 -1.21094716e+00 6.17550194e-01
2.84575194e-01 -7.29242146e-01 -1.09009302e+00 1.81297350e+00
5.94894946e-01 3.91943693e-01 5.91817051e-02 1.14434206e+00
6.89322472e-01 1.51578665e+00 7.89879560e-02 9.24364775e-02
9.42960858e-01 -9.81106222e-01 -4.29470122e-01 -3.83718997e-01
5.72111607e-01 -4.19339597e-01 1.25348842e+00 -1.54865012e-01
-7.31572926e-01 -5.59936702e-01 -7.05230534e-01 -4.17637685e-03
-1.12583244e+00 -6.91382959e-02 7.31331408e-01 1.15492508e-01
-1.45560801e+00 7.90632725e-01 -1.11424839e+00 -1.93730563e-01
-8.61784965e-02 -3.01615801e-02 6.34703115e-02 4.83893603e-01
-1.58438182e+00 9.68482852e-01 3.06302637e-01 4.67647403e-01
-1.04934216e+00 -1.41093206e+00 -7.96658933e-01 3.97527874e-01
-1.54315948e-01 -8.02337348e-01 1.07341456e+00 -6.71600044e-01
-1.26764274e+00 4.62907016e-01 -2.78544068e-01 -5.09787619e-01
3.15321892e-01 -9.40674245e-02 -6.17668092e-01 -2.27712288e-01
-6.16470166e-02 5.66742480e-01 3.85895610e-01 -1.09077048e+00
-4.38126832e-01 6.25589490e-02 9.50113088e-02 2.17303991e-01
-2.43215412e-01 -8.74605682e-03 -9.56531391e-02 -5.58376312e-01
-1.90465704e-01 -9.25526619e-01 -6.05319560e-01 2.21316501e-01
-5.81218041e-02 1.78863332e-01 7.95887947e-01 -8.54199946e-01
1.20386136e+00 -1.78904057e+00 1.95697576e-01 4.39746827e-02
9.69017148e-02 3.06871772e-01 -1.71615779e-01 6.21005297e-01
-8.64482820e-02 5.10708570e-01 -7.85255015e-01 -4.41693366e-01
6.96864128e-02 5.12167931e-01 -8.53685915e-01 4.18304831e-01
4.05910015e-01 8.82434487e-01 -6.08757734e-01 -1.86824709e-01
3.92731130e-01 5.15186071e-01 -4.27132964e-01 5.98335087e-01
-8.49859238e-01 8.23512614e-01 -3.57499659e-01 3.12289268e-01
1.01058984e+00 -4.57189351e-01 -2.97693629e-02 2.55270123e-01
-6.00588858e-01 5.65569699e-01 -1.03103077e+00 1.38326597e+00
-9.21202123e-01 6.46826327e-01 3.53151441e-01 -9.00742233e-01
7.33790100e-01 3.71657968e-01 3.29481065e-01 -8.11263740e-01
-4.72420812e-01 2.26747990e-01 -2.68756628e-01 -3.78935188e-01
5.63926280e-01 -3.24975461e-01 -7.00186659e-03 5.60780644e-01
-3.04540485e-01 -5.92343569e-01 -1.27953932e-01 -5.86815923e-02
5.57146072e-01 5.46174765e-01 -2.18536146e-02 -6.67405784e-01
3.15592527e-01 5.77886581e-01 5.81062078e-01 6.16097689e-01
1.76755607e-01 6.28935814e-01 1.54704079e-01 -1.11959982e+00
-1.23666441e+00 -1.01274252e+00 -5.70590615e-01 1.30968392e+00
-2.42630497e-01 -3.25232983e-01 -2.58647680e-01 -1.20063230e-01
3.06852698e-01 5.67238808e-01 -7.50492454e-01 2.65644670e-01
-6.18276715e-01 -1.33661044e+00 5.95809698e-01 3.93647492e-01
3.87153566e-01 -8.74207973e-01 -6.34986579e-01 2.75464296e-01
-1.31682128e-01 -9.40003395e-01 1.68956056e-01 2.84312099e-01
-8.63605320e-01 -6.06378913e-01 -4.87862051e-01 3.80751267e-02
2.14682847e-01 -1.58083439e-01 1.59282732e+00 1.75088849e-02
5.77819496e-02 -7.50345662e-02 -2.01517716e-03 -4.18752939e-01
-2.88017690e-01 5.49288809e-01 2.28595659e-01 -4.03774977e-01
-5.56242727e-02 -9.92492199e-01 -7.37923086e-01 8.71594772e-02
-8.72928917e-01 2.84841210e-01 -6.26015961e-02 6.96329832e-01
5.55349946e-01 -4.55291532e-02 4.38449472e-01 -5.06234705e-01
2.29554206e-01 -9.87929642e-01 -1.04784131e+00 9.02030319e-02
-7.33525217e-01 1.20766155e-01 8.86515498e-01 -9.39029008e-02
-1.24471009e+00 -1.77298233e-01 -1.77430809e-01 -3.18030477e-01
-3.92796904e-01 1.09096611e+00 3.44421238e-01 1.27951518e-01
9.05493796e-01 3.34018350e-01 -2.81254858e-01 -6.73631072e-01
5.20606875e-01 3.38745177e-01 5.31421781e-01 -9.28157032e-01
1.06545424e+00 5.17668903e-01 1.25518337e-01 -6.30176663e-01
-9.67502058e-01 -1.42208219e-01 -4.62323815e-01 -1.71999827e-01
8.17339361e-01 -1.32358968e+00 -2.88645774e-01 5.63951850e-01
-8.96153927e-01 -1.20991838e+00 -1.59319147e-01 4.51197535e-01
-1.41387478e-01 -1.73501909e-01 -4.86858726e-01 -8.38184834e-01
-5.25307894e-01 -7.18256772e-01 1.05444801e+00 1.19077042e-01
-1.19135335e-01 -1.53007424e+00 6.35844827e-01 -2.89300561e-01
1.16403782e+00 7.64086962e-01 8.91963720e-01 6.70000240e-02
-5.85645437e-01 3.76189977e-01 -3.02516341e-01 1.38756484e-01
2.44364351e-01 2.73546129e-01 -1.19450057e+00 -3.50076795e-01
-2.53503144e-01 -4.90052193e-01 1.17507303e+00 4.82041627e-01
1.13943899e+00 -5.94040394e-01 -1.37567282e-01 1.33486128e+00
1.55904400e+00 -3.13543350e-01 6.05877526e-02 3.09320003e-01
7.10041642e-01 6.17063999e-01 1.65767461e-01 3.68103951e-01
6.10514700e-01 3.30321342e-01 6.45762265e-01 -7.79184043e-01
2.30304748e-01 7.20235109e-02 9.08194259e-02 5.88431001e-01
-1.47819668e-01 -1.88918769e-01 -1.68044949e+00 7.61033297e-01
-1.88345790e+00 -9.56371665e-01 -2.50218570e-01 1.95769179e+00
1.12627339e+00 -5.27015887e-02 -1.51783615e-01 -4.97893244e-01
1.41173258e-01 7.14615047e-01 -8.07557285e-01 -6.08683467e-01
-3.90648752e-01 5.13182759e-01 7.88244128e-01 7.67908573e-01
-1.20743537e+00 1.05959189e+00 6.43401909e+00 1.72322839e-01
-1.83976138e+00 3.31722230e-01 9.85143185e-01 -4.08321232e-01
-4.02747720e-01 3.70949805e-02 -7.87465692e-01 3.10063541e-01
1.76630914e+00 -1.11684509e-01 5.55472076e-01 6.16418183e-01
8.22814763e-01 1.39490932e-01 -8.82051349e-01 2.07005858e-01
-8.37302864e-01 -1.95056224e+00 -1.74160942e-01 2.67888512e-02
1.08209956e+00 9.15017724e-01 1.96444035e-01 4.56998587e-01
7.92830944e-01 -1.40799439e+00 5.98893762e-01 7.47088909e-01
9.81701434e-01 -3.52849573e-01 3.75200719e-01 4.17041004e-01
-1.51553011e+00 1.73944116e-01 -3.99867713e-01 -6.25741541e-01
2.35111207e-01 8.82247746e-01 -1.59281373e-01 6.31047249e-01
1.18140006e+00 7.25363195e-01 -1.98281124e-01 5.65622628e-01
9.68256295e-02 1.08248782e+00 -8.93499255e-01 4.72167075e-01
6.78234100e-01 -3.22205305e-01 2.77130783e-01 1.31975687e+00
4.85591829e-01 2.51675457e-01 2.49174356e-01 1.18916082e+00
1.11925885e-01 -8.60898569e-02 -6.56807482e-01 2.16986224e-01
2.77727157e-01 1.20846379e+00 -8.74541998e-02 -5.22948980e-01
-4.78749096e-01 3.88883621e-01 4.55959797e-01 7.88098693e-01
-1.03553462e+00 1.71802193e-01 1.42659688e+00 1.80078074e-01
3.64944577e-01 -6.28600419e-01 -4.66895163e-01 -1.21528566e+00
-1.27651125e-01 -8.07234764e-01 2.61382669e-01 -9.26114857e-01
-1.06733060e+00 7.13840067e-01 1.92404091e-01 -9.58146632e-01
-3.66057038e-01 -6.44407570e-01 -1.08137619e+00 1.75012958e+00
-2.04950786e+00 -1.22391856e+00 -4.23799634e-01 2.83939987e-01
2.67792679e-02 2.13334024e-01 1.15392256e+00 3.85081805e-02
-6.38085783e-01 -1.69548824e-01 6.19137645e-01 -1.30661130e-01
5.81817329e-01 -1.45290542e+00 9.95653391e-01 1.10629165e+00
-3.57117116e-01 4.51641768e-01 8.08167040e-01 -4.44184422e-01
-1.18466508e+00 -1.86365008e+00 1.03689909e+00 -3.83900642e-01
8.47119689e-01 -4.52997029e-01 -1.33215141e+00 6.80378675e-01
2.45764002e-01 5.71073771e-01 2.76609093e-01 2.79976130e-01
-4.34842765e-01 -3.21012765e-01 -7.42702782e-01 3.31637472e-01
6.88844979e-01 -7.44542062e-01 -3.25959027e-01 7.06898630e-01
8.15364182e-01 -6.66511118e-01 -1.09285045e+00 5.91131985e-01
3.22597206e-01 -8.50708485e-01 9.43798125e-01 -9.23120856e-01
8.51025164e-01 -5.03098726e-01 -2.04277366e-01 -1.66404700e+00
-4.72858638e-01 -4.40035492e-01 -2.78828233e-01 9.29650605e-01
7.98840642e-01 -7.22510040e-01 4.00073558e-01 6.42524660e-01
-9.99092758e-02 -6.75370038e-01 -8.48118901e-01 -6.26130342e-01
1.00969243e+00 -3.97220820e-01 1.13481617e+00 1.39857721e+00
-5.12259007e-01 -6.46414757e-02 -4.53661859e-01 9.45109367e-01
3.40049356e-01 8.36447895e-01 5.83084226e-01 -1.16968060e+00
-3.08112741e-01 -5.31211257e-01 4.24534619e-01 -7.66007662e-01
1.63309783e-01 -5.97576380e-01 7.51862526e-02 -1.44845867e+00
-3.67023945e-01 -7.79249489e-01 -4.41466808e-01 7.36283362e-01
-2.55911171e-01 6.68760091e-02 -2.00404584e-01 3.30707163e-01
2.40468085e-01 7.72633612e-01 8.64994645e-01 -5.65733947e-02
-2.49873951e-01 -2.44640708e-01 -6.68038055e-02 4.53025341e-01
1.20408785e+00 -4.60417479e-01 -2.38663092e-01 -9.26491201e-01
5.27142584e-01 7.90668726e-02 5.78527272e-01 -1.02693582e+00
7.04267770e-02 -8.73224556e-01 4.28033382e-01 -6.20007694e-01
2.87481666e-01 -5.20552158e-01 4.41765815e-01 2.71036506e-01
-1.99466079e-01 4.47365075e-01 8.24026108e-01 3.16814110e-02
-2.92210281e-01 6.06471598e-01 7.17663527e-01 -2.83446610e-01
-6.76589191e-01 7.20319331e-01 -5.20197511e-01 4.11293395e-02
5.75848699e-01 5.27491868e-01 -4.69730020e-01 -3.67447942e-01
-6.67609930e-01 5.53151131e-01 4.82612103e-01 2.03892246e-01
-2.19170190e-02 -1.02617502e+00 -1.06886184e+00 1.32772073e-01
1.87635713e-03 3.97886038e-01 3.59912604e-01 4.60132450e-01
-7.82384813e-01 5.71257889e-01 -5.12062497e-02 -5.97086310e-01
-5.05497634e-01 4.90940869e-01 1.24733341e+00 -4.30870652e-01
-6.48330569e-01 7.30461240e-01 4.32460070e-01 -1.27562404e+00
-3.58538598e-01 -8.87222111e-01 2.15669006e-01 -1.52805373e-01
4.03763205e-01 -8.48733038e-02 1.17747419e-01 -5.81032097e-01
-3.14231753e-01 4.16777909e-01 7.48136282e-01 -1.65418386e-01
1.56044567e+00 -2.46113032e-01 -1.43249273e-01 6.51916802e-01
9.09324765e-01 -2.91967660e-01 -1.71708798e+00 -6.34527862e-01
-3.55607361e-01 -6.44207420e-03 5.15417218e-01 -1.03249919e+00
-1.37803543e+00 1.25170088e+00 4.95275766e-01 2.37463057e-01
1.16405237e+00 -3.16942245e-01 7.06461549e-01 4.44068074e-01
-2.33386591e-01 -8.11446309e-01 -7.08781064e-01 8.92458856e-01
9.86644626e-01 -1.39431345e+00 -4.54359539e-02 1.50399730e-01
-1.71676710e-01 1.00983870e+00 6.89797163e-01 9.63580981e-02
1.15673530e+00 6.34740233e-01 4.23306763e-01 -2.24797934e-01
-1.10551965e+00 -2.27076873e-01 2.79320687e-01 1.91010430e-01
4.56164867e-01 4.04001564e-01 3.76435369e-01 2.05944195e-01
-3.95395011e-01 -1.06370002e-01 1.61355942e-01 7.31128931e-01
-2.74888963e-01 -6.53716862e-01 -5.72974622e-01 3.00559670e-01
-2.71781594e-01 -7.01644599e-01 8.12233388e-02 5.54486811e-01
1.51211351e-01 6.25285745e-01 4.61683929e-01 -4.11570631e-02
-1.13088973e-01 1.50029615e-01 -3.82520020e-01 -4.22729105e-01
-6.03986681e-01 -3.93983632e-01 8.59351382e-02 -6.62814379e-01
-4.11230266e-01 -6.04456007e-01 -1.06037498e+00 -9.60637867e-01
4.45827365e-01 3.08327287e-01 6.33214355e-01 9.44925487e-01
7.20562160e-01 5.95968246e-01 5.52921951e-01 -1.28870380e+00
-3.75272095e-01 -1.07578456e+00 -3.75593722e-01 1.53653637e-01
8.98871839e-01 -4.61782813e-01 -4.22726870e-01 7.68053979e-02] | [6.564457416534424, 2.9846603870391846] |
e4f2e8e3-6f7f-4246-b56d-640e32e43548 | doodle-to-search-practical-zero-shot-sketch | 1904.03451 | null | https://arxiv.org/abs/1904.03451v2 | https://arxiv.org/pdf/1904.03451v2.pdf | Doodle to Search: Practical Zero-Shot Sketch-based Image Retrieval | In this paper, we investigate the problem of zero-shot sketch-based image retrieval (ZS-SBIR), where human sketches are used as queries to conduct retrieval of photos from unseen categories. We importantly advance prior arts by proposing a novel ZS-SBIR scenario that represents a firm step forward in its practical application. The new setting uniquely recognizes two important yet often neglected challenges of practical ZS-SBIR, (i) the large domain gap between amateur sketch and photo, and (ii) the necessity for moving towards large-scale retrieval. We first contribute to the community a novel ZS-SBIR dataset, QuickDraw-Extended, that consists of 330,000 sketches and 204,000 photos spanning across 110 categories. Highly abstract amateur human sketches are purposefully sourced to maximize the domain gap, instead of ones included in existing datasets that can often be semi-photorealistic. We then formulate a ZS-SBIR framework to jointly model sketches and photos into a common embedding space. A novel strategy to mine the mutual information among domains is specifically engineered to alleviate the domain gap. External semantic knowledge is further embedded to aid semantic transfer. We show that, rather surprisingly, retrieval performance significantly outperforms that of state-of-the-art on existing datasets that can already be achieved using a reduced version of our model. We further demonstrate the superior performance of our full model by comparing with a number of alternatives on the newly proposed dataset. The new dataset, plus all training and testing code of our model, will be publicly released to facilitate future research | ['Yi-Zhe Song', 'Josep Llados', 'Sounak Dey', 'Pau Riba', 'Anjan Dutta'] | 2019-04-06 | doodle-to-search-practical-zero-shot-sketch-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Dey_Doodle_to_Search_Practical_Zero-Shot_Sketch-Based_Image_Retrieval_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Dey_Doodle_to_Search_Practical_Zero-Shot_Sketch-Based_Image_Retrieval_CVPR_2019_paper.pdf | cvpr-2019-6 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 3.62211794e-01 -1.60288453e-01 -2.41136983e-01 -1.14902295e-01
-1.05026948e+00 -9.42311347e-01 9.95391250e-01 -4.59979147e-01
-2.07964897e-01 4.68186468e-01 2.46558219e-01 1.15124032e-01
-4.34952438e-01 -7.11773396e-01 -6.06330454e-01 -4.05693382e-01
3.41376781e-01 6.23174131e-01 8.35233182e-02 -4.36830848e-01
5.19992113e-01 5.94984055e-01 -1.83389676e+00 2.11714104e-01
3.84135455e-01 1.06849849e+00 4.01646286e-01 4.37049776e-01
-1.78025648e-01 4.81745601e-01 -3.76351506e-01 -6.65458143e-01
6.84065640e-01 -1.98654264e-01 -7.93409348e-01 9.97313634e-02
1.02431452e+00 -5.57110786e-01 -7.37410486e-01 8.10842454e-01
6.39766812e-01 2.86403060e-01 8.26764047e-01 -1.43790901e+00
-1.29353774e+00 5.85715473e-02 -3.22241366e-01 -2.13774025e-01
4.37544823e-01 -1.53379068e-01 1.27766168e+00 -1.28023171e+00
1.12146413e+00 1.28559148e+00 3.57005000e-01 7.52666354e-01
-1.15597093e+00 -7.51850069e-01 -6.28373101e-02 2.37512007e-01
-1.72149754e+00 -4.74707276e-01 1.00070763e+00 -3.56486052e-01
4.70847279e-01 3.60625714e-01 4.67033058e-01 1.24333656e+00
-7.22422242e-01 1.05720186e+00 8.19109440e-01 -5.49615622e-01
2.15627328e-01 4.15367365e-01 -1.68703586e-01 4.91747648e-01
4.44279425e-02 -4.61791568e-02 -7.38855660e-01 -2.81244367e-01
1.17867184e+00 2.20360994e-01 -1.19888827e-01 -8.91510844e-01
-1.35475206e+00 7.72871792e-01 5.28847754e-01 3.80244613e-01
-4.16274145e-02 1.28056869e-01 8.21682960e-02 5.80908358e-01
3.48058343e-01 5.95723629e-01 -6.17306158e-02 5.27801625e-02
-1.09052849e+00 4.02696401e-01 8.44454825e-01 1.38735354e+00
7.83886909e-01 -3.59561533e-01 -1.30336046e-01 1.31121349e+00
1.24623496e-02 6.53725684e-01 2.47152671e-01 -1.14071965e+00
1.85913965e-01 5.28577626e-01 3.24346423e-01 -1.14439261e+00
6.35747612e-01 9.38089639e-02 -5.68772972e-01 -9.03195962e-02
2.01137051e-01 6.05707109e-01 -8.19010973e-01 1.62536430e+00
-8.97351000e-03 9.69857052e-02 -3.98645625e-02 1.04967237e+00
9.25852418e-01 4.18876320e-01 -1.12153165e-01 3.39228839e-01
1.31206179e+00 -1.03688562e+00 -4.70862538e-01 -1.28358856e-01
-1.96427833e-02 -9.16561723e-01 1.30191600e+00 2.10146666e-01
-1.09823120e+00 -4.42467123e-01 -1.04505324e+00 -4.80546951e-01
-8.40950370e-01 4.10998225e-01 5.93274415e-01 4.52176362e-01
-1.34493577e+00 6.27972364e-01 -3.35080251e-02 -9.02932882e-01
5.62698305e-01 1.77676365e-01 -6.37956679e-01 -5.86421728e-01
-1.09233308e+00 7.71064639e-01 -3.87697332e-02 -2.71413296e-01
-8.04931641e-01 -7.79830694e-01 -7.14097261e-01 -6.80011734e-02
5.06162167e-01 -7.87580788e-01 9.05895472e-01 -9.46030796e-01
-1.27148151e+00 1.40550530e+00 -4.82106470e-02 1.22531772e-01
6.39041960e-01 -1.19983628e-01 -2.40985468e-01 5.33885419e-01
2.23078087e-01 1.08310902e+00 9.93570149e-01 -1.66973627e+00
-1.43172204e-01 -4.02872115e-01 3.17040354e-01 2.47558743e-01
-7.60833561e-01 -4.44331132e-02 -1.02177167e+00 -9.44942951e-01
-8.46888721e-02 -1.08311200e+00 1.36454955e-01 7.64812410e-01
-6.25900328e-02 -3.45626235e-01 9.17038977e-01 -3.78780007e-01
9.22913074e-01 -2.09820724e+00 3.31357718e-01 8.26411024e-02
3.55766974e-02 2.79693365e-01 -6.84332252e-01 1.00189209e+00
-2.59714443e-02 7.59962127e-02 -2.11052984e-01 -4.64623481e-01
2.57492721e-01 2.72712201e-01 -8.18664551e-01 2.06044540e-01
2.00029075e-01 1.15101171e+00 -1.06449330e+00 -4.88526911e-01
3.27499300e-01 4.80131030e-01 -3.02446067e-01 2.42585391e-01
-4.68312837e-02 1.23437010e-02 -4.04800922e-01 9.81576681e-01
7.36026883e-01 -4.66054738e-01 4.78849672e-02 -2.76273310e-01
1.99932769e-01 -1.39630139e-01 -1.05620027e+00 2.26112843e+00
-2.89271474e-01 6.45923495e-01 -2.51463000e-02 -9.94978428e-01
1.03365493e+00 1.40449047e-01 4.88262683e-01 -7.00376570e-01
-3.41939002e-01 2.24954814e-01 -7.72926211e-01 -2.63775468e-01
8.51043999e-01 -2.73148060e-01 -1.51366115e-01 5.47862589e-01
2.26802617e-01 -4.31223720e-01 1.49446636e-01 5.81461489e-01
9.53315854e-01 1.33636624e-01 3.69412303e-02 -2.41005912e-01
3.72905433e-01 1.01903081e-02 -6.86372295e-02 9.94704187e-01
-1.19919106e-01 1.00095272e+00 1.46379977e-01 -4.71972257e-01
-1.31061375e+00 -1.39136362e+00 -9.53509286e-02 1.08137441e+00
6.02686882e-01 -3.81957471e-01 -2.54966915e-01 -5.67094564e-01
3.74117762e-01 3.28147948e-01 -6.74759269e-01 -5.50644426e-03
-2.68684626e-01 -5.67739718e-02 5.71530879e-01 3.98607254e-01
3.50417852e-01 -1.09298360e+00 -1.92332834e-01 -2.55602151e-01
-1.88412189e-01 -1.09528899e+00 -6.44322634e-01 -7.09762931e-01
-5.49876332e-01 -1.02340996e+00 -1.39136004e+00 -8.15899193e-01
6.26417696e-01 8.81098032e-01 1.25925672e+00 2.89959252e-01
-6.88462913e-01 1.21079934e+00 -4.54184771e-01 -2.48558283e-01
4.24004681e-02 -9.69981328e-02 -8.93199369e-02 -6.53288653e-03
3.86535376e-01 -6.31736934e-01 -9.67691064e-01 5.48500657e-01
-1.34465218e+00 -4.00646217e-02 6.81323230e-01 8.52242768e-01
2.56417811e-01 -5.48023582e-01 6.33832157e-01 -5.75124979e-01
5.10766566e-01 -4.00856256e-01 -3.67424369e-01 6.68065071e-01
-4.41689730e-01 -4.21143658e-02 1.93995059e-01 -6.55874431e-01
-8.33149493e-01 -8.97070169e-02 3.40807766e-01 -9.07623649e-01
9.87954065e-02 9.08055305e-02 4.40169908e-02 -3.75850976e-01
4.13626194e-01 3.60686809e-01 1.08304791e-01 -5.99998951e-01
8.02566051e-01 7.92587996e-01 4.97517973e-01 -8.65151703e-01
9.82756793e-01 8.51492286e-01 -7.69910887e-02 -1.00970078e+00
-4.71207947e-01 -8.43844891e-01 -4.98827159e-01 -2.00250760e-01
3.79726917e-01 -1.09219027e+00 -5.50523639e-01 2.25175872e-01
-1.13536048e+00 -7.71397278e-02 -4.60274845e-01 -2.91518420e-02
-6.91170454e-01 5.95937550e-01 -4.86138254e-01 -9.03506756e-01
-3.65340382e-01 -8.03174675e-01 1.66896355e+00 4.38814797e-02
6.79066218e-03 -8.67023230e-01 1.24879360e-01 4.57364053e-01
4.92482245e-01 -1.55897096e-01 7.93967545e-01 -5.15170455e-01
-9.41214979e-01 -2.41567433e-01 -8.26707721e-01 2.75893927e-01
1.22385379e-02 -2.10579723e-01 -1.00162220e+00 -4.44925338e-01
-6.45590246e-01 -7.65997291e-01 9.78119373e-01 -1.67733938e-01
1.09279537e+00 -4.59994636e-02 -3.82470727e-01 2.90357351e-01
1.70030642e+00 -1.23805076e-01 9.48453546e-01 2.16886908e-01
4.94181484e-01 7.49583840e-01 6.53893411e-01 4.03805614e-01
4.71401423e-01 8.62968385e-01 1.44848377e-01 -5.29024117e-02
-4.26291198e-01 -7.15594113e-01 6.02704808e-02 6.77165031e-01
-1.42847046e-01 -2.12647036e-01 -5.87784946e-01 9.82624888e-01
-1.78699374e+00 -1.06401730e+00 5.32213748e-01 2.29895520e+00
6.22058630e-01 -6.76151454e-01 6.90787509e-02 -1.57464683e-01
6.62797749e-01 3.01514179e-01 -4.42715317e-01 1.70385793e-01
-1.27445057e-01 4.76091743e-01 1.79458499e-01 1.68573111e-01
-1.01405501e+00 1.15675628e+00 6.53374147e+00 1.13159502e+00
-8.78391922e-01 -7.05833286e-02 1.47831112e-01 -9.43678319e-02
-5.98190665e-01 1.20961078e-01 -4.65529889e-01 2.37735033e-01
2.72036284e-01 -3.13599557e-01 6.99478149e-01 9.30429578e-01
-4.73441690e-01 2.18746737e-01 -1.33211219e+00 1.45555365e+00
5.49900830e-01 -1.52447534e+00 5.14347613e-01 1.49391890e-01
8.13714623e-01 -2.37267107e-01 2.85607994e-01 3.00829619e-01
2.44676843e-01 -8.67718995e-01 6.07522070e-01 6.12159014e-01
1.31663072e+00 -3.27577740e-01 1.21766396e-01 1.89177804e-02
-1.18538785e+00 -9.52441320e-02 -7.26673126e-01 1.90887451e-01
-7.18587637e-02 2.00540125e-01 -4.58858997e-01 7.23767400e-01
6.12595737e-01 9.44167614e-01 -6.32295072e-01 8.46710980e-01
1.34759843e-01 -2.02196389e-01 -2.02912912e-01 4.92238440e-02
7.01112896e-02 -2.90647477e-01 4.93802190e-01 9.45379198e-01
5.40480793e-01 1.70310825e-01 -1.19938157e-01 1.12016129e+00
-3.85851353e-01 4.63303998e-02 -1.20469630e+00 -3.53833228e-01
6.51015878e-01 1.29901576e+00 -4.95942116e-01 -5.07492721e-01
-3.39503169e-01 1.40349472e+00 4.33988214e-01 5.77935994e-01
-3.87849122e-01 -4.38230723e-01 7.71596789e-01 7.40354210e-02
4.96849477e-01 1.29667865e-02 9.04274955e-02 -1.48257935e+00
1.87802717e-01 -7.17703044e-01 3.96944731e-01 -1.20204794e+00
-1.96013045e+00 3.67350549e-01 2.45219935e-02 -1.38854373e+00
-9.87348482e-02 -6.03943110e-01 -2.82434970e-01 7.66760468e-01
-1.72147465e+00 -1.63066185e+00 -4.23091531e-01 7.50052571e-01
7.83017993e-01 -2.31462464e-01 9.57260430e-01 5.33461034e-01
-4.43686615e-04 6.67100966e-01 1.90088823e-01 1.21325888e-01
1.15311086e+00 -9.59950686e-01 4.00970072e-01 3.20842117e-01
4.87349004e-01 8.27358246e-01 3.95905048e-01 -5.22054374e-01
-1.80533099e+00 -8.21709096e-01 8.13524902e-01 -8.09450924e-01
7.48516202e-01 -6.19844615e-01 -6.78382754e-01 5.92130840e-01
4.26513329e-02 1.41374901e-01 4.61201578e-01 -7.92559981e-02
-8.00128818e-01 -1.27175182e-01 -1.17615092e+00 6.80660725e-01
1.51034510e+00 -1.02375340e+00 -7.59637833e-01 3.21716249e-01
5.96342862e-01 7.02052489e-02 -8.11731756e-01 1.87012076e-01
1.13310730e+00 -7.97029555e-01 1.60088539e+00 -6.21310890e-01
7.74603486e-01 3.50721274e-03 -4.97030020e-01 -9.56954122e-01
-1.34525076e-01 -4.27466631e-01 1.65468395e-01 1.21215022e+00
-1.91918954e-01 -3.66599977e-01 9.17944312e-01 7.95525610e-01
3.94344985e-01 -3.60414267e-01 -7.13611603e-01 -1.01977491e+00
1.47619277e-01 -1.43793032e-01 4.77035224e-01 1.10175014e+00
-1.73548982e-01 1.63244784e-01 -5.76585591e-01 -2.67656565e-01
8.40323985e-01 3.76869977e-01 1.02993321e+00 -1.39292562e+00
-1.31725565e-01 -4.40461546e-01 -5.16318440e-01 -1.31681967e+00
1.35021716e-01 -8.28822076e-01 -1.71834737e-01 -1.38250291e+00
5.15460908e-01 -6.12127244e-01 -3.10248345e-01 3.27706128e-01
6.50665835e-02 7.94038594e-01 6.08176053e-01 5.69764495e-01
-7.03414559e-01 6.09002709e-01 1.31137311e+00 -2.54293054e-01
2.76082397e-01 -4.59774852e-01 -7.45692074e-01 2.90232509e-01
1.59001067e-01 -1.92261294e-01 -6.04262888e-01 -3.68847728e-01
1.08777434e-01 1.33198023e-01 8.20031047e-01 -6.10289395e-01
3.04119617e-01 -8.87893513e-02 2.22121552e-01 -5.79836905e-01
8.25580180e-01 -9.49856222e-01 2.01911286e-01 -1.55926198e-01
-5.56941628e-01 -1.26951933e-01 -4.70262803e-02 8.83585870e-01
-2.39637986e-01 -1.32173195e-01 3.55827421e-01 -3.67157608e-01
-9.33192551e-01 4.45636332e-01 2.14119807e-01 -1.18389487e-01
8.97854269e-01 -2.14189917e-01 -4.86065120e-01 -6.94759130e-01
-4.42477077e-01 -2.32402775e-02 9.14284110e-01 8.08119774e-01
8.85208130e-01 -1.71255660e+00 -4.77957249e-01 1.80216894e-01
7.63872147e-01 -5.31271994e-01 3.54288340e-01 3.12589943e-01
-2.71373242e-01 5.07175624e-01 -3.27601522e-01 -2.97489733e-01
-1.22459507e+00 7.33387172e-01 -2.64681518e-01 -6.99119046e-02
-5.37791848e-01 7.12787390e-01 4.45597380e-01 -5.95622063e-01
2.02257946e-01 4.03364301e-01 2.60934144e-01 9.25070494e-02
5.41959882e-01 2.64373392e-01 -1.16696686e-01 -5.44055998e-01
-3.52515340e-01 8.70521605e-01 -5.49229044e-05 -4.02570635e-01
1.44081497e+00 -2.13375971e-01 -1.89652935e-01 3.41479421e-01
1.45480180e+00 -1.97011083e-01 -1.00962174e+00 -5.26957929e-01
-1.10166728e-01 -9.95759845e-01 -1.91349238e-01 -8.74127984e-01
-8.55913460e-01 9.65336025e-01 5.49546123e-01 2.81512359e-04
9.73554909e-01 3.67397189e-01 8.39087248e-01 5.74181139e-01
7.19975948e-01 -9.77322757e-01 5.66840291e-01 -8.32834747e-03
1.25163150e+00 -1.37852442e+00 2.57158335e-02 -2.99380660e-01
-6.15223765e-01 9.88395452e-01 2.36382142e-01 -4.02126044e-01
4.20906812e-01 -3.03967327e-01 -1.39833778e-01 -3.08688223e-01
-6.78087533e-01 -3.29208195e-01 5.30072331e-01 6.81584060e-01
5.23573495e-02 -1.24324255e-01 -9.72210914e-02 3.95924389e-01
2.24541187e-01 2.37302393e-01 1.74115106e-01 9.74522948e-01
-1.21964641e-01 -1.38374507e+00 -1.89523205e-01 2.36648723e-01
1.07352808e-01 -1.59850374e-01 -8.55549693e-01 9.55619872e-01
-2.62312621e-01 6.55496716e-01 8.41846764e-02 -2.63817161e-01
3.14877152e-01 8.39297697e-02 7.76505828e-01 -3.06727678e-01
-1.34759918e-01 -3.21853012e-01 -2.05798298e-01 -5.60199499e-01
-5.22105038e-01 -2.68844783e-01 -3.66157502e-01 -3.40073019e-01
-1.27897426e-01 -1.35243580e-01 7.84659743e-01 4.80009645e-01
6.56086028e-01 -2.55036265e-01 5.32349944e-01 -1.15614295e+00
-5.84007561e-01 -6.58403933e-01 -7.85612881e-01 9.80863392e-01
1.18685514e-01 -9.03304219e-01 -4.07647163e-01 7.71712139e-02] | [11.6128568649292, 0.6520737409591675] |
cdc009f3-1717-4f15-92bd-acee905da4e9 | iqiyi-vid-a-large-dataset-for-multi-modal | 1811.07548 | null | http://arxiv.org/abs/1811.07548v2 | http://arxiv.org/pdf/1811.07548v2.pdf | iQIYI-VID: A Large Dataset for Multi-modal Person Identification | Person identification in the wild is very challenging due to great variation
in poses, face quality, clothes, makeup and so on. Traditional research, such
as face recognition, person re-identification, and speaker recognition, often
focuses on a single modal of information, which is inadequate to handle all the
situations in practice. Multi-modal person identification is a more promising
way that we can jointly utilize face, head, body, audio features, and so on. In
this paper, we introduce iQIYI-VID, the largest video dataset for multi-modal
person identification. It is composed of 600K video clips of 5,000 celebrities.
These video clips are extracted from 400K hours of online videos of various
types, ranging from movies, variety shows, TV series, to news broadcasting. All
video clips pass through a careful human annotation process, and the error rate
of labels is lower than 0.2\%. We evaluated the state-of-art models of face
recognition, person re-identification, and speaker recognition on the iQIYI-VID
dataset. Experimental results show that these models are still far from being
perfect for the task of person identification in the wild. We proposed a
Multi-modal Attention module to fuse multi-modal features that can improve
person identification considerably. We have released the dataset online to
promote multi-modal person identification research. | ['Tingwei Gao', 'Ganwen Wang', 'Danming Xie', 'Bo Peng', 'Yuanliu Liu', 'Yong Zhou', 'Peipei Shi', 'Jianbin Jiang', 'Chao Lin', 'Yi Zheng', 'Yin Fan', 'Xiangju Lu', 'Jian Liu', 'He Yan', 'Bing Han'] | 2018-11-19 | null | null | null | null | ['person-identification', 'multi-modal-person-identification'] | ['computer-vision', 'miscellaneous'] | [-1.71344534e-01 -6.47019744e-01 -1.44485489e-01 -5.51711142e-01
-7.61009574e-01 -7.20288038e-01 5.24154305e-01 -4.42921489e-01
-3.44335377e-01 5.57933629e-01 3.69651079e-01 3.24423611e-01
1.31057516e-01 -2.07020104e-01 -3.86983901e-01 -7.07279086e-01
2.09835351e-01 3.69762361e-01 -2.18263566e-01 5.47628216e-02
9.46452841e-02 3.79919112e-01 -1.79261625e+00 2.42251992e-01
3.73587489e-01 1.00357282e+00 -3.47227842e-01 5.29653668e-01
2.69043613e-02 2.18408227e-01 -6.80843651e-01 -9.34480429e-01
-3.67863104e-02 -1.96782783e-01 -7.36041427e-01 9.06839371e-02
6.87429905e-01 -6.38043821e-01 -5.41904271e-01 1.09755540e+00
1.03656030e+00 4.83722389e-02 6.05194986e-01 -1.57451606e+00
-7.96340585e-01 5.59887171e-01 -6.55745089e-01 4.21272099e-01
8.45274448e-01 -3.66601311e-02 4.32242244e-01 -9.46282446e-01
1.14645667e-01 1.58129573e+00 8.60489070e-01 9.20184433e-01
-8.10928106e-01 -1.16432357e+00 4.09257337e-02 4.85268205e-01
-1.96311748e+00 -1.02962244e+00 6.46272242e-01 -3.84137452e-01
3.79838228e-01 3.14487875e-01 4.44945067e-01 1.42699933e+00
-2.37168401e-01 7.50401020e-01 8.67933750e-01 -3.00744265e-01
-3.31962585e-01 4.38624889e-01 3.82735878e-01 4.54036415e-01
-1.27281785e-01 -1.00544868e-02 -8.68565083e-01 -2.42695242e-01
4.71630603e-01 2.49044612e-01 -4.10308063e-01 3.73875648e-01
-1.24349666e+00 5.52925527e-01 -1.10137962e-01 2.94418365e-01
-1.49028478e-02 -2.47655571e-01 6.04949534e-01 2.31464699e-01
2.77400911e-01 -2.91474760e-01 -8.45452175e-02 -1.79405347e-01
-7.53429949e-01 1.60815850e-01 6.34009182e-01 1.02911150e+00
5.00778735e-01 -3.41059007e-02 -2.00877860e-01 1.14971411e+00
3.83485049e-01 7.84395278e-01 8.53641331e-01 -7.17615187e-01
2.40979075e-01 3.75790358e-01 1.86019167e-01 -1.15628195e+00
-3.87757510e-01 9.69850942e-02 -9.38219070e-01 -3.97945583e-01
2.84509689e-01 -2.49711022e-01 -6.68865919e-01 1.77535367e+00
3.48905325e-01 4.55158263e-01 -1.94339398e-02 8.12563479e-01
1.43175411e+00 5.61802387e-01 7.34139234e-02 -1.63751692e-01
1.65322137e+00 -7.71429777e-01 -8.98046970e-01 -3.95144336e-02
9.19198543e-02 -7.44894683e-01 7.47695684e-01 2.13744566e-01
-7.02031434e-01 -9.22895372e-01 -7.29779899e-01 2.69996345e-01
-3.74147743e-01 2.32419685e-01 3.02553207e-01 1.25209224e+00
-9.61648941e-01 2.41870239e-01 -1.54372245e-01 -4.78158921e-01
3.04982483e-01 4.23204005e-01 -8.48403037e-01 -2.72224277e-01
-1.22046793e+00 4.96443331e-01 2.41946682e-01 2.79623419e-01
-8.79366457e-01 -4.19413745e-01 -8.90929937e-01 -1.88524768e-01
2.30693817e-01 -3.60816389e-01 9.92809594e-01 -9.81609285e-01
-1.44801164e+00 1.12521434e+00 -5.46770096e-01 3.11752465e-02
1.69897422e-01 -1.01155959e-01 -1.08970487e+00 3.55709754e-02
1.18290566e-01 2.98453063e-01 1.18965840e+00 -1.07698119e+00
-5.78855455e-01 -9.52612042e-01 -1.02125615e-01 3.98453884e-02
-7.15558290e-01 6.84201539e-01 -7.98339963e-01 -4.81505632e-01
-2.01430336e-01 -9.97920513e-01 5.96394658e-01 -3.84758174e-01
-5.14449477e-01 -3.58912975e-01 9.55171883e-01 -9.94497001e-01
1.14358115e+00 -2.43240190e+00 8.59694928e-02 5.02737425e-02
-9.18380171e-02 3.40364277e-01 -8.70800540e-02 2.50915010e-02
-4.13068756e-02 2.08784074e-01 3.02501857e-01 -6.77734733e-01
7.68280253e-02 -1.63373619e-01 -5.76405674e-02 5.48219144e-01
-2.73675561e-01 7.15220451e-01 -4.29949760e-01 -4.83297378e-01
1.54136360e-01 6.88496113e-01 -1.58883333e-01 2.16392249e-01
6.40902162e-01 5.67201197e-01 -1.61749870e-01 1.10563314e+00
7.36277163e-01 8.29739571e-02 -1.43820226e-01 -5.33926904e-01
4.20764759e-02 -4.12402689e-01 -1.47416508e+00 1.55172658e+00
-8.00924748e-02 4.88683939e-01 1.98360443e-01 -7.58761108e-01
7.63757706e-01 7.18202829e-01 4.30837959e-01 -3.41647804e-01
2.82420188e-01 1.25195980e-02 -2.64552087e-01 -4.94480729e-01
4.01433378e-01 -3.12781124e-03 -2.75774062e-01 3.23383808e-01
2.36160606e-01 6.42258644e-01 1.86830774e-01 -6.45033717e-02
4.38289762e-01 -3.15752834e-01 2.41366774e-02 -7.21788257e-02
9.36809599e-01 -5.23805976e-01 6.44244373e-01 5.62615752e-01
-6.23491287e-01 6.21133208e-01 -5.92749827e-02 -3.47335696e-01
-7.00752139e-01 -8.57953966e-01 -2.66893685e-01 1.46554458e+00
1.67564809e-01 -3.34938228e-01 -8.94449234e-01 -3.16442519e-01
5.35978153e-02 1.22570373e-01 -5.63159168e-01 -2.35361680e-01
-4.06099111e-01 -9.62578714e-01 1.01705098e+00 5.01387537e-01
8.24163377e-01 -8.98887515e-01 2.21391350e-01 2.77420860e-02
-4.59349334e-01 -1.31790960e+00 -1.20667064e+00 -5.96826613e-01
-1.62248656e-01 -9.58129883e-01 -1.22202218e+00 -1.04618764e+00
4.09927487e-01 4.74638134e-01 8.52673948e-01 -7.82662630e-02
-2.71692336e-01 8.70947599e-01 -3.39060873e-01 -3.82791311e-01
-1.43130675e-01 -1.13746271e-01 8.04642498e-01 6.92345023e-01
6.16749883e-01 -3.06898922e-01 -4.07303423e-01 7.83627987e-01
-5.60677707e-01 -5.42126477e-01 4.33810353e-02 8.61433804e-01
3.03052336e-01 3.02644700e-01 7.76661754e-01 -3.29255909e-01
4.33404148e-01 -1.95011750e-01 -8.23112354e-02 4.11718816e-01
-4.27220352e-02 -6.74146593e-01 3.87161642e-01 -7.60216951e-01
-1.08863044e+00 -2.72179581e-02 -2.49667317e-01 -3.71028185e-01
-5.55380523e-01 1.78724870e-01 -7.22690284e-01 -3.45083535e-01
1.96739674e-01 4.64834690e-01 5.77140972e-02 -5.80297768e-01
9.31176590e-04 1.29891610e+00 9.60434616e-01 -4.11346257e-01
6.46715820e-01 3.28522891e-01 -6.53042316e-01 -1.08034992e+00
-6.12908661e-01 -6.66938961e-01 -6.84230268e-01 -4.43413526e-01
8.63509834e-01 -1.23337722e+00 -1.30343556e+00 1.23431313e+00
-1.01354098e+00 2.66607940e-01 4.27934796e-01 4.76628065e-01
-8.66559744e-02 6.51715457e-01 -7.26816416e-01 -1.11140585e+00
-4.86904740e-01 -9.74794447e-01 1.13713741e+00 5.86659193e-01
6.76688850e-02 -7.48035192e-01 -1.85043395e-01 7.40681291e-01
3.00790519e-01 -5.33072725e-02 2.77282655e-01 -5.94887078e-01
1.42337373e-02 -4.31064039e-01 -1.00217499e-01 2.82146543e-01
3.71215224e-01 -1.36758849e-01 -1.25174677e+00 -5.94649613e-01
-1.92978948e-01 -4.09325331e-01 6.70005620e-01 2.35564992e-01
1.17972803e+00 -3.14629883e-01 -4.33617115e-01 6.90824807e-01
8.70800197e-01 3.70033652e-01 4.88899678e-01 1.06819160e-01
9.38797116e-01 6.03540182e-01 3.01769614e-01 5.64894974e-01
7.12279439e-01 9.66254413e-01 7.12048188e-02 1.78810641e-01
5.98429181e-02 -9.60217863e-02 6.23555481e-01 7.00661182e-01
-2.22836897e-01 -3.10210675e-01 -8.46655190e-01 5.15341103e-01
-1.53615022e+00 -1.37676203e+00 2.85135508e-01 2.30156136e+00
6.68394506e-01 -3.78835708e-01 7.04108655e-01 2.88861036e-01
1.25892258e+00 -1.31241679e-01 -4.69443828e-01 8.65827054e-02
-2.39284962e-01 -4.96282011e-01 2.29315579e-01 1.64781779e-01
-1.42608571e+00 6.53344214e-01 6.37364817e+00 7.71987557e-01
-1.03686750e+00 1.74749926e-01 6.10443830e-01 -1.66468248e-01
1.71932966e-01 -7.11961746e-01 -1.49690998e+00 1.01738644e+00
1.11206007e+00 -2.19666600e-01 5.52456260e-01 6.78338885e-01
-6.92426860e-02 2.65960008e-01 -1.14351308e+00 1.90183794e+00
8.31424296e-01 -8.47037315e-01 -7.73590505e-02 -7.40272226e-03
3.04902405e-01 -4.09638584e-01 2.10983500e-01 3.85681659e-01
-1.94732100e-01 -1.17771935e+00 7.02550530e-01 5.75458825e-01
9.73256648e-01 -9.16468620e-01 9.63376343e-01 2.57704169e-01
-1.39479733e+00 -2.87016004e-01 -1.38336346e-01 3.63464296e-01
3.42926353e-01 9.11066756e-02 -3.34452689e-01 4.71422493e-01
1.21441591e+00 8.66578639e-01 -5.65506518e-01 9.30495977e-01
5.65366626e-01 3.95640999e-01 -3.38606685e-01 2.28813723e-01
-3.69960397e-01 7.39155989e-03 3.52303386e-01 1.17446792e+00
4.63913411e-01 1.40283450e-01 2.50344604e-01 2.55361468e-01
-8.90696049e-02 1.12089925e-01 -5.13581753e-01 7.73905292e-02
5.13507366e-01 1.09187496e+00 -2.06726372e-01 -3.62412542e-01
-6.25729561e-01 1.20169234e+00 -1.41666889e-01 2.61820972e-01
-9.58661497e-01 -1.12375878e-01 7.78411508e-01 -7.99456909e-02
5.13364114e-02 4.61621843e-02 3.25133711e-01 -1.31040859e+00
1.50936440e-01 -1.08695006e+00 6.72303557e-01 -5.49956501e-01
-1.68039632e+00 7.41161406e-01 -2.33101137e-02 -1.04545271e+00
-3.08154792e-01 -5.70931375e-01 -4.63528365e-01 8.51826966e-01
-1.24016964e+00 -1.35065651e+00 -5.06022990e-01 1.10059369e+00
4.52056885e-01 -7.46285677e-01 9.96169627e-01 1.00954294e+00
-1.18701780e+00 1.32282233e+00 2.70339400e-01 5.72378218e-01
1.08155453e+00 -7.30678797e-01 1.58420786e-01 5.93083739e-01
4.61170301e-02 6.85880899e-01 3.74415010e-01 -3.60278904e-01
-1.54526818e+00 -9.36383426e-01 6.99854732e-01 -4.71468151e-01
2.44913906e-01 -2.30922848e-01 -8.10847104e-01 7.03554332e-01
7.18197785e-03 -4.02505286e-02 1.22849011e+00 1.19990319e-01
-4.52958316e-01 -4.23858225e-01 -1.22831333e+00 2.19002619e-01
1.00506079e+00 -8.17196190e-01 -3.21325451e-01 2.61417776e-01
2.92374313e-01 -1.88328639e-01 -1.05072463e+00 2.23649070e-01
9.07110810e-01 -7.82599509e-01 1.35435784e+00 -4.73674178e-01
-3.18754286e-01 -2.62904942e-01 -3.65201116e-01 -1.08937430e+00
-3.90503049e-01 -4.15201724e-01 -2.66050138e-02 1.99952984e+00
6.32157773e-02 -6.59063935e-01 5.47917485e-01 9.04111385e-01
1.01322033e-01 6.69480860e-02 -1.08097720e+00 -8.04316461e-01
-3.14940959e-01 -2.24118069e-01 9.53153312e-01 8.60707283e-01
-5.52643463e-02 3.43129724e-01 -1.14072478e+00 2.11743638e-01
8.67492020e-01 1.10372663e-01 8.63234580e-01 -1.42232752e+00
-1.16703957e-01 -2.14114040e-01 -6.63605392e-01 -8.00323427e-01
5.09586215e-01 -6.06540203e-01 -2.21361220e-01 -9.07684565e-01
6.67860389e-01 -3.86359431e-02 -2.01918140e-01 4.40655529e-01
-2.15025783e-01 7.01731384e-01 1.93679675e-01 3.65057707e-01
-5.79786539e-01 5.33256948e-01 7.97232866e-01 -5.59959948e-01
-7.09328204e-02 1.13942839e-01 -7.70565867e-01 7.25810170e-01
6.32448852e-01 -1.41264722e-02 -2.24070828e-02 -5.42441130e-01
-5.69418848e-01 8.74963254e-02 5.07179201e-01 -1.05828130e+00
3.90587121e-01 1.78284407e-01 7.17759311e-01 -5.39197564e-01
7.89864719e-01 -7.02831924e-01 4.12882507e-01 3.70105505e-02
-8.07896033e-02 2.23851074e-02 2.95323551e-01 5.49391091e-01
-3.72846603e-01 -1.48451313e-01 6.98213339e-01 -1.53508663e-01
-9.90319729e-01 7.15948284e-01 -2.07374156e-01 -8.13103840e-02
9.22223985e-01 -3.34036559e-01 -2.48058423e-01 -4.40379143e-01
-9.63344872e-01 7.10261390e-02 2.93792397e-01 6.70438588e-01
5.44334531e-01 -1.76736188e+00 -9.66352403e-01 3.51586044e-01
1.56709611e-01 -6.44110560e-01 7.27156103e-01 5.74763179e-01
1.32501572e-01 3.66683960e-01 -4.50161666e-01 -6.04951203e-01
-1.87811196e+00 6.65926754e-01 3.62856895e-01 4.20457214e-01
-4.25258040e-01 8.28816593e-01 9.24340934e-02 -3.45825136e-01
5.43439507e-01 4.55041289e-01 -5.80832243e-01 3.44722927e-01
1.37190175e+00 4.45084959e-01 -1.57586157e-01 -1.53698349e+00
-7.73711324e-01 8.88764203e-01 -1.70204952e-01 1.80143826e-02
9.07725215e-01 -5.09055614e-01 -2.30145343e-02 3.69244993e-01
1.27479196e+00 -1.59476757e-01 -7.01275051e-01 -4.73271787e-01
-4.83097523e-01 -7.19890058e-01 -1.54107496e-01 -4.22665417e-01
-1.14105773e+00 6.75739050e-01 9.88258004e-01 2.05588251e-01
1.00333750e+00 -1.54799572e-03 8.77884448e-01 1.79060116e-01
6.11184955e-01 -1.15546966e+00 -1.31372086e-04 3.00171345e-01
8.63219678e-01 -1.61429012e+00 -2.77130455e-01 -2.53686935e-01
-6.76363409e-01 8.76177430e-01 6.33065224e-01 5.26267529e-01
9.18630302e-01 -1.13453902e-02 1.68294162e-01 2.66809642e-01
-1.51821077e-01 -9.60978791e-02 3.46752703e-01 8.51388156e-01
3.34880739e-01 8.91115516e-02 3.38394552e-01 1.05055809e+00
-2.20961183e-01 -2.56578296e-01 1.72045439e-01 4.39226955e-01
-1.89454570e-01 -8.46362591e-01 -1.01236653e+00 2.94440240e-01
-6.53376520e-01 1.49199173e-01 -2.71881193e-01 2.95820475e-01
9.40359756e-02 1.40289259e+00 4.84624654e-02 -6.44722283e-01
2.85987496e-01 2.60173410e-01 3.46166104e-01 -1.88675791e-01
-3.80971283e-01 -8.04406628e-02 -4.35877666e-02 -3.30490679e-01
-7.15096116e-01 -8.78382444e-01 -6.43051684e-01 -8.61525416e-01
-9.59903821e-02 3.13671455e-02 4.82370526e-01 8.36099446e-01
2.51577079e-01 1.71651706e-01 7.29894876e-01 -1.20110488e+00
-3.75676751e-01 -9.98076737e-01 -9.32840526e-01 5.84485829e-01
3.84516001e-01 -8.89904618e-01 -3.55234325e-01 2.79414564e-01] | [14.467227935791016, 1.108917236328125] |
d27b0399-f9d1-4bbd-8ea7-2054581a07db | the-knowledge-graph-track-at-oaei-gold | 2002.10283 | null | https://arxiv.org/abs/2002.10283v1 | https://arxiv.org/pdf/2002.10283v1.pdf | The Knowledge Graph Track at OAEI -- Gold Standards, Baselines, and the Golden Hammer Bias | The Ontology Alignment Evaluation Initiative (OAEI) is an annual evaluation of ontology matching tools. In 2018, we have started the Knowledge Graph track, whose goal is to evaluate the simultaneous matching of entities and schemas of large-scale knowledge graphs. In this paper, we discuss the design of the track and two different strategies of gold standard creation. We analyze results and experiences obtained in first editions of the track, and, by revealing a hidden task, we show that all tools submitted to the track (and probably also to other tracks) suffer from a bias which we name the golden hammer bias. | ['Sven Hertling', 'Heiko Paulheim'] | 2020-02-24 | null | null | null | null | ['ontology-matching'] | ['knowledge-base'] | [-1.10707290e-01 8.03176641e-01 -2.61082381e-01 -6.45102486e-02
-1.59366071e-01 -7.12797046e-01 6.22012854e-01 5.19308507e-01
-2.33968064e-01 4.28394198e-01 3.47535700e-01 -2.36251235e-01
-6.48353100e-01 -8.38520646e-01 -4.72016573e-01 4.31758553e-01
-1.34146675e-01 1.04183948e+00 6.60495698e-01 -5.17160416e-01
1.60651192e-01 3.00836921e-01 -1.77832985e+00 3.28840673e-01
8.59008193e-01 4.95801002e-01 -3.30921948e-01 4.33940664e-02
-5.12808800e-01 9.62306023e-01 -4.43164706e-01 -1.03146100e+00
1.96428642e-01 -9.33915228e-02 -1.63065243e+00 -5.99920928e-01
8.16140950e-01 4.72480088e-01 -1.44463331e-01 1.24843752e+00
4.69744474e-01 -1.78092808e-01 2.19374761e-01 -1.99695230e+00
-4.56581861e-01 9.39902425e-01 1.53901860e-01 2.22269237e-01
9.20586467e-01 -5.56683779e-01 1.15303552e+00 -7.70666480e-01
1.58130181e+00 9.29004908e-01 1.17693877e+00 3.64551753e-01
-9.25429583e-01 -6.16223931e-01 -3.38549912e-01 7.16317415e-01
-1.74546695e+00 -8.80023062e-01 2.21057713e-01 -5.98774135e-01
1.58656025e+00 3.60356420e-01 6.22044861e-01 5.61487377e-01
-2.73230281e-02 1.24907993e-01 6.22211099e-01 -8.08308303e-01
-1.01890072e-01 3.34225178e-01 2.54613757e-01 8.59585583e-01
8.63439202e-01 -1.20252438e-01 -5.13415039e-01 -5.52698076e-01
1.03569254e-01 -8.07762742e-01 -8.58228356e-02 -7.71663964e-01
-1.01323783e+00 1.57535255e-01 3.04697286e-02 5.55646181e-01
-2.73194641e-01 -3.16253424e-01 3.23863238e-01 4.75727856e-01
-1.58825517e-01 8.44623268e-01 -3.39502633e-01 -1.38751641e-01
-4.45566416e-01 3.95057678e-01 1.35945475e+00 1.63460493e+00
7.35598981e-01 -6.25545144e-01 3.23558778e-01 6.58024788e-01
3.07826340e-01 -1.38080508e-01 2.67935246e-01 -9.84295666e-01
5.53393781e-01 1.23866010e+00 2.78489292e-01 -1.04437482e+00
-6.16657436e-01 -3.00632834e-01 5.60582206e-02 5.33302166e-02
3.66753012e-01 1.39417127e-01 -4.37743813e-01 1.49493158e+00
3.39889228e-01 -1.17550120e-01 2.75809526e-01 3.02651346e-01
1.20820987e+00 -4.14152324e-01 3.15309495e-01 -3.51180625e-03
1.36216903e+00 -9.09226060e-01 -9.92226779e-01 2.67374162e-02
9.97665823e-01 -9.40629721e-01 6.64317131e-01 1.06168509e-01
-1.29704368e+00 -3.10115069e-01 -1.24802125e+00 4.42810776e-03
-1.04840612e+00 -6.06275916e-01 6.66007817e-01 7.11172998e-01
-1.12192690e+00 6.51608944e-01 -4.13235992e-01 -1.14995837e+00
1.02255054e-01 1.64823920e-01 -8.58557880e-01 1.44226059e-01
-1.43906093e+00 1.62795460e+00 8.31534445e-01 -6.45446777e-01
-9.51427314e-03 -9.07048106e-01 -5.79960883e-01 -1.06489860e-01
6.17212355e-01 -8.32193613e-01 1.04527664e+00 -4.51410383e-01
-6.73063338e-01 1.36269617e+00 -4.17421348e-02 -3.78453195e-01
3.72196525e-01 -5.96824326e-02 -1.39006305e+00 -1.66386247e-01
4.44731772e-01 2.84888715e-01 -1.96011692e-01 -9.53872800e-01
-1.01500380e+00 -3.30051631e-01 2.97794849e-01 3.77731174e-02
-2.60298461e-01 5.85602701e-01 -6.22625947e-01 -1.57344058e-01
-8.68686810e-02 -6.36237144e-01 -1.60733832e-03 -4.73147124e-01
1.45632960e-02 -5.04474521e-01 3.93057048e-01 -7.63051033e-01
1.87437522e+00 -1.81677854e+00 2.14350358e-01 4.89419341e-01
4.29129660e-01 -1.16990302e-02 1.40253022e-01 9.59191084e-01
-4.13344145e-01 4.86509651e-01 1.19469509e-01 1.35209441e-01
3.99638355e-01 2.07491726e-01 -8.49659815e-02 -8.06923117e-03
-2.28875279e-01 9.34226632e-01 -1.06850779e+00 -9.20780838e-01
-2.10268185e-01 -1.79800555e-01 -3.25048566e-01 -2.06777036e-01
-1.05764695e-01 -2.53221899e-01 -7.13827908e-02 8.70312393e-01
3.37847412e-01 -4.08054262e-01 8.57568681e-01 -5.47858596e-01
-3.22879761e-01 4.04387861e-01 -1.28836668e+00 2.02129173e+00
7.63535351e-02 3.72785538e-01 -2.47764707e-01 -3.77060682e-01
6.19107187e-01 7.35799909e-01 8.08601499e-01 -9.25461769e-01
-3.53938460e-01 6.32685125e-01 1.69429183e-01 -5.36456645e-01
6.58342242e-01 4.32547361e-01 4.76722494e-02 3.06936294e-01
5.29344022e-01 1.83541521e-01 5.47527611e-01 3.25287372e-01
1.46040082e+00 3.67572069e-01 7.10537851e-01 -4.43402827e-01
3.74150932e-01 6.02660000e-01 5.69724321e-01 6.68284833e-01
-8.08993801e-02 -8.50125104e-02 2.11115375e-01 -4.94467556e-01
-1.08186674e+00 -9.27318215e-01 -3.83717269e-01 7.20820665e-01
2.62825161e-01 -1.46733940e+00 -5.40497124e-01 -8.22706878e-01
7.21322149e-02 7.67121553e-01 -4.36977714e-01 -1.73477009e-01
-3.88573587e-01 -2.91247219e-01 1.02915347e+00 3.23114216e-01
3.22375596e-02 -9.74342585e-01 -6.80298567e-01 3.45376849e-01
-2.97681242e-01 -1.46998966e+00 1.38925031e-01 -1.67564064e-01
-5.25093436e-01 -1.50718570e+00 2.45879203e-01 -6.01938665e-01
2.75680959e-01 -9.15238708e-02 1.79716611e+00 3.83152008e-01
-2.97397524e-01 6.90724194e-01 -4.49615806e-01 -8.28408122e-01
-4.26609516e-01 4.87654001e-01 6.07686006e-02 -6.54288173e-01
1.02891171e+00 -7.18089342e-01 -6.61656037e-02 6.40621066e-01
-8.51544321e-01 -5.37781119e-02 9.70989019e-02 8.71797502e-02
4.21200275e-01 9.90289729e-03 5.63680530e-01 -9.57549453e-01
5.66270947e-01 -6.42119229e-01 -7.47729480e-01 1.02691162e+00
-1.40189385e+00 -6.17900491e-02 -1.49630010e-01 3.22837442e-01
-7.74693072e-01 -2.61410892e-01 2.99638254e-03 -5.65958992e-02
1.28515184e-01 9.01230276e-01 -2.51168698e-01 -4.44585949e-01
7.69193769e-01 -5.39365888e-01 -3.51467341e-01 -7.35157728e-01
3.31082225e-01 5.45665503e-01 7.69119024e-01 -8.28089952e-01
7.98559010e-01 2.82987177e-01 -6.74313074e-03 -2.90607810e-01
-4.97239858e-01 -5.25144041e-01 -6.99986458e-01 -3.67414296e-01
6.86558723e-01 -8.31466019e-01 -3.72015804e-01 7.43346140e-02
-1.11065602e+00 -1.98835973e-02 -4.75201815e-01 1.47276282e-01
-2.58673519e-01 2.80724555e-01 5.91256320e-02 -3.43225718e-01
-3.15073133e-01 -6.13692939e-01 4.86798376e-01 2.14979351e-01
-7.13386953e-01 -1.11981332e+00 4.68659520e-01 2.47742698e-01
4.47739840e-01 3.23057562e-01 9.47493732e-01 -1.17189097e+00
-2.33972058e-01 -2.91198909e-01 -1.54713973e-01 -2.01178625e-01
4.33761291e-02 1.78241089e-01 -6.58227205e-01 -7.32918084e-02
-6.78969502e-01 9.20192748e-02 -1.44744307e-01 -4.75691676e-01
4.80408639e-01 5.32372529e-03 -7.41833389e-01 4.33156341e-01
1.48060846e+00 2.89458394e-01 9.05661106e-01 1.16203296e+00
5.37950873e-01 9.36489880e-01 7.32278883e-01 1.37561569e-02
8.48188281e-01 1.30277932e+00 1.86912000e-01 2.17720628e-01
-3.81914943e-01 -3.08702767e-01 -1.65404037e-01 1.02228022e+00
-5.88169813e-01 3.53227817e-02 -1.51571250e+00 7.24353313e-01
-2.13894272e+00 -9.76140499e-01 -4.73390430e-01 2.25511789e+00
7.43466377e-01 5.13966642e-02 3.94576825e-02 -2.45217547e-01
5.84321320e-01 -4.71647054e-01 7.07306489e-02 -1.17923483e-01
-4.42825377e-01 5.46947896e-01 5.14542162e-01 4.89963531e-01
-5.91296315e-01 1.01557171e+00 7.68234301e+00 3.35498929e-01
-2.75940776e-01 3.75101507e-01 -7.81099916e-01 2.52548724e-01
-5.24436235e-01 7.11914062e-01 -8.11221719e-01 1.75061941e-01
1.17024207e+00 -1.04006124e+00 4.42226917e-01 7.25582480e-01
-5.82446039e-01 2.40988374e-01 -1.21896195e+00 5.98744333e-01
-6.04705662e-02 -1.65139091e+00 4.06958871e-02 1.72512814e-01
5.77644825e-01 4.46575373e-01 -9.07880127e-01 3.87355626e-01
6.79504097e-01 -9.22281384e-01 7.99957633e-01 9.09625769e-01
5.79167962e-01 -4.04010683e-01 8.42788815e-01 -3.81510824e-01
-1.27031541e+00 1.62110105e-01 -2.13306934e-01 1.57719448e-01
1.71024397e-01 8.14478099e-02 -4.39101160e-01 1.55047095e+00
1.05216694e+00 6.91785574e-01 -1.01586533e+00 1.47757971e+00
3.14655416e-02 -5.02876751e-02 -2.77137756e-01 5.20507693e-01
-4.02513415e-01 -2.19777733e-01 7.47116446e-01 1.25411594e+00
2.90979832e-01 -1.82983845e-01 1.23367347e-02 5.85751593e-01
-2.41416350e-01 1.84224024e-01 -7.26299047e-01 -1.50239334e-01
1.10718203e+00 1.02974939e+00 -2.04946488e-01 -4.21944946e-01
-6.12450600e-01 5.95412314e-01 4.72685516e-01 2.27944121e-01
-5.65338016e-01 -8.30882072e-01 6.62570834e-01 3.73653583e-02
-6.18579313e-02 2.11033449e-01 -1.45217925e-01 -1.00383770e+00
2.19312102e-01 -9.78090048e-01 1.05089617e+00 -1.08500755e+00
-1.08098781e+00 5.67227423e-01 4.07529771e-01 -9.07779217e-01
-2.33937472e-01 -2.66705692e-01 -2.69009918e-01 8.43778253e-01
-1.22279656e+00 -1.00225043e+00 -6.07848823e-01 5.73031008e-01
-4.40310180e-01 -3.30295861e-01 1.35584664e+00 1.02256274e+00
-3.73980939e-01 6.71912432e-01 -2.71233559e-01 8.74301270e-02
9.35880899e-01 -1.25883496e+00 7.70297289e-01 6.73930764e-01
2.92930990e-01 9.97165918e-01 8.96563649e-01 -8.05908024e-01
-1.19514227e+00 -7.90445387e-01 1.59542370e+00 -1.04559660e+00
1.17720473e+00 2.61733141e-02 -9.09105301e-01 1.26165593e+00
4.53544259e-01 -3.13813806e-01 7.39544094e-01 3.13728839e-01
-5.52541375e-01 5.24597801e-02 -1.09007192e+00 4.36928928e-01
1.78467178e+00 -6.60822630e-01 -1.16144180e+00 4.10941154e-01
4.61849332e-01 -5.55969119e-01 -1.74097109e+00 7.47466862e-01
9.79264677e-01 -9.66660023e-01 8.92926812e-01 -1.26309764e+00
-1.60498936e-02 -5.67482173e-01 -2.26857424e-01 -9.45575893e-01
-2.87564516e-01 -7.03088999e-01 -2.79897451e-01 1.42561781e+00
5.73528469e-01 -9.76281404e-01 4.81389642e-01 5.68267643e-01
-4.28758323e-01 -1.58241894e-02 -9.98603344e-01 -1.14679444e+00
-3.55065495e-01 -3.17526281e-01 1.12877560e+00 1.67039979e+00
7.19327748e-01 1.42135143e-01 2.07563192e-01 3.10431033e-01
6.89145327e-01 -2.03242376e-01 9.15127993e-01 -1.77875280e+00
4.60204519e-02 -5.48981667e-01 -8.77699971e-01 5.63524477e-02
-2.19929978e-01 -1.20572829e+00 -7.09340572e-01 -2.04741311e+00
1.26615418e-02 -3.60554695e-01 -6.13627672e-01 7.27963686e-01
2.78605763e-02 -1.19649597e-01 -1.25352830e-01 6.38360858e-01
-8.61080229e-01 -2.70216346e-01 4.57963139e-01 8.90820175e-02
1.68190360e-01 -4.93078202e-01 -7.19990432e-01 5.82221508e-01
6.13804579e-01 -6.96439266e-01 -3.48672092e-01 -3.97787929e-01
9.11417663e-01 -5.29807389e-01 2.30353296e-01 -1.25697470e+00
6.14894748e-01 -2.96775941e-02 -3.19369286e-01 -3.40133339e-01
-1.89817324e-01 -9.35072064e-01 1.31643403e+00 2.80937016e-01
2.22056713e-02 4.91223007e-01 3.44575822e-01 -9.43780392e-02
-3.30345213e-01 -6.54226720e-01 1.51961997e-01 -9.05335769e-02
-1.12314856e+00 3.35584283e-02 4.25395593e-02 3.09427828e-01
8.87338817e-01 -3.58839333e-01 -8.76879752e-01 2.34123141e-01
-7.16184497e-01 4.32066321e-01 8.55731189e-01 6.90195024e-01
-1.01986192e-01 -1.58818913e+00 -4.96004730e-01 -3.94823313e-01
9.43778753e-01 -5.58504581e-01 -2.06374049e-01 1.05818820e+00
-4.65136945e-01 4.99156475e-01 -6.30140007e-01 -4.21938524e-02
-1.20271516e+00 4.19648170e-01 5.61342299e-01 -3.68906528e-01
-6.32365763e-01 3.44083667e-01 -6.88174605e-01 -7.18169510e-01
9.58512649e-02 3.68018091e-01 -4.17436272e-01 1.53306752e-01
4.53188062e-01 6.35594010e-01 6.76239371e-01 -4.53708678e-01
-7.60130465e-01 3.83801341e-01 1.10006995e-01 1.40082352e-02
1.23861969e+00 6.18602186e-02 -6.52803063e-01 3.62194419e-01
5.82339406e-01 3.47562492e-01 1.27405915e-02 -3.68675053e-01
1.03057039e+00 -3.29291582e-01 -2.96084344e-01 -1.05109537e+00
-6.95926487e-01 3.07906251e-02 5.42492747e-01 5.15677631e-01
5.72722912e-01 1.22019552e-01 3.16410989e-01 4.43781585e-01
8.04728746e-01 -1.30089343e+00 -8.82442236e-01 4.50971127e-01
8.10020626e-01 -6.43776000e-01 1.93827763e-01 -8.05584610e-01
-1.72858313e-01 1.08429873e+00 7.28132129e-01 5.86613476e-01
4.83421892e-01 2.30873466e-01 1.26093000e-01 -1.02778292e+00
-6.99536741e-01 -4.38236177e-01 4.26621228e-01 9.86293674e-01
4.38047290e-01 -3.77533406e-01 -7.06530333e-01 5.68711460e-01
-3.65353912e-01 4.85521764e-01 2.48441905e-01 1.20717168e+00
-1.55980229e-01 -1.62515485e+00 -7.44955987e-02 2.96613067e-01
-3.66390318e-01 -1.78528845e-01 -9.41878378e-01 1.31882370e+00
3.61761659e-01 8.02591383e-01 3.09055541e-02 -7.02151835e-01
1.30803192e+00 5.26581585e-01 7.50803947e-01 -4.96435285e-01
-8.80439043e-01 -8.75847697e-01 1.20820475e+00 -8.52361143e-01
-6.10350847e-01 -6.55880272e-01 -1.11267841e+00 -4.15205091e-01
-2.72976965e-01 7.18999684e-01 5.25267124e-01 7.28047669e-01
5.93686283e-01 5.41582227e-01 -3.36429775e-01 1.21105708e-01
4.96566892e-02 -7.68009901e-01 -2.56802469e-01 5.82916498e-01
-7.40968704e-01 -9.29402888e-01 8.81033689e-02 1.14610970e-01] | [9.21380615234375, 8.055939674377441] |
463f0a06-f2ee-4c1a-ba34-abd240ec331c | tsmixer-lightweight-mlp-mixer-model-for | 2306.09364 | null | https://arxiv.org/abs/2306.09364v3 | https://arxiv.org/pdf/2306.09364v3.pdf | TSMixer: Lightweight MLP-Mixer Model for Multivariate Time Series Forecasting | Transformers have gained popularity in time series forecasting for their ability to capture long-sequence interactions. However, their high memory and computing requirements pose a critical bottleneck for long-term forecasting. To address this, we propose TSMixer, a lightweight neural architecture exclusively composed of multi-layer perceptron (MLP) modules. TSMixer is designed for multivariate forecasting and representation learning on patched time series, providing an efficient alternative to Transformers. Our model draws inspiration from the success of MLP-Mixer models in computer vision. We demonstrate the challenges involved in adapting Vision MLP-Mixer for time series and introduce empirically validated components to enhance accuracy. This includes a novel design paradigm of attaching online reconciliation heads to the MLP-Mixer backbone, for explicitly modeling the time-series properties such as hierarchy and channel-correlations. We also propose a Hybrid channel modeling approach to effectively handle noisy channel interactions and generalization across diverse datasets, a common challenge in existing patch channel-mixing methods. Additionally, a simple gated attention mechanism is introduced in the backbone to prioritize important features. By incorporating these lightweight components, we significantly enhance the learning capability of simple MLP structures, outperforming complex Transformer models with minimal computing usage. Moreover, TSMixer's modular design enables compatibility with both supervised and masked self-supervised learning methods, making it a promising building block for time-series Foundation Models. TSMixer outperforms state-of-the-art MLP and Transformer models in forecasting by a considerable margin of 8-60%. It also outperforms the latest strong benchmarks of Patch-Transformer models (by 1-2%) with a significant reduction in memory and runtime (2-3X). | ['Jayant Kalagnanam', 'Phanwadee Sinthong', 'Nam Nguyen', 'Arindam Jati', 'Vijay Ekambaram'] | 2023-06-14 | null | null | null | null | ['multivariate-time-series-forecasting'] | ['time-series'] | [ 2.34716937e-01 -2.87455767e-01 -3.53459090e-01 -2.97080278e-01
-1.03578758e+00 -4.73128945e-01 6.19024873e-01 1.24310799e-01
-7.16984943e-02 4.01057363e-01 2.82935470e-01 -6.28533185e-01
7.20806494e-02 -4.89064842e-01 -9.45767879e-01 -7.49522388e-01
-5.19414365e-01 1.70327108e-02 6.12988174e-02 -2.23083496e-01
-5.57805002e-02 4.14338887e-01 -1.73365021e+00 6.74303293e-01
5.02493203e-01 1.55874026e+00 1.00717157e-01 6.77589715e-01
-2.35520139e-01 1.22818196e+00 -4.73662645e-01 -2.04948395e-01
6.66831210e-02 -1.40401512e-01 -3.80977601e-01 -4.29512680e-01
4.23144281e-01 -2.57254809e-01 -5.10363460e-01 2.97479361e-01
7.11621284e-01 -1.38661146e-01 3.16447109e-01 -1.27268624e+00
-2.17284977e-01 1.08738554e+00 -3.01368386e-01 3.03449243e-01
-6.55712634e-02 1.46000370e-01 1.22968984e+00 -7.71696925e-01
-4.26444085e-03 1.10529089e+00 1.27359045e+00 3.87995869e-01
-1.47045028e+00 -8.43495965e-01 6.56502903e-01 3.18622321e-01
-1.06827819e+00 -6.81032181e-01 9.18974400e-01 -3.75211209e-01
1.53386962e+00 4.48333323e-01 5.65146208e-01 1.51111245e+00
1.36011884e-01 1.08443737e+00 9.93362904e-01 -1.62448213e-01
3.97670269e-01 -8.89012963e-02 2.10840508e-01 2.95389444e-01
-6.08302414e-01 1.67392358e-01 -7.31299281e-01 -3.46255064e-01
5.73343217e-01 2.69011140e-01 -6.14701547e-02 2.06993639e-01
-1.25642979e+00 4.18452471e-01 5.01091599e-01 1.76742300e-01
-4.71674800e-01 6.98589087e-01 6.66003823e-01 4.92292374e-01
4.62466091e-01 2.22753286e-01 -8.73862445e-01 -4.57456231e-01
-1.04213858e+00 7.97757953e-02 9.46821511e-01 7.87370622e-01
4.47978675e-01 4.25911933e-01 -2.38737792e-01 6.97680116e-01
2.60055810e-01 6.04310155e-01 5.03057897e-01 -7.37616181e-01
4.94993210e-01 4.47324663e-01 -3.78530025e-01 -7.98488081e-01
-5.65780222e-01 -1.13774264e+00 -9.99360502e-01 -2.02182531e-01
1.93004295e-01 1.03997782e-01 -7.81419158e-01 1.68994868e+00
-5.34345917e-02 6.89231634e-01 -6.11696877e-02 4.40833211e-01
6.88550711e-01 8.42198610e-01 1.90572545e-01 -2.94042110e-01
1.15004325e+00 -1.18988311e+00 -3.17991346e-01 -1.88234389e-01
4.91256386e-01 -6.40112102e-01 9.36179757e-01 4.83903646e-01
-8.36548924e-01 -6.17084146e-01 -9.96428192e-01 4.84719165e-02
-2.66793072e-01 1.18146956e-01 8.44769061e-01 5.76695502e-01
-8.83180678e-01 8.06659102e-01 -1.06605983e+00 2.10090950e-02
3.77391845e-01 3.85972559e-01 9.94727463e-02 3.15562457e-01
-1.02834439e+00 5.28609872e-01 -2.05994025e-02 2.34085351e-01
-1.09696460e+00 -1.19209719e+00 -4.96595979e-01 2.95486003e-01
-8.19400176e-02 -3.71336311e-01 1.45052028e+00 -9.85896826e-01
-1.84164143e+00 1.77812397e-01 -3.07506174e-01 -8.11286926e-01
3.93782079e-01 -1.85764223e-01 -6.85172081e-01 -8.76433849e-02
-2.14585140e-01 5.53520501e-01 1.03745580e+00 -8.98829639e-01
-6.27903759e-01 -5.73114343e-02 -2.00480253e-01 -1.66522875e-01
-5.47770500e-01 -1.67041376e-01 -4.71434355e-01 -9.60490167e-01
-3.49099487e-02 -7.85584331e-01 -4.73343313e-01 -2.02790320e-01
-2.70942271e-01 6.56447187e-02 9.63465095e-01 -5.74588418e-01
1.56656635e+00 -2.13073778e+00 -2.77133971e-01 2.61635959e-01
4.04691070e-01 6.64511397e-02 -2.31683567e-01 7.04967618e-01
-1.00081652e-01 -2.87453175e-01 -1.31932795e-01 -8.36653411e-01
1.08974591e-01 2.66867727e-01 -9.79869008e-01 1.87173247e-01
1.09315358e-01 1.06994903e+00 -6.83956563e-01 1.17373139e-01
2.99046785e-01 7.72057772e-01 -5.26945949e-01 3.46229151e-02
-3.65671694e-01 5.50776660e-01 -7.53889531e-02 9.38216746e-01
4.47368592e-01 -6.32720411e-01 2.18644753e-01 -3.26352477e-01
-1.65813342e-01 5.96449494e-01 -8.09080958e-01 1.45089984e+00
-8.41932416e-01 7.55018115e-01 -5.36604971e-02 -1.07664073e+00
9.43947434e-01 4.21694666e-01 6.88194692e-01 -9.94866729e-01
8.39619711e-02 4.21149462e-01 -2.60325402e-01 -1.70392796e-01
2.99134582e-01 8.08229223e-02 -2.88392790e-02 3.75103086e-01
9.28637981e-02 4.21068221e-01 -1.16777010e-01 -1.20237865e-01
1.24590969e+00 -6.67636888e-03 -2.48252556e-01 -5.75664975e-02
2.73170978e-01 -3.16559076e-01 5.13727486e-01 8.55271518e-01
6.98893070e-02 4.11963582e-01 4.32789236e-01 -6.53910220e-01
-8.32552433e-01 -9.27328229e-01 -9.82632041e-02 1.47687078e+00
-3.50493312e-01 -7.80248284e-01 -3.34372491e-01 -3.45332623e-01
1.62200332e-01 3.41833353e-01 -4.95809913e-01 4.88023944e-02
-5.67208886e-01 -8.72939825e-01 9.06806111e-01 9.55926001e-01
3.82929504e-01 -7.44607329e-01 -5.06514192e-01 7.06755161e-01
-1.50900260e-01 -1.04803491e+00 -2.37914354e-01 7.43534446e-01
-1.08592260e+00 -6.04726255e-01 -6.28698766e-01 -6.14727855e-01
1.83253780e-01 4.31953073e-02 1.11485028e+00 -3.42762291e-01
-1.07168749e-01 2.66589701e-01 -2.03418061e-01 -3.94359648e-01
-3.11698824e-01 2.91921109e-01 -5.91073409e-02 2.72157699e-01
1.76677629e-01 -1.31047606e+00 -7.20492601e-01 4.31165338e-01
-5.07706285e-01 -2.67891251e-02 7.26166248e-01 8.35783362e-01
5.54546893e-01 -2.17763722e-01 6.08225286e-01 -5.34632504e-01
5.22718310e-01 -6.76206887e-01 -6.22354567e-01 7.89244100e-02
-8.18211019e-01 2.61338037e-02 1.02184176e+00 -8.29042673e-01
-6.87917352e-01 1.20348834e-01 -3.62738431e-01 -6.55754268e-01
1.72036946e-01 6.66922092e-01 1.76736459e-01 -1.56504720e-01
5.70471883e-01 4.65625763e-01 -7.27552408e-03 -8.01649690e-01
2.40314543e-01 6.83793783e-01 5.40288031e-01 -5.78716815e-01
2.74540037e-01 5.45426011e-01 -7.67331645e-02 -7.90576398e-01
-5.35703182e-01 -4.67394084e-01 -1.77928403e-01 -1.29168868e-01
2.19047502e-01 -1.26422822e+00 -9.98877823e-01 7.23845363e-01
-1.00500047e+00 -4.89062220e-01 -1.49298042e-01 4.39088553e-01
-3.29451054e-01 1.11498415e-01 -9.43083882e-01 -7.80728459e-01
-5.57112873e-01 -1.10688615e+00 1.18541598e+00 -1.71520650e-01
-1.68184936e-01 -1.07704091e+00 -1.55455451e-02 1.50953054e-01
9.38254178e-01 6.67525306e-02 8.59384596e-01 -5.61037421e-01
-7.57393539e-01 -1.60129175e-01 -8.83628204e-02 4.21147376e-01
-2.55365968e-01 -2.44389907e-01 -1.62292683e+00 -2.80626506e-01
-2.30668902e-01 -7.37017840e-02 1.12898254e+00 5.29588819e-01
1.24459267e+00 -3.35135013e-01 -4.82913494e-01 8.65206420e-01
1.16611493e+00 1.38826042e-01 3.95590037e-01 1.34812981e-01
8.52317572e-01 2.07087174e-01 -1.43684313e-01 6.48422182e-01
5.93116879e-01 5.45365632e-01 4.23031867e-01 -1.75259307e-01
5.55358529e-02 -4.71316278e-01 7.03312159e-01 1.34176898e+00
1.09098904e-01 -8.68680999e-02 -8.52801323e-01 4.06792849e-01
-2.01054907e+00 -9.59534764e-01 -8.55870321e-02 2.06370139e+00
7.45412230e-01 2.30690479e-01 1.33558914e-01 3.93724471e-01
1.14106335e-01 3.39881122e-01 -6.93145454e-01 -3.30251008e-01
-2.11910933e-01 2.28872955e-01 6.29028797e-01 2.59386659e-01
-1.19171274e+00 6.13326371e-01 7.03516626e+00 9.53826904e-01
-1.76553881e+00 1.27369389e-01 6.59977674e-01 -2.34743923e-01
-2.98160225e-01 7.45776435e-03 -9.94542181e-01 6.62542045e-01
1.48008609e+00 2.64104694e-01 5.99916518e-01 8.22665513e-01
1.72858790e-01 4.39554811e-01 -1.17334402e+00 1.24244618e+00
-1.91100016e-01 -1.73230374e+00 -1.28205314e-01 -6.62401021e-02
4.59381819e-01 5.64631224e-01 2.44508550e-01 5.54719269e-01
-2.61489805e-02 -1.05987263e+00 9.80991244e-01 3.71392548e-01
6.55968368e-01 -4.15389925e-01 4.37136650e-01 2.75720716e-01
-1.59516883e+00 -5.55757344e-01 5.89667521e-02 -3.82383972e-01
2.07305819e-01 9.07785833e-01 -5.51967204e-01 3.11194181e-01
8.23823452e-01 9.98997808e-01 -3.64540339e-01 9.12263691e-01
1.06075816e-01 1.10437441e+00 -6.76610589e-01 2.85677668e-02
2.88484216e-01 3.68214041e-01 4.08061028e-01 1.37899292e+00
2.59716839e-01 -4.48257387e-01 2.47642428e-01 4.94751841e-01
1.81067754e-02 -5.67517150e-03 -2.10641339e-01 -2.28887275e-02
7.06831813e-01 1.11575520e+00 -4.85445231e-01 -2.70491004e-01
-6.00695133e-01 7.26105511e-01 1.61871389e-01 3.77743572e-01
-9.57989156e-01 -1.95569143e-01 6.75106585e-01 -1.38114078e-03
6.43205225e-01 -1.69652045e-01 -2.63175935e-01 -1.04933453e+00
2.38898218e-01 -1.04340541e+00 2.69637644e-01 -3.58920276e-01
-1.39538121e+00 9.35903847e-01 -3.74463439e-01 -1.46639466e+00
-2.13851288e-01 -6.05070651e-01 -5.67470670e-01 8.18955421e-01
-1.64811945e+00 -1.55301189e+00 -1.58667266e-01 7.04396665e-01
4.42313343e-01 -1.64153084e-01 1.00961494e+00 7.21307158e-01
-5.38336396e-01 7.68856585e-01 4.24700081e-01 -2.02586100e-01
4.42582220e-01 -1.08339429e+00 8.13550830e-01 6.33928061e-01
1.11158028e-01 7.03382969e-01 4.65777755e-01 -2.59084046e-01
-1.73557329e+00 -1.26805878e+00 7.46972203e-01 -3.97214174e-01
9.07503307e-01 -7.62048721e-01 -8.19499135e-01 6.51907742e-01
-1.90100819e-02 1.36278898e-01 7.89912403e-01 4.67847615e-01
-9.08403158e-01 -6.39600217e-01 -4.92469370e-01 5.02012730e-01
9.87937570e-01 -9.04709160e-01 -5.62382601e-02 4.20076139e-02
6.57924533e-01 -3.30817670e-01 -9.97591197e-01 2.88997918e-01
9.48915303e-01 -8.55688453e-01 1.01172686e+00 -1.52665392e-01
2.72638917e-01 -2.05347463e-01 -3.77151757e-01 -1.14430249e+00
-3.50921214e-01 -1.26651657e+00 -7.30301678e-01 1.10832405e+00
6.59512997e-01 -8.33792925e-01 8.18224013e-01 1.76505134e-01
-2.64960736e-01 -9.26289022e-01 -9.67382669e-01 -7.19863832e-01
-1.66331232e-01 -1.05493999e+00 6.34689271e-01 8.45121741e-01
5.98724224e-02 4.44702327e-01 -6.53522372e-01 2.32348979e-01
6.02184713e-01 1.45808354e-01 5.51131725e-01 -1.16511238e+00
-7.50923693e-01 -6.97027802e-01 -2.09781677e-01 -1.57605088e+00
-6.37700558e-02 -8.67562771e-01 -1.73189819e-01 -9.26526368e-01
-2.81610847e-01 -7.04924285e-01 -6.30182743e-01 5.95698833e-01
2.47671649e-01 2.59084642e-01 1.21443398e-01 4.72536266e-01
-4.41966027e-01 5.97622335e-01 6.66478932e-01 -2.58082688e-01
-4.34137732e-01 2.38429263e-01 -6.24877274e-01 4.31811541e-01
6.89038217e-01 -2.32814983e-01 -5.30034661e-01 -4.99032617e-01
1.99909791e-01 1.02074169e-01 4.41763103e-01 -1.29398811e+00
4.58339334e-01 3.59770536e-01 3.17640990e-01 -6.52779520e-01
5.79091609e-01 -9.51904178e-01 3.17230642e-01 2.85436690e-01
-4.42433923e-01 1.70015097e-01 5.91862500e-01 6.79884613e-01
-3.15285116e-01 6.81130469e-01 2.53832310e-01 7.04934075e-02
-7.93474615e-01 3.34627509e-01 -5.03265321e-01 -3.35723311e-01
6.87684894e-01 -1.66212156e-01 -3.66401434e-01 -3.46036017e-01
-6.41946197e-01 2.41203010e-01 -1.99466988e-01 5.31443059e-01
3.55280757e-01 -1.12251997e+00 -3.66169333e-01 5.26704311e-01
2.03685552e-01 -3.85806024e-01 3.82692873e-01 1.18217564e+00
-1.76160187e-01 5.60930371e-01 -5.78923561e-02 -8.97503674e-01
-9.71772969e-01 4.43337291e-01 3.72729987e-01 -2.94657856e-01
-7.23477721e-01 8.97577047e-01 2.90868543e-02 -4.45789456e-01
8.06283951e-01 -9.43029761e-01 9.04661715e-02 7.63435736e-02
6.94429934e-01 3.66109073e-01 3.49155337e-01 -2.67715931e-01
-3.43467295e-01 4.28456783e-01 -1.53146625e-01 3.41617242e-02
1.53234410e+00 -4.92274389e-02 -1.97451171e-02 7.52205849e-01
1.35561538e+00 -1.87762901e-02 -1.44994307e+00 -4.51028705e-01
-6.30257279e-02 1.49318367e-01 2.55379587e-01 -9.77130175e-01
-8.87341261e-01 1.04077661e+00 5.92302561e-01 3.18589568e-01
1.36387944e+00 -1.32598951e-01 1.02288854e+00 3.96733493e-01
3.41422319e-01 -7.48850346e-01 -2.26776734e-01 6.21280253e-01
6.93492293e-01 -8.76857400e-01 -4.30722505e-01 -1.58793405e-01
-4.18237478e-01 1.10686648e+00 1.83849514e-01 1.22554645e-01
9.66516137e-01 7.75742352e-01 2.76194602e-01 5.67940399e-02
-1.35280871e+00 1.03604928e-01 2.64349997e-01 4.63934511e-01
4.76712406e-01 1.31123707e-01 3.55332434e-01 7.73715675e-01
-5.48626930e-02 -3.55138928e-02 -1.30335152e-01 9.27314579e-01
-9.90441889e-02 -1.16194201e+00 -2.14618996e-01 5.08316875e-01
-4.55475897e-01 -3.94155800e-01 7.58149028e-02 3.88496161e-01
-7.22052455e-02 8.27073276e-01 1.59279972e-01 -7.35891998e-01
1.83129191e-01 -5.64962886e-02 1.17474839e-01 -1.47917522e-02
-1.07816827e+00 4.30417478e-01 5.84927239e-02 -7.99599111e-01
-2.55454302e-01 -3.78852785e-01 -8.60561550e-01 -4.01970655e-01
-8.09612870e-02 1.54046584e-02 7.02859223e-01 7.79373467e-01
8.03553760e-01 8.33342314e-01 6.42831326e-01 -1.11363256e+00
-6.89395070e-01 -8.35062563e-01 -3.57033759e-01 -1.40780032e-01
8.20706427e-01 -3.04643363e-01 -3.21652025e-01 1.12533025e-01] | [7.076905727386475, 2.9522225856781006] |
c40cce24-9492-46dd-925c-5757876fff92 | context-aware-text-based-binary-image | 1810.03767 | null | http://arxiv.org/abs/1810.03767v1 | http://arxiv.org/pdf/1810.03767v1.pdf | Context-Aware Text-Based Binary Image Stylization and Synthesis | In this work, we present a new framework for the stylization of text-based
binary images. First, our method stylizes the stroke-based geometric shape like
text, symbols and icons in the target binary image based on an input style
image. Second, the composition of the stylized geometric shape and a background
image is explored. To accomplish the task, we propose legibility-preserving
structure and texture transfer algorithms, which progressively narrow the
visual differences between the binary image and the style image. The
stylization is then followed by a context-aware layout design algorithm, where
cues for both seamlessness and aesthetics are employed to determine the optimal
layout of the shape in the background. Given the layout, the binary image is
seamlessly embedded into the background by texture synthesis under a
context-aware boundary constraint. According to the contents of binary images,
our method can be applied to many fields. We show that the proposed method is
capable of addressing the unsupervised text stylization problem and is superior
to state-of-the-art style transfer methods in automatic artistic typography
creation. Besides, extensive experiments on various tasks, such as
visual-textual presentation synthesis, icon/symbol rendering and
structure-guided image inpainting, demonstrate the effectiveness of the
proposed method. | ['Shuai Yang', 'Jiaying Liu', 'Zongming Guo', 'Wenhan Yang'] | 2018-10-09 | null | null | null | null | ['layout-design', 'image-stylization'] | ['computer-vision', 'computer-vision'] | [ 9.74809170e-01 -2.41427988e-01 2.55119968e-02 -4.35070023e-02
-3.55719067e-02 -7.33576238e-01 6.61630511e-01 -2.27681417e-02
-3.67784947e-02 5.06230772e-01 -3.40410247e-02 -3.03983688e-01
-2.28165295e-02 -9.00669158e-01 -7.90300608e-01 -5.79504430e-01
7.88674474e-01 3.25929701e-01 2.51570910e-01 -2.62670487e-01
6.65607154e-01 6.09136283e-01 -1.42666662e+00 2.86853075e-01
1.14908731e+00 9.88975823e-01 4.78402078e-01 6.11556888e-01
-7.50436544e-01 4.65739042e-01 -6.46972537e-01 -4.48020041e-01
2.23907888e-01 -8.04642200e-01 -5.48736453e-01 7.41045833e-01
5.09903610e-01 -1.19227976e-01 -1.15760036e-01 1.24244535e+00
1.15349829e-01 6.62326217e-02 8.18511128e-01 -9.38278615e-01
-1.00286448e+00 3.52920502e-01 -8.14710140e-01 -3.60471904e-01
2.37553224e-01 3.58039886e-02 8.12277615e-01 -1.00217581e+00
9.75897074e-01 1.37752187e+00 2.16677383e-01 3.85643065e-01
-1.57421935e+00 -4.61959869e-01 1.60097286e-01 1.13257440e-02
-1.20543289e+00 -1.78472042e-01 1.31592166e+00 -5.04619181e-01
1.02118932e-01 5.25837421e-01 7.80427516e-01 7.91469157e-01
2.47694924e-01 7.70008445e-01 1.23482513e+00 -8.70984674e-01
2.93456554e-01 2.24688813e-01 -3.77456754e-01 7.31099665e-01
9.21444595e-02 -1.98181629e-01 -3.66549700e-01 4.00851332e-02
1.20154500e+00 5.39576076e-02 -2.99650609e-01 -5.81869543e-01
-1.10204136e+00 4.89905179e-01 3.45237374e-01 4.20892596e-01
-2.28759527e-01 -3.12901568e-03 2.28203148e-01 2.98302509e-02
5.12779415e-01 6.10344827e-01 2.09864229e-01 9.94802639e-02
-1.20062256e+00 1.82203934e-01 5.48694849e-01 1.19406760e+00
7.63262391e-01 6.45878613e-02 -4.11121637e-01 1.05733538e+00
7.05018565e-02 6.52057230e-01 2.25229815e-01 -7.77953982e-01
5.15841067e-01 7.21011758e-01 2.11864933e-01 -1.24494743e+00
2.41999164e-01 8.12925324e-02 -8.19336057e-01 4.91923362e-01
2.90885627e-01 1.40010953e-01 -9.53850746e-01 1.25790477e+00
1.97731376e-01 -8.33210573e-02 -2.35257715e-01 6.78473890e-01
5.71286440e-01 6.05632901e-01 -1.04816355e-01 -5.50710708e-02
1.48058498e+00 -9.53014612e-01 -1.06847370e+00 -2.35248297e-01
-5.25603220e-02 -1.25904047e+00 1.48113346e+00 2.72116005e-01
-1.23660731e+00 -5.98493397e-01 -1.11660910e+00 -8.13745931e-02
-5.99615946e-02 5.27460337e-01 3.76367643e-02 7.19369471e-01
-7.98751533e-01 4.78788376e-01 -3.15547794e-01 -4.07863826e-01
3.24201018e-01 -1.53882638e-01 4.86674719e-02 1.21252134e-01
-7.39268243e-01 5.88762283e-01 3.98008049e-01 -2.88966727e-02
-2.08327562e-01 -5.30831873e-01 -7.13040113e-01 5.76377846e-02
3.67045134e-01 -7.33049035e-01 9.02819753e-01 -1.53392363e+00
-1.91164851e+00 8.44370663e-01 -2.64868945e-01 -1.83484294e-02
7.24338591e-01 -7.01721460e-02 -1.81209728e-01 2.42987871e-01
-2.37815874e-03 5.77085793e-01 1.48267138e+00 -1.75673187e+00
-6.96099043e-01 -1.20039647e-02 -2.72419363e-01 4.05921310e-01
-3.30087125e-01 -2.26399064e-01 -9.13314581e-01 -1.32433236e+00
1.67190284e-01 -6.68993115e-01 -9.17585492e-02 2.50574976e-01
-6.77882969e-01 4.39284682e-01 1.20983267e+00 -7.62287140e-01
1.33806086e+00 -2.25007796e+00 3.60965222e-01 6.96024895e-01
-1.74144879e-02 1.76226139e-01 -1.77822709e-01 3.48596036e-01
1.08977005e-01 1.42974988e-01 -4.57854420e-01 -4.52802896e-01
-1.92085147e-01 6.05444461e-02 -5.19138336e-01 -3.74543574e-03
1.56484917e-01 1.02426541e+00 -7.96960652e-01 -6.70548975e-01
4.18255627e-01 2.54186183e-01 -4.26339209e-01 3.10116023e-01
-5.27641475e-01 7.17534840e-01 -4.90854412e-01 5.48752904e-01
8.41160536e-01 1.08576082e-01 2.94163525e-01 -2.86889434e-01
-1.54432401e-01 -1.77138895e-01 -1.16067219e+00 1.57235968e+00
-5.82069218e-01 7.57198930e-01 9.52599421e-02 -4.93573397e-01
1.23870754e+00 -2.24408507e-01 4.92466763e-02 -7.54544616e-01
2.58857399e-01 2.38063186e-01 -3.74159515e-01 -3.02483052e-01
9.60325480e-01 -1.01619484e-02 6.85412362e-02 5.41379929e-01
-4.76723492e-01 -6.55002058e-01 2.49345079e-01 5.42771481e-02
4.21132952e-01 2.97740251e-01 1.24496043e-01 -2.94459760e-01
6.46196425e-01 -1.43865153e-01 5.13155162e-02 7.15363920e-01
4.51272786e-01 7.86944151e-01 6.25540555e-01 -2.56392002e-01
-1.40334690e+00 -9.22430575e-01 2.03735568e-02 8.52760494e-01
4.60269481e-01 -3.45309258e-01 -1.38676059e+00 -4.73594606e-01
-1.38765126e-01 7.90505707e-01 -6.65980816e-01 2.45750267e-02
-6.73667490e-01 -1.94598541e-01 2.93510973e-01 3.51243645e-01
6.51240468e-01 -1.27610910e+00 -5.17209709e-01 2.10481673e-01
-1.91420883e-01 -1.00597274e+00 -1.03456914e+00 -2.54102200e-01
-1.01471639e+00 -7.48628259e-01 -9.75858688e-01 -1.00859487e+00
1.13290584e+00 2.66363233e-01 8.38510871e-01 2.20463857e-01
-3.20500404e-01 2.51015365e-01 -3.31982225e-01 -2.55247969e-02
-6.72429323e-01 -1.77093536e-01 -4.53911781e-01 6.03132129e-01
-3.93950731e-01 -5.35111129e-01 -6.11983657e-01 3.87514859e-01
-1.39462590e+00 6.23745084e-01 5.78867018e-01 8.83107066e-01
8.25837791e-01 3.88543494e-02 -9.72177833e-03 -1.11563063e+00
8.86397839e-01 1.96628660e-01 -6.36034727e-01 4.69055861e-01
-3.30570012e-01 3.62700373e-01 7.27374673e-01 -7.29465485e-01
-1.29292858e+00 7.51701891e-02 2.09212914e-01 -5.36468923e-01
-3.92543375e-02 1.08947754e-01 -4.36808020e-01 -1.55525595e-01
4.99292940e-01 4.20560539e-01 -1.42656835e-02 -5.71599901e-01
8.78713071e-01 5.76809287e-01 7.27702916e-01 -7.57204115e-01
1.09137619e+00 5.97939789e-01 -9.87775158e-03 -9.53576744e-01
-2.79855192e-01 1.88117981e-01 -9.16894674e-01 -1.92470685e-01
9.23632264e-01 -2.05026224e-01 -3.47476661e-01 5.34735739e-01
-1.33858967e+00 -4.17892992e-01 -4.87668395e-01 -9.30752084e-02
-6.41161263e-01 7.27968931e-01 -3.68203282e-01 -7.59855628e-01
-1.96728513e-01 -1.30312312e+00 1.16499722e+00 2.23752156e-01
-2.62439370e-01 -8.86090875e-01 -9.46458802e-02 7.22527504e-02
2.59400785e-01 3.05167377e-01 1.47206199e+00 1.59217432e-01
-7.89158762e-01 1.45569175e-01 -5.40233195e-01 2.31068537e-01
6.10461116e-01 2.29156256e-01 -6.07829273e-01 4.75997813e-02
-3.99203718e-01 1.10239148e-01 7.18901634e-01 2.16470584e-01
1.36541808e+00 -3.83827657e-01 -2.70597398e-01 7.42141724e-01
1.27368402e+00 5.68714917e-01 8.30186367e-01 4.21506017e-01
1.05804205e+00 6.06382012e-01 5.98350883e-01 4.16514546e-01
-1.14840187e-01 8.87402296e-01 1.35488063e-01 -5.25573134e-01
-3.80216092e-01 -5.85940003e-01 2.21690647e-02 4.03768808e-01
2.52812672e-02 -3.07215393e-01 -4.87286627e-01 2.77525663e-01
-1.69685400e+00 -7.39514291e-01 1.31593216e-02 2.14316893e+00
8.51543784e-01 9.50248018e-02 -5.43153360e-02 1.69982657e-01
1.07600510e+00 2.22271755e-01 -4.59482551e-01 -7.68082619e-01
-2.13934705e-01 2.79297769e-01 4.12898749e-01 6.22905195e-01
-8.64967227e-01 1.37342727e+00 5.68964577e+00 1.33800590e+00
-1.29054439e+00 -2.75198579e-01 7.96271145e-01 3.99196357e-01
-7.14080572e-01 -6.36102334e-02 -3.94018412e-01 5.47366142e-01
1.10271424e-01 -6.99777305e-02 8.31408203e-01 6.00180566e-01
3.44261825e-01 -2.75431305e-01 -9.13048089e-01 1.08494592e+00
2.13141695e-01 -1.39674234e+00 5.64249337e-01 -8.47539082e-02
1.02082992e+00 -1.20761859e+00 5.47465622e-01 -3.57068151e-01
-1.67928159e-01 -9.73698139e-01 1.26578331e+00 5.93068421e-01
1.31599057e+00 -8.51210415e-01 1.68811321e-01 -2.54068505e-02
-1.32132351e+00 1.87818378e-01 -1.25711814e-01 2.72432417e-01
8.47841576e-02 4.18783396e-01 -6.71628654e-01 5.00761569e-01
3.52305949e-01 4.71566409e-01 -6.86787844e-01 8.48305166e-01
-4.20379162e-01 2.81154305e-01 7.55954087e-02 -1.21260464e-01
2.28137881e-01 -7.45451212e-01 5.78677237e-01 1.35358751e+00
3.20274055e-01 5.24976104e-03 -7.06590526e-03 1.29829490e+00
-1.52858287e-01 4.92955714e-01 -6.02607846e-01 -1.89935580e-01
4.21927422e-01 1.10710132e+00 -1.27437866e+00 -4.19192225e-01
-2.76466347e-02 1.49374545e+00 7.32350275e-02 5.74060619e-01
-6.83649957e-01 -6.79046392e-01 1.81371346e-01 1.50971785e-01
4.72753018e-01 -1.53151140e-01 -1.05425513e+00 -9.16962802e-01
9.63716134e-02 -1.04041398e+00 -2.20612317e-01 -9.72816765e-01
-8.96256924e-01 6.99196160e-01 -1.40139908e-01 -1.47537482e+00
1.67832062e-01 -5.24853110e-01 -9.30888116e-01 9.67965901e-01
-1.14733112e+00 -1.15459776e+00 -3.76983374e-01 3.54519457e-01
7.98501909e-01 -1.46072373e-01 4.41906393e-01 7.12795407e-02
-3.12444061e-01 4.29583579e-01 4.71017480e-01 -1.74015611e-02
7.41227031e-01 -1.16338789e+00 6.80289507e-01 9.41942215e-01
1.07693404e-01 6.07303381e-01 6.29454613e-01 -1.09309101e+00
-1.16985416e+00 -1.10896242e+00 5.32530308e-01 -1.61095798e-01
4.35879380e-01 -5.87698996e-01 -8.38080347e-01 2.11226761e-01
3.89220059e-01 -6.04866803e-01 1.50364533e-01 -4.70892131e-01
-2.11535931e-01 6.72209486e-02 -1.03401732e+00 1.18057108e+00
9.20745015e-01 -4.06038046e-01 -5.12002587e-01 2.89739072e-02
5.03078222e-01 -5.05666375e-01 -3.37168276e-01 4.84810024e-02
6.80989921e-01 -8.01866531e-01 9.25264955e-01 -2.41638303e-01
6.50087595e-01 -5.05075932e-01 1.86727971e-01 -1.34519494e+00
-2.96270788e-01 -9.23808515e-01 2.69919693e-01 1.37555993e+00
2.53341377e-01 -2.05668971e-01 6.42631471e-01 3.26036304e-01
1.10891625e-01 -5.07916927e-01 -5.21498322e-01 -4.73677367e-01
-2.41694525e-01 -1.09067306e-01 7.27112412e-01 6.16035342e-01
-4.13727462e-01 6.03501946e-02 -5.27412295e-01 -2.56045163e-01
6.06276631e-01 4.63923097e-01 8.82842183e-01 -9.12472963e-01
-2.91160345e-01 -7.64068723e-01 -1.64703950e-02 -1.17233241e+00
-2.77437419e-02 -5.56690753e-01 8.61693546e-02 -1.42734277e+00
4.74991165e-02 -4.57477778e-01 2.41846278e-01 5.98070137e-02
-3.93186897e-01 3.98204178e-01 5.90174675e-01 1.93357497e-01
-2.35033393e-01 6.67283833e-01 1.83016467e+00 -2.93991387e-01
-4.90812510e-01 -7.21118525e-02 -6.70022666e-01 7.09744751e-01
5.90663493e-01 -1.77898288e-01 -3.15170109e-01 -2.83683538e-01
-1.36503547e-01 -1.37268298e-03 2.29387581e-01 -6.25087500e-01
-9.84829962e-02 -2.73137331e-01 5.97844362e-01 -7.24813461e-01
1.79699361e-01 -7.91225433e-01 1.16225958e-01 4.62753683e-01
-4.63919789e-01 6.17503189e-02 1.40475661e-01 6.11990988e-01
1.23442493e-01 -4.74303097e-01 1.03167105e+00 6.76586702e-02
-4.40170765e-01 -1.16099054e-02 -2.87143946e-01 1.45037798e-02
7.96433747e-01 -5.85506439e-01 -9.51378942e-02 -3.43744874e-01
-4.19413418e-01 -2.89745599e-01 9.13496196e-01 3.60870183e-01
7.69288898e-01 -1.51614165e+00 -3.40778500e-01 4.65370953e-01
2.34105275e-03 -2.58895695e-01 2.17966422e-01 5.18699825e-01
-9.33210969e-01 1.15356624e-01 -4.36676919e-01 -4.91587132e-01
-1.36894572e+00 6.49656117e-01 3.53920460e-02 -1.32447511e-01
-7.18406856e-01 4.26482856e-01 7.99614251e-01 8.44122693e-02
2.68507242e-01 -6.46938145e-01 -1.02896065e-01 -1.44052178e-01
5.14548481e-01 3.43513459e-01 -1.25547290e-01 -5.37886441e-01
1.33192971e-01 1.07492638e+00 -2.19889393e-04 -6.04847252e-01
8.43097091e-01 -3.83841187e-01 -2.22679287e-01 3.71234089e-01
8.77799869e-01 4.07316357e-01 -1.49893916e+00 -2.13491186e-01
-3.89699973e-02 -8.69362235e-01 -2.01217413e-01 -6.68152690e-01
-8.79668236e-01 8.63677740e-01 4.11399215e-01 1.22492284e-01
1.30107844e+00 -3.80290747e-01 8.00869346e-01 5.32243773e-02
1.02591597e-01 -1.18003833e+00 4.51850832e-01 3.35471928e-01
1.17331755e+00 -6.51465952e-01 -1.50315106e-01 -6.53240681e-01
-7.89350152e-01 1.22483957e+00 3.65886956e-01 -1.38157427e-01
8.56241509e-02 2.30038226e-01 5.21137007e-02 7.33814761e-02
-4.67511937e-02 -6.56903461e-02 7.04537690e-01 5.75438857e-01
1.31657735e-01 -4.09975164e-02 -3.55314672e-01 2.99307585e-01
-1.93159446e-01 -2.98449010e-01 4.45437938e-01 8.52571368e-01
-4.41480547e-01 -1.36684966e+00 -8.77667308e-01 6.81875497e-02
-6.34679943e-02 -2.85447687e-01 -7.27408230e-01 4.56063092e-01
1.89139828e-01 6.31124139e-01 -3.93442661e-02 -2.13632986e-01
5.30344069e-01 -9.94561240e-04 6.52160406e-01 -4.30911750e-01
-4.34738100e-01 5.39766788e-01 -2.34663740e-01 -3.06155235e-01
-2.81838048e-02 -4.39898223e-01 -8.77533197e-01 -1.27847984e-01
-6.12156466e-02 -1.04729421e-01 6.83694303e-01 6.85991645e-01
1.11900978e-01 6.54186845e-01 7.43049324e-01 -1.14658046e+00
3.96074466e-02 -6.25051320e-01 -6.54891193e-01 6.58081651e-01
1.61880814e-02 -6.19249582e-01 -3.18967625e-02 5.61777472e-01] | [11.593738555908203, -0.6308208107948303] |
ed09c982-8987-428e-9491-6b701e0ddb99 | evidenceminer-textual-evidence-discovery-for | null | null | https://aclanthology.org/2020.acl-demos.8 | https://aclanthology.org/2020.acl-demos.8.pdf | EVIDENCEMINER: Textual Evidence Discovery for Life Sciences | Traditional search engines for life sciences (e.g., PubMed) are designed for document retrieval and do not allow direct retrieval of specific statements. Some of these statements may serve as textual evidence that is key to tasks such as hypothesis generation and new finding validation. We present EVIDENCEMINER, a web-based system that lets users query a natural language statement and automatically retrieves textual evidence from a background corpora for life sciences. EVIDENCEMINER is constructed in a completely automated way without any human effort for training data annotation. It is supported by novel data-driven methods for distantly supervised named entity recognition and open information extraction. The entities and patterns are pre-computed and indexed offline to support fast online evidence retrieval. The annotation results are also highlighted in the original document for better visualization. EVIDENCEMINER also includes analytic functionalities such as the most frequent entity and relation summarization. EVIDENCEMINER can help scientists uncover important research issues, leading to more effective research and more in-depth quantitative analysis. The system of EVIDENCEMINER is available at https://evidenceminer.firebaseapp.com/. | ['John Caufield', 'Dibakar Sigdel', 'Enyi Jiang', 'David Liem', 'Aabhas Chauhan', 'Xuan Wang', 'Yingjun Guan', 'Weili Liu', 'Qi Li', 'Jiawei Han', 'Peipei Ping'] | 2020-07-01 | null | null | null | acl-2020-6 | ['open-information-extraction'] | ['natural-language-processing'] | [ 1.39314532e-01 2.02290162e-01 -1.14506519e+00 -2.63773315e-02
-1.00875247e+00 -8.02079082e-01 4.67222393e-01 1.47157919e+00
-4.55893397e-01 1.17938900e+00 3.60401332e-01 -7.99334526e-01
-2.48355702e-01 -7.99594581e-01 -6.64637029e-01 -4.00786489e-01
1.27608970e-01 4.92178440e-01 2.12498993e-01 1.92412153e-01
7.76473999e-01 5.66134334e-01 -1.43617785e+00 4.53738064e-01
1.13619411e+00 5.11029959e-01 3.21918458e-01 4.17894274e-01
-6.06844187e-01 5.56779444e-01 -7.56729424e-01 -3.23646843e-01
-5.19686401e-01 -2.22809985e-01 -7.86544144e-01 -7.33124375e-01
9.85787436e-03 4.64481749e-02 4.59804013e-02 1.07083416e+00
6.71692014e-01 -3.45475256e-01 5.66044152e-01 -1.09153271e+00
-3.73658270e-01 6.63076460e-01 -3.56786638e-01 3.30523342e-01
8.03389966e-01 2.43327841e-02 7.97915459e-01 -1.18967760e+00
1.30348194e+00 7.54731119e-01 1.88044325e-01 2.24810898e-01
-8.40801120e-01 -8.47779870e-01 -3.12623948e-01 4.17307377e-01
-1.32262802e+00 -3.20483983e-01 3.41144323e-01 -5.50036192e-01
1.19812405e+00 7.48293877e-01 7.76398897e-01 8.80409062e-01
4.02985871e-01 6.27129018e-01 7.50774801e-01 -5.41714072e-01
4.06238347e-01 2.00740948e-01 4.63945448e-01 8.78186524e-01
8.56543601e-01 -4.42774415e-01 -8.30462277e-01 -6.39431179e-01
1.51732072e-01 3.81091312e-02 -3.45777422e-01 3.17078710e-01
-1.38656533e+00 4.91219521e-01 1.23245694e-01 4.42437500e-01
-6.99078381e-01 -2.67287672e-01 7.73793340e-01 5.66035248e-02
5.24141133e-01 7.56403506e-01 -6.28471315e-01 5.64055815e-02
-1.15919101e+00 2.15798855e-01 9.53549504e-01 9.19158995e-01
4.96503592e-01 -7.87324846e-01 -4.00538862e-01 6.90351963e-01
3.35307419e-01 2.74269134e-01 7.87198901e-01 -6.98919773e-01
1.96374968e-01 1.16226470e+00 6.32693619e-02 -8.97070289e-01
-4.92567688e-01 -3.19810271e-01 -5.06651759e-01 -3.50979298e-01
9.35269669e-02 1.33312047e-01 -6.65814817e-01 1.18713510e+00
6.18316829e-01 5.00977878e-03 1.41612858e-01 5.55322289e-01
1.71780312e+00 5.62283933e-01 3.52921844e-01 -5.16284525e-01
1.90352571e+00 -5.77800155e-01 -1.34047198e+00 1.94203719e-01
1.00183785e+00 -8.16309690e-01 7.47044802e-01 4.89484817e-01
-1.03885245e+00 2.08536729e-01 -1.21257257e+00 -3.95971924e-01
-1.18929458e+00 2.12307841e-01 6.63078308e-01 8.39612186e-02
-6.39322579e-01 3.02135289e-01 -5.01362205e-01 -4.80930150e-01
6.63210988e-01 7.25252181e-02 -6.47588313e-01 -1.68718055e-01
-1.16318452e+00 1.02773476e+00 8.41474593e-01 -2.09136426e-01
-5.33119977e-01 -1.09590745e+00 -8.50346804e-01 4.17647697e-02
7.37470388e-01 -9.23922837e-01 1.02344275e+00 2.14073852e-01
-7.44806767e-01 1.14781773e+00 -4.19223219e-01 -3.99226904e-01
-8.12252536e-02 -1.88261792e-01 -4.66943562e-01 5.66489041e-01
5.22334874e-01 2.24350184e-01 -1.10990100e-01 -6.69128954e-01
-3.29989463e-01 -5.83393276e-01 -5.88308394e-01 -1.24401107e-01
-3.23073328e-01 3.96772265e-01 -7.00953662e-01 -7.73786128e-01
-2.03630596e-01 -4.11674768e-01 -2.67451495e-01 2.60033786e-01
-7.94655979e-01 -6.08094811e-01 4.81942534e-01 -7.09023476e-01
1.69880986e+00 -1.45038879e+00 -4.20417130e-01 2.72750199e-01
4.67137605e-01 2.55554974e-01 2.85939723e-01 7.98015475e-01
-2.84853697e-01 6.62349761e-01 -8.66999477e-02 4.99861866e-01
-3.10681820e-01 -5.70498146e-02 -1.52881205e-01 4.69756395e-01
1.93402454e-01 1.07725286e+00 -1.25674033e+00 -1.08916497e+00
-4.56416309e-02 1.18063971e-01 -1.11381978e-01 5.79454452e-02
-6.27652287e-01 1.34636074e-01 -8.30512702e-01 1.07471120e+00
3.74592662e-01 -7.25002050e-01 3.28784496e-01 -2.18830645e-01
-4.74138975e-01 6.33079588e-01 -7.40483105e-01 1.75255823e+00
-5.94724566e-02 4.40356016e-01 -1.91948265e-01 -8.22731853e-01
7.45873034e-01 5.67864835e-01 5.12281299e-01 -4.09660161e-01
-5.90381213e-02 5.06153703e-01 -4.27481622e-01 -8.68155420e-01
3.73793274e-01 1.29155234e-01 5.88807054e-02 4.15407091e-01
-8.64857733e-02 1.50448233e-01 7.19592452e-01 6.16686106e-01
1.44412982e+00 -1.12690471e-01 9.76954818e-01 -1.67801201e-01
4.09949154e-01 5.13861120e-01 4.78103697e-01 6.12532139e-01
5.46149909e-01 1.90719087e-02 2.38950759e-01 -2.21626565e-01
-5.97552419e-01 -7.06641555e-01 -5.23823142e-01 5.54417372e-01
-8.44497830e-02 -1.08588076e+00 -2.90656507e-01 -4.56429482e-01
8.72573480e-02 7.91538596e-01 -5.36655128e-01 -5.21937609e-02
4.31894697e-02 -5.95030010e-01 6.12771094e-01 8.66090730e-02
2.87253648e-01 -1.07147455e+00 -6.62943482e-01 2.28055522e-01
-4.38395470e-01 -7.05845118e-01 -1.40474990e-01 1.40820712e-01
-7.74488866e-01 -1.56851256e+00 -7.44431078e-01 -5.02076566e-01
8.13482344e-01 1.66546796e-02 9.58010554e-01 3.57625097e-01
-7.45691776e-01 8.36392399e-03 -2.63212562e-01 -9.97070014e-01
-4.41786230e-01 -1.02676719e-01 -2.45107472e-01 -8.32675040e-01
5.40499806e-01 -1.38459086e-01 -7.81669259e-01 2.54052505e-02
-1.08010519e+00 5.30104004e-02 8.03596020e-01 7.66965687e-01
1.05397034e+00 -1.26537174e-01 8.63723218e-01 -1.05157745e+00
1.05401802e+00 -9.23849821e-01 -5.85987031e-01 6.88276231e-01
-1.15239799e+00 1.07583426e-01 2.53159881e-01 -5.55549487e-02
-7.52954423e-01 -4.08458203e-01 -2.80989051e-01 8.09171051e-02
-1.30580798e-01 1.54744983e+00 -3.35477814e-02 5.37507057e-01
7.62697339e-01 1.51516885e-01 -5.60184270e-02 -5.91674268e-01
2.21514851e-01 8.18774521e-01 4.54201490e-01 -2.43073583e-01
1.37676060e-01 1.47337750e-01 1.71415851e-01 -7.12550759e-01
-7.79841959e-01 -9.34088051e-01 -5.21004722e-02 -2.30253965e-01
5.79982936e-01 -6.82927907e-01 -1.00683367e+00 -2.58895308e-01
-1.25251806e+00 1.01344183e-01 -2.50520915e-01 5.41931450e-01
-3.91224287e-02 2.51094013e-01 -3.83168846e-01 -4.42754894e-01
-8.97870541e-01 -5.94055772e-01 1.09554672e+00 3.62718403e-01
-5.50786316e-01 -7.85038650e-01 3.90959471e-01 2.32111841e-01
-1.38720244e-01 3.50145191e-01 1.11717165e+00 -1.18189502e+00
-2.67952621e-01 -6.15292549e-01 3.20432596e-02 -5.66002369e-01
2.18672693e-01 3.65043163e-01 -5.65289617e-01 2.94625342e-01
-5.64702511e-01 -2.33464673e-01 8.25620949e-01 1.83908239e-01
1.51405835e+00 -9.43321824e-01 -1.14740455e+00 7.82341883e-02
1.16847372e+00 1.60979316e-01 4.95367408e-01 5.47568738e-01
1.23518698e-01 7.65807569e-01 9.73746359e-01 3.93365145e-01
1.11477047e-01 4.74964589e-01 1.03768848e-01 -1.09043948e-01
1.39554963e-01 -2.67617494e-01 -2.95798808e-01 5.43497860e-01
1.29978597e-01 -3.69151205e-01 -1.17936361e+00 6.22104585e-01
-1.76008070e+00 -1.08306611e+00 -4.26326156e-01 2.07388544e+00
1.55252147e+00 1.15400888e-01 -9.66535583e-02 -8.40487853e-02
5.15236378e-01 -4.58467543e-01 -6.58730984e-01 -3.92031431e-01
-2.40394011e-01 3.87960285e-01 3.12691063e-01 1.27219081e-01
-5.23197591e-01 6.89748824e-01 6.14509153e+00 9.97536778e-01
-9.71048892e-01 -4.05002087e-02 5.24395823e-01 -1.26299530e-01
-4.43058223e-01 1.52424440e-01 -7.61852086e-01 5.28488457e-01
1.03743863e+00 -9.62542713e-01 -4.89724815e-01 8.17730188e-01
5.41616857e-01 -4.96321023e-01 -8.63850713e-01 6.81125700e-01
-3.77555132e-01 -2.33422256e+00 2.05005065e-01 1.87194243e-01
2.50872284e-01 6.09615371e-02 -4.52384889e-01 -1.78947628e-01
1.11464709e-01 -8.43254566e-01 2.98456877e-01 8.25370252e-01
1.03218532e+00 -3.28333229e-01 7.96052933e-01 4.04996246e-01
-6.72580540e-01 2.91118979e-01 -1.28361523e-01 4.61591572e-01
2.09315438e-02 1.36407006e+00 -1.41530704e+00 8.73066604e-01
5.80781639e-01 7.41071284e-01 -6.06360257e-01 1.46050906e+00
-4.48251337e-01 7.98457444e-01 -2.93798476e-01 -5.04951119e-01
-4.28045422e-01 1.54136553e-01 7.00590134e-01 1.65453994e+00
3.10074806e-01 2.71209031e-01 6.89929863e-03 8.19750309e-01
-3.55732381e-01 5.85281074e-01 -6.76024973e-01 -7.03237951e-01
9.01320338e-01 1.33906031e+00 -1.00924134e+00 -7.12691605e-01
-1.15493327e-01 3.71424794e-01 -2.46992428e-03 2.05283761e-02
-3.45040828e-01 -6.53107524e-01 8.85832980e-02 2.78872669e-01
-8.63503069e-02 1.72718287e-01 -5.58043122e-02 -9.39911544e-01
-1.27244875e-01 -8.29267204e-01 9.12158370e-01 -9.62863624e-01
-1.09493959e+00 1.94368646e-01 1.61756024e-01 -1.02475786e+00
-2.87554741e-01 -4.20884639e-01 -4.61861610e-01 6.70699120e-01
-1.26082420e+00 -7.34369576e-01 -2.40045749e-02 2.86158323e-01
3.64597768e-01 2.08732441e-01 1.13979828e+00 2.42530733e-01
-7.60096014e-01 2.40142748e-01 1.08892433e-01 -1.23734161e-01
9.38114107e-01 -1.16602504e+00 -2.10633688e-02 4.11444128e-01
-1.36306867e-01 1.26551962e+00 8.80063713e-01 -1.28885293e+00
-1.38797367e+00 -8.46226454e-01 1.30388379e+00 -1.91901952e-01
7.07161307e-01 -4.99033034e-02 -1.04593468e+00 1.75606191e-01
2.54635930e-01 -1.15222275e-01 1.52731049e+00 1.02603614e-01
-1.97563469e-01 2.39432126e-01 -1.11054552e+00 7.60119617e-01
7.75733948e-01 -1.17974646e-01 -6.53278589e-01 7.55767822e-01
6.41421139e-01 -2.09595412e-01 -1.10241020e+00 1.46469593e-01
6.07560039e-01 -1.76366419e-01 8.41140628e-01 -9.33393538e-01
6.22564137e-01 -4.33972150e-01 4.54099417e-01 -7.47031629e-01
6.68600202e-02 -5.38149655e-01 -5.27773678e-01 1.01058066e+00
9.77639139e-01 -5.05343437e-01 4.11997080e-01 4.95209485e-01
-2.11698607e-01 -1.05177188e+00 -7.91913211e-01 -4.56006676e-01
-3.55868995e-01 -2.78096467e-01 3.75089049e-01 1.13696694e+00
8.49685848e-01 3.95715773e-01 4.78330463e-01 -3.53559628e-02
3.93212527e-01 4.68511611e-01 4.35972035e-01 -1.40290463e+00
2.24981636e-01 -5.04163802e-01 -9.88016874e-02 -3.77286881e-01
-1.42236024e-01 -1.35825479e+00 -1.60285547e-01 -2.18887258e+00
6.12410009e-01 -1.75519541e-01 -3.42976928e-01 8.51826787e-01
-2.58806676e-01 -1.11738503e-01 -7.59242654e-01 3.48307520e-01
-7.78879642e-01 1.97219327e-01 8.69569898e-01 -2.11362749e-01
-4.96499650e-02 -3.89955848e-01 -1.07246995e+00 5.43319225e-01
8.56095970e-01 -9.39959526e-01 -4.27185930e-02 3.43533278e-01
8.82302284e-01 -1.07366599e-01 2.71458715e-01 -4.10229772e-01
6.21182859e-01 -4.46832627e-01 3.48567456e-01 -1.14010608e+00
-1.84296161e-01 -3.88378978e-01 3.01610559e-01 6.59063518e-01
-5.55254579e-01 6.26887456e-02 5.18340290e-01 4.88104701e-01
-1.28127784e-01 -5.91589153e-01 4.03477205e-03 -2.35976368e-01
-3.58267128e-01 1.61047988e-02 -6.67556107e-01 -2.22319886e-01
7.97816992e-01 -5.84900677e-02 -9.79132473e-01 -1.00747898e-01
-5.60781240e-01 5.00101924e-01 2.26771444e-01 2.33013079e-01
8.25826645e-01 -1.02818942e+00 -7.25900769e-01 -4.95471448e-01
6.44646108e-01 3.94834345e-03 -2.35545896e-02 1.06023979e+00
-4.95025843e-01 9.17021215e-01 1.02742493e-01 -2.28019834e-01
-1.59380090e+00 8.83977175e-01 -3.49114060e-01 -5.92822611e-01
-6.46822393e-01 4.77498978e-01 -1.65055975e-01 -1.86261073e-01
1.80820897e-01 -2.88241476e-01 -3.60321373e-01 1.01029530e-01
1.02228332e+00 3.62616062e-01 3.68348897e-01 7.05219507e-02
-7.24618196e-01 -9.21460018e-02 -3.53760719e-01 -9.75249335e-02
1.63856196e+00 2.33568549e-01 -6.00781560e-01 4.93291706e-01
8.88668180e-01 5.22417724e-01 1.41510163e-02 4.73685861e-02
4.96822059e-01 -1.08041264e-01 -6.02478758e-02 -1.33131313e+00
-3.67936432e-01 3.10663611e-01 9.41516906e-02 8.21812078e-02
8.37781370e-01 4.46876854e-01 3.09418499e-01 6.11389220e-01
-1.76944546e-02 -9.89648223e-01 -1.05912700e-01 3.75867635e-02
1.12471139e+00 -9.96556699e-01 6.60545886e-01 -5.90639651e-01
-2.21941597e-03 1.20039868e+00 1.31638974e-01 7.95586109e-01
5.95020413e-01 5.07793069e-01 -3.48474532e-01 -9.38526988e-01
-1.14845717e+00 1.18560903e-02 4.80122536e-01 2.40375638e-01
7.96594799e-01 -2.32086822e-01 -1.05054104e+00 1.03035152e+00
1.42958194e-01 4.16854858e-01 2.53003478e-01 1.24919605e+00
-4.43933070e-01 -1.14515591e+00 -4.46413934e-01 9.49137986e-01
-8.47043037e-01 -4.43607092e-01 -8.39877903e-01 4.62493062e-01
-8.04262683e-02 9.26972330e-01 -3.49038661e-01 1.24439813e-01
1.24219894e-01 1.66999727e-01 1.92559272e-01 -7.77020633e-01
-4.75899398e-01 1.41678929e-01 4.74481285e-01 -3.74125510e-01
-3.93539578e-01 -6.33623898e-01 -1.66406369e+00 5.91943786e-02
-6.88035786e-01 7.73874462e-01 1.05659473e+00 7.12694168e-01
8.06450844e-01 5.19124031e-01 -8.23013708e-02 7.76333436e-02
1.63108021e-01 -9.35849607e-01 7.01856846e-03 -2.29888961e-01
-1.07022494e-01 -5.48611760e-01 1.05490074e-01 2.06897154e-01] | [8.737812042236328, 8.598203659057617] |
0d88b6d8-b6fa-4c4e-9a3c-84caaac05f67 | e-t-entity-transformers-coreference-augmented | 2011.05431 | null | https://arxiv.org/abs/2011.05431v1 | https://arxiv.org/pdf/2011.05431v1.pdf | E.T.: Entity-Transformers. Coreference augmented Neural Language Model for richer mention representations via Entity-Transformer blocks | In the last decade, the field of Neural Language Modelling has witnessed enormous changes, with the development of novel models through the use of Transformer architectures. However, even these models struggle to model long sequences due to memory constraints and increasing computational complexity. Coreference annotations over the training data can provide context far beyond the modelling limitations of such language models. In this paper we present an extension over the Transformer-block architecture used in neural language models, specifically in GPT2, in order to incorporate entity annotations during training. Our model, GPT2E, extends the Transformer layers architecture of GPT2 to Entity-Transformers, an architecture designed to handle coreference information when present. To that end, we achieve richer representations for entity mentions, with insignificant training cost. We show the comparative model performance between GPT2 and GPT2E in terms of Perplexity on the CoNLL 2012 and LAMBADA datasets as well as the key differences in the entity representations and their effects in downstream tasks such as Named Entity Recognition. Furthermore, our approach can be adopted by the majority of Transformer-based language models. | ['Ioannis Vlahavas', 'Nikolaos Stylianou'] | 2020-11-10 | null | https://aclanthology.org/2020.crac-1.1 | https://aclanthology.org/2020.crac-1.1.pdf | coling-crac-2020-12 | ['lambada'] | ['natural-language-processing'] | [ 3.91965844e-02 5.46708703e-01 -1.84097379e-01 -4.52242881e-01
-5.62644839e-01 -7.03086972e-01 8.84884834e-01 -5.57571650e-03
-6.39612019e-01 7.10525811e-01 5.28450131e-01 -6.03382528e-01
3.68085206e-02 -8.03889155e-01 -6.30870044e-01 -2.16056138e-01
-1.14814624e-01 7.70157933e-01 2.60510504e-01 -4.14276034e-01
-1.27019972e-01 3.95450890e-01 -1.02234018e+00 5.92520058e-01
6.16338372e-01 7.24187434e-01 2.14191675e-01 3.01380783e-01
-5.28836370e-01 1.12152493e+00 -3.57698500e-01 -8.23785245e-01
-1.59597352e-01 5.09066768e-02 -1.06236982e+00 -6.32556498e-01
3.40488762e-01 -4.12040241e-02 -6.47170722e-01 4.84211445e-01
6.84749782e-01 2.55042583e-01 5.43053091e-01 -1.10073435e+00
-7.42892802e-01 1.32148743e+00 -1.61466852e-01 2.91039407e-01
-4.93925922e-02 -3.17270130e-01 1.29683554e+00 -1.03848004e+00
8.02328765e-01 1.53707886e+00 1.09250271e+00 8.11791003e-01
-1.28731513e+00 -6.76492572e-01 5.12140870e-01 2.41418928e-01
-1.21076477e+00 -6.27920687e-01 4.46676999e-01 -2.01845407e-01
1.94469857e+00 9.33311228e-03 2.15875864e-01 1.42982996e+00
-4.58564935e-03 1.18122923e+00 5.63066006e-01 -4.65689033e-01
-1.53933898e-01 -9.75618362e-02 3.08685511e-01 5.65303504e-01
9.88108441e-02 1.58036366e-01 -4.90810871e-01 -1.16552040e-01
6.66377366e-01 -2.99103677e-01 -1.19545884e-01 -3.55300933e-01
-1.15085900e+00 7.77105927e-01 5.70531011e-01 6.04785085e-01
-3.89850825e-01 3.65256965e-01 6.67419732e-01 1.38681546e-01
4.44041580e-01 5.20083487e-01 -8.95010114e-01 -5.49608991e-02
-9.23455536e-01 8.37578699e-02 9.49277878e-01 1.09681046e+00
3.59217614e-01 1.62415341e-01 -2.67721415e-01 9.73920286e-01
4.30790812e-01 3.54250148e-02 6.19835317e-01 -7.24393010e-01
8.06520045e-01 4.98622149e-01 -1.73462316e-01 -5.79644680e-01
-5.52267730e-01 -7.82301366e-01 -6.67839587e-01 -4.13821161e-01
2.34790325e-01 -1.29490569e-01 -8.75248075e-01 2.20026565e+00
-1.14722058e-01 1.48232207e-01 3.56208771e-01 3.69671971e-01
1.12374020e+00 5.92059255e-01 6.77429736e-01 1.81348696e-01
1.41844463e+00 -1.07722914e+00 -6.89440846e-01 -6.08030260e-01
1.16198051e+00 -4.28729802e-01 5.10944724e-01 -4.07003984e-02
-1.12024999e+00 -3.53143930e-01 -7.22497642e-01 -4.31831568e-01
-5.84362507e-01 1.38223425e-01 7.84633517e-01 4.95520592e-01
-1.25141633e+00 5.78868806e-01 -9.39390540e-01 -6.68126225e-01
2.75739968e-01 4.92034763e-01 -4.72530693e-01 1.99286625e-01
-1.65763474e+00 1.57245672e+00 8.35334659e-01 2.19642520e-01
-6.45753324e-01 -8.69306982e-01 -1.20774686e+00 4.14251953e-01
1.22717492e-01 -8.65176201e-01 1.60535538e+00 -6.12164378e-01
-1.08165419e+00 9.70932603e-01 -3.57482910e-01 -8.91596138e-01
2.52100527e-01 -3.37907255e-01 -3.49428713e-01 -3.01915944e-01
-2.64069382e-02 9.79398608e-01 -2.17404112e-01 -9.95247841e-01
-5.37867308e-01 -1.62336320e-01 1.12772517e-01 2.20748052e-01
-1.43881515e-01 2.89431304e-01 -4.31461304e-01 -6.58802032e-01
-2.17875555e-01 -8.66614640e-01 -3.98027807e-01 -4.10679311e-01
-3.45183700e-01 -5.51652789e-01 5.16140044e-01 -6.24395669e-01
1.44222522e+00 -1.93075967e+00 2.12587520e-01 -1.82063684e-01
6.81183636e-02 4.98277247e-01 -2.84850836e-01 7.31061637e-01
-3.71823072e-01 3.65860611e-01 -2.90132761e-01 -6.40399396e-01
2.36582622e-01 5.53211808e-01 -5.34550190e-01 -6.56858310e-02
4.76429135e-01 1.32818735e+00 -8.45520675e-01 -3.80139887e-01
-1.26963511e-01 6.67468488e-01 -5.83304465e-01 -9.18744318e-03
-1.93832248e-01 1.12741306e-01 -2.77413189e-01 2.82295734e-01
2.91007042e-01 -2.80902147e-01 5.33155739e-01 -1.31293580e-01
-1.77813262e-01 1.11794436e+00 -7.32303560e-01 1.71486962e+00
-6.74497068e-01 7.21752644e-01 3.89994122e-02 -9.20829654e-01
9.02572572e-01 7.80308545e-01 2.01523930e-01 -7.97951996e-01
-1.58790182e-02 3.77249569e-01 1.40138984e-01 -2.48706490e-01
6.17352486e-01 -2.25240126e-01 -1.83805153e-01 4.30692494e-01
5.03877223e-01 4.01602417e-01 2.28590339e-01 2.43992016e-01
9.29901242e-01 4.23963517e-01 3.95153254e-01 -1.84982657e-01
3.36167753e-01 -6.76477328e-02 5.72286487e-01 6.52991652e-01
4.46348377e-02 2.10621700e-01 3.42414916e-01 -4.13046837e-01
-1.00586426e+00 -7.27000713e-01 -1.93780825e-01 1.36075723e+00
-4.42241520e-01 -6.36893690e-01 -4.98021066e-01 -7.63727486e-01
-6.89586923e-02 1.14678681e+00 -5.68496585e-01 -2.99117118e-01
-1.16959405e+00 -7.23449409e-01 1.24996293e+00 9.81881261e-01
4.41367626e-01 -1.33213449e+00 -4.99385864e-01 5.54941475e-01
-4.47668105e-01 -1.17995727e+00 -2.37961262e-01 7.04239845e-01
-1.06724465e+00 -8.00218940e-01 -6.18012130e-01 -1.01507461e+00
1.87745064e-01 -3.46986771e-01 1.62599671e+00 -1.30361885e-01
2.36714378e-01 2.07157493e-01 -1.75228044e-01 -4.60169226e-01
-3.43356341e-01 5.99755943e-01 -1.39613509e-01 -4.81627136e-01
6.77006245e-01 -6.17504179e-01 -2.12324455e-01 9.24606435e-03
-6.72554433e-01 -3.33738253e-02 6.67746663e-01 1.01041460e+00
2.00511351e-01 -4.14771289e-01 6.47683144e-01 -1.07416785e+00
5.95848083e-01 -7.23808825e-01 -1.63222313e-01 3.76910001e-01
-6.06234014e-01 4.23191220e-01 3.12329799e-01 -3.88346940e-01
-1.32144713e+00 -3.06068093e-01 -4.84040976e-01 -2.78750986e-01
-4.93104458e-02 8.86210144e-01 -1.96823463e-01 2.62659311e-01
5.02344370e-01 2.46583492e-01 -4.92916912e-01 -8.47026825e-01
5.16029179e-01 4.13177431e-01 5.80162883e-01 -7.21365750e-01
3.22722226e-01 -6.45449907e-02 -2.11010888e-01 -5.06272256e-01
-8.95982325e-01 -2.26076767e-01 -8.52234542e-01 4.69356120e-01
6.83898509e-01 -9.66569066e-01 -3.66215289e-01 3.05796146e-01
-1.55846727e+00 -3.40497017e-01 -1.99289590e-01 4.97465879e-01
-4.19286579e-01 2.64204592e-01 -9.96576309e-01 -7.27559924e-01
-5.66539466e-01 -9.62962747e-01 9.87677634e-01 -8.36246274e-03
-3.54528844e-01 -1.47023213e+00 1.84401348e-01 4.02881168e-02
5.96122622e-01 -1.58432461e-02 1.31706679e+00 -1.32999015e+00
-3.17917258e-01 8.24382082e-02 -1.11016259e-01 4.21650596e-02
-3.44656140e-01 -4.46488917e-01 -1.09086704e+00 -2.13604182e-01
-3.68353099e-01 -1.76109895e-01 1.06826961e+00 1.02961265e-01
5.93961656e-01 -3.55724305e-01 -7.99037635e-01 6.85373783e-01
1.21039939e+00 2.22921297e-01 5.98319650e-01 7.11692810e-01
5.63766897e-01 5.87658346e-01 2.46526957e-01 -1.11185327e-01
6.80223465e-01 7.09743917e-01 4.17441994e-01 -1.37520647e-02
-1.93602726e-01 -6.67207420e-01 2.48981163e-01 8.60996842e-01
3.50997853e-03 -5.76317847e-01 -1.21937060e+00 8.85828197e-01
-1.86184204e+00 -1.05835795e+00 -1.26723185e-01 1.86844611e+00
9.23986316e-01 6.28302693e-02 -3.06017131e-01 -1.84951887e-01
5.74113250e-01 1.28985867e-01 -3.59839410e-01 -6.07844532e-01
-3.31607252e-01 3.28564435e-01 4.55491483e-01 3.71237755e-01
-1.09015012e+00 1.22694528e+00 7.04234219e+00 4.73800659e-01
-9.93085086e-01 1.87881708e-01 2.45763466e-01 7.95321986e-02
-2.71271288e-01 1.38219863e-01 -1.16436684e+00 1.90846100e-01
1.44286239e+00 -1.05847768e-01 -3.21311168e-02 6.05535150e-01
-1.61683083e-01 5.46517611e-01 -1.44431877e+00 7.01564252e-01
-4.09866050e-02 -1.27469122e+00 2.04449072e-01 -4.88634631e-02
3.07420403e-01 5.21901429e-01 -1.43974513e-01 9.92290080e-01
6.42327607e-01 -1.17287481e+00 6.78143501e-01 3.39206725e-01
6.06103480e-01 -6.21667922e-01 8.77044022e-01 5.20913959e-01
-1.20036888e+00 -2.33077615e-01 -3.04717779e-01 7.20572993e-02
4.63901937e-01 8.34425166e-02 -1.05521035e+00 6.06098175e-01
6.04140639e-01 6.98566973e-01 -4.42090303e-01 9.98788834e-01
-3.76996249e-01 6.94923103e-01 -3.24513197e-01 2.21288368e-01
5.39783657e-01 1.97608158e-01 4.09445018e-01 1.77571952e+00
3.13581854e-01 -4.52249907e-02 -5.52407689e-02 7.87102580e-01
-3.90152276e-01 7.04208063e-03 -7.49720275e-01 -1.37994245e-01
8.63994896e-01 9.64208543e-01 -1.32087857e-01 -4.39087540e-01
-5.10930419e-01 7.56396174e-01 8.14459443e-01 3.61457199e-01
-7.06213474e-01 -3.12044352e-01 6.55979693e-01 -1.64449625e-02
6.14375532e-01 -2.56058037e-01 -1.79165497e-03 -1.26138473e+00
-1.97047725e-01 -8.40242386e-01 5.70077777e-01 -7.41999626e-01
-1.28422749e+00 9.60798264e-01 1.65762693e-01 -6.26458347e-01
-6.31540120e-01 -8.55974019e-01 -6.20123863e-01 1.10039508e+00
-1.63130450e+00 -1.59050393e+00 3.93883675e-01 4.21937674e-01
4.53326076e-01 -1.38647109e-01 1.10044706e+00 5.80209911e-01
-4.82249469e-01 7.98887372e-01 6.96018189e-02 6.39779210e-01
6.87994838e-01 -1.19920731e+00 1.02993238e+00 7.85205007e-01
2.58821636e-01 1.14784789e+00 4.85998303e-01 -5.14815927e-01
-7.97617376e-01 -1.18474853e+00 1.85545480e+00 -7.54761696e-01
6.93812072e-01 -5.45468688e-01 -1.02582586e+00 1.50842774e+00
4.32653099e-01 -1.32609054e-01 7.71567762e-01 5.53826809e-01
-7.93435156e-01 4.98305827e-01 -8.18345606e-01 5.24369180e-01
1.37423968e+00 -6.96383834e-01 -1.17760766e+00 -5.48610874e-02
9.17382360e-01 -4.29626673e-01 -1.03426862e+00 6.40090883e-01
4.28136617e-01 -6.10455155e-01 1.04871309e+00 -1.17544281e+00
8.03192034e-02 2.20424626e-02 -2.72282243e-01 -1.34858012e+00
-5.76999307e-01 -3.25109541e-01 -3.95225912e-01 1.61206913e+00
8.63783062e-01 -6.59354866e-01 4.84516472e-01 6.27196848e-01
-3.56891304e-01 -7.31555283e-01 -9.46403146e-01 -6.92975461e-01
5.96781552e-01 -4.22926664e-01 5.94925880e-01 1.08010471e+00
6.57087937e-02 7.37398267e-01 -1.54840335e-01 -1.56639423e-02
2.43498251e-01 -1.34029403e-01 1.35320380e-01 -1.45676363e+00
-2.71295249e-01 -5.85373044e-01 -2.21518174e-01 -1.35603070e+00
5.14315128e-01 -1.29946017e+00 -1.08116351e-01 -1.74979210e+00
1.42459780e-01 -6.06119633e-01 -3.81862640e-01 9.32445467e-01
-2.64998414e-02 2.77023874e-02 4.22684014e-01 2.82609016e-01
-3.94857675e-01 5.05199134e-01 7.06616521e-01 -7.24112317e-02
1.01567470e-01 -2.33363420e-01 -7.71644056e-01 6.80134952e-01
5.82307398e-01 -4.19903219e-01 -2.40877897e-01 -1.08301365e+00
4.34028089e-01 -5.00011370e-02 1.91544533e-01 -6.76173151e-01
3.58638704e-01 3.70842397e-01 3.20745051e-01 -5.57611406e-01
4.11386818e-01 -7.63050497e-01 1.20433398e-01 2.14599773e-01
-6.65711284e-01 3.57768625e-01 4.87725824e-01 3.43067706e-01
-4.20749515e-01 -2.14948446e-01 4.17404383e-01 -4.39401478e-01
-8.87369156e-01 1.05611622e-01 -4.46189821e-01 1.44779280e-01
6.03844345e-01 3.90474009e-03 -4.87892091e-01 -6.57482892e-02
-1.03413701e+00 4.45468217e-01 1.44426614e-01 7.60311663e-01
2.38999233e-01 -1.32847762e+00 -7.42662728e-01 2.55849026e-02
7.52980858e-02 -2.29626894e-04 -2.38188058e-02 9.19417083e-01
-2.57863067e-02 1.10268044e+00 -9.94592756e-02 -3.36342543e-01
-1.12663352e+00 5.57054102e-01 7.06988811e-01 -8.63473475e-01
-4.71708775e-01 8.84284854e-01 5.63523412e-01 -1.06795323e+00
2.50705123e-01 -3.84465158e-01 -5.53562105e-01 1.93883806e-01
3.29734266e-01 9.51679870e-02 3.06690067e-01 -8.30858767e-01
-5.31657875e-01 1.96484432e-01 -3.04407656e-01 -5.21509871e-02
1.47835267e+00 -5.36204688e-02 -3.44634801e-03 3.50633621e-01
1.00994647e+00 -2.09329098e-01 -7.62784421e-01 -5.30998826e-01
6.85303271e-01 3.52117270e-01 -1.53767550e-02 -1.11721516e+00
-8.67370009e-01 1.15925193e+00 2.18083888e-01 -1.92909002e-01
7.33581603e-01 1.10964216e-01 8.02744746e-01 6.69866860e-01
4.37311709e-01 -5.45784831e-01 -6.17181778e-01 1.20701134e+00
6.57758057e-01 -8.29181492e-01 -6.14437103e-01 -9.48809162e-02
-5.56119084e-01 1.01078010e+00 6.94758713e-01 5.79147078e-02
4.38779414e-01 4.31754798e-01 2.34618951e-02 -2.26327047e-01
-1.33119643e+00 -1.06657587e-01 1.64069653e-01 5.63694775e-01
1.12579548e+00 -2.58200556e-01 -2.19324902e-01 8.04828644e-01
-2.08972424e-01 -6.32239953e-02 1.62333414e-01 9.09749448e-01
-2.34014377e-01 -1.33985174e+00 3.11980303e-02 7.48477504e-02
-8.16371024e-01 -7.04181552e-01 -4.62015778e-01 9.62393463e-01
6.08768985e-02 7.27608979e-01 1.44769147e-01 -9.94893759e-02
6.25028431e-01 6.86903596e-01 2.22129539e-01 -7.56802261e-01
-1.07030427e+00 -2.76925415e-01 6.66668355e-01 -4.52258587e-01
-4.62365806e-01 -4.47938710e-01 -1.34075463e+00 -1.03831992e-01
-3.14790100e-01 4.69221801e-01 4.92287308e-01 9.06022608e-01
6.04750574e-01 6.30598664e-01 -1.96336269e-01 -7.36871362e-01
-6.05432034e-01 -1.28123820e+00 -1.94445357e-01 1.45688534e-01
2.24445313e-01 -6.51869178e-01 -1.63668897e-02 -1.66942507e-01] | [10.05649471282959, 9.3326416015625] |
9862cec6-a9dd-4a57-89f9-5639e16dd104 | learning-large-scale-network-embedding-from | 2112.01442 | null | https://arxiv.org/abs/2112.01442v1 | https://arxiv.org/pdf/2112.01442v1.pdf | Learning Large-scale Network Embedding from Representative Subgraph | We study the problem of large-scale network embedding, which aims to learn low-dimensional latent representations for network mining applications. Recent research in the field of network embedding has led to significant progress such as DeepWalk, LINE, NetMF, NetSMF. However, the huge size of many real-world networks makes it computationally expensive to learn network embedding from the entire network. In this work, we present a novel network embedding method called "NES", which learns network embedding from a small representative subgraph. NES leverages theories from graph sampling to efficiently construct representative subgraph with smaller size which can be used to make inferences about the full network, enabling significantly improved efficiency in embedding learning. Then, NES computes the network embedding from this representative subgraph, efficiently. Compared with well-known methods, extensive experiments on networks of various scales and types demonstrate that NES achieves comparable performance and significant efficiency superiority. | ['Shengyu Zhang', 'Jinhui Zhu', 'Yi Cai', 'Hsieh', 'Chang-Yu', 'Jiezhong Qiu', 'Ben Liao', 'Weizhao Li', 'Junsheng Kong'] | 2021-12-02 | null | null | null | null | ['graph-sampling', 'network-embedding'] | ['graphs', 'methodology'] | [-1.35747612e-01 4.61243749e-01 -6.94012284e-01 -1.27993986e-01
-8.50379542e-02 -3.68351579e-01 3.95226598e-01 2.50517726e-01
3.60120609e-02 4.78445828e-01 3.41661870e-01 -3.40117484e-01
-5.54712713e-01 -1.35138297e+00 -3.04656178e-01 -6.01857185e-01
-6.11718774e-01 5.08607805e-01 1.15189046e-01 1.81446925e-01
3.41938138e-02 5.53789794e-01 -8.87417912e-01 -1.14343219e-01
3.70140254e-01 6.48691237e-01 -4.11355168e-01 6.24650121e-01
-1.88035250e-01 6.83812499e-01 -2.58704722e-01 -4.41194832e-01
2.22365111e-01 -1.90843999e-01 -8.83010328e-01 6.37039691e-02
1.77023187e-01 -5.93642473e-01 -1.26264966e+00 1.08714247e+00
3.79755348e-01 -4.78915311e-02 6.66287303e-01 -1.88197255e+00
-5.79259574e-01 9.48705792e-01 -5.30843258e-01 1.50436670e-01
2.59092212e-01 -4.97236252e-02 1.48328555e+00 -6.28570497e-01
6.31094337e-01 1.50342369e+00 6.52237773e-01 3.03223163e-01
-1.68134189e+00 -8.26655746e-01 1.88280508e-01 -5.22479601e-03
-1.29460251e+00 -1.06329367e-01 1.06330466e+00 -1.64278492e-01
7.47456610e-01 1.31739408e-01 9.33620691e-01 8.93151999e-01
-1.64733693e-01 7.17382908e-01 7.50165701e-01 -2.68032104e-01
1.08656861e-01 1.97766557e-01 1.74981356e-01 9.44186389e-01
5.50399363e-01 -8.39567930e-02 -1.91155240e-01 -5.58117032e-01
9.39418256e-01 4.89698678e-01 -1.79572254e-01 -7.31134534e-01
-1.31892204e+00 1.21997821e+00 9.25451517e-01 1.86076865e-01
-2.77595907e-01 5.74484646e-01 6.86585724e-01 6.90398395e-01
7.24010646e-01 1.80899769e-01 -2.74241537e-01 5.88557683e-02
-5.61922193e-01 7.44082266e-03 1.26790452e+00 9.62696016e-01
9.53854382e-01 -1.14509329e-01 1.69431284e-01 7.23493278e-01
3.92292529e-01 6.28561229e-02 1.01429008e-01 -8.11989009e-01
7.18882263e-01 1.08796322e+00 -5.61169326e-01 -1.55126321e+00
-1.80432305e-01 -3.76559347e-02 -1.12621963e+00 -1.93770394e-01
1.80769861e-02 -1.53967053e-01 -2.55962402e-01 1.47366226e+00
4.37842309e-01 6.34115994e-01 1.70440599e-02 4.41766351e-01
6.62476480e-01 6.80780470e-01 -1.76901430e-01 1.98145192e-02
1.11095309e+00 -1.16172290e+00 -4.17056203e-01 -3.14449780e-02
9.12329793e-01 -7.45690018e-02 7.17844248e-01 -7.31200278e-02
-5.23684859e-01 -1.61100924e-01 -1.04611993e+00 1.07800908e-01
-5.48732579e-01 -2.04719841e-01 1.27181935e+00 5.05980015e-01
-1.03939056e+00 8.52677047e-01 -6.93795860e-01 -5.46816945e-01
6.72153115e-01 4.86776948e-01 -7.47397542e-01 -4.42719042e-01
-1.13651001e+00 2.29154035e-01 8.13965857e-01 -5.23704477e-02
-8.24691236e-01 -7.67431796e-01 -1.15287971e+00 4.25287127e-01
6.15474641e-01 -6.33867502e-01 5.11644125e-01 -5.31946838e-01
-1.31521618e+00 4.52144712e-01 -3.32845859e-02 -5.69558442e-01
1.68803111e-01 7.42640626e-03 -5.01688242e-01 3.76319468e-01
-7.24877343e-02 4.79559869e-01 7.17526853e-01 -8.29494417e-01
-1.00015618e-01 -2.28411332e-01 5.45798540e-01 -1.81728214e-01
-9.83348250e-01 -2.46445358e-01 -4.24833924e-01 -4.97034281e-01
2.17123747e-01 -9.31505561e-01 -5.05073845e-01 3.36967021e-01
-7.91221619e-01 -5.26608288e-01 1.12375200e+00 -2.01568455e-01
1.49173617e+00 -2.03686309e+00 2.52953708e-01 5.59688032e-01
1.14143491e+00 2.07040552e-02 -4.19488132e-01 9.97713029e-01
-3.46104205e-01 3.97123009e-01 4.68049385e-02 -7.29606226e-02
1.19087420e-01 3.70005131e-01 -1.48891598e-01 5.33090711e-01
1.48524195e-01 1.06876838e+00 -1.17754316e+00 -5.94513953e-01
2.55427986e-01 4.01238859e-01 -7.18066931e-01 2.06947222e-01
4.43606637e-02 -2.07045823e-01 -5.38796544e-01 4.70832646e-01
3.68483722e-01 -6.98302984e-01 6.57269895e-01 -3.48815709e-01
4.78468746e-01 2.18120784e-01 -1.11498952e+00 1.36610627e+00
-3.73064250e-01 8.98616850e-01 -1.25800699e-01 -1.49752116e+00
9.01232064e-01 3.45575362e-01 5.85975170e-01 -1.39732495e-01
-1.06752142e-02 -1.79642141e-01 -8.87540728e-02 -3.23465079e-01
7.57090673e-02 1.48532078e-01 9.26597565e-02 9.21558797e-01
5.90913892e-02 3.36781740e-01 3.18953961e-01 8.10288310e-01
1.43013406e+00 -6.52070522e-01 4.43257987e-01 -1.37883127e-01
3.69560927e-01 -3.67876589e-01 3.56509238e-01 3.33866179e-01
-1.03085347e-01 -1.02778681e-01 1.20239067e+00 -5.15259624e-01
-1.00876188e+00 -1.28007603e+00 1.65647328e-01 6.54039860e-01
-2.15582382e-02 -1.04056454e+00 -4.15606976e-01 -1.15760195e+00
4.73861188e-01 -2.42050495e-02 -5.84925592e-01 -4.59411889e-01
-4.15170282e-01 -5.55785060e-01 3.63031060e-01 5.87177992e-01
2.43819341e-01 -7.24277914e-01 5.43895364e-01 2.61593640e-01
1.46863833e-01 -1.04511654e+00 -4.94028896e-01 -2.67257482e-01
-1.32225311e+00 -1.41378915e+00 -2.96326607e-01 -9.50890779e-01
1.13227355e+00 4.90977973e-01 1.02672446e+00 3.85516018e-01
-4.79113132e-01 3.24343592e-01 -1.44755781e-01 2.15012863e-01
-2.11332485e-01 3.28607976e-01 2.22165599e-01 -4.75087427e-02
5.35220683e-01 -1.05790150e+00 -5.27031541e-01 6.67016134e-02
-1.01051438e+00 -1.76497146e-01 7.90180385e-01 8.29873979e-01
3.81241858e-01 6.49168491e-01 6.60420656e-01 -1.44536912e+00
8.85688424e-01 -9.16856825e-01 -5.00738442e-01 1.10995568e-01
-9.76048172e-01 2.20792562e-01 6.05085671e-01 -4.95900214e-01
-1.78457394e-01 -4.83365119e-01 2.74136603e-01 -6.06690228e-01
2.25507572e-01 7.48914599e-01 -1.62873328e-01 -1.44218832e-01
1.97307125e-01 1.20369911e-01 3.07294875e-01 -4.30004597e-01
6.65662885e-01 5.96101701e-01 -1.16612308e-01 -3.00569415e-01
1.32037580e+00 5.92443824e-01 2.57086575e-01 -8.44673336e-01
-7.80206144e-01 -5.15784800e-01 -7.62564003e-01 2.24359423e-01
2.88934827e-01 -8.32299531e-01 -9.76609886e-01 -2.40824342e-01
-7.97119319e-01 -3.23707005e-03 -2.76695341e-01 5.82540691e-01
-2.58032560e-01 5.86109519e-01 -8.91658008e-01 -4.53789145e-01
-4.35558230e-01 -6.28060460e-01 6.30822003e-01 -7.03218356e-02
-2.51907736e-01 -1.73632622e+00 3.23643267e-01 4.30974700e-02
2.75562912e-01 4.28536892e-01 1.19385958e+00 -6.22984409e-01
-7.90600777e-01 -5.98991215e-01 -6.73614502e-01 4.30626810e-01
3.04826617e-01 1.85264245e-01 -4.72755134e-01 -7.16657102e-01
-7.13338971e-01 -2.37111747e-01 7.73912191e-01 8.33722204e-02
1.25688565e+00 -6.67965710e-01 -7.51030147e-01 7.48496473e-01
1.62854719e+00 -5.57909906e-01 4.99847919e-01 1.21721901e-01
9.34097230e-01 5.02951026e-01 2.56474912e-01 3.50338012e-01
3.28872234e-01 2.60392159e-01 3.55880201e-01 -1.21361323e-01
3.74165699e-02 -7.11458266e-01 2.84664184e-01 1.33135784e+00
1.74416482e-01 -4.00068201e-02 -6.89449847e-01 5.68551183e-01
-1.86158514e+00 -9.14429545e-01 1.93930477e-01 1.74793732e+00
5.09349167e-01 3.98318559e-01 1.90073833e-01 3.47097844e-01
7.82063127e-01 5.50608635e-01 -5.68855166e-01 -2.18223780e-01
2.46430084e-01 4.33457680e-02 4.94356126e-01 2.79657215e-01
-9.89767432e-01 8.98221254e-01 6.90197229e+00 7.71839142e-01
-7.33328104e-01 -1.11691177e-01 2.25820959e-01 3.54069509e-02
-4.93012786e-01 5.42140901e-01 -6.21642828e-01 3.59902650e-01
1.11542988e+00 -6.57140732e-01 4.98375863e-01 1.13358402e+00
1.10497354e-02 5.67052782e-01 -1.35302138e+00 9.69726205e-01
-1.30406126e-01 -1.52635908e+00 3.88327777e-01 4.68944520e-01
7.29824126e-01 3.01774349e-02 -1.45035401e-01 3.26161504e-01
5.03720880e-01 -9.57167745e-01 -3.85257453e-01 2.73180008e-01
9.64382946e-01 -1.05927515e+00 5.86921751e-01 1.47620142e-01
-1.70394182e+00 -1.03116885e-01 -8.47065032e-01 -1.50325134e-01
-4.90764976e-02 9.28054631e-01 -1.03825498e+00 5.33709109e-01
3.84091556e-01 1.46434832e+00 -4.28004682e-01 8.01663399e-01
-4.01004702e-01 7.56317735e-01 -2.15572998e-01 -1.19838081e-02
3.39991897e-01 -5.36291301e-01 5.79223394e-01 8.34938407e-01
-6.86671138e-02 -2.32666716e-01 3.91442537e-01 9.17515576e-01
-6.60305023e-01 4.75668274e-02 -1.14532650e+00 -8.90246391e-01
8.36642146e-01 1.57173443e+00 -7.03426480e-01 -3.66702765e-01
-5.26332259e-01 8.24081242e-01 7.12919772e-01 3.63894641e-01
-6.43325746e-01 -8.52572143e-01 9.99440253e-01 1.96135074e-01
1.80581957e-01 -2.99081981e-01 4.79246438e-01 -1.19649971e+00
-1.79554239e-01 -4.82126623e-01 5.32263696e-01 -2.19420880e-01
-1.60365081e+00 4.04573232e-01 -1.55035079e-01 -1.20087504e+00
-2.74436213e-02 -5.81368029e-01 -1.00890672e+00 4.84896094e-01
-1.57321608e+00 -1.05963302e+00 -1.92652747e-01 5.80138385e-01
1.42036611e-02 -2.35812098e-01 9.44536150e-01 3.88421535e-01
-8.43901396e-01 6.43137634e-01 4.52147007e-01 7.56142676e-01
2.98421055e-01 -1.30539143e+00 5.66519141e-01 3.20800960e-01
3.76812756e-01 7.81662524e-01 6.96164891e-02 -5.21402419e-01
-1.60965943e+00 -1.23872471e+00 7.28388488e-01 -7.70326927e-02
1.11544442e+00 -4.90640610e-01 -7.26380229e-01 9.94633794e-01
-1.95284307e-01 4.72849965e-01 9.41346407e-01 5.00977635e-01
-6.38856590e-01 -4.23077911e-01 -1.00730944e+00 7.57615685e-01
1.22374725e+00 -9.81248617e-01 -3.15477848e-01 5.51106870e-01
1.16290939e+00 3.32943946e-01 -1.38483000e+00 3.47987711e-02
5.08770347e-01 -4.96683508e-01 1.29317749e+00 -1.00248516e+00
3.57890815e-01 -6.89658709e-03 2.00788245e-01 -1.25484312e+00
-4.97044384e-01 -6.01924539e-01 -7.28127718e-01 1.29244459e+00
1.02163740e-01 -1.02111244e+00 1.30457866e+00 4.02157933e-01
6.39159918e-01 -9.62232113e-01 -6.87069476e-01 -9.01061952e-01
-2.73104638e-01 -1.09619617e-01 8.71163785e-01 1.22186387e+00
2.52412528e-01 4.86183852e-01 -3.41173083e-01 1.65617093e-01
9.88689065e-01 3.23489040e-01 9.95093882e-01 -1.67294741e+00
-2.04836741e-01 -2.84494311e-01 -9.95747805e-01 -9.70126152e-01
5.74418962e-01 -1.34122205e+00 -9.72385168e-01 -1.74360788e+00
4.43357259e-01 -5.55282831e-01 -4.06732857e-01 3.84412378e-01
9.16143432e-02 1.43498287e-01 -8.79768357e-02 1.53295159e-01
-7.31716752e-01 7.55701780e-01 1.14944625e+00 -3.62044990e-01
3.20336670e-02 -4.28589582e-02 -7.75684416e-01 6.21001184e-01
7.29359269e-01 -7.19121456e-01 -6.72146440e-01 -1.27219427e-02
2.56084770e-01 -1.06546246e-02 2.28289247e-01 -7.41295278e-01
1.99858531e-01 1.04801096e-01 2.28753790e-01 -4.65016425e-01
2.65641242e-01 -1.07580233e+00 -4.12389971e-02 6.75475538e-01
-3.57666612e-01 4.78167944e-02 -2.72053689e-01 1.17115939e+00
-3.48429739e-01 -9.25963074e-02 3.69913489e-01 6.26439452e-02
-5.18532753e-01 1.05160391e+00 1.66947767e-01 1.03729673e-01
1.20185459e+00 -1.50323972e-01 -2.97683716e-01 -4.14037496e-01
-7.02338278e-01 4.39463764e-01 3.11207592e-01 4.06301707e-01
8.67706954e-01 -1.84137928e+00 -5.31688333e-01 1.13301352e-01
2.36100703e-01 -1.04257137e-01 6.40415074e-03 7.88488686e-01
-4.87609029e-01 3.63545567e-01 7.20298588e-02 -3.46989542e-01
-1.33279347e+00 7.48541176e-01 -5.71974218e-02 -4.98876691e-01
-9.25639033e-01 6.02518380e-01 1.68521013e-02 -6.66474044e-01
1.82734802e-01 -4.44764178e-03 -6.10870980e-02 4.00664173e-02
5.54409206e-01 7.15262532e-01 -6.12308741e-01 -2.33457297e-01
-1.66588560e-01 3.94933075e-01 -3.09650421e-01 4.09818649e-01
1.59269524e+00 3.65780457e-03 -6.74483180e-01 3.43448102e-01
1.98316848e+00 -1.63603693e-01 -9.38221931e-01 -7.71720529e-01
1.14381060e-01 -7.50207365e-01 3.58177945e-02 2.06801683e-01
-1.41992867e+00 8.83507967e-01 8.13099518e-02 4.67195809e-01
7.13872254e-01 7.98556134e-02 1.00456011e+00 8.31825435e-01
2.91078597e-01 -7.84363627e-01 3.32235843e-01 2.48739034e-01
4.32136029e-01 -1.22084618e+00 3.21921617e-01 -5.21823287e-01
-1.13885470e-01 1.36514282e+00 5.81954718e-01 -5.07281303e-01
1.22671056e+00 -7.37243220e-02 -6.23458862e-01 -6.00172997e-01
-9.43361402e-01 1.15698345e-01 2.41015688e-01 4.95934516e-01
9.26891044e-02 1.60740614e-01 1.99396744e-01 8.24755058e-02
4.05937545e-02 -2.41428792e-01 3.06362331e-01 6.92446947e-01
-3.82110775e-01 -1.35350883e+00 3.24695975e-01 1.04028893e+00
-1.33887798e-01 -1.55568257e-01 -3.42547089e-01 8.97339046e-01
-4.90066200e-01 5.30203998e-01 4.77912389e-02 -4.76980299e-01
6.41951486e-02 -1.98199883e-01 1.68572545e-01 -8.41239929e-01
-6.02836497e-02 -4.63543862e-01 6.32481277e-02 -7.76060462e-01
-3.15325379e-01 -1.39129251e-01 -1.02014410e+00 -9.46103096e-01
-5.15837789e-01 4.02974606e-01 3.04755807e-01 6.00587785e-01
4.68503863e-01 4.00351971e-01 9.40075696e-01 -5.95497549e-01
-5.08210301e-01 -8.16153884e-01 -1.02182043e+00 4.09659505e-01
2.30272308e-01 -5.51809072e-01 -7.13558853e-01 -4.50189292e-01] | [7.170361518859863, 6.20354700088501] |
e96c1378-7e59-435c-8770-1026e948f8a2 | joint-image-compression-and-denoising-via | 2205.01874 | null | https://arxiv.org/abs/2205.01874v2 | https://arxiv.org/pdf/2205.01874v2.pdf | Joint Image Compression and Denoising via Latent-Space Scalability | When it comes to image compression in digital cameras, denoising is traditionally performed prior to compression. However, there are applications where image noise may be necessary to demonstrate the trustworthiness of the image, such as court evidence and image forensics. This means that noise itself needs to be coded, in addition to the clean image itself. In this paper, we present a learning-based image compression framework where image denoising and compression are performed jointly. The latent space of the image codec is organized in a scalable manner such that the clean image can be decoded from a subset of the latent space (the base layer), while the noisy image is decoded from the full latent space at a higher rate. Using a subset of the latent space for the denoised image allows denoising to be carried out at a lower rate. Besides providing a scalable representation of the noisy input image, performing denoising jointly with compression makes intuitive sense because noise is hard to compress; hence, compressibility is one of the criteria that may help distinguish noise from the signal. The proposed codec is compared against established compression and denoising benchmarks, and the experiments reveal considerable bitrate savings compared to a cascade combination of a state-of-the-art codec and a state-of-the-art denoiser. | ['Ivan V. Bajić', 'Hyomin Choi', 'Mateen Ulhaq', 'Saeed Ranjbar Alvar'] | 2022-05-04 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 6.89059675e-01 -1.95030540e-01 1.10493004e-01 -1.14178836e-01
-8.38532448e-01 -2.79125392e-01 3.94309789e-01 3.79671901e-02
-3.82404625e-01 3.69755536e-01 3.47034395e-01 -2.32237041e-01
-4.04549576e-02 -8.27921391e-01 -7.53265202e-01 -1.14450026e+00
3.49577181e-02 -1.68247133e-01 -1.16434798e-01 -3.02371681e-02
1.33653715e-01 1.22467071e-01 -1.51896298e+00 5.56279659e-01
7.55482197e-01 1.21310258e+00 6.13205194e-01 8.26476812e-01
-8.14467072e-02 9.53292906e-01 -5.84620953e-01 -4.71452236e-01
5.83901942e-01 -7.52955317e-01 -4.29255068e-01 4.37493861e-01
2.97397494e-01 -6.25740528e-01 -6.21933341e-01 1.59084535e+00
2.08611548e-01 -1.40090242e-01 3.21235150e-01 -7.73751438e-01
-3.60535532e-01 4.61123794e-01 -3.98930967e-01 -4.61945608e-02
3.25240076e-01 1.91321418e-01 7.90973186e-01 -3.61989230e-01
5.72154760e-01 1.24263799e+00 3.62575829e-01 3.01666588e-01
-1.42610884e+00 -6.49046123e-01 -2.40205348e-01 3.21785033e-01
-1.24380696e+00 -7.37120688e-01 9.43724990e-01 -1.15533262e-01
4.21253055e-01 3.15904260e-01 5.28768241e-01 1.08636546e+00
3.57104331e-01 5.35344362e-01 1.15181780e+00 -4.34882402e-01
5.36862016e-01 -2.57082254e-01 -3.71996850e-01 2.98858821e-01
2.40928888e-01 1.30816791e-02 -6.17391527e-01 8.13924670e-02
7.20626056e-01 1.45050764e-01 -5.85878074e-01 -1.18733205e-01
-8.20063949e-01 7.00535893e-01 4.59436208e-01 1.65632948e-01
-7.58505225e-01 3.09806854e-01 5.20164371e-01 4.07579005e-01
3.70993525e-01 -1.04760081e-02 2.21695840e-01 -2.80043066e-01
-1.52508163e+00 -1.38700865e-02 8.15779448e-01 7.08842635e-01
5.13658345e-01 2.32127070e-01 7.77068883e-02 6.50662184e-01
2.88261086e-01 3.24747324e-01 5.10865808e-01 -1.32786584e+00
6.45524502e-01 2.62103140e-01 -1.99475735e-01 -1.07885742e+00
5.24692833e-01 -4.29761529e-01 -1.52618456e+00 5.30387044e-01
1.32522568e-01 2.91585445e-01 -1.00024962e+00 1.40674043e+00
-1.18894711e-01 4.52277988e-01 3.41275960e-01 9.11239505e-01
3.99353087e-01 9.11969841e-01 -2.59574056e-01 -5.23493171e-01
1.20968223e+00 -5.56399465e-01 -1.00204515e+00 -3.41730773e-01
3.78397591e-02 -8.43666315e-01 6.55425131e-01 8.00406218e-01
-1.32408202e+00 -7.19530106e-01 -1.36860597e+00 -1.81630284e-01
1.25718281e-01 -2.05131650e-01 5.86694404e-02 5.44820070e-01
-9.65297639e-01 6.80986524e-01 -8.59237313e-01 5.24856672e-02
5.51225424e-01 -3.13618854e-02 -4.62919265e-01 -7.24808693e-01
-9.76492047e-01 6.25311255e-01 3.49934071e-01 1.79771096e-01
-9.89262760e-01 -1.72544509e-01 -8.98252010e-01 3.88340116e-01
2.03772768e-01 -4.35005933e-01 8.10557842e-01 -8.91485512e-01
-1.13003981e+00 5.38295150e-01 -3.05244476e-01 -8.35726082e-01
7.98458457e-01 -2.41282210e-01 -3.27392310e-01 7.17019558e-01
-9.50282440e-04 3.46249580e-01 1.47325194e+00 -1.48290956e+00
-4.64214146e-01 -1.35155529e-01 3.82251851e-03 2.50783768e-02
-2.01695591e-01 -1.81329221e-01 -7.03205347e-01 -8.00251722e-01
5.00523746e-01 -6.20083332e-01 -1.58130482e-01 1.83265507e-01
-2.46436238e-01 4.44660634e-01 1.06140339e+00 -1.10025525e+00
1.34945679e+00 -2.46567774e+00 2.77262986e-01 3.46397936e-01
4.09640729e-01 2.09855199e-01 -1.25950277e-01 3.82138133e-01
-1.94186687e-01 1.83114409e-01 -5.21996498e-01 -6.46612644e-01
-4.00924027e-01 4.33949530e-01 -3.99553299e-01 7.12953448e-01
-3.18955332e-02 5.38046241e-01 -8.07427883e-01 -5.07067084e-01
3.35187793e-01 6.95333123e-01 -4.47340488e-01 2.57520318e-01
1.61462277e-01 4.16329443e-01 -1.37364507e-01 4.58716631e-01
8.25610518e-01 -1.11492693e-01 2.35927805e-01 -4.70643252e-01
1.61132783e-01 1.84335008e-01 -1.34598291e+00 1.56917810e+00
-3.80609989e-01 7.37391829e-01 5.41522622e-01 -1.20800686e+00
7.88288176e-01 5.90966284e-01 2.82964975e-01 -7.49634385e-01
1.06765509e-01 1.27753943e-01 -1.43704005e-02 -5.44958293e-01
4.01243716e-01 -1.88691497e-01 2.18730569e-01 3.09095323e-01
1.77553371e-02 -2.33622849e-01 2.75758862e-01 2.27046758e-01
1.26280725e+00 -6.35857433e-02 2.73805231e-01 1.38527960e-01
5.83418489e-01 -4.26208138e-01 5.71372390e-01 8.47365081e-01
2.06510350e-03 9.44006324e-01 4.16630089e-01 -2.59479463e-01
-1.37837195e+00 -9.17405128e-01 -7.75435418e-02 3.94219249e-01
2.52346873e-01 -6.70043707e-01 -1.05831897e+00 -3.44237745e-01
-4.25721109e-01 6.22048497e-01 -2.99509645e-01 -2.41137862e-01
-5.71644008e-01 -2.69789845e-01 4.15048838e-01 1.00018971e-01
9.88178849e-01 -7.38606930e-01 -7.34887838e-01 2.45465919e-01
-5.78113258e-01 -1.22600961e+00 -2.84835190e-01 3.17900777e-01
-1.06903279e+00 -1.02204692e+00 -3.89854461e-01 -4.61548537e-01
7.36135960e-01 5.21116495e-01 8.31633568e-01 4.70017910e-01
-2.30789017e-02 1.75464854e-01 -5.63205659e-01 3.22451144e-02
-9.89468992e-01 -4.73649681e-01 -1.75957367e-01 2.96024263e-01
5.28156236e-02 -8.52533937e-01 -6.92263842e-01 -4.33940515e-02
-1.66318035e+00 2.08887130e-01 6.84518695e-01 1.01173663e+00
5.99963844e-01 8.75741720e-01 -7.00824857e-02 -6.45933688e-01
4.77666855e-01 -3.70186031e-01 -5.91009617e-01 -9.93170589e-03
-4.19184804e-01 1.19598582e-01 8.51904094e-01 -3.36435258e-01
-9.31160450e-01 1.29156053e-01 -1.50791556e-01 -7.08524108e-01
-6.59413785e-02 6.08142257e-01 -3.93900961e-01 7.85016939e-02
4.95191991e-01 6.17801547e-01 4.08690274e-01 -6.91586137e-01
2.59548008e-01 8.59480143e-01 8.65778804e-01 -1.28014460e-01
9.74898577e-01 6.19910955e-01 2.38729402e-01 -9.05162990e-01
-4.47530121e-01 -4.27601874e-01 -6.09423578e-01 -1.84659064e-01
7.15782166e-01 -1.12655067e+00 -4.36332077e-01 3.61270070e-01
-1.16996241e+00 8.38680789e-02 -3.43680710e-01 3.38995904e-01
-5.18245757e-01 7.88207591e-01 -7.80590773e-01 -8.11929286e-01
-1.81253284e-01 -1.61105180e+00 9.62993264e-01 -1.32906824e-01
-3.63620333e-02 -6.92889750e-01 -5.46036959e-01 5.57467937e-01
3.76498967e-01 2.36345649e-01 8.94689441e-01 -2.20821992e-01
-7.33681679e-01 -4.98582244e-01 1.13311876e-02 1.03118086e+00
1.98478967e-01 -4.32778388e-01 -9.96760964e-01 -4.94999617e-01
8.01178813e-01 -1.28423139e-01 1.26127076e+00 3.69575679e-01
1.22177386e+00 -5.90105653e-01 2.11846650e-01 8.75835955e-01
1.55705845e+00 6.16888776e-02 1.24078763e+00 4.50605720e-01
4.98722464e-01 4.38707024e-01 1.69728905e-01 3.91382635e-01
-2.64567770e-02 3.99432927e-01 7.04337001e-01 -7.33859539e-02
-3.73584181e-01 -2.23958582e-01 5.41362286e-01 9.55217540e-01
-8.32509398e-02 -3.09691429e-01 -4.98445600e-01 2.06740052e-01
-1.61370552e+00 -1.22020030e+00 -1.61575824e-02 2.36083937e+00
9.25130367e-01 2.12865412e-01 -4.68517780e-01 8.83698225e-01
7.73038208e-01 4.65387881e-01 -3.87541264e-01 -2.01088876e-01
-2.94781476e-01 1.11946791e-01 5.71531594e-01 4.66759622e-01
-1.19194651e+00 3.20501894e-01 6.05565071e+00 1.14657879e+00
-1.14175439e+00 -2.68143043e-02 7.90355921e-01 -7.85446316e-02
-9.67982709e-02 1.64638996e-01 -1.94446027e-01 8.07278752e-01
1.00101066e+00 7.75897782e-03 7.90581942e-01 6.22506320e-01
3.47551703e-01 -3.06075990e-01 -1.08729291e+00 1.29790545e+00
1.66440263e-01 -1.26521230e+00 5.18085882e-02 4.53078568e-01
6.01884067e-01 -3.25334966e-01 8.30608681e-02 -1.72917530e-01
-1.71600387e-01 -1.01145566e+00 8.94623220e-01 4.85691428e-01
8.68691564e-01 -6.73305094e-01 1.01651859e+00 6.13889217e-01
-1.00908852e+00 -2.21823364e-01 -5.78944683e-01 7.23070279e-03
3.08796614e-01 1.05221963e+00 -3.04347962e-01 6.18859053e-01
8.87161732e-01 6.50725961e-01 -4.09031957e-01 9.70960975e-01
-4.89813656e-01 9.61883426e-01 -6.19748309e-02 9.01999712e-01
1.25701040e-01 -4.39765990e-01 6.46351278e-01 1.18358612e+00
5.81062138e-01 1.25705600e-01 1.03048250e-01 6.87300444e-01
-4.66505110e-01 -2.50087410e-01 -5.71248651e-01 7.60620236e-02
3.30326796e-01 1.09892559e+00 -6.48020029e-01 -4.22575146e-01
-3.37447464e-01 1.26950121e+00 -2.71510631e-01 4.28764433e-01
-5.02035141e-01 -1.43726453e-01 4.53951061e-01 9.87013578e-02
5.04277945e-01 -1.91665202e-01 -3.60449314e-01 -1.23215544e+00
1.67317301e-01 -1.20382142e+00 8.22063461e-02 -8.01981211e-01
-1.11527801e+00 5.01431465e-01 -1.95297062e-01 -1.74470842e+00
-2.96393514e-01 -1.45289198e-01 -5.22545099e-01 9.76792336e-01
-1.71330667e+00 -8.33090007e-01 -4.11237061e-01 4.38629270e-01
6.44407630e-01 -1.54338688e-01 6.86392069e-01 2.08224207e-01
-1.60719022e-01 3.05208087e-01 4.41255033e-01 1.70190364e-01
4.27974403e-01 -9.96293187e-01 1.97851524e-01 1.38355589e+00
1.03452861e-01 6.57445431e-01 8.09956193e-01 -8.05060089e-01
-1.58863282e+00 -1.00381827e+00 5.71737468e-01 3.62820685e-01
3.43566537e-01 -3.74521911e-01 -1.18077421e+00 2.14656740e-01
3.31428945e-01 1.29203960e-01 4.17844743e-01 -5.83332896e-01
-4.36771184e-01 -4.04573262e-01 -1.29803181e+00 4.05293763e-01
5.30039907e-01 -8.58764172e-01 -4.64935899e-01 3.17833899e-03
5.93327463e-01 -3.52418542e-01 -6.26932621e-01 6.42839745e-02
3.82325590e-01 -1.13411534e+00 1.08593535e+00 1.50486052e-01
9.84387994e-01 -4.55519199e-01 -5.26064456e-01 -1.12262011e+00
-1.96288109e-01 -6.39044404e-01 -4.33799207e-01 1.23068404e+00
-2.41750717e-01 -1.22718692e-01 6.27128422e-01 4.37504411e-01
9.27410051e-02 -4.13173795e-01 -1.18901300e+00 -7.88101017e-01
-3.80013824e-01 -6.59382701e-01 3.61726642e-01 6.10964119e-01
-4.18173432e-01 -3.61576080e-02 -7.79452443e-01 2.73374736e-01
9.83499467e-01 -1.48742422e-01 5.83662868e-01 -8.03649604e-01
-3.99693191e-01 -8.83903950e-02 -6.32619083e-01 -1.28192008e+00
-1.11486100e-01 -7.72321820e-01 2.02875286e-01 -1.41801620e+00
2.52891034e-01 3.01342458e-02 -2.77756363e-01 1.64529473e-01
-6.03859313e-02 4.16790158e-01 5.55948496e-01 7.07850695e-01
-2.47176871e-01 4.78828609e-01 1.12795734e+00 -3.26112747e-01
1.87330544e-01 -2.08967194e-01 -6.07023716e-01 6.78969860e-01
4.74668115e-01 -5.72663069e-01 -3.24397326e-01 -4.35319066e-01
-9.03318971e-02 2.95889318e-01 5.34140766e-01 -1.08659375e+00
5.83660543e-01 1.32469639e-01 5.38361728e-01 -4.01234090e-01
5.67437768e-01 -1.10869610e+00 2.59502769e-01 5.28185010e-01
-3.16012621e-01 -2.68369853e-01 -1.93210691e-01 7.59026110e-01
-7.10779428e-01 -5.21453381e-01 9.22007084e-01 -2.70742893e-01
-4.10168052e-01 -1.09559260e-01 -2.35890090e-01 -3.10013711e-01
6.31937444e-01 -5.98910332e-01 -1.16863884e-01 -8.32368255e-01
-4.89938259e-01 -2.43581980e-01 5.95718145e-01 3.58569697e-02
1.03417051e+00 -1.08775258e+00 -8.15002799e-01 3.96120906e-01
-2.73650527e-01 3.47220488e-02 3.53358626e-01 5.76962292e-01
-6.21799827e-01 -2.78358366e-02 -1.29846353e-02 -6.67903185e-01
-1.36458445e+00 5.99186063e-01 -9.49214473e-02 -2.67812461e-01
-9.52559352e-01 4.77544129e-01 1.29582644e-01 4.16415930e-01
3.67637396e-01 -1.90687090e-01 1.22289155e-02 -4.65514064e-02
8.27251971e-01 3.55507493e-01 8.08046833e-02 -8.44512224e-01
1.61144659e-01 3.95977467e-01 1.68392360e-01 -1.33157909e-01
1.67294073e+00 -4.91433144e-01 -4.30536628e-01 9.26704332e-02
1.37204099e+00 8.33582692e-03 -1.55186296e+00 -3.26484859e-01
-1.66842759e-01 -9.07347083e-01 6.23549461e-01 -3.90721709e-01
-1.32662463e+00 9.02967691e-01 6.87363148e-01 3.60564560e-01
1.80240202e+00 -2.81462342e-01 8.72008562e-01 1.92117989e-01
3.15031290e-01 -1.06311321e+00 7.39096552e-02 7.30302185e-02
9.53799963e-01 -1.18375611e+00 3.90775353e-01 -2.17600614e-01
-5.49341917e-01 1.22829235e+00 -2.41676778e-01 -8.83941874e-02
5.01111448e-01 4.69136298e-01 -1.58504024e-01 6.63689896e-02
-5.57013631e-01 2.36621760e-02 1.61389649e-01 4.34089094e-01
4.38448079e-02 -2.63283134e-01 -7.11123347e-02 4.08713996e-01
-2.46782973e-01 8.24575871e-03 7.42832422e-01 9.99161661e-01
-4.13146257e-01 -1.22877264e+00 -8.02067637e-01 4.60894763e-01
-7.46414900e-01 -2.42383480e-01 2.04176977e-01 2.10251227e-01
3.04668188e-01 1.25080025e+00 -1.40133783e-01 -3.69670093e-01
1.03948982e-02 -7.60300905e-02 1.37640119e-01 -3.79039377e-01
-2.74167001e-01 6.48057878e-01 -2.15170771e-01 -7.10872293e-01
-5.43532729e-01 -4.59851444e-01 -9.11630511e-01 -4.16373163e-01
-2.51315236e-01 6.85637444e-02 9.85985875e-01 9.09134030e-01
7.87386671e-06 4.73858237e-01 7.92357445e-01 -9.81513679e-01
-5.62518656e-01 -7.47392178e-01 -6.57499194e-01 6.63590729e-01
6.52499259e-01 -6.88714087e-02 -6.44466400e-01 7.91985154e-01] | [11.442732810974121, -2.004211187362671] |
d5dab21c-5380-4365-bb66-632271efa13c | clues-before-answers-generation-enhanced | 2205.00274 | null | https://arxiv.org/abs/2205.00274v1 | https://arxiv.org/pdf/2205.00274v1.pdf | Clues Before Answers: Generation-Enhanced Multiple-Choice QA | A trending paradigm for multiple-choice question answering (MCQA) is using a text-to-text framework. By unifying data in different tasks into a single text-to-text format, it trains a generative encoder-decoder model which is both powerful and universal. However, a side effect of twisting a generation target to fit the classification nature of MCQA is the under-utilization of the decoder and the knowledge that can be decoded. To exploit the generation capability and underlying knowledge of a pre-trained encoder-decoder model, in this paper, we propose a generation-enhanced MCQA model named GenMC. It generates a clue from the question and then leverages the clue to enhance a reader for MCQA. It outperforms text-to-text models on multiple MCQA datasets. | ['Gong Cheng', 'Yue Zhao', 'Yu Gu', 'Jiaying Zhou', 'Ao Wu', 'Zixian Huang'] | 2022-04-30 | null | https://aclanthology.org/2022.naacl-main.239 | https://aclanthology.org/2022.naacl-main.239.pdf | naacl-2022-7 | ['multiple-choice-qa'] | ['natural-language-processing'] | [ 5.37089407e-01 6.00421309e-01 2.14348853e-01 -4.78442222e-01
-1.63392913e+00 -7.13400245e-01 7.42960930e-01 -9.70162451e-02
-9.51254889e-02 8.70104134e-01 7.01424360e-01 -5.03637373e-01
4.04158980e-01 -9.26606894e-01 -7.21600890e-01 -2.22302914e-01
7.64936447e-01 8.28512907e-01 1.41291037e-01 -7.83533871e-01
1.56609565e-01 -4.41554278e-01 -1.35206902e+00 9.01479304e-01
1.25041342e+00 7.92027593e-01 4.76229340e-01 1.20065045e+00
-3.99646252e-01 1.09218979e+00 -7.49880135e-01 -1.11946166e+00
2.07366943e-02 -9.13559675e-01 -9.62976635e-01 -1.02947317e-01
5.73757887e-01 -4.11435694e-01 -3.28994751e-01 5.16348898e-01
4.86900508e-01 -4.46825586e-02 8.20610881e-01 -1.19278347e+00
-1.14328349e+00 7.95186579e-01 1.59663230e-01 -6.04132116e-02
6.18060768e-01 2.21750975e-01 1.42597818e+00 -7.18529940e-01
6.76060319e-01 1.27082145e+00 1.76226407e-01 9.13854778e-01
-9.30212796e-01 -2.70008981e-01 1.54735148e-02 4.77444679e-02
-8.42784286e-01 -6.13487244e-01 3.62909317e-01 -2.38065720e-01
1.00737607e+00 3.48748893e-01 2.39826128e-01 1.36029279e+00
3.18973184e-01 1.20692527e+00 8.51625025e-01 -6.06640339e-01
3.80672701e-02 -1.02993287e-01 -1.04122736e-01 5.33583045e-01
-9.77089554e-02 -4.50429171e-02 -6.62225783e-01 -8.18673596e-02
3.43093991e-01 -3.09450299e-01 -2.73381144e-01 1.94348991e-01
-9.14776266e-01 1.16657114e+00 1.07847363e-01 -3.21378894e-02
-2.14109391e-01 1.58881217e-01 1.99249640e-01 5.66700876e-01
4.59975183e-01 8.62529695e-01 -4.97117370e-01 -3.87395889e-01
-9.05095279e-01 4.55059737e-01 1.06483054e+00 1.18416786e+00
6.14990711e-01 -1.12337835e-01 -9.19320166e-01 6.85585082e-01
5.10050535e-01 7.28094697e-01 6.27528906e-01 -8.56690347e-01
7.16980696e-01 6.57558739e-01 -9.90771875e-02 -2.50079751e-01
3.20560560e-02 -7.34385490e-01 -3.33516002e-01 -2.30780780e-01
3.93003702e-01 -3.15700591e-01 -9.44242358e-01 1.79784489e+00
6.35152459e-02 -4.73938346e-01 3.96058410e-01 6.45247519e-01
8.59771550e-01 7.53248453e-01 1.47316143e-01 3.07970166e-01
1.34276354e+00 -1.35489953e+00 -6.56189442e-01 -6.54840291e-01
9.27547038e-01 -7.27979839e-01 1.22607362e+00 2.16339663e-01
-1.19974351e+00 -5.23601949e-01 -9.16592538e-01 -6.68584526e-01
-2.30407894e-01 7.72303250e-03 1.19680591e-01 7.43087292e-01
-1.13443375e+00 3.65567654e-02 -3.40871334e-01 -1.08290888e-01
1.89128488e-01 -1.92365065e-01 -4.76619415e-02 -5.04786730e-01
-1.31921554e+00 9.54055309e-01 3.50179702e-01 -3.67555380e-01
-1.03131592e+00 -7.18254805e-01 -9.39754605e-01 3.80095877e-02
4.42927033e-01 -1.34377646e+00 1.86980247e+00 -9.18641627e-01
-1.67232978e+00 4.44009423e-01 -4.09627467e-01 -3.89450669e-01
3.03857654e-01 -3.22229624e-01 -4.53409761e-01 6.14595972e-02
3.79222065e-01 8.54063809e-01 9.99570191e-01 -1.17956853e+00
-6.90920830e-01 -1.58349991e-01 2.38764659e-01 3.56490999e-01
2.65157539e-02 -2.60185480e-01 -4.41620708e-01 -6.51444972e-01
-2.29878366e-01 -5.47697544e-01 5.32284714e-02 -4.95445311e-01
-3.18794698e-01 -3.64864260e-01 5.47074378e-01 -1.05799258e+00
1.35982049e+00 -1.81694818e+00 6.72046468e-02 -2.88488030e-01
2.87259132e-01 4.11751978e-02 -6.69345856e-01 8.65901351e-01
3.24419647e-01 1.53217494e-01 -2.15992868e-01 -5.68214059e-01
2.59630263e-01 2.86594898e-01 -6.70299411e-01 -3.73810232e-01
7.26850688e-01 1.58454049e+00 -1.07423174e+00 -4.78013188e-01
-3.40870857e-01 1.78871080e-01 -9.13411260e-01 6.61254108e-01
-1.03993428e+00 3.94123912e-01 -5.42472959e-01 6.06400907e-01
3.10659945e-01 -5.13252079e-01 9.62835774e-02 2.39143506e-01
2.85556287e-01 9.43820834e-01 -4.52691317e-01 1.92045701e+00
-6.51486695e-01 5.50004780e-01 -1.88823029e-01 -4.25852954e-01
7.85831153e-01 5.06291807e-01 -1.65349662e-01 -1.06166255e+00
-5.33499420e-02 4.56715137e-01 -7.76363304e-03 -5.98687112e-01
9.94187534e-01 -2.15056539e-01 -2.67369241e-01 8.57676864e-01
4.77928758e-01 -4.03995275e-01 3.90551746e-01 6.47082329e-01
1.24956012e+00 3.13376784e-01 1.76058542e-02 1.37780815e-01
1.50705189e-01 6.25872463e-02 2.50358302e-02 9.56174731e-01
3.43881637e-01 8.79931450e-01 4.04707998e-01 3.08601826e-01
-1.12743509e+00 -1.12486446e+00 2.21289203e-01 1.64847064e+00
-4.99184430e-01 -6.73074961e-01 -6.70152664e-01 -1.09322345e+00
-2.04724535e-01 1.30139089e+00 -6.38206959e-01 -3.75283748e-01
-4.30311739e-01 -4.56479788e-01 7.73393810e-01 5.29322624e-01
3.13534856e-01 -8.12292635e-01 -3.23545665e-01 4.64816183e-01
-6.88541651e-01 -1.00770545e+00 -6.35208845e-01 -2.35656053e-02
-6.58974707e-01 -7.43763387e-01 -7.28272200e-01 -5.07150352e-01
3.82600754e-01 -1.11392315e-03 1.92985034e+00 1.64222773e-02
2.25941181e-01 5.26971579e-01 -9.10002470e-01 -5.97944677e-01
-1.05183530e+00 4.99286711e-01 -8.43934119e-01 -1.27791882e-01
2.03818932e-01 -1.16156809e-01 -3.93620282e-01 1.18054710e-02
-1.23392808e+00 4.56677169e-01 9.54359412e-01 8.02394986e-01
2.52887487e-01 -6.89065516e-01 1.07880151e+00 -8.46744239e-01
9.86691773e-01 -8.01365495e-01 -2.00818717e-01 6.51921928e-01
-5.16976893e-01 5.41451216e-01 4.50635165e-01 -1.64782092e-01
-1.23011506e+00 -4.55237001e-01 -5.68381250e-01 1.55153498e-01
8.73926654e-02 6.73949301e-01 -4.59497839e-01 5.15989304e-01
7.02719450e-01 4.19736058e-01 -1.08405240e-01 -4.75158244e-01
1.05590916e+00 7.25404203e-01 5.55180907e-01 -5.38230300e-01
6.25328362e-01 -1.04437865e-01 -4.43303823e-01 -2.85759062e-01
-1.19934547e+00 -2.87414461e-01 -2.03245759e-01 -2.02510610e-01
1.10064828e+00 -9.22231376e-01 -1.78767025e-01 2.83403963e-01
-1.43636477e+00 -4.22587395e-01 -4.55577165e-01 1.13379126e-02
-3.87017667e-01 2.16191515e-01 -4.17832583e-01 -7.44412780e-01
-6.02787137e-01 -8.99002075e-01 1.21448970e+00 2.05348283e-01
-2.33356267e-01 -1.10061014e+00 1.68351203e-01 9.50443864e-01
6.28298163e-01 -2.33445480e-01 1.08251131e+00 -1.07142770e+00
-7.14293420e-01 -1.76204875e-01 -8.80288929e-02 4.23509538e-01
-1.79167122e-01 -4.18322176e-01 -1.02220047e+00 -6.43620566e-02
7.78237209e-02 -8.13016057e-01 9.56224382e-01 -1.05028555e-01
9.34510410e-01 -5.19787788e-01 1.49737867e-02 2.54474491e-01
1.15607369e+00 7.23265335e-02 8.63447964e-01 1.72070518e-01
5.42016804e-01 5.22869885e-01 3.03829163e-01 1.50919318e-01
1.11150026e+00 4.30095762e-01 4.71791655e-01 1.74884379e-01
-4.46406901e-01 -7.85319507e-01 6.05206251e-01 1.11426401e+00
4.05641854e-01 -8.54629338e-01 -7.52343237e-01 6.75114512e-01
-1.59855008e+00 -7.95618176e-01 -2.27167428e-01 1.98176718e+00
1.32528710e+00 -9.17150453e-02 -2.16260582e-01 -2.75652587e-01
2.96333700e-01 5.31634130e-02 -3.49661201e-01 -6.08327687e-01
-3.08542818e-01 5.44250488e-01 1.16743900e-01 6.35719836e-01
-5.32923162e-01 7.86237061e-01 6.71790123e+00 9.89440441e-01
-6.42970920e-01 4.36393738e-01 5.27068913e-01 7.68846720e-02
-9.69653368e-01 9.32216048e-02 -8.80530119e-01 3.62960577e-01
1.27203536e+00 -2.27709189e-01 2.15577349e-01 5.37788510e-01
-4.28644925e-01 -5.87927364e-02 -1.27789414e+00 4.51413393e-01
5.42614162e-01 -1.22883976e+00 6.55374348e-01 -3.64007056e-02
7.67842472e-01 -2.01295558e-02 1.28334403e-01 7.74916947e-01
6.42616570e-01 -1.29146564e+00 8.29084814e-01 5.68451226e-01
7.31306672e-01 -3.32470387e-01 7.40483403e-01 5.34715295e-01
-7.45388985e-01 -9.69858542e-02 -1.66279033e-01 -1.51091293e-01
3.66612494e-01 5.53472042e-01 -1.22242236e+00 8.45016360e-01
4.95642424e-02 1.93627134e-01 -8.50201547e-01 6.82155192e-01
-5.14446318e-01 7.09093511e-01 1.40556157e-01 -1.79313704e-01
3.27741981e-01 2.51703542e-02 3.63526672e-01 1.11394835e+00
5.48253417e-01 1.15490193e-02 -1.67403460e-01 1.01763856e+00
-4.23492610e-01 7.31685907e-02 -3.60602438e-01 -5.03149271e-01
3.91724467e-01 1.11734152e+00 1.45041957e-01 -5.66383600e-01
-6.03650928e-01 1.21317589e+00 3.96522522e-01 3.30994010e-01
-5.35413623e-01 -2.76181817e-01 1.59358352e-01 -2.70883888e-02
3.08052838e-01 -6.17888048e-02 -3.19107801e-01 -1.34636164e+00
9.44848359e-02 -1.49591279e+00 5.41090786e-01 -1.14442146e+00
-1.36234891e+00 6.96405709e-01 -1.71480566e-01 -7.99565136e-01
-8.91399503e-01 -4.60876554e-01 -5.13609827e-01 1.11698246e+00
-1.68139064e+00 -1.46727800e+00 -1.27773464e-01 5.29732466e-01
8.41337264e-01 -2.22860560e-01 8.40385914e-01 1.35102779e-01
-3.67296860e-02 7.76081681e-01 1.23479784e-01 1.80952862e-01
8.50535929e-01 -1.52756250e+00 8.21781754e-01 8.91749561e-01
3.08410287e-01 3.90059590e-01 4.74370301e-01 -6.59837186e-01
-1.43048191e+00 -1.23342860e+00 1.49490118e+00 -1.10112679e+00
5.01357019e-01 -5.48361063e-01 -8.37448120e-01 7.91531324e-01
7.36821830e-01 -8.10870528e-01 9.17247057e-01 -1.12645745e-01
-5.81400096e-01 2.92443782e-01 -6.86437786e-01 6.85515642e-01
6.58735275e-01 -8.64609480e-01 -1.04660654e+00 6.34888485e-02
1.16817629e+00 -4.95046645e-01 -6.71260595e-01 6.40794486e-02
4.22043443e-01 -6.59941733e-01 6.84346020e-01 -7.63051927e-01
1.02220178e+00 2.27158237e-02 -3.25203806e-01 -1.61595619e+00
-1.57058239e-01 -8.02459240e-01 -2.60034412e-01 1.17621553e+00
8.98986518e-01 -5.31384647e-01 3.41830641e-01 5.48297405e-01
-6.66121423e-01 -5.86161673e-01 -9.38478053e-01 -4.86567259e-01
4.73511428e-01 -5.66084564e-01 9.65783358e-01 5.02852261e-01
-1.47625834e-01 9.64021444e-01 -1.92832053e-01 -1.28789738e-01
3.81222010e-01 6.65159822e-02 7.72688985e-01 -8.72552514e-01
-6.75285459e-01 -1.67961597e-01 2.98552155e-01 -1.71695828e+00
-1.39525190e-01 -1.26717055e+00 1.86627761e-01 -1.81286693e+00
2.40437835e-01 -1.82962045e-01 1.88710913e-01 4.11421716e-01
-6.66244805e-01 1.27610013e-01 3.26202899e-01 -1.04727708e-01
-8.24535966e-01 7.83941448e-01 1.57427299e+00 -1.56117156e-02
1.70648903e-01 -1.57352269e-01 -1.25028491e+00 4.14283760e-02
6.49113178e-01 -4.79483128e-01 -5.98477185e-01 -9.90329027e-01
6.95311666e-01 4.45754498e-01 3.02293479e-01 -6.88821077e-01
1.76842749e-01 1.21879637e-01 1.64331287e-01 -6.28173649e-01
3.82923216e-01 -9.59310308e-02 -3.40359569e-01 1.33313954e-01
-7.30079770e-01 2.45959550e-01 -2.14448757e-02 5.02528012e-01
-3.21950555e-01 -4.59701270e-01 2.00452179e-01 -2.81031907e-01
-4.46120113e-01 7.42277578e-02 -4.61100787e-01 7.99563885e-01
3.82361174e-01 7.92538151e-02 -8.91137302e-01 -1.02758193e+00
-2.65578121e-01 4.37700331e-01 2.22810671e-01 6.41003907e-01
5.82105994e-01 -1.26014507e+00 -1.25515389e+00 1.67926207e-01
5.07315278e-01 -1.85919061e-01 1.30109191e-01 3.91507804e-01
-9.74018797e-02 6.32820427e-01 3.81990187e-02 -2.46058539e-01
-7.95088053e-01 1.23200648e-01 2.43198872e-01 -6.43920720e-01
-1.21978022e-01 9.55285490e-01 2.85002708e-01 -5.45363545e-01
-2.82697916e-01 -2.14439213e-01 -1.20544858e-01 -4.66813929e-02
5.02813101e-01 2.06167445e-01 1.94654465e-01 -1.75626040e-01
2.18188226e-01 -1.34001657e-01 -1.28352180e-01 -6.61082745e-01
9.61488008e-01 -3.60378623e-01 -6.47492707e-02 1.87667400e-01
1.08479643e+00 2.83315722e-02 -1.14135027e+00 -2.69783616e-01
-9.04731173e-03 -1.35169819e-01 -7.27192611e-02 -1.41560268e+00
-5.95206857e-01 1.05830622e+00 -2.52162367e-02 1.87559262e-01
9.93801534e-01 1.93240896e-01 1.17699945e+00 4.33483899e-01
3.47422846e-02 -1.03747821e+00 5.51569581e-01 9.26554382e-01
1.07445407e+00 -1.16377354e+00 -5.66107392e-01 -9.28335041e-02
-9.92036283e-01 1.04569411e+00 6.67327642e-01 4.91745055e-01
1.14318803e-01 -2.94499137e-02 2.34087914e-01 -1.21128105e-01
-1.31138992e+00 -4.31336790e-01 6.10712051e-01 7.45719790e-01
6.31499708e-01 -2.36681234e-02 -1.62789166e-01 1.00068724e+00
-6.37349904e-01 2.60176859e-03 4.99947101e-01 9.28481758e-01
-5.42246878e-01 -1.42706668e+00 -1.58350065e-01 6.07052028e-01
-4.49299127e-01 -5.80472350e-01 -5.09205937e-01 3.99713218e-01
7.61134503e-03 1.58487403e+00 4.08027694e-02 -3.82694364e-01
2.48999167e-02 5.78983545e-01 3.23359579e-01 -9.63706076e-01
-8.55905294e-01 -2.26051167e-01 5.86308420e-01 -2.58678347e-01
4.23320532e-02 -3.43436778e-01 -8.43220711e-01 1.65113937e-02
-3.38574857e-01 2.96995789e-01 5.52987337e-01 1.20934522e+00
6.46163404e-01 6.24444544e-01 4.09989357e-01 3.84955928e-02
-7.48029053e-01 -1.24552250e+00 -2.00909764e-01 3.49344194e-01
3.19888711e-01 -3.63496020e-02 -1.30924940e-01 9.66194645e-02] | [11.452658653259277, 8.284036636352539] |
1de6662d-ef65-4902-a3f1-376f9ed7ac7d | comparing-phonemes-and-visemes-with-dnn-based | 1805.02924 | null | http://arxiv.org/abs/1805.02924v1 | http://arxiv.org/pdf/1805.02924v1.pdf | Comparing phonemes and visemes with DNN-based lipreading | There is debate if phoneme or viseme units are the most effective for a
lipreading system. Some studies use phoneme units even though phonemes describe
unique short sounds; other studies tried to improve lipreading accuracy by
focusing on visemes with varying results. We compare the performance of a
lipreading system by modeling visual speech using either 13 viseme or 38
phoneme units. We report the accuracy of our system at both word and unit
levels. The evaluation task is large vocabulary continuous speech using the
TCD-TIMIT corpus. We complete our visual speech modeling via hybrid DNN-HMMs
and our visual speech decoder is a Weighted Finite-State Transducer (WFST). We
use DCT and Eigenlips as a representation of mouth ROI image. The phoneme
lipreading system word accuracy outperforms the viseme based system word
accuracy. However, the phoneme system achieved lower accuracy at the unit level
which shows the importance of the dictionary for decoding classification
outputs into words. | ['Helen L. Bear', 'Kwanchiva Thangthai', 'Richard Harvey'] | 2018-05-08 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 1.75533548e-01 1.30989417e-01 -4.68182027e-01 7.23931286e-03
-9.00176644e-01 -1.14814997e-01 7.97557533e-01 -4.42230999e-01
-5.92914760e-01 4.32440519e-01 7.62221158e-01 -6.47758007e-01
8.22469115e-01 -1.34911403e-01 -4.58765090e-01 -6.00119889e-01
7.89376974e-01 1.19693272e-01 2.55980194e-01 8.42067003e-02
5.29170752e-01 2.48482987e-01 -1.93195653e+00 4.65232730e-01
4.53363419e-01 7.78786004e-01 7.72308171e-01 1.33671701e+00
-4.74374384e-01 3.92645985e-01 -5.47330081e-01 -2.19298556e-01
7.27985874e-02 -4.68052924e-01 -3.63849461e-01 1.33550167e-01
6.27884626e-01 -3.19779634e-01 -1.68468297e-01 1.11977494e+00
1.13811624e+00 -2.09097803e-01 1.08939314e+00 -1.06119335e+00
-7.80168533e-01 3.81712675e-01 3.03556565e-02 -6.37266561e-02
4.58920717e-01 3.13831627e-01 6.13614261e-01 -1.31938803e+00
3.27880740e-01 1.56062555e+00 4.84581500e-01 9.56249654e-01
-1.02998221e+00 -4.98055011e-01 -4.53294247e-01 5.42218506e-01
-1.60253537e+00 -1.62987518e+00 4.45270747e-01 -4.16639149e-01
1.67404187e+00 2.05944270e-01 4.53121722e-01 1.26484299e+00
2.60045290e-01 6.90255940e-01 1.09603488e+00 -9.58234906e-01
7.74405431e-03 5.98396599e-01 3.59252878e-02 5.78108013e-01
-6.16066344e-02 2.29176521e-01 -7.89210439e-01 2.24433586e-01
5.22295237e-01 -6.68650985e-01 -3.96859258e-01 2.49604159e-03
-9.94742990e-01 9.17467892e-01 -5.28332293e-01 2.19026402e-01
-3.87174040e-01 1.14643902e-01 6.32311881e-01 1.26246840e-01
1.52550787e-01 -4.10064399e-01 -2.12775797e-01 -4.31510031e-01
-1.19715607e+00 -6.93787217e-01 7.89104164e-01 7.94422209e-01
1.35687545e-01 4.15483236e-01 -7.71669820e-02 1.61543250e+00
1.02998757e+00 8.90598059e-01 1.07749665e+00 -7.51531601e-01
2.56919861e-01 -1.86077207e-01 -3.39270204e-01 -3.34558874e-01
9.62637439e-02 5.42824864e-01 -2.96060681e-01 6.89157546e-01
2.94939399e-01 4.53004725e-02 -1.48040116e+00 1.40275979e+00
-1.52538180e-01 -1.02226742e-01 2.97033340e-01 5.06278753e-01
1.17804956e+00 9.84812796e-01 2.98279554e-01 -4.05755818e-01
1.59138727e+00 -9.97145653e-01 -1.42253244e+00 -2.09403589e-01
4.71583188e-01 -1.27840352e+00 1.15716290e+00 3.95200789e-01
-1.46112478e+00 -7.71124601e-01 -1.02079558e+00 -3.95530015e-01
-6.92646027e-01 5.55718064e-01 -7.97339454e-02 1.38410449e+00
-1.70540249e+00 -2.66837358e-01 -4.82346714e-01 -6.65780306e-01
-7.40214065e-02 4.08321142e-01 -4.11136091e-01 3.03278267e-01
-1.05722225e+00 1.48004961e+00 2.70683110e-01 -3.24739456e-01
-8.19753170e-01 -5.99547103e-02 -1.36148083e+00 5.76439574e-02
-3.32720697e-01 -2.96653122e-01 1.40292656e+00 -6.50377929e-01
-2.10992360e+00 1.00884795e+00 -9.55808699e-01 -4.47431117e-01
2.88588673e-01 5.62574446e-01 -7.23695934e-01 3.52210313e-01
-4.11747575e-01 1.33134449e+00 1.31962144e+00 -1.20010448e+00
-6.10060155e-01 7.13776425e-02 -7.10351288e-01 3.67165953e-01
-1.49824068e-01 2.71500975e-01 -3.80179554e-01 -4.22925353e-01
-2.22207010e-01 -6.25180840e-01 8.06012452e-01 1.29919946e-01
-1.94780812e-01 -4.86321449e-01 7.44154871e-01 -1.15972066e+00
1.29887366e+00 -2.32672262e+00 -3.67386937e-01 -2.54865259e-01
-5.91774918e-02 6.58132195e-01 -2.08556175e-01 3.72800261e-01
-1.12942964e-01 3.92462730e-01 3.51720713e-02 -8.90829921e-01
2.75804996e-01 2.21761271e-01 -1.25572473e-01 3.99202228e-01
-3.34684730e-01 1.00430965e+00 -5.35912402e-02 -1.05867600e+00
7.73184657e-01 9.58845198e-01 -2.91584998e-01 -1.07425615e-01
1.36373639e-01 -6.35443032e-01 4.99687582e-01 7.97957301e-01
7.49459207e-01 3.88006330e-01 -1.48620501e-01 -3.52826387e-01
-3.01850200e-01 6.33442402e-01 -7.87666738e-01 1.45508242e+00
-7.32282639e-01 1.48005700e+00 3.51473302e-01 -4.89723474e-01
7.42472053e-01 9.55284953e-01 -7.72111937e-02 -6.55779243e-01
2.74523020e-01 3.34373355e-01 3.17539787e-03 -8.31472576e-01
6.46372557e-01 -5.03476679e-01 7.36326277e-01 1.27765700e-01
2.81007856e-01 -2.53110081e-01 -2.54718423e-01 -8.31984207e-02
1.86462730e-01 -3.21135819e-01 3.59276950e-01 -1.88197300e-01
7.77599275e-01 -5.59285283e-01 -2.33471707e-01 4.35012281e-01
-6.89540803e-01 7.50048935e-01 1.21151358e-01 6.32950902e-01
-1.04693925e+00 -1.21176839e+00 -5.07982075e-01 1.13917744e+00
-1.17877223e-01 -4.06014681e-01 -1.05719566e+00 -5.48774451e-02
-2.86130458e-01 1.03592646e+00 -3.60520571e-01 4.95081469e-02
-7.08684996e-02 -1.78691655e-01 1.15067828e+00 2.99767494e-01
3.23594004e-01 -1.38547540e+00 -1.85976580e-01 1.47553220e-01
-3.30499768e-01 -1.13055968e+00 -9.75926220e-01 -1.75794318e-01
-3.55404764e-01 -4.29317653e-01 -1.22213602e+00 -1.54259324e+00
2.08797440e-01 8.76264498e-02 5.00721335e-01 -1.37890816e-01
-1.90156564e-01 6.01634741e-01 -1.32076204e-01 -6.49056435e-01
-1.18462396e+00 -2.28609487e-01 4.29757506e-01 -2.55685240e-01
8.62957776e-01 6.89938143e-02 -3.26438010e-01 2.63385415e-01
-6.40379310e-01 4.87487614e-02 6.15053236e-01 7.37285793e-01
1.01737335e-01 -5.15267849e-01 5.66437900e-01 3.49428564e-01
1.09773684e+00 2.33791657e-02 -2.67466575e-01 3.42787236e-01
-6.95021808e-01 -1.92738674e-03 -6.72446340e-02 -7.72690773e-01
-1.06885016e+00 -8.37008655e-02 -7.60690212e-01 -1.81423828e-01
-3.63102943e-01 -3.78621183e-02 -2.72284895e-01 -9.55752358e-02
4.96884406e-01 7.72234142e-01 7.43989527e-01 -6.60516083e-01
3.73597741e-01 1.78418481e+00 4.48484808e-01 7.81350955e-02
-1.51196957e-01 1.30563170e-01 -6.04641318e-01 -1.67840731e+00
7.09133685e-01 -7.29519904e-01 -2.87890375e-01 -3.81808430e-01
1.24663723e+00 -8.67205024e-01 -1.12825882e+00 9.09354806e-01
-1.42862725e+00 -3.55247319e-01 -8.55380520e-02 6.88197970e-01
-7.09119916e-01 7.98377097e-01 -7.07075000e-01 -1.15931797e+00
-2.96892941e-01 -1.64390278e+00 1.14342701e+00 -5.38508631e-02
-1.06190220e-01 -1.02869368e+00 1.35856554e-01 4.77165848e-01
4.14452910e-01 -9.57812488e-01 7.10510015e-01 -4.68797684e-01
7.19722658e-02 1.35302007e-01 -3.57838035e-01 8.52923989e-01
3.87629122e-01 6.53784052e-02 -1.39846969e+00 7.98137020e-03
-4.88034710e-02 -3.16646993e-01 9.91087139e-01 9.31812286e-01
6.24402046e-01 -4.08430487e-01 -3.34627688e-01 1.82378158e-01
1.28409910e+00 4.91254091e-01 1.04408765e+00 -2.28476003e-01
3.31407189e-01 5.99942029e-01 1.90484188e-02 1.03656445e-02
4.19937760e-01 7.28089333e-01 -1.31357126e-02 -2.59897318e-02
-1.05816793e+00 -4.89680946e-01 1.09347391e+00 1.17329204e+00
2.48123944e-01 -6.95324421e-01 -9.00650203e-01 7.92223573e-01
-1.02270043e+00 -9.46626127e-01 1.47230968e-01 2.01410365e+00
9.85947192e-01 3.95711772e-02 2.30096444e-01 3.51588875e-01
1.04769981e+00 1.91763788e-01 -1.49042308e-01 -1.14339280e+00
-1.15799963e-01 1.17801338e-01 6.68263078e-01 1.32060540e+00
-4.05429512e-01 1.32172036e+00 7.26345634e+00 1.34139967e+00
-1.41979825e+00 2.52900094e-01 8.64875987e-02 8.72531533e-02
1.03395045e-01 -5.28862536e-01 -1.13300788e+00 6.92518294e-01
1.51314592e+00 1.38531923e-01 4.51753259e-01 4.86120075e-01
6.42656028e-01 -4.57483351e-01 -9.48779285e-01 1.58086693e+00
7.50180423e-01 -1.11822009e+00 2.83739418e-01 3.44703108e-01
7.45760277e-02 2.13301644e-01 3.05225790e-01 1.11271970e-01
-4.27182972e-01 -1.43074822e+00 7.78688848e-01 4.68282253e-01
1.57737052e+00 -1.87234759e-01 2.47085333e-01 1.67798579e-01
-1.23322952e+00 3.49088937e-01 -4.65091109e-01 4.22743052e-01
3.90285730e-01 -4.68769044e-01 -1.55124509e+00 -4.81831789e-01
4.31558266e-02 2.64656663e-01 -3.06520820e-01 1.11477983e+00
1.89331055e-01 6.44139886e-01 -3.41714859e-01 -4.00807887e-01
-1.99116301e-02 1.55939028e-01 5.95314145e-01 1.76843119e+00
4.05158907e-01 -1.07127376e-01 -6.31585777e-01 3.15861732e-01
1.41764030e-01 3.84234577e-01 -6.12614691e-01 -2.31271729e-01
4.25550282e-01 6.33705616e-01 -1.98107854e-01 -5.76053679e-01
-4.33338970e-01 1.01973140e+00 -6.00642800e-01 4.37282354e-01
-4.14944261e-01 -4.67790991e-01 9.92124915e-01 7.63516501e-02
3.10742676e-01 -2.76576549e-01 -4.03375268e-01 -9.29346859e-01
-3.46410066e-01 -1.04432249e+00 -2.85640299e-01 -1.40888941e+00
-6.99609458e-01 4.83484268e-01 -2.43312553e-01 -1.02832019e+00
-6.05108261e-01 -1.21032310e+00 -1.97155759e-01 1.13009167e+00
-1.51029420e+00 -9.65156376e-01 1.78945750e-01 6.84052289e-01
1.27366328e+00 -4.29474115e-01 8.89234364e-01 1.65079191e-01
-9.73137245e-02 8.79238129e-01 2.06289049e-02 1.65419579e-01
9.66239810e-01 -8.76370668e-01 2.31694967e-01 5.90789795e-01
3.01029254e-02 5.71619570e-01 7.27479577e-01 -6.07414901e-01
-9.34319794e-01 -3.38734269e-01 1.51059878e+00 -4.07702297e-01
2.92697549e-01 -5.50446510e-01 -6.45295203e-01 3.24007690e-01
9.05942976e-01 -6.82442963e-01 6.34645283e-01 -6.20432377e-01
-1.48762405e-01 3.28127235e-01 -1.38501978e+00 7.65058160e-01
6.11938953e-01 -1.24393284e+00 -9.48373377e-01 5.72541244e-02
6.98589981e-01 7.38957897e-02 -5.48120916e-01 2.07708031e-02
9.25546646e-01 -6.80371642e-01 1.02064788e+00 -4.50278968e-02
-3.91720422e-02 -1.01867862e-01 -3.87671530e-01 -1.23336327e+00
2.17547715e-01 -3.77653688e-01 7.32983649e-02 1.10271633e+00
6.20377064e-01 -8.40522408e-01 4.29701746e-01 2.59689748e-01
-3.62346470e-02 -1.70016214e-01 -1.17846000e+00 -7.81730056e-01
1.86220348e-01 -7.24685609e-01 1.65982798e-01 4.25315589e-01
4.93712336e-01 3.01132351e-01 -4.47385877e-01 -1.89945504e-01
4.86565053e-01 -9.40862417e-01 1.84479073e-01 -7.76506305e-01
1.71331629e-01 -7.65110373e-01 -4.19388562e-01 -1.45170200e+00
3.42887700e-01 -7.46417761e-01 4.23315823e-01 -1.69819403e+00
-1.66987255e-01 3.04194927e-01 2.00676143e-01 4.48994428e-01
2.95949608e-01 3.54009211e-01 3.59320164e-01 -4.43144180e-02
2.08783641e-01 3.45996708e-01 9.04891610e-01 -3.64351541e-01
-2.54038244e-01 -2.17543185e-01 -8.63137320e-02 6.01822376e-01
7.76230574e-01 -5.92503361e-02 -4.00334001e-01 -2.37586632e-01
-3.90440434e-01 1.97109178e-01 2.14876086e-01 -8.29947233e-01
5.94672859e-01 3.35480869e-01 3.76443267e-02 -7.79704511e-01
1.07648253e+00 -5.63870132e-01 -3.05861861e-01 3.54646862e-01
-3.92225146e-01 -1.52260721e-01 4.40280735e-01 1.77519634e-01
-2.13378593e-01 -5.90173423e-01 1.06913853e+00 1.31699577e-01
-8.04366112e-01 -1.56412408e-01 -1.41909695e+00 -2.88609177e-01
5.36331713e-01 -1.05155289e+00 -2.69731343e-01 -7.92018831e-01
-9.13341224e-01 -4.26778883e-01 5.65582395e-01 4.69818920e-01
1.08141518e+00 -9.88538861e-01 -6.83409572e-01 5.98637998e-01
4.83208150e-02 -1.06999493e+00 -1.86867341e-01 6.57067478e-01
-5.80939412e-01 9.69214916e-01 -3.06511790e-01 -5.94449401e-01
-1.94156516e+00 3.48909199e-01 5.76484144e-01 8.16926122e-01
-2.66643226e-01 9.18371618e-01 8.52586031e-02 7.93531090e-02
8.02350104e-01 -4.65042204e-01 -4.75687295e-01 3.63847613e-01
5.63185155e-01 5.66465974e-01 3.69340293e-02 -1.15519845e+00
-3.61070842e-01 7.39375770e-01 2.14209303e-01 -9.83738482e-01
4.26856726e-01 -7.29050696e-01 2.85080314e-01 5.81190228e-01
1.53126740e+00 3.31291884e-01 -7.13454664e-01 3.23212892e-01
-4.86929536e-01 -1.19478330e-01 3.61996680e-01 -9.35342848e-01
-4.49506074e-01 1.33430469e+00 1.07611609e+00 1.37708321e-01
8.06874633e-01 -7.57666975e-02 8.11146200e-01 3.07925697e-02
-1.03786923e-01 -1.45817077e+00 -2.36980304e-01 4.66779590e-01
8.73043716e-01 -1.27532601e+00 -5.78058600e-01 -3.81569564e-01
-9.49495554e-01 1.11148548e+00 7.90093020e-02 6.04474247e-01
7.38682628e-01 3.37220579e-01 5.56974530e-01 4.38637346e-01
-6.43031597e-01 -5.47865927e-01 4.29019958e-01 1.00846076e+00
4.33689415e-01 7.30697662e-02 -5.13853192e-01 -3.57129797e-02
-3.34257573e-01 -3.81527059e-02 6.39257431e-01 3.61330807e-01
-8.35203290e-01 -1.06862462e+00 -5.58146238e-01 2.00279772e-01
-4.08729911e-01 -8.27463984e-01 -3.86774600e-01 5.61973691e-01
3.09765227e-02 1.52053046e+00 1.06906869e-01 -2.74893433e-01
-2.40161717e-01 8.20515871e-01 4.11459327e-01 -4.32674259e-01
-2.07397461e-01 3.76544476e-01 2.15328544e-01 -1.19989328e-01
-2.50767112e-01 -7.32068658e-01 -1.15197980e+00 -2.95047611e-01
-5.20468950e-01 -2.65365448e-02 1.55897462e+00 7.48609304e-01
8.96086171e-02 -4.83306218e-03 9.62666422e-02 -6.81076229e-01
-3.80752027e-01 -1.43814433e+00 -5.60484707e-01 -2.24159479e-01
8.16773057e-01 -4.25659657e-01 -8.58430207e-01 4.59941894e-01] | [14.204079627990723, 4.892361640930176] |
b47e7c34-0028-4fb0-abc6-df6292e83911 | metal-artifact-reduction-with-intra-oral-scan | 2202.03571 | null | https://arxiv.org/abs/2202.03571v1 | https://arxiv.org/pdf/2202.03571v1.pdf | Metal Artifact Reduction with Intra-Oral Scan Data for 3D Low Dose Maxillofacial CBCT Modeling | Low-dose dental cone beam computed tomography (CBCT) has been increasingly used for maxillofacial modeling. However, the presence of metallic inserts, such as implants, crowns, and dental filling, causes severe streaking and shading artifacts in a CBCT image and loss of the morphological structures of the teeth, which consequently prevents accurate segmentation of bones. A two-stage metal artifact reduction method is proposed for accurate 3D low-dose maxillofacial CBCT modeling, where a key idea is to utilize explicit tooth shape prior information from intra-oral scan data whose acquisition does not require any extra radiation exposure. In the first stage, an image-to-image deep learning network is employed to mitigate metal-related artifacts. To improve the learning ability, the proposed network is designed to take advantage of the intra-oral scan data as side-inputs and perform multi-task learning of auxiliary tooth segmentation. In the second stage, a 3D maxillofacial model is constructed by segmenting the bones from the dental CBCT image corrected in the first stage. For accurate bone segmentation, weighted thresholding is applied, wherein the weighting region is determined depending on the geometry of the intra-oral scan data. Because acquiring a paired training dataset of metal-artifact-free and metal artifact-affected dental CBCT images is challenging in clinical practice, an automatic method of generating a realistic dataset according to the CBCT physics model is introduced. Numerical simulations and clinical experiments show the feasibility of the proposed method, which takes advantage of tooth surface information from intra-oral scan data in 3D low dose maxillofacial CBCT modeling. | ['Jin Keun Seo', 'Hyoung Suk Park', 'Tae Jun Jang', 'Hye Sun Yun', 'Taigyntuya Bayaraa', 'Chang Min Hyun'] | 2022-02-08 | null | null | null | null | ['metal-artifact-reduction'] | ['medical'] | [ 2.56368637e-01 2.90562063e-01 1.00214928e-01 -3.87953520e-01
-7.82185078e-01 3.06487292e-01 -8.42712075e-02 1.21143281e-01
-5.02256691e-01 1.83970928e-01 -6.77310377e-02 -2.60799736e-01
1.69420131e-02 -8.27930927e-01 -5.31725883e-01 -8.62206876e-01
3.78088444e-01 6.75857425e-01 4.31036711e-01 9.97195989e-02
1.67759657e-01 6.57574832e-01 -1.41701996e+00 2.88973361e-01
7.93207943e-01 9.09383237e-01 6.63581967e-01 1.25864282e-01
-3.02270710e-01 3.87728959e-01 -2.09137157e-01 -2.06475317e-01
2.85679966e-01 -3.69087130e-01 -3.58222008e-01 5.39079487e-01
9.95749608e-02 -6.28557682e-01 9.25171934e-03 9.10602927e-01
6.90360188e-01 -1.52489558e-01 6.41723990e-01 -7.55073607e-01
-3.25431526e-01 1.90315962e-01 -7.62866855e-01 1.28396899e-01
-9.91887003e-02 3.56137872e-01 3.42241526e-01 -9.16318119e-01
3.56627554e-01 1.09802699e+00 5.34076512e-01 7.51913428e-01
-1.04912686e+00 -7.34634697e-01 -1.15731426e-01 5.68839125e-02
-9.78987634e-01 -1.21558562e-01 1.04354131e+00 -4.88146245e-01
4.68618274e-01 4.93929684e-02 9.36958671e-01 4.15261358e-01
5.26036024e-01 5.56671321e-01 1.11131394e+00 -6.66498482e-01
4.89509284e-01 -1.07471250e-01 5.27496859e-02 7.69098580e-01
2.42239207e-01 4.06183749e-01 1.00233465e-01 -2.09577546e-01
8.23001623e-01 3.39951754e-01 -3.32490057e-01 -2.45088205e-01
-4.30340588e-01 7.06198990e-01 7.69957840e-01 3.93279374e-01
-6.08078480e-01 -1.19643718e-01 3.34396452e-01 -2.71526903e-01
5.45373738e-01 -3.72778714e-01 -1.54570922e-01 5.19625008e-01
-7.38724113e-01 -1.31730437e-01 1.47515789e-01 4.38276052e-01
4.96598154e-01 2.29367334e-02 1.81257635e-01 1.23284388e+00
1.01941538e+00 5.38201988e-01 7.97300518e-01 -5.64487696e-01
1.75161988e-01 5.64637423e-01 -4.59851235e-01 -3.37107897e-01
-3.61706883e-01 -7.37951621e-02 -7.27430522e-01 4.82892066e-01
4.12243485e-01 3.20422530e-01 -1.77184761e+00 1.18870747e+00
9.82709825e-01 -6.74586669e-02 -2.86683977e-01 1.01532817e+00
8.01500857e-01 4.04458940e-01 3.61616254e-01 -5.88019907e-01
1.54192340e+00 -5.59615076e-01 -6.57676816e-01 -4.67175633e-01
4.54979450e-01 -8.28487396e-01 1.34002805e+00 1.71590835e-01
-1.20708048e+00 -3.23442459e-01 -1.22035944e+00 -1.22977927e-01
1.32666424e-01 -2.17141092e-01 3.97455543e-01 8.26443613e-01
-4.59827691e-01 2.71785796e-01 -1.28905833e+00 3.11221063e-01
7.87653029e-01 4.69115257e-01 -5.54163381e-02 -5.07935405e-01
-7.23711848e-01 9.08678114e-01 -1.58459306e-01 4.27447855e-01
-6.69078946e-01 -8.08701575e-01 -9.30447578e-01 -2.79038221e-01
1.05332702e-01 -6.96066439e-01 1.46777916e+00 -7.03369737e-01
-1.95438099e+00 9.91409779e-01 1.48472413e-02 1.98269933e-01
6.21336758e-01 7.78067112e-02 -1.88817754e-01 2.50797182e-01
3.28408092e-01 1.52719945e-01 7.47796595e-01 -1.51327026e+00
-2.46165499e-01 -1.25525880e+00 -4.86641794e-01 -4.56497772e-03
2.36223444e-01 -7.94094428e-02 -7.47110128e-01 -5.02335370e-01
9.68588889e-01 -8.43083739e-01 -6.63428009e-01 3.18549246e-01
-3.29685658e-01 7.18727559e-02 4.99487311e-01 -8.29292655e-01
6.38777852e-01 -2.17511106e+00 -6.56242728e-01 4.70413387e-01
-9.09243450e-02 -8.52150023e-02 1.35614991e-01 -2.09512323e-01
-1.19715013e-01 -2.44010404e-01 -7.25253105e-01 -5.54503024e-01
-5.77601552e-01 2.94572353e-01 3.44407707e-01 7.36968696e-01
-4.38775659e-01 5.02430856e-01 -4.08203632e-01 -8.27334642e-01
4.36207086e-01 7.94133425e-01 -4.81603354e-01 -3.75636481e-02
-2.80989736e-01 6.85592413e-01 -7.21821964e-01 7.95369029e-01
1.22163332e+00 -7.16154352e-02 1.05316035e-01 -3.26014847e-01
1.09545924e-01 1.78787261e-01 -9.31479335e-01 1.60456061e+00
-9.91296232e-01 -1.85437158e-01 4.46125776e-01 -5.26020944e-01
4.97145951e-01 4.91424471e-01 8.18614781e-01 -1.40883327e+00
4.28223401e-01 6.20029449e-01 4.46532220e-02 -6.93882644e-01
-3.74702096e-01 -9.03336823e-01 4.98370826e-01 5.75267613e-01
-4.59533840e-01 -7.11382926e-01 -3.23859364e-01 -3.20602983e-01
4.75990802e-01 -5.62038347e-02 5.56100123e-02 4.03238460e-03
5.88395178e-01 -1.36558741e-01 6.37120664e-01 2.18700022e-02
-6.67551532e-03 7.87499487e-01 -2.78364748e-01 -5.75885713e-01
-8.97912443e-01 -1.03504539e+00 -6.99996471e-01 3.55577469e-01
1.53175801e-01 4.14473534e-01 -9.61448550e-01 -3.08475405e-01
-1.48664325e-01 6.19828105e-01 -4.77456689e-01 -7.99223185e-02
-1.13127339e+00 -1.18044698e+00 -1.78183898e-01 4.34415966e-01
4.26106334e-01 -1.18570340e+00 -7.70273030e-01 5.77537417e-01
2.15598894e-03 -9.14400160e-01 -4.64746892e-01 3.18866223e-01
-1.65142524e+00 -1.18346369e+00 -7.67284811e-01 -9.64123487e-01
9.45870340e-01 4.96860556e-02 7.85419822e-01 4.89105046e-01
-6.51804030e-01 1.65699899e-01 2.40239307e-01 -3.97964388e-01
-7.56084204e-01 -6.32729828e-01 -3.58628035e-01 -1.69710591e-01
2.43588984e-01 -6.31471932e-01 -9.88605738e-01 3.79580677e-01
-1.14575326e+00 -1.66199319e-02 6.61195993e-01 9.20660734e-01
1.11584222e+00 1.87623072e-02 2.48934492e-01 -8.40099633e-01
3.77489656e-01 -1.32384673e-02 -7.88115203e-01 1.17072985e-01
-6.53770983e-01 -3.69841829e-02 2.73152143e-01 -3.15228462e-01
-1.47991562e+00 3.81654859e-01 -7.26601720e-01 -3.54751013e-02
-6.72649294e-02 2.72359520e-01 -4.86921549e-01 1.80977117e-02
4.67147201e-01 2.33297735e-01 2.78621167e-01 -5.34835696e-01
2.46462915e-02 6.84124291e-01 4.86096829e-01 -3.68916273e-01
4.09514874e-01 8.62692595e-01 -1.42285542e-04 -8.49366367e-01
-5.11489332e-01 -3.55200201e-01 -6.29141867e-01 -2.35109672e-01
1.03111160e+00 -8.02912951e-01 -8.31273496e-01 6.49501145e-01
-1.06754887e+00 -2.97519177e-01 -2.21876293e-01 9.04451728e-01
-5.07538199e-01 5.63321114e-01 -9.62348342e-01 -7.79090226e-01
-6.05938852e-01 -1.67022407e+00 1.05263197e+00 1.34721041e-01
2.43736461e-01 -6.29666030e-01 5.21392711e-02 8.81610572e-01
1.04200910e-03 8.04828256e-02 1.41254985e+00 -2.55406136e-03
-6.25561655e-01 -1.92861468e-01 2.45973334e-01 5.14289141e-01
4.07189220e-01 -3.51345569e-01 -1.04376769e+00 1.44864440e-01
8.90020788e-01 -2.53479332e-01 5.98282993e-01 1.01016021e+00
1.22634268e+00 2.12615490e-01 -4.55490291e-01 6.32781625e-01
1.62905300e+00 7.88053453e-01 6.59857452e-01 3.26107651e-01
6.68933094e-01 5.42617619e-01 4.33411688e-01 1.77085668e-01
4.22521323e-01 6.07996702e-01 5.85825264e-01 -1.89375460e-01
-4.10014838e-01 -9.56420898e-02 -1.03740796e-01 7.15454280e-01
-7.36915842e-02 2.82019317e-01 -7.63284683e-01 3.91871572e-01
-1.04189420e+00 -4.51199412e-01 -3.83255273e-01 2.39394283e+00
9.71883237e-01 9.84145999e-02 -2.04253629e-01 3.85064483e-01
6.60082221e-01 -4.63968158e-01 -8.08039427e-01 1.91110407e-03
3.96834195e-01 6.84031963e-01 2.55700260e-01 7.60466993e-01
-4.03957963e-01 4.03984934e-01 5.13698053e+00 7.40359426e-01
-1.34202743e+00 4.55426604e-01 7.52292395e-01 -4.12037931e-02
-3.66549850e-01 -3.22254449e-01 -4.77844238e-01 5.84289551e-01
4.76898342e-01 4.93020475e-01 -1.68517232e-01 8.80755484e-01
5.55924833e-01 -6.11506224e-01 -9.33919191e-01 8.61822069e-01
-2.36857161e-01 -9.33001399e-01 -1.61670282e-01 3.05615425e-01
3.96858007e-01 -4.16266806e-02 -1.58288255e-01 -7.69818127e-02
9.80945453e-02 -7.33922899e-01 6.06437564e-01 -5.41302860e-02
5.19280076e-01 -5.41233063e-01 4.74968106e-01 3.28069210e-01
-8.25317502e-01 6.60548434e-02 -2.28581965e-01 4.17834222e-01
5.74461341e-01 9.90746558e-01 -8.88771474e-01 1.90538898e-01
7.28536904e-01 -1.35378197e-01 -1.41002849e-01 1.19273698e+00
-2.28380099e-01 2.99276650e-01 -6.04128361e-01 4.30724919e-01
-9.40610245e-02 -3.17834496e-01 -5.27040809e-02 4.35439676e-01
3.17721903e-01 5.34620702e-01 1.52549399e-02 7.97560215e-01
-6.33208007e-02 4.50469077e-01 -1.93237156e-01 9.77438211e-01
1.61882192e-01 8.54671717e-01 -1.13271856e+00 -2.00481579e-01
-5.05674958e-01 7.82169223e-01 -1.74861655e-01 4.47547771e-02
-5.33819258e-01 5.90058208e-01 6.21306561e-02 5.80245614e-01
8.77397358e-02 6.08460456e-02 -8.68841350e-01 -5.65366387e-01
2.17879444e-01 -5.02391994e-01 3.45239103e-01 -6.37724340e-01
-1.13242126e+00 6.25956655e-01 -9.99989137e-02 -9.94109988e-01
-5.19333184e-02 -3.74383986e-01 -7.55954027e-01 7.68537223e-01
-1.44388676e+00 -1.10587442e+00 -3.39357644e-01 5.95540524e-01
6.49382651e-01 4.85861033e-01 5.29694736e-01 5.04563987e-01
-5.45744002e-01 3.24729174e-01 2.07509864e-02 -1.29987136e-01
1.86881378e-01 -8.70618522e-01 -6.61345795e-02 4.04842108e-01
-2.16527954e-01 2.69279391e-01 4.26424265e-01 -1.00367880e+00
-9.72232699e-01 -7.86417067e-01 4.58149821e-01 7.44650653e-03
3.80373895e-02 -1.15427956e-01 -8.75134408e-01 5.13055444e-01
-1.94526523e-01 3.31485063e-01 8.25510919e-01 -5.93021750e-01
9.05130059e-02 -3.92598622e-02 -1.66956592e+00 4.45310891e-01
5.46607256e-01 -2.65223354e-01 -5.32814264e-01 5.43289125e-01
3.64134520e-01 -6.75916851e-01 -8.48741710e-01 5.42624712e-01
5.60181320e-01 -8.97101521e-01 1.23733997e+00 7.84666985e-02
4.49099243e-01 -1.00882180e-01 -6.14157021e-02 -1.06709886e+00
1.69205010e-01 1.57899573e-01 4.56013173e-01 6.93963945e-01
2.31087685e-01 -5.20095468e-01 1.16827869e+00 8.99112105e-01
-7.05386817e-01 -7.60669291e-01 -1.35665166e+00 -5.46772242e-01
3.51249278e-01 -6.51691616e-01 3.29357326e-01 6.20471239e-01
-4.66212779e-01 2.95949019e-02 4.25783724e-01 4.48753357e-01
8.94360363e-01 2.89428890e-01 3.25671583e-01 -1.30968368e+00
-1.27921343e-01 -1.41455486e-01 2.91799959e-02 -9.37895596e-01
-2.15758085e-01 -8.44590485e-01 2.61467338e-01 -1.69145763e+00
1.22729577e-01 -7.24751234e-01 5.05088530e-02 2.05664024e-01
-1.06368877e-01 4.01801080e-01 -1.89751193e-01 2.92462260e-01
6.30489171e-01 7.81088769e-01 2.01905727e+00 -3.54763657e-01
-4.94972855e-01 5.38462877e-01 -1.07091106e-01 1.03327441e+00
4.06713217e-01 -7.02009082e-01 -5.07914424e-01 -4.33999449e-01
-3.52406979e-01 3.31947744e-01 4.28676307e-01 -8.13964069e-01
1.45001754e-01 4.46517691e-02 3.83971065e-01 -8.17057550e-01
4.91276145e-01 -1.26695013e+00 2.20481157e-01 1.03567731e+00
2.04242215e-01 -3.46520215e-01 2.66718596e-01 3.90947044e-01
-2.33698592e-01 -4.66807067e-01 1.31717205e+00 -5.79327345e-01
-9.66058373e-02 4.23273504e-01 -2.92638630e-01 -3.06254566e-01
8.40682983e-01 -5.85286498e-01 3.84486526e-01 1.60545543e-01
-1.02480602e+00 -2.56709546e-01 2.66231179e-01 -5.48650026e-02
9.24876690e-01 -1.14773619e+00 -4.80483711e-01 4.10911798e-01
5.13603501e-02 6.03792906e-01 7.91016400e-01 7.90297449e-01
-7.39967465e-01 4.44919094e-02 -8.36495757e-02 -8.86075616e-01
-1.09219611e+00 4.64822948e-01 7.26326108e-01 -2.72810519e-01
-9.14133012e-01 6.20630980e-01 3.70807827e-01 -3.91511679e-01
-6.43276349e-02 -8.66294742e-01 1.21422499e-01 -2.54278004e-01
3.21183771e-01 5.46132214e-02 7.66540170e-01 -7.46767581e-01
-3.17042142e-01 1.10058022e+00 -2.93600917e-01 -1.72773972e-02
1.57195628e+00 -1.74520746e-01 8.13247412e-02 -6.28210306e-02
1.10308862e+00 -3.77790071e-02 -1.14032602e+00 -1.86687872e-01
-2.13028803e-01 -3.61981750e-01 5.36877632e-01 -7.58372128e-01
-1.67146647e+00 8.88379872e-01 1.26493084e+00 -3.91622543e-01
1.15764761e+00 5.37423678e-02 1.12075555e+00 -3.02575022e-01
5.27429879e-01 -1.07517040e+00 2.47856423e-01 -3.28839600e-01
8.31452250e-01 -1.39546990e+00 1.63016602e-01 -9.91345048e-01
-1.73399597e-01 1.04703367e+00 7.04174995e-01 2.47285292e-01
1.16694665e+00 3.35932434e-01 4.67731655e-01 -6.50705159e-01
1.33421093e-01 1.50617078e-01 -1.16301365e-01 6.70072615e-01
3.07945579e-01 -2.02459320e-01 -4.45348710e-01 4.14254099e-01
-9.66044813e-02 2.33168468e-01 3.71641546e-01 1.12796652e+00
-2.59299457e-01 -1.07901800e+00 -7.69971073e-01 8.26512277e-02
-5.50936520e-01 -7.23904446e-02 3.45630199e-01 8.03817451e-01
1.86408758e-01 8.80892277e-01 -1.05987728e-01 3.19306463e-01
5.19650638e-01 -1.17137589e-01 6.56601369e-01 -7.89122283e-01
-6.33240581e-01 5.61103046e-01 -4.23954278e-01 -3.38225812e-01
-4.01353627e-01 -4.92794901e-01 -1.71128309e+00 1.75260231e-01
-5.77622235e-01 -2.92201012e-01 1.16821003e+00 1.08876050e+00
-2.13392362e-01 5.24329424e-01 7.23715127e-01 -9.32224631e-01
-3.03260714e-01 -9.29811478e-01 -7.06005096e-01 3.54564935e-01
1.84702098e-01 -6.81784511e-01 -4.73738819e-01 8.53177086e-02] | [13.733798027038574, -2.2509491443634033] |
bc2b838b-7bf7-42d7-8759-d7cc2f651700 | ssorn-self-supervised-outlier-removal-network | 2208.14093 | null | https://arxiv.org/abs/2208.14093v1 | https://arxiv.org/pdf/2208.14093v1.pdf | SSORN: Self-Supervised Outlier Removal Network for Robust Homography Estimation | The traditional homography estimation pipeline consists of four main steps: feature detection, feature matching, outlier removal and transformation estimation. Recent deep learning models intend to address the homography estimation problem using a single convolutional network. While these models are trained in an end-to-end fashion to simplify the homography estimation problem, they lack the feature matching step and/or the outlier removal step, which are important steps in the traditional homography estimation pipeline. In this paper, we attempt to build a deep learning model that mimics all four steps in the traditional homography estimation pipeline. In particular, the feature matching step is implemented using the cost volume technique. To remove outliers in the cost volume, we treat this outlier removal problem as a denoising problem and propose a novel self-supervised loss to solve the problem. Extensive experiments on synthetic and real datasets demonstrate that the proposed model outperforms existing deep learning models. | ['Zhenyu He', 'Wenjie Pei', 'Yi Li'] | 2022-08-30 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [-5.03486320e-02 -2.44260654e-02 3.80698740e-01 -2.53513545e-01
-7.20122755e-01 -1.74255967e-01 6.84142113e-01 1.04653180e-01
-3.18525583e-01 2.63777107e-01 3.50478031e-02 1.72624856e-01
1.94861323e-01 -7.35095978e-01 -1.18638945e+00 -4.97914791e-01
2.65715867e-01 4.75268483e-01 2.08360627e-01 1.20327719e-01
5.51076233e-01 5.03098547e-01 -1.17487144e+00 -2.13906229e-01
1.03511703e+00 9.97515678e-01 -3.67018968e-01 4.40367341e-01
1.77631840e-01 6.07549608e-01 -4.55616772e-01 -4.98426974e-01
7.55553663e-01 -4.03074980e-01 -2.89328992e-01 2.47524977e-01
9.54236090e-01 -9.97087419e-01 -6.31572008e-01 1.24705112e+00
5.54787338e-01 9.08461437e-02 4.72586900e-01 -1.47205365e+00
-4.76012915e-01 1.38403535e-01 -7.49226213e-01 -3.45159948e-01
2.31518000e-01 7.99852982e-02 8.71247888e-01 -1.09266376e+00
5.29472113e-01 1.10598063e+00 1.13200128e+00 -1.27430215e-01
-1.07603204e+00 -6.38433874e-01 -3.70195925e-01 7.51179177e-03
-1.25163352e+00 -4.38699096e-01 9.98759270e-01 -5.04761398e-01
1.01737726e+00 -1.96301401e-01 7.74804771e-01 7.34179854e-01
5.15772700e-01 8.16448629e-01 5.91762722e-01 -7.95654655e-02
5.98003902e-02 -3.89993936e-01 -7.58937597e-02 7.91940749e-01
4.46165770e-01 1.95283800e-01 -3.50589693e-01 -1.25810161e-01
1.14369166e+00 2.90955603e-01 -7.55461827e-02 -9.42354858e-01
-1.20769668e+00 8.23436797e-01 6.02069914e-01 -6.31785840e-02
-3.35676819e-01 5.07564902e-01 4.94154185e-01 3.68812293e-01
4.05475706e-01 2.66691536e-01 -1.11608371e-01 5.72630689e-02
-1.30415940e+00 3.43663067e-01 9.70572054e-01 1.19947410e+00
1.19953513e+00 8.94513354e-02 1.43793672e-01 5.58262110e-01
3.33652169e-01 1.10732131e-01 5.39684892e-01 -1.03690577e+00
5.25496364e-01 7.74659157e-01 -8.68502110e-02 -1.42576766e+00
-3.37229341e-01 -6.01245463e-01 -9.98390734e-01 2.87020892e-01
3.02795529e-01 -1.22031532e-02 -1.06679058e+00 1.29645491e+00
4.51366514e-01 7.21623123e-01 -2.29149878e-01 7.85177588e-01
6.66516364e-01 3.53679180e-01 -5.47083378e-01 1.31550595e-01
9.09052849e-01 -1.34572029e+00 -7.35832810e-01 -2.74249464e-01
6.59568429e-01 -1.17643106e+00 7.22454667e-01 2.92876273e-01
-1.20132113e+00 -4.47116137e-01 -1.50452137e+00 -7.84626782e-01
-2.26508245e-01 2.82273114e-01 5.20595312e-01 1.81469724e-01
-8.23890865e-01 9.65852678e-01 -9.28856313e-01 -1.55063093e-01
2.43276060e-01 4.17691290e-01 -5.23854256e-01 1.69958249e-01
-8.86063457e-01 6.38026834e-01 3.38019490e-01 2.59547085e-01
-7.72642016e-01 -6.23020232e-01 -1.08539212e+00 3.17496777e-01
2.88102657e-01 -1.08806646e+00 1.03142881e+00 -6.20619953e-01
-1.62099910e+00 7.88873553e-01 4.78196032e-02 -5.36994755e-01
1.06033003e+00 -7.45238483e-01 1.62710994e-01 -5.12077548e-02
2.05184326e-01 3.05998683e-01 1.20408964e+00 -1.11960661e+00
-4.77136493e-01 -2.78214991e-01 -2.76705325e-01 6.78015873e-02
3.80725190e-02 -3.25337440e-01 -7.40108073e-01 -7.29926884e-01
5.92853248e-01 -8.67802262e-01 -2.08139405e-01 9.60061625e-02
-5.20400643e-01 1.86300546e-01 8.47982466e-01 -8.71964931e-01
9.01309371e-01 -2.05451584e+00 -9.92353633e-02 4.39272463e-01
5.01204610e-01 1.31686985e-01 8.97453874e-02 5.37834406e-01
-8.69892538e-02 -3.68621349e-01 -3.45776737e-01 -7.81569839e-01
2.77942091e-01 -6.91499263e-02 -2.47186452e-01 8.51333976e-01
1.77942678e-01 8.62833321e-01 -8.33298743e-01 -2.90267318e-01
6.12767756e-01 6.87775433e-01 -8.75733078e-01 3.37632120e-01
1.78597897e-01 2.59178340e-01 -1.65180750e-02 5.82747042e-01
1.07949615e+00 -1.21639119e-02 -7.87131935e-02 -5.65143466e-01
-2.92585880e-01 1.76293284e-01 -1.55812907e+00 1.84427965e+00
-4.29767013e-01 5.40332377e-01 4.17530276e-02 -9.49717462e-01
1.20292866e+00 8.92274976e-02 8.23516846e-01 -3.57943386e-01
5.15706539e-01 5.22090137e-01 1.34352654e-01 -2.09358424e-01
5.39834142e-01 1.96774587e-01 1.12507790e-01 2.46968880e-01
2.66231745e-01 -3.67319137e-01 1.17335528e-01 1.97368160e-01
1.02985549e+00 3.88909400e-01 4.31794494e-01 2.64584497e-02
5.13699472e-01 -1.23590320e-01 8.86818469e-01 5.90326965e-01
-4.22755778e-01 1.12855375e+00 6.56808197e-01 -6.50954127e-01
-1.43452060e+00 -9.36921120e-01 7.75989667e-02 2.93648988e-01
3.01598430e-01 -4.69410777e-01 -8.14451098e-01 -4.12334174e-01
3.20733994e-01 1.41909242e-01 -5.11155665e-01 -4.02213216e-01
-7.25953996e-01 -4.56693381e-01 3.25656205e-01 6.34065986e-01
1.00778842e+00 -7.27887452e-01 -4.84063387e-01 4.21391636e-01
1.28515884e-01 -1.15124750e+00 -7.25608170e-01 9.96596366e-02
-8.48663330e-01 -1.17541158e+00 -4.13957715e-01 -8.46635520e-01
8.52083743e-01 1.46427765e-01 9.37067986e-01 1.87250987e-01
-2.51907200e-01 1.45862117e-01 -4.31495123e-02 -1.78478107e-01
-1.11919180e-01 2.04624962e-02 -5.54265454e-02 1.25284210e-01
5.01777470e-01 -7.91919649e-01 -8.05476367e-01 2.23778531e-01
-6.90711319e-01 -1.26081079e-01 8.70890200e-01 1.02958548e+00
7.21748114e-01 -1.52894244e-01 1.44885451e-01 -7.21222162e-01
4.41395223e-01 -2.17022538e-01 -9.94545043e-01 -4.15236913e-02
-5.17849624e-01 -4.11572051e-04 9.17942703e-01 -2.71882892e-01
-6.77125573e-01 6.83045924e-01 -2.31300011e-01 -9.20184016e-01
5.13086729e-02 3.52047056e-01 -4.10192400e-01 -5.40644348e-01
1.38678059e-01 2.71194369e-01 9.02901515e-02 -6.78129137e-01
1.45238832e-01 2.38289684e-01 9.98982251e-01 -8.68255347e-02
1.48579478e+00 6.05096519e-01 1.33042723e-01 -6.18563533e-01
-4.35906142e-01 -6.61266387e-01 -1.08449638e+00 1.04509860e-01
6.54638767e-01 -1.01324916e+00 -7.15917647e-01 6.90117717e-01
-1.30003417e+00 1.06094964e-01 -2.27045044e-01 5.81885934e-01
-7.38019407e-01 8.63184929e-01 -4.63647455e-01 -3.04846525e-01
-5.62296152e-01 -1.36063337e+00 1.37392342e+00 7.81594515e-02
-2.74943672e-02 -8.95608664e-01 2.99429864e-01 1.43842384e-01
2.58678943e-01 4.55416799e-01 3.84289473e-01 -6.53822184e-01
-9.13099825e-01 -7.68162370e-01 -2.96089709e-01 5.24949491e-01
-7.74916783e-02 1.53216675e-01 -7.72289693e-01 -3.54989499e-01
2.90138900e-01 -2.03284398e-01 1.01728654e+00 3.48579317e-01
9.68916178e-01 -1.67596489e-01 3.25565003e-02 1.59479833e+00
1.50478554e+00 -7.24347355e-03 1.06140578e+00 6.95021272e-01
1.11388588e+00 2.31627733e-01 4.40583289e-01 4.76096839e-01
2.65094280e-01 4.54209656e-01 7.01792002e-01 -2.27955252e-01
-5.89296333e-02 -5.94704330e-01 1.40312880e-01 1.07724893e+00
2.03938216e-01 2.40588740e-01 -7.65243351e-01 4.76271659e-01
-2.16442966e+00 -7.60688722e-01 -2.50760615e-01 2.19841552e+00
3.73174876e-01 1.10446148e-01 -1.74547926e-01 2.71287799e-01
3.93663436e-01 2.52396464e-01 -6.21320188e-01 -3.24721128e-01
-3.52127291e-03 -1.80507362e-01 4.95601058e-01 6.36174202e-01
-1.38478994e+00 1.04790938e+00 5.90868711e+00 3.16862732e-01
-1.05640006e+00 -3.21961731e-01 7.54409507e-02 3.05452883e-01
1.75973251e-02 4.81187284e-01 -6.49885237e-01 3.36278588e-01
1.69955805e-01 -1.22842655e-01 3.33687782e-01 1.09632123e+00
1.06135644e-01 -7.88978264e-02 -1.27268732e+00 1.24556851e+00
3.58237088e-01 -1.23358905e+00 1.47627786e-01 -1.10820867e-01
7.05940485e-01 -1.43223599e-01 -1.70099199e-01 1.29635721e-01
3.79866213e-02 -6.93106949e-01 4.22591567e-01 5.70368648e-01
2.89342970e-01 -8.36458504e-01 8.23094845e-01 1.87227547e-01
-1.32984078e+00 6.42205700e-02 -6.70513809e-01 -5.18746711e-02
3.54314685e-01 7.83927381e-01 -7.05704987e-01 7.22640991e-01
6.35326445e-01 9.80178833e-01 -5.56287825e-01 1.52266002e+00
-2.51518548e-01 6.34363815e-02 -3.55156690e-01 6.56643987e-01
2.34388769e-01 -7.55880058e-01 5.61193764e-01 7.66982377e-01
4.86983061e-01 -3.86394411e-01 2.03344613e-01 9.66148019e-01
-4.62972939e-01 1.11058593e-01 -7.14857578e-01 2.10662454e-01
3.17650914e-01 1.18795335e+00 -4.90423322e-01 -3.15432221e-01
-7.01482058e-01 1.26675534e+00 2.92269588e-01 1.84176847e-01
-7.96660006e-01 -1.05287039e+00 4.98134106e-01 9.48138759e-02
2.34410152e-01 -3.99092108e-01 -6.19268358e-01 -1.42231750e+00
2.19064564e-01 -7.73756325e-01 -4.75439616e-02 -5.25236070e-01
-1.06646717e+00 1.43902391e-01 -2.53930777e-01 -1.38018346e+00
-3.24524909e-01 -4.80947018e-01 -1.16871488e+00 6.66116059e-01
-1.54793060e+00 -1.25363350e+00 -7.39777684e-01 5.84126294e-01
4.36192989e-01 -1.81833431e-01 2.30712950e-01 6.10634923e-01
-7.17558801e-01 6.38745904e-01 4.29732800e-01 3.39719474e-01
9.95465100e-01 -1.31777763e+00 7.69806683e-01 1.21760428e+00
-1.87103987e-01 9.11473751e-01 6.85099125e-01 -9.15670872e-01
-1.39464080e+00 -1.04402649e+00 1.03080249e+00 -1.54016227e-01
7.67083406e-01 -4.52209800e-01 -1.00521970e+00 1.03651631e+00
1.85688511e-01 1.01922460e-01 1.13584548e-01 -4.43140626e-01
-2.06619710e-01 -1.26675144e-01 -1.04616058e+00 4.28159982e-01
1.02614701e+00 -4.56582546e-01 -5.54970145e-01 2.17129871e-01
3.09547573e-01 -6.80347264e-01 -8.39044869e-01 5.39748073e-01
6.85146928e-01 -1.20318842e+00 1.05285668e+00 -7.34281838e-02
4.59787995e-01 -4.75426048e-01 2.69122005e-01 -1.23574114e+00
-2.71614969e-01 -9.06759322e-01 -4.96574938e-01 1.10526693e+00
-8.98962319e-02 -6.36343896e-01 8.85868311e-01 3.60848784e-01
-4.50487614e-01 -4.77055281e-01 -7.92623818e-01 -7.25628436e-01
-7.92704672e-02 1.01353908e-02 6.51201904e-01 1.04302883e+00
-4.48458642e-01 1.97302923e-01 -6.45224273e-01 2.87643284e-01
1.03903067e+00 1.83798954e-01 1.32067561e+00 -1.25580883e+00
-1.68576092e-01 -2.85038829e-01 -6.85845852e-01 -1.27399528e+00
1.58913478e-01 -6.70428097e-01 2.26722732e-01 -1.63203025e+00
9.68514010e-03 1.22411475e-01 2.58288562e-01 2.24965394e-01
-1.55818373e-01 1.56379655e-01 -3.61889638e-02 4.22794431e-01
-2.28402123e-01 8.42339754e-01 1.02909935e+00 -7.60617107e-02
-3.40113491e-01 -9.02313665e-02 -1.98024213e-01 1.05375385e+00
6.76762521e-01 -3.67261797e-01 -3.43533874e-01 -6.98515773e-01
-3.09997331e-03 -3.12725395e-01 4.50087667e-01 -1.39151955e+00
7.22129583e-01 2.04104915e-01 6.66119397e-01 -1.09337497e+00
2.88608015e-01 -1.08991039e+00 1.64088458e-02 5.11516750e-01
1.48675919e-01 4.28867161e-01 -7.63215944e-02 3.63135278e-01
-3.65888298e-01 -2.46338174e-02 8.27819645e-01 -5.04716597e-02
-4.37029511e-01 5.72884083e-01 7.66881406e-02 -1.61190271e-01
9.92650151e-01 -4.35041428e-01 -1.93200678e-01 -4.10210431e-01
-5.14676332e-01 2.21204296e-01 7.33563244e-01 2.57007986e-01
8.56359959e-01 -1.29436386e+00 -5.24883151e-01 5.35208881e-01
-9.02953073e-02 3.41379076e-01 2.14249399e-02 1.02635837e+00
-1.12476861e+00 -4.15427834e-02 -3.99408966e-01 -4.54473674e-01
-8.66495371e-01 4.21009094e-01 5.60648322e-01 -3.69245738e-01
-8.00682068e-01 5.02115250e-01 6.21787608e-02 -5.08117437e-01
6.03699803e-01 -3.77263039e-01 9.95720848e-02 -1.75581768e-01
9.82484072e-02 6.38461530e-01 1.46847889e-01 -7.56453633e-01
-1.25472620e-01 9.28838849e-01 -1.46272093e-01 1.55303806e-01
1.35234964e+00 -1.46456942e-01 -1.84052616e-01 8.00827593e-02
1.55274332e+00 -1.89723074e-01 -1.36987054e+00 -1.06831200e-01
8.85486081e-02 -6.98021889e-01 1.44305974e-01 -1.43395692e-01
-1.24412370e+00 1.10195529e+00 4.79241699e-01 -2.35008299e-01
9.24099743e-01 -7.06305385e-01 1.37434435e+00 5.02568722e-01
-1.53177112e-01 -1.47661805e+00 2.59791404e-01 8.56761396e-01
9.21218872e-01 -1.29204726e+00 3.71356875e-01 -4.45290029e-01
-1.69505358e-01 1.29336226e+00 9.51051652e-01 -8.20042908e-01
6.98370636e-01 2.91441083e-01 3.87275368e-02 -1.37436882e-01
-3.48932087e-01 -1.18918121e-01 2.59032220e-01 3.74998063e-01
4.13387716e-01 -5.90507567e-01 -5.09221673e-01 1.51765317e-01
-3.02937001e-01 -5.07272929e-02 6.00524724e-01 1.02184594e+00
-4.31801230e-01 -1.15431261e+00 -3.66793096e-01 3.42863888e-01
-1.67771742e-01 -5.07203005e-02 -6.85064375e-01 9.18695211e-01
-5.65065481e-02 4.76361424e-01 1.30374968e-01 -5.67179680e-01
6.10047281e-01 -1.00361802e-01 1.89273238e-01 -2.73447722e-01
-7.24808395e-01 3.90589744e-01 -1.72955930e-01 -9.04806852e-01
-5.41073382e-02 -3.47060025e-01 -1.13928831e+00 -5.26966810e-01
-3.01195949e-01 -2.75465429e-01 6.01922452e-01 8.66333425e-01
2.71217823e-01 4.13044602e-01 7.78207242e-01 -1.12290668e+00
-7.93242872e-01 -7.54209578e-01 -5.18076599e-01 7.40427136e-01
5.73558033e-01 -6.94643259e-01 -5.42804360e-01 -1.05014130e-01] | [8.6321439743042, -2.2521753311157227] |
f3a75e76-9aec-41ac-a837-ab9ef9572207 | rethinking-portrait-matting-with-privacy | 2203.16828 | null | https://arxiv.org/abs/2203.16828v2 | https://arxiv.org/pdf/2203.16828v2.pdf | Rethinking Portrait Matting with Privacy Preserving | Recently, there has been an increasing concern about the privacy issue raised by identifiable information in machine learning. However, previous portrait matting methods were all based on identifiable images. To fill the gap, we present P3M-10k, which is the first large-scale anonymized benchmark for Privacy-Preserving Portrait Matting (P3M). P3M-10k consists of 10,421 high resolution face-blurred portrait images along with high-quality alpha mattes, which enables us to systematically evaluate both trimap-free and trimap-based matting methods and obtain some useful findings about model generalization ability under the privacy preserving training (PPT) setting. We also present a unified matting model dubbed P3M-Net that is compatible with both CNN and transformer backbones. To further mitigate the cross-domain performance gap issue under the PPT setting, we devise a simple yet effective Copy and Paste strategy (P3M-CP), which borrows facial information from public celebrity images and directs the network to reacquire the face context at both data and feature level. Extensive experiments on P3M-10k and public benchmarks demonstrate the superiority of P3M-Net over state-of-the-art methods and the effectiveness of P3M-CP in improving the cross-domain generalization ability, implying a great significance of P3M for future research and real-world applications. | ['DaCheng Tao', 'He Zhang', 'Jing Zhang', 'Jizhizi Li', 'Sihan Ma'] | 2022-03-31 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 2.76688606e-01 8.56942236e-02 -1.43270448e-01 -5.56829989e-01
-7.71273315e-01 -6.57186687e-01 5.86801946e-01 -6.04881048e-01
-6.89598620e-02 6.76053345e-01 3.34160149e-01 -1.58329800e-01
-3.74799259e-02 -7.36567318e-01 -9.86822784e-01 -6.20415688e-01
2.28247121e-01 8.69129039e-03 -3.42040420e-01 -2.92005893e-02
-2.94664204e-01 3.51993948e-01 -1.07756841e+00 6.95955932e-01
7.93920696e-01 1.17890620e+00 -4.73814696e-01 -9.90953073e-02
2.69244909e-01 2.77865380e-01 -4.30719972e-01 -1.50870264e+00
9.85569954e-01 -3.27729225e-01 -8.06937039e-01 2.04619646e-01
1.19219327e+00 -6.46688282e-01 -7.28659332e-01 1.16738331e+00
3.46183807e-01 -3.10621262e-01 3.77414525e-01 -1.89569271e+00
-1.16353369e+00 5.64051926e-01 -7.35347569e-01 -2.15976581e-01
9.43849050e-03 5.36640942e-01 6.76307678e-01 -1.03223050e+00
5.97678125e-01 1.49313390e+00 1.02422321e+00 7.67423511e-01
-1.52941513e+00 -1.14589751e+00 1.45650469e-02 6.21779040e-02
-1.61527896e+00 -7.06459641e-01 6.06944323e-01 -1.41409382e-01
2.88111597e-01 5.54826856e-01 5.02959609e-01 1.43862915e+00
2.10039213e-01 8.65789354e-01 1.35800970e+00 1.36215925e-01
-6.08086698e-02 2.82307982e-01 -1.92185432e-01 5.85652828e-01
2.93951094e-01 -5.51540703e-02 -8.57568204e-01 -5.29320300e-01
8.07395041e-01 3.97286713e-02 -5.23207366e-01 -3.98807138e-01
-1.05230403e+00 5.67882121e-01 5.70358515e-01 -6.97734207e-02
-1.27909854e-01 1.60613641e-01 3.22431356e-01 6.12597704e-01
7.84585714e-01 3.70223492e-01 -1.71749249e-01 2.59582281e-01
-1.23526824e+00 2.82755613e-01 4.95290250e-01 1.20832407e+00
5.49289048e-01 -1.48416236e-01 -3.77015978e-01 7.55835652e-01
5.17174974e-02 6.28977418e-01 6.86958879e-02 -1.00933278e+00
8.06184947e-01 4.67674494e-01 -1.46690503e-01 -1.24174726e+00
4.07840341e-01 -6.40389547e-02 -1.29197371e+00 7.33766865e-05
3.79485279e-01 5.80681600e-02 -8.32553267e-01 2.05762458e+00
2.26465955e-01 2.71752149e-01 2.02928800e-02 7.62870431e-01
7.64742136e-01 3.25928241e-01 -2.09374160e-01 1.42930135e-01
1.41126311e+00 -8.65050077e-01 -6.69414878e-01 -4.17593122e-02
2.33031213e-01 -5.09729326e-01 1.09463120e+00 2.69443784e-02
-9.26750720e-01 -2.89066195e-01 -8.61739755e-01 -2.05278516e-01
-4.23935354e-01 -1.22450665e-01 5.88763177e-01 9.48799014e-01
-1.36211586e+00 7.25041568e-01 -5.54700553e-01 -3.73971552e-01
1.39475620e+00 5.77905655e-01 -1.14229548e+00 -5.07579386e-01
-1.04857135e+00 4.90932375e-01 1.08764255e-02 2.74342149e-01
-1.03567159e+00 -1.24810278e+00 -8.03354740e-01 -1.33029908e-01
2.37835392e-01 -8.40442955e-01 8.05232763e-01 -1.12285709e+00
-1.03785479e+00 1.22581732e+00 -8.62815157e-02 -5.79114377e-01
9.55776691e-01 -2.53259510e-01 -4.94089842e-01 1.33328363e-01
7.14807436e-02 1.10618377e+00 1.17360950e+00 -1.29325366e+00
-2.18666613e-01 -4.94215339e-01 -8.20148662e-02 -9.38020647e-02
-7.51704633e-01 -1.84038430e-01 -4.85467911e-01 -9.13799286e-01
-2.32658386e-01 -8.74424100e-01 1.85174733e-01 5.65080285e-01
-7.51588523e-01 2.82025605e-01 1.08613634e+00 -8.36066961e-01
8.48728716e-01 -2.33774495e+00 -1.72112267e-02 2.38513678e-01
5.63134313e-01 5.28892159e-01 -5.69608510e-01 3.69976312e-01
-2.76214123e-01 4.13101405e-01 -4.07580405e-01 -9.52052653e-01
-6.91585913e-02 1.91003874e-01 -6.46738052e-01 5.68726897e-01
3.84368956e-01 1.31918824e+00 -6.08617544e-01 -4.27802920e-01
-9.43042487e-02 5.55173457e-01 -5.19253731e-01 -8.06941316e-02
2.28379369e-02 4.21897769e-01 -1.35813415e-01 9.52802598e-01
1.42662680e+00 -1.06902517e-01 3.92976105e-02 -2.66323775e-01
4.95403498e-01 -1.55849591e-01 -7.15554297e-01 1.73887014e+00
-2.54963748e-02 6.89902246e-01 3.45065564e-01 -3.98269773e-01
7.25528777e-01 2.00435862e-01 3.95351112e-01 -6.37841344e-01
4.70237732e-02 -2.65990160e-02 -5.97954214e-01 -8.70124698e-02
6.14988863e-01 5.61195277e-02 1.06219180e-01 2.33391032e-01
8.71747583e-02 3.43552440e-01 -5.46819746e-01 2.60388613e-01
1.07630837e+00 -2.15772405e-01 -1.41178578e-01 -5.27637839e-01
1.98779240e-01 -4.66370732e-01 7.63275683e-01 7.16669023e-01
-3.10492337e-01 7.97637343e-01 5.18461645e-01 -4.79130238e-01
-1.03303099e+00 -1.21467447e+00 -2.20132008e-01 5.95688045e-01
2.13778272e-01 -7.18360186e-01 -1.07918656e+00 -9.64807570e-01
5.06906211e-01 4.16671515e-01 -1.07893217e+00 -4.27414536e-01
-3.82127941e-01 -7.60124922e-01 8.96097600e-01 2.58788913e-01
1.17967832e+00 -6.94751322e-01 3.45588714e-01 -3.42081457e-01
-3.75907332e-01 -1.02561235e+00 -1.04126120e+00 -5.41412175e-01
-7.83259571e-01 -9.26982880e-01 -6.90978825e-01 -5.03431320e-01
9.30186749e-01 5.46987951e-01 1.10444224e+00 5.76888546e-02
-8.82655904e-02 5.75359821e-01 -7.36545697e-02 -3.13836604e-01
-3.77867043e-01 1.92603052e-01 6.15542196e-02 7.43621230e-01
3.87739658e-01 -9.69295859e-01 -7.44815409e-01 6.76152825e-01
-1.02896023e+00 2.43327245e-01 4.77376550e-01 7.76550949e-01
4.67572778e-01 4.41429317e-02 3.96352738e-01 -9.84855175e-01
3.47847044e-01 -4.58832294e-01 -4.05346096e-01 2.66531467e-01
-6.78135574e-01 -3.08366239e-01 3.65532517e-01 -5.03010929e-01
-7.81706631e-01 -1.32943019e-02 2.29797028e-02 -8.83836031e-01
2.12428883e-01 -6.75776154e-02 -9.43507612e-01 -4.14219350e-01
2.97483116e-01 4.11365002e-01 4.57474172e-01 -5.52166045e-01
5.26449919e-01 5.77923179e-01 9.17111039e-01 -5.67831576e-01
1.36489308e+00 7.59106755e-01 8.46649781e-02 -5.74636281e-01
-7.04276741e-01 1.40490904e-01 -4.68409330e-01 -1.08379818e-01
5.90453744e-01 -1.19774520e+00 -8.34492028e-01 9.67709959e-01
-9.11765397e-01 -1.27088264e-01 -3.13066572e-01 -1.77444264e-01
-3.70727807e-01 4.56719488e-01 -6.24441445e-01 -4.69263941e-01
-6.47679329e-01 -1.01416981e+00 1.15448546e+00 -5.25280200e-02
-1.27071217e-01 -6.56280279e-01 -4.05240387e-01 7.83174098e-01
4.63847280e-01 3.79503638e-01 7.90382683e-01 -3.76395255e-01
-9.14357126e-01 -1.89621300e-01 -3.68662775e-01 6.49261594e-01
3.12367409e-01 -4.94799495e-01 -1.30967522e+00 -6.75892532e-01
9.52706113e-03 -2.62837827e-01 1.00645852e+00 1.80263016e-02
1.51223934e+00 -9.40037310e-01 -2.88564056e-01 1.22014558e+00
1.20301890e+00 -3.72050434e-01 1.37304866e+00 2.29307860e-01
9.86858487e-01 5.07553458e-01 4.19971198e-01 2.33733773e-01
3.52752686e-01 6.08631134e-01 7.54285753e-01 -3.12183321e-01
-7.63714463e-02 -8.00717354e-01 3.92912149e-01 2.74646789e-01
2.73626775e-01 -2.41063446e-01 -2.65555859e-01 6.18454993e-01
-1.81408048e+00 -9.57592428e-01 2.53796756e-01 2.20341706e+00
1.11077619e+00 -2.72683114e-01 1.71518996e-01 -1.28461003e-01
6.82933033e-01 4.25422698e-01 -6.55825913e-01 2.09632181e-02
-3.96417350e-01 -1.33606130e-02 7.65093803e-01 1.52241200e-01
-1.21852660e+00 8.66766036e-01 6.09454393e+00 1.28789997e+00
-9.87291694e-01 1.43088028e-01 1.08344281e+00 -3.05771977e-01
-4.74203527e-01 -1.87208384e-01 -7.78861105e-01 8.04370522e-01
7.04730034e-01 -1.54845998e-01 7.32355893e-01 9.19179440e-01
-1.78923607e-01 3.97349536e-01 -1.49584615e+00 1.36964655e+00
8.58126506e-02 -1.69099653e+00 4.40842777e-01 5.34796000e-01
6.98302925e-01 -1.91276506e-01 6.94350779e-01 -2.09137592e-02
1.36332005e-01 -1.40736139e+00 5.68215787e-01 4.39371467e-01
1.25241041e+00 -7.39954591e-01 4.80367869e-01 -1.18629783e-01
-1.01912785e+00 1.14277877e-01 -4.46053892e-01 3.79585624e-01
-1.39811829e-01 5.66045403e-01 -6.55914307e-01 7.97948956e-01
8.31735611e-01 9.46073711e-01 -7.88763463e-01 7.43134081e-01
1.82938635e-01 6.37602806e-01 -4.39243495e-01 6.33325338e-01
6.68261498e-02 -2.12903351e-01 6.14394188e-01 8.73050213e-01
2.48148263e-01 -1.27549231e-01 -3.41597706e-01 1.15533423e+00
-8.48485291e-01 -1.36573493e-01 -7.97235250e-01 -3.79483476e-02
7.62091219e-01 1.28027380e+00 -2.47853324e-01 -4.11834791e-02
-2.17035040e-01 1.31279910e+00 1.31761640e-01 2.35803798e-01
-8.61810923e-01 2.12473288e-01 1.44584644e+00 4.52199221e-01
3.62736315e-01 1.14658192e-01 -2.97996640e-01 -1.32564080e+00
4.41322923e-01 -1.26331604e+00 3.24423850e-01 -6.72857881e-01
-1.76893711e+00 5.61456621e-01 -7.93683752e-02 -1.14904511e+00
3.97455007e-01 -2.84025043e-01 -4.67297733e-01 8.27298820e-01
-1.33702362e+00 -1.80752838e+00 -3.54361087e-01 9.52743351e-01
-3.62111814e-02 -3.08894336e-01 7.59054840e-01 4.33195919e-01
-6.29583299e-01 1.50576138e+00 3.34772289e-01 4.67766434e-01
9.69894469e-01 -8.39655042e-01 8.58696640e-01 8.59398901e-01
1.38691619e-01 7.55355656e-01 2.46942863e-01 -5.95089853e-01
-1.58443880e+00 -1.80331838e+00 5.50032198e-01 -9.63308930e-01
2.10414246e-01 -9.05995846e-01 -9.70501065e-01 9.70179141e-01
3.02846193e-01 1.63656339e-01 7.34112501e-01 -2.75784820e-01
-9.33069885e-01 -6.62175357e-01 -1.73955464e+00 7.41884887e-01
1.30991793e+00 -7.36467123e-01 -2.27014691e-01 3.08793247e-01
1.02140260e+00 -3.18848372e-01 -1.12657189e+00 3.08567882e-01
5.70960939e-01 -9.08908248e-01 1.11357510e+00 -5.03712773e-01
4.95113999e-01 -9.40024257e-02 -3.52808863e-01 -1.03190279e+00
-2.51426995e-01 -1.17606342e+00 1.14486478e-01 2.01916552e+00
1.63185760e-01 -9.42803979e-01 1.04508173e+00 1.06313097e+00
2.76179403e-01 -6.80220783e-01 -1.24064744e+00 -9.56134081e-01
2.71883160e-01 -1.36667088e-01 1.28516805e+00 1.20430148e+00
-6.87104404e-01 -2.22358525e-01 -1.09005427e+00 1.62576512e-01
9.36339736e-01 8.64301696e-02 1.09188604e+00 -8.15893054e-01
-3.62644196e-02 -8.43921378e-02 -5.15631676e-01 -9.66122627e-01
2.92056203e-01 -1.04641533e+00 -4.14551437e-01 -7.46536672e-01
4.13596898e-01 -2.19606563e-01 -2.36123472e-01 9.17043090e-01
-6.42297342e-02 7.11430132e-01 3.43675673e-01 3.30783337e-01
-3.13955814e-01 8.52124214e-01 1.32632494e+00 -2.95301378e-01
1.97729021e-01 -1.54379308e-01 -1.19844961e+00 2.63773173e-01
4.62743312e-01 -5.04547596e-01 -4.94002581e-01 -6.30878210e-01
-1.57107726e-01 -3.71084720e-01 8.45376134e-01 -8.70573640e-01
4.58630137e-02 -1.75078437e-01 5.35513818e-01 -5.57820439e-01
6.25021040e-01 -1.10559022e+00 6.02892280e-01 4.24647741e-02
-3.00395697e-01 4.25199186e-03 2.94115365e-01 6.64616942e-01
-7.46691003e-02 5.06664574e-01 8.87281477e-01 1.05207741e-01
-4.80989993e-01 9.56260920e-01 3.10891330e-01 1.42115578e-01
9.90070820e-01 -4.03716952e-01 -7.14135766e-01 -4.59472239e-01
-1.79836661e-01 3.50666434e-01 9.18788970e-01 5.13910234e-01
6.40176237e-01 -1.77333248e+00 -8.60352695e-01 4.18467224e-01
2.72358119e-01 -6.68614879e-02 3.05491745e-01 6.03652775e-01
-9.22456309e-02 4.59477715e-02 -4.48515117e-01 -3.68910670e-01
-1.48518670e+00 6.68658614e-01 3.09384912e-01 1.41865075e-01
-8.73022616e-01 9.98153150e-01 5.49325407e-01 -3.22910994e-01
1.73225388e-01 -8.83208364e-02 6.59381270e-01 -2.56087482e-01
6.98001683e-01 8.14233199e-02 -1.02129236e-01 -7.39879131e-01
-5.04559278e-01 2.65034735e-01 -2.91587293e-01 9.84764099e-02
1.35827124e+00 -2.13634863e-01 -4.65099961e-01 -2.13510737e-01
1.35901523e+00 -9.90527868e-02 -1.47145617e+00 -3.69321793e-01
-4.59727734e-01 -1.12658954e+00 -2.80721784e-01 -7.54584134e-01
-1.61818731e+00 5.44912517e-01 4.77913678e-01 -3.73272270e-01
1.38120627e+00 -1.70389608e-01 1.26746857e+00 9.09851491e-02
4.82001215e-01 -6.35345280e-01 -3.16316485e-02 -1.63518399e-01
1.07683957e+00 -1.07947040e+00 9.76425502e-03 -6.61470592e-01
-6.48313642e-01 5.47228277e-01 6.55728281e-01 7.07101002e-02
7.29587376e-01 1.26385033e-01 1.30757652e-02 4.44280356e-02
-5.48055351e-01 6.47246122e-01 2.62316227e-01 7.90282667e-01
-3.74861747e-01 -2.82675098e-03 3.27236503e-01 8.34256053e-01
-4.40184027e-01 8.15114006e-02 1.50530547e-01 7.11632788e-01
4.87158179e-01 -1.23032260e+00 -3.73716176e-01 4.92368996e-01
-3.80635768e-01 -1.82683051e-01 -8.13748956e-01 7.86916077e-01
2.22910181e-01 7.31065094e-01 -1.61102451e-02 -7.99630523e-01
1.09246694e-01 -2.15604883e-02 4.35458958e-01 -8.99347439e-02
-7.25354195e-01 -5.28886795e-01 -1.35792673e-01 -8.27103913e-01
-1.32133186e-01 -5.84381700e-01 -2.69459099e-01 -1.23162472e+00
3.51698697e-03 -8.27796832e-02 2.54004866e-01 6.05272949e-01
1.02395856e+00 -8.36272389e-02 7.50245690e-01 -5.73938429e-01
-5.36202788e-01 -6.27506077e-01 -6.64795756e-01 6.83529794e-01
3.13145518e-01 -3.62311780e-01 -1.32349521e-01 4.89483178e-02] | [12.702186584472656, 0.5652596354484558] |
f15666d9-7b8a-48b8-8228-eeef4f0b7c39 | reprogramming-large-pretrained-language | 2210.07144 | null | https://arxiv.org/abs/2210.07144v2 | https://arxiv.org/pdf/2210.07144v2.pdf | Reprogramming Pretrained Language Models for Antibody Sequence Infilling | Antibodies comprise the most versatile class of binding molecules, with numerous applications in biomedicine. Computational design of antibodies involves generating novel and diverse sequences, while maintaining structural consistency. Unique to antibodies, designing the complementarity-determining region (CDR), which determines the antigen binding affinity and specificity, creates its own unique challenges. Recent deep learning models have shown impressive results, however the limited number of known antibody sequence/structure pairs frequently leads to degraded performance, particularly lacking diversity in the generated sequences. In our work we address this challenge by leveraging Model Reprogramming (MR), which repurposes pretrained models on a source language to adapt to the tasks that are in a different language and have scarce data - where it may be difficult to train a high-performing model from scratch or effectively fine-tune an existing pre-trained model on the specific task. Specifically, we introduce ReprogBert in which a pretrained English language model is repurposed for protein sequence infilling - thus considers cross-language adaptation using less data. Results on antibody design benchmarks show that our model on low-resourced antibody sequence dataset provides highly diverse CDR sequences, up to more than a two-fold increase of diversity over the baselines, without losing structural integrity and naturalness. The generated sequences also demonstrate enhanced antigen binding specificity and virus neutralization ability. Code is available at https://github.com/IBM/ReprogBERT | ['Devleena Das', 'Inkit Padhi', 'Amit Dhurandhar', 'Payel Das', 'Pin-Yu Chen', 'Vijil Chenthamarakshan', 'Igor Melnyk'] | 2022-10-05 | null | null | null | null | ['text-infilling'] | ['natural-language-processing'] | [ 5.15667379e-01 -3.80345225e-01 -4.89342175e-02 -4.02500331e-01
-6.89070702e-01 -9.34620142e-01 3.16848904e-01 9.24577117e-02
-4.70897287e-01 1.25926352e+00 4.66918387e-02 -6.51929736e-01
4.12578940e-01 -4.68375832e-01 -1.18419492e+00 -7.68155277e-01
2.37800434e-01 5.76636016e-01 -4.84557264e-02 -6.25105739e-01
1.25583112e-01 6.19589984e-01 -1.16353059e+00 4.84892517e-01
1.07391322e+00 3.38625222e-01 4.67050463e-01 6.64922953e-01
-4.87449497e-01 4.82328624e-01 -5.19702196e-01 -3.02774280e-01
1.14371851e-01 -6.10031486e-01 -7.12212026e-01 -4.56611305e-01
2.81193078e-01 -3.36758164e-03 2.57469118e-01 6.09563529e-01
6.26112461e-01 -2.14585349e-01 5.86594522e-01 -6.89832330e-01
-1.21690440e+00 8.36154222e-02 -4.52002138e-01 2.51039684e-01
4.93355602e-01 5.05320847e-01 8.28045130e-01 -8.44968557e-01
9.06595588e-01 9.69524801e-01 6.75213337e-01 9.57404077e-01
-1.59807277e+00 -7.11107433e-01 1.16146140e-01 4.27261181e-02
-1.36212671e+00 -1.94150984e-01 4.37464118e-01 -7.08911598e-01
1.37157905e+00 2.57280260e-01 5.34874439e-01 1.47754061e+00
7.11416364e-01 3.55001658e-01 9.98114944e-01 -1.42325014e-01
4.76617180e-02 -1.85698662e-02 -4.02501598e-02 6.43794298e-01
-7.99230263e-02 1.56514913e-01 -4.17540044e-01 -2.82157749e-01
4.28912610e-01 2.75902808e-01 -1.54648021e-01 -5.52360356e-01
-1.02058768e+00 9.99221504e-01 5.68575621e-01 1.95868105e-01
-5.06990016e-01 -4.45213139e-01 2.35487670e-01 5.50020874e-01
-1.09811239e-02 7.65954554e-01 -7.94780016e-01 1.79395288e-01
-2.38004327e-01 3.37053806e-01 5.93097866e-01 8.87471795e-01
8.06953251e-01 -1.25864416e-01 -1.32747337e-01 8.58635306e-01
-1.37800321e-01 5.56977153e-01 4.84739810e-01 -5.23040164e-03
9.73797962e-03 5.21683276e-01 7.85907358e-02 -6.37208939e-01
-6.39658868e-01 -6.83295310e-01 -8.18708301e-01 -3.96204777e-02
1.98685169e-01 -9.53971129e-03 -1.09079373e+00 2.16299725e+00
4.47977334e-01 -2.09334984e-01 2.45612800e-01 8.52269471e-01
7.43743360e-01 7.71575212e-01 2.60629773e-01 -9.15583447e-02
1.42746627e+00 -9.75520432e-01 -2.99799412e-01 -3.10435474e-01
3.67021739e-01 -9.33773458e-01 1.16346562e+00 3.43244106e-01
-6.71950698e-01 -5.55736721e-01 -1.21057689e+00 -7.36392662e-02
-5.00734270e-01 -6.05376959e-01 5.77098668e-01 4.62668896e-01
-9.88037765e-01 1.37968630e-01 -4.68988150e-01 -3.25087875e-01
2.12616414e-01 6.00495756e-01 -6.62636518e-01 -1.57560781e-01
-1.26768100e+00 1.12187862e+00 4.67030942e-01 -1.54010952e-01
-9.30139899e-01 -1.11015677e+00 -6.00046098e-01 -4.19334561e-01
1.88520625e-01 -1.01811254e+00 1.07095158e+00 -1.07266557e+00
-1.56179357e+00 7.36551046e-01 -1.53523609e-01 -4.44283396e-01
2.45316774e-01 -2.89129303e-03 -4.36992824e-01 -3.77200454e-01
-4.24751714e-02 8.95073175e-01 6.09242916e-01 -9.05240059e-01
-5.66352785e-01 -1.78981170e-01 4.09735180e-02 -1.83365613e-01
1.66849971e-01 1.99355092e-02 -2.55357385e-01 -7.39961386e-01
-5.27561009e-01 -9.87516403e-01 -4.19008851e-01 -5.16335189e-01
-4.40106019e-02 3.08172125e-02 2.30079412e-01 -7.35006452e-01
9.06539500e-01 -1.64723444e+00 5.15645266e-01 1.78505853e-01
1.76431775e-01 5.31513453e-01 -6.08526289e-01 7.94175386e-01
-2.95831740e-01 -2.13101320e-02 -3.31744432e-01 4.19218093e-01
-2.79846787e-01 5.26815839e-02 -1.72799259e-01 2.28700221e-01
6.53476596e-01 1.18543375e+00 -8.41990769e-01 9.54617038e-02
-5.08422703e-02 7.92968214e-01 -8.31812382e-01 4.05824155e-01
-5.88092506e-01 9.51197267e-01 -4.02469993e-01 5.68990827e-01
7.84789205e-01 -4.58534390e-01 5.70427775e-01 -8.96026939e-02
1.63282350e-01 1.70204356e-01 -5.41756809e-01 1.64973092e+00
-6.74047589e-01 6.63121641e-02 -3.17445010e-01 -6.18867040e-01
1.00034475e+00 -1.67288519e-02 2.90393203e-01 -1.12849486e+00
-7.64172450e-02 2.10875422e-01 4.21324819e-01 -4.08533603e-01
2.73905601e-02 -5.38790703e-01 -4.36633267e-02 2.46447265e-01
-1.22887641e-01 2.19547838e-01 3.44719775e-02 6.28893524e-02
1.14950252e+00 5.87825254e-02 4.03567225e-01 -1.85473919e-01
8.38647008e-01 8.85385573e-02 7.29162753e-01 4.25001144e-01
1.39647484e-01 4.21058685e-01 2.84679413e-01 -7.60732472e-01
-1.37866557e+00 -1.19580543e+00 -6.54382408e-02 1.33689916e+00
-2.35211402e-01 -5.21173663e-02 -5.64793110e-01 -6.24232471e-01
1.70906842e-01 5.18973649e-01 -6.74157619e-01 -2.75530726e-01
-7.14634836e-01 -8.88342977e-01 4.61535156e-01 2.72811264e-01
-1.65375486e-01 -1.07227290e+00 -2.33845577e-01 4.39793944e-01
5.69846891e-02 -7.07590580e-01 -7.77708888e-01 2.28367761e-01
-3.78513187e-01 -8.84975374e-01 -7.03902662e-01 -1.05885601e+00
6.63018525e-01 -2.26223525e-02 1.26746154e+00 1.53848737e-01
-4.38692093e-01 -3.79434854e-01 -2.20754296e-01 -4.83087033e-01
-7.73441672e-01 3.81804794e-01 1.60319582e-01 -1.49418786e-01
6.85662270e-01 -4.48667258e-01 -7.13561118e-01 2.81788558e-01
-1.13061571e+00 2.19155177e-01 9.55008566e-01 1.23440194e+00
7.83581495e-01 -6.52833700e-01 1.12890029e+00 -1.26286685e+00
6.00378931e-01 -4.60621864e-01 -5.55765033e-01 3.37522596e-01
-5.34309745e-01 4.43832457e-01 1.00752628e+00 -5.07113814e-01
-8.52387249e-01 1.79875061e-01 -6.56067908e-01 1.47497460e-01
1.73225906e-02 3.53443921e-01 -2.17614427e-01 -1.85429230e-01
7.98034430e-01 4.20097321e-01 4.00135219e-01 -4.79107291e-01
3.58347595e-01 6.17523849e-01 2.93389231e-01 -6.70951426e-01
5.38524628e-01 1.58837125e-01 -2.67398179e-01 -5.32368481e-01
-6.26966715e-01 -2.95445621e-01 -3.95414591e-01 4.64568555e-01
7.65293479e-01 -7.99997151e-01 -7.89760590e-01 3.02088052e-01
-1.03005016e+00 -3.26530278e-01 3.01447242e-01 1.84961587e-01
-2.89398611e-01 2.19771832e-01 -6.60883009e-01 -9.27033350e-02
-6.52684987e-01 -1.53627539e+00 8.33192766e-01 -9.80155542e-03
-4.30318981e-01 -8.35165679e-01 5.22020042e-01 4.78689671e-01
6.80060565e-01 3.35236758e-01 1.37701201e+00 -8.36239994e-01
-4.84241247e-01 8.41040313e-02 4.72551733e-02 3.37149352e-01
5.13263702e-01 -3.66512805e-01 -5.67052364e-01 -7.14363277e-01
-3.04167032e-01 -4.62395728e-01 7.05506682e-01 -4.31074686e-02
7.58320928e-01 -2.39650458e-01 -2.77054220e-01 6.09988928e-01
1.36445522e+00 5.19757390e-01 4.98385668e-01 3.50366324e-01
6.64514601e-01 5.53909719e-01 3.76219481e-01 4.78230743e-03
2.05282062e-01 7.41025209e-01 9.14450809e-02 -2.77895302e-01
9.46207643e-02 -3.16435307e-01 2.47345448e-01 7.01990306e-01
2.27808222e-01 -2.57333398e-01 -9.46601093e-01 3.69653434e-01
-1.52077198e+00 -7.31055439e-01 2.63624340e-01 2.04953909e+00
1.44963908e+00 -1.10894047e-01 1.90417871e-01 -8.12551439e-01
5.60731471e-01 -2.85002142e-01 -1.00911510e+00 -5.80019712e-01
-4.13276047e-01 5.94704568e-01 3.85416567e-01 7.39635766e-01
-6.59409761e-01 9.96869087e-01 6.35603333e+00 7.57848501e-01
-1.40238357e+00 6.43077940e-02 5.49145758e-01 -3.14847946e-01
-5.91535866e-01 -1.51742071e-01 -8.07322502e-01 5.96723735e-01
1.03486419e+00 9.78262052e-02 4.79137540e-01 4.80680496e-01
1.23784663e-02 6.13854766e-01 -1.21289146e+00 7.37154663e-01
3.15590799e-02 -1.78105259e+00 4.55455214e-01 3.55637223e-02
7.53273010e-01 1.21445142e-01 1.87453896e-01 3.44871432e-01
4.46289450e-01 -1.32790542e+00 2.80174762e-01 5.05016148e-01
7.30870605e-01 -9.71879244e-01 7.37462521e-01 1.58818007e-01
-7.20417798e-01 2.27882490e-01 -4.20188755e-01 1.91840649e-01
6.63659275e-02 2.50872940e-01 -1.14057183e+00 4.49533522e-01
4.21636671e-01 2.84860134e-01 -5.85620999e-01 4.09595191e-01
3.84305939e-02 1.03439897e-01 4.33745533e-02 -2.04854935e-01
1.61318138e-01 -3.02747756e-01 1.96446970e-01 1.33308041e+00
-1.67228598e-02 -5.57809621e-02 3.37003320e-01 6.26107693e-01
-2.55554259e-01 2.85293043e-01 -5.16770422e-01 -1.81938395e-01
2.81743735e-01 1.01706648e+00 -1.48174182e-01 -8.60995576e-02
-5.71529448e-01 1.04752898e+00 7.13592350e-01 3.76857847e-01
-1.01337659e+00 -2.00117230e-01 1.19956243e+00 -5.08427396e-02
4.39497530e-01 8.51343349e-02 1.24951154e-01 -8.13063920e-01
-1.76658452e-01 -1.72342920e+00 4.16699201e-01 -5.22713304e-01
-1.64015043e+00 7.87268400e-01 -5.77865362e-01 -7.09279478e-01
-2.00308800e-01 -7.48456717e-01 -2.39002816e-02 1.21518898e+00
-1.36732900e+00 -1.28690553e+00 2.43174702e-01 2.13536739e-01
5.96538007e-01 -4.90669131e-01 7.93879569e-01 5.09128690e-01
-5.75443268e-01 9.09721732e-01 3.30276221e-01 -3.52819651e-01
1.03505814e+00 -9.98230398e-01 7.28261292e-01 5.33700824e-01
-3.29928607e-01 1.19675148e+00 8.82454813e-01 -8.55940342e-01
-1.71627343e+00 -1.13504517e+00 7.48222589e-01 -8.12664211e-01
6.47723019e-01 -6.52337432e-01 -1.29655206e+00 6.09162152e-01
1.69253632e-01 -5.92504442e-01 1.18948424e+00 2.94680037e-02
-7.43544877e-01 4.87453490e-02 -1.09973872e+00 7.83152640e-01
1.14175606e+00 -6.03534043e-01 -4.44267690e-01 3.98170590e-01
1.02278209e+00 -4.84803200e-01 -7.76575685e-01 4.59103554e-01
7.87081003e-01 -7.53030598e-01 1.17817390e+00 -1.32648695e+00
2.53561169e-01 -5.92811763e-01 -8.20091590e-02 -1.26359272e+00
-7.50873804e-01 -4.82754141e-01 3.05956960e-01 9.29978788e-01
9.18917120e-01 -8.90322626e-01 7.29099929e-01 1.90112591e-01
-9.27844942e-02 -9.02473509e-01 -6.40781879e-01 -8.03252935e-01
3.49004805e-01 1.29883245e-01 1.00190234e+00 9.53403294e-01
-2.37711504e-01 8.81476879e-01 -4.65740442e-01 7.71457404e-02
5.94056882e-02 2.64336050e-01 9.96184230e-01 -8.01117361e-01
-5.70377409e-01 -4.26346779e-01 -1.76593572e-01 -9.53682840e-01
2.37007618e-01 -1.22401762e+00 -8.03577304e-02 -1.31368995e+00
4.75536853e-01 -2.89100260e-01 -4.86453295e-01 3.61095279e-01
-3.85095537e-01 1.36747420e-01 -8.88265073e-02 -5.36941513e-02
-3.06479990e-01 4.45489705e-01 1.16854107e+00 -3.68071347e-01
-2.00694367e-01 -3.34260166e-01 -1.00494945e+00 1.78017661e-01
8.80476475e-01 -4.18401927e-01 -3.65200818e-01 -4.55513895e-01
4.19565886e-01 -3.86147290e-01 3.81940678e-02 -4.35277820e-01
-1.13684244e-01 -4.29412454e-01 5.16145408e-01 -3.14084828e-01
1.56205118e-01 -6.37327850e-01 9.07142580e-01 9.14185524e-01
-1.66763052e-01 5.73481023e-01 5.04453778e-01 5.45989573e-01
1.24507956e-01 1.45149440e-01 8.49401653e-01 -5.33618554e-02
-7.09134400e-01 3.14584464e-01 -4.06273156e-01 -1.05349980e-02
8.60949039e-01 -2.18009055e-01 -3.82417470e-01 9.92414057e-02
-2.98004270e-01 1.80501878e-01 9.03467119e-01 8.31855059e-01
5.86021602e-01 -1.04410577e+00 -9.70164478e-01 5.62585354e-01
5.34433186e-01 -3.53233963e-01 5.17735064e-01 4.19157475e-01
-8.87330115e-01 6.47229671e-01 -4.83970374e-01 -5.82269609e-01
-1.19411731e+00 1.06729329e+00 4.68351126e-01 -2.57529825e-01
-2.52870917e-01 8.83196175e-01 7.39169538e-01 -9.36286151e-01
-2.80053526e-01 5.01018465e-02 5.80713302e-02 -2.77458996e-01
6.08391047e-01 -2.69941032e-01 3.13978821e-01 -5.42420685e-01
-6.61739528e-01 5.44715822e-01 -7.03488588e-01 4.48951572e-01
1.35985720e+00 1.54247850e-01 -2.35930756e-01 -2.75196936e-02
1.27647471e+00 5.82143217e-02 -1.17353511e+00 -4.16740805e-01
2.28408929e-02 -2.32413635e-01 -5.61269403e-01 -1.15181005e+00
-6.07030690e-01 5.41549027e-01 7.71121740e-01 -5.42174518e-01
9.77331400e-01 -8.45004544e-02 9.71555293e-01 4.91800487e-01
5.45157611e-01 -7.09755242e-01 6.44004121e-02 6.45459354e-01
9.42530513e-01 -1.24637389e+00 -3.03252459e-01 -1.33920284e-02
-7.14534819e-01 5.88711321e-01 7.00935185e-01 2.93318093e-01
1.84435248e-01 3.24720204e-01 3.75666916e-01 -6.35298863e-02
-9.50291753e-01 9.41256061e-02 2.79405475e-01 8.45523417e-01
7.41303742e-01 8.25682059e-02 -3.33214968e-01 5.32556534e-01
-2.57677048e-01 -6.19370043e-02 -1.57783981e-02 9.82182324e-01
-2.74398804e-01 -1.71487427e+00 -1.40985921e-01 1.38060167e-01
-4.66593027e-01 -2.76309520e-01 -7.69035101e-01 5.57622194e-01
1.39637887e-01 6.85560763e-01 -7.73937777e-02 -3.86177182e-01
5.82880497e-01 1.06844373e-01 5.38285077e-01 -6.39049590e-01
-7.80586064e-01 -9.21076238e-02 -1.13543555e-01 -2.45212242e-01
-3.22860852e-02 -2.70146132e-01 -1.39157224e+00 -4.37038422e-01
1.04750857e-01 2.60197192e-01 4.20817733e-01 6.86994314e-01
8.71714115e-01 6.86812282e-01 4.51715112e-01 -2.73113310e-01
-3.91851634e-01 -6.92554057e-01 -7.63190314e-02 4.59062576e-01
5.76468170e-01 -4.38751310e-01 2.04215363e-01 9.54821035e-02] | [4.742801189422607, 5.592767238616943] |
1d201ebb-7797-4e2f-aaca-f32026b37a3e | data-quality-estimation-framework-for-faster | null | null | https://aclanthology.org/2022.ecnlp-1.4 | https://aclanthology.org/2022.ecnlp-1.4.pdf | Data Quality Estimation Framework for Faster Tax Code Classification | This paper describes a novel framework to estimate the data quality of a collection of product descriptions to identify required relevant information for accurate product listing classification for tax-code assignment. Our Data Quality Estimation (DQE) framework consists of a Question Answering (QA) based attribute value extraction model to identify missing attributes and a classification model to identify bad quality records. We show that our framework can accurately predict the quality of product descriptions. In addition to identifying low-quality product listings, our framework can also generate a detailed report at a category level showing missing product information resulting in a better customer experience. | ['Nicolas Nicolov', 'Allen Williams', 'Ravi Kondadadi'] | null | null | null | null | ecnlp-acl-2022-5 | ['code-classification', 'attribute-value-extraction'] | ['computer-code', 'natural-language-processing'] | [-1.37388885e-01 3.16677839e-02 -5.93072116e-01 -8.01905096e-01
-1.34760296e+00 -6.68873787e-01 -1.71502516e-01 1.08106446e+00
3.49392235e-01 4.56189513e-01 1.69173002e-01 -2.51681447e-01
-3.51889759e-01 -1.40949607e+00 -4.87335473e-01 1.65632516e-01
3.15016985e-01 6.54433370e-01 -1.18601725e-01 -5.42065427e-02
5.44535279e-01 4.88601714e-01 -1.82299471e+00 8.38826120e-01
1.27780855e+00 1.37489653e+00 -1.14351258e-01 4.80924219e-01
-5.90313315e-01 1.05991042e+00 -7.17355549e-01 -8.96846890e-01
2.95534730e-01 4.54992428e-02 -1.01843941e+00 4.38974530e-01
4.03208554e-01 -1.95686787e-01 4.73473549e-01 1.14604867e+00
-8.19144845e-02 -3.05175781e-01 6.40786290e-01 -1.74274528e+00
-6.03638709e-01 3.18451047e-01 7.64495134e-02 -1.86079025e-01
8.38535786e-01 1.21358633e-01 1.54667246e+00 -1.02081740e+00
6.29644513e-01 7.96306074e-01 6.20478272e-01 9.87198390e-03
-1.28322661e+00 -4.48055953e-01 -2.34147787e-01 3.57654810e-01
-1.43894589e+00 -3.70252877e-01 5.88575602e-01 -7.53571928e-01
1.06723344e+00 2.04198822e-01 6.62423670e-01 1.84583366e-01
2.33509839e-01 5.62804103e-01 7.73308635e-01 -3.15688312e-01
5.09192407e-01 5.44304669e-01 4.92189974e-01 5.80041528e-01
7.44696498e-01 -5.60892895e-02 -5.78585327e-01 -2.68620789e-01
2.36478284e-01 1.39465541e-01 3.89930427e-01 -3.35981965e-01
-6.17654324e-01 8.24955940e-01 1.35010645e-01 7.38109946e-02
-8.59285235e-01 -7.95845240e-02 3.95409316e-01 6.07113242e-01
2.93557316e-01 8.77323508e-01 -7.80112207e-01 -1.84912488e-01
-7.22611487e-01 5.23769140e-01 1.13707113e+00 1.48051679e+00
1.15019321e+00 -2.54110038e-01 -4.55069453e-01 5.30351460e-01
3.30096155e-01 4.36730117e-01 1.79836329e-03 -1.09164631e+00
5.76585233e-01 1.26992178e+00 6.02848828e-01 -9.78722990e-01
-2.17687726e-01 -3.43648672e-01 -6.95239529e-02 2.00085685e-01
7.00653940e-02 4.61577892e-01 -5.91350734e-01 9.64685619e-01
-1.16964132e-01 -5.49472690e-01 2.74619639e-01 3.57713252e-01
8.70859802e-01 4.07885551e-01 2.29189545e-01 -2.24384561e-01
1.61374199e+00 -4.63822514e-01 -1.07407856e+00 1.66595548e-01
1.03402150e+00 -8.11344922e-01 1.09586012e+00 8.20100486e-01
-1.12614274e+00 -6.67700768e-01 -9.95055914e-01 -8.25779811e-02
-6.93842530e-01 4.19758074e-02 8.02108109e-01 7.99890101e-01
-5.36623776e-01 7.56250858e-01 -3.00062150e-01 -6.40757149e-03
6.51628792e-01 2.10018247e-01 -4.28935260e-01 -4.38123226e-01
-1.00476277e+00 9.00419176e-01 4.95115936e-01 -5.29806495e-01
-7.29908526e-01 -1.03917897e+00 -1.17433918e+00 3.14149916e-01
3.50276530e-01 -7.43646026e-01 1.38262737e+00 -7.28021801e-01
-7.23694324e-01 3.97575289e-01 -1.49898261e-01 -3.61920208e-01
-3.22714031e-01 1.12584196e-01 -1.04564464e+00 -9.33983773e-02
4.06526119e-01 2.91184425e-01 3.25974464e-01 -1.17755353e+00
-1.23006999e+00 -6.79655433e-01 -6.96587786e-02 3.80737730e-03
9.16067325e-03 -2.69440021e-02 -1.14144385e-01 -2.67100304e-01
1.69771150e-01 -2.23909646e-01 -1.50624871e-01 -3.67636025e-01
-1.78041697e-01 -4.47102487e-01 1.28202394e-01 -7.51527786e-01
1.91489899e+00 -1.86417675e+00 -5.76637983e-01 5.03430903e-01
2.86699086e-01 -1.95923939e-01 -2.22820342e-01 6.33690178e-01
-7.17059895e-02 4.28220600e-01 1.69424042e-01 8.86926055e-02
1.17409088e-01 1.22904047e-01 -7.10226372e-02 -1.59350917e-01
6.53382838e-01 9.02616858e-01 -8.08837116e-01 -4.19381469e-01
3.13572064e-02 -8.17380846e-02 -6.08994305e-01 4.85649139e-01
-3.74226868e-01 -1.96036488e-01 -3.44087422e-01 1.32827258e+00
8.17135334e-01 -2.99578980e-02 -2.09225059e-01 -5.43724358e-01
6.37044311e-02 6.73489511e-01 -1.46820760e+00 1.31404054e+00
-6.79980755e-01 4.29687276e-02 -3.94509077e-01 -5.11382282e-01
1.39169049e+00 2.04091415e-01 7.35308230e-01 -9.55434144e-01
-1.35762617e-01 2.48194173e-01 -3.34471822e-01 -7.58713961e-01
8.42443168e-01 -8.24523941e-02 -4.99285698e-01 2.83644289e-01
1.08919866e-01 -2.21313149e-01 3.93905640e-01 2.20742196e-01
1.23265398e+00 -1.98894143e-01 5.05294979e-01 5.97686134e-03
5.21733403e-01 5.99111021e-01 7.74525106e-01 3.35935116e-01
-2.43029684e-01 3.71693969e-01 4.42698687e-01 -6.69385612e-01
-1.41590428e+00 -8.34080815e-01 -8.43407139e-02 6.20982707e-01
-7.36929476e-02 -1.14174259e+00 -4.48915422e-01 -7.65563965e-01
4.41784471e-01 9.85224247e-01 -2.64830559e-01 -3.52708191e-01
1.11952439e-01 -1.47808805e-01 -6.24341778e-02 5.38698614e-01
9.70926359e-02 -9.73778248e-01 -7.48917237e-02 5.09847999e-01
-2.27962583e-01 -7.84180999e-01 -2.49424100e-01 2.15684399e-01
-7.52298474e-01 -1.28154802e+00 2.99584627e-01 -6.16636872e-01
5.73314369e-01 -1.85908034e-01 1.83884621e+00 8.78746435e-02
6.40126457e-03 1.05252892e-01 -4.88775164e-01 -5.97157001e-01
-8.27856362e-01 -5.58318011e-02 -1.79569632e-01 -1.28963962e-01
1.11390901e+00 6.97180033e-02 -3.54556292e-01 3.99475604e-01
-9.03222442e-01 -5.14642656e-01 4.67700362e-01 4.69889581e-01
9.34816480e-01 8.30855191e-01 8.00827742e-01 -1.04932749e+00
9.94372129e-01 -5.48760653e-01 -7.49181807e-01 4.34301883e-01
-1.57345045e+00 3.01592082e-01 4.11219418e-01 2.13825852e-01
-9.12119269e-01 3.66375118e-01 -6.08224452e-01 2.37471148e-01
-4.17414010e-01 6.83563411e-01 -5.26311338e-01 3.21976572e-01
6.76813304e-01 -2.32310131e-01 -2.37199351e-01 -8.83769929e-01
2.60625780e-01 1.08500326e+00 5.00070572e-01 -3.37187618e-01
5.99684000e-01 3.27743888e-02 -3.47937569e-02 1.11639373e-01
-9.43679988e-01 -9.70875442e-01 -9.60554540e-01 -1.87419932e-02
4.43629324e-01 -9.81979370e-01 -9.37380314e-01 -2.44513944e-01
-8.96503031e-01 5.61195672e-01 -7.66778409e-01 2.54989117e-01
-6.84701800e-01 -1.21877208e-01 -2.84340322e-01 -9.34861183e-01
-2.33015269e-01 -1.02323639e+00 1.11959994e+00 -1.86972126e-01
-5.06348789e-01 -5.29211938e-01 3.72940190e-02 7.13791907e-01
1.43650115e-01 1.27313539e-01 1.26780689e+00 -7.70183742e-01
-7.66003907e-01 -6.24289095e-01 -1.67616233e-01 4.07728165e-01
8.39701965e-02 3.72829288e-02 -5.63301444e-01 -3.43001969e-02
-2.95262069e-01 -7.39201084e-02 4.38528240e-01 1.98687360e-01
8.45967829e-01 -6.78714871e-01 -9.57207605e-02 1.81461066e-01
1.81466174e+00 5.20980895e-01 8.76615644e-01 3.39763284e-01
3.33061993e-01 8.50479126e-01 1.52121687e+00 6.44856155e-01
5.18318832e-01 8.62689734e-01 4.95974034e-01 4.33606580e-02
1.40285343e-01 -4.95328248e-01 1.46075875e-01 2.95114666e-01
5.00202417e-01 7.35893846e-02 -9.10517573e-01 7.49210596e-01
-1.73513949e+00 -9.27936852e-01 -4.39363211e-01 2.29590154e+00
6.84867799e-01 2.65750170e-01 5.80110431e-01 5.22015989e-01
3.41706157e-01 -7.33694315e-01 -3.63472909e-01 -9.18672860e-01
2.92469382e-01 -9.00666490e-02 7.34939218e-01 3.58480334e-01
-5.79577923e-01 5.93498409e-01 7.63712502e+00 5.79285741e-01
1.62232339e-01 6.67815581e-02 3.72433603e-01 4.08853441e-01
-9.08684731e-01 9.11144614e-02 -1.07928729e+00 3.87398064e-01
1.15097451e+00 -5.05247056e-01 4.20437008e-02 1.21381235e+00
2.24899650e-01 -1.63139984e-01 -1.22425699e+00 9.50971663e-01
-1.48927063e-01 -1.49704707e+00 4.25386131e-01 3.28208804e-01
5.76317132e-01 -7.25054324e-01 -1.89453796e-01 2.33067468e-01
5.60025513e-01 -1.02058601e+00 5.97609580e-01 8.16095352e-01
7.91521847e-01 -1.41286492e+00 1.23344064e+00 1.47948191e-01
-1.29559183e+00 -4.47291881e-01 -6.03567183e-01 -1.60551071e-03
4.79096957e-02 1.05300438e+00 -1.24600673e+00 7.87746131e-01
7.66844809e-01 6.89865112e-01 -8.31572950e-01 1.24582493e+00
1.93501666e-01 1.99491367e-01 4.16388631e-01 2.27827162e-01
-2.92071164e-01 -1.00777879e-01 5.65320663e-02 9.12803352e-01
4.35939848e-01 -5.75190634e-02 1.69118822e-01 1.10438204e+00
-8.90429020e-02 3.47463578e-01 -6.92798138e-01 -1.64355472e-01
6.14172161e-01 1.08522785e+00 -1.39502361e-01 -4.03727889e-01
-4.98098731e-01 6.18518651e-01 1.05790213e-01 -1.20770581e-01
-2.01052144e-01 -5.81138432e-01 9.67639983e-01 4.44803566e-01
1.88081056e-01 9.32883099e-02 -6.48601234e-01 -8.62675369e-01
1.39252514e-01 -7.76412487e-01 7.01371670e-01 -9.41635013e-01
-1.49756861e+00 5.57883024e-01 -4.13660854e-01 -1.66018629e+00
-3.80899668e-01 -3.51464033e-01 3.82394306e-02 9.04035270e-01
-1.38738859e+00 -9.33132291e-01 -3.41701210e-01 3.18699032e-01
3.16743344e-01 -2.94912636e-01 8.38810205e-01 4.10257548e-01
-2.15652496e-01 3.72122318e-01 -2.47016251e-01 9.88464803e-02
4.24130827e-01 -1.62740564e+00 3.04028988e-01 5.70944786e-01
6.11782707e-02 4.68848407e-01 6.73403084e-01 -9.12982047e-01
-1.15190828e+00 -1.26368904e+00 1.65081275e+00 -1.00806689e+00
4.03835177e-01 -1.80324882e-01 -1.04235852e+00 3.39248091e-01
-2.57287294e-01 -3.23462218e-01 1.05307114e+00 3.36124510e-01
-6.27137542e-01 -8.15876544e-01 -1.47992206e+00 -2.04488233e-01
6.43723547e-01 -5.27809262e-01 -5.37938952e-01 2.42328167e-01
9.13192749e-01 1.57985449e-01 -1.69144297e+00 2.16475710e-01
4.17041421e-01 -8.18911493e-01 7.28432775e-01 -7.56486773e-01
2.37156048e-01 -6.61858976e-01 -1.99074879e-01 -1.33850956e+00
-6.04998052e-01 -1.90149263e-01 -1.13857426e-02 1.47306252e+00
8.35079789e-01 -2.49504391e-02 8.74953985e-01 1.24667132e+00
2.13207994e-02 -4.01850462e-01 -5.40871680e-01 -1.02448332e+00
-3.54234099e-01 -7.36993849e-01 1.39962232e+00 6.81825042e-01
2.86879450e-01 4.44451183e-01 -1.04720622e-01 2.68863708e-01
4.59872782e-01 3.33764434e-01 6.49834454e-01 -1.64961362e+00
1.86401561e-01 -1.85074091e-01 -6.98171437e-01 -3.73578995e-01
-1.38856605e-01 -9.57849860e-01 -1.79779887e-01 -1.98018837e+00
3.19808841e-01 -5.43048382e-01 -3.06034923e-01 4.25751001e-01
1.46267757e-01 1.43228590e-01 -2.13224277e-01 1.85139537e-01
-7.78186560e-01 -3.20174508e-02 7.77758539e-01 -2.14792401e-01
-1.97135314e-01 4.92008775e-01 -1.15993440e+00 1.79747924e-01
6.96596861e-01 -8.93088996e-01 -1.58838615e-01 1.44627199e-01
1.12172127e+00 1.14413351e-01 1.49272084e-01 -8.39291334e-01
-7.17843510e-03 -3.80609781e-01 3.20373893e-01 -1.00847888e+00
-3.99785131e-01 -1.34316838e+00 5.06407142e-01 4.92998600e-01
-5.20080090e-01 4.55038399e-01 -1.63278461e-01 3.55776101e-01
-6.07313931e-01 -6.71451330e-01 3.97902161e-01 -1.67713523e-01
-9.01637137e-01 3.54508549e-01 -2.31199309e-01 -4.78639543e-01
8.80295932e-01 -2.76987791e-01 -1.59515962e-01 -2.33524412e-01
-8.43989432e-01 1.86362550e-01 6.04373574e-01 4.35539633e-01
6.53829396e-01 -1.76924801e+00 -7.38901913e-01 3.96874040e-01
1.14644468e+00 -4.32164878e-01 -1.51058108e-01 3.88320424e-02
-2.79032677e-01 6.42672718e-01 -3.09239537e-01 -1.81653857e-01
-1.10195541e+00 9.86079991e-01 1.59777194e-01 -4.98463094e-01
-5.82016148e-02 8.67882669e-02 -5.61860323e-01 -2.99780697e-01
1.82752684e-01 -4.68002856e-01 -5.20369112e-01 1.11453809e-01
7.61038601e-01 4.96855527e-01 8.65204871e-01 -5.68470776e-01
-3.95978332e-01 1.77312180e-01 2.26087980e-02 8.56407210e-02
1.15068710e+00 -3.69522452e-01 1.01870157e-01 1.77445859e-01
1.29752445e+00 -7.24319071e-02 -8.52676153e-01 -2.55458146e-01
6.95260167e-01 -8.86961937e-01 1.98500529e-01 -1.31529725e+00
-8.69208694e-01 4.89928633e-01 5.18437922e-01 5.39330006e-01
1.29031515e+00 1.64206758e-01 6.64553404e-01 1.87710539e-01
5.48855186e-01 -1.37378931e+00 6.07817732e-02 -7.55638331e-02
8.21122766e-01 -1.51017523e+00 -1.14048021e-02 -7.05695629e-01
-8.97556901e-01 1.05490077e+00 5.49173295e-01 3.65369767e-01
6.76128209e-01 3.01937014e-01 -1.07459351e-01 -6.24527812e-01
-1.00543547e+00 -6.65475011e-01 5.77702641e-01 1.16144609e+00
3.11897159e-01 1.77838072e-01 -2.02950671e-01 1.08020818e+00
-1.69509694e-01 2.83633262e-01 7.40787148e-01 7.32125878e-01
-7.39315927e-01 -1.48390746e+00 -2.50055134e-01 8.88629377e-01
-4.15050000e-01 -1.93735212e-01 -3.58776689e-01 3.92543644e-01
4.40669656e-01 1.33368528e+00 2.68209249e-01 -5.60527146e-01
1.08510876e+00 1.69779375e-01 1.61083162e-01 -9.79772270e-01
-7.91772723e-01 -2.76365757e-01 4.91954446e-01 -8.63325238e-01
-2.64680147e-01 -6.76770270e-01 -1.05898595e+00 -4.64211881e-01
-3.45298260e-01 4.36846912e-01 8.69747341e-01 5.73902369e-01
7.30933845e-01 4.47693169e-01 9.99010801e-01 4.26885009e-01
-3.35096002e-01 -6.83015764e-01 -1.00653028e+00 9.49612260e-01
3.04853797e-01 -4.40874130e-01 8.71191081e-03 2.37519532e-01] | [9.942928314208984, 6.284664154052734] |
256a235c-166d-4f8f-9936-8b71a899f51c | a-simple-yet-effective-corpus-construction | null | null | https://aclanthology.org/2022.lrec-1.742 | https://aclanthology.org/2022.lrec-1.742.pdf | A Simple Yet Effective Corpus Construction Method for Chinese Sentence Compression | Deletion-based sentence compression in the English language has made significant progress over the past few decades. However, there is a lack of large-scale and high-quality parallel corpus (i.e., (sentence, compression) pairs) for the Chinese language to train an efficient compression system. To remedy this shortcoming, we present a dependency-tree-based method to construct a Chinese corpus with 151k pairs of sentences and compression based on Chinese language-specific characteristics. Subsequently, we trained both extractive and generative neural compression models using the constructed corpus. The experimental results show that our compression model can generate high-quality compressed sentences on both automatic and human evaluation metrics compared with the baselines. The results of the faithfulness evaluation also indicated that the Chinese compression model trained on our constructed corpus can produce more faithful compressed sentences. Furthermore, a dataset with 1,000 pairs of sentences and ground truth compression was manually created for automatic evaluation, which, we believe, will benefit future research on Chinese sentence compression. | ['Akiko Aizawa', 'Masayasu Muraoka', 'Issei Yoshida', 'Hiroshi Kanayama', 'Yang Zhao'] | null | null | null | null | lrec-2022-6 | ['sentence-compression'] | ['natural-language-processing'] | [ 2.62562424e-01 -3.62223014e-02 -9.91887003e-02 -5.43811917e-01
-1.17298937e+00 8.06947891e-03 4.85119790e-01 8.66369456e-02
-4.42098618e-01 8.36622298e-01 9.36503947e-01 -3.15958172e-01
4.58157957e-01 -9.55531895e-01 -7.02120543e-01 -2.95802146e-01
9.51503813e-02 3.41933995e-01 1.37084480e-02 -2.92527020e-01
4.65144098e-01 -2.61320472e-01 -1.18491435e+00 5.37390232e-01
1.31768310e+00 6.08887792e-01 7.98301637e-01 7.88178384e-01
-1.64050668e-01 6.93734765e-01 -8.98581505e-01 -7.71317124e-01
1.10412508e-01 -9.89439368e-01 -1.00923097e+00 -2.36719530e-02
-4.51822728e-02 -5.72906435e-01 -4.75368023e-01 1.15807283e+00
4.62101251e-01 -2.41711631e-01 3.79408032e-01 -3.85101259e-01
-8.79111588e-01 1.25831413e+00 -3.46305281e-01 4.95568626e-02
3.69070500e-01 -1.65372565e-02 9.45016682e-01 -6.56892598e-01
7.29012311e-01 1.09731054e+00 3.41508329e-01 5.36515296e-01
-7.73282409e-01 -5.31388879e-01 -4.40508127e-01 1.91974416e-01
-1.30262172e+00 -6.03140116e-01 6.14779413e-01 3.03015709e-01
1.35948086e+00 4.47374970e-01 7.13756800e-01 8.97786379e-01
4.84269530e-01 7.43065536e-01 6.92120790e-01 -9.23374593e-01
6.27454296e-02 3.64643522e-02 -1.91056743e-01 3.17112654e-01
3.21525097e-01 -1.27759174e-01 -2.26934031e-01 2.44150415e-01
2.91360110e-01 -3.33361149e-01 -2.94247091e-01 4.65985030e-01
-1.01110923e+00 9.31579053e-01 1.35592014e-01 6.72512472e-01
-2.16822520e-01 -9.85353068e-02 8.25609446e-01 3.06307733e-01
5.05464792e-01 5.98285079e-01 -1.21093526e-01 -6.06986761e-01
-1.40848279e+00 2.30206683e-01 8.67174745e-01 1.58390832e+00
3.84824276e-01 1.57912806e-01 -2.63544470e-01 9.45050180e-01
4.16958928e-02 6.32501423e-01 9.19850767e-01 -8.71124268e-01
1.10028803e+00 2.44837388e-01 -4.08388734e-01 -1.08232367e+00
1.86433390e-01 -2.84220815e-01 -1.22629690e+00 -8.34654331e-01
-3.14500600e-01 -1.22507200e-01 -5.13551712e-01 1.62249625e+00
-3.31971616e-01 -2.77369857e-01 5.11133015e-01 6.71694994e-01
4.94803756e-01 1.06649220e+00 -2.51742005e-01 -6.93010747e-01
1.00909424e+00 -1.15087104e+00 -1.11393070e+00 -1.41038910e-01
9.64723468e-01 -1.06006622e+00 1.33679581e+00 9.78747308e-02
-1.56619072e+00 -6.66260123e-01 -1.29759014e+00 -1.40205100e-01
1.41953886e-01 -1.67486956e-03 3.22395384e-01 7.51428545e-01
-9.37065959e-01 5.78743398e-01 -8.80270839e-01 -2.81243324e-01
2.46389106e-01 -2.64841467e-01 -1.44788533e-01 -4.59874898e-01
-1.37085736e+00 9.09979701e-01 7.87362516e-01 -1.93107665e-01
-6.38326347e-01 -2.46280909e-01 -9.50172961e-01 3.65176469e-01
7.48617342e-03 -4.06216294e-01 1.37416172e+00 -6.18619382e-01
-1.29637122e+00 3.87357056e-01 -2.43165106e-01 -9.08989310e-01
1.07933320e-01 -3.01852971e-01 -5.88304400e-01 4.67566222e-01
3.26144346e-03 6.86531186e-01 4.41182852e-01 -9.09698486e-01
-5.34722984e-01 3.83851714e-02 -3.71736616e-01 2.51654506e-01
-6.71853483e-01 2.27310166e-01 -9.21932042e-01 -1.02572715e+00
-9.95527655e-02 -6.42618477e-01 -2.42772803e-01 -7.33460903e-01
-6.29404485e-01 5.78458700e-03 7.69342661e-01 -1.19517004e+00
1.97083533e+00 -1.97481501e+00 -8.76361355e-02 -8.71504098e-02
-3.52322161e-01 4.78016764e-01 -3.96786064e-01 9.06388164e-01
2.70884156e-01 5.24054229e-01 -7.40518093e-01 -6.22422814e-01
-2.07897797e-01 2.55213231e-01 -4.86028522e-01 -3.24150026e-02
1.01514734e-01 1.02591372e+00 -7.50952125e-01 -1.03260756e+00
-1.58908412e-01 3.01321894e-01 -8.20660233e-01 5.35794497e-01
-1.75404102e-01 3.68165299e-02 -2.69624293e-01 5.09126961e-01
6.81501567e-01 1.26420110e-01 3.78126234e-01 1.20285921e-01
1.96540073e-01 6.89309478e-01 -4.35040981e-01 2.08182383e+00
-5.24933398e-01 6.40026391e-01 -3.24112415e-01 -8.52554679e-01
1.14418519e+00 4.03252125e-01 6.10623583e-02 -1.05922878e+00
9.94126275e-02 2.14466900e-01 6.53008372e-03 -6.61871433e-01
1.29426980e+00 -1.92779467e-01 -3.97241205e-01 6.19806945e-01
-8.19973350e-02 -7.15458989e-01 7.44682193e-01 6.19048655e-01
9.37457323e-01 -6.08806387e-02 8.47761184e-02 -2.64797974e-02
5.64445674e-01 3.45272161e-02 5.98560810e-01 5.30498385e-01
-1.76954493e-02 1.09235787e+00 3.35464627e-01 9.42010358e-02
-1.74530625e+00 -4.50266033e-01 -7.86980242e-02 3.98863107e-01
-1.62254363e-01 -1.08199191e+00 -1.25786734e+00 -2.62983114e-01
-6.37143493e-01 1.11620200e+00 1.23527952e-01 -4.41811323e-01
-1.00919998e+00 -5.85498631e-01 1.01081979e+00 5.66966176e-01
8.25185061e-01 -1.23441529e+00 -4.28393096e-01 3.14353168e-01
-8.87915432e-01 -1.28866696e+00 -8.69524121e-01 -3.01446468e-01
-1.08833063e+00 -7.19068050e-01 -5.52415490e-01 -1.07782769e+00
4.96888548e-01 2.89984614e-01 1.24199378e+00 3.55447382e-01
1.39589041e-01 -2.07967341e-01 -9.37704325e-01 -3.31465513e-01
-1.14888191e+00 3.25376123e-01 -1.31436586e-01 -7.13680208e-01
4.70040292e-01 -4.73506004e-01 -2.89916992e-01 -1.45236596e-01
-1.33259737e+00 3.18847656e-01 7.98634768e-01 8.13376248e-01
6.22833967e-01 2.54225075e-01 6.52923465e-01 -7.83129513e-01
1.16036355e+00 -1.14654206e-01 -3.19752961e-01 4.65779364e-01
-7.92127669e-01 2.39045531e-01 9.53318357e-01 -1.35025743e-03
-1.21517026e+00 -3.86519343e-01 -6.25319183e-01 -6.16147220e-02
1.44980073e-01 1.02901137e+00 -1.86778292e-01 6.72851861e-01
5.62420487e-01 8.07891190e-01 7.24601224e-02 -5.71124315e-01
3.02596092e-01 1.22092128e+00 8.62691581e-01 -5.41534722e-01
4.97303277e-01 -1.32894576e-01 -4.19582188e-01 -7.79715955e-01
-5.10793388e-01 4.46494967e-02 -5.31246543e-01 1.82309896e-01
7.58809805e-01 -1.04457378e+00 1.68398097e-01 3.10401440e-01
-1.29258645e+00 -4.61789742e-02 -2.58784741e-01 5.92578709e-01
-5.24449348e-01 1.02546287e+00 -1.13946331e+00 -6.01808727e-01
-9.87869799e-01 -1.10915303e+00 7.73661256e-01 -8.91165137e-02
-1.92012459e-01 -5.76433897e-01 1.64147496e-01 3.82772595e-01
5.32401800e-01 -2.16255292e-01 9.71279085e-01 -2.79606730e-01
-4.03880507e-01 -5.70716262e-01 -3.67516465e-02 7.73278058e-01
-5.86031042e-02 -1.23832887e-02 -5.00523329e-01 -4.04687047e-01
5.13032556e-01 -6.90806866e-01 8.46409261e-01 1.43935114e-01
1.45294058e+00 -5.75908303e-01 3.02942656e-02 4.98347580e-01
1.31828046e+00 2.37424642e-01 1.35772908e+00 2.30911933e-02
3.10502261e-01 3.39862615e-01 6.71013296e-01 2.91083276e-01
4.20208901e-01 5.17184734e-01 -2.34822389e-02 1.79457560e-01
-2.44104743e-01 -7.21683025e-01 4.07659680e-01 2.03700709e+00
6.59748241e-02 -4.59961146e-01 -5.77202022e-01 3.79766315e-01
-1.34848213e+00 -1.15612733e+00 4.13066596e-02 2.09543252e+00
1.35408986e+00 3.82585377e-01 -2.89674193e-01 2.68504828e-01
7.26847947e-01 1.16924100e-01 -9.81097892e-02 -6.26443565e-01
-3.73534739e-01 2.22803697e-01 2.40028232e-01 4.64539319e-01
-7.71259427e-01 9.86483693e-01 6.71200085e+00 1.12954831e+00
-9.48500752e-01 -2.89020725e-02 8.89481246e-01 3.97165716e-02
-6.34914696e-01 8.71639624e-02 -8.15590858e-01 7.25518823e-01
1.58890653e+00 -5.24997354e-01 3.31796646e-01 7.08949566e-01
3.09878141e-01 -5.46062812e-02 -7.42571115e-01 6.34019494e-01
3.59317243e-01 -1.56636393e+00 2.08025828e-01 -3.40637926e-04
8.26688468e-01 -1.32633865e-01 -4.13327783e-01 4.92798477e-01
-3.91140394e-02 -9.32440042e-01 7.40290105e-01 3.84633780e-01
1.13891351e+00 -9.22950089e-01 1.00125051e+00 8.04318786e-01
-1.09278429e+00 2.58510411e-01 -8.71004879e-01 -6.62053302e-02
6.14011467e-01 7.17744470e-01 -6.14354968e-01 8.26102257e-01
4.20777470e-01 7.91491926e-01 -6.44876897e-01 7.84509063e-01
-2.88617492e-01 1.08216119e+00 -9.94990617e-02 -1.42676994e-01
1.21938139e-01 -2.81029314e-01 2.05839872e-01 1.59379125e+00
5.74056506e-01 1.54493451e-01 -1.60122141e-01 8.13644171e-01
-2.78842509e-01 2.06531093e-01 -5.45976877e-01 -4.45425600e-01
6.45817578e-01 9.37878549e-01 -5.07511258e-01 -5.10535359e-01
-4.83202904e-01 9.63447511e-01 3.60119075e-01 8.31629857e-02
-6.76345825e-01 -6.67184770e-01 -3.99901085e-02 -1.89235777e-01
2.87186563e-01 -3.61589670e-01 -5.09473979e-01 -1.54882550e+00
2.28903890e-01 -1.04238057e+00 9.17684287e-02 -6.22630954e-01
-1.00561380e+00 9.46233511e-01 1.71774149e-01 -1.36535144e+00
-6.86639845e-01 2.41057724e-01 -7.07324207e-01 9.17903244e-01
-1.34687793e+00 -7.72316754e-01 7.92649239e-02 1.53074503e-01
8.70243490e-01 -2.05959156e-01 1.01279068e+00 4.05338794e-01
-5.90626359e-01 9.26431060e-01 1.28662363e-01 1.13599524e-01
4.18761611e-01 -7.37012446e-01 7.28136003e-01 1.36623955e+00
5.80253601e-02 6.79056585e-01 5.97764611e-01 -8.87115121e-01
-1.32505584e+00 -1.19589400e+00 1.64282024e+00 1.22516759e-01
8.97791609e-02 -4.58724916e-01 -1.09325171e+00 3.90271336e-01
8.30137253e-01 -6.83259547e-01 6.26262367e-01 -4.01356101e-01
1.49058998e-02 -1.53313801e-02 -9.67708826e-01 5.51165104e-01
9.37870085e-01 -4.95592207e-01 -7.67455399e-01 3.39892119e-01
1.45446157e+00 -4.86982495e-01 -8.45571280e-01 3.23722959e-01
1.69451237e-01 -8.85699093e-01 5.60939014e-01 -3.01006228e-01
1.36894834e+00 1.05344296e-01 -4.34875429e-01 -1.10665357e+00
-1.96150243e-01 -4.00732785e-01 -2.05041751e-01 1.46851921e+00
5.70347309e-01 -4.17253152e-02 6.12368762e-01 5.22917569e-01
-6.36968672e-01 -9.45015013e-01 -5.76515138e-01 -8.41052830e-01
4.56048161e-01 -5.76987743e-01 9.39417005e-01 3.89940351e-01
4.95852917e-01 5.12377024e-01 -6.63806617e-01 -5.17529547e-01
6.63642436e-02 2.17119083e-01 6.02432966e-01 -2.47739032e-01
-3.51754993e-01 -2.38200292e-01 5.16503714e-02 -1.35625899e+00
1.73129961e-01 -1.08131218e+00 2.89150923e-01 -1.50098670e+00
6.09653354e-01 -2.83112168e-01 3.54224533e-01 2.94369627e-02
-3.70554864e-01 -4.83805761e-02 3.76980066e-01 4.40091103e-01
-3.88466597e-01 1.11949086e+00 1.52307546e+00 -9.38265100e-02
5.44685908e-02 -2.38001302e-01 -9.08254981e-01 1.16727971e-01
8.68376553e-01 -3.96046489e-01 -4.03847665e-01 -8.06419551e-01
-1.13197058e-01 3.23214114e-01 -5.27263165e-01 -1.04735970e+00
2.00131744e-01 -1.24272928e-01 9.03249756e-02 -8.48279536e-01
6.40833154e-02 -5.10742307e-01 -1.09220268e-02 6.80840015e-01
-6.67498767e-01 3.49856555e-01 -2.13447753e-02 2.16766745e-01
-7.05138505e-01 -6.07074201e-01 7.73216069e-01 -2.50082105e-01
-1.18898481e-01 7.66992598e-05 -1.88832536e-01 3.12405497e-01
6.38528287e-01 -8.58858228e-02 -1.89662725e-01 -7.54621089e-01
1.88956141e-01 8.47295970e-02 4.88236040e-01 3.12269807e-01
9.74265516e-01 -1.35033917e+00 -1.11558092e+00 3.41976136e-01
-1.43243998e-01 1.01828843e-01 1.11504421e-01 3.17754388e-01
-9.00742888e-01 8.92039955e-01 -3.53829190e-02 -2.68657893e-01
-1.05616426e+00 5.99312186e-01 -1.89335823e-01 -6.09982848e-01
-7.92802930e-01 4.27777201e-01 -5.27952909e-01 -2.59640604e-01
1.60917848e-01 -3.92832398e-01 -3.36052515e-02 -6.69047058e-01
8.00375044e-01 1.28450707e-01 1.43776119e-01 -5.92030346e-01
1.49688035e-01 5.18880645e-03 -2.33250618e-01 -3.97225142e-01
1.38052082e+00 -2.76469201e-01 -3.11096966e-01 -1.55572653e-01
1.50910258e+00 9.26919468e-03 -6.69703364e-01 -1.08045392e-01
-1.35996997e-01 -7.03735054e-01 -1.68638140e-01 -4.40999359e-01
-1.02518582e+00 7.98790336e-01 5.47944047e-02 1.31270066e-01
1.50639665e+00 -2.54973769e-01 1.54940081e+00 5.30039132e-01
3.05899352e-01 -1.26484895e+00 -3.99694853e-02 8.19028735e-01
1.02208388e+00 -9.06990409e-01 4.53174077e-02 -3.39805275e-01
-8.01268578e-01 1.09454286e+00 5.37585676e-01 7.37308487e-02
3.64439428e-01 3.02253932e-01 -1.45324260e-01 2.73941249e-01
-8.12918961e-01 2.79321015e-01 -4.94298600e-02 4.53171283e-01
8.06497395e-01 9.43126604e-02 -1.13672495e+00 6.44746959e-01
-9.23797011e-01 8.35234858e-03 7.52458692e-01 1.18440163e+00
-6.19833827e-01 -1.52706432e+00 -1.29321352e-01 6.74292266e-01
-4.78698730e-01 -5.14392376e-01 -8.27708915e-02 4.04456466e-01
-2.88647115e-01 1.12056315e+00 2.03338519e-01 -6.81508899e-01
1.10641569e-01 -9.27104875e-02 2.18718484e-01 -6.13996804e-01
-4.29473817e-01 2.78443307e-01 1.70856506e-01 -2.40406945e-01
-3.11542928e-01 -5.20128727e-01 -1.22447407e+00 -5.26326478e-01
-5.20169199e-01 4.37477291e-01 5.46998382e-01 7.74809480e-01
3.82052690e-01 2.91089833e-01 7.80781984e-01 -5.21302879e-01
-7.47601688e-01 -1.49979496e+00 -4.73902434e-01 4.67623889e-01
-2.88099796e-01 5.21062553e-01 -6.06580600e-02 2.95473933e-01] | [12.162492752075195, 9.327052116394043] |
f93dd1f8-aee7-4e0d-8915-abc4ce6e0845 | data-free-evaluation-of-user-contributions-in | 2108.10623 | null | https://arxiv.org/abs/2108.10623v1 | https://arxiv.org/pdf/2108.10623v1.pdf | Data-Free Evaluation of User Contributions in Federated Learning | Federated learning (FL) trains a machine learning model on mobile devices in a distributed manner using each device's private data and computing resources. A critical issues is to evaluate individual users' contributions so that (1) users' effort in model training can be compensated with proper incentives and (2) malicious and low-quality users can be detected and removed. The state-of-the-art solutions require a representative test dataset for the evaluation purpose, but such a dataset is often unavailable and hard to synthesize. In this paper, we propose a method called Pairwise Correlated Agreement (PCA) based on the idea of peer prediction to evaluate user contribution in FL without a test dataset. PCA achieves this using the statistical correlation of the model parameters uploaded by users. We then apply PCA to designing (1) a new federated learning algorithm called Fed-PCA, and (2) a new incentive mechanism that guarantees truthfulness. We evaluate the performance of PCA and Fed-PCA using the MNIST dataset and a large industrial product recommendation dataset. The results demonstrate that our Fed-PCA outperforms the canonical FedAvg algorithm and other baseline methods in accuracy, and at the same time, PCA effectively incentivizes users to behave truthfully. | ['Chengfei Lv', 'Rongfei Jia', 'Lifeng Hua', 'Shaojie Tang', 'Fan Wu', 'Tie Luo', 'Zhenzhe Zheng', 'Hongtao Lv'] | 2021-08-24 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [-1.32823795e-01 -1.13886476e-01 -5.77927291e-01 -3.80924612e-01
-1.00967062e+00 -6.10427678e-01 2.94155747e-01 -1.29365325e-01
-8.91442597e-03 8.35603476e-01 -9.82744023e-02 -1.11898385e-01
-2.76954859e-01 -7.97677040e-01 -8.28450382e-01 -7.59314716e-01
7.06266388e-02 5.20916939e-01 -2.30188176e-01 1.22986749e-01
-7.89838135e-02 -9.66299921e-02 -1.27777326e+00 2.82718390e-01
1.07164621e+00 1.22038877e+00 -7.52505660e-02 5.03176212e-01
1.07641824e-01 8.86224926e-01 -4.56156462e-01 -1.12572277e+00
7.06864893e-01 -4.01013911e-01 -5.73713481e-01 -6.76983735e-03
2.27422956e-02 -6.03714705e-01 -1.06046595e-01 1.19424796e+00
5.18573284e-01 -5.70586957e-02 2.14943245e-01 -1.76410532e+00
-7.44485259e-01 1.14770484e+00 -4.64115858e-01 -3.50055873e-01
1.44552246e-01 2.92810768e-01 1.18585348e+00 -4.95704740e-01
2.73546457e-01 8.02028298e-01 5.77024102e-01 4.80880380e-01
-1.05158961e+00 -9.20302808e-01 2.80210041e-02 1.14872165e-01
-1.14384639e+00 -3.58785987e-01 9.91869092e-01 -3.10130239e-01
-7.91878067e-03 5.00945330e-01 5.07615983e-01 1.23722005e+00
-2.39541486e-01 1.09832501e+00 1.05618024e+00 -1.27596095e-01
7.08169103e-01 4.41097587e-01 2.45590564e-02 4.36223358e-01
5.17083824e-01 3.21837980e-03 -6.12540364e-01 -8.73151481e-01
2.84778446e-01 5.04510462e-01 -1.99964643e-01 -5.25006115e-01
-9.38873172e-01 7.78214276e-01 1.02615587e-01 9.55775157e-02
-6.34773672e-01 -6.76273135e-03 3.67359608e-01 3.67067844e-01
5.65248072e-01 1.13860592e-01 -9.29180205e-01 -2.52389610e-01
-8.00365865e-01 1.73040181e-01 1.06135607e+00 7.51058877e-01
7.80170023e-01 -2.24555925e-01 -1.85592398e-01 6.25417590e-01
3.15512598e-01 5.64521849e-01 6.10677958e-01 -1.35166419e+00
4.77989525e-01 7.07490981e-01 5.62707484e-01 -1.07301104e+00
4.15455520e-01 -6.23209536e-01 -8.48210335e-01 -2.40223631e-02
3.72381032e-01 -4.35402632e-01 1.96415797e-01 1.68241596e+00
4.85949010e-01 3.48787248e-01 -5.54973893e-02 9.31594253e-01
2.87895769e-01 3.48809391e-01 -2.85432488e-01 -4.55328763e-01
7.75770724e-01 -1.00772047e+00 -6.88679039e-01 3.41051638e-01
6.04335070e-01 -4.58698720e-01 1.17994535e+00 6.94846153e-01
-1.08005953e+00 -3.43109310e-01 -9.45268631e-01 5.38968384e-01
1.39562488e-01 2.13648543e-01 7.16418445e-01 1.14199328e+00
-7.33436525e-01 9.21752334e-01 -5.56171656e-01 2.67304569e-01
9.47988629e-01 6.57985687e-01 -3.45950097e-01 -1.57282144e-01
-8.90271902e-01 -3.54924873e-02 -1.42907247e-01 -3.40007424e-01
-9.86655354e-01 -7.01954901e-01 -3.22430909e-01 2.28338912e-01
5.39400518e-01 -7.48646796e-01 1.53780258e+00 -1.37516403e+00
-1.66445851e+00 4.85057771e-01 6.31003156e-02 -5.42720735e-01
8.83019507e-01 -1.08495437e-01 -3.26569825e-01 -1.49248615e-01
5.41681424e-02 9.48249642e-03 9.44596350e-01 -1.37497699e+00
-9.17634308e-01 -6.02027476e-01 9.18166619e-03 -9.27718505e-02
-9.28121686e-01 -1.25056326e-01 -3.77483457e-01 -4.50547129e-01
-3.58493924e-01 -8.88882697e-01 -3.17777336e-01 -1.73120558e-01
-3.20812047e-01 -1.99575216e-01 9.61602628e-01 -6.77998304e-01
1.18617976e+00 -2.11197972e+00 -3.79673421e-01 4.67995971e-01
4.78675961e-01 3.19049507e-01 -9.38068405e-02 4.41117659e-02
4.70636815e-01 2.81828910e-01 -1.60843302e-02 -5.71893334e-01
1.40289262e-01 2.34500930e-01 -1.56943962e-01 5.13982058e-01
-5.89098573e-01 8.30165267e-01 -1.08517206e+00 -3.01650256e-01
-1.65449351e-01 1.31377533e-01 -8.24282110e-01 3.99664402e-01
-1.82357088e-01 4.71778035e-01 -6.04932666e-01 7.47215450e-01
7.73976862e-01 -6.53532565e-01 5.90523243e-01 4.86441888e-02
4.54567432e-01 8.27662349e-02 -1.30626535e+00 1.39801002e+00
-3.89914751e-01 -8.26707110e-02 3.39270353e-01 -6.15386903e-01
1.03885078e+00 4.49640453e-01 9.58803713e-01 -3.05219591e-01
3.75292659e-01 3.43033969e-01 -3.13747585e-01 -2.52417803e-01
2.87409335e-01 5.50199509e-01 1.38961838e-03 1.19182611e+00
-7.30665177e-02 5.74639380e-01 -3.35961550e-01 3.49860668e-01
1.23833597e+00 -1.93681821e-01 3.01530421e-01 1.02914266e-01
5.28542221e-01 -4.45035160e-01 9.46187794e-01 8.29495728e-01
-4.47025508e-01 1.02269627e-01 3.22454929e-01 -4.32002604e-01
-8.60033453e-01 -6.42602742e-01 4.85799611e-01 1.17777288e+00
8.13316256e-02 -6.20555758e-01 -8.75051081e-01 -1.36695910e+00
4.09692407e-01 4.69898462e-01 -4.55128461e-01 -2.07259525e-02
-1.48809686e-01 -4.12159175e-01 2.65536875e-01 2.60307282e-01
7.28626847e-01 -5.97425222e-01 -1.85696527e-01 1.13446064e-01
-3.78053159e-01 -8.14262509e-01 -8.16760302e-01 -2.43428558e-01
-6.73143923e-01 -1.28030682e+00 -3.64241093e-01 -2.99251437e-01
7.33020842e-01 6.43166661e-01 8.22890818e-01 3.76980938e-03
5.45788407e-01 3.87869418e-01 -3.95486623e-01 -3.29190671e-01
-4.43990767e-01 5.22467215e-03 3.66625756e-01 6.33344412e-01
4.66769308e-01 -5.94876766e-01 -6.85572207e-01 8.63999069e-01
-5.76725304e-01 -3.10585767e-01 3.24593753e-01 9.68917370e-01
5.92252612e-01 2.84599483e-01 7.21652985e-01 -1.27873683e+00
7.65443742e-01 -7.43114412e-01 -6.43548191e-01 2.65741646e-01
-1.09242737e+00 -3.65342587e-01 8.66976440e-01 -5.81272840e-01
-1.05599153e+00 2.15945512e-01 2.50083208e-01 -7.49491870e-01
1.03727013e-01 1.85188413e-01 -6.11828923e-01 -1.41261056e-01
7.21990645e-01 -1.58879712e-01 7.72066861e-02 -5.98235488e-01
4.93289679e-01 1.23295248e+00 3.99234414e-01 -6.22470379e-01
6.85579360e-01 5.02315223e-01 -3.57627779e-01 3.77230463e-03
-6.85118735e-01 -4.16866809e-01 -2.36611158e-01 -2.40492642e-01
-8.82192329e-03 -1.02268565e+00 -1.02425599e+00 3.05329055e-01
-8.92187655e-01 -1.84093252e-01 -5.59024036e-01 4.69423741e-01
-5.06895959e-01 3.38263452e-01 -5.76209486e-01 -1.06852138e+00
-7.57286668e-01 -1.01265073e+00 5.68987012e-01 1.21502884e-01
-6.52395412e-02 -6.41496420e-01 3.35931145e-02 9.28647578e-01
4.50626045e-01 2.23888345e-02 3.42422247e-01 -9.26277339e-01
-5.96261799e-01 -5.78654826e-01 9.98298749e-02 6.44970536e-01
3.16742003e-01 -1.34974167e-01 -1.12708700e+00 -4.89315420e-01
2.47613981e-01 -2.90279239e-01 1.61795139e-01 1.58938587e-01
1.46444690e+00 -1.03478515e+00 -1.92131713e-01 4.11003500e-01
9.83823001e-01 -4.50274982e-02 2.52788872e-01 -4.95957807e-02
5.68133712e-01 4.33508277e-01 5.28078258e-01 8.17245841e-01
5.57817996e-01 5.62448382e-01 5.85719824e-01 2.06393868e-01
3.70322973e-01 -7.44300544e-01 4.70017821e-01 1.03081071e+00
-1.02947861e-01 4.26792651e-02 -3.10993433e-01 2.85226107e-01
-2.42789459e+00 -9.43131268e-01 -9.46944058e-02 2.58714342e+00
8.58403981e-01 -1.19401962e-01 3.75822097e-01 3.41721743e-01
8.15636933e-01 -2.72729844e-01 -7.72585213e-01 4.41405885e-02
-2.00093105e-01 -1.05052121e-01 6.28270268e-01 -5.24704382e-02
-8.78145218e-01 3.75994772e-01 5.58075142e+00 8.09021831e-01
-8.30693245e-01 8.50120723e-01 1.03782272e+00 -2.24065065e-01
-3.66467983e-01 2.19590981e-02 -4.22500610e-01 8.63199115e-01
9.51936185e-01 -4.66379732e-01 8.09122026e-01 1.54918993e+00
1.59506634e-01 3.55133891e-01 -9.84302878e-01 1.39995110e+00
-2.78611891e-02 -1.35650933e+00 -3.71650904e-01 3.77660066e-01
1.24064434e+00 1.40155122e-01 2.80492548e-02 3.85819107e-01
9.22274828e-01 -4.57120329e-01 6.56260133e-01 6.05470181e-01
4.80534762e-01 -8.25363159e-01 7.73643374e-01 6.72799587e-01
-7.18434513e-01 -6.17847979e-01 -3.76909494e-01 -2.53090188e-02
-1.87536880e-01 8.41372073e-01 -5.30664802e-01 5.82123280e-01
7.12786973e-01 6.77231312e-01 -3.06141406e-01 1.09594095e+00
-3.31002295e-01 1.06415951e+00 -7.49888271e-02 -3.09931133e-02
-3.29955459e-01 -2.85779238e-01 3.57306987e-01 5.32593668e-01
4.97927159e-01 -1.39579982e-01 2.29458138e-01 7.51516402e-01
-8.32100987e-01 3.68088573e-01 -3.58802468e-01 8.58782977e-02
7.27899730e-01 1.58213508e+00 -1.22174092e-01 -2.02820554e-01
-3.42817366e-01 8.78999770e-01 2.01123014e-01 1.21135347e-01
-6.25621140e-01 1.37849376e-01 7.18525887e-01 1.23348467e-01
2.99332887e-01 1.49568975e-01 -3.53059769e-01 -1.26525295e+00
8.44961554e-02 -1.12847066e+00 5.43881714e-01 -4.48325396e-01
-1.83002865e+00 1.88682690e-01 -5.97213328e-01 -1.35503888e+00
-2.33542144e-01 -1.07704543e-01 -5.85031033e-01 6.14990234e-01
-1.27524996e+00 -1.15759802e+00 -2.85240084e-01 1.06644869e+00
3.77666168e-02 -5.69960773e-01 8.64266753e-01 4.84202981e-01
-5.89925706e-01 1.02481711e+00 3.37772548e-01 -2.16226019e-02
7.56432235e-01 -9.25023496e-01 7.25083873e-02 6.11718714e-01
2.91491419e-01 5.72299063e-01 2.38511205e-01 -6.70671761e-01
-1.61055470e+00 -1.29163527e+00 6.81154191e-01 -6.47858620e-01
4.84734058e-01 -2.63836414e-01 -6.38177216e-01 5.34611166e-01
-1.18592471e-01 5.27395844e-01 1.18057287e+00 2.43942499e-01
-3.63250971e-01 -5.93747973e-01 -1.62976241e+00 1.87651142e-01
7.87186205e-01 -4.84861404e-01 1.47957072e-01 5.02201259e-01
8.61174285e-01 -1.39090970e-01 -8.19116056e-01 1.01469196e-02
5.53440750e-01 -9.12062168e-01 3.85285854e-01 -8.14889789e-01
4.28945944e-02 -3.02467078e-01 -3.95998776e-01 -1.15158498e+00
-4.05741930e-01 -1.28834045e+00 -8.17670107e-01 1.48627937e+00
1.84287727e-01 -7.34659612e-01 1.11975789e+00 8.86513352e-01
4.73972112e-01 -7.23478854e-01 -7.89498150e-01 -8.04478824e-01
-4.01616782e-01 -5.92745781e-01 1.34146488e+00 1.20160377e+00
6.99180812e-02 1.53493518e-02 -8.74596655e-01 -3.56754735e-02
8.42743158e-01 3.57007906e-02 1.19232786e+00 -1.40031135e+00
-8.24866951e-01 -3.41716073e-02 -4.39216420e-02 -7.28410661e-01
1.30112901e-01 -7.47172177e-01 -4.33587432e-01 -9.32471275e-01
4.02047366e-01 -8.72520089e-01 -6.24546230e-01 6.68745279e-01
-9.55268368e-02 4.51253019e-02 2.13952065e-01 7.73700655e-01
-1.00274885e+00 3.94800067e-01 9.06061351e-01 -2.14053869e-01
-3.04576844e-01 7.37111270e-01 -1.11516750e+00 4.77437854e-01
9.49639022e-01 -5.86217284e-01 -4.19639915e-01 -7.61850178e-02
3.32317293e-01 5.00571467e-02 3.16269606e-01 -6.42238975e-01
3.45198184e-01 -1.42930061e-01 2.21084788e-01 -2.28789553e-01
-1.30541533e-01 -1.06790435e+00 3.51929456e-01 3.21561545e-01
-2.31301203e-01 -1.79784670e-01 -6.57121718e-01 8.05500090e-01
2.10967317e-01 -2.39919475e-03 4.04958397e-01 -7.54523873e-02
-2.88976785e-02 6.29521549e-01 2.68002212e-01 -2.48309001e-01
1.16774130e+00 2.02006787e-01 -3.62189651e-01 -6.85268641e-01
-3.46177727e-01 2.98160464e-01 6.32594287e-01 3.47204566e-01
3.21445763e-01 -1.54271877e+00 -5.94839036e-01 1.98202386e-01
3.21275480e-02 -3.83149385e-01 2.06051454e-01 8.24103415e-01
1.79470167e-01 2.48792227e-02 1.57306856e-03 -3.98989052e-01
-1.19518888e+00 6.54284298e-01 1.54702857e-01 -2.85061508e-01
-7.32005611e-02 5.29231548e-01 -4.89127517e-01 -7.45260239e-01
4.35851306e-01 3.51287991e-01 1.29989907e-01 -3.11507612e-01
6.54406011e-01 6.59316719e-01 2.15348691e-01 -5.37954211e-01
-9.69048939e-04 -2.69816339e-01 8.15482363e-02 1.93223044e-01
1.25218022e+00 -1.83645815e-01 1.48037523e-02 8.28538314e-02
1.08258688e+00 4.32556033e-01 -1.44099736e+00 -6.46974206e-01
-3.55641037e-01 -9.30680454e-01 1.68628797e-01 -8.68857324e-01
-1.53351414e+00 3.10113758e-01 7.16328979e-01 2.65588462e-01
9.88705575e-01 -2.80173391e-01 1.02443230e+00 1.71637386e-01
9.29108024e-01 -1.12746930e+00 -4.42032106e-02 -1.78486377e-01
2.52679378e-01 -1.35932577e+00 -1.31601721e-01 -1.44960791e-01
-9.38463807e-01 6.58121526e-01 4.70856279e-01 2.78081805e-01
7.57136464e-01 1.17474258e-01 -7.45878518e-02 2.54065037e-01
-8.25028002e-01 2.94122607e-01 6.34714440e-02 5.93531072e-01
-8.18247497e-02 5.33286989e-01 -2.28589699e-01 1.64641333e+00
-5.96260317e-02 5.37881196e-01 2.44871989e-01 8.00840735e-01
-1.50006294e-01 -1.49834621e+00 -3.08838814e-01 7.97685802e-01
-5.98331392e-01 4.95432645e-01 -3.64868104e-01 -5.57867996e-02
4.29872602e-01 1.28752971e+00 -4.66001600e-01 -9.03335452e-01
6.83299005e-02 -2.38264292e-01 -3.46073061e-02 -3.00695479e-01
-8.61674428e-01 -4.73001003e-02 -1.99877664e-01 -7.57836282e-01
-4.09600616e-01 -6.60260677e-01 -7.77219355e-01 -6.27823174e-01
-6.35747552e-01 5.22699177e-01 8.50987911e-01 7.89260745e-01
7.54765093e-01 5.61429299e-02 1.55202460e+00 -5.79455912e-01
-1.24109077e+00 -5.57425976e-01 -8.10712695e-01 6.63197100e-01
-1.51047185e-01 -3.24949443e-01 -4.32000905e-01 -5.69534954e-03] | [5.863729476928711, 6.324940204620361] |
58bd4ed9-1719-425b-b2f9-7c227b910556 | finki-at-semeval-2016-task-4-deep-learning | null | null | https://aclanthology.org/S16-1022 | https://aclanthology.org/S16-1022.pdf | Finki at SemEval-2016 Task 4: Deep Learning Architecture for Twitter Sentiment Analysis | null | ['Gjorgji Strezoski', 'Ivica Dimitrovski', 'Gjorgji Madjarov', 'Dario Stojanovski'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['twitter-sentiment-analysis'] | ['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] | [-7.425168037414551, 3.8160078525543213] |
67f80593-b9fc-4ca9-a618-e14c867d4bf9 | aerk-aligned-entropic-reproducing-kernels | 2303.03396 | null | https://arxiv.org/abs/2303.03396v1 | https://arxiv.org/pdf/2303.03396v1.pdf | AERK: Aligned Entropic Reproducing Kernels through Continuous-time Quantum Walks | In this work, we develop an Aligned Entropic Reproducing Kernel (AERK) for graph classification. We commence by performing the Continuous-time Quantum Walk (CTQW) on each graph structure, and computing the Averaged Mixing Matrix (AMM) to describe how the CTQW visit all vertices from a starting vertex. More specifically, we show how this AMM matrix allows us to compute a quantum Shannon entropy for each vertex of a graph. For pairwise graphs, the proposed AERK kernel is defined by computing a reproducing kernel based similarity between the quantum Shannon entropies of their each pair of aligned vertices. The analysis of theoretical properties reveals that the proposed AERK kernel cannot only address the shortcoming of neglecting the structural correspondence information between graphs arising in most existing R-convolution graph kernels, but also overcome the problem of neglecting the structural differences between pairs of aligned vertices arising in existing vertex-based matching kernels. Moreover, unlike existing classical graph kernels that only focus on the global or local structural information of graphs, the proposed AERK kernel can simultaneously capture both global and local structural information through the quantum Shannon entropies, reflecting more precise kernel based similarity measures between pairs of graphs. The above theoretical properties explain the effectiveness of the proposed kernel. The experimental evaluation on standard graph datasets demonstrates that the proposed AERK kernel is able to outperform state-of-the-art graph kernels for graph classification tasks. | ['Edwin R. Hancock', 'Lu Bai', 'Yue Wang', 'Ming Li', 'Lixin Cui'] | 2023-03-04 | null | null | null | null | ['graph-classification'] | ['graphs'] | [-3.74193825e-02 1.34858370e-01 1.24114320e-01 9.43756942e-03
-2.71205753e-01 -5.42354524e-01 6.16110206e-01 6.78154826e-01
-2.68559366e-01 2.92376727e-01 -3.20935935e-01 -4.24599111e-01
-5.75211167e-01 -1.19352174e+00 -4.78062451e-01 -1.02997112e+00
-5.65316558e-01 1.35254860e-01 2.45655715e-01 -2.79269695e-01
1.52745068e-01 4.51926678e-01 -1.11392486e+00 -3.25461447e-01
1.03160608e+00 8.80416274e-01 -1.17127532e-02 6.92141712e-01
7.50389621e-02 4.61274058e-01 -9.92788747e-02 -4.86048818e-01
4.89035547e-02 -7.29908347e-01 -8.68738592e-01 -2.19849497e-01
2.39881799e-01 1.85379341e-01 -8.54494572e-01 1.47720623e+00
1.97092444e-01 2.67546415e-01 8.99055481e-01 -1.23597383e+00
-1.04740834e+00 6.53765917e-01 -2.13641346e-01 1.77947357e-01
5.10380089e-01 -7.49176592e-02 1.45809197e+00 -3.93745333e-01
5.10197282e-01 7.63834298e-01 6.78880334e-01 1.16688810e-01
-1.55341721e+00 -3.19511324e-01 -4.65345085e-01 3.64973456e-01
-1.70231843e+00 1.49758413e-01 9.93023217e-01 -4.42924321e-01
7.97912836e-01 3.58575583e-01 8.52818489e-01 5.22967339e-01
4.72278297e-01 1.58950210e-01 1.18144560e+00 -4.35119957e-01
2.45051593e-01 -6.45902753e-02 5.48237085e-01 1.15319455e+00
4.71382231e-01 9.94492620e-02 -1.15523405e-01 -3.81585240e-01
6.66835308e-01 -1.14066258e-01 -5.09858787e-01 -7.41558790e-01
-1.37394834e+00 8.27091634e-01 8.74725044e-01 6.05924189e-01
-2.65091360e-01 2.94019014e-01 4.05637920e-01 4.30688262e-01
2.14368284e-01 2.86915213e-01 2.64201283e-01 1.97892874e-01
-4.95590925e-01 -1.93119571e-01 9.54611063e-01 9.32415545e-01
1.39017487e+00 -3.29253405e-01 -1.46445796e-01 2.39872754e-01
1.51958987e-01 5.81691444e-01 1.20983459e-01 -4.13618207e-01
1.57806933e-01 8.00567329e-01 -2.88259923e-01 -1.30383706e+00
-4.54963863e-01 -2.92889208e-01 -1.23308766e+00 -2.18651459e-01
4.59482789e-01 4.06597078e-01 -4.78437275e-01 1.58732152e+00
2.43848637e-01 4.09200758e-01 1.30447686e-01 7.24069774e-01
8.92237663e-01 3.00441384e-01 -1.57729074e-01 -1.17877945e-01
1.24904263e+00 -5.30558407e-01 -5.46291709e-01 4.40911144e-01
8.38866293e-01 -4.94144171e-01 7.07618952e-01 -2.33431861e-01
-5.59492230e-01 -2.43011802e-01 -1.11250603e+00 3.09198588e-01
-5.69386482e-01 -3.75402346e-02 8.68120074e-01 8.74906600e-01
-1.24872255e+00 1.05019867e+00 -5.77466667e-01 -5.49566507e-01
-2.14392215e-01 2.18943760e-01 -7.12258816e-01 -4.10551280e-02
-1.40436625e+00 6.10403001e-01 4.93404746e-01 2.43455753e-01
-4.92071897e-01 -3.14789593e-01 -9.58698392e-01 1.78447172e-01
3.21269691e-01 -5.28832018e-01 5.44498146e-01 -5.31106770e-01
-1.45465517e+00 6.66356504e-01 2.06574649e-01 -3.07174653e-01
1.74225513e-02 5.87571144e-01 -5.11650801e-01 4.32444692e-01
-2.74588186e-02 -2.94278026e-01 7.92419314e-01 -7.28212297e-01
1.86112374e-01 -3.30734789e-01 1.94559857e-01 -2.47527510e-02
-2.05647260e-01 -3.45023274e-01 -1.19546115e-01 -2.02410772e-01
3.10686320e-01 -1.02420163e+00 -4.90621924e-02 -5.46631098e-01
-5.16461790e-01 -2.43511379e-01 3.23336869e-01 -3.34100127e-01
1.30374479e+00 -2.17132807e+00 2.91248113e-01 6.28801107e-01
6.38581097e-01 1.56468987e-01 -2.46089160e-01 1.09646583e+00
-3.11045170e-01 -3.16916630e-02 -2.69894063e-01 2.82729656e-01
1.01765521e-01 2.28818059e-02 7.09586814e-02 8.73469055e-01
1.06217928e-01 1.11532772e+00 -1.24638677e+00 -2.72120059e-01
3.82533401e-01 6.57156050e-01 -2.53336489e-01 -3.15812789e-02
3.76904160e-01 2.37286106e-01 -4.84001666e-01 1.43895641e-01
9.76049304e-01 -6.21395588e-01 3.81028205e-01 -3.91231149e-01
1.11847602e-01 1.17665857e-01 -1.13216734e+00 1.44910038e+00
-9.84692425e-02 2.73575306e-01 -2.14181617e-01 -1.21762860e+00
8.13655555e-01 1.34901643e-01 3.92567188e-01 -5.04191577e-01
7.85395503e-02 2.20793813e-01 1.86590999e-01 -1.87399499e-02
3.03251624e-01 -1.68974474e-02 -1.13341667e-01 5.49760044e-01
3.97131622e-01 2.29018897e-01 2.45818853e-01 7.16644526e-01
1.53701043e+00 -4.27408993e-01 5.30484796e-01 -6.35346711e-01
9.13186550e-01 -5.09259582e-01 -1.46031296e-02 8.21220458e-01
-3.15188080e-01 3.02279353e-01 7.41409779e-01 -2.67678440e-01
-7.92034268e-01 -1.37775493e+00 -7.26010650e-02 5.28212786e-01
4.89286453e-01 -7.63957262e-01 -8.69632065e-01 -6.61356151e-01
1.40137091e-01 2.18663216e-01 -7.87716985e-01 -5.94561458e-01
-1.90955121e-02 -7.79085815e-01 6.38372064e-01 5.60989566e-02
6.01641953e-01 -7.07176447e-01 -8.90360773e-02 1.17859721e-01
-1.30497441e-01 -1.13151658e+00 -7.31578708e-01 4.58537377e-02
-6.86298966e-01 -1.37572658e+00 -5.11436760e-01 -5.46752334e-01
6.15948617e-01 6.54538512e-01 8.38353157e-01 1.58037320e-01
-3.98164958e-01 7.42803812e-01 -6.41120493e-01 4.76726264e-01
-5.79505026e-01 7.49124214e-02 -4.36319364e-03 4.36282635e-01
2.34038010e-01 -7.29596615e-01 -6.64930820e-01 2.59268671e-01
-9.49644268e-01 -5.87989502e-02 5.93205333e-01 7.50762105e-01
4.63257372e-01 3.67424995e-01 2.33051315e-01 -6.96789980e-01
7.33219028e-01 -3.74859422e-01 -6.08283103e-01 5.73122501e-01
-7.55492866e-01 5.81780255e-01 7.47831643e-01 -2.28038177e-01
-4.30140406e-01 -2.25317016e-01 3.44665736e-01 -1.57579377e-01
1.59598812e-01 6.72544658e-01 9.92914513e-02 -7.70104527e-01
4.02577817e-01 6.36360407e-01 2.77333818e-02 -9.18530002e-02
5.71779072e-01 3.51362109e-01 2.45854080e-01 -5.15347540e-01
9.71298039e-01 4.75471199e-01 6.49089158e-01 -1.11073983e+00
-2.56484866e-01 -7.80179441e-01 -7.60662973e-01 -2.36328050e-01
9.35653985e-01 -4.92888808e-01 -1.21733153e+00 5.01112342e-01
-9.21959102e-01 1.14183865e-01 -1.38640851e-01 7.29704559e-01
-5.43460429e-01 1.06984651e+00 -9.19257700e-01 -9.32084382e-01
-3.71484816e-01 -9.58494127e-01 8.65384758e-01 3.90850753e-02
3.53927612e-01 -1.35700762e+00 3.86018574e-01 -1.10361159e-01
2.91149408e-01 3.37123454e-01 1.15558600e+00 -5.82540512e-01
-6.66321754e-01 -6.34601057e-01 -5.52630544e-01 2.28243664e-01
1.55663699e-01 -1.54985711e-01 -5.16827822e-01 -5.34642816e-01
-2.84785360e-01 1.35391474e-01 7.90533066e-01 5.96814230e-02
4.61416572e-01 -5.39629254e-03 -2.07567915e-01 5.52783012e-01
1.69498420e+00 -2.82021999e-01 6.25242710e-01 -1.88782722e-01
9.13440943e-01 2.72393495e-01 1.42308578e-01 2.89026856e-01
2.79839188e-01 5.52303672e-01 3.18159580e-01 2.46935681e-01
2.17476368e-01 -2.82341331e-01 3.55000943e-01 1.41744816e+00
-5.63325286e-01 5.49163893e-02 -7.53423333e-01 1.41174912e-01
-1.94127905e+00 -8.99473608e-01 -5.79620779e-01 2.71494532e+00
1.80646479e-01 -1.58562303e-01 8.16868432e-03 1.52687924e-02
9.92890060e-01 3.43523443e-01 -1.57146409e-01 -3.45703512e-01
-1.73486292e-01 4.63163227e-01 8.53931844e-01 5.30560613e-01
-9.35937047e-01 7.87539840e-01 5.38400841e+00 9.21423614e-01
-7.56371677e-01 1.20418161e-01 -2.17863128e-01 6.80673361e-01
-3.75951439e-01 5.80721140e-01 -1.10031605e-01 2.47939602e-01
8.38101029e-01 -4.84352350e-01 9.64562953e-01 6.98618472e-01
-2.70111829e-01 -9.56169367e-02 -1.00215852e+00 1.04963839e+00
-6.81299865e-02 -1.10777771e+00 1.18490808e-01 4.53510731e-01
4.70276862e-01 7.85249937e-03 -1.46396831e-01 2.17865527e-01
-1.60272308e-02 -9.43755567e-01 1.57070473e-01 6.39450550e-01
6.87209904e-01 -8.06253314e-01 7.78756380e-01 1.04185380e-01
-1.85342765e+00 3.84947866e-01 -6.00320816e-01 4.12301626e-03
-2.92203605e-01 8.63841772e-01 -6.04414821e-01 1.36106622e+00
1.84925243e-01 7.11148143e-01 -6.88400507e-01 8.96485031e-01
-1.17119417e-01 3.35771561e-01 -9.20124650e-02 -2.85211891e-01
3.55568141e-01 -1.10126579e+00 7.43938684e-01 9.64163065e-01
2.00397238e-01 -2.60214927e-03 -7.63760358e-02 1.00183237e+00
5.48241399e-02 4.32630509e-01 -9.21090901e-01 -5.79135656e-01
1.34176791e-01 1.55160940e+00 -9.53069687e-01 -2.21396551e-01
-3.78045082e-01 1.48946774e+00 6.89598680e-01 5.01682281e-01
-7.67571867e-01 -8.30610216e-01 5.11721253e-01 -1.80698261e-01
1.77588880e-01 -3.99932265e-01 4.11820203e-01 -1.44890714e+00
1.10951412e-04 -2.81510323e-01 2.99766570e-01 -4.13752198e-01
-1.34276831e+00 5.18704057e-01 -1.46548048e-01 -1.07951355e+00
2.89116740e-01 -7.18198717e-01 -6.41091466e-01 9.27872539e-01
-1.23794353e+00 -1.06081367e+00 -3.61871928e-01 6.97820246e-01
-7.40537226e-01 2.61970162e-01 1.15592086e+00 6.42126501e-02
-4.80738342e-01 5.55621505e-01 4.71995980e-01 3.81417096e-01
3.53863597e-01 -1.58026314e+00 5.23182929e-01 5.16697466e-01
3.22462916e-01 8.85123432e-01 5.04869103e-01 -7.19560206e-01
-1.94879901e+00 -7.36559212e-01 6.74994409e-01 -2.48647407e-01
1.04228747e+00 -4.97153014e-01 -9.25744057e-01 3.60640079e-01
-4.70132045e-02 3.80200446e-01 6.13752782e-01 1.47940561e-01
-7.98321187e-01 -2.94947531e-02 -9.37273562e-01 3.74835879e-01
1.11227000e+00 -1.14345181e+00 -2.45237425e-01 3.57882768e-01
5.91887295e-01 6.53530955e-02 -1.19214344e+00 4.28854674e-01
5.93863130e-01 -1.16552889e+00 8.02169681e-01 -5.37548184e-01
-1.06870838e-01 -3.12810719e-01 -7.90128037e-02 -1.29063523e+00
-7.59839356e-01 -6.76890612e-01 1.13485776e-01 8.90496075e-01
8.02421197e-02 -1.23563206e+00 3.78170490e-01 3.37826192e-01
2.16404572e-01 -4.96880352e-01 -1.10843527e+00 -1.12563145e+00
-1.53875068e-01 -2.81237364e-01 4.37481731e-01 1.08469307e+00
5.48486173e-01 3.44264984e-01 -1.59776121e-01 4.45873708e-01
9.75760221e-01 2.61856079e-01 4.92682308e-01 -1.32619226e+00
-4.66487497e-01 -5.40357351e-01 -1.30190289e+00 -4.27119613e-01
2.30379090e-01 -1.56057608e+00 -4.05677646e-01 -1.32333958e+00
4.79234934e-01 -1.64149627e-01 -6.16728902e-01 -5.54900579e-02
-1.97009817e-01 1.05432309e-01 5.48915528e-02 6.05860129e-02
-6.57229841e-01 6.02029979e-01 1.19432688e+00 8.77642911e-03
-6.18287362e-02 -2.44051829e-01 -9.06677097e-02 3.24378252e-01
4.47049886e-01 -2.98750579e-01 -2.95344591e-01 4.33183908e-01
3.99074852e-01 6.16111830e-02 7.17032850e-01 -1.16758049e+00
1.93734512e-01 1.36355564e-01 -2.28659526e-01 -1.59242302e-01
-3.64845730e-02 -5.83393395e-01 3.80359709e-01 6.98104143e-01
6.59758300e-02 -1.89915061e-01 -2.56879658e-01 1.13430047e+00
-1.40848622e-01 -2.07151815e-01 6.99373722e-01 1.10345751e-01
-5.07234633e-01 4.92143512e-01 -1.02970041e-01 -3.99787687e-02
1.09118736e+00 -3.94878015e-02 -3.80644560e-01 -3.71102512e-01
-6.16247058e-01 -2.46878520e-01 7.51818120e-01 -2.99898814e-03
6.65893137e-01 -1.60370100e+00 -3.07816535e-01 1.95234239e-01
5.32756865e-01 -7.08709002e-01 4.91387963e-01 1.31958902e+00
-6.06748521e-01 5.42355776e-01 -3.28543931e-02 -3.43469173e-01
-1.04303503e+00 7.83387423e-01 3.71175170e-01 -5.33154309e-01
-7.10160851e-01 3.86216938e-01 3.41176480e-01 -4.91918027e-01
-3.98564368e-01 -3.73276085e-01 6.66481629e-02 -1.87993467e-01
2.42713526e-01 4.07785445e-01 1.04902275e-01 -9.07789707e-01
-5.21325469e-01 7.89749324e-01 1.96040124e-01 1.17887773e-01
7.47330427e-01 3.18446159e-02 -4.99450803e-01 5.64837158e-01
1.56886661e+00 6.93515837e-02 -4.47853446e-01 -2.85312653e-01
-1.38443083e-01 -2.27733701e-01 -1.83734745e-01 -8.86354372e-02
-6.43121660e-01 8.01198065e-01 3.47596616e-01 7.43903339e-01
9.10889208e-01 2.04371661e-01 7.04705954e-01 3.81416440e-01
7.95206666e-01 -6.58617616e-01 -1.86317995e-01 4.69504803e-01
3.58470708e-01 -9.32631731e-01 -1.77088901e-01 -7.79094338e-01
-3.33959401e-01 1.35460496e+00 -3.62057686e-02 -4.26342279e-01
7.92418897e-01 -5.37474096e-01 -5.92736423e-01 -4.79600340e-01
-1.15583308e-01 -7.17793167e-01 6.40474379e-01 4.12678152e-01
3.59596044e-01 5.92044115e-01 -6.15076482e-01 1.35092586e-01
-4.01251093e-02 -4.53768551e-01 5.40187418e-01 5.47057986e-01
-3.24397981e-01 -1.09254086e+00 3.36135775e-02 4.35534596e-01
1.13404721e-01 -3.18873167e-01 -6.31603479e-01 7.76689827e-01
-3.72936070e-01 7.36946285e-01 -2.62826324e-01 -8.39302421e-01
5.62078506e-02 9.27579775e-02 7.40012288e-01 -4.74094391e-01
-2.31847137e-01 -3.66320223e-01 -2.57265449e-01 -6.14281535e-01
-3.62605751e-01 -2.34670818e-01 -1.29190302e+00 -4.49677110e-01
-6.18398547e-01 5.11695385e-01 5.26549935e-01 8.11901748e-01
3.20693910e-01 2.75442630e-01 6.51517510e-01 -5.16011298e-01
-5.94454288e-01 -7.48353302e-01 -1.30391848e+00 6.51058912e-01
2.21005678e-01 -6.81443155e-01 -7.81100452e-01 -6.49387002e-01] | [7.052361488342285, 5.693635940551758] |
ea637100-6259-47b0-9864-3b72d4e35503 | rono-robust-discriminative-learning-with | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Feng_RONO_Robust_Discriminative_Learning_With_Noisy_Labels_for_2D-3D_Cross-Modal_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Feng_RONO_Robust_Discriminative_Learning_With_Noisy_Labels_for_2D-3D_Cross-Modal_CVPR_2023_paper.pdf | RONO: Robust Discriminative Learning With Noisy Labels for 2D-3D Cross-Modal Retrieval | Recently, with the advent of Metaverse and AI Generated Content, cross-modal retrieval becomes popular with a burst of 2D and 3D data. However, this problem is challenging given the heterogeneous structure and semantic discrepancies. Moreover, imperfect annotations are ubiquitous given the ambiguous 2D and 3D content, thus inevitably producing noisy labels to degrade the learning performance. To tackle the problem, this paper proposes a robust 2D-3D retrieval framework (RONO) to robustly learn from noisy multimodal data. Specifically, one novel Robust Discriminative Center Learning mechanism (RDCL) is proposed in RONO to adaptively distinguish clean and noisy samples for respectively providing them with positive and negative optimization directions, thus mitigating the negative impact of noisy labels. Besides, we present a Shared Space Consistency Learning mechanism (SSCL) to capture the intrinsic information inside the noisy data by minimizing the cross-modal and semantic discrepancy between common space and label space simultaneously. Comprehensive mathematical analyses are given to theoretically prove the noise tolerance of the proposed method. Furthermore, we conduct extensive experiments on four 3D-model multimodal datasets to verify the effectiveness of our method by comparing it with 15 state-of-the-art methods. Code is available at https://github.com/penghu-cs/RONO. | ['Peng Hu', 'Xi Peng', 'Dezhong Peng', 'Hongyuan Zhu', 'Yanglin Feng'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['learning-with-noisy-labels', 'cross-modal-retrieval', 'learning-with-noisy-labels'] | ['computer-vision', 'miscellaneous', 'natural-language-processing'] | [-8.69554132e-02 -3.69767815e-01 -1.88366950e-01 -2.31158227e-01
-1.41639924e+00 -6.15266740e-01 4.09962088e-01 9.83038396e-02
-1.71043947e-01 2.34474853e-01 1.92238867e-01 2.82995790e-01
-5.36573231e-01 -2.47110084e-01 -3.44505876e-01 -1.04865992e+00
2.29493022e-01 2.61414915e-01 -1.67610884e-01 -1.54879261e-02
2.14931414e-01 2.21272588e-01 -1.54188895e+00 -1.70836300e-02
1.01527429e+00 1.26064444e+00 2.69406915e-01 -1.61958281e-02
-1.24366641e-01 2.50531614e-01 -4.54361439e-01 -2.61396676e-01
3.12042147e-01 -2.25612193e-01 -3.68379831e-01 3.51496071e-01
4.06004637e-01 -7.51078650e-02 -4.48340893e-01 1.38469183e+00
8.98423016e-01 3.91693026e-01 6.43890083e-01 -1.34880197e+00
-7.92273760e-01 2.49998286e-01 -7.77234018e-01 -2.03237176e-01
4.73865241e-01 -7.39959925e-02 1.09031951e+00 -1.30403352e+00
6.16500735e-01 1.34989035e+00 2.60812819e-01 4.62544709e-01
-1.09105861e+00 -6.03961110e-01 2.08339542e-01 1.54080391e-01
-1.73155415e+00 -2.51363039e-01 1.29801714e+00 -2.44938746e-01
7.83883408e-02 3.98210496e-01 4.26409960e-01 1.17629194e+00
-2.47400120e-01 1.18414068e+00 1.08249724e+00 -3.41733575e-01
8.19012597e-02 3.75216603e-02 4.16777879e-02 5.88767707e-01
9.02269706e-02 -1.87861234e-01 -6.35681272e-01 -2.67833382e-01
5.01216173e-01 5.51396087e-02 -3.70006710e-01 -5.56535661e-01
-1.27693021e+00 4.86880630e-01 4.18429554e-01 2.58975834e-01
-4.96946201e-02 -2.06191465e-01 4.40200001e-01 2.53073685e-03
5.26122093e-01 -5.07346392e-02 -4.33523729e-02 1.02091730e-01
-6.51769757e-01 1.24442354e-01 2.22333193e-01 1.11193883e+00
7.01730669e-01 -2.70398647e-01 -7.39462823e-02 1.17585289e+00
6.49711609e-01 7.54413307e-01 4.25334722e-01 -1.01232541e+00
6.18603349e-01 7.26574242e-01 9.12248865e-02 -1.52246416e+00
-3.33428651e-01 -4.86353427e-01 -9.82504129e-01 -4.69067395e-01
2.18197107e-01 2.48221308e-01 -6.74519718e-01 1.72151971e+00
7.20586538e-01 5.70328906e-02 -8.22001025e-02 1.46680033e+00
8.98231924e-01 6.09174550e-01 -7.86197409e-02 -1.95413858e-01
1.08340466e+00 -7.32096374e-01 -8.27215195e-01 -1.44923367e-02
6.34078026e-01 -9.42456603e-01 1.07335615e+00 2.99298704e-01
-8.11560690e-01 -5.63341677e-01 -9.05969381e-01 -3.19594920e-01
-2.70121217e-01 4.27910298e-01 3.04746360e-01 3.02826703e-01
-4.77438003e-01 -3.97471040e-02 -5.89593410e-01 -1.16485290e-01
2.98381865e-01 6.45590201e-02 -4.23863828e-01 -4.26322162e-01
-1.21799588e+00 5.65310061e-01 4.39385742e-01 5.31448126e-01
-7.01233268e-01 -4.08578724e-01 -8.11290562e-01 -3.28013748e-01
5.39776981e-01 -4.20714051e-01 7.76708901e-01 -6.62093878e-01
-1.18520486e+00 9.40092027e-01 -1.18033871e-01 4.23045635e-01
5.62402844e-01 -2.40452394e-01 -3.46303642e-01 1.81334466e-01
2.57125497e-01 5.38240254e-01 7.66276717e-01 -1.78737593e+00
-3.13163996e-01 -7.02682793e-01 -9.96079892e-02 5.21371722e-01
-3.54036570e-01 -3.43674242e-01 -9.10626829e-01 -7.38916874e-01
7.85527885e-01 -8.74524415e-01 9.91763920e-02 1.00361638e-01
-5.36841214e-01 -4.23597127e-01 6.49827361e-01 -6.54343247e-01
1.22644579e+00 -2.37546420e+00 5.32080173e-01 4.68147635e-01
8.35133716e-02 -4.01118398e-02 -3.32860500e-01 2.26453125e-01
2.15387449e-01 -6.90482706e-02 -4.00761038e-01 -5.14105737e-01
2.56647855e-01 1.57957554e-01 -1.89146906e-01 7.71571934e-01
-6.05891719e-02 6.79847240e-01 -9.57603276e-01 -7.59198844e-01
1.71565890e-01 4.13112134e-01 -1.37712717e-01 2.40994021e-01
-1.06904864e-01 6.19932592e-01 -7.60934055e-01 1.14406443e+00
1.04561174e+00 -2.21551895e-01 1.13862060e-01 -5.93603790e-01
9.79677588e-02 -2.01666042e-01 -1.46930563e+00 2.12656927e+00
-2.75579691e-01 1.70617849e-01 2.61568546e-01 -1.07317948e+00
1.03602540e+00 2.13508576e-01 5.95890343e-01 -9.74167824e-01
3.57037663e-01 6.16174102e-01 -5.79679251e-01 -7.61572897e-01
4.20090705e-01 1.76224727e-02 -2.93748230e-01 1.34020820e-01
-3.78536545e-02 -7.58015811e-02 1.49425701e-03 2.46681213e-01
4.13643181e-01 1.62723854e-01 -1.52548954e-01 -1.15192078e-01
7.78368711e-01 -2.34568194e-01 7.78195798e-01 4.69302326e-01
-3.00429195e-01 7.86475241e-01 2.35078707e-01 6.66785762e-02
-8.27407300e-01 -9.91273701e-01 -2.36527130e-01 7.94539332e-01
1.01148438e+00 -3.59304726e-01 -4.93223220e-01 -6.29848242e-01
3.64622287e-02 2.95299470e-01 -4.11262423e-01 -2.05720425e-01
-3.19544464e-01 -7.42783964e-01 4.16323721e-01 1.09116681e-01
4.99492973e-01 -3.49589139e-01 5.79696260e-02 -2.28889838e-01
-7.29111731e-01 -1.02701044e+00 -7.63829648e-01 -2.38571167e-01
-6.34193003e-01 -1.11517930e+00 -8.17296505e-01 -9.00392294e-01
7.74523437e-01 7.27010965e-01 6.56443000e-01 1.99761093e-01
-2.10076883e-01 6.67510629e-01 -4.57100481e-01 3.44305299e-02
9.11461480e-04 4.13849615e-02 1.98467106e-01 3.17575783e-01
3.09237003e-01 -2.79486656e-01 -6.51070058e-01 6.08444452e-01
-1.25928617e+00 5.11592776e-02 4.04131323e-01 1.01352715e+00
9.82646465e-01 -7.62438029e-02 4.88702446e-01 -2.37097785e-01
4.99367923e-01 -7.24546850e-01 -6.80377066e-01 4.07700717e-01
-3.75664294e-01 -9.26098004e-02 2.72209793e-01 -4.39807802e-01
-9.16733265e-01 -1.23632506e-01 2.27432102e-01 -6.17926896e-01
-9.12702903e-02 5.75021386e-01 -6.82850420e-01 1.04358219e-01
2.00336173e-01 3.47410470e-01 7.57679567e-02 -7.23413229e-01
3.63560736e-01 8.73299539e-01 4.00765866e-01 -8.37513506e-01
8.62845719e-01 6.06962144e-01 1.20320909e-01 -7.18859851e-01
-1.09243524e+00 -7.89014578e-01 -4.43132579e-01 -4.43690479e-01
5.30061424e-01 -1.10972726e+00 -5.70814550e-01 5.93263328e-01
-9.38582182e-01 1.58848420e-01 1.63781956e-01 4.78062838e-01
-2.50185668e-01 6.99395776e-01 -4.04484451e-01 -8.46694589e-01
-1.61952212e-01 -1.21740460e+00 1.47672594e+00 3.88423026e-01
2.73164034e-01 -8.78801525e-01 8.39724764e-02 7.25854635e-01
7.37052336e-02 1.20283805e-01 7.51429915e-01 -4.46212083e-01
-6.59737289e-01 -2.86962897e-01 -3.90743464e-01 4.44442809e-01
8.18062872e-02 -2.04796568e-01 -8.17464948e-01 -3.22310776e-01
-6.70125987e-03 -4.51193601e-01 6.83316290e-01 1.35497823e-02
1.09896207e+00 3.62415239e-02 -2.63899237e-01 4.16919380e-01
1.37401938e+00 -1.46196336e-01 1.97488680e-01 2.95616090e-01
8.09926629e-01 7.33626246e-01 9.63924170e-01 4.50456381e-01
4.95711237e-01 8.48627448e-01 4.81356680e-01 2.08683416e-01
-2.72893719e-02 -3.40560287e-01 1.15048140e-01 1.26188350e+00
2.59040892e-01 -3.66426885e-01 -7.21982956e-01 4.92139280e-01
-2.03487611e+00 -5.49354315e-01 -2.58643534e-02 2.07954621e+00
7.90045142e-01 -2.62533218e-01 -2.30091304e-01 5.52769890e-03
8.78865480e-01 2.67019928e-01 -5.36596119e-01 5.77075064e-01
-6.29579484e-01 -4.75264877e-01 1.48095950e-01 4.39966440e-01
-1.18261862e+00 5.49858093e-01 4.25557041e+00 1.33220422e+00
-9.28567231e-01 9.46875662e-02 4.69941348e-01 -1.47048861e-01
-6.93329990e-01 -1.67453021e-01 -6.41158462e-01 6.43814623e-01
1.42118573e-01 1.01675205e-01 3.43572706e-01 7.11510360e-01
2.70581871e-01 -9.94138718e-02 -7.08913565e-01 1.45889676e+00
3.40998918e-01 -9.19990242e-01 1.27660349e-01 -8.57727826e-02
8.75047147e-01 -1.52318031e-01 1.58599406e-01 6.07396523e-03
-3.27613324e-01 -5.34036398e-01 7.97567129e-01 8.16092074e-01
5.98630488e-01 -7.27833271e-01 8.33456099e-01 3.41025710e-01
-1.05650365e+00 -2.64374781e-02 -3.68657172e-01 6.23278201e-01
1.08933941e-01 7.98060894e-01 -9.56027396e-03 9.87683713e-01
6.17600203e-01 8.84700775e-01 -6.44235432e-01 1.16219699e+00
-1.27213195e-01 1.66614041e-01 -4.56422150e-01 1.55319840e-01
1.97284594e-01 -3.74435842e-01 8.73067260e-01 8.80616009e-01
4.67439830e-01 1.29262179e-01 3.45745087e-01 6.09269440e-01
-2.32051551e-01 2.82317311e-01 -2.85338223e-01 2.82082379e-01
6.89324200e-01 1.30526149e+00 -4.12294000e-01 -3.18379072e-03
-1.99369073e-01 8.46885622e-01 1.80413485e-01 5.50100088e-01
-6.40311062e-01 -3.12150538e-01 3.97898763e-01 -3.25730652e-01
-1.29513308e-01 -3.76520216e-01 -3.87903482e-01 -1.39846075e+00
4.46440250e-01 -8.14487338e-01 4.25760925e-01 -7.62619495e-01
-1.75010693e+00 3.02145153e-01 7.54566863e-02 -1.58672762e+00
4.57983345e-01 -4.40816194e-01 -3.83742005e-02 6.99360251e-01
-1.54267943e+00 -1.17048442e+00 -5.88437617e-01 5.36409974e-01
3.22773874e-01 -1.43434942e-01 4.19390827e-01 6.99443817e-01
-6.46025419e-01 6.62902951e-01 5.11845171e-01 3.21576223e-02
1.01081645e+00 -8.64556611e-01 -4.59986836e-01 5.57342649e-01
4.88371290e-02 5.25291979e-01 4.33103561e-01 -4.83888388e-01
-1.77144170e+00 -8.67823660e-01 5.13297915e-01 -2.28376180e-01
6.00923240e-01 -3.10461134e-01 -1.04131269e+00 9.38369408e-02
-1.46784663e-01 2.84885131e-02 6.41113281e-01 -2.23121822e-01
-5.68237662e-01 -3.64222914e-01 -9.69305396e-01 5.58494389e-01
1.00315356e+00 -7.30708659e-01 -2.82630980e-01 4.56832170e-01
6.47365749e-01 -5.93998253e-01 -8.40885639e-01 6.34734929e-01
4.02159691e-01 -6.76501095e-01 8.32939327e-01 -8.93534273e-02
1.84304371e-01 -6.69316113e-01 -7.49095440e-01 -1.10895026e+00
6.36310726e-02 -2.82330006e-01 -1.59990758e-01 1.43463635e+00
1.30245343e-01 -4.24736619e-01 4.48924094e-01 4.62838739e-01
-1.24826550e-01 -7.16015279e-01 -1.17005122e+00 -6.86126947e-01
-1.49210870e-01 -4.70332891e-01 4.47395205e-01 9.70079243e-01
-1.10908553e-01 1.56124189e-01 -4.34148043e-01 3.32303584e-01
8.67747426e-01 2.88848758e-01 6.29051626e-01 -1.01054978e+00
1.43522933e-01 -2.96675950e-01 -3.74341369e-01 -1.34681249e+00
2.68042386e-01 -1.02346385e+00 1.05930984e-01 -1.23159552e+00
3.44912231e-01 -7.05745697e-01 -3.88025790e-01 1.57060415e-01
-3.16987634e-01 3.66285771e-01 2.07172379e-01 6.39124990e-01
-9.71340954e-01 1.07719088e+00 1.40396774e+00 -3.19415033e-01
4.08748537e-02 -2.30652854e-01 -4.82404381e-01 4.25180227e-01
5.83616674e-01 -3.20729613e-01 -4.24556345e-01 -6.64463222e-01
3.69709730e-01 -1.39684528e-01 5.49941599e-01 -6.56341612e-01
3.28504890e-01 -2.08424721e-02 2.07170367e-01 -7.53928244e-01
4.79645938e-01 -9.62104082e-01 -3.94288041e-02 -1.40040115e-01
-4.52514827e-01 -1.04791194e-01 -6.36954084e-02 8.24007690e-01
-5.04793584e-01 -1.74992085e-01 5.52701175e-01 1.42032191e-01
-5.61513007e-01 3.70417178e-01 1.94988191e-01 1.52582005e-01
7.38072336e-01 1.05128005e-01 -2.93703973e-01 -2.68493176e-01
-5.80901623e-01 6.69677496e-01 5.36086917e-01 5.87563872e-01
7.23319829e-01 -1.71712577e+00 -5.98179936e-01 5.38648590e-02
4.86333460e-01 2.12114617e-01 8.23892832e-01 9.49684441e-01
-2.34347761e-01 3.38079274e-01 3.17230016e-01 -8.57333362e-01
-1.10841680e+00 4.75098968e-01 3.04298997e-01 6.87403232e-02
-3.15714300e-01 6.19932592e-01 -7.25879148e-02 -7.45231569e-01
4.92985964e-01 2.51520753e-01 5.09850830e-02 3.30181986e-01
3.07219058e-01 4.55085933e-01 -3.00941132e-02 -9.20294166e-01
-4.65559989e-01 7.82377660e-01 9.26966295e-02 -9.78745744e-02
1.05750465e+00 -6.51598811e-01 -2.89826304e-01 5.72173238e-01
1.54682159e+00 -7.90338963e-02 -1.11790478e+00 -6.23252153e-01
-4.00689133e-02 -7.03516960e-01 6.61302805e-02 -5.01342475e-01
-1.02868211e+00 7.62821555e-01 7.85597920e-01 5.93284257e-02
1.14610887e+00 1.72851160e-01 8.77965331e-01 2.12498799e-01
3.03930789e-01 -1.39448059e+00 3.44684839e-01 1.98382258e-01
8.83293092e-01 -1.48029101e+00 -1.27276117e-02 -3.72072190e-01
-5.93690574e-01 8.19187403e-01 5.31513453e-01 7.13433102e-02
6.91348970e-01 -3.26205432e-01 1.94992229e-01 -2.30960190e-01
-3.39782476e-01 -1.23186745e-02 5.15747309e-01 3.49447578e-01
1.10750347e-01 -1.24152573e-02 -3.70056391e-01 6.35056257e-01
3.24899673e-01 -3.74051154e-01 7.97161311e-02 9.24418926e-01
-1.16502181e-01 -1.02505493e+00 -6.51246727e-01 -3.06303147e-02
-1.35171860e-01 2.12481737e-01 -2.13929415e-01 6.34351552e-01
2.20718458e-01 1.03626633e+00 -2.59526342e-01 -3.98588747e-01
2.78960854e-01 -9.81512368e-02 3.46867681e-01 -9.19023305e-02
7.68661797e-02 5.58123052e-01 -3.36904049e-01 -3.94556165e-01
-7.45961189e-01 -6.42246366e-01 -1.12522912e+00 -1.03702256e-02
-5.99060714e-01 2.91401237e-01 7.82541037e-01 8.14733386e-01
4.76950735e-01 5.53443618e-02 9.38316941e-01 -8.11981738e-01
-6.30557716e-01 -8.44112217e-01 -6.56253994e-01 5.63912809e-01
3.07690799e-01 -9.45692360e-01 -6.09298229e-01 -2.57188857e-01] | [11.115046501159668, 1.1651995182037354] |
3a894838-34ae-45b8-8644-ebcbbdfe70c2 | latent-class-conditional-noise-model | 2302.09595 | null | https://arxiv.org/abs/2302.09595v1 | https://arxiv.org/pdf/2302.09595v1.pdf | Latent Class-Conditional Noise Model | Learning with noisy labels has become imperative in the Big Data era, which saves expensive human labors on accurate annotations. Previous noise-transition-based methods have achieved theoretically-grounded performance under the Class-Conditional Noise model (CCN). However, these approaches builds upon an ideal but impractical anchor set available to pre-estimate the noise transition. Even though subsequent works adapt the estimation as a neural layer, the ill-posed stochastic learning of its parameters in back-propagation easily falls into undesired local minimums. We solve this problem by introducing a Latent Class-Conditional Noise model (LCCN) to parameterize the noise transition under a Bayesian framework. By projecting the noise transition into the Dirichlet space, the learning is constrained on a simplex characterized by the complete dataset, instead of some ad-hoc parametric space wrapped by the neural layer. We then deduce a dynamic label regression method for LCCN, whose Gibbs sampler allows us efficiently infer the latent true labels to train the classifier and to model the noise. Our approach safeguards the stable update of the noise transition, which avoids previous arbitrarily tuning from a mini-batch of samples. We further generalize LCCN to different counterparts compatible with open-set noisy labels, semi-supervised learning as well as cross-model training. A range of experiments demonstrate the advantages of LCCN and its variants over the current state-of-the-art methods. | ['Ivor W. Tsang', 'Ya zhang', 'Zhihan Zhou', 'Bo Han', 'Jiangchao Yao'] | 2023-02-19 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 3.61928284e-01 2.94707596e-01 -4.60606441e-02 -5.30927718e-01
-1.27050579e+00 -3.34107518e-01 7.52645612e-01 -6.11841232e-02
-5.91868222e-01 6.57635689e-01 -1.81022227e-01 4.96632280e-03
-2.40943469e-02 -7.70767450e-01 -6.56587720e-01 -1.15962219e+00
3.85886133e-01 8.68265390e-01 2.37910092e-01 4.07811970e-01
-9.70575809e-02 7.30313882e-02 -1.64497244e+00 -6.74602687e-02
7.57830441e-01 9.93799686e-01 1.64582506e-01 4.67802763e-01
-1.99825034e-01 4.00633454e-01 -4.78678524e-01 -4.24600899e-01
6.62110299e-02 -2.06039473e-01 -6.01152182e-01 8.29744712e-02
1.33910803e-02 1.30277127e-01 3.43959183e-01 1.37948012e+00
4.93978202e-01 1.14267699e-01 7.48740315e-01 -1.03405035e+00
-2.74099588e-01 6.83312833e-01 -4.65238541e-01 -5.04681468e-01
7.00065121e-02 -3.36917229e-02 8.56891692e-01 -9.33205068e-01
5.82384050e-01 1.28375864e+00 1.16291642e+00 6.03787363e-01
-1.63523483e+00 -5.34125090e-01 2.27442130e-01 -1.05449505e-01
-1.66478050e+00 -3.31911534e-01 6.35625064e-01 -6.12831533e-01
4.14426327e-01 1.12792820e-01 5.17295539e-01 1.50413811e+00
-1.75022766e-01 6.77161336e-01 1.31723905e+00 -6.07579112e-01
8.10048938e-01 2.56226003e-01 1.28643289e-01 4.66636121e-01
2.04763606e-01 -7.88724646e-02 -4.24117178e-01 -4.21925634e-01
3.74070048e-01 -7.92425126e-02 -9.98375416e-02 -4.97971117e-01
-6.82947218e-01 8.14661980e-01 -8.24507326e-02 2.70887185e-02
-1.88836753e-01 2.72430837e-01 3.46406579e-01 -4.40109707e-02
9.01790679e-01 -1.35198981e-01 -5.20324349e-01 -2.22582534e-01
-1.30370820e+00 -1.47511426e-04 1.06073642e+00 8.30056310e-01
1.09370732e+00 -2.77851790e-01 -1.44069687e-01 9.08983052e-01
8.57424021e-01 4.52372909e-01 1.86368167e-01 -9.28973138e-01
2.29392499e-01 2.85253882e-01 2.26249099e-01 -5.63138187e-01
-5.02406299e-01 -7.19805777e-01 -9.43157136e-01 1.60128340e-01
7.68409073e-01 -2.19543889e-01 -1.02622902e+00 1.94670165e+00
7.70753145e-01 3.21357995e-01 -2.35822037e-01 4.42671776e-01
1.00868739e-01 5.23775220e-01 1.83503136e-01 -4.15498465e-01
1.25411427e+00 -7.97835112e-01 -9.03972268e-01 -1.92487225e-01
8.02903712e-01 -4.74578530e-01 1.17669415e+00 7.44470596e-01
-7.59879470e-01 -4.08280730e-01 -7.90971816e-01 6.79357722e-02
-3.48405510e-01 1.55160695e-01 3.57557535e-01 1.04421079e+00
-9.60264623e-01 6.70304418e-01 -1.31356728e+00 -1.73304588e-01
5.50250947e-01 3.10007334e-01 -7.31187314e-03 -1.50553480e-01
-1.11391592e+00 8.26031864e-01 4.50432092e-01 5.60344517e-01
-9.97039378e-01 -5.52204907e-01 -7.22511649e-01 -2.09161967e-01
7.23273993e-01 -5.14856279e-01 1.25067210e+00 -7.89439082e-01
-1.89364457e+00 8.97069097e-01 -1.99655190e-01 -2.50610262e-01
8.33209932e-01 -2.03571007e-01 -1.02758944e-01 -2.23484576e-01
3.46677415e-02 3.88832390e-01 1.02545500e+00 -1.57476366e+00
-4.08597261e-01 -2.72821844e-01 -3.67799342e-01 -2.74047237e-02
-2.12126553e-01 -8.06082189e-02 -7.73745537e-01 -4.30171013e-01
4.35388386e-01 -1.10122979e+00 -4.22675341e-01 -2.31430922e-02
-5.43478251e-01 -2.83034325e-01 5.58046639e-01 -2.84212351e-01
1.26806796e+00 -2.06826234e+00 -2.59989984e-02 4.43318814e-01
9.40218642e-02 3.67662981e-02 3.04014653e-01 2.21597195e-01
1.22262366e-01 5.80702499e-02 -6.72113419e-01 -1.03007734e+00
3.34650934e-01 6.22283638e-01 -7.25002065e-02 8.13589573e-01
7.82009885e-02 5.41958690e-01 -9.35371935e-01 -5.50476611e-01
-1.31460994e-01 6.80195212e-01 -5.20172656e-01 8.48743394e-02
-4.92636144e-01 4.97062147e-01 -3.05788726e-01 5.13378441e-01
8.75723183e-01 -3.45561564e-01 2.63708621e-01 5.52489273e-02
4.88141887e-02 7.32354075e-02 -1.73426151e+00 1.84817529e+00
-3.90454799e-01 4.71320078e-02 2.84537554e-01 -1.06487310e+00
8.84949684e-01 2.47438669e-01 3.39860886e-01 7.13395551e-02
4.59574610e-02 3.32485974e-01 -6.46689951e-01 -4.25348490e-01
3.36333476e-02 -4.92261797e-01 -8.14561695e-02 3.22829038e-01
3.60785723e-01 -1.54327929e-01 1.76605303e-02 -1.14817522e-01
9.05449569e-01 6.89207613e-01 7.37995729e-02 -2.81900078e-01
3.72577697e-01 -5.12455285e-01 8.00270200e-01 1.06080961e+00
-8.48064572e-02 8.21977317e-01 6.39194131e-01 -1.27695560e-01
-8.01254988e-01 -9.42316175e-01 -6.24405742e-01 1.11907792e+00
-2.64818668e-01 -4.44162965e-01 -1.17943192e+00 -7.97786653e-01
-2.98831791e-01 7.43070781e-01 -6.58605039e-01 1.76094007e-02
-1.78638965e-01 -1.32390130e+00 6.25028491e-01 1.07148416e-01
2.10610211e-01 -7.36970186e-01 -1.95349768e-01 2.51231879e-01
-2.84733087e-01 -9.43948150e-01 -1.05991296e-01 8.01940322e-01
-6.61368847e-01 -7.21579671e-01 -5.20084858e-01 -4.87685293e-01
7.30013430e-01 -6.17436171e-01 9.95140076e-01 -3.11571985e-01
-1.05711818e-01 3.89778674e-01 -6.14554249e-02 -3.06400090e-01
-5.39941251e-01 1.25704855e-01 8.12838376e-02 3.20565552e-01
6.50275409e-01 -6.07804656e-01 -3.20158899e-01 2.35185817e-01
-9.72697318e-01 -7.42115304e-02 2.97596276e-01 9.20578837e-01
9.02818501e-01 1.55819684e-01 5.04745126e-01 -1.25657845e+00
2.31864691e-01 -6.80147946e-01 -9.64353740e-01 2.26144388e-01
-9.49696481e-01 2.96992987e-01 3.36888015e-01 -4.92684066e-01
-1.30404508e+00 4.61754173e-01 -3.85607332e-01 -2.36608803e-01
-3.57884526e-01 4.62418139e-01 -4.43012089e-01 3.06109369e-01
7.18180299e-01 -1.43054053e-01 -2.58775294e-01 -7.19581306e-01
4.93308157e-01 7.91144609e-01 4.31499362e-01 -9.24636304e-01
6.95644438e-01 6.32207572e-01 2.61165593e-02 -5.34222782e-01
-1.35707295e+00 -5.91175079e-01 -8.56155932e-01 -1.79811448e-01
9.23228920e-01 -9.09094989e-01 -6.03382409e-01 6.78032875e-01
-1.01910019e+00 -5.14055550e-01 -6.69855416e-01 3.34439099e-01
-6.65194631e-01 4.26465541e-01 -5.44498444e-01 -1.13354874e+00
1.10841237e-01 -1.14653420e+00 1.27187312e+00 -4.41230722e-02
-4.18398194e-02 -1.20909286e+00 3.55512857e-01 3.10419649e-01
1.24194264e-01 1.80775478e-01 6.26832604e-01 -7.62644053e-01
-3.80293667e-01 -4.67892021e-01 2.22821650e-03 5.51216006e-01
-2.16111928e-01 8.72865412e-03 -1.59185302e+00 -1.15276925e-01
3.24690044e-01 -4.48560208e-01 9.52491224e-01 4.51006562e-01
1.11687398e+00 -8.48732591e-02 -3.81408304e-01 7.02507794e-01
1.52042317e+00 -3.28249961e-01 4.91142333e-01 2.33327568e-01
6.55030608e-01 5.85433424e-01 2.90271044e-01 5.92875838e-01
4.58423406e-01 4.93815184e-01 4.26191777e-01 7.83765018e-02
1.57027990e-01 -4.02270496e-01 3.24888438e-01 9.52977777e-01
1.62171990e-01 -2.11786479e-01 -1.12895787e+00 2.85240322e-01
-1.94528759e+00 -5.77661157e-01 -1.97186261e-01 2.35894442e+00
1.45489442e+00 2.66969562e-01 -1.73013613e-01 2.37897262e-01
7.22576499e-01 -4.96170260e-02 -5.42351246e-01 1.32973149e-01
-6.30085468e-02 3.53279198e-03 3.99779886e-01 7.88805664e-01
-1.19551146e+00 8.65462601e-01 6.31950378e+00 1.13564873e+00
-6.69535995e-01 5.37152886e-01 6.70773268e-01 3.82891148e-02
-3.26750040e-01 2.60034144e-01 -1.33780849e+00 6.08883262e-01
1.20170057e+00 5.80902874e-01 3.06600600e-01 9.86650586e-01
1.60861745e-01 -4.13307130e-01 -1.18614626e+00 8.01659763e-01
-1.11031577e-01 -7.41863430e-01 -3.59694570e-01 1.38356522e-01
8.82288456e-01 1.00530073e-01 -1.14195436e-01 4.25903887e-01
7.27582455e-01 -9.09990013e-01 7.68334746e-01 9.32407677e-01
9.39400017e-01 -4.87614691e-01 8.33241045e-01 7.12889075e-01
-7.77471423e-01 -5.18613793e-02 -4.67138857e-01 4.98125926e-02
1.95179105e-01 1.32457554e+00 -6.12554550e-01 2.97190398e-01
6.08054578e-01 4.34726030e-01 -3.82343233e-01 7.46820688e-01
-2.91830987e-01 1.00822437e+00 -8.13163221e-01 2.43882731e-01
8.09802860e-02 -5.19966722e-01 2.96746731e-01 1.33424795e+00
2.28693977e-01 -2.08995476e-01 3.53961349e-01 1.00822628e+00
-7.05394372e-02 -5.16050048e-02 -2.07881585e-01 2.47330621e-01
3.41708452e-01 1.30826962e+00 -1.03702271e+00 -3.13623518e-01
-3.02312434e-01 7.97577083e-01 4.42860901e-01 5.73387504e-01
-7.32822061e-01 -4.13640812e-02 2.28449434e-01 1.46203293e-02
2.94853985e-01 -5.95184229e-02 -3.80957216e-01 -1.18971705e+00
7.85913989e-02 -6.99006081e-01 2.40091592e-01 -4.22269374e-01
-1.38875449e+00 2.97012359e-01 7.55357593e-02 -9.27854717e-01
-3.30595046e-01 -4.25697237e-01 -2.15021938e-01 8.65584075e-01
-1.41163957e+00 -1.05769277e+00 -9.34805945e-02 3.80701452e-01
1.85610294e-01 2.07159579e-01 1.01954889e+00 2.67522991e-01
-7.20434785e-01 4.93712991e-01 6.49796903e-01 -1.33496374e-01
8.15197289e-01 -1.45002985e+00 -5.26528172e-02 7.15188146e-01
2.53728516e-02 3.74420375e-01 7.59938240e-01 -7.65976965e-01
-7.93247163e-01 -1.01363158e+00 7.70565093e-01 -7.35477388e-01
7.81338096e-01 -8.23362708e-01 -1.04784298e+00 6.17890835e-01
-3.72817844e-01 4.45949763e-01 7.06279457e-01 3.49648386e-01
-3.58423382e-01 -1.11702017e-01 -1.13646448e+00 2.78849274e-01
8.09648752e-01 -5.81122935e-01 -9.96751413e-02 4.75516528e-01
6.22064054e-01 -3.30765218e-01 -9.07370150e-01 1.59241274e-01
3.08789253e-01 -9.30276155e-01 7.25013912e-01 -2.63107330e-01
-3.10645462e-03 -3.37218106e-01 -2.09289446e-01 -1.11261189e+00
-6.47309818e-04 -8.18422318e-01 -1.76524907e-01 1.56603444e+00
4.39057827e-01 -4.77996707e-01 8.54519010e-01 8.16841245e-01
-5.79354875e-02 -6.34792924e-01 -1.23139548e+00 -5.36880195e-01
-6.08592154e-03 -1.07618582e+00 2.81686604e-01 8.76764953e-01
-2.76396155e-01 1.33516103e-01 -3.34295988e-01 3.07410091e-01
9.90242004e-01 -5.61266005e-01 4.75280493e-01 -1.62216365e+00
-3.77639979e-01 -3.37007493e-02 -7.95356482e-02 -1.00440824e+00
2.88694024e-01 -7.89636075e-01 6.73138916e-01 -1.25662827e+00
1.52206227e-01 -8.54160905e-01 -1.96375594e-01 4.82498974e-01
-2.53791600e-01 1.78614184e-01 -2.88408518e-01 2.17004701e-01
-9.09810901e-01 6.39276624e-01 8.25090885e-01 2.06917003e-01
-1.23119920e-01 3.41779888e-01 -3.58170241e-01 1.17280483e+00
4.73978072e-01 -1.08884907e+00 -3.40116829e-01 -1.11878112e-01
7.11592138e-01 -1.98178634e-01 2.49053121e-01 -8.99713218e-01
4.43182945e-01 -1.03538425e-03 1.49637580e-01 -4.93044078e-01
3.26739818e-01 -9.77612674e-01 2.92668998e-01 6.95344731e-02
-4.33550984e-01 -6.74574494e-01 -2.20915049e-01 1.06999040e+00
1.56211453e-02 -8.22752416e-01 8.86531770e-01 -1.22942887e-01
3.33246700e-02 1.30787194e-01 -4.84192759e-01 3.23643535e-01
7.13161230e-01 -8.08611214e-02 2.62796909e-01 -2.02512071e-01
-1.09122777e+00 8.09868425e-02 3.36849928e-01 -1.88540250e-01
7.63852000e-02 -1.15093386e+00 -4.21251267e-01 3.01038384e-01
-9.48543251e-02 6.93932235e-01 1.15995310e-01 8.57564270e-01
-8.54166970e-02 -6.95445538e-02 5.93893409e-01 -9.33768690e-01
-6.97753608e-01 4.42442745e-01 3.92718911e-01 -5.69666445e-01
-5.15748441e-01 9.74511385e-01 -1.02304809e-01 -8.61620069e-01
7.36992240e-01 -3.14409345e-01 1.42609738e-02 4.08753186e-01
2.76963174e-01 3.84989500e-01 1.80764556e-01 -3.17227632e-01
-4.14683856e-02 6.06695712e-01 1.76402807e-01 -3.45740348e-01
1.30848300e+00 -5.25928140e-01 -2.26199731e-01 1.11444736e+00
1.11407471e+00 -1.10891581e-01 -1.70234966e+00 -7.10705400e-01
4.88108188e-01 -2.11039428e-02 1.63143486e-01 -7.13085175e-01
-5.61858296e-01 9.27694738e-01 7.12031484e-01 2.27938935e-01
8.91983807e-01 6.03580661e-03 2.52522171e-01 4.41955775e-01
2.63030320e-01 -1.57147515e+00 -1.39625460e-01 5.17859221e-01
3.02501142e-01 -1.42875373e+00 -9.55739468e-02 -2.62974650e-01
-3.65518540e-01 9.99500334e-01 2.24582195e-01 2.12608188e-01
1.19573629e+00 5.46474457e-01 1.99878767e-01 -9.59315673e-02
-6.92853630e-01 -1.18428431e-01 -3.02348682e-03 6.27825916e-01
9.49589908e-02 -1.06212497e-01 1.17686942e-01 8.49996328e-01
6.26689047e-02 5.55494688e-02 3.21257591e-01 7.39869654e-01
-4.96639401e-01 -1.06273353e+00 -5.50707221e-01 1.43589899e-01
-5.66357493e-01 -1.41320750e-01 1.82888508e-01 4.40913320e-01
5.24548948e-01 7.95416176e-01 -9.42909420e-02 8.74837115e-02
7.43577629e-02 8.07587743e-01 5.80885187e-02 -7.16982305e-01
-2.95996040e-01 5.52479804e-01 -6.95146620e-02 -3.62557650e-01
-5.42205513e-01 -8.72103274e-01 -1.08894944e+00 7.41949305e-02
-7.52799749e-01 1.59725651e-01 9.21762586e-01 1.10192204e+00
-5.21758795e-02 2.86152691e-01 4.22588199e-01 -9.57306564e-01
-1.02666306e+00 -1.17271268e+00 -6.75305545e-01 2.30159163e-01
2.22818360e-01 -6.47464097e-01 -7.35796392e-01 2.80476302e-01] | [9.242593765258789, 4.005716323852539] |
1a9279f5-8305-4da7-88cc-a7ad5c3766a4 | exploring-exploiting-high-order-graph | 2306.17034 | null | https://arxiv.org/abs/2306.17034v1 | https://arxiv.org/pdf/2306.17034v1.pdf | Exploring & Exploiting High-Order Graph Structure for Sparse Knowledge Graph Completion | Sparse knowledge graph (KG) scenarios pose a challenge for previous Knowledge Graph Completion (KGC) methods, that is, the completion performance decreases rapidly with the increase of graph sparsity. This problem is also exacerbated because of the widespread existence of sparse KGs in practical applications. To alleviate this challenge, we present a novel framework, LR-GCN, that is able to automatically capture valuable long-range dependency among entities to supplement insufficient structure features and distill logical reasoning knowledge for sparse KGC. The proposed approach comprises two main components: a GNN-based predictor and a reasoning path distiller. The reasoning path distiller explores high-order graph structures such as reasoning paths and encodes them as rich-semantic edges, explicitly compositing long-range dependencies into the predictor. This step also plays an essential role in densifying KGs, effectively alleviating the sparse issue. Furthermore, the path distiller further distills logical reasoning knowledge from these mined reasoning paths into the predictor. These two components are jointly optimized using a well-designed variational EM algorithm. Extensive experiments and analyses on four sparse benchmarks demonstrate the effectiveness of our proposed method. | ['Bing Qin', 'Zheng Chu', 'Zihao Zheng', 'Zekun Wang', 'Yixin Cao', 'Ming Liu', 'Tao He'] | 2023-06-29 | null | null | null | null | ['knowledge-graph-completion', 'logical-reasoning'] | ['knowledge-base', 'reasoning'] | [-1.56809136e-01 5.37029803e-01 -2.75596976e-01 -1.86458871e-01
-3.96169275e-01 -1.01936497e-01 3.76437724e-01 2.69410938e-01
-5.82553931e-02 7.45267510e-01 5.71178794e-01 -9.62285474e-02
-5.43200850e-01 -1.17154968e+00 -7.75350153e-01 -5.81756175e-01
-1.58844925e-02 6.62594497e-01 1.34514719e-01 -1.41390786e-01
6.63229004e-02 7.35580474e-02 -1.29249787e+00 1.45646602e-01
1.44387269e+00 8.01449835e-01 3.17160308e-01 3.65765579e-02
-6.01021409e-01 1.22745562e+00 6.05336484e-03 -6.44220412e-01
3.02455761e-02 -1.01593800e-01 -8.01362097e-01 -3.52611532e-03
9.28203836e-02 1.35730933e-02 -5.14019668e-01 1.26496494e+00
8.45826268e-02 3.59284014e-01 3.86786819e-01 -1.14498663e+00
-4.52747226e-01 1.18533206e+00 -8.25381398e-01 -5.18615320e-02
3.24530959e-01 -9.64969620e-02 1.28176904e+00 -1.02008283e+00
7.71683693e-01 1.35481596e+00 7.73993373e-01 9.67116728e-02
-8.58723760e-01 -5.26466727e-01 5.47764957e-01 5.01363099e-01
-1.52300978e+00 -8.87239724e-02 1.08387184e+00 -3.56154263e-01
8.17753196e-01 2.31496245e-02 8.40481281e-01 7.71601975e-01
-1.79720253e-01 8.97714496e-01 7.08397329e-01 -1.26843333e-01
1.85781687e-01 4.18052189e-02 3.18326265e-01 1.10760331e+00
5.57774305e-01 -3.71737748e-01 -6.76627636e-01 -1.26103416e-01
7.06221581e-01 1.54481679e-02 -3.83662224e-01 -5.29453218e-01
-1.01765847e+00 7.64399648e-01 6.45410955e-01 4.13189568e-02
-6.93749845e-01 2.34035477e-01 2.20786646e-01 -8.19668099e-02
3.11071724e-01 1.81288943e-01 -4.26432014e-01 -1.60870761e-01
-8.32278073e-01 2.76397347e-01 1.01995301e+00 1.09635437e+00
9.79178667e-01 -5.81673086e-02 -1.62649035e-01 8.39363158e-01
4.57383275e-01 3.44714165e-01 3.43087167e-02 -7.69406438e-01
7.81822681e-01 1.16046965e+00 -2.41872236e-01 -1.55938435e+00
-3.74938637e-01 -7.58919775e-01 -1.08049607e+00 -5.92656493e-01
-1.66640070e-03 -2.88857631e-02 -6.72928452e-01 1.70829391e+00
7.54818439e-01 4.13951337e-01 2.09826641e-02 8.79693985e-01
1.07817793e+00 4.65884030e-01 1.07698813e-01 -9.71300304e-02
1.37507987e+00 -1.09001195e+00 -7.92955220e-01 -1.22190274e-01
5.51742196e-01 -3.65901709e-01 8.44297588e-01 2.19974414e-01
-9.16548908e-01 -1.97658479e-01 -7.73247361e-01 -1.84440896e-01
-1.22870661e-01 4.54828423e-03 1.08250511e+00 2.82983989e-01
-8.10799479e-01 4.58082169e-01 -7.51783490e-01 -5.55725023e-02
6.42236710e-01 2.47874297e-02 -2.31837064e-01 -5.26485920e-01
-1.35050941e+00 6.17817283e-01 8.62470806e-01 4.69563097e-01
-6.25557780e-01 -9.47724164e-01 -1.14579785e+00 3.34440082e-01
1.07848251e+00 -9.19317544e-01 5.83304107e-01 -2.69564867e-01
-1.04573262e+00 3.77472728e-01 -1.74784690e-01 -2.95340568e-01
3.90198410e-01 -2.55498409e-01 -4.88718718e-01 1.57348827e-01
1.75742969e-01 3.18455637e-01 8.21717501e-01 -1.34223282e+00
-4.70923394e-01 -3.98966849e-01 1.66871712e-01 3.90359700e-01
-3.89444023e-01 -7.65503109e-01 -1.07851636e+00 -6.99695647e-01
3.70345086e-01 -5.54360271e-01 -3.48007143e-01 -4.85778481e-01
-7.46218503e-01 -2.65672058e-01 7.30859935e-01 -8.05727839e-01
1.48453999e+00 -1.79574549e+00 5.69204688e-01 7.22723901e-01
6.59757912e-01 6.14118241e-02 -9.83957946e-02 5.36386847e-01
1.94228277e-01 -1.27171636e-01 -3.10068399e-01 -3.25825036e-01
1.69583738e-01 5.80491424e-01 -2.40260869e-01 1.03892766e-01
4.47053425e-02 1.15763783e+00 -1.19562125e+00 -8.10392022e-01
2.21711267e-02 3.37652653e-01 -8.35166574e-01 -9.09672379e-02
-5.14990032e-01 2.42938265e-01 -7.82294691e-01 7.65020370e-01
9.14175987e-01 -6.44367635e-01 5.38934767e-01 -4.43388999e-01
1.75052032e-01 1.39966547e-01 -1.45208204e+00 2.02696466e+00
-2.42874235e-01 -5.16492911e-02 1.92715228e-01 -1.26725280e+00
8.82131040e-01 -2.02300400e-01 5.29180229e-01 -5.78445852e-01
3.84541005e-02 4.80223820e-02 -3.74180436e-01 -4.78224456e-01
7.18030989e-01 6.61693290e-02 2.69759502e-02 2.26179972e-01
-1.78388692e-02 5.46683110e-02 4.31485653e-01 9.58344162e-01
1.23046291e+00 1.83289513e-01 7.46820122e-02 -3.99586022e-01
8.71864140e-01 1.05575278e-01 9.90011215e-01 4.84400719e-01
1.93877473e-01 -1.88001096e-02 7.19498098e-01 -1.45803437e-01
-4.36259329e-01 -9.40902591e-01 2.67404884e-01 4.85760510e-01
4.16680098e-01 -1.01935565e+00 -5.16678035e-01 -8.02327693e-01
2.70460576e-01 5.86909354e-01 -4.76697892e-01 -2.78207988e-01
-5.00991344e-01 -5.87221384e-01 3.05960804e-01 5.13817549e-01
6.47423148e-01 -9.21633542e-01 5.59110977e-02 1.65107325e-01
-5.87258518e-01 -1.37250245e+00 -1.62551984e-01 -1.78336829e-01
-7.67012894e-01 -1.42876720e+00 -2.91826278e-01 -6.91903472e-01
9.78767335e-01 1.86598346e-01 1.24761379e+00 4.08594847e-01
-2.03773052e-01 5.36728263e-01 -4.77023751e-01 9.02718753e-02
-4.57774661e-02 5.89853190e-02 -2.37413391e-01 4.05899957e-02
2.69300997e-01 -8.83534551e-01 -4.35415179e-01 -2.25683991e-02
-6.48033798e-01 4.46707845e-01 8.20932806e-01 7.14198053e-01
8.93532932e-01 7.54445374e-01 3.87646616e-01 -1.37470877e+00
6.04230046e-01 -7.90169477e-01 -5.15031219e-01 4.29103225e-01
-7.30699241e-01 3.33768636e-01 5.46081781e-01 -5.93949072e-02
-1.35635662e+00 -1.25470474e-01 7.32945725e-02 -5.92591465e-01
5.12862742e-01 1.32218492e+00 -3.13716978e-01 1.05406322e-01
1.34914070e-01 2.31090337e-01 -2.30727017e-01 -5.39304256e-01
6.59165025e-01 -2.37703752e-02 5.70897698e-01 -1.02375054e+00
1.13504541e+00 5.04795074e-01 2.03334570e-01 -8.69961143e-01
-1.01979113e+00 -5.62572837e-01 -3.12828749e-01 -1.75313964e-01
6.18661106e-01 -1.09245789e+00 -7.32829750e-01 2.92550534e-01
-7.98772275e-01 -1.07770689e-01 -3.01927537e-01 4.10342872e-01
-1.41198650e-01 7.16315210e-01 -6.03955746e-01 -6.40557766e-01
-3.39019835e-01 -8.28369975e-01 8.74684870e-01 2.42745131e-01
7.02020079e-02 -1.17675745e+00 -6.79900264e-03 7.73790002e-01
1.71660438e-01 3.57931286e-01 1.16534269e+00 -1.82623446e-01
-1.08729231e+00 9.24726203e-02 -5.86096525e-01 8.10139403e-02
-5.04167229e-02 -3.44724387e-01 -4.86812502e-01 -6.80554286e-02
-2.68062770e-01 -1.86279133e-01 1.03428602e+00 2.00494125e-01
1.12705493e+00 -2.97382593e-01 -4.88845497e-01 6.10824227e-01
1.55665374e+00 -4.02414352e-01 5.19716799e-01 -1.91317927e-02
1.40415251e+00 4.83749092e-01 6.48906291e-01 5.35463274e-01
1.02346826e+00 3.12553614e-01 4.30552512e-01 5.08790649e-02
-1.93143100e-01 -7.52688229e-01 1.02746323e-01 1.27330112e+00
-2.37824425e-01 -1.05205327e-02 -9.43042278e-01 4.73439544e-01
-2.25687051e+00 -7.38112986e-01 -5.55947781e-01 1.73220623e+00
8.70301723e-01 -4.65069190e-02 -3.03501427e-01 3.02256290e-02
6.38886571e-01 2.29740679e-01 -5.77472031e-01 2.09921226e-01
-1.42609671e-01 1.32178590e-01 2.47155294e-01 5.35293460e-01
-6.70782804e-01 1.03907740e+00 4.60484076e+00 9.83765960e-01
-3.78529638e-01 -4.63227145e-02 5.32105677e-02 1.76405519e-01
-9.13203657e-01 4.01027650e-01 -7.35366285e-01 3.48847210e-01
4.08700019e-01 -1.62495494e-01 4.95843768e-01 9.30225074e-01
-2.89264768e-02 -3.34461570e-01 -7.32623100e-01 1.05639994e+00
5.65257967e-02 -1.55571258e+00 2.27676749e-01 9.24811512e-02
9.55562711e-01 -3.22360486e-01 -3.77387732e-01 7.37136602e-01
6.40379131e-01 -6.69825673e-01 5.13256788e-01 7.64865756e-01
4.69918966e-01 -9.15742517e-01 6.08990014e-01 3.69561315e-01
-1.71458185e+00 6.62789792e-02 -3.86747688e-01 1.10389732e-01
3.35405499e-01 1.26035976e+00 -6.93454504e-01 1.26821315e+00
5.41336954e-01 1.20758152e+00 -4.64452654e-01 6.70752227e-01
-6.35390699e-01 4.61149544e-01 -2.10113287e-01 1.14702769e-01
3.16509545e-01 -6.15320504e-01 6.60203874e-01 9.09095287e-01
-3.22840773e-02 3.14377308e-01 5.05742192e-01 1.17715192e+00
-2.41027102e-01 5.34158461e-02 -2.31038362e-01 -3.44389588e-01
6.70037270e-01 1.39287663e+00 -7.38901198e-01 -3.14860076e-01
-4.22810793e-01 8.26923072e-01 7.91766465e-01 5.00514925e-01
-9.70311284e-01 -1.54978842e-01 5.11096358e-01 -1.26589322e-02
3.90234977e-01 -2.19060719e-01 -1.56875491e-01 -1.46459293e+00
3.91674608e-01 -7.48542070e-01 6.68660581e-01 -5.82861185e-01
-1.25384808e+00 1.08292602e-01 -1.10108703e-02 -6.27538323e-01
1.43514633e-01 -2.07282156e-02 -4.05268729e-01 8.19053590e-01
-1.81445444e+00 -1.33561897e+00 -7.12901950e-01 8.73873651e-01
1.79698810e-01 1.30016819e-01 4.63068485e-01 5.35956144e-01
-8.24186444e-01 3.35253775e-01 -1.31360799e-01 3.23698996e-03
1.06985904e-01 -1.24098647e+00 -7.16923177e-02 9.97457743e-01
1.27203077e-01 8.79105806e-01 3.54231328e-01 -1.13836884e+00
-1.83560419e+00 -1.23776507e+00 5.94685733e-01 -1.43698663e-01
7.80945241e-01 -1.90633491e-01 -1.03289711e+00 7.18418181e-01
-4.49389368e-01 -1.27549529e-01 4.59957421e-01 7.55843878e-01
-4.47752774e-01 -1.94964111e-02 -8.43641222e-01 5.15213728e-01
1.41203403e+00 -5.77793539e-01 -6.47902608e-01 3.24513972e-01
7.53952980e-01 -4.94594395e-01 -9.04984355e-01 4.45527285e-01
2.49813169e-01 -7.01677740e-01 1.06720817e+00 -3.30333143e-01
5.66046000e-01 -4.01529461e-01 1.83976050e-02 -1.14350045e+00
-3.66160631e-01 -5.61736226e-01 -7.60402799e-01 1.41320300e+00
3.43438447e-01 -5.43980956e-01 1.06316531e+00 5.03866494e-01
-1.91511586e-01 -8.92816961e-01 -4.61303711e-01 -6.10620260e-01
-4.40128356e-01 -3.77073079e-01 6.24271452e-01 1.27533531e+00
1.51821569e-01 4.63586032e-01 -3.68504524e-01 3.63650620e-01
9.99736726e-01 4.71422642e-01 8.52158010e-01 -1.35814846e+00
-3.40021461e-01 -2.20073283e-01 -3.40655357e-01 -1.17280138e+00
3.09527934e-01 -1.31315720e+00 -1.48805231e-01 -2.10458660e+00
4.85831916e-01 -8.69322777e-01 -1.57385156e-01 4.93929327e-01
-5.48897624e-01 -4.58914757e-01 -8.62511694e-02 1.31695226e-01
-8.71069729e-01 9.73383248e-01 1.42383039e+00 -1.65134817e-01
-1.92693070e-01 -5.36830366e-01 -8.76719892e-01 7.02613652e-01
3.52781355e-01 -3.83922100e-01 -9.74363685e-01 -3.50021124e-01
6.33491755e-01 1.42904893e-01 4.37311411e-01 -7.76645839e-01
5.48986137e-01 -9.26548019e-02 -2.87191439e-02 -8.98663580e-01
1.33133426e-01 -8.79025042e-01 4.60623205e-01 1.78666919e-01
5.72312400e-02 -3.56732607e-01 -1.60640348e-02 1.09482849e+00
-3.76325160e-01 7.68543407e-02 3.59259576e-01 -1.52198553e-01
-1.04669583e+00 4.21513736e-01 3.64340752e-01 4.51448649e-01
7.88891494e-01 -7.81091601e-02 -2.54936308e-01 -1.67492017e-01
-8.04500818e-01 9.41912711e-01 2.52550989e-01 2.99979568e-01
6.82996511e-01 -1.41252327e+00 -5.50072551e-01 1.23471417e-01
3.11432686e-02 5.66720307e-01 7.57883012e-01 1.21671224e+00
-2.88159698e-01 2.62137860e-01 3.00461888e-01 -3.22686613e-01
-9.21608925e-01 4.74590480e-01 -1.04009129e-01 -7.65127957e-01
-1.13397574e+00 1.14395714e+00 4.89565469e-02 -2.94399828e-01
1.84458375e-01 -3.90977114e-01 -2.04432249e-01 8.59637186e-03
1.23648465e-01 5.14489830e-01 -7.16492534e-02 -4.57188487e-01
-3.17374676e-01 3.03786635e-01 -2.03847304e-01 3.11709195e-01
1.48910558e+00 -1.76235586e-01 -5.04368305e-01 8.39716792e-02
7.50337243e-01 2.08040804e-01 -9.43309605e-01 -5.63924432e-01
9.64051634e-02 -4.35820699e-01 1.66467056e-01 -6.20961070e-01
-1.42057216e+00 4.16615814e-01 -4.25680995e-01 -9.79960635e-02
1.10734975e+00 4.00100872e-02 1.06047916e+00 4.73623186e-01
5.28122783e-01 -1.18824601e+00 1.08012423e-01 5.10515869e-01
6.29257619e-01 -1.01508653e+00 2.95262337e-01 -1.17423046e+00
-6.16614163e-01 6.49664462e-01 6.87197745e-01 2.34997831e-02
7.69391596e-01 -1.18817836e-01 -6.35982811e-01 -8.39983284e-01
-4.85682070e-01 -3.11089814e-01 3.70458186e-01 3.97966921e-01
-1.17306106e-01 2.08894983e-01 -4.30571318e-01 9.56426442e-01
-2.30746537e-01 5.79032004e-02 2.30478764e-01 8.13003421e-01
-2.34835446e-01 -9.08303678e-01 1.13084707e-02 5.28613567e-01
6.36787266e-02 -3.22125882e-01 -1.89195931e-01 7.87274539e-01
1.33670419e-01 7.63082087e-01 -3.63444388e-01 -3.67217958e-01
2.91011781e-01 -1.32762119e-01 3.42558026e-01 -5.37022412e-01
-1.41257375e-01 -1.09070443e-01 3.67004395e-01 -9.17620599e-01
-3.59042078e-01 -5.01490891e-01 -1.63899124e+00 -4.23580527e-01
-3.93732697e-01 4.51235831e-01 2.66203105e-01 9.94879842e-01
5.49848139e-01 8.79051805e-01 1.14504255e-01 -2.15240180e-01
-3.30309361e-01 -6.21474266e-01 -8.89718771e-01 5.26826262e-01
-1.75955102e-01 -9.51438487e-01 -1.92789420e-01 -2.40779847e-01] | [8.748922348022461, 7.881193161010742] |
af59d92e-c46f-4621-8a44-ab28a6652f73 | autofocusformer-image-segmentation-off-the | 2304.12406 | null | https://arxiv.org/abs/2304.12406v1 | https://arxiv.org/pdf/2304.12406v1.pdf | AutoFocusFormer: Image Segmentation off the Grid | Real world images often have highly imbalanced content density. Some areas are very uniform, e.g., large patches of blue sky, while other areas are scattered with many small objects. Yet, the commonly used successive grid downsampling strategy in convolutional deep networks treats all areas equally. Hence, small objects are represented in very few spatial locations, leading to worse results in tasks such as segmentation. Intuitively, retaining more pixels representing small objects during downsampling helps to preserve important information. To achieve this, we propose AutoFocusFormer (AFF), a local-attention transformer image recognition backbone, which performs adaptive downsampling by learning to retain the most important pixels for the task. Since adaptive downsampling generates a set of pixels irregularly distributed on the image plane, we abandon the classic grid structure. Instead, we develop a novel point-based local attention block, facilitated by a balanced clustering module and a learnable neighborhood merging module, which yields representations for our point-based versions of state-of-the-art segmentation heads. Experiments show that our AutoFocusFormer (AFF) improves significantly over baseline models of similar sizes. | ['Li Fuxin', 'Alex Colburn', 'Alex Schwing', 'Zhile Ren', 'Alvin Wan', 'Shuangfei Zhai', 'Kaushik Patnaik', 'Chen Ziwen'] | 2023-04-24 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ziwen_AutoFocusFormer_Image_Segmentation_off_the_Grid_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ziwen_AutoFocusFormer_Image_Segmentation_off_the_Grid_CVPR_2023_paper.pdf | cvpr-2023-1 | ['panoptic-segmentation'] | ['computer-vision'] | [ 1.87799275e-01 1.24348745e-01 -1.41307354e-01 -2.73560584e-01
-3.75464171e-01 -2.49703169e-01 5.08288920e-01 8.97469819e-02
-3.35877389e-01 7.48706400e-01 2.87824869e-01 1.89086333e-01
1.29486918e-01 -1.10020781e+00 -1.14548004e+00 -1.08666754e+00
3.88024151e-01 3.79673392e-01 5.91621280e-01 -3.55132893e-02
1.82152629e-01 4.19632167e-01 -1.74505007e+00 3.11240286e-01
1.19403553e+00 9.24824417e-01 6.38726711e-01 1.69423655e-01
-1.95233718e-01 8.68718326e-01 -6.38427854e-01 -1.08946033e-01
2.90650070e-01 -3.86956304e-01 -7.01568604e-01 3.13732952e-01
8.45634580e-01 -2.97471911e-01 -2.86156327e-01 1.39041948e+00
2.78900832e-01 5.08468807e-01 5.30168593e-01 -8.04054320e-01
-8.41870904e-01 9.30758655e-01 -1.15923810e+00 5.02075851e-01
-3.83528173e-01 5.75570948e-02 9.99172986e-01 -9.01773572e-01
4.91718441e-01 1.20470607e+00 5.92699766e-01 1.98944420e-01
-1.30872166e+00 -6.39405966e-01 4.22334701e-01 3.21425527e-01
-1.46899915e+00 -2.16085494e-01 9.14039731e-01 4.99475896e-02
6.87155068e-01 3.19112450e-01 7.14169443e-01 6.67365611e-01
1.98584810e-01 9.09170866e-01 8.43083918e-01 -1.93739519e-01
3.12897533e-01 7.90707096e-02 3.27703580e-02 2.14482933e-01
4.84717369e-01 -5.30699432e-01 -5.35757601e-01 3.22992057e-01
9.98997331e-01 5.41371286e-01 -5.36216915e-01 -6.48547828e-01
-1.10489142e+00 6.62516773e-01 9.35646653e-01 5.54594696e-01
-5.20019829e-01 9.12590548e-02 -1.77721605e-02 -3.02493144e-02
8.44796658e-01 3.48092288e-01 -2.20657051e-01 2.57853508e-01
-1.24490309e+00 3.86537164e-01 1.43456578e-01 8.66982698e-01
1.25645828e+00 -8.98127109e-02 -2.22612143e-01 1.09870291e+00
-5.97424954e-02 1.67072460e-01 5.49019516e-01 -9.87954915e-01
3.44314694e-01 8.07373583e-01 -1.26571590e-02 -1.08779287e+00
-3.92856508e-01 -1.07867444e+00 -1.21722889e+00 2.87451148e-02
3.98443699e-01 1.44974381e-01 -1.02877283e+00 1.77494502e+00
3.85224909e-01 3.58383179e-01 -3.76050144e-01 1.19782054e+00
4.19272929e-01 6.55707955e-01 -9.14265811e-02 -1.42688707e-01
1.38882470e+00 -1.28230560e+00 -5.70680559e-01 -4.07527506e-01
3.38858187e-01 -4.37174171e-01 1.31952667e+00 2.94764012e-01
-1.31679893e+00 -7.82119155e-01 -9.13299978e-01 -4.61768031e-01
-3.09573442e-01 -2.28547513e-01 5.34730613e-01 3.66140246e-01
-1.14232624e+00 7.43881583e-01 -7.31038034e-01 -2.03017443e-01
1.01888275e+00 2.29331404e-02 -1.91939279e-01 -2.76644021e-01
-8.30692947e-01 5.03731847e-01 2.92147607e-01 -1.35250703e-01
-6.67554855e-01 -1.03045344e+00 -8.37228835e-01 5.21706104e-01
3.61446351e-01 -6.57782555e-01 1.02219975e+00 -1.20328975e+00
-9.54068720e-01 7.50795603e-01 -4.13596004e-01 -5.42314172e-01
3.30810606e-01 -4.43617970e-01 7.27896541e-02 1.98410094e-01
4.27348047e-01 8.28027070e-01 1.01188195e+00 -1.31479073e+00
-7.94096053e-01 -7.08946645e-01 -1.04516149e-01 4.34612155e-01
-6.02357209e-01 -3.40146601e-01 -5.75898826e-01 -9.09365296e-01
3.33743900e-01 -5.05679309e-01 -3.62827688e-01 1.00195045e-02
-3.83074164e-01 -1.10654093e-01 8.75944018e-01 -3.51345599e-01
1.21965992e+00 -2.25530338e+00 8.10097530e-02 3.84092480e-02
5.65778017e-01 -5.73637933e-02 3.12129017e-02 -1.11889318e-01
-3.11409801e-01 3.07568815e-02 -4.15128440e-01 -4.41858619e-01
-2.51323164e-01 1.29238516e-01 -4.64079559e-01 6.03658438e-01
1.97393358e-01 8.74925911e-01 -1.04553711e+00 -2.94999510e-01
4.22409534e-01 3.82405102e-01 -7.28193283e-01 -1.18951924e-01
-2.14191169e-01 3.07790756e-01 -3.78095597e-01 5.28858244e-01
1.18692994e+00 -4.92332906e-01 -2.44569063e-01 -2.42160365e-01
-3.40350300e-01 1.01071768e-01 -1.04906452e+00 1.84339976e+00
-2.69825876e-01 5.49720585e-01 1.53328136e-01 -1.07343435e+00
5.92168629e-01 -2.48017490e-01 5.47626019e-01 -6.44603729e-01
1.15188099e-01 -9.35546830e-02 -1.62018850e-01 -1.01158641e-01
6.93687320e-01 -1.46420738e-02 2.45704085e-01 3.97814155e-01
1.39899608e-02 -1.05651848e-01 1.46917045e-01 2.88663507e-01
1.03299439e+00 -6.02019280e-02 2.75593340e-01 -5.29052913e-01
1.57993093e-01 8.01242050e-03 6.48654580e-01 9.47828889e-01
-2.59801775e-01 1.34812057e+00 3.65714341e-01 -5.30395091e-01
-1.02683413e+00 -8.21583986e-01 -2.78804541e-01 1.20673037e+00
6.42529488e-01 -3.07990402e-01 -1.07480991e+00 -4.83035296e-01
-1.51294157e-01 5.79127371e-01 -8.88728201e-01 -2.15050697e-01
-7.18690574e-01 -1.01534796e+00 3.14906181e-04 6.33122146e-01
8.07474971e-01 -1.10866499e+00 -6.73094034e-01 2.37222284e-01
-3.49468976e-01 -6.85237288e-01 -6.59668684e-01 4.42256451e-01
-8.06822419e-01 -7.52859294e-01 -1.20330596e+00 -8.10721636e-01
7.69918501e-01 1.00909925e+00 1.28543186e+00 -6.52129278e-02
-6.13816492e-02 -3.20414633e-01 -3.63760024e-01 -3.46060008e-01
3.33094060e-01 2.18894660e-01 -1.86372712e-01 2.62693286e-01
3.76837999e-01 -7.61818409e-01 -9.41660106e-01 2.66335934e-01
-1.03230357e+00 7.66958594e-02 7.16204107e-01 9.39394772e-01
9.05221879e-01 3.72308254e-01 1.68823123e-01 -9.78907168e-01
1.42180935e-01 -5.54925799e-01 -2.60072201e-01 -7.22759888e-02
-2.09793329e-01 -6.20967597e-02 8.76416922e-01 -3.74737531e-01
-9.56031978e-01 -2.46191144e-01 -2.51139924e-02 -9.16279018e-01
-3.25933963e-01 1.60398364e-01 -3.73854160e-01 7.67447650e-02
6.10698760e-01 4.69213307e-01 -1.86268404e-01 -6.84680045e-01
4.29745317e-01 3.70021433e-01 6.55026495e-01 -2.16186836e-01
6.29743040e-01 8.85792911e-01 -4.22972679e-01 -9.53140318e-01
-8.56049359e-01 -5.68914294e-01 -4.57098335e-01 -2.55792178e-02
7.07343638e-01 -1.01924574e+00 -2.09857494e-01 6.54302180e-01
-7.95228183e-01 -4.06027108e-01 -8.21615875e-01 1.18230902e-01
-3.99221689e-01 1.96387053e-01 -3.94742280e-01 -4.97840136e-01
-8.43959898e-02 -1.05804241e+00 1.42171991e+00 5.04394889e-01
-1.35054022e-01 -5.17947614e-01 -1.09602317e-01 2.22533479e-01
4.24311876e-01 -1.74408823e-01 7.83936322e-01 -2.79021591e-01
-7.45817423e-01 3.13059717e-01 -4.97350961e-01 2.07510337e-01
1.26179993e-01 -3.57973933e-01 -1.07208121e+00 -2.86315352e-01
2.79337615e-01 -1.76278323e-01 1.45486355e+00 8.45877409e-01
1.82097375e+00 -2.27516487e-01 -5.61605453e-01 1.04303300e+00
1.17118573e+00 -1.63176417e-01 7.74653494e-01 4.05043036e-01
8.78796279e-01 6.33483171e-01 4.28402364e-01 3.42206478e-01
1.96119621e-01 5.47963321e-01 5.61779439e-01 -3.25590849e-01
-4.24172401e-01 -2.73495555e-01 -6.93703666e-02 3.19938093e-01
-3.87673751e-02 -2.73326755e-01 -5.99340022e-01 8.37819457e-01
-1.82267797e+00 -1.00624001e+00 1.27846047e-01 2.00476289e+00
8.94709289e-01 1.12158261e-01 -3.55479978e-02 4.49492931e-02
8.58341038e-01 5.93399405e-01 -7.89236546e-01 3.04303497e-01
-4.41514939e-01 1.94992170e-01 5.97465575e-01 2.77009070e-01
-1.26018429e+00 1.03485799e+00 5.47154713e+00 1.35335279e+00
-1.08638251e+00 1.42609209e-01 1.30426502e+00 -4.22311842e-01
-3.82894754e-01 -2.59380370e-01 -6.98680282e-01 7.21799076e-01
2.63352662e-01 9.70846340e-02 3.14484239e-01 1.02686048e+00
3.17254812e-01 -1.51567683e-01 -5.81716835e-01 1.08587956e+00
1.20605215e-01 -1.39916956e+00 2.72212893e-01 -1.17942337e-02
1.06465757e+00 2.53329545e-01 2.58910030e-01 5.84480576e-02
2.03786105e-01 -9.07120645e-01 9.68061805e-01 2.92630345e-01
6.75312161e-01 -8.60880435e-01 5.89080393e-01 4.06542569e-01
-1.07039475e+00 -1.33843169e-01 -8.72495711e-01 5.56836873e-02
-1.78191699e-02 1.08148241e+00 -9.08740535e-02 2.71033049e-01
1.25924838e+00 9.54199910e-01 -4.77811575e-01 1.01747489e+00
5.30221760e-02 4.44779903e-01 -4.46672916e-01 2.41526663e-01
3.96315724e-01 -3.42267632e-01 5.04916966e-01 8.90483320e-01
3.08950722e-01 -3.57689038e-02 -2.13444650e-01 1.13338685e+00
-3.61817986e-01 1.14202490e-02 -4.93585408e-01 4.04851794e-01
3.77736121e-01 1.25984991e+00 -1.13908446e+00 -4.93341506e-01
-4.88748223e-01 1.05966830e+00 5.86140573e-01 4.29094642e-01
-7.59496987e-01 -3.83010477e-01 7.40671992e-01 4.71611977e-01
6.66804731e-01 2.99493015e-01 -7.47157872e-01 -1.36727691e+00
9.64922532e-02 -9.01606917e-01 1.50073498e-01 -7.37591386e-01
-1.13784385e+00 5.53788543e-01 -3.12611610e-01 -1.29166067e+00
2.30246231e-01 -1.52288884e-01 -6.72988117e-01 6.78284466e-01
-1.56278944e+00 -1.03458476e+00 -6.52828634e-01 5.37712693e-01
9.54697073e-01 2.07401425e-01 1.25760913e-01 3.35659772e-01
-6.10017061e-01 4.48689878e-01 2.01519713e-01 -7.11564273e-02
8.14862370e-01 -1.41515183e+00 3.33653539e-01 1.13993704e+00
3.05779546e-01 6.29822314e-01 5.96626222e-01 -6.11359835e-01
-7.87727714e-01 -1.50218689e+00 5.96530914e-01 -2.40433842e-01
3.60814661e-01 -3.74942750e-01 -1.25238216e+00 4.72613364e-01
3.28216940e-01 2.08451822e-01 2.39317179e-01 -7.06051812e-02
-1.89009622e-01 -2.67300308e-01 -8.80977690e-01 7.66485333e-01
1.13033938e+00 -1.13570966e-01 -4.77070093e-01 3.15936685e-01
6.97522581e-01 -2.85049170e-01 -3.26695830e-01 3.18944126e-01
2.09522620e-01 -1.40302229e+00 1.04370463e+00 -2.04339832e-01
6.17646575e-01 -4.26139176e-01 1.19418234e-01 -1.52952600e+00
-7.54028320e-01 -5.22586286e-01 -1.56023189e-01 9.78021502e-01
7.36316592e-02 -6.24798298e-01 1.02633440e+00 2.48340338e-01
-4.30644363e-01 -8.04747343e-01 -8.52351844e-01 -3.66099536e-01
4.17333730e-02 -1.60704866e-01 7.79581368e-01 9.07651544e-01
-2.91298360e-01 2.41565272e-01 -1.88279986e-01 2.13234834e-02
6.68750584e-01 4.32405174e-01 6.17308080e-01 -9.71821487e-01
-1.31172195e-01 -7.64650345e-01 -2.24098206e-01 -1.58706117e+00
-1.89937204e-01 -4.84087676e-01 6.23649172e-02 -1.23114872e+00
5.13781905e-01 -3.73410463e-01 -4.17055935e-01 4.08782423e-01
-5.88602841e-01 7.66135514e-01 -1.31461015e-02 4.38437551e-01
-6.92575514e-01 7.71894991e-01 1.51320755e+00 -2.32328624e-01
-1.40141234e-01 -1.33638352e-01 -1.14759350e+00 8.98441374e-01
5.68576455e-01 -3.70415241e-01 -4.12591517e-01 -7.00036526e-01
-3.84831280e-02 -6.60521388e-01 3.40635866e-01 -1.19633400e+00
3.19485664e-01 -1.42890951e-02 7.70510197e-01 -8.67251992e-01
1.20321229e-01 -7.18281865e-01 -9.84035209e-02 2.50427544e-01
-1.46417037e-01 -2.73458809e-01 -5.27460948e-02 5.17976701e-01
-4.18272525e-01 2.92501859e-02 1.08791351e+00 -3.05790722e-01
-6.35934234e-01 4.85521436e-01 -1.56470791e-01 2.15128642e-02
1.04699254e+00 -4.27683979e-01 -3.99695516e-01 -2.67624021e-01
-5.06702006e-01 1.14494912e-01 8.07555735e-01 2.77614802e-01
4.16473180e-01 -1.01629281e+00 -5.43469906e-01 6.06258988e-01
-1.27496451e-01 7.49517024e-01 5.94218075e-01 8.99549723e-01
-5.64067483e-01 2.30445892e-01 -2.67738134e-01 -6.34019494e-01
-7.76141167e-01 5.59538424e-01 1.93059668e-01 -6.65219352e-02
-9.52226639e-01 1.15606534e+00 1.07466269e+00 -5.82418814e-02
1.64324015e-01 -6.59125447e-01 -3.42678815e-01 1.39576271e-01
8.50719094e-01 2.28552029e-01 1.35841966e-01 -6.16230905e-01
-1.16738327e-01 5.62049448e-01 -4.53052759e-01 3.75829935e-01
1.52848291e+00 -3.75482351e-01 -2.74422169e-01 2.51322925e-01
9.13501203e-01 3.80210206e-02 -1.65292764e+00 -3.55445772e-01
-6.11157954e-01 -7.86394954e-01 3.92087311e-01 -2.15004101e-01
-1.40839052e+00 9.24465001e-01 4.03103560e-01 3.33494693e-01
1.41903245e+00 2.84409165e-01 8.20381999e-01 1.91256423e-02
3.00343424e-01 -9.53125000e-01 -6.76376745e-02 4.09861803e-01
7.42125928e-01 -1.11040545e+00 -5.32453172e-02 -4.86811668e-01
-5.17131448e-01 7.02189922e-01 9.66617227e-01 -3.75480890e-01
6.37252212e-01 4.13349092e-01 -1.25408575e-01 -6.62329942e-02
-6.11535490e-01 -2.60783643e-01 7.52159134e-02 5.26648819e-01
4.99089323e-02 -1.52508333e-01 2.93623824e-02 6.22302830e-01
-2.02579185e-01 -3.32277030e-01 2.60307223e-01 6.56677365e-01
-7.23870516e-01 -4.33698565e-01 -4.87994999e-01 7.74872065e-01
-4.73020852e-01 -2.72434413e-01 -1.53484344e-01 5.71715653e-01
3.82720888e-01 4.77991909e-01 7.05504954e-01 -1.52564704e-01
7.69662112e-02 -3.40226352e-01 2.11629912e-01 -6.17563188e-01
-3.72623056e-01 4.26576346e-01 -6.62497997e-01 -7.93127477e-01
-4.60212082e-01 -6.62970066e-01 -9.78990972e-01 -3.09833974e-01
-3.93085331e-01 -9.91927758e-02 -7.76111009e-03 8.30972075e-01
4.15833563e-01 8.28891516e-01 5.27260900e-01 -1.17556655e+00
-2.44984999e-01 -1.14262199e+00 -1.00378215e+00 5.25737762e-01
4.84810770e-01 -7.24464118e-01 -3.81320864e-01 2.51438208e-02] | [10.387979507446289, -0.9990102648735046] |
19fcef1d-f76a-46df-b13a-bb588ef6189d | combined-scaling-for-zero-shot-transfer | 2111.10050 | null | https://arxiv.org/abs/2111.10050v3 | https://arxiv.org/pdf/2111.10050v3.pdf | Combined Scaling for Zero-shot Transfer Learning | We present a combined scaling method - named BASIC - that achieves 85.7% top-1 accuracy on the ImageNet ILSVRC-2012 validation set without learning from any labeled ImageNet example. This accuracy surpasses best published similar models - CLIP and ALIGN - by 9.3%. Our BASIC model also shows significant improvements in robustness benchmarks. For instance, on 5 test sets with natural distribution shifts such as ImageNet-{A,R,V2,Sketch} and ObjectNet, our model achieves 84.3% top-1 average accuracy, only a small drop from its original ImageNet accuracy. To achieve these results, we scale up the contrastive learning framework of CLIP and ALIGN in three dimensions: data size, model size, and batch size. Our dataset has 6.6B noisy image-text pairs, which is 4x larger than ALIGN, and 16x larger than CLIP. Our largest model has 3B weights, which is 3.75x larger in parameters and 8x larger in FLOPs than ALIGN and CLIP. Finally, our batch size is 65536 which is 2x more than CLIP and 4x more than ALIGN. We encountered two main challenges with the scaling rules of BASIC. First, the main challenge with implementing the combined scaling rules of BASIC is the limited memory of accelerators, such as GPUs and TPUs. To overcome the memory limit, we propose two simple methods which make use of gradient checkpointing and model parallelism. Second, while increasing the dataset size and the model size has been the defacto method to improve the performance of deep learning models like BASIC, the effect of a large contrastive batch size on such contrastive-trained image-text models is not well-understood. To shed light on the benefits of large contrastive batch sizes, we develop a theoretical framework which shows that larger contrastive batch sizes lead to smaller generalization gaps for image-text models such as BASIC. | ['Quoc V. Le', 'Yonghui Wu', 'Minh-Thang Luong', 'Yi-Ting Chen', 'Jiahui Yu', 'Kenji Kawaguchi', 'Mingxing Tan', 'Adams Wei Yu', 'Hanxiao Liu', 'Golnaz Ghiasi', 'Zihang Dai', 'Hieu Pham'] | 2021-11-19 | null | null | null | null | ['zero-shot-transfer-image-classification'] | ['computer-vision'] | [-1.19625047e-01 -2.61190176e-01 -1.41199097e-01 -3.71094912e-01
-8.04882646e-01 -4.79600191e-01 4.65785533e-01 -1.53631300e-01
-8.38293314e-01 4.11924988e-01 -4.33668979e-02 -5.35836697e-01
2.47906178e-01 -4.61297750e-01 -9.48443115e-01 -4.73699719e-01
-6.18942752e-02 2.13023394e-01 4.03313488e-01 -1.90854654e-01
2.36143976e-01 4.11045820e-01 -1.46817183e+00 4.47521597e-01
6.64524138e-01 9.68568206e-01 1.04929954e-01 9.21000242e-01
5.86756878e-02 6.18675470e-01 -6.72272563e-01 -3.61723006e-01
5.74413419e-01 -2.51814783e-01 -5.81938863e-01 -2.17972115e-01
1.24328923e+00 -5.42180002e-01 -4.08217877e-01 8.88381839e-01
7.72897005e-01 4.18409519e-02 4.86146718e-01 -1.38148797e+00
-4.26131934e-01 7.02620566e-01 -8.89900923e-01 1.64751664e-01
-1.76479295e-01 3.90756488e-01 8.29336584e-01 -7.29498982e-01
5.35034657e-01 1.22648537e+00 9.65201080e-01 7.90177524e-01
-1.25628042e+00 -1.07890606e+00 2.00690046e-01 1.26602948e-01
-1.37225771e+00 -3.39794427e-01 1.53882280e-01 -2.68870890e-01
1.13600326e+00 4.15871412e-01 3.91674966e-01 1.15141821e+00
1.54143527e-01 7.94387519e-01 1.28864193e+00 -2.66728938e-01
1.25509024e-01 2.80588478e-01 2.84327358e-01 6.18866742e-01
2.67052770e-01 1.29082292e-01 -3.90957713e-01 -6.12313077e-02
5.97318172e-01 -2.01357841e-01 -1.76059306e-01 4.96245846e-02
-1.00496161e+00 5.86188734e-01 5.59961915e-01 2.43232027e-01
9.64829177e-02 5.81103861e-01 7.25018084e-01 3.76846731e-01
4.26153481e-01 5.35336494e-01 -6.38971090e-01 -2.56047040e-01
-1.17207587e+00 2.75583327e-01 7.28996515e-01 9.41212177e-01
5.91099799e-01 1.26833707e-01 -6.96763024e-03 8.48386765e-01
-3.25516239e-02 7.94014394e-01 4.82132375e-01 -8.59958649e-01
7.67014921e-01 3.83559376e-01 -2.14055821e-01 -8.28745544e-01
-6.91246450e-01 -4.48786318e-01 -9.29708302e-01 3.58532339e-01
6.00476027e-01 -1.81525901e-01 -1.09203589e+00 1.79943287e+00
3.95512022e-02 1.19660109e-01 -1.12759754e-01 7.90493965e-01
7.77491450e-01 6.97930753e-01 2.63294458e-01 2.44195446e-01
1.38062835e+00 -1.00267279e+00 -2.45051280e-01 -1.88325599e-01
1.10300231e+00 -9.61686909e-01 1.53041911e+00 4.41710562e-01
-1.02471268e+00 -6.76482201e-01 -1.34229934e+00 -2.02031478e-01
-3.80715281e-01 2.78057545e-01 4.32499588e-01 8.49972188e-01
-1.13009679e+00 7.32663691e-01 -7.68691003e-01 -4.63766247e-01
4.47694719e-01 4.88609523e-01 -4.31181014e-01 -1.08409107e-01
-8.34108591e-01 8.24893296e-01 3.45609188e-01 -2.28790745e-01
-5.60548961e-01 -1.03894520e+00 -5.95278680e-01 2.00515598e-01
2.11966440e-01 -3.30359131e-01 9.70905542e-01 -8.21521461e-01
-1.36412323e+00 7.96060085e-01 1.07457153e-01 -7.31945753e-01
8.27969372e-01 -2.97875077e-01 -2.64963508e-01 -4.98227291e-02
-1.81117594e-01 1.21014023e+00 8.55912685e-01 -1.21956182e+00
-3.80106241e-01 -4.17254210e-01 -1.38573423e-01 2.13928357e-01
-5.12990892e-01 4.07833830e-02 -6.66617632e-01 -5.65672278e-01
-1.01880461e-01 -1.26195776e+00 -2.60326594e-01 1.32045552e-01
-3.55337858e-01 8.93014148e-02 9.16860223e-01 -4.12496924e-01
9.47639227e-01 -2.32206488e+00 -4.48638439e-01 1.73376665e-01
1.81543484e-01 5.36476493e-01 -5.25647104e-01 8.88807625e-02
-1.36284977e-01 3.63274127e-01 1.45730764e-01 -4.56892967e-01
-1.48689412e-02 1.58619985e-01 -3.18063647e-01 4.82225060e-01
2.13300586e-02 7.69282639e-01 -4.35259521e-01 -3.14478159e-01
2.38873601e-01 4.05555487e-01 -7.18635321e-01 -2.23600298e-01
-3.27504352e-02 1.18510956e-02 -1.01061955e-01 3.04875344e-01
1.12871695e+00 -3.09288830e-01 3.95438634e-02 -4.01650667e-01
6.21140096e-03 6.70514926e-02 -1.17776418e+00 1.40412974e+00
-5.85339665e-01 8.84397388e-01 5.74251376e-02 -7.68070281e-01
7.32582629e-01 -3.00609320e-02 2.91637331e-01 -9.77484822e-01
1.69910938e-01 1.80609703e-01 -7.77685316e-03 -3.47206801e-01
6.14569187e-01 1.76700145e-01 1.42315686e-01 3.65433276e-01
1.55315027e-02 -3.28401059e-01 2.21918926e-01 2.65941679e-01
1.21757853e+00 -8.65088180e-02 -1.49984211e-01 -4.55403358e-01
2.66682595e-01 -9.37551484e-02 2.45973527e-01 1.05674255e+00
-2.44417280e-01 6.46298051e-01 6.95946336e-01 -6.46115541e-01
-1.28755200e+00 -7.94036686e-01 -3.30276757e-01 1.22804451e+00
6.67634234e-02 -7.09409475e-01 -8.42365801e-01 -8.62169027e-01
-6.93283789e-03 6.07312799e-01 -5.43222487e-01 -5.79141974e-02
-7.17706978e-01 -1.07323313e+00 9.83819783e-01 7.71887600e-01
8.44712198e-01 -6.44823551e-01 -5.06561637e-01 -1.64207473e-01
1.23329587e-01 -1.27209425e+00 -5.45612335e-01 1.65009975e-01
-8.85070503e-01 -9.59124386e-01 -5.94141722e-01 -6.07921183e-01
6.44807756e-01 1.85379863e-01 1.10087442e+00 3.21430296e-01
-2.96066970e-01 1.17144428e-01 -1.99220955e-01 -4.79023457e-01
-4.94734287e-01 2.59579092e-01 5.41290864e-02 -5.09994686e-01
9.25324708e-02 -3.58809292e-01 -6.67857409e-01 6.24479890e-01
-1.02186263e+00 1.95556298e-01 5.59019506e-01 8.91784608e-01
2.71430820e-01 -2.50849754e-01 2.63080627e-01 -5.96989930e-01
4.26678807e-01 -1.68503806e-01 -7.58028984e-01 1.00328192e-01
-6.75655484e-01 8.98137316e-02 8.05826843e-01 -8.24471533e-01
-6.88150346e-01 1.19429283e-01 -4.20727074e-01 -2.64630914e-01
1.31934071e-02 5.44999093e-02 1.60785660e-01 -2.23174274e-01
1.04996049e+00 -3.41051407e-02 8.90814234e-03 -3.59917969e-01
3.87070358e-01 6.37531161e-01 3.71387273e-01 -7.08706558e-01
4.64605302e-01 4.15069938e-01 -1.07620664e-01 -8.41034293e-01
-5.59641480e-01 -3.18643004e-01 -2.92932004e-01 9.76634305e-03
7.41000235e-01 -1.07569754e+00 -9.94956732e-01 7.27154076e-01
-9.64107990e-01 -7.58671343e-01 -1.52970076e-01 5.71472764e-01
-3.94410104e-01 2.41264910e-01 -8.66808593e-01 -3.79980117e-01
-5.27396142e-01 -1.40601516e+00 1.15573967e+00 1.77112892e-02
-2.62131710e-02 -7.24229872e-01 -2.21658885e-01 2.76033193e-01
5.82942426e-01 -5.55495843e-02 8.39102209e-01 -7.58179784e-01
-3.19068670e-01 -9.42710713e-02 -6.26989186e-01 6.59342945e-01
-3.99606138e-01 -8.23366418e-02 -9.10905600e-01 -5.44440448e-01
1.49181411e-02 -5.21720767e-01 9.66710746e-01 3.21399808e-01
1.20778179e+00 -2.13643253e-01 -3.49849850e-01 7.77653277e-01
1.50390518e+00 -4.11911495e-03 7.69598246e-01 3.72695148e-01
7.03006566e-01 1.53788924e-01 3.94439071e-01 3.03860247e-01
1.67590275e-01 7.96025753e-01 3.22868109e-01 -2.80678093e-01
-4.53605622e-01 -1.94069371e-02 4.77154940e-01 6.86331391e-01
3.07130218e-02 -2.01580688e-01 -9.83818114e-01 3.01975638e-01
-1.60436797e+00 -7.50358880e-01 -2.23764136e-01 2.22240829e+00
8.90157402e-01 2.72162110e-01 1.54669508e-02 -2.61205584e-02
5.49620926e-01 1.62852015e-02 -6.39267981e-01 -4.61380124e-01
-1.01974063e-01 2.14786306e-01 9.42928016e-01 5.12096941e-01
-9.89270806e-01 1.02811491e+00 6.59371519e+00 1.15158176e+00
-1.42290807e+00 1.31698951e-01 9.03813124e-01 -5.07131636e-01
3.01359713e-01 -2.49163166e-01 -1.18561852e+00 5.64921021e-01
1.10888994e+00 1.82575703e-01 3.36700141e-01 1.09167099e+00
3.48263942e-02 -2.14868397e-01 -1.17487299e+00 1.17637038e+00
1.88786373e-01 -1.38016307e+00 -3.04366816e-02 1.65361404e-01
9.44630921e-01 5.70953608e-01 1.88195631e-01 4.83449399e-01
3.04282844e-01 -9.93901730e-01 5.92590272e-01 -8.05595964e-02
8.83372784e-01 -5.83695829e-01 7.25417912e-01 2.58827835e-01
-8.68587136e-01 -5.52551672e-02 -6.38569236e-01 1.70729205e-01
-2.21423075e-01 5.91134906e-01 -6.88059807e-01 1.02663092e-01
1.03038788e+00 4.13914561e-01 -8.80704880e-01 8.40326905e-01
9.45693776e-02 6.70661271e-01 -7.92575955e-01 -1.65561840e-01
3.48293573e-01 1.43617809e-01 2.98192441e-01 1.42070770e+00
2.00322211e-01 -1.35938972e-01 9.31567624e-02 5.56893647e-01
-4.05078739e-01 1.18233673e-01 -4.03602540e-01 4.49400067e-01
4.08154577e-01 1.29168868e+00 -6.95757985e-01 -6.37017071e-01
-2.63902664e-01 8.78941476e-01 2.57731140e-01 2.54254520e-01
-1.13650596e+00 -3.13751489e-01 4.72171485e-01 1.36085823e-01
2.93060124e-01 -3.39776158e-01 -5.47996521e-01 -1.08289254e+00
1.90195441e-01 -1.25063598e+00 2.24593863e-01 -6.57025278e-01
-1.06997311e+00 7.67024815e-01 5.68312295e-02 -9.80230510e-01
1.76409781e-01 -8.99505675e-01 -4.07914013e-01 6.85903728e-01
-1.16559303e+00 -1.06291211e+00 -5.22786677e-01 6.13216400e-01
4.99019742e-01 -9.18145478e-02 6.44421399e-01 4.30425704e-01
-5.84801555e-01 1.18631470e+00 3.60901892e-01 2.21534997e-01
9.37170267e-01 -9.80810881e-01 8.08378220e-01 6.90429211e-01
4.31310646e-02 5.02663553e-01 5.92402101e-01 -4.60593283e-01
-1.26851678e+00 -1.03291738e+00 5.40350795e-01 -5.02865732e-01
7.19814479e-01 -7.49350250e-01 -9.11748707e-01 5.81799865e-01
1.64366126e-01 3.68429989e-01 2.38527983e-01 9.50324312e-02
-7.62456298e-01 -3.25259298e-01 -1.25293219e+00 9.63399589e-01
1.01271689e+00 -2.87720233e-01 -6.01330884e-02 5.66386402e-01
5.78056574e-01 -6.75149441e-01 -7.71944404e-01 4.56896275e-01
5.76088786e-01 -1.01376677e+00 8.04586887e-01 -4.42471385e-01
2.86645830e-01 1.05242655e-02 -2.38268346e-01 -1.15227783e+00
-6.49207234e-02 -4.56996024e-01 1.66012481e-01 9.89213467e-01
4.67703789e-01 -8.23490500e-01 9.45396245e-01 6.57699525e-01
4.45568338e-02 -9.03731823e-01 -9.21191335e-01 -1.08843613e+00
4.05121535e-01 -6.18401170e-01 3.76800418e-01 8.17558169e-01
-1.86836228e-01 3.83319855e-01 -2.90026486e-01 -1.48025692e-01
2.86303103e-01 -4.12557900e-01 9.45005715e-01 -8.27812076e-01
-3.02660793e-01 -5.19577205e-01 -4.32743132e-01 -1.21686006e+00
3.50702591e-02 -7.19310105e-01 -1.64899409e-01 -1.01912928e+00
2.80948371e-01 -8.05205703e-01 -6.13463037e-02 8.01544130e-01
-8.46376419e-02 5.47032356e-01 7.01309264e-01 1.76644012e-01
-4.51146424e-01 2.69826055e-01 1.08358824e+00 -2.66994655e-01
-2.15647906e-01 -2.92874694e-01 -4.54805017e-01 7.16091931e-01
9.76448774e-01 -5.55527866e-01 -3.89394641e-01 -6.32379353e-01
2.54273027e-01 -2.66339302e-01 3.45737517e-01 -1.28037572e+00
2.08306223e-01 2.35863060e-01 5.37764013e-01 -5.43594182e-01
3.08254391e-01 -7.26350725e-01 -1.67227075e-01 5.41776657e-01
-4.73319709e-01 3.43382001e-01 8.06394696e-01 2.14675859e-01
9.60507393e-02 -2.39893094e-01 9.88947690e-01 6.89671412e-02
-4.59566772e-01 1.84891224e-01 -2.94848680e-01 2.30201110e-01
7.96362638e-01 -3.88938971e-02 -6.92427754e-01 -2.09583178e-01
-4.42706227e-01 7.65219852e-02 3.80153924e-01 6.57256544e-01
3.15586299e-01 -1.18268824e+00 -7.97971904e-01 2.26056755e-01
-6.67841360e-02 -2.18710482e-01 4.05983657e-01 9.37349796e-01
-6.80101752e-01 2.60250658e-01 -2.15939596e-01 -8.18730831e-01
-1.62938058e+00 5.69653153e-01 4.35925364e-01 -3.95662516e-01
-6.70038700e-01 7.51999497e-01 2.68350691e-01 -4.93004084e-01
3.92549515e-01 -4.52067465e-01 3.06508899e-01 -3.12630683e-01
7.74075747e-01 4.52188224e-01 3.57953191e-01 -3.40525210e-01
-3.75247180e-01 7.30901241e-01 -5.67406476e-01 -1.14218391e-01
1.08949411e+00 2.09359497e-01 6.11124150e-02 -9.83844995e-02
1.47039628e+00 -6.06980063e-02 -1.30426538e+00 6.75580092e-03
-1.86053082e-01 -3.59899819e-01 5.26065864e-02 -9.09875929e-01
-1.14196432e+00 8.40982497e-01 1.01524150e+00 -1.17638275e-01
1.09555280e+00 -1.68990508e-01 7.86046386e-01 5.06327510e-01
2.06001103e-01 -1.27881706e+00 1.85338587e-01 6.38917565e-01
8.67405534e-01 -1.30352616e+00 1.76519468e-01 -1.40621036e-01
-5.01517773e-01 9.03695464e-01 9.44794714e-01 -1.82105571e-01
5.24983466e-01 6.37060225e-01 1.69718146e-01 1.54557908e-02
-7.50340462e-01 8.10919181e-02 2.14447156e-01 3.72578204e-01
2.36274004e-01 -9.80995446e-02 -2.19401494e-01 2.76953846e-01
-2.87677109e-01 -1.54972509e-01 3.72646481e-01 6.23643398e-01
-1.96499646e-01 -8.78093541e-01 -3.67857724e-01 3.90930563e-01
-6.05350494e-01 -3.53831023e-01 -1.48189232e-01 1.09165299e+00
2.68373564e-02 8.46138358e-01 2.73286432e-01 -6.00939870e-01
4.80895698e-01 -2.27751717e-01 5.30315340e-01 -2.44526148e-01
-8.30394030e-01 -1.48877323e-01 -8.89863893e-02 -6.39418185e-01
-2.70201173e-02 -2.63027757e-01 -1.21239877e+00 -7.83790469e-01
-5.47577858e-01 -2.73982793e-01 1.15191066e+00 7.01248884e-01
5.61327100e-01 2.98229039e-01 3.58218938e-01 -8.98507953e-01
-7.55172908e-01 -9.43528116e-01 -3.39961171e-01 4.44860816e-01
6.04757704e-02 -2.12334111e-01 -4.73328859e-01 -1.48801342e-01] | [9.227837562561035, 2.2668113708496094] |
87fe3203-198c-4ef7-ad14-d542056be2c5 | style-erd-responsive-and-coherent-online | 2203.02574 | null | https://arxiv.org/abs/2203.02574v2 | https://arxiv.org/pdf/2203.02574v2.pdf | Style-ERD: Responsive and Coherent Online Motion Style Transfer | Motion style transfer is a common method for enriching character animation. Motion style transfer algorithms are often designed for offline settings where motions are processed in segments. However, for online animation applications, such as realtime avatar animation from motion capture, motions need to be processed as a stream with minimal latency. In this work, we realize a flexible, high-quality motion style transfer method for this setting. We propose a novel style transfer model, Style-ERD, to stylize motions in an online manner with an Encoder-Recurrent-Decoder structure, along with a novel discriminator that combines feature attention and temporal attention. Our method stylizes motions into multiple target styles with a unified model. Although our method targets online settings, it outperforms previous offline methods in motion realism and style expressiveness and provides significant gains in runtime efficiency | ['Michiel Van de Panne', 'Zhongquan Chen', 'Xiaohang Zhan', 'Tianxin Tao'] | 2022-03-04 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Tao_Style-ERD_Responsive_and_Coherent_Online_Motion_Style_Transfer_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Tao_Style-ERD_Responsive_and_Coherent_Online_Motion_Style_Transfer_CVPR_2022_paper.pdf | cvpr-2022-1 | ['motion-style-transfer'] | ['computer-code'] | [ 1.70198455e-01 -1.95541710e-01 -3.53329927e-01 -1.17342986e-01
-6.10772908e-01 -7.92388558e-01 5.79371750e-01 -4.04741138e-01
-4.18553114e-01 5.39653599e-01 4.17876631e-01 -2.09307149e-01
5.06559134e-01 -5.69232762e-01 -6.67899311e-01 -3.70582491e-01
3.15101475e-01 3.14890355e-01 3.14424425e-01 -2.74692118e-01
5.80156036e-02 4.77301151e-01 -9.88972545e-01 1.30817324e-01
6.45322561e-01 4.43706900e-01 2.65189409e-01 1.27264452e+00
-2.11620376e-01 9.09218907e-01 -5.48908412e-01 -4.14933175e-01
1.71559468e-01 -8.58993173e-01 -9.59573805e-01 6.03460073e-02
3.89994711e-01 -9.64552462e-01 -7.55141258e-01 7.55295694e-01
6.24756157e-01 4.34736878e-01 5.16851068e-01 -1.20073819e+00
-7.08883226e-01 5.29443622e-01 -6.25145495e-01 3.97540070e-03
5.33895850e-01 4.46707428e-01 9.20918941e-01 -7.08859086e-01
9.09213901e-01 1.48443544e+00 5.56099892e-01 1.17298174e+00
-1.42166734e+00 -5.99934518e-01 3.02309901e-01 7.91168585e-02
-8.41807723e-01 -2.66315490e-01 9.56461191e-01 -1.34940088e-01
5.97073257e-01 2.69088894e-01 1.00232697e+00 1.45731390e+00
3.48778851e-02 1.30690145e+00 5.50199687e-01 1.28561277e-02
1.86968744e-01 -4.63574946e-01 -3.98999214e-01 3.07946146e-01
-4.47002560e-01 -1.91064149e-01 -6.06582701e-01 -1.51560590e-01
1.50827587e+00 -9.09484625e-02 -1.71666041e-01 -2.70295925e-02
-1.48631239e+00 7.73855031e-01 2.30937839e-01 -9.51824561e-02
-2.82537520e-01 6.87459290e-01 7.67665744e-01 4.34015304e-01
4.95132625e-01 3.76309693e-01 -5.83543330e-02 -7.85817087e-01
-9.61355686e-01 5.84886014e-01 8.11698198e-01 1.26888919e+00
4.39600706e-01 3.79105866e-01 -5.81825376e-01 5.63622475e-01
-5.84861189e-02 5.54092705e-01 5.51136136e-01 -1.34390652e+00
3.79105419e-01 4.53629605e-02 2.76172161e-01 -9.13922310e-01
-5.26699312e-02 1.67396381e-01 -9.22378838e-01 1.45385414e-01
2.47770786e-01 -4.42058831e-01 -6.77393973e-01 2.01525664e+00
5.29425025e-01 6.84322357e-01 -3.20862502e-01 1.15879512e+00
6.03504002e-01 9.44452107e-01 2.91364163e-01 5.96943088e-02
1.13782942e+00 -1.34901226e+00 -8.47400904e-01 1.68060679e-02
5.89126825e-01 -7.88917005e-01 1.62044942e+00 2.05825523e-01
-1.39835894e+00 -5.72596133e-01 -6.55821562e-01 -5.13574064e-01
4.93740976e-01 2.09947620e-02 5.54443002e-01 3.18158507e-01
-9.40275431e-01 1.12490082e+00 -1.06934738e+00 -2.73375809e-01
2.88806558e-01 2.59031624e-01 -9.97040570e-02 4.52373803e-01
-8.91222417e-01 3.96416038e-01 -1.94458384e-02 -1.70682415e-01
-8.43132973e-01 -8.73882115e-01 -8.92714977e-01 5.00422157e-02
2.60361642e-01 -1.17203164e+00 1.65751588e+00 -1.34271836e+00
-2.58544970e+00 4.09213364e-01 -3.50116700e-01 -1.64743915e-01
8.28618646e-01 -7.36768961e-01 -2.85987630e-02 3.57552767e-01
-7.80474767e-02 8.87254953e-01 9.57998395e-01 -1.06199753e+00
-5.57424724e-01 2.40549520e-01 7.42020682e-02 3.99467379e-01
-4.09394830e-01 1.69435218e-01 -7.99261689e-01 -1.29116821e+00
-4.64365363e-01 -1.09635603e+00 -4.06452745e-01 4.01550055e-01
-2.45204896e-01 -1.09043039e-01 1.15466583e+00 -6.51857555e-01
1.27465415e+00 -2.07978749e+00 7.32731223e-01 -2.05780581e-01
3.11424822e-01 4.05081719e-01 -2.95928776e-01 3.62512946e-01
2.09746897e-01 -2.49536969e-02 -2.50876516e-01 -6.85279310e-01
4.98697646e-02 2.74813086e-01 -5.12686551e-01 1.11581445e-01
3.96555871e-01 1.08123934e+00 -1.20466888e+00 -4.73124653e-01
4.59574610e-02 4.32476252e-01 -1.00399196e+00 6.29768908e-01
-3.60458940e-01 8.82828772e-01 -5.63076138e-01 3.32392663e-01
3.42604429e-01 -1.45220920e-01 6.44694269e-02 -1.48153994e-02
-5.43602444e-02 7.69457445e-02 -1.01423419e+00 2.25283623e+00
-7.09155858e-01 8.11180532e-01 1.00687385e-01 -3.22978705e-01
7.21693099e-01 3.96729618e-01 4.91679698e-01 -2.50703514e-01
1.16251349e-01 -7.71233514e-02 -2.88840443e-01 -4.00707543e-01
1.02146161e+00 -4.97338511e-02 -2.19156668e-01 7.77964711e-01
-1.18099591e-02 -2.97479570e-01 -2.75128130e-02 1.96656585e-01
1.12321842e+00 7.88574040e-01 6.32391199e-02 6.54250458e-02
2.50478625e-01 -2.85056140e-02 6.66745365e-01 4.20295447e-01
-1.71909377e-01 8.87501776e-01 5.22776365e-01 -3.48573148e-01
-1.55762231e+00 -9.03351188e-01 5.91532826e-01 1.32939708e+00
2.08170533e-01 -6.07078969e-01 -7.83028364e-01 -4.91297394e-01
-1.83481723e-01 3.67655575e-01 -3.61359835e-01 -2.12507844e-01
-1.25781703e+00 5.68356141e-02 8.05369258e-01 9.44141209e-01
3.63955051e-01 -1.18781519e+00 -7.03178108e-01 5.50077617e-01
-1.49978086e-01 -1.08032131e+00 -1.52899110e+00 -5.12692273e-01
-9.98712420e-01 -4.57595348e-01 -1.32707632e+00 -7.08909035e-01
2.22930491e-01 2.15861440e-01 1.06955171e+00 -9.48897079e-02
-6.05121143e-02 3.58492345e-01 -3.79234970e-01 3.94709371e-02
-4.69235003e-01 2.33577222e-01 -4.99320999e-02 1.65305302e-01
-2.15703234e-01 -7.56814122e-01 -9.43183303e-01 1.65199280e-01
-1.09522343e+00 4.51764375e-01 3.05897236e-01 1.09452045e+00
4.91126597e-01 -9.78149414e-01 5.22913039e-01 -8.66866767e-01
6.77718937e-01 -3.58856648e-01 -2.86735862e-01 -9.73932892e-02
-9.68910679e-02 1.37134969e-01 1.08733666e+00 -1.05891478e+00
-1.22667253e+00 1.51476398e-01 -2.77652264e-01 -1.05791509e+00
-6.80367798e-02 1.36262402e-01 -1.85629264e-01 -6.06535263e-02
4.71825719e-01 1.53040752e-01 5.31198457e-02 -5.87659895e-01
7.63580024e-01 3.02394509e-01 9.49679494e-01 -8.17066371e-01
9.05563831e-01 4.51080233e-01 8.74153897e-02 -6.85578644e-01
-2.12723881e-01 -2.82349199e-01 -4.88931328e-01 -2.46779352e-01
8.14264536e-01 -8.52315843e-01 -9.43449557e-01 6.02534056e-01
-1.32465518e+00 -9.23812330e-01 -4.47682798e-01 4.27977026e-01
-1.00967395e+00 7.48987615e-01 -1.20978701e+00 -5.36999583e-01
-6.06587708e-01 -1.17889655e+00 1.29737055e+00 2.26167396e-01
-7.55933225e-01 -9.46378887e-01 2.42068395e-01 -2.18163848e-01
4.07255411e-01 4.04494643e-01 5.03468156e-01 2.71222722e-02
-4.24985707e-01 2.49860868e-01 -1.49086699e-01 -9.51929986e-02
3.12918991e-01 1.69100478e-01 -6.05010152e-01 -2.56373256e-01
-5.79337656e-01 -4.05453295e-01 6.17968440e-01 1.41520321e-01
1.38799846e+00 -5.60499609e-01 1.84925869e-02 1.03110611e+00
1.02843356e+00 2.21983150e-01 6.04694068e-01 9.95303541e-02
1.22058773e+00 2.35511452e-01 4.48931694e-01 7.31992900e-01
2.83217281e-01 9.69108641e-01 -2.85584647e-02 -9.72353145e-02
-3.57823044e-01 -6.73953891e-01 6.10248566e-01 1.07811379e+00
-4.24088448e-01 -2.10057095e-01 -4.36283052e-01 5.72757900e-01
-2.02780581e+00 -1.11361074e+00 1.15267195e-01 1.86537278e+00
1.15429521e+00 -1.78271607e-01 7.34056711e-01 -1.90427341e-02
5.17705679e-01 2.79490888e-01 -7.35879183e-01 -6.96238637e-01
1.86155602e-01 3.71850878e-01 2.82986760e-01 4.58815247e-01
-9.93438423e-01 1.25711656e+00 6.52130556e+00 9.66841817e-01
-1.31186044e+00 -7.43160537e-03 3.63946706e-01 -3.25138241e-01
-7.23415375e-01 -2.47678235e-02 -4.07886863e-01 5.85043788e-01
7.34781742e-01 -2.73900777e-01 4.79008913e-01 9.18636084e-01
4.43450838e-01 3.94824773e-01 -1.17621863e+00 1.14646459e+00
-2.13688776e-01 -1.52294409e+00 3.61642212e-01 -2.03149512e-01
6.51099801e-01 -5.98100722e-01 1.01797014e-01 1.59783497e-01
3.48386645e-01 -8.68136942e-01 9.64266956e-01 3.97330552e-01
1.26830459e+00 -8.54474425e-01 4.19991538e-02 9.02974457e-02
-1.43724310e+00 2.97580063e-01 -3.91619690e-02 -4.59565707e-02
6.74808681e-01 -1.07556164e-01 -2.14901730e-01 5.19245744e-01
3.61006498e-01 9.67524588e-01 -9.77344364e-02 8.55657578e-01
7.41278892e-03 8.19745243e-01 -5.54328971e-02 -7.02932179e-02
4.04979825e-01 -3.65552753e-01 7.46958315e-01 1.40982997e+00
4.15575475e-01 3.28577727e-01 1.75188005e-01 8.38503778e-01
-6.57260790e-02 2.58944649e-02 -5.82324803e-01 -1.39956981e-01
3.66675168e-01 1.15495956e+00 -5.16001463e-01 -4.85833824e-01
-3.86548430e-01 1.63345945e+00 1.82830811e-01 3.14509153e-01
-1.07284701e+00 -5.59966922e-01 9.64403927e-01 -8.91489834e-02
2.64845312e-01 -6.58832967e-01 -3.94786954e-01 -1.45425022e+00
-1.98464543e-01 -8.91755462e-01 -2.67886315e-02 -6.76261425e-01
-1.10610235e+00 5.48330069e-01 -5.21742254e-02 -1.56053507e+00
-5.01619756e-01 -3.30238700e-01 -1.00117242e+00 6.52092576e-01
-9.63890493e-01 -1.36440897e+00 -4.18513358e-01 7.33503103e-01
1.03531480e+00 2.91817952e-02 6.14512026e-01 3.02638203e-01
-4.62797970e-01 9.38599646e-01 -1.99459884e-02 1.46646515e-01
1.03277671e+00 -1.37124598e+00 9.18050706e-01 6.24559462e-01
-1.26309365e-01 3.26281279e-01 8.32993090e-01 -6.92350984e-01
-1.63744926e+00 -1.22289610e+00 3.73493075e-01 -1.24262482e-01
6.56059027e-01 -2.14216918e-01 -1.01318014e+00 7.20490754e-01
4.79291469e-01 -1.44809440e-01 6.02686882e-01 -3.77127200e-01
-1.38052836e-01 2.88377434e-01 -7.35954702e-01 1.27815592e+00
1.46581030e+00 -6.01569712e-01 -3.47852528e-01 -3.53290029e-02
9.53986049e-01 -7.59154558e-01 -9.18443918e-01 -7.24716764e-03
6.80583954e-01 -3.95926267e-01 7.84781337e-01 -9.04477000e-01
8.31518173e-01 -2.81781465e-01 2.40947142e-01 -1.46017182e+00
-4.70467120e-01 -1.54510272e+00 -3.94585073e-01 1.22671008e+00
5.08076809e-02 -2.10873023e-01 8.36006284e-01 5.44457018e-01
-2.35844657e-01 -4.70481873e-01 -6.62287176e-01 -7.63498366e-01
1.13201775e-01 -2.58002192e-01 6.90544963e-01 7.99144328e-01
-1.36997014e-01 3.67824465e-01 -9.33547437e-01 -2.82910228e-01
5.04073858e-01 2.18022093e-01 1.17769504e+00 -6.18082285e-01
-7.16851115e-01 -5.92932522e-01 -9.12441239e-02 -1.76153660e+00
3.45064610e-01 -5.10982931e-01 -6.49574772e-02 -1.12544727e+00
1.14427425e-01 -1.74227193e-01 3.55972975e-01 1.32996723e-01
-4.70007598e-01 1.99150875e-01 5.85799038e-01 3.48177373e-01
-4.42699909e-01 1.04388618e+00 1.83561790e+00 -6.96621463e-02
-3.92685026e-01 -9.40809026e-03 -1.91938788e-01 7.89700508e-01
6.11664534e-01 -1.55208111e-01 -4.34858829e-01 -6.31477177e-01
2.98668575e-02 5.27898252e-01 2.17963159e-01 -7.97415614e-01
1.61212370e-01 -4.57585663e-01 2.64045715e-01 -4.83797103e-01
4.71288264e-01 -4.30386722e-01 2.23404676e-01 3.07137460e-01
-4.94959474e-01 4.81005341e-01 1.66314349e-01 7.21067369e-01
-2.07373992e-01 6.19831905e-02 6.48169696e-01 -1.03588672e-02
-6.96446717e-01 7.30314314e-01 -5.39925337e-01 3.69760022e-02
8.60556781e-01 -1.25814110e-01 3.22430022e-02 -8.30407083e-01
-8.29252660e-01 1.35346711e-01 6.73287690e-01 4.14787143e-01
6.38042390e-01 -2.00778532e+00 -6.39664948e-01 -1.60095409e-01
-1.95663214e-01 1.81760162e-01 2.92373508e-01 4.96873885e-01
-8.66401076e-01 -1.49996400e-01 -4.94768739e-01 -4.64526266e-01
-1.19327950e+00 4.89615053e-01 -9.43488255e-02 -1.33921906e-01
-1.07641864e+00 5.49898922e-01 4.69303846e-01 5.01154102e-02
4.88477871e-02 -3.88214797e-01 9.37290676e-03 -2.37164751e-01
7.07967103e-01 5.79517603e-01 -6.37184203e-01 -4.80428666e-01
1.51859716e-01 7.19607294e-01 1.77113369e-01 -5.00435412e-01
1.06016541e+00 -2.65502572e-01 1.90090716e-01 6.51857197e-01
1.01533782e+00 1.12846315e-01 -1.96083295e+00 -1.27210081e-01
-2.28319764e-01 -6.44074321e-01 -4.56083596e-01 1.05530053e-01
-1.04440498e+00 8.28903556e-01 1.49485826e-01 -2.03817755e-01
1.16301155e+00 -3.13232899e-01 1.61021543e+00 1.53326839e-01
1.23746201e-01 -1.01210415e+00 5.64204097e-01 6.84229374e-01
9.59623039e-01 -9.46795940e-01 -5.30701518e-01 -2.89883643e-01
-1.03365242e+00 1.23683512e+00 7.45030165e-01 -5.01177430e-01
3.47862154e-01 3.70302945e-01 -4.55502830e-02 4.48168725e-01
-8.20404172e-01 5.10625765e-02 2.00564429e-01 6.05231285e-01
4.53503370e-01 -1.50185842e-02 -3.60109001e-01 6.69459701e-01
-1.67328060e-01 1.86431065e-01 5.68971694e-01 7.98569381e-01
-5.44858985e-02 -1.27889132e+00 -2.31782272e-01 -7.54392073e-02
-3.90802652e-01 4.21965383e-02 -2.84772098e-01 5.91733694e-01
-4.92262274e-01 5.21456957e-01 2.43846714e-01 -5.63432813e-01
2.62394667e-01 -1.05837457e-01 5.85742414e-01 -5.03660440e-01
-8.61065328e-01 3.41092318e-01 7.74125010e-02 -7.99022615e-01
-2.14122429e-01 -5.05086958e-01 -1.14186656e+00 -6.97481930e-01
1.95693180e-01 -1.18911475e-01 6.15849830e-02 6.80654585e-01
5.36600173e-01 6.80485606e-01 9.09660637e-01 -1.45691729e+00
-4.20963734e-01 -8.37328434e-01 -3.08366686e-01 7.43132353e-01
3.68131518e-01 -3.63854676e-01 1.79272994e-01 4.74936873e-01] | [10.811979293823242, -0.6785629987716675] |
c3775171-4f0f-4208-ab90-b6b24082e906 | federated-neural-compression-under | 2305.16416 | null | https://arxiv.org/abs/2305.16416v1 | https://arxiv.org/pdf/2305.16416v1.pdf | Federated Neural Compression Under Heterogeneous Data | We discuss a federated learned compression problem, where the goal is to learn a compressor from real-world data which is scattered across clients and may be statistically heterogeneous, yet share a common underlying representation. We propose a distributed source model that encompasses both characteristics, and naturally suggests a compressor architecture that uses analysis and synthesis transforms shared by clients. Inspired by personalized federated learning methods, we employ an entropy model that is personalized to each client. This allows for a global latent space to be learned across clients, and personalized entropy models that adapt to the clients' latent distributions. We show empirically that this strategy outperforms solely local methods, which indicates that learned compression also benefits from a shared global representation in statistically heterogeneous federated settings. | ['Shirin Saeedi Bidokhti', 'Hamed Hassani', 'Eric Lei'] | 2023-05-25 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [-3.74175131e-01 2.68108044e-02 -6.11778378e-01 -5.01650155e-01
-1.24494016e+00 -4.98377472e-01 6.16037250e-01 -8.68010148e-03
2.74679303e-01 4.59082782e-01 1.09188116e+00 3.19267064e-01
-3.16547930e-01 -9.42886174e-01 -9.14275944e-01 -7.46532202e-01
-4.24292535e-01 1.03068185e+00 -2.76986927e-01 2.44254041e-02
2.69247424e-02 3.23642641e-01 -1.48444319e+00 7.96884358e-01
3.11827719e-01 9.62942719e-01 2.51440197e-01 7.01981425e-01
-2.53843337e-01 1.26777864e+00 -3.03042471e-01 -5.48222005e-01
6.24141276e-01 -4.41781610e-01 -9.22811270e-01 -2.45423615e-02
1.87703893e-01 -7.20959425e-01 -8.26382339e-01 7.23397017e-01
2.28126392e-01 2.50428855e-01 6.42145336e-01 -1.08438873e+00
-8.18673074e-01 1.35041273e+00 8.40891749e-02 -8.91268998e-02
8.33579227e-02 3.54615241e-01 1.32639921e+00 -3.17114949e-01
9.21672165e-01 9.72933114e-01 5.77388346e-01 2.67225981e-01
-1.42248988e+00 -2.99034059e-01 -4.73444201e-02 3.92208636e-01
-1.17117584e+00 -6.26056790e-01 6.27903640e-01 8.93934630e-03
1.02493775e+00 3.94107997e-01 5.40393054e-01 1.20674384e+00
3.06108654e-01 9.28285778e-01 6.14886761e-01 8.74472596e-03
8.18206370e-01 3.38133574e-01 -4.57923204e-01 3.00048321e-01
7.44529143e-02 -1.49780745e-02 -1.02999413e+00 -7.88980365e-01
1.39600068e-01 6.86219931e-01 -3.38888079e-01 -7.90122390e-01
-8.54641378e-01 1.09238565e+00 1.42491996e-01 1.87811896e-01
-7.20116973e-01 3.19101900e-01 7.69568503e-01 7.05773771e-01
4.24556553e-01 2.32117862e-01 -7.12730229e-01 -4.23865527e-01
-1.19126379e+00 4.15116876e-01 1.56079161e+00 1.42022645e+00
1.23737347e+00 -1.69713140e-01 -3.38586926e-01 6.81230545e-01
3.51927169e-02 6.39870584e-01 9.50589359e-01 -1.57892370e+00
5.64273536e-01 2.58114934e-01 -2.77007967e-01 -6.86342359e-01
4.68649417e-01 -3.20942640e-01 -7.69960344e-01 -1.27205819e-01
-2.83232540e-01 9.04325992e-02 -2.07396939e-01 1.67041373e+00
9.41886604e-02 2.49050140e-01 1.64300263e-01 6.13901794e-01
-9.75874066e-02 5.26438653e-01 -2.68644989e-01 -1.69246808e-01
6.40949667e-01 -1.06138897e+00 -3.74290466e-01 3.41874331e-01
5.10645449e-01 -2.30126575e-01 9.32579219e-01 4.53688025e-01
-1.29212666e+00 2.02621818e-01 -6.91449523e-01 -1.76781580e-01
-6.10352345e-02 -7.99187243e-01 5.35403013e-01 3.81660759e-01
-1.24797499e+00 1.10905015e+00 -1.00322592e+00 -2.59502262e-01
5.74943900e-01 1.32434711e-01 -4.66250718e-01 -4.16452438e-01
-5.87245822e-01 4.14930135e-01 2.31629714e-01 -1.07404935e+00
-1.44430482e+00 -8.27570617e-01 -4.25511569e-01 8.92540455e-01
1.84716046e-01 -1.08152223e+00 1.52160764e+00 -7.86917746e-01
-1.60744870e+00 1.36097506e-01 -1.09210759e-01 -6.29394770e-01
3.22121590e-01 4.25221659e-02 -3.02499533e-02 4.16592389e-01
-2.40600839e-01 -2.80576706e-01 9.04516935e-01 -1.14965212e+00
-6.32807434e-01 -6.03631496e-01 -5.28680384e-01 -8.97829048e-03
-7.63003707e-01 -1.78771421e-01 -3.77536327e-01 -5.96946120e-01
-1.28900886e-01 -6.38763607e-01 -1.60472125e-01 3.37601043e-02
-6.56366944e-02 -1.51661783e-01 9.85226989e-01 -6.00665152e-01
1.07969940e+00 -1.96183515e+00 3.21774095e-01 5.03872752e-01
5.04067063e-01 -5.57010889e-01 -4.20536488e-01 9.30038750e-01
3.63327712e-01 2.01697312e-02 -1.22580260e-01 -9.65130866e-01
4.25062686e-01 6.52468324e-01 -7.28040338e-01 5.54066479e-01
-5.52720666e-01 8.34975421e-01 -7.74242759e-01 -4.26501125e-01
-2.57542133e-01 4.83109564e-01 -1.16726351e+00 8.15161347e-01
-7.55951107e-01 3.76999527e-02 -5.24065614e-01 6.94519758e-01
7.46026456e-01 -5.20429432e-01 5.93700409e-01 1.19590268e-01
3.25487256e-01 4.42702323e-01 -1.12182140e+00 2.09238935e+00
-6.10355914e-01 2.28094816e-01 4.61226285e-01 -1.06992865e+00
7.46395826e-01 5.17817795e-01 1.00910687e+00 -3.29646945e-01
-1.05396852e-01 2.58363903e-01 -7.84771085e-01 -5.17272770e-01
4.79368240e-01 -1.55858099e-01 3.55512537e-02 1.49125171e+00
4.75379586e-01 3.55542451e-02 -3.73718470e-01 6.52267873e-01
1.65912843e+00 -3.84263813e-01 7.38449991e-02 -1.08433448e-01
-1.61004618e-01 -3.71606857e-01 5.92447817e-01 9.54263926e-01
5.82832359e-02 6.86588585e-01 3.72680038e-01 -5.52865982e-01
-1.47166491e+00 -1.05094147e+00 1.83587074e-01 1.45103419e+00
-1.43845066e-01 -1.15039015e+00 -4.95785654e-01 -6.05272353e-01
3.72311473e-01 5.42415679e-01 -5.57582200e-01 -1.21550135e-01
-4.81754899e-01 -7.14946911e-02 2.73069084e-01 4.80438232e-01
4.02915142e-02 -8.08804154e-01 -5.23513317e-01 2.78831184e-01
-3.44011694e-01 -6.53757393e-01 -6.90788031e-01 4.21216577e-01
-1.12862158e+00 -6.94508433e-01 -3.71532053e-01 -2.18536302e-01
3.89818065e-02 3.19342524e-01 1.27484322e+00 -1.66775197e-01
-1.16427779e-01 9.53057647e-01 -4.37336832e-01 3.17311063e-02
-5.99699020e-01 4.48483020e-01 3.48209143e-02 9.15902555e-02
3.02658737e-01 -1.31336319e+00 -6.91511273e-01 -6.56757131e-02
-1.31949329e+00 -3.26000035e-01 3.70104253e-01 7.32060075e-01
6.90759957e-01 1.99414082e-02 1.62214026e-01 -9.51858222e-01
6.30087018e-01 -1.29651117e+00 -1.70890875e-02 5.42817175e-01
-9.81750906e-01 6.00725234e-01 1.15671635e+00 -2.95829475e-01
-1.05898452e+00 -2.97364175e-01 2.98733324e-01 -1.05925369e+00
-7.70044401e-02 2.85170645e-01 -2.46300966e-01 3.13134998e-01
7.71985710e-01 4.70611572e-01 3.67722303e-01 -1.02790451e+00
7.97914028e-01 7.54895866e-01 5.34480274e-01 -1.14657795e+00
6.55848026e-01 6.46532774e-01 -2.97491848e-01 -2.09763020e-01
-1.49824545e-01 -2.87503511e-01 -1.71142817e-01 4.40579057e-01
1.33571267e-01 -9.51219618e-01 -7.02903986e-01 5.24089895e-02
-7.93871760e-01 -6.26721025e-01 -1.26411390e+00 4.03234273e-01
-1.24504864e+00 3.44516098e-01 -9.14249539e-01 -4.32385802e-01
-7.77443111e-01 -8.85576606e-01 8.46122086e-01 -1.87112585e-01
-1.79556191e-01 -1.02220035e+00 6.53073490e-01 -1.63399857e-02
1.20659792e+00 -2.03693181e-01 1.01386917e+00 -1.08061492e+00
-1.10590160e+00 -8.97519290e-02 2.25438714e-01 3.04824650e-01
1.20034188e-01 -2.81890213e-01 -7.47909963e-01 -6.42279565e-01
5.54142356e-01 -6.52465582e-01 7.63280928e-01 -2.48515960e-02
1.69334400e+00 -1.29399872e+00 -1.80296779e-01 1.38750589e+00
1.55838823e+00 -4.68236923e-01 3.15833896e-01 1.67095676e-01
2.97941715e-01 4.76345718e-01 -2.85471171e-01 1.43260992e+00
6.00952625e-01 4.89072412e-01 5.95652044e-01 6.63962722e-01
2.34275423e-02 -7.23847508e-01 4.71450865e-01 1.03279841e+00
1.50109559e-01 -7.33269602e-02 -3.62756312e-01 5.78138888e-01
-1.95489562e+00 -1.42823064e+00 7.67778277e-01 2.06208801e+00
1.03045678e+00 -7.49688447e-01 2.66887277e-01 -2.67857313e-01
2.98587620e-01 2.41987735e-01 -1.03544116e+00 -4.58366394e-01
-2.55040020e-01 5.81549644e-01 7.10671425e-01 1.34737805e-01
-4.00449812e-01 3.71060997e-01 7.15550518e+00 7.42860258e-01
-9.77330625e-01 4.83021289e-01 6.92276418e-01 -7.03364015e-01
-1.10415041e+00 1.84490591e-01 -3.07222039e-01 6.45309031e-01
1.49782979e+00 -9.19635594e-01 1.36789620e+00 1.43276942e+00
-3.00031245e-01 6.01697862e-01 -1.54280293e+00 9.66783166e-01
1.43227100e-01 -1.81369662e+00 1.93992257e-01 4.17019427e-01
1.00054395e+00 7.84145772e-01 2.87224919e-01 3.06591570e-01
9.14014935e-01 -9.18354571e-01 5.04053533e-01 6.94358170e-01
7.51492858e-01 -6.50794625e-01 2.56516874e-01 6.54630184e-01
-1.05350435e+00 -6.10100389e-01 -5.01772523e-01 5.00048161e-01
4.98913080e-02 2.92198092e-01 -5.78455448e-01 3.26935649e-01
9.97319639e-01 6.77577019e-01 -1.73118338e-01 8.53827357e-01
4.07218218e-01 6.12573087e-01 -4.83560681e-01 4.35424715e-01
-2.19321817e-01 -2.33836010e-01 3.73153418e-01 1.21542549e+00
7.35865772e-01 -2.52114348e-02 1.05653375e-01 8.89059067e-01
-5.31570435e-01 3.09954494e-01 -6.96757734e-01 3.03940941e-02
7.17788041e-01 8.88462186e-01 2.28642270e-01 -6.10573471e-01
-4.11019027e-01 1.13043499e+00 6.58431470e-01 4.63713259e-01
-4.20752376e-01 2.66890787e-03 1.01122546e+00 2.96414256e-01
7.42973924e-01 4.35735751e-03 -1.57734618e-01 -1.73907995e+00
6.63389545e-03 -1.26084459e+00 8.07650745e-01 -1.67672530e-01
-1.85963309e+00 5.57319045e-01 1.02242492e-01 -8.01184952e-01
-7.25679338e-01 6.14523850e-02 -7.98367739e-01 7.99301565e-01
-1.28805518e+00 -1.04296422e+00 -1.03107698e-01 1.39274859e+00
3.85085911e-01 -7.49515533e-01 1.04433036e+00 5.20870388e-02
-1.06207892e-01 1.02472651e+00 8.85074735e-01 -4.22158360e-01
6.20906830e-01 -9.87675071e-01 3.86705855e-03 3.30455542e-01
2.08842665e-01 8.13173592e-01 5.68956673e-01 -4.60133761e-01
-2.12080812e+00 -1.18100858e+00 7.75206506e-01 -2.25841761e-01
5.97108424e-01 -4.32245769e-02 -1.10380042e+00 1.06780648e+00
3.94665360e-01 3.32082778e-01 1.03638029e+00 8.07589144e-02
-1.04856920e+00 -5.20269096e-01 -1.66534960e+00 1.07644148e-01
1.05020845e+00 -9.97708917e-01 -1.33859515e-01 6.22488320e-01
8.41125965e-01 4.37661074e-02 -1.24799991e+00 -1.83938399e-01
4.53606844e-01 -1.09720850e+00 6.87540889e-01 -9.90618348e-01
3.87556851e-01 4.26896036e-01 -9.84721065e-01 -1.25711727e+00
-5.56915939e-01 -1.07450914e+00 -1.17164850e+00 1.15590727e+00
-1.17912062e-01 -6.97654009e-01 9.82812583e-01 1.18948245e+00
9.79778916e-03 -6.61646426e-01 -1.07454431e+00 -9.22991514e-01
2.75344163e-01 -2.68870026e-01 1.41136003e+00 7.02554822e-01
5.00060201e-01 -2.72780150e-01 -4.45177048e-01 -1.85198262e-01
8.49878848e-01 4.63228285e-01 9.79588628e-01 -7.53340840e-01
-1.21299481e+00 -3.93287480e-01 -6.48581013e-02 -1.14655757e+00
3.66334051e-01 -1.25135624e+00 -3.16206068e-01 -9.50970411e-01
8.85175824e-01 -6.24504685e-01 -4.50699151e-01 4.71726596e-01
4.97183055e-01 -4.00082946e-01 2.83356547e-01 9.04150903e-01
-7.19348550e-01 1.04917359e+00 4.83140707e-01 -2.36959234e-01
-1.24020062e-01 2.54448541e-02 -1.09683430e+00 -1.31541014e-01
6.80880725e-01 -5.96224308e-01 -4.68563586e-01 -6.25997841e-01
1.34404935e-02 3.09795618e-01 -8.44619200e-02 -7.05931008e-01
5.82657456e-01 -6.31714940e-01 -7.70049095e-02 -1.35332420e-01
1.55518949e-01 -9.74247932e-01 5.80851734e-01 2.30686530e-01
-7.24561751e-01 6.94792941e-02 -6.64863825e-01 7.63721406e-01
-1.13647744e-01 7.26343170e-02 6.53189003e-01 -4.17047501e-01
1.71927363e-01 1.03532958e+00 5.10823391e-02 8.59730095e-02
8.30434382e-01 2.36833215e-01 -4.18771237e-01 -9.77310300e-01
-5.20982385e-01 1.28212497e-02 9.78111863e-01 4.87041660e-02
4.91770893e-01 -1.34914899e+00 -8.21149528e-01 3.50111365e-01
7.28154778e-02 -4.62579355e-02 2.57410049e-01 4.19774383e-01
-2.27262065e-01 9.53635946e-02 -5.12888804e-02 -3.46593529e-01
-6.67450011e-01 9.94533300e-01 1.14667259e-01 -3.51076573e-01
-1.11756635e+00 5.38313448e-01 -3.43345582e-01 -2.75845468e-01
5.39465249e-01 2.25775735e-03 7.43202865e-01 -3.98449957e-01
7.39556730e-01 6.41927779e-01 9.11008343e-02 -2.66937196e-01
1.59998298e-01 -7.68304393e-02 -1.30305395e-01 -2.64873683e-01
2.02648044e+00 -2.81093210e-01 -3.32280308e-01 3.03739816e-01
1.71874475e+00 2.76437262e-03 -1.35225570e+00 -9.50785518e-01
-3.12646151e-01 -1.00414121e+00 -1.59273401e-01 -4.80053335e-01
-1.52405286e+00 5.17699897e-01 3.03728461e-01 7.96481967e-02
1.08554912e+00 4.10886317e-01 1.21641243e+00 6.04756236e-01
6.46829486e-01 -1.01059878e+00 2.64249682e-01 3.89558017e-01
7.59036183e-01 -8.08244586e-01 -1.31631598e-01 4.39537048e-01
-6.84876680e-01 1.20811450e+00 2.00441927e-01 -1.22895777e-01
8.91010284e-01 5.30445814e-01 -4.80283469e-01 -4.15338464e-02
-1.63868999e+00 4.58824158e-01 -3.85063410e-01 7.24259555e-01
5.47026955e-02 1.64974988e-01 2.62952268e-01 8.00412953e-01
-3.75456482e-01 2.81114969e-02 1.83139175e-01 9.79776621e-01
-3.28363806e-01 -1.53635514e+00 -2.71408886e-01 7.02775717e-01
-3.24355185e-01 1.88452557e-01 6.23859987e-02 -2.19122797e-01
-2.81618927e-02 6.84769094e-01 1.34668246e-01 -5.73163509e-01
-2.10092559e-01 2.52965927e-01 1.93822384e-01 -5.66665828e-01
-7.69223869e-01 -1.40293077e-01 -5.23484707e-01 -1.03628266e+00
1.94906339e-01 -8.00198555e-01 -8.92244518e-01 -1.04772985e+00
2.38657609e-01 4.28645015e-01 7.66496301e-01 4.73221749e-01
9.13242161e-01 -1.22583792e-01 1.31644809e+00 -7.68312335e-01
-1.49348652e+00 -5.91797829e-01 -1.04845107e+00 8.40964913e-01
3.80544752e-01 -2.47417223e-02 -5.85986137e-01 1.89416036e-01] | [5.819883823394775, 6.281915187835693] |
0acb0280-24d6-40f3-8e87-40aa5be1d7b7 | reward-teaching-for-federated-multi-armed | 2305.02441 | null | https://arxiv.org/abs/2305.02441v1 | https://arxiv.org/pdf/2305.02441v1.pdf | Reward Teaching for Federated Multi-armed Bandits | Most of the existing federated multi-armed bandits (FMAB) designs are based on the presumption that clients will implement the specified design to collaborate with the server. In reality, however, it may not be possible to modify the client's existing protocols. To address this challenge, this work focuses on clients who always maximize their individual cumulative rewards, and introduces a novel idea of "reward teaching", where the server guides the clients towards global optimality through implicit local reward adjustments. Under this framework, the server faces two tightly coupled tasks of bandit learning and target teaching, whose combination is non-trivial and challenging. A phased approach, called Teaching-After-Learning (TAL), is first designed to encourage and discourage clients' explorations separately. General performance analyses of TAL are established when the clients' strategies satisfy certain mild requirements. With novel technical approaches developed to analyze the warm-start behaviors of bandit algorithms, particularized guarantees of TAL with clients running UCB or epsilon-greedy strategies are then obtained. These results demonstrate that TAL achieves logarithmic regrets while only incurring logarithmic adjustment costs, which is order-optimal w.r.t. a natural lower bound. As a further extension, the Teaching-While-Learning (TWL) algorithm is developed with the idea of successive arm elimination to break the non-adaptive phase separation in TAL. Rigorous analyses demonstrate that when facing clients with UCB1, TWL outperforms TAL in terms of the dependencies on sub-optimality gaps thanks to its adaptive design. Experimental results demonstrate the effectiveness and generality of the proposed algorithms. | ['Jing Yang', 'Cong Shen', 'Wei Xiong', 'Chengshuai Shi'] | 2023-05-03 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [-1.80920675e-01 2.21571088e-01 -7.74985373e-01 -2.69328862e-01
-9.16542768e-01 -7.78996885e-01 2.22181261e-01 -2.04060271e-01
-2.47937083e-01 9.29744422e-01 -8.70900378e-02 -7.52745688e-01
-7.78806806e-01 -6.66097999e-01 -9.17178869e-01 -1.29433548e+00
-8.44629630e-02 7.80344307e-01 -9.49712172e-02 1.61177032e-02
1.92580000e-01 5.97588718e-01 -1.22734320e+00 1.51518434e-01
7.84981132e-01 1.19189036e+00 3.63710038e-02 5.71116984e-01
-4.56353098e-01 8.61783564e-01 -1.74947575e-01 -4.68800932e-01
5.68720877e-01 -2.95960754e-01 -8.11662614e-01 3.40258986e-01
-1.78284168e-01 -7.22931683e-01 -1.59848779e-02 8.47996175e-01
2.54280776e-01 5.25071733e-02 2.97259063e-01 -1.48733449e+00
-1.72874421e-01 1.02182829e+00 -8.79739344e-01 1.39721423e-01
-5.97498566e-02 -8.24921802e-02 1.33020616e+00 -2.02026501e-01
1.41910657e-01 9.30687249e-01 2.68396229e-01 5.38262606e-01
-1.37526095e+00 -6.20695353e-01 6.33956432e-01 1.44328326e-01
-7.20993638e-01 -2.86602497e-01 7.90974855e-01 3.43693309e-02
3.74013960e-01 8.56562853e-01 5.45600593e-01 7.17249036e-01
-5.70174754e-01 1.21174943e+00 1.20413923e+00 -5.66139698e-01
5.29605627e-01 5.75105250e-01 3.72211337e-01 2.72443354e-01
4.02862549e-01 1.34701937e-01 -4.29248184e-01 -4.76088256e-01
5.77734292e-01 1.65910974e-01 -4.04879481e-01 -9.11033213e-01
-5.61105609e-01 9.15281534e-01 1.82066321e-01 9.76332128e-02
-6.82350397e-01 2.13154361e-01 7.89253712e-02 5.67504525e-01
4.45697486e-01 -2.39188480e-03 -3.93648297e-01 -2.04414725e-01
-7.77612567e-01 1.51843354e-01 9.48577762e-01 1.15811145e+00
7.29623735e-01 -1.60590827e-01 -1.91396266e-01 6.09022558e-01
8.96571353e-02 2.60300189e-01 7.49064907e-02 -8.50072682e-01
6.58379734e-01 2.17184871e-01 6.61908686e-01 -3.53807092e-01
-2.21156478e-01 -9.28902864e-01 -3.97584140e-01 2.51663476e-01
6.03826880e-01 -2.22794965e-01 -4.62276518e-01 1.76261759e+00
7.30390489e-01 -8.51765573e-02 5.10484427e-02 9.57441568e-01
-4.97142598e-02 6.22541666e-01 -2.73251474e-01 -8.21717620e-01
9.50412393e-01 -1.02105093e+00 -5.50085008e-01 6.80861250e-02
7.72018075e-01 -4.97900397e-01 9.32929695e-01 4.85424101e-01
-1.41803312e+00 4.05413717e-01 -5.26262045e-01 7.71317601e-01
1.42679274e-01 -3.61003548e-01 8.88680518e-01 1.20128679e+00
-7.63742745e-01 3.42191398e-01 -6.98125362e-01 -1.87257066e-01
4.29734647e-01 5.64134598e-01 1.86128139e-01 -4.69623283e-02
-4.35440212e-01 3.44080299e-01 1.53924916e-02 2.62248576e-01
-9.28466022e-01 -8.40704620e-01 -6.47398156e-06 4.88274664e-01
8.02459180e-01 -5.21594226e-01 1.80060828e+00 -1.40908241e+00
-1.70778191e+00 5.29366732e-01 -1.85585488e-02 -4.96605963e-01
1.04880452e+00 -2.40143478e-01 2.72736937e-01 2.32449859e-01
-1.97475284e-01 -2.06052765e-01 5.39747596e-01 -1.45615923e+00
-1.14777267e+00 -4.01705503e-01 5.70395112e-01 2.85754263e-01
-7.41758764e-01 8.00797567e-02 -1.65522277e-01 -2.37686276e-01
9.14901271e-02 -8.49282980e-01 -4.46978301e-01 -2.54211605e-01
-2.40512416e-01 -2.67058760e-01 6.11549616e-01 1.29616156e-01
8.38298619e-01 -1.93641150e+00 -2.12809920e-01 6.64422870e-01
4.80196811e-02 -9.98219289e-03 -2.26963311e-01 6.29958093e-01
6.59101754e-02 1.39038041e-02 3.80473107e-01 -4.05059487e-01
2.64025897e-01 3.08776528e-01 -6.37742579e-01 7.74811983e-01
-4.35072899e-01 3.62676293e-01 -6.83695793e-01 -2.76700467e-01
-4.49504554e-02 -2.27488667e-01 -1.01851070e+00 5.22895932e-01
-4.53501970e-01 3.08358848e-01 -1.01926863e+00 6.08982861e-01
7.65042365e-01 -4.77473050e-01 6.26264870e-01 2.91314036e-01
-3.94671112e-01 2.91339099e-01 -1.24140286e+00 8.67570341e-01
-5.43329179e-01 -3.48580703e-02 7.48434961e-01 -1.61417425e+00
4.57255870e-01 4.95613813e-01 7.58075893e-01 -4.24822986e-01
1.54418588e-01 4.26241070e-01 -2.39677832e-01 -3.12411934e-01
1.69334058e-02 -1.86817214e-01 3.85810912e-01 8.63329709e-01
-4.18429047e-01 4.60168839e-01 7.11028054e-02 2.31117323e-01
9.72127378e-01 -1.03936605e-01 -2.61887703e-02 -2.47421622e-01
3.16359729e-01 8.87758359e-02 6.64768338e-01 1.15494370e+00
-1.31247669e-01 -4.64211404e-02 8.44849586e-01 -2.68654197e-01
-6.49358273e-01 -9.12027955e-01 1.47586718e-01 1.57751846e+00
1.82172760e-01 1.74249694e-01 -5.08089066e-01 -7.28900433e-01
1.01491325e-01 8.03522944e-01 -5.71002960e-01 2.78953344e-01
-3.49252939e-01 -7.65581310e-01 -1.44205719e-01 1.27652094e-01
3.70839566e-01 -5.87008595e-01 -6.88200593e-01 3.92285883e-01
-6.24015331e-02 -7.18934417e-01 -4.65663582e-01 4.51499581e-01
-9.53026235e-01 -9.81881201e-01 -8.30211461e-01 -4.26764101e-01
7.26677001e-01 7.83029795e-01 6.71990514e-01 -4.05151658e-02
1.86391696e-01 7.81300902e-01 -4.81592327e-01 -2.33558446e-01
-4.30671237e-02 1.68497592e-01 -1.26203939e-01 4.46909279e-01
1.68933362e-01 -1.00798082e+00 -8.62256229e-01 4.69541490e-01
-8.29658926e-01 1.20723620e-01 7.93437123e-01 8.88340056e-01
3.46577406e-01 -1.58709958e-01 7.72351801e-01 -8.67111802e-01
4.15506274e-01 -5.73791206e-01 -1.05215490e+00 4.54329461e-01
-9.02287304e-01 1.30567342e-01 6.90180838e-01 -6.98578358e-01
-1.20975792e+00 -9.92735401e-02 3.63037080e-01 -3.94516826e-01
1.98405132e-01 3.35644096e-01 -6.27905875e-02 -1.07542947e-01
1.38936415e-01 3.85910213e-01 -8.04180875e-02 -6.05663836e-01
3.62654328e-01 7.85418510e-01 1.11121852e-02 -1.16968191e+00
6.80063546e-01 5.36204159e-01 -3.09665799e-02 -4.30521429e-01
-1.01907277e+00 -4.77459729e-01 2.04403788e-01 -5.07142305e-01
-1.77592561e-02 -5.03482997e-01 -1.53798795e+00 -1.32091613e-02
-8.53293657e-01 -3.44955385e-01 -1.80669174e-01 6.37734950e-01
-9.23734248e-01 1.87941328e-01 -4.51591432e-01 -1.60551989e+00
-3.80810231e-01 -8.49138021e-01 2.14121521e-01 3.88540268e-01
2.23037675e-01 -6.14592016e-01 -2.12286375e-02 6.59768641e-01
5.00157535e-01 -2.31429815e-01 8.26723635e-01 -9.00115013e-01
-1.06220174e+00 -3.66530605e-02 -1.85480103e-01 3.93463410e-02
-3.15204486e-02 -3.34155649e-01 -7.34092355e-01 -5.58791339e-01
2.93997340e-02 -4.24711317e-01 3.80526394e-01 5.56583822e-01
1.25828087e+00 -9.69476759e-01 -3.43813986e-01 5.47822714e-01
1.44485319e+00 3.64881366e-01 1.67653903e-01 7.78342545e-01
-3.32205631e-02 7.89412677e-01 5.23825109e-01 1.07941484e+00
8.28239471e-02 4.91867781e-01 9.60357189e-01 2.52092242e-01
7.10282624e-01 -4.40569334e-02 3.61962318e-01 3.08247685e-01
-2.33589858e-01 -1.37062654e-01 -4.45958734e-01 6.89539313e-01
-2.08362865e+00 -9.86625433e-01 4.46724240e-03 2.48412418e+00
1.11311340e+00 1.70559123e-01 5.26190281e-01 5.91217391e-02
6.42419696e-01 -1.95262402e-01 -7.73866177e-01 -5.80526233e-01
1.16691589e-01 -6.96876459e-03 9.54581022e-01 3.44878256e-01
-6.37869418e-01 4.82962549e-01 5.35196257e+00 8.54361773e-01
-9.70761061e-01 1.81961313e-01 8.05338144e-01 -7.08698452e-01
-4.57708538e-01 2.37841442e-01 -7.59813130e-01 3.61330032e-01
9.43516672e-01 -4.46228206e-01 9.06055510e-01 1.30854678e+00
5.03916860e-01 -2.60283481e-02 -1.04733026e+00 5.90423644e-01
-5.77764213e-01 -1.39982867e+00 -3.74615103e-01 3.31860691e-01
6.83381319e-01 -3.13061208e-01 1.83512345e-01 2.56778419e-01
8.53781283e-01 -4.79456127e-01 7.73269594e-01 1.79073364e-01
3.62682968e-01 -1.17189729e+00 3.79650056e-01 6.73886061e-01
-8.44430327e-01 -7.32442617e-01 -9.05870870e-02 -7.81006068e-02
6.76824301e-02 5.34759760e-01 -5.72111905e-01 6.28811359e-01
6.42511964e-01 -8.81200135e-02 3.40940356e-01 1.18700719e+00
1.58268452e-01 8.61356556e-01 -6.30322456e-01 -3.02446842e-01
6.20716989e-01 -6.02592587e-01 4.71830308e-01 8.17169487e-01
3.07999998e-01 3.58973473e-01 1.69520617e-01 6.39592350e-01
-7.77904224e-03 4.33723569e-01 -5.86919934e-02 -6.75197765e-02
4.80453670e-01 1.22658670e+00 -4.56327200e-01 -2.37493381e-01
-4.49571252e-01 4.71834868e-01 4.25605267e-01 4.87733006e-01
-9.93023276e-01 1.29188681e-02 7.95443773e-01 1.04487598e-01
5.84705234e-01 2.50249565e-01 -3.31405759e-01 -7.17890799e-01
2.35142499e-01 -7.72404373e-01 6.28140509e-01 -3.19904298e-01
-1.23559809e+00 1.89278707e-01 -6.27309456e-02 -9.81691122e-01
-1.28034413e-01 -1.82925269e-01 -6.66462541e-01 5.98997593e-01
-1.78383398e+00 -1.00558984e+00 1.87750086e-01 7.62230515e-01
3.96824956e-01 1.34784803e-01 4.31145549e-01 2.94216007e-01
-8.26439857e-01 9.68655348e-01 8.37763369e-01 -4.43174303e-01
1.92041129e-01 -9.95643616e-01 -7.75155902e-01 5.08902311e-01
-1.37043417e-01 4.43361163e-01 1.03594589e+00 -1.02146104e-01
-1.73851824e+00 -8.53780866e-01 3.77166569e-01 3.63725245e-01
9.84233022e-01 -4.92226407e-02 -4.79746342e-01 7.84322500e-01
1.11743826e-02 -1.26506656e-01 6.20100915e-01 3.86647582e-01
-2.55973130e-01 -6.27330065e-01 -1.11205244e+00 6.62884235e-01
9.76369083e-01 9.43759922e-03 -1.94349110e-01 7.53541708e-01
5.52221417e-01 -7.12848604e-02 -4.13227826e-01 3.69155928e-02
6.16757929e-01 -1.09265602e+00 6.14817381e-01 -1.13277876e+00
-8.79298372e-04 2.02521041e-01 -1.96475163e-01 -1.01420295e+00
-1.46372378e-01 -1.37460601e+00 -2.79332221e-01 1.30521297e+00
2.72191972e-01 -9.76810873e-01 1.41940153e+00 8.10315132e-01
-2.39797477e-02 -8.66809666e-01 -1.33220267e+00 -1.16744840e+00
1.19446173e-01 -2.49038696e-01 6.31493270e-01 6.08522832e-01
2.22814351e-01 -1.59701094e-01 -5.62288702e-01 4.50755149e-01
8.46265018e-01 7.52361894e-01 9.19912875e-01 -8.46939206e-01
-8.26037586e-01 -7.21462607e-01 5.65109789e-01 -1.46596301e+00
-8.22012797e-02 -4.57780957e-01 -1.02949888e-01 -1.21912897e+00
4.61125851e-01 -9.09822345e-01 -5.45427978e-01 3.90703648e-01
2.45672375e-01 -4.25906748e-01 3.06917489e-01 4.13886517e-01
-6.77783489e-01 6.06037796e-01 1.01854312e+00 8.08338672e-02
-4.32842553e-01 8.86533558e-01 -7.23178864e-01 3.43014687e-01
8.82703364e-01 -5.95664442e-01 -4.76238072e-01 -7.74653479e-02
1.19412735e-01 5.72115123e-01 1.20098762e-01 -2.49372751e-01
1.97840527e-01 -8.60766768e-01 -4.36869919e-01 -7.52104402e-01
2.08054438e-01 -1.21539664e+00 1.45572379e-01 6.01335406e-01
-5.13077378e-01 -3.00476164e-01 -3.06324929e-01 7.40566313e-01
3.13525677e-01 -3.27545196e-01 6.88476622e-01 1.09622955e-01
2.43347690e-01 3.07701945e-01 -5.03465652e-01 -2.78799474e-01
1.30641794e+00 -8.14542174e-02 -2.50656009e-01 -8.19836020e-01
-5.76968849e-01 5.88130891e-01 -1.30355641e-01 -3.11558515e-01
4.22469713e-02 -1.09031761e+00 -4.38161612e-01 -2.19296366e-01
-3.04932624e-01 -3.27680439e-01 3.33819479e-01 1.38437080e+00
-4.52152602e-02 6.84542716e-01 1.10279590e-01 -3.54434192e-01
-1.35779738e+00 8.01166415e-01 4.20217484e-01 -5.86828172e-01
-4.31453019e-01 7.13600516e-01 3.05318505e-01 -5.09691099e-03
7.13668704e-01 -4.87724021e-02 2.24115267e-01 2.28667464e-02
3.85984272e-01 3.88408303e-01 -5.60461432e-02 2.54305094e-01
-5.24621569e-02 2.37248354e-02 -4.18930203e-01 -2.38198623e-01
1.70836878e+00 -3.89388770e-01 -1.10786818e-02 1.93663295e-02
6.96821094e-01 1.22031160e-01 -1.28981674e+00 -6.20684147e-01
1.79827929e-01 -6.15407825e-01 2.53986064e-02 -5.73592901e-01
-1.24164569e+00 4.62257326e-01 4.07522678e-01 7.12143958e-01
1.31900561e+00 -2.85131373e-02 6.12073123e-01 4.86615002e-01
5.78310072e-01 -1.06062627e+00 -8.32589045e-02 5.66520402e-03
5.36893189e-01 -8.46200585e-01 -2.49021351e-01 -2.65051961e-01
-2.81366408e-01 1.05689883e+00 5.07959366e-01 1.50583491e-01
3.74410331e-01 7.58301616e-02 -2.96602875e-01 2.49404714e-01
-1.03132343e+00 -2.57772446e-01 -3.84455293e-01 1.83154225e-01
1.53334225e-02 1.21041521e-01 -6.75374568e-01 8.19426656e-01
8.01097602e-02 -5.18044122e-02 5.03766894e-01 1.05533218e+00
-7.07094193e-01 -1.22338080e+00 -7.79053569e-01 4.13629204e-01
-6.47216141e-01 1.49649739e-01 -7.67234564e-02 7.63712466e-01
-5.79690039e-01 1.05157936e+00 -4.58186306e-02 2.25880593e-01
8.41041431e-02 -1.33824840e-01 3.08252841e-01 -1.92217946e-01
-5.07146776e-01 5.17910838e-01 1.58940166e-01 -5.73568285e-01
-4.82222021e-01 -6.19543076e-01 -8.24378967e-01 -7.35343039e-01
-5.74695885e-01 7.65243709e-01 5.35741985e-01 9.40826714e-01
1.18280895e-01 1.49366319e-01 1.40283930e+00 -4.12648141e-01
-1.42856598e+00 -7.38102138e-01 -7.39895701e-01 -1.26017839e-01
4.13160920e-01 -4.78296608e-01 -6.90078199e-01 -3.25721532e-01] | [4.558437824249268, 3.34870982170105] |
a466f475-eb29-439e-a2a1-adc4db1d19ee | oo-dmvmt-a-deep-multi-view-multi-task | 2304.05956 | null | https://arxiv.org/abs/2304.05956v1 | https://arxiv.org/pdf/2304.05956v1.pdf | OO-dMVMT: A Deep Multi-view Multi-task Classification Framework for Real-time 3D Hand Gesture Classification and Segmentation | Continuous mid-air hand gesture recognition based on captured hand pose streams is fundamental for human-computer interaction, particularly in AR / VR. However, many of the methods proposed to recognize heterogeneous hand gestures are tested only on the classification task, and the real-time low-latency gesture segmentation in a continuous stream is not well addressed in the literature. For this task, we propose the On-Off deep Multi-View Multi-Task paradigm (OO-dMVMT). The idea is to exploit multiple time-local views related to hand pose and movement to generate rich gesture descriptions, along with using heterogeneous tasks to achieve high accuracy. OO-dMVMT extends the classical MVMT paradigm, where all of the multiple tasks have to be active at each time, by allowing specific tasks to switch on/off depending on whether they can apply to the input. We show that OO-dMVMT defines the new SotA on continuous/online 3D skeleton-based gesture recognition in terms of gesture classification accuracy, segmentation accuracy, false positives, and decision latency while maintaining real-time operation. | ['Marco Cristani', 'Andrea Giachetti', 'Marco Emporio', 'Andrea Avogaro', 'Federico Girella', 'Federico Cunico'] | 2023-04-12 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.73543262e-01 -6.86218560e-01 -1.51374742e-01 -2.88072228e-01
-9.04557705e-01 -6.24604940e-01 4.47925478e-01 -2.29633689e-01
-4.45973337e-01 5.71452118e-02 -1.19624212e-01 -3.72459441e-02
-2.02857971e-01 -4.69818383e-01 -2.76002496e-01 -8.07581306e-01
1.04654513e-01 9.15100455e-01 6.95938170e-01 -1.29294619e-02
1.44463107e-01 7.96055973e-01 -1.60604870e+00 5.04892826e-01
3.66186053e-01 9.68901634e-01 2.92239130e-01 1.21939778e+00
-2.83071697e-01 2.01981753e-01 -6.64905369e-01 -2.25866288e-02
4.35538799e-01 -3.91187161e-01 -5.47369957e-01 1.03020889e-03
3.88944834e-01 -8.07130694e-01 -2.07646415e-01 5.32085419e-01
9.61629927e-01 3.65820289e-01 5.65596998e-01 -1.28130674e+00
3.65238219e-01 6.35825917e-02 -5.09983957e-01 1.29877359e-01
5.83829045e-01 3.89840245e-01 7.38603413e-01 -7.97677159e-01
6.58775926e-01 1.10539865e+00 1.26751721e-01 7.41034746e-01
-9.82890129e-01 -6.48195446e-01 4.82224792e-01 3.32253724e-01
-1.23516464e+00 -2.07797259e-01 6.14405692e-01 -5.16859353e-01
9.05830920e-01 7.19563246e-01 7.42630363e-01 1.43911660e+00
-9.28139910e-02 1.12160766e+00 9.61122692e-01 -1.96888834e-01
3.50930661e-01 -6.32687092e-01 2.52446830e-01 4.79945123e-01
-1.04034573e-01 -6.93665519e-02 -9.13984179e-01 -4.12995927e-02
1.16256344e+00 4.61469084e-01 -3.44831973e-01 -3.00200194e-01
-1.38618577e+00 3.05088431e-01 -7.36273127e-03 4.64214683e-01
-4.80980575e-01 -1.65522657e-02 4.90696371e-01 3.82572860e-01
2.03862220e-01 -1.28624469e-01 -4.84053552e-01 -4.79665756e-01
-1.17118621e+00 3.60180050e-01 6.90530896e-01 9.61608887e-01
2.60094583e-01 -1.36039272e-01 -3.26385498e-01 3.75874728e-01
2.79330969e-01 6.21176600e-01 5.63633919e-01 -5.15495539e-01
7.01317906e-01 6.59641743e-01 -1.37944475e-01 -5.00629604e-01
-6.00116670e-01 -4.11435813e-02 -7.01283455e-01 5.65055251e-01
5.40062606e-01 3.70130055e-02 -1.19156921e+00 1.31227958e+00
5.27218699e-01 1.19932741e-01 -2.01766983e-01 1.41225576e+00
7.63518393e-01 3.06489736e-01 -3.81029211e-02 -1.74832001e-01
1.27323556e+00 -7.46415377e-01 -6.27595365e-01 -3.28320824e-02
5.15643239e-01 -6.40718460e-01 1.28424370e+00 7.97736943e-01
-8.23869467e-01 -5.36325097e-01 -8.63265693e-01 5.45101501e-02
-8.36687386e-02 1.06395729e-01 4.97471541e-01 8.09750259e-01
-7.02644944e-01 3.55263352e-01 -1.37082422e+00 -2.14123949e-01
1.78703621e-01 6.66466892e-01 -3.61839533e-01 8.75480380e-03
-5.21365762e-01 4.06982839e-01 1.75070763e-01 5.06912649e-01
-8.07535648e-01 -1.86675712e-01 -4.99589592e-01 -2.05276936e-01
7.48440683e-01 -5.41244626e-01 1.19764066e+00 -7.55931199e-01
-1.99499285e+00 4.41149503e-01 -4.42989171e-01 2.32046515e-01
1.00297272e+00 -3.71412247e-01 -6.92130625e-02 3.49512368e-01
-1.70261398e-01 3.35002124e-01 9.78928089e-01 -1.24640357e+00
-7.26411045e-01 -1.02966619e+00 6.86126053e-02 5.09865344e-01
-2.47556105e-01 -4.84262258e-02 -7.81176388e-01 -4.94305223e-01
4.73870486e-01 -9.93472159e-01 -4.79011089e-02 -8.13507661e-02
-3.19087714e-01 -2.70109266e-01 1.15227580e+00 -7.03492582e-01
9.19537187e-01 -1.90553141e+00 6.13606930e-01 3.70601386e-01
2.34031394e-01 3.84207070e-01 -9.69698057e-02 2.08152294e-01
5.08394182e-01 -2.27044016e-01 -4.73116301e-02 -3.88265431e-01
-1.16072036e-01 3.04092616e-01 -7.32273236e-02 4.10666347e-01
-2.21204042e-01 8.22905779e-01 -8.39346826e-01 -6.01703227e-01
6.78856850e-01 3.90039563e-01 -2.54976183e-01 5.30945659e-01
-2.42148012e-01 8.85484278e-01 -7.06536174e-01 8.30075681e-01
4.55990255e-01 -9.70917940e-02 3.54002744e-01 -8.58855993e-02
-2.43088678e-02 -1.33581072e-01 -1.58231866e+00 1.98442006e+00
-5.07584333e-01 4.11273062e-01 2.86208302e-01 -5.93797803e-01
5.23170292e-01 7.06757307e-01 6.68808877e-01 -5.46285629e-01
2.81877548e-01 4.27023321e-01 -3.89685333e-02 -6.92211032e-01
2.77432233e-01 2.50063688e-01 1.73534706e-01 7.45106816e-01
-1.30204797e-01 1.52757853e-01 -6.52596820e-03 -2.21998557e-01
9.93470073e-01 5.75194836e-01 9.27966684e-02 3.86269420e-01
2.22804278e-01 9.44679603e-02 1.07501365e-01 6.90328598e-01
-1.17886372e-01 7.92850971e-01 1.04700573e-01 -2.21628994e-01
-5.60534000e-01 -1.01508105e+00 2.98441887e-01 1.35167432e+00
4.12585914e-01 -1.16537519e-01 -5.82768142e-01 -7.29451597e-01
-3.32404375e-01 1.79803148e-02 -3.39235723e-01 5.03074527e-01
-1.01490021e+00 -4.81570840e-01 4.49658602e-01 8.23374867e-01
5.00186265e-01 -1.10362124e+00 -1.52289939e+00 1.70355678e-01
3.09480783e-02 -1.06044507e+00 -5.81946313e-01 1.90187216e-01
-1.09642565e+00 -1.04838336e+00 -1.19145775e+00 -5.11309862e-01
3.70866597e-01 2.86228687e-01 3.76281112e-01 -1.27317443e-01
-4.37651873e-01 6.02924585e-01 -7.81497061e-01 1.39768822e-02
-4.90255989e-02 3.07827115e-01 -1.69295982e-01 2.12330341e-01
2.09650993e-01 -6.82923555e-01 -5.98954976e-01 4.99507427e-01
-9.22923565e-01 1.40844747e-01 7.98448384e-01 7.14926600e-01
6.16507769e-01 -7.19826102e-01 1.51894614e-01 -5.00148773e-01
2.91259259e-01 5.88366129e-02 -2.14444473e-01 4.18059558e-01
-2.94150591e-01 1.24592423e-01 2.28725225e-01 -8.84241521e-01
-8.92747283e-01 5.24944961e-01 -3.61793697e-01 -9.29365039e-01
-3.95376086e-01 1.36828139e-01 -4.39940780e-01 -1.15349116e-02
2.75179148e-01 4.47167367e-01 -1.00636847e-01 -5.15106797e-01
3.21400940e-01 9.26542699e-01 4.30778652e-01 -4.40049469e-01
3.26256037e-01 5.70740104e-01 7.27482587e-02 -8.63085926e-01
-9.45424810e-02 -8.57941687e-01 -1.19737720e+00 -6.17186129e-01
1.07833695e+00 -4.02918488e-01 -9.86856401e-01 6.99487627e-01
-1.24445474e+00 -6.90427423e-01 1.07533634e-01 6.93667948e-01
-5.69071472e-01 2.77320594e-01 -3.09762836e-01 -1.13688850e+00
-4.67730790e-01 -1.60565078e+00 1.57554483e+00 -4.54678051e-02
-2.42957160e-01 -5.06810725e-01 -1.32701203e-01 1.78009346e-01
8.26561078e-02 3.66539389e-01 6.21425390e-01 -7.88279235e-01
-6.84272885e-01 -2.88732529e-01 -2.25273930e-02 -2.21504211e-01
1.72570735e-01 -3.00619692e-01 -9.85000551e-01 -3.35885048e-01
4.17213095e-03 -1.22156493e-01 5.93244314e-01 3.52825731e-01
1.02544248e+00 1.61610872e-01 -4.11620826e-01 5.18774986e-01
1.15407741e+00 5.48704803e-01 2.34028399e-01 1.03527708e-02
1.04146457e+00 5.68823218e-01 5.96794605e-01 4.25427914e-01
1.90696239e-01 1.17010498e+00 5.09906709e-01 1.24251492e-01
-3.73959867e-03 1.30860060e-01 3.20540339e-01 7.67111659e-01
-6.51714623e-01 -3.63848597e-01 -9.13931787e-01 1.79873526e-01
-1.97010410e+00 -7.80841947e-01 -2.39819452e-01 2.39547682e+00
2.83513904e-01 -1.84508543e-02 5.39205015e-01 3.46087158e-01
7.82955468e-01 2.65204549e-01 -7.44651735e-01 -8.77768248e-02
2.57362068e-01 4.95071679e-01 2.78163224e-01 3.39438587e-01
-9.27542806e-01 7.59221375e-01 5.27759790e+00 6.43182039e-01
-1.48261857e+00 4.48513716e-01 9.17660296e-02 -6.78650856e-01
2.50969213e-02 -4.17987406e-01 -6.42898679e-01 4.66200635e-02
4.86784130e-01 5.07309318e-01 3.24582160e-01 6.53511584e-01
2.10424125e-01 -1.02147110e-01 -1.33926356e+00 1.28036427e+00
2.19646081e-01 -7.21605122e-01 2.72942871e-01 2.60142148e-01
1.47116989e-01 -2.27521718e-01 -2.34965980e-01 1.13202058e-01
-3.79773587e-01 -7.58946300e-01 7.78411925e-01 2.92618722e-01
1.08071196e+00 -4.55940336e-01 4.62666303e-01 6.91631615e-01
-1.74258149e+00 1.37139812e-01 4.21531230e-01 1.70643464e-01
5.16548872e-01 -7.98853412e-02 -7.46783853e-01 7.68815935e-01
5.19187391e-01 2.71912724e-01 -2.26194188e-01 6.22792721e-01
-3.60875577e-02 1.35719821e-01 -5.02117217e-01 -2.72614419e-01
6.54515177e-02 5.29555716e-02 7.34907210e-01 1.16659725e+00
3.25991780e-01 2.80467987e-01 5.62136590e-01 4.15653348e-01
4.63447303e-01 8.48397613e-02 -2.62375593e-01 1.78746864e-01
2.82507807e-01 9.38335478e-01 -1.21375835e+00 -3.54946852e-01
-1.80125400e-01 1.64572644e+00 -2.27733940e-01 2.30470315e-01
-6.22836411e-01 -3.34422588e-01 5.56128979e-01 -6.91239536e-02
2.75001824e-01 -7.49112308e-01 -4.76283133e-01 -1.12766194e+00
4.52991605e-01 -7.54563808e-01 3.77492696e-01 -3.34783256e-01
-8.27011883e-01 6.59532785e-01 8.46565962e-02 -1.48227024e+00
-5.91124535e-01 -7.63668716e-01 -3.66707087e-01 7.69325078e-01
-1.05489182e+00 -1.30214822e+00 -7.11790383e-01 8.41414809e-01
1.19652855e+00 -2.06944905e-02 9.41040039e-01 2.65841872e-01
-5.19951522e-01 6.29241824e-01 -4.22570080e-01 1.19009703e-01
5.23779333e-01 -1.10005414e+00 4.05425668e-01 7.28702247e-01
3.98651689e-01 3.38861108e-01 2.55013734e-01 -8.42068195e-01
-1.90626597e+00 -5.94662607e-01 2.74262190e-01 -4.82332200e-01
3.69667113e-02 -4.28267479e-01 -7.60991454e-01 5.06439030e-01
-3.52607369e-01 8.31788629e-02 4.41590071e-01 -1.23936245e-02
-2.80315936e-01 7.00570121e-02 -9.65133369e-01 5.56486189e-01
1.59601128e+00 -4.95481431e-01 -4.30588394e-01 9.65497196e-02
2.99325258e-01 -9.02383506e-01 -6.46915853e-01 2.40341321e-01
1.13692474e+00 -9.34310138e-01 8.25040519e-01 -3.92923057e-01
1.56877309e-01 -1.44677311e-01 -2.81954646e-01 -8.25599790e-01
2.58889288e-01 -5.79507947e-01 -3.43933940e-01 6.99662983e-01
3.12655680e-02 -4.72395569e-01 1.06504941e+00 4.27476972e-01
-1.37330562e-01 -6.79282308e-01 -1.33731747e+00 -7.37236857e-01
-6.08095467e-01 -8.14128220e-01 6.22012854e-01 5.34083486e-01
-2.40935698e-01 -4.62089851e-02 -2.27851734e-01 3.23674858e-01
4.77578968e-01 5.66482604e-01 1.07754290e+00 -1.28353155e+00
-4.53548670e-01 -5.86553395e-01 -5.44318318e-01 -1.41975462e+00
-2.34706521e-01 -7.18304694e-01 1.95021302e-01 -1.47836185e+00
3.40391807e-02 -2.88266808e-01 4.02974663e-03 3.81265610e-01
-5.93058253e-03 2.35740557e-01 4.58965242e-01 6.96763694e-01
-5.02585411e-01 2.24885404e-01 1.32000959e+00 -3.81440260e-02
-6.84991181e-01 3.43340635e-01 2.83343911e-01 4.87659365e-01
3.86087626e-01 -3.02020669e-01 -3.67527515e-01 -5.38288057e-01
-3.39118034e-01 6.17994428e-01 4.65545654e-01 -1.02297127e+00
3.75874639e-01 -4.81582582e-01 2.01220468e-01 -7.73966789e-01
7.47888207e-01 -8.32028627e-01 2.09860131e-01 5.39550662e-01
-1.86390355e-01 2.70644307e-01 -2.03004211e-01 5.69818735e-01
-1.07759364e-01 4.93847616e-02 3.02926600e-01 -3.05403858e-01
-9.12477612e-01 4.88306791e-01 -2.99513191e-01 -2.78401434e-01
1.16073060e+00 -6.58439934e-01 8.94867629e-02 -1.25010684e-01
-9.94300485e-01 5.69662042e-02 8.42385069e-02 4.78244245e-01
9.14648414e-01 -1.04450238e+00 -5.41280270e-01 4.58407164e-01
1.03886895e-01 3.24877471e-01 4.37462360e-01 8.57377112e-01
-4.90197092e-01 4.53487277e-01 -3.29498053e-01 -1.08990979e+00
-1.65033400e+00 2.87726447e-02 1.62973359e-01 -1.86625630e-01
-9.17626917e-01 7.02082932e-01 -1.63659632e-01 -5.19516356e-02
4.75131840e-01 -4.19529468e-01 -1.50418550e-01 2.08499387e-01
4.97842431e-01 5.49584448e-01 3.59577924e-01 -5.44177592e-01
-4.27580506e-01 8.62693012e-01 2.72389472e-01 -7.87852824e-01
1.10652196e+00 1.70940652e-01 3.49753201e-01 8.27584267e-01
9.55835402e-01 -2.36924827e-01 -1.32793939e+00 3.97294424e-02
-2.67346911e-02 -7.60530710e-01 4.74733301e-02 -7.78097689e-01
-1.07287824e+00 1.15964532e+00 1.01252949e+00 -1.30498648e-01
1.05348802e+00 -9.87870917e-02 1.01560605e+00 3.89099866e-01
8.76586199e-01 -9.15793657e-01 2.32061982e-01 3.68914843e-01
8.37426484e-01 -1.00714993e+00 -2.51378596e-01 -3.81302923e-01
-6.77431464e-01 1.36814833e+00 7.47243345e-01 1.18063726e-01
4.45763558e-01 4.68488604e-01 2.33360767e-01 -1.49654463e-01
-2.18487293e-01 -3.71627212e-01 2.47567594e-01 5.91958880e-01
1.01008110e-01 2.14283496e-01 -2.29485899e-01 6.14016831e-01
1.36594087e-01 8.66590664e-02 -3.14095058e-02 1.28821993e+00
-1.00825086e-01 -1.29314959e+00 -4.65661883e-01 6.40315652e-01
-7.34703802e-03 4.26159173e-01 -3.80372584e-01 7.08615422e-01
-3.57627496e-02 7.41484463e-01 9.64934602e-02 -8.42856109e-01
5.19814909e-01 4.03730363e-01 8.12244117e-01 -7.50019491e-01
-9.43543911e-01 3.68235797e-01 -3.02967459e-01 -7.14162171e-01
-5.02220750e-01 -8.30207527e-01 -1.36727393e+00 -2.84178113e-03
-2.81956881e-01 -3.68267566e-01 8.46955717e-01 1.30904138e+00
1.50650308e-01 5.33636391e-01 3.50083888e-01 -1.67408037e+00
-7.24386454e-01 -9.87215459e-01 -5.11838734e-01 3.60491723e-01
4.45231259e-01 -7.61597574e-01 -1.40738234e-01 -1.61241919e-01] | [6.653153419494629, -0.27552396059036255] |
d144db84-8990-4fbd-9732-66e4c43bde99 | the-proof-is-in-the-pudding-using-automated | 2203.02683 | null | https://arxiv.org/abs/2203.02683v1 | https://arxiv.org/pdf/2203.02683v1.pdf | The Proof is in the Pudding: Using Automated Theorem Proving to Generate Cooking Recipes | This paper presents FASTFOOD, a rule-based Natural Language Generation Program for cooking recipes. Recipes are generated by using an Automated Theorem Proving procedure to select the ingredients and instructions, with ingredients corresponding to axioms and instructions to implications. FASTFOOD also contains a temporal optimization module which can rearrange the recipe to make it more time-efficient for the user, e.g. the recipe specifies to chop the vegetables while the rice is boiling. The system is described in detail, using a framework which divides Natural Language Generation into 4 phases: content production, content selection, content organisation and content realisation. A comparison is then made with similar existing systems and techniques. | ['Carl Vogel', 'Louis Mahon'] | 2022-03-05 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [ 2.97200650e-01 5.99268258e-01 -2.70557106e-01 -1.26606703e-01
3.92902046e-02 -1.04182422e+00 6.98853374e-01 5.49018145e-01
1.57685101e-01 7.02169597e-01 5.04484832e-01 -6.26057923e-01
9.62940007e-02 -1.12503159e+00 -4.00366187e-01 -3.28675151e-01
-1.05582930e-01 4.14357603e-01 3.69176447e-01 -4.70835775e-01
2.32397586e-01 9.20305625e-02 -1.73352110e+00 6.59961700e-01
7.76344359e-01 5.37369251e-01 2.34560758e-01 1.04251826e+00
-5.50191641e-01 1.34014809e+00 -2.29464754e-01 -1.42604426e-01
3.44258338e-01 -1.04978216e+00 -9.90801513e-01 2.47385234e-01
-2.46971682e-01 -4.58891660e-01 4.15627152e-01 1.15982556e+00
-3.07253212e-01 1.77727073e-01 4.61469740e-01 -1.69429111e+00
-3.17689449e-01 1.43783522e+00 2.34880656e-01 -5.62502921e-01
1.13872755e+00 3.89403522e-01 9.05350983e-01 -3.05390865e-01
6.95842326e-01 1.27037728e+00 2.36135557e-01 6.88412607e-01
-1.34832203e+00 -2.16224521e-01 1.14989705e-01 -7.74620473e-02
-1.18091249e+00 -4.79563236e-01 6.43868744e-01 -3.34630281e-01
1.30831289e+00 7.18298793e-01 1.08082390e+00 2.72666752e-01
5.80646336e-01 7.25451827e-01 9.22727466e-01 -9.07291949e-01
6.16974473e-01 3.21620762e-01 -6.17359648e-04 7.80021369e-01
4.11593467e-01 2.51087338e-01 -1.51507407e-01 -1.24953859e-01
7.34205902e-01 -5.17086029e-01 7.61155486e-02 -4.90800053e-01
-1.48953378e+00 9.77103710e-01 -2.76869237e-01 2.62376428e-01
-6.72899902e-01 2.60092854e-01 7.64971197e-01 3.12351644e-01
1.84125341e-02 7.88961649e-01 -6.84753299e-01 1.66459054e-01
-8.14917445e-01 1.08659053e+00 1.45646787e+00 1.20620692e+00
4.16842222e-01 2.33807974e-02 -2.57973313e-01 1.94297254e-01
7.28626549e-01 4.04513955e-01 -1.15711354e-01 -1.27076340e+00
5.08933328e-02 5.57064176e-01 5.86442709e-01 -9.08756018e-01
-3.02557796e-01 6.09656036e-01 -4.43771034e-01 4.19482142e-01
4.36839312e-01 -4.30966467e-01 -4.25127417e-01 1.35008097e+00
5.10106623e-01 -4.47921246e-01 2.96148568e-01 6.92343295e-01
7.85625935e-01 1.15072012e+00 5.69446683e-01 -6.52548611e-01
1.56594002e+00 -7.61082113e-01 -1.00513697e+00 4.53538805e-01
7.40528107e-01 -7.15153754e-01 7.63187230e-01 8.63840163e-01
-1.32706690e+00 -2.30070010e-01 -1.02949584e+00 1.30215466e-01
-7.99045324e-01 -7.05199242e-02 9.82532144e-01 6.78704143e-01
-1.00096893e+00 5.40655375e-01 -3.08727890e-01 -4.38510597e-01
-5.58724940e-01 3.88195217e-01 1.16928279e-01 -1.32338805e-02
-1.26555192e+00 8.50852191e-01 1.45496035e+00 -1.20780654e-01
-4.33706164e-01 -7.95915484e-01 -1.48446906e+00 -5.01029519e-03
6.10646784e-01 -8.50665987e-01 1.81985509e+00 -9.05024052e-01
-2.15601921e+00 6.30765319e-01 3.38791370e-01 -5.52939415e-01
1.83355257e-01 1.18002757e-01 -7.21648395e-01 1.68949127e-01
2.99767494e-01 9.71078694e-01 3.67034346e-01 -1.10500586e+00
-9.11292732e-01 3.26442093e-01 5.63871622e-01 2.05967687e-02
7.64185846e-01 5.18028319e-01 -5.70166633e-02 -4.83825117e-01
-1.60623580e-01 -5.81580758e-01 -5.02757072e-01 -2.93777734e-01
-6.51410937e-01 -5.75453639e-01 2.76251078e-01 -6.23652756e-01
1.49356580e+00 -1.54223347e+00 -1.12430453e-01 4.50304240e-01
-1.94975704e-01 -2.59069622e-01 2.92706322e-02 9.85132158e-01
-3.59980106e-01 1.27253398e-01 9.90516916e-02 6.89178824e-01
7.85249591e-01 1.26921639e-01 -2.06943616e-01 1.15014398e-02
3.68306577e-01 6.04693115e-01 -1.31010985e+00 -7.53989339e-01
6.93687916e-01 -1.21996939e-01 -6.42825961e-01 1.15475617e-01
-1.42868996e+00 -1.82168156e-01 -3.97046894e-01 3.71868998e-01
4.73132282e-01 3.11097622e-01 6.24960482e-01 -1.83178484e-01
-7.26846218e-01 6.12640738e-01 -1.67742348e+00 1.67752612e+00
-1.85255229e-01 -1.27651885e-01 2.16702819e-01 -4.18666869e-01
8.42749119e-01 7.70088136e-01 4.90368634e-01 -3.51789773e-01
2.44858071e-01 -1.14639029e-02 -2.33061880e-01 -7.54622579e-01
8.43687713e-01 -4.54810411e-01 -5.42463064e-01 8.35047841e-01
-3.62165630e-01 -7.28666186e-01 1.31941032e+00 4.96960729e-01
7.03965902e-01 9.91086304e-01 9.11967576e-01 -7.04379141e-01
8.03370774e-01 9.49649811e-01 3.66406024e-01 3.13073277e-01
1.59267098e-01 -2.25760385e-01 5.55033684e-01 -3.46121520e-01
-1.38617730e+00 -8.29013228e-01 5.16175926e-01 1.14471412e+00
-1.63430646e-02 -8.84246945e-01 -8.88120651e-01 -4.86229330e-01
-3.03137988e-01 1.70089912e+00 -3.52174699e-01 1.14527568e-01
-7.06122160e-01 -2.03619465e-01 8.10390636e-02 2.23797575e-01
3.92326087e-01 -1.51392114e+00 -1.17954445e+00 6.41849041e-01
-3.59955102e-01 -5.61755121e-01 -6.57541454e-01 9.16282609e-02
-6.21645749e-01 -1.14251339e+00 1.06837504e-01 -6.95761323e-01
5.32654226e-01 -6.39885366e-02 1.39323652e+00 2.48504728e-01
-1.55055925e-01 4.39430863e-01 -7.31694818e-01 -6.79724514e-01
-1.26996386e+00 -3.70908409e-01 -3.07741016e-01 -5.79926193e-01
2.88081795e-01 -5.70365004e-02 -2.89732486e-01 -1.15186818e-01
-1.13585138e+00 8.28438878e-01 5.44040725e-02 3.95923585e-01
3.87598246e-01 7.58100271e-01 2.90207684e-01 -8.18005979e-01
6.50081873e-01 -4.96782184e-01 -7.48100758e-01 4.62019086e-01
-4.53751951e-01 4.79113847e-01 9.86113727e-01 -2.54961878e-01
-1.08702052e+00 5.36345124e-01 -8.20713211e-03 6.18566275e-01
-5.50125241e-01 3.06246459e-01 -4.70802635e-01 7.01868892e-01
4.85843360e-01 6.40674382e-02 -7.97965229e-02 -6.28563836e-02
1.00916398e+00 -1.04867974e-02 6.21297777e-01 -1.02239966e+00
5.54260910e-01 -2.49383613e-01 -1.11461453e-01 -3.57462466e-01
-1.61021635e-01 -3.36977839e-01 -4.63409603e-01 -5.37424743e-01
9.00202870e-01 -3.93447369e-01 -9.87404704e-01 -4.11816612e-02
-1.13524628e+00 -8.63237083e-01 -6.30080044e-01 2.31830925e-01
-9.42779124e-01 1.18276946e-01 -5.45444250e-01 -1.00698423e+00
-5.19043386e-01 -7.63403296e-01 6.73127234e-01 7.41109475e-02
-9.19282436e-01 -8.42656374e-01 2.58066833e-01 -3.49428564e-01
3.26077431e-03 7.55591393e-01 1.22943795e+00 -3.34128320e-01
-4.39048469e-01 2.07134411e-01 7.67612755e-02 -2.43733540e-01
1.68071970e-01 4.87938553e-01 -1.28222853e-01 2.23426908e-01
-4.18938309e-01 -1.65566280e-01 -1.91182981e-03 3.00361902e-01
2.52516091e-01 -8.30733597e-01 -2.36667097e-01 -1.97632149e-01
1.71632624e+00 6.29162550e-01 6.65088177e-01 3.94043297e-01
-2.89283488e-02 8.49979818e-01 9.97670650e-01 4.58780289e-01
3.75521034e-01 4.93402123e-01 1.44349739e-01 3.28102894e-02
-2.12749485e-02 -4.89090145e-01 6.47816122e-01 2.19008550e-01
1.39699727e-01 -1.49827048e-01 -4.42283422e-01 5.88225663e-01
-1.84908164e+00 -1.46210754e+00 -4.59422916e-01 1.78863597e+00
1.16768420e+00 -1.94290265e-01 6.78474247e-01 5.49055457e-01
3.98234189e-01 -2.13293910e-01 -7.66371116e-02 -1.10231543e+00
3.26950163e-01 2.52869755e-01 4.30840641e-01 7.74787962e-01
-8.71805012e-01 9.10819292e-01 7.34856987e+00 2.90417075e-01
-4.10733640e-01 -3.94261152e-01 1.28055707e-01 2.78311789e-01
-5.81824183e-01 1.67312086e-01 -4.41777647e-01 4.85977605e-02
8.51265788e-01 -4.57726628e-01 8.54527116e-01 5.31921983e-01
7.96689212e-01 -6.48091435e-01 -1.42293870e+00 2.90324718e-01
-1.63675487e-01 -1.41693652e+00 6.90943375e-02 -5.74083626e-01
4.41941917e-01 -9.21902299e-01 -8.47463965e-01 2.48919636e-01
1.16363657e+00 -5.98360538e-01 1.46054685e+00 3.91347110e-01
4.59941506e-01 -9.78372514e-01 2.93248892e-01 5.04864097e-01
-1.51005733e+00 -5.03024869e-02 -8.45868215e-02 -3.92089844e-01
3.29051167e-01 3.77473116e-01 -9.71660435e-01 7.30803370e-01
4.42678899e-01 1.70019761e-01 -1.12296596e-01 8.89358103e-01
-5.26102304e-01 2.36532673e-01 -4.19990212e-01 -5.95837235e-01
1.55286297e-01 -5.22379398e-01 4.48998243e-01 1.55459237e+00
1.61026105e-01 6.75850093e-01 3.95862311e-01 1.27584541e+00
7.34986842e-01 5.49891591e-01 -5.22036791e-01 -9.08350796e-02
7.58428052e-02 1.08965302e+00 -1.23557520e+00 -7.58388281e-01
-1.15153700e-01 7.92155325e-01 -6.05725348e-01 2.57646263e-01
-8.35536063e-01 -5.08700609e-01 2.66991884e-01 1.02650516e-01
2.44745016e-01 -9.82862040e-02 -7.47706294e-02 -8.48597944e-01
-5.25546730e-01 -1.27105486e+00 4.56167668e-01 -9.99629438e-01
-6.96067691e-01 1.23560116e-01 8.40676963e-01 -5.44721603e-01
-7.72516847e-01 -5.44881463e-01 -6.96488738e-01 5.50865233e-01
-6.98004007e-01 -9.18211102e-01 3.38203609e-01 4.74924803e-01
6.89388335e-01 3.16417843e-01 1.15130103e+00 -1.72197640e-01
-1.68851972e-01 -1.90562949e-01 -6.27789438e-01 -1.60021588e-01
4.48226146e-02 -1.49845159e+00 2.03085542e-01 1.00023246e+00
-2.80149758e-01 6.76239491e-01 1.31338263e+00 -7.96538293e-01
-1.64851677e+00 -9.65967774e-01 1.26749420e+00 9.92745757e-02
5.70410788e-01 -3.80704015e-01 -2.82881200e-01 4.39180881e-01
8.03836644e-01 -1.07042515e+00 4.40836787e-01 -3.69719833e-01
-1.71809405e-01 7.30565339e-02 -1.46596265e+00 9.07768846e-01
5.14352620e-01 2.28399374e-02 -7.85906136e-01 4.49151158e-01
8.97962213e-01 -2.03882560e-01 -9.50753450e-01 -1.53076872e-01
5.03829956e-01 -7.51253963e-01 6.78975821e-01 -3.93931925e-01
3.65453094e-01 -1.10035765e+00 2.53985859e-02 -9.77225006e-01
-4.43181098e-01 -1.17266488e+00 -7.61653706e-02 1.19309366e+00
5.55697739e-01 -2.96142310e-01 4.07353997e-01 1.05315280e+00
-7.61333704e-02 -1.00867234e-01 -4.48532254e-02 -4.93435830e-01
-1.19995147e-01 -4.24920738e-01 1.16961408e+00 8.41498375e-01
1.03202856e+00 4.03915673e-01 -3.13331932e-01 -3.30713876e-02
3.34392965e-01 3.79990846e-01 4.83043134e-01 -8.20842206e-01
-4.04425949e-01 -6.70957208e-01 4.60695028e-01 -6.45924985e-01
-2.32505903e-01 -8.75310302e-01 6.82400167e-01 -1.97128212e+00
-1.46938711e-01 -9.93435606e-02 2.83777654e-01 8.50534856e-01
2.53093362e-01 -3.28075618e-01 5.28221726e-01 -5.26284516e-01
-5.86302400e-01 -1.36701822e-01 1.16249919e+00 -2.58999139e-01
-7.34914601e-01 -1.68026984e-01 -7.41751790e-01 3.50635737e-01
1.11613345e+00 -4.40638423e-01 -6.60421908e-01 3.08238596e-01
8.25119376e-01 2.42828205e-01 2.88814753e-01 -7.02727914e-01
8.92669931e-02 -8.89696777e-01 -8.12061355e-02 -8.07753325e-01
-4.51261431e-01 -1.04837751e+00 9.14860427e-01 7.94424117e-01
-6.14854157e-01 2.72572249e-01 4.55254495e-01 -6.78528845e-02
2.67716020e-01 -5.93155205e-01 3.66490722e-01 -4.45191264e-01
-9.46823895e-01 -2.61298656e-01 -1.13638663e+00 -5.34041524e-01
1.21037412e+00 -1.79844767e-01 7.92921856e-02 -1.90689191e-01
-6.60318077e-01 3.93488616e-01 3.78355145e-01 2.48594746e-01
3.62326562e-01 -1.40792787e+00 -7.53384113e-01 6.35216162e-02
-6.57620952e-02 -2.72589475e-01 -9.90352631e-02 4.17488396e-01
-1.06957841e+00 5.26837230e-01 -1.98132947e-01 5.53364307e-02
-1.06050098e+00 1.15518141e+00 2.33140841e-01 -5.30683875e-01
-6.80703163e-01 6.29320890e-02 -2.27666527e-01 -3.48018855e-01
-3.00066564e-02 -1.15158117e+00 -2.52891511e-01 -3.84381205e-01
9.00081992e-01 3.12152237e-01 -4.56072420e-01 -1.71195030e-01
-2.09308013e-01 1.69659615e-01 4.84218806e-01 -2.88262755e-01
9.97995615e-01 -8.82268846e-02 -6.70611143e-01 4.35160428e-01
2.94302970e-01 -9.05773640e-02 -7.94106543e-01 2.97728390e-01
2.33935028e-01 1.73838511e-01 -2.52653152e-01 -1.09624779e+00
-3.55729938e-01 -7.97080174e-02 -4.07910086e-02 8.41075718e-01
1.30430400e+00 -1.90701663e-01 7.07856774e-01 3.17910075e-01
5.10667324e-01 -1.32365799e+00 -5.04704058e-01 4.06732172e-01
1.00125659e+00 -4.12894756e-01 3.10869426e-01 -7.42134690e-01
-5.56787729e-01 1.49634075e+00 9.28931534e-02 -1.43786326e-01
4.18177426e-01 6.14063323e-01 -2.84055561e-01 -3.41284215e-01
-9.16562617e-01 3.52817364e-02 -5.44387549e-02 7.32620001e-01
4.03704554e-01 4.55418438e-01 -8.17388177e-01 5.00864923e-01
-4.09167558e-01 4.32311833e-01 5.01216829e-01 9.20785367e-01
-6.59334302e-01 -1.52303851e+00 -8.35886061e-01 -3.73267345e-02
-2.19996870e-01 -2.48039693e-01 -3.43046486e-01 1.15127087e+00
7.07035899e-01 1.57435143e+00 -1.52173281e-01 -5.75059913e-02
3.86513829e-01 2.97194660e-01 7.85856009e-01 -8.23219061e-01
-1.10715997e+00 1.77977622e-01 8.75419855e-01 -4.70742106e-01
-7.42571890e-01 -8.00425887e-01 -1.78914201e+00 -6.27104044e-01
-1.23520479e-01 5.30075371e-01 4.01758254e-01 9.01929855e-01
-4.23068643e-01 3.67241114e-01 3.76064420e-01 -5.47109663e-01
-1.55264437e-01 -4.97812659e-01 -5.41304529e-01 1.73326001e-01
-2.02274024e-01 -9.78296995e-03 3.07519376e-01 9.25755322e-01] | [11.351414680480957, 4.613583087921143] |
587fd034-0f9b-4161-a298-7c494255d342 | refined-commonsense-knowledge-from-large | 2112.04596 | null | https://arxiv.org/abs/2112.04596v2 | https://arxiv.org/pdf/2112.04596v2.pdf | Refined Commonsense Knowledge from Large-Scale Web Contents | Commonsense knowledge (CSK) about concepts and their properties is helpful for AI applications. Prior works, such as ConceptNet, have compiled large CSK collections. However, they are restricted in their expressiveness to subject-predicate-object (SPO) triples with simple concepts for S and strings for P and O. This paper presents a method called ASCENT++ to automatically build a large-scale knowledge base (KB) of CSK assertions, with refined expressiveness and both better precision and recall than prior works. ASCENT++ goes beyond SPO triples by capturing composite concepts with subgroups and aspects, and by refining assertions with semantic facets. The latter is essential to express the temporal and spatial validity of assertions and further qualifiers. Furthermore, ASCENT++ combines open information extraction (OpenIE) with judicious cleaning and ranking by typicality and saliency scores. For high coverage, our method taps into the large-scale crawl C4 with broad web contents. The evaluation with human judgments shows the superior quality of the ASCENT++ KB, and an extrinsic evaluation for QA-support tasks underlines the benefits of ASCENT++. A web interface, data, and code can be accessed at https://ascentpp.mpi-inf.mpg.de/. | ['Gerhard Weikum', 'Julien Romero', 'Simon Razniewski', 'Tuan-Phong Nguyen'] | 2021-11-30 | null | null | null | null | ['open-information-extraction'] | ['natural-language-processing'] | [-2.94027895e-01 3.35157514e-01 -4.74083990e-01 -2.61606574e-01
-5.73299646e-01 -7.91408122e-01 6.31062090e-01 7.13224590e-01
-2.20250919e-01 9.21580195e-01 4.38861340e-01 -7.55733550e-02
-4.63015050e-01 -9.56349134e-01 -5.76582849e-01 -7.14621022e-02
-1.52362466e-01 5.50642014e-01 6.71856582e-01 -6.33502126e-01
3.62208515e-01 -4.57977206e-02 -1.74790680e+00 6.38423920e-01
1.18522334e+00 1.22901130e+00 -4.24389541e-02 8.63877088e-02
-3.51272225e-01 1.05148053e+00 -2.94776380e-01 -1.13016415e+00
-3.05000097e-02 1.61591191e-02 -1.18174744e+00 -5.29401898e-01
3.15578580e-01 -1.13068921e-02 9.54785496e-02 1.20664513e+00
1.67421997e-01 -6.86807483e-02 4.32772309e-01 -1.65072346e+00
-1.05874527e+00 1.07820487e+00 -1.46477535e-01 8.09652358e-02
7.62761354e-01 -1.59210458e-01 1.71038723e+00 -1.03037751e+00
8.39595318e-01 1.00222135e+00 8.39764774e-01 2.77954400e-01
-1.03994441e+00 -7.15609193e-01 2.77889799e-02 8.82000268e-01
-1.59314191e+00 -2.92506158e-01 6.76018596e-01 -2.98226774e-01
1.08630455e+00 5.64409375e-01 8.77412081e-01 9.78969157e-01
-3.67964655e-01 8.43721330e-01 1.11634648e+00 -5.43394148e-01
4.49309468e-01 5.31793654e-01 6.01959884e-01 7.05457509e-01
5.90070963e-01 -3.23336542e-01 -8.41114998e-01 -3.30721915e-01
3.15828770e-01 -2.11411938e-01 -3.27385575e-01 -3.80053878e-01
-1.32396328e+00 6.48958266e-01 4.70185578e-01 2.21650124e-01
-3.95665407e-01 -1.02210149e-01 6.18226409e-01 8.80956128e-02
3.58406544e-01 1.02194762e+00 -7.63562858e-01 -2.17026070e-01
-9.63952303e-01 5.65745234e-01 1.13967848e+00 1.36786866e+00
7.09492087e-01 -6.12704217e-01 -1.84599534e-01 7.66136169e-01
-6.39694035e-02 4.57298309e-01 4.86752391e-01 -1.18546379e+00
4.66599494e-01 9.81838107e-01 1.91476107e-01 -1.24413836e+00
-3.78312439e-01 -1.71311766e-01 -4.40044075e-01 -3.81950349e-01
-2.74661686e-02 1.13603741e-01 -3.69314075e-01 1.61918294e+00
4.75465000e-01 -2.41766777e-02 2.38581762e-01 8.01461756e-01
1.17631698e+00 4.22831595e-01 1.76819563e-01 -2.07024552e-02
1.76670456e+00 -6.49693787e-01 -8.85737479e-01 -2.63028461e-02
7.06050754e-01 -2.21292466e-01 1.19057155e+00 4.07865047e-01
-9.11438465e-01 2.46251095e-02 -9.31523561e-01 -3.07352543e-01
-9.39164281e-01 -1.30672708e-01 9.93349612e-01 3.61984819e-01
-9.22047377e-01 4.03967351e-01 -3.60065669e-01 -4.06600446e-01
5.80712736e-01 -6.37472197e-02 -4.58225787e-01 -6.22083549e-04
-2.01370239e+00 1.09929240e+00 9.98110890e-01 -4.66698885e-01
-5.08002102e-01 -1.18876266e+00 -9.88616705e-01 2.69164652e-01
1.11272001e+00 -5.82780480e-01 9.91000056e-01 -4.12062287e-01
-9.08268332e-01 8.22808146e-01 6.58803657e-02 -7.62592435e-01
5.82890697e-02 -3.64082545e-01 -7.09472418e-01 4.76986051e-01
5.59005082e-01 6.08026743e-01 3.91875595e-01 -1.29516423e+00
-7.47472942e-01 -3.46335024e-01 5.87199450e-01 1.84243605e-01
-5.80069959e-01 1.15825266e-01 -3.41527641e-01 -5.14436245e-01
6.62425607e-02 -3.64863098e-01 2.32238457e-01 -2.30915979e-01
-5.29874384e-01 -4.60518897e-01 4.64119107e-01 -6.89690232e-01
1.71904099e+00 -1.84590483e+00 -3.97514850e-02 3.36085469e-01
5.60156882e-01 3.30780260e-02 2.70896673e-01 7.24174976e-01
3.15116793e-02 2.41162583e-01 -1.46716774e-01 2.61814356e-01
3.99751425e-01 2.27788180e-01 -5.92080951e-01 -1.15935870e-01
1.45238325e-01 1.13667750e+00 -1.18825734e+00 -9.06341493e-01
-2.02050075e-01 -1.75909862e-01 -7.02097118e-01 -3.41096491e-01
-7.16223657e-01 -3.62586379e-01 -4.15364593e-01 7.90312111e-01
1.70832366e-01 -5.71595728e-01 1.94153279e-01 -5.02266228e-01
-2.02239421e-03 6.16017103e-01 -1.08779490e+00 1.52131236e+00
-2.17198849e-01 1.77529469e-01 -9.46753547e-02 -7.21063137e-01
7.15134084e-01 2.47200012e-01 5.49271166e-01 -6.13691390e-01
-7.55802691e-02 1.34789988e-01 -4.89660054e-01 -7.86978066e-01
9.70325768e-01 -1.11478724e-01 -1.95934325e-01 2.71734029e-01
3.82404715e-01 -3.26519579e-01 6.29923224e-01 9.58182395e-01
9.27037358e-01 1.20395921e-01 7.75166988e-01 -5.35373449e-01
3.83224845e-01 5.37850976e-01 6.18482292e-01 4.55070406e-01
-1.19992316e-01 -7.21054822e-02 7.15444505e-01 -8.89009908e-02
-9.81179953e-01 -8.42875063e-01 -3.75358462e-01 1.13689482e+00
3.07853162e-01 -1.16498911e+00 -4.52832937e-01 -7.81991720e-01
2.68939734e-01 9.88364160e-01 -6.80987716e-01 7.67920464e-02
-2.44044140e-02 -3.35847318e-01 9.54820693e-01 5.21988750e-01
5.26083708e-01 -9.38746154e-01 -4.08305407e-01 7.00291693e-02
-8.48979712e-01 -1.31921959e+00 1.27476066e-01 3.81199829e-02
-4.04175848e-01 -1.43197596e+00 -2.68990677e-02 -4.80416298e-01
2.60233730e-01 -2.08713338e-01 1.45978928e+00 5.73883802e-02
-1.42747136e-02 3.53253335e-01 -9.13873553e-01 -6.46824777e-01
-9.81748626e-02 -2.19605356e-01 3.04103047e-01 -3.27359825e-01
7.68507183e-01 -5.95341265e-01 -2.43938863e-01 2.97083706e-01
-9.03902113e-01 1.76335976e-01 2.14543954e-01 6.36508942e-01
5.29889822e-01 2.92558312e-01 7.14165509e-01 -9.71478820e-01
6.49178565e-01 -7.52346754e-01 -3.49851668e-01 4.60887939e-01
-9.83024418e-01 -1.00153089e-01 4.71336693e-01 -7.64173865e-02
-9.56316590e-01 -6.67636156e-01 -9.64315012e-02 -9.26008970e-02
-9.37344208e-02 8.81812036e-01 -2.10141227e-01 4.10915822e-01
1.02928948e+00 1.43544301e-01 -3.07267368e-01 -2.84966797e-01
6.56818748e-01 6.47322536e-01 7.43600488e-01 -1.13562405e+00
6.28451228e-01 5.93686879e-01 -2.27773443e-01 -6.11268461e-01
-1.38408458e+00 -6.00259006e-01 -4.84128445e-01 -1.85738932e-02
4.48195994e-01 -9.94822264e-01 -8.11076462e-01 -8.93577114e-02
-7.61725068e-01 -5.52448332e-02 -6.34614766e-01 2.35593855e-01
-4.49080378e-01 5.22359729e-01 -5.42410433e-01 -5.65326810e-01
-4.50811774e-01 -4.00284290e-01 7.06296384e-01 -8.59609544e-02
-4.07251209e-01 -8.40956569e-01 -8.07201564e-02 5.19859135e-01
2.91702569e-01 8.25865492e-02 9.11860883e-01 -1.18841457e+00
-2.66423613e-01 -1.73050508e-01 -3.43292654e-01 2.17521146e-01
-1.38356566e-01 -2.50064135e-01 -7.65871763e-01 1.68860078e-01
-3.26630086e-01 -6.37528241e-01 7.13053167e-01 7.80284824e-03
1.24737799e+00 -9.46159959e-01 -3.97298604e-01 1.46016076e-01
1.40762520e+00 -4.01706725e-01 5.56031466e-01 5.71439207e-01
5.27956903e-01 6.32751286e-01 8.96581829e-01 5.78009605e-01
8.72335792e-01 4.70697999e-01 4.41931725e-01 4.91299599e-01
8.61152783e-02 -3.65477800e-01 1.73918098e-01 5.81879497e-01
-1.68762356e-01 1.12491295e-01 -1.07475746e+00 9.24713850e-01
-1.76496625e+00 -1.35383379e+00 -8.97388682e-02 1.73288107e+00
1.53267682e+00 2.32142329e-01 1.35192290e-01 4.39592540e-01
4.81753081e-01 -1.89473569e-01 -2.35632956e-01 1.38519719e-01
-3.47143710e-01 1.57744214e-01 1.98810577e-01 3.65944743e-01
-7.40281224e-01 1.10478675e+00 5.33916664e+00 1.25360990e+00
-4.64656681e-01 1.28360108e-01 -3.66054103e-02 5.01228264e-04
-6.39158070e-01 3.54751199e-01 -1.07492423e+00 4.42275733e-01
6.99133515e-01 -4.17284042e-01 2.75320262e-01 1.02045071e+00
-2.21732870e-01 -1.75058469e-01 -1.19072449e+00 7.28271902e-01
1.54139891e-01 -1.71390915e+00 2.15878814e-01 -3.32721740e-01
4.96198446e-01 -2.94714663e-02 -4.19638276e-01 5.88753879e-01
5.27437031e-01 -7.16733634e-01 9.91573215e-01 4.83465046e-01
8.56013477e-01 -4.69490886e-01 7.18892097e-01 3.15240055e-01
-1.11515355e+00 -1.75714076e-01 -1.62575781e-01 -1.44344658e-01
5.08712195e-02 8.81879389e-01 -7.92169511e-01 8.79882336e-01
9.70273793e-01 8.87597322e-01 -7.22233951e-01 7.32952535e-01
-5.68206012e-01 7.30000854e-01 -3.24909896e-01 -3.38451326e-01
3.99098806e-02 7.06357956e-02 6.27177775e-01 1.51618159e+00
-8.16665217e-02 4.18655515e-01 1.78156108e-01 1.04224348e+00
-1.73226744e-01 2.17182338e-01 -3.01554918e-01 -1.42941624e-01
9.62726951e-01 1.22560930e+00 -4.49856699e-01 -7.94519365e-01
-1.66856378e-01 4.77863491e-01 4.90707338e-01 1.57494515e-01
-7.42261529e-01 -6.75397635e-01 5.81467807e-01 3.24239165e-01
1.66859284e-01 -1.12809129e-02 -3.82029444e-01 -1.33266485e+00
2.49165773e-01 -7.12136745e-01 8.96532714e-01 -1.02706432e+00
-1.59402919e+00 3.78091007e-01 5.38671613e-01 -8.86243641e-01
4.42933850e-02 -5.64974427e-01 -1.68077722e-02 4.86896038e-01
-1.63178730e+00 -1.36105478e+00 -3.16780359e-01 8.53023529e-01
8.97662714e-02 2.55280405e-01 8.23510170e-01 2.21338838e-01
-1.90011680e-01 3.15064579e-01 -3.56167823e-01 3.51323992e-01
6.62827373e-01 -1.47663414e+00 5.02821207e-02 6.77803218e-01
-3.83239947e-02 9.91534293e-01 7.75308967e-01 -9.41231370e-01
-1.01987541e+00 -9.36015129e-01 1.15468884e+00 -8.85637641e-01
1.18927789e+00 -2.07147613e-01 -9.75582719e-01 8.07534277e-01
1.00741602e-01 -6.32337257e-02 8.58585477e-01 5.83500266e-01
-9.03402686e-01 -9.44336802e-02 -1.23859918e+00 5.20840287e-01
1.16593778e+00 -5.26151717e-01 -1.18000007e+00 3.91900182e-01
1.06047106e+00 -3.58301252e-01 -1.28876567e+00 4.72574413e-01
4.25236493e-01 -7.39350021e-01 1.04983974e+00 -8.61115515e-01
7.64013469e-01 -4.37549889e-01 -3.98968399e-01 -1.13443053e+00
-4.05690938e-01 -1.58280015e-01 -6.12372160e-01 1.36698365e+00
6.59634888e-01 -5.06584406e-01 2.98118860e-01 7.00371444e-01
-1.77143306e-01 -7.40409613e-01 -6.97372258e-01 -8.32354069e-01
-1.53547779e-01 -7.92882860e-01 8.57312560e-01 1.42875636e+00
1.01783538e+00 3.48278850e-01 -5.68396300e-02 2.88646489e-01
5.54145098e-01 4.39318091e-01 2.59451538e-01 -1.47392988e+00
-1.29783796e-02 -3.43646348e-01 -4.13466036e-01 -4.90080118e-01
1.62618741e-01 -1.27634358e+00 -2.95568228e-01 -1.66038275e+00
4.15928245e-01 -7.94507384e-01 -1.27694309e-01 1.02906466e+00
-1.99134514e-01 2.79656470e-01 -2.58084070e-02 1.59929574e-01
-1.18673503e+00 4.63030010e-01 9.80510414e-01 -1.38478940e-02
-1.22923091e-01 -6.72389507e-01 -1.09031749e+00 7.34950602e-01
8.65392864e-01 -2.45438889e-01 -2.54779816e-01 -1.16564527e-01
1.05690551e+00 -3.01419348e-01 7.23875225e-01 -9.00394559e-01
2.79372931e-01 -3.02067101e-01 2.94350423e-02 -3.97873938e-01
3.61915261e-01 -7.66216636e-01 -1.22900762e-01 4.17923555e-02
-5.26202917e-01 -2.60748595e-01 2.12775767e-01 4.66494828e-01
-2.90020108e-01 -2.83443660e-01 4.00190592e-01 -3.13464940e-01
-1.21494043e+00 4.83150035e-02 2.04693049e-01 6.68004572e-01
8.58015060e-01 7.68406019e-02 -7.69129813e-01 -1.80961683e-01
-7.84994781e-01 5.14419138e-01 3.72528613e-01 3.74831200e-01
7.11284935e-01 -1.35574675e+00 -6.70145094e-01 -2.82612056e-01
7.96368182e-01 -7.03236014e-02 1.22465976e-01 9.09601986e-01
-1.90454751e-01 5.22916377e-01 -1.69320628e-01 -1.54168442e-01
-1.02912593e+00 8.45047534e-01 -8.53269249e-02 -2.45934084e-01
-7.37726510e-01 7.78660417e-01 -3.78055900e-01 -3.70172173e-01
1.85154378e-01 -3.58819604e-01 -4.31884199e-01 1.58161581e-01
6.67391121e-01 3.85475874e-01 1.67511731e-01 -3.51394892e-01
-6.15042567e-01 9.89747345e-02 5.79416230e-02 3.40291928e-03
1.32042480e+00 -1.34894013e-01 -6.36103868e-01 2.82715261e-01
5.04326940e-01 2.33149469e-01 -4.55766201e-01 -5.36301553e-01
3.12332511e-01 -3.03741723e-01 -2.55347311e-01 -1.26119685e+00
-3.39948922e-01 2.33187288e-01 -3.84249359e-01 4.73343462e-01
9.10671711e-01 4.87371027e-01 5.70973277e-01 6.16135180e-01
6.50640547e-01 -1.24964988e+00 -2.94450205e-02 5.73816538e-01
1.06502128e+00 -1.10222363e+00 8.44382197e-02 -6.90069318e-01
-9.84021008e-01 7.10701942e-01 6.35604560e-01 1.94113940e-01
6.05417669e-01 2.24457011e-01 -3.39209259e-01 -7.21847117e-01
-8.19751918e-01 -3.72262925e-01 3.58184785e-01 4.61025655e-01
1.66557312e-01 1.95064962e-01 -4.58870530e-01 1.22684360e+00
-7.20789671e-01 2.43258268e-01 4.32568640e-01 8.65939319e-01
-6.14773095e-01 -6.30497932e-01 -3.03872973e-01 6.36595309e-01
-3.69989604e-01 -5.04491568e-01 -4.04819310e-01 6.19896233e-01
3.26111525e-01 9.74722683e-01 -3.81358534e-01 -3.84392142e-01
3.09115887e-01 2.68096119e-01 2.09305421e-01 -6.35901690e-01
-4.15190905e-01 -7.35713422e-01 5.49317181e-01 -7.64479280e-01
-5.78598440e-01 -3.70261252e-01 -1.46012330e+00 -3.61787170e-01
-4.78214145e-01 7.24906027e-01 5.36591411e-01 1.00334501e+00
5.29203176e-01 3.37340295e-01 -5.48433401e-02 -1.66206911e-01
-5.01317859e-01 -8.84445846e-01 -6.58843696e-01 6.52031183e-01
1.21308133e-01 -7.71264315e-01 -2.01582476e-01 1.03697248e-01] | [9.671456336975098, 8.272083282470703] |
41236739-62c3-4bc9-ba70-05fc51c4ca3e | judge-the-judges-a-large-scale-evaluation | 1901.00398 | null | https://arxiv.org/abs/1901.00398v2 | https://arxiv.org/pdf/1901.00398v2.pdf | Judge the Judges: A Large-Scale Evaluation Study of Neural Language Models for Online Review Generation | We conduct a large-scale, systematic study to evaluate the existing evaluation methods for natural language generation in the context of generating online product reviews. We compare human-based evaluators with a variety of automated evaluation procedures, including discriminative evaluators that measure how well machine-generated text can be distinguished from human-written text, as well as word overlap metrics that assess how similar the generated text compares to human-written references. We determine to what extent these different evaluators agree on the ranking of a dozen of state-of-the-art generators for online product reviews. We find that human evaluators do not correlate well with discriminative evaluators, leaving a bigger question of whether adversarial accuracy is the correct objective for natural language generation. In general, distinguishing machine-generated text is challenging even for human evaluators, and human decisions correlate better with lexical overlaps. We find lexical diversity an intriguing metric that is indicative of the assessments of different evaluators. A post-experiment survey of participants provides insights into how to evaluate and improve the quality of natural language generation systems. | ['Shiyan Yan', 'Cristina Garbacea', 'Samuel Carton', 'Qiaozhu Mei'] | 2019-01-02 | judge-the-judges-a-large-scale-evaluation-1 | https://aclanthology.org/D19-1409 | https://aclanthology.org/D19-1409.pdf | ijcnlp-2019-11 | ['review-generation'] | ['natural-language-processing'] | [ 1.86880291e-01 3.90049964e-01 -2.76220918e-01 -3.28321487e-01
-1.23032129e+00 -1.20638108e+00 9.61073339e-01 5.39082408e-01
-4.34107035e-01 7.25045502e-01 5.49704373e-01 -2.87293702e-01
4.90361333e-01 -8.38493645e-01 -2.73014307e-01 -1.49469405e-01
6.33676648e-01 6.45515323e-01 -1.77468717e-01 -6.34176612e-01
5.25998831e-01 -3.53874527e-02 -1.32507837e+00 4.95847732e-01
1.19813454e+00 5.47323108e-01 -2.44471133e-01 6.98857605e-01
-1.55609008e-02 6.21082425e-01 -1.08785772e+00 -9.97981787e-01
2.14214310e-01 -7.36620486e-01 -9.08277750e-01 -1.33682653e-01
7.70313799e-01 -1.16789207e-01 2.52487421e-01 1.06556988e+00
6.61611497e-01 -1.75845712e-01 9.61764157e-01 -1.12670815e+00
-1.31410384e+00 1.02774942e+00 -2.10109487e-01 1.86448038e-01
8.23529541e-01 4.34036881e-01 1.43765771e+00 -8.57596934e-01
8.73163581e-01 1.31070399e+00 5.37593722e-01 4.96315509e-01
-1.26303172e+00 -6.17881536e-01 -1.77382484e-01 -3.70295197e-01
-1.03715909e+00 -3.09957892e-01 2.48530984e-01 -8.25863719e-01
6.88740492e-01 3.35245170e-02 1.44526511e-01 1.17169654e+00
4.07838106e-01 4.54122335e-01 1.17268622e+00 -4.60411638e-01
3.97852629e-01 6.52620256e-01 2.64731973e-01 3.30434859e-01
5.75247645e-01 2.36946847e-02 -5.55183709e-01 -5.04965544e-01
1.46908343e-01 -7.12643564e-01 -1.78458899e-01 1.51491836e-01
-1.16443431e+00 1.30827427e+00 5.25487103e-02 2.32063964e-01
-3.82314742e-01 -2.13644892e-01 6.27775848e-01 3.39378208e-01
5.73998451e-01 1.40856612e+00 -2.86993235e-01 -1.42847076e-01
-8.35817993e-01 6.56895697e-01 1.10526669e+00 9.20727909e-01
6.05964243e-01 6.01128526e-02 -6.22140408e-01 6.35968328e-01
-4.07961123e-02 5.47729075e-01 9.36717987e-01 -7.80599594e-01
4.95136410e-01 6.20387018e-01 4.67293292e-01 -1.04413855e+00
1.66014452e-02 -3.81393015e-01 -4.70406204e-01 3.75469565e-01
3.92865062e-01 -4.37974989e-01 -5.19343615e-01 1.62246907e+00
-1.30583957e-01 -8.82180095e-01 5.26175722e-02 6.21082723e-01
7.04400241e-01 6.38286233e-01 1.38013214e-01 -8.93466547e-02
1.50526857e+00 -9.16095018e-01 -5.06552756e-01 -3.83657962e-01
8.13937843e-01 -1.25473893e+00 1.39416826e+00 4.61114377e-01
-1.07799172e+00 -9.24445331e-01 -1.30241144e+00 1.23255253e-01
-5.18997133e-01 2.12409049e-01 2.92242795e-01 9.94562685e-01
-9.47010934e-01 6.85867727e-01 -1.00040436e-01 -2.14621261e-01
-4.74070571e-02 -5.28250225e-02 -2.59042382e-01 3.07293504e-01
-1.47849798e+00 1.32163894e+00 2.06506059e-01 -3.57921392e-01
-8.04553449e-01 -6.17807925e-01 -9.10465181e-01 -3.04658890e-01
-4.15485213e-03 -6.58179700e-01 1.86644995e+00 -1.14622426e+00
-1.21154308e+00 1.03785264e+00 -1.12556346e-01 -5.08454263e-01
4.96603519e-01 -9.30647403e-02 -5.40966332e-01 -1.54706672e-01
5.89732826e-01 6.14008427e-01 5.25918543e-01 -1.17575812e+00
-6.67106092e-01 -6.62660301e-02 1.42888144e-01 2.48905048e-01
-8.73265117e-02 6.03930503e-02 4.50889945e-01 -6.92620337e-01
-5.73055983e-01 -8.75174463e-01 -3.37753236e-01 -4.63088721e-01
-5.27194083e-01 -5.54571450e-01 1.54064462e-01 -5.79760253e-01
1.44391203e+00 -1.77927029e+00 -3.77257615e-01 1.21729568e-01
3.28255117e-01 3.22014838e-01 -4.98529077e-01 8.70924234e-01
-3.45008820e-02 8.02195370e-01 1.07978091e-01 -1.55254051e-01
4.34211135e-01 -5.66999614e-01 -6.40610576e-01 1.78685877e-02
4.08350229e-01 1.01813185e+00 -1.18504345e+00 -2.61155218e-01
-3.31650853e-01 1.19471394e-01 -1.90485984e-01 2.47059315e-02
-3.43693852e-01 -2.59345416e-02 -3.60095918e-01 3.06326300e-01
2.14506179e-01 -1.99053630e-01 -6.34348169e-02 -1.02834087e-02
9.91332345e-03 8.35799456e-01 -7.88024783e-01 1.06734860e+00
-5.59321523e-01 8.43039453e-01 -3.86739016e-01 -8.72465074e-02
1.12284815e+00 2.90930271e-01 -4.58494127e-01 -7.92838991e-01
1.18803032e-01 4.96573836e-01 2.12327778e-01 -8.42245743e-02
1.15165424e+00 -2.14991003e-01 -4.71475244e-01 1.05414224e+00
-7.30215013e-02 -5.11783600e-01 7.21493661e-01 4.50933725e-01
1.04003704e+00 -2.01419935e-01 5.55175066e-01 -4.31057483e-01
2.80140162e-01 2.38278285e-01 1.04259692e-01 9.48655427e-01
-1.28950596e-01 7.22161591e-01 6.44544959e-01 -3.10023516e-01
-1.11820543e+00 -1.02026379e+00 1.64135903e-01 1.15008640e+00
-5.48417913e-03 -7.32849419e-01 -1.06326270e+00 -8.47063541e-01
-1.08617067e-01 1.23913467e+00 -8.24561059e-01 -2.98457891e-01
-1.25192478e-01 -5.28534770e-01 6.35456026e-01 4.26338553e-01
-1.59122586e-01 -1.29001021e+00 -4.45356101e-01 1.67469725e-01
-2.57653564e-01 -8.52905571e-01 -8.94445956e-01 -1.87694460e-01
-3.85551006e-01 -9.04037356e-01 -6.83254540e-01 -7.91210234e-01
5.72596312e-01 9.89027098e-02 1.96238995e+00 -7.19367117e-02
-2.21636854e-02 5.58489375e-02 -5.88186324e-01 -6.83717787e-01
-1.32681191e+00 4.59493011e-01 -2.29122862e-02 -3.85788649e-01
6.24694467e-01 -2.63007972e-02 -5.18601537e-01 5.89420438e-01
-9.36590135e-01 -3.10131848e-01 5.91296732e-01 6.07951224e-01
2.42120281e-01 -1.07466854e-01 8.84433866e-01 -1.13965118e+00
1.67490697e+00 -3.36251765e-01 -2.87727892e-01 4.19021457e-01
-8.97691369e-01 3.24261606e-01 7.85834014e-01 -5.08062303e-01
-6.98720813e-01 -4.07255173e-01 1.41312942e-01 4.93707865e-01
2.91586593e-02 6.16543829e-01 1.76907957e-01 3.29312086e-01
1.33578742e+00 -1.69235587e-01 -8.66127014e-02 1.32247195e-01
6.56864583e-01 6.73254430e-01 2.34521568e-01 -4.90963697e-01
8.61843050e-01 -3.93661670e-02 -6.64764047e-01 -3.26550037e-01
-8.95357430e-01 -2.40995198e-01 -3.51102054e-01 2.76099537e-02
8.78675103e-01 -9.34367478e-01 -3.02633077e-01 2.92526394e-01
-1.44470048e+00 -3.35166663e-01 -5.95697165e-01 1.81223661e-01
-1.62050560e-01 1.71935081e-01 -5.56989729e-01 -6.71772778e-01
-7.63593435e-01 -1.21204543e+00 1.31405818e+00 1.89551011e-01
-1.34136796e+00 -1.04683328e+00 5.11500120e-01 4.20493543e-01
4.31389362e-01 3.62131119e-01 8.62170339e-01 -1.00448811e+00
-4.03035618e-02 -6.30912364e-01 1.03946880e-01 5.25192559e-01
9.18361098e-02 2.58216262e-01 -7.92369127e-01 1.49517525e-02
-1.43926561e-01 -5.45316279e-01 5.71545541e-01 8.82371888e-03
2.65455097e-01 -5.14344692e-01 -7.66720250e-02 -3.82244825e-01
1.12525856e+00 2.65321806e-02 7.54583716e-01 2.20926121e-01
4.51652408e-01 9.71660495e-01 7.19376028e-01 2.57519007e-01
2.62148917e-01 5.10594249e-01 -7.63451755e-02 -2.36885739e-03
9.21123028e-02 -6.82443321e-01 6.65444255e-01 7.67817736e-01
4.08826560e-01 -6.42949164e-01 -7.08742499e-01 4.91320848e-01
-1.39284205e+00 -1.00677562e+00 -1.46651551e-01 2.15393090e+00
1.11320603e+00 3.67931724e-01 3.52846324e-01 -1.15102068e-01
9.28653002e-01 2.61447459e-01 -2.96472013e-01 -9.12142098e-01
-1.85774833e-01 4.33470845e-01 4.28064287e-01 6.07755661e-01
-7.46436059e-01 6.90003872e-01 7.46700668e+00 9.86351848e-01
-7.72510529e-01 -1.03785887e-01 9.63901639e-01 1.93732396e-01
-7.64742672e-01 8.90307780e-03 -9.67391729e-01 6.00923419e-01
1.12662756e+00 -8.54677856e-01 5.12380674e-02 1.01491523e+00
1.88521445e-01 7.98743218e-02 -1.47857451e+00 5.61114371e-01
3.23346466e-01 -1.16959476e+00 3.59928191e-01 1.76097482e-01
1.15213573e+00 -3.27829301e-01 1.71775460e-01 3.56316328e-01
8.12117517e-01 -1.37351620e+00 1.06445587e+00 1.44242465e-01
7.37035275e-01 -6.00286245e-01 9.08895195e-01 2.64634728e-01
-8.52544725e-01 2.86992759e-01 -2.09156424e-01 -9.56001133e-02
3.95167768e-01 7.66170382e-01 -1.11687827e+00 1.52513698e-01
7.78529346e-02 2.41985634e-01 -9.55263078e-01 3.52239460e-01
-5.69983244e-01 4.46393460e-01 1.62412748e-01 -4.82409626e-01
2.12337285e-01 1.99605129e-03 2.72167623e-01 1.17832243e+00
4.18363005e-01 -3.98576766e-01 1.49690434e-02 1.08184636e+00
-2.48039976e-01 3.20491672e-01 -6.67482436e-01 -5.73179483e-01
5.43412805e-01 1.42069936e+00 -5.87784111e-01 -3.82641971e-01
4.98592332e-02 7.77601302e-01 5.98414466e-02 8.15876350e-02
-5.96734226e-01 -5.87379217e-01 4.02877599e-01 3.09327334e-01
-4.59145345e-02 1.26007907e-02 -5.80403209e-01 -7.82423735e-01
1.50132403e-01 -1.49710393e+00 8.59753266e-02 -9.57666159e-01
-1.73033392e+00 1.09886706e+00 -3.28978360e-01 -1.39036787e+00
-7.19761908e-01 -7.20827162e-01 -8.49466562e-01 1.06522512e+00
-7.99277365e-01 -5.59266508e-01 -2.44392112e-01 -2.11460933e-01
6.10166013e-01 -3.47754300e-01 7.87232935e-01 -1.00744829e-01
-1.88805446e-01 8.97180319e-01 -2.96111792e-01 1.12629220e-01
9.50978577e-01 -1.41044176e+00 1.04421091e+00 7.64899552e-01
2.64221936e-01 8.52004886e-01 9.49813604e-01 -7.98425734e-01
-7.29128957e-01 -9.64530945e-01 1.17357659e+00 -9.69394624e-01
7.65673041e-01 -2.29936123e-01 -5.65411270e-01 1.92719087e-01
6.62993848e-01 -7.52846539e-01 9.49851751e-01 1.32821292e-01
-6.90745652e-01 1.68251708e-01 -9.80605960e-01 7.21096337e-01
7.15147078e-01 -8.81166160e-01 -5.70630610e-01 4.06177729e-01
7.73522556e-01 -1.09317109e-01 -6.73545957e-01 1.16406403e-01
4.86212850e-01 -8.30584764e-01 4.55126613e-01 -5.20563126e-01
1.13143718e+00 -2.84036249e-01 2.24379763e-01 -1.68870342e+00
-3.38258386e-01 -8.51220787e-01 5.71145117e-01 1.47714972e+00
1.14622986e+00 -6.70042872e-01 4.87260699e-01 7.80779898e-01
8.13905597e-02 -5.96200228e-01 -9.72693413e-02 -6.92338288e-01
4.41260934e-01 4.71571684e-02 5.15962660e-01 7.61645079e-01
1.95115939e-01 1.02459633e+00 9.27402154e-02 -4.02657837e-01
5.81745915e-02 -7.83988312e-02 8.73134494e-01 -9.61930811e-01
-3.02409232e-01 -7.66606212e-01 -4.12937790e-01 -5.88439882e-01
2.41888613e-01 -9.26402628e-01 2.03479871e-01 -1.50032198e+00
1.39129445e-01 1.15697728e-02 2.27099881e-01 4.61417176e-02
-4.89237189e-01 3.26480091e-01 1.32605150e-01 6.70886710e-02
-5.50587714e-01 2.59880900e-01 1.15731037e+00 -1.76290825e-01
-1.37739018e-01 -7.68130794e-02 -1.47493267e+00 4.14962560e-01
8.24866533e-01 -3.64144683e-01 -6.66854322e-01 -3.22086275e-01
8.35763454e-01 -4.76468205e-01 1.94100633e-01 -7.45312154e-01
-1.22955710e-01 -5.15470952e-02 1.89292848e-01 -1.24320656e-01
-3.46145779e-01 2.09161490e-02 6.53821975e-02 2.32946366e-01
-9.12419260e-01 7.14802921e-01 -1.55087356e-02 2.31894121e-01
-3.94548088e-01 -6.29563808e-01 5.86002052e-01 -3.86561602e-01
-2.55793512e-01 -3.46045867e-02 -4.90216374e-01 6.36694849e-01
7.11870253e-01 -2.07965568e-01 -7.17414558e-01 -7.65020370e-01
-1.92563593e-01 -5.01454733e-02 6.86871707e-01 5.12251556e-01
1.45359129e-01 -1.26883686e+00 -1.18052542e+00 -3.88337642e-01
3.77685606e-01 -5.44929624e-01 -2.58706719e-01 2.59892344e-01
-5.45838952e-01 3.65883023e-01 7.21156597e-02 -9.72674191e-02
-8.87861907e-01 2.41898462e-01 1.68357179e-01 -8.49988580e-01
2.94661462e-01 7.32451737e-01 2.88028598e-01 -3.46388787e-01
-1.93620116e-01 -2.12781936e-01 -3.58463824e-02 4.67532545e-01
7.85023868e-01 3.70157242e-01 2.55727232e-01 -5.59935331e-01
-5.23975082e-02 2.82001585e-01 -3.49708498e-01 -6.73642039e-01
5.69306493e-01 1.62546769e-01 3.68189253e-02 5.53304374e-01
1.11250663e+00 3.99622470e-01 -5.36149144e-01 1.29007727e-01
-5.36891744e-02 -1.12777844e-01 -5.41893125e-01 -9.96388137e-01
-4.81242597e-01 6.88996553e-01 1.63014159e-01 7.32156813e-01
5.56168258e-01 -1.37304947e-01 8.50107610e-01 2.09604919e-01
4.43193644e-01 -1.24808776e+00 3.46548587e-01 5.09157598e-01
1.18208599e+00 -1.33745122e+00 1.86966255e-01 -3.96122903e-01
-1.13970339e+00 8.28410029e-01 6.55596852e-01 -1.98089182e-01
2.79683858e-01 8.85962602e-03 3.67826015e-01 -5.49552962e-02
-7.39970148e-01 1.12768508e-01 5.10896444e-01 6.29608989e-01
8.77682865e-01 1.90073207e-01 -6.47053182e-01 5.88417232e-01
-1.09637702e+00 -3.00599158e-01 7.13967383e-01 4.93305802e-01
-2.45155871e-01 -1.25202513e+00 -7.18344525e-02 4.46454078e-01
-5.07136881e-01 -3.48461092e-01 -9.46521103e-01 5.73791265e-01
-1.41187727e-01 1.33893824e+00 -1.34185329e-01 -5.38686097e-01
3.54701936e-01 1.20467246e-02 1.43092796e-01 -1.13539970e+00
-1.10121167e+00 -4.68862146e-01 5.40424466e-01 -1.74382880e-01
-9.84170511e-02 -5.59118688e-01 -8.29029918e-01 -2.37232566e-01
-5.59661806e-01 5.38540363e-01 5.36510348e-01 8.36386204e-01
3.46152127e-01 -4.22218479e-02 7.19876111e-01 -6.08137071e-01
-1.01621568e+00 -1.10327101e+00 -3.05876046e-01 7.81353295e-01
7.08385855e-02 -6.01034462e-02 -7.30436802e-01 1.44135490e-01] | [11.79723072052002, 8.955818176269531] |
54df8c9c-657c-40e1-a8e8-ea05c069789e | dynamic-adaptive-spatio-temporal-graph | 2109.12517 | null | https://arxiv.org/abs/2109.12517v1 | https://arxiv.org/pdf/2109.12517v1.pdf | Dynamic Adaptive Spatio-temporal Graph Convolution for fMRI Modelling | The characterisation of the brain as a functional network in which the connections between brain regions are represented by correlation values across time series has been very popular in the last years. Although this representation has advanced our understanding of brain function, it represents a simplified model of brain connectivity that has a complex dynamic spatio-temporal nature. Oversimplification of the data may hinder the merits of applying advanced non-linear feature extraction algorithms. To this end, we propose a dynamic adaptive spatio-temporal graph convolution (DAST-GCN) model to overcome the shortcomings of pre-defined static correlation-based graph structures. The proposed approach allows end-to-end inference of dynamic connections between brain regions via layer-wise graph structure learning module while mapping brain connectivity to a phenotype in a supervised learning framework. This leverages the computational power of the model, data and targets to represent brain connectivity, and could enable the identification of potential biomarkers for the supervised target in question. We evaluate our pipeline on the UKBiobank dataset for age and gender classification tasks from resting-state functional scans and show that it outperforms currently adapted linear and non-linear methods in neuroimaging. Further, we assess the generalizability of the inferred graph structure by transferring the pre-trained graph to an independent dataset for the same task. Our results demonstrate the task-robustness of the graph against different scanning parameters and demographics. | ['Guido van Wingen', 'Rajat Mani Thomas', 'Ahmed El-Gazzar'] | 2021-09-26 | null | null | null | null | ['age-and-gender-classification', 'graph-structure-learning'] | ['computer-vision', 'graphs'] | [ 2.01123804e-01 3.17115575e-01 7.47489855e-02 -6.21820807e-01
1.39475897e-01 -5.17276943e-01 9.02945459e-01 3.90083611e-01
-4.87206221e-01 5.01782060e-01 3.17745537e-01 -3.63885909e-01
-6.78863108e-01 -7.71302581e-01 -4.61546272e-01 -3.91777605e-01
-1.02007794e+00 5.90989292e-01 1.20378107e-01 -5.64254709e-02
3.79840545e-02 5.74370682e-01 -9.67493713e-01 1.11211926e-01
7.07179427e-01 7.24930227e-01 -5.25832660e-02 4.18806463e-01
1.46518871e-01 1.62657350e-01 -2.26080269e-01 -5.24795413e-01
8.23827311e-02 -4.45799798e-01 -7.57858157e-01 -3.33672762e-01
5.49104333e-01 2.09537864e-01 -4.00004894e-01 8.75448227e-01
6.28836334e-01 -8.82599205e-02 6.38424277e-01 -1.07780850e+00
-3.10456604e-01 7.60246277e-01 -4.89148349e-01 8.50815773e-01
2.01898798e-01 2.91773677e-01 9.36110795e-01 -3.72629702e-01
8.94802809e-01 1.29788244e+00 9.14002478e-01 3.24047625e-01
-1.79536974e+00 -6.81743324e-01 1.07256502e-01 1.26961559e-01
-1.12217808e+00 -3.66271406e-01 3.79162848e-01 -7.64475167e-01
1.28646004e+00 -2.35498641e-02 1.19790781e+00 1.17597365e+00
4.45981830e-01 -4.91948463e-02 1.37004900e+00 1.06952395e-02
-4.12849598e-02 -6.82682872e-01 1.76790848e-01 1.05528903e+00
2.57859528e-01 2.17991486e-01 -6.21194065e-01 -3.06535989e-01
6.33234859e-01 -1.84602454e-01 -1.74503680e-02 -3.02565098e-01
-1.45771921e+00 6.40810132e-01 8.27076256e-01 5.10780513e-01
-3.92793804e-01 2.44392559e-01 7.77992725e-01 2.26530373e-01
8.49657059e-01 2.38342151e-01 -4.66657817e-01 2.18800530e-01
-1.37361121e+00 1.16838366e-02 6.27448380e-01 1.85038462e-01
4.08690691e-01 -6.26287013e-02 -1.94236174e-01 7.27867723e-01
3.19904834e-01 2.64654160e-02 5.11272013e-01 -4.98595089e-01
3.01371634e-01 7.58225262e-01 -7.81655014e-01 -1.11170423e+00
-1.26472771e+00 -6.50352061e-01 -9.07452703e-01 4.06584330e-02
5.81417024e-01 -1.82477012e-01 -6.64076865e-01 1.95346689e+00
3.63674164e-01 3.83903116e-01 -6.70131505e-01 7.56317675e-01
6.49409950e-01 -1.82973772e-01 4.43271816e-01 5.98553903e-02
1.40448594e+00 -3.10435116e-01 -2.52076924e-01 -2.25313276e-01
6.18139386e-01 2.75836848e-02 5.98283768e-01 1.24229208e-01
-7.72771895e-01 -3.55951399e-01 -8.99021208e-01 -4.98745069e-02
-5.43138862e-01 -7.86821768e-02 1.08200109e+00 6.73407733e-01
-1.41164911e+00 8.07022154e-01 -1.13839638e+00 -6.99930608e-01
8.29775035e-01 6.87171221e-01 -7.99777389e-01 -2.43618870e-05
-1.07295072e+00 1.02240825e+00 5.11545658e-01 2.38558501e-01
-5.75824320e-01 -1.16580129e+00 -6.40485704e-01 1.33673787e-01
-9.13821533e-02 -1.00396597e+00 2.38675937e-01 -6.75469875e-01
-9.25198913e-01 1.12139082e+00 9.42073879e-04 -6.63991868e-01
5.35120964e-01 2.36987770e-01 -5.73652744e-01 3.50861162e-01
6.99091330e-02 6.47999525e-01 5.88549316e-01 -3.73971909e-01
1.38664722e-01 -7.49187529e-01 -2.02232599e-01 -2.24901676e-01
-7.69607797e-02 6.95370659e-02 1.06231235e-02 -4.54756588e-01
1.69263124e-01 -7.78991163e-01 -1.31635368e-01 2.51832873e-01
-4.35882449e-01 -1.85820702e-02 3.15086395e-01 -8.95592690e-01
9.74725366e-01 -1.85723484e+00 2.48228461e-01 5.59086800e-01
7.73181975e-01 -1.35111153e-01 -1.99641421e-01 4.54005688e-01
-6.15589797e-01 2.18903035e-01 -4.13395494e-01 -5.09546511e-02
-2.34490022e-01 -2.14047223e-01 3.27685356e-01 1.04573643e+00
4.17476267e-01 1.32164705e+00 -1.01114750e+00 -3.59917819e-01
-4.77901027e-02 7.72394598e-01 -6.44502640e-01 -1.82907388e-01
8.99942070e-02 7.22611129e-01 -1.58233196e-01 3.74426067e-01
4.12573040e-01 -9.34151933e-02 6.33554101e-01 -2.67241389e-01
6.61064610e-02 2.87061870e-01 -4.46520984e-01 1.96304679e+00
-1.43875927e-01 6.27432883e-01 2.20091343e-02 -1.45631981e+00
9.53712106e-01 1.12033121e-01 7.05771327e-01 -7.63952017e-01
1.00225270e-01 -5.18069696e-03 9.03448701e-01 -4.67262715e-01
-5.04288256e-01 -1.23610690e-01 2.26970807e-01 4.84728545e-01
5.22934318e-01 3.14824104e-01 2.82736003e-01 2.70937979e-01
1.67901015e+00 2.98714757e-01 3.11219603e-01 -7.37567365e-01
3.90800744e-01 -2.20712483e-01 2.05854163e-01 3.82377297e-01
-1.72468826e-01 2.48307154e-01 9.44682240e-01 -2.62926519e-01
-8.18442762e-01 -1.00067687e+00 -3.29609185e-01 9.41504300e-01
-8.17000270e-01 -4.42471772e-01 -5.37188768e-01 -6.42549932e-01
1.85943499e-01 2.11977556e-01 -1.10517335e+00 -3.84464294e-01
-5.24413764e-01 -1.13388598e+00 9.31227446e-01 1.65569171e-01
3.58892590e-01 -8.08691144e-01 -4.91081625e-01 1.14920639e-01
1.39120474e-01 -1.08808470e+00 -2.56377995e-01 6.48419112e-02
-1.18311381e+00 -1.45217144e+00 -4.16527778e-01 -3.34538609e-01
7.86392093e-01 -3.12136084e-01 1.23087609e+00 1.62012145e-01
-7.07775235e-01 4.29091543e-01 1.55404508e-02 -5.58283320e-03
-3.01937051e-02 2.28619531e-01 7.59067833e-02 -1.14431139e-02
2.01184750e-01 -1.19113111e+00 -1.10808885e+00 -6.99512884e-02
-5.65326989e-01 -5.02924621e-02 6.59978032e-01 8.70331764e-01
2.13656992e-01 -3.24479908e-01 7.43923008e-01 -1.11737454e+00
7.46792853e-01 -9.49666202e-01 -3.67559195e-01 6.65818378e-02
-8.20178270e-01 1.45732075e-01 3.45894068e-01 -4.36272264e-01
-6.72422349e-01 -2.25402266e-02 1.31972358e-01 8.11301768e-02
-2.41931267e-02 9.65223908e-01 8.28676522e-02 -1.29752651e-01
6.82937741e-01 -6.23316541e-02 4.43438411e-01 -2.90370226e-01
5.77422619e-01 -7.15551898e-02 5.05923510e-01 -5.60909629e-01
4.14226472e-01 5.48053026e-01 7.01706529e-01 -6.15783691e-01
-3.99065673e-01 -2.52708703e-01 -1.31477869e+00 -3.76431316e-01
8.33558440e-01 -7.37904310e-01 -8.79834831e-01 3.57563384e-02
-7.49305964e-01 -5.37707388e-01 8.95408243e-02 4.32741344e-01
-2.06700355e-01 2.99143463e-01 -3.31121981e-01 -4.60129261e-01
-5.00936747e-01 -8.24251235e-01 7.73610294e-01 -2.18167230e-01
-4.96813446e-01 -1.53902650e+00 2.39390135e-01 1.54618606e-01
5.14349580e-01 8.74523222e-01 1.22160721e+00 -6.99485421e-01
-6.53320029e-02 -2.23351091e-01 -3.83256376e-01 -1.57105535e-01
-5.04861213e-02 -1.75570771e-01 -6.61913455e-01 -2.46748835e-01
-4.71167535e-01 -3.77540812e-02 9.71883118e-01 4.28027600e-01
1.05118656e+00 3.44000524e-03 -4.46537793e-01 6.22096181e-01
1.21583092e+00 -4.95646477e-01 4.02930945e-01 2.73573697e-02
7.07211792e-01 9.26324368e-01 -2.58942217e-01 1.26553252e-01
4.87038761e-01 5.69789827e-01 3.99595171e-01 -2.23743767e-01
-3.35668206e-01 4.02405411e-02 3.19970131e-01 5.11606097e-01
-3.12694401e-01 3.01604092e-01 -1.14378834e+00 4.53483790e-01
-1.63625109e+00 -9.75228786e-01 -5.46884537e-01 2.09091902e+00
6.64051831e-01 3.14625680e-01 4.87137049e-01 -2.36696765e-01
6.24721169e-01 2.88401041e-02 -6.85243905e-01 -3.64186168e-01
-1.35529995e-01 6.48565948e-01 3.43586564e-01 1.18224680e-01
-6.84228182e-01 7.05380380e-01 6.78772926e+00 1.53157219e-01
-1.26512253e+00 3.30721766e-01 6.12378597e-01 -2.02842638e-01
4.94632386e-02 1.90975368e-01 -1.24456681e-01 3.91053915e-01
1.40703583e+00 -1.49865076e-01 8.13251257e-01 1.20192058e-01
5.11627138e-01 -1.25126362e-01 -1.31569147e+00 6.25389516e-01
-2.83865388e-02 -1.16041625e+00 -3.86001229e-01 3.16841006e-01
1.93598703e-01 5.93364000e-01 -1.33558571e-01 6.28579482e-02
1.40263394e-01 -1.27615082e+00 4.07687455e-01 9.90850389e-01
9.72043574e-01 -3.91832560e-01 2.61497349e-01 1.31753698e-01
-1.08186340e+00 -6.48960005e-03 -4.27613668e-02 -5.17256074e-02
-3.51341139e-03 9.20426488e-01 -1.00491440e+00 5.78732371e-01
4.75882888e-01 8.84056032e-01 -1.14631379e+00 1.10739493e+00
-6.95585459e-02 8.34719002e-01 -2.51604348e-01 3.51239800e-01
-6.73954487e-02 -4.39223081e-01 3.16960752e-01 1.44093692e+00
2.15372369e-01 -1.87651232e-01 -1.40269548e-01 1.29290390e+00
9.35749561e-02 8.06461275e-02 -6.33487403e-01 -3.47004831e-01
9.15112421e-02 1.75582063e+00 -1.17502916e+00 5.34740975e-03
-5.52579701e-01 7.62753427e-01 8.12304974e-01 2.03542233e-01
-5.23717105e-01 4.67562377e-02 5.02409279e-01 4.24652725e-01
-7.56409205e-03 -5.68009436e-01 -1.50653899e-01 -1.03041613e+00
-1.96659088e-01 -4.25306857e-01 6.13610804e-01 -5.10290205e-01
-1.58812678e+00 3.79959822e-01 1.82847112e-01 -5.36771357e-01
-9.97915864e-02 -6.00088894e-01 -8.83780181e-01 1.13609540e+00
-1.13797617e+00 -1.33335018e+00 -1.09699346e-01 4.80308205e-01
-1.43694967e-01 -4.13431227e-02 8.60793710e-01 2.18517840e-01
-6.43945515e-01 2.90390372e-01 -4.28165764e-01 2.17918143e-01
5.16427994e-01 -1.30222058e+00 4.79940444e-01 8.17076743e-01
4.48277667e-02 1.07515883e+00 4.95153964e-01 -9.72860277e-01
-1.27076948e+00 -9.71925735e-01 7.36904740e-01 -3.71589869e-01
1.12906325e+00 -7.38236189e-01 -8.57875288e-01 5.58421671e-01
5.15417494e-02 3.93269151e-01 8.45012069e-01 4.65304047e-01
-5.12031198e-01 3.55137400e-02 -1.17404747e+00 2.99234569e-01
1.60643625e+00 -6.08626723e-01 -3.20425808e-01 3.47022444e-01
1.89053133e-01 1.24846749e-01 -1.31508768e+00 2.55182028e-01
6.90158427e-01 -8.97505283e-01 1.03640449e+00 -7.57509291e-01
3.84084642e-01 1.07486404e-01 4.63575959e-01 -1.57015812e+00
-4.99503195e-01 -3.03897560e-01 -7.60426894e-02 1.10984433e+00
4.65494126e-01 -8.29382300e-01 4.53087240e-01 6.03938341e-01
-4.17527854e-02 -8.64666045e-01 -1.13958549e+00 -5.41202903e-01
2.92172819e-01 -2.25619331e-01 3.76883864e-01 1.09217727e+00
2.80500591e-01 4.39581394e-01 1.25990689e-01 -7.90480077e-02
5.71605086e-01 -2.86042929e-01 2.07579225e-01 -1.75794029e+00
-2.45488986e-01 -7.44125068e-01 -8.58441651e-01 -5.28156161e-02
6.15109086e-01 -1.80733013e+00 -6.30821586e-01 -1.62121201e+00
3.11315864e-01 -2.77945638e-01 -4.12183225e-01 6.26698375e-01
3.40163223e-02 2.12834269e-01 -2.99331427e-01 -5.85463978e-02
-2.17192739e-01 1.31635472e-01 1.04215121e+00 -7.44580552e-02
-2.89796423e-02 -3.94628406e-01 -5.64978480e-01 3.16943258e-01
8.94122481e-01 -3.95785600e-01 -5.59650064e-01 -1.46966442e-01
3.20510298e-01 -4.95834872e-02 6.94284856e-01 -8.61843467e-01
1.46492288e-01 3.53754222e-01 7.83108830e-01 -2.34659594e-02
1.59741879e-01 -5.86048841e-01 4.01094139e-01 7.39408612e-01
-3.29666704e-01 3.62494439e-01 1.10563204e-01 6.09350920e-01
3.05460453e-01 3.25934440e-01 5.10596454e-01 -2.35918343e-01
-2.07138762e-01 5.03652394e-01 -1.88703254e-01 3.86326723e-02
5.77636421e-01 -1.11298107e-01 -4.25221920e-01 3.03062331e-03
-1.07097697e+00 3.65028232e-02 3.03187165e-02 2.52363086e-01
3.58084977e-01 -9.47623670e-01 -7.72711515e-01 8.78886133e-03
-5.52949198e-02 -6.11481249e-01 3.08908373e-01 1.56042790e+00
-3.57434660e-01 2.90159047e-01 -6.22590542e-01 -5.32992482e-01
-1.16180050e+00 3.60113949e-01 5.64472497e-01 -2.83099025e-01
-8.47710013e-01 5.04877508e-01 -1.39074773e-01 -3.90125245e-01
-2.65343547e-01 -2.54200995e-01 -4.44708645e-01 3.52863222e-01
8.60483721e-02 3.37765306e-01 1.93419993e-01 -8.30187023e-01
-5.26241302e-01 1.32904634e-01 1.18080609e-01 -1.15584962e-01
1.83967757e+00 -6.61407113e-02 -7.41618574e-01 4.70311165e-01
1.03942430e+00 -2.96557814e-01 -9.99620497e-01 -6.86692744e-02
4.04728889e-01 -9.29808021e-02 9.72120985e-02 -1.06764519e+00
-1.36907780e+00 7.79908836e-01 7.68033683e-01 8.29652045e-03
8.05343270e-01 8.77095386e-02 1.62899062e-01 9.70292836e-02
2.49341488e-01 -6.14985585e-01 -1.77050442e-01 2.64899135e-01
9.04734850e-01 -8.10950100e-01 4.63649750e-01 -3.90171230e-01
-1.69603795e-01 1.25920200e+00 4.69153076e-01 -2.91162908e-01
9.79419649e-01 -1.46639675e-01 -4.41333592e-01 -9.57777798e-01
-6.50208950e-01 -3.09608698e-01 6.18440926e-01 7.76795268e-01
7.78603494e-01 1.32638440e-01 -7.52869427e-01 4.25558507e-01
-4.97699678e-01 -2.33769700e-01 1.05789632e-01 3.54845434e-01
1.36702344e-01 -1.18714821e+00 1.80910856e-01 1.01281667e+00
-6.28593564e-01 -2.70830691e-01 -5.63119590e-01 7.93245196e-01
2.91384310e-01 5.82489550e-01 2.29069293e-02 -1.59350961e-01
1.23071782e-01 4.56164271e-01 7.83315718e-01 -6.93158031e-01
-8.30702066e-01 -6.47808835e-02 3.21919501e-01 -7.07543731e-01
-5.37782311e-01 -9.20720994e-01 -1.21778202e+00 -5.76175489e-02
-7.66706914e-02 -3.52567822e-01 7.40520895e-01 1.00682008e+00
6.71867728e-01 7.88451493e-01 4.40457165e-02 -8.54927301e-01
6.08276017e-02 -1.17837965e+00 -4.59238112e-01 4.25290704e-01
-3.58221238e-03 -9.16980386e-01 4.15154696e-02 -1.20286979e-02] | [12.4049711227417, 3.3747920989990234] |
1b5dce28-fb05-4899-9494-b3c5f393ecdb | few-shot-named-entity-recognition-with-joint | null | null | https://openreview.net/forum?id=CvGtHdgukK | https://openreview.net/pdf?id=CvGtHdgukK | Few-shot Named Entity Recognition with Joint Token and Sentence Awareness | Few-shot learning has been proposed and rapidly emerging as a viable means for completing various tasks. Recently, few-shot models have been used for Named Entity Recognition (NER). Prototypical network shows high efficiency on few-shot NER. However, existing prototypical methods only consider the similarity of tokens in query sets and support sets and ignore the semantic similarity among the sentences which contain these entities. We present a novel model, Few-shot Named Entity Recognition with \textbf{J}oint \textbf{T}oken and \textbf{S}entence \textbf{A}wareness (\textbf{JTSA}), to address the issue. The sentence awareness is introduced to probe the semantic similarity among the sentences. The Token awareness is used to explore the similarity of the tokens. To further improve the robustness and results of the model, we adopt the joint learning scheme on the few-shot NER. Experimental results demonstrate that our model outperforms state-of-the-art models on two standard Few-shot NER datasets. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['few-shot-ner'] | ['natural-language-processing'] | [-2.79346079e-01 -2.00259715e-01 4.23460593e-03 -3.97445142e-01
-6.05006397e-01 -1.72518581e-01 5.94503701e-01 3.34283203e-01
-8.21664333e-01 6.20862961e-01 3.79184961e-01 3.33276898e-01
-2.95829594e-01 -9.51607049e-01 -2.04559922e-01 -5.08789539e-01
2.11567342e-01 1.25210986e-01 8.18513751e-01 -4.50815707e-01
3.95957947e-01 1.40101537e-01 -1.44949520e+00 2.96083093e-02
7.82491624e-01 7.53681481e-01 3.83223921e-01 4.23311055e-01
-8.59179676e-01 7.56898224e-01 -7.27528989e-01 -5.35641074e-01
-8.44421387e-02 -3.44618678e-01 -8.43615949e-01 -3.75280708e-01
5.46816830e-03 -1.13661930e-01 -6.48355424e-01 1.35072649e+00
8.64262104e-01 9.57260191e-01 6.38431370e-01 -9.67881322e-01
-6.92889154e-01 6.92696691e-01 -2.81083256e-01 7.27168679e-01
3.01932245e-01 -1.34780243e-01 9.53418553e-01 -1.20329750e+00
7.49801219e-01 1.09982324e+00 5.88982403e-01 6.36268377e-01
-4.03621733e-01 -5.89973330e-01 -6.38617948e-02 5.86713314e-01
-1.49123025e+00 -5.34868658e-01 3.44494671e-01 -1.01738758e-01
1.24295712e+00 4.49991971e-03 1.22049697e-01 9.24431622e-01
1.70528948e-01 6.74344480e-01 7.13586867e-01 -6.36272430e-01
5.67663252e-01 8.73225927e-02 7.57556558e-01 5.16209722e-01
1.30700454e-01 -2.03468189e-01 -4.87194031e-01 -1.74205139e-01
2.02199891e-01 4.49880779e-01 -8.75837058e-02 9.75976810e-02
-9.86242414e-01 7.51517594e-01 2.71069348e-01 9.39031601e-01
-4.39714104e-01 -2.89327919e-01 6.96201622e-01 -3.82108130e-02
2.62825131e-01 2.74171233e-01 -2.46643931e-01 -3.39077145e-01
-7.62538970e-01 -2.58464694e-01 9.11377072e-01 1.24343526e+00
8.78089964e-01 1.66971490e-01 -6.55810654e-01 9.99010205e-01
4.68908176e-02 1.08761877e-01 9.02437031e-01 -5.05902052e-01
4.72351968e-01 4.59245443e-01 1.79149330e-01 -8.45235765e-01
-2.23214492e-01 1.15825869e-01 -6.53819978e-01 -4.31440949e-01
-2.11661458e-01 -4.25418586e-01 -1.16466928e+00 1.29339135e+00
4.61912870e-01 5.67004740e-01 2.40655839e-01 6.76865280e-01
1.24891543e+00 9.31099832e-01 5.88347256e-01 -2.79095501e-01
1.47541010e+00 -8.76877725e-01 -9.59988236e-01 -1.06843978e-01
8.46536040e-01 -7.31479704e-01 7.14846253e-01 -3.03057015e-01
-5.32769263e-01 -5.06326556e-01 -9.35624063e-01 -1.34959919e-02
-9.25158918e-01 -3.15149397e-01 3.37732285e-01 8.16352129e-01
-6.33615613e-01 8.62835169e-01 -5.10810077e-01 -9.15834546e-01
1.39152169e-01 8.26670304e-02 -2.45034084e-01 -3.21220279e-01
-1.82549369e+00 9.51095998e-01 1.03890598e+00 -5.50186038e-02
-6.94577456e-01 -3.98793042e-01 -1.08970773e+00 5.18513203e-01
6.69781327e-01 -2.87657052e-01 1.10552323e+00 -1.38871044e-01
-1.24982071e+00 5.54227769e-01 -9.85657498e-02 -2.28730291e-01
-5.14384024e-02 -3.23081277e-02 -1.06315184e+00 8.86289552e-02
5.09904683e-01 1.97642952e-01 3.45869780e-01 -6.38565004e-01
-6.31708086e-01 -1.92439973e-01 -1.88503131e-01 2.61428475e-01
-8.09545934e-01 6.26620293e-01 -2.49695480e-01 -5.18200040e-01
-7.40799606e-02 -3.04015756e-01 -1.22290000e-01 -5.66929698e-01
-4.20758605e-01 -6.90894306e-01 7.71024108e-01 -1.40263319e-01
1.48989534e+00 -2.38610220e+00 -3.80048841e-01 -1.55370876e-01
5.78860380e-02 6.25654995e-01 -1.49637610e-01 8.74665439e-01
4.87281978e-02 2.35446990e-01 4.60058711e-02 6.73595145e-02
1.05053239e-01 1.62770152e-01 -1.73737392e-01 1.86554864e-01
-3.52247506e-02 6.23497367e-01 -1.11518538e+00 -8.73297870e-01
2.20746815e-01 1.20020166e-01 1.91505760e-01 3.45252544e-01
2.25905165e-01 -2.07807586e-01 -8.48202169e-01 4.67556179e-01
2.75933027e-01 9.30230040e-03 -2.33717099e-01 -1.37147292e-01
-2.69243956e-01 -1.95659157e-02 -1.27995884e+00 1.62042379e+00
-7.71525279e-02 1.38395324e-01 -3.40690583e-01 -8.51275086e-01
1.03163397e+00 6.84621096e-01 2.54310668e-01 -4.59973186e-01
3.87906790e-01 1.40843287e-01 -9.44841132e-02 -7.00310946e-01
6.73920989e-01 -5.73104799e-01 -2.42660075e-01 3.46427232e-01
5.97346842e-01 7.17895180e-02 5.10261774e-01 3.16434771e-01
1.25954604e+00 -1.96823776e-01 7.34997213e-01 -2.71489974e-02
3.75883967e-01 -3.84394862e-02 8.45986605e-01 8.69847536e-01
-7.39724636e-01 3.47118884e-01 6.52269498e-02 -1.95248663e-01
-9.03493404e-01 -8.63651514e-01 1.66040920e-02 1.53255498e+00
4.42588478e-01 -5.09295225e-01 -6.26143992e-01 -7.02593803e-01
-4.67567980e-01 1.20608509e+00 -5.74452341e-01 -4.56504613e-01
-2.22687259e-01 -6.04950368e-01 8.61605167e-01 6.14091754e-01
6.84549987e-01 -1.34944510e+00 -4.83456761e-01 3.10217530e-01
-5.92022128e-02 -9.35431838e-01 -5.24902463e-01 2.47840226e-01
-5.74007034e-01 -7.81747520e-01 -8.82111967e-01 -9.23889697e-01
1.24805070e-01 3.61471593e-01 6.72970831e-01 -2.65775919e-01
-5.59256732e-01 4.13933694e-01 -7.46116042e-01 -5.40097952e-01
4.06500585e-02 -1.29813313e-01 1.73886895e-01 -1.50344864e-01
9.90895748e-01 -3.58891547e-01 -2.27507323e-01 2.57651985e-01
-8.79093587e-01 -6.62395239e-01 4.67224300e-01 8.23472261e-01
4.68864202e-01 2.71911800e-01 7.03298092e-01 -8.67239356e-01
6.79763556e-01 -8.28617573e-01 2.01271027e-01 6.04348242e-01
-5.00428081e-01 -1.01054572e-01 7.42350280e-01 -4.24321800e-01
-1.46300149e+00 -5.25530398e-01 1.79846566e-02 -5.84936559e-01
-4.90170479e-01 5.76526999e-01 -1.35002926e-01 3.32153052e-01
6.57290816e-01 4.75308567e-01 -7.64721870e-01 -5.22343874e-01
3.38722616e-01 9.40099180e-01 3.73312265e-01 -3.48133445e-01
4.10264164e-01 2.27436364e-01 -3.56509835e-01 -1.09298122e+00
-1.09797609e+00 -1.22675776e+00 -5.37001729e-01 -1.04263037e-01
1.11864090e+00 -7.58458793e-01 -2.66009539e-01 1.86529607e-01
-1.30797708e+00 4.96678621e-01 -4.03940558e-01 7.33251572e-01
-3.08146775e-01 6.35098517e-01 -6.58131897e-01 -9.32409585e-01
-5.00392914e-01 -4.02815372e-01 7.34501779e-01 7.98420548e-01
-5.75311184e-02 -8.38931680e-01 3.28484148e-01 1.22390676e-03
2.20592469e-01 -2.50470281e-01 8.86970460e-01 -1.70262992e+00
-5.94889820e-02 -4.83000606e-01 -2.31210172e-01 4.04510759e-02
1.54130191e-01 -2.47808903e-01 -8.84045959e-01 -1.82332769e-02
2.93480963e-01 -3.18695962e-01 7.73667991e-01 -5.69969155e-02
6.48961663e-01 -2.77793985e-02 -4.53159422e-01 4.52664681e-02
1.39300787e+00 3.63333970e-01 6.16741419e-01 2.21289083e-01
6.12097979e-01 5.19194663e-01 8.40939879e-01 7.34611034e-01
2.17052266e-01 1.57698959e-01 -8.82848278e-02 4.14570153e-01
7.45308725e-03 -2.22738847e-01 1.93457007e-01 1.17634976e+00
-1.31837353e-02 -3.69366974e-01 -1.05237913e+00 8.01880777e-01
-1.84715331e+00 -1.42206275e+00 1.57705083e-01 1.92971361e+00
6.24659121e-01 1.63047716e-01 -3.52444977e-01 -4.03267115e-01
1.43123055e+00 4.88219857e-01 -4.65453863e-01 -2.70583510e-01
7.67277816e-05 2.65715569e-01 1.36625797e-01 -1.88109338e-01
-1.13381016e+00 1.14733112e+00 5.49511480e+00 1.28656530e+00
-5.73969722e-01 5.06527841e-01 1.02651842e-01 1.01305865e-01
8.15402940e-02 8.74031484e-02 -1.20555878e+00 5.64664662e-01
1.05194020e+00 -7.12062359e-01 -1.20736599e-01 1.08903205e+00
-1.33093130e-02 -2.51879413e-02 -7.66792119e-01 7.22932994e-01
4.50290591e-01 -1.02824390e+00 -2.18970166e-03 -4.53920096e-01
6.83630288e-01 1.26328155e-01 -6.55501366e-01 9.94265616e-01
2.88239688e-01 -5.21850884e-01 2.73060501e-01 7.52383351e-01
5.82429290e-01 -7.83837140e-01 1.09273458e+00 6.13277912e-01
-1.54856336e+00 -1.04581974e-01 -9.06105220e-01 1.03148371e-01
2.89814085e-01 4.30455953e-01 -6.47413254e-01 8.80767047e-01
8.63449872e-01 6.03824615e-01 -3.00914705e-01 1.19246030e+00
-1.32123381e-01 3.89333606e-01 -8.82405862e-02 -5.89239538e-01
3.82414877e-01 4.07911316e-02 5.90702116e-01 1.32906592e+00
4.22380626e-01 6.18466020e-01 4.30863827e-01 5.72836280e-01
-2.00419217e-01 6.27091169e-01 -7.43227243e-01 -4.80756938e-01
8.04074943e-01 1.23610616e+00 -8.88740480e-01 -7.04518616e-01
-5.84381163e-01 1.05474675e+00 4.75905925e-01 2.03182206e-01
-6.13255620e-01 -1.44957376e+00 2.36010864e-01 -3.56670648e-01
6.66626930e-01 1.04745977e-01 1.05445892e-01 -1.26117992e+00
-2.77106881e-01 -1.88887596e-01 6.81989014e-01 -7.30509639e-01
-1.72253895e+00 4.43722457e-01 -4.59095575e-02 -1.18512046e+00
3.71956155e-02 -2.17966482e-01 -1.15186810e+00 7.10349917e-01
-1.20592618e+00 -6.94456100e-01 -1.02774493e-01 6.87760174e-01
8.58970046e-01 -3.46882582e-01 9.87391293e-01 1.87739298e-01
-7.14277267e-01 2.37555832e-01 3.16341937e-01 5.31189561e-01
8.45324993e-01 -9.44291651e-01 1.60093471e-01 9.76145148e-01
-2.38800589e-02 9.22541499e-01 5.42529643e-01 -1.04680836e+00
-1.03825641e+00 -9.96792495e-01 1.24616396e+00 -4.96938117e-02
8.42237473e-01 1.25164866e-01 -1.02112830e+00 4.85784590e-01
3.38485837e-01 1.45149156e-01 1.04590476e+00 1.09308258e-01
-3.25152785e-01 1.38160661e-01 -1.02424312e+00 5.23920894e-01
8.58619750e-01 -6.00263596e-01 -1.44939125e+00 2.60404468e-01
9.77185667e-01 1.33359969e-01 -9.49008346e-01 2.03654423e-01
1.27196208e-01 -8.60036790e-01 6.78998351e-01 -9.13719833e-01
6.62525445e-02 -1.61318839e-01 -2.21022055e-01 -1.20430529e+00
-4.18552786e-01 -3.35223556e-01 -1.24654673e-01 1.58659768e+00
1.26731908e-02 -3.92890960e-01 3.69648665e-01 7.26666570e-01
-2.03085825e-01 -2.35552847e-01 -1.12044787e+00 -1.14082682e+00
-2.72384703e-01 -9.92610604e-02 4.32166040e-01 1.23760104e+00
4.17980283e-01 5.66098332e-01 -2.83120364e-01 -8.90996754e-02
4.31475997e-01 -1.58033252e-01 1.39631286e-01 -1.30980611e+00
9.11385864e-02 -1.24275550e-01 -5.56445062e-01 -5.69545925e-01
2.65719563e-01 -7.66816020e-01 5.25701106e-01 -1.63436067e+00
6.33015156e-01 4.64424603e-02 -8.22009265e-01 4.67952073e-01
-4.42094177e-01 -2.99703628e-01 2.13938087e-01 2.62311220e-01
-1.15360618e+00 8.15672576e-01 8.13851893e-01 8.49176347e-02
-7.30646700e-02 -3.09503078e-01 -2.77321100e-01 7.58745313e-01
5.41579723e-01 -7.43194282e-01 -2.44955406e-01 -6.02667555e-02
-6.07438162e-02 3.86403836e-02 -1.75013185e-01 -1.13914788e+00
8.95269990e-01 -1.98760822e-01 2.63681203e-01 -4.56709772e-01
2.92094588e-01 -5.66491067e-01 -2.77092665e-01 2.64807642e-01
-2.59504944e-01 -1.33809343e-01 -1.91160709e-01 1.03828990e+00
-3.23958278e-01 -9.37855065e-01 6.53117836e-01 -5.11244535e-01
-1.46628273e+00 5.11637807e-01 -3.77148837e-01 3.93125176e-01
1.22697997e+00 -2.59001583e-01 -3.66787702e-01 -8.52237269e-02
-6.73478961e-01 2.18551084e-01 -1.93180994e-03 4.66066122e-01
8.05379868e-01 -1.23503017e+00 -4.24830407e-01 -2.69579083e-01
5.47504902e-01 -4.60391104e-01 6.10907972e-01 5.53183019e-01
-6.24010805e-03 3.04427713e-01 -8.57919529e-02 2.70787589e-02
-1.01830697e+00 8.13623846e-01 4.42438535e-02 -1.48454055e-01
-5.75765550e-01 8.33873868e-01 -5.90255670e-02 -4.07127500e-01
1.47103012e-01 3.70430589e-01 -6.10704958e-01 3.11522037e-01
7.50219047e-01 8.10545504e-01 -1.84073031e-01 -6.34578764e-01
-4.07846868e-01 2.45324746e-01 -4.20694053e-01 -7.31099350e-03
1.27942002e+00 -2.91893572e-01 -2.19655454e-01 9.49746192e-01
9.85874116e-01 -3.08719367e-01 -5.65190911e-01 -4.58050132e-01
4.26370084e-01 -3.11530322e-01 -3.97324674e-02 -5.42407930e-01
-3.61287504e-01 8.39299083e-01 5.54223657e-01 2.45748967e-01
5.69203317e-01 -8.00760835e-03 1.09913766e+00 9.53158975e-01
6.67682886e-01 -1.49944878e+00 -4.95896898e-02 8.70861709e-01
1.06431492e-01 -1.15395296e+00 -1.06940329e-01 -3.70433211e-01
-7.39411235e-01 1.18060219e+00 7.25984037e-01 3.84088680e-02
7.99581409e-01 -1.82029828e-01 -1.42547876e-01 -2.85555601e-01
-5.48300982e-01 -5.29238820e-01 6.69260770e-02 3.92822832e-01
3.24793786e-01 -1.68355808e-01 -6.71846151e-01 9.00243819e-01
3.38934273e-01 -1.60956604e-03 4.47686404e-01 1.31438565e+00
-1.17828298e+00 -8.19292605e-01 -1.05749264e-01 4.19541717e-01
-4.80336845e-01 -4.32373941e-01 -1.97482526e-01 4.62201446e-01
9.49807167e-02 1.14522171e+00 -1.53197870e-01 -2.39366174e-01
3.66686672e-01 6.21905088e-01 -3.66598777e-02 -1.25743425e+00
-6.31204426e-01 -1.31242618e-01 7.88833499e-02 -2.17586845e-01
-3.33555937e-01 -3.66646856e-01 -1.57633746e+00 -6.81248158e-02
-6.99283004e-01 5.44433057e-01 3.63338143e-01 1.27970743e+00
3.90191346e-01 3.57686371e-01 6.37190402e-01 -3.56948376e-01
-8.00686479e-01 -1.15322828e+00 -1.07152605e+00 4.52444315e-01
-3.59061569e-01 -6.41627550e-01 -3.01809996e-01 -2.46099845e-01] | [9.713356971740723, 9.344440460205078] |
46e4f35c-1b15-47b3-b4ce-93495d2b7460 | dynamical-models-for-metabolomics-data | 2105.10365 | null | https://arxiv.org/abs/2105.10365v1 | https://arxiv.org/pdf/2105.10365v1.pdf | Dynamical models for metabolomics data integration | As metabolomics datasets are becoming larger and more complex, there is an increasing need for model-based data integration and analysis to optimally leverage these data. Dynamical models of metabolism allow for the integration of heterogeneous data and the analysis of dynamical phenotypes. Here, we review recent efforts in using dynamical metabolic models for data integration, focusing on approaches that are not restricted to steady-state measurements or that require flux distributions as inputs. Furthermore, we discuss recent advances and current challenges. We conclude that much progress has been made in various areas, such as the development of scalable simulation tools, and that, although challenges remain, dynamical modeling is a powerful tool for metabolomics data analysis that is not yet living up to its full potential. | ['Daniel Weindl', 'Polina Lakrisenko'] | 2021-05-21 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 9.54531431e-02 -6.60386622e-01 -1.30171105e-01 -8.09169561e-02
5.80756217e-02 -9.10076678e-01 3.93394232e-01 6.82745576e-01
2.27568373e-02 6.93524301e-01 -1.19194351e-02 -2.25768149e-01
-4.15305197e-01 -4.70056474e-01 -2.61486858e-01 -7.65748143e-01
-5.16648471e-01 4.41238195e-01 -4.24088957e-03 -9.15102139e-02
-2.09263772e-01 7.22618341e-01 -1.22746861e+00 -7.74050727e-02
7.44665623e-01 6.30468249e-01 2.17075586e-01 9.15459156e-01
4.07364368e-02 3.20623994e-01 -3.81375849e-01 1.84221849e-01
1.33532435e-01 -9.44050133e-01 -4.79209214e-01 -9.45892632e-02
-3.99583071e-01 4.12445873e-01 -1.49683714e-01 6.50172770e-01
6.96543813e-01 2.70840406e-01 4.59492326e-01 -1.02752197e+00
-3.21469486e-01 2.82979548e-01 -2.12946564e-01 5.02453089e-01
2.32758537e-01 5.04488945e-01 1.01861648e-01 -3.44595432e-01
7.65875161e-01 1.09694302e+00 7.25361049e-01 2.19689459e-01
-1.66047287e+00 -2.93742210e-01 2.05352038e-01 5.77470325e-02
-1.51444364e+00 -6.03057921e-01 3.30812365e-01 -7.49762356e-01
1.33810735e+00 5.41995108e-01 1.27273583e+00 8.04107368e-01
6.41813278e-01 -2.86348369e-02 1.19591081e+00 -1.69763789e-01
5.21445334e-01 -6.89102635e-02 2.18731221e-02 3.78230155e-01
5.59942961e-01 1.04675323e-01 -5.68915784e-01 -1.83315888e-01
7.20984638e-01 -4.98595387e-02 -1.51982114e-01 -6.37185037e-01
-1.36649847e+00 4.96771038e-01 -1.59821317e-01 3.48688215e-01
-5.81119776e-01 1.72133252e-01 4.43644881e-01 3.32556456e-01
4.99698400e-01 6.60550177e-01 -9.73476410e-01 -2.29230940e-01
-8.39205205e-01 3.77384037e-01 1.01283395e+00 6.36288583e-01
2.82281011e-01 -2.43868139e-02 3.08373153e-01 4.90002960e-01
4.11779732e-01 3.58593971e-01 3.14621478e-01 -1.09577942e+00
-2.66480237e-01 6.60630405e-01 2.03721032e-01 -6.75412774e-01
-8.39678764e-01 -2.19707370e-01 -8.51276040e-01 7.74174929e-02
4.75238979e-01 -3.93417150e-01 -7.19752252e-01 1.77460754e+00
5.71950972e-01 5.89256808e-02 2.94217557e-01 6.49097919e-01
4.02672917e-01 6.66407645e-01 5.18820584e-01 -7.35623300e-01
1.11429739e+00 -3.98682863e-01 -8.60415876e-01 2.69551277e-01
2.87126809e-01 -7.03763127e-01 3.41328233e-01 3.41444135e-01
-1.20184517e+00 -9.87186134e-02 -8.52792144e-01 8.86343196e-02
-8.70023012e-01 -6.07940555e-01 8.18713784e-01 5.94723523e-01
-9.57902908e-01 8.80570114e-01 -1.69789243e+00 -1.10622847e+00
3.06906216e-02 4.93886709e-01 -1.37364939e-01 2.57769287e-01
-9.93772507e-01 1.09744000e+00 4.03709471e-01 -4.10916284e-02
-9.06402469e-01 -1.19668770e+00 -7.18062282e-01 3.62331867e-02
-8.04088190e-02 -1.06496334e+00 9.99977767e-01 -3.09697121e-01
-1.81847560e+00 1.33147135e-01 -2.08704636e-01 -2.52753168e-01
4.52763736e-01 4.91346359e-01 -5.01361787e-01 2.15356108e-02
-3.17121804e-01 3.54394585e-01 -2.11927131e-01 -8.30807030e-01
4.71055321e-02 -3.67319643e-01 -3.30202729e-01 2.91236103e-01
7.60370959e-03 3.21114212e-01 2.00042143e-01 -8.79675001e-02
2.06189319e-01 -1.01255023e+00 -5.56270778e-01 1.47636279e-01
-1.66869253e-01 2.98858970e-01 6.09614491e-01 -4.19251233e-01
9.00589287e-01 -1.65337837e+00 6.24293864e-01 -1.19978584e-01
9.27596074e-03 2.17784747e-01 8.52886513e-02 1.09557688e+00
-1.91191375e-01 4.84535843e-01 -1.87425360e-01 2.64736600e-02
-4.84290868e-02 -1.25208303e-01 3.40634249e-02 5.78679800e-01
3.15842777e-01 9.82256770e-01 -1.10086083e+00 9.06160325e-02
6.97035134e-01 5.39040565e-01 -1.97192371e-01 -2.68506743e-02
-3.79464775e-01 1.01704085e+00 -2.18995988e-01 6.37114763e-01
4.92900193e-01 -4.04418886e-01 7.91279554e-01 -3.51626961e-03
-8.39935422e-01 -1.83231737e-02 -1.18910539e+00 1.67494798e+00
-1.51839674e-01 2.17329577e-01 2.35599205e-01 -7.55450845e-01
4.89382982e-01 5.17574728e-01 1.15717208e+00 -2.46979654e-01
7.17669800e-02 2.26874545e-01 5.49998939e-01 -5.87665498e-01
-9.39424485e-02 -4.52353746e-01 1.83232531e-01 3.43162119e-01
1.83361903e-01 -3.46545845e-01 6.05616987e-01 -2.58205533e-01
9.92627800e-01 4.06449139e-01 6.43980205e-01 -8.44807506e-01
4.07402605e-01 5.19250214e-01 5.33255100e-01 1.26229465e-01
-2.50505626e-01 1.45406604e-01 4.74438518e-01 -2.28975281e-01
-1.04569685e+00 -8.40241432e-01 -4.59868908e-01 5.81649721e-01
2.93200433e-01 -5.45197368e-01 -5.51131904e-01 1.96335390e-01
2.20543489e-01 1.34595811e-01 -4.54426199e-01 -6.04147278e-02
-1.97540194e-01 -1.70512986e+00 3.78381044e-01 1.91803694e-01
6.01063529e-03 -5.30865133e-01 -5.60804248e-01 6.28968418e-01
1.79440513e-01 -8.27288747e-01 9.86908227e-02 6.30501390e-01
-1.13648999e+00 -1.48429871e+00 -4.55027252e-01 -2.56714880e-01
3.85083020e-01 3.00569773e-01 8.62075567e-01 -2.65752614e-01
-6.91978693e-01 2.49967620e-01 -5.39684482e-02 -6.18509650e-01
-5.25744498e-01 -5.10962605e-02 3.62091213e-01 -6.44951522e-01
1.36786386e-01 -8.36325586e-01 -9.44971979e-01 4.19545561e-01
-8.14545393e-01 1.12115890e-02 2.46316895e-01 5.27338922e-01
8.72927845e-01 5.29613756e-02 8.69195342e-01 -4.86089975e-01
5.79805493e-01 -8.94257009e-01 -7.03077495e-01 2.02681422e-01
-8.17820907e-01 -1.93530828e-01 7.63022900e-01 -4.02531713e-01
-8.94241631e-01 1.35407016e-01 1.38597950e-01 -4.91254292e-02
-2.07040399e-01 9.22694981e-01 -1.52748236e-02 -1.28381342e-01
5.62708020e-01 1.11973360e-01 5.14948606e-01 -6.25039339e-01
2.31110930e-01 1.72327429e-01 -1.96637079e-01 -5.02215385e-01
2.03846216e-01 3.49776655e-01 6.00699186e-01 -1.08944452e+00
6.89551532e-02 -2.90550977e-01 -6.95132196e-01 -9.26985815e-02
1.00839078e+00 -7.39676356e-01 -9.86102521e-01 6.43462479e-01
-4.91654396e-01 -7.86391199e-01 -3.30230236e-01 7.14849114e-01
-6.22593462e-01 2.79216170e-01 -7.64968395e-01 -8.10552061e-01
-1.79970145e-01 -1.28180420e+00 4.64859158e-01 2.82796949e-01
-4.48775083e-01 -1.33807600e+00 5.88618517e-01 -2.74919599e-01
8.23150337e-01 8.03445756e-01 1.02740264e+00 -5.36192596e-01
-5.79606712e-01 -2.41288245e-02 3.08717459e-01 -3.01042140e-01
3.59315634e-01 5.15782773e-01 -7.36362219e-01 -3.31071138e-01
1.00120120e-01 -1.30717263e-01 2.89819062e-01 6.78998113e-01
7.79422224e-01 2.55687535e-01 -6.80794716e-01 7.61224866e-01
1.50262868e+00 6.14223957e-01 1.84240371e-01 -3.15636061e-02
1.92715839e-01 6.48570716e-01 2.43417576e-01 3.08527410e-01
3.69446874e-01 5.84568381e-01 2.04486474e-01 -2.10884169e-01
-1.31933019e-01 4.51166052e-05 1.04586467e-01 1.00362551e+00
-2.33484149e-01 -5.59225738e-01 -7.10751116e-01 1.79173015e-02
-1.81253707e+00 -1.03628838e+00 -5.08814216e-01 2.30671215e+00
6.45268917e-01 -5.80390036e-01 3.07454765e-01 -2.17166066e-01
5.55466056e-01 -2.18256786e-01 -1.14269257e+00 -3.50766689e-01
-1.13896705e-01 4.06213626e-02 3.49518955e-01 4.18000221e-01
-8.12134862e-01 5.77331901e-01 8.72443867e+00 -2.62678951e-01
-1.37298810e+00 9.71969217e-02 5.34181595e-01 -1.96296051e-01
1.39015168e-01 1.10532880e-01 -4.47417468e-01 5.48674047e-01
1.52146292e+00 -8.49351645e-01 7.62191057e-01 2.87932009e-01
9.67222214e-01 -4.12950516e-01 -1.11510527e+00 4.57874864e-01
-4.90641057e-01 -1.13091910e+00 -4.18976694e-01 4.30387706e-01
5.21628737e-01 2.15381324e-01 -3.14453900e-01 -7.66710490e-02
4.23866630e-01 -7.55993009e-01 4.07430977e-01 7.42528677e-01
5.21357477e-01 -4.73326176e-01 4.23803389e-01 2.95450658e-01
-1.33891189e+00 1.08744070e-01 -6.04739904e-01 -3.78338158e-01
4.20693606e-01 7.75512040e-01 -4.94706154e-01 7.93811798e-01
3.34900647e-01 8.98772776e-01 -3.31231833e-01 1.36643839e+00
4.43967789e-01 2.86827594e-01 -4.49375659e-01 3.43603157e-02
-5.50002337e-01 -6.34034872e-01 2.94854462e-01 1.17965388e+00
6.90381229e-01 1.64907854e-02 1.78915486e-01 9.43887830e-01
3.50113600e-01 5.34843132e-02 -5.93281925e-01 -6.68988764e-01
3.50246876e-01 1.36998296e+00 -1.04289114e+00 -3.18690062e-01
-3.57349306e-01 6.19535983e-01 -6.44437298e-02 2.19220176e-01
-6.65042758e-01 -1.28773076e-03 1.42423093e+00 -6.99152723e-02
-2.66114950e-01 -6.50184035e-01 3.52363251e-02 -1.21157157e+00
-6.37010515e-01 -5.13476074e-01 4.53366131e-01 -6.47449493e-01
-1.12206304e+00 1.31104022e-01 3.97581995e-01 -7.04785049e-01
-2.85219729e-01 -7.30511308e-01 -3.99985284e-01 1.00190890e+00
-1.30824673e+00 -5.37304223e-01 -1.88915804e-01 2.69211959e-02
3.60875785e-01 4.24883515e-01 1.11142910e+00 2.05066770e-01
-1.06687617e+00 -3.40275139e-01 5.50739348e-01 -8.22378337e-01
4.98026788e-01 -1.22058880e+00 3.08034867e-01 7.23759174e-01
-4.46505934e-01 9.96145964e-01 9.50227380e-01 -8.77931714e-01
-2.04830337e+00 -1.04295325e+00 2.35818163e-01 -6.11667454e-01
8.50945711e-01 -4.58209217e-01 -7.21849859e-01 6.03864551e-01
2.98740387e-01 -1.22873172e-01 1.28303349e+00 -1.03037022e-01
3.83482844e-01 1.06470704e-01 -1.11781824e+00 4.11078185e-01
1.02076590e+00 -1.02152094e-01 8.60104263e-02 5.77022195e-01
2.77302593e-01 -4.83087480e-01 -1.60738397e+00 2.48021305e-01
6.76472008e-01 -5.69724083e-01 9.85127091e-01 -7.36981750e-01
-1.31467715e-01 -5.62900662e-01 1.07178211e-01 -1.58924079e+00
-4.83392328e-01 -7.92120218e-01 -2.85395861e-01 1.01234615e+00
3.04773241e-01 -8.00693274e-01 2.62254328e-01 8.08953762e-01
-2.96029866e-01 -4.69036013e-01 -6.10868335e-01 -8.74070644e-01
2.85957992e-01 1.10880964e-01 7.51413405e-01 1.02549970e+00
7.04639077e-01 6.77889735e-02 1.40951529e-01 -1.39191924e-02
3.05895746e-01 1.14378355e-01 3.09024036e-01 -1.30120564e+00
-1.24145202e-01 -5.05438149e-01 -5.15115559e-01 -3.55098009e-01
-2.53012925e-01 -9.63633776e-01 -1.76836297e-01 -1.71975601e+00
3.67275357e-01 -1.16142012e-01 -4.79123980e-01 1.18423276e-01
1.78097766e-02 1.87472329e-01 -2.02575568e-02 3.75080369e-02
-2.96996713e-01 1.81733161e-01 1.29970133e+00 2.10794285e-01
-4.81858969e-01 -4.28439558e-01 -7.40736485e-01 2.97097296e-01
9.57029581e-01 -2.77207226e-01 -5.67088068e-01 7.32538477e-03
2.16600060e-01 2.94887006e-01 1.20362185e-01 -9.52908456e-01
-2.67614555e-02 -6.49330199e-01 7.96123981e-01 -1.52257964e-01
2.78536677e-01 -5.75827539e-01 1.35315073e+00 8.35268140e-01
-1.26255304e-01 4.89868283e-01 4.82467532e-01 6.57231331e-01
2.86797494e-01 1.48506820e-01 8.29646587e-01 -4.60740983e-01
-1.57808051e-01 3.91163617e-01 -9.55144644e-01 -1.62533149e-01
1.20793211e+00 5.52410260e-02 -5.89143217e-01 5.19187897e-02
-1.14210105e+00 1.27694964e-01 1.01721585e+00 1.05470501e-01
-2.00886041e-01 -9.89049673e-01 -3.01482499e-01 -6.18917532e-02
-1.74149945e-02 -3.02524716e-01 2.56213635e-01 1.15462148e+00
-7.27464974e-01 6.75337255e-01 -3.36436778e-01 -4.92527097e-01
-9.70148146e-01 9.95128870e-01 6.56494737e-01 -1.73752867e-02
-3.86183769e-01 8.82397667e-02 7.94082880e-02 -2.44996116e-01
-6.72622800e-01 -5.31111121e-01 1.28885314e-01 2.53058504e-02
4.60023642e-01 7.72316635e-01 2.08188951e-01 -4.55896825e-01
-5.42051613e-01 3.54717821e-01 7.18595684e-01 1.47996366e-01
1.37887084e+00 -4.43309903e-01 -5.16315341e-01 8.11707914e-01
6.49037838e-01 -2.18173489e-01 -1.14909315e+00 6.40339077e-01
-1.89071000e-01 -1.04994394e-01 -1.68261260e-01 -1.14917934e+00
-7.22377777e-01 7.38262177e-01 6.35206401e-01 6.82756901e-01
8.47235501e-01 -2.55450636e-01 2.84930348e-01 2.42026597e-01
2.59401590e-01 -7.61104703e-01 -3.25303942e-01 3.16858590e-01
6.26433611e-01 -6.97365046e-01 2.11487338e-01 -6.68983936e-01
-4.76696938e-02 1.11705363e+00 2.90876091e-01 2.21737519e-01
8.50007772e-01 5.13383329e-01 1.59618407e-01 -2.62845099e-01
-1.04650021e+00 -1.77245408e-01 -3.47677410e-01 8.22650731e-01
8.07062268e-01 1.63521498e-01 -6.40794039e-01 1.44155696e-01
2.74962962e-01 3.16206425e-01 5.96480310e-01 1.11612856e+00
-4.03315693e-01 -1.42527831e+00 -2.77907223e-01 5.13991714e-01
-4.14230376e-01 1.69762671e-01 -4.05649126e-01 7.55162179e-01
-1.37510002e-01 1.14756167e+00 -2.11926103e-01 -6.15091398e-02
1.63994312e-01 6.55188084e-01 6.93836987e-01 -6.36062086e-01
-4.88567173e-01 2.48198330e-01 -2.27700472e-02 -4.37548578e-01
-4.06304896e-01 -1.04562974e+00 -1.04068899e+00 -4.72238898e-01
-3.64622712e-01 2.35229656e-01 1.20327461e+00 6.01662755e-01
8.15156400e-01 9.58674014e-01 3.62734437e-01 -9.62034643e-01
-2.34598279e-01 -8.15918267e-01 -6.25969708e-01 -7.44167641e-02
-8.66568647e-03 -8.66082489e-01 -2.81975538e-01 2.42890775e-01] | [5.899750709533691, 4.470616340637207] |
a3904bdb-bdeb-4aac-bc6d-0a6ff07ecdf7 | bridging-clip-and-stylegan-through-latent | 2210.04506 | null | https://arxiv.org/abs/2210.04506v1 | https://arxiv.org/pdf/2210.04506v1.pdf | Bridging CLIP and StyleGAN through Latent Alignment for Image Editing | Text-driven image manipulation is developed since the vision-language model (CLIP) has been proposed. Previous work has adopted CLIP to design a text-image consistency-based objective to address this issue. However, these methods require either test-time optimization or image feature cluster analysis for single-mode manipulation direction. In this paper, we manage to achieve inference-time optimization-free diverse manipulation direction mining by bridging CLIP and StyleGAN through Latent Alignment (CSLA). More specifically, our efforts consist of three parts: 1) a data-free training strategy to train latent mappers to bridge the latent space of CLIP and StyleGAN; 2) for more precise mapping, temporal relative consistency is proposed to address the knowledge distribution bias problem among different latent spaces; 3) to refine the mapped latent in s space, adaptive style mixing is also proposed. With this mapping scheme, we can achieve GAN inversion, text-to-image generation and text-driven image manipulation. Qualitative and quantitative comparisons are made to demonstrate the effectiveness of our method. | ['Zhongyuan Wang', 'Pengfei Wan', 'Xiaoyan Guo', 'Qiang Li', 'Wanfeng Zheng'] | 2022-10-10 | null | null | null | null | ['image-manipulation'] | ['computer-vision'] | [ 6.23898745e-01 -3.05532031e-02 -3.49449128e-01 -3.29729110e-01
-6.58235669e-01 -4.09995824e-01 8.27781796e-01 -7.41057158e-01
9.02052745e-02 4.63375360e-01 2.08928749e-01 2.15607602e-02
-2.35562846e-01 -6.79598570e-01 -6.51845098e-01 -8.00028622e-01
4.75402117e-01 3.33925873e-01 -1.18482850e-01 -1.02176256e-01
3.68512630e-01 1.73234627e-01 -1.35095799e+00 1.32361099e-01
1.14380801e+00 8.84712994e-01 6.94980860e-01 5.76783955e-01
-3.25909197e-01 8.75616729e-01 -4.18476641e-01 -1.59747630e-01
4.44267571e-01 -9.46094096e-01 -6.20384872e-01 4.67569798e-01
3.27819824e-01 -3.74812901e-01 4.25796509e-02 1.07091081e+00
5.14944971e-01 -1.40896663e-01 6.45627260e-01 -1.38824737e+00
-7.02147782e-01 6.24593616e-01 -7.60427117e-01 -6.21284842e-01
-7.87169032e-04 5.05789280e-01 7.28180051e-01 -8.14742029e-01
7.62347400e-01 1.15596032e+00 2.83438027e-01 5.58510423e-01
-1.14998841e+00 -8.29412103e-01 1.38183117e-01 8.43805149e-02
-1.32602513e+00 -4.76111501e-01 1.15673780e+00 -5.20359814e-01
4.66147512e-01 3.64092082e-01 8.87068093e-01 1.17730522e+00
7.78416768e-02 8.46364439e-01 1.18764782e+00 -7.16407776e-01
1.69150800e-01 2.94836998e-01 -4.84082967e-01 8.14551592e-01
-5.29189967e-02 7.60191306e-02 -7.25000858e-01 3.10549825e-01
1.02300882e+00 -1.58710152e-01 -2.94224322e-01 -3.88263881e-01
-1.36636007e+00 6.18915081e-01 1.44266963e-01 3.65754634e-01
-2.97242910e-01 2.38154441e-01 2.37538204e-01 2.53497213e-01
4.88320410e-01 6.48040652e-01 -2.05890045e-01 -2.87167639e-01
-1.28256023e+00 1.24671973e-01 4.39409137e-01 1.51063716e+00
7.49848723e-01 3.77385288e-01 -5.52708745e-01 6.74852431e-01
4.72677022e-01 6.25791013e-01 5.30964732e-01 -1.24700773e+00
6.28105164e-01 6.48357451e-01 -2.62398303e-01 -9.72645760e-01
1.81152493e-01 -2.92073250e-01 -1.09564745e+00 1.23951569e-01
-3.84262614e-02 -2.34705489e-02 -1.11661029e+00 1.71465552e+00
1.70522675e-01 4.80388924e-02 -1.22074783e-01 7.70511150e-01
3.35957766e-01 6.81799173e-01 -2.39444375e-01 -4.20708716e-01
1.12276161e+00 -1.11095965e+00 -1.08285809e+00 -1.73110202e-01
4.81244922e-01 -8.98745418e-01 1.36821628e+00 1.46631688e-01
-1.35903549e+00 -6.38091087e-01 -1.16093600e+00 -1.60089225e-01
-3.98392454e-02 3.52647424e-01 4.38068658e-01 5.59072018e-01
-8.71942163e-01 1.03123404e-01 -8.14746261e-01 -3.50977004e-01
4.48781610e-01 9.75532159e-02 9.62772779e-03 -2.34553125e-02
-9.27321076e-01 6.61600590e-01 4.79311913e-01 1.90533951e-01
-1.04270542e+00 -7.36408472e-01 -6.99386299e-01 -1.99826047e-01
4.25938934e-01 -9.76924479e-01 8.29918981e-01 -1.22716236e+00
-1.89955068e+00 7.98236012e-01 -2.74758726e-01 -8.10691267e-02
6.96537435e-01 -4.23673168e-02 -6.35334179e-02 -2.55937446e-02
2.82939196e-01 9.64546442e-01 1.19435012e+00 -1.45368671e+00
-5.59300661e-01 -1.77759409e-01 -1.24660999e-01 3.75108629e-01
-4.92088348e-01 -2.08498776e-01 -1.10861850e+00 -1.04594612e+00
1.39583260e-01 -9.29132819e-01 -8.06517527e-02 9.33731422e-02
-5.50933182e-01 7.92148039e-02 8.98534715e-01 -7.91741669e-01
1.34049213e+00 -2.19574261e+00 5.88907063e-01 2.42983714e-01
2.41586074e-01 9.49399695e-02 -1.88682407e-01 2.87665039e-01
2.33224377e-01 1.13306351e-01 -3.96338254e-01 -6.22741401e-01
1.78909376e-02 2.12135926e-01 -4.97786522e-01 2.80338198e-01
2.37573862e-01 1.12285221e+00 -5.89586020e-01 -8.20708334e-01
4.51564014e-01 8.91195238e-02 -6.93386793e-01 4.51475233e-01
-5.52735925e-01 7.85179794e-01 -5.83655715e-01 8.73172879e-01
6.66974306e-01 -2.09740847e-01 2.15570748e-01 -6.12163246e-01
-2.72771984e-01 -2.23031506e-01 -9.13044393e-01 2.28420496e+00
-5.71568906e-01 6.02426171e-01 7.75093511e-02 -8.48935962e-01
1.01048362e+00 3.18386368e-02 4.62095737e-01 -8.66406560e-01
1.33602098e-01 6.11892529e-02 -2.52433717e-01 -6.85468256e-01
6.22892976e-01 1.11308463e-01 7.94051662e-02 3.33221048e-01
-9.84039456e-02 -3.86332631e-01 1.08849376e-01 6.38941303e-02
5.67895055e-01 6.64080322e-01 -1.64842755e-01 -2.85773665e-01
4.74102706e-01 1.69743612e-01 7.31556833e-01 6.92727864e-01
1.44539446e-01 7.32926846e-01 3.50425214e-01 1.60855398e-01
-1.31710494e+00 -9.03004289e-01 1.08132116e-01 6.33120298e-01
4.16526198e-01 -3.93714845e-01 -8.86838913e-01 -3.73838067e-01
-5.06327629e-01 9.01352882e-01 -4.66875106e-01 -1.52801439e-01
-7.32548654e-01 -5.01176298e-01 4.33271945e-01 3.26858968e-01
1.00496709e+00 -8.10727239e-01 -3.58379543e-01 -3.72564420e-02
-3.26106250e-01 -8.47433567e-01 -7.54808426e-01 -1.29747421e-01
-6.87492311e-01 -7.98937857e-01 -7.98354328e-01 -8.74656320e-01
9.07324076e-01 1.96055800e-01 7.91962445e-01 -1.12086676e-01
-2.53226727e-01 4.22396719e-01 -4.41567779e-01 -2.80730903e-01
-3.63724709e-01 2.44270228e-02 -1.58147290e-01 -7.73596950e-03
-1.09256074e-01 -6.09956503e-01 -6.06789112e-01 5.16958714e-01
-1.03124797e+00 8.03430557e-01 8.72818291e-01 1.00720644e+00
6.59820318e-01 1.52682230e-01 1.63865820e-01 -6.05059206e-01
4.36942577e-01 -1.66698202e-01 -8.29155624e-01 7.11187899e-01
-8.92704725e-01 2.02333897e-01 4.24168408e-01 -4.90897924e-01
-1.38333142e+00 -4.64335866e-02 2.01672912e-01 -6.94295526e-01
9.61299166e-02 4.61834908e-01 -6.17157698e-01 9.58176032e-02
3.79924953e-01 7.70790040e-01 2.35877320e-01 -2.37740904e-01
6.92039251e-01 6.28207922e-01 3.82486314e-01 -7.54111350e-01
9.95511711e-01 5.39475083e-01 -3.73357721e-02 -6.41592860e-01
-5.53151250e-01 8.75997096e-02 -5.88873386e-01 -3.51134032e-01
1.27394176e+00 -9.80975807e-01 -3.83848995e-01 5.93278706e-01
-1.04285634e+00 -5.96442461e-01 -2.49050409e-01 4.22047287e-01
-8.03279161e-01 3.81785065e-01 -2.93537736e-01 -6.84771419e-01
-4.29733247e-01 -1.39608836e+00 1.24184334e+00 6.22051135e-02
-4.17372026e-02 -8.82393658e-01 -1.06354013e-01 5.56945026e-01
5.60084939e-01 1.57831267e-01 8.24624360e-01 1.82366565e-01
-1.09684217e+00 1.91980377e-01 -2.88148671e-01 3.94721329e-01
3.79630744e-01 -5.41712064e-03 -7.35220134e-01 -4.28408235e-01
1.91006228e-01 -3.39091212e-01 5.49429655e-01 4.91023540e-01
1.43094540e+00 -4.19980794e-01 -2.70810962e-01 1.11135614e+00
1.28592861e+00 3.39100420e-01 7.44116664e-01 4.12501603e-01
9.81310368e-01 5.85207820e-01 9.49633837e-01 2.46250629e-01
3.59840095e-01 8.60368192e-01 1.83622137e-01 -2.96397805e-01
-4.26721007e-01 -5.72572351e-01 5.62168658e-01 1.03581583e+00
1.10130971e-02 -3.23749363e-01 -5.56453824e-01 3.82209897e-01
-2.01435137e+00 -7.62147307e-01 1.21142387e-01 1.89929378e+00
9.28856850e-01 4.55260724e-02 -3.10704976e-01 -1.04762144e-01
6.21953785e-01 1.67014346e-01 -5.33434212e-01 1.48851112e-01
-1.20308049e-01 -1.14186577e-01 4.97112602e-01 3.36290896e-01
-7.29969621e-01 1.04126465e+00 5.73539591e+00 1.27962410e+00
-1.38323307e+00 1.47840917e-01 7.35755682e-01 -1.27978593e-01
-6.83764458e-01 2.17632815e-01 -5.96700370e-01 6.36476934e-01
2.13955179e-01 -5.32349162e-02 5.80902159e-01 6.35870695e-01
3.78668576e-01 -1.17673330e-01 -9.09487605e-01 1.23387349e+00
1.90450177e-01 -1.51123512e+00 3.47170442e-01 1.33765265e-01
9.61840808e-01 -4.86982882e-01 1.36956155e-01 1.34599820e-01
3.82243022e-02 -8.58772814e-01 1.04993999e+00 7.65327990e-01
1.22989154e+00 -5.43098390e-01 1.56337306e-01 3.36540669e-01
-1.16804051e+00 8.86652712e-03 -1.26961678e-01 3.28064084e-01
3.54954571e-01 6.06330156e-01 -5.55097103e-01 8.47980917e-01
4.67675865e-01 7.64425397e-01 -4.11417067e-01 3.74923497e-01
-3.98126483e-01 5.90598702e-01 6.01993687e-02 2.52949059e-01
5.40181883e-02 -5.48958182e-01 6.49687529e-01 8.26814353e-01
4.90595490e-01 -2.24898383e-01 1.79132402e-01 1.49174035e+00
2.42216662e-01 -2.72310555e-01 -5.52612841e-01 -2.99251407e-01
5.01313150e-01 1.18872893e+00 -6.19211495e-01 -2.70018458e-01
-9.69662145e-02 1.25133526e+00 -8.46856087e-02 4.54058111e-01
-1.15578055e+00 -2.28030905e-01 3.76579285e-01 6.56257421e-02
2.62107879e-01 -5.45051455e-01 -6.70699060e-01 -1.20457017e+00
1.45435542e-01 -9.21121001e-01 -2.39998683e-01 -8.43906224e-01
-9.94534612e-01 3.60108435e-01 2.30233803e-01 -1.31175840e+00
-2.93442458e-01 -1.49933636e-01 -6.00445867e-01 1.02596593e+00
-1.47087514e+00 -1.75442851e+00 -5.18376887e-01 6.25899971e-01
7.99661696e-01 -4.35179144e-01 5.01620352e-01 4.02491987e-01
-7.67883658e-01 7.68842518e-01 -6.34692162e-02 -7.37791210e-02
7.23507226e-01 -9.57207501e-01 2.10960031e-01 1.10739553e+00
-9.74853635e-02 5.76532125e-01 5.20856977e-01 -8.71981442e-01
-1.83088171e+00 -1.28859162e+00 4.29493725e-01 -2.63827056e-01
3.67202312e-01 -5.58247268e-01 -4.53990430e-01 6.09238207e-01
3.29268754e-01 -5.36123395e-01 3.55427802e-01 -2.33242199e-01
-7.72043988e-02 -2.98707843e-01 -8.46718550e-01 8.26505661e-01
1.22681832e+00 -5.31563103e-01 -2.15704858e-01 1.33955196e-01
8.02313745e-01 -4.00825649e-01 -7.28123009e-01 4.91904110e-01
4.75274622e-01 -6.99644685e-01 6.55770004e-01 -1.51589466e-02
6.87237501e-01 -7.17150807e-01 -1.48372903e-01 -9.98287261e-01
-3.32900703e-01 -7.43565857e-01 -3.25505175e-02 1.61881161e+00
1.88134596e-01 -2.66828686e-01 7.37625003e-01 3.33977014e-01
-2.28442594e-01 -6.19727433e-01 -5.39991856e-01 -7.15475798e-01
-2.90497214e-01 -4.42017823e-01 5.93388975e-01 1.01621127e+00
-5.16255736e-01 2.11394921e-01 -7.54755795e-01 -3.42753157e-02
6.09069169e-01 3.16189438e-01 1.08334458e+00 -4.36806023e-01
-4.45297837e-01 -5.46134412e-01 -6.21783771e-02 -1.33722472e+00
-4.93023992e-02 -7.57726192e-01 1.09236419e-01 -1.40410268e+00
2.79713064e-01 -7.63619423e-01 2.08480015e-01 4.44294095e-01
-7.68534765e-02 1.52114853e-01 2.52578467e-01 4.82295305e-01
-3.52406144e-01 9.04650092e-01 1.72641981e+00 -3.42088282e-01
-2.05989257e-01 -4.43072140e-01 -6.73942268e-01 2.55570590e-01
7.06586838e-01 -1.97514296e-01 -8.81721258e-01 -7.53405631e-01
2.64523745e-01 1.63634121e-01 3.12163591e-01 -7.86054432e-01
2.16831148e-01 -5.47260225e-01 2.46470422e-01 -6.64076447e-01
2.55391866e-01 -7.99124122e-01 5.03121555e-01 3.92030656e-01
-3.89518231e-01 -2.99961060e-01 -1.79951653e-01 4.78017926e-01
-2.53342718e-01 1.87870435e-04 7.44859815e-01 -1.07670575e-01
-6.28564715e-01 3.46854657e-01 -1.12153627e-01 -2.52741724e-01
1.14419341e+00 -4.24462378e-01 -1.06820755e-01 -2.80369967e-01
-2.11110562e-01 3.45249295e-01 6.87603414e-01 5.02978027e-01
5.91458738e-01 -1.47585845e+00 -6.52365983e-01 3.40214550e-01
1.56375915e-01 3.68193775e-01 2.38520443e-01 9.52330470e-01
-5.79628706e-01 2.97871351e-01 -2.38062069e-01 -8.85939121e-01
-1.07547009e+00 3.36286068e-01 1.43626288e-01 -2.50372887e-01
-4.67972308e-01 7.82331049e-01 3.58095258e-01 -2.68398583e-01
2.12321818e-01 -8.69790912e-02 1.90087289e-01 -1.32330447e-01
2.37635955e-01 2.36653045e-01 -3.52887630e-01 -2.04000130e-01
4.83830497e-02 6.81321979e-01 -3.18185836e-02 -3.77708912e-01
1.08983994e+00 -4.03568089e-01 -4.24533874e-01 3.86593878e-01
1.00875032e+00 -1.04358971e-01 -1.39228547e+00 -8.20954621e-04
-2.26340488e-01 -6.57579601e-01 1.25992239e-01 -6.66694880e-01
-1.11586869e+00 6.80681050e-01 7.37525403e-01 -2.52381265e-01
1.46853697e+00 -2.00684935e-01 8.14561248e-01 -6.72718212e-02
1.69403747e-01 -1.23825562e+00 2.64480084e-01 2.86127955e-01
9.95810747e-01 -1.08721733e+00 -3.01553234e-02 -3.30014944e-01
-6.32104754e-01 9.44500864e-01 8.61839652e-01 3.24081928e-01
3.81991863e-01 2.46544391e-01 -8.91684964e-02 -1.89214706e-01
-6.09966755e-01 -1.86026664e-04 4.34908688e-01 4.25779074e-01
2.01940775e-01 -8.35250691e-02 -3.43577564e-01 1.74347252e-01
-1.41947985e-01 1.10252656e-01 1.25314891e-01 1.09201503e+00
-1.21956669e-01 -1.35458398e+00 -2.05101773e-01 2.53778160e-01
1.86085328e-01 -1.04370974e-01 -1.54201984e-01 4.90646571e-01
3.69554669e-01 7.07256913e-01 -7.36819673e-03 -4.97175485e-01
-6.82009980e-02 -1.56171277e-01 7.02900648e-01 -3.68255794e-01
-1.14887148e-01 3.99771124e-01 -2.34473273e-01 -6.37394786e-01
-6.55319870e-01 -3.87292325e-01 -7.61968970e-01 -1.79523945e-01
-5.57344139e-01 -1.40320495e-01 8.96034241e-01 1.03613341e+00
5.00710964e-01 6.86223984e-01 8.50010753e-01 -7.29409933e-01
-2.74068683e-01 -8.17836940e-01 -4.89560723e-01 2.10648090e-01
2.42855381e-02 -3.82254392e-01 -1.16694562e-01 3.83481503e-01] | [11.452862739562988, -0.6104161739349365] |
2209ba79-73e5-43d3-939a-1673c5fbcab4 | approximate-knowledge-graph-query-answering | 2102.11389 | null | https://arxiv.org/abs/2102.11389v1 | https://arxiv.org/pdf/2102.11389v1.pdf | Approximate Knowledge Graph Query Answering: From Ranking to Binary Classification | Large, heterogeneous datasets are characterized by missing or even erroneous information. This is more evident when they are the product of community effort or automatic fact extraction methods from external sources, such as text. A special case of the aforementioned phenomenon can be seen in knowledge graphs, where this mostly appears in the form of missing or incorrect edges and nodes. Structured querying on such incomplete graphs will result in incomplete sets of answers, even if the correct entities exist in the graph, since one or more edges needed to match the pattern are missing. To overcome this problem, several algorithms for approximate structured query answering have been proposed. Inspired by modern Information Retrieval metrics, these algorithms produce a ranking of all entities in the graph, and their performance is further evaluated based on how high in this ranking the correct answers appear. In this work we take a critical look at this way of evaluation. We argue that performing a ranking-based evaluation is not sufficient to assess methods for complex query answering. To solve this, we introduce Message Passing Query Boxes (MPQB), which takes binary classification metrics back into use and shows the effect this has on the recently proposed query embedding method MPQE. | ['Michael Cochez', 'Dimitrios Alivanistos', 'Daniel Daza', 'Teodor Aleksiev', 'Ruud van Bakel'] | 2021-02-22 | null | null | null | null | ['complex-query-answering'] | ['knowledge-base'] | [ 1.61308590e-02 5.19202650e-01 -1.89771995e-01 -2.21325308e-01
-5.65245211e-01 -6.75098479e-01 7.57623672e-01 1.25030506e+00
-4.22635466e-01 9.06439126e-01 3.49337012e-01 -9.55826715e-02
-6.64664507e-01 -1.43605936e+00 -5.80779612e-01 -2.14365065e-01
-1.25515506e-01 8.66260350e-01 6.62892163e-01 -4.85470206e-01
2.90364504e-01 3.70250463e-01 -1.69621754e+00 4.15384918e-01
7.97632813e-01 8.11962962e-01 -2.71489114e-01 3.41319084e-01
-7.00021386e-01 9.96049404e-01 -9.03300226e-01 -1.00124288e+00
-1.61359444e-01 -1.79640174e-01 -1.11594713e+00 -4.44017977e-01
4.13375080e-01 -9.97490212e-02 -2.77307302e-01 1.27498662e+00
2.91239530e-01 -2.92478144e-01 6.03151381e-01 -1.42466033e+00
-1.92348480e-01 8.25744748e-01 -5.23360074e-02 6.08677827e-02
9.20930207e-01 -4.74547505e-01 1.41583133e+00 -1.03602028e+00
1.08455050e+00 1.19862497e+00 6.30995572e-01 -5.41663766e-02
-1.38078320e+00 -6.32644519e-02 -2.36214995e-01 5.14079809e-01
-1.52062631e+00 -7.61367232e-02 7.39461005e-01 -3.29555720e-01
7.86347210e-01 5.57273149e-01 4.90521044e-01 7.24874914e-01
-1.27993196e-01 3.68038476e-01 9.93191779e-01 -4.93559390e-01
4.38003272e-01 5.87503672e-01 4.58416343e-01 5.70148826e-01
9.58263457e-01 -1.19622998e-01 -5.56371331e-01 -5.36587656e-01
5.19667342e-02 -4.33925539e-02 -4.59827036e-01 -5.50292790e-01
-1.08819282e+00 8.46201718e-01 6.12856507e-01 7.96297371e-01
-4.55424368e-01 -1.54644912e-02 4.34741944e-01 7.14247227e-01
2.92180181e-01 6.59298718e-01 -2.96994865e-01 1.23932585e-01
-8.95818233e-01 3.44152331e-01 1.25352240e+00 6.66884840e-01
1.22804666e+00 -6.23836160e-01 -1.52526617e-01 5.87609231e-01
1.28901199e-01 1.87766571e-02 2.26910949e-01 -8.08045089e-01
4.86743867e-01 1.26794767e+00 3.44882131e-01 -1.76216245e+00
-3.30089539e-01 -5.34132600e-01 -6.57643139e-01 1.24847069e-02
5.22338688e-01 4.01470423e-01 -2.80043721e-01 1.60008705e+00
4.49505061e-01 -4.67171997e-01 7.08847716e-02 7.37793684e-01
7.80673206e-01 3.74211818e-01 -5.56643754e-02 -4.72088419e-02
1.33909702e+00 -4.23410624e-01 -8.54981244e-01 1.66883886e-01
7.96141803e-01 -5.98450005e-01 7.48007238e-01 3.62926692e-01
-7.33034074e-01 -2.55261004e-01 -1.02038884e+00 1.55355623e-02
-1.07980704e+00 -1.05458498e-01 3.26770842e-01 7.41580129e-01
-1.07006490e+00 7.25497842e-01 -3.65667373e-01 -4.86417800e-01
2.86547877e-02 1.87453419e-01 -5.17650425e-01 -3.19030166e-01
-1.51786900e+00 1.02364314e+00 6.09824955e-01 -1.48391640e-02
-1.41663954e-01 -4.75488067e-01 -6.56709671e-01 3.23909700e-01
6.97140694e-01 -7.59963393e-01 7.03900576e-01 -7.00552642e-01
-5.48888266e-01 9.20425832e-01 -7.95274526e-02 -4.52066511e-01
6.28724396e-01 1.62897184e-01 -4.76047903e-01 3.94051343e-01
1.67865977e-01 2.89012611e-01 5.64822674e-01 -1.35347950e+00
-3.55658501e-01 -7.22260058e-01 4.27259982e-01 -3.63733947e-01
-3.60198736e-01 -7.94215277e-02 -1.45399377e-01 -4.46643412e-01
2.36487463e-01 -5.62010288e-01 3.85635570e-02 -2.16848142e-02
-3.66167456e-01 -5.48004150e-01 6.42564178e-01 -5.78004122e-01
1.65571272e+00 -1.67885780e+00 1.11670487e-01 5.63669264e-01
7.05817223e-01 1.63026512e-01 1.85549006e-01 1.21435106e+00
5.36097810e-02 4.30058509e-01 -2.05596298e-01 -3.73023674e-02
3.50302994e-01 3.31070274e-01 -3.40167522e-01 1.53303906e-01
1.85464621e-01 7.92281866e-01 -9.70093608e-01 -7.01278031e-01
-7.25696236e-02 3.98227781e-01 -4.21778500e-01 -1.50876969e-01
-3.22303772e-01 -1.41989037e-01 -5.21753073e-01 6.12779617e-01
5.65259635e-01 -3.94596696e-01 2.23286375e-01 -4.40158159e-01
1.49672419e-01 3.54155213e-01 -1.46523035e+00 1.26758397e+00
-1.34671077e-01 4.12104636e-01 -1.26488894e-01 -1.15316224e+00
8.26534033e-01 3.83744419e-01 4.77755070e-01 -7.27923810e-01
-1.77774593e-01 6.33522570e-01 -3.47676218e-01 -5.07112563e-01
9.65445876e-01 4.92958762e-02 -2.98474431e-02 2.59428918e-01
-5.72745949e-02 1.05672985e-01 6.08594775e-01 6.05797946e-01
1.45724535e+00 -3.31881464e-01 2.55559832e-01 -8.18959773e-02
7.39357293e-01 3.68762285e-01 1.56921983e-01 9.02566493e-01
1.02109827e-01 5.13711333e-01 9.58212852e-01 -3.88148785e-01
-9.27438676e-01 -7.57723451e-01 2.39001550e-02 5.93793154e-01
1.51837066e-01 -9.54762042e-01 -4.99966741e-01 -8.41296971e-01
2.11841226e-01 5.30702829e-01 -6.60560668e-01 -3.30006063e-01
-2.55225569e-01 -3.90438765e-01 4.42615658e-01 8.91489908e-02
1.83658808e-01 -9.44968343e-01 -6.22904301e-01 3.25939566e-01
-4.77628738e-01 -9.23392475e-01 1.37884438e-01 -4.98428233e-02
-9.08958554e-01 -1.40669906e+00 -5.72488129e-01 -3.11169803e-01
5.62445343e-01 -5.94842769e-02 1.64851105e+00 4.14714724e-01
-9.63651836e-02 6.91220284e-01 -7.29225993e-01 -1.55172229e-01
-5.37414551e-01 9.46953669e-02 -3.00022870e-01 1.53728381e-01
4.49794739e-01 -4.90497023e-01 -6.41644180e-01 3.07219297e-01
-1.47306967e+00 -5.86793184e-01 5.13455749e-01 7.30101168e-01
4.54207778e-01 1.24742389e-01 7.31443763e-01 -1.17665565e+00
8.13510358e-01 -5.99273682e-01 -6.49141252e-01 5.63606203e-01
-9.10527945e-01 5.37004113e-01 5.17265499e-01 -1.14122652e-01
-4.85779554e-01 -3.68356049e-01 -2.37827674e-01 -1.41776234e-01
8.66881087e-02 1.09890962e+00 5.06374575e-02 -1.02869660e-01
8.20361197e-01 6.43418878e-02 -3.34808677e-01 -4.83294308e-01
3.71193618e-01 4.56697941e-01 1.14820011e-01 -2.66252488e-01
8.40592921e-01 5.24572134e-01 1.17621124e-01 -6.81374192e-01
-5.55829525e-01 -7.14850426e-01 -2.75188297e-01 -2.00748444e-01
4.47123110e-01 -4.21317101e-01 -5.22087872e-01 -8.56667683e-02
-1.28534245e+00 4.48037386e-01 -7.42012799e-01 2.91018873e-01
-1.37809649e-01 5.61453879e-01 -2.18198463e-01 -8.07524145e-01
1.96125843e-02 -7.99271822e-01 9.43223894e-01 -1.92087859e-01
-3.75162154e-01 -9.11545396e-01 3.17204654e-01 1.45419583e-01
6.85390532e-01 2.83752531e-01 1.33062971e+00 -1.00503254e+00
-7.36015081e-01 -6.94413602e-01 -2.80627787e-01 1.16982773e-01
-1.33035094e-01 -3.13283622e-01 -7.26463437e-01 -2.36442685e-01
-2.08467364e-01 -2.06531599e-01 8.14282954e-01 -1.96771383e-01
5.85047066e-01 -4.23098058e-01 -3.91570002e-01 -2.25555643e-01
1.85896587e+00 -4.79101390e-01 7.33913720e-01 2.92223155e-01
3.53907257e-01 1.10455823e+00 4.16555852e-01 4.05607671e-01
3.83938819e-01 8.11632752e-01 6.31667197e-01 2.46863902e-01
-4.46965806e-02 -5.96481740e-01 5.24238572e-02 5.67382157e-01
2.04493582e-01 -4.10046369e-01 -7.63051689e-01 6.16609335e-01
-1.56191158e+00 -1.08700585e+00 -4.42421794e-01 2.61336660e+00
6.63434803e-01 1.52851135e-01 8.89191553e-02 7.02798903e-01
6.28991783e-01 2.03981280e-01 2.04383731e-02 -1.82793349e-01
-2.08668932e-01 1.99389964e-01 3.16442102e-01 4.86544281e-01
-4.50093508e-01 3.97598714e-01 5.33236933e+00 7.79625535e-01
-7.39835560e-01 3.11108436e-02 2.09464118e-01 5.39136887e-01
-8.82140219e-01 4.60445762e-01 -3.83601516e-01 3.93131405e-01
8.33402514e-01 -2.12222308e-01 8.89599323e-02 5.64634085e-01
-2.43624330e-01 -3.77830923e-01 -1.10178292e+00 9.23093021e-01
8.06885064e-02 -1.23479378e+00 2.62292385e-01 9.95456502e-02
2.71955431e-01 -1.78494081e-01 -3.89877915e-01 3.25583100e-01
9.86888446e-03 -8.78109038e-01 5.16506791e-01 6.87910557e-01
4.50412005e-01 -4.62894320e-01 9.58909392e-01 2.78869301e-01
-1.08652174e+00 -2.06103595e-03 -3.70869398e-01 1.11639939e-01
-5.36963809e-03 9.41991329e-01 -7.50006020e-01 8.80765617e-01
4.32051361e-01 1.86367393e-01 -8.83207798e-01 1.27868748e+00
-8.30508769e-02 4.30742532e-01 -5.15840828e-01 -3.32269788e-01
-1.09182455e-01 -1.64828762e-01 7.64850974e-01 1.01185727e+00
3.60971510e-01 -2.52176136e-01 -1.73254952e-01 9.48715270e-01
-2.75380671e-01 4.03663665e-01 -8.78723562e-01 -1.94345444e-01
4.46380347e-01 1.15429497e+00 -7.48347044e-01 -5.10818243e-01
-5.69436371e-01 8.57002318e-01 3.73719603e-01 3.78984481e-01
-4.60587740e-01 -6.12858057e-01 1.96605295e-01 3.79714698e-01
1.42857909e-01 -1.34891346e-02 1.60570294e-01 -1.16680038e+00
5.34014165e-01 -7.02110350e-01 6.69106126e-01 -6.34410501e-01
-1.29481053e+00 5.75710833e-01 1.50113530e-03 -1.08467519e+00
-2.80319929e-01 -3.39429110e-01 -2.95965225e-02 5.37028730e-01
-1.56486094e+00 -6.54283464e-01 -3.90251726e-01 4.46047843e-01
-1.44790336e-01 2.45639220e-01 8.85872602e-01 7.71390319e-01
-5.51319271e-02 3.19730729e-01 -1.13360703e-01 2.76319589e-02
5.84639311e-01 -1.52214742e+00 -1.04499407e-01 4.76348281e-01
7.90071905e-01 4.93398935e-01 1.07462430e+00 -5.29005766e-01
-1.39853227e+00 -7.21889913e-01 1.72636890e+00 -7.47769833e-01
7.49981701e-01 -9.46557894e-02 -1.13161755e+00 3.44603837e-01
2.95289271e-02 1.01522677e-01 3.94010335e-01 1.98582169e-02
-6.82600260e-01 -1.81326196e-01 -1.16503882e+00 2.83968687e-01
8.47833216e-01 -6.53585076e-01 -8.74470353e-01 3.64825547e-01
5.96776903e-01 1.73523590e-01 -9.26444292e-01 5.32068431e-01
3.11788052e-01 -1.34901071e+00 8.23476374e-01 -5.39776027e-01
2.40615159e-01 -4.12170857e-01 -1.71578854e-01 -1.02791202e+00
1.96895897e-01 -3.55974585e-01 -1.07789874e-01 1.14254940e+00
6.81217611e-01 -8.01961601e-01 8.94327521e-01 3.99410963e-01
3.22824717e-01 -6.34208918e-01 -1.11280298e+00 -6.98417425e-01
-3.07749599e-01 -2.39272892e-01 6.11631274e-01 9.61113691e-01
1.23539343e-01 2.51350641e-01 5.29812761e-02 -3.25927250e-02
4.82012302e-01 3.18341702e-01 6.76880836e-01 -1.68729913e+00
-6.49814680e-02 -3.51514250e-01 -9.03408468e-01 -5.64965427e-01
-2.06928238e-01 -9.74637270e-01 -5.53284466e-01 -1.75138831e+00
5.00040129e-02 -5.05125582e-01 -1.82652339e-01 1.35253116e-01
5.03202900e-02 2.04693809e-01 -6.13824725e-02 2.62851745e-01
-8.47142696e-01 4.48018819e-01 6.94757938e-01 -3.04474890e-01
1.48646817e-01 -1.28706962e-01 -4.07816321e-01 4.90883946e-01
5.32917202e-01 -9.15955722e-01 -1.83320984e-01 -7.10873380e-02
1.31741023e+00 2.16365620e-01 6.62560344e-01 -9.30863202e-01
5.52744627e-01 4.38733578e-01 -1.87490180e-01 -5.35131752e-01
2.86356509e-01 -1.28017485e+00 3.44404638e-01 5.74090183e-01
-2.60547876e-01 3.57575625e-01 -4.14970368e-01 8.39856625e-01
-8.58321726e-01 -5.24348617e-01 1.33346960e-01 -1.77915558e-01
-4.42396373e-01 -2.54257005e-02 -7.51913488e-02 2.87202209e-01
4.89944667e-01 -2.35004961e-01 -4.05920446e-01 -5.53513527e-01
-7.38853753e-01 1.76507056e-01 4.50721174e-01 1.32422045e-01
6.65847242e-01 -1.22073281e+00 -6.80313706e-01 -2.51408458e-01
6.21472597e-01 -3.68128240e-01 -4.63863797e-02 1.07596195e+00
-5.94054639e-01 4.76884604e-01 2.12865442e-01 -3.55559498e-01
-1.01691663e+00 6.59656882e-01 2.66042143e-01 -5.86814821e-01
-3.32572311e-01 3.88420671e-01 -5.01537919e-01 -4.66673821e-01
1.60240844e-01 -4.37111557e-01 -4.45545882e-01 6.18219495e-01
2.01005980e-01 5.16575933e-01 4.60085481e-01 -5.26552141e-01
-3.69165987e-01 4.36384797e-01 1.57736987e-01 -1.11111306e-01
9.95132565e-01 -1.36143997e-01 -5.23986995e-01 4.55470353e-01
1.21967936e+00 5.86611629e-01 -1.49651602e-01 -3.64960343e-01
5.84870040e-01 -4.51929480e-01 -3.16997111e-01 -6.74535096e-01
-7.13846922e-01 6.14100099e-01 3.72676760e-01 1.07269061e+00
9.41334367e-01 2.23881096e-01 5.99088013e-01 6.96695030e-01
6.94584012e-01 -1.00077021e+00 -5.87364659e-02 6.04181662e-02
8.96896839e-01 -1.30728841e+00 9.68741067e-03 -5.63291252e-01
-6.09710626e-02 1.05891407e+00 -1.93389156e-03 1.97477285e-02
4.98144656e-01 -2.27310807e-01 -2.61015832e-01 -7.22495854e-01
-5.14250159e-01 -3.32813501e-01 1.96238756e-01 2.92027861e-01
2.36737296e-01 -1.87236935e-01 -8.95730257e-01 3.42490882e-01
-1.42405048e-01 1.72355045e-02 4.32171494e-01 8.73484790e-01
-4.06462133e-01 -1.39879739e+00 -3.53882104e-01 5.43701768e-01
-6.18629873e-01 2.74943132e-02 -8.32483411e-01 7.62178540e-01
-1.13130175e-01 1.05232775e+00 -3.73210341e-01 -2.86630809e-01
4.88512844e-01 2.70685226e-01 3.05883557e-01 -3.85480374e-01
-6.33066297e-01 -6.79319620e-01 3.23315442e-01 -6.46590292e-01
-4.50400978e-01 -1.60898924e-01 -8.81795228e-01 -5.15409559e-02
-5.40355206e-01 6.74445271e-01 6.52206361e-01 6.22520626e-01
3.76806229e-01 3.14073622e-01 3.81779671e-01 -1.11558102e-01
-7.24887967e-01 -7.21260905e-01 -5.45436323e-01 7.94503748e-01
2.47511089e-01 -6.19557023e-01 -6.67474747e-01 -4.06029403e-01] | [9.19325065612793, 7.972320556640625] |
af26e7c6-6aad-457f-aaa6-e3f3a849ade3 | inferencing-based-on-unsupervised-learning-of | 1803.02627 | null | http://arxiv.org/abs/1803.02627v1 | http://arxiv.org/pdf/1803.02627v1.pdf | Inferencing Based on Unsupervised Learning of Disentangled Representations | Combining Generative Adversarial Networks (GANs) with encoders that learn to
encode data points has shown promising results in learning data representations
in an unsupervised way. We propose a framework that combines an encoder and a
generator to learn disentangled representations which encode meaningful
information about the data distribution without the need for any labels. While
current approaches focus mostly on the generative aspects of GANs, our
framework can be used to perform inference on both real and generated data
points. Experiments on several data sets show that the encoder learns
interpretable, disentangled representations which encode descriptive properties
and can be used to sample images that exhibit specific characteristics. | ['Tobias Hinz', 'Stefan Wermter'] | 2018-03-07 | null | null | null | null | ['unsupervised-image-classification', 'unsupervised-mnist'] | ['computer-vision', 'methodology'] | [ 3.63742262e-01 7.22896278e-01 -3.28499794e-01 -5.98257482e-01
-8.38520348e-01 -7.60115683e-01 1.01726890e+00 -2.96540201e-01
2.26774439e-01 1.00283909e+00 4.13488746e-01 -3.20813544e-02
3.58149186e-02 -1.19999278e+00 -8.71612668e-01 -1.03584123e+00
-2.48991866e-02 9.11745191e-01 -5.66836834e-01 -1.35051429e-01
-1.52207032e-01 4.63729680e-01 -1.39595735e+00 2.92892814e-01
5.66066563e-01 5.19624829e-01 -3.63155514e-01 8.74577105e-01
2.39165962e-01 9.69898283e-01 -8.46555889e-01 -5.98424196e-01
3.54192138e-01 -6.99857771e-01 -5.07294357e-01 6.94662854e-02
2.85073191e-01 -3.99631053e-01 -5.19212484e-01 9.41113949e-01
3.38379800e-01 -1.75443247e-01 1.14233661e+00 -1.54813981e+00
-1.26331770e+00 6.02547646e-01 -1.79037556e-01 -1.82597563e-01
1.34050637e-01 1.61668733e-01 1.22481692e+00 -4.97385234e-01
5.67522526e-01 1.23503327e+00 3.79894584e-01 9.54846084e-01
-1.81670022e+00 -6.55188918e-01 -4.09373105e-01 -1.25081450e-01
-8.95189643e-01 -3.20036054e-01 1.02733827e+00 -5.07371306e-01
3.45652014e-01 2.99541324e-01 6.90554142e-01 1.75239384e+00
-5.42437285e-02 9.07608509e-01 1.14948821e+00 -2.32412159e-01
4.66046244e-01 5.50681688e-02 -2.13708013e-01 6.38023674e-01
6.27022684e-01 4.60972667e-01 -4.25348669e-01 -1.57139853e-01
8.86865258e-01 1.21411949e-01 -1.59096092e-01 -7.00628996e-01
-1.09461081e+00 1.44652891e+00 6.46224618e-01 -4.47727256e-02
-4.03861403e-01 6.28901124e-01 3.34341586e-01 2.42737398e-01
5.70468962e-01 8.28082800e-01 -1.60368845e-01 -7.60992765e-02
-6.67693138e-01 3.42460334e-01 6.69413149e-01 9.95421886e-01
9.71631706e-01 4.53762621e-01 -2.39176318e-01 3.40533048e-01
1.50873765e-01 7.94804275e-01 4.41747218e-01 -9.16694045e-01
2.02874124e-01 5.56006074e-01 -1.80613145e-01 -6.35816932e-01
3.32989007e-01 -2.16765419e-01 -9.46688652e-01 5.67809999e-01
7.25018978e-02 -2.63179302e-01 -1.10251939e+00 1.94892240e+00
-1.98386759e-01 2.75558770e-01 4.06703323e-01 5.60820878e-01
7.40257084e-01 5.88615119e-01 -1.41713843e-01 4.24097806e-01
9.82205510e-01 -4.50774014e-01 -6.59412324e-01 -2.30476223e-02
2.27085084e-01 -1.65135130e-01 9.21208322e-01 2.26245657e-01
-1.15679812e+00 -3.48161787e-01 -1.17515147e+00 -2.38666296e-01
-4.20783758e-01 7.48633221e-02 8.17622483e-01 7.92520344e-01
-9.36445296e-01 3.57484579e-01 -8.85528684e-01 1.44642532e-01
1.08910120e+00 3.42178822e-01 -5.68589926e-01 -2.72391234e-02
-8.40314269e-01 5.97797036e-01 5.18977821e-01 -2.96572864e-01
-1.45261252e+00 -7.11044312e-01 -1.06917226e+00 1.46213502e-01
1.00376178e-02 -1.15744042e+00 9.97296810e-01 -9.78579104e-01
-1.52545524e+00 8.20673645e-01 1.16765663e-01 -5.73486805e-01
3.09768677e-01 -4.21430692e-02 -9.95450988e-02 8.67710486e-02
-1.33219704e-01 6.84075296e-01 1.15928996e+00 -1.55974984e+00
3.05040833e-02 -3.10934454e-01 2.47124776e-01 -8.38228539e-02
-1.56505704e-01 -5.68269134e-01 4.33803469e-01 -9.34727728e-01
-2.51460135e-01 -9.53342378e-01 -2.07268730e-01 3.18619944e-02
-8.96533251e-01 -5.15542813e-02 8.61739278e-01 -3.29204053e-01
4.85679656e-01 -1.85498786e+00 6.19548738e-01 2.11107675e-02
5.06841362e-01 1.26211867e-01 -2.01110825e-01 5.44762671e-01
-3.21504772e-01 3.79256666e-01 -2.74648875e-01 -4.75691915e-01
1.68942004e-01 7.53044069e-01 -6.92994833e-01 3.37804586e-01
6.40750945e-01 1.38175797e+00 -1.07477868e+00 -4.55737449e-02
1.21628061e-01 6.05038881e-01 -8.38792980e-01 8.59384835e-01
-4.32882160e-01 6.89021111e-01 -2.78287679e-01 1.55146107e-01
4.71827537e-01 -1.48508728e-01 8.86047035e-02 -3.89633849e-02
6.53914630e-01 4.01404679e-01 -4.97863024e-01 1.69185460e+00
-5.97534180e-01 1.04566967e+00 -3.43337893e-01 -1.22581828e+00
1.09172893e+00 4.20759141e-01 1.11753821e-01 -1.32738233e-01
6.59337267e-02 -2.16051236e-01 -1.04079418e-01 -5.54893911e-01
2.51352698e-01 -5.31604648e-01 -2.68466324e-01 8.41739058e-01
4.91448224e-01 -5.12498498e-01 -2.58894637e-02 3.23265761e-01
1.00298321e+00 2.07489952e-01 3.13756496e-01 -8.35864544e-02
1.26517832e-01 -3.63339603e-01 3.22092235e-01 5.40008783e-01
6.40725434e-01 9.29532826e-01 8.13770652e-01 -5.45764267e-01
-1.34397316e+00 -1.65507138e+00 1.63394764e-01 6.60035133e-01
-3.51142615e-01 -4.30214822e-01 -6.41378045e-01 -9.05355871e-01
-7.54752532e-02 1.07424724e+00 -1.32075012e+00 -5.20259440e-01
-2.78809249e-01 -5.09540915e-01 7.18594491e-01 8.50858271e-01
1.31017014e-01 -1.07489109e+00 -5.25329649e-01 -3.60438883e-01
1.49851784e-01 -8.23875248e-01 -3.93375456e-02 3.17322612e-01
-8.10370207e-01 -1.09586072e+00 -4.54931587e-01 -3.73786867e-01
1.23682165e+00 -2.03735322e-01 1.39635384e+00 -3.30132782e-01
-2.19029397e-01 4.13153589e-01 -2.79655457e-01 -7.03880727e-01
-9.92561877e-01 1.84133410e-01 -2.82737553e-01 2.00400385e-03
2.74043471e-01 -9.63123500e-01 -3.93398315e-01 -4.52734120e-02
-1.20879352e+00 2.35135227e-01 6.13564909e-01 1.10427439e+00
2.77976900e-01 -2.05895185e-01 7.28061855e-01 -1.56665266e+00
7.83268750e-01 -7.39145458e-01 -3.64242584e-01 1.46633610e-01
-4.96462762e-01 7.98150003e-01 8.09303403e-01 -4.13583338e-01
-9.34779227e-01 -4.03431132e-02 -1.30768614e-02 -6.04254425e-01
-6.10916078e-01 2.49555483e-01 -4.41255122e-01 3.12175542e-01
8.34826350e-01 2.80066103e-01 3.22235912e-01 -3.77582043e-01
9.88750160e-01 3.21506083e-01 6.78401530e-01 -8.48975420e-01
1.20771873e+00 4.89878148e-01 2.68974841e-01 -2.86610335e-01
-7.56760299e-01 4.52492625e-01 -7.20656455e-01 3.26065987e-01
9.12795842e-01 -8.34419310e-01 -5.22011817e-01 -3.39167602e-02
-9.53903079e-01 -1.66655466e-01 -1.24062240e+00 1.84246778e-01
-1.04165339e+00 -2.21455157e-01 -2.89400548e-01 -5.53798139e-01
-1.45976484e-01 -9.38996255e-01 1.18050420e+00 6.14436157e-02
-4.47808266e-01 -1.41934037e+00 3.57727408e-01 1.34680152e-01
3.37834537e-01 7.95913696e-01 1.27954149e+00 -9.03356194e-01
-9.05341327e-01 -4.64242786e-01 8.74225348e-02 5.33542991e-01
5.02479613e-01 -2.04380862e-02 -1.19831026e+00 -2.49490872e-01
-1.28943741e-01 -7.89323211e-01 7.77397573e-01 8.77210647e-02
1.46981776e+00 -9.32364464e-01 4.49436270e-02 8.56595695e-01
1.22408581e+00 -1.36920244e-01 8.98715079e-01 -3.43115747e-01
7.71661460e-01 3.82366240e-01 -3.25459480e-01 3.13524812e-01
1.52313322e-01 3.90527934e-01 8.13039601e-01 -1.07364424e-01
-2.75255650e-01 -8.85417640e-01 2.91108429e-01 5.18799424e-01
-3.71916085e-01 -4.59528834e-01 -4.97570723e-01 4.21176136e-01
-1.50490141e+00 -1.19599974e+00 2.24484384e-01 1.96986568e+00
8.96403372e-01 -1.28013775e-01 6.15897216e-02 1.32836804e-01
3.08240801e-01 3.92106533e-01 -7.54714370e-01 -5.17054915e-01
-5.62065244e-02 7.64001906e-01 1.70135498e-01 2.45946229e-01
-7.10215390e-01 5.06224096e-01 7.66709900e+00 4.86853242e-01
-9.35426712e-01 -8.60725567e-02 5.70547163e-01 -1.18024997e-01
-1.16332889e+00 1.03104278e-01 -5.97532913e-02 3.32826257e-01
8.65547836e-01 -4.48218703e-01 6.67562306e-01 7.99507260e-01
-3.04259777e-01 4.13333505e-01 -1.73889852e+00 1.01075411e+00
1.62412778e-01 -1.55179822e+00 6.08344376e-01 4.17187333e-01
1.00791252e+00 -2.93276280e-01 6.84020877e-01 6.73020035e-02
1.02728176e+00 -1.73484111e+00 3.98627341e-01 8.41226578e-01
9.70366955e-01 -7.53711998e-01 5.70035756e-01 4.19350415e-01
-3.53794873e-01 2.41679415e-01 -3.72271866e-01 -7.72828534e-02
-2.04359874e-01 4.59835112e-01 -1.18794310e+00 6.02720380e-01
2.40102541e-02 9.58052635e-01 -6.62334800e-01 3.64303499e-01
-8.57923448e-01 7.83654928e-01 8.49144161e-02 -7.98287336e-04
-9.67176557e-02 -3.93160403e-01 4.73916411e-01 7.80694485e-01
3.49308908e-01 -1.82532370e-01 -2.62061536e-01 1.58818662e+00
-4.46333200e-01 -5.44113159e-01 -1.16797650e+00 -5.05835712e-01
1.86540559e-01 1.11878812e+00 -3.24751019e-01 -3.42526406e-01
-1.34892896e-01 8.78262043e-01 5.31152904e-01 5.96023500e-01
-5.83100438e-01 -7.04437420e-02 9.16283131e-01 -6.03026769e-04
2.17620432e-01 -2.75222838e-01 -4.28827494e-01 -1.39664853e+00
-3.43857557e-01 -9.34724748e-01 1.32270038e-01 -8.53252053e-01
-1.68984842e+00 6.29717708e-01 2.14517325e-01 -1.21339405e+00
-1.03437614e+00 -6.33139372e-01 -8.29031885e-01 9.21214700e-01
-9.83404577e-01 -1.36656678e+00 -2.22384542e-01 5.22663414e-01
3.23206514e-01 -5.59311032e-01 1.37845981e+00 -3.79924148e-01
-3.37011069e-01 5.98621666e-01 1.12600356e-01 3.69702131e-01
1.49370819e-01 -1.71071005e+00 4.99614149e-01 5.78365505e-01
7.75093138e-01 7.54565954e-01 7.32058644e-01 -2.48684779e-01
-1.17061412e+00 -9.47778761e-01 2.10281372e-01 -7.48699129e-01
1.82686478e-01 -7.94294834e-01 -5.96955180e-01 1.07166421e+00
6.94673836e-01 3.60552251e-01 1.17793334e+00 1.26448378e-01
-6.89874291e-01 1.56618699e-01 -1.22550762e+00 3.43992531e-01
8.00361454e-01 -8.04132402e-01 -6.48794949e-01 1.73677817e-01
5.65932512e-01 -3.56135160e-01 -7.77485669e-01 1.54190511e-01
3.94799769e-01 -9.62184310e-01 1.03046691e+00 -1.25894094e+00
8.49626124e-01 -1.62033772e-03 -1.12750925e-01 -1.85629332e+00
-2.02423245e-01 -5.93366802e-01 -4.83265191e-01 1.17463136e+00
2.76893765e-01 -6.94616139e-01 8.85330021e-01 6.81744576e-01
2.28959143e-01 -7.65317202e-01 -6.89199746e-01 -5.92810929e-01
2.81980574e-01 -1.76637724e-01 1.07465374e+00 8.50840628e-01
-3.74306977e-01 6.03378236e-01 -5.77183127e-01 8.61435430e-04
6.38673425e-01 2.90923238e-01 1.05706465e+00 -1.26515353e+00
-3.40126336e-01 -1.24536812e-01 -7.95059681e-01 -8.39780271e-01
5.23168981e-01 -1.27505374e+00 -2.31945246e-01 -1.31738353e+00
2.64147043e-01 -4.04881984e-01 -1.21344663e-02 5.29597342e-01
-7.03560412e-02 3.30456883e-01 2.11824966e-03 -2.78962404e-02
-9.81112719e-02 1.01550055e+00 1.17384982e+00 -2.13882402e-01
4.21044081e-01 1.65933922e-01 -9.56326306e-01 4.73216563e-01
6.82611883e-01 -6.50701940e-01 -9.65225339e-01 -4.53511834e-01
4.66153741e-01 -1.72061503e-01 7.72409499e-01 -7.03071177e-01
-3.40710729e-01 -1.75315619e-01 6.40036345e-01 -1.14578485e-01
3.73795271e-01 -8.23754728e-01 3.70220095e-01 2.58318096e-01
-8.95384371e-01 -1.06385164e-01 -1.95178375e-01 7.23062754e-01
-2.61037380e-01 -4.11904454e-01 5.66553652e-01 -1.56875685e-01
-3.06465514e-02 5.46835780e-01 -1.60126299e-01 3.13775957e-01
1.12843370e+00 1.27508655e-01 -3.22874993e-01 -9.47428882e-01
-7.96455443e-01 -2.35949814e-01 6.48145974e-01 4.28502411e-01
7.08634615e-01 -1.72081435e+00 -9.12306607e-01 6.57455683e-01
9.23967659e-02 2.86290556e-01 -2.11475968e-01 -2.16307119e-01
-3.52120817e-01 7.24308789e-02 -5.08513510e-01 -3.58596474e-01
-6.28483593e-01 6.88807070e-01 1.87209457e-01 -1.51352093e-01
-5.29492855e-01 6.09512091e-01 3.38403106e-01 -3.58208954e-01
-3.90535504e-01 -2.11694166e-01 1.03908181e-01 -7.38063604e-02
2.75335163e-01 1.41904876e-01 -3.27012599e-01 -5.76378047e-01
1.60526335e-01 2.61464920e-02 8.96262228e-02 -1.68966562e-01
1.71635163e+00 3.73141438e-01 -1.79635342e-02 5.88106871e-01
1.57156825e+00 -2.45710812e-03 -1.54610038e+00 -3.19141187e-02
-4.26262617e-01 -7.03600526e-01 -3.09821665e-01 -6.50135338e-01
-1.30599177e+00 1.32644629e+00 3.31130594e-01 5.41504622e-01
1.01247740e+00 2.73130238e-01 4.45129573e-01 1.75559282e-01
2.40131378e-01 -2.44111940e-01 5.17185152e-01 2.95616295e-02
1.12938821e+00 -1.03389573e+00 -1.02845021e-01 -2.94510037e-01
-6.50762618e-01 1.33053803e+00 3.97554904e-01 -6.75092995e-01
4.41975743e-01 4.29165572e-01 -2.46773213e-01 -2.73270667e-01
-9.01970088e-01 -7.59556890e-02 5.35346627e-01 1.22572720e+00
3.95559609e-01 3.07606936e-01 3.02802324e-01 3.71794343e-01
-6.22614443e-01 -1.20816223e-01 6.44736707e-01 5.50881326e-01
2.65891314e-01 -1.37694085e+00 1.05237044e-01 5.81920445e-01
-1.70669138e-01 6.08933009e-02 -5.09403825e-01 6.72244906e-01
1.68245450e-01 3.59303892e-01 2.56650269e-01 -3.73841643e-01
-9.90258306e-02 2.29859382e-01 6.90565825e-01 -8.99700642e-01
-1.15945652e-01 -4.88472641e-01 -1.34402081e-01 -3.42165649e-01
-5.78712046e-01 -6.23118997e-01 -9.14370120e-01 -1.54585809e-01
-3.00367307e-02 1.64363295e-01 4.17800486e-01 7.54017413e-01
3.55779201e-01 6.81378186e-01 7.69740164e-01 -6.97049618e-01
-6.25818551e-01 -8.97921920e-01 -5.04295886e-01 8.48289669e-01
6.07192636e-01 -4.08220917e-01 -2.25265965e-01 3.67654622e-01] | [11.5714750289917, -0.05522846058011055] |
30315405-a653-4df6-b44a-02d75a7c57d8 | a-literature-review-on-length-of-stay | 2201.00005 | null | https://arxiv.org/abs/2201.00005v1 | https://arxiv.org/pdf/2201.00005v1.pdf | A Literature Review on Length of Stay Prediction for Stroke Patients using Machine Learning and Statistical Approaches | Hospital length of stay (LOS) is one of the most essential healthcare metrics that reflects the hospital quality of service and helps improve hospital scheduling and management. LOS prediction helps in cost management because patients who remain in hospitals usually do so in hospital units where resources are severely limited. In this study, we reviewed papers on LOS prediction using machine learning and statistical approaches. Our literature review considers research studies that focus on LOS prediction for stroke patients. Some of the surveyed studies revealed that authors reached contradicting conclusions. For example, the age of the patient was considered an important predictor of LOS for stroke patients in some studies, while other studies concluded that age was not a significant factor. Therefore, additional research is required in this domain to further understand the predictors of LOS for stroke patients. | ['Ayman Alahmar', 'Ola Alkhatib'] | 2021-12-30 | null | null | null | null | ['length-of-stay-prediction'] | ['medical'] | [-7.34448314e-01 -6.03465736e-01 -8.27542007e-01 -3.33604634e-01
-4.42513645e-01 -2.07834944e-01 -2.09769607e-01 9.12950099e-01
-8.12403321e-01 8.12431157e-01 7.91123807e-01 -6.63092136e-01
-4.19614792e-01 -6.38067842e-01 9.65139419e-02 -6.48039758e-01
-1.12496108e-01 5.58296800e-01 -2.25065514e-01 1.49705485e-01
4.93039936e-01 1.01291847e+00 -1.04482388e+00 3.30622286e-01
7.33822405e-01 5.65632045e-01 1.89000010e-01 3.99150223e-01
-1.99998185e-01 1.11782742e+00 -4.90676194e-01 1.02717310e-01
-3.84634244e-03 -6.15686417e-01 -7.29690909e-01 -5.96183419e-01
-5.32252133e-01 -8.45391631e-01 -5.47932327e-01 1.97560146e-01
8.61232936e-01 1.51102869e-02 1.07326698e+00 -1.00496221e+00
-1.44233719e-01 7.79454708e-01 1.21439658e-01 6.77357972e-01
7.51599371e-02 -1.58704400e-01 4.11049485e-01 -5.37944317e-01
1.52532328e-02 7.93136001e-01 4.03093994e-01 2.02212185e-01
-7.67556667e-01 -6.86864853e-01 -3.01877916e-01 5.72004855e-01
-1.05155325e+00 -9.46261585e-02 4.10561353e-01 -5.47427654e-01
7.71395504e-01 1.99954763e-01 1.04889166e+00 9.02658463e-01
9.38698351e-01 9.57964957e-02 1.05473864e+00 -5.16125798e-01
1.50178343e-01 -2.16272566e-02 5.85901201e-01 6.09393902e-02
1.02433503e+00 1.90507799e-01 -1.55665893e-02 -3.89464140e-01
8.01555276e-01 9.95375335e-01 -2.45845228e-01 -2.80729026e-01
-1.20695448e+00 9.54533517e-01 4.80129063e-01 5.70248187e-01
-6.11143470e-01 -3.52438718e-01 7.96158731e-01 4.51738417e-01
-1.16993785e-01 4.14234966e-01 -6.05614066e-01 -6.61286294e-01
-7.44227350e-01 -7.18922019e-02 8.23516428e-01 5.39508760e-01
-9.71965641e-02 -3.19926023e-01 -3.66735488e-01 6.72454476e-01
-1.07733093e-01 4.59600240e-01 6.53652608e-01 -6.38704181e-01
2.96372652e-01 7.70585060e-01 1.74952701e-01 -5.38475215e-01
-7.64213324e-01 -3.09872717e-01 -7.64960229e-01 3.97263803e-02
3.80126387e-01 -4.78746146e-01 -3.73020738e-01 9.66707289e-01
-6.40353382e-01 -3.76185805e-01 2.09186211e-01 7.66168237e-01
9.08440590e-01 2.40675360e-01 4.03677374e-01 -6.91652358e-01
1.32762051e+00 -8.84706259e-01 -6.64401054e-01 -2.63778359e-01
1.20790005e+00 -7.37262726e-01 9.60121393e-01 2.93779820e-01
-9.50203955e-01 7.63671007e-03 -5.21843910e-01 3.99602979e-01
-7.01297298e-02 2.83471853e-01 5.15347660e-01 4.52223659e-01
-4.15084004e-01 4.29453611e-01 -9.20408607e-01 -8.79051328e-01
2.54675388e-01 -4.42754813e-02 -1.21426687e-01 -2.59492159e-01
-8.72970760e-01 1.35112524e+00 2.60139525e-01 -3.30545813e-01
-5.12312233e-01 -4.46843266e-01 -3.80006790e-01 1.01772837e-01
-1.65163070e-01 -8.79815996e-01 9.61902618e-01 -4.23181832e-01
-8.06578040e-01 5.77114165e-01 -4.15671468e-01 -2.65537918e-01
2.58795410e-01 -5.42232871e-01 -2.98824519e-01 2.30527446e-01
-2.02487689e-02 -2.80254155e-01 1.15839161e-01 -8.09357762e-01
-6.56866670e-01 -4.94170010e-01 3.52750272e-02 2.27066070e-01
-2.19741076e-01 2.90516227e-01 5.55606842e-01 -5.90835571e-01
3.78184080e-01 -8.65916908e-01 -6.39453948e-01 -5.00696063e-01
3.94766837e-01 -1.41521752e-01 4.59065586e-01 -6.34304762e-01
1.73545825e+00 -2.00271010e+00 -3.06318671e-01 -1.09162619e-02
1.49310187e-01 1.42744854e-01 4.65268433e-01 9.66521442e-01
-1.86186150e-01 5.34366108e-02 1.13038868e-01 4.35931146e-01
-5.42497039e-01 3.72818172e-01 8.42292756e-02 4.66618448e-01
-3.74439180e-01 7.14749515e-01 -6.73911214e-01 -6.59736276e-01
6.79672182e-01 3.15481722e-01 -3.79876256e-01 6.48974240e-01
8.02240610e-01 5.96899092e-01 -5.27642965e-01 3.18186432e-01
3.25974405e-01 -5.04478998e-02 -8.68139639e-02 3.49398628e-02
-4.03929263e-01 2.34824255e-01 -5.84609687e-01 1.14182973e+00
-2.88947374e-01 4.76128399e-01 -6.22489810e-01 -1.36034620e+00
8.27694118e-01 4.97921169e-01 9.08641100e-01 -4.43710417e-01
7.23796606e-01 2.14075401e-01 2.69432217e-01 -1.07569695e+00
-3.32182884e-01 -1.47677258e-01 1.03072025e-01 4.05590087e-01
-9.19812024e-01 8.87614265e-02 8.59806538e-02 1.00317225e-01
9.65219617e-01 -6.02504790e-01 8.82866621e-01 -2.47459456e-01
3.45369548e-01 1.12411745e-01 5.73084950e-01 9.41477537e-01
-5.87609231e-01 4.41815197e-01 5.78019977e-01 -5.96223891e-01
-8.53848755e-01 -9.25886810e-01 -4.58390862e-01 6.38198078e-01
-1.89598426e-01 -5.57642095e-02 -5.71865380e-01 -2.85804421e-01
-1.97502032e-01 8.27743113e-01 -2.14221418e-01 -4.03871953e-01
-5.81484675e-01 -7.22388446e-01 1.76996872e-01 9.53141034e-01
3.16757977e-01 -8.74799371e-01 -1.43023205e+00 3.12561452e-01
-4.49601203e-01 -5.71087956e-01 -2.81689912e-02 3.06316644e-01
-1.52661097e+00 -1.45851445e+00 -9.37777698e-01 -9.52868283e-01
7.98601151e-01 5.86258888e-01 6.91293955e-01 4.12358731e-01
-4.00394142e-01 2.66962618e-01 -7.34128475e-01 -7.43831098e-01
-1.36395529e-01 1.51235357e-01 -5.25311055e-03 -9.57412243e-01
8.60161841e-01 -4.63343501e-01 -1.22188497e+00 1.10903956e-01
-5.24699867e-01 -4.10949677e-01 1.02298403e+00 8.64342451e-01
8.95623192e-02 -5.86632937e-02 8.66167426e-01 -4.73971069e-01
6.44995689e-01 -7.04172373e-01 7.28239343e-02 7.19186813e-02
-1.17479348e+00 -1.17064282e-01 3.75530601e-01 -9.05227140e-02
-7.28270411e-01 -5.19918442e-01 2.13620886e-01 3.98058087e-01
-7.42617786e-01 7.59080350e-01 1.68158844e-01 2.55828083e-01
4.24509764e-01 -1.26173273e-01 2.35227183e-01 -3.47237080e-01
-6.81731939e-01 1.07358932e+00 -3.09923828e-01 -2.48194292e-01
1.93850603e-02 2.08025336e-01 1.73251376e-01 -7.09297478e-01
-7.89496064e-01 -1.09931111e+00 -7.37005115e-01 -1.78961739e-01
7.64094472e-01 -7.99747288e-01 -7.09595740e-01 7.28065893e-02
-7.50688016e-01 -2.33397499e-01 2.12724105e-01 1.64288187e+00
-4.70333278e-01 2.31446400e-01 -4.06950831e-01 -7.18413591e-01
-3.36282462e-01 -1.34373355e+00 4.19166647e-02 1.01669841e-01
-6.59607053e-01 -9.36921060e-01 1.66388795e-01 3.15745652e-01
5.92420995e-01 4.05265868e-01 1.45215201e+00 -4.59630221e-01
9.41524655e-02 -5.74318886e-01 -3.60364437e-01 1.77286550e-01
4.13948953e-01 -2.79514730e-01 -1.26171142e-01 -2.82988995e-01
2.01993808e-01 1.78338036e-01 4.37963754e-01 1.23474371e+00
1.16531730e+00 -1.24464512e-01 -4.65806961e-01 4.35689420e-01
1.59049559e+00 9.59078491e-01 6.92898929e-01 8.32981706e-01
2.54796714e-01 4.98112440e-01 7.02925742e-01 7.32124269e-01
4.16630596e-01 6.07588626e-02 2.70419359e-01 -1.37950450e-01
3.64730209e-01 4.83989209e-01 -6.61196699e-03 5.76557815e-01
-5.44145346e-01 -2.11904310e-02 -1.24021971e+00 3.63754570e-01
-1.80821598e+00 -1.06996977e+00 -5.92437744e-01 2.20670509e+00
2.00576365e-01 3.84855969e-03 3.55810016e-01 3.82637143e-01
2.73723036e-01 -3.14803928e-01 -2.91813403e-01 -7.51006961e-01
1.41190097e-01 -2.26565544e-02 8.34894657e-01 7.12933838e-02
-6.41424417e-01 3.54889542e-01 7.26312542e+00 -3.94165397e-01
-9.96811807e-01 -1.88254356e-01 4.76077318e-01 -3.80424291e-01
2.84232140e-01 2.38718003e-01 -1.31162360e-01 4.90775138e-01
1.04234624e+00 -3.37418646e-01 -9.21061411e-02 1.05287313e+00
1.04217112e+00 -6.21079862e-01 -9.11704957e-01 1.05985403e+00
-3.27842456e-04 -8.77378583e-01 -4.64438535e-02 -9.78963729e-03
3.07441354e-01 -2.53596485e-01 -2.89073199e-01 1.29533201e-01
-2.60460794e-01 -9.66737807e-01 8.23364109e-02 8.96117389e-01
4.77670133e-01 -1.00081253e+00 1.57579172e+00 3.93527657e-01
-7.26617813e-01 -4.82471645e-01 -4.47635084e-01 -8.51110816e-01
4.37804788e-01 5.92046797e-01 -7.08095074e-01 5.38786128e-02
1.11222851e+00 6.46276653e-01 -2.76153117e-01 1.41878211e+00
-1.69233568e-02 1.01396549e+00 2.12832570e-01 -1.24412946e-01
3.28471392e-01 -4.08345759e-01 1.75200835e-01 9.03986573e-01
3.65121514e-01 7.55149782e-01 -4.92214710e-02 -9.29632112e-02
6.54997408e-01 5.88015437e-01 -8.33428979e-01 1.70895457e-01
3.87482196e-01 3.58198434e-01 -7.41774917e-01 -2.76479661e-01
-7.64585197e-01 6.10126972e-01 6.15596399e-02 3.90665412e-01
-6.12503171e-01 -3.59276056e-01 5.75671017e-01 4.15365100e-01
-3.03791195e-01 -2.68324405e-01 -9.24062252e-01 -8.67675006e-01
-2.85478532e-01 -3.76328409e-01 6.46493852e-01 -2.93388903e-01
-8.42661142e-01 1.39670119e-01 2.70262569e-01 -1.15110052e+00
1.12983927e-01 -4.49627399e-01 -5.23040473e-01 7.94214368e-01
-1.25254142e+00 -4.30616796e-01 -4.86551434e-01 4.14978236e-01
5.24271667e-01 -2.13176996e-01 1.20417273e+00 1.46195963e-01
-7.84379601e-01 2.11973652e-01 3.89428467e-01 2.05367401e-01
1.09700215e+00 -7.83029199e-01 -8.61140192e-01 3.57755065e-01
-1.02979600e+00 8.35149527e-01 5.62223494e-01 -7.24699736e-01
-7.71914482e-01 -6.48348391e-01 1.38578451e+00 -3.88916463e-01
4.96192984e-02 6.51744902e-01 -5.68126082e-01 6.07339561e-01
1.65055022e-01 -8.34947765e-01 1.31871319e+00 1.07723348e-01
3.67635339e-01 -5.25784157e-02 -8.96116376e-01 7.54890740e-01
8.27578843e-01 7.07000270e-02 -8.66193712e-01 9.76444855e-02
-1.30358171e-02 -3.66920643e-02 -1.08728278e+00 4.87149447e-01
1.00881720e+00 -1.23784816e+00 1.03592682e+00 -9.05587852e-01
4.95828450e-01 5.14001787e-01 1.55693799e-01 -1.10418546e+00
-5.45601845e-01 1.85685515e-01 4.41511929e-01 4.85845715e-01
1.90274432e-01 -9.23515677e-01 6.72290027e-01 8.98550034e-01
-2.42440134e-01 -1.05273950e+00 -5.18320560e-01 -5.96169233e-01
2.55825907e-01 -1.59931630e-01 8.07237089e-01 6.82327628e-01
5.23042321e-01 -8.60055014e-02 -5.70936017e-02 -9.91315395e-02
3.55754584e-01 -1.45984799e-01 4.90110785e-01 -1.27676296e+00
5.04491568e-01 -6.13078535e-01 -2.97880560e-01 -1.51706666e-01
-2.53088251e-02 -5.51741600e-01 -5.26633501e-01 -2.52668548e+00
6.82493627e-01 -5.80426991e-01 -4.48685914e-01 1.66749477e-01
-2.40915477e-01 -7.00450778e-01 -2.39680424e-01 5.13318837e-01
2.29270563e-01 9.26362574e-02 9.84471500e-01 3.05658698e-01
-6.04850888e-01 3.30925077e-01 -7.23601282e-01 7.46806145e-01
1.52472067e+00 -7.80917764e-01 -6.10649109e-01 -3.18348408e-01
-9.08096358e-02 4.33515102e-01 -6.78014755e-02 -1.04873204e+00
4.89592664e-02 -9.53383148e-01 4.35700148e-01 -7.96498954e-01
-1.15224011e-01 -1.27741361e+00 -1.14625454e-01 1.29578364e+00
-3.33461702e-01 4.94798720e-01 -1.15351915e-01 -5.65755889e-02
-1.79452300e-01 -3.73643041e-01 6.75486088e-01 -3.30106467e-02
-3.43277842e-01 2.23268121e-01 -9.75171268e-01 -5.51042259e-02
1.36643970e+00 -4.95446593e-01 -6.44242838e-02 -4.52201009e-01
-6.97481751e-01 1.65498182e-02 3.63051534e-01 1.34205163e-01
8.36416125e-01 -1.14389181e+00 -6.84698999e-01 -1.04860641e-01
5.41444123e-01 -2.45384142e-01 4.38239992e-01 1.46548009e+00
-1.21765172e+00 1.06977689e+00 -4.73945379e-01 8.10045842e-03
-1.29973018e+00 5.77315271e-01 3.59167717e-02 -1.16276942e-01
-1.06299055e+00 2.27605417e-01 -2.02176198e-01 3.91842514e-01
6.26496494e-01 -5.21259487e-01 -5.77903390e-01 6.42904863e-02
6.44495010e-01 1.09657085e+00 -1.76061735e-01 -4.26183850e-01
-6.19563699e-01 4.28926736e-01 8.76820236e-02 1.88789189e-01
1.29239273e+00 -3.79224479e-01 -2.38544732e-01 7.11276174e-01
1.09580100e+00 -4.38721985e-01 -4.70440388e-01 6.86253607e-02
-8.76322687e-02 -4.84592557e-01 1.60983220e-01 -6.77729130e-01
-6.40297711e-01 7.90752351e-01 8.35650444e-01 -4.25250053e-01
1.17220736e+00 5.58964945e-02 9.43385780e-01 5.48076987e-01
3.77982885e-01 -1.07628310e+00 8.68595243e-02 3.69194031e-01
8.30643773e-01 -1.53432703e+00 -3.44178826e-01 1.18109338e-01
-7.88706660e-01 1.36196649e+00 1.55333728e-01 -1.68386862e-01
1.27111804e+00 -1.14042731e-02 4.22681093e-01 -1.67001024e-01
-1.24071933e-01 2.20831946e-01 -2.16291904e-01 4.65738714e-01
9.86817181e-01 5.74992418e-01 -1.53336084e+00 8.26080024e-01
-3.15472126e-01 2.51875818e-01 6.80172026e-01 1.35644412e+00
-7.25402355e-01 -1.28723347e+00 -5.45814514e-01 1.20996511e+00
-5.40094495e-01 -1.94065168e-01 -4.02221754e-02 6.13879681e-01
2.58042403e-02 1.21540940e+00 -1.66077986e-01 9.42966938e-02
6.09729707e-01 2.30519503e-01 2.01381207e-01 -4.95887697e-01
-4.59947050e-01 -2.67667085e-01 -3.87048461e-02 -3.84172112e-01
-3.69372696e-01 -8.37337971e-01 -1.58144629e+00 -5.35205364e-01
-1.69618912e-02 5.40926456e-01 6.94363117e-01 1.05337155e+00
9.84282419e-02 7.38625526e-01 5.23460150e-01 1.30978808e-01
-1.58582017e-01 -9.81519938e-01 -7.94114947e-01 7.33644096e-03
4.31348652e-01 -6.70892656e-01 -6.74839497e-01 -9.90760401e-02] | [8.015809059143066, 6.1444830894470215] |
6dea8412-2013-4921-ac58-6fc14bbe6a36 | poet-a-generative-model-of-protein-families | 2306.06156 | null | https://arxiv.org/abs/2306.06156v1 | https://arxiv.org/pdf/2306.06156v1.pdf | PoET: A generative model of protein families as sequences-of-sequences | Generative protein language models are a natural way to design new proteins with desired functions. However, current models are either difficult to direct to produce a protein from a specific family of interest, or must be trained on a large multiple sequence alignment (MSA) from the specific family of interest, making them unable to benefit from transfer learning across families. To address this, we propose $\textbf{P}$r$\textbf{o}$tein $\textbf{E}$volutionary $\textbf{T}$ransformer (PoET), an autoregressive generative model of whole protein families that learns to generate sets of related proteins as sequences-of-sequences across tens of millions of natural protein sequence clusters. PoET can be used as a retrieval-augmented language model to generate and score arbitrary modifications conditioned on any protein family of interest, and can extrapolate from short context lengths to generalize well even for small families. This is enabled by a unique Transformer layer; we model tokens sequentially within sequences while attending between sequences order invariantly, allowing PoET to scale to context lengths beyond those used during training. PoET outperforms existing protein language models and evolutionary sequence models for variant function prediction in extensive experiments on deep mutational scanning datasets, improving variant effect prediction across proteins of all MSA depths. | ['Tristan Bepler', 'Timothy F. Truong Jr'] | 2023-06-09 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 7.06288099e-01 1.70614153e-01 1.13997169e-01 -5.46379089e-01
-9.74979818e-01 -8.42324495e-01 2.54536897e-01 9.47548673e-02
-3.04706782e-01 1.23813832e+00 -9.11135450e-02 -6.52463734e-01
1.11682341e-02 -7.11997688e-01 -1.37309790e+00 -8.74142051e-01
-5.87140694e-02 9.27522480e-01 3.26935649e-01 -4.56717223e-01
2.19478924e-02 4.70546186e-01 -1.28410470e+00 7.02329874e-01
9.79659081e-01 2.91307300e-01 8.34083676e-01 6.76876009e-01
-3.10386300e-01 2.48924971e-01 -5.48145473e-01 -2.31571376e-01
1.33780122e-01 -6.41931534e-01 -7.42424428e-01 -2.67625391e-01
1.93604350e-01 8.80625658e-03 -1.01379722e-01 6.52341723e-01
4.90588933e-01 6.74600303e-02 9.56932962e-01 -6.40423000e-01
-1.09402502e+00 5.25050998e-01 -4.81637329e-01 2.60818124e-01
2.94850141e-01 6.86342299e-01 1.24024880e+00 -9.41902220e-01
1.05190074e+00 1.24097371e+00 5.37637293e-01 6.55344188e-01
-1.82350791e+00 -6.99197054e-01 1.98246136e-01 -8.19720551e-02
-1.09100187e+00 -4.12430130e-02 6.40437230e-02 -4.53695685e-01
2.00021935e+00 -1.03912637e-01 5.13443530e-01 1.21564186e+00
5.93591154e-01 4.09761876e-01 4.94941622e-01 -7.01854676e-02
1.26617536e-01 -5.92553616e-01 6.10910617e-02 7.02514529e-01
-5.31355739e-02 -1.57059953e-01 -6.90604746e-01 -6.19777739e-01
6.23257279e-01 2.11303711e-01 -2.00250924e-01 -3.53803128e-01
-1.10616100e+00 9.03057635e-01 3.72758955e-01 7.87512362e-02
-6.64330304e-01 2.30438441e-01 1.82168722e-01 4.28584546e-01
1.60024658e-01 6.37959361e-01 -9.32797253e-01 1.45473303e-02
-8.33581448e-01 6.10307992e-01 3.87391806e-01 9.73698199e-01
1.02197886e+00 -4.79532778e-02 -1.13579864e-02 8.48722756e-01
1.48975328e-01 5.40583789e-01 6.17339611e-01 -4.77096617e-01
1.84456334e-01 4.26375002e-01 1.64356381e-01 -1.70905635e-01
-3.02669048e-01 -1.34444937e-01 -3.79396468e-01 4.87183407e-02
3.35385114e-01 -1.29367243e-02 -1.15462720e+00 2.07907510e+00
4.92074601e-02 -1.06306039e-01 9.75001752e-02 5.85424304e-01
4.32883680e-01 1.08164418e+00 4.81976122e-01 -2.07357749e-01
1.31960535e+00 -6.16803646e-01 2.64214575e-02 -2.62738407e-01
7.08836794e-01 -6.46897554e-01 1.17993665e+00 4.49274421e-01
-1.07210541e+00 -4.08884555e-01 -8.75389040e-01 -2.08504722e-01
-3.65431935e-01 -4.37173814e-01 8.40313494e-01 2.94632196e-01
-1.14555430e+00 9.09703970e-01 -9.12170470e-01 -3.15488636e-01
4.04360354e-01 7.80122459e-01 -3.83643329e-01 -8.54368210e-02
-1.17768574e+00 9.22161698e-01 7.26967931e-01 -2.41570264e-01
-9.93512213e-01 -9.21428561e-01 -6.11190856e-01 2.86784589e-01
-7.72488192e-02 -1.01783347e+00 9.27005231e-01 -9.60539997e-01
-1.24204957e+00 7.92336702e-01 -5.36235929e-01 -7.50053406e-01
-2.39606023e-01 -1.56779885e-01 -2.19263494e-01 3.37852463e-02
7.57356435e-02 1.05067062e+00 6.46711528e-01 -5.59156299e-01
-3.12370032e-01 -2.58383065e-01 -2.53971130e-01 -7.38364737e-03
2.62219518e-01 1.65876836e-01 2.42550060e-01 -6.38712883e-01
-8.71579647e-02 -9.11200047e-01 -6.30373836e-01 -1.82704762e-01
-5.98915527e-03 -3.38050485e-01 4.28868562e-01 -6.34188592e-01
7.36513078e-01 -1.76689816e+00 5.13110638e-01 2.29254887e-01
2.36761615e-01 4.49700296e-01 -5.31412661e-01 7.27319837e-01
-6.33875787e-01 8.42117742e-02 -4.54691619e-01 4.45136428e-01
-1.70579761e-01 3.36288273e-01 -5.12787223e-01 2.80556887e-01
8.30541015e-01 1.17786276e+00 -8.54773939e-01 8.44637975e-02
-1.64713681e-01 6.39417827e-01 -8.50706398e-01 1.59106657e-01
-1.01130223e+00 3.03809911e-01 -5.84569454e-01 4.41864371e-01
5.38740635e-01 -7.36980379e-01 4.81195062e-01 1.13072023e-01
3.24490607e-01 4.83673394e-01 -5.47919452e-01 1.66732037e+00
-8.54840055e-02 4.40000324e-03 -3.78834993e-01 -1.12103903e+00
1.00980651e+00 2.73502409e-01 5.79489887e-01 -1.54669002e-01
-1.53487295e-01 2.90997982e-01 4.63453323e-01 -9.07610655e-02
2.61725485e-01 -5.94884515e-01 -4.53169122e-02 5.41713059e-01
5.56978881e-01 5.51934130e-02 3.29836607e-01 1.41514897e-01
1.62918353e+00 5.06497502e-01 2.62684673e-01 -2.19240591e-01
3.30299109e-01 1.89222902e-01 7.17757463e-01 5.08477032e-01
1.52246565e-01 4.53682512e-01 4.50967908e-01 -4.95244861e-01
-1.50418127e+00 -1.42683661e+00 1.30598266e-02 1.59693205e+00
-2.62813598e-01 -3.51011783e-01 -6.32191062e-01 -6.60862803e-01
-2.24157032e-02 6.10633433e-01 -1.57590196e-01 -3.75694811e-01
-8.42142165e-01 -1.18074596e+00 5.98427773e-01 4.02795732e-01
-3.64815086e-01 -1.54316008e+00 -2.71027803e-01 6.57954514e-01
-7.08148405e-02 -4.53824461e-01 -6.92577243e-01 8.09959710e-01
-7.61493444e-01 -9.38698769e-01 -9.14490938e-01 -1.02823889e+00
7.02663779e-01 -7.18167871e-02 1.27785945e+00 -2.21089825e-01
-4.42887545e-01 -2.10473269e-01 -1.44397378e-01 -3.26210797e-01
-6.94309533e-01 7.54882321e-02 2.85983294e-01 -4.28586811e-01
9.56990302e-01 -1.07446790e+00 -5.40810406e-01 2.38011271e-01
-1.03246868e+00 4.54107719e-03 8.19963276e-01 1.39016509e+00
9.13360178e-01 -6.23998463e-01 9.04165030e-01 -1.04393566e+00
5.06688356e-01 -6.56438351e-01 -6.60599589e-01 3.95931631e-01
-3.70833308e-01 6.24565303e-01 9.73442972e-01 -7.50312269e-01
-9.03081179e-01 2.88095206e-01 -4.04904455e-01 -1.72879398e-01
-1.40130535e-01 3.64272028e-01 -2.06886381e-01 3.84912580e-01
1.05655849e+00 7.55848050e-01 1.40373902e-02 -3.83612007e-01
6.09244883e-01 3.18542600e-01 4.57029521e-01 -7.87520647e-01
5.42399883e-01 3.26753817e-02 -6.01108782e-02 -6.28109336e-01
-2.68856525e-01 -2.85036474e-01 -5.92775285e-01 6.01450384e-01
7.91895390e-01 -9.31113005e-01 -7.88704991e-01 8.58359039e-02
-1.04089499e+00 -4.19822127e-01 -1.41187891e-01 2.22488463e-01
-8.72901857e-01 5.03499925e-01 -8.59133542e-01 -3.72752845e-01
-5.67144871e-01 -1.26593339e+00 1.25910330e+00 5.04029393e-02
-6.55683994e-01 -6.56248629e-01 6.70667365e-02 8.28441679e-02
1.53414845e-01 5.79300486e-02 1.41338158e+00 -8.65222514e-01
-7.23087788e-01 1.90729022e-01 1.10623718e-03 1.41806841e-01
1.15428507e-01 -1.52676418e-01 -4.14932191e-01 -4.61571515e-01
-3.59264374e-01 -6.05253994e-01 1.19472349e+00 3.48369420e-01
9.65299785e-01 -2.66221404e-01 -6.48556590e-01 4.28950548e-01
1.15922129e+00 3.54862183e-01 6.84188008e-01 -6.70035649e-03
4.17362124e-01 2.81659365e-01 5.50480247e-01 1.52891919e-01
-2.42663205e-01 4.56970125e-01 1.00648656e-01 4.47032079e-02
1.55355677e-01 -5.20412982e-01 6.70856476e-01 2.13804781e-01
4.13671024e-02 -5.27278602e-01 -7.47286558e-01 4.94446188e-01
-1.68631375e+00 -1.11357975e+00 1.95317894e-01 2.07101893e+00
1.21844399e+00 1.09717563e-01 5.92575073e-02 -6.19761586e-01
6.72872365e-01 -1.82778403e-01 -1.08765244e+00 -4.26058531e-01
-3.37723613e-01 1.05601108e+00 4.68073130e-01 3.90880346e-01
-5.93510211e-01 1.28526318e+00 6.44052982e+00 9.00398791e-01
-1.17853260e+00 -9.58868340e-02 5.84503770e-01 -2.49628872e-01
-6.35034204e-01 2.24659607e-01 -9.67903733e-01 6.12804115e-01
1.30988216e+00 -1.09312207e-01 5.85897744e-01 7.67208457e-01
1.18508369e-01 4.12555546e-01 -1.33961225e+00 6.26110554e-01
-2.49047354e-01 -1.65710354e+00 2.95269340e-01 1.03336364e-01
4.10986513e-01 5.40504396e-01 1.79639831e-01 2.94384629e-01
1.01062059e+00 -1.46585500e+00 2.46845692e-01 4.36047435e-01
8.64126444e-01 -6.83387101e-01 1.75208613e-01 4.49580431e-01
-8.24235201e-01 1.34824127e-01 -8.10932636e-01 2.20417038e-01
3.02725136e-01 4.51239347e-01 -1.33983588e+00 3.71415876e-02
5.18841803e-01 4.83707726e-01 -1.60402685e-01 2.83338338e-01
-2.09573042e-02 6.12286747e-01 -4.10936624e-01 -1.78439602e-01
9.68075246e-02 -2.29592741e-01 3.74522239e-01 1.19774294e+00
4.41218406e-01 3.18263113e-01 1.23905137e-01 1.18295801e+00
-1.81960195e-01 1.55949995e-01 -4.46414858e-01 -5.04069626e-01
2.94887692e-01 8.73786867e-01 -4.78900462e-01 -3.47281843e-01
-2.07763553e-01 1.23006928e+00 6.75318480e-01 5.12573361e-01
-8.75111461e-01 -4.23857570e-01 1.07515800e+00 2.73853362e-01
7.30352879e-01 -1.76693901e-01 4.22615111e-01 -1.07319927e+00
-1.71837255e-01 -1.05746877e+00 2.81509101e-01 -9.11690831e-01
-1.59348190e+00 4.95454252e-01 -3.37960571e-01 -9.09899890e-01
-4.18671697e-01 -8.97020876e-01 -2.31719702e-01 1.33547640e+00
-9.52725887e-01 -1.23157334e+00 7.78890669e-01 2.45825440e-01
5.76847672e-01 -4.33884859e-01 1.10317171e+00 -1.66215608e-03
-1.33379385e-01 5.99889040e-01 3.04476857e-01 -4.54296529e-01
8.14899087e-01 -1.16179585e+00 1.10931611e+00 5.75656116e-01
2.51738913e-02 1.28335261e+00 9.20972049e-01 -8.64152551e-01
-1.18608928e+00 -1.15647447e+00 9.28238750e-01 -6.65029347e-01
4.72379923e-01 -5.57471097e-01 -1.38236809e+00 9.01862085e-01
-8.67878720e-02 -1.45292118e-01 1.00171924e+00 3.54698971e-02
-6.60302579e-01 3.64725053e-01 -1.23547435e+00 5.74396968e-01
1.33854485e+00 -5.11902392e-01 -6.67302966e-01 5.73782146e-01
1.10554588e+00 -2.51665086e-01 -1.07901931e+00 2.97560722e-01
5.26839197e-01 -6.09849870e-01 1.25088191e+00 -1.28072298e+00
5.43446183e-01 -4.30667728e-01 -4.89422940e-02 -1.22801089e+00
-7.21314192e-01 -7.00607359e-01 1.08971618e-01 8.04785252e-01
1.08041942e+00 -7.81460881e-01 7.76382744e-01 2.65408397e-01
-3.16169530e-01 -6.26465797e-01 -7.60824740e-01 -5.46198130e-01
5.31016409e-01 -1.14670403e-01 7.46506691e-01 6.69436514e-01
2.70256639e-01 6.36162698e-01 -1.61128759e-01 4.55778167e-02
1.15852170e-01 1.26543537e-01 5.86310983e-01 -8.94156277e-01
-9.65820432e-01 -3.49905521e-01 -2.07896709e-01 -1.40429938e+00
1.89309925e-01 -1.21954286e+00 1.59542248e-01 -1.32830763e+00
4.35797423e-01 -2.42987499e-01 -4.89091843e-01 6.25323415e-01
-4.00478959e-01 -2.28784401e-02 -3.04865301e-01 1.41901374e-01
-3.76112819e-01 4.55728650e-01 8.73650968e-01 -1.71159908e-01
-1.00958876e-01 -1.85198888e-01 -8.48543823e-01 4.14933711e-01
5.83079457e-01 -5.42624354e-01 -4.64042306e-01 -4.83897589e-02
4.54099327e-01 1.17911987e-01 1.20851822e-01 -4.58869249e-01
-3.01094145e-01 -5.23051977e-01 6.53383315e-01 -5.14101624e-01
4.14154291e-01 -3.38372737e-01 5.20498335e-01 4.67890203e-01
-4.57752913e-01 1.55871674e-01 1.20403975e-01 7.05042422e-01
3.12398285e-01 6.19573146e-02 7.63607383e-01 -4.34376806e-01
-4.34774578e-01 4.18020785e-01 -7.03632653e-01 -1.77819163e-01
7.92380929e-01 -9.06967595e-02 -2.19754159e-01 -6.64874017e-02
-8.92633796e-01 1.30982786e-01 8.30940187e-01 4.42315936e-01
5.58499217e-01 -8.04896533e-01 -7.08984613e-01 3.50178599e-01
3.44864100e-01 -5.11087961e-02 2.69226223e-01 1.49806812e-01
-6.23454690e-01 5.83008289e-01 -2.75075614e-01 -6.03626966e-01
-1.23800790e+00 8.51801157e-01 3.19033563e-01 -4.70357001e-01
-4.54533637e-01 1.14906573e+00 7.14005411e-01 -6.98318303e-01
-5.01838505e-01 -3.34219366e-01 1.08880036e-01 -7.08619893e-01
5.30557692e-01 -3.51649404e-01 -4.14604805e-02 -6.35391653e-01
-1.44587621e-01 2.42344379e-01 -4.96678799e-01 2.19348043e-01
1.57878637e+00 2.78112799e-01 -1.78371444e-01 1.19308056e-02
1.01445913e+00 -3.78311694e-01 -1.43156719e+00 -1.92954361e-01
1.19102389e-01 -9.01431516e-02 -8.75809908e-01 -9.41395998e-01
-3.38225275e-01 6.47690773e-01 5.12020946e-01 -5.06998956e-01
7.75539637e-01 2.60584533e-01 1.16363227e+00 6.75432861e-01
5.87857127e-01 -6.34358048e-01 2.01210249e-02 5.32167852e-01
5.39062917e-01 -8.53530049e-01 -3.89017910e-01 -2.13848338e-01
-4.39479828e-01 8.94376099e-01 3.75758260e-01 -1.70903832e-01
2.57668942e-01 1.52494475e-01 -2.82616138e-01 -1.59521803e-01
-1.08693850e+00 5.67662381e-02 7.39167780e-02 8.92498732e-01
8.78974915e-01 3.39221247e-02 -4.68089253e-01 6.89384699e-01
-2.48674884e-01 1.64518645e-03 1.80579513e-01 9.29007649e-01
-7.58255303e-01 -1.77866280e+00 -6.11072071e-02 6.21565402e-01
-6.41942441e-01 -5.04149020e-01 -3.77936959e-01 2.01057807e-01
2.53322870e-01 5.22578299e-01 -9.51345339e-02 -5.72607629e-02
9.49818119e-02 6.43659472e-01 5.94338596e-01 -9.43364739e-01
-5.27134299e-01 1.59382194e-01 -2.09809691e-01 -2.59319097e-01
5.21826819e-02 -7.57456660e-01 -1.82985997e+00 -6.79942518e-02
-2.83256024e-01 2.91029066e-01 1.19346820e-01 6.66297853e-01
7.37446070e-01 4.66858268e-01 1.76769018e-01 -6.22211039e-01
-5.46105564e-01 -8.96713972e-01 -5.34023464e-01 4.34032381e-01
1.32311016e-01 -3.94434661e-01 1.93047255e-01 4.74651128e-01] | [4.690733909606934, 5.611966133117676] |
8b0fbe71-7122-479e-bf71-17ef1ce765db | enhancing-user-behavior-sequence-modeling-by | 2208.10846 | null | https://arxiv.org/abs/2208.10846v1 | https://arxiv.org/pdf/2208.10846v1.pdf | Enhancing User Behavior Sequence Modeling by Generative Tasks for Session Search | Users' search tasks have become increasingly complicated, requiring multiple queries and interactions with the results. Recent studies have demonstrated that modeling the historical user behaviors in a session can help understand the current search intent. Existing context-aware ranking models primarily encode the current session sequence (from the first behavior to the current query) and compute the ranking score using the high-level representations. However, there is usually some noise in the current session sequence (useless behaviors for inferring the search intent) that may affect the quality of the encoded representations. To help the encoding of the current user behavior sequence, we propose to use a decoder and the information of future sequences and a supplemental query. Specifically, we design three generative tasks that can help the encoder to infer the actual search intent: (1) predicting future queries, (2) predicting future clicked documents, and (3) predicting a supplemental query. We jointly learn the ranking task with these generative tasks using an encoder-decoder structured approach. Extensive experiments on two public search logs demonstrate that our model outperforms all existing baselines, and the designed generative tasks can actually help the ranking task. Besides, additional experiments also show that our approach can be easily applied to various Transformer-based encoder-decoder models and improve their performance. | ['Ji-Rong Wen', 'Xiaohua Cheng', 'Zhao Cao', 'Yutao Zhu', 'Zhicheng Dou', 'Haonan Chen'] | 2022-08-23 | null | null | null | null | ['session-search', 'document-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.81401122e-01 -2.60952234e-01 -5.70380569e-01 -8.65861416e-01
-1.16774070e+00 -5.10571778e-01 7.82477796e-01 -3.46605182e-01
-2.94728905e-01 4.13284808e-01 7.31171131e-01 -4.60727692e-01
-3.38014774e-02 -6.32507563e-01 -6.75662994e-01 -1.35577455e-01
-7.43244663e-02 6.15876913e-01 3.59737217e-01 -3.72189343e-01
4.01925176e-01 -3.35649997e-01 -1.67463827e+00 6.85620487e-01
1.11532807e+00 1.12653482e+00 7.38413811e-01 9.41391945e-01
-1.19209625e-01 1.05207396e+00 -6.18035078e-01 -3.16524893e-01
-8.45507085e-02 -5.59993982e-01 -8.65384340e-01 -1.84216142e-01
1.13599621e-01 -9.02101696e-01 -7.94624925e-01 6.92003906e-01
3.63866001e-01 1.61521271e-01 3.44960839e-01 -9.32876229e-01
-9.44478810e-01 7.84303725e-01 -7.16211721e-02 2.99931914e-01
7.39916861e-01 -8.01241547e-02 1.49583900e+00 -5.44449210e-01
4.76748526e-01 1.09468126e+00 7.88281336e-02 4.24940318e-01
-8.97048593e-01 -4.92849827e-01 4.05288130e-01 5.26906908e-01
-1.07499504e+00 -3.83109987e-01 7.15720654e-01 -2.29337886e-01
9.79874372e-01 4.86830205e-01 3.58602136e-01 1.55054343e+00
3.99135333e-03 1.25022233e+00 6.88192487e-01 -9.48188603e-02
-7.43536800e-02 9.71483216e-02 5.75796485e-01 5.70284903e-01
-2.59886235e-01 1.65287808e-01 -6.07144356e-01 -5.92957497e-01
2.20265210e-01 4.11135048e-01 -2.21661821e-01 -7.47945681e-02
-9.03846860e-01 8.30886960e-01 4.66056049e-01 1.23830520e-01
-3.50343496e-01 3.57248932e-01 1.99151397e-01 2.30912715e-01
2.97899276e-01 1.60865873e-01 -4.83440697e-01 -5.14138460e-01
-8.45733881e-01 2.65667170e-01 1.14187181e+00 1.20630789e+00
8.13918054e-01 -4.38752592e-01 -6.19579196e-01 9.77615058e-01
5.57568908e-01 4.87949371e-01 8.64218831e-01 -6.20291591e-01
6.31260335e-01 4.37390596e-01 2.23724276e-01 -6.20717347e-01
2.41874233e-01 -6.86416745e-01 -3.05082202e-01 -1.02547467e+00
-1.37300164e-01 1.38352096e-01 -1.09659541e+00 1.82722473e+00
-3.65121514e-01 7.77091607e-02 -2.08080903e-01 6.96381927e-01
5.59094012e-01 7.89612055e-01 5.93892373e-02 -1.21070527e-01
1.22341931e+00 -1.12821627e+00 -7.77486205e-01 -6.40976548e-01
6.35432780e-01 -6.54601574e-01 1.20421278e+00 -8.54749456e-02
-9.64727879e-01 -6.76183105e-01 -8.22107494e-01 -3.99414122e-01
-2.62519598e-01 5.06890774e-01 8.21693718e-01 3.09138238e-01
-9.79635000e-01 1.44808009e-01 -8.84812951e-01 -1.71757355e-01
-8.04877579e-02 3.67396295e-01 4.66318935e-01 -3.50528024e-02
-1.60437953e+00 3.79130214e-01 2.08006293e-01 -1.13660265e-02
-9.53384399e-01 -4.78720725e-01 -5.41063845e-01 5.40006280e-01
4.07517940e-01 -7.66582429e-01 1.81907630e+00 -4.50747073e-01
-1.14914429e+00 4.46275294e-01 -1.04829717e+00 -4.33810413e-01
2.77085543e-01 -5.08524954e-01 -5.08167922e-01 -4.23894346e-01
1.72441453e-01 2.45382592e-01 5.35158634e-01 -9.90876496e-01
-8.91895592e-01 -4.82920498e-01 1.54379383e-01 4.06789750e-01
-4.08547461e-01 -2.03939483e-01 -1.24322951e+00 -3.77644241e-01
-9.64217186e-02 -9.17106986e-01 2.68374532e-02 -6.41404390e-01
-6.63189113e-01 -6.07438803e-01 5.67442596e-01 -9.56272900e-01
1.95034552e+00 -2.11978126e+00 -1.23645201e-01 1.51434630e-01
-3.55636664e-02 2.20892922e-04 -8.07152689e-02 6.77570283e-01
4.84453440e-01 1.20390832e-01 2.72936434e-01 -2.15821877e-01
2.15382010e-01 1.26441181e-01 -9.44414973e-01 -3.38679731e-01
-3.86384547e-01 1.37447786e+00 -9.37035263e-01 -2.92481035e-01
-2.73975015e-01 2.97808886e-01 -7.02424526e-01 7.86821306e-01
-6.05852306e-01 -3.57510783e-02 -9.76282775e-01 4.00380820e-01
2.25210503e-01 -8.74633551e-01 3.73024970e-01 2.39442289e-02
3.14278215e-01 1.22594571e+00 -5.37783563e-01 1.77434742e+00
-8.68829608e-01 4.67759907e-01 -3.82820427e-01 -6.21588707e-01
6.08340740e-01 7.16092736e-02 2.43471667e-01 -1.20168495e+00
-4.23284531e-01 1.80980831e-01 -2.35987648e-01 -6.04564846e-01
6.80965126e-01 6.39437377e-01 -2.17847154e-01 8.80061209e-01
-1.23486198e-01 6.60409570e-01 4.23074245e-01 6.00508630e-01
1.24988616e+00 6.52182028e-02 -8.86095967e-03 3.79725069e-01
5.13681829e-01 -1.89850435e-01 2.94107407e-01 1.25264347e+00
3.10425162e-01 1.94812566e-01 3.22979569e-01 -1.39414325e-01
-7.78294683e-01 -9.83904302e-01 2.20733255e-01 1.84937656e+00
3.53355497e-01 -7.19642460e-01 -3.92143667e-01 -7.86457837e-01
-9.84332860e-02 1.05423212e+00 -5.66533506e-01 -5.59054911e-01
-6.82069182e-01 -5.40936470e-01 2.26067156e-01 6.19434655e-01
5.92007935e-01 -8.24349940e-01 -1.03556283e-01 3.26908350e-01
-7.71736860e-01 -8.99078012e-01 -1.16693830e+00 1.44164741e-01
-7.99030244e-01 -9.66232777e-01 -3.94448936e-01 -8.67153704e-01
4.51284438e-01 5.81662238e-01 1.26816607e+00 3.54685456e-01
-3.12683322e-02 3.76269937e-01 -2.65547872e-01 3.01518645e-02
-2.56774008e-01 4.62031335e-01 -4.61045444e-01 -1.54087424e-01
7.85702288e-01 -4.75676805e-01 -1.03001320e+00 5.63510239e-01
-8.10642660e-01 1.38056681e-01 8.53227437e-01 8.11810195e-01
2.41255715e-01 -1.21281989e-01 4.37050253e-01 -1.15212774e+00
1.17535734e+00 -8.25571239e-01 -2.52894074e-01 8.01594436e-01
-1.06028163e+00 4.82363790e-01 4.93716121e-01 -4.31854069e-01
-1.32545173e+00 -2.25969344e-01 -3.78436685e-01 1.67127892e-01
1.11485094e-01 6.64530575e-01 3.94589268e-03 6.46987259e-01
5.41001141e-01 1.00291038e+00 -4.13592249e-01 -8.17789614e-01
3.32660973e-01 1.05926752e+00 2.64985204e-01 -2.44649902e-01
5.98711848e-01 5.02785258e-02 -7.68792510e-01 -3.50710124e-01
-1.21684384e+00 -8.76357436e-01 -1.76785603e-01 1.13643825e-01
3.96638304e-01 -9.44410264e-01 -7.40519881e-01 -7.66489247e-04
-1.07829857e+00 -1.39552131e-01 1.74774364e-01 1.67621464e-01
-5.17837405e-01 2.72729486e-01 -6.79106951e-01 -7.76065886e-01
-3.54997575e-01 -1.35846853e+00 1.47255409e+00 2.03094736e-01
-1.31225571e-01 -1.03810966e+00 2.80402273e-01 6.60638332e-01
5.59111118e-01 -8.51679265e-01 1.25650680e+00 -1.06016672e+00
-1.04976737e+00 -4.35265064e-01 -1.71641693e-01 -1.52140670e-03
2.47868612e-01 -5.49083591e-01 -8.89360964e-01 -1.85214609e-01
-5.33950143e-02 -3.52321565e-01 1.02634037e+00 2.39737276e-02
1.59311652e+00 -6.45773470e-01 -7.12144673e-01 4.84268844e-01
1.08792520e+00 6.59692943e-01 6.76851690e-01 4.37376313e-02
2.82651126e-01 2.10430235e-01 5.57315767e-01 2.84900606e-01
6.85067296e-01 1.04508829e+00 1.43137395e-01 4.09385324e-01
-1.73846498e-01 -1.21540809e+00 6.12721264e-01 8.69648159e-01
2.83642411e-01 -6.68570161e-01 -3.33125412e-01 3.07276338e-01
-1.98250270e+00 -1.11086369e+00 3.49800676e-01 2.17585778e+00
1.03615797e+00 5.83762415e-02 -1.10658363e-01 -4.09906298e-01
4.44347173e-01 4.42132384e-01 -9.32138264e-01 -6.67413697e-02
4.49956000e-01 -2.27959864e-02 1.59309611e-01 6.59518898e-01
-7.94444442e-01 8.22872341e-01 6.86662436e+00 7.04813659e-01
-7.31010199e-01 -7.43667129e-03 4.44024473e-01 -3.11131962e-02
-8.30877423e-01 2.40793467e-01 -1.29268241e+00 9.88043785e-01
1.08325171e+00 -5.58900356e-01 7.37865388e-01 1.11779475e+00
1.16191618e-01 1.53008908e-01 -1.37054610e+00 8.15658331e-01
2.62046516e-01 -1.04273927e+00 3.83129656e-01 2.68399358e-01
3.69005650e-01 4.87628998e-03 2.72479862e-01 9.26749527e-01
5.21035552e-01 -7.83694804e-01 1.40426621e-01 7.94986188e-01
6.03269339e-01 -2.12606221e-01 4.39504027e-01 7.59721816e-01
-1.26941812e+00 -2.57870495e-01 -2.55358189e-01 2.95679361e-01
4.01301920e-01 4.46599156e-01 -1.09104896e+00 3.17231208e-01
3.80597383e-01 8.63797843e-01 -8.69355142e-01 9.12993789e-01
-2.58423507e-01 8.01562965e-01 -1.24761604e-01 -5.20897567e-01
9.81619954e-02 -7.26957470e-02 2.17884064e-01 1.01431632e+00
4.03445542e-01 -2.30256215e-01 1.22256801e-01 9.00461495e-01
-3.55438739e-01 -1.45267099e-01 -4.32506293e-01 -3.30167770e-01
4.65163469e-01 7.99926162e-01 2.92801615e-02 -4.74147797e-01
-5.53634882e-01 1.41554451e+00 3.79644692e-01 7.72918999e-01
-8.53967547e-01 -4.23686475e-01 6.27419651e-01 2.38627955e-01
2.69414514e-01 -7.04702213e-02 2.42866248e-01 -1.42894149e+00
3.43944997e-01 -7.61196434e-01 5.89303911e-01 -8.81345570e-01
-1.04036260e+00 5.99037647e-01 -1.24724932e-01 -9.53765690e-01
-1.13712370e+00 -4.46386933e-02 -5.17335415e-01 9.51175272e-01
-1.58469594e+00 -8.85914743e-01 -4.13704485e-01 4.29243118e-01
1.11544693e+00 -1.20269626e-01 6.06576741e-01 3.67106587e-01
-2.57199943e-01 7.03305364e-01 2.12407306e-01 1.03255220e-01
5.81187367e-01 -1.14937115e+00 5.52325130e-01 5.54167747e-01
4.54886049e-01 1.09528267e+00 3.17704022e-01 -8.82226467e-01
-1.62431502e+00 -8.47500205e-01 1.44254804e+00 -6.00623071e-01
5.93192041e-01 -5.58593035e-01 -7.84378111e-01 9.70498085e-01
9.70766842e-02 -6.87660635e-01 8.75947177e-01 5.05823731e-01
-3.14250797e-01 -4.04532664e-02 -2.54485488e-01 7.36747921e-01
1.39858019e+00 -1.01473808e+00 -8.09996307e-01 5.96526265e-01
9.70155299e-01 -2.45465606e-01 -4.15419221e-01 1.85813978e-01
8.01743031e-01 -8.43223214e-01 1.12477612e+00 -7.42072701e-01
3.12152952e-01 8.23272988e-02 -4.00132835e-02 -1.08808422e+00
-3.99887025e-01 -5.11044085e-01 -7.14613676e-01 8.44916344e-01
6.33277059e-01 -3.17834884e-01 9.26996291e-01 8.84030759e-01
1.06407097e-02 -7.76598454e-01 -4.24546689e-01 -4.93000031e-01
-6.49034798e-01 -3.55845779e-01 6.82110369e-01 1.37967259e-01
-7.72579610e-02 7.42962480e-01 -6.31446719e-01 -3.60798649e-02
4.33263391e-01 6.45857573e-01 6.79046690e-01 -1.16865897e+00
-6.28272057e-01 -1.47840247e-01 2.32816026e-01 -2.47354698e+00
3.41392234e-02 -9.90333676e-01 7.85467997e-02 -1.62830031e+00
8.36484373e-01 -2.22953349e-01 -3.17795992e-01 2.57855564e-01
-4.64151412e-01 -4.32669610e-01 -2.55771279e-01 6.63727760e-01
-1.06456614e+00 6.19612396e-01 1.06766319e+00 -1.97111651e-01
-2.29507208e-01 6.06318653e-01 -9.94364798e-01 1.34075984e-01
3.31307739e-01 -2.95462161e-01 -9.17919934e-01 -6.10487342e-01
5.69934070e-01 3.58889222e-01 1.24751024e-01 -4.67428774e-01
6.14476979e-01 3.20300311e-02 2.20442563e-01 -8.37313592e-01
4.27444607e-01 -6.84934199e-01 -1.34807020e-01 4.38687474e-01
-1.19754052e+00 -6.54373765e-02 -4.76524144e-01 1.06983709e+00
-4.71849471e-01 -1.95133507e-01 -8.85618255e-02 -3.96564230e-02
-8.11895847e-01 5.70759296e-01 -2.18039349e-01 9.17805433e-02
4.75499362e-01 5.47088608e-02 -2.90491700e-01 -1.05699611e+00
-6.72529936e-01 5.89688540e-01 6.21692128e-02 9.35591698e-01
5.48026681e-01 -1.26692665e+00 -2.07499474e-01 4.39269602e-01
3.24182481e-01 -4.98507559e-01 -8.07376858e-03 1.90114111e-01
1.80746317e-01 1.19535589e+00 5.19404650e-01 -4.41272914e-01
-1.19224918e+00 3.62502337e-01 9.00984183e-02 -8.11446607e-01
-9.03587863e-02 7.05772102e-01 3.28684270e-01 -1.64944321e-01
6.82705104e-01 -1.67751670e-01 -2.89529830e-01 -1.96522340e-01
6.26064301e-01 1.36435062e-01 -2.06999704e-02 -7.09985346e-02
-1.04138717e-01 1.66739032e-01 -7.64915705e-01 -1.41277641e-01
1.12335277e+00 -3.80998284e-01 2.09593356e-01 3.48115623e-01
1.59830749e+00 -1.95791945e-01 -1.01605582e+00 -6.25027835e-01
2.14513391e-01 -6.11920536e-01 -5.48083149e-02 -1.09043276e+00
-7.65849710e-01 8.04109871e-01 3.27478617e-01 3.48652303e-01
1.15733087e+00 2.72385389e-01 1.33138990e+00 8.12454045e-01
5.29004335e-01 -9.94578600e-01 3.22570443e-01 6.34024143e-01
6.73751295e-01 -1.16762793e+00 -2.73629397e-01 -2.31370628e-01
-6.80378616e-01 8.52367342e-01 5.95875204e-01 4.04775590e-01
6.56494796e-01 -2.57760435e-01 -2.47130111e-01 -1.69754609e-01
-1.40864801e+00 -3.86905074e-01 7.13853121e-01 1.30772635e-01
7.35761583e-01 -1.83355689e-01 -3.61194909e-01 6.45188451e-01
-3.03858370e-01 1.14350699e-01 -6.42130598e-02 9.19215739e-01
-5.21859765e-01 -1.46263659e+00 4.42978442e-01 1.05349088e+00
-3.08045924e-01 -4.20667529e-01 -4.65316236e-01 1.75630957e-01
-5.35100639e-01 9.99180615e-01 1.10445514e-01 -7.37718046e-01
2.98460335e-01 5.50571501e-01 -1.36667162e-01 -7.48280942e-01
-4.09510225e-01 -1.38448726e-03 2.12072790e-01 -1.04269874e+00
4.95596044e-02 -3.95492077e-01 -6.69740260e-01 -4.03295942e-02
-2.68574566e-01 6.88640058e-01 6.30247056e-01 7.45413423e-01
8.04920256e-01 4.02436465e-01 9.83469605e-01 -7.07514733e-02
-9.75889623e-01 -1.14928126e+00 -2.81306922e-01 6.19944990e-01
2.13788405e-01 -4.98185575e-01 -5.52746356e-01 9.47303325e-02] | [11.741608619689941, 7.641104698181152] |
bb601f03-be83-4ad8-b1b1-56af61d63894 | towards-multi-sense-cross-lingual-alignment-1 | 2103.06459 | null | https://arxiv.org/abs/2103.06459v4 | https://arxiv.org/pdf/2103.06459v4.pdf | Towards Multi-Sense Cross-Lingual Alignment of Contextual Embeddings | Cross-lingual word embeddings (CLWE) have been proven useful in many cross-lingual tasks. However, most existing approaches to learn CLWE including the ones with contextual embeddings are sense agnostic. In this work, we propose a novel framework to align contextual embeddings at the sense level by leveraging cross-lingual signal from bilingual dictionaries only. We operationalize our framework by first proposing a novel sense-aware cross entropy loss to model word senses explicitly. The monolingual ELMo and BERT models pretrained with our sense-aware cross entropy loss demonstrate significant performance improvement for word sense disambiguation tasks. We then propose a sense alignment objective on top of the sense-aware cross entropy loss for cross-lingual model pretraining, and pretrain cross-lingual models for several language pairs (English to German/Spanish/Japanese/Chinese). Compared with the best baseline results, our cross-lingual models achieve 0.52%, 2.09% and 1.29% average performance improvements on zero-shot cross-lingual NER, sentiment classification and XNLI tasks, respectively. | ['Luo Si', 'Lidong Bing', 'Shafiq Joty', 'Thien Hai Nguyen', 'Linlin Liu'] | 2021-03-11 | towards-multi-sense-cross-lingual-alignment | https://aclanthology.org/2022.coling-1.386 | https://aclanthology.org/2022.coling-1.386.pdf | coling-2022-10 | ['cross-lingual-ner'] | ['natural-language-processing'] | [-3.98488268e-02 -3.65015209e-01 -3.91940922e-01 -5.18975258e-01
-1.11589587e+00 -8.24639678e-01 4.46177185e-01 4.69127476e-01
-1.15283537e+00 6.81342125e-01 3.29445899e-01 -3.44528705e-01
1.50989264e-01 -6.26565874e-01 -4.33132023e-01 -4.67131793e-01
3.71345371e-01 1.99726075e-01 -1.79705322e-01 -6.20222807e-01
-6.04383424e-02 -1.08015768e-01 -1.08599925e+00 6.35864064e-02
1.19187641e+00 7.86120653e-01 4.02314812e-01 1.12496063e-01
-3.98070276e-01 1.48599735e-02 -1.53637335e-01 -6.80234253e-01
1.83434322e-01 -1.81393147e-01 -6.88283622e-01 -5.16883492e-01
4.58165050e-01 5.64976096e-01 1.55376762e-01 1.36054492e+00
6.89087331e-01 1.45320043e-01 4.02836531e-01 -7.84519970e-01
-1.23789155e+00 9.71233547e-01 -6.08081520e-01 3.06355327e-01
2.19644725e-01 -1.97649494e-01 1.62399530e+00 -1.25622022e+00
5.37775099e-01 1.20421660e+00 9.18350935e-01 5.02910256e-01
-1.25710166e+00 -7.85247445e-01 3.99523169e-01 3.68162304e-01
-1.53387737e+00 1.66204758e-03 7.41216660e-01 -1.89042583e-01
1.36981285e+00 -1.19243354e-01 4.55417484e-01 1.44418824e+00
3.75822693e-01 6.87938869e-01 1.31363153e+00 -7.29216576e-01
-3.58273927e-03 2.77901113e-01 4.70049858e-01 3.97845566e-01
4.04960454e-01 3.56472954e-02 -5.16226888e-01 5.10993227e-02
1.60666034e-01 -1.34109363e-01 -2.98352377e-03 -2.58973718e-01
-1.14743471e+00 9.29211080e-01 3.31334144e-01 7.42396235e-01
-1.93864942e-01 -1.66833282e-01 7.30354011e-01 3.62791359e-01
7.16018140e-01 7.23916769e-01 -9.78142381e-01 7.59858754e-04
-6.48954272e-01 -2.24158168e-01 4.24113452e-01 9.30759072e-01
1.00333643e+00 1.36155203e-01 -3.62396315e-02 1.40496647e+00
3.80987853e-01 7.50940025e-01 9.89941597e-01 -2.10176781e-02
3.25360835e-01 4.84455347e-01 -2.86690354e-01 -7.17486203e-01
-3.04978728e-01 -6.07779443e-01 -5.85034907e-01 -4.58693147e-01
-2.76815563e-01 -4.35462147e-01 -8.10017228e-01 2.11447573e+00
2.41510227e-01 3.84346336e-01 4.48916703e-01 6.72891140e-01
4.47699964e-01 4.33862925e-01 5.18196762e-01 -1.93885770e-02
1.78359413e+00 -9.59486306e-01 -8.85914266e-01 -6.52219296e-01
8.37114334e-01 -1.06016016e+00 1.66679370e+00 1.85418844e-01
-5.91187477e-01 -6.17311537e-01 -1.33094561e+00 -3.91554773e-01
-9.52489197e-01 2.80370593e-01 4.99081910e-01 5.79245627e-01
-7.29146361e-01 1.59462005e-01 -6.84669137e-01 -5.02173781e-01
-1.09507471e-01 6.50181547e-02 -4.10829931e-01 -1.83564037e-01
-1.86473095e+00 1.20626831e+00 7.21878707e-01 -1.74133033e-01
-3.37132156e-01 -9.81002450e-01 -1.24106348e+00 -1.15748264e-01
2.89559633e-01 -5.13250947e-01 8.92551482e-01 -8.24490070e-01
-9.87267554e-01 1.02430463e+00 -2.20458984e-01 -4.86891419e-01
-9.38855633e-02 -6.86672270e-01 -9.55496073e-01 -4.34334338e-01
6.33141577e-01 4.84903395e-01 -1.95393600e-02 -1.12370086e+00
-7.63062894e-01 -2.79759139e-01 -3.09011750e-02 3.30543250e-01
-6.26676559e-01 -4.58982401e-02 -3.71919423e-01 -1.08850586e+00
-2.70667821e-01 -9.54566121e-01 -2.56529540e-01 -6.21873438e-01
-2.42313981e-01 -6.25284791e-01 5.08455515e-01 -3.97956103e-01
1.37615180e+00 -2.06300664e+00 -1.65183783e-01 -1.14226855e-01
-4.08350974e-01 4.62926507e-01 -4.40036237e-01 4.60276753e-01
-2.08750844e-01 2.54165351e-01 -3.86042625e-01 -6.06623411e-01
3.50939244e-01 4.99708802e-01 -2.96805352e-01 2.28083460e-03
3.55429530e-01 8.62056494e-01 -1.25285232e+00 -3.06479812e-01
1.54189900e-01 7.98753500e-01 -7.06883132e-01 -4.49023172e-02
6.69964775e-02 -4.86013219e-02 -2.50850827e-01 4.18087035e-01
5.47284484e-01 1.92047298e-01 6.16970003e-01 -4.37639654e-01
-1.29197136e-01 7.25408375e-01 -8.70871902e-01 2.21928644e+00
-1.26570499e+00 3.35512787e-01 -2.88986951e-01 -7.11158216e-01
8.80191207e-01 3.84543121e-01 3.40859294e-01 -7.92941093e-01
7.80723542e-02 4.01094973e-01 -1.10950477e-01 -2.59458989e-01
8.38384449e-01 -6.58463597e-01 -6.75847709e-01 3.22284669e-01
4.96050447e-01 8.49071518e-02 3.06999117e-01 -2.23441303e-01
6.24394774e-01 8.53640810e-02 5.93555331e-01 -7.30959356e-01
5.74253321e-01 -6.59045875e-02 8.25270236e-01 4.37366724e-01
-1.58961445e-01 3.71876538e-01 2.49401256e-02 -3.19395036e-01
-7.89031982e-01 -1.27280402e+00 -3.71783704e-01 1.39489257e+00
3.93342674e-01 -7.23385155e-01 -5.59857607e-01 -8.73280168e-01
-4.46202829e-02 9.84858572e-01 -5.75638115e-01 -1.94017664e-01
-4.55220133e-01 -1.01683354e+00 6.27785921e-01 6.10514820e-01
5.03077984e-01 -8.55627835e-01 8.48246664e-02 3.71277332e-01
-1.10838883e-01 -1.46546829e+00 -9.72421050e-01 4.64032412e-01
-3.79815727e-01 -6.99540257e-01 -1.32501960e-01 -1.01848924e+00
1.92157835e-01 -3.37560624e-02 1.36245918e+00 -3.43734205e-01
-1.67320997e-01 2.00107530e-01 -6.76262856e-01 -4.46671039e-01
1.48507044e-01 4.76704419e-01 6.75730228e-01 -6.08565621e-02
1.22240567e+00 -5.61465442e-01 -4.34560090e-01 3.17012309e-05
-9.63997781e-01 -5.99076688e-01 5.93398094e-01 1.14974058e+00
1.13314474e+00 -4.68028933e-01 7.06210494e-01 -1.13069630e+00
8.97659957e-01 -5.56258500e-01 -3.50609839e-01 4.56098229e-01
-1.01909852e+00 5.29501140e-01 6.08956873e-01 -3.38380456e-01
-1.12547064e+00 -8.49449858e-02 -4.88731980e-01 -1.82645589e-01
-1.17071472e-01 6.76811159e-01 -3.41546535e-01 3.69226605e-01
5.24884760e-01 1.11579835e-01 -7.88443387e-01 -5.94260216e-01
7.30044484e-01 6.80912495e-01 4.15570855e-01 -7.90334105e-01
4.41768855e-01 2.17442811e-02 -7.14755058e-01 -5.51558137e-01
-1.44603980e+00 -7.53094792e-01 -9.21699703e-01 5.64569771e-01
1.31870759e+00 -1.11015868e+00 -6.62972853e-02 3.86096835e-02
-1.14420176e+00 9.82845947e-02 -2.32076898e-01 7.11982667e-01
-6.22015521e-02 2.58649588e-01 -2.86863685e-01 -3.74141186e-01
-6.31833613e-01 -1.07818067e+00 1.07332647e+00 4.11102315e-03
-3.21483314e-01 -1.62569666e+00 6.09101534e-01 9.78790149e-02
4.65332568e-01 -3.11553143e-02 8.63497913e-01 -9.82609630e-01
1.59959197e-01 1.05159588e-01 -7.74706155e-02 6.56357765e-01
4.69208091e-01 -5.69312036e-01 -9.72370923e-01 -3.89642030e-01
-2.24092305e-01 -3.70063454e-01 1.01763558e+00 -2.53943503e-02
7.18411565e-01 -8.45652744e-02 -3.12648207e-01 7.21592426e-01
1.93658638e+00 9.82400477e-02 7.97388554e-02 3.54753703e-01
8.19853783e-01 1.35111824e-01 7.28065848e-01 1.01267420e-01
8.15778136e-01 6.38154626e-01 -1.25558853e-01 -9.53078791e-02
-1.15871117e-01 -3.54186207e-01 5.17245650e-01 1.63129473e+00
4.06745851e-01 -1.29895642e-01 -9.67688680e-01 1.19884408e+00
-1.49858701e+00 -5.94033182e-01 2.14096457e-01 1.98302042e+00
1.26803970e+00 9.58379358e-02 -3.47724646e-01 -3.49508703e-01
5.21480262e-01 2.72977144e-01 -3.70257288e-01 -9.07599807e-01
-3.81814748e-01 7.45604277e-01 6.51157320e-01 8.34100306e-01
-1.34990609e+00 1.52400839e+00 5.15895844e+00 1.02656209e+00
-1.14469516e+00 6.40760481e-01 8.96818340e-02 6.80347830e-02
-6.97171211e-01 5.52898347e-02 -1.10642219e+00 3.64673525e-01
8.73843014e-01 -3.39071453e-01 1.21490836e-01 8.59996319e-01
-3.50501686e-01 3.39280188e-01 -1.04693663e+00 1.07981348e+00
3.00537318e-01 -8.94224524e-01 5.63712157e-02 -2.43479565e-01
1.07233465e+00 6.13211572e-01 1.39800766e-02 6.90634608e-01
7.29812086e-01 -8.38691294e-01 4.00012404e-01 6.37876466e-02
9.42523479e-01 -1.12658072e+00 1.05800867e+00 -1.74224377e-01
-1.51760626e+00 3.33229572e-01 -5.57159662e-01 1.91637665e-01
2.91675925e-01 6.87240303e-01 -5.71129441e-01 8.87854815e-01
5.61265528e-01 1.04235625e+00 -4.90348101e-01 3.57669264e-01
-5.39453685e-01 5.86165965e-01 -1.84731781e-01 6.40641004e-02
4.01070833e-01 -1.46488741e-01 3.91300917e-01 1.69632626e+00
3.70343417e-01 -1.60437599e-01 4.11372691e-01 6.29215598e-01
-2.76819617e-01 4.59084302e-01 -5.28051972e-01 -5.33722267e-02
7.64806688e-01 1.11357057e+00 -1.29942805e-01 -3.26037109e-01
-6.05579317e-01 1.10783052e+00 5.24626791e-01 2.30565026e-01
-6.15321994e-01 -7.44399190e-01 1.47930372e+00 -4.97948110e-01
3.24676275e-01 -3.34040612e-01 -3.77883464e-01 -1.50737667e+00
5.88608198e-02 -6.00401878e-01 5.66631913e-01 -3.11352640e-01
-1.90003049e+00 1.01545012e+00 -1.57456219e-01 -1.11682808e+00
-2.52372682e-01 -1.01291108e+00 -3.66210341e-01 9.60144639e-01
-2.15095353e+00 -1.35470450e+00 3.09536487e-01 3.17641765e-01
5.72217822e-01 -2.96481431e-01 1.13514519e+00 6.55264437e-01
-3.98227721e-01 1.20903945e+00 1.47216111e-01 2.77307361e-01
1.08726275e+00 -1.55927241e+00 4.84547377e-01 9.11802828e-01
5.79834402e-01 1.06877720e+00 4.73001361e-01 -4.08923030e-01
-1.12099528e+00 -1.28605247e+00 1.53002095e+00 -6.64667130e-01
1.03789830e+00 -5.92481613e-01 -8.40415835e-01 8.02968800e-01
7.95970678e-01 7.39980340e-02 1.14979279e+00 6.93461180e-01
-7.95921028e-01 -2.38582745e-01 -8.32136989e-01 6.68169022e-01
1.12572396e+00 -9.45494592e-01 -1.04391193e+00 1.59484427e-02
1.07575703e+00 -5.32750227e-02 -1.14065206e+00 3.41594666e-01
4.74238783e-01 -5.13918400e-01 9.92647588e-01 -6.36463404e-01
1.64468899e-01 -2.35238656e-01 -5.56290984e-01 -1.84355962e+00
-1.04187600e-01 -2.94865489e-01 6.61848426e-01 1.41171432e+00
6.60891533e-01 -7.46704280e-01 1.88165195e-02 -2.25481048e-01
-3.46414864e-01 -7.23791063e-01 -9.90211964e-01 -9.98450160e-01
7.55343676e-01 -8.43775034e-01 6.18596077e-01 1.49754500e+00
1.63737312e-01 8.03157866e-01 -2.87707955e-01 2.25869358e-01
4.46913987e-01 1.10410154e-01 6.41802698e-02 -9.09034312e-01
-9.02322382e-02 -4.41308677e-01 -3.23791325e-01 -1.02362132e+00
6.50252104e-01 -1.40311456e+00 8.87911469e-02 -1.31414175e+00
1.31439880e-01 -4.65658218e-01 -1.08708596e+00 7.65270650e-01
-6.76057279e-01 2.50131130e-01 1.00634828e-01 -4.28546816e-01
-5.96796811e-01 7.36351609e-01 7.78498828e-01 -1.48212031e-01
1.06146321e-01 -7.87438035e-01 -8.39478254e-01 6.09930813e-01
7.89589345e-01 -7.27836072e-01 -3.87947083e-01 -8.64495218e-01
4.24557388e-01 -6.90245688e-01 -1.49437279e-01 -7.49899149e-01
7.58297816e-02 -8.41496512e-02 -7.26991296e-02 -2.73295790e-01
1.77090436e-01 -7.77253211e-01 -4.52818334e-01 4.68028933e-02
-3.86139661e-01 5.24328232e-01 4.23929095e-01 4.75499868e-01
-6.31259203e-01 -9.07262266e-02 6.04032040e-01 8.51274654e-02
-1.07403421e+00 1.44094199e-01 9.42143891e-03 6.52633667e-01
7.51770437e-01 1.95385188e-01 2.16456857e-02 3.10414076e-01
-6.72486365e-01 3.88935119e-01 3.52246106e-01 1.08461571e+00
2.53227472e-01 -1.82862890e+00 -7.22384810e-01 3.80052805e-01
7.12385237e-01 -4.35700804e-01 1.17846325e-01 5.78264892e-01
1.45964012e-01 5.57609081e-01 -4.56743017e-02 -3.44802350e-01
-1.05485976e+00 5.80501616e-01 2.30302528e-01 -5.00072002e-01
4.02934030e-02 1.10075080e+00 3.08869988e-01 -1.05409658e+00
-2.61551499e-01 -4.95085180e-01 -8.22350979e-02 2.11784735e-01
3.23953688e-01 -1.97541923e-03 3.04345191e-01 -8.09228420e-01
-7.42876291e-01 1.00055790e+00 -1.87960833e-01 2.36625839e-02
1.14703500e+00 -2.53657818e-01 -2.61685867e-02 6.90911651e-01
1.47969198e+00 4.96673048e-01 -6.83305442e-01 -6.01494133e-01
4.40330714e-01 -7.91120604e-02 -1.21162636e-02 -9.59956050e-01
-8.81247163e-01 1.09258485e+00 7.46339917e-01 -1.80464268e-01
1.11590695e+00 4.85355668e-02 1.16209698e+00 2.14470327e-01
5.04333258e-01 -1.35188591e+00 -3.18300664e-01 1.02798200e+00
5.12351990e-01 -1.38955307e+00 -4.91170734e-01 -7.88591355e-02
-8.09562385e-01 7.46321023e-01 5.64992785e-01 -2.54100323e-01
8.30989540e-01 3.36990356e-01 5.24099231e-01 -8.21041837e-02
-8.59887362e-01 -6.16208315e-01 5.98160088e-01 4.38997179e-01
9.34429049e-01 3.31246495e-01 -8.40741158e-01 1.00843465e+00
-4.69220787e-01 -3.58255774e-01 2.06072237e-02 4.76410389e-01
-2.30129316e-01 -1.82301593e+00 2.04475269e-01 -6.24083318e-02
-7.20734060e-01 -8.17532182e-01 -2.03992158e-01 6.44265234e-01
4.67523932e-01 8.72439086e-01 1.05707183e-01 -5.30314088e-01
4.46539015e-01 4.45806623e-01 2.07957625e-01 -8.98446679e-01
-5.55084467e-01 -1.04413465e-01 1.27708718e-01 -3.20239246e-01
-3.94913644e-01 -5.33515215e-01 -1.26487911e+00 2.25635961e-01
-2.67647386e-01 3.20608467e-01 4.51237947e-01 1.12653017e+00
5.27217925e-01 4.66214329e-01 4.60711598e-01 -7.45868608e-02
-4.29415673e-01 -1.08711720e+00 -4.91857827e-01 6.90645993e-01
1.61264583e-01 -7.24040508e-01 -2.31623128e-01 -7.16617554e-02] | [10.887657165527344, 9.771105766296387] |
af055459-5ee7-4b5d-bd95-3ca2f2f5198b | explicit3d-graph-network-with-spatial | 2302.06494 | null | https://arxiv.org/abs/2302.06494v1 | https://arxiv.org/pdf/2302.06494v1.pdf | Explicit3D: Graph Network with Spatial Inference \\for Single Image 3D Object Detection | Indoor 3D object detection is an essential task in single image scene understanding, impacting spatial cognition fundamentally in visual reasoning. Existing works on 3D object detection from a single image either pursue this goal through independent predictions of each object or implicitly reason over all possible objects, failing to harness relational geometric information between objects. To address this problem, we propose a dynamic sparse graph pipeline named Explicit3D based on object geometry and semantics features. Taking the efficiency into consideration, we further define a relatedness score and design a novel dynamic pruning algorithm followed by a cluster sampling method for sparse scene graph generation and updating. Furthermore, our Explicit3D introduces homogeneous matrices and defines new relative loss and corner loss to model the spatial difference between target pairs explicitly. Instead of using ground-truth labels as direct supervision, our relative and corner loss are derived from the homogeneous transformation, which renders the model to learn the geometric consistency between objects. The experimental results on the SUN RGB-D dataset demonstrate that our Explicit3D achieves better performance balance than the-state-of-the-art. | ['Wenming Yang', 'Qingmin Liao', 'Yehu Shen', 'Yanjun Liu'] | 2023-02-13 | null | null | null | null | ['scene-graph-generation', 'visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'computer-vision', 'reasoning'] | [ 6.67009279e-02 2.18619421e-01 7.26143597e-04 -5.14330983e-01
-1.65225416e-01 -4.27961975e-01 6.29325211e-01 3.31299365e-01
-2.05162883e-01 1.26939967e-01 4.40851003e-02 -2.04002172e-01
-2.16400594e-01 -8.27817559e-01 -9.29073036e-01 -4.77403939e-01
3.15906629e-02 4.83963579e-01 7.16630101e-01 -4.86186184e-02
3.31553459e-01 6.91912949e-01 -1.54388201e+00 1.45073801e-01
8.16209733e-01 1.17373776e+00 6.45314634e-01 3.36804181e-01
-1.93624958e-01 9.14448082e-01 -2.19475999e-01 -2.96372771e-01
6.03942752e-01 -1.06466413e-01 -5.45129359e-01 3.46310914e-01
7.67600238e-01 -3.42749953e-01 -5.52230120e-01 1.13254976e+00
3.36909920e-01 2.09843710e-01 5.75271368e-01 -1.28730905e+00
-7.20763028e-01 3.25004905e-01 -7.13589370e-01 4.62215766e-03
4.39870328e-01 2.30532080e-01 1.04429233e+00 -9.27889824e-01
6.77743375e-01 1.35768783e+00 5.76062620e-01 1.43439919e-01
-1.27559543e+00 -3.92746180e-01 8.18212450e-01 3.87353837e-01
-1.58634973e+00 -2.40101308e-01 1.11029971e+00 -4.65195835e-01
1.09371865e+00 3.39877129e-01 9.30119693e-01 8.47380459e-01
-3.37768018e-01 7.46408880e-01 9.93429184e-01 -3.16423178e-01
3.91750664e-01 8.26268941e-02 1.97133079e-01 9.90614533e-01
5.00395536e-01 -5.20265754e-03 -5.84283173e-01 9.93958041e-02
8.64608049e-01 1.75350800e-01 -2.02266976e-01 -1.22580409e+00
-1.13403308e+00 4.83639002e-01 9.99699831e-01 -1.67292997e-01
-1.31341219e-01 1.70436621e-01 -1.24619953e-01 -1.96500689e-01
2.28258118e-01 2.16561824e-01 -1.52588263e-01 3.93269747e-01
-6.17985785e-01 4.49302703e-01 6.37038350e-01 1.37844515e+00
1.04358673e+00 -2.41378143e-01 -2.77126431e-01 5.00636220e-01
4.84919757e-01 6.18928790e-01 -3.11216027e-01 -9.95400429e-01
5.83811581e-01 1.13604105e+00 -5.36368713e-02 -1.51624417e+00
-4.56135392e-01 -5.82405210e-01 -7.14558244e-01 2.30943024e-01
2.39153177e-01 6.67151928e-01 -1.14405966e+00 1.55466354e+00
7.52199531e-01 4.16069739e-02 -4.26201433e-01 1.35391724e+00
9.01030660e-01 2.74833530e-01 -9.52544883e-02 2.13044494e-01
1.09659338e+00 -1.07209110e+00 -2.66312301e-01 -3.95699263e-01
5.25977135e-01 -5.21513164e-01 1.14204144e+00 1.16177157e-01
-9.31725144e-01 -4.78025496e-01 -9.86748695e-01 -5.33695161e-01
-4.82397050e-01 8.89196321e-02 9.77497876e-01 4.86158133e-01
-8.42899919e-01 1.27594486e-01 -8.37563872e-01 -3.84048164e-01
5.77653706e-01 2.60382444e-01 -3.60812902e-01 -3.81599724e-01
-5.23094833e-01 8.69007707e-01 4.23761874e-01 4.00438113e-03
-8.45611989e-01 -7.95442164e-01 -9.25206721e-01 -3.01929973e-02
6.46506906e-01 -1.20128763e+00 7.74322093e-01 -4.20806825e-01
-1.18579566e+00 1.05077851e+00 -6.30846545e-02 -3.51436555e-01
5.54836154e-01 -2.52522022e-01 1.79189444e-01 1.49619788e-01
2.98879027e-01 7.20963180e-01 6.79558456e-01 -1.64636803e+00
-4.78932559e-01 -5.81787288e-01 4.51266736e-01 5.64965308e-01
3.11400220e-02 -5.54865122e-01 -7.54733384e-01 -3.43525052e-01
7.66822040e-01 -9.32359576e-01 -4.44050789e-01 2.99610883e-01
-6.92932427e-01 -1.41491428e-01 6.00632071e-01 -3.21823686e-01
7.93727934e-01 -2.19971347e+00 2.84953773e-01 4.28067684e-01
5.14823854e-01 -2.01332331e-01 -9.14192572e-02 5.59779182e-02
2.05546722e-01 -1.96036756e-01 -2.07404986e-01 -5.27300894e-01
7.55413175e-02 1.17569610e-01 -2.87857115e-01 5.34117579e-01
2.53569394e-01 9.22201693e-01 -1.16878581e+00 -5.71339488e-01
5.67337513e-01 4.14198965e-01 -9.36616004e-01 1.85391486e-01
-4.95463997e-01 2.43574798e-01 -7.44299352e-01 7.98904538e-01
9.96989310e-01 -5.05421460e-01 -7.49261603e-02 -4.27542537e-01
-5.09527959e-02 4.56415325e-01 -1.30424714e+00 2.34912872e+00
-7.73085803e-02 1.02031872e-01 -7.42479339e-02 -8.34644794e-01
8.07402074e-01 -4.64145124e-01 2.90832937e-01 -6.26335204e-01
1.62400305e-02 -2.48307036e-03 -2.41572335e-01 -2.72236615e-01
3.81883204e-01 3.34799588e-01 1.59135237e-01 1.32568881e-01
4.02664095e-02 -5.70259869e-01 -5.52785732e-02 5.00909388e-01
1.16518307e+00 5.80460310e-01 1.00934200e-01 -1.24434568e-01
3.23539525e-01 2.04774007e-01 3.93360794e-01 9.78157997e-01
4.94885680e-05 7.49349833e-01 2.64956772e-01 -4.09480095e-01
-8.02054703e-01 -1.40916109e+00 7.04419911e-02 6.58094227e-01
8.80283952e-01 -5.28965592e-01 -4.19080228e-01 -7.63425171e-01
2.64403462e-01 6.94804907e-01 -5.03825903e-01 -1.44062012e-01
-3.25343519e-01 -5.97692370e-01 -8.92374516e-02 2.80371010e-01
6.45773053e-01 -4.67867851e-01 -7.43162751e-01 -1.73906535e-01
-1.82714984e-01 -1.31287289e+00 -2.26119101e-01 2.46766925e-01
-7.10936666e-01 -1.09359896e+00 -2.43457898e-01 -5.94228864e-01
9.31700826e-01 9.55408514e-01 1.09657228e+00 2.17655711e-02
-5.97621083e-01 6.46110833e-01 -3.43895376e-01 -3.00836265e-01
2.47712851e-01 1.12020068e-01 -6.12572804e-02 3.44783952e-03
2.87787318e-01 -8.17857265e-01 -8.46129596e-01 9.42050666e-02
-4.21186686e-01 4.97841805e-01 7.16667652e-01 3.82365346e-01
8.61662090e-01 -1.61745176e-01 -1.83724731e-01 -6.01360500e-01
-1.19024098e-01 -4.08348560e-01 -7.89616287e-01 3.17135751e-01
-4.76010114e-01 1.22992553e-01 1.26413777e-01 -1.94148228e-01
-1.06748652e+00 4.65374887e-01 3.81107450e-01 -8.75038087e-01
-1.15657084e-01 7.80803859e-02 -3.76840800e-01 -3.17826092e-01
6.10290110e-01 2.65012264e-01 -3.28890800e-01 -4.96370733e-01
7.30090857e-01 -7.76435714e-03 4.36560243e-01 -5.09836495e-01
1.12397480e+00 8.15696657e-01 4.06028599e-01 -8.40456605e-01
-1.21737802e+00 -6.53052568e-01 -7.61577427e-01 -3.19971472e-01
9.29207683e-01 -1.19448245e+00 -7.77058542e-01 1.84651613e-01
-1.25299239e+00 -2.75674850e-01 -2.68100560e-01 4.89883989e-01
-5.22399664e-01 2.21385077e-01 -1.45486906e-01 -8.35480452e-01
2.26727635e-01 -8.29595566e-01 1.39381194e+00 -4.66697551e-02
3.87353264e-02 -5.55902481e-01 -2.25652888e-01 3.19028407e-01
4.01531719e-02 3.34198296e-01 1.00117290e+00 -1.75031245e-01
-1.58534849e+00 4.90943044e-02 -6.46630824e-01 -1.09164946e-01
9.79285091e-02 -4.19494063e-01 -9.14275527e-01 7.56302625e-02
9.67741385e-02 -1.13215186e-01 9.80126381e-01 2.06043079e-01
1.36051333e+00 -6.63060695e-02 -4.70729291e-01 9.65438247e-01
1.45645165e+00 -3.19843292e-01 3.20032537e-01 3.05079520e-01
1.29227352e+00 6.49329185e-01 4.81497318e-01 3.96704406e-01
7.94677973e-01 6.62029147e-01 8.21395636e-01 -8.52256641e-02
-5.32952607e-01 -6.36693418e-01 -7.39831477e-02 4.82016772e-01
-1.07273459e-01 -2.38562003e-02 -1.01082861e+00 3.93114686e-01
-1.90412688e+00 -7.78665066e-01 -3.20330232e-01 2.19919538e+00
5.44327796e-01 2.75455236e-01 -1.87501460e-01 -6.78022653e-02
3.94259721e-01 1.74564794e-01 -6.93367600e-01 4.33415234e-01
-3.31946343e-01 -1.45408690e-01 6.69731081e-01 6.39813066e-01
-9.10381436e-01 1.19493115e+00 5.16934156e+00 5.28448462e-01
-9.65991914e-01 -3.27862799e-02 3.78294200e-01 -3.40848833e-01
-4.01993006e-01 3.79491776e-01 -9.73211467e-01 -4.24822904e-02
-2.44893897e-02 7.89012760e-02 5.30852318e-01 1.00544202e+00
1.02539649e-02 -2.97991276e-01 -1.22598767e+00 1.25479662e+00
2.17289641e-01 -1.22192585e+00 4.16155308e-01 1.01033166e-01
6.73989832e-01 1.03730924e-01 2.92218905e-02 -2.38416754e-02
4.09860700e-01 -7.04141140e-01 1.28344870e+00 6.02356672e-01
2.57414699e-01 -4.09607083e-01 3.44412588e-02 4.26592320e-01
-1.20390713e+00 1.06680917e-03 -5.48337936e-01 -2.74372458e-01
3.35583836e-03 8.42687547e-01 -9.27492678e-01 6.34422779e-01
8.54083717e-01 8.38345349e-01 -9.56451654e-01 1.12848783e+00
-2.79933512e-01 7.94857740e-02 -5.85973918e-01 -1.32374484e-02
8.27899650e-02 -2.70416141e-01 7.49822438e-01 7.97880948e-01
1.21097751e-01 2.81104565e-01 3.98297876e-01 1.26610327e+00
1.29791647e-01 -1.03629850e-01 -6.88309848e-01 3.13904792e-01
4.33143437e-01 1.18441045e+00 -1.03257251e+00 -1.27763197e-01
-3.86129349e-01 1.07798386e+00 7.11565256e-01 4.18449402e-01
-7.79077291e-01 -1.32016400e-02 5.98627150e-01 3.79889011e-01
4.97535229e-01 -6.10615969e-01 -5.96484721e-01 -1.26760840e+00
4.29333121e-01 -3.17243338e-01 6.66846037e-02 -9.30421710e-01
-1.25221348e+00 2.80215830e-01 2.55376726e-01 -1.19110703e+00
3.12401712e-01 -6.10845685e-01 -1.88659415e-01 6.61095858e-01
-1.69777155e+00 -1.46820486e+00 -7.39174843e-01 6.14789426e-01
2.16414437e-01 2.82022953e-01 4.81527388e-01 6.59858286e-02
-2.74736494e-01 2.08870426e-01 -5.29624701e-01 -8.77328590e-02
3.81501317e-01 -1.13590920e+00 4.12206709e-01 8.63526404e-01
3.22537720e-01 6.82006776e-01 3.96438032e-01 -7.82347500e-01
-1.74275589e+00 -1.15001070e+00 7.30143249e-01 -8.01920950e-01
3.25776935e-01 -9.73091483e-01 -7.35180676e-01 4.91567045e-01
-3.42637450e-01 2.93742895e-01 2.63670027e-01 1.61958352e-01
-7.06935287e-01 -1.39937773e-01 -8.52353394e-01 8.41996074e-01
2.07691717e+00 -4.56777126e-01 -3.97766173e-01 4.67808098e-01
1.06134748e+00 -6.51886821e-01 -3.79080623e-01 5.20710766e-01
3.03346723e-01 -1.28468740e+00 1.33410215e+00 -3.53847891e-01
2.08187744e-01 -7.54807293e-01 -4.31055576e-01 -7.69757509e-01
-4.72702533e-01 -2.70676523e-01 -3.44550878e-01 9.49920535e-01
1.87572554e-01 -2.80376107e-01 9.82062936e-01 7.25304604e-01
-3.00324559e-01 -6.25382543e-01 -8.29114974e-01 -8.05530310e-01
-5.96529841e-01 -6.07075751e-01 6.12163842e-01 8.13135743e-01
-4.38098937e-01 3.65515530e-01 -6.19300604e-02 6.98654294e-01
1.13399160e+00 5.11282206e-01 1.10702789e+00 -1.21356940e+00
-2.07196459e-01 -5.56964397e-01 -7.34771073e-01 -1.56084490e+00
-3.10781728e-02 -1.08657730e+00 3.33226435e-02 -1.63012779e+00
2.19290763e-01 -7.42507339e-01 -1.12185687e-01 4.07657444e-01
-1.42718032e-01 4.07505296e-02 2.98891127e-01 9.21117589e-02
-9.91575062e-01 9.28505301e-01 1.27010226e+00 -2.39040792e-01
-3.55413049e-01 -3.76933455e-01 -5.48833668e-01 7.91324914e-01
2.96143711e-01 -3.91063213e-01 -6.42328203e-01 -7.73184717e-01
3.55664045e-01 -4.29109603e-01 9.13786829e-01 -1.13030767e+00
4.01693046e-01 -3.66968960e-01 6.18199885e-01 -8.87777388e-01
5.64137042e-01 -1.02872264e+00 -4.44057807e-02 2.66553998e-01
-1.08984463e-01 -3.61234069e-01 -4.86487597e-02 8.66199434e-01
2.10231498e-01 2.34604940e-01 5.50057352e-01 -2.76222974e-01
-9.88904417e-01 6.04311764e-01 3.45081717e-01 -3.08302101e-02
1.09740388e+00 -4.36061442e-01 -1.47389784e-01 -3.12613733e-02
-5.75591445e-01 3.44612777e-01 7.34017432e-01 5.54638326e-01
8.75711441e-01 -1.40445042e+00 -4.19781566e-01 3.78451258e-01
3.89421970e-01 6.48354053e-01 2.43418097e-01 7.22346246e-01
-3.95465791e-01 3.69006991e-01 1.23163432e-01 -1.07494545e+00
-9.37886655e-01 6.81662858e-01 3.26851785e-01 1.46073461e-01
-7.45271325e-01 1.13395524e+00 7.29593158e-01 -5.07036805e-01
4.15341377e-01 -5.66022098e-01 2.09922150e-01 -3.28801781e-01
5.24779335e-02 2.43667036e-01 4.69845049e-02 -4.83154595e-01
-5.42068958e-01 7.80916989e-01 9.54727829e-02 5.72888739e-02
1.28933859e+00 -3.41019422e-01 -1.15661792e-01 5.23337722e-01
9.20364797e-01 -1.13250464e-01 -1.44678104e+00 -3.79487306e-01
-5.98009862e-02 -9.09458101e-01 1.88721761e-01 -6.00968540e-01
-7.00070679e-01 6.81402385e-01 5.19202888e-01 -6.52069822e-02
9.06998396e-01 2.95457095e-01 3.30343634e-01 6.23646796e-01
7.18820453e-01 -7.24139392e-01 2.75170356e-01 5.01294553e-01
9.28231061e-01 -1.41214716e+00 2.65183806e-01 -9.68502164e-01
-2.43422106e-01 7.26892233e-01 8.72947991e-01 -2.24399105e-01
7.43702114e-01 -1.04834042e-01 -3.29381734e-01 -3.34510118e-01
-4.12925303e-01 -4.86123919e-01 5.35930037e-01 7.06533670e-01
-1.44461885e-01 -1.14036545e-01 2.13881984e-01 4.06029880e-01
-2.67382711e-01 -2.95855224e-01 -1.18278317e-01 9.73300993e-01
-3.83619517e-01 -8.33329797e-01 -2.29640797e-01 3.26089233e-01
3.02310348e-01 -1.56327590e-01 -5.49343705e-01 6.41932309e-01
2.72503078e-01 6.99604511e-01 1.49398586e-02 -3.93818945e-01
4.81882691e-01 -2.34725371e-01 9.10124362e-01 -8.63843143e-01
-2.69845817e-02 -2.05495253e-01 -1.99236020e-01 -8.79634798e-01
-5.17019093e-01 -5.84097624e-01 -1.22869456e+00 -4.72157933e-02
-3.32648456e-01 -3.25819463e-01 8.26574862e-01 7.87925959e-01
5.86046219e-01 3.78357857e-01 5.10915875e-01 -1.11574602e+00
-2.97124237e-01 -5.07389367e-01 -4.08126771e-01 4.74286288e-01
3.90622437e-01 -9.96251106e-01 -3.41307014e-01 -6.76992461e-02] | [8.199882507324219, -2.852332592010498] |
ac0cd072-c9af-408a-9593-f4d9a39bf9c3 | robust-end-to-end-focal-liver-lesion | 2112.01535 | null | https://arxiv.org/abs/2112.01535v2 | https://arxiv.org/pdf/2112.01535v2.pdf | Robust End-to-End Focal Liver Lesion Detection using Unregistered Multiphase Computed Tomography Images | The computer-aided diagnosis of focal liver lesions (FLLs) can help improve workflow and enable correct diagnoses; FLL detection is the first step in such a computer-aided diagnosis. Despite the recent success of deep-learning-based approaches in detecting FLLs, current methods are not sufficiently robust for assessing misaligned multiphase data. By introducing an attention-guided multiphase alignment in feature space, this study presents a fully automated, end-to-end learning framework for detecting FLLs from multiphase computed tomography (CT) images. Our method is robust to misaligned multiphase images owing to its complete learning-based approach, which reduces the sensitivity of the model's performance to the quality of registration and enables a standalone deployment of the model in clinical practice. Evaluation on a large-scale dataset with 280 patients confirmed that our method outperformed previous state-of-the-art methods and significantly reduced the performance degradation for detecting FLLs using misaligned multiphase CT images. The robustness of the proposed method can enhance the clinical adoption of the deep-learning-based computer-aided detection system. | ['Sungroh Yoon', 'Jung Hoon Kim', 'Jae Seok Bae', 'Eunji Kim', 'Sang-gil Lee'] | 2021-12-02 | null | null | null | null | ['medical-object-detection', 'liver-segmentation', 'automatic-liver-and-tumor-segmentation'] | ['computer-vision', 'medical', 'medical'] | [-0.09696203 -0.16188528 -0.06011811 -0.3366642 -1.4846491 -0.40395322
0.41549268 0.37339112 -0.34721008 0.01722383 0.30348513 -0.49722221
-0.28141147 -0.40678284 -0.238929 -0.8376237 -0.5511697 0.7230072
0.21980192 0.29344591 -0.10559823 0.7355355 -0.55845004 0.4162025
0.5422789 0.8419171 0.5322615 0.73175365 0.2812032 0.8605303
-0.22573411 0.27830535 0.47131443 -0.5839511 -0.67711127 0.14719902
0.57037604 -0.6011222 -0.45312902 0.727589 1.2253683 -0.42020556
0.5511882 -0.66038233 -0.22893676 0.3951468 -0.5025648 0.6937844
0.24675442 0.56018686 0.41076347 -0.9761644 0.297372 0.59395653
1.2992622 0.16689828 -0.937633 -0.45323244 -0.5859532 0.1101234
-1.3022548 -0.20424397 0.39141268 -0.7271558 0.9478296 0.13533327
1.0442551 0.27413136 0.74134636 0.7196377 1.1747563 -0.4911635
-0.19143939 -0.28496912 -0.15600105 1.1550955 0.46711883 0.3503754
-0.00772353 -0.31277153 0.9885453 0.07574724 -0.6149821 -0.85874134
-1.6130751 0.83699405 0.7813285 0.5510445 -0.7549686 -0.030619
0.5324443 -0.05184504 0.22879417 0.5387787 -0.2309122 0.11264758
-1.2088354 -0.2664644 0.5823685 0.4666336 -0.08329713 -0.01134455
-0.40533814 0.41696066 0.42418742 0.42205602 0.85237616 -0.32132125
-0.02841769 0.6164743 -0.29880404 -0.62732613 -1.2414871 -0.878645
-0.93124807 0.2311971 0.5061277 -0.24471807 -1.0874834 0.9776374
0.23354414 0.06528513 -0.25399423 1.3532654 1.0083532 -0.02396515
0.17838322 -0.26831192 1.4455128 -0.96968937 -0.47586685 -0.11615356
0.9859033 -0.9685975 0.70221126 0.24631931 -0.93251354 -0.17863515
-1.0894827 0.4110845 0.20156665 0.4166229 0.54247135 0.7634757
-0.99126273 0.30718735 -1.2867881 -0.44997296 0.59429884 0.6477251
-0.46571913 -0.24104574 -0.75779945 1.4805592 0.20687056 0.2300041
-1.053962 -1.1639211 -0.7886388 -0.14561142 0.29870924 -0.82236415
1.3942072 -0.41946593 -1.1144183 0.97772914 0.26838058 -0.5831122
0.8656698 0.10227817 -0.13094936 0.46907443 0.02908946 0.30341086
0.6649198 -0.7723898 -0.44843388 -0.23564892 -0.55153817 0.10789503
0.20610134 -0.01983027 -0.14490186 -0.49928108 0.4591388 -0.9762112
-0.5692777 0.28172997 -0.01823565 0.05390253 0.72671694 -0.7046569
0.84752357 -1.7423807 -0.45575157 0.05780369 0.62163913 0.5839121
0.13947773 -0.10061649 -0.36082157 -0.11433894 0.01892297 0.01969905
-0.35287374 -0.13114695 0.56303066 1.2425963 -0.05651901 1.2027739
-1.0570735 -0.5241327 0.9397083 0.65580547 -0.32815596 0.41006058
0.39016733 0.8751568 -0.2394929 0.54355794 0.66445917 -0.67113966
0.32735363 -0.63724625 -0.22484173 0.17096655 -0.83828914 1.8267907
-0.6020222 0.7407662 0.03199085 -0.705066 0.4439752 0.7567757
1.1017563 -0.67666596 0.26034692 0.3848863 0.46146068 -0.82960266
-0.27150312 -0.34456995 0.24676476 0.59300435 -0.02327585 -0.28317448
-0.11011698 -0.14329366 1.2290522 -0.30898863 0.7720335 -0.5235162
0.5776447 0.20293973 0.25717643 0.897409 -0.68345547 0.8417998
-0.07610776 -1.0483295 -0.8593237 -0.8340801 -0.63126516 0.5628734
-0.12704326 -0.08443008 -0.41189614 -1.0439544 0.06196341 0.06628555
-0.46574232 0.07613317 -0.71273625 -1.2060713 0.5017489 0.505101
0.40382275 -0.7695753 -1.1832945 0.48944253 -0.25440153 -0.9681468
-0.40174276 0.17377356 -0.83369726 -1.6070988 -1.0381343 -0.9047575
0.7217563 0.31299832 1.175172 0.18005078 -0.9461479 0.43389776
-0.08297728 -0.22873525 -0.69232804 -0.14785622 -0.05494514 -0.56140065
0.3567465 0.01850396 -1.1198901 0.09483995 -0.47111592 0.10392964
1.1257049 1.03674 0.44086725 -0.22557007 0.31462422 -0.75280607
0.4810255 -0.19491537 -0.5625221 0.2971255 -0.8148376 -0.14202586
0.26489407 -0.20179662 -0.642608 0.31958792 -0.18531868 -0.31573004
-0.17163436 0.6741759 0.74622244 -0.90241283 0.8267306 0.27294183
0.22586349 -0.22506057 -0.054424 0.40212896 0.45416552 0.05662714
0.38465983 0.3440914 0.38085967 -0.57045454 -0.43668345 -0.79681796
-0.8520012 -0.38512564 0.6514688 -1.0312239 -0.4146355 0.44069862
-0.8472232 -0.13828133 -0.16065086 0.77274936 -0.46484008 0.55640274
-0.8046532 -0.29058018 -0.8443807 -1.7592205 0.94884604 -0.02319507
-0.20417777 -1.1596533 0.21621104 0.09983395 0.9660665 0.4317291
0.7538189 -0.76566505 -0.58004147 -0.50363445 -0.3478746 0.10305965
0.38022858 -0.40644 -0.6678268 -0.7006013 0.06467607 -0.10412737
0.6349779 0.8574664 0.7164635 0.0285299 -0.29057816 0.8262068
1.5161806 -0.02002196 0.05239956 0.5079802 0.41759485 -0.21772912
0.31594262 0.5447057 0.2681744 0.4109837 0.56559604 -0.88908803
-0.5606222 0.29438972 -0.23110609 0.6231364 0.15486684 0.29582644
-1.6468942 0.60897076 -1.4908315 -0.52727425 -0.28113744 1.9975055
0.5566317 -0.2164117 -0.14555833 -0.07630018 0.4276363 -0.08730745
-0.26698557 0.31878594 0.3126422 0.19462898 0.7485391 0.32423803
-1.453153 0.31027293 7.2322106 0.31820774 -1.6561127 0.46062598
0.3871464 0.05079919 0.4485097 -0.50038284 -0.21384872 0.12913398
0.44699976 -0.05843156 -0.03899398 0.7767312 0.4197836 -0.32228762
-1.308727 0.9514622 0.15063612 -1.6232822 -0.2482392 -0.130869
0.58128774 0.6148165 0.07900111 0.10529823 0.01414275 -1.030966
0.0832191 0.14536826 1.0876774 -0.33433828 1.1888291 0.3182137
-1.072005 0.11793263 -0.04154367 0.29667565 0.04531372 0.6425263
-1.8499197 0.54512566 0.3685023 0.54873794 -0.7318336 1.5938405
-0.02910287 0.44333124 -0.2778379 0.4263671 0.25729004 0.37813696
0.514298 1.7878835 0.35661942 0.16697375 0.59121597 0.3920968
0.3453643 0.08149169 -0.41725615 0.32269198 0.12061261 1.6793276
-0.90379363 -0.29288593 -0.5667247 0.47375998 -0.18761694 -0.09639539
-0.63576317 0.31096807 0.02824582 0.49355653 0.09344379 -0.14946915
-0.23995538 -1.0002712 -0.32481056 -0.9397021 0.69085085 -0.5457127
-1.2042377 0.8352021 -0.37045845 -1.5686109 -0.42866448 -0.61935395
-0.4978216 0.8563605 -1.8198948 -1.428526 -0.6504064 0.51213014
0.49284238 -0.0849393 1.0842208 0.31126457 -0.11016607 0.44311243
-0.08470513 0.43539277 0.8865422 -1.3457518 -0.12385087 1.0240861
-0.25641784 0.37680417 0.46977964 -0.44714162 -1.6265365 -1.0775754
0.54750615 -0.43189532 0.32823494 0.34925523 -0.5756389 0.72892886
0.17164405 0.73479694 0.9555724 -0.46293277 0.2524908 0.14234187
-1.5054071 0.04310352 0.26569885 -0.21386838 -0.711621 0.5172457
-0.0763443 -1.1074749 -1.1082355 0.91699696 0.58779114 -0.9648842
1.0596088 -0.1830007 0.05538971 -0.15952103 0.36203083 -1.42159
-0.76940936 -0.2729576 -0.06437776 0.10011216 0.09920339 -0.4716784
0.86305577 0.3838633 -0.3924692 -0.8150996 -1.0746987 -0.19457415
0.08775224 -0.18307588 0.1815525 0.98352826 -0.03161833 -0.24026509
0.13932241 0.61066157 0.74884385 0.10930342 0.35127026 -0.98888
-0.22185461 -0.5792565 -0.5781503 -0.7084198 -0.5641648 -1.1616014
0.08781313 -1.8542281 0.44941276 -0.6224433 -0.47599033 0.5377522
-0.17753462 0.5287206 0.23571311 0.35804632 -0.57210356 -0.05802308
1.5372486 -0.03405409 0.15525797 0.04770002 -0.23223928 0.8985925
0.3973849 -0.524605 0.07587115 -0.34205964 -0.330447 0.42499256
0.4447579 -1.3230555 0.58459425 0.42637107 0.95329624 -0.68914455
-0.18161702 -1.0756004 0.05579399 1.2247124 -0.0231551 0.0842522
0.36811635 0.06218064 -0.10531219 0.12236992 1.1429951 -0.51760674
-0.6537657 0.4720196 -0.69383144 -0.1060554 1.1277031 0.01884031
-0.05264411 -0.07163751 -0.8200282 0.18844065 0.04844043 0.02705966
0.7041902 -1.1156365 -1.0365603 0.72523457 -0.05095992 0.25288028
0.2982296 1.6977679 -1.0919718 0.80975854 -0.19577459 -1.2639134
-1.3035452 0.31809926 1.0451659 -0.7501954 -1.0475048 0.72038984
0.05536806 -0.3138408 0.14577873 -0.44357243 -0.02646353 -0.21281634
0.512947 0.08071689 0.76039207 -0.5009223 -0.6825202 0.48684332
-0.32817745 0.40859514 1.2597929 0.0468963 0.12656067 -0.08276907
0.8681545 -0.32304198 -1.1736851 -0.29739222 -0.16574347 -0.43429688
0.6902875 -1.2005533 -1.2611525 0.8144056 1.2040764 -0.11744274
1.1004983 -0.11665621 0.6201793 -0.06420635 0.44630167 -0.26041856
0.01842684 0.2558371 0.6279367 -1.6290098 0.43805674 -0.28670117
-0.46812692 1.4826843 0.30076143 -0.12624083 0.67784256 0.6453187
0.63018864 -0.3843721 -0.44263119 -0.01606571 0.44984928 0.5506046
0.83609957 0.24983983 -0.10287675 0.24102995 0.2842775 0.28622237
0.40444103 1.0623292 -0.46826863 -0.7274152 -0.42361805 0.49150726
-0.669613 -0.09584979 0.2869059 1.0469357 0.07041738 0.516468
-0.24259613 0.11957987 0.3495056 -0.04298664 0.7479425 -0.576609
-0.95708615 0.3670683 -0.28431222 -0.56659186 -0.30311516 -0.6190626
-1.0492992 -0.04111286 -0.32785988 -0.17659596 0.63570565 0.92922825
0.3142814 0.7621915 0.56821775 -1.0042654 -0.7974346 -1.0193317
-0.10669579 0.325574 0.72920424 -0.56063354 -0.14821665 -0.14763066] | [14.620542526245117, -2.568955659866333] |
cfa18a19-161c-49f8-b981-c54f61100274 | unsupervised-clustering-of-time-series | 2102.09200 | null | https://arxiv.org/abs/2102.09200v1 | https://arxiv.org/pdf/2102.09200v1.pdf | Unsupervised Clustering of Time Series Signals using Neuromorphic Energy-Efficient Temporal Neural Networks | Unsupervised time series clustering is a challenging problem with diverse industrial applications such as anomaly detection, bio-wearables, etc. These applications typically involve small, low-power devices on the edge that collect and process real-time sensory signals. State-of-the-art time-series clustering methods perform some form of loss minimization that is extremely computationally intensive from the perspective of edge devices. In this work, we propose a neuromorphic approach to unsupervised time series clustering based on Temporal Neural Networks that is capable of ultra low-power, continuous online learning. We demonstrate its clustering performance on a subset of UCR Time Series Archive datasets. Our results show that the proposed approach either outperforms or performs similarly to most of the existing algorithms while being far more amenable for efficient hardware implementation. Our hardware assessment analysis shows that in 7 nm CMOS the proposed architecture, on average, consumes only about 0.005 mm^2 die area and 22 uW power and can process each signal with about 5 ns latency. | ['John Paul Shen', 'José M. F. Moura', 'Harideep Nair', 'Shreyas Chaudhari'] | 2021-02-18 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 4.29262072e-01 -4.24693286e-01 8.67524594e-02 -2.94426292e-01
-3.04929286e-01 -4.51398492e-01 9.23337489e-02 5.76669395e-01
-7.69214988e-01 4.57993835e-01 -5.71190536e-01 -2.80366898e-01
-4.36812222e-01 -3.92870843e-01 -6.52009130e-01 -7.18593299e-01
-4.61116731e-01 3.04188430e-01 3.22332352e-01 1.79087698e-01
2.51790047e-01 3.94855112e-01 -1.78604960e+00 3.20863165e-02
5.14121175e-01 1.69594967e+00 -1.79128274e-01 4.01351541e-01
2.37708554e-01 4.04466003e-01 -4.74725455e-01 2.74391789e-02
3.14907193e-01 -1.27467841e-01 1.51071429e-01 -3.01341206e-01
5.30517772e-02 2.14928448e-01 -3.35785061e-01 1.12831974e+00
6.38243735e-01 1.35190174e-01 5.38913190e-01 -1.27427804e+00
7.88589939e-02 8.58372390e-01 -3.98577929e-01 3.74287099e-01
-9.48826969e-02 -1.83174431e-01 3.48373383e-01 -3.31690460e-01
3.72938573e-01 4.35728222e-01 1.02513492e+00 5.50943077e-01
-1.35585344e+00 -8.71265411e-01 -1.62957981e-01 3.64297628e-01
-1.37987280e+00 -5.05528390e-01 1.10402560e+00 -6.29192963e-02
1.29203165e+00 1.84520856e-01 6.36502564e-01 1.00833130e+00
6.59759521e-01 4.93222356e-01 1.13776982e+00 2.01000720e-02
1.07451785e+00 -4.13974047e-01 3.32847029e-01 1.91602036e-01
4.23214495e-01 -4.09537405e-01 -6.98740482e-01 -2.27766857e-01
4.16331321e-01 3.86785626e-01 1.86654121e-01 -5.87831289e-02
-8.81157398e-01 1.00676022e-01 2.22204234e-02 4.29364532e-01
-3.62558246e-01 7.56292105e-01 8.28380525e-01 3.89989913e-01
2.71988720e-01 2.55144358e-01 -5.44347525e-01 -5.79038501e-01
-1.20341086e+00 -2.70732790e-01 6.95501447e-01 8.97569895e-01
3.22364807e-01 4.22888547e-01 4.45750177e-01 3.84378552e-01
-6.04546517e-02 2.17523411e-01 8.97242069e-01 -7.28772521e-01
1.21446624e-02 7.25781560e-01 -3.64868015e-01 -7.45491505e-01
-7.44519949e-01 -2.68949091e-01 -1.19117773e+00 1.34258136e-01
2.53257930e-01 -1.97747052e-01 -9.46051657e-01 1.45961785e+00
9.70402807e-02 4.94653881e-01 -1.42921194e-01 5.98320782e-01
1.70160532e-01 5.50274134e-01 2.66094655e-02 -7.01642931e-01
9.64244604e-01 -2.23974466e-01 -8.82260978e-01 -1.19170407e-02
1.69770420e-01 -4.71638471e-01 6.89993918e-01 7.63472617e-01
-1.13928938e+00 -4.23232287e-01 -1.60014772e+00 2.20360979e-01
-4.62154478e-01 7.82006234e-02 8.32080603e-01 9.45719779e-01
-9.81118143e-01 1.15446401e+00 -1.56011963e+00 -4.33671623e-01
3.49218428e-01 1.04610991e+00 2.06849892e-02 5.55957794e-01
-4.83924955e-01 1.88106820e-01 2.30661452e-01 1.39103547e-01
-7.40427852e-01 -6.40450716e-01 -4.69439030e-01 1.68931577e-02
-4.35743295e-02 -1.20435044e-01 9.78476465e-01 -8.70181561e-01
-1.65797472e+00 4.56029326e-01 -1.24459729e-01 -1.08809853e+00
1.63363785e-01 -1.02495275e-01 -7.34486461e-01 2.14991748e-01
-4.62582171e-01 2.06514999e-01 9.90583003e-01 -4.59625244e-01
-2.89589137e-01 -8.17356110e-01 -9.44288254e-01 -5.40715396e-01
-8.35644484e-01 -7.41799846e-02 -9.23307836e-02 -5.30641258e-01
4.54328090e-01 -1.06396484e+00 -3.51847410e-01 -9.47625563e-02
-1.03915617e-01 -1.67709053e-01 1.30052936e+00 1.78345281e-03
1.26993525e+00 -2.19943309e+00 -2.47321889e-01 5.38117707e-01
3.74070145e-02 -1.62197828e-01 4.93658125e-01 2.86260158e-01
-9.82409790e-02 -1.88486472e-01 -4.13545758e-01 -1.75161511e-01
-6.08441159e-02 -9.16327834e-02 -1.86331525e-01 8.29189181e-01
-3.21021378e-01 5.68592072e-01 -5.69288313e-01 -4.42565344e-02
3.06609094e-01 4.32000309e-01 -1.86182469e-01 -2.91942358e-01
-9.27983259e-04 2.45260134e-01 -1.48020178e-01 8.26628864e-01
1.84722483e-01 -1.20317675e-01 4.09356266e-01 -4.07217264e-01
-3.27744752e-01 1.66650653e-01 -1.08780181e+00 2.03404474e+00
-9.68707651e-02 9.57404077e-01 2.33398210e-02 -1.11814559e+00
1.01555824e+00 3.46061885e-01 1.21174824e+00 -8.16559434e-01
7.22002864e-01 4.62700993e-01 2.28940040e-01 -1.36851996e-01
2.52260208e-01 1.73402399e-01 -3.03414553e-01 6.94709420e-01
9.77674648e-02 4.53429848e-01 2.84912903e-02 -2.02028334e-01
1.78066480e+00 -1.40737519e-01 8.76892209e-02 -5.79661131e-01
1.68890804e-01 -1.22622989e-01 6.98984683e-01 4.61374968e-01
-4.70915467e-01 1.06333196e-01 -3.45099568e-02 -4.02047843e-01
-1.04018676e+00 -1.27985275e+00 -7.21177682e-02 7.54708529e-01
1.02645665e-01 -4.64056581e-01 -9.20755744e-01 1.19364224e-01
-1.81360140e-01 4.04921830e-01 -2.78254509e-01 -2.74926811e-01
-5.18645644e-01 -8.63948226e-01 7.51104534e-01 6.54894769e-01
3.32552373e-01 -8.67476404e-01 -1.45322645e+00 5.07050931e-01
5.38555980e-01 -1.06518316e+00 -1.19374745e-01 8.25842857e-01
-1.54301238e+00 -4.26780373e-01 2.21536737e-02 -7.86074519e-01
7.32941866e-01 -2.68264592e-01 7.25567698e-01 -4.28413808e-01
-8.28806996e-01 4.95077342e-01 -6.73030168e-02 -6.54152036e-01
4.11032259e-01 -1.84312522e-01 7.34932184e-01 1.17201030e-01
8.44180286e-01 -1.56190801e+00 -8.42329502e-01 4.64002639e-02
-7.69449413e-01 -6.54382825e-01 4.00300592e-01 3.98039699e-01
1.01540649e+00 6.38214409e-01 7.91571081e-01 -7.11570621e-01
5.05942702e-01 -3.20961922e-01 -7.42397547e-01 -2.00256348e-01
-7.71120489e-01 -3.20664681e-02 1.10352933e+00 -8.89959097e-01
-5.23029923e-01 7.61424661e-01 2.69817114e-01 -6.89538777e-01
-5.66509292e-02 2.03772441e-01 -7.51061738e-02 -2.16698617e-01
6.89649999e-01 2.70865709e-01 -2.40201086e-01 -2.82366008e-01
-8.70672390e-02 6.93797231e-01 8.54774714e-01 -4.73570526e-01
4.66400981e-01 8.35393131e-01 2.08468005e-01 -1.02402139e+00
2.51408190e-01 -5.58627129e-01 -3.93194616e-01 -4.53507483e-01
6.52603507e-01 -7.76157975e-01 -1.26547647e+00 3.34395945e-01
-5.08509636e-01 -2.84860492e-01 -1.88919619e-01 4.82871056e-01
-7.08762109e-01 1.13570191e-01 -6.99098110e-01 -1.21429157e+00
-9.17429924e-01 -6.51448965e-01 7.56541371e-01 4.03237641e-01
-5.50756335e-01 -6.63092017e-01 -2.34588400e-01 -2.78903157e-01
4.33863640e-01 4.89218563e-01 7.01189160e-01 -7.10847676e-01
-4.15462315e-01 -3.83043200e-01 3.26862723e-01 -1.02288192e-02
9.02810618e-02 -2.14550778e-01 -1.17340064e+00 -4.34855610e-01
6.52397096e-01 -9.97691974e-02 6.24509692e-01 5.43929040e-01
1.64113784e+00 1.59352452e-01 -3.99643779e-01 6.86088920e-01
1.74025106e+00 8.26547205e-01 6.44727111e-01 -4.26580086e-02
5.77930808e-01 2.79604852e-01 3.22148174e-01 4.62865293e-01
-2.15156928e-01 7.86067247e-02 2.31109694e-01 3.72884572e-01
5.58534741e-01 1.81880683e-01 5.68124294e-01 1.43810642e+00
1.40948683e-01 -4.84948134e-04 -8.94231558e-01 7.63964295e-01
-2.06236315e+00 -9.31139886e-01 -1.07965991e-01 2.46499157e+00
7.20623016e-01 5.50832510e-01 1.02181546e-01 7.86020458e-01
7.02286661e-01 -4.21661943e-01 -1.17862678e+00 -4.71687585e-01
2.10601270e-01 6.77110612e-01 8.35948884e-01 -4.01502162e-01
-1.04672956e+00 3.63086224e-01 5.83415079e+00 6.79808497e-01
-1.31368208e+00 8.57931748e-02 5.09561241e-01 -7.00579762e-01
3.63925070e-01 -3.61577421e-01 -4.40581292e-01 7.91350543e-01
1.71585453e+00 -3.41252327e-01 4.11255807e-01 9.68070745e-01
1.33813158e-01 -6.58929050e-02 -1.44317257e+00 1.64581966e+00
-1.15642540e-01 -1.06621253e+00 -4.77926821e-01 -9.25954655e-02
6.27472162e-01 1.00656293e-01 1.04413822e-01 -1.47625670e-01
-2.77361095e-01 -8.60904813e-01 5.00962436e-01 2.44396970e-01
7.97962248e-01 -1.16424239e+00 4.05612051e-01 1.69699937e-01
-1.35346162e+00 -2.63112754e-01 -3.28854322e-01 -3.25591058e-01
3.49670835e-02 9.41922486e-01 -3.20607990e-01 3.04623842e-02
1.12537992e+00 5.99413335e-01 -2.53946841e-01 1.10383022e+00
4.95295197e-01 7.18847752e-01 -8.96132886e-01 -4.80739117e-01
-1.34200475e-03 -1.38888121e-01 4.13248241e-01 1.12269604e+00
5.72315037e-01 5.11757024e-02 -6.13266230e-02 5.06069899e-01
-1.73704997e-01 -7.63615221e-02 -6.55027926e-01 -3.12279969e-01
7.67523468e-01 1.38728797e+00 -1.54684222e+00 -1.88379943e-01
-1.43665254e-01 1.09109962e+00 -1.65907681e-01 -1.66291535e-01
-8.39449584e-01 -8.84647489e-01 6.06676936e-01 -4.74360660e-02
1.49438068e-01 -5.56486189e-01 -7.69981861e-01 -7.31578112e-01
3.02905887e-01 -4.13423270e-01 2.73837000e-01 -3.60385418e-01
-1.28464794e+00 4.68419135e-01 -4.91251200e-01 -1.54777169e+00
-2.63605744e-01 -8.16878140e-01 -6.54905558e-01 7.35914633e-02
-7.17711568e-01 -3.37593108e-01 -1.49840713e-01 7.55648613e-01
4.54406321e-01 -1.02657981e-01 6.55304253e-01 5.67636490e-01
-6.51222348e-01 6.58411801e-01 3.97515178e-01 -1.23876467e-01
5.45697331e-01 -1.07304537e+00 5.45541704e-01 8.45198750e-01
2.56866869e-02 8.45578253e-01 7.39339650e-01 -7.94309199e-01
-2.13697362e+00 -1.26614654e+00 2.79007077e-01 -6.71162531e-02
8.69433880e-01 -7.42818654e-01 -7.21642137e-01 2.14771971e-01
1.37964129e-01 -1.39955124e-02 1.05938375e+00 -1.31685629e-01
8.53980985e-03 -4.90570217e-01 -1.31631362e+00 7.49866545e-01
1.23416662e+00 -6.94021225e-01 -3.11917484e-01 1.50530681e-01
2.91851759e-01 -3.25271077e-02 -1.03405559e+00 4.22022730e-01
5.97050786e-01 -7.91612625e-01 6.64955020e-01 -6.08634911e-02
-9.84652862e-02 -5.61299920e-01 -1.63283929e-01 -6.03578329e-01
4.50150017e-03 -1.27994835e+00 -4.74612385e-01 1.07977760e+00
3.27318400e-01 -4.79609281e-01 1.19757450e+00 4.91548419e-01
-1.63584843e-01 -5.67491531e-01 -1.26305473e+00 -1.10486591e+00
-5.85940659e-01 -8.18887711e-01 4.34602685e-02 7.43978977e-01
5.04326940e-01 3.36040795e-01 -1.97083149e-02 -6.76036179e-02
1.08777833e+00 -1.70517638e-02 1.10775582e-01 -1.35888958e+00
-1.53746307e-01 -3.09826761e-01 -1.00151455e+00 -4.01269853e-01
-3.66644189e-02 -6.74837768e-01 2.08992541e-01 -9.66292500e-01
-2.92110384e-01 -2.99396981e-02 -6.99556589e-01 2.30752066e-01
5.21195292e-01 4.37872499e-01 -3.69577736e-01 -2.60937330e-03
-7.77983963e-01 2.55204320e-01 4.68126275e-02 -3.54382545e-02
-3.46938163e-01 -2.93802232e-01 -3.90113071e-02 8.79372239e-01
1.03948843e+00 -6.32336795e-01 -7.08184838e-01 -2.68109411e-01
1.41896412e-01 -2.39410520e-01 -1.48422057e-02 -1.88060701e+00
1.10841894e+00 4.63629991e-01 8.24627936e-01 -8.27648580e-01
4.77992564e-01 -1.23693442e+00 3.22246194e-01 6.72485352e-01
-6.46865647e-03 4.69635397e-01 5.02349854e-01 7.08853364e-01
8.91925022e-02 9.10567418e-02 6.64530277e-01 1.52602941e-02
-6.70080066e-01 1.87960073e-01 -7.54864573e-01 -2.75942445e-01
1.12833881e+00 -5.42195737e-01 -1.56111166e-01 -2.77036410e-02
-6.40686810e-01 -4.37307991e-02 4.70015973e-01 1.85259849e-01
6.11720502e-01 -1.23466539e+00 1.69962674e-01 1.65925920e-01
-2.33396053e-01 -3.80225390e-01 3.30062479e-01 7.45050967e-01
-3.22911948e-01 2.78331995e-01 -6.24755085e-01 -8.63781512e-01
-1.16657841e+00 5.79918861e-01 6.62657991e-03 3.52310330e-01
-7.32542574e-01 5.25075853e-01 -7.39278793e-01 3.12397033e-01
4.85165030e-01 -3.95746797e-01 2.18542144e-01 6.00426123e-02
5.71477711e-01 7.08873093e-01 3.36947590e-01 2.60465462e-02
-5.65368116e-01 6.53224349e-01 2.72098538e-02 -3.34357649e-01
1.46932089e+00 1.64109260e-01 -3.42635661e-01 1.07885146e+00
1.24664617e+00 -3.76890332e-01 -1.05430913e+00 1.58514291e-01
6.39807761e-01 3.43767852e-01 7.20824972e-02 -2.85737306e-01
-8.23103249e-01 6.60924792e-01 1.26806509e+00 2.47578010e-01
1.76356459e+00 -5.44146121e-01 1.10036314e+00 9.56714451e-01
7.85637259e-01 -1.80822921e+00 -1.20437056e-01 2.46239066e-01
1.20635005e-02 -6.68691635e-01 5.04204892e-02 -9.99674723e-02
6.92949966e-02 1.06542325e+00 5.91390371e-01 -5.65515757e-01
8.38553190e-01 9.44930375e-01 -1.85972899e-01 -1.44448102e-01
-9.42669034e-01 -7.97154324e-04 -3.19115035e-02 5.33354700e-01
2.53723890e-01 1.47202061e-02 -3.52204651e-01 6.98473334e-01
-1.71963662e-01 -4.82217595e-02 2.63071418e-01 1.29503250e+00
-4.25264657e-01 -7.55340636e-01 -6.52330443e-02 8.30385804e-01
-7.02363551e-01 1.45329416e-01 -4.49340254e-01 2.82424599e-01
-2.14738920e-02 8.96501243e-01 5.37447631e-01 -8.07238817e-01
8.58756602e-02 3.15977156e-01 4.78825897e-01 -1.94647193e-01
-8.97032499e-01 3.13114285e-01 -3.78064811e-01 -8.68848503e-01
-3.65116566e-01 -7.18047917e-01 -1.66723871e+00 -5.52959293e-02
1.63771026e-02 -5.47348894e-02 1.22419202e+00 6.61761522e-01
6.17966473e-01 7.69084573e-01 4.28364247e-01 -8.23552549e-01
-6.83270842e-02 -6.62423909e-01 -5.47166228e-01 6.81425706e-02
9.88311693e-03 -3.37388754e-01 -2.85362333e-01 1.41625568e-01] | [8.203685760498047, 2.4849863052368164] |
09f7a0f9-1a88-4a36-876b-0b746a430de8 | online-signature-verification-using-deep | 1806.09986 | null | http://arxiv.org/abs/1806.09986v1 | http://arxiv.org/pdf/1806.09986v1.pdf | Online Signature Verification using Deep Representation: A new Descriptor | This paper presents an accurate method for verifying online signatures. The
main difficulty of signature verification come from: (1) Lacking enough
training samples (2) The methods must be spatial change invariant. To deal with
these difficulties and modeling the signatures efficiently, we propose a method
that a one-class classifier per each user is built on discriminative features.
First, we pre-train a sparse auto-encoder using a large number of unlabeled
signatures, then we applied the discriminative features, which are learned by
auto-encoder to represent the training and testing signatures as a self-thought
learning method (i.e. we have introduced a signature descriptor). Finally,
user's signatures are modeled and classified using a one-class classifier. The
proposed method is independent on signature datasets thanks to self-taught
learning. The experimental results indicate significant error reduction and
accuracy enhancement in comparison with state-of-the-art methods on SVC2004 and
SUSIG datasets. | ['Mohammad Sabokrou', 'Mahmood Fathy', 'Mohsen Fayyaz', 'Mohammad Hajizadeh Saffar'] | 2018-06-24 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [ 4.12409365e-01 -5.62277138e-01 -2.85050780e-01 -6.95696712e-01
-7.59149790e-01 -5.59933782e-01 8.10165167e-01 -1.76297486e-01
-8.02528113e-02 6.02631509e-01 1.63474865e-02 -8.81496370e-02
-3.38764369e-01 -8.28652978e-01 -6.89787209e-01 -7.40282595e-01
-3.62128228e-01 4.49027032e-01 2.83736169e-01 -2.60605048e-02
4.76266265e-01 7.52243698e-01 -1.82807934e+00 4.54361349e-01
8.84522378e-01 1.22411609e+00 -9.73152965e-02 5.78646421e-01
2.19008829e-02 7.31089830e-01 -3.92784089e-01 -7.06338361e-02
3.19976717e-01 -5.07247627e-01 -7.18851268e-01 1.40043437e-01
4.63485211e-01 -6.41768038e-01 -7.80935884e-01 1.05432951e+00
2.73818970e-01 1.56517811e-02 7.20082521e-01 -1.33776164e+00
-8.42081070e-01 5.47657609e-01 -3.99685174e-01 -2.39890277e-01
2.86437422e-01 -1.99616402e-01 9.64558363e-01 -6.65360868e-01
7.03591526e-01 8.87701213e-01 8.06914628e-01 5.69928765e-01
-7.89065599e-01 -7.74282217e-01 -1.99197695e-01 8.41994703e-01
-1.56503975e+00 -3.94802839e-01 9.48271751e-01 -2.06509203e-01
2.32760757e-01 3.18351060e-01 2.89260268e-01 9.90398169e-01
-7.40769580e-02 9.89360273e-01 1.29480791e+00 -5.67985356e-01
1.95265859e-01 3.77033323e-01 5.36409080e-01 8.47323775e-01
1.62256323e-02 5.77822089e-01 -3.69508415e-01 -2.08570018e-01
2.59226590e-01 1.75873116e-01 -1.95699617e-01 -4.76636112e-01
-8.76541853e-01 6.19199753e-01 1.77768171e-01 6.22826397e-01
-1.15363561e-01 -1.02944560e-02 6.35291994e-01 3.62940699e-01
-2.08015129e-01 -1.92929238e-01 -3.88498127e-01 -1.23244703e-01
-1.20458996e+00 8.23998302e-02 6.88125312e-01 1.03983414e+00
9.12484288e-01 2.56879866e-01 -1.94015503e-02 5.34911335e-01
2.12839872e-01 7.74387956e-01 7.93492556e-01 -2.92947710e-01
6.87928498e-02 7.41145372e-01 1.96771696e-01 -9.98240709e-01
-6.69682696e-02 -3.36173743e-01 -8.38996351e-01 1.46535318e-02
1.65310830e-01 1.37241855e-01 -1.02128589e+00 1.34892821e+00
3.01045030e-02 5.82756102e-01 6.05760477e-02 7.28442252e-01
4.76754189e-01 6.40307426e-01 -3.18602175e-01 8.58057141e-02
1.04156387e+00 -7.01797962e-01 -7.86900818e-01 3.91304761e-01
5.63431799e-01 -5.56840122e-01 5.84888995e-01 4.09595162e-01
-3.91711026e-01 -9.10459459e-01 -1.30057406e+00 4.20450121e-01
-9.50220287e-01 6.63281858e-01 5.63760102e-01 1.09763551e+00
-7.72791743e-01 8.25346649e-01 -6.42292976e-01 -2.95731455e-01
3.79043788e-01 3.90063584e-01 -4.89470184e-01 -1.64310530e-01
-1.19113314e+00 8.63388777e-01 6.83926761e-01 3.43847871e-02
-1.10745645e+00 -3.99205387e-01 -8.86943579e-01 2.33624011e-01
-1.34702295e-01 5.87148964e-02 9.55165625e-01 -1.12216830e+00
-1.53398871e+00 6.06256843e-01 -1.32801309e-01 -4.95899022e-01
4.63129431e-01 -2.70119682e-02 -1.07962799e+00 9.45499763e-02
-6.97922781e-02 3.36506963e-02 1.06185126e+00 -1.63418090e+00
-7.79007256e-01 -3.89685452e-01 -3.80388498e-01 -4.64065075e-01
-5.61787486e-01 1.58181593e-01 -1.89643115e-01 -4.74368006e-01
1.87818557e-01 -9.40613329e-01 2.30487168e-01 -1.40086949e-01
-1.25443399e-01 3.55630405e-02 1.86671102e+00 -9.51858461e-01
1.10149014e+00 -2.12831306e+00 -3.91352475e-01 8.03109288e-01
-4.28909808e-01 6.47296369e-01 -3.16882700e-01 4.49731469e-01
-1.09088346e-01 -3.00005227e-01 -1.92747682e-01 -1.67072535e-01
2.34092593e-01 2.61472046e-01 -5.52123785e-01 8.24341297e-01
8.74133036e-03 7.86730111e-01 -9.79487121e-01 -4.82394636e-01
3.63972455e-01 5.28991938e-01 -1.53222516e-01 2.02780247e-01
1.24255836e-01 2.70493478e-01 -5.07149994e-01 9.23903883e-01
1.12710726e+00 -1.55168131e-01 3.20244730e-01 -3.26987892e-01
-1.10968109e-02 -1.26817390e-01 -1.55051470e+00 1.48114610e+00
-4.73021597e-01 5.34855843e-01 -1.10171907e-01 -1.37700963e+00
1.11891544e+00 1.69440359e-01 5.38718343e-01 -6.55351102e-01
2.60025948e-01 4.31329280e-01 -1.76352978e-01 -2.20312282e-01
4.26669985e-01 1.14623412e-01 -2.25857850e-02 7.61944532e-01
4.23607379e-01 3.09602141e-01 9.92666930e-02 1.23574764e-01
9.65861261e-01 1.89562231e-01 1.65941417e-02 -1.14893645e-01
1.00263953e+00 -2.82056093e-01 7.37085223e-01 7.10534453e-01
-8.89991596e-02 2.75434852e-01 -7.04483241e-02 -5.52022219e-01
-8.77704203e-01 -7.17010736e-01 -4.36630368e-01 7.10265398e-01
2.13420138e-01 -4.13021415e-01 -4.87993032e-01 -9.37296808e-01
4.40606654e-01 5.78407049e-01 -6.45156801e-01 -2.37232193e-01
-4.57656056e-01 -3.76829237e-01 9.08945501e-01 7.64178336e-01
7.64354646e-01 -7.37678170e-01 -3.35861683e-01 2.51054436e-01
3.33836913e-01 -8.86703968e-01 -2.78509200e-01 1.53024971e-01
-7.08384752e-01 -1.30072391e+00 -5.11500359e-01 -8.13295186e-01
9.00173068e-01 3.46082300e-01 3.53363723e-01 2.46837676e-01
-5.06806791e-01 4.47085440e-01 -6.90502882e-01 -3.47155005e-01
-5.12720108e-01 -2.31948733e-01 4.00835901e-01 7.58709311e-01
4.73453313e-01 -3.80786687e-01 -1.88914448e-01 4.37733948e-01
-8.30484033e-01 -5.59220731e-01 8.22312355e-01 1.37765944e+00
3.22125286e-01 3.58336955e-01 6.19084477e-01 -1.00768363e+00
2.11201027e-01 -4.85949181e-02 -1.00587130e+00 7.36103117e-01
-9.13702309e-01 6.10106066e-02 8.06365788e-01 -1.16923258e-01
-1.11860204e+00 4.74131137e-01 2.80887187e-01 -4.71862078e-01
-1.64272919e-01 2.98596859e-01 -2.01038599e-01 -7.44590342e-01
4.82180804e-01 1.03804457e+00 1.25482053e-01 -3.82533848e-01
3.44433159e-01 1.15970612e+00 5.66646278e-01 -7.02103376e-01
1.36154950e+00 5.63117802e-01 -5.54619730e-03 -7.00710475e-01
-4.07198846e-01 -7.41158068e-01 -1.06845844e+00 -1.91639990e-01
4.05296594e-01 -7.37994730e-01 -7.74564624e-01 9.13861215e-01
-9.43245530e-01 4.19801064e-02 -2.16692891e-02 3.66797954e-01
-4.52943712e-01 1.04297531e+00 -5.28949201e-01 -9.91548896e-01
-3.53484720e-01 -8.87199104e-01 1.10157824e+00 4.35200274e-01
2.55667180e-01 -8.79741907e-01 9.93679240e-02 1.27309546e-01
6.64810956e-01 -6.58381148e-04 6.42589331e-01 -9.22533333e-01
-7.24335134e-01 -9.63616908e-01 -4.68870640e-01 5.41121423e-01
3.97598505e-01 4.46702875e-02 -1.15700257e+00 -5.58796346e-01
-2.79559761e-01 -5.13442934e-01 7.59912848e-01 -1.94590449e-01
1.64377010e+00 -3.01700681e-01 -4.95586187e-01 5.25143802e-01
1.49481773e+00 2.79409021e-01 7.45670855e-01 1.50737271e-01
3.14868391e-01 1.40535459e-01 7.14880764e-01 5.35130322e-01
1.19384699e-01 6.37526512e-01 4.98842523e-02 1.27712116e-01
5.92627609e-03 -5.07659197e-01 3.11831862e-01 7.86650598e-01
-2.47928705e-02 1.04109861e-01 -7.13523269e-01 4.89372820e-01
-1.90508759e+00 -1.30804002e+00 7.07300529e-02 2.03671718e+00
5.65705478e-01 -2.06260577e-01 -1.37444675e-01 6.88649237e-01
6.51375294e-01 1.03912592e-01 -3.34858567e-01 -3.76387127e-02
-2.30692655e-01 3.23114097e-01 7.10461080e-01 3.22817892e-01
-1.41114008e+00 7.75360048e-01 6.26798296e+00 7.24539936e-01
-1.39147949e+00 -8.19329172e-04 9.94660705e-02 6.90159798e-01
-1.22097842e-01 1.22400917e-01 -9.08858061e-01 7.06974983e-01
9.29456234e-01 -2.46881917e-01 2.40700513e-01 1.09484601e+00
-2.00480804e-01 2.16202259e-01 -1.10669518e+00 9.97444272e-01
5.80839574e-01 -1.34023082e+00 7.11432770e-02 -1.60163671e-01
9.78231192e-01 -3.72300386e-01 -1.08986691e-01 4.53898579e-01
2.03469604e-01 -8.30071568e-01 4.84528154e-01 6.63636386e-01
1.01959336e+00 -5.50672174e-01 1.08262122e+00 2.40318701e-01
-1.28483021e+00 -4.53484952e-01 -5.27000964e-01 3.89130026e-01
-6.01570942e-02 4.28365827e-01 -8.13691139e-01 1.10048652e+00
3.26823652e-01 7.59057939e-01 -6.06524050e-01 1.23476160e+00
-1.83351487e-01 6.70082390e-01 -5.34411706e-02 -7.09486380e-02
5.60846142e-02 -1.87324241e-01 1.21303640e-01 1.41714716e+00
4.27777827e-01 9.76093635e-02 2.73728758e-01 6.26277685e-01
4.76115346e-02 3.30525748e-02 -5.65753818e-01 -1.85946882e-01
5.79629600e-01 1.05629539e+00 -2.44348258e-01 -6.10614002e-01
-3.72764021e-01 1.28444946e+00 -3.07444364e-01 1.14597797e-01
-8.95313561e-01 -8.04698110e-01 -2.79785618e-02 -8.78390372e-02
4.63731408e-01 -5.28107285e-02 -1.06680259e-01 -1.27583778e+00
1.31193355e-01 -8.68293822e-01 4.54282463e-01 -4.41865802e-01
-1.27235842e+00 2.82811850e-01 -4.04603869e-01 -1.75343215e+00
-3.02327543e-01 -8.55223894e-01 -6.73126400e-01 7.40659356e-01
-1.89908910e+00 -1.74660194e+00 -5.95968544e-01 7.79354036e-01
1.28688440e-01 -7.97082722e-01 1.29774356e+00 6.90445006e-01
-2.84076482e-01 1.15734923e+00 6.24540269e-01 6.34849429e-01
8.14727962e-01 -1.07473803e+00 -6.65184855e-02 9.05855775e-01
4.02869105e-01 6.75335050e-01 4.81196046e-02 -4.59015518e-01
-1.84254575e+00 -1.09725010e+00 1.04545772e+00 -9.65598524e-02
6.94539785e-01 -1.37734726e-01 -7.98447430e-01 6.71236873e-01
-1.97585821e-01 4.14309293e-01 8.93100500e-01 2.60437001e-02
-7.69262910e-01 -5.35928071e-01 -1.31681490e+00 -5.45247644e-02
8.15146565e-01 -1.05067003e+00 -6.63889647e-01 3.45943868e-01
-1.54073790e-01 -2.89739668e-01 -6.66579902e-01 2.88250595e-01
8.32696140e-01 -6.80736661e-01 5.81675947e-01 -9.12794054e-01
-9.25920084e-02 -7.31369078e-01 -3.42207432e-01 -1.02783322e+00
-4.60617751e-01 -2.95894295e-01 -3.43306303e-01 1.21673214e+00
6.63836300e-02 -3.69441807e-01 8.97030056e-01 2.77836502e-01
2.11748062e-03 -2.70816118e-01 -8.17861855e-01 -1.18586218e+00
-4.15193826e-01 -3.85140926e-02 7.60712802e-01 1.23253274e+00
1.03859425e-01 -4.06657636e-01 -6.34493947e-01 4.41419244e-01
8.42301667e-01 5.42217374e-01 8.10256064e-01 -1.29345262e+00
-2.93109715e-01 -4.02947478e-02 -1.03225911e+00 -9.53830123e-01
4.59893018e-01 -8.68378162e-01 1.34003945e-02 -9.65008736e-01
2.67333150e-01 -5.67089081e-01 -6.31450951e-01 6.93516970e-01
1.89679444e-01 -3.50947157e-02 2.16865838e-02 -4.62883078e-02
-5.55778146e-01 5.61595142e-01 6.34669781e-01 -5.67472517e-01
2.09992990e-01 2.31962636e-01 -2.83153087e-01 1.85245737e-01
4.99782532e-01 -1.05689034e-01 -2.19318733e-01 5.03802635e-02
-6.40549183e-01 -1.35805592e-01 4.50015217e-01 -1.13051796e+00
5.43714583e-01 -1.55594587e-01 7.16055691e-01 -8.47495139e-01
5.43840788e-02 -1.10173357e+00 -1.03019141e-01 7.33219683e-01
-1.48004159e-01 -3.55607986e-01 -1.47558721e-02 5.73372900e-01
-5.64262807e-01 -4.55446392e-01 9.56742585e-01 1.21166959e-01
-1.22995865e+00 2.66567171e-01 -9.94869396e-02 -7.08031058e-01
1.06690383e+00 -3.26875776e-01 -1.39405802e-01 -3.87881666e-01
-4.39347774e-01 1.19176142e-01 3.33198339e-01 4.19043660e-01
7.78051972e-01 -1.73359025e+00 -5.94630599e-01 5.81823111e-01
3.03901762e-01 -5.71988344e-01 4.42547888e-01 2.57991135e-01
-3.88408512e-01 5.85168302e-01 -5.82993209e-01 -5.52738667e-01
-1.26023018e+00 4.96157855e-01 2.63536245e-01 -1.90784231e-01
-4.63552564e-01 4.68461663e-01 -6.82161331e-01 -5.76470435e-01
5.02299309e-01 1.32477000e-01 -1.87852252e-02 -2.82144477e-03
8.40524495e-01 3.01949024e-01 2.57317424e-01 -7.58675694e-01
-5.33507407e-01 6.74121439e-01 -2.52753943e-01 1.15822233e-01
1.42884600e+00 4.46944803e-01 3.15825343e-02 -4.12310213e-02
1.47672760e+00 -4.47918102e-02 -7.96743691e-01 -3.25626373e-01
2.68749386e-01 -9.22444880e-01 -2.86271237e-02 -7.53769457e-01
-1.07179737e+00 7.71446884e-01 1.06032765e+00 -1.08839832e-01
1.06151330e+00 -6.37504876e-01 8.92385483e-01 6.62542105e-01
7.01305091e-01 -1.45776236e+00 -1.25132307e-01 6.22537673e-01
5.79925120e-01 -1.20249999e+00 5.11439033e-02 -1.03472598e-01
-3.86946827e-01 1.38523936e+00 3.29396516e-01 -2.14730263e-01
9.70146298e-01 1.71432495e-01 6.31023571e-02 1.08871482e-01
-2.53675640e-01 -1.02542795e-01 4.07510996e-01 7.94216037e-01
2.34767184e-01 8.54696333e-02 -4.05780852e-01 7.69328058e-01
1.12288212e-02 3.73499900e-01 3.11422408e-01 1.06859446e+00
-2.49996528e-01 -1.47440946e+00 -2.77346373e-01 4.11587507e-01
1.54635295e-01 4.57207084e-01 -4.47346419e-01 5.03080606e-01
2.24584907e-01 5.06897569e-01 -3.36586028e-01 -8.72827411e-01
2.57872760e-01 3.26140851e-01 5.44398725e-01 -2.37734944e-01
-2.91649759e-01 -5.45587122e-01 -1.53920174e-01 -4.63051915e-01
-2.69818395e-01 -7.28837073e-01 -1.16595519e+00 -7.25176930e-02
-5.43922007e-01 3.11714977e-01 7.77657449e-01 8.45455647e-01
3.64126772e-01 2.65760601e-01 1.32423973e+00 -5.92256308e-01
-1.36719227e+00 -7.96972632e-01 -9.14174557e-01 7.25239336e-01
1.12499952e-01 -6.70340121e-01 -3.28072727e-01 1.60256118e-01] | [12.62831974029541, 1.2272861003875732] |
daf8c406-33e1-4d96-b7c8-2f5c1fab6bc4 | swinrdm-integrate-swinrnn-with-diffusion | 2306.03110 | null | https://arxiv.org/abs/2306.03110v1 | https://arxiv.org/pdf/2306.03110v1.pdf | SwinRDM: Integrate SwinRNN with Diffusion Model towards High-Resolution and High-Quality Weather Forecasting | Data-driven medium-range weather forecasting has attracted much attention in recent years. However, the forecasting accuracy at high resolution is unsatisfactory currently. Pursuing high-resolution and high-quality weather forecasting, we develop a data-driven model SwinRDM which integrates an improved version of SwinRNN with a diffusion model. SwinRDM performs predictions at 0.25-degree resolution and achieves superior forecasting accuracy to IFS (Integrated Forecast System), the state-of-the-art operational NWP model, on representative atmospheric variables including 500 hPa geopotential (Z500), 850 hPa temperature (T850), 2-m temperature (T2M), and total precipitation (TP), at lead times of up to 5 days. We propose to leverage a two-step strategy to achieve high-resolution predictions at 0.25-degree considering the trade-off between computation memory and forecasting accuracy. Recurrent predictions for future atmospheric fields are firstly performed at 1.40625-degree resolution, and then a diffusion-based super-resolution model is leveraged to recover the high spatial resolution and finer-scale atmospheric details. SwinRDM pushes forward the performance and potential of data-driven models for a large margin towards operational applications. | ['Zhibin Wang', 'Fan Wang', 'Yuan Hu', 'Fei Du', 'Lei Chen'] | 2023-06-05 | null | null | null | null | ['super-resolution', 'weather-forecasting'] | ['computer-vision', 'miscellaneous'] | [-4.10097212e-01 -2.71158427e-01 9.60748196e-02 -2.94008285e-01
-6.05902255e-01 -5.56913018e-01 8.37007999e-01 -1.51929008e-02
5.05299866e-02 1.12172759e+00 4.62608129e-01 -7.57309318e-01
-1.90738633e-01 -1.37900269e+00 -7.87652622e-04 -9.37093735e-01
-4.98922706e-01 3.37035730e-02 2.55391985e-01 -9.04976070e-01
1.42772153e-01 7.64184892e-01 -1.59493506e+00 -8.17558244e-02
1.41430211e+00 1.04234862e+00 5.03633395e-02 6.74150407e-01
-4.62454930e-02 4.55809742e-01 -2.65485942e-01 3.36469859e-01
3.76799911e-01 -1.45718902e-01 -5.62878132e-01 -3.51279885e-01
1.10383809e-01 -4.76916879e-01 -9.28196404e-03 6.78954303e-01
5.09187758e-01 4.98014659e-01 3.50487977e-01 -7.23828852e-01
-2.76415199e-01 1.48941770e-01 -5.49451649e-01 5.08382618e-01
-2.19352335e-01 2.54217774e-01 7.57868648e-01 -1.14477646e+00
3.29901576e-01 9.26446021e-01 8.64952862e-01 1.86415184e-02
-1.33916104e+00 -5.97543359e-01 1.40183613e-01 -4.92336452e-01
-1.49281585e+00 -4.55526084e-01 2.13772327e-01 -8.43272686e-01
1.31400204e+00 3.75628978e-01 7.34284699e-01 3.88818115e-01
8.54391575e-01 -3.30067813e-01 1.24780667e+00 4.50349264e-02
4.35709774e-01 -3.23915511e-01 1.81518704e-01 -2.58285832e-02
2.11019337e-01 7.33851969e-01 -6.14275038e-01 -3.33504200e-01
7.80324578e-01 2.18195856e-01 -4.20107871e-01 6.77092433e-01
-7.65294850e-01 1.03915656e+00 5.90411305e-01 4.12355691e-01
-1.01746035e+00 -2.92260021e-01 -3.94159406e-02 3.54885668e-01
1.61350799e+00 5.86581290e-01 -8.76478732e-01 -1.41565785e-01
-1.68526244e+00 7.92973220e-01 5.00890851e-01 6.10063314e-01
6.93478644e-01 7.77564108e-01 -3.93632762e-02 6.41328216e-01
3.36063206e-01 1.22598934e+00 4.25522983e-01 -9.84055579e-01
3.42439741e-01 2.49971226e-01 6.05916977e-01 -1.06689906e+00
-7.35029399e-01 -7.72222281e-01 -1.50766242e+00 4.88645107e-01
-7.63638169e-02 -8.72442782e-01 -8.25569153e-01 1.26662302e+00
3.61186683e-01 3.30371559e-01 5.30921459e-01 1.08220840e+00
3.74952406e-01 1.66618156e+00 3.03782821e-02 -3.45505089e-01
1.06356466e+00 -9.60967422e-01 -8.42226028e-01 -3.88535231e-01
8.96261573e-01 -7.80717790e-01 3.77936274e-01 -1.93183377e-01
-8.67247164e-01 -6.97869003e-01 -8.19365501e-01 3.19221824e-01
-6.10944927e-01 -2.58239716e-01 3.81317079e-01 6.04867972e-02
-1.34185624e+00 1.13341224e+00 -8.82141173e-01 -4.78074215e-02
-3.59143287e-01 -2.19497353e-01 1.22754574e-02 1.89680561e-01
-1.91577768e+00 9.88527894e-01 -4.93709296e-02 6.99285328e-01
-5.18296897e-01 -1.26329052e+00 -7.64102280e-01 1.39882505e-01
-3.78786713e-01 -6.32095277e-01 9.71759319e-01 -3.28500688e-01
-1.52380538e+00 7.96387568e-02 -7.00720489e-01 -7.24276960e-01
1.64510950e-01 -3.59541655e-01 -1.17207098e+00 -2.00582877e-01
7.29363784e-02 1.76400572e-01 6.34586751e-01 -7.24677563e-01
-8.26059937e-01 -3.63826036e-01 -5.08156776e-01 3.37654501e-01
5.73164113e-02 -2.40871295e-01 2.70378500e-01 -9.35481787e-01
2.24219382e-01 -1.04026115e+00 -8.06788802e-01 -4.31846887e-01
1.05884820e-01 2.35489309e-01 6.42207026e-01 -1.03871596e+00
1.57247126e+00 -2.12082362e+00 -2.40927920e-01 3.46848875e-01
-8.05982389e-03 3.76283735e-01 -1.66701004e-01 8.37069154e-01
-1.21319462e-02 2.24393964e-01 -5.66906810e-01 -2.71279067e-01
-4.84180301e-01 4.82794404e-01 -1.19115043e+00 3.04434031e-01
4.05105919e-01 6.94864511e-01 -6.45895481e-01 2.86726713e-01
2.54138678e-01 6.91330016e-01 -3.14699858e-01 3.59519333e-01
-1.62872598e-01 8.58023345e-01 -4.72862333e-01 3.20484519e-01
1.30609608e+00 -3.13218981e-01 -6.08199239e-02 5.13775468e-01
-1.12292266e+00 5.19918084e-01 -1.16338706e+00 1.08161378e+00
-6.20715320e-01 5.13743877e-01 2.15189859e-01 -1.35684356e-01
1.39981592e+00 4.88704681e-01 1.23601049e-01 -8.26741099e-01
-7.38112092e-01 5.09626567e-01 -1.60467951e-03 -1.78051978e-01
9.93447542e-01 -3.46727699e-01 3.41465414e-01 2.46232122e-01
-7.69373536e-01 -4.45541412e-01 -8.58161673e-02 -1.52343558e-02
3.52685004e-01 1.78729519e-01 1.10426843e-01 -1.04152679e+00
3.07215214e-01 2.99462646e-01 6.03042066e-01 5.15642583e-01
2.20258832e-01 6.14238083e-01 1.83449998e-01 -7.08170891e-01
-9.94148612e-01 -4.73933101e-01 -5.83255708e-01 1.05548859e+00
-2.54253149e-01 -3.40642273e-01 -9.00281593e-03 2.80718356e-01
3.90130937e-01 9.25028205e-01 -6.87255025e-01 3.55020106e-01
-3.90155852e-01 -1.30595791e+00 3.00064743e-01 3.20852578e-01
5.76268971e-01 -6.53658509e-01 -4.27417815e-01 6.44764960e-01
1.68540329e-01 -7.57124126e-01 1.52785391e-01 9.96471867e-02
-1.45706642e+00 -8.41688737e-02 -9.69635725e-01 -1.90491360e-02
2.12624267e-01 2.54435837e-01 1.28900361e+00 -1.83385879e-01
5.70575118e-01 -4.90636677e-01 -2.67999589e-01 -8.90341699e-02
-5.93927316e-02 2.94669002e-01 2.67875940e-01 -2.59938296e-02
-1.43847063e-01 -9.27987933e-01 -7.40507662e-01 2.49818549e-01
-6.44085884e-01 2.72435009e-01 1.61920786e-01 5.38647652e-01
3.77588779e-01 1.91800017e-02 5.75065374e-01 -6.95300102e-01
5.09054363e-01 -8.23875010e-01 -1.25482261e+00 -1.84664339e-01
-1.18386948e+00 -3.73527966e-02 6.12425208e-01 1.33966222e-01
-1.36004126e+00 -3.93262655e-01 -3.76535833e-01 -1.27258092e-01
-7.56074190e-02 1.17410934e+00 6.35007560e-01 1.09049678e-01
5.38930774e-01 5.06803274e-01 -2.26245537e-01 -8.42684567e-01
3.25910658e-01 4.68310922e-01 3.72367114e-01 -3.45821649e-01
8.27106476e-01 2.91692883e-01 6.11257963e-02 -1.05164587e+00
-8.26495647e-01 -2.95072198e-01 -5.41068554e-01 7.81677291e-03
4.96908426e-01 -1.37748647e+00 -1.29499316e-01 8.29861939e-01
-1.04349947e+00 -5.56416988e-01 -1.21675283e-01 4.64982271e-01
3.01531941e-01 -1.62788615e-01 -6.90591037e-01 -1.11798739e+00
-7.37704575e-01 -6.03539050e-01 8.82613957e-01 1.89781010e-01
-2.20338091e-01 -1.11555171e+00 5.82312405e-01 -4.60930824e-01
1.32200205e+00 5.53622186e-01 6.74990714e-01 -1.86022624e-01
-3.43973041e-01 -1.29319549e-01 -3.82031351e-01 -6.07237965e-02
9.81387198e-02 3.18805277e-01 -9.39287245e-01 -2.51800060e-01
9.79920775e-02 1.65383860e-01 8.20822656e-01 7.28072047e-01
4.93744254e-01 -5.05713642e-01 -2.17903167e-01 9.90509987e-01
1.42781830e+00 1.15640312e-01 3.63919467e-01 2.19576448e-01
2.84074366e-01 5.19004583e-01 6.43737435e-01 8.52052569e-01
3.19775194e-01 3.01187664e-01 2.42848277e-01 -3.52759451e-01
3.89019549e-01 7.42565170e-02 -1.31851779e-02 1.03233671e+00
-6.07428730e-01 -9.69929434e-03 -1.52809083e+00 5.39184928e-01
-1.72438979e+00 -9.37203169e-01 -5.49832940e-01 2.02051187e+00
7.15783775e-01 1.17895044e-01 -5.72665215e-01 -1.66332468e-01
2.97548562e-01 1.05607986e+00 -5.51390827e-01 -5.03245294e-01
-2.65246302e-01 2.53386974e-01 8.44802916e-01 1.00975788e+00
-8.45473588e-01 8.94067824e-01 6.31011772e+00 2.67658472e-01
-1.60630107e+00 6.40921146e-02 7.06266582e-01 1.04363889e-01
-3.06407064e-01 6.79616556e-02 -1.42337430e+00 4.90037322e-01
1.85356367e+00 -2.86594748e-01 2.53687620e-01 8.06694508e-01
8.39736164e-01 -1.63962170e-01 -2.66887426e-01 2.90237367e-01
-8.20258915e-01 -1.94132829e+00 -7.23073259e-02 2.55460978e-01
1.23336375e+00 5.24917245e-01 7.06630424e-02 4.35251832e-01
4.59627241e-01 -1.00130403e+00 1.23374052e-01 9.50726211e-01
8.83586526e-01 -7.93140948e-01 8.91529799e-01 5.62545598e-01
-1.59674442e+00 2.52991199e-01 -5.01866043e-01 -8.61781776e-01
3.98100257e-01 1.32500374e+00 -2.05515087e-01 8.65682364e-01
8.75532627e-01 1.12831903e+00 -3.10424678e-02 5.06325841e-01
6.69841692e-02 9.56352115e-01 -5.31087995e-01 4.98879582e-01
6.57134175e-01 -5.42977095e-01 3.22769016e-01 9.67468381e-01
9.27980721e-01 8.73566747e-01 6.19030967e-02 4.02492642e-01
3.92303467e-01 -4.40167710e-02 -6.69316351e-01 4.93914753e-01
6.56118691e-01 1.06406474e+00 1.85688034e-01 -6.13122642e-01
-3.03763598e-01 5.48062325e-01 -5.09708598e-02 7.97982216e-01
-3.95639896e-01 -1.01446100e-01 1.43922198e+00 3.28901917e-01
3.44802082e-01 -6.27045751e-01 -6.07360601e-01 -1.09779823e+00
-4.11884606e-01 -6.03034914e-01 9.83681381e-02 -8.74270856e-01
-1.09664094e+00 1.08347988e+00 -1.99491084e-01 -1.53966701e+00
-6.35638595e-01 -1.88812166e-01 -7.48602927e-01 2.08297825e+00
-2.19267011e+00 -6.12506270e-01 -1.52046800e-01 2.57565916e-01
3.20348024e-01 8.02318305e-02 1.38594460e+00 -7.08390176e-02
-6.20119452e-01 -4.20514554e-01 1.06490600e+00 -1.20833300e-01
4.55650121e-01 -9.43454981e-01 1.25066173e+00 9.21883166e-01
-8.08870733e-01 6.43445611e-01 7.80956626e-01 -8.50153267e-01
-1.10124230e+00 -1.62139940e+00 1.57565796e+00 1.05134130e-01
8.65348458e-01 9.96699482e-02 -1.36141145e+00 4.53543812e-01
3.37256819e-01 5.87162197e-01 3.31656843e-01 2.37980351e-01
-3.05513069e-02 -4.48785961e-01 -9.12527084e-01 4.34741110e-01
2.43206874e-01 -4.07297343e-01 -5.05910277e-01 3.05517673e-01
7.81283200e-01 -6.31658018e-01 -1.51144612e+00 8.25928390e-01
3.27805281e-01 -9.41821933e-01 4.41710591e-01 -3.98010433e-01
5.72875559e-01 -2.64184773e-01 -2.24512294e-01 -1.85812926e+00
-7.06726074e-01 -7.11850464e-01 -1.21921271e-01 9.77734745e-01
6.25812888e-01 -1.30253124e+00 1.50752649e-01 9.19074833e-01
7.56284371e-02 -6.79543972e-01 -1.09290981e+00 -6.35712385e-01
5.45407951e-01 -3.15408826e-01 9.38341200e-01 1.20091510e+00
-2.31923446e-01 6.32901266e-02 -6.33782864e-01 1.02029347e+00
4.56222832e-01 8.27090085e-01 3.09217066e-01 -1.49993682e+00
3.00314754e-01 -3.58185887e-01 3.69847029e-01 -1.21878815e+00
-3.11458617e-01 -1.30782440e-01 -1.78181872e-01 -1.25116503e+00
-7.18899786e-01 -5.26580215e-01 -3.13003302e-01 3.59160841e-01
-2.03437522e-01 6.48201779e-02 -7.47379586e-02 5.82896352e-01
5.32730937e-01 8.59592617e-01 1.07106769e+00 4.10301179e-01
-3.59419376e-01 4.20713127e-02 1.28747404e-01 4.26242411e-01
8.45046520e-01 -2.58262962e-01 -5.04477136e-03 -6.85269713e-01
1.25467896e-01 7.29944825e-01 -2.68260166e-02 -1.15629709e+00
1.40375853e-01 -5.85874438e-01 4.77324188e-01 -9.98368442e-01
3.39339972e-01 -4.60408270e-01 3.70829910e-01 4.34668392e-01
-1.12090275e-01 4.02762800e-01 7.07745433e-01 4.50710580e-02
-6.42253280e-01 4.16563600e-01 7.79207945e-01 7.78384507e-02
-8.68110299e-01 6.05567694e-01 -7.29762495e-01 -1.35463983e-01
5.79306126e-01 4.10523504e-01 -3.83570969e-01 -3.54509085e-01
-6.75998986e-01 6.12600327e-01 3.05023223e-01 2.84402847e-01
3.28625590e-01 -1.01307595e+00 -1.15532100e+00 6.56197190e-01
-1.99818656e-01 9.82563794e-02 5.61188519e-01 4.79602665e-01
-5.45805097e-01 6.64104044e-01 7.54844993e-02 -4.21908110e-01
-6.47890031e-01 1.58485338e-01 5.85060060e-01 -4.81437594e-01
-7.97053456e-01 6.12747252e-01 -1.41293779e-01 -5.33273578e-01
-6.58442199e-01 -5.33486724e-01 -1.81902379e-01 3.74894202e-01
8.76729608e-01 3.17231774e-01 1.62484661e-01 -6.25717700e-01
-2.22742826e-01 7.81291604e-01 4.11823392e-01 -3.18061173e-01
1.57688344e+00 -4.39254582e-01 -1.15417391e-01 7.08639979e-01
8.25436652e-01 -3.81366946e-02 -1.62189960e+00 -4.61193949e-01
-1.29993975e-01 -4.13737148e-01 6.95765734e-01 -7.06616521e-01
-1.15552354e+00 1.04131675e+00 3.92348588e-01 5.06523013e-01
9.75114167e-01 -8.52565348e-01 1.05355489e+00 1.13954380e-01
4.30002809e-02 -6.75649762e-01 -1.05145872e+00 1.21364617e+00
1.01881862e+00 -1.18078673e+00 1.21726781e-01 3.14832628e-02
-5.49753428e-01 1.10013556e+00 1.67046323e-01 2.00633686e-02
1.16149080e+00 5.94259381e-01 4.50986564e-01 6.90742061e-02
-1.20028567e+00 2.08547916e-02 3.23759347e-01 -1.70453772e-01
6.49418652e-01 3.23822111e-01 -1.75991561e-02 3.47363085e-01
-1.79038301e-01 1.51427731e-01 1.40370563e-01 7.85169125e-01
-6.55030012e-01 -5.16550303e-01 -6.82982206e-01 2.71007270e-01
-8.26104283e-02 -8.32228124e-01 6.94521606e-01 3.41030210e-01
-3.44543248e-01 8.87667894e-01 4.00570333e-01 -8.42558395e-04
1.47409379e-01 2.16120273e-01 -8.40126157e-01 -2.91690439e-01
-6.09249175e-01 1.29669145e-01 3.77973393e-02 -5.71989655e-01
-2.39871696e-01 -4.81403261e-01 -9.96340454e-01 -1.02332485e+00
1.18266828e-01 7.00469494e-01 4.69174623e-01 9.47417736e-01
8.48565996e-01 3.53317559e-01 9.11431253e-01 -1.44213343e+00
-4.00656581e-01 -1.24479890e+00 -1.06798398e+00 -2.33239517e-01
9.16791439e-01 -2.49247491e-01 -4.87559080e-01 -3.70734781e-01] | [6.554004669189453, 2.9579529762268066] |
8357ef85-efc6-4c4e-beec-cd0e0c6ec74e | unsupervised-attention-based-instance | 2011.01888 | null | https://arxiv.org/abs/2011.01888v1 | https://arxiv.org/pdf/2011.01888v1.pdf | Unsupervised Attention Based Instance Discriminative Learning for Person Re-Identification | Recent advances in person re-identification have demonstrated enhanced discriminability, especially with supervised learning or transfer learning. However, since the data requirements---including the degree of data curations---are becoming increasingly complex and laborious, there is a critical need for unsupervised methods that are robust to large intra-class variations, such as changes in perspective, illumination, articulated motion, resolution, etc. Therefore, we propose an unsupervised framework for person re-identification which is trained in an end-to-end manner without any pre-training. Our proposed framework leverages a new attention mechanism that combines group convolutions to (1) enhance spatial attention at multiple scales and (2) reduce the number of trainable parameters by 59.6%. Additionally, our framework jointly optimizes the network with agglomerative clustering and instance learning to tackle hard samples. We perform extensive analysis using the Market1501 and DukeMTMC-reID datasets to demonstrate that our method consistently outperforms the state-of-the-art methods (with and without pre-trained weights). | ['Benjamin S. Riggan', 'Kshitij Nikhal'] | 2020-11-03 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 3.56969424e-02 -3.79918844e-01 1.79253325e-01 -7.67693281e-01
-6.71609581e-01 -4.99178529e-01 5.38437426e-01 1.34326756e-01
-1.09267354e+00 6.20916069e-01 2.80638784e-01 8.64427984e-02
-1.45942166e-01 -3.91581446e-01 -6.35886788e-01 -4.73267436e-01
9.10916105e-02 5.98446548e-01 -2.10261047e-01 6.18745685e-02
1.28199682e-01 3.47482562e-01 -1.48603404e+00 -1.26251593e-01
1.23912215e+00 6.36390805e-01 -2.40608435e-02 6.12615287e-01
4.49919909e-01 2.76434600e-01 -4.42973673e-01 -7.38852501e-01
3.85892957e-01 -6.09380519e-03 -6.55004144e-01 2.51017153e-01
8.31260324e-01 -4.18425947e-01 -3.64248455e-01 1.09052742e+00
9.21605408e-01 4.45355266e-01 6.11886442e-01 -9.66398954e-01
-7.90729880e-01 2.51601994e-01 -7.82331169e-01 2.26037949e-01
1.49795413e-01 2.97740489e-01 5.70830762e-01 -8.14065158e-01
2.76694030e-01 1.32840168e+00 9.14310932e-01 6.65143192e-01
-1.35274220e+00 -8.33024383e-01 3.57458651e-01 3.10792148e-01
-1.47923338e+00 -6.28439903e-01 5.52929640e-01 -4.53419000e-01
8.94292355e-01 -2.13508531e-02 2.06068397e-01 1.24688900e+00
-4.04676795e-01 6.14296794e-01 9.19780552e-01 -3.25434119e-01
-2.12504133e-03 2.19675332e-01 2.87497699e-01 3.91041547e-01
3.93039256e-01 1.46498531e-01 -3.07812870e-01 1.47551343e-01
5.67044735e-01 1.73646033e-01 -1.36584928e-02 -1.85294151e-01
-1.06517756e+00 5.08995533e-01 5.29911160e-01 1.07603557e-01
2.84479558e-02 -8.51135235e-03 4.44822609e-01 -4.59377281e-02
4.13234144e-01 4.03046787e-01 -4.49636221e-01 -1.73067912e-01
-8.86371791e-01 2.49719292e-01 4.28575546e-01 8.65494847e-01
5.94867885e-01 -2.74530798e-01 -1.13843173e-01 1.08469486e+00
2.02002347e-01 4.14812863e-01 5.74580371e-01 -6.08499527e-01
7.54482985e-01 7.12056100e-01 3.23824048e-01 -9.58635092e-01
-7.08773255e-01 -7.78459132e-01 -1.05659389e+00 -4.77167293e-02
6.06970310e-01 -2.87689358e-01 -9.11178350e-01 1.88857043e+00
2.98561752e-01 3.21253210e-01 -1.87171191e-01 9.85054672e-01
6.17048144e-01 9.47878808e-02 3.42104644e-01 2.34216511e-01
1.24634099e+00 -1.06689787e+00 -5.33508837e-01 -3.42501253e-01
5.19645751e-01 -4.11671996e-01 9.81971502e-01 5.23289815e-02
-1.00811279e+00 -1.01840103e+00 -1.00108480e+00 -2.67648965e-01
-5.53036273e-01 4.45236415e-01 3.78764212e-01 8.93537521e-01
-1.05791008e+00 4.31105375e-01 -7.61817753e-01 -7.12267220e-01
4.78047937e-01 8.23229611e-01 -5.48382163e-01 1.11827708e-03
-8.55241060e-01 6.21552944e-01 3.53983939e-01 3.84224683e-01
-4.55708802e-01 -7.13606715e-01 -8.15748632e-01 1.62753165e-01
1.01889193e-01 -7.87861645e-01 8.38409007e-01 -9.12201047e-01
-1.46337545e+00 8.89141798e-01 -2.34273672e-01 -3.92023593e-01
7.37509608e-01 -5.36732733e-01 -3.59419733e-01 -1.30734652e-01
2.24850416e-01 8.57434273e-01 5.69278002e-01 -1.15323925e+00
-5.71817398e-01 -6.37332857e-01 -8.15660506e-02 4.18261051e-01
-6.66338384e-01 2.03104138e-01 -6.60125017e-01 -6.70760334e-01
-2.12510258e-01 -1.03823674e+00 -1.75466895e-01 -1.96773127e-01
-5.29246569e-01 -1.64282903e-01 5.56178272e-01 -8.61649334e-01
8.28317761e-01 -2.20853853e+00 1.15455061e-01 2.62173355e-01
1.24738745e-01 4.08489883e-01 -5.03471717e-02 2.09641103e-02
-1.83765352e-01 8.99198502e-02 -2.62668103e-01 -1.17150414e+00
1.39498591e-01 -1.96794689e-01 2.07559377e-01 5.86847246e-01
4.47648913e-02 9.95744050e-01 -6.81424499e-01 -4.19953406e-01
4.02506173e-01 5.64722598e-01 -6.30110383e-01 2.73551166e-01
3.59975636e-01 7.90795267e-01 -9.05063301e-02 5.49657822e-01
7.90317655e-01 -3.63080144e-01 3.89159955e-02 -1.55283809e-01
-7.75985196e-02 5.02908491e-02 -1.30910993e+00 1.73838580e+00
-1.33722961e-01 4.95719761e-01 -5.92450872e-02 -1.03918147e+00
5.86925566e-01 2.93267723e-02 3.60635638e-01 -6.54130697e-01
1.64350346e-01 -1.22143224e-01 -7.10585415e-02 -3.91143799e-01
5.76773942e-01 4.33911115e-01 -9.25093889e-02 3.76030833e-01
-3.37824188e-02 8.03076565e-01 1.32605597e-01 1.09435707e-01
7.40498364e-01 2.69264951e-02 -1.27975494e-01 -3.66549939e-01
5.78066468e-01 -4.31577802e-01 6.72559619e-01 7.98515081e-01
-3.71135652e-01 6.55413926e-01 -4.58815955e-02 -3.80777061e-01
-1.07664597e+00 -8.97814751e-01 -2.15537734e-02 1.24540687e+00
3.42304975e-01 -1.14006326e-01 -1.01623595e+00 -5.50307572e-01
1.56795070e-01 2.69077539e-01 -7.17192709e-01 -2.37292387e-02
-5.03906131e-01 -1.18178201e+00 7.09367037e-01 8.34203124e-01
8.86345923e-01 -6.79072559e-01 -1.81237608e-01 1.21449612e-01
-1.90186337e-01 -1.47517133e+00 -8.90504479e-01 -2.53665954e-01
-5.87081790e-01 -9.14359152e-01 -7.82806277e-01 -8.92013967e-01
1.03246832e+00 3.68321538e-01 8.52392375e-01 1.19587690e-01
-3.47679794e-01 4.83703822e-01 -2.38055244e-01 -2.02094421e-01
2.69791484e-01 4.55675781e-01 4.06947821e-01 3.66042972e-01
5.07861078e-01 -5.17439663e-01 -8.81794930e-01 4.99401748e-01
-4.87771183e-01 3.22116092e-02 4.15623039e-01 8.08664799e-01
3.66554826e-01 -8.56702309e-03 6.34602189e-01 -7.07331657e-01
4.97529328e-01 -1.66493431e-01 -5.22054672e-01 2.45060250e-01
-6.10245049e-01 -1.45261124e-01 3.78772378e-01 -5.72445750e-01
-1.20907724e+00 6.54025078e-02 2.90255621e-02 -5.47063611e-02
-4.66121644e-01 1.78876817e-01 -4.37265843e-01 -1.55361101e-01
5.55029213e-01 1.16891533e-01 -2.76024669e-01 -5.69340587e-01
3.47216696e-01 7.63869703e-01 9.74862933e-01 -5.01079559e-01
9.89675343e-01 5.69257617e-01 -3.42612267e-01 -6.39670491e-01
-6.16556168e-01 -6.62510335e-01 -1.07081199e+00 3.32730226e-02
1.12639868e+00 -1.25958312e+00 -9.68082368e-01 8.33305538e-01
-9.34292972e-01 -4.57260489e-01 2.40757465e-01 5.48065245e-01
-1.10858038e-01 6.02867484e-01 -6.73317254e-01 -7.13492274e-01
-4.91452605e-01 -1.07573926e+00 9.08336222e-01 6.73542798e-01
-1.85021117e-01 -8.10996056e-01 -2.35255018e-01 6.50982440e-01
4.09720361e-01 1.48614973e-01 4.00204450e-01 -6.71637058e-01
-4.38760281e-01 -2.85780400e-01 -6.07324123e-01 8.07569474e-02
1.53375164e-01 -3.37430060e-01 -1.01956499e+00 -7.04901636e-01
-5.55443823e-01 -2.21123308e-01 8.07780325e-01 3.23439419e-01
1.33751643e+00 -1.49113446e-01 -3.99662286e-01 1.04909849e+00
1.00878322e+00 -1.97543591e-01 5.07631719e-01 5.60650706e-01
1.03275204e+00 5.97405732e-01 1.48449242e-01 3.94390434e-01
9.16904569e-01 8.29536796e-01 1.63734317e-01 -2.90969104e-01
3.68209071e-02 -1.47513196e-01 2.15493571e-02 4.30864900e-01
-3.62009257e-01 -5.89089803e-02 -6.91303432e-01 5.77962220e-01
-2.13242340e+00 -1.01837754e+00 1.55432254e-01 2.34682536e+00
5.89299262e-01 -6.01051971e-02 5.18478930e-01 -2.84101199e-02
9.20167267e-01 -2.59322941e-01 -7.65038192e-01 1.68257296e-01
-9.93208289e-02 -1.46734104e-01 6.86218381e-01 4.76449937e-01
-1.49476695e+00 9.24631298e-01 5.82043076e+00 2.93242067e-01
-8.66425455e-01 2.31548082e-02 7.63401568e-01 -1.76170468e-01
2.78889626e-01 -5.06597161e-01 -9.88424838e-01 7.61824369e-01
7.47865438e-01 2.63943672e-01 4.83564496e-01 6.52642369e-01
1.82808176e-01 5.52215539e-02 -1.26126647e+00 1.40609527e+00
1.94184303e-01 -7.88902164e-01 -3.57407719e-01 1.72649011e-01
9.30575490e-01 -1.04358546e-01 2.75004923e-01 3.04783821e-01
3.16094398e-01 -1.08628881e+00 5.07932365e-01 4.49220270e-01
7.25584090e-01 -8.34164381e-01 9.39502656e-01 1.53753147e-01
-1.18812120e+00 -2.15280548e-01 -3.07931304e-01 4.69096527e-02
7.54913613e-02 3.09920847e-01 -4.78541881e-01 4.76713687e-01
1.03520596e+00 6.95224881e-01 -8.96899939e-01 1.19258380e+00
-1.07633621e-01 3.46753955e-01 -3.60626578e-01 2.41563484e-01
-1.35551542e-01 -9.06452462e-02 2.19945952e-01 1.39913130e+00
1.07516244e-01 1.41072348e-01 2.68673092e-01 6.23816013e-01
-2.50940293e-01 -1.50293425e-01 8.36515799e-02 2.38398820e-01
5.36060035e-01 1.16688001e+00 -5.23436844e-01 -2.68829793e-01
-4.81452346e-01 1.36221850e+00 6.26759052e-01 6.53746128e-01
-8.52416277e-01 -2.62654901e-01 7.47131288e-01 8.92735645e-02
2.36237749e-01 -4.66509908e-01 -3.86149943e-01 -1.27549577e+00
2.40181163e-02 -7.62861848e-01 5.32408416e-01 -4.11800504e-01
-1.64987814e+00 4.25758600e-01 -1.12820990e-01 -7.67716646e-01
-1.86559230e-01 -5.40049493e-01 -4.31809515e-01 9.68025804e-01
-1.50788510e+00 -1.40053916e+00 -7.10534453e-01 7.66489208e-01
3.02621931e-01 -2.61992127e-01 6.36277378e-01 7.52281845e-01
-9.84680295e-01 1.36241722e+00 1.67660192e-01 6.40736043e-01
1.07194161e+00 -1.31136143e+00 5.44214606e-01 9.60281491e-01
-2.73076773e-01 8.16713929e-01 4.38261122e-01 -4.77450788e-01
-1.10089970e+00 -1.18056953e+00 8.63292992e-01 -5.27122378e-01
2.61113316e-01 -7.03905523e-01 -7.74732590e-01 7.39009619e-01
-8.00247714e-02 9.93549675e-02 9.04958904e-01 5.10938168e-01
-4.43961710e-01 -3.41429710e-01 -1.18084919e+00 5.72286308e-01
1.29181671e+00 -6.20063186e-01 -3.36684048e-01 2.44078323e-01
3.27809423e-01 -3.90405446e-01 -7.01194465e-01 3.89477193e-01
7.10728526e-01 -8.10873866e-01 1.21693206e+00 -5.62678933e-01
-1.92697346e-01 -3.25590611e-01 6.65472671e-02 -1.20909548e+00
-6.32518351e-01 -6.08201265e-01 7.64698312e-02 1.40758467e+00
2.80087084e-01 -7.82264829e-01 8.62121046e-01 1.18216157e+00
1.55520394e-01 -2.89067924e-01 -8.14017892e-01 -6.48685992e-01
-1.46617398e-01 -1.08074136e-01 6.19508088e-01 9.83206749e-01
-1.72888726e-01 2.88125604e-01 -5.49526453e-01 4.35363173e-01
8.38620067e-01 -2.48621359e-01 1.03611434e+00 -1.29627764e+00
-2.64146805e-01 -3.77070516e-01 -4.49264467e-01 -1.07537043e+00
1.37880757e-01 -6.80248499e-01 -5.35733178e-02 -1.26484346e+00
5.12038469e-01 -4.59593236e-01 -4.86519724e-01 4.18718219e-01
-6.80007041e-01 3.73243421e-01 1.85640901e-01 3.70566994e-01
-8.90979767e-01 5.22297859e-01 7.06655622e-01 -2.12229878e-01
-3.44634742e-01 -1.08075172e-01 -8.79366517e-01 5.75296938e-01
7.46836722e-01 -2.19217420e-01 -1.75448015e-01 -8.16650093e-01
-8.95896778e-02 -7.07615018e-01 5.79045713e-01 -1.20024955e+00
4.94613230e-01 2.46564031e-01 9.03817475e-01 -4.85711515e-01
3.78750592e-01 -6.64583564e-01 -9.74356309e-02 8.52586702e-02
-3.91986489e-01 3.06506932e-01 1.58240750e-01 6.42715693e-01
1.57902762e-01 6.81618378e-02 7.46637702e-01 3.78952594e-03
-5.47695577e-01 5.62745869e-01 4.10826430e-02 -7.47994632e-02
7.80087292e-01 -2.76876807e-01 -2.86143601e-01 -3.09446752e-01
-6.94042444e-01 4.69893098e-01 5.96532464e-01 4.72596973e-01
1.84766307e-01 -1.26877666e+00 -8.15327287e-01 2.27783084e-01
1.01796895e-01 1.35057583e-01 4.51545477e-01 6.61720335e-01
-1.06291167e-01 3.24779332e-01 -9.65473503e-02 -6.04611158e-01
-1.23910677e+00 4.59838808e-01 3.85730982e-01 -3.13779414e-01
-5.17841101e-01 8.68432522e-01 1.95456564e-01 -6.64240837e-01
4.62165445e-01 1.68714181e-01 -1.47712708e-01 -6.71279132e-02
7.33833969e-01 5.00507593e-01 -1.39770225e-01 -9.07761514e-01
-5.13242483e-01 6.57995999e-01 -3.75525624e-01 5.74356131e-02
1.28933668e+00 -3.14058721e-01 1.42179444e-01 -8.05837139e-02
1.17656696e+00 -2.10883543e-01 -1.51663601e+00 -2.76764363e-01
9.93002057e-02 -4.58557844e-01 -2.40606144e-01 -8.60967398e-01
-8.50201011e-01 7.47874618e-01 9.34415817e-01 -2.18033984e-01
8.41689765e-01 -2.97927797e-01 7.06004918e-01 3.57575953e-01
1.95697069e-01 -1.47833359e+00 1.56325728e-01 3.46673548e-01
5.08010328e-01 -1.61503232e+00 1.52354585e-02 -2.22558901e-01
-4.03821945e-01 7.37933099e-01 7.60064185e-01 -4.33013290e-02
4.56004620e-01 4.71356474e-02 1.05815046e-02 3.00858140e-01
1.66731712e-03 -2.64134139e-01 5.65381050e-01 7.04244912e-01
3.07213545e-01 4.02279422e-02 7.59197250e-02 7.45667279e-01
-2.44392425e-01 -8.02437514e-02 -8.86646658e-02 5.86255908e-01
-2.38612648e-02 -9.58333671e-01 -4.03939813e-01 2.73456424e-01
-3.16354573e-01 9.22797620e-02 -2.58509666e-01 6.36337340e-01
3.83678317e-01 1.04414845e+00 1.60622805e-01 -3.87025803e-01
3.03863645e-01 -8.06180388e-02 4.58895832e-01 -4.41199660e-01
-5.82076967e-01 -1.32070705e-01 -6.61417842e-02 -2.92481869e-01
-6.39504850e-01 -8.19496810e-01 -8.91345322e-01 -4.56405699e-01
-2.03083515e-01 -2.05308735e-01 4.80755985e-01 1.19637680e+00
6.05321169e-01 4.21526074e-01 3.23574781e-01 -1.07714128e+00
-4.01500732e-01 -1.23214281e+00 -2.23730624e-01 7.68601537e-01
2.05265269e-01 -6.59842789e-01 -1.39517531e-01 2.28248954e-01] | [14.683653831481934, 0.9695492386817932] |
21b390c2-ff20-4bfd-9019-f2d4aa5bbb1c | scale-selective-extended-local-binary-pattern | 1812.04174 | null | http://arxiv.org/abs/1812.04174v1 | http://arxiv.org/pdf/1812.04174v1.pdf | Scale Selective Extended Local Binary Pattern for Texture Classification | In this paper, we propose a new texture descriptor, scale selective extended
local binary pattern (SSELBP), to characterize texture images with scale
variations. We first utilize multi-scale extended local binary patterns (ELBP)
with rotation-invariant and uniform mappings to capture robust local micro- and
macro-features. Then, we build a scale space using Gaussian filters and
calculate the histogram of multi-scale ELBPs for the image at each scale.
Finally, we select the maximum values from the corresponding bins of
multi-scale ELBP histograms at different scales as scale-invariant features. A
comprehensive evaluation on public texture databases (KTH-TIPS and UMD) shows
that the proposed SSELBP has high accuracy comparable to state-of-the-art
texture descriptors on gray-scale-, rotation-, and scale-invariant texture
classification but uses only one-third of the feature dimension. | [] | 2018-12-11 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 1.60710230e-01 -8.49542201e-01 -4.18464422e-01 -3.57018888e-01
-8.83673728e-01 -3.63352507e-01 5.01183987e-01 2.20441237e-01
-1.81772962e-01 5.08662403e-01 -1.07654594e-01 8.76738355e-02
-5.12691677e-01 -1.04179454e+00 -1.51844025e-01 -1.01924431e+00
-5.96671849e-02 9.46314558e-02 9.97273982e-01 -2.54227698e-01
6.27318680e-01 1.17279291e+00 -1.64620364e+00 6.32327378e-01
3.61913234e-01 1.70435393e+00 -7.66071379e-02 6.75232828e-01
-1.59815408e-03 3.25601548e-01 -4.64325041e-01 4.92434986e-02
7.24055022e-02 -1.47620901e-01 -7.82183290e-01 8.18947628e-02
4.54862326e-01 5.21445051e-02 -3.70124102e-01 1.12250495e+00
4.83740747e-01 2.60160983e-01 7.96382368e-01 -7.35746861e-01
-7.90327132e-01 -3.49848032e-01 -8.56700122e-01 5.41550219e-01
3.43416482e-01 -5.43380380e-01 4.54963923e-01 -9.19694364e-01
9.66682374e-01 1.38542008e+00 7.23979592e-01 -2.88314641e-01
-1.33721876e+00 -3.74151379e-01 -6.27326846e-01 2.92268783e-01
-1.69313574e+00 -1.97943777e-01 5.60455084e-01 -1.87711641e-02
7.56204545e-01 6.64023578e-01 3.33325446e-01 5.79782188e-01
1.22247970e+00 2.45798975e-01 2.05833316e+00 -6.98251843e-01
2.10423023e-01 -2.13665783e-01 1.32734358e-01 8.97575319e-01
-4.38038707e-02 -1.26379311e-01 -5.96261561e-01 -3.77997130e-01
1.39445150e+00 -1.08666904e-01 2.12850407e-01 -3.85473371e-01
-1.07426739e+00 5.39627731e-01 2.09775656e-01 5.76738060e-01
-3.67605090e-01 -3.03949684e-01 3.06044966e-01 5.26006162e-01
7.31144249e-01 -4.36084494e-02 -3.83644313e-01 -3.25026721e-01
-6.41956806e-01 -1.11245416e-01 4.69152272e-01 4.88703936e-01
1.01380563e+00 -4.06194657e-01 -2.61932790e-01 1.35820401e+00
-2.29625449e-01 1.04014695e+00 6.54269814e-01 -5.87016165e-01
-2.41181567e-01 4.49890465e-01 7.15480819e-02 -1.72575879e+00
-3.93700242e-01 2.74553984e-01 -7.79331267e-01 2.51255870e-01
4.76650119e-01 7.95143068e-01 -9.85892177e-01 8.89881909e-01
4.06720161e-01 -2.21500948e-01 -2.84159362e-01 6.46309495e-01
5.62010050e-01 4.41332012e-01 -6.51470572e-02 1.48147689e-02
1.62509406e+00 -6.33180320e-01 -4.13017094e-01 2.00365216e-01
-1.41023725e-01 -1.21136200e+00 1.01780009e+00 1.96615741e-01
-6.08652353e-01 -8.13906848e-01 -9.12595153e-01 6.97671250e-02
-6.43445373e-01 4.73919868e-01 4.60835427e-01 6.97168589e-01
-9.41301942e-01 8.23978782e-01 -1.04230773e+00 -6.92328751e-01
-2.26087198e-01 1.14834294e-01 -8.51120353e-01 -6.36433735e-02
-7.89332032e-01 6.69683456e-01 -2.67399043e-01 -2.75332063e-01
1.40797896e-02 -7.90351443e-03 -8.50195587e-01 -2.13316441e-01
-1.09416813e-01 3.49378675e-01 4.66765195e-01 -7.54068077e-01
-1.77373159e+00 1.18937159e+00 -6.16815925e-01 2.22889200e-01
2.30654944e-02 4.96466964e-01 -8.30653667e-01 6.74758494e-01
2.09821448e-01 2.16691777e-01 1.07624769e+00 -8.09364378e-01
-7.03446448e-01 -2.88664043e-01 -4.86026436e-01 1.84555445e-02
-5.54871336e-02 4.38545048e-01 -5.35132706e-01 -9.22284126e-01
7.63314366e-01 -7.04990625e-01 1.83615312e-01 -6.83209524e-02
7.29607232e-03 -2.94821739e-01 8.21037710e-01 -9.87389982e-01
1.22088265e+00 -2.60829520e+00 -1.55965239e-01 7.34345555e-01
-2.12869480e-01 -2.30554387e-01 -3.62046957e-01 1.48255497e-01
3.84134054e-02 -1.62971243e-01 4.11841303e-01 1.52062073e-01
-2.86790561e-02 2.58368224e-01 -2.01396957e-01 6.33610666e-01
3.77861783e-02 5.36442757e-01 -5.58668256e-01 -6.18661404e-01
4.31853861e-01 3.67009342e-01 1.08486101e-01 -3.54887903e-01
7.14395821e-01 1.64289773e-01 -5.58367789e-01 1.13697004e+00
1.19407868e+00 -1.28674284e-02 6.19698092e-02 -4.64465052e-01
-3.74975532e-01 -3.41367334e-01 -1.16869819e+00 8.05810690e-01
-3.36940467e-01 5.18846631e-01 -2.33516067e-01 -7.99173713e-01
1.48475289e+00 -1.44162253e-01 5.00663400e-01 -1.08286500e+00
8.71196240e-02 5.03771722e-01 -6.25446081e-01 -2.13317499e-01
6.86404705e-01 4.45560031e-02 -2.86873817e-01 6.33795932e-02
1.09188177e-01 -2.74654068e-02 3.08847278e-01 -6.32169783e-01
9.60763454e-01 2.28950381e-02 6.21861398e-01 -7.73510098e-01
8.01713288e-01 -3.82103652e-01 3.30486923e-01 8.52182388e-01
-4.63550389e-01 6.66981816e-01 6.31835043e-01 -7.00039625e-01
-1.07041037e+00 -1.11607718e+00 -8.21093559e-01 1.30053473e+00
4.97837096e-01 -1.51983038e-01 -5.34682333e-01 -4.04625058e-01
3.34226340e-01 -2.63971984e-01 -9.11525846e-01 2.49085665e-01
-4.75609571e-01 -7.14369357e-01 3.50355536e-01 3.83129030e-01
7.27986932e-01 -9.99736905e-01 -4.57834691e-01 9.94224548e-02
5.01375757e-02 -9.93822217e-01 -5.59141040e-01 1.63444951e-01
-5.31055570e-01 -9.77572918e-01 -8.53593767e-01 -9.98409390e-01
6.46733403e-01 2.59725481e-01 5.93026996e-01 -5.59878886e-01
-6.07085764e-01 3.34681451e-01 -5.79725981e-01 2.82616585e-01
-1.63945287e-01 -3.23535502e-01 -7.41616786e-02 3.73165131e-01
4.30896044e-01 -2.00747147e-01 -5.16533792e-01 1.10809386e+00
-5.84591210e-01 -2.29569450e-01 6.32102668e-01 1.02130818e+00
1.44008958e+00 5.19059956e-01 1.69360563e-01 -5.58740534e-02
4.43212062e-01 5.18291816e-02 -5.11338115e-01 6.40587091e-01
-9.93538648e-02 -6.23425003e-03 5.42159140e-01 -5.95184982e-01
-1.01380086e+00 -4.19263452e-01 1.76737100e-01 1.53998673e-01
-3.79707426e-01 3.55096012e-01 3.13862473e-01 -1.07409000e+00
5.92643321e-01 6.90895379e-01 -1.21664349e-03 -5.24788201e-01
-3.25354226e-02 8.48115027e-01 5.47875345e-01 -8.70028675e-01
2.42914990e-01 8.97833765e-01 4.35376674e-01 -1.23214948e+00
-2.79106528e-01 -6.29136980e-01 -8.12824547e-01 6.17754683e-02
7.15507269e-01 -3.71197462e-01 -4.33439910e-01 1.04723954e+00
-3.30502003e-01 -3.19799751e-01 -1.48756534e-01 1.48562416e-01
-6.85795426e-01 6.17267072e-01 -8.35017264e-01 -3.24478328e-01
-1.47749603e-01 -1.30666995e+00 1.43773043e+00 4.21853364e-01
1.48968995e-02 -9.82448936e-01 -2.63053719e-02 -2.97908008e-01
8.95323813e-01 4.68990505e-01 8.75689685e-01 -9.12628174e-02
-5.28300367e-02 -5.07751644e-01 -6.45794868e-01 3.87906879e-01
5.89214861e-01 9.27283764e-02 -2.25506023e-01 -4.22862500e-01
-2.74118483e-01 -1.94961637e-01 6.13435328e-01 6.93514407e-01
1.27722454e+00 1.94786429e-01 -3.21945578e-01 8.44014764e-01
1.51311922e+00 2.09258616e-01 6.26185298e-01 6.75755918e-01
-3.89683917e-02 1.17313765e-01 8.73123348e-01 5.13321459e-01
5.53307161e-02 8.16090405e-01 -6.48977816e-01 -3.83637488e-01
-3.39103222e-01 6.71034977e-02 3.61384079e-02 4.65909928e-01
-3.42033416e-01 4.08080548e-01 -5.67119241e-01 2.05189794e-01
-1.43116939e+00 -7.83151984e-01 2.64075428e-01 2.17998290e+00
7.66349733e-01 2.62363930e-04 -1.67724088e-01 3.79983634e-02
9.62179840e-01 2.21665829e-01 -9.30800140e-02 -6.05374157e-01
-6.67113006e-01 1.02851009e+00 9.19975042e-01 2.59704918e-01
-1.62853062e+00 1.01974559e+00 6.76864004e+00 1.59431410e+00
-1.76394928e+00 -1.80493866e-03 8.14772964e-01 8.32908571e-01
4.08041388e-01 -1.90141276e-01 -6.56559348e-01 3.47590268e-01
5.45951486e-01 -5.13992533e-02 4.90162849e-01 8.62696111e-01
-9.95659456e-02 -6.45372510e-01 -2.69809157e-01 1.01855230e+00
4.54457253e-02 -8.57114851e-01 -1.01664059e-01 -2.58665159e-02
9.10915077e-01 -2.25657269e-01 3.97760957e-01 -6.04314171e-02
-1.75842628e-01 -7.39402711e-01 5.39417207e-01 6.45971417e-01
1.46936846e+00 -6.85366929e-01 8.03113937e-01 -5.75909853e-01
-1.64165020e+00 1.37747481e-01 -1.05862176e+00 3.38186592e-01
-4.01772469e-01 3.09585184e-01 -1.90827534e-01 5.77434182e-01
9.45573926e-01 3.56953770e-01 -1.05837727e+00 9.85802472e-01
1.22533776e-01 2.31574669e-01 -5.57063699e-01 -4.73178923e-02
8.02516639e-02 -1.09561481e-01 1.52452767e-01 1.32016313e+00
6.60569429e-01 -8.45069736e-02 1.99101120e-01 1.00599185e-01
4.73939717e-01 4.83446270e-01 -7.63565078e-02 2.05810800e-01
5.08768678e-01 1.37643611e+00 -1.38809478e+00 -4.40919906e-01
-3.87267470e-01 1.12771487e+00 1.26270549e-02 3.40703934e-01
-4.12616014e-01 -8.91757667e-01 3.51793826e-01 -1.02467969e-01
5.87263525e-01 -2.97991037e-01 -2.39187196e-01 -9.30164635e-01
-1.53805673e-01 -8.87152553e-01 4.38376695e-01 -6.99166358e-01
-1.41521466e+00 6.13034010e-01 -1.60291959e-02 -1.18779683e+00
2.88444638e-01 -1.06528533e+00 -1.46408528e-01 1.01842201e+00
-1.27737880e+00 -1.50603616e+00 -3.52051735e-01 7.05756307e-01
7.75376558e-02 -2.07251877e-01 1.06816185e+00 -1.14001684e-01
3.98530848e-02 8.61689031e-01 7.63024449e-01 4.68491055e-02
1.13682652e+00 -1.13842154e+00 2.44968712e-01 4.30174619e-01
-4.01579857e-01 4.77306396e-01 3.93239111e-01 -7.45992124e-01
-1.26194870e+00 -6.63043797e-01 7.05336154e-01 -1.55694410e-01
7.96476245e-01 3.64334956e-02 -6.10438049e-01 1.60031796e-01
-3.19382548e-01 5.44489920e-01 5.08747160e-01 -1.38875246e-01
-4.53265250e-01 -4.06021863e-01 -1.41755998e+00 2.11339384e-01
4.66206938e-01 -7.59552062e-01 -3.37524652e-01 -3.16269174e-02
-1.91404179e-01 -6.21084750e-01 -1.34043503e+00 6.28116548e-01
1.14091539e+00 -9.76061463e-01 1.02966464e+00 -3.70441563e-02
1.19821720e-01 -2.58011490e-01 -5.54630756e-01 -1.04252088e+00
-9.47400749e-01 -8.24624076e-02 5.99929333e-01 6.93612218e-01
-1.23236939e-01 -9.73207355e-01 4.54657197e-01 -1.17182389e-01
2.31174275e-01 -7.42130637e-01 -1.36140800e+00 -8.59463692e-01
-2.41392136e-01 2.41017252e-01 5.67381084e-01 7.38363385e-01
2.60504663e-01 -6.47901654e-01 -2.26620615e-01 7.87932351e-02
6.72472239e-01 4.96158093e-01 5.39579213e-01 -7.89915144e-01
-1.03157885e-01 -5.95973849e-01 -1.15619719e+00 -6.85917258e-01
-8.23478475e-02 -4.98765200e-01 -1.90205023e-01 -1.05548000e+00
3.31001282e-01 -5.84800243e-01 -7.62044907e-01 5.92816889e-01
4.34801765e-02 9.22633469e-01 -3.40450525e-01 2.14542270e-01
-5.38081050e-01 -2.34414507e-02 1.49602771e+00 -5.36464490e-02
-3.35490443e-02 -2.63957322e-01 -5.83936572e-02 4.78766859e-01
5.39315701e-01 -1.07951760e-01 3.98846805e-01 3.60774547e-01
-3.66497934e-01 1.23460233e-01 1.23217121e-01 -1.07924485e+00
6.67741150e-03 -6.35671079e-01 8.45603466e-01 -4.65934962e-01
1.99101254e-01 -2.46895671e-01 -1.14653260e-01 2.69747317e-01
8.65231007e-02 1.01161346e-01 1.46249905e-01 2.20164269e-01
-6.41671836e-01 2.71409541e-01 1.19449317e+00 2.20191285e-01
-1.18726647e+00 7.97276050e-02 -4.91922498e-01 -6.91444457e-01
7.76070058e-01 -3.91652405e-01 -5.84390640e-01 9.93273780e-02
-8.43843400e-01 -6.80611134e-01 8.91704321e-01 4.78633255e-01
6.35298431e-01 -1.63986647e+00 -3.80121082e-01 8.10726464e-01
3.30902904e-01 -9.36723411e-01 7.15760648e-01 9.81759548e-01
-1.01623893e+00 4.51107144e-01 -9.26724613e-01 -6.73110485e-01
-1.53566563e+00 2.23528892e-01 2.59969294e-01 -3.12146068e-01
-3.72534245e-01 6.81824327e-01 2.15737638e-03 -2.63072342e-01
-4.06160951e-01 -4.87206429e-01 -1.51643515e-01 -1.52514055e-01
6.51490986e-01 5.30967712e-01 3.86577994e-01 -1.03648770e+00
-4.23650980e-01 1.30564797e+00 2.29198501e-01 3.23201455e-02
1.07322526e+00 -1.94386244e-01 -7.36933589e-01 2.10662633e-01
1.43078315e+00 3.71913373e-01 -1.01996791e+00 -3.11756134e-01
-2.10342094e-01 -9.24739778e-01 -9.14834440e-02 -5.89331925e-01
-4.40232247e-01 2.98921227e-01 1.10365593e+00 2.05983758e-01
1.33585572e+00 -4.53489050e-02 4.62237924e-01 1.35780126e-01
7.95673490e-01 -1.35281718e+00 1.81290969e-01 7.27737248e-01
9.71711695e-01 -8.11275005e-01 2.58141290e-02 -6.08151436e-01
-3.78345460e-01 1.64619029e+00 1.58025235e-01 -3.96101654e-01
9.72104430e-01 2.80020088e-01 2.04080939e-01 1.38088420e-01
-1.17230177e-01 7.50313103e-02 5.75050175e-01 5.38814127e-01
3.55950832e-01 5.08546770e-01 -6.59440160e-01 1.57219410e-01
-5.53939827e-02 -1.12686910e-01 2.09279396e-02 1.13447022e+00
-7.13617444e-01 -1.27416289e+00 -8.93834949e-01 4.62487996e-01
-4.35428113e-01 3.24177414e-01 1.03641607e-01 5.78999221e-01
6.28050193e-02 6.56762064e-01 3.04329365e-01 -5.43488026e-01
1.69650719e-01 9.06070136e-03 9.43056881e-01 1.13583826e-01
2.75353223e-01 4.22682464e-01 -1.71659783e-01 -7.87781417e-01
-3.15178454e-01 -5.34644902e-01 -7.91952372e-01 -4.02337909e-01
-5.54668866e-02 -1.03859395e-01 5.63163400e-01 5.33766389e-01
3.47328722e-01 1.77759510e-02 8.17894757e-01 -1.02844250e+00
-6.31957173e-01 -9.32749748e-01 -1.33043325e+00 6.56609237e-01
9.52839777e-02 -9.67122018e-01 -1.96620643e-01 -1.08880915e-01] | [10.431018829345703, -0.3597431480884552] |
7bdd3c69-b5bc-419c-b05a-00b4cd4b815e | vision-transformer-for-contrastive-clustering | 2206.12925 | null | https://arxiv.org/abs/2206.12925v2 | https://arxiv.org/pdf/2206.12925v2.pdf | Vision Transformer for Contrastive Clustering | Vision Transformer (ViT) has shown its advantages over the convolutional neural network (CNN) with its ability to capture global long-range dependencies for visual representation learning. Besides ViT, contrastive learning is another popular research topic recently. While previous contrastive learning works are mostly based on CNNs, some recent studies have attempted to combine ViT and contrastive learning for enhanced self-supervised learning. Despite the considerable progress, these combinations of ViT and contrastive learning mostly focus on the instance-level contrastiveness, which often overlook the global contrastiveness and also lack the ability to directly learn the clustering result (e.g., for images). In view of this, this paper presents a novel deep clustering approach termed Vision Transformer for Contrastive Clustering (VTCC), which for the first time, to our knowledge, unifies the Transformer and the contrastive learning for the image clustering task. Specifically, with two random augmentations performed on each image, we utilize a ViT encoder with two weight-sharing views as the backbone. To remedy the potential instability of the ViT, we incorporate a convolutional stem to split each augmented sample into a sequence of patches, which uses multiple stacked small convolutions instead of a big convolution in the patch projection layer. By learning the feature representations for the sequences of patches via the backbone, an instance projector and a cluster projector are further utilized to perform the instance-level contrastive learning and the global clustering structure learning, respectively. Experiments on eight image datasets demonstrate the stability (during the training-from-scratch) and the superiority (in clustering performance) of our VTCC approach over the state-of-the-art. | ['Jian-Huang Lai', 'Chang-Dong Wang', 'Ding-Hua Chen', 'Dong Huang', 'Bowen Zhu', 'Hua-Bao Ling'] | 2022-06-26 | null | null | null | null | ['image-clustering'] | ['computer-vision'] | [ 1.50522545e-01 -3.00600469e-01 9.87974033e-02 -2.67992139e-01
-4.54591304e-01 -2.86858141e-01 7.73732662e-01 -1.63140178e-01
-2.73916692e-01 2.86835432e-01 -9.81645957e-02 -4.13965657e-02
-2.07907706e-01 -6.64272726e-01 -9.30481315e-01 -1.20656872e+00
1.46446586e-01 3.17970425e-01 2.13623866e-01 5.80223557e-03
2.00534046e-01 3.29165131e-01 -1.87827432e+00 4.13112372e-01
8.29578221e-01 9.78671193e-01 5.65512955e-01 2.47268170e-01
-2.53094703e-01 9.13489759e-01 -4.91444498e-01 -2.12698534e-01
2.31584281e-01 -4.55396384e-01 -5.47606051e-01 4.72684592e-01
4.82977360e-01 2.26649195e-01 -2.60437727e-01 9.20220852e-01
3.90801281e-01 1.14618666e-01 6.21726751e-01 -1.39594901e+00
-7.66048253e-01 4.72641885e-01 -1.01268792e+00 -7.10078180e-02
-8.06491077e-02 2.51188368e-01 9.58777487e-01 -1.09391499e+00
4.43351924e-01 9.70944464e-01 7.82584071e-01 4.17267174e-01
-1.12137246e+00 -7.03773975e-01 3.25695544e-01 3.55071545e-01
-1.43802571e+00 -8.51630792e-02 1.16116858e+00 -4.83626425e-01
8.16525042e-01 2.33383235e-02 8.13151240e-01 9.43192780e-01
-1.04503334e-01 8.63659978e-01 1.34147859e+00 -4.50292736e-01
2.64715135e-01 1.50132179e-01 1.19267009e-01 7.69393563e-01
-1.29266754e-01 1.49624720e-01 -2.23033547e-01 1.82164237e-01
7.14001536e-01 4.67763722e-01 -2.86167204e-01 -5.99444866e-01
-9.72023904e-01 7.67873406e-01 8.88938189e-01 4.21851397e-01
-2.49020845e-01 1.74268410e-01 2.43425101e-01 1.52882874e-01
3.78450990e-01 3.02356124e-01 -3.75858635e-01 2.16578260e-01
-1.05406404e+00 -1.36565998e-01 5.76381683e-01 7.49987900e-01
1.17282534e+00 6.31965995e-02 -2.30029270e-01 1.01193571e+00
3.86772230e-02 2.22361252e-01 5.80795467e-01 -5.53521991e-01
2.90529102e-01 9.30173397e-01 -5.51089942e-01 -9.46879625e-01
-3.21749449e-01 -8.00038218e-01 -1.33116937e+00 4.46845293e-01
7.13927597e-02 4.75415662e-02 -1.23722827e+00 1.78598166e+00
1.81931704e-01 6.51856899e-01 -7.82355200e-03 8.11555743e-01
7.28865921e-01 7.48296916e-01 -1.22945741e-01 -2.69147962e-01
1.17627251e+00 -1.29875457e+00 -4.43501294e-01 -4.62660380e-02
4.50929195e-01 -7.90628314e-01 1.18989527e+00 4.73382771e-01
-9.78547394e-01 -9.22451377e-01 -1.16114020e+00 2.20319033e-02
-6.04079962e-01 1.88474923e-01 5.43334484e-01 3.39515150e-01
-1.23083854e+00 5.84032655e-01 -9.57708478e-01 -3.44089180e-01
5.55029452e-01 4.20375824e-01 -2.60713965e-01 -2.91930109e-01
-7.26163149e-01 6.17851913e-01 2.11266831e-01 1.05406627e-01
-8.65559638e-01 -6.57482862e-01 -6.53814793e-01 2.29668140e-01
3.76553297e-01 -5.70775151e-01 5.86864054e-01 -1.41655481e+00
-1.50444412e+00 6.62580371e-01 -6.24314137e-02 -2.48295605e-01
2.13836536e-01 3.72511223e-02 4.59169112e-02 2.39025474e-01
8.18953365e-02 8.00470889e-01 1.03934646e+00 -1.64760518e+00
-5.54459453e-01 -3.16248745e-01 -1.26112491e-01 3.70206207e-01
-6.84899449e-01 -2.23227322e-01 -8.17748666e-01 -6.47186995e-01
4.53615598e-02 -9.57494378e-01 -2.54082680e-01 -1.43500388e-01
-2.69064188e-01 -2.89728522e-01 1.30574918e+00 -2.79637456e-01
1.07851660e+00 -2.29268861e+00 2.96633363e-01 2.50025809e-01
2.73647964e-01 4.35630351e-01 -2.92608917e-01 3.86998504e-01
-3.53578836e-01 -5.26709557e-02 -7.00751305e-01 -7.23830104e-01
-2.64670372e-01 4.01229948e-01 1.27619758e-01 3.92634064e-01
4.20749396e-01 1.08774221e+00 -7.40013659e-01 -4.42544937e-01
5.56463003e-01 6.71731412e-01 -5.44646144e-01 3.49932253e-01
-1.29998654e-01 4.71204042e-01 1.45189390e-02 3.91847074e-01
9.57713902e-01 -4.79921967e-01 1.75312445e-01 -3.12214434e-01
-3.40708613e-01 -2.31187552e-01 -1.06420612e+00 1.72397578e+00
-3.74601066e-01 3.16104531e-01 6.12832755e-02 -1.56024861e+00
8.63906205e-01 1.89403221e-01 6.02823257e-01 -6.88960552e-01
-1.54868320e-01 8.52063149e-02 -1.28870144e-01 -4.33093965e-01
1.77416727e-01 -1.35634929e-01 2.00136632e-01 3.73715281e-01
2.57078439e-01 1.12000331e-01 -3.46466973e-02 4.71753217e-02
8.48868966e-01 2.37522677e-01 -2.24287473e-02 -3.02531987e-01
5.76044917e-01 -1.10758394e-01 5.36274254e-01 4.97101426e-01
1.26925543e-01 9.85975921e-01 3.48791063e-01 -3.06493849e-01
-1.06959546e+00 -9.28633809e-01 1.76836532e-02 8.10756862e-01
2.07563967e-01 -4.04834658e-01 -8.34846675e-01 -7.30767548e-01
-3.36906880e-01 1.78226307e-01 -8.27616572e-01 -2.90844411e-01
-6.18591845e-01 -9.05573010e-01 3.88501316e-01 6.62529171e-01
9.12281573e-01 -1.14083934e+00 -3.40926886e-01 5.28181568e-02
-1.07008889e-01 -9.26746368e-01 -3.43430340e-01 5.83170533e-01
-6.99387670e-01 -1.09862649e+00 -6.69923782e-01 -1.27444470e+00
6.79594040e-01 6.68709755e-01 8.57833505e-01 3.35363865e-01
-3.03446084e-01 3.47391725e-01 -4.34285402e-01 -5.68988733e-02
7.40583092e-02 2.29943633e-01 -3.21643740e-01 3.24810028e-01
2.57612497e-01 -8.85824859e-01 -7.01801777e-01 2.32205838e-01
-1.05755746e+00 7.67041072e-02 8.33092630e-01 1.20490825e+00
6.86794162e-01 2.66190112e-01 4.04778719e-01 -8.73256385e-01
3.16430479e-01 -5.69420159e-01 -3.89030874e-01 2.41023883e-01
-6.58307076e-01 -8.20508674e-02 8.59841466e-01 -4.81558740e-01
-9.98713613e-01 3.93287241e-01 -1.10943042e-01 -9.92080808e-01
-2.21458465e-01 6.28771007e-01 -1.43822223e-01 -1.86122864e-01
4.47430015e-01 4.45858538e-01 1.40184298e-01 -3.69548351e-01
5.38132846e-01 4.85951692e-01 5.36209106e-01 -5.27730942e-01
9.07873034e-01 6.37133479e-01 -1.27344951e-01 -8.28271985e-01
-7.69648850e-01 -6.71249151e-01 -8.24965954e-01 -1.01729028e-01
1.13615251e+00 -9.29849923e-01 -5.08648098e-01 6.19048238e-01
-8.10116470e-01 -3.58116448e-01 -2.94756949e-01 3.33751261e-01
-4.37016487e-01 5.11257052e-01 -4.90490377e-01 -5.76718867e-01
-1.29075140e-01 -1.20313323e+00 9.19811785e-01 1.50555775e-01
3.91630441e-01 -1.10154474e+00 -6.47701621e-02 2.28421390e-01
3.42259228e-01 3.56044948e-01 8.60010505e-01 -2.84332335e-01
-6.05114281e-01 1.93025678e-01 -3.02572787e-01 6.58567607e-01
1.10318601e-01 1.84305459e-01 -1.16844869e+00 -4.07050282e-01
1.26470089e-01 -3.96657944e-01 1.20242763e+00 4.53219265e-01
1.41295612e+00 -2.05298737e-01 -3.70243341e-01 9.07785296e-01
1.79338622e+00 1.55950382e-01 8.29984128e-01 2.34687939e-01
1.07965374e+00 5.05273342e-01 2.34001815e-01 2.84765989e-01
3.66749674e-01 6.35122657e-01 6.18656039e-01 -6.34454668e-01
-3.78562093e-01 -1.36368006e-01 2.10913301e-01 1.03856277e+00
-3.21895868e-01 -5.55711146e-03 -7.29505122e-01 6.11405492e-01
-1.98101366e+00 -9.83223796e-01 1.28327440e-02 2.11395979e+00
5.92477798e-01 -1.39435962e-01 9.19282902e-03 2.93231189e-01
8.42343807e-01 1.68716833e-01 -4.04015690e-01 -1.30947992e-01
-2.83153951e-01 4.63749647e-01 1.33129910e-01 2.21792743e-01
-1.12377965e+00 1.02085280e+00 5.21191931e+00 1.04308879e+00
-1.46802962e+00 1.04674920e-01 6.99280500e-01 1.10729471e-01
-7.98494592e-02 1.05852976e-01 -4.66322392e-01 4.30746526e-01
5.24865389e-01 5.24939179e-01 4.93391305e-01 7.28241026e-01
-5.36702871e-02 1.80194452e-01 -8.81827414e-01 1.10575175e+00
2.17985973e-01 -1.32408559e+00 2.32135370e-01 1.52690724e-01
1.11297369e+00 3.73584218e-02 4.29891050e-01 4.17488903e-01
3.55313607e-02 -1.15334296e+00 6.09561741e-01 4.58049685e-01
6.01336598e-01 -9.67677355e-01 8.94738495e-01 2.99979120e-01
-1.42781281e+00 -3.43302995e-01 -5.37560582e-01 -1.29288167e-01
-2.72783130e-01 6.30180776e-01 -4.03152287e-01 7.34774232e-01
9.81576741e-01 1.09384823e+00 -7.96172202e-01 8.92842412e-01
-1.62902012e-01 7.12643564e-01 -1.05164826e-01 1.50610134e-01
5.81077397e-01 -4.15963501e-01 1.86139852e-01 1.30645061e+00
2.37313181e-01 -4.58124988e-02 4.12499189e-01 9.27387774e-01
-1.28992796e-01 -1.20248680e-03 -5.25180519e-01 2.70549566e-01
2.71183819e-01 1.60192692e+00 -7.55396605e-01 -3.90402704e-01
-5.88565528e-01 1.01386094e+00 5.85388601e-01 4.13214773e-01
-7.18338788e-01 -2.52786756e-01 4.14503276e-01 4.13367748e-02
9.92941380e-01 -1.09548114e-01 -4.77452338e-01 -1.01003408e+00
1.60672083e-01 -8.83572340e-01 2.89511263e-01 -6.12381160e-01
-1.49968970e+00 6.55418694e-01 -5.73096387e-02 -1.39565063e+00
8.67338553e-02 -4.49188769e-01 -8.72474372e-01 5.25963187e-01
-1.74517274e+00 -1.54235983e+00 -5.70452929e-01 1.11013162e+00
5.12726843e-01 -1.03122011e-01 5.50626099e-01 1.83132544e-01
-6.38261139e-01 6.06273174e-01 2.37356216e-01 4.44491394e-02
6.12128973e-01 -1.37429571e+00 -1.00492634e-01 6.93108678e-01
9.32964459e-02 7.33869255e-01 1.66876048e-01 -2.56613791e-01
-1.11651230e+00 -1.41243804e+00 3.97886544e-01 -1.88606486e-01
3.78466547e-01 -5.59232414e-01 -1.05096817e+00 4.74762291e-01
7.19691038e-01 1.64504170e-01 5.52948356e-01 5.77437989e-02
-5.07539332e-01 -3.60090524e-01 -9.28485513e-01 5.04739583e-01
9.92242694e-01 -4.26215440e-01 -4.86333221e-01 1.01310506e-01
6.73435867e-01 5.42075522e-02 -7.95486629e-01 4.85035837e-01
2.72480190e-01 -1.21702361e+00 8.64256978e-01 -4.99752304e-03
6.05595171e-01 -6.63060308e-01 -8.17024112e-02 -1.32479119e+00
-6.32468998e-01 -3.24530363e-01 9.25076827e-02 1.44072759e+00
2.19999567e-01 -4.22583908e-01 7.43803918e-01 -8.93766508e-02
-4.97855544e-01 -9.46111143e-01 -7.61819363e-01 -5.96926689e-01
1.99290097e-01 -1.32947147e-01 4.76905107e-01 1.24807906e+00
-2.19478264e-01 6.59542263e-01 -3.94365251e-01 1.67332202e-01
6.38247907e-01 2.24957779e-01 6.90642059e-01 -1.33242416e+00
-4.35117722e-01 -4.21176106e-01 -3.65406215e-01 -9.45166469e-01
4.19024043e-02 -9.68020201e-01 9.51491445e-02 -1.49901462e+00
5.25300145e-01 -5.00999272e-01 -5.28388262e-01 6.02183402e-01
-3.44620436e-01 4.54730541e-01 2.55977958e-01 3.34952086e-01
-6.56324387e-01 6.97337210e-01 1.30289555e+00 -3.24298769e-01
-1.95066482e-01 -1.79450110e-01 -7.56394744e-01 5.31717956e-01
7.32212782e-01 -3.33780557e-01 -6.69003963e-01 -3.24069023e-01
2.18458381e-02 -3.60688746e-01 4.60332334e-01 -1.34075522e+00
5.12749493e-01 9.03013647e-02 4.82649297e-01 -6.99265063e-01
3.99834126e-01 -7.72231042e-01 7.19671026e-02 4.27681506e-01
-1.16745055e-01 6.58164844e-02 5.98781034e-02 6.54866934e-01
-5.69602728e-01 -2.00147834e-02 9.49508011e-01 -4.25741822e-01
-6.63449705e-01 5.77176869e-01 -3.23837280e-01 -2.33819604e-01
1.00471210e+00 -4.39513922e-01 -2.75556594e-01 -1.38821274e-01
-5.99374890e-01 2.04780579e-01 5.97370625e-01 1.93574354e-01
8.43169332e-01 -1.28397572e+00 -5.84053338e-01 3.47210705e-01
1.44531980e-01 3.42524171e-01 2.84118384e-01 9.72407937e-01
-1.83669329e-01 6.98087439e-02 -3.46764714e-01 -1.01126039e+00
-1.07629538e+00 1.01305687e+00 3.01883250e-01 -4.45769221e-01
-6.88465655e-01 8.13374996e-01 6.70363724e-01 -5.43256760e-01
3.24640602e-01 -5.74595295e-02 -4.44573224e-01 5.14797159e-02
3.29601496e-01 4.56348583e-02 9.68275741e-02 -3.60551745e-01
-2.25536644e-01 9.30692017e-01 -2.00781181e-01 2.95734294e-02
1.62423670e+00 -1.40256703e-01 -3.75154763e-01 4.31351602e-01
1.50554550e+00 -3.14830929e-01 -1.55634141e+00 -2.08529085e-01
-1.50543645e-01 -1.37182981e-01 -1.64381247e-02 -5.65506518e-01
-1.53885400e+00 1.23750675e+00 6.74477458e-01 8.99577513e-02
1.41971040e+00 -8.37443247e-02 6.13086998e-01 2.21429348e-01
-1.23479534e-02 -1.07273412e+00 6.41336143e-01 3.38902861e-01
6.56198502e-01 -1.18544745e+00 -1.89010695e-01 -3.80535275e-01
-7.13059723e-01 1.09662604e+00 7.72713423e-01 -4.49274331e-01
8.28991115e-01 2.78491437e-01 1.89380512e-01 -4.32495713e-01
-7.82647789e-01 -4.89207804e-01 7.34452084e-02 8.73891830e-01
3.51383328e-01 -1.28509134e-01 -6.75717220e-02 1.77608192e-01
2.08354946e-02 -2.14633077e-01 3.31809133e-01 8.14730167e-01
-2.77996689e-01 -1.00063038e+00 -2.00730294e-01 2.93296337e-01
-2.23325029e-01 -2.29114532e-01 -4.45335805e-01 8.21608961e-01
5.22821546e-01 9.67919827e-01 2.39045754e-01 -6.99848831e-01
1.26347512e-01 -1.44834369e-01 4.07239765e-01 -5.77003658e-01
-9.50341523e-01 2.60599762e-01 -5.52075803e-01 -4.36152101e-01
-9.62370813e-01 -5.18154502e-01 -1.06998086e+00 -5.02823666e-02
-3.40593219e-01 1.35367200e-01 4.90600646e-01 9.41779375e-01
2.35540316e-01 7.47643054e-01 9.97823596e-01 -9.00756836e-01
-2.11848959e-01 -9.84321415e-01 -6.15977705e-01 5.09825289e-01
3.11948448e-01 -6.35122418e-01 -4.61537421e-01 1.00346930e-01] | [9.188694953918457, 2.9387574195861816] |
d92cc409-d444-4419-979c-1bb5c1b8b3e2 | a-deep-analysis-of-transfer-learning-based | 2304.05022 | null | https://arxiv.org/abs/2304.05022v1 | https://arxiv.org/pdf/2304.05022v1.pdf | A Deep Analysis of Transfer Learning Based Breast Cancer Detection Using Histopathology Images | Breast cancer is one of the most common and dangerous cancers in women, while it can also afflict men. Breast cancer treatment and detection are greatly aided by the use of histopathological images since they contain sufficient phenotypic data. A Deep Neural Network (DNN) is commonly employed to improve accuracy and breast cancer detection. In our research, we have analyzed pre-trained deep transfer learning models such as ResNet50, ResNet101, VGG16, and VGG19 for detecting breast cancer using the 2453 histopathology images dataset. Images in the dataset were separated into two categories: those with invasive ductal carcinoma (IDC) and those without IDC. After analyzing the transfer learning model, we found that ResNet50 outperformed other models, achieving accuracy rates of 90.2%, Area under Curve (AUC) rates of 90.0%, recall rates of 94.7%, and a marginal loss of 3.5%. | ['Ahmed Abdelgawad', 'Muntasir Mamun', 'Md Ishtyaq Mahmud'] | 2023-04-11 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 4.31049941e-03 2.20803738e-01 -4.95781064e-01 -2.05540717e-01
-7.30653524e-01 -2.25950569e-01 3.77504498e-01 5.87179840e-01
-6.50154948e-01 7.58296490e-01 -7.45142773e-02 -7.72128701e-01
1.24942929e-01 -9.53481495e-01 -3.18429023e-01 -8.52618575e-01
-1.43277019e-01 3.59385043e-01 2.13742226e-01 -1.39886355e-02
-1.11393981e-01 6.33980870e-01 -8.80539119e-01 4.71854806e-01
9.26680982e-01 1.17536688e+00 -8.13545007e-03 6.60997629e-01
1.55936424e-02 8.35464060e-01 -3.56496096e-01 -4.37438577e-01
-1.77346542e-01 -1.65624470e-01 -7.39273310e-01 -3.77462476e-01
3.16904932e-02 -4.11696285e-01 -4.50718462e-01 1.12999034e+00
5.82337677e-01 -5.93155324e-01 8.06092262e-01 -7.68845499e-01
-4.13655192e-01 3.61940056e-01 -8.48830640e-01 2.24498197e-01
-3.20740938e-01 8.71891528e-02 4.72188771e-01 -5.25427401e-01
7.43216574e-01 8.83329391e-01 1.05628693e+00 9.01176512e-01
-9.92013335e-01 -9.69186604e-01 -7.86318779e-01 2.30563685e-01
-1.37306869e+00 -2.89135784e-01 2.74533361e-01 -3.35198045e-01
5.53966343e-01 2.27363080e-01 6.95731699e-01 8.34655464e-01
9.24577236e-01 6.94660544e-01 9.85243261e-01 -4.01165903e-01
2.01027200e-01 1.20407358e-01 1.09451994e-01 7.55036354e-01
5.68004489e-01 -1.63892042e-02 9.48522091e-02 -1.72468394e-01
7.78920233e-01 3.14932585e-01 2.81360112e-02 1.26162782e-01
-6.79683387e-01 8.62145483e-01 9.21851993e-01 5.31846344e-01
-2.58319110e-01 1.10479593e-01 6.55470788e-01 1.59027204e-01
3.78168017e-01 1.07534505e-01 -2.08538949e-01 3.72856230e-01
-5.66085756e-01 -1.77704602e-01 4.01575834e-01 4.72119659e-01
4.62815434e-01 -2.47416466e-01 -1.55101389e-01 9.59044933e-01
2.69670069e-01 3.95556897e-01 8.90174448e-01 -4.88633037e-01
-2.25597136e-02 9.92645442e-01 -2.74021059e-01 -1.00265443e+00
-7.81844318e-01 -6.75987065e-01 -1.48551476e+00 -9.76308733e-02
3.48810315e-01 -3.22908927e-05 -1.28839445e+00 1.48656082e+00
9.08122659e-02 -1.44552246e-01 2.60920346e-01 4.65802610e-01
1.01801682e+00 2.85233468e-01 6.18083239e-01 1.62122071e-01
1.45609105e+00 -3.96050990e-01 -5.88708222e-01 -4.05705497e-02
1.28141212e+00 -2.82209069e-01 3.87172312e-01 -1.38621375e-01
-6.42026186e-01 -2.58717895e-01 -1.04759169e+00 -8.25387239e-02
-4.17803347e-01 3.97857755e-01 9.79976416e-01 6.08998060e-01
-1.23645175e+00 3.90246391e-01 -1.21063924e+00 -8.31897974e-01
9.60179448e-01 4.48007137e-01 -8.26391816e-01 -2.93832093e-01
-9.74634349e-01 8.24108362e-01 3.73856395e-01 -1.01476945e-01
-1.19401205e+00 -8.82549882e-01 -5.64786971e-01 -2.26035360e-02
-2.23942429e-01 -5.74462414e-01 1.17813754e+00 -7.81352520e-01
-8.26458573e-01 1.40212810e+00 2.22799461e-02 -6.42026663e-01
4.98941481e-01 2.49341801e-01 -4.45610255e-01 3.58165830e-01
1.90729335e-01 8.44786584e-01 -1.57922775e-01 -9.77160037e-01
-8.61661911e-01 -7.07588017e-01 -2.79684126e-01 -1.94703251e-01
-4.59983230e-01 -1.33500308e-01 -4.09745008e-01 -2.45394066e-01
2.26341963e-01 -9.53842103e-01 -4.71324742e-01 2.61153847e-01
-4.11276847e-01 -2.71533102e-01 6.65367305e-01 -9.44538176e-01
7.89763451e-01 -2.20433211e+00 -5.32650590e-01 1.42471552e-01
4.70234096e-01 5.13504565e-01 1.69600472e-02 -3.89095880e-02
8.23313743e-02 5.67534149e-01 1.84843406e-01 2.21288726e-01
-5.45112789e-01 -2.21420024e-02 4.94121820e-01 7.06799030e-01
2.59611040e-01 1.06141782e+00 -8.48281443e-01 -6.70497835e-01
-3.92460376e-02 4.71434295e-01 -3.50744367e-01 -4.62563075e-02
2.65905231e-01 1.27646655e-01 -5.23405790e-01 1.06856751e+00
5.77978134e-01 -5.06738544e-01 4.43108082e-01 -1.57520831e-01
4.89781439e-01 -2.74085104e-01 -1.86723426e-01 1.11238766e+00
-1.61957324e-01 7.90109277e-01 -8.76567501e-04 -7.99209356e-01
9.74512756e-01 2.86990851e-01 5.40733278e-01 -9.88458514e-01
4.21598762e-01 3.18857968e-01 4.26614702e-01 -4.58736986e-01
-3.73475924e-02 -1.80403441e-01 5.86478263e-02 -1.23225920e-01
3.99198718e-02 4.66822326e-01 -7.27160648e-03 9.52867493e-02
1.48237956e+00 -7.16921449e-01 6.19196534e-01 -4.96758580e-01
3.44899386e-01 1.42282993e-01 7.59874463e-01 5.70572913e-01
-3.96676034e-01 4.21479285e-01 6.70785606e-01 -6.39465153e-01
-7.84068346e-01 -9.03929889e-01 -4.79248285e-01 4.79201078e-01
-6.49401397e-02 4.54712026e-02 -5.05331457e-01 -6.21388137e-01
1.84529215e-01 2.72420079e-01 -1.01004934e+00 -4.83183146e-01
-3.44570071e-01 -1.05413556e+00 1.02034116e+00 7.78002977e-01
9.29162562e-01 -7.35291302e-01 -3.54899734e-01 6.14178255e-02
8.06262493e-02 -8.47678721e-01 3.09397876e-01 1.99505702e-01
-1.03866005e+00 -1.51581573e+00 -1.00729227e+00 -1.05018580e+00
1.06449807e+00 -4.59386408e-02 8.42102349e-01 4.21695054e-01
-7.42539346e-01 -2.56797403e-01 -1.32937342e-01 -7.11159766e-01
-7.62091577e-01 1.42230913e-01 -2.94285953e-01 -1.41411141e-01
6.98180199e-01 -3.69994231e-02 -7.31308043e-01 1.00570016e-01
-7.56511629e-01 1.70478463e-01 1.13432419e+00 9.12067473e-01
5.74199557e-01 7.24274367e-02 7.10306287e-01 -1.07919359e+00
2.92015344e-01 -5.46233594e-01 -1.75431579e-01 1.14302188e-01
-5.95463812e-01 -6.21428370e-01 3.73742938e-01 -1.97203547e-01
-8.15821528e-01 2.02638526e-02 -2.18386695e-01 -7.77899623e-02
-1.53904125e-01 6.50372744e-01 6.62703514e-02 -1.39022663e-01
7.01123238e-01 -7.81511664e-02 4.63898838e-01 -2.19544590e-01
-5.74868798e-01 8.06102514e-01 6.84863806e-01 1.65964887e-02
2.70002961e-01 5.45217812e-01 2.62189895e-01 -7.95102954e-01
-6.69317365e-01 -4.32209194e-01 -1.60479233e-01 -2.39173040e-01
9.35695112e-01 -1.03883803e+00 -6.32844687e-01 8.10821533e-01
-6.36778533e-01 -4.05747026e-01 2.06790030e-01 5.96819758e-01
6.12524748e-02 -7.59698078e-02 -1.18880463e+00 -3.10256153e-01
-6.14608943e-01 -8.58297706e-01 7.96058059e-01 5.58400989e-01
-2.21516579e-01 -1.14894831e+00 -2.07897753e-01 7.04674870e-02
6.58904374e-01 6.61938012e-01 1.30940616e+00 -8.55721414e-01
6.10723756e-02 -8.30893099e-01 -5.83520949e-01 1.00309394e-01
4.53148574e-01 1.20742381e-01 -9.20869827e-01 -4.99404103e-01
-5.35687327e-01 -3.76162350e-01 1.11266518e+00 6.11957073e-01
1.45135403e+00 -7.52819255e-02 -1.24011445e+00 7.40243554e-01
1.57989490e+00 6.75794125e-01 9.34883714e-01 4.54533249e-01
3.31180185e-01 3.20425719e-01 1.70693472e-01 -1.39052242e-01
1.95095465e-01 -5.78632504e-02 4.69703704e-01 -5.76202571e-01
-2.94319242e-01 -1.68347016e-01 -3.02090496e-01 1.99749291e-01
-1.36904880e-01 -4.00896549e-01 -1.42898035e+00 7.42855310e-01
-1.36603951e+00 -6.64599478e-01 -7.33211264e-02 1.70119238e+00
8.21459293e-01 1.55030563e-01 -3.62025350e-01 2.32089743e-01
8.84223998e-01 -4.50506896e-01 -6.04869246e-01 -1.21106640e-01
2.11282354e-02 1.68684453e-01 7.21384466e-01 -6.84514567e-02
-1.17506444e+00 6.49132192e-01 6.70643187e+00 7.22000062e-01
-1.44071925e+00 -1.78478912e-01 1.48250890e+00 4.46771890e-01
1.43534541e-01 -4.07679349e-01 -7.26142704e-01 3.47513199e-01
1.00853622e+00 -9.70853791e-02 -4.75554973e-01 7.17452466e-01
-3.33015546e-02 -2.77815580e-01 -1.00114369e+00 5.83832920e-01
-2.92636007e-01 -1.38651192e+00 -2.75922450e-03 3.76764834e-01
6.54527605e-01 -4.76893187e-02 5.26929740e-04 4.79890347e-01
4.22934175e-01 -1.45049679e+00 -2.69231439e-01 4.83947068e-01
1.21967506e+00 -6.67744994e-01 1.51161408e+00 2.90820926e-01
-5.02949536e-01 2.53167152e-02 -4.77911890e-01 2.30103135e-01
-6.19109154e-01 5.85667551e-01 -1.26042068e+00 2.91775763e-01
9.17953074e-01 6.46074235e-01 -7.90875375e-01 1.07797611e+00
2.26024762e-01 8.24283600e-01 -1.65269807e-01 -2.61755675e-01
1.35663211e-01 4.80300337e-01 -2.10705101e-01 1.10194337e+00
4.64519322e-01 1.11677140e-01 -8.57635438e-02 5.15471876e-01
-2.85573781e-01 1.14392554e-02 -3.12170535e-01 -1.43197803e-02
5.06529391e-01 1.49341297e+00 -1.01150334e+00 -2.23587155e-01
-2.51835078e-01 3.87482882e-01 1.18370458e-01 1.09485261e-01
-6.09598160e-01 -5.63476384e-01 1.90397516e-01 2.15582222e-01
-1.57768592e-01 5.79345703e-01 -3.41348678e-01 -4.53558326e-01
-4.98584241e-01 -6.66056097e-01 5.73909760e-01 -4.43730414e-01
-1.25684631e+00 3.00216764e-01 -3.14819723e-01 -9.63751912e-01
9.87725891e-03 -7.86032140e-01 -7.07920194e-01 7.08878100e-01
-1.51586628e+00 -1.22373903e+00 -6.26408756e-01 2.23940939e-01
2.08645072e-02 -3.07503581e-01 1.04816771e+00 2.69945979e-01
-5.79859614e-01 8.53697956e-01 2.03555226e-01 7.47963011e-01
5.68165064e-01 -1.18068826e+00 -1.24332801e-01 2.24920243e-01
-9.57479239e-01 3.76496166e-01 1.53746486e-01 -5.61873853e-01
-1.05229390e+00 -1.62057018e+00 9.04173553e-01 2.48163298e-01
4.54699785e-01 1.34223951e-02 -7.35619366e-01 7.14107633e-01
-5.08946441e-02 2.54470915e-01 9.83931601e-01 -1.18592374e-01
-1.06425442e-01 -3.16318065e-01 -1.64091969e+00 4.87965912e-01
4.15797621e-01 -1.95724562e-01 1.07629038e-01 6.99180663e-02
2.76265532e-01 -4.61412609e-01 -1.22686040e+00 8.89820337e-01
7.71078169e-01 -7.61467397e-01 7.83440948e-01 -5.52599072e-01
7.14900136e-01 1.16282985e-01 -7.21551701e-02 -1.16398716e+00
-6.54952705e-01 2.51775265e-01 3.35904568e-01 1.04798162e+00
5.67824006e-01 -6.85284555e-01 1.07152259e+00 4.39797044e-01
-1.45389140e-01 -9.83512759e-01 -8.00173581e-01 -5.72907925e-01
4.79369611e-01 1.94892019e-01 3.99988472e-01 9.09922779e-01
-1.44936979e-01 -7.24180341e-02 3.29957247e-01 -6.38398109e-03
3.86355162e-01 -4.13792163e-01 3.92230242e-01 -1.32631290e+00
2.69289315e-01 -4.88721550e-01 -8.96159768e-01 -1.33293986e-01
-1.28172830e-01 -1.06080747e+00 -3.48248303e-01 -1.75166798e+00
8.39079559e-01 -5.25903046e-01 -6.50647521e-01 9.07465160e-01
-1.42847374e-01 5.90448141e-01 -3.33106160e-01 1.61150172e-01
-2.82848507e-01 -8.38713571e-02 1.29307592e+00 -5.45828640e-01
1.61480024e-01 -1.08753718e-01 -9.29844022e-01 8.02241206e-01
1.13727355e+00 -2.83346951e-01 2.46844273e-02 -3.03690374e-01
-1.29034474e-01 1.66266903e-01 3.88432801e-01 -1.13469863e+00
2.02756345e-01 -3.16163003e-02 1.20125973e+00 -6.13653243e-01
4.62442301e-02 -4.69128281e-01 2.96068609e-01 1.20983803e+00
-4.55040663e-01 -4.74845558e-01 4.58080918e-01 3.43507469e-01
-2.68739283e-01 4.19114679e-02 9.20448244e-01 -1.54814065e-01
-6.56414628e-01 3.57283235e-01 -6.55572712e-01 -5.26783347e-01
1.33567846e+00 -1.43546164e-01 -6.30785286e-01 -3.27127650e-02
-5.05211413e-01 4.01791304e-01 2.88756549e-01 1.36173904e-01
5.23424745e-01 -1.32192659e+00 -9.07119691e-01 2.25917753e-02
4.16680932e-01 4.22139950e-02 4.10637170e-01 8.35472465e-01
-9.35090125e-01 5.97408414e-01 -4.18023467e-01 -7.26156414e-01
-1.38825250e+00 9.79767069e-02 7.00926244e-01 -6.32578611e-01
-4.37886208e-01 8.72754931e-01 2.61496782e-01 -3.54361236e-01
2.56741375e-01 -2.22487494e-01 -4.65223014e-01 -3.55138659e-01
5.35080016e-01 2.99515575e-01 2.46129453e-01 -2.98666984e-01
-4.37256813e-01 2.23199993e-01 -5.68407178e-01 6.55332625e-01
1.14038372e+00 5.21678329e-01 -1.68454111e-01 8.61172099e-03
1.37491000e+00 -4.86190557e-01 -5.42823970e-01 -1.12409126e-02
5.80337122e-02 4.94116805e-02 2.43597552e-01 -1.05781353e+00
-1.19995022e+00 7.11661994e-01 9.99528468e-01 6.20334223e-02
1.19648921e+00 1.16272420e-01 6.59326494e-01 2.65188903e-01
2.81915426e-01 -6.11335218e-01 -1.68757841e-01 9.61551741e-02
3.79333109e-01 -1.29212642e+00 -4.96585965e-02 -3.02317828e-01
-1.76066577e-01 1.19575846e+00 7.69597054e-01 -2.35091820e-01
6.91312790e-01 2.84089386e-01 2.19657496e-01 -3.04756641e-01
-8.74315858e-01 1.36558801e-01 -5.57091683e-02 6.13648713e-01
6.14627242e-01 4.24303532e-01 -5.20738065e-01 9.49160874e-01
-1.79729089e-01 2.45060101e-01 3.08726996e-01 9.30718958e-01
-6.17819846e-01 -5.88913858e-01 -6.19369280e-03 1.01759398e+00
-1.10646045e+00 3.14469218e-01 -6.63753450e-01 1.20843899e+00
-1.15732230e-01 6.88308775e-01 3.72925371e-01 -4.51299548e-01
-3.03818583e-02 -2.12839872e-01 2.23486260e-01 -2.98213094e-01
-3.78906250e-01 -5.59601746e-02 -2.99043115e-03 -1.68736756e-01
-2.03474447e-01 -2.53378987e-01 -1.38880575e+00 -3.17871124e-01
-2.81998783e-01 6.87043741e-02 7.64617622e-01 6.40451014e-01
3.23069334e-01 7.70464957e-01 4.10540342e-01 -1.51806876e-01
-3.39725465e-01 -1.27182925e+00 -7.82944679e-01 1.27244622e-01
3.43283147e-01 -2.79908389e-01 -1.29286036e-01 7.59995682e-03] | [15.179503440856934, -2.852385997772217] |
fdb9e8b7-2ca7-42fe-bea1-8be87bf3d1d1 | probing-for-predicate-argument-structures-in | null | null | https://aclanthology.org/2022.acl-long.316 | https://aclanthology.org/2022.acl-long.316.pdf | Probing for Predicate Argument Structures in Pretrained Language Models | Thanks to the effectiveness and wide availability of modern pretrained language models (PLMs), recently proposed approaches have achieved remarkable results in dependency- and span-based, multilingual and cross-lingual Semantic Role Labeling (SRL). These results have prompted researchers to investigate the inner workings of modern PLMs with the aim of understanding how, where, and to what extent they encode information about SRL. In this paper, we follow this line of research and probe for predicate argument structures in PLMs. Our study shows that PLMs do encode semantic structures directly into the contextualized representation of a predicate, and also provides insights into the correlation between predicate senses and their structures, the degree of transferability between nominal and verbal structures, and how such structures are encoded across languages. Finally, we look at the practical implications of such insights and demonstrate the benefits of embedding predicate argument structure information into an SRL model. | ['Roberto Navigli', 'Simone Conia'] | null | null | null | null | acl-2022-5 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 3.81580770e-01 4.09804881e-01 -6.58199549e-01 -4.12481815e-01
-4.50753421e-01 -9.78732228e-01 7.55303741e-01 7.64902353e-01
-4.38603967e-01 5.50384521e-01 1.02517438e+00 -3.94298017e-01
-3.09543937e-01 -7.01577544e-01 -6.03142500e-01 -2.71456897e-01
-1.80450585e-02 3.91001374e-01 2.80339837e-01 -5.48798859e-01
2.17877358e-01 2.71512479e-01 -1.32757926e+00 5.30076742e-01
7.63988853e-01 4.92487431e-01 3.02508324e-01 -1.41949281e-01
-3.20925981e-01 9.76524115e-01 -3.97682399e-01 -4.21191990e-01
-1.05283163e-01 -3.72682005e-01 -1.03927040e+00 -2.21768782e-01
2.86842108e-01 2.29419485e-01 -1.73542961e-01 7.49583244e-01
1.20489888e-01 1.28915772e-01 3.71466905e-01 -1.02803969e+00
-9.70656633e-01 1.18109620e+00 -1.52614906e-01 4.34927016e-01
5.91571629e-01 -1.93717316e-01 1.71990144e+00 -5.30507624e-01
1.06336010e+00 1.59177959e+00 4.86495793e-01 4.35046822e-01
-1.19800603e+00 -2.46163443e-01 4.34140354e-01 3.78909707e-01
-9.34663534e-01 -4.98281419e-01 7.79765546e-01 -2.79019296e-01
9.41611886e-01 -2.92766243e-02 5.81289053e-01 1.05766261e+00
-2.89453357e-01 8.40836883e-01 1.36587703e+00 -8.26560736e-01
-1.56878784e-01 9.08947065e-02 5.08702457e-01 5.80035210e-01
3.66531700e-01 6.17580861e-02 -9.18867230e-01 -3.33102793e-02
7.38014162e-01 -5.12336016e-01 -2.58153856e-01 -1.80625871e-01
-1.27595699e+00 9.18937922e-01 4.08126920e-01 8.91627967e-01
-5.97935133e-02 9.18374024e-03 6.84615016e-01 2.98940837e-01
2.27412105e-01 5.53836942e-01 -7.74787843e-01 1.75314471e-02
-2.84585387e-01 1.41351387e-01 5.42939782e-01 6.35099649e-01
5.32112479e-01 -1.14086360e-01 -8.89075696e-02 1.09434450e+00
3.29608232e-01 1.12820387e-01 6.25482321e-01 -1.16583061e+00
4.94940877e-01 6.71841919e-01 1.38836005e-03 -8.41482997e-01
-3.98855716e-01 -2.03920305e-01 1.80535927e-01 -6.06176019e-01
4.86224115e-01 2.74704218e-01 -2.41290331e-01 2.30271363e+00
2.53649443e-01 -3.09612155e-01 3.06095302e-01 5.86853623e-01
6.19621098e-01 4.48711425e-01 7.63972461e-01 -1.39122546e-01
1.55599511e+00 -6.02590680e-01 -4.99124765e-01 -5.73054671e-01
1.05263925e+00 -4.23719913e-01 1.70713425e+00 -3.50197017e-01
-9.56631303e-01 -3.77600640e-01 -8.83615553e-01 -5.15765429e-01
-4.33886379e-01 -7.60636106e-02 9.87415433e-01 5.26668906e-01
-9.11250532e-01 5.35697460e-01 -6.58789098e-01 -4.45178688e-01
2.78645635e-01 -8.62759799e-02 -4.59857076e-01 -2.86446035e-01
-1.62559652e+00 1.21010911e+00 7.57006109e-01 -8.06143731e-02
-4.49116439e-01 -7.69629419e-01 -1.20447981e+00 -5.03613465e-02
4.26896989e-01 -3.18691730e-01 1.06217873e+00 -1.03661048e+00
-9.24523652e-01 1.38817012e+00 -4.40885276e-01 -4.31741029e-01
-2.46633589e-01 -2.52132416e-01 -4.59278733e-01 1.70026064e-01
4.85962093e-01 5.92424512e-01 2.08172977e-01 -1.14174902e+00
-6.85917020e-01 -7.65602052e-01 5.06946862e-01 5.28073192e-01
-1.74571544e-01 5.01253903e-01 1.95630908e-01 -5.85769892e-01
2.06780180e-01 -7.61160433e-01 1.84002861e-01 -2.49310017e-01
-2.14720704e-02 -8.20126414e-01 3.97646368e-01 -7.19370663e-01
1.11834085e+00 -2.25694776e+00 2.24544723e-02 -3.90937701e-02
-2.40087241e-01 -1.01363296e-02 -1.78132236e-01 8.38392019e-01
-1.97150677e-01 2.89284796e-01 -3.05849642e-01 9.28952843e-02
3.87308538e-01 8.40106785e-01 -4.17557418e-01 4.39353049e-01
6.99148402e-02 1.24359834e+00 -9.00893986e-01 -2.72874266e-01
7.50493407e-02 3.21100205e-01 -4.53119099e-01 -1.53558254e-01
-4.11537737e-01 4.37651426e-01 -2.89606273e-01 5.80264926e-01
5.91913126e-02 -5.50915971e-02 9.07518029e-01 -2.18832731e-01
-2.61261612e-01 1.31717086e+00 -7.64865339e-01 1.54895115e+00
-4.50684816e-01 5.53741693e-01 8.14002082e-02 -1.14511514e+00
4.05819148e-01 3.75981212e-01 3.42765972e-02 -9.55923021e-01
-7.88335279e-02 4.27144736e-01 5.54455161e-01 -9.00376439e-02
3.51495177e-01 -4.52800333e-01 -3.11484396e-01 4.71076310e-01
-7.63961747e-02 1.69494048e-01 3.63488317e-01 1.13043733e-01
6.80883944e-01 3.08615357e-01 6.53123021e-01 -5.29105663e-01
7.07850277e-01 1.26242056e-01 6.68313086e-01 3.28661382e-01
-1.46211311e-01 -7.01349452e-02 6.24462187e-01 -1.18693382e-01
-8.03329408e-01 -1.08005190e+00 -5.83079815e-01 1.47484040e+00
1.02777421e-01 -5.08619010e-01 -4.66550857e-01 -5.53195894e-01
3.80443521e-02 1.04057443e+00 -5.92732906e-01 -1.53879017e-01
-1.20059538e+00 -5.01408219e-01 7.44012952e-01 8.29956710e-01
2.85073429e-01 -1.41785586e+00 -6.66994631e-01 2.41753235e-01
-1.62532151e-01 -1.42795682e+00 4.19874340e-02 1.20080644e-02
-8.67632627e-01 -1.27465451e+00 2.18787059e-01 -8.79407167e-01
2.89406270e-01 -2.77612098e-02 1.29844844e+00 1.78904578e-01
2.51666188e-01 3.57756168e-01 -4.30060446e-01 -2.97387332e-01
-4.25546914e-01 2.43757311e-02 -1.85383931e-01 -3.43892246e-01
3.25928986e-01 -5.85840642e-01 -9.13016424e-02 -3.68306413e-02
-8.86225402e-01 -2.82880098e-01 1.20366640e-01 4.89294976e-01
4.25505608e-01 -1.68167815e-01 4.98398662e-01 -1.30435312e+00
6.84879959e-01 -4.91862416e-01 -3.90346169e-01 3.96308869e-01
-4.18191642e-01 3.35836232e-01 3.46304089e-01 9.40794796e-02
-1.34613991e+00 -6.37213111e-01 -2.36124083e-01 5.76607466e-01
-2.73430888e-02 7.56379187e-01 -4.67580616e-01 3.12558115e-01
5.19717693e-01 7.79657140e-02 -7.16005936e-02 -4.84477997e-01
6.37952089e-01 1.74436554e-01 2.70314872e-01 -1.28987265e+00
6.29468799e-01 5.56975722e-01 -1.56178817e-01 -8.21094036e-01
-1.46522832e+00 -2.57322460e-01 -6.01906419e-01 4.56760734e-01
8.23052049e-01 -1.02311885e+00 -3.55055243e-01 1.17685862e-01
-1.11286461e+00 -5.04196107e-01 -4.32852626e-01 2.76076108e-01
-4.50357378e-01 5.08751750e-01 -9.96564567e-01 -2.84957647e-01
9.81631037e-03 -7.01742709e-01 7.55195379e-01 -1.57399580e-01
-6.26304567e-01 -1.48656535e+00 -1.46491095e-01 3.50857228e-01
3.02782029e-01 9.30999685e-03 1.74634802e+00 -8.14952254e-01
-3.61837626e-01 3.51814240e-01 -2.84905553e-01 1.39805004e-01
1.92216203e-01 -5.69911599e-01 -6.64867580e-01 6.22934848e-02
1.29995912e-01 -2.94080019e-01 5.34335673e-01 2.79255807e-01
6.76015973e-01 -3.21265727e-01 -2.19507858e-01 4.95504647e-01
1.40747237e+00 2.03254614e-02 4.16632831e-01 6.06152713e-01
5.77884138e-01 9.57425535e-01 8.06369483e-01 -1.26334146e-01
7.94440627e-01 4.88299817e-01 6.38452321e-02 4.15728688e-01
-4.10292685e-01 -6.92118347e-01 4.26780701e-01 7.73037672e-01
1.75217748e-01 -3.01308818e-02 -8.45737278e-01 7.94939458e-01
-1.69536567e+00 -6.40258729e-01 -1.51646361e-01 1.86567652e+00
1.17300630e+00 -5.22060655e-02 -1.29052445e-01 1.39163882e-01
5.80924928e-01 6.98632538e-01 -1.60857514e-01 -4.21566129e-01
-4.83120143e-01 4.26791698e-01 2.41096914e-01 5.21682680e-01
-7.78042614e-01 1.62876320e+00 6.70729923e+00 5.82930267e-01
-8.00296247e-01 2.87332565e-01 1.01923481e-01 5.40386260e-01
-7.53934443e-01 2.88283467e-01 -8.07680607e-01 1.62803978e-01
9.38950479e-01 -7.82091767e-02 4.27720755e-01 5.83390415e-01
6.03066199e-03 -9.53337178e-02 -1.28130984e+00 4.80394989e-01
3.57828103e-02 -1.27960026e+00 1.24424659e-01 -1.89540744e-01
3.52971077e-01 5.04348911e-02 -2.28418618e-01 2.33057693e-01
3.89071852e-01 -9.57678378e-01 9.75023031e-01 -3.22234750e-01
5.77564776e-01 -4.39262211e-01 3.93058598e-01 2.21628144e-01
-1.16102946e+00 -1.70577928e-01 -3.48894030e-01 -4.49036628e-01
2.27483958e-01 2.25797333e-02 -5.31979382e-01 3.00509244e-01
1.82279751e-01 8.44003797e-01 -4.89415169e-01 2.51224875e-01
-1.23711014e+00 7.90056229e-01 -2.42373385e-02 2.44815156e-01
2.06660405e-01 -6.28488287e-02 4.28114861e-01 9.90536034e-01
-1.93695709e-01 1.46447629e-01 2.46010035e-01 7.33144343e-01
1.26725286e-01 3.54662538e-01 -4.41752285e-01 -4.53518927e-01
7.72297502e-01 8.22985590e-01 -8.94809186e-01 -2.00827390e-01
-3.92344415e-01 7.10873246e-01 6.53606236e-01 1.97884113e-01
-5.44175148e-01 3.20350975e-01 8.04862320e-01 3.30052257e-01
4.60677706e-02 -6.07838452e-01 -4.05588210e-01 -9.67172086e-01
4.07182872e-02 -7.10988700e-01 7.09703684e-01 -6.29824996e-01
-1.33938956e+00 2.45664418e-01 3.01217645e-01 -3.07618350e-01
-2.07342207e-01 -8.10747564e-01 -1.22915596e-01 6.90055847e-01
-1.81981933e+00 -1.40845895e+00 4.43111867e-01 5.88361979e-01
5.11905074e-01 1.51443318e-01 1.02908826e+00 3.66055109e-02
-1.84474990e-01 3.94055694e-01 -4.06316519e-01 4.04149622e-01
5.02761245e-01 -1.04578567e+00 4.26117361e-01 7.51022160e-01
4.31517482e-01 1.03524947e+00 4.43976820e-01 -6.01221144e-01
-1.07434773e+00 -6.59357607e-01 1.33058846e+00 -8.11763942e-01
1.05526531e+00 -2.21292317e-01 -1.10082388e+00 1.09741521e+00
2.84321457e-01 6.42408878e-02 9.53253984e-01 6.44694030e-01
-7.85097778e-01 2.14331374e-01 -1.00647914e+00 5.19903958e-01
1.62710917e+00 -1.00965536e+00 -1.36349237e+00 4.43177000e-02
1.23851287e+00 -2.15053499e-01 -6.37836933e-01 3.73056918e-01
2.80258268e-01 -7.82852590e-01 9.15856779e-01 -7.96342432e-01
4.40841138e-01 -2.73424000e-01 -4.70614702e-01 -1.32620347e+00
-3.12847435e-01 2.73800716e-02 1.94650874e-01 1.16296113e+00
4.98777956e-01 -8.08503926e-01 2.35853702e-01 5.94639301e-01
-4.77075845e-01 -3.51952493e-01 -8.51009429e-01 -5.92624843e-01
2.15883479e-01 -6.30351424e-01 4.99364287e-01 1.21144664e+00
6.77213371e-02 7.02226937e-01 4.68976885e-01 -2.08060760e-02
3.60325158e-01 3.87920469e-01 -8.74748528e-02 -1.38120437e+00
-3.29196781e-01 -1.74007744e-01 -2.20646024e-01 -5.98271012e-01
1.01507258e+00 -1.35061502e+00 -3.62385899e-01 -1.70500231e+00
-8.20618272e-02 -8.56884301e-01 -2.71677077e-01 7.52330005e-01
-1.89301252e-01 5.25357649e-02 4.37094420e-01 2.02246711e-01
-3.98966104e-01 2.27303535e-01 9.36312795e-01 4.25240725e-01
-9.62120071e-02 -5.16364992e-01 -1.13998556e+00 9.33728516e-01
7.19216287e-01 -3.38974237e-01 -5.88214159e-01 -7.64137983e-01
5.56322277e-01 -1.83643505e-01 2.87234962e-01 -3.61161113e-01
-2.50093728e-01 -3.10328573e-01 1.10337496e-01 2.59257387e-02
1.47227898e-01 -6.65431499e-01 -3.50597382e-01 2.41120741e-01
-5.26708782e-01 1.26859158e-01 2.11890921e-01 2.73381621e-01
-4.37018931e-01 -3.21612298e-01 4.94513482e-01 -5.20326376e-01
-1.18712986e+00 -1.06385693e-01 -1.78776771e-01 7.38669753e-01
5.79679251e-01 -2.00744584e-01 -3.48280311e-01 8.29005167e-02
-6.97314858e-01 4.95337918e-02 4.94789213e-01 6.95773542e-01
1.70835942e-01 -1.19583535e+00 -4.38352346e-01 2.59027015e-02
2.49427155e-01 -3.79797459e-01 -1.05671465e-01 5.50019801e-01
-1.65711209e-01 6.00100338e-01 -3.18382829e-02 -1.44086823e-01
-9.90697742e-01 4.52695668e-01 1.91386923e-01 -2.58898139e-01
-3.94703597e-01 8.66222680e-01 3.70668799e-01 -6.11609042e-01
-2.13970795e-01 -2.05652192e-01 -4.14087236e-01 3.07975799e-01
1.16093233e-01 3.14053260e-02 -1.69581503e-01 -1.05079269e+00
-5.49277067e-01 4.28067029e-01 1.87066391e-01 -3.16699535e-01
1.33524835e+00 -1.56117886e-01 -6.65426672e-01 6.89845085e-01
9.50270236e-01 5.25623202e-01 -8.53719413e-01 -4.19307530e-01
7.51153588e-01 -2.15403691e-01 -4.18905854e-01 -8.43160987e-01
-5.05461872e-01 6.35036349e-01 -1.72741592e-01 -5.58895804e-02
7.07880497e-01 6.76747561e-01 6.47825003e-01 -7.10554123e-02
4.46941584e-01 -1.21988034e+00 -2.08301708e-01 9.49887335e-01
7.03358948e-01 -7.64452875e-01 -2.80944735e-01 -8.56419086e-01
-6.41286850e-01 6.57312155e-01 4.24315780e-01 -9.20137689e-02
3.91272217e-01 1.21396281e-01 -5.77408485e-02 -4.16737765e-01
-6.65003061e-01 -4.60337311e-01 9.64698568e-02 6.28633857e-01
9.81499970e-01 4.23509866e-01 -1.00968206e+00 5.22783816e-01
-6.52147412e-01 -6.20754361e-01 2.52091080e-01 9.01984453e-01
-3.18760991e-01 -1.58524215e+00 -1.59571514e-01 -9.33449864e-02
-7.33260930e-01 -6.02334619e-01 -4.22945827e-01 1.00564516e+00
2.52859533e-01 7.94909060e-01 9.05506611e-02 3.91898572e-01
1.81353450e-01 5.81683517e-01 8.44862103e-01 -1.07099986e+00
-4.38695848e-01 -2.91430682e-01 6.24645650e-01 -5.22754312e-01
-6.14855528e-01 -8.02545428e-01 -1.87637377e+00 2.71195292e-01
2.26609372e-02 3.14654797e-01 3.91735613e-01 1.28885555e+00
1.26006693e-01 2.42702633e-01 4.22255620e-02 -1.73288494e-01
-4.34999287e-01 -7.28877842e-01 -5.36323667e-01 4.80293781e-01
-2.66821776e-02 -6.50719166e-01 -2.69838065e-01 -1.84988901e-01] | [10.36921501159668, 9.3471040725708] |
5cd1c53e-c10e-415f-bafa-7f05fdfbbf0c | handcrafted-localized-phase-features-for | null | null | https://doi.org/10.1016/j.imavis.2022.104465 | https://www.sciencedirect.com/science/article/pii/S0262885622000944/pdfft?md5=513fcc71e4d80888039a419003552cb4&pid=1-s2.0-S0262885622000944-main.pdf | Handcrafted localized phase features for human action recognition | Human action recognition is one of the most important topics in computer vision. Monitoring elderly people and children, smart surveillance systems and human-computer interaction are a few examples of its applications. The aim of this study is to recognize human activities by utilizing the phase information extracted from the frequency domain of the video data as handcrafted features. Rather than estimating optical flow or computing motion vectors, we aim to utilize the localized phase information as descriptors of the motion dynamics of the scene. Phase correlation information extracted from each two co-sited blocks from each two consecutive frames of video clips were used to train a model using KNN classifier to model the action. To evaluate the performance of our method, an extensive work has been done on three large and complex datasets: UCF101, Kinetics-400 and Kinetics-700. The results show that our approach succeeds on recognizing human actions across all these datasets with high accuracy. | ['Charith Abhayaratne', 'Seyed Mostafa Hejazi'] | 2022-05-05 | null | null | null | image-and-vision-computing-2022-5 | ['action-classification'] | ['computer-vision'] | [ 2.98860580e-01 -4.40672249e-01 -3.60057265e-01 -1.50999382e-01
-1.05047442e-01 -3.15514266e-01 7.42696166e-01 -1.65382653e-01
-7.31079340e-01 7.56111085e-01 4.85782385e-01 4.19835597e-01
-2.69861948e-02 -4.80880529e-01 -2.20246837e-01 -9.46065962e-01
-3.48186910e-01 3.17866430e-02 6.27379477e-01 1.38368130e-01
4.77881223e-01 4.92937535e-01 -1.73810077e+00 3.92824590e-01
3.90799433e-01 7.78401077e-01 -9.36933383e-02 1.05007327e+00
3.03467721e-01 1.21608317e+00 -4.37228590e-01 1.69134840e-01
1.44050568e-01 -4.54799443e-01 -7.63452470e-01 2.38529578e-01
4.09604192e-01 -2.64846355e-01 -7.87612081e-01 7.60098338e-01
4.25613165e-01 3.04950744e-01 7.17230737e-01 -1.28692985e+00
8.70041847e-02 -9.57580134e-02 -5.44143081e-01 9.90132809e-01
9.77884531e-01 1.68588683e-01 7.13199377e-01 -5.02272666e-01
7.45420516e-01 1.04031527e+00 5.34551442e-01 5.31385124e-01
-8.35309088e-01 -3.10637176e-01 -1.41090184e-01 9.26194906e-01
-1.23072064e+00 -4.46028262e-01 8.52228820e-01 -6.84101224e-01
1.13763261e+00 8.73862393e-03 1.15871119e+00 1.09769857e+00
3.06767255e-01 1.04589617e+00 9.00228322e-01 -5.24729192e-01
2.34304518e-01 -3.77101809e-01 4.07286227e-01 9.54477370e-01
2.20789999e-01 1.38511389e-01 -8.16424191e-01 1.27556976e-02
6.72798693e-01 2.81419400e-02 -5.62471569e-01 -5.18308282e-01
-1.54285336e+00 5.39747596e-01 3.31414640e-02 5.66848099e-01
-7.16475129e-01 1.86977655e-01 4.86162007e-01 -8.07784498e-04
2.39087731e-01 -3.51961739e-02 -2.56837398e-01 -9.57913041e-01
-7.20131278e-01 2.50654161e-01 8.46065342e-01 4.79067594e-01
5.16847551e-01 -1.16942942e-01 -1.60195097e-01 3.56921613e-01
3.36145967e-01 5.94119549e-01 7.49039829e-01 -1.12661278e+00
5.09803832e-01 6.24462843e-01 3.06841403e-01 -1.16237497e+00
-6.46403551e-01 4.36667860e-01 -6.90131366e-01 5.30255362e-02
7.42073417e-01 -1.50249839e-01 -8.76752794e-01 1.40090227e+00
3.97629529e-01 6.59638762e-01 5.76950721e-02 7.05955565e-01
6.01368785e-01 5.88818550e-01 9.97228101e-02 -5.62790751e-01
1.24838889e+00 -8.96507502e-01 -8.42185736e-01 3.00573241e-02
6.87102795e-01 -4.69623566e-01 3.06353450e-01 5.85620403e-01
-6.26595020e-01 -5.73586047e-01 -1.04715276e+00 4.86304164e-01
-2.24214762e-01 1.57654643e-01 8.39164734e-01 5.22719562e-01
-6.78881943e-01 6.32023156e-01 -1.46081078e+00 -6.37947857e-01
5.23054004e-01 4.06687438e-01 -8.10186327e-01 -1.07962921e-01
-8.47413659e-01 6.54978037e-01 4.29198831e-01 -8.13740864e-02
-5.95761359e-01 -2.21987933e-01 -8.33750010e-01 -3.55356216e-01
-1.49550974e-01 -2.71834612e-01 1.02112412e+00 -1.19370461e+00
-1.40439939e+00 8.06259215e-01 -2.79087722e-01 -4.66621071e-01
3.81934553e-01 -3.67141396e-01 -4.13934737e-01 7.63897955e-01
1.10574290e-02 4.04103726e-01 8.39343846e-01 -3.48531425e-01
-9.66616929e-01 -5.17446101e-01 -2.98884392e-01 9.91437137e-02
-2.94795752e-01 1.69539571e-01 -2.04611257e-01 -6.07920110e-01
1.45471662e-01 -1.07371986e+00 1.45672783e-01 -2.78097957e-01
9.56600457e-02 -3.68632436e-01 1.02366638e+00 -6.90739691e-01
1.16620731e+00 -2.05191231e+00 1.50347903e-01 -8.65494013e-02
-5.44259846e-02 5.55986047e-01 7.84875900e-02 3.15736026e-01
-6.51590601e-02 -5.94635069e-01 4.99427458e-03 1.81023553e-01
-4.76975322e-01 2.80178249e-01 1.57138914e-01 8.07094991e-01
8.25377107e-02 7.30471969e-01 -1.09782863e+00 -6.39440775e-01
4.81942713e-01 6.37766063e-01 -4.31219310e-01 1.78170949e-01
2.08672032e-01 8.05073619e-01 -7.55201936e-01 3.73534411e-01
1.76945686e-01 -3.06023490e-02 -4.34186421e-02 -3.30874652e-01
-9.84788779e-03 -6.57405928e-02 -1.27447367e+00 1.62091160e+00
1.81193829e-01 1.06696117e+00 -5.55667222e-01 -1.49129391e+00
4.31471497e-01 7.17744827e-01 1.31247866e+00 -8.30588698e-01
1.23183064e-01 -1.76427647e-01 6.64429041e-03 -1.23955274e+00
2.27248222e-02 3.31617564e-01 2.55497754e-01 3.74297440e-01
2.55644917e-01 5.37981808e-01 5.33648551e-01 -1.67560324e-01
1.65362775e+00 3.37717116e-01 7.13148534e-01 6.85242005e-03
9.78972793e-01 8.61904621e-02 6.59324288e-01 5.38638890e-01
-8.70714128e-01 5.68018973e-01 1.93098187e-01 -1.01537907e+00
-7.95514703e-01 -8.66056204e-01 1.66587874e-01 6.36220634e-01
-1.21778943e-01 -4.09746587e-01 -8.39937150e-01 -6.69223189e-01
-2.71648705e-01 6.86916038e-02 -7.31028199e-01 -3.98968719e-02
-1.06861234e+00 -7.07164109e-01 4.03901726e-01 7.32118964e-01
9.48822796e-01 -1.36245906e+00 -1.31974185e+00 3.00168991e-01
-4.27620351e-01 -1.52733910e+00 -3.51303279e-01 -3.29210907e-01
-8.43025684e-01 -1.63936734e+00 -7.30165422e-01 -6.67140484e-01
4.85105962e-01 4.17507827e-01 7.06363857e-01 -1.99895576e-01
-8.16844106e-01 9.07173514e-01 -5.94488263e-01 -1.53999269e-01
8.07031691e-02 -4.02086675e-01 2.47467741e-01 3.73694539e-01
7.73902237e-01 -6.60805941e-01 -7.29874790e-01 2.46957868e-01
-5.73429346e-01 -2.20627263e-01 2.88914979e-01 7.02916682e-01
2.19554305e-01 6.83179200e-02 -6.22606166e-02 -3.54023635e-01
1.50674775e-01 -1.33283496e-01 -3.96203488e-01 1.72555566e-01
1.68105550e-02 1.21473558e-01 2.91590184e-01 -8.35472941e-01
-9.11333323e-01 5.32635212e-01 2.72631377e-01 -1.94283217e-01
-4.24464345e-01 1.03633620e-01 -2.01884061e-02 8.24389234e-02
6.09482467e-01 3.51825714e-01 -9.22499970e-02 -7.01292604e-02
-8.67174380e-03 5.50408900e-01 7.62782395e-01 -2.11304352e-01
3.80423933e-01 7.91991591e-01 2.13748142e-01 -1.32557940e+00
-6.73911810e-01 -9.70060706e-01 -1.07623553e+00 -5.24360299e-01
1.45811915e+00 -8.01864207e-01 -1.01758778e+00 1.01452792e+00
-1.09971976e+00 -1.90212969e-02 2.38330085e-02 1.07822788e+00
-8.71049464e-01 5.31364262e-01 -4.19422239e-01 -8.53868961e-01
-2.41773352e-01 -9.60005224e-01 9.31534648e-01 5.35588145e-01
-2.67145604e-01 -9.05528963e-01 5.34741998e-01 6.00242674e-01
5.28771989e-03 5.68088055e-01 2.47030035e-01 -6.92491591e-01
-5.64692140e-01 -5.14100552e-01 1.59450680e-01 2.25152716e-01
4.61352378e-01 1.48990303e-01 -8.27447832e-01 -4.83296886e-02
7.86174648e-03 -1.91835493e-01 7.50476778e-01 5.50951183e-01
6.73889756e-01 -1.88183799e-01 -5.18066585e-01 4.30117637e-01
1.04960155e+00 5.68933845e-01 8.80623937e-01 3.31509680e-01
7.64279902e-01 4.47917521e-01 7.01195538e-01 6.89600766e-01
2.26847589e-01 8.76205802e-01 -6.38393760e-02 6.26151621e-01
1.02720343e-01 -3.86910588e-02 6.60497487e-01 5.65672278e-01
-7.36479640e-01 5.18891104e-02 -9.54431772e-01 3.05225253e-01
-2.19149256e+00 -1.73115385e+00 -5.19010983e-02 2.06718969e+00
3.31619918e-01 5.13812006e-02 3.70783240e-01 6.21484816e-01
5.10736704e-01 3.08475792e-01 -9.84412506e-02 1.78503349e-01
1.94559187e-01 2.91844457e-01 2.47564942e-01 2.16027990e-01
-1.65531850e+00 7.74097443e-01 6.53973198e+00 2.28238374e-01
-1.08457124e+00 -1.77433968e-01 1.57920152e-01 7.84679726e-02
7.99142182e-01 -2.06445739e-01 -6.68892324e-01 5.35367548e-01
8.92043054e-01 1.19583748e-01 1.59551784e-01 4.94769424e-01
5.77623785e-01 -5.56483448e-01 -1.12370586e+00 1.27549958e+00
2.15154856e-01 -1.03405285e+00 -3.08523983e-01 -5.09665273e-02
5.66698492e-01 -1.40551373e-01 -4.91204560e-01 5.26369270e-03
6.03704154e-03 -8.63500535e-01 2.81302512e-01 9.37951565e-01
1.20230861e-01 -5.91468215e-01 5.80661356e-01 5.05609751e-01
-1.37847042e+00 -1.89675212e-01 -3.02491430e-02 -3.19522560e-01
9.14978534e-02 1.33867145e-01 -7.12333739e-01 1.47913784e-01
8.71463120e-01 1.31113255e+00 -4.56462115e-01 1.22229326e+00
-1.20804101e-01 6.89272523e-01 -2.91039795e-01 -1.01148024e-01
5.22140227e-02 -2.30061725e-01 5.01674175e-01 1.20944691e+00
8.93849507e-02 5.64486384e-01 1.44119337e-01 4.45480756e-02
5.48111320e-01 -1.45530179e-01 -8.47447932e-01 -3.01565856e-01
-5.34078032e-02 1.08023083e+00 -8.03927779e-01 -4.14951563e-01
-7.93317378e-01 1.07796955e+00 1.87916551e-02 3.35370570e-01
-6.98761165e-01 -1.84849054e-01 6.49817109e-01 -2.99484283e-02
4.65796858e-01 -5.13665378e-01 4.49888945e-01 -1.43218100e+00
3.44144268e-04 -7.71956980e-01 4.89903659e-01 -6.34958684e-01
-6.52857065e-01 2.21690178e-01 2.21324190e-01 -1.49111187e+00
-5.53656697e-01 -6.91781342e-01 -7.25111067e-01 2.41363466e-01
-9.23723638e-01 -9.05522466e-01 -7.36868322e-01 9.59717691e-01
6.51693285e-01 -4.90094572e-01 6.68584883e-01 3.01024318e-01
-5.65304697e-01 4.01770957e-02 -1.38704523e-01 5.71455181e-01
4.76725370e-01 -9.85886753e-01 1.19284734e-01 8.86270106e-01
4.38946486e-01 6.52158484e-02 7.65023589e-01 -5.25978982e-01
-1.31514883e+00 -6.52588904e-01 8.40041518e-01 -5.30001283e-01
5.45752704e-01 1.21163972e-01 -6.23224616e-01 5.56818604e-01
1.36221290e-01 5.29305041e-01 4.41909045e-01 -6.01393402e-01
-4.39837500e-02 -1.36352062e-01 -9.23506260e-01 1.55648127e-01
1.15962017e+00 -3.06462407e-01 -7.70403683e-01 2.54328638e-01
7.93727785e-02 -8.58926177e-02 -7.13627219e-01 4.64047045e-01
7.41609395e-01 -1.22173691e+00 1.09030223e+00 -7.97799826e-01
1.00363486e-01 -2.86567748e-01 -2.50409245e-01 -9.82668281e-01
-2.79225379e-01 -4.91545677e-01 -4.83706594e-01 8.04046988e-01
-2.52129465e-01 -4.51002985e-01 1.13664198e+00 4.45727110e-01
3.96695107e-01 -3.30158323e-01 -1.04035115e+00 -5.15818417e-01
-6.48322999e-01 -2.46303126e-01 -4.85659689e-02 6.43988788e-01
1.04349136e-01 1.65971234e-01 -3.74890119e-01 4.38536145e-02
7.52379715e-01 -2.92204648e-01 7.65355706e-01 -1.23442400e+00
-2.32772544e-01 -1.43228605e-01 -1.30890083e+00 -7.51912415e-01
2.78702617e-01 -1.85495630e-01 -4.28455276e-03 -1.33399868e+00
2.03563958e-01 2.41284326e-01 -2.30242893e-01 2.41699263e-01
-1.27007291e-01 1.64533392e-01 5.89170977e-02 2.46681944e-01
-6.45549595e-01 4.19260859e-01 9.98630285e-01 -2.85105914e-01
-2.20271543e-01 4.19171780e-01 1.86552241e-01 9.69321191e-01
5.09768546e-01 -2.73659259e-01 -4.34263140e-01 6.66084001e-03
-3.61797154e-01 2.10218012e-01 4.37530845e-01 -1.64006209e+00
3.67824644e-01 -3.30183238e-01 5.77831209e-01 -5.90020359e-01
4.49068487e-01 -8.07836771e-01 2.42131837e-02 8.06559682e-01
-9.39123034e-02 2.75894940e-01 -2.52768517e-01 6.45267367e-01
-4.45076197e-01 -7.04234168e-02 8.38642776e-01 -1.47261873e-01
-1.28348005e+00 3.37045103e-01 -5.78441441e-01 2.08637610e-01
1.38226473e+00 -4.56129640e-01 -1.70112133e-01 -3.17112178e-01
-6.77460134e-01 -1.52965322e-01 9.44277495e-02 5.46558022e-01
6.31088436e-01 -1.38934016e+00 -3.90074193e-01 3.81534249e-01
1.27723485e-01 -4.61123496e-01 1.00701071e-01 9.70418036e-01
-8.17710221e-01 7.68898189e-01 -6.43297851e-01 -7.66739368e-01
-1.65896654e+00 4.46095169e-01 4.18347478e-01 -2.19851747e-01
-7.93525696e-01 4.96145964e-01 -2.03739434e-01 3.73155087e-01
4.29431289e-01 -1.37770534e-01 -8.52327704e-01 1.18283108e-01
9.04571116e-01 6.76111341e-01 -2.89923340e-01 -1.09633482e+00
-6.70168400e-01 8.14047873e-01 1.71999782e-01 -2.71209627e-01
1.43026483e+00 1.22386612e-01 1.27012268e-01 3.06907147e-01
1.37689555e+00 -4.48698044e-01 -1.36167359e+00 -1.96591541e-01
3.81440550e-01 -5.01662433e-01 -3.17980468e-01 -9.81316641e-02
-1.06991673e+00 8.37991774e-01 1.02702212e+00 2.48937085e-02
1.17405784e+00 -2.68627435e-01 5.58234990e-01 5.77141643e-01
5.95745325e-01 -1.29289722e+00 6.33091748e-01 4.40296680e-01
4.47886556e-01 -1.25558043e+00 9.03837085e-02 -2.05760390e-01
-7.14851558e-01 1.23529732e+00 5.82982361e-01 -3.80616158e-01
8.80309045e-01 -7.20785782e-02 3.11489915e-03 -8.57047215e-02
-5.88439226e-01 -2.68489063e-01 3.37708563e-01 8.90171409e-01
4.43112344e-01 -1.28338873e-01 -3.71509463e-01 -4.78975940e-03
1.58793971e-01 4.27448928e-01 4.46949571e-01 1.31402540e+00
-7.18014777e-01 -9.60948288e-01 -3.53310972e-01 2.81899005e-01
-5.57413161e-01 4.90068942e-01 -1.00812562e-01 6.73706293e-01
4.49125618e-02 9.13297236e-01 2.77060401e-02 -3.33964884e-01
3.44866753e-01 2.92063057e-01 9.15559173e-01 -1.98396891e-01
-1.60983160e-01 -2.88441241e-01 8.61032754e-02 -1.06471717e+00
-1.25299442e+00 -9.75126863e-01 -1.27442265e+00 5.97754158e-02
2.76192039e-01 -1.55040219e-01 3.54470462e-01 1.27687633e+00
4.34960499e-02 1.70127764e-01 5.87565303e-01 -9.58356082e-01
-1.83755204e-01 -9.06316698e-01 -5.08915722e-01 8.10505569e-01
5.11893153e-01 -9.36264634e-01 -1.96302414e-01 5.93108118e-01] | [7.99215841293335, 0.36951756477355957] |
42735d33-a6ea-4176-a2c6-dc3b23f20377 | homogeneous-learning-self-attention-1 | 2110.05290 | null | https://arxiv.org/abs/2110.05290v1 | https://arxiv.org/pdf/2110.05290v1.pdf | Homogeneous Learning: Self-Attention Decentralized Deep Learning | Federated learning (FL) has been facilitating privacy-preserving deep learning in many walks of life such as medical image classification, network intrusion detection, and so forth. Whereas it necessitates a central parameter server for model aggregation, which brings about delayed model communication and vulnerability to adversarial attacks. A fully decentralized architecture like Swarm Learning allows peer-to-peer communication among distributed nodes, without the central server. One of the most challenging issues in decentralized deep learning is that data owned by each node are usually non-independent and identically distributed (non-IID), causing time-consuming convergence of model training. To this end, we propose a decentralized learning model called Homogeneous Learning (HL) for tackling non-IID data with a self-attention mechanism. In HL, training performs on each round's selected node, and the trained model of a node is sent to the next selected node at the end of each round. Notably, for the selection, the self-attention mechanism leverages reinforcement learning to observe a node's inner state and its surrounding environment's state, and find out which node should be selected to optimize the training. We evaluate our method with various scenarios for an image classification task. The result suggests that HL can produce a better performance compared with standalone learning and greatly reduce both the total training rounds by 50.8% and the communication cost by 74.6% compared with random policy-based decentralized learning for training on non-IID data. | ['Hideya Ochiai', 'Yuwei Sun'] | 2021-10-11 | homogeneous-learning-self-attention | https://openreview.net/forum?id=BvowzJp_Yl6 | https://openreview.net/pdf?id=BvowzJp_Yl6 | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [-4.63420093e-01 3.29294086e-01 -3.06583256e-01 -2.86749661e-01
-6.37627125e-01 -5.00639677e-01 2.03725040e-01 3.35096329e-01
-6.63007021e-01 8.12370479e-01 -1.60276026e-01 -2.65944570e-01
2.42577735e-02 -7.61675715e-01 -7.03574300e-01 -1.16276050e+00
-3.92052919e-01 4.02884513e-01 -2.57645044e-02 6.55072033e-02
-3.82497549e-01 5.30918062e-01 -9.04155374e-01 -1.59378499e-01
5.92709720e-01 1.24345994e+00 4.09333706e-02 5.66765428e-01
1.82078302e-01 1.06225872e+00 -8.76071692e-01 -2.72206157e-01
4.32138056e-01 -1.65233865e-01 -6.44670308e-01 -2.45242164e-01
-1.49363950e-01 -6.18730366e-01 -5.02745032e-01 1.05466866e+00
7.14990139e-01 -6.94206059e-02 1.17783725e-01 -1.64725220e+00
-2.51954049e-01 7.59299278e-01 -3.63971561e-01 -9.10062566e-02
-2.47311756e-01 2.56214529e-01 5.84864020e-01 -2.05974374e-02
5.15412807e-01 9.91052032e-01 6.16193414e-01 9.79161382e-01
-1.02809405e+00 -1.03125155e+00 3.03817123e-01 6.12283014e-02
-9.74001944e-01 -3.84264052e-01 8.54999006e-01 8.92685950e-02
5.49066842e-01 3.23480606e-01 7.54843950e-01 1.10493791e+00
4.78127122e-01 9.04009402e-01 8.25059116e-01 -2.29457067e-03
8.34054112e-01 2.69353122e-01 -3.20427775e-01 6.39624178e-01
1.18441530e-01 5.84065132e-02 -4.96527702e-01 -5.86830437e-01
5.05497754e-01 2.83315510e-01 -5.52083142e-02 -2.97032356e-01
-7.43465722e-01 7.80079424e-01 7.43237376e-01 -1.90237556e-02
-6.88609183e-01 4.38030601e-01 5.68254292e-01 7.17598498e-01
6.24577463e-01 2.93093324e-01 -8.50868225e-01 7.20226541e-02
-6.47996426e-01 3.37901749e-02 1.11671066e+00 6.26562774e-01
1.04670608e+00 -9.86251161e-02 -2.29017418e-02 2.59805292e-01
3.30083728e-01 4.61289227e-01 5.33830762e-01 -1.21125400e+00
2.24838257e-01 6.09106600e-01 -8.80205166e-03 -1.02216196e+00
-3.98007214e-01 -6.40604138e-01 -1.30741000e+00 1.15854189e-01
-5.62954275e-03 -9.57491636e-01 -4.07753646e-01 1.95849490e+00
9.58783388e-01 3.26027632e-01 3.60452414e-01 6.28466070e-01
4.75997120e-01 5.23603916e-01 4.49861512e-02 -3.62725317e-01
9.54709470e-01 -1.07725215e+00 -4.54902738e-01 -9.02632624e-02
8.26449215e-01 -3.63701582e-02 3.58451039e-01 1.76632494e-01
-7.44947195e-01 4.68743928e-02 -6.80406272e-01 3.72300744e-01
-8.45474005e-02 -2.17167363e-01 6.88513517e-01 3.04644078e-01
-1.25697660e+00 6.55215144e-01 -1.24892068e+00 -2.46479169e-01
9.04114783e-01 7.63924181e-01 -5.35461009e-01 2.50742584e-01
-1.06234658e+00 3.36008698e-01 9.18774605e-02 -1.65457025e-01
-1.31436610e+00 -9.35334980e-01 -5.99267304e-01 2.29328051e-01
2.64688581e-01 -8.01863253e-01 1.24938071e+00 -9.15494204e-01
-1.49619353e+00 3.40853125e-01 -9.74719319e-03 -1.03197145e+00
6.99572563e-01 8.28633085e-02 -1.28501892e-01 2.44210258e-01
8.44500586e-02 5.30829847e-01 9.18583393e-01 -1.19670081e+00
-6.27836943e-01 -4.55323696e-01 1.19592741e-01 2.96930104e-01
-9.62310374e-01 -1.18060581e-01 -3.08921605e-01 -3.74800980e-01
-2.05401257e-01 -9.75350261e-01 -5.73392451e-01 4.11427557e-01
-3.98747921e-01 -3.21129292e-01 1.43141317e+00 -4.18666720e-01
9.23525274e-01 -2.23110342e+00 -1.65651187e-01 2.06335202e-01
6.02272868e-01 2.49290422e-01 -1.40782997e-01 5.82100034e-01
3.91419977e-01 2.31191233e-01 1.89379510e-02 -8.09472203e-01
-2.16373846e-01 4.34518933e-01 -1.43209815e-01 6.38954282e-01
-2.21577913e-01 8.78265798e-01 -8.58774185e-01 -6.13131404e-01
-1.01686589e-01 4.07771975e-01 -6.79268360e-01 2.81485558e-01
-2.07251340e-01 7.52108157e-01 -9.26481307e-01 4.46486920e-01
5.55374265e-01 -6.17705584e-01 3.14388454e-01 1.13870263e-01
2.89295375e-01 -9.97602493e-02 -9.04366076e-01 1.50981677e+00
-4.27577555e-01 4.60228533e-01 8.55813384e-01 -9.81132686e-01
7.18279064e-01 5.26954889e-01 9.77440178e-01 -5.15357792e-01
1.38362497e-01 -4.13880944e-02 -1.89106986e-01 -3.54728639e-01
-1.60295695e-01 1.64799586e-01 -6.13348782e-02 8.64603877e-01
-6.58045486e-02 3.27089757e-01 -4.97905642e-01 4.10946965e-01
1.57609797e+00 -5.29903948e-01 7.40774423e-02 -1.85029656e-01
3.49623442e-01 -1.47565693e-01 1.07825220e+00 9.30626214e-01
-5.62628865e-01 -1.12433828e-01 4.77847993e-01 -7.18711972e-01
-6.89944148e-01 -6.59347177e-01 3.30347747e-01 1.04973292e+00
3.35680693e-01 -3.18541348e-01 -9.62131023e-01 -1.06761563e+00
3.12142193e-01 3.06982309e-01 -6.94043398e-01 -4.92854118e-01
-3.89358073e-01 -5.42042673e-01 5.80181062e-01 -3.01473588e-02
8.18588138e-01 -1.17742968e+00 -6.98477447e-01 3.07092398e-01
9.08151865e-02 -7.46350706e-01 -4.47843671e-01 4.28510040e-01
-6.54891014e-01 -9.94887114e-01 -2.40383506e-01 -6.17969751e-01
9.25559700e-01 1.55245543e-01 6.94492698e-01 3.09644878e-01
-1.03560694e-01 4.63770449e-01 -1.09622933e-01 -6.88573480e-01
-6.16222680e-01 3.68391514e-01 3.06240082e-01 2.98274726e-01
-1.81553029e-02 -6.21350825e-01 -9.42479730e-01 2.83940732e-01
-9.17525053e-01 -3.32449675e-01 4.19685364e-01 8.27280998e-01
4.60520983e-01 1.70712918e-01 9.06126380e-01 -9.03905928e-01
6.66643202e-01 -7.62230337e-01 -5.02054751e-01 3.39765668e-01
-7.41087496e-01 -2.39939969e-02 1.24186194e+00 -5.61843276e-01
-7.85714209e-01 2.12360471e-01 3.25891942e-01 -7.61604786e-01
-6.02671057e-02 2.12880194e-01 -5.72374985e-02 -3.44964385e-01
5.67424417e-01 3.17827314e-01 6.63454831e-01 -3.21095884e-01
-4.45037186e-02 8.11489284e-01 1.53240770e-01 -2.97360212e-01
8.79817903e-01 7.57137001e-01 7.08183572e-02 -5.45512676e-01
-6.35398567e-01 -5.90667389e-02 3.82902625e-04 -9.07668173e-02
6.23546779e-01 -1.18107808e+00 -1.20156550e+00 7.71841228e-01
-1.02639437e+00 -6.23922825e-01 -4.96413589e-01 3.13421994e-01
-1.30810648e-01 1.39113933e-01 -7.29616523e-01 -7.75872469e-01
-9.82024729e-01 -9.69455063e-01 6.56886876e-01 5.86276829e-01
1.73323065e-01 -1.27319551e+00 -9.92060080e-02 4.20127928e-01
7.88551509e-01 2.02040538e-01 5.42447209e-01 -9.72399294e-01
-4.04305696e-01 -3.81472766e-01 2.68630475e-01 4.75371093e-01
2.51512498e-01 -2.75532097e-01 -9.20280755e-01 -8.98950398e-01
3.08725595e-01 -5.41863918e-01 3.62032533e-01 1.64268568e-01
1.31854272e+00 -9.91083801e-01 -5.19152045e-01 8.42067599e-01
1.15740860e+00 5.94848245e-02 -8.44943076e-02 2.71411240e-01
6.12326264e-01 3.36159259e-01 3.47529799e-01 8.35904837e-01
6.93917513e-01 7.51161948e-02 9.50194120e-01 -3.25492710e-01
3.03613991e-01 -3.33819777e-01 3.34794194e-01 6.61108613e-01
7.40013421e-01 -1.32442877e-01 -6.43264711e-01 3.78896385e-01
-1.91455603e+00 -7.48815894e-01 5.57833374e-01 2.05385900e+00
8.65654707e-01 -3.20763707e-01 -3.96272391e-02 -3.88147444e-01
5.58323026e-01 1.49645686e-01 -1.34945095e+00 -9.97080579e-02
-1.04691833e-01 -1.78161159e-01 7.99809396e-01 2.28427380e-01
-1.13837194e+00 9.26716924e-01 5.12324524e+00 6.58697963e-01
-1.58062184e+00 2.54241168e-01 9.52151716e-01 -4.08437133e-01
-7.94653874e-03 4.93666157e-02 -5.82450688e-01 5.37850618e-01
9.38374221e-01 -4.72351521e-01 7.03141809e-01 1.14863443e+00
2.18132988e-01 1.19187891e-01 -8.23612988e-01 8.45340490e-01
-3.24719071e-01 -1.48398662e+00 -2.80750781e-01 3.17380935e-01
9.45249438e-01 6.48061275e-01 -9.13540050e-02 1.48790538e-01
7.60474563e-01 -7.80067861e-01 3.02477926e-01 2.67772079e-01
3.21940064e-01 -9.19904172e-01 5.53457975e-01 9.69690979e-01
-8.38191032e-01 -5.75202405e-01 -4.02861029e-01 1.54339731e-01
-2.55819410e-01 4.87478524e-01 -7.39871442e-01 4.40178633e-01
1.00384986e+00 6.52546167e-01 -3.27676475e-01 8.31499219e-01
-1.66085288e-02 7.26172507e-01 -6.99118674e-01 -7.70101994e-02
2.80962437e-01 1.33858051e-03 6.95293427e-01 5.30949175e-01
-2.15200018e-02 -1.66550040e-01 3.05586457e-01 4.38374698e-01
-5.89150012e-01 3.18821892e-02 -5.49625933e-01 3.59302908e-01
1.01630056e+00 1.41826093e+00 -3.12295377e-01 -2.96145022e-01
-7.21638873e-02 9.70363081e-01 6.35477006e-01 4.04400587e-01
-5.09725034e-01 -2.11510912e-01 9.99383450e-01 -2.73374617e-01
2.62857884e-01 1.98131353e-01 -3.83818373e-02 -1.06136072e+00
1.22930832e-01 -9.66925442e-01 6.17772639e-01 -4.66904789e-02
-1.41909969e+00 5.43796897e-01 -5.16503751e-01 -1.04505396e+00
-1.84630707e-01 1.18504919e-01 -8.97467673e-01 4.43533897e-01
-1.46595347e+00 -1.09682930e+00 -1.82582900e-01 9.98930752e-01
-9.02795345e-02 -4.24693197e-01 1.02589524e+00 1.15421496e-01
-9.39949751e-01 1.02188230e+00 5.44731975e-01 4.54878539e-01
6.16441846e-01 -1.02309608e+00 1.38976753e-01 6.10249221e-01
3.37232128e-02 4.36158240e-01 3.60815376e-01 -5.38547337e-01
-1.73394203e+00 -1.51853395e+00 4.89521891e-01 -1.16213318e-02
5.26572227e-01 -4.89763349e-01 -7.17943728e-01 5.63305855e-01
2.77978070e-02 7.16877162e-01 5.98138750e-01 -3.75799298e-01
-1.60864487e-01 -7.66851902e-01 -1.69066083e+00 4.98183906e-01
5.49348831e-01 -4.52571929e-01 4.27579850e-01 7.13094890e-01
1.00857306e+00 -2.22133353e-01 -1.11078453e+00 1.03451662e-01
1.05784491e-01 -8.14286053e-01 6.12317324e-01 -7.54257381e-01
-2.50662476e-01 -9.75775421e-02 2.09897682e-01 -1.40846586e+00
-8.69663656e-02 -1.37390280e+00 -4.00227547e-01 1.10872614e+00
2.77123064e-01 -1.23545718e+00 1.31649530e+00 8.89558434e-01
2.50630200e-01 -1.11635947e+00 -1.52266347e+00 -5.39637029e-01
2.00961277e-01 3.69103290e-02 8.95186961e-01 1.02018368e+00
-2.63781339e-01 7.45917633e-02 -4.00219679e-01 4.05908138e-01
9.58319724e-01 -9.20860320e-02 1.09842277e+00 -1.06793046e+00
-3.53411555e-01 3.03688534e-02 -3.40059757e-01 -7.69649327e-01
4.58766580e-01 -7.02210903e-01 -2.30779260e-01 -1.26627362e+00
-1.29650071e-01 -8.93303216e-01 -4.68448132e-01 1.01213515e+00
5.02880365e-02 -2.45728806e-01 3.49316061e-01 4.34875786e-01
-9.94243979e-01 6.65734231e-01 1.22682357e+00 -2.03522444e-01
-3.07474107e-01 4.01641399e-01 -8.42486858e-01 4.07740980e-01
1.09334409e+00 -8.63662899e-01 -4.79452729e-01 -5.35351455e-01
-1.70194983e-01 3.61758500e-01 4.06697154e-01 -9.71918523e-01
8.28507364e-01 -1.53310552e-01 1.77184314e-01 -8.35043713e-02
1.77828759e-01 -1.21095240e+00 1.80745646e-01 9.28523123e-01
-3.58556807e-01 -6.51397482e-02 -1.30885288e-01 8.05216551e-01
-4.55173552e-02 3.11859369e-01 8.07351708e-01 -6.45561069e-02
-1.28515214e-01 8.57063174e-01 -1.19528122e-01 4.00667712e-02
1.54280472e+00 2.35005036e-01 -4.67282295e-01 -7.63286114e-01
-6.07023597e-01 8.79382610e-01 5.96322238e-01 2.37007648e-01
2.74266899e-01 -9.40572917e-01 -6.55095637e-01 3.66899133e-01
-3.08415413e-01 6.11878037e-01 3.62536103e-01 8.49014163e-01
-1.97794601e-01 -3.50140408e-02 1.67028382e-01 -5.40349841e-01
-1.24247372e+00 4.38260198e-01 6.63130581e-01 -3.82430732e-01
-6.07935071e-01 8.50703418e-01 4.16409830e-03 -7.06381738e-01
5.96109331e-01 1.62287146e-01 1.99619681e-01 -1.55046284e-01
3.34306836e-01 3.14991325e-01 -6.55861050e-02 -2.43751064e-01
-3.87449205e-01 -1.43426806e-01 -3.71750474e-01 2.14934900e-01
1.54209924e+00 -5.77716306e-02 -4.74819839e-01 5.61153218e-02
1.54702151e+00 -4.06289160e-01 -1.67502558e+00 -5.23130834e-01
-4.21323508e-01 4.30372544e-03 1.91348970e-01 -6.70118630e-01
-1.65343201e+00 3.50642949e-01 4.81616318e-01 2.66354144e-01
1.15808761e+00 1.97186526e-02 9.48688447e-01 6.78715229e-01
7.26346970e-01 -1.06348145e+00 -1.87725648e-01 2.34022826e-01
4.27602857e-01 -1.31680560e+00 -4.77095619e-02 1.96172938e-01
-6.37438834e-01 7.90018082e-01 8.26207042e-01 -2.00723350e-01
1.14786541e+00 4.21049953e-01 3.54880810e-01 -2.46675033e-02
-1.28991318e+00 6.31743729e-01 -3.84619236e-01 4.74202216e-01
-4.21919852e-01 8.16183463e-02 3.41616273e-01 4.38793331e-01
-7.81450421e-02 -1.89839438e-01 1.56842336e-01 1.11563075e+00
-1.15421750e-01 -1.11277783e+00 -4.72348295e-02 6.92895532e-01
-6.48709774e-01 3.41587633e-01 -3.74488324e-01 4.02334034e-01
-1.67502314e-01 1.10159147e+00 5.99796809e-02 -4.91387427e-01
-1.28754720e-01 -3.81108373e-01 -2.66918749e-01 -3.78300995e-01
-1.23295569e+00 -2.08025202e-01 -3.75519246e-01 -7.54511535e-01
-7.11638182e-02 -5.24097264e-01 -1.39099884e+00 -4.77192134e-01
-2.10891441e-01 6.00563526e-01 8.58013332e-01 6.40353620e-01
1.02930832e+00 1.02297477e-01 1.49327600e+00 -6.33662641e-01
-1.21894145e+00 -6.14607632e-01 -4.90006089e-01 2.96518803e-01
7.54026890e-01 -1.11563066e-02 -6.23054206e-01 -5.45291603e-01] | [5.934664726257324, 6.381450653076172] |
6255b9dd-ad70-4fb0-943a-640a71e08e7b | constrained-r-cnn-a-general-image | 1911.08217 | null | https://arxiv.org/abs/1911.08217v3 | https://arxiv.org/pdf/1911.08217v3.pdf | Constrained R-CNN: A general image manipulation detection model | Recently, deep learning-based models have exhibited remarkable performance for image manipulation detection. However, most of them suffer from poor universality of handcrafted or predetermined features. Meanwhile, they only focus on manipulation localization and overlook manipulation classification. To address these issues, we propose a coarse-to-fine architecture named Constrained R-CNN for complete and accurate image forensics. First, the learnable manipulation feature extractor learns a unified feature representation directly from data. Second, the attention region proposal network effectively discriminates manipulated regions for the next manipulation classification and coarse localization. Then, the skip structure fuses low-level and high-level information to refine the global manipulation features. Finally, the coarse localization information guides the model to further learn the finer local features and segment out the tampered region. Experimental results show that our model achieves state-of-the-art performance. Especially, the F1 score is increased by 28.4%, 73.2%, 13.3% on the NIST16, COVERAGE, and Columbia dataset. | ['Hao Zhao', 'Fangting Lin', 'Chao Yang', 'Bin Jiang', 'Huizhou Li'] | 2019-11-19 | null | null | null | null | ['image-manipulation-detection', 'image-forensics'] | ['computer-vision', 'computer-vision'] | [ 8.94577429e-02 -6.38401985e-01 -5.79372227e-01 -1.45044535e-01
-9.47256982e-01 -3.39552402e-01 2.99795985e-01 -1.28021389e-01
-2.38591611e-01 1.19508728e-01 1.39707327e-01 6.98061362e-02
-1.07174240e-01 -7.12378204e-01 -6.98337793e-01 -8.38432908e-01
-1.37153305e-02 -1.99081928e-01 2.65016228e-01 -1.54158756e-01
6.47488594e-01 7.33060300e-01 -1.32219505e+00 4.63142902e-01
5.49786925e-01 1.31370807e+00 2.43350923e-01 3.88989210e-01
2.51866877e-01 8.51379275e-01 -8.13747823e-01 -1.21056335e-02
2.84773290e-01 6.68301061e-02 -4.35949713e-01 1.63262740e-01
3.38314801e-01 -9.32321250e-01 -9.78916943e-01 1.25524068e+00
3.47448498e-01 -1.34820733e-02 4.99027461e-01 -1.11789536e+00
-1.22983837e+00 4.69811887e-01 -1.07093251e+00 3.74412745e-01
1.80347517e-01 6.50606811e-01 5.56622148e-01 -8.63963962e-01
5.42450070e-01 1.30117083e+00 3.41669768e-01 4.27598089e-01
-7.71909118e-01 -1.02182853e+00 3.11001509e-01 4.67364281e-01
-1.42011404e+00 -4.05289203e-01 1.04237044e+00 -3.25509787e-01
7.11837053e-01 -2.29498595e-02 9.04351920e-02 1.13960516e+00
2.09470317e-01 9.76795614e-01 1.02372038e+00 -7.96748921e-02
-3.30105871e-01 -3.03473949e-01 3.57521802e-01 1.04875338e+00
3.08163524e-01 1.47874981e-01 -5.26677489e-01 2.71487504e-01
9.53741014e-01 3.28784376e-01 -4.62005109e-01 -2.16124821e-02
-1.34027326e+00 6.28567040e-01 7.16665149e-01 3.31135958e-01
-4.59877938e-01 1.90382153e-01 6.37306869e-01 4.20688046e-03
4.41374518e-02 4.34776813e-01 -3.70430648e-01 -1.36861935e-01
-8.71679962e-01 2.15692118e-01 1.75087526e-02 1.06035352e+00
8.59493852e-01 1.04928203e-01 -5.13139963e-01 4.17176723e-01
1.33461645e-02 5.11034846e-01 2.86244750e-01 -9.88143623e-01
8.43154609e-01 7.04368293e-01 8.57953131e-02 -1.45499063e+00
-2.14063585e-01 -2.74737060e-01 -1.00205386e+00 5.58202229e-02
3.50763887e-01 3.58441502e-01 -1.12578642e+00 1.28899884e+00
1.65965587e-01 5.05838245e-02 -4.64225203e-01 1.04171312e+00
6.81663513e-01 3.61934721e-01 8.65146443e-02 8.94738510e-02
1.43535709e+00 -1.19323921e+00 -7.21704960e-01 -2.35478301e-02
3.08057755e-01 -6.41811728e-01 1.17986274e+00 5.99236965e-01
-9.09814775e-01 -9.34556007e-01 -1.07896912e+00 -3.43355358e-01
-5.61260343e-01 6.26935422e-01 5.69253445e-01 5.81964076e-01
-5.40858865e-01 5.35061121e-01 -6.19571030e-01 2.28904575e-01
1.00574577e+00 3.41335535e-01 -4.28296000e-01 -2.74815142e-01
-9.49116409e-01 5.23688614e-01 4.44748908e-01 3.38330686e-01
-1.43822765e+00 -5.50651670e-01 -7.10309505e-01 3.43687415e-01
5.69324195e-01 9.29667521e-03 8.22182417e-01 -2.83332288e-01
-1.30535066e+00 7.77054191e-01 4.32283580e-02 -5.61036170e-02
2.71052599e-01 -3.34297597e-01 -3.19667518e-01 4.21812743e-01
1.40269116e-01 4.79414552e-01 1.16732156e+00 -1.61075652e+00
-6.18470967e-01 -4.46107298e-01 3.01809132e-01 -4.20916617e-01
-5.46080768e-01 3.59421343e-01 -7.59335339e-01 -9.27869260e-01
-1.16255262e-03 -4.19357479e-01 4.85954992e-03 1.06898211e-01
-6.19205117e-01 -1.43355027e-01 1.17981291e+00 -1.04708922e+00
1.38477337e+00 -2.29629803e+00 4.98696715e-02 9.47407261e-02
4.18907672e-01 4.83316571e-01 -4.33953762e-01 -4.29977030e-02
-1.20994411e-01 2.62378037e-01 6.29215613e-02 -1.29602402e-01
8.76082927e-02 -3.86849135e-01 -5.32180727e-01 9.56768095e-01
2.98844635e-01 1.10969913e+00 -8.42247367e-01 -4.41102803e-01
5.54123998e-01 3.77680242e-01 -4.91896242e-01 2.41329521e-01
-5.32174408e-02 4.83726978e-01 -7.10726619e-01 1.27665710e+00
8.95475864e-01 -1.01405717e-01 -3.65475714e-01 -6.47756100e-01
3.87357399e-02 -4.54545617e-02 -8.28759372e-01 1.91377831e+00
-3.48375946e-01 4.85715121e-01 2.63160497e-01 -9.97504473e-01
8.06159496e-01 -1.42101809e-01 4.32439059e-01 -8.69552135e-01
5.55228353e-01 -7.13907881e-04 -1.90040514e-01 -9.03785646e-01
6.68628633e-01 5.02415538e-01 -3.38786721e-01 2.88026571e-01
-3.81854661e-02 3.71508121e-01 -1.99616954e-01 -2.50371452e-02
1.05903125e+00 1.37172803e-01 4.84211296e-02 9.60284173e-02
7.60777354e-01 -3.87627661e-01 5.64298689e-01 8.26510906e-01
-6.55217886e-01 6.44330323e-01 3.69262844e-01 -2.98854709e-01
-7.85674453e-01 -9.10659850e-01 4.64716144e-02 9.45194244e-01
7.39651322e-01 -3.30354929e-01 -8.81089270e-01 -8.96975696e-01
-1.07368164e-01 2.62665629e-01 -7.84875393e-01 -5.29661298e-01
-1.07588363e+00 -2.93127954e-01 7.48986900e-01 6.54042900e-01
1.05978441e+00 -1.23815858e+00 -2.84877181e-01 -6.49410859e-02
-3.74524951e-01 -1.19257081e+00 -5.22504985e-01 -2.98890352e-01
-5.33906996e-01 -1.24176776e+00 -4.71399248e-01 -9.06287134e-01
7.04617202e-01 6.96790159e-01 4.53282684e-01 3.79180402e-01
-5.69385767e-01 -7.35751633e-03 -4.51318443e-01 -7.95813873e-02
1.21445082e-01 -2.23114155e-02 -1.52600184e-01 3.57748531e-02
5.10912061e-01 -1.34960905e-01 -7.38385320e-01 3.89413863e-01
-8.72763097e-01 -5.69498539e-01 8.50178778e-01 9.91187930e-01
5.25002003e-01 4.05224055e-01 4.51601446e-01 -2.60725200e-01
4.82325077e-01 -3.61528784e-01 -3.78085107e-01 3.51746261e-01
-2.90982693e-01 -2.37919256e-01 5.80593050e-01 -6.05652094e-01
-9.96889710e-01 -1.63195774e-01 1.18473992e-01 -9.90903020e-01
-3.63198191e-01 1.25034109e-01 -6.51117444e-01 -4.41208690e-01
5.40541351e-01 5.26740432e-01 -4.00616080e-01 -5.94732046e-01
4.03602034e-01 8.26907635e-01 7.63726592e-01 -6.69941604e-01
1.21581733e+00 6.54533029e-01 -2.16507718e-01 -4.17166144e-01
-6.59693182e-01 -5.29772222e-01 -6.52999222e-01 -1.86871424e-01
1.10584760e+00 -8.41697812e-01 -1.15064406e+00 9.63647723e-01
-1.23875511e+00 -1.53040692e-01 -4.29809317e-02 1.81935802e-01
-3.89465213e-01 9.41863179e-01 -8.20738912e-01 -7.10486650e-01
-3.64985406e-01 -1.56215334e+00 1.63458753e+00 3.55095416e-01
3.30041140e-01 -3.02293360e-01 -6.96681023e-01 7.05814064e-01
3.69453579e-01 3.64902079e-01 8.77654195e-01 -1.91678822e-01
-1.05545855e+00 -5.78946173e-01 -8.64581466e-01 2.87104279e-01
1.87181935e-01 -1.68241426e-01 -1.07423973e+00 -3.32987487e-01
1.15903348e-01 -5.64510167e-01 1.19147742e+00 7.90208206e-02
2.34415007e+00 -9.81804505e-02 -3.83565068e-01 8.07609439e-01
1.20581460e+00 9.85598564e-02 8.89495611e-01 5.10742903e-01
9.42541778e-01 3.19912702e-01 9.61795688e-01 4.40379620e-01
2.10798010e-01 5.39093018e-01 7.46166527e-01 2.19231769e-01
-2.82513738e-01 -2.68197685e-01 4.27594662e-01 5.75706422e-01
-4.90297452e-02 -8.99534225e-02 -6.43106163e-01 5.83880544e-01
-1.59482300e+00 -1.02487051e+00 1.72836512e-01 1.92289758e+00
6.26714885e-01 2.80600727e-01 -1.00245245e-01 3.22971523e-01
9.94291604e-01 6.29990458e-01 -5.10788083e-01 1.33350357e-01
-1.68667082e-02 3.07370961e-01 7.27879584e-01 6.89190179e-02
-1.48013210e+00 1.22160542e+00 5.36145926e+00 1.50691438e+00
-1.04852700e+00 1.39720902e-01 5.48012257e-01 3.66179496e-02
1.46709174e-01 -2.31051624e-01 -9.07561779e-01 6.52043343e-01
3.78559619e-01 4.10323352e-01 6.91468596e-01 7.88423657e-01
1.51890129e-01 5.50414622e-02 -6.59888506e-01 1.11140871e+00
3.43524307e-01 -1.31273329e+00 1.37115076e-01 1.98479488e-01
5.94143987e-01 -2.76089847e-01 3.09616119e-01 6.15537822e-01
-1.47887170e-01 -1.21156120e+00 7.49777734e-01 5.56770682e-01
9.36507702e-01 -9.61937666e-01 5.09324193e-01 3.96434426e-01
-1.40265107e+00 -4.42259908e-01 -3.58841658e-01 3.28635186e-01
9.22789052e-02 3.03493887e-01 -1.52171537e-01 4.93558735e-01
5.71238875e-01 5.46689808e-01 -5.58893144e-01 9.62673426e-01
-4.93233442e-01 4.00049716e-01 9.37509388e-02 3.02563965e-01
3.27604115e-01 1.74115583e-01 2.98704833e-01 1.16996789e+00
9.34516359e-03 -1.49699496e-02 4.43786323e-01 8.92875016e-01
-4.88251656e-01 -3.09139252e-01 -3.29694629e-01 -6.71834722e-02
5.81484258e-01 1.32821488e+00 -5.66171050e-01 -2.70229191e-01
-2.70682991e-01 9.65699673e-01 5.32159746e-01 4.14844811e-01
-1.26881540e+00 -1.01899815e+00 6.79363549e-01 -1.14421189e-01
5.90548038e-01 -3.37269723e-01 -3.23282778e-01 -1.38573194e+00
3.41133624e-01 -1.16962004e+00 4.68822420e-02 -5.82450569e-01
-1.12285781e+00 3.23776990e-01 -1.98098496e-01 -1.23661435e+00
3.20121735e-01 -9.40043330e-01 -5.86625755e-01 7.96821713e-01
-1.72994423e+00 -1.46263671e+00 -7.53191829e-01 6.71185315e-01
6.59345925e-01 -2.77495265e-01 6.30122900e-01 5.05210698e-01
-8.95745695e-01 1.04935610e+00 -1.09732233e-01 7.91184843e-01
7.49128461e-01 -7.96826661e-01 4.33683008e-01 9.83215332e-01
-2.31686518e-01 9.47932184e-01 -5.23614511e-02 -7.62789428e-01
-1.61761820e+00 -1.41908264e+00 3.49117458e-01 -5.16233504e-01
6.81123793e-01 -4.27674055e-01 -9.65279937e-01 4.63151187e-01
-1.15438730e-01 4.30419296e-01 1.72152430e-01 -2.80793607e-01
-9.06473219e-01 -1.29365504e-01 -1.47470295e+00 4.26260471e-01
1.00274122e+00 -1.00399423e+00 -2.89624542e-01 3.57093453e-01
7.31932759e-01 -4.83971000e-01 -1.07850301e+00 3.88790339e-01
5.71587324e-01 -7.71407783e-01 1.15603375e+00 -6.51875913e-01
6.23552024e-01 -4.08380687e-01 -2.78870493e-01 -6.87296450e-01
-6.32971048e-01 -5.56968451e-01 -6.05100036e-01 1.22932029e+00
-1.51996553e-01 -2.72917718e-01 8.04574430e-01 1.90746382e-01
-2.33540878e-01 -9.63382483e-01 -6.53198957e-01 -7.84478545e-01
-1.67693831e-02 -3.86191249e-01 8.26748669e-01 9.78447616e-01
-1.91481605e-01 -2.89130539e-01 -6.00772977e-01 4.42762911e-01
8.11326504e-01 3.14298153e-01 9.03970242e-01 -5.71022332e-01
-2.69656628e-01 -4.70919758e-01 -6.59830391e-01 -1.45626509e+00
3.53121519e-01 -5.33286273e-01 1.33240551e-01 -1.07238066e+00
4.86462742e-01 -3.80232245e-01 -4.48323280e-01 5.21056056e-01
-5.12444139e-01 3.18761796e-01 2.12834433e-01 2.71683276e-01
-7.56167889e-01 4.61705804e-01 1.45980489e+00 -5.18583357e-01
1.58072740e-01 -5.05615234e-01 -7.45758116e-01 6.11120701e-01
1.04838991e+00 -2.90284991e-01 -4.16811742e-02 -7.30955899e-01
-4.82966959e-01 -1.73973754e-01 6.05903625e-01 -9.63505030e-01
9.82980579e-02 -4.37601238e-01 7.27670074e-01 -8.58684778e-01
3.18058491e-01 -5.80477178e-01 -6.52458966e-01 4.45991158e-01
-3.09027970e-01 -3.54608536e-01 5.63213155e-02 7.77802765e-01
-1.87727079e-01 -2.00046360e-01 8.57505679e-01 -6.09602220e-02
-8.67487371e-01 6.96225166e-01 -9.26392823e-02 -6.01262338e-02
1.12398517e+00 -6.78823963e-02 -7.79033601e-01 -8.21666718e-02
-3.27787399e-01 1.17286429e-01 4.12574261e-01 7.44080424e-01
8.13233376e-01 -1.31278384e+00 -4.12716955e-01 3.15954357e-01
2.84163713e-01 1.30622283e-01 4.01570588e-01 6.55902386e-01
-5.35729945e-01 4.61377561e-01 -3.15931439e-01 -4.26268965e-01
-1.04325056e+00 9.80836332e-01 6.07449524e-02 -2.48930126e-01
-4.29083288e-01 9.08265769e-01 1.35616824e-01 -8.08692053e-02
4.33527738e-01 -4.33783472e-01 -9.15322900e-02 -2.72399873e-01
8.60858560e-01 5.95009327e-01 -2.98493624e-01 -8.69020998e-01
-3.77502829e-01 7.60890007e-01 -3.99133921e-01 4.79057938e-01
1.20405984e+00 6.93252385e-02 -1.19674914e-01 -4.01491284e-01
1.49146020e+00 6.40849993e-02 -1.31676853e+00 -2.44069219e-01
-2.58241355e-01 -9.87908006e-01 1.59142867e-01 -5.95571995e-01
-1.56525671e+00 1.20219290e+00 4.99202192e-01 -8.42782110e-02
1.32395983e+00 -1.75098538e-01 1.10797238e+00 4.62655634e-01
5.02177417e-01 -8.80177915e-01 3.67018729e-01 4.70400095e-01
9.40830648e-01 -1.22730672e+00 2.36556977e-01 -3.88418376e-01
-2.57592291e-01 1.16360414e+00 1.04120922e+00 -5.61640620e-01
4.24814641e-01 2.07394347e-01 -3.02112341e-01 -3.79669607e-01
-1.00037448e-01 -3.80529538e-02 1.47839785e-01 8.17402959e-01
-1.85051896e-02 -1.40657574e-01 1.04176462e-01 7.88400471e-01
1.99329346e-01 -1.21584162e-01 4.22036834e-02 8.56933117e-01
-5.78722656e-01 -6.92360997e-01 -4.01443988e-01 4.63896185e-01
-5.53665578e-01 8.76963884e-02 -3.50815326e-01 8.49566460e-01
5.35924792e-01 9.13963795e-01 -2.02845842e-01 -6.71606839e-01
2.27648929e-01 -5.52390397e-01 6.08480632e-01 -3.21951270e-01
-7.04399407e-01 -1.32865056e-01 -2.47367829e-01 -9.77974653e-01
-1.11547433e-01 -4.29778188e-01 -1.06003284e+00 -2.80340970e-01
-7.07058907e-01 -2.25620821e-01 5.43193340e-01 7.61383653e-01
4.07937825e-01 6.04770243e-01 9.68013883e-01 -1.18105900e+00
-8.55720818e-01 -9.93097067e-01 -4.57981616e-01 4.19504285e-01
4.74578977e-01 -8.07401955e-01 -1.99306339e-01 3.14467251e-02] | [12.256134033203125, 0.8996975421905518] |
86f24687-b19d-4973-8db8-a065f25320f9 | computing-nonlinear-eigenfunctions-via | 1902.10414 | null | http://arxiv.org/abs/1902.10414v1 | http://arxiv.org/pdf/1902.10414v1.pdf | Computing Nonlinear Eigenfunctions via Gradient Flow Extinction | In this work we investigate the computation of nonlinear eigenfunctions via
the extinction profiles of gradient flows. We analyze a scheme that recursively
subtracts such eigenfunctions from given data and show that this procedure
yields a decomposition of the data into eigenfunctions in some cases as the
1-dimensional total variation, for instance. We discuss results of numerical
experiments in which we use extinction profiles and the gradient flow for the
task of spectral graph clustering as used, e.g., in machine learning
applications. | ['Daniel Tenbrinck', 'Martin Burger', 'Leon Bungert'] | 2019-02-27 | null | null | null | null | ['spectral-graph-clustering'] | ['graphs'] | [-1.41254365e-01 -1.47807762e-01 3.57381195e-01 -9.28210001e-03
-1.98783781e-02 -7.94025898e-01 2.18777403e-01 -1.86823100e-01
-3.52173805e-01 6.61793768e-01 -1.75324958e-02 -5.35365641e-01
-3.54362100e-01 -6.19173169e-01 -4.00459826e-01 -9.50114548e-01
-3.96996617e-01 4.75159168e-01 1.16354339e-01 -2.90420324e-01
3.12176943e-01 8.78690064e-01 -1.12638366e+00 -8.66504461e-02
9.07642305e-01 6.16183758e-01 -3.78369391e-01 8.11308444e-01
-2.21124932e-01 1.68989450e-01 -2.61445820e-01 -1.57623693e-01
5.71648359e-01 -6.17184401e-01 -1.00987399e+00 6.24199688e-01
1.71389654e-01 3.70349079e-01 -1.78076178e-01 1.26647425e+00
1.80114210e-01 5.02974331e-01 1.01245117e+00 -1.13558304e+00
-2.93069631e-01 9.84387025e-02 -5.62579453e-01 3.67023051e-01
-3.64757106e-02 -7.97320455e-02 7.19791770e-01 -7.66033053e-01
7.51765072e-01 1.06768000e+00 6.70553982e-01 3.30990255e-01
-1.90468776e+00 -8.63528997e-02 -4.53459680e-01 -2.95473654e-02
-1.28854322e+00 -3.25250983e-01 1.08811629e+00 -8.09445143e-01
5.44316709e-01 3.85847300e-01 4.98160243e-01 3.66152674e-01
6.70716912e-02 5.89981079e-01 8.84168386e-01 -4.84627843e-01
3.14741731e-01 1.29705340e-01 5.28057456e-01 1.09546196e+00
3.77276421e-01 -1.49764940e-01 1.55422717e-01 -5.31242847e-01
6.40840709e-01 -2.97391176e-01 -6.34117484e-01 -7.24685133e-01
-7.11150289e-01 1.08465672e+00 5.15672445e-01 5.17833054e-01
-4.19464767e-01 -1.99921265e-01 1.07183576e-01 4.23499256e-01
7.02418327e-01 5.79503119e-01 -1.94320250e-02 2.59766310e-01
-6.99281156e-01 -1.97281495e-01 1.02101541e+00 4.23744887e-01
1.25108147e+00 1.48908764e-01 3.95116061e-01 8.53821576e-01
-6.32258877e-03 3.84811640e-01 2.61477351e-01 -9.54958677e-01
1.84615672e-01 7.67021239e-01 -1.95909925e-02 -1.11666441e+00
-7.17561483e-01 -2.53432423e-01 -9.54645574e-01 1.38866305e-01
5.86186647e-01 -4.09822315e-01 -8.48641753e-01 1.47979832e+00
2.91406542e-01 3.99813950e-01 6.43747896e-02 9.89376366e-01
3.11226964e-01 4.99095351e-01 -2.11890370e-01 -6.30925715e-01
5.16533613e-01 -6.16221130e-01 -4.64831412e-01 5.09826958e-01
6.80557251e-01 -7.40657926e-01 8.23652923e-01 4.22066808e-01
-9.84574378e-01 -2.82354832e-01 -7.32461810e-01 3.86487901e-01
-3.21691841e-01 1.45327032e-01 4.50962573e-01 7.31416702e-01
-1.19390929e+00 8.33789647e-01 -9.17624474e-01 -5.00512302e-01
-8.73091258e-03 3.03230315e-01 -3.82137716e-01 4.21049803e-01
-9.28075135e-01 4.63111579e-01 -7.09014535e-02 2.29599267e-01
-4.30775881e-01 -6.96780443e-01 -7.51296997e-01 5.54579534e-02
1.21612944e-01 -3.26579034e-01 6.46064937e-01 -1.10705769e+00
-1.20948350e+00 9.24664974e-01 -1.17500924e-01 -1.33385554e-01
5.23546517e-01 3.35196286e-01 -2.44868070e-01 2.12830722e-01
-1.29117161e-01 -2.50536591e-01 9.33182359e-01 -1.20561945e+00
1.28898725e-01 -4.50305760e-01 -2.36403212e-01 -8.61850604e-02
-3.18004161e-01 -2.89877027e-01 -2.93517057e-02 -1.99534714e-01
2.15565234e-01 -1.06919181e+00 -6.09967947e-01 -3.72778922e-01
-7.47100234e-01 -1.01095036e-01 8.31229925e-01 -5.15783072e-01
1.07430470e+00 -2.04171681e+00 7.06248879e-01 1.00573349e+00
4.64025825e-01 8.75106752e-02 4.71889675e-02 6.15348458e-01
-3.31782192e-01 2.12245673e-01 -7.25638211e-01 -1.39420211e-01
-2.41765380e-01 5.37290946e-02 -1.39354557e-01 1.04378819e+00
6.10220293e-03 5.10442853e-01 -6.62076831e-01 -2.95337915e-01
1.23311624e-01 3.15269917e-01 -5.14244020e-01 1.04143992e-01
6.38446361e-02 1.74020126e-01 -3.81242931e-01 1.96930021e-02
6.85759127e-01 -3.05163056e-01 1.95253432e-01 -2.79458523e-01
-1.45312011e-01 -3.48743588e-01 -1.19116771e+00 1.33677471e+00
-3.28924596e-01 8.54197443e-01 6.63173795e-01 -1.42869568e+00
7.38863170e-01 2.30798200e-02 1.01226473e+00 3.43169980e-02
2.57705778e-01 2.01206356e-01 4.31464128e-02 -3.88130754e-01
3.56517822e-01 -4.90755260e-01 1.29961103e-01 5.56718230e-01
-2.37797685e-02 1.05789647e-01 6.20633423e-01 5.73206484e-01
1.00785172e+00 -4.83431280e-01 1.01325177e-02 -8.61198485e-01
8.46898556e-01 2.87082583e-01 -4.98541147e-02 3.59894753e-01
-2.63435990e-02 5.19839585e-01 8.83270442e-01 -3.32580388e-01
-8.83945167e-01 -9.33807969e-01 -1.86804533e-01 4.43306118e-01
8.91007483e-02 -4.43878710e-01 -9.61210310e-01 -5.03187001e-01
2.30580911e-01 1.66047841e-01 -4.19635862e-01 -3.17435861e-01
-4.89115715e-01 -1.17671800e+00 3.24029088e-01 5.26276156e-02
3.00287992e-01 -8.79561424e-01 -3.24625492e-01 9.39778835e-02
1.14551783e-01 -7.40149617e-01 -5.06693065e-01 1.95533261e-02
-9.70845222e-01 -1.41789865e+00 -7.05903649e-01 -5.05185843e-01
8.02306592e-01 1.00558251e-01 9.87540960e-01 2.86412477e-01
-6.76658809e-01 7.65508771e-01 5.28637767e-02 4.12195385e-01
-5.10388613e-01 8.61942023e-02 1.09793641e-01 4.69484776e-01
5.88608831e-02 -7.53573895e-01 -3.92519146e-01 2.49825761e-01
-1.08959270e+00 -4.59791869e-01 2.35789746e-01 5.82742274e-01
3.64018559e-01 2.89198518e-01 2.64627934e-01 -1.14945412e+00
1.29136467e+00 -6.06641769e-01 -9.90580022e-01 1.67840868e-01
-5.79696059e-01 3.54819238e-01 9.01449621e-01 -1.63232386e-01
-7.20951200e-01 4.18620110e-01 1.33526176e-01 -4.75712121e-01
8.19366351e-02 6.13896608e-01 3.00277680e-01 -5.32631516e-01
9.11460638e-01 7.17493296e-02 2.26439655e-01 -6.47634804e-01
3.27867210e-01 6.64352417e-01 3.60620826e-01 -6.30687833e-01
6.55942678e-01 4.59061891e-01 6.20551765e-01 -1.35666144e+00
-2.45793179e-01 -9.00680065e-01 -5.77229738e-01 -5.45805275e-01
7.26847768e-01 -2.52225045e-02 -9.07365918e-01 2.53065467e-01
-9.64268386e-01 -2.66624093e-01 -1.35872751e-01 3.58810127e-01
-7.07567871e-01 5.74834287e-01 -6.94676697e-01 -9.25420463e-01
4.54677362e-03 -8.23711276e-01 6.34475768e-01 1.94896623e-01
2.15557039e-01 -1.48415625e+00 6.01610839e-01 -1.61800385e-01
2.30456173e-01 2.36638948e-01 1.09678054e+00 -3.53147656e-01
-2.29834244e-01 -4.17443477e-02 6.59198388e-02 4.14664835e-01
5.41036315e-02 3.38727623e-01 -5.11154652e-01 -2.24358261e-01
6.58880398e-02 7.18990564e-02 1.04985297e+00 5.48171699e-01
8.56253445e-01 -8.83149430e-02 -3.16257060e-01 7.67177522e-01
1.69331193e+00 -1.95668340e-01 3.83514285e-01 -3.40683997e-01
7.99976230e-01 8.44728351e-01 -2.14485481e-01 2.06892088e-01
-3.16078752e-01 3.07885289e-01 -1.26568452e-01 -1.55601755e-01
2.22665798e-02 1.46118820e-01 1.23387396e-01 8.08632374e-01
-2.55836815e-01 5.39791211e-02 -1.03049695e+00 6.75018609e-01
-1.91872132e+00 -9.96167123e-01 -7.28010833e-01 2.14944506e+00
4.55215067e-01 -1.69805482e-01 4.18277502e-01 2.34393880e-01
9.47387993e-01 6.31455779e-02 -3.36814791e-01 -5.95995009e-01
7.94192776e-02 2.07264096e-01 6.54835343e-01 1.23778939e+00
-8.61319005e-01 5.84783733e-01 7.26784801e+00 2.78101981e-01
-1.11314189e+00 -1.69313505e-01 2.73338079e-01 2.62090147e-01
-2.29896665e-01 -2.08132733e-02 -4.71465848e-02 2.60260522e-01
7.33252645e-01 -2.26150170e-01 6.92471027e-01 5.00985622e-01
2.13121455e-02 -2.20471293e-01 -6.68985248e-01 8.55832517e-01
-3.06357861e-01 -1.17724335e+00 -1.67419776e-01 2.71501452e-01
8.39412808e-01 -1.43111512e-01 -9.67645198e-02 -2.62694210e-01
2.03017801e-01 -7.82256126e-01 -1.31271705e-01 4.74764824e-01
5.07655203e-01 -8.24518859e-01 1.78514302e-01 2.49902263e-01
-1.02533150e+00 1.24761485e-01 -2.76810229e-01 9.85328294e-03
1.91098928e-01 9.31747615e-01 -6.85144186e-01 6.15171015e-01
1.62622035e-01 7.11570680e-01 -2.67239809e-01 1.21943343e+00
2.96332031e-01 7.74350047e-01 -6.29829705e-01 -6.81795180e-02
2.35532209e-01 -1.16455197e+00 8.81490767e-01 1.22720325e+00
2.43895710e-01 2.49359682e-01 1.68094292e-01 8.73484433e-01
-1.08219400e-01 3.74221772e-01 -7.11117208e-01 -2.54981220e-01
-2.80862898e-01 1.47459376e+00 -1.13523138e+00 -1.09110281e-01
-1.66584894e-01 8.52064669e-01 4.57970232e-01 9.34059918e-01
-1.71865776e-01 -6.74244285e-01 7.42119372e-01 2.21369028e-01
2.17727810e-01 -4.80749637e-01 -2.15854391e-01 -1.35513115e+00
-9.65711623e-02 -9.79716405e-02 4.78890151e-01 -4.64899629e-01
-1.38108814e+00 7.23432422e-01 1.03368685e-01 -5.92787862e-01
-3.49279344e-01 -1.00404298e+00 -8.39689910e-01 1.04370999e+00
-9.70184147e-01 -1.58150017e-01 -7.55710900e-02 8.59946191e-01
3.66967916e-02 -9.24312547e-02 2.73188978e-01 1.36552468e-01
-6.69475079e-01 -9.36757103e-02 4.36968565e-01 6.53743511e-03
2.01727360e-01 -1.44358861e+00 -3.65970582e-02 8.93250108e-01
3.19821566e-01 4.66852427e-01 9.78675485e-01 -6.22421324e-01
-1.17307436e+00 -4.20542628e-01 5.48338830e-01 -1.46412030e-01
8.04226398e-01 -3.05716634e-01 -9.96867120e-01 5.49722910e-01
7.51425326e-02 2.83489525e-01 4.33768749e-01 5.70437610e-02
9.00121555e-02 -1.02848113e-01 -1.19295740e+00 4.27757114e-01
8.90415609e-01 -4.87741441e-01 -1.57782272e-01 4.63243783e-01
3.05480864e-02 9.90718529e-02 -5.92097938e-01 2.25961626e-01
1.58463180e-01 -1.12142360e+00 8.14565778e-01 -1.11376071e+00
7.71872848e-02 -2.39686295e-01 1.97915941e-01 -1.73233855e+00
-2.79591650e-01 -9.39774513e-01 2.53238857e-01 6.26734257e-01
4.50421423e-01 -7.19664216e-01 9.30635989e-01 6.76674604e-01
1.64092541e-01 -5.00800908e-01 -8.21358085e-01 -6.18827641e-01
7.25198910e-02 -1.10897809e-01 -1.01091295e-01 9.53621268e-01
3.05970043e-01 4.89135563e-01 -9.25528333e-02 -3.36140729e-02
8.59235525e-01 2.34852552e-01 3.66502911e-01 -1.32522428e+00
-1.77858353e-01 -6.32746160e-01 -6.34689689e-01 -7.76333869e-01
6.25046015e-01 -1.15448570e+00 -2.32309550e-02 -1.18902922e+00
3.40618938e-02 -3.83709908e-01 -8.40452388e-02 -1.04263365e-01
-4.15401012e-02 1.76603496e-01 -2.05803402e-02 1.46254063e-01
-2.18217209e-01 2.85164088e-01 1.04874444e+00 2.22800016e-01
-5.06708622e-01 1.97589412e-01 -2.04268143e-01 4.56366986e-01
8.63606453e-01 -4.54155803e-01 -4.98619348e-01 1.30520567e-01
2.02783763e-01 1.36433825e-01 2.19908476e-01 -7.48457193e-01
2.08319768e-01 -1.59161717e-01 5.12890853e-02 7.39296526e-03
6.17519729e-02 -9.41319466e-01 1.46087825e-01 4.77815121e-01
-2.39342377e-01 1.55731276e-01 1.23902429e-02 5.72996378e-01
-1.59087062e-01 -5.38045585e-01 7.11136580e-01 -6.62837476e-02
-2.60034472e-01 1.03657447e-01 -6.33263409e-01 9.76893082e-02
7.71273375e-01 4.89437468e-02 -1.71387911e-01 -5.95183313e-01
-1.00224566e+00 9.59658399e-02 5.34206867e-01 -3.87497246e-01
5.17231941e-01 -1.28792799e+00 -5.22112131e-01 3.66527289e-01
-4.55739796e-01 -4.86118436e-01 -1.21555373e-01 1.25913787e+00
-7.53933251e-01 -7.99612049e-03 -2.05839947e-01 -5.09034157e-01
-1.12524545e+00 8.05077434e-01 7.94509888e-01 -1.10482961e-01
-4.20454383e-01 2.91532338e-01 2.09004954e-01 -2.60539293e-01
-1.66940898e-01 7.01767579e-02 3.35871801e-02 -2.74480833e-03
1.52489424e-01 7.98227787e-01 -4.79447236e-03 -7.31465518e-01
-3.65829587e-01 8.17359984e-01 4.20760095e-01 -4.16254491e-01
1.15164280e+00 -4.86462340e-02 -6.01111412e-01 4.27134991e-01
1.75456905e+00 2.24075988e-01 -9.72502053e-01 -3.90213728e-02
1.75516650e-01 -2.98426449e-01 1.40095819e-02 -2.42724404e-01
-1.45567179e+00 7.84644723e-01 1.48221761e-01 1.06218803e+00
1.28240788e+00 5.17505780e-02 5.11258662e-01 4.11603451e-01
-1.57153577e-01 -1.02804935e+00 -2.04165295e-01 3.12964797e-01
7.18101680e-01 -7.90729880e-01 -2.38317102e-02 -5.94257653e-01
-2.45568037e-01 1.43139958e+00 1.32010445e-01 -7.40379155e-01
1.22778988e+00 1.53055295e-01 1.34499341e-01 -4.21166480e-01
-4.60353434e-01 -4.91118699e-01 5.51397741e-01 3.29785734e-01
2.87748098e-01 1.30107090e-01 -7.47829318e-01 -1.17838562e-01
1.37870843e-02 -3.94729704e-01 7.71229446e-01 5.87262571e-01
-4.21403915e-01 -8.97636473e-01 -3.19488913e-01 4.38489228e-01
-1.59281433e-01 2.11801335e-01 -9.04226899e-01 5.38746476e-01
-4.16475236e-01 8.49809051e-01 1.58222150e-02 -3.12493950e-01
3.90991330e-01 3.86318505e-01 5.73680699e-01 -1.55586854e-01
-4.70036387e-01 -1.48085477e-02 9.20886621e-02 -4.68119293e-01
-6.10304713e-01 -5.33704519e-01 -1.26575935e+00 -4.33011979e-01
-2.09552824e-01 6.30833328e-01 6.06691599e-01 6.14650011e-01
4.66442294e-02 1.54094428e-01 7.98775971e-01 -6.99450552e-01
-4.65325028e-01 -7.22028196e-01 -9.93202746e-01 6.97795212e-01
6.18663967e-01 -4.66579199e-01 -8.79582107e-01 7.26337805e-02] | [7.040802955627441, 5.171727180480957] |
3209af84-c797-4b3e-9e9c-8060ce69e486 | search-based-repair-of-deep-neural-networks | 1912.12463 | null | https://arxiv.org/abs/1912.12463v2 | https://arxiv.org/pdf/1912.12463v2.pdf | Arachne: Search Based Repair of Deep Neural Networks | The rapid and widespread adoption of Deep Neural Networks (DNNs) has called for ways to test their behaviour, and many testing approaches have successfully revealed misbehaviour of DNNs. However, it is relatively unclear what one can do to correct such behaviour after revelation, as retraining involves costly data collection and does not guarantee to fix the underlying issue. This paper introduces Arachne, a novel program repair technique for DNNs, which directly repairs DNNs using their input-output pairs as a specification. Arachne localises neural weights on which it can generate effective patches and uses Differential Evolution to optimise the localised weights and correct the misbehaviour. An empirical study using different benchmarks shows that Arachne can fix specific misclassifications of a DNN without reducing general accuracy significantly. On average, patches generated by Arachne generalise to 61.3% of unseen misbehaviour, whereas those by a state-of-the-art DNN repair technique generalise only to 10.2% and sometimes to none while taking tens of times more than Arachne. We also show that Arachne can address fairness issues by debiasing a gender classification model. Finally, we successfully apply Arachne to a text sentiment model to show that it generalises beyond Convolutional Neural Networks. | ['Shin Yoo', 'Sungmin Kang', 'Jeongju Sohn'] | 2019-12-28 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [ 1.38914958e-01 4.65697616e-01 -5.58646545e-02 -4.60611165e-01
-1.55168874e-02 -8.40840638e-01 4.26378012e-01 3.48712923e-03
-5.13509810e-01 9.16172087e-01 -3.67736936e-01 -6.68462873e-01
-7.98155442e-02 -1.01437712e+00 -1.27252853e+00 -8.87573957e-01
1.24616116e-01 1.07378893e-01 3.84107530e-01 -2.30035782e-01
4.04007286e-01 3.78399670e-01 -1.74039555e+00 2.96962589e-01
7.60873377e-01 7.99590111e-01 -4.53322083e-01 7.25661337e-01
2.06931740e-01 8.68079662e-01 -1.30936885e+00 -1.03633225e+00
5.19610405e-01 -1.84180945e-01 -8.30935895e-01 -3.50689292e-01
8.58766556e-01 -5.68745673e-01 -1.34734526e-01 1.52470124e+00
3.74565691e-01 -9.47807655e-02 1.69140339e-01 -1.81376958e+00
-8.48242402e-01 1.13107717e+00 -4.18119043e-01 4.33902889e-01
-4.90700454e-01 2.36963719e-01 9.97093022e-01 1.87403131e-02
3.90507251e-01 1.28919017e+00 8.72554302e-01 1.10420597e+00
-1.37352371e+00 -1.05491257e+00 -2.84332156e-01 -3.19722801e-01
-1.07334578e+00 -3.36144567e-01 2.47271046e-01 -1.08791932e-01
1.23908770e+00 4.21826661e-01 4.15129155e-01 1.38106239e+00
6.43542051e-01 5.27101636e-01 7.73749232e-01 -1.03393339e-01
4.79952574e-01 1.78453643e-02 3.68259072e-01 6.12626791e-01
6.96932256e-01 4.71077561e-01 -1.88710406e-01 -5.52286446e-01
3.60028625e-01 -1.11882180e-01 -1.62066408e-02 2.55548004e-02
-4.38810229e-01 1.12104189e+00 4.35256779e-01 1.37193233e-01
-1.85677737e-01 6.12452090e-01 7.22894192e-01 4.43979651e-01
3.81704777e-01 7.89926827e-01 -6.81490242e-01 -3.27708513e-01
-8.53769422e-01 4.99626309e-01 9.94469881e-01 5.23105204e-01
6.18048429e-01 3.19454610e-01 -7.07083791e-02 8.95510554e-01
-2.94758826e-01 1.80562437e-01 5.00485718e-01 -1.34450614e+00
5.90122025e-03 7.56352127e-01 -5.02255261e-01 -1.07844734e+00
-2.02058226e-01 -4.53448534e-01 -7.99036264e-01 6.01365685e-01
2.99506664e-01 -6.22778475e-01 -8.77777934e-01 1.83203709e+00
-1.48398623e-01 1.51843682e-01 5.39698415e-02 5.14819443e-01
4.46431547e-01 4.07274127e-01 -1.47236377e-01 3.50099713e-01
8.48271191e-01 -7.55465448e-01 -2.12441161e-01 -3.10578525e-01
8.29836190e-01 -1.25738740e-01 6.38609827e-01 7.44548440e-01
-9.52453732e-01 -2.66391709e-02 -1.37341774e+00 3.83770227e-01
-3.21488827e-01 -3.16125661e-01 5.97691894e-01 1.35980046e+00
-1.54664493e+00 9.73447859e-01 -8.24340761e-01 -4.99215722e-02
8.28298867e-01 7.78661609e-01 -1.85921013e-01 9.05787051e-02
-1.11439717e+00 7.33667254e-01 5.27932703e-01 -1.35074586e-01
-1.15371621e+00 -8.34448934e-01 -5.47675133e-01 1.51889995e-01
2.30440214e-01 -3.32454532e-01 1.33993518e+00 -1.36469316e+00
-1.07109618e+00 7.29358673e-01 3.92603368e-01 -1.05464423e+00
5.28735518e-01 1.98216259e-01 -3.11653376e-01 -3.58770341e-01
-5.65217361e-02 1.01607394e+00 6.92126215e-01 -1.07631004e+00
-8.64156783e-01 -1.80742517e-01 4.41250592e-01 -4.27988023e-01
-6.43522978e-01 1.26080841e-01 1.81257099e-01 -6.64077580e-01
-4.58347470e-01 -9.74866748e-01 -2.09959578e-02 -6.55878857e-02
-4.76987749e-01 -2.90535122e-01 7.96537280e-01 -3.82166266e-01
1.02463555e+00 -2.14795995e+00 -1.07847050e-01 3.31026793e-01
4.83210772e-01 6.06203735e-01 -2.74424732e-01 -3.25153381e-01
-3.53548646e-01 5.84499896e-01 -5.10747731e-01 -1.26620933e-01
1.57547742e-01 6.78741992e-01 -3.30404103e-01 5.16675472e-01
2.62573123e-01 7.66339898e-01 -6.23446167e-01 3.26997131e-01
-4.42654759e-01 3.88553649e-01 -1.00729382e+00 -3.02129060e-01
-2.97824651e-01 -5.28997540e-01 2.36674741e-01 6.88688636e-01
8.61357808e-01 1.08862586e-01 -1.10562313e-02 3.75415534e-01
1.80545390e-01 2.04821333e-01 -6.73865318e-01 6.19428158e-01
-4.59450968e-02 9.78821576e-01 7.57519528e-02 -1.12454283e+00
1.12733078e+00 -1.96665585e-01 -1.13991857e-01 -8.36529016e-01
3.49178791e-01 2.77523071e-01 4.19161737e-01 -1.47071525e-01
6.72337055e-01 1.35145843e-01 -2.19832063e-01 6.98644638e-01
1.33936718e-01 1.83707222e-01 -1.00798540e-01 -1.43375155e-03
1.84538829e+00 -6.77672863e-01 -1.60296440e-01 -4.52591985e-01
6.63015991e-02 -4.20216210e-02 7.71533132e-01 1.19175744e+00
-3.67422521e-01 5.09537637e-01 1.26572394e+00 -7.26271152e-01
-1.15372694e+00 -6.02429986e-01 -1.47527844e-01 1.24779880e+00
-3.10583174e-01 -1.49170801e-01 -1.38380384e+00 -1.07501233e+00
2.64499754e-01 1.11002445e+00 -1.09765553e+00 -7.99635053e-01
-1.59439117e-01 -9.67385232e-01 1.47257006e+00 4.03014898e-01
8.32099974e-01 -1.31563902e+00 -8.50737870e-01 4.89637144e-02
3.37581366e-01 -6.19900525e-01 -2.82821544e-02 5.52323461e-01
-8.53758335e-01 -9.68379855e-01 -4.26423818e-01 -4.81267482e-01
7.33069599e-01 -9.55059230e-02 1.11057043e+00 8.64022076e-01
-1.95558041e-01 -4.72867750e-02 -2.33974561e-01 -4.17965978e-01
-1.00912774e+00 2.69364536e-01 2.42162813e-02 -3.25976133e-01
6.19585216e-01 -5.22458136e-01 -3.79553325e-02 4.29628789e-01
-1.18200624e+00 -6.25407755e-01 4.85757142e-01 9.39307868e-01
-2.55706489e-01 4.85791236e-01 4.09763277e-01 -1.23109567e+00
1.02834201e+00 -5.55143535e-01 -6.93350434e-01 1.75886869e-01
-9.07982111e-01 1.47048727e-01 3.76457781e-01 -6.21728003e-01
-5.47394216e-01 -2.94482768e-01 -2.43089776e-02 -5.42196751e-01
-2.36098662e-01 2.44217187e-01 2.30864845e-02 -2.75611967e-01
1.24614501e+00 -1.63751841e-01 2.16811195e-01 1.29814185e-02
-1.31717384e-01 5.15771091e-01 7.03409970e-01 -5.87574959e-01
3.97859186e-01 7.42646605e-02 -3.12062711e-01 -3.17332238e-01
-4.61614996e-01 4.60077524e-01 -1.05010621e-01 -3.45610678e-02
4.48752284e-01 -4.27351594e-01 -9.91145313e-01 8.70019853e-01
-1.32798612e+00 -7.34107971e-01 -3.68458591e-02 -4.33972597e-01
4.35642228e-02 -7.85248503e-02 -6.22902930e-01 -5.40996134e-01
-2.91585416e-01 -1.27771282e+00 4.73705143e-01 1.54649898e-01
-4.22259092e-01 -7.57332027e-01 -8.69442299e-02 2.14875847e-01
5.22281110e-01 7.34500811e-02 1.12253535e+00 -1.09214723e+00
-2.07445338e-01 -1.73410922e-01 -2.70784765e-01 8.74615788e-01
-1.63540870e-01 4.43556577e-01 -1.04390526e+00 -2.09308177e-01
-5.97432554e-02 -2.57475764e-01 8.70110333e-01 3.71360242e-01
1.43138850e+00 -8.59074295e-01 -2.38729760e-01 5.01341701e-01
1.03499126e+00 3.85999203e-01 9.66077268e-01 7.70607054e-01
6.16492271e-01 5.35639763e-01 5.23943119e-02 3.70604187e-01
6.57802448e-02 2.30269924e-01 1.11143017e+00 1.91055745e-01
3.46023053e-01 2.20037326e-01 5.24982154e-01 -1.03856869e-01
1.48888528e-01 -4.14259076e-01 -1.04288661e+00 6.09089911e-01
-1.78164256e+00 -9.47692096e-01 -1.64546072e-01 1.89714313e+00
1.08853900e+00 5.15454054e-01 9.68553051e-02 2.81555206e-01
9.44934666e-01 -8.92255381e-02 -8.23396385e-01 -1.15999091e+00
-2.02173352e-01 3.20177138e-01 1.09916377e+00 3.02736223e-01
-8.91726553e-01 9.31079984e-01 6.40663719e+00 7.59309590e-01
-1.16820192e+00 7.57711008e-02 1.03608716e+00 -2.71028638e-01
-3.71841699e-01 -9.37357694e-02 -9.20443833e-01 5.81687033e-01
1.20698917e+00 -1.57600924e-01 7.37924278e-01 1.06100261e+00
-2.47374609e-01 -1.91323906e-01 -1.08285606e+00 6.48661256e-01
2.47947454e-01 -1.56182086e+00 -8.78904909e-02 2.76123881e-01
9.55483496e-01 1.16569191e-01 1.98580801e-01 4.31624204e-01
9.16807353e-01 -1.28658795e+00 9.56733227e-01 2.98911899e-01
9.12579969e-02 -1.46751225e+00 1.06617498e+00 4.86367464e-01
6.88330978e-02 -5.01200914e-01 -6.97656274e-01 -1.90837443e-01
-5.02702475e-01 6.79994881e-01 -1.03302395e+00 -3.83974463e-01
1.24179721e+00 3.29239935e-01 -1.10238051e+00 1.07285094e+00
-2.60908246e-01 8.35293233e-01 -2.27523357e-01 -3.48619014e-01
2.48083413e-01 3.72693688e-01 4.55130041e-01 9.90274847e-01
2.35882580e-01 -2.21696869e-01 -6.16578698e-01 1.23716223e+00
-5.47535479e-01 -6.66626155e-01 -6.59766734e-01 -8.48210976e-02
5.29146969e-01 9.82161820e-01 -7.37927854e-01 -1.32572830e-01
2.16561109e-01 9.70002055e-01 3.99555027e-01 2.09708326e-02
-9.64399993e-01 -5.10306358e-01 1.01016891e+00 -5.54208346e-02
3.78527492e-01 3.03233534e-01 -6.43105924e-01 -5.62850296e-01
-1.11208625e-01 -1.35793424e+00 2.32755899e-01 -7.04145968e-01
-1.21530223e+00 6.46153569e-01 -1.18435510e-01 -5.42527258e-01
-1.50443539e-01 -6.48728967e-01 -7.27603436e-01 4.93362904e-01
-9.28770721e-01 -4.03305411e-01 -1.66199356e-02 2.56291687e-01
1.24050342e-01 -4.59105462e-01 7.25602210e-01 2.41787523e-01
-8.49750280e-01 1.20539498e+00 -1.95065051e-01 3.25636595e-01
5.75336218e-01 -1.08652592e+00 7.76133120e-01 9.49926615e-01
-2.41191089e-01 7.43032634e-01 8.33098769e-01 -7.12327540e-01
-7.71219432e-01 -1.25930500e+00 8.13433111e-01 -3.77019525e-01
6.45963132e-01 -4.70204741e-01 -1.19407725e+00 6.55416071e-01
4.16193366e-01 8.72318372e-02 2.81119138e-01 -1.61504254e-01
-5.70024967e-01 -2.11565480e-01 -1.49909663e+00 7.72205949e-01
9.63607669e-01 -3.38331938e-01 -3.42618525e-01 -7.56736379e-03
8.26121747e-01 -4.50666934e-01 -3.00686032e-01 2.48402148e-01
2.50933856e-01 -1.46757889e+00 5.13608277e-01 -5.64765096e-01
9.31913733e-01 1.39689267e-01 -5.98231070e-02 -1.39578712e+00
2.85968278e-02 -5.93992710e-01 1.95332572e-01 1.50646794e+00
7.25355387e-01 -9.64092255e-01 1.15761733e+00 1.13426709e+00
-1.45905674e-01 -4.02117163e-01 -9.75895643e-01 -7.42451608e-01
3.95732343e-01 -5.27081370e-01 8.21379125e-01 1.25931537e+00
-3.78113836e-01 -2.99874783e-01 -2.38590822e-01 1.08007640e-01
5.20361185e-01 -7.29320884e-01 6.09335721e-01 -1.17822444e+00
-3.09118867e-01 -9.15817916e-01 -6.96723461e-01 -3.44283700e-01
6.99432731e-01 -9.00619566e-01 1.67339042e-01 -6.37483060e-01
1.57447830e-01 -3.28130364e-01 -2.89009549e-02 1.36753154e+00
2.36856341e-01 4.72473592e-01 -1.99083328e-01 -2.99364775e-01
-2.18753770e-01 2.15026662e-02 4.34603184e-01 -4.38696861e-01
-6.97170570e-02 1.02117501e-01 -1.22736299e+00 5.45195699e-01
1.01303530e+00 -1.11966729e+00 -2.18528703e-01 -4.57373410e-01
7.41715550e-01 -4.58299190e-01 1.08048177e+00 -9.41210806e-01
3.71072471e-01 3.47523354e-02 8.48985314e-02 -2.38961756e-01
-3.70828986e-01 -6.96858823e-01 1.80703998e-01 5.66398323e-01
-5.67679703e-01 3.32072079e-01 6.25376940e-01 3.65562320e-01
1.49274975e-01 -8.23508382e-01 7.80966580e-01 1.96977593e-02
-5.58612108e-01 1.60466880e-02 -7.01173782e-01 2.01997720e-02
1.02917457e+00 -2.79381871e-01 -8.39687884e-01 -1.89934582e-01
-2.15838179e-01 5.52258641e-02 6.16633058e-01 5.63993633e-01
4.47260290e-01 -1.02431881e+00 -3.68386328e-01 4.75562871e-01
-2.58716136e-01 -9.32245031e-02 -5.28823547e-02 3.80758047e-01
-6.31553113e-01 -7.83060342e-02 -5.31144977e-01 -6.04873240e-01
-1.51983404e+00 2.03280210e-01 8.63230050e-01 3.61863643e-01
-3.21344674e-01 1.29784977e+00 -2.24400789e-01 -7.71958649e-01
3.93311858e-01 -3.76045197e-01 2.09792748e-01 4.19569295e-03
4.45322007e-01 4.44381684e-01 5.18415332e-01 -1.68417633e-01
-3.42406541e-01 -2.51052856e-01 -3.11767191e-01 -6.06794171e-02
1.63019073e+00 4.67444301e-01 -6.54850841e-01 -2.43070647e-01
1.20811427e+00 -3.63304824e-01 -1.46246707e+00 2.56090075e-01
1.89722683e-02 -2.99392819e-01 4.51334804e-01 -1.03818607e+00
-1.68661654e+00 7.60775149e-01 4.36936706e-01 5.54603040e-01
9.63671148e-01 -3.11849415e-01 5.99218071e-01 6.22861087e-01
2.28080049e-01 -8.74046266e-01 -9.09290835e-02 7.73522973e-01
6.46181762e-01 -8.36661398e-01 -3.26770544e-01 3.09584856e-01
-5.34131587e-01 1.12042022e+00 9.26137507e-01 -3.44609827e-01
5.82300484e-01 6.83087468e-01 -1.18666358e-01 -1.68711066e-01
-9.94371414e-01 7.92059660e-01 -3.65848064e-01 6.34406745e-01
-2.01868936e-01 -3.24812345e-02 1.69109255e-02 8.70070636e-01
-6.06580853e-01 -4.21693847e-02 1.01360881e+00 8.75756979e-01
-2.49647930e-01 -9.24186051e-01 -3.76141548e-01 5.86888134e-01
-8.22080255e-01 -1.71921775e-01 -7.38373876e-01 7.33968139e-01
3.96975517e-01 8.27698410e-01 3.80455762e-01 -8.04590821e-01
2.48768196e-01 -1.17868654e-01 1.41328856e-01 -3.83052200e-01
-1.19629681e+00 -4.90344495e-01 9.73731279e-02 -5.21670878e-01
-1.01530835e-01 -6.98984087e-01 -1.19334936e+00 -1.09623599e+00
-2.83864498e-01 -1.51790828e-01 5.77861905e-01 8.40844035e-01
4.07060623e-01 5.62787175e-01 3.87421995e-01 -5.93315959e-01
-8.29973519e-01 -5.40475011e-01 -5.46011269e-01 1.87010560e-02
4.80834872e-01 -4.86365527e-01 -8.49050760e-01 -2.84077823e-01] | [6.4157185554504395, 7.702718734741211] |
979915ae-2e40-4199-aaa4-6b07fa901722 | disir-deep-image-segmentation-with | 2003.14200 | null | https://arxiv.org/abs/2003.14200v2 | https://arxiv.org/pdf/2003.14200v2.pdf | DISIR: Deep Image Segmentation with Interactive Refinement | This paper presents an interactive approach for multi-class segmentation of aerial images. Precisely, it is based on a deep neural network which exploits both RGB images and annotations. Starting from an initial output based on the image only, our network then interactively refines this segmentation map using a concatenation of the image and user annotations. Importantly, user annotations modify the inputs of the network - not its weights - enabling a fast and smooth process. Through experiments on two public aerial datasets, we show that user annotations are extremely rewarding: each click corrects roughly 5000 pixels. We analyze the impact of different aspects of our framework such as the representation of the annotations, the volume of training data or the network architecture. Code is available at https://github.com/delair-ai/DISIR. | ['Adrien Chan Hon Tong', 'Nicola Luminari', 'Bertrand Le Saux', 'Gaston Lenczner', 'Guy Le Besnerais'] | 2020-03-31 | null | null | null | null | ['remote-sensing-image-classification'] | ['miscellaneous'] | [ 2.92609364e-01 2.44969204e-01 1.68786421e-01 -5.20106316e-01
-3.51394594e-01 -1.12558806e+00 1.35275558e-01 -5.76662133e-03
-7.87441730e-01 4.89331692e-01 -2.09441736e-01 -2.80524194e-01
1.86387971e-01 -7.70397723e-01 -8.89308751e-01 -4.72348690e-01
4.02670987e-02 1.97523728e-01 6.22640729e-01 -1.89487904e-01
-5.29413559e-02 4.69894409e-01 -1.47253442e+00 3.19331735e-01
6.45472169e-01 1.05909705e+00 1.96599379e-01 1.08830297e+00
2.69350857e-01 4.60162640e-01 -6.60717368e-01 -4.43230152e-01
7.36868322e-01 1.57468691e-02 -1.05218673e+00 1.60420716e-01
4.15633082e-01 -6.19709969e-01 -1.54230624e-01 1.15738559e+00
2.21906543e-01 2.40676329e-01 2.81356484e-01 -1.16037107e+00
-4.61158752e-01 1.04131639e+00 -4.98253435e-01 2.75685620e-02
-2.23996583e-02 5.00480771e-01 1.10027707e+00 -6.56100512e-01
6.07479095e-01 6.67879403e-01 8.02909315e-01 4.07650739e-01
-1.20764482e+00 -4.73545313e-01 3.28849822e-01 -1.87212780e-01
-1.37243545e+00 -1.89357251e-01 4.17119086e-01 -5.94632149e-01
6.02353394e-01 5.07798016e-01 9.99815941e-01 6.90728724e-01
-2.40669861e-01 7.37362623e-01 8.34789574e-01 -2.02182114e-01
6.72698244e-02 1.49044856e-01 8.23993981e-02 8.48602831e-01
-4.05047797e-02 -5.89951463e-02 -6.34231120e-02 1.04421377e-02
9.59551573e-01 7.38800690e-02 -2.90355593e-01 -4.10667419e-01
-9.89643931e-01 5.89524329e-01 9.14431810e-01 9.60055515e-02
-3.42991084e-01 4.58535731e-01 8.90875533e-02 -3.82945836e-02
2.32529715e-01 6.70367062e-01 -7.85718322e-01 2.40054592e-01
-1.15650916e+00 3.13862786e-02 5.75076520e-01 7.39419162e-01
9.44907665e-01 -3.12781453e-01 -2.20540538e-01 3.69936883e-01
1.77531391e-01 2.83224016e-01 3.66485119e-02 -1.22367930e+00
3.18514332e-02 6.71771109e-01 4.60682184e-01 -6.67347968e-01
-5.79187334e-01 -3.03766936e-01 -4.08784181e-01 6.02886558e-01
7.14253962e-01 -6.56647384e-01 -1.24629045e+00 1.48180783e+00
3.45259160e-01 -2.40761220e-01 -3.52921903e-01 1.13339078e+00
8.27370405e-01 3.47693175e-01 2.48932794e-01 5.61253726e-01
1.10525119e+00 -1.11908424e+00 -2.15465724e-01 -3.70374531e-01
4.01947677e-01 -4.70084876e-01 1.26739931e+00 3.58213365e-01
-1.20539021e+00 -5.60979962e-01 -8.85316789e-01 -1.09599382e-02
-6.99750543e-01 9.07604516e-01 4.41483378e-01 4.48413998e-01
-1.27557898e+00 9.43042397e-01 -8.76135111e-01 -3.15678239e-01
7.07686543e-01 5.91147184e-01 -1.51096463e-01 2.96092987e-01
-9.19566929e-01 4.69381481e-01 6.88136101e-01 3.06682646e-01
-9.53888655e-01 -4.86597121e-01 -5.72522044e-01 1.12388149e-01
5.32577991e-01 -5.62429667e-01 1.41901934e+00 -1.44266248e+00
-1.44577587e+00 8.31540346e-01 3.12942952e-01 -3.86857510e-01
7.76419103e-01 -6.09441161e-01 1.87638685e-01 1.35178521e-01
-1.28186911e-01 1.35098767e+00 7.70332038e-01 -1.22283912e+00
-8.58359933e-01 -1.95819065e-01 6.15465879e-01 1.67531177e-01
-1.94306627e-01 -2.26926059e-01 -8.22614908e-01 -3.09462011e-01
-1.85285583e-01 -1.07599914e+00 -6.08911633e-01 2.87839800e-01
-7.78632045e-01 2.08702371e-01 5.45908332e-01 -6.80313170e-01
9.01034355e-01 -2.33734989e+00 1.80957392e-01 3.79277974e-01
2.41394132e-01 8.24773312e-02 -2.84616232e-01 -3.18455696e-02
-2.73585290e-01 5.61627150e-01 -5.35482347e-01 -2.57483959e-01
-5.00858836e-02 -9.66069326e-02 -1.65715143e-01 3.62968892e-01
2.25121006e-01 9.32490110e-01 -6.67099833e-01 -1.82554752e-01
3.26376975e-01 4.84456956e-01 -5.72415769e-01 2.71559566e-01
-5.12545228e-01 4.63957906e-01 -1.76093176e-01 6.80126309e-01
5.41960120e-01 -4.20049608e-01 1.34922862e-01 -4.07330364e-01
-3.28402400e-01 -1.39138877e-01 -1.12124813e+00 1.84946537e+00
-2.34959126e-02 6.67477608e-01 9.59036201e-02 -4.94811505e-01
4.47798103e-01 -3.04469801e-02 5.07088065e-01 -1.13588080e-01
4.11854446e-01 -1.37790307e-01 -2.52468109e-01 -2.47288823e-01
7.46507525e-01 5.80977380e-01 -2.43367220e-04 4.36655879e-01
1.52928635e-01 -3.36570650e-01 4.63202208e-01 3.03895473e-01
9.37649250e-01 5.61086357e-01 -8.17467831e-03 -1.09825879e-01
4.07281667e-02 3.23276043e-01 1.56611681e-01 7.30043828e-01
-1.27241910e-01 6.80183709e-01 6.40591204e-01 -5.70652366e-01
-9.42198396e-01 -7.87896037e-01 -1.63028855e-02 1.32634032e+00
2.74414957e-01 -5.30179441e-01 -1.06702161e+00 -9.68296409e-01
-5.98976947e-02 5.67589998e-01 -1.05000710e+00 1.68696076e-01
-1.07760288e-01 -7.79752970e-01 5.74615180e-01 4.23457742e-01
5.42644083e-01 -1.10798883e+00 -1.23623323e+00 -2.21561670e-01
-1.19244717e-01 -1.01948678e+00 -2.52623707e-01 5.67533791e-01
-8.15996110e-01 -1.17200541e+00 -5.74546695e-01 -3.08952093e-01
9.72290218e-01 3.19007449e-02 1.09305680e+00 4.31285173e-01
-5.10926068e-01 4.88492280e-01 -4.06075209e-01 -4.41106468e-01
-3.87449861e-02 5.26958823e-01 -5.13703763e-01 -1.95327178e-01
-2.95086484e-02 -2.97872275e-01 -7.90601194e-01 2.58707702e-01
-1.17797029e+00 3.55256021e-01 4.36694533e-01 4.05080527e-01
7.00104952e-01 1.93890221e-02 -2.96108902e-01 -9.60320354e-01
3.06024611e-01 -2.33586848e-01 -1.02788687e+00 2.49682173e-01
-1.28900871e-01 -8.23365375e-02 2.18928829e-01 -1.78029135e-01
-5.70455134e-01 8.77952754e-01 -2.54242837e-01 -2.69068241e-01
-6.20531261e-01 3.31196398e-01 2.68488944e-01 -2.00847656e-01
8.30173373e-01 -3.37696403e-01 -4.32802200e-01 -1.58141509e-01
7.00853407e-01 2.69131601e-01 6.11692965e-01 -1.31474376e-01
9.08239067e-01 4.52174962e-01 -4.89246279e-01 -3.75067025e-01
-1.04881012e+00 -1.99158877e-01 -1.13766706e+00 -5.48379481e-01
1.20127380e+00 -7.51798093e-01 -7.20292211e-01 4.36652035e-01
-1.13459575e+00 -1.20464683e+00 -5.85639596e-01 2.00019255e-01
-3.17356914e-01 -1.20890081e-01 -4.67484444e-01 -5.78783929e-01
-2.30242237e-01 -1.36474907e+00 9.80045617e-01 5.64350605e-01
-1.93269223e-01 -7.58795857e-01 -2.02982873e-01 7.18348920e-02
3.33901614e-01 4.55883592e-01 3.64689559e-01 -5.85495055e-01
-8.16075206e-01 -2.82566667e-01 -4.12271082e-01 2.69081801e-01
-4.31635194e-02 4.99660969e-01 -1.22153330e+00 -8.96393787e-03
-6.01477206e-01 -5.44833004e-01 9.59294736e-01 5.49507260e-01
1.57992816e+00 -2.25451559e-01 -2.37290248e-01 8.26545358e-01
1.54919493e+00 -2.38185063e-01 5.37395000e-01 4.83746171e-01
7.83194125e-01 4.95439887e-01 5.83667278e-01 6.27847373e-01
2.32927233e-01 2.58904338e-01 1.05346477e+00 -6.48885071e-01
2.89597601e-01 4.47355881e-02 1.22351505e-01 -2.10500881e-01
-5.19807220e-01 -3.75580728e-01 -9.35491025e-01 3.19427580e-01
-1.83891559e+00 -6.41265631e-01 -5.38763888e-02 2.09080648e+00
9.77307379e-01 4.76184897e-02 2.28694171e-01 -9.98058170e-02
4.91355538e-01 1.24303959e-01 -7.06898868e-01 -9.95210558e-02
5.26810437e-02 1.71765655e-01 1.16967130e+00 5.53536832e-01
-1.49901497e+00 1.27195597e+00 6.58087683e+00 3.68831038e-01
-1.10796142e+00 -3.36090662e-02 7.57929444e-01 -2.68493235e-01
-2.70290263e-02 -1.66031465e-01 -6.24895990e-01 2.52278626e-01
5.40597141e-01 3.08366746e-01 7.20470488e-01 9.69891250e-01
4.07149196e-02 -5.68821609e-01 -8.19947779e-01 5.49639285e-01
-1.76707819e-01 -1.27868462e+00 -2.07587242e-01 -1.37632474e-01
6.67283237e-01 4.43286061e-01 7.04086870e-02 -5.39846309e-02
9.08730686e-01 -9.46081042e-01 1.05683947e+00 6.07401192e-01
8.93434823e-01 -5.54628968e-01 7.60857522e-01 3.65700498e-02
-8.97696495e-01 -1.04344718e-01 -1.79561153e-01 6.64907992e-02
-5.83918840e-02 4.06530797e-01 -7.02281415e-01 1.74472570e-01
1.21344793e+00 5.27773201e-01 -1.08312428e+00 1.14619517e+00
-5.73565066e-01 5.39451718e-01 -4.65970337e-01 3.01464319e-01
3.07891846e-01 -1.54271469e-01 2.42375910e-01 1.29605305e+00
2.45746374e-01 2.60416090e-01 3.43979031e-01 9.51968670e-01
-1.51452139e-01 -5.97128533e-02 -4.56163734e-01 -3.68936136e-02
2.19349086e-01 1.71316123e+00 -1.22231913e+00 -3.24252903e-01
-6.33450225e-02 1.31458676e+00 3.27609807e-01 5.68800390e-01
-1.08416104e+00 -4.10655707e-01 4.14895415e-01 -8.26138724e-03
5.60935557e-01 -3.12273949e-01 -4.55190420e-01 -7.90459692e-01
-3.31555963e-01 -6.14773870e-01 3.66390377e-01 -1.13721538e+00
-5.24767160e-01 7.30220079e-01 -1.45786360e-01 -7.20938921e-01
1.98094621e-01 -6.42759144e-01 -5.27640760e-01 7.42031634e-01
-1.34420967e+00 -1.11492479e+00 -8.61472130e-01 3.95887852e-01
4.03386742e-01 3.29721630e-01 7.85059929e-01 1.47399098e-01
-6.25902236e-01 2.41122589e-01 -3.37075770e-01 4.02693272e-01
5.13035655e-01 -1.45221543e+00 6.25407577e-01 9.38538313e-01
2.66080171e-01 1.30135208e-01 6.34889483e-01 -4.80506480e-01
-7.02089846e-01 -1.20619917e+00 1.58132449e-01 -6.94797099e-01
7.21048236e-01 -2.41127178e-01 -5.94100595e-01 9.22945917e-01
4.95519578e-01 -1.25469975e-02 5.46532631e-01 -2.09434956e-01
1.13634251e-01 1.56936124e-01 -9.15730178e-01 5.33854783e-01
8.72166872e-01 -1.63378954e-01 2.09360734e-01 5.20047843e-01
7.55942762e-01 -9.26852882e-01 -6.73573196e-01 3.68771225e-01
5.66317856e-01 -1.26030028e+00 8.46010745e-01 -3.95484328e-01
5.33378839e-01 -5.41262448e-01 -6.96135759e-02 -1.29284370e+00
-4.20281112e-01 -2.43208393e-01 2.90807784e-01 7.90034235e-01
8.38346720e-01 -2.68012017e-01 6.63521588e-01 9.61159587e-01
3.73219065e-02 -7.47530937e-01 -2.59492874e-01 -1.79959044e-01
-3.96632820e-01 -3.43285233e-01 6.36949420e-01 6.98471606e-01
-5.76225460e-01 -2.13263966e-02 -2.30535924e-01 5.92226744e-01
3.02323222e-01 -7.11277202e-02 8.48975003e-01 -1.20937514e+00
-3.58737707e-01 -3.45410496e-01 1.09350450e-01 -8.55621874e-01
-2.29819834e-01 -5.88613629e-01 2.76974559e-01 -1.62349176e+00
-1.11438401e-01 -5.17943621e-01 -4.03454527e-02 1.03439355e+00
-2.06843138e-01 7.92268336e-01 3.84671986e-01 1.90811604e-01
-7.74348497e-01 9.60527360e-02 1.00203049e+00 -1.15499035e-01
-5.72352767e-01 6.83621168e-02 -4.29035693e-01 1.02140093e+00
1.24445319e+00 -4.32866782e-01 -1.78148337e-02 -8.88908744e-01
3.67468566e-01 -4.06503439e-01 7.42325783e-01 -1.08633327e+00
2.51266569e-01 -1.18480809e-01 7.66825020e-01 -5.23575306e-01
2.78004944e-01 -9.81280267e-01 1.87212765e-01 3.91627699e-01
-5.42018354e-01 -3.59728262e-02 5.57577252e-01 2.93832839e-01
2.38776535e-01 -5.00985563e-01 6.97766960e-01 -4.26109076e-01
-7.50166297e-01 3.40144187e-01 -2.59172082e-01 -3.01626772e-01
1.03267205e+00 -1.66547775e-01 -1.89944893e-01 -9.49245095e-02
-1.19568110e+00 3.31776977e-01 7.63871789e-01 1.73076689e-02
1.97695985e-01 -8.28991234e-01 -3.05560708e-01 1.90504730e-01
-1.43571243e-01 3.85433048e-01 7.21398965e-02 7.40729213e-01
-7.52241611e-01 -8.80151093e-02 -3.42443407e-01 -6.05419517e-01
-1.36376178e+00 1.50219038e-01 6.74647748e-01 -7.77699053e-02
-3.92371327e-01 1.11445963e+00 4.68498021e-02 -5.28798342e-01
4.94794607e-01 -5.95146239e-01 -3.06264728e-01 1.30770892e-01
2.33443007e-01 1.01778150e-01 -1.44459099e-01 -2.20626920e-01
-1.94053531e-01 3.84038746e-01 2.41774112e-01 -3.36030692e-01
1.45482945e+00 -1.37745842e-01 -3.07493564e-02 4.63983864e-01
4.40597475e-01 -1.40478134e-01 -1.73036528e+00 1.35041952e-01
-3.05483222e-01 -4.08559859e-01 1.24206543e-01 -1.17154765e+00
-1.33843124e+00 7.39718854e-01 6.92146480e-01 3.94185334e-01
1.20769680e+00 -8.79689753e-02 3.93781334e-01 5.38133085e-01
9.90191326e-02 -1.20695662e+00 8.94535035e-02 3.85348350e-01
7.84986734e-01 -1.32435334e+00 7.86224902e-02 -3.34729463e-01
-7.52029181e-01 1.25692785e+00 8.25221062e-01 -1.20012179e-01
4.94292855e-01 3.15964103e-01 2.55313158e-01 -2.78871387e-01
-2.80481160e-01 -4.94832724e-01 1.52812168e-01 3.50574046e-01
3.76808017e-01 1.08038085e-02 3.49406511e-01 5.46444058e-01
-2.87274748e-01 1.97798610e-01 5.11544824e-01 8.05855811e-01
-5.87118566e-01 -6.59291387e-01 -2.69055068e-01 4.13344234e-01
-4.56917644e-01 -2.27123588e-01 -7.68661320e-01 6.71220839e-01
3.37677717e-01 5.28834760e-01 5.44842742e-02 -4.71438706e-01
3.10244888e-01 -1.23168670e-01 2.04620659e-01 -7.27899194e-01
-8.07032168e-01 -1.04924142e-01 -9.88118947e-02 -7.80791879e-01
-5.68662643e-01 -4.26681161e-01 -1.23647070e+00 -1.03749126e-01
-3.67865562e-01 1.39396247e-02 7.43138969e-01 6.86630726e-01
2.08781734e-01 7.93909609e-01 2.94786543e-01 -1.26764822e+00
-2.20511928e-02 -7.74456978e-01 -3.59496444e-01 3.60091589e-02
7.36960694e-02 -2.41573632e-01 -2.69196033e-01 5.41745603e-01] | [9.455677032470703, 0.07851049304008484] |
6504c2b4-55e7-4f36-87ea-6fb29b4617af | graph-partitioning-and-graph-neural-network | 2005.08008 | null | https://arxiv.org/abs/2005.08008v3 | https://arxiv.org/pdf/2005.08008v3.pdf | Graph Partitioning and Graph Neural Network based Hierarchical Graph Matching for Graph Similarity Computation | Graph similarity computation aims to predict a similarity score between one pair of graphs to facilitate downstream applications, such as finding the most similar chemical compounds similar to a query compound or Fewshot 3D Action Recognition. Recently, some graph similarity computation models based on neural networks have been proposed, which are either based on graph-level interaction or node-level comparison. However, when the number of nodes in the graph increases, it will inevitably bring about reduced representation ability or high computation cost. Motivated by this observation, we propose a graph partitioning and graph neural network-based model, called PSimGNN, to effectively resolve this issue. Specifically, each of the input graphs is partitioned into a set of subgraphs to extract the local structural features directly. Next, a novel graph neural network with an attention mechanism is designed to map each subgraph into an embedding vector. Some of these subgraph pairs are automatically selected for node-level comparison to supplement the subgraph-level embedding with fine-grained information. Finally, coarse-grained interaction information among subgraphs and fine-grained comparison information among nodes in different subgraphs are integrated to predict the final similarity score. Experimental results on graph datasets with different graph sizes demonstrate that PSimGNN outperforms state-of-the-art methods in graph similarity computation tasks using approximate Graph Edit Distance (GED) as the graph similarity metric. | ['Zhongbin Xu', 'Ziheng Duan', 'Qianru Zhang', 'Haoyan Xu', 'Jie Feng', 'Yueyang Wang', 'Runjian Chen'] | 2020-05-16 | null | null | null | null | ['3d-human-action-recognition', 'graph-similarity'] | ['computer-vision', 'graphs'] | [ 3.66359323e-01 1.05959186e-02 -3.54155123e-01 -3.91084135e-01
-3.10202271e-01 -4.67771441e-01 4.21435207e-01 9.79878724e-01
-1.00184390e-02 4.21252161e-01 -1.57851377e-03 -3.43797088e-01
-2.92278349e-01 -1.30245447e+00 -4.77610826e-01 -6.04600668e-01
-1.44542366e-01 1.82096660e-01 2.51426429e-01 -1.18567824e-01
2.66185164e-01 5.29029191e-01 -8.76880467e-01 1.27337456e-01
1.13551569e+00 8.85370135e-01 3.19008410e-01 1.57085165e-01
-3.50018829e-01 4.28464144e-01 -2.75020570e-01 -3.76843601e-01
2.55555660e-01 -4.97232288e-01 -6.32694423e-01 -2.48359516e-01
3.79233360e-01 7.47671649e-02 -6.36512041e-01 1.62517810e+00
6.94834530e-01 3.06806058e-01 5.17970085e-01 -1.21380591e+00
-8.13800931e-01 6.47389531e-01 -6.93584323e-01 1.66475743e-01
6.36688650e-01 2.11826891e-01 1.26359391e+00 -8.69999290e-01
6.48441672e-01 1.29585814e+00 4.98032182e-01 -3.75406630e-02
-1.14369571e+00 -5.86815357e-01 4.09113139e-01 4.11217630e-01
-1.54864752e+00 1.16139434e-01 1.05201209e+00 -3.01461965e-01
1.23717451e+00 2.48574033e-01 7.66143441e-01 5.41858971e-01
3.94680411e-01 1.93382040e-01 6.37273610e-01 -5.67170558e-03
1.39622137e-01 -4.00144041e-01 3.38355005e-01 1.02038574e+00
4.38774437e-01 -8.26009139e-02 -2.23462135e-01 -2.39899635e-01
4.64333713e-01 4.03767496e-01 -4.51993197e-01 -3.09245139e-01
-1.15154219e+00 9.93835092e-01 1.23016357e+00 4.78771061e-01
-4.24026787e-01 -3.49883586e-02 5.72667301e-01 4.22817141e-01
4.71751243e-01 7.46522665e-01 -1.89005166e-01 5.12810051e-01
-4.98634338e-01 2.19584052e-02 7.19009340e-01 9.91708040e-01
1.09274626e+00 -1.76942527e-01 -3.75044048e-01 5.75697422e-01
3.82073462e-01 -1.70665711e-01 7.12685943e-01 -2.76442677e-01
6.17625475e-01 1.54320967e+00 -6.96994126e-01 -1.87467873e+00
-3.17706198e-01 -3.64795029e-01 -1.22096026e+00 -2.30410233e-01
4.87294272e-02 3.18476349e-01 -8.92236054e-01 1.68998003e+00
4.98989463e-01 4.60714638e-01 -2.19191402e-01 7.49089003e-01
1.40308344e+00 5.83875954e-01 1.53985739e-01 -9.77582708e-02
1.19542646e+00 -8.65413904e-01 -3.77100468e-01 -6.54504402e-03
9.62723553e-01 -3.93406570e-01 9.59518790e-01 -2.04409659e-01
-7.34040737e-01 -5.54991841e-01 -1.23870850e+00 -4.71235327e-02
-8.66799414e-01 -3.83090883e-01 5.84449410e-01 3.08494806e-01
-7.78424323e-01 1.20098519e+00 -4.44330126e-01 -3.46615285e-01
4.13822025e-01 5.01023531e-01 -5.22621930e-01 -2.39466235e-01
-1.55600500e+00 5.59451401e-01 8.39267075e-01 3.99803892e-02
-5.07061779e-01 -6.41786873e-01 -1.21531522e+00 3.28322679e-01
3.88394326e-01 -6.65449381e-01 4.68619794e-01 -6.24893248e-01
-1.09225988e+00 6.24526381e-01 4.56667505e-02 -1.77458763e-01
-5.15258834e-02 6.34432554e-01 -5.02875745e-01 1.05864219e-01
2.84328789e-01 4.66903955e-01 4.37468439e-01 -5.97298145e-01
-3.30811858e-01 -7.01198220e-01 2.40644559e-01 4.30971891e-01
-4.18693602e-01 -2.18154341e-01 -5.22631586e-01 -6.71924710e-01
2.99268723e-01 -7.00848937e-01 -4.57324505e-01 -1.45618245e-01
-7.25003242e-01 -4.17663038e-01 5.52584589e-01 -5.02281487e-01
1.43613613e+00 -1.69768834e+00 3.71005207e-01 6.30896151e-01
8.97325218e-01 2.31357157e-01 -6.37619555e-01 6.62950635e-01
-6.28131449e-01 1.89205095e-01 -2.92478502e-01 4.32575196e-01
-2.22250566e-01 -3.27703565e-01 1.25252485e-01 4.17478412e-01
3.31683964e-01 1.19101608e+00 -1.30491173e+00 -3.54571104e-01
1.47120297e-01 1.61243901e-01 -3.01516622e-01 3.26237708e-01
-1.06391862e-01 1.92410231e-01 -8.26125562e-01 6.54978096e-01
7.92344391e-01 -5.60851514e-01 4.89983916e-01 -6.94857955e-01
2.25091890e-01 2.44892970e-01 -1.01091480e+00 1.72094357e+00
-3.88626680e-02 6.58898875e-02 -4.14281636e-01 -1.34701633e+00
1.16744614e+00 -6.67235181e-02 5.51031590e-01 -6.15067720e-01
1.79765835e-01 3.19539122e-02 3.45374405e-01 -1.00380667e-01
1.27304137e-01 2.50376314e-01 -4.06085178e-02 4.24888998e-01
-9.68897194e-02 -3.88362631e-02 2.83135474e-01 4.84396309e-01
1.44614196e+00 -3.18451226e-01 6.13431334e-01 -2.70284325e-01
9.20643806e-01 -2.33896941e-01 6.12709224e-01 2.65140355e-01
-1.97814658e-01 2.55703062e-01 5.85215390e-01 -5.70569754e-01
-5.36361575e-01 -8.63950670e-01 3.35304052e-01 7.77705669e-01
5.79195619e-01 -8.62787902e-01 -7.76966810e-01 -8.62958372e-01
1.48091555e-01 1.28248900e-01 -5.93368113e-01 -8.98740768e-01
-2.34306276e-01 -6.65291905e-01 3.91607821e-01 2.61606336e-01
5.18782437e-01 -9.86325622e-01 3.12105417e-01 4.78845775e-01
1.97378039e-01 -7.12709308e-01 -1.04072309e+00 1.13296852e-01
-7.80959249e-01 -1.42569602e+00 -3.96057755e-01 -1.10510993e+00
9.06362712e-01 6.76766276e-01 9.81600821e-01 3.68672699e-01
-1.85299978e-01 -2.09433198e-01 -3.05737644e-01 -1.19847387e-01
-2.22600624e-01 1.12436906e-01 -1.03484318e-01 4.91445921e-02
4.79754299e-01 -7.84389317e-01 -6.25596881e-01 2.70172507e-01
-8.65286529e-01 1.50916968e-02 6.38029397e-01 8.66967142e-01
8.90693963e-01 1.49657235e-01 7.00044096e-01 -1.02826858e+00
1.16255724e+00 -5.83403349e-01 -4.57324535e-01 4.40113634e-01
-8.14478934e-01 2.19853029e-01 1.03323340e+00 -3.84514809e-01
-4.62404937e-01 1.07266419e-01 1.79251328e-01 -5.46036720e-01
1.96506441e-01 1.14694929e+00 -5.73999584e-01 -3.84426296e-01
4.77377832e-01 4.38531905e-01 -1.04224712e-01 -1.79475009e-01
4.89436954e-01 4.34849560e-01 1.78772599e-01 -2.10997909e-01
7.59519696e-01 -7.47515112e-02 3.57518047e-01 -5.14607847e-01
-4.20456618e-01 -4.37556744e-01 -7.33054459e-01 1.04428992e-01
8.78409743e-01 -7.79246151e-01 -8.63184750e-01 4.29804057e-01
-1.18882382e+00 1.48284771e-02 4.12805825e-02 4.54611689e-01
-2.68981427e-01 8.71567965e-01 -6.69354916e-01 -6.98917359e-02
-6.59208059e-01 -1.29412889e+00 9.53508973e-01 3.12465698e-01
-7.41908252e-02 -9.52724636e-01 2.77570754e-01 2.80307662e-02
-1.29779845e-01 5.54304779e-01 1.28491914e+00 -9.56018388e-01
-5.96710563e-01 -4.53781754e-01 -6.06569171e-01 -1.41510695e-01
6.38951063e-01 -1.63198605e-01 -3.73946071e-01 -5.52083135e-01
-4.18835491e-01 -7.84408450e-02 8.06753278e-01 1.65005580e-01
1.42333877e+00 -2.85676450e-01 -7.29897141e-01 7.77481437e-01
1.35481572e+00 2.73591936e-01 3.30031067e-01 -7.40206242e-02
1.34524608e+00 3.70877296e-01 5.01900077e-01 1.13582797e-01
3.01538855e-01 5.13063788e-01 4.88077164e-01 -1.45340756e-01
1.10111982e-01 -5.55426002e-01 6.53464645e-02 9.69499826e-01
6.06726669e-02 -3.39198142e-01 -5.46713710e-01 -6.32946491e-02
-1.88941860e+00 -8.03153217e-01 -1.23056993e-01 2.21378422e+00
6.69341564e-01 4.00138944e-02 -5.18233068e-02 -1.19904719e-01
1.18837512e+00 3.88555914e-01 -9.29029405e-01 -3.44632238e-01
5.53071536e-02 1.39526069e-01 2.84164995e-01 2.33096734e-01
-8.71297121e-01 9.68538165e-01 4.88148499e+00 1.06184876e+00
-1.02649081e+00 -3.05412620e-01 8.25722039e-01 4.66944903e-01
-4.66574192e-01 5.02743386e-02 -3.19469720e-01 7.07966328e-01
6.44657075e-01 -5.80046237e-01 5.58018148e-01 6.19194567e-01
1.14217408e-01 3.93713146e-01 -1.27629757e+00 1.05815470e+00
6.99781775e-02 -1.54056621e+00 5.27453899e-01 1.74551383e-01
5.67661107e-01 8.99557769e-03 -3.26031446e-01 2.63626754e-01
3.26929778e-01 -1.08238244e+00 -1.02003820e-01 3.79317582e-01
8.43279958e-01 -9.00519729e-01 5.21007657e-01 3.61348130e-02
-2.05646801e+00 2.16857836e-01 -7.19168484e-01 7.33987689e-02
-2.04610750e-01 6.19426250e-01 -8.51163268e-01 1.05150795e+00
3.23742449e-01 1.23323905e+00 -7.15947390e-01 8.57675850e-01
-2.03392878e-01 2.37554554e-02 -1.02422023e-02 -2.84572035e-01
4.63440835e-01 -7.99434602e-01 4.78080243e-01 8.13010931e-01
2.63836235e-01 2.41731003e-01 5.37024617e-01 1.04843521e+00
-5.29623032e-01 3.93052042e-01 -1.00421560e+00 -3.80494267e-01
6.99791670e-01 1.58458924e+00 -7.93504298e-01 -3.97929281e-01
-5.35727620e-01 1.13050199e+00 8.36053073e-01 1.81974962e-01
-6.68602705e-01 -1.07625425e+00 7.18668938e-01 -5.77666201e-02
-5.04625812e-02 -9.64391232e-02 1.25584781e-01 -9.46384490e-01
-1.62928775e-01 -7.69790590e-01 7.41645038e-01 -6.69948757e-01
-1.51774538e+00 6.60317838e-01 -3.70476902e-01 -1.33777285e+00
1.70586050e-01 -5.34035981e-01 -1.17386270e+00 9.61098313e-01
-1.21662045e+00 -1.00324345e+00 -5.39535284e-01 5.51723123e-01
2.38762304e-01 -2.30780885e-01 7.93111861e-01 2.25846753e-01
-8.74362051e-01 7.83077598e-01 6.10769354e-02 2.27900296e-01
5.32388508e-01 -1.21186280e+00 8.20719957e-01 4.95982528e-01
1.37308955e-01 8.12232137e-01 7.31602237e-02 -9.23679352e-01
-1.64486980e+00 -1.64584672e+00 7.51968265e-01 -2.00934321e-01
6.94938958e-01 -1.85176104e-01 -1.15781772e+00 3.28984648e-01
-2.39104062e-01 1.98607162e-01 7.07525373e-01 -2.12427571e-01
-4.79610294e-01 -1.00588426e-01 -1.07975805e+00 7.49182701e-01
1.50187290e+00 -7.75146127e-01 -3.89545292e-01 5.37019849e-01
1.21132028e+00 -2.73805290e-01 -1.37142289e+00 4.92655218e-01
2.89068907e-01 -6.20818615e-01 1.02755404e+00 -9.87858057e-01
3.15480441e-01 -4.97903436e-01 3.57195809e-02 -1.61643219e+00
-8.09980810e-01 -4.34778333e-01 1.46457672e-01 9.31635916e-01
4.00477380e-01 -7.56189764e-01 8.55869591e-01 3.27266335e-01
-3.03820848e-01 -1.01659322e+00 -6.62170947e-01 -7.17947066e-01
-2.46249899e-01 1.65292576e-01 1.07804990e+00 1.37452257e+00
3.76851499e-01 8.55391443e-01 1.53526336e-01 6.66914694e-03
2.76449770e-01 6.11558020e-01 5.95032215e-01 -1.32174397e+00
-2.22619578e-01 -7.04782069e-01 -9.37393069e-01 -7.82724023e-01
2.79352993e-01 -1.66500378e+00 -3.20006460e-01 -1.56915414e+00
4.28655982e-01 1.36192814e-02 -5.69836676e-01 4.63609040e-01
-5.37295103e-01 9.79303382e-04 -1.19366795e-01 -4.32357080e-02
-5.38590252e-01 7.38504648e-01 1.45649970e+00 -6.92113817e-01
-4.62667584e-01 -3.27894300e-01 -9.27421987e-01 5.18576682e-01
6.62151873e-01 -2.83085048e-01 -6.24933302e-01 2.39501856e-02
1.17412850e-01 -5.19484207e-02 -5.46555445e-02 -7.71175861e-01
2.79522449e-01 -2.99253404e-01 4.32058036e-01 -4.17228073e-01
2.88058780e-02 -6.52026951e-01 2.93441027e-01 8.56209099e-01
-4.30473447e-01 9.84031335e-02 -9.50374156e-02 1.05712616e+00
-3.03784072e-01 7.46131539e-02 6.60678089e-01 -1.14683807e-01
-6.08533919e-01 1.25492620e+00 7.59831667e-02 -1.01456761e-01
1.02085102e+00 -5.19513249e-01 -2.77621806e-01 -1.17641576e-01
-5.30477405e-01 2.87729234e-01 3.89919579e-01 3.81297946e-01
7.90738523e-01 -1.82495999e+00 -4.57894653e-01 6.58354908e-02
5.94670057e-01 -1.18152453e-02 1.38645053e-01 6.07463419e-01
-2.54800200e-01 3.31415713e-01 -2.44870529e-01 -3.48777026e-01
-1.44623876e+00 7.81342149e-01 2.99041539e-01 -5.83705783e-01
-2.15725929e-01 7.76764631e-01 5.29324353e-01 -6.86293900e-01
-1.81059778e-01 -5.03121853e-01 -3.69609028e-01 3.74571350e-03
3.42559397e-01 2.92147875e-01 1.46330014e-01 -6.14983141e-01
-4.46323484e-01 7.91978300e-01 -2.52285063e-01 7.68026173e-01
1.12065184e+00 1.90201566e-01 -3.69318128e-01 -1.79962501e-01
1.62273335e+00 -4.44842249e-01 -7.52270758e-01 -4.41152513e-01
-9.78930146e-02 -4.80181098e-01 3.01650371e-02 -3.68488044e-01
-1.15558767e+00 7.65738487e-01 3.87994498e-01 1.52066723e-01
1.07610047e+00 -6.29296601e-02 8.38794708e-01 7.78791845e-01
1.14005387e-01 -8.22242916e-01 3.12068075e-01 4.26701099e-01
9.00530040e-01 -1.29404473e+00 1.68933541e-01 -7.24368215e-01
-2.99988359e-01 1.00145018e+00 9.12008584e-01 -2.67251492e-01
7.58522570e-01 -4.77182806e-01 -6.33921146e-01 -7.16428936e-01
-3.41043323e-01 -1.80471495e-01 6.74870968e-01 6.80060744e-01
5.29983759e-01 9.50271487e-02 -4.51458484e-01 6.49547219e-01
1.19416304e-01 -4.15083170e-01 3.53414230e-02 6.36002600e-01
-4.99453068e-01 -1.16891491e+00 3.09686959e-01 9.65932012e-01
8.00683647e-02 -5.06296337e-01 -1.01641750e+00 4.76974219e-01
6.32303953e-03 7.12008893e-01 -1.28183469e-01 -8.87416124e-01
3.03570747e-01 -4.17013675e-01 2.69393176e-01 -9.94666338e-01
-6.18173778e-01 -3.25275868e-01 -5.72849996e-02 -7.93449700e-01
2.35666875e-02 -1.56307921e-01 -1.47588909e+00 -3.24363172e-01
-5.45014322e-01 1.85785383e-01 8.41263980e-02 6.58747375e-01
7.15486348e-01 5.04290879e-01 9.80876327e-01 -6.77302778e-01
-3.14172953e-01 -8.32926393e-01 -8.63570690e-01 6.54969573e-01
-2.54869640e-01 -4.02689666e-01 -5.56286126e-02 -4.13589299e-01] | [7.165014266967773, 6.234344482421875] |
04fefac7-9db3-4cb9-bc40-dd1695acd4c9 | knowledge-based-multimodal-music-similarity | 2306.12249 | null | https://arxiv.org/abs/2306.12249v1 | https://arxiv.org/pdf/2306.12249v1.pdf | Knowledge-based Multimodal Music Similarity | Music similarity is an essential aspect of music retrieval, recommendation systems, and music analysis. Moreover, similarity is of vital interest for music experts, as it allows studying analogies and influences among composers and historical periods. Current approaches to musical similarity rely mainly on symbolic content, which can be expensive to produce and is not always readily available. Conversely, approaches using audio signals typically fail to provide any insight about the reasons behind the observed similarity. This research addresses the limitations of current approaches by focusing on the study of musical similarity using both symbolic and audio content. The aim of this research is to develop a fully explainable and interpretable system that can provide end-users with more control and understanding of music similarity and classification systems. | ['Andrea Poltronieri'] | 2023-06-21 | null | null | null | null | ['retrieval'] | ['methodology'] | [ 2.96327509e-02 -4.23284501e-01 -2.17885762e-01 -5.82064539e-02
-2.01668933e-01 -7.64067650e-01 5.25215149e-01 6.32949531e-01
6.37904108e-02 2.00470388e-01 2.31897458e-01 -2.89653987e-02
-8.31074834e-01 -7.30444372e-01 -1.45362958e-01 -1.94741070e-01
-7.13806525e-02 1.48610488e-01 -3.72649217e-03 -3.68579507e-01
7.65674114e-01 4.70207274e-01 -1.94893575e+00 2.90325582e-01
6.26365960e-01 9.41757441e-01 2.36417636e-01 5.42826712e-01
-1.64669037e-01 4.05582458e-01 -6.17252171e-01 -3.13833714e-01
1.67709619e-01 -9.39972997e-01 -5.37087142e-01 -4.05388623e-01
4.06848341e-01 7.53476471e-02 -6.52604997e-02 1.12030804e+00
4.77701634e-01 3.03939700e-01 5.21886528e-01 -1.14590037e+00
-6.35586798e-01 1.02603281e+00 6.37168884e-02 1.77174002e-01
9.39332962e-01 -3.35197866e-01 1.36399591e+00 -2.73138702e-01
4.23844546e-01 9.40666080e-01 8.05091381e-01 1.87998340e-01
-1.18042040e+00 -7.40894139e-01 -1.97861969e-01 5.63717842e-01
-1.35975921e+00 -2.47624904e-01 1.15041518e+00 -5.62657714e-01
3.24638844e-01 6.34531438e-01 1.17088187e+00 7.54894912e-01
-1.27592891e-01 5.28252840e-01 1.00684369e+00 -5.94972491e-01
1.19685411e-01 1.33584067e-01 3.79424393e-02 9.60456729e-02
8.59547406e-02 1.64043218e-01 -9.28112626e-01 -9.93952006e-02
8.18568349e-01 4.10315692e-02 -3.32516164e-01 -7.55057782e-02
-1.13933718e+00 7.73163080e-01 1.95672303e-01 1.11448526e+00
-3.76019984e-01 -3.27008814e-02 3.32050204e-01 6.26668930e-01
7.76130632e-02 1.15058315e+00 -4.91975173e-02 -6.54520035e-01
-1.32215405e+00 4.31829214e-01 7.69993722e-01 4.92005855e-01
3.50910306e-01 5.23528755e-02 2.34922439e-01 6.85391903e-01
3.20434570e-01 3.08680922e-01 7.75114357e-01 -9.93680477e-01
-3.02702546e-01 6.34697080e-01 -1.71805322e-01 -1.57347333e+00
-2.83995122e-01 -6.00276172e-01 -4.24562961e-01 3.01418722e-01
6.02811694e-01 6.68820739e-01 -4.81495708e-02 1.49876928e+00
-6.87900558e-02 2.86757112e-01 -1.79622322e-01 9.79438663e-01
8.85329902e-01 2.10088700e-01 -2.65206903e-01 -3.19705461e-03
1.43709886e+00 -4.79195684e-01 -6.09454215e-01 1.02126993e-01
3.51487994e-01 -1.34605074e+00 1.41820478e+00 7.05905735e-01
-1.05538225e+00 -7.91050613e-01 -1.06940806e+00 5.19493222e-02
-1.80778295e-01 -8.72942656e-02 6.52112484e-01 7.27699578e-01
-4.49432015e-01 1.29467309e+00 -4.47269917e-01 -3.99563372e-01
-2.02475503e-01 4.79888618e-02 -6.59791082e-02 3.68758172e-01
-1.19948792e+00 7.79576957e-01 3.47515285e-01 -2.56006122e-01
-8.88178423e-02 -9.19421256e-01 -4.90586698e-01 1.16395533e-01
1.56572923e-01 -4.77334529e-01 1.36341226e+00 -1.09501231e+00
-1.51751757e+00 5.65840960e-01 1.42465204e-01 -3.66644740e-01
1.32240027e-01 -1.88323796e-01 -5.22799134e-01 -2.61697010e-03
-3.30055542e-02 -7.01384321e-02 6.52370512e-01 -8.00853789e-01
-3.06489646e-01 -3.66441131e-01 2.91958600e-02 8.91953483e-02
-4.10311759e-01 1.40708774e-01 -8.93972889e-02 -1.01112747e+00
4.77251768e-01 -9.91686165e-01 2.72282094e-01 -8.31816941e-02
-1.30112424e-01 1.14321135e-01 4.94616926e-01 -7.32589543e-01
1.71117270e+00 -2.28536272e+00 6.06746674e-02 4.62775618e-01
-1.02072492e-01 1.15133025e-01 2.02163175e-01 7.92583168e-01
-4.74502780e-02 4.55025397e-02 3.25130343e-01 3.57776076e-01
2.65948653e-01 -2.52312198e-02 -5.46549857e-01 6.46455511e-02
-4.85620886e-01 5.98518729e-01 -9.36970532e-01 -3.44001502e-01
3.68855506e-01 3.05310696e-01 -5.28900266e-01 4.36050221e-02
-4.87921759e-02 4.52412337e-01 -4.65870678e-01 6.63662374e-01
-4.90615144e-02 8.49235058e-02 2.80070662e-01 -4.19952959e-01
-3.41574252e-01 6.33270979e-01 -1.44918430e+00 1.68580568e+00
-3.93073380e-01 8.83849680e-01 -3.98282349e-01 -8.52799416e-01
1.27477729e+00 4.51048642e-01 5.13229251e-01 -7.04847753e-01
1.75765857e-01 3.73464376e-01 3.49801481e-01 -4.68984157e-01
7.56563067e-01 -3.62279087e-01 2.00562943e-02 6.77225947e-01
-3.59811902e-01 -4.55242634e-01 2.56708890e-01 -1.97917134e-01
7.28290737e-01 1.10668421e-01 6.26137435e-01 -8.25581998e-02
5.40392518e-01 -6.64879903e-02 3.86221051e-01 3.88594568e-01
2.06313118e-01 5.35370350e-01 2.07219198e-01 -1.94965288e-01
-9.43579495e-01 -1.03807878e+00 -3.35373543e-02 1.02591407e+00
2.72632807e-01 -8.58032763e-01 -5.03553808e-01 1.95416108e-01
1.03405073e-01 7.31395483e-01 -3.78622472e-01 -2.05636397e-01
-2.43146762e-01 3.23350020e-02 7.37659633e-01 4.83534724e-01
-2.25617457e-02 -1.10762930e+00 -6.54591143e-01 2.22649202e-01
-3.61076027e-01 -5.88680506e-01 -2.66463488e-01 -2.42726579e-01
-1.26983023e+00 -1.24037588e+00 -3.42953473e-01 -3.58071953e-01
-2.18974333e-02 3.41758192e-01 1.13384986e+00 1.86178446e-01
-2.60709345e-01 3.78548831e-01 -5.66465616e-01 -6.57930255e-01
-5.58017135e-01 2.00116813e-01 1.84229329e-01 -1.78381503e-01
3.18700284e-01 -9.47124422e-01 -4.45602208e-01 5.66066563e-01
-8.27206016e-01 -1.34633541e-01 3.81198794e-01 4.69931811e-01
4.33475554e-01 1.02714762e-01 6.70536458e-01 -4.94206876e-01
9.80533957e-01 -2.27157429e-01 -1.77107126e-01 1.93699598e-01
-8.81253779e-01 -1.55921048e-02 5.58188319e-01 -7.67360330e-01
-5.00642776e-01 -2.96615392e-01 5.12227416e-02 -4.02567208e-01
-1.19101144e-01 7.13926733e-01 3.92330550e-02 -1.41685903e-01
7.69377112e-01 1.68550253e-01 -1.82790440e-02 -7.75037408e-01
2.86846429e-01 6.98061287e-01 7.63676763e-01 -6.63416982e-01
6.07815325e-01 1.98316261e-01 -3.55461836e-02 -9.31280792e-01
-5.55470586e-01 -6.78212285e-01 -5.87071121e-01 -5.57988763e-01
4.77160841e-01 -3.75360608e-01 -8.87770057e-01 3.71140614e-02
-7.99052477e-01 1.83780611e-01 -5.30085504e-01 8.88497114e-01
-6.09202445e-01 6.32437825e-01 -3.12780291e-01 -8.52878630e-01
-2.27462873e-01 -6.17518961e-01 5.99423945e-01 2.81683594e-01
-1.22382283e+00 -9.01429892e-01 1.34676293e-01 5.65153897e-01
4.00554895e-01 9.51149315e-02 1.03286171e+00 -6.65060341e-01
-3.67542475e-01 -3.17084998e-01 3.53050143e-01 -8.33822042e-02
2.64836937e-01 2.09595874e-01 -8.25483978e-01 5.10446802e-02
-5.11546247e-02 -1.50997296e-01 2.00673297e-01 1.31030589e-01
8.51142883e-01 6.19407110e-02 1.61606848e-01 2.71696329e-01
1.02950847e+00 4.03945506e-01 5.99168837e-01 5.41754544e-01
1.44371986e-01 7.10651755e-01 7.43265867e-01 5.17411828e-01
-8.53773430e-02 1.10701573e+00 2.82385200e-01 3.35432351e-01
-8.88522938e-02 -5.46313941e-01 1.10449515e-01 1.16175127e+00
-5.27150333e-01 2.44258612e-01 -7.04930663e-01 4.53113973e-01
-1.82595825e+00 -1.30343413e+00 -2.14683905e-01 2.32391477e+00
6.32378399e-01 -9.20832157e-02 2.34218985e-01 1.03015494e+00
6.74160600e-01 -1.88715652e-01 -1.78450733e-01 -4.08511132e-01
2.10093185e-02 3.07583272e-01 -1.93605900e-01 1.33155629e-01
-6.21836126e-01 6.73804700e-01 6.69032145e+00 7.13597834e-01
-1.23653686e+00 -3.06162387e-01 -4.80253547e-01 -1.41765207e-01
-3.86284649e-01 1.11395977e-01 -2.50586625e-02 4.32543665e-01
6.95084572e-01 -4.84796613e-01 6.83935404e-01 7.51641035e-01
4.08077657e-01 -2.13644151e-02 -1.27674735e+00 1.15846550e+00
2.31746789e-02 -1.06257749e+00 1.51271820e-02 -6.06511831e-02
3.93982351e-01 -5.13298512e-01 8.37647393e-02 -1.22382671e-01
-3.54914784e-01 -1.01081455e+00 9.74501193e-01 9.75703597e-01
1.72568589e-01 -6.21956170e-01 5.31716347e-01 1.59693122e-01
-1.15735078e+00 -6.37399498e-03 -1.26956359e-01 -6.56332552e-01
1.89105272e-01 3.40896070e-01 -7.04902828e-01 7.10978866e-01
7.17971027e-01 7.25431144e-01 -3.56537402e-01 1.22333074e+00
-3.12099189e-01 6.73600733e-01 -1.93917856e-01 -2.34057054e-01
-2.45271921e-01 -4.73546535e-01 6.23453856e-01 8.36975813e-01
7.64412045e-01 -4.57928553e-02 -6.03111349e-02 9.52878535e-01
4.69256520e-01 5.08964777e-01 -6.30826950e-01 -5.66565931e-01
5.54528236e-01 8.46753538e-01 -9.41627800e-01 -3.53955626e-02
-2.45006084e-01 6.53570354e-01 -3.42012852e-01 -1.46563813e-01
-4.97828454e-01 -2.95848787e-01 7.99698949e-01 3.86499852e-01
5.46358712e-02 -3.87440681e-01 -5.09418547e-01 -8.50477159e-01
-1.74598798e-01 -1.22803426e+00 2.88261443e-01 -8.77574742e-01
-1.32396317e+00 2.96515346e-01 -5.99740818e-02 -1.52298808e+00
-7.01476514e-01 -3.50123316e-01 -6.51249528e-01 7.34269977e-01
-7.92970359e-01 -9.58180904e-01 -3.59975487e-01 5.41027427e-01
2.17891961e-01 -2.93978423e-01 1.13915765e+00 3.90274316e-01
5.31907678e-02 4.95151520e-01 6.24526739e-02 -2.72586584e-01
7.93571472e-01 -1.00451946e+00 -8.99263285e-03 3.69694501e-01
8.74367356e-01 7.92075872e-01 1.08413005e+00 -5.19621253e-01
-1.23203576e+00 -2.42305994e-01 9.84438419e-01 -2.04011500e-01
9.00410175e-01 2.35178128e-01 -9.05742049e-01 6.72773719e-02
-1.76701918e-01 -8.15738916e-01 1.28891563e+00 5.76307714e-01
-4.47694808e-01 -1.00931928e-01 -6.76585078e-01 6.59747422e-01
9.37842369e-01 -1.08398235e+00 -1.04197097e+00 -2.47503251e-01
1.46591976e-01 2.98835486e-02 -1.26119614e+00 2.06041172e-01
1.27404773e+00 -1.15139151e+00 1.05035555e+00 -4.46143061e-01
3.41304451e-01 -4.64349836e-01 -2.78906286e-01 -1.09471679e+00
-3.10915381e-01 -5.23712695e-01 2.24246562e-01 1.31712210e+00
2.48523340e-01 -2.02837348e-01 6.85879648e-01 3.21986377e-01
7.90989473e-02 -3.00906837e-01 -4.25812662e-01 -1.07060015e+00
-3.59309107e-01 -9.87625957e-01 7.77570724e-01 1.31985557e+00
2.89947182e-01 1.88840896e-01 -4.02959496e-01 -9.19942856e-02
2.25179598e-01 7.79563427e-01 8.57039273e-01 -1.68740976e+00
-6.60567522e-01 -8.61551881e-01 -9.91298914e-01 -4.32907850e-01
-1.73689038e-01 -1.01400924e+00 -4.34391171e-01 -1.19325125e+00
-1.52938217e-01 -3.35430026e-01 -4.12524909e-01 1.47594467e-01
2.46779352e-01 6.35947227e-01 5.52468002e-01 5.14182389e-01
-1.66394129e-01 2.97040612e-01 1.03347170e+00 -3.41643579e-02
-4.31957394e-01 4.41952944e-01 -7.10357606e-01 1.01719737e+00
9.36263204e-01 -4.59990084e-01 -5.47001660e-01 1.62289087e-02
6.22636795e-01 -1.37507975e-01 4.22539741e-01 -1.17388833e+00
3.63241434e-01 -1.47258312e-01 -8.61829612e-03 -3.68123055e-01
2.65459388e-01 -7.49903321e-01 7.43715882e-01 5.06652117e-01
-3.24165493e-01 9.58495513e-02 2.71541532e-02 2.69467354e-01
-5.61116159e-01 -7.11659014e-01 2.97752589e-01 8.53429586e-02
-5.26027858e-01 -3.48124743e-01 -2.18487471e-01 -1.01276189e-01
6.97677493e-01 -4.03948098e-01 2.97179550e-01 -6.75485671e-01
-7.62577236e-01 -4.98218685e-01 6.14274144e-01 6.42869413e-01
4.16684091e-01 -1.48944819e+00 -4.88762021e-01 -5.04832789e-02
2.23639131e-01 -8.48619342e-01 1.69680998e-01 6.47179306e-01
-5.02412260e-01 2.87084043e-01 -4.05072749e-01 -3.98449212e-01
-1.75635898e+00 4.46460038e-01 3.24610574e-03 1.44118115e-01
-5.51573873e-01 3.42714608e-01 -2.69034803e-01 -1.56762943e-01
2.38776818e-01 -3.00769687e-01 -3.77135903e-01 3.37564975e-01
5.25668502e-01 6.73007309e-01 -1.21161841e-01 -4.84493613e-01
-1.82388470e-01 9.19690490e-01 4.86056030e-01 -3.66561651e-01
1.05011272e+00 5.76402061e-02 -1.03763245e-01 1.08585978e+00
5.25906503e-01 3.28256100e-01 -3.85487795e-01 -7.41375163e-02
1.79167315e-01 -7.98999608e-01 -4.56028618e-02 -5.59332550e-01
-6.20980144e-01 8.29414129e-01 3.96797031e-01 5.73359787e-01
1.17504740e+00 5.31511568e-02 7.03605890e-01 4.20011044e-01
3.44053984e-01 -1.10769677e+00 7.12541416e-02 2.37653285e-01
1.00083911e+00 -7.25316107e-01 2.12987736e-01 -3.69030565e-01
-3.61775011e-01 1.31337607e+00 -7.39440098e-02 1.10239111e-01
7.32857466e-01 -6.13547210e-03 1.30073488e-01 -1.27210200e-01
-2.48579577e-01 -2.80956417e-01 8.67539704e-01 4.43340927e-01
1.04266322e+00 1.00258999e-01 -9.27783906e-01 9.52157557e-01
-1.03129840e+00 7.04133883e-02 2.35780895e-01 7.32495844e-01
-4.12350416e-01 -1.55222058e+00 -6.32013023e-01 2.00549945e-01
-6.08140469e-01 -4.78212908e-02 -7.90120065e-01 6.42422497e-01
3.28778356e-01 9.86858428e-01 -9.38022137e-02 -9.01963234e-01
3.53610456e-01 1.09825164e-01 6.93463802e-01 -4.25317287e-01
-7.72684932e-01 9.44798440e-02 -2.20292644e-03 -4.17773366e-01
-5.67865431e-01 -8.41005743e-01 -1.07758260e+00 -5.74149430e-01
-4.03760403e-01 5.52610636e-01 9.94654953e-01 8.13809335e-01
1.40136749e-01 5.11088133e-01 5.97183347e-01 -6.93369806e-01
-3.34850013e-01 -1.11542892e+00 -8.98052394e-01 5.88268280e-01
-4.55895156e-01 -5.91468155e-01 -1.41334325e-01 1.24783844e-01] | [15.92845344543457, 5.2680840492248535] |
8aa5d24a-b67f-4264-bf59-0951b421a600 | pifu-pixel-aligned-implicit-function-for-high | 1905.05172 | null | https://arxiv.org/abs/1905.05172v3 | https://arxiv.org/pdf/1905.05172v3.pdf | PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitization | We introduce Pixel-aligned Implicit Function (PIFu), a highly effective implicit representation that locally aligns pixels of 2D images with the global context of their corresponding 3D object. Using PIFu, we propose an end-to-end deep learning method for digitizing highly detailed clothed humans that can infer both 3D surface and texture from a single image, and optionally, multiple input images. Highly intricate shapes, such as hairstyles, clothing, as well as their variations and deformations can be digitized in a unified way. Compared to existing representations used for 3D deep learning, PIFu can produce high-resolution surfaces including largely unseen regions such as the back of a person. In particular, it is memory efficient unlike the voxel representation, can handle arbitrary topology, and the resulting surface is spatially aligned with the input image. Furthermore, while previous techniques are designed to process either a single image or multiple views, PIFu extends naturally to arbitrary number of views. We demonstrate high-resolution and robust reconstructions on real world images from the DeepFashion dataset, which contains a variety of challenging clothing types. Our method achieves state-of-the-art performance on a public benchmark and outperforms the prior work for clothed human digitization from a single image. | ['Hao Li', 'Angjoo Kanazawa', 'Zeng Huang', 'Shigeo Morishima', 'Shunsuke Saito', 'Ryota Natsume'] | 2019-05-13 | pifu-pixel-aligned-implicit-function-for-high-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Saito_PIFu_Pixel-Aligned_Implicit_Function_for_High-Resolution_Clothed_Human_Digitization_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Saito_PIFu_Pixel-Aligned_Implicit_Function_for_High-Resolution_Clothed_Human_Digitization_ICCV_2019_paper.pdf | iccv-2019-10 | ['3d-shape-reconstruction-from-a-single-2d', '3d-object-reconstruction', '3d-object-reconstruction-from-a-single-image', '3d-human-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.39927763e-01 -2.52305210e-01 3.84707689e-01 -3.64357293e-01
-7.01407254e-01 -7.44013488e-01 2.17950180e-01 -3.93904269e-01
1.33769400e-03 4.75504875e-01 1.77917659e-01 4.18431342e-01
2.83560485e-01 -1.04830015e+00 -1.16284251e+00 -5.62679589e-01
2.64896035e-01 7.12922454e-01 2.33944416e-01 -1.87001064e-01
-3.26143771e-01 7.52770245e-01 -1.18798029e+00 2.87504107e-01
4.22365576e-01 1.16615129e+00 1.99921176e-01 7.17951000e-01
7.40153790e-02 -9.26225483e-02 -2.75193363e-01 -5.93288362e-01
4.58371669e-01 7.72752939e-03 -5.98094165e-01 7.29874372e-01
1.28287125e+00 -7.11400211e-01 -4.63395447e-01 9.17447150e-01
4.98267770e-01 -1.43740177e-01 6.85221434e-01 -6.11041725e-01
-1.09672034e+00 -1.88937671e-02 -7.43168175e-01 -4.94070441e-01
5.63706756e-01 3.06557477e-01 5.74878037e-01 -9.27934170e-01
1.03442800e+00 1.53064024e+00 1.11815464e+00 4.56938356e-01
-1.69915247e+00 -1.60146043e-01 2.76548117e-01 -4.50026393e-01
-1.32305324e+00 -1.19908519e-01 1.16101563e+00 -3.67885351e-01
6.18427336e-01 3.35194677e-01 9.63303566e-01 1.30125141e+00
1.53633595e-01 8.01621020e-01 1.27449584e+00 -1.14093557e-01
-1.00787506e-01 -4.67497826e-01 -2.10277021e-01 1.00433636e+00
1.97992787e-01 1.04935821e-02 -2.16370195e-01 -2.15667233e-01
1.75634074e+00 2.14669421e-01 -2.74802953e-01 -8.37093532e-01
-1.44241440e+00 3.66059303e-01 5.27933300e-01 -3.80081922e-01
-5.69587350e-01 4.74796116e-01 2.37472117e-01 8.31103399e-02
5.88217556e-01 1.78204030e-02 -4.81493622e-01 1.72294229e-01
-7.73264349e-01 6.10249221e-01 7.68349230e-01 1.09710610e+00
6.84853137e-01 -3.93420756e-02 2.55946144e-02 7.51207829e-01
1.70964181e-01 8.96983922e-01 -3.60278189e-01 -1.21999514e+00
3.26760739e-01 5.75242400e-01 2.89780200e-01 -1.10193706e+00
-2.55796254e-01 -6.61313161e-02 -1.25917876e+00 3.73282790e-01
4.67841923e-01 1.22593604e-01 -1.25447011e+00 1.48816371e+00
6.46047711e-01 -1.80477560e-01 -4.05535161e-01 1.40418839e+00
1.05378723e+00 5.18866062e-01 -2.22632393e-01 2.76759505e-01
1.41360855e+00 -8.41784596e-01 -4.58344012e-01 -1.48603961e-01
-3.97408634e-01 -8.00802886e-01 1.10201180e+00 5.85522473e-01
-1.51460147e+00 -7.11765707e-01 -8.47960830e-01 -7.15376377e-01
-1.31949499e-01 -1.03658931e-02 7.07333267e-01 1.99805379e-01
-1.00122154e+00 6.98462129e-01 -8.68051231e-01 -3.48607242e-01
5.70936322e-01 1.94082037e-01 -6.52032852e-01 -3.53820741e-01
-7.26221561e-01 4.50322300e-01 -2.28181839e-01 3.04896384e-01
-9.15646195e-01 -7.84805417e-01 -9.96684968e-01 -2.99090564e-01
4.11294848e-01 -1.19296408e+00 7.83100784e-01 -9.70659792e-01
-1.49006307e+00 1.32244611e+00 4.51712012e-02 -3.33786756e-03
9.72911417e-01 -4.20968324e-01 1.83515642e-02 5.64470477e-02
-1.66541100e-01 6.54809475e-01 1.06462121e+00 -1.55805230e+00
1.00474551e-01 -7.07028747e-01 1.83527157e-01 2.34126121e-01
4.00924742e-01 -3.72090787e-01 -8.02551270e-01 -1.04770863e+00
4.38038915e-01 -8.05535972e-01 -2.86218196e-01 9.65382218e-01
-4.53592688e-01 5.05423695e-02 7.25349188e-01 -1.05306089e+00
5.43231189e-01 -2.11546087e+00 8.24626625e-01 6.94435537e-02
2.69299448e-01 -1.48804650e-01 -6.02128059e-02 1.07033394e-01
4.01763380e-01 5.09739593e-02 -5.53776920e-01 -5.70163906e-01
2.74084598e-01 5.06742895e-01 -7.96784088e-02 6.51842356e-01
1.31698444e-01 1.09628904e+00 -7.88732111e-01 -4.15669948e-01
2.88334638e-01 9.61718082e-01 -4.10907716e-01 3.28718603e-01
-3.83862615e-01 3.78869057e-01 -1.93756729e-01 1.08848834e+00
1.01561153e+00 -1.61164030e-01 2.08483875e-01 -7.07799554e-01
2.45411813e-01 -1.57435283e-01 -1.30787039e+00 2.30845046e+00
-3.15021366e-01 2.23262906e-01 6.45553350e-01 -4.29761082e-01
9.79061246e-01 2.88751215e-01 3.77769768e-01 -3.86755496e-01
-2.19704192e-02 -4.25982103e-02 -8.48848104e-01 -2.46195748e-01
3.96136373e-01 -8.13804790e-02 -2.79742390e-01 3.39647502e-01
2.69037038e-02 -4.71890599e-01 -3.07448983e-01 -1.72292620e-01
7.00821817e-01 7.59275377e-01 1.73913285e-01 -1.70144528e-01
9.08020437e-02 -2.20721081e-01 4.70975071e-01 3.25331420e-01
3.48755457e-02 1.20603704e+00 1.53922558e-01 -9.83714342e-01
-1.45320845e+00 -1.76314974e+00 -1.94496363e-01 8.22267413e-01
4.13737893e-01 -1.96935445e-01 -7.49701023e-01 -6.38409317e-01
5.63719869e-01 -3.81744616e-02 -8.04540813e-01 3.70164871e-01
-7.46814549e-01 -1.43324867e-01 2.69889414e-01 7.95474589e-01
7.26383984e-01 -1.02151215e+00 -5.30174851e-01 1.73531041e-01
-1.14398062e-01 -1.08540118e+00 -1.00465858e+00 -2.86751062e-01
-9.03177261e-01 -1.06881392e+00 -8.94972801e-01 -9.25922334e-01
8.47376168e-01 2.60139883e-01 1.53914261e+00 -3.87161821e-02
-5.60759664e-01 7.50803709e-01 8.70732740e-02 4.32560034e-02
-1.97951600e-01 -3.49512517e-01 2.60018613e-02 5.72533570e-02
-1.02788970e-01 -6.52881086e-01 -7.49854922e-01 4.49313492e-01
-8.47084224e-01 2.79995382e-01 4.30164844e-01 8.68346393e-01
1.42552304e+00 -1.31405354e-01 1.64150551e-01 -8.99662197e-01
2.97585994e-01 -8.35536048e-02 -4.45358127e-01 3.90494823e-01
1.10265233e-01 -1.82324111e-01 3.71571451e-01 -6.96736038e-01
-9.90297496e-01 4.20304358e-01 -1.92889929e-01 -7.95533061e-01
-4.48984176e-01 -1.97226312e-02 -4.08797204e-01 -8.04035738e-02
5.09523153e-01 1.34338573e-01 2.01997697e-01 -9.21693742e-01
6.57672644e-01 1.81454182e-01 9.07150567e-01 -9.37259018e-01
7.71515369e-01 7.61928499e-01 -1.32200187e-02 -8.44310403e-01
-5.84817290e-01 -1.68800466e-02 -1.32872391e+00 -2.33505532e-01
8.90089035e-01 -1.00357032e+00 -5.60681999e-01 8.53448629e-01
-1.18420613e+00 -5.82993388e-01 -3.66938263e-01 -9.88753885e-02
-6.53572321e-01 4.61280704e-01 -8.23235869e-01 -4.28899348e-01
-6.04936004e-01 -9.46322024e-01 1.76074553e+00 -1.10678792e-01
-4.02758092e-01 -1.12492681e+00 -1.84010863e-01 4.93612796e-01
1.67938784e-01 1.03250217e+00 8.01792204e-01 5.83167791e-01
-7.64954209e-01 -1.15277454e-01 -1.26660243e-01 4.36196029e-01
3.34144711e-01 -1.12985671e-01 -8.31503630e-01 -5.33216178e-01
-2.54235476e-01 -4.84599501e-01 7.06185102e-01 4.35703695e-01
1.28133285e+00 -4.44410264e-01 -1.00771829e-01 9.74493921e-01
1.45689189e+00 -5.22788763e-01 6.20394528e-01 -1.75286010e-01
1.19592941e+00 3.24574858e-01 2.80360967e-01 3.72243315e-01
4.91667747e-01 7.83201754e-01 4.98491496e-01 -5.45186400e-01
-4.78567988e-01 -3.83971006e-01 2.30682194e-01 6.42663360e-01
-5.31318247e-01 2.49548107e-01 -3.44644994e-01 3.71164978e-01
-1.55723178e+00 -7.93449283e-01 -4.16288599e-02 2.22173786e+00
8.65007162e-01 -1.06601797e-01 3.39517683e-01 -1.30345464e-01
4.45229053e-01 2.48571947e-01 -9.78174388e-01 -2.90620446e-01
-3.31474900e-01 3.58046383e-01 5.12707770e-01 3.92496347e-01
-1.23597932e+00 7.94020295e-01 6.67158651e+00 4.32514310e-01
-1.03929818e+00 7.18357414e-02 4.21350449e-01 -2.31011391e-01
-4.93606329e-01 -5.93823612e-01 -3.96670341e-01 2.41194859e-01
-2.32249107e-02 3.90148044e-01 8.22528958e-01 7.96748698e-01
-2.26146221e-01 2.07515970e-01 -1.18036878e+00 1.30117059e+00
1.72647238e-01 -1.32428336e+00 2.50067443e-01 1.66091830e-01
8.60563636e-01 -2.90836155e-01 1.05304480e-01 -1.34595692e-01
2.23928183e-01 -1.19893336e+00 1.11367941e+00 8.85600090e-01
1.37136781e+00 -5.56897402e-01 3.18412900e-01 1.19989671e-01
-1.38739383e+00 3.79685313e-01 -5.65341771e-01 9.86269042e-02
2.21370220e-01 4.82210577e-01 -3.84034812e-01 5.92496753e-01
9.24670398e-01 7.37328768e-01 -3.56255323e-01 5.98994613e-01
-8.94953683e-02 2.06557915e-01 -4.55082327e-01 3.96361887e-01
-2.57055294e-02 -2.94660777e-01 4.26360101e-01 1.11224127e+00
1.20551005e-01 2.67103821e-01 4.32539165e-01 1.15601599e+00
-3.47912073e-01 -2.53276229e-01 -6.56952322e-01 2.04715624e-01
1.36740834e-01 1.16695535e+00 -5.05831897e-01 -2.10557312e-01
-3.04351240e-01 1.61190867e+00 4.73651499e-01 4.56208050e-01
-7.75519609e-01 6.70082644e-02 8.13753903e-01 5.34590721e-01
6.06457889e-01 -6.29643142e-01 -3.90507370e-01 -1.09698224e+00
3.48204046e-01 -7.16028333e-01 9.00863856e-02 -9.71035004e-01
-1.86792958e+00 4.54607517e-01 -1.82518482e-01 -9.43667591e-01
3.63246083e-01 -7.05314636e-01 -3.08948994e-01 9.93781209e-01
-1.05313849e+00 -1.69899058e+00 -6.58264041e-01 7.11640716e-01
6.43066525e-01 3.67230147e-01 9.71883357e-01 1.24799021e-01
-9.15567428e-02 3.50403696e-01 -1.31569669e-01 3.70585203e-01
6.77936554e-01 -1.43407798e+00 8.38452876e-01 4.35078591e-01
4.29129638e-02 3.07105482e-01 2.09512964e-01 -8.45787525e-01
-2.07283044e+00 -1.24601436e+00 1.58425033e-01 -6.90220416e-01
-2.39887349e-02 -6.58484578e-01 -9.83842969e-01 7.66315401e-01
7.15513229e-02 3.26996982e-01 2.65967071e-01 1.61480248e-01
-5.23434937e-01 -2.02340469e-01 -1.48066175e+00 5.19550860e-01
1.63531113e+00 -5.64623594e-01 -5.73897123e-01 2.91004777e-01
5.77815711e-01 -1.10427201e+00 -1.39393449e+00 3.06564271e-01
9.85084414e-01 -8.48038793e-01 1.46838617e+00 -4.40019310e-01
4.01425332e-01 -2.47865468e-01 -4.38109368e-01 -1.19533813e+00
-7.26803899e-01 -5.19838631e-01 -4.02937204e-01 9.87781346e-01
-2.17739165e-01 -1.75837874e-01 6.90406382e-01 6.39164150e-01
9.04331077e-03 -1.05498278e+00 -7.69270778e-01 -4.99068499e-01
5.03791906e-02 -1.08275324e-01 8.61988485e-01 7.48600543e-01
-8.39428008e-01 -1.50359794e-02 -6.27775908e-01 3.81714821e-01
1.22617257e+00 5.74361324e-01 1.03792393e+00 -1.31288731e+00
-2.15537503e-01 -1.49892271e-01 -4.28231627e-01 -1.44900107e+00
-7.73933828e-02 -7.75997758e-01 1.39150703e-02 -1.71235538e+00
9.95338932e-02 -5.17153323e-01 1.35116279e-01 3.98347616e-01
1.60917062e-02 6.81854904e-01 8.25009868e-02 -5.46406284e-02
-4.00422662e-01 4.45816487e-01 1.90421236e+00 -4.32714790e-01
-2.81147622e-02 -1.24791481e-01 -4.22515601e-01 8.50696385e-01
2.98121661e-01 6.96892217e-02 -8.68689120e-02 -8.53825092e-01
8.31189752e-03 8.56373385e-02 9.32401240e-01 -6.30979776e-01
-2.01868817e-01 -1.59618899e-01 1.14321220e+00 -8.24439824e-01
7.26204395e-01 -8.89730453e-01 6.84754968e-01 -1.88613497e-02
-3.16240527e-02 -1.09809898e-02 1.23836167e-01 7.43185043e-01
2.94262230e-01 4.49772149e-01 8.55581760e-01 -6.66502357e-01
-5.29547930e-01 7.88483262e-01 2.59202033e-01 3.02354116e-02
7.42741585e-01 -3.46156597e-01 1.20775364e-02 -2.79457718e-02
-9.42594230e-01 -5.20022810e-02 1.20427501e+00 4.58542287e-01
1.10743523e+00 -1.66211998e+00 -6.55003369e-01 4.03783351e-01
-3.34290445e-01 7.37362802e-01 4.76392150e-01 4.09210384e-01
-8.19053769e-01 -1.46659523e-01 -5.28407633e-01 -8.85096371e-01
-1.26209021e+00 5.61923862e-01 4.69780266e-01 7.35977069e-02
-1.26237059e+00 7.23012447e-01 4.46566254e-01 -7.78021097e-01
2.83516049e-01 -6.78876340e-01 4.87128168e-01 -4.58519757e-01
2.53921270e-01 2.30231062e-01 -1.87851220e-01 -7.59828925e-01
-1.34322792e-01 1.18345022e+00 1.87862396e-01 2.33553410e-01
1.43640912e+00 -4.99414988e-02 -2.30799600e-01 4.60714370e-01
1.12295151e+00 6.39607459e-02 -2.08326411e+00 -4.60845590e-01
-7.48364151e-01 -8.52486849e-01 -1.08718224e-01 -8.99221659e-01
-1.30604482e+00 7.94938743e-01 4.73711401e-01 -6.34828284e-02
1.03563416e+00 1.18579060e-01 1.14972067e+00 8.66026208e-02
8.40297222e-01 -8.20665777e-01 7.40392506e-02 8.60245749e-02
1.48438275e+00 -9.86547410e-01 3.69085580e-01 -6.18958712e-01
-6.16476059e-01 1.19032192e+00 4.83174831e-01 -5.00067294e-01
5.88262618e-01 3.13724935e-01 -1.56040266e-01 -3.59794140e-01
-3.18432838e-01 1.62097290e-01 5.78603387e-01 7.43371248e-01
1.30152225e-01 4.91295159e-01 3.52049679e-01 6.24504328e-01
-2.64329668e-02 -1.48914963e-01 2.72479746e-02 8.10384214e-01
1.57938059e-02 -8.44703794e-01 -6.86990619e-01 2.65151381e-01
-1.08935572e-02 3.01998228e-01 -5.02962112e-01 7.70029306e-01
3.17555606e-01 2.15319529e-01 2.19126299e-01 -1.48414731e-01
7.11866975e-01 -1.90696478e-01 1.20807183e+00 -4.22545016e-01
-4.37455297e-01 1.82721540e-01 -2.12415140e-02 -7.61701465e-01
-4.47363198e-01 -6.71032369e-01 -8.98973823e-01 -5.95798969e-01
2.55474240e-01 -6.20391190e-01 4.61692333e-01 6.03212774e-01
3.80905032e-01 3.70574415e-01 3.55456382e-01 -1.57164848e+00
-2.96236932e-01 -6.70529962e-01 -8.01418543e-01 8.41367066e-01
4.13701624e-01 -9.18468058e-01 1.35792866e-01 3.95706713e-01] | [7.157039165496826, -1.2865161895751953] |
01cedb45-8ba0-454b-ba98-521525380821 | audio-source-separation-via-multi-scale | 1904.04161 | null | http://arxiv.org/abs/1904.04161v1 | http://arxiv.org/pdf/1904.04161v1.pdf | Audio Source Separation via Multi-Scale Learning with Dilated Dense U-Nets | Modern audio source separation techniques rely on optimizing sequence model
architectures such as, 1D-CNNs, on mixture recordings to generalize well to
unseen mixtures. Specifically, recent focus is on time-domain based
architectures such as Wave-U-Net which exploit temporal context by extracting
multi-scale features. However, the optimality of the feature extraction process
in these architectures has not been well investigated. In this paper, we
examine and recommend critical architectural changes that forge an optimal
multi-scale feature extraction process. To this end, we replace regular $1-$D
convolutions with adaptive dilated convolutions that have innate capability of
capturing increased context by using large temporal receptive fields. We also
investigate the impact of dense connections on the extraction process that
encourage feature reuse and better gradient flow. The dense connections between
the downsampling and upsampling paths of a U-Net architecture capture
multi-resolution information leading to improved temporal modelling. We
evaluate the proposed approaches on the MUSDB test dataset. In addition to
providing an improved performance over the state-of-the-art, we also provide
insights on the impact of different architectural choices on complex
data-driven solutions for source separation. | ['Andreas Spanias', 'Huan Song', 'Sameeksha Katoch', 'Vivek Sivaraman Narayanaswamy', 'Jayaraman J. Thiagarajan'] | 2019-04-08 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 1.42588988e-01 -3.75646114e-01 6.94573596e-02 -3.25842708e-01
-6.95928693e-01 -4.98051524e-01 4.07784522e-01 -1.36123776e-01
-3.91862959e-01 5.16182125e-01 5.67434072e-01 -4.74754423e-02
-4.19141799e-01 -4.56368297e-01 -5.28510332e-01 -6.15654528e-01
-5.61200440e-01 -3.77388656e-01 1.25745058e-01 -1.51946992e-01
1.02862969e-01 5.22739828e-01 -1.60065496e+00 6.60674036e-01
5.31794906e-01 1.19539309e+00 1.02784835e-01 1.02205217e+00
4.44776155e-02 6.17085218e-01 -5.90990663e-01 8.90700370e-02
2.67479658e-01 -5.82381248e-01 -5.57540059e-01 -1.12741679e-01
6.58896983e-01 -3.88082862e-01 -4.73188668e-01 7.97431052e-01
9.05352950e-01 3.81805301e-01 5.98432004e-01 -9.26671326e-01
-4.40254003e-01 9.21565533e-01 -4.48078990e-01 8.99022639e-01
3.50832120e-02 2.16470227e-01 1.13797355e+00 -8.46825659e-01
4.15620565e-01 1.06629467e+00 7.89059579e-01 4.67755914e-01
-1.31279695e+00 -7.33064294e-01 2.77304947e-01 3.00022185e-01
-1.17327058e+00 -1.04149461e+00 1.09055233e+00 -2.53949016e-01
1.38591433e+00 1.22684851e-01 4.88540649e-01 1.22843075e+00
-3.06408666e-02 7.60713220e-01 8.99999559e-01 -3.97229731e-01
1.93760455e-01 -1.00648187e-01 9.73998755e-02 2.21473634e-01
-1.80032432e-01 3.05039495e-01 -1.03690183e+00 7.78466789e-03
8.45277667e-01 -1.54661283e-01 -5.65544903e-01 4.12201546e-02
-9.23622131e-01 4.85290200e-01 4.56799537e-01 6.91664696e-01
-4.26692247e-01 3.53067636e-01 4.87206548e-01 5.12138844e-01
5.59421539e-01 5.36627650e-01 -5.21780491e-01 -3.75538588e-01
-1.48596299e+00 1.23122230e-01 4.30766642e-01 7.31712580e-01
4.69152063e-01 6.81909025e-01 -3.04702550e-01 9.41543162e-01
6.32673576e-02 -4.91920263e-02 8.17366660e-01 -1.02433610e+00
3.66333723e-01 1.93406701e-01 -2.23284453e-01 -6.78571880e-01
-4.60695356e-01 -1.05951285e+00 -7.58630633e-01 1.79806501e-01
4.48549360e-01 -4.15919214e-01 -9.10803676e-01 1.85141110e+00
4.09319773e-02 6.89122915e-01 1.27939746e-01 9.33958411e-01
6.03802562e-01 5.64624131e-01 -2.23637447e-01 -9.21466295e-03
1.25183964e+00 -9.20832396e-01 -7.01470375e-01 -2.52623141e-01
3.53280574e-01 -9.06133294e-01 7.64834881e-01 4.50698137e-01
-1.32882559e+00 -9.00927126e-01 -1.23777163e+00 1.42549142e-01
-1.70670941e-01 1.69787034e-01 2.34319583e-01 5.30537367e-01
-1.17734802e+00 1.29661894e+00 -7.93803632e-01 -3.36290419e-01
3.42830002e-01 4.36028808e-01 -2.51177642e-02 2.65214264e-01
-1.22966850e+00 5.31134546e-01 1.34551868e-01 5.89031130e-02
-9.05832171e-01 -8.74080360e-01 -5.94730675e-01 2.59508401e-01
-1.13304835e-02 -5.58522761e-01 1.32980692e+00 -1.04496527e+00
-1.55406189e+00 2.62064278e-01 -2.28547618e-01 -9.20860410e-01
1.05932325e-01 -4.01369929e-01 -7.53634155e-01 3.72373998e-01
-2.99765825e-01 7.23308802e-01 1.14042222e+00 -9.32071745e-01
-7.32528448e-01 -1.04098029e-01 -1.14405274e-01 1.34532034e-01
-5.53627551e-01 8.54886398e-02 3.80988121e-02 -1.15692079e+00
-1.45391136e-01 -6.86549306e-01 -2.32393190e-01 -3.39660585e-01
-7.39893019e-02 2.06911922e-01 8.14447224e-01 -6.11922443e-01
1.55321705e+00 -2.52376223e+00 -7.45068173e-05 7.72974268e-02
1.09565623e-01 3.56213838e-01 -3.73894572e-01 3.47278208e-01
-2.81374991e-01 -1.51268458e-02 -9.91247743e-02 -4.04877543e-01
-7.46961236e-02 -1.69528350e-01 -4.70705330e-01 3.82564276e-01
5.27496934e-01 5.74895978e-01 -5.97418845e-01 -1.89366892e-01
4.00866158e-02 6.92041874e-01 -9.19278800e-01 1.15306064e-01
-2.34510098e-02 4.59056050e-01 -1.64662406e-01 3.39948684e-01
4.63010490e-01 6.94042118e-03 4.49547917e-03 -2.55442917e-01
-4.26944226e-01 7.83616006e-01 -1.28720915e+00 1.92844093e+00
-5.96164107e-01 9.36585486e-01 3.83989811e-01 -8.76028180e-01
8.78521740e-01 7.23683655e-01 5.95780849e-01 -6.44809544e-01
2.37491831e-01 5.10635197e-01 4.06178355e-01 -1.76239178e-01
6.06752574e-01 -1.77611396e-01 3.48990619e-01 3.69307756e-01
4.28830683e-01 1.33974895e-01 1.32886156e-01 -1.73979491e-01
1.18077552e+00 1.75126702e-01 -8.30602869e-02 -4.85411614e-01
2.89896935e-01 -3.97288889e-01 4.41191107e-01 6.16000473e-01
-2.95271277e-01 7.60798514e-01 1.26072764e-01 -9.41544473e-02
-8.23426068e-01 -9.59449887e-01 -1.44130707e-01 1.13283062e+00
-3.10720593e-01 -5.19289196e-01 -7.02427983e-01 -2.18551889e-01
-2.18555823e-01 5.56394398e-01 -3.70960563e-01 -3.02289486e-01
-8.76166701e-01 -6.21966422e-01 9.19417441e-01 6.69967771e-01
4.67401177e-01 -9.67840552e-01 -9.66609001e-01 7.66659796e-01
-1.61209211e-01 -1.08297515e+00 -6.10214233e-01 8.35268199e-01
-1.16362667e+00 -5.33411205e-01 -8.15615356e-01 -6.86028719e-01
-2.25019194e-02 2.05543101e-01 1.10374737e+00 -5.08792043e-01
-1.39518589e-01 2.68638998e-01 -4.51785654e-01 -2.87768334e-01
2.19346266e-02 3.70465100e-01 3.16851676e-01 1.36888862e-01
3.03028435e-01 -1.30272639e+00 -8.29893708e-01 2.37678140e-01
-9.80664372e-01 -3.64732683e-01 6.15924358e-01 6.45462453e-01
3.40458840e-01 2.11237460e-01 8.05615664e-01 -2.86562055e-01
8.61519098e-01 -4.41288322e-01 -7.87050575e-02 -1.73103556e-01
-3.86251152e-01 4.00810957e-01 7.63479710e-01 -7.38560319e-01
-1.05603755e+00 -3.05933654e-01 -3.23257059e-01 -8.01713228e-01
-1.25883117e-01 1.92710429e-01 1.24328978e-01 1.84002474e-01
7.81185210e-01 1.32844403e-01 -3.77299666e-01 -7.27616012e-01
2.69329101e-01 6.28283978e-01 4.00644600e-01 -6.04370177e-01
4.42703635e-01 5.71673810e-01 -2.65513450e-01 -1.07602632e+00
-5.45619905e-01 -5.70542753e-01 -5.72589993e-01 -6.96647912e-02
8.79139781e-01 -9.11419034e-01 -2.81165451e-01 3.76077801e-01
-1.12816954e+00 -4.48680609e-01 -5.69912732e-01 7.00039566e-01
-5.52673459e-01 -1.29455641e-01 -8.10064554e-01 -8.72778416e-01
-3.61479282e-01 -9.83634114e-01 8.41308832e-01 2.53822684e-01
-5.68838298e-01 -7.07926154e-01 2.28425011e-01 -2.60433733e-01
8.88249218e-01 -4.90477830e-02 7.48903215e-01 -6.83624685e-01
-5.25101304e-01 4.38909650e-01 2.47475840e-02 4.73347425e-01
3.90220322e-02 2.49101408e-02 -1.59869802e+00 -2.36234665e-01
1.78304285e-01 9.09573510e-02 1.08943939e+00 7.10599482e-01
8.40568364e-01 -3.49178553e-01 -1.31428093e-01 8.82799685e-01
1.07278621e+00 4.12216902e-01 4.39857155e-01 1.67683750e-01
3.00194442e-01 5.85210979e-01 2.15933666e-01 5.02659082e-01
-1.93239018e-01 6.22921705e-01 2.38091588e-01 -8.00086558e-02
-5.33960223e-01 -4.25234959e-02 5.44584155e-01 1.02280068e+00
-7.53021091e-02 -7.59747252e-02 -5.99959314e-01 7.96473742e-01
-1.63773179e+00 -1.06324208e+00 1.79162800e-01 1.98185575e+00
8.44742835e-01 4.11941975e-01 2.21484110e-01 4.64269102e-01
5.24964929e-01 5.27031541e-01 -3.44409108e-01 -4.84776229e-01
-1.37096703e-01 6.18479848e-01 2.73670703e-01 3.11836541e-01
-9.34334874e-01 6.34426534e-01 6.22860336e+00 1.04421115e+00
-1.48772669e+00 1.74009189e-01 4.28444892e-01 -7.88124740e-01
-1.95439681e-01 -1.66299790e-01 -7.47519493e-01 2.21753538e-01
1.25321746e+00 1.35097325e-01 4.89065260e-01 5.42982578e-01
2.51341671e-01 2.02661902e-01 -1.28576148e+00 9.87098157e-01
-1.95151135e-01 -1.32122469e+00 -3.06211233e-01 4.03854400e-02
4.80657816e-01 1.62991986e-01 3.26403707e-01 2.39283979e-01
1.06096171e-01 -9.22352910e-01 1.07259154e+00 3.38317752e-01
6.80668354e-01 -8.24549019e-01 1.93554536e-01 -5.65876365e-02
-1.55018723e+00 -2.77571648e-01 -1.34965166e-01 -1.79194957e-01
2.04766780e-01 7.70838022e-01 -7.84961402e-01 3.93583983e-01
9.31550920e-01 6.75428867e-01 -3.10093671e-01 9.54967558e-01
2.25639790e-01 7.34287500e-01 -3.80336255e-01 2.06527844e-01
5.27398944e-01 8.04893225e-02 7.99337924e-01 1.62311232e+00
4.65542734e-01 -1.34133935e-01 -2.44119644e-01 7.06679404e-01
7.75546208e-02 -1.90406084e-01 -5.82268298e-01 2.90595973e-03
3.59446883e-01 9.42863703e-01 -5.48718333e-01 -2.32010130e-02
-3.35544974e-01 6.80699706e-01 2.67348200e-01 5.57026207e-01
-7.23547220e-01 -4.45215881e-01 1.22773659e+00 3.91211390e-01
7.06911147e-01 -4.74883318e-01 -2.08128288e-01 -9.47821736e-01
-9.65513363e-02 -9.15793419e-01 3.09961587e-01 -4.92462307e-01
-1.09192169e+00 1.08661532e+00 -1.35745555e-01 -1.33470201e+00
-4.27420348e-01 -4.30889249e-01 -7.49806821e-01 9.66763079e-01
-1.62650907e+00 -6.21269107e-01 2.74677068e-01 7.51299739e-01
7.52333701e-01 -1.72505379e-01 5.78592837e-01 6.46320760e-01
-4.01840329e-01 6.41554117e-01 1.21021857e-02 4.94952351e-02
8.14418137e-01 -9.80114758e-01 4.45476800e-01 1.16864991e+00
4.65424567e-01 7.75531530e-01 7.99958646e-01 -2.48927340e-01
-9.01213288e-01 -1.09972942e+00 5.58236659e-01 -1.30231008e-01
6.79334581e-01 -3.66368949e-01 -7.84399509e-01 4.67835218e-01
4.73036498e-01 1.84709136e-03 8.21001232e-01 2.10666955e-01
-4.52292919e-01 -3.36819142e-01 -8.02185595e-01 6.09468520e-01
1.21543777e+00 -7.85585284e-01 -5.10728478e-01 -5.33749223e-01
6.31488323e-01 -1.10259550e-02 -6.75968230e-01 3.24379861e-01
6.79302514e-01 -1.32328820e+00 1.17422616e+00 -6.53928578e-01
4.54214633e-01 -1.10311769e-01 -3.64382029e-01 -1.41758990e+00
-5.00604212e-01 -8.17544580e-01 -3.62288237e-01 1.32333040e+00
4.31187838e-01 -4.09319639e-01 5.91436446e-01 5.63808046e-02
-3.35747957e-01 -8.36110890e-01 -1.12226808e+00 -9.04661834e-01
5.80818858e-03 -7.03385293e-01 5.48541367e-01 7.07489610e-01
4.65224963e-03 3.38860929e-01 -3.66112858e-01 8.17986205e-02
3.03959101e-01 6.40345216e-02 2.88306087e-01 -8.90944004e-01
-7.15124667e-01 -8.15456450e-01 -1.80076018e-01 -1.14918613e+00
-1.20745018e-01 -6.68154478e-01 -1.04856841e-01 -1.02437329e+00
-5.36386192e-01 -3.08634490e-01 -6.83378339e-01 1.49338514e-01
1.04007855e-01 2.46945918e-01 3.12066168e-01 1.41192421e-01
-2.80494720e-01 5.29099703e-01 1.01135468e+00 -5.50791398e-02
-5.38463295e-01 -2.41854712e-02 -5.84781587e-01 7.44729936e-01
9.95921135e-01 -4.85063106e-01 -5.93273818e-01 -7.17623115e-01
-9.45508853e-02 1.53296441e-01 2.45400161e-01 -1.49410403e+00
1.72652125e-01 2.64732778e-01 3.77308488e-01 -4.32568610e-01
6.50450110e-01 -8.13515544e-01 7.45924655e-03 1.72006935e-01
-4.57824826e-01 6.71678558e-02 4.22453791e-01 4.77217376e-01
-5.28406620e-01 -2.00903177e-01 8.55183601e-01 -1.53273150e-01
-4.91982490e-01 1.58493593e-01 -4.99719977e-01 6.66153729e-02
5.11759758e-01 -3.33398461e-01 1.23096012e-01 -3.07866275e-01
-8.70271027e-01 -2.23267853e-01 4.01426591e-02 4.81871367e-01
4.41128552e-01 -1.39748096e+00 -6.42512560e-01 2.82298595e-01
-3.22434157e-01 -2.52487481e-01 4.02207941e-01 8.08022916e-01
-7.66465738e-02 4.84347135e-01 -3.01080644e-01 -4.91709620e-01
-1.09846377e+00 4.33220416e-01 5.73735118e-01 -4.23856497e-01
-5.45448363e-01 1.12275243e+00 2.01837286e-01 1.68658718e-01
3.75019461e-01 -6.66444659e-01 -2.48130664e-01 4.10442293e-01
5.39077044e-01 5.05324841e-01 2.99893409e-01 -4.62230206e-01
-4.20279622e-01 4.19883341e-01 -4.79065161e-03 -5.72168708e-01
1.52213156e+00 -1.97877035e-01 3.36767167e-01 5.36074877e-01
1.32404327e+00 1.53640851e-01 -1.51455510e+00 -3.27073306e-01
5.54046780e-03 -3.02201599e-01 2.64350235e-01 -4.98389125e-01
-1.29902565e+00 1.15060890e+00 6.81639969e-01 3.24612260e-01
1.45802879e+00 -3.07562023e-01 8.23994756e-01 3.40457284e-03
3.42893191e-02 -1.06492484e+00 2.36803636e-01 5.27321339e-01
7.84536541e-01 -7.33194113e-01 -4.33371425e-01 -2.21760534e-02
-3.03819805e-01 1.20111215e+00 4.60692972e-01 -3.73205155e-01
8.66149426e-01 6.26292706e-01 7.21932277e-02 -1.14164643e-01
-8.00798237e-01 -4.04263318e-01 3.69282037e-01 4.53951955e-01
6.90150261e-01 -2.04075888e-01 1.54284686e-01 8.51476490e-01
-3.16236466e-01 -1.37356082e-02 2.65399933e-01 8.90060842e-01
-2.98643202e-01 -1.08646667e+00 -4.00075853e-01 4.96462733e-01
-7.44247377e-01 -4.05659378e-01 -7.81854242e-02 4.45744723e-01
1.32460013e-01 1.11859465e+00 2.03151181e-01 -4.92558002e-01
3.46579015e-01 3.63236874e-01 4.21857983e-01 -4.72857654e-01
-1.03935921e+00 4.92742300e-01 1.40950531e-01 -3.58882695e-01
-5.55835426e-01 -6.64097786e-01 -9.00931299e-01 -1.37103247e-02
-1.91016302e-01 6.59513324e-02 3.94897461e-01 7.73929417e-01
5.58758616e-01 1.19184589e+00 4.28636134e-01 -1.09696829e+00
-4.99299705e-01 -1.10054326e+00 -5.66575527e-01 1.90108985e-01
7.93105483e-01 -4.50711131e-01 -3.31670493e-01 1.06881835e-01] | [15.350550651550293, 5.572744846343994] |
a0b687a5-a800-4aab-9b9f-df14339a70ee | explainable-artificial-intelligence-on | 2305.07511 | null | https://arxiv.org/abs/2305.07511v1 | https://arxiv.org/pdf/2305.07511v1.pdf | eXplainable Artificial Intelligence on Medical Images: A Survey | Over the last few years, the number of works about deep learning applied to the medical field has increased enormously. The necessity of a rigorous assessment of these models is required to explain these results to all people involved in medical exams. A recent field in the machine learning area is explainable artificial intelligence, also known as XAI, which targets to explain the results of such black box models to permit the desired assessment. This survey analyses several recent studies in the XAI field applied to medical diagnosis research, allowing some explainability of the machine learning results in several different diseases, such as cancers and COVID-19. | ['Claudio Filipi Gonçalves dos Santos', 'Vitor Yukio Kondo', 'Danilo Xavier Silva', 'Francisco Alves de Souza Neto', 'Fabiana Cristina Queiroz de Oliveira Marucci', 'Jose Victor Nogueira Alves da Silva', 'Ana Claudia Akemi Matsuki de Faria', 'Nayara Rossi Brito da Silva', 'Vitor Lopes Fabris', 'Rodrigo Dória Villaça', 'Lucas Cesar Ferreira Domingos', 'Renata De Paris', 'Natalia Backhaus Pereira', 'Mateus Antonio Chinelatto', 'Felipe Souza Tânios', 'Jhessica Victoria Santos da Silva', 'Rodrigo Reis Arrais', 'Matteus Vargas Simão da Silva'] | 2023-05-12 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 2.47449845e-01 8.12308431e-01 -4.54907298e-01 -6.36362970e-01
-3.54936980e-02 1.03082672e-01 3.30649495e-01 1.18327901e-01
1.93626195e-01 9.53517497e-01 -5.04761562e-02 -7.54188061e-01
-7.92640090e-01 -7.34986961e-01 -7.22651958e-01 -6.07332349e-01
-1.36117697e-01 7.02551782e-01 -6.42900407e-01 -1.80165425e-01
3.34567912e-02 5.14694273e-01 -1.23382235e+00 6.78945482e-01
8.47361445e-01 1.00009906e+00 -1.93045929e-01 7.04675317e-01
-2.91193634e-01 9.65136707e-01 -7.52114296e-01 -3.31230223e-01
-4.12377059e-01 -5.94507575e-01 -1.21852696e+00 -2.04188973e-01
-9.38235410e-03 7.03719109e-02 -1.88813433e-01 7.05437779e-01
-1.68278180e-02 -2.81762958e-01 8.85155797e-01 -1.36588836e+00
-7.71930933e-01 7.76150525e-01 -1.93549171e-01 2.68060625e-01
1.51191443e-01 -6.38718978e-02 7.91773379e-01 -5.10730445e-01
5.57843387e-01 1.15055561e+00 6.59298718e-01 9.06570911e-01
-5.37286043e-01 -4.97125059e-01 -2.76923954e-01 4.85618681e-01
-9.86981630e-01 4.42544222e-01 7.43721724e-01 -2.87761152e-01
9.12314594e-01 7.98884392e-01 7.54014730e-01 1.27617109e+00
9.39921796e-01 8.57230961e-01 1.12083125e+00 -5.52598536e-01
9.09398645e-02 2.76642293e-01 7.86252916e-01 8.65656555e-01
4.80752438e-01 4.40841556e-01 -5.38642645e-01 2.18529463e-01
6.95310891e-01 2.68875122e-01 -3.71392556e-02 9.44038853e-02
-1.23814631e+00 1.03798890e+00 8.16533089e-01 7.98461676e-01
-2.69366920e-01 8.18326399e-02 3.76156151e-01 1.56229243e-01
3.80407006e-01 1.09929478e+00 -8.79181981e-01 1.70600682e-01
-7.12456703e-01 3.05655986e-01 4.88351226e-01 6.35641694e-01
3.34994376e-01 1.48157880e-01 2.57477928e-02 9.84142125e-02
3.42384636e-01 2.02812806e-01 7.97131121e-01 -4.29409921e-01
1.59463584e-01 8.79169762e-01 -4.00945961e-01 -1.01719058e+00
-1.00490773e+00 -9.09996629e-01 -1.44513881e+00 3.95821452e-01
4.37431991e-01 -1.23368017e-01 -1.15269017e+00 1.23915172e+00
1.36509404e-01 1.20160148e-01 1.77846700e-01 7.77510107e-01
1.52553880e+00 4.08045202e-01 1.08943269e-01 4.37613353e-02
1.42093921e+00 -9.12525952e-01 -1.16951966e+00 2.74029020e-02
7.32476294e-01 -4.98318434e-01 7.02933431e-01 7.57885754e-01
-1.05205452e+00 -7.78585732e-01 -1.09017479e+00 -1.40392825e-01
-7.01796949e-01 6.10808842e-02 1.29695857e+00 6.07183635e-01
-7.69409895e-01 8.14005196e-01 -1.03127646e+00 -2.92197287e-01
6.77693069e-01 7.91628599e-01 -2.30259225e-01 1.14952959e-01
-1.44145310e+00 1.22844827e+00 3.76787037e-01 2.17849374e-01
-7.30732858e-01 -1.07570469e+00 -7.44010210e-01 2.88894385e-01
-4.34536152e-02 -1.20339966e+00 9.36518192e-01 -8.87409389e-01
-8.37076366e-01 1.04988885e+00 -5.19104823e-02 -8.09414685e-01
4.83711958e-01 -1.24168515e-01 -4.96305466e-01 -1.35138810e-01
-3.30812633e-01 6.93825722e-01 3.94189090e-01 -1.09125650e+00
-5.41055977e-01 -4.90877420e-01 -5.89267025e-03 -3.26022804e-01
-5.54363579e-02 -3.53405327e-01 7.69160017e-02 -5.42303681e-01
7.61037809e-05 -9.82006371e-01 -3.37853581e-01 -9.00966376e-02
-9.83367324e-01 -4.66987222e-01 7.22145438e-01 -5.28411388e-01
8.84831607e-01 -1.73878169e+00 2.71224737e-01 1.89343750e-01
7.22842038e-01 2.26723120e-01 2.70035118e-01 4.57855649e-02
-6.98344171e-01 4.76210147e-01 -1.34974420e-01 -1.04906298e-02
-3.32799047e-01 4.62720364e-01 -2.36630231e-01 3.54280055e-01
4.26788151e-01 1.16294134e+00 -6.94424212e-01 -4.33489621e-01
5.04475415e-01 7.52379417e-01 -2.91075677e-01 2.73623973e-01
-2.44210407e-01 9.28508043e-01 -7.45169580e-01 6.34064138e-01
4.40455228e-01 -6.59881949e-01 -2.49415174e-01 3.81141752e-02
2.51242310e-01 1.05223782e-01 -4.09095526e-01 1.51030600e+00
-2.95381695e-01 8.87659550e-01 -4.07325178e-01 -1.47838092e+00
6.73226833e-01 6.69959545e-01 7.40657270e-01 -3.28810990e-01
4.49888259e-01 2.38864705e-01 5.40123105e-01 -7.09381461e-01
-1.40345708e-01 -3.39318842e-01 1.93816364e-01 2.54069269e-01
-1.21540934e-01 -6.82334155e-02 -3.24480653e-01 -3.46215703e-02
8.10305774e-01 -1.36426196e-01 6.84834898e-01 -3.75638872e-01
6.57633245e-01 4.71431911e-01 1.71828315e-01 7.76479602e-01
4.02695872e-03 5.88171899e-01 2.13500276e-01 -1.38518178e+00
-7.63516605e-01 -7.00005293e-01 -7.33546257e-01 2.57750660e-01
7.63652623e-02 8.10573921e-02 -9.28429067e-01 -8.20906460e-01
-5.69823794e-02 7.15331912e-01 -1.30779016e+00 -4.24315363e-01
-4.49892789e-01 -9.81688261e-01 4.48780596e-01 5.61272621e-01
2.32837945e-01 -1.69666624e+00 -6.76041663e-01 -4.02566008e-02
-8.59130099e-02 -5.19313753e-01 5.12878835e-01 4.96429861e-01
-1.13394845e+00 -1.18422830e+00 -5.99080801e-01 -6.08322263e-01
6.27921581e-01 -1.46191806e-01 1.46594155e+00 8.87076855e-01
-9.53919947e-01 1.13460861e-01 -3.52487445e-01 -1.19058526e+00
-7.50651300e-01 5.32245696e-01 -1.92204252e-01 -3.68338555e-01
5.72495580e-01 -1.06862508e-01 -4.42868143e-01 -1.38120754e-02
-1.05640543e+00 3.80416751e-01 6.90524936e-01 1.11666405e+00
5.74365675e-01 1.45196589e-02 4.23033565e-01 -1.16087270e+00
2.20391870e-01 -6.76558197e-01 1.72683839e-02 3.00108343e-01
-7.81768978e-01 2.19585031e-01 5.11304677e-01 -1.03516564e-01
-8.42937827e-01 -2.99905807e-01 -5.19345999e-01 1.57090966e-02
-6.02068663e-01 7.00143456e-01 5.48057212e-03 -9.88423750e-02
7.52840102e-01 -1.08223736e-01 -1.18758110e-02 -3.98327261e-01
1.36478245e-01 4.45335627e-01 4.96763438e-01 9.54801217e-02
7.62173116e-01 3.98427218e-01 5.57391942e-01 -5.91336310e-01
-1.39310646e+00 1.19022176e-01 -7.03812480e-01 -1.15361407e-01
1.32763529e+00 -1.72376379e-01 -1.06748724e+00 8.65544826e-02
-1.25921190e+00 7.62877911e-02 -2.85624564e-01 7.78074980e-01
-5.89736640e-01 -1.65633187e-01 -5.63695252e-01 -3.53825212e-01
-6.15196705e-01 -1.35557020e+00 8.33599925e-01 4.12244290e-01
-6.57606244e-01 -1.57099164e+00 -1.39103383e-01 4.19973820e-01
4.68275934e-01 5.11830211e-01 1.52400720e+00 -8.90913486e-01
-4.68976557e-01 -2.97522038e-01 -1.19744591e-01 -1.27420917e-01
2.55236179e-01 -1.13626979e-01 -1.16995895e+00 1.16747960e-01
3.58317882e-01 -1.33478954e-01 8.51358831e-01 9.94181454e-01
1.79767823e+00 -3.87136675e-02 -8.15036356e-01 7.00797677e-01
1.21723843e+00 4.34263259e-01 6.71831071e-01 4.09845144e-01
7.22455323e-01 6.08025253e-01 6.94168150e-01 5.83898555e-03
-8.57718885e-02 5.57832778e-01 1.00068057e+00 -9.28509414e-01
-2.43717000e-01 2.30367944e-01 -4.67457414e-01 5.49278557e-01
-3.58619958e-01 -2.19237819e-01 -1.11404467e+00 3.21077853e-01
-1.57756650e+00 -8.74558806e-01 -6.65489018e-01 1.46197069e+00
5.63778818e-01 1.32704601e-01 -3.86571199e-01 5.57272077e-01
3.83562297e-01 -3.43125522e-01 -4.47371453e-01 -7.26783991e-01
-1.11252542e-04 3.75736535e-01 -1.60077050e-01 5.30257106e-01
-1.08653843e+00 5.17542899e-01 7.88166571e+00 5.45909822e-01
-1.26069379e+00 -2.32519552e-01 1.37254894e+00 3.09480608e-01
-4.18583751e-01 -4.50421542e-01 -5.82309783e-01 1.26529276e-01
1.06882417e+00 -1.17663793e-01 9.34615210e-02 9.46809828e-01
2.65971571e-01 3.05665910e-01 -1.39030719e+00 8.61321807e-01
-5.38721010e-02 -1.80387414e+00 6.98893443e-02 1.93649217e-01
7.17709363e-01 -3.59168887e-01 4.37727839e-01 2.30092332e-01
-1.38967946e-01 -1.75782955e+00 1.16459720e-01 8.15784514e-01
6.25313461e-01 -7.94858396e-01 1.13826692e+00 2.03084916e-01
-3.84554029e-01 -2.99920724e-03 -4.57470357e-01 -3.19617718e-01
-2.90089369e-01 6.84696376e-01 -1.43066919e+00 8.66389453e-01
6.53543651e-01 6.69894338e-01 -5.15712440e-01 9.22058284e-01
-3.15814763e-01 9.60067034e-01 2.18315467e-01 -2.10343897e-01
3.15141410e-01 -5.69039397e-02 3.34906846e-01 1.04696941e+00
2.86830783e-01 -5.36839031e-02 -2.93299079e-01 1.03848588e+00
2.39785612e-01 -1.72965080e-02 -6.26203656e-01 1.52580798e-01
-3.39175582e-01 1.24019277e+00 -6.20998979e-01 -5.22200465e-01
-1.25422478e-01 6.88058019e-01 -5.10922521e-02 -2.70299297e-02
-1.17405117e+00 -1.91331673e-02 5.68062305e-01 1.12335071e-01
-3.43286663e-01 3.63461167e-01 -1.12642145e+00 -8.10766637e-01
-4.13411647e-01 -1.05460787e+00 4.67046112e-01 -9.66513991e-01
-1.16151524e+00 1.01094711e+00 1.35966130e-02 -1.10282660e+00
-6.25983715e-01 -1.03360474e+00 -6.16998136e-01 8.34675491e-01
-1.56153870e+00 -1.34313571e+00 -4.68673706e-01 3.82365048e-01
4.58548099e-01 -4.25214618e-01 1.22184265e+00 9.93318483e-02
-3.94716591e-01 4.13661271e-01 -2.30736628e-01 -1.45939980e-02
3.08127165e-01 -1.44112623e+00 2.53798544e-01 1.73450217e-01
3.90474975e-01 7.27499306e-01 9.84088361e-01 -4.32163954e-01
-1.11873329e+00 -9.53466117e-01 1.23439014e+00 -8.18455935e-01
1.30584478e-01 1.69518426e-01 -9.78855491e-01 7.56649673e-01
5.21432817e-01 -6.13254942e-02 8.05774391e-01 2.31864557e-01
2.55981892e-01 -1.69242658e-02 -1.08510363e+00 2.80460447e-01
6.05125308e-01 2.46162862e-01 -7.04491258e-01 6.36437654e-01
6.89607382e-01 -5.53403080e-01 -8.75879943e-01 6.21632576e-01
4.70386147e-01 -8.11254799e-01 1.04911470e+00 -1.40410435e+00
9.70475435e-01 1.11261688e-01 5.65151870e-01 -1.27942753e+00
-2.45142609e-01 -1.99499130e-01 -6.25237301e-02 5.03875434e-01
7.56560147e-01 -2.37628013e-01 1.12894416e+00 7.31887221e-01
-3.91020983e-01 -1.30186224e+00 -8.30783069e-01 -3.80307019e-01
3.86826605e-01 -4.70096469e-01 7.43006408e-01 1.09475625e+00
1.96982753e-02 4.33837995e-03 -4.80957091e-01 8.43758136e-02
2.24224404e-01 1.93828702e-01 6.24227762e-01 -1.63339245e+00
-1.76212341e-01 -3.75969499e-01 -6.20001197e-01 -3.13425869e-01
2.57894881e-02 -1.03417218e+00 -3.61495107e-01 -1.73811567e+00
3.22684020e-01 -2.03346148e-01 -4.70713258e-01 6.05057478e-01
-4.42544132e-01 1.70761526e-01 -2.62462012e-02 8.02598819e-02
-5.82769550e-02 1.71539247e-01 1.51558924e+00 -4.92158622e-01
-1.31003447e-02 4.54259783e-01 -7.72732139e-01 9.67843533e-01
7.80709326e-01 -6.67675912e-01 -3.56183589e-01 -2.85717189e-01
3.32238883e-01 1.64766163e-01 2.98608422e-01 -1.14656234e+00
-2.68585622e-01 -1.49398193e-01 9.02523696e-01 -6.87004745e-01
8.36679861e-02 -8.98346186e-01 2.82861441e-01 1.09181333e+00
-6.07772231e-01 5.23768589e-02 1.61210984e-01 1.23617664e-01
-5.09330988e-01 -3.32572877e-01 7.95533597e-01 -3.35673630e-01
-3.27028781e-01 3.05896491e-01 -3.48793000e-01 -3.63963306e-01
1.15331089e+00 1.61766000e-02 -2.09944353e-01 -4.13896412e-01
-1.20450914e+00 -1.30799353e-01 -3.02974522e-01 4.11690503e-01
6.24687314e-01 -1.04987967e+00 -6.75115407e-01 3.88828735e-03
9.57106426e-03 6.76696247e-04 3.41018856e-01 8.35975409e-01
-8.07860851e-01 1.19263256e+00 -4.75227386e-01 -6.10406160e-01
-1.39043510e+00 9.31509852e-01 6.85985506e-01 -8.59614909e-01
-7.73364484e-01 7.28871763e-01 3.22069854e-01 -3.87761265e-01
2.89979666e-01 -7.17538059e-01 -7.30663121e-01 -4.13214922e-01
7.24826396e-01 2.96545237e-01 1.84522539e-01 -1.93064839e-01
-3.78556073e-01 3.87164295e-01 -8.40012431e-02 3.88665706e-01
1.56932735e+00 2.88278401e-01 -1.05169505e-01 4.28379804e-01
1.00259244e+00 -6.16290510e-01 -3.61730784e-01 3.32908750e-01
5.74493036e-02 7.43481666e-02 -5.38925454e-02 -1.16118824e+00
-1.12809265e+00 1.40750003e+00 7.63435900e-01 4.37446713e-01
1.03530443e+00 4.34914194e-02 4.89505947e-01 2.52627939e-01
6.65589496e-02 -3.59710574e-01 -2.23952681e-02 7.49759525e-02
9.99638498e-01 -1.41529262e+00 1.41689122e-01 -5.52733600e-01
-5.67175329e-01 1.50714552e+00 5.08750021e-01 1.58237107e-02
7.96432376e-01 1.43419117e-01 2.65385419e-01 -6.88489139e-01
-5.03388703e-01 3.39664221e-01 8.24036121e-01 5.73410749e-01
8.24279130e-01 9.52059850e-02 -4.05098826e-01 7.72064924e-01
-3.86212856e-01 1.96584865e-01 4.84388620e-01 3.45103860e-01
-1.22090295e-01 -1.05007815e+00 -6.20234370e-01 5.41580200e-01
-8.42849612e-01 5.95584512e-02 -6.33691013e-01 1.25711751e+00
2.75043339e-01 8.70743632e-01 -5.91411442e-02 -1.46063507e-01
-4.56336960e-02 1.08211385e-02 3.07793796e-01 -4.77694631e-01
-7.75673747e-01 -2.96649516e-01 -1.75714374e-01 -4.00587469e-01
-4.97866869e-01 -2.25551054e-01 -1.68509150e+00 -1.95743099e-01
-2.29461178e-01 1.67903110e-01 8.92698467e-01 1.26964247e+00
-8.28005001e-02 1.16307986e+00 2.32034355e-01 -2.48447880e-01
-1.77722022e-01 -1.10421073e+00 -4.85704035e-01 5.19875526e-01
5.58974504e-01 -5.58670998e-01 -2.21095636e-01 4.00011279e-02] | [8.71788215637207, 5.740085601806641] |
ad73491f-3254-473d-adaa-738a88c0eb0b | a-comprehensive-survey-on-image-dehazing | 2106.03323 | null | https://arxiv.org/abs/2106.03323v5 | https://arxiv.org/pdf/2106.03323v5.pdf | A Comprehensive Survey and Taxonomy on Single Image Dehazing Based on Deep Learning | With the development of convolutional neural networks, hundreds of deep learning based dehazing methods have been proposed. In this paper, we provide a comprehensive survey on supervised, semi-supervised, and unsupervised single image dehazing. We first discuss the physical model, datasets, network modules, loss functions, and evaluation metrics that are commonly used. Then, the main contributions of various dehazing algorithms are categorized and summarized. Further, quantitative and qualitative experiments of various baseline methods are carried out. Finally, the unsolved issues and challenges that can inspire the future research are pointed out. A collection of useful dehazing materials is available at \url{https://github.com/Xiaofeng-life/AwesomeDehazing}. | ['DaCheng Tao', 'Jiuxin Cao', 'Jing Zhang', 'Jun Zhang', 'Wenqi Ren', 'Yuan Cao', 'Xiaofeng Cong', 'Jie Gui'] | 2021-06-07 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [-1.38735317e-03 -2.44761735e-01 -1.37640685e-01 -1.22541197e-01
-4.45186645e-01 -1.52374074e-01 5.03541231e-01 9.01144743e-02
-1.96211815e-01 7.47644544e-01 8.13803002e-02 -5.94223849e-02
-1.69900656e-01 -9.71619844e-01 -4.60726976e-01 -1.20266902e+00
-1.85580209e-01 -4.51621801e-01 3.22113752e-01 -2.67947823e-01
4.77173477e-01 5.03202081e-01 -1.59332359e+00 1.38743997e-01
8.33462834e-01 1.27547073e+00 -1.38721973e-01 6.02481365e-01
4.60972190e-01 8.68448853e-01 -1.02053356e+00 -4.63176847e-01
2.13215485e-01 -5.15157342e-01 -6.28080904e-01 -1.93385452e-01
4.72159475e-01 -8.63726974e-01 -1.19285965e+00 1.18421364e+00
8.13940048e-01 3.23010415e-01 1.05404913e+00 -1.23192489e+00
-1.05411220e+00 3.66779357e-01 -4.07536983e-01 4.29570913e-01
-3.04043859e-01 1.87560230e-01 6.95482314e-01 -8.49555135e-01
2.16091629e-02 6.23047173e-01 6.00410461e-01 7.05880404e-01
-6.21684015e-01 -1.03374875e+00 -2.98403800e-01 4.91712034e-01
-1.68437052e+00 -3.29133987e-01 1.14047158e+00 -3.53073448e-01
9.28845644e-01 1.02036431e-01 2.10337207e-01 1.05279148e+00
4.53470409e-01 6.89810514e-01 7.28880048e-01 -4.62384254e-01
2.08857074e-01 1.50301337e-01 2.20473096e-01 8.36144388e-01
5.28702915e-01 5.00595093e-01 -6.31577194e-01 3.27954516e-02
5.49212694e-01 -5.50448410e-02 -2.50261396e-01 -1.15417264e-01
-7.74447322e-01 8.04974377e-01 4.94955719e-01 4.30079967e-01
-1.51363686e-01 2.66609043e-01 4.18879092e-01 5.74128106e-02
6.87934816e-01 6.65128708e-01 -1.07998699e-02 2.73503661e-01
-9.64818180e-01 1.97927251e-01 3.01543474e-01 9.96829569e-01
8.16868484e-01 4.89283860e-01 -6.88338280e-02 9.17332768e-01
-7.27240592e-02 2.39494160e-01 1.51742220e-01 -9.17272091e-01
-1.07028462e-01 7.70661049e-03 -7.80261531e-02 -1.17958272e+00
-1.72621146e-01 -1.70295373e-01 -1.11471617e+00 4.02995288e-01
-1.51290745e-01 -3.90921831e-01 -9.43880141e-01 8.99094045e-01
-1.41655818e-01 2.34910399e-01 -2.47903660e-01 7.12814093e-01
1.23315537e+00 7.20529020e-01 -1.61498617e-02 1.25754252e-01
1.11255014e+00 -1.15811074e+00 -9.21072483e-01 7.42552653e-02
2.16257542e-01 -7.12934136e-01 7.40600646e-01 6.10294104e-01
-1.07162440e+00 -4.32149619e-01 -1.63020825e+00 -2.54173219e-01
-6.88408732e-01 3.83283943e-02 4.51129049e-01 7.93844700e-01
-1.05115521e+00 8.11306953e-01 -8.00689101e-01 -3.20802033e-01
9.06343699e-01 7.33023733e-02 2.04999641e-01 5.54195270e-02
-1.48261213e+00 7.98821032e-01 5.07953227e-01 3.41370180e-02
-1.42981434e+00 -7.32333362e-01 -6.86700046e-01 -2.16099545e-01
1.44112200e-01 -2.98684150e-01 1.24535358e+00 -5.12423933e-01
-1.49998236e+00 9.00895536e-01 1.71985775e-01 -6.48856997e-01
1.73956037e-01 -4.03941512e-01 -7.16591299e-01 5.04595816e-01
-4.66897309e-01 6.13827884e-01 9.82231975e-01 -1.43593276e+00
-6.37525499e-01 -3.26879174e-02 1.96190193e-01 -2.06663296e-01
-6.79204881e-01 3.90682399e-01 -2.66297907e-01 -1.28231680e+00
-5.02065599e-01 -2.74048716e-01 8.79144222e-02 2.35113546e-01
-5.75353324e-01 -3.78283560e-02 1.00315726e+00 -7.64139891e-01
1.80962026e+00 -2.43168259e+00 -1.57154769e-01 -1.43082943e-02
5.46976864e-01 6.70069098e-01 8.98197442e-02 8.67436528e-01
-1.82614774e-01 4.22648400e-01 -6.06432498e-01 -3.38707149e-01
-1.11072227e-01 -2.73340762e-01 -3.54751021e-01 7.56013811e-01
1.04689650e-01 7.81915367e-01 -6.36617601e-01 -2.81189382e-01
5.73881209e-01 7.14044333e-01 -9.47168097e-02 3.21607858e-01
-6.44453838e-02 2.92266756e-01 -1.28773674e-01 1.03293860e+00
8.56986880e-01 8.83567110e-02 -6.28898501e-01 -3.68247449e-01
-2.40030780e-01 2.58198738e-01 -6.64848387e-01 1.32308340e+00
-2.47847438e-01 8.94905746e-01 2.92779561e-02 -1.13077915e+00
8.36027682e-01 2.98474252e-01 4.45963770e-01 -6.16255641e-01
3.48000288e-01 3.27473283e-01 -4.47975159e-01 -6.36555493e-01
5.35063267e-01 -8.14800784e-02 1.65944830e-01 4.32369947e-01
2.99337089e-01 -1.09766491e-01 1.05994910e-01 1.37728825e-01
9.33129787e-01 -2.05975741e-01 2.37342402e-01 -5.06299853e-01
5.85314631e-01 8.29927251e-03 2.80585676e-01 6.23702824e-01
-6.00291491e-01 6.89621508e-01 1.23025544e-01 -6.77675068e-01
-1.04617953e+00 -1.03947735e+00 -3.46499979e-01 6.95874810e-01
7.16429472e-01 -5.40870547e-01 -9.65770543e-01 -5.11716902e-01
-1.72607720e-01 7.82988966e-01 -7.84934759e-01 -6.72555983e-01
-3.64208579e-01 -1.05035901e+00 9.45641696e-01 2.64269918e-01
1.13356066e+00 -1.02782965e+00 -3.18813264e-01 -1.44054607e-01
-1.74732581e-02 -6.82328165e-01 -2.60017484e-01 5.23080416e-02
-6.87271476e-01 -8.23841333e-01 -7.56632030e-01 -9.62152541e-01
3.90680492e-01 6.32993162e-01 8.75486016e-01 5.64697862e-01
-5.70306480e-01 1.71571761e-01 -4.92808551e-01 -7.77820945e-01
-3.35590810e-01 4.50332239e-02 1.90532371e-01 -2.27143019e-01
7.45646358e-01 -7.63110697e-01 -1.07580638e+00 2.78780192e-01
-1.27924299e+00 -2.16570571e-01 3.85449797e-01 4.80439454e-01
4.59907770e-01 7.00690508e-01 3.36682856e-01 -6.51981592e-01
4.76310194e-01 -4.21339422e-01 -5.13053894e-01 2.26165261e-02
-7.84048796e-01 -1.90373018e-01 4.28241223e-01 -2.17515662e-01
-1.02645171e+00 -6.22976184e-01 -3.99503380e-01 -3.23888242e-01
-5.33811688e-01 5.99771850e-02 -1.92587048e-01 -4.23697114e-01
1.03418887e+00 1.81802019e-01 5.32111228e-02 -5.53436935e-01
2.03137591e-01 8.61233056e-01 3.90306532e-01 -4.12189901e-01
1.05868971e+00 7.57460952e-01 -4.09251660e-01 -1.35239530e+00
-9.02832448e-01 -3.52173716e-01 -4.79922682e-01 -4.42335576e-01
1.02302039e+00 -7.16147721e-01 -1.02610372e-01 1.33363366e+00
-1.04651594e+00 -6.48988128e-01 -5.74230142e-02 2.67854303e-01
-3.03134710e-01 5.24789631e-01 -9.33655977e-01 -5.65001786e-01
-5.40989697e-01 -1.00027990e+00 8.23004067e-01 3.95841062e-01
-3.04526165e-02 -1.26189995e+00 7.10375747e-03 3.98092866e-01
6.75667107e-01 3.27706277e-01 9.39525843e-01 -3.77245605e-01
-5.66954553e-01 -1.49932429e-01 -1.86654061e-01 8.09679747e-01
3.65470171e-01 2.92476207e-01 -1.27511466e+00 -4.15944397e-01
1.49118289e-01 -3.07540357e-01 1.17775762e+00 5.07112563e-01
1.71323931e+00 -3.22735697e-01 -3.43791902e-01 1.17550480e+00
1.49969900e+00 3.59939188e-01 1.03841376e+00 4.96431619e-01
6.53212845e-01 4.99309599e-01 3.38103116e-01 3.31149071e-01
-2.10117802e-01 2.07732648e-01 6.50480390e-01 -2.08967075e-01
-3.70152295e-01 -6.38340414e-02 1.74786836e-01 9.01099145e-01
-2.08858714e-01 -6.56069279e-01 -8.42431128e-01 4.62023407e-01
-1.22925997e+00 -8.61251533e-01 1.05161518e-01 1.89675295e+00
5.62047362e-01 -2.97670922e-04 1.39495000e-01 2.19848141e-01
1.02704692e+00 5.75921357e-01 -6.97170556e-01 -1.69997051e-01
-2.29762807e-01 4.05146211e-01 8.08939815e-01 4.29236501e-01
-1.67511225e+00 9.45986927e-01 7.06485033e+00 1.12973177e+00
-1.03444314e+00 -1.61857363e-02 6.28895760e-01 -1.47267848e-01
-1.88213177e-02 -3.82399142e-01 -6.49756968e-01 5.52368581e-01
9.57947731e-01 -3.12521845e-01 4.83566374e-01 5.26180804e-01
9.67726260e-02 -8.03620741e-02 -6.05059505e-01 1.01100826e+00
2.56220490e-01 -1.60991478e+00 1.49440527e-01 -3.78852598e-02
7.73618340e-01 1.62340999e-01 3.50368440e-01 -5.47230728e-02
1.58386379e-01 -9.37435269e-01 5.08848131e-01 4.09489185e-01
1.05591106e+00 -1.00821924e+00 7.31133163e-01 -3.11130472e-02
-9.60776567e-01 -3.65943974e-03 -4.77964759e-01 -4.74396162e-03
-1.18239820e-01 7.85485208e-01 7.01090544e-02 7.62624204e-01
1.27996147e+00 1.04570234e+00 -5.21859825e-01 1.13630629e+00
-6.51479959e-01 9.39292967e-01 7.09846169e-02 4.55471985e-02
1.37407616e-01 -2.04261005e-01 3.81476820e-01 1.28597379e+00
1.79457933e-01 2.09772587e-01 -3.96030754e-01 8.89735401e-01
-1.63244456e-01 -2.97641546e-01 -6.15751207e-01 -2.21899390e-01
5.98897755e-01 1.18365753e+00 -4.40550983e-01 -1.77833647e-01
-4.88508493e-01 8.77978086e-01 -2.36802793e-04 6.05281532e-01
-1.02307642e+00 -1.27938879e+00 1.07289815e+00 8.74364227e-02
-8.85123312e-02 -4.85067487e-01 -2.35385671e-01 -1.00147831e+00
-1.72880650e-01 -7.05597997e-01 3.17938805e-01 -6.42695069e-01
-1.50841570e+00 3.46522391e-01 3.20989043e-01 -1.19537771e+00
5.59160471e-01 -9.39871907e-01 -8.54572058e-01 5.82432747e-01
-1.61937284e+00 -7.22334623e-01 -6.04150593e-01 2.45345414e-01
4.99586254e-01 -3.81966799e-01 6.26587451e-01 6.18708730e-01
-9.24150825e-01 7.30834782e-01 4.10753101e-01 4.71434861e-01
9.00340199e-01 -7.76193142e-01 5.31157374e-01 1.07624078e+00
-2.07743645e-01 3.75275373e-01 6.41836286e-01 -4.44175333e-01
-8.59259367e-01 -1.27519441e+00 3.74490529e-01 -2.28086382e-01
8.30748737e-01 -3.98077458e-01 -1.24573815e+00 3.55139613e-01
6.39610946e-01 -2.72316903e-01 7.61336684e-01 -7.26076484e-01
-2.58690715e-01 -1.10120848e-01 -1.12565053e+00 6.48578882e-01
1.13371062e+00 -5.99771559e-01 -2.11214885e-01 3.13991100e-01
7.58058906e-01 -1.66343793e-01 -7.43090093e-01 3.45742196e-01
2.65965223e-01 -1.12618303e+00 1.02284348e+00 -3.58900577e-01
6.26415908e-01 -5.83683178e-02 1.30397692e-01 -1.11942506e+00
-4.28170770e-01 -8.40787828e-01 -4.25696552e-01 9.83469605e-01
1.99107230e-01 -7.42169917e-01 7.98706710e-01 1.86970383e-01
-2.24641785e-01 -8.58458698e-01 -4.91998762e-01 -8.61199260e-01
6.11020505e-01 -1.93373621e-01 5.44548154e-01 9.27741110e-01
-1.38916507e-01 -8.30763280e-02 -6.44280016e-01 3.70666593e-01
9.09251988e-01 -4.70448673e-01 3.11483830e-01 -8.17029357e-01
1.15799099e-01 -4.75344718e-01 -4.13935363e-01 -8.91816258e-01
4.32342961e-02 -6.70429051e-01 -3.05535346e-02 -1.61212397e+00
-1.01692058e-01 -8.18306804e-02 -5.42314351e-01 2.79962242e-01
-1.62492499e-01 4.04379785e-01 -3.40656102e-01 3.46578151e-01
-2.78549194e-01 7.03220129e-01 9.54426765e-01 -3.27187985e-01
1.62656948e-01 -1.50757894e-01 -7.58016109e-01 6.00463748e-01
1.66005731e+00 -3.82848710e-01 -3.47667694e-01 -8.00572574e-01
-2.24396214e-01 -7.93613970e-01 4.10628498e-01 -1.23805118e+00
2.19965011e-01 -1.50788933e-01 1.15625687e-01 -4.84327704e-01
2.18724713e-01 -6.44445896e-01 -2.01091304e-01 2.73615569e-01
-3.05354863e-01 -8.62935036e-02 2.64739692e-01 3.90801877e-01
-4.33678061e-01 -4.22237813e-01 1.24665105e+00 -1.03625320e-02
-8.83079410e-01 7.46538997e-01 -6.22355878e-01 -4.69196625e-02
1.22194576e+00 -1.59090295e-01 -5.06873429e-01 -3.62292111e-01
-2.53422737e-01 8.36661011e-02 4.86760765e-01 3.09576422e-01
7.89159715e-01 -1.29334950e+00 -5.92477202e-01 1.59421936e-01
1.31750599e-01 2.59932708e-02 4.69441712e-01 5.38896203e-01
-1.05773795e+00 2.93572694e-01 -3.08128417e-01 1.05188610e-02
-1.03881335e+00 8.80222499e-01 7.84016192e-01 3.88509005e-01
-5.78865469e-01 1.06008720e+00 3.09862256e-01 1.20739594e-01
5.01656175e-01 1.29302651e-01 -1.03760689e-01 -3.49747151e-01
8.36573303e-01 7.62552559e-01 2.82647282e-01 -4.08468515e-01
-1.12551317e-01 3.61504048e-01 -2.35635847e-01 2.09261268e-01
1.27781606e+00 -1.24463245e-01 -3.12326670e-01 8.30970183e-02
1.33845997e+00 -2.27909312e-01 -1.36932766e+00 -2.39650339e-01
-2.16665700e-01 -4.10988986e-01 2.61620820e-01 -4.11391348e-01
-1.38165569e+00 1.37372971e+00 6.22340918e-01 1.56642959e-01
1.44714653e+00 7.41976351e-02 1.00394630e+00 3.56303275e-01
2.65217483e-01 -1.14517605e+00 3.93960595e-01 3.35593462e-01
8.62997234e-01 -9.98190343e-01 2.56151974e-01 -5.09095609e-01
-4.13376629e-01 9.55019832e-01 7.59954035e-01 -3.75285208e-01
1.23715401e+00 4.02921557e-01 2.15245411e-01 -3.73821735e-01
-3.38419676e-01 8.16623941e-02 -1.94722004e-02 9.29777086e-01
3.66186917e-01 -2.93363690e-01 -4.03235927e-02 3.02936882e-01
-1.29405677e-01 -3.01849306e-01 4.31199372e-01 1.10361898e+00
-6.03957534e-01 -1.01270807e+00 -1.88126236e-01 4.52227265e-01
-5.34378648e-01 -1.09852493e-01 -4.91258085e-01 6.54117048e-01
1.65607706e-01 1.02528822e+00 2.50950293e-03 -7.64898419e-01
2.19326481e-01 -2.09336370e-01 2.71865398e-01 -5.21244466e-01
-2.41083369e-01 -2.45296404e-01 -5.91704138e-02 -2.78020233e-01
-5.58989227e-01 -1.54110879e-01 -1.00852132e+00 -6.38031900e-01
-3.03805530e-01 2.69763451e-02 3.18223506e-01 5.65043569e-01
3.49451184e-01 6.24721408e-01 5.66446185e-01 -8.46797466e-01
-1.89592466e-01 -8.37564349e-01 -7.88373351e-01 -6.97886944e-02
6.40783370e-01 -7.98628032e-01 -9.29981411e-01 2.21014783e-01] | [10.939800262451172, -3.0624380111694336] |
60a67bdf-bdba-484e-bd82-07f633a1a94a | mplm-sim-unveiling-better-cross-lingual | 2305.13684 | null | https://arxiv.org/abs/2305.13684v1 | https://arxiv.org/pdf/2305.13684v1.pdf | mPLM-Sim: Unveiling Better Cross-Lingual Similarity and Transfer in Multilingual Pretrained Language Models | Recent multilingual pretrained language models (mPLMs) have been shown to encode strong language-specific signals, which are not explicitly provided during pretraining. It remains an open question whether it is feasible to employ mPLMs to measure language similarity, and subsequently use the similarity results to select source languages for boosting cross-lingual transfer. To investigate this, we propose mPLM-Sim, a new language similarity measure that induces the similarities across languages from mPLMs using multi-parallel corpora. Our study shows that mPLM-Sim exhibits moderately high correlations with linguistic similarity measures, such as lexicostatistics, genealogical language family, and geographical sprachbund. We also conduct a case study on languages with low correlation and observe that mPLM-Sim yields more accurate similarity results. Additionally, we find that similarity results vary across different mPLMs and different layers within an mPLM. We further investigate whether mPLM-Sim is effective for zero-shot cross-lingual transfer by conducting experiments on both low-level syntactic tasks and high-level semantic tasks. The experimental results demonstrate that mPLM-Sim is capable of selecting better source languages than linguistic measures, resulting in a 1%-2% improvement in zero-shot cross-lingual transfer performance. | ['Hinrich Schütze', 'André F. T. Martins', 'Zheyu Zhang', 'Chengzhi Hu', 'Peiqin Lin'] | 2023-05-23 | null | null | null | null | ['zero-shot-cross-lingual-transfer', 'open-question', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.82847917e-01 -3.94536167e-01 -4.91405308e-01 -6.55945063e-01
-1.25226927e+00 -6.71464264e-01 8.70206356e-01 3.82418007e-01
-9.89024758e-01 6.71342969e-01 4.76526648e-01 -1.91245958e-01
1.61291882e-01 -7.18555033e-01 -8.94843757e-01 -3.34885836e-01
-7.26501718e-02 5.15680671e-01 2.75412709e-01 -5.22232115e-01
3.09372824e-02 1.03807516e-01 -1.14043260e+00 6.71696126e-01
1.06651640e+00 4.34635997e-01 3.42845201e-01 1.84424117e-01
-4.87398118e-01 4.47198570e-01 -3.00934672e-01 -5.89549005e-01
1.29611626e-01 -6.36425853e-01 -9.45613086e-01 -6.01361454e-01
7.72662938e-01 3.46829027e-01 2.36233070e-01 1.11874282e+00
5.24920523e-01 1.22438021e-01 8.93793941e-01 -8.01833093e-01
-8.61471891e-01 1.18274260e+00 -5.48703015e-01 4.08334702e-01
3.80642235e-01 2.83104889e-02 1.35582161e+00 -1.03217137e+00
7.52651632e-01 1.77080655e+00 9.69788790e-01 3.60564411e-01
-1.51541150e+00 -8.87131155e-01 5.97081259e-02 1.85685888e-01
-1.41629708e+00 -4.34028208e-01 6.20492697e-01 -1.99078918e-01
1.08019590e+00 -2.60078851e-02 6.15035892e-02 1.12118232e+00
3.20922375e-01 8.27281296e-01 1.61180460e+00 -6.93347454e-01
-1.20703854e-01 5.18569767e-01 2.67667145e-01 7.01269805e-01
-1.05635952e-02 4.57055978e-02 -6.93099260e-01 2.04260871e-02
4.53502238e-01 -6.14900410e-01 -3.07072643e-02 -1.69636980e-01
-1.38227344e+00 1.16313589e+00 7.23767817e-01 1.03263295e+00
-7.38535672e-02 -4.54487763e-02 7.95644045e-01 8.14106703e-01
6.53519809e-01 6.53456867e-01 -4.75782424e-01 2.23448828e-01
-8.27105045e-01 -1.10806212e-01 5.18817365e-01 7.36676037e-01
9.88718033e-01 1.60122328e-02 -1.12836681e-01 1.38494778e+00
-5.60271135e-03 5.14283299e-01 9.01386678e-01 -4.73302960e-01
5.88531613e-01 2.72383958e-01 -5.81671178e-01 -7.14094520e-01
-4.59195554e-01 -2.30216339e-01 -5.79006016e-01 -2.30683193e-01
2.52723336e-01 -1.09648287e-01 -4.34518009e-01 2.32920432e+00
-5.41868061e-02 -1.17334640e-02 3.23592246e-01 6.19437575e-01
7.75688708e-01 6.80257618e-01 5.61712146e-01 -2.24497896e-02
1.42937338e+00 -8.80248070e-01 -1.90400735e-01 -5.11124194e-01
1.30511415e+00 -9.16667104e-01 1.66366351e+00 -2.53148794e-01
-8.73356044e-01 -7.60699511e-01 -7.88548291e-01 -1.37102112e-01
-6.49753034e-01 -1.50419042e-01 6.40728593e-01 4.32588190e-01
-1.08629620e+00 5.54301739e-01 -4.47117031e-01 -7.72574842e-01
-2.51575038e-02 1.58178747e-01 -5.26113272e-01 -2.66917706e-01
-1.71428406e+00 1.35822356e+00 7.32665002e-01 -7.59756565e-01
-4.53366458e-01 -9.07742083e-01 -1.10741460e+00 -4.27501462e-02
-2.74704278e-01 -5.42234480e-01 9.01304603e-01 -1.18676603e+00
-1.18641126e+00 1.35323870e+00 -6.56020865e-02 -4.74224895e-01
6.57447875e-02 1.04852907e-01 -4.53172177e-01 -2.14702055e-01
4.95385438e-01 9.26874697e-01 1.54703736e-01 -1.06320417e+00
-7.18046248e-01 -2.29715198e-01 -1.96214557e-01 3.54675502e-01
-6.44667089e-01 4.91431773e-01 -1.02876253e-01 -7.61751652e-01
-3.76389831e-01 -9.10571098e-01 4.92839841e-03 -4.88259971e-01
3.01848173e-01 -5.96697509e-01 1.45961000e-02 -5.72718799e-01
9.44423735e-01 -1.90697384e+00 8.15976486e-02 1.42668322e-01
-5.65329790e-01 2.46218681e-01 -7.32254446e-01 4.63050336e-01
3.48015204e-02 1.02486774e-01 -2.82818139e-01 -1.82048902e-01
8.49911477e-03 1.84755474e-01 -1.20562375e-01 3.23036253e-01
1.29144117e-01 1.12478316e+00 -1.05612874e+00 -6.45107627e-01
9.08061303e-03 4.31555152e-01 -6.15066528e-01 -1.51932031e-01
3.85307595e-02 2.71965504e-01 -3.06445397e-02 2.99125642e-01
3.65313023e-01 3.09101790e-02 2.64154255e-01 -1.59804925e-01
-8.30345005e-02 7.79209435e-01 -5.43378532e-01 1.90767992e+00
-1.10861802e+00 5.35843194e-01 -2.20434800e-01 -1.01237881e+00
9.85742629e-01 1.20807916e-01 9.83455628e-02 -9.90686476e-01
1.28436834e-01 5.99412501e-01 4.13286209e-01 -2.48082839e-02
4.62580264e-01 -6.95679307e-01 -4.94507849e-01 5.67748129e-01
6.19847953e-01 -1.35938719e-01 2.65038073e-01 1.53727969e-02
7.38015652e-01 -4.54128459e-02 4.41755682e-01 -9.09767747e-01
7.56553113e-01 2.81529743e-02 2.75557041e-01 6.34837449e-01
-1.14049964e-01 1.37770578e-01 -9.25591066e-02 -2.25423463e-02
-8.28037858e-01 -1.34855723e+00 -4.30475175e-01 1.76432323e+00
-7.93912262e-02 -2.41146907e-01 -6.45750761e-01 -7.73626864e-01
1.37857154e-01 1.07596910e+00 -3.80963564e-01 -2.96052366e-01
-6.71885014e-01 -9.17248905e-01 7.87338853e-01 4.61220950e-01
3.58220458e-01 -1.18324494e+00 1.15029953e-01 8.64149854e-02
-3.17279547e-01 -1.12417996e+00 -6.58305883e-01 2.16587737e-01
-7.46334016e-01 -4.42917317e-01 -8.17169845e-01 -1.37543261e+00
3.14799547e-01 2.41695464e-01 1.50456595e+00 -2.63509750e-01
1.13033578e-01 2.22355187e-01 -2.47748733e-01 -1.76993638e-01
-7.33044803e-01 5.29155731e-01 4.04424697e-01 -2.35867649e-01
8.75131071e-01 -5.64818263e-01 -5.09337112e-02 2.56291091e-01
-4.82259184e-01 -2.30299607e-01 4.95430738e-01 7.84567058e-01
4.75192130e-01 -3.51255208e-01 9.27595317e-01 -8.90177488e-01
8.88468981e-01 -5.60028374e-01 -2.53941566e-01 4.84054148e-01
-3.85428816e-01 3.55363071e-01 8.53994548e-01 -5.48057556e-01
-1.02557385e+00 -3.35359335e-01 -2.23922253e-01 -1.44079328e-01
-1.52274936e-01 6.91746294e-01 1.16177641e-01 -1.24416672e-01
8.54136586e-01 1.52606979e-01 -2.12559000e-01 -4.95736957e-01
5.65248430e-01 7.76956320e-01 4.54085529e-01 -9.23793495e-01
5.24105728e-01 1.05596431e-01 -5.22523224e-01 -9.27483022e-01
-1.19091165e+00 -5.09031832e-01 -8.39357555e-01 1.73641816e-01
9.34351683e-01 -1.20465112e+00 -2.82684416e-01 3.22089717e-02
-1.12563646e+00 -3.31530422e-01 8.38489830e-03 8.26114476e-01
-4.37454373e-01 3.78240421e-02 -9.07788217e-01 -1.95114166e-01
-4.30497736e-01 -9.85140145e-01 9.35965896e-01 -2.16961518e-01
-6.40593648e-01 -1.73780453e+00 3.78740221e-01 2.36247525e-01
4.32849795e-01 -2.68708408e-01 1.32621086e+00 -9.96110260e-01
-6.65227398e-02 2.48645410e-01 -1.54978558e-01 3.98460001e-01
3.49289030e-01 -5.76744020e-01 -8.40061128e-01 -5.64527690e-01
-2.07036749e-01 -6.74815476e-01 9.63417232e-01 2.28088275e-01
4.89697993e-01 -7.85802156e-02 -1.53096318e-01 6.82919443e-01
1.40917301e+00 -1.91926196e-01 1.83039397e-01 4.51989532e-01
6.91187680e-01 8.58612657e-01 5.35854638e-01 -4.38984513e-01
6.22251749e-01 8.19663048e-01 -4.38297540e-01 -9.64637548e-02
-4.03585792e-01 -4.93328005e-01 9.98665810e-01 1.77129781e+00
2.39500076e-01 3.38203385e-02 -1.06835043e+00 8.32457483e-01
-1.38848305e+00 -8.52495492e-01 1.68507807e-02 2.24135947e+00
1.29129672e+00 4.34280671e-02 1.37614593e-01 -4.89980221e-01
7.28258729e-01 1.38930455e-01 -1.97896421e-01 -8.86189222e-01
-3.85495931e-01 4.83438522e-01 5.04689336e-01 7.83524930e-01
-9.42772686e-01 1.74281585e+00 6.49273252e+00 1.14055312e+00
-1.31288791e+00 4.08009648e-01 3.90689045e-01 1.43183753e-01
-4.83173758e-01 -1.36149108e-01 -1.00843811e+00 5.04247129e-01
1.22750652e+00 -5.57984710e-01 2.83721089e-01 6.13223732e-01
-1.98866606e-01 1.94144443e-01 -1.38035452e+00 7.55511165e-01
3.43534321e-01 -1.03092575e+00 4.13212210e-01 -2.23769754e-01
8.98393929e-01 6.79462373e-01 -6.88392520e-02 6.73880517e-01
6.65560126e-01 -9.69293296e-01 4.74212438e-01 7.87937641e-03
7.73886740e-01 -9.55433071e-01 6.35743618e-01 3.57217103e-01
-1.23951757e+00 3.68521750e-01 -6.29244804e-01 1.35869682e-01
2.12963656e-01 8.60108063e-02 -9.50642407e-01 5.09103477e-01
5.12468994e-01 7.20696211e-01 -7.48344004e-01 4.85189170e-01
-2.28336930e-01 7.20450282e-01 -1.52555302e-01 -4.84811626e-02
6.03901625e-01 -8.65001157e-02 2.82268703e-01 1.74678671e+00
3.86576444e-01 -5.18275678e-01 3.73632729e-01 7.62088597e-01
-1.64107561e-01 8.38908374e-01 -8.59710455e-01 3.84050701e-03
4.19248313e-01 9.01224017e-01 -4.71524686e-01 -4.18212950e-01
-7.00157166e-01 1.08685350e+00 7.89287865e-01 4.55051251e-02
-6.39944196e-01 -2.83440173e-01 8.27875078e-01 5.04039675e-02
-4.07722928e-02 -2.80347377e-01 -1.86643735e-01 -1.16690183e+00
-4.48444039e-01 -8.91331553e-01 6.07978821e-01 -4.58033472e-01
-1.98664904e+00 7.08120227e-01 3.95554677e-03 -9.20995057e-01
-4.16570514e-01 -7.30896354e-01 -4.80901212e-01 1.15034533e+00
-1.54682171e+00 -1.28380382e+00 3.69263530e-01 7.05357850e-01
6.38731718e-01 -5.16365826e-01 9.54962850e-01 3.91469419e-01
-1.72227845e-01 1.03399742e+00 9.90986973e-02 2.28716329e-01
1.14589393e+00 -1.06520760e+00 5.37950099e-01 6.13578618e-01
6.06718719e-01 8.15878749e-01 4.13429290e-01 -5.04100919e-01
-9.05792356e-01 -1.20701551e+00 1.42828405e+00 -4.28547144e-01
1.07207215e+00 -3.98505211e-01 -1.27666008e+00 7.69469559e-01
5.12752414e-01 -2.22396180e-01 9.06847417e-01 5.35147488e-01
-7.46681452e-01 -8.38170946e-02 -9.07655120e-01 7.43191242e-01
1.11032712e+00 -1.03358281e+00 -1.00520349e+00 3.96215260e-01
8.59808564e-01 2.04187125e-01 -1.12995994e+00 4.23138022e-01
2.80810326e-01 -6.81856930e-01 1.05573964e+00 -8.01200628e-01
4.39172864e-01 1.46111533e-01 -3.68486434e-01 -1.94509363e+00
-5.87076247e-01 5.16382046e-02 8.95987391e-01 1.53835738e+00
7.56870329e-01 -7.84756064e-01 6.67069331e-02 -1.40264511e-01
-1.13805227e-01 -3.39716405e-01 -8.46751988e-01 -1.37010419e+00
9.51451957e-01 -4.04306889e-01 2.36154452e-01 1.57431352e+00
1.35757998e-01 9.60331619e-01 -8.64991024e-02 -8.14121440e-02
5.65196872e-01 1.40427724e-01 4.58408952e-01 -1.02412462e+00
-2.67830282e-01 -7.12150753e-01 -2.15789258e-01 -6.98402166e-01
9.73588049e-01 -2.05558896e+00 1.57319054e-01 -1.30692101e+00
3.02541733e-01 -5.96126854e-01 -7.37309635e-01 4.11766559e-01
-2.68460900e-01 2.82460153e-01 2.93570369e-01 3.73959988e-02
-4.38444167e-01 4.94207859e-01 9.07379270e-01 -7.17106834e-02
-1.58286780e-01 -4.12452132e-01 -6.46761060e-01 8.88739944e-01
8.37193310e-01 -5.25422215e-01 -2.12567165e-01 -7.44157493e-01
-9.87441316e-02 -3.58594656e-01 -1.00167885e-01 -8.43286693e-01
-8.46100226e-02 -1.68512031e-01 7.98474923e-02 -6.99352175e-02
1.33659333e-01 -3.31314325e-01 -3.60016286e-01 4.44466382e-01
-6.55523360e-01 2.47405618e-01 3.76340419e-01 -8.87350738e-02
-4.42434102e-01 -2.82239288e-01 1.09894526e+00 -3.02710861e-01
-8.40533018e-01 3.58358249e-02 -2.67366588e-01 5.68417370e-01
6.14265323e-01 2.51102269e-01 -1.91916987e-01 -5.35819605e-02
-2.29079932e-01 2.02425987e-01 5.97152889e-01 9.44042325e-01
4.39182334e-02 -1.76914692e+00 -1.11857176e+00 3.14545780e-02
4.60844547e-01 -8.56974244e-01 -1.38152137e-01 7.23567426e-01
-1.06167540e-01 6.13617957e-01 -3.44885439e-01 -6.67904913e-01
-1.36203992e+00 5.59672892e-01 1.77997217e-01 -3.02312315e-01
-3.14247072e-01 1.02938318e+00 5.71268976e-01 -1.08060360e+00
-1.39426202e-01 -1.91398740e-01 -5.87438792e-02 2.36346051e-01
3.75329345e-01 3.30171920e-02 -2.83803288e-02 -1.15704131e+00
-5.88896632e-01 8.75638664e-01 -1.55117556e-01 -3.27850223e-01
1.30677080e+00 -1.24606863e-01 -1.83424696e-01 9.39853013e-01
1.61263597e+00 2.06847876e-01 -4.18686628e-01 -7.02716112e-01
5.21817386e-01 -1.47668868e-01 -1.61827713e-01 -7.24863112e-01
-7.18049467e-01 9.31385636e-01 4.13857907e-01 -3.19497138e-01
7.38523722e-01 2.74300307e-01 9.27640676e-01 2.58606344e-01
7.60772169e-01 -1.09093618e+00 -8.91954005e-02 9.39379454e-01
8.35718036e-01 -1.39660716e+00 -3.63260061e-01 -1.85615093e-01
-7.60212183e-01 5.52348375e-01 6.42370999e-01 -2.08268970e-01
4.17191267e-01 1.42499864e-01 2.16778636e-01 -1.93200205e-02
-7.78842390e-01 -4.09392655e-01 6.31341994e-01 3.56036037e-01
1.12321210e+00 3.25217038e-01 -6.89598083e-01 5.20870388e-01
-6.62007391e-01 -4.66908783e-01 -8.64393935e-02 5.27060032e-01
-5.48399329e-01 -1.37345397e+00 -2.95919925e-01 1.71658680e-01
-3.77783120e-01 -5.90663731e-01 -5.10764241e-01 8.13839555e-01
2.12273180e-01 5.23218274e-01 3.29051048e-01 -3.30824465e-01
1.51898265e-01 5.09204686e-01 7.87208259e-01 -9.08358097e-01
-8.62935543e-01 -2.43321151e-01 2.49265254e-01 -3.62858504e-01
-4.15171176e-01 -6.35453522e-01 -1.32920384e+00 -1.66180953e-01
-7.50243142e-02 2.01389179e-01 5.17436504e-01 8.92045557e-01
8.30547884e-02 2.36418143e-01 5.04224300e-01 -5.37311375e-01
-4.24939126e-01 -1.09779119e+00 -2.94273645e-01 6.86002314e-01
-7.21965954e-02 -5.48443794e-01 -3.55794966e-01 -1.95642993e-01] | [10.905125617980957, 9.911439895629883] |
e6a56e09-5def-4eff-842e-4b5494d5e00a | combining-referring-expression-generation-and | null | null | https://aclanthology.org/P13-1152 | https://aclanthology.org/P13-1152.pdf | Combining Referring Expression Generation and Surface Realization: A Corpus-Based Investigation of Architectures | null | ['Sina Zarrie{\\ss}', 'Jonas Kuhn'] | 2013-08-01 | null | null | null | acl-2013-8 | ['referring-expression-generation'] | ['computer-vision'] | [-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] | [-7.347914695739746, 3.742560625076294] |
70c1d6f9-557a-4cdb-b145-146d69c7600d | an-exploration-of-encoder-decoder-approaches | 2305.05627 | null | https://arxiv.org/abs/2305.05627v1 | https://arxiv.org/pdf/2305.05627v1.pdf | An Exploration of Encoder-Decoder Approaches to Multi-Label Classification for Legal and Biomedical Text | Standard methods for multi-label text classification largely rely on encoder-only pre-trained language models, whereas encoder-decoder models have proven more effective in other classification tasks. In this study, we compare four methods for multi-label classification, two based on an encoder only, and two based on an encoder-decoder. We carry out experiments on four datasets -two in the legal domain and two in the biomedical domain, each with two levels of label granularity- and always depart from the same pre-trained model, T5. Our results show that encoder-decoder methods outperform encoder-only methods, with a growing advantage on more complex datasets and labeling schemes of finer granularity. Using encoder-decoder models in a non-autoregressive fashion, in particular, yields the best performance overall, so we further study this approach through ablations to better understand its strengths. | ['Ilias Chalkidis', 'Yova Kementchedjhieva'] | 2023-05-09 | null | null | null | null | ['multi-label-text-classification', 'multi-label-text-classification'] | ['methodology', 'natural-language-processing'] | [ 6.07527375e-01 2.91748077e-01 -2.96398550e-01 -5.53908765e-01
-1.39844024e+00 -5.12371957e-01 7.23120034e-01 3.70258301e-01
-5.56872427e-01 8.25404644e-01 3.36465955e-01 -3.91702145e-01
2.05073416e-01 -2.83822715e-01 -3.02067846e-01 -4.91885662e-01
2.76927710e-01 6.26137018e-01 -6.80310503e-02 2.47939043e-02
-4.06177193e-02 -2.51125962e-01 -1.31512415e+00 8.47585738e-01
5.47881782e-01 6.76666975e-01 -2.38939479e-01 6.83569551e-01
-5.21799102e-02 1.14557469e+00 -5.38955152e-01 -6.84537172e-01
1.11795612e-01 -5.05629778e-01 -9.46573198e-01 -2.95086019e-02
4.86372381e-01 -9.71049145e-02 -6.06196076e-02 6.76495194e-01
5.69460511e-01 -2.71914124e-01 1.02560675e+00 -9.11936104e-01
-6.21583223e-01 7.88851798e-01 -6.55747712e-01 -2.59622812e-01
2.71681875e-01 -2.85747319e-01 1.31860828e+00 -5.37771404e-01
6.02299929e-01 1.35034573e+00 1.15516448e+00 7.62706161e-01
-1.67825699e+00 -7.06382513e-01 9.91019905e-02 -1.15879409e-01
-1.35777247e+00 -5.34914672e-01 4.74433541e-01 -6.05422080e-01
1.14078057e+00 5.27240373e-02 -3.11767645e-02 1.42029059e+00
4.30134386e-01 7.60242522e-01 1.55240464e+00 -7.57387161e-01
-1.17836259e-01 3.53832334e-01 2.84861535e-01 5.49261570e-01
2.06715688e-01 2.56467089e-02 -3.21252108e-01 -3.79725695e-01
2.22082645e-01 -2.36118168e-01 8.86797979e-02 -1.70553535e-01
-1.13108051e+00 1.15050590e+00 -2.34614447e-01 5.77628851e-01
-1.17705114e-01 2.77265429e-01 7.84163296e-01 6.00049078e-01
9.57486272e-01 4.72325563e-01 -9.53127444e-01 -1.12368792e-01
-1.39490330e+00 1.58497229e-01 8.36706221e-01 1.01510739e+00
6.20124400e-01 -2.63932198e-01 -3.76809269e-01 1.15016091e+00
2.91113496e-01 -1.95294879e-02 6.78790927e-01 -5.93665659e-01
4.57638830e-01 4.87538487e-01 -1.41576782e-01 -3.37436229e-01
-7.54718661e-01 -5.75940967e-01 -6.11492455e-01 -2.91589624e-03
3.97257805e-01 -4.27482873e-01 -1.04152429e+00 1.90007329e+00
-1.21933118e-01 -3.33479699e-03 1.45719573e-01 2.21346140e-01
6.67155147e-01 3.24467748e-01 6.78576171e-01 -9.39989164e-02
1.76259184e+00 -1.12071192e+00 -7.00068831e-01 -5.02851367e-01
1.57858527e+00 -6.76810682e-01 8.71776700e-01 3.49235326e-01
-7.16977477e-01 -2.51985520e-01 -9.14071679e-01 -3.87109071e-01
-4.70047206e-01 5.49356937e-01 3.44545066e-01 1.01242447e+00
-1.12171984e+00 5.23551702e-01 -5.61519682e-01 -4.04645503e-01
3.53176951e-01 1.87185049e-01 -4.36279088e-01 -7.09025636e-02
-1.39276636e+00 1.29426908e+00 5.16997755e-01 -5.61454535e-01
-7.95936167e-01 -6.56379104e-01 -9.03562784e-01 -6.00926839e-02
8.71376246e-02 -6.17880404e-01 1.51852083e+00 -1.01316845e+00
-1.27997327e+00 1.36694109e+00 -8.69714692e-02 -4.36878949e-01
4.82933134e-01 -8.87586847e-02 -3.49725425e-01 -2.01191038e-01
2.08547503e-01 7.44141698e-01 7.36080945e-01 -1.29911411e+00
-7.53496885e-01 -1.50934577e-01 1.17269810e-02 7.76553974e-02
-4.14462477e-01 3.59960943e-01 -1.07667394e-01 -7.52673268e-01
-3.54582071e-01 -1.08372486e+00 -8.07150379e-02 -4.99629050e-01
-4.62193400e-01 -3.46618593e-01 4.62616891e-01 -6.79173887e-01
1.37641633e+00 -2.00443530e+00 1.20906256e-01 -3.23228866e-01
2.31604874e-01 1.01094335e-01 -3.16474915e-01 5.09410083e-01
-5.55052817e-01 4.54471707e-01 -2.92960554e-01 -1.00833714e+00
6.04634173e-02 2.01946303e-01 -3.58692929e-02 4.37627405e-01
3.17497641e-01 8.45208287e-01 -8.58104050e-01 -6.77992702e-01
-1.26449645e-01 4.00779009e-01 -4.98512983e-01 -1.27365097e-01
-2.93147027e-01 3.71367365e-01 -2.52535164e-01 4.61303085e-01
2.73749352e-01 -5.05294204e-01 3.39046061e-01 -5.83582819e-02
-8.41384102e-03 4.97449249e-01 -8.82088482e-01 1.94336987e+00
-9.19886112e-01 5.88235080e-01 1.99100319e-02 -9.17434275e-01
5.19095838e-01 8.48378122e-01 5.32651126e-01 -5.11727691e-01
1.79623619e-01 2.95766115e-01 -7.38916323e-02 -3.48210305e-01
3.76966238e-01 -5.87026179e-01 -2.60963321e-01 7.45931327e-01
3.17135900e-01 1.19023621e-01 3.38866413e-01 1.44709274e-01
1.29444635e+00 1.46039933e-01 3.68730664e-01 -3.37690324e-01
2.89652884e-01 -2.69762903e-01 3.98346990e-01 8.77307951e-01
-9.43432152e-02 6.34568572e-01 7.77473629e-01 -1.55528933e-01
-1.06709146e+00 -4.21565086e-01 -5.05713642e-01 1.41726756e+00
-3.48021686e-01 -8.92665446e-01 -6.10375583e-01 -1.18326521e+00
1.13854036e-01 1.16149592e+00 -8.37642431e-01 -1.17319658e-01
-3.41044664e-01 -1.04096532e+00 9.98590052e-01 3.54182541e-01
1.09354325e-01 -8.13680291e-01 -3.44548672e-01 3.27631205e-01
-1.94693834e-01 -1.02416801e+00 -2.37123594e-01 8.14278424e-01
-6.22308969e-01 -7.72785425e-01 -7.96134531e-01 -6.73875809e-01
4.08066243e-01 -2.72962272e-01 1.37759447e+00 -8.22612867e-02
-5.23477644e-02 2.73714453e-01 -4.74064440e-01 -4.70295429e-01
-9.21434581e-01 7.48695135e-01 -3.37990701e-01 -1.03333481e-01
4.79166150e-01 -3.03808272e-01 -1.11373022e-01 2.53915489e-01
-8.32435668e-01 2.91252762e-01 6.85036838e-01 9.59703803e-01
1.82811365e-01 -6.58213943e-02 7.67486632e-01 -1.63105690e+00
7.91339755e-01 -6.28880441e-01 -2.79947501e-02 4.19796824e-01
-9.86849368e-01 4.27420080e-01 6.09523416e-01 -3.46275955e-01
-8.59166622e-01 8.17844272e-02 -4.30604964e-01 -4.30150405e-02
-4.22175527e-01 4.44132566e-01 7.17436150e-02 1.13893688e-01
8.56954277e-01 -2.02019420e-02 -2.11800769e-01 -6.92140698e-01
4.92990345e-01 9.64156032e-01 -8.98446068e-02 -5.27797580e-01
3.26712638e-01 2.53403876e-02 -1.18684195e-01 -3.36192459e-01
-1.27882421e+00 -5.12149513e-01 -7.41528988e-01 6.36828840e-02
1.04318976e+00 -1.14430141e+00 -2.67931551e-01 3.79457325e-01
-1.17462468e+00 -4.16815937e-01 -1.30935013e-01 3.63820404e-01
-4.90313858e-01 2.33154014e-01 -1.08828771e+00 -5.10707378e-01
-2.12617457e-01 -1.34651661e+00 1.55502903e+00 -3.12043458e-01
-5.27951360e-01 -1.35767019e+00 2.51121908e-01 3.06833416e-01
3.56206357e-01 7.51860859e-03 1.18719423e+00 -1.13248003e+00
1.65539846e-01 -2.13190719e-01 -3.05195868e-01 2.04115093e-01
2.69143760e-01 -3.98668617e-01 -1.30957544e+00 -4.44802821e-01
-4.56743352e-02 -8.40721130e-01 1.21618176e+00 1.24319769e-01
9.85361397e-01 1.03408419e-01 -6.44934535e-01 1.80481672e-01
1.44605935e+00 -1.17523782e-01 4.58197892e-01 3.20618778e-01
7.34920561e-01 6.09302640e-01 2.61396229e-01 1.13602608e-01
6.59658790e-01 7.78120399e-01 9.70363170e-02 -2.06036001e-01
-2.27051526e-01 -8.22355680e-04 3.78678948e-01 8.87434006e-01
3.77162009e-01 -5.90214193e-01 -9.97886717e-01 4.23753679e-01
-1.70844471e+00 -6.40130639e-01 -1.99969068e-01 1.93597865e+00
1.34414983e+00 1.88085869e-01 3.93607803e-02 7.17704818e-02
5.16376495e-01 2.61287153e-01 -2.52185971e-01 -5.67130148e-01
-7.20554963e-02 2.47794837e-01 6.65589273e-01 4.51971948e-01
-1.55240083e+00 8.79225969e-01 7.22576141e+00 1.02058268e+00
-8.83958459e-01 5.37192643e-01 6.74687743e-01 -8.56065676e-02
-2.12036088e-01 1.07428357e-01 -1.09715176e+00 5.21973491e-01
1.57729721e+00 3.59806240e-01 1.04189642e-01 5.70381165e-01
-2.46121183e-01 -5.38636893e-02 -1.42476618e+00 8.39051187e-01
1.74341112e-01 -9.64574337e-01 -1.02434695e-01 1.42602608e-01
6.17273271e-01 1.34729981e-01 -1.25409514e-01 6.08138740e-01
5.91159523e-01 -1.08056200e+00 7.11770952e-01 2.68339187e-01
1.14472950e+00 -4.31279093e-01 7.53814518e-01 4.69435453e-01
-9.70753491e-01 7.13802129e-02 -6.38738200e-02 1.33787200e-01
1.25206605e-01 7.13179767e-01 -5.66572487e-01 7.31530845e-01
2.23832056e-01 8.82696927e-01 -8.65527332e-01 4.39910680e-01
-1.52550384e-01 9.07043993e-01 -7.75592923e-02 2.27386519e-01
1.65908471e-01 1.97663665e-01 1.10335276e-01 1.73114264e+00
3.69982719e-02 -2.57170945e-01 2.18811169e-01 3.27158988e-01
-2.20636740e-01 2.52205342e-01 -5.86389542e-01 -6.17973506e-02
1.86005011e-01 1.40340841e+00 -8.10880125e-01 -4.69842583e-01
-1.00901639e+00 8.75551045e-01 6.05181694e-01 3.09091676e-02
-8.88419151e-01 -2.99927711e-01 3.58865440e-01 -7.21749738e-02
6.43158928e-02 -9.92449149e-02 -3.97371173e-01 -1.31257749e+00
-3.65000308e-01 -1.21481538e+00 5.77126682e-01 -4.72155243e-01
-1.28533220e+00 6.52195394e-01 -5.55520765e-02 -9.93613601e-01
-5.74942052e-01 -5.77515602e-01 1.01939015e-01 9.37551916e-01
-1.55097926e+00 -1.40801346e+00 2.36166239e-01 1.18154801e-01
5.58194041e-01 -7.05534080e-03 1.41836536e+00 7.32283115e-01
-5.66529810e-01 9.13066626e-01 4.10972953e-01 1.28092051e-01
1.26630914e+00 -1.40317059e+00 2.64476389e-01 3.90004724e-01
3.98547381e-01 3.57330114e-01 4.94428188e-01 -6.43009901e-01
-7.34011948e-01 -1.21727121e+00 1.38635206e+00 -8.69512260e-01
5.31005085e-01 -7.85444736e-01 -9.31431413e-01 1.01152718e+00
5.40368378e-01 -4.09250706e-01 1.11920869e+00 5.94852567e-01
-4.64460254e-01 3.71342987e-01 -8.92351449e-01 3.66338909e-01
8.25112879e-01 -7.86924362e-01 -5.39146185e-01 5.86953938e-01
5.51380098e-01 -2.67026663e-01 -1.14037597e+00 3.27395558e-01
6.21136010e-01 -7.14888752e-01 7.26988316e-01 -5.86325586e-01
5.49609661e-01 6.22913092e-02 -1.46219414e-02 -1.48723507e+00
-4.99484092e-01 -2.27320880e-01 1.70418248e-01 1.36914313e+00
8.62320900e-01 -6.62589014e-01 5.68096817e-01 4.54165488e-01
-1.38395861e-01 -7.80726135e-01 -7.26468861e-01 -6.73625648e-01
5.45331478e-01 -3.82791847e-01 2.66372889e-01 1.18908882e+00
-7.99643546e-02 8.08662415e-01 -6.30435348e-01 -2.15181664e-01
3.91713053e-01 -1.05752870e-01 2.96028137e-01 -1.35069191e+00
-2.72547722e-01 -4.31948513e-01 -1.08531572e-01 -9.34194624e-01
6.43521667e-01 -1.26484823e+00 2.45679282e-02 -1.46159244e+00
4.71751958e-01 -5.92628837e-01 -3.52954179e-01 1.03730774e+00
-3.07256401e-01 2.85215497e-01 -1.59731925e-01 2.29497716e-01
-6.56353414e-01 2.56741971e-01 7.25288630e-01 -2.43710995e-01
1.43266782e-01 -1.66645527e-01 -7.80395508e-01 4.94877160e-01
7.23441899e-01 -1.11095715e+00 -3.98126096e-01 -5.47739148e-01
2.08882868e-01 -5.97872622e-02 -1.21138707e-01 -8.22540283e-01
-5.40908240e-03 3.39339346e-01 1.52556643e-01 -5.29599786e-02
1.17134772e-01 -6.45077169e-01 7.08243102e-02 3.56181830e-01
-1.02106595e+00 -2.62684599e-02 3.43013048e-01 4.39335972e-01
-1.15978830e-01 -6.59713566e-01 8.18119109e-01 -1.26343211e-02
-4.04173672e-01 4.19461280e-02 -6.83728039e-01 2.90577680e-01
8.88649881e-01 1.36623621e-01 -3.25298131e-01 -2.18142122e-01
-1.01447606e+00 1.32362977e-01 3.98832202e-01 4.71988201e-01
-1.02187999e-01 -1.32432723e+00 -1.04289973e+00 1.37843639e-01
3.63374382e-01 -5.45959651e-01 -1.54550046e-01 8.70905101e-01
-2.28496566e-01 7.39628851e-01 2.20037967e-01 -4.77022946e-01
-1.26786256e+00 4.55293328e-01 4.79324818e-01 -9.93899643e-01
-3.86839241e-01 7.00558126e-01 3.05423468e-01 -6.47385955e-01
6.18508048e-02 -1.96173355e-01 -2.51835108e-01 4.40903932e-01
2.17871562e-01 3.98318246e-02 3.02831173e-01 -7.21729159e-01
-3.37500691e-01 4.69985783e-01 -4.08385515e-01 -2.13978827e-01
9.74154413e-01 -1.76221281e-01 -1.50093390e-02 7.96189427e-01
1.38480771e+00 -1.15617961e-01 -8.70949149e-01 -2.76255995e-01
5.07990241e-01 -7.85445049e-03 2.33373135e-01 -9.48477745e-01
-7.03570604e-01 9.25971210e-01 3.41735065e-01 4.44362342e-01
9.44867730e-01 2.38905028e-02 4.55052078e-01 7.49356579e-03
3.40558469e-01 -9.71827567e-01 -2.10346684e-01 5.29356658e-01
5.02322674e-01 -1.28444850e+00 3.91205214e-02 -2.46265277e-01
-6.40550613e-01 1.06763422e+00 3.10032248e-01 3.11078340e-01
7.35131979e-01 5.62408626e-01 2.25893408e-01 -7.68785924e-02
-1.07266116e+00 -2.45968938e-01 2.01568961e-01 1.45312011e-01
1.23753965e+00 -3.92462723e-02 -3.03613037e-01 4.76210713e-01
1.32176980e-01 4.41966653e-02 1.82373509e-01 1.12569606e+00
-1.17141474e-02 -1.67102790e+00 -7.73479939e-02 8.81485343e-01
-8.31151664e-01 -4.24258530e-01 -3.97024035e-01 7.42012620e-01
1.82514712e-01 1.08758283e+00 -3.82158048e-02 -4.23819005e-01
1.66396022e-01 7.36434579e-01 4.70269650e-01 -1.03093159e+00
-8.78149986e-01 1.25136361e-01 6.32166028e-01 -2.43358284e-01
-5.18129885e-01 -9.26725090e-01 -9.51517403e-01 -8.63006189e-02
-6.46228909e-01 1.93293884e-01 5.66458285e-01 1.03426480e+00
4.87260103e-01 7.16435373e-01 3.03297490e-01 -4.33618039e-01
-5.96352398e-01 -1.41339719e+00 -6.49074852e-01 4.31551307e-01
2.45524183e-01 -6.84702277e-01 -3.12464535e-01 2.33142585e-01] | [9.518298149108887, 4.541554927825928] |
1ebdcecc-3527-4772-9ce5-83e66457a7b4 | multilingual-pre-training-with-language-and-1 | 2203.08552 | null | https://arxiv.org/abs/2203.08552v1 | https://arxiv.org/pdf/2203.08552v1.pdf | Multilingual Pre-training with Language and Task Adaptation for Multilingual Text Style Transfer | We exploit the pre-trained seq2seq model mBART for multilingual text style transfer. Using machine translated data as well as gold aligned English sentences yields state-of-the-art results in the three target languages we consider. Besides, in view of the general scarcity of parallel data, we propose a modular approach for multilingual formality transfer, which consists of two training strategies that target adaptation to both language and task. Our approach achieves competitive performance without monolingual task-specific parallel data and can be applied to other style transfer tasks as well as to other languages. | ['Malvina Nissim', 'Antonio Toral', 'Huiyuan Lai'] | 2022-03-16 | null | https://aclanthology.org/2022.acl-short.29 | https://aclanthology.org/2022.acl-short.29.pdf | acl-2022-5 | ['text-style-transfoer'] | ['natural-language-processing'] | [ 2.48333126e-01 -7.95747191e-02 -2.57800490e-01 -4.11593288e-01
-1.34222960e+00 -9.37429547e-01 7.62126625e-01 -1.77071273e-01
-9.44620252e-01 1.27224588e+00 2.96245903e-01 -5.87242365e-01
4.14360464e-01 -3.89798224e-01 -8.35866332e-01 -1.77637115e-01
3.07075292e-01 9.81968045e-01 7.27866590e-02 -6.97417676e-01
-8.67736191e-02 1.64208919e-01 -5.85377753e-01 5.47050595e-01
1.05864632e+00 3.22537959e-01 3.99742723e-01 5.22227645e-01
-1.33616596e-01 4.45749491e-01 -4.79567319e-01 -7.24400342e-01
2.45703503e-01 -7.40140676e-01 -1.16188622e+00 -3.75489682e-01
6.40875280e-01 -3.39408778e-02 -2.32623190e-01 9.87315059e-01
6.98834896e-01 3.28162499e-03 5.33392787e-01 -3.81502450e-01
-7.16199160e-01 7.16186762e-01 -3.51972759e-01 3.11210424e-01
3.63156825e-01 1.26463011e-01 1.08690810e+00 -1.13224399e+00
1.07245791e+00 1.42029881e+00 5.45548081e-01 9.20700669e-01
-1.46167600e+00 -4.75270838e-01 1.87145889e-01 1.01914980e-01
-1.07084870e+00 -5.94629526e-01 4.26284730e-01 -1.97193503e-01
1.23001337e+00 -1.49039194e-01 2.74427116e-01 1.56591856e+00
4.82578546e-01 9.07209218e-01 1.51809812e+00 -8.20893228e-01
-2.20290661e-01 3.42496216e-01 -3.70673001e-01 5.37367105e-01
-1.43786013e-01 1.06651776e-01 -5.70224822e-01 2.03311950e-01
5.18481612e-01 -6.60741746e-01 -2.22438708e-01 -1.04469985e-01
-1.58472490e+00 8.79200339e-01 1.67389795e-01 5.67479670e-01
-8.62477273e-02 -8.65103677e-02 9.19953048e-01 8.39949131e-01
8.93418908e-01 5.92891932e-01 -1.01224482e+00 -1.30585596e-01
-9.24888492e-01 8.54652300e-02 7.91081369e-01 1.23200321e+00
7.84443021e-01 1.86368838e-01 -2.62012810e-01 1.10597563e+00
-1.49058253e-01 8.79542470e-01 5.24500251e-01 -3.95058006e-01
1.11838317e+00 3.19461781e-03 -2.74443805e-01 -1.78174376e-01
-1.39794901e-01 -4.62164193e-01 -7.20869005e-01 -7.71658123e-02
4.04456258e-01 -4.75294381e-01 -5.82180798e-01 1.98467135e+00
-3.20031680e-02 -4.50187743e-01 4.96066719e-01 5.12597740e-01
5.14986515e-01 7.55056679e-01 1.32711828e-01 -3.76408622e-02
1.37829602e+00 -1.22557092e+00 -4.55217123e-01 -4.52463776e-01
8.94238472e-01 -1.11096764e+00 1.51992953e+00 1.87541619e-01
-1.19757688e+00 -5.21426976e-01 -7.50017107e-01 -2.71092772e-01
-4.09621805e-01 1.56636223e-01 2.99373060e-01 3.29114735e-01
-1.26798379e+00 5.28647184e-01 -6.91331804e-01 -8.59972596e-01
1.60469338e-01 2.77501225e-01 -4.70001906e-01 -2.53656417e-01
-1.44802392e+00 1.34292150e+00 5.78186274e-01 -9.44537371e-02
-8.32375944e-01 -7.56929338e-01 -9.02921617e-01 -3.83003831e-01
1.60587832e-01 -8.24009240e-01 1.32540524e+00 -1.41808140e+00
-1.90320742e+00 1.25755632e+00 -5.75733520e-02 -3.23324829e-01
8.68489146e-01 -5.79510272e-01 -4.71712142e-01 -2.07924604e-01
2.24170923e-01 8.20362389e-01 4.27774698e-01 -7.13267148e-01
-6.58724487e-01 -1.23017363e-01 -8.22396725e-02 5.90543389e-01
-4.52834249e-01 5.31246662e-01 -3.24256480e-01 -8.92192364e-01
-7.69383073e-01 -1.04086828e+00 -3.53784680e-01 -5.34104943e-01
-1.95795298e-01 -2.18377531e-01 3.72712344e-01 -1.14136529e+00
7.79816747e-01 -1.76855755e+00 6.82812750e-01 -1.55425400e-01
-6.21593356e-01 3.51338863e-01 -6.75646305e-01 6.93372369e-01
2.10722342e-01 -3.51378880e-02 -3.19555789e-01 -5.16230345e-01
-1.63528651e-01 2.36649841e-01 -2.75088131e-01 2.25530580e-01
5.20783186e-01 1.25500858e+00 -1.07141089e+00 -4.93177742e-01
-5.52206151e-02 3.36724073e-01 -6.30747318e-01 3.94917011e-01
-3.55531275e-01 1.03619790e+00 -2.51577795e-01 2.49134392e-01
3.40060502e-01 2.99049914e-01 3.69310349e-01 1.02091715e-01
-2.13340610e-01 7.98296869e-01 -4.20362234e-01 2.51057339e+00
-1.05854797e+00 5.07615685e-01 -2.64770631e-02 -7.47262299e-01
8.21526170e-01 5.63585639e-01 -1.23192616e-01 -7.83364058e-01
-8.85199010e-02 6.91835880e-01 2.11187340e-02 -9.99790505e-02
5.39891362e-01 -5.42467773e-01 -4.81393933e-01 5.04615009e-01
7.63035834e-01 -3.57941210e-01 2.78273553e-01 -1.31138250e-01
7.04268873e-01 6.66754186e-01 5.76246202e-01 -9.30440068e-01
8.02377939e-01 -2.32773964e-05 4.02830392e-01 5.11428058e-01
1.32811993e-01 3.25802416e-01 1.07042179e-01 -4.60709065e-01
-1.27383804e+00 -1.05907536e+00 1.04776099e-01 1.54609096e+00
-4.81775135e-01 -1.76458895e-01 -7.83601940e-01 -1.17121923e+00
-3.31540555e-01 8.87526274e-01 -4.76442277e-01 1.05907492e-01
-1.15909445e+00 -4.34809834e-01 7.80620456e-01 3.71893466e-01
1.32718250e-01 -1.44853330e+00 1.44330993e-01 4.78436321e-01
-3.30517113e-01 -1.35457981e+00 -8.95671070e-01 1.27765775e-01
-7.63050973e-01 -5.79820275e-01 -1.04849195e+00 -1.06004596e+00
3.04697871e-01 -2.96575814e-01 1.63833320e+00 -3.51187050e-01
3.11682820e-01 1.82850599e-01 -2.18331844e-01 -3.59061539e-01
-9.31434035e-01 9.57900763e-01 1.66466072e-01 -1.62583470e-01
2.12323442e-01 -4.57891494e-01 -7.22961575e-02 6.75118864e-02
-6.16392612e-01 1.14944614e-01 7.66233027e-01 9.54640865e-01
3.85284036e-01 -8.70979905e-01 8.47599387e-01 -1.19985437e+00
6.36000812e-01 -2.29120538e-01 -5.79545975e-01 5.65857351e-01
-3.63591760e-01 2.55012721e-01 1.05049467e+00 -3.06414276e-01
-1.22722030e+00 -2.33016908e-01 -4.11021471e-01 1.05380841e-01
-3.31361741e-01 4.04363215e-01 -3.56553644e-01 6.65048435e-02
7.07471430e-01 2.89116323e-01 -2.38237575e-01 -7.14915097e-01
5.72492659e-01 5.88406086e-01 4.92580444e-01 -8.56314719e-01
6.49656355e-01 -1.01747038e-02 -1.27236426e-01 -8.70454252e-01
-1.01333368e+00 -5.52373528e-02 -1.04924178e+00 1.98970631e-01
9.86234069e-01 -1.12496817e+00 -1.74812581e-02 2.88058281e-01
-1.61813486e+00 -7.00114727e-01 -2.93337345e-01 5.33917010e-01
-7.75966406e-01 1.51607543e-01 -1.06267750e+00 -3.53529975e-02
-7.37014532e-01 -1.10137331e+00 1.02361643e+00 -3.26186776e-01
-3.76063496e-01 -1.61139905e+00 5.66510022e-01 6.63798302e-02
5.30306935e-01 -1.79862708e-01 9.92095888e-01 -8.76134098e-01
-3.39925259e-01 1.90797791e-01 3.99000011e-03 6.75887346e-01
1.84008360e-01 -5.96916497e-01 -6.25443578e-01 -7.14239120e-01
-2.08835080e-01 -7.97821999e-01 6.19551480e-01 -1.55610546e-01
3.39329243e-01 -3.33492160e-01 4.17350326e-03 6.05441689e-01
1.29491210e+00 -1.86567783e-01 3.42810154e-01 3.87466431e-01
8.96787167e-01 6.81791961e-01 5.07859170e-01 -2.83224136e-01
4.86577421e-01 9.07582343e-01 -2.81439573e-01 -4.03452158e-01
-3.07550877e-01 -4.30134565e-01 8.42434347e-01 1.55634630e+00
-1.61853002e-03 -2.42092341e-01 -1.05760527e+00 6.91688418e-01
-1.58883119e+00 -5.36886573e-01 1.32311061e-01 2.17210293e+00
1.30253208e+00 1.39437371e-03 1.71547875e-01 -5.74721754e-01
5.67960322e-01 9.68006477e-02 -1.64126605e-01 -8.17456603e-01
-3.55277181e-01 7.12477863e-01 4.53085124e-01 7.93328285e-01
-9.07377779e-01 1.80492353e+00 6.75995255e+00 1.08378661e+00
-1.38563573e+00 6.56361043e-01 5.62225401e-01 -2.69882148e-03
-4.51389015e-01 -3.98401776e-03 -9.84781981e-01 1.82698056e-01
1.18584239e+00 -1.55657262e-01 6.94293022e-01 3.84417206e-01
-6.71240091e-02 4.05006438e-01 -1.43426669e+00 5.45150876e-01
1.03118546e-01 -9.66451287e-01 3.69473308e-01 -1.66820705e-01
1.05469108e+00 5.14495909e-01 -3.73025611e-02 5.19208550e-01
5.69685280e-01 -9.63973463e-01 8.34101081e-01 1.41743407e-01
1.36731279e+00 -8.23154032e-01 7.48067200e-01 3.69638354e-01
-9.28176343e-01 5.44315279e-01 -3.10390621e-01 -9.27439705e-03
3.49047601e-01 1.92624018e-01 -7.31116652e-01 1.05042899e+00
4.02204216e-01 1.00677109e+00 -4.61379170e-01 2.79121727e-01
-7.04593480e-01 8.71454298e-01 -1.17794693e-01 1.19999543e-01
4.21248347e-01 -1.54927552e-01 4.86895055e-01 1.82506323e+00
3.00507367e-01 -7.28635669e-01 3.86658341e-01 5.21080852e-01
-1.77999809e-01 8.01874638e-01 -9.30077970e-01 1.63221452e-02
8.38459730e-02 1.02770770e+00 -2.50190645e-01 -3.75898540e-01
-6.88865602e-01 1.45209539e+00 1.06649053e+00 4.64422673e-01
-2.63292253e-01 -3.24687004e-01 3.81611913e-01 -2.38363862e-01
2.68518031e-01 -4.11535800e-01 -1.04946479e-01 -1.48751533e+00
1.42403143e-02 -1.33379185e+00 4.13100243e-01 -2.00605199e-01
-1.53474021e+00 1.20844448e+00 -1.66029096e-01 -1.07750762e+00
-5.06689787e-01 -7.77092814e-01 -5.69926202e-01 1.33144701e+00
-1.64629531e+00 -1.66231751e+00 4.90156770e-01 6.02359116e-01
8.46835554e-01 -5.98952413e-01 1.21121264e+00 3.35835844e-01
-2.70063341e-01 8.28024983e-01 2.38646671e-01 1.16745569e-01
1.33725452e+00 -1.27451658e+00 9.61393297e-01 8.39079618e-01
3.12842071e-01 4.16237473e-01 5.60259283e-01 -4.74110961e-01
-9.73246634e-01 -1.25972319e+00 1.54605913e+00 -6.15036011e-01
9.04587209e-01 -8.40691686e-01 -7.63172090e-01 1.02378893e+00
9.89444971e-01 -2.86812752e-01 6.57950938e-01 4.10988986e-01
-2.58775324e-01 -7.09128231e-02 -7.39722610e-01 7.66901016e-01
1.07596922e+00 -7.51598299e-01 -6.17927015e-01 4.46691930e-01
6.75427973e-01 -4.65383649e-01 -9.71738398e-01 4.02194560e-01
2.97054231e-01 -5.15969872e-01 4.99722570e-01 -9.82398510e-01
5.25256515e-01 2.13337511e-01 -2.31689319e-01 -1.99278688e+00
-2.66817838e-01 -7.79958487e-01 5.04843295e-01 1.41899848e+00
9.27359283e-01 -8.00263882e-01 1.09982923e-01 -4.47918177e-01
-2.15554655e-01 -5.76125205e-01 -1.06715667e+00 -1.02725804e+00
1.00080597e+00 -7.97997266e-02 2.94257075e-01 1.07349217e+00
6.62351493e-03 1.17855179e+00 -7.09120572e-01 -3.55265766e-01
3.80225986e-01 1.60143562e-02 8.77688885e-01 -9.27581787e-01
-5.10462463e-01 -3.77810866e-01 1.11859754e-01 -1.03191602e+00
8.52917016e-01 -1.57177830e+00 1.99572846e-01 -1.29497516e+00
1.34680033e-01 -3.81492674e-01 -3.36834908e-01 4.30635303e-01
-4.18605864e-01 5.12223244e-01 2.69994408e-01 1.22868955e-01
-5.56187451e-01 8.01756740e-01 1.24159670e+00 -6.00850768e-03
-1.77284926e-02 -1.27420500e-01 -4.86118078e-01 3.79463553e-01
7.50583112e-01 -5.39963186e-01 -1.44032523e-01 -9.62474465e-01
1.41188920e-01 -7.93197155e-02 -2.81610698e-01 -6.77777827e-01
-3.54970723e-01 -1.16370618e-01 1.06142826e-01 -1.74741223e-02
-1.43786632e-02 -5.06903827e-01 -2.97180295e-01 5.13957739e-01
-4.71740752e-01 4.68040019e-01 6.81928933e-01 9.95831266e-02
-2.84373999e-01 -1.85214698e-01 8.98310602e-01 -3.02112728e-01
-3.78436029e-01 2.02561229e-01 -4.78884578e-01 5.99253714e-01
6.50334358e-01 3.21341813e-01 -1.08142853e-01 -2.54573941e-01
-4.11452234e-01 1.92498282e-01 5.86944997e-01 7.69479215e-01
-1.13649651e-01 -1.60531592e+00 -1.44588184e+00 1.97357804e-01
2.56534100e-01 -4.18749809e-01 -3.51223387e-02 7.44725287e-01
-4.11140501e-01 7.61307716e-01 -5.12790918e-01 -5.94686806e-01
-9.82147992e-01 6.19697750e-01 2.76242644e-01 -8.43720615e-01
-3.71255457e-01 5.39433181e-01 3.88257802e-01 -1.20096695e+00
-2.80066878e-01 8.36331099e-02 1.40328914e-01 -2.60259062e-01
3.35195452e-01 1.15281984e-01 2.78991908e-01 -8.15877020e-01
-3.83584946e-01 5.63144505e-01 -3.71139854e-01 -6.01749241e-01
1.16907966e+00 -2.14856640e-01 -2.59043962e-01 8.11551630e-01
1.11725032e+00 3.42393339e-01 -9.94189262e-01 -5.93730569e-01
2.97223896e-01 1.45723848e-02 -4.85611707e-01 -9.59666193e-01
-6.80909634e-01 1.07748437e+00 2.01566309e-01 -5.67962170e-01
8.08920920e-01 -3.12662125e-02 7.49897599e-01 4.26625699e-01
5.41670859e-01 -1.14855719e+00 -8.40403736e-02 1.30627131e+00
1.06467581e+00 -1.15206087e+00 -3.78997922e-01 -1.74718648e-01
-6.70617044e-01 9.29962277e-01 5.53715289e-01 -2.37048179e-01
2.83812553e-01 1.08201250e-01 4.82188284e-01 4.04727817e-01
-8.78493726e-01 -3.11252810e-02 6.43147647e-01 5.43641865e-01
1.05115819e+00 1.11044884e-01 -7.32013285e-01 3.62315059e-01
-3.67938042e-01 -6.86836317e-02 6.89996332e-02 7.90078163e-01
5.85560091e-02 -1.81625211e+00 -6.23911992e-02 -4.36361693e-03
-8.33715320e-01 -6.94386661e-01 -4.70573872e-01 8.32726359e-01
-1.50610536e-01 5.21800935e-01 -1.17267139e-01 -2.14536786e-02
4.61364120e-01 3.68768871e-01 8.26348841e-01 -9.19007540e-01
-1.00519824e+00 7.05874413e-02 3.69658053e-01 -3.64740610e-01
-1.64621368e-01 -6.71860635e-01 -6.24070644e-01 -1.53112248e-01
8.76946151e-02 1.25429675e-01 4.36467618e-01 1.18313479e+00
2.39165634e-01 3.58387321e-01 3.76965970e-01 -9.74789202e-01
-5.76485276e-01 -1.46775544e+00 -2.19407305e-01 3.63492191e-01
1.19243167e-01 -5.74155040e-02 1.76107526e-01 9.74996164e-02] | [11.392507553100586, 10.10200023651123] |
4eac61b3-06a9-43c6-9867-ee97eac3f2a0 | joint-2d-3d-breast-cancer-classification | 2002.12392 | null | https://arxiv.org/abs/2002.12392v1 | https://arxiv.org/pdf/2002.12392v1.pdf | Joint 2D-3D Breast Cancer Classification | Breast cancer is the malignant tumor that causes the highest number of cancer deaths in females. Digital mammograms (DM or 2D mammogram) and digital breast tomosynthesis (DBT or 3D mammogram) are the two types of mammography imagery that are used in clinical practice for breast cancer detection and diagnosis. Radiologists usually read both imaging modalities in combination; however, existing computer-aided diagnosis tools are designed using only one imaging modality. Inspired by clinical practice, we propose an innovative convolutional neural network (CNN) architecture for breast cancer classification, which uses both 2D and 3D mammograms, simultaneously. Our experiment shows that the proposed method significantly improves the performance of breast cancer classification. By assembling three CNN classifiers, the proposed model achieves 0.97 AUC, which is 34.72% higher than the methods using only one imaging modality. | ['Xiaoqin Wang', 'Hunter Blanton', 'Nathan Jacobs', 'Yu Zhang', 'Xin Xing', 'Tawfiq Salem', 'Gongbo Liang'] | 2020-02-27 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 3.99424046e-01 3.18751037e-01 -6.89523935e-01 -4.14953440e-01
-4.84283268e-01 4.44845669e-02 4.60905790e-01 3.01955342e-01
-4.03573692e-01 4.19446826e-01 -1.04353197e-01 -9.71915841e-01
8.77136663e-02 -1.16467452e+00 -5.12222350e-01 -5.23836255e-01
-3.18833962e-02 4.47625101e-01 2.14512721e-01 9.95535683e-03
-3.68709683e-01 5.42009115e-01 -1.10342467e+00 3.64139199e-01
4.42462623e-01 1.52467704e+00 5.04532196e-02 7.48610914e-01
-1.31412268e-01 7.55736887e-01 -1.25288591e-01 -5.72443843e-01
1.61162212e-01 -5.55711091e-01 -4.19285178e-01 1.62624106e-01
2.58284330e-01 -8.62506449e-01 -6.42558634e-01 1.08163273e+00
4.97827709e-01 -8.34304988e-01 5.78623593e-01 -5.42129874e-01
-6.60730839e-01 5.88493943e-01 -9.47833121e-01 4.72005844e-01
-1.41709238e-01 -3.40602487e-01 1.16492249e-01 -6.98810339e-01
1.87675923e-01 1.03943264e+00 8.74778032e-01 7.33634353e-01
-4.27280068e-01 -7.51026928e-01 -5.65493405e-01 1.07275918e-01
-1.10224617e+00 -1.65510163e-01 3.92363876e-01 -1.15865976e-01
5.24198294e-01 3.95822138e-01 1.05415595e+00 7.88910270e-01
8.51127088e-01 9.43084776e-01 9.96708333e-01 -4.93289381e-01
9.52103883e-02 -5.35186604e-02 2.03431904e-01 8.48779023e-01
9.31017160e-01 2.95177072e-01 -3.55439819e-02 -1.88102722e-01
9.10480261e-01 4.97988820e-01 2.38922648e-02 6.05471199e-03
-9.84849453e-01 7.62625575e-01 6.64668024e-01 5.43877184e-01
-4.11173582e-01 2.34491348e-01 4.98900354e-01 1.26332277e-02
2.29555041e-01 -2.08127692e-01 1.67911034e-02 4.22025561e-01
-8.23364258e-01 3.48811634e-02 4.04750645e-01 6.83013856e-01
9.31642875e-02 -8.44876245e-02 -4.13892865e-02 8.19132447e-01
3.53832990e-01 6.62206173e-01 9.83595371e-01 -3.05485427e-01
1.41551346e-02 7.44763196e-01 -1.22257635e-01 -1.11421931e+00
-7.17350900e-01 -7.42578328e-01 -1.47937560e+00 -3.50605994e-01
1.48774102e-01 1.14716142e-01 -1.28851426e+00 9.25788343e-01
2.74275988e-01 -1.04313619e-01 2.07966164e-01 7.03880310e-01
1.37441087e+00 1.07745059e-01 2.79402405e-01 -1.84183851e-01
1.78353596e+00 -6.57650948e-01 -1.03589308e+00 -2.98882335e-01
9.16330040e-01 -3.73776466e-01 1.83096811e-01 3.92294049e-01
-1.00339067e+00 -4.93184656e-01 -1.39254773e+00 -5.52286878e-02
-2.03981847e-01 6.20115995e-01 9.53084886e-01 1.00804746e+00
-6.98926806e-01 1.75681397e-01 -1.23738682e+00 -4.54665780e-01
8.81274641e-01 2.84758210e-01 -4.75123078e-01 -5.22205174e-01
-9.97769177e-01 8.30141485e-01 3.65836382e-01 2.85848975e-01
-9.98051822e-01 -3.78451496e-01 -7.31864214e-01 -6.66428506e-02
1.76818326e-01 -4.34568763e-01 1.52293515e+00 -9.58172441e-01
-8.84870052e-01 1.30109084e+00 -4.84814482e-05 -7.35209644e-01
6.73235714e-01 -8.56161118e-02 -5.29130638e-01 3.62334639e-01
1.13793518e-02 7.13490844e-01 3.69209260e-01 -8.93267632e-01
-7.80691385e-01 -6.29593432e-01 -4.24811393e-01 -1.13316126e-01
-5.31488299e-01 -1.14310361e-01 -7.40737259e-01 -3.48880678e-01
6.22460365e-01 -5.97896397e-01 -3.44439805e-01 3.14312816e-01
-4.24577236e-01 3.80993932e-02 7.93117106e-01 -9.90239620e-01
1.05944896e+00 -2.04490089e+00 -5.53580999e-01 2.86926366e-02
4.68373328e-01 4.40201253e-01 3.61713171e-01 -3.03233325e-01
-1.52940705e-01 3.59938711e-01 -6.41190261e-02 4.16107960e-02
-4.88799602e-01 1.35600105e-01 5.40574551e-01 2.95026809e-01
2.61253327e-01 1.26717532e+00 -8.05722415e-01 -6.59341514e-01
1.83065146e-01 3.34209740e-01 1.29096165e-01 -1.03022508e-01
2.85258979e-01 2.20634282e-01 -6.48242891e-01 1.19116414e+00
1.03375876e+00 -6.69959843e-01 5.98327398e-01 -3.43187332e-01
5.22261441e-01 -2.02481419e-01 -5.92755079e-01 1.16894901e+00
2.39736540e-03 4.24424291e-01 -1.83451042e-01 -1.08254898e+00
9.93661940e-01 4.62755233e-01 4.89513040e-01 -8.62517595e-01
5.60404778e-01 4.63939488e-01 3.27081203e-01 -8.39810729e-01
-6.40526563e-02 -2.22156346e-01 2.49115214e-01 1.25881419e-01
-2.82420903e-01 9.74528417e-02 -6.00579008e-02 -1.65027797e-01
1.49527895e+00 -5.21352649e-01 1.09633863e+00 -1.54328719e-01
2.45324329e-01 1.62005201e-01 5.74188113e-01 8.11835706e-01
-3.57184023e-01 4.02159750e-01 3.95520151e-01 -1.00607109e+00
-1.02262509e+00 -7.80635834e-01 -5.63243449e-01 1.26950249e-01
1.54002964e-01 2.90028691e-01 -4.02491599e-01 -7.57631004e-01
1.73528209e-01 9.13175419e-02 -9.50051308e-01 -1.28868952e-01
-7.12451875e-01 -1.17721546e+00 7.53269196e-01 8.45757782e-01
1.29931629e+00 -5.90040088e-01 -7.86205471e-01 1.62479892e-01
7.53933415e-02 -9.47497427e-01 1.84197545e-01 2.10645139e-01
-1.28156471e+00 -1.48357165e+00 -1.07723320e+00 -9.04775918e-01
7.19531596e-01 4.17814046e-01 9.12657261e-01 5.06283164e-01
-6.71594501e-01 -4.61177900e-02 -4.43419307e-01 -7.29889989e-01
-7.22574472e-01 -7.50564784e-02 -2.70753860e-01 -5.94235510e-02
7.46908307e-01 -2.00764522e-01 -6.90055251e-01 -8.36117193e-02
-1.16471970e+00 4.02289182e-01 1.26920104e+00 9.45873082e-01
6.83218658e-01 2.32618228e-01 5.74283063e-01 -1.10254931e+00
3.55174899e-01 -6.58266187e-01 -2.19273582e-01 1.77848279e-01
-2.83588409e-01 -6.90963089e-01 2.09878251e-01 -4.38300520e-01
-8.60302389e-01 9.84364003e-02 -2.46794149e-01 -1.89935833e-01
-4.45328474e-01 8.39213371e-01 3.29428315e-02 -1.11733705e-01
6.86430514e-01 2.76497811e-01 6.94597047e-03 -3.55906457e-01
-5.79665720e-01 8.47781360e-01 7.15164244e-01 4.03816432e-01
2.22100630e-01 5.06619751e-01 3.10008168e-01 -6.50806785e-01
-7.61162519e-01 -2.29463398e-01 -3.96274179e-01 -2.31602684e-01
1.02153492e+00 -1.05373275e+00 -5.63187957e-01 6.24900579e-01
-8.34746957e-01 7.06827715e-02 3.22656631e-01 7.72064567e-01
3.76661345e-02 2.55694270e-01 -7.67689705e-01 -8.17811310e-01
-4.69724298e-01 -9.67438996e-01 9.95916605e-01 4.55294102e-01
5.25713190e-02 -7.51787484e-01 -5.55801749e-01 3.05992335e-01
5.07821262e-01 5.83545864e-01 9.75415528e-01 -7.86421418e-01
-3.60118747e-01 -9.79458272e-01 -5.30753732e-01 1.45689905e-01
3.66127133e-01 -4.28036541e-01 -7.84542024e-01 -1.63617536e-01
1.20514147e-01 -2.30383560e-01 1.02555335e+00 8.74249458e-01
1.67222059e+00 2.74705321e-01 -9.75624323e-01 5.57590723e-01
1.43446672e+00 9.18187618e-01 7.57192731e-01 3.28536004e-01
3.16384971e-01 9.25562754e-02 3.73316556e-01 2.29719713e-01
1.87830985e-01 -3.54769058e-03 6.67029977e-01 -5.49463987e-01
-8.66629109e-02 -9.89419371e-02 -5.04642546e-01 5.60375452e-01
-3.22942697e-02 -2.77400762e-01 -1.26046813e+00 5.18733859e-01
-1.49310231e+00 -5.14835775e-01 -2.16161415e-01 1.82049227e+00
5.86811960e-01 2.85144717e-01 -3.64448249e-01 4.14343268e-01
6.87827468e-01 -4.19919401e-01 -6.35553837e-01 6.36820719e-02
9.77677107e-02 4.59946424e-01 8.22257578e-01 -2.81671524e-01
-1.29407513e+00 1.67106643e-01 7.11539316e+00 5.72189152e-01
-1.35331178e+00 1.64855376e-01 1.25079262e+00 2.75915414e-01
2.15313002e-01 -5.88113666e-01 -6.66271806e-01 3.01737756e-01
7.62778878e-01 3.57276917e-01 -5.55085361e-01 8.63012910e-01
-4.95049777e-03 -5.40753961e-01 -1.18839717e+00 8.64177644e-01
-9.37456787e-02 -1.44214165e+00 3.71304676e-02 1.64738938e-01
5.18117130e-01 -4.46409136e-01 1.26680639e-02 1.89213604e-01
-5.10503463e-02 -1.32114804e+00 -9.75358579e-03 5.67237079e-01
1.04669726e+00 -5.02166808e-01 1.46582150e+00 4.49871331e-01
-7.97465324e-01 -1.24866515e-01 -1.31263733e-01 1.63021162e-01
-2.37758279e-01 8.12786341e-01 -1.03926265e+00 5.76596379e-01
1.06765151e+00 5.68296552e-01 -6.58036888e-01 1.20732296e+00
3.85684878e-01 9.09958601e-01 -8.42009708e-02 -1.27567798e-01
6.34517744e-02 4.89395231e-01 -5.77848665e-02 1.00279713e+00
7.18474090e-01 3.20272237e-01 1.02855474e-01 5.16759276e-01
-2.72285104e-01 -9.53668356e-02 -5.64390242e-01 -9.91866887e-02
4.41262484e-01 1.15549934e+00 -9.11948621e-01 -5.94612122e-01
-5.19509971e-01 5.43468475e-01 -2.82758236e-01 -2.05842555e-01
-6.96112812e-01 -9.87176895e-02 -2.99121380e-01 1.14856079e-01
1.23271212e-01 3.45412076e-01 -4.48127210e-01 -6.03901863e-01
-1.13742717e-01 -7.65608788e-01 2.30555311e-01 -6.28007650e-01
-1.12733495e+00 3.37332755e-01 2.89408825e-02 -1.18534458e+00
-1.48839038e-03 -9.19737875e-01 -4.97082591e-01 4.42305088e-01
-1.61121702e+00 -1.31731021e+00 -8.16010773e-01 5.19919433e-02
3.50946456e-01 -2.76870251e-01 9.36338902e-01 4.90161598e-01
-5.65635264e-01 7.49989331e-01 1.20568275e-01 5.34515381e-01
3.85719329e-01 -1.02253640e+00 1.04722947e-01 2.70987928e-01
-8.92771959e-01 1.79519609e-01 1.45060763e-01 -8.24316978e-01
-1.39297056e+00 -1.22885394e+00 8.00364852e-01 2.94949919e-01
-1.51562346e-02 3.05742353e-01 -5.68275213e-01 6.69940948e-01
-3.44595835e-02 3.46271127e-01 8.81818116e-01 -4.91551965e-01
5.25512472e-02 -2.92588353e-01 -1.49296653e+00 1.72934636e-01
6.32954180e-01 1.66604683e-01 -1.16019182e-01 4.75836605e-01
4.37088132e-01 -6.95877016e-01 -1.05143523e+00 1.01389647e+00
8.01175952e-01 -1.00767374e+00 7.47831821e-01 -3.15469533e-01
9.66906548e-01 1.94566399e-01 -1.96880624e-01 -7.06341386e-01
-2.25061402e-01 3.33276480e-01 -1.84248969e-01 5.00936210e-01
2.17857167e-01 -6.04414940e-01 9.70475554e-01 1.86958939e-01
-7.08487108e-02 -1.18471193e+00 -8.79607916e-01 -4.32779312e-01
1.40680596e-01 -2.42540628e-01 7.05820799e-01 8.93623531e-01
-3.76235992e-01 -2.74283648e-01 -1.42426267e-01 5.31072281e-02
4.02743578e-01 -3.38220030e-01 4.99693483e-01 -1.20099354e+00
-3.42110783e-01 -4.16898519e-01 -7.23705828e-01 -9.19454634e-01
-6.80017054e-01 -7.60456860e-01 -2.03384757e-01 -1.48962557e+00
7.88212061e-01 -3.22812587e-01 -3.49628210e-01 5.05160868e-01
-2.29992494e-01 6.23103976e-01 -3.39820355e-01 9.29237455e-02
-4.54253368e-02 -9.81894433e-02 1.72008109e+00 -5.06097913e-01
2.55538285e-01 7.10696653e-02 -8.18019271e-01 9.12839651e-01
8.21576774e-01 -2.39349991e-01 8.27502981e-02 -4.85161304e-01
-1.12945132e-01 7.35183954e-01 4.08229291e-01 -1.25607538e+00
1.41771123e-01 -1.00964278e-01 1.29342389e+00 -1.02316713e+00
3.34722877e-01 -7.47329950e-01 2.41866872e-01 1.06634784e+00
-3.90573204e-01 -3.14585000e-01 3.50798577e-01 5.98622859e-01
-2.69146383e-01 -3.05410087e-01 9.34973359e-01 -5.64648449e-01
-4.37245876e-01 4.04736817e-01 -6.89464569e-01 -9.00389373e-01
1.10845101e+00 -3.56240749e-01 -1.59059912e-01 -1.94000691e-01
-7.32219279e-01 6.28267899e-02 -2.30671600e-01 1.31521493e-01
9.08732474e-01 -1.75553715e+00 -8.26207697e-01 2.42600396e-01
5.18009514e-02 1.97729245e-01 2.71564662e-01 1.08676612e+00
-9.15530384e-01 6.37993634e-01 -2.64119625e-01 -6.62549198e-01
-1.52528846e+00 1.56222016e-01 8.01109314e-01 -5.03494859e-01
-5.78221798e-01 8.32914829e-01 3.69093083e-02 -2.11520866e-01
3.93795192e-01 -3.50001454e-01 -4.24279362e-01 -3.84788960e-01
7.90507138e-01 2.11725906e-01 3.70741904e-01 -2.19718352e-01
-2.65678525e-01 1.55937478e-01 -5.83303630e-01 3.24304700e-01
1.17774832e+00 2.63748735e-01 -2.08999693e-01 1.29003286e-01
1.07722402e+00 -5.97102344e-01 -3.60206991e-01 -2.91911334e-01
-1.88796267e-01 -4.37828422e-01 3.71961892e-01 -9.62546349e-01
-1.18000984e+00 9.08860922e-01 1.13800907e+00 3.65551412e-01
1.38491488e+00 -5.76773025e-02 9.63510036e-01 5.75250566e-01
1.92839488e-01 -7.15782166e-01 7.43860900e-02 -2.63101161e-02
6.01978302e-01 -1.63820064e+00 2.90126324e-01 -4.78361100e-01
-1.73422456e-01 1.48873365e+00 8.17257166e-01 -1.33670047e-01
9.71397996e-01 2.54143625e-01 9.64318439e-02 -3.29542726e-01
-4.95921642e-01 1.18317425e-01 1.39087781e-01 5.45969069e-01
7.98440456e-01 4.08376306e-01 -6.61230743e-01 9.91538227e-01
5.68250492e-02 5.52875161e-01 3.32957774e-01 1.36495936e+00
-7.19777465e-01 -5.46196580e-01 -5.99143922e-01 1.26706576e+00
-7.76096582e-01 2.13142350e-01 -5.95619202e-01 1.05283129e+00
4.85963345e-01 8.47891152e-01 3.08153033e-01 -4.06269073e-01
-7.10600168e-02 -6.07886761e-02 4.61461753e-01 -3.70430410e-01
-2.11405471e-01 2.92640239e-01 -9.15349349e-02 -1.20153002e-01
-5.71756721e-01 -2.48120248e-01 -1.07816756e+00 -3.00153613e-01
-5.22119999e-01 -4.15873319e-01 7.03606606e-01 9.50210571e-01
-3.66595387e-02 8.90969038e-01 3.59642953e-01 -5.46008527e-01
-4.83208239e-01 -1.36174572e+00 -6.69827640e-01 -1.69668525e-01
3.46280426e-01 -5.48192918e-01 1.34794295e-01 1.16798736e-01] | [15.24258041381836, -2.5528173446655273] |
c14cc15b-9f84-47bb-aec7-7cf4d037917a | clinical-concept-extraction-with-contextual | 1810.10566 | null | http://arxiv.org/abs/1810.10566v2 | http://arxiv.org/pdf/1810.10566v2.pdf | Clinical Concept Extraction with Contextual Word Embedding | Automatic extraction of clinical concepts is an essential step for turning
the unstructured data within a clinical note into structured and actionable
information. In this work, we propose a clinical concept extraction model for
automatic annotation of clinical problems, treatments, and tests in clinical
notes utilizing domain-specific contextual word embedding. A contextual word
embedding model is first trained on a corpus with a mixture of clinical reports
and relevant Wikipedia pages in the clinical domain. Next, a bidirectional
LSTM-CRF model is trained for clinical concept extraction using the contextual
word embedding model. We tested our proposed model on the I2B2 2010 challenge
dataset. Our proposed model achieved the best performance among reported
baseline models and outperformed the state-of-the-art models by 3.4% in terms
of F1-score. | ['Ioannis Ch. Paschalidis', 'Henghui Zhu', 'Amir Tahmasebi'] | 2018-10-24 | null | null | null | null | ['clinical-concept-extraction'] | ['medical'] | [ 3.77257138e-01 5.60598314e-01 -4.23349380e-01 -3.73027921e-01
-1.17340088e+00 -4.88917939e-02 3.61913800e-01 1.04153168e+00
-8.42863381e-01 7.01224566e-01 7.85028756e-01 -4.67315972e-01
-1.51476279e-01 -4.36305404e-01 -2.65388191e-02 -5.78703880e-01
-8.08309540e-02 5.83165467e-01 -2.00055674e-01 2.69973457e-01
4.48411144e-02 1.97206866e-02 -4.20652658e-01 5.71353197e-01
7.89443731e-01 8.39564085e-01 2.32506350e-01 6.39098525e-01
-2.15144843e-01 9.77810621e-01 -4.97817367e-01 -2.68539727e-01
-3.31742048e-01 -2.86000907e-01 -1.00146306e+00 -6.13891445e-02
-4.42725420e-02 -1.56627908e-01 -2.30751246e-01 9.03115511e-01
5.89566886e-01 -7.16843680e-02 6.74031854e-01 -3.68938416e-01
-8.14610898e-01 5.69824815e-01 -1.85346797e-01 1.25996023e-01
9.53922719e-02 -1.68128416e-01 1.15142405e+00 -1.07755721e+00
1.01444304e+00 6.10090435e-01 5.06838322e-01 7.94246495e-01
-1.08828712e+00 -4.55950946e-01 1.52246311e-01 5.16208299e-02
-1.26935732e+00 -3.31465676e-02 2.46561497e-01 -6.18865609e-01
1.39421153e+00 3.94062176e-02 6.33584440e-01 1.08135366e+00
7.00407088e-01 5.26236594e-01 7.52822101e-01 -4.81865346e-01
1.36183023e-01 1.80936664e-01 4.49605733e-01 6.72439456e-01
2.25236759e-01 -3.80448878e-01 -2.28986844e-01 -5.86517394e-01
2.95515358e-01 4.07744497e-01 -3.08217913e-01 3.01839679e-01
-1.07352293e+00 9.04189587e-01 4.68492508e-01 4.76930529e-01
-6.42359495e-01 -2.33769305e-02 6.08656466e-01 -2.34724775e-01
7.50928938e-01 6.47809088e-01 -6.50279760e-01 -2.87372898e-02
-1.00264347e+00 -1.41770363e-01 6.75432503e-01 8.61924589e-01
-2.96694636e-01 -5.07896304e-01 -5.17573476e-01 9.59246278e-01
4.81992334e-01 1.60700306e-01 7.69787252e-01 -3.44192326e-01
5.95656991e-01 6.92184627e-01 -1.29332513e-01 -7.47490108e-01
-6.62340581e-01 -6.25925958e-01 -7.22693264e-01 -5.60591578e-01
-2.17230916e-01 -3.80144864e-01 -1.17444193e+00 1.34811962e+00
3.05697858e-01 2.24874169e-01 3.92278731e-01 5.43668568e-01
1.32116234e+00 4.98524278e-01 7.09631979e-01 -2.73072749e-01
1.81466842e+00 -1.11067319e+00 -1.29172778e+00 -2.84798324e-01
1.06057680e+00 -7.10563242e-01 3.07470441e-01 1.35542631e-01
-7.53276527e-01 -8.09846073e-03 -8.96131098e-01 -2.24509612e-01
-3.61199915e-01 3.44215602e-01 2.47595251e-01 6.81177676e-02
-8.46619785e-01 3.00627410e-01 -1.26359534e+00 -4.20702964e-01
7.47851431e-01 2.43518949e-01 -6.51931584e-01 -3.30673993e-01
-1.00646889e+00 9.31095898e-01 4.06223804e-01 1.82862669e-01
-6.85842097e-01 -9.37205851e-01 -1.05685544e+00 2.40550831e-01
3.52582574e-01 -9.04311001e-01 1.28803837e+00 2.63793450e-02
-1.11453712e+00 9.18078661e-01 -2.81595200e-01 -4.19868141e-01
-4.99772020e-02 -4.02084023e-01 -6.49971068e-01 3.31168562e-01
-2.20379513e-02 5.18206537e-01 1.26069143e-01 -7.75951624e-01
-5.35589695e-01 -3.24489415e-01 -3.22761565e-01 -7.16812536e-02
-5.12234688e-01 5.98358661e-02 -5.49291611e-01 -6.83527589e-01
-1.08056903e-01 -9.96591270e-01 -8.12391758e-01 -1.24012172e-01
-6.32857680e-01 -3.68070692e-01 2.90269196e-01 -1.04334342e+00
1.61064923e+00 -2.00643253e+00 -9.98339280e-02 3.43563892e-02
6.90910757e-01 3.66339058e-01 -6.80972040e-02 6.91882670e-01
-2.16158837e-01 3.47238839e-01 -2.06531867e-01 -4.91029471e-01
-2.30368033e-01 2.14296374e-02 -1.03108883e-01 3.04759383e-01
5.26671648e-01 1.01313889e+00 -1.16888416e+00 -9.18477476e-01
-1.02826290e-01 8.63573015e-01 -6.77244782e-01 4.10995573e-01
-3.98587510e-02 2.35165775e-01 -6.92279935e-01 5.88191271e-01
2.38869846e-01 -7.20121920e-01 8.13179255e-01 -1.62174642e-01
2.97215134e-01 7.48647273e-01 -4.41866785e-01 1.90044689e+00
-6.60716891e-01 3.39489520e-01 -1.43928126e-01 -7.81583011e-01
6.84123755e-01 9.96958613e-01 9.82543588e-01 -2.39670068e-01
8.18835422e-02 5.97089268e-02 2.32605357e-02 -7.87020504e-01
3.38629663e-01 -3.85804355e-01 -9.03296769e-02 2.74488956e-01
3.22305024e-01 2.27826267e-01 -1.58866242e-01 4.08624887e-01
1.54088819e+00 -1.44150019e-01 8.35415065e-01 -1.96283147e-01
4.47061121e-01 -2.31818035e-02 9.20207500e-01 3.23593915e-01
-2.31375575e-01 4.95699912e-01 4.20833945e-01 -3.65773052e-01
-4.85328525e-01 -8.13344240e-01 -3.99627000e-01 4.55199659e-01
-5.49666524e-01 -1.02080345e+00 -4.14329857e-01 -9.49462891e-01
-2.65133679e-01 5.96717000e-01 -8.56354356e-01 -8.21006969e-02
-4.02302057e-01 -7.91761458e-01 2.77876288e-01 7.91562915e-01
-9.21499655e-02 -1.05611610e+00 -4.73610252e-01 6.68545723e-01
-2.17092887e-01 -1.53502381e+00 -6.68376148e-01 3.08286607e-01
-1.06001997e+00 -1.21142185e+00 -5.60260534e-01 -1.12816584e+00
9.25564349e-01 -4.18853819e-01 1.04259181e+00 8.75220075e-02
-5.49806237e-01 1.53581306e-01 -4.54041094e-01 -5.78829825e-01
-1.58246592e-01 2.06711024e-01 -4.16887194e-01 -3.24648201e-01
7.86676288e-01 -6.36518002e-02 -8.86364460e-01 -4.34142888e-01
-1.02085352e+00 8.99894536e-02 6.10610485e-01 1.26672900e+00
9.06812012e-01 -6.10067189e-01 6.68989718e-01 -1.45158923e+00
7.99608171e-01 -5.88090122e-01 -1.11242514e-02 3.46047074e-01
-9.81130898e-01 1.08379364e-01 3.68776202e-01 -1.31998524e-01
-9.39675748e-01 1.05334058e-01 -3.93833041e-01 3.28273699e-02
4.44288831e-03 1.08473229e+00 2.70310134e-01 5.83945870e-01
4.29868966e-01 -6.21759482e-02 -3.08863491e-01 -6.59416318e-01
3.77712578e-01 1.07824242e+00 2.55785972e-01 -1.74062207e-01
1.75233141e-01 2.37701371e-01 -6.23152331e-02 -5.55262268e-01
-1.20306861e+00 -8.31506312e-01 -6.11561596e-01 3.83346766e-01
1.44818366e+00 -9.18714345e-01 -3.09205115e-01 -2.89235681e-01
-1.31167495e+00 -1.03645064e-01 -2.87491471e-01 8.07590127e-01
-1.41717076e-01 9.37669277e-02 -1.12210774e+00 -3.85830581e-01
-9.14351761e-01 -1.19939339e+00 9.97329354e-01 -1.48495641e-02
-4.63603169e-01 -1.35676110e+00 5.90741456e-01 3.25822830e-01
2.01811641e-01 3.42451960e-01 1.16851914e+00 -1.12169158e+00
7.10583990e-03 -4.32744831e-01 -2.13457689e-01 1.49796233e-01
5.79435170e-01 -2.55203396e-01 -7.80596375e-01 -2.64866382e-01
-1.26425505e-01 -2.13966250e-01 1.09368491e+00 3.72056961e-01
1.33389211e+00 -4.74072546e-01 -8.36280465e-01 5.50144911e-01
1.55239701e+00 2.50437886e-01 3.33251625e-01 1.60269901e-01
5.87361753e-01 4.06344533e-01 4.91671920e-01 4.28369015e-01
3.77305478e-01 4.25242633e-01 1.15061454e-01 -2.83315301e-01
-1.49626415e-02 -1.11123115e-01 -1.00544512e-01 1.37409294e+00
1.39691144e-01 -2.59410411e-01 -1.22645450e+00 1.02076948e+00
-1.74441743e+00 -4.87165719e-01 1.64194077e-01 1.69918311e+00
1.28785038e+00 7.06334412e-02 -5.64117789e-01 -4.87825781e-01
4.16571379e-01 -7.58877322e-02 -1.49253488e-01 -5.37982941e-01
4.65447098e-01 7.50514746e-01 4.26029772e-01 4.98818457e-01
-1.20439708e+00 8.29756081e-01 6.24432325e+00 5.55468798e-01
-9.25285101e-01 4.68726307e-01 5.36152363e-01 -3.07374537e-01
-8.17699954e-02 -2.33885229e-01 -9.08545315e-01 3.68418425e-01
1.16512942e+00 -1.09993652e-01 -4.97768670e-01 7.36916780e-01
1.76090434e-01 2.76334852e-01 -1.14278662e+00 6.23088717e-01
1.96519241e-01 -1.63395858e+00 -1.51850149e-01 3.23143184e-01
7.81705678e-01 2.22405374e-01 -2.68903617e-02 2.36491203e-01
4.80926096e-01 -1.30637062e+00 -2.37140223e-01 5.93721986e-01
1.22185409e+00 -2.50842780e-01 1.34913146e+00 -9.69294086e-02
-1.01404488e+00 9.83791649e-02 -1.17122434e-01 5.66812575e-01
3.74837279e-01 5.97120583e-01 -1.50540531e+00 7.25581706e-01
4.21525210e-01 1.02288353e+00 -3.37211698e-01 1.01366973e+00
-5.02084434e-01 9.56155062e-01 1.71715334e-01 1.10775366e-01
4.57825273e-01 1.69256870e-02 3.86639014e-02 1.72308660e+00
1.84974238e-01 5.21467507e-01 3.70666176e-01 4.49424416e-01
-4.66675907e-01 5.51797330e-01 -4.94271129e-01 -4.72737908e-01
2.58786947e-01 1.58172774e+00 -6.01676106e-01 -5.82860947e-01
-6.34376585e-01 6.57939196e-01 3.97278070e-01 2.11080611e-01
-6.20388746e-01 -4.50688928e-01 4.60583955e-01 -6.64338544e-02
4.16955709e-01 4.07995135e-02 -6.34592548e-02 -1.11314166e+00
-1.90125152e-01 -7.03580141e-01 7.99264014e-01 -4.62200731e-01
-1.44478881e+00 9.80008185e-01 -3.96399528e-01 -1.26553249e+00
-3.60731781e-01 -7.43838906e-01 -2.95672923e-01 8.40402663e-01
-1.60430562e+00 -1.18613613e+00 -1.40905783e-01 1.49785295e-01
4.91588861e-01 -2.49956101e-01 1.50598073e+00 4.12801206e-01
-5.30038714e-01 4.72281873e-01 1.75070584e-01 6.76096022e-01
1.02695453e+00 -1.21955514e+00 1.69816017e-01 4.28492695e-01
1.24819474e-02 1.08356619e+00 2.10758597e-01 -7.75712788e-01
-6.91772759e-01 -1.41319644e+00 1.60513282e+00 -4.41163361e-01
7.43032753e-01 -1.16626643e-01 -9.94643509e-01 8.77625883e-01
4.30420935e-01 1.90326408e-01 1.59608233e+00 2.36461997e-01
-2.23211423e-01 1.76914051e-01 -9.61576164e-01 4.26307738e-01
6.77516699e-01 -7.35207438e-01 -9.27968562e-01 6.23216808e-01
1.18396115e+00 -3.21915686e-01 -1.60858941e+00 3.36213022e-01
3.58506292e-01 2.45201617e-01 8.97371709e-01 -1.25075853e+00
8.07473421e-01 1.94760755e-01 -9.27517638e-02 -1.33439398e+00
-3.67022634e-01 -2.00779125e-01 -9.00565907e-02 7.80880988e-01
8.34657371e-01 -4.57199335e-01 5.11113405e-01 5.80177903e-01
-2.80563742e-01 -1.40192533e+00 -7.72439420e-01 -2.27623850e-01
2.27653787e-01 -2.43996799e-01 9.06268805e-02 1.10570920e+00
6.24046445e-01 4.91290003e-01 6.53364658e-02 -2.77919974e-02
2.59892076e-01 1.11096926e-01 -8.83569494e-02 -1.02915585e+00
-1.92613363e-01 -9.68591645e-02 -2.99486905e-01 -7.13321149e-01
1.86498031e-01 -1.22751832e+00 7.99461678e-02 -2.20005703e+00
6.69241667e-01 -2.53897309e-01 -9.61813092e-01 7.94606805e-01
-5.53635240e-01 4.01809998e-02 -2.54942011e-02 6.15818873e-02
-6.88923776e-01 3.98572147e-01 9.95755732e-01 -3.92027289e-01
-2.58499831e-01 -4.02472168e-01 -7.78222263e-01 4.39170450e-01
6.42254591e-01 -1.11840129e+00 -1.05222344e-01 -5.40865898e-01
9.05918181e-02 2.12506607e-01 -7.14080036e-02 -5.79238594e-01
4.46424484e-01 -8.21806937e-02 3.28859426e-02 -4.93491977e-01
2.06276491e-01 -7.39630520e-01 -4.85455215e-01 7.00982273e-01
-8.37423086e-01 9.55651551e-02 3.60195458e-01 6.87209427e-01
-4.54936296e-01 -2.30960771e-01 2.89953858e-01 -4.48052818e-03
-1.40687972e-01 4.31827664e-01 -5.18599212e-01 2.81382114e-01
7.71569610e-01 4.47546184e-01 -3.47895056e-01 -1.58132259e-02
-1.14715075e+00 3.80765706e-01 -1.73439369e-01 2.91600376e-01
8.24614108e-01 -1.13295054e+00 -8.88201237e-01 -2.52925247e-01
4.63174552e-01 8.95458311e-02 2.96713978e-01 1.04738939e+00
-5.73687315e-01 9.48811650e-01 2.23528862e-01 -2.79699206e-01
-1.55364132e+00 5.80366194e-01 1.47893474e-01 -1.15030134e+00
-7.79322505e-01 7.02305496e-01 2.47728452e-01 -3.31185669e-01
1.72752947e-01 -7.05332935e-01 -5.75719953e-01 -1.26079470e-01
6.35981500e-01 -4.40950930e-01 2.77999163e-01 -4.04290169e-01
-6.57162488e-01 3.68537754e-01 -4.97739404e-01 -1.27788201e-01
1.90627742e+00 3.49133402e-01 -6.80122226e-02 1.97005674e-01
1.43682587e+00 1.43070165e-02 -5.91092646e-01 -6.30573511e-01
3.12770486e-01 1.30024388e-01 3.52687269e-01 -9.80556667e-01
-8.98667991e-01 1.04894352e+00 4.38816279e-01 -3.72094959e-01
1.03734875e+00 1.54309586e-01 8.40018332e-01 4.06753242e-01
-2.22412571e-01 -1.00711536e+00 7.48635903e-02 3.50383937e-01
6.12980425e-01 -1.29162610e+00 2.77964175e-01 -5.13750196e-01
-9.00445819e-01 1.08219147e+00 2.85498619e-01 1.04333095e-01
1.05440474e+00 2.89556384e-01 2.75001824e-01 -6.01762533e-01
-1.17947781e+00 -1.00588868e-03 6.29647017e-01 2.95554161e-01
9.16157484e-01 1.24857694e-01 -6.89186931e-01 1.14577925e+00
3.48148525e-01 3.67678463e-01 3.50649983e-01 1.10763955e+00
1.24887619e-02 -1.29962480e+00 3.42765480e-01 4.67361540e-01
-1.03553402e+00 -6.70544982e-01 -2.66540229e-01 6.10274911e-01
1.31544128e-01 8.82346690e-01 -8.20850506e-02 -3.11940044e-01
1.73340559e-01 4.52734113e-01 7.11014196e-02 -1.38906085e+00
-8.00571561e-01 2.26884753e-01 2.46524349e-01 -5.45863271e-01
-4.77632344e-01 -3.77523988e-01 -1.51661801e+00 5.55454075e-01
-3.15901697e-01 3.53679240e-01 6.43963575e-01 8.80187869e-01
7.43439257e-01 1.05184817e+00 1.04526907e-01 2.25963950e-01
-3.50570530e-01 -1.22744906e+00 -1.70536503e-01 3.44734669e-01
3.11000735e-01 -1.82201371e-01 1.15140498e-01 3.40941608e-01] | [8.452741622924805, 8.711801528930664] |
cd975421-ca40-452b-8537-bad868af2861 | correlation-driven-multi-level-multimodal | 2305.02323 | null | https://arxiv.org/abs/2305.02323v1 | https://arxiv.org/pdf/2305.02323v1.pdf | Correlation-Driven Multi-Level Multimodal Learning for Anomaly Detection on Multiple Energy Sources | Advanced metering infrastructure (AMI) has been widely used as an intelligent energy consumption measurement system. Electric power was the representative energy source that can be collected by AMI; most existing studies to detect abnormal energy consumption have focused on a single energy source, i.e., power. Recently, other energy sources such as water, gas, and heating have also been actively collected. As a result, it is necessary to develop a unified methodology for anomaly detection across multiple energy sources; however, research efforts have rarely been made to tackle this issue. The inherent difficulty with this issue stems from the fact that anomalies are not usually annotated. Moreover, existing works of anomaly definition depend on only individual energy sources. In this paper, we first propose a method for defining anomalies considering not only individual energy sources but also correlations between them. Then, we propose a new Correlation-driven Multi-Level Multimodal Learning model for anomaly detection on multiple energy sources. The distinguishing property of the model incorporates multiple energy sources in multi-levels based on the strengths of the correlations between them. Furthermore, we generalize the proposed model in order to integrate arbitrary new energy sources with further performance improvement, considering not only correlated but also non-correlated sources. Through extensive experiments on real-world datasets consisting of three to five energy sources, we demonstrate that the proposed model clearly outperforms the existing multimodal learning and recent time-series anomaly detection models, and we observe that our model makes further the performance improvement as more correlated or non-correlated energy sources are integrated. | ['Hyuk-Yoon Kwon', 'Taehee Kim'] | 2023-05-01 | null | null | null | null | ['time-series-anomaly-detection'] | ['time-series'] | [-1.11650437e-01 -5.63571274e-01 -1.83710381e-01 -1.04635425e-01
-8.01803946e-01 -6.10202610e-01 6.34051561e-01 8.53120923e-01
-4.80539128e-02 5.53377330e-01 2.16829658e-01 -2.18484905e-02
-2.36005321e-01 -9.28378344e-01 -2.93248326e-01 -9.94348586e-01
-1.69740185e-01 1.62305772e-01 -8.91640186e-02 -9.33931619e-02
5.83772697e-02 4.82018948e-01 -1.43591154e+00 -1.92575604e-01
1.10747242e+00 1.21712267e+00 -3.52364421e-01 2.25595251e-01
-2.35742643e-01 4.59324092e-01 -8.76454055e-01 1.44489378e-01
1.41965941e-01 -6.11349821e-01 -4.34234709e-01 1.92887112e-01
-3.77783805e-01 -8.55394974e-02 -3.28070670e-02 1.29041576e+00
4.01389331e-01 1.80125266e-01 6.64722860e-01 -1.84207022e+00
-5.35887599e-01 8.29374075e-01 -8.73873770e-01 4.32588875e-01
3.23547691e-01 1.24144712e-02 1.03934801e+00 -5.71852624e-01
-3.07832330e-01 7.22915828e-01 1.78829387e-01 1.02966532e-01
-9.23921406e-01 -7.06471384e-01 4.65448022e-01 7.01409340e-01
-1.35020494e+00 8.90253559e-02 1.54416382e+00 -1.47247910e-01
9.24191117e-01 4.06860769e-01 8.37177753e-01 9.64724898e-01
-8.19465145e-02 8.28674495e-01 9.43953931e-01 -4.23262537e-01
5.33008218e-01 -2.84227766e-02 1.67290837e-01 2.16383353e-01
5.16472101e-01 -3.33325237e-01 -2.14157507e-01 -3.83627951e-01
-2.50456650e-02 2.36408800e-01 -1.43345237e-01 -2.79271305e-01
-1.10128391e+00 6.26914442e-01 1.77100390e-01 9.62429106e-01
-3.19236010e-01 -1.82062954e-01 5.17934740e-01 1.41148433e-01
5.19928396e-01 1.07064247e-01 -3.26193333e-01 -3.51586282e-01
-8.00500453e-01 -3.72596502e-01 6.73895419e-01 6.67653799e-01
7.70715475e-01 5.49832106e-01 1.67682260e-01 6.94573641e-01
3.72886568e-01 5.93578994e-01 6.96020722e-01 -3.56530786e-01
5.33976912e-01 1.15028799e+00 7.26209432e-02 -1.18816197e+00
-7.42824256e-01 -3.25688273e-01 -1.37241375e+00 -6.91603720e-02
2.12113202e-01 -2.08342835e-01 -5.18972456e-01 1.67931747e+00
3.24800223e-01 4.10335839e-01 2.05217227e-01 7.07007587e-01
4.69855368e-01 6.50116920e-01 2.60772109e-01 -6.79548800e-01
1.20184886e+00 -4.98246551e-01 -1.00105667e+00 2.36381561e-01
6.90946043e-01 -4.26734895e-01 5.82149088e-01 3.64559144e-01
-7.92534828e-01 -1.85121015e-01 -1.22577345e+00 5.33074081e-01
-9.08515453e-01 -7.48772100e-02 4.69365180e-01 6.74279690e-01
-5.48144221e-01 3.65627170e-01 -8.96978915e-01 -3.60060900e-01
1.62377745e-01 8.30976665e-02 -9.58209112e-03 3.46479446e-01
-1.31977797e+00 8.79918575e-01 5.99149823e-01 7.99048468e-02
-3.58907133e-01 -4.28883225e-01 -9.11589861e-01 1.00536801e-01
3.96339178e-01 -2.44190678e-01 7.65775621e-01 -8.20322394e-01
-1.12082005e+00 3.55324484e-02 -2.29450390e-01 -2.26682499e-01
2.01825559e-01 -2.05818769e-02 -1.39710581e+00 -7.15681491e-03
-5.44419251e-02 -2.11479291e-01 4.33905572e-01 -1.34853280e+00
-8.46219122e-01 -4.59299833e-01 -1.46971822e-01 7.41779199e-03
-7.53872335e-01 -1.54021233e-01 1.72700897e-01 -5.89041054e-01
1.20576613e-01 -5.43841779e-01 2.42142417e-02 -6.79138482e-01
-5.25952756e-01 -6.86233699e-01 1.10765135e+00 -5.06153166e-01
1.56033337e+00 -2.12862158e+00 1.18733838e-01 6.48112655e-01
1.14689425e-01 1.17365293e-01 1.61286220e-01 5.93591511e-01
-3.33490908e-01 1.09491609e-01 -6.09229088e-01 -9.08136889e-02
4.38576728e-01 2.82655358e-01 -1.03653178e-01 4.39690918e-01
1.81456938e-01 7.60208607e-01 -1.15864432e+00 -3.28819960e-01
6.14596546e-01 3.14413935e-01 2.85979211e-01 3.33836824e-02
3.53466123e-02 5.35560846e-01 -5.43470681e-01 1.05074036e+00
6.98862731e-01 -1.03916138e-01 -7.65273497e-02 -4.40059036e-01
-1.74340248e-01 -3.98610592e-01 -1.39496529e+00 1.46004736e+00
-3.45928520e-01 1.21748157e-01 -2.65997946e-01 -1.46554518e+00
8.78005624e-01 6.25889421e-01 1.31310391e+00 -7.00875342e-01
2.16610596e-01 2.67917335e-01 2.14778595e-02 -5.72903752e-01
2.48982549e-01 6.76169321e-02 -3.35105032e-01 4.92967069e-01
-2.84212232e-01 2.96758085e-01 5.34357488e-01 -1.89344678e-02
1.01433933e+00 -3.59294385e-01 5.92560947e-01 -1.44107966e-02
9.23344672e-01 -1.72230944e-01 6.13949239e-01 3.22968125e-01
-2.66189575e-01 1.86954841e-01 1.94346428e-01 -1.25020325e-01
-8.20098758e-01 -1.07785904e+00 -1.69230342e-01 8.03097486e-01
2.63967633e-01 -4.68670696e-01 -3.79104823e-01 -9.18544054e-01
3.92400660e-02 1.06347048e+00 -3.16876441e-01 -2.36069128e-01
-3.93149704e-01 -1.48233581e+00 6.04273021e-01 6.05378211e-01
6.45440876e-01 -8.20942581e-01 -4.19642776e-01 2.52951652e-01
-3.89819115e-01 -9.80011106e-01 -2.34021172e-01 1.98984802e-01
-6.53871059e-01 -1.17836833e+00 -4.81482923e-01 -2.58607417e-01
6.49629056e-01 3.27873975e-02 9.81554687e-01 -5.78330085e-02
-1.18083425e-01 7.62831926e-01 -5.09049594e-01 -5.25762737e-01
-2.78522760e-01 -8.50040540e-02 2.73749679e-01 4.98788148e-01
7.57485926e-01 -9.20210123e-01 -6.08096182e-01 2.43616506e-01
-1.21245563e+00 -5.03008962e-01 6.21128976e-01 4.60541815e-01
4.58940655e-01 6.17123067e-01 1.25510478e+00 -3.12941700e-01
6.43467963e-01 -1.05914450e+00 -5.20044029e-01 2.46946126e-01
-9.19798493e-01 -9.90750119e-02 8.82781804e-01 -2.98440397e-01
-8.46246004e-01 -1.03582650e-01 -1.53738618e-01 -2.41992906e-01
-7.21601903e-01 4.62454259e-01 -5.61666429e-01 1.67744532e-01
-3.33091542e-02 4.29205060e-01 -6.23118699e-01 -4.67891484e-01
3.09009761e-01 5.67961931e-01 2.71984339e-01 -5.62775493e-01
1.13389552e+00 3.30423176e-01 3.00019562e-01 -8.32581282e-01
-3.73118848e-01 -6.49903953e-01 -6.77171290e-01 -1.53448910e-01
7.44224072e-01 -6.96430862e-01 -8.01274717e-01 5.95859468e-01
-8.89488280e-01 4.41614032e-01 -4.21832174e-01 5.49203932e-01
-1.23949358e-02 7.89691806e-01 -2.09018558e-01 -1.09470081e+00
-3.23849857e-01 -8.47830594e-01 6.89351320e-01 3.19229871e-01
-1.91968799e-01 -1.39016831e+00 2.17302278e-01 -1.08233370e-01
5.72510183e-01 6.23921573e-01 1.17304707e+00 -1.14475787e+00
-4.10547554e-01 -2.39063174e-01 -9.98743922e-02 2.66455978e-01
7.92676389e-01 -2.66751021e-01 -6.53811634e-01 -4.15375322e-01
1.75719380e-01 4.21131253e-02 5.81156135e-01 -1.42953202e-01
1.25054526e+00 -2.78708130e-01 -2.36567348e-01 2.30359674e-01
1.38650334e+00 5.62984169e-01 3.92387480e-01 4.25423086e-01
9.47554290e-01 1.96616247e-01 2.46326402e-01 6.48674786e-01
5.96110344e-01 4.59043503e-01 6.11205876e-01 -1.40943855e-01
5.46594679e-01 6.21607676e-02 4.36219484e-01 1.32464230e+00
-2.04044908e-01 -3.78385574e-01 -6.77547038e-01 7.59841800e-01
-1.97221851e+00 -1.04853559e+00 -1.05555050e-01 2.13683081e+00
4.89542454e-01 -1.49117529e-01 3.30261260e-01 6.66335821e-01
4.38851237e-01 2.49919146e-01 -8.52276385e-01 -1.39099538e-01
-4.55799729e-01 -6.96373582e-02 1.85242385e-01 6.88024685e-02
-1.17779362e+00 6.44967752e-03 5.23872042e+00 7.50254393e-01
-9.46905851e-01 8.83287489e-02 3.98581743e-01 -1.16986267e-01
-4.64914113e-01 -3.68047476e-01 -2.97985554e-01 9.53099191e-01
7.45415926e-01 -4.98354077e-01 1.77476391e-01 6.07906759e-01
2.30845332e-01 -2.08089069e-01 -1.14788556e+00 1.00367856e+00
1.73688963e-01 -4.42078352e-01 4.20966186e-02 7.23172948e-02
8.05815458e-01 -1.50500059e-01 -2.87972242e-01 6.09076843e-02
1.67470202e-01 -7.80925333e-01 3.07119995e-01 5.38924396e-01
8.85428190e-02 -1.00797927e+00 1.04849148e+00 3.05880934e-01
-1.64191663e+00 -2.43140370e-01 3.44638266e-02 2.01449856e-01
2.42485672e-01 8.46580565e-01 -2.33351409e-01 1.38152385e+00
7.23068416e-01 7.79227078e-01 -6.20202422e-01 1.06291592e+00
-1.21346228e-01 7.50524640e-01 -5.26952207e-01 6.53704628e-02
3.08237355e-02 -3.64842534e-01 5.60871303e-01 1.18702686e+00
8.61274362e-01 1.77838951e-01 3.99309814e-01 6.25093639e-01
1.28751025e-01 2.68859446e-01 -6.77814126e-01 1.97580338e-01
4.95288193e-01 1.52876437e+00 -7.26359487e-01 -2.80488461e-01
-7.84908175e-01 6.79245591e-01 -1.62375778e-01 5.69217384e-01
-8.79666150e-01 -3.52183342e-01 6.05065703e-01 -3.32879692e-01
6.58538640e-02 -2.10940525e-01 -1.66122764e-01 -1.40152085e+00
1.74655825e-01 -5.21601081e-01 7.36299813e-01 -3.17515016e-01
-1.86769307e+00 3.05818617e-01 3.11950237e-01 -1.48987138e+00
-3.95538181e-01 -1.40465736e-01 -9.24220204e-01 6.70589566e-01
-1.72589660e+00 -1.05164981e+00 -4.02063608e-01 8.97943795e-01
3.11262935e-01 -1.58600703e-01 9.30047572e-01 4.48355258e-01
-8.81500363e-01 5.89940131e-01 3.16335201e-01 3.20830286e-01
2.74120808e-01 -1.58829319e+00 -1.86671838e-01 1.07458198e+00
2.83278555e-01 9.28966627e-02 5.41308105e-01 -5.45819581e-01
-1.30477393e+00 -1.01416457e+00 4.77519065e-01 -2.31588736e-01
9.10360515e-01 -6.44837469e-02 -1.21059787e+00 5.67135513e-01
6.41803622e-01 -7.46971071e-02 9.26263273e-01 -6.24639876e-02
-6.53868541e-02 -4.60465193e-01 -1.33687973e+00 2.89608687e-01
6.67937934e-01 -2.39440471e-01 -6.64130986e-01 1.19480520e-01
1.41553387e-01 8.53170007e-02 -1.12936771e+00 6.16318882e-01
-5.02547286e-02 -7.05249488e-01 7.33243227e-01 -1.92486688e-01
-1.31202102e-01 -7.73877859e-01 -2.93310016e-01 -1.71375239e+00
-2.22180918e-01 -2.23513559e-01 -6.84701562e-01 1.76909876e+00
1.94156259e-01 -1.04161775e+00 3.69690537e-01 4.18067664e-01
-1.31243393e-01 -5.71298659e-01 -1.23781371e+00 -8.41719091e-01
4.39684093e-02 -6.72163129e-01 1.07000732e+00 1.14099169e+00
5.80680549e-01 7.51306862e-02 -3.11646402e-01 6.66038394e-01
9.55229282e-01 2.72613198e-01 2.73141116e-01 -1.27663696e+00
1.09648049e-01 -6.70531929e-01 -4.91350949e-01 -5.94198704e-01
2.42258888e-02 -8.16953063e-01 -2.46188149e-01 -1.52834463e+00
6.68668672e-02 -2.02996880e-01 -1.02083266e+00 5.39281666e-01
-1.75173372e-01 1.77684322e-01 3.76976877e-02 1.47091776e-01
-6.65616870e-01 8.04184794e-01 5.32988906e-01 -3.99025291e-01
-1.38569221e-01 6.18430711e-02 -5.12775302e-01 1.01255810e+00
1.24382019e+00 -2.08211973e-01 -3.66043121e-01 -8.95963833e-02
1.08790606e-01 -2.19429240e-01 2.96039790e-01 -1.16908503e+00
4.82190579e-01 -1.95439413e-01 5.73602974e-01 -9.50770199e-01
1.30323976e-01 -1.36848414e+00 2.69234449e-01 9.10476968e-02
1.60132706e-01 5.33763945e-01 1.64508745e-01 6.72545969e-01
-3.06922734e-01 -3.00646182e-02 3.31567138e-01 3.47244292e-02
-9.05400932e-01 1.48850784e-01 -5.52876830e-01 -1.19829476e-01
1.34762025e+00 1.26770303e-01 -3.96808982e-01 -3.79764766e-01
-7.53161073e-01 6.63092077e-01 1.68364227e-01 6.46185458e-01
4.87005144e-01 -1.77833080e+00 -5.97339213e-01 1.84416085e-01
3.73834968e-01 -3.06168169e-01 2.31054977e-01 9.77426589e-01
2.97024012e-01 3.31654727e-01 2.24133562e-02 -6.61881030e-01
-9.67577040e-01 6.70681775e-01 2.93780953e-01 -3.37514430e-01
-5.11506259e-01 -1.57120273e-01 -3.02795291e-01 -2.79901773e-02
2.62560666e-01 -4.37257469e-01 -5.80999315e-01 4.61562455e-01
3.75449598e-01 7.72501767e-01 2.05301210e-01 -8.73093009e-01
-4.96982098e-01 8.94984484e-01 3.70595366e-01 2.92883992e-01
1.11310697e+00 -4.69562888e-01 -2.20537692e-01 1.04470277e+00
8.87250960e-01 2.27139238e-02 -5.48644662e-01 -3.08068812e-01
4.40298557e-01 -2.14997903e-01 -2.74460077e-01 -9.35901523e-01
-1.34382343e+00 5.73403537e-01 7.65517235e-01 9.61841524e-01
1.56244254e+00 4.28329371e-02 9.37594175e-01 2.25554973e-01
5.28083205e-01 -1.17030919e+00 -5.43137677e-02 5.99591844e-02
4.09841418e-01 -1.31003726e+00 -9.26361308e-02 -1.02331370e-01
-2.76360363e-01 1.06956756e+00 4.73364890e-01 3.45059454e-01
8.14675927e-01 1.05547905e-01 5.80007732e-02 -9.74431932e-02
-2.30566040e-01 -3.94677848e-01 3.00326437e-01 4.82710987e-01
2.01689214e-01 1.72092319e-01 -2.45393351e-01 8.37041795e-01
1.40633464e-01 -2.53297806e-01 3.73430699e-01 8.90768349e-01
-2.88587183e-01 -1.23246777e+00 -5.80856383e-01 4.28908646e-01
-5.67390084e-01 2.47230649e-01 -4.29714322e-02 6.37662292e-01
4.98095989e-01 1.46613753e+00 3.22456472e-02 -2.60273188e-01
6.01170719e-01 3.78868014e-01 -2.18366482e-03 -1.15502737e-01
-2.99513966e-01 9.28589201e-04 -3.91085744e-01 -3.65144998e-01
-7.53073335e-01 -6.68625474e-01 -1.39865792e+00 -2.12449104e-01
-2.17165723e-01 6.25162780e-01 5.24682283e-01 1.22005630e+00
2.03029364e-01 7.12560475e-01 1.00106180e+00 -6.63112104e-01
-2.36948475e-01 -1.06599367e+00 -8.01163375e-01 6.89042270e-01
4.47528988e-01 -6.91124260e-01 -8.13219965e-01 -2.05117330e-01] | [6.276374340057373, 2.632408857345581] |
2fcc8b9a-4012-4a30-a622-8e0cf1f6003f | deepface-emd-re-ranking-using-patch-wise | 2112.04016 | null | https://arxiv.org/abs/2112.04016v2 | https://arxiv.org/pdf/2112.04016v2.pdf | DeepFace-EMD: Re-ranking Using Patch-wise Earth Mover's Distance Improves Out-Of-Distribution Face Identification | Face identification (FI) is ubiquitous and drives many high-stake decisions made by law enforcement. State-of-the-art FI approaches compare two images by taking the cosine similarity between their image embeddings. Yet, such an approach suffers from poor out-of-distribution (OOD) generalization to new types of images (e.g., when a query face is masked, cropped, or rotated) not included in the training set or the gallery. Here, we propose a re-ranking approach that compares two faces using the Earth Mover's Distance on the deep, spatial features of image patches. Our extra comparison stage explicitly examines image similarity at a fine-grained level (e.g., eyes to eyes) and is more robust to OOD perturbations and occlusions than traditional FI. Interestingly, without finetuning feature extractors, our method consistently improves the accuracy on all tested OOD queries: masked, cropped, rotated, and adversarial while obtaining similar results on in-distribution images. | ['Anh Nguyen', 'Hai Phan'] | 2021-12-07 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Phan_DeepFace-EMD_Re-Ranking_Using_Patch-Wise_Earth_Movers_Distance_Improves_Out-of-Distribution_Face_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Phan_DeepFace-EMD_Re-Ranking_Using_Patch-Wise_Earth_Movers_Distance_Improves_Out-of-Distribution_Face_CVPR_2022_paper.pdf | cvpr-2022-1 | ['face-identification'] | ['computer-vision'] | [ 1.81647167e-01 -3.65931779e-01 -5.85712604e-02 -4.56440121e-01
-5.38743734e-01 -1.01949275e+00 8.53139341e-01 -1.79267198e-01
-4.52614546e-01 2.54069090e-01 -1.80815235e-02 -6.17200322e-02
-1.08385265e-01 -5.66720963e-01 -6.52046978e-01 -6.27074778e-01
-1.22755393e-01 1.71783060e-01 8.94384366e-03 -3.39906663e-02
3.93310606e-01 9.43150342e-01 -1.81829190e+00 1.34372309e-01
2.49325931e-01 1.33126628e+00 -3.77534926e-01 5.23957253e-01
2.29376584e-01 3.76626074e-01 -8.68958414e-01 -7.25645244e-01
7.74091899e-01 -1.08280167e-01 -4.46442097e-01 -9.72871333e-02
1.38731539e+00 -4.73028451e-01 -5.66827416e-01 1.27746093e+00
6.81428611e-01 -2.72344928e-02 9.48167324e-01 -1.49764431e+00
-1.23833907e+00 -3.14629734e-01 -9.13522363e-01 3.42271656e-01
6.06123507e-01 2.61680543e-01 5.21711349e-01 -1.23033679e+00
6.89914286e-01 1.59277129e+00 6.06657565e-01 6.80588186e-01
-1.52123201e+00 -9.96702671e-01 -3.70364971e-02 2.37222224e-01
-1.75108457e+00 -7.64538407e-01 6.07673585e-01 -5.81083536e-01
5.21683455e-01 3.40978056e-01 -1.25317024e-02 1.08703089e+00
3.40163074e-02 2.36196250e-01 1.25584018e+00 -3.57593358e-01
1.18657209e-01 2.04339013e-01 -2.45579645e-01 4.47180092e-01
3.85888308e-01 1.92937106e-01 -6.09584987e-01 -3.71163219e-01
6.20401740e-01 6.66649193e-02 -3.62326354e-01 -4.23210979e-01
-7.71244347e-01 7.08364427e-01 3.66754293e-01 2.20678493e-01
-1.17545098e-01 -8.33245665e-02 7.16418326e-02 4.02040750e-01
4.04650301e-01 4.14875209e-01 -2.40694210e-01 2.61964295e-02
-9.81773734e-01 2.50335455e-01 5.23842514e-01 6.87412798e-01
9.90850985e-01 -2.18665197e-01 -3.34167272e-01 7.92265892e-01
6.92139715e-02 7.73768783e-01 3.96575719e-01 -7.17278898e-01
3.15980613e-01 3.89549255e-01 6.93421811e-02 -1.52840519e+00
-3.15675400e-02 -5.32982424e-02 -7.32151568e-01 6.26524866e-01
4.86251324e-01 3.62956002e-02 -1.02306283e+00 1.90627527e+00
4.89383012e-01 1.66610554e-01 -2.62086183e-01 8.34353685e-01
8.70945930e-01 6.80357143e-02 -3.15250101e-04 3.68732587e-02
1.43365824e+00 -5.04620552e-01 -5.26744008e-01 -2.79682875e-01
1.63959250e-01 -8.68994236e-01 9.57262337e-01 1.92087159e-01
-8.45366120e-01 -7.41242051e-01 -9.94120181e-01 4.46860529e-02
-6.70966327e-01 4.15129680e-03 1.38928980e-01 1.00926423e+00
-1.35603392e+00 6.47738218e-01 -2.56691754e-01 -4.01751131e-01
7.83513665e-01 6.29488826e-01 -1.04988575e+00 -4.87846166e-01
-1.01571083e+00 8.05187821e-01 -2.21819684e-01 -1.61070094e-01
-8.48886132e-01 -7.42952049e-01 -8.72361898e-01 -1.28384188e-01
7.07354248e-02 -2.11605474e-01 7.94630229e-01 -1.02295566e+00
-1.16826248e+00 1.42790425e+00 -3.58341396e-01 4.47899327e-02
4.80606526e-01 -1.05886608e-01 -6.52632952e-01 1.79845095e-01
2.11756393e-01 8.85824800e-01 1.50569260e+00 -1.20201719e+00
-2.07780734e-01 -9.15777564e-01 4.36292030e-02 4.66166772e-02
-6.07797146e-01 2.32796848e-01 -4.37564641e-01 -7.28998363e-01
-1.84953168e-01 -1.10739923e+00 4.10777509e-01 4.12529051e-01
-3.56714934e-01 -1.74778506e-01 1.18270004e+00 -5.49820065e-01
1.16076374e+00 -2.65732193e+00 -1.29906133e-01 2.86386192e-01
8.47278386e-02 4.51029032e-01 -3.50626767e-01 1.83448285e-01
-4.88486648e-01 3.97360265e-01 -2.14224774e-02 -3.58068764e-01
2.35313565e-01 -1.75140783e-01 -6.20500110e-02 7.28927612e-01
3.68899018e-01 7.05928504e-01 -6.59207284e-01 -3.16583246e-01
-3.93219572e-03 5.09142756e-01 -4.79807943e-01 1.88383341e-01
3.68531644e-01 4.06831443e-01 -2.30887327e-02 7.11902559e-01
1.12216008e+00 7.96949193e-02 -1.31394058e-01 -4.81234670e-01
2.66781658e-01 -9.73381251e-02 -1.14456093e+00 1.38685358e+00
-2.40007907e-01 8.46112132e-01 -1.45508843e-02 -7.08593547e-01
8.53609920e-01 1.28778324e-01 2.18430221e-01 -8.24951053e-01
4.76381816e-02 1.31135345e-01 3.81018519e-02 -3.07356775e-01
2.72568703e-01 8.31353888e-02 6.68639168e-02 5.13594508e-01
-6.97010458e-02 1.04925998e-01 1.06899375e-02 -1.00197077e-01
9.68433499e-01 -2.05645040e-01 1.97030634e-01 -3.19565117e-01
4.64482993e-01 -4.77233738e-01 3.00231308e-01 7.30968237e-01
-4.92663890e-01 9.26263273e-01 4.36324567e-01 -3.34692150e-01
-7.59155810e-01 -1.07005513e+00 -5.31802416e-01 1.04761744e+00
2.62150079e-01 -1.84785411e-01 -9.66734529e-01 -9.58032250e-01
6.14534259e-01 2.09178478e-01 -1.11258054e+00 -3.90163571e-01
-2.98514158e-01 -5.11104405e-01 8.96392405e-01 2.41220742e-01
6.00870132e-01 -8.45553637e-01 -2.56424367e-01 -2.87127137e-01
1.51007921e-01 -1.04699230e+00 -9.69224811e-01 -3.71075809e-01
-2.44201809e-01 -1.18965364e+00 -8.16304386e-01 -7.82809734e-01
7.02282667e-01 4.47117150e-01 9.15731072e-01 -1.23884670e-01
-6.36731982e-01 5.59647858e-01 -3.80643201e-03 -2.31396750e-01
-7.87857994e-02 -4.60509092e-01 5.36885858e-01 5.36173820e-01
7.60614753e-01 -4.44160461e-01 -8.33251536e-01 7.13356674e-01
-9.84060109e-01 -8.45995367e-01 3.30856979e-01 6.15193009e-01
3.96276623e-01 1.36750504e-01 4.08425331e-01 -4.95734304e-01
6.40994191e-01 -2.15945184e-01 -4.02763039e-01 3.21465343e-01
-3.35157752e-01 -8.45154002e-02 4.02281374e-01 -8.01659942e-01
-6.55326128e-01 -3.12428504e-01 2.45102733e-01 -7.43263185e-01
-4.14249808e-01 -1.75815508e-01 -4.12575871e-01 -6.65299058e-01
8.58737826e-01 -1.18168801e-01 4.83203381e-02 -4.28487837e-01
2.66945332e-01 8.39691281e-01 6.81110978e-01 -3.23743284e-01
1.15978253e+00 5.89433312e-01 5.42871580e-02 -8.33525896e-01
-3.84647250e-01 -2.13345394e-01 -6.90208375e-01 -5.58101945e-02
8.32771063e-01 -7.30858624e-01 -8.86667430e-01 7.32234240e-01
-8.63037109e-01 -1.62544940e-02 2.63257846e-02 2.51396477e-01
-3.56528051e-02 4.22046632e-01 -2.57293016e-01 -4.11678135e-01
-1.31018773e-01 -1.23419058e+00 1.36202514e+00 2.99919397e-01
-2.79441208e-01 -6.89882278e-01 -6.22766279e-02 2.42039785e-01
5.32955587e-01 3.76459479e-01 8.44021022e-01 -5.07631660e-01
-2.21274197e-01 -4.48852360e-01 -4.64087814e-01 4.64477152e-01
6.15833223e-01 1.28940001e-01 -1.35193694e+00 -5.48926413e-01
-1.44213751e-01 -2.87895173e-01 6.58946037e-01 2.03365877e-01
1.25853503e+00 -4.60793912e-01 -2.84069002e-01 6.36784852e-01
1.13840675e+00 1.80992603e-01 7.33353913e-01 1.77049980e-01
4.90842015e-01 7.25887239e-01 5.52199662e-01 2.74181575e-01
-2.84427144e-02 9.87067282e-01 3.60417187e-01 -2.11481098e-02
-2.06037760e-01 -1.38317406e-01 3.70049179e-01 -2.29296148e-01
2.93742090e-01 -2.52749562e-01 -6.54640973e-01 4.22377646e-01
-1.28774297e+00 -1.09113765e+00 4.85795379e-01 2.53786325e+00
6.66466653e-01 -1.17405571e-01 -2.38468461e-02 1.03194691e-01
1.02551651e+00 1.91770285e-01 -7.88227379e-01 -3.63428324e-01
-2.03344911e-01 3.66863221e-01 4.56691712e-01 2.54136950e-01
-1.38502502e+00 8.51062715e-01 6.14311409e+00 9.32189822e-01
-1.33755839e+00 -2.25874204e-02 8.76841724e-01 -1.81960151e-01
-6.11691028e-02 -5.07881761e-01 -7.94442177e-01 5.23410976e-01
4.76671934e-01 9.75278839e-02 6.28696144e-01 7.27093279e-01
-2.36162379e-01 -7.13267103e-02 -1.35991704e+00 1.49871743e+00
5.62335908e-01 -1.00507736e+00 -7.34500811e-02 3.55610281e-01
7.24681973e-01 -2.68308938e-01 7.40807533e-01 4.17827582e-03
8.04978311e-02 -1.48080766e+00 6.04870379e-01 2.44641080e-01
1.43118143e+00 -5.69897890e-01 4.53099847e-01 -1.90583155e-01
-1.04807484e+00 -6.34841025e-02 -3.56657922e-01 2.41024777e-01
-4.53107476e-01 3.13334584e-01 -5.66074312e-01 1.96956739e-01
9.69002068e-01 4.96076971e-01 -9.75512207e-01 5.75340211e-01
8.40871558e-02 1.52934462e-01 -3.29308748e-01 4.05924797e-01
1.96084548e-02 -4.87961695e-02 3.93397093e-01 9.42282200e-01
2.51182765e-01 -1.21351488e-01 -1.27357185e-01 6.34694278e-01
-3.89446646e-01 -1.00743264e-01 -8.94937336e-01 5.72792662e-04
6.63967729e-01 1.23522294e+00 -2.75306284e-01 -1.89559057e-01
-3.14168513e-01 1.23312521e+00 2.50627548e-01 5.07795036e-01
-6.68526471e-01 -4.26355720e-01 1.35454416e+00 1.84456304e-01
4.09434289e-01 5.50964810e-02 1.84561849e-01 -8.68584752e-01
2.73389935e-01 -1.14771903e+00 2.25787550e-01 -6.33549869e-01
-1.52599430e+00 7.35499918e-01 4.39119078e-02 -1.14793038e+00
-2.14813337e-01 -8.31291735e-01 -5.38988292e-01 9.42526579e-01
-1.38090050e+00 -7.84094214e-01 -4.30168599e-01 7.87442863e-01
-9.86853465e-02 -3.02447826e-01 7.47900128e-01 6.38581038e-01
-5.55008054e-01 1.20284295e+00 1.45374805e-01 5.79870224e-01
1.14862168e+00 -9.56336677e-01 4.51253176e-01 6.34685040e-01
2.12970808e-01 7.70344198e-01 3.83432299e-01 -3.04944873e-01
-1.32450211e+00 -1.12223864e+00 6.71833277e-01 -5.48724890e-01
3.29280913e-01 -4.91535068e-01 -8.02457452e-01 4.12306100e-01
-8.35576206e-02 5.73483229e-01 6.77968323e-01 -1.77579701e-01
-8.84031117e-01 -3.86456013e-01 -1.69012797e+00 5.89170456e-01
1.23824024e+00 -1.03256011e+00 -2.51609981e-01 1.76637262e-01
1.78431213e-01 -1.34668693e-01 -7.68189788e-01 4.28198487e-01
8.37705493e-01 -1.11302698e+00 1.26088071e+00 -6.91015899e-01
1.02521077e-01 -3.72067928e-01 -4.25784558e-01 -1.15994573e+00
-3.87136400e-01 -5.45451164e-01 6.15768619e-02 1.22952223e+00
8.80261734e-02 -7.76233375e-01 6.71803296e-01 5.31407475e-01
4.34058785e-01 -5.25052726e-01 -1.03116882e+00 -1.01447940e+00
-8.26493055e-02 -8.65879878e-02 8.52087557e-01 9.26657975e-01
-2.63732284e-01 -9.92528126e-02 -2.35876933e-01 3.33663493e-01
6.25488937e-01 5.11454185e-03 9.95682776e-01 -1.20226145e+00
-1.72506701e-02 -5.19851685e-01 -1.04701793e+00 -7.82661319e-01
2.82658905e-01 -7.84177482e-01 -2.61521667e-01 -8.99740875e-01
3.88152711e-02 -3.35556418e-01 -1.72866791e-01 4.50455576e-01
-2.70395786e-01 1.06891680e+00 2.57588565e-01 2.29585454e-01
-2.62332559e-01 3.16153675e-01 1.14666736e+00 -3.65274698e-01
2.43576154e-01 -3.13489854e-01 -7.98944831e-01 5.93181729e-01
6.28592432e-01 -4.40714896e-01 -1.67811334e-01 -4.40825701e-01
-2.34167159e-01 -3.99905682e-01 6.20655000e-01 -1.01102424e+00
3.18667218e-02 -1.11730732e-02 7.37534106e-01 -1.00899294e-01
4.68892038e-01 -8.63303602e-01 2.33928531e-01 2.15721428e-01
-1.72594249e-01 3.13347727e-01 4.07113731e-01 4.48619217e-01
-2.62573719e-01 -1.27348781e-01 1.07317722e+00 1.27384216e-01
-4.83239740e-01 6.66766524e-01 9.94977131e-02 2.72344351e-01
1.05790460e+00 -6.73682213e-01 -4.24505979e-01 -4.21147436e-01
-3.73576343e-01 -2.70274431e-01 7.12006271e-01 6.76666915e-01
5.75504422e-01 -1.53872955e+00 -7.14224100e-01 6.31946802e-01
3.30537856e-01 -4.68803912e-01 1.86753482e-01 5.99624157e-01
-2.69436568e-01 3.00691366e-01 -1.81653410e-01 -5.73778450e-01
-1.61405301e+00 6.33155107e-01 4.59063798e-01 7.54868984e-02
-6.75797388e-02 1.02391064e+00 6.90264761e-01 -3.64360988e-01
2.74196625e-01 1.44218817e-01 -8.52896646e-02 2.75703311e-01
9.38509643e-01 1.41041830e-01 2.67630488e-01 -1.09699726e+00
-7.45191753e-01 1.03227854e+00 -2.01163158e-01 -8.34365143e-04
7.46019661e-01 5.78991063e-02 1.64710563e-02 -3.72391604e-02
1.80379403e+00 1.81425989e-01 -1.18758559e+00 -2.74829358e-01
-1.47345319e-01 -1.09115791e+00 -1.70633867e-01 -5.35901546e-01
-1.25911403e+00 9.72878754e-01 1.15429115e+00 1.55408695e-01
1.00000548e+00 -7.24242926e-02 4.08980012e-01 2.26234287e-01
2.77061939e-01 -6.83710456e-01 2.87283003e-01 1.11623093e-01
1.02843785e+00 -1.48960662e+00 -3.06655802e-02 -1.08822666e-01
-4.87280369e-01 8.91639352e-01 6.25482023e-01 -7.24832900e-03
8.51643324e-01 4.24205251e-02 1.69618517e-01 -1.90525487e-01
-1.79907992e-01 -4.88162227e-02 6.16791666e-01 9.18747604e-01
2.00469285e-01 -1.91626906e-01 2.85151392e-01 -5.55608384e-02
-1.97013602e-01 -3.97115290e-01 4.93854061e-02 7.99873710e-01
-1.43137291e-01 -8.81139874e-01 -6.33237839e-01 4.64376390e-01
-5.46002150e-01 -6.14442453e-02 -6.41757667e-01 7.94790924e-01
3.21217090e-01 1.01179552e+00 2.95153379e-01 -4.94466484e-01
3.44543993e-01 -2.26414278e-02 6.12570345e-01 -4.46610004e-01
-4.05633897e-01 -4.28956717e-01 -3.48086178e-01 -5.81744373e-01
-2.59092391e-01 -6.53415501e-01 -5.49954236e-01 -5.10821104e-01
-2.48480901e-01 -1.30223289e-01 5.75977087e-01 6.30818665e-01
7.01052010e-01 -7.35188052e-02 9.07800317e-01 -9.82567728e-01
-5.08215666e-01 -8.42953622e-01 -6.82939649e-01 9.70719039e-01
5.82963645e-01 -9.04194474e-01 -6.49389803e-01 -6.45633712e-02] | [13.11854362487793, 0.9180476069450378] |
de41bfe3-8261-4661-a755-a71d2e034c41 | imposition-implicit-backdoor-attack-through | 2306.15755 | null | https://arxiv.org/abs/2306.15755v1 | https://arxiv.org/pdf/2306.15755v1.pdf | IMPOSITION: Implicit Backdoor Attack through Scenario Injection | This paper presents a novel backdoor attack called IMPlicit BackdOor Attack through Scenario InjecTION (IMPOSITION) that does not require direct poisoning of the training data. Instead, the attack leverages a realistic scenario from the training data as a trigger to manipulate the model's output during inference. This type of attack is particularly dangerous as it is stealthy and difficult to detect. The paper focuses on the application of this attack in the context of Autonomous Driving (AD) systems, specifically targeting the trajectory prediction module. To implement the attack, we design a trigger mechanism that mimics a set of cloned behaviors in the driving scene, resulting in a scenario that triggers the attack. The experimental results demonstrate that IMPOSITION is effective in attacking trajectory prediction models while maintaining high performance in untargeted scenarios. Our proposed method highlights the growing importance of research on the trustworthiness of Deep Neural Network (DNN) models, particularly in safety-critical applications. Backdoor attacks pose a significant threat to the safety and reliability of DNN models, and this paper presents a new perspective on backdooring DNNs. The proposed IMPOSITION paradigm and the demonstration of its severity in the context of AD systems are significant contributions of this paper. We highlight the impact of the proposed attacks via empirical studies showing how IMPOSITION can easily compromise the safety of AD systems. | ['Amir Rasouli', 'Mohammad Sabokrou', 'Mozhgan PourKeshavarz'] | 2023-06-27 | null | null | null | null | ['backdoor-attack', 'trajectory-prediction'] | ['adversarial', 'computer-vision'] | [ 5.28732724e-02 4.10780668e-01 -1.52317852e-01 -1.36794463e-01
-2.82622606e-01 -1.08751619e+00 8.69550645e-01 -4.01788741e-01
-2.62891591e-01 4.11581844e-01 -4.54199970e-01 -8.86862934e-01
1.64553463e-01 -8.92561078e-01 -1.22459459e+00 -7.91603744e-01
-2.75552720e-01 -2.35619262e-01 4.31722909e-01 -2.40685865e-01
1.99066907e-01 9.23300803e-01 -1.33966422e+00 -1.68365179e-04
3.15384239e-01 8.54904473e-01 -4.75180835e-01 5.95829964e-01
6.08821273e-01 6.98774874e-01 -1.04416704e+00 -6.22735143e-01
6.80742919e-01 1.36456862e-02 -2.53062636e-01 -6.46257460e-01
5.04559934e-01 -9.84101117e-01 -8.47807765e-01 1.07434607e+00
2.30351254e-01 -1.69776767e-01 2.31289327e-01 -2.23930240e+00
3.67791246e-04 5.51554143e-01 -9.32392702e-02 2.06230551e-01
-2.18745470e-01 6.55455649e-01 3.64454716e-01 -3.28745157e-01
8.03109333e-02 1.07523322e+00 5.40994048e-01 8.33509684e-01
-8.77908409e-01 -1.42325127e+00 9.06526893e-02 5.06966896e-02
-1.34852707e+00 -5.75267673e-01 7.31790423e-01 -3.50374281e-01
7.84200907e-01 4.40291911e-01 3.98481250e-01 1.85913110e+00
8.47808480e-01 5.02818286e-01 8.20900500e-01 -4.62984405e-02
5.24231136e-01 3.26900184e-01 4.09372598e-01 4.31515217e-01
6.86416447e-01 1.28817892e+00 -3.98147047e-01 -6.83852732e-01
3.37519497e-01 -1.43399432e-01 4.70411964e-02 -9.47301760e-02
-8.39799762e-01 8.08516860e-01 2.34190419e-01 -3.72199744e-01
-1.34485915e-01 7.41792381e-01 5.19314885e-01 7.43389353e-02
-1.10668398e-01 2.34355763e-01 -2.30796918e-01 -1.31086871e-01
-6.70234859e-01 5.90951562e-01 7.66555071e-01 8.44022393e-01
2.31232852e-01 4.62181389e-01 -1.83761403e-01 -3.48032683e-01
5.13205826e-01 7.99013078e-01 -5.48738800e-02 -7.28510201e-01
4.34045613e-01 1.44111529e-01 1.43873423e-01 -1.07473993e+00
-1.97939530e-01 -2.71268994e-01 -9.03400034e-02 7.14213192e-01
1.61691234e-01 -5.85674763e-01 -7.53343821e-01 1.96920240e+00
5.47119081e-01 8.73265803e-01 2.65391856e-01 6.70080602e-01
5.01512408e-01 5.82307279e-01 2.30370924e-01 2.83928216e-01
1.06600082e+00 -4.93965745e-01 -6.62689686e-01 2.25542653e-02
5.28683364e-01 -1.86344668e-01 6.91660821e-01 2.93421537e-01
-5.61843157e-01 -2.47975916e-01 -1.56816506e+00 5.54412842e-01
-5.93677044e-01 -3.23692381e-01 3.28829616e-01 1.28503299e+00
-5.97756326e-01 1.94960788e-01 -9.49000239e-01 -6.66379705e-02
4.24325764e-01 5.55593371e-01 -1.74942151e-01 2.66415536e-01
-1.61174583e+00 1.04025483e+00 3.49989176e-01 4.75770161e-02
-1.97500098e+00 -8.62778127e-01 -7.07985461e-01 -2.39204556e-01
2.59887636e-01 -2.23901451e-01 1.20317554e+00 -3.16806674e-01
-1.43006754e+00 4.50942039e-01 3.60161602e-01 -1.18875515e+00
7.34901071e-01 -3.97404224e-01 -9.29365098e-01 1.38656959e-01
-2.19739482e-01 5.79501867e-01 1.27356637e+00 -1.32737732e+00
-3.68057072e-01 2.96791103e-02 4.97960985e-01 -5.01866698e-01
-3.32981706e-01 4.44169939e-02 4.54696864e-01 -5.80875993e-01
-6.76493108e-01 -1.25437307e+00 -1.05971411e-01 -1.19444482e-01
-8.01323116e-01 2.90148258e-01 1.66506898e+00 -1.10547513e-01
1.13817859e+00 -2.42881846e+00 -7.38440275e-01 2.66187578e-01
3.68894845e-01 7.91384816e-01 7.30786622e-02 6.16016030e-01
2.42235214e-02 1.00082539e-01 -2.52771884e-01 1.25661502e-02
1.90814048e-01 4.22135472e-01 -1.36905968e+00 6.77104175e-01
1.04177631e-01 9.24917519e-01 -6.22846067e-01 6.82449639e-02
2.56316274e-01 3.75596225e-01 -4.77052659e-01 2.91616559e-01
-1.83936521e-01 2.49458402e-01 -4.87976462e-01 4.40664589e-01
7.87675440e-01 6.60000265e-01 -1.98015600e-01 -9.07519907e-02
-3.67566273e-02 2.62332320e-01 -6.20416999e-01 5.83408833e-01
-7.55173266e-02 6.71037614e-01 -2.48114944e-01 -6.05646431e-01
8.92962277e-01 4.86256301e-01 -5.29114045e-02 -2.02627540e-01
5.37036479e-01 -1.67575538e-01 2.50407338e-01 -4.11110789e-01
3.85782748e-01 -1.07900485e-01 -4.39672440e-01 6.56534612e-01
-3.07400256e-01 1.20648764e-01 -5.71254671e-01 2.99022406e-01
1.35914946e+00 -1.41125143e-01 3.18163563e-03 -2.05277592e-01
3.58038515e-01 1.21600538e-01 4.25608456e-01 9.69821990e-01
-6.88183129e-01 -1.18838333e-01 5.70004463e-01 -3.72835636e-01
-6.65953338e-01 -1.38079357e+00 -1.91697448e-01 3.89749467e-01
2.96814293e-01 -3.61752480e-01 -1.12097549e+00 -1.26045108e+00
4.04415816e-01 1.36938310e+00 -7.94265926e-01 -1.12242365e+00
-3.69100004e-01 -3.97021174e-01 1.92861903e+00 5.89776337e-01
7.66571641e-01 -5.48454463e-01 -1.07043576e+00 -2.59382963e-01
1.76461801e-01 -1.50572670e+00 -1.53744027e-01 -6.12187162e-02
-6.67140484e-01 -1.13440514e+00 2.52225786e-01 -9.77317393e-02
5.35616934e-01 4.40482259e-01 4.70909148e-01 -1.13181947e-02
7.39662722e-03 3.91573250e-01 1.61899123e-02 -8.83357406e-01
-9.54502285e-01 -3.67695391e-01 6.82402730e-01 1.05606675e-01
6.90511942e-01 -5.17206907e-01 -2.37980664e-01 6.07792258e-01
-1.05167055e+00 -5.94750166e-01 2.38130078e-01 3.52334619e-01
-2.40505673e-02 2.80002296e-01 7.16543674e-01 -6.76688552e-01
6.36878550e-01 -7.98246503e-01 -1.07814765e+00 -7.74108022e-02
-4.98934954e-01 -2.37745732e-01 8.94908011e-01 -6.68778300e-01
-6.26031518e-01 -1.05545722e-01 -7.71559849e-02 -8.53858829e-01
-4.04507667e-01 4.00802940e-02 -5.07104218e-01 -5.09613097e-01
7.94741273e-01 6.68189153e-02 1.56059280e-01 -3.80486213e-02
3.95159304e-01 5.10779381e-01 5.99045217e-01 -5.28446734e-01
1.35705602e+00 8.31066608e-01 2.42977217e-01 -8.02901268e-01
-2.59242922e-01 4.30607498e-01 -2.42563248e-01 -4.99584258e-01
6.18568659e-01 -7.80082643e-01 -1.00357389e+00 6.16293669e-01
-1.21680975e+00 -3.82110327e-01 -5.57637727e-03 3.69948655e-01
-2.75146633e-01 2.19971016e-01 -4.22968179e-01 -8.00217450e-01
-8.15805048e-02 -1.32140064e+00 6.03686690e-01 -1.55354455e-01
-2.60834485e-01 -8.03053677e-01 -6.67469800e-02 2.19024971e-01
2.64014155e-01 6.64478600e-01 7.69523263e-01 -9.57840800e-01
-8.42152357e-01 -7.16187716e-01 2.63742328e-01 5.12257516e-01
-1.87444612e-01 2.76585758e-01 -1.52179968e+00 -2.23058000e-01
2.56816804e-01 -5.03336377e-02 3.61219257e-01 9.10754688e-03
9.41049993e-01 -6.51661158e-01 -4.80656266e-01 5.15955448e-01
1.02559173e+00 3.78980428e-01 6.56523526e-01 2.32030615e-01
7.05316722e-01 5.50243616e-01 7.01604247e-01 2.89021611e-01
-6.30255369e-03 6.30327106e-01 1.06504285e+00 2.30025962e-01
2.69352525e-01 -6.44203246e-01 1.02499175e+00 -1.12179697e-01
5.10634065e-01 -2.69697905e-01 -8.28544974e-01 1.89085320e-01
-1.61142111e+00 -9.44006801e-01 -2.00994805e-01 2.33259821e+00
1.94580078e-01 5.11256278e-01 2.27760315e-01 2.43927598e-01
6.07439399e-01 1.03872612e-01 -6.35450482e-01 -7.52711058e-01
3.25062633e-01 -1.14733659e-01 9.52051222e-01 5.16579688e-01
-1.16526818e+00 9.51521099e-01 7.04697752e+00 6.12172246e-01
-1.27028799e+00 2.02926382e-01 2.06540272e-01 -2.20424026e-01
-1.74273044e-01 6.13459349e-02 -1.00935769e+00 6.72408104e-01
1.52601230e+00 -3.15807253e-01 3.16954181e-02 9.80839968e-01
5.08148372e-01 1.85500249e-01 -1.36304402e+00 3.89703333e-01
-9.50472653e-02 -1.31180048e+00 1.44411013e-01 4.28860188e-01
1.82867005e-01 8.74621347e-02 5.34644127e-01 2.78302103e-01
5.41193485e-01 -1.14518547e+00 1.06652772e+00 6.42279536e-03
6.43320441e-01 -1.35137987e+00 4.53821272e-01 5.90434015e-01
-7.02511966e-01 -2.26840988e-01 -1.65153787e-01 -1.63306221e-01
1.81587234e-01 2.78451920e-01 -9.96895611e-01 1.45265326e-01
4.92187053e-01 2.18171328e-01 -2.61991709e-01 4.52641279e-01
-4.54776019e-01 1.20530236e+00 -4.13141042e-01 2.04389378e-01
5.16643882e-01 3.24667841e-01 1.11541498e+00 1.09043181e+00
-5.38568869e-02 2.42886208e-02 -1.78627893e-01 1.14986289e+00
1.85377032e-01 -8.31068516e-01 -1.57113886e+00 1.44322300e-02
7.73465157e-01 9.99428093e-01 -3.57599199e-01 -5.24301343e-02
-5.23668565e-02 3.98142487e-01 -8.67195129e-02 2.94945240e-01
-1.58715892e+00 -2.14686513e-01 1.23452640e+00 1.83948740e-01
8.16426873e-02 -3.91859323e-01 -3.58259171e-01 -5.85415184e-01
-9.55809467e-03 -8.75461280e-01 1.36915818e-01 -3.15858513e-01
-8.96044612e-01 5.53914607e-01 4.50797915e-01 -1.41882157e+00
-1.61974922e-01 -4.24372405e-01 -8.72230589e-01 4.76776451e-01
-1.20433724e+00 -1.14648187e+00 -1.22905970e-02 6.67710125e-01
-5.58611704e-03 -4.90491331e-01 6.73082411e-01 2.35739619e-01
-9.14692521e-01 1.23125446e+00 -2.92069554e-01 1.17009111e-01
3.34691733e-01 -4.58310992e-01 8.91815960e-01 1.42468131e+00
-1.28470570e-01 8.27101052e-01 9.44128692e-01 -7.99692512e-01
-1.48322940e+00 -1.59453762e+00 2.57618874e-01 -9.37955856e-01
6.70798838e-01 -7.65629113e-01 -6.19574428e-01 9.84113932e-01
-1.36985248e-02 9.37430635e-02 7.42024422e-01 -6.82433963e-01
-5.64527810e-01 -7.98770413e-02 -1.62726259e+00 8.80564928e-01
6.07220590e-01 -6.98386431e-01 -5.06724298e-01 5.26686385e-02
1.03546715e+00 -2.71661580e-01 -3.21082413e-01 3.91859502e-01
6.68417990e-01 -8.29677045e-01 9.75186467e-01 -8.40863109e-01
1.77718669e-01 -5.06033301e-01 -4.30066884e-01 -9.14266527e-01
1.53810829e-01 -8.47853541e-01 -6.82005942e-01 1.08230603e+00
2.88436085e-01 -1.21377194e+00 6.14646554e-01 7.85730898e-01
-7.67956972e-02 -4.71208304e-01 -1.15862691e+00 -1.22291768e+00
3.02337199e-01 -8.96297872e-01 9.38705027e-01 8.32614601e-01
-8.14421400e-02 -2.33923301e-01 -3.93632829e-01 1.09260070e+00
8.72203529e-01 -6.45621121e-01 1.07707071e+00 -5.65803051e-01
-1.86149657e-01 6.78018481e-02 -7.74416983e-01 -5.81756294e-01
3.84046495e-01 -7.41840005e-01 -5.28167514e-03 -3.60011488e-01
-4.30532962e-01 -3.29437703e-01 -3.15999091e-01 4.74592596e-01
2.89563268e-01 2.04399332e-01 4.00103837e-01 1.67512372e-01
3.18356417e-02 4.73136842e-01 4.67774004e-01 -1.94029346e-01
2.34721497e-01 3.17955256e-01 -5.66103995e-01 7.43988097e-01
8.84016514e-01 -9.40176368e-01 -8.69215310e-01 -4.73030396e-02
-2.41597906e-01 -1.85451776e-01 1.07858109e+00 -1.13242090e+00
1.50881931e-01 -1.64574862e-01 -1.91431805e-01 -5.04898012e-01
3.39373857e-01 -1.31203818e+00 1.58831447e-01 8.22240889e-01
-1.78990364e-01 1.69700131e-01 7.52689660e-01 8.05787921e-01
2.03742951e-01 -8.89318530e-03 6.98955119e-01 3.73885155e-01
-6.31591678e-01 1.67640343e-01 -8.58658433e-01 -2.92236239e-01
1.83845913e+00 -2.31337905e-01 -5.77245474e-01 -2.00355992e-01
-2.60758340e-01 1.33885473e-01 3.08199078e-01 6.17251933e-01
8.43701661e-01 -1.13750601e+00 -3.35596204e-01 4.84457016e-01
1.73380777e-01 -4.44426596e-01 9.40543134e-04 5.50151825e-01
-3.65280151e-01 4.35869932e-01 -1.12932131e-01 -5.07263720e-01
-1.23987842e+00 9.45750654e-01 6.63258374e-01 3.38230371e-01
-7.97565877e-01 5.37711918e-01 5.37094951e-01 -4.26743984e-01
3.73723298e-01 -1.49035200e-01 1.88728884e-01 -5.50965488e-01
6.50099516e-01 4.12843734e-01 1.37868807e-01 -5.43349683e-01
-6.43883705e-01 -1.01929389e-01 -1.63937673e-01 -2.14135543e-01
5.24744451e-01 2.23702788e-01 2.57714301e-01 2.01866165e-01
9.54556286e-01 -1.06136546e-01 -1.47125387e+00 3.14460903e-01
-3.05649161e-01 -5.14302433e-01 1.18688650e-01 -8.88385415e-01
-9.42083538e-01 8.06650102e-01 5.58215857e-01 1.13856152e-01
5.87095022e-01 -3.87903094e-01 1.20899367e+00 5.10838985e-01
5.86205482e-01 -6.31927788e-01 -1.56420231e-01 2.88085252e-01
6.21909440e-01 -7.43571281e-01 -4.05781627e-01 -2.28417292e-01
-4.97583270e-01 8.57754469e-01 7.83458948e-01 -4.31160718e-01
1.06980860e+00 7.56646395e-01 3.47279996e-01 -2.56292969e-01
-7.52497435e-01 6.37659490e-01 -2.54748255e-01 8.23545754e-01
-4.83426124e-01 2.94434488e-01 1.54178888e-01 7.65778601e-01
-3.20278347e-01 -6.51462525e-02 9.26741421e-01 1.08377957e+00
-2.21260190e-01 -8.65041673e-01 -6.96238279e-01 1.12423748e-01
-4.01673287e-01 1.41258880e-01 -4.92811739e-01 9.11020577e-01
1.23322383e-01 1.24744976e+00 -1.57513365e-01 -1.03898001e+00
2.20354617e-01 -2.37117931e-02 2.45250333e-02 -1.95408240e-01
-6.73570275e-01 -7.29188919e-01 1.92989245e-01 -8.96185219e-01
2.39747345e-01 -5.13384402e-01 -1.23830104e+00 -7.91997910e-01
-2.10180819e-01 -2.99042165e-02 7.22133219e-01 1.03998232e+00
5.02056360e-01 5.19913971e-01 9.32041168e-01 -5.19062102e-01
-1.01446557e+00 -2.79750019e-01 -4.62454140e-01 -1.08733289e-01
6.81860983e-01 -1.09682381e+00 -6.56289995e-01 -3.67905617e-01] | [5.541797161102295, 7.640066623687744] |
7fe8d794-b370-4d34-9a9b-6701de8a5a7f | a-multi-stage-model-based-on-yolov3-for | 2111.11709 | null | https://arxiv.org/abs/2111.11709v2 | https://arxiv.org/pdf/2111.11709v2.pdf | A Multi-Stage model based on YOLOv3 for defect detection in PV panels based on IR and Visible Imaging by Unmanned Aerial Vehicle | As solar capacity installed worldwide continues to grow, there is an increasing awareness that advanced inspection systems are becoming of utmost importance to schedule smart interventions and minimize downtime likelihood. In this work we propose a novel automatic multi-stage model to detect panel defects on aerial images captured by unmanned aerial vehicle by using the YOLOv3 network and Computer Vision techniques. The model combines detections of panels and defects to refine its accuracy and exhibits an average inference time per image of 0.98 s. The main novelties are represented by its versatility to process either thermographic or visible images and detect a large variety of defects, to prescript recommended actions to O&M crew to give a more efficient data-driven maintenance strategy and its portability to both rooftop and ground-mounted PV systems and different panel types. The proposed model has been validated on two big PV plants in the south of Italy with an outstanding AP@0.5 exceeding 98% for panel detection, a remarkable AP@0.4 (AP@0.5) of roughly 88.3% (66.9%) for hotspots by means of infrared thermography and a mAP@0.5 of almost 70% in the visible spectrum for detection of anomalies including panel shading induced by soiling and bird dropping, delamination, presence of puddles and raised rooftop panels. The model predicts also the severity of hotspot areas based on the estimated temperature gradients, as well as it computes the soiling coverage based on visual images. Finally an analysis of the influence of the different YOLOv3's output scales on the detection is discussed. | ['Benedetto Michelozzi', 'Giacomo Fontanelli', 'Alessandro Betti', 'Antonio Di Tommaso'] | 2021-11-23 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 2.38860279e-01 5.64853251e-02 5.39783120e-01 2.61412472e-01
-1.36442766e-01 -1.09098864e+00 4.07679468e-01 4.42227662e-01
3.74018312e-01 6.02136314e-01 -5.66276550e-01 -4.22096550e-01
-4.30190772e-01 -9.71374393e-01 -4.08604771e-01 -8.78627777e-01
-1.57689929e-01 1.30250677e-01 4.96831924e-01 -3.03821057e-01
3.45443166e-03 8.53093922e-01 -1.91261756e+00 5.99930994e-02
1.40151334e+00 1.07185316e+00 4.93794173e-01 7.86914289e-01
7.44277537e-01 3.29309344e-01 -8.98568451e-01 2.59144604e-01
3.43019426e-01 -1.64953440e-01 -2.10815087e-01 3.80859435e-01
4.83287603e-01 -3.57698500e-01 5.38936377e-01 8.81069005e-01
1.91810757e-01 -2.76730180e-01 9.20032620e-01 -1.13271379e+00
-6.09731115e-02 -1.71743497e-01 -5.47383606e-01 3.09264064e-02
4.34529632e-01 2.96579987e-01 7.07003713e-01 -3.05491865e-01
2.61200488e-01 3.85847539e-01 6.70206249e-01 -3.35434467e-01
-1.15443254e+00 1.11896344e-01 -3.44617367e-01 2.48584643e-01
-1.15435278e+00 2.57188737e-01 5.17217517e-01 -8.73820603e-01
8.97305131e-01 5.53806484e-01 8.08368921e-01 3.20827723e-01
2.96053588e-01 -1.23448819e-01 1.37553144e+00 -5.63353181e-01
3.50607306e-01 5.27077794e-01 -2.24763542e-01 7.95926452e-01
5.39359033e-01 5.71613424e-02 8.85577947e-02 -9.69169289e-02
5.45655489e-01 -3.28469962e-01 -5.79443216e-01 -6.27331138e-02
-5.26211381e-01 4.85397100e-01 4.62826908e-01 4.18260187e-01
-5.04318416e-01 -3.58116269e-01 8.90627876e-02 4.97761704e-02
3.23994786e-01 2.82042980e-01 -5.12299299e-01 4.65910137e-01
-8.26911330e-01 -2.18889281e-01 5.67345917e-01 4.80969191e-01
7.23964751e-01 2.84529775e-01 1.31768897e-01 8.79163444e-01
3.78627151e-01 1.10415876e+00 -8.40885341e-02 -7.81008899e-01
1.78106595e-02 9.05285120e-01 4.01730478e-01 -1.07957816e+00
-4.52591002e-01 -3.36985707e-01 -7.55231202e-01 8.34599435e-01
1.31231889e-01 -4.69159216e-01 -6.99925542e-01 9.72058535e-01
4.49486464e-01 -5.18849075e-01 -2.76996434e-01 6.87921524e-01
2.43713975e-01 9.36527491e-01 -3.81016254e-01 -2.85466343e-01
1.45689511e+00 -5.17304838e-01 -3.09916854e-01 -1.15423188e-01
4.64380056e-01 -6.61779344e-01 6.25932753e-01 8.30587566e-01
-7.40643919e-01 -6.39839292e-01 -1.20239425e+00 7.45430529e-01
-4.38091397e-01 9.91963863e-01 -2.72551458e-02 9.73415494e-01
-1.31746590e+00 6.75194621e-01 -5.41751087e-01 -7.45305955e-01
4.91590425e-03 1.32634535e-01 -1.06557459e-01 2.79451370e-01
-5.82927227e-01 1.18031406e+00 1.22353747e-01 5.39376497e-01
-7.71874428e-01 -5.64315557e-01 -6.09793246e-01 1.97006822e-01
1.97226942e-01 -3.51667106e-01 6.89598501e-01 -1.01871169e+00
-1.37931490e+00 6.49986982e-01 1.96021482e-01 -4.93746817e-01
4.81422901e-01 -1.26718283e-01 -6.65485144e-01 6.07339740e-01
-7.72349834e-02 -7.36100320e-03 7.58126020e-01 -1.35950506e+00
-7.92117476e-01 -3.26662600e-01 -4.85587269e-02 -1.62230730e-02
-3.07070762e-01 -3.14031988e-01 4.55546468e-01 -8.45084637e-02
-1.29058242e-01 -8.55218589e-01 6.39967918e-02 -3.06387432e-04
-4.88946527e-01 2.15459727e-02 1.00603747e+00 -1.07945609e+00
6.77390993e-01 -1.72415853e+00 -7.66255260e-02 5.75129449e-01
-2.92488247e-01 5.19902885e-01 2.69014865e-01 8.55778277e-01
4.00587730e-03 -1.91353858e-01 -7.11686552e-01 3.11382592e-01
-1.43905252e-01 2.72799400e-03 -1.71142027e-01 5.74573994e-01
2.16178805e-01 -6.23836815e-02 -5.38732350e-01 -5.49304113e-02
9.27156866e-01 4.45471406e-01 6.84293434e-02 2.60490537e-01
-2.61846501e-02 -2.23709666e-03 -1.23770997e-01 8.58231306e-01
9.04236376e-01 2.34318331e-01 3.89606148e-01 -1.62227914e-01
-8.31072688e-01 -3.46206963e-01 -1.14400363e+00 7.13909268e-01
-4.86085057e-01 7.08668828e-01 4.51844007e-01 -9.28239882e-01
1.35838604e+00 5.41482568e-01 3.75401050e-01 -3.18915665e-01
2.66614966e-02 2.63893545e-01 -3.74863982e-01 -6.54411137e-01
3.94213170e-01 9.22677964e-02 4.39275920e-01 -2.15317726e-01
-5.93285076e-04 -4.59011465e-01 1.88670710e-01 -6.95054755e-02
9.39775944e-01 1.31446853e-01 5.58690429e-02 -7.91853428e-01
9.08376038e-01 9.99478996e-02 3.92704368e-01 2.30768725e-01
7.34995902e-02 4.13318008e-01 5.60061455e-01 -7.56350607e-02
-8.32333267e-01 -9.09575939e-01 -2.82436609e-01 3.63247693e-01
1.09166667e-01 1.66639492e-01 -8.94468188e-01 -2.43227765e-01
3.84974070e-02 8.21972311e-01 -3.85133982e-01 -6.13351306e-03
5.94013184e-02 -9.95250583e-01 -1.09964147e-01 1.60887420e-01
6.82051659e-01 -1.07403481e+00 -1.12685835e+00 -1.09004907e-01
3.93160917e-02 -9.19684052e-01 5.32551050e-01 1.05293691e-01
-1.11142194e+00 -1.40030682e+00 -6.16680920e-01 -5.05090415e-01
7.97192514e-01 2.58504122e-01 1.09238470e+00 1.43049344e-01
-7.24574983e-01 6.88647151e-01 -5.66483736e-01 -3.52359891e-01
-6.36601567e-01 -3.28961074e-01 -2.52526462e-01 -5.70785953e-03
-2.72889704e-01 -4.34764922e-01 -6.60834014e-01 6.59819543e-01
-7.87850916e-01 -3.16061407e-01 2.62260497e-01 4.62393582e-01
1.80283368e-01 4.37978178e-01 3.60174626e-01 -6.59153640e-01
1.77460507e-01 -3.57671767e-01 -1.41511738e+00 4.35878992e-01
-8.10934067e-01 -7.05869913e-01 6.77601814e-01 2.79822350e-01
-1.19183397e+00 3.67057212e-02 6.36718273e-02 1.67832851e-01
-7.85297036e-01 1.01914994e-01 -4.44400646e-02 -2.12751955e-01
7.06251502e-01 -1.28529370e-01 -2.77459651e-01 -3.53882372e-01
-8.93053561e-02 3.74673665e-01 4.25566524e-01 -3.80764492e-02
1.25676990e+00 5.51188469e-01 2.19809264e-01 -1.55693305e+00
-2.29377121e-01 -3.42397958e-01 -5.38527489e-01 -8.78695786e-01
8.96940768e-01 -6.85711920e-01 -8.19209993e-01 6.19215012e-01
-1.02316463e+00 -3.93883526e-01 -1.29686520e-01 1.57649264e-01
-8.61510932e-02 7.09173262e-01 -2.94734180e-01 -1.20781255e+00
-5.69477141e-01 -7.77864218e-01 8.13267648e-01 5.90199471e-01
1.81818709e-01 -1.10129440e+00 1.27567962e-01 4.11496699e-01
2.18162075e-01 7.46934056e-01 7.50396609e-01 3.50919425e-01
-4.43216622e-01 -1.95590526e-01 -5.20625189e-02 8.13263416e-01
2.29287773e-01 7.34108627e-01 -1.38098109e+00 -2.92708993e-01
5.45162447e-02 1.04138240e-01 6.06760442e-01 5.89367270e-01
6.64793074e-01 1.04298227e-01 -2.60653704e-01 2.08277792e-01
2.09267068e+00 4.24871534e-01 6.54733181e-01 2.93615907e-01
2.05293283e-01 8.27497661e-01 8.07146966e-01 6.34558022e-01
-2.22235411e-01 5.31248450e-01 1.25762379e+00 -4.06453699e-01
2.60149688e-01 2.41528124e-01 5.44972539e-01 2.71834999e-01
-3.75814468e-01 -5.41540921e-01 -7.21710861e-01 7.47105837e-01
-1.33056772e+00 -8.90688419e-01 -9.76275802e-01 2.46325040e+00
-7.43503943e-02 8.88250694e-02 8.74993578e-02 5.40450096e-01
9.65406239e-01 4.08532098e-02 -7.34385103e-02 -6.24650896e-01
-2.84379750e-01 3.82925607e-02 7.42376268e-01 4.93870229e-01
-6.90167010e-01 1.70158595e-01 5.48222828e+00 2.52359092e-01
-8.67146850e-01 -1.86083138e-01 3.94177854e-01 4.35659617e-01
-9.71755907e-02 1.81956850e-02 -3.44769597e-01 5.35014451e-01
6.40560448e-01 3.99443001e-01 2.24277228e-01 6.50794089e-01
3.76032740e-01 -1.05007768e+00 -4.97454107e-01 5.43462634e-01
1.10650711e-01 -8.01608443e-01 -4.25079435e-01 7.68544599e-02
6.74930990e-01 8.03670436e-02 -2.59353220e-01 -3.83143783e-01
1.22877751e-02 -4.23224926e-01 4.08925474e-01 5.40217221e-01
6.58681095e-01 -5.19786239e-01 8.41486692e-01 2.41236120e-01
-1.21001422e+00 -3.70896965e-01 -4.11507696e-01 -1.36533484e-01
2.27389067e-01 9.98961210e-01 -1.04585838e+00 1.08469605e+00
9.00646448e-01 3.36088896e-01 -6.26824498e-01 1.20752048e+00
-6.27365232e-01 8.53138447e-01 -7.30846822e-01 1.12711318e-01
1.57517165e-01 -7.31454909e-01 4.54140335e-01 8.78789008e-01
8.76584709e-01 -2.58330137e-01 -3.35702777e-01 6.95471644e-01
6.29192650e-01 -2.13689327e-01 -7.83464313e-01 4.45368141e-01
4.16163206e-01 1.77769494e+00 -8.51869345e-01 -1.21008761e-01
-9.28576067e-02 9.16797400e-01 -4.96490121e-01 3.10242355e-01
-8.07424545e-01 -4.72316980e-01 2.44555190e-01 2.84281850e-01
4.34259206e-01 3.08988560e-02 -2.53716677e-01 -6.04991436e-01
1.44322395e-01 -2.34448701e-01 3.82890701e-01 -1.35499620e+00
-8.96823704e-01 5.31031072e-01 -8.31686109e-02 -1.17649519e+00
1.84185922e-01 -1.04158628e+00 -1.16448247e+00 9.56941247e-01
-1.22207749e+00 -8.00724745e-01 -8.44294488e-01 2.65746564e-01
3.07266772e-01 2.79373322e-02 8.34164858e-01 -8.79227591e-04
-7.76064634e-01 -1.37373209e-01 4.34235126e-01 -4.72501725e-01
2.93190628e-01 -1.48710537e+00 -1.99260056e-01 1.01294470e+00
-2.40066022e-01 -3.94723266e-01 8.91934395e-01 -5.25676668e-01
-1.05848944e+00 -8.22655678e-01 4.90276009e-01 -2.06343949e-01
5.03552496e-01 -1.81174348e-03 -7.14149654e-01 2.56422430e-01
6.40565991e-01 -4.14164335e-01 2.58872718e-01 -1.93750650e-01
1.98802173e-01 -5.86382389e-01 -1.41073072e+00 1.73403516e-01
1.78849176e-01 -3.43490809e-01 -2.65929103e-01 7.23361373e-01
-4.47475314e-02 1.05297565e-01 -9.74850833e-01 5.28984487e-01
3.11964244e-01 -1.59879899e+00 6.36482298e-01 5.62234163e-01
1.04812741e-01 -5.47165930e-01 1.98387265e-01 -1.30332482e+00
-3.41049671e-01 -4.89537805e-01 5.01582921e-01 1.53327096e+00
3.92109841e-01 -8.14111888e-01 5.02027750e-01 9.84746870e-03
-3.74249250e-01 -2.30624542e-01 -7.59958327e-01 -5.83594799e-01
-6.23100519e-01 7.50876144e-02 -3.33789997e-02 6.11449003e-01
-1.51955366e-01 -3.20067629e-02 9.41159762e-03 1.09754014e+00
6.03592992e-01 2.33704120e-01 4.90402400e-01 -1.67022228e+00
-1.93779513e-01 -4.95662801e-02 -2.75021851e-01 -1.71241403e-01
-5.74632525e-01 -2.99952418e-01 5.09815514e-02 -1.74918520e+00
-3.72294515e-01 -2.09104493e-01 -8.83204937e-02 3.72610599e-01
1.96145013e-01 2.62350380e-01 -1.41333491e-02 -1.35972768e-01
2.62197345e-01 3.32415551e-01 5.43842554e-01 4.16760705e-02
-1.02381684e-01 1.54088765e-01 1.05102276e-02 8.25575233e-01
1.00615811e+00 -1.48192123e-01 -3.07723254e-01 -3.07459325e-01
3.98761928e-01 8.20121169e-02 7.28957415e-01 -1.52049792e+00
-2.78192341e-01 8.99863318e-02 4.23115283e-01 -7.57884204e-01
2.16907024e-01 -1.32033825e+00 3.65620434e-01 8.20978701e-01
5.81005931e-01 -5.61470315e-02 2.80692250e-01 4.09126014e-01
1.62312612e-02 -6.15306437e-01 7.84828901e-01 1.12880521e-01
-5.28385222e-01 -3.28918040e-01 -8.90651345e-01 -7.17087269e-01
1.38359511e+00 -3.43837738e-01 -5.99324286e-01 -1.07749149e-01
-2.96096861e-01 2.84740657e-01 7.98763931e-01 -1.16266735e-01
1.66473418e-01 -5.76975584e-01 -4.80444551e-01 1.41536012e-01
-8.32712185e-03 -3.31848562e-01 8.38287711e-01 7.84028888e-01
-1.04694104e+00 1.35306671e-01 -5.24189889e-01 -8.04970205e-01
-1.72803843e+00 1.72342092e-01 5.94652832e-01 8.62330720e-02
-3.69726390e-01 5.26673555e-01 -1.26770046e-02 -7.00808987e-02
-1.95294455e-01 -4.70712692e-01 -3.84462804e-01 1.87495887e-01
4.13625799e-02 9.07032132e-01 3.09731752e-01 -2.81154573e-01
-2.94813544e-01 9.17967260e-01 7.51922071e-01 2.08059683e-01
1.21653032e+00 2.71777762e-03 -3.54160011e-01 3.62569362e-01
3.37033123e-01 1.57759547e-01 -1.38837373e+00 6.43697619e-01
-1.64740086e-01 -1.02654941e-01 1.08875155e-01 -1.37354410e+00
-9.10305619e-01 8.55546415e-01 9.43988621e-01 1.07913053e+00
1.68288696e+00 -3.37906241e-01 1.41118899e-01 1.93456098e-01
2.69651592e-01 -1.49598777e+00 -3.93138796e-01 -7.14335591e-02
9.58280802e-01 -9.42199051e-01 3.08727324e-01 -7.86869764e-01
-5.44752419e-01 1.36390233e+00 5.01992762e-01 -1.99056745e-01
3.00595462e-01 6.85730428e-02 1.44230336e-01 -2.78395295e-01
-5.30197084e-01 -1.54790536e-01 1.25433570e-02 8.72136354e-01
1.72102734e-01 2.26847216e-01 -3.08381319e-01 -2.36632466e-01
3.27687472e-01 -2.92753369e-01 7.74015844e-01 7.32817471e-01
-7.86924422e-01 -7.44203150e-01 -8.82113814e-01 4.13055331e-01
-3.91994156e-02 4.71725971e-01 -4.62430418e-01 8.57334614e-01
5.10815024e-01 1.29573309e+00 -6.95362361e-03 -2.43417576e-01
5.32666504e-01 -1.25567213e-01 3.28653812e-01 -3.86081696e-01
-8.63769054e-01 9.77625549e-02 3.23193252e-01 -1.27587661e-01
-4.39278603e-01 -7.05950677e-01 -6.08091474e-01 7.88764358e-02
-7.32598782e-01 2.65665203e-01 1.03592110e+00 5.41176021e-01
1.08794205e-01 6.77043080e-01 9.89362895e-01 -7.23392010e-01
-1.90373525e-01 -1.23064351e+00 -1.15518081e+00 -3.14060628e-01
1.19078591e-01 -6.34907782e-01 -6.11987531e-01 -6.86107669e-03] | [7.170417785644531, 1.9297723770141602] |
77d2786b-d0b0-44af-99f4-5b3fc3e3f7e5 | dimensionapp-android-app-to-estimate-object | 1609.07597 | null | http://arxiv.org/abs/1609.07597v1 | http://arxiv.org/pdf/1609.07597v1.pdf | DimensionApp : android app to estimate object dimensions | In this project, we develop an android app that uses on computer vision
techniques to estimate an object dimension present in field of view. The app
while having compact size, is accurate upto +/- 5 mm and robust towards touch
inputs. We use single-view metrology to compute accurate measurement. Unlike
previous approaches, our technique does not rely on line detection and can be
generalize to any object shape easily. | ['Vijay Kumar', 'Suriya Singh'] | 2016-09-24 | null | null | null | null | ['line-detection'] | ['computer-vision'] | [ 5.61164394e-02 -6.99125603e-02 -5.57045732e-03 -2.00809300e-01
-2.48218670e-01 -9.00554538e-01 3.30599695e-01 -4.62970883e-01
-4.42926288e-02 1.41616434e-01 -4.76389647e-01 -5.51345646e-01
7.03168288e-02 -5.62066495e-01 -7.36324668e-01 4.74560410e-02
6.16403759e-01 3.54926825e-01 5.00361621e-01 7.05518341e-03
8.34756970e-01 5.06379247e-01 -1.29974306e+00 -1.54475287e-01
1.92155540e-01 1.39916480e+00 1.40114188e-01 1.00965917e+00
2.74971783e-01 9.97845083e-02 -5.35038292e-01 -5.24506450e-01
6.03298604e-01 2.85060436e-01 -2.59212367e-02 1.42099289e-02
1.00580192e+00 -8.88192117e-01 5.53754568e-02 8.17359567e-01
5.00406146e-01 -7.47536004e-01 5.05012631e-01 -9.55762029e-01
-4.33991581e-01 7.53352791e-02 -9.48342264e-01 5.54764830e-02
1.07316327e+00 -1.90144703e-01 3.40445668e-01 -1.10138643e+00
4.90942359e-01 9.32174504e-01 1.35384965e+00 3.01529557e-01
-8.33628714e-01 -3.85758817e-01 -5.28414071e-01 -4.37874705e-01
-1.10602474e+00 -2.80826718e-01 8.49157989e-01 -6.54778242e-01
9.36545491e-01 3.59017134e-01 7.22922266e-01 6.46222293e-01
9.55156982e-01 3.37887220e-02 1.42667842e+00 -5.62703431e-01
2.16076210e-01 4.95190114e-01 9.61285830e-03 5.99819899e-01
5.49715459e-01 -1.11543285e-02 -1.58395812e-01 -1.03392377e-01
1.36309195e+00 2.79523522e-01 2.83017419e-02 -7.27137864e-01
-1.10307443e+00 4.67476159e-01 -8.30092728e-02 -1.99609026e-01
-4.56836894e-02 3.45542699e-01 5.64340986e-02 -5.86593263e-02
1.98806688e-01 3.90894473e-01 -6.08096600e-01 -8.10700893e-01
-8.30988407e-01 -1.58090115e-01 9.11319733e-01 1.24491405e+00
4.47799921e-01 -2.03369096e-01 6.37821555e-01 4.55489337e-01
6.39343441e-01 1.05387282e+00 1.56245008e-01 -1.50537252e+00
5.93749166e-01 6.71060264e-01 3.43562812e-01 -8.69633555e-01
-3.51882309e-01 4.21488047e-01 6.00434490e-04 8.62086952e-01
2.42780045e-01 -3.11738789e-01 -7.91142225e-01 3.56050253e-01
5.37543356e-01 -2.14219779e-01 -3.82624298e-01 5.98588467e-01
9.85099673e-01 -2.76124179e-01 -8.24885070e-01 2.18242198e-01
1.20752883e+00 -6.36949599e-01 -5.85622251e-01 -1.23763993e-01
9.39688534e-02 -1.38583124e+00 1.19828355e+00 1.03589880e+00
-1.17112947e+00 -4.89556551e-01 -1.50490391e+00 -1.28902167e-01
-1.25126317e-01 2.54907846e-01 7.85566986e-01 1.35977197e+00
-7.94376910e-01 3.74942154e-01 -6.93554461e-01 -1.98554814e-01
1.48097903e-01 5.92161894e-01 -1.81237370e-01 2.14696363e-01
-1.18095197e-01 9.87012386e-01 -3.13900203e-01 -2.71247625e-01
-2.20525060e-02 -7.89086401e-01 -4.38477665e-01 -5.78614533e-01
2.08716080e-01 -5.17036855e-01 1.46691608e+00 -9.19673368e-02
-2.18454146e+00 8.95478308e-01 3.95210348e-02 -2.74197251e-01
3.52131993e-01 -7.17286706e-01 -4.09680605e-01 5.66318691e-01
-1.52993843e-01 1.72388241e-01 1.09058797e+00 -1.16226542e+00
-3.00456494e-01 -5.73969543e-01 4.14645702e-01 1.10231578e-01
-2.08810833e-03 -1.55002534e-01 -2.49017939e-01 -1.42513365e-01
6.59226179e-01 -1.03969336e+00 1.37009874e-01 5.63263416e-01
-2.35581651e-01 3.96609932e-01 1.42777884e+00 -4.01942790e-01
7.84307778e-01 -1.69755900e+00 -6.31351590e-01 2.40824535e-01
2.29371160e-01 3.62046570e-01 8.10983539e-01 4.04176086e-01
5.70044339e-01 -2.74300635e-01 3.08506638e-01 -1.70334168e-02
-1.97016209e-01 -1.70117214e-01 -1.53176308e-01 4.77143228e-01
-4.45784062e-01 7.61179507e-01 -2.73356497e-01 -3.40617180e-01
7.14166164e-01 5.89710295e-01 -2.82383382e-01 -3.51678699e-01
1.93775222e-01 1.33750916e-01 -3.68026853e-01 1.17802179e+00
1.10930884e+00 -5.57591282e-02 -1.22441091e-01 -2.40261763e-01
-2.47882411e-01 5.10531329e-02 -1.47636712e+00 1.53806520e+00
-8.59887242e-01 8.74493659e-01 3.04705560e-01 -3.55639875e-01
1.18447518e+00 1.24385189e-02 6.28161311e-01 -3.83371353e-01
2.70604521e-01 1.35815278e-01 -3.29986006e-01 -6.94503427e-01
4.31215018e-01 6.96147382e-02 1.20764792e-01 5.39782286e-01
-3.58102083e-01 -7.90856302e-01 -4.74851519e-01 -1.01172432e-01
8.53568554e-01 5.53357244e-01 5.96825361e-01 1.21875051e-02
2.86860794e-01 -4.69775945e-02 -1.42186582e-02 3.76931131e-01
-4.24063504e-02 8.12299430e-01 1.99217070e-02 -3.45902860e-01
-9.63963628e-01 -1.33363438e+00 -6.42703891e-01 4.52204794e-01
5.48977196e-01 -1.20499931e-01 -8.30182195e-01 -5.76858163e-01
4.25636530e-01 2.06506908e-01 -2.06919968e-01 6.25446320e-01
-2.38055006e-01 2.21385315e-01 3.76665831e-01 9.26156461e-01
3.85271937e-01 -3.43829304e-01 -1.31169832e+00 6.19297773e-02
7.11536229e-01 -1.43052375e+00 -4.43862289e-01 -3.63777578e-01
-1.20413208e+00 -1.25713694e+00 -6.70476019e-01 -2.69477040e-01
5.73749065e-01 5.90064406e-01 1.17513800e+00 -3.87570113e-01
-5.83754063e-01 1.45797873e+00 -7.87889361e-02 -1.18481088e+00
9.21420828e-02 -5.62123001e-01 2.01490477e-01 -5.82596838e-01
7.09642589e-01 -5.91152132e-01 -8.81996274e-01 6.63149536e-01
2.07961607e-03 -3.12683553e-01 3.75003129e-01 2.59120972e-03
5.13971627e-01 -4.75582540e-01 3.30500416e-02 -5.75669289e-01
3.40587288e-01 -7.45143294e-02 -8.53274047e-01 -1.31562680e-01
-7.40300596e-01 -5.89587688e-01 1.20704353e-01 -5.09097099e-01
-8.51854026e-01 2.98697859e-01 1.34351701e-01 -3.38473529e-01
-2.49846056e-01 -1.28354907e-01 1.58421814e-01 -7.36477137e-01
6.51381016e-01 -3.30636859e-01 2.54039139e-01 -3.94542694e-01
3.47420245e-01 1.15722871e+00 4.95770365e-01 9.25553590e-03
6.42373919e-01 1.18652260e+00 2.49598071e-01 -1.11761725e+00
-2.94251680e-01 -4.96053159e-01 -1.05473304e+00 -3.59941721e-01
4.05864954e-01 -8.55974317e-01 -1.12254679e+00 2.60665387e-01
-1.07671654e+00 -7.28124678e-02 2.92507671e-02 8.68692517e-01
-6.70370638e-01 3.41786146e-01 -2.67946303e-01 -1.03681934e+00
-4.41703349e-01 -8.99125576e-01 1.31763363e+00 4.32391971e-01
-2.69898921e-01 -1.09965825e+00 -9.21039134e-02 5.46744943e-01
2.89434165e-01 2.15239257e-01 -2.40791291e-01 3.36958528e-01
-5.60712814e-01 -6.42445803e-01 -1.39670640e-01 9.59334150e-02
1.74557954e-01 4.76416081e-01 -9.88154113e-01 -2.35882308e-02
3.45133632e-01 1.02608331e-01 1.87894423e-03 6.99619532e-01
7.95398533e-01 9.15866718e-02 -6.82297111e-01 5.33009946e-01
1.55844474e+00 2.73327470e-01 7.68133700e-01 3.15179408e-01
7.71917462e-01 7.10123926e-02 9.87156987e-01 4.80776161e-01
5.00724077e-01 1.02725101e+00 6.15995288e-01 3.07788312e-01
-5.65508530e-02 8.53727534e-02 2.21845984e-01 6.38197422e-01
-3.91985714e-01 4.87224758e-01 -9.05386448e-01 2.88992561e-02
-1.14172280e+00 -6.78403020e-01 -8.06334794e-01 2.21595907e+00
2.44572461e-01 4.40197140e-01 2.42603257e-01 1.49736956e-01
4.02812511e-01 -5.54651916e-01 -5.51302612e-01 -1.01794660e+00
7.06192970e-01 2.46236935e-01 1.01187015e+00 4.35315728e-01
-8.75220001e-01 3.88918519e-01 8.01576710e+00 8.61738846e-02
-1.53671956e+00 8.44318569e-02 -3.65103960e-01 -1.82206184e-01
-3.38508189e-03 2.55449358e-02 -1.10571921e+00 4.29373801e-01
4.33237165e-01 2.64457345e-01 2.28276476e-01 1.21026993e+00
-1.52343139e-01 -8.50176334e-01 -9.65938210e-01 1.38300288e+00
4.33365315e-01 -1.00822854e+00 -7.78464735e-01 2.35702246e-01
5.85919559e-01 -6.58378145e-03 2.83699602e-01 -2.55358875e-01
-2.93969333e-01 -7.42490590e-01 5.00751793e-01 5.91832459e-01
1.10088003e+00 -4.20933455e-01 5.14783204e-01 4.78358030e-01
-1.04103339e+00 1.72177762e-01 -5.36786258e-01 -5.47220290e-01
3.63464132e-02 6.42747045e-01 -1.37806296e+00 2.72164829e-02
8.89609933e-01 2.66429991e-01 -5.42395890e-01 9.25924957e-01
1.63921192e-01 2.78681695e-01 -7.43637443e-01 -2.43998691e-01
-2.96902508e-01 -3.10173631e-01 4.35293108e-01 7.62403607e-01
7.49589443e-01 -1.48870036e-01 -1.58098295e-01 4.71506208e-01
4.98832971e-01 -4.70060632e-02 -1.14570355e+00 4.66564804e-01
5.38773596e-01 1.43107402e+00 -8.18470538e-01 1.19606713e-02
-7.22759545e-01 9.36486542e-01 -3.27869058e-01 -1.43907905e-01
-7.20727086e-01 -6.87422752e-01 -2.98236101e-03 5.78701138e-01
7.67344058e-01 -7.06066430e-01 -1.13959026e+00 -8.47190142e-01
5.39140701e-01 -3.11196178e-01 -4.67362553e-01 -8.56766820e-01
-7.55896032e-01 1.39157668e-01 -1.14658102e-01 -1.77273083e+00
-1.66802615e-01 -1.12366426e+00 -7.85166502e-01 4.69573617e-01
-7.83662260e-01 -1.39804351e+00 -3.93285930e-01 2.88943648e-01
3.99460644e-01 -2.90254653e-01 7.30189681e-01 1.33090749e-01
1.51715487e-01 5.20829082e-01 1.51869515e-02 -3.11063498e-01
5.90240777e-01 -1.33975625e+00 4.64935720e-01 2.83564419e-01
1.06675796e-01 8.66846561e-01 6.59076333e-01 -8.26792240e-01
-1.83667481e+00 -3.46775442e-01 5.15416563e-02 -1.24970651e+00
5.31121016e-01 -3.65912199e-01 -2.75238097e-01 7.92240322e-01
2.07203299e-01 3.19083452e-01 5.13688922e-01 8.03528279e-02
-2.83855885e-01 -9.13186297e-02 -1.53845906e+00 1.76965132e-01
7.67090559e-01 -4.35477853e-01 -6.54028773e-01 4.29908708e-02
1.31876737e-01 -1.27803898e+00 -1.23674679e+00 3.21036398e-01
1.48223245e+00 -1.36781609e+00 1.14238846e+00 2.41626143e-01
-1.06157577e-02 -1.40230730e-01 -1.46841705e-01 -7.30139256e-01
1.02208547e-01 -7.67473698e-01 -5.85080683e-01 9.15402889e-01
3.50418121e-01 -8.27149630e-01 1.04514360e+00 5.98104358e-01
8.65158997e-03 -8.08982849e-01 -1.02485931e+00 -8.17688107e-01
-2.86790490e-01 -6.69645011e-01 3.63868117e-01 3.79401267e-01
2.08702222e-01 2.28531629e-01 -1.33347258e-01 2.25609049e-01
7.33005583e-01 1.86958611e-01 1.22923863e+00 -1.44034421e+00
-4.66537386e-01 -1.18204495e-02 -6.93125486e-01 -1.06876099e+00
-7.34219253e-01 -1.56768531e-01 -3.82351637e-01 -1.39244938e+00
-2.67515123e-01 -3.18301350e-01 4.88164753e-01 -1.00365773e-01
4.10376042e-01 9.19844508e-01 -6.19789548e-02 -1.84288204e-01
-2.45311439e-01 -4.87452090e-01 1.13684225e+00 4.12631541e-01
-2.65218109e-01 5.46059132e-01 -5.77748179e-01 1.48090708e+00
8.13297808e-01 -2.73551077e-01 -3.58000964e-01 -3.48924220e-01
4.78119791e-01 -1.47439644e-01 4.51378286e-01 -1.33398569e+00
9.72230062e-02 -3.61287482e-02 6.43263936e-01 -1.01645041e+00
7.24957049e-01 -1.18287945e+00 -7.79638514e-02 2.98119694e-01
5.00313103e-01 1.34847760e-01 1.58630535e-01 2.34692052e-01
6.29998446e-01 -6.18193090e-01 4.93225038e-01 -3.09040666e-01
-4.66340929e-01 -3.13745961e-02 -7.75425732e-02 -3.16177368e-01
1.36160350e+00 -1.08915198e+00 -2.65104264e-01 -4.93851572e-01
-4.66407418e-01 -4.14549112e-01 1.01697516e+00 2.42308348e-01
8.78917992e-01 -1.24333501e+00 3.36886346e-02 2.95516729e-01
-7.23369047e-02 -3.46756876e-01 -2.12626725e-01 8.32975268e-01
-1.08455765e+00 4.88521069e-01 -4.68608364e-02 -1.05480373e+00
-1.55513310e+00 6.63552225e-01 1.83862057e-02 6.26454949e-01
-6.93653941e-01 3.00625473e-01 -5.50344825e-01 -3.33823740e-01
-6.55409619e-02 -4.46535736e-01 1.44668985e-02 -4.43507880e-01
6.09436512e-01 9.64006305e-01 1.50175750e-01 -4.17260319e-01
-4.25884068e-01 1.66669714e+00 2.17750818e-01 -4.11742806e-01
8.85885358e-01 -4.48115766e-01 2.47547060e-01 6.65084541e-01
9.13493097e-01 8.48344922e-01 -1.29999614e+00 4.19333696e-01
-2.76153415e-01 -7.19503880e-01 1.03870593e-01 -6.27992988e-01
-5.67103624e-01 1.05079103e+00 9.82181907e-01 3.11912954e-01
5.28142512e-01 -5.12346551e-02 7.69016445e-01 4.45436716e-01
7.94629037e-01 -1.40913033e+00 1.70544937e-01 3.24656218e-01
9.72268164e-01 -1.30975246e+00 7.56015897e-01 -9.48141098e-01
-3.89360249e-01 1.60060978e+00 7.15955853e-01 -4.06589329e-01
9.74707842e-01 1.13801837e+00 3.63412708e-01 -4.03390259e-01
2.67759617e-02 2.46469676e-01 4.03438509e-01 1.23433757e+00
3.23713899e-01 1.45397484e-01 -3.29484314e-01 -2.17595417e-03
-4.53701645e-01 2.90093660e-01 6.76909745e-01 1.29414630e+00
-6.32709861e-01 -8.47843885e-01 -7.86127388e-01 6.08468652e-01
-7.39155471e-01 4.41336989e-01 -2.44121000e-01 7.93396294e-01
1.11555666e-01 1.03632438e+00 1.37691451e-02 -4.16573107e-01
5.69334149e-01 -4.83788222e-01 1.02739334e+00 -5.40331006e-01
-1.11429200e-01 6.45022392e-02 -6.78521469e-02 -7.63537824e-01
-4.17487651e-01 -7.04123378e-01 -1.07117105e+00 -4.00594771e-02
-5.95273256e-01 -7.82731295e-01 1.36694586e+00 7.03478873e-01
3.54730278e-01 -1.06301486e-01 7.75133014e-01 -1.18472183e+00
-4.62499797e-01 -8.68221581e-01 -5.81668079e-01 -2.91192293e-01
7.98787400e-02 -8.34935844e-01 -2.90107243e-02 -3.41373459e-02] | [7.197399139404297, -1.853243112564087] |
9e8c17d6-4295-446c-9412-94dd421acb27 | box-adapt-domain-adaptive-medical-image | 2108.08432 | null | https://arxiv.org/abs/2108.08432v2 | https://arxiv.org/pdf/2108.08432v2.pdf | Box-Adapt: Domain-Adaptive Medical Image Segmentation using Bounding BoxSupervision | Deep learning has achieved remarkable success in medicalimage segmentation, but it usually requires a large numberof images labeled with fine-grained segmentation masks, andthe annotation of these masks can be very expensive andtime-consuming. Therefore, recent methods try to use un-supervised domain adaptation (UDA) methods to borrow in-formation from labeled data from other datasets (source do-mains) to a new dataset (target domain). However, due tothe absence of labels in the target domain, the performance ofUDA methods is much worse than that of the fully supervisedmethod. In this paper, we propose a weakly supervised do-main adaptation setting, in which we can partially label newdatasets with bounding boxes, which are easier and cheaperto obtain than segmentation masks. Accordingly, we proposea new weakly-supervised domain adaptation method calledBox-Adapt, which fully explores the fine-grained segmenta-tion mask in the source domain and the weak bounding boxin the target domain. Our Box-Adapt is a two-stage methodthat first performs joint training on the source and target do-mains, and then conducts self-training with the pseudo-labelsof the target domain. We demonstrate the effectiveness of ourmethod in the liver segmentation task. Weakly supervised do-main adaptation | ['Shaoan Xie', 'Kayhan Batmanghelich', 'Mingming Gong', 'Yanwu Xu'] | 2021-08-19 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 6.94248825e-02 2.77795762e-01 -7.43962884e-01 -4.64359343e-01
-9.94218826e-01 -5.78198016e-01 4.32509094e-01 1.49962187e-01
-5.30058682e-01 8.78993452e-01 5.20933904e-02 -1.95907161e-01
2.89228767e-01 -7.91233659e-01 -6.51002049e-01 -9.55905318e-01
3.68231595e-01 7.56886840e-01 4.73157823e-01 1.22252606e-01
-3.89325619e-01 3.48008543e-01 -8.59769404e-01 2.69929230e-01
1.01949859e+00 8.02106321e-01 2.75372893e-01 3.87076616e-01
-2.44783580e-01 3.35029513e-01 -5.56991696e-01 -7.56006911e-02
1.57021880e-01 -8.73962879e-01 -1.07469177e+00 5.86428821e-01
4.46925871e-02 -2.17251316e-01 2.45958511e-02 1.15510428e+00
2.28374436e-01 -5.56951948e-02 7.00275481e-01 -8.53480220e-01
-2.83423662e-01 5.39160311e-01 -7.70058632e-01 2.29507327e-01
-3.19528282e-01 -7.99759477e-02 5.08061886e-01 -8.23338270e-01
6.61673307e-01 7.55543768e-01 7.36856580e-01 6.66545093e-01
-1.42508805e+00 -8.81593764e-01 2.62575030e-01 -2.79023796e-01
-1.22000694e+00 -1.95622116e-01 9.29773927e-01 -5.48549652e-01
2.58667022e-01 -3.12727354e-02 5.04237294e-01 8.35421741e-01
-1.18158907e-01 9.34239388e-01 1.51639998e+00 -4.17339593e-01
3.01454902e-01 2.77757704e-01 2.63820916e-01 8.15935731e-01
4.43136096e-02 2.55591273e-01 4.18818668e-02 -1.35297682e-02
1.10557985e+00 8.59374404e-02 -4.74582836e-02 -6.13800943e-01
-1.25759971e+00 7.97382653e-01 5.21624804e-01 5.59441090e-01
-3.49312901e-01 -3.17925125e-01 4.40134943e-01 1.91193268e-01
5.82111537e-01 2.44316801e-01 -6.16865277e-01 3.43005031e-01
-1.11668694e+00 -2.28472963e-01 5.93876600e-01 8.64937425e-01
9.42597687e-01 -2.02765062e-01 -9.55835879e-02 9.22239840e-01
1.69955730e-01 2.92056590e-01 8.07608008e-01 -5.68973660e-01
1.69773549e-01 6.46580398e-01 -3.14055220e-03 -4.09032911e-01
-6.42603397e-01 -4.03291494e-01 -1.10011530e+00 9.66257453e-02
8.07461262e-01 -5.12952387e-01 -1.42553365e+00 1.57024527e+00
7.07552969e-01 1.40672982e-01 -2.88510993e-02 9.97355640e-01
9.18823600e-01 3.56401235e-01 2.70860434e-01 -2.68734962e-01
1.30751765e+00 -1.34685028e+00 -6.81306601e-01 -2.33132303e-01
9.30004060e-01 -5.87162495e-01 1.13477612e+00 1.54425070e-01
-8.34069431e-01 -6.72663510e-01 -8.68677616e-01 1.24156818e-01
-1.20551459e-01 1.81957290e-01 4.91564959e-01 4.10759360e-01
-8.99324238e-01 3.58058691e-01 -8.75384986e-01 -3.29281360e-01
8.66320014e-01 2.70574033e-01 -4.13029730e-01 -7.36542279e-03
-9.06638861e-01 6.40823424e-01 7.53913999e-01 -3.70337367e-01
-9.71453428e-01 -6.70078635e-01 -9.35237110e-01 -2.21582741e-01
6.91238344e-01 -3.60497922e-01 1.22466564e+00 -1.26968980e+00
-1.42000306e+00 1.22783172e+00 6.71301410e-02 -2.64805108e-01
4.48953390e-01 1.53425902e-01 -1.50191277e-01 3.29937547e-01
2.01771215e-01 7.79782295e-01 9.11077678e-01 -1.29892242e+00
-5.20504892e-01 -3.71151417e-01 4.49663997e-02 1.93545356e-01
-2.45849371e-01 -1.89355850e-01 -6.37043059e-01 -1.22333086e+00
2.74590939e-01 -8.85620415e-01 -4.70005393e-01 2.89000832e-02
-7.12058008e-01 -3.30608904e-01 7.85418332e-01 -5.63904881e-01
1.10945165e+00 -2.36433172e+00 4.06450406e-02 1.21352106e-01
3.83806348e-01 4.86311823e-01 -1.67246595e-01 -2.93753028e-01
-3.66676390e-01 8.80049355e-03 -6.74062908e-01 -2.25271076e-01
-2.53413171e-01 5.59538126e-01 6.40233979e-02 3.85961562e-01
1.18585676e-01 9.08669472e-01 -1.15521669e+00 -9.60041463e-01
2.04976305e-01 -5.60883805e-02 -4.45033967e-01 4.76414829e-01
-3.52488369e-01 8.10831428e-01 -5.69795489e-01 6.18844688e-01
6.20697439e-01 -4.92128849e-01 1.77196562e-01 -1.33107960e-01
1.61045209e-01 -1.02864280e-01 -9.38377380e-01 1.88231921e+00
-3.55069101e-01 1.82600856e-01 1.79074734e-01 -1.40854108e+00
8.32177341e-01 3.41520995e-01 7.70582438e-01 -4.72704440e-01
2.99705029e-01 3.30725670e-01 -1.29120588e-01 -5.58980286e-01
-3.27055156e-01 -7.73714006e-01 -2.70771474e-01 5.50448358e-01
2.03885272e-01 -3.28671515e-01 1.54465556e-01 -1.02638237e-01
6.03227019e-01 3.19454521e-01 5.88874817e-01 -4.24726993e-01
6.79836750e-01 2.36140385e-01 8.05869043e-01 6.66484714e-01
-4.89033639e-01 6.48011565e-01 4.37264800e-01 -5.11193097e-01
-1.05354691e+00 -8.30074549e-01 -3.97617638e-01 1.14468575e+00
2.23094299e-01 -7.78728127e-02 -1.11414456e+00 -1.60177875e+00
-9.82779562e-02 1.53507978e-01 -7.47394383e-01 -1.32895663e-01
-5.81574261e-01 -8.73102367e-01 5.81126094e-01 8.42959464e-01
9.27548826e-01 -9.26432729e-01 -3.28098357e-01 2.75210410e-01
-1.36873022e-01 -1.02704585e+00 -8.26093256e-01 4.25272882e-01
-1.12203574e+00 -1.11972654e+00 -1.34338772e+00 -1.29163396e+00
1.08999109e+00 -8.26344043e-02 1.01707554e+00 1.04250781e-01
1.00030936e-01 -6.10204637e-02 -2.96515226e-01 -3.43518674e-01
-5.10508955e-01 1.48331553e-01 -2.75593460e-01 3.48350219e-03
5.25618613e-01 -2.69705087e-01 -4.51626211e-01 5.12887001e-01
-8.65157306e-01 2.51596838e-01 7.71377861e-01 1.22356880e+00
1.18096924e+00 9.64932740e-02 6.70548737e-01 -1.46019924e+00
1.55946702e-01 -4.02847826e-01 -5.12750506e-01 1.34806812e-01
-3.75504225e-01 -6.26886263e-02 8.51711810e-01 -5.93866467e-01
-1.03215158e+00 4.00091141e-01 -3.29731286e-01 -3.50642890e-01
-6.80705428e-01 4.16723847e-01 -3.51104796e-01 -2.25211326e-02
8.01022112e-01 -2.97713522e-02 7.68520087e-02 -7.31097996e-01
3.59217644e-01 5.62793732e-01 5.78249335e-01 -4.35077846e-01
8.50792706e-01 5.81560194e-01 -2.59194255e-01 -4.87823725e-01
-1.19267273e+00 -5.12821198e-01 -1.04742897e+00 2.05403090e-01
1.04458976e+00 -7.09532380e-01 1.64871648e-01 6.87305152e-01
-7.50474215e-01 -8.84410441e-01 -7.52343714e-01 5.58316886e-01
-4.63462621e-01 4.79054779e-01 -6.57677054e-01 -4.59556058e-02
-1.26615807e-01 -1.13382816e+00 1.00348270e+00 3.10575217e-01
-1.85188539e-02 -1.35786462e+00 1.04149012e-02 2.02042207e-01
4.58183400e-02 3.74759793e-01 9.06829596e-01 -9.88570273e-01
-1.68882042e-01 -1.51663899e-01 -4.90764618e-01 5.79850614e-01
3.40757221e-01 -8.02705109e-01 -7.62174308e-01 -3.16549510e-01
2.01835725e-02 -3.63356918e-01 8.88869584e-01 5.52467823e-01
1.32125199e+00 -1.50664672e-01 -6.16839409e-01 8.27986360e-01
1.23882639e+00 1.01724990e-01 8.11484307e-02 1.82214364e-01
7.17598498e-01 4.99436587e-01 5.94495654e-01 1.74821943e-01
4.38409150e-01 6.26975656e-01 -2.32581813e-02 -9.96404052e-01
-4.83181924e-01 -1.19501904e-01 -3.38732265e-02 5.22781432e-01
1.63357973e-01 1.51380599e-01 -9.40404415e-01 6.88598990e-01
-1.56165493e+00 -2.42509767e-01 2.14339957e-01 1.99802756e+00
1.32683802e+00 2.08524123e-01 5.54813802e-01 -1.24802984e-01
7.46356428e-01 -5.03468663e-02 -8.03493857e-01 -5.48704453e-02
6.94530904e-02 1.72991246e-01 3.58866513e-01 4.36648220e-01
-1.61147726e+00 9.55410659e-01 6.20855808e+00 8.12681317e-01
-1.10258150e+00 3.82581770e-01 8.78756940e-01 2.88018972e-01
-5.61901834e-03 -2.46312007e-01 -6.23677075e-01 5.79464853e-01
6.10234857e-01 5.91574199e-02 6.22874275e-02 9.57795262e-01
-5.60719520e-02 -1.28752977e-01 -1.04040062e+00 7.72218406e-01
1.56772450e-01 -1.16621935e+00 -2.43568361e-01 -1.36327092e-02
9.25454617e-01 -3.66222039e-02 -2.02596724e-01 4.48443234e-01
4.00613666e-01 -6.61180139e-01 1.48318365e-01 -4.99222754e-03
1.18991792e+00 -6.15153909e-01 9.59906220e-01 6.09084964e-01
-9.71710503e-01 6.52177483e-02 -2.04306200e-01 4.44698334e-01
6.41232058e-02 8.32978189e-01 -8.94094467e-01 2.72702008e-01
7.15308666e-01 7.05401182e-01 -3.04008991e-01 1.04226053e+00
-5.17619491e-01 7.32476592e-01 -2.06816256e-01 4.12340343e-01
2.76992321e-01 -2.15366811e-01 3.26954722e-01 1.33368373e+00
-8.30424577e-02 2.76316881e-01 8.44402909e-01 6.63487434e-01
-3.42395604e-01 1.27813160e-01 -2.83690840e-01 -8.64232406e-02
9.78518873e-02 1.16026199e+00 -1.11869824e+00 -7.69563198e-01
-4.70442683e-01 9.97162342e-01 2.37449147e-02 4.58935231e-01
-7.60563672e-01 -4.27001655e-01 1.96160272e-01 1.26942068e-01
2.58476108e-01 1.18208215e-01 -5.53174913e-01 -1.17679417e+00
-2.64483422e-01 -7.50266016e-01 9.01031733e-01 -2.91188061e-01
-1.23051810e+00 6.53537989e-01 2.11685658e-01 -1.08954597e+00
-5.05223982e-02 -3.85815382e-01 -2.88589835e-01 6.86255753e-01
-1.88801813e+00 -1.30602026e+00 -3.82779598e-01 6.89116478e-01
7.06363678e-01 -1.99007019e-02 8.30937505e-01 3.52964520e-01
-5.66928566e-01 6.70864820e-01 2.08966762e-01 5.65510452e-01
8.61847997e-01 -1.48621619e+00 2.94257700e-02 7.30958760e-01
-7.66918659e-02 2.49873638e-01 2.32460350e-01 -7.40577042e-01
-5.08981228e-01 -1.23481810e+00 6.79944992e-01 -1.51442394e-01
3.76730442e-01 -3.26153457e-01 -1.19508314e+00 7.04464674e-01
-8.05036444e-03 5.50154209e-01 8.46213102e-01 -4.84943986e-02
-3.88318449e-02 -4.40342762e-02 -1.36260712e+00 2.46733055e-01
7.52807736e-01 -4.21194434e-01 -8.66393447e-01 7.23117709e-01
6.12466395e-01 -7.37192452e-01 -8.57203186e-01 3.97679806e-01
6.05963767e-02 -6.21895492e-01 8.96006048e-01 -6.77817822e-01
1.48606583e-01 -2.78421283e-01 4.24300104e-01 -1.37377155e+00
-2.44754836e-01 -4.18292522e-01 -9.99859273e-02 9.08023179e-01
3.04835737e-01 -5.78885734e-01 1.23512590e+00 2.72695363e-01
-4.10202324e-01 -8.34145725e-01 -6.50085807e-01 -8.24919283e-01
6.48838520e-01 4.78294306e-02 7.36730516e-01 1.22687924e+00
-9.16244984e-02 2.29654148e-01 1.27010199e-03 3.88911404e-02
5.47713399e-01 3.33406776e-01 5.59522390e-01 -1.23294127e+00
-3.06722045e-01 -2.57534862e-01 -1.10828258e-01 -1.22777474e+00
1.98414475e-01 -9.11799073e-01 3.75745326e-01 -1.45971847e+00
2.29244158e-01 -7.33615756e-01 -4.03297573e-01 9.67450619e-01
-4.02482241e-01 5.88446558e-01 -1.23290814e-01 3.84065628e-01
-5.77257454e-01 1.74919903e-01 1.72755599e+00 -1.79686263e-01
-4.05786842e-01 2.21951511e-02 -5.65594018e-01 1.02685821e+00
7.55602002e-01 -4.97923881e-01 -3.83887529e-01 -2.51597881e-01
-4.97071981e-01 7.92331100e-02 9.05687585e-02 -7.16201246e-01
4.71847244e-02 -2.11837083e-01 5.50293803e-01 -5.75835526e-01
-9.31577757e-02 -8.22375298e-01 -3.00353765e-01 4.67727274e-01
-2.24524751e-01 -4.62988287e-01 9.23152715e-02 4.45306242e-01
-2.87408769e-01 -2.86887228e-01 1.22254765e+00 -4.49491918e-01
-7.66569197e-01 5.05771518e-01 -2.78009921e-01 2.58375615e-01
1.30549085e+00 -2.01966599e-01 1.75103590e-01 6.55030534e-02
-1.27855194e+00 2.14815453e-01 6.01564169e-01 -7.90905673e-03
4.23355728e-01 -1.25571966e+00 -6.24591887e-01 4.59132910e-01
9.98611227e-02 5.74006855e-01 1.95995346e-01 1.03474116e+00
-3.81920397e-01 2.61490226e-01 2.33386224e-03 -7.43113816e-01
-1.03373504e+00 9.17544842e-01 5.61182320e-01 -4.86869097e-01
-8.39100420e-01 1.10906565e+00 7.48067439e-01 -7.18457580e-01
1.45836025e-01 -3.71860385e-01 -3.41160357e-01 8.86210939e-04
4.72476006e-01 1.70188490e-04 -6.58830181e-02 -5.70734620e-01
-2.39370480e-01 4.89200324e-01 -2.48989090e-01 3.20942700e-01
1.13663530e+00 -8.66122544e-02 5.21904323e-03 6.24367744e-02
1.08097994e+00 -2.68959820e-01 -1.53502500e+00 -6.61886215e-01
1.57685712e-01 -3.52660090e-01 8.05362761e-02 -8.82794082e-01
-1.23273575e+00 9.23720896e-01 6.82356238e-01 -5.85269891e-02
1.33443689e+00 2.73213863e-01 1.09094906e+00 3.33435982e-02
2.99015760e-01 -1.20826769e+00 3.70171845e-01 3.27288598e-01
4.59354997e-01 -1.47976506e+00 -2.14449808e-01 -5.68092823e-01
-7.13353276e-01 8.98645341e-01 8.10450673e-01 -4.86640148e-02
8.74047935e-01 2.73389339e-01 3.71111572e-01 -4.13137004e-02
-5.95452124e-03 -3.50071073e-01 3.74951333e-01 7.98815072e-01
2.47036785e-01 1.90139517e-01 -3.39494079e-01 8.40637207e-01
2.43714720e-01 1.53902233e-01 2.48295516e-01 9.66006398e-01
-3.20905447e-01 -1.34628773e+00 -4.57434833e-01 5.08568287e-01
-4.12886322e-01 8.40618238e-02 -2.70374954e-01 8.87046456e-01
5.73866129e-01 6.67076051e-01 -5.04274219e-02 -1.14231020e-01
2.56998956e-01 -6.20833263e-02 2.53057152e-01 -1.05517697e+00
-3.85863394e-01 6.20115459e-01 -1.41498864e-01 -3.01022321e-01
-6.04263365e-01 -6.07591808e-01 -1.69042611e+00 3.22214335e-01
-5.00306308e-01 3.47376525e-01 2.79453754e-01 1.26320755e+00
3.18934321e-02 4.11782205e-01 5.33304214e-01 -5.46210587e-01
-3.76605719e-01 -7.34137595e-01 -5.78224301e-01 6.78011954e-01
4.16845769e-01 -6.63820207e-01 6.20863214e-02 3.36321324e-01] | [14.579072952270508, -2.014826536178589] |
8ddb9a75-635b-4e6f-bc95-e46c4e9e6ddd | precise-wifi-indoor-positioning-using-deep | 2307.02011 | null | https://arxiv.org/abs/2307.02011v1 | https://arxiv.org/pdf/2307.02011v1.pdf | Precise WiFi Indoor Positioning using Deep Learning Algorithms | This study demonstrates a WiFi indoor positioning system using Deep Learning algorithms. A new method using fitting function in MATLAB will be utilized to compute the path loss coefficient and log-normal fading variance. To reduce the error, a new hybrid localization approach utilizing Received Signal Strength Indicator (RSSI) and Angle of Arrival (AoA) has been created. Three Deep Learning algorithms would be utilized to decrease the adverse influence of the noise and interference. This paper compares the performance of two models in three different indoor environments. The average error of our hybrid positioning model trained by CNN in the big classroom is less than 250 mm. | ['Zihuai Lin', 'Minxue Cai'] | 2023-07-05 | null | null | null | null | ['hybrid-positioning'] | ['computer-vision'] | [-4.29914027e-01 -2.91608483e-01 4.18885767e-01 -4.54781562e-01
-4.31278110e-01 -3.70887697e-01 7.30519369e-02 2.17082173e-01
-3.58429074e-01 1.00457561e+00 -1.00678012e-01 -8.18606019e-01
-6.78308725e-01 -1.23143065e+00 -8.94196510e-01 -7.88933873e-01
-2.83550262e-01 -4.38835680e-01 2.41108462e-01 1.73678342e-03
3.09123456e-01 4.32471454e-01 -1.24969578e+00 -3.51536840e-01
1.18770134e+00 8.57923269e-01 3.25562626e-01 1.07691407e+00
-8.67556259e-02 6.44465864e-01 -1.18547416e+00 3.76976311e-01
1.24544874e-01 7.91162923e-02 -1.33236185e-01 -1.11330557e+00
5.63872874e-01 -3.66430789e-01 -5.50241709e-01 6.52461708e-01
1.33598518e+00 4.45462227e-01 2.42397189e-01 -1.34674990e+00
-3.95254284e-01 7.63854086e-01 -2.36887366e-01 5.02652943e-01
9.32140887e-01 -5.15072346e-01 -3.77661049e-01 -3.56590956e-01
-2.41633818e-01 7.90506542e-01 1.52880955e+00 -1.92915380e-01
-4.03708309e-01 -1.21614337e+00 -3.09602410e-01 1.38793930e-01
-1.44342661e+00 1.50948167e-01 5.98742425e-01 -4.63708714e-02
7.14847863e-01 1.34721711e-01 7.85477698e-01 1.18710852e+00
1.12503004e+00 2.58954495e-01 1.17110634e+00 -6.83907509e-01
2.90375590e-01 -7.20811412e-02 2.19281856e-02 7.87034810e-01
3.56358677e-01 9.75118950e-02 -3.69211107e-01 1.05109222e-01
8.39150071e-01 3.60119343e-01 -1.48377553e-01 1.16012499e-01
-6.67194009e-01 3.11251163e-01 8.71464550e-01 6.86303675e-01
-2.65217602e-01 6.88496411e-01 -3.33676964e-01 3.05710584e-01
-3.34262289e-02 3.58401656e-01 -7.50994205e-01 -5.45718372e-01
-6.62813127e-01 1.04942597e-01 8.73396456e-01 1.27144349e+00
4.76319790e-01 1.20908439e-01 2.27693170e-01 5.47463655e-01
8.98531258e-01 8.22768688e-01 4.73673105e-01 -1.00376260e+00
2.20377699e-01 -1.31289884e-01 1.50825649e-01 -1.40778840e+00
-6.96843266e-01 -1.16754043e+00 -4.67969507e-01 -9.73830558e-03
5.19709051e-01 -8.03801119e-01 -7.70939171e-01 1.40455687e+00
4.31266904e-01 9.87183690e-01 -3.77265841e-01 3.01947314e-02
1.03605139e+00 7.53486276e-01 -2.43027776e-01 4.26136255e-01
6.37648880e-01 -9.72103059e-01 -9.08373356e-01 5.15877493e-02
5.83939731e-01 -1.15301049e+00 4.34806824e-01 4.85368103e-01
-7.28485227e-01 -8.44449282e-01 -1.56713593e+00 2.13402897e-01
-8.87294233e-01 -2.09197938e-01 6.66659355e-01 1.45792508e+00
-1.25812805e+00 6.49086475e-01 -9.47469711e-01 -3.66620660e-01
2.23260894e-01 8.45480621e-01 -1.48289159e-01 -9.64468494e-02
-1.10160565e+00 4.83885318e-01 -3.02453816e-01 2.03467607e-01
-3.22747201e-01 -9.86736774e-01 -6.26653731e-01 1.92501485e-01
-7.30183184e-01 -7.49909639e-01 1.15171897e+00 -3.82810920e-01
-2.01426411e+00 -3.52910101e-01 9.09833908e-02 -1.38846278e-01
2.23485336e-01 -5.62963188e-01 -6.83620572e-01 -4.70657408e-01
1.56262845e-01 -1.51285976e-02 4.01167683e-02 -1.01903677e+00
-8.36781919e-01 -2.50416100e-01 -1.80581674e-01 -7.02398866e-02
1.48678318e-01 -4.98333961e-01 7.51624778e-02 -2.71113336e-01
6.97489858e-01 -7.65644372e-01 -4.18754429e-01 -4.72508252e-01
-1.17057152e-01 1.56150041e-02 7.11744308e-01 -5.36380589e-01
1.32730997e+00 -1.73570120e+00 -1.21990430e+00 7.32613921e-01
-1.28479078e-01 6.12526238e-02 9.29323509e-02 2.24111348e-01
2.83842713e-01 -1.02890246e-01 6.75755322e-01 1.10553242e-01
-1.66451782e-01 6.85300380e-02 3.40552926e-01 5.47323704e-01
-9.21587825e-01 2.56125212e-01 -1.22187972e+00 -1.62324488e-01
5.60548604e-01 9.88863170e-01 -5.23266673e-01 1.87451929e-01
9.64698136e-01 6.53653622e-01 -6.18565917e-01 3.41109812e-01
1.26015115e+00 3.04313779e-01 -4.25533265e-01 2.60672092e-01
-4.86069828e-01 5.29974759e-01 -1.43987906e+00 1.79191625e+00
-1.18224156e+00 8.36708128e-01 1.54670894e-01 -5.48323274e-01
9.36252415e-01 9.17723551e-02 7.61128664e-01 -7.87768364e-01
4.28054512e-01 2.41718501e-01 -2.78103296e-02 -6.61716104e-01
2.68538576e-02 7.06450939e-01 3.13181847e-01 5.63762859e-02
4.47211228e-02 1.94921136e-01 -7.91308999e-01 -2.57935435e-01
1.58334363e+00 2.88020492e-01 -1.06764898e-01 -3.25013816e-01
4.53873038e-01 -4.72170055e-01 5.11951923e-01 1.26460242e+00
-5.40541232e-01 2.89335996e-01 -4.21711385e-01 -5.26750207e-01
-1.20330669e-01 -1.26891220e+00 -2.83719182e-01 1.05951452e+00
3.30744296e-01 -6.66997433e-02 -7.86806405e-01 -3.82503033e-01
2.43206881e-02 3.42084795e-01 -3.71970564e-01 2.14361951e-01
-7.15632260e-01 -4.82873619e-01 8.39886010e-01 4.91430640e-01
8.85833144e-01 -4.10704345e-01 -1.83007926e-01 4.88318980e-01
1.15501791e-01 -8.38439763e-01 1.03575408e-01 4.94701445e-01
-5.80986857e-01 -8.76543880e-01 -3.34365070e-01 -1.16955125e+00
5.74462295e-01 7.30386138e-01 5.65082371e-01 1.11344188e-01
1.69690669e-01 5.89199007e-01 -1.35112703e-01 -9.43993926e-01
2.99361199e-01 3.32156807e-01 2.09261701e-01 -8.10465813e-01
4.87967253e-01 -1.09765351e+00 -1.06587434e+00 1.21138198e-03
1.00016028e-01 -6.01913869e-01 4.95002806e-01 2.88937092e-01
3.65990967e-01 4.27834123e-01 5.45262694e-01 -4.98713732e-01
5.00487804e-01 -8.12094629e-01 -6.51289046e-01 -1.33464500e-01
-6.73028231e-01 -4.78313386e-01 7.56968975e-01 -2.03232989e-01
-9.15414453e-01 1.89863518e-01 -7.45530009e-01 4.78317887e-01
-4.49561208e-01 1.52096689e-01 -2.00259194e-01 -9.10586238e-01
5.33132255e-01 -1.00484282e-01 -6.27177358e-01 -3.48464489e-01
4.47350405e-02 9.53872740e-01 4.63907063e-01 -2.81551987e-01
5.73879361e-01 -5.04140407e-02 3.24264348e-01 -9.38662112e-01
-6.02461040e-01 -4.60338086e-01 -4.44400609e-01 -4.52158272e-01
5.69861293e-01 -1.14288104e+00 -1.15213406e+00 4.74941999e-01
-1.02902317e+00 -2.72014886e-01 6.08176231e-01 1.10881698e+00
-1.46398678e-01 -1.06951073e-01 -5.05816340e-01 -4.74269032e-01
-1.98526025e-01 -1.06531894e+00 3.88946861e-01 1.11935902e+00
3.03458244e-01 -1.14202631e+00 4.08931702e-01 2.55096257e-01
1.00730765e+00 3.58444452e-01 1.78946003e-01 -5.80965281e-01
-5.74706018e-01 -5.98985791e-01 2.72142798e-01 9.75150801e-03
9.45120230e-02 1.54120903e-02 -1.14125156e+00 -1.05267130e-01
-1.17068300e-02 5.67335784e-01 3.90900783e-02 9.71846163e-01
1.43179643e+00 -2.17027161e-02 -6.37156665e-01 1.07429516e+00
1.61030793e+00 7.64805317e-01 6.23502791e-01 9.29127455e-01
7.37348735e-01 -2.45300204e-01 2.15578556e-01 1.90942377e-01
5.80980361e-01 3.39462280e-01 3.35467428e-01 1.60033464e-01
-8.44930187e-02 -5.22859395e-01 2.15688169e-01 1.17368710e+00
2.10120827e-02 -3.88726920e-01 -9.34252977e-01 9.52010751e-02
-1.46537435e+00 -8.17841470e-01 -7.83405960e-01 1.87591946e+00
2.79835343e-01 1.07160084e-01 -4.89965945e-01 2.12553874e-01
4.07579333e-01 -2.44380608e-01 1.90958291e-01 -5.63430607e-01
5.33644140e-01 6.88190043e-01 1.08753431e+00 1.00688922e+00
-1.19388211e+00 3.97251189e-01 6.61967230e+00 4.66753632e-01
-1.43977392e+00 5.68364449e-02 1.39282435e-01 3.23694378e-01
-1.77109502e-02 -4.81117547e-01 -6.47946119e-01 9.12074506e-01
1.27390420e+00 4.64946300e-01 3.25682312e-01 1.18336391e+00
2.61890739e-01 -4.66899961e-01 -4.85091537e-01 1.01430643e+00
-3.11554670e-01 -9.95937109e-01 -8.71448100e-01 -6.49627075e-02
1.01154864e+00 3.51068169e-01 3.52397263e-01 6.34130001e-01
6.03525043e-01 -8.94557655e-01 -1.30987121e-02 6.41828477e-01
6.60687014e-02 -9.06879842e-01 1.26864588e+00 1.86725080e-01
-1.23355484e+00 -2.92647302e-01 -4.26514685e-01 -4.47980940e-01
-4.61825848e-01 7.01346934e-01 -1.01569307e+00 4.84660327e-01
1.08465362e+00 1.75270557e-01 -6.47191882e-01 1.85257304e+00
-4.69569147e-01 8.37748110e-01 -9.42965806e-01 -3.23491216e-01
2.74002939e-01 3.54598910e-02 1.60837527e-02 1.16724694e+00
1.18113446e+00 -2.17329755e-01 -3.86436358e-02 2.08377868e-01
1.00051552e-01 1.73651837e-02 -1.03428602e+00 1.11281550e+00
1.27420270e+00 1.22303712e+00 -5.01699209e-01 2.59960324e-01
-2.25708291e-01 4.40043151e-01 -2.85107225e-01 4.04091567e-01
-9.50459301e-01 -1.02179825e+00 3.91096711e-01 1.82335123e-01
1.82925329e-01 -5.70531189e-01 -5.93374252e-01 -3.78524959e-01
-6.57899737e-01 -1.15548648e-01 -4.33201671e-01 -6.93098485e-01
-6.11100852e-01 4.29344177e-03 -4.44901258e-01 -9.78593528e-01
9.53475684e-02 -4.74681139e-01 -1.27359486e+00 7.33722866e-01
-1.29378653e+00 -8.61453295e-01 -7.31890559e-01 6.02801681e-01
1.40140042e-01 -2.60826517e-02 7.28618920e-01 1.14905930e+00
-7.54951894e-01 9.36316609e-01 1.01535690e+00 9.25525054e-02
7.53793955e-01 -1.53581321e+00 -6.60980865e-02 6.87468052e-01
-1.59716651e-01 1.14284539e+00 8.61344635e-01 -2.09592059e-01
-1.35239375e+00 -9.28704381e-01 9.58228827e-01 -5.25553465e-01
2.65369207e-01 -2.35440165e-01 -2.03985974e-01 7.55242348e-01
4.04884756e-01 1.92507833e-01 1.05146968e+00 2.72151858e-01
3.96473438e-01 -6.96440041e-01 -1.54839170e+00 2.78350651e-01
6.40003383e-01 -8.72665197e-02 -5.00469208e-02 2.54980084e-02
4.85465050e-01 -7.02240884e-01 -9.63006556e-01 1.77407831e-01
9.51969624e-01 -1.05194497e+00 1.12665069e+00 2.85271734e-01
-2.33187914e-01 -4.61362511e-01 -4.76292334e-02 -1.48458850e+00
-5.01890182e-01 -7.04656959e-01 -1.26345947e-01 1.25090837e+00
2.11976275e-01 -8.11526895e-01 8.88222218e-01 6.64127246e-02
-3.73041183e-01 -1.00178015e+00 -1.04206038e+00 -4.45034504e-01
-3.83333117e-02 -5.90172052e-01 1.09844923e+00 9.16687489e-01
-3.27826887e-01 5.71529120e-02 6.95011467e-02 8.43399942e-01
2.85675913e-01 -6.36418641e-01 9.61704016e-01 -1.07297981e+00
2.31934100e-01 4.13405858e-02 -5.05317152e-01 -1.31215203e+00
-2.96377778e-01 -2.56715894e-01 2.12493718e-01 -1.85701990e+00
-8.11111867e-01 -8.99298668e-01 -1.08642364e+00 -2.21409742e-02
1.12140737e-01 -8.17010179e-02 -4.05005991e-01 -7.17132092e-01
-6.66149974e-01 2.49458373e-01 1.14500642e+00 5.28533682e-02
-3.61406773e-01 7.48353302e-01 -3.24232340e-01 1.07201517e+00
1.12658143e+00 -6.45018518e-01 -4.46775287e-01 -7.55517602e-01
3.86416674e-01 1.26281492e-02 -1.27408773e-01 -1.90807784e+00
8.61950397e-01 1.11682408e-01 1.26827300e+00 -7.39986479e-01
-1.35669887e-01 -1.25973582e+00 3.99972759e-02 5.61453223e-01
4.53566350e-02 4.29714739e-01 2.64120817e-01 4.20050144e-01
2.56824911e-01 1.09657161e-01 4.57707793e-01 2.50590950e-01
-2.32611090e-01 6.27286211e-02 -7.30918050e-01 -9.31692302e-01
8.76129150e-01 -4.42928910e-01 -3.11532855e-01 -5.98345876e-01
-3.33900154e-01 6.56794235e-02 -1.25588596e-01 2.66900033e-01
5.25848866e-01 -1.55620170e+00 2.76065111e-01 5.20638347e-01
-6.59222484e-01 -1.05569981e-01 1.97498322e-01 7.15747416e-01
-1.29002857e+00 6.28436327e-01 -4.74216491e-01 -2.87261277e-01
-1.03014648e+00 -2.28812173e-01 7.23878443e-01 1.61980279e-02
6.89191595e-02 1.33250117e+00 -8.04961622e-01 -9.53416049e-01
7.78391063e-01 -6.89912200e-01 -3.48675191e-01 -6.77132308e-01
4.52328980e-01 1.08906949e+00 3.29783298e-02 -2.76460409e-01
-5.48816741e-01 6.77685499e-01 3.87885004e-01 3.24194223e-01
1.32679272e+00 -4.88899320e-01 7.72592947e-02 4.98326542e-03
1.10972309e+00 9.31734145e-01 -8.71155322e-01 3.95734340e-01
-6.74033612e-02 -5.61465859e-01 3.52980942e-01 -1.04562354e+00
-8.11612368e-01 6.68440759e-01 1.58414376e+00 2.61286110e-01
8.87158871e-01 -9.12238181e-01 1.08411610e+00 6.58388257e-01
6.27487421e-01 -9.10666168e-01 -2.32190087e-01 7.88606882e-01
-7.40067884e-02 -1.24793196e+00 -5.54213440e-03 -8.73067975e-02
7.44060397e-01 1.27539909e+00 8.75288248e-01 -3.67813826e-01
1.50861394e+00 7.51102149e-01 5.08942246e-01 1.62857175e-01
1.37849376e-01 2.88566023e-01 -1.46595225e-01 1.19087327e+00
9.55298305e-01 -4.37666960e-02 -1.77435443e-01 7.59888232e-01
-8.98459971e-01 8.18890035e-02 5.38132727e-01 1.21258342e+00
-8.09249043e-01 -1.04271317e+00 -7.20342517e-01 1.10336028e-01
-1.02456367e+00 3.59693587e-01 2.51282334e-01 6.17502153e-01
9.62067127e-01 1.53189981e+00 -2.11465552e-01 -7.84990132e-01
1.41583696e-01 -4.79269713e-01 6.84408009e-01 -2.59268701e-01
-7.80623794e-01 -3.22838008e-01 -4.05793667e-01 -6.72896981e-01
-1.22640118e-01 6.55515566e-02 -1.48600686e+00 -7.83878088e-01
-4.69979197e-01 2.08851889e-01 1.28089559e+00 8.11136484e-01
2.83194691e-01 1.09703696e+00 7.51358330e-01 -5.40921748e-01
2.03536093e-01 -8.14512908e-01 -4.23163891e-01 -6.25541806e-01
5.48085749e-01 -4.72849905e-01 -5.61534464e-02 -7.46442735e-01] | [6.415289878845215, 0.9167177677154541] |
886611f8-e26d-4b87-b6bf-09e4a9d3fa74 | weakly-supervised-one-stage-vision-and | 2007.15778 | null | https://arxiv.org/abs/2007.15778v1 | https://arxiv.org/pdf/2007.15778v1.pdf | Weakly supervised one-stage vision and language disease detection using large scale pneumonia and pneumothorax studies | Detecting clinically relevant objects in medical images is a challenge despite large datasets due to the lack of detailed labels. To address the label issue, we utilize the scene-level labels with a detection architecture that incorporates natural language information. We present a challenging new set of radiologist paired bounding box and natural language annotations on the publicly available MIMIC-CXR dataset especially focussed on pneumonia and pneumothorax. Along with the dataset, we present a joint vision language weakly supervised transformer layer-selected one-stage dual head detection architecture (LITERATI) alongside strong baseline comparisons with class activation mapping (CAM), gradient CAM, and relevant implementations on the NIH ChestXray-14 and MIMIC-CXR dataset. Borrowing from advances in vision language architectures, the LITERATI method demonstrates joint image and referring expression (objects localized in the image using natural language) input for detection that scales in a purely weakly supervised fashion. The architectural modifications address three obstacles -- implementing a supervised vision and language detection method in a weakly supervised fashion, incorporating clinical referring expression natural language information, and generating high fidelity detections with map probabilities. Nevertheless, the challenging clinical nature of the radiologist annotations including subtle references, multi-instance specifications, and relatively verbose underlying medical reports, ensures the vision language detection task at scale remains stimulating for future investigation. | ['Yuhong Wen', 'Leo K. Tam', 'Xiaosong Wang', 'Kevin Lu', 'Daguang Xu', 'Evrim Turkbey'] | 2020-07-31 | null | null | null | null | ['head-detection'] | ['computer-vision'] | [ 4.06985819e-01 4.31991935e-01 -2.26059496e-01 -4.40996647e-01
-1.61399877e+00 -6.29034936e-01 5.95071733e-01 2.66713679e-01
-5.62427163e-01 4.10317004e-01 4.81453955e-01 -5.90416312e-01
2.99097866e-01 -1.72134161e-01 -4.98657107e-01 -5.03331482e-01
1.47077098e-01 6.10270083e-01 3.03176254e-01 1.53319702e-01
-2.78192282e-01 4.76740688e-01 -8.19840908e-01 9.11391973e-01
-2.90070996e-02 7.62111068e-01 3.28824997e-01 1.25149286e+00
1.66275635e-01 1.71435666e+00 -2.36306086e-01 -7.28612095e-02
1.55517936e-01 -3.84377182e-01 -9.49254215e-01 1.39681727e-01
1.01477742e+00 -5.46341658e-01 -3.06046009e-01 7.02332735e-01
7.29888678e-01 -3.44425142e-01 7.87332237e-01 -7.16984391e-01
-8.78939867e-01 4.74175721e-01 -6.35396540e-01 7.33392656e-01
4.76555973e-01 8.13966274e-01 9.13663089e-01 -1.28650415e+00
6.68471456e-01 1.03817308e+00 9.08103228e-01 7.95087039e-01
-1.21926236e+00 -3.18318635e-01 4.48964275e-02 -2.63307393e-01
-1.47170985e+00 -2.80311465e-01 7.29960352e-02 -8.00338984e-01
1.35766304e+00 4.16690618e-01 3.48672241e-01 1.07979488e+00
1.86195802e-02 7.70031214e-01 1.32398951e+00 -4.87600386e-01
9.10502672e-02 5.08909523e-01 4.26822186e-01 1.36044145e+00
5.10611152e-03 2.37240493e-01 -3.02896142e-01 -3.80582541e-01
7.23140836e-01 4.61537838e-02 -5.37962735e-01 -2.56711155e-01
-1.61041117e+00 7.19614983e-01 7.22000122e-01 2.99756318e-01
-4.07296717e-01 4.98839766e-01 4.97180641e-01 -3.06349546e-01
1.55957460e-01 4.03215915e-01 -7.69446194e-02 2.28843898e-01
-1.07571733e+00 -1.91475004e-01 7.15336919e-01 1.10734904e+00
2.63770223e-01 -2.33891249e-01 -9.32811022e-01 4.61440533e-01
3.68987560e-01 5.11890769e-01 3.21815670e-01 -7.91175723e-01
2.60417581e-01 5.65192997e-01 -2.88693100e-01 -4.69632804e-01
-9.29045618e-01 -6.08560383e-01 -5.71196556e-01 2.31469497e-01
3.35856974e-01 6.82438463e-02 -1.34017420e+00 1.41732371e+00
1.18353784e-01 -4.10860330e-02 1.07844278e-01 1.06916511e+00
1.45082164e+00 2.61894792e-01 5.70470095e-01 5.74421324e-02
1.89014018e+00 -1.11117995e+00 -3.26955914e-01 -4.35474575e-01
9.71798956e-01 -6.55111730e-01 1.50722432e+00 -4.20570411e-02
-1.24126804e+00 -3.64023238e-01 -7.66414821e-01 -5.08320570e-01
-2.90894061e-01 5.48790932e-01 3.55102956e-01 4.35680956e-01
-1.57433224e+00 -3.21417302e-01 -9.32349443e-01 -6.99337006e-01
7.58580506e-01 3.61769944e-01 -2.57180423e-01 -1.78186134e-01
-7.54115343e-01 1.35802901e+00 2.59792060e-01 -1.91752344e-01
-1.24362695e+00 -1.05322862e+00 -7.92289495e-01 -1.58841804e-01
2.69448251e-01 -1.09759080e+00 1.69064224e+00 -4.44467425e-01
-7.48040378e-01 1.58480477e+00 7.81754851e-02 -5.94061077e-01
9.53785896e-01 1.65304735e-01 1.42234089e-02 6.92204118e-01
3.66172373e-01 1.23594105e+00 6.54237688e-01 -1.04952991e+00
-5.93963265e-01 -2.27683827e-01 2.46608667e-02 2.06097364e-01
3.79142491e-03 2.35120922e-01 -5.67179799e-01 -5.39827704e-01
-1.89918086e-01 -7.88000047e-01 -4.42284495e-01 5.63574374e-01
-4.20322359e-01 -1.40444770e-01 6.31045580e-01 -3.91303092e-01
8.56180608e-01 -2.19960308e+00 -4.16353613e-01 -9.67146978e-02
7.39399195e-01 -1.29555538e-01 1.24478266e-01 -2.95260847e-01
-1.71745747e-01 -1.00399449e-01 -4.08464074e-01 -3.57555270e-01
-3.10958743e-01 2.26041645e-01 -2.69139230e-01 6.25540316e-01
4.25274014e-01 1.32606947e+00 -1.03137994e+00 -1.17425311e+00
4.84055251e-01 4.42760617e-01 -6.90639138e-01 4.18775350e-01
1.43258134e-03 3.92400980e-01 -3.74300629e-01 9.70010340e-01
1.70700908e-01 -9.15678918e-01 -3.69874567e-01 -5.58005989e-01
-1.14846013e-01 -1.08793862e-02 -6.61317050e-01 1.71782887e+00
-6.02714658e-01 5.80938578e-01 3.69236141e-01 -4.27367419e-01
2.59456158e-01 5.24673462e-01 3.81003886e-01 -2.94363230e-01
-6.93884939e-02 1.57854229e-01 -6.80414811e-02 -8.23475957e-01
5.73456883e-02 -4.47311819e-01 -2.17293911e-02 3.51982385e-01
-2.59883478e-02 -6.40159786e-01 -7.64696524e-02 6.07637703e-01
1.56699193e+00 -5.31325340e-02 5.43906748e-01 -2.62408048e-01
5.45290887e-01 6.82497025e-01 -3.85132432e-02 1.48237908e+00
-7.13182747e-01 1.33065331e+00 5.54993600e-02 -2.99783349e-01
-9.13551033e-01 -1.32231033e+00 -5.73958635e-01 1.32994890e+00
-3.31955016e-01 -1.88762754e-01 -4.92149174e-01 -9.27502692e-01
-1.06995858e-01 5.97999990e-01 -8.12531471e-01 7.91976079e-02
-6.87375605e-01 -7.52018571e-01 1.03625298e+00 9.91105139e-01
1.62232801e-01 -1.05221105e+00 -1.13854313e+00 1.19724683e-01
7.29190037e-02 -1.45322633e+00 -5.95514655e-01 6.69080317e-01
-4.50049639e-01 -1.00020945e+00 -1.15590274e+00 -1.13002241e+00
7.60690808e-01 4.48693521e-02 1.38203430e+00 -3.21591124e-02
-1.23312509e+00 1.16825700e+00 5.85395023e-02 -3.88602734e-01
-7.26529837e-01 -3.04463983e-01 -2.85556644e-01 -6.60395980e-01
4.65628713e-01 2.16947213e-01 -8.65804970e-01 -2.09647179e-01
-6.38460338e-01 3.92071664e-01 7.45865583e-01 1.09485626e+00
6.95905268e-01 -9.05545890e-01 2.63203055e-01 -1.06484115e+00
3.35455179e-01 -3.11781168e-01 -2.14556143e-01 4.55718845e-01
-3.70127171e-01 -1.87988222e-01 3.16394925e-01 -1.89332470e-01
-9.38822091e-01 5.58007002e-01 -7.41557823e-03 -4.79698807e-01
-4.71352488e-01 3.80520523e-02 7.73295820e-01 -1.15225419e-01
1.23801315e+00 2.71746486e-01 -2.30127022e-01 -8.08827579e-03
7.88324416e-01 7.54391074e-01 1.15988994e+00 -4.66773510e-01
4.35882449e-01 9.12678361e-01 -3.76476012e-02 -5.98123193e-01
-1.15381551e+00 -9.70481157e-01 -7.02011108e-01 1.78828035e-02
1.34451246e+00 -1.26729608e+00 -4.10336643e-01 -3.75719219e-01
-1.30261350e+00 -3.29494178e-01 -7.17225373e-01 4.66039091e-01
-6.35559201e-01 2.87515700e-01 -1.06922460e+00 -7.02145576e-01
-7.76420593e-01 -1.36924219e+00 1.58070695e+00 -2.04538926e-01
-5.38907230e-01 -8.91730666e-01 -1.05463371e-01 4.60369080e-01
4.36339676e-01 2.98290759e-01 9.91934717e-01 -5.74813008e-01
-5.58324277e-01 -2.28518426e-01 -8.55427146e-01 8.73461962e-02
2.57321373e-02 -2.33358622e-01 -1.37566006e+00 -3.68859440e-01
8.01258385e-02 -7.89725363e-01 9.60334063e-01 5.74279428e-01
1.11965656e+00 9.56384540e-02 -3.83300960e-01 5.85773706e-01
1.17314339e+00 -1.30328417e-01 -2.71429062e-01 -1.62297394e-02
9.25564408e-01 3.86030078e-01 1.24343999e-01 1.38530388e-01
3.73240680e-01 3.08940083e-01 2.01023564e-01 -1.02180064e+00
-7.57850409e-01 -1.51084142e-03 9.43762287e-02 3.04510742e-01
3.64675701e-01 1.23303108e-01 -1.56406224e+00 5.46378911e-01
-1.57827461e+00 -4.72403497e-01 -2.44275406e-01 1.48711741e+00
1.07472599e+00 1.09512217e-01 -1.66605972e-02 -4.74611074e-01
4.44319516e-01 -5.54556847e-02 -4.85909194e-01 -1.19147226e-01
-1.09465919e-01 2.36086667e-01 7.30396390e-01 5.80574989e-01
-1.32573116e+00 6.75357461e-01 6.84803438e+00 3.89560312e-01
-1.07531333e+00 7.15075076e-01 8.64928544e-01 -3.16907883e-01
1.52567953e-01 -2.42065191e-01 -8.00370097e-01 -1.60228014e-01
6.33025765e-01 1.80231243e-01 -1.38280481e-01 9.99776065e-01
2.82903552e-01 -4.81287949e-02 -1.71708429e+00 1.20277727e+00
4.26343113e-01 -1.60150993e+00 1.95124000e-01 -3.27149540e-01
4.33283538e-01 6.51826084e-01 2.68703818e-01 3.79605204e-01
3.60643655e-01 -1.28568161e+00 6.91887379e-01 1.98460177e-01
1.32606673e+00 1.86866418e-01 5.70958138e-01 1.13293417e-01
-1.05218124e+00 -4.55243550e-02 6.76886067e-02 2.35005558e-01
1.01756886e-01 8.96247625e-02 -1.58032715e+00 -9.60813835e-02
7.04781711e-01 4.99851108e-01 -8.68360221e-01 7.69180894e-01
-2.15877622e-01 7.01200306e-01 -3.68351012e-01 1.27365604e-01
7.39988148e-01 7.42085934e-01 4.20200855e-01 2.11838579e+00
-3.19775492e-01 3.84338856e-01 7.28278100e-01 1.25641131e+00
-1.04265787e-01 4.42260839e-02 -6.18460774e-01 3.43533128e-01
-6.57079071e-02 1.44076025e+00 -8.15970361e-01 -7.41688371e-01
-7.76383936e-01 8.25200796e-01 1.05568975e-01 3.45926702e-01
-7.77737439e-01 3.65829706e-01 -2.93724895e-01 3.29538137e-01
-1.74168751e-01 2.24101767e-01 -6.51218295e-01 -9.40910757e-01
-2.60589689e-01 -6.95923865e-01 9.16567802e-01 -1.09544337e+00
-1.37545979e+00 6.82013988e-01 -7.12701306e-02 -1.07563436e+00
-2.32342094e-01 -9.03829873e-01 -4.27112728e-01 7.84259081e-01
-1.60713291e+00 -1.54910016e+00 -7.34109581e-01 7.29028761e-01
7.68146873e-01 6.87188804e-02 9.41318095e-01 2.76547641e-01
-3.58516604e-01 5.76764405e-01 -5.56171894e-01 2.92403549e-01
7.57754028e-01 -1.68915606e+00 7.18494058e-02 6.48875415e-01
1.31843761e-01 7.17010856e-01 3.28990906e-01 -3.65089297e-01
-1.13299072e+00 -1.34657919e+00 4.69112903e-01 -1.24346852e+00
7.01668978e-01 -2.91617930e-01 -7.96542883e-01 8.46927345e-01
2.41499752e-01 7.16112971e-01 5.22888958e-01 -4.56637412e-01
-3.06364924e-01 4.49096143e-01 -1.25421131e+00 5.08385897e-01
8.17429543e-01 -9.41607773e-01 -9.84348536e-01 1.03439927e+00
7.04793036e-01 -7.60359466e-01 -5.57292521e-01 6.71988070e-01
1.08020753e-01 -4.25024241e-01 1.11543131e+00 -6.33290946e-01
2.61557579e-01 -3.59247774e-01 -6.66584671e-02 -3.60598236e-01
-4.04531121e-01 -2.51994699e-01 4.81775887e-02 6.09740496e-01
5.72687149e-01 -1.81981146e-01 8.82748902e-01 7.80355036e-01
-1.68140143e-01 -7.49389052e-01 -9.47456658e-01 -2.05047399e-01
6.82992190e-02 -4.97575194e-01 -3.70061815e-01 8.84688735e-01
4.09280211e-02 4.09102947e-01 2.13308737e-01 2.12573647e-01
5.67424536e-01 -3.16352909e-03 3.75837684e-01 -4.56641257e-01
-7.56263196e-01 -4.10774767e-01 -3.01501542e-01 -8.07792902e-01
-2.44762480e-01 -1.42612708e+00 2.89461344e-01 -1.67297375e+00
8.78515601e-01 -3.49423885e-01 -4.34330791e-01 8.70944023e-01
-2.13521183e-01 6.36551499e-01 1.21560447e-01 4.25928116e-01
-7.08180308e-01 -2.34421670e-01 9.60325956e-01 -3.69086653e-01
1.55254588e-01 -2.79174805e-01 -5.38922012e-01 1.17113256e+00
1.66282117e-01 -5.20938277e-01 -2.04657435e-01 -4.48068142e-01
-2.42046580e-01 1.41757905e-01 1.01902926e+00 -9.82657015e-01
4.87829626e-01 2.99460202e-01 5.23305595e-01 -6.94709778e-01
3.24480921e-01 -6.72695756e-01 -6.25017524e-01 8.20850611e-01
-8.33461881e-01 7.23461807e-02 3.30317616e-01 6.37266219e-01
7.81998336e-02 7.21892789e-02 1.20764351e+00 -6.92349672e-01
-7.06567585e-01 4.85455431e-03 -4.60533142e-01 3.74116689e-01
9.74125504e-01 -2.10464373e-01 -4.40157086e-01 -1.30068555e-01
-1.15674031e+00 3.70912284e-01 2.76385069e-01 8.53083134e-02
7.62890995e-01 -7.28626788e-01 -1.04491150e+00 -1.97115485e-02
7.56859779e-01 3.21884513e-01 5.92397787e-02 1.18048608e+00
-7.99588084e-01 6.78307652e-01 3.41583312e-01 -1.31182957e+00
-1.24851358e+00 5.22660017e-01 7.87475586e-01 -3.42840314e-01
-1.08006823e+00 1.14593923e+00 9.11441207e-01 -2.57429928e-01
4.49102640e-01 -9.42174256e-01 1.35441318e-01 -3.37365299e-01
5.19584298e-01 -1.74128279e-01 2.70465374e-01 -5.04296362e-01
-7.86389947e-01 4.44169044e-01 -4.50406343e-01 -6.40889406e-02
8.81774187e-01 -6.59386395e-03 2.70927757e-01 2.53211677e-01
1.05633473e+00 -2.50894964e-01 -9.54761386e-01 -3.43986034e-01
9.79846120e-02 1.22527905e-01 3.28784257e-01 -1.23109138e+00
-5.62907755e-01 1.10064375e+00 1.18135321e+00 -3.22712183e-01
8.56401503e-01 7.14445114e-01 2.31937811e-01 4.86884266e-01
7.42542669e-02 -7.43683994e-01 1.68860465e-01 1.60218418e-01
8.46630692e-01 -1.64430857e+00 4.17016968e-02 -2.99658120e-01
-8.24711859e-01 8.83999825e-01 8.31564009e-01 1.21790268e-01
4.35385555e-01 8.49419594e-01 5.53436577e-01 -6.90727949e-01
-6.99900568e-01 -3.98485482e-01 2.85149723e-01 5.46672583e-01
8.65087450e-01 5.42887039e-02 9.96683612e-02 1.46715865e-01
2.01393589e-01 6.21317141e-02 5.09002328e-01 1.08130741e+00
-3.11082155e-01 -2.19127908e-01 -5.21185994e-01 5.46707988e-01
-6.76641464e-01 -5.59553802e-01 -3.98388267e-01 8.39465737e-01
4.91652563e-02 7.32637346e-01 -7.92707205e-02 3.64301175e-01
4.42284286e-01 -7.47719109e-02 2.78388053e-01 -1.41477370e+00
-9.83898520e-01 1.36700898e-01 3.17054391e-02 -4.83026028e-01
-2.73156136e-01 -3.11656922e-01 -1.73076880e+00 6.21892869e-01
-1.55655488e-01 -2.86073506e-01 2.56866217e-01 8.01046669e-01
4.92607541e-02 8.22731495e-01 5.18912822e-03 -5.91949463e-01
-7.52870023e-01 -7.21877575e-01 -8.43724087e-02 5.97194254e-01
7.73716569e-01 -2.25951448e-01 -2.84561455e-01 2.65722603e-01] | [15.026883125305176, -1.8292726278305054] |
06482b8b-adaa-448f-b7f4-73bd5f4a9b57 | advancing-full-text-search-lemmatization | 2305.10848 | null | https://arxiv.org/abs/2305.10848v1 | https://arxiv.org/pdf/2305.10848v1.pdf | Advancing Full-Text Search Lemmatization Techniques with Paradigm Retrieval from OpenCorpora | In this paper, we unveil a groundbreaking method to amplify full-text search lemmatization, utilizing the OpenCorpora dataset and a bespoke paradigm retrieval algorithm. Our primary aim is to streamline the extraction of a word's primary form or lemma - a crucial factor in full-text search. Additionally, we propose a compact dictionary storage strategy, significantly boosting the speed and precision of lemma retrieval. | ['Dmitriy Kalugin-Balashov'] | 2023-05-18 | null | null | null | null | ['lemmatization'] | ['natural-language-processing'] | [ 2.17943013e-01 -2.14404359e-01 -7.39534855e-01 4.66124564e-01
-8.57154787e-01 -1.14715683e+00 7.30132818e-01 4.99550313e-01
-7.62161791e-01 7.39429593e-01 4.82578635e-01 -1.14721262e+00
-1.38528079e-01 -7.87369490e-01 -3.86521757e-01 -2.56873310e-01
5.32220781e-01 3.37380856e-01 4.77325059e-02 -4.98232961e-01
4.96380955e-01 4.86613482e-01 -1.30291665e+00 2.51315117e-01
8.42143893e-01 8.63357008e-01 1.50990888e-01 4.64964807e-01
-4.98915493e-01 3.64606559e-01 -4.90112931e-01 -9.50421393e-01
2.35253915e-01 -2.18014568e-01 -1.07961786e+00 -7.02244163e-01
3.65733832e-01 -8.09741244e-02 -8.05317163e-01 1.15166509e+00
4.99980092e-01 -2.76561201e-01 4.82636899e-01 -7.21220315e-01
-4.00303900e-01 1.30749023e+00 -1.71310633e-01 6.97759092e-01
7.04756916e-01 -2.76424021e-01 1.71459436e+00 -9.45887268e-01
9.53361690e-01 8.31748366e-01 3.96409661e-01 2.06709057e-01
-7.58299828e-01 -6.39199197e-01 -2.32258111e-01 5.47287285e-01
-1.67504704e+00 -4.65243727e-01 8.66026580e-01 -6.36568516e-02
1.33414042e+00 4.02258873e-01 9.44542289e-01 1.10372186e+00
3.56752932e-01 9.82142627e-01 8.87628675e-01 -9.50639904e-01
-2.11002961e-01 -6.76387846e-02 2.35013455e-01 6.91593289e-01
7.53895462e-01 7.28989094e-02 -9.30260420e-01 -3.60804647e-01
4.01436776e-01 -3.93360019e-01 -3.71468544e-01 1.44217148e-01
-1.20987880e+00 8.04106236e-01 -2.37520725e-01 7.21741021e-01
-2.24941254e-01 -4.53371294e-02 8.63598347e-01 4.93434072e-01
2.88165659e-01 8.54543924e-01 -5.39122105e-01 -4.69733238e-01
-1.14517772e+00 2.00766087e-01 8.33454907e-01 8.35148156e-01
3.95737618e-01 -4.51216161e-01 1.29455686e-01 3.56876016e-01
1.69593992e-03 9.77427006e-01 5.97303629e-01 -3.71815473e-01
6.41497195e-01 4.54799652e-01 -2.82756597e-01 -1.02812505e+00
-9.00637507e-02 -3.19442540e-01 -5.54937601e-01 -8.88651788e-01
1.46148980e-01 2.33614981e-01 -5.28558671e-01 1.27969468e+00
1.35795608e-01 -4.48393196e-01 5.23262396e-02 5.15927196e-01
6.87178016e-01 5.42479873e-01 -1.84264392e-01 -4.41454053e-01
1.67591035e+00 -7.18556821e-01 -1.04293537e+00 -7.32246414e-02
8.77231658e-01 -1.04440391e+00 1.18381345e+00 7.19364226e-01
-1.13512731e+00 -5.31867445e-02 -1.26431835e+00 -3.48818213e-01
-7.92890191e-01 -1.02622025e-01 8.71014118e-01 7.54481375e-01
-7.19395101e-01 4.61893409e-01 -4.53092456e-01 -1.23628661e-01
2.17327550e-01 -6.29613996e-02 -4.37611997e-01 3.52056551e-04
-1.62011492e+00 9.75188494e-01 6.45932496e-01 -3.93841147e-01
-2.28696764e-01 -6.95047140e-01 -4.97989774e-01 -4.77827452e-02
4.92649853e-01 -6.53776765e-01 9.17462230e-01 -2.13894039e-01
-1.24518192e+00 1.16990352e+00 -1.37593433e-01 -4.79200959e-01
-1.08274318e-01 -3.49530548e-01 -6.50394678e-01 5.55552125e-01
-3.31015550e-02 2.63373721e-02 1.14308619e+00 -5.25342464e-01
-6.38401091e-01 -3.09309810e-01 -2.46284217e-01 -3.24323811e-02
-8.28174293e-01 3.21132362e-01 -4.67655391e-01 -1.16669881e+00
-5.86448759e-02 -6.03570461e-01 2.14511082e-01 -4.78406787e-01
-2.58350968e-01 -6.57146037e-01 6.24535918e-01 -8.41791809e-01
2.23653722e+00 -2.24832678e+00 4.75170165e-01 4.68984097e-01
3.20392072e-01 6.42432690e-01 -5.34144603e-02 9.04183030e-01
2.50724684e-02 4.99337554e-01 2.59024799e-01 -3.67314704e-02
1.54022455e-01 4.19302136e-01 -1.18390048e+00 4.48604047e-01
-1.99621767e-01 1.43001652e+00 -8.07380438e-01 -8.81511927e-01
8.34303349e-03 -3.18611972e-03 -2.97185451e-01 -7.70550296e-02
-4.36254069e-02 -4.47677433e-01 -4.19277966e-01 9.30580318e-01
2.84164548e-01 9.85293463e-02 1.18284829e-01 -2.15660319e-01
-2.28843674e-01 9.04577315e-01 -9.01614726e-01 2.02279520e+00
-5.37157893e-01 6.25697017e-01 2.75993533e-02 -6.97035849e-01
4.97017175e-01 2.19305411e-01 3.46748352e-01 -9.99083698e-01
5.78113616e-01 5.53381801e-01 -4.61270690e-01 -1.34844705e-01
1.04753065e+00 -4.82577495e-02 -4.89604414e-01 7.05956340e-01
1.43270150e-01 -4.46166962e-01 4.24279273e-01 6.99884892e-01
1.19204426e+00 -4.84862804e-01 4.88255113e-01 -4.39632624e-01
5.23063004e-01 1.56220049e-01 -5.48327453e-02 6.96649194e-01
1.89030111e-01 8.35585073e-02 3.68076354e-01 -1.88225225e-01
-1.17896914e+00 -9.77800727e-01 -3.89832109e-01 9.78753269e-01
4.34510559e-02 -1.21636021e+00 -4.65244591e-01 -9.65521514e-01
1.28527537e-01 5.35795867e-01 -4.35593188e-01 -3.54482532e-01
-8.30802321e-01 -2.77199894e-01 1.21478939e+00 1.83154091e-01
9.00805965e-02 -8.99309516e-01 -5.38119137e-01 1.89296082e-02
-4.62159067e-01 -1.08243120e+00 -5.58607042e-01 6.73741817e-01
-6.54986918e-01 -1.08079052e+00 -4.27639574e-01 -8.68575335e-01
2.75991321e-01 1.70362636e-01 1.05916989e+00 3.23761374e-01
-4.41183418e-01 6.21959753e-02 -7.26050138e-01 -1.13499671e-01
-2.56933749e-01 6.21482909e-01 1.10260405e-01 -7.09450841e-01
4.81286407e-01 -5.48054755e-01 -2.16480017e-01 -2.58292198e-01
-1.24330473e+00 -3.61176670e-01 5.69738209e-01 7.86797345e-01
4.25968349e-01 1.74656436e-01 2.14765266e-01 -6.87494993e-01
1.07690215e+00 -1.84772223e-01 -6.44377232e-01 4.81689364e-01
-1.19065940e+00 4.01631236e-01 6.88724875e-01 -6.45599127e-01
-4.81727779e-01 -4.44281906e-01 -3.15479010e-01 -3.45894456e-01
4.95653927e-01 8.52362633e-01 -1.00185812e-01 -8.87518376e-03
6.74296260e-01 5.24454951e-01 -3.36461872e-01 -6.47746146e-01
1.09222746e+00 7.39331126e-01 9.16121006e-01 -8.04285705e-01
1.08411884e+00 3.84654880e-01 -7.90148750e-02 -6.64867997e-01
-5.91506898e-01 -8.91858339e-01 -5.29647827e-01 1.83436990e-01
3.22450429e-01 -7.39163101e-01 -3.62435699e-01 2.66257942e-01
-1.13733828e+00 -6.08728938e-02 -5.45197785e-01 4.94355895e-02
-4.84417230e-02 1.02035999e+00 -7.41040468e-01 -3.04120213e-01
-1.04294133e+00 -4.64722216e-01 1.12641919e+00 -8.22706670e-02
-4.67047542e-01 -8.83082747e-01 4.62950796e-01 1.23759739e-01
1.91172063e-01 -6.15434408e-01 1.18073404e+00 -8.67778718e-01
-3.07251010e-02 -2.69482046e-01 -2.69803196e-01 -1.06335513e-03
-9.81835574e-02 -1.28096193e-01 -3.62379700e-01 -2.48013064e-01
-2.46875256e-01 -3.65035087e-01 8.07311594e-01 -4.43095237e-01
8.25327396e-01 -4.98779476e-01 -4.40324545e-01 8.56394827e-01
1.27209520e+00 -8.73114169e-02 6.26428246e-01 7.04377711e-01
2.98857361e-01 9.48081613e-02 6.66559398e-01 4.63718623e-01
2.58611366e-02 3.61194372e-01 -1.55058846e-01 2.01513067e-01
-4.86202121e-01 -8.63526642e-01 1.80248991e-01 1.33158290e+00
4.60736245e-01 -2.98209071e-01 -9.57302332e-01 7.24881947e-01
-1.36664975e+00 -7.70694613e-01 1.21743707e-02 1.63220739e+00
1.38347876e+00 2.35967398e-01 -1.97304368e-01 6.23857498e-01
3.37535977e-01 2.20575139e-01 5.22206388e-02 -5.13595521e-01
-5.14965892e-01 1.20684421e+00 7.67114043e-01 5.38532615e-01
-7.32712448e-01 1.63008273e+00 7.24776745e+00 1.52993143e+00
-1.26696193e+00 3.92186716e-02 -2.60256857e-01 1.09202221e-01
-7.92206764e-01 1.63542360e-01 -1.00622797e+00 3.20199102e-01
7.45315373e-01 -6.92710459e-01 6.33101821e-01 7.62434483e-01
-4.45579171e-01 -1.18917137e-01 -7.43245900e-01 1.14629877e+00
2.85790861e-01 -1.42945707e+00 4.83806312e-01 3.73280011e-02
2.08877623e-01 -1.29303858e-01 1.51596904e-01 2.82699704e-01
6.81672711e-03 -8.71393323e-01 8.79375935e-01 -7.75968051e-03
1.25351715e+00 -8.48714292e-01 5.66322744e-01 7.97862411e-02
-1.13891983e+00 8.51075649e-02 -7.80189559e-02 -9.12723765e-02
2.32717365e-01 7.96958864e-01 -6.52751982e-01 5.26859522e-01
4.08604503e-01 3.87812555e-01 -6.40253901e-01 5.30236304e-01
-7.14157104e-01 6.95863366e-01 -6.45888507e-01 -4.28340644e-01
1.93493590e-01 3.63494009e-02 7.75361180e-01 1.58663833e+00
1.51128963e-01 4.44570392e-01 -1.63980097e-01 2.89359957e-01
-2.34126732e-01 5.87011695e-01 -6.19962931e-01 -7.72532403e-01
7.35536218e-01 1.18860996e+00 -8.38394523e-01 -6.36339664e-01
7.91535899e-02 1.11605060e+00 2.78712481e-01 8.71162638e-02
-6.31934881e-01 -9.94114816e-01 4.64105636e-01 -4.91197854e-02
6.53522611e-01 -4.43860024e-01 -4.40914124e-01 -1.56826174e+00
1.75433591e-01 -1.26615536e+00 6.59322262e-01 -4.93192464e-01
-1.10575676e+00 4.62232739e-01 -1.03752203e-01 -6.36968434e-01
-1.45880654e-01 -5.83355010e-01 -1.00676112e-01 6.00117207e-01
-1.46602523e+00 -1.10201085e+00 4.35901940e-01 9.64099944e-01
2.40888372e-01 -1.38072073e-01 8.71917367e-01 6.31563365e-01
-4.15891051e-01 9.30657387e-01 1.08175911e-01 1.38628930e-01
6.99981093e-01 -7.01021373e-01 3.32424909e-01 1.21009278e+00
8.22417855e-01 1.11453295e+00 6.77227914e-01 -9.71870005e-01
-2.15337515e+00 -3.35486650e-01 1.59960270e+00 -3.76568377e-01
1.60273659e+00 -4.33927774e-01 -4.63986218e-01 2.92594135e-01
3.13078970e-01 -4.50496197e-01 5.05053282e-01 3.27984989e-01
-8.47360730e-01 2.05028299e-02 -7.87753582e-01 8.76696169e-01
1.21400058e+00 -1.26618993e+00 -1.26188970e+00 1.11216940e-01
9.06970501e-01 -3.14770371e-01 -6.20959997e-01 4.07801926e-01
6.93746746e-01 6.17807209e-02 9.48264718e-01 -6.88947439e-01
2.88899928e-01 2.07863167e-01 -2.76836365e-01 -1.02615833e+00
-1.95141405e-01 -1.16771853e+00 -6.03329599e-01 1.11141777e+00
4.79710281e-01 -4.94067580e-01 6.01678789e-01 9.63943601e-02
5.67034036e-02 -6.75707459e-01 -1.24218452e+00 -7.19449639e-01
3.38711888e-01 -7.86485195e-01 4.61078674e-01 1.02455723e+00
1.01961601e+00 7.46299684e-01 -4.79254983e-02 -3.43207538e-01
2.91834325e-01 1.67701274e-01 4.71142024e-01 -7.69898951e-01
-7.59705529e-02 -7.29877889e-01 -1.54822946e-01 -1.33555651e+00
3.58369529e-01 -1.23209727e+00 -4.15065467e-01 -9.46452498e-01
2.05040455e-01 -4.09196436e-01 -2.48782724e-01 4.78856415e-01
-2.32975613e-02 2.94853598e-01 6.96306005e-02 1.68277547e-01
-5.70551932e-01 3.04039389e-01 7.46744275e-01 -1.46229014e-01
2.33048797e-02 -4.34632391e-01 -8.55372488e-01 2.09042341e-01
5.52972615e-01 -6.59976482e-01 -2.58216530e-01 -3.04495394e-01
1.12920630e+00 -8.75370651e-02 -3.13493311e-02 -3.95947337e-01
5.32455981e-01 -2.47045215e-02 -3.02066088e-01 -1.00447547e+00
-5.13956361e-02 -6.96051896e-01 -2.58421391e-01 4.90290940e-01
-1.81909114e-01 6.02216363e-01 3.67062539e-01 2.36901909e-01
-1.18798956e-01 -4.09893781e-01 2.40189865e-01 8.19404796e-02
-3.89949828e-01 1.12208664e-01 -5.91742158e-01 5.44581592e-01
4.37548757e-01 2.24365667e-01 -2.62639254e-01 7.43359402e-02
-2.13895932e-01 -1.69092566e-01 5.65457642e-01 3.72153491e-01
6.81641519e-01 -9.84835923e-01 -4.23737109e-01 3.02612901e-01
-5.77901751e-02 -5.95267355e-01 -6.19995534e-01 5.40301561e-01
-6.11284018e-01 9.90856767e-01 1.42419830e-01 5.16961478e-02
-1.44145787e+00 1.07468462e+00 -3.58787477e-01 -6.39386475e-01
-7.83434987e-01 9.65303242e-01 -7.21280158e-01 2.44741201e-01
2.40672097e-01 -3.76469195e-01 1.48642346e-01 3.46673459e-01
6.47956192e-01 2.72925030e-02 4.87564743e-01 -2.96339631e-01
-3.75170469e-01 3.84012252e-01 -1.53598890e-01 -5.29387832e-01
9.89695072e-01 -1.99099615e-01 -7.28013158e-01 1.55322105e-01
1.25684738e+00 8.07116270e-01 2.99111724e-01 -4.24786747e-01
1.11549981e-01 -3.89763921e-01 3.86700779e-01 -7.57067263e-01
-6.10845983e-01 4.20151621e-01 -1.88668184e-02 1.27253920e-01
1.09661651e+00 2.66635805e-01 1.67669010e+00 8.57827485e-01
6.13785207e-01 -1.46405220e+00 -2.30120663e-02 9.31479096e-01
2.78673023e-01 -6.53959036e-01 4.09020960e-01 -2.29642019e-01
-1.62906628e-02 1.18640745e+00 -1.76534444e-01 1.75273463e-01
7.00100780e-01 6.35520756e-01 -7.20396563e-02 -4.68130440e-01
-6.17858827e-01 -3.47578049e-01 2.70706654e-01 2.17363417e-01
-3.29584815e-02 -2.20364645e-01 -1.06505501e+00 7.04964101e-01
-6.96695328e-01 -3.03922862e-01 -2.81535815e-02 1.05247903e+00
-3.17675769e-01 -1.50448191e+00 -3.53822410e-01 2.40479261e-01
-9.36083734e-01 -9.40456867e-01 -8.11059535e-01 5.63665748e-01
-5.27561232e-02 7.90347159e-01 -2.36934751e-01 -5.35803080e-01
1.80367809e-02 3.06779474e-01 7.29337335e-01 -2.41600916e-01
-6.71960294e-01 -1.72901884e-01 2.25365788e-01 -5.14581263e-01
1.61129445e-01 -5.30150890e-01 -9.36118662e-01 -8.35897326e-01
-8.50161135e-01 6.74917817e-01 7.11259723e-01 9.26097274e-01
2.74655432e-01 -2.13019013e-01 5.12431502e-01 -1.34496450e-01
-3.96146148e-01 -7.15361953e-01 -4.19375628e-01 9.90282670e-02
-1.45765273e-02 -3.08809131e-01 -7.37999499e-01 -1.81771070e-01] | [11.578847885131836, 9.078139305114746] |
9af7fe5a-9f7d-45ea-a8c2-1d24868dbb2d | structcoder-structure-aware-transformer-for | 2206.05239 | null | https://arxiv.org/abs/2206.05239v2 | https://arxiv.org/pdf/2206.05239v2.pdf | StructCoder: Structure-Aware Transformer for Code Generation | There has been a recent surge of interest in automating software engineering tasks using deep learning. This paper addresses the problem of code generation where the goal is to generate target code given source code in a different language or a natural language description. Most of the state-of-the-art deep learning models for code generation use training strategies primarily designed for natural language. However, understanding and generating code requires a more rigorous comprehension of the code syntax and semantics. With this motivation, we develop an encoder-decoder Transformer model where both the encoder and decoder are explicitly trained to recognize the syntax and data flow in the source and target codes, respectively. We not only make the encoder structure-aware by leveraging the source code's syntax tree and data flow graph, but we also support the decoder in preserving the syntax and data flow of the target code by introducing two novel auxiliary tasks: AST (Abstract Syntax Tree) paths prediction and data flow prediction. To the best of our knowledge, this is the first work to introduce a structure-aware Transformer decoder that models both syntax and data flow to enhance the quality of generated code. The proposed StructCoder model achieves state-of-the-art performance on code translation and text-to-code generation tasks in the CodeXGLUE benchmark, and improves over baselines of similar size on the APPS code generation benchmark. Our code is publicly available at https://github.com/reddy-lab-code-research/StructCoder/. | ['Chandan K. Reddy', 'Ming Zhu', 'Sindhu Tipirneni'] | 2022-06-10 | null | null | null | null | ['code-translation', 'text-to-code-generation'] | ['computer-code', 'computer-code'] | [ 2.78058559e-01 2.21262738e-01 -4.11461174e-01 -3.21794122e-01
-7.65232563e-01 -7.73000598e-01 3.86405140e-01 1.84894934e-01
4.15198177e-01 1.98821098e-01 3.05441618e-01 -9.76207018e-01
6.35921061e-01 -8.90672863e-01 -1.07740903e+00 1.96363628e-01
2.26631269e-01 -5.99373020e-02 1.35964394e-01 -2.55461335e-01
4.00834471e-01 -5.64766005e-02 -1.43332386e+00 6.76760316e-01
1.16864026e+00 4.97224420e-01 2.98952758e-01 8.31522942e-01
-4.51994032e-01 1.13132882e+00 -3.93287271e-01 -6.63924575e-01
9.72550586e-02 -6.83845997e-01 -1.14081478e+00 -2.87790239e-01
2.19461694e-01 -3.40073228e-01 -2.94328898e-01 9.78836775e-01
2.27182865e-01 -6.14087284e-01 2.71337122e-01 -1.31641293e+00
-9.71840680e-01 1.21203232e+00 -5.26429772e-01 -1.27253070e-01
3.33090901e-01 3.51026922e-01 1.12934875e+00 -8.43534172e-01
5.02575934e-01 8.41041207e-01 4.87173289e-01 8.62017393e-01
-1.20405006e+00 -6.21702790e-01 -9.45395157e-02 -1.44711241e-01
-1.08286786e+00 -4.60996836e-01 6.62762165e-01 -9.72429693e-01
1.31655681e+00 -4.85944301e-02 3.79072547e-01 1.04851067e+00
5.10886371e-01 5.58357954e-01 5.57553887e-01 -4.78076875e-01
7.89872855e-02 1.72767520e-01 2.27551654e-01 1.05507123e+00
1.79464087e-01 2.08179146e-01 -2.73231030e-01 -1.83910698e-01
3.88822466e-01 -8.65890905e-02 -1.94003209e-01 -5.27549863e-01
-1.18978918e+00 8.38207781e-01 2.95364052e-01 1.68481931e-01
1.48625016e-01 6.49954140e-01 7.62305915e-01 4.12753075e-01
1.01771668e-01 4.94384557e-01 -7.18879104e-01 -6.35816395e-01
-1.03979135e+00 3.07832122e-01 1.02064466e+00 1.58710361e+00
8.70209157e-01 2.82067060e-01 -2.06711724e-01 5.15655935e-01
4.13660765e-01 4.08894390e-01 6.44874871e-01 -6.52648926e-01
9.98738110e-01 9.75398242e-01 -2.65680313e-01 -6.72892332e-01
1.60714775e-01 -3.28408003e-01 -4.67202485e-01 3.34670126e-01
1.33821651e-01 -2.04846486e-01 -6.70685172e-01 1.59321058e+00
-1.03830710e-01 -1.44502565e-01 2.55038112e-01 3.08825523e-01
7.63821423e-01 7.03483701e-01 -3.37953925e-01 2.63756186e-01
1.12984228e+00 -1.27198207e+00 -3.05781603e-01 -4.42282617e-01
1.26245940e+00 -9.05509472e-01 1.23402882e+00 -6.34285435e-02
-1.12930858e+00 -6.95483208e-01 -1.12332630e+00 -4.56679493e-01
-1.79929927e-01 5.56001961e-01 4.72873837e-01 5.59028506e-01
-1.13610113e+00 2.92360961e-01 -8.89802814e-01 -1.95504025e-01
5.08968651e-01 1.24452405e-01 -1.63103253e-01 4.53286655e-02
-7.29418457e-01 5.02800703e-01 4.45805699e-01 -3.57725233e-01
-1.24201787e+00 -1.12916255e+00 -1.19628084e+00 2.77855963e-01
1.62037745e-01 -6.56205416e-01 1.78399634e+00 -1.12772632e+00
-1.34554386e+00 7.96390116e-01 -2.58118391e-01 -5.00166416e-01
2.12838069e-01 -2.07660660e-01 -2.92485684e-01 -4.79198843e-01
2.24210337e-01 4.39387977e-01 5.97329497e-01 -1.16005170e+00
-5.62224030e-01 9.50618368e-03 2.24919066e-01 -5.05477130e-01
-2.36848459e-01 2.41772875e-01 -2.39348307e-01 -7.60505497e-01
-6.80430233e-01 -9.80616510e-01 1.33588850e-01 2.34391280e-02
-5.67331493e-01 9.01126638e-02 8.09134841e-01 -9.34980810e-01
1.55654323e+00 -2.25194097e+00 2.10622981e-01 -1.34711817e-01
3.66880029e-01 3.57047528e-01 -3.26753318e-01 6.14007950e-01
-3.27620625e-01 4.75827068e-01 -6.64720058e-01 -1.51987389e-01
2.30845705e-01 -6.42706752e-02 -8.01083028e-01 1.32408915e-02
4.14044470e-01 1.16211295e+00 -9.70321059e-01 -2.60607690e-01
-1.71106055e-01 3.04948449e-01 -1.12557769e+00 5.25107145e-01
-6.11007035e-01 1.92921489e-01 -4.00205851e-01 4.59843725e-01
3.91054451e-01 -3.20606172e-01 1.25734091e-01 1.20627366e-01
-3.28171074e-01 8.78397465e-01 -5.78605533e-01 2.02856946e+00
-1.11414635e+00 9.09557045e-01 -2.68579841e-01 -6.42478764e-01
9.91779506e-01 3.73384953e-01 1.19793667e-02 -5.75100541e-01
-1.86687097e-01 4.85282511e-01 2.48521611e-01 -5.11522532e-01
4.85596806e-01 2.64405638e-01 -4.13253158e-01 8.79462004e-01
-7.62756988e-02 -3.21117699e-01 3.30128759e-01 3.66316170e-01
1.34185374e+00 6.92808807e-01 3.20475817e-01 -1.29790813e-01
5.57073116e-01 9.52499509e-02 3.28942925e-01 4.69293028e-01
5.11937439e-01 5.99366486e-01 9.19277966e-01 -2.83623427e-01
-1.28376389e+00 -9.79047060e-01 2.39615589e-01 1.07053673e+00
-3.09865803e-01 -1.07067513e+00 -1.00755310e+00 -1.05296385e+00
-2.62042046e-01 1.11785793e+00 -6.09714091e-01 -5.22080421e-01
-9.08332944e-01 -2.13133633e-01 9.42020595e-01 7.33450949e-01
3.62020731e-01 -1.01886141e+00 -7.77340353e-01 1.45597309e-01
-1.85537040e-01 -9.23151314e-01 -9.38127100e-01 -9.75681320e-02
-7.31243432e-01 -1.06290972e+00 -3.25896978e-01 -8.51159573e-01
7.91015923e-01 -2.05095068e-01 1.45907176e+00 5.47470331e-01
-1.48351803e-01 -2.65502930e-02 -5.19992054e-01 -2.39430279e-01
-1.42588842e+00 5.35445750e-01 -7.55723536e-01 -3.53947639e-01
7.39549324e-02 -5.78372061e-01 -2.65005708e-01 5.61923049e-02
-1.10143399e+00 6.74232483e-01 5.03571689e-01 6.21280491e-01
1.91467404e-01 -2.87041366e-01 2.07735866e-01 -1.01864827e+00
4.75965410e-01 -5.05430400e-01 -9.21884537e-01 1.53537020e-01
-6.33556843e-01 5.15779793e-01 1.12721455e+00 -1.55956373e-02
-1.14497674e+00 5.45577705e-02 -3.56267959e-01 3.20810871e-03
1.65862255e-02 7.07121372e-01 -2.40993395e-01 2.05938935e-01
7.88820386e-01 3.56159866e-01 -1.42601639e-01 -4.83258039e-01
4.51443911e-01 8.19406688e-01 3.94427091e-01 -8.54773462e-01
1.03888416e+00 1.61639489e-02 -2.38310039e-01 -4.93422896e-02
-4.45941269e-01 1.16216257e-01 -5.99395454e-01 2.97307372e-01
7.15419352e-01 -8.22786748e-01 -1.81195468e-01 5.23742557e-01
-1.63706183e+00 -6.78364992e-01 -4.24801379e-01 -2.04614684e-01
-6.09801114e-01 2.89893389e-01 -5.60105920e-01 -1.86772510e-01
-5.46313167e-01 -1.71137035e+00 1.33752394e+00 -1.66515708e-01
-3.09067488e-01 -8.84709001e-01 1.85897782e-01 1.39347892e-02
6.28334522e-01 1.53972268e-01 1.41181970e+00 -4.63562906e-01
-8.55247021e-01 8.50327164e-02 -3.03725809e-01 4.57085997e-01
4.41514373e-01 2.58487374e-01 -6.72515094e-01 -1.10718749e-01
-1.62758216e-01 -2.02601016e-01 5.31056881e-01 -2.84755319e-01
1.14174891e+00 -6.44098759e-01 -2.75568306e-01 8.75952244e-01
1.48586881e+00 2.13331506e-01 7.42281318e-01 1.10446751e-01
9.63809848e-01 3.74879688e-01 1.01553284e-01 4.38240439e-01
7.06791043e-01 7.74562895e-01 5.04320621e-01 -2.94293649e-03
-4.81793553e-01 -6.45101428e-01 8.66987407e-01 8.71744633e-01
5.91613293e-01 -2.35636353e-01 -1.39455891e+00 7.66416669e-01
-1.69100869e+00 -7.10185587e-01 -4.35378283e-01 1.99280226e+00
1.30993176e+00 -1.86551325e-02 -1.05973147e-01 -5.86312823e-02
4.23775285e-01 -4.71881889e-02 -3.64560127e-01 -7.68154562e-01
5.71737707e-01 3.86997461e-01 3.09987724e-01 5.46188354e-01
-6.84519827e-01 9.08877611e-01 5.08612633e+00 5.69540620e-01
-1.19513988e+00 1.56872943e-01 2.97466189e-01 1.62322491e-01
-6.10766590e-01 5.03772020e-01 -8.80760908e-01 6.32645428e-01
1.35470998e+00 -5.80718219e-01 6.26189709e-01 1.02972150e+00
1.85143724e-01 3.32253516e-01 -1.69637597e+00 5.22227168e-01
-7.01618269e-02 -1.58837545e+00 7.79367015e-02 2.37888284e-02
7.76152432e-01 1.36083113e-02 -1.16000079e-01 4.99245554e-01
4.81102079e-01 -1.07591522e+00 1.12618196e+00 3.71715844e-01
1.04400194e+00 -4.70863640e-01 5.51734447e-01 2.10969016e-01
-1.43658161e+00 -1.28208280e-01 4.93320264e-02 -8.09382871e-02
-4.77344058e-02 5.72323024e-01 -8.43478978e-01 4.54678297e-01
3.27924490e-01 1.02998888e+00 -1.02661371e+00 7.84566522e-01
-4.39649493e-01 7.49321461e-01 3.82151842e-01 1.11488104e-01
8.71934404e-04 2.83370435e-01 2.92186499e-01 1.45344889e+00
7.13460326e-01 -5.38010240e-01 5.42062684e-04 1.71975625e+00
-4.17059332e-01 -1.23121766e-02 -6.83851898e-01 -5.25362074e-01
2.95134902e-01 9.74610567e-01 -3.34624738e-01 -3.67673367e-01
-7.85608709e-01 8.44329536e-01 2.42935374e-01 1.66655660e-01
-1.05584097e+00 -6.60353243e-01 6.67311907e-01 3.06017369e-01
4.13109183e-01 -3.18424553e-01 -5.31724393e-01 -1.29292691e+00
3.60149950e-01 -1.26271701e+00 -4.49520200e-02 -7.90965259e-01
-6.43847406e-01 6.89583838e-01 4.25403602e-02 -1.21731043e+00
-5.65647423e-01 -4.01035905e-01 -7.80497074e-01 1.07959139e+00
-1.45190549e+00 -1.21557415e+00 -2.07755789e-01 1.36014327e-01
7.16633737e-01 -4.68145967e-01 7.74120510e-01 3.05764705e-01
-3.48262876e-01 7.77966380e-01 -7.30129182e-02 5.09303570e-01
2.92449683e-01 -1.29974997e+00 1.41108298e+00 1.32921827e+00
-1.34878129e-01 9.76711154e-01 5.15229881e-01 -6.76714897e-01
-1.66903162e+00 -1.58199990e+00 1.04833508e+00 -5.82422495e-01
8.30333292e-01 -7.92866886e-01 -8.84186208e-01 8.89231026e-01
3.88394386e-01 -3.01975422e-02 5.46307087e-01 -4.95297372e-01
-7.39974499e-01 5.68600409e-02 -6.19095623e-01 5.13213694e-01
9.17835236e-01 -7.77849436e-01 -4.23426181e-01 2.27611512e-01
1.08072639e+00 -7.08726168e-01 -8.26104224e-01 4.81951162e-02
4.47709978e-01 -8.95205021e-01 6.31747484e-01 -3.69036287e-01
1.36983049e+00 -5.49496353e-01 -1.08256400e-01 -1.23531687e+00
1.31495282e-01 -7.08012819e-01 -3.47342998e-01 1.55461597e+00
7.66602218e-01 -3.05054158e-01 5.49750805e-01 3.91177803e-01
-6.08130693e-01 -6.70235097e-01 -2.66897768e-01 -6.57006681e-01
3.48163009e-01 -5.06874442e-01 9.41038609e-01 6.46859527e-01
1.36256099e-01 4.30797726e-01 -1.92371473e-01 -1.41650379e-01
1.14723757e-01 5.14300883e-01 1.02157652e+00 -7.93289483e-01
-7.47175336e-01 -5.16720951e-01 1.11510474e-02 -1.06198442e+00
5.76705456e-01 -1.44406629e+00 1.67218715e-01 -1.63225043e+00
1.96441397e-01 -1.94384426e-01 4.38993752e-01 7.78929055e-01
1.11564815e-01 -2.40411475e-01 2.08332941e-01 6.48503900e-02
-1.32567361e-01 5.71216106e-01 8.60045612e-01 -3.59247476e-01
-9.34349000e-02 -1.36871813e-02 -1.03089929e+00 3.19054216e-01
9.52197134e-01 -7.93940127e-01 -5.99814057e-01 -9.38758433e-01
5.47549605e-01 2.17725933e-01 3.66560847e-01 -8.88740122e-01
-3.86550613e-02 -1.06716536e-01 -3.44562769e-01 -1.02981411e-01
-3.01715851e-01 -5.86965382e-01 2.31664881e-01 5.91566265e-01
-6.80572808e-01 4.50673491e-01 4.97939348e-01 1.13140598e-01
-1.93897679e-01 -5.26614666e-01 7.84863293e-01 -1.91620022e-01
-6.42281771e-01 2.52182603e-01 -4.33309853e-01 4.06744510e-01
8.19570184e-01 3.97309065e-02 -6.19149148e-01 -9.09707025e-02
-4.64914404e-02 -1.10653929e-01 9.00607586e-01 9.19779837e-01
6.23331547e-01 -1.23582196e+00 -8.58017743e-01 4.16533709e-01
4.96560484e-01 -1.89567134e-01 -2.25074887e-01 5.55183530e-01
-7.46074855e-01 4.43831801e-01 -1.06049649e-01 -2.27648258e-01
-1.11904144e+00 6.04336381e-01 5.01329184e-01 -3.93524796e-01
-4.18763459e-01 5.08483291e-01 4.53093290e-01 -6.68325484e-01
-2.66048521e-01 -6.98744714e-01 3.66805524e-01 -6.61893606e-01
5.58181405e-01 1.19771004e-01 2.07544297e-01 -6.13416255e-01
-2.97590107e-01 3.74743402e-01 -2.38946557e-01 1.67299226e-01
1.14955676e+00 3.20406914e-01 -6.00311339e-01 1.26157358e-01
1.54148388e+00 2.13777244e-01 -9.78704751e-01 -3.22186239e-02
1.15817048e-01 -3.62733752e-01 -1.28009096e-01 -9.43492115e-01
-1.14518452e+00 1.26959229e+00 2.68031925e-01 9.91192833e-02
9.72912967e-01 -6.00723587e-02 1.04218531e+00 2.16482058e-01
4.14893150e-01 -4.68122780e-01 2.19180688e-01 8.55281651e-01
9.55160141e-01 -8.55755329e-01 -4.78381813e-01 -3.77627850e-01
-3.81009489e-01 1.30838311e+00 8.44957829e-01 1.19070671e-01
4.61456209e-01 8.56460273e-01 -2.00775787e-01 1.43403336e-02
-9.52007771e-01 1.65210605e-01 2.90787846e-01 6.66905761e-01
1.09435093e+00 -2.01664969e-01 -1.73615724e-01 6.07949436e-01
-4.44007635e-01 3.43438983e-01 1.11711085e+00 1.23421216e+00
-7.61811212e-02 -1.77940571e+00 -6.67411238e-02 5.64423084e-01
-4.74349469e-01 -7.08846569e-01 -4.44048434e-01 5.59070587e-01
5.88096073e-03 8.23507428e-01 -1.69447064e-01 -4.28738952e-01
2.69198716e-01 7.00704902e-02 3.07553232e-01 -1.28196323e+00
-9.51735854e-01 -4.72147554e-01 4.21636403e-02 -5.71208954e-01
5.19795865e-02 -4.61130738e-01 -1.54132557e+00 -3.07904303e-01
-3.00140511e-02 1.68853343e-01 5.80102921e-01 6.67050600e-01
8.95327926e-01 8.21864486e-01 4.06903625e-01 -4.43183988e-01
-2.85593808e-01 -6.68855190e-01 -4.98250909e-02 2.49945670e-01
5.11802614e-01 -1.81610081e-02 7.65279755e-02 7.74138927e-01] | [7.738667964935303, 7.882440090179443] |
fc7d8c9b-aa51-414a-89cb-da7d0bc2ca7b | a-simple-yet-effective-baseline-for-robust-1 | null | null | https://openreview.net/forum?id=S1xnKi5BOV | https://openreview.net/pdf?id=S1xnKi5BOV | A Simple yet Effective Baseline for Robust Deep Learning with Noisy Labels | Recently deep neural networks have shown their capacity to memorize training data, even with noisy labels, which hurts generalization performance. To mitigate this issue, we propose a simple but effective method that is robust to noisy labels, even with severe noise. Our objective involves a variance regularization term that implicitly penalizes the Jacobian norm of the neural network on the whole training set (including the noisy-labeled data), which encourages generalization and prevents overfitting to the corrupted labels. Experiments on noisy benchmarks demonstrate that our approach achieves state-of-the-art performance with a high tolerance to severe noise. | ['Anonymous'] | 2019-03-25 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 2.48655155e-01 -7.21063092e-02 -2.92660668e-02 -6.49753213e-01
-7.83334613e-01 -4.91852790e-01 2.92189330e-01 -1.09229796e-01
-8.17164838e-01 8.74759674e-01 -1.45584896e-01 -1.23057969e-01
-7.41762593e-02 -7.37035394e-01 -8.20151329e-01 -8.64084005e-01
3.80355455e-02 8.33433941e-02 -3.16996649e-02 3.69667038e-02
-6.41499683e-02 2.16683626e-01 -1.45982814e+00 7.25104138e-02
7.04074025e-01 1.35281956e+00 -1.45217642e-01 2.59287030e-01
6.97335750e-02 1.03612256e+00 -1.01860404e+00 -2.84015715e-01
3.95351589e-01 -2.53722109e-02 -9.47121203e-01 1.31612003e-01
5.91897666e-01 1.58167761e-02 -4.08589035e-01 1.28439331e+00
5.60307682e-01 5.14041603e-01 4.75226730e-01 -8.34323645e-01
-9.05616760e-01 5.57868540e-01 -2.24596664e-01 8.79463851e-02
-2.91518539e-01 2.43121549e-01 8.71545792e-01 -9.53082025e-01
3.53927195e-01 1.06914580e+00 9.81783092e-01 7.11645782e-01
-1.45902276e+00 -6.17500961e-01 2.70563185e-01 -1.08956449e-01
-1.54000008e+00 -2.45040864e-01 6.41634881e-01 -2.66440839e-01
7.78953552e-01 5.19691929e-02 1.59073666e-01 1.20845842e+00
4.23996039e-02 6.99141681e-01 1.25347590e+00 -1.55184433e-01
3.90072852e-01 -1.30579934e-01 5.16042650e-01 5.75185537e-01
1.58090055e-01 -7.22221285e-03 -2.14283332e-01 -4.88974191e-02
3.17397088e-01 -9.36144143e-02 -2.87434310e-01 4.44328785e-03
-9.99073386e-01 6.90621674e-01 7.52580523e-01 2.05018550e-01
-2.92967349e-01 3.51041049e-01 5.96132517e-01 5.09966195e-01
7.16727197e-01 5.74934542e-01 -6.43333733e-01 -4.05998044e-02
-6.39779687e-01 -1.36632025e-01 7.32255578e-01 6.51243091e-01
7.68875241e-01 4.22259450e-01 -1.59140721e-01 1.30814004e+00
-1.00155927e-01 5.76607108e-01 6.63905323e-01 -9.82908785e-01
2.84335464e-01 5.53483844e-01 -1.79232344e-01 -9.15319681e-01
-6.11275077e-01 -9.64429259e-01 -1.39121842e+00 1.69024408e-01
3.14199895e-01 -2.64577091e-01 -1.36241913e+00 2.08778882e+00
7.31777251e-02 2.55718201e-01 7.73637295e-02 8.40284348e-01
9.58541155e-01 5.26650488e-01 2.44039774e-01 -9.69908163e-02
7.40461111e-01 -8.96461725e-01 -7.64884710e-01 -6.31332040e-01
7.03753650e-01 -3.50933224e-01 1.23224127e+00 5.07827997e-01
-9.00167346e-01 -6.26290143e-01 -1.09563112e+00 4.31168638e-02
-4.08905298e-01 6.94094151e-02 4.24783289e-01 6.29618704e-01
-1.03618741e+00 7.33089566e-01 -6.71949506e-01 1.45161182e-01
5.14752209e-01 5.20193100e-01 -4.47369069e-01 -1.41211733e-01
-1.36075342e+00 9.14364100e-01 8.19973409e-01 3.82054448e-01
-8.03427815e-01 -5.56530774e-01 -8.26033413e-01 1.93271384e-01
5.24934530e-01 -3.12454551e-01 1.28228688e+00 -9.96439517e-01
-1.53865230e+00 7.63473928e-01 8.67490619e-02 -3.90525699e-01
6.15136266e-01 -3.50682139e-01 -3.65701228e-01 -3.20005298e-01
-1.52222946e-01 7.38159835e-01 6.98131561e-01 -1.42985761e+00
-2.53197044e-01 -2.40035936e-01 -2.86954492e-01 -1.00015067e-01
-7.08146751e-01 -4.62246776e-01 -3.12142789e-01 -8.26619089e-01
4.27220523e-01 -1.02150810e+00 -2.82035559e-01 -2.50975937e-01
-6.37134433e-01 -3.17884177e-01 6.67434454e-01 -2.32785925e-01
1.07588196e+00 -2.39541960e+00 -2.48178169e-01 4.09134865e-01
1.46570653e-01 5.40099323e-01 -4.59195912e-01 -1.30905941e-01
-7.55168274e-02 3.33186150e-01 -2.90375710e-01 -6.06441736e-01
-1.05618043e-02 6.39780939e-01 -3.54988426e-01 5.21053433e-01
3.39253873e-01 7.96050906e-01 -9.38491881e-01 4.37141992e-02
-2.07575232e-01 5.15462339e-01 -7.24490881e-02 2.60178447e-02
-1.69873342e-01 4.25052315e-01 -1.98878571e-01 4.14406478e-01
7.81774521e-01 -6.08750045e-01 1.63050219e-01 6.02075905e-02
5.07236242e-01 3.50967139e-01 -1.35505736e+00 1.38753653e+00
-4.92442250e-01 5.71649849e-01 3.33761536e-02 -1.01320255e+00
1.15909576e+00 1.26593962e-01 2.21181408e-01 -6.21964157e-01
4.27722067e-01 2.87639052e-01 -2.50553489e-01 -2.74591267e-01
3.37309241e-01 -6.33693561e-02 -8.01203921e-02 4.04121727e-01
1.70831844e-01 1.22592360e-01 6.58140332e-02 -1.48648426e-01
1.05883086e+00 -3.56310099e-01 -2.03097109e-02 -3.16969365e-01
2.79386669e-01 -3.86938602e-01 8.05389524e-01 1.13264775e+00
-1.21821024e-01 6.24224365e-01 2.71773249e-01 -6.86873734e-01
-8.21437657e-01 -8.08069646e-01 -1.91407025e-01 1.32740557e+00
5.37256226e-02 -4.23137397e-02 -6.94726408e-01 -9.03869867e-01
2.28530057e-02 5.69022894e-01 -8.38121235e-01 -5.73348463e-01
-4.94983166e-01 -9.30250406e-01 7.51741230e-01 8.28830123e-01
6.74779534e-01 -8.64375293e-01 -6.78367466e-02 1.85397848e-01
-1.06695384e-01 -1.14093733e+00 -1.52437314e-01 6.82850480e-01
-7.98345685e-01 -7.84098029e-01 -5.57083488e-01 -9.39493060e-01
8.51784587e-01 4.36788090e-02 1.15592241e+00 4.18824255e-01
6.76487833e-02 -1.16368877e-02 -1.34190142e-01 -2.25443989e-01
-4.80888039e-01 1.83529466e-01 2.35559285e-01 -9.08373296e-02
2.47085541e-01 -3.97099435e-01 -2.88750589e-01 5.69259286e-01
-1.09540963e+00 -4.67236996e-01 4.01129425e-01 1.32397449e+00
6.61478460e-01 9.73725021e-02 7.88478136e-01 -9.90203977e-01
8.30058992e-01 -5.34579575e-01 -6.00902498e-01 2.33610705e-01
-7.15801895e-01 2.41913408e-01 9.10389721e-01 -8.14448953e-01
-7.53521204e-01 1.06380939e-01 -1.97645247e-01 -2.92021424e-01
-1.94631949e-01 6.46415055e-01 8.11859593e-02 -2.27080002e-01
1.05872071e+00 2.05365196e-01 -2.35068098e-01 -7.19139099e-01
4.18695778e-01 7.27349222e-01 7.32648611e-01 -6.75596297e-01
4.68116254e-01 3.15354377e-01 2.02240631e-01 -3.90688270e-01
-1.49963176e+00 -3.18038434e-01 -5.00875711e-01 -2.15198211e-02
3.69354993e-01 -9.23090637e-01 -7.34361947e-01 7.24096596e-01
-1.01231754e+00 -5.48229635e-01 -4.15475667e-01 4.43643391e-01
-2.00496092e-01 1.54366627e-01 -9.20784056e-01 -6.70306623e-01
-3.35789263e-01 -1.20293260e+00 6.12725258e-01 1.62076041e-01
-5.73508404e-02 -1.05439019e+00 -1.49680033e-01 7.02252015e-02
5.53054869e-01 2.88052768e-01 8.22043836e-01 -1.00101018e+00
-2.12169483e-01 -4.61585611e-01 -2.76501715e-01 1.23446822e+00
-8.59199315e-02 -1.79756105e-01 -1.24923003e+00 -4.05557275e-01
1.14332840e-01 -7.93183327e-01 1.28966248e+00 2.50423968e-01
1.48156357e+00 -3.26150060e-01 -2.79536238e-03 6.50008798e-01
1.28020823e+00 -1.89900696e-01 3.45807165e-01 4.17762309e-01
6.10973656e-01 2.41119727e-01 2.56139338e-01 3.06988031e-01
-5.07193105e-03 2.96321541e-01 5.75087547e-01 -1.28395721e-01
-3.47809456e-02 7.85191823e-03 -1.99249350e-02 9.04363275e-01
2.14473084e-01 -4.81282115e-01 -9.70167994e-01 4.66150731e-01
-1.77210784e+00 -4.88876134e-01 -1.36443168e-01 2.04160810e+00
1.08125317e+00 3.86781722e-01 -4.29233760e-01 4.41959560e-01
6.75687969e-01 1.60072356e-01 -7.15712130e-01 -3.97567183e-01
-3.31049740e-01 2.21793368e-01 6.45587325e-01 3.29459548e-01
-1.36880159e+00 8.98153365e-01 8.17341518e+00 9.67309177e-01
-1.19693446e+00 9.23984051e-02 9.48856473e-01 -9.37762633e-02
8.95948634e-02 -5.23274362e-01 -6.47616625e-01 5.09178281e-01
8.45691442e-01 2.55405039e-01 3.60875100e-01 1.20825791e+00
-9.50391740e-02 -2.26295907e-02 -9.99031186e-01 7.24453628e-01
-8.00810009e-02 -1.23359478e+00 -2.54080266e-01 -2.93881267e-01
1.35589588e+00 3.39030147e-01 4.18414176e-01 5.15329123e-01
6.78066015e-01 -1.17656755e+00 6.24665976e-01 3.99983197e-01
7.86353409e-01 -7.76137531e-01 1.08135092e+00 4.95677322e-01
-6.29002988e-01 -3.07951093e-01 -5.89264452e-01 -2.47900337e-01
-2.25020364e-01 1.24389780e+00 -7.79332876e-01 -1.18990634e-02
5.90456545e-01 4.99928027e-01 -6.59967780e-01 9.57877159e-01
-6.55498683e-01 8.83303940e-01 -4.10442561e-01 -1.23789376e-02
2.93025583e-01 2.25085318e-01 2.38177583e-01 1.27038014e+00
1.63126141e-01 -6.24177828e-02 6.26664102e-01 5.79171896e-01
-7.63004839e-01 -2.83357017e-02 -3.64256799e-01 2.71177620e-01
7.04398930e-01 9.67086792e-01 -6.58027232e-01 -4.16503787e-01
-6.15431294e-02 6.39187455e-01 7.84589589e-01 6.12599909e-01
-3.65211993e-01 -2.87678272e-01 5.54268718e-01 -3.52557808e-01
9.92072299e-02 -2.47158363e-01 -8.12114596e-01 -8.37825656e-01
2.32261524e-01 -8.49629700e-01 3.98979008e-01 -4.65003401e-01
-1.66401708e+00 8.00799906e-01 -4.50591266e-01 -9.64514434e-01
-5.83523363e-02 -6.68672800e-01 -3.56119692e-01 8.31836939e-01
-1.42157495e+00 -6.45468950e-01 -3.34727108e-01 4.15400088e-01
1.59700140e-01 -1.81058317e-01 8.17280054e-01 3.77636909e-01
-7.46276498e-01 9.29202378e-01 5.84066510e-01 3.94594580e-01
6.83017671e-01 -1.19059527e+00 3.20596963e-01 7.12800741e-01
-4.27963585e-02 6.39870584e-01 5.11872947e-01 -4.45488602e-01
-8.48685145e-01 -1.33876026e+00 6.63474143e-01 -2.14123413e-01
5.69326282e-01 -2.80257881e-01 -1.32302105e+00 6.43046558e-01
-2.37359554e-01 5.57021081e-01 7.91840613e-01 2.79927850e-01
-7.85024762e-01 -2.10814372e-01 -1.16707480e+00 3.27736199e-01
8.86120081e-01 -4.81635600e-01 -4.56955224e-01 4.81226712e-01
6.70563102e-01 -5.41439354e-01 -8.22719216e-01 7.43048608e-01
1.91676617e-01 -5.56329489e-01 7.57991016e-01 -6.81030810e-01
1.85120925e-01 -5.55578321e-02 -1.26785755e-01 -1.68091989e+00
-2.87521571e-01 -4.95506912e-01 -1.25600904e-01 9.91852462e-01
4.83634651e-01 -6.42546713e-01 7.85023272e-01 7.47262001e-01
-1.56208977e-01 -9.16253448e-01 -1.08371890e+00 -1.24878383e+00
1.04914106e-01 -6.37550235e-01 3.68547827e-01 1.24546981e+00
-2.33558089e-01 1.49962902e-01 -5.81352413e-01 5.65009890e-03
3.66944581e-01 -4.94513541e-01 3.95988792e-01 -1.39013851e+00
-8.29549059e-02 -3.94933432e-01 -5.31738877e-01 -9.91893589e-01
5.85867465e-01 -8.56687903e-01 5.21182656e-01 -1.31192327e+00
-2.04182770e-02 -5.79388976e-01 -6.98370874e-01 9.66906607e-01
-4.86092627e-01 7.23567069e-01 -1.91009622e-02 2.41352364e-01
-7.87707746e-01 4.83387291e-01 1.24639487e+00 -2.35543773e-01
-5.37482612e-02 1.03528798e-03 -5.72559714e-01 7.99141586e-01
8.99975479e-01 -9.33415234e-01 -2.36134186e-01 -7.38954425e-01
2.47462332e-01 -7.47686565e-01 1.81227490e-01 -1.13558769e+00
1.43174857e-01 -3.96367647e-02 4.40213799e-01 -1.63924545e-01
3.42446953e-01 -6.88268006e-01 -7.01437816e-02 4.11495894e-01
-9.63056207e-01 -1.77509949e-01 1.69227406e-01 6.29758835e-01
-3.85663807e-01 -4.53723818e-01 1.04713440e+00 8.91333669e-02
-3.82585406e-01 1.73336938e-01 -3.29924852e-01 3.51231307e-01
6.81040347e-01 2.30569869e-01 -5.65027893e-01 -4.42600310e-01
-9.72866952e-01 3.52565765e-01 2.81390697e-01 2.15124846e-01
3.95115465e-01 -1.52668631e+00 -6.64309800e-01 1.72181115e-01
-2.28819884e-02 1.48569778e-01 1.10814072e-01 4.07471836e-01
-2.02427521e-01 8.59988257e-02 1.69666007e-01 -6.04284704e-01
-9.92607772e-01 5.97747087e-01 4.77193773e-01 -2.97448486e-01
-3.96219641e-01 1.20254338e+00 -3.18487763e-01 -5.95033288e-01
7.88849235e-01 -3.92015845e-01 -4.24099341e-02 -6.43083602e-02
5.23579895e-01 4.88724709e-01 3.83738399e-01 -5.94458580e-01
-2.65930384e-01 4.10170615e-01 -1.41537130e-01 2.23538831e-01
1.23507166e+00 1.18335009e-01 3.73105198e-04 5.88173687e-01
1.64549005e+00 -5.17306924e-01 -1.45840812e+00 -8.30732107e-01
2.15670675e-01 -3.63718837e-01 2.71839231e-01 -1.02322268e+00
-1.31589484e+00 8.93206298e-01 6.64219558e-01 5.37850261e-01
1.04703867e+00 -3.76801878e-01 8.64887834e-01 1.06652510e+00
8.59556571e-02 -1.45598066e+00 2.21156806e-01 1.04948461e+00
5.40757775e-01 -1.41345775e+00 -1.35436743e-01 -2.43929729e-01
-5.91513991e-01 1.10542405e+00 5.62190473e-01 -2.03850836e-01
7.61076808e-01 2.98126668e-01 5.43733954e-01 8.20417479e-02
-6.75299764e-01 -4.67327721e-02 3.20021629e-01 4.86757666e-01
2.05299124e-01 2.13598963e-02 -2.14833677e-01 5.54434121e-01
9.97019783e-02 1.67608657e-03 3.28048825e-01 1.00394177e+00
-5.50453722e-01 -8.96762788e-01 -3.92782271e-01 4.22277957e-01
-6.61075592e-01 -5.77206053e-02 -2.88930774e-01 4.68216032e-01
2.33198926e-01 1.07328677e+00 -3.68738584e-02 -4.48806763e-01
5.00592589e-01 2.74696380e-01 -2.27913503e-02 -5.76559961e-01
-7.92313159e-01 -8.83928165e-02 -1.45010233e-01 -4.92308974e-01
-2.47821584e-01 -7.06497952e-02 -1.25804710e+00 -2.45929867e-01
-4.86899436e-01 7.13438317e-02 7.64027894e-01 9.72841859e-01
3.13221365e-01 6.03716552e-01 8.30970109e-01 -6.60233319e-01
-1.13379061e+00 -1.21227491e+00 -6.29742265e-01 7.81312227e-01
4.47097778e-01 -6.18450046e-01 -5.47170401e-01 -1.95794255e-01] | [9.264730453491211, 3.812629222869873] |
f449ec8e-42f6-4e41-8a63-c44d8e99e0de | a-study-of-latent-monotonic-attention | 2103.16710 | null | https://arxiv.org/abs/2103.16710v1 | https://arxiv.org/pdf/2103.16710v1.pdf | A study of latent monotonic attention variants | End-to-end models reach state-of-the-art performance for speech recognition, but global soft attention is not monotonic, which might lead to convergence problems, to instability, to bad generalisation, cannot be used for online streaming, and is also inefficient in calculation. Monotonicity can potentially fix all of this. There are several ad-hoc solutions or heuristics to introduce monotonicity, but a principled introduction is rarely found in literature so far. In this paper, we present a mathematically clean solution to introduce monotonicity, by introducing a new latent variable which represents the audio position or segment boundaries. We compare several monotonic latent models to our global soft attention baseline such as a hard attention model, a local windowed soft attention model, and a segmental soft attention model. We can show that our monotonic models perform as good as the global soft attention model. We perform our experiments on Switchboard 300h. We carefully outline the details of our training and release our code and configs. | ['Hermann Ney', 'Ralf Schlüter', 'Albert Zeyer'] | 2021-03-30 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 3.78587574e-01 2.14009374e-01 -2.42942110e-01 -3.11331749e-01
-1.24239206e+00 -3.96866769e-01 5.80183744e-01 -1.85729533e-01
-1.78530589e-01 7.58081079e-01 5.86102903e-01 -5.57909846e-01
-3.37768942e-01 -1.30713910e-01 -6.86180353e-01 -7.55131364e-01
-1.72276512e-01 5.36968172e-01 5.29954195e-01 -1.57990158e-02
2.22988859e-01 1.18481167e-01 -1.51806808e+00 4.09217983e-01
5.81113398e-01 9.88513052e-01 2.80411184e-01 1.03152418e+00
-1.02284759e-01 8.52147102e-01 -7.80141294e-01 -1.77340493e-01
1.31639495e-01 -5.66280246e-01 -1.10132790e+00 1.37497336e-01
3.70412737e-01 -3.48684907e-01 -8.84677023e-02 8.25125813e-01
6.75835788e-01 1.54015437e-01 3.00268412e-01 -1.12731791e+00
-6.31855667e-01 7.88662255e-01 -3.35554570e-01 5.39370298e-01
3.77307147e-01 -1.14027150e-01 1.15805686e+00 -8.72910500e-01
2.43836284e-01 1.18431532e+00 7.11527526e-01 5.36990643e-01
-1.03893507e+00 -3.75465095e-01 4.64875370e-01 3.63981754e-01
-1.02650201e+00 -1.02606952e+00 4.80291396e-01 -1.55376375e-01
1.25801790e+00 6.79445267e-01 4.51663166e-01 1.27569354e+00
-1.71768263e-01 1.05863965e+00 9.37506616e-01 -6.53542936e-01
1.29969195e-01 -1.39867783e-01 2.96540469e-01 3.62010926e-01
-3.20027024e-01 5.32450527e-02 -9.11814988e-01 -2.78163373e-01
5.92297256e-01 -3.90967935e-01 -4.02967513e-01 -1.52530104e-01
-1.05618548e+00 6.13534391e-01 1.54262945e-01 2.68084466e-01
-2.60062873e-01 1.35119930e-01 4.31684285e-01 6.41061962e-01
5.47366738e-01 2.07600802e-01 -7.49839723e-01 -7.37094581e-01
-1.22192311e+00 1.71630397e-01 8.15641165e-01 8.50300312e-01
1.81088030e-01 2.29994655e-01 -9.15378481e-02 1.21216285e+00
3.38126659e-01 1.25026464e-01 6.45846486e-01 -1.08657527e+00
4.91500080e-01 -3.06393504e-01 -1.59184411e-01 -4.05174762e-01
-3.18270206e-01 -4.72794145e-01 -5.56398869e-01 -4.21438515e-02
3.54636639e-01 -7.36239478e-02 -8.64203393e-01 1.64163113e+00
-2.24992663e-01 3.72014105e-01 -1.00031845e-01 9.07198071e-01
7.21692324e-01 8.16235483e-01 -9.92712155e-02 -6.94038987e-01
1.00918162e+00 -1.40920317e+00 -9.22899961e-01 -4.13602889e-01
3.55082273e-01 -9.14407670e-01 1.01548612e+00 6.76748693e-01
-1.72160769e+00 -3.56762499e-01 -9.56099987e-01 -4.77132760e-02
-9.51788388e-03 -2.61942625e-01 5.84431946e-01 7.56900489e-01
-1.29753530e+00 4.08087373e-01 -1.18046510e+00 -3.87766629e-01
-1.28222965e-02 5.01881659e-01 1.56394765e-02 2.16059074e-01
-1.12303460e+00 7.83638299e-01 9.15431529e-02 1.80280060e-02
-8.14914286e-01 -4.21735287e-01 -6.84997320e-01 1.93163887e-01
4.33560193e-01 -5.63767970e-01 1.83181512e+00 -1.23330688e+00
-2.14780784e+00 6.38094902e-01 -4.34131324e-01 -5.79168200e-01
2.69440621e-01 -2.74987161e-01 -1.44099176e-01 -2.34225485e-03
-2.45435312e-01 5.31341910e-01 9.24955428e-01 -1.00598502e+00
-5.54517210e-01 4.94071655e-02 2.07006615e-02 1.62124351e-01
-2.55133629e-01 5.51694095e-01 -6.58437848e-01 -8.72867525e-01
2.60577798e-01 -8.38483870e-01 -7.12591410e-02 -3.34071994e-01
-2.28425056e-01 -1.84468552e-01 4.57029849e-01 -7.13508248e-01
1.56384504e+00 -2.15228415e+00 2.43422762e-01 -4.60458770e-02
-2.76910305e-01 4.33856577e-01 -1.06857553e-01 3.20350081e-01
-3.15202564e-01 2.22298369e-01 -4.73808199e-01 -7.01656759e-01
2.07309395e-01 3.74511957e-01 -6.19385898e-01 2.71483302e-01
-8.05930048e-02 7.26315141e-01 -9.07359719e-01 -2.20684990e-01
-3.05598751e-02 5.08666158e-01 -8.09160352e-01 1.04400076e-01
-1.04079179e-01 -7.48438984e-02 9.34473276e-02 6.88320577e-01
3.93074453e-01 -2.88721938e-02 1.55495077e-01 4.53437448e-01
-1.33981228e-01 1.15940154e+00 -1.07247126e+00 1.46821284e+00
-3.12050402e-01 8.43463778e-01 4.62112397e-01 -1.07480657e+00
5.83155334e-01 7.04974711e-01 1.78782478e-01 -3.45009476e-01
-2.06004217e-01 2.57794142e-01 3.52545194e-02 -3.92608762e-01
5.52070677e-01 -3.35813344e-01 1.83574930e-01 4.85505700e-01
1.07223570e-01 1.25573233e-01 -1.40299618e-01 9.50174406e-02
1.13354301e+00 -1.67507809e-02 2.04986230e-01 -1.49760067e-01
3.78831029e-01 -5.50198793e-01 4.75538313e-01 7.82475471e-01
-3.14419746e-01 1.03003502e+00 7.40125179e-01 -1.10702574e-01
-7.29924679e-01 -1.02099860e+00 -1.40015230e-01 1.52244878e+00
-3.68288398e-01 -6.53598189e-01 -7.17153609e-01 -4.75044936e-01
-4.61975098e-01 4.34185535e-01 -2.43366107e-01 4.71933186e-02
-5.68826556e-01 -6.96603179e-01 5.21135509e-01 7.07914889e-01
1.08805828e-01 -1.37667513e+00 -3.45256627e-01 3.23565274e-01
-3.37581486e-01 -9.53755438e-01 -6.13144100e-01 4.94693607e-01
-7.94820130e-01 -3.23694587e-01 -7.54214048e-01 -7.66148746e-01
2.73938566e-01 2.44873434e-01 9.94860709e-01 1.43787265e-01
2.91906774e-01 1.80700347e-01 -4.30265784e-01 -2.89785743e-01
-4.17455822e-01 2.99611241e-01 9.32570100e-02 -1.25738919e-01
8.89702216e-02 -4.07770663e-01 -2.52869517e-01 2.35902280e-01
-6.88859105e-01 1.11985411e-02 4.81758952e-01 9.79972124e-01
3.16584319e-01 -2.30922699e-01 7.25259364e-01 -6.70183659e-01
6.74421191e-01 -4.07419264e-01 -3.07159543e-01 -3.38867493e-02
-3.61270428e-01 -3.80259417e-02 4.25167054e-01 -2.18453497e-01
-8.97749782e-01 2.20567919e-02 -7.24821985e-01 -4.07399327e-01
-1.52980387e-01 7.19481409e-01 -1.86319336e-01 2.42838979e-01
2.94705927e-01 1.75701141e-01 -1.43996151e-02 -6.42496228e-01
2.63167053e-01 9.92068350e-01 5.90226591e-01 -4.67802107e-01
3.40985626e-01 1.20831728e-01 -5.78378439e-01 -1.14392924e+00
-7.74874270e-01 -5.47980249e-01 -3.99782091e-01 6.18348233e-02
5.35492122e-01 -6.72698379e-01 -2.76348263e-01 3.98152679e-01
-1.14353919e+00 -7.49059498e-01 -2.00041085e-01 3.83889914e-01
-9.50513542e-01 5.58243513e-01 -9.50940967e-01 -1.05996537e+00
-1.61768556e-01 -1.08744287e+00 1.00860119e+00 -3.36389393e-01
-4.38374102e-01 -9.47130203e-01 7.96362087e-02 4.74443763e-01
5.29451311e-01 -3.12935412e-01 5.36330760e-01 -4.70749468e-01
-5.22479475e-01 -5.36835054e-03 7.31268302e-02 5.03241301e-01
-1.18218549e-01 2.04655737e-01 -1.18900907e+00 -4.88494486e-01
8.10050219e-02 -1.50831774e-01 1.17454743e+00 7.23780751e-01
1.22480834e+00 -4.58472043e-01 -9.76101235e-02 5.84907532e-01
8.10320854e-01 1.67931601e-01 6.74600482e-01 2.90901810e-01
2.74383426e-01 4.92765248e-01 3.13476652e-01 3.16329598e-01
2.17955351e-01 9.40614820e-01 2.25446895e-01 3.08235437e-02
-8.70722011e-02 1.38926521e-01 8.90814126e-01 1.40921664e+00
-3.05437535e-01 -4.59269673e-01 -8.40736985e-01 7.37399697e-01
-1.88644671e+00 -1.29182613e+00 -9.38170478e-02 2.38860750e+00
8.12946379e-01 6.60931826e-01 4.70423341e-01 5.50610065e-01
5.53428173e-01 1.72603458e-01 9.39208716e-02 -8.46973598e-01
-1.00923562e-02 2.57100314e-01 3.86996120e-01 1.14295149e+00
-1.08569252e+00 8.44312429e-01 7.74295187e+00 7.39413738e-01
-1.04197872e+00 4.26719427e-01 3.54023337e-01 -5.12116730e-01
-1.99845612e-01 -4.55325022e-02 -6.75218463e-01 3.47131670e-01
1.38156199e+00 1.29248708e-01 5.21225750e-01 6.62527978e-01
1.68435931e-01 3.96703392e-01 -9.71739769e-01 7.86083341e-01
2.18666255e-01 -1.01432931e+00 -2.12490410e-01 1.35877147e-01
5.48727334e-01 3.23486030e-01 1.28926232e-01 4.10691530e-01
-2.68248413e-02 -8.24454486e-01 8.53363514e-01 2.78354406e-01
6.10449791e-01 -7.05224991e-01 6.10911846e-01 5.11708915e-01
-9.07058418e-01 -1.04581892e-01 -3.18111330e-01 -3.61682922e-01
3.40903997e-01 1.87564805e-01 -5.27282119e-01 4.35636342e-01
6.78295732e-01 4.45221484e-01 -2.53671944e-01 1.15883017e+00
-3.15254927e-01 1.07408178e+00 -5.26809216e-01 6.00675493e-03
4.92015481e-01 4.18792993e-01 7.20503747e-01 1.46158683e+00
1.72657102e-01 8.73014480e-02 -2.26738844e-02 3.98459554e-01
1.15618370e-01 2.68413946e-02 -2.13354483e-01 1.15711138e-01
1.56724125e-01 7.22389460e-01 -5.82812250e-01 -5.11479616e-01
-7.14388788e-01 9.33119357e-01 3.19378197e-01 5.14344215e-01
-9.06442344e-01 -3.69009316e-01 6.03720605e-01 1.67985469e-01
5.56567669e-01 -2.90700376e-01 -3.03773522e-01 -1.22093344e+00
1.39871081e-02 -1.06312263e+00 4.59139824e-01 -7.61527419e-01
-1.14781499e+00 7.42826164e-01 4.01864387e-02 -8.57196748e-01
-5.15301228e-01 -5.04230559e-01 -6.64330125e-01 7.76782691e-01
-1.40415263e+00 -6.14440143e-01 1.92394238e-02 2.79766798e-01
1.05854309e+00 9.10835434e-03 8.90703440e-01 4.68792886e-01
-4.85300004e-01 8.32144558e-01 -1.83939971e-02 -1.73511922e-01
7.40112901e-01 -1.34984088e+00 5.51259339e-01 9.80071306e-01
3.29305023e-01 7.30272710e-01 8.18058193e-01 -2.32077807e-01
-1.01975107e+00 -5.08496046e-01 1.29357386e+00 -6.04436636e-01
7.62656927e-01 -3.06395978e-01 -1.18003166e+00 9.99664307e-01
4.73229080e-01 -2.16234386e-01 7.21791148e-01 5.83032608e-01
-1.57065079e-01 1.05570788e-02 -4.91683155e-01 4.64514911e-01
1.08517063e+00 -6.37324274e-01 -8.67552936e-01 3.40714276e-01
8.51117194e-01 -4.46939111e-01 -5.34728050e-01 2.97630727e-01
6.23808563e-01 -1.11490190e+00 9.69909906e-01 -4.17389214e-01
1.84717312e-01 5.05819619e-02 -1.50959551e-01 -1.16759789e+00
-5.85167706e-01 -1.22057521e+00 -5.03682196e-01 1.15049255e+00
6.55915320e-01 -7.50034630e-01 6.67572498e-01 4.35938835e-01
-7.72611678e-01 -9.58151102e-01 -1.24679399e+00 -1.09084654e+00
1.41935602e-01 -8.94110322e-01 4.45459843e-01 7.21307635e-01
5.63062370e-01 3.51148933e-01 -4.38224792e-01 1.77454129e-01
2.25512281e-01 -5.51139303e-02 5.40394843e-01 -7.91544974e-01
-7.53039300e-01 -1.05007923e+00 -2.40987852e-01 -1.74951327e+00
7.01119304e-02 -7.77339816e-01 1.94529161e-01 -1.44019175e+00
3.96672226e-02 -1.54237911e-01 -6.56027615e-01 5.87073505e-01
-2.82802526e-02 2.03983381e-01 1.73924029e-01 3.63441110e-02
-6.15690470e-01 4.08556163e-01 7.90907860e-01 -2.76585873e-02
-4.15682197e-01 9.79969725e-02 -6.14825904e-01 7.22673655e-01
8.43012512e-01 -2.91204929e-01 -3.32060874e-01 -4.85279113e-01
9.55451205e-02 1.80249467e-01 1.58463880e-01 -9.50582206e-01
3.79859000e-01 8.31455067e-02 -2.45261088e-01 -5.40878594e-01
4.59182471e-01 -4.03666794e-01 -1.73352286e-01 2.77302176e-01
-4.69511598e-01 -1.54395729e-01 2.27650404e-01 3.15920860e-01
-5.19750416e-01 -2.87969708e-01 7.77231693e-01 1.18393444e-01
-3.77529919e-01 2.05143586e-01 -6.26635134e-01 -6.36442900e-02
4.55822617e-01 -1.97354019e-01 -1.08351946e-01 -6.78674400e-01
-1.07875204e+00 1.13547005e-01 8.02679509e-02 6.77471876e-01
3.79080415e-01 -1.22219813e+00 -7.90588319e-01 4.92904812e-01
-2.22505972e-01 -2.99629718e-01 -3.53774540e-02 1.06854105e+00
-3.67199689e-01 6.83955967e-01 2.43379533e-01 -5.12907326e-01
-1.50343275e+00 5.85203409e-01 2.01223329e-01 -1.29583016e-01
-4.93026435e-01 1.07620382e+00 1.00381777e-01 -1.94662794e-01
7.66518176e-01 -7.40450442e-01 1.94667637e-01 -2.94668842e-02
6.60553932e-01 3.36176068e-01 1.71880484e-01 -6.02890611e-01
-5.85450351e-01 4.23723161e-01 5.63592538e-02 -4.06900376e-01
1.20880115e+00 -1.64450228e-01 4.13661338e-02 7.17109203e-01
1.00504136e+00 5.19890748e-02 -1.23294365e+00 -1.53133720e-01
-3.20551209e-02 -4.02046204e-01 1.95030540e-01 -6.88711584e-01
-7.52424598e-01 1.14124918e+00 1.70054197e-01 8.86767149e-01
1.24071515e+00 1.74749881e-01 8.04922104e-01 1.37015522e-01
9.03850719e-02 -9.88787651e-01 -1.43624812e-01 9.33808923e-01
1.18225467e+00 -9.76109028e-01 -1.81948170e-01 -3.19708556e-01
-4.25886214e-01 9.63109493e-01 2.74811029e-01 1.70644313e-01
6.18110180e-01 4.58516091e-01 8.42223614e-02 1.62916467e-01
-1.30905640e+00 -3.32699180e-01 2.50598818e-01 3.81419241e-01
6.63480341e-01 2.03491212e-03 -1.26007125e-01 7.59172678e-01
-4.37898874e-01 -2.20852211e-01 3.68103027e-01 7.75380373e-01
-6.12155497e-01 -1.14561188e+00 -4.47577596e-01 1.25689328e-01
-8.72617066e-01 -2.73311138e-01 -2.86322713e-01 4.44530219e-01
-7.92437941e-02 1.00436783e+00 5.48679382e-02 -1.29259199e-01
3.05677295e-01 5.00689447e-01 4.77583766e-01 -5.77420115e-01
-5.72172225e-01 4.57657635e-01 2.11399719e-01 -3.62248302e-01
-2.63104230e-01 -1.00086987e+00 -1.02291584e+00 -3.27267081e-01
-4.57129955e-01 3.87276188e-02 5.64556837e-01 9.11696792e-01
2.13462129e-01 6.08884573e-01 3.63247097e-01 -1.02021098e+00
-5.48339248e-01 -1.18521380e+00 -2.61932522e-01 1.24341637e-01
8.17381144e-01 -4.12281781e-01 -7.73604870e-01 1.04766577e-01] | [14.538613319396973, 6.33204984664917] |
2d52b2bc-e3d5-4c02-8768-eb7154c67ab0 | improving-pairwise-ranking-for-multi-label | 1704.03135 | null | http://arxiv.org/abs/1704.03135v3 | http://arxiv.org/pdf/1704.03135v3.pdf | Improving Pairwise Ranking for Multi-label Image Classification | Learning to rank has recently emerged as an attractive technique to train
deep convolutional neural networks for various computer vision tasks. Pairwise
ranking, in particular, has been successful in multi-label image
classification, achieving state-of-the-art results on various benchmarks.
However, most existing approaches use the hinge loss to train their models,
which is non-smooth and thus is difficult to optimize especially with deep
networks. Furthermore, they employ simple heuristics, such as top-k or
thresholding, to determine which labels to include in the output from a ranked
list of labels, which limits their use in the real-world setting. In this work,
we propose two techniques to improve pairwise ranking based multi-label image
classification: (1) we propose a novel loss function for pairwise ranking,
which is smooth everywhere and thus is easier to optimize; and (2) we
incorporate a label decision module into the model, estimating the optimal
confidence thresholds for each visual concept. We provide theoretical analyses
of our loss function in the Bayes consistency and risk minimization framework,
and show its benefit over existing pairwise ranking formulations. We
demonstrate the effectiveness of our approach on three large-scale datasets,
VOC2007, NUS-WIDE and MS-COCO, achieving the best reported results in the
literature. | ['Yuncheng Li', 'Jiebo Luo', 'Yale Song'] | 2017-04-11 | improving-pairwise-ranking-for-multi-label-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Li_Improving_Pairwise_Ranking_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Li_Improving_Pairwise_Ranking_CVPR_2017_paper.pdf | cvpr-2017-7 | ['multi-label-image-classification'] | ['computer-vision'] | [ 2.69741625e-01 -3.41341585e-01 -2.51666069e-01 -7.42509723e-01
-1.17614007e+00 -5.41150808e-01 3.91753018e-01 5.84314883e-01
-7.12913334e-01 5.39429843e-01 -2.21030340e-01 -7.34687299e-02
-3.22379053e-01 -4.26848948e-01 -6.75183892e-01 -8.77104521e-01
5.40161133e-02 4.40182477e-01 1.77254379e-01 3.06113273e-01
3.07060122e-01 3.04305166e-01 -1.64958596e+00 2.19209418e-01
8.32433343e-01 1.46240902e+00 2.99340151e-02 3.70433867e-01
2.11941987e-01 8.69703770e-01 -3.96329701e-01 -5.68207145e-01
2.70531535e-01 -1.51591033e-01 -8.58416915e-01 -6.62678555e-02
9.07036185e-01 -2.37213820e-01 1.97556708e-02 1.13328695e+00
4.25221741e-01 9.08750668e-02 9.34166908e-01 -1.21117198e+00
-5.82303405e-01 3.70754540e-01 -8.15933704e-01 -2.07435846e-01
-2.97979176e-01 -4.36103284e-01 1.44164157e+00 -7.32317746e-01
2.49360755e-01 1.09775352e+00 7.64411449e-01 3.85402799e-01
-1.25049603e+00 -8.18971038e-01 1.08440220e-01 3.91892195e-01
-1.43612874e+00 -1.46992415e-01 7.66344607e-01 -4.85453844e-01
3.34125400e-01 2.28941604e-01 1.46843702e-01 7.76560724e-01
1.00947633e-01 8.86110902e-01 1.41411090e+00 -4.00869787e-01
-3.26502919e-02 1.49445385e-01 2.51431733e-01 8.81256044e-01
-8.02532136e-02 -6.55873120e-02 -4.22652513e-01 -2.92519301e-01
3.95944864e-01 2.32116878e-02 2.52891984e-02 -6.06928229e-01
-1.05057335e+00 1.04439580e+00 6.27603650e-01 8.38927459e-04
-6.65650591e-02 5.40827930e-01 6.75597072e-01 1.11744776e-01
8.65458965e-01 3.00943404e-01 -4.45797920e-01 3.76393378e-01
-9.53414679e-01 2.11801797e-01 2.94101030e-01 5.28005838e-01
5.87765396e-01 -7.54712462e-01 -3.78490806e-01 1.44275045e+00
5.35311699e-01 1.43772960e-01 1.40991926e-01 -1.05280614e+00
4.03991103e-01 4.29073155e-01 1.00451820e-02 -1.00469017e+00
-5.93522310e-01 -5.65691352e-01 -8.55948567e-01 2.40960121e-01
3.89088750e-01 1.72776610e-01 -8.21645141e-01 1.94624889e+00
1.89040676e-01 2.31147110e-01 -3.13495040e-01 9.04604554e-01
8.09440613e-01 3.13211560e-01 1.95854634e-01 -1.71290398e-01
1.21126962e+00 -1.16668212e+00 -2.94583082e-01 5.79499528e-02
6.82020247e-01 -5.72666824e-01 8.89615238e-01 4.61025834e-01
-8.28684866e-01 -3.41335505e-01 -1.12665057e+00 -5.56285866e-02
-2.79112339e-01 3.89192075e-01 6.05856419e-01 4.57007647e-01
-1.22151852e+00 8.55081260e-01 -5.11091113e-01 -4.34491821e-02
6.62970901e-01 4.49414492e-01 -3.09770077e-01 -1.62056416e-01
-1.21802759e+00 8.47125411e-01 2.91527748e-01 1.82632253e-01
-8.63220811e-01 -5.33169746e-01 -7.12780714e-01 4.51254882e-02
4.77232397e-01 -4.98414159e-01 1.23733735e+00 -8.57351422e-01
-1.32313347e+00 1.12243664e+00 5.48542365e-02 -2.67925739e-01
7.44722009e-01 -1.27944812e-01 -1.80242211e-02 1.41600013e-01
1.79823473e-01 9.26412761e-01 6.68132186e-01 -1.42021871e+00
-8.35086524e-01 -3.15858245e-01 2.29792133e-01 2.44630262e-01
-5.70715785e-01 2.08897620e-01 -5.20916104e-01 -5.70363224e-01
4.40662429e-02 -1.06065536e+00 -3.36811632e-01 1.55689985e-01
-4.81189787e-01 -7.42833018e-01 3.64745378e-01 -3.04115713e-01
1.05294061e+00 -2.04228544e+00 6.02443218e-02 2.28896260e-01
3.89399648e-01 1.73572063e-01 -1.23259574e-01 1.45059386e-02
1.49980098e-01 3.03549260e-01 -1.89484432e-01 -7.71226346e-01
2.59800013e-02 -8.71196687e-02 3.68689746e-02 5.92190802e-01
1.78966492e-01 6.70776904e-01 -9.08077896e-01 -7.42360175e-01
1.97332382e-01 5.02334177e-01 -3.52096409e-01 3.14852856e-02
-3.00721079e-02 1.70304358e-01 -2.21268535e-01 4.44268227e-01
6.39792681e-01 -6.35803223e-01 2.51289785e-01 -3.83587837e-01
3.19117606e-02 9.51492190e-02 -1.07035851e+00 1.37610948e+00
-5.65579832e-01 5.46200275e-01 -1.26124099e-01 -1.10705113e+00
7.18627572e-01 1.24881260e-01 6.91500962e-01 -5.08742094e-01
2.72611082e-01 3.11417878e-01 -3.88436437e-01 -1.46558344e-01
3.26534867e-01 -1.66420728e-01 -6.51650876e-02 2.98759401e-01
-7.73910582e-02 2.57721782e-01 3.77015412e-01 8.39512944e-02
8.53320837e-01 -1.33919895e-01 2.27703363e-01 -3.20747018e-01
6.29294395e-01 -4.25775051e-01 6.00039840e-01 8.37573409e-01
-3.27624857e-01 7.67019093e-01 5.50561428e-01 -4.32133138e-01
-6.60152853e-01 -8.90005589e-01 -3.79534543e-01 1.32304418e+00
3.49400520e-01 -1.92943275e-01 -6.01061881e-01 -1.15959537e+00
1.55822085e-02 1.61349759e-01 -7.34795570e-01 5.25281429e-02
-3.81356031e-01 -8.18068981e-01 4.18017209e-01 5.54522216e-01
4.56515014e-01 -8.49945128e-01 -2.80685723e-01 1.46880046e-01
-2.78112501e-01 -9.69484210e-01 -6.61375761e-01 5.80019236e-01
-6.76453054e-01 -1.19324780e+00 -8.07807684e-01 -8.48477900e-01
7.74950981e-01 3.33934963e-01 1.02168202e+00 2.10770071e-01
-1.20086968e-01 1.14523441e-01 -2.67327547e-01 -1.16090856e-01
-5.01752198e-02 1.40077472e-01 -1.61198437e-01 2.04071850e-01
3.50272395e-02 -1.37511820e-01 -6.17263556e-01 6.34810507e-01
-8.30468953e-01 -4.26949002e-03 5.85981846e-01 1.12915325e+00
8.87339771e-01 5.82222193e-02 6.56839371e-01 -1.08139527e+00
4.09605801e-01 -3.06414396e-01 -6.56934679e-01 5.51297903e-01
-8.00889671e-01 5.53808436e-02 6.21726632e-01 -2.71907926e-01
-5.62248588e-01 1.52218923e-01 -1.00686453e-01 -4.19642508e-01
1.98204070e-03 5.61532497e-01 1.17904730e-02 -3.43184412e-01
2.64290571e-01 -2.23766565e-01 -1.77538902e-01 -4.24242318e-01
3.04840326e-01 6.05910242e-01 3.17536771e-01 -5.90688229e-01
4.35325503e-01 4.17527020e-01 2.73425430e-01 -3.60628128e-01
-1.48835468e+00 -8.01517665e-01 -5.60019016e-01 -3.35137963e-01
9.32154059e-01 -8.65276217e-01 -8.76533866e-01 7.15107441e-01
-1.08579278e+00 -2.10675597e-01 3.84273440e-01 3.71642321e-01
-5.10718524e-01 3.91261280e-01 -7.33613551e-01 -6.13405466e-01
-3.25007200e-01 -1.45386505e+00 1.25517118e+00 1.91961482e-01
1.38109863e-01 -9.91029143e-01 -5.60073480e-02 6.05057597e-01
2.54844546e-01 2.93825567e-01 1.07516956e+00 -5.75218081e-01
-3.06685239e-01 -1.35714486e-02 -7.57176220e-01 6.73500955e-01
-5.86897284e-02 -1.22250490e-01 -1.06832647e+00 -4.91434485e-01
-3.09369057e-01 -8.01209390e-01 1.30174899e+00 4.59126920e-01
1.74177325e+00 -2.90731937e-02 -3.68461281e-01 5.62564373e-01
1.54071820e+00 -7.50219747e-02 2.61213154e-01 4.13161278e-01
7.99752593e-01 6.99514031e-01 7.83488989e-01 2.42490023e-01
6.23737574e-01 1.01676822e+00 6.39135778e-01 -1.60257161e-01
1.31110987e-02 1.55434117e-01 3.55926082e-02 8.25670958e-01
1.04181409e-01 -3.02953124e-01 -6.94332540e-01 3.35313320e-01
-2.08926439e+00 -5.78116059e-01 -1.27726402e-02 2.31668615e+00
8.56614470e-01 9.73630548e-02 4.21110988e-02 -2.80229710e-02
8.58106971e-01 1.46973297e-01 -5.99165618e-01 -1.83025181e-01
7.48906583e-02 4.15597893e-02 6.80547118e-01 4.13056105e-01
-1.81294751e+00 6.98925495e-01 5.97954130e+00 1.05505240e+00
-1.02355909e+00 2.29612544e-01 1.09746957e+00 -1.61121786e-01
3.93864475e-02 -2.60749370e-01 -8.73153865e-01 4.55799282e-01
6.73664451e-01 3.33455771e-01 1.46123633e-01 8.22679162e-01
-1.49229765e-01 -6.72880784e-02 -1.16985583e+00 1.08582664e+00
1.77608430e-01 -1.09495187e+00 -1.75752714e-01 1.16675705e-01
8.02233160e-01 1.16510935e-01 2.15901017e-01 1.43539473e-01
3.61574709e-01 -1.05187273e+00 8.49099040e-01 2.45241106e-01
8.55815172e-01 -1.06085801e+00 8.86879027e-01 1.06224023e-01
-1.18557894e+00 -1.63111240e-01 -4.58412051e-01 3.92772049e-01
-4.38321382e-03 9.01654720e-01 -3.94515008e-01 5.50141394e-01
7.55101085e-01 7.73125768e-01 -8.17472339e-01 1.23604822e+00
-7.11184591e-02 4.58107293e-01 -1.86201259e-01 -7.76356980e-02
3.42920363e-01 -1.10212885e-01 7.51992911e-02 1.12720978e+00
1.68655038e-01 -3.44027519e-01 5.83893359e-01 3.28042507e-01
-5.10624409e-01 2.92854548e-01 -2.15971947e-01 2.90246844e-01
2.17506498e-01 1.49007988e+00 -9.77638841e-01 -7.28713870e-02
-3.98132741e-01 7.44951129e-01 8.49550843e-01 -2.83194147e-02
-9.50091839e-01 -3.79818082e-01 5.15515089e-01 -2.77938515e-01
1.66261941e-01 5.19238859e-02 -2.37821236e-01 -8.46576691e-01
8.47521722e-02 -6.77727044e-01 5.77409387e-01 -3.96456480e-01
-1.55294406e+00 4.96427923e-01 -1.08327903e-02 -1.21976781e+00
5.15446812e-02 -8.15758407e-01 -1.14650533e-01 5.65033495e-01
-1.74963689e+00 -1.14321673e+00 -1.53553069e-01 2.88607270e-01
3.28758061e-01 3.94106209e-02 5.92050850e-01 7.42743731e-01
-5.81878364e-01 9.63114798e-01 4.26587999e-01 9.58678126e-02
1.08170795e+00 -1.39180529e+00 -6.87115863e-02 3.19530785e-01
2.26471096e-01 3.53711724e-01 4.11472201e-01 -1.84244931e-01
-8.43248487e-01 -1.18205500e+00 8.55785251e-01 -2.74760425e-01
5.28813481e-01 -3.76829386e-01 -7.62227237e-01 1.44354284e-01
-2.29842253e-02 2.40262568e-01 8.62696171e-01 2.12727308e-01
-6.77271605e-01 -3.25196445e-01 -1.07276225e+00 3.90426487e-01
7.93226063e-01 -4.69224453e-01 7.05722049e-02 5.23911357e-01
5.58008492e-01 -1.94805980e-01 -9.06335354e-01 6.92789018e-01
8.94773006e-01 -9.97366309e-01 9.90224242e-01 -3.50720286e-01
6.15375102e-01 -3.54934961e-01 -2.57135272e-01 -1.34419489e+00
-3.84227991e-01 -4.92914096e-02 1.56641990e-01 1.29958665e+00
5.20010352e-01 -5.75918913e-01 6.66252911e-01 3.22026372e-01
-8.42336342e-02 -1.34265924e+00 -8.62093687e-01 -7.35183299e-01
9.90160853e-02 -2.72049993e-01 2.42242724e-01 8.63820076e-01
-4.84620214e-01 1.18271008e-01 -5.85894704e-01 6.42732978e-02
8.11235428e-01 4.01852615e-02 2.80291289e-01 -1.51856267e+00
-2.93228924e-01 -7.39783287e-01 -3.18024844e-01 -8.93260241e-01
5.19213676e-01 -1.06217146e+00 4.32836205e-01 -1.53710413e+00
7.05784440e-01 -7.78939366e-01 -9.96697903e-01 6.18453026e-01
-2.25543231e-01 6.25986338e-01 2.27752522e-01 2.70922691e-01
-1.04437768e+00 4.48363900e-01 1.03661323e+00 -4.05766875e-01
2.16988459e-01 2.14851294e-02 -8.06043506e-01 7.40551233e-01
6.72925532e-01 -7.80269921e-01 -2.68452585e-01 -4.16452110e-01
3.46986592e-01 -2.17827290e-01 3.53642225e-01 -9.26853776e-01
1.50558054e-01 -6.62891716e-02 2.85626471e-01 -5.97985744e-01
1.49245441e-01 -7.21922874e-01 -2.20495731e-01 2.65965790e-01
-7.90740430e-01 -2.81175431e-02 -1.23988852e-01 6.04755044e-01
-2.39054605e-01 -4.40466285e-01 1.17957520e+00 1.42581314e-01
-5.58745742e-01 5.31524718e-01 1.23926163e-01 3.25681157e-02
9.59926605e-01 3.15241575e-01 -3.99275899e-01 -1.31309509e-01
-3.12291503e-01 4.96890903e-01 3.76625180e-01 3.80792916e-01
3.92606318e-01 -1.60205317e+00 -6.72656655e-01 -2.94152707e-01
4.21815872e-01 -1.44340485e-01 1.84654981e-01 8.83389115e-01
-4.21278596e-01 3.20347041e-01 1.33440658e-01 -7.83443570e-01
-1.50384521e+00 3.73899728e-01 2.57057518e-01 -6.76350415e-01
-2.61260480e-01 8.89528036e-01 3.24768841e-01 -4.56237048e-01
5.05437016e-01 -5.39721511e-02 -5.01507044e-01 3.25684518e-01
4.53733921e-01 2.66025066e-01 2.33353645e-01 -6.08259082e-01
-5.19810796e-01 8.77299726e-01 -4.18930024e-01 8.02932829e-02
1.13732815e+00 -1.00068659e-01 -3.33388418e-01 4.46658939e-01
1.70124197e+00 -3.85849118e-01 -1.30113196e+00 -2.82664180e-01
2.29756683e-01 -3.00971568e-01 2.24653989e-01 -7.23379254e-01
-1.16637242e+00 8.19771528e-01 8.10801864e-01 2.01493889e-01
1.06012309e+00 7.79986307e-02 6.66738212e-01 3.69570881e-01
4.05607492e-01 -1.16852880e+00 1.77080303e-01 4.65472460e-01
5.98141611e-01 -1.66539252e+00 2.35437043e-02 -3.42864990e-01
-5.53293586e-01 9.24802780e-01 5.07927120e-01 7.46772066e-02
7.28961706e-01 -1.91777721e-01 2.59842038e-01 -1.17714494e-01
-5.88663042e-01 -1.35100454e-01 6.35460854e-01 8.53246003e-02
5.25131583e-01 2.16267109e-01 -5.53587735e-01 2.40011185e-01
3.51729542e-01 -2.54775256e-01 1.35214120e-01 5.24163604e-01
-3.79050791e-01 -1.23676431e+00 -6.72902074e-03 7.89219022e-01
-7.70980656e-01 -2.39737071e-02 -2.34404653e-01 4.48072135e-01
1.97306633e-01 1.09634280e+00 -1.88300818e-01 -4.06110555e-01
2.79606655e-02 -8.15142244e-02 4.20011699e-01 -5.38713574e-01
-5.43302417e-01 -9.48879719e-02 -2.31752414e-02 -4.82485324e-01
-6.33183420e-01 -6.74684405e-01 -9.38028336e-01 5.01239710e-02
-6.51269436e-01 -3.77060310e-03 7.89613128e-01 1.07635999e+00
2.00986356e-01 3.59234184e-01 8.77827227e-01 -8.35475564e-01
-7.08811283e-01 -8.02771091e-01 -5.48223853e-01 5.00273049e-01
2.83472419e-01 -1.08302712e+00 -3.88162881e-01 -4.40632820e-01] | [9.514890670776367, 4.022329330444336] |
7e9669a8-ec64-4552-8470-787bdad51ce6 | see-plan-predict-language-guided-cognitive | 2210.03825 | null | https://arxiv.org/abs/2210.03825v1 | https://arxiv.org/pdf/2210.03825v1.pdf | See, Plan, Predict: Language-guided Cognitive Planning with Video Prediction | Cognitive planning is the structural decomposition of complex tasks into a sequence of future behaviors. In the computational setting, performing cognitive planning entails grounding plans and concepts in one or more modalities in order to leverage them for low level control. Since real-world tasks are often described in natural language, we devise a cognitive planning algorithm via language-guided video prediction. Current video prediction models do not support conditioning on natural language instructions. Therefore, we propose a new video prediction architecture which leverages the power of pre-trained transformers.The network is endowed with the ability to ground concepts based on natural language input with generalization to unseen objects. We demonstrate the effectiveness of this approach on a new simulation dataset, where each task is defined by a high-level action described in natural language. Our experiments compare our method again stone video generation baseline without planning or action grounding and showcase significant improvements. Our ablation studies highlight an improved generalization to unseen objects that natural language embeddings offer to concept grounding ability, as well as the importance of planning towards visual "imagination" of a task. | ['Animesh Garg', 'Igor Gilitschenski', 'Wei Yu', 'Ziyi Zhou', 'Advaya Gupta', 'Maria Attarian'] | 2022-10-07 | null | null | null | null | ['video-generation', 'video-prediction'] | ['computer-vision', 'computer-vision'] | [ 6.10796392e-01 7.04238772e-01 -1.52782932e-01 -9.58848447e-02
-2.58862764e-01 -5.65749466e-01 1.28786922e+00 -4.68486287e-02
-2.83647120e-01 4.37567383e-01 1.14505255e+00 -2.95431256e-01
1.22042000e-01 -7.97546804e-01 -1.05410767e+00 -3.58714670e-01
-3.44211251e-01 3.73755962e-01 3.73352729e-02 -2.45266899e-01
2.16456950e-01 1.27552226e-01 -1.70822728e+00 7.69254863e-01
3.68517399e-01 7.64582753e-01 5.16987920e-01 8.09846938e-01
2.23185122e-01 1.59668899e+00 -1.12778947e-01 -1.51369467e-01
4.03620452e-01 -5.32582283e-01 -8.76492977e-01 4.31205690e-01
4.77437854e-01 -6.05513752e-01 -5.95430017e-01 5.93902469e-01
-7.65350237e-02 6.20962501e-01 5.46986043e-01 -1.47772300e+00
-9.01944876e-01 5.90727508e-01 2.16452733e-01 1.13871954e-02
9.20219421e-01 7.11057186e-01 1.10624301e+00 -6.81819677e-01
1.05439246e+00 1.35879254e+00 3.26970935e-01 9.44803417e-01
-1.44809198e+00 -1.29553109e-01 6.13903522e-01 2.58518696e-01
-9.20398772e-01 -4.27946031e-01 5.34225166e-01 -7.88528860e-01
1.34764957e+00 -2.50170022e-01 1.10184705e+00 1.57537532e+00
3.64955932e-01 1.01256859e+00 9.08076406e-01 -3.44859302e-01
5.03561020e-01 -2.46360838e-01 -1.81324005e-01 1.10522568e+00
-5.26497699e-02 6.26620114e-01 -1.14126849e+00 2.34726936e-01
7.96812713e-01 1.97283372e-01 -4.84664559e-01 -7.75653541e-01
-1.62095964e+00 7.21874952e-01 4.27293509e-01 -2.87193824e-02
-5.62706351e-01 5.88719845e-01 2.64879972e-01 1.74013346e-01
8.53537694e-02 8.15315902e-01 -1.40606135e-01 -1.69951111e-01
-7.41846144e-01 4.17918980e-01 6.91822588e-01 1.32277822e+00
7.54221082e-01 8.36490691e-02 -6.51302695e-01 -9.31194797e-02
1.56406179e-01 1.51844531e-01 6.09283507e-01 -1.52203500e+00
4.06770259e-01 6.40208960e-01 2.58308381e-01 -8.93498182e-01
-4.64406282e-01 -7.86225572e-02 -3.09283286e-01 5.20143747e-01
1.93342716e-01 -9.38623329e-04 -1.02649319e+00 1.97193527e+00
-2.30445236e-01 4.25573289e-01 3.49476308e-01 8.81631017e-01
2.34558597e-01 6.88479125e-01 5.01973510e-01 1.89527012e-02
1.27659202e+00 -1.13200796e+00 -4.31647092e-01 -5.86460829e-01
5.53239882e-01 8.45182389e-02 1.54711759e+00 3.23086619e-01
-1.04600358e+00 -6.24864995e-01 -1.06798470e+00 -2.25565419e-01
-2.26608783e-01 -2.81455010e-01 7.63741612e-01 2.12611407e-01
-1.49725235e+00 6.59411371e-01 -1.08669531e+00 -6.14192843e-01
6.72997355e-01 8.53675455e-02 -5.75122535e-01 -2.48529106e-01
-8.95811439e-01 9.39879954e-01 5.49424708e-01 -3.13389271e-01
-1.59887159e+00 -6.55118942e-01 -1.20256233e+00 2.82531142e-01
6.09281600e-01 -1.21467018e+00 1.29372334e+00 -1.15526021e+00
-1.36037183e+00 7.35150933e-01 -6.73805699e-02 -9.61938024e-01
2.84075886e-01 -3.20583582e-01 5.98413236e-02 4.24415529e-01
8.95246640e-02 1.28868210e+00 7.97427237e-01 -1.12314188e+00
-5.72816432e-01 -2.96541248e-02 8.26299548e-01 5.07245243e-01
-1.44541517e-01 -5.20487905e-01 -1.95335627e-01 -5.19913435e-01
-1.59364194e-01 -1.05972755e+00 -2.94454038e-01 1.10208601e-01
-8.36950094e-02 -1.41935527e-01 4.33861464e-01 -5.05713761e-01
8.71680260e-01 -2.06758881e+00 4.36463803e-01 -1.99127182e-01
3.84846151e-01 -2.69354105e-01 -3.67215425e-01 5.95975280e-01
9.21728928e-03 1.84747204e-01 -4.01041061e-02 -2.51643866e-01
3.69680136e-01 3.03644687e-01 -5.29818833e-01 1.75705999e-01
4.78873998e-01 1.25303876e+00 -1.15673411e+00 -2.74130851e-01
1.61082342e-01 3.96500766e-01 -1.21681154e+00 3.80472869e-01
-8.36492717e-01 3.67144406e-01 -3.88410270e-01 4.20833260e-01
-1.65153295e-01 -2.59501845e-01 3.38985294e-01 -1.00871259e-02
3.24381255e-02 3.60168666e-01 -5.78414202e-01 2.23297215e+00
-4.93787944e-01 6.59751952e-01 -3.19702774e-01 -7.85097957e-01
3.07103992e-01 3.95579904e-01 3.05793911e-01 -6.14072263e-01
-2.45028615e-01 -4.78819400e-01 -4.17159386e-02 -6.07623756e-01
3.89562637e-01 -3.72773230e-01 -1.56186685e-01 7.24385202e-01
2.62227237e-01 -2.43986532e-01 9.77386832e-02 5.46428561e-01
1.46416593e+00 9.94155169e-01 3.98645312e-01 -2.71542251e-01
8.82042944e-03 4.92802888e-01 2.73468375e-01 9.22669351e-01
-3.77384812e-01 3.64202172e-01 5.24043322e-01 -6.60461068e-01
-9.31356192e-01 -1.06630826e+00 6.18148804e-01 1.57059443e+00
-1.14413816e-02 -6.83836043e-01 -7.82415926e-01 -4.71185952e-01
-2.80619830e-01 1.13582551e+00 -8.76111329e-01 -4.72272038e-01
-7.20374763e-01 -1.10474668e-01 2.48164609e-01 8.08323801e-01
2.79119849e-01 -1.47727454e+00 -1.31812978e+00 1.50225073e-01
-1.15974970e-01 -1.33954608e+00 -4.40073937e-01 5.72514236e-02
-7.48472512e-01 -1.04399562e+00 -2.98002392e-01 -7.58311927e-01
5.63691258e-01 2.86201358e-01 1.51457489e+00 1.23696670e-01
7.28857964e-02 1.35411346e+00 -4.55121458e-01 -2.65914977e-01
-4.63342637e-01 -2.32436225e-01 1.90614492e-01 -1.48752958e-01
2.96406001e-01 -6.64767623e-01 -5.17727017e-01 -1.47831306e-01
-8.45158160e-01 7.21353769e-01 5.62545121e-01 7.95956254e-01
3.60144764e-01 -1.57529354e-01 6.61667362e-02 -5.75183570e-01
6.68593466e-01 -3.51732999e-01 -3.71815264e-01 1.76461086e-01
-3.90053213e-01 4.83021200e-01 4.56472337e-01 -3.45481545e-01
-1.11501002e+00 4.17666733e-01 4.68559474e-01 -5.07541180e-01
-3.99252295e-01 6.13114953e-01 -1.37919143e-01 2.47883543e-01
6.91888094e-01 3.87972653e-01 4.81035523e-02 1.89631298e-01
7.95096934e-01 -2.80731618e-01 6.10017955e-01 -9.06446099e-01
7.53790736e-01 4.90372390e-01 1.88029051e-01 -5.49368978e-01
-7.55293310e-01 -6.10510912e-03 -6.50318384e-01 -3.07146937e-01
1.36204338e+00 -1.04857850e+00 -9.36530590e-01 -2.33935669e-01
-1.12457049e+00 -8.42726946e-01 -5.67184567e-01 3.46586406e-01
-1.35134697e+00 7.03751743e-02 -4.08972502e-01 -5.54624140e-01
1.86128184e-01 -1.20066500e+00 1.03434062e+00 -1.74527124e-01
-3.98588032e-01 -9.21311736e-01 -9.05090421e-02 1.27037123e-01
2.43549168e-01 2.86441386e-01 1.14694726e+00 -4.77374166e-01
-1.18516111e+00 1.77304104e-01 6.89797029e-02 -1.15524940e-01
-3.88294607e-01 -5.61498821e-01 -8.92756879e-01 -9.98853594e-02
-3.69865424e-03 -7.36096263e-01 9.15418983e-01 1.70644283e-01
1.09089434e+00 -4.15120423e-01 -3.29291165e-01 4.41970050e-01
1.31398439e+00 1.84656918e-01 6.87773705e-01 4.29034770e-01
5.94591200e-01 7.27107167e-01 4.44070786e-01 3.24541867e-01
4.57911432e-01 5.87536395e-01 6.59413218e-01 4.23839927e-01
-1.27883300e-01 -8.46259594e-01 8.25240374e-01 1.03219174e-01
-3.45751882e-01 -2.60097474e-01 -1.19247925e+00 2.93951213e-01
-1.81532979e+00 -1.47633195e+00 6.69684708e-01 1.94701529e+00
5.96430898e-01 2.36984670e-01 8.65943264e-03 -3.70174229e-01
1.83595121e-01 2.15423077e-01 -4.67738986e-01 -1.95611894e-01
2.22127542e-01 7.65934810e-02 1.05014920e-01 6.52972579e-01
-1.04437745e+00 1.19531572e+00 6.43684912e+00 1.49062976e-01
-7.21200764e-01 1.04678616e-01 4.17641014e-01 -2.09925383e-01
-3.42534333e-01 8.57125223e-02 -4.41003501e-01 -8.54849666e-02
1.07777023e+00 -2.95166552e-01 7.58103371e-01 7.07111478e-01
2.11623415e-01 -1.24027446e-01 -1.84816194e+00 7.21427262e-01
3.10107201e-01 -1.61767852e+00 4.47087824e-01 3.59433778e-02
4.03054833e-01 -1.87591344e-01 5.68448417e-02 8.43554616e-01
6.41080260e-01 -1.34284675e+00 1.06715059e+00 7.07631469e-01
5.79244435e-01 -2.29575172e-01 6.46024197e-02 5.37516236e-01
-1.03527617e+00 -4.28162068e-01 -2.49414816e-01 -5.68363130e-01
3.39687228e-01 -2.77435958e-01 -9.84966755e-01 6.26549423e-02
2.03968361e-01 9.05649722e-01 -5.31489372e-01 3.73965144e-01
-4.27728504e-01 4.26275402e-01 8.72449577e-02 5.97545654e-02
4.32780564e-01 -4.31890041e-02 3.80177706e-01 1.01197350e+00
3.42033446e-01 4.29189742e-01 6.10511899e-01 9.17183697e-01
1.19056385e-02 -3.65918368e-01 -1.04161537e+00 -4.15558487e-01
8.95080566e-02 6.35150850e-01 -6.20509982e-01 -4.01726097e-01
-5.70740104e-01 1.06857193e+00 4.33533847e-01 5.60156941e-01
-7.70567358e-01 3.71258140e-01 8.00410271e-01 3.31137985e-01
1.52313262e-01 -5.09067714e-01 -3.53031270e-02 -1.27009404e+00
-2.00670019e-01 -9.05650258e-01 1.79837465e-01 -1.28128421e+00
-6.85069978e-01 5.67515731e-01 1.19158126e-01 -1.19024742e+00
-6.59533024e-01 -9.64964569e-01 -3.51038218e-01 4.45047081e-01
-1.11439061e+00 -1.32513678e+00 -3.58319074e-01 7.45768428e-01
9.72080350e-01 -3.62726390e-01 1.07583880e+00 -4.53784078e-01
-3.71601768e-02 7.02927960e-03 -7.03999817e-01 -2.57059895e-02
2.20476866e-01 -1.21058083e+00 3.65232587e-01 9.17094827e-01
3.58306050e-01 6.68667436e-01 7.61046529e-01 -5.96750438e-01
-1.54444087e+00 -1.00557876e+00 6.46471918e-01 -9.19150829e-01
7.19052672e-01 -7.47478545e-01 -4.28306937e-01 1.31582630e+00
5.49153268e-01 -9.36453044e-02 6.87021434e-01 7.41066858e-02
-4.22770202e-01 4.40144390e-01 -6.96521878e-01 1.03658032e+00
1.66640735e+00 -8.20040107e-01 -1.02889752e+00 4.22039151e-01
1.15315747e+00 -2.37381682e-01 -4.01550978e-01 -1.18673220e-01
5.57933509e-01 -1.04743862e+00 1.09841979e+00 -1.11782181e+00
1.04639685e+00 -2.20505849e-01 -4.08957511e-01 -1.32781887e+00
-6.12989962e-01 -5.81755936e-01 -2.45962471e-01 6.07665896e-01
2.27223337e-01 3.50411562e-03 9.49213624e-01 9.34597731e-01
-4.74529296e-01 -4.00135845e-01 -5.81453741e-01 -6.58034444e-01
-1.21784955e-01 -7.00484931e-01 4.63912815e-01 6.65289342e-01
5.19395828e-01 4.77735400e-01 -3.61869663e-01 6.76256046e-02
3.27060163e-01 9.22161788e-02 6.82927310e-01 -9.42467391e-01
-4.95072037e-01 -2.40404293e-01 -6.66947484e-01 -1.23361862e+00
6.94113374e-01 -1.07030332e+00 2.58877844e-01 -1.61807156e+00
2.51796901e-01 1.89650550e-01 -2.14074776e-01 6.30732834e-01
1.76844731e-01 -1.41055197e-01 6.84903443e-01 1.44057080e-01
-7.87804723e-01 6.86461449e-01 1.35384262e+00 -2.16889754e-01
-6.31577820e-02 -3.80339026e-01 -7.32881308e-01 7.47818947e-01
5.83872378e-01 -1.47095144e-01 -1.08288801e+00 -7.47613072e-01
3.33110988e-01 2.85371572e-01 6.85791075e-01 -1.35799205e+00
2.33732060e-01 -4.96909946e-01 1.93094775e-01 1.96304452e-02
4.52651441e-01 -9.90023553e-01 -6.48634732e-02 5.38902938e-01
-8.05011153e-01 2.69023657e-01 1.76914021e-01 9.93250132e-01
-4.25645597e-02 1.20849103e-01 2.39470318e-01 -5.14854372e-01
-1.39601147e+00 2.28125796e-01 -8.06036413e-01 -1.72713734e-02
1.13916838e+00 -3.94989759e-01 -3.96972239e-01 -7.19234228e-01
-1.16235077e+00 1.45654291e-01 6.26710892e-01 4.01055425e-01
8.04212451e-01 -1.28022683e+00 -3.69651735e-01 2.43928611e-01
3.63583058e-01 -3.56550545e-01 -3.58298793e-02 5.71776688e-01
-4.30600017e-01 5.28140306e-01 -6.66057050e-01 -3.52149636e-01
-6.30437434e-01 1.05201972e+00 3.00499737e-01 -1.04286909e-01
-8.36011291e-01 8.17432761e-01 7.31657684e-01 -1.94094896e-01
2.01261103e-01 -6.72951221e-01 -2.20565081e-01 -1.84018880e-01
5.40944278e-01 -8.59804600e-02 -3.89929473e-01 -4.80162323e-01
-1.30063787e-01 4.97104898e-02 2.36603945e-01 -4.26466852e-01
1.27898419e+00 -1.13351651e-01 2.31838867e-01 2.89761931e-01
7.06546605e-01 -3.61818641e-01 -1.90611923e+00 9.48107317e-02
1.00842968e-01 -2.34026700e-01 -4.38661352e-02 -8.46709490e-01
-5.41729271e-01 1.03349137e+00 3.09258074e-01 -2.16930285e-01
1.06260610e+00 -2.16162708e-02 2.78648168e-01 9.01414096e-01
7.81006694e-01 -8.66960585e-01 7.23871827e-01 7.03587830e-01
1.10989833e+00 -1.15724468e+00 -1.34994507e-01 -1.39082536e-01
-1.03804731e+00 1.04745889e+00 7.95251012e-01 -1.71844557e-01
3.74116808e-01 1.72201127e-01 -3.05984825e-01 -3.03472936e-01
-1.26924992e+00 -2.87644118e-01 1.02152333e-01 1.01550961e+00
2.66089559e-01 5.55302911e-02 2.66166776e-01 6.61958635e-01
-1.75442442e-01 3.14111859e-01 8.20002615e-01 1.15835071e+00
-5.44992447e-01 -6.55003428e-01 2.75234468e-02 2.94492483e-01
9.77895707e-02 -3.16257209e-01 -2.55995721e-01 7.85824060e-01
7.84048215e-02 7.52714217e-01 2.81769156e-01 -5.42552531e-01
1.63609922e-01 5.25030136e-01 6.34965479e-01 -1.11618793e+00
-2.43569806e-01 -3.28900576e-01 8.06502625e-02 -1.24835670e+00
-6.64978743e-01 -7.56078482e-01 -1.32025599e+00 -1.01026721e-01
6.03544474e-01 -2.96892673e-01 4.21739407e-02 1.03622329e+00
3.96755397e-01 5.55290580e-01 -1.89996243e-01 -1.29564881e+00
-3.24374557e-01 -5.33133984e-01 -2.41965413e-01 7.19182432e-01
3.32510859e-01 -8.18549454e-01 -7.38233328e-02 7.27148116e-01] | [4.452603340148926, 0.832120954990387] |
311ee02c-04a9-4db5-9db6-61031b9d7fc5 | bagging-bert-models-for-robust-aggression | null | null | https://aclanthology.org/2020.trac-1.9 | https://aclanthology.org/2020.trac-1.9.pdf | Bagging BERT Models for Robust Aggression Identification | Modern transformer-based models with hundreds of millions of parameters, such as BERT, achieve impressive results at text classification tasks. This also holds for aggression identification and offensive language detection, where deep learning approaches consistently outperform less complex models, such as decision trees. While the complex models fit training data well (low bias), they also come with an unwanted high variance. Especially when fine-tuning them on small datasets, the classification performance varies significantly for slightly different training data. To overcome the high variance and provide more robust predictions, we propose an ensemble of multiple fine-tuned BERT models based on bootstrap aggregating (bagging). In this paper, we describe such an ensemble system and present our submission to the shared tasks on aggression identification 2020 (team name: Julian). Our submission is the best-performing system for five out of six subtasks. For example, we achieve a weighted F1-score of 80.3{\%} for task A on the test dataset of English social media posts. In our experiments, we compare different model configurations and vary the number of models used in the ensemble. We find that the F1-score drastically increases when ensembling up to 15 models, but the returns diminish for more models. | ['Julian Risch', 'Ralf Krestel'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['aggression-identification'] | ['natural-language-processing'] | [-3.75581086e-01 -1.44935817e-01 4.31985967e-02 -4.12523419e-01
-9.29966867e-01 -4.81493741e-01 4.97708052e-01 1.82505280e-01
-6.61438823e-01 7.27030158e-01 2.51440048e-01 -1.39922038e-01
-2.03271404e-01 -4.76351202e-01 -2.08977267e-01 -3.86217207e-01
-8.65431726e-02 9.36624527e-01 2.61764020e-01 -6.44134045e-01
2.64109433e-01 2.43769601e-01 -1.22845137e+00 5.87607622e-01
6.93060577e-01 7.23694921e-01 -5.48073888e-01 7.52890289e-01
1.70875221e-01 8.91245008e-01 -8.83743823e-01 -1.09989917e+00
-3.72913107e-02 -3.16533260e-02 -1.05967796e+00 -5.98684251e-01
5.39374769e-01 -5.44494808e-01 -2.12800220e-01 8.12102914e-01
8.12397420e-01 3.23183611e-02 8.18278432e-01 -1.34628952e+00
-6.36949986e-02 1.22777259e+00 -6.25472188e-01 4.74930525e-01
1.81814775e-01 1.66793495e-01 1.28596544e+00 -7.44638741e-01
6.41938150e-02 1.38033247e+00 1.07300556e+00 7.46462464e-01
-1.34424460e+00 -1.15466189e+00 -9.17066634e-02 2.56219923e-01
-1.33307731e+00 -7.05621183e-01 4.72301602e-01 -6.13736391e-01
1.07592118e+00 2.61939317e-01 4.28732663e-01 1.79700339e+00
1.42322080e-02 7.78972268e-01 8.42312932e-01 3.86023149e-02
4.74328734e-02 9.64204296e-02 6.11187458e-01 6.16513133e-01
2.85964727e-01 -7.57487267e-02 -7.89594710e-01 -7.23995030e-01
4.33532372e-02 -4.04956907e-01 2.16590494e-01 1.66173756e-01
-7.58998334e-01 1.15439916e+00 5.19646853e-02 2.53844798e-01
-1.41373858e-01 1.49813831e-01 8.50351095e-01 2.23062098e-01
5.78767538e-01 6.39425099e-01 -5.06221116e-01 -6.48523211e-01
-8.92849267e-01 6.23385668e-01 1.04061782e+00 4.65872854e-01
1.15923963e-01 1.35964945e-01 -3.94719779e-01 1.41800117e+00
-2.28960797e-01 1.91742942e-01 7.83805251e-01 -8.56325626e-01
5.43234527e-01 2.48040661e-01 -2.35232245e-02 -8.36491883e-01
-8.07451427e-01 -7.26002753e-01 -7.99347162e-01 -7.80934989e-02
7.02082336e-01 -3.45503271e-01 -3.97098839e-01 1.87564695e+00
4.34304476e-02 -1.47884106e-02 -3.94731402e-01 4.88851756e-01
8.16601515e-01 2.06521064e-01 3.58874112e-01 4.94593903e-02
1.08669615e+00 -9.00973618e-01 -3.65021408e-01 -3.53907138e-01
1.00738788e+00 -5.69002390e-01 1.01216578e+00 7.50795782e-01
-1.04257655e+00 -2.27212176e-01 -8.40697527e-01 1.07070938e-01
-2.97340780e-01 -2.50721097e-01 6.39318168e-01 9.42536592e-01
-9.29716945e-01 6.86877847e-01 -5.52857459e-01 -3.06507438e-01
3.75370294e-01 4.82579648e-01 -3.51840943e-01 4.33818698e-01
-1.31003308e+00 1.03308582e+00 2.02379093e-01 -3.61002207e-01
-8.67580056e-01 -6.72021687e-01 -3.11756194e-01 2.30697110e-01
6.81034178e-02 -6.15603507e-01 1.50612915e+00 -5.78907013e-01
-1.28817940e+00 9.13230062e-01 1.85894445e-01 -7.03886688e-01
7.65926421e-01 -3.86031508e-01 -2.10090905e-01 -4.35007751e-01
-1.19120013e-02 3.54194492e-01 6.77119672e-01 -7.39940405e-01
-5.43819487e-01 -3.89270335e-01 2.81742476e-02 5.83858043e-03
-8.48524749e-01 6.55889153e-01 5.94371706e-02 -5.46761215e-01
-5.51679552e-01 -8.76658320e-01 -7.29053319e-02 -5.99810898e-01
-4.89293426e-01 -4.50577587e-01 4.37423646e-01 -7.74749756e-01
1.69918203e+00 -1.96876907e+00 1.36121348e-01 -4.80185123e-03
4.50699538e-01 2.73857445e-01 2.62269545e-02 3.55065107e-01
-1.63514420e-01 2.93958634e-01 -4.24134657e-02 -6.86456144e-01
1.04936749e-01 -1.78123284e-02 -1.76511198e-01 1.56370178e-01
-7.38025382e-02 6.34889007e-01 -6.92654431e-01 -3.45529914e-01
-2.13078529e-01 3.45528983e-02 -9.48699594e-01 7.07705989e-02
2.30057359e-01 2.40645736e-01 -2.58677661e-01 3.37152123e-01
2.53534436e-01 -1.24283709e-01 -8.94939303e-02 2.78961629e-01
2.25022420e-01 5.47736943e-01 -7.66261280e-01 1.08209300e+00
-4.80231076e-01 5.14196754e-01 1.57196730e-01 -9.18302059e-01
7.71470249e-01 1.18390448e-01 4.13957268e-01 -2.32219234e-01
2.21722364e-01 2.84374952e-01 4.77005333e-01 -3.58481437e-01
4.39567775e-01 -2.40202054e-01 -3.46452236e-01 5.68727791e-01
1.73203461e-02 -1.00961484e-01 1.38769343e-01 3.49315464e-01
1.51311946e+00 -5.40218294e-01 3.30865443e-01 -1.06777817e-01
2.85831422e-01 -2.62051195e-01 3.09127033e-01 1.02973616e+00
-3.78754169e-01 3.49798530e-01 6.90618157e-01 -5.50351858e-01
-1.09392321e+00 -6.94430172e-01 -4.91284758e-01 1.70151103e+00
-5.39390206e-01 -8.82164776e-01 -9.50404584e-01 -7.69131601e-01
2.14065164e-01 9.60587204e-01 -6.23215675e-01 -4.92520452e-01
-4.49800700e-01 -1.37188745e+00 1.30217874e+00 5.66553593e-01
2.41606921e-01 -8.70531917e-01 -1.98714361e-01 1.97375029e-01
-5.06152749e-01 -9.50870454e-01 -4.37819511e-02 2.01207966e-01
-5.39801419e-01 -8.45810294e-01 -2.66481996e-01 -1.45771474e-01
-1.46621600e-01 -3.14460188e-01 1.29014778e+00 2.11124673e-01
-1.34378701e-01 -8.48259777e-02 -3.70786041e-01 -7.39446938e-01
-6.22401953e-01 4.26469833e-01 5.12668133e-01 -1.16792470e-01
6.78915203e-01 -5.73228776e-01 -1.31530836e-01 4.41971928e-01
-2.49243289e-01 -4.30279411e-02 1.43214434e-01 1.23099542e+00
-2.74920970e-01 -1.49207219e-01 7.01878548e-01 -8.41576874e-01
1.07166266e+00 -6.91431105e-01 -1.39630765e-01 -2.05979347e-01
-4.17202741e-01 -2.42544934e-01 7.67907143e-01 -5.76114297e-01
-5.75161397e-01 -4.07223493e-01 -5.35180509e-01 -2.35448793e-01
-1.18635535e-01 2.66809434e-01 3.91542047e-01 8.30774978e-02
1.07408834e+00 -1.97479546e-01 -1.65983751e-01 -7.49491394e-01
-7.13714361e-02 1.09343147e+00 1.59995914e-01 -6.42289042e-01
6.67212129e-01 -6.20301962e-02 -2.89643735e-01 -6.15269721e-01
-1.15317273e+00 -3.27268511e-01 -4.63236958e-01 -1.01282082e-01
6.71374261e-01 -7.85192370e-01 -1.06257129e+00 7.21752107e-01
-1.00460804e+00 -5.06316662e-01 1.04626544e-01 3.54596823e-01
-4.79430914e-01 2.05982402e-01 -8.95136416e-01 -9.90064681e-01
-6.47568643e-01 -1.00891805e+00 1.10302210e+00 -2.89630145e-01
-8.56076837e-01 -7.27729023e-01 2.27537721e-01 7.28827178e-01
5.63058734e-01 -7.62418509e-02 1.06014717e+00 -1.32493091e+00
5.30365467e-01 -3.75604570e-01 -7.82501698e-02 3.35513115e-01
-2.28662387e-01 1.45638566e-02 -1.32731891e+00 -2.82521784e-01
-3.32145602e-01 -7.94017136e-01 9.58180666e-01 1.81489691e-01
1.69693148e+00 -3.23875457e-01 -2.27195144e-01 3.70868921e-01
5.41185558e-01 -9.90226492e-02 3.37183774e-01 4.55033123e-01
6.68971002e-01 6.02765501e-01 2.50830024e-01 7.29977608e-01
3.92906994e-01 1.07648253e+00 2.95005679e-01 2.47761816e-01
1.90158620e-01 -2.51383275e-01 4.66662854e-01 6.79970562e-01
-1.79722026e-01 -2.09124774e-01 -1.28505278e+00 1.93307385e-01
-1.74474967e+00 -1.26843917e+00 -9.74760354e-02 2.21895385e+00
9.71352100e-01 4.31421727e-01 8.98179114e-01 3.28252912e-01
5.60499191e-01 -6.35405928e-02 -4.98541802e-01 -8.49919081e-01
-8.02567825e-02 2.50698000e-01 2.43340746e-01 5.50019622e-01
-1.18420005e+00 1.04332650e+00 7.31394482e+00 1.21717536e+00
-9.03153896e-01 3.20367038e-01 8.91215682e-01 -6.38342798e-01
2.34307036e-01 -5.00848532e-01 -9.85214055e-01 6.91362321e-01
1.45223844e+00 -3.55820388e-01 4.74957138e-01 9.19775784e-01
9.04176980e-02 2.07690224e-01 -1.12116945e+00 1.14184618e+00
9.88885611e-02 -7.95250595e-01 -4.45639268e-02 2.39803597e-01
3.82558346e-01 3.88772726e-01 7.78400004e-02 8.57631683e-01
7.26678193e-01 -1.40810466e+00 7.01357245e-01 2.47754216e-01
4.79940742e-01 -7.17702985e-01 8.36837351e-01 8.09766293e-01
-3.75469059e-01 -7.34074533e-01 -3.32967252e-01 -2.52545297e-01
-9.77258831e-02 5.04759014e-01 -8.86005878e-01 4.70068417e-02
1.04876161e+00 4.85416770e-01 -8.86544645e-01 9.97184813e-01
2.06550792e-01 1.12132692e+00 -4.88962054e-01 -4.91285831e-01
4.99628633e-02 1.01830937e-01 6.76815450e-01 1.30002582e+00
3.89823541e-02 3.93091217e-02 -2.63618231e-02 4.29243743e-01
-1.69368848e-01 3.53328168e-01 -5.78026593e-01 1.34307474e-01
3.40375155e-01 1.32817400e+00 6.23447858e-02 -3.86677176e-01
-4.63954546e-02 5.91526508e-01 9.74139929e-01 -1.48308367e-01
-9.49621856e-01 -6.97523728e-02 6.84442818e-01 1.48691252e-01
-2.67732233e-01 5.97256534e-02 -3.97634923e-01 -1.12343919e+00
-2.49244303e-01 -1.23318481e+00 6.89443171e-01 -5.90771079e-01
-1.68962789e+00 5.78786731e-01 2.05940485e-01 -8.66024435e-01
-5.84863245e-01 -7.54618883e-01 -5.69309413e-01 6.01156235e-01
-6.10999525e-01 -1.00699699e+00 -9.90649536e-02 3.23565573e-01
5.93198538e-01 -5.00115097e-01 9.88000929e-01 4.79350507e-01
-9.38180149e-01 1.12011302e+00 1.37964725e-01 3.90851527e-01
1.09062374e+00 -1.27172828e+00 4.32526469e-01 3.30986887e-01
9.15014967e-02 5.85267246e-01 8.68247867e-01 -3.24687392e-01
-6.04377508e-01 -8.41970086e-01 7.98139036e-01 -1.12726998e+00
1.04962862e+00 -6.11890435e-01 -7.99572706e-01 7.49620497e-01
-1.54808477e-01 -4.30306882e-01 9.88321602e-01 8.16103399e-01
-7.40209401e-01 -3.62707451e-02 -1.06513548e+00 6.04033947e-01
1.23053277e+00 -2.74597645e-01 -4.45507586e-01 5.84654689e-01
1.93702847e-01 -2.37179026e-01 -8.50477993e-01 4.24913615e-01
8.89127612e-01 -1.38338125e+00 8.60698164e-01 -1.40623009e+00
6.79566383e-01 5.47776759e-01 -1.70138374e-01 -1.56250560e+00
-6.90106213e-01 -4.36479032e-01 -8.96536335e-02 1.15017164e+00
6.40764832e-01 -7.11753964e-01 6.69244409e-01 7.84961462e-01
1.24705032e-01 -8.31885159e-01 -1.01512003e+00 -7.23303199e-01
6.24191463e-01 -5.27456284e-01 5.35140812e-01 1.04223394e+00
3.87037009e-01 6.87913835e-01 -6.00929558e-01 -3.69327158e-01
5.49119651e-01 -4.30385828e-01 1.05352747e+00 -1.50969446e+00
-6.32943034e-01 -8.84348869e-01 -4.72478807e-01 -5.24718463e-01
4.49002951e-01 -8.87614191e-01 -4.77596492e-01 -8.22758973e-01
7.40442455e-01 -5.51311255e-01 -1.99744940e-01 8.32425058e-01
-3.48639518e-01 4.36312944e-01 2.86916140e-02 3.92806083e-02
-5.37635028e-01 4.13707495e-01 5.63120961e-01 -1.68285504e-01
3.61726321e-02 1.94202915e-01 -9.87526298e-01 9.27461922e-01
1.05289721e+00 -5.05344570e-01 -3.06113679e-02 -3.69159311e-01
1.80782586e-01 -2.26629302e-01 2.63248861e-01 -9.68366742e-01
-8.34337771e-02 -1.12885535e-02 4.35479164e-01 -2.29564741e-01
6.89624012e-01 -1.91649005e-01 -4.52714115e-02 4.89274830e-01
-6.62761211e-01 1.56435907e-01 1.84859902e-01 1.88476384e-01
8.34472403e-02 -3.07610720e-01 8.33699524e-01 3.96935605e-02
-3.11244503e-02 2.15241909e-01 -4.73959655e-01 2.61313289e-01
7.44189143e-01 9.77044553e-02 -6.05019689e-01 -7.41784871e-01
-7.89807498e-01 1.76015332e-01 5.97426407e-02 6.15607381e-01
6.71538413e-02 -1.01832044e+00 -1.18036425e+00 6.93205325e-03
2.65827954e-01 -5.54820418e-01 8.73824731e-02 1.03352427e+00
-1.70644552e-01 2.09948882e-01 -1.58922866e-01 -3.73339534e-01
-1.49432671e+00 1.56920418e-01 6.35232747e-01 -5.19562125e-01
-2.13453054e-01 1.11796868e+00 9.59157348e-02 -6.81675017e-01
1.18115149e-01 3.53519738e-01 -4.29730475e-01 1.68752462e-01
7.80302286e-01 6.81254685e-01 3.03802073e-01 -7.65659750e-01
-3.98089886e-01 1.06000699e-01 -5.36204934e-01 -4.32587340e-02
1.25309348e+00 4.29633558e-01 -8.17101598e-02 4.35030609e-01
9.57102537e-01 1.68140307e-01 -3.58546555e-01 -1.23772083e-03
-5.80486469e-02 -4.04314727e-01 1.77267417e-02 -9.47609127e-01
-6.52480781e-01 9.21292782e-01 2.93957472e-01 4.22201693e-01
6.17217898e-01 -1.02369614e-01 5.92733383e-01 5.24373412e-01
4.90519524e-01 -1.01840341e+00 1.04380131e-01 8.70812953e-01
8.77030432e-01 -1.23428905e+00 -1.06176376e-01 -1.16871409e-01
-6.99697733e-01 8.86064470e-01 9.68689144e-01 1.63771689e-01
4.37110543e-01 3.93739194e-01 -1.64296240e-01 -1.19937114e-01
-1.21863198e+00 1.39570177e-01 1.68716423e-02 4.41264868e-01
5.79811931e-01 2.08425656e-01 -3.39591324e-01 1.09508872e+00
-9.91105735e-01 -3.33628774e-01 3.87838542e-01 2.47601375e-01
-4.42484617e-01 -9.07321036e-01 -2.56445348e-01 1.03289628e+00
-7.80930579e-01 -3.07736278e-01 -9.07613158e-01 5.56130886e-01
-2.48677894e-01 1.05858254e+00 8.26611668e-02 -8.68501008e-01
2.77547747e-01 3.75907302e-01 3.46158475e-01 -5.25322855e-01
-1.08605635e+00 -3.46712261e-01 8.19882035e-01 -5.39215684e-01
2.69622415e-01 -6.82132125e-01 -7.23822176e-01 -8.76033127e-01
-2.83276707e-01 1.89627945e-01 4.80473429e-01 8.29408884e-01
7.14783370e-02 1.36349380e-01 5.93388617e-01 -7.35713720e-01
-1.28483629e+00 -1.47957706e+00 -5.39160788e-01 5.34734726e-01
1.24913417e-01 -8.57627332e-01 -6.66339040e-01 -5.14678895e-01] | [8.849098205566406, 10.588687896728516] |
ea914403-f7fa-47d7-adad-84f4da4f23e5 | proton-probing-schema-linking-information | 2206.14017 | null | https://arxiv.org/abs/2206.14017v2 | https://arxiv.org/pdf/2206.14017v2.pdf | Proton: Probing Schema Linking Information from Pre-trained Language Models for Text-to-SQL Parsing | The importance of building text-to-SQL parsers which can be applied to new databases has long been acknowledged, and a critical step to achieve this goal is schema linking, i.e., properly recognizing mentions of unseen columns or tables when generating SQLs. In this work, we propose a novel framework to elicit relational structures from large-scale pre-trained language models (PLMs) via a probing procedure based on Poincar\'e distance metric, and use the induced relations to augment current graph-based parsers for better schema linking. Compared with commonly-used rule-based methods for schema linking, we found that probing relations can robustly capture semantic correspondences, even when surface forms of mentions and entities differ. Moreover, our probing procedure is entirely unsupervised and requires no additional parameters. Extensive experiments show that our framework sets new state-of-the-art performance on three benchmarks. We empirically verify that our probing procedure can indeed find desired relational structures through qualitative analysis. Our code can be found at https://github.com/AlibabaResearch/DAMO-ConvAI. | ['Yongbin Li', 'Luo Si', 'Fei Huang', 'Binhua Li', 'Bailin Wang', 'Min Yang', 'Bowen Li', 'Binyuan Hui', 'Bowen Qin', 'Lihan Wang'] | 2022-06-28 | null | null | null | null | ['text-to-sql'] | ['computer-code'] | [ 1.48990542e-01 6.22822940e-01 -4.60323304e-01 -5.28181016e-01
-1.06936693e+00 -9.25064802e-01 6.22195780e-01 6.74878120e-01
-2.40530837e-02 4.81193572e-01 2.14479178e-01 -5.48372567e-01
-9.00878981e-02 -1.37880075e+00 -1.07561135e+00 1.07407443e-01
6.52294084e-02 8.35131943e-01 4.35046405e-01 -3.54187250e-01
1.42156854e-01 4.38998878e-01 -1.32103789e+00 5.47499776e-01
9.29224551e-01 6.16538525e-01 -1.71661004e-01 2.48033538e-01
-6.73970997e-01 7.60335505e-01 -2.63134807e-01 -1.05238152e+00
1.04712255e-01 -2.80791909e-01 -1.28694558e+00 -2.13063896e-01
3.93771112e-01 1.70317650e-01 -3.16358805e-01 1.20197475e+00
7.99438953e-02 -3.13354045e-01 2.61434734e-01 -1.13388014e+00
-7.46234059e-01 1.20533895e+00 -1.77142918e-01 9.64657292e-02
7.40856171e-01 -2.81684905e-01 1.59692955e+00 -9.57790434e-01
1.05391335e+00 1.28243148e+00 5.73369026e-01 4.55019832e-01
-1.31854320e+00 -5.25045156e-01 5.07409796e-02 1.06675945e-01
-1.32914448e+00 -5.59765637e-01 7.77179837e-01 -2.60141999e-01
9.77309585e-01 4.74968404e-01 1.51052922e-01 1.02046335e+00
-2.41183832e-01 6.72983766e-01 8.27841759e-01 -6.10216975e-01
-5.61059937e-02 4.09602493e-01 1.99574932e-01 9.72230613e-01
7.34730124e-01 -3.51709545e-01 -4.86227900e-01 -1.86769947e-01
5.77792227e-01 -4.06826317e-01 3.22316736e-02 -5.85709870e-01
-1.23705900e+00 7.61343598e-01 3.65831971e-01 4.45666045e-01
-1.14301316e-01 -1.14107177e-01 3.74163151e-01 1.78864121e-01
2.47638702e-01 6.93052888e-01 -5.40463269e-01 -4.54692356e-02
-5.68942130e-01 1.43831998e-01 1.11290908e+00 1.47014630e+00
9.07987177e-01 -6.47501707e-01 1.81966141e-01 7.15403914e-01
2.83510625e-01 3.28185558e-01 4.43151221e-02 -8.62709939e-01
9.54270720e-01 1.21515846e+00 -1.30746216e-01 -1.08653104e+00
-3.23541909e-01 -2.80079603e-01 -5.22689283e-01 -5.28038681e-01
4.88968968e-01 1.99402153e-01 -4.99651253e-01 1.64887631e+00
4.79272693e-01 -2.22844362e-01 3.25068921e-01 5.20745277e-01
1.04993534e+00 3.47474277e-01 1.75733745e-01 4.68863286e-02
1.47605145e+00 -6.20739520e-01 -5.73845208e-01 -4.60139513e-01
1.08026171e+00 -6.33977056e-01 1.26743448e+00 2.29594857e-03
-1.16988730e+00 -4.10269201e-01 -8.26257527e-01 -2.69085616e-01
-7.16615438e-01 -1.74485147e-02 8.75321150e-01 5.37325621e-01
-7.74304032e-01 5.52517414e-01 -7.75191128e-01 -8.50981116e-01
4.44945365e-01 1.25525892e-01 -6.13148928e-01 -5.44914417e-02
-1.23008609e+00 8.34373236e-01 9.05118167e-01 7.59196207e-02
-3.06992382e-01 -6.45403981e-01 -1.03395092e+00 5.37154004e-02
1.06397295e+00 -7.56872892e-01 1.16593540e+00 -2.84937531e-01
-1.03345764e+00 1.15542483e+00 -3.63770187e-01 -4.12812144e-01
2.30978459e-01 -1.99693590e-01 -5.62442422e-01 7.04484209e-02
3.74535054e-01 4.89906073e-01 -1.82152838e-01 -1.50928032e+00
-4.80307639e-01 -3.38646680e-01 3.06758106e-01 -1.46693746e-02
-2.71233290e-01 2.03073665e-01 -8.29507411e-01 -4.45363313e-01
5.01557469e-01 -6.46251678e-01 -8.99978727e-02 -2.16324449e-01
-9.20790911e-01 -4.29679185e-01 2.79646963e-01 -5.51020086e-01
1.41664660e+00 -1.77706933e+00 1.26785412e-02 5.31655431e-01
2.02261195e-01 1.96309552e-01 -4.45975661e-02 9.25074160e-01
-9.40414295e-02 5.86056173e-01 -4.31054264e-01 1.26947701e-01
3.20064366e-01 3.62161487e-01 -4.68394697e-01 -2.19185740e-01
4.75077689e-01 1.21615410e+00 -8.83442044e-01 -8.33844364e-01
-1.74224272e-01 2.68441085e-02 -5.21135151e-01 2.84568876e-01
-5.55189788e-01 3.33144814e-02 -3.73973995e-01 8.07262957e-01
4.25676376e-01 -5.69397926e-01 7.60438621e-01 -3.88576537e-01
2.20583126e-01 9.11333978e-01 -1.33946550e+00 1.82342172e+00
-2.96717614e-01 1.26115039e-01 -4.29767817e-01 -9.61171508e-01
1.24554396e+00 5.70192598e-02 2.01201662e-01 -7.31099248e-01
-1.88000679e-01 3.81388098e-01 -2.36074418e-01 -7.13451028e-01
3.86260986e-01 1.88980907e-01 -3.64294022e-01 2.76481271e-01
1.43001243e-01 -1.02884866e-01 6.09827280e-01 6.06991589e-01
1.25609338e+00 2.55394220e-01 3.89905691e-01 -1.52875409e-01
6.26013756e-01 3.64486039e-01 6.31344497e-01 6.29365861e-01
3.84840012e-01 2.61692673e-01 9.11806703e-01 -2.99513429e-01
-9.71629322e-01 -1.18730259e+00 -2.63868719e-02 8.85525703e-01
1.60609379e-01 -8.65483105e-01 -6.61224782e-01 -1.13157821e+00
1.17536932e-02 8.30105960e-01 -4.96456355e-01 1.14002153e-01
-8.49826872e-01 -5.29792964e-01 7.20779181e-01 6.28434122e-01
1.31420761e-01 -1.03449178e+00 -1.20086148e-01 2.11250141e-01
-3.63774061e-01 -1.51021850e+00 3.82394567e-02 1.86477527e-01
-7.58735180e-01 -1.41378760e+00 1.93374932e-01 -8.80509317e-01
7.81533301e-01 -1.28553659e-01 1.54617834e+00 2.24394321e-01
-7.50076398e-03 1.86729416e-01 -3.64008933e-01 -1.54859737e-01
-7.07138598e-01 4.99846756e-01 -4.53900725e-01 -3.74134928e-01
6.57528818e-01 -4.94138360e-01 -2.04660118e-01 3.36370885e-01
-1.08484042e+00 2.90266573e-01 5.88617146e-01 4.21837062e-01
7.00797379e-01 -6.26413748e-02 5.16637921e-01 -1.74106705e+00
4.99127477e-01 -4.63119566e-01 -7.14445174e-01 7.89392710e-01
-6.95841670e-01 4.03306067e-01 5.50612926e-01 1.32780954e-01
-9.93186831e-01 3.11480295e-02 -1.82788402e-01 2.08191916e-01
-3.18552196e-01 9.20571923e-01 -5.31218290e-01 2.12481350e-01
6.49966240e-01 -1.01089954e-01 -4.01363283e-01 -5.92038333e-01
7.28791893e-01 4.25312430e-01 7.24106848e-01 -1.05577421e+00
1.14975524e+00 3.90705764e-01 -1.02656186e-02 -2.26829991e-01
-9.33216810e-01 -3.67394388e-01 -9.84539986e-01 2.52700835e-01
5.71170211e-01 -7.98081815e-01 -5.71190476e-01 -9.12728086e-02
-1.23475456e+00 -2.05598071e-01 -2.18264550e-01 5.74709363e-02
-1.94011360e-01 4.07659948e-01 -6.57863975e-01 -3.63909006e-01
-2.41897717e-01 -7.82904744e-01 8.77913296e-01 1.50981650e-01
-3.84053320e-01 -1.04187667e+00 1.52342066e-01 5.47973871e-01
1.04273697e-02 2.26255178e-01 1.35792005e+00 -1.13923156e+00
-8.77209783e-01 -7.62401149e-02 -4.90336508e-01 -1.26269013e-01
1.50282919e-01 1.14902496e-01 -6.10645413e-01 1.51679829e-01
-5.64838529e-01 -2.88153112e-01 4.77371842e-01 -5.30030668e-01
9.90291119e-01 -4.59810615e-01 -4.77928519e-01 5.02882063e-01
1.53640151e+00 1.29568219e-01 7.11040735e-01 5.42912364e-01
8.27037573e-01 8.45593572e-01 6.23405755e-01 1.43986344e-02
9.55309629e-01 6.27991676e-01 3.49975973e-01 1.78749766e-02
-1.26634121e-01 -8.06376934e-01 -2.48046797e-02 9.84364748e-01
3.17731857e-01 -1.96746424e-01 -1.32522488e+00 6.43357873e-01
-1.82503808e+00 -6.80947363e-01 -3.23407829e-01 1.88698435e+00
1.27690542e+00 3.99452269e-01 -8.04452598e-02 -1.04764923e-02
6.19052410e-01 -9.55174044e-02 -3.21833014e-01 -2.44107738e-01
-1.51914135e-01 4.39651936e-01 4.08093989e-01 4.83172208e-01
-9.06053126e-01 1.26545894e+00 5.27463484e+00 5.77848315e-01
-6.58931375e-01 -1.03949733e-01 3.44171584e-01 3.24777842e-01
-8.81720245e-01 3.91276330e-01 -1.08808148e+00 1.20588742e-01
8.94046307e-01 -3.03612024e-01 3.21014196e-01 7.30665028e-01
-1.59444824e-01 7.73380399e-02 -1.39330018e+00 6.01003706e-01
-1.56269133e-01 -1.58927441e+00 1.02622867e-01 -1.32767767e-01
3.48697722e-01 -1.17872462e-01 -4.33196932e-01 3.38996917e-01
6.04631722e-01 -8.43763053e-01 5.52284718e-01 5.08129776e-01
5.92544198e-01 -4.77848411e-01 6.87644541e-01 7.01875985e-02
-1.30653095e+00 2.88904309e-01 -2.91253358e-01 4.13641155e-01
1.93212628e-02 7.17485905e-01 -9.81485069e-01 1.05823505e+00
6.28809512e-01 6.32050514e-01 -1.04806852e+00 7.78478384e-01
-6.43508255e-01 4.95181739e-01 -2.88374722e-01 5.36703654e-02
-1.22843921e-01 -8.73925835e-02 2.76922047e-01 1.32265174e+00
2.75259405e-01 8.65622908e-02 6.54826611e-02 1.15254974e+00
-4.93133545e-01 2.78878391e-01 -7.31376946e-01 -3.00192416e-01
8.64937544e-01 1.30127215e+00 -7.77301371e-01 -4.38863486e-01
-5.73271394e-01 5.29381573e-01 6.90387189e-01 2.11547658e-01
-7.31296718e-01 -5.50549507e-01 3.19957495e-01 2.99316317e-01
3.12582105e-01 -1.67476714e-01 -4.46260482e-01 -1.37006998e+00
4.70674574e-01 -1.08502805e+00 7.28716671e-01 -7.88940489e-01
-1.22788715e+00 6.86357677e-01 8.66157487e-02 -8.31575990e-01
-1.73305884e-01 -4.26147580e-01 -3.75629067e-01 6.11627221e-01
-1.50207567e+00 -1.31502914e+00 -3.53115290e-01 3.94241929e-01
1.41620979e-01 1.27594888e-01 9.69231725e-01 4.87020820e-01
-7.21941352e-01 5.95004559e-01 -5.01350701e-01 6.87232018e-01
7.19124436e-01 -1.35200596e+00 8.26899946e-01 1.26066613e+00
7.18414485e-01 9.94834781e-01 5.70709884e-01 -8.21583331e-01
-1.60549259e+00 -1.16282344e+00 1.43806148e+00 -6.26800299e-01
9.84437585e-01 -6.81906939e-01 -1.39582777e+00 1.05928028e+00
1.84655741e-01 4.04026406e-03 6.64897561e-01 3.65218192e-01
-8.35210145e-01 -2.06881896e-01 -8.32104385e-01 5.56144357e-01
1.53761041e+00 -7.16518760e-01 -8.11871886e-01 4.57310230e-01
8.58621418e-01 -6.89480126e-01 -1.27982998e+00 6.43493056e-01
2.37274796e-01 -8.12237203e-01 8.67038429e-01 -9.30144489e-01
5.24831831e-01 -3.58659238e-01 -2.18393952e-01 -9.01292682e-01
-8.65977705e-02 -5.82957625e-01 -2.39065811e-01 1.77928936e+00
9.09247994e-01 -7.25147486e-01 8.33946407e-01 6.59559071e-01
-1.05695613e-01 -6.98401630e-01 -5.32492459e-01 -7.68411338e-01
-2.54731835e-03 -4.22670931e-01 8.94694507e-01 1.20856142e+00
3.29208493e-01 3.73562336e-01 2.48965502e-01 5.68976402e-01
4.72492129e-01 3.44245940e-01 8.51793289e-01 -1.30103707e+00
-2.15439260e-01 -2.82498777e-01 -1.12258501e-01 -8.36247206e-01
2.56776303e-01 -1.31820989e+00 -1.88377276e-01 -2.00694156e+00
1.94154099e-01 -8.09071839e-01 -1.09427638e-01 8.12480152e-01
-1.83592156e-01 -1.03928605e-02 -1.64905842e-02 1.53892532e-01
-6.46771133e-01 1.11260496e-01 8.61676633e-01 -1.19953021e-01
-6.31633475e-02 -2.63859093e-01 -1.09760988e+00 5.59020936e-01
8.97633493e-01 -8.06817532e-01 -3.66648257e-01 -4.76470441e-01
8.24482799e-01 1.84304658e-02 3.12182575e-01 -7.23401427e-01
4.41481739e-01 -1.76507577e-01 -1.15672335e-01 -4.79449809e-01
-1.05677351e-01 -7.24744916e-01 3.14251453e-01 1.61168590e-01
-5.49346805e-01 2.78374791e-01 2.29828387e-01 3.03663224e-01
-3.19350272e-01 -3.30605149e-01 3.10405612e-01 -1.75498128e-01
-7.06477165e-01 7.16348588e-02 1.82310447e-01 5.70590794e-01
8.24403167e-01 2.71380365e-01 -7.98173666e-01 -1.75600778e-02
-5.78168571e-01 1.99589431e-01 4.94476140e-01 5.16161144e-01
4.36916888e-01 -1.29802120e+00 -5.61515927e-01 5.44392280e-02
4.86550868e-01 2.61964679e-01 -3.70480180e-01 6.05608821e-01
-5.64278603e-01 5.03934979e-01 1.52442768e-01 -3.97329807e-01
-1.22258067e+00 6.57413661e-01 1.08059518e-01 -5.55941164e-01
-4.54286307e-01 7.13513494e-01 -1.05486006e-01 -8.04542661e-01
8.71963203e-02 -5.10708869e-01 -9.24585909e-02 -1.43818125e-01
2.58859061e-02 1.92849971e-02 3.46085936e-01 -3.92538756e-01
-5.47594070e-01 4.35865283e-01 -1.09335221e-01 1.20053351e-01
1.39190590e+00 7.43092503e-03 -5.05142450e-01 2.93128997e-01
9.36134040e-01 3.43875080e-01 -5.61856747e-01 -5.56934237e-01
7.89056063e-01 -3.82361293e-01 -5.58272302e-01 -1.00263739e+00
-8.78373504e-01 5.99161863e-01 -1.74233094e-01 6.20013654e-01
8.65010500e-01 5.51450253e-01 8.04264665e-01 6.53450251e-01
4.18776631e-01 -7.71512389e-01 -1.23424016e-01 3.26885641e-01
7.36352205e-01 -1.26634538e+00 -1.01589248e-01 -1.08520699e+00
-5.19475520e-01 1.15433717e+00 6.92700386e-01 2.56764650e-01
3.56121093e-01 3.73537213e-01 -7.03330850e-03 -5.05665541e-01
-8.02350521e-01 -3.92569125e-01 3.52327973e-01 4.68425363e-01
6.16671145e-01 -8.08669925e-02 -1.95518866e-01 6.95068598e-01
-3.68015438e-01 -1.80438146e-01 4.20081735e-01 9.18366015e-01
-1.95920140e-01 -1.73803127e+00 -4.22254838e-02 3.39317381e-01
-4.20616239e-01 -3.30144078e-01 -5.49685061e-01 9.82815623e-01
-8.87616873e-02 9.19061959e-01 -1.95489720e-01 -3.02449554e-01
7.76337922e-01 3.44846010e-01 4.77403522e-01 -9.50764358e-01
-5.44137001e-01 -2.65974164e-01 6.43204093e-01 -7.43681014e-01
-4.26841199e-01 -6.37986243e-01 -1.38724804e+00 -3.24043632e-01
-1.89677805e-01 4.20289963e-01 4.67528999e-01 7.78024673e-01
7.18486249e-01 4.02704298e-01 2.96815306e-01 2.52172470e-01
-2.01577559e-01 -5.72237015e-01 -2.26580817e-02 6.36384130e-01
-4.50903952e-01 -4.27301586e-01 -3.76382843e-02 2.20525667e-01] | [9.489997863769531, 8.137606620788574] |
1b0d74fc-402b-4401-b689-75a6104ec7d9 | mptv-matching-pursuit-based-total-variation | 1810.05438 | null | http://arxiv.org/abs/1810.05438v1 | http://arxiv.org/pdf/1810.05438v1.pdf | MPTV: Matching Pursuit Based Total Variation Minimization for Image Deconvolution | Total variation (TV) regularization has proven effective for a range of
computer vision tasks through its preferential weighting of sharp image edges.
Existing TV-based methods, however, often suffer from the over-smoothing issue
and solution bias caused by the homogeneous penalization. In this paper, we
consider addressing these issues by applying inhomogeneous regularization on
different image components. We formulate the inhomogeneous TV minimization
problem as a convex quadratic constrained linear programming problem. Relying
on this new model, we propose a matching pursuit based total variation
minimization method (MPTV), specifically for image deconvolution. The proposed
MPTV method is essentially a cutting-plane method, which iteratively activates
a subset of nonzero image gradients, and then solves a subproblem focusing on
those activated gradients only. Compared to existing methods, MPTV is less
sensitive to the choice of the trade-off parameter between data fitting and
regularization. Moreover, the inhomogeneity of MPTV alleviates the
over-smoothing and ringing artifacts, and improves the robustness to errors in
blur kernel. Extensive experiments on different tasks demonstrate the
superiority of the proposed method over the current state-of-the-art. | ['Anton Van Den Hengel', 'Yanning Zhang', 'Dong Gong', 'Qinfeng Shi', 'Mingkui Tan'] | 2018-10-12 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 3.23690236e-01 -2.39937618e-01 1.39125541e-01 -1.88018650e-01
-5.87192118e-01 -1.82877392e-01 3.27579409e-01 -3.07717294e-01
-4.05397326e-01 5.08937418e-01 1.05776675e-01 8.85172561e-02
-1.77798718e-01 -2.88177073e-01 -4.80689853e-01 -1.06931674e+00
4.56282377e-01 -1.53518543e-01 3.46022069e-01 1.32863462e-01
5.41123927e-01 2.26257354e-01 -1.03361332e+00 -3.29695717e-02
1.34602749e+00 9.22075093e-01 4.77285743e-01 -3.56514081e-02
-1.10829867e-01 6.12607956e-01 -1.55809224e-01 -1.05040289e-01
3.58314514e-01 -2.35964552e-01 -4.55586404e-01 6.47267520e-01
4.50908422e-01 -1.53535590e-01 -1.22593790e-01 1.39594901e+00
3.70970100e-01 3.79158765e-01 5.31847417e-01 -9.01513040e-01
-8.86322737e-01 7.79856443e-02 -1.23147321e+00 3.49278808e-01
-1.26850680e-01 1.05544508e-01 6.77302659e-01 -1.31085432e+00
4.57105935e-01 1.02416289e+00 7.59819508e-01 3.21083516e-01
-1.50637794e+00 -2.41275236e-01 2.03894913e-01 8.16098452e-02
-1.37197685e+00 -3.87397707e-01 1.21848893e+00 -7.11666703e-01
3.28679651e-01 2.27932826e-01 2.37165108e-01 4.90251750e-01
1.07819237e-01 7.52916873e-01 1.19378531e+00 -4.04007912e-01
1.64851770e-01 2.14677081e-01 4.56295997e-01 6.16919756e-01
2.82512575e-01 -1.31210968e-01 -1.62189782e-01 -3.02678138e-01
1.03887463e+00 -7.00620934e-02 -8.95067751e-01 -6.52446628e-01
-1.09107125e+00 7.20115960e-01 2.32658759e-01 2.91499794e-01
-5.97867131e-01 -1.09261587e-01 1.53581113e-01 -1.39243245e-01
8.93167257e-01 2.38145307e-01 -2.18594611e-01 3.82201791e-01
-1.05117404e+00 2.00634867e-01 2.44276375e-01 6.31520629e-01
8.07096839e-01 2.01241195e-01 -5.33928394e-01 1.27714467e+00
5.48717380e-01 2.16681033e-01 3.29286814e-01 -9.98123467e-01
4.11838889e-01 4.29917753e-01 2.35487744e-01 -1.31843495e+00
-2.77466834e-01 -5.17240584e-01 -8.07517648e-01 2.16623098e-01
4.98956054e-01 -1.90478534e-01 -9.49784219e-01 1.63252330e+00
4.71518993e-01 3.57547402e-01 -2.44390234e-01 1.35792887e+00
7.39522755e-01 6.45959377e-01 -1.41476363e-01 -7.03363061e-01
1.07501698e+00 -1.12541997e+00 -1.01879573e+00 -3.44412446e-01
3.04781407e-01 -9.29490864e-01 8.22202146e-01 2.29222864e-01
-1.21873879e+00 -4.36545104e-01 -7.58762121e-01 -6.39422983e-02
3.67873728e-01 3.02273005e-01 3.67654622e-01 4.35681939e-01
-8.98389161e-01 5.05888402e-01 -6.96569502e-01 -4.43146601e-02
2.30816051e-01 2.21134350e-01 -1.74242556e-01 -2.32483428e-02
-7.68451393e-01 7.73336649e-01 -2.69583184e-02 5.55086136e-01
-3.74818414e-01 -8.47372293e-01 -8.33218217e-01 -3.67949232e-02
4.87844408e-01 -5.13776302e-01 7.79233694e-01 -1.16285574e+00
-1.59977961e+00 8.09458494e-01 -4.07450199e-01 -7.86019564e-02
6.39170766e-01 -2.86686271e-01 -2.33032554e-02 6.84438497e-02
6.67599663e-02 8.74362811e-02 1.37263966e+00 -1.43608582e+00
-8.95407870e-02 -4.10324067e-01 -2.94348747e-01 1.95907667e-01
-2.71944791e-01 4.15094197e-02 -8.84748459e-01 -8.87295187e-01
4.49446797e-01 -9.87172306e-01 -4.24858361e-01 5.71723878e-02
-2.85075635e-01 -3.50411460e-02 7.06678927e-01 -7.11772144e-01
1.26593244e+00 -2.27727437e+00 3.92901868e-01 1.06299788e-01
2.35475242e-01 3.64072472e-01 7.32123256e-02 5.20751514e-02
-4.39117476e-02 -2.64429748e-01 -7.75728524e-01 -4.07100111e-01
-3.76557142e-01 3.84706557e-02 4.62098084e-02 9.27049220e-01
3.21082622e-02 3.77994269e-01 -7.11229980e-01 -6.10869229e-01
3.66321802e-01 6.92587376e-01 -5.02464294e-01 1.86941549e-01
2.19665617e-02 7.66959965e-01 -5.48612058e-01 2.54433453e-01
1.29711509e+00 -1.56579852e-01 -2.13845298e-02 -5.16736925e-01
-6.05366707e-01 -3.54697585e-01 -1.32959843e+00 1.58650804e+00
-1.71484783e-01 6.69651151e-01 5.09173274e-01 -1.22261548e+00
8.09440017e-01 1.98770911e-01 6.78520620e-01 -2.56710112e-01
2.19889104e-01 2.91256040e-01 -2.18247905e-01 -8.48248124e-01
3.37977052e-01 -2.08310321e-01 6.29420996e-01 -3.48742120e-02
-2.33883381e-01 2.75501665e-02 9.94803980e-02 -5.06321620e-03
6.52249873e-01 2.62907863e-01 5.06677944e-03 -6.05267406e-01
8.94212961e-01 -1.54131606e-01 9.18798327e-01 4.48526293e-01
-3.89326751e-01 8.76551926e-01 2.52587885e-01 -2.93882400e-01
-7.50218809e-01 -6.25333667e-01 -5.76315343e-01 5.77973783e-01
4.00871485e-01 -1.88166909e-02 -9.21944559e-01 -3.55695158e-01
-1.04936562e-01 4.41630274e-01 -4.37257886e-01 7.12542087e-02
-7.45308280e-01 -1.13281488e+00 -3.24592367e-02 1.14469722e-01
8.06139767e-01 -7.29192436e-01 -3.74183476e-01 2.63089925e-01
-3.58673662e-01 -1.21102309e+00 -8.68462145e-01 -3.03965211e-01
-1.02043736e+00 -8.59969020e-01 -1.31382132e+00 -9.13332641e-01
1.03915846e+00 7.79498994e-01 6.67348027e-01 1.08552411e-01
-1.68507025e-01 3.56402814e-01 -1.62804630e-02 7.79561922e-02
7.38750794e-04 -4.84215051e-01 -5.84338382e-02 7.64909208e-01
-2.97329109e-02 -2.98507899e-01 -6.84717774e-01 4.20275480e-01
-9.34788227e-01 -1.61904693e-02 5.17261147e-01 9.35633421e-01
7.30160475e-01 1.26677617e-01 1.88785344e-01 -1.01783478e+00
6.93447530e-01 -2.85040319e-01 -7.16575801e-01 2.09411785e-01
-6.99391723e-01 7.97477141e-02 4.11961555e-01 -4.45651263e-01
-1.63084328e+00 1.71787784e-01 2.56615877e-01 -7.30496228e-01
2.47207507e-01 5.86703598e-01 -3.92971793e-03 -5.47335386e-01
3.59045982e-01 3.55049074e-01 9.69500616e-02 -6.46097660e-01
1.73720583e-01 3.72470200e-01 4.54835206e-01 -4.38896328e-01
6.97995722e-01 6.81278348e-01 5.40295653e-02 -1.18921649e+00
-7.28894532e-01 -6.72447383e-01 -4.54540163e-01 -3.80726010e-01
9.60649669e-01 -6.56831324e-01 -5.14488816e-01 7.13059306e-01
-1.16248584e+00 -1.76288128e-01 1.15091942e-01 5.39604306e-01
-3.33566725e-01 1.01443899e+00 -5.58806062e-01 -8.74122739e-01
-2.86612988e-01 -1.24990511e+00 5.95235825e-01 2.78390408e-01
2.77479857e-01 -1.05050182e+00 1.47119030e-01 3.94149572e-01
5.03424942e-01 1.97541028e-01 5.39440095e-01 1.59700066e-01
-4.54321563e-01 5.02519943e-02 -4.86275494e-01 4.96488988e-01
7.55908340e-02 -5.16268017e-04 -7.41334200e-01 -2.80395240e-01
6.53491080e-01 1.90048963e-01 1.01344252e+00 1.03815675e+00
1.05160427e+00 -2.37099349e-01 -1.72724873e-01 8.97082210e-01
1.70280385e+00 -6.78298324e-02 5.71505308e-01 5.15956223e-01
8.94390762e-01 7.39258647e-01 5.94710827e-01 3.86904687e-01
2.10546449e-01 8.35048616e-01 4.14494216e-01 -4.02615577e-01
-3.31979170e-02 4.00444329e-01 1.07128277e-01 6.63234770e-01
-3.08261693e-01 9.42503586e-02 -6.33232176e-01 6.47396505e-01
-2.07223535e+00 -7.75052249e-01 -8.20373952e-01 2.31848288e+00
8.00047100e-01 -4.72569503e-02 -1.48429543e-01 -2.06511002e-02
9.83373940e-01 3.28161091e-01 -4.54680711e-01 -1.58057585e-01
-2.09466685e-02 -3.08589846e-01 6.14854813e-01 5.91224074e-01
-1.14561427e+00 7.51909733e-01 5.69797850e+00 8.39267671e-01
-1.09437299e+00 1.36349708e-01 2.74150401e-01 1.09622754e-01
-1.95194036e-01 1.24069132e-01 -5.02451956e-01 6.32482052e-01
8.88129994e-02 3.94364595e-02 4.73813981e-01 4.94350523e-01
6.15319848e-01 -1.56152129e-01 -5.36553979e-01 1.05729473e+00
1.22201897e-01 -1.01764655e+00 -1.47804916e-01 -4.46451381e-02
8.42801452e-01 -1.23351656e-01 2.14725018e-01 -1.48649395e-01
-2.46431947e-01 -5.62007308e-01 6.62312984e-01 5.15410781e-01
3.66849661e-01 -3.97133738e-01 6.03016853e-01 2.06561297e-01
-9.80589867e-01 3.95626724e-02 -4.88758832e-01 1.50870800e-01
3.10395509e-01 1.11774862e+00 -8.87354612e-02 3.95131826e-01
6.28209651e-01 7.13802099e-01 -1.81036741e-01 1.17617738e+00
-2.22816765e-01 4.61579084e-01 -9.42892730e-02 4.58446473e-01
1.69281378e-01 -8.49500120e-01 9.79073167e-01 1.17133749e+00
9.79231447e-02 4.28904921e-01 1.76589698e-01 1.09469104e+00
1.64292261e-01 2.69935519e-01 -2.58745432e-01 3.73561561e-01
1.43186718e-01 1.39299035e+00 -6.48799539e-01 -1.00937091e-01
-6.65631711e-01 9.97535229e-01 3.32553267e-01 7.94293702e-01
-7.27200270e-01 -1.63078710e-01 3.20368409e-01 1.04524016e-01
3.98655117e-01 -3.09978217e-01 -4.86491591e-01 -1.44072354e+00
3.06470245e-01 -6.48991287e-01 1.25939950e-01 -6.85846329e-01
-1.11186481e+00 4.73785460e-01 -1.22459449e-01 -1.29056621e+00
5.22592962e-01 -3.91228974e-01 -7.50315428e-01 8.93674850e-01
-1.74386466e+00 -8.57457638e-01 -3.63824219e-01 6.67278647e-01
7.67717361e-01 3.65916491e-01 7.51018524e-02 4.29987073e-01
-9.88883555e-01 2.14121580e-01 3.04909319e-01 -1.99287251e-01
7.66994119e-01 -1.01160192e+00 -3.29233140e-01 1.15361416e+00
-4.31566060e-01 8.08060527e-01 9.85847414e-01 -6.05192363e-01
-1.26420128e+00 -9.35517848e-01 6.88626826e-01 1.12950027e-01
5.18306375e-01 1.81137472e-01 -1.33760333e+00 6.04442775e-01
1.53001353e-01 2.08756432e-01 3.44660461e-01 -2.44620278e-01
-3.32374536e-02 -4.53009009e-02 -1.13079786e+00 5.02285302e-01
5.78159213e-01 -7.54577145e-02 -4.49654490e-01 2.63658613e-01
3.56853455e-01 -4.55482483e-01 -7.33218670e-01 4.29983288e-01
2.63943672e-01 -9.88419116e-01 1.08176601e+00 -8.31771120e-02
3.96001518e-01 -4.81398851e-01 1.58015355e-01 -1.36793804e+00
-7.97856808e-01 -6.57606304e-01 1.35752186e-01 1.32451200e+00
1.52913079e-01 -6.03889942e-01 5.06103873e-01 8.93458903e-01
-3.11600059e-01 -5.33862591e-01 -7.30724454e-01 -6.62845552e-01
-2.52556026e-01 -1.94339722e-01 -1.83731690e-01 1.20283222e+00
-5.31257316e-02 1.45489886e-01 -6.92322850e-01 3.68925780e-01
1.22982824e+00 3.34954970e-02 5.34496427e-01 -1.03207481e+00
-3.21126640e-01 -3.60299736e-01 -7.93035626e-02 -1.16579604e+00
-6.23177737e-02 -6.56351149e-01 4.25006092e-01 -1.39957798e+00
2.04802796e-01 -3.87006640e-01 -2.49137208e-01 3.23959030e-02
-5.50598085e-01 1.88604087e-01 1.65811315e-01 7.06151366e-01
-2.11761042e-01 5.73595583e-01 1.49160683e+00 -1.63698830e-02
-4.99478757e-01 2.80302558e-02 -4.42527443e-01 1.02829695e+00
5.37607670e-01 -4.00021106e-01 -3.81909430e-01 -8.55000377e-01
-1.24109231e-01 5.50948046e-02 3.25423211e-01 -6.34907663e-01
2.85465211e-01 -3.08764905e-01 3.01861931e-02 -3.26031238e-01
2.39277557e-01 -7.68357694e-01 5.73554561e-02 1.91409826e-01
-2.21423760e-01 -5.08061290e-01 1.18729442e-01 6.88795328e-01
-1.76811203e-01 -4.82709169e-01 1.27780187e+00 -1.55080199e-01
-5.69088697e-01 1.87827066e-01 -4.36473757e-01 -1.36971893e-03
9.10882592e-01 -2.43022174e-01 -4.92865145e-02 -7.96862543e-02
-5.75049162e-01 1.58848301e-01 3.96375954e-01 5.37249085e-04
7.00939059e-01 -1.05913746e+00 -7.78734446e-01 2.31620267e-01
-2.66539454e-01 -1.06000798e-02 5.24599612e-01 1.50138056e+00
-3.97612035e-01 1.56873211e-01 1.84223801e-02 -6.82171345e-01
-1.32852566e+00 7.82865345e-01 3.78512830e-01 -1.50868267e-01
-8.94647300e-01 8.92795622e-01 6.04762256e-01 3.28669213e-02
7.68121481e-02 -1.62536889e-01 -4.41413134e-01 -2.15073302e-01
3.72901946e-01 6.83155000e-01 -1.22884482e-01 -8.36944163e-01
-3.76429141e-01 1.07096577e+00 -7.01030716e-02 -1.42269405e-02
1.24018514e+00 -5.21852732e-01 -3.90442997e-01 1.66263059e-01
1.16502547e+00 1.03672139e-01 -1.48618579e+00 -4.28476781e-01
-4.44623493e-02 -6.95691526e-01 6.63199425e-01 -1.57745361e-01
-1.46339154e+00 6.27570391e-01 4.79658067e-01 2.92392761e-01
1.17372525e+00 -4.87836599e-01 8.15950572e-01 -3.03170979e-01
1.05722226e-01 -1.03954184e+00 -9.70341116e-02 1.97604924e-01
8.99602890e-01 -1.29473472e+00 3.14036667e-01 -8.47652495e-01
-6.87062383e-01 9.51069117e-01 4.31135386e-01 -4.23109233e-01
6.75169945e-01 -9.14094225e-02 7.44230300e-02 -2.18605757e-01
-2.92398632e-01 -7.57898688e-02 5.57608128e-01 2.74508208e-01
4.39270020e-01 -2.93290406e-01 -1.03585112e+00 3.07419777e-01
7.26362646e-01 2.40488470e-01 4.78328615e-01 7.76003003e-01
-5.07230222e-01 -7.56848752e-01 -7.62689173e-01 8.52563456e-02
-7.52185643e-01 -9.48561281e-02 -4.34640013e-02 3.73972744e-01
5.22883013e-02 1.11854422e+00 -1.54738724e-01 1.31233826e-01
3.41769338e-01 -3.41391712e-01 4.25746173e-01 -3.00842941e-01
-4.34048742e-01 6.75427735e-01 -2.74757743e-01 -4.41906720e-01
-7.09984183e-01 -7.37271607e-01 -1.11655986e+00 7.15505704e-02
-7.12936997e-01 1.07632801e-01 5.76988757e-01 8.82218897e-01
2.53128260e-01 4.29462612e-01 6.65396094e-01 -1.01351893e+00
-5.82554519e-01 -8.70424092e-01 -7.44575441e-01 5.22632897e-01
4.85814631e-01 -8.38260531e-01 -8.33649158e-01 1.82170331e-01] | [11.509711265563965, -2.574632167816162] |
b44b7f6a-ffa4-409c-a2c0-8f7e2c8c55d0 | multivariate-time-series-early-classification | 2306.14606 | null | https://arxiv.org/abs/2306.14606v1 | https://arxiv.org/pdf/2306.14606v1.pdf | Multivariate Time Series Early Classification Across Channel and Time Dimensions | Nowadays, the deployment of deep learning models on edge devices for addressing real-world classification problems is becoming more prevalent. Moreover, there is a growing popularity in the approach of early classification, a technique that involves classifying the input data after observing only an early portion of it, aiming to achieve reduced communication and computation requirements, which are crucial parameters in edge intelligence environments. While early classification in the field of time series analysis has been broadly researched, existing solutions for multivariate time series problems primarily focus on early classification along the temporal dimension, treating the multiple input channels in a collective manner. In this study, we propose a more flexible early classification pipeline that offers a more granular consideration of input channels and extends the early classification paradigm to the channel dimension. To implement this method, we utilize reinforcement learning techniques and introduce constraints to ensure the feasibility and practicality of our objective. To validate its effectiveness, we conduct experiments using synthetic data and we also evaluate its performance on real datasets. The comprehensive results from our experiments demonstrate that, for multiple datasets, our method can enhance the early classification paradigm by achieving improved accuracy for equal input utilization. | ['Henri Bal', 'Mark Hoogendoorn', 'Kees Verstoep', 'Leonardos Pantiskas'] | 2023-06-26 | null | null | null | null | ['classification-1', 'time-series'] | ['methodology', 'time-series'] | [ 1.21681772e-01 -5.09432316e-01 -4.76807892e-01 -5.99478893e-02
-2.91730970e-01 -3.84067953e-01 3.58623683e-01 3.20795834e-01
-3.67521167e-01 4.38580424e-01 -2.52522737e-01 -6.16910219e-01
-4.80807602e-01 -6.78339124e-01 -3.51141214e-01 -5.67162335e-01
-2.82362044e-01 -9.35686901e-02 7.73828477e-02 9.48735923e-02
2.59349346e-01 4.33095753e-01 -1.41257298e+00 2.23991841e-01
8.28328609e-01 1.56852007e+00 -9.47684646e-02 5.49296916e-01
-1.12832926e-01 8.33410144e-01 -7.25995660e-01 -9.06182900e-02
3.57915193e-01 -2.13940233e-01 -5.07450283e-01 1.08559594e-01
-4.47307438e-01 -2.36019611e-01 -2.48309731e-01 7.08822072e-01
5.26971877e-01 -1.31746441e-01 3.17647815e-01 -1.45172691e+00
-1.64819896e-01 4.36847776e-01 -7.04556704e-01 5.68896532e-01
-7.75568187e-02 -2.49404497e-02 1.04130578e+00 -5.64182103e-01
2.09527284e-01 4.74081129e-01 7.17503011e-01 2.66213924e-01
-1.09684241e+00 -6.96524560e-01 4.13376272e-01 5.17057478e-01
-1.25874281e+00 -3.57656926e-01 1.10101593e+00 -3.41357857e-01
6.98697746e-01 1.39356285e-01 6.82655275e-01 8.77330244e-01
2.36571476e-01 1.08492410e+00 8.87771785e-01 -4.26316619e-01
7.50168204e-01 -3.33713107e-02 4.13750559e-01 4.10921872e-02
2.62132660e-02 2.64711559e-01 -4.74849552e-01 -1.28088165e-02
4.54638630e-01 4.67425704e-01 -1.33089781e-01 -1.68422744e-01
-9.00825620e-01 4.69060034e-01 3.03397000e-01 2.74763077e-01
-6.29976273e-01 8.44672993e-02 6.34434283e-01 5.25692165e-01
6.11127794e-01 9.66245905e-02 -5.23468196e-01 -5.88796258e-01
-7.48895645e-01 -2.00185761e-01 6.60331845e-01 8.12668622e-01
3.51081073e-01 -6.88435063e-02 -1.82208687e-01 4.48989809e-01
-6.72188625e-02 1.61005165e-02 3.95254493e-01 -6.43172204e-01
3.68346751e-01 6.99485898e-01 -2.58500464e-02 -8.22824478e-01
-5.43578506e-01 -9.20224607e-01 -1.00357044e+00 8.44882950e-02
3.66589308e-01 -3.89640749e-01 -6.14366949e-01 1.56555557e+00
2.45618314e-01 5.12445986e-01 1.58418626e-01 8.66485178e-01
2.45597616e-01 7.05013752e-01 1.83643222e-01 -4.91809309e-01
1.13017225e+00 -8.50955248e-01 -5.51486492e-01 -6.03393242e-02
7.26052582e-01 -3.95779222e-01 7.06434727e-01 5.84866524e-01
-6.14273429e-01 -6.19721949e-01 -1.03453970e+00 4.42431420e-01
-2.52257556e-01 3.02495241e-01 9.93741214e-01 7.33816564e-01
-8.24174821e-01 5.49485981e-01 -8.77242506e-01 -1.79419681e-01
5.30616045e-01 4.95883346e-01 1.81158811e-01 1.48484349e-01
-1.01477289e+00 1.56599149e-01 1.92199707e-01 1.97004646e-01
-4.11651760e-01 -7.03952134e-01 -2.95178682e-01 3.70855689e-01
2.92522490e-01 -3.80578220e-01 1.18212783e+00 -1.19546640e+00
-1.58627343e+00 1.25853881e-01 -8.95080939e-02 -5.89821696e-01
3.96221489e-01 -3.08202091e-03 -6.15574121e-01 6.56238496e-02
-2.81352550e-01 1.60863459e-01 8.29386771e-01 -8.24453294e-01
-1.04393518e+00 -4.23077703e-01 3.31164777e-01 2.98391655e-02
-9.45208311e-01 -2.10883066e-01 -2.12815508e-01 -6.30563080e-01
-7.94962719e-02 -8.32765698e-01 -3.54725510e-01 -9.40563437e-03
1.52859792e-01 -2.82954276e-01 9.41932321e-01 -3.28584582e-01
1.55815613e+00 -2.46707225e+00 -9.24966261e-02 3.19983691e-01
3.91264319e-01 3.16598326e-01 1.60970405e-01 2.87970155e-01
-4.74328967e-03 -7.00689927e-02 9.77167860e-02 -4.02021885e-01
-2.22143918e-01 8.19434077e-02 -3.50361705e-01 1.95492700e-01
5.51334992e-02 8.37855577e-01 -8.59667301e-01 -3.70467044e-02
8.09156969e-02 2.29648128e-01 -5.27895272e-01 1.10038042e-01
-3.65653075e-02 4.48590577e-01 -4.98896033e-01 5.96256077e-01
4.63603348e-01 -6.96203709e-01 1.18664593e-01 -1.37617663e-01
-1.51179805e-01 -1.56670898e-01 -9.92938042e-01 1.37529969e+00
-6.53361320e-01 6.23888850e-01 -1.14459567e-01 -1.34298515e+00
7.68006325e-01 4.07433659e-01 1.04540741e+00 -1.17965126e+00
2.97297865e-01 1.52416736e-01 1.70383751e-01 -3.67039382e-01
3.00691068e-01 2.24053621e-01 -1.15070827e-02 5.74647427e-01
-5.19816577e-01 6.84971809e-01 1.40597820e-01 -6.00363389e-02
1.22478569e+00 -1.18149422e-01 7.45093822e-02 -1.53605388e-02
3.63086134e-01 -9.20329392e-02 7.14842498e-01 5.88393807e-01
-3.73775065e-01 6.49914965e-02 3.68190944e-01 -5.45670629e-01
-7.29344666e-01 -6.99365199e-01 -3.08686160e-02 8.09707820e-01
3.89042258e-01 -5.32927454e-01 -4.28971022e-01 -6.64188802e-01
-1.06500842e-01 3.75587434e-01 -5.12152612e-01 -3.08935821e-01
-1.55941442e-01 -1.03828740e+00 1.72129363e-01 8.22179735e-01
5.00569105e-01 -7.21764386e-01 -1.12828672e+00 3.44227731e-01
1.14764713e-01 -1.17746735e+00 -7.00834692e-02 5.02167523e-01
-9.54366028e-01 -9.46447253e-01 -7.79135823e-01 -5.85772932e-01
5.89748919e-01 3.11333477e-01 6.79758191e-01 -6.27850667e-02
-2.19777301e-01 3.41801494e-01 -7.11015701e-01 -3.52748752e-01
-8.53052363e-02 2.74777681e-01 -1.40908897e-01 3.86004329e-01
2.98786670e-01 -5.46487331e-01 -7.97434866e-01 3.21247399e-01
-8.42704475e-01 1.06031299e-01 6.63428903e-01 8.86051834e-01
3.57885242e-01 5.20926833e-01 9.96918380e-01 -7.40838647e-01
6.09789610e-01 -5.67230284e-01 -5.13515651e-01 2.54742414e-01
-8.88442159e-01 -6.11722209e-02 1.01075637e+00 -6.84159875e-01
-7.36719906e-01 -6.17033020e-02 -1.15910605e-01 -4.09652770e-01
8.53360593e-02 8.21072042e-01 7.32982019e-03 -7.66205564e-02
3.59548032e-01 3.39299291e-01 -4.04839180e-02 -2.06762910e-01
-2.17506122e-02 1.06348336e+00 2.80801188e-02 -5.38572371e-01
3.03475022e-01 5.24720669e-01 5.80056161e-02 -7.02767968e-01
-4.24135417e-01 -5.72559178e-01 -1.89405531e-01 -5.53190529e-01
3.66159171e-01 -8.06349158e-01 -1.03920829e+00 4.67803687e-01
-8.16390991e-01 -4.64394540e-01 -1.65072501e-01 4.65054423e-01
-2.75079519e-01 1.66514337e-01 -5.52084863e-01 -9.08497930e-01
-4.13716614e-01 -1.10262418e+00 1.02128220e+00 3.54294300e-01
-2.29912058e-01 -1.05837655e+00 -2.76652873e-01 -1.18548468e-01
6.89711034e-01 8.83841235e-03 9.54454541e-01 -5.43545425e-01
-5.20693064e-01 -2.68475205e-01 -2.29749426e-01 5.92182949e-02
3.03251058e-01 -1.36588588e-01 -8.64768207e-01 -4.18902487e-01
3.19173113e-02 5.25819026e-02 6.46143734e-01 2.82292932e-01
1.57516325e+00 9.31203514e-02 -4.94300336e-01 5.36765635e-01
1.37896013e+00 6.90486968e-01 4.87147272e-01 3.28958452e-01
3.71177703e-01 3.03512335e-01 6.62043154e-01 8.09481561e-01
3.06840628e-01 5.83543181e-01 5.28975308e-01 -2.13330597e-01
1.33632585e-01 1.84065357e-01 2.04228714e-01 7.47291684e-01
6.13203831e-02 -4.27799165e-01 -7.32012928e-01 5.32206953e-01
-1.97266328e+00 -8.28671217e-01 2.83594579e-01 2.26445365e+00
4.34344709e-01 5.74291289e-01 1.72762126e-01 7.58081138e-01
6.22602522e-01 -1.39186889e-01 -7.75514603e-01 -1.43477112e-01
2.75267541e-01 1.58892035e-01 3.44810039e-01 -1.70647383e-01
-9.41473842e-01 4.35059249e-01 5.88575554e+00 6.79943383e-01
-1.73479056e+00 -1.35442838e-01 1.00140786e+00 6.11339463e-03
-1.62844330e-01 5.74654043e-02 -4.34794307e-01 6.85125113e-01
1.03400278e+00 -3.48938733e-01 4.54997867e-01 7.82328606e-01
4.04096991e-01 -3.60316113e-02 -1.22313547e+00 1.37744844e+00
-4.51534659e-01 -1.34765792e+00 -3.48953575e-01 3.82796079e-02
5.98452747e-01 -1.21348426e-01 1.87401529e-02 5.10932982e-01
-4.34404582e-01 -4.72511083e-01 6.91935241e-01 3.53864580e-01
5.07847607e-01 -8.60571265e-01 5.75871587e-01 5.18937051e-01
-1.35689640e+00 -6.53380036e-01 1.79471761e-01 -4.25809830e-01
1.10291705e-01 9.21039104e-01 -5.12931347e-01 6.19194806e-01
6.37980223e-01 9.46772158e-01 -4.61667180e-01 1.30874288e+00
2.87312448e-01 6.15153611e-01 -3.03151369e-01 -2.54631698e-01
-3.57859805e-02 7.36156851e-02 2.26387918e-01 7.83575177e-01
4.10583496e-01 1.53267473e-01 3.12280685e-01 3.89169663e-01
3.66411544e-02 1.85884424e-02 -2.81423897e-01 -2.51692295e-01
4.94739145e-01 1.12080395e+00 -9.54607546e-01 -1.72374576e-01
-6.01594031e-01 7.47214019e-01 6.09149523e-02 4.29281682e-01
-9.84940171e-01 -3.35362941e-01 4.18934911e-01 -7.80369267e-02
9.93100330e-02 -3.03974658e-01 -5.81867635e-01 -9.93486404e-01
1.67601734e-01 -6.17266238e-01 4.81166869e-01 -3.92198890e-01
-1.07307100e+00 5.79743028e-01 -3.28080595e-01 -1.57359445e+00
-1.85964257e-01 -6.42787635e-01 -6.12128317e-01 4.95031655e-01
-1.49907982e+00 -7.19692647e-01 -3.79368812e-01 4.40003961e-01
6.60382271e-01 2.20065601e-02 5.71217716e-01 7.16492772e-01
-7.56366253e-01 7.51168907e-01 1.06555730e-01 2.62549445e-02
3.03221494e-01 -9.10053313e-01 2.14700192e-01 8.61256003e-01
-7.10013881e-02 3.38925481e-01 3.58440995e-01 -4.03387368e-01
-1.56240869e+00 -1.04293132e+00 3.92170370e-01 2.72334903e-01
5.64975441e-01 -2.99317151e-01 -5.62912762e-01 2.72882819e-01
-2.10760981e-02 1.10077664e-01 7.29052961e-01 1.02195062e-01
2.02297434e-01 -6.15625739e-01 -8.55175495e-01 7.25029051e-01
9.56004679e-01 -4.42757905e-01 1.97814092e-01 3.35676335e-02
4.39931542e-01 -1.87491477e-01 -9.03573275e-01 7.07939029e-01
5.42494178e-01 -7.79200554e-01 4.34140116e-01 -2.25307167e-01
8.79393592e-02 -2.70786583e-01 9.24637020e-02 -1.18629503e+00
-1.86687902e-01 -6.95915580e-01 -6.23138428e-01 1.05141151e+00
3.24929386e-01 -6.95020556e-01 1.16413617e+00 3.80912691e-01
-1.11460753e-01 -1.05010414e+00 -7.79057741e-01 -8.03331971e-01
-4.57274407e-01 -7.15534985e-01 5.43560922e-01 7.42640913e-01
1.88643068e-01 2.62014359e-01 -2.93287963e-01 2.54235268e-01
4.68113452e-01 5.10376513e-01 5.68626702e-01 -1.27984059e+00
-5.37968993e-01 -6.18784010e-01 -4.51453507e-01 -1.19695938e+00
-8.00440684e-02 -6.02155387e-01 -1.05517492e-01 -1.20372212e+00
-4.48231287e-02 -7.79571533e-01 -6.49494231e-01 4.03025329e-01
-7.52672777e-02 7.52910748e-02 -1.18093975e-01 3.23777407e-01
-6.64845288e-01 5.87590933e-01 8.42411697e-01 -1.83820114e-01
-3.34365338e-01 4.29947853e-01 -5.52020550e-01 2.59248883e-01
7.85085082e-01 -1.57064050e-01 -7.45413601e-01 -4.77812380e-01
1.49345189e-01 1.88585833e-01 1.30919546e-01 -1.41667092e+00
6.03825748e-01 7.71719068e-02 3.08569968e-01 -3.70492727e-01
7.89107308e-02 -1.27355504e+00 1.99702699e-02 5.63826263e-01
-1.35420755e-01 1.85252488e-01 2.62428135e-01 6.78794265e-01
-2.52145767e-01 2.85400480e-01 4.48455483e-01 5.06232917e-01
-1.17948747e+00 5.66471398e-01 -2.08451509e-01 -2.86124915e-01
1.41495526e+00 -2.82842427e-01 -8.65660161e-02 -2.99791217e-01
-8.03874910e-01 4.30606216e-01 2.35171139e-01 5.31331480e-01
4.48043376e-01 -1.17534566e+00 -2.16390431e-01 4.03430343e-01
3.08177114e-01 -2.75184095e-01 1.25106931e-01 1.03392780e+00
-3.46826524e-01 2.83866584e-01 -1.10823646e-01 -9.22263801e-01
-8.92675996e-01 7.48031139e-01 2.77771324e-01 -1.15730561e-01
-5.16179323e-01 2.50114202e-01 -9.96462554e-02 2.94971526e-01
5.13363838e-01 -3.65610331e-01 -2.12921709e-01 1.92179047e-02
5.20937085e-01 3.58119130e-01 3.13584447e-01 -6.59054965e-02
-3.86711627e-01 2.88472503e-01 -5.04035279e-02 1.67380020e-01
1.26565564e+00 -2.35010624e-01 2.47703686e-01 4.25862730e-01
1.06282198e+00 -1.45569161e-01 -1.44370914e+00 -2.11443037e-01
2.66209513e-01 -1.53621525e-01 3.07630837e-01 -6.38650417e-01
-1.27817822e+00 7.98569798e-01 9.79622722e-01 6.32125914e-01
1.72594023e+00 -3.42775434e-01 8.84495735e-01 1.51100636e-01
7.37019897e-01 -1.09544253e+00 5.79683520e-02 3.26208770e-01
9.13668126e-02 -1.05140150e+00 -4.00264323e-01 -5.25481105e-01
-2.32546046e-01 1.21686530e+00 5.51455379e-01 2.01393113e-01
8.71824622e-01 4.16410089e-01 -2.07103211e-02 9.18461680e-02
-8.61472011e-01 -1.67636976e-01 -1.66214362e-01 1.75241202e-01
2.28967607e-01 -7.19607025e-02 -4.65633303e-01 7.47468114e-01
4.15386260e-01 4.95334119e-01 2.81877846e-01 1.09794068e+00
-4.63833734e-02 -1.25061238e+00 -2.16489118e-02 4.44566518e-01
-4.64898705e-01 1.65856063e-01 5.88988177e-02 4.85870570e-01
7.52216876e-02 9.47172344e-01 2.82591969e-01 -6.85569406e-01
1.40777081e-01 -1.86532259e-01 2.06107989e-01 -3.84817481e-01
-2.90465534e-01 8.45826864e-02 -1.49496928e-01 -3.84831101e-01
-3.09233725e-01 -4.86919463e-01 -1.26006150e+00 -1.70960590e-01
-3.44674915e-01 3.77940357e-01 8.23072970e-01 1.17066801e+00
7.54633069e-01 9.79197145e-01 1.11529624e+00 -7.52322972e-01
-5.88918030e-01 -5.26045322e-01 -4.54747140e-01 2.31129259e-01
3.30476284e-01 -4.89661872e-01 -2.71789283e-01 -8.45883042e-02] | [7.171858787536621, 2.6752352714538574] |
87763037-7821-4ed5-a218-b7a596957776 | occlusion-robust-face-recognition-based-on-1 | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Song_Occlusion_Robust_Face_Recognition_Based_on_Mask_Learning_With_Pairwise_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Song_Occlusion_Robust_Face_Recognition_Based_on_Mask_Learning_With_Pairwise_ICCV_2019_paper.pdf | Occlusion Robust Face Recognition Based on Mask Learning With Pairwise Differential Siamese Network | Deep Convolutional Neural Networks (CNNs) have been pushing the frontier of face recognition over past years. However, existing CNN models are far less accurate when handling partially occluded faces. These general face models generalize poorly for occlusions on variable facial areas. Inspired by the fact that human visual system explicitly ignores the occlusion and only focuses on the non-occluded facial areas, we propose a mask learning strategy to find and discard corrupted feature elements from recognition. A mask dictionary is firstly established by exploiting the differences between the top conv features of occluded and occlusion-free face pairs using innovatively designed pairwise differential siamese network (PDSN). Each item of this dictionary captures the correspondence between occluded facial areas and corrupted feature elements, which is named Feature Discarding Mask (FDM). When dealing with a face image with random partial occlusions, we generate its FDM by combining relevant dictionary items and then multiply it with the original features to eliminate those corrupted feature elements from recognition. Comprehensive experiments on both synthesized and realistic occluded face datasets show that the proposed algorithm significantly outperforms the state-of-the-art systems.
| [' Wei Liu', ' Changsong Liu', ' Zhifeng Li', ' Dihong Gong', 'Lingxue Song'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['robust-face-recognition'] | ['computer-vision'] | [ 1.25969857e-01 -7.72060603e-02 4.04843651e-02 -6.19617164e-01
-4.29014638e-02 -5.74362874e-02 3.72888923e-01 -5.49005389e-01
-1.88159898e-01 5.17911553e-01 1.45938277e-01 3.25567603e-01
-1.27975687e-01 -6.14355028e-01 -6.33290887e-01 -9.01242077e-01
2.17620991e-02 2.30683908e-01 -2.22616389e-01 -9.84800011e-02
9.32532623e-02 9.73880768e-01 -1.80457413e+00 3.34550261e-01
5.70926249e-01 1.23533762e+00 -2.37624124e-01 -2.67767996e-01
-3.94902855e-01 2.45204329e-01 -6.01642787e-01 -3.81657153e-01
8.24095011e-01 -3.66231292e-01 -3.86351705e-01 3.87059361e-01
9.18266237e-01 -3.37902427e-01 -5.85681975e-01 1.25930071e+00
4.56252515e-01 6.80748522e-02 5.95986545e-01 -1.38578081e+00
-7.73970664e-01 1.40131503e-01 -9.47509348e-01 2.03222021e-01
1.84614375e-01 -7.76521638e-02 4.87492591e-01 -1.48154175e+00
7.20365107e-01 1.61195922e+00 5.10820448e-01 6.75607979e-01
-1.05835366e+00 -8.21187317e-01 3.42976958e-01 2.75767714e-01
-1.80897069e+00 -7.74612427e-01 1.06973648e+00 -2.53032744e-01
5.74469268e-01 2.33558416e-01 5.03200591e-01 7.03823626e-01
1.29682615e-01 6.77597821e-01 7.20869720e-01 -2.74443477e-01
-7.47287571e-02 -5.76223619e-02 -1.36185393e-01 9.43351209e-01
2.63469964e-01 3.45511846e-02 -6.48325264e-01 -2.03534722e-01
7.23495543e-01 2.84279525e-01 -3.26087713e-01 -6.21763349e-01
-8.38633299e-01 7.04679489e-01 3.37469548e-01 3.73697132e-01
-5.24000883e-01 -2.76634276e-01 1.10960238e-01 3.76124769e-01
3.16172481e-01 -2.54547864e-01 -4.25925255e-01 6.53574347e-01
-1.08442891e+00 7.23699555e-02 6.89978778e-01 8.16105247e-01
1.09368193e+00 3.14899594e-01 -2.31095612e-01 1.07182324e+00
2.98731297e-01 3.13273728e-01 2.58657068e-01 -7.69706845e-01
2.35995620e-01 8.04113030e-01 -2.28267461e-01 -1.42258584e+00
-1.98731929e-01 -5.69652796e-01 -1.06909096e+00 3.13646495e-01
3.07320088e-01 -6.18883483e-02 -1.04454768e+00 1.76601648e+00
6.16172194e-01 2.17657387e-01 -1.41097650e-01 8.86112034e-01
1.01811397e+00 3.27313125e-01 -7.42762461e-02 -4.69971240e-01
1.18049872e+00 -8.03996503e-01 -7.67869234e-01 2.70858277e-02
2.20857218e-01 -8.07937145e-01 5.14340937e-01 2.57737994e-01
-9.41991150e-01 -8.49541426e-01 -1.05046928e+00 9.27080587e-02
-2.95930535e-01 4.93689567e-01 3.66303444e-01 6.35226011e-01
-1.00477946e+00 5.08260846e-01 -5.92918634e-01 -3.73639949e-02
1.04831421e+00 7.62934268e-01 -9.10965741e-01 -3.48286927e-01
-8.46068919e-01 6.56735957e-01 -1.28976524e-01 6.40007079e-01
-1.05308783e+00 -6.99743450e-01 -9.24907982e-01 2.05942675e-01
4.30267006e-01 -2.39759520e-01 4.01300758e-01 -1.47641003e+00
-1.16825271e+00 7.38168538e-01 -4.13729072e-01 -2.04681251e-02
4.23611134e-01 4.00999449e-02 -5.67658484e-01 -5.64771816e-02
-8.18485767e-03 9.43939567e-01 1.33345199e+00 -1.10797834e+00
-3.10178339e-01 -5.51944792e-01 -2.84895867e-01 1.08231008e-01
-4.21281487e-01 9.12663639e-02 -4.09020454e-01 -6.25055313e-01
6.34446144e-01 -6.69367075e-01 6.64774477e-02 6.37266576e-01
-2.14030966e-01 -3.64466965e-01 1.35852444e+00 -6.08569384e-01
8.03360760e-01 -2.44102788e+00 1.04482755e-01 3.52696538e-01
3.36081088e-02 5.34126937e-01 -6.25550687e-01 1.36123776e-01
-4.95248467e-01 -2.56112397e-01 -3.83613110e-01 -4.59718436e-01
-1.33171096e-01 3.81263703e-01 -1.35057494e-01 7.26715922e-01
5.27920306e-01 5.21058202e-01 -5.16873300e-01 -4.70399618e-01
1.48885205e-01 7.90402472e-01 -6.81430042e-01 -1.02299251e-01
-1.51296444e-02 5.73181212e-01 -2.86610991e-01 1.00788307e+00
1.46293092e+00 3.67550790e-01 1.88063309e-01 -4.12155479e-01
-8.36145729e-02 -2.34950393e-01 -1.60732830e+00 1.48819232e+00
2.19761711e-02 6.00517511e-01 3.14539403e-01 -1.13940001e+00
1.17422831e+00 3.00952166e-01 6.27464652e-01 -4.67603356e-01
3.52137566e-01 1.98364124e-01 1.51538149e-01 -5.31086743e-01
-1.83235675e-01 1.31328225e-01 6.20298862e-01 1.02836780e-01
3.88394177e-01 3.24707240e-01 9.19262320e-02 -1.64042309e-01
7.48763978e-01 9.12541673e-02 9.46807489e-02 -5.62264621e-01
1.23400164e+00 -5.88148773e-01 9.33944881e-01 2.83670753e-01
-4.98074174e-01 7.18773782e-01 5.15978813e-01 -1.03600132e+00
-5.51730633e-01 -9.23273623e-01 -3.02400231e-01 7.50098586e-01
1.20674133e-01 -3.20948027e-02 -9.14480448e-01 -7.95045912e-01
2.21849918e-01 -6.99350685e-02 -8.00557911e-01 -8.63559172e-02
-6.71018183e-01 -5.57714701e-01 2.37810791e-01 1.73274055e-01
9.00730431e-01 -1.33086431e+00 3.18129212e-02 3.17090075e-03
1.40066877e-01 -8.88087511e-01 -5.13394892e-01 -2.71932751e-01
-4.89785463e-01 -1.13519943e+00 -6.65929437e-01 -1.20746839e+00
1.11774504e+00 4.67739969e-01 6.44432366e-01 4.30025667e-01
-5.58135569e-01 -5.67585528e-02 8.12628269e-02 -1.11693479e-01
6.96183220e-02 -3.77997696e-01 3.16889256e-01 7.66588569e-01
5.65528452e-01 -6.69805706e-01 -5.67549944e-01 4.98673499e-01
-8.64209116e-01 -3.87104630e-01 4.81636763e-01 9.13320363e-01
7.02357829e-01 1.29944146e-01 4.98937368e-01 -5.34106433e-01
2.22487211e-01 -3.81554186e-01 -5.65182209e-01 2.93500572e-01
1.63844461e-03 -2.54421905e-02 5.17973304e-01 -6.42402232e-01
-9.89183903e-01 4.24089432e-01 -1.01908423e-01 -7.15907156e-01
-1.96563989e-01 4.41984870e-02 -7.55998254e-01 -6.04246616e-01
2.21443176e-01 1.94263697e-01 1.83259249e-01 -5.36829352e-01
5.80546595e-02 3.80566716e-01 4.80020642e-01 -3.67782444e-01
8.54750037e-01 8.61201942e-01 2.85367668e-01 -9.11586881e-01
-6.05882287e-01 -8.29853863e-02 -7.37051845e-01 -2.75072038e-01
5.51566124e-01 -7.52419174e-01 -7.76394963e-01 4.98301864e-01
-1.07230961e+00 4.28286701e-01 -2.73257822e-01 3.53041440e-01
-1.28288224e-01 3.78237456e-01 -1.56708241e-01 -6.75960660e-01
-1.00981265e-01 -1.32279980e+00 8.93418014e-01 3.78543556e-01
1.59714252e-01 -4.16762531e-01 -2.95572400e-01 7.61720315e-02
4.25076485e-01 1.84681565e-01 7.49054730e-01 -5.41158557e-01
-6.07388854e-01 -3.08516234e-01 -1.64347932e-01 6.50571406e-01
3.81897956e-01 1.47228137e-01 -1.00556397e+00 -2.85370886e-01
3.41155261e-01 5.88262873e-03 9.70497966e-01 3.65421802e-01
1.25030053e+00 -2.75988340e-01 -3.63913506e-01 8.42313707e-01
1.19219995e+00 1.91986397e-01 7.91776240e-01 -2.09450662e-01
5.88693261e-01 7.98689246e-01 3.07489514e-01 2.83382535e-01
-1.55878708e-01 5.18755436e-01 5.14471173e-01 -7.26202726e-02
-3.04720134e-01 -5.01552820e-02 1.97417185e-01 4.40946966e-01
1.27514035e-01 5.48419368e-04 -3.44266325e-01 5.05419314e-01
-1.45474970e+00 -8.50883842e-01 2.28118971e-01 1.99307966e+00
4.79089856e-01 -1.76732019e-01 -3.60361040e-01 1.75266802e-01
8.80400121e-01 3.34484905e-01 -5.39878309e-01 -1.40139222e-01
-5.17770946e-01 5.25554240e-01 -4.45978045e-02 4.05156285e-01
-1.08242726e+00 8.96571398e-01 5.38271141e+00 9.87372339e-01
-1.03575349e+00 1.13497555e-01 6.63440645e-01 -1.68583751e-01
-7.01037198e-02 4.45932709e-02 -9.43526804e-01 3.71247858e-01
2.08395407e-01 3.24722618e-01 3.51657599e-01 8.35117161e-01
-3.77828226e-04 -1.92311667e-02 -1.02419913e+00 1.06550312e+00
4.57322478e-01 -1.08959019e+00 4.13864672e-01 2.20840927e-02
1.02181828e+00 -3.40153784e-01 3.10785532e-01 1.05453856e-01
-3.13124239e-01 -1.14698768e+00 5.10903478e-01 7.46184349e-01
5.63879609e-01 -7.72607923e-01 7.39702463e-01 5.86903319e-02
-1.23566031e+00 -1.77359819e-01 -7.49687016e-01 4.33316603e-02
-2.91346461e-01 5.19963264e-01 -2.62555391e-01 3.41670364e-01
6.44757628e-01 6.11591339e-01 -3.37826848e-01 8.82116735e-01
3.99268083e-02 -3.99038531e-02 -4.80455965e-01 4.29825932e-01
1.56383738e-01 -4.54599500e-01 5.15591323e-01 6.87692523e-01
3.48296106e-01 1.18687853e-01 -5.61396405e-02 9.24021840e-01
-4.54844058e-01 2.56212026e-01 -5.89239836e-01 3.45266849e-01
4.65043545e-01 1.39778841e+00 -5.90791583e-01 -2.71515518e-01
-6.16087675e-01 8.80213559e-01 2.84623653e-01 3.58353198e-01
-6.05603218e-01 -3.38725522e-02 1.00131416e+00 -9.56586301e-02
6.11735582e-01 -1.41628176e-01 -9.52318013e-02 -9.00516033e-01
2.94192880e-01 -9.45958972e-01 1.77615449e-01 -3.71867537e-01
-1.16473663e+00 8.02177250e-01 -1.88838929e-01 -1.26405668e+00
3.68739247e-01 -5.77841699e-01 -5.58564782e-01 9.23682928e-01
-1.47277117e+00 -9.59906220e-01 -4.26710010e-01 8.27648461e-01
6.66789770e-01 -4.89765614e-01 6.39845431e-01 5.64498723e-01
-7.36001730e-01 8.42128515e-01 -8.10662806e-02 2.59128600e-01
5.37632346e-01 -2.03935727e-01 2.88953837e-02 7.36039996e-01
1.99171737e-01 7.58797824e-01 3.60638261e-01 -5.87158263e-01
-1.26887929e+00 -1.05368185e+00 9.14523900e-01 9.07599106e-02
3.21773700e-02 -2.78300762e-01 -1.13511825e+00 4.34376329e-01
8.80845785e-02 6.01076841e-01 3.67461979e-01 -4.76571470e-01
-4.95850503e-01 -5.41032374e-01 -1.46005607e+00 5.96409976e-01
1.28903687e+00 -3.02962929e-01 -3.39692831e-01 3.95945966e-01
2.20317349e-01 -1.47684008e-01 -4.13867325e-01 7.83856392e-01
6.98005021e-01 -1.10152209e+00 9.09016728e-01 -5.74824572e-01
-9.15411562e-02 -5.16939163e-01 -2.31102243e-01 -8.74435067e-01
-2.75809050e-01 -5.64310670e-01 -5.30969650e-02 1.10193515e+00
3.60373221e-03 -6.20411694e-01 8.97221327e-01 3.48598450e-01
-1.23332597e-01 -8.12227249e-01 -1.28667951e+00 -7.20398307e-01
-2.32429266e-01 1.73105791e-01 9.02121365e-01 9.12873507e-01
-5.78417599e-01 -2.58301497e-01 -3.76822919e-01 2.57992625e-01
6.11812234e-01 9.25327744e-03 5.43546438e-01 -1.36193383e+00
4.04094487e-01 -3.56814474e-01 -6.67185605e-01 -8.41112375e-01
5.95244348e-01 -6.66822195e-01 -1.35393858e-01 -8.75522494e-01
8.21285322e-02 -2.61908263e-01 -2.77329922e-01 5.92250526e-01
1.80322424e-01 7.23782480e-01 5.06945588e-02 2.07170472e-01
-1.51630998e-01 7.69130170e-01 1.42904294e+00 -2.55452335e-01
-1.66099235e-01 -1.37457848e-01 -4.67811674e-01 9.20632958e-01
5.79798043e-01 -5.10380864e-01 -1.91739976e-01 -4.87337440e-01
-5.15868485e-01 -2.74581313e-01 2.93534160e-01 -1.38771605e+00
3.10445338e-01 -3.61549780e-02 8.87579739e-01 -6.18199348e-01
5.44571459e-01 -1.07665765e+00 3.85598093e-01 4.62868273e-01
-7.03970566e-02 -2.03617662e-01 2.17398167e-01 3.76906872e-01
-3.90769064e-01 -2.18008578e-01 1.00713038e+00 -1.05652912e-02
-5.43377101e-01 7.71943212e-01 -9.62207913e-02 -4.23681974e-01
9.82243121e-01 -4.91190553e-01 3.71735319e-02 -1.67880997e-01
-9.23689783e-01 -1.81008518e-01 1.21634305e-01 5.93235433e-01
8.35391343e-01 -1.45237064e+00 -8.22494984e-01 1.25335670e+00
-2.78303117e-01 -1.91429123e-01 6.33040428e-01 7.61013031e-01
-4.47283208e-01 3.55109513e-01 -5.31992614e-01 -4.78119314e-01
-1.41788769e+00 5.57371855e-01 5.29697597e-01 2.46329904e-01
-3.66089076e-01 1.05282974e+00 4.96305376e-01 -9.26707983e-02
4.20787841e-01 -5.12463078e-02 -3.50866139e-01 1.21404365e-01
5.67989588e-01 1.58229977e-01 2.27557227e-01 -1.20275009e+00
-5.24720490e-01 8.52864504e-01 -1.96113616e-01 2.61135221e-01
1.33680689e+00 7.19525144e-02 -4.51483220e-01 -2.93450177e-01
1.51546431e+00 -3.82895116e-03 -1.34494472e+00 -4.58244711e-01
-3.24835747e-01 -8.32764208e-01 -1.13011807e-01 -2.74506062e-01
-1.88266420e+00 7.68501461e-01 7.59442270e-01 -2.21367195e-01
1.31500089e+00 -3.36685121e-01 5.97078264e-01 2.43047059e-01
2.05874860e-01 -7.79703915e-01 1.41398683e-01 5.08347392e-01
1.22993124e+00 -1.05214489e+00 4.16242741e-02 -6.65757656e-01
-2.19007045e-01 1.13158441e+00 8.21845472e-01 -4.75069255e-01
1.13922966e+00 3.10142208e-02 -6.46753982e-02 -2.26371750e-01
-4.14295912e-01 -1.05437428e-01 3.76050025e-01 6.50220275e-01
4.42807078e-02 -3.15824091e-01 -4.40650851e-01 3.75513315e-01
6.99383542e-02 -9.84949619e-02 -4.07683756e-03 8.47803295e-01
-3.89761120e-01 -1.05272973e+00 -4.79889512e-01 4.17308748e-01
-3.29086989e-01 3.11984476e-02 -2.98573673e-01 8.42317045e-01
8.08373332e-01 6.93135917e-01 3.30046266e-01 -2.13394880e-01
2.12920532e-01 1.99124560e-01 5.58997929e-01 -4.67452705e-01
-4.30778891e-01 1.02795742e-01 -4.10200506e-01 -6.07279420e-01
-4.29254025e-01 -7.72770464e-01 -9.39787447e-01 -3.05619419e-01
-3.79075468e-01 6.65905625e-02 5.99638820e-01 8.69424999e-01
4.84605968e-01 2.59780943e-01 7.69792855e-01 -9.73741055e-01
-4.08352464e-01 -9.55329776e-01 -6.79566264e-01 3.57627600e-01
4.42419022e-01 -1.11935997e+00 -5.50790668e-01 4.60035838e-02] | [13.201336860656738, 0.4604114294052124] |
565f51a1-a0f6-4cba-b78b-7134be42b340 | spatio-temporal-lstm-with-trust-gates-for-3d | 1607.07043 | null | http://arxiv.org/abs/1607.07043v1 | http://arxiv.org/pdf/1607.07043v1.pdf | Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition | 3D action recognition - analysis of human actions based on 3D skeleton data -
becomes popular recently due to its succinctness, robustness, and
view-invariant representation. Recent attempts on this problem suggested to
develop RNN-based learning methods to model the contextual dependency in the
temporal domain. In this paper, we extend this idea to spatio-temporal domains
to analyze the hidden sources of action-related information within the input
data over both domains concurrently. Inspired by the graphical structure of the
human skeleton, we further propose a more powerful tree-structure based
traversal method. To handle the noise and occlusion in 3D skeleton data, we
introduce new gating mechanism within LSTM to learn the reliability of the
sequential input data and accordingly adjust its effect on updating the
long-term context information stored in the memory cell. Our method achieves
state-of-the-art performance on 4 challenging benchmark datasets for 3D human
action analysis. | ['Gang Wang', 'Amir Shahroudy', 'Jun Liu', 'Dong Xu'] | 2016-07-24 | null | null | null | null | ['action-analysis', '3d-human-action-recognition'] | ['computer-vision', 'computer-vision'] | [ 3.38381767e-01 -3.36723775e-01 -3.09278518e-01 -4.68930632e-01
-3.27524215e-01 2.86877751e-02 4.20271069e-01 -1.48967439e-02
-5.25322020e-01 4.95448649e-01 7.61946917e-01 -4.32960950e-02
-8.52761716e-02 -6.18290782e-01 -5.11688948e-01 -6.18015468e-01
-3.08636248e-01 9.03290287e-02 7.03752160e-01 -5.11008017e-02
3.50711912e-01 7.00647473e-01 -1.58222163e+00 6.53170586e-01
2.51019537e-01 9.48161960e-01 2.45833442e-01 5.86050808e-01
-1.99275702e-01 1.22666943e+00 -4.25924897e-01 3.09093118e-01
7.32907429e-02 -4.75670099e-01 -8.36129904e-01 2.79770404e-01
1.63091674e-01 -3.97764713e-01 -7.00178385e-01 6.01933777e-01
7.05373049e-01 4.90362912e-01 4.47277516e-01 -9.46785569e-01
-1.44517288e-01 1.76216170e-01 -6.62993670e-01 6.37507319e-01
4.81586963e-01 2.54903227e-01 8.22296798e-01 -5.93631506e-01
6.57549381e-01 1.44125330e+00 3.09900701e-01 6.82268798e-01
-8.51448953e-01 -2.68917412e-01 8.29985082e-01 7.61083603e-01
-8.43981326e-01 -2.73277521e-01 9.96787965e-01 -3.13367724e-01
1.56121528e+00 -1.15720935e-01 1.06493497e+00 1.51632822e+00
5.63690662e-01 1.26085341e+00 1.00565374e+00 -3.91341954e-01
1.31789833e-01 -9.86449957e-01 1.21667184e-01 8.72374833e-01
-3.11158925e-01 8.35640803e-02 -9.42912281e-01 1.34086698e-01
9.66512859e-01 2.02073932e-01 1.64247826e-01 -5.76813161e-01
-1.14326143e+00 3.70974958e-01 2.72304088e-01 3.97119164e-01
-5.08590758e-01 4.14968461e-01 7.07653880e-01 3.73310968e-02
3.54458392e-01 -1.32044703e-01 -6.96272314e-01 -4.47779834e-01
-5.20848513e-01 1.62397921e-01 1.90604225e-01 6.29638433e-01
1.61597699e-01 6.82517290e-02 -4.38917011e-01 6.07028246e-01
4.75413084e-01 1.40110329e-01 7.22563863e-01 -8.19667101e-01
7.27325439e-01 9.64675069e-01 -1.98318854e-01 -8.19404662e-01
-8.24137688e-01 -2.97180861e-01 -7.70364761e-01 3.76256794e-01
3.74409169e-01 3.87192786e-01 -1.27802038e+00 1.86055338e+00
5.59313893e-01 9.32507962e-02 -1.34328753e-01 8.36277962e-01
6.68698430e-01 3.07055324e-01 3.72473150e-01 -4.77581136e-02
1.19989550e+00 -9.30655062e-01 -7.54207492e-01 -2.02243000e-01
6.91112459e-01 -2.39575967e-01 9.27078724e-01 2.74583399e-01
-9.27298486e-01 -9.02394593e-01 -9.94824588e-01 -3.33564848e-01
-2.70787686e-01 7.78858438e-02 3.76833856e-01 6.85010403e-02
-6.33336127e-01 6.84949100e-01 -1.39531875e+00 -4.91128743e-01
5.17377794e-01 3.24913114e-01 -5.34712613e-01 7.49190822e-02
-1.13997817e+00 8.74240339e-01 3.89975607e-01 2.53467262e-01
-1.05603230e+00 1.51697904e-01 -9.24076557e-01 -3.56695205e-01
6.00414991e-01 -6.78253829e-01 9.78549004e-01 -7.16067731e-01
-1.66034448e+00 8.94853830e-01 -2.81075567e-01 -4.22641844e-01
2.74037808e-01 -4.50602144e-01 -2.01802939e-01 1.05859965e-01
-8.61223321e-03 5.02009153e-01 9.47863758e-01 -6.94362164e-01
-6.06131613e-01 -9.51387644e-01 3.32147069e-02 4.56230193e-01
2.62953099e-02 -1.53783590e-01 -5.95602930e-01 -9.11710262e-01
5.92712045e-01 -9.10852849e-01 -4.32460427e-01 -2.43224427e-02
-1.04001544e-01 -3.64247143e-01 6.72333419e-01 -7.29325473e-01
1.26480663e+00 -2.10911298e+00 6.04348302e-01 -7.95738306e-03
-7.59232640e-02 3.35735410e-01 -2.13203102e-01 3.27836096e-01
-1.35134190e-01 -3.20850998e-01 -1.95723414e-01 -3.73857111e-01
-1.47441000e-01 5.81610382e-01 -1.89240411e-01 4.41428185e-01
2.04177856e-01 1.03580153e+00 -7.59497046e-01 -6.08847857e-01
4.14718300e-01 4.92043048e-01 -2.56755263e-01 2.69515872e-01
-2.29278475e-01 6.70988262e-01 -8.20829153e-01 7.59653091e-01
1.62036672e-01 -9.88971740e-02 2.44610280e-01 -1.47877887e-01
4.38804626e-02 3.92736524e-01 -1.12544203e+00 2.37962556e+00
-1.99148774e-01 2.30702922e-01 -2.64831007e-01 -1.10989952e+00
7.68078566e-01 2.61157483e-01 6.50326967e-01 -1.07082438e+00
1.69497579e-01 -1.90604836e-01 3.73882018e-02 -9.45429623e-01
1.43277079e-01 -3.15023400e-02 -1.00879269e-02 4.43937004e-01
1.16594555e-02 4.72299308e-01 7.74868065e-03 -1.73043504e-01
1.37662911e+00 9.88140225e-01 3.88163328e-01 1.59351662e-01
7.28305936e-01 -3.73958081e-01 6.40136063e-01 4.37069595e-01
-5.50388694e-01 2.68602908e-01 4.27500278e-01 -8.30765843e-01
-5.82272410e-01 -8.86076152e-01 3.41024637e-01 1.28449190e+00
6.91191852e-03 -3.66105765e-01 -5.22489607e-01 -9.81422186e-01
-1.89034343e-01 4.25895184e-01 -9.72924829e-01 -3.04443598e-01
-9.96892869e-01 -4.43816036e-01 4.41819131e-01 1.07532358e+00
7.85797477e-01 -1.33105886e+00 -1.17102146e+00 4.44080800e-01
-3.22000742e-01 -1.05068696e+00 -1.52675733e-01 3.92610133e-01
-1.48119199e+00 -1.25349724e+00 -5.01578450e-01 -5.67955077e-01
5.22308886e-01 5.10908850e-02 7.95698583e-01 -3.10066976e-02
-1.67548656e-01 5.95558524e-01 -3.89052451e-01 -7.28772134e-02
-3.52185518e-02 2.39061136e-02 1.75693989e-04 1.04702249e-01
4.43165183e-01 -7.13901341e-01 -4.50730562e-01 3.72031331e-01
-9.05104756e-01 4.06535305e-02 5.10709524e-01 7.13202953e-01
8.76582801e-01 1.06602404e-02 2.67557859e-01 -4.14797395e-01
1.76704288e-01 -6.16427278e-03 -1.45840347e-01 3.17017615e-01
-1.83607861e-01 4.55872476e-01 2.29577437e-01 -4.19656575e-01
-1.24523044e+00 4.38049644e-01 -1.80951267e-01 -5.37594080e-01
-3.77995223e-01 4.42028999e-01 -4.12186980e-01 3.98067445e-01
3.37470382e-01 3.53144825e-01 -1.67163804e-01 -6.53693438e-01
1.27551585e-01 1.60000846e-01 3.78096223e-01 -5.49264789e-01
2.25658372e-01 8.04898560e-01 4.14598227e-01 -4.92431819e-01
-9.57913637e-01 -2.64058560e-01 -1.15805411e+00 -4.76257890e-01
1.25342977e+00 -8.25585485e-01 -6.33720040e-01 8.70096207e-01
-1.22123790e+00 -4.77100909e-01 -1.97247818e-01 6.23510897e-01
-7.99964011e-01 5.18952668e-01 -6.13188148e-01 -7.12335587e-01
-2.15078704e-02 -1.08750129e+00 1.20325410e+00 -2.36144930e-01
-1.43142045e-01 -7.83411741e-01 3.18074554e-01 3.45668972e-01
-1.20740585e-01 4.17512655e-01 9.87435699e-01 -3.76326948e-01
-4.47863758e-01 -1.41839936e-01 3.58960219e-02 3.19005221e-01
1.78995848e-01 -3.01634341e-01 -7.53729701e-01 -7.98529834e-02
1.82533085e-01 -2.82256514e-01 1.16105998e+00 5.61200440e-01
1.36998594e+00 7.86561891e-02 -2.60524154e-01 3.72497380e-01
8.54081750e-01 9.89312381e-02 7.61218548e-01 3.48343551e-01
8.19545150e-01 7.25446165e-01 7.29900956e-01 5.44383943e-01
2.41691500e-01 7.58363307e-01 5.46832263e-01 2.08931074e-01
-1.69263124e-01 -5.41289389e-01 5.41839838e-01 5.89181483e-01
-5.56046844e-01 -2.15163156e-01 -6.18820965e-01 2.21423760e-01
-2.28241992e+00 -1.12925732e+00 3.01558618e-03 2.07790160e+00
5.74532449e-01 5.88682830e-01 1.76491991e-01 1.75703555e-01
5.24424076e-01 6.25013828e-01 -7.12353051e-01 -2.02711523e-01
-1.43931005e-02 1.51249155e-01 2.51785249e-01 2.68556178e-01
-1.19684649e+00 9.72463787e-01 5.97151804e+00 7.18115211e-01
-8.76662195e-01 -6.93041924e-03 3.70663881e-01 -3.24510992e-01
1.23461910e-01 -2.01161832e-01 -7.37680733e-01 2.45742798e-01
5.90809941e-01 5.64687312e-01 -9.34291631e-02 6.01322949e-01
5.37746251e-01 -4.11045611e-01 -1.24051046e+00 1.03264403e+00
9.18459296e-02 -9.41810191e-01 2.20558107e-01 1.06721908e-01
2.92700648e-01 -2.86508352e-02 -1.43966794e-01 2.43308157e-01
-1.22587994e-01 -9.38262284e-01 5.90619385e-01 8.74951422e-01
4.64006901e-01 -6.24817252e-01 3.45774174e-01 4.68835652e-01
-1.40940928e+00 -2.98354894e-01 -9.35038477e-02 -4.23042297e-01
3.80458385e-01 1.28206253e-01 -2.44695395e-01 4.25171137e-01
7.78416038e-01 1.21137524e+00 -7.10782349e-01 5.95569015e-01
-5.15429914e-01 3.06576550e-01 -1.64007902e-01 1.36405066e-01
3.72541666e-01 9.33816582e-02 5.71584284e-01 9.30604279e-01
6.45992607e-02 3.41735333e-01 1.79024488e-01 2.33030751e-01
3.02999586e-01 -3.03323150e-01 -6.44946694e-01 1.16764344e-01
-1.87542111e-01 7.16128528e-01 -8.49444866e-01 -2.04129919e-01
-3.80271584e-01 1.08635545e+00 4.61155623e-01 2.95593441e-01
-9.08015072e-01 1.47230580e-01 4.97056544e-01 1.06948480e-01
5.01274467e-01 -7.22082973e-01 -1.14166014e-01 -9.87403691e-01
3.26862037e-01 -8.68085802e-01 8.01365197e-01 -6.62955642e-01
-9.85147297e-01 4.21631694e-01 1.28172174e-01 -1.41003501e+00
-2.51084298e-01 -6.60305321e-01 -2.72098899e-01 4.34502870e-01
-1.28707111e+00 -1.23720229e+00 -1.42550915e-01 9.57374513e-01
8.69801044e-01 -8.23065266e-02 7.81011701e-01 3.89203578e-02
-4.51257229e-01 1.19795747e-01 -4.08751786e-01 1.73487380e-01
4.00382906e-01 -8.79167318e-01 5.97157538e-01 8.07051778e-01
3.34349126e-01 2.77397037e-01 2.88441628e-01 -9.20203567e-01
-1.22144485e+00 -7.55971909e-01 9.52001572e-01 -6.84674144e-01
4.05292094e-01 -1.95705906e-01 -7.48139083e-01 6.23317540e-01
-9.57234427e-02 1.88233498e-02 6.15479887e-01 -7.61546893e-03
-4.27090228e-01 7.85334874e-03 -7.66876459e-01 5.65990627e-01
1.72941053e+00 -5.44205308e-01 -1.00952876e+00 -2.77092531e-02
5.32915354e-01 -4.94317174e-01 -5.33030093e-01 6.33316219e-01
7.57293761e-01 -1.09720016e+00 8.91320169e-01 -1.06563735e+00
3.31580698e-01 -4.49856430e-01 -3.83405536e-01 -7.39306390e-01
-4.89127576e-01 -3.28657836e-01 -4.67337579e-01 5.61159790e-01
-4.98339124e-02 -1.01798229e-01 8.95646393e-01 4.25491691e-01
-2.15361342e-01 -1.02673864e+00 -1.34664571e+00 -5.19800365e-01
-4.76222038e-01 -7.35846221e-01 2.73133606e-01 4.59316015e-01
-1.89766258e-01 3.30156893e-01 -4.52593565e-01 -4.83361147e-02
3.94440025e-01 6.28523156e-02 7.29032636e-01 -9.52856243e-01
-2.95388192e-01 -1.50917009e-01 -7.11763263e-01 -1.41164899e+00
3.67978096e-01 -6.11828446e-01 -5.48886433e-02 -1.67506635e+00
-1.45610748e-02 9.50438604e-02 -6.51385427e-01 7.48688161e-01
6.62866533e-02 2.79174955e-03 1.47303462e-01 4.03736997e-03
-8.63714755e-01 8.34463954e-01 1.39547932e+00 -9.41291302e-02
-3.34594339e-01 2.31724828e-01 1.12895578e-01 8.57819736e-01
5.55930436e-01 -3.82464349e-01 -5.11923373e-01 -8.25794697e-01
9.88557637e-02 5.25958091e-02 5.40271819e-01 -1.22124076e+00
1.49777427e-01 -2.55001724e-01 7.01021075e-01 -1.02438176e+00
5.31815708e-01 -8.50338995e-01 -2.05935776e-01 4.94999260e-01
-5.55411935e-01 2.69185513e-01 2.35646948e-01 9.25293148e-01
-1.57383010e-01 8.82421285e-02 5.39429307e-01 -5.52820206e-01
-1.07046247e+00 4.80014980e-01 -6.93755627e-01 7.63495043e-02
8.38470459e-01 -4.14715648e-01 2.24602580e-01 -1.07806161e-01
-1.05241060e+00 4.62448597e-03 2.09436417e-01 6.13479376e-01
9.32003975e-01 -1.44839275e+00 -2.86744446e-01 4.04054910e-01
1.20770589e-01 -1.41328990e-01 5.77980995e-01 7.86977530e-01
-4.05088589e-02 3.07704359e-01 -6.07488215e-01 -6.13249838e-01
-1.39398241e+00 5.30770600e-01 4.28412646e-01 -5.96806705e-01
-7.34770596e-01 8.56452763e-01 5.27253561e-02 -1.17660768e-01
5.87196171e-01 -6.57951832e-01 -3.89160335e-01 1.40838668e-01
4.11898553e-01 5.03091156e-01 -2.53899042e-02 -8.15596402e-01
-6.27341211e-01 7.66820133e-01 1.28258541e-01 -2.06700996e-01
1.31362343e+00 -1.01516530e-01 6.37254193e-02 7.89006472e-01
1.07659662e+00 -5.86192310e-01 -1.58455765e+00 -3.71008337e-01
1.41187206e-01 -4.83289748e-01 1.57848131e-02 -7.11568058e-01
-1.01732111e+00 1.03845882e+00 7.60480404e-01 -2.64042318e-01
1.15899789e+00 -7.17854649e-02 7.87050486e-01 5.09294629e-01
3.01939070e-01 -1.47821212e+00 5.21396816e-01 7.82086790e-01
8.46608400e-01 -1.10932672e+00 2.87539750e-01 2.34450370e-01
-6.34172618e-01 1.17946112e+00 7.57667303e-01 -3.09314523e-02
7.19830096e-01 -1.38697907e-01 -6.23426735e-02 -3.26552510e-01
-8.06747615e-01 -4.38029945e-01 2.92261690e-01 6.12207949e-01
3.01736176e-01 -2.85459250e-01 -1.13230139e-01 4.41404194e-01
5.30204952e-01 1.64872080e-01 -3.72206420e-02 1.40830064e+00
-3.71312290e-01 -1.27002835e+00 -5.06168604e-02 2.05960408e-01
-2.84186870e-01 2.59814769e-01 -3.49826306e-01 6.81251287e-01
1.75959840e-01 6.69992685e-01 -9.43660289e-02 -4.63606954e-01
6.51403785e-01 3.81082356e-01 8.99567127e-01 -4.75888491e-01
-4.35466230e-01 3.66054624e-01 1.86155066e-02 -1.18134701e+00
-1.15497828e+00 -8.63695860e-01 -1.57223308e+00 3.48006219e-01
8.51192251e-02 -4.35732365e-01 2.86926478e-01 1.28957665e+00
5.22685170e-01 6.26589894e-01 3.37985545e-01 -9.38115418e-01
-4.90396798e-01 -9.32798982e-01 -4.87960994e-01 3.72778118e-01
3.20518404e-01 -1.07696521e+00 2.66141351e-02 -5.16517237e-02] | [7.879940032958984, 0.4545818865299225] |
0dbc529f-b183-4f1b-913f-f52406a84afe | 190909986 | 1909.09986 | null | https://arxiv.org/abs/1909.09986v1 | https://arxiv.org/pdf/1909.09986v1.pdf | Improving Quality and Efficiency in Plan-based Neural Data-to-Text Generation | We follow the step-by-step approach to neural data-to-text generation we proposed in Moryossef et al (2019), in which the generation process is divided into a text-planning stage followed by a plan-realization stage. We suggest four extensions to that framework: (1) we introduce a trainable neural planning component that can generate effective plans several orders of magnitude faster than the original planner; (2) we incorporate typing hints that improve the model's ability to deal with unseen relations and entities; (3) we introduce a verification-by-reranking stage that substantially improves the faithfulness of the resulting texts; (4) we incorporate a simple but effective referring expression generation module. These extensions result in a generation process that is faster, more fluent, and more accurate. | ['Yoav Goldberg', 'Ido Dagan', 'Amit Moryossef'] | 2019-09-22 | improving-quality-and-efficiency-in-plan | https://aclanthology.org/W19-8645 | https://aclanthology.org/W19-8645.pdf | ws-2019-10 | ['referring-expression-generation'] | ['computer-vision'] | [ 3.96195263e-01 1.12913752e+00 -2.02522725e-01 -3.90560180e-01
-1.10667717e+00 -7.21985459e-01 1.18264389e+00 3.12638223e-01
-2.21569419e-01 1.05189705e+00 1.07854533e+00 -4.74674731e-01
8.84374902e-02 -1.19628167e+00 -5.69530249e-01 2.23054960e-01
1.02968067e-01 9.13793445e-01 3.22312176e-01 -5.95682859e-01
2.60663331e-01 2.54986405e-01 -1.18135047e+00 3.72636974e-01
8.04627538e-01 5.03548145e-01 -1.43371895e-01 6.33181810e-01
-1.52248874e-01 1.27618802e+00 -5.28668821e-01 -6.81297600e-01
1.28773615e-01 -5.64289212e-01 -1.45368373e+00 1.04122445e-01
1.40362857e-02 -6.07009709e-01 -1.89733043e-01 6.07731521e-01
2.33748898e-01 3.98836523e-01 6.35068417e-01 -9.37492251e-01
-6.56584442e-01 1.41118407e+00 -1.61865786e-01 -2.45581448e-01
7.64229476e-01 2.22371340e-01 1.14265001e+00 -6.43016458e-01
1.05824256e+00 1.23169565e+00 5.30048013e-01 9.03996050e-01
-1.03034329e+00 -3.69727969e-01 1.75412223e-01 -1.11397624e-01
-1.32916188e+00 -7.29605436e-01 5.02205729e-01 -1.00988410e-01
1.49233949e+00 2.20476091e-01 7.32182860e-01 8.67918313e-01
-5.89118488e-02 7.81730711e-01 5.39305031e-01 -7.99938560e-01
1.32412210e-01 -2.34109581e-01 7.63854533e-02 7.83529878e-01
1.50621086e-01 1.11688420e-01 -4.20332879e-01 -1.29647344e-01
7.65443802e-01 -6.05140567e-01 -1.10186018e-01 3.58021706e-02
-1.30598772e+00 9.36370134e-01 3.12211156e-01 3.98396224e-01
-4.39596862e-01 1.57669380e-01 4.59938943e-01 6.96965382e-02
2.59453714e-01 1.04055989e+00 -1.78762898e-01 -2.54631996e-01
-9.94994223e-01 8.42333913e-01 1.17770457e+00 1.23347044e+00
4.90769684e-01 7.29480460e-02 -5.17855644e-01 5.21991372e-01
2.45501518e-01 6.24636821e-02 4.47657466e-01 -1.23622096e+00
8.26444030e-01 5.87641776e-01 4.25529569e-01 -6.64658010e-01
-4.63948369e-01 -1.28554016e-01 -3.67389798e-01 -1.11634798e-01
4.12490040e-01 -4.65013415e-01 -8.86954784e-01 1.93800271e+00
2.78515145e-02 -3.90939474e-01 2.84862012e-01 4.78007704e-01
8.19211423e-01 6.31129622e-01 1.55902073e-01 -1.30949080e-01
1.18057418e+00 -1.07647169e+00 -7.28791773e-01 -2.27320120e-01
9.50913668e-01 -4.50503379e-01 1.05316544e+00 6.97671324e-02
-1.68978310e+00 -3.44297767e-01 -9.85417306e-01 -4.29117858e-01
-3.67713988e-01 -1.17759272e-01 1.07517350e+00 4.01472926e-01
-1.31295526e+00 7.08102643e-01 -8.79040420e-01 -3.94353300e-01
3.23387086e-01 3.59589875e-01 -3.06662381e-01 3.16168554e-02
-1.30859959e+00 1.29038000e+00 9.67520475e-01 -1.43373773e-01
-6.56104445e-01 -2.75783509e-01 -1.23094070e+00 7.11068064e-02
6.68363333e-01 -1.13976073e+00 1.85950029e+00 -7.11859822e-01
-1.78557813e+00 4.85174954e-01 -4.26725209e-01 -4.37704861e-01
4.33619857e-01 -1.97487801e-01 -2.48781033e-03 5.63496761e-02
1.63771302e-01 1.12866402e+00 7.82438740e-02 -1.15739632e+00
-5.47889292e-01 -3.83835323e-02 3.85635018e-01 5.11825085e-01
1.78429186e-01 1.49528340e-01 -4.57572788e-01 -6.59023643e-01
2.29419060e-02 -8.56127501e-01 -3.23508561e-01 -3.38671029e-01
-8.66197765e-01 -4.04183328e-01 2.06902176e-01 -7.42816925e-01
1.30779219e+00 -1.46729159e+00 2.15665713e-01 1.17424347e-01
1.85851559e-01 2.74126619e-01 -3.25239509e-01 8.83401155e-01
-1.11854203e-01 3.48258913e-01 -2.51655072e-01 -4.66810137e-01
1.53072491e-01 1.68957248e-01 -4.02735651e-01 -3.46982747e-01
5.32724500e-01 1.36117125e+00 -1.02666426e+00 -5.36468446e-01
-1.27708822e-01 4.34222668e-02 -7.04630673e-01 2.46352568e-01
-7.13843644e-01 2.91566253e-01 -5.44349790e-01 5.12771964e-01
-1.24123678e-01 -1.93434969e-01 1.98629096e-01 9.19566378e-02
-1.90792009e-01 9.95285511e-01 -1.08895087e+00 1.64521182e+00
-4.05594110e-01 4.30523664e-01 -3.01940322e-01 -2.64761806e-01
6.20947063e-01 5.98030746e-01 -9.67404395e-02 -2.19750896e-01
8.75171423e-02 7.19648376e-02 2.19515394e-02 -3.54301453e-01
1.13394845e+00 -3.79194349e-01 -2.92973846e-01 9.50679421e-01
-5.15595116e-02 -5.24777412e-01 6.60387933e-01 5.50011694e-01
1.16765201e+00 6.05850220e-01 5.54129958e-01 8.89343321e-02
2.16697603e-01 3.61393005e-01 4.46761310e-01 7.74491012e-01
2.35808969e-01 5.54210186e-01 6.46496475e-01 -2.36178026e-01
-1.18562424e+00 -5.97083390e-01 2.79200226e-01 9.52526748e-01
-3.29618722e-01 -6.87904179e-01 -9.26660955e-01 -6.62031889e-01
-3.72676343e-01 1.40916681e+00 -5.61315417e-01 7.53962845e-02
-9.59107637e-01 -3.95584434e-01 9.15151715e-01 9.89909530e-01
4.88775045e-01 -1.38825929e+00 -5.89611590e-01 3.40767860e-01
-5.34952462e-01 -9.75342095e-01 -4.42697287e-01 2.73200989e-01
-8.33920658e-01 -8.15813601e-01 -4.16290969e-01 -7.79346347e-01
8.27790856e-01 -3.47271442e-01 1.17618418e+00 1.95138916e-01
3.41908485e-01 1.89431608e-02 -5.34051538e-01 -4.32714790e-01
-8.90993178e-01 3.56787920e-01 -4.04643118e-01 -7.15993762e-01
-2.49075405e-02 -3.24352473e-01 9.16262642e-02 -1.92620069e-01
-8.22104096e-01 6.14375472e-01 4.49667394e-01 7.71529257e-01
3.08877200e-01 -7.93559477e-03 4.98843670e-01 -1.36990035e+00
1.01046252e+00 -1.04155205e-01 -4.94314253e-01 2.44115338e-01
-7.71862149e-01 4.58910614e-01 7.49132156e-01 -3.38425077e-02
-1.23360634e+00 2.28520617e-01 -4.89196926e-01 3.37911755e-01
-2.39765555e-01 6.61907554e-01 -1.88691050e-01 2.05475107e-01
8.11125457e-01 6.82649314e-02 -2.05547690e-01 -1.39331296e-01
8.29294443e-01 5.08943379e-01 7.33160138e-01 -8.39623809e-01
1.02813542e+00 -1.40704930e-01 -1.11509174e-01 -2.41573036e-01
-7.37521172e-01 2.74796635e-02 -6.95243835e-01 1.05317786e-01
8.03805113e-01 -7.95320153e-01 -5.38335741e-01 2.12505341e-01
-1.51226270e+00 -9.11342919e-01 -6.01659715e-01 1.91027313e-01
-7.85128236e-01 3.94929707e-01 -1.08436382e+00 -7.26743162e-01
-5.28891027e-01 -8.70263457e-01 1.07738757e+00 2.93369323e-01
-8.59002173e-01 -1.01623809e+00 3.16153206e-02 2.33433336e-01
3.75780523e-01 2.91502088e-01 1.32870519e+00 -9.36252058e-01
-6.98015094e-01 -1.84368983e-01 -2.16359898e-01 -1.38416132e-02
-1.63348138e-01 8.52982029e-02 -6.14675105e-01 2.07662776e-01
-3.81349325e-01 -5.06765187e-01 3.92794460e-01 -3.10318568e-03
6.24796689e-01 -9.48261857e-01 -3.91295761e-01 2.64395058e-01
9.17723358e-01 2.20223427e-01 8.54181230e-01 3.06735098e-01
5.78386188e-01 7.28162348e-01 4.64706779e-01 7.96332508e-02
9.31143522e-01 6.61256909e-01 2.29707047e-01 -5.48906215e-02
-2.35952497e-01 -8.22750807e-01 1.85040489e-01 7.15319395e-01
-2.87227958e-01 -4.69499767e-01 -7.75980949e-01 8.32725048e-01
-2.03495693e+00 -1.20744026e+00 9.63221863e-02 1.80807602e+00
1.26273417e+00 2.74528742e-01 2.33946919e-01 1.29527465e-01
3.82034391e-01 1.10226996e-01 -1.31302178e-01 -5.30776858e-01
1.56789109e-01 4.19300973e-01 2.56703585e-01 9.47395802e-01
-7.71750033e-01 1.50519109e+00 7.29836369e+00 3.88734221e-01
-5.84601939e-01 -2.57043451e-01 3.77020746e-01 2.21553236e-01
-7.19202459e-01 2.84339994e-01 -7.71824300e-01 -6.53335527e-02
8.77155542e-01 -3.89361441e-01 5.79004943e-01 6.60457790e-01
-1.43568814e-02 -9.82053652e-02 -1.38267994e+00 1.74425900e-01
9.51082781e-02 -1.58221507e+00 2.84575999e-01 -1.29937276e-01
5.79848528e-01 -2.53104240e-01 -5.60199380e-01 4.66232568e-01
9.00633752e-01 -1.09474015e+00 9.36055899e-01 3.87161434e-01
6.76329315e-01 -7.42787302e-01 6.77547991e-01 5.81403017e-01
-1.01420259e+00 -1.04746900e-01 -5.16102463e-02 -1.95539758e-01
6.75637007e-01 2.50158221e-01 -1.28909659e+00 7.57886529e-01
-1.72240436e-01 4.37321693e-01 -4.59081799e-01 7.93665767e-01
-8.93278778e-01 3.62217337e-01 -8.24770257e-02 -2.83804387e-01
1.10466860e-01 1.98449016e-01 5.08228302e-01 1.17256320e+00
1.65542573e-01 3.64715368e-01 2.50832587e-01 1.01789141e+00
-1.82061359e-01 -7.66156539e-02 -5.47240376e-01 -1.71500787e-01
7.04868257e-01 1.08764446e+00 -6.09712422e-01 -5.71482182e-01
-1.01559870e-01 9.83008921e-01 4.20743674e-01 3.24690044e-01
-4.41244125e-01 -7.20108747e-01 -9.13383886e-02 1.37482643e-01
2.57245660e-01 -2.18934327e-01 -3.65913808e-01 -1.05652642e+00
-1.88761100e-01 -1.03657317e+00 3.03662688e-01 -1.18513227e+00
-8.28882158e-01 9.67839539e-01 3.34663987e-01 -5.61604679e-01
-1.14213324e+00 -2.90385216e-01 -4.92040604e-01 1.14243507e+00
-1.22051573e+00 -1.46254170e+00 -1.30907223e-02 2.84838945e-01
5.23112118e-01 4.17168625e-02 1.23811960e+00 -3.01051531e-02
-3.68523449e-01 4.33079094e-01 -8.25726748e-01 5.48334837e-01
2.72970289e-01 -1.37688565e+00 1.03925848e+00 1.09944379e+00
6.82027712e-02 1.14976645e+00 6.17913485e-01 -9.26484406e-01
-1.03388441e+00 -1.20170105e+00 1.69839537e+00 -6.75410032e-01
5.72971582e-01 -2.01943323e-01 -6.38565898e-01 1.27840912e+00
4.27005529e-01 -6.04385912e-01 4.57731336e-01 2.99650609e-01
-3.00872654e-01 3.51451337e-01 -1.03312552e+00 9.34551775e-01
1.00203347e+00 -5.79815507e-01 -9.72703993e-01 4.21349704e-01
8.69239748e-01 -9.91121650e-01 -7.17481077e-01 1.65626630e-01
5.08700848e-01 -5.70978105e-01 4.19027686e-01 -7.25904763e-01
6.57550931e-01 -3.15514058e-01 6.65489733e-02 -1.26086736e+00
-4.92478490e-01 -1.07315385e+00 -1.49754405e-01 1.25809181e+00
9.16302085e-01 -4.66505200e-01 5.09559631e-01 1.06455052e+00
-5.37641764e-01 -8.18574250e-01 -5.12934148e-01 -3.68700147e-01
1.19124092e-01 -4.12119925e-01 9.52302396e-01 7.17562139e-01
6.64664567e-01 7.42773533e-01 -1.91402406e-01 -3.49853694e-01
-6.02504872e-02 4.33044322e-02 6.93798244e-01 -1.03641450e+00
-6.82799816e-01 -4.53007847e-01 2.13354126e-01 -1.24265754e+00
1.04170442e-01 -1.02696490e+00 4.93239105e-01 -2.04229474e+00
1.42870113e-01 -4.04433399e-01 3.30514967e-01 1.07003021e+00
-1.94221184e-01 -6.23749606e-02 2.68327236e-01 1.76580742e-01
-5.04488230e-01 1.99874178e-01 1.20430434e+00 2.64176279e-01
-3.36155623e-01 3.67031880e-02 -1.34054744e+00 5.21129310e-01
5.06996274e-01 -3.95380676e-01 -4.53915834e-01 -5.56953847e-01
3.90556931e-01 3.94431472e-01 -3.19345295e-02 -7.52203941e-01
1.79671004e-01 -2.80092090e-01 1.78158194e-01 -3.08921754e-01
1.68844268e-01 -1.02049604e-01 1.35169253e-01 2.49508023e-01
-8.84743929e-01 3.22961777e-01 2.04544291e-01 6.35713190e-02
6.78961873e-02 -4.73618180e-01 4.68387127e-01 -4.31184292e-01
-4.17382151e-01 -4.75819707e-02 -5.50239503e-01 5.80631420e-02
6.81829691e-01 -7.41909742e-02 -6.17058635e-01 -7.63217926e-01
-4.60516572e-01 1.64288625e-01 5.65708876e-01 2.09519535e-01
2.16341078e-01 -1.16038573e+00 -6.18360519e-01 -8.25323388e-02
3.35779786e-02 3.50519061e-01 -5.45354605e-01 3.71103138e-01
-4.87502545e-01 7.37668931e-01 1.17890254e-01 2.17241630e-01
-9.25466776e-01 2.19606847e-01 3.16181719e-01 -7.02314198e-01
-6.92891598e-01 8.22372317e-01 -2.87736237e-01 -5.06860554e-01
2.38871291e-01 -4.77480054e-01 -2.87374765e-01 -1.52273893e-01
6.27351165e-01 1.86786681e-01 -9.27111804e-02 -6.42314672e-01
-1.93952858e-01 1.49562910e-01 -1.87004447e-01 -6.83079362e-01
1.22150457e+00 1.14561498e-01 5.94737707e-03 1.97399393e-01
3.22183251e-01 2.40713313e-01 -9.19817507e-01 -5.14276773e-02
1.69149145e-01 1.13960721e-01 -2.55876482e-01 -1.38021874e+00
-4.52083975e-01 5.27624428e-01 -4.60010648e-01 2.16297850e-01
9.09959614e-01 -2.91444119e-02 1.10160196e+00 5.75958729e-01
3.74457330e-01 -1.04370344e+00 -1.09858789e-01 9.41928327e-01
9.30219114e-01 -5.67146003e-01 8.16772357e-02 -4.37352866e-01
-8.83267164e-01 1.22842920e+00 5.40658534e-01 3.23553056e-01
-1.95441455e-01 4.18173701e-01 3.30576748e-02 -2.49453746e-02
-9.98140395e-01 -2.95594841e-01 1.60177201e-01 8.32716525e-01
6.47032142e-01 -1.44987121e-01 -4.09979463e-01 7.25056827e-01
-5.99879265e-01 2.42703721e-01 7.41711617e-01 1.03378999e+00
-4.14854884e-01 -1.53385103e+00 2.53064018e-02 3.89840394e-01
-2.23476812e-01 -4.60257083e-01 -6.71330631e-01 9.15467322e-01
-1.14503294e-01 1.12892640e+00 -1.69242516e-01 -4.07350898e-01
5.25072515e-01 5.06561697e-01 6.18309140e-01 -1.20463657e+00
-8.72618556e-01 -8.57739598e-02 9.82192159e-01 -6.37777150e-01
-8.61089900e-02 -6.57302558e-01 -1.67719591e+00 -2.79562920e-01
-3.94820750e-01 3.20783705e-01 3.62422794e-01 1.18692708e+00
5.22400200e-01 5.63707054e-01 1.49307355e-01 -7.82787085e-01
-5.81424296e-01 -1.09655631e+00 -1.13416560e-01 1.63860619e-01
3.11692618e-02 -2.64523715e-01 4.85978611e-02 2.20267117e-01] | [11.458874702453613, 8.988122940063477] |
1618f024-7216-4bbd-9ce7-4bfb71ddd81a | rms-net-regression-and-masking-for-soccer | 2102.07624 | null | https://arxiv.org/abs/2102.07624v1 | https://arxiv.org/pdf/2102.07624v1.pdf | RMS-Net: Regression and Masking for Soccer Event Spotting | The recently proposed action spotting task consists in finding the exact timestamp in which an event occurs. This task fits particularly well for soccer videos, where events correspond to salient actions strictly defined by soccer rules (a goal occurs when the ball crosses the goal line). In this paper, we devise a lightweight and modular network for action spotting, which can simultaneously predict the event label and its temporal offset using the same underlying features. We enrich our model with two training strategies: the first one for data balancing and uniform sampling, the second for masking ambiguous frames and keeping the most discriminative visual cues. When tested on the SoccerNet dataset and using standard features, our full proposal exceeds the current state of the art by 3 Average-mAP points. Additionally, it reaches a gain of more than 10 Average-mAP points on the test set when fine-tuned in combination with a strong 2D backbone. | ['Rita Cucchiara', 'Simone Bronzin', 'Simone Calderara', 'Lorenzo Baraldi', 'Matteo Tomei'] | 2021-02-15 | null | null | null | null | ['action-spotting'] | ['computer-vision'] | [ 3.03762436e-01 -8.91715139e-02 -3.34483862e-01 -6.79411665e-02
-5.69321513e-01 -4.25512105e-01 6.30995870e-01 3.75274211e-01
-8.36828470e-01 6.61094666e-01 2.12286860e-01 2.77513593e-01
-1.79701239e-01 -4.22006696e-01 -6.19172454e-01 -6.29064977e-01
-2.82049090e-01 3.98870558e-01 8.45502913e-01 -1.60258606e-01
3.49031031e-01 5.32428026e-01 -1.77564168e+00 5.47881901e-01
2.97184139e-01 1.28941607e+00 1.54107213e-01 4.94793385e-01
3.42768043e-01 1.09019136e+00 -6.16643071e-01 -1.81137919e-01
3.49718511e-01 -3.82494092e-01 -5.72310805e-01 1.11333154e-01
5.06487668e-01 -2.27733493e-01 -3.89634460e-01 5.72915614e-01
4.79129881e-01 4.23541844e-01 2.19416291e-01 -1.30710757e+00
2.33971715e-01 5.41492701e-01 -5.34630418e-01 5.35704970e-01
5.67962289e-01 2.06444144e-01 9.40566421e-01 -8.33780527e-01
9.86008048e-01 8.79817545e-01 6.74446106e-01 2.71207362e-01
-1.21452832e+00 -4.72143054e-01 1.81097195e-01 7.00677097e-01
-1.28555608e+00 -4.61952031e-01 6.53671443e-01 -4.31934804e-01
9.18728471e-01 1.15629226e-01 8.91242564e-01 1.29627705e+00
1.52658880e-01 8.76030982e-01 9.70947862e-01 -3.25932801e-01
4.57825452e-01 -3.01238358e-01 -1.39653042e-01 2.16629773e-01
2.71305200e-02 1.52675629e-01 -9.82950687e-01 1.28788084e-01
7.37918019e-01 2.51683453e-03 -3.32574360e-02 -5.98059297e-01
-1.46550572e+00 5.37424028e-01 1.46150917e-01 4.73673075e-01
-6.00625932e-01 2.73908377e-01 6.59708798e-01 1.55242890e-01
2.18619078e-01 4.30230230e-01 -3.77026588e-01 -4.87246931e-01
-1.37330961e+00 5.51043451e-01 4.53790873e-01 5.21595836e-01
5.54184675e-01 -1.00792073e-01 -6.27828419e-01 3.44273567e-01
-2.52140820e-01 7.94825852e-02 2.97335923e-01 -1.02192318e+00
3.45867485e-01 6.32675588e-01 4.16201144e-01 -9.37902689e-01
-7.07820415e-01 -4.71019298e-01 -3.53357881e-01 4.21609968e-01
8.68891537e-01 -3.87864225e-02 -7.60105133e-01 1.64340818e+00
3.71683896e-01 5.12903571e-01 -4.44613457e-01 1.15188789e+00
4.23692524e-01 2.41437286e-01 1.08045414e-01 -9.65078399e-02
1.47818506e+00 -7.31303155e-01 -5.33996224e-01 -3.84839058e-01
3.55945498e-01 -4.96582329e-01 8.32098484e-01 6.72389090e-01
-9.36711371e-01 -6.40319526e-01 -1.08780849e+00 -1.60789751e-02
-2.22214967e-01 3.50190878e-01 5.61901510e-01 3.93448412e-01
-8.50849569e-01 8.24148417e-01 -1.03218031e+00 -3.86579007e-01
2.18423948e-01 1.95625529e-01 -6.56522334e-01 2.18386129e-01
-9.89371419e-01 9.29487884e-01 7.69408047e-01 -2.39044894e-02
-1.00419581e+00 -5.37263930e-01 -8.44412923e-01 -1.77538805e-02
7.77705491e-01 -2.52802044e-01 1.02365386e+00 -9.65192020e-01
-1.46129394e+00 1.00938129e+00 5.10685854e-02 -1.01388097e+00
1.02303159e+00 -3.66440117e-01 -2.43556827e-01 5.10077357e-01
1.72372147e-01 7.45648563e-01 7.85622656e-01 -7.62845516e-01
-9.24185634e-01 -3.21217716e-01 2.59100735e-01 6.30990118e-02
-1.16024472e-01 1.63779691e-01 -6.60706580e-01 -6.76646352e-01
6.69978485e-02 -7.63074636e-01 -2.88393199e-01 8.26385096e-02
-1.87941924e-01 -2.08045676e-01 5.59752643e-01 -6.14301682e-01
1.23162925e+00 -2.24327826e+00 2.12658018e-01 1.42451376e-01
8.87915865e-03 8.32174569e-02 1.21006407e-01 4.10887182e-01
-2.12596416e-01 -5.58178723e-01 1.55363709e-01 -2.49169096e-01
1.45980462e-01 -1.61311273e-02 -3.05281758e-01 7.01323450e-01
2.06322774e-01 7.87490368e-01 -9.55804884e-01 -5.55682778e-01
5.39151013e-01 3.29729080e-01 -6.12533391e-01 6.25039265e-02
-2.45148301e-01 5.12857795e-01 -8.56293738e-02 4.37997252e-01
3.98051798e-01 -1.18943237e-01 4.02764231e-01 -1.27533928e-01
-2.89954215e-01 3.66582334e-01 -1.70433211e+00 2.08956838e+00
9.42239910e-03 6.37809157e-01 -8.55436027e-02 -9.58313763e-01
8.39114726e-01 3.10711563e-01 8.41849923e-01 -7.47094035e-01
1.26850590e-01 7.67512918e-02 -2.08783939e-01 -4.29087937e-01
6.83102548e-01 1.57999590e-01 -2.40567297e-01 3.05405051e-01
2.89700806e-01 4.20908928e-01 7.97714293e-01 -2.62319930e-02
1.31722355e+00 6.88723087e-01 5.11542141e-01 -2.08206922e-01
3.48013908e-01 -9.92991179e-02 6.22356355e-01 7.52587855e-01
-3.09602946e-01 7.27181137e-01 8.68938684e-01 -6.89984977e-01
-8.69895935e-01 -9.26401734e-01 8.47703740e-02 1.28376544e+00
3.84676695e-01 -5.07307827e-01 -7.17922688e-01 -7.85363078e-01
-7.55993873e-02 6.03392661e-01 -8.60758066e-01 -1.15561984e-01
-7.97087133e-01 -3.75868082e-01 4.65079188e-01 6.12997830e-01
3.81123155e-01 -1.22199500e+00 -1.29695165e+00 3.50038618e-01
-3.82650763e-01 -1.35358334e+00 -3.37396234e-01 4.20989424e-01
-5.50505817e-01 -1.18780017e+00 -5.55321336e-01 -3.40070337e-01
1.48309276e-01 7.12362155e-02 1.11286700e+00 -2.18026415e-01
-3.33425730e-01 3.51922624e-02 -4.94476885e-01 -6.59821853e-02
1.00987842e-02 7.93671533e-02 -2.61536967e-02 2.85685748e-01
4.57946658e-01 -5.09113550e-01 -7.57721305e-01 5.09590030e-01
-8.67047668e-01 7.17078075e-02 4.39140826e-01 5.46040773e-01
6.24355912e-01 -2.94382751e-01 3.15922886e-01 -1.16664805e-01
1.25268230e-03 -2.93943405e-01 -5.00134766e-01 6.66431263e-02
-1.47969842e-01 -6.07067309e-02 3.96272659e-01 -4.13928449e-01
-6.60936415e-01 3.63692582e-01 -6.24532811e-03 -4.85760570e-01
-4.42413449e-01 7.20668063e-02 1.15456961e-01 1.65097099e-02
7.23509848e-01 1.70682952e-01 -3.38746637e-01 -5.20742238e-01
1.37009516e-01 7.29990005e-02 6.91667318e-01 -3.69646043e-01
3.79843265e-01 7.66353190e-01 8.07541460e-02 -5.15158057e-01
-8.45391452e-01 -6.01711035e-01 -8.80191803e-01 -6.22822702e-01
9.01594222e-01 -8.67770493e-01 -8.85843217e-01 3.24839026e-01
-9.13951635e-01 -2.28884161e-01 -5.88956058e-01 7.79973924e-01
-8.95086765e-01 3.69994611e-01 -4.25761878e-01 -5.93775570e-01
1.40703872e-01 -9.41848695e-01 1.18649077e+00 3.81048396e-02
-3.69634062e-01 -4.63025630e-01 3.77486855e-01 9.31988508e-02
-1.43853761e-02 7.20202386e-01 2.85260946e-01 -7.72403717e-01
-4.54638869e-01 -2.56830424e-01 -4.16715965e-02 3.07778828e-03
-1.82400316e-01 -2.81383485e-01 -7.86286056e-01 -3.00755680e-01
-2.71423250e-01 -2.85206825e-01 1.12216604e+00 4.38880265e-01
9.68556881e-01 3.17386925e-01 -2.85940111e-01 4.52071697e-01
1.19999027e+00 -2.02299301e-02 5.29244125e-01 5.32789052e-01
3.06321472e-01 4.25745130e-01 8.76761615e-01 6.87150598e-01
7.97707066e-02 1.17386317e+00 7.47180998e-01 -5.88368550e-02
-1.54914409e-01 -2.57575989e-01 4.37069029e-01 -9.22770053e-02
-4.29088414e-01 -1.29115373e-01 -6.79848969e-01 6.82194531e-01
-2.18178391e+00 -1.32980835e+00 1.87932983e-01 2.42556787e+00
5.60889423e-01 5.91710150e-01 5.41275501e-01 4.05250460e-01
6.52090132e-01 4.90815878e-01 -3.64834279e-01 -7.98280165e-02
-1.49101689e-01 2.00207502e-01 5.42777956e-01 1.74805969e-01
-1.49738514e+00 7.53619969e-01 6.22064877e+00 8.87711823e-01
-1.18258405e+00 5.75874560e-02 3.37384731e-01 -6.21017635e-01
3.18276405e-01 -8.02600831e-02 -8.60641956e-01 4.82141316e-01
5.85204542e-01 1.24280743e-01 1.03506610e-01 7.48987198e-01
4.74979311e-01 -4.47110802e-01 -1.22764337e+00 9.92282510e-01
2.24330917e-01 -1.37132537e+00 -3.09308439e-01 -1.23919904e-01
3.97075474e-01 -1.01715773e-01 -2.85691231e-01 2.79722661e-01
3.59840319e-02 -8.61744404e-01 1.18431568e+00 4.69128430e-01
5.76232135e-01 -7.24758983e-01 4.14592206e-01 1.74636275e-01
-1.20599103e+00 -1.40675962e-01 3.31397839e-02 -2.92792439e-01
4.74925488e-01 3.52342278e-01 -7.03759313e-01 6.89756513e-01
6.98823452e-01 9.58361149e-01 -6.30509019e-01 1.45495355e+00
-3.39422405e-01 3.51602048e-01 -4.67943013e-01 2.06797644e-01
4.45045918e-01 6.60982206e-02 7.01741278e-01 1.31988728e+00
3.16379815e-01 -3.31768066e-01 4.52703804e-01 4.48955923e-01
4.15497780e-01 3.92842665e-02 -3.01996976e-01 3.78384054e-01
2.29247026e-02 1.28763509e+00 -1.15524256e+00 -2.48551458e-01
-1.24446392e-01 1.03776944e+00 2.81468362e-01 1.27263561e-01
-1.12588549e+00 -2.55732149e-01 6.06970012e-01 3.24057162e-01
8.14689398e-01 -2.66954601e-01 -2.90100463e-03 -1.19168329e+00
1.85767263e-01 -7.16889739e-01 6.06839240e-01 -7.15347648e-01
-6.32539392e-01 5.48086166e-01 1.58269957e-01 -1.52094781e+00
-5.60130298e-01 -6.60372376e-01 -3.36250037e-01 3.14321101e-01
-1.14530683e+00 -9.92510915e-01 -3.14574629e-01 6.06152236e-01
5.28091252e-01 1.73013300e-01 4.85621244e-01 5.06056607e-01
-4.20341611e-01 3.18654925e-01 -2.77467430e-01 1.13861106e-01
7.85575211e-01 -1.06679678e+00 2.42454916e-01 1.03974450e+00
4.08574402e-01 -2.29830272e-03 1.17655337e+00 -4.59427297e-01
-9.24576283e-01 -6.65809929e-01 8.96401763e-01 -3.71481866e-01
7.31185675e-01 -5.34518003e-01 -6.28804862e-01 7.13504195e-01
7.63806179e-02 2.53841039e-02 2.18776926e-01 -1.90421566e-01
-2.16503307e-01 -1.85648695e-01 -8.47875655e-01 4.89828676e-01
1.10630953e+00 -1.86268479e-01 -6.57878041e-01 2.33628646e-01
2.48708054e-01 -6.98030710e-01 -5.53176939e-01 3.46317112e-01
6.74673915e-01 -1.43770802e+00 1.00885344e+00 -6.05386794e-01
4.58299637e-01 -4.99806106e-01 2.76732109e-02 -9.52290058e-01
-3.01888019e-01 -6.31589890e-01 -4.93141450e-02 7.40262210e-01
1.53333936e-02 2.44853296e-03 9.10430312e-01 1.10190123e-01
2.84086931e-02 -5.74255168e-01 -1.38726544e+00 -7.26570487e-01
-6.31110907e-01 -5.54160237e-01 3.72385442e-01 6.51566982e-01
7.32216239e-02 -6.95802867e-02 -6.52465463e-01 -1.84393719e-01
4.90308106e-01 1.77876785e-01 7.19317555e-01 -1.09222496e+00
-4.14834470e-01 -4.60173577e-01 -8.80391300e-01 -1.06098461e+00
-5.40335476e-02 -6.06351912e-01 5.77935204e-02 -1.09779036e+00
8.23555067e-02 1.87556222e-01 -6.74810171e-01 5.83405256e-01
4.38776501e-02 5.79052091e-01 3.80462199e-01 -1.04957789e-01
-1.05120754e+00 2.09253475e-01 7.84731388e-01 2.35408396e-01
-1.72331318e-01 8.42004269e-03 -9.96726602e-02 7.66080201e-01
5.34449220e-01 -4.48780149e-01 -2.22962946e-02 -4.88300882e-02
3.35265666e-01 2.23640218e-01 7.47789621e-01 -1.38164890e+00
2.62031257e-01 -1.88657761e-01 4.73497212e-01 -7.16338754e-01
5.29074073e-01 -8.66108298e-01 2.21651167e-01 5.00679016e-01
-4.61257279e-01 -5.96652739e-02 2.06031471e-01 5.36583543e-01
-3.53300840e-01 -7.48999566e-02 7.18134105e-01 -4.87773195e-02
-1.27402925e+00 1.09115548e-01 -2.96089888e-01 -4.05780710e-02
1.42692935e+00 -4.86568362e-01 -1.28955403e-02 -2.62181580e-01
-1.09614134e+00 -1.77194662e-02 3.99216533e-01 6.12644196e-01
3.68443102e-01 -1.36263156e+00 -7.00613379e-01 1.90885231e-01
2.06855386e-01 -5.35846472e-01 3.48933548e-01 1.23288119e+00
-3.78828049e-01 3.23909253e-01 -6.04347229e-01 -7.17591584e-01
-1.23228240e+00 4.26314384e-01 3.14586073e-01 -5.49132168e-01
-9.21661377e-01 5.39928257e-01 3.86965764e-03 1.03077568e-01
3.02847475e-01 -3.38320285e-01 -4.47295219e-01 4.32350457e-01
7.58700669e-01 5.81611812e-01 1.93595052e-01 -6.90739453e-01
-6.34308100e-01 3.74520808e-01 2.47071639e-01 -2.26320282e-01
1.28953564e+00 1.54327586e-01 1.12906665e-01 4.60333198e-01
6.54615104e-01 -1.82812363e-02 -1.82767689e+00 -1.77756891e-01
2.24260107e-01 -4.83581364e-01 -1.36011124e-01 -7.34299183e-01
-9.19796765e-01 5.78606248e-01 5.47224820e-01 1.45399585e-01
1.04379010e+00 -1.07641742e-01 4.42276746e-01 9.30173695e-02
5.51707447e-01 -1.12894392e+00 1.72800452e-01 4.41910774e-01
7.59522498e-01 -9.62555408e-01 4.28373441e-02 -5.24469033e-05
-7.15435028e-01 1.04594076e+00 4.97457623e-01 -5.20894110e-01
2.64373541e-01 2.96718329e-01 -1.42522156e-01 -1.83474958e-01
-7.58887768e-01 -4.07641530e-01 3.33717763e-01 3.09681177e-01
4.07765284e-02 -8.68948773e-02 -5.50357163e-01 4.40203667e-01
9.87706333e-03 3.41801822e-01 3.39625955e-01 9.60006535e-01
-5.92598021e-01 -7.23759592e-01 -3.25089365e-01 1.25412762e-01
-6.04645193e-01 2.15228081e-01 -2.14655295e-01 9.40369666e-01
4.45884407e-01 7.76130080e-01 3.46533597e-01 -2.94331342e-01
5.87615311e-01 8.12961906e-02 4.90425348e-01 -3.22360337e-01
-7.58422434e-01 1.43671617e-01 2.17317998e-01 -1.18896496e+00
-7.09373951e-01 -8.63757670e-01 -1.07665598e+00 5.72018931e-03
-4.90592560e-03 -4.04862948e-02 3.33054483e-01 1.04348934e+00
3.20164561e-01 5.16345382e-01 3.01733196e-01 -1.20512736e+00
-4.07115072e-01 -9.23701763e-01 -6.41831100e-01 5.22629261e-01
1.65851802e-01 -9.57760870e-01 -2.22189739e-01 1.21462829e-01] | [8.125190734863281, 0.2779960632324219] |
92871e90-c3e2-42e1-9a51-ebbd63414e7f | modelling-fixated-discourse-in-chats-with | null | null | https://aclanthology.org/W12-0413 | https://aclanthology.org/W12-0413.pdf | Modelling Fixated Discourse in Chats with Cyberpedophiles | null | ['Thamar Solorio', 'Paolo Rosso', 'Dasha Bogdanova'] | 2012-04-01 | null | null | null | ws-2012-4 | ['deception-detection'] | ['miscellaneous'] | [-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] | [-7.355007648468018, 3.684802532196045] |
43b06b31-df0f-4c61-947e-041bbce2422d | vita-visual-linguistic-translation-by | 2106.00250 | null | https://arxiv.org/abs/2106.00250v3 | https://arxiv.org/pdf/2106.00250v3.pdf | ViTA: Visual-Linguistic Translation by Aligning Object Tags | Multimodal Machine Translation (MMT) enriches the source text with visual information for translation. It has gained popularity in recent years, and several pipelines have been proposed in the same direction. Yet, the task lacks quality datasets to illustrate the contribution of visual modality in the translation systems. In this paper, we propose our system under the team name Volta for the Multimodal Translation Task of WAT 2021 from English to Hindi. We also participate in the textual-only subtask of the same language pair for which we use mBART, a pretrained multilingual sequence-to-sequence model. For multimodal translation, we propose to enhance the textual input by bringing the visual information to a textual domain by extracting object tags from the image. We also explore the robustness of our system by systematically degrading the source text. Finally, we achieve a BLEU score of 44.6 and 51.6 on the test set and challenge set of the multimodal task. | ['Radhika Mamidi', 'Devansh Gautam', 'Kshitij Gupta'] | 2021-06-01 | null | https://aclanthology.org/2021.wat-1.19/ | https://aclanthology.org/2021.wat-1.19.pdf | workshop-on-asian-translation-2021-8 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 3.73361737e-01 -1.73669711e-01 6.99975193e-02 -2.76321918e-01
-1.31899047e+00 -9.41353381e-01 1.05238807e+00 -4.64718997e-01
-4.52169627e-01 6.69440925e-01 2.73968339e-01 -4.06225890e-01
6.60391390e-01 -1.41655028e-01 -7.91476786e-01 -5.67621946e-01
6.16316140e-01 6.31331146e-01 -1.62365630e-01 -2.67633796e-01
-1.14507049e-01 7.05134645e-02 -1.01220512e+00 1.02052176e+00
8.84344518e-01 6.91750467e-01 3.46836030e-01 4.89732563e-01
-1.45045444e-01 3.64081472e-01 -4.39444900e-01 -1.01111746e+00
3.68153691e-01 -7.38244355e-01 -7.14182496e-01 1.38177097e-01
6.05435789e-01 -2.31440172e-01 -3.49029094e-01 9.70740736e-01
8.06110978e-01 -3.49876761e-01 6.59443021e-01 -1.42436492e+00
-9.12505150e-01 6.73927426e-01 -6.57553256e-01 -2.55203098e-01
4.86297458e-01 3.03477049e-01 8.37713003e-01 -1.56082869e+00
1.17075562e+00 1.38797784e+00 1.88069254e-01 5.56421399e-01
-1.17535853e+00 -4.24867630e-01 -2.34824389e-01 2.62849718e-01
-1.18233776e+00 -7.10390866e-01 5.51801980e-01 -3.44753832e-01
7.64497459e-01 2.97798008e-01 3.60041291e-01 1.76222944e+00
-1.99339930e-02 1.20400381e+00 1.34990954e+00 -6.28250301e-01
-2.28149250e-01 2.54167855e-01 -5.47063053e-01 5.92172801e-01
-3.38503599e-01 -4.09841426e-02 -7.57991970e-01 2.63257682e-01
4.21130508e-01 -3.03959340e-01 -1.79928362e-01 -4.31931108e-01
-1.92359793e+00 5.33909321e-01 2.03197584e-01 3.01492393e-01
-1.34572715e-01 1.03148110e-01 5.08990824e-01 4.52235311e-01
4.81347114e-01 2.54278779e-01 -2.33208373e-01 -3.01282346e-01
-8.11519325e-01 3.65965720e-03 3.92926127e-01 1.25099242e+00
4.67103064e-01 -1.77614301e-01 -6.60429239e-01 9.87084687e-01
4.57626730e-01 1.06611931e+00 2.17955217e-01 -7.78457880e-01
1.24269414e+00 4.60656673e-01 1.69112056e-01 -4.59815055e-01
-9.08260643e-02 -2.58669734e-01 -6.46215260e-01 -1.10194995e-03
5.14440477e-01 -2.29470283e-01 -1.01663411e+00 1.57357550e+00
1.70671448e-01 -4.46096987e-01 3.10861588e-01 1.24024761e+00
8.71808171e-01 8.42599034e-01 3.92065234e-02 -1.16698392e-01
1.39377606e+00 -1.36548138e+00 -9.28323388e-01 -2.24293321e-01
5.37579834e-01 -1.49992394e+00 1.34094739e+00 8.29497427e-02
-1.26996911e+00 -5.31538546e-01 -8.15757453e-01 -2.90204406e-01
-2.27481246e-01 5.97726703e-01 1.21939212e-01 4.56408918e-01
-1.11361802e+00 1.25323206e-01 -6.67454600e-01 -6.55269384e-01
1.10657334e-01 7.52420574e-02 -5.63997805e-01 -3.51744890e-01
-1.26796556e+00 1.20983458e+00 1.12892553e-01 1.36895195e-01
-9.31311727e-01 -2.84366667e-01 -8.22915494e-01 -2.13292331e-01
3.11643109e-02 -9.75801587e-01 1.18637466e+00 -1.23532534e+00
-1.51825738e+00 1.04051864e+00 -3.43149602e-01 -9.42664295e-02
1.01782000e+00 -1.91982999e-01 -2.67047137e-01 3.17113578e-01
1.56718381e-02 1.21463287e+00 9.31940258e-01 -1.42785835e+00
-5.81231654e-01 -3.22986096e-01 -1.17077813e-01 5.27015984e-01
-2.81744182e-01 3.62024575e-01 -9.52320814e-01 -8.20484161e-01
-1.81032091e-01 -1.19601631e+00 1.75838247e-01 -2.15555176e-01
-3.92688543e-01 1.92539215e-01 9.62899327e-01 -1.05079126e+00
7.99416900e-01 -2.38296485e+00 6.58743262e-01 -9.25737023e-02
-4.72692885e-02 -2.27484163e-02 -6.42556489e-01 6.94312632e-01
1.09737217e-01 5.42429611e-02 -4.14799511e-01 -8.47965240e-01
3.38316739e-01 6.37378767e-02 -3.22362512e-01 2.59980470e-01
2.11746052e-01 1.49815750e+00 -7.73235202e-01 -6.40846312e-01
1.21293974e-03 5.12574375e-01 6.10804893e-02 2.84900337e-01
-2.02638417e-01 6.68942153e-01 6.25875033e-03 7.64272332e-01
6.82865024e-01 -5.41792177e-02 1.92731485e-01 -3.55472058e-01
-5.31445928e-02 8.58230963e-02 -4.59414929e-01 2.18940091e+00
-2.86967158e-01 9.92887199e-01 1.07346825e-01 -4.60844755e-01
6.43774211e-01 5.50178468e-01 2.34128520e-01 -1.10447657e+00
9.47735757e-02 3.90143514e-01 6.25623949e-03 -6.72174871e-01
8.09129655e-01 -1.14399858e-01 -1.61713362e-01 4.96354610e-01
1.24885388e-01 -1.65052921e-01 2.84834832e-01 3.22837949e-01
7.31689632e-01 6.00463569e-01 -2.55390704e-01 2.12279841e-01
9.50851068e-02 2.64550686e-01 -2.18499340e-02 4.69806820e-01
-2.52360910e-01 9.48027909e-01 1.60041854e-01 -7.75494799e-02
-1.49830103e+00 -1.10529840e+00 1.47617176e-01 1.10350728e+00
1.91276982e-01 -2.55806535e-01 -8.03819239e-01 -8.83907557e-01
-3.92944127e-01 5.67376971e-01 -4.47743088e-01 4.14577546e-03
-6.50113821e-01 -5.75136900e-01 8.09341133e-01 3.79800498e-01
4.61960524e-01 -9.64934647e-01 -2.06323296e-01 -4.36367169e-02
-1.04813933e+00 -1.53862202e+00 -8.05371702e-01 -1.82266727e-01
-5.74896157e-01 -5.01394153e-01 -1.35918617e+00 -9.48287249e-01
4.55122769e-01 2.99735427e-01 1.12542593e+00 -3.73124272e-01
1.17598191e-01 5.32724202e-01 -5.09966373e-01 -1.23149939e-01
-6.85669839e-01 2.58121431e-01 -1.49017558e-01 1.86932236e-01
-1.12040397e-02 -1.61674991e-02 -3.98455203e-01 5.76086581e-01
-8.57049882e-01 6.80977941e-01 8.12759638e-01 6.87326372e-01
3.36077899e-01 -9.48184788e-01 3.40986311e-01 -2.50898987e-01
5.63631058e-01 -9.91934165e-02 -3.17630410e-01 4.85278040e-01
-1.43367484e-01 -6.83912560e-02 1.87104970e-01 -5.45620024e-01
-1.08651888e+00 1.87618911e-01 5.46709225e-02 -4.24155504e-01
-7.11784735e-02 5.34805238e-01 -4.57165569e-01 4.20757458e-02
4.40500528e-01 3.25114012e-01 -5.28419763e-02 -3.10021251e-01
7.33877182e-01 8.34293365e-01 5.61785102e-01 -5.51063240e-01
9.33576643e-01 3.08701247e-01 -1.30732507e-01 -4.35210854e-01
-2.51042187e-01 -1.77871928e-01 -6.05218589e-01 -3.66509169e-01
9.93002594e-01 -1.09045649e+00 -3.11294913e-01 4.03732151e-01
-1.44928932e+00 -4.11569297e-01 2.09891170e-01 5.37788570e-01
-6.54144228e-01 4.77688968e-01 -6.79865062e-01 -5.07004678e-01
-3.88733655e-01 -1.51145267e+00 1.42858481e+00 -1.91934019e-01
-8.34470093e-02 -6.33595467e-01 3.12733613e-02 7.58945227e-01
3.69983345e-01 5.45681529e-02 8.93170834e-01 -4.07288462e-01
-5.92872322e-01 1.46177128e-01 -5.78596234e-01 9.14227515e-02
-2.20451638e-01 -8.19090083e-02 -9.62366581e-01 -3.32309872e-01
-4.98001367e-01 -4.68198627e-01 9.42757487e-01 -1.19755808e-02
3.53563845e-01 -1.94982048e-02 -1.75727785e-01 3.00641596e-01
9.13033426e-01 5.53420633e-02 8.01117241e-01 4.17247951e-01
8.86521101e-01 8.26369584e-01 7.85866439e-01 3.94737124e-02
5.03806472e-01 1.06478679e+00 4.90084559e-01 -5.10175705e-01
-4.95041996e-01 -2.51357704e-01 8.38626862e-01 9.70159829e-01
-1.34401858e-01 -5.40308714e-01 -8.44064176e-01 5.55806875e-01
-1.98501670e+00 -8.09731305e-01 -2.91888803e-01 1.91162705e+00
8.42754602e-01 -4.03426647e-01 1.18178591e-01 -3.69173914e-01
7.71221519e-01 -1.01675794e-01 2.75073270e-03 -2.70188928e-01
-4.33496892e-01 -2.92276263e-01 2.02656165e-01 5.20375192e-01
-9.51760948e-01 1.21592569e+00 5.64060116e+00 8.27349424e-01
-1.10328650e+00 2.97469020e-01 5.29315770e-01 -9.27271247e-02
-4.65762198e-01 -5.92937507e-02 -4.01870489e-01 3.72937202e-01
8.51409197e-01 1.99666932e-01 4.83529985e-01 2.41806328e-01
2.69573957e-01 3.83007005e-02 -1.24841797e+00 1.13788080e+00
3.63993824e-01 -1.04260492e+00 3.11419815e-01 1.64594218e-01
8.73681128e-01 2.27576733e-01 4.16452348e-01 4.52710122e-01
-1.39292806e-01 -9.57086563e-01 1.09027481e+00 3.87553245e-01
1.18514693e+00 -5.52439570e-01 7.45622218e-01 1.66263521e-01
-9.57478285e-01 3.98171693e-01 -9.35553536e-02 3.98559898e-01
4.77248073e-01 8.80834758e-02 -1.03836536e+00 8.46137941e-01
4.86129194e-01 5.51239550e-01 -7.67870426e-01 9.05500174e-01
-3.67167890e-01 4.75399822e-01 -2.88221184e-02 1.09333254e-01
3.60974103e-01 -2.82797396e-01 6.52658284e-01 1.37770259e+00
6.27287865e-01 -5.88783741e-01 3.60779539e-02 6.30711377e-01
-5.12454271e-01 2.33673349e-01 -6.05951965e-01 -3.37131649e-01
8.03029761e-02 1.34238648e+00 -4.22816366e-01 -3.99311215e-01
-4.47028726e-01 1.61646461e+00 1.05527483e-01 7.87038207e-01
-1.00860810e+00 -1.52265221e-01 2.62506187e-01 -2.86189586e-01
1.23414762e-01 -2.16770709e-01 -3.13845932e-01 -1.40489423e+00
1.91488847e-01 -1.30495036e+00 4.79237288e-02 -1.34981441e+00
-1.04640734e+00 7.32655644e-01 -1.73144832e-01 -1.45756161e+00
-1.98223576e-01 -6.27457559e-01 -5.74100167e-02 1.07245016e+00
-1.26825285e+00 -1.75695181e+00 -1.14793926e-01 6.31841063e-01
6.86258137e-01 -2.58016288e-01 6.40570164e-01 5.22005379e-01
-1.76765278e-01 7.31048286e-01 1.40027851e-01 1.36483535e-01
1.54315090e+00 -9.53109026e-01 5.27802646e-01 8.36225033e-01
3.61702561e-01 4.88887101e-01 7.05584407e-01 -5.54715872e-01
-1.78219926e+00 -1.02883291e+00 1.33580339e+00 -8.10696125e-01
6.31201684e-01 -6.47833884e-01 -5.64382315e-01 6.49084389e-01
1.08042419e+00 -4.60755885e-01 4.31026906e-01 -3.36996496e-01
-4.28031713e-01 2.63858229e-01 -9.11397815e-01 9.04783547e-01
8.91594291e-01 -6.03354454e-01 -5.12705803e-01 2.71711320e-01
8.83469522e-01 -6.17940068e-01 -6.60092533e-01 3.74291897e-01
6.57195330e-01 -4.70762551e-01 8.44667256e-01 -5.28587580e-01
1.01567006e+00 -5.90534091e-01 -4.12502736e-01 -1.57208312e+00
9.42070968e-03 -5.95619559e-01 2.02826321e-01 1.29252493e+00
9.45765138e-01 -1.38656706e-01 3.33993554e-01 1.12148501e-01
-2.24450693e-01 -2.88584709e-01 -1.08908737e+00 -6.08789384e-01
5.57058267e-02 -2.56666720e-01 2.80278504e-01 1.07466972e+00
1.00892864e-01 7.29833424e-01 -8.04315686e-01 -2.59400476e-02
3.80591989e-01 2.64695615e-01 1.00661516e+00 -5.73026419e-01
-1.81969106e-01 -4.13509876e-01 1.26226112e-01 -1.02304113e+00
-1.34763410e-02 -1.16008949e+00 6.48447871e-02 -1.86170590e+00
7.91472495e-01 3.93632561e-01 1.07034281e-01 5.17174542e-01
-2.80106068e-01 7.42500246e-01 5.83376765e-01 4.18977648e-01
-8.37921798e-01 6.69875622e-01 1.58607793e+00 -5.32050669e-01
9.42027718e-02 -4.55565542e-01 -2.90108323e-01 2.09360152e-01
7.32045710e-01 -2.10233882e-01 -3.91950943e-02 -9.68343139e-01
3.62950295e-01 5.95070422e-02 4.14911151e-01 -2.56909728e-01
-1.63048163e-01 -8.16302598e-02 4.91794080e-01 -7.34999120e-01
4.98364568e-01 -9.56734538e-01 1.37369141e-01 1.92184880e-01
-3.74406606e-01 5.33753633e-01 3.28442633e-01 1.80467531e-01
-3.06942940e-01 8.48815590e-02 5.47529101e-01 1.39475495e-01
-4.82656538e-01 -2.89732628e-02 -4.22286659e-01 -1.48185715e-01
7.31612504e-01 9.47884172e-02 -7.15039134e-01 -4.66693670e-01
-6.56468213e-01 2.79793113e-01 7.17818975e-01 6.63503230e-01
5.79792976e-01 -1.69399583e+00 -1.10525644e+00 -2.45708004e-01
6.34098768e-01 -8.81901860e-01 9.44432467e-02 1.30157840e+00
-4.19056803e-01 3.19217145e-01 -3.69613796e-01 -7.84173667e-01
-1.51631200e+00 5.68017840e-01 -1.46987354e-02 -2.45049834e-01
-2.72281796e-01 4.01500672e-01 2.47735605e-01 -6.53683424e-01
4.14140783e-02 4.64788117e-02 6.94433600e-02 6.87861294e-02
4.00489122e-01 3.10429811e-01 1.37344718e-01 -9.52046692e-01
-3.97025704e-01 4.90765661e-01 1.18732691e-01 -9.50438797e-01
9.52058911e-01 -5.79263270e-01 -1.69858590e-01 3.18566054e-01
1.17849112e+00 2.24884227e-01 -9.67118680e-01 -2.14706421e-01
-9.97765884e-02 -3.02108973e-01 -4.23200965e-01 -1.40960729e+00
-7.61705399e-01 1.19947410e+00 7.20673144e-01 -3.30479413e-01
9.73167956e-01 6.34303093e-02 7.69053698e-01 3.42115730e-01
2.17353746e-01 -9.85896707e-01 6.49709553e-02 7.80131698e-01
1.23496521e+00 -1.53545308e+00 -4.32729065e-01 -2.12348029e-01
-1.04982817e+00 1.01119471e+00 3.86854351e-01 7.17462718e-01
-5.74685514e-01 1.96857035e-01 6.52400017e-01 2.90072262e-01
-6.89263046e-01 -3.38947266e-01 6.69021547e-01 6.42329454e-01
5.47444224e-01 1.17614679e-01 -4.47569221e-01 2.32432693e-01
-1.96112096e-02 -2.52211064e-01 2.32643992e-01 7.67192006e-01
-1.82248250e-01 -1.36556613e+00 -6.71468914e-01 -3.49971116e-01
-2.96073616e-01 -4.37206298e-01 -9.64681387e-01 7.17253506e-01
-1.47811890e-01 1.09519684e+00 -3.33416879e-01 -4.56352115e-01
3.01242918e-01 4.19650286e-01 7.77438164e-01 -1.93178505e-01
-6.76663101e-01 5.95398784e-01 3.19976300e-01 -4.30089116e-01
-4.54351485e-01 -6.10005617e-01 -8.92500937e-01 -2.03233436e-01
2.50989288e-01 5.29328175e-02 9.76012170e-01 8.59320045e-01
4.13218915e-01 3.65552187e-01 3.70187432e-01 -9.58349049e-01
-1.10018775e-01 -1.21549964e+00 6.19059987e-02 4.97535974e-01
1.75892130e-01 -3.05352747e-01 -9.83253941e-02 4.59936708e-01] | [11.47209358215332, 1.5214899778366089] |
3b9f74f8-04ff-4535-ac5f-656b88db235a | classification-and-understanding-of-cloud | 2009.12931 | null | https://arxiv.org/abs/2009.12931v4 | https://arxiv.org/pdf/2009.12931v4.pdf | Classification and understanding of cloud structures via satellite images with EfficientUNet | Climate change has been a common interest and the forefront of crucial political discussion and decision-making for many years. Shallow clouds play a significant role in understanding the Earth's climate, but they are challenging to interpret and represent in a climate model. By classifying these cloud structures, there is a better possibility of understanding the physical structures of the clouds, which would improve the climate model generation, resulting in a better prediction of climate change or forecasting weather update. Clouds organise in many forms, which makes it challenging to build traditional rule-based algorithms to separate cloud features. In this paper, classification of cloud organization patterns was performed using a new scaled-up version of Convolutional Neural Network (CNN) named as EfficientNet as the encoder and UNet as decoder where they worked as feature extractor and reconstructor of fine grained feature map and was used as a classifier, which will help experts to understand how clouds will shape the future climate. By using a segmentation model in a classification task, it was shown that with a good encoder alongside UNet, it is possible to obtain good performance from this dataset. Dice coefficient has been used for the final evaluation metric, which gave the score of 66.26\% and 66.02\% for public and private (test set) leaderboard on Kaggle competition respectively. | ['Noor Hossain Nuri Sabab', 'Tashin Ahmed'] | 2020-09-27 | null | null | null | null | ['satellite-image-classification'] | ['computer-vision'] | [-6.69986382e-02 -2.12229624e-01 3.08110237e-01 -5.12609482e-01
-1.53163001e-01 -9.31039453e-01 7.57504940e-01 3.85161728e-01
-3.05916786e-01 7.98245549e-01 4.89980541e-02 -6.41583323e-01
-1.35490426e-03 -1.09186566e+00 -7.40787625e-01 -8.39514256e-01
-1.06024541e-01 3.92512977e-01 1.54446617e-01 -2.26528227e-01
4.45410460e-01 5.20732999e-01 -1.52811205e+00 5.06816089e-01
9.24903572e-01 9.44331765e-01 4.51181084e-01 8.08980346e-01
-3.19099307e-01 4.89782602e-01 -6.19978666e-01 -2.03992963e-01
3.60659063e-01 -1.97735980e-01 -8.01696122e-01 -2.74824798e-01
4.70221102e-01 2.86384284e-01 1.61011204e-01 9.01340663e-01
3.31281722e-01 -8.18488598e-02 7.00532377e-01 -6.62512124e-01
-4.03690457e-01 2.50736058e-01 -2.85155654e-01 3.51768970e-01
-1.65545344e-01 1.78445682e-01 1.06575859e+00 -4.16913003e-01
5.20934463e-01 1.08431387e+00 4.85350549e-01 2.41954863e-01
-7.58174062e-01 -7.89633989e-01 6.76229820e-02 2.45789856e-01
-8.64562869e-01 -1.18528455e-01 3.67545515e-01 -7.62548566e-01
1.03760862e+00 5.51538289e-01 1.04852116e+00 4.36582595e-01
4.11172122e-01 4.10873085e-01 1.72600174e+00 -3.22414845e-01
2.71471530e-01 5.79094112e-01 4.83883871e-03 5.03137290e-01
3.71815115e-01 3.46663147e-01 -2.57164627e-01 2.30055645e-01
3.08710754e-01 -1.55831158e-01 -1.42022610e-01 1.29394278e-01
-8.54042947e-01 1.01633310e+00 1.14416409e+00 4.85365570e-01
-3.62364441e-01 2.61940241e-01 1.19406700e-01 3.97127599e-01
5.82411289e-01 6.34736061e-01 -5.32354116e-01 1.99340619e-02
-1.03302705e+00 5.34402788e-01 7.32871056e-01 2.16986567e-01
6.53574884e-01 3.40411924e-02 2.21530169e-01 2.97505736e-01
3.39825004e-01 8.52690756e-01 4.90837604e-01 -6.05011761e-01
2.76349485e-01 7.93450654e-01 9.47127566e-02 -1.18294144e+00
-3.16346467e-01 -9.60120261e-01 -1.04504251e+00 4.73069161e-01
2.28734806e-01 -2.50134885e-01 -1.17624247e+00 1.35438156e+00
3.22381824e-01 7.19171166e-02 8.00264180e-02 1.07747626e+00
6.14259779e-01 9.95762408e-01 -2.66070008e-01 2.24373683e-01
1.38768661e+00 -5.82858562e-01 -2.34856203e-01 -1.82106584e-01
4.27758962e-01 -7.58582234e-01 6.40302300e-01 4.18851256e-01
-4.00711924e-01 -5.59240878e-01 -1.04619312e+00 4.03824806e-01
-1.07470524e+00 -1.52037948e-01 9.78636324e-01 7.78119087e-01
-1.11430597e+00 6.52205467e-01 -5.66433370e-01 -2.88536757e-01
5.38795114e-01 2.61438280e-01 -2.98127592e-01 1.03422314e-01
-1.27644658e+00 1.19161880e+00 4.30799127e-01 4.75348502e-01
-9.21421230e-01 -4.08693045e-01 -2.73683846e-01 9.49302614e-02
-7.09672198e-02 -5.62244952e-01 8.29091251e-01 -1.03212571e+00
-1.15852261e+00 7.35864162e-01 -1.25163883e-01 -7.37745285e-01
3.47052097e-01 6.84139086e-03 -2.93768018e-01 -2.13485524e-01
1.82469990e-02 6.25592113e-01 5.70882499e-01 -1.32167947e+00
-9.87979054e-01 -3.96606535e-01 -4.75672297e-02 2.92641640e-01
1.33109987e-01 -2.34446600e-01 1.86528072e-01 -2.41294548e-01
5.02113141e-02 -1.20706058e+00 -2.82307118e-01 -3.71216506e-01
-2.45017663e-01 -4.35629897e-02 9.08419311e-01 -7.70603359e-01
8.61430168e-01 -1.82010758e+00 4.31321897e-02 4.59864557e-01
2.74992913e-01 5.03681302e-01 2.71282762e-01 3.60385507e-01
-1.30048543e-01 5.80461323e-01 -4.96451318e-01 7.11695552e-02
-1.87810838e-01 2.61179179e-01 -3.44655395e-01 3.11402798e-01
1.50628820e-01 6.48524761e-01 -6.37528777e-01 3.70231941e-02
1.87671721e-01 2.99038917e-01 -3.89654309e-01 2.69492149e-01
-5.08407354e-01 6.77591324e-01 -3.97697628e-01 3.73336911e-01
7.63232112e-01 2.81730071e-02 -1.48101849e-02 3.26995105e-01
-5.58339357e-01 1.89533889e-01 -9.97655034e-01 1.13861763e+00
-6.46260977e-01 1.05911887e+00 3.50085576e-03 -9.99890804e-01
9.95420992e-01 3.14595759e-01 -8.38622972e-02 -8.68400693e-01
2.02479940e-02 3.73555034e-01 4.57044750e-01 -4.33187991e-01
4.63102251e-01 -4.36699063e-01 -1.05053283e-01 2.43508503e-01
-5.02185166e-01 -6.92403495e-01 -9.41493213e-02 -4.56802882e-02
7.12595940e-01 -1.30893335e-01 -1.60515886e-02 -3.86410117e-01
4.62052166e-01 3.21995169e-01 3.26261669e-01 4.50645119e-01
-2.08297130e-02 5.40315092e-01 3.11896205e-01 -9.28632498e-01
-1.10138941e+00 -6.59207880e-01 -2.58867860e-01 5.12982488e-01
-1.89277813e-01 -9.33995878e-04 -5.98783433e-01 -2.68058985e-01
2.70366203e-02 6.81579828e-01 -7.94356763e-01 1.56166792e-01
-3.04235548e-01 -1.00456178e+00 3.56490225e-01 9.91068482e-02
7.97612607e-01 -1.11153376e+00 -8.84744763e-01 4.59313206e-02
-6.27077222e-02 -7.87815452e-01 2.37270519e-01 6.10648632e-01
-8.82173061e-01 -1.05701828e+00 -3.25965494e-01 -3.50932449e-01
3.67201596e-01 2.40631014e-01 1.11729228e+00 9.85740796e-02
-3.69646728e-01 -4.25603747e-01 -3.83313537e-01 -1.07940483e+00
-2.98169523e-01 3.26806784e-01 -2.78970569e-01 -1.53120443e-01
9.65897664e-02 -4.55266476e-01 -7.78655052e-01 -1.23691887e-01
-9.19046164e-01 1.90782085e-01 6.69924617e-01 6.65884376e-01
1.81821108e-01 3.70962292e-01 3.45698118e-01 -1.11153865e+00
4.89346743e-01 -6.48229241e-01 -8.76527607e-01 1.33574707e-02
-6.62228584e-01 6.64464310e-02 5.94968796e-01 4.56541806e-01
-1.05310905e+00 2.30768323e-02 -1.34266600e-01 7.15903565e-03
-3.51688385e-01 6.94034338e-01 7.03685135e-02 -7.61069125e-03
4.63402867e-01 3.13999623e-01 -3.36249858e-01 -4.57304269e-01
3.12081695e-01 1.00130951e+00 2.54106790e-01 -3.64071012e-01
1.01099563e+00 4.48721737e-01 -7.80414278e-03 -8.67633045e-01
-7.74269581e-01 -6.41626596e-01 -6.90211058e-01 -2.54363060e-01
1.17739618e+00 -1.06131685e+00 -5.58298826e-01 4.48014677e-01
-1.09476781e+00 -7.07894638e-02 1.46851614e-01 2.69494981e-01
1.06545530e-01 -1.84628531e-01 1.27802506e-01 -9.73864436e-01
-5.28981209e-01 -1.05335498e+00 6.60825968e-01 5.46286166e-01
1.60648748e-02 -1.13786769e+00 2.14079842e-01 5.76958299e-01
8.31571877e-01 6.87132657e-01 1.02209032e+00 -4.49929833e-01
-9.18172300e-01 -8.22285488e-02 -3.08165133e-01 4.70833957e-01
1.04784042e-01 1.37867317e-01 -1.35979629e+00 -1.49958774e-01
-3.12635265e-02 -4.78968844e-02 1.14029539e+00 2.18299001e-01
1.00249958e+00 -2.44164258e-01 -1.56697318e-01 7.60868788e-01
1.78507686e+00 3.82751137e-01 6.71697915e-01 5.37987113e-01
5.86530089e-01 7.03577042e-01 2.45812714e-01 1.74577028e-01
3.19967538e-01 3.71970505e-01 7.86509454e-01 -1.66471764e-01
5.53402491e-02 2.30339736e-01 1.32933065e-01 9.37199891e-01
-2.56779552e-01 -2.03940153e-01 -1.19963646e+00 6.79755151e-01
-1.52301133e+00 -9.95375335e-01 -4.19457078e-01 2.02973652e+00
6.95617437e-01 3.15888435e-01 -3.16110700e-01 4.56110351e-02
3.58502805e-01 2.44289711e-01 -2.90774286e-01 -9.76436794e-01
-1.81555480e-01 6.49194419e-01 7.82427788e-01 4.52552140e-01
-1.09540820e+00 1.03415394e+00 5.65395689e+00 4.83340412e-01
-1.70945323e+00 6.84156641e-02 8.18886995e-01 2.27579907e-01
-3.15946132e-01 2.10738182e-01 -7.32204735e-01 7.29720950e-01
1.28948486e+00 3.42547931e-02 3.31342369e-01 4.47152168e-01
5.76954007e-01 -4.89930779e-01 -5.26667595e-01 5.04738092e-01
-4.13244843e-01 -1.60451722e+00 9.59612522e-03 9.95171666e-02
9.44776952e-01 5.02454460e-01 -3.69144566e-02 7.73841813e-02
4.71061558e-01 -1.55559146e+00 6.36145771e-01 6.55359268e-01
5.07178485e-01 -6.74068868e-01 1.06700861e+00 5.09972095e-01
-1.11893201e+00 -6.96237907e-02 -5.26920378e-01 -4.28248852e-01
-4.11814332e-01 8.01288903e-01 -1.11459124e+00 7.17371106e-01
7.34755754e-01 4.31506991e-01 -5.86078465e-01 1.21850729e+00
-8.40820298e-02 9.46293235e-01 -6.04345322e-01 -1.55926973e-01
6.48678601e-01 -2.61899710e-01 4.35924500e-01 1.24700916e+00
2.89594561e-01 7.59320660e-03 -1.02101631e-01 7.09563076e-01
-9.88650844e-02 -1.16846999e-02 -5.98625839e-01 -8.23675543e-02
3.14075440e-01 1.24367058e+00 -8.00396025e-01 -2.29979441e-01
-9.48287621e-02 5.55620313e-01 1.28117934e-01 -4.58022468e-02
-6.50241494e-01 -3.06562573e-01 7.19748735e-01 2.68855214e-01
3.54675829e-01 -7.12456107e-02 -3.71989816e-01 -9.72099423e-01
-8.40983316e-02 -8.75075877e-01 -1.27943292e-01 -7.09939957e-01
-8.53964448e-01 8.27224076e-01 -3.24116498e-01 -1.04893911e+00
-1.01577602e-01 -6.58128560e-01 -1.07860398e+00 1.41863644e+00
-1.86306596e+00 -1.23224342e+00 -5.41047275e-01 6.82739392e-02
2.59523183e-01 -1.46099180e-01 8.97647619e-01 7.36880973e-02
-2.92340010e-01 -2.28942800e-02 5.89796305e-01 3.03841271e-02
4.23145533e-01 -1.64694309e+00 4.12265152e-01 8.14369023e-01
2.10008562e-01 3.74331683e-01 7.72258341e-01 -6.23256922e-01
-1.28756380e+00 -1.53309751e+00 7.81271517e-01 -4.11168098e-01
3.14226717e-01 -3.63053441e-01 -8.78979564e-01 1.36401266e-01
4.04772699e-01 -1.01493932e-01 5.91593444e-01 -1.26548167e-02
-1.17777012e-01 -2.96640486e-01 -9.95355070e-01 2.26938277e-02
3.16194803e-01 -4.23698813e-01 -5.36334991e-01 3.70769680e-01
5.15331209e-01 -2.63810128e-01 -4.62052613e-01 -7.47103915e-02
5.56856930e-01 -1.26606441e+00 4.75927532e-01 -5.31748354e-01
6.43509269e-01 -4.52289701e-01 -1.98800638e-01 -1.79640269e+00
-2.63610721e-01 -1.18327238e-01 4.00028139e-01 8.93452704e-01
6.24308109e-01 -7.47713029e-01 9.30779994e-01 2.27219239e-01
-2.22810164e-01 -7.35744774e-01 -9.86206949e-01 -4.42387551e-01
5.29679716e-01 -3.04882705e-01 6.88709080e-01 1.04336166e+00
-6.75074875e-01 1.50271550e-01 6.79702545e-03 3.49293053e-01
4.29235786e-01 4.32693064e-01 6.50437772e-01 -1.42916596e+00
-1.13498732e-01 -4.53394562e-01 -3.40601146e-01 -3.26988220e-01
-1.79163143e-01 -1.06870830e+00 -4.84578125e-02 -1.72072673e+00
2.20674928e-02 -8.70317876e-01 -2.33879671e-01 1.98765919e-01
-5.06749414e-02 1.71144202e-01 3.10288638e-01 2.25904062e-01
1.27751455e-01 3.05145383e-01 1.28349531e+00 -2.31907770e-01
1.48062676e-01 1.11971952e-01 -6.81195438e-01 5.68166792e-01
1.01972210e+00 -6.08208895e-01 -1.66372702e-01 -5.39738595e-01
5.98154247e-01 -1.75131634e-01 3.24120998e-01 -1.21870637e+00
1.50735065e-01 -2.39638031e-01 7.58497834e-01 -5.89852750e-01
1.17383666e-01 -8.90830517e-01 4.68449831e-01 6.33705735e-01
-1.04363710e-01 1.11733422e-01 2.01527610e-01 4.05305237e-01
-4.99249250e-01 -1.36989504e-01 7.85304189e-01 -5.30717850e-01
-7.08458066e-01 1.73406765e-01 -2.38432273e-01 -2.11416528e-01
8.38934839e-01 -1.95839748e-01 -2.24510252e-01 -2.77658939e-01
-5.82022965e-01 2.93506235e-01 3.55857342e-01 3.69837970e-01
3.24298739e-01 -8.05467844e-01 -9.65680361e-01 1.14118390e-01
-1.71007544e-01 2.24445492e-01 8.72235373e-02 3.64328653e-01
-1.04844546e+00 7.94109523e-01 -2.01915562e-01 -5.34786642e-01
-1.17951298e+00 -9.47619304e-02 7.14436650e-01 -4.26507473e-01
-1.60448194e-01 1.00292134e+00 5.80131561e-02 -5.39601028e-01
-2.67760158e-01 -6.07229292e-01 -4.53840911e-01 2.99935818e-01
2.36239478e-01 -5.24621904e-02 3.73625845e-01 -5.13292789e-01
-2.81454504e-01 5.30834913e-01 9.51124802e-02 -5.00542112e-02
1.70385885e+00 5.23005724e-02 -3.11025947e-01 5.98496199e-01
9.87856150e-01 -3.30304503e-02 -9.46393967e-01 1.72151234e-02
6.18060082e-02 -5.49006462e-01 2.50566989e-01 -1.29901063e+00
-1.23554361e+00 1.39192319e+00 1.00684845e+00 3.91212583e-01
9.06380296e-01 -4.32870775e-01 6.52830780e-01 4.86499846e-01
1.91764608e-01 -8.50429833e-01 -4.46406394e-01 7.21739471e-01
7.79144585e-01 -1.52441609e+00 4.73740883e-02 -7.64232203e-02
-4.37423378e-01 1.12832189e+00 3.72226954e-01 -1.89613163e-01
7.89306998e-01 1.30918548e-01 1.77052572e-01 -4.28414524e-01
-1.01808369e+00 -2.96950907e-01 2.44231865e-01 4.52026457e-01
6.24322712e-01 6.59589887e-01 9.51364711e-02 2.68608600e-01
-6.61321282e-01 -1.45374447e-01 5.44819653e-01 7.78706789e-01
-8.24083567e-01 -1.00188124e+00 -5.65192342e-01 6.98416173e-01
-5.73695421e-01 -2.80454993e-01 -4.52330828e-01 6.08060300e-01
6.50589824e-01 9.13340449e-01 1.68602660e-01 -4.30199593e-01
-2.00658552e-02 1.26398548e-01 1.56409681e-01 -6.56999648e-01
-1.01302731e+00 -3.02226096e-01 2.47102648e-01 -2.14717403e-01
-3.63430470e-01 -5.11929035e-01 -9.77769196e-01 -4.90724921e-01
-3.62818956e-01 6.51063025e-01 1.33092058e+00 9.81352627e-01
2.43640929e-01 6.33769453e-01 8.69126260e-01 -6.26143336e-01
-3.12205404e-01 -1.11330879e+00 -3.80295753e-01 2.99838573e-01
3.54758918e-01 -4.31276411e-01 -2.08047241e-01 1.43121481e-01] | [9.740118980407715, -1.5892373323440552] |
0ecff8f4-757c-43bb-b52a-82b245db366a | reference-production-in-human-computer | null | null | https://aclanthology.org/L18-1474 | https://aclanthology.org/L18-1474.pdf | Reference production in human-computer interaction: Issues for Corpus-based Referring Expression Generation | null | ["r{\\'e}", 'Iv Paraboni', 'Danillo Rocha'] | 2018-05-01 | reference-production-in-human-computer-1 | https://aclanthology.org/L18-1474 | https://aclanthology.org/L18-1474.pdf | lrec-2018-5 | ['referring-expression-generation'] | ['computer-vision'] | [-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] | [-7.3108696937561035, 3.785184144973755] |
084668c5-7dca-4e15-b936-23e76fee4572 | model-agnostic-meta-learning-for-multilingual | 2303.02513 | null | https://arxiv.org/abs/2303.02513v1 | https://arxiv.org/pdf/2303.02513v1.pdf | Model-Agnostic Meta-Learning for Multilingual Hate Speech Detection | Hate speech in social media is a growing phenomenon, and detecting such toxic content has recently gained significant traction in the research community. Existing studies have explored fine-tuning language models (LMs) to perform hate speech detection, and these solutions have yielded significant performance. However, most of these studies are limited to detecting hate speech only in English, neglecting the bulk of hateful content that is generated in other languages, particularly in low-resource languages. Developing a classifier that captures hate speech and nuances in a low-resource language with limited data is extremely challenging. To fill the research gap, we propose HateMAML, a model-agnostic meta-learning-based framework that effectively performs hate speech detection in low-resource languages. HateMAML utilizes a self-supervision strategy to overcome the limitation of data scarcity and produces better LM initialization for fast adaptation to an unseen target language (i.e., cross-lingual transfer) or other hate speech datasets (i.e., domain generalization). Extensive experiments are conducted on five datasets across eight different low-resource languages. The results show that HateMAML outperforms the state-of-the-art baselines by more than 3% in the cross-domain multilingual transfer setting. We also conduct ablation studies to analyze the characteristics of HateMAML. | ['Tanmoy Chakraborty', 'Tanmay Garg', 'Eshaan Tanwar', 'Roy Ka-Wei Lee', 'Md Rabiul Awal'] | 2023-03-04 | null | null | null | null | ['hate-speech-detection', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [-2.49348715e-01 -3.38783592e-01 -2.65638530e-01 5.12791574e-02
-7.78753281e-01 -6.14814699e-01 4.83809531e-01 -1.52054220e-01
-4.40665394e-01 5.38750887e-01 3.89659464e-01 -1.88079387e-01
6.53972626e-01 -2.34117821e-01 -3.85363847e-01 -5.51023245e-01
2.79037356e-01 2.24444922e-02 4.29303087e-02 -2.06921101e-01
1.60582997e-02 1.09519161e-01 -1.08054435e+00 3.32343072e-01
1.39183331e+00 3.48060094e-02 1.23174667e-01 4.17879522e-01
-1.01768253e-02 9.60842848e-01 -1.09122300e+00 -7.87632167e-01
-2.14578867e-01 -4.26170528e-01 -6.83383942e-01 6.25724345e-02
6.53437376e-01 -5.92691600e-01 -3.97138268e-01 1.14884675e+00
8.52693856e-01 -1.11174867e-01 5.92273712e-01 -1.21539056e+00
-1.38634050e+00 5.61648309e-01 -6.66140020e-01 4.37734425e-01
1.10373579e-01 2.10044965e-01 5.52180290e-01 -1.11060452e+00
4.84274089e-01 1.49316156e+00 4.33371037e-01 9.26840961e-01
-1.00923240e+00 -1.21306860e+00 -7.54779801e-02 8.85829106e-02
-1.48363543e+00 -5.78797162e-01 9.47582185e-01 -7.38904536e-01
9.96662199e-01 -1.07694007e-01 6.38842806e-02 1.93740308e+00
-1.77957684e-01 1.16903639e+00 1.25784671e+00 -4.70044553e-01
-8.94348323e-02 6.62833571e-01 -8.90460797e-03 4.48484331e-01
1.91654280e-01 -4.24029857e-01 -8.55632663e-01 -2.33169496e-01
1.57553658e-01 4.23672237e-03 -2.08392739e-01 3.32013041e-01
-6.44303918e-01 1.28098476e+00 5.58821745e-02 5.77973843e-01
-9.22204182e-03 -5.47559083e-01 7.98446119e-01 2.42142454e-01
1.24039161e+00 5.42253613e-01 -1.72218129e-01 -3.72902691e-01
-8.82993460e-01 1.94680355e-02 6.52499437e-01 8.41184616e-01
4.32727784e-01 3.14488858e-01 -2.63184793e-02 1.52497232e+00
4.06189486e-02 1.02381134e+00 6.29419923e-01 -3.23517114e-01
9.52118933e-01 3.83318484e-01 -1.74690530e-01 -9.35360849e-01
-1.40352517e-01 -3.78077388e-01 -6.19545877e-01 -3.41865420e-01
1.56046182e-01 -6.52229607e-01 -8.48550856e-01 1.75984681e+00
1.93757296e-01 1.58647329e-01 -7.52734542e-02 7.25929499e-01
6.81113243e-01 8.63320589e-01 4.61071223e-01 -1.59016460e-01
1.14574611e+00 -1.15093386e+00 -9.52476025e-01 -5.00869274e-01
1.30006838e+00 -9.25073266e-01 1.52367699e+00 2.65562952e-01
-5.95884800e-01 -7.61525333e-02 -8.81659031e-01 -2.43671328e-01
-7.62781322e-01 -3.52685601e-02 2.41089031e-01 1.01089919e+00
-5.82747400e-01 -9.65982676e-02 -4.17217880e-01 -5.98996341e-01
4.22857136e-01 -2.06017822e-01 -4.63408649e-01 -2.90538251e-01
-1.59515119e+00 1.25863683e+00 1.67199776e-01 -2.42836550e-01
-9.49295461e-01 -8.31613243e-01 -9.62893069e-01 -2.82203615e-01
2.50453532e-01 8.82313848e-02 1.09513259e+00 -1.11779654e+00
-1.06613207e+00 1.15290582e+00 -1.75007835e-01 -2.90185928e-01
1.39173836e-01 -7.06264079e-01 -7.69200444e-01 -1.57101795e-01
2.37240642e-01 3.08593690e-01 1.10544872e+00 -1.12054873e+00
-3.52423787e-01 -3.24211150e-01 -1.12039506e-01 1.16193332e-01
-1.45296371e+00 7.64289021e-01 -5.28575368e-02 -8.70663524e-01
-9.44131911e-01 -1.04678273e+00 6.07212067e-01 -7.12959468e-01
-5.85475862e-01 -3.84395808e-01 1.43576229e+00 -1.14997780e+00
2.01692533e+00 -2.39125538e+00 4.30791713e-02 -2.98653275e-01
1.50089681e-01 1.02214062e+00 -1.25287205e-01 6.65054440e-01
8.89119729e-02 4.59724426e-01 -2.92061329e-01 -7.15117753e-01
-6.50430471e-02 -6.11845143e-02 -4.65886444e-01 5.06848812e-01
3.80771101e-01 5.18233836e-01 -9.67176318e-01 -5.99397182e-01
7.87719786e-02 6.74891829e-01 -5.32517791e-01 5.40498555e-01
1.98796690e-01 1.94936231e-01 -2.40290284e-01 8.45529497e-01
5.77210724e-01 1.95746794e-01 -1.93197563e-01 3.65367264e-01
-2.60281384e-01 4.28087562e-01 -4.00629818e-01 1.32483363e+00
-7.17228115e-01 9.25125241e-01 2.48399585e-01 -4.10689652e-01
9.13452804e-01 4.15603727e-01 3.84201705e-02 -7.21752465e-01
1.39693812e-01 3.36873263e-01 1.37859240e-01 -7.95095742e-01
4.51347709e-01 -1.19119033e-01 -8.77540037e-02 4.29100394e-01
4.58786823e-02 1.17033280e-01 1.63934395e-01 1.96377248e-01
1.11777663e+00 -2.30326235e-01 7.22832754e-02 -4.05011065e-02
5.02222300e-01 -6.06024452e-02 3.57343733e-01 4.17304605e-01
-7.47977972e-01 3.29726547e-01 1.57885134e-01 -4.77759093e-02
-1.00041139e+00 -7.91714430e-01 -1.65155143e-01 1.77610385e+00
-3.62034976e-01 -5.30710936e-01 -1.05196846e+00 -8.84723544e-01
-3.00848186e-02 9.29783642e-01 -4.90180343e-01 -3.53647709e-01
-6.28915548e-01 -8.63095582e-01 9.26794589e-01 1.79499090e-01
4.49489236e-01 -1.32185638e+00 -1.55928612e-01 -1.40447449e-02
-2.62468547e-01 -1.35819614e+00 -6.23662174e-01 6.28833994e-02
-1.74599633e-01 -6.63478673e-01 -9.84520257e-01 -9.40026045e-01
4.59547758e-01 6.09269679e-01 8.60584676e-01 1.07653670e-01
-9.18281972e-02 -7.07441196e-03 -6.69657350e-01 -7.13558972e-01
-6.76864207e-01 5.81367195e-01 2.76122898e-01 5.51690310e-02
9.85219002e-01 -3.29228044e-01 -3.09398603e-02 1.65587634e-01
-9.59433436e-01 -1.36288270e-01 3.84883374e-01 7.13694751e-01
-1.83492392e-01 1.31982282e-01 6.99562788e-01 -1.16207862e+00
9.62049723e-01 -1.04289806e+00 -6.17811047e-02 1.16067320e-01
-4.75247204e-02 -4.31025237e-01 1.00183976e+00 -8.84479582e-01
-1.21210527e+00 -4.72517580e-01 3.68478820e-02 -6.71731889e-01
-3.08739185e-01 4.27596718e-01 -2.10252404e-01 2.73939043e-01
7.08943784e-01 1.95356715e-03 -3.83609623e-01 -7.58348763e-01
2.11529091e-01 1.35358071e+00 2.60312170e-01 -5.46193004e-01
1.19472337e+00 -2.27131248e-02 -8.22503209e-01 -1.40453851e+00
-1.18765712e+00 -6.93307519e-01 -6.69782341e-01 -1.50327876e-01
8.94412577e-01 -1.11564553e+00 -2.65682824e-02 9.06285644e-01
-1.28822410e+00 -3.72269303e-01 5.52538931e-01 3.22465062e-01
2.61566281e-01 2.16389731e-01 -9.45627034e-01 -1.21911037e+00
-4.89785343e-01 -9.92411613e-01 1.02166378e+00 -3.55606936e-02
-4.50794369e-01 -1.21157610e+00 4.67229366e-01 6.61328375e-01
3.59329432e-01 5.26281223e-02 1.03291154e+00 -1.00267482e+00
1.60040885e-01 -9.81009658e-03 -1.85368896e-01 7.18498409e-01
1.20169923e-01 1.95332110e-01 -1.25138128e+00 -4.72235918e-01
-1.76545307e-01 -1.03349578e+00 5.43234706e-01 -1.76314890e-01
8.64068806e-01 -5.16106844e-01 -1.62201717e-01 2.74929076e-01
9.09981847e-01 3.89655903e-02 3.85042399e-01 4.68664408e-01
1.12010276e+00 7.66904950e-01 4.63366598e-01 4.07871962e-01
3.80829036e-01 6.28675759e-01 1.53643703e-02 -1.98487565e-01
-3.07247847e-01 -5.92213154e-01 1.01067889e+00 1.44430888e+00
3.75543058e-01 -3.39654624e-01 -1.23666155e+00 8.59397471e-01
-1.45805681e+00 -1.06187534e+00 -1.33041693e-02 1.86906707e+00
1.11482418e+00 -5.07929213e-02 5.15989840e-01 -1.68257207e-01
9.81354892e-01 4.91379857e-01 -4.66112226e-01 -7.49958515e-01
-2.81170100e-01 -3.86057608e-02 2.43224710e-01 3.01602602e-01
-1.43885767e+00 1.52945948e+00 5.33604956e+00 1.00416148e+00
-1.24122012e+00 5.75164378e-01 3.66125494e-01 -1.31758645e-01
7.25040659e-02 -4.81208980e-01 -8.78514469e-01 8.24179947e-01
1.09886622e+00 -4.64583337e-02 5.04944980e-01 1.07546592e+00
1.58081636e-01 3.06675792e-01 -6.82596385e-01 1.06395388e+00
6.52896881e-01 -5.68494737e-01 -1.10851578e-01 1.82578117e-01
8.63451779e-01 1.86174020e-01 5.53862631e-01 7.55491614e-01
2.76815653e-01 -1.04027045e+00 5.69667101e-01 -2.27029964e-01
6.84229851e-01 -9.92426276e-01 6.98121190e-01 6.38685822e-01
-5.85798621e-01 -1.68225646e-01 -3.77029121e-01 2.35253740e-02
4.69680019e-02 4.12652075e-01 -9.57625329e-01 -1.43662423e-01
7.78529108e-01 7.61413872e-01 -8.79812956e-01 4.75237757e-01
-4.44927484e-01 1.26557815e+00 -2.56655756e-02 -3.07091093e-03
3.76264274e-01 6.08079657e-02 5.17399251e-01 1.79741096e+00
1.07250810e-01 -2.71747351e-01 1.26121581e-01 7.70050585e-01
-3.94132912e-01 3.92661154e-01 -1.03442049e+00 -6.22192562e-01
8.27927172e-01 1.10536814e+00 7.38093778e-02 -2.44563147e-02
-8.42996359e-01 1.19154894e+00 9.15035844e-01 2.28895664e-01
-8.19673121e-01 -4.61445332e-01 8.36224318e-01 1.03280656e-01
-9.07992050e-02 -3.17380667e-01 7.62237534e-02 -1.20602906e+00
-1.05975620e-01 -1.42653275e+00 3.20227295e-01 -5.29208362e-01
-1.76130140e+00 4.94181156e-01 -7.19791278e-02 -7.96052694e-01
-3.29682171e-01 -5.57410955e-01 -4.78344440e-01 9.51993942e-01
-1.48732746e+00 -1.44841790e+00 1.82288960e-01 6.46845996e-01
7.76733994e-01 -3.35060835e-01 8.16400886e-01 6.17355704e-01
-1.27784920e+00 9.15623009e-01 3.34513783e-02 7.42621005e-01
1.27385283e+00 -9.63419676e-01 3.19603205e-01 1.22286654e+00
-8.33806843e-02 8.05142045e-01 5.37012041e-01 -9.18253303e-01
-1.11807168e+00 -1.21792734e+00 1.24048245e+00 -9.36538696e-01
1.13350189e+00 -7.81178474e-01 -1.32181215e+00 6.60763383e-01
5.08684933e-01 -3.11046064e-01 9.99194741e-01 4.27994162e-01
-8.66990030e-01 2.82656789e-01 -9.24804449e-01 7.06031859e-01
9.53226328e-01 -1.20329738e+00 -5.83739161e-01 4.32093173e-01
7.37840891e-01 -1.33436218e-01 -6.55267656e-01 -5.61934337e-02
2.65455484e-01 -6.91024005e-01 4.99554038e-01 -8.50167811e-01
5.44441640e-01 4.75213341e-02 5.27154543e-02 -1.72446966e+00
-3.90025467e-01 -7.03725636e-01 -3.34208280e-01 1.84790170e+00
3.04498017e-01 -3.28088492e-01 2.43114978e-01 5.35131991e-01
-2.17488080e-01 -3.75696957e-01 -6.76211059e-01 -8.43735456e-01
5.24261713e-01 -2.13808343e-01 2.32539997e-01 1.67534530e+00
1.13538571e-01 7.43060708e-01 -9.51792479e-01 2.79943466e-01
4.38589096e-01 -4.63838607e-01 9.73042309e-01 -8.25029910e-01
2.06381559e-01 -2.55018860e-01 -2.61283517e-02 -5.50729811e-01
1.03658783e+00 -9.00199175e-01 -1.49391159e-01 -9.81442511e-01
5.31631708e-01 -1.31414041e-01 -4.38607894e-02 6.96625173e-01
-5.41785896e-01 1.99820757e-01 4.31346267e-01 1.99375913e-01
-5.25857985e-01 4.89510208e-01 1.00927055e+00 -2.32114911e-01
-1.77408814e-01 -3.68308544e-01 -5.33100307e-01 8.66054296e-01
8.51164520e-01 -6.68817699e-01 -3.80778730e-01 -6.68303668e-01
-1.15197971e-01 -7.09598958e-01 8.25119987e-02 -8.17238748e-01
-7.17755854e-02 -1.08091161e-01 8.63193050e-02 -2.43398681e-01
3.15107763e-01 -5.84373891e-01 -6.08978212e-01 1.06474385e-01
-3.18360865e-01 1.33977801e-01 1.84036374e-01 -6.46206662e-02
-1.85847268e-01 -3.55263650e-01 8.80907714e-01 1.11983657e-01
-6.22768879e-01 2.28158668e-01 -5.26966393e-01 7.25178838e-01
7.21217930e-01 1.55158490e-01 -1.00944555e+00 -4.69677836e-01
-2.84789111e-02 -1.15548316e-02 6.04604125e-01 9.22616661e-01
3.14577848e-01 -1.20026994e+00 -9.92879152e-01 2.49604248e-02
5.29858589e-01 -5.93511939e-01 1.46418750e-01 6.38842821e-01
-4.84407879e-02 4.80383813e-01 -4.22093533e-02 -1.29228190e-01
-1.37233317e+00 6.85430408e-01 1.37683049e-01 -1.91529140e-01
-2.44068056e-01 6.98169947e-01 3.49237412e-01 -5.24658561e-01
2.36725494e-01 4.68078047e-01 -3.58478241e-02 1.92127571e-01
7.72966623e-01 4.60474223e-01 -1.36350706e-01 -1.24733484e+00
-5.12292802e-01 -1.37797068e-03 -3.75299215e-01 8.88485312e-02
1.13055265e+00 -1.02228686e-01 -1.01675689e-01 6.02120042e-01
1.50377560e+00 5.81514060e-01 -8.53966832e-01 -3.72444361e-01
1.42502114e-01 -6.40579522e-01 1.11260466e-01 -8.28781247e-01
-7.69933462e-01 1.17996442e+00 3.76437038e-01 2.63514102e-01
7.86189497e-01 -5.41668711e-03 1.26070344e+00 2.35671610e-01
1.75519213e-01 -1.44419646e+00 3.29749614e-01 9.43079591e-01
8.64297450e-01 -1.43717539e+00 -2.28339627e-01 -2.69956261e-01
-9.99805748e-01 6.75456583e-01 1.21981645e+00 2.97429949e-01
4.31597084e-01 2.26012796e-01 4.60132271e-01 -5.68984039e-02
-5.42786241e-01 -2.19220206e-01 2.98051924e-01 8.04246604e-01
9.52753007e-01 1.84272021e-01 1.21561261e-02 5.58509409e-01
-3.68245333e-01 -6.90126598e-01 3.68201196e-01 8.98813665e-01
-4.63726729e-01 -9.38317835e-01 -5.78177691e-01 5.64836226e-02
-7.46831834e-01 -4.84301776e-01 -1.02779078e+00 9.60059643e-01
2.78127521e-01 1.25743270e+00 -2.11976454e-01 -5.05166590e-01
3.05862397e-01 3.26131821e-01 1.27354220e-01 -9.59969163e-01
-7.66886711e-01 5.12536615e-02 2.64700741e-01 -7.46243298e-02
-2.89996326e-01 -4.80943948e-01 -6.69800341e-01 -6.35420918e-01
-4.14824784e-01 1.60676811e-03 4.68863726e-01 9.28854704e-01
3.88813645e-01 5.60112782e-02 8.01784515e-01 -4.59066838e-01
-2.62966901e-01 -1.29171705e+00 -5.83551466e-01 6.86962008e-01
5.22452593e-01 -7.50007153e-01 -5.81788123e-01 -1.05304174e-01] | [8.82630729675293, 10.558805465698242] |
d3dff89c-58e0-40af-99c6-cbb5185f1b6d | robust-and-explainable-contextual-anomaly | 2302.11239 | null | https://arxiv.org/abs/2302.11239v2 | https://arxiv.org/pdf/2302.11239v2.pdf | Explainable Contextual Anomaly Detection using Quantile Regression Forests | Traditional anomaly detection methods aim to identify objects that deviate from most other objects by treating all features equally. In contrast, contextual anomaly detection methods aim to detect objects that deviate from other objects within a context of similar objects by dividing the features into contextual features and behavioral features. In this paper, we develop connections between dependency-based traditional anomaly detection methods and contextual anomaly detection methods. Based on resulting insights, we propose a novel approach to inherently interpretable contextual anomaly detection that uses Quantile Regression Forests to model dependencies between features. Extensive experiments on various synthetic and real-world datasets demonstrate that our method outperforms state-of-the-art anomaly detection methods in identifying contextual anomalies in terms of accuracy and interpretability. | ['Matthijs van Leeuwen', 'Zhong Li'] | 2023-02-22 | null | null | null | null | ['contextual-anomaly-detection'] | ['miscellaneous'] | [ 3.02222759e-01 -3.69704694e-01 7.20013604e-02 -6.77698195e-01
-4.06609416e-01 -3.32659572e-01 7.67329991e-01 8.44981968e-01
7.15566501e-02 4.35164541e-01 -2.09658012e-01 -4.28474188e-01
-4.39590931e-01 -7.79399633e-01 -4.74327534e-01 -5.18073440e-01
-5.09406149e-01 2.44330198e-01 5.32390535e-01 6.58554137e-02
7.01509058e-01 8.18756998e-01 -1.73799443e+00 3.93777519e-01
9.14263785e-01 1.26062965e+00 -9.08519030e-01 6.08601332e-01
-3.15520644e-01 8.83427501e-01 -7.02716053e-01 -1.57165170e-01
1.93520084e-01 -4.67717767e-01 -7.17358708e-01 1.29863590e-01
7.27129340e-01 -4.10152793e-01 5.93558252e-02 9.49600518e-01
-2.77740538e-01 3.27413499e-01 8.68125200e-01 -1.87567616e+00
-8.16064298e-01 2.50543714e-01 -7.35506058e-01 9.27430272e-01
5.89572966e-01 2.31700353e-02 1.28769243e+00 -8.90528142e-01
3.47319208e-02 1.35176098e+00 7.52378166e-01 3.61576319e-01
-1.63845646e+00 -3.74434680e-01 9.13194180e-01 5.52650154e-01
-1.11727059e+00 -6.64928406e-02 7.11298168e-01 -6.09002471e-01
1.03555429e+00 7.11832166e-01 4.95425940e-01 1.20936179e+00
3.47723067e-01 6.64426506e-01 8.78449082e-01 -3.89042646e-01
5.27994990e-01 -4.82738674e-01 6.49117649e-01 3.36692065e-01
7.87768126e-01 2.27063656e-01 -4.36482221e-01 -9.18247104e-01
8.15057755e-02 6.27399445e-01 4.18948352e-01 -3.44716281e-01
-7.61982381e-01 9.15106475e-01 1.86989740e-01 1.21704467e-01
-5.40568531e-01 2.30996147e-01 4.56067622e-01 4.34334755e-01
7.00624108e-01 5.50918043e-01 -4.18003738e-01 1.12539388e-01
-4.43679720e-01 6.33064568e-01 4.82357383e-01 6.35131836e-01
8.12056601e-01 1.73684418e-01 -3.63875568e-01 5.84336221e-01
4.73370224e-01 3.55965793e-01 3.63209069e-01 -5.70769489e-01
1.45807536e-02 1.07769442e+00 1.15804806e-01 -1.12107980e+00
-3.09180170e-01 -1.40970275e-01 -5.02588987e-01 2.54899859e-01
3.51364046e-01 2.06474364e-01 -1.07177436e+00 1.43153894e+00
5.02502859e-01 5.26861966e-01 -2.65337408e-01 4.72300470e-01
2.54640967e-01 1.89439774e-01 4.09546942e-01 -1.94990716e-03
9.25227523e-01 -5.11679709e-01 -5.65893292e-01 -1.77429557e-01
6.47891283e-01 -4.18105930e-01 1.03296483e+00 2.31200069e-01
-4.51936126e-01 -1.33003131e-01 -7.19362617e-01 6.49593353e-01
-6.29089236e-01 -7.09988177e-01 6.38643742e-01 6.97177112e-01
-5.58053553e-01 5.85314691e-01 -1.08407390e+00 -3.78880501e-01
5.24257362e-01 4.68112305e-02 -3.11624050e-01 3.15648854e-01
-7.72809267e-01 5.90937972e-01 2.95408100e-01 -4.54269759e-02
-7.94946432e-01 -6.72834992e-01 -8.13429356e-01 -7.46586993e-02
6.51053011e-01 -2.41376922e-01 1.32691228e+00 -1.06859231e+00
-6.56806588e-01 5.02343774e-01 -3.97587121e-01 -6.17856264e-01
4.39442843e-01 -6.73591912e-01 -9.88023818e-01 -1.63950950e-01
2.92492092e-01 -2.55560577e-01 9.56455767e-01 -1.22940433e+00
-1.05278850e+00 -6.25775039e-01 -3.00681561e-01 -2.92787462e-01
-9.97794718e-02 1.14499815e-01 3.62720639e-01 -7.85305798e-01
4.00601596e-01 -7.39930391e-01 -4.39810306e-01 -1.59911066e-01
-9.14465189e-01 -4.19110090e-01 1.46860814e+00 -2.14919616e-02
1.41522622e+00 -2.00857687e+00 -6.58049643e-01 9.05307412e-01
3.99861753e-01 -1.04499713e-01 3.31744063e-03 3.65326643e-01
-3.09964716e-01 3.43606383e-01 -4.96796638e-01 -6.12762533e-02
-1.32411048e-01 5.45559764e-01 -6.72926009e-01 5.53635418e-01
6.29015207e-01 6.74274206e-01 -1.01565254e+00 -2.31956512e-01
1.49149746e-01 -3.23821872e-01 -5.45075536e-01 1.82956934e-01
-2.06689179e-01 5.13912082e-01 -7.74743021e-01 1.17630076e+00
5.91880620e-01 1.63817063e-01 -1.58604637e-01 8.09593558e-01
-4.60828878e-02 1.97585329e-01 -1.08275771e+00 8.00673187e-01
1.77049398e-01 4.79534060e-01 -5.49384713e-01 -1.35337496e+00
8.81262839e-01 -8.44120607e-02 6.24461055e-01 -6.29064918e-01
-3.18233192e-01 2.27491990e-01 9.63752046e-02 -4.98963714e-01
3.92725885e-01 1.77553654e-01 -2.59086728e-01 6.65098011e-01
-3.51325393e-01 3.48875821e-01 1.74382273e-02 5.34580909e-02
1.65328491e+00 -2.07578003e-01 5.74375451e-01 -2.42225543e-01
7.35245526e-01 -1.55979648e-01 8.22227955e-01 1.39679599e+00
-4.78518069e-01 3.37009698e-01 8.63930404e-01 -8.95705640e-01
-7.50591218e-01 -1.66668999e+00 -4.02398765e-01 1.42500246e+00
-1.15006737e-01 -4.49360371e-01 -4.19510812e-01 -1.34478951e+00
6.02622509e-01 9.83482957e-01 -1.00602901e+00 -2.20930323e-01
-5.30248284e-01 -7.38093257e-01 4.58150506e-01 6.76129282e-01
8.46336335e-02 -1.17866850e+00 -7.17208266e-01 8.29284042e-02
1.61893204e-01 -7.18617857e-01 6.65500313e-02 3.25645596e-01
-9.59478498e-01 -1.30070031e+00 4.61353809e-01 -7.14239627e-02
8.39618027e-01 1.59930751e-01 1.24073124e+00 3.07883143e-01
-5.73458314e-01 5.43326080e-01 -5.38878322e-01 -8.43294084e-01
-3.29885811e-01 -1.59649700e-01 2.96900511e-01 2.55794615e-01
9.35645461e-01 -6.94650888e-01 -3.66881222e-01 6.10223114e-01
-1.19760180e+00 -8.86821210e-01 1.86940491e-01 6.96975410e-01
7.10967243e-01 -8.31699222e-02 6.36112571e-01 -1.19544125e+00
5.44865847e-01 -1.05842149e+00 -5.72397888e-01 1.84648409e-01
-1.02655458e+00 5.15363701e-02 4.66719508e-01 -3.78175884e-01
-9.17775691e-01 -1.01870894e-01 2.60196120e-01 -1.98883161e-01
-8.09546649e-01 2.85275817e-01 -4.86693494e-02 2.89937288e-01
9.55544353e-01 2.45134994e-01 -3.09777707e-01 -3.91014159e-01
1.24524839e-01 4.20175642e-01 4.90458161e-01 -8.19705963e-01
1.04849648e+00 5.40701687e-01 2.28474766e-01 -7.08719373e-01
-8.93637955e-01 -5.44551849e-01 -8.27718616e-01 -2.54573673e-01
4.46426570e-01 -3.09699863e-01 -3.49222183e-01 4.37138915e-01
-5.65979302e-01 1.56237707e-01 -5.48526347e-01 1.03236437e-01
-4.04290974e-01 3.77126336e-01 -1.00972764e-01 -1.21619976e+00
7.42594898e-02 -6.20129108e-01 9.12735522e-01 -1.57663420e-01
-6.93178475e-01 -9.32303548e-01 4.28898185e-01 -6.21999390e-02
4.94365722e-01 7.20835090e-01 1.06957269e+00 -1.57677925e+00
-3.93743813e-01 -5.45364738e-01 -3.04483250e-02 2.58992404e-01
4.95716602e-01 4.69930559e-01 -9.42926526e-01 8.60647392e-03
-1.88746944e-01 1.25104561e-01 8.14649224e-01 2.24061683e-01
1.70605814e+00 -5.43297708e-01 -3.28195333e-01 2.57930785e-01
1.13808298e+00 3.87922287e-01 4.25844491e-01 6.80365026e-01
6.40874326e-01 5.65494537e-01 8.35348785e-01 5.53195834e-01
-1.49865463e-01 3.00790489e-01 8.30689371e-01 2.34948173e-01
7.30996311e-01 -1.24247178e-01 2.25515515e-01 -9.90102813e-02
-7.61399418e-02 -2.39881612e-02 -1.29835093e+00 7.35173166e-01
-2.27660656e+00 -1.27860904e+00 -4.41454440e-01 2.14357185e+00
2.31602639e-01 3.33171278e-01 4.43231523e-01 2.21254170e-01
8.29021811e-01 1.45588238e-02 -8.38802695e-01 -8.69661033e-01
5.19502163e-02 -4.50362749e-02 1.25816882e-01 7.33386874e-02
-1.49121463e+00 6.36811435e-01 7.43577909e+00 3.60309869e-01
-5.58297694e-01 -2.16531903e-01 7.69078076e-01 -2.60568619e-01
-3.20456445e-01 5.67694455e-02 -4.17460352e-01 4.52189565e-01
9.41612422e-01 -9.96292159e-02 -5.78798316e-02 1.24855697e+00
4.21540961e-02 -2.10416153e-01 -1.47440267e+00 4.09867972e-01
-1.86988369e-01 -7.77583599e-01 3.13831836e-01 1.47378966e-01
7.63915896e-01 -7.84950778e-02 1.94729805e-01 1.76784128e-01
4.79574263e-01 -9.94301438e-01 5.72907984e-01 7.75296569e-01
5.50915413e-02 -9.39123690e-01 7.14239836e-01 4.44255732e-02
-1.05161166e+00 -4.49791551e-01 -1.51800543e-01 -5.63134611e-01
-3.86402071e-01 9.00179863e-01 -6.19906783e-01 4.22923148e-01
1.10403395e+00 6.28895223e-01 -8.50533426e-01 1.01019585e+00
-2.38660768e-01 8.69777262e-01 -2.63868243e-01 2.67983615e-01
2.79396743e-01 -8.83559734e-02 9.02596831e-01 1.10371947e+00
1.87173471e-01 -2.66911000e-01 2.83959031e-01 9.37093079e-01
3.20435554e-01 1.84891477e-01 -1.08715665e+00 2.71981299e-01
4.40439582e-01 9.10478294e-01 -6.88586533e-01 -2.01674685e-01
-7.45098948e-01 5.92562318e-01 1.18843108e-01 2.96609104e-01
-7.59955883e-01 -3.11852068e-01 1.29565573e+00 2.64083296e-01
2.19033152e-01 9.49706063e-02 -4.44030315e-01 -1.10382020e+00
3.44508767e-01 -6.76130235e-01 8.29549134e-01 5.08284313e-04
-2.06042242e+00 2.96686530e-01 1.61196694e-01 -1.43228757e+00
-4.40205216e-01 -5.85987389e-01 -1.24555480e+00 4.48741794e-01
-1.15150404e+00 -8.49330842e-01 -3.12215596e-01 8.05624306e-01
3.60414445e-01 -3.01597863e-01 8.85539412e-01 -3.87066841e-01
-6.25812054e-01 5.34805179e-01 2.23435134e-01 2.89953947e-01
5.22117496e-01 -1.59261203e+00 8.15555096e-01 1.22771168e+00
2.66006470e-01 6.43749595e-01 6.81887448e-01 -7.86686718e-01
-6.99709713e-01 -1.36473382e+00 6.29617572e-01 -7.96725035e-01
1.03860390e+00 -1.90712094e-01 -1.37440670e+00 9.90693569e-01
-3.02583218e-01 6.84983552e-01 1.24527502e+00 6.25387490e-01
-7.63998151e-01 -1.41907884e-02 -1.32503974e+00 5.79303145e-01
1.31148493e+00 -4.40772980e-01 -9.41858411e-01 -4.35751723e-03
3.11271518e-01 7.29559585e-02 -5.48632681e-01 6.11261785e-01
5.43103755e-01 -1.23039711e+00 8.13306570e-01 -1.26529562e+00
3.08294803e-01 -3.72083426e-01 -3.57266486e-01 -1.10990191e+00
-4.83794183e-01 -2.88736105e-01 -5.38274884e-01 1.16322052e+00
2.66103625e-01 -1.17786658e+00 6.22422934e-01 7.65515268e-01
2.92486642e-02 -5.76261699e-01 -1.05836999e+00 -1.11285114e+00
2.09743660e-02 -9.73852813e-01 9.22289670e-01 1.08096802e+00
-6.01657033e-02 -9.55166072e-02 -4.77967523e-02 3.66634667e-01
6.78970039e-01 3.10619354e-01 6.73299909e-01 -1.90972662e+00
9.86083969e-02 -4.97020960e-01 -1.07839906e+00 -1.27508670e-01
3.23789716e-01 -4.67944413e-01 -2.80513917e-03 -8.42026532e-01
1.17885470e-01 -4.63252187e-01 -9.49232399e-01 7.93291688e-01
-4.64666337e-01 1.44248426e-01 -4.15175050e-01 2.18435779e-01
-8.22372139e-01 3.68312389e-01 3.45761895e-01 -4.25276570e-02
-2.97277540e-01 4.94013131e-01 -8.02982926e-01 1.05769527e+00
8.73155713e-01 -8.34949911e-01 -2.77644366e-01 1.28813595e-01
2.08655015e-01 -5.17890930e-01 6.31803036e-01 -1.00081336e+00
-2.25468457e-01 -7.51149237e-01 6.87749743e-01 -5.98960519e-01
-3.17691535e-01 -9.13446009e-01 -3.30434561e-01 4.56579119e-01
-4.53892797e-01 4.68224168e-01 1.40575111e-01 1.15042269e+00
-2.13844061e-01 -4.86258194e-02 4.55290884e-01 2.58306623e-01
-7.51440644e-01 3.17953676e-01 -9.20291781e-01 1.44818738e-01
1.41635180e+00 -1.94977254e-01 -4.39101905e-01 -2.66928792e-01
-9.78008509e-01 1.67227060e-01 3.11670601e-01 5.74915528e-01
7.30881035e-01 -1.43213856e+00 -7.79753745e-01 4.65049744e-01
7.00219870e-01 -1.32866561e-01 7.80897797e-04 7.80772269e-01
-2.34803602e-01 2.17969596e-01 -2.78585136e-01 -9.02173400e-01
-1.35476124e+00 5.62366545e-01 2.26535976e-01 -7.78270736e-02
-6.70362413e-01 3.67704183e-01 2.63449013e-01 -3.49304587e-01
6.80531040e-02 -4.67251241e-01 8.85498673e-02 -5.60075194e-02
5.58128476e-01 7.02343941e-01 7.57843703e-02 -4.42393512e-01
-7.59844601e-01 8.94714296e-02 -4.19718266e-01 1.58729672e-01
1.10456300e+00 1.07839748e-01 -2.65273660e-01 8.54064167e-01
6.69941545e-01 4.70647253e-02 -1.03760719e+00 -5.07624388e-01
9.10305679e-01 -9.96594071e-01 -4.83815014e-01 -6.54701710e-01
-7.93927550e-01 4.02318478e-01 6.37125731e-01 8.54663193e-01
1.29396701e+00 1.18964672e-01 1.26291826e-01 6.83110535e-01
-1.30754113e-02 -1.05950081e+00 2.46891260e-01 5.72859526e-01
7.03453660e-01 -1.36839294e+00 3.66416313e-02 -3.46849054e-01
-5.91048300e-01 1.12271893e+00 9.91661429e-01 -4.93649274e-01
8.55845451e-01 -5.58160543e-02 3.73425637e-03 -1.78298563e-01
-7.82329381e-01 -2.24436417e-01 4.16755229e-01 6.79458618e-01
1.45693377e-01 1.52434111e-01 -6.23903610e-02 4.42984790e-01
6.83013424e-02 -8.16141605e-01 4.89011347e-01 1.27362955e+00
-5.91324687e-01 -9.15540099e-01 -6.51595116e-01 9.78440106e-01
-9.33216155e-01 6.62421882e-02 -7.32545972e-01 7.84681380e-01
-1.13154799e-02 1.06790733e+00 7.03173935e-01 -3.25552583e-01
4.67225105e-01 5.20033479e-01 -2.44096965e-01 -6.40822589e-01
-4.67326820e-01 -2.59217173e-01 -4.57846113e-02 -8.91913354e-01
-5.29994033e-02 -1.04541898e+00 -1.14982855e+00 -2.58135647e-01
-2.90072441e-01 2.46491358e-02 2.71478117e-01 1.18227768e+00
3.11944902e-01 5.73749304e-01 8.45367670e-01 -3.09022307e-01
-4.86724406e-01 -8.65206599e-01 -6.69316292e-01 7.88284004e-01
7.23923028e-01 -8.23228538e-01 -5.91941893e-01 -3.29910487e-01] | [7.580821514129639, 2.5488498210906982] |
2f9fa8f7-0784-4382-b088-1c5aaef033ec | global-local-stepwise-generative-network-for | 2207.08808 | null | https://arxiv.org/abs/2207.08808v2 | https://arxiv.org/pdf/2207.08808v2.pdf | Global-Local Stepwise Generative Network for Ultra High-Resolution Image Restoration | While the research on image background restoration from regular size of degraded images has achieved remarkable progress, restoring ultra high-resolution (e.g., 4K) images remains an extremely challenging task due to the explosion of computational complexity and memory usage, as well as the deficiency of annotated data. In this paper we present a novel model for ultra high-resolution image restoration, referred to as the Global-Local Stepwise Generative Network (GLSGN), which employs a stepwise restoring strategy involving four restoring pathways: three local pathways and one global pathway. The local pathways focus on conducting image restoration in a fine-grained manner over local but high-resolution image patches, while the global pathway performs image restoration coarsely on the scale-down but intact image to provide cues for the local pathways in a global view including semantics and noise patterns. To smooth the mutual collaboration between these four pathways, our GLSGN is designed to ensure the inter-pathway consistency in four aspects in terms of low-level content, perceptual attention, restoring intensity and high-level semantics, respectively. As another major contribution of this work, we also introduce the first ultra high-resolution dataset to date for both reflection removal and rain streak removal, comprising 4,670 real-world and synthetic images. Extensive experiments across three typical tasks for image background restoration, including image reflection removal, image rain streak removal and image dehazing, show that our GLSGN consistently outperforms state-of-the-art methods. | ['Guangming Lu', 'Fanglin Chen', 'Wenjie Pei', 'Haobo Ji', 'Xin Feng'] | 2022-07-16 | null | null | null | null | ['image-dehazing', 'reflection-removal'] | ['computer-vision', 'computer-vision'] | [ 6.23767376e-01 -5.52143931e-01 4.52884197e-01 -1.64718196e-01
-1.06776965e+00 -7.46955127e-02 3.59224021e-01 -2.39012837e-01
1.15564108e-01 8.29802096e-01 4.44183409e-01 -2.09437922e-01
-1.94826588e-01 -1.02880621e+00 -6.57871366e-01 -1.18103778e+00
2.69170642e-01 -3.81864071e-01 2.88319409e-01 -7.17093885e-01
1.10590927e-01 7.04154134e-01 -1.86944485e+00 5.05573928e-01
1.20786572e+00 6.47428930e-01 4.52521324e-01 8.39839935e-01
8.01270455e-02 9.75861728e-01 -2.80515254e-01 9.71163288e-02
2.48230532e-01 -5.30201733e-01 -4.52666521e-01 2.77434498e-01
7.71815121e-01 -4.31556135e-01 -3.44565749e-01 1.41329324e+00
7.83237755e-01 2.92481065e-01 2.45303929e-01 -5.37995219e-01
-1.23732829e+00 3.02581072e-01 -9.32754993e-01 6.64067328e-01
3.77627134e-01 1.92576334e-01 5.81175923e-01 -1.01390755e+00
2.77095020e-01 1.33935308e+00 6.47158206e-01 4.35287625e-01
-1.45940185e+00 -5.85603654e-01 2.42580011e-01 6.68850541e-02
-1.23455381e+00 -4.67860490e-01 5.64764440e-01 -1.51252970e-01
6.80699348e-01 5.19067705e-01 2.50976920e-01 7.42243171e-01
3.60748559e-01 -1.81293096e-02 1.71062541e+00 -2.95399964e-01
9.04802233e-02 -3.22196871e-01 3.81630927e-01 3.85820389e-01
2.59625733e-01 2.26542741e-01 -4.29209411e-01 -2.26784628e-02
1.00272524e+00 5.71242087e-02 -7.54769504e-01 2.77163535e-01
-1.01343215e+00 3.69749457e-01 5.97762227e-01 1.98063433e-01
-5.22783816e-01 -1.65861800e-01 -1.15250237e-01 2.95848429e-01
8.43440533e-01 -6.44971207e-02 -3.31927627e-01 4.00836557e-01
-7.05331266e-01 2.97348976e-01 3.13048929e-01 5.17215371e-01
9.81932402e-01 4.19614822e-01 -3.34747165e-01 1.16650617e+00
9.13050473e-02 7.03067899e-01 1.21309139e-01 -9.11862791e-01
3.24583620e-01 9.42471810e-03 3.83089393e-01 -1.12703502e+00
-2.33525813e-01 -4.67141002e-01 -1.54526436e+00 4.91058648e-01
9.66961682e-02 2.45922059e-01 -1.23289883e+00 1.53764999e+00
3.10211778e-01 3.64440590e-01 2.20331058e-01 1.23225439e+00
1.29552519e+00 8.70178223e-01 -7.77940229e-02 -4.62201238e-01
1.55703104e+00 -8.38713169e-01 -9.05451775e-01 -3.74464720e-01
-2.66388178e-01 -1.02999246e+00 1.25006568e+00 1.95698246e-01
-1.23599744e+00 -8.32087219e-01 -9.08587754e-01 -1.08562343e-01
-1.70402050e-01 -1.76227331e-01 5.22326827e-01 3.52088183e-01
-1.41746891e+00 6.08402908e-01 -3.92973393e-01 -3.85639042e-01
3.83925647e-01 -1.94964290e-01 -3.67550671e-01 -6.95291877e-01
-1.18049204e+00 6.87954545e-01 2.39717634e-03 7.15720057e-01
-8.61662686e-01 -8.69144082e-01 -8.18051517e-01 -1.09966025e-02
6.87359422e-02 -8.03775907e-01 5.23199201e-01 -8.05368781e-01
-1.30026245e+00 9.52710688e-01 -1.83458135e-01 -1.98266283e-01
1.50795937e-01 -1.00437783e-01 -7.66581774e-01 9.51658264e-02
5.00299186e-02 1.47905812e-01 1.15860748e+00 -1.85180140e+00
-6.25233054e-01 -4.08515722e-01 -6.79325238e-02 3.99146795e-01
2.52229691e-01 2.68505931e-01 -4.11145061e-01 -9.63525355e-01
5.07619262e-01 -3.98008138e-01 -4.00501937e-01 -3.10278445e-01
-2.97560751e-01 5.82737982e-01 6.81818068e-01 -1.15133059e+00
1.02223110e+00 -2.11225009e+00 8.87005925e-02 -2.56373674e-01
3.61583531e-01 3.30376834e-01 -2.14200050e-01 2.46627405e-02
-3.48509371e-01 -2.55282596e-03 -6.22348726e-01 -5.19972146e-02
-4.27005857e-01 1.95766807e-01 -6.11322045e-01 6.05488241e-01
4.91134776e-03 6.75081074e-01 -7.52034903e-01 -7.39295781e-02
4.27555919e-01 1.01097977e+00 -1.62662879e-01 3.16889286e-01
7.16417357e-02 6.49599016e-01 -1.05527125e-01 9.57476377e-01
1.54306519e+00 -1.07287958e-01 -2.10073739e-01 -5.37709832e-01
-3.22341323e-01 -3.00244808e-01 -1.25068450e+00 1.27587616e+00
-4.89719719e-01 3.32006991e-01 4.51521158e-01 -5.89094460e-01
9.83644187e-01 3.67960557e-02 2.20883355e-01 -1.25006723e+00
-3.09490979e-01 2.43675821e-02 -4.26479220e-01 -3.33959967e-01
7.39715636e-01 -4.26686734e-01 2.81644106e-01 3.34550619e-01
-4.02132690e-01 1.61778912e-01 3.74214351e-02 2.24557012e-01
8.16981077e-01 7.83724636e-02 6.35469481e-02 -2.89118946e-01
6.30459845e-01 -1.36395052e-01 6.03503406e-01 1.10829318e+00
-2.69128770e-01 1.23920023e+00 -2.89810132e-02 -4.64914322e-01
-1.05601954e+00 -1.34257150e+00 -2.99804777e-01 1.00179577e+00
5.41663468e-01 -5.97052537e-02 -6.98061883e-01 1.44399121e-01
-1.89308658e-01 4.22749341e-01 -6.28610075e-01 -7.29200318e-02
-7.14901328e-01 -1.62569177e+00 2.49855250e-01 2.50495344e-01
9.59046423e-01 -1.26659179e+00 -1.48728460e-01 5.38992835e-03
-5.08154333e-01 -1.19870090e+00 -1.02930382e-01 1.51186278e-02
-8.01652193e-01 -9.57500756e-01 -6.06223226e-01 -6.52321339e-01
6.55080497e-01 1.02527690e+00 1.45745122e+00 1.43789962e-01
-7.07085788e-01 7.85800368e-02 -4.22343671e-01 3.03808033e-01
-2.25802541e-01 -9.11705554e-01 -2.45799035e-01 2.68899411e-01
-2.08157703e-01 -8.51181746e-01 -8.95744562e-01 2.73666531e-01
-1.04833865e+00 1.76064909e-01 5.48312902e-01 7.99906552e-01
1.09924948e+00 5.43744147e-01 3.95059079e-01 -8.84016037e-01
6.49793029e-01 -3.10091376e-01 -5.20100951e-01 6.51691109e-02
-3.49433571e-01 -4.67355937e-01 6.22968495e-01 -1.33611888e-01
-1.81984484e+00 -4.84783202e-01 -1.93267539e-01 -1.18203245e-01
-3.73775572e-01 2.09726959e-01 -4.13371295e-01 -4.85751212e-01
7.16807604e-01 5.19166172e-01 -3.05969566e-01 -6.97832525e-01
5.82375228e-01 4.24658865e-01 1.14788008e+00 -4.08394605e-01
8.77164006e-01 9.63037968e-01 -1.46869019e-01 -1.05862832e+00
-1.12105370e+00 -3.82991940e-01 -5.30537963e-01 -1.14748925e-01
9.16261971e-01 -1.17831826e+00 -4.24776137e-01 8.60500693e-01
-7.46364474e-01 -5.95548093e-01 -2.85108954e-01 -6.52596503e-02
-3.38931769e-01 5.26142836e-01 -8.97051096e-01 -7.79664993e-01
-5.11356175e-01 -9.04216230e-01 1.05427325e+00 4.20481026e-01
6.89978063e-01 -6.77190065e-01 -3.08040027e-02 5.34002662e-01
8.56329918e-01 3.68219465e-01 8.68143976e-01 6.05493784e-01
-9.76720631e-01 2.29058877e-01 -1.03640747e+00 5.39273381e-01
3.54712337e-01 -1.11889899e-01 -1.09372175e+00 -4.56966072e-01
2.11563706e-01 -1.08509809e-01 1.39367557e+00 7.63677061e-01
1.13096082e+00 4.41046804e-03 1.34365231e-01 1.16989410e+00
1.81694245e+00 -2.07822964e-01 1.42407131e+00 5.66160500e-01
9.90104616e-01 3.68160486e-01 8.61937344e-01 3.26393604e-01
2.61016846e-01 4.66625333e-01 6.14216924e-01 -7.93987632e-01
-7.63503373e-01 1.76198214e-01 3.55213910e-01 5.05212724e-01
-2.83360481e-01 -1.31949723e-01 -4.87182111e-01 3.58714759e-01
-1.86678004e+00 -1.15248740e+00 -3.92927200e-01 2.23683429e+00
7.00625002e-01 -2.87704259e-01 -5.72896361e-01 -1.36577217e-02
6.81690931e-01 6.80581331e-01 -3.17043871e-01 -2.24009931e-01
-9.10313606e-01 8.71503502e-02 4.23219532e-01 7.71018863e-01
-1.14880323e+00 9.95843053e-01 6.11368847e+00 9.11588252e-01
-7.97232032e-01 1.32588446e-01 8.63061368e-01 2.84803867e-01
-3.38666737e-01 -1.73662230e-02 -7.96615660e-01 4.74778175e-01
5.94041049e-01 3.42849076e-01 7.75387406e-01 2.12876741e-02
6.15690649e-01 -3.53390664e-01 -2.08236977e-01 1.07882202e+00
3.32201958e-01 -1.21347141e+00 3.73629153e-01 -1.91407010e-01
9.47225690e-01 5.94579130e-02 7.97613785e-02 4.21031788e-02
2.28827298e-01 -1.22938037e+00 2.56640762e-01 9.46349084e-01
7.93305874e-01 -7.79215753e-01 4.88920718e-01 2.26164907e-01
-1.28745496e+00 5.83619997e-02 -7.11817384e-01 1.57881960e-01
2.21151546e-01 9.61694360e-01 2.25633278e-01 7.70469308e-01
1.18726790e+00 6.61818981e-01 -5.15167356e-01 7.56018281e-01
-2.62959778e-01 4.00663972e-01 4.08857800e-02 1.13339198e+00
-8.00589249e-02 -7.29925573e-01 6.69664979e-01 1.28435779e+00
2.06915766e-01 5.08487761e-01 2.92093873e-01 8.35127294e-01
1.22165985e-01 -2.24492759e-01 -3.14486891e-01 5.49203455e-01
1.39375664e-02 1.56239820e+00 -6.45243108e-01 -2.44374186e-01
-3.01564276e-01 1.25609887e+00 6.11860640e-02 8.43185246e-01
-8.46288919e-01 -1.62209705e-01 1.02834272e+00 2.18860731e-01
2.85590410e-01 -3.79529521e-02 -3.98331136e-01 -1.33620775e+00
1.01087108e-01 -1.01426220e+00 4.11292136e-01 -1.29545069e+00
-1.28582442e+00 7.24414349e-01 -3.60487670e-01 -9.54666257e-01
5.36406457e-01 -2.77778417e-01 -5.66668093e-01 1.40138149e+00
-2.28393626e+00 -1.31454909e+00 -1.02053726e+00 9.04409051e-01
4.83601958e-01 3.93740356e-01 6.27279401e-01 6.12913787e-01
-6.99074984e-01 -7.55892247e-02 3.50393474e-01 -3.02609622e-01
8.91106009e-01 -1.10835195e+00 2.32962698e-01 1.40554869e+00
-3.25655460e-01 4.90247607e-01 8.90368640e-01 -7.97831535e-01
-1.37349117e+00 -1.25942612e+00 5.18396199e-01 -8.13179761e-02
4.14163202e-01 -6.71193898e-02 -1.42543924e+00 4.51287061e-01
-8.53935629e-02 6.40283823e-01 1.94451645e-01 -1.83512911e-01
-4.95188326e-01 -4.99610603e-01 -1.28624189e+00 6.61542773e-01
9.52270329e-01 -6.16706431e-01 -1.69483051e-01 4.17180657e-01
7.81918883e-01 -5.16234875e-01 -7.58949995e-01 6.57031953e-01
2.50629812e-01 -1.45802808e+00 1.63665545e+00 -2.17680216e-01
5.47406554e-01 -8.83492470e-01 -5.63523769e-01 -1.27580559e+00
-8.40225220e-01 -4.36133474e-01 5.47505319e-02 1.41345310e+00
2.43753344e-02 -6.10995948e-01 2.07158253e-01 1.03246041e-01
-2.91415602e-01 -3.41943413e-01 -6.07049227e-01 -2.47620687e-01
-2.68903822e-01 -2.59749293e-01 4.29173142e-01 9.17308450e-01
-9.57049012e-01 8.52939487e-02 -8.68438721e-01 7.63773680e-01
1.24144733e+00 5.85144043e-01 5.75264633e-01 -9.72882986e-01
-2.05782413e-01 -1.61041275e-01 -5.34019396e-02 -6.94648027e-01
-2.18633816e-01 -5.40426373e-01 3.17243367e-01 -1.91316986e+00
6.15264237e-01 -1.94905162e-01 -4.80720758e-01 1.35566026e-01
-5.41863382e-01 9.22219157e-01 -1.50381327e-01 3.16763431e-01
-6.02857709e-01 4.78167564e-01 1.41277397e+00 7.88146555e-02
-1.33041024e-01 -2.14482412e-01 -1.11501181e+00 8.38274240e-01
6.00822270e-01 -4.55660112e-02 -1.10392183e-01 -4.61660832e-01
8.97781327e-02 7.49570727e-02 7.84414530e-01 -7.08198488e-01
9.20192525e-02 -2.86895394e-01 5.64738512e-01 -5.85095763e-01
4.11835492e-01 -4.09255743e-01 1.85926661e-01 -7.02336207e-02
9.77214053e-02 -3.66011977e-01 1.61214650e-01 9.11709428e-01
-4.49515641e-01 4.36590075e-01 1.46927321e+00 -3.22904706e-01
-1.01303458e+00 4.39291745e-01 -2.43154258e-01 -1.12403944e-01
6.23434365e-01 -3.05032551e-01 -9.48768437e-01 -2.03767940e-01
-7.28598475e-01 -1.26371579e-02 5.38114905e-01 1.54509500e-01
9.35169935e-01 -1.01648295e+00 -1.10780001e+00 4.06084925e-01
1.93911195e-02 -1.04796417e-01 1.13409805e+00 9.47852135e-01
-5.54027379e-01 -1.34343326e-01 -4.08862919e-01 -4.79658723e-01
-1.50214958e+00 3.89193565e-01 5.33843696e-01 -3.07703435e-01
-1.39798176e+00 7.86654830e-01 6.87067628e-01 -3.34674925e-01
9.74357035e-03 -8.11123475e-02 -3.69290084e-01 -2.13321254e-01
1.19936585e+00 4.99571443e-01 1.64383739e-01 -8.49607527e-01
-1.32562250e-01 9.01477516e-01 3.16193290e-02 2.93195873e-01
1.49674249e+00 -7.52229512e-01 -7.61153698e-01 1.93023637e-01
5.46257973e-01 2.08258983e-02 -1.34653080e+00 -4.46971804e-01
-6.07654274e-01 -1.10296643e+00 4.02176648e-01 -9.24519122e-01
-1.24195504e+00 6.34893656e-01 8.00329149e-01 1.82961702e-01
1.66933978e+00 -2.55699754e-01 6.80580258e-01 -1.03574902e-01
1.80504560e-01 -7.27219522e-01 1.33161739e-01 5.82575798e-01
1.02046978e+00 -1.36705124e+00 1.82580858e-01 -6.29570127e-01
-3.42556626e-01 8.96343112e-01 5.90716660e-01 -1.05586872e-01
5.25108516e-01 4.45137173e-01 2.01129124e-01 -2.67543823e-01
-4.50379342e-01 -4.92759258e-01 7.24891946e-03 8.54030609e-01
4.59431589e-01 -1.96293765e-03 4.59117591e-02 4.42905188e-01
2.94318825e-01 -2.16427937e-01 6.79065943e-01 5.20970106e-01
-7.51118958e-01 -6.41478479e-01 -9.47772324e-01 2.78015882e-01
-5.00150681e-01 -5.74409306e-01 1.74524516e-01 2.88668007e-01
5.34022935e-02 1.14350986e+00 -3.26772258e-02 2.08285023e-02
5.00958562e-01 -6.11319005e-01 3.92489582e-01 -3.46439749e-01
-3.22220951e-01 2.21360520e-01 1.42391315e-02 -8.00095737e-01
-7.27263093e-01 -1.83846101e-01 -9.16141510e-01 -5.65143585e-01
-5.40859029e-02 -2.51339704e-01 3.63309622e-01 5.99363089e-01
3.24688047e-01 9.73170400e-01 5.51418364e-01 -1.20005870e+00
-1.16365328e-01 -8.71518195e-01 -1.18531477e+00 5.00425935e-01
7.57841289e-01 -2.84614325e-01 -5.79600394e-01 8.12624916e-02] | [11.029095649719238, -2.3310329914093018] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.